Advances in
CANCER RESEARCH Volume 83
This Page Intentionally Left Blank
Advances in
CANCER RESEARCH Volume 83
E...
74 downloads
784 Views
2MB Size
Report
This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!
Report copyright / DMCA form
Advances in
CANCER RESEARCH Volume 83
This Page Intentionally Left Blank
Advances in
CANCER RESEARCH Volume 83
Edited by
George F. Vande Woude Van Andel Research Institute Grand Rapids, Michigan
George Klein Microbiology and Tumor Biology Center Karolinska Institute Stockholm, Sweden
San Diego
San Francisco New York Boston London Sydney Tokyo
∞ This book is printed on acid-free paper.
C 2001 by ACADEMIC PRESS Copyright
All Rights Reserved. No part of this publication may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopy, recording, or any information storage and retrieval system, without permission in writing from the Publisher. The appearance of the code at the bottom of the first page of a chapter in this book indicates the Publisher’s consent that copies of the chapter may be made for personal or internal use of specific clients. This consent is given on the condition, however, that the copier pay the stated per copy fee through the Copyright Clearance Center, Inc. (222 Rosewood Drive, Danvers, Massachusetts 01923), for copying beyond that permitted by Sections 107 or 108 of the U.S. Copyright Law. This consent does not extend to other kinds of copying, such as copying for general distribution, for advertising or promotional purposes, for creating new collective works, or for resale. Copy fees for pre-2001 chapters are as shown on the title pages. If no fee code appears on the title page, the copy fee is the same as for current chapters. 0065-230X/2001 $35.00 Explicit permission from Academic Press is not required to reproduce a maximum of two figures or tables from an Academic Press chapter in another scientific or research publication provided that the material has not been credited to another source and that full credit to the Academic Press chapter is given.
Academic Press A Harcourt Science and Technology Company 525 B Street, Suite 1900, San Diego, California 92101-4495, USA http://www.academicpress.com
Academic Press Harcourt Place, 32 Jamestown Road, London NW1 7BY, UK http://www.academicpress.com International Standard Book Number: 0-12-006683-1 PRINTED IN THE UNITED STATES OF AMERICA 01 02 03 04 05 06 QW 9 8 7 6 5 4
3
2
1
Contents
Contributors to Volume 83 ix
Suppressor and Oncogenic Roles of Transforming Growth Factor-β and Its Signaling Pathways in Tumorigenesis Ester Piek and Anita B. Roberts I. II. III. IV. V.
Introduction 2 The TGF-β Signaling Pathway 3 Bimodal Action of TGF-β in Tumorigenesis 14 Activities of TGF-β Important for Oncogenesis 18 Dysregulated Expression or Activity of Components of TGF-β Signaling Pathways in Oncogenesis 33 VI. Summary 41 References 42
Hereditary Diffuse Gastric Cancer Anita Dunbier and Parry Guilford I. II. III. IV. V. VI. VII. VIII.
Introduction 55 Hereditary Diffuse Gastric Cancer 56 Mutations in CDH-1 58 The Tumor Spectrum of HDGC 60 Inactivation of the Second CDH-1 Allele 60 Molecular Mechanism of HDGC Susceptibility 61 Clinical Criteria and Management of HDGC 62 Conclusion 63 References 63
Role of Heparan Sulfate Proteoglycans in Cell Signaling and Cancer Erica M. Selva and Norbert Perrimon I. Introduction 67 II. HSPGs and Cancer 69
v
vi III. IV. V. VI. VII. VIII.
Contents
FGF Signaling and HSPGs in Drosophila 71 Dpp Signaling and HSPGs in Drosophila 73 Wg and Hh Signaling and HSPGs in Drosophila 73 The Role of Glypicans in Wg Signaling 75 HSPGs Are Involved in Hh Movement 77 Conclusion 78 References 79
The Occurrence and Significance of V Gene Mutations in B Cell–Derived Human Malignancy Freda K. Stevenson, Surinder S. Sahota, Christian H. Ottensmeier, Delin Zhu, Francesco Forconi, and Terry J. Hamblin I. II. III. IV. V. VI. VII. VIII. IX.
Introduction 82 Immunoglobulin Genes in Normal B Cell Development 83 Immunoglobulin Genes in B Cell Tumors 90 Somatic Mutation in B Cell Tumors 92 Chronic Lymphocytic Leukemia 95 Follicular Lymphoma 100 Diffuse Large B Cell Lymphoma 102 Plasma Cell Tumors 104 Conclusion 109 References 110
MHC Antigens and Tumor Escape from Immune Surveillance Federico Garrido and Ignacio Algarra I. II. III. IV. V. VI. VII. VIII. IX.
Introduction 117 HLA Class I Antigen Expression in Primary Tumors 119 Changes in MHC Class I Antigen Expression during Metastatic Colonization 139 T Cell Immunoselection of MHC Class I–Negative Tumor Clones 141 Expression of Nonclassical HLA Class I Molecules in Tumors 143 Tumor NK Escape Mechanisms 145 HLA Class I Loss and T Cell–Based Immunotherapy 146 HLA Class II Antigens in Tumors 147 Conclusions 149 References 151
The Role of Selection in Progressive Neoplastic Transformation Harry Rubin I. Introduction 160 II. Spontaneous Neoplastic Transformation in Cell Culture 162 III. Selection in Transformation of Established Cell Lines 168
Contents
vii
IV. The Contribution of Apoptosis to Selection in Neoplastic Transformation 184 V. Inhibition of Growth of Transformed Cells by Surrounding Nontransformed Cells 187 VI. Confounding Effects of Variable Cell Behavior on the Dynamics of Transformation 188 VII. Summary of the Major Features of Spontaneous Transformation 188 VIII. Evidence for Selection in Experimental and Human Cancer 190 IX. Sources of Genetic Variation for Possible Selection in Tumor Development 195 X. The Nature of Selection in Vivo 196 XI. Selection in Carcinogenesis by Polycyclic Aromatic Hydrocarbons 199 XII. Conclusions 201 References 202
ATM: Genome Stability, Neuronal Development, and Cancer Cross Paths Yosef Shiloh and Michael B. Kastan I. II. III. IV. V. VI. VII.
Introduction 210 Ataxia-Telangiectasia: A Disease Caused by ATM Deficiency 211 The ATM Gene and Its Mutations 216 The ATM Protein: From Sequence to Function 218 ATM Functions: Lessons from Knockout Mice 231 ATM Deficiency Leads to Increased Oxidative Stress 238 ATM: Interplay with Signaling Pathways Associated with Growth and Differentiation 240 VIII. Defects in DNA Damage Response and Cancer Predisposition 241 IX. Conclusions 243 References 244
Index 255
This Page Intentionally Left Blank
Contributors
Numbers in parentheses indicate the page on which the authors’ contributions begin.
Ignacio Algarra, Departamento de Ciencias de la Salud, Universidad de Jaen, Jaen, Spain (117) Anita Dunbier, Cancer Genetics Laboratory, Department of Biochemistry, University of Otago, Dunedin, New Zealand (55) Francesco Forconi, Molecular Immunology Group, Tenovous Laboratory, Southampton University Hospitals Trust, Southampton SO16 6YD, United Kingdom (81) Federico Garrido, Departmento de Analisis Clinicos, Hospital Universitario Virgen de las Nieves, 18014 Granada (117) Parry Guilford, Cancer Genetics Laboratory, Department of Biochemistry, University of Otago, Dunedin, New Zealand (55) Terry J. Hamblin, Molecular Immunology Group, Tenovus Laboratory, Southampton University Hospitals Trust, Southampton SO16 6YD, United Kingdom (81) Michael B. Kastan, Department of Hematology-Oncology, St. Jude Children’s Research Hospital, Memphis, Tennessee 38105 (209) Christian H. Ottensmeier, Molecular Immunology Group, Tenovus Laboratory, Southampton University Hospitals Trust, Southampton SO16 6YD, United Kingdom (81) Norbert Perrimon, Department of Genetics and Howard Hughes Medical Institute, Harvard Medical School, Boston, Massachusetts 02115 (67) Ester Piek, Laboratory of Cell Regulation and Carcinogenesis, National Cancer Institute, Bethesda, Maryland 20892 (1) Anita B. Roberts, Laboratory of Cell Regulation and Carcinogenesis, National Cancer Institute, Bethesda, Maryland 20892 (1) Harry Rubin, Department of Molecular and Cell Biology and Virus Laboratory, Life Sciences Addition, University of California at Berkeley, Berkeley, California 94720 (159) Surinder S. Sahota, Molecular Immunology Group, Tenovus Laboratory, Southampton University Hospitals Trust, Southampton SO16 6YD, United Kingdom (81)
ix
x
Contributors
Erica M. Selva, Department of Genetics and Howard Hughes Medical Institute, Harvard Medical School, Boston, Massachusetts 02115 (67) Yosef Shiloh, Department of Human Genetics and Molecular Medicine, Sackler School of Medicine, Tel Aviv University, Tel Aviv 69978, Israel (209) Freda K. Stevenson, Molecular Immunology Group, Tenovus Laboratory, Southampton University Hospitals Trust, Southampton SO16 6YD, United Kingdom (81) Delin Zhu, Molecular Immunology Group, Tenovus Laboratory, Southampton University Hospitals Trust, Southampton SO16 6YD, United Kingdom (81)
Suppressor and Oncogenic Roles of Transforming Growth Factor-β and Its Signaling Pathways in Tumorigenesis Ester Piek and Anita B. Roberts Laboratory of Cell Regulation and Carcinogenesis National Cancer Institute Bethesda, MD 20892-8395
I. Introduction II. The TGF-β Signaling Pathway A. TGF-β Ligands B. TGF-β Receptors C. Downstream Signaling Pathways III. Bimodal Action of TGF-β in Tumorigenesis A. Tumor-Suppressor Activities of TGF-β B. Tumor-Promoting Activities of TGF-β IV. Activities of TGF-β Important for Oncogenesis A. TGF-β Isoform-Specific Activities during Tumorigenesis B. Increased Activation of Latent TGF-β Associated with Tumorigenesis C. Effects of TGF-β on Epithelial–Mesenchymal Transdifferentiation D. Effects of TGF-β on Genomic Instability E. Effects of TGF-β on Invasion and Metastasis F. Indirect Effects of TGF-β on Tumorigenesis V. Dysregulated Expression or Activity of Components of TGF-β Signaling Pathways in Oncogenesis A. Receptors B. Functional Implications of Smad Mutations Identified in Tumors C. Alterations in Smad-Interacting Proteins D. Regulation of TGF-β Signal Transduction Pathways by Oncogenes VI. Summary References
Transforming growth factor-β (TGF-β) has been implicated in oncogenesis since the time of its discovery almost 20 years ago. The complex, multifunctional activities of TGF-β endow it with both tumor suppressor and tumor promoting activities, depending on the stage of carcinogenesis and the responsivity of the tumor cell. Dysregulation or alteration of TGF-β signaling in tumorigenesis can occur at many different levels, including activation of the ligand, mutation or transcriptional suppression of the
Advances in CANCER RESEARCH 0065-230X/01 $35.00
1
C 2001 by Academic Press. Copyright All rights of reproduction in any form reserved.
2
Ester Piek and Anita B. Roberts
receptors, or alteration of downstream signal transduction pathways resulting from mutation or changes in expression patterns of signaling intermediates or from changes in expression of other proteins which modulate signaling. New insights into signaling from the TGF-β receptors, including the identification of Smad signaling pathways and their interaction with mitogen-activated protein (MAP) kinase pathways, are providing an understanding of the changes involved in the change from tumor suppressor to tumor promoting activities of TGF-β. It is now appreciated that loss of sensitivity to inhibition of growth by TGF-β by most tumor cells is not synonymous with complete loss of TGF-β signaling but rather suggests that tumor cells gain advantage by selective inactivation of the tumor suppressor activities of TGF-β with retention of its tumor promoting activities, especially those dependent on cross talk with MAP kinase pathways and AP-1. C 2001 Academic Press.
I. INTRODUCTION Growth factor receptors have been likened to switches on the cell surface that activate complex circuitry within the cell which impinges, ultimately, on the nucleus to effect gene transcription. The rapidly expanding knowledge of the complexity of cellular signaling networks consisting of dynamic cross talk between various signaling pathways makes it clear that any extracellular signal or growth factor can have a defined function only in the theoretical state of “hard wiring” or, experimentally, only under defined conditions of a particular cell type in culture. Given the pleiotropic nature of cells and the plasticity of the cellular phenotype as a cell progresses through the stages of oncogenesis from a nonneoplastic to a fully malignant, invasive tumorigenic cell, it follows that there must be a corresponding plasticity of the wiring of the signal transduction pathways. Transforming growth factor-β (TGF-β) is arguably the paradigmatic multifunctional growth factor. It is the prototypic member of a large superfamily of structurally and functionally related cytokines, which play key roles in embryonic development, normal physiology, and disease pathogenesis (Blobe et al., 2000; Kingsley, 1994; Massague et al., 2000; Roberts and Sporn, 1990). Prominent members of the family include the activins, inhibin, bone ¨ morphogenetic proteins (BMPs), and Mullerian inhibitory substance. Since most cells can express both TGF-β and its receptors, any understanding of its roles in oncogenesis must include not only effects on the tumor cell but also effects on stromal elements, which contribute to angiogenesis, suppression of immune surveillance, and desmoplasia (Wakefield et al., 2001; Akhurst and Balmain, 1999; Gold, 1999; Reiss, 1999). Moreover, since many tumor cells secrete increased levels of the active TGF-β ligand compared to their nontransformed counterparts, these effects include both autocrine effects on the tumor cell and autocrine and paracrine effects on stromal components (Fig. 1). In this review, we discuss recent developments concerning the
Suppressor and Oncogenic Roles of TGF-β
3
Fig. 1 Models of the dual tumor suppressor (dashed lines) and prooncogenic activities (solid lines) of the TGF-β signaling pathways. (A) Complete loss of TGF-β receptor function in neoplastic cells precludes any direct effects of TGF-β on the transformed cells and limits the prooncogenic activities of TGF-β to indirect effects on stromal elements. (B) Partial loss of receptor function or signaling activity allows for direct prooncogenic effects of TGF-β on tumor cells, such as those which mediate epithelial-to-mesenchymal transformation (EMT), invasion, and metastasis.
pleiotropic effects of TGF-β on tumor cells and stroma during oncogenesis and put these into the context of its dual tumor suppressor and oncogenic activities. Whenever possible, we provide mechanistic insights based on the recently described signaling pathways which mediate these effects of TGF-β. Where appropriate, we cite reviews in lieu of the many primary reports.
II. THE TGF-β SIGNALING PATHWAY TGF-β signaling can be regulated at any of many levels ranging from activation or sequestration of the ligand to alteration of the expression of the signaling receptor complex or receptor-modifying proteins including endoglin and betaglycan, modulation of the signal transduction pathways either directly or indirectly via regulation of Smad-interacting proteins, control of proteasomal degradation of signaling components, or cross talk with other receptor pathways. Cells typically undergo changes in their responses to TGF-β, coincident with changes in their phenotype, as they progress from nonneoplastic, growth-regulated cells to fully malignant, proliferative, and invasive tumor cells (see Section V). Dysregulation of TGF-β signaling associated with these changes can be found at any of these levels, although the most common alterations are found in the transcriptional regulation of receptor expression.
4
Ester Piek and Anita B. Roberts
A. TGF-β Ligands Mammals express three highly homologous isoforms of TGF-β—TGF-β 1, TGF-β 2, and TGF-β 3, localized to human chromosomes 19q13, 1q41, and 14q24, respectively (Roberts and Sporn, 1990). Expression of these isoforms is under the control of distinct promoters: The TGF-β 2 and TGF-β 3 promoters have hormone-responsive CREB binding sites and TATA boxes, whereas the TGF-β 1 promoter is TATA-less and is characterized instead by a myriad of response elements including those regulated by immediate early response genes, oncogenes, and the retinoblastoma gene product (Roberts and Sporn, 1992). TGF-β 1 is the most abundant isoform in most cells and tissues and, consistent with the complex regulation of its promoter, it is the isoform most often dysregulated in disease pathogenesis, including tumorigenesis. Each of these ligands is secreted from cells in a latent form in which the N-terminal domain of the TGF-β transcript, called the latency-associated protein (LAP), is noncovalently associated with the mature, C-terminal domain of the protein (Taipale et al., 1998). Latent TGF-β is unable to bind or activate its signaling receptors. In certain cases, LAP is covalently linked via disulfide bonds to any member of a family of larger fibrillin-like glycoproteins called latent TGF-β binding proteins (LTBPs), forming a large latent complex which preferentially localizes to matrix, depending on crosslinking by transglutaminase. The processes governing activation of these latent forms of TGF-β are understandably diverse and likely play a very important role in not only normal physiology but also disease pathogenesis (Fig. 2). Mechanisms of activation include deglycosylation of LAP, exposure of the latent complex to reactive oxygen or acidic microenvironments, and proteolysis of the latency proteins mediated by specific proteases localized in the extracellular matrix or targeted to the cell surface by the mannose 6-phosphate/insulin-like growth factor II receptor (Taipale et al., 1998). Most prominent in proteolytic activation of latent TGF-β is the system involving urokinase plasminogen activator (uPA) and its cell surface receptor (uPAR), which together with plasminogen activator inhibitor-1 (PAI-1) regulate the conversion of plasminogen into plasmin, a protease involved in degradation of extracellular matrix. Expression of all these genes is controlled either directly or indirectly by TGF-β, thus establishing feedback mechanisms for regulation of its bioactivation (Rifkin et al., 1997). Other mechanisms include activation by conformational changes in LAP, such as that mediated by binding of the latent complex to thrombospondin (TSP-1), a large homotrimeric protein secreted by many cell types including tumor cells (Murphy-Ullrich and Poczatek, 2000; Tuszynski and Nicosia, 1996). Again, many feedback loops exist since TSP-1 is also an important regulator of the
Suppressor and Oncogenic Roles of TGF-β
5
Fig. 2 TGF-β is secreted in a latent form involving noncovalent association of the N-terminal portion of pro-TGF-β, called LAP (latency-associated protein), with the mature, bioactive C-terminal fragment. Activation of latent TGF-β is highly regulated and can be achieved by many different mechanisms, several of which are dysregulated in tumorigenesis and others which mediate effects of chemopreventive agents (see Sections II.A and IV.B).
plasminogen–plasmin system (Albo et al., 2000) and since TGF-β in certain cells can stimulate induction of TSP-1 (Majack et al., 1990) (see Section IV.B). TSP-1-dependent activation of TGF-β may be especially important in normal physiology since TSP-1 null and TGF-β1 null mice phenocopy each other (Crawford et al., 1998). Another mechanism thought to result in localized activation of TGF-β on the cell surface involves binding of LAP to cell surface α vβ 6 integrin (Munger et al., 1999). The importance of this mechanism has been demonstrated in a model of bleomycin-induced pulmonary fibrosis known to be dependent, in part, on TGF-β because mice null for the epithelial-restricted integrin β 6 are protected from fibrosis coincident with a repression of induction of a large cohort of ECM genes dependent on TGF-β (Kaminski et al., 2000). Still other mechanisms of activation of latent TGF-β result from binding of the latent complex to IgG, which is especially important in autoimmune disease (Letterio and Roberts, 1998). Clearly, these many modes of activation of latent TGF-β provide the cell with fine-tuned control over the process and suggest that it might be an important focus of dysregulation in disease (see Section IV.B).
6
Ester Piek and Anita B. Roberts
B. TGF-β Receptors TGF-βs bind and activate a heteromeric complex of related transmembrane receptors with intrinsic cytoplasmic serine/threonine kinase domains (Derynck and Feng, 1997; Massague, 1998; Piek et al., 1999a). TGF-β binds the type II receptor (TβRII), which in turn recruits a TGF-β type I receptor, principally TβRI /ALK5 (activin-like kinase) or alternatively ALK1 in vascular endothelial cells (Oh et al., 2000), to form a heterotetrameric receptor complex. Other members of the TGF-β superfamily interact with different combinations of homologous type I and II receptor serine/threonine kinases. The ligand-binding type II receptor kinase is constitutively active and phosphorylates type I receptors on serine and threonine residues in the GS box, a conserved stretch of glycine and serine residues preceding the receptor kinase domain. This phosphorylation event activates the type I receptor kinase and downstream signaling (Massague, 1998). Two other cell surface TGF-β binding proteins also play prominent roles in modulating the interaction of TGF-β with its signaling receptor complex. Betaglycan, also called the TGF-β type III receptor or TβRIII, binds all isoforms of TGF-β and is of particular importance in “presenting” TGF-β 2 to the receptor complex since this isoform is uniquely unable to bind to TβRII in its absence (Massague, 1998). Another cell surface binding protein, endoglin (also called CD105), has limited homology to betaglycan in its N-terminal domain and binds only TGF-β 1 and TGF-β 3 (Barbara et al., 1999). In contrast to betaglycan, it is preferentially expressed on proliferating vascular endothelial cells. Comparison of the effects of overexpression of these proteins shows that endoglin decreases TGF-β responses, whereas betaglycan enhances responsiveness to TGF-β and correlates with increased binding of ligand to TβRII (Letamendia et al., 1998). Endoglin, in cooperation with ALK1, might control TGF-β responsiveness in vascular endothelial cells (see Section IV.F.1). Examples of dysregulated expression or activity of each of these receptor types in tumor cells are discussed in Section V.A.
C. Downstream Signaling Pathways The identification of specific signaling pathways downstream of TGF-β receptors provides an opportunity to understand how changes in the balance and the “circuitry” of signal transduction pathways might underlie changes in cellular behavior which contribute to the tumor suppressor and oncogenic activities of TGF-β. There is general agreement that the majority of target genes of TGF-β are regulated by a recently identified family of intracellular signaling molecules called Smad proteins and that a smaller percentage of
Suppressor and Oncogenic Roles of TGF-β
7
TGF-β end points might result from activation of any of several mitogenactivated protein (MAP) kinase pathways or possibly by interaction between these two pathways. Other pathways might be important in specific cellular contexts.
1. THE Smad SIGNALING PATHWAY Originally discovered as downstream mediators of signals from the TGF-β superfamily receptor serine/threonine kinases as a result of genetic screens in Drosophila and Caenorhabditis elegans, the Smad proteins are now known to function as signal transducers of TGF-β superfamily members in both invertebrates and vertebrates, including mammals. These proteins mediate a direct pathway to transduce TGF-β signals from the receptors to nuclear target genes (Fig. 3A). Many excellent reviews of Smad signaling have been published, so we outline only the features important to our discussion of the roles of these proteins in mediating the complex effects of TGF-β in oncogenesis (de Caestecker et al., 2000a; Massague et al., 2000; Piek et al., 1999a). The family of mammalian Smad proteins includes eight proteins which are categorized into three functional classes: receptor-activated Smads (R-Smads), co-Smads, and inhibitory Smads (Fig. 3B). The N-terminal and C-terminal domains of R-Smads are relatively highly conserved and are called MH1 and MH2, respectively, based on the homology of these domains to that of the Drosophila MAD protein. For R-Smads and Smad4, the MH1 domain harbors the DNA binding activity and the MH2 domain the transcriptional activating activity and the protein–protein binding activity important in formation of homomeric and heteromeric Smad complexes. R-Smads are “pathway restricted” in that their activation is specified only by certain receptors. Thus, in the broadest sense, Smads 2 and 3 are activated by TGF-β and activin receptors and Smads 1, 5, and 8 by BMP receptors, although exceptions to this rule have been reported (Liu et al., 1998; Oh et al., 2000; Yue et al., 1999). Whereas the receptor specificity lies, in part, in a domain called the L3 loop (Chen et al., 1998), each of these R-Smads shares a critˇ Sˇ motif at its extreme C terminus, where the two marked serines ical SSX are phosphorylated directly by the type I receptor kinase, typically within 15 min of treatment of cells with ligand. R-Smads are not catalytic in that they have no identified enzymatic activity that could amplify the signal. Rather, the activity of the Smad signaling pathway in particular cells is controlled, in part, by the absolute amount and the relative concentrations of the various R-Smads (see Section V.B). Although Smad2 and Smad3 are each activated by TGF-β receptors, recent data suggest that they have both distinct and overlapping effects on gene activation (Fig. 4). Smad2 has a unique 30-amino acid insertion in its MH1
8
Ester Piek and Anita B. Roberts
Fig. 3 The most prominent signaling pathway activated by the TGF-β receptors is that mediated by Smad proteins. (A) The basic scheme of Smad signaling involves phosphorylation of Smad2 or -3 by the activated TβRI kinase, formation of a heteromeric complex with Smad4, and translocation to the nucleus, where, together with transcription factors, coactivator and corepressor Smads regulate the transcriptional activity of target genes. Smad7 can block signal transduction through this pathway either by binding to an R-Smad and preventing its partnering with Smad4 or by binding to TβRI, thereby blocking phosphorylation of the R-Smad. (B) Smads can be categorized into three classes. The conserved MH1 domains of R-Smads and Smad4 mediate their DNA binding, whereas the MH2 domains harbor the transcriptional activating activity and are the site of protein–protein interactions, including (for R-Smads) interaction with TβRI. Smad4 and inhibitory Smads lack the C-terminal phosphorylation (SSXS) motif found in R-Smads. PKC (䊉) and MAP kinase (∗) sites are indicated in the MH1 and middle linker domains, respectively. The SAD domain in Smad4 is required for its transcriptional activating activity.
Suppressor and Oncogenic Roles of TGF-β
9
Fig. 4 The DNA binding and transcription factor binding of Smad2 and Smad3 can be modulated both by alternative splicing of Smad2 resulting in loss of a 30-amino acid insertion in its MH1 domain and by phosphorylation of Smads 2 and 3 in their MH1 domain (asterisks).
domain which precludes its binding to DNA. An alternatively spliced form of Smad2 missing this inserted exon can bind DNA and activate transcription similar to Smad3 (Yagi et al., 1999). Also contributing to altered patterns of gene activation by these two R-Smads is the recently described protein kinase C (PKC)-dependent phosphorylation of key serine residues in the MH1 domain which abrogates the DNA binding of Smad3, thus shifting its pattern of gene activation to mimic that of Smad2 (Yakymovych et al., 2001). The latter may contribute to mechanisms of tumor promotion by agents such as phorbol esters which activate PKC. For R-Smads, phosphorylation by the type I receptor serves to relieve autoinhibitory interactions between the MH1 and MH2 domains, allowing these proteins to multimerize and to partner with the co-Smad, Smad4, first identified in the context of its putative tumor suppressor activity and called deleted in pancreatic carcinoma locus 4 (Hahn et al., 1996). Smad4 is distinguished from R-Smads by the absence of the C-terminal phosphorylation motif and by the fact that it does not interact with the TGF-β superfamily receptors and requires a unique domain in the middle linker region called the Smad activation domain for its transcriptional activating activity (de Caestecker et al., 2000b) (Fig. 3B). Although two highly related coSmads, Smad4 and Smad4β, are found in Xenopus (Howell et al., 1999), only Smad4 has been found in mammals. Since Smad4 is an obligate partner for R-Smads in control of transcriptional activation of target genes, it serves as a convergent node for signaling from all TGF-β superfamily receptors, suggesting that competition of activated R-Smads for partnering with Smad4 might contribute to the regulation of target gene expression in particular cells (Candia et al., 1997). Moreover, given the central position of
10
Ester Piek and Anita B. Roberts
Smad4 in the signal transduction pathway, it is expected that mutation of Smad4, as found in particular human cancers, can affect multiple pathways (S. Zhou et al., 1999). Also important in control of signaling through Smad pathways are the inhibitory Smads, Smad6 and Smad7 (Piek et al., 1999a). These two proteins lack both the MH1 domain and the C-terminal phosphorylation motif of the R-Smads, suggesting that they are not autorepressed but, rather, can interact with other Smad proteins and with the receptors (Fig. 3B). Although the specific mechanisms of action of these proteins are still not clear, Smad7 appears to be a nuclear protein which exits the nucleus in response to certain stimuli such as TGF-β signals (Itoh et al., 1998). Moreover, the regulation of Smad7 expression not only by TGF-β—in an autoinhibitory Smad-dependent feedback loop (Brodin et al., 2000; Hayashi et al., 1997; Imamura et al., 1997; Nakao et al., 1997)—but also by tumor necrosis factor-α (TNF-α) and interleukin-1β (IL-1β) through NF-κB binding sites in its promoter (Bitzer et al., 2000; Nagarajan et al., 2000) and by interferon-γ (IFN-γ ) and epidermal growth factor (EGF) through Stat1 binding sites (Ulloa et al., 1999), suggests that this molecule plays an important role in integration of signals from opposing pathways (Fig. 5; see color insert). Intuitively, Smad7 may also be found to be important in certain cancers, in which it might contribute to reduced sensitivity of tumor cells to TGF-β (see Section V.B). Whether Smad6, a nuclear protein like Smad7 (Imamura et al., 1997), implicated in inhibition of BMP signaling (Hata et al., 1998a), also plays a role in inhibition of TGF-β signaling is not clear. Ultimately, Smad signaling pathways converge on transcriptional complexes in the nucleus to regulate target gene expression (Fig. 3A). Again, this has been the subject of numerous reviews and only its rudimentary features are discussed (de Caestecker et al., 2000a; Derynck et al., 1998; Massague and Wotton, 2000). Since the Smad proteins bind DNA only weakly through Smad binding elements (SBEs) consisting either of CAGA- or GC-rich sequences, the principal mode of transcriptional activation appears to be via their ability to interact with and stabilize the transcriptional complex of a diverse set of transcription factors which bind to a variety of specific binding sites in promoters. This stabilization depends on at least three aspects: the ability of the Smad proteins to bind and thereby activate the transcription factor directly, the ability of the Smad proteins to interact weakly with SBE sites adjacent to the transcription factor binding sites, and the ability of the Smad proteins to bind the transcriptional coactivators CBP/p300. Smad proteins also bind to inhibitory proteins, such as TGIF, which recruits histone deacetylases (HDAC) to the transcriptional complex (Wotton et al., 1999a); SNIP1, which competes for the binding of Smad4 to p300/CBP (R. H. Kim et al., 2000); or SIP (Verschueren et al., 1999). They also bind to oncogenes, including Evi-1 (Kurokawa et al., 1998), Ski (Sun et al., 1999), and
Suppressor and Oncogenic Roles of TGF-β
11
SnoN (Stroschein et al., 1999), which inhibit their activity, or the hepatitis B virus pX, which enhances their transcriptional activating activity (Lee et al., 2001) (see Sections V.C. and V.D). Overall, the plasticity of Smad–protein and Smad–DNA interactions provides for a very pleiotropic, contextual effect of Smad proteins on transcription consistent with the cell-specific effects of TGF-β. In this manner the cellular context of transcriptional activators and inhibitors contributes to the way in which a cell “reads” the signal of an active Smad complex in the nucleus.
2. SIGNALING THROUGH MITOGEN-ACTIVATED PROTEIN KINASE CASCADES MAP kinase cascades play pivotal roles in cellular signaling from a wide variety of stimuli, resulting in phosphorylation and activation of transcription factors (Choi, 2000; Denhardt, 1996a; Hartsough and Mulder, 1997). Distinct MAP kinase pathways lead to downstream activation of either extracellular signal-regulated kinases, ERK1/2 (also known as p44/p42 MAPKs), or two stress-activated protein kinases (Sapks)—the c-Jun N-terminal kinase (JNK) and the p38 MAP kinase. These enzymatic pathways, consisting of kinase cascades that transduce signals by sequential phosphorylation and activation of the next kinase in their respective pathway, contrast with the nonenzymatic, unamplified Smad signaling pathways. Depending on the cell, TGF-β can rapidly activate MAP kinase (Hartsough and Mulder, 1995; Ravanti et al., 1999), JNK (Atfi et al., 1997; Engel et al., 1999), or p38 MAP kinase (Hanafusa et al., 1999), and these pathways have been shown, in certain contexts, to be required for both Smad-dependent and Smadindependent transcriptional responses to TGF-β (Hocevar et al., 1999; Mulder, 2000) (Fig. 6; see color insert). In perhaps the best studied example, TGF-β can activate the cascade Ras/Raf-1/MEK/ERK1/2 within minutes, leading to inhibition of growth, autoinduction of TGF-β 1, induction of p21, or even effects on apoptosis (Choi, 2000; Mulder, 2000). Exactly how these MAP kinase pathways are activated by TGF-β is unknown. For example, TGF-β-induced activation of JNK signaling is dependent on the upstream activation of Rho-like GTPases (Atfi et al., 1997; Engel et al., 1999), but linkage of this pathway to the TGF-β receptor complex is unclear. Some insights into upstream links of other MAP kinase pathways to TβRI have been provided with the identification of TAK1 and TAB1, a novel MAP kinase/kinase/kinase and its activator, which have been shown to be important for TGF-β signaling through p38 MAP kinase (Hanafusa et al., 1999). TAK1 is a substrate for the hematopoetic progenitor kinase-1 (HPK-1), which is activated following treatment with TGF-β (G. Zhou et al., 1999), but direct linkage between the TAB/TAK1 cascade and TGF-β receptors is unclear. In the case of another TGF-β superfamily type I receptor,
12
Ester Piek and Anita B. Roberts
BMPR1A, the direct physical linkage of BMPR1A–XIAP–TAB1–TAK1 has been shown, where XIAP is a human X-chromosome-linked member of the inhibitor of apoptosis (IAP) family (Yamaguchi et al., 1999). We have also shown that XIAP interacts with TβRI and activates TGF-β-dependent reporters as well as JNK via a Smad4-dependent pathway (Birkey-Reffey et al., 2001). Whereas the ability of XIAP to activate JNK was blocked by dnTAK1, its ability to activate TGF-β-dependent reporters was not. Since a large variety of growth factors and external stimuli are capable of activating MAP kinase pathways, they represent an important cellular mechanism for integration of signaling inputs. Some reports suggest that persistent activation of MAP kinase pathways by oncogenic Ras can inactivate Smad signaling by retention of R-Smads in the cytoplasm dependent on phosphorylation of consensus ERK sites (PXSP) and JNK/p38 sites (XXSP) in the middle linker regions of the R-Smads (Calonge and Massague, 1999; Kretzschmar et al., 1997, 1999). However, an increasing number of reports have failed to detect any effects of activated Ras or Raf on nuclear localization of Smad proteins (Hu et al., 1999; Lehmann et al., 2000; Liu et al., 2000) and instead show that these pathways are required for Smaddependent induction of an epithelial-to-mesenchymal transition (see Section IV.C) as well as for autoinduction of TGF-β in MDCK dog kidney epithelial cells (Lehmann et al., 2000) and IEC1.4 rat intestinal epithelial cells (Mulder, 2000; Yue and Mulder, 2000). Activation of MAP kinase pathways by ligands signaling through receptor tyrosine kinases can also engage Smad signaling. Thus, EGF or hepatocyte growth factor can stimulate Smad2 phosphorylation and nuclear translocation via an ERK-dependent pathway in epithelial cells (de Caestecker et al., 1998). In hepatocytes, EGF also potentiated Smad3 activation of AP-1 activity mediated in part via activation of p38 and phosphatidylinositol-3′ kinase (PI3 kinase) (Peron et al., 2000). Activated MEKK1, an upstream activator of the JNK pathway in endothelial cells, also induces phosphorylation and nuclear translocation of Smad2 (Brown et al., 1999). Studies in mink lung epithelial Mv1Lu cells have clearly shown that TGF-β can activate JNK by both Smad-independent (early) and Smad-dependent (late) pathways and, conversely, that the TGF-β-dependent phosphorylation of Smad3 by JNK may enhance its C-terminal phosphorylation by the TβRI kinase (Engel et al., 1999). Together, these contrasting findings indicate that TGF-β can activate MAP kinase pathways directly and that these pathways may have positive or negative regulatory effects on R-Smads depending on the cell and the nature of MAP kinase activation. Downstream components of MAP kinase signaling pathways may also interact with the Smad complex in the nucleus, providing an additional level of transcriptional cross talk between these pathways (Fig. 6; see color insert). Smads regulate the transcription of both c-Jun and JunB (Jonk et al., 1998; Wong et al., 1999), downstream substrates of JNK and components of the
Suppressor and Oncogenic Roles of TGF-β
13
AP-1 complex. Moreover, Smads interact functionally with both Jun and Fos to enhance the activity of AP-1 complexes bound to their cognate cis element, the TRE, or to composite sites with juxtaposed AP-1 and SBE sites (W. Tang et al., 1998; Liberati et al., 1999; Wong et al., 1999; Zhang et al., 1998). ATF-2, a constitutively expressed member of the c-Jun family and a downstream substrate of both JNK and p38 MAP kinases, is also transcriptionally activated as a result of its interaction with Smad3 and Smad4 (Hanafusa et al., 1999; Sano et al., 1999). Together, this provides for a complex model for integration of signaling involving the activation of AP-1 via a MAP kinase/Smad-interdependent amplification loop. This has special significance for oncogenesis since AP-1 activity is critical both for the autoinduction of TGF-β 1 (Kim et al., 1990; Yue and Mulder, 2000), and for the acquisition of a metastatic phenotype (Denhardt, 1996b; see Section IV.E).
3. OTHER PATHWAYS PKA and PKC have been shown to play a role in the regulation of TGFβ-dependent gene targets, but the effects appear to be cell specific (Hirota et al., 2000; Sylvia et al., 2000). Some of these effects could be dependent on differential expression of Smad3-dependent gene targets resulting from PKC phosphorylation of Smad3 in its MH1 domain (Yakymovych et al., 2001; see Section II.C.1). Studies in mouse mammary epithelial NMuMg cells have shown TGF-β-dependent phosphorylation of Akt mediated by RhoA and PI3 kinase (Fig. 6), possibly bound to TGF-β receptor complexes (Krymskaya et al., 1997). This pathway is shown to play a role in TGF-βdependent epithelial-to-mesenchymal transformation and migration of tumor cells (Bakin et al., 2000). PI3 kinase is also involved in potentiation of Smad transactivation by Jun proteins mediated by both EGF and TGF-β (Peron et al., 2000). Although inhibitors of PI3 kinase block Smad signaling, the effect is likely indirect and possibly involves interference of Smad binding to SARA, a FYVE motif-containing protein that recruits Smad2 to the activated receptors and that depends on binding of phosphatidylinositol 3-phosphates (Tsukazaki et al., 1998). Smad-independent inhibition of p70 S6 kinase (p70s6k) has also been shown to play a role in inhibition of growth by TGF-β in certain cells (Petritsch et al., 2000). This pathway depends on the activation by TGF-β of the protein phosphatase 2A (PP2A) Bα subunit, previously shown to bind to TβRI (Griswold-Prenner et al., 1998) (Fig. 6). The association of PP2A Bα with the Aβ and A–C subunits and their binding to p70s6k results in dephosphorylation and inactivation of p70s6k (Petritsch et al., 2000). It is suggested that this alternative pathway is sufficient for inhibition of growth by TGF-β, even in the absence of Smad signaling, but that the S6K pathway and Smad signaling must both be inactivated for cells to escape from G1 arrest by
14
Ester Piek and Anita B. Roberts
TGF-β. The potential importance of this mechanism in the tumor suppressor activities of TGF-β depends on whether it can be shown to function in Smad null cells that remain sensitive to inhibition of growth by TGF-β.
III. BIMODAL ACTION OF TGF-β IN TUMORIGENESIS Nearly all cells, including tumor cells and the surrounding stromal, immune, and endothelial cells that play important support roles in tumorigenesis, express both TGF-β ligands and their receptors. Negative regulation of cellular proliferation by TGF-β has been shown to constitute a tumor suppressor pathway (Markowitz, 2000; Markowitz and Roberts, 1996). However, reduction or alteration of TGF-β signaling in tumor cells as they progress through the stages of tumorigenesis is often accompanied by increased secretion and activation of the ligand, which functions both in a paracrine fashion through its effects on accessory cells and, as recently appreciated, in an autocrine manner on the tumor cells to promote tumorigenesis and increase metastasis (Wakefield et al., 2001; Akhurst and Balmain, 1999; Gold, 1999; Reiss, 1999). The latter effects are of particular importance for the many tumor cells in which certain TGF-β signaling pathways remain functional even though growth control by TGF-β may be lost (Chen et al., 1993; Lehmann et al., 2000; Lu et al., 1999; Oft et al., 1996, 1998; Zhao and Buick, 1993). Understanding the function of those TGF-β signaling pathways that remain operative in tumor cells, including potential reorganization of the circuitry such as to change the patterns of cross talk between pathways, is of paramount importance to understanding the development of the malignant phenotype.
A. Tumor-Suppressor Activities of TGF-β Although the purification and characterization of TGF-β was based on its ability to act as a proximal effector of transformation and to stimulate growth of colonies of nontransformed NRK rat kidney fibroblasts in soft agar in combination with EGF (Roberts and Sporn, 1990), it is commonly acknowledged that TGF-β and its signaling receptors and certain downstream signaling components exert tumor suppressor activity on nontransformed cells (Markowitz and Roberts, 1996). TGF-β is a potent inhibitor of cellular proliferation of nonneoplastic epithelial and lymphoid cells, which form the basis of the majority of human cancers (de Caestecker et al., 2000b). TGF-β arrests cells in the G1 phase of the cell cycle (Alexandrow and Moses, 1997; Massague et al., 2000), in part, by suppression of the expression of the protooncogene, c-Myc, a promoter of cell growth and proliferation (Hueber
Suppressor and Oncogenic Roles of TGF-β
15
and Evan, 1998). TGF-β rapidly downregulates expression of c-Myc in most cells sensitive to its effects on inhibition of growth (Alexandrow et al., 1995), and overexpression of c-Myc interferes with TGF-β-dependent inhibition of growth of cells (Blain and Massague, 2000). Recent data suggest that suppression of c-Myc is required for the rapid induction by TGF-β of the expression of cyclin-dependent kinase (cdk) inhibitors p15INK4B (Warner et al., 1999) and p21WAF1/ CIP1 (Claassen and Hann, 2000). The induction of p15 expression by TGF-β requires functional cooperativity of Sp1 with a complex of Smads 2–4 (Feng et al., 2000), whereas that of p21 is dependent on both Sp1 and the Ras/Raf/MEK pathway (Hu et al., 1999; Moustakas and Kardassis, 1998; Pardali et al., 2000). Enhanced expression of p15 increases its binding to the cyclin D-dependent kinases, cdk4 and cdk6, inhibiting their activity and preventing the association of p27Kip1A with the D-type cyclins, thereby stabilizing its binding to and inhibition of cyclinE/cdk2 (Massague et al., 2000). In this manner, TGF-β inhibits both the cyclin D and cyclin E kinases, which are required for progression of cells through the G1/S transition. Although it was initially described as a primary response to TGF-β, it is now understood that inhibition of phosphorylation of the retinoblastoma gene product, Rb, by TGF-β is secondary to suppression of the activity of the G1 cdks (Laiho et al., 1990; Munger et al., 1992). Finally, recent data suggest that another contributing mechanism to TGF-β-dependent growth arrest is Smad-independent inhibition of p70 S6 kinase via PP2A (Petritsch et al., 2000). TGF-β-dependent induction of apoptosis, or programmed cell death, also likely contributes to its tumor suppressor activity since it is an important mechanism for the elimination of preneoplastic and, in some cases, neoplastic cells (Rosfjord and Dickson, 1999). TGF-β induces apoptosis in many cell types, including uterine epithelial cells, hepatoma cells, gastric carcinoma cells, prostatic carcinoma cells, myeloid leukemia cells, and B cell lymphomas (Guo and Kyprianou, 1999; Lin and Chou, 1992; Rotello et al., 1991; Saltzman et al., 1998; Yanagihara and Tsumuraya, 1992). Although suppression of expression of the apoptotic inhibitor Bcl-xL has been implicated in TGF-β-mediated apoptosis (Larisch-Bloch et al., 2000; Saltzman et al., 1998), mechanisms of TGF-β-dependent apoptosis are largely unknown. An experimental approach in which serine 165 in TβRI was mutated has shown that effects of TGF-β on growth inhibitory and apoptotic responses can be segregated in the absence of any differences in effects on transcriptional activation of the reporter 3TP-Lux (Souchelnytskyi et al., 1996). Consistent with this finding, insertional mutagenesis of NRP-154 cells, exquisitely sensitive to TGF-β-induced apoptosis (Hsing et al., 1996), showed that mutant cells could be isolated in which growth inhibitory and apoptotic pathways downstream of TGF-β were interrupted but in which signaling to Smad-dependent reporters was unaltered (Larisch-Bloch et al., 2000). Although these experiments suggest that Smad signaling is not sufficient to induce apoptosis, Smad
16
Ester Piek and Anita B. Roberts
pathways have nonetheless been shown to play a role, although there is no consistent pattern. Thus, induction of expression of the inhibitor Smad7 has been shown to be critical in TGF-β-dependent apoptosis of several prostatic cell lines (Landstrom et al., 2000), whereas Smad7 blocks TGF-β-dependent apoptosis of a variety of other cell types (Patil et al., 2000; Yamamura et al., 2000). The recent discovery of a novel mitochondrial septin-like protein, ARTS (apoptosis-related protein in the TGF-β signaling pathway), which translocates from mitochondria to the nucleus coincident with the induction of apoptosis and activation of caspases, suggests that novel pathways might be involved (Larisch et al., 2000). This protein makes cells competent to undergo apoptosis mediated by TGF-β and has also been shown to be required for apoptosis in cells intrinsically sensitive to TGF-β-mediated cell death, such as the rat prostatic epithelial cell line NRP-154 (Hsing et al., 1996; Larisch et al., 2000). Studies are now focusing on possible dysregulation of the expression or activity of ARTS in tumor cells resistant to effects of TGF-β on apoptosis. Supporting the tumor suppressor activity of TGF-β in vivo are experiments showing that controlled overexpression of a TGF-β 1 transgene in the epidermis of bigenic mice dramatically decreased the proliferative index in the epidermis and resulted in resistance to the tumor promoter phorbol 12-myristate 13-acetate (X. J. Wang et al., 1999). Conversely, genetic inactivation of the TGF-β pathway through loss of TGF-β 1 leads to accelerated progression of multistage skin carcinogenesis in vivo (Cui et al., 1994; Glick et al., 1993), consistent with the observation that deletion of TGF-β 1 by homologous recombination increases the propensity of keratinocytes to undergo transformation to squamous carcinoma cells (Glick et al., 1994). Other tissues also show increased sensitivity to transformation when the TGF-β pathway is compromised. Thus, TGF-β 1 null mice bred into either a p21 null or a Rag2 null background develop colon carcinomas, showing that germline loss of this ligand increases the frequency of tumor initiation (Letterio et al., 1999; Engle et al., 1999). TGF-β 1 heterozygous mice also show an increased frequency of chemically induced tumors of the lung and liver compared to wildtype littermates (Tang et al., 1998b). Significantly, tumors formed in these mice do not lose the remaining allele of TGF-β 1, showing that the TGF-β 1 gene is haplo-insufficient with respect to its tumor suppressor activity and suggesting that retention of the remaining allele likely bestows a selective advantage on the tumor cell due to the coexisting tumor promoting activities of this cytokine (B. Tang et al., 1998). The TGF-β receptors also show tumor suppressor activity (S. J. Kim et al., 2000; Markowitz and Roberts, 1996) as shown convincingly in model systems by the ability of dominant negative (dn) forms of the receptors to increase tumorigenicity of tumor cells which retain responsiveness to the
Suppressor and Oncogenic Roles of TGF-β
17
growth inhibitory effects of TGF-β (Amendt et al., 1998; Bottinger et al., 1997; Go et al., 2000; Wang et al., 1997) and by restoration of the expression of TβRI in cells expressing limiting amounts of this receptor (Wang et al., 1996). In human cancers, mutation or transcriptional repression of TβRII shows a strong correlation to cancer progression (see Section V.A). Consistent with the tumor suppressor activity of the TGF-β ligands and their signaling receptors, Smad4, the common Smad partner for all R-Smads, was initially described in the context of its tumor suppressor locus on chromosome 18q21.1, a locus frequently deleted in pancreatic cancer (Hahn et al., 1996). Moreover, Smad2 mutations have been described in a small subset of human cancers (see Section V.B) and Smad3, although not found to be mutated in human cancers, can be functionally inactivated by a variety of nuclear oncogenes including Evi-1, SnoN, and Ski (see Section V.D). Interestingly, germline loss of Smad3, possibly in collaboration with environmental factors, can lead to colon carcinogenesis (Zhu et al., 1998), although this was not seen in two other studies (Datto et al., 1999; Yang et al., 1999a). Regardless, these studies suggest that in the early stages of carcinogenesis the Smad proteins also function in a tumor suppressor pathway downstream of the TGF-β receptors.
B. Tumor-Promoting Activities of TGF-β In direct opposition to the designation of TGF-β as constituting a tumor suppressor pathway, are both experiments using animal models and clinical data clearly showing it to have tumor promoting activity and even to be essential for the progression of a tumor in terms of its escape from immune surveillance. How can these two opposing actions be reconciled? In addition to its important effects on inhibition of growth of cells and on apoptosis, TGF-β also controls many other aspects of cellular behavior that endow tumor cells with metastatic and invasive potential, including conversion of epithelial cells to an invasive fibroblastoid phenotype at more advanced stages of tumorigenesis (Caulin et al., 1995; Cui et al., 1996; Miettinen et al., 1994; Oft et al., 1996, 1998; Portella et al., 1998) (see Section IV.C). As they progress to fully malignant, invasive cells, the majority of tumor cells lose their ability to be growth inhibited by TGF-β or to be sensitive to TGF-β-induced apoptosis (Wakefield et al., 2001; Akhurst and Balmain, 1999; de Caestecker et al., 2000a). Initially, this loss of sensitivity to TGF-β-induced growth inhibition was considered equivalent to loss of all responsivity to TGF-β. However, it is now appreciated that only in a very small percentage of human cancers are signaling responses to TGF-β actually “lost.” Rather, selection processes may result in outgrowth of tumor cells with an altered response pattern to TGF-β in which disadvantageous
18
Ester Piek and Anita B. Roberts
effects, such as inhibition of growth or stimulation of apoptosis by TGF-β, are lost and in which advantageous effects, such as the acquisition of a metastatic, invasive, fibroblastoid phenotype, are retained (Wakefield et al., 2001; Akhurst and Balmain, 1999; de Caestecker et al., 2000a) (see Section IV.C). We are now beginning to understand some of the changes in tumor cells which result in this altered response pattern. Mutational inactivation and transcriptional repression of the TβRII gene in human cancers change the relative expression levels of TβRII compared to TβRI and result in altered response patterns of cells to TGF-β. High TβRII expression levels are important for inhibition of cellular proliferation, whereas production of extracellular matrix proteins and induction of other potentially important tumorpromoting effects of TGF-β are retained at low TβRII levels, suggesting that different activation thresholds control the induction of different TGF-β responses (Chen et al., 1993; Cui et al., 1995; Geiser et al., 1992; Portella et al., 1998). Moreover, the downstream signal mediators Smad2 and Smad3 exert different roles in TGF-β signal transduction (Piek et al., 2001; Datto et al., 1999; Waldrip et al., 1998; Weinstein et al., 1998; Yang et al., 1999a; Heyer et al., 1999; see Section II.C.1). Therefore, it is likely that the relative balance of activated Smad2 and Smad3 will affect the proximal signaling targets. Contributing to the biological outcome of TGF-β signaling in cells are the relative expression levels of Smad2 and Smad3 and their relative affinities for the different TGF-β receptors. Also, the levels of expression of other intracellular signal regulators, such as oncogenes and inhibitors that can selectively affect the activity of Smad2 or Smad3, will affect the biological outcome of TGF-β signaling in cells. Finally, it is also anticipated that, in addition to possible changes in signaling through the Smad pathway resulting from altered patterns of receptor expression, there might be shifts in the relative activities of the Smad signaling pathway and the catalytic MAP kinase pathways resulting in enhanced activation or expression of various oncogenes, such as AP-1, shown to play key roles in invasion and metastasis (see Section IV.E).
IV. ACTIVITIES OF TGF-β IMPORTANT FOR ONCOGENESIS A. TGF-β Isoform-Specific Activities during Tumorigenesis The three isoforms TGF-β 1, TGF-β 2, and TGF-β 3 generally mediate similar in vitro activities, at least under conditions in which expression of TβRIII is sufficient to efficiently present TGF-β 2 to the signaling receptors. However,
Suppressor and Oncogenic Roles of TGF-β
19
evidence suggests that these three isoforms not only play nonredundant roles during embryonic development (Proetzel et al., 1995; Roberts and Sporn, 1992; Sanford et al., 1997) but, also may exert specific activities during tumorigenesis. Suggestive of this is the demonstration that tumors derived from human HaCaT keratinocytes show expression of TGF-β 1 in differentiated cells, TGF-β 2 in malignant and invading cells, and TGF-β 3 in tumor stroma including tumor blood vessels (Gold et al., 2000). In a different tumor model based on spontaneous colon tumorigenesis developing from an inflammation-associated hyperplasia in TGF-β 1 null mice crossed onto an immunodeficient Rag2−/− background, it has been proposed that TGF-β 1 suppresses the early stages of colon tumorigenesis not by its effects on cellular proliferation but by maintenance of normal crypt architecture (Engle et al., 1999). Moreover, since these mice developed only nonmetastatic colon cancer, these authors propose that expression of TGF-β 2 and TGF-β 3, which is unimpaired in these mice, might actually inhibit metastasis. However, until this can actually be tested in conditional mutants of TGF-β 2 and TGF-β 3, and until we begin to understand how the different isoforms of TGF-β might differentially affect receptor activation or downstream signaling, the mechanisms underlying putative differences in effects of the TGF-β isoforms will remain elusive.
B. Increased Activation of Latent TGF-β Associated with Tumorigenesis Although activation of latent TGF-β is clearly a key regulatory event in the control of TGF-β signaling, specific mechanisms underlying dysregulation of this process in disease pathogenesis are not well understood. However, the importance of activation of latent TGF-β in mediating both its tumor suppressive and tumor promoting roles is underscored by the observations that a variety of chemopreventive agents activate latent TGF-β, which is proposed to partially mediate the inhibition of induction of tumors by these agents, and that tumor cells generally express increased levels of active TGF-β, which is proposed to be critical for both autocrine and paracrine roles of TGF-β in promoting the late stages of carcinogenesis. Various ligands belonging to the steroid hormone receptor family, including retinoids, vitamin D3, tamoxifen, and the synthetic progestin gestodene, as well as the monoterpene perillyl alcohol, have all been shown to enhance cellular secretion of active TGF-β and expression of cellular receptors for TGF-β (Ariazi et al., 1999; Jung et al., 1999; Roberts and Sporn, 1986; Turley et al., 1996). In the case of retinoids, it has been proposed that the increased secretion of active TGF-β results from the induction by retinoids of plasminogen activator and plasmin (Kojima and Rifkin, 1993). In contrast, protection from tumorigenesis by perillyl alcohol has been shown to be
20
Ester Piek and Anita B. Roberts
linked, in part, to its ability to induce expression of the M6P/IGFII receptor, a putative tumor suppressor which is thought to activate TGF-β by binding the latent complex and presenting it to surface-bound plasminogen–plasmin (Devi et al., 1999; Godar et al., 1999; Jirtle et al., 1993). The ability of these agents to enhance expression of active TGF-β and its receptors in target cells which are sensitive to its tumor suppressive effects is consistent with the chemopreventive and, in certain cases, the chemotherapeutic activity of these agents. At later stages of tumorigenesis, when most tumor cells are no longer sensitive to suppression of growth by TGF-β, they develop intrinsic mechanisms to express active TGF-β. The mechanisms underlying tumor cell activation of TGF-β are possibly quite varied and cell specific (Fig. 2). Examples include downregulation in certain tumor types of the expression of LTBPs, which enhance secretion of the latent complex and target it for deposition to the extracellular matrix or to the cell surface for activation, potentially favoring tumor expansion by reducing growth inhibitory constraints (Taipale et al., 1998). Metalloproteinases MMP-2 and MMP-9, localized at the cell surface through binding to the α vβ 3 integrin or hyaluronan receptor CD44, respectively, have also been suggested to enhance activation of TGF-β 1 and thereby to contribute to tumor invasion and angiogenesis (Yu and Stamenkovic, 2000). Overexpression of TSP-1 by several types of tumors can also lead to an increase in circulating levels of bioactive TGF-β (Tuszynski and Nicosia, 1996). TSP-1, similar to TGF-β, can strongly induce cell-secreted PAI-1, as well as cell-associated uPA and/or uPAR, in lung, breast, or pancreatic carcinoma cells (Albo et al., 1999, 2000). PAI-1 prevents uPA-mediated turnover of plasminogen into plasmin and thereby promotes in vitro cell adhesion and cell spreading in these cells (Arnoletti et al., 1995). Neutralizing TGF-β antibodies abolish the TSP-1 effect, suggesting that TSP-1 regulates cellular adhesion through activation of endogenous TGF-β. In turn, TGF-β can promote TSP-1 production by stromal cells, thereby creating a positive feedback loop (Majack et al., 1990). Thus, TSP-1 is an important regulator of the plasminogen–plasmin system that controls extracellular matrix degradation and invasion in different carcinomas, in part through activation of TGF-β. Recent data demonstrate that TGF-β may also contribute to its activation by upregulating uPAR and expression through Ras/MKK4/JNK signaling (Yue and Mulder, 2000).
C. Effects of TGF-β on Epithelial–Mesenchymal Transdifferentiation The ability of epithelial cells to transform into fibroblastoid cells during a process described as epithelial–mesenchymal transdifferentiation (EMT)
Suppressor and Oncogenic Roles of TGF-β
21
is essential during embryogenesis, in development of organs and tissues, in healing of wounded epithelia, and in the malignant transformation of tumor cells into fully malignant, invasive carcinomas. TGF-β plays a profound role in this dedifferentiation of epithelial cells, causing depolarization, disruption of epithelial interactions, altered expression of extracellular matrix proteins, rearrangement of the cytoskeleton, and formation of actin stress fibers, thus providing the cells with metastatic potential and increased motility (Caulin et al., 1995; Lehmann et al., 2000; Miettinen et al., 1994; Oft et al., 1996, 1998; Portella et al., 1998). For example, targeted overexpression of TGF-β 1 in keratinocytes reduces the frequency of chemically induced benign papillomas in the multistage model of skin carcinogenesis in the mouse, but those papillomas that do develop are highly malignant and change from squamous carcinoma cells to spindle cell carcinomas during the progression of skin carcinogenesis (Caulin et al., 1995; Cui et al., 1996; Portella et al., 1998). Thus, although overexpression of TGF-β 1 inhibits the formation of benign tumors, likely by its ability to arrest the epithelial cells in G1, it aggravates later stages of tumorigenesis by transforming epithelial cells into the more aggressive and highly invasive mesenchymal spindle cell type (Cui et al., 1996). Upregulation of TGF-β 3 in fibroblastoid cells might also be of importance to maintaining the spindle cell phenotype (Cui et al., 1996). TGF-β has also been implicated in the malignant progression and metastatic behavior of many other tumor types through induction of EMT (Hojo et al., 1999; Oft et al., 1996, 1998). Further supporting a requirement for autocrine TGF-β signaling in EMT, overexpression of dnTβRII prevented EMT in EpRas cells in vivo and reversed the mesenchymal phenotype of highly metastatic mouse colon carcinoma CT26 cells to an epithelial phenotype accompanied by loss of metastatic and invasive capacities both in vitro and in vivo (Oft et al., 1996, 1998). Similarly, overexpression of dnTβRII in squamous carcinoma cells prevented their conversion into spindle cells in vivo (Portella et al., 1998), while neutralizing TGF-β antibodies or overexpression of the soluble, extracellular domain of TβRII could prevent metastasis and/or invasion of different carcinoma cell lines (Oft et al., 1998). The molecular mechanisms that contribute to TGF-β-induced cellular transformation have been extensively studied in the mouse mammary epithelial cell line NMuMg. Thus, activin receptor-like kinase 2 (ALK2) was initially implicated in the TGF-β-mediated EMT process based on the observations that antisense ALK2 oligonucleotides or overexpression of dnALK2 prevent TGF-β-induced transdifferentiation of these cells (Miettinen et al., 1994). However, evidence that ALK2 can propagate TGF-β signals to downstream targets is lacking, and studies by Piek et al. (1999b) suggest instead that TβRI, in cooperation with TβRII, is the signaling type I receptor activated by TGF-β in these cells. Adenoviral-based overexpression of constitutively active (ca)TβRI, but not caALK2 (A. Moustakas, unpublished
22
Ester Piek and Anita B. Roberts
results), can trigger the transdifferentation process in the absence of TGF-β (Piek et al., 1999b). Further support for the role of TβRI (ALK5) in EMT derives from experiments demonstrating that suboptimal levels of the constitutively active form of TβRI synergize with Smad proteins at levels which, by themselves, are not sufficient to induce EMT (Piek et al., 1999b). Other studies show that additional pathways may be activated by TGF-β in these cells in that TGF-β-dependent phosphorylation of Akt through PI3 kinase and RhoA (see Section II.C.3) results in delocalization of E-cadherin from adherens junctions, ZO-1 from tight junctions, and integrin β 1 from the cell surface, thereby controlling the early events that precede the acquisition of the spindle cell morphology (Bakin et al., 2000). Akt can also play other roles in tumorigenic behavior such as in TGF-β-induced migration of the metastatic breast tumor cell lines 4T1 and EMT6 (Bakin et al., 2000). Many human tumors harbor activated Ras mutations (Bos, 1989; Fearon and Vogelstein, 1990) and the Ras/MAPK signaling pathway has also been implicated in fibroblastoid conversion of carcinomas and tumor invasiveness (Hay, 1995). Ras signaling ultimately leads to activation of AP-1 complexes and several transcription factors of the AP-1 family are involved in EMT and tumor invasiveness (Kustikova et al., 1998; Lamb et al., 1997). Since a wealth of data support the activation of AP-1 complexes and AP-1-dependent promoters by TGF-β via Smad-dependent pathways, as well as the functional cooperation between Smads and AP-1 complexes to drive certain TGF-β responses (W. Tang et al., 1998; Liberati et al., 1999; Wong et al., 1999; Zhang et al., 1998), it is not surprising that the Ras/MAPK/AP-1 pathway contributes to or enhances the effects of TGF-β on EMT. Recent data using an unbiased multigene analysis approach to study gene expression patterns involved in the TGF-β-induced EMT of HaCaT keratinocytes also implicate Ras/AP-1 pathways (Zavadil et al., 2001). Consistent with the loss of TGF-β-dependent tumor suppressor activities accompanying the malignant transformation of cells, v-Ha-Ras transformants (EpRas cells) of the parental nontumorigenic EpH4 mouse mammary epithelial cells are resistant to growth arrest by TGF-β. However, v-Ha-Ras does not block all TGF-β signaling but instead sensitizes the cells to TGF-βinduced transformation into highly invasive fibroblastoid cells in vitro and to TGF-β-dependent maintenance of a highly invasive fibroblastoid phenotype in vivo (Oft et al., 1996). In contrast, activation of Raf in MDCK dog kidney epithelial cells blocks TGF-β-dependent apoptosis, whereas the cells retain sensitivity to TGF-β-induced growth arrest (Lehmann et al., 2000). Induction of Raf expression again triggers EMT in these cells and invasion into collagen gels by a mechanism involving uPA-dependent activation of TGF-β (Lehmann et al., 2000). In contrast to some studies showing that Ras overexpression inhibits the nuclear translocation of Smad proteins (Kretzschmar et al., 1997, 1999), no inhibition of TGF-β-dependent Smad activation or nuclear translocation was seen in either MDCK Raf-ER or EpRas cells
Suppressor and Oncogenic Roles of TGF-β
23
(Lehmann et al., 2000). Rather, suppression of effects of TGF-β on cell growth by oncogenic Ras has been shown to result from mislocalization of p27Kip1 and the cyclin-dependent kinase CDK6 from the nucleus to the cytoplasm, resulting in loss of the TGF-β-dependent switching of the p27 partner from CDK6 to CDK2 (Liu et al., 2000). Together, these data implicate autocrine TGF-β signaling in collaboration with Ras/Raf activation in the induction and maintenance of the invasive and metastatic fibroblastoid phenotype during late-stage carcinogenesis (Oft et al., 1998) and distinguish this signaling pathway from that involved in growth inhibition or apoptosis by TGF-β.
D. Effects of TGF-β on Genomic Instability Many studies suggest that loss of TGF-β ligand or receptor function predisposes to carcinogenesis in the skin, consistent with the tumor suppressor activities of TGF-β (see Section III.A). Chemical-induced multistage skin carcinogenesis is accelerated in mice with loss of TGF-β 1 or overexpression of dnTβRII in the skin, and oncogenic Ras can aggravate the development of TGF-β 1 null keratinocytes into squamous cell carinoma (Amendt et al., 1998; Glick et al., 1993, 1994). Comparison of the frequencies of gene amplification of keratinocyte cell lines in response to treatment with the drug N-phosphonoacetyl-L-aspartate showed that keratinocytes with a targeted deletion of the TGF-β 1 gene have significantly increased frequencies of gene amplification compared to control keratinocyte cell lines. This suggested that effects of TGF-β on genomic stability, independent of its effects on cell proliferation, might also contribute to its tumor suppressor activity (Glick et al., 1996). Although studies using immortalized keratinocytes that overexpress dnTβRII or that lack the TGF-β 1 genes demonstrate that these cells show increased aneuploidy and chromosomal aberrations in vitro (Glick et al., 1999), in vivo studies in mice expressing a dnTβRII in the epidermis suggest that loss of TGF-β receptor function primarily affects cellular proliferation in the absence of effects on chromosome instability (Go et al., 2000). Similarly, in spontaneous colon carcinomas in TGF-β 1 null/Rag2 null mice, there was no evidence of genomic instability resulting from loss of TGF-β 1 (Engle et al., 1999), suggesting that the loss of genomic stability coincident with loss of TGF-β signaling may be limited to cells in culture.
E. Effects of TGF-β on Invasion and Metastasis Tumor cells that metastasize are endowed with characteristics that enable them to escape from the primary tumor site, invade stroma, transverse
24
Ester Piek and Anita B. Roberts
endothelial cells to enter the blood stream and be transported to distant tissues, adhere to and transverse endothelial basement membranes, and subsequently invade the host tissue (Boyce et al., 1999; Guise, 2000; Koeneman et al., 1999). The metastatic capacity of tumor cells depends to a large extent on a delicate and tightly regulated balance in expression of cell adhesion molecules and proteolytic enzymes, including matrix metalloproteinases (MMPs), tissue inhibitors of matrix metalloproteinases (TIMPs), and the plasminogen–plasmin protease system. TGF-β contributes significantly to the metastatic and invasive properties of tumor cells by regulating expression of several of these adhesion proteins and proteases, some of which in turn can modulate the bioactivity of TGF-β. The fact that certain types of tumors preferentially metastasize to particular tissues was first noticed by Paget (1889), who proposed the seed–soil theory, suggesting that certain host tissues provide fertile soils that foster the growth of particular tumor cells. We focus on the role of TGF-β in the multistep process of tumor cell migration and host penetrance.
1. EXTRACELLULAR MATRIX PROTEINS TGF-β has been shown to control adhesion of tumor cells to basement membranes and extracellular matrix (ECM) proteins by regulating expression of several adhesion molecules, including fibronectin, laminin, vitronectin, types I and IV collagen, tenascin, and several integrins such as the collagen receptor α 2β 1 and the fibronectin receptor α 5β 1, each of which has been implicated in metastasis and invasion (Arrick et al., 1992; Heino and Massague, 1989; Keski-Oja et al., 1988; Koli et al., 1991). In addition, several of these ECM proteins are also involved in cellular motility induced by TGF-β as a result of their chemotactic and chemoinvasive properties (Festuccia et al., 1999; Woodhouse et al., 1997). Enhanced cellular adhesion and migration of hepatocellular carcinoma cells by TGF-β was accompanied by TGF-β-mediated induction of α 5β 1, downregulation of phosphatase PTEN protein expression, and increased tyrosine phosphorylation of focal adhesion kinase (Cai et al., 2000), but it is not clear whether these events are required for TGF-β-mediated migration. Studies focused on the cross talk between TGF-β signaling and integrin signaling will provide major insight into the molecular mechanisms by which TGF-β can promote tumor cell adhesion and metastasis.
2. MATRIX METALLOPROTEINASES AND THEIR INHIBITORS One of the mechanisms by which TGF-β can control the level of extracellular matrix turnover, a prerequisite for metastasis and invasion of host tissue, is by regulating expression of MMPs and their inhibitors. MMPs are zinc-dependent endopeptidases that undergo extracellular proteolytic
Suppressor and Oncogenic Roles of TGF-β
25
cleavage for activation. The expression and activation levels of MMPs and TIMPs have been correlated with aggressiveness and metastatic potential of tumors (Ray and Stetler-Stevenson, 1994). TGF-β can induce the expression of several MMPs in multiple types of tumors (Festuccia et al., 2000; Samuel et al., 1992; Welch et al., 1990; Farina et al., 1998). Posttranscriptional and posttranslational mechanisms can also contribute to enhanced expression of MMPs by TGF-β, including stabilization of MMP-9 mRNA or stabilization of secreted MMP-2 proenzyme, as observed in prostate tumor cells (Sehgal and Thompson, 1999). Supporting the role of MAP kinase pathways and AP-1 in invasion and metastasis, upregulation of MMP-13/collagenase-3 by TGF-β in transformed human squamous epithelial cells is dependent on p38 MAPK signaling and possibly involves interaction of JunB with an AP-1 promoter element. Both p38 MAPK and ERK1,2 signaling are also involved in TGF-β-induced expression of MMP-1/collagenase-1 and MMP-9, possibly by promoting interaction of c-fos with the AP-1 promoter element (Johansson et al., 1999, 2000). Consistent with the role of the MMPs in metastasis, inhibition of p38 activity prevents TGF-β-induced migration of ras-transformed HaCaT cells in vitro (Johansson et al., 2000). TGF-β induces MMP-9 activity in five of six cell lines isolated from metastatic prostate tumors but only in one of six cell cultures derived from primary tumors, thereby strongly correlating MMP-9 induction with the metastatic potential of tumor cells (Sehgal et al., 1996). Furthermore, exposure of mammary adenocarcinoma cells to TGF-β or overexpression of TGF-β in prostate tumor cells or fibrosarcoma cells increases invasiveness in vitro and enhances metastatic potential in vivo, which is partly due to enhanced collagenolytic activity as a result of increased expression of MMP-2 and/or MMP-9 by TGF-β (Samuel et al., 1992; Stearns et al., 1999; Welch et al., 1990). MMP-2 and MMP-9 can also promote invasion by proteolytic activation of TGF-β by their interaction with α vβ 3 integrin or hyaluronan receptor CD44, respectively, on the cell surface (Yu and Stamenkovic, 2000). TIMPs control the extent of extracellular matrix degradation by MMPs, and the four TIMPs identified to date (TIMP-1 through 4) have all been implicated in suppression of metastasis (Khokha, 1994; Sun et al., 1994). TGF-β induces TIMP expression in several tumor cell types (Farina et al., 1998; Festuccia et al., 2000; Kordula et al., 1992). Induction of TIMP expression in HT1080 fibrosarcoma cells by TGF-β is correlated with reduced collagenolytic activity in the conditioned medium and with the antiinvasive effects of TGF-β in these cells in vitro (Kubota et al., 1991).
3. THE PLASMINOGEN–PLASMIN PROTEASE SYSTEM In addition to MMPs and TIMPs, the plasminogen–plasmin protease system is also of importance for cancer progression by mediating pericellular
26
Ester Piek and Anita B. Roberts
proteolysis (Fig. 2). TGF-β and its major activator TSP-1 (see Section IV.B) have been shown to regulate expression of several components of this protease system in multiple types of tumors, including uPA and the uPA receptor uPAR as well as PAI-1 (Farina et al., 1998; Festuccia et al., 2000; Keski-Oja et al., 1991). Moreover, several clinical studies show a strong correlation between elevated levels of TGF-β and high uPA and PAI-1 levels, for example, in breast tumors (Foekens et al., 1992; Murray et al., 1993). uPA, which is inhibited by PAI-1, can control bioactivation of TGF-β through conversion of plasminogen into plasmin, thereby establishing a mutually stimulatory feedback loop between TGF-β and uPA/plasmin (Godar et al., 1999; Odekon et al., 1994; Rifkin et al., 1997; Teti et al., 1997) (see Section II.A). Plasmin is also involved in degradation of extracellular matrix and induction of MMP secretion and thereby plays an important role in invasion of metastatic cells into stroma and tissues (Andreasen et al., 2000). The functional importance of the plasminogen–plasmin system in tumor cell invasion induced by TGF-β is shown in different tumor cell types. Thus, in prostate tumor cells, transformed keratinocytes, and breast tumor cells, TGF-β increases uPA and MMP-9 levels; despite elevated levels of the protease inhibitors PAI-1 and TIMP, this correlates with uPA and plasmindependent stimulation of matrigel invasion by TGF-β (Festuccia et al., 2000; Farina et al., 1998; Santibanez et al., 1999). In other cells such as human HT-1080 tumor cells, TGF-β strongly induces PAI-1 expression and inhibits anchorage-independent growth of these tumor cells, suggesting that a shift in the balance of this system might prevent pericellular proteolysis in certain tumor cells in vitro (Laiho et al., 1987). Recent studies have provided insight into the signal transduction molecules involved in induction of uPAR and uPA expression by TGF-β and emphasize the importance of MAP kinase signaling in control of expression of this proteolytic system. In GEO colon cancer cells, TGF-β induces uPAR expression through activation of the Ras/MKK4/JNK1 pathway, leading to formation of a complex of JunD and Fra-2 which activates a distal AP-1 promoter element in the uPAR gene (J. Yue and K. Mulder, unpublished data). Induction of uPA expression and invasion and metastasis by TGF-β in transformed keratinocytes involve tyrosine kinase signaling as shown by treatment of these cells with the tyrosine kinase inhibitors genistein or curcumin (Santibanez et al., 2000). In contrast, in SW480 Smad4 null colon carcinoma cells reconstitution of Smad4 expression resulted in suppression of endogenous uPA and PAI-1 gene expression and was associated with accelerated cell adhesion and cell spreading in vitro and reduced tumorigenicity in vivo, suggesting that prominence of the Smad pathway may shift the balance to the tumor suppressor activity of TGF-β (Schwarte-Waldhoff et al., 1999).
Suppressor and Oncogenic Roles of TGF-β
27
4. SYMBIOTIC CROSS TALK BETWEEN INVADING CELLS AND HOST TISSUE Prostate and breast tumor cells preferentially disseminate to bone, which is the third most frequent metastatic site (Boyce et al., 1999; Guise, 2000; Koeneman et al., 1999; Woodhouse et al., 1997). The bone marrow, which is first invaded by tumor cells metastasizing to the skeleton, is rich in hematopoietic growth factors and cytokines, including TGF-β and BMPs. TGF-β and BMPs are also secreted by osteoblasts in the bone matrix and are activated during the normal process of bone remodeling as well as during degradation of the bone matrix caused by tumor cell invasion (Boyce et al., 1999; Goltzman, 1997). Bone-derived TGF-β 1 is chemotactic and chemoinvasive to prostate cells in vitro and regulates the expression of cell adhesion molecules and proteases on these cells (Festuccia et al., 1999, 2000). Although direct evidence is lacking, TGF-β produced in bone is likely one of the factors that attracts circulating and metastasizing prostate cells in vivo. Prostate tumor cells produce a wide variety of growth factors, including TGF-β and BMPs, and express several of the corresponding receptors (Barrack, 1997; Ide et al., 1997). Prostate tumor cells engineered to overexpress TGF-β 1 have lost sensitivity to the growth inhibitory effects mediated by TGF-β and instead display enhanced in vivo tumorigenicity and metastatic potential compared to control cells (Steiner and Barrack, 1992), possibly through altering the balance in expression of cell adhesion molecules and proteases or by enhancing motility of the tumor cells (Barrack, 1997; Morton and Barrack, 1995). Moreover, prostate tumor cells trigger osteoblasts to form new bone (osteosclerosis), possibly by release of bone promoting factors such as TGF-β and BMPs from the tumor cells (Boyce et al., 1999; Goltzman, 1997). In contrast, breast tumor cells that metastasize to the skeleton cause osteolytic destruction of bone, and several studies implicate parathyroid hormonerelated peptide (PTH-rP) in this process (Guise et al., 1996; Powell et al., 1991). TGF-β has been shown to induce PTH-rP expression in renal and squamous cell carcinoma and in breast tumor cells in vitro (Kiriyama et al., 1993; Merryman et al., 1994; Yin et al., 1999) in a Smad-dependent fashion (Guise, 2000). The importance of TGF-β signaling in the osteolytic bone destruction triggered by metastasizing breast cancer cells was shown by overexpression of dnTβRII in MDA-MB-231 tumor cells, which resulted in suppressed PTH-rP levels and concomitant reduced bone destruction (Guise, 2000; Yin et al., 1999). Overexpression of PTH-rP in dnTβRII expressing cells restored the severity of bone metastasis. In addition, overexpression of caTβRI increased PTH-rP levels concomitant with increased osteolytic lesions and reduced survival (Yin et al., 1999), whereas concomitant addition of PTH-rP antibodies reversed the effect induced by caTβRI (Guise, 2000).
28
Ester Piek and Anita B. Roberts
Thus, autocrine TGF-β receptor signaling through the Smad pathway leads to induction of PTH-rP secretion by the breast tumor cells, which in turn causes osteolytic lesions of the bone. TGF-β, which is released and activated during bone destruction, further induces PTH-rP expression, thereby creating a vicious cycle between these two proteins (Guise, 2000). The propensity of breast tumor cells to metastasize to bone might also result from the cooperative induction of PTH-rP expression by TGF-β and a constitutively active estrogen receptor α mutant (ER-α Tyr537Asn) identified from a human bone metastasis (Guise, 2000). Cooperation between TGFβ-activated Smads and the vitamin D hormone receptor in transcriptional regulation of gene responses has been reported as well (Yanagisawa et al., 1999), and the cross talk between Smads and steroid hormone receptors may endow certain tumor cells with additional oncogenic properties.
F. Indirect Effects of TGF-β on Tumorigenesis The ability of TGF-β to act on nearly every cell type enables its effects in carcinogenesis to be extended from those on the tumor cell to those on stromal elements, including vascular endothelial cells, immune cells, and components of extracellular matrix, each of which modulates growth of the tumor. These effects can include both paracrine effects mediated by TGF-β secreted by tumor cells and cell-autonomous effects (Fig. 1).
1. ANGIOGENESIS Tumors require an adequate supply of nutrients in order to grow. This is accomplished by tumor-induced neovascularization, also referred to as angiogenesis. Angiogenesis also increases the rate of metastasis by enabling cells that have detached from the primary tumor to reach the blood system for transport to distant sites in the body. Although early studies in which TGF-β was injected subcutaneously showed it to be proangiogenic (Roberts et al., 1986), other studies in which TGF-β was overexpressed under control of tissue-specific promoters in transgenic mice, as might model secretion of active TGF-β from a tumor cell, have revealed contradictory roles for TGF-β in angiogenesis. Depending on the system, TGF-β has been shown to have no effect on neovascularization (Cui et al., 1995; Nabel et al., 1993; Pierce et al., 1993; Sanderson et al., 1995), to be proangiogenic (Wang et al., 2000; X. J. Wang et al., 1999), or to be antiangiogenic (Pierce et al., 1995). These effects can be mediated by direct action of TGF-β on endothelial cells or vascular smooth muscle cells (VSMCs) and pericytes, which stabilize the vessels. Alternatively, these effects can be indirect, resulting from its ability to recruit inflammatory cells, connective tissue cells, and epithelial cells
Suppressor and Oncogenic Roles of TGF-β
29
that express angiogenic factors, such as vascular endothelial growth factor (VEGF) and basic fibroblast growth factor, in response to TGF-β (Pepper, 1997; Roberts and Sporn, 1989). In support of this finding, in vivo studies showed that increased neovascularization correlated with effects of TGF-β resulting in increased expression of proangiogenic VEGF and reduced levels of antiangiogenic TSP-1 (Go et al., 1999), whereas reduced vascularization such as that resulting from reintroduction of Smad4 into pancreatic tumor cells was accompanied by suppression of VEGF expression and an increase in expression of TSP-1 (Schwarte-Waldhoff et al., 2000). In addition, in vitro experiments examining endothelial cell invasion of three-dimensional collagen or fibrin gels suggest that TGF-β can modulate the effects of these factors on endothelial cells (Pepper et al., 1993; Pepper, 1997). Direct effects of TGF-β on endothelial cells include inhibition of cellular proliferation and inhibition or stimulation of migration, depending on the assay conditions (Madri et al., 1988; Muller et al., 1987; Vernon and Sage, 1999). In addition, TGF-β can induce endothelial cells to form tubelike structures and deposit ECM proteins necessary to form a basement membrane, as shown in three-dimensional assays in vitro (Madri et al., 1988; Merwin et al., 1990). Moreover, there is a synergistic interplay of endothelial cells and pericytes/VSMC resulting in activation of latent TGF-β, which acts to induce both endothelial cell quiescence and differentiation of mesenchymal precursor cells into VSMC and pericytes (Folkman and D’Amore, 1996; Nunes et al., 1998; Pepper, 1997; Tada et al., 1994). The latter is accompanied by the induction of several markers of VSMC differentiation in vitro, including α-SM actin, SM myosin heavy chain, SM22α, and telokin, all of which harbor a positive-acting TGF-β control element in their promoter regions (Adam et al., 2000; Hautmann et al., 1997; Owens, 1998). Quantitative immunofluorescence analysis in wild-type or TGF-β +/− rat arteries in vivo showed a strong correlation in expression levels of TGF-β and markers of VSMC differentiation, implicating TGF-β in the in vivo control of VSMC differentiation (Grainger et al., 1998). Evidence for the importance of the TGF-β signal transduction pathway in the control of angiogenesis in vivo is provided by the phenotype of mice lacking expression of TGF-β 1, TβRII, ALK1, endoglin, Smad1, or Smad5, which die on approximately embryonic day 11 from impaired angiogenesis (Goumans and Mummery, 2000; Lechleider et al., 2001). These studies implicate TGF-β signaling in terminal differentiation of endothelial cells, recruitment and differentiation of VSMC and pericytes, and in integrity of the vessel walls (Chang et al., 1999; Dickson et al., 1995; Li et al., 1999; Oh et al., 2000; Oshima et al., 1996; Yang et al., 1999b). The importance of ALK1 and endoglin in vascular development is underscored by the high incidence of loss-of-function mutations in their gene loci associated with hereditary hemorrhagic telangiectasia, an autosomal-dominant
30
Ester Piek and Anita B. Roberts
disorder characterized by multisystemic vascular dysplasia (Johnson et al., 1996; McAllister et al., 1994). ALK1 is an alternative TGF-β receptor in endothelial cells which, together with endoglin and likely TβRII, can regulate angiogenesis by activation of Smads 1 and 5 (Oh et al., 2000). Interestingly, both ALK1 and endoglin negatively regulate TβRI signaling in different cell types, possibly serving to protect the endothelium from the suppressive effects of TGF-β on proliferation. Thus, ALK1 can suppress TβRI signaling in HepG2 cells through a mechanism dependent on Smads 1 and 5 (Oh et al., 2000), and endoglin counteracts the inhibitory effect of TGF-β 1 on cellular proliferation, migration, and capillary structure formation of HUVEC cells as well as other cell types (Letamendia et al., 1998; Li et al., 2000b). Many reports document increased expression of endoglin in endothelial cells of tumors, suggesting that it might promote angiogenesis by protecting the cells from the suppressive effects of TGF-β (see Section V.A). Overall these findings suggest that effects of TGF-β on endothelial cells are regulated by the relative levels of signaling from two different type I receptors that activate distinct Smad pathways, requiring integration of these signals by the common mediator Smad4. Additional intracellular cross talk may involve interactions between the Smad pathway and the MAPK pathway, which has also been shown to play an important role in angiogenesis (Pages et al., 2000). In this regard, MEKK1, an upstream activator of the SAPK/JNK pathway, has been shown to activate Smad2 in endothelial cells, independent of TGF-β (Brown et al., 1999). In addition, the recent demonstration that VSMCs express an alternative splice variant of TβRI that lacks four amino acids in the extracellular domain adjacent to the transmembrane region raises the question of whether relative levels of expression of this receptor and the wild-type TβRI, which signal with different potency, might contribute to differential responsivity of VSMC to TGF-β in cancer (Agrotis et al., 2000).
2. IMMUNOSUPPRESSION Immune surveillance is a mechanism whereby immune cells home to tumor cells to effect their elimination. As tumors progress, they evolve mechanisms to escape from immune surveillance, and TGF-β is key to that process. TGF-β has profound suppressive effects on hematopoietic cells, regulating their proliferation and differentiation as well as controlling activation of the differentiated cell lineages (Fortunel et al., 2000). It is the most potent endogenous immunosuppressive factor identified to date, with broad actions on a variety of lineages, including activated macrophages, lymphocytes, and neutrophils. Each of these cell types has the ability to both express receptors for TGF-β and synthesize and secrete the ligand. Studies in the mid-1980s showed that TGF-β 1 plays a critical role in immune cell homeostasis in that
Suppressor and Oncogenic Roles of TGF-β
31
in vitro activation of T cells results in increased expression of TGF-β receptors and increased secretion of the ligand, resulting in a self-limitation of clonal expansion of activated T cells (Kehrl et al., 1986, 1991). Moreover, TGF-β is also a potent inhibitor of the generation of allospecific cytotoxic T lymphocytes (CTLs) from mixed lymphocyte cultures and inhibits the proliferation of these cells, although it does not inhibit their cytolytic activity (Ranges et al., 1987). Data suggest that immune suppression by TGF-β is due in part to impairment of the expression and function of both IL-2 and its high-affinity receptors, which are generally considered critical for initiation and maintenance of T-cell activation and function (Kehrl et al., 1986). In addition to suppressing the response to IL-2, TGF-β also inhibits expression of other mitogenic cytokines, such as TNF-α and IFN-γ , by immune cells. The IL-2-dependent expansion of lymphokine-activated killer (LAK) cells (Mule et al., 1988), tumor infiltrating lymphocytes (Merogi et al., 1997), and natural killer (NK) cells (Rook et al., 1986), each of which is thought to be important in the natural immune surveillance of tumor cells, is also inhibited by TGF-β. Cancer patients often present with defective immune responses. Since many tumor cells secrete active TGF-β, it has been proposed, based on earlier studies of the suppressive effects of TGF-β on immune cells, that this might contribute to the immunosuppression generally characteristic of tumor-bearing hosts (Tada et al., 1991), and that it might provide a mechanism by which tumor cells escape elimination by tumor-specific T lymphocytes (de Visser and Kast, 1999; Wojtowicz-Praga, 1997). Indeed, experiments in which highly immunogenic tumor cells become less immunogenic and more proliferative when transfected with an expression vector for TGF-β 1 have led to the hypothesis that TGF-β production by a tumor may be important in determining whether the tumor disappears or grows progressively (Torre-Amione et al., 1990). Supporting the concept that advanced tumors secrete substantial amounts of TGF-β that can contribute to a generally suppressed immune response are data showing that plasma levels of TGF-β often correlate with disease progression and decrease following surgical resection of the tumor (Tsushima et al., 1996; Wunderlich et al., 1998). In animal models, antibodies to TGF-β have been shown to decrease the tumorigenicity of breast cancer cells, in part, by increasing NK cell activity (Arteaga et al., 1993a,b). Other studies using a rat model of glioma, in which immunization with rat 9L gliosarcoma cells engineered to express antisense TGF-β 2 resulted in complete remissions in all animals, point to putative immune suppressive effects of tumor cell-derived TGF-β (Fakhrai et al., 1996). An exciting feature of these studies was that rats in which a first tumor was eliminated by immunization with antisense-expressing tumor cells showed significant resistance to the establishment of secondary tumors upon repeated challenge. Moreover, these immunized rats showed higher cytolytic activity than rats
32
Ester Piek and Anita B. Roberts
immunized with unmodified tumor cells, and this increased activity was also observed using NK cell-sensitive target cells. Similar results have been obtained using a model of C6 gliomas engineered to express decorin, a small leucine-rich proteoglycan which can sequester TGF-β (Stander et al., 1999). Decorin-expressing tumors regressed and showed a fourfold increase in infiltration by activated T cells (Stander et al., 1998). These groundbreaking studies were extended using the human glioma cell line Onda10 and EMT6 murine mammary tumor cells to show that retroviral transduction with antisense TGF-β 1 increased immunogenicity of the tumor cells and enhanced their susceptibility to LAK cells both in vitro and in vivo (Park et al., 1997; Yamanaka et al., 1999). Recent studies show that TGF-β and IL-10 secreted by T-regulatory γ δ T cells in early tumor lesions also contribute to the attenuation of the antitumor activity of CTLs and NK cells (Seo et al., 1999). These findings have been expanded to include other animal models, such as mammary cancer, thymoma, and melanoma, and collectively suggest that efficient generation of antitumor immune activity requires the concurrent reduction of suppressive factors such as TGF-β that potentially mitigate immune cell reactivity (Conrad et al., 1999; McEarchern et al., 1999; Won et al., 1999). Correlative data supporting a role of tumor-secreted TGF-β in immunotherapy are accumulating from clinical trials. Thus, in colorectal cancer patients, in a cohort of patients who responded to immunotherapy with IL-1/IL-2 and adoptive cellular therapy no TGF-β reactivity was found in their tumors, whereas tumors of a group of nonresponders in the same trial did express TGF-β (Doran et al., 1997). No significant correlations could be seen with other cytokines. In summary, although many approaches to sequester or restrict tumor cell-secreted TGF-β are being tested in animal models and appear to be promising, there are still many problems to overcome before any of them might be used effectively in humans (de Visser and Kast, 1999).
3. DESMOPLASIA Desmoplasia is the formation of highly cellular, excessive connective tissue stroma associated with certain cancers. The process shares many features with fibrosis, involving activation of fibroblasts to secrete increased levels of matrix proteins and matrix degrading enzymes. TGF-β affects every aspect of extracellular matrix production, from the induction of the expression of matrix proteins and protease inhibitors to suppression of expression of matrix-degrading proteases (Border and Noble, 1994; Roberts and Sporn, 1990). Strong expression of TGF-β is seen in a variety of cancers with associated desmoplasia, including desmoplastic ameloblastoma and hepatocellular carcinoma associated with lamellar fibrosis (Orsatti et al., 1997;
Suppressor and Oncogenic Roles of TGF-β
33
Takata et al., 2000), suggesting the TGF-β secreted by tumor cells is important in the formation of desmoplastic matrix. Additional support for this conclusion derives from studies of human invasive mammary ductal carcinomas showing a correlation of expression of TGF-β and connective tissue growth factor (CTGF), a proximal inducer of collagen synthesis under regulation by TGF-β (Grotendorst, 1997). Tumors in which there was extensive connective tissue involvement showed strong staining for TGF-β in the tumor epithelial cells and for CTGF in the stromal fibroblasts (Frazier and Grotendorst, 1997). The mechanisms involved in this process have been reviewed extensively in the context of the role of TGF-β in fibrotic disease (Border and Noble, 1994) and will not be discussed here.
V. DYSREGULATED EXPRESSION OR ACTIVITY OF COMPONENTS OF TGF-β SIGNALING PATHWAYS IN ONCOGENESIS Although there are many examples of mutations in the TGF-β receptors and in Smads 2 and 4 in human cancers, some form of the TGF-β signaling pathway likely remains intact in most cancer cells (Wakefield et al., 2001; Akhurst and Balmain, 1999; de Caestecker et al., 2000a). Altered levels of the TGF-β receptors or Smad proteins, altered interplay between Smad and MAP kinase pathways, and either mutationally altered function of receptor and Smad proteins or altered cellular context all may contribute to modification of signaling from predominance of a tumor suppressor phenotype to that of a prooncogenic phenotype (Table I).
A. Receptors There is a strong correlation between malignant progression and loss of sensitivity to the antiproliferative effects of TGF-β, which is frequently associated with reduced expression or inactivation of TGF-β receptors (S. J. Kim et al., 2000; Markowitz and Roberts, 1996). Although there are only sporadic reports of mutations or deletions in TβRI in malignancy (Goggins et al., 1998), TβRII is a frequent locus of inactivating mutations. The most common mutation is seen in replication error repair positive (RER+) colon (usually in the proximal colon most distant from the rectum) and gastric carcinomas that harbor a deficient DNA mismatch repair system (Markowitz, 2000). In these tumors, genomic instability is associated with frameshift mutations in a 10-base pair polyadenine tract (big polyadenine tract; BAT) in exon 3
Table I Regulators of the TGF-β Signal Transduction Cascade Dysregulated in Tumorigenesis Function Ligand interactors TSP-1/uPA/plasmin MMP-2 MMP-9 Steroid receptor ligands M6P/IGFII receptor LTBP
34
Decorin Ligand TGF-β 1 TGF-β 2 TGF-β3 Receptors TβRII TβRI ALK1 BAMBI Endoglin Betaglycan
Expression in tumors
Inducer of plasminogen–plasmin system; proteolytic activation of TGF-β Interacts with α vβ 3 integrin, activates TGF-β Interacts with hyaluronan receptor CD44; activates TGF-β Increases secretion of bioactive TGF-β and expression of TGF-β receptors Presents latent TGF-β to plasmin; activates TGF-β Enhances secretion and ECM deposition of TGF-β; facilitates activation of TGF-β Sequesters TGF-β on cell surface
Induced Induced Induced Induced Induced Suppressed
Multiplicity of actions on both tumor cells and stroma Regulator of invasiveness of keratinocytes? Maintains fibroblastoid phenotype of transformed epithelial cells
Induced Induced Induced
Ligand binding receptor Signaling receptor; activator of Smads and MAPK Inhibits TGF-β signaling through TβRI on endothelial cells Kinase-deficient type I receptor; inhibits TGF-β signaling Accessory receptor for ALK1, TβRI; does not bind TGF-β 2; inhibits TGF-β signaling Accessory receptor for TβRII / TβRI
BAT-RII mutations, suppressed expression Transcriptionally repressed ? Downregulated Induced expression in tumor vasculature
Suppressed in tumors, upregulated in stroma
Dysregulated processing by oncogenic Ras
35
Smads Smad2 Smad3 Smad4 Smad6, Smad7 Smad interactors TGIF2 Evi-1 Ski, Sno-N Oncoprotein pX v-src, v-abl SNIP1 MAPK TAB/TAK1 Ras, Raf Other pathways PI3K PKC
Signal transducer Signal transducer; synergizes with MAPK signaling Common mediator Inhibits R-Smad activation
Mutational inactivation, genetic loss ? Mutational inactivation, genetic loss Induced
Related to TGIF, which inhibits Smad signaling by recruitment of HDACs Interacts with Smad3 and prevents its binding to DNA Repressor of Smad transcriptional activity Enhances activity of Smads by stabilizing Smad complex with TFIIB Induces receptor expression Competes for binding of Smad4 to p300; inhibits signaling
Overexpressed Overexpressed Overexpressed Overexpressed Overexpressed ?
Mediates TGF-β signals by activation of p38 Phosphorylates R-Smads in linker region; activates AP-1 complexes
Oncogenic/hyperactive mutations Oncogenic/hyperactive mutations
Activates Akt-2 and Jun Phosphorylates Smad3 in MH1 domain; inhibits binding to DNA
? ?
36
Ester Piek and Anita B. Roberts
of the TβRII gene (BAT-RII), resulting in truncated receptors that lack the serine/threonine kinase domain (Grady et al., 1999; Markowitz et al., 1995; Markowitz, 2000). Other TGF-β receptor mutations, such as the Thr315Met germline mutation of TβRII in a family with hereditary nonpolyposis colorectal cancer, do not interfere with the kinase activity of the type II receptor but alter the responsivity and enhance the metastatic potential of tumor cells by specifically impeding TGF-β-mediated growth arrest without affecting induction of extracellular matrix formation (Lu et al., 1999). Whereas TGF-β receptor mutations other than the BAT-RII frameshift mutations are rare events in tumorigenesis, repression of TGF-β receptor expression is a common mechanism that enables tumor cells to escape from negative regulation of growth by TGF-β (S. J. Kim et al., 2000). Thus, aberrant receptor trafficking can reduce cell surface receptor expression, for example, in pristane-induced mouse plasmacytomas (Amoroso et al., 1998) and a human cutaneous T cell lymphoma (Knaus et al., 1996), although this frequently involves transcriptional silencing of the TGF-β receptor promotors. Several members of the ETS family of transcriptional transactivators, including ERT/ESX/ESE-1/ELF3/jen and Fli-1, are critical for the expression of TβRII and correlate with reduced receptor expression in gastric cancers (Choi et al., 1998; Hahm et al., 1999). In Ewing sarcoma (EWS), any of several members of the ETS family are fused to the EWS gene as a result of chromosomal translocations (Im et al., 2000). The EWS/Fli-1 fusion protein represses TβRII expression as it retains the DNA-binding activity of Fli-1 but is unable to activate transcription due to deletion of the transactivating amino-terminus of Fli-1 (Hahm et al., 1999). TβRII expression is also negatively regulated by several oncogenes frequently overexpressed in human tumors, including H-Ras (Zhao and Buick, 1995), the adenoviral oncoprotein E1A (Kim et al., 1997), and cyclin D1 (Okamoto et al., 1994). Another mechanism of transcriptional repression observed in some gastric cancer cell lines involves hypermethylation of CpG islands in the TβRI promoter (Kang et al., 1999). Methylation of gene promoters results in chromatin condensation, limiting accessibility of transcription factors to the DNA. Although this is of critical importance for gene silencing in development, the mechanisms underlying abnormal methylation in cancer are unknown (Jones and Wolffe, 1999). Finally, there are mutations in the TβRII promoter that can contribute to reduced receptor expression in tumors by interfering with binding of transcriptional regulators to the mutated promoter (Hougaard et al., 1999; Jackson et al., 1999; Munoz-Antonia et al., 1996). Changes in expression of endoglin and betaglycan in tumorigenesis have also been reported. Endoglin/CD105 is overexpressed in the vasculature of different types of tumors in advanced stages of tumorigenesis (Bodey et al., 1998; Burrows et al., 1995; Kumar et al., 1999; Matsuno et al., 1999). These elevated levels of endoglin expression in the vasculature potentially
Suppressor and Oncogenic Roles of TGF-β
37
contribute to a permissive environment for tumor cell growth by protecting the endothelium from the suppressive effects of active TGF-β secreted by tumor cells (Li et. al., 2000b). In addition, elevated levels of soluble endoglin or endoglin/TGF-β complexes are found in plasma of breast cancer patients, correlating with the presence of distant metastases and with poor prognosis for patient survival (Kumar et al., 1999; Li et al., 2000a). Intriguing new studies show that overexpression of oncogenic but not wild-type Ras in HD6-4 colon cancer cells changes them from being insensitive to effects of TGF-β on growth to being growth stimulated by TGF-β, giving rise to a more aggressive phenotype. This correlates specifically with posttranslational modification of betaglycan and suggests that dysregulated processing of betaglycan can also contribute to tumorigenicity (Yan et al., 2001). The pseudoreceptor BAMBI, a naturally occurring truncated type I receptor that lacks a kinase domain, may also play a role in tumorigenesis (Onichtchouk et al., 1999; Tsang et al., 2000). BAMBI interacts with TGF-β receptors, disrupting ligand-induced receptor heteromerization and inhibiting TGF-β-dependent responses (Onichtchouk et al., 1999). Interestingly, nma, the mammalian ortholog of BAMBI, was identified from a differential display analysis in which it was downregulated in metastatic melanoma (Degen et al., 1996), suggesting that de-repression of TGF-β signaling following downregulation of nma expression could account for the increased metastatic potential of these cells.
B. Functional Implications of Smad Mutations Identified in Tumors Smad4 is located on chromosome 18q21.1, a locus with a particularly high frequency of deletion in pancreatic and colorectal carcinomas, and Smad4 was initially identified as a potential tumor suppressor gene in pancreatic carcinomas (Hahn et al., 1996). Although 90% of pancreatic tumors show allelic loss of chromosome 18q, only about 40% of the tumors display inactivation of Smad4. Smad4 mutations are frequently observed in advanced stages of human colorectal cancer (Zhou et al., 1998), and inactivation of both Smad4 alleles occurs in 95% of highly invasive and metastatic carcinomas that harbor Smad4 mutations (Miyaki et al., 1999). The importance of Smad4 in colon tumorigenesis is underscored by observations that Smad4 heterozygous mice develop gastric and duodenal polyps similar to those characteristic of human juvenile polyposis, an autosomal-dominant, inherited syndrome associated with hamartomatous polyps and increased risk for gastrointestinal cancer (Friedl et al., 1999; Howe et al., 1998; Taketo and Takaku, 2000). Other studies (Friedl et al., 1999) show that concomitant
38
Ester Piek and Anita B. Roberts
heterozygosity of Smad4 and APC strongly aggravates malignant progression of colon tumorigenesis (Takaku et al., 1998). Moreover, Smad4 heterozygous mice developed gastric carcinoma with loss of the remaining wild-type allele at late stages of tumorigenesis (Xu et al., 2000), similar to the finding of LOH of Smad4 at late stages in human colon carcinogenesis. It is important to note that in many cell lines lacking Smad4, such as Smad4-deficient Vaco-235 colon tumor cells or certain pancreatic carcinoma cells, TGF-β is still able to mediate growth inhibition (Fink et al., 2001; Dai et al., 1999). Although this could possibly occur through TGF-β-dependent activation of MAPK signaling (Mulder, 2000), recent data suggest that TGFβ-dependent G1 growth arrest might alternatively involve a parallel Smadindependent pathway emanating directly from TβRI via PP2A and resulting in inhibition of p70 S6 kinase (Petritsch et al., 2000; see Section II.C.3). It is important to note that reconstitution of Smad4 expression does not unequivocally restore the growth inhibitory response to TGF-β in Smad4deficient cells, such as in SW480.7 colon carcinoma cells that express hyperactive K-Ras or in Hs766T pancreatic adenocarcinoma cells (Calonge and Massague, 1999; Schwarte-Waldhoff et al., 2000). Possible explanations for these results include that TGF-β can mediate certain responses through other signal transduction pathways which are independent of Smad4, such as the S6 kinase pathway; that these cells express proteins with functional homology to Smad4, as recently identified in Xenopus laevis (Howell et al., 1999), although this is unlikely; and that other changes in the cells interfere with TGF-β-dependent growth control downstream of the Smad pathway. Multiple mutations have been identified in Smad4 and Smad2, which share the same chromosomal locus at 18q21; these are reviewed elsewhere (de Caestecker et al., 2000a; Hata et al., 1998b). These inactivating Smad mutations often involve residues important for Smad structure or intermolecular Smad–Smad interactions. Thus, mutations in the MH1 domain cause defective DNA binding, whereas mutations in the MH2 domain affect nuclear translocation and impair transcriptional activation properties (Hata et al., 1998b; Shi et al., 1997). Specific mechanisms underlying the inactivity of particular Smad2 and Smad4 mutants have been identified and represent interesting models that show that all aspects of Smad signaling can be affected. Thus, the corresponding N-terminal R133C Smad2 and R100T Smad4 missense mutations observed in colon and pancreatic carcinoma, respectively, enhance the affinity of the MH1 domain for the MH2 domain, thereby increasing the autoinhibitory interaction of these domains (Hata et al., 1997). Although Hata et al. (1997) and Moren et al. (2000) found strongly impaired activation and transcriptional activity of these Smad mutants, Xu and Attisano (2000) observed normal Smad activation and functioning but instead reported that these mutations render Smad2 and Smad4 more sensitive to ubiquitination and subsequent proteasomal degradation (Fig. 5; see color
Suppressor and Oncogenic Roles of TGF-β
39
insert). Mutation of G65V, R100T, and P130S in the MH1 domain of Smad4 leads to reduced binding to DNA and reduced protein stability, whereas L43S Smad4 shows decreased nuclear translocation and transcriptional activation (Moren et al., 2000). Six different mutations in Smad2 and Smad4, identified in lung tumors, all failed to transduce growth inhibitory signals following TGF-β stimulation and were incapable of eliciting certain transcriptional responses, correlating with impaired heteromerization with wild-type Smads and interaction with transcriptional coactivators (Yanagisawa et al., 2000). Interestingly, although Smad3 is located on chromosome 15q21–22, which is a frequent site of allelic loss in breast, colorectal, lung, and pancreatic tumors (Hahn et al., 1995; Park et al., 2000; Wick et al., 1996), mutation or deletion of Smad3 has not been found in any human cancer (Arai et al., 1998; Bevan et al., 1999; Roth et al., 1999; D. Wang et al., 1999). Expression levels of Smad proteins are altered in tumor cells. Since these pathways are stoichiometric, and since relative levels of Smad2 and Smad3 are expected to alter gene targets, these changes are likely to affect cell phenotypes. Thus, expression of Smad2–Smad4 was downregulated in epithelial cells of many skin tumors and in rat prostate tumors (Brodin et al., 1999; Lange et al., 1999). In contrast, immunohistochemical analysis of Smad protein expression in colon cancer specimens revealed that although the common mediator Smad4 and the inhibitory Smads showed similar expression patterns in tumor and normal tissue, receptor-activated Smads were induced in tumor tissues but were barely detectable in tumor stroma (Korchynskyi et al., 1999). Analysis of TGF-β, TGF-β receptor, and Smad expression levels in gliomas with different degrees of malignancy indicated that TGF-β 1 and TGF-β 2 expression was particularly enhanced in the most malignant glioblastoma multiforme grade IV, concomitant with increased TβRI and TβRII levels, but that levels of expression of Smad2–Smad4 were reduced (Kjellman et al., 2000). It is important to examine whether tissues in which Smad signaling is suggested to be downregulated exhibit a shift to TGF-β signal transduction via MAP kinase pathways, which might support a more aggressive phenotype. Smad6 and Smad7, which inhibit TGF-β signaling through an autocrine feedback loop (Hayashi et al., 1997; Imamura et al., 1997; Nakao et al., 1997) and which are also induced by cross talk with other signaling pathways including those dependent on MAP kinase, Stat1, and NF-κB (Afrakhte et al., 1998; Bitzer et al., 2000; Nagarajan et al., 2000; Ulloa et al., 1999), might be expected to have tumor promoting activity and to be overexpressed in tumor cells (Fig. 5; see color insert). Smads 6 and 7 show enhanced expression in pancreatic carcinoma cells, and their overexpression in COLO-357 cells increases the tumorigenicity of the cells by rendering them insensitive to the growth inhibitory effects of TGF-β but still enabling expression of genes that promote tumor metastasis
40
Ester Piek and Anita B. Roberts
and invasion (Kleeff et al., 1999a,b). Overexpression of Smad7 in erythroid leukemia cells prevents activin-induced differentiation, again supporting an oncogenic function for this inhibitory Smad (Kitamura et al., 2000). It remains to be seen whether inhibition of Smad signaling might lead to enhanced TGF-β signaling through MAP kinase pathways, possibly resulting in activation of AP-1 complexes (Keeton et al., 1991; Zhang et al., 1998).
C. Alterations in Smad-Interacting Proteins The TALE superclass homeodomain transcription corepressor TGIF inhibits TGF-β-induced transcriptional activation by interaction with receptoractivated Smad2 and Smad3, followed by recruitment of HDACs (Wotton et al., 1999a,b). TGIF competes with the coactivators p300/CBP to regulate Smad function at the promoter of TGF-β responsive genes. TGIF2, localized on chromosome 20q11.2–12 and overexpressed in ovarian cancers with amplification of region 20q, shows sequence homology to TGIF, but its function has yet to be identified. Although the TGF-β receptors or Smads are infrequently inactivated in ovarian cancer, it is tempting to speculate that, analogous to the role of TGIF in TGF-β signal transduction, upregulation of TGIF2 might render ovarian cells insensitive to the growth inhibitory action of TGF-β (Imoto et al., 2000). It is not known whether other inhibitors of TGF-β signaling, such as SNIP1 or SIP (R. H. Kim et al., 2000; Verschueren et al., 1999), are overexpressed in human cancers.
D. Regulation of TGF-β Signal Transduction Pathways by Oncogenes Apart from the tumor promoting effects that oncogenes exert directly, there are also indirect mechanisms by which they affect tumorigenesis, for example, by regulating the expression or activity of components of the TGF-β signal transduction pathway or by direct cross talk with the TGF-β signal transduction pathway. Mutational activation of MAP kinase pathways is frequently found in human cancers, most notably the activation of Ras, an upstream activator of the ERK1/2 pathway, which is an early event in many cancers (Akhurst and Balmain, 1999; Denhardt, 1996a; Imoto et al., 2000; Dumont, 1999). Oncogenic Ras and other components of MAP kinase signaling pathways have important effects on Smad signaling, as discussed in Sections II.C.2, IV.C, and IV.E. Certain other oncoproteins, such as the nuclear proteins Evi-1 (Kurokawa et al., 1998; Sood et al., 1999), Ski (Luo et al., 1999; Sun et al., 1999), and SnoN (Stroschein et al., 1999), interact directly with Smad proteins and
Suppressor and Oncogenic Roles of TGF-β
41
repress Smad signaling. However, in the case of the latter, there is a complex interplay between suppression by SnoN and Smad-mediated proteasomal degradation of SnoN which might be important in oncogenesis (Miyazono, 2000; Sun et al., 1999; Stroschein et al., 1999). Fusion proteins involving Evi-1 and c-Ski are expressed in various hematopoietic malignancies and may contribute to altered TGF-β responsiveness in these cancers (Kurokawa et al., 1998; Chaganti et al., 1986; Kees et al., 1990; Sood et al., 1999). Consistent with this finding, germline loss of Smad3 in T cells confers resistance to inhibition by TGF-β, possibly a result of the loss of Smad3-dependent repression of IL-2 by TGF-β in these cells (S. McKarns and J. Letterio, unpublished data). Together, these data suggest that either loss of Smad3 or blocking of Smad3 activity by oncogenes may contribute to the proliferative capacity of leukemic cells and that Smad3 may not confer an advantage to nonmetastatic tumor cells. In contrast, in hepatocellular carcinoma, which is closely associated with chronic hepatitis and cirrhosis, recent data suggest that the hepatitis B virusencoded oncoprotein pX, often expressed in liver of patients, selectively inhibits TGF-β-dependent apoptosis through activation of PI3 kinase signaling (Shih et al., 2000). The fact that pX has also been shown to amplify Smad-mediated signaling suggests a mechanism for the role of pX in disease progression (Lee et al., 2001). pX binds Smad4 and is presumed to act by enhancing the stability of the transcriptional complex by linking the Smad complex with TFIIB of the basal transcriptional machinery (Lee et al., 2001). Other oncogenes, such as v-Src and v-Abl, also affect TGF-β signaling by amplifying expression of the signaling receptors (Birchenall-Roberts et al., 1991). Clearly, expression of these oncogenes at later stages of tumorigenesis, when growth inhibitory effects of TGF-β have been lost by mechanisms involving alterations of either upstream or downstream elements of the signaling pathway, has the potential to enhance the tumor promoting (oncogenic) effects of TGF-β.
VI. SUMMARY It is beginning to be appreciated that complete loss of TGF-β responsivity is a relatively rare event in tumorigenesis, even though tumor cells frequently exhibit altered responsivity to TGF-β, having become insensitive either to its growth inhibitory effects or, in rare cases, to its proapoptotic effects. Rather, the comparatively small percentage of tumors in which either TGF-β receptors or Smad4 have been totally lost, as with biallelic loss or mutational inactivation, suggests that there is selective advantage in tumorigenesis to loss of the tumor suppressor activities of TGF-β but to retention of aspects of its
42
Ester Piek and Anita B. Roberts
signaling which promote expression of the transformed phenotype. Consistent with the tumor suppressor activities of TGF-β, experimental blockade of the TGF-β pathway enhances tumorigenesis, as seen in TGF-β 1 null mice, mice engineered to overexpress dnTβRII, Smad3 null mice, or mice heterozygous for Smad4. However, since TGF-β signaling is usually not completely eliminated in early stages of human carcinogenesis, it is far more likely that other changes in tumor cells, such as altered expression of Smad-interacting proteins or oncogenes, may interfere with downstream events in TGF-βdependent tumor suppressor pathways, allowing the prooncogenic effects of TGF-β to remain intact. In addition to these changes, altered patterns of expression of the TGF-β receptors or Smad proteins may also contribute to a shift toward signaling to prooncogenic gene targets or possibly even enable new signaling targets. Data suggest that the Ras–Raf–MAP kinase pathway, and possibly other MAP kinase pathways, synergizes with Smad3 to autoinduce expression of TGF-β 1 and to activate expression of AP1-dependent genes while simultaneously suppressing the inhibitory and apoptotic effects of TGF-β important in tumor suppression. Other oncogenes, such as Abl, Src, and pX, may further enhance signaling through these oncogenic pathways by inducing TGF-β receptor expression or enhancing Smad signaling. Thus, we propose that alternative “wiring” patterns of the signal transduction pathways underlie the tumor suppressor and prooncogenic activities of TGF-β (Fig. 7A; see color insert), such that suppressor pathways mediated in part by Smad2 and Smad3 and prominent in nonneoplastic cells give way to prooncogenic pathways mediated by Smad3 and MAP kinase pathways. The latter lead to activation of AP-1 and result in epithelial-to-mesenchymal transition of cells and more invasive, metastatic behavior (Fig. 7B; see color insert). Clearly, there are exceptions to this model, as in hematopoietic malignancies which lack invasive behavior. Regardless, the original designation of TGF-β as a transforming growth factor and a proximal effector of transformation has come full circle with new insights gained from understanding of its signal transduction pathways.
REFERENCES Adam, P. J., Regan, C. P., Hautmann, M. B., and Owens, G. K. (2000). J. Biol. Chem. 275, 37798–37806. Afrakhte, M., Moren, A., Jossan, S., Itoh, S., Sampath, K., Westermark, B., Heldin, C. H., Heldin, N. E., and ten Dijke, P. (1998). Biochem. Biophys. Res. Commun. 249, 505–511. Agrotis, A., Condron, M., and Bobik, A. (2000). FEBS Lett. 467, 128–132. Akhurst, R. J., and Balmain, A. (1999). J. Pathol. 187, 82–90. Albo, D., Berger, D. H., Vogel, J., and Tuszynski, G. P. (1999). J. Gastrointest. Surg. 3, 411– 417. Albo, D., Rothman, V. L., Roberts, D. D., and Tuszynski, G. P. (2000). Br. J. Cancer 83, 298– 306.
Suppressor and Oncogenic Roles of TGF-β
43
Alexandrow, M. G., and Moses, H. L. (1997). J. Cell. Biochem. 66, 427– 432. Alexandrow, M. G., Kawabata, M., Aakre, M., and Moses, H. L. (1995). Proc. Natl. Acad. Sci. USA 92, 3239–3243. Amendt, C., Schirmacher, P., Weber, H., and Blessing, M. (1998). Oncogene 17, 25–34. Amoroso, S. R., Huang, N., Roberts, A. B., Potter, M., and Letterio, J. J. (1998). Proc. Natl. Acad. Sci. USA 95, 189–194. Andreasen, P. A., Egelund, R., and Petersen, H. H. (2000). Cell. Mol. Life Sci. 57, 25– 40. Arai, T., Akiyama, Y., Okabe, S., Ando, M., Endo, M., and Yuasa, Y. (1998). Cancer Lett. 122, 157–163. Ariazi, E. A., Satomi, Y., Ellis, M. J., Haag, J. D., Shi, W., Sattler, C. A., and Gould, M. N. (1999). Cancer Res. 59, 1917–1928. Arnoletti, J. P., Albo, D., Granick, M. S., Solomon, M. P., Castiglioni, A., Rothman, V. L., and Tuszynski, G. P. (1995). Cancer 76, 998–1005. Arrick, B. A., Lopez, A. R., Elfman, F., Ebner, R., Damsky, C. H., and Derynck, R. (1992). J. Cell Biol. 118, 715–726. Arteaga, C. L., Hurd, S. D., Winnier, A. R., Johnson, M. D., Fendly, B. M., and Forbes, J. T. (1993a). J. Clin. Invest. 92, 2569–2576. Arteaga, C. L., Dugger, T. C., Winnier, A. R., and Forbes, J. T. (1993b). J. Cell Biochem. Suppl. 17G, 187–193. Atfi, A., Djelloul, S., Chastre, E., Davis, R., and Gespach, C. (1997). J. Biol. Chem. 272, 1429– 1432. Bakin, A. V., Tomlinson, A. K., Bhowmick, N. A., Moses, H. L., and Arteaga, C. L. (2000). J. Biol. Chem. 275, 36803–36810. Barbara, N. P., Wrana, J. L., and Letarte, M. (1999). J. Biol. Chem. 274, 584–594. Barrack, E. R. (1997). Prostate 31, 61–70. Bevan, S., Woodford-Richens, K., Rozen, P., Eng, C., Young, J., Dunlop, M., Neale, K., Phillips, R., Markie, D., Rodriguez-Bigas, M., Leggett, B., Sheridan, E., Hodgson, S., Iwama, T., Eccles, D., Bodmer, W., Houlston, R., and Tomlinson, I. (1999). Gut 45, 406– 408. Birchenall-Roberts, M. C., Falk, L. A., Kasper, J., Keller, J., Faltynek, C. R., and Ruscetti, F. W. (1991). J. Biol. Chem. 266, 9617–9621. Birkey-Reffey, S., Wurthner, J. U., Roberts, A. B., and Duckett, C. S. (2001). J. Biol. Chem., in press. Bitzer, M., von Gersdorff, G., Liang, D., Dominguez-Rosales, A., Beg, A. A., Rojkind, M., and Bottinger, E. P. (2000). Genes Dev. 14, 187–197. Blain, S. W., and Massague, J. (2000). J. Biol. Chem. 275, 32066–32070. Blobe, G. C., Schiemann, W. P., and Lodish, H. F. (2000). N. Engl. J. Med. 342, 1350–1358. Bodey, B., Bodey, B. J., Siegel, S. E., and Kaiser, H. E. (1998). Anticancer Res. 18, 2701–2710. Border, W. A., and Noble, N. A. (1994). N. Engl. J. Med. 331, 1286–1292. Bos, J. L. (1989). Cancer Res. 49, 4682– 4689. Bottinger, E. P., Jakubczak, J. L., Haines, D. C., Bagnall, K., and Wakefield, L. M. (1997). Cancer Res. 57, 5564–5570. Boyce, B. F., Yoneda, T., and Guise, T. A. (1999). Endocrinol. Relat. Cancer 6, 333–347. Brodin, G., ten Dijke, P., Funa, K., Heldin, C. H., and Landstrom, M. (1999). Cancer Res. 59, 2731–2738. Brodin, G., Ahgren, A., ten Dijke, P., Heldin, C. H., and Heuchel, R. (2000). J. Biol. Chem. 275, 29023–29030. Brown, J. D., DiChiara, M. R., Anderson, K. R., Gimbrone, M. A., Jr., and Topper, J. N. (1999). J. Biol. Chem. 274, 8797–8805. Burrows, F. J., Derbyshire, E. J., Tazzari, P. L., Amlot, P., Gazdar, A. F., King, S. W., Letarte, M., Vitetta, E. S., and Thorpe, P. E. (1995). Clin. Cancer Res. 1, 1623–1634. Cai, T., Lei, Q. Y., Wang, L. Y., and Zha, X. L. (2000). Biochem. Biophys. Res. Commun. 274, 519–525.
44
Ester Piek and Anita B. Roberts
Calonge, M. J., and Massague, J. (1999). J. Biol. Chem. 274, 33637–33643. Candia, A. F., Watabe, T., Hawley, S. H., Onichtchouk, D., Zhang, Y., Derynck, R., Niehrs, C., and Cho, K. W. (1997). Development 124, 4467– 4480. Caulin, C., Scholl, F. G., Frontelo, P., Gamallo, C., and Quintanilla, M. (1995). Cell Growth Differ. 6, 1027–1035. Chaganti, R. S., Balazs, I., Jhanwar, S. C., Murty, V. V., Koduru, P. R., Grzeschik, K. H., and Stavnezer, E. (1986). Cytogenet. Cell Genet. 43, 181–186. Chang, H., Huylebroeck, D., Verschueren, K., Guo, Q., Matzuk, M. M., and Zwijsen, A. (1999). Development 126, 1631–1642. Chen, R. H., Ebner, R., and Derynck, R. (1993). Science 260, 1335–1338. Chen, Y. G., Hata, A., Lo, R. S., Wotton, D., Shi, Y., Pavletich, N., and Massague, J. (1998). Genes Dev. 12, 2144–2152. Choi, M. E. (2000). Kidney Int. 58, 53–58. Choi, S. G., Yi, Y., Kim, Y. S., Kato, M., Chang, J., Chung, H. W., Hahm, K. B., Yang, H. K., Rhee, H. H., Bang, Y. J., and Kim, S. J. (1998). J. Biol. Chem. 273, 110 –117. Claassen, G. F., and Hann, S. R. (2000). Proc. Natl. Acad. Sci. USA 97, 9498–9503. Conrad, C. T., Ernst, N. R., Dummer, W., Brocker, E. B., and Becker, J. C. (1999). J. Exp. Clin. Cancer Res. 18, 225–232. Crawford, S. E., Stellmach, V., Murphy-Ullrich, J. E., Ribeiro, S. M., Lawler, J., Hynes, R. O., Boivin, G. P., and Bouck, N. (1998). Cell 93, 1159–1170. Cui, W., Kemp, C. J., Duffie, E., Balmain, A., and Akhurst, R. J. (1994). Cancer Res. 54, 5831–5836. Cui, W., Fowlis, D. J., Cousins, F. M., Duffie, E., Bryson, S., Balmain, A., and Akhurst, R. J. (1995). Genes Dev. 9, 945–955. Cui, W., Fowlis, D. J., Bryson, S., Duffie, E., Ireland, H., Balmain, A., and Akhurst, R. J. (1996). Cell 86, 531–542. Dai, J. L., Schutte, M., Bansal, R. K., Wilentz, R. E., Sugar, A. Y., and Kern, S. E. (1999). Mol. Carcinogen. 26, 37– 43. Datto, M. B., Frederick, J. P., Pan, L., Borton, A. J., Zhuang, Y., and Wang, X. F. (1999). Mol. Cell. Biol. 19, 2495–2504. de Caestecker, M. P., Parks, W. T., Frank, C. J., Castagnino, P., Bottaro, D. P., Roberts, A. B., and Lechleider, R. J. (1998). Genes Dev. 12, 1587–1592. de Caestecker, M. P., Piek, E., and Roberts, A. B. (2000a). J. Natl. Cancer Inst. 92, 1388–1402. de Caestecker, M. P., Yahata, T., Wang, D., Parks, W. T., Huang, S., Hill, C. S., Shioda, T., Roberts, A. B., and Lechleider, R. J. (2000b). J. Biol. Chem. 275, 2115–2122. Degen, W. G., Weterman, M. A., van Groningen, J. J., Cornelissen, I. M., Lemmers, J. P., Agterbos, M. A., Geurts, V. K., Swart, G. W., and Bloemers, H. P. (1996). Int. J. Cancer 65, 460– 465. Denhardt, D. T. (1996a). Biochem. J. 318, 729–747. Denhardt, D. T. (1996b). Crit. Rev. Oncogene. 7, 261–291. Derynck, R., and Feng, X. H. (1997). Biochim. Biophys. Acta 1333, F105–F150. Derynck, R., Zhang, Y., and Feng, X. H. (1998). Cell 95, 737–740. Devi, G. R., De Souza, A. T., Byrd, J. C., Jirtle, R. L., and MacDonald, R. G. (1999). Cancer Res. 59, 4314– 4319. de Visser, K. E., and Kast, W. M. (1999). Leukemia 13, 1188–1199. Dickson, M. C., Martin, J. S., Cousins, F. M., Kulkarni, A. B., Karlsson, S., and Akhurst, R. J. (1995). Development 121, 1845–1854. Doran, T., Stuhlmiller, H., Kim, J. A., Martin, E. W., Jr., and Triozzi, P. L. (1997). J. Immunother. 20, 372–376. Dumont, N. (1999). APMIS 107, 711–722.
Suppressor and Oncogenic Roles of TGF-β
45
Engel, M. E., McDonnell, M. A., Law, B. K., and Moses, H. L. (1999). J. Biol. Chem. 274, 37413–37420. Engle, S. J., Hoying, J. B., Boivin, G. P., Ormsby, I., Gartside, P. S., and Doetschman, T. (1999). Cancer Res. 59, 3379–3386. Fakhrai, H., Dorigo, O., Shawler, D. L., Lin, H., Mercola, D., Black, K. L., Royston, I., and Sobol, R. E. (1996). Proc. Natl. Acad. Sci. USA 93, 2909–2914. Farina, A. R., Coppa, A., Tiberio, A., Tacconelli, A., Turco, A., Colletta, G., Gulino, A., and Mackay, A. R. (1998). Int. J. Cancer 75, 721–730. Fearon, E. R., and Vogelstein, B. (1990). Cell 61, 759–767. Feng, X. H., Lin, X., and Derynck, R. (2000). EMBO J. 19, 5178–5193. Festuccia, C., Bologna, M., Gravina, G. L., Guerra, F., Angelucci, A., Villanova, I., Millimaggi, D., and Teti, A. (1999). Int. J. Cancer 81, 395– 403. Festuccia, C., Angelucci, A., Gravina, G. L., Villanova, I., Teti, A., Abini, A., and Bologna, M. (2000). Int. J. Cancer 85, 407– 415. Fink, S. P., Swinler, S. E., Lutterbaugh, J. D., Massagu´e, J., Thiagalingam, S., Kinzler, K. W., Vogelstein, B., Willson, J. K. V., and Markowitz, S. (2001). Cancer Res., 61, 256– 260. Foekens, J. A., Schmitt, M., van Putten, W. L., Peters, H. A., Bontenbal, M., Janicke, F., and Klijn, J. G. (1992). Cancer Res. 52, 6101–6105. Folkman, J., and D’Amore, P. A. (1996). Cell 87, 1153–1155. Fortunel, N. O., Hatzfeld, A., and Hatzfeld, J. A. (2000). Blood 96, 2022–2036. Frazier, K. S., and Grotendorst, G. R. (1997). Int. J. Biochem. Cell Biol. 29, 153–161. Friedl, W., Kruse, R., Uhlhaas, S., Stolte, M., Schartmann, B., Keller, K. M., Jungck, M., Stern, M., Loff, S., Back, W., Propping, P., and Jenne, D. E. (1999). Genes Chromosomes Cancer 25, 403– 406. Geiser, A. G., Burmester, J. K., Webbink, R., Roberts, A. B., and Sporn, M. B. (1992). J. Biol. Chem. 267, 2588–2593. Glick, A., Popescu, N., Alexander, V., Ueno, H., Bottinger, E., and Yuspa, S. H. (1999). Proc. Natl. Acad. Sci. USA 96, 14949–14954. Glick, A. B., Kulkarni, A. B., Tennenbaum, T., Hennings, H., Flanders, K. C., O’Reilly, M., Sporn, M. B., Karlsson, S., and Yuspa, S. H. (1993). Proc. Natl. Acad. Sci. USA 90, 6076– 6080. Glick, A. B., Lee, M. M., Darwiche, N., Kulkarni, A. B., Karlsson, S., and Yuspa, S. H. (1994). Genes Dev. 8, 2429–2440. Glick, A. B., Weinberg, W. C., Wu, I. H., Quan, W., and Yuspa, S. H. (1996). Cancer Res. 56, 3645–3650. Go, C., Li, P., and Wang, X. J. (1999). Cancer Res. 59, 2861–2868. Go, C., He, W., Zhong, L., Li, P., Huang, J., Brinkley, B. R., and Wang, X. J. (2000). Oncogene 19, 3623–3631. Godar, S., Horejsi, V., Weidle, U. H., Binder, B. R., Hansmann, C., and Stockinger, H. (1999). Eur. J. Immunol. 29, 1004 –1013. Goggins, M., Shekher, M., Turnacioglu, K., Yeo, C. J., Hruban, R. H., and Kern, S. E. (1998). Cancer Res. 58, 5329–5332. Gold, L. I. (1999). Crit. Rev. Oncogene 10, 303–360. Gold, L. I., Jussila, T., Fusenig, N. E., and Stenback, F. (2000). J. Pathol. 190, 579–588. Goltzman, D. (1997). Cancer 80, 1581–1587. Goumans, M. J., and Mummery, C. (2000). Int. J. Dev. Biol. 44, 253–265. Grady, W. M., Myeroff, L. L., Swinler, S. E., Rajput, A., Thiagalingam, S., Lutterbaugh, J. D., Neumann, A., Brattain, M. G., Chang, J., Kim, S. J., Kinzler, K. W., Vogelstein, B., Willson, J. K., and Markowitz, S. (1999). Cancer Res. 59, 320 –324. Grainger, D. J., Metcalfe, J. C., Grace, A. A., and Mosedale, D. E. (1998). J. Cell Sci. 111, 2977–2988.
46
Ester Piek and Anita B. Roberts
Griswold-Prenner, I., Kamibayashi, C., Maruoka, E. M., Mumby, M. C., and Derynck, R. (1998). Mol. Cell. Biol. 18, 6595–6604. Grotendorst, G. R. (1997). Cytokine Growth Factor Rev. 8, 171–179. Guise, T. A. (2000). Cancer 88, 2892–2898. Guise, T. A., Yin, J. J., Taylor, S. D., Kumagai, Y., Dallas, M., Boyce, B. F., Yoneda, T., and Mundy, G. R. (1996). J. Clin. Invest. 98, 1544 –1549. Guo, Y., and Kyprianou, N. (1999). Cancer Res. 59, 1366–1371. Hahm, K. B., Cho, K., Lee, C., Im, Y. H., Chang, J., Choi, S. G., Sorensen, P. H., Thiele, C. J., and Kim, S. J. (1999). Nature Genet. 23, 222–227. Hahn, S. A., Seymour, A. B., Hoque, A. T., Schutte, M., da Costa, L. T., Redston, M. S., Caldas, C., Weinstein, C. L., Fischer, A., and Yeo, C. J. (1995). Cancer Res. 55, 4670– 4675. Hahn, S. A., Schutte, M., Hoque, A. T., Moskaluk, C. A., da Costa, L. T., Rozenblum, E., Weinstein, C. L., Fischer, A., Yeo, C. J., Hruban, R. H., and Kern, S. E. (1996). Science 271, 350–353. Hanafusa, H., Ninomiya-Tsuji, J., Masuyama, N., Nishita, M., Fujisawa, J., Shibuya, H., Matsumoto, K., and Nishida, E. (1999). J. Biol. Chem. 274, 27161–27167. Hartsough, M. T., and Mulder, K. M. (1995). J. Biol. Chem. 270, 7117–7124. Hartsough, M. T., and Mulder, K. M. (1997). Pharmacol. Ther. 75, 21– 41. Hata, A., Lo, R. S., Wotton, D., Lagna, G., and Massague, J. (1997). Nature 388, 82–87. Hata, A., Lagna, G., Massague, J., and Hemmati-Brivanlou, A. (1998a). Genes Dev. 12, 186–197. Hata, A., Shi, Y., and Massague, J. (1998b). Mol. Med. Today 4, 257–262. Hautmann, M. B., Madsen, C. S., and Owens, G. K. (1997). J. Biol. Chem. 272, 10948–10956. Hay, E. D. (1995). Acta Anat. 154, 8–20. Hayashi, H., Abdollah, S., Qiu, Y., Cai, J., Xu, Y. Y., Grinnell, B. W., Richardson, M. A., Topper, J. N., Gimbrone, M. A., Jr., Wrana, J. L., and Falb, D. (1997). Cell 89, 1165–1173. Heino, J., and Massague, J. (1989). J. Biol. Chem. 264, 21806–21811. Heyer, J., Escalante-Alcalde, D., Lia, M., Boettinger, E., Edelmann, W., Stewart, C. L., and Kucherlapati, R. (1999). Proc. Natl. Acad. Sci. USA 96, 12595–12600. Hirota, Y., Tsukazaki, T., Yonekura, A., Miyazaki, Y., Osaki, M., Shindo, H., and Yamashita, S. (2000). Osteoarthritis Cartilage 8, 241–247. Hocevar, B. A., Brown, T. L., and Howe, P. H. (1999). EMBO J. 18, 1345–1356. Hojo, M., Morimoto, T., Maluccio, M., Asano, T., Morimoto, K., Lagman, M., Shimbo, T., and Suthanthiran, M. (1999). Nature 397, 530–534. Hougaard, S., Norgaard, P., Abrahamsen, N., Moses, H. L., Spang-Thomsen, M., and Skovgaard, P. H. (1999). Br. J. Cancer 79, 1005–1011. Howe, J. R., Roth, S., Ringold, J. C., Summers, R. W., Jarvinen, H. J., Sistonen, P., Tomlinson, I. P., Houlston, R. S., Bevan, S., Mitros, F. A., Stone, E. M., and Aaltonen, L. A. (1998). Science 280, 1086 –1088. Howell, M., Itoh, F., Pierreux, C. E., Valgeirsdottir, S., Itoh, S., ten Dijke, P., and Hill, C. S. (1999). Dev. Biol. 214, 354 –369. Hsing, A. Y., Kadomatsu, K., Bonham, M. J., and Danielpour, D. (1996). Cancer Res. 56, 5146–5149. Hu, P. P., Shen, X., Huang, D., Liu, Y., Counter, C., and Wang, X. F. (1999). J. Biol. Chem. 274, 35381–35387. Hueber, A. O., and Evan, G. I. (1998). Trends Genet. 14, 364 –367. Ide, H., Katoh, M., Sasaki, H., Yoshida, T., Aoki, K., Nawa, Y., Osada, Y., Sugimura, T., and Terada, M. (1997). Oncogene 14, 1377–1382. Im, Y. H., Kim, H. T., Lee, C., Poulin, D., Welford, S., Sorensen, P. H., Denny, C. T., and Kim, S. J. (2000). Cancer Res. 60, 1536–1540. Imamura, T., Takase, M., Nishihara, A., Oeda, E., Hanai, J., Kawabata, M., and Miyazono, K. (1997). Nature 389, 622–626.
Suppressor and Oncogenic Roles of TGF-β
47
Imoto, I., Pimkhaokham, A., Watanabe, T., Saito-Ohara, F., Soeda, E., and Inazawa, J. (2000). Biochem. Biophys. Res. Commun. 276, 264 –270. Itoh, S., Landstrom, M., Hermansson, A., Itoh, F., Heldin, C. H., Heldin, N. E., and ten Dijke, P. (1998). J. Biol. Chem. 273, 29195–29201. Jackson, R. J., Antonia, S. J., Wright, K. L., Moon, N. S., Nepveu, A., and Mu (1999). Arch. Biochem. Biophys. 371, 290–300. Jirtle, R. L., Haag, J. D., Ariazi, E. A., and Gould, M. N. (1993). Cancer Res. 53, 3849–3852. Johansson, N., Vaalamo, M., Grenman, S., Hietanen, S., Klemi, P., Saarialho-Kere, U., and Kahari, V. M. (1999). Am. J. Pathol. 154, 469– 480. Johansson, N., Ala-aho, R., Uitto, V., Grenman, R., Fusenig, N. E., Lopez-Otin, C., and Kahari, V. M. (2000). J. Cell Sci. 113, 227–235. Johnson, D. W., Berg, J. N., Baldwin, M. A., Gallione, C. J., Marondel, I., Yoon, S. J., Stenzel, T. T., Speer, M., Pericak-Vance, M. A., Diamond, A., Guttmacher, A. E., Jackson, C. E., Attisano, L., Kucherlapati, R., Porteous, M. E., and Marchuk, D. A. (1996). Nature Genet. 13, 189–195. Jones, P. L., and Wolffe, A. P. (1999). Semin. Cancer Biol. 9, 339–347. Jonk, L. J., Itoh, S., Heldin, C. H., ten Dijke, P., and Kruijer, W. (1998). J. Biol. Chem. 273, 21145–21152. Jung, C. W., Kim, E. S., Seol, J. G., Park, W. H., Lee, S. J., Kim, B. K., and Lee, Y. Y. (1999). Leukocyte Res. 23, 1105–1112. Kaminski, N., Allard, J. D., Pittet, J. F., Zuo, F., Griffiths, M. J., Morris, D., Huang, X., Sheppard, D., and Heller, R. A. (2000). Proc. Natl. Acad. Sci. USA 97, 1778–1783. Kang, S. H., Bang, Y. J., Im, Y. H., Yang, H. K., Lee, D. A., Lee, H. Y., Lee, H. S., Kim, N. K., and Kim, S. J. (1999). Oncogene 18, 7280–7286. Kees, U. R., Lukeis, R., Ford, J., Willoughby, M. L., and Garson, O. M. (1990). Cancer Genet. Cytogenet. 46, 201–208. Keeton, M. R., Curriden, S. A., van Zonneveld, A. J., and Loskutoff, D. J. (1991). J. Biol. Chem. 266, 23048–23052. Kehrl, J. H., Wakefield, L. M., Roberts, A. B., Jakowlew, S., Alvarez-Mon, M., Derynck, R., Sporn, M. B., and Fauci, A. S. (1986). J. Exp. Med. 163, 1037–1050. Kehrl, J. H., Taylor, A., Kim, S. J., and Fauci, A. S. (1991). Ann. N.Y. Acad. Sci. 628, 345–353. Keski-Oja, J., Raghow, R., Sawdey, M., Loskutoff, D. J., Postlethwaite, A. E., Kang, A. H., and Moses, H. L. (1988). J. Biol. Chem. 263, 3111–3115. Keski-Oja, J., Koli, K., Lohi, J., and Laiho, M. (1991). Semin. Thromb. Hemost. 17, 231–239. Khokha, R. (1994). J. Natl. Cancer Inst. 86, 299–304. Kim, D. H., Chang, J. H., Lee, K. H., Lee, H. Y., and Kim, S. J. (1997). J. Biol. Chem. 272, 688–694. Kim, R. H., Wang, D., Tsang, M., Martin, J., Huff, C., de Caestecker, M. P., Parks, W. T., Meng, X., Lechleider, R. J., Wang, T., and Roberts, A. B. (2000). Genes Dev. 14, 1605–1616. Kim, S. J., Angel, P., Lafyatis, R., Hattori, K., Kim, K. Y., Sporn, M. B., Karin, M., and Roberts, A. B. (1990). Mol. Cell. Biol. 10, 1492–1497. Kim, S. J., Im, Y. H., Markowitz, S. D., and Bang, Y. J. (2000). Cytokine Growth Factor Rev. 11, 159–168. Kingsley, D. M. (1994). Genes Dev. 8, 133–146. Kiriyama, T., Gillespie, M. T., Glatz, J. A., Fukumoto, S., Moseley, J. M., and Martin, T. J. (1993). Mol. Cell. Endocrinol. 92, 55–62. Kitamura, K., Aota, S., Sakamoto, R., Yoshikawa, S. I., and Okazaki, K. (2000). Blood 95, 3371–3379. Kjellman, C., Olofsson, S. P., Hansson, O., Von Schantz, T., Lindvall, M., Nilsson, I., Salford, L. G., Sjogren, H. O., and Widegren, B. (2000). Int. J. Cancer 89, 251–258. Kleeff, J., Maruyama, H., Friess, H., Buchler, M. W., Falb, D., and Korc, M. (1999a). Biochem. Biophys. Res. Commun. 255, 268–273.
48
Ester Piek and Anita B. Roberts
Kleeff, J., Ishiwata, T., Maruyama, H., Friess, H., Truong, P., Buchler, M. W., Falb, D., and Korc, M. (1999b). Oncogene 18, 5363–5372. Knaus, P. I., Lindemann, D., DeCoteau, J. F., Perlman, R., Yankelev, H., Hille, M., Kadin, M. E., and Lodish, H. F. (1996). Mol. Cell. Biol. 16, 3480–3489. Koeneman, K. S., Yeung, F., and Chung, L. W. (1999). Prostate 39, 246–261. Kojima, S., and Rifkin, D. B. (1993). J. Cell. Physiol. 155, 323–332. Koli, K., Lohi, J., Hautanen, A., and Keski-Oja, J. (1991). Eur. J. Biochem. 199, 337–345. Korchynskyi, O., Landstrom, M., Stoika, R., Funa, K., Heldin, C. H., ten Dijke, P., and Souchelnytskyi, S. (1999). Int. J. Cancer 82, 197–202. ¨ Kordula, T., Guttgemann, I., Rose-John, S., Roeb, E., Osthues, A., Tschesche, H., Koj, A., Heinrich, P. C., and Graeve, L. (1992). FEBS Lett. 313, 143–147. Kretzschmar, M., Doody, J., and Massague, J. (1997). Nature 389, 618– 622. Kretzschmar, M., Doody, J., Timokhina, I., and Massague, J. (1999). Genes Dev. 13, 804 –816. Krymskaya, V. P., Hoffman, R., Eszterhas, A., Ciocca, V., and Panettieri, R. A., Jr. (1997). Am. J. Physiol. 273, L1220–L1227. Kubota, S., Fridman, R., and Yamada, Y. (1991). Biochem. Biophys. Res. Commun. 176, 129– 136. Kumar, S., Ghellal, A., Li, C., Byrne, G., Haboubi, N., Wang, J. M., and Bundred, N. (1999). Cancer Res. 59, 856–861. Kurokawa, M., Mitani, K., Imai, Y., Ogawa, S., Yazaki, Y., and Hirai, H. (1998). Blood 92, 4003– 4012. Kustikova, O., Kramerov, D., Grigorian, M., Berezin, V., Bock, E., Lukanidin, E., and Tulchinsky, E. (1998). Mol. Cell. Biol. 18, 7095–7105. Laiho, M., Saksela, O., and Keski-Oja, J. (1987). J. Biol. Chem. 262, 17467–17474. Laiho, M., DeCaprio, J. A., Ludlow, J. W., Livingston, D. M., and Massague, J. (1990). Cell 62, 175–185. Lamb, R. F., Hennigan, R. F., Turnbull, K., Katsanakis, K. D., MacKenzie, E. D., Birnie, G. D., and Ozanne, B. W. (1997). Mol. Cell. Biol. 17, 963–976. Landstrom, M., Heldin, N. E., Bu, S., Hermansson, A., Itoh, S., Dijke, P., and Heldin, C. H. (2000). Curr. Biol. 10, 535–538. Lange, D., Persson, U., Wollina, U., ten Dijke, P., Castelli, E., Heldin, C. H., and Funa, K. (1999). Int. J. Oncol. 14, 1049–1056. Larisch, S., Yi, Y., Lotan, R., Kerner, H., Eimerl, S., Tony, P. W., Gottfried, Y., Birkey, R. S., de Caestecker, M. P., Danielpour, D., Book-Melamed, N., Timberg, R., Duckett, C. S., Lechleider, R. J., Steller, H., Orly, J., Kim, S. J., and Roberts, A. B. (2000). Nat. Cell Biol. 2, 915–921. Larisch-Bloch, S., Danielpour, D., Roche, N. S., Lotan, R., Hsing, A. Y., Kerner, H., Hajouj, T., Lechleider, R. J., and Roberts, A. B. (2000). Cell Growth Differ. 11, 1–10. Lechleider, R. J., Ryan, J. L., Garrett, L., Eng, C., Deng, C., Wynshaw-Boris, A., and Roberts, A. B. (2001). Submitted for publication. Lee, D. K., Park, S. H., Yi, Y., Choi, S.-G., Lee, C., Parks, T., Cho, H., de Caestecker, M., Shaul, Y., Roberts, A. B., and Kim, S.-J. (2001). Genes Dev., 15, 455–466. Lehmann, K., Janda, E., Pierreux, C. E., Rytomaa, M., Schulze, A., McMahon, M., Hill, C. S., Beug, H., and Downward, J. (2000). Genes Dev. 14, 2610–2622. Letamendia, A., Lastres, P., Botella, L. M., Raab, U., Langa, C., Velasco, B., Attisano, L., and Bernabeu, C. (1998). J. Biol. Chem. 273, 33011–33019. Letterio, J. J., and Roberts, A. B. (1998). Annu. Rev. Immunol. 16, 137–161. Letterio, J., Guarino, J., Lucero, T., Anver, M., Deng, C., and Roberts, A. B. (1999). Proc. Am. Assoc. Cancer Res. 40, 200–201. Li, C., Guo, B., Wilson, P. B., Stewart, A., Byrne, G., Bundred, N., and Kumar, S. (2000a). Int. J. Cancer 89, 122–126.
Suppressor and Oncogenic Roles of TGF-β
49
Li, C., Hampson, I. N., Hampson, L., Kumar, P., Bernabeu, C., and Kumar, S. (2000b). FASEB J. 14, 55–64. Li, D. Y., Sorensen, L. K., Brooke, B. S., Urness, L. D., Davis, E. C., Taylor, D. G., Boak, B. B., and Wendel, D. P. (1999). Science 284, 1534–1537. Liberati, N. T., Datto, M. B., Frederick, J. P., Shen, X., Wong, C., Rougier-Chapman, E. M., and Wang, X. F. (1999). Proc. Natl. Acad. Sci. USA 96, 4844– 4849. Lin, J. K., and Chou, C. K. (1992). Cancer Res. 52, 385–388. Liu, X., Yue, J., Frey, R. S., Zhu, Q., and Mulder, K. M. (1998). Cancer Res. 58, 4752– 4757 Liu, X., Sun, Y., Ehrlich, M., Lu, T., Kloog, Y., Weinberg, R. A., Lodish, H. F., and Henis, Y. I. (2000). Oncogene 19, 5926–5935. Lu, S. L., Kawabata, M., Imamura, T., Miyazono, K., and Yuasa, Y. (1999). Biochem. Biophys. Res. Commun. 259, 385–390. Luo, K., Stroschein, S. L., Wang, W., Chen, D., Martens, E., Zhou, S., and Zhou, Q. (1999). Genes Dev. 13, 2196–2206. Madri, J. A., Pratt, B. M., and Tucker, A. M. (1988). J. Cell Biol. 106, 1375–1384. Majack, R. A., Majesky, M. W., and Goodman, L. V. (1990). J. Cell Biol. 111, 239–247. Markowitz, S. (2000). Biochim. Biophys. Acta 1470, M13–M20. Markowitz, S. D., and Roberts, A. B. (1996). Cytokine Growth Factor Rev. 7, 93–102. Markowitz, S., Wang, J., Myeroff, L., Parsons, R., Sun, L., Lutterbaugh, J., Fan, R. S., Zborowska, E., Kinzler, K. W., and Vogelstein, B. (1995). Science 268, 1336–1338. Massague, J. (1998). Annu. Rev. Biochem. 67, 753–791. Massague, J., and Wotton, D. (2000). EMBO J. 19, 1745–1754. Massague, J., Blain, S. W., and Lo, R. S. (2000). Cell 103, 295–309. Matsuno, F., Haruta, Y., Kondo, M., Tsai, H., Barcos, M., and Seon, B. K. (1999). Clin. Cancer Res. 5, 371–382. McAllister, K. A., Grogg, K. M., Johnson, D. W., Gallione, C. J., Baldwin, M. A., Jackson, C. E., Helmbold, E. A., Markel, D. S., McKinnon, W. C., and Murrell, J. (1994). Nature Genet. 8, 345–351. McEarchern, J. A., Besselsen, D. G., and Akporiaye, E. T. (1999). Cancer Immunol. Immunother. 48, 63–70. Merogi, A. J., Marrogi, A. J., Ramesh, R., Robinson, W. R., Fermin, C. D., and Freeman, S. M. (1997). Hum. Pathol. 28, 321–331. Merryman, J. I., DeWille, J. W., Werkmeister, J. R., Capen, C. C., and Rosol, T. J. (1994). Endocrinology 134, 2424 –2430. Merwin, J. R., Anderson, J. M., Kocher, O., Van Itallie, C. M., and Madri, J. A. (1990). J. Cell. Physiol. 142, 117–128. Miettinen, P. J., Ebner, R., Lopez, A. R., and Derynck, R. (1994). J. Cell Biol. 127, 2021–2036. Miyaki, M., Iijima, T., Konishi, M., Sakai, K., Ishii, A., Yasuno, M., Hishima, T., Koike, M., Shitara, N., Iwama, T., Utsunomiya, J., Kuroki, T., and Mori, T. (1999). Oncogene 18, 3098–3103. Miyazono, K. (2000). Biol. Pharm. Bull. 23, 1125–1130. Moren, A., Itoh, S., Moustakas, A., Dijke, P., and Heldin, C. H. (2000). Oncogene 19, 4396– 4404. Morton, D. M., and Barrack, E. R. (1995). Cancer Res. 55, 2596–2602. Moustakas, A., and Kardassis, D. (1998). Proc. Natl. Acad. Sci. USA 95, 6733–6738. Mulder, K. M. (2000). Cytokine Growth Factor Rev. 11, 23–35. Mule, J. J., Schwarz, S. L., Roberts, A. B., Sporn, M. B., and Rosenberg, S. A. (1988). Cancer Immunol. Immunother. 26, 95–100. Muller, G., Behrens, J., Nussbaumer, U., Bohlen, P., and Birchmeier, W. (1987). Proc. Natl. Acad. Sci. USA 84, 5600–5604.
50
Ester Piek and Anita B. Roberts
Munger, J. S., Huang, X., Kawakatsu, H., Griffiths, M. J., Dalton, S. L., Wu, J., Pittet, J. F., Kaminski, N., Garat, C., Matthay, M. A., Rifkin, D. B., and Sheppard, D. (1999). Cell 96, 319–328. Munger, K., Pietenpol, J. A., Pittelkow, M. R., Holt, J. T., and Moses, H. L. (1992). Cell Growth Differ. 3, 291–298. Munoz-Antonia, T., Li, X., Reiss, M., Jackson, R., and Antonia, S. (1996). Cancer Res. 56, 4831– 4835. Murphy-Ullrich, J. E., and Poczatek, M. (2000). Cytokine Growth Factor Rev. 11, 59– 69. Murray, P. A., Barrett-Lee, P., Travers, M., Luqmani, Y., Powles, T., and Coombes, R. C. (1993). Br. J. Cancer 67, 1408–1412. Nabel, E. G., Shum, L., Pompili, V. J., Yang, Z. Y., San, H., Shu, H. B., Liptay, S., Gold, L., Gordon, D., Derynck, R., et al. (1993). Proc. Natl. Acad. Sci. USA 90, 10759–10763. Nagarajan, R. P., Chen, F., Li, W., Vig, E., Harrington, M. A., Nakshatri, H., and Chen, Y. (2000). Biochem. J. 348, 591–596. Nakao, A., Afrakhte, M., Moren, A., Nakayama, T., Christian, J. L., Heuchel, R., Itoh, S., Kawabata, M., Heldin, N. E., Heldin, C. H., and ten Dijke, P. (1997). Nature 389, 631–635. Nunes, I., Munger, J., Harpel, J. G., Nagano, Y., Shapiro, R., Gleizes, P. E., and Rifkin, D. B. (1998). J. Am. Optom. Assoc. 69, 643– 648. Odekon, L. E., Blasi, F., and Rifkin, D. B. (1994). J. Cell. Physiol. 158, 398–407. Oft, M., Peli, J., Rudaz, C., Schwarz, H., Beug, H., and Reichmann, E. (1996). Genes Dev. 10, 2462–2477. Oft, M., Heider, K. H., and Beug, H. (1998). Curr. Biol. 8, 1243–1252. Oh, S. P., Seki, T., Goss, K. A., Imamura, T., Yi, Y., Donahoe, P. K., Li, L., Miyazono, K., ten Dijke, P., Kim, S., and Li, E. (2000). Proc. Natl. Acad. Sci. USA 97, 2626–2631. Okamoto, A., Jiang, W., Kim, S. J., Spillare, E. A., Stoner, G. D., Weinstein, I. B., and Harris, C. C. (1994). Proc. Natl. Acad. Sci. USA 91, 11576–11580. Onichtchouk, D., Chen, Y. G., Dosch, R., Gawantka, V., Delius, H., Massague, J., and Niehrs, C. (1999). Nature 401, 480– 485. Orsatti, G., Hytiroglou, P., Thung, S. N., Ishak, K. G., and Paronetto, F. (1997). Liver 17, 152–156. Oshima, M., Oshima, H., and Taketo, M. M. (1996). Dev. Biol. 179, 297–302. Owens, G. K. (1998). Acta Physiol. Scand. 164, 623– 635. Pages, G., Milanini, J., Richard, D. E., Berra, E., Gothie, E., Vinals, F., and Pouyssegur, J. (2000). Ann. N.Y. Acad. Sci. 902, 187–200. Paget, S. (1889). Lancet 1, 571–573. Pardali, K., Kurisaki, A., Moren, A., Dijke, P., Kardassis, D., and Moustakas, A. (2000). J. Biol. Chem. 275, 29244–29256. Park, J. A., Wang, E., Kurt, R. A., Schluter, S. F., Hersh, E. M., and Akporiaye, E. T. (1997). Cancer Gene Ther. 4, 42–50. Park, W. S., Park, J. Y., Oh, R. R., Yoo, N. J., Lee, S. H., Shin, M. S., Lee, H. K., Han, S., Yoon, S. K., Kim, S. Y., Choi, C., Kim, P. J., Oh, S. T., and Lee, J. Y. (2000). Cancer Res. 60, 70 –73. Patil, S., Wildey, G. M., Brown, T. L., Choy, L., Derynck, R., and Howe, P. H. (2000). J. Biol. Chem. 275, 38363–38370. Pepper, M. S. (1997). Cytokine Growth Factor Rev. 8, 21– 43. Pepper, M. S., Vassalli, J. D., Orci, L., and Montesano, R. (1993). Exp. Cell Res. 204, 356–363 Peron, P., Rahmani, M., Zagar, Y., Durand-Schneider, A. M., Lardeux, B., and Bernuau, D. (2001). J. Biol. Chem., 276, 10524–10531. Petritsch, C., Beug, H., Balmain, A., and Oft, M. (2000). Genes Dev. 14, 3093–3101. Piek, E., Heldin, C. H., and ten Dijke, P. (1999a). FASEB J. 13, 2105–2124.
Suppressor and Oncogenic Roles of TGF-β
51
Piek, E., Moustakas, A., Kurisaki, A., Heldin, C. H., and ten Dijke, P. (1999b). J. Cell Sci. 112, 4557– 4568. Piek, E., Ju, W. J., Heyer, J., Escalante-Alcalde, D., Stewart, C. L., Weinstein, M., Deng, ¨ C., Kucherlapati, R., Bottinger, E. P., and Roberts, A. B. (2001). J. Biol. Chem. 276, 19945– 19953. Pierce, D. F., Jr., Johnson, M. D., Matsui, Y., Robinson, S. D., Gold, L. I., Purchio, A. F., Daniel, C. W., Hogan, B. L., and Moses, H. L. (1993). Genes Dev. 7, 2308–2317. Pierce, D. F., Jr., Gorska, A. E., Chytil, A., Meise, K. S., Page, D. L., Coffey, R. J., Jr., and Moses, H. L. (1995). Proc. Natl. Acad. Sci. USA 92, 4254– 4258. Portella, G., Cumming, S. A., Liddell, J., Cui, W., Ireland, H., Akhurst, R. J., and Balmain, A. (1998). Cell Growth Differ. 9, 393– 404. Powell, G. J., Southby, J., Danks, J. A., Stillwell, R. G., Hayman, J. A., Henderson, M. A., Bennett, R. C., and Martin, T. J. (1991). Cancer Res. 51, 3059–3061. Proetzel, G., Pawlowski, S. A., Wiles, M. V., Yin, M., Boivin, G. P., Howles, P. N., Ding, J., Ferguson, M. W., and Doetschman, T. (1995). Nature Genet. 11, 409– 414. Ranges, G. E., Figari, I. S., Espevik, T., and Palladino, M. A., Jr. (1987). J. Exp. Med. 166, 991–998. Ravanti, L., Heino, J., Lopez-Otin, C., and Kahari, V. M. (1999). J. Biol. Chem. 274, 2446– 2455. Ray, J. M., and Stetler-Stevenson, W. G. (1994). Eur. Respir. J. 7, 2062–2072. Reiss, M. (1999). Microbes Infect. 1, 1327–1347. Rifkin, D. B., Gleizes, P. E., Harpel, J., Nunes, I., Munger, J., Mazzieri, R., and Noguera, I. (1997). CIBA Foundation Symp. 212, 105–115 [discussion 116–118]. Roberts, A. B., and Sporn, M. B. (1986). Cancer Surv. 5, 405– 412. Roberts, A. B., and Sporn, M. B. (1989). Am. Rev. Respir. Dis. 140, 1126–1128. Roberts, A. B., and Sporn, M. B. (1990). In “Handbook of Experimental Pharmacology. Peptide Growth Factor and Their Receptors” (M. B. Sporn and A. B. Roberts, eds.), pp. 419– 472. Springer-Verlag, Berlin. Roberts, A. B., and Sporn, M. B. (1992). Mol. Reprod. Dev. 32, 91–98. Roberts, A. B., Sporn, M. B., Assoian, R. K., Smith, J. M., Roche, N. S., Wakefield, L. M., Heine, U. I., Liotta, L. A., Falanga, V., and Kehrl, J. H. (1986). Proc. Natl. Acad. Sci. USA 83, 4167– 4171. Rook, A. H., Kehrl, J. H., Wakefield, L. M., Roberts, A. B., Sporn, M. B., Burlington, D. B., Lane, H. C., and Fauci, A. S. (1986). J. Immunol. 136, 3916–3920. Rosfjord, E. C., and Dickson, R. B. (1999). J. Mammary Gland Biol. Neoplasia 4, 229–237 Rotello, R. J., Lieberman, R. C., Purchio, A. F., and Gerschenson, L. E. (1991). Proc. Natl. Acad. Sci. USA 88, 3412–3415. Roth, S., Sistonen, P., Salovaara, R., Hemminki, A., Loukola, A., Johansson, M., Avizienyte, E., Cleary, K. A., Lynch, P., Amos, C. I., Kristo, P., Mecklin, J. P., Kellokumpu, I., Jarvinen, H., and Aaltonen, L. A. (1999). Genes Chromosomes Cancer 26, 54–61. Saltzman, A., Munro, R., Searfoss, G., Franks, C., Jaye, M., and Ivashchenko, Y. (1998). Exp. Cell Res. 242, 244–254. Samuel, S. K., Hurta, R. A., Kondaiah, P., Khalil, N., Turley, E. A., Wright, J. A., and Greenberg, A. H. (1992). EMBO J. 11, 1599–1605. Sanderson, N., Factor, V., Nagy, P., Kopp, J., Kondaiah, P., Wakefield, L., Roberts, A. B., Sporn, M. B., and Thorgeirsson, S. S. (1995). Proc. Natl. Acad. Sci. USA 92, 2572–2576. Sanford, L. P., Ormsby, I., Gittenberger-de Groot, A. C., Sariola, H., Friedman, R., Boivin, G. P., Cardell, E. L., and Doetschman, T. (1997). Development 124, 2659–2670. Sano, Y., Harada, J., Tashiro, S., Gotoh-Mandeville, R., Maekawa, T., and Ishii, S. (1999). J. Biol. Chem. 274, 8949–8957.
52
Ester Piek and Anita B. Roberts
Santibanez, J. F., Frontelo, P., Iglesias, M., Martinez, J., and Quintanilla, M. (1999). J. Cell. Biochem. 74, 61–73. Santibanez, J. F., Quintanilla, M., and Martinez, J. (2000). Nutr. Cancer 37, 49–54. Schwarte-Waldhoff, I., Klein, S., Blass-Kampmann, S., Hintelmann, A., Eilert, C., Dreschers, S., Kalthoff, H., Hahn, S. A., and Schmiegel, W. (1999). Oncogene 18, 3152–3158. Schwarte-Waldhoff, I., Volpert, O. V., Bouck, N. P., Sipos, B., Hahn, S. A., Klein-Scory, S., Luttges, J., Kloppel, G., Graeven, U., Eilert-Micus, C., Hintelmann, A., and Schmiegel, W. (2000). Proc. Natl. Acad. Sci. USA 97, 9624–9629. Sehgal, I., and Thompson, T. C. (1999). Mol. Biol. Cell 10, 407– 416. Sehgal, I., and Baley, P. A. (1996). Cancer Res. 56, 3359–3365. Seo, N., Tokura, Y., Takigawa, M., and Egawa, K. (1999). J. Immunol. 163, 242–249. Shi, Y., Hata, A., Lo, R. S., Massague, J., and Pavletich, N. P. (1997). Nature 388, 87–93. Shih, W. L., Kuo, M. L., Chuang, S. E., Cheng, A. L., and Doong, S. L. (2000). J. Biol. Chem. 275, 25858–25864. Sood, R., Talwar-Trikha, A., Chakrabarti, S. R., and Nucifora, G. (1999). Leukemia 13, 348– 357. Souchelnytskyi, S., ten Dijke, P., Miyazono, K., and Heldin, C. H. (1996). EMBO J. 15, 6231– 6240. Stander, M., Naumann, U., Dumitrescu, L., Heneka, M., Loschmann, P., Gulbins, E., Dichgans, J., and Weller, M. (1998). Gene Ther. 5, 1187–1194. Stander, M., Naumann, U., Wick, W., and Weller, M. (1999). Cell Tissue Res. 296, 221–227. Stearns, M. E., Garcia, F. U., Fudge, K., Rhim, J., and Wang, M. (1999). Clin. Cancer Res. 5, 711–720. Steiner, M. S., and Barrack, E. R. (1992). Mol. Endocrinol. 6, 15–25. Stroschein, S. L., Wang, W., Zhou, S., Zhou, Q., and Luo, K. (1999). Science 286, 771–774. Sun, Y., Hegamyer, G., and Colburn, N. H. (1994). Cancer Res. 54, 1139–1144. Sun, Y., Liu, X., Ng-Eaton, E., Lodish, H. F., and Weinberg, R. A. (1999). Proc. Natl. Acad. Sci. USA 96, 12442–12447. Sylvia, V. L., Schwartz, Z., Dean, D. D., and Boyan, B. D. (2000). Biochim. Biophys. Acta 1496, 311–324. Tada, K., Fukunaga, T., Wakabayashi, Y., Masumi, S., Sato, Y., Izumi, H., Kohno, K., and Kuwano, M. (1994). Biochim. Biophys. Acta 1201, 135–142. Tada, T., Ohzeki, S., Utsumi, K., Takiuchi, H., Muramatsu, M., Li, X. F., Shimizu, J., Fujiwara, H., and Hamaoka, T. (1991). J. Immunol. 146, 1077–1082. Taipale, J., Saharinen, J., and Keski-Oja, J. (1998). Adv. Cancer Res. 75, 87–134. Takaku, K., Oshima, M., Miyoshi, H., Matsui, M., Seldin, M. F., and Taketo, M. M. (1998). Cell 92, 645– 656. Takata, T., Miyauchi, M., Ogawa, I., Kudo, Y., Takekoshi, T., Zhao, M., Sato, S., Nikai, H., and Komiyama, K. (2000). Virchows Arch. 436, 319–323. Taketo, M. M., and Takaku, K. (2000). Cytokine Growth Factor Rev. 11, 147–157. Tang, B., Bottinger, E. P., Jakowlew, S. B., Bagnall, K. M., Mariano, J., Anver, M. R., Letterio, J. J., and Wakefield, L. M. (1998). Nat. Med. 4, 802–807. Tang, B., de Castro, K., Barnes, H. E., Parks, W. T., Stewart, L., Bottinger, E. P., Danielpour, D., and Wakefield, L. M. (1999). Cancer Res. 59, 4834– 4842. Tang, W., Yang, L., Yang, Y. C., Leng, S. X., and Elias, J. A. (1998). J. Biol. Chem. 273, 5506–5513. Teti, A., De Giorgi, A., Spinella, M. T., Migliaccio, S., Canipari, R., Onetti Muda, A., and Faraggiana, T. (1997). Int. J. Cancer 72, 1013–1020. Torre-Amione, G., Beauchamp, R. D., Koeppen, H., Park, B. H., Schreiber, H., Moses, H. L., and Rowley, D. A. (1990). Proc. Natl. Acad. Sci. USA 87, 1486–1490. Tsang, M., Kim, R., de Caestecker, M. P., Kudoh, T., Roberts, A. B., and Dawid, I. B. (2000). Genesis 28, 47–57.
Suppressor and Oncogenic Roles of TGF-β
53
Tsukazaki, T., Chiang, T. A., Davison, A. F., Attisano, L., and Wrana, J. L. (1998). Cell 95, 779–791. Tsushima, H., Kawata, S., Tamura, S., Ito, N., Shirai, Y., Kiso, S., Imai, Y., Shimomukai, H., Nomura, Y., Matsuda, Y., and Matsuzawa, Y. (1996). Gastroenterology 110, 375–382. Turley, J. M., Falk, L. A., Ruscetti, F. W., Kasper, J. J., Francomano, T., Fu, T., Bang, O. S., and Birchenall-Roberts, M. C. (1996). Cell Growth Differ. 7, 1535– 1544. Tuszynski, G. P., and Nicosia, R. F. (1996). Bioessays 18, 71–76. Ulloa, L., Doody, J., and Massague, J. (1999). Nature 397, 710 –713. Vernon, R. B., and Sage, E. H. (1999). Microvasc. Res. 57, 118–133. Verschueren, K., Remacle, J. E., Collart, C., Kraft, H., Baker, B. S., Tylzanowski, P., Nelles, L., Wuytens, G., Su, M. T., Bodmer, R., Smith, J. C., and Huylebroeck, D. (1999). J. Biol. Chem. 274, 20489–20498. ¨ Wakefield, L. M., Piek, E., and Bottinger, E. P. (2001). J Mammary Gland Biol. Neoplasia, in press. Waldrip, W. R., Bikoff, E. K., Hoodless, P. A., Wrana, J. L., and Robertson, E. J. (1998). Cell 92, 797–808. Wang, D., Kanuma, T., Takama, F., Mizumuma, H., Ibuki, Y., Wake, N., Mogi, A., Shitara, Y., Hagiwara, K., and Takenoshita, S. (1999). Int. J. Oncol. 15, 949–953. Wang, D., Kanuma, T., Mizunuma, H., Takama, F., Ibuki, Y., Wake, N., Mogi, A., Shitara, Y., and Takenoshita, S. (2000). Cancer Res. 60, 4507–4512. Wang, J., Han, W., Zborowska, E., Liang, J., Wang, X., Willson, J. K. V., Sun, L., and Brattain, M. G. (1996). J. Biol. Chem. 271, 17366–17371. Wang, X. J., Greenhalgh, D. A., Bickenbach, J. R., Jiang, A., Bundman, D. S., Krieg, T., Derynck, R., and Roop, D. R. (1997). Proc. Natl. Acad. Sci. USA 94, 2386–2391. Wang, X. J., Liefer, K. M., Tsai, S., O’Malley, B. W., and Roop, D. R. (1999). Proc. Natl. Acad. Sci. USA 96, 8483–8488. Warner, B. J., Blain, S. W., Seoane, J., and Massagu, (1999). Mol. Cell. Biol. 19, 5913–5922. Weinstein, M., Yang, X., Li, C., Xu, X., Gotay, J., and Deng, C. X. (1998). Proc. Natl. Acad. Sci. USA 95, 9378–9383. Welch, D. R., Fabra, A., and Nakajima, M. (1990). Proc. Natl. Acad. Sci. USA 87, 7678–7682 Wick, W., Petersen, I., Schmutzler, R. K., Wolfarth, B., Lenartz, D., Bierhoff, E., Hummerich, J., Muller, D. J., Stangl, A. P., Schramm, J., Wiestler, O. D., and von Deimling, A. (1996). Oncogene 12, 973–978. Wojtowicz-Praga, S. (1997). J. Immunother. 20, 165–177. Won, J., Kim, H., Park, E. J., Hong, Y., Kim, S. J., and Yun, Y. (1999). Cancer Res. 59, 1273– 1277. Wong, C., Rougier-Chapman, E. M., Frederick, J. P., Datto, M. B., Liberati, N. T., Li, J. M., and Wang, X. F. (1999). Mol. Cell. Biol. 19, 1821–1830. Woodhouse, E. C., Chuaqui, R. F., and Liotta, L. A. (1997). Cancer 80 , 1529–1537. Wotton, D., Lo, R. S., Lee, S., and Massague, J. (1999a). Cell 97, 29–39. Wotton, D., Lo, R. S., Swaby, L. A., and Massague, J. (1999b). J. Biol. Chem. 274, 37105– 37110. Wunderlich, H., Steiner, T., Kosmehl, H., Junker, U., Reinhold, D., Reichelt, O., Zermann, D. H., and Schubert, J. (1998). Urol. Int. 60, 205–207. Xu, J., and Attisano, L. (2000). Proc. Natl. Acad. Sci. USA 97, 4820–4825. Xu, X., Brodie, S. G., Yang, X., Im, Y. H., Parks, W. T., Chen, L., Zhou, Y. X., Weinstein, M., Kim, S. J., and Deng, C. X. (2000). Oncogene 19, 1868–1874. Yagi, K., Goto, D., Hamamoto, T., Takenoshita, S., Kato, M., and Miyazono, K. (1999). J. Biol. Chem. 274, 703–709. Yakymovych, I., ten Dijke, P., Heldin, C. H., and Souchelnytskyi, S. (2001). FASEB J., 15, 553–555.
54
Ester Piek and Anita B. Roberts
Yamaguchi, K., Nagai, S., Ninomiya-Tsuji, J., Nishita, M., Tamai, K., Irie, K., Ueno, N., Nishida, E., Shibuya, H., and Matsumoto, K. (1999). EMBO J. 18, 179–187. Yamamura, Y., Hua, X., Bergelson, S., and Lodish, H. F. (2000). J. Biol. Chem. 275, 36295– 36302. Yamanaka, R., Tanaka, R., Yoshida, S., Saitoh, T., Fujita, K., and Naganuma, H. (1999). J. Neurooncol. 43, 27–34. Yan, Z., Deng, X., and Friedman, E. (2001). J. Biol. Chem. 276, 1555–1563. Yanagihara, K., and Tsumuraya, M. (1992). Cancer Res 52, 4042–4045. Yanagisawa, J., Yanagi, Y., Masuhiro, Y., Suzawa, M., Watanabe, M., Kashiwagi, K., Toriyabe, T., Kawabata, M., Miyazono, K., and Kato, S. (1999). Science 283, 1317–1321. Yanagisawa, K., Uchida, K., Nagatake, M., Masuda, A., Sugiyama, M., Saito, T., Yamaki, K., Takahashi, T., and Osada, H. (2000). Oncogene 19, 2305–2311. Yang, X., Letterio, J. J., Lechleider, R. J., Chen, L., Hayman, R., Gu, H., Roberts, A. B., and Deng, C. (1999a). EMBO J. 18, 1280–1291. Yang, X., Castilla, L. H., Xu, X., Li, C., Gotay, J., Weinstein, M., Liu, P. P., and Deng, C. X. (1999b). Development 126, 1571–1580. Yin, J. J., Selander, K., Chirgwin, J. M., Dallas, M., Grubbs, B. G., Wieser, R., Massague, J., Mundy, G. R., and Guise, T. A. (1999). J. Clin. Invest. 103, 197–206. Yu, Q., and Stamenkovic, I. (2000). Genes Dev. 14, 163–176. Yue, J., and Mulder, K. M. (2000). J. Biol. Chem. 275, 35656. Yue, J., Frey, R. S., and Mulder, K. M. (1999). Oncogene 18, 2033–2037. ¨ Zavadil, J., Bitzer, M., Liang, D., Yang, Y.-C., Massimi, A., Kneitz, S., Piek, E., and Bottinger, E. P. (2001). Proc. Natl. Acad. Sci. USA 98, 6686–6691. Zhang, Y., Feng, X. H., and Derynck, R. (1998). Nature 394, 909–913. Zhao, J., and Buick, R. N. (1993). Exp. Cell Res. 204, 82–87. Zhao, J., and Buick, R. N. (1995). Cancer Res. 55, 6181–6188. Zhou, G., Lee, S. C., Yao, Z., and Tan, T. H. (1999). J. Biol. Chem. 274, 13133–13138. Zhou, S., Buckhaults, P., Zawel, L., Bunz, F., Riggins, G., Le Dai, J., Kern, S. E., Kinzler, K. W., and Vogelstein, B. (1998). Proc. Natl. Acad. Sci. USA 95, 2412–2416. Zhou, S., Kinzler, K. W., and Vogelstein, B. (1999). N. Engl. J. Med. 341, 1144–1146. Zhu, Y., Richardson, J. A., Parada, L. F., and Graff, J. M. (1998). Cell 94, 703–714.
Hereditary Diffuse Gastric Cancer Anita Dunbier and Parry Guilford Cancer Genetics Laboratory Department of Biochemistry University of Otago Dunedin, New Zealand
I. II. III. IV. V. VI. VII. VIII.
Introduction Hereditary Diffuse Gastric Cancer Mutations in CDH-1 The Tumor Spectrum of HDGC Inactivation of the Second CDH-1 Allele Molecular Mechanism of HDGC Susceptibility Clinical Criteria and Management of HDGC Conclusion References
Hereditary diffuse gastric cancer (HDGC) is a cancer predisposition syndrome caused by germline mutation of the gene for the cell-to-cell adhesion protein E-cadherin. The syndrome is dominated by predisposition to the histologically diffuse, poorly differentiated form of gastric cancer. It is not associated with intestinal-type gastric cancer, but families may have an elevated risk of lobular breast cancer. Here, we review the identified families, mutations, and proposed mechanisms by which E-cadherin loss promotes tumorigenesis. C 2001 Academic Press.
I. INTRODUCTION Despite a significant decline in incidence during the 20th century, gastric cancer still ranks second in terms of the global cancer burden, accounting for more than half a million deaths each year (Boyle, 1997; Pisani et al., 1999). Wide variation is observed in the rates of gastric cancer among different populations. Japan has an incidence of nearly 80 cases per 100,000 males, whereas the United States and most Western countries have incidence rates of between 10 and 40 per 100,000 (Parkin et al., 1999). Although environmental and dietary effects play a critical role in gastric cancer incidence (Howson et al., 1986), recent documentation of familial clustering has highlighted the importance of inherited predisposition to this
Advances in CANCER RESEARCH 0065-230X/01 $35.00
55
C 2001 by Academic Press. Copyright All rights of reproduction in any form reserved.
56
Anita Dunbier and Parry Guilford
form of cancer. Approximately 10% of gastric cancer cases show familial clustering (La Vecchia et al., 1992; Zanghieri et al., 1990) and epidemiologic studies have shown that the risk of gastric cancer in first-degree relatives is increased two- or three-fold (Goldgar et al., 1994). More than 90% of gastric cancers can be classified as either intestinal or diffuse-type adenocarcinomas based on their degree of differentiation (Lauren, 1965). Intestinal tumors are well differentiated and characterized by metaplasia to an intestinal cell type. Their incidence is usually associated with severe atrophic gastritis. Diffuse-type tumors are poorly differentiated and are characterized by infiltrative growth and peritoneal dissemination. Analysis of epidemiological data in light of this classification reveals that environmental factors have a relatively greater effect on the incidence of intestinal-type gastric cancer than the diffuse form. This difference was first suggested by a study of Japanese migrants to Hawaii which showed the rates of intestinal-type gastric cancer halved in the migrants although the rates of diffuse-type carcinoma were similar before and after migration (Correa et al., 1973). Interestingly, this effect was most pronounced in younger males: The intestinal rate decreased more than fivefold in the 15- to 49-year-old age group, contrasting strongly with the absence of a significant change in the rate of the diffuse disease. Similarly, the worldwide decrease in gastric cancer incidence does not appear to be mirrored by a decease in the rate of the diffuse type (Borch et al., 2000; Howson et al., 1986). In contrast, evidence for a significant hereditary component in diffuse gastric cancer has been found by many authors. A study of 16 Japanese gastric cancer families (defined by the existence of three or more family members with gastric cancer in at least two successive generations) found that compared to nonfamilial cases, the probands were more likely to develop poorly differentiated gastric cancer (Kakiuchi et al., 1999). Similarly, data from familial gastric cancer registry databases in the United Kingdom showed that 17 of 20 gastric cancer families for which pathological information was available showed cancer of the diffuse type (Caldas et al., 1999). A population-based study also found that the diffuse form is overrepresented in patients under the age of 40 years (Theuer et al., 1998). Although this tendency toward early onset may reflect the aggressive nature of diffuse gastric cancer, early onset is one hallmark of a genetic predisposition to cancer.
II. HEREDITARY DIFFUSE GASTRIC CANCER The first evidence for a specific gastric cancer susceptibility locus was the localization of a gene predisposing to diffuse gastric cancer in a large New Zealand Maori family (Guilford et al., 1998). The predisposing gene mapped
57
Hereditary Diffuse Gastric Cancer
to an interval on chromosome 16q22.1 containing the gene for the cell-tocell adhesion protein E-cadherin (CDH-1). Subsequent mutation analysis identified inactivating germline CDH-1 mutations in this family and two other families of Maori ethnicity with familial diffuse gastric cancer (Guilford et al., 1998). E-cadherin is a member of the cadherin family of homophilic cell adhesion proteins that are central to the processes of development, cell differentiation, and the maintenance of tissue integrity (Grunwald, 1993). It is the predominant cadherin family member expressed in epithelial tissue and is localized at the adherens junctions on the cell’s basolateral surface. Inactivating germline CDH-1 mutations have been identified in numerous gastric cancer families from diverse ethnic groups (Table I). These families are predisposed predominantly to diffuse-type gastric cancer, with linitis plastica in advanced cases. Histologically, the tumors are highly invasive, poorly differentiated, and display occasional signet ring cells. Two patients with germline CDH-1 mutations have presented with the mixed type of gastric tumor that contains components of both diffuse and intestinal histology (Caldas et al., 1999). However, there is no association with the pure intestinal type. This inherited cancer susceptibility has been designated hereditary diffuse gastric cancer (HDGC) (Guilford et al., 1999). Table I Germline Mutations Described in HDGC Families Mutation Truncating 49-2A→G 59G→A 70G→T 187C→T 190C→T 372delC 586G→T 1008G→T 1137+1G→A 1488del7 1588insC 1711insG 1792C→T 2095C→T 2381insC Missense 185G→T 731A→G 1018A→G 1460T→C 1796C→G
Effect
Exon
Reference
Splice Nonsense Nonsense Nonsense Nonsense Frameshift Nonsense Splice Splice Frameshift Frameshift Frameshift Nonsense Nonsense Frameshift
2 2 2 3 3 3 5 7 8 10 11 11 12 13 15
Richards et al. (1999) Richards et al. (1999) Guilford et al. (1999) Gayther et al. (1998) Guilford et al. (1999) Keller et al. (1999) Guilford et al. (1999) Guilford et al. (1998) Guilford et al. (1999) Guilford et al. (1999) Guilford et al. (1999) Gayther et al. (1998) Gayther et al. (1998) Guilford et al. (1998) Guilford et al. (1998)
G62V D244G T340A V487A T599S
3 6 8 10 12
Shinmura et al. (1999) Yoon et al. (1999) Kim et al. (2000a) Yoon et al. (1999) Kim et al. (2000a)
58
Anita Dunbier and Parry Guilford
The lifetime penetrance of HDGC is about 70%, and age of onset shows marked variation between and within families, beginning as young as 14 years of age (Guilford et al., 1998). It is notable that the median age of onset in Maori CDH-1 mutation carriers (n = 30) is significantly younger (32 years) than the median age in carriers of other ethnicity (43 years, n = 27) (P. Guilford, unpublished observations). It is not known if this difference reflects the effect of environmental triggers or differing genetic backgrounds.
III. MUTATIONS IN CDH-1 The CDH-1 coding sequence is 2.6 kB in length and divided over 16 exons. Germline mutations identified in HDGC families consist predominantly of frameshift mutations, premature termination codons, and exon/intron splice site mutations that are distributed throughout the gene without any apparent hot spots (Fig. 1) (Gayther et al.,1998; Guilford et al.,1998, 1999; Keller et al., 1999; Richards et al., 1999). Missense germline CDH-1 mutations have also been observed in some families, and they are predicted to be responsible for gastric cancer predisposition if they disrupt critical functional domains of the protein (Kim et al.,2000a; Shinmura et al.,1999; Yoon et al., 1999). Somatic CDH-1 mutations have been identified in about 50% of sporadic diffuse gastric tumors and lobular breast cancers but occur very rarely in other tumors, including intestinal-type gastric cancers (Becker et al., 1994; Berx et al., 1996, 1998a). In contrast to the germline mutations found in the familial syndrome, mutations in sporadic tumors typically result in in-frame deletions removing partial or complete exon sequences from the extracellular portion of the transmembrane protein or point mutations resulting in amino acid substitutions (Berx et al., 1998a). In-frame deletions of exon 8 or 9 are the most frequent of these events and result in disruption of one of the extracellular calcium binding domains of E-cadherin. The significance of the apparent differences in reported sporadic and germline mutations is difficult to ascertain. Because somatic CDH-1 mutation screening in sporadic tumors has been undertaken largely using reverse transcriptase-polymerase chain reaction from tumor RNA, a bias toward the identification of shorter, exon-skipped products may exist. In contrast, searches for germline CDH-1 mutations are typically conducted by direct sequencing of individual exons amplified from genomic DNA and hence will lack this bias toward exon-skipped transcripts. However, the observed differences in mutation type may have a biological basis. In vitro analyses of cells expressing E-cadherin lacking exon 8 or 9 suggested that these mutant proteins may act in a transdominant-active manner, increasing cell motility
Fig. 1 Pedigrees of HDGC families and reported germline CDH-1 mutations. Squares, males; circles, females; all symbols with a diagonal, deceased. Solid symbols, gastric cancer; dotted symbols, extragastric cancer.
60
Anita Dunbier and Parry Guilford
(Handschuh et al., 1999; Luber et al., 2000). Such mutations are unlikely to be observed in the germline because they would be embryonically lethal.
IV. THE TUMOR SPECTRUM OF HDGC Since CDH-1 is somatically mutated in a substantial portion of invasive lobular breast carcinomas (Berx et al., 1996), it is reasonable to propose that HDGC families with CDH-1 mutations may also have an elevated risk of lobular breast cancer. In one reported case, a patient carrying a germline mutation developed metachronous lobular breast and diffuse-type gastric carcinoma at the ages of 49 and 58, respectively (Keller et al., 1999). Further lobular breast carcinomas, colorectal carcinomas, and prostate carcinomas have also been documented in CDH-1 mutation carriers (Caldas et al., 1999), although because of the small number of cases it is not possible to determine whether these rates are higher than that of the general population. However, the absence of observed somatic CDH-1 mutations in colorectal and prostate cancer argues against these cancers being significant contributors to the HDGC syndrome phenotype.
V. INACTIVATION OF THE SECOND CDH-1 ALLELE Immunohistochemical staining of HDGC tumors with anti-E-cadherin antibodies has shown that the second CDH-1 allele is inactivated somatically (Grady et al., 2000). HDGC therefore resembles other inherited cancer syndromes caused by germline mutation of tumor suppressor genes in requiring somatic inactivation of the wild-type allele to enable tumor progression. However, unlike other cancer syndromes, loss of heterozygosity does not appear to be frequent in HDGC (Grady et al., 2000; Guilford et al., 1999; Richards et al., 1999). Instead, Grady et al. (2000) demonstrated that hypermethylation of the CDH-1 promoter is likely to be a common cause of the “second hit” in HDGC tumors. CDH-1 promoter hypermethylation has also been reported to occur in about 80% of sporadic diffuse gastric cancers and about one-third of other gastric cancer types (Tamura et al., 2000). The demonstration that the second hit on CDH-1 expression need not be an irreversible mutation or deletion event suggests that sustained transcription factor-mediated CDH-1 downregulation or posttranslational modification of E-cadherin may also be sufficient to promote gastric cancer progression. Since the expression of E-cadherin is downregulated by many
Hereditary Diffuse Gastric Cancer
61
factors, including ulceration (Hanby et al., 1996), H. pylori infection (Terr´es et al., 1998), and possibly dietary factors (Meng et al., 2000a,b), the control of E-cadherin expression provides an access point for environmental factors to influence the genetics of tumor progression. It is also possible that the retention of trace levels of E-cadherin expression, rather than complete irreversible loss through mutation, may provide a survival advantage for the tumor by inhibiting apoptotic pathways that are associated with the loss of cell adhesion. In addition, the ability of a tumor to transiently reexpress E-cadherin may facilitate the establishment of distant metastases.
VI. MOLECULAR MECHANISM OF HDGC SUSCEPTIBILITY The invasive phenotype of malignant epithelial tumor cells can be abrogated in model systems by transfection with E-cadherin cDNA (Frixen et al., 1991; Vleminckx et al., 1991). Perl et al. (1998) demonstrated, using a mouse model of pancreatic β cell tumorigenesis, that downregulation of E-cadherin-mediated cell adhesion by expression of a dominant-negative form of E-cadherin coincides with the transition from well-differentiated adenoma to invasive carcinoma. CDH-1 can thus be regarded as a tumor invasion suppressor gene (Berx et al., 1995). The close relationship between cell proliferation and cell migration during development, wound repair, and stem cell proliferation would necessitate cross talk between the molecular pathways for cell adhesion and cell proliferation. E-cadherin loss in the gastric epithelium would therefore also be predicted to contribute to tumorigenesis not only by enhancing tumor invasion but also by stimulating cell proliferative pathways. One of the key proliferative pathways associated with E-cadherin is the Wnt signaling pathway, which is strongly implicated in the pathogenesis of human gastrointestinal and hepatocellular cancers (Berx et al., 1998b; Polakis, 1999). The cytoplasmic domain of E-cadherin interacts with a complex of proteins at the adherens junction, including the central mediator of Wnt signaling, β-catenin (Kikuchi, 2000). Evidence suggests that loss of functional E-cadherin may shift the cellular equilibrium of β-catenin away from the adherens junction and toward the pool of free β-catenin (Orsulic et al., 1999; Sadot et al., 1998). Increased free β-catenin would activate the transcriptional targets of the Wnt pathway which include the oncogene c-myc, the regulator of cell proliferation cyclin D1, and components of the AP1 transcription complex (He et al., 1998; Mann et al., 1999; Shtutman et al., 1999).
62
Anita Dunbier and Parry Guilford
Disruption of other “outside–in” signaling pathways involving E-cadherin may also promote tumorigenesis in HDGC families. The formation of E-cadherin-mediated cell adhesion leads to the colocalization of epidermal growth factor receptor with the E-cadherin complex (Pece and Gutkind, 2000). This colocalization activates the receptor, leading to stimulation of the mitogen-activated protein kinase signaling pathway (Pece and Gutkind, 2000). E-cadherin-mediated cell adhesion also activates the phoshatidylinositol 3-kinase/Akt survival pathway (Pece et al., 1999) and Cdc42, a member of the Rho family of small GTPases (Kim et al., 2000b). It is therefore possible that E-cadherin loss also contributes to the poorly differentiated cancer phenotype by perturbation of these fundamental mechanisms for proliferation, survival, and cell differentiation (Aplin et al., 1999; Seger and Krebs, 1995).
VII. CLINICAL CRITERIA AND MANAGEMENT OF HDGC The distinction between the histological subtypes of gastric cancer is a key feature of the clinical criteria used as a screen for HDGC families. Currently, the following criteria are used (Caldas et al., 1999): (i) two or more documented cases of diffuse gastric cancer in first- or second-degree relatives, with at least one diagnosed before the age of 50 years, or (ii) three or more cases of documented diffuse gastric cancer in first- or second-degree relatives, independent of age of onset. About 75% of families meeting these criteria have identifiable germline mutations in the CDH-1 coding region. The remainder of families may have undetected mutations in CDH-1 regulatory sequences or germline mutations in unidentified genes that also contribute to a diffuse, poorly differentiated phenotype. Alternatively, families lacking CDH-1 mutations may simply represent chance clusters of sporadic cancer. Five-year survival rates of gastric cancer patients following complete gastrectomy are up to 90% for stage Ia cancers but decrease dramatically for stage IV cancers. Intensive clinical surveillance of germline CDH-1 mutation carriers therefore provides an opportunity to improve outcome. Diffuse gastric cancer can be difficult to detect due to a tendency toward infiltrative submucosal spread; nevertheless, regular endoscopy (every 6–12 months) provides the best method for early detection of HDGC. For some HGDC families, prophylactic total gastrectomy may constitute an appropriate preventative strategy for gastric cancer. However, when deciding whether to carry out a prophylactic gastrectomy, the 1 or 2% risk of mortality following
Hereditary Diffuse Gastric Cancer
63
surgery, the 100% long-term morbidity, the incomplete penetrance of HDGC, and the unknown risk of later development of extragastric tumors must be considered.
VIII. CONCLUSION HDGC is a classic cancer susceptibility syndrome resulting from inheritance of a germline mutation in the E-cadherin tumor suppressor gene CDH-1. Inactivation or downregulation of the second allele is required for tumor progression, and promoter hypermethylation is likely to constitute a major mechanism for this second hit. The absence of normal E-cadherin levels facilitates progression of a tumor from a benign state to an invasive, malignant one. However, loss of E-cadherin may also disrupt outside–in signaling pathways, leading to increased cell proliferation or survival.
REFERENCES Aplin, A. E., Howe, A. K., and Juliano, R. L. (1999). Curr. Opin. Cell Biol. 11, 737–744. Becker, K.-F., Atkinson, M. J., Reich, U., Brecker, I., Nekarda, H., Siewart, J. R., and Hofler, H. (1994). Cancer Res. 54, 3845–3852. Berx, G., Staes, K., Hengel, J. V., Molemans, F., Bussemakers, M. J. G., Bokhoven, A. V., and Roy, F. V. (1995). Genomics 26, 281–289. Berx, G., Cleton-Jansen, A.-M., Strumane, K., de Leeuw, W. J. F., Nollet, F., van Roy, F., and Cornelisse, C. (1996). Oncogene 13, 1919–1925. Berx, G., Becker, K.-F., Hofler, H., and Roy, F. V. (1998a). Hum. Mutat. 12, 226–237. Berx, G., Nollet, F., and Roy, F. V. (1998b). Cell Adhesion Commun. 6, 171–184. Borch, K., Jonsson, B., Tarpila, E., Franzen, T., Berglund, J., Kullman, E., and Franzen, L. (2000). Br. J. Surg. 87, 618–626. Boyle, P. (1997). Lancet 349, 23–26. Caldas, C., Carneiro, F., Lynch, H. T., Yokota, J., Wiesner, G. L., Powell, S. M., Lewis, F. R., Huntsman, D. G., Pharoah, P. D., Jankowski, J. A., MacLeod, P., Vogelsang, H., Keller, G., Park, K. G., Richards, F. M., Maher, E. R., Gayther, S. A., Oliveira, C., Grehan, N., Wight, D., Seruca, R., Roviello, F., Ponder, B. A., and Jackson, C. E. (1999). J. Med. Genet. 36, 873–880. Correa, P., Sasano, N., Stemmermann, G. N., and Haenszel, W. (1973). J. Natl. Cancer Inst. 51, 1449–1459. Frixen, E. H., Behrens, J., Sachs, M., Eberle, G., Voss, B., Warda, A., Lochner, D., and Birchmeier, W. (1991). J. Cell Biol. 113, 173–185. Gayther, S. A., Gorringe, K. L., Ramus, S. J., Huntsman, D., Roviello, F., Grehan, N., Machado, J. C., Pinto, E., Seruca, R., Halling, K., MacLeod, P., Powell, S. M., Jackson, C. E., Ponder, B. A. J., and Caldas, C. (1998). Cancer Res. 58, 4086– 4089.
64
Anita Dunbier and Parry Guilford
Goldgar, D. E., Easton, D. F., Cannon-Albright, L. A., and Skolnick, M. H. (1994). J. Natl. Cancer Inst. 86, 1600–1608. Grady, W. M., Willis, J., Guilford, P. J., Dunbier, A. K., Toro, T. T., Lynch, H., Wiesner, G., Ferguson, K., Eng, C., Park, J. G., Kim, S. J., and Markowitz, S. (2000). Nature Genet. 26, 16–17. Grunwald, G. B. (1993). Curr. Opin. Cell Biol. 5, 797–805. Guilford, P., Hopkins, J., Harraway, J., McLeod, M., McLeod, N., Harawira, P., Taite, H., Scoular, R., Miller, A., and Reeve, A. E. (1998). Nature 392, 402–405. Guilford, P., Hopkins, J., Grady, W., Markowitz, S., Willis, J., Lynch, H., Rajput, A., Wiesner, G., Lindor, N., Burgart, L., Toro, T., Lee, D., Limacher, J.-M., Shaw, D., Findlay, M., and Reeve, A. (1999). Hum. Mutat. 14, 249–255. Hanby, A. M., Chinery, R., Poulsom, R., Playford, R. J., and Pignatelli, M. (1996). Am. J. Pathol. 148, 723–729. Handschuh, G., Candidus, S., Luber, B., Reich, U., Schott, C., Oswald, S., Becke, H., Hutzler, ¨ P., Birchmeier, W., Hofler, H., and Becker, K.-F. (1999). Oncogene 18, 4301–4312. He, T.-C., Sparks, A. B., Rago, C., Hermeking, H., Zawel, L., Costa, L. T. d., Morin, P. J., Vogelstein, B., and Kinzler, K. W. (1998). Science 281, 1509–1512. Howson, C. P., Hiyama, T., and Wynder, E. L. (1986). Epidemiol. Rev. 8, 1–27. Kakiuchi, H., Itoh, F., Kusano, M., Adachi, Y., Mita, H., Mihara, M., Matsuno, K., Endo, T., Hinoda, Y., Hosokawa, M., and Imai, K. (1999). Tumour Biol. 20, 235–241. Keller, G., Vogelsang, H., Becker, I., Hutter, J., Ott, K., Candidus, S., Grundei, T., Becker, K.-F., Mueller, J., Siewert, J., and Hofler, H. (1999). Am. J. Pathol. 155, 337–342. Kikuchi, A. (2000). Biochem. Biophys. Res. Commun. 268, 243–248. Kim, H., Wheeler, J., Kim, J., Ilyas, M., Beck, N., Kim, B., Park, K., and Bodmer, W. (2000a). Gut 47, 262–267. Kim, S. H., Li, Z., and Sacks, D. B. (2000b). J. Biol. Chem. 275, 36999–37005. Lauren, P. (1965). Acta Pathol. Microbiol. Scand. 64, 31–49. La Vecchia, C., Negri, E., Franceschi, S., and Gentile, A. (1992). Cancer 70, 50–55. Luber, B., Candidus, S., Handschuh, G., Mentele, E., Hutzler, P., Feller, S., Voss, J., Hofler, H., and Becker, K. F. (2000). Cell Adhesion Commun. 7, 391– 408. Mann, B., Gelos, M., Siedow, A., Hanski, M. L., Gratchev, A., Ilyas, M., Bodmer, W. F., Moyer, M. P., Riecken, E. O., Buht, H. J., and Hanski, C. (1999). Proc. Natl. Acad. Sci. USA 96, 1603–1608. Meng, Q., Gao, B., Goldberg, I. D., Rosen, E. M., and Fan, S. (2000a). Biochem. Biophys. Res. Commun. 273, 448– 453. Meng, Q., Qi, M., Chen, D. Z., Yuan, R., Goldberg, I. D., Rosen, E. M., Auborn, K., and Fan, S. (2000b). J. Mol. Med. 78, 155–165. Orsulic, S., Huber, O., Aberle, H., Arnold, S., and Kemler, R. (1999). J. Cell Sci. 112, 1237– 1245. Parkin, D. M., Pisani, P., and Ferlay, J. (1999). Int. J. Cancer 80, 827–841. Pece, S., and Gutkind, J. S. (2000). J. Biol. Chem. 275, 41227– 41233. Pece, S., Chiariello, M., Murga, C., and Gutkind, J. S. (1999). J. Biol. Chem. 274, 19347–19351. Perl, A. K., Wilenbus, P., Dahl, U., Semb, H., and Christofori, G. (1998). Nature 392, 190– 193. Pisani, P., Parkin, D. M., Bray, F., and Ferlay, J. (1999). Int. J. Cancer 83, 18–29. Polakis, P. (1999). Curr. Opin. Genet. Dev. 9, 15–21. Richards, F. M., McKee, S. A., Rajpar, M. H., Cole, T. R. P., Evans, D. G. R., Jankowski, J. A., McKeown, C., Sanders, D. S. A., and Maher, E. R. (1999). Hum. Mol. Genet. 8, 607–610. Sadot, E., Simcha, I., Shtutman, M., Ben-Ze’ev, A., and Geiger, B. (1998). Proc. Natl. Acad. Sci. USA 95, 15339–15344. Seger, R., and Krebs, E. G. (1995). FASEB J. 9, 726–735.
Hereditary Diffuse Gastric Cancer
65
Shinmura, K., Kohno, T., Takahashi, M., Sasaki, A., Ochiai, A., Guilford, P., Hunter, A., Reeve, A. E., Sugimura, H., Yamaguchi, N., and Yokota, J. (1999). Carcinogenesis 20, 1127–1131. Shtutman, M., Zhurinsky, J., Simcha, I., Albanese, C., D’Amico, M., Pestell, R., and Ben-Ze’ev, A. (1999). Proc. Natl. Acad. Sci. USA 96, 5522–5527. Tamura, G., Yin, J., Wang, S., Fleisher, A. S., Zou, T., Abraham, J. M., Kong, D., Smolinski, K. N., Wilson, K. T., James, S. P., Silverberg, S. G., Nishizuka, S., Terashima, M., Motoyama, T., and Meltzer, S. J. (2000). J. Natl. Cancer Inst. 92, 569–573. Terr´es, A. M., Pajares, J. M., O’Toole, D., Ahern, S., and Kelleher, D. (1998). J. Clin. Pathol. 51, 410– 412. Theuer, C. P., Kurosaki, T., Taylor, T. H., and Anton-Culver, H. (1998). Cancer 83, 25–33. Vleminckx, K., Vakaet, L., Mareel, M., Fiers, W., and van Roy, F. (1991). Cell 66, 107–119. Yoon, K., Ku, J., Yang, H., Kim, W., Park, S., and Park, J. (1999). J. Hum. Genet. 44, 177–180. Zanghieri, G., Di Gregorio, C., Sacchetti, C., Fante, R., Sassatelli, R., Cannizzo, G., Carriero, A., and Ponz de Leon, M. (1990). Cancer 66, 2047–2051.
This Page Intentionally Left Blank
Role of Heparan Sulfate Proteoglycans in Cell Signaling and Cancer Erica M. Selva and Norbert Perrimon Department of Genetics and Howard Hughes Medical Institute Harvard Medical School, Boston, MA 02115
I. II. III. IV. V. VI. VII. VIII.
Introduction HSPGs and Cancer FGF Signaling and HSPGs in Drosophila Dpp Signaling and HSPGs in Drosophila Wg and Hh Signaling and HSPGs in Drosophila The Role of Glypicans in Wg Signaling HSPGs Are Involved in Hh Movement Conclusion References
I. INTRODUCTION The correct growth and development of multicellular organisms depends on the reception of numerous extracellular signals that activate various signal transduction cascades within the target cells. The activity of these pathways, such as the receptor tyrosine kinases (RTKs), transforming growth factor-β (TGF-β), Wnts, and Hedgehogs (Hh), are usually regulated by the binding of extracellular ligands to their transducing receptors. Aberrant regulation of these pathways has been linked to many human cancers. Although we have a fair understanding of the structure of these signaling pathways downstream of the receptors, we are only beginning to understand the complexity of the regulatory mechanisms that operate at the extracellular level. Recently, it has become clear that heparan sulfate proteoglycans (HSPGs), a diverse group of cell surface and extracellular matrix proteins, play a key role in modulating a wide range of signaling pathways at this level (Perrimon and Bernfield, 2000). HSPGs are composed of a protein core modified on specific serine residues by the addition of heparan sulfate (HS) glucosaminoglycans (GAGs) synthesized in the Golgi (Fig. 1; Salmivirta et al., 1996). The HS-GAGs are
Advances in CANCER RESEARCH 0065-230X/01 $35.00
67
C 2001 by Academic Press. Copyright All rights of reproduction in any form reserved.
Fig. 1 HSPGs biosynthesis. The substrates for HS biosynthesis, UDP sugars, are synthesized in the cytoplasm and transported into the Golgi by nucleotide sugar transporters. In Drosophila, this activity is encoded by fringe connection (E. M. Selva, unpublished results). See text for all other details and Salmivirta et al. (1996). (Modified with permission from Baeg and Perrimon, 2000.)
Role of Heparan Sulfate Proteoglycans in Cancer
69
defined by10–200 linear repeating disaccharide units of N-acetylglucosamine (GlcNAc) and glucuronic acid (GlcA) which are added to the growing HS chain by glycosyltransferases as UDP–sugar substrates. The serine residue of the core protein is attached directly to a xylose–galactose–galactose– GlcA tetrasaccharide linker which serves as the scaffold for the growth of the HS-GAG. Three distinct types of proteins can serve as the HSPG cores: the transmembrane proteins encoded by the syndecan genes, the glycosylphosphatidylinositol (GPI) membrane-bound glypicans, and the extracellular matrix secreted perlecan proteins. Extensive postsynthetic modification of the linear HS sugar chains further increases the complexity of HSPGs. Alterations of the linear HS-GAG chains occur in a stepwise manner beginning with N-sulfation of GlcNAc catalyzed by an N-deacetylase/ N-sulfotransferase, followed by epimerization events and O-sulfation at the C2, C6, and C3 positions of the hexose sugar backbone. Thus, HS sugar chains can be heterogeneously decorated with negatively charged sulfate groups and sugar epimers within any given sugar chain, leading to the potential for a wide range of molecular diversity even among a common protein core. It is thought that the molecular diversity of HSPGs allows for their participation in a wide range of different signaling events and allows them to exert their influence on individual signaling pathways through unique mechanisms. In the past 3 years, two major advances have been made in understanding the role of HSPGs in development and cancer. First, studies in Drosophila have identified many mutations in either the biosynthetic enzymes [e.g., Sugarless (Sgl), Sulfateless (Sfl), and Tout velu (Ttv); Fig. 1] or the protein cores (e.g., syndecans and glypicans). Analysis of the mutant phenotypes has revealed the critical roles of HSPGs in modulating various growth factor signaling pathways. Second, many mutations linked to human cancers have been isolated and shown to correspond to defects in the biosynthesis of HSPGs. Altogether, the studies reviewed here underscore the importance of HSPGs in cell signaling and provide insights into their functions. Possibly, modulating the activity of HSPGs could influence the development of cancers caused by aberrant cell signaling. Thus, HSPGs may constitute important targets for therapeutics to treat some human tumors.
II. HSPGs AND CANCER The fidelity of HSPG biosynthesis has been shown to be important for the proper activity of many signaling pathways. These include the pathways regulated by the fibroblast growth factor (FGF), TGF-β, Wnt/Wingless (Wg),
70
Erica M. Selva and Norbert Perrimon
and Hh ligands. In Drosophila, as well as in other organisms, these signaling pathways participate in a multitude of proliferative and differentiation events during growth and development. Indeed, aberrant regulation of many of these signaling pathways leads to uncontrolled cell growth associated with the various types of cancers found in vertebrates. For example, ectopic expression of the Wnt-1 oncogene has long been known to lead to mammary tumors in mice (Nusse and Varmus, 1982). Loss-of-function mutations in adenomatous polyposis coli, a downstream effector and negative regulator of the Wnt signaling pathway, are the most common genetic lesions found in colon cancers (Kinzler and Vogelstein, 1996; Polakis, 1997). Mutations in the Hh pathway have also been implicated in human cancers. For example, human carcinomas have been linked to loss-of-function mutations in patched (ptc). Ptc encodes the Hh receptor and acts as a negative regulator of the pathway such that ptc loss-of-function mutations are associated with constitutive signaling (Johnson et al., 1996). Consistent with these findings, Cubitus Interruptus (Ci/Gli), a transcription factor and positive transducer of Hh signaling, is amplified in various types of cancer (Kinzler et al., 1987; Roberts et al., 1989). Finally, TGF-β signaling in mammals acts to both suppress and promote tumorigenesis. Therefore, loss-of-function mutations at any given step in this pathway could result in uncontrolled cell proliferation leading to cancer, whereas aberrant expression of positive regulators of the pathway, such as TGF-β, could result in the same outcome depending on the context. Indeed, loss-of-function mutations in the receptor and downstream effectors of TGF-β signaling have been observed in various cancers, as have high levels of TGF-β expression (Massague et al., 2000). Mutations in genes involved in the biosynthesis of HSPGs have also been directly implicated in human tumors, consistent with the critical role of HSPGs in regulating the previously mentioned pathways. Two multiple hereditary exostosis (Ext) genes have been identified as putative tumor suppressor genes. Loss-of-function mutations in either the Ext1 or Ext2 genes are associated with bony outgrowths (exostosis) that can undergo malignant transformation into chondrosarcomas (Hennekam, 1991; Leone et al., 1987) and osteosarcomas (Schmale et al., 1994; Wicklund et al., 1995). Ext1 and Ext2 have subsequently been shown to encode a Golgi-localized glycosyltransferase complex required to catalyze the polymerization of UDP–GlcA and UDP–GlcNAc into linear HS chains (McCormick et al., 2000). Furthermore, loss of glypican-3, a GPI-linked HSPG core protein, results in the Simpson–Golabi syndrome, which is characterized by pre- and postnatal overgrowths and a variety of dysmorphisms (Pilia et al., 1996). Finally, a direct potentiating role has been ascribed to HSPGs in Wnt-1-mediated tumorigenesis since mammary hyperplasias were significantly reduced in wnt-1/syndecan-1 double-knockout mice (Alexander et al., 2000).
Role of Heparan Sulfate Proteoglycans in Cancer
71
III. FGF SIGNALING AND HSPGs IN Drosophila The FGF signaling cascade is prototypical of the RTK class of signaling pathways. The pathway is initiated by the binding of an extracellular FGF ligand to its cognate tyrosine kinase receptor (FGFR). Ligand binding induces receptor dimerization and subsequent transphosphorylation, which activates a phosphorylation cascade that includes mitogen-activated protein kinase. A large body of work has shown that HS is required for both ligand binding and signal transduction, underscoring the importance of HSPGs in mediating the FGF signal (Ornitz et al., 1992; Rapraeger et al., 1991; Yayon et al., 1991). However, it has only recently been demonstrated through genetic studies in Drosophila that HSPGs are required in vivo to promote FGF-dependent developmental signaling events (Lin et al., 1999). In the Drosophila embryo, homologs of both FGF and FGFR are required for two important developmental events; dorsolateral migration of mesodermal cells (Figs. 2A and 2E) (Beiman et al., 1996; Gisselbrecht et al., 1996; Michelson et al., 1998; Shishido et al., 1993) and tracheal morphogenesis (Lee et al., 1996; Sutherland et al., 1996). Mesodermal migration requires an unknown FGF ligand and the FGFR Heartless (Htl), whereas tracheal migration involves the FGF ligand Branchless (Bnl) and the FGFR Breathless (Btl). In htl mutant embryos, mesodermal cells pile up at the ventral midline (Figs. 2B and 2F) (Beiman et al., 1996; Gisselbrecht et al., 1996), a phenotype that can be partially rescued by the expression of an activated form of Htl (Fig. 2J). Embryos that lack both the maternal and zygotic activities of sgl or sfl exhibit mesodermal defects identical to those observed in htl mutant embryos (Figs. 2C, 2D, 2G, 2H, 2K, and 2L). Furthermore, the tracheal defects in sgl mutants can be partially rescued by activated Htl, supporting the notion that HSPGs are required upstream of the receptor. The general requirement for HSPG biosynthesis in FGF signaling is also revealed by the observation that zygotic sgl and sfl mutations yield tracheal branching phenotypes similar to those found in btl and bnl mutants (Lin et al., 1999). The nature of the protein core that carries the HS-GAG chains is not known, and it will be interesting to determine whether the same protein is involved in both FGF pathways. The recent crystallographic structure of an FGF–FGFR–HS ternary complex further demonstrates the importance of HSPGs in FGF signaling (Pellegrini et al., 2000; Plotnikov et al., 1999; Schlessinger et al., 2000). Negatively charged heparin (a highly sulfated HS) is bound in a canyon of positive charge (Fig. 3A, shown in blue; see color insert) that is created by the dimerization of FGF receptor–ligand binary complexes. HS forms hydrogen bonds with the FGF–FGFR binary complex and with FGFR from the opposing binary complex [Fig. 3B (see color insert); Plotnikov et al., 1999;
72
Erica M. Selva and Norbert Perrimon
Fig. 2 HSPGs biosynthesis is required for dorsolateral migration of mesodermal cells. (A–D) Ventral view of late stage 9 embryos stained with Twist (Twi), a marker of mesodermal cells. Embryos devoid of both maternal and zygotic expression (germline clone embryos) of sgl and sfl are identified as mat + zyg. Anterior is to the left. (E–H) Transverse sections of Twi-stained early stage 10 embryos. In wild-type embryos, Twi-positive mesodermal cells have completed their dorsolateral migration, whereas in zygotic null htl embryos these cells accumulate at the ventral midline. The same phenotype is also observed for sgl and sfl germline clone-derived embryos. (I–L) Transverse sections of Twi-stained embryos expressing an activated allele of htl. (Reproduced with permission from Lin et al., 1999.)
Schlessinger et al., 2000]. Furthermore, these hydrogen bonds arise primarily from the HS N- and O-sulfate groups (Schlessinger et al., 2000), demonstrating the significance of postsynthetic modification in the specificity of HS molecular interactions. The previous data suggest a model for the role of HSPGs in the activation of FGF signaling in vivo. In the signal-receiving cells, HSPGs are likely to both stabilize the FGF–FGFR binary complexes and promote dimerization to yield the active ternary complex. Thus, in the
Role of Heparan Sulfate Proteoglycans in Cancer
73
context of FGF signaling, the HSPG acts as a coreceptor to facilitate the interaction between the FGF ligands and the FGFR transducing receptors.
IV. Dpp SIGNALING AND HSPGs IN Drosophila In Drosophila, decapentaplegic (dpp) encodes a member of the TGF-β/ bone morphogenetic protein (BMP) family, and numerous studies have established that the signaling pathway activated by Dpp is evolutionary conserved. The extracellular ligand, Dpp, binds to its heterodimeric types I and II serine/threonine kinase receptor to initiate signaling (Brummel et al., 1994; Penton et al., 1994; Ruberte et al., 1995). The glypican protein, called Dally, has been shown to play a role in Dpp-dependent imaginal disc patterning (Jackson et al., 1997). Through genetic interaction studies, Dally was found to be required downstream of Dpp for signal transduction; when overexpressed, Dally was able to amplify the outcome of Dpp signaling. These observations suggest that Dally can potentiate the activity of Dpp; however, the precise mechanism of HSPG action in Dpp signaling is not understood. Interestingly, in vertebrate cells, cross-linking studies have shown that the HSPG betaglycan interacts with TGF-β and promotes binding to the signaling receptor (Lopez-Casillas et al., 1993). Based on these results, it will be of interest to determine whether vertebrate glypicans can regulate some aspects of TGF-β /BMP signaling.
V. Wg AND Hh SIGNALING AND HSPGs IN Drosophila A member of the Wnt family of secreted glycoproteins has been implicated in many events during Drosophila embryogenesis, including segmentation of the epidermis, segmental patterning of the midgut epithelium, formation of the stomatogastric nervous system, neuroblast determination and differentiation, the control of cellular proliferation during Malpighian tubule formation, and generation of epithelial cell type diversity. One function of Wg during embryonic segmentation is to stabilize the expression of both the homeobox gene engrailed (en) (Ingham and Martinez Arias, 1992; Perrimon, 1994) and the signaling molecule Hh. In the early embryo, wg is expressed in stripes of epidermal cells that are immediately adjacent and anterior to cells expressing both en and hh. The juxtaposition of en/hh and wg-expressing cells is crucial for the formation of alternating bands of naked cuticle and denticles within each segmental unit (Fig. 4; for Figure 4A, see color insert). Expression of both en/hh and wg is first initiated as the result of complex
74
Erica M. Selva and Norbert Perrimon
Fig. 4 Model for the role of Wg and En / Hh signaling during embryonic segmentation and the concurrent role of HSPGs. (A) See color insert. (B) Table summarizing the effect of mutations that disrupt the Wg or En/Hh signaling cascade. A “+” indicates that a given phase can occur in the absence of the indicated genes, whereas a “−” indicates a block.
regulatory interactions between the pair-rule genes (Fig. 4A, phase 1). Subsequently, the maintenance of both en/hh and wg expression becomes mutually dependent at stage 8 of embryonic development until early stage 11 (Fig. 4A, phase 2). In wg mutant embryos, en/hh expression fades from the epidermis, and in en and hh mutant embryos epidermal wg expression disappears due to the absence of Hh signaling (Fig. 4B; Bejsovec and Martinez Arias, 1991; DiNardo et al., 1988; Heemskerk et al., 1991; Martinez-Arias et al., 1988). At a later stage of embryonic development, Wg also signals anteriorly to promote the differentiation of epithelial cells to secrete cuticle that lacks denticle bands (Fig. 4A, phase 3; Sanson et al., 1999). Thus, the absence of either wg or en/hh function at early stages of embryonic development disrupts overall patterning, generating embryos with a lawn of denticles and reduced size along their anterior–posterior (AP) axis. Loss of Wg activity at later stages of embryogenesis affects long-range patterning and prevents the deposition of naked cuticle to yield a lawn of denticles (Fig. 4B). Mutations in sgl, sfl, and ttv were originally identified based on their phenotypic similarities to the wg/hh loss-of-function mutations (Bellaiche et al., ¨ 1998; Binari and Perrimon, 1994; Hacker et al., 1997; Haerry et al., 1997; Lin et al., 1999; Perrimon et al., 1996). Mutations in these genes that remove both maternal and zygotic activities are associated with segmentation phenotypes that resemble the loss of either Wg or Hh signaling. However, because of the interdependence of the Hh and Wg signaling pathways, it was necessary to analyze the role of these genes in tissues in which they do not regulate each other in order to determine which pathway(s) requires HSPGs for signaling. Such a situation is found in the imaginal disc in which these genes
Role of Heparan Sulfate Proteoglycans in Cancer
75
control different processes. In the wing disc, Wg organizes patterning along the dorsoventral (DV) boundary and Hh controls AP axis formation (Fig. 5; see color insert).
VI. THE ROLE OF GLYPICANS IN Wg SIGNALING In the wing imaginal disc, wg is expressed at the DV boundary, where it acts as an organizer to control the growth of the wing blade and differentiation of the wing margin. When groups of homozygous wg mutant cells are generated in the wing disc (also known as clones), wings develop nicks that overlap the wing margin. A similar phenotype is generated when clones of sfl mutant cells are induced (Baeg et al., 2001). However, this is not the case when clones of sgl mutant cells are generated, presumably because GlcA, the product of Sgl activity, is able to freely diffuse between cells. The phenotype associated with sfl mutant clones, together with its maternal effect phenotype, suggests that HSPGs are required for Wg signaling. Further support for this model derives from the analysis of both the loss- and gain-of-function phenotypes associated with the glypican molecules Dally and Dally-like (Dly) (Baeg et al., 2001; Lin and Perrimon, 1999; Tsuda et al., 1999). For example, dally homozygous mutants show a low penetrance of wing nicks at the wing margin. Furthermore, this phenotype can be enhanced by reducing the amount of extracellular Wg, and it can be suppressed by introducing an activated downstream Wg signaling component (Lin and Perrimon, 1999). How do the HSPGs work in the context of Wg signaling? Interestingly, homozygous sfl mutant clones that span the DV boundary of the wing disc did not disrupt the expression (and presumably secretion) of Wg. The distribution of Wg in sfl mutant clones is indistinguishable from surrounding wild-type tissue. However, a highly sensitive staining method that detects extracellular Wg showed that Wg was not present at the surface within clones (Baeg et al., 2001). This suggests that HSPGs are required to restrict Wg diffusion and thus may serve to trap extracellular Wg (Fig. 6). In support of this model, sfl mutant cells located near wild-type Wg-secreting cells display Wg staining to some extent and thereby show local nonautonomy in mosaic analyses. Another possibility is that HSPGs may increase the local concentration of Wg by preventing its degradation by extracellular proteases. Furthermore, the local nonautonomy of sfl mutant clones suggests that the HSPGs are not absolutely required for Wg association with Frizzled, its transducing receptor (Bhanot et al., 1996; Chen and Struhl, 1999). Finally, consistent with this model, overexpression of Dly along the DV boundary yields wg loss-offunction phenotype, presumably because Wg is not free to diffuse to its site of action but rather becomes sequestered by the high local concentration of Dly
76
Erica M. Selva and Norbert Perrimon
Fig. 6 The influence of HSPGs on Wg signaling in imaginal discs. The binding of extracellular Wg to HSPGs appears to serve two functions. HSPGs facilitate the organization of the extracellular Wg gradient in the wing pouch and promote the interaction of Wg with its signaling receptor, Frizzled (Dfz2).
(Baeg et al., 2001). The proposed role of HSPGs in the context of Wg signaling is distinct from the role of HSPGs in FGF signaling because it does not require an association between the HS-GAG and the ligand/receptor complex. In the embryo, the segmentation phenotype associated with loss of maternal and zygotic sgl and sfl suggests that a core HSPG protein must also be important for phase 2 wg maintenance and phase 3 Wg-dependent secretion of naked cuticle. The glypican Dally has been proposed to correspond to this protein because weak dally alleles and dally RNA interference (RNAi) experiments generate embryos reminiscent of partial loss of wg function (Baeg et al., 2001; Lin and Perrimon, 1999; Tsuda et al., 1999). Furthermore, dally misexpression results in an expansion of the en/hh domain (Tsuda et al., 1999). These observations suggest that Dally may be involved in posterior Wg signaling to maintain en/hh expression as well as to promote Wg signaling anteriorly. Recently, RNAi experiments with dly implicated HSPG in long-range Wg patterning but not during the initiation or maintenance phase (Baeg et al., 2001). Interestingly, both dally and dly are expressed at high levels anterior to the wg-expressing cells (Khare and Baumgartner, 2000; Lin and Perrimon, 1999), which is consistent with these genes playing a major role in the organization of the anterior Wg activity. Together, these data suggest a model in which wg and en/hh expression is initiated by pair-rule gene expression (Fig. 7; see color insert). Subsequently, maintenance of wg and en/hh may require the HSPG Dally. Finally, anterior long-range Wg signaling to pattern the naked cuticle may require the action of Dly and, to a lesser extent, Dally. It should be noted that some of these conclusions are not as definitive as one would like since they were based on partial loss-of-function alleles or overexpression experiments. Thus, it will be critical to reexamine these issues once null alleles for either dally or dly become available.
Role of Heparan Sulfate Proteoglycans in Cancer
77
VII. HSPGs ARE INVOLVED IN Hh MOVEMENT Much of what we know about the role of HSPGs in Hh signaling comes from the analysis of the ttv mutant phenotype in wing imaginal discs. Ttv/ Ext has been shown to play an important role in the movement of Hh from its site of synthesis in the posterior compartment of the wing disc to its site of action in the anterior compartment (Bellaiche et al., 1998). Homozygous mutant clones of ttv along the AP boundary show reduced Ptc expression and Ci stabilization, two targets of Hh signaling. Hh is a very unusual protein because it undergoes an autoprocessing event whereby a cholesterol moiety is attached to its N terminus to produce the active ligand (Lee et al., 1994; Porter et al., 1996). The linkage of cholesterol to Hh decreases its solubility and tethers the molecule to the membrane, presumably limiting its long-range diffusion. Recently, dispatched (disp), which encodes an extracellular membrane protein with a sterol-sensing domain, was found to be required for the release of Hh from sending cells since Hh was retained in clones of disp mutant cells (Burke et al., 1999). The current model suggests that Disp is required in the posterior cells to transfer cholesterol-bound Hh to an unidentified anterior compartment HSPG which requires Ttv for its appropriate biosynthesis (Fig. 8). In turn, this complex directly or indirectly transfers Hh to its receptor Ptc to transduce the Hh signal. Whether the ttv-dependent HSPG is sufficient for Hh movement within the anterior compartment awaits further investigation. Finally, it is possible that HSPGs also play a more direct role in Hh signaling. In ttv embryos that are devoid of both maternal and zygotic gene activity, wg expression decays because of defective Hh signaling. Thus, even when Hh signals to immediate neighboring cells, HSPGs may be required for regulation of Ptc by Hh. Furthermore, Ttv, which encodes a glycosyltransferase by analogy to mammalian Exts, is surprisingly specific to Hh signaling (The et al., 1999). Careful characterization of the ttv mutant phenotype failed to reveal a function for Ttv in either the Wg or FGF signaling pathways. This is unexpected because Ttv encodes a polymerase involved in HS-GAG chain biosynthesis; therefore, it should have a phenotype similar to those of either sgl or sfl mutants (see Fig. 1). The reason why Ttv is specific to Hh signaling is not resolved, and many possibilities that include either qualitative or quantitative models can be considered. For example, one quantitative model is that in the absence of Ttv activity, a reduced amount of HSPGs are synthesized by other Ext enzymes and Hh signaling is much more sensitive to this reduction than either the FGF or Wg pathways. Alternatively, a qualitative model suggests that specific Exts may only modify a subset of protein cores such that in ttv mutants the HSPGs involved in Wg and FGF signaling pathways are modified properly, but the HSPGs implicated in Hh signaling are not.
78
Erica M. Selva and Norbert Perrimon
Fig. 8 The role of HSPGs in Hh signaling in imaginal discs. The biologically active Hh ligand is synthesized in the posterior compartment of the wing imaginal disc. The N-terminal region of Hh is linked to a cholesterol moiety which may be localized to microdomains (rafts) within the membrane (Rietveld et al., 1999). Appropriate movement of Hh from the posterior compartment to the anterior compartment requires the activity of Disp with its sterol-sensing domain. Subsequently, Hh requires the glycosyltransferase activity of Ttv for movement within the anterior compartment, presumably by adding HS to an unknown core protein. The movement of Hh from the posterior to the anterior compartment might require a direct transfer between Disp and the HSPG. Once in the anterior compartment, Hh must interact with its receptor Ptc to initiate the downstream signaling cascade; this interaction may also involve a Ttv-dependent HSPG. (Adapted from Ingham, 2000.)
VIII. CONCLUSION Recent studies of HSPGs have implicated these molecules as key players in regulation of cell–cell communication events. Interestingly, in all the pathways examined to date they appear to positively regulate signaling events. Although the requirement for HSPGs is well documented, the precise mechanisms by which they act remains obscure. An understanding of the molecular interactions between HS-GAGs and growth factors could potentially provide an excellent means to interfere with specific pathway activities and to develop therapeutics that act extracellularly.
Role of Heparan Sulfate Proteoglycans in Cancer
79
ACKNOWLEDGMENTS We thank David Bilder, Inge The, Peter Rapiejko, and Gyeong-Hun Baeg for comments on the manuscript and S. Hubbard for the materials shown in Fig. 3. This work is supported by a National Institutes of Health grant. N.P is an investigator of the Howard Hughes Medical Institute.
REFERENCES Alexander, C. M., Reichsman, F., Hinkes, M. T., Lincecum, J., Becker, K. A., Cumberledge, S., and Bernfield, M. (2000). Nature Genet. 25, 329–332. Baeg, G., and Perrimon, N. (2000). Curr. Opin. Cell Biol. 12, 575–580. Baeg, G.-H., Lin, X., Khare, N., Baumgartner, S., and Perrimon, N. (2001). Development 128, in press. Beiman, M., Shilo, B. Z., and Volk, T. (1996). Genes Dev. 10, 2993–3002. Bejsovec, A., and Martinez Arias, A. (1991). Development 113, 471– 485. Bellaiche, Y., The, I., and Perrimon, N. (1998). Nature 394, 85–88. Bhanot, P. B. M., Samos, C. H., Hsieh, J. C., Wang, Y., Macke, J. P., Andrew, D., Nathans, J., and Nusse, R. (1996). Nature 382, 225–230. Binari, R., and Perrimon, N. (1994). Genes Dev. 8, 300–312. Brummel, T. J., Twombly, V., Marques, G., Wrana, J. L., Newfeld, S. J., Attisano, L., Massague, J., O’Connor, M. B., and Gelbart, W. M. (1994). Cell 78, 251–261. Burke, R., Nellen, D., Bellotto, M., Hafen, E., Senti, K. A., Dickson, B. J., and Basler, K. (1999). Cell 99, 803–815. Chen, C. M., and Struhl, G. (1999). Development 126, 5441–5452. DiNardo, S., Sher, E., Heemskerk, J. J., Kassis, J. A., and O’Farrell, P. H. (1988). Nature 332, 604–609. Gisselbrecht, S., Skeath, J. B., Doe, C. Q., and Michelson, A. M. (1996). Genes Dev. 10, 3003– 3017. ¨ Hacker, U., Lin, X., and Perrimon, N. (1997). Development 124, 3565–3573. Haerry, T. E., Heslip, T. R., Marsh, J. L., and O’Connor, M. B. (1997). Development 124, 3055–3064. Heemskerk, J., DiNardo, S., Kostriken, R., and O’Farrell, P. H. (1991). Nature 352, 404–410. Hennekam, R. C. (1991). J. Med. Genet. 28, 262–266. Ingham, P. W. (2000). Curr. Biol. 10, R180–183. Ingham, P. W., and Martinez Arias, A. (1992). Cell 68, 221–235. Jackson, S. M., Nakato, H., Sugiura, M., Jannuzi, A., Oakes, R., Kaluza, V., Golden, C., and Selleck, S. B. (1997). Development 124, 4113–4120. Johnson, R. L., Rothman, A. L., Xie, J., Goodrich, L. V., Bare, J. W., Bonifas, J. M., Quinn, A. G., Myers, R. M., Cox, D. R., Epstein, E. H., Jr., and Scott, M. P. (1996). Science 272, 1668–1671. Khare, N., and Baumgartner, S. (2000). Mech. Dev. 99, 199–202. Kinzler, K. W., and Vogelstein, B. (1996). Cell 87, 159–170. Kinzler, K. W., Bigner, S. H., Bigner, D. D., Trent, J. M., Law, M. L., O’Brien, S. J., Wong, A. J., and Vogelstein, B. (1987). Science 236, 70–73. Lee, J. J., Ekker, S. C., von Kessler, D. P., Porter, J. A., Sun, B. I., and Beachy, P. A. (1994). Science 266, 1528–1537.
80
Erica M. Selva and Norbert Perrimon
Lee, T., Hacohen, N., Krasnow, M., and Montell, D. J. (1996). Genes Dev. 10, 2912–2921. Leone, N. C., Shupe, J. L., Gardner, E. J., Millar, E. A., Olson, A. E., and Phillips, E. C. (1987). J. Hered. 78, 171–177. Lin, X., and Perrimon, N. (1999). Nature 400, 281–284. Lin, X., Buff, E. M., Perrimon, N., and Michelson, A. M. (1999). Development 126, 3715– 3723. Lopez-Casillas, F., Wrana, J. L., and Massague, J. (1993). Cell 73, 1435–1444. Martinez-Arias, A., Baker, N., and Ingham, P. W. (1988). Development 103, 157–170. Massague, J., Blain, S. W., and Lo, R. S. (2000). Cell 103, 295–309. McCormick, C., Duncan, G., Goutsos, K. T., and Tufaro, F. (2000). Proc. Natl. Acad. Sci. USA 97, 668–673. Michelson, A. M., Gisselbrecht, S., Zhou, Y., Baek, K. H., and Buff, E. M. (1998). Dev. Genet. 22, 212–229. Nusse, R., and Varmus, H. E. (1982). Cell 31, 99–109. Ornitz, D. M., Yayon, A., Flanagan, J. G., Svahn, C. M., Levi, E., and Leder, P. (1992). Mol. Cell. Biol. 12, 240–247. Pellegrini, L., Burke, D. F., von Delft, F., Mulloy, B., and Blundell, T. L. (2000). Nature 407, 1029–1034. Penton, A., Chen, Y., Staehling-Hampton, K., Wrana, J. L., Attisano, L., Szidonya, J., Cassill, J. A., Massague, J., and Hoffmann, F. M. (1994). Cell 78, 239–250. Perrimon, N. (1994). Cell 76, 781–784. Perrimon, N., and Bernfield, M. (2000). Nature 404, 725–729. Perrimon, N., Lanjuin, A., Arnold, C., and Noll, E. (1996). Genetics 144, 1681–1692. Pilia, G., Hughes-Benzie, R. M., MacKenzie, A., Baybayan, P., Chen, E. Y., Huber, R., Neri, G., Cao, A., Forabosco, A., and Schlessinger, D. (1996). Nature Genet. 12, 241–247. Plotnikov, A. N., Schlessinger, J., Hubbard, S. R., and Mohammadi, M. (1999). Cell 98, 641– 650. Polakis, P. (1997). Biochim. Biophys. Acta 1332, F127–F147. Porter, J. A., Young, K. E., and Beachy, P. A. (1996). Science 274, 255–259. Rapraeger, A. C., Krufka, A., and Olwin, B. B. (1991). Science 252, 1705–1708. Rietveld, A., Neutz, S., Simons, K., and Eaton, S. (1999). J. Biol. Chem. 274, 12049–12054. Roberts, W. M., Douglass, E. C., Peiper, S. C., Houghton, P. J., and Look, A. T. (1989). Cancer Res. 49, 5407–5413. Ruberte, E., Marty, T., Nellen, D., Affolter, M., and Basler, K. (1995). Cell 80, 889–897. Salmivirta, M., Lidholt, K., and Lindahl, U. (1996). FASEB J. 10, 1270–1279. Sanson, B., Alexandre, C., Fascetti, N., and Vincent, J. P. (1999). Cell 98, 207–216. Schlessinger, J., Plotnikov, A. N., Ibrahimi, O. A., Eliseenkova, A. V., Yeh, B. K., Yayon, A., Linhardt, R. J., and Mohammadi, M. (2000). Mol. Cell. 6, 743–750. Schmale, G. A., Conrad, E. U. III, and Raskind, W. H. (1994). J. Bone Joint Surg. Am. 76, 986–992. Shishido, E., Higashijima, S., Emori, Y., and Saigo, K. (1993). Development 117, 751–761. Sutherland, D., Samakovlis, C., and Krasnow, M. A. (1996). Cell 87, 1091–1101. The, I., Bellaiche, Y., and Perrimon, N. (1999). Mol. Cell. 4, 633–639. Tsuda, M., Kamimura, K., Nakato, H., Archer, M., Staatz, W., Fox, B., Humphrey, M., Olson, S., Futch, T., Kaluza, V., Siegfried, E., Stam, L., and Selleck, S. B. (1999). Nature 400, 276– 780. Wicklund, C. L., Pauli, R. M., Johnston, D., and Hecht, J. T. (1995). Am. J. Med. Genet. 55, 43– 46. Yayon, A., Klagsbrun, M., Esko, J. D., Leder, P., and Ornitz, D. M. (1991). Cell 64, 841–848.
The Occurrence and Significance of V Gene Mutations in B Cell–Derived Human Malignancy Freda K. Stevenson, Surinder S. Sahota, Christian H. Ottensmeier, Delin Zhu, Francesco Forconi, and Terry J. Hamblin Molecular Immunology Group, Tenovus Laboratory Southampton University Hospitals Trust Southampton SO16 6YD, United Kingdom
I. Introduction II. Immunoglobulin Genes in Normal B Cell Development A. V Gene Recombination and Selection B. Somatic Mutation and Isotype Switch III. Immunoglobulin Genes in B Cell Tumors A. VH Gene Usage by B Cell Tumors IV. Somatic Mutation in B Cell Tumors V. Chronic Lymphocytic Leukemia A. V Gene Mutational Status B. Prognostic Value of V Gene Status C. VH Gene Usage VI. Follicular Lymphoma VII. Diffuse Large B Cell Lymphoma VIII. Plasma Cell Tumors A. V Gene Mutational Status B. V Gene Usage C. Ig Locus and Chromosomal Translocations IX. Conclusion References
The classification of B cell tumors has relevance for refining and improving clinical strategies. However, consensus has been difficult to establish, and although a scheme is now available, objective criteria are desirable. Genetic technology will underpin and extend current knowledge, and it is certain to reveal further subdivisions of current tumor categories. The Ig variable region genes of B cell tumors present a considerable asset for this area of investigation. The unique sequences carried in neoplastic B cells are easily isolated and sequenced. In addition to acting as clone-specific markers of each tumor, they indicate where the cell has come from and track its history following transformation. There is emerging clinical value in knowing whether the cell of origin has encountered antigen and has moved from the naive compartment
Advances in CANCER RESEARCH 0065-230X/01 $35.00
81
C 2001 by Academic Press. Copyright All rights of reproduction in any form reserved.
82
Freda K. Stevenson et al.
to the germinal center, where somatic mutation is activated. This is amply illustrated by the subdivision of chronic lymphocytic leukemia into two subsets, unmutated or mutated, each with very different prognosis. Other tumors may be subdivided in a similar way. Microarray technology is developing rapidly to probe gene expression and to further divide tumor categories. All these genetic analyses will provide objective data to enhance both our understanding of B cell tumors and our ability to treat them. C
2001 Academic Press.
I. INTRODUCTION In clinical practice, B cell tumors include two broad categories, leukemia and lymphoma—terms which relate to the main location of tumor in blood or tissue, respectively. The leukemias have always been attractive for study because of easy availability of tumor cells. In fact, only recently has the tendency to consider B cell tumors as two separate entities been modified. For B cell-derived lymphomas, pathologists and clinicians have for many years searched for a clear system of classification. The need for this has been twofold, with pathologists usually trying both to make a diagnosis and to understand the nature of the tumor cell and clinicians generally seeking advice on which treatment protocol to use for the patient. Classifications have reflected this duality and have sometimes differed between Europe and the United States. Early classification relied on cell morphology, from which differences of opinion between pathologists could arise. More objective criteria were supplied by immunophenotypic analysis, with a wide range of monoclonal antibodies available for investigating most biopsy material. This approach was also quickly adopted for the study of the leukemias. Eventually, the REAL classification, which includes both leukemias and lymphomas, was agreed by the International Lymphoma Study Group (Chan et al., 1995), with subsequent modification and extension into the World Health Organization classification (Harris et al., 1999). There are currently 16 categories of B cell tumors, and since the malignant cells of Hodgkin’s disease have recently been recognized to be derived from B cells (Hummel et al., 1995; Kuppers et al., 1994), this number will increase. Although recent classifications have incorporated some genetic features, including chromosomal aberrations, the wealth of information becoming available from gene-based technology has not yet been included. The most obvious genes to be analyzed in B cell tumors are the immunoglobulin genes, which undergo a series of changes during B cell differentiation. B cell tumors largely preserve these changes after transformation, and analysis of the Ig gene status provides a clonal history of the tumor, both pre and posttransformation (Stevenson et al., 1998). In addition to acting as a
The Occurrence and Significance of V Gene Mutations
83
genetic marker for tracking tumor cells, Ig gene sequence information is leading to further subdivisions of tumors previously considered as single entities. Some of these are demonstrating clear prognostic value. This review focuses on the information emerging from Ig gene analysis of B cell tumors and discusses how it is changing our view of tumor cell behavior, both in the clinic and in the laboratory. However, Ig gene analysis is only the beginning of the application of gene-based technology to tumors. Expression profiles of a range of normal or abnormal genes will certainly differ among or within tumor categories (Alizadeh et al., 2000). Microarray technology should add further prognostic value and provide data relevant to disease progression or response to treatment. The future is likely to reveal subdivision of tumors to such an extent that treatment may eventually be tailored to the individual patient rather than to the current perceived category.
II. IMMUNOGLOBULIN GENES IN NORMAL B CELL DEVELOPMENT A. V Gene Recombination and Selection Expression of immunoglobulin (Ig) is the defining feature of B cells. In normal B lymphocytes, Ig embedded in the surface membrane has the clear function of recognizing and responding to exogenous antigens. Recognition is via the variable (V) regions, which differ in sequence from one B cell to another and provide a wide protective antibody cover against invading pathogens. The binding range of the available Ig heavy-chain variable region gene repertoire is vastly extended by cutting and pasting of the component VH, D, and JH gene segments (Alt et al., 1987; Tonegawa, 1983). A similar rearrangement occurs for the Ig light-chain variable regions, involving singlestep recombinations of Vκ/Jκ or Vλ/Jλ gene segments but with no D segment genes. Selection of each component of the heavy chain takes place from the potentially functional genes in the unrearranged repertoire, with ∼51 VH genes divided into seven families (VH1–VH7), ∼27 D segment genes, and 6 JH genes available. There are two types of polymorphisms at the VH locus involving insertions or deletions of gene segments, with the number of functional segments dependent on the haplotype. In addition, there is evidence for a low degree of allelic polymorphism (Cook and Tomlinson, 1995). The sequential steps involved in the recombinatorial processes (Fig. 1) lead to transcriptional VH–D–JH–CH and VL–JL–CL units. These are translated into heavy- and light-chain proteins, which are able to combine to form whole IgM in the mature B cell.
84
Freda K. Stevenson et al.
Fig. 1 Schematic representation of the genetic recombinatorial events involved in generating an individual Ig molecule. For the Ig heavy chain, the VH and D genes in unrearranged DNA undergo a two-step rearrangement process, with excision of the intervening DNA. The first step places a D segment gene next to a JH gene, and the second step places a VH gene upstream of the D–JH union, forming a VHDJH sequence. Imprecision in the joints, with deletion and insertion of nucleotides, leads to a unique third complementarity determining sequence (CDR3). Transcription is now activated due to the influence of the upstream VH promoter and the intronic enhancer located between JH and Cµ. A similar process occurs for the Ig light chain, with generation of a VL–JL transcriptional sequence. Following processing of the RNA, translation leads to heavy and light chains with unique variable region genes, which combine to form an Ig molecule.
Analysis of the rearranged VH gene repertoire in normal B cells in the blood of three individuals revealed that there is preferential usage of certain genes in the primary Ig repertoire (Brezinschek et al., 1995). Using single-cell analysis, overrepresentation of the VH3 family was observed, largely due to preferential usage of a small number of specific family members. Within the VH3 family, the V3-23 gene (DP-47) was the most commonly used, possibly due to duplication of this segment in some haplotypes (Rubinstein et al., 1993). This was followed by the V3-30.3 (DP-46), V3-30 (DP-49), and V3-07 (DP-79) genes (Brezinschek et al., 1997). Knowledge of the normal B cellexpressed repertoire is providing an essential basis for assessment of biased usage of VH genes by tumor cells. Expansion of B cells expressing selected V genes may reflect an influence of B cell superantigens. These are similar to T cell superantigens in that they are able to bind to conserved framework sequences outside the conventional antigen binding site (Goodglick and Braun, 1994). Regarding T cells, several known B cell superantigens are
The Occurrence and Significance of V Gene Mutations
85
derived from pathogens. Biased usage of V genes in tumors may therefore provide a clue to the origin and pathogenesis of the tumor. Recombination of Ig gene elements is initiated in the bone marrow at the pro-B cell stage when selected DH and VH genes are moved, in two sequential steps, to the position upstream of the JH-constant region gene sequence (Fig. 1) (Wang et al., 1998). This process is mediated by proteins encoded by the recombination-activating genes RAG1 and RAG2. These proteins bind the two recombination signals, consisting of conserved heptamer and nonamer sequences separated by spacers, which flank the component genes (Lewis and Gellert, 1989; Schatz et al., 1992). The mechanism appears to parallel genetic transposition, with the excised transposon in this case being inactivated by joining of the two ends (Agrawal et al., 1998; Hiom et al., 1998). The maturational steps occurring in the bone marrow, and subsequently in the secondary lymphoid organs, are illustrated in Fig. 2. Also indicated are the points of differentiation reached by the cell of origin of the various
Fig. 2 Changes in immunoglobulin occurring during B cell maturation. Differentiation of B cells from pluripotential stem cells to Ig-secreting plasma cells involves rearrangement of Ig genes, with D–JH combination in pro-B cells being followed by VHDJH formation in pre-B cells. Prior to VL–JL rearrangement, heavy chains are expressed with surrogate light chain. Selected Ig-expressing B cells leave the bone marrow, and when antigen is encountered they locate in a germinal center where somatic mutation and isotype switching occur. Tumors can arise at various points of differentiation, and they carry the imprint of the Ig gene status of the cell of origin.
86
Freda K. Stevenson et al.
B cell tumors. Pre-B cells express Ig heavy chains together with the surrogate light chain encoded by the VpreB and λ5/14.1 genes (Wang et al., 1998). The µ heavy chain, surrogate light chain, and the α/β heterodimers form the pre-B receptor complex, which is able to sense the environment and is likely to act as a checkpoint for further differentiation. Heterogeneity in the resulting Ig binding site is generated not only due to combinatorial events but also due to the fact that the gene combinations are imprecise, with gain and loss of nucleotides. Addition of nucleotides can be via the terminal deoxynucleotide transferase (N nucleotides) or can result from hairpin structures in cleavage intermediates [palindromic (P) nucleotides]. The consequence of this process is that resulting sequences can be out of frame and therefore cannot make protein. The availability of a second allele allows another attempt at recombination, and many B cells have evidence in both VH and VL sequences of a first nonfunctional rearrangement which has occurred prior to a second successful attempt. Since nonfunctional sequences can accumulate somatic mutations, analysis of the nonfunctional sequences has been useful in assessing mutational patterns generated in the absence of the influence of antigen (Dorner et al., 1997). Analysis of functional recombinations has revealed that the D segment genes can be read in more than one reading frame (Corbett et al., 1997). The consequent protein sequence encoded by the D segment gene and the combination joints is therefore unique to each B cell. It is of variable sequence and length and is known as the complementarity determining region 3 (CDR3). It lies at the center of the antibody combining site and provides a clonal sequence signature for the B cell. Successful rearrangement of the light-chain genes then extinguishes expression of the pre-B cell complex and suppresses further rearrangement. Cells that fail to produce functional Ig will undergo apoptosis since Ig expression appears to be mandatory for survival in the periphery (Lam et al., 1997). Even cells expressing Ig must face a selective checkpoint in the bone marrow which removes autoreactive specificities. Escape from death can occur by changing specificity via receptor editing, a process involving secondary rearrangements of V genes, most evident in light chains (Gay et al., 1993; Nussenzweig, 1998). Although the majority of B cells show evidence for allelic exclusion, with expression of a single Ig structure, this may not be absolute. Normal B cells can occasionally express both κ and λ light-chain types, indicative of a failure of allelic exclusion in light chains (Giachino et al., 1995). Among B cell tumors, ∼5% of cases of chronic lymphocytic leukemia (CLL) express Ig of more than one VH subgroup, suggestive of a failure of allelic exclusion in heavy chains (Rassenti and Kipps, 1997). Following completion of the maturation program, the Ig-expressing B cell exits from the bone marrow and is ready to encounter an antigen which can bind to the selected V region sequence.
The Occurrence and Significance of V Gene Mutations
87
B. Somatic Mutation and Isotype Switch Once interaction with antigen has occurred, the B cell Ig undergoes affinity maturation, a process which generally occurs in the germinal centers of secondary lymphoid organs. In these sites, somatic mutation in the encoding V genes is activated under the influence of CD40L+ve T cells, cytokines, and antigen-bearing follicular dendritic cells (Berek, 1992; Kelsoe, 1995; Kosco et al., 1992; MacLennan, 1994). The rate of introduction of nontemplated base pair (bp) changes into the rearranged V genes is high, estimated at 10−4–10−3 bp per generation. The failure of this process has been revealed in hyper-IgM syndrome, which is associated with defective expression of CD40 ligand by T cells and the absence of germinal center formation. In these patients, B cells have only low levels of both somatic mutation and isotype switching events (Chu et al., 1995). Mutational activity is targeted to the V(D)J sequence, and it appears to be site specific rather than sequence specific since substitution of a V gene sequence by a β-globin sequence leads to accumulation of mutations in the latter (Yelamos et al., 1995). Mutations begin ∼180 bp downstream of the transcription initiation site (Klix et al., 1998), and activity is controlled by flanking regulatory elements (Neuberger et al., 1998). The process involves mainly single nucleotide changes, although deletions and duplications can also occur (Goossens et al., 1998; Wilson et al., 1998). Analysis of the process of somatic mutation has revealed that strand polarity is unlikely (Dorner et al., 1999) and that mutational activity is highly dependent on cell division (Toellner et al., 1996). During replication of B cells in germinal centers, when high levels of somatic mutation occur, transcription of Ig genes is diminished (Toellner et al., 1996). This may argue against the apparent link between transcription and mutational activity (Bross et al., 2000) and suggest instead that mutational activity is linked to cell division rather than to transcription (Dorner et al., 1999). Gene knockout mice have indicated that there may be two separate mutational mechanisms involving mismatch repair enzymes, affecting either frequency of mutations or targeting to “hot spots” (Dorner et al., 1999). As expected, mutational activity independent of antigen influence is especially evident in the nonfunctional sequences (Klein et al., 1998a,b). Clustering of replaced amino acids tends to occur in the CDRs for intrinsic structural reasons and possibly also due to antigen selection (Dorner et al., 1998). There have been many attempts to determine the influence of antigen selection from the distribution of mutations in either CDRs or framework regions (FWRs) (Chang and Casali, 1994; Lossos et al., 2000b). However, these analyses have been questioned due to the intrinsic hot spots (Betz et al., 1993) and to the fact that a single amino acid change can have a major effect on affinity
88
Freda K. Stevenson et al.
(Cook et al., 1982). The feature most consistently associated with antigen selection appears to be conservation of FWRs (Dorner et al., 1998). The outcome of the selective process is generation of antibody-secreting plasma cells and memory B cells, each with increased binding strength for antigen. The choice between generation of plasma cells or memory cells appears to be directed by cytokines, with interleukin-10 (IL-10) a critical factor favoring plasma cells (Choe and Choi, 1998). Unselected cells will again die by apoptosis, although there is controversial evidence that receptor editing can also occur in the periphery (Kelsoe, 1996; Meffre et al., 1998). Recently, it has been shown that RAG1 and RAG2 genes are coordinately expressed in tonsillar B cells, some of which are postswitched memory B cells which can exit to the blood (Girschick et al., 2001). Evidence for this process in rare normal B cells is provided by the finding of different light chains in combination with a common identical mutated heavy-chain sequence (de Wildt et al., 1999). A further characteristic of receptor editing is discordance between the level of somatic mutation in VH compared to that in VL, and there is an indication of this phenomenon in the autoimmune disease systemic lupus erythematosus (Mockridge et al., 1998). To date, this discordance has not been observed in B cell tumors, but a more comprehensive analysis of VL genes may be required to detect such cells. Recently, it has been suggested that receptor editing can occur in VH (Wilson et al., 2000), but the extent of this process is not known. A final genetic rearrangement is required for Ig class switching from the initial IgM(D) isotype to IgG, IgA, or IgE, each with a different antibody effector capacity. This process occurs in the germinal center (MacLennan, 1994), although it may not be confined to this site. The choice of isotype depends on the cytokine milieu, which influences transcription of the individual constant region genes (Stavnezer et al., 1985). The Ig locus includes an array of constant (CH) genes, each flanked at its 5′ region by a switch (S) region composed of tandem repetitive unit sequences with many palindromic motifs (Zhang et al., 1995). Conventional isotype switching occurs between two S regions, leading to a deletional looping out of intervening constant region genes (Bentley and Rabbits, 1980; Matsuoka et al., 1990). However, there is also evidence that RNA splicing may be used to generate multiple isotypes, and it has been suggested that this process may precede conventional deletional switching (Fujieda et al., 1996a; Perlmutter and Gilbert, 1984; Weiss et al., 1987). It is currently unclear whether somatic mutation continues at a significant level following isotype switch. However, study of a B cell line in vitro has shown that somatic mutation can be induced both before and after isotype switching and that the stimuli for each can be separated (Zan et al., 1999). In normal adults, the balance of the maturational pathways is obviously influenced by the status of the immune system and will be disturbed during
The Occurrence and Significance of V Gene Mutations
89
infection. However, in the blood, there appears to be a relatively stable cell profile in which naive cells with unmutated V genes represent ∼60% of B lymphocytes (Klein et al., 1998b). A small subpopulation of the naive population expresses CD5 (Brezinschek et al., 1997), and this proportion is high in early life (Bhat et al., 1992; Kipps, 1989). The Ig secreted by these so-called B1a cells tends to be polyreactive, with some autoreactivity. However, not all B1a cells are naive, indicating heterogeneity in this population (Schettino et al., 1997). There is a tendency for CD5-expressing B cells to show an increase in autoimmune disease such as rheumatoid arthritis (Youinou et al., 1993), and in a few cases they can make pathological autoantibodies (Mantovani et al., 1993). However, the role of these cells and their relationship to CD5+ve B cell tumors such as CLL and small lymphocytic lymphoma remain unclear. Analysis of normal CD5+ B cells has revealed that the vast majority carry unmutated V genes, as expected in naive B cells (Fischer et al., 1997). The remaining ∼40% of blood B cells carry mutated V genes and express CD27, both features of memory B cells (Klein et al., 1998b). Interestingly, the majority express IgM, some with coexpression of IgD, with the remainder having undergone isotype switch (Klein et al., 1997). Memory B cells are also found in secondary lymphoid organs, in which they locate to sites such as the marginal zone of the spleen (Liu et al., 1991), ready to deal with invading organisms. Plasma cells containing mutated V genes migrate to the bone marrow but can also be found in spleen, lymph nodes, and mucosa-associated lymphoid tissue. There is also evidence that IgM+ memory B cells can migrate to the bone marrow apparently in clonally related waves (Paramithiotis and Cooper, 1997). These cells have an activated phenotype and are able to differentiate further only under the influence of T cells. The possibility of local maturation and isotype switch occurring in the bone marrow has relevance to our understanding of the pathogenesis of multiple myeloma. Interference with the genome during combinatorial and mutational events is potentially dangerous. It is likely to contribute to the chromosome changes leading to B cell tumors, many of which involve translocations in chromosome 14 at the position of the recombined VH genes. There are several translocations characteristic of certain lymphomas, such as the t(14; 18) (q32;q21) of follicular lymphoma, which couples the bcl-2 protooncogene on chromosome 18 to the Ig heavy-chain joining region, JH (Yunis et al., 1987). The breakpoints on chromosome 14 are usually at the 5′ border of JH (Cleary et al., 1986a; Tsujimoto et al., 1985a,b). The consequence of translocation is to deregulate production of Bcl-2 protein, likely due to the proximity of the Ig enhancer (Graninger et al., 1987). Surprisingly, similar translocations have been detected in 13 of 24 normal lymph nodes and tonsils with follicular hyperplasia (Limpens et al., 1991). They have also been found in blood B cells in six of nine normal individuals (Limpens et al., 1995),
90
Freda K. Stevenson et al.
indicating that these translocations are quite frequent but not sufficient for lymphomagenesis. Translocations may also arise due to the somatic mutation process such as in endemic Burkitt’s lymphoma, in which the c-myc gene may be translocated into the V region or the J intronic region (Klein, 1989). Since there is no detectable association with recombination signal sequences at the breakpoints, and the V genes are somatically mutated, translocation could have arisen via the strand breaks occurring during this process (Kuppers et al., 1999).
III. IMMUNOGLOBULIN GENES IN B CELL TUMORS When B cell tumors develop, the V gene sequences of the cell of origin are preserved. It is technically relatively simple to amplify the sequence across the CDR3 region and to determine if the B cell population in a diagnostic biopsy is clonal. Interestingly, in some clinical situations, such as the early stages of posttransplantation lymphomas, oligoclonal B cell proliferations have been observed. These are largely Ebstein–Barr virus (EBV) induced and can be a prelude to development of a monoclonal lymphoma (Sklar et al., 1984). Analysis of V gene sequences can also be carried out at the singlecell level, and this strategy has revealed that the Reed-Sternberg cells in the vast majority of cases of Hodgkin’s disease are clonal B cell populations (Kuppers and Rajewsky, 1998). The unique CDR3 sequence of B cell tumors facilitates tracking of the tumor clone, and V gene sequence analysis allows insight into the point of differentiation reached by the cell of origin. Comparison of V gene sequences with the known germline gene sequences can quickly reveal whether the cell has undergone somatic mutation, indicative of antigen encounter. Further analysis of VHDJH-constant region transcripts can show if isotype switch events have taken place in the tumor cells. It also reveals changes which have occurred posttransformation. This wealth of information is leading to a biologically relevant extension of the classification of B cell tumors (Fig. 2) (Stevenson et al., 1998). Translocations involving chromosome 14 at the site of the VH gene locus are highly associated with certain tumors. It is likely that these occur at the stages at which either genetic recombination or somatic mutation occur. Those which are found in all cells of the clone are likely to have contributed to tumorgenesis. Examples include the t(8;14)(q24;q32) translocation, involving the myc gene, characteristic of Burkitt’s lymphoma; the t(14;18) (q32;q21) translocation, involving the bcl-2 gene, characteristic of follicular lymphoma (FL); and the t(11;14)(q13;q32), involving the cyclin D1 gene, characteristic of mantle cell lymphoma (Willis and Dyer, 2000). However, tumor cells harboring these translocations need not have features associated
The Occurrence and Significance of V Gene Mutations
91
with arrest at the stage of VHDJH recombination. For example, tumors such as FL which have translocations involving D–JH or VH–D rearrangements, expected to have occurred in the bone marrow, are not necessarily located at that site. In fact, analysis of CD34+/CD19+ bone marrow cells from patients with FL using sensitive polymerase chain reaction (PCR) failed to detect the tumor-related bcl-2–JH sequence, arguing against the concept that the disease arose in progenitor cells (Voso et al., 1997). It is possible that either the tumor cell can differentiate further before arresting in the germinal center or secondary recombinatorial events can occur in other sites (Nussenzweig, 1998). Many other chromosomal aberrations occur in B cell tumors, but few are found at an early stage of growth and in all cells of the clone. Genetic instability of tumors is common and can operate either at the nucleotide level, possibly due to faulty DNA repair, or at the chromosome level (Lengauer et al., 1998). Tumors of B cells often appear to involve both of these processes, and the effect on malignant behavior is complex and variable.
A. VH Gene Usage by B Cell Tumors Since the repertoire of VH genes used by normal B cells is known, it is possible to detect bias in VH gene usage among B cell tumors. Bias is likely to reflect a proliferative drive on the B cell of origin through binding of a putative antigen to a FWR of the VH gene. Most of the amino acid sequences of FWR1 and FWR3 of the variable region of Ig are exposed to solvent and therefore have the potential to interact directly with antigen outside the conventional CDR sequences (Kirkham and Schroeder, 1994). Antigens able to bind via FWRs could act as B cell superantigens, stimulating large numbers of B cells expressing defined V gene segments (Goodglick and Braun, 1994). The most studied B cell superantigen is staphylococcal protein A, which binds to the majority of Igs with heavy chains derived from the VH3 family (Silverman, 1997). Other superantigens include gp120 of HIV-1, pFv (a gut-associated sialoprotein), and protein L from Peptostreptococcus magnus (Silverman, 1997). Clearly, the expansion of B cells resulting from stimulation with a superantigen carries a risk of neoplastic transformation, and the resulting tumors would show a bias in V gene usage. A dramatic example of bias in VH gene usage is seen in the cold agglutinins, in which monoclonal IgM paraproteins with specificity for the I/i carbohydrate antigen of red blood cells may be produced (Roelcke, 1974). All these IgMs have heavy chains encoded by the V4-34 gene (Pascual et al., 1992; Silberstein et al., 1991), and it has been shown that interaction with the red cell antigen is mediated via a sequence in the first FWR (Li et al., 1996). Interestingly, the serum Ig of patients following infection with EBV, cytomegalovirus, or Mycoplasma pneumoniae shows high levels of
92
Freda K. Stevenson et al.
V4-34-encoded Ig, and patients can occasionally develop cold agglutination (Chapman et al., 1993). This indicates a superantigenic binding to V4-34expressing B cells by these organisms, paralleled by binding of the red cell antigen, possibly via molecular mimicry. The same V4-34 gene is overrepresented in primary central nervous system lymphomas (Montesinos-Rongen et al., 1999; Thompsett et al., 1999), raising the possibility of involvement of a pathogen. Bias in VH gene usage has also been observed in CLL, with an increased representation of the V1-69 (51p1) gene, often together with JH6 (Kipps, 1989). Intriguingly, overrepresentation of this gene appears to be largely confined to one of the newly defined subsets of CLL (Damle et al., 1999; Hamblin et al., 1999). Bias can provide clues to the pathogenesis of tumors; however, with the notable exception of the tight association of V4-34 with cold agglutinin disease, it is rarely absolute. Selective involvement of VH (or VL) in tumors may be shown more convincingly once disease subsets are more clearly defined. Until then, many cases usually need to be analyzed to assess genuine bias.
IV. SOMATIC MUTATION IN B CELL TUMORS One of the critical steps in B cell differentiation is the initiation of somatic mutation which follows encounter with antigen. V gene analysis has shown that, within current tumor classifications, differences in mutational status can occur, suggestive of previously unsuspected heterogeneity. A new map of the clonal history of B cell tumors is emerging which takes into account the mutational features of the cell of origin and the influence of the somatic mutational mechanism posttransformation. Figure 3 illustrates the relationship of a range of B cell tumors to the site of somatic mutation, generally assumed to be a germinal center-like environment. Recent data have revealed that CLL comprises two distinct subsets, one of which has unmutated V genes and the other has mutated V genes. The two subsets evidently have arisen at different points of differentiation and, interestingly, display wide differences in clinical behavior (Damle et al., 1999; Hamblin et al., 1999). In this case, heterogeneity confirmed clinical suspicions and will provide objective criteria to help in tailoring treatment. It is also possible that mantle cell lymphoma, although mainly derived from naive B cells with unmutated V genes (Hummel et al., 1994), may include a subcategory with mutated genes (Nakamura et al., 1999). The majority of B cell tumors have mutated V genes, and many have apparently accumulated additional mutations posttransformation. This may be important since it indicates that tumor cells are susceptible to environmental influences, and it may open the possibility of therapeutic manipulation. It is a feature of tumors sited in the germinal center, and the consequent intraclonal sequence variation is found in FL (Bahler et al., 1991; Cleary et al.,
The Occurrence and Significance of V Gene Mutations
93
Fig. 3 The relationship of B cell tumors to the site of somatic mutation. Mutational patterns in V genes allow assignment of differentiation status and subdivision of known categories. Straight arrows indicate that the cell of origin has not accumulated mutations, whereas arrows entering and leaving indicate a stable mutational pattern with no intraclonal variation. Chronic lymphocytic leukemia and mantle cell lymphoma include subsets of both these categories. Tumors within the site continue to accumulate mutations and in some cases such as diffuse large cell lymphoma, show both stable and ongoing mutations. Benign gammopathies and hairy cell leukemia have limited intraclonal variation, and most cases of lymphoplasmacytoid lymphoma and all cases of multiple myeloma have stable mutational patterns.
1986b; Zhu et al., 1994), diffuse large B cell lymphoma (Kuppers et al., 1997; Ottensmeier et al., 1998), Burkitt’s lymphoma (Chapman et al., 1995; Klein et al., 1995; Tamaru et al., 1995), AIDS-associated Burkitt’s-like lymphoma (Ng and McGrath, 1998; Riboldi et al., 1994), and mucosa-associated lymphoid tissue (MALT) lymphomas (Bahler et al., 1997; Du et al., 1996). Surprisingly, not all ongoing mutation in B cell tumors relies on location in a germinal center site. Primary central nervous lymphomas, for example, which show no evidence of association with a germinal center, have very high levels of somatic mutation, with intraclonal heterogeneity (Thompsett et al., 1999). Even normal B cells can undergo affinity maturation in the absence of germinal centers, as shown in lymphotoxin-α-deficient mice (Matsumoto et al., 1996). However, high doses of antigen (Matsumoto et al., 1996) or persistent antigen (Wang et al., 2000) are required to overcome the lack of the germinal center environment. In normal human B cells, induction of somatic mutation appears to be dependent on ligation of surface Ig (Liu et al., 1997; Razanajaona et al., 1997). If this is the case for tumors, it could indicate a role for persisting
94
Freda K. Stevenson et al.
antigen in stimulating tumor cell growth. A prominent example is in gastric MALT lymphomas, in which infection with Helicobacter pylori plays a clear role in pathogenesis (Isaacson, 1999). The majority of MALT lymphomas have somatically mutated V genes, often with ongoing mutation, consistent with derivation from germinal center B cells (Qin et al., 1997). Interestingly, MALT lymphoma of the salivary gland has similar features but also shows bias both in VH gene usage and in the amino acid composition of the CDR3, strongly suggesting a role for superantigenic drive (Miklos et al., 2000). However, it is unclear whether B cell tumor growth requires continued stimulation by specific antigen, and chromosomal events are likely to liberate some cells from this requirement (Bemark and Neuberger, 2000). A common characteristic of B cell tumors is an ability to avoid apoptosis, and tumor cells surviving at different times of differentiation are providing insight into features of normal B cells reaching those stages, in which apoptosis is the usual default pathway. Interestingly, analysis of V genes in the Reed–Sternberg cells of classical Hodgkin’s disease (HD), now revealed to be largely a B cell-derived disease, has shown that the V genes are somatically mutated but often “crippled” and unable to encode functional Ig (Hummel et al., 1996; Kanzler et al., 1996). HD is therefore an exception to the observation that most B cell tumors express Ig, despite frequent loss of one allele due to translocations involving chromosome 14. Retention of Ig may reflect the fact that normal B cell development is dependent on Ig expression (Rajewsky, 1996). Clearly, cells of classical HD, with crippled V genes, have avoided the death pathway possibly due to the presence of either EBV or mutations in IκBα (Jungnickel et al., 2000) and/or CD95 (Muschen et al., 2000). Many tumors arise from B cells which have undergone somatic mutation but which do not accumulate further mutations posttransformation. In these cases, tumor cells have apparently exited from the site of somatic mutation and generally do not reactivate this process. Ongoing mutations are uncommon in lymphoplasmacytoid lymphomas (Sahota et al., 1998), including Waldenstrom’s macroglobulinemia (Wagner et al., 1994) and splenic lymphoma with villous lymphocytes (Zhu et al., 1995). Marginal zone lymphoma is a heterogeneous entity even in the current classifications, and the pattern of somatic mutation reflects this categorization, with evidence for unmutated sequences and mutated sequences with or without ongoing mutation (Tierens et al., 1998). The picture is clearer for multiple myeloma, with the vast majority of these plasma cell tumors showing a high level of somatic mutation and no ongoing mutational activity—features consistent with a postfollicular tumor (Bakkus et al., 1992; Sahota et al., 1994; Vescio et al., 1995). However, benign plasma cell tumors, termed monoclonal gammopathies of undetermined significance (MGUS), show evidence of residual ongoing mutation, at least in some cases (Sahota et al., 1996). A similar low level of intraclonal variation is seen in hairy cell leukemia
The Occurrence and Significance of V Gene Mutations
95
(Maloum et al., 1998; Forconi et al., 2001), and both these tumor categories are at the border of the site of somatic mutation (Fig. 3). The status of V genes is one clear biological marker of disease subsets. This review focuses on how emerging knowledge of V gene status is influencing our understanding of the pathogenesis of B cell tumors. We concentrate on examples of disease categories in which such knowledge is relatively advanced, focusing on CLL, FL, diffuse large B cell lymphoma, and multiple myeloma, each of which has been derived from B cells at distinct stages of maturation.
V. CHRONIC LYMPHOCYTIC LEUKEMIA CLL is an example of a B cell tumor for which knowledge of V genes is having a significant impact both at the immediate clinical level and on our understanding of the pathogenesis of the disease. Strikingly, the relatively simple analysis of V gene somatic mutational status has allowed CLL to emerge from diagnostic confusion. Two subsets of the disease exist, with either unmutated or mutated V genes, each with a very different prognosis (Damle et al., 1999; Hamblin et al., 1999). Assignment to either subset can be done using objective V gene-based criteria, and this should inform clinical management. Until recently, CLL was considered a single disease entity characterized by the relentless accumulation in the blood and bone marrow of monoclonal B cells with the appearance of small mature lymphocytes. All cases of CLL have common features which distinguish this tumor from many similar B cell tumors. Typically, the cells are positive for CD5, CD23, and CD19 and negative for surface CD22 and FMC7 (Matutes et al., 1994). Surface Ig (usually IgM + IgD) is sparse and the immunoglobulin-associated molecule CD79b is low or absent (Zomas et al., 1996). Most cells are in the G0 phase of the cell cycle and may be unresponsive to mitogenic stimuli (Andreef et al., 1980). The cells overexpress the bcl-2 gene product and are resistant to apoptosis (Pezella et al., 1990).
A. V Gene Mutational Status Because CD5+ B cells are found in the fetal spleen (Antin et al., 1986) and surface IgD is a feature of cells that have not yet met antigen in the germinal center (Nicholson et al., 1995), it has been suggested that CLL is a tumor of naive B cells, possibly arising in the follicular mantle zone (CaligarisCappio, 1996). Such naive cells are expected to lack somatic mutations in the Ig variable domain genes since this process occurs in the germinal center
96
Freda K. Stevenson et al.
environment (Berek and Milstein, 1987). Early sequences of IgV genes of tumor cells from patients with CLLs were in a germline configuration (Deane and Norton, 1991; Ebeling et al., 1992; Kipps et al., 1989), thus tending to confirm this concept. However, reports began to appear in the literature detailing cases with evidence of somatic mutation, culminating in 1994 with a review of the literature by Schroeder and Dighiero which found that 36 of 75 reported cases had IgV genes with less than 98% sequence homology to the appropriate germline gene. This percentage was chosen because polymorphisms, which are quite common in IgV genes, might account for this degree of disparity (Matsuda et al., 1993). Schroeder and Dighiero suspected that CLL might be a heterogeneous disorder but were unable to cull from the literature the comprehensive clinical detail needed to establish this hypothesis. It is clear, however, that some of the cases with mutated V genes were CD5−, some had high levels of monoclonal immunoglobulins in their sera, and some expressed surface IgG rather than IgM on their cells. None of these are features of classical CLL. Unfortunately, the CLL literature includes cases of mantle cell lymphoma and splenic marginal zone lymphoma, which because of their more florid nature often draw attention more readily than cases of true CLL. Because of the risk of confusion between true CLL and other low-grade lymphomas with a leukemic phase, we took great pains to study only cases scoring 4 or 5 on the Royal Marsden Score (Matutes et al., 1994). We produced convincing evidence that a subset with somatically mutated V genes exists (Hamblin et al., 1999). In 38 cases, at least five separate clones were analyzed. No intraclonal heterogeneity was found. This is in contrast to the finding in FL and implies that cells that showed somatic mutations were no longer under the influence of mutational mechanism. The mutational pattern was stable. In 3 cases of CLL from this series, a second blood sample was analyzed 5 years, 3 years, and 18 months, respectively, after the first. In each case, the clonal sequence was identical with the original sequence (with 0, 2, and 19 mutations, respectively).
B. Prognostic Value of V Gene Status Beyond establishing the existence of two subtypes of CLL arising at different stages of lymphocyte maturation, we were also able to demonstrate that the two subtypes behaved differently clinically. In 1997, our group examined the V genes of 22 patients with classical B cell CLL segregated according to karyotype. In general, tumors with trisomy 12 had unmutated IgV genes, but those with 13q14 abnormalities detected by conventional cytogenetics had evidence of somatic mutations (Oscier et al., 1997). Since it had been previously shown that CLL patients with trisomy 12 have a poorer survival
The Occurrence and Significance of V Gene Mutations
97
rate than those with 13q14 abnormalities (Juliusson et al., 1990), this pointed to an association between clinical status and degree of somatic mutation. In extending this series to 84 patients, we were able to show a survival difference between the two subtypes (Hamblin et al., 1999). Binet stage A patients (Binet et al., 1997), whose tumor cells lacked somatic mutations in their V genes, had a median survival of 8 years compared with 25 years for patients whose cells had somatic mutations. Our current analysis of 156 patients remains consistent with this finding (Fig. 4) (Oscier et al., 2000). Contemporaneously with our work, a group in New York studied 64 patients with surface IgM+, CD5+ CLL. They also found two subtypes of approximately equal numbers of mutated and unmutated V genes (Fais et al., 1998). Although no clinical details were available in the original paper, the authors subsequently published survival curves that were very similar to ours (Damle et al., 1999). There are clear differences between the two subsets. The gender ratio was close to unity for patients with evidence of somatic mutations, whereas there was a threefold male preponderance among patients with unmutated V genes. This latter subtype had characteristics associated with a more malignant type of disease. Those lacking mutations were significantly more likely to have advanced stage disease (p = 0.0009), progressive disease (p < 0.0001), and atypical morphology (p < 0.0001). Trisomy 12 as an isolated karyotypic abnormality was significantly associated with a lack of somatic mutations (p = 0.0019), and deletions or translocations at 13q14 were significantly associated with their presence (p = 0.023).
Fig. 4 Prognostic value of VH gene mutational status in chronic lymphocytic leukemia. Survival curves for 156 patients with CLL at all stages are shown in relation to the presence (<98% homology with germline sequence) or absence (>98% homology) of somatic mutations in VH genes.
98
Freda K. Stevenson et al.
Damle et al. (1999) suggested that the two groups might be distinguished simply by measuring the expression of CD38 on the surface of the CLL cells. In our hands, CD38 expression is certainly a marker of poor prognosis in CLL, but it is discordant with V gene mutational status in 30% of cases (Oscier et al., 2000). Moreover, CD38 expression varies over time in 25% of patients, sometimes reflecting changes in the speed of progression of the disease or responses to treatment. We find V gene mutations and CD38 expression to be independent prognostic variables. Patients whose cells show no mutations and do not express CD38 live approximately twice as long as those whose cells show no mutations and do express CD38. The discovery of the two subtypes of CLL and the possibility of diagnosing them at an early stage have important clinical significance. Current best practice delays treatment in early stage disease until progression occurs. Fewer than half will ever need treatment. Were it possible to know which patients would progress, then early treatment would be possible, at a time of low tumor bulk, before the acquisition of subsequent genetic damage, when cure might be a reasonable objective. Furthermore, some slowly progressive tumors might escape treatment altogether were it known that they had mutated VH genes. Microarray technology is extending our knowledge of CLL. There is evidence for a “CLL signature” which distinguishes gene expression profiles in CLL from those in other B cell tumors (Alizadeh et al., 2000). There is also emerging data on the genes that are differentially expressed between the two defined subsets that might provide clues to explain the malignant behavior of the unmutated subset.
C. VH Gene Usage Prior to the description of the two subsets of CLL, it had been observed that there was bias in usage of VH genes, with as many as 20% of cases being derived from the V1-69 gene (Kipps et al., 1989). Subsequent studies (Hamblin et al., 1999; Oscier et al., 1997; Fais, 1998) confirmed overrepresentation of V1-69, but at a lower level of 10–12%. The profile of VH gene usage by our cases of CLL compared to the normal repertoire is shown in Fig. 5. Interestingly, all studies found that the majority of cases using the V1-69 gene were of germline sequence. The reason for the disparity in the level of usage compared with that of earlier data likely reflects the fact that the later studies were of patients presenting with CLL rather than tertiary referrals. Presenting patients include both good and poor prognoses, whereas tertiary referrals include more of the poorer prognosis cases, which are generally unmutated. Intriguingly, this unmutated subset generally contains those derived from V1-69 (Damle et al., 1999; Hamblin et al., 1999).
The Occurrence and Significance of V Gene Mutations
99
Fig. 5 Bias in VH gene usage in chronic lymphocytic leukemia. The VH gene usage in a cohort of 84 patients with CLL was analyzed and gene usage compared with that of the normal adult population. The V1-69 gene showed clear overrepresentation, and this is largely in the unmutated subset. The V4-34 gene was also overrepresented, although the normal incidence has generally been found to be higher than that reported by Brezinchek et al. (1995), being 5–10% in normal blood. This gene is largely confined to the mutated subset.
Among the CLLs, V3-23 is the most commonly used VH gene in most studies, as expected from its high usage in the normal B cell repertoire, possibly due to the presence of multiple copies of the gene. The other gene which has a notable distribution in CLL is the V4-34 gene, which is used by 11% of our patients (Fig. 5) (Hamblin et al., 1999). There is some dispute as to how commonly V4-34 is used by normal B cells, with most studies (including our own) finding between 5 and 10% positivity (Stevenson et al., 1989). Overrepresentation in the total cases is therefore not very striking, but the interesting feature is that the V4-34 gene is virtually confined to the subset with mutated VH genes (Damle et al., 1999; Hamblin et al., 1999). Therefore, it appears that the V1-69 gene and the V4-34 gene are used by separate subsets of CLL, perhaps suggesting a different superantigenic drive on each cohort. One reason for the increased overall representation of the V1-69 and V4-34 genes may lie in an intriguing observation that these genes are more commonly used in elderly individuals (Wang and Stollar, 1999). It may well be that the usage of VH genes simply follows their normal usage for the age group. However, there is no doubt that their usage is uneven between the two subsets of CLLs. An explanation for this divergence might well provide a clue to their different origins. Figure 6 summarizes our current perception of the origin and nature of the two subsets of CLLs.
100
Freda K. Stevenson et al.
Fig. 6 The origin and nature of the cell of origin in the two subsets of chronic lymphocytic leukemia (CLL). Analysis of VH gene mutational status in CLL indicates that one subset (CLL subset 1) arises from a naive B cell, with unmutated VH genes, a bias to usage of the V1-69 gene, and a poor prognosis. A second subset (CLL subset 2) derives from a cell which has undergone somatic mutation with no ongoing mutation, a bias to usage of the V4-34 gene, and a good prognosis.
VI. FOLLICULAR LYMPHOMA FL represents the largest component of non-Hodgkin’s lymphoma and is clearly different from CLL in location, morphology, and immunophenotype. The architecture of a lymph node infiltrated with FL tends to retain some follicular structure reminiscent of a reactive lymph node (Fig. 7), and most patients have disseminated disease at diagnosis (Portlock, 1990). Histologically, it is a low-grade tumor, but it is a frustrating disease to manage, because it is essentially incurable, with a median survival of 7–10 years. Modern chemotherapeutic schedules have had little impact on survival, although patients often achieve clinical remission. Partially for this reason, FL has always been an attractive candidate for immunotherapy, and treatment with monoclonal anti-CD20 antibody is showing some encouraging results (Maloney et al., 1997). Vaccination against tumor antigens is also being tested, with the idiotypic determinants of surface Ig being a candidate target (George and Stevenson, 1989; Hsu et al., 1997). The tumor cells of FL tend to locate in lymph nodes, spleen, and bone marrow and have a mixed morphology of centrocytes and centroblasts, reflecting
The Occurrence and Significance of V Gene Mutations
101
Fig. 7 Histological features of cases of follicle center lymphoma (FCL) and primary diffuse large B cell lymphoma (DLBCL). The low-power view of FCL (a) shows some preservation of follicular structure, and the high-power view (b) indicates the cellular features of the predominant centrocytic and accompanying centroblastic tumor cell populations. (c) DLBCL consists of sheets of mainly centroblasts or immunoblasts with large nuclei, prominent nucleoli, and a high growth fraction.
their origin from germinal center B cells (Fig. 7). The majority of cases of FL carry the characteristic t(14;18) chromosomal translocation at the site of the bcl-2 gene, usually involving the JH sequence of the Ig gene. This translocation fuses the bcl-2 gene at 18q21 to the Ig heavy-chain locus at 14q32, leading to deregulated expression of bcl-2 and production of increased levels of the antiapoptotic Bcl-2 protein (Tsujimoto et al., 1985a,b,c; Yunis et al., 1987). In some cases, there is evidence of involvement of RAG-1/-2-mediated transposition events (Vaandrager et al., 2000). Deregulation of gene expression appears to be due to the proximity of the powerful Ig transcriptional enhancers, which apparently can act over extreme distances of DNA. Tumor cells of FL have undergone V(D)J recombination, and gene usage is similar to the normal B cell repertoire, with a possible small excess of genes from the VH4 family (Bahler et al., 1991; Stevenson et al., 1995). Most VH genes in FL are somatically mutated, consistent with an origin from a germinal center B cell, with a distribution of mutations consistent with antigen selection (Bahler and Levy, 1992). As expected, the nature of the mutational changes reflects that of normal B cells, with mainly single base substitutions, but there is also evidence for insertions and deletions
102
Freda K. Stevenson et al.
(Noppe et al., 1999). In the majority of cases there is clear evidence for mutational activity posttransformation (Zelenetz et al., 1992; Zhu et al., 1994), which leads to intraclonal sequence variation (Noppe et al., 1999). Perhaps this is not surprising since the environment of the germinal center includes follicular dendritic cells and activated T cells, considered to be required for continuing mutational activity. However, it is unclear whether antigen plays a role at this stage. Studies of neoplastic B cell lines in vitro have shown dependence of continuing mutational activity on engagement of the B cell receptor in some (Denepoux et al., 1997; Sale and Neuberger, 1998; Wu and Kaartinen 1995; Zan et al., 1999) but not all cases (Sale and Neuberger, 1998; Wu and Kaartinen, 1995). For tumors in vivo, it is not known how long this mutational activity persists within the clone. Interestingly, somatic mutations can also be found in the bcl-2 gene, and it has been suggested that these may have been introduced by the mutational mechanism normally targeted to the V genes (Tanaka et al., 1992). A similar accumulation of somatic mutations is apparent in other protooncogenes translocated to 14q32, such as the c-myc gene derived from chromosome 8 (Klein, 1989; Nowell and Croce, 1988; Zajac-Kaye et al., 1988), characteristic of Burkitt’s lymphoma. A significant number of cases of FL (40–80% over a period of 8 years from diagnosis) will transform to a more aggressive form, with or without chemotherapy. Typically, FL transforms to a diffuse large B cell lymphoma, and it is accompanied by a diverse range of chromosomal and genetic changes; no defining abnormalities have been identified (Symmans et al., 1995). After treatment, a single VH sequence can emerge, with loss of the previous intraclonal heterogeneity. This narrowing of heterogeneity strongly suggests that a single cell has escaped from control, and that further somatic mutation does not occur at this stage of disease (Zelenetz et al., 1992; Zhu et al., 1994). In addition to intraclonal sequence variation, there is evidence from analysis of VDJ-constant region transcripts for isotype switch events occurring within the tumor clone (Ottensmeier et al., 1998). In some cases, there is also indication from DNA fiber fluorescence of complex rearrangements of the constant region genes downstream of Cµ–Cδ, which might reflect tumorspecific deregulation of the class-switch machinery (Vaandrager et al., 1998).
VII. DIFFUSE LARGE B CELL LYMPHOMA Diffuse large B cell lymphoma (DLBCL) is a term used to describe lymphomas which destroy the structure of the invaded lymphoid organ in which they arise, replacing it with sheets of centroblastic or immunoblastic tumor cells (Fig. 7). DLBCL can also develop at extranodal sites, such as the liver,
The Occurrence and Significance of V Gene Mutations
103
thyroid, bone, and gastrointestinal tract. The central nervous system or testis may also be involved or may be the primary sites of disease. Tumor cells have large nuclei with prominent nucleoli, and they have a moderately high proliferation fraction (Chan et al., 1995). Although pathologists have recognized variants based on morphologic and immunophenotypic criteria, there has been no agreement on a subclassification of DLBCL. Primary DLBCL accounts for ∼40% of all B cell malignancies and differs from transformed FL in presentation and clinical course. Response of DLBCL to combination chemotherapy is good, with an overall cure rate of ∼35%. A substantial proportion of cases (30–40%) carry the t(3;14)(q27;q32) chromosomal translocation, involving the 5′ regulatory region of the bcl-6 gene and the Ig heavy-chain switch region (Ye et al., 1995a,b). Translocations can lead to deregulation of the bcl-6 gene with increased expression of Bcl-6 protein, which appears multifunctional since it is able to act as a transcriptional repressor and to contribute to nuclear organization, replication, and chromatin-mediated regulation (Albagli et al., 2000). The protein is highly expressed in normal germinal center B cells, and it appears to be essential for formation of germinal centers (Fukuda et al., 1997). Point mutations in the region containing the first exon and intron are also common in B cell tumors, but since they are also evident in normal germinal center B cells (Shen et al., 1998), it is unclear if they play a role in lymphomagenesis. There is intriguing evidence that transformation of FL to DLBCL may be accompanied by additional mutations in the noncoding region of the bcl-6 gene (Lossos and Levy, 2000). The mechanism for introducing somatic mutations into this site of the bcl-6 gene is unknown. It is tempting to link it with mutation of the V genes, especially since mutational activity in bcl-6 tends to increase in tumors undergoing somatic mutations in V genes (Migliazza et al., 1995) and there are similarities in the nucleotide substitution patterns. However, the correlation between levels of mutation in the two target genes is not strong, and the finding of intraclonal heterogeneity in bcl-6 in tumor cells with no such heterogeneity in VH indicates that the two mechanisms are not temporally linked (Rothwell et al., 1999; Sahota et al., 2000). The majority of cases of DLBCL have undergone conventional V(D)J recombination and express surface Ig. Regarding FL, usage of VH genes appears to be similar to that of the normal repertoire, although one study (not confirmed in larger surveys) appeared to indicate an increased usage of the V4-34 gene (Hsu and Levy, 1995). Bias to this gene is more certain for the subset comprising the primary central nervous system lymphomas, but more cases are needed (Montesinos-Rongen et al., 1999; Thompsett et al., 1999). Most subtypes of DLBCL show somatic mutations (Hsu and Levy, 1995; Kume et al., 1997; Kuppers et al., 1997; Kuze et al., 1998; Ottensmeier et al., 1998; Taniguchi et al., 1998). Many also show evidence for ongoing mutational activity, with consequent intraclonal heterogeneity which can be narrowed by intensive treatment (Zelenetz et al., 1992). Interestingly, isotype
104
Freda K. Stevenson et al.
switch variants can also be identified with a common CDR3, but with differing patterns of mutation indicative of subpopulations undergoing switch to different isotypes within the clone (Ottensmeier and Stevenson, 2000). Recently, gene expression profiling using microarray technology has been applied to DLBCL (Alizadeh et al., 2000). In a striking demonstration of the power of this approach, two major subsets were clearly identified. The first subset expressed genes associated with those normally expressed in germinal center B cells, and the second had a profile similar to that of activated B cells (Alizadeh et al., 2000). The clinical value of the technology was demonstrated by the finding that patients with the germinal center-like profile had a significantly better overall survival rate than those with activated B celllike profile. In an interesting extension to the study which related to V gene status, seven of seven cases defined as germinal center-like showed evidence for ongoing mutations, whereas five of seven cases with an activated B cell profile had few or no ongoing mutations (Lossos et al., 2000a). This again demonstrates that V gene status has significance for prognosis, and that combining it with microarray technology should be fruitful.
VIII. PLASMA CELL TUMORS Multiple myeloma (MM) is a clinically aggressive tumor characterized by the accumulation of malignant plasma cells in the bone marrow compartment (Durie and Salmon, 1985). Clinical features include osteolytic lesions, anemia, and renal impairment; disease outcome remains poor despite modern chemotherapy. However, recent approaches using high-dose chemotherapy combined with autologous or allogeneic transplantation have provided encouraging results (Cunningham et al., 1994; Lokhorst et al., 2000), and new drugs such as thalidomide (Singhal et al., 1999) are showing early promise. The benign counterpart of MM, MGUS, is much less severe, but there is difficulty at presentation in distinguishing it from early stage MM (Kyle and Lust, 1990). The relationship between the two diseases is also unclear, although ∼15% of patient with MGUS have been found to develop MM over a period of 9.6 years (Kyle and Lust, 1990). Of particular importance for the treatment of MM is whether there exists a “precursor” less mature cell within the clone which feeds the plasma cell compartment. It is here that V gene analysis has provided useful clarification and has established MM as the benchmark postfollicular tumor. MM is mainly derived from a B cell which has undergone isotype switch, with >95% being IgG or IgA, 1.5% secreting IgM, and 0.2% secreting IgD (MacLennan, 1992). Interestingly, the clinical outcome in the different switch variants is essentially the same. In fact, the cell of origin in the rare
The Occurrence and Significance of V Gene Mutations
105
IgM-secreting MM has somatically mutated V genes indistinguishable from those of the more common isotype-switched tumor cells, suggesting similar derivation and behavior. It has been more difficult to delineate the origin of the very rare IgD myeloma, which, with its more aggressive course, has been considered clinically as a variant of Bence–Jones MM (Blade and Kyle, 1999). One interesting suggestion is that it may be derived from an IgM−ve, IgD+ve germinal center cell isolated from tonsil (Arpin et al., 1998). The features of this cell are similar to those of IgD MM in having a Cµ–Cδ isotype switch, a strong association with λ light chains, and a high level of somatic mutation (Arpin et al., 1998). A further emerging feature of MM cells is the frequent involvement of aberrant translocations of chromosome 14 at the Ig locus, indicating that Ig rearrangements and somatic mutations may contribute to tumorgenesis (Hallek et al., 1998). The bulk of tumors have undergone isotype switch, and aberrant involvement of switch events appears to contribute to neoplastic behavior in a significant proportion of cases (Bergsagel et al., 1996).
A. V Gene Mutational Status In MM V genes are invariably somatically mutated, consistent with an origin from a B cell which has encountered the germinal center site (Bakkus et al., 1992; Ralph et al., 1993; Sahota et al., 1994; Vescio et al., 1995). For IgM-secreting MM, this excludes an origin from the pathway generating primary, unmutated plasma cells (MacLennan, 1992). The level of mutation in MM cells appears to be higher in VH than in VL, also a feature of V genes in normal B cells (Sahota et al., 1997). Doublet mutations have been observed, but both deletions and insertion events appear to be rare in functional V genes. When the previously accepted criterion for antigen selection as clustering of replacement amino acids in CDRs was applied, ∼40% of VH genes in MM appeared selected (Vescio et al., 1995). When both VH and VL were analyzed together, ∼70% of cases in our study showed significant clustering (Sahota et al., 1997). Using the current criterion of preservation of the FWR sequence (Dorner et al., 1997), VH genes in MM show evidence for antigen selection in 38 of 67 (57%) cases (Sahota et al., 1994, 1997; Vescio et al., 1995). It is likely that the cell of origin in MM has been influenced by antigen selection, although antigen is less likely to play a continuing role in stimulating growth of the surface Ig-negative plasma cell. The nature of the inducing antigens is unknown, and lack of significant bias in VH gene use argues against a role for superantigenic drive in MM. However, several myeloma proteins display autoantibody activity (Dighiero et al., 1983). A common feature of VH and VL sequences in MM is the complete lack of intraclonal variation (Bakkus et al., 1992; Kiyoi et al., 1998; Kosmas et al.,
106
Freda K. Stevenson et al.
1996; Ralph et al., 1993; Sahota et al., 1994, 1997; Vescio et al., 1995), indicating that the cell of origin in MM, considered to be a memory B cell, undergoes neoplastic arrest at a stage in which the mutation mechanism is silent. Stability of sequence is maintained from presentation to plateau phase of disease (Ralph et al., 1993). Shared stable V gene sequences within tumor-derived clones provide an important molecular backdrop to assess subclonal genetic events. There is evidence for chromosomal translocations (Avet-Loiseau et al., 1999), with oncogene mutations and bcl-6 mutations (Rothwell et al., 1999) being present in subpopulations, indicating heterogeneity of disease which may influence tumor outgrowth. However, the V gene sequence homogeneity characteristic of myeloma cells may not be a feature of benign plasma cell tumors. In cases of MGUS, intraclonal heterogeneity in tumor VH sequence was observed in three of seven patients (Sahota et al., 1996). Heterogeneity was subsequently confirmed in a further minority of cases (S. S. Sahota, unpublished data) and detected in a nonpassaged case of benign monoclonal gammopathy in the mouse (Zhu et al., 1998). Therefore, MGUS appears to arise from a less mature B cell than does MM. It may indicate a continuing influence of the somatic mutation mechanism on cells of MGUS (Sahota et al., 1996) in a manner generally associated with tumors of the germinal center (Stevenson et al., 1998). For MM, it has not been clear at which stage of B cell differentiation neoplastic transformation occurs. The existence of less mature cells was suggested from early phenotypic analysis, but the issue became more accessible when V gene analysis became available. The fact that MM cells had somatically mutated V gene sequences, with no ongoing mutational activity, indicated that the final neoplastic event had occurred at a postfollicular stage (Bakkus et al., 1992; Sahota et al., 1994; Vescio et al., 1995). However, the detection of transcripts containing a tumor-derived CDR3 sequence linked to a Cµ sequence, in some cases of isotype-switched myeloma, raised the possibility of coexistence of a preswitched IgM+ve B cell (Billadeau et al., 1993; Corradini et al., 1993). The few available sequences appeared to indicate that the somatic mutational pattern was identical to the postswitched cell, suggesting that the precursor cell was arrested just prior to isotype switch. However, it has not been possible to isolate these cells, and the frequency appears low, suggesting that these cells may not be important in feeding the plasma cell compartment (Bakkus et al., 1994; Berenson et al., 1995; S. S. Sahota, unpublished observations). The question of the contribution of circulating B cells to the tumor cell clone has been controversial. With regard to a precursor cell, it has been difficult to identify tumor-related CDR3–Cµ transcripts in circulating B cells. In one report, although these were detected in bone marrow in five of five cases, they could not be identified in blood in any patient (Corradini et al., 1993). Subsequently, tumor-derived Cµ transcripts could be detected in CD19+ve B cells in only one of five MM cases (Bakkus et al., 1994).
The Occurrence and Significance of V Gene Mutations
107
Investigations of tumor-related CD19+B cells in peripheral blood have also been sought using PCR for VH–JH sequence, independent of constant region, but results have been conflicting (Chen and Epstein, 1996; Kay et al., 1997; Szczepek et al., 1998). The tumorigenic potential of such cells is not clear, although it has been suggested that in NOD SCID mice, grafted PBMNCs from myeloma patients are able to home to the marrow and survive (Pilarski et al., 2000). One clue to the nature of the IgM+ve precursor cells might be derived from the study of IgM-secreting MM. In our study, the VH gene pattern in IgM+ve MM was very similar to that of the putative precursor cell. In six of six cases, we detected somatically mutated VH genes with no intraclonal variation (Sahota et al., 1999), and it is possible that this rare tumor is derived from the preswitched IgM+ve cell occasionally detected in typical MM. Interestingly, we found coexisting tumor-derived VHDJH–Cγ transcripts in two of four cases, suggesting that these cells were starting to undergo isotype switch (Sahota et al., 1999). Neoplastic transformation in MM may therefore occur during isotype switch events in an IgM+ memory cell, with arrest possible on either side of the switch point. Somatic mutation is silent at this stage. It is possible that a precursor IgM+ memory B cell homes to the bone marrow and is able to isotype switch in situ (Paramithiotis and Cooper, 1997). A second explanation for the detection of CDR3–Cµ transcripts in MM is that some cells have not undergone conventional deletional switch recombination. There is evidence from the mouse that the mechanism for expressing multiple isotypes in a subpopulation of B cells involves RNA processing (Shimizu et al., 1991). It also appears that this mechanism may operate as a prelude to conventional isotype switching (Fujieda et al., 1996 a,b). We have found that multiple isotypes expressed in hairy cell leukemia may derive from a similar mechanism (Forconi et al., 2001), and it will be of interest to determine if this is the case in MM.
B. V Gene Usage Currently, there are ∼200 MM V gene sequences available for analysis that have consistent features. Although VH and VL gene use is comparable with the expressed normal B cell repertoires at the level of the V gene families, discordance occurs at the level of individual V genes. The most striking feature is the exclusion of the V4-34 gene in MM (Rettig et al., 1996), whereas there is a frequency of 5–10% in normal B cells (Brezinschek et al., 1997; Kraj et al., 1995; Stevenson et al., 1989). The V4-34 gene is also used by normal isotype-switched cells both in tonsil (Chapman et al., 1996) and in plasma cells of the gut (Dunn-Walters et al., 2000). However, only a single, unusual case of IgD-secreting MM has been reported to use V4-34
108
Freda K. Stevenson et al.
(Kiyoi et al., 1998). As discussed previously, this gene is mandatory for encoding heavy chains used by IgM proteins with cold agglutinin anti-red cell activity. There is no obvious explanation for the lack of involvement of V4-34-encoded Ig in MM, but one could speculate that this Ig may bind to a molecule, possibly similar to the red cell antigen in being carbohydrate in nature, which may be required for tumor maintenance. There is no significant bias in usage of VL genes in MM, with the Vκl gene family members O8/O18, O2-12, and L11 commonly used, in line with usage by normal B cells (Kosmas et al., 1998; Kosmas et al., 1996; Sahota et al., 1997). Gene segments 2A2, 2C, and 3R from the Vλ repertoire are also used in ∼30% of cases, in accordance with expected normal incidence (Kiyoi et al., 1998; Kosmas et al., 1997; Sahota et al., 1997).
C. Ig Locus and Chromosomal Translocations It is becoming clear that the Ig locus is frequently involved in translocations in MM, with involvement of switch regions in 14q32 common in both cell lines and in primary myeloma cells (Hallek et al., 1998). The identification of downstream switch regions in the CH locus as being the major sites of promiscuous translocations on the nonfunctional allele in myeloma suggests a late event, occurring at the stage of a mature B cell undergoing isotype switch. Recurrent translocations at 14q32 involve chromosome 11 (11q13) at the site of the cyclin D1 gene and chromosome 4 (4p16) at the site of FGFR3, with each accounting for 20–25% of cases (Chesi et al., 1998a,b). These tend to be in all tumor cells of the clone and are likely to contribute to malignant behavior (Avet-Loiseau et al., 1999). Both of these have also been identified in MGUS, indicating a potential oncogenic role in MM and MGUS but suggesting that further events are required for full malignant behavior (Avet-Loiseau et al., 1999). One of these may involve chromosome 13 since monosomy 13 is associated with de novo MM and with transition from MGUS to MM (Avet-Loiseau et al., 1999). In addition, excision elements from isotype switch events have been identified as transposon elements, integrating adjacent to c-myc (Shou et al., 2000) and cyclin D1 (Gabrea et al., 1999) and potentially acting as transforming elements in MM. MM targeted by these transposons could possibly deregulate a variety of additional genes. Illegitimate IgH isotype switch events appear to have a widespread consequence in MM biology and have recently been implicated in the rare Ig heavy-chain loss variants. In 9 of 12 such cases, lack of IgH synthesis has been attributed to the deletion of the functional VHDJH allele, most likely a result of CH deletion events in the IgH locus (Szczepanski et al., 2000). The functional VL genes in the IgH− MM cases display somatic mutation, consistent
The Occurrence and Significance of V Gene Mutations
109
with the final oncogenic event being postgerminal center and occurring at the isotype switch stage. Clearly, there is great complexity consequent on rearrangements and perhaps mutations at the Ig locus, and those events critical for tumorgenesis and progression are gradually being delineated.
IX. CONCLUSION Molecular genetics is having an impact on our understanding of tumor development that is clearly visible for B cell tumors where genetic rearrangement and mutation of Ig genes are part of the normal developmental program. This genetic lability has a price in that translocations commonly involve the Ig heavy-chain locus at chromosome 14q32. Although such translocations may not be sufficient for tumorgenesis, some are likely to contribute to the escape of a B cell from normal control mechanisms. The emerging tumors carry their Ig genes, with the clonal history imprinted in the sequences. We can determine the point of differentiation reached by the cell of origin and the response of the transformed cell to environmental influences. A new classification will soon be available which subdivides current categories. However, rather than adding to the confusion, this biologically relevant information should help us to understand more about the pathogenesis of B cell tumors. Importantly, it will also have relevance for management of patients. However, Ig V genes are not just passive historical indicators. The functional genes encode idiotypic Ig protein which is clonally distinct from normal Ig. It can act as a target for immune attack either by exogenous antiidiotypic antibody or following vaccination with idiotypic protein (George and Stevenson, 1989; Hsu et al., 1997). The variable genes can also be placed into plasmid vectors to generate DNA vaccines capable of inducing immunity against lymphoma (King et al., 1998). Undoubtedly, microarray technology will provide additional information about genes which are activated in B cell tumors, and this can be linked to knowledge available from V gene analysis. New genes can also be placed into vaccines to improve attack on tumors. We are now in position to exploit genetic technology for a rational approach to therapy.
ACKNOWLEDGMENTS This work was supported by the Leukaemia Research Fund, Tenovus, the Cancer Research Campaign (UK), and the Multiple Myeloma Research Fund (USA). We thank Prof. Dennis Wright for providing the immunohistology of the lymphomas.
110
Freda K. Stevenson et al.
REFERENCES Agrawal, A., Eastman, Q. M., and Schatz, D. G. (1998). Nature 394, 744–751. Albagli, O., Lindon, C., Lantoine, D., Quief, S., Puvion, E., et al. (2000). Mol. Cell 20, 8560– 8570. Alizadeh, A. A., Eisen, M. B., Davis, R. E., Ma, C., Lossos, I. S., et al. (2000). Nature 403, 503–511. Alt, F. W., Blackwell, T. K., and Yancopoulos, G. D. (1987). Science 238, 1079–1087. Andreef, M., Darzynkiewicz, Z., Sharpless, T., Clarkson, B., and Melamed, M. (1980). Blood 55, 282–293. Antin, J. H., Emerson, S. G., Martin, P., Gadol, N., and Ka, A. (1986). J. Immunol. 136, 505. Arpin, C., de Bouteiller, O., Razanajaona, D., Fugier-Vivier, I., Briere, F., et al. (1998). J. Exp. Med. 187, 1169–1178. Avet-Loiseau, H., Li, J. Y., Morineau, N., Facon, T., Brigaudeau, C., et al. (1999). Blood 94, 2583–2589. Bahler, D. W., and Levy, R. (1992). Proc. Natl. Acad. Sci. USA 89, 6770–6774. Bahler, D. W., Campbell, M. J., Hart, S., Miller, R. A., Levy, S., and Levy, R. (1991). Blood 78, 1561–1568. Bahler, D. W., Miklos, J. A., and Swerdlow, S. H. (1997). Blood 89, 3335–3344. Bakkus, M. H., Heirman, C., Van Riet, I., Van Camp, B., and Thielemans, K. (1992). Blood 80, 2326–2335. Bakkus, M. H., Van Riet, I., Van Camp, B., and Thielemans, K. (1994). Br. J. Haematol. 87, 68–74. Bemark, M., and Neuberger, M. S. (2000). Oncogene 19, 3404–3410. Bentley, D. L., and Rabbits, T. H. (1980). Nature 288, 730–733. Berek, C. (1992). Immunol. Rev. 126, 5–19. Berek, C., and Milstein, C. (1987). Immunol. Rev. 96, 23–41. Berenson, J. R., Vescio, R. A., Hong, C. H., Cao, J., Kim, A., et al. (1995). Curr. Topics Microbiol. Immunol. 194, 25–33. Bergsagel, P. L., Chesi, M., Nardini, E., Brents, L. A., Kirby, S. L., and Kuehl, W. M. (1996). Proc. Natl. Acad. Sci. USA 93, 13931–13936. Betz, A. G., Neuberger, M. S., and Milstein, C. (1993). Immunol. Today 14, 405–411. Bhat, N. M., Kantor, A. B., Bieber, M. M., Stall, A. M., Herzenberg, L. A., and Teng, N. N. (1992). Int. Immunol. 4, 243–252. Billadeau, D., Ahmann, G., Greipp, P., and Van Ness, B. (1993). J. Exp. Med. 178, 1023–1031. Binet, J.-L., Lepoprier, M., Dighiero, G., Charron, D., D’Athis, P., et al. (1977). Cancer 40, 855–864. Blade, J., and Kyle, R. A. (1999). Hematol. Oncol. Clin. North Am. 13, 1259–1272. Brezinschek, H. P., Brezinschek, R. I., and Lipsky, P. E. (1995). J. Immunol. 155, 190–202. Brezinschek, H. P., Foster, S. J., Brezinschek, R. I., Dorner, T., Domiati-Saad, R., and Lipsky, P. E. (1997). J. Clin. Invest. 99, 2488–2501. Bross, L., Fukita, Y., McBlane, F., Demolliere, C., Rajewsky, K., and Jacobs, H. (2000). Immunity 13, 589–597. Caligaris-Cappio, F. (1996). Blood 87, 2615–2620. Chan, J. K., Banks, P. M., Cleary, M. L., Delsol, G., De Wolf-Peeters, C., et al. (1995). Am. J. Clin. Pathol. 103, 543–560. Chang, B., and Casali, P. (1994). Immunol. Today 15, 367–373. Chapman, C. J., Spellerberg, M. B., Smith, G. A., Carter, S. J., Hamblin, T. J., and Stevenson, F. K. (1993). J. Immunol. 151, 1051–1061. Chapman, C. J., Spellerberg, M. B., Hamblin, T. J., and Stevenson, F. K. (1995). Mol. Immunol. 32, 347–353.
The Occurrence and Significance of V Gene Mutations
111
Chapman, C. J., Mockridge, C. I., Hamblin, T. J., and Stevenson, F. K. (1996). Clin. Exp. Immunol. 105, 360–368. Chen, B. J., and Epstein, J. (1996). Blood 87, 1972–1976. Chesi, M., Bergsagel, P. L., Shonukan, O. O., Martelli, M. L., Brents, L. A., et al. (1998a). Blood 91, 4457– 4463. Chesi, M., Nardini, E., Lim, R. S., Smith, K. D., Kuehl, W. M., and Bergsagel, P. L. (1998b). Blood 92, 3025–3034. Choe, J., and Choi, Y. S. (1998). Eur. J. Immunol. 28, 508–515. Chu, Y. W., Marin, E., Fuleihan, R., Ramesh, N., Rosen, F. S., et al. (1995). J. Clin. Invest. 95, 1389–1393. Cleary, M. L., Galili, N., and Sklar, J. (1986a). J. Exp. Med. 164, 315–320. Cleary, M. L., Meeker, T. C., Levy, S., Lee, E., Trela, M., et al. (1986b). Cell 44, 97–106. Cook, G. P., and Tomlinson, I. M. (1995). Immunol. Today 16, 237–242. Cook, W. D., Rudikoff, S., Giusti, A. M., and Scharff, M. D. (1982). Proc. Natl. Acad. Sci. USA 79, 1240–1244. Corbett, S. J., Tomlinson, I. M., Sonnhammer, E. L. L., Buck, D., and Winter, G. (1997). J. Mol. Biol. 270, 587–597. Corradini, P., Boccadoro, M., Voena, C., and Pileri, A. (1993). J. Exp. Med. 178, 1091–1096. Cunningham, D., Paz-Ares, L., Gore, M. E., Malpas, J., Hickish, T., et al. (1994). J. Clin. Oncol. 12, 764–768. Damle, R. N., Wasil, T., Fais, F., Ghiotto, F., Valetto, A., et al. (1999). Blood 94, 1840–1847. Deane, M., and Norton, J. (1991). Leukemia 5, 646–650. Denepoux, S., Razanajaona, D., Blanchard, D., Meffre, G., Capra, J. D., et al. (1997). Immunity 6, 35– 46. de Wildt, R. M., Hoet, R. M. A., van Venrooij, W. J., Tomlinson, I. M., and Winter, G. (1999). J. Mol. Biol. 285, 895–901. Dighiero, G., Guilbert, B., Fermand, J. P., Lymberi, P., Danon, F., and Avrameas, S. (1983). Blood 62, 264–270. Dorner, T., Brezinschek, H. P., Brezinschek, R. I., Foster, S. J., Domiati-Saad, R., and Lipsky, P. E. (1997). J. Immunol. 158, 2779–2789. Dorner, T., Foster, S. J., Brezinschek, H. P., and Lipsky, P. E. (1998). Immunol. Rev. 162, 161–171. Dorner, T., Farner, N. L., and Lipsky, P. E. (1999). Immunologist 7, 153–162. Du, M., Diss, T. C., Xu, C., Peng, H., Isaacson, P. G., and Pan, L. (1996). Leukemia 10, 1190–1197. Dunn-Walters, D. K., Hackett, M., Boursier, L., Ciclitira, P. J., Morgan, P. (2000). J. Immunol 164, 1595–1601. Durie, B., and Salmon, S. I. (1985). “Staging, Kinetics and Flow Cytometry of Multiple Myeloma,” Vol. 2, pp. 513–528. Churchill Livingstone, New York. Ebeling, S., Schutte, M., Akkermans-Koolhaas, K., Bloem, A., Gmelig-Meyling, F., and Logtenberg, T. (1992). Int. Immunol. 4, 313–320. Fais, F., Ghiotto, F., et al. (1998). J. Clin. Invest. 102, 1515–1525. Fischer, M., Klein, U., and Kuppers, R. (1997). J. Clin. Invest. 100, 1667–1676. Forconi, F., Sahota, S. S., Raspadori, D., Mockridge, C. I., Lauria, F., et al. (2001). Blood 98, in press. Fujieda, S., Lin, Y. Q., Saxon, A., and Zhang, K. (1996a). J. Immunol. 157, 3450–3459. Fujieda, S., Saxon, A., and Zhang, K. (1996b). Mol. Immunol. 33, 1335–1343. Fukuda, T., Yoshida, T., Okada, S., Hatano, M., Miki, T., et al. (1997). J. Exp. Med. 186, 439–448. Gabrea, A., Bergsagel, P. L., Chesi, M., Shou, Y., and Kuehl, W. M. (1999). Mol. Cell. 3, 119–123. Gay, D., Saunders, T., Camper, S., and Weigert, M. (1993). J. Exp. Med. 177, 999–1008.
112
Freda K. Stevenson et al.
George, A. J., and Stevenson, F. K. (1989). Int. Rev. Immunol. 4, 271–310. Giachino, C., Padovan, E., and Lanzavecchia, A. (1995). J. Exp. Med. 181, 1245–1250. Girschick, H. J., Grammer, A. C., Nanki, T., Mayo, M., and Lipsky, P. E. (2001). J. Immunol. 166, 377–386. Goodglick, L., and Braun, J. (1994). Am. J. Pathol. 144, 623–636. Goossens, T., Klein, U., and Kuppers, R. (1998). Proc. Natl. Acad. Sci. USA 95, 2463–2468. Graninger, W. B., Seto, M., Boutain, B., Goldman, P., and Korsmeyer, S. J. (1987). J. Clin. Invest. 80, 1512–1515. Hallek, M., Bergsagel, P. L., and Anderson, K. C. (1998). Blood 91, 3–21. Hamblin, T. J., Davis, Z., Gardiner, A., Oscier, D. G., and Stevenson, F. K. (1999). Blood 94, 1848–1854. Harris, N. L., Jaffe, E. S., Diebold, J., Flandrin, G., Muller-Hermelink, H. K., et al. (1999). Ann. Oncol. 10, 1419–1432. Hiom, K., Melek, M., and Gellert, M. (1998). Cell 94, 463– 470. Hsu, F. J., and Levy, R. (1995). Blood 86, 3072–3082. Hsu, F. J., Caspar, C. B., Czerwinski, D., Kwak, L. W., Liles, T. M., et al. (1997). Blood 89, 3129–3135. Hummel, M., Tamaru, J., Kalvelage, B., and Stein, H. (1994). Blood 84, 403– 407. Hummel, M., Ziemann, K., Lammert, H., Pileri, S., Sabattini, E., and Stein, H. (1995). N. Engl. J. Med. 333, 901–906. Hummel, M., Marafioti, T., Ziemann, K., and Stein, H. (1996). Ann. Oncol. 4(suppl. 7), 31– 33. Isaacson, P. G. (1999). Ann. Oncol. 10, 637–645. Juliusson, G., Oscier, D., et al. (1990). N. Engl. J. Med. 323, 720–724. Jungnickel, B., Staratschek-Jox, A., Brauninger, A., Spieker, T., Wolf, J., et al. (2000). J. Exp. Med. 191, 395– 402. Kanzler, H., Kuppers, R., Hansmann, M. L., and Rajewsky, K. (1996). J. Exp. Med. 184, 1495–1505. Kay, N. E., Leong, T., Kyle, R. A., Greipp, P., Billadeau, D., et al. (1997). Blood 90, 340–345. Kelsoe, G. (1995). Immunol. Today 16, 324–326. Kelsoe, G. (1996). Immunity 4, 107–111. King, C. A., Spellerberg, M. B., Zhu, D., Rice, J., Sahota, S. S., et al. (1998). Nat. Med. 4, 1281–1286. Kipps, T., Tomhave, E., Pratt, L., Duffey, S., Chen, P., and Carson, D. (1989). Proc. Natl. Acad. Sci. USA 86, 5913–5917. Kipps, T. J. (1989). Adv. Immunol. 47, 117–185. Kirkham, P. M., and Schroeder, H. W., Jr. (1994). Semin. Immunol. 6, 347–360. Kiyoi, H., Naito, K., Ohno, R., Saito, H., and Naoe, T. (1998). Leukemia 12, 601–609. Klein, G. (1989). Genes Chromosomes Cancer 1, 3–8. Klein, U., Klein, G., Ehlin-Henriksson, B., Rajewsky, K., and Kuppers, R. (1995). Mol. Med. 1, 495–505. Klein, U., Kuppers, R., and Rajewsky, K. (1997). Blood 89, 1288–1298. Klein, U., Goossens, T., Fischer, M., Kanzler, H., Braeuninger, A., et al. (1998a). Immunol. Rev. 162, 261–280. Klein, U., Rajewsky, K., and Kuppers, R. (1998b). J. Exp. Med. 188, 1679–1689. Klix, N., Jolly, C. J., Davies, S. L., Bruggemann, M., Williams, G. T., and Neuberger, M. S. (1998). Eur. J. Immunol. 28, 317–326. Kosco, M. H., Pflugfelder, E., and Gray, D. (1992). J. Immunol. 148, 2331–2339. Kosmas, C., Viniou, N. A., Stamatopoulos, K., Courtenay-Luck, N. S., Papadaki, T., et al. (1996). Br. J. Haematol. 94, 306–317. Kosmas, C., Stamatopoulos, K., and Loukopoulos, D. (1997). Blood 90, 1334–1335.
The Occurrence and Significance of V Gene Mutations
113
Kosmas, C., Stamatopoulos, K., Papadaki, T., Belessi, C., Yataganas, X., et al. (1998). Immunol. Rev. 162, 281–292. Kraj, P., Friedman, D. F., Stevenson, F., and Silberstein, L. E. (1995). J. Immunol. 154, 6406– 6420. Kume, M., Suzuki, R., Yatabe, Y., Kagami, Y., Miura, I., et al. (1997). Jpn. J. Cancer Res. 88, 1087–1093. Kuppers, R., and Rajewsky, K. (1998). Annu. Rev. Immunol. 16, 471– 493. Kuppers, R., Rajewsky, K., Zhao, M., Simons, G., Laumann, R., et al. (1994). Proc. Natl. Acad. Sci. USA 91, 10962–10966. Kuppers, R., Rajewsky, K., and Hansmann, M. L. (1997). Eur. J. Immunol. 27, 1398–1405. Kuppers, R., Goossens, T., and Klein, U. (1999). Curr. Topics Microbiol. Immunol. 246, 193– 198. Kuze, T., Nakamura, N., Hashimoto, Y., and Abe, M. (1998). Leukemia 12, 753–757. Kyle, R. A., and Lust, J. A. (1990). Adv. Clin. Chem. 28, 145–218. Lam, K. P., Kuhn, R., and Rajewsky, K. (1997). Cell 90, 1073–1083. Lengauer, C., Kinzler, K. W., and Vogelstein, B. (1998). Nature 396, 643–649. Lewis, S., and Gellert, M. (1989). Cell 59, 585–588. Li, Y., Spellerberg, M. B., Stevenson, F. K., Capra, J. D., and Potter, K. N. (1996). J. Mol. Biol. 256, 577–589. Limpens, J., de Jong, D., van Krieken, J. H., Price, C. G., Young, B. D., et al. (1991). Oncogene 6, 2271–2276. Limpens, J., Stad, R., Vos, C., de Vlaam, C., de Jong, D., et al. (1995). Blood 85, 2528– 2536. Liu, Y. J., Zhang, J., Lane, P. J., Chan, E. Y., and MacLennan, I. C. (1991). Eur. J. Immunol. 21, 2951–2962. Liu, Y. J., de Bouteiller, O., and Fugier-Vivier, I. (1997). Curr. Opin. Immunol. 9, 256–262. Lokhorst, H. M., Schattenberg, A., Cornelissen, J. J., van Oers, M. H., Fibbe, W., et al. (2000). J. Clin. Oncol. 18, 3031–3037. Lossos, I. S., and Levy, R. (2000). Blood 96, 635–639. Lossos, I. S., Alizadeh, A. A., Eisen, M. B., Chan, W. C., Brown, P. O., et al. (2000a). Proc. Natl. Acad. Sci. USA 97, 10209–10213. Lossos, I. S., Tibshirani, R., Narasimhan, B., and Levy, R. (2000b). J. Immunol. 165, 5122– 5126. MacLennan, I. C. (1992). Curr. Topics Microbiol. Immunol. 182, 209–214. MacLennan, I. C. (1994). Annu. Rev. Immunol. 12, 117–139. Maloney, D. G., Grillo-Lopez, A. J., Bodkin, D. J., White, C. A., Liles, T. M., et al. (1997). J. Clin. Oncol. 15, 3266–3274. Maloum, K., Magnac, C., Azgui, Z., Cau, C., Charlotte, F., et al. (1998). Br. J. Haematol. 101, 171–178. Mantovani, L., Wilder, R. L., and Casali, P. (1993). J. Immunol. 151, 473– 488. Matsuda, F., Shin, E., Nagaoka, H., Matsumura, R., Haino, M., et al. (1993). Nature Genet. 3, 88–94. Matsumoto, M., Lo, S. F., Carruthers, C. J., Min, J., Mariathasan, S., et al. (1996). Nature 382, 462– 466. Matsuoka, M., Yoshida, K., Maeda, T., Usuda, S., and Sakano, H. (1990). Cell 62, 135–142. Matutes, E., Owusu-Ankomah, K., Morilla, R., et al. (1994). Leukemia 8, 1640–1645. Meffre, E., Papavasiliou, F., Cohen, P., de Bouteiller, O., Bell, D., et al. (1998). J. Exp. Med. 188, 765–772. Migliazza, A., Martinotti, S., Chen, W., Fusco, C., Ye, B. H., et al. (1995). Proc. Natl. Acad. Sci. USA 92, 12520–12524. Miklos, J. A., Swerdlow, S. H., and Bahler, D. W. (2000). Blood 95, 3878–3884.
114
Freda K. Stevenson et al.
Mockridge, C. I., Chapman, C. J., Spellerberg, M. B., Sheth, B., Fleming, T. P., et al. (1998). Clin. Exp. Immunol. 114, 129–136. Montesinos-Rongen, M., Kuppers, R., Schluter, D., Spieker, T., Van Roost, D., et al. (1999). Am. J. Pathol. 155, 2077–2086. Muschen, M., Re, D., Brauninger, A., Wolf, J., Hansmann, M. L., et al. (2000). Cancer Res 60, 5640–5643. Nakamura, N., Kuze, T., Hashimoto, Y., Tasaki, K., Hojo, H., et al. (1999). Pathol. Int. 49, 595–600. Neuberger, M. S., Ehrenstein, M. R., Klix, N., Jolly, C. J., Yelamos, J., et al. (1998). Immunol. Rev. 162, 107–116. Ng, V. L., and McGrath, M. S. (1998). Immunol. Rev. 162, 293–298. Nicholson, I., Brisco, M., and Zola, H. (1995). J. Immunol. 154, 1105–1113. Noppe, S. M., Heirman, C., Bakkus, M. H., Brissinck, J., Schots, R., and Thielemans, K. (1999). Br. J. Haematol. 107, 625–640. Nowell, P. C., and Croce, C. M. (1988). FASEB J. 2, 3054–3060. Nussenzweig, M. C. (1998). Cell 95, 875–878. Oscier, D. G., Thompsett, A., Zhu, D., and Stevenson, F. K. (1997). Blood 89, 4153– 4160. Oscier, D., Orchard, J., Ibbotson, R., Davis, Z., Stevenson, F., and Hamblin, T. (2000). Blood (suppl) in press. Ottensmeier, C. H., Thompsett, A. R., Zhu, D., Wilkins, B. S., Sweetenham, J. W., and Stevenson, F. K. (1998). Blood 91, 4292– 4299. Paramithiotis, E., and Cooper, M. D. (1997). Proc. Natl. Acad. Sci. USA 94, 208–212. Pascual, V., Victor, K., Spellerberg, M., Hamblin, T. J., Stevenson, F. K., and Capra, J. D. (1992). J. Immunol. 149, 2337–2344. Perlmutter, A. P., and Gilbert, W. (1984). Proc. Natl. Acad. Sci. USA 81, 7189–7193. Pezzella, F., Tse, A. G. D., Cordell, J. L., Pulford, K. A. F., Gatter, K. C., and Dy, M. (1990). Am. J. Pathol. 137, 225–232. Pilarski, L. M., Hipperson, G., Seeberger, K., Pruski, E., Coupland, R. W., and Belch, A. R. (2000). Blood 95, 1056–1065. Portlock, C. S. (1990). Semin. Oncol. 17, 51–59. Qin, Y., Greiner, A., Hallas, C., Haedicke, W., and Muller-Hermelink, H. K. (1997). Lab. Invest. 76, 447– 485. Rajewsky, K. (1996). Nature 381, 751–758. Ralph, Q. M., Brisco, M. J., Joshua, D. E., Brown, R., Gibson, J., and Morley, A. A. (1993). Blood 82, 202–206. Rassenti, L. Z., and Kipps, T. J. (1997). J. Exp. Med. 185, 1435–1445. Razanajaona, D., Denepoux, S., Blanchard, D., de Bouteiller, O., Liu, Y. J., et al. (1997). J. Immunol. 159, 3347–3353. Rettig, M. B., Vescio, R. A., Cao, J., Wu, C. H., Lee, J. C., et al. (1996). Blood 87, 2846–2852. Riboldi, P., Gaidano, G., Schettino, E. W., Steger, T. G., Knowles, D. M., et al. (1994). Blood 83, 2952–2961. Roelcke, D. (1974). Clin. Immunol. Immunopathol. 2, 266–280. Rothwell, D., Kalakonda, N., Scaffe, J., and Norton, J. (1999). Blood 94(Suppl. 1), 543a. Rubinstein, D. B., Symann, M., Stewart, A. K., and Guillaume, T. (1993). Mol. Immunol. 30, 403– 412. Sahota, S., Hamblin, T., Oscier, D. G., and Stevenson, F. K. (1994). Leukemia 8, 1285–1289. Sahota, S. S., Leo, R., Hamblin, T. J., and Stevenson, F. K. (1996). Blood 87, 746–755. Sahota, S. S., Leo, R., Hamblin, T. J., and Stevenson, F. K. (1997). Blood 89, 219–226. Sahota, S. S., Garand, R., Bataille, R., Smith, A. J., and Stevenson, F. K. (1998). Blood 91, 238–243. Sahota, S. S., Garand, R., Mahroof, R., Smith, A., Juge-Morineau, N., et al. (1999). Blood 94, 1070–1076.
The Occurrence and Significance of V Gene Mutations
115
Sahota, S. S., Davis, Z., Hamblin, T. J., and Stevenson, F. K. (2000). Blood 95, 3534–3540. Sale, J. E., and Neuberger, M. S. (1998). Immunity 9, 859–869. Schatz, D. G., Oettinger, M. A., and Schlissel, M. S. (1992). Annu. Rev. Immunol. 10, 359–383. Schettino, E. W., Chai, S. K., Kasaian, M. T., Schroeder, H. W., Jr., and Casali, P. (1997). J. Immunol. 158, 2477–2489. Schroeder, H. W., Jr., and Dighiero, G. (1994). Immunol. Today 15, 288–294. Shen, H. M., Peters, A., Baron, B., Zhu, X., and Storb, U. (1998). Science 280, 1750–1752. Shimizu, A., Nussenzweig, M. C., Han, H., Sanchez, M., and Honjo, T. (1991). J. Exp. Med. 173, 1385–1393. Shou, Y., Martelli, M. L., Gabrea, A., Qi, Y., Brents, L. A., et al. (2000). Proc. Natl. Acad. Sci. USA 97, 228–233. Silberstein, L. E., Jefferies, L. C., Goldman, J., Friedman, D., Moore, J. S., et al. (1991). Blood 78, 2372–2386. Silverman, G. J. (1997). Immunol. Today 18, 379–386. Singhal, S., Mehta, J., Desikan, R., Ayers, D., Roberson, P., et al. (1999). N. Engl. J. Med. 341, 1565–1571. Sklar, J., Cleary, M. L., Thielemans, K., Gralow, J., Warnke, R., and Levy, R. (1984). N. Engl. J. Med. 311, 20–27. Stavnezer, J., Sirlin, S., and Abbott, J. (1985). J. Exp. Med. 161, 577–601. Stevenson, F., Sahota, S., Zhu, D., Ottensmeier, C., Chapman, C., et al. (1998). Immunol. Rev. 162, 247–259. Stevenson, F. K., Smith, G. J., North, J., Hamblin, T. J., and Glennie, M. J. (1989). Br. J. Haematol. 72, 9–15. Stevenson, F. K., Spellerberg, M. B., Chapman, C. J., and Hamblin, T. J. (1995). Leukocytes Lymphoma 16, 379–384. Symmans, W. F., Katz, R. L., Ordonez, N. G., Dalton, H., Romaguera, J. E., and Cabanillas, F. (1995). Acta Cytol. 39, 673–682. Szczepanski, T., van ’t Veer, M. B., Wolvers-Tettero, I. L., Langerak, A. W., and van Dongen, J. J. (2000). Blood 96, 1087–1093. Szczepek, A. J., Seeberger, K., Wizniak, J., Mant, M. J., Belch, A. R., and Pilarski, L. M. (1998). Blood 92, 2844–2855. Tamaru, J., Hummel, M., Marafioti, T., Kalvelage, B., Leoncini, L., et al. (1995). Am. J. Pathol. 147, 1398–1407. Tanaka, S., Louie, D. C., Kant, J. A., and Reed, J. C. (1992). Blood 79, 229–237. Taniguchi, M., Oka, K., Hiasa, A., Yamaguchi, M., Ohno, T., et al. (1998). Blood 91, 1145– 1151. Thompsett, A. R., Ellison, D. W., Stevenson, F. K., and Zhu, D. (1999). Blood 94, 1738–1746. Tierens, A., Delabie, J., Pittaluga, S., Driessen, A., and DeWolf-Peeters, C. (1998). Blood 91, 2381–2386. Toellner, K. M., Gulbranson-Judge, A., Taylor, D. R., Sze, D. M., and MacLennan, I. C. (1996). J. Exp. Med. 183, 2303–2312. Tonegawa, S. (1983). Nature 302, 575–581. Tsujimoto, Y., Cossman, J., Jaffe, E., and Croce, C. M. (1985a). Science 228, 1440–1443. Tsujimoto, Y., Gorham, J., Cossman, J., Jaffe, E., and Croce, C. M. (1985b). Science 229, 1390–1393. Tsujimoto, Y., Jaffe, E., Cossman, J., Gorham, J., Nowell, P. C., and Croce, C. M. (1985c). Nature 315, 340–343. Vaandrager, J. W., Schuuring, E., Kluin-Nelemans, H. C., Dyer, M. J., Raap, A. K., and Kluin, P. M. (1998). Blood 92, 2871–2878. Vaandrager, J. W., Schuuring, E., Philippo, K., and Kluin, P. M. (2000). Blood 96, 1947–1952. Vescio, R. A., Cao, J., Hong, C. H., Lee, J. C., Wu, C. H., et al. (1995). J. Immunol. 155, 2487–2497.
116
Freda K. Stevenson et al.
Voso, M. T., Hohaus, S., Moos, M., and Haas, R. (1997). Blood 89, 3763–3768. Wagner, S. D., Martinelli, V., and Luzzatto, L. (1994). Blood 83, 3647–3653. Wang, X., and Stollar, B. (1999). Clin. Immunol. 93, 132–142. Wang, Y., Huang, G., Wang, J., Molina, H., Chaplin, D. D., and Fu, Y. X. (2000). J. Immunol. 30, 2226–2234. Wang, Y. H., Nomura, J., Faye-Petersen, O. M., and Cooper, M. D. (1998). J. Immunol. 161, 1132–1139. Weiss, E. A., Tucker, P. W., and Yuan, D. (1987). J. Mol. Cell. Immunol. 3, 69–81. Willis, T. G., and Dyer, M. J. (2000). Blood 96, 808–822. Wilson, P. C., de Bouteiller, O., Liu, Y. J., Potter, K., Banchereau, J., et al. (1998). J. Exp. Med. 187, 59–70. Wilson, P. C., Wilson, K., Liu, Y. J., Banchereau, J., Pascual, V., and Capra, J. D. (2000). J. Exp. Med. 191, 1881–1894. Wu, H. Y., and Kaartinen, M. (1995). Scand. J. Immunol. 42, 52–59. Ye, B. H., Chaganti, S., Chang, C. C., Niu, H., Corradini, P., et al. (1995a). EMBO J. 14, 6209–6217. Ye, B. H., Lo Coco, F., Chang, C. C., Zhang, J., Migliazza, A., et al. (1995b). Curr. Topics Microbiol. Immunol. 194, 101–108. Yelamos, J., Klix, N., Goyenechea, B., Lozano, F., Chui, Y. L., et al. (1995). Nature 376, 225– 259. Youinou, P., Mackenzie, L. E., Lamour, A., Mageed, R. A., and Lydyard, P. M. (1993). Eur. J. Clin. Invest. 23, 139–150. Yunis, J. J., Frizzera, G., Oken, M. M., McKenna, J., Theologides, A., and Arnesen, M. (1987). N. Engl. J. Med. 316, 79–84. Zajac-Kaye, M., Gelmann, E. P., and Levens, D. (1988). Science 240, 1776–1780. Zan, H., Cerutti, A., Dramitinos, P., Schaffer, A., Li, Z., and Casali, P. (1999). J. Immunol. 162, 3437–3447. Zelenetz, A. D., Chen, T. T., and Levy, R. (1992). J. Exp. Med. 176, 1137–1148. Zhang, J., Alt, F., and Honjo, T. (1995). “Regulation of class switch recombination of the immunoglobulin heavy chain genes,” pp. 235–265. Academic Press, London. Zhu, D., Hawkins, R. E., Hamblin, T. J., and Stevenson, F. K. (1994). Br. J. Haematol. 86, 505–512. Zhu, D., Oscier, D. G., and Stevenson, F. K. (1995). Blood 85, 1603–1607. Zhu, D., van Arkel, C., King, C. A., Meirvenne, S. V., de Greef, C., et al. (1998). Immunology 93, 162–170. Zomas, A. P., Matutes, E., Morilla, R., Owusu-Ankomah, K., Seon, B. K., and Catousky, D. (1996). Leukemia 10, 1966–1970.
MHC Antigens and Tumor Escape from Immune Surveillance Federico Garrido1 and Ignacio Algarra2 1 Departamento de Analisis Clinicos Hospital Universitario Virgen de las Nieves 18014 Granada 2 Departamento de Ciencias de la Salud Universidad de Jaen Jaen, Spain
I. Introduction II. HLA Class I Antigen Expression in Primary Tumors A. Altered HLA Class I Phenotypes B. HLA Expression in Different Tumor Tissues III. Changes in MHC Class I Antigen Expression during Metastatic Colonization IV. T Cell Immunoselection of MHC Class I–Negative Tumor Clones V. Expression of Nonclassical HLA Class I Molecules in Tumors VI. Tumor NK Escape Mechanisms VII. HLA Class I Loss and T Cell–Based Immunotherapy VIII. HLA Class II Antigens in Tumors IX. Conclusions References
I. INTRODUCTION It has long been proposed that our immune system subjects the cells of our body to constant surveillance, distinguishing those that are normal from those that have undergone aberrant transformation (Ehrlich, 1909; Thomas, 1959). It was also proposed that T cell-mediated immunity performs this surveillance in order to destroy the transformed cells, thus avoiding the spread of disease. The term immune surveillance was coined to describe the natural immunological host resistance to the development of cancer (Burnet, 1970). This theory received significant support when the antigen presentation pathway to T lymphocytes was discovered (Townsend et al., 1986), when the tumor antigens recognized by T lymphocytes were identified (Boon, 1983), and when the molecular basis of natural killer (NK) cell function was partially elucidated (Moretta et al., 1996).
Advances in CANCER RESEARCH 0065-230X/01 $35.00
117
C 2001 by Academic Press. Copyright All rights of reproduction in any form reserved.
118
Federico Garrido and Ignacio Algarra
Tumor cells grow, invade, and metastasize despite an active and a priori normal immune response of a healthy immune system (Klein et al., 1960; Boon et al., 1994). Data obtained in different laboratories in the past 20 years indicate that this is due in many instances to a highly sophisticated and poorly understood process of selection of MHC class I-deficient tumor escape variants (Festenstein and Garrido, 1986; Garrido et al., 1997a). This widespread escape strategy used by tumor cells allows them to behave as stealth targets to immune effectors (Ljunggren and Karre, 1985; Karre et al., 1986). This is not surprising since MHC genes control the synthesis of molecules that are at the center of the immune function mediated by T lymphocytes and NK cells. Immune surveillance against cancer has thus been demonstrated by the detection of these escape routes used by tumor cells; findings to date suggest that without an escape mechanism in a particular tumor, there is no tumor growth or growth is confined and controlled by the host. MHC class I downregulation in experimental or spontaneous tumors is a mechanism used frequently by tumor cells to escape recognition and destruction by cytotoxic T lymphocytes (Garrido et al., 1976, 1993; Ferrone et al., 1995). Similarly, viruses have evolved strategies to interfere with antigen presentation by HLA class I molecules (Alcami and Koszinowski, 2000). These parallel strategies for selection have probably evolved in different ways and converge at the level of antigen presentation; however, the mechanisms are diverse and not necessarily overlapping. Viruses carry the appropriate genes that code for the escape proteins selected during evolution, whereas tumors appear to increase genetic instability, which probably affects HLA and other genes responsible for key immunological functions. This mechanism provides a diversity of tumor phenotypes that are presumably selected by active immune surveillance. Interestingly, NK cells seem to have evolved to destroy MHC-deficient target cells and therefore may play an important role when they encounter these T cell-resistant target cells. It is therefore important to precisely define these MHC-deficient tumor cells, especially when heterogeneous populations exist in tumor tissues. Our group has been developing new strategies and extensively analyzing the MHC altered phenotypes found in a variety of human tumors. This review summarizes evidence that primary and metastatic tumor cell growth results from the development of sophisticated molecular and biological mechanisms that allow tumor cells to escape immune surveillance. These escape mechanisms are selected by the cancer cells after a period of interaction with the immune system. Among these mechanisms, the MHC class I phenotypic alteration that occurs in tumor cells plays a leading role in the tumor–host scenario since these are crucial molecules for antigen presentation to T cells and modulation of NK activity.
MHC Antigens and Tumor Escape from Immune Surveillance
119
This review also provides a description of the major HLA class I phenotypic alterations found in human tumors, with special attention paid to the molecular mechanisms responsible for their generation. It also describes the HLA class I alterations that are detected in tumors derived from different tissues of the body, such as the bladder, breast, colon and rectum, larynx, lung, kidney, melanocytes, pancreas, and prostate. The issue of nonclassical HLA class I molecule expression in tumor cell lines and tumor tissues and its possible role in immune escape is also discussed, as are the implications of these findings for T cell-based immunotherapy. Finally, we summarize experimental data supporting the hypothesis that the MHC class I-negative tumor variants are selected in vivo by cytotoxic lymphocyte (CTL) responses against MHC class I-positive tumor cells.
II. HLA CLASS I ANTIGEN EXPRESSION IN PRIMARY TUMORS Data have accumulated in recent years indicating that alterations in HLA class I expression are a widespread finding in most tumors analyzed (Festenstein, 1987; Garrido et al., 1997a). The frequency of such phenomena is evaluated by studying series of tumor samples by immunohistological techniques or flow cytometry in disrupted tumor cell suspensions, and with monoclonal antibodies (mAbs) directed against HLA class I monomorphic, HLA-A or -B locus-specific, or HLA allelic epitopes (Garrido et al., 1997b; Koopman et al., 2000) (Fig. 1, see color insert). Most anti-HLA mAbs do not work properly in tissues or in cell lines since they have been produced with and selected against peripheral blood lymphocytes. They need to be carefully evaluated when studying HLA antigens in tumor tissue samples. This is particularly important when trying to define HLA alleles since these anti-HLA antibodies are difficult to find. A consensus by different research groups is that the more anti-HLA mAbs defining HLA alleles are available, the more HLA class I losses are detected. The rates of HLA class I loss in some tumors are near 100%; for example, the rate is 96% in cervix carcinomas (Koopman et al., 2000), 96% in breast carcinomas (Cabrera et al., 1996), 87% in colorectal carcinomas (Cabrera et al., 1998), and 70% in laringeal carcinomas (Cabrera et al., 2000) (Table I). However, it should be remembered that reactivity with one or two mAbs defining a particular HLA epitope may not be sufficient to establish the loss of a particular cell surface molecule. In our experience and that of other groups, the correlation between such reactions in tissues and the presence or absence of HLA cell surface expression is very high. This is observed when the tumor tissue and
120
Federico Garrido and Ignacio Algarra
Table I Percentage of HLA Class I Losses in Human Tumorsa Tumor
Percent
Bladder Breast Cervix Colorectal Head and neck Kidney Lung Melanoma Pancreas Prostate
85 96 96 87 70 38b 38b 63 39b 85
a These data include HLA total loss, HLA-A and -B locus loss, HLA haplotype loss, and HLA allelic losses. b The frequency of HLA losses in these tumors was analyzed only with mAbs directed against HLA monomorphic determinants.
the corresponding tumor cell line of a particular patient are available and can be carefully analyzed (Torres et al., 1996; Benitez et al., 1998). Information about the mechanism(s) responsible for these HLA alterations is also available from studies that analyzed cell lines derived from a particular tumor tissue (Garrido et al., 1995; Real et al., 1998). Recent analyses of cervix tumors by two independent groups concluded that multiple mechanisms are responsible for HLA class I downregulation (Koopman et al., 2000; Brady et al., 2000). These mechanisms include β 2 microglobulin mutations for HLA total loss, loss of heterozygosity (LOH) associated with chromosome 6 for HLA haplotype loss, HLA class I allele mutations for HLA allele loss, and HLA-A and -B locus alteration in the downregulation of transcription of HLA-A or -B locus products. These mechanisms are not unique for cervix carcinomas but are found in tumors of different histological origin. However, tumors derived from certain tissues appear to exhibit mechanisms that occur more frequently than others. For instance, β 2 microglobulin mutations seem to occur quite frequently in melanomas to generate HLA class I total loss (P´erez et al., 1999), but the same altered HLA class I phenotype found in laryngeal carcinomas is not associated with such molecular lesion (Fernandez et al., 2000; Feenstra et al., 1999a). In this context, LOH associated with chromosome 6p.21 is the most widespread mechanism used by tumor cells to lose HLA class I antigens. HLA haplotype loss occurs relatively frequently in all tumors analyzed to date (e.g., colorectal, laryngeal, melanoma, and cervix; Jimenez et al., 1999; Brady et al., 2000; Feenstra et al., 1999b). Studies of LOH by analysis of polymorphic short tamden
MHC Antigens and Tumor Escape from Immune Surveillance
121
repeats (STRs) in chromosome 6 and also by the direct HLA class I typing in microdissected tumor samples are helping to elucidate the percentage of LOH for different tumor tissues (Ramal et al., 2000a,b). Altered HLA class I phenotypes are found in primary tumors and also in ´ metastatic colonies (Lopez-Nevot et al., 1989). In the former, heterogeneous tumor cell populations are present. In the latter, colonies frequently composed of tumor cells with homogeneous phenotypes are detected (Benitez et al., 1998; Garcia Lora et al., 2001). The altered HLA class I phenotypes found in primary tumors and metastases are described in the next section.
A. Altered HLA Class I Phenotypes The classification of HLA class I altered phenotypes in human tumors has proved to be useful in establishing additional strategies to analyze the mechanisms responsible for such alterations (Garrido et al., 1995, 1997a). Such classifications may have useful clinical implications. A summary of the major HLA class I altered phenotypes seen in tumors derived from different tissues is presented here and in Fig. 2.
1. PHENOTYPE NO. I: HLA CLASS I TOTAL LOSS This alteration was originally described with the use of mAbs directed against HLA class I monomorphic determinants on tumor tissue sections, especially in melanoma lesions (Natali et al., 1983; Ruiter et al., 1984; ´ Lopez-Nevot et al., 1986; Parmiani et al., 1986) but also in other tumor samples, such as breast and colorectal carcinomas (P´erez et al., 1986; Momburg et al., 1986). The antibody W6/32, which defines a cell surface monomorphic determinant formed by a conformational epitope of the heavy chain and β 2 microglobulin, was widely used. Later, monoclonal antibodies against β 2 microglobulin light chain also came into use. This phenotype is characterized by the absence of any HLA class I antigen expression in tumor cells (Fig. 1) and is present to a different extent in different tumors. Its frequency is low in laryngeal carcinomas (10%), colorectal carcinomas (18%), and melanomas (17%) and higher in breast (52%), prostate (40%), and bladder (35%) carcinomas (Garrido et al., 1997a ). The mechanisms underlying these HLA class I total loss phenotypes are not fully understood. There are clear indications in melanomas that β 2 microglobulin mutations are involved in producing this phenotype (Wang et al., 1996), and a summary of such mutations was recently published (P´erez et al., 1999) (Fig. 3). There are also indications that β 2 microglobulin mutations are also involved in some cases of HLA class I total loss observed in colorectal carcinomas (Browning et al., 1996). However, our group obtained data
122
Federico Garrido and Ignacio Algarra
Fig. 2 Altered HLA class I phenotypes found in human tumors. A hypothetical normal cell expressing the HLA class I alleles (A2, A24, B7, B44, Cw2, and Cw7) can produce the following altered HLA phenotypes found in tumor tissues: No. I (HLA total loss), No. II (HLA haplotype loss) No. III (HLA-A, -B, or -C locus downregulation), No. IV (HLA allelic loss), and No. V (compound phenotype: phenotypes II + III).
indicating that mutations in this gene may be less frequently involved in producing HLA class I total loss in colorectal tumors (Fernandez et al., 2000). In some laryngeal carcinomas, we and others reported HLA class I total loss without β 2 microglobulin mutations (Feenstra et al., 1999a; Fernandez et al., 2000).
MHC Antigens and Tumor Escape from Immune Surveillance
123
Fig. 3 Summary of the β 2 microglobulin mutations described in human tumors. A hot spot region seems to be present in the highly repetitive CT zone in exon 1. Co, colorectal; Me, melanoma; L, lung; Ly, lymphoma.
A single defect in one β 2 microglobulin gene is not sufficient to produce a total HLA loss phenotype. LOH associated with chromosome 15 near the β 2 microglobulin gene is present in 30% of tumor tissues but does not affect the other β 2 gene and therefore expression of HLA class I products is normal (Ramal et al., 2000b). It has been reported, however, that expression of some HLA alleles requires higher levels of β 2 microglobulin and therefore may be more sensitive to LOH of one β 2 gene (Bicknell et al., 1994). These findings suggest that LOH in the β 2 gene is an early event in tumor development, and when combined with mutation in the other β 2 gene it will produce the HLA class I total loss phenotype (Benitez et al., 1998).
Table II Tumor cell line
124
Phenotype I β 2 microglobulin mutations Daudi LoVo HCT15/DLD1 SW48 HRA19 C84 H630 H2009 FO-1 SK-MEL BB74-MEL LB1622-MEL Me 1386 ME 18105 Me 9922 GR 34 TAP alterations PPC-1 MZ1851 MZ1879 MZ1940 H1436 H1092 H82 Phenotype II 915 877
Tumor
Molecular alteration
Reference
Burkitt’s lymphoma Colorectal Colorectal Colorectal Colorectal Colorectal Colorectal Lung adenocarcinoma Melanoma Melanoma Melanoma Melanoma Melanoma Melanoma Melanoma Melanoma
G ⇒ C at initation codon CT deletion in leader sequence C ⇒ A exon 2: G ⇒ T intron 1 CTCT deletion in leader sequence; deletion in exon 2 TCTT deletion in exon 2 G ⇒ A exon 2 CT deletion in exon 1 A ⇒ G at initation codon Deletion first exon and a segment of first intron G deletion in codon 76 C ⇒ G exon 2 T ⇒ A exon 1 CT deletion in exon 1 A ⇒ G in splice acceptor site of intron 1 14-bp deletion in exon 2 TTCT deletion in leader sequence
Rosa et al. (1983) Bicknell et al. (1994) Bicknell et al. (1994) Bicknell et al. (1994) Browning et al. (1996) Browning et al. (1996) Chen et al. (1996a) Chen et al. (1996a) D’Urso et al. (1991) Wang et al. (1993) Benitez et al. (1998) Benitez et al. (1998) Hicklin et al. (1998) Hicklin et al. (1998) Hicklin et al. (1998) P´ erez et al. (1999)
Prostate Renal Renal Renal Small cell lung Small cell lung Small cell lung
Regulatory defect NDa ND ND CGG ⇒ CAG exon 10 Regulatory defect Regulatory defect
Sanda et al. (1995) Seliger et al. (1997) Seliger et al. (1997) Seliger et al. (1997) Chen et al. (1994) Restifo et al. (1993) Restifo et al. (1993)
Cervix Cervix
LOH at 6p21 LOH at 6p21
Brady et al. (2000) Brady et al. (2000)
NW145 LB33-MEL.A IMIM-PC2 OCM-3
125
Melanoma Melanoma Pancreas Uveal melanoma
LOH at 6p21 LOH at 6p21 LOH at 6p21 LOH at 6p21
Mendez et al. (2000) Lehmann et al. (1995) Torres et al. (1996) Hurks et al. (2000)
Phenotype III CC10 603 634 823 136-2 453A FM55 NW16
Cervix Melanoma Melanoma Melanoma Melanoma Melanoma Melanoma Melanoma
Regulatory defect Regulatory defect Regulatory defect Regulatory defect Regulatory defect Regulatory defect Regulatory defect Regulatory defect
Koopman et al. (1998) Schier et al. (1991) Schier et al. (1991) Schier et al. (1991) Schier et al. (1991) Schier et al. (1991) Real et al. (1998) Mendez et al. (2000)
Phenotype IV HLA class I gene mutations CSCC7 808 778 LS 411 624MEL28
Cervical Cervical Cervical Colorectal Melanoma
TGGG insertion at codon 32 in exon 2 of HLA-B15 CAG ⇒ TAG in exon 3 of HLA-A2 G ⇒ C at the 3′ acceptor site of intron 1 of HLA-A2 Chromosomal breakpoint in HLA-A11 Base substitution at the 5′ donor site of intron 2 of HLA-A2
Koopman et al. (1999) Brady et al. (2000) Brady et al. (2000) Browning et al. (1993) Wang et al. (1999)
Phenotype V Compound phenotype CC11 LB33-MEL.B R22.2 FM37
Cervical Melanoma Melanoma Melanoma
G ⇒ T exon 2 of HLA-A24 + LOH LOH + not determined mechanism LOH + downregulation of locus B LOH + downregulation of locus B
Koopman et al. (1999) Lehmann et al. (1995) Real et al. (1998) Real et al. (1998)
Phenotype VI Unresponsivenes to γ -IFN AGS Caki2
Gastric Renal
Downregulation of transcriptional factor binding to IRSE Alterated TAP-1 and LMP2 expression by a defective IFN-γ signaling pathway
Abril et al. (1996) Dovhey et al. (2000)
a
ND, not determined.
126
Federico Garrido and Ignacio Algarra
Defects in the transporters associated with antigen processing (TAPs) or the proteosome components of low-molecular-weight proteins (LMPs) can also produce MHC class I total loss (Restifo et al., 1993; Sanda et al., 1995; Chen et al., 1996a; and Seliger et al., 1997), but the relevance of such mechanisms in different human tumors is unknown. Mechanisms such as hypermethylation of HLA class I promoter genes can also repress HLA class I expression producing phenotype No. I, as is demonstrated in melanoma cell line MSR3 (A. Serrano, C. Traversari, and F. Garrido, unpublished results). In other words, this is a well-established altered HLA phenotype that can be produced by several molecular mechanisms, most of which cannot be regulated by cytokines. For details of tumor cell lines with phenotype No. I, see Table II.
2. PHENOTYPE NO. II: HLA HAPLOTYPE LOSS This tumor phenotype is produced by LOH associated with chromosome 6, and it may affect a large portion of 6p, including the 6q region. It can be found in tumors derived from any tissue studied to date. However, the frequency varies in different tumors. Current percentages of LOH for the HLA region in chromosome 6 are 46% in cervix carcinomas, 15–49% in head and neck, 17% in colorectal carcinomas, and 14% in breast (Jimenez et al., 1999; Feenstra et al., 1999b; Koopman et al., 2000). This is a mechanism producing HLA class I alteration that is present in most tumors (Ramal et al., 2000b). The greatest obstacle to determining how often LOH affects the HLA region in solid tumor samples is the contaminating stroma present in most tumor tissues analyzed. To overcome this problem it is necessary to develop strategies that use microdissected material from cryopreserved tumor tissues and that make it possible to distinguish between stroma and tumor tissue. We recently used this approach. A cooperative study involving two laboratories (that of Dr. Tilanus in Utrecht and ours in Granada) defined basic criteria to establish HLA haplotype loss in microdisected tumor tissues (Ramal et al., 2000a). A nonradioactive method using polymerase chain reaction amplification of a selected panel of polymorphic STRs located in or near the HLA region was found to be a useful tool for identifying this HLA altered phenotype (Feenstra et al., 2000). Results from our group indicate that this could be a reliable method to define this HLA altered phenotype in microdissected human tumors (Ramal et al., 2000b). For details of tumor cell lines that have been reported to have HLA haplotype loss, see Table II. In summary, this phenotype can be precisely diagnosed using microdissected tumor tissue DNA. It probably arises in early stages of tumor
MHC Antigens and Tumor Escape from Immune Surveillance
127
development and can appear in combination with other HLA genetic lesions to produce compound phenotypes.
3. PHENOTYPE NO. III: HLA-A, -B, OR -C LOCUS PRODUCT DOWNREGULATION This altered phenotype is found when both products of HLA-A, -B, or -C loci are coordinately downregulated (Versteeg et al., 1989; Schrier et al., 1991). The mechanism of locus downregulation is often transcriptional since the levels of mRNA found in these tumor cell lines can frequently be upregulated with cytokines, and low expression of transcription factors that bind to locus-specific DNA motifs can induce HLA-B locus downregulation (Soong and Hui, 1992). In melanomas, selective HLA-B locus downregulation correlates with increased c-myc transcription (Peltenburg and Schrier, 1994). The definition of this altered phenotype in tumor tissues requires the use of anti-HLA-A, -B, or -C locus-specific mAbs. The HLA and cancer component of the XII International Histocompatibility Wokshop recommended the use of a selected mAb panel for HLA-A and HLA-B locus products (Garrido et al., 1997b). The definition of HLA-C expression in tissues will require further study since mAbs that define these gene products are lacking (Setini et al., 1996). In summary, to precisely define this phenotype, a tumor cell line is required since in most of the reported examples HLA expression can be recovered after cytokine treatment (Table II). In tissues, use of anti-HLA-A and -B locus-specific mAbs can help to establish the diagnosis.
4. PHENOTYPE NO. IV: HLA ALLELIC LOSS This alteration is defined as the loss of a single HLA class I allele (Garrido et al., 1997a). The use of anti-HLA class I mAbs that define HLA class I individual alleles is required to establish this diagnosis in tissues. In cell lines, immunoprecipitation with the w6/32 mAb and isoelectric focusing can be used to detect single allelic losses (Ruiz-Cabello et al., 1991; Real et al., 1998). Several examples of tumor cell lines with molecular defects have been reported, such as the colorectal carcinoma LS411, with a chromosomal break point in the HLA-A11 allele (Browning et al., 1993), or the cervical cell lines CC11 and CSCC7 (Koopman et al., 1999) or 808 and 778 (Brady et al., 2000; Serrano et al., 2000). The latter cell lines present a G ⇒ T substitution in exon 2 of HLA-A24, a TGGG insertion at codon 32 in exon 2 of HLA-B15, a CAG ⇒ TAG substitution in exon 3 of HLA-A2, and a point mutation (G ⇒ C) at the 3′ acceptor site of intron 1 in HLA-A2,
128
Federico Garrido and Ignacio Algarra
respectively. The melanoma cell line 624 MEL28 contains a base substitution at the 5′ donor site of intron 2 in HLA-A2 (Wang et al., 1999) (Table II). The frequency of these altered HLA phenotypes in different tumors is unknown since only a few anti-HLA class I mAbs are available for working with cryopreserved tissue samples (Cabrera et al., 1996, 1998, 2000).
5. PHENOTYPE NO. V: COMPOUND PHENOTYPES It is clear that many tumors display compound phenotypes. Thus, there are several examples in which the tumor cells only express a single HLA class I allele: melanomas expressing only HLA-A1 (Real et al., 1998) or HLA-A24 (Ikeda et al., 1997) or cervical carcinomas expressing only the HLA-A24 allele (Brady et al., 2000; Koopman et al., 2000). This phenotype requires a combination of at least two different alterations, for instance, an HLA haplotype loss and an HLA-B and -C locus dowregulation (a combination of phenotypes II and III) (Algarra et al., 2000).
6. PHENOTYPE NO. VI: UNRESPONSIVENESS TO INTERFERON Some tumor cells express basal levels of HLA class I antigens but have lost the capacity to upregulate these molecules in response to different cytokines, including α and γ interferons (IFNs). Examples have been reported in which the IFN signaling pathway is defective. For instance, the renal cell carcinoma Caki-2 does not have DNA binding activity for IFN regulatory factor-1 or signal transducer and activator of transcription (Stat-1) (Dovhey et al., 2000). We have also analyzed a gastric carcinoma cell line (AGS) that is completely defective in MHC class I response to IFN-α and -γ . We found that AGS had a low level of transcriptional factor binding to an IFN-responsive sequence element when compared with other IFN-responsive tumor cell lines (Abril et al., 1996). Defects in overlapping factors in the signal transduction pathway of both type I and type II IFNs may be a frequent cause of nonresponsiveness of tumor cells to these cytokines. These alterations represent an advantage for tumor growth and immune escape. Table II summarizes published tumor cell lines that present a particular alteration in HLA class I phenotype with a defined molecular mechanism.
B. HLA Expression in Different Tumor Tissues Altered HLA class I phenotypes can be found in tumor tissues derived from a variety of epithelia. Nevertheless, carcinogenesis in each tissue presents particular characteristics. The following sections describe the different
MHC Antigens and Tumor Escape from Immune Surveillance
129
characteristics found to date in HLA class I expression in tumors derived from different tissues.
1. BLADDER CARCINOMAS Transitional cell carcinoma (TCC) of the bladder is the second most common malignancy of the urinary tract. As the bladder undergoes malignant transformation, several phenotypic changes can be detected on the cell surface. Among these, the loss of normally expressed MHC class I molecules (Fig. 1) has been found to be important since survival of the patient correlated with grade, stage, and the HLA class I expression of the tumors (Levin et al., 1991, 1992; Tomita et al., 1990; Klein et al., 1996). The term “superficial bladder carcinoma” encompasses a spectrum of diseases that range from the innocuous Ta grade I tumor to the life-threatening grade III T1 tumor (Niall and Heney, 1992). Grade I tumors are usually papillary and confined to the mucosal surface (Heney et al., 1983). The behavior of grade II tumors is difficult to predict. It has been reported that approximately 40% of these tumors showed invasion of the lamina propia at inicial diagnosis, with a 5-year survival of up to 65%. Carcinoma of the bladder in situ is a treacherous entity. It is defined as a high-grade noninvasive flat cancer confined to the epithelium. However, it is an enigmatic condition with unpredictable behavior and with the potential to invade and metastasize. The definition of grade III TCC is less controversial. This is a high-grade cancer, with marked nuclear pleomorphism, many mitotic figures, and conspicuous cellular atypicality (Heney, 1983). Normal bladder mucosa expresses HLA class I antigens (Daar et al., 1984, Levin et al., 1991; Witjes et al., 1995). The expression of β 2 microglobulin and heavy chain (Tomita et al., 1990) in TCC seems to correlate directly with both grade and stage (Amirghofran et al., 1997). Moderate (grade II) and high-grade (grade III) tumors showed losses (32%) or reductions in class I antigens to a much greater extent than did low-grade (grade I) tumors when tumor cells where stained with W6/32 mAb (Tomita et al., 1990). A study of 68 tumor sections by Nouri et al. (1994) showed that 42% of the cases had reduced or absent expression of HLA molecules. The use of polymorphic antibodies for HLA-A2, HLA-A3, Bw4, and Bw6 increased this frequency of defects to 73%. A similar frequency of HLA negativity (75%) was found in metastasis (Cordon-Cardo et al., 1991). Other studies using malignant (grade III) human urothelial cell lines reported low or absent expression of HLA-B molecules compared with premalignant grade II or grade I tumor cells (Ottensen and Kieler, 1991). Our laboratory recently analyzed a series of bladder tumor samples and found a high frequency of HLA class I antigen loss (85%). Phenotype No. I (HLA class I total loss) was found in 45% of the samples and phenotype No. IV (HLA allele loss) in approximately 25% (T. Cabrera,
130
Federico Garrido and Ignacio Algarra
J. M. Lopez-Cozar, and F. Garrido, unpublished results). Very little is known about the molecular mechanism responsible for the altered HLA phenotypes found in bladder carcinomas.
2. BREAST CARCINOMAS Among breast tumors, ductal adenocarcinoma is the most common histological presentation of malignant breast cancer (80%). This proliferation of epithelial cells from galactophoral ducts may be preceeded by in situ carcinoma, characterized by the proliferation of cells within the ducts without interruption of the basal membrane. Lobular carcinoma is the second most frequent (5–10%). The third type of tumor (medullary carcinoma) is a rare entity (1%) with a better prognosis than was previously believed (Eisen and Weber, 1998). Studies at the molecular level have demonstrated many chromosomal aberrations in breast cancer. Among these, LOH in several chromosomes is seen in almost all breast tumors (Mertens et al., 1997). Normal epithelia and low-risk proliferative lesions of the breast tissue are HLA class I positive (Garrido et al., 1993) (Fig. 1). Previous studies by different groups, including ours, have demonstrated that HLA class I antigens are downregulated in tumor tissues of patients diagnosed as having breast car¨ cinomas (Fleming et al., 1981; P´erez et al., 1986; Natali et al., 1983; Moller et al., 1989; Wintzer et al., 1990; Concha et al., 1991a,b; Maiorana et al., 1995). In these cases a maximum of 40–50% HLA class I dowregulation was reported. These studies were performed using cryopreserved tumor sections and anti-HLA antibodies that define HLA monomorphic and locus-specific determinants. However, the introduction of anti-HLA mAbs that define HLA alleles and that work with tissue sections has made a more detailed analyses of the phenomenon possible. Data from our laboratory have shown that the frequency of total or selective loss of HLA class I antigens in patients diagnosed as having breast cancer is 88.5%, which breaks down as follows: HLA-A, -B, and -C, 52%; HLA-A, 4%; HLA-B, 8%; HLA-A and -B, 9%; and HLA allelic loses, 15% (Cabrera et al., 1996; Algarra et al., 2000). In some breast tumor cell lines we have also observed that the level of HLA expression is under steroid hormone control (Rodriguez et al., 1994). A concordant downregulation of HLA class I antigen and TAP1 and TAP2 staining was observed in 22% of 37 high-grade breast carcinoma lesions (Vitale et al., 1998), and complete HLA loss of HLA class I antigens TAP1 and TAP2 was seen in 8% of the cases (Vitale et al., 1998). Defective expression of class I genes, TAP, and β 2 microglobulin was simultaneously found in primary and metastatic breast carcinomas (Kaklamanis et al., 1995). Forty-four percent of lymph node metastases have complete class I loss. These results suggest the existence of common transacting regulatory mechanisms for HLA and β 2 microglobulin genes.
MHC Antigens and Tumor Escape from Immune Surveillance
131
The high incidence of HLA class I loss in breast cancer patients shows that adjuvant immunotherapy to induce HLA class I expression may be of value in a subgroup of patients with reversible HLA class I defects.
3. CERVIX CARCINOMAS Carcinoma of the cervix usually arises in the transitional zone between squamous and columnar cell epithelia. Approximately 80% of all malignant tumors of the cervix are squamous cell carcinomas, 10% are adenocarcinomas, and 10% are adenoacanthoma (Atkin, 1997). Three grades of premalignant lesion, cervical intraephitelial neoplasia (CIN) are recognized (I–III). Carcinomas are staged as follows: IA; early invasive, not grossly visible; IB, IIA, IIB, and IIIA; confined to the cervix; and IIIB, IVA, and IVB; involvement of the pelvis, bladder, and distant metastase, respectively. A crucial event in the malignant progression of CIN appears to be the upregulation of high-risk human papillomavirus (HPV) early gene expression (Stern, 1996). Types 16 and 18 are associated with approximately 70% of all cervical carcinomas. These HPV high-risk types are invariably detected in the moderate and severe stages of preinvasive malignancy (CIN II and III). Steroid hormones have also been linked to the downregulation of HLA in cervical tumor cells with integrated HPV sequences (Bartholomew et al., 1997). Tumors of the cervix are among the most thoroughly analyzed types for HLA expression (Torres et al., 1993; Keating et al., 1995; Brady et al., 2000; Koopman et al., 2000). Cervical carcinoma is associated with infection by different types of human papilloma virus in more than 99% of cases (Walboomers et al., 1999), and the well-established multistep process of carcinogenesis in the natural history of tumor development has provided well-classified material for the study of MHC alterations. Normal cervix epithelium and premalignant lesions are HLA class I positive (Garrido et al., 1993) (Fig. 1) and show no alteration with monomorphic or locus- or allele-specific anti-HLA mAbs (Hilders et al., 1994). On the other hand, the frequency of loss of HLA class I expression measured immunohistochemically in cervical carcinoma is very high (Connor and Stern, 1990; Keating et al., 1995; Garrido et al., 1997a). Phenotypes observed range from complete absence of all HLA-A, -B and -C alleles and β2 microglobulin to loss of expression of a single allele (Keating et al., 1995). The frequency of altered HLA class I phenotypes analyzed by monomorphic or allele-specific antibodies is as follows: HLA-A -B, and -C 18%; HLA-A, 3%; HLA-B, 19%; HLA-A and -B, 2%; and HLA allelic loses, 21% (Garrido et al., 1997a). It has recently been reported that more than 96% of all cervical carcinomas show some HLA class I alteration in the tumor cells, and almost 70% are caused by
132
Federico Garrido and Ignacio Algarra
multiple genetic alterations at chromosome 6p21.3, including genomic deletions and mutations in class I genes (Koopman et al., 1998). Four altered pheno/genotype categories have been described (Koopman et al., 2000): 1. Total HLA class I loss (10%) and retention of heterozygosity (ROH) at 6p21.3: This category includes HLA losses caused by β 2 microglobulin gene mutations. 2. HLA haplotype loss (50%): In these cases, HLA losses are caused by LOH at chromosome 6p21.3 (20–30%) and represent a common mechanism by which HLA genes and their products are abolished (Kersemaekers et al., 1998; Mazunrenko et al., 1999; Ramal et al., 2000a). 3. B locus or HLA-A and -B downregulation associated with ROH and/or allelic imbalance at 6p21.3 (10%). 4. HLA-A or -B allelic loss (17%), mostly associated with gene mutations. HLA-B antigens have been found to be more frequently downregulated than HLA-A antigens (Brady et al., 2000), suggesting that the molecular basis of this particular alteration in HLA-A and -B locus expression may be different. The relationship between the presence of high-risk HPV in cervical lesions and HLA class I downregulation has not been clearly stablished, but recent reports indicate that integration of high-risk HPV 16 and 18 sequences in cervix tumor DNA is linked to the downregulation of HLA class I antigens by steroid hormones (Bartholomew et al., 1997).
4. COLORECTAL CARCINOMAS Colorectal cancer is a commonly diagnosed disease in both men and women. It represents a broad spectrum of neoplasms, ranging from benign growths to invasive cancer. Pathologists have classified the lesions into three groups: nonneoplastic polyps, neoplastic polyps (adenomatous polyps and adenomas), and cancers. More than 95% of all colorectal cancers are carcinomas, and about 95% of these are adenocarcinomas. The transition from normal epithelium to adenoma and carcinoma is associated with the acquisition of different but cumulative molecular events. At least five to seven major molecular alterations need to occur for a normal epithelial cell to progress in a clonal fashion to carcinoma. Some key changes include loss of chromosomes 5q, 17p, and 18q and mutation of the K-ras oncogene (Fearon and Volgestein, 1990). Normal epithelia as well as benign and premalignant lesions of the colon are HLA class I positive. Although other genetic and morphological changes occur at this stage of tumor development, the expression of HLA class I antigens is apparently not altered (Garrido et al., 1993). In some cases, expression is stronger than in normal epithelia. For instance, colon adenomas
MHC Antigens and Tumor Escape from Immune Surveillance
133
that are known to accumulate several activated protooncogens such as K-ras are strongly HLA class I-positive (Gutierrez et al., 1987) (Fig. 1). Previous studies of anti-HLA mAbs directed against monomorphic HLA ¨ determinats revealed HLA class I losses of 30 –40% of the tumors (Moller et al., 1991). However, when a broad panel of mAbs defining monomorphic, locus-specific, and allele-specific determinants was used, HLA losses were found in 73% of the cases (Cabrera et al., 1998). These HLA alterations were classified as follows: total HLA loss, 18%; HLA-A locus-specific loss, 9%; HLA-B locus-specific loss, 8%; HLA-A and -B locus losses, 2%; and HLA allelic losses, 36%. Despite the high frequency of altered HLA expression detected in these tumors, it was not possible to study all the HLA alleles. This means that the percentage of HLA losses is still underestimated. In recent years there have been considerable efforts to define, at the molecular level, the mechanisms by which HLA expression in colorectal tumors is downregulated or lost. Complete loss of expression of HLA class I antigens in colorectal tumors may be associated with a lack of β 2 microglobulin synthesis or with the synthesis of truncated β 2 microglobulin. Any genetic or posttranslational event that impairs β 2 microglobulin production results in failure to form peptide heavy chain-β 2 microglobulin complexes on the cell surface (Browning et al., 1996; Cabrera et al., 1998). This is the case for the analysis of the β 2 microglobulin gene in a series of colorectal tumors in which the absence of HLA class I expression was associated with mutations in these genes (Browning et al., 1996; Bicknell et al., 1994). We recently analyzed the β 2 microglobulin gene in 31 cases of tumors with HLA class I total loss selected from 162 tumor samples. We found that colorectal carcinomas with HLA total loss had no β 2 microglobulin mutations; this result indicated that other mechanisms might be involved in the generation of this ´ HLA class I total loss phenotype (Fernandez et al., 2000). We also analyzed microsatellite instability in HLA-A, -B, and -C-negative colorectal tumors and found no relationship between a replication error phenotype (RER+) and HLA class I alteration. Only one tumor out of 15 HLA class I-negative colorectal carcinomas showed a RER+ phenotype (Jimenez et al., 2000). Other groups, however, found colorectal tumor cell lines which failed to express β 2 microglobulin but did exhibit microsatellite instability (Branch et al., 1995). These HLA genetic alterations belong to the first category of HLA class I altered phenotypes (phenotype No. I, HLA total loss). Colorectal carcinomas with selective HLA alterations have also been reported: The LS411 colorectal cell line shows a chromosomal breakpoint in the HLA-A11 allele (HLA allele loss, phenotype No. IV) (Browning et al., 1993); the C08, C012, and C067 tumors display concordant HLA altered phenotypes compatible with HLA haplotype loss (phenotype No. II) (Ramal et al., 2000b) and HLAA,B locus-specific downregulation (phenotype No. III) (Smith et al., 1988; ´ Lopez-Nevot et al., 1989).
134
Federico Garrido and Ignacio Algarra
5. HEAD AND NECK CARCINOMAS Head and neck cancer is the sixth most common form of cancer worldwide (Vokes et al., 1993). More than 90% of head and neck cancers are squamous cell carcinomas. The oral cavity is the most common site within the head and neck, and the larynx is the second most common site. Since Broders initial classification, many histopathological criteria have been introduced as differentiation and prognostic parameters (Pera et al., 1986). Indeed, a strong correlation between the degree of differentiation and ´ class I antigen expression has been found in laryngeal carcinomas (LopezNevot et al., 1989). HLA class I total loss and tumor aggressiveness have also been correlated in laryngeal carcinomas (Esteban et al., 1989, 1990a). Recently, it has been suggested that the loss of expression of HLA class I alleles may have prognostic implications (Grandis et al., 2000). Head and neck squamous cell carcinomas are derived from HLA class I– positive epithelia. Benign (squamous metaplasic epithelium and squamous papilloma) and in situ carcinomas are HLA class I positive (Garrido et al., 1993) (Fig. 1). However, many HLA altered phenotypes can be found in invasive carcinomas (Garrido et al., 1997a). The frequency (%) and distribution of the losses is as follows: HLA total loss, 9%; HLA-A loss, 19%; HLA-B loss, 16%; HLA-A+ -B loss, 9%; and HLA allelic losses, 26%. Seventy-nine percent of the laryngeal tumors present an HLA class I altered phenotype (Cabrera et al., 2000). Concerning the molecular basis of the defect in HLA class I expression in these tumors, we previously reported that transcriptional regulation of HLA expression is likely to be involved (Esteban et al., 1989). A recent study has shown that downregulation of HLA class I expression in head and neck squamous cell carcinomas is correlated with loss of chromosomal region 6p21.3 (including the HLA complex) (Feenstra et al., 1999a). Data show that LOH for 6p21.3 is a frequent event, occurring in 49% of the tumors with downregulated HLA class I expression. In addition, it has been reported by two independent groups that β 2 microglobulin mutations are not involved in the generation of HLA class I total loss (phenotype No. I) ´ in laryngeal carcinomas (Feenstra et al., 1999b; Fernandez et al., 2000). It is therefore necessary to further characterize the molecular basis for HLA class I total or allelic loses detected in head and neck carcinomas.
6. LUNG CANCER Carcinoma of the lung is the most common cancer in males. The subdivision of bronchogenic carcinoma into different clinicopathological entities, such as small cell carcinoma of aggressive behavior and non-small cell carcinoma of slower evolution, is well established and of practical use. The available evidence suggests that all the major subtypes of bronchogenic
MHC Antigens and Tumor Escape from Immune Surveillance
135
carcinoma arise in the basal cells of the bronchial epithelium, which is of endodermal origin. During the early phases of neoplastic development the malignant cells tend to differentiate as squamous cells, glandular cells, large undifferentiated epithelial cells, or small cells (Jeffery and Reid, 1977). The prognosis in patients with lung cancer is determined by several factors, the most important of which are the specific histological diagnosis, tumor stage, and host performance status (Carbone, 1997). We have shown that the loss of HLA class I antigen in lung cancer bears a significant relationship with two markers of biological aggressiveness—the degree of differentiation and the presence of aneuploidy (Redondo et al., 1991a). Normal lung tissue (pneumocytes and epithelial respiratory cells) expresses HLA class I antigens (Redondo et al., 1991b). However, HLA class I antigen expression is frequently modified in lung tumors. HLA class I total loss is the most frequent phenotype in this type of tumor. The loss of HLA-A, -B, and -C molecules is found in 38% of all lung tumors and is usually accompanied by loss of β 2 microglobulin and heavy-chain A locus (Korkolopoulou et al., 1996; Redondo et al., 1997). Selective loss of A and B locuses has also been found in bronchogenic carcinomas (Redondo et al., 1991a; Korkolopoulou et al., 1996). The available data indicate a frequency of 8% for selective HLA-A locus loss, although this percentage would probably be higher in a more thorough analysis. This was implied by a study of HLA class I allelic loss: In a series of 93 specimens of non-small cell lung carcinomas, HLA-A2 allele had been lost in 27% of the cases (Korkolopoulou et al., 1996). The molecular mechanisms for these alterations have not been fully elucidated, but data from different laboratories suggest that mutations in the β 2 microglobulin gene (Chen et al., 1996a) and in TAP1 (Chen et al., 1996b) are mechanisms responsible for HLA alterations, as is the case for lung carcinoma cell lines H2009 and H1436, respectively. Loss of TAP1 is also involved in the abnormal HLA class I expression in other lung carcinoma tissues and cell lines (Korkolopoulou et al., 1996; Singal et al., 1998).
7. MELANOMA Melanoma is the least common cancer among the main types of skin cancer (basal cell carcinoma and squamous cell carcinoma), but it accounts for three-fourths of all deaths from skin cancer. Currently, it is the most rapidly increasing form of cancer, and recent statistics have shown a doubling in the number of new cases during the past 15 years. The most important risk factor that has been identified is exposure to sunlight, especially during childhood. There are two distinct phases of growth in malignant melanoma. During the first noninvasive phase the neoplastic cells spread horizontally without breaking the basal membrane. This phase may last 2–5 years. The second phase is characterized by rapid vertical growth, leading to the invasion of the dermis (McGovern et al., 1979). Clark’s levels establish the degree of
136
Federico Garrido and Ignacio Algarra
dermal invasion in melanoma, and Breslow’s thickness quantifies the local growth of the tumor (Clark et al., 1969). It has been proposed that many immunological events are involved in the pathogenesis of malignant melanoma, and that abnormalities in HLA class I antigen expression by melanoma cells may have adverse effects on the clinical course of the disease. Many laboratories have identified genes expressed within melanoma cells that code for melanoma-associated tumor antigens (MAGEs) (Van der Bruggen et al., 1991). These genes encode proteins that contribute a peptide tumor antigen presented by the HLA products to cytotoxic T lymphocytes. HLA class I antigen expression has been demonstrated in normal melanocytes by electron microscopy (Van Duinen et al., 1984). Benign nevi are class I negative, whereas dysplastic nevi are class I positive (Ruiter et al., 1982). HLA class I losses in malignant melanoma have been shown in many tumors, both metastastic tissues and cultured tumor cell lines. Most of the initial studies of HLA and melanoma were done with mAbs that recognized nonpolymorphic determinants of the HLA-A, -B, and -C molecules. The differences observed in the data reported (Ruiter et al., 1982; D’Alessandro ´ et al., 1987; Lopez-Nevot et al., 1986; Holzmann et al., 1987; Ernstoff et al., 1985) can be explained by the different sensitivities of the immunohistochemical staining procedures used, the different nature of the lesions tested, differences in antibody specificities, and differences in the criteria used to define a positive or negative lesion. In these studies HLA class I losses were observed in both primary and metastatic lesions (Holzmann et al., 1987). The phenotypic alterations were related to histopathological malignancy cri´ teria and tumor progression (Lopez-Nevot et al., 1988; Ruiter et al., 1991). Using a large battery of mAbs (locus or allele specific), which define HLA losses more accurately, studies of melanoma tumors (Marincola et al., 1994; Garrido et al., 1997b) have shown that approximately 63% of all melanoma lesions present HLA phenotypic alterations with the following frequency: HLA total loss, 16%; HLA haplotype loss, 14%; HLA locus loss, 8%; and HLA allelic losses, 25%. In this context, a high frequency of allelespecific downregulation of HLA class I expression has been reported in uveal melanoma cell lines (Hurks et al., 2000). Several molecular mechanisms underlie the HLA phenotypes found in melanoma tumors. The most common molecular alteration responsible for total HLA class I loss in melanoma cells is a mutation in the β 2 microglobulin gene, which results in loss of functional β 2 microglobulin expression (Table II) (Hicklin et al., 1998; D’Urso et al., 1991; Wang et al., 1993, 1996). Our laboratory recently described β 2 microglobulin mutations in a melanoma tumor cell line (GR34) (P´erez et al., 1999) and in melanoma tissues of two patients immunized with MAGE peptides (BB74-MEL and LB1622-MEL) (Benitez et al., 1998). These mutations are the reason why the tumors of both patients progressed despite peptide therapy.
MHC Antigens and Tumor Escape from Immune Surveillance
137
Loss of heterozigocity at 6p21 has been described as one of the most frequent karyotypic abnormalities that appear in human malignant melanoma (Real et al., 1998; Mendez et al., 2001). We identified a new HLA class I altered phenotype that is the result of a combination of HLA-B and -C locus downregulation and HLA haplotype loss. The alteration was found in two melanoma cell lines generated from two patients; one derived from an in vivo lesion (FM37 cell line) and the other was obtained after in vitro immunoselection (R22.2 cell line) (Real et al., 1998). The final result was a melanoma tumor cell that expressed a single HLA class I allele. An example that adds further complexity to the pattern of HLA altered phenotypes in melanoma tumors is represented by human melanoma cell line FO-1 (Martayan et al., 1999). A minimum of three defects hinder class I expression in FO-1 cells: the lack of β 2 microglobulin, a low expression of TAP, and poor class I–calreticulin interaction. All these defects can impair normal cell surface recognition by cytotoxic T cells (Hicklin et al., 1999). All the altered HLA class I phenotypes described in this review have been identified in melanomas (Real et al., 1998; Benitez et al., 1998).
8. PANCREAS CANCER Cancer of the exocrine pancreas (referred to here as pancreas cancer) has a very poor prognosis in humans (Warshaw and Fernandez del Castillo, 1992). Difficulties in the diagnosis, late detection after the tumor has already affected neighboring or distant organs, and the lack of effective therapy are responsible for the low 1-year survival rate (<5%). This aggressive clinical behavior resulted in the hypothesis that loss of MHC class I expression might contribute to the biological and clinical behavior of this tumor. Our group found evidence that altered MHC class I expression is frequent in pancreas cancer (Torres et al., 1996). The frequency of HLA class I losses according to analyses with monomorphic and locus-specific antibodies is as follows: HLA total loss, 12%; HLA-A loss, 19%; HLA-B loss, 8%. Using a combination of immunohistochemical, biochemical, and recombinant DNA approaches, we determined the prevalence and molecular basis of altered HLA class I expression in a panel of pancreas cancer tissues and cell lines. The results obtained in one solid tumor tissue and its corresponding pancreatic tumor cell line (IMIM-PC2) indicate that LOH was responsible for the selective loss of MHC class I alleles in this tumor (Torres et al., 1996).
9. PROSTATE CANCER Prostate cancer is the second most common cause of death in most developed countries. In general, prostate cancer is a disease which is rarely seen in men in their 30s and 40s, but it increases in frequency after the age of 55. In fact, about 80% of all cases occur in men who are 65 years of age or older.
138
Federico Garrido and Ignacio Algarra
Studies have revealed differences in the incidence of prostate cancer and suggest that the development of the disease is probably influenced by genetic factors. One of the factors that may influence prostate cancer progression is the expression of human leukocyte antigens in tumor cells (Natali et al., 1989; Levin et al., 1994). HLA class I and II expression have been studied in benign (benign prostatic hyperplasia), malignant (prostatic adenocarcinoma), and metastatic prostate disease to define the extent of altered HLA expression and to determine whether HLA expression is related to disease progression (Blades et al., 1995; Garrido et al., 1997a; Lu et al., 2000). These studies used mAbs which recognized both monomorphic determinants and HLA allelic products. In contrast to the normal HLA class I expression of the benign tissue and benign hyperplasias (Sharpe et al., 1994), complete loss of HLA class I expression was observed in 34% of primary prostate cancers and 80% of lymph node metastases (Blades et al., 1995). The frequency of HLA allelic loss in primary and metastatic prostate cancer has been estimated to be 51% (Garrido et al., 1997a; Natali et al., 1989). Some tumor cell lines have been used in the search for the mechanisms that underlie these alterations in MHC class I expression in prostate tumors. Among these (Sanda et al., 1995), the metastasis-derived human prostate cancer cell line PPC-1 shows a regulatory defect in the antigen transport machinery that leads to specific underexpression of the TAP-2 gene product.
10. RENAL CELL CARCINOMA Renal cell carcinoma (RCC) is the most common malignancy of the kidney and accounts for approximately 3% of all adult malignancies. The most frequent presentation of RCC is a solitary renal mass that arises in the renal cortex. It may invade the renal vein and even extend into the inferior vena cava. Most analyses of HLA class I antigen expression in these tumors have been performed using inmunohistochemical techniques with mAbs that recognize a monomorphic determinant of HLA class I antigens (W6/32). A decrease in or loss of HLA class I expression was demonstrated in about ´ 38% of the tumors (Cordon-Cardo et al., 1991). Others studies reported lower frequencies of HLA class I losses in primary tumors, with losses in up to 50% of the metastases (Buszello and Ackermann, 1994). This frequency of HLA class I losses in primary RCCs has been correlated with the stage and tumor diameter (Brasanac et al., 1999). Indeed, selective loss of HLA expression has been observed to occur predominantly in advanced-stage disease. One-dimensional isoelectric focusing and Western blot analysis identified the selective (HLA-A1) or combined (HLA-A2 and HLA-B38) antigen variations in HLA class I expression in 5 of 23 advanced-stage renal tumors (Luboldt et al., 1996).
MHC Antigens and Tumor Escape from Immune Surveillance
139
The mechanisms that underlie the reduced expression of HLA class I antigens in RCCs appear to be associated in most cases with deficiencies in the expression and function of different components of the MHC class I antigen processing pathway and poor recognition by cytotoxic T lymphocytes (Kallfelz et al., 1999; Seliger et al., 1997). Several RCC cell lines (MZ1851, MZ1879, and MZ1940) show strong reduction in TAP and LMP molecules (Seliger et al., 1997). This reduction was even more pronounced in a metastatic cell line derived from the MZ1851 primary tumor (Seliger et al., 1996). This type of molecular alteration is also present in the Caki2 RCC line (Dovhey et al., 2000). This tumor cell line appears to be defective for the IFN-γ signaling pathway that regulates TAP1 and LMP2 expression.
III. CHANGES IN MHC CLASS I ANTIGEN EXPRESSION DURING METASTATIC COLONIZATION Few studies have carefully analyzed and compared in-depth the MHC class I changes in tumor cells during tumor development. In human tumors only a single long-term follow-up of one melanoma patient has been reported (Lehmann et al., 1995; Coulie et al., 1999). This patient showed different patterns of HLA expression in response to T cell-mediated immunity, including HLA haplotype loss (phenotype No. II) and, later, HLA-B and -C locus dowregulation (phenotype No. III). The final result was a melanoma cell that expressed a single HLA class I antigen (phenotype No. V)—HLA A24. These phenotypic HLA alterations paralleled different T cell responses directed against tumor antigens that are presented by different HLA alleles. The cancer cell changes its HLA phenotype to escape, and the host changes the T cell repertoire to control tumor growth (Ruiz-Cabello and Garrido, 1998). This dynamic process is not easy to see when a single tumor sample is taken from a patient. Nevertheless, we investigated changes in HLA class I expression in primary tumors versus autologous metastases in a variety of ´ tissue samples, including laryngeal, colon, and gastric tumors (Lopez-Nevot et al., 1986). These changes included several possibilities: HLA-positive primary tumor with HLA-negative metastases (+/−), positive tumor/positive metastases (+/+), negative primary tumor/positive metastasis (−/+), and negative primary and negative metastases (−/−). These data were obtained using monomorphic anti-HLA mAbs and therefore do not indicate which alleles are downregulated or upregulated. However, they do indicate that a decrease in HLA class I expression during metastatic spread may not always occur (Cromme et al., 1994). We analyzed several examples of melanoma metastases that presented HLA class I total loss (Benitez et al., 1998) or HLA haplotype loss (Real et al., 1998;
140
Federico Garrido and Ignacio Algarra
Mendez et al., 2001). These lesions were highly homogeneous, and in some cases different metastases from the same patient presented the same HLA alteration (Mendez et al., 2001). The following question must be addressed: Are the changes observed in MHC class I expression during tumor development randomly produced or do they follow a predictable pattern in individual cancer types and patients? We analyzed the MHC class I expression of metastatic colonies produced in immunocompetent BALB/c mice using an H-2 negative primary tumor clone (B9) derived from a syngeneic fibrosarcoma produced and extensively analyzed in our laboratory (Garrido et al., 1986; P´erez et al., 1990; Algarra et al., 1989, 1991). This clone was injected into the footpad in a standard
Fig. 4 New altered MHC class I phenotypes are found in metastatic colonies. These phenotypes derive from clonal populations present in the primary tumor. The changes observed in MHC class I expression in metastases are not random but can be reproduced in syngeneic mouse tumor systems (Garcia Lora et al., 2001).
MHC Antigens and Tumor Escape from Immune Surveillance
141
spontaneous metastasis assay to investigate whether the metastatic colonies presented the same H-2 class I expression as the primary tumor clone. The data indicated that 83% of the metastases obtained in different syngeneic BALB/c mice repeatedly exhibited a phenotype different from that of the original B9 clone, with no induction of the H-2Ld antigen after IFN-γ treatment (Garcia Lora et al., 2001). These metastatic colonies therefore exhibited qualitative changes in MHC antigens compared with the B9 primary tumor clone, and we concluded that these alterations were not random but identical in different colonies obtained in different syngeneic BALB/c mice. A microdeletion present near the Ld gene was responsible for the absence of expression of this molecule. The rest of the metastases (17%) presented the same phenotype as that of the primary tumor clone. We concluded that each metastases does indeed follow a particular pattern of HLA mutation in each individual. These findings are the first indication that the changes in MHC class I profiles during metastatic colonization are not random but can be reproduced in different syngeneic animals (Fig. 4). It will be of interest to analyze the HLA expression in multiple metastases obtained from the same patient to determine if similar, identical, or different alterations in HLA phenotypes are found. We recently obtained data showing that identical HLA class I phenotypes appear in different metastatic colonies in melanoma patients (Mendez et al., 2000, 2001). However, other authors have found different HLA alterations in different autologous metastases (Rebmann et al., 2000).
IV. T CELL IMMUNOSELECTION OF MHC CLASS I–NEGATIVE TUMOR CLONES It has been proposed that the major force contributing to the appearence of MHC class I–negative tumor clones is T cell immunoselection (Garrido ¨ et al., 1986; Tomlinson and Bodmer, 1999; Jager et al., 1997). This hypothesis implies that T cells can recognize tumor antigens presented by HLA class I–positive tumor cells and thus perform effective immune surveillance; however, when HLA class I-defective tumor variants appear, T cells cannot recognize these targets and these tumor clones acquire a growth advantage that allows them to take over the other clonal tumor populations (Fig. 5, see color insert). Nevertheless, the origin of these MHC-negative tumor variants is unknown, and there is no reported experimental or clinical data to substantiate this hypothesis. Hence, this issue is currently the subject of debate. Our group recently obtained data indicating that an H-2 class I–negative fibrosarcoma tumor clone generated MHC class I-negative spontaneous lung metastases in inmunocompetent syngeneic BALB/c mice. In contrast, the same tumor clone produced MHC class I–positive metastatic nodes in athymic nude/nude mice (Fig. 6, see color insert). (Garcia Lora et al., 2001).
142
Federico Garrido and Ignacio Algarra
This phenomenon was observed in metastatic nodules generated after a period of in vivo growth but not in the primary tumors growing locally in the footpad. These findings support the hypothesis that the MHC phenotype of metastatic nodes is influenced by the T cell repertoire of the host since in the absence of this T cell pressure, the metastatic nodes “recovered” H-2 class I expression. Figure 6 summarizes these experiments and shows the MHC class I phenotypes found in metastatic nodes arising in T cellimmunodeficient nude/nude and in immunocompetent BALB/c mice. Similarly, it can be predicted in humans that tumors arising in immunodeficient patients, such as HIV-infected individuals or kidney transplant recipients who have received long-term treatment with immunosuppressive drugs, will not require the production of HLA-deficient clones to escape T cell responses since T cell function is markedly decreased. In this context, it was recently described that chemically induced sarcomas produced in nude mice were more immunogenic than similar sarcomas induced in congenic immunocompetent mice (Engel et al., 1997). Tumors derived from nude mice were transplanted and rejected significantly in normal mice, indicating that the interaction of tumor cells with an intact immune system leads to a progressive low immunogenicity of tumor cells. Another prediction from this hypothesis is that the number of HLA class I–negative cells present in a particular tumor will increase as the tumor develops, and therefore the presence of homogeneous HLA class I–negative tumors implies a prolonged period of growth. In contrast, tumor tissue samples with small numbers of HLAdeficient cells or heterogeneous populations suggest an early diagnosis and a short period of evolution (Garrido et al., 1993) (Fig. 5). It is important to remember that HLA class I downregulation is not the only mechanism of tumor escape to avoid T cell responses. Other mechanisms, ¨ such as downregulation of the tumor antigens (Jager et al., 1997), downregulation of the antigen processing machinery (Cromme et al., 1994), alteration of the apoptosis program (Hahne et al., 1996), expression of inhibitory cytokines (Chouaib et al., 1997), and immunological ignorance (Ochsenbein et al., 1999), have also been described. During tumor progression each step may involve activation, mutation, or loss of different genes. New cell variants arise and those with growth advantage over earlier forms are selected. Whether an antigenic loss is advantageous to the tumor depends on the particular tumor antigen. For instance, human solid tumors appear to conserve mutant oncoproteins throughout their course. It is very likely that the expression of tumor antigens derived from such oncoproteins in tumor tissue is homogeneous because tumor cells require the continuous presence of these proteins to maintain their transformed phenotype. In fact, only tumor cells that have acquired mutant copies of the protein become malignant. The clearest examples of these antigens are the oncogenic DNA viruses which carry their own transforming genes. Therefore, antigen loss variants occur less frequently in these neoplasms. However, tumor antigens derived from many
MHC Antigens and Tumor Escape from Immune Surveillance
143
other proteins (tumor-specific shared antigens, differentiation antigens, etc.) that are not intrinsically involved in cell transformation mechanisms can be lost in the course of tumor progression. Some of these antigens have been used in clinical trials of peptides-derived vaccines (MAGE, tyrosinase, gp100, and Melan A/MART-1), and the selection of antigen loss variants is often encountered in nonresponsive patients, even in the presence of antigen-specific ¨ CTLs (Jager et al., 1997; Thurner et al., 1999). Many of the relevant proteins involved in neoplastic transformation and tumor growth are intracellular mutant oncoproteins that can generate individual tumor antigens recognized by T cells. Peptides derived from the mutant intracellular protein can be presented by class I MHC. It is also possible that mutations occurring during tumor progression may be located outside the consensus binding motif for a given MHC haplotype. Cancer cells might develop and grow unrestrained by the immune system only when these mutations occur outside an MHC binding motif. This is expected to occur only as a process of selection during the development of cancer, and it is possible that the immune system selects those oncoprotein-derived mutated peptides that cannot be presented by the host’s MHC molecules. This seems to be the case for p53 mutations in lung cancer. Wiedenfield et al. (1994) found that p53 mutations fell within the HLA A0201 motif less often than expected, probably because of the selection of tumor clones with mutant forms of p53 that do not bind to MHC. A similar finding was reported by Nistico et al. (1997) for breast cancer patients whose tumors overexpressed ErbB-2 molecules. T lymphocyte antitumor immune responses play a direct role in the generation of HLA and antigen loss variants. Direct evidence in support of this hypothesis derives from our work with an experimental mouse fibrosarcoma (Garcia Lora et al., 2001) and suggests that the search for clinical data to substantiate this hypothesis may well yield important insights into the natural history of tumor MHC phenotypes. It is not easy to reconcile the benefits of an efficient T cell antitumor immune response with the selection of escaped MHC class I loss variants by the same antitumor T lymphocytes (Fig. 5). This concept is under investigation (Breivik and Gaudernack, 1999; Tomlinson and Bodmer, 1999), and the Darwinian view of cancer development may point to new ways to manipulate the tumor–host relationship.
V. EXPRESSION OF NONCLASSICAL HLA CLASS I MOLECULES IN TUMORS The nonclassical HLA class I molecules, such as HLA-G, -E, and -F, are also coded by HLA genes in the HLA complex. Like their classical counterparts, they are composed of a heavy chain noncovalently linked to β 2 microglobulin
144
Federico Garrido and Ignacio Algarra
(Geraghty, 1993). Important characteristics of these molecules in comparison with the classical HLA-A, -B, and -C elements is their high degree of homology with classical HLA I molecules but their low degree of polymorphism. Only a few alleles have been described for HLA-G, and very little is known about HLA-E and -F. It is well established that HLA-G is selectively expressed in the extravillous trophoblast of the human placenta (Le Boutellier, 1997), in which HLA-A and -B molecules are not present. This peculiar tissue distribution of HLA class I molecules in the trophoblast (HLA-G and -C) is thought to reflect its role in protecting the fetus from attack by the mother’s NK cells. It was recently reported that HLA-G can interact with NK inhibitory receptors KIR2DL4 and LIR (ILT2) (Ugolini and Vivier, 2000), and it is known that HLA-C is a potent inhibitor of NK cells (Moretta et al., 1996). It is tempting to speculate that the aberrant expression of HLA-G in HLA-A and -B–negative tumors could represent a way to escape NK immunosurveillance (Paul et al., 1998). Paul et al. analyzed a small number of melanoma tissue sections with an immunofluorescence technique and antibodies that were assumed to react specifically with HLA-G. However, careful analysis of a broad panel of tumor tissues and tumor cell lines, and the use of specific anti-HLA-G mAbs in immunohistologic and flow cytometric techniques, did not confirm these findings (Real et al., 1999). Recently, an independent research group also obtained similar results—namely, that HLA-G is not expressed at the cell surface in human melanomas and melanoma cell lines (Frumento et al., 2000) but is frequently transcribed and translated (Real et al., 1999). An important clinical application of these findings was recently reported (Wagner et al., 2000). A group of melanoma patients included in a clinical trial of IFN-α-2b treatment showed detectable intracytoplasmic levels of HLA-G in metastatic melanoma lesions with total loss of classical HLA-A, -B, and -C molecules. Patients in this group relapsed, whereas patients whose tumors expressed classical HLA class I molecules but not HLA-G did not relapse. In this connection, several isoforms of HLA-G have been described, including soluble forms. It is therefore intriguing to speculate that any of these HLA-G isoforms may play a role in NK suppressor activity induced by tumor cells. HLA-G seems to be expressed in monocyte-derived and macrophagederived cells. Positive tissue staining has been found in alveolar infiltrated histiocytes and also in cytomegalovirus-infected macrophages (Onno et al., 2000). We have also detected HLA-G in the myelomonocite-derived cell line U937 after IFN-γ treatment (Real et al., 1999). The function that may be reflected by the expression of HLA-G in this cell line is not known. HLA-E is another nonclassical HLA class Ib molecule that also recently attracted the interest of different laboratories, especially when it was
MHC Antigens and Tumor Escape from Immune Surveillance
145
discovered that surface expression of HLA-E depends on the binding of conserved peptides derived from different MHC class Ia molecules (Llano et al. 1998) and that HLA-E was interacting with the lectin-type derived NK inhibitory receptor CD94/NKG2A (Braud et al., 1998). HLA-E cell surface expression was found to be dependent on HLA class Ia expression. These findings have shed new light on a mechanism that NK cells may use to sense the level of expression of classical HLA class I molecules by using the interaction of HLA-E with the CD94/NKG2A inhibitory receptor. The aberrant expression of this HLA class Ib molecule could be used by virus-infected cells and tumor cells to escape NK cell attack. In this connection, data obtained in HIV-1-infected cells showed selective HLA-A and -B downregulation but no significant effect on HLA-C or -E expression. This HLA imbalance may represent a way to escape from specific anti-HIV-1 CTLs (HLA-A and -B loss) while maintaining protection from NK cell killing (HLA-C and -E expression) (Cohen et al., 1999). Recently, it has been reported that the cytomegalovirus gpUL40 carries the same conserved peptide in the leader sequence of the glycoprotein UL40 (gpUL40) as that requires for HLA-E expression. gpUL40 upregulates the expression of HLA-E independently of TAP. This upregulation protects cytomegalovirus-infected cells from being attacked by NK cells and represents a newly described escape route for human cytomegalovirus (Tomasec et al., 2000). HLA-E may play an important role in strategies adopted by viruses to escape T and NK cell immune surveillance. Viral homologs of the peptide that binds to Qa-1 (the murine counterpart of HLA-E) can also be found in databases (Tomasec et al., 2000). The role of HLA-E in tumor immune escape remains unclear. We recently screened many tumor cell lines with defined HLA class Ia alterations for HLA-E expression and found that only those cell lines derived from the myelomonocytic lineage were HLA-E positive (S. Pedrinaci, D. Geraghty, and F. Garrido, unpublished results).
VI. TUMOR NK ESCAPE MECHANISMS Much still remains to be elucidated about the mechanisms of NK escape used by HLA class Ia-deficient tumor targets during tumor development. In theory, the loss of some or all HLA class I molecules during tumor development activates NK cell killing. It is well established that NK cells are constantly inhibited by the interaction of HLA class I antigens expressed by the target cells with a set of inhibitory receptors (KIRs) that maintain inhibitory signals to NK cells (Moretta et al., 1996). Why aren’t MHC-deficient tumor cells destroyed, especially when there is total loss of HLA class I antigens (Garrido et al., 1997a)? There must be a new source of selective
146
Federico Garrido and Ignacio Algarra
pressure to produce NK escape variants. Partial tumor HLA class I losses might not be sufficient to induce NK killing since the remaining HLA class I alleles will interact with NK-KIRs. The important role of HLA-E in NK escape routes in virus-infected cells was discussed previously. In addition, the absence from the tumor cell surface of the ligands that interact with the newly described NK natural cytotoxicity receptors, which trigger NK cell killing, probably helps virus-infected and tumor cells to avoid NK immune surveillance (Moretta et al., 2000). Another possible explanation for NK escape mechanisms was suggested by the discovery of the interaction of the MHC class I–related chain A (MICA) molecules, a distant homolog of MHC class I antigens, with the NK receptor NKG2D expressed in NK and γ /δ T cells (Bauer et al., 1999). The stressinducible MICAs stimulate NK cells by binding to the NKG2D receptor and promoting antitumor NK and T cell responses. The absence of MICA inducibility in tumor targets could also favor an escape strategy in actively growing tumors.
VII. HLA CLASS I LOSS AND T CELL–BASED IMMUNOTHERAPY Cytotoxic T cells require an intact and functionally active HLA class I molecule expressed on the target cell surface for antigen presentation and T cell killing. The HLA class I losses observed during the development and spread of human tumors have not evolved to undermine T cell-based immunotherapy currently being performed in several centers (Marchand et al., 1995, 1999; Rosenberg et al., 1998). Rather, they are a requirement for tumor escape from the host’s immune response. This situation might parallel that observed in the early days of kidney transplantion, when surgeons realized that the immune system was frustrating their work. However, the immune system was simply doing what it had evolved to do (i.e., react against non-self-peptides). Currently, careful immunosuppressive therapy keeps this undesirable immune reaction under control. Similarly, we must learn to cope with the selection of HLA loss tumor variants present in many tumors when T cell–based immunotherapy is proposed for a patient. The coexistence of this immune escape mechanism and the development of strategies to diagnose HLA tumor phenotypic alterations promise to provide clinical benefits in the future. The blind use of peptide vaccination strategies in patients who are assumed to have an HLA class I–deficient tumor lacking the particular HLA allele that carries the correspondent peptide is therefore not only ineffective but also wasteful as a therapeutic approach. It can be compared with prescribing an
MHC Antigens and Tumor Escape from Immune Surveillance
147
antibiotic for an infection produced by a bacteria that is resistant to that particular drug. What does one do then? Some patients seem to benefit from antitumor peptide immunotherapy: Between 30% and 40% show clinical response (Marchand et al., 1995, 1999). This percentage is similar to that obtained with interleukin-2 or interferons in melanoma (Nestle et al., 1998, Wagner et al., 2000). We propose that patients who respond have MHC class I “soft” lesions that can be reversed with cytokine treatment, such as the HLA loss variants that correspond to phenotype No. III. Some patients with melanoma who did not respond to MAGE-1–3 HLA-A1 peptide were found to have β 2 microglobulin mutations and HLA total loss (phenotype No. I) in a careful retrospective analysis of HLA expression (Benitez et al., 1998). A similar study reported loss of functional β 2 microglobulin in patients with metastatic melanoma who were receiving cytokine gene therapy (Restifo et al., 1996). We recently found that HLA class I loss and antigen loss can also coexist in metastatic lesions of patients with melanoma who do not respond to peptide (Gp100 and melant/A-Mart1) immunotherapy (Mendez et al., 2001). There are few immunohistological laboratories that can perform reliable HLA analyses in tumor tissues. The main goal of the XII International Histocompatibility Workshop was to test the reliability of anti-HLA mAbs in different tissues and to standarize the techniques (Garrido et al., 1997b)—an effort that will be continued in the 13th workshop.
VIII. HLA CLASS II ANTIGENS IN TUMORS Under physiological conditions the tissue distribution of HLA class II molecules (DR, DP, and DQ) is restricted. They are expressed on cells of the immune system, including B lymphocytes, antigen-presenting cells (dendritic and Langerhans cells, monocytes, and macrophages), and activated T cells (Janeway et al., 1985; Daar et al., 1984). Although most cells do not normally express HLA class II antigens constitutively, in some pathological and nonpathological circumstances class II-negative cells can be induced to express greatly increased amounts of class II molecules (Basham et al., 1985, Cabrera et al., 1995). We and others have found that cells of the colon, larynx, skin, muscle epithelium, and squamous cells of the cervix do not express HLA-DR antigens (Gutierrez et al., 1987; Esteban et al., 1990b; Glew et al., 1992; Cabrera et al., 1992, 1995; Fernandez et al., 1991). In contrast, weak HLA-DR expression has been observed in stomach, lung, and breast epithelia and in columnar epithelium of the cervix (Ferron et al., 1989; Redondo et al., 1991a; Concha et al., 1995). These epithelia are constantly exposed to enviromental
148
Federico Garrido and Ignacio Algarra
antigens, and they may behave as accessory antigen-presenting cells. It is paradoxical, however, that these epithelia, which are strongly affected by external stimuli, can be either HLA-DR positive (lung and stomach) or HLA-DR negative (colon, larynx, cervix, and skin). In studies of HLA-DR expression in tumors we observed two different situations. In one group, tumors derived from HLA-DR–negative tissue acquired de novo DR expression in the early stages of tumor development (colon, larynx, skin, and cervix). Another group comprised tumors derived from epithelia that expressed DR antigen (although weakly), only a certain percentage of which retained this positivity (stomach, breast, and lung) (Table III). MHC class II molecules are involved in antigen presentation and in the regulation of T cell activation. However, their function when expressed on the surface of epithelial cells is not clear. It is reasonable to think that tumor cells that express MHC class II antigens might present tumor peptides directly to CD4+ T helper cells, a feature thought to be associated in most tumors with a good prognosis (Ostrand-Rosenberg et al., 1996). We have found this to be the case in some tumors that express HLA class II molecules. For instance, of a total of 69 larynx carcinomas, HLA-DR expression was present in only 8 well-differentiated, highly keratinizing squamous cell tumors. These neoplasms had in common a slow rate of growth and were classified as lowgrade carcinomas. The expression of DR locus may be responsible for the excellent prognosis in these neoplasms (Esteban et al., 1990b). Similarly, in
Table III Abnormal HLA-DR Expression in Tumors and Prognosis
Normal tissue HLA-DR
Tumor HLA-DR+ (%)
Cervix Colon
− −
83 36
Larynx
−
12
Melanoma
−
25
Breast
+
36
Lung
+
18
Tumor
Histopathology and prognosis Good prognosis Less invasiveness, good prognosis Low-grade differentiation, good prognosis Bad prognosis Low-grade differentiation, good prognosis Low-grade differentiation
Reference Cabrera et al. (1995) Gutierrez et al. (1987) Esteban et al. (1990a)
¨ Brocker et al. (1985), ´ Lopez-Nevot et al. (1988) Brunner et al. (1991), Natali et al. (1983), Concha et al. (1991a) Redondo et al. (1991a)
MHC Antigens and Tumor Escape from Immune Surveillance
149
colon carcinomas we observed a correlation between the absence of HLA-DR antigen expression and invasiveness. Tumors with a worse prognosis (Jass’s stages IV and V) were HLA-DR negative, in contrast to the HLA-DR positivity in tumors with a more favorable prognosis (Jass’s stages I–II) (Gutierrez et al., 1987, 1990). In lung and breast carcinomas there was a clear relationship between HLA-DR expression and well-differentiated tumors (Redondo et al. 1991a; Concha et al. 1991b). Other authors have also found an association between positive HLA-DR expression and a more favorable prognosis in breast carcinoma (Natali et al., 1983; Brunner et al., 1991), gastric carcinoma (Hilton and Kevin, 1990), and large bowel carcinomas (Andersen et al., 1993). Constitutive expression of HLA-DR antigens in high-degree B cell lymphomas has been associated with less aggressive behavior in these neoplasms (Momburg et al., 1987; Spier et al., 1988). The expression of HLA-DR in melanomas seems to be an exception. In this type of neoplasm we and others have found that HLA-DR expression is correlated with a ¨ more aggressive phenotype and high risk of metastasis (Brocker et al., 1985; ´ Lopez-Nevot et al., 1988). It should be noted, however, that other authors found no relation between HLA-DR expression and parameters of tumor ¨ aggressiveness or prognosis in different types of tumors (Moller et al., 1991; Wintzer et al., 1990). The biological implications of HLA class II expression in tumor cells are not clear. It can be hypothesized that in some types of tumors this expression gives the cell the ability to present antigens, thereby conferring to the tumor a low degree of aggressiveness and a better prognosis. This theory is supported by the finding that transfectants obtained from a variety of mouse tumors transfected with syngeneic MHC class II genes are very effective vaccines against subsequent challenge with wild-type class II tumors (OstrandRosenberg et al., 1990). Also, in vitro transfection assays in a melanoma system showed that class II-positive tumor cells were able to function as antigen-presenting cells (Chen et al., 1994). The following possibilities may explain the expression of MHC class II molecules in tumors: (i) DR expression may be a marker of tumor differentiation, and (ii) de novo DR expression may be induced by lymphokines released by tumor-infiltrating lymphocytes. Currently, however, data support both possibilities.
IX. CONCLUSIONS HLA class I altered phenotypes are frequent in a variety of human tumors derived from HLA class I–positive epithelia. Total or selective losses of HLA class I antigens have been reported in more than 90% of some tumor
150
Federico Garrido and Ignacio Algarra
samples. Multiple molecular mechanisms have been identified as responsible for the generation of these altered tumor HLA phenotypes. These mechanisms are not exclusive to a particular tumor but are found repeatedly in tumors derived from a variety of different tissues. HLA class I alteration therefore represents a major mechanism of tumor escape from T cell immune responses. MHC class I-deficient tumor cells are selected by antitumor T cell-mediated responses, and altered MHC phenotypes produced during metastatic colonization in experimental models are not random but can be reproduced in different syngeneic animals (Garcia Lora et al., 2001). Possible routes of tumor escape from NK immune surveillance were reviewed, including the role of virus HLA class I homologs that induce HLA-E expression. However, preliminary work by different groups has failed to confirm aberrant cell surface expression of HLA-G and -E molecules in tumors. Indirect evidence of immunosurveillance against the growth and dissemination of tumors is provided each time a particular route of tumor escape from T and NK cell attack is defined. The new science of “escapology” frequently centers on HLA class I molecules. In order to transform this view of tumor immunology into a potentially powerful tool for clinical application, it will be necessary to define individual HLA altered phenotypes and identify the molecular mechanisms used by tumor cells in individual cancer patients. This review also proposes that a careful analysis of HLA expression in tumors should be done routinely in patients who are candidates for T cell-based immunotherapy. This individual diagnosis will probably generate unique treatment aimed at modifying the altered HLA phenotype. MHC class I downregulation should not be seen as an obstacle to performing T cell-based immunotherapy, it can be considered a crucial step in the natural history of tumor development. This view is an oversimplification of the situation since the complexity of cancer and the discovery of a variety of immune escape routes along with the generation of multiple tumor variants in a particular individual will probably continue to prolong the search for better anticancer strategies.
ACKNOWLEDGMENTS We thank all the member of the Departamento de Analisis Clinicos at Hospital Universitario Virgen de las Nieves, Granada, Spain, for their contribution to this work. We thank Dr. Teresa Cabrera for providing the immunohistological images of normal and tumor tissues. The comments and suggestions made by Dr. Peter Stern are gratefully acknowledged. This work was partially supported by the Fondo de Investigaciones Sanitarias and Consejeria de Salud de la Junta de Andalucia, Spain.
MHC Antigens and Tumor Escape from Immune Surveillance
151
REFERENCES Abril, E., Mendez, R. E., Garcia, A., Serrano, A., Cabrera, T., Garrido, F., and Ruiz-Cabello, F. (1996). Tissue Antigens 47, 391–398. Alcami, A., and Koszinowski, U. H. (2000). Immunol. Today 21, 447– 454. ¨ ¨ Algarra, I., Ohlen, C., Perez, M., Ljunggren, H. G., Klein, G., Garrido, F., and Karre, K. (1989). Int. J. Cancer 44, 675–680. Algarra, I., Gaforio, J. J., Garrido, A., Mialdea, M. J., Perez, M., and Garrido, F. (1991). Int. J. Cancer Suppl. 6, 73–81. Algarra, I., Cabrera, T., and Garrido, F. (2000). Hum. Immunol. 61, 65–73. Amirghofran, Z., Khezri, A. A., and Mohammadi, S. (1997). Irn. J. Med. Sci. 22, 145. Andersen, S. N., Rognum, T. O., Lund, E., Meling, G. I., and Hauge, S. (1993). Br. J. Cancer 68, 80–85. Atkin, N. B. (1997). Cytogenetics 95, 33–39. Bartholomew, J. S., Glenville, S., Sarkar, S., Burt, D. J., Stanley, M. A., Ruiz-Cabello, F., Chengang, J., Garrido, F., and Stern, P. L. (1997). Cancer Res. 57, 937–942. Basham, T. Y., Nickoloff, B. F., and Merigan, T. C. (1985). J. IFN Res. 5, 23–32. Bauer, S., Groh, V., Wu, J., Steinle, A., Phillips, J. H., Lanier, L. L., and Spies, T. (1999). Science 285, 727–729. ´ Benitez, R., Godelaine, D., Lopez Nevot, M. A., Brasseur, F., Jimenez, P., Marchand, M., Oliva, M. R., van Baren, N., Cabrera, T., Andry, G., Landry, C., Ruiz-Cabello, F., Boon, T., and Garrido, F. (1998). Tissue Antigens 52, 520–529. Bicknell, D. C., Rowan, A., and Bodmer, W. F. (1994). Proc. Natl. Acad. Sci. USA 91, 4751– 4756. Blades, R. A., Keating, P. J., McWillian, L. J., George, N. J., and Stern, P. L. (1995). Urology 46, 681–686. Boon, T. (1983). Adv. Cancer Res. 39, 121–151. Boon, T., Cerottini, J. C., Van den Eynde, B., van der Bruggen, P., and Van Pel, A. (1994). Annu. Rev. Immunol. 12, 337–365. Brady, C. S., Burt, D. J., Bartholomew, J. S., Duggan-Keen, M. F., Glenville, S., Telford, N., Little, A. M., Davison, J., Jimenez, P., Ruiz-Cabello, F., Garrido, F., and Stern, P. L. (2000). Tissue Antigens 55, 401– 411. Branch, P., Rowan, A., Bicknell, D. C., Bodmer, W. F., and Karran, P. (1995). Nature Genet. 9, 231–232. Brasanac, D., Muller, C. A., Muller, G. A., Hadzi-Dzokic, J., and Markovic-Lipkovski, J. (1999). J. Exp. Clin. Cancer Res. 18, 505–510. ¨ ¨ Braud, V. M., Allan, D. S., O’Callaghan, C. A., Soderstr om, K., D’Andrea, A., Ogg, G. S., Lazetic, S., Young, N. T., Bell, J. I., Phillips, J. H., Lanier, L. L., and McMichael, A. J. (1998). Nature 391, 795–799. Breivik, J., and Gaudernack, G. (1999). Adv. Cancer Res. 76, 187–209. ¨ ¨ Brocker, E. B., Uter, L., Bruggen, J., Ruiter, D. J., Macher, E., and Sorg, C. (1985). Int. J. Cancer 36, 29–35. Browning, M. J., Krausa, P., Rowan, A., Hill, A. B., Bicknell, D. C., Bodmer, J. G., and Bodmer, W. F. (1993). J. Immunother. 14, 163–168. Browning, M., Petronzelli, F., Bicknell, D., Krausa, P., Rowan, A., Tonks, S., Murray, N., Bodmer, J., and Bodmer, W. (1996). Tissue Antigens 47, 364–371. Brunner, C. A., Gokel, J. M., Riethmuller, G., and Johnson, J. P. (1991). Eur. J. Cancer 27, 411– 416. Burnet, F. M. (1970). Prog. Exp. Tumor Res. 13, 1–27. Buszello, H., and Ackermann, R. (1994). Eur. Urol. 25, 158–163.
152
Federico Garrido and Ignacio Algarra
Cabrera, T., Garrido, V., Concha, A., Martin, J., Esquivias, J., Oliva, M. R., Ruiz-Cabello, F., Serrano, S., and Garrido, F. ( 1992). Immunobiology 185, 440– 452. Cabrera, T., Ruiz-Cabello, F., and Garrido, F. (1995). Scand. J. Immunol. 41, 398–406. ´ Cabrera, T., Sierra, A., Fernandez, M. A., Herruzo, A., Fabra, A., and Garrido, F. (1996). Hum. Immunol. 50, 127–134. ´ ´ A., Sancho, J., Ruiz-Cabello, F., and Cabrera, T., Collado, A., Fernandez, M. A., Ferron, Garrido, F. (1998). Tissue Antigens 52, 114–123. ´ Cabrera, T., Salinero, J., Fernandez, M. A., Garrido, A., Esquivias, J., and Garrido, F. (2000). Hum. Immunol. 61, 499–506. Carbone, D. P. (1997). Semin. Oncol. 24, 388– 401. Chen, H. L., Gabrilovich, D., Virmani, A., Ratnani, I., Girgis, K. R., Nadaf-Rahrov, S., Fernandez-Vina, M., and Carbone, D. P. (1996a). Int. J. Cancer 67, 756–763. Chen, H. L., Gabrilovich, D., Tamp´e, R., Girgis, K. R., Nadaf, S., and Carbone, D. P. (1996b). Nature Genet. 13, 210–213. Chen, P., Ullrich, S., and Ananthaswamy, H. (1994). J. Leukocyte Biol. 56, 469– 474. Chouaib, S., Asselin-Paturel, C., Mami-Chouaib, F., Caignard, A., and Blay, Y. J. (1997). Immunol. Today 18, 493– 497. Clark, W. H., From, L., Bernardino, E. A., and Mihm, M. C. (1969). Cancer Res. 29, 705–711. Cohen, G. B., Gandhi, R. T., Davis, D. M., Mandelboim, O., Chen, B. K., Strominger, J. L., and Baltimore, D. (1999). Immunity 10(6), 661–671. Concha, A., Cabrera, T., Ruiz-Cabello, F., and Garrido, F. (1991a). Int. J. Cancer Suppl. 6, 146–154. Concha, A., Esteban, F., Ruiz-Cabello, F., and Garrido, F. (1991b). Semin. Cancer Biol. 2, 47–54. Concha, A., Nogales, F., Cabrera, T., Collado, A., Garrido, F., and Ruiz-Cabello, F. (1995). Eur. J. Immunogen. 22, 299–310. Connor, M. E., and Stern, P. (1990). Int. J. Cancer 46, 1029–1034. ´ Cordon-Cardo, C., Fuks, Z., Drobnjak, M., Moreno, C., Eisenbach, L., and Feldman, M. (1991). Cancer Res. 51, 6372–6380. Coulie, P. G., Ikeda, H., Baurain, J. F., and Chiari, R. (1999). Adv. Cancer Res. 76, 213–242. Cromme, F. V., van Bommel, P. F., Walboomers, J. M., Gallee M. P., Stern, P., Kenemans, P., Helmerhorst, T. J., Stukart, M. J., and Meijer, C. J. (1994). Br. J. Cancer 69, 1176–1181. Daar, A. S., Fuggle, S. V., Fabre, J. V., Ting, A., and Morris, P. J. (1984). Transplantation 38, 287–292. D’Alessandro, G., Zardawi, I., Grace, J., McCarthy, W. H., and Hersey, P. (1987). Pathology 19, 339–346. Dovhey, S. E., Ghosh, N. S., and Wright, K. L. (2000). Cancer Res. 60, 5789–5796. D’Urso, C. M., Wang, Z. G., Cao, Y., Tatake, R., Zeff, R. A., and Ferrone, S. (1991). J. Clin. Invest. 87, 284–292. Ehrlich, P. (1909). Ned. Tijdschr. Geneeskd. 5, 273–290. Eisen, A., and Weber, B. (1998). Curr. Opin. Oncol. 10, 486– 491. Engel, A. M., Svane, I. M., Rygaard, J., and Wedelin, O. (1997). J. Immunol. 45, 463– 470. Ernstoff, M. S., Duray, P., Stenn, K., and Kirkwood, J. (1985). J. Invest. Dermatol. 84, 430– 432. Esteban, F., Concha, A., Huelin, C., P´erez-Ayala, M., Pedrinaci, S., Ruiz-Cabello, F., and Garrido, F. (1989). Int. J. Cancer 43, 436– 439. Esteban, F., Concha, A., Delgado, M., P´erez-Ayala, M., Ruiz-Cabello, F., and Garrido, F. (1990a). Br. J. Cancer 62, 1047–1051. Esteban, F., Ruiz-Cabello, F., Concha, A., Perez-Ayala, M., Sanchez-Rozas, J. A., and Garrido, F. (1990b). Clin. Exp. Metastasis 8, 319–328. Fearon, E. R., and Vogelstein, B. (1990). Cell 61, 759–767. Feenstra, M., Rozemuller, E., Duran, K., Stuy, I., van den Tweel, J., Slootweg, P., Weger, R., and Tilanus, M. (1999a). Hum. Immunol. 60, 697–706.
MHC Antigens and Tumor Escape from Immune Surveillance
153
Feenstra, M., Veltkamp, M., Van Kuik, J., Wiertsma, S., Slootweg, P., van Den Tweel, J., De Weger, R., and Tilanus, M. (1999b). Tissue Antigens 54, 235–245. Feenstra, M., Verdaasdonk, M., van der Zwan, A. W., Weger, R., Slootweg, P., and Tilanus, M. (2000). Lab. Invest. 80, 405– 414. Fernandez, J. E., Concha, A., Aranega, A., Ruiz-Cabello, F., Cabrera, T., and Garrido, F. (1991). Immunobiology 182, 440– 448. ´ ´ Fernandez, M. A., Oliva, M. R., Cabrera, T., Lopez Nevot, M. A., Jimenez, P., Ruiz-Cabello, F., and Garrido, F. (2000). Int. J. Clin. Lab. Res. 30, 87–92. Ferron, A., Perez-Ayala, M., Concha, A., Cabrera, T., Redondo, M., Oliva, M. R., Ruiz-Cabello, F., and Garrido, F. (1989). J. Immunogen. 16, 413– 423. Ferrone, S., and Marincola, F. M. (1995). Immunol. Today 16, 487– 494. Festenstein, H. (1987). Br. Med. Bull. 43, 217–225. Festenstein, H., and Garrido, F. (1986). Nature 322, 502–503. Fleming, K. A., McMichael, A., Morton, J. A., Woods, J., and McGee, J. O. (1981). J. Clin. Pathol. 34, 779–784. Frumento, G., Franchello, S., Palmisano, G. L., Nicotra, M. R., Giacomini, P., Loke, Y. W., Geragthy, D. E., Maio, M., Manzo, C., Natali, P. G., and Ferrara, G. B. (2000). Tissue Antigens 56, 30–37. Garc´ıa, Lora, A., Algarra, I., Gaforio, J. J., Ruiz-Cabello, F., and Garrido, F. (2001). Int. J. Cancer 91, 109–119. Garrido, A., Perez, M., Delgado, C., Garrido, M. L., Rojano, J., Algarra, I., and Garrido, F. (1986). Exp. Clin. Immunogen. 3, 98–110. Garrido, F., Festenstein, H., and Schirrmacher, V. (1976). Nature 261, 705–707. Garrido, F., Cabrera, T., Concha, A., Glew, S., Ruiz-Cabello, F., and Stern, P. (1993). Immunol. Today 14(10) 491– 499. Garrido, F., Cabrera, T., Lopez-Nevot, M. A., and Ruiz-Cabello, F. (1995). Adv. Cancer Res. 67, 155–195. ´ Garrido, F., Ruiz-Cabello, F., Cabrera, T., Perez-Villar, J. J., Lopez-Botet, M., Duggan-Keen, M., and Stern, P. L. (1997a). Immunol. Today 18, 2, 89–95. Garrido, F., Cabrera, T., Accola, R. S., Bensa, J. C., Wodmer, W., Dohr, G., Drenou, B., Drouet, M., Fauchet, R., Ferrara, G. B., Ferrone, S., Giacomini, P., Kageshita, T., Koopman, L., Maio, M., Marincola, F., Mazzilli, C., Morell, P. A., Murray, A., Papasteriades, Crh., Salvaneschi, L., Stern, P. L., and Ziegler, A. (1997b). In “HLA, Genetic Diversity of HLA. Functional and Medical Implications” (D. Charron, Ed.), Vol. 1, pp. 445– 452. EDK. S`evres, France. Geraghty, D. E. (1993). Curr. Opin. Immunol. 5, 3–7. Glew, S. S., Duggan-Keen, M., Cabrera, T., and Stern, P. L. (1992). Cancer Res. 52, 1–8. Grandis, J. R., Falkner, D. M., Melhem, M. F., Gooding, W. E., Drenning, S. D., and Morel, P. A. (2000). Clin. Cancer Res. 6, 2794–2802 . ´ Gutierrez, J., Lopez-Nevot, M. A., Cabrera, T., Oliva, M. R., Esquivias, J., Ruiz-Cabello, F., and Garrido, F. (1987). Exp. Clin. Immunogen. 4, 144–152. Gutierrez, J., Ruiz-Cabello, F., Lopez-Nevot, M. A., Cabrera, T., Esquivias, J., and Garrido, F. (1990). Immunobiology 180, 138–148. ¨ Hahne, M., Rimoldi, D., Schroter, M., Romero, P., Schreier, M., French, L. E., Schneider, P., Bornand, T., Fontana, A., Lienard, D., Cerottini, J. C., and Tschopp, J. (1996). Science 274, 1363–1366. Heney, N. M., Ahmed, S., Flanagan, M. J., Frable, W., Corder, M. P., Hafermann, M. D., and Hawkins, I. R. (1983). J. Urol. 130, 1083–1086. Hicklin, D. J., Wang, Z., Arienti, F., Rivoltini, L., Parmiani, G., and Ferrone, S. (1998). J. Clin. Invest. 101, 2720–2729. Hicklin, D. J., Marincola, F. M., and Ferrone, S. (1999). Mol. Med. Today 5, 178–186 . Hilders, C. G., Houbiers, J. G., Krul, E. J., and Fleuren, G. J. (1994). Am. J. Clin. Pathol. 101, 5–12.
154
Federico Garrido and Ignacio Algarra
Hilton, D. A., and Kevin, P. W. (1990). Cancer 66, 1154–1157. ¨ Holzmann, B., Brocker, E. B., Lehmann, J. M., Ruiter, D. J., Sorg, C., Riethmuller, G., and Johnson, J. P. (1987). Int. J. Cancer 39, 466– 471. Hurks, H. M., Metzelaar-Blok, J. A., Mulder, A., Claas, F. H., and Jager, M. J. (2000). Int. J. Cancer 85, 697–702. Ikeda, H., Leth´e, B., Lehmann, F., van Baren, N., Baurain, J. F., De Smet, C., Chambost, H., Vitale, M., Moretta, A., Boon, T., and Couli, P. G. (1997). Immunity 6, 199–208. ¨ ¨ Jager, E., Ringhoffer, M., Altmannsberger, M., Arand, M., Karbach, J., Jager, D., Oesch, F., and Knuth, A. (1997). Int. J. Cancer 71, 142–147. Janeway, C. A., Bottomly, K., Horowitz, Z., Kaye, J., Jones, B., and Tite, J. (1985). J. Immunol. 135, 739–742. Jeffery, P. K., and Reid, L. M. (1977). In “Development of the Lung. Lung Biology in Health and Disease” (W. A. Hodson, Ed.), Vol. 38, pp. 215–268. Dekker, New York. ´ J., Collado, A., Cabrera, T., Serrano, A., Real, L. M., Garc´ıa, A., RuizJimenez, P., Canton, Cabello, F., and Garrido, F. (1999). Int. J. Cancer 83, 91–97. Jimenez, P., Canton, J., Concha, A., Cabrera, T., Fernandez, M., Real, L. M., Garcia, A., Serrano, A., Garrido, F., and Ruiz-Cabello, F. (2000). Cancer Immunol. Immunother. 48, 684–690. Kaklamanis, L., Leek, R., Koukourakis, M., Gatter, K. C., and Harris, A. L. (1995). Cancer Res. 55, 5191–5194. Kallfelz, M., Jung, D., Hilmes, C., Knuth, A., Jaeger, E., Huber, C., and Seliger, B. (1999). Int. J. Cancer 81, 125–133. Karre, K., Ljunggren, H. G., Pionteck, G., and Kiessling, R. (1986). Nature 261, 705–707. Keating, P. J., Cromme, F. V., Duggan-Keen, M., Snijders, P. J., Walboomers, J. M., Hunter, R. D., Dyer, P. A., and Stern, P. L. (1995). Br. J. Cancer 72, 405– 411. Kersemaekers, A. M., Kenter, G. G., Hermans, J., Fleuren, G. J., and van de Vijver, M. J. (1998). Int. J. Cancer 79, 411– 417. Klein, B., Levin, I., and Klein, T. (1996). Isr. J. Med. Sci. 33, 1238–1243. ¨ ¨ Klein, G., Sjogren, H. O., Klein, E., and Hellstrom, K. E. (1960). Cancer Res. 20, 1561–1572. Koopman, L. A., Mulder, A., Corver, W. E., Anholts, J. D. H., Giphart, M. J., Claas, F. H. J., and Fleuren, G. J. (1998). Tissue Antigens 51, 623–636. Koopman, L. A., van der Slik, A. R., Giphart, M. J., and Fleuren, G. J. (1999). J. Natl. Cancer Inst. 91, 1669–1677. Koopman, L. A., Corver, W. E., van der Slik, A. R., Giphart, M. J., and Fleuren, G. J. (2000). J. Exp. Med. 191, 961–975. Korkolopoulou, P., Kaklamanis, L., Pezzella, F., Harris, A. L., and Gatter, K. C. (1996). Br. J. Cancer 73, 148–153. Le Boutellier, P. (1997). Eur. J. Immunogen. 24, 397– 408. Lehmann, F., Marchand, M., Hainaut, P., Sastre, X., Ikeda, H., Boon, T., and Coulie, P. G. (1995). Eur. J. Immunol. 25, 340–347. Levin, I., Klein, T., Goldstein, J., Kuperman, O., Kanetti, J., and Klein, B. (1991). Cancer 68, 2591–2594. Levin, I., Klein, B., Segal, S., Eyal, A., Gopas, J., Kanetti, J., Saadon, M., and Kuperman, O. (1992). Tissue Antigens 39, 19–22. Levin, I., Klein, T., Kuperman, O., Segal, S., Shapira, J., Gal, R., Hart, Y., and Klein, B. (1994). Cancer Detection Prevention 18, 443–445. Ljunggren, H. G., and Karre, K. (1985). J. Exp. Med. 180, 2287–2295. Llano, M., Lee, N., Navarro, F., Garcia, P., Albar, J. P., Geraghty, D. E., and Lopez-Botet, M. (1998). Eur. J. Immunol. 28, 2854–2863. ´ Lopez-Nevot, M. A., Garc´ıa, E., Pareja, E., Bonal, F. J., Mart´ın, J., Ruiz-Cabello, F., and Garrido, F. (1986). J. Immunogen. 13, 219–227. ´ Lopez-Nevot, M. A., Garcia, E., Romero, C., Oliva, M. R., Serrano, S., and Garrido, F. (1988). Exp. Clin. Immunogen. 5, 203–212.
MHC Antigens and Tumor Escape from Immune Surveillance
155
´ ´ A., Gutierrez, J., Oliva, M. R., Romero, C., Huelin, Lopez-Nevot, M. A., Esteban, F., Ferron, C., Ruiz-Cabello, F., and Garrido, F. (1989). Br. J. Cancer 59, 221–226. Lu, Q. L., Abel, P., Mitchell, S., Foster, C., and Lalani, E. N. (2000). J. Pathol. 190, 169–176. ¨ Luboldt, H. J., Kubens, B. S., Rubben, H., and Wilde, H. G. (1996). Cancer Res. 56, 826–830. Maiorana, A., Cesinaro, A. M., Fano, A., and Collina, G. (1995). Clin. Exp. Metastasis 13, 43– 48. Marchand, M., Weynants, P., Rankin, E., Arienti, F., Belli, F., Parmiani, G., Cascinelli, N., Bourlond, A., Vanwijck, R., Humblet, Y., Canon, J. L., Laurent, C., Naeyaert, J. M., Plagne, ¨ R., Deraemaeker, R., Knuth, A., Jager, E., Brasseur, F., Herman, J., Couli, P. G., and Boon, T. (1995). Int. J. Cancer 63, 883–885. Marchand, M., Von Baren, N., Weynants, P., Brichard, V., Dreno, B., Tessier, M. H., Rankin, E., Parmiani, G., Arienti, F., Humblet, Y., Bourlond, A., Vanwijck, R., Lienard, D., Beauduin, M., Dietrich, P. Y., Russo, V., Kerger, J., Masucci, G., Jager, E., De Greve, J., Atzpodien, J., Brasseur, F., Coulie, P., van der Bruggen., P., and Boon, T. (1999). Int. J. Cancer 80, 219–222. Marincola, F. M., Shamamian, P., and Alexander, R. B. (1994). J. Immunol. 153, 1225–1237. Martayan, A., Fraioli, R., Giorda, E., Setini, A., Ciccarelli, G., Delfino, L., Ferrara, G. B., and Giacomini, P. (1999). Br. J. Cancer 80, 639–649. Mazurenko, N., Attaleb, M., Gritsko, T., Semjonova, L., Pavlova, L., Sakharova, O., and Kisseljov, F. (1999). Oncol. Rep. 6, 859–863. McGovern, V. J., Shaw, H. M., Milton, G. W., and Farago, G. A. (1979). Histopathology 3, 385–393. ¨ Mendez, R., Serrano, A., Jager, E., Ramal, L. M., Ruiz-Cabello, F., Knuth, A., and Garrido, F. (2000). Hum. Immunol. 61(suppl. 1), 226. Mendez, R., Serrano, A., Jaeger, E., Maleno, I., Ruiz-Cabello, F., Knuth, A., and Garrido, F. (2001). Tissue Antigens, in press. Mertens, F., Johansson, B., Husglund, M., and Mitelman, F. (1997). Cancer Res. 57, 2765– 2780. ¨ Moller, P., Mattfeldt, T., Gross, C., Schlosshauer, P., Kock, A., Koretz, K., Moldenhauer, G., Kaufmann, M., and Otto, H. F. (1989). Am. J. Pathol. 135, 73–78. ¨ Moller, P., Koretz, K., Schlag, P., and Momburg, F. (1991). Int. J. Cancer Suppl. 6, 155–162. ¨ Momburg, F., Moller, P., Moldehauer, G., and Hammerling, G. J. (1986). J. Immunogen. 13, 195–199. Momburg, F., Hermann, B., Moldenhauer, G., and Moller, P. B. (1987). Int. J. Cancer 40, 598–603. Moretta, A., Bottino, C., Vitale, M., Pende, D., Biassoni, R., Mingari, M. C., and Moretta, L. ( 1996). Annu. Rev. Immunol. 14, 619–648. Moretta, A., Biassoni, R., Bottino, C., Mingari, M. C., and Moretta, L. (2000). Immunol. Today 21(5), 228–234. Natali, P. G., Giacomini, P., Bigotti, A., Imai, K., Nicotra, M. R., and Ferrone, S. (1983). Cancer Res. 43, 660–668. Natali, P. G., Nicotra, M. R., Bigotti, A., Venturo, I., Marcenaro, L., Giacomini, P., and Russo, C. (1989). Proc. Natl. Acad. Sci. USA 86, 6719–6723. Nestle, F. O., Alijagic, S., Gilliet, M., Sun, Y., Grabbe, S., Dummer, R., Burg, G., and Schadendorf, D. (1998). Nature Med. 4(3), 328–332. Niall, M., and Heney, M. D. (1992). Urol. Clin. North Am. 19, 429– 433. Nistico, P., Mottolese, M., Mammi, C., Benevolo, M., Del Bello, D., Gentile, P., Botti, C., Venturo, I., and Natali, P. G. (1997). J. Natl. Cancer Inst. USA 89, 319–321. Nouri, A. M., Hussain, R. F., and Oliver, R. T. (1994). Cancer Gene Ther. 2, 119–123. Ochsenbein, A. F., Klenerman, P., Karrer, U., Ludewig, B., Pericin, M., Hengartner, H., and Zinkernagel, R. M. (1999). Proc. Natl. Acad. Sci. USA 96, 2233–2238. Onno, M., Pangault, C., Le Friec, G., Guilloux, V., Andre, P., and Fauchet, R. (2000). J. Immunol. 164, 6426–6434.
156
Federico Garrido and Ignacio Algarra
Ostrand-Rosenberg, S., Thakur, A., and Clements, V. (1990). J. Immunol. 144, 4068– 4071. Ostrand-Rosenberg, S., Baskar, S., Pattersson, N., and Clements, K. (1996). Tissue Antigens 47, 414– 421. Ottensen, S. S., and Kieler, J. (1991). Anticancer Res. 11, 217–223. Parmiani, G., Taramelli, A., Anichini, A., Andreola, S., Cascinelli, N., and Fossati, G. (1986). J. Immunogen. 13, 235–240. Paul, P., Rouas-Freis, N., Khalil-Daher, I., Moreau, P., Riteau, B., Le Gal, F. A., Avril, M. F., Dausset, J., Guillet, J. G., and Carosella, E. (1998). Proc. Natl. Acad. Sci. USA 95, 4510 – 4515. Peltenburg, L. T., and Schrier, P. I. (1994). Immunogenetics 40, 54–61. Pera, E., Moreno, A., and Galindo, L. (1986). Cancer 58, 928–934. ´ ´ P´erez, B., Benitez, R., Fernandez, M. A., Oliva, M. R., Soto, J. L., Serrano, S., Lopez-Nevot, M. A., and Garrido, F. (1999). Tissue Antigens 53, 569–572. ´ ´ F., Ruiz-Cabello, F., and P´erez, M., Cabrera, T., Lopez-Nevot, M. A., Gomez, M., Peran, Garrido, F. (1986). J. Immunogen. 13, 247–253. P´erez, M., Algarra, I., Ljunggren, H. G., Caballero, A., Mialdea, M. J., Gaforio, J. J., Klein, G., ¨ Karre, K., and Garrido, F. (1990). Int. J. Cancer 46, 258–261. ´ Ramal, L. M., Fenestra, M., van der Zwan, A. W., Collado, A., Lopez-Nevot, M. A., Tilanus, M., and Garrido, F. (2000a). Tissue Antigens 55, 443– 448. ´ Ramal, L. M., Maleno, I., Cabrera, T., Collado, A., Ferron, A., Lopez-Nevot, M. A., and Garrido, F. (2000b). Hum. Immunol. 61(10), 1001–1012. Real, L. M., Cabrera, T., Canton, J., Kirkin, A. F., Garcia, A., Abril, E., Zeuthen, J., Ruiz Cabello, F., and Garrido, F. (1998). Int. J.Cancer 75, 317–323. Real, L. M., Cabrera, T., Collado, A., Jimenez, P., Garcia, A., Ruiz-Cabello, F., and Garrido, F. (1999). Int. J. Cancer 8, 512–518. Rebmann, V., Wagner, S., Pabler, M., Goos, M., and Grosse-Wilde, H. (2000). Hum. Immunol. 61(Suppl. 1), 130. Redondo, M., Concha, A., Oldiviela, R., Cueto, A., Gonzalez, A., Garrido, F., and Ruiz-Cabello, F. (1991a). Cancer Res. 51, 4948– 4954. Redondo, M., Ruiz-Cabello, F., Concha, A., Cabrera, T., P´erez-Ayala, M., Oliva, M. R., and Garrido, F. (1991b). Cancer Res. 51, 2463–2468. Redondo, M., Concha, A., Ruiz-Cabello, F., Morell, M., Esteban, F., Talavera, P., and Garrido, F. (1997). Cancer Detection Prevention 21, 22–28. Restifo, N. P., Esquivel, F., Kawakami, Y., Yewdell, J. W., Mul´e, J. J., Rosenberg, S. A., and Bennink, J. R. (1993). J. Exp. Med. 177, 265–272. Restifo, N., Marincola, F., Kawakami, Y., Taubenberger, J., Yannelli, J. R., and Rosenberg, S. A. (1996). J. Natl. Cancer Inst. 88, 100–108. Rodriguez, F., Peran, F., Garrido, F., and Ruiz-Cabello, F. (1994). Immunogenetics 39, 161–167. Rosa, F., Berissi, H., Weisssenbach, J., Maroteaux, L., Fellous, M., and Revel, M. (1983). EMBO J. 2, 239–243. Rosenberg, S. A., Yang, J. C., Schwartzentruber, D. J., Hwu, P., Marincola, F., Topalian, S. L., Restifo, N. P., Dudley, M. E., Schwarz, S. L., Spiess, P. J., Wunderlich, J. R., Parkhurst, M. R., Kawakami, Y., Seipp, C. A., Einhorn, J. H., and White, D. E. (1998). Nature Med. 4, 321–327. Ruiter, D. J., Bhan, A. K., Harrist, T. J., Sober, A. J., and Mihm, M. C. (1982). J. Immunol. 129, 2808–2815. Ruiter, D. J., Bergman, W., Welvaart, K., Scheffer, E., van Vloten, W. A., Russo, C., and Ferrone, S. (1984). Cancer Res. 44, 3930–3935. Ruiter, D. J., Mattijssen, V., Broecker, E. B., and Ferrone, S. (1991). Semin. Cancer Biol. 2, 35– 45. Ruiz-Cabello, F., and Garrido, F. (1998). Immunol. Today 12, 539–542.
MHC Antigens and Tumor Escape from Immune Surveillance
157
Ruiz-Cabello, F., P´erez-Ayala, M., Gomez, O., Redondo, M., Concha, A., Cabrera, T., and Garrido, F. (1991). Int. J. Cancer Suppl. 6, 123–130. Sanda, M. G., Restifo, N. P., Walsh, J. C., Kawakami, Y., Nelson, W. G., Pardoll, D. M., and Simons, J. W. (1995). J. Natl. Cancer Inst. 87, 280–285. ¨ Schrier, P., Versteeg, R., Peltemburg, L., Plomp, A. C., van’t Veer, L. J., and Kruse-Wolters, K. M. (1991). Semin. Cancer Biol. 2, 73–83. Seliger, B., Hohne, A., Knuth, A., Bernhard, H., Meyer, T., Tampe, R., Momburg, F., and Huber, C. (1996). Cancer Res. 56, 1756–1760. Seliger, B., Hohne, A., Jung, D., Kallfelz, M., Knuth, A., Jaeger, E., Bernhard, H., Momburg, F., Tampe, R., and Huber, C. (1997). Exp. Hematol. 25, 608–614. Serrano, A., Brady, C. S., Jimenez, P., Duggan-Keen, M. F., Mendez, R., Stern, P., Garrido, F., and Ruiz-Cabello, F. (2000). Immunogenetics 51, 1047–1052. Setini, A., Beretta, A., De Santis, C., Meneveri, R., Martayan, A., Mazilli, M. C., Apella, E., Siccardi, A. G., Natali, P. G., and Giacomini, P. (1996). Hum. Immunol. 46, 69–81. Sharpe, J. C., Abel, P. D., Gilbertson, J. A., Brawn, P., and Foster, C. S. (1994). Br. J. Urol. 74, 609–616. Singal, D. P., Ye, M., and Bienzle, D. (1998). Int. J. Cancer 75, 112–116. Smith, M., Bodmer, W., and Bodmer, J. (1988). Lancet 1, 823–824. Soong, T. W., and Hui, K. M. (1992). J. Immunol. 149, 2008–2020. Spier, C. M., Grogan, T. M., Lippman, S. M., Slymen, D. J., Rybskki, J. A., and Miller, T. P. (1988). Am. J. Pathol. 133, 118–126. Stern, P. (1996). Adv. Cancer Res. 67, 175–211. Thomas, L. (1959). “Discussion of cellular and humoral aspects of the hypersensitivity states.”In (H. S., Lawrence, Ed.), pp. 529–532. Hoeber-Harper, New York. Tomasec, P., Braus, V., Rickards, C., Powell, M., McSharry, B., Gadola, S., Cerundolo, V., Borysiewicz, L., McMichael, A., and Wilkinson, G. (2000). Science 287, 1031–1033. Tomita, Y., Matsumoto, Y., Nishiyama, T., and Fujiwara, M. (1990). J. Pathol. 162, 157–164. Tomlinson, I., and Bodmer, W. (1999). Nature Med. 5, 11–12. Torres, L. M., Cabrera, T., Concha, A., Oliva, M. R., Ruiz-Cabello, F., and Garrido, F. (1993). Tissue Antigens 41, 65–71. Torres, M. J., Ruiz-Cabello, F., Skoudy, A., Berrozpe, G., Jimenez, P., Cabrera, T., Real, F. X., and Garrido, F. (1996). Tissue Antigens 47, 372–381. Townsend, A., Rothbard, J., and Gotch, F. M. (1986). Cell 44, 959–968. Thurner, B., Haendle, I., Roder, C., Dieckmann, D., Keikavoussi, P., Jonuleit, H., Bender, A., Maczek, C., Schreiner, D., von den Driesch, P., Brocker, E. B., Steinman, R. M., Enk, A., Kampgen, E., and Schuler, G. (1999). J. Exp. Med. 190, 1669–1678. Ugolini, S., and Vivier, E. (2000). Curr. Opin. Immunol. 12, 295–300. Van der Bruggen, P., Traversari, C., Chomez, P., Lurguin, C. D., Plaen, E., Van den Eynde, B., Knuth, A., and Boon, T. (1991). Science 254, 1643–1647. Van Duinen, S. G., Mauw, B. J., Graaff-Reitsma, C. B., and Ruiter, D. J. (1984). Lab. Invest. 50, 733–741. Versteeg, R., Kruse-Wolters, K. M., Plomp, A. C., van Leeuwen, A., Stam, N. J., Ploegh, H. L., Ruiter, D. J., and Schrier, P. I. (1989). J. Exp. Med. 170, 621–635. Vitale, M., Rezzani, R., Rodella, L., Zauli, G., Grigolato, P., Cadei, M., Hicklin, D. J., and Ferrone, F. (1998). Cancer Res. 15(4), 737–742. Vokes, E. E., Weichselbaum, R. R., Lippman, S. M., and Hong, W. K. (1993). N. Engl. J. Med. 328, 184–194. Wagner, S., Rebmann, V., Willers, C., Grosse-Wilde, H., and Goos, M. (2000). Lancet 356, 220–221. Walboomers, J. M., Jacobs, M. V., Manos, M. M., Bosch, F. X., Kummer, J. A., Shah, K. V., ˜ Snijders, P. J., Peto, J., Meijer, C. J., and Munoz, N. (1999). J. Pathol. 189, 12–19.
158
Federico Garrido and Ignacio Algarra
Wang, Z., Cao, Y., Albino, A. P., Zeff, R. A., Houghton, A., and Ferrone, S. (1993). J. Clin. Invest. 91, 684–692. Wang, Z., Margulies, L., Hicklin, D. J., and Ferrone, S. (1996). Tissue Antigens 47, 382–390. Wang, Z., Marincola, F. M., Rivoltini, L., Parmiani, G., and Ferrone, S. (1999). J. Exp. Med. 190, 205–215. Warshaw, A. L., and Fernandez del Castillo, C. N. (1992). N. Engl. J. Med. 326, 455– 465. ˜ M., Berzofsky, J. A., and Carbone, D. P. (1994). Cancer Wiendenfield, E. A., Fernandez-Vina, Res. 54, 1175–1177. Wintzer, H. O., Beuzing, M., and von Kleist, S. (1990). Br. J. Cancer 62, 289–295. Witjes, J. A., Umbas, R., Debruyne, F. M. J., and Schalken, J. A. (1995). J. Urol. 154, 2185– 2189.
FOUNDATIONS OF CANCER RESEARCH The Role of Selection in Progressive Neoplastic Transformation Harry Rubin Department of Molecular and Cell Biology and Virus Laboratory Life Sciences Addition University of California at Berkeley Berkeley, CA 94720
I. Introduction II. Spontaneous Neoplastic Transformation in Cell Culture A. Primary Cell Strains B. Established Cell Lines III. Selection in Transformation of Established Cell Lines A. Conditions for Eliciting and Expressing Spontaneous Transformation B. Interclonal Heterogeneity of Competence for Transformation C. Subliminal, Selectable Stages of Transformation D. The Role of Selective Clonal Expansion in Transformation IV. The Contribution of Apoptosis to Selection in Neoplastic Transformation V. Inhibition of Growth of Transformed Cells by Surrounding Nontransformed Cells VI. Confounding Effects of Variable Cell Behavior on the Dynamics of Transformation VII. Summary of the Major Features of Spontaneous Transformation VIII. Evidence for Selection in Experimental and Human Cancer IX. Sources of Genetic Variation for Possible Selection in Tumor Development X. The Nature of Selection in Vivo XI. Selection in Carcinogenesis by Polycyclic Aromatic Hydrocarbons XII. Conclusions References
Mathematical modeling indicates that selective growth of cells with biallelic mutations in tumor suppressor genes is the driving force in the development of most human tumors, and that increased mutation rate is not required. Spontaneous neoplastic transformation of cells in culture offers the opportunity for quantitative analysis of all stages of neoplastic progression, the cellular variation that underlies it, and the selective conditions that promote it. Most of the early work on spontaneous transformation was done in primary cultures of mouse embryo cells, but established mouse cell lines have been used more in recent years. The main criteria for transformation have been tumorigenesis in mice, increase in saturation density, and production of discrete, multilayered foci in confluent cell cultures. Spontaneous transformation in NIH 3T3
Advances in CANCER RESEARCH 0065-230X/01 $35.00
159
C 2001 by Academic Press. Copyright All rights of reproduction in any form reserved.
160
Harry Rubin
mouse fibroblasts is efficiently evoked by progressive selection under prolonged contact inhibition at high population density or during multiplication at low population density in suboptimal concentrations or types of serum. In general, it is a multistep process with some stages of progression occurring before there is any visible sign of transformed foci. There is a high degree of heritable heterogeneity in the original NIH 3T3 cell population for susceptibility to transformation. Isolation and expansion of minority susceptible clones from a relatively refractory population exhibit transformation long before the polyclonal parental population does because of the increased proportion of susceptible cells in these clones. There are indications that the selective conditions induce selectable variants. Tumor development in animals and man shares important characteristics with spontaneous transformation in culture, including a major role for selection, but the selective conditions for clonal expansion probably vary with the dynamics of differentiation in each tissue. These considerations support a role for an altered microenvironment (as in the aging process) in selective growth of rogue clones. C 2001 Academic Press.
I. INTRODUCTION It is widely accepted that tumors arise from genetic alterations in cells and that many such alterations are required before the capacity for malignant (i.e., invasive) growth develops (Armitage and Doll, 1954; Fisher, 1958; Nowell, 1976). Given generally assumed rates of classical point mutations (∼10−7 per gene per cell generation) and estimates of five to seven mutations for development of cancer, it was proposed that there had to be an increase in mutation rate to account for the frequency of cancer in the human population [Loeb, 1991; but see Simpson (1997) for much higher frequencies of somatic mutation]. This genetic instability was referred to as a mutator phenotype, and clear evidence for such a state is found in hereditary nonpolyposis colorectal cancer (HNPCC) of humans. In that case, mutation in one allele of a mismatch repair (MMR) gene is transmitted in the germline, and another mutation which occurs in the second allele of the colorectal epithelium results in an MMR defect, with an increase of several thousandfold in mutations of the simple sequence repeats that constitute microsatellites but are also found in structural genes. However, most colorectal tumors are sporadic and any increase in mutation rates arising from defects at an MMR locus would require two mutations to create a mutator phenotype. There is evidence that MMR defects do not occur in the early stages of tumor development (Homfray et al., 1998; Young et al., 1993). Studies have been done in culture to determine whether there are differences in mutation rates between malignant and normal cells, and variable results have been obtained. Some studies have failed to find a difference
Role of Selection in Neoplastic Transformation
161
in mutation rates between the two cell types (Eldridge and Gould, 1992; Elmore et al., 1983; Wittenkeller et al., 1997). Others have found the mutation rates to be significantly higher in malignant compared to normal cells (Eshleman et al., 1995; Seshadri et al., 1987). Some of the cases of increased mutation rates in malignant cells are associated with defects in MMR genes (Eshleman et al., 1995; Glaab and Tindall, 1997). The rate of gene amplification is much higher in some tumor cells than in their normal counterparts (Tlsty et al., 1989). A tumorigenic line of Chinese hamster cells with complex chromosome rearrangements had a mutation rate that was approximately five times higher than that of a diploid non-tumorigenic line, but there was overlapping of mutation rates among sublines of these cells (Kaden et al., 1989). The authors concluded that there was no simple relationship between spontaneous mutation rate and the malignant phenotype, and that mutation rate per se is not a sensitive index of malignancy. The overall results indicate that the many genetic changes found in human cancers cannot be simply explained by increased mutation rates, a point emphasized by the findings of MMR deficiency in some tumors and not in others (Kinzler and Vogelstein, 1996a). Recent computer modeling investigations have studied the relative roles of selection and increased mutation rate in the development of sporadic human colorectal cancer (Tomlinson and Bodmer, 1999; Tomlinson et al., 1996). They assume that selective growth advantage arises from inactivating mutations at both alleles of a tumor suppressor locus such as the adenomatous polyposis coli (APC) gene, and that the mutator phenotype requires mutations at both alleles of a mismatch repair gene. Even when the mutator phenotype has a mutation rate 104 times higher than that of normal cells, it is likely that the tumor starts to grow with a normal mutation rate no matter how many DNA repair loci are present (Tomlinson et al., 1996). The model does not address other forms of genetic change, such as chromosome rearrangements, or the effects that cause thousands of allelic losses in colorectal cancer (CRC) (Vogelstein et al., 1989) and an even greater number of inter-SSR abnormalities in CRC (Stoler et al., 1999). However, this is the first model that sets out two distinct alternatives and provides a plausible quantitative treatment that arrives at a clear-cut answer. The mathematical model, however, does not examine the diversity of neoplastic phenotypes, including the incipient stages of tumor progression, nor does it consider the variety of microenvironments that might confer selectivity. These have in fact been experimentally documented in studies of transformation in cell culture. The possible role of selection in neoplastic development was first suggested (Aaronson and Todaro, 1968a) 25 years after the discovery of spontaneous transformation of cells in culture (Earle, 1943a,b). There have since been many examples of spontaneous
162
Harry Rubin
transformation both in cultures freshly explanted from animals and in established cell lines. The easy accessibility and manipulation of cells in culture make it possible to observe the morphological changes during neoplastic development, the conditions that engender the transformation, and the quantitative growth properties of the cells at all stages. In most of the work on neoplastic transformation in cell culture, specific agents such as tumor viruses, chemical carcinogens, or putative oncogenes served as transforming agents, and attention was directed to the nature of those agents and the biochemical and behavioral changes they entailed in culture. However, studies on chemical carcinogenesis led indirectly to the discovery of spontaneous transformation in untreated cultures (Earle, 1943a) and to elaborating the dynamics of selection in the process.
II. SPONTANEOUS NEOPLASTIC TRANSFORMATION IN CELL CULTURE A. Primary Cell Strains In the course of experiments on chemical carcinogenesis in primary cultures of mouse fibroblasts from the inbred C3H line, it was discovered that some of the untreated cultures exhibited morphologic changes characteristic of cancer cells, although they did so much later than did the carcinogentreated cells (Earle, 1943a). Cells from both groups also produced sarcomas when inoculated into syngeneic mice, but the inoculations were not frequent enough to determine the time of origin of this property (Earle, 1943b; Sanford et al., 1950). At first, it was considered that the transformation of the untreated controls resulted from contamination with the carcinogen, but this possibility was later ruled out (Sanford et al., 1950). Quantitative studies in monolayer cultures of fibroblasts from partially inbred Swiss mice revealed a progressive decline in growth rate in successive transfers that was steeper in cultures maintained at low than at high population density (Todaro and Green, 1963). After a crisis period of little or no growth, the cultures exhibited a gradual increase in growth rate which reached a level as high as or higher than that of the originally explanted culture. The growth rate of the established cultures slowed down as they became crowded, and net growth ceased at various times after the cultures became confluent. The saturation density of the established cultures increased with the seeding density of cells used in serial passages but was independent of the seeding density used in the test to determine their saturation density. It was later shown
Role of Selection in Neoplastic Transformation
163
with cells from BALB/c mouse embryos that serial passage at high density resulted in both an increase in saturation density and a capacity to produce sarcomas in syngeneic mice; cells passaged at low density to minimize contact among the cells remained nontumorigenic indefinitely (Aaronson and Todaro, 1968a). However, clonal cultures of cells from hybrid mice passaged at low density in another laboratory did develop tumorigenic capacity, but they did so a few months and many population doublings later than those passaged at high density (Taylor et al., 1978). Tumors from the cells passaged at low density also had longer latent periods than those from the high-density passages. Both passage variations gave rise to cells that produced colonies after suspension in agar, but the cells from the high-density passages produced more and larger colonies. The authors concluded that carrying cells at high density accelerates neoplastic evolution of a population, but carrying them at low density does not prevent transformation (Taylor et al., 1978). Since high-density passages favor contact inhibition of nontransformed cells, it seemed that selection might play a dominant role in accelerating transformation. Problems arose with this interpretation in studies with rat fibroblasts. Although they first appeared to be resistant to transformation (Krooth et al., 1964), longer term cultivation resulted in their transformation (Jackson et al., 1970). The time to spontaneous transformation was significantly longer in the rat embryo cells than had been reported for mouse embryo cells in the same type of medium. Rat cells maintained without subculturing but with regular medium changes for almost 3 years as a dense, multilayered network with cells sloughing into the medium did not produce sarcomas in syngeneic rats at a time when cells which had undergone many subcultures from the same source had been tumorigenic for 9–14 months (Jackson et al., 1970). In studies with cells from another inbred rat strain, cultures in which cell contact was minimal showed a definite trend toward earlier transformation than those in which cell contact was high (Kirkland et al., 1975). There was a good correlation between the ability to form colonies in soft agar and the ability to form tumors in young syngeneic animals. The effects of a carcinogen could not be distinguished from spontaneous transformation. Interpretation of the results is complicated by changes in the type and concentration of serum during the cultivation period, with serum regimes used in the low-density cultures differing from those used in the high-density cultures. Effects of population density in spontaneous transformation of rat cells were not pursued further and the role of selection in that case has as inconclusive. Quantitative results with mouse cells show that selective conditions have narrow optima that have to be precisely defined, which was not done with the rat cells.
164
Harry Rubin
B. Established Cell Lines Studies on spontaneous transformation as a function of cell density using primary cultures obviously yielded conflicting results, especially with regard to rat cells. There are many complications in working with transformation in primary cultures, such as (i) the continuous loss of cells and decrease in growth rate, until variants appear which establish a permanent line of cells; (ii) the difficulty in performing clonal studies since the efficiency of cloning is very low and varies continuously with passage; and (iii) the great irregularity of transformation and the many months sometimes necessary to achieve it. In all likelihood, neoplastic transformation does not occur until a permanent line is established.
1. MOUSE FIBROBLASTS Permanent nontumorigenic cell lines have been established from embryo fibroblasts of various mouse strains (Aaronson and Todaro, 1968b; Jainchill et al., 1969; Todaro and Green, 1963). These cell lines are highly aneuploid with wide variations in chromosome number within each line (Rubin, 1993b; Rubin et al., 1984). They exhibit the growth properties of normal cells, including low saturation density at confluence and failure to produce tumors in syngeneic or immune-compromised mice. Spontaneous morphological transformation was reported in a few cells of the BALB 3T3 line after 4 months of weekly low-density passages (LDPs)(Rubin, 1981). A clone of transformed cells had a saturation density about 10 times higher than that of a nontransformed clone (Rubin, 1981) and produced sarcomas upon subcutaneous inoculation into nude mice (Rubin et al., 1986). Although the first observation of transformed cells occurred during the weekly LDPs at maximal growth rate on plastic substrates, transformed foci were rarely seen in continuous passages of the cells over approximately 7 years of culture under the same conditions. However, regular monitoring of the cultures for transformation by suspending them for colony formation in soft agar occasionally yielded large transformed colonies that produced sarcomas in nude mice (Rubin, 1988). Since the transformed cells from the large agar colonies were easily recognized by their rounded morphology when seeded on plastic even when they were surrounded by large numbers of nontransformed cells, it was evident that the transformation actually occurred during the incubation in agar and not when they were rapidly multiplying on the plastic surface. This was the first indication that transformation of an established cell line of fibroblasts was favored by conditions that selectively limit the multiplication of nontransformed cells. Although transformation occurred more frequently under conditions that restrained rather than allowed maximal expression of growth, it still occurred
Role of Selection in Neoplastic Transformation
165
rarely and irregularly, which frustrated a systematic, quantitative analysis of the process. A screen of other mouse cell lines revealed that the original NIH 3T3 cell line, which was widely used as a target for transformation by putative cellular oncogenes, regularly produced dense foci of transformed cells if left at confluence for more than 2 weeks (Rubin and Xu, 1989) or in a shorter period of time if they had been previously maintained in LDP on plastic in low calf serum (CS) concentration before assay at confluence (Rubin et al., 1990b). The transformed cells were tumorigenic in nude mice (Rubin et al., 1990a). The susceptibility of the cells to transformation at confluence was gradually lost during frequent LDPs of the nontransformed cells in high CS (Rubin, 1992). However, their susceptibility was heightened in parallel LDPs in high fetal bovine serum (FBS), which had a weaker growthpromoting activity than cs. The original procedures for LDP for maintaining the cells and the higher density assays for transformation are shown in Fig. 1. The results of maintaining the cells in LDPs with CS or FBS and assaying them periodically for focus formation are shown in Fig. 2. The transforming effect of cultivation in sera that slowed growth of nontransformed cells again raised the question of the role of selection in spontaneous transformation. The facility of obtaining spontaneous transformation in the NIH 3T3 cells prompted a broader investigation of the process, which is described later.
2. RAT EPITHELIAL CELLS The foregoing studies of selective conditions in spontaneous transformation involved the use of fibroblasts. Since most human neoplasms are of epithelial origin, it is of interest to examine the results obtained on the role of selection in the transformation of epithelial cells. Epithelial cells were obtained by dissociating rat liver, and their epithelial character was maintained in LDPs (Williams et al., 1971). The cells multiplied in small islands of closely adherent sheets; therefor, unlike fibroblasts, the cells were in contact with one another even in passage at low density. They multiplied in these islands at their maximum rate, which was slower than that of fibroblasts in LDPs. The rate of proliferation slowed at confluence of the islands with one another, and after 1 week at confluence foci of cells appeared that were smaller and more densely crowded than the surrounding cells. In this foundational epithelial study, cells from the rapidly growing cultures failed to produce tumors in syngeneic newborn rats or in irradiated weanlings (Williams et al., 1971). There was no report of inoculation into animals of cells from the confluent cultures that contained foci of densely crowded small cells. Diploid rat liver epithelial cells derived by the same method were later passaged weekly as they attained a confluent density (nonselective conditions) or were maintained at confluence for 3 weeks between monthly passages (selective conditions) (Lee et al., 1989). The capacity for tumor production
166
Harry Rubin
Fig. 1 Procedure for passaging cells and testing them for focus formation in primary and sec-
ondary assays. The large circles represent 21-cm2 plastic Petri dishes. The left column represents the standard 2- or 3-day low-density passages (LPDs) in either 10% calf serum (CS) or 10% fetal bovine serum (FBS). At intervals, a primary assay (1◦ ) was done in 2% CS by seeding 105 of the passaged cells or 102 – 104 of these cells together with 105 cells of a line that no longer produced foci in either 1◦ or 2◦ assays. The latter then formed a confluent lawn to display the focus-forming capacity of the minority. The cells were incubated 14 days with medium changes every 3 or 4 days. Some cultures were fixed and stained for transformed focus counts. Foci are represented as small circles; shading indicates degrees of progressive transformation. Others were trypsinized, counted for saturation density, and reseeded in a 2◦ assay at various cell densities, as in the primary assay. LDPs for subsequent experiments were done in larger (56-cm2) dishes, the cell concentrations were reduced five-fold, and only CS (10%) was used in order to reduce the chance of transformation before the 1◦ assays for focus formation at confluence. In addition, the serial assays were extended up to as many as six successive rounds. (Reproduced with permission from Rubin, H. Proc. Natl. Acad. Sci. USA. 1992;89(3):977–981. Copyright 1992, National Academy of Sciences, U.S.A.)
in young syngeneic rats was detectable under selective conditions in one-sixth of cell divisions required for the same effect under nonselective conditions. A similar relation between selective and nonselective conditions obtained for cells treated with a highly mutagenic carcinogen, although tumorigenic capacity appeared a few cell divisions earlier under both conditions than in the
Fig. 2 Effects of long-term passage in calf serum (CS) or Fetal bovine serum (FBS) on the morphology of foci in 2◦ assays. The total number of days in low-density passages preceding the 1◦ assay is shown between the two sets of dishes. Cells had been passaged in 10% CS (left) and in 10% FBS (right), but all were assayed in 2% CS. Note that all the 2◦ assays of the CS passages were of 105 cells, whereas only the earliest 2◦ assay of the FBS cells used 105 cells; the latter assays used only 103 + 105 nontransforming cells to form a confluent lawn. Regardless of the serum used in the passages, the cells of all the cultures shown were in medium of the same composition of 2% CS for two assays, or a total of 28 days, before fixing and staining. Thus, the differences in focus formation represented a stable difference produced by their prior passage history in 10% CS or 10% FBS. Apparently, there was selection against transformed and transformable cells during the low-density passage in 10% CS, whereas there was selection for such cells during the slower growth at low density in 10% FBS. Obviously, there is selection for transformed cells in the assays at high density in 2% CS. Focus formation was slower at higher density in FBS than in CS (not shown). (Reproduced with permission from Rubin, H. Proc. Natl. Acad. Sci. USA. 1992;89(3):977–981. Copyright 1992, National Academy of Sciences, U.S.A.)
168
Harry Rubin
untreated cultures under comparable conditions. The results indicate that selective conditions favor spontaneous transformation in epithelial cells just as they do in fibroblastic cells. In addition, they show that the selective conditions alone are far more efficient in eliciting transformation than carcinogen treatment under nonselective conditions (i.e., treatment with chemical carcinogen increased the rate of spontaneous transformation only slightly, and was a minor factor relative to the effect of selection). Spontaneously and chemically transformed rat liver epithelial lines have both common and differing features (Tsao et al., 1990). They both show extensive phenotypic heterogeneity in a variety of characteristics, including response to growth factors. Unlike most chemically induced tumor lines, the spontaneously induced lines do not exhibit the resistance phenotype to a variety of chemicals. However, they are, resistant to the growth inhibitory effects of transforming growth factor-β 1 (Haggett et al., 1991). The chemical resistance of most chemically induced lines and the resistance to a physiological inhibitor of the spontaneous lines suggest that these resistant phenotypes are generated selectively during the tumorigenic process. A comparison was made of the metastatic capacity of two lines of mouse melanoma cells inoculated into mice after harvest at confluence and at low density (Bosmann and Lione, 1974). The confluent cells produced two to four times as many metastases as the sparse cells. Unfortunately, there was no mention of how long the cells were kept at confluence or of growth rates of the cells, but the implication is that the same conditions that select for transformation of normal cells also select for metastatic progression of melanoma cells.
III. SELECTION IN TRANSFORMATION OF ESTABLISHED CELL LINES The results described previously left many questions unanswered. Transformation at high population density occurred when there was no net growth: Could transformation be occurring in the absence of cell multiplication? Selective conditions such as high population density must have variants to select in the population: Why were they often not observed in early assays at confluence? Was the assay method limiting in detecting preneoplastic variants? Could it be ruled out that high population density actually induced transformation, as a chemical carcinogen might do, rather than selecting preexisting variants? Given the progressive nature of neoplastic development, was it possible that there were preexisting variants that had to undergo several successive steps of progression before their transformed phenotype was expressed in focus formation? Could there be selection against
Role of Selection in Neoplastic Transformation
169
transformed cells in the exponential growth under the nonselective condition of LDPs? To what extent did the susceptibility to transformation change in passaging the cultures, and would this provide sublines of different capacities for transformation? These and other questions could only be approached by a more detailed and quantitative study of the dynamics of spontaneous transformation. The effectiveness of such studies required cell populations in which spontaneous transformation could be regularly produced in a quantifiable manner and in a relatively short time. The NIH 3T3 cells had these characteristics, although a full appreciation of the results obtained for answering the preceding questions was only realized after a decade of use.
A. Conditions for Eliciting and Expressing Spontaneous Transformation As noted earlier, neoplastic transformation in NIH 3T3 cells has been elicited under three conditions: (i) prolonged incubation under the contact inhibition of multiplication after cultures grow to confluence, (ii) successive LDPs of the cells in suboptimal concentrations of CS, and (iii) successive LDPs in FBS. It will be shown later that all these conditions are selective for transformed cells, which are operationally defined here as those with an increased capacity to multiply at confluence and in suboptimal concentrations of serum. The definition derives in part from the methods for assaying transformation, which is the production of foci of cells with increased capacity to overcome the contact inhibition of confluent cultures and its corollary of a significantly increased saturation density when the proportion of transformed cells in a population is high. To avoid the excessive repetition of the percentages 2, 5, and 10% CS used in the experiments, they will be referred to as low, intermediate, and high CS, respectively. The growth rate of nontransformed NIH 3T3 cells is at a maximum in 10% CS, with graded reductions at lower concentrations. The saturation density of the cultures is directly proportional to the concentration of CS in the medium. Progression of transformation is expressed in the increasing size and density of individual foci and increasing levels of saturation density (Fig. 3). Tumorigenesis is associated with the more advanced stages of transformation (Reznikoff et al., 1973; Rubin et al., 1990a). Dense multilayered foci among the original NIH 3T3 cells were first observed against a confluent background in high CS concentration at about 2 weeks after seeding 105 cells (Rubin and Xu, 1989). The foci increased in size with further incubation. Foci developed more slowly in equally high FBS and in low CS. Frequent LDP of the cells in high CS to minimize contact among cells and maximize growth rate resulted in a gradual loss of
Fig. 3 Progression of size, number, and density of transformed foci in four serial assays (1◦ – 4◦ , top to bottom). The saturation density (×10−5 cells) of each assay is shown alongside the appropriate dish. All assays were seeded with 105 cells. (Reproduced with permission from Rubin, H. Experimental control of neoplastic progression in cell populations: Foulds’ rules revisited. Proc. Natl. Acad. Sci. USA, 1994;91:6619–6623. Copyright 1994, National Academy of Sciences, U.S.A.)
Role of Selection in Neoplastic Transformation
171
focus-forming capacity, but parallel passage in high FBS increased focus formation (Rubin, 1992; Rubin and Xu, 1989) (Fig. 2), as did passage in low CS (Rubin et al., 1990b; Yao et al., 1990). Since the aim was to study both the origin and the progression of focus-forming capacity, an assay was developed using the low concentration (2%) of CS, in which few foci were seen in confluent cultures at 14 days in the early passages of cells and none were seen after they had undergone many LDPs in high CS. An assay procedure was adopted of serial 14-day assays of 105 cells designated 1◦ , 2◦ , 3◦ , etc. (Fig. 1) to observe progressive development of the transformed state. Cultures became confluent in 4 or 5 days and reached saturation density at about 7 days. The procedure allowed determination of saturation density at each passage dilution of cells in an excess of nontransformed cells if necessary to determine the number of transformed cells present at each time, and observation of the population density within individual foci as an indicator of progression toward autonomous growth. Although the use of low CS in LDPs is an obvious rationale for selecting transformed cells, the reason for their accumulation and progression in high FBS could only be surmised. It was found, however, that the nontransformed cells in LDP multiplied at about a 20% lower rate in high FBS than in an equal concentration of CS and had only half the saturation density (Yao and Rubin, 1992). It can be assumed, therefore, that the growth restraint during LDPs in FBS was selective for transformed cells, which are known to have a lower requirement for growth factors. It may seem paradoxical that high CS in the assay for transformation at confluence maximized the number of foci but LDP in high CS actually reduced and ultimately eliminated the number of cells that registered as foci in primary and even secondary serial assays for focus formation (Rubin, 1992). It can be understood as follows: Cells multiply to five times higher saturation density in high rather than in low CS (Yao and Rubin, 1992). The increase in cell division that occurs at confluence in high CS creates a greater opportunity for transformative genetic change. Even partial growth constraint at confluence favors selection of progressively transforming cells. Each transformed cell can also form a larger and more visible focus, given the additional rounds of multiplication. In the LDPs, however, high CS minimizes the selective advantage of the transformed cells. In fact, most clones of transformed cells multiply at a lower rate at low density when they are in high CS than do nontransformed cells (Grisham et al., 1988a; Grundel and Rubin, 1992; Reznikoff et al., 1973; Rubin et al., 1995; Smith et al., 1993). Therefore, it is not surprising that LDP in high CS leads to the disappearance of transformed and even easily transformable cells. The same logic can be applied to the apparent paradox of an increase in capacity for focus formation during LDPs in FBS (Fig. 2) with reduced focus formation in the assay in FBS than in CS (Rubin and Xu, 1989).
172
Harry Rubin
The effect of high population density is easily seen by growing cells in the standard 1◦ assay in low CS and transferring a few cultures every few days into a 2◦ assay for 14 days (Rubin and Xu, 1989). During the early stages of multiplication in the 1◦ assay, no focus formers are produced when the cells are transferred for a 2◦ assay, but once the 1◦ culture becomes crowded and approaches saturation density, selection begins and foci appear in the 2◦ assays in increasing numbers (Rubin and Xu, 1989). A critical test later showed it was the increase in population density and not the increase in cell number that was responsible for the appearance of focus-forming cells upon transfer of the cells for a 2◦ assay (Yao and Rubin, 1994). It is difficult to quantitate the relation between varying degrees of contact interaction between cells and transformation by simply relying on growth to confluence and a subsequent plateau of no net growth. However, cell growth can be initiated by parallel sets of graduated increments in seeding number in separate dishes and the cell number kept from excessive crowding by thrice weekly transfers (Yao and Rubin, 1992). The sensitivity of growth regulation to individual cell contacts is heightened by the presence of low CS. The percentage of cells with at least one contact with other cells is recorded daily. Change in the transformed state of the cells is determined every third passage by setting aside aliquots to measure saturation density. The saturation density of all the populations increased at each successive measurement, but it increased more rapidly at the higher than at the lower seeding densities in proportion to the number of contacts between individual cells. This demonstrates a remarkable sensitivity of transformation to selective conditions and suggests extensive heterogeneity of selectable variants in the population. The extent to which the variants are transient (physiological) or stable (genetic) will be considered later. The effect of different degrees of growth regulation on the transforming process is illustrated by a comparison of the effects on saturation density of cells grown under moderate regulation by LDPs in low CS versus those grown to high density in the same concentration of CS and maintained in a stationary state without passage (Rubin et al., 1990b). The saturation density of cells in LDP with low CS increased with successive passages to a level more than five times higher than that of the starting population (Fig. 4). The saturation density of cells sampled from the population initially maintained at high density without passage increased at about the same rate as that of cells from the LDPs but reached a maximum only twice that of the starting population a few days after net growth ceased. Those maintained by LDP in high serum showed no increase in saturation density. The implication is that continuous multiplication under growth-moderating conditions is a more effective selective condition for transformed cells than is the more stringent cessation of net multiplication under the combined effect of low CS and confluence. Confirmation of this conclusion came from counting transformed foci in cells serially assayed after successive postconfluent passages (Rubin, 1994c)
Role of Selection in Neoplastic Transformation
173
Fig. 4 Changes in saturation density of cells passaged frequently in high or low concentrations of calf serum or maintained without passage in a low concentration of calf serum. NIH 3T3 cells which had been maintained at their maximal growth rate by low-density passages every 2 or 3 days were seeded in the amount of 105 cells into each of many 21-cm2 culture dishes in medium with 10% (A) or 2% (B, C) calf serum and treated as follows: (A) passaged at 2- or 3-day intervals in 10% calf serum. At each passage 105 cells were seeded into 20 dishes in 2% calf serum for growth curves and saturation densities; (B) passaged at 2- or 3-day intervals in 2% calf serum with seeding of 105 cells as in A in 2% calf serum for growth curves and saturation densities; (C) cells were maintained without passage in 2% calf serum. On the same days that A and B were passaged, some of the C cultures were trypsinized to set up growth curves in 2% calf serum. The day on which the cells were transferred to set up growth curves is shown next to the appropriate symbol in B and C. The corresponding curves in A are effectively indistinguishable from one another, with the cells having undergone no change on frequent passage in 10% calf serum. (Reproduced with permission from Rubin, H. Selected cell and selective microenvironment in neoplastic development. Cancer Research, 2001;61: 799–807.)
versus cells kept at confluence without passage (unpublished part of (Rubin, 1994c). Foci were produced in replicate cultures in the fourth round of the serial assays at a time when none could be seen in the nonpassaged culture. Therefore, measurements of increased saturation density and focus-forming capacity are more efficient in suboptimally multiplying cells than in those in which multiplication has been more strongly suppressed. The implication is that the selective conditions must allow some continued albeit reduced cell multiplication for genetic change to occur and be amplified by selection.
174
Harry Rubin
Since the action of tumor promoters is reversible, it is generally accepted that their effects are produced by selection of initiated (mutated) cells (Berenblum, 1974). This conclusion is reinforced by the finding that the classic promoting agent 12-O-tetradacanoyl-phorbol-13-acetate (TPA), also known as phorbol myristate acetate, does not enhance neoplastic transformation of NIH 3T3 cells if applied continuously when the cells are maintained under the nonselective conditions of exponential multiplication in LDPs (Rubin and Rubin, 1994). A pronounced promoting effect is produced by TPA in sublines that are relatively resistant to neoplastic transformation under the selective condition of contact inhibition at confluence. An antitransforming effect of TPA is produced when it is added to a subline of NIH 3T3 cells that is primed for rapid transformation at confluence (Rubin, 1995). This suppression of transformation apparently occurs because TPA overrides the selective, inhibitory effect of high population density. Thus, the results in cell culture are consonant with the conclusion derived from in vivo studies that carcinogenesis is not augmented when it is already operating under optimal conditions, and it is not inhibited when operating at minimal efficiency (Berenblum, 1974). Recent evidence that promoters act in a selective manner in vivo is considered later. Progressive transformation is driven to extremes by culturing cells in stepwise reductions of concentration from 10 to 0.25% CS (Yao et al., 1990). The progressive change results in a capacity to produce transformed foci at confluence at the lowest concentration of CS that does not occur in single stepdown from the highest to the lowest concentration of CS. Evidently, the NIH 3T3 cells have an almost unlimited capacity to adapt to reductions in the availability of growth factors—a property associated with progressive increases in neoplastic behavior. In this sense, they differ from the Swiss 3T3 line of mouse cells, a few of which multiplied for several generations when switched from high to low (1%) CS but the property was not heritable in the long term (Brooks et al., 1984). However, the minority population of cells with a physiological capacity for some multiplication under selective conditions is the most likely to generate mutations under these conditions and therefore to become transformed. The capacity of the NIH 3T3 cultures to develop long-term heritable capacity to multiply at high density apparently accounts for the ease of their transformation under selective conditions.
B. Interclonal Heterogeneity of Competence for Transformation The appearance of foci of increasing density under selective growth conditions implies cell populations that continuously generate variants with the capacity to overcome contact inhibition and ultimately to produce tumors. Such variability was observed among clones obtained from NIH 3T3 cell
Role of Selection in Neoplastic Transformation
175
populations that had been passaged at low density and cloned with various in CS concentrations at both stages (Rubin et al., 1990b). Very few clones obtained from cultures passaged many times in high CS and cloned in high CS produced any foci at confluence in 1◦ assay, and the few foci were very small and lightly stained. In contrast, all clones from cultures passaged and cloned in low CS produced foci, mainly large and dense, but the number of foci varied widely from clone to clone. Some produced thick multilayered confluent cultures from multiple overlapping foci, others had numerous discrete foci or a few lighter foci. Clones from cultures that had been switched from high to low CS or the reverse were intermediate in focus formation but also heterogeneous in this property. After more extensive LDPs in high CS, 1◦ assays from among many parallel passages of the uncloned cultures produced only light foci (Grundel and Rubin, 1991). All the 2◦ assays from each 1◦ assay produced mixtures of light and dense foci, but the numbers and ratios of each type varied significantly among the assays. When the original culture was divided into groups of 1000 cells or less, and each group was grown to large numbers, the 1◦ assays revealed little transformation but the 2◦ assays had extremely wide ranges of focus formation, from many dense foci to light foci or no foci at all. This indicated a very low proportion of cells that could progress rapidly to multiple dense focus formation in two serial assays at confluence and also great heterogeneity in susceptibility to transformation. Clearly, the starting population presented great diversity for selection under the growthconstraining conditions of low CS and contact inhibition at confluence. The high rate of variation in transformed clones is seen in differences in focus formation among subclones (Rubin et al., 1992). Not only do the subclones differ from one another in numbers of foci produced but also they differ in size and density of the foci. In contrast, the foci within a culture from any subclone are very similar. Independent transforming events from the same starting population can be isolated by repeated splitting of cultures in serial rounds of confluence. At the 5◦ assay of such a procedure, all sublines produced large, well-defined, dense foci, similar to each other within a subline but distinctly different from one subline to another (Fig. 5) (Rubin, 1994b). All had progressed from light focus formation in the earlier assays in which no distinction could be made among the sublines. The results indicate either that the cell that was the site of each independent transformation differed before the final event that led to dense foci or the final event differed in each transformation. Whatever the explanation, the results illustrate the high degree of variability in the population. The capacity for focus formation of any clone has some stability in repeated passages, but it inevitably changes over many LDPs in high CS. Both aspects were illustrated with clones made soon after thawing the original frozen stock of NIH 3T3 cells which had a high sensitivity to spontaneous transformation
176
Harry Rubin
Fig. 5 Differences in morphological detail of dense foci from parallel lineages. The A and B lineages were started from the same LDP, the C lineage was started from the 1◦ assay of B, and the D lineage was started from the 2◦ assay of C. The photographs are segments of cultures from the 5◦ assay of each lineage. Cultures of A, C, and D were seeded with 104 cells from the lineage shown plus 1.8 × 105 nontransformed cells for background. The saturation densities (×10−5 cells) shown in parentheses for A, C, and D were determined on cultures seeded with 1.8 × 105 cells of the same lineage with no background cells (not shown). Culture B was seeded with 1.8 × 105 cells of the lineage, and the saturation density is that of the sister culture to the one shown. The saturation density of controls in the 1◦ assay was about 5 × 105 cells per culture. The photograph shows that a group of foci arising from an independent transforming event in one lineage can be morphologically distinguished from a group arising independently in a parallel lineage. The foci in any single culture arising from a single event in a lineage are morphologically alike. The heightened saturation densities quantitatively reflect the increased capacity of the transformed cells to multiply at high density. (Reproduced with permission from Rubin, H. Experimental control of neoplastic progression in cell populations: Foulds’ rules revisited. Proc. Natl. Acad. Sci. USA. 1994;91:6619–6623. Copyright 1994, National Academy of Sciences, U.S.A.)
(Chow and Rubin, 1999a). The uncloned culture made broad, light foci in 1◦ assay; dense foci appeared in the 2◦ assay. One clone from a total of 29 clones isolated from the parental culture made a few dense, pinpoint foci in the 1◦ assay and many large, dense foci in the 2◦ assay (Fig. 6). Even more of its cells were focus formers in subsequent assays. It repeated this behavior in periodic tests made over 21 LDPs but lost the capacity for making dense, pinpoint foci in the 1◦ assay by the 37th LDP, though large, dense foci continued to appear in the 2◦ assay (Chow and Rubin, 1999b). The results indicated that some cells in the clone were selected for progression from light to dense focus formation toward the end of the 1◦ assay, and similar cells were
Role of Selection in Neoplastic Transformation
177
Fig. 6 Progressive transformation of diverse individual clones in serial assays at confluence. The numbers on the left are the serial assays. The assays shown are of 105 cells unless there is an attached number (500, 103, or 104) representing the clonal cells seeded together with 105 nontransformed cells that make a flat confluent background to display the foci. Clone 1A regularly produced one or two tiny dense foci in the 1◦ assay, indicating an advanced stage of transformation that occurred without intermediate stages late in each 1◦ assay. Clone 6H showed no foci in the 1◦ assay and many light foci in the 2◦ assay, some of which progressed to dense foci in later assays. Clone 4B was refractory to formation of any but a few very small but light foci and then only in the 5◦ assay. The group of three clones illustrates the great heterogeneity of transformability of cells derived from the same uncloned parental culture and, therefore, of their selectability under growth constraint. (Reproduced with permission from Chow, M., and Rubin, H. Clonal dynamics of progressive neoplastic transformation. Proc. Natl. Acad. Sci. USA. 1999;96;6976–6981. Copyright 1999, National Academy of Sciences, U.S.A.)
retained through 21 LDPs. Further passage selected against them, however, although other cells in the population retained a capacity for progression at only a slower rate. Most other clones obtained from the original population produced no foci in the 1◦ assay but did produce increasing numbers of light foci in serial assays with later progression to dense foci (Fig. 6). These clones tended to maintain this behavior with some variation through many LDPs, but one of them completely lost the capacity for any transformed behavior (Chow and Rubin, 1999b, 2000b). Unexpectedly, one clone switched between 41 and 56 LDPs to producing dense foci in the 1◦ assay. Again, great diversity was seen in the growth behavior of clones from the same parental culture. Long-term trends in focus-forming ability during LDPs follow certain patterns but such behavior is not invariant (Rubin, 1992, 1993a).
178
Harry Rubin
The precise genetic basis of variability in the cultures has not been established, but it is known that there is a high degree of karyotypic heterogeneity among cells in the same culture. The stemline of chromosome number changes during serial passage of the NIH 3T3 cultures, as does the average number of marker chromosomes (Rubin, 1993b). When individual chromosomes were identified by G banding in three cells at each of three different passage levels, every one of the nine cells examined proved to be karyotypically unique. This chromosomal instability, however, is not unique to the NIH 3T3 cell or to its transformability since similar instability was found in the BALB 3T3 cells, which are difficult to transform (Rubin et al., 1984). Furthermore, sublines of NIH 3T3 cells have been developed which are relatively resistant to transformation (Chow and Rubin, 1999a,c), and one clone exhibited little or no sign of transformation in extensive tests after it had been through numerous LDPs (Fig. 6) (Chow and Rubin, 1999b). It therefore appears that transformability is a complex function of many genes and is difficult to predict from mutational analysis of only a few genes.
C. Subliminal, Selectable Stages of Transformation Most studies of transformation have treated it as an ungraded unit phenomenon whether using tumor production, saturation density, or focus formation as criteria for transformation (Aaronson and Todaro, 1968a; Earle, 1943a; Grisham et al., 1988b; Sanford et al., 1950). It was recognized that there were grades of focus formation with regard to size and density (e.g., I–III) (Reznikoff et al., 1973), or light and dense (Rubin and Xu, 1989), but the actual counting was usually based on large dense foci because they were easy to quantitate and were tumorigenic. It is well-known, however, that cancer is progressive, passing through many stages, some of which are clinically undetectable (Foulds, 1969). The presence of visually undetectable stages in spontaneous transformation (and even in chemically or physically induced transformation) was apparent from a prolonged one-step assay in which it could be assumed that a long-term process had to occur before any phenotypic evidence of neoplastic behavior could be detected (Haber et al., 1977). The multistep serial assay method at relatively short (2-week) intervals showed that there was progression from light to dense focus formation within cell lines. This, however, said nothing about any subliminal stage that might have preceded light focus formation, and it left unanswered the question of what was being selected if many LDPs gave 1◦ assays with no sign of even the lightest foci (Rubin, 1992). The problem of subliminal selectable stages was resolved in an experiment combining focus formation and saturation density in serial assays in which the main variable was the concentration of CS used in the 1◦ assay (Rubin, 1994c). The procedure is outlined in Fig. 7. The starting culture, using the
Role of Selection in Neoplastic Transformation
179
Fig. 7 Flow chart of 1◦ assays in 2, 5, and 10% CS followed by 2◦ , 3◦ , and 4◦ assays of each lineage in 2% CS. The number 4 at the top indicates that each set was done with four quadruplicate lineages to assess the degree of internal variation in focus formation and saturation density within each original CS concentration. The quadruplicates were also used to evaluate the variation between sets differing in CS concentration in the 1◦ assay. See Fig. 8 for a sampling of results from the assays. The large circles represent 21-cm2 Petri dishes, the small circles represent foci of different sizes, and degrees of progression are represented by darkness of shading. (Reproduced with permission from Rubin, H. Incipient and overt stages of neoplastic transformation. Proc. Natl. Acad. Sci. USA. 1994;91:12076–12080. Copyright 1994, National Academy of Sciences, U.S.A.)
180
Harry Rubin
standard low CS (2%) in all assays including the 1◦ assay, showed no little sign of transformation until the 4◦ assay (Fig. 8). However, the saturation density of the cultures increased significantly in each successive assay, indicating that many cells had incremental capacities to overcome contact inhibition without producing discrete, visible foci. When higher concentrations of CS were used in the 1◦ assay, the saturation density was proportionally increased. Indeed, in high (10%) CS, many dense, pinpoint foci were visible in the 1◦ assay, but the 2◦ assay, which downshifted to low CS, had only very small, light foci uniformly distributed over the whole culture. In addition, the saturation density was higher in the 2◦ assay derived from the 1◦ assay in high CS than in the 2◦ assay derived from the 1◦ assay in low CS. There was a progressive increase in saturation
Fig. 8 Increasing the concentration of CS in a 1◦ assay increases the size and density of transformed foci produced in remote assays in a constant CS concentration. A 1◦ assay (not shown) was done in 2, 5, or 10% CS (left, middle, and right columns, respectively). Each group was downshifted to 2% CS in the following 2◦ , 3◦ , and 4◦ serial assays (see Fig. 7). Significant formation of foci was not seen until the 4◦ assay in which their size and density increased with the concentration of CS present in the 1◦ assay 4 weeks earlier. Since the number of cells at confluence in the 1◦ assay was proportional to the CS concentration, the result shows that transformation progressed with the number of cell divisions that occurred under the selective condition of confluence, but the transformation did not manifest in discrete, tumor-like foci until 4 weeks after all the cultures had been downshifted to the same low concentration of CS. It suggests that most neoplastic development usually occurs before clinically apparent tumors can be detected. (Reproduced with permission from Rubin, H. Incipient and overt stages of neoplastic transformation. Proc. Natl. Acad. Sci. USA. 1994;91:12076–12080. Copyright 1994, National Academy of Sciences, U.S.A.)
Role of Selection in Neoplastic Transformation
181
density in all groups in the later assays, but the values were always correlated with the CS concentration used in the 1◦ assay. This showed that the advantage gained by high serum in the 1◦ assay was heritable. The most plausible explanation is that there was selection for cells with small increments in genetic capacity to overcome contact inhibition in the 1◦ assay, but there were more cycles of selection in high CS 1◦ assays because of the additional rounds of cell replication afforded by the increased concentration of growth factors. The progressive increase in saturation density through the 3◦ serial assay of cultures started from the 1◦ assay in low CS showed unequivocally that selection can precede the appearance of discrete transformed foci. When foci did appear in the 4◦ assay, they were larger and denser in the cultures originating in high CS than in low CS, reinforcing the selective and heritable nature of changes in the 1◦ assays. These conclusions were further reinforced in groups of 1◦ assays given an extra week in different CS concentrations but otherwise using same procedure of low CS for 2 weeks in all subsequent assays. Progression as measured by focus formation and saturation density was accelerated in all groups compared to those with only 2 weeks in 1◦ assay described previously. This indicated that selective growth continued between the second and third weeks of the 1◦ assay and drove further progression of the transformed state. The increases in saturation density (before the appearance of substantial foci) were highly reproducible among the quadruplicate lineages in each group. The dense foci, however, first appeared at different times and in different sizes among the lineages. This made for sharp differences in saturation density among parallel lineages. The results suggested that there was considerable diversity in growth properties among cells in the starting population, with equal numbers of selectable cells when the culture was subdivided for the 1◦ assays in varying concentrations of CS. This would account for the uniform cell sheets and the reproducible increases in saturation density among lineages of the same CS concentration in the early assays. The progression to large, dense focus formation, however, probably occurred in single, random events at a low rate characteristic of mutations.
D. The Role of Selective Clonal Expansion in Transformation Models of the clonal evolution of tumors have stressed the importance of acquired genetic lability as an early, if not the first, step in that evolution (Nowell, 1976). It is clear that certain inherited mutations involving defective DNA repair, such as HNPCC, lead to genetic lability as an early step in tumor development when the normal allele of the gene mutates (Kinzler and Vogelstein, 1996a; Tomlinson et al., 1996). However, mathematical
182
Harry Rubin
models predict that genetic lesions leading to clonal expansion precede genetic destabilization in the majority of solid tumors (Tomlinson and Bodmer, 1999; Tomlinson et al., 1996). The role of selective clonal expansion in neoplastic development can be readily examined in cell culture by isolating and expanding clones from a polyclonal parental culture and comparing the course of transformation in the clones with that of the parental culture by subjecting all to serial rounds of selection at confluence. A suggestion that clonal expansion might play a dominant role came from clonal expansion of the original transformation-sensitive line of NIH 3T3 cells (Chow and Rubin, 1999a). One of the expanded clones consistently made small, dense foci in 1◦ assays at confluence, whereas the parental culture made no dense foci until the 2◦ assay. The results were not impressive because a 1-week extension of the parental 1◦ assay produced dense foci within the light ones (Chow and Rubin, 1999b). Clonal studies in the transformation-refractory subline of NIH 3T3 cells gave a much clearer demonstration of the role of clonal expansion in transformation (Fig. 9). The parental culture exhibited no sign of transformation in five serial assays at confluence and only a slight sign of irregularities in the sixth serial assay (Chow and Rubin, 2000a). However, about one-fifth of 25 clones consistently exhibited moderate focus formation from the fourth serial assay onward (Chow and Rubin, 2000a). This seems to be a surprising result since the parental culture should have about 2 × 104 cells from the productive clones in the regular seeding of 105 cells for assay. However, it could be easily explained since the parental culture had only one-fifth the number of productive cells to select from at confluence as each of the individual productive clones had under the same condition, given that the saturation density was approximately the same for the parental culture as for the isolated expanded clones (Chow and Rubin, 2000a). Hence, the parental cultures would not be expected to produce moderate foci until far beyond the sixth and final serial assay of the experiment. One remote alternative is that the isolated clones were much more unstable than the polyclonal parental culture. No evidence of this possibility was found in an experiment in which six isolated clones from LDPs of the transformation-sensitive line were mixed and then cocultured in successive assays at confluence (Chow and Rubin, 2000b). Half the clones produced no dense foci in three serial assays, and the other half produced varying numbers of well-developed dense foci by the second assay. However, the mixture of all six clones produced no dense foci through the second assay and hundreds of times less than the average of the productive clones in the third assay. Since dense focus production originates in a single cell (Ellison and Rubin, 1992), and the first of these appeared in the second assay of the clones, it would have taken an average of four assays for them to appear in the mixed population in which productive cells formed only half of the population. As with the transformation-refractory subline, the discrepancy in focus-formation could
Role of Selection in Neoplastic Transformation
183
Fig. 9 Flow chart of clonal isolation from a relatively refractory subline of cells and serial assays of each clone and of the uncloned parental culture. Only 5 clones are shown, although 25 were run through the same procedure. The basic result was that about one-fifth of the clones exhibited moderate focus formation at the 3◦ or 4◦ assay, one-fifth had smaller foci in the later assays, and the rest had few or no foci through the 6◦ assay. The uncloned parental culture showed no significant focus formation in any assay, with only a few very small light foci in the last assay. The results indicate that focus formation is correlated with the minority proportion of selectable, transformable cells in a culture. See Chow and Rubin (2000a) for details and photographs of the relatively weak foci produced by the refractory line of NIH 3T3 cells which conclusively revealed the role of selection and clonal expansion in neoplastic transformation. (Reproduced with permission from Chow, M., and Rubin, H. Cancer Research, 2000;60:6510–6518.)
be most simply explained by the differences in numbers of potential focus forming cells available for selection at confluence. A variation of the clonal coculturing experiment displayed the dominating effect of the most highly transformed cells under selective conditions. Six clones were mixed after the 1◦ assay at a time that half of them would be
184
Harry Rubin
expected to produce dense foci in the 2◦ assay (Chow and Rubin, 2000b). In this case, dense foci did appear in a parallel assay of the mixture, and their number was about the average of the numbers in the six clones. The original progression of a single cell to dense focus formation had already occurred in three of the clones and had been expanded to many dense focus formers before the mixture was made. The novel feature of the experiment was that the number and size of the foci in the 4◦ assay of the mixture were like those of the individual clone that produced the largest, densest foci, indicating that selective outgrowth of the focus-forming cells from that clone had led to their dominance in the three rounds of confluence following the mixture. The foregoing results show that selective clonal expansion is a major factor in spontaneous neoplastic transformation in culture. The importance of selective conditions had previously been demonstrated in spontaneous transformation of cultured rat liver epithelial cells (Lee et al., 1989). An additional aspect of the experiments with the epithelial cells was the effect of treatment by a powerful carcinogen: The carcinogen induced only a slightly earlier transformation than the untreated culture under selective conditions and a statistically insignificant acceleration of the delayed transformation that occurred under nonselective conditions. Perhaps the most striking aspect of the comparison was that the untreated cultures became tumorigenic under selective conditions in one-sixth the number of cell divisions that were required for the carcinogen-treated cultures to become tumorigenic under nonselective conditions. The sheer magnitude of this difference reinforces the major role of selective forces acting on spontaneous mutations as well as any increase in mutations produced by the carcinogen. These results do not rule out some contribution from genetic damage during prolonged incubation of cells under the combined inhibitory effect of contact inhibition and low CS. Populationwide heritable damage to cells often accompanies these conditions, as indicated by death and detachment of a fraction of the population and an inherited decrease in growth rate of the surviving cells on repeated LDP (Rubin et al., 1995). Such damage is usually caused by large-scale deletions and chromosome rearrangements (Chow and Rubin, 1999c; Hozier et al., 1985). These lesions could lead to loss of tumor suppressor genes and overexpression of oncogenes, which would increase the risk of transformation.
IV. THE CONTRIBUTION OF APOPTOSIS TO SELECTION IN NEOPLASTIC TRANSFORMATION A few days after the NIH 3T3 cells came under the influence of contact inhibition at confluence, many cells (most of which were incapable of replication) were seen floating in the medium (Rubin et al., 1995). Although they
Role of Selection in Neoplastic Transformation
185
represented only about 1% of cells in the culture, they indicated that confluence was not only inhibiting multiplication but also damaging cells. Microscopic examination of the cells remaining in the attached monolayer revealed some nuclei with marginated chromatin typical of early stage apoptosis (Wyllie, 1980). The growth rate of the cells from the monolayer given a period of recovery in LDP and passaged again was about 15% lower than that of cells which had not been subject to prolonged confluence. When seeded for cloning at very low density, their colonies had fewer and more scattered cells than controls and contained varying numbers of giant cells (Fig. 10). This suggested that the cells had undergone genetic damage that might destabilize
Fig. 10 Heritable damage in transformable cultures after three rounds of selection of transformed cells at confluence. Cells were seeded at 100 cells per culture for colony formation in 10% CS and incubated for 5 days. Clones seeded from LDPs are shown on the left; low power on top (scale bar = 400 µm), and higher power on bottom (scale bar = 100 µm). Clones on the right are derived from the 3◦ assay of the cells at the same magnifications as that of the controls. The resulting transformed cells initially multiplied more slowly than control cells during colony formation in 10% CS. The presence of many giant cells is a prominent feature of the transformed colony at the lower right, as is the abnormality of most of the other cells. Figure 11 shows that such cells have a growth advantage at higher density in low CS concentration. (Reproduced with permission from Rubin, H., Yao, A., and Chow, M. Heritable, populationwide damage to cells as the driving force of neoplastic transformation. Proc. Natl. Acad. Sci. USA. 1995;92:4843– 4847. Copyright 1995, National Academy of Sciences, U.S.A.)
186
Harry Rubin
the population and contribute to an increase in cell transformation. However, the accelerated transformation in minority susceptible clones over that in the uncloned parental population, as described previously, showed that genetic destabilization was not a major force in transformation (Chow and Rubin, 2000a). The damage and death at confluence of those cells most sensitive
Fig. 11 Initial growth disadvantage of transformed colonies in high CS concentration followed by selective advantage when downshifted to low CS concentration at 6 days. Cells from LDPs (controls) or from two transformed cultures (after a 2-day recovery from a 4◦ assay) were seeded at 50 cells per dish in 10% CS and incubated for 6 days, at which time some cultures were fixed and stained and the remaining cultures were downshifted to 2% CS for 3 and 9 more days of incubation before fixation and staining. C, controls; #6 and #8, two transformed lines; 6 d, 6 days in 10% CS; 6d + 3d and 6 d + 9 d, 6 days in 10% CS plus 3 or 9 days in 2% CS. There were only one-fifth as many cells at 6 days in the transformed cultures as in the control cultures, but there were about 10 times as many at 9 days after the downshift in CS concentration. (Reproduced with permission from Rubin, H., Chow, M., and Yao, A. Cellular aging, destabilization and cancer. Proc. Natl. Acad. Sci. USA. 1996;93:1825–1830. Copyright 1996, National Academy of Sciences, U.S.A.)
Role of Selection in Neoplastic Transformation
187
to contact inhibition would favor selection of transformed cells which, by definition, are favored at high population densities or at low serum concentrations. The combination of slower growth of transformed cells at low density in high CS concentration and more prolific growth at higher densities in low CS concentration is shown in Fig. 11. The control colonies appear after 6 days in 10% CS, when the transformed colonies are not yet visible. When switched to 2% CS at 6 days, the control colonies spread and flatten out, showing little sign of further growth, whereas the transformed colonies proliferate into dense, multilayered populations. Apoptosis might be considered an expression of terminal differentiation in confluent NIH 3T3 cultures. A more obvious example of terminal differentiation is seen in basal epidermal cells (keratinocytes) cultured in physiological concentrations of calcium (Kulesz-Martin et al., 1980). A small fraction of the population is resistant to terminal differentiation and is considered to be in the first stage of transformation because the fraction is increased by carcinogen treatment. The extent of that increase after exposure to various carcinogens is correlated with their capacity to initiate tumor development when painted on mouse skin in vivo (Kilkenny et al., 1985). However, there is also a relation between the carcinogenicity of the agents and their toxicity, especially evident among isomers of polycyclic aromatic hydrocarbons; therefore, it is not unlikely that they are acting as selective agents for spontaneous variants that are resistant to terminal differentiation, just as confluence does in transformation of liver epithelial cells (Lee et al., 1989). Although a contribution from mutagenesis by the carcinogens in terminally resistant keratinocyte colonies cannot be ruled out, their significant presence in untreated cultures and the dominant role of selection in the liver epithelial system suggest that selection also plays a major role in the epidermal system.
V. INHIBITION OF GROWTH OF TRANSFORMED CELLS BY SURROUNDING NONTRANSFORMED CELLS Selective clonal expansion multiplies the opportunity for successive mutations to further progression in a clone already exhibiting a growth advantage. Inhibition of the expansion by surrounding cells would of course reduce the chance of progression. Such inhibition was first described for virustransformed cells in the midst of nontransformed cells (Rubin, 1960; Stoker, 1964; Weiss, 1970). Growth inhibition of the transformed cells correlated with their junctional communication with normal cells (Mehta et al., 1986). Similar interactions were demonstrated between spontaneously transformed NIH 3T3 cells and their nontransformed counterparts, but the inhibitory capacity of the latter varied widely in different sublines from the same source
188
Harry Rubin
(Rubin, 1994a). Although there was a suggestion that the inhibitory capacity of the nontransformed cells was related to their relative resistance to transformation, this was never established. The fact that such inhibition has been demonstrated in many combinations of transformed and normal cells, however, indicates that the cellular microenvironment can modulate expression of the neoplastic phenotype. The role that such factors play in tumor development in vivo is considered later.
VI. CONFOUNDING EFFECTS OF VARIABLE CELL BEHAVIOR ON THE DYNAMICS OF TRANSFORMATION Several aspects of cell behavior masked the true nature of the transforming process. First, was the continuing variation in the transformability of the cultures as they were maintained by LDP. This is exemplified by the gradual loss of transformability by LDP in high CS and by variations in transformability during passage with slight differences in cell density (Rubin, 1993a). As a result, experiments would give only a hint of the sustained transformation-accelerating effects of a single round of confluence in high CS that could only be resolved by many repetitions of the experiments until unequivocal results could be obtained (Rubin, 1994c). Only then was it apparent that neoplastic progression was occurring even before there was any sign of focus formation. The reason for the usual selection against transformation in LDP only became obvious with the finding that transformation was usually accompanied by a reduced growth rate of cells at low density (Rubin et al., 1995). The role of expansion of productive clones in progressive transformation was hinted at in experiments with the transformationsensitive line of NIH 3T3 cells but became crystal clear only with the more refractory subline (Chow and Rubin, 1999a, 2000a). Although the NIH 3T3 cell culture system provides quantitative results of the dynamics of selection in neoplastic transformation, an understanding of its continuing variation during repeated passage is required for a meaningful interpretation of these results.
VII. SUMMARY OF THE MAJOR FEATURES OF SPONTANEOUS TRANSFORMATION The following are the major features of spontaneous transformation: First, spontaneous neoplastic transformation occurs in cultures over a period of months in newly explanted rodent cells or established nontransformed lines
Role of Selection in Neoplastic Transformation
189
of rodent cells. The newly explanted cells presumably become established as a permanent line before they exhibit transformed behavior. In recent years, the most commonly used cells have been the established NIH 3T3 mouse fibroblasts and an established line of rat liver epithelial cells. Second, the indicators most commonly used for transformation have been (i) tumorigenesis by cells inoculated into syngeneic mice or athymic mice, (ii) increased saturation density of confluent cultures, and (iii) formation of densely populated or multilayered foci against a monolayered confluent background of nontransformed cells. Third, the transformation is greatly accelerated under suboptimal growth conditions, such as (i) long-term incubation of cells under contact inhibition at confluence, (ii) LDP in reduced concentrations of calf serum, and (iii) LDP in standard concentrations of fetal bovine serum. If the selective conditions are too stringent (e.g., too low a concentration of serum or a combination of low CS and confluence) the rate of transformation is reduced: there is an optimal zone for selection that combines some multiplication with constraint. Fourth, frequent subculture (“passage”) of the cells at LP density in maximally stimulating concentrations of calf serum usually results in a gradual decrease in their capacity to produce transformed foci when grown to confluence. It then requires repeated rounds of incubation at confluence to elicit transformed foci. However, this partially refractory state of the cells brings out the progressive nature of the transformation from barely detectable overgrowth to small, light foci and then large, dense ones. Only the latter induce tumors in nude mice in less than 1 month. Fifth, in a transformation-sensitive culture, a mixture of light and dense foci is seen in the first or second round of prolonged confluence. There is selection at confluence of the cells that produce the largest and densest foci so that a single type of dense focus will predominate after several rounds of prolonged confluence. In a subline that produces large, well-defined foci after several rounds of confluence, parallel lineages from the same source undergo single independent transforming events in each lineage. In some experiments, foci in the same lineage are morphologically similar, and foci in parallel lineages are morphologically different. This indicates either that the cell undergoing transformation in each lineage was originally genetically different from the cells transforming in the other lineages or that the later transforming events were different. Sixth, populations of cells originating from single cells of a polyclonal parental culture and cultured apart from one another exhibit variation in the timing of focus formation and morphological appearance of foci when assayed in serial rounds of confluence. This indicates great genetic heterogeneity among cells of the parental culture in the capacity to undergo transformation.
190
Harry Rubin
Seventh, in a particularly informative experiment using cultures that did not produce large, dense foci in the first few rounds of confluence, there was diffuse selective growth at confluence of many cells that was detected as a uniform multireplicate increase in saturation density. Large, dense foci arose in later rounds of confluence in a random fashion, leading to large differences in saturation density between replicate cell lineages. The time of appearance and size of the foci were related to the total number of cell divisions under selective conditions. Eighth, a minority fraction of clones derived from a polyclonal culture that was relatively refractory to transformation produced transformed foci before the parental culture produced any. Each of the minority clones had many more cells competent to undergo transformation under selective conditions than the parental culture, and each therefore produced recognizable foci before the parental culture. None of the clones produced foci before the third round of confluence, indicating that subliminal stages of transformation were occurring before their morphological expression. The results highlighted the importance of selective clonal expansion in expression of the transformed state. Nineth, spontaneous neoplastic transformation results from progressive selection of genetic variants that are continuously generated in culture. The selective conditions are those that impose physiological limits on growth rate and the variants are heterogeneous in their capacity to overcome the limits. There is no evidence that the selective conditions increase the ongoing rate of variation. Finally, the earliest stages of transformation involve selection of spontaneous variants that progress to full neoplastic behavior only under selective conditions. The analysis of spontaneous transformation in culture raises the following questions about carcinogenesis in the organism: How much weight should be given to selection versus increased mutation rate? What are the selective conditions that promote tumor development? What are the frequency and type of variant cells that are selected? These questions are considered next.
VIII. EVIDENCE FOR SELECTION IN EXPERIMENTAL AND HUMAN CANCER Studies of chemical carcinogenesis of the liver of rats suggested an important role of selection in tumor development. The appearance of microscopic foci of enzyme-altered cells is an early step in various models of liver carcinogenesis (Campbell et al., 1986). These are thought to progress to macroscopic
Role of Selection in Neoplastic Transformation
191
nodules, most of which eventually remodel to normal-appearing areas, but a few of which progress to carcinoma (Enomoto and Farber, 1982; Squire and Levitt, 1975). Between 10,000 and 50,000 enzyme-altered microscopic foci appear following a regimen of initiation and promotion, depending on the promoter used and the age and sex of the rat at the time of treatment (Xu et al., 1990). About 1000 of these grow into nodules, and only a few nodules become carcinomas (Farber and Sarma, 1987). Careful examination of the liver of untreated rats revealed the presence of enzyme-altered foci, with the number increasing sharply with age to a maximum of about 100 (Ogawa et al., 1981; Schulte-Hermann et al., 1983). The percentage of hepatocytes synthesizing DNA is 5–10 times higher in the foci than in the normal surrounding areas (Schulte-Hermann et al., 1983), indicating their selective growth. Treatment with promoters alone increases the number of foci to about 1000 (Xu et al., 1990); it increases DNA synthesis in the normal areas but does so to a higher degree in the spontaneous foci. Nodules formed during full carcinogenic treatment also exhibit a large increase in the number of DNA-synthesizing cells over the number in the surrounding liver (Enomoto and Farber, 1982). The proportion of such cells is reduced when the nodule remodels but remains higher than that of the surrounding cells. There is also residual patchy staining for nodule-related enzymes in the remodeled areas, indicating a continuing distinction from normal cells, although the overall morphology and architecture blend imperceptibly with the surrounding liver. Treatment with promoters alone also induces the formation of about 10 hyperplastic nodules, considered a true neoplastic development, but without invasive or metastatic capacity (Rossi et al., 1977). None of the promoting agents used in these experiments have any known mutagenic action and they are generally regarded as growth stimulants. The implication of these results is that spontaneous mutation normally leads to the formation of the enzyme-altered foci in untreated animals, and these are stimulated to form hyperplastic nodules by treatment with promoters alone. Treatment with a carcinogen and a promoter increases the number of foci and nodules many-fold. The promoter acts essentially as a selective agent in both the untreated and the carcinogen-treated rat. The role of the carcinogen is presumably to greatly increase the number of mutated cells that can then be selected for expansion by the promoter. That the carcinogen may also introduce selective growth is suggested in the resistant hepatocyte model based on the finding that most carcinogens inhibit the growth of normal hepatocytes but not of cells in the hyperplastic nodules (Tsuda et al., 1980). Microscopic altered hepatic foci are the earliest visible indicators of neoplasia in liver carcinogenesis, but the fact that repeated administration of promoters increases their number approximately 10-fold (Ogawa et al., 1981; Schulte-Hermann et al., 1983; Xu et al., 1990) suggests that there are
192
Harry Rubin
preneoplastic cells which can be detected only by the selective growth induced by promoters. The situation is analogous to the incipient neoplasia observed in cultured cells which is most clearly manifest by extended exposure to differentially selective growth conditions (Rubin, 1994c). There is evidence that spontaneously initiated cells also occur in the epidermis. Long-term repetitive promoter treatment of uninitiated skin of 40 Sencar mice induced six papillomas and three carcinomas (Pelling et al., 1988). Seven of the nine tumors contained a point mutation in the 61st codon of one allele of the c-Ha-ras oncogene. This suggests that the promoters selectively induced this expansion of clones bearing a spontaneous mutation of the Ha-ras gene, and this expansion increased the chance of further mutations that drove progression to papillomas and carcinomas. Brookes (1989) noted many other cases in which chemically induced tumors in experimental animals contained a mutation in one of the ras genes that was apparently of spontaneous rather than carcinogen-induced origin. These included transformation induced in mouse bladder epithelium in culture which increased markedly in frequency with the age of the donor. All the transformants, including an untreated control, had a Ki-ras mutation at the same site (Brookes et al., 1988). There were also liver tumors in rats induced by different carcinogens which contained the same Ha-ras mutation found in a spontaneous liver tumor (Stowers et al., 1988), and a similar situation was noted with guinea pig fibroblast lines in culture (Doniger et al., 1987). Evidence indicated that the carcinogen was selecting cells with spontaneous ras mutations for growth into tumors and possibly inducing additional mutations at other sites. Ha-ras-1 mutations were actually found in small patches of mammary epithelium in young untreated Fisher female rats. These developed into mammary carcinomas when the rats were treated with N-nitroso-N-methylurea (NMU), which apparently selected for growth of the spontaneously mutated cells and drove further progression into carcinoma (Cha et al., 1994, 1996). As a result, over 90% of the mammary cancers had Ha-ras-1 mutations. In contrast, mammary tumors induced by dimethylbenzanthrene rarely had such mutations, indicating that NMU was specifically selective for cells carrying a Ha-ras-1 mutation, thereby enabling progression to carcinoma. These observations raise questions about the nature of the microenvironment that influences selectivity for potentially cancerous cells. For example, does the structure of a tissue change with age in such a way to contribute to the marked increase with age in the incidence of cancer? The rate of entry into DNA synthesis of cells in the mouse intestinal crypt decreases markedly with age and becomes more heterogeneous (Fry et al., 1966; Lesher, 1966; Lesher et al., 1961). This suggests accumulation of damage in the epithelial cells over time which makes the cells more likely to undergo transformation
Role of Selection in Neoplastic Transformation
193
and/or makes the microenvironment lose its regulatory power over rogue cells. Evidence for the diminished regulatory power of some tissues with age derives from the observation that rat hepatocarcinoma cells are more likely to form tumors when inoculated into the livers of old syngeneic rats than those of young rats (McCullough et al., 1997). That this is a local rather than a systemic effect is indicated by the facility of the cells to form tumors when inoculated subcutaneously in both young and old rats. An observation in human cancer also indicates the significance of the local environment in tumor development. Metastases presented in the cervical lymph nodes of several patients in whom no primary tumors could be found in the mucosa of the upper aerodigestive tract (Califano et al., 1999). However, the same genetic lesions were found in defined spots of the normalappearing mucosa as in the lymph node metastases, and in some cases these later developed into carcinoma. Apparently, the lymph nodes offered a more permissive environment for growth of the tumor cells than did the mucosa. Damaging a tissue by radiation increases the likelihood that tumors will be established there by later-inoculated neoplastic cells. This was found to be true for the development of metastatic lesions in the lungs of mice (Brown, 1973; Milas and Peters, 1984). Mammary cancer cells that produce no tumors subcutaneously do so in the cleared mammary fat pad but do so more efficiently if the fat pad is preirradiated (Barcellos-Hoff and Ravani, 2000). This indicates that the undisturbed subcutaneous connective tissue is more regulatory than the mammary fat pad divested of its mammary epithelium, but further damage to the remaining connective tissue makes the fat pad an even more favorable environment. Perhaps the best known case of environmental effects on tumor development concerns teratocarcinoma. These tumors are induced by inoculating young mouse embryos into the scrotum of adult mice (Mintz and Illmensee, 1975). There they develop into teratocarcinomas which can be maintained by transfer into the adult or in culture. However, when inoculated into the blastocyst of the developing embryo, a normal chimeric mouse emerges. The previously carcinomatous cells occur in a wide variety of tissues, including the germ cells, through which they can give rise to normal progeny. In this system, the microenvironment determines both the origin of the tumor and its normalization. Most preneoplastic lesions induced during experimental carcinogenesis revert back to an apparently normal state or disappear entirely when administration of the carcinogen or the promoter is terminated, with only a small minority progressing to more neoplastic behavior. This is the case for papillomas of the skin (Shubik, 1950), hyperplastic nodules of the liver (Tatematsu et al., 1983), and aberrant crypt foci of the colon (Shpitz et al., 1996).
194
Harry Rubin
Tumors whose persistence is dependent on maintaining the condition used in their induction were designated as conditional (Rous and Kidd, 1941), and the concept was extended to many experimental neoplasms of endocrine origin (Furth, 1953). In humans with familial adenomatous polyposis, there is extensive resolution of rectal polyps following abdominal colectomy and ileorectal anastomosis (Feinberg et al., 1988), indicating that the contents of the colon contribute to the development of the rectal polyps. In all these cases persistence of the preneoplastic lesions is dependent on maintenance of the conditions that evoked them, which can therefore be considered a selective environment. Since the lesions have been shown to be of clonal origin, it is likely they arise from mutations. As noted previously, those mutations in some cases are spontaneous, although in other cases they may actually be induced by the carcinogenic treatment. As noted earlier, spontaneous transformation in culture is driven by selection of spontaneous variants, which raises the possibility that there are more instances of such origin in experimental and human cancer than have been identified to date. If so, the importance of selection assumes increased significance in tumor development. Although there is ample evidence for the role of topographical relations and of hormones in creating selective conditions for tumor development, there is also evidence for cellular autonomy in modifying the response to oncogenic mutations. As previously noted, biallelic inactivating mutations in the APC gene are widely accepted as the key to the initiation of colorectal tumor growth in humans (Kinzler and Vogelstein, 1996a; Levy et al., 1994; Tomlinson et al., 1996) and tumors of the small intestine and colon in mice (Luongo et al., 1994). The genetic background dramatically modifies the mouse tumor phenotype, and approximately 50% of the genetic variation is determined by a modifier gene, Mom-1 (Dietrich et al., 1993). The Mom-1 gene encodes a secretory phospholipase and it was thought to modify polyp number by altering the cellular environment within the intestinal crypt (MacPhee et al., 1993). However, analysis of mouse aggregation chimeras shows that the actions of both APC and Mom-1 are localized within the cell lineage that gives rise to intestinal tumors (Gould and Dove, 1997; Novelli et al., 1999). The cellular autonomy of biallelic APC mutations is consistent with the evidence of their sufficiency for the growth of early colorectal adenomas (Lamlum et al., 2000). However, the sufficiency of APC mutations for early colorectal adenoma growth has to be qualified in terms of modifier genes in the same cell that can suppress excessive growth (Dietrich et al., 1993). This is consistent with higher level controls at the tissue level as indicated by the following: It is likely that at least as many biallelic APC mutations occur in the small intestine crypt epithelium of humans as in colorectal epithelium since the former multiplies at an even a faster rate, but it only rarely forms tumors (Potten et al., 1992). In this regard, it is noteworthy that the structures of the small and large intestines differ considerably:
Role of Selection in Neoplastic Transformation
195
Epithelial crypt cells of the small intestine ascend into elongated villi that extend into the lumen, whereas those of the large intestine are extruded directly into the lumen. The distribution of DNA-synthesizing cells also differs in the two tissues (Potten et al., 1992). It is very important to establish the frequency of APC mutations in both tissues to evaluate the role of tissue structure in the neoplastic expression of this genetic alteration.
IX. SOURCES OF GENETIC VARIATION FOR POSSIBLE SELECTION IN TUMOR DEVELOPMENT In cell culture it appears well established that there is a high degree of variation to provide material for progressive transformation under selective conditions. It has been suggested that the frequency of mutated cells in epithelial cells in vivo is high enough to provide cells for clonal expansion (Simpson, 1997). For instance, an average of about 1 in 4000 cells of kidney from older humans harbors a mutation in the HPRT gene (Martin et al., 1996). The APC gene, which is prominently implicated as a gatekeeper in the early stages of human colorectal cancer development (Kinzler and Vogelstein, 1996a), is four times the length of the HPRT gene, and APC mutations would therefore be expected to be present in a higher frequency than HPRT mutations. Mutation in both APC alleles is required for adenoma formation, but mutation of a single allele causes a decrease in both migration rate and apoptosis of intestinal epithelium (Mahmoud et al., 1997). This suggests that there is already selection for cells with a single APC mutation in normal colon, which would then increase the probability for a mutation in the normal allele with ensuing adenoma formation. Microsatellite mutations occur in much higher frequencies than structural gene mutations, and when enough of them are examined all tumors appear to have some level of alteration (Gleeson et al., 1996). Simpson (1997) suggested that tumors represent a clonal selection of mutations that occur widely in normal tissue, thereby adding further weight to the role of selection in neoplastic development. The full extent of mutations in normal tissues has yet to be accurately determined. It has been estimated that there are approximately 28 large genetic rearrangements in the average liver cell of aging mice (Vijg et al., 1997; see also Doll´e et al., 2000). It has been reported that there are about 11,000 inter-simple sequence repeat mutations per cell in both polyps and carcinomas of the human colon (Stoler et al., 1999), which suggests many such mutations may occur clonally in individual normal crypts of the colon. Given the thousands of divisions that occur in stem cells of the intestines in a human lifetime (Potten and Loeffler, 1990), the presence of many mutations in such cells can be rationalized.
196
Harry Rubin
Use of a panel of nine highly informative microsatellite markers in histologically normal human breast ducts or terminal ductal-lobular revealed at least one genetic abnormality in one-half of the women examined (Larson et al., 1998). Half the markers were in chromosomal regions known to be lost or mutated in breast cancer, but none of the tissue samples revealed any histological abnormality. The presence of such genetic lesions does not necessarily lead to cancer, even many years later (Kasami et al., 1997). The results indicate the importance of selective conditions for the development of such clones into tumors. Selectivity could arise from additional mutations in the same cells or from a change in the selective microenvironment.
X. THE NATURE OF SELECTION IN VIVO The selective conditions in cell culture for the emergence of neoplastic growth are well defined. They reflect the fundamental property of transformed cells to overcome the physiological regulators of cell multiplication, such as contact inhibition in crowded cultures, the requirement for attachment to a solid substratum, and the need for a relatively high concentration of serum growth factors. The most commonly used of these criteria for transformation is the overcoming of contact inhibition in crowded cultures (Aaronson and Todaro, 1968a; Lee et al., 1989), sometimes combined with reduced serum concentrations (Rubin and Xu, 1989). However, epithelial cells in multicellular animals are always in intimate contact with one another, with many of them multiplying at a rapid rate. A representative example of such cells in a continuously renewing tissue is found in the crypts of the colon, in which stem cells multiply near the base of the crypts and their descendent transit cells continue to multiply through the bottom half of the crypt (Chang and Leblond, 1971). They then differentiate to carry out their function at the surface, lose the capacity to multiply, and are desquamated into the lumen. In the early stages of chemically induced carcinogenesis, the zone of multiplication of many crypts moves further toward the surface, and there is an increase in their length (Chang, 1978, 1982). At least part of this increase in thickness of the mucosa is a persistent compensatory hyperplasia in response to damage inflicted on the multiplying cells by repeated application of the carcinogen. This is then followed by focal multiplication throughout a single crypt, resulting in the formation of a polyp protruding into the lumen. There has been a loss of control by those signals that propel the differentiation and replicative sterility of the cells. This could occur through a genetic lesion in the stem cells that interferes with the recognition of topographical signal for differentiation or it may be a loss of the
Role of Selection in Neoplastic Transformation
197
signal. Perhaps it is a combination of both since the altered stem cells seed the remainder of the crypt, thereby changing both the signals and the signaled. The polyps represent a clonal expansion but are unlikely to progress to cancer unless there is biallelic mutation at a “gatekeeper” gene such as the APC locus. This produces a dysplastic adenomatous polyp with further growth and progression to carcinoma (Kinzler and Vogelstein, 1998). The dysplasia of an adenoma is an alteration of the normal architecture of the crypt, and the cells exhibit internal changes such as increased nuclear size and greater intensity of staining by dyes. Therefore, in addition to the loss of response to topographical differentiating signals of the crypt, there is a loss of the regular relation of cells to one another which normally modulates their growth, even in a polyp. In this case, the APC or gatekeeper mutation changes both the selective environment and the selective advantage of the cells harboring the mutation. Additional evidence for the role of normal topographical relations in regulating growth and differentiation derives from experiments in which the architecture of a tissue is physically disrupted (Rubin, 1985). Mammary epithelium of virgin female mice is removed, and the cells are enzymatically dissociated from one another and inoculated into the cleared fat pad of syngeneic females (DeOme et al., 1978). The mice then develop transformed hyperplastic nodules in 2 months rather than the average 9 months in undisturbed mammary glands. If the architecture is undisturbed as in undissociated fragments, the development of nodules takes about 4 months longer than if the cells are dispersed (Medina et al., 1978). Apparently, the controls that maintain cells in a regulated process of growth and differentiation are breached when the architecture of the tissue is disrupted, and tumor development is accelerated. In organs that do not undergo a regular localized renewal from stem cells, regulation may be hormonal or some other form of systemic control. Disruption may occur through either excess or deficiency of hormones. Radiation damage to the thyroid of mice with large doses of iodine131 abolishes thyroid hormone production, causes sustained stimulation of the pituitary, and terminates in multifocal development of pituitary adenomas (Furth, 1953). The lack of thyroid hormone is a selective signal for proliferation of pituitary cells and those with the greatest proliferative potential develop into tumors. Treatment of mice with thiouracil blocks the synthesis of thyroid hormone without damaging thyroid cells (Furth, 1953). This causes an overproduction of thyroid-stimulating hormone from the pituitary, which causes hyperplasia of the thyroid. If sustained long enough, there is focal outgrowth of thyroid adenomas and, later, carcinomas. The tumors can only be successfully grafted in mice whose thyroids are similarly blocked by thiouracil, but not in untreated mice. In the course of subpassages in thiouracil-treated
198
Harry Rubin
mice, the tumors become autonomous and can grow in normal mice. In one case of thyroid hormone deficiency, selective outgrowth of pituitary tumors occurs, and in the other case selective outgrowth of thyroid tumors occurs. A variety of tumors are induced in rats by prolonged treatment with estrogen (Noble, 1977). These tumors can only be transplanted in estrogenized rats, and they maintain their hormone-dependent status. They regress when estrogen treatment stops, but spontaneous regrowth often occurs. When regrowth occurs after complete regression, most of the tumors progress to hormone independence and rapidly kill the host. This is a clear case of selection driving progression to a highly malignant state. It bears resemblance to the selective progression in transformation seen in culture when the concentration of CS is reduced or FBS is substituted for CS. Prolonged hyperplasia in a tissue is the common precursor of neoplastic growth (Chang, 1978; Cramer and Stowell, 1942; Foulds, 1969, 1975; Furth, 1953; Laird and Barton, 1959) and is the site for the development of clonal neoplastic growth (Rabes et al., 1982; Reddy et al., 1982; Siu et al., 1999). The sequence is similar to that seen in culture when NIH 3T3 cultures are grown to confluence in high concentrations of CS which produces hyperplasia in comparison to the thin monolayered sheets seen with low concentrations of CS (Rubin, 1994c). Many small neoplastic foci appear in the former within 2 weeks, whereas no such growths are seen in the latter even after 7 weeks. Subcultures made in low CS from the high CS cultures progress rapidly at confluence to large, dense tumorigenic foci, whereas similar subcultures from the low CS progress very slowly. The results in cell culture show the importance of hyperplastic growth under selective conditions in neoplastic development and provide a quantitative model for progression in vivo. In both cases, clonal expansion provides the opportunity for accumulation of tumor suppressor mutations which drive tumor progression (Tomlinson and Bodmer, 1999; Tomlinson et al., 1996). Deficiency of growth factors may be a common selective agency in the development of cancer and its relation to aging. For example, there is a sharp increase in the incidence of prostate cancer with age at the same time androgen levels decrease (Lamberts and Van der Lely, 1997; Prehn, 1999). The declining levels of androgen are associated with varying degrees of atrophy in the prostate (Bruchovsky et al., 1975). Compensatory hyperplasia occurs in various parts of the prostate of aging men, but the most significant of the hyperplastic areas are foci of surviving cells that are less dependent on androgen stimulation. Continuing decline in androgen levels probably exerts selective pressure for cells in hyperplastic foci that are increasingly independent of androgens and more malignant in their behavior. Confirmation of this pathway of prostatic cancer would suggest that androgen administration is an appropriate therapy in such cases (Prehn, 1999).
Role of Selection in Neoplastic Transformation
199
Another accompaniment of aging is a continuous decrease in the functional capacity of most tissues (Shock, 1983). The growth rate of intestinal epithelium decreases with age in mice, as does the cellular heterogeneity for time of onset of DNA synthesis (Lesher et al., 1961). This suggests the accumulation of genetic damage in the fast growing stem cells of the intestine. Not only does this increase the chances for oncogenic mutations but also it might lessen the growth regulatory capacity of the intestinal microenvironment and allow the selective outgrowth of the genetically altered cells.
XI. SELECTION IN CARCINOGENESIS BY POLYCYCLIC AROMATIC HYDROCARBONS Modern study of chemical carcinogenesis began with the demonstration that chemically synthesized dibenzanthracene and benzpyrene isolated from coal tar caused tumors when painted repeatedly on the skin over a long period of time (Cook et al., 1933; Kennaway, 1930). Many other polycyclic aromatic hydrocarbons (PAHs) were later found to cause skin tumors by the same procedure (Phillips, 1983). The first indication that selection might play a role in the development of tumors induced by PAHs occurred when the growth inhibitory effect of these compounds on cells was discovered (Haddow, 1938). PAHs caused a prolonged growth inhibition and even regression in spontaneous tumors and long-transplanted tumors of rats. In contrast, PAHs had little effect on the growth of primary tumors induced by these and other compounds. There was little or no specificity involved between the causative agent and the inhibitory PAH, indicating that some general property of tumors newly induced by chemicals was selected during development of the tumors. The role of selection took a different form when it was discovered that a single application of a subcarcinogenic dose of PAH followed by repeated application of a noncarcinogenic promoter, such as croton oil or its active ingredient TPA, also induced tumors (Berenblum, 1974; Berenblum and Shubik, 1947). It was generally accepted that the PAH acted in this procedure as an initiator by inducing mutations and the promoter acted essentially as a selective agent (Hennings et al., 1983). It has since been shown that both croton oil and TPA act in two stages in promotion of skin carcinogenesis ¨ (Boutwell, 1964; Furstenberger et al., 1981; Slaga et al., 1980). If TPA is followed by many repeated applications of phorbol-12-retinoate-13-acetate (RPA), which induces hyperplasia but does not by itself act as a promoter, ¨ many tumors are formed (Furstenberger et al., 1981). The limited TPA treatment of initiated skin produces a stable state that can be promoted
200
Harry Rubin
to tumor formation by repeated application of RPA beginning up to 8 weeks ¨ later and perhaps even beyond that interval (Furstenberger et al., 1983). TPA by itself causes chromosome breaks in the basal cells of the epidermis within 48 hr of a single application, whereas RPA induces no chromosome aberra¨ tions (Furstenberger et al., 1989). The well-known alkylating and clastogenic agent methyl methanesulfonate plays the same role as TPA in the first stage of promotion. The first stage is considered conversion, and the second stage is true promotion which probably involves clonal expansion driven by the ¨ chronic growth stimulation of RPA (Furstenberger et al., 1981) or mezerein (Slaga et al., 1980). The second stage would then be the stage of selection of cells which had undergone tumorigenic mutation and chromosomal aberrations during initiation and conversion. Selection has become an issue in lung cancer, which is the leading cause of cancer death in the United States. PAHs in tobacco smoke are strongly implicated as causative agents of these cancers, with benzpyrene considered the most important of the PAHs in tobacco smoke. Benzpyrene diol epoxides (BPDEs) are assumed to be the ultimate carcinogenic form of benzpyrene (Phillips, 1983). BPDEs form adducts with guanine in specified codons of the p53 gene after treatment of cells in culture, resulting in G → T transversions (Denissenko et al., 1996). These positions are also mutational hot spots in human lung cancers, and it was concluded that targeted adduct formation rather than phenotypic selection appears to shape the p53 mutational spectrum in lung cancer. A contrary view arose from analysis of complementary base substitutions of the p53 gene in lung cancer. The analysis indicated that strand-specific repair of primary lesions and sitespecific selection of the resulting mutations determines lung cancer-specific hot spots in DNA (Rodin and Rodin, 2000). This provides a selectionbased explanation of why lung cancers with the specific p53 mutations are more common in smokers than in nonsmokers. This conclusion concurs with the observation that base pair substitutions in the p53 gene in different human cancers are similar to germline mutations and suggests that selection is the critical factor in determining which p53 mutations are associated with human cancer (Krawczak and Cooper, 1998; Krawczak et al., 1995). A mechanism of selection involving the p53 gene was described in cultured cells and supported by findings in tumors (Graeber et al., 1996; Kinzler and Vogelstein, 1996b). It was found that hypoxia induced apoptosis in a cell line and that mutational loss of p53 tumor-suppressor gene function substantially reduced the apoptotic loss of cells. Highly apoptotic regions correlated with hypoxic regions in transplanted tumors expressing wild-type p53. In contrast, little apoptosis occurred in hypoxic regions of p53-deficient tumors. Hence, it was proposed that hypoxia in tumors selects for cells with a defect in apoptosis often produced by loss of p53 function. This loss drives
Role of Selection in Neoplastic Transformation
201
progression to a more malignant prototype. Selection under stressful conditions may therefore explain the common occurrence of p53 mutations in human cancer.
XII. CONCLUSIONS Recent data from human tumors has focused interest on the role of selection in the carcinogenic process. A strong argument has been made that selection plays a far more important role than genetic instability in the development of sporadic colon cancer of humans (Tomlinson and Bodmer, 1999; Tomlinson et al., 1996). It has also been argued that the selection of endogenous mutations by benzpyrene in tobacco smoke can better account for the base substitutions in specific codons of the p53 gene than can the direct induction of those changes by the specific binding of benzpyrene to those sites (Rodin and Rodin, 2000). Although both points of view may still be considered controversial in the case of human cancers, there is collateral evidence that indicates a major role for selection of endogenous mutations in experimental cancer of the skin and of the liver (Pelling et al., 1988; Schulte-Hermann et al., 1981). However, the most unequivocal experimental support for a dominant role of selection derives from spontaneous neoplastic transformation of cells in culture. A variety of findings derived from studies of spontaneous transformation in the NIH 3T3 line of mouse fibroblasts leave no doubt that selection is the driving force in the process (Chow and Rubin, 2000a). This evidence is supported by the finding that selection drives transformation of a diploid line of rat liver epithelial cells to almost the same extent in carcinogen-treated and in untreated cells (Lee et al., 1989). The experiments in cell culture also reveal aspects of the selective process that have not been clearly established in the organism. They show that there is a high rate of spontaneous variation in proliferative potential of cells at high population density (Rubin et al., 1990b) and that selection occurs under mild constraint of proliferation before there are any discrete neoplastic lesions (Chow and Rubin, 2000a; Rubin, 1994c). A similar absence of visible precursors occurs in the early progression of mutator-type colorectal tumors of humans (Tsao et al., 2000). The quantitative results in cell culture arise from analysis of the dynamics of cell proliferation and transformation under controlled conditions. However, there has been no attempt to establish the molecular basis for the underlying variation, which should prove a fruitful area for future investigation. The insights gained from studies of cell culture indicate that more attention should be given to the biology of tumor development to provide a context
202
Harry Rubin
for the rapidly accumulating data from molecular genetics for properly evaluating the role of selection in carcinogenesis. A full consideration of carcinogenesis in the organism has to take account of regulatory forces that differ from those in cell culture. One feature of growth regulation that plays a major role in maintaining homeostasis of the organism is the architecture of tissues. This was indicated by the marked enhancement of tumorigenesis when explanted mammary epithelial cells of young mice were enzymatically dissociated and reinjected into the gland-free mammary fat pad (DeOme et al., 1978). In contrast, reinsertion of intact 1-mm pieces elicited little or no enhancement of tumorigenesis (Medina et al., 1978). A general view of the role of tissue organization in homeostasis has been expressed as macrodeterminism in embryonic development (Weiss, 1973), top-down control of consciousness over excitation of cerebral neurons (Sperry, 1969), and ordered heterogeneity as a basic principle in a comprehensive theory of organisms (Elsasser, 1998). Terminal differentiation serves as an operational arm of such principles to deter unregulated proliferation of rogue clones that arise from endogenous mutations. A breakdown in such hierarchical regulatory forces with age, carcinogen exposure, or cell dispersal will have to be considered along with the cellular dynamics of tumor development and molecular changes in the genome to achieve a balanced understanding of neoplastic development.
ACKNOWLEDGMENTS Much of the clonal work that led to the major conclusions about the role of selection in neoplastic transformation was done in collaboration with Dr. Ming Chow. I thank Dorothy M. Rubin for assistance with the various revisions of the manuscript. The expenses in support of preparing the manuscript were partly underwritten by the Elsasser Family Fund.
REFERENCES Aaronson, S. A., and Todaro, G. J. (1968a). Science 162, 1024–1026. Aaronson, S. A., and Todaro, G. J. (1968b). J. Cell. Physiol. 72, 141–148. Armitage, P., and Doll, R. (1954). Br. J. Cancer 8, 1–12. Barcellos-Hoff, M. H., and Ravani, S. A. (2000). Cancer Res. 60, 1254–1260. Berenblum, I. (1974). “Carcinogenesis as a Biological Problem.” North-Holland, Amsterdam. Berenblum, I., and Shubik, P. (1947). Br. J. Cancer 1, 383–391. Bosmann, H. B., and Lione, A. (1974). Biochem. Biophys. Res. Commun. 61(2), 564–567. Boutwell, R. K. (1964). Prog. Exp. Tumor Res. 4, 207–250. Brookes, P. (1989). Mol. Carcinogen. 2, 305–307.
Role of Selection in Neoplastic Transformation
203
Brookes, P., Cooper, C. S., Ellis, M. V., Warren, W., Gardner, E., and Summerhayes, I. C. (1988). Mol. Carcinogen. 1, 82–88. Brooks, R. F., Richmond, F. N., Riddle, P. N., and Richmond, K. M. V. (1984). J. Cell. Physiol. 121, 341–350. Brown, J. M. (1973). Br. J. Radiol. 46, 613–618. Bruchovsky, N., Lesser, B., Van Doorn, E., and Craven, S. (1975). Vitamins Horm. 33, 61– 102. Califano, J., Westra, W. H., Koch, W., Meininger, G., Reed, A., Yip, L., Boyle, J. O., Lonardo, F., Sidransky, D. (1999). J. Natl. Cancer Inst. 91, 599–604. Campbell, H. A., Xu, Y.-D., Hanigan, M. H., and Pitot, H. C. (1986). J. Natl. Cancer Inst. 76, 751–767. Cha, R. S., Thilly, W. G., and Zarbl, H. (1994). Proc. Natl. Acad. Sci. USA 91, 3749–3753. Cha, R. S., Guerra, L., Thilly, W. G., and Zarbl, H. (1996). Carcinogenesis 17, 2519–2524. Chang, W. (1978). J. Natl. Cancer Inst. 60, 1405–1418. Chang, W. W. L. (1982). Virchows Arch. [Cell Pathol.] 41, 17–37. Chang, W. W. I., and Leblond, C. P. (1971). Am. J. Anat. 131, 73–100. Chow, M., and Rubin, H. (1999a). Proc. Natl. Acad. Sci. USA 96, 2093–2098. Chow, M., and Rubin, H. (1999b). Proc. Natl. Acad. Sci. USA 96, 6976– 6981. Chow, M., and Rubin, H. (1999c). In Vitro 35, 449– 458. Chow, M., and Rubin, H. (2000a). Cancer Res. 60, 6510– 6518. Chow, M., and Rubin, H. (2000b). Proc. Natl. Acad. Sci. USA 97, 174–178. Cook, J. W., Hewett, C. L., and Hieger, I. (1933). J. Chem. Soc. (London) 395– 405. Cramer, W., and Stowell, R. E. (1942). J. Natl. Cancer Inst. 2, 369–378. Denissenko, M. F., Pao, A., Tang, M.-S., and Pfeifer, G. P. (1996). Science 274, 430 – 432. DeOme, K., Miyamoto, M., Osborn, R., Guzman, R., and Lum, K. (1978). Cancer Res. 38, 2103–2111. Dietrich, W. F., Lander, E. S., Smith, J. S., Moser, A. R., Gould, K. A., Lungo, C., Borenstein, N., and Dove, W. F. (1993). Cell 75, 631–639. Dolle, M., Snyder, W., Crossen, J., Lohman, P., and Vijg, J. (2000). Proc. Natl. Acad. Sci. USA. 97, 8403–8408. Doniger, J., Notario, V., and DiPaolo, J. A. (1987). J. Biol. Chem. 262, 3813–3819. Earle, W. R. (1943a). J. Natl. Cancer Inst. 4, 165–212. Earle, W. R. (1943b). J. Natl. Cancer Inst. 4, 213–227. Eldridge, S. R., and Gould, M. N. (1992). Int. J. Cancer. 50, 321–324. Ellison, B. J., and Rubin, H. (1992). Cancer Res. 52, 667– 673. Elmore, E., Kakanaga, T., and Barrett, J. C. (1983). Cancer Res. 43, 1650–1655. Elsasser, W. M. (1998). “Reflections on a Theory of Organisms.” Johns Hopkins Univ. Press, Baltimore, MD. Enomoto, K., and Farber, E. (1982). Cancer Res. 42, 2330–2335. Eshleman, J. R., Lang, E. Z., Bowerfield, G. K., Parsons, R., Vogelstein, B., Willson, J. K. V., Veigl, M. L., Sedgwick, W. D., Markowitz, S. D. (1995). Oncogene 10, 33–37. Farber, E., and Sarma, D. S. R. (1987). Lab. Invest. 56, 4–22. Feinberg, S. M., Sarre, R. G., Fazio, V. W., and Weakly, F. L. (1988). Dis. Col. Rect. 31, 169–175. Fisher, J. C. (1958). Nature 181, 651–652. Foulds, L. (1969). “Neoplastic Development, Vol. 1.” Academic Press, New York. Foulds, L. (1975). “Neoplastic Development, Vol. 2.” Academic Press, New York. Fry, R. J. M., Tyler, S. A., and Lesher, S. (1966). In “Radiation and Ageing. (P. J. Lindop and G. A. Sacher, Eds.). Taylor & Francis, Semmering, Austria. ¨ Furstenberger, G., Berry, D. L., Sorg, B., and Marks, F. (1981). Proc. Natl. Acad. Sci. USA 78, 7722–7726. ¨ Furstenberger, G., Sorg, B., and Marks, F. (1983). Science. 220, 89–91.
204
Harry Rubin
¨ Furstenberger, G., Schurich, B., Kaina, B., Petrusevska, R. T., Fusenig, N. E., and Marks, F. (1989). Carcinogenesis. 10, 749–752. Furth, J. (1953). Cancer Res. 13, 477– 492. Glaab, W. E., and Tindall, K. R. (1997). Carcinogenesis 18, 1–8. Gleeson, C. M., Sloan, J. M., Mcguigan, J. A., Ritchie, A. J., Weber, J. L., and Russell, S. E. H. (1996). Cancer Res. 56, 259–263. Gould, K. A., and Dove, W. F. (1997). Proc. Natl. Acad. Sci. USA 94, 5848–5853. Graeber, T. G., Osmanian, C., Jacks, T., Housman, D. E., Koch, C. J., Lowe, S. W., and Giaccia, A. (1996). Nature 379, 88–91. Grisham, J. W., Smith, G. J., Lee, L. W., Bentley, K. S., and Fatteh, M. V. (1988a). Cancer Res. 48, 5977–5983. Grisham, J. W., Smith, G. J., Lee, L. W., Bentley, K. S., and Fatteh, M. V. (1988b). Cancer Res. 48, 5969–5976. Grundel, R., and Rubin, H. (1991). Cancer Res. 51, 1003–1013. Grundel, R., and Rubin, H. (1992). Carcinogenesis 13, 1873–1877. Haber, D. A., Fox, D. A., Dynan, W. S., and Thilly, W. G. (1977). Cancer Res. 37, 1644 –1648. Haddow, A. (1938). J. Pathol. Bacteriol. 47, 581–591. Haggett, A. C., Ellis, P. A., Ford, C. P., Hampton, L. L., Rimoldi, D., and Thorgeirsson, S. S. (1991). Cancer Res. 51, 5929–5936. Hennings, H., Shores, R., Wenk, M. L., Spangler, E. F., Tarone, R., and Yuspa, S. H. (1983). Nature 304, 67– 69. Homfray, T. F. R., Cottrell, S. E., Ilyas, M., Rowan, A., Talbot, I. C., Bodmer, W. F., and Tomlinson, I. P. M. (1998). Hum. Mutat. 11, 114–120. Hozier, J., Sawyer, J., Clive, D., and Moore, M. (1985). Mutat. Res. 147, 237–242. Jackson, J. L., Sanford, K. K., and Dunn, T. B. (1970). J. Natl. Cancer Inst. 45, 11–23. Jainchill, J. L., Aaronson, S. A., and Todaro, G. J. (1969). J. Virol. 4, 549–553. Kaden, D., Gadi, I. K., Bardwell, L., Gelman, R., and Sager, R. (1989). Cancer Res. 49, 3374– 3379. Kasami, M., Vnencak-Jones, C. L., Manning, S., Dupont, W. D., and Page, D. L. (1997). Am. J. Pathol. 150, 1925–1932. Kennaway, E. L. (1930). Biochem. J. 24, 497–504. Kilkenny, A. E., Morgan, D., Spangler, E. F., and Yuspa, S. H. (1985). Cancer Res. 45, 2219– 2225. Kinzler, K. W., and Vogelstein, B. (1996a). Cell 87, 159–170. Kinzler, K. W., and Vogelstein, B. (1996b). Nature 379, 19–20. Kinzler, K. W., and Vogelstein, B. (1998). In “Genetic Basis of Human Cancer.” (B. Vogelstein and K. W. Kinzler, Eds.). McGraw-Hill, New York. Kirkland, D. J., Armstrong, C., and Harris, R. J. C. (1975). B. J. Cancer 31, 329–337. Krawczak, M., and Cooper, D. N. (1998). Mutagenesis 13, 319–320. Krawczak, M., Smith-Sorenson, B., Schmidtke, J., Kakkar, V. V., Cooper, D. N., and Hovig, E. (1995). Hum. Mutat. 5, 48–57. Krooth, R. S., Shaw, M. W., and Campbell, B. K. (1964). J. Natl. Cancer Inst. 32, 1031–1044. Kulesz-Martin, M. F., Koehler, B., Hennings, H., and Yuspa, S. H. (1980). Carcinogenesis 1, 995–1006. Laird, A. K., and Barton, A. D. (1959). Nature 183, 1655–1657. Lamberts, S., and Van der Lely, A.-J. (1997). Science 278, 419– 414. Lamlum, H., Papadopolou, A., Ilyas, M., Rowan, A., Gillet, C., Hanby, A., Talbot, I., Bodmer, W., Tomlinson, I. (2000). Proc. Natl. Acad. Sci. USA 97, 2225–2228. Larson, P. S., de las Morenas, A., Cupples, L. A., Huang, K., and Rosenberg, C. L. (1998). Am. J. Pathol. 152, 1591–1598.
Role of Selection in Neoplastic Transformation
205
Lee, L. W., Tsao, M.-S., Grisham, J. W., and Smith, G. J. (1989). Am. J. Pathol. 135, 63–71. Lesher, S. (1966). Radiation Ageing, 183–206. Lesher, S., Fry, R. J. M., and Kohn, H. I. (1961). Exp. Cell Res. 24, 334–343. Levy, D. B., Smith, K. J., Beazer-Barclay, Y., Hamilton, S. R., Vogelstein, B., and Kinzler, K. W. (1994). Cancer Res. 54, 5953–5958. Loeb, L. A. (1991). Cancer Res. 51, 3075–3079. Luongo, C., Moser, A. R., Gledhill, S., and Dove, W. F. (1994). Cancer Res. 54, 5947–5952. MacPhee, M., Chepenik, K. P., Liddell, R. A., Nelson, K. K., Siracusa, L. D., and Buchberg, A. M. (1993). Cell 81, 957–966. Mahmoud, N. N., Boolbol, S. K., Bilinski, R. T., Martucci, C., Chadburn, A., and Bertagnolli, M. M. (1997). Cancer Res. 57, 5045–5050. Martin, G. M., Ogburn, C. E., Colgin, L. M., Gown, A. M., Edland, S. D., and Monnat, R. J. (1996). Hum. Mol. Genet. 5, 215–221. McCullough, K. D., Coleman, W. R., Smith, G. J., and Grisham, J. W. (1997). Cancer Res. 57, 1807–1813. Medina, D., Shepherd, F., and Gropp, T. (1978). J. Natl. Cancer Inst. 60, 1121–1126. Mehta, P. P., Bertram, J. S., and Loewenstein, W. R. (1986). Cell 44, 187–196. Milas, L., and Peters, L. J. (1984). In “Cancer Invasion and Metastasis: Biologic and Therapeutic Aspects.” (G. L. Vichols and L. Milas, Eds.). Raven Press, New York. Mintz, B., and Illmensee, K. (1975). Proc. Natl. Acad. Sci. USA 72, 3585–3589. Noble, R. L. (1977). Cancer Res. 37, 82–94. Novelli, M. R., Wasan, H., Rosewell, I., Bee, J., Tomlinson, I. P. M., Wright, N. A., and Bodmer, W. F. (1999). Proc. Natl. Acad. Sci. USA 96, 1255–1258. Nowell, P. (1976). Science 194, 23–28. Ogawa, K., Ono´e, T., and Takeuchi, M. (1981). J. Natl. Cancer Inst. 67, 407– 412. Pelling, J. C., Neades, R., and Strawhecker, J. (1988). Carcinogenesis 9, 665– 667. Phillips, D. H. (1983). Nature 303, 468– 472. Potten, C. S., and Loeffler, M. (1990). Development 110, 1001–1020. Potten, C. S., Rew, D. A., and Roberts, S. A. (1992). Gut 33, 524–529. Prehn, R. T. (1999). Cancer Res. 59, 4161– 4164. Rabes, H. M., Bucher, T., Hartman, A., Linke, I., and Dunnwald, M. (1982). Cancer Res., 3220–3227. Reddy, E. P., Reynolds, R. K., Santos, E., and Barbacid, M. (1982). Nature 300, 149–152. Reznikoff, C. A., Bertram, J. S., Brankow, D. W., and Heidelberger, C. (1973). Cancer Res. 33, 3239–3249. Rodin, S. N., and Rodin, A. S. (2000). Proc. Natl. Acad. Sci. USA 97, 12244–12249. Rossi, L., Ravera, M., Repetti, G., and Santi, L. (1977). Int. J. Cancer 19, 179–197. Rous, P., and Kidd, J. G. (1941). J. Exp. Med. 73, 365–390. Rubin, A. L. (1995). In Vitro Cell. Dev. Biol. 31, 183–189. Rubin, A. L., and Rubin, H. (1994). Proc. Natl. Acad. Sci. USA 91, 2320–2323. Rubin, A. L., Arnstein, P., and Rubin, H. (1990a). Proc. Natl. Acad. Sci. USA 87, 10005–10009. Rubin, A. L., Yao, A., and Rubin, H. (1990b). Proc. Natl. Acad. Sci. USA 87, 482– 486. Rubin, A. L., Sneade-Koenig, A., and Rubin, H. (1992). Proc. Natl. Acad. Sci. USA 89, 4183– 4186. Rubin, H. (1960). Virology 12, 14 –31. Rubin, H. (1981). Proc. Natl. Acad. Sci. USA 78, 328–332. Rubin, H. (1985). Cancer Res. 45, 2935–2942. Rubin, H. (1988). Cancer Res. 48, 2512–2518. Rubin, H. (1992). Proc. Natl. Acad. Sci. USA 89(3), 977–981. Rubin, H. (1993a). Proc. Natl. Acad. Sci. USA 90, 10715–10719.
206
Harry Rubin
Rubin, H. (1993b). Differentiation 53, 123–137. Rubin, H. (1994a). Proc. Natl. Acad. Sci. USA 91, 1039–1043. Rubin, H. (1994b). Proc. Natl. Acad. Sci. USA 91, 6619–6623. Rubin, H. (1994c). Proc. Natl. Acad. Sci. USA 91, 12076 –12080. Rubin, H. (2001). Cancer Res. 61, 799 –807. Rubin, H., and Xu, K. (1989). Proc. Natl. Acad. Sci. USA 86, 1860–1864. Rubin, H., Arnstein, P., and Chu, B. M. (1984). Cancer Res. 44, 5242–5248. Rubin, H., Chu, B. M., and Arnstein, P. (1986). Cancer Res. 46, 2027–2034. Rubin, H., Yao, A., and Chow, M. (1995). Proc. Natl. Acad. Sci. USA 92, 4843– 4847. Rubin, H., Chow, M., and Yao, A. (1996). Proc. Natl. Acad. Sci. USA 93, 1825 –1830. Sanford, K. K., Earle, W. R., Shelton, E., Schilling, E. L., Duchesne, E. M., Likely, G. D., and Becker, M. M. (1950). J. Natl. Cancer Inst. 11, 351–375. Schulte-Hermann, R., Ohde, G., Schuppler, J., and Timmermann-Trosiener, I. (1981). Cancer Res. 41, 2556–2562. Schulte-Hermann, R., Timmerman-Trosiener, I., and Schuppler, J. (1983). Cancer Res. 43, 839–844. Seshadri, R., Kutlaca, R. J., Trainor, K., Matthews, C., and Morley, A. A. (1987). Cancer Res. 47, 407– 409. Shock, N. (1983). In “Fundamentals of Geriatric Medicine.” (R. D. T. Cape, R. M. Coe, and I. Rossman, Eds.). Raven Press, New York. Shpitz, B., Hay, K., Medline, A., Bruce, W. R., Bull, S. B., Gallinger, S., and Stern, H. (1996). Dis. Colon Rectum 39, 763–767. Shubik, P. (1950). Cancer Res. 10, 713–717. Simpson, A. J. G. (1997). Adv. Cancer Res. 71, 209–240. Siu, I.-M., Robinson, D. R., Schwartz, S., Kung, H.-J., Pretlow, T. G., Peterson, R. B., and Pretlow, T. P. (1999). Cancer Res. 59, 63– 66. Slaga, T. J., Fischer, S. M., Nelson, K., and Gleason, G. L. (1980). Proc. Natl. Acad. Sci. USA 77, 3659–3663. Smith, G. J., Bell, W. N., and Grisham, J. W. (1993). Cancer Res. 53, 500–508. Sperry, R. W. (1969). Psychol. Rev. 76, 532–536. Squire, R. A., and Levitt, M. H. (1975). Cancer Res. 35, 3214 –3223. Stoker, M. (1964). Virology 24, 165–174. Stoler, D. L., Chen, N., Basik, M., Kahlenberg, M. S., Rodriguez-Bigas, M. A., Petrelli, N. J., and Anderson, G. R. (1999). Proc. Natl. Acad. Sci. USA 96, 15121–15126. Stowers, S. J., Wiseman, R. W., Ward, J. M., Miller, E. C., Miller, J. A., Anderson, M. W., and Eva, A. (1988). Carcinogenesis 9, 271–276. Tatematsu, M., Nagamine, Y., and Farber, E. (1983). Cancer Res. 43, 5049–5058. Taylor, W. G., Camalier, R. F., Tucker, R. W., and Sanford, K. K. (1978). Cancer Res. 38, 3840–3846. Tlsty, T., Margolin, B. H., and Lum, K. (1989). Proc. Natl. Acad. Sci. USA 86, 9441–9445. Todaro, G. J., and Green, H. (1963). J. Cell Biol. 17, 299–313. Tomlinson, I. P. M., and Bodmer, W. F. (1999). Nature Med. 5, 11–12. Tomlinson, I. P. M., Novelli, M. R., and Bodmer, W. F. (1996). Proc. Natl. Acad. Sci. USA 93, 14800–14803. Tsao, M.-S., Shepherd, J., and Batist, G. (1990). Cancer Res. 50, 1941–1947. ¨ Tsao, J.-L., Yatabe, Y., Salovaara, R., Jarvinen, H. J., Mecklin, J.-P., Aaltonen, L. A., Tavare, S., and Shibata, D. (2000). Proc. Natl. Acad. Sci. USA 97, 1236–1241. Tsuda, H., Lee, G., and Farber, E. (1980). Cancer Res. 40, 1157–1164. Vijg, J., Dolle, M. E. T., Martus, H.-J., and Boerrigter, M. (1997). Mech. Ageing Dev. 99, 257–271.
Role of Selection in Neoplastic Transformation
207
Vogelstein, B., Fearon, E. C., Kern, S. E., Hamilton, S. R., Preisinger, A. C., Nakamura, Y., and White, R. (1989). Science 244, 207–211. Weiss, P. (1973). “The Science of Life.” Futura, Mt. Kisco, NY. Weiss, R. A. (1970). Exp. Cell Res. 63, 1–18. Williams, G. M., Weisburger, E. K., and Weisburger, J. H. (1971). Exp. Cell Res. 69, 106–112. Wittenkeller, J. L., Storer, B., Bittner, G., and Schiller, J. H. (1997). Oncology 54, 335–341. Wyllie, A. H. (1980). Nature 284, 555–556. Xu, Y.-H., Campbell, H. A., Sattler, G. L., Hendrich, S., Maronpot, R., Sato, K., and Pitot, H. C. (1990). Cancer Res. 50, 472– 479. Yao, A., and Rubin, H. (1992). Proc. Natl. Acad. Sci. USA 89, 7486–7490. Yao, A., and Rubin, H. (1994). Proc. Natl. Acad. Sci. USA 91, 7712–7716. Yao, A., Rubin, A. L., and Rubin, H. (1990). Cancer Res. 50, 5171–5176. Young, J., Leggett, B., Gustafson, C., Ward, M., Searle, J., Thomas, L., Buttenshaw, R., and Chenevix-Trench, G. (1993). Hum. Mutat. 2, 351–354.
This Page Intentionally Left Blank
ATM: Genome Stability, Neuronal Development, and Cancer Cross Paths Yosef Shiloh1 and Michael B. Kastan2 1
Department of Human Genetics and Molecular Medicine Sackler School of Medicine Tel Aviv University Tel Aviv 69978, Israel 2 Department of Hematology–Oncology St. Jude Children’s Research Hospital Memphis, TN 38105
I. Introduction II. Ataxia-Telangiectasia: A Disease Caused by ATM Deficiency A. Clinical–Pathological Phenotype B. The Cellular Phenotype III. The ATM Gene and Its Mutations A. The A-T Gene, ATM B. Somatic ATM Mutations in Cancer IV. The ATM Protein: From Sequence to Function A. ATM Sequence Motifs: A Membership Card in the PIK-Related Kinase Family B. The ATM Protein: Initial Characterization C. ATM’s Kinase Activity and Dynamic Response to DNA Damage D. ATM Substrates and Downstream Pathways V. ATM Functions: Lessons From Knockout Mice A. Phenotype of Atm-Deficient Mice B. Insights into Atm-Mediated Damage Responses C. Maturation of Immune System Genes, Immunodeficiency, and Lymphoid Tumors D. Role of ATM in Meiosis and Telomere Dynamics E. Atm, Neurogenesis, and Neurodegeneration VI. ATM Deficiency Leads to Increased Oxidative Stress VII. Interplay with Signaling Pathways Associated with Growth and Differentiation VIII. Defects in DNA Damage Response and Cancer Predisposition IX. Conclusions References
One of the cornerstones of the web of signaling pathways governing cellular life and differentiation is the DNA damage response. It spans a complex network of pathways, ranging from DNA repair to modulation of numerous processes in the cell. DNA double-strand breaks (DSBs), which are formed as a result of genotoxic stress or normal
Advances in CANCER RESEARCH 0065-230X/01 $35.00
209
C 2001 by Academic Press. Copyright All rights of reproduction in any form reserved.
210
Yosef Shiloh and Michael B. Kastan
recombinational processes, are extremely lethal lesions that rapidly mobilize this intricate defense system. The master controller that pilots cellular responses to DSBs is the ATM protein kinase, which turns on this network by phosphorylating key players in its various branches. ATM is the protein product of the gene mutated in the human genetic disorder ataxia-telangiectasia (A-T), which is characterized by neuronal degeneration, immunodeficiency, sterility, genomic instability, cancer predisposition, and radiation sensitivity. The clinical and cellular phenotype of A-T attests to the numerous roles of ATM, on the one hand, and to the link between the DNA damage response and developmental processes on the other hand. Recent studies of this protein and its effectors, combined with a thorough investigation of animal models of A-T, have led to new insights into the mode of action of this master controller of the DNA damage response. The evidence that ATM is involved in signaling pathways other than those related to damage response, particularly ones relating to cellular growth and differentiation, reinforces the multifaceted nature of this protein, in which genome stability, developmental processes, and cancer cross paths. C 2001 Academic Press.
I. INTRODUCTION Cell life is governed by a highly structured network of biochemical pathways that evolved to maintain its metabolism and, in higher organisms, to allow it to carry out specific functions according to the tissue context. This carefully laid out plan of operation may be perturbed, however, by unexpected environmental stimuli or by physical and chemical agents that alter or damage cellular constituents. Notable among these are agents that damage the DNA. Cellular responses to such damage include the activation of repair mechanisms as well as signaling pathways that alert the cellular systems, such as the cell cycle machinery, to the presence of critical damage. Our understanding of cellular responses to environmental damage has advanced considerably due to careful dissection of these pathways and elucidation of the players. Particularly important in this regard are proteins responsible for the initial sensing of the damage and that mobilize the entire system of responses. The ATM protein is a an example of such a master controller that acts in the context of a specific type of DNA damage—the DNA double-strand break (DSB). This protein is a member of a diversified family of proteins in various eukaryotes which share functions and structural motifs. ATM was discovered by virtue of being the product of a gene responsible for a human genetic disorder ataxia-telangiectasia (A-T). This is a typical case of the discovery of a novel physiological function following an attempt to understand the molecular basis of a genetic disorder. Since its discovery (Savitsky et al., 1995a,b), a significant amount of knowledge has been obtained about ATM, illuminating mainly its central role in the DNA damage response. However, recent information points also to its possible
ATM
211
involvement in other cellular processes, portraying ATM as an unusually multifaceted player in the web of signaling pathways. This review concentrates primarily on the biochemical and functional aspects of the ATM protein, its downstream pathways, and its relationships with various cellular processes. Genetic aspects of ATM sequence variations are discussed briefly.
II. ATAXIA-TELANGIECTASIA: A DISEASE CAUSED BY ATM DEFICIENCY A. Clinical–Pathological Phenotype The genetic disorder A-T, which is caused by lack or inactivation of the ATM protein, was the starting point in understanding ATM’s functions. A-T is inherited in an autosomal-recessive manner and is found in frequencies of 1:40,000–1:100,000 live births in various populations. The complex clinical and cellular phenotype of A-T indicates a defect in a multi-branched junction of physiological pathways, some of which are intimately connected, whereas others diverge in apparently unrelated directions (Crawford, 1998; Lavin and Shiloh, 1997, 1999; Meyn, 1999; Shiloh, 1997) Clearly, the defective protein is uncommonly multifaceted. The clinical manifestations of A-T affect numerous organs. The major feature is cerebellar ataxia (lack of balance), which appears in early infancy and gradually develops into severe neuromotor dysfunction. This ataxia is a reflection of progressive degeneration of the cerebellar cortex, particularly the loss of Purkinje and granule cells. Other parts of the central nervous system may show degenerative changes at a later age. Telangiectasias (dilated blood vessels) variably appear in the eyes and sometimes on the facial skin or ears. Primary immunodeficiency is expressed as a general defect in humoral and cellular immune responses, reduction in circulating T cells, and deficiencies of immunoglobulin classes IgA, IgG2, IgG4, and IgE. About 50–80% of the patients in various populations exhibit recurrent sinopulmonary infections, but their occurrence is not correlated with the degree of immunodeficiency. A-T patients also exhibit variable signs of premature aging, severe degeneration of the thymus and gonads, and growth retardation. Another hallmark of A-T is a marked predisposition to malignancies (Taylor et al., 1996; Xu, 1999). Although lymphoid malignancies of both B and T cell origin and several forms of leukemias are predominant in younger patients, older ones also develop epithelial tumors. Malignancy is the cause of death in about 10% of A-T patients. Early attempts at radiotherapy revealed a striking sensitivity of
212
Yosef Shiloh and Michael B. Kastan
A-T patients to the cytotoxic effects of ionizing radiation (IR), precluding this treatment altogether. Other laboratory findings include a high level of serum α-fetoprotein and carcinoembryonic antigen, insulin-resistant diabetes, and occasional mild hepatic dysfunction evidenced by abnormal enzyme levels. Clearly, A-T provides a striking combination of tissue-specific degenerative processes with malfunction of major cellular responses to environmental agents and acute cancer predisposition. The cellular phenotype of this disease adds another dimension to this complex picture.
B. The Cellular Phenotype 1. GENERAL CHARACTERISTICS AND GENOMIC INSTABILITY Cell lines derived from A-T patients have been valuable for further characterization of the complex physiological defect in this disorder (Shiloh, 1995; Lavin and Shiloh, 1997, 1999). The two hallmarks of the cellular A-T phenotype are genomic instability and abnormal response to IR and radiomimetic chemicals. A striking characteristic is premature senescence of primary A-T fibroblasts. A-T fibroblasts also show abnormalities of gross morphology and cytoskeletal organization due to alterations in actin stress fibers (McKinnon and Burgoyne, 1985; Shalev et al., submitted). Genomic instability is exhibited in various types of cells from A-T patients as a high rate of chromosomal breaks. Importantly, peripheral lymphocytes from these patients may show clonal chromosomal translocations involving primarily the sites of the immunoglobulin and T cell receptor genes at 7p14, 7q35, 14q11, and 14q32. These translocations often herald the appearance of malignancy (Taylor et al., 1996; Xu, 1999). Shortened telomeres and increased end-to-end association between chromosomes are an important facet of genomic instability in A-T (Metcalfe et al., 1996; Smilenov et al., 1997). Recent studies of telomere dynamics disclosed abnormalities in telomere clustering and telomere–nuclear matrix association in spermatocytes of Atmdeficient mice and A-T fibroblasts (Pandita et al., 1999; Pandita and Dhar, 2000; Scherthan et al., 2000; Smilenov et al., 1997, 1999). These observations extend previous results showing abnormal nuclear matrix organization in A-T cells (Pandita and Dhar, 2000).
2. DEFECTIVE RESPONSE TO DNA DSBs The most extensively studied aspect of the cellular phenotype of A-T is the abnormal response to IR and radiomimetic chemicals. Cells respond to DNA damage by activating a broad array of physiological responses. Although the primary responses are aimed at repairing the damage and allowing cell
ATM
213
survival, the cells may also opt for programmed cell death (apoptosis) when the damage load is incompatible with continuation of cellular life (Lowndes and Murguia, 2000; Schmid-Ullrich et al., 2000). A critical response in the survival pathway is temporary arrest of the cell cycle caused by activation of the cell cycle checkpoints. These checkpoints are normally activated at different phases of the cell cycle to delay its progression and provide time for repair before the important phases of DNA replication and mitosis are initiated (Dasika et al., 1999; Zhou and Elledge, 2000). The critical DNA lesion induced by IR and radiomimetic chemicals is the DSB. Most if not all cellular responses to this type of DNA damage are defective in A-T cells, most notably the activation of the cell cycle checkpoints. DSBs may be induced by exogenous agents or formed in the course of oxidative metabolism or natural processes, such as meiotic recombination or the maturation of the immune system genes. Elaborate repair systems have evolved to rapidly repair this extremely toxic lesion (Karran, 2000). The major repair systems for DSBs are based on either homologous recombination (HR) or nonhomologous end joining (NHEJ), which are carried out by multiprotein complexes (Haber, 2000; Karran, 2000). Major players in the NHEJ pathway are the DNA-dependent protein kinase and its associated Ku proteins, DNA ligase IV and XRCC4; HR is conducted primarily by a protein complex associated with the Rad51 protein. Another protein complex containing the Rad50, Mre11, and Nbs1 proteins appears to be involved in both processes (Petrini, 1999). Of note, the Brcal protein and possibly some of the mismatch repair proteins may also be involved in DSB repair (Karran, 2000). Although it has always been suspected that a defect in DSB repair underlies the radiation sensitivity of A-T cells, many studies reported no differences between A-T and normal cells in rejoining DSBs (Hariharan et al., 1981; Lavin and Davidson, 1981). Despite the absence of a measurable gross defect in closing DSBs in A-T, there is clearly an increase in spontaneous and radiation-induced chromosomal breaks (Taylor et al., 1976; Cornforth and Bedford, 1985; Morgan et al., 1997; Takao et al., 1999). The common notion today is that A-T cells are indeed defective in repairing a certain fraction of DSBs, and that this defect is probably responsible for the exquisite sensitivity of these cells to DSB-inducing agents (Jeggo et al., 1998). A subtle defect in recombination processes may provide one mechanistic explanation. Using reactivation of reporter plasmids dependent on recombination as the readout, an increased frequency of error-prone DNA recombination has been reported in A-T cells (Powell et al., 1993; Luo et al., 1996). There have also been reports of high rates of spontaneous intrachromosomal recombination in A-T cells (Meyn, 1993) and Atm-deficient mice (Bishop et al., 2000). The Rad51-associated protein complex has a major role in the HR mode of DSB repair. Chen et al. (1999) reported defective c-Abl-mediated phosphorylation of the Rad51 protein and subsequent association of the Rad51 and Rad52
214
Yosef Shiloh and Michael B. Kastan
proteins in A-T cells. On the other hand, direct functional links were recently established between ATM and important players in DSB repair pathways. Although the DSB repair defect in A-T cells is hardly detectable experimentally, the defective activation of the cell cycle checkpoints is readily observed. The arrest of A-T cells at the G1/S and G2/M boundaries is severely retarded. The reduction in semiconservative DNA synthesis that normally follows DNA damage induction and represents the S-phase checkpoint is greatly diminished in irradiated A-T cells, a phenomenon termed “radioresistant DNA synthesis.” A major player in the induction of the G1/S checkpoint (and to a certain extent also in the maintenance of the G2/M checkpoint) is the p53 protein, which is activated and stabilized following DNA damage. This process is mediated by many posttranslational modifications of this protein, including phosphorylations, dephosphorylation, and acetylation (Oren, 1999; Colman et al., 2000). Notably, p53 activation and accumulation as well as these modifications are severaly defective in A-T cells. Other mediators of these pathways are replication protein A (RPA), which is thought to be involved in the S-phase checkpoint (Iftode et al., 1999), and the Chk1 and Chk2 protein kinases that play a crucial role in the G2/M checkpoint; Chk2 is also involved in the G1/S checkpoint (Hirao et al., 2000; Chehab et al., 1999, 2000; Shieh et al., 2000; Tominaga et al., 1999). Interestingly, the Nbs1 protein, which is part of the DSB repair complex Rad50/Mre11/Nbs1, also seems to be involved in the S-phase checkpoint pathway (Petrini, 2000; Lim et al., 2000). The activation of these pathways is mediated by ATM when the damage is caused by DSBs. In most cases, this functional link is based on ATM-mediated phosphorylation of key players in these pathways. Although the initial cellular responses to DSBs are directed toward the cell’s survival, excessive amounts or specific types of damage may instead initiate the apoptotic response. Karlseder et al. (1999) showed that exposed chromosome ends, created by interfering with telomere formation, induce apoptosis in wild-type cells but not in A-T cells or cells lacking the p53 protein. This interesting response indicates a role of the A-T gene product, ATM, in a pathway that rather than attempting to save the cells leads them directly to programmed cell death in response to the exposed telomeric DNA ends. The cellular response to DNA damage is not confined to DNA repair and cell cycle arrest. Although these two processes predominate, the presence of DNA damage affects numerous aspects of cellular signaling, including the stress responses mediated by the transcription factor NF-κB and the mitogen-activated protein (MAP) kinases. Again, when the trigger is DSBs, these responses are ATM dependent and therefore impaired in A-T cells. NF-κB is best known for activating gene expression in response to external stress-related agents, such as proinflammatory cytokines, bacterial lipids,
ATM
215
and viruses (Mercurio and Manning, 1999); however, it also responds to DNA damage, which is clearly a nuclear signal. Activation of NF-κB is mediated by its translocation into the nucleus from the cytoplasm, where it is sequestered by its inhibitor, IκBα. The NF-κB is released from this inhibitor upon IκBα degradation following its phosphorylation by the IκB kinase IKK (Karin, 1999). This process is defective in irradiated A-T cells (Lee et al., 1998; Piret et al., 1999). Li et al. (2000) traced this defect back to IKK activation following DSBs, which is defective in human A-T cells and tissues of Atm-deficient mice. It is noteworthy that in A-T cells this pathway responds normally to proinflammatory triggers. MAP kinases are involved in cellular signaling: They play a pivotal role in cellular responses to extracellular stimuli induced by cytokines, growth factors, and environmental stress; they mediate inflammatory responses and specific developmental processes; and they respond to DNA damages. MAP kinases are activated via phosphorylation cascades and in turn phosphorylate transcription factors or other kinases (Ichijo, 1999; Widmann et al., 1999). The three major MAP kinase pathways are mediated by the extracellular signal-regulated protein kinase, the c-Jun N-terminal kinases (JNKs; also termed stress-activated protein kinases), and the p38 MAP kinase. The latter two branches of this system can respond specifically to cellular stress, including DNA damage, and their response to DSBs is defective in A-T cells. Thus, the activation of JNK, which normally leads to phosphorylation of the c-Jun protein and increased DNA binding activity of the transcription factor AP-I, is not observed in A-T cells (Lee et al., 1998; Shafman et al., 1995). Bar-Shira et al. (submitted for publication) recently showed that the induction of the JNK activator SEK by DSBs is retarded in A-T cells. One of ATM’s downstream effectors is the c-Abl protein. Kharbanda et al. (2000) reported that MEKK1, the protein kinase that activates SEK, is activated directly by c-Abl following genotoxic stress, indicating the possible link between ATM and the JNK pathway in the DNA damage response. Similarly, the activation of the gamma isoform of the p38 MAP kinase by IR is defective in A-T cells (Wang et al., 2000). Interestingly, Bar-Shira et al. noticed in A-T cells a significant delay in the activation of the gene encoding MAP kinase phosphatase-5, which is a dual-specificity phosphatase that inactivates, and therefore terminates, the response of JNK and p38 (Tanoue et al., 1999; Theodosiou et al., 1999). Thus, the entire cycle of activation and deactivation of these MAP kinases following DSB induction is ATM dependent. JNK activation by DNA-damaging agents is part of the apoptotic response (Chen et al., 1996; Verheij et al., 1998). Of note, JNK is an important mediator of apoptosis in the central nervous system (Mielke and Herdegen, 2000). Herzog et al. (1998) observed resistance to IR-induced apoptosis in the central nervous system of Atm−/− mice. On the other hand, the involvement of JNK and p38 in T cell-mediated immune response and thymocyte
216
Yosef Shiloh and Michael B. Kastan
maturation (Rincon et al., 2000), which are defective in A-T patients, may suggest links between ATM and the MAP kinases in processes involving cellular responses to DSBs that are not due to DNA-damaging agents. As our understanding of the cellular response to DNA damage evolves, it has begun to take the form of a multibranched network linked to major signaling processes. Clearly, different pathways and processes are activated by different types of damage. The strong control exerted by ATM on all branches of the response to DSBs is well established and mechanistically understood.
3. DEFECTS IN OTHER SIGNALING PATHWAYS Although most of the literature describing the cellular phenotype in A-T has focused on the defective DNA damage response, sporadic reports have repeatedly indicated defects in other signaling pathways not associated with DNA damage. The defective response to mitogens of T cells from A-T patients is well documented (Lavin and Shiloh, 1997; Kondo et al., 1993), suggesting that these cells have a defect in a calcium-dependent pathway initiated at the CD3 complex. Interestingly, defective mobilization of calcium ions from internal stores was demonstrated in A-T cells following mitogenic stimulation (Famulski and Paterson, 1999; Khanna et al., 1997) and IR treatment (Yan et al., 2000). Abnormal calcium metabolism was also observed in Purkinje cells of Atm-deficient mice (Chiesa et al., 2000). Khanna et al. (1997) identified in B cells from A-T patients a broad defect in cellular responses to cross-linking of the B cell receptor. Rhodes et al. (1998) demonstrated in cultured A-T cells a defect in depolarization in response to extracellular potassium ions and defective potassium currents across the cellular membrane. Although the defects in these pathways may be secondary to the disturbed function of the cellular membrane of A-T cells due to oxidative damage, they may represent bonafide involvement of the ATM protein in these pathways. ATM’s involvement in the insulin response is an example of this thesis.
III. THE ATM GENE AND ITS MUTATIONS A. The A-T Gene, ATM The gene responsible for A-T, ATM, was identified by positional cloning (Savitsky et al., 1995a,b) and found to extend over 150 kb spanning 66 exons (Platzer et al., 1997; Uziel et al., 1996). This gene encodes multiple mRNA species that share a single open reading frame encoding the ATM protein,
ATM
217
flanked by a variety of 3′ and 5′ untranslated regions (UTRs) (Savitsky et al., 1997). Although the significance of these multiple UTRs has not been studied experimentally, their existence suggests posttrancriptional control of ATM levels under certain physiological conditions. An extensive repertoire of ATM mutations was found in A-T patients. The majority of these mutations truncate the ATM protein, whereas amino acid substitutions are in the minority among ATM alterations leading to the disease (Broeks et al., 1998; Castellvi-Bel et al., 1999; Concannon and Gatti, 1997; Gilad et al., 1996, 1998a,b; Laake et al., 2000; Sandoval et al., 1999; Sasaki et al., 1998; Savitsky et al., 1995a; Telatar et al., 1996; Teraoka et al., 1999; Wright et al., 1996). Since truncated ATM derivatives are extremely unstable, A-T cells are usually completely devoid of ATM protein (BeckerCatania et al., 2000). Various genetic combinations involving milder mutations that leave residual amounts of ATM give rise to less severe clinical phenotypes termed “A-T variants” (Gilad et al., 1998a; McConville et al., 1996). Additional experiments confirmed that the deficiency of ATM’s protein product is indeed responsible for the A-T cellular phenotype. Thus, ectopic expression of recombinant ATM protein in A-T cells complemented various features of this phenotype (Zhang et al., 1997; Ziv et al., 1997), whereas downregulation of ATM using antisense strategies conferred such features to various cell lines (Fan et al., 2000; Uhrhammer et al., 1999; Zhang et al., 1998). Furthermore, expression of ATM fragments containing the leucine zipper of this protein abrogated the S-phase checkpoint and increased the radiosensitivity of the human tumor cell line RKO (Morgan et al., 1997). Such protein fragments appear to act in a dominant-negative fashion, possibly by competing with ATM for a specific interactor(s).
B. Somatic ATM Mutations in Cancer An important new link was found between ATM sequence variations and cancer by searching for such variations in tumor tissues. These investigations were driven by the hypothesis that ATM could act as a tumor suppressor gene in certain tumors, particularly the lymphoreticular malignancies that are common among A-T patients. ATM was expected to be selectively inactivated in the tumor cells via a mechanism involving sequence alterations and loss of heterozygosity. These studies indeed revealed such a phenomemon in three specific hematopoietic malignancies: T cell prolymphocytic leukemia, B cell chronic lymphocytic leukemia (B-CLL), and mantle cell lymphoma (Bullrich et al., 1999; Schaffner et al., 1999, 2000; Starostik et al., 1998; Stilgenbauer et al., 1997, 2000; Stankovic et al., 1999; Stoppa-Lyonnet et al., 1998; Vorechovsky et al., 1997; Yuille et al., 1998). In all three malignancies, significant portions of various patient cohorts showed inactivation of both
218
Yosef Shiloh and Michael B. Kastan
ATM alleles entailing only the malignant cells. Absent or reduced expression of ATM protein was noticed in a significant fraction of B-CLL tumors (Stankovic et al., 1999; Starostik et al., 1998). Starostik et al. also noted lower survival among patients with reduced or absent ATM in the tumors. These observations indicate a critical role for ATM in maintaining proper growth and differentiation of lymphocytes of both lineages. This function could conceivably tie in with ATM’s important role in protecting the normal cellular life cycle during the genomic rearrangements that accompany the maturation of the immune system genes. The ATM gene was inactivated in these tumors by gross rearrangements, by truncating mutations, and by amino acid substitutions. Importantly, the proportion of the last type of gene alteration was significantly higher than that of the germline mutations that cause A-T, indicating the importance of missense mutations in ATM-driven malignancies (Gatti et al., 1999).
IV. THE ATM PROTEIN: FROM SEQUENCE TO FUNCTION A. ATM Sequence Motifs: A Membership Card in the PIK-Related Kinase Family The open reading frame of ATM transcripts predicts a large protein containing 3056 amino acids (Savitsky et al., 1996). The predominant domain of this protein is the carboxy-terminal region of about 350 residues, which contains sequence signatures similar to those of the catalytic subunit of phosphatidylinositol 3-kinases (PI3Ks) (Fig. 1A). These motifs, which are shared with the lipid kinases that play major roles in various signaling pathways, suggest a signaling role for ATM. They also place it within a multibranched family of large proteins with the carboxy-terminal PI3K signatures (PIK-related kinases), which is represented in all eukaryotes Kuruvilla Fig. 1 Bioinformatic analysis of ATM’s amino acid sequence. (A) Basic motifs and protein domains identified in the ATM protein. NLS, nuclear localization signal; HEAT, sequence element common to the Huntington protein, elongation factor 3 (EF3), the regulatory A subunit of protein phosphatase 2a (Pp2A), and the Tor1p protein; FAT, a domain found in FRAP, ATM, and TRRAP proteins; PI3K, a region containing signatures of the catalytic subunit of phosphatidylinositol 3-phosphate kinases; FATC, a carboxy-terminal domain common to the PIK-related kinases. (B) A phylogenetic tree of the PIK-related protein kinase superfamily. The tree was generated using the neighbor-joining method, based on multiple alignment of the proteins generated by CLUSTALW (Thompson et al., 1994) and drawn using TREEVIEW software (Page, 1996). The bar marked “0.1” corresponds to 10% sequence divergence.
220
Yosef Shiloh and Michael B. Kastan
and (Schreiber, 1999; Lavin and Khanna, 1999; Rotman and Shiloh, 1999; Fig. 1B). A few other motifs identified along the rest of the large ATM molecule (Fig. 1A) include several versions of a nuclear localization signal and also a leucine zipper that probably represents a region of interaction with other proteins. Two domains, FATC (about 30 amino acids) and FAT (about 500 amino acids), whose functional significance is not clear, are common to many PIK-related kinases (Bosotti et al., 2000; Keith and Schreiber, 1995). An interesting motif is the HEAT repeat, which is common to the Huntington protein, elongation factor 3, the regulatory A subunit of protein phosphatase 2a, and the Tor1p protein. These repeats of 37–43 amino acids, which contain two α helices, were identified in many nuclear and cytoplasmic proteins with diverse functions, some entailing protein trafficking (Andrade and Bork, 1995). Interestingly, numerous putative phosphorylation sites extend over most of the molecule but do not span the PI3K domain (E. Levanon, unpublished observation). Although ATM’s sequence provides relatively scant information about its cellular functions, identification of its catalytic activity as a protein kinase has been the major clue to understanding its mode of action. Most members of the PIK-related kinases possess a protein kinase activity directed at serine/threonine residues. These kinases function at the top of various signaling cascades that are involved in sensing specific types of stresses or stimuli related to cellular growth, and are critical for the activation of cellular responses to these stimuli. Importantly, the majority of these proteins are involved in sensing and responding to DNA damage and in maintaining genomic stability. The current phylogenetic tree of the PIK-related kinases (Fig. 1B) distributes them into several subfamilies, each of which can be referred to by its human member. The ATM-related subfamily includes ATM orthologs in various species ranging from mouse to yeast. Murine ATM shows 84% identity and 91% similarity with the human protein (Pecker et al., 1996); judging from the phenotype of Atm knockout mice, the human and murine proteins share many functions. The Xenopus laevis ortholog (xAtm) (Robertson et al., 1999) is involved in the activation by DSBs of an S-phase checkpoint pathway that acts at the formation of the prereplication complex (Costanzo et al., 2000). In view of the similarity between human ATM and xAtm (Robertson et al., 1999), it is expected that the latter will be involved in additional damage-induced pathways. Functional analyses of the Drosophila melanogaster ortholog (CG6535 gene product) and the Arabidopsis thaliana ortholog (Garcia et al., 2000) have not been reported. An extensively studied ATM ortholog is Tel1p in the budding yeast Saccharomyces cerevisiae, whose ectopic expression in human A-T cells partially complements their phenotype (Fritz et al., 2000). Tel1p and another PIK-related protein kinase in the budding yeast, Mec1p, closely collaborate
ATM
221
in maintaining telomere length and mediating cellular responses to DNA damage. Tel1p, which has protein kinase activity (Mallory and Petes, 2000), has been studied mainly with regard to its role in maintaining telomere length (Morrow et al., 1995; Greenwell et al., 1995; Ritchie et al., 1999). TEL1 mutations lead to shortened telomeres, a typical characteristic of human A-T cells. Interestingly, in this function Tel1p cooperates with the DSB repair complex Mre11/Rad50p/Xrs (Ritchie and Petes, 2000), whereas in human cells ATM functionally interacts with the orthologous protein complex in the DNA damage response. However, Tel1p does not seem to be directly involved in DSB repair (Boulton and Jackson, 1998). The synergism between Tel1p and Mec1p in both telomere maintenance and cellular responses to DNA damage has been clearly documented in double-mutant tel1/mec1, which are extremely sensitive to DNA damage and senesce prematurely (Greenwell et al., 1995; Ritchie et al., 1999; Vialard et al., 1998). Although tel1 mutants are not hypersensitive to DNA-damaging agents, an extra copy of TEL1 can largely complement such sensitivity in mec1 mutants (Morrow et al., 1995). The Tel1 protein in the fission yeast Schizosaccharomyces pombe has similar roles (Matsuura et al., 1999). The closest human ATM homolog is the ATR protein, which represents a separate branch of the PIK-related family (Fig. 1B). ATR (a homolog of ATM and Rad3) is a protein kinase that shares many substrates with ATM and acts in partial redundancy with ATM in many damage response pathways. ATR orthologs are typically involved in maintaining genome stability and DNA damage responses. Thus, mei-41 mutants in the fruit fly exhibit chromosomal instability, radiation sensitivity, and defective activation of cell cycle checkpoints by IR (Hari et al., 1995). Recombination-defective lines of this organism reveal a role for Mei-41 protein in meiotic precocious anaphase in females (McKim et al., 2000). Complete deficiency of the Mei-41 protein is embryonic lethal since the checkpoint function of this protein is essential at midblastula transition (Sibon et al., 1999). The protein Ce-atl-1 in the nematode C. elegans was found to be essential in early embryogenesis and seems to be involved in controlling mitotic chromosome segregation (Aoki et al., 2000). The Aspergillus nidulans ortholog in this group, UvsB, controls several damage-induced pathways, including cell cycle and septum formation checkpoints (Harris and Kraus, 1998; Hofmann and Harris, 2000) as well as a mitotic checkpoint that is not associated with DNA damage (De Souza et al., 1999). The yeast ATR orthologs Mec1p in the budding yeast and Rad3p in the fission yeast are pivotal in elaborate networks that maintain genomic stability and activate DNA damage responses, exerting their effect by phosphorylating key proteins in these pathways (Zhou and Elledge, 2000; Humphrey, 2000). As mentioned previously, the S. cerevisiae ortholog of human ATR, Mec1p, collaborates with the ATM ortholog, Tel1p, in the damage response
222
Yosef Shiloh and Michael B. Kastan
pathway and telomere maintenance (Neecke et al., 1999; Longhese et al., 2000). Notably, Mec1p-dependent processes include cell cycle checkpoints induced by DNA damage or replication arrest (Clarke et al., 1999; Vallen and Cross, 1999; Santocanale and Diffley, 1998), redistribution of the Sir3 silencing protein from telomeres to DSBs (Mills et al., 1999), meiotic recombination (Gruschcow et al., 1999), and silencing gene expression at telomeres (Craven and Petes, 2000). The Rad3p of S. pombe is critical for the immediate activation of cell cycle checkpoints following DNA damage and for the activation and maintenance of the responses to replication arrest (Humphrey, 2000). The mechanisms in each response are different and carried out by different effectors (Martinho et al., 1998). Many of the relevant pathways are conserved through evolution; hence, in mammalian cells, ATM and ATR phosphorylate orthologs of Mec1p and Rad3p substrates. The catalytic subunit of the DNA-dependent protein kinase (DNA-PKcs) is centered in another branch of the PIK-related kinase phylogenetic tree (Fig. 1B). This large protein kinase is involved in the major mode of DSB repair in mammalian cells via nonhomologous end joining (Karran, 2000; Haber, 2000). DNA-PKcs is recruited to DSB sites by the Ku70/Ku80 heterodimer and appears to play a central role in the rejoining process (Smith and Jackson, 1999). DNA-PK exhibits vigorous protein kinase activity in vitro, which is strongly stimulated by DNA ends. The role of this kinase activity in vivo is unclear, however, since its numerous in vitro substrates do not seem to be phosphorylated in cells in a DNA-PK-dependent manner. Moreover, ATM- and ATR-dependent pathways appear to be completely independent of DNA-PK (Araki et al., 1999; Khosravi et al., 1999). However, the critical role of DNA-PK in DSB response is underscored by the scid phenotype of DNA-PK-deficient mice, which includes radiosensitivity, chromosomal instability, immunodeficiency, and cancer predisposition (Jhappan et al., 1997). Clearly, DNA-PK has a unique and critical role in the cellular response to DSBs which is distinct from the roles of ATM and ATR. The TOR subfamily of the PIK-related kinases (Dennis et al., 1999; Kuruvilla and Schreiber, 1999) represents a group of proteins whose duty is to convey, via their kinase activity, extracellular signals rather than DNA damage. The TOR proteins signal nutrient availability (e.g., amino acids levels) and certain mitogenic stimuli to the mRNA translation machinery. One known mechanism is TOR-dependent phosphorylation of the p70s6k kinase. The pathways controlled by the mammalian homolog in this group, mTOR /FRAP/RAFT, may regulate both protein translation and transcription of rRNA and tRNA. Importantly, TOR-mediated pathways appear to intersect with certain PI3K pathways. Finally, members of the TRRAP branch of the PIK-related kinase superfamily are devoid of kinase activity due to inactivating sequence alterations in their kinase domain. The mammalian (TRRAP/PAF400) and yeast (Tra1p)
ATM
223
homologs are part of large protein complexes involved in transcriptional regulation and chromatin remodeling via their histone actetyltransferase activity (Grant et al., 1998; Saleh et al., 1998; Vassilev et al., 1998). TRRAP/PAF400 is a coactivator of the c-Myc and E2F transcription factors and recruits the histone acetyltransferase GCN5 to c-Myc (McMahon et al., 1998, 2000). The existence of the TRRAP group in the PIK-related kinase family raises the possibility that members residing in other branches of the PIK-related kinase phylogenetic tree are involved in gene regulation and shaping the chromatin. In general, despite the apparently diverse functions of the different branches of this important protein family, they may share many more functions.
B. The ATM Protein: Initial Characterization ATM is an abundant, highly phosphorylated, 370-kDa protein (Chen et al., 1996). Conceivably, the PI3K-like carboxy-terminal domain harbors the catalytic site of this protein. Laboratory studies have not been informative in elucidating the functions of the remaining portion of the protein, which represents about 90% of its sequence, other than binding of certain substrates. The nonkinase portions of the protein may participate in one or more of the following activities: (i) intracellular localization of the protein, (ii) regulation of the activity of the kinase in response to particular stimuli, (iii) maintenance of protein stability, (iv) facilitation of ATM proximity to or binding to its substrates, or (v) docking of other ATM-interacting proteins. Elucidation of the function of these other regions is complicated by the fact that ATM protein-containing deletions or ATM fragments are unstable in cells and thus difficult to express and study. Although ATM generally exhibits ubiquitous tissue expression (Savitsky et al., 1995a), cell lines in culture express detectable levels of ATM protein as a general rule, although the relative amounts can vary (unless the ATM alleles are mutated). In vivo distribution of ATM protein has not been fully characterized, however, partly because of the limited availability of anti-ATM antibodies that specifically recognize the ATM protein in well-controlled immunohistochemical studies. This deficiency is especially problematic in mouse tissues, with which such in vivo protein studies are typically done. Given the predicted role of ATM in cellular responses to DNA damage, it is not surprising that it is localized in the nucleus of cells (Brown et al., 1997). Subsequent studies continued to support the view that ATM is predominantly nuclear in proliferating cells, both in culture and in vivo, although there are exceptions to this generalization. Initial clues that ATM may have extranuclear functions in some cell types derived from the finding that it is associated with cytoplasmic vesicles (Watters et al., 1997) and bound to the
224
Yosef Shiloh and Michael B. Kastan
cytoplasmic vesicular protein β-adaptin (Lim et al., 1998). Further support for this notion was obtained by the following results: Some ATM protein is found in peroxisomes (Watters et al., 1999); ATM protein appears to be in cytoplasmic vesicles according to immunohistochemical analysis of human brain tissues (Oka and Takashima, 1998); and ATM is predominantly cytoplasmic in mouse cerebellar Purkinje cells and other neurons (Barlow et al., 2000). Barlow et al. (2000) suggested that ATM protein is required to prevent lysosomal accumulation in neurons. Recent reports that ATM is activated by insulin treatment in certain settings, and that it phosphorylates and regulates the cytoplasmic protein 4E-BP1 after insulin treatment (Yang and Kastan, 2000), further suggest extranuclear functions for ATM protein. If the view of distinct, and possibly multiple, nuclear and cytoplasmic functions holds, they almost certainly represent cell-type specific differences and it will be interesting to speculate why one protein would have such seemingly disparate functional roles in the mammalian cell.
C. ATM’s Kinase Activity and Dynamic Response to DNA Damage Since the kinase domain of ATM is homologous to the lipid kinase, PI-3K, it was not clear initially whether it would phosphorylate lipids or proteins. Although PI-3K has both lipid and protein kinase functions, many of the PIK-related kinases have protein kinase activity and no detectable lipid kinase activity (Smith and Jackson, 1999). Therefore, it seemed likely that ATM would also be a protein kinase. This prediction was borne out when ATM was first shown to phosphorylate serine 15 of the p53 protein, both in vitro and in vivo (Banin et al., 1998; Canman et al., 1998; Khanna et al., 1998). Lipid kinase activity has not been reported to date, but the possibility is difficult to rule out absolutely. Optimization of in vitro conditions for assessing ATM kinase activity revealed differences between DNA-PK and ATM. In contrast to DNA-PK, ATM exhibited a dependence on Mn2+ and a lack of dependence on exogenously added DNA ends and Ku proteins for optimal basal kinase activity (Canman et al., 1998; Kim et al., 1999). Interestingly, in addition to its lipid kinase activity, the p110 catalytic subunit of PI-3K also has Mn2+-dependent protein kinase activity (Dhand et al., 1994). Similar to DNA-PK, ATM is inhibited by the fungal metabolite wortmannin (Banin et al., 1998; Sarkaria et al., 1998). Another inhibitor of ATM kinase activity is the radiosensitizing agent caffeine (Sarkaria et al., 1999; Zhou et al., 2000). The role of DNA ends in modulating ATM function is unclear. The in vitro phosphorylation of RPA by ATM was reported to be stimulated by the addition of exogenous DNA ends (Chan et al., 2000; Gately et al.,
ATM
225
1998), and atomic-force microscopy suggested that biochemically purified ATM binds to DNA ends (Smith et al., 1999). How these in vitro characteristics translate into in vivo cofactor requirements remains to be clarified. It is also conceivable that the presence of DNA ends can alter the conformation of a substrate, such as the RPA protein, and make it a better target rather than affect the ATM kinase directly. Recent findings on the dynamics of ATM association with nuclear structures following DNA damage, however, may indicate an association of ATM with DSBs. Early studies based on biochemical and immunofluorescence analysis suggested that activation of ATM is not accompanied by a change in its abundance or subcellular distribution (Brown et al., 1997; Gately et al., 1998; Lakin et al., 1997; Watters et al., 1997). Furthermore, the amount of ATM and its activity do not vary throughout the cell cycle (Gately et al., 1998; Pandita et al., 2000). ATM was shown to associate with the chromatin and the nuclear matrix (Gately et al., 1998), but the association was not altered following cellular exposure to IR. It is conceivable, however, that only a fraction of ATM molecules associate with sites of damage while the rest remain unbound, thus masking detection of the damage-associated fraction. Andegeko et al. (submitted for publication) tested this hypothesis by applying a stepwise detergent extraction protocol to cellular extracts and fixed cells; biochemical and immunofluorescence studies identified a fraction of ATM that was rapidly converted to an extraction-resistant form after damage infliction. This ATM fraction was retained in nuclear aggregates, which colocalized with foci of two proteins previously shown to occur at sites of DSBs, Nbs1 (Maser et al., 1997; Nelms et al., 1998), and γ -H2AX (the phosphorylated form of histone H2AX) (Rogakou et al., 1999). These findings indicate a rapid association of a fraction of ATM with sites of DSBs, in agreement with its role in the immediate, early signaling of this DNA damage. A well-documented, dynamic response of ATM to DSBs is enhancement of its kinase activity. This enhancement, commonly referred to as ATM activation, was observed following treatment of cells with IR, the radiomimetic NCS, and the topoisomerase inhibitor etoposide, but not UV radiation (Banin et al., 1998; Canman et al., 1998; Uzeil et al., unpublished results). Since DSB-inducing agents do not increase the levels of ATM protein (Brown et al., 1997; Canman et al., 1998), increased ATM kinase activity in cells after irradiation results, at least in part, from an increase in the specific activity of the enzyme. Similarly, it was recently reported that insulin treatment of starved cells can increase the activity, but not the level, of ATM (Yang and Kastan, 2000). The mechanisms by which DNA strand breaks or insulin increase the specific activity of ATM remain to be elucidated. Plausible mechanisms include posttranslational modifications of ATM protein or alterations in proteins that bind to ATM. Although the ATM-related protein ATR is also
226
Yosef Shiloh and Michael B. Kastan
important in DNA damage responses (Tibbetts et al., 1999) and shares many of the aforementioned biochemical characteristics with ATM (Kim et al., 1999; Gatei et al., 2001; Tibbetts et al., 1999, 2000), it does not appear to increase its intrinsic specific activity in response to DNA damage. However, ATR shows a dramatic relocalization at stalled replication forks following DNA damage or treatment with replication inhibitors (Tibbetts et al., 2000). Presumably, the cellular activity of ATR might be altered following this relocalization due to protein–protein interactions at the sites of the stalled replication forks, which are not maintained in the in vitro kinase assays done with immunoprecipitated protein. Since the measured enhancement of ATM activity after IR is only two- or three-fold by in vitro kinase assays (Banin et al., 1998; Canman et al., 1998), whereas the enhanced phosphorylation of cellular substrates seems much greater, it is likely that these other types of mechanisms also contribute to the activation of ATM in cells. Alternatively, only a portion of ATM might be activated (e.g., the portion that undergoes nuclear retention) at a rate much higher than three- or four-fold, but this enhancement of activity is masked by nonactivated ATM molecules. Similar to DNA-PK (Smith and Jackson, 1999) and ATR, the ATM kinase appears to phosphorylate only serine or threonine residues that are immediately followed by a glutamine (Kim et al., 1999; O’Neill et al., 2000). Furthermore, the presence of positively charged amino acids near the target serine or threonine in the substrate protein tends to diminish activity, whereas hydrophobic or negatively charged residues enhance it (Kim et al., 1999). The clarification of these target site preferences has greatly facilitated identification of sites within proteins that are phosphorylated by ATM in response to irradiation or other stimuli (Fig. 2). Proteins that have been validated as targets of the ATM kinase in cells, and whose target sites have been identified, include p53, Mdm2, Nbs1, Brca1, CtIP, Chk2, and 4E-BP1. The physiologic significance of these ATM targets is discussed next.
D. ATM Substrates and Downstream Pathways The first protein shown to interact with ATM was the c-Ab1 oncoprotein (Shafman et al., 1997), and it was suggested that c-Ab1 is a direct target of ATM’s kinase activity (Baskaran et al., 1997). Although c-Ab1 appears to mediate DNA damage signals to downstream effectors, such as p73, in an ATM-dependent manner (Yuan et al., 1999), it is not clear whether c-Ab1 is indeed a direct substrate of ATM in vitro or in vivo. The first validated target of the ATM kinase was the p53 protein. This protein is activated and its levels increase in the cell following the introduction of DNA strand breaks (Nelson and Kastan, 1994), in response to hypoxia (Graeber et al., 1994),
Fig. 2 A schematic diagram of signaling pathways activated in an ATM-dependent manner immediately following the induction of DSBs. (see text for details). A solid arrow indicated activation, whereas a dotted “T” bar indicates inhibition. A single arrow indicates the direct interaction between two proteins or a clearly defined, single step in a pathway. Two successive arrows indicates several steps, not all of which are clear. P, phosphorylation; PTM, posttranslational modifications.
228
Yosef Shiloh and Michael B. Kastan
and in response to decreases in ribonucleotide triphosphate pools (Linke et al., 1996). This increase results from a combination of enhanced translation and increased half-life of p53 protein (Giaccia and Kastan, 1998). The induced p53 protein causes either cell cycle arrest in the G1 phase of the cycle (Kastan et al., 1991) or programmed cell death (Lowe et al., 1993; Clarke et al., 1993). Cells from A-T patients fail to optimally induce p53 protein following IR (Kastan et al., 1992; Khanna and Lavin, 1993; Canman et al., 1994), thus implicating the ATM protein in p53 induction following DNA damage. Once it was reported that p53 becomes phosphorylated on serine 15 in response to DNA damage (Shieh et al., 1997; Siliciano et al., 1997), it was natural to inquire whether ATM was required for this posttranslational modification. First, in addition to exhibiting poor induction of p53, cells from A-T patients also displayed suboptimal phosphorylation of serine 15 on p53 (Siliciano et al., 1997). ATM was then shown to directly phosphorylate p53–serine 15 in vitro and to be activated by DNA damage (Banin et al., 1998; Canman et al., 1998; Khanna et al., 1998). Thus, the introduction of DNA strand breaks in cells induces phosphorylation of p53 on serine 15 in an ATM-dependent manner, and ATM protein directly phosphorylates p53 protein on this same site in vitro. Since ATM is required for both the induction of p53 (Kastan et al., 1992) and its phosphorylation on serine 15 after irradiation (Siliciano et al., 1997), it was conceivable that this phosphorylation event was important for the induction of p53 following DNA damage. However, mutation of this site did not affect p53 levels or its half-life, nor did it affect the ATM dependence of this process (Ashcroft et al., 1999; Khosravi et al., 1999). It has been suggested that serine 15 phosphorylation may facilitate the introduction of other posttranslational modifications in p53, such as phosphorylation or acetylation of other sites in the protein (Sakaguchi et al., 1998; Dumaz et al., 1999). A functional role for serine 15 phosphorylation was recently demonstrated by enhancing the binding of p53 protein to p300 and thus its transcriptional activity (Dumaz and Meek, 1999; Lambert et al., 1998). Mutation of the serine 18 site in a mouse model (which is equivalent to serine 15 in human p53) indeed affected p53 transcription activity (Chao et al., 2000a). Since ATM was required for the increase in p53 protein levels after IR, but serine 15 phosphorylation was apparently not involved in this induction, the mechanistic link between ATM activity and p53 induction remained a mystery. An explanation for this conundrum was recently suggested by the identification of two other ATM substrates that affect the half-life of p53 protein. ATM phosphorylates the checkpoint protein kinase Chk2 on threonine 68—a phosphorylation event that is thought to enhance the kinase activity of Chk2 (Zhou et al., 2000; Melchionna et al., 2000; Ahn et al., 2000; Matsuoka et al., 2000). In turn, Chk2 kinase phosphorylates p53 on serine 20. In contrast to serine 15 phosphorylation, this particular
ATM
229
posttranslational modification was reported to inhibit the binding of Mdm2 protein to p53, which would inhibit the targeting of p53 for degradation by Mdm2, thereby increasing the level of p53 protein (Chehab et al., 1999, 2000; Shieh et al., 2000). In addition, Mdm2 protein is a direct target of the ATM kinase. Mdm2 becomes phosphorylated after IR in an ATM-dependent manner (Khosravi et al., 1999), and the phosphorylation of Mdm2 on serine 395 by ATM appears to inhibit the ability of Mdm2 to export p53 from the nucleus (Maya et al., 2001). Thus, in response to DNA damage, ATM kinase simultaneously phosphorylates three different proteins that control p53 function: It phosphorylates p53 directly on serine 15 to enhance its transcriptional activity, it phosphorylates Chk2 (which in turn phosphorylates p53 on serine 20), and it phosphorylates Mdm2 to decrease the ability of Mdm2 to degrade p53, which leads to an increased level of p53 in the cell. By screening the protein sequence databases for proteins that contain potential ATM target sequences, the protein Nbs1 was identified as a possible ATM substrate (Kim et al., 1999). Nbs1 was a particularly attractive candidate substrate because it had been characterized as the product of the gene mutated in the Nijmegen breakage syndrome (NBS) (Carney et al., 1998; Varon et al., 1998). NBS and A-T share many similarities, including a markedly increased incidence of lymphoid malignancies, radiosensitivity, radioresistant DNA synthesis, and increased chromosomal breakage (Shiloh, 1997). An enzyme–substrate relationship was shown to exist between these two gene products when ATM phosphorylated Nbs1 protein on serine 343 in vitro (Kim et al., 1999; Lim et al., 2000), and phosphorylation of this same serine occurred in an ATM-dependent manner following irradiation of cells (Lim et al., 2000; Zhao et al., 2000; Wu et al., 2000; Gatei et al., 2000). It has also been suggested that ATM phosphorylates one or two additional sites in Nbs1 (Zhao et al., 2000; Wu et al., 2000). The functional significance of the phosphorylation of serine 343 by ATM was shown in the pathway leading to inhibition of DNA synthesis (i.e., the S-phase checkpoint) following IR (Lim et al., 2000; Zhao et al., 2000). How this particular phosphorylation event affects the initiation of replication forks in S phase is unknown. Inherited germline alterations in one copy of the BRCA1 gene strongly predispose women to develop breast and ovarian cancers (Martin and Weber, 2000). Although the exact function(s) of the Brca1 protein remains to be clarified, like ATM it clearly plays an important role in cellular responses to DNA damage (Chen et al., 1999; Deng and Brodie, 2000; Zhong et al., 1999). Cells with altered Brca1 function are radiosensitive and exhibit increased chromosomal aberrations. In addition, Brca1 protein exists in focal complexes with other proteins that are altered in the cellular responses to IR (Chen et al., 1999; Scully and Livingston, 2000). Brca1 is found in a large protein complex, BASC, that includes numerous other proteins involved in
230
Yosef Shiloh and Michael B. Kastan
cellular stress responses, including ATM, MLH1, and Nbs1 (Wang et al., 2000). Brca1 protein becomes hyperphosphorylated when cells are exposed to IR, and this phosphorylation is markedly diminished in cells from A-T patients (Tibbetts et al., 2000). ATM was found to phosphorylate Brca1 directly on several residues in vitro, and some of them were phosphorylated in vivo in an ATM-dependent manner (Cortez et al., 1999; Gatei et al., 2000a). The functional significance of ATM phosphorylation of Brca1 remains to be clarified, but attractive possibilities include involvement in the irradiation-induced S phase or G2 checkpoints and in determining radiosensitivity. Of note, the Chk2/Cds1 kinase, which is activated following DSB induction via ATM-mediated phosphorylation, phosphorylated Brca1 on another site (Lee et al., 2000). In addition, ATM has also been shown to directly phosphorylate CtIP, a protein that binds to and inhibits Brca1 function (Li et al., 2000). The phosphorylation of CtIP by ATM inhibits its binding to Brca1, presumably freeing up Brca1 to function after a cell has been irradiated. Since ATM phosphorylates multiple proteins to affect p53 function and the G1 checkpoint and phosphorylates at least two proteins to alter Brca1 function, it may become a recurring scenario that ATM phosphorylates multiple proteins in a complex to achieve its functional goals. A-T patients exhibit many symptoms that are not obviously attributable to alterations in DNA damage responses, suggesting that ATM has functional roles other than in damage responses. It can be explained how ATM’s role in DNA damage responses contributes to cancer predisposition, radiosensitivity, and even neurodegeneration and telangiectasias. It is more difficult to explain how abnormalities in DNA damage responses account for the cell membrane abnormalities and cytoskeletal alterations reported in A-T cells (Lavin and Shiloh, 1997) and insulin resistance and increased α-fetoprotein levels in A-T patients. The aforementioned observations that ATM can be found in the cytoplasm of some cells and can bind to the vesicular protein β-adaptin are further evidence of additional functions for the ATM protein. One extranuclear function and substrate of ATM was recently reported—The translational regulatory protein 4E-BP1, which was identified by in vitro screens. The site of in vitro phosphorylation was serine 111, the same site phosphorylated in cells in response to insulin stimulation in an ATM-dependent manner (Yang and Kastan, 2000). In addition, insulin treatment of growtharrested cells was found to increase the specific activity of ATM in certain cell types, analogous to the activation by irradiation. Finally, a functional significance of this insulin-stimulated, ATM-dependent phosphorylation event was demonstrated: It appears to facilitate the insulin-stimulated release of 4E-BP1 from the translation factor eIF-4E. Whether the participation of ATM in this signaling pathway will eventually explain the insulin resistance or other abnormalities in A-T patients is not known, but it does demonstrate that ATM has substrates and functions outside of DNA damage response pathways.
ATM
231
A discussion of ATM substrates and downstream pathways is not complete without mention of the ATR protein. ATR appears to phosphorylate some of ATM’s protein targets, such as p53 (Tibbetts et al., 1999) or Brcal (Canman et al., 1998; Gatei et al., 2001; Tibbetts et al., 2000), in response to UV damage and DNA replication inhibitors.Tibbets et al. (1999) showed that ATM and ATR may act sequentially following IR damage, and they proposed that although ATM responds first to the damage signal, ATR-mediated phosphorylation later serves to maintain the activation of downstream pathways. This important contention may explain why ATM-dependent pathways are not completely abolished but are only attenuated in A-T cells, which are completely devoid of ATM protein activity, and underscores the importance of redundancy in the network of DNA damage response (Shiloh, 2000). It must be noted, however, that ATR also acts independently in pathways that are completely ATM independent (Liu et al., 2000).
V. ATM FUNCTIONS: LESSONS FROM KNOCKOUT MICE A. Phenotype of Atm-Deficient Mice Animal models of human genetic diseases, usually obtained by gene targeting in mice, may provide a close simulation of the human disorder or diverge in unexpected directions due to physiological differences between mice and men. The murine Atm gene and protein are highly homologous to the corresponding human ones (Pecker et al., 1996), and the large Atm gene can be readily targeted, providing a viable phenotype in animals homozygous for a null Atm allele. Atm knockout (KO) mice manifest the typical organismal and cellular phenotype of A-T, with one striking difference: They barely show the most cardinal feature of the human disease—cerebellar degeneration and the associated neuromotor dysfunction. The main characteristics of most mice homozygous for Atm null mutations are growth retardation, immunodeficiency with reduced lymphocyte numbers and defective maturation of the T-lymphocytes, male and female infertility, chromosomal instability, a striking predisposition to thymic lymphomas, and acute sensitivity to ionizing radiation. Mouse embryo fibroblasts (MEFs) from Atm-deficient mice have a short life span in culture and exhibit a cellular phenotype similar in many respects to that of human A-T cells (Barlow et al., 1996; Borghesani et al., 2000; Elson et al., 1996; Herzog et al., 1998; Xu and Baltimore, 1996; Xu et al., 1996). In most of the Atm KO mouse lines, no gross abnormalities were observed in cerebellar architecture, and no overt ataxia was reported. Stringent behavioral studies of one
232
Yosef Shiloh and Michael B. Kastan
Atm KO line, however, did reveal mild motor deficits (Barlow et al., 1996). Of particular interest in this regard is the Atmy/y line generated by Borghesani et al. (2000). This mouse does exhibit a cerebellar phenotype, expressed as defects in motor ability and ectopically localized, abnormally differentiated Purkinje cells. Interestingly, this mouse lives longer than other Atm KO mice due to later appearance of the thymic lymphomas, which are the main cause of death of Atm-deficient mice. Since Atmy/y mice are homozygous for a null Atm allele, similar to other Atm Ko lines, these differences could possibly be attributed to the different genetic backgrounds of these animals. The first murine Atm mutant to produce the Atm protein was recently described by Spring et al. (2001). The mutant protein contains an in-frame 3-amino acid deletion, it is kinase inactive, and it is produced at a low level. These mice, called “delta-SRI,” are radiosensitive and immunodeficient but live significantly longer than the classical Atm KO animals and develop a variety of tumors including B cell lymphomas. This interesting phenotype indicates that the presence of the Atm protein, even in a catalytically inactive form, makes a difference compared to its complete absence.
B. Insights into Atm-Mediated Damage Responses The phenotypic hallmarks of mice without Atm protein reflect mainly the roles of this protein in the cellular response to DSBs. Importantly, heterozygous (Atm+/−) mice show premature aging and shortened life span after treatment with sublethal doses of IR (Barlow et al., 1999a). These results corroborate the early observations of intermediate radiation sensitivity of cell lines from human A-T heterozygotes (Shiloh, 1995). The possibility of creating compound genotypes by combining different genetic alterations in the same animal provides a valuable tool for the study of functional interactions between proteins and pathways. Studies of animals deficient in both Atm and p53, or Atm and the tumor suppressor protein p21Cip1/Waf1, confirmed the upstream position of Atm in the damageinduced G1/S checkpoint mediated by p53 and p21 (Barlow et al., 1996; Westphal et al., 1997a; Xu et al., 1998; Wang et al., 1997). Importantly, p53 deficiency rescued Atm-deficient MEFs from the severe cell cycle arrest that accounted for their short life span in culture, indicating that this arrest was indeed p53 mediated (Westphal et al., 1997a; Xu et al., 1998). On the other hand, Atm and p53 deficiencies significantly synergized in tumor formation (Westphal et al., 1997b; Xu et al., 1998). Another important pathway in which the cooperation between Atm and p53 is evident is the apoptotic response of thymocytes to IR. Although Atm-deficient thymocytes are partially resistant to radiation-induced apoptosis, this resistance is enhanced in Atm
ATM
233
and p53 null thymocytes (Westphal et al., 1997a). Similar experiments combining Atm and p21 deficiencies indicated the position of p21 downstream of Atm in growth control and damage response pathways. Loss of p21 on the background of Atm deficiency led to a delay in the appearance of thymic lymphomas but enhanced cellular sensitivity to acute radiation (Wang et al., 1997). Although double deficiencies of Atm and p53 reconfirmed their interaction in damage response pathways affecting cellular survival and apoptosis, Liao et al. (1999) showed that the tumor suppressor activity of p53 in oncogenic pathways is Atm independent. Thus, in a brain tumor model obtained by ectopic expression of a truncated SV40 large T-antigen, suppression of tumor formation by p53 was not affected by removing Atm. These results underscore the value of compound genotypes of animal models in the dissection of signaling pathways and the identification of their points of convergence or divergence, which may be tissue and process specific. An interesting genotype in this regard was generated by Kamijo et al. (1999), who combined Atm and p19ARF deficiencies. p19ARF, an alternative reading frame product of the INK4a/ARF locus, is induced by oncoproteins and functions in a checkpoint pathway that responds to hyperproliferative signals (Sherr, 2000). In this pathway, p19ARF binds and inhibits p53’s inhibitor Mdm2, thus promoting p53-mediated cell cycle arrest, but it is thought not to be involved in the damage-induced activation of p53. Loss of p19ARF relieved the cell cycle arrest observed in Atm-deficient MEFs, but it did not alter the radiation sensitivity and the propensity to develop thymic lymphomas compared to those of Atm KO animals. Thus, it appears that Atm and p19ARF signal to p53 through distinct pathways but also interact in p53-dependent regulation of cellular proliferation.
C. Maturation of Immune System Genes, Immunodeficiency, and Lymphoid Tumors Three hallmarks of A-T are prominently expressed in the immune system: immunodeficiency, chromosomal instability, and tumor formation. In view of the known roles of ATM in DSB responses, these phenomena should converge at the maturation of the immune system genes, a process that involves DNA strand breakage and repair and that is unique to the immune system. The high propensity of A-T patients and Atm-deficient mice to lymphoid malignancies is intimately associated with the appearance of clonal chromosomal translocations that involve the loci of the immune system genes (Taylor et al., 1996; Barlow et al., 1996; Elson et al., 1996; Xu et al., 1996). Conceivably, defective maturation of these genes via V(D)J recombination may lead to chromosomal aberrations, which in turn alter the
234
Yosef Shiloh and Michael B. Kastan
expression of growth-promoting protooncogenes. In order to test this notion, Liao and Van Dyke (1999) inactivated the Rag-1 gene in Atm KO mice. Since the recombination-activating genes (RAGs) Rag-1 and Rag-2 are essential for the first step of V(D)J recombination, the absence of either one of these proteins will block the process (Mombaerts et al., 1992b; Shinkai et al., 1992). Indeed, introducing the Rag-1−/− genotype into Atmdeficient animals considerably reduced the appearance of thymomas compared to that in Atm KO animals. This reduction persisted when thymocyte differentiation, which had been blocked by the arrest of V(D)J recombination, was reconstituted by the introduction of a T cell receptor transgene. In contrast to these results, Petiniot et al. (2000) were not able to completely prevent the development of thymomas in Atm−/− mice in which the Rag-2 gene was inactivated, although the tumors occurred at lower frequencies and after longer latency periods compared to Atm KO animals. Importantly, the numerous chromosomal aberrations that characterized these tumors did not involve the T cell receptor locus. These authors concluded that chromosomal translocations stemming from erroneous processing of DNA DSBs are indeed critical in tumor formation in the absence of Atm; however, although V(D)J recombination may contribute to the formation of part of the breaks, it is not the sole cause of them. In support of this conclusion, Liyanage et al. (2000) showed that although certain chromosomal translocations in tumors from Atm-deficient mice do indeed involve the T cell receptor α/δ locus, others involve many other random loci. Together, the aforementioned studies indicate that the striking lymphomagenesis in Atm-deficient mice is closely associated with chromosomal instability. The translocations typical of Atm-deficient lymphoid tissue probably represent a defective response to DSBs induced as a result of V(D)J recombination and other events. The basic mechanism of this instability is unclear. It may represent the well-documented defective signaling induced by DSBs in Atm-deficient cells as well as aberrant processing of DNA discontinuities, or the lack of an ATM-dependent mechanism that eliminates cells with illegitimate rejoining of DSBs from the immune system. To what extent is the T cell deficiency of Atm KO mice a result of the defect in the maturation of the T cell receptor (TCR) genes? It is known that the two major facets of this immunodeficiency—thymic hypoplasia and defective T cell maturation—may result from defective rearrangement of the TCR-α and -β chains (Mombaerts et al., 1992a; Shinkai et al., 1993; Kisielow and von Boehmer, 1995). Chao et al. (2000) examined the role of TCR gene rearrangement in these processes and found that a functional TCR-α/β transgene rescued thymic hypoplasia and partially corrected the cell maturation defect in Atm KO mice. This observation suggests that a mechanistic defect in V(D)J recombination is at the root of the immunodeficiency in A-T.
ATM
235
T cell maturation is a biologically complex process that involves alterations in the response of major signaling pathways, including the cellular response to DSBs. Immature T cells, in which the rearrangement of the TCR genes is accompanied by DNA breakage and reunion, should be exceptionally tolerant to this damage. Bhandoola et al. (2000) found that murine immature (CD4+/CD8+, double-positive) thymocytes are more resistant than mature (single-positive) cells to apoptotic death induced by DNA intercalating agents. Apoptosis in cells treated with such agents (e.g., actinomycin D) follows the appearance of DSBs in the DNA, probably resulting from attempts to unwind the DNA cross-links. Importantly, this apoptotic process is Atm dependent but p53 independent. These results could conceivably reflect an ability of immature thymocytes to sustain the DNA breakage that accompanies the process of TCR gene rearrangements. Thus, Atm may be involved both in the mechanistic aspects of TCR gene maturation and in signaling pathways that defend the cells against harmful consequences of DSBs formed in the course of this process.
D. Role of ATM in Meiosis and Telomere Dynamics Severe gonadal atrophy is an important characteristic of A-T. Studies in Atm-deficient mice allowed a close look at the progress of meiosis in the absence of Atm. Dramatic meiotic arrest was noticed early in prophase I (Barlow et al., 1996; Xu and Baltimore, 1996), as early as the leptotene stage (Barlow et al., 1998). This arrest is followed by massive chromosomal fragmentation. Although this striking phenomenon attributes an essential role for ATM in maintaining chromosomal integrity at early meiosis I, probably at the stage of recombination, the mechanism behind ATM’s involvement in this process remains obscure. Early studies suggested that ATM binds directly to synapsed chromosomal axes, together with replication protein A and the Mlh1 and Rad51 protein, at sites defined as “meiotic nodules” (Keegan et al., 1996; Plug et al., 1997, 1998). This localization was challenged by Barlow et al. (1998), who localized Atm to the cytoplasm of the ova in the developing ovarian follicles and to the nucleus of spermatogonia but not to the synaptonemal complex (SC). Barlow et al. noticed, however, mislocalization of Atr, Dmc1, and Rad51 to the chromatin in Atm KO mice, with reduced localization of these three proteins to the SC. Flaggs et al. (1999) observed reduced localization of the Chk1 protein to meiotic chromosomes in Atm KO mice. Clearly, Atm’s absence causes reduced binding to meiotic chromosomes of several proteins that are probably essential for the progression of meiossis I in its very early stage. Further insight into the defect in the meiotic chromosomal dynamics caused by Atm deficiency was obtained by Pandita et al. (1999). They noticed that
236
Yosef Shiloh and Michael B. Kastan
the typical telomere clustering at the nuclear envelope during meiosis was defective in Atm KO mice, as was the interaction between the telomeres and the nuclear matrix. These defects can be partially rescued by eliminating the p53 protein on the background of Atm deficiency (Scherthan et al., 2000). Furthermore, Atm deficiency was found by Scherthan et al. (2000) to cause immature nuclear architecture and heterochromatin distribution in Sertoli cells, the supportive somatic cells in the seminefrous epithelium. These observations indicate a basic defect caused by the absence of ATM in the mechanism and dynamics of chromatin and chromosomal reorganization at meiosis.
E. Atm, Neurogenesis, and Neurodegeneration One of the major obstacles to our understanding of A-T is deciphering ATM’s functions in the nervous system, particularly its apparently critical role in the cerebellum. It is not immediately obvious how a role for ATM in DNA damage responses could account for the progressive neurologic degeneration seen in A-T patients; perhaps the neurodegeneration is due to some other function of ATM. The strong cytoplasmic presence of ATM in human (Oka and Takashima, 1998) and murine (Barlow et al., 1999b) Purkinje cells may indicate such a special role for ATM in these cells. However, many different syndromes that result from defective DNA damage responses also exhibit neurologic abnormalities. In addition to A-T, some sort of neuropathology is exhibited by patients with xeroderma pigmentosum, Fanconi anemia, Werner syndrome, Bloom syndrome, and Rothmund–Thomson syndrome, all due to defects in DNA damage responses (Rolig and McKinnon, 2000). Although the effect of Atm deficiency on the central nervous system of the mouse is clearly less dramatic than in humans, careful analysis yielded several interesting observations on Atm involvement in central nervous system development. Herzog et al. (1998) and Chong et al. (2000) reported that Atm and the proapoptotic protein Bax are required for radiation-induced apoptosis in the developing central nervous system. Another damage-induced apoptotic pathway was dependent on p53, with Atm- and Bax-dependent branches. Interestingly, the levels of Atm in the central nervous system are highest during embryonic development (Soares et al., 1998). These observations led to the suggestion that lack of ATM results in problems with handling the generation of DNA strand breaks during neuronal development, which in turn results in the continued survival of damaged, abnormal neurons that are destined to die later in life. The demonstration that mice missing the DNA ligase IV protein, which is critical for optimal repair of BSBs via NHEJ, die at birth because of abnormalities of neuronal development supports the
ATM
237
concept that DNA strand breaks normally occur during neuronal development (Gao et al., 1998). Mice deficient in either of the Ku proteins (but not mice lacking the catalytic subunit of DNA-PK) show increased apoptotic death of neurons during development (Gao et al., 1998; Gu et al., 2000). Importantly, when Atm KO mice were crossed with the DNA ligase IV KO mice, a partial rescue of this neuronal phenotype was observed (Y. Lee et al., 2000). This result further supports the notion that DNA strand breaks occur during neuronal development, and that ATM function is important for signaling the appropriate death of damaged neurons during development. Important support for this notion came from a rare human disorder that is very similar to A-T—the ataxia-telangiectasia-like disorder (ATLD). Two families with this disease were described by Stewart et al. (1999)—one with clinical features of an “A-T variant” and the other with apparently classical A-T. Although these patients exhibited normal levels of ATM protein and no ATM mutations, they showed reduced levels or absence of the Mre11 proteins, a component of the DSB repair complex Rad50/Mre11/Nbs1. Accordingly, mutations were identified in the corresponding gene, hMRE11, truncating the protein in severe cases and leaving a residual amount of mutant protein in milder cases. This important observation underscores the role of the DSB repair defect in the development of the A-T phenotype, including its neurological component. Interestingly, the different phenotypes associated with defective Nbs1 (the disorder NBS) and defective Mre11 (ATLD) point to different roles of the two proteins in neuronal development and suggest that Mre11 may have a function upstream of ATM in the DSB response pathway. Why cerebellar Purkinje cells would be particularly sensitive to this process is not clear. Potential explanations are that more strand breaks normally occur during development in Purkinje cells than in other types of neurons, Purkinje cells are more sensitive than other neurons to this process, or this process occurs in many types of neurons but is more clinically evident in Purkinje cells because of their relatively small numbers. The cell type specificity of the effect of Atm deficiency on radiation responses was demonstrated by Gosink et al. (1999), who showed that although murine Atm−/− fibroblasts are radiation sensitive, their astrocytes show normal radiosensitivity but still exhibit premature senescence and impaired damage-induced p53 stabilization. Although the cerebellum and its Purkinje cells have naturally attracted most of the attention of investigators in this field, Eilam et al. (1998) observed in Atm-deficient mice severe degeneration of dopaminergic neurons in various parts of the brain. Interestingly, locomotor abnormalities exhibited by these mice could be corrected by peripheral application of the dopamine precursor L-dopa, and the mice were hypersensitive to the dopamine-releasing drug D-amphetamine. Although these effects may be secondary to other
238
Yosef Shiloh and Michael B. Kastan
defects in cellular metabolism, they indicate possible causes of certain elements of the neurological phenotype of A-T. Further support to the notion that ATM’s presence may be important in certain developmental stages of the nervous system was recently obtained from a study by Allen et al. (2001). This work showed that ATM was abundant in dividing neural progenitor cells but was downregulated when these cells differentiated. Importantly, in Atm KO mice, neural progenitor cells of the dentate gyrus showed abnormally high rates of proliferation and genomic stability. Furthermore, in vitro such Atm−/− neuronal cells showed abnormal response to environmental stimuli that promote neural progenitor cell proliferation, survival, and differentiation. These findings establish an important role for ATM in maintaining proper survival and differentiation along the neuronal lineage, perhaps not only by guarding genome stability but also by being involved in cellular responses to growth cues which are essential for normal development. This model may have serious implications for A-T patients. It suggests that certain neuronal cells in these patients may be destined to abnormal differentiation or death because of damage sustained during development and possibly abnormal response to developmental cue. If this is the case, postnatal interventions to block the neurodegeneration may be futile.
VI. ATM DEFICIENCY LEADS TO INCREASED OXIDATIVE STRESS Cellular metabolism, particularly mitochondrial respiration, inflammation, and IR, leads in cells to the production of reactive oxygen species (ROS): the superoxide anion radical ( ·O2− ), the hydroxyl radical ( ·OH), and hydrogen peroxide (H2O2). These highly reactive radicals react with cellular lipids, proteins, and nucleic acids and induce in them a variety of chemical alterations (Kehrer, 1993; Gutteridge, 1994; Yu, 1994). Secondary intermediates of ROS metabolism may also be harmful to cellular macromolecules. A prominent example is the peroxinitrite anion (ONOO−), which results from the reaction between the neurotransmitter nitric oxide (NO) and the superoxide anion and is a potent damaging agent for lipids and proteins. The brain is especially exposed to the production of ROS due to its high oxygen consumption. Cells possess an extensive battery of defense mechanisms that scavenge ROS or convert them to other species, including low-molecularweight compounds such as glutathione and antioxidant enzymes such as catalase, the superoxide dismutases, and the glutathione-metabolizing enzymes (Yu, 1994). Insufficient action of these mechanisms leads to a state of oxidative stress. Thus, the extent of oxidative stress is determined by the balance between the formation of ROS and the activity of detoxifying systems.
ATM
239
Rotman and Shiloh (1997) hypothesized that part of the clinical phenotype in A-T—neuronal degeneration, premature aging, and possibly telangiectases—might be associated with increased oxidative stress in this disease. The contention that ATM deficiency might result in increased oxidative stress was based on previous evidence of oxidative stress in aging, neurodegenerative disorders, and several chromosomal instability syndromes (including several observations in A-T patients). Support for this idea was provided by several recent studies in Atmdeficient mice and human A-T cell lines. Barlow et al. (1999b) demonstrated increased NO-mediated damage to proteins in brains of Atm-deficient mice. This finding suggested that Atm-deficient brains contained elevated levels of superoxide anion, which reacts with NO to form the powerful oxidant peroxynitrite. This study also showed higher levels of lipid peroxidation in Atm-deficient testes and elevated activity of heme oxygenase (HO) in Atm-deficient Purkinje cells. Such levels of HO probably represent a response to oxidative stress and may result in increased availability of prooxidant iron released from heme, leading to tissue damage through iron-catalyzed oxidation. Quick et al. (2001) observed significantly higher ROS levels in the cerebellum (specifically in Purkinje cells) and basal ganglia, but not in the cortex, of Atm-deficient mice compared to control animals. Kamsler et al. (2001) found a reduced amount of ROS scavengers in Atm-deficient cerebella, such as thiol compounds, but there was also a higher level of thioredoxin, a general oxireductase that catalyzes NADPH-dependent reduction of exposed S–S bridges in a variety of proteins. High thioredoxin is a typical indication of oxidative stress. Takao et al. (2000) found that Atmdeficient chicken DT40 cells were highly susceptible to apoptosis-induced IR or bleomycin as well as to C2-ceramide and hydrogen peroxide. This effect was blocked by antioxidants. ROS production was significantly increased in Atm(−/−) DT40 cells compared to wild-type cells. In cultured cells, a fraction of cytoplasmic ATM was detected outside the nucleus in peroxisomes, in which it was colocalized with catalase (Watters et al., 1999). Indeed, significantly decreased catalase activity and increased lipid peroxidation were observed by Watters et al. (1999) in several A-T cell lines, indicating continuous oxidative stress in these cells. Similarly, red blood cells from A-T patients showed a three-fold increase in lipid peroxidation and decrease in the GSH content (Rybczynska et al., 1996). Shackelford et al. (2001) showed that the response of human cells to the oxidative stress-inducing agent t-butyl hydroperoxide is ATM dependent. Accordingly, A-T fibroblasts were hypersensitive to this agent and exhibited defective activation of cell cycle checkpoints in response to this treatment. This agent also activated ATM kinase activity in a manner similar to IR. The tenous oxidative stress in A-T cells may be associated with another phenomenon that is often observed in these cells—a constitutive, low-level
240
Yosef Shiloh and Michael B. Kastan
stress response. The constitutively activated stress responses are normally activated by DSBs in an ATM-dependent manner. This apparently paradoxical phenomenon was noted by Gatei et al. (2001), who showed that the basal levels of p53, p53 serine 15 phosphorylation, p21, and activated Cdc2 were constitutively elevated (two- or threefold) in the majority of A-T lymphoblasts. This constitutive activation of ATM-dependent damage responses, which is observed in ATM-deficient cells, may be driven by redundant kinases such as ATR in response to continuous low-level damage. Importantly, treating the cells with the versatile antixodiant α-lipoic acid significantly reduced the constitutive activation of these stress responses (Gatei et al., 2001). Higher levels of p53 and p21 were also found in the testes of Atm-deficient mice (Barlow et al., 2000), and constitutively high levels of p21 were reported in Atm−/− mouse embryo fibroblasts (Xu et al., 1996; Westphal et al., 1997b). In further attempts to identify the components of the A-T phenotype which might be associated with increased oxidative stress, Peter et al. (2001) created a line of mice (“SAT mice”) which are deficient for Atm and overexpress human Cu/Zn superoxide dismutase (SOD-1). Elevated levels of SOD-1 (which lead to excessive production of hydrogen peroxide) were found to exacerbate specific features of the murine Atm-deficient phenotype, including radiosensitivity, somatic growth retardation, and abnormalities in hematopoiesis, but did not affect the rate of tumor formation. These results suggest that specific features of the Atm-deficient phenotype might indeed be associated with increased oxidative stress. Although evidence is mounting for oxidative stress in ATM-deficient cells, the mechanism underlying ATM’s involvement in this phenomenon remains obscure. It is not clear whether ATM is directly involved in sensing the increase in ROS or whether oxidative stress in A-T cells is associated with unrepaired DSBs continuously present in the DNA. It is also possible that ATM regulates the expression of genes whose products are involved in various aspects of oxidative stress responses. ATM thus seems to affect the regulation of oxidative stress, possibly by direct interaction with proteins involved in ROS detoxification or by regulating the production of such proteins.
VII. INTERPLAY WITH SIGNALING PATHWAYS ASSOCIATED WITH GROWTH AND DIFFERENTIATION Certain characteristics of the cellular phenotype of A-T (see Section II.B) clearly indicate the effect of ATM absence on signaling pathways that are not associated with DNA damage but respond to various external cues. The
ATM
241
functional link recently established between ATM and the insulin response pathway (Yang and Kastan, 2000) is the first demonstration of a mechanism bridging ATM and such a metabolic pathway. Recent studies indicate an interesting interplay between ATM expression and its activation following damage, on the one hand, and the function of certain mitogenic signaling pathways on the other hand. Gueven et al. (2001) observed transcriptional downregulation of the ATM gene by epidermal growth factor, which has long been known to alter cellular sensitivity to IR. Interestingly, reduction in the amount of ATM diminished the activity of the wide specificity transcription factor Sp1 and, in turn, modulation of Sp1 activity combined with IR treatment influenced ATM levels. Two studies noted interesting relationships between ATM and the insulin-like growth factor-1 receptor (IGF-1R), a transmembrane tyrosine kinase receptor involved in regulating cell differentiation, growth, transformation, and apoptosis. Peretz et al. (2001) noted that transcriptional expression of IGF-1R was ATM dependent and hence low in A-T cells. On the other hand, overexpression of IGF-1R in A-T cells increased their radioresistance, leading to the conclusion that reduced expression of IGF-1R in A-T cells contributed to their radiosensitivity. Since IGF-1R plays a role in somatic growth and neuronal development, Peretz et al. (2001) suggested that abnormal IGF1-R expression might contribute to certain characteristics of the clinical phenotype in A-T. Further evidence of the interrelationships between ATM and IGF-1R was obtained by Macaulay et al. (2001). In this study, downregulation of IGF-1R expression using antisense strategy led to a reduction in ATM protein levels and inhibited ATM activation by IR. The mechanisms of these intriguing relationships between ATM and growth factors or their receptors await elucidation. However, these initial observations indicate several levels of complexity in the fine-tuned network of cellular responses to DNA damage as well as the involvement of ATM in various signaling pathways involved in cellular growth and differentiation.
VIII. DEFECTS IN DNA DAMAGE RESPONSE AND CANCER PREDISPOSITION There is little doubt about the importance of DNA damage in the genesis of human malignancies. Epidemiology studies clearly demonstrate a link between exposure to agents that damage DNA and the development of cancer (Doll and Peto, 1981). In addition, many known inherited cancer susceptibility syndromes arise because of germline mutations in genes involved in DNA damage responses (Table I). For example, Fanconi anemia patients,
242
Yosef Shiloh and Michael B. Kastan
Table I Cancer Susceptibility and DNA Damage/Repaira
Syndrome
DNA damage
Ataxia-telangiectasia Nijmegen breakage Xeroderma pigmentosum Fanconi’s anemia Li–Fraumeni
IR responses IR responses Excision repair Cross-link repair mp53—multiple stress responses IR responses Mismatch repair
Brca1/Brca2 HNPCC
Mutated alleles (no.)
Cancer predisposition
2 2 2 2 1
Lymphoid Lymphoid UV-induced skin cancer Myeloid leukemia Lymphoma, sarcoma
1 1
Breast, ovarian Colon, endometrial, other
a Selected cancer susceptibility syndromes representing situations in which germline mutations occur in genes involved in DNA damage responses. The diseases are inherited either in an autosomal-recessive (two mutated alleles) or -dominant (one mutated allele) manner. CA, cancer; mp53, mutant p53; HNPCC, hereditary nonpolyposis colon cancer; IR, ionizing radiation; UV, ultraviolet radiation.
defective in repair of DNA cross-links, have a very high incidence of myeloid leukemias (D’Andrea and Grompe, 1997). Also, xeroderma pigmentosum patients, defective in DNA excision repair, have a very high incidence of skin cancer (Cleaver et al., 1999). Both of these diseases, like A-T and NBS, are autosomal-recessive disorders and require inheritance of two mutant alleles of the particular gene to exhibit the phenotypic abnormalities and the marked cancer predisposition. In contrast, inheritance of a single mutated copy of p53, Brca1, Brca2, or mismatch repair genes is sufficient to predispose patients to the development of particular malignancies (Table I). Since A-T and NBS patients almost exclusively develop lymphoid malignancies, it is reasonable to speculate that this reflects important roles for ATM and Nbs 1 in cellular responses to the introduction of DNA DSBs during lymphoid development. It is curious, however, that inherited mutations in Brca1 and Brca2, which participate in the same cellular responses to DNA strand breaks as ATM and Nbs1, predispose women to develop breast and ovarian cancers but do not increase the risk of lymphoid malignancies. Perhaps there is a functionally significant difference between loss of a single copy of one of these genes, as is the case with Brca1 and Brca2 mutations, and homozygous loss of both copies, as occurs with A-T and NBS. The markedly increased incidence of breast cancer in patients with germline mutations in a single copy of the p53 gene (Li–Fraumeni syndrome) could reflect this same situation. The possible increased incidence of breast cancer among A-T heterozygotes (Khanna, 2000; Gatti et al., 1999; Janin et al., 1999; Olsen et al., 2001; Su and Swift, 2000) could also result from functional differences between loss of one versus loss of both copies of genes in these damage response pathways. Clarifying why
ATM
243
breast epithelial cells would be particularly sensitive to transformation in this setting is of significant interest. The marked sensitivity of breast epithelium to radiation-induced cancers may reflect a particularly important role of cellular responses to DNA strand breaks in the genesis of cancers in this tissue.
IX. CONCLUSIONS In the intricate circuitry of cellular signaling, several pathways sometimes converge at the same multibranched junction. Such junctions are often controlled by extremely versatile, multifaceted proteins that regulate the traffic at these busy crossroads. ATM has all the requisite features of such proteins. What is known about its functions entitles it to the position of such a superpower. ATM clearly has a pivotal role in guarding the cells from the consequences of DNA DSBs. However, evidence is mounting that ATM is engaged in other cellular processes as well—ones involving signaling by external stimuli. It is these other “jobs” that remain to be elucidated before we obtain a full picture of ATM’s enormous capacities and roles. A potential drawback of such master controllers is that the genetic defects that inactivate or eliminate them are most likely extremely harmful. The emerging relationship between ATM and ATR demonstrates a functional redundancy that probably relieves this problem to some extent and explains why A-T is a slowly progressing disease rather than embryonic lethal. A-T, the disease caused by ATM deficiency, is one of many genetic disorders that indicate the intimate connection between genome stability, developmental processes, and cancer formation. The recently gained understanding of ATM’s mode of operation allows us to obtain an in-depth look at all three.
ACKNOWLEDGMENTS We thank Rani Elkon for preparing the phylogenetic tree of the PIK-related protein kinase family. We are grateful to Ari Barzilai, Galit Rotman, Yael Ziv, Dganit Shkedy, Sharon RashiElkeles, and Yuval Landau for helpful comments on the manuscript. Work in the laboratory of Y.S. was supported by the A-T Medical Research Foundation, the A-T Children’s Project, the National Institutes of Health (Grant RO1 NS 31763), and the Thomas Appeal. M.B.K. was supported in part by Grants CA71387, ES05777, and CA21765 from the National Institutes of Health and by the American Lebanese Syrian Associated Charities of the St. Jude Children’s Research Hospital.
244
Yosef Shiloh and Michael B. Kastan
REFERENCES Ahn, J.-Y., Schwarz, J. K., Piwnica-Worms, H., and Canman, C. E. (2000). Cancer Res. 60, 5934–5936. Allen, D. M., van Praag, H., Ray, J., Weaver, Z., Winrow, C. J., Carter, T. A., Braquet, R., Harrington, E., Ried, T., Brown, K. D., Gage, F. H., and Barlow, C. (2001). Genes Dev. 15, 554–566. Andrade, M. A., and Bork, P. (1995). Nature Genet. 11, 115–116. Aoki, H., Sato, S., Takanami, T., Ishihara, T., Katsuraa, I., Takahashi, H., and Higashitani, A. (2000). Mol. Gen. Genet. 264, 119–126. Araki, R., Fukumura, R., Fujimori, A., Taya, Y., Shiloh, Y., Kurimasa, A., Burma, S., Li, G. C., Chen, D. J., Sato, K., Hoki, Y., Tatsumi, K., and Abe, M. (1999). Cancer Res. 59, 3543– 3546. Ashcroft, M., Kubbutat, M. H. G., and Vousden, K. H. (1999). Mol. Cell. Biol. 19, 1751–1758. Banin, S., Moyal, L., Shieh, S.-Y., Taya, Y., Anderson, C. W., Chessa, L., Smorodinsky, N. I., Prives, C., Reiss, Y., Shiloh, Y., and Ziv, Y. (1998). Science 281, 1674–1677. Barlow, C., Hirotsune, S., Paylor, R., Liyanage, M., Eckhaus, M., Collins, F., Shiloh, Y., Crawley, J. N., Ried, T., Tagle, D., and Wynshaw-Boris, A. (1996). Cell 86, 159–171. Barlow, C., Liyanage, M., Moens, P. B., Deng, C.-X., Ried, T., and Wynshaw-Boris, A. (1997a). Nature Genet. 17, 462–466. Barlow, C., Brown, K. D., Deng, C.-X., Tagle, D. A., and Wynshaw-Boris, A. (1997b). Nature Genet. 17, 453–456. Barlow, C., Liyanage, M., Moens, P. B., Tarsounas, M., Nagashima, K., Brown, K., Rottinghaus, S., Jackson, S. P., Tagle, D., Ried, T., and Wynshaw-Boris, A. (1998). Development 125, 4007–4017. Barlow, C., Ribaut-Barassin, C., Zwingman, T. A., Pope, A. J., Brown, K. D., Owens, J. W., Larson, D., Harrington, E. A., Haeberle, A.-M., Mariani, J., Eckhaus, M., Herrup, K., Bailly, Y., and Wynshaw-Boris, A. (2000). Proc. Natl. Acad. Sci. USA 97, 871–876. ¨ Barlow, C., Eckhaus, M. A., Schaffer, A. A., and Wynshaw-Boris, A. (1999a). Nature Genet. 21, 359–360. Barlow, C., Dennery, P. A., Shigenaga, M. K., Smith, M. A., Morrow, J. D., Roberts, L. J., Wynshaw-Boris, A., and Levine, R. L. (1999b). Proc. Natl. Acad. Sci. USA 96, 9915–9919. Baskaran, R., Wood, L. D., Whitaker, L. L., Canman, C. E., Morgan, S. E., Xu, Y., Barlow, C., Baltimore, D., Wynshaw-Boris, A., Kastan, M. B., and Wang, J. Y. (1997). Nature 387, 516–519. Becker-Catania, S. G., Chen, G., Hwang, M. J., Wang, Z., Sun, X., Sanal, O., BernatowskaMatuszkiewicz, E., Chessa, L., Lee, E. Y.-H. P., and Gatti, R. A. (2000). Mol. Genet. Metab. 70, 122–133. Bhandoola, A., Dolnick, B., Fayad, N., Nussenzweig, A., and Singer, A. (2000). J. Exp. Med. 192, 891–897. Bishop, A. J., Barlow, C., Wynshaw-Boris, A., and Schiestl, R. H. (2000). Cancer Res. 60, 395–399. Borghesani, P. R., Alt, F. W., Bottaro, A., Davidson, L., Aksoy, S., Rathbun, G. A., Roberts, T. M., Swat, W., Segal, R. A., and Gu, Y. (2000). Proc. Natl. Acad. Sci. USA 97, 3336–3341. Bosotti, R., Isacchi, A., and Sonnhammer, E. L. L. (2000). TIBS 25, 225–227. Boulton, S. J., and Jackson, S. P. (1998). EMBO J. 17, 1819–1828. Broeks, A., de Klein, A. A., Floore, A. N., Muijtjens, M., Kleijer, W. J., Jaspers, N. G., and van’t Veer, L. J. (1998). Hum. Mutat. 12, 330–337. Brown, K. D., Ziv, Y., Sadanandan, S. N., Chessa, L., Collins, F. S., Shiloh, Y., and Tagle, D. A. (1997). Proc. Natl. Acad. Sci. USA 94, 1840–1845.
ATM
245
Bullrich, F., Rasio, D., Kitada, S., Starostik, P., Kipps, T., Keating, M., Albitar, M., Reed, J. C., and Croce, C. M. (1999). Cancer Res. 59, 24–27. Canman, C. E., Wolff, A. C., Chen, C., Fornace, A. J., and Kastan, M. B. (1994). Cancer Res. 54, 5054–5058. Canman, C. E., Lim, D.-S., Cimprich, K. A., Taya, Y., Tamai, K., Appella, E., Kastan, M., and Siliciano, J. D. (1998). Science 281, 1677–1679. Carney, J. P., Maser, R. S., Olivares, H., Davis, E. M., Le Beau, M., Yates, J. R., III, Hays, L., Morgan, W. F., and Petrini, J. H. J. (1998). Cell 93, 477–486. Castellvi-Bel, S., Sheikhavandi, S., Telatar, M., Tai, L.-Q., Hwang, M., Wang, Z., Yang, Z., Cheng, R., and Gatti, R. A. (1999). Hum. Mutat. 14, 156–162. Chan, D. W., Son, S.-C., Block, W., Ye, R., Khanna, K. K., Wold, M. S., Douglas, P., Goodarzi, A. A., Pelley, J., Taya, Y., Lavin, M. F., and Less-Miller, S. P. (2000). J. Biol. Chem. 275, 7803–7810. Chao, C., Saito, S., Anderson, C. W., Appella, E., and Xu, Y. (2000a). Proc. Natl. Acad. Sci. USA 97, 11936–11941. Chao, C., Yang, E. M., and Xu, Y. (2000b). J. Immunol. 164, 345–349. Chehab, N. H., Malikzay, A., Stavridi, E. S., and Halazonetis, T. D. (1999). Proc. Natl. Acad. Sci. USA 96, 13777–13782. Chehab, N. H., Malikzay, A., Appel, M., and Halazonetis, T. D. (2000). Genes Dev. 14, 278–288. Chen, G., Yuan, S.-S. F., Liu, W., Xu, Y., Trujillo, K., Song, B., Cong, F., Goff, S. P., Wu, Y., Arlinghaus, R., Baltimore, D., Gasser, P. J., Park, M. S., Sung, P., and Lee, E. Y. H. P. (1999). J. Biol. Chem. 274, 12748–12752. Chen, J.-J., Silver, D., Cantor, S., Livingston, D. M., and Scully, R. (1999). Cancer Res. 59, 1752s–1756s. Chen, Y.-R., Meyer, C. F., and Tan, T.-H. (1996). J. Biol. Chem. 271, 631–634. Chiesa, N., Barlow, C., Wynshaw-Boris, A., Strata, P., and Tempia, F. (2000). Neuroscience 96, 575–583. Chong, M. J., Murray, M. R., Gosink, E. C., Russell, H. R. C., Srinivasan, A., Kapsetaki, M., Korsmeyer, S. J., and McKinnon, P. J. (2000). Proc. Natl. Acad. Sci. USA 97, 889–894. Clarke, A. R., Purdie, C. A., Harrison, D. J., Morris, R. G., Bird, C. C., Hopper, M. L., and Wyllie, A. H. (1993). Nature 362, 849–852. Clarke, D. J., Segal, M., Mondesert, G., and Reed, S. I. (1999). Curr. Biol. 9, 365–368. Cleaver, J. E., Afzal, V., Feeney, L., McDowell, M., Sadinski, W., Volpe, J. P. G., Busch, D. B., Coleman, D. M., Ziffer, D. W., Yu, Y., Nagasawa, H., and Little, J. B. (1999). Cancer Res. 59, 1102–1108. Colman, M. S., Afshari, C. A., and Arett, J. C. (2000). Mutat. Res. 462, 179–188. Concannon, P., and Gatti, R. A. (1997). Hum. Mutat. 10, 100–107. Cornforth, M. N., and Bedford, J. S. (1985). Science 227, 1589–1591. Cortez, D., Wang, Y., Qin, J., and Elledge, S. J. (1999). Science 286, 1162–1166. Costanzo, V., Robertson, K., Ying, C. Y., Kim, E., Avvedimento, E., Gottesman, M., Grieco, D., and Gautier, J. (2000). Mol. Cell 6, 649–659. Craven, R. J., and Petes, T. D. (2000). Mol. Cell. Biol. 20, 2378–2384. Crawford, T. O. (1998). Semin. Pediatr. Neurol. 5, 287–294. D’Andrea, A. D., and Grompe, M. (1997). Blood 90, 1725–1736. Dar, M. E., Winters, T. A., and Jorgensen, T. J. (1997). Mutat. Res. 384, 169–179. Dasika, G. K., Lin, S. C., Zhao, S., Sung, P., Tomkinson, A., and Lee, E. Y. (1999). Oncogene 18, 7883–7899. Deng, C.-X., and Brodie, S. G. (2000). BioEssays 22, 728–737. Dennis, P. B., Fumagalli, S., and Thomas, G. (1999). Curr. Opin. Genet. Dev. 9, 49–54. De Souza, C. P. C., Ye, X. S., and Osmani, S. A. (1999). Mol. Biol. Cell 10, 3661–3674.
246
Yosef Shiloh and Michael B. Kastan
Dhand, R., Panayotou, G., Roche, S., Fry, M. J., Gout, I., Totty, N. F., Truong, O., Vicendo, P., and Yonezawa, K. (1994). EMBO J. 13, 522–533. Doll, R., and Peto, R. (1981). J. Natl. Cancer Inst. 66, 1192–1308. Dumaz, N., and Meek, D. W. (1999). EMBO J. 18, 7002–7010. Dumaz, N., Milne, D. M., and Meek, D. W. (1999). FEBS Lett. 463, 312–316. Eilam, R., Peter, Y., Elson, A., Rotman, G., Shiloh, Y., Groner, Y., and Segal, M. (1998). Proc. Natl. Acad. Sci. USA 95, 12653–12656. Elson, A., Wang, Y., Daugherty, C. J., Morton, C. C., Zhou, F., Campos-Torres, J., and Leder, P. (1996). Proc. Natl. Acad. Sci. USA 93, 13084–13089. Famulski, K. S., and Paterson, M. C. (1999). FEBS Lett. 453, 183–186. Fan, Z., Chakravarty, P., Alfieri, A., Pandita, T. K., Vikram, B., and Guha, C. (2000). Cancer Gene Ther. 7, 1307–1314. Flaggs, G., Plug, A. W., Dunks, K. M., Mundt, K. E., Ford, J. C., Quiggle, M. R., Taylor, E. M., Westphal, C. H., Ashley, T., Hoekstra, M. F., and Carr, A. M. (1999). Curr. Biol. 7, 977–986. Fritz, E., Friedl, A. A., Zwacka, R. M., Eckardt-Schupp, F., and Meyn, M. S. (2000). Mol. Biol. Cell 11, 2605–2616. Gao, Y., Sun, Y., Frank, K. M., Dikkes, P., Fujiwara, Y., Seidl, K. J., Sekiguchi, J. M., Rathbun, G. A., Swat, W., Wang, J. et al. (1998). Cell 95, 891–902. Garcia, V., Salanoubat, M., Choisne, N., and Tissier, A. (2000). Nucleic Acids Res. 28, 1692–1699. Gatei, M., Scott, S. P., Filippovitch, I., Soronika, N., Lavin, M. F., Weber, B., and Khanna, K. K. (2000a). Cancer Res. 60, 3299–3304. Gatei, M., Young, D., Cerosaletti, K. M., Desai-Mehta, A., Spring, K., Kozlov, S., Lavin, M. F., Gatti, R. A., Concannon, P., and Khanna, K. K. (2000b). Nature Genet. 25, 115–119. Gatei, M., Zhou, B. B., Hobson, K., Scott, S., Young, D., and Khanna, K. K. (2001). J. Biol. Chem., in press. Gately, D. P., Hittle, J. C., Chan, G. K. T., and Yen, T. J. (1998). Mol. Biol. Cell 9, 2361–2374. Gatti, R. A., Tward, A., and Concannon, P. (1999). Mol. Genet. Metab. 68, 419– 423. Genesh, A., North, P., and Thacker, J. (1993). Mutat. Res. 299, 251–259. Giaccia, A. J., and Kastan, M. B. (1998). Genes Dev. 12, 2973–2983. Gilad, S., Khosravi, R., Uziel, T., Ziv, Y., Rotman, G., Savitsky, K., Smith, S., Chessa, L., Harnik, R., Shkedi, D., Frydman, M., Sanal, O., Portnoi, S., Goldwicz, Z., Jaspers, N. G. J., Gatti, R. A., Lenoir, G., Lavin, M. F., Tatsumi, K., Wegner, R. D., Shiloh, Y., and Bar-Shira, A. (1996). Hum. Mol. Genet. 5, 433– 439. Gilad, S., Chessa, L., Khosravi, R., Russell, P., Galanty, Y., Piane, M., Gatti, R. A., Jorgensen, T. J., Shiloh, Y., and Bar-Shira, A. (1998a). Am. J. Hum. Genet. 62, 551–561. Gilad, S., Khosravi, R., Harnik, R., Ziv, Y., Shkedi, D., Galanty, Y., Frydman, M., Levi, J., Sanal, O., Chessa, L., Smeets, D., Shiloh, Y., and Bar-Shira, A. (1998b). Hum. Mutat. 11, 69–75. Gosink, E. C., Chong, M. J., and McKinnon, P. J. (1999). Cancer Res. 59, 5294–5298. Graeber, T. G., Peterson, J. F., Tsai, M., Monica, K., Fornace, A. J. Jr., and Giaccia, A. J. (1994). Mol. Cell. Biol. 14, 6264–6277. Grant, P. A., Schieltz, D., Pray-Grant, M. G., Yates, J. R., III, and Workman, J. L. (1998). Mol. Cell 2, 863–867. Green, M. H. L., Marcovitch, A. J., Harcourt, S. A., Lowe, J. E., Green, I. C., and Arlett, C. F. (1997). Free Radical Biol. Med. 22, 343–347. Greenwell, P. W., Kronmal, S. L., Porter, S. E., Gassenhuber, J., Obermaier, B., and Petes, T. D. (1995). Cell 82, 823–829. Gruschcow, J. M., Holzen, T. M., Park, K. J., Weinert, T., Lichten, M., and Bishop, D. K. (1999). Genetics 153, 607–620.
ATM
247
Gu, Y., Sekiguchi, J., Gao, Y., Dikkes, P., Frank, K., Ferguson, D., Hasty, P., Vhun, J., and Alt, F. W. (2000). Proc. Natl. Acad. Sci. USA 97, 2668–2673. Gueven, N., Keating, K., Chen, P., Fukao, T., Khanna, K. K., Watters, D., Rodemann, P. H., and Lavin, M. F. (2001). J. Biol. Chem., 276, 8884–8891. Gutteridge, J. M. (1994). Chem. Biol. Interact. 91, 133–140. Haber, J. E. (2000). TIG 16, 259–264. Hari, K. L., Santerre, A., Sekelsky, J. J., McKim, K. S., Boyd, J. B., and Hawley, R. S. (1995). Cell 82, 815–821. Hariharan, P. V., Eleczko, S., Smith, B.P., and Patterson, M. C. (1981). Radiation Res. 86, 589–597. Harris, S. D., and Kraus, P. R. (1998). Genetics 148, 1055–1067. Herzog, K.-H., Chong, M. J., Kapsetaki, M., Morgan, J. I., and McKinnon, P. J. (1998). Science 280, 1089–1091. Hirao, A., Kong, Y.-Y., Matsuoka, S., Wakeham, A., Ruland, J., Yoshida, H., Liu, D., Elledge, S. J., and Mak, T. W. (2000). Science 287, 1824–1827. Hofmann, A. F., and Harris, S. D. (2000). Genetics 154, 1577–1586. Humphrey, T. (2000). Mutat. Res. 451, 211–226. Ichijo, H. (1999). Oncogene 18, 6807–6093. Iftode, C., Daniely, Y., and Borowiec, J. A. (1999). Crit. Rev. Biochem. Mol. Biol. 34, 141– 180. Ichijo, H. (1999). Oncogene 18, 6807–6093. Janin, N., Andrieu, N., Ossian, K., Laug´e, A., Croquette, M.-F., Griscelli, C., Debre, M., Bressacde-Paillerets, B., Aurias, A., and Stoppa-Lyonnet, D. (1999). Br. J. Cancer 80, 1042–1045. Jeggo, P. A., Carr, A. M., and Lehmann, A. R. (1998). TIG 14, 312–316. Jhappan, C., Morse, H. C., Fleischmann, R. D., Gottesman, M. M., and Merlino, G. (1997). Nature Genet. 17, 483–486. Kamijo, T., van de Kamp, E., Chong, M. J., Zindy, F., Diehl, J. A., Sherr, C. J., and McKinnon, P. J. (1999). Cancer Res. 59, 3464–3469. Kamsler, A., Daily, D., Hochman, A., Stern, N., Shiloh, Y., Rotman, G., and Barzilai, A. (2001). Cancer Res. 61, 1849–1854. Karin, M. (1999). Oncogene 18, 6867–6874. Karlseder, J., Broccoli, D., Dai, Y., Hardy, S., and de Lange, T. (1999). Science 283, 1321– 1325. Karran, P. (2000). Curr. Opin. Genet. Dev. 10, 144–150. Kastan, M. B., and Lim, D.-S. (2000). Nature Rev. Mol. Cell. Biol. 1, 179–186. Kastan, M. B., Onyekwere, O., Sidransky, D., Vogelstein, B., and Craig, R. W. (1991). Cancer Res. 51, 6304–6311. Kastan, M. B., Zhan, Q., El-Deiry, W. S., Carrier, F., Jacks, T., Walsh, W. V., Plunkett, B. S., Vogelstein, B., and Fornace, A. J., Jr. (1992). Cell 71, 587–597. Keegan, K. S., Holtzman, D. A., Plug, A. W., Christenson, E. R., Brainerd, E. E., Flaggs, G., Bentley, N. J., Taylor, E. M., Meyn, M. S., Moss, S. B., Carr, A. M., Ashley, T., and Hoekstra, M. F. (1996). Genes Dev. 10, 2423–2437. Kehrer, J. P. (1993). Crit. Rev. Toxicol. 23, 21–48. Keith, C. T., and Schreiber, S. L. (1995). Science 270, 50–51. Khanna, K. K. (2000). J. Natl. Cancer Inst. 92, 795–802. Khanna, K. K., and Lavin, M. F. (1993). Oncogene 8, 3307–3312. Khanna, K. K., Yan, J., Watters, D., Hobson, K., Beamish, H., Spring, K., Shiloh, Y., Gatti, R. A., and Lavin, M. F. (1997). J. Biol. Chem. 272, 9489–9495. Khanna, K. K., Keating, K. E., Kozlov, S., Scott, S., Gatei, M., Hobson, K., Taya, Y., Gabrielli, B., Chan, D., Lees-Miller, S. P., and Lavin, M. F. (1998). Nature Genet. 20, 398–400.
248
Yosef Shiloh and Michael B. Kastan
Kharbanda, S., Pandey, P., Yamauchi, T., Kumar, S., Kaneki, M., Kumar, V., Bharti, A., Yuan, Z.-M., Ghanem, L., Rana, A., Weichselbaum, R., Johnson, G., and Kufe, D. (2000). Mol. Cell. Biol. 20, 4979–4989. Khosravi, R., Maya, R., Gottlieb, T., Oren, M., Shiloh, Y., and Shkedy, D. (1999). Proc. Natl. Acad. Sci. USA 96, 14973–14977. Kim, S.-T., Lim, D.-S., Canman, C. E., and Kastan, M. B. (1999). J. Biol. Chem. 274, 37538–37543. Kisielow, P., and von Boehmer, H. (1995). Adv. Immunol. 58, 87–209. Kondo, N., Inoue, R., Nishimura, S., Kasahara, K., Kameyama, T., Miwa, Y., Lorenzo, P. R., and Orii, T. (1993). Scand. J. Immunol. 38, 45–48. Kuruvilla, F. G., and Schreiber, S. L. (1999). Chem. Biol. 6, R129–R136. Laake, K., Jansen, L., Hahnemann, J. M., Brondum-Nielsen, K., Lonnqvist, T., Kaariainen, H., Saankila, R., Lahdesmaki, A., Hammarstrom, L., Yuen, J., Tretli, S., Heiberg, A., Olsen, J. H., Tucker, M., Kleinerman, R., and Borresen-Dale, A. L. (2000). Hum. Mutat. 16, 232–246. Lakin, N. D., Weber, P., Stankovic, T., Rottinghaus, S. T., Taylor, A. M., and Jackson, S. P. (1997). Oncogene 13, 2707–2716. Lambert, P. F., Kashanchi, F., Radonovich, M. F., Shiekhatter, R., and Brady, J. N. (1998). J. Biol. Chem. 273, 33048–33053. Lavin, M. F., and Davidson, M. (1981). J. Cell Sci. 48, 383–391. Lavin, M. F., and Shiloh, Y. (1996). Curr. Opin. Immunol. 8, 459–464. Lavin, M. F., and Shiloh, Y. (1997). Annu. Rev. Immunol. 15, 177–202. Lavin, M. F., and Shiloh, Y. (1999). In “Primary Immunodeficiency Diseases, a Molecular and Genetic Approach” (H. D. Ochs, C. I. E. Smith, and J. Puck, Eds.), pp. 306–323. Oxford Univ. Press, Oxford. Lee, J.-S., Collins, K. M., Brown, A. L., Lee, C.-H., and Chung, J. J. (2000). Nature 404, 201–204. Lee, S. A., Dritschillo, A., and Jung, M. (1998). J. Biol. Chem. 273, 32889–32894. Lee, S. J., Dimtchev, A., Lavin, M. F., Dritschilo, A., and Jung, M. (1998). Oncogene 17, 1821–1826. Lee, Y., Barnes, D. E., Lindahl, T., and McKinnon, P. J. (2000). Genes Dev. 14, 2576–2580. Li, N., Banin, S., Ouyang, H., Li, G. C., Courtois, G., Shiloh, Y., Karin, M., and Rotman, G. (2000). J. Biol. Chem. 276, 8898–8903. Li, S., Ting, N. S. Y., Zheng, L., Chen, P.-L., Ziv, L., Shiloh, Y., Lee, E. Y. H. P., and Lee, W.-H. (2000). Nature 406, 210–215. Liao, M.-J., and Van Dyke, T. (1999). Genes Dev. 13, 1246–1250. Liao, M.-J., Yin, C., Barlow, C., Wynshaw-Boris, A., and van Dyke, T. (1999). Mol. Cell. Biol. 19, 3095–3102. Lim, D.-S., Kirsch, D. G., Canman, C. E., Ahn, J. H., Ziv, Y., Newman, L. S., Darnell, R. B., Shiloh, Y., and Kastan, M. B. (1998). Proc. Natl. Acad. Sci. USA 95, 10146–10151. Lim, D.-S., Kim, S.-T., Xu, B., Maser, R. S., Lin, J., Petrini, J. H. J., and Kastan, M. B. (2000). Nature 404, 613–617. Linke, S. P., Clarkin, K. C., DiLeonardo, A., Tsou, A., and Wahl, G. M. (1996). Genes Dev. 10, 934–947. Liu, Q., Guntuku, S., Cui, X.-S., Matsuoka, S., Cortez, D., Tamai, K., Luo, G., Carattini-Rivera, S., DeMayo, F., Bradley, A., Donehowever, L. A., and Elledge, S. J. (2000). Genes Dev. 14, 1448–1459. Liyanage, M., Weaver, Z., Barlow, C., Coleman, A., Pankratz, D. G., Anderson, S., WynshawBoris, A., and Ried, T. (2000). Blood 96, 1940–1946. Longhese, M. P., Paciotti, V., Neecke, H., and Lucchini, G. (2000). Genetics 155, 1577–1591. Lowe, S. W., Schmitt, S. W., Smith, S. W., Osborne, B. A., and Jacks, T. (1993). Nature 362, 847–849.
ATM
249
Lowndes, N. F., and Murgia, J. R. (2000). Curr. Opin. Genet. Dev. 10, 17–25. Luo, C. M., Tang, W., Mekeel, K. L., DeFrank, J. S., Anne, R. P., and Powell, S. N. (1996). J. Biol. Chem. 271, 4497–4503. Macaulay, V., Salisbury, A. J., Bohula, E. A., Playford, M. P., Smorodinsky, N. I., and Shiloh, Y. (2001). Oncogene, in press. Mallory, J. C., and Petes, T. D. (2000). Proc. Natl. Acad. Sci. USA 97, 13749–13754. Martin, A.-M., and Weber, B. L. (2000). J. Natl. Cancer Inst. 92, 1126–1135. Martinho, R. G., Lindsay, H. D., Flaggs, G., DeMaggio, A. J., Hoekstra, M. F., Carr, A. M., and Bentley, N. J. (1998). EMBO J. 17, 7239–7249. Maser, R. S., Monsen, K. J., Nelms, B. E., and Petrini, J. H. (1997). Mol. Cell. Biol. 17, 6087–6096. Matsuoka, S., Rotman, G., Ogawa, A., Shiloh, Y., Tamai, K., and Elledge, S. J. (2000). Proc. Natl. Acad. Sci. USA 97, 10389–10394. Matsuura, A., Naito, T., and Ishikawa, F. (1999). Genetics 152, 1501–1512. Maya, R., Balass, M., Kim, S.-T., Shkedy, D., Leal, J., Shifman, O., Moas, M., Buschmann, T., Ronai, Z., Shiloh, Y., Kastan, M. B., Katzir, E., and Oren, M. (2001). Genes Dev. 15, 1067–1077. McConville, C. M., Stankovic, T., Byrd, P. J., McGuire, G. M., Yao, Q.-Y., Lennox, G. G., and Taylor, A. M. R. (1996). Am. J. Hum. Genet. 59, 320–330. McKim, K. S., Jang, J. K., Sekelsky, J. J., Laurencon, A., and Hawley, R. S. (2000). Chromosoma 109, 44 –49 . McKinnon, P. J., and Burgoyne, L. A. (1985). Eur. J. Cell Biol. 39, 161–166. McMahon, S. B., Van Buskirk, H. A., Dugan, K. A., Copeland, T. D., and Cole, M. D. (1998). Cell 94, 363–374. McMahon, S. B., Wood, M. A., and Cole, M. D. (2000). Mol. Cell. Biol. 20, 556–562. Mercurio, F., and Manning, A. M. (1999). Oncogene 18, 6163–6171. Melchionna, R., Chen, X.-B., Blasina, A., and McGowan, C. H. (2000). Nat. Cell Biol. 2, 762–765. Metcalfe, J. A., Parkhill, J., Campbell, L., Stacey, M., Biggs, P., Byrd, P. J., and Taylor, A. M. (1996). Nature Genet. 13, 350–353. Meyn, M. S. (1993). Science 260, 1327–1330. Meyn, M. S. (1999). Clin. Genet. 55, 289–304. Mielke, K., and Herdegen, T. (2000). Prog. Neurobiol. 61, 45–60. Mills, K. D., Sinclair, D. A., and Guarente, L. (1999). Cell 97, 609–620. Mombaerts, P., Clarke, A. R., Rudnicki, M. A., Iacomini, J., Itohara, S., Lafaille, J. J., Wang, L., Ichikawa, Y., Jaenisch, R., and Hooper, M. L., et al. (1992a). Nature 360, 225–229. Mombaerts, P., Iacomini, J., Johnson, R. S., Herrup, K., Tonegawa, S., and Papaioannou, V. (1992b). Cell 68, 869–877. Morgan, S. E., Lovly, C., Pandita, T. K., Shiloh, Y., and Kastan, M. B. (1997). Mol. Cell. Biol. 17, 2020–2039. Morrison, C., Sonoda, E., Takao, N., Shinohara, A., Yamamoto, K., and Takeda, S. (2000). EMBO J. 19, 463–471. Morrow, D. M., Tagle, D. A., Shiloh, Y., Collins, F. S., and Hieter, P. (1995). Cell 82, 831–840. Neecke, H., Lucchini, G., and Longhese, M. P. (1999). EMBO J. 18, 4485–4497. Nelms, B. E., Masser, R. S., MacKay, J. F., Lagally, M. G., and Petrini, J. H. (1998). Science 280, 590–592. Nelson, W. G., and Kastan, M. B. (1994). Mol. Cell. Biol. 14, 1815–1823. Oka, A., and Takashima, S. (1998). Neurosci. Lett. 252, 195–198. Olsen, J. H., Hahnemann, J. M., Borresen-Dale, A. L., Brondum-Nielsen, K., Hammarstrom, L., Kleinerman, R., Kaariainen, H., Lonnqvist, T., Sankila, R., Seersholm, N., Tretli, S., Yuen, J., Boice, J. D., and Tucker, M. (2001). J. Natl. Cancer Inst. 93, 121–127.
250
Yosef Shiloh and Michael B. Kastan
O’Neill, T., Dwyer, A. J., Ziv, Y., Chan, D. W., Lees-Miller, S. P., Abraham, R. H., Lai, J. H., Hill, D., Shiloh, Y., Cantley, L. C., and Rathbun, G. A. (2000). J. Biol. Chem. 275, 22719–22727. Page, R. D. M. (1996). Comp. Appl. Biosci. 12, 357–358. Pandita, T. K., and Dhar, S. (2000). Radiation Res. 154, 133–139. Pandita, T. K., Hall, E. J., Hei, T. K., Piatyszek, M. A., Wright, W. E., Piao, C. Q., Pandita, R. K., Willey, J. C., Geard, C. R., Kastan, M. B., and Shay, J. W. (1996). Oncogene 13, 1423–1430. Pandita, T. K., Westphal, C. H., Anger, M., Sawant, S. G., Geard, C. R., Pandita, R. K., and Scherthan, H. (1999). Mol. Cell. Biol. 19, 5096–5105. Pandita, T. K., Lieberman, H. B., Lim, D. S., Dhar, S., Zheng, W., Taya, Y., and Kastan, M. B. (2000). Oncogene 19, 1386–1391. Pecker, I., Avraham, K., Gilbert, D. J., Savitsky, K., Rotman, G., Harnik, R., Fukao, T., Schrock, E., Hirotsume, S., Tagle, D. A., Collins, F. S., Wynshaw-Boris, A., Ried, T., Copeland, N. G., Jenkins, N. A., Shiloh, Y., and Ziv, Y. (1996). Genomics 35, 39–45. Peretz, S., Jensen, R., Baserga, R., and Glazer, P. M. (2001). Proc. Natl. Acad. Sci. USA 98, 1676–1681. Petiniot, L. K., Weaver, Z., Barlow, C., Shen, R., Eckhaus, M., Steinberg, S. M., Ried, T., Wynshaw-Boris, A., and Hodes, R. J. (2000). Proc. Natl. Acad. Sci. USA 97, 6664–6669. Petrini, J. H. (2000). Curr. Opin. Cell Biol. 12, 293–296. Piret, B., Schoonbroodt, S., and Piette, J. (1999). Oncogene 18, 2261–2271. Platzer, M., Rotman, G., Bauer, D., Uziel, T., Savitsky, K., Bar-Shira, A., Gilad, S., Shiloh, Y., and Rosenthal, A. (1997). Genome Res. 7, 592–605. Plug, A. W., Peters, A. H., Xu, Y., Keegan, K. S., Hoekstra, M. F., Baltimore, D., de Boer, P., and Ashley, T. (1997). Nature Genet. 17, 457–461. Plug, A. W., Peters, A. H., Xu, Y., Keegan, K. S., Hoekstra, M. F., de Boer, P., and Ashley, T. (1998). J. Cell. Sci. 111, 413–423. Powell, S., Whitaker, S., Peacock, J., and McMillan, T. (1993). Mutat. Res. 294, 9–20. Quick, K. L., and Dugan, L. L. (2001). Ann. Neurol. 49, 627–635. Rhodes, N., D’Souza, T., Foster, C. D., Ziv, Y., Kirsch, D. G., Shiloh, Y., Kastan, M. B., Reinhart, P. H., and Gilmer, T. M. (1998). Genes Dev. 12, 3686–3692. Rincon, M., Flavell, R. A., and Davis, R. A. (2000). Free Radical Biol. Med. 28, 1328–1337. Ritchie, K. B., and Petes, T. D. (2000). Genetics 155, 475–479. Ritchie, K. B., Mallory, J. C., and Petes, T. D. (1999). Mol. Cell. Biol. 19, 6065–6075. Rivero-Carmena, M., Porras, O., Pelaez, B., Pacheco-Castro, A., Gatti, R. A., and Regueiro, J. R. (2000). Int. Immunol. 12, 927–935. Robertson, K., Hensey, C., and Gautier, J. (1999). Oncogene 18, 7070–7079. Rogakou, E. P., Boon, C., Redon, C., and Bonner, W. M. (1999). J. Cell Biol. 146, 905–916. Rolig, R. L., and McKinnon, P. J. (2000). TINS 23, 417–424. Rotman, G., and Shiloh, Y. (1997). BioEssays 19, 911–917. Rotman, G., and Shiloh, Y. (1999). Oncogene 18, 6135–6144. Rybczynska, M., Pawlak, A. L., Sikorska, E., and Ignatowicz, R. (1996). Biochem. Biophys. Acta 1302, 231–235. Sakaguchi, K., Herrera, J. E., Saito, S., Miki, T., Bustin, M., Vassilev, A., Anderson, C. W., and Appella, E. (1998). Genes Dev. 12, 2831–2841. Saleh, A., Schieltz, D., Ting, N., McMahon, S. B., Litchfield, D. W., Yates, J. R., III, Lees-Miller, S. P., Cole, M. D., and Brandl, C. J. (1998). J. Biol. Chem. 273, 26559–26565. Sandoval, N., Platzer, M., Rosenthal, A., Dork, T., Bendix, R., Skawran, B., Stuhrmann, M., Wegner, R. D., Sperling, K., Banin, S., Shiloh, Y., Baumer, A., Bernthaler, U., Sennefelder, H., Brohm, M., Weber, B. H. F., and Schindler, D. (1999). Hum. Mol. Genet. 8, 69–79. Santocanale, C., and Diffley, J. F. (1998). Nature 395, 615–618. Sarkaria, J. N., Tibbetts, R. S., Busby, E. C., Kennedy, A. P., Hill, D. E., and Abraham, R. T. (1998). Cancer Res. 58, 4375–4382.
ATM
251
Sarkaria, J. N., Busby, E. C., Tibbetts, R. S., Roos, P., Taya, Y., Karnitz, L. M., and Abraham, R. T. (1999). Cancer Res. 59, 4375–4382. Sasaki, T., Tian, H., Kukita, Y., Inazuka, M., Tahira, T., Imai, T., Yamauchi, M., Saito, T., Hori, T., Hashimoto-Tamaoki, T., Komatsu, K., Nikaido, O., and Hayashi, K. (1998). Hum. Mutat. 12, 186–195. Savitsky, K., Bar-Shira, A., Gilad, S., Rotman, G., Ziv, Y., Vanagaite, L., Tagle, D. A., Smith, S., Uziel, T., Sfez, S., Ashkenazi, M., Pecker, I., Frydman, M., Harnik, R., Patanjali, S. R., Simmons, A., Clines, G. A., Sartiel, A., Gatti, R. A., Chessa, L., Sanal, O., Lavin, M. F., Jaspers, N. G. J., Taylor, A. M. R., Arlett, C. F., Miki, T., Weissman, S. M., Lovett, M., Collins, F. S., and Shiloh, Y. (1995a). Science 268, 1749–1753. Savitsky, K., Sfez, S., Tagle, D. A., Ziv, Y., Sartiel, A., Collins, F. S., Shiloh, Y., and Rotman, G. (1995b). Hum. Mol. Genet. 4, 2025–2032. Savitsky, K., Ziv, Y., Bar-Shira, A., Gilad, S., Tagle, D. A., Smith, S., Uziel, T., Sfez, S., Nahmias, J., Sartiel, A., Eddy, R. L., Shows, T. B., Povey, S., Collins, F. S., Shiloh, Y., and Rotman, G. (1996). Genomics 33, 199–206. Savitsky, K., Platzer, M., Uziel, T., Gilad, S., Sartiel, A., Rosenthal, A., Elroy-Stein, O., Shiloh, Y., and Rotman, G. (1997). Nucleic Acids Res. 25, 1678–1684. ¨ Schaffner, C., Stilgenbauer, S., Rappold, G. A., Dohner, H., and Lichter, P. (1999). Blood 94, 748–753. ¨ Schaffner, C., Idler, I., Stilgenbauer, S., Dohner, H., and Lichter, P. (2000). Proc. Natl. Acad. Sci. USA 97, 2773–2778. Schackelford, R. E., Innes, C. L., Sieber, S. O., Heinloth, A. N., Leadon, S. A., and Paules, R. S. (2001). J. Biol. Chem. 276, 21951–21959. Scherthan, H., Jerratsch, M., Dhar, S., Wang, Y. A., Goff, S. P., and Pandita, T. K. (2000). Mol. Cell. Biol. 20, 7773–7783. Schmidt-Ullrich, R. K., Dent, P., Grant, S., Mikkelsen, R. B., and Valerie, K. (2000). Radiat. Res. 153, 245–257. Schubert, R., Reinchenbach, J., Royer, N., Pichler, M., and Zielen, S. (2000). Clin. Exp. Immunol. 119, 140–147. Scully, R., and Livingston, D. M. (2000). Nature 408, 429–432. Shackelformd, R. E., Innes, C. L., Sieber, S. O., Leadon, S. A., and Paules, R. S. (2001). J. Biol. Chem., in press. Shafman, T. D., Saleem, A., Kyriakis, J., Weichselbaum, R., Kharbanda, S., and Kufe, D. W. (1995). Cancer Res. 55, 3242–3245. Shafman, T., Khanna, K. K., Kedar, P., Spring, K., Kozlov, S., Yen, T., Hobson, K., Gatei, M., Zhang, N., Watters, D., Egerton, M., Shiloh, Y., Kharbanda, S., Kufe, D., and Lavin, M. F. (1997). Nature 387, 520–523. Sherr, C. J. (2000). Cancer Res. 60, 3689–3695. Shieh, S.-Y., Ikeda, M., Taya, Y., and Prives, C. (1997). Cell 91, 325–334. Shieh, S.-Y., Ahn, J., Tamai, K., Taya, Y., and Prives, C. (2000). Genes Dev. 14, 289–300. Shiloh, Y. (1995). Eur. J. Hum. Genet. 3, 116–138. Shiloh, Y. (1997). Annu. Rev. Genet. 31, 635–662. Shiloh, Y. (2000). Curr. Opin. Genet. Dev. 11, 71–77. Shinkai, Y., Koyasu, S., Nakayama, K., Murphy, K. M., Loh, D. Y., Reinherz, E. L., and Alt, F. W. (1993). Science 259, 822–825. Sibon, O. C., Laurencon, A., Hawley, R., and Theurkauf, W. E. (1999). Curr. Biol. 9, 302– 312. Siliciano, J. D., Canman, C. E., Taya, Y., Sakaguchi, K., Appella, E., and Kastan, M. B. (1997). Genes Dev. 11, 3471–3481. Smilenov, L. B., Morgan, S. E., Mellado, W., Sawant, S. G., Kastan, M. B., and Pandita, T. K. (1997). Oncogene 15, 2659–2665. Smilenov, L. B., Dhar, S., and Pandita, T. K. (1999). Mol. Cell. Biol. 19, 6963–6971.
252
Yosef Shiloh and Michael B. Kastan
Smith, G. C. M., and Jackson, S. P. (1999). Genes Dev. 13, 916–934. Smith, G. C. M., Cary, R. B., Lakin, N. D., Hann, B. C., Teo, S.-H., Chen, D. J., and Jackson, S. P. (1999). Proc. Natl. Acad. Sci. USA 96, 11134–11139. Soares, H. D., Morgan, J. I., and McKinnon, P. J. (1998). Neuroscience 86, 1045–1054. Spring, K., Cross, S., Li, C., Ben-Senior, L., Watters, D., Waring, P., Ahangari, F., Lu, S., Chen, P., Misko, I., Paterson, C., Kay, G., Smorodinsky, N. I., Shiloh, Y., and Lavin, M. F. (2001). Cancer Res. 61, 4561– 4568. Stankovic, T., Weber, P., Stewart, G., Bedenham, T., Murray, J., Byrd, P. J., Moss, P. A. H., and Taylor, A. M. R. (1999). Lancet 353, 26–29. Starostik, P., Manshouri, T., O’Brien, S., Freireich, E., Kantarjian, H., Haidar, M., Lerner, S., Keating, M., and Albitar, M. (1998). Cancer Res. 58, 4552–4557. Stewart, G. S., Maser, R., Stankovic, T., Bressan, D. A., Kaplan, M., Jaspers, N. G. J., Raama, A., Byrd, P. J., Petrini, J. H., and Taylor, M. R. (1999). Cell 99, 577–587. Stilgenbauer, S., Chaffner, C., Litterst, A., Liebisch, P., Gilad, S., Bar-Shira, A., James, M. R., ¨ Lichter, P., and Dohner, H. (1997). Nature Med. 3, 1155–1159. ¨ ¨ Stilgenbauer, S., Chaffner, C., Winkler, D., Ott, G., Leupolt, E., Bentz, M., Moller, P., Muller¨ Hermelink, H. K., James, M. R., Lichter, P., and Dohner, H. (2000). Ann. Oncol. 11, S127–S130. Stoppa-Lyonnet, D., Soulier, J., Laug´e, A., Dastot, H., Garand, R., Sigaux, F., and Stern, M.-H. (1998). Blood 91, 3920–3926. Su, Y., and Swift, M. (2000). Ann. Intern. Med. 133, 770–778. Swift, M., Reitnauer, P. J., Morrell, D., and Chase, C. L. (1987). N. Engl. J. Med. 316, 1289–1294. Swift, M., Morrell, D., Massey, R. B., and Chase, C. L. (1991). N. Engl. J. Med. 325, 1831–1836. Takao, N., Kato, H., Mori, R., Morrison, C., Sonada, E., Sun, X., Shimizu, H., Yoshioka, K., Rakeda, S., and Yamamoto, K. (1999). Oncogene 18, 7002–7009. Takao, N., Li, Y., and Yamamoto, K. (2000). FEBS Lett. 427, 133–136. Tanoue, T., Moriguchi, T., and Nishida, E. (1999). J. Biol. Chem. 274, 19949–19956. Taylor, A. M. R., Metcalfe, J. A., Oxford, J. M., and Harnden, D. G. (1976). Nature 260, 441–443. Taylor, A. M. R., Metcalfe, J. A., Thick, J., and Mak, Y.-F. (1996). Blood 87, 423 – 438. Telatar, M., Wang, Z., Udar, N., Liang, T., Bernatowska-Matuszkiewicz, E., Lavin, M., Shiloh, Y., Concannon, P., Good, R. A., and Gatti, R. A. (1996). Am. J. Hum. Genet. 59, 40 – 44. ¨ ¨ S., Tolun, A., Chessa, L., Teraoka, S. N., Telatar, M., Becker-Catania, S., Liang, T., Oneng ut, Sanal, O., Bernatowska, E., Gatti, R. A., and Concannon, P. (1999). Am. J. Hum. Genet. 64, 1617–1631. Thacker, J. (1989). Mutat. Res. 220, 187–204. Theodosiou, A., Smith, A., Gillieron, C., Arkinstall, S., and Ashworth, A. (1999). Oncogene 18, 6981–6988. Thompson, J. D., Higgins, D. G., and Gibson, T. J. (1994). Nucleic Acids Res. 22, 4673 – 4680. Tibbetts, R. S., Brumbaugh, K. M., Williams, J. M., Sarkaria, J. N., Cliby, W. A., Shieh, S. Y., Taya, Y., Prives, C., and Abraham, R. T. (1999). Genes Dev. 13, 152–157. Tibbetts, R. S., Cortez, D., Brumbaugh, K. M., Scully, R., Livingston, D., Elledge, S. J., and Abraham, R. T. (2000). Genes Dev. 14, 2989–3002. Tibbles, L. A., and Woodgett, J. R. (1999). Cell. Mol. Life Sci. 55, 1230–1254. Tominaga, K., Morisaki, H., Kenako, Y., Fujimoto, A., Tanaka, T., Ohtsubo, M., Hirai, M., Okayama, H., Ikeda, K., and Nakanishi, M. (1999). J. Biol. Chem. 274, 31463–31467. Uhrhammer, N., Fritz, E., Boyden, L., and Meyn, M. S. (1999). Int. J. Mol. Med. 4, 43–47. Uziel, T., Savitsky, K., Platzer, M., Ziv, Y., Helbitz, T., Nehls, M., Boehm, T., Rosenthal, A., Shiloh, Y., and Rotman, G. (1996). Genomics 33, 317–320.
ATM
253
Vallen, E. A., and Cross, F. R. (1999). Genetics 151, 459–471. Varon, R., Vissinga, C., Platzer, M., Cerosaletti, K. M., Chrzanowska, K. H., Saar, K., Beckmann, G., Seemanova, E., Cooper, P. R., Nowak, N. J., Stumm, M., Weemaes, C. M. R., Gatti, R. A., Wilson, R. K., Digweed, M., Rosenthal, A., Sperling, K., Concannon, P., and Reis, A. (1998). Cell 93, 467–476. Vassilev, A., Yamauchi, J., Kotani, T., Prives, C., Avantaggiati, M. L., Qin, J., and Nakatani, Y. (1998). Mol. Cell 2, 869–875. Vaziri, H. (1997). Biochemistry 62, 1306–1310. Vaziri, H., West, M. D., Allsopp, R. C., Davison, T. S., Wu, Y. S., Arrowsmith, C. H., Poirier, G. G., and Benchimol, S. (1997). EMBO J. 16, 6018–6033. Verheij, M., Ruiter, G. A., Zerp, S. F., van Blitterswijk, W. J., Fuks, Z., Haimovitz-Friedman, A., and Bartelink, H. (1998). Radiother. Oncol. 47, 225–232. Vialard, J. E., Gilbert, C. S., Green, C. M., and Lowndes, N. F. (1998). EMBO J. 17, 5679–5688. von Boehmer, H. (1994). Cell 76, 219–225. Vorechovsky, I., Luo, L., Dyer, M. J. S., Catovsky, D., Amlot, P. L., Yaxley, J. C., Foroni, L., ¨ L., Webster, A. D. B., and Yuille, M. A. R. (1997). Nature Genet. 17, 96–99. Hammarstrom, Ward, A. J., Olive, P. L., Burr, A. H., and Rosin, M. P. (1994). Environ. Mol. Mutagen. 24, 103–111. Wang, X., McGowan, C. H., Zhao, M., He, L., Downey, J. S., Fearns, C., Wang, Y., Huang, S., and Han, J. (2000). Mol. Cell. Biol. 20, 4543–4552. Wang, Y., Cortez, D., Yazdi, P., Neff, N., Elledge, S. J., and Qin, J. (2000). Genes Dev. 14, 927–939. Wang, Y. A., Elson, A., and Leder, P. (1997). Proc. Natl. Acad. Sci. USA 94, 14590–14595. Watters, D., Khanna, K. K., Beamish, H., Birrell, G., Spring, K., Kedar, P., Gatei, M., Stenzel, D., Hobson, K., Kozlov, S., Zhang, N., Farrell, A., Ramsay, J., Gatti, R., and Lavin, M. (1997). Oncogene 14, 1911–1921. Watters, D., Kedar, P., Spring, K., Bjorkman, J., Chen, P., Gatei, M., Birrell, G., Garrone, B., Srinivasa, P., Crane, D. I., and Lavin, M. F. (1999). J. Biol. Chem. 274, 34277–34282. Westphal, C. H., Schmaltz, C., Rowan, S., Elson, A., Fisher, D. E., and Leder, P. (1997a). Cancer Res. 57, 1664–1667. Westphal, C. H., Rowan, S., Schmaltz, C., Elson, A., Fisher, D. E., and Leder, P. (1997b). Nature Genet. 16, 397– 401. Westphal, C. H., Hoyes, K. P., Canman, C. E., Huang, X., Kastan, M. B., Hendry, J. H., and Leder, P. (1998). Cancer Res. 58, 5637–5639. Widmann, C., Gibson, S., Jarpe, M. B., and Johnson, G. L. (1999). Physiol. Rev. 79, 143– 180. Wright, J., Teraoka, S., Onengut, S., Tolun, A., Gatti, R. A., Ochs, H. D., and Concannon, P. (1996). Am. J. Hum. Genet. 59, 839–846. Wu, X., Ranganathan, V., Weisman, D. S., Heine, W. F., Ciccone, D. N., O’Neill, T. B., Crick, K. E., Pierce, K. A., Lane, W. S., Rathbun, G., Livingston, D. M., and Weaver, D. T. (2000). Nature 405, 477– 482. Xia, S. J., Shammas, M. A., and Shmookler Reis, R. J. (1996). Mutat. Res. 364, 1–11. Xu, Y. (1999). Adv. Immunol. 72, 179–189. Xu, Y., and Baltimore, D. (1996). Genes Dev. 10, 2401–2410. Xu, Y., Ashley, T., Brainerd, E. E., Bronson, R. T., Meyn, M. S., and Baltimore, D. (1996). Genes Dev. 10, 2411–2422. Xu, Y., Yang, E. M., Brugarolas, J., Jacks, T., and Baltimore, D. (1998). Mol. Cell. Biol. 18, 4385–4390. Yan, J., Khanna, K. K., and Lavin, M. F. (2000). Int. J. Radiat. Biol. 76, 1025–1035. Yang, D., and Kastan, M. B. (2000). Nat. Cell Biol. 2, 893–898. Yu, B. P. (1994). Physiol. Rev. 74, 139–162.
254
Yosef Shiloh and Michael B. Kastan
Yuan, Z.-M., Shioya, H., Ishiko, T., Sun, X., Gu, J., Huang, Y., Lu, H., Kharbanda, S., Weichselbaum, R., and Kufe, D. (1999). Nature 399, 814–817. Yuille, M. A. R., Coignet, L. J. A., Abraham, S. M., Yaqub, F., Luo, L., Matutes, E., BritoBabapulle, V., Vorechovsky, I., Dyer, M. J. S., and Catovsky, D. (1998). Oncogene 16, 789–796. Zhang, N., Chen, P., Khanna, K. K., Scott, S., Gatei, M., Kozlov, S., Watters, D., Spring, K., Yen, T., and Lavin, M. F. (1997). Proc. Natl. Acad. Sci. USA 94, 8021–8026. Zhang, N., Chen, P., Gatei, M., Scott, S., Khanna, K. K., and Lavin, M. F. (1998). Oncogene 17, 811–818. Zhao, S., Weng, Y.-C., Yuan, S.-S. F., Lin, Y.-T., Hsu, H.-C., Lin, S.-C. J., Gerbino, E., Song, M.-H., Zdzlenicka, M. Z., Gatti, R. A., Shay, J. W., Ziv, Y., Shiloh, Y., and Lee, E. Y. H. P. (2000). Nature 405, 473–477. Zhong, Q., Chen, C.-F., Li, S., Chen, Y., Wang, C.-C., Xiao, J., Chen, P.-L., Sharp, Z. D., and Lee, W.-H. (1999). Science 285, 747–750. Zhou, B. B., Chaturvedi, P., Spring, K., Scott, S. P., Johanson, R. A., Mishra, R., Mattern, M. R., Winkler, J. D., and Khanna, K. K. (2000). J. Biol. Chem. 275, 10342–10348. Zhou, B.-B. S., and Elledge, S. J. (2000). Nature 408, 433–439. Ziv, Y., Bar-Shira, A., Jorgensen, T. J., Russell, P. S., Sartiel, A., Shows, T. B., Eddy, R. L., Buchwald, M., Legerski, R., Schimke, R. T., and Shiloh, Y. (1995). Somatic Cell Mol. Genet. 21, 99–111. Ziv, Y., Bar-Shira, A., Pecker, I., Russell, P. S., Jorgensen, T. J., Tsarfati, I., and Shiloh, Y. (1997). Oncogene 15, 159–167.
Index
A Adenomatous polyposis coli, 194 –195 Alleles CDH-1, inactivation, 60–61 HLA, loss, 127–128 Angiogenesis, TGF- role, 28–30 Antigens HLA class I phenotype I, 121–126 phenotype II, 126–127 phenotype III, 127 phenotype IV, 127–128 phenotype V, 128 phenotype VI, 128 primary tumor expression, 119–121 HLA class II, 147–149 MHC class I, 139–141 APC, see Adenomatous polyposis coli Apoptosis, in neoplastic transformation, 184 –187 Aromatic hydrocarbons, in carcinogenesis, 199–201 A–T, see Ataxia–telangiectasia Ataxia–telangiectasia ATM gene, 216–217 ATM mutations in cancer, 217–218 ATM role in meiosis, 235–236 cellular phenotype, 212–216 clinical–pathological phenotype, 211–212 immune system, 233–235 neurogenesis and neurodegeneration, 236–238 signaling pathways, 240–241 ATM protein A–T cellular phenotype, 212–216 A–T clinical–pathological phenotype, 211–212 Atm-deficient mouse phenotype, 231–232 Atm-mediated damage response, 232–233 ATM mutations, 216–218
characterization, 223–224 in DNA damage response, 241–243 gene mutation, ATM, 216–218 immune system gene maturation, 233–235 kinase activity, 224–226 in meiosis, 235–236 neurodegeneration, 236–238 neurogenesis, 236–238 in oxidative stress, 238–240 sequence motifs, 218–223 in signaling pathways, 240–241 substrates, downstream pathways, 226–231
B B cells diffuse large B cell lymphoma, 102–104 Ig gene role, 83–86 tumors Ig genes, 90–91 somatic mutations, 92–95 VH gene, 91–92 Behavior, cell in transformation dynamics, 188 Benzpyrene diol epoxides, 200 Bladder carcinomas, 129–130 BPDEs, see Benzpyrene diol epoxides Breast carcinomas, 130–131
C E-Cadherin, 57–58 Calf serum in interclonal heterogeneity studies, 174 –178 in NIH 3T3 transformation, 169–174 spontaneous transformation stages, 178–181
255
256 Cancer ATM mutations, 217–218 DNA damage response defects, 241–243 hereditary diffuse gastric cancer, 56–58, 60–63 hereditary nonpolyposis colorectal cancer, 181 and HSPGs, 69–70 pancreas, 137 prostate, 137–138 selection, 190–195 Carcinogenesis, selection, 199–201 Carcinomas, HLA expression bladder, 129–130 breast, 130–131 cervix, 131–132 colorectal, 132–133 head, 134 lung, 134 –135 melanoma, 135–137 neck, 134 renal cell, 138–139 transitional cell, 129–131 CDH-1 allele inactivation, 60–61 in HDGC, 57–58 mutations, 58–60 Cell behavior, in transformation dynamics, 188 Cell culture, spontaneous transformation mouse fibroblasts, 164–165 primary cell strains, 162–163 rat epithelial cells, 165–168 Cell differentiation, ATM role, 240–241 Cell growth, inhibition, 187–188 Cell lines NK cells, 145–146 spontaneous transformation apoptosis effects, 184 –187 elicitation and expression conditions, 169–174 interclonal heterogeneity, 174 –178 mouse fibroblasts, 164 –165 rat epithelial cells, 165–168 stages, 178–181 Cellular phenotype, A–T DNA DSBs, 212–216 genomic instability, 212 signaling pathway defects, 216 Cervix carcinoma, 131–132 c-Ha-ras, 192
Index
Chromosomal translocations, Ig locus, 108–109 Chronic lymphocytic leukemia, V gene mutational status, 95–96 status prognostic value, 96–98 usage, 98–99 Clinical criteria, HDGC, 62–63 Clinical–pathological phenotype, A–T, 211–212 CLL, see Chronic lymphocytic leukemia Clonal expansion, in spontaneous transformation, 181–184 Colonization, MHC class I antigen expression, 139–141 Colorectal carcinoma, 132–133 CS, see Calf serum
D Damage response Atm-mediated, 232–233 ATM role, 241–243 DNA, 224 –226 Decapentaplegic, signaling, in Drosophila, 73 Desmoplasia, TGF- role, 32–33 Differentiation ATM role, 240–241 epithelial–mesenchymal, TGF- effects, 20–23 Diffuse large B cell lymphoma, 102–104 Disease, A–T cellular phenotype, 212–216 clinical–pathological phenotype, 211–212 DLBCL, see Diffuse large B cell lymphoma DNA damage effect on ATM kinase activity, 224 –226 damage response, ATM role, 241–243 DSBs, 212–216 Double-stranded breaks, A–T DNA, 212–216 Downstream signaling pathways, TGF- identification, 6–7 Smad proteins, 7–11 Dpp, see Decapentaplegic Drosophila, and HSPGs Dpp signaling, 73 FGF signaling, 71–73 Wg and Hh signaling, 73–75 DSBs, see Double-stranded breaks
257
Index
E
H
ECM proteins, see Extracellular matrix proteins EMT, see Epithelial–mesenchymal transdifferentiation Epithelial cells, rat, 165–168 Epithelial–mesenchymal transdifferentiation, TGF- effects, 20–23 Extracellular matrix proteins, TGF- effects, 24
Haplotype, HLA, loss, 126–127 Ha-ras-1, 192 HDGC, see Hereditary diffuse gastric cancer Head carcinoma, HLA expression, 134 Heparan sulfate proteoglycans and cancer, 69–70 in Drosophila and Dpp signaling, 73 Wg and Hh signaling, 73–75 and FGF signaling, in Drosophila, 71–73 in Hh movement, 77 Hereditary diffuse gastric cancer CDH-1 allele inactivation, 60–61 clinical criteria, 62–63 first evidence, 56–58 susceptibility mechanisms, 61–62 tumor spectrum, 60 Hereditary nonpolyposis colorectal cancer, 181 Heterogeneity, interclonal, spontaneous transformation, 174–178 Hh movement, HSPG role, 77 signaling, in Drosophila, 73–75 HLA in bladder carcinomas, 129–130 in breast carcinomas, 130–131 in cervix carcinomas, 131–132 class I nonclassical molecules, 143–145 and T cell-based immunotherapy, 146–147 class I antigen expression in primary tumors, 119–121 phenotype I, 121–126 phenotype II, 126–127 phenotype III, 127 phenotype IV, 127–128 phenotype V, 128 phenotype VI, 128 class II antigens, 147–149 in colorectal carcinomas, 132–133 in head carcinomas, 134 HLA-A, 127 HLA-B, 127 HLA-C, 127 in lung carcinomas, 134 –135 in melanoma, 135–137 in neck carcinomas, 134 pancreas cancer, 137
F FBS, see Fetal bovine serum Fetal bovine serum, spontaneous transformation, 169–171 FGF, see Fibroblast growth factor Fibroblast growth factor, signaling and HSPGs, 71–73 Fibroblasts, mouse, spontaneous transformation, 164 –165 FL, see Follicular lymphoma Follicular lymphoma, 100–102
G Genes APC, 194–195 ATM, 216–218 CDH-1, 57–61 Ig, 83–92 immune system, maturation, 233–235 Mom-1, 194 oncogenes, in TGF- pathways, 40–41 p53, 200–201 V mutational status, 95–96, 105–107 prognostic value, 96–98 recombination and selection, 83–86 usage, 98–99, 107–108 VH, in B cell tumors, 91–92 Genetic variation, selection in tumor development, 195–196 Genomic instability A–T cellular phenotype, 212 TGF- effects, 23 Glypicans, in Wg signaling, 75–76 Growth, ATM role, 240–241
258 HLA (continued ) prostate cancer, 137–138 renal cell carcinoma, 138–139 in tumor tissues, 128–129 HNPCC, see Hereditary nonpolyposis colorectal cancer Host tissue, invading cell cross talk, 27–28 HPRT, gene mutation, 195 HSPGs, see Heparan sulfate proteoglycans Humans, selection in cancer, 190–195 Hydrocarbons, in carcinogenesis selection, 199–201
I Ig, see Immunoglobulin Immune system, A–T, 233–235 Immunodeficiency, A–T, 233–235 Immunoglobulin chromosomal translocations, 108–109 genes in B cell tumors, 90–92 somatic mutation and isotype switch, 87–90 V gene recombination and selection, 83–86 Immunoselection, T cell, MHC class I-negative tumor clones, 141–143 Immunosuppression, TGF- role, 30–32 Immunotherapy, T cell-based, and HLA class I loss, 146–147 Interclonal heterogeneity, spontaneous transformation, 174–178 Interferon, and HLA class I antigen phenotype VI, 128 Invasion, tumor cells, TGF- effects, 23–28 Isotype switch, Ig genes, 87–90
K Knockout mouse Atm-deficient phenotype, 231–232 Atm-mediated damage response, 232–233
L LAP, see Latency-associated protein Latency-associated protein, 4 LDPs, see Low-density passages Ligands, TGF-, 4 –5 Low-density passages
Index
and interclonal heterogeneity, 174–178 mouse fibroblasts, 164–165 NIH 3T3 cells, 169–174 rat epithelial cells, 165–168 and spontaneous transformation dynamics, 188 Lung carcinoma, HLA expression, 134 –135 Lymphoid tumors, A–T, 233–235 Lymphoma A–T tumors, 233–235 chronic lymphocytic leukemia, 95–99 diffuse large B cell lymphoma, 102–104 follicular, 100–102
M MAP kinase, see Mitogen-activated protein kinase Matrix metalloproteinases, TGF- effects, 24–25 Maturation, immune system genes in A–T, 233–235 Meiosis, ATM role, 235–236 Melanoma, HLA expression, 135–137 Metastasis, TGF- effects, 23–28 MHC class I antigen, expression, 139–141 negative tumor clones, 141–143 Mitogen-activated protein kinase, TGF- signaling pathways, 11–13 MM, see Multiple myeloma MMPs, see Matrix metalloproteinases Mom-1, 194 Mouse fibroblasts, spontaneous transformation, 164 –165 knockout Atm-deficient phenotype, 231–232 Atm-mediated damage response, 232–233 Multiple myeloma plasma cell tumors, 104 –105 V gene mutational status, 105–107 Mutations ATM, 216–218 in B cell tumors, 92–95 CDH-1, 58–60 HPRT gene, 195 Ig genes, 87–90 p53 gene, 200–201
259
Index
Oncogenes, in TGF- signal transduction pathways, 40– 41 Oncogenesis latent TGF-, 19–20 TGF- isoform activities, 18–19 TGF- signaling pathways oncogene role, 40– 41 receptors, 33–37 Smad interacting proteins, 40 Smad mutations, 37– 40 Oxidative stress, ATM role, 238–240
phenotype V, 128 phenotype VI, 128 Phosphatidylinositol kinase related family ATM activity, 224 –226 ATM sequence motif, 218–223 PIK related family, see Phosphatidylinositol kinase related family Plasma cell tumors Ig locus and chromosomal translocations, 108–109 multiple myeloma, 104 –105 V gene mutational status, 105–107 V gene usage, 107–108 Plasminogen–plasmin protease system, TGF- role, 25–26 Polycyclic aromatic hydrocarbons, in carcinogenesis selection, 199–201 Primary tumors, HLA class I antigen expression, 119–121 phenotype I, 121–126 phenotype II, 126–127 phenotype III, 127 phenotype IV, 127–128 phenotype V, 128 phenotype VI, 128 Prostate cancer, HLA expression, 137–138 Proteins ATM, see ATM protein extracellular matrix proteins, 24 latency-associated protein, 4 matrix metalloproteinases, 24 –25 mitogen-activated protein kinase, 11–13 Smad interacting proteins, 40 Smad proteins, 7–11, 37– 40
P
R
p53, 200–201 PAHs, see Polycyclic aromatic hydrocarbons Pancreas cancer, HLA expression, 137 Phenotypes A–T cellular phenotype, 212–216 clinical–pathological phenotype, 211–212 Atm-deficient mouse, 231–232 HLA class I phenotype I, 121–126 phenotype II, 126–127 phenotype III, 127 phenotype IV, 127–128
Rat epithelial cells, 165–168 NMU treatment, 192 Recombination, V gene, 83–86 Renal cell carcinoma, HLA expression, 138–139
Smad, in tumors, 37– 40 V gene status, 95–96, 105–107 Myeloma, multiple plasma cell tumors, 104 –105 V gene mutational status, 105–107
N Neck carcinoma, HLA expression, 134 Neurodegeneration, in A–T, 236–238 Neurogenesis, in A–T, 236–238 NIH 3T3, spontaneous transformation apoptosis effects, 184 –187 elicitation and expression conditions, 169–174 N-Nitroso-N-methylurea, rat treatment, 192 NK cells, escape from tumor, 145–146 NMU, see N-Nitroso-N-methylurea Nontransformed cells, growth inhibition, 187–188
O
S Selection in cancer, 190–195 in carcinogenesis, 199–201 in neoplastic transformation, 184 –187 in tumor development, 195–196
260 Selection (continued ) V gene, 83–86 in vivo, 196–199 Signaling pathways A–T, defects, 216 ATM effect, 240–241 downstream ATM substrates, 226–231 TGF- identification, 6–7 TGF- Smad proteins, 7–11 Dpp, in Drosophila, 73 FGF, and HSPGs, in Drosophila, 71–73 Hh, in Drosophila, 73–75 TGF- alternative pathways, 13–14 ligands, 4–5 MAP kinase signaling, 11–13 TGF-, in oncogenesis oncogene role, 40– 41 receptors, 33–37 Smad interacting proteins, 40 Smad mutations, 37– 40 Wg in Drosophila, 73–75 glypicans role, 75–76 Smad interacting proteins, 40 Smad proteins mutations in tumors, 37– 40 signaling pathway, 7–11 Somatic mutations in B cell tumors, 92–95 Ig genes, 87–90 Spontaneous neoplastic transformation cell growth inhibition, 187–188 elicitation and expression conditions, 169–174 interclonal heterogeneity, 174 –178 major features, 188–190 mouse fibroblasts, 164 –165 primary cell strains, 162–163 rat epithelial cells, 165–168 selection in vivo, 196–199 selective clonal expansion role, 181–184 stages, 178–181 variable cell behavior effects, 188 Stress, oxidative, ATM role, 238–240
T TCC, see Transitional cell carcinoma T cells
Index
immunoselection, 141–143 immunotherapy, 146–147 Telomere, ATM role, 235–236 TGF-, see Transforming growth factor- Tissues, host, invading cell cross talk, 27–28 Tobacco smoke, in carcinogenesis selection, 200 Transdifferentiation, see Epithelial–mesenchymal transdifferentiation Transformation, see Spontaneous neoplastic transformation Transformed cells, growth inhibition, 187–188 Transforming growth factor- alternative signaling pathways, 13–14 in angiogenesis, 28–30 bimodal action in tumorigenesis, 14 in desmoplasia, 32–33 downstream signaling pathways, 6–11 in ECM, 24 in EMT, 20–23 in genomic instability, 23 in immunosuppression, 30–32 in invading cell–host tissue cross talk, 27–28 isoforms, 18–19 latent, 19–20 ligands, 4 –5 in MAPK, 11–13 in metastasis, 23–28 in MMPs, 24–25 and oncogenes, 40– 41 in plasminogen–plasmin protease system, 25–26 signaling pathway ligands, 4 –5 Smad interacting proteins, 40 Smad mutations, 37– 40 TGF- receptors, 6, 33–37 in tumor cell invasion, 23–28 in tumor promotion, 17–18 in tumor suppression, 14 –17 Transforming growth factor- receptors in oncogenesis, 33–37 TGF- binding, 6 Transitional cell carcinoma, HLA expression, 129–131 Translocations, chromosomal, Ig locus, 108–109
261
Index
Tumor cells invasion and metastasis, 23–28 plasma Ig locus and chromosomal translocations, 108–109 multiple myeloma, 104 –105 V gene mutational status, 105–107 V gene usage, 107–108 Tumorigenesis, TGF- role in angiogenesis, 28–30 in bimodal action, 14 in desmoplasia, 32–33 in immunosuppression, 30–32 isoform activities, 18–19 latent TGF-, 19–20 signaling pathways alternative pathways, 13–14 downstream, 6–11 ligands, 4 –5 MAP kinase signaling, 11–13 oncogene role, 40– 41 receptors, 33–37 Smad interacting proteins, 40 Smad mutations, 37– 40 tumor promotion, 17–18 tumor suppression, 14 –17 Tumors B cell Ig genes, 90–91 somatic mutations, 92–95 VH gene, 91–92 development, genetic variation, 195–196 HDGC, 60 HLA class II antigens, 147–149 HLA expression, 128–129
bladder carcinomas, 129–130 breast carcinomas, 130–131 cervix carcinoma, 131–132 colorectal carcinoma, 132–133 head carcinoma, 134 lung carcinoma, 134 –135 melanoma, 135–137 neck carcinoma, 134 pancreas cancer, 137 prostate cancer, 137–138 renal cell carcinoma, 138–139 lymphoid, A–T, 233–235 MHC class I-negative clones, 141–143 NK escape mechanisms, 145–146 nonclassical HLA class I molecule expression, 143–145 primary, see Primary tumors promotion by TGF-, 17–18 Smad mutations, 37– 40 Tumor suppression, TGF- role, 14 –17
V V genes mutational status, 95–96, 105–107 prognostic value, 96–98 recombination and selection, 83–86 usage, 98–99, 107–108 VH, in B cell tumors, 91–92
W Wg signaling in Drosophila, 73–75 glypicans role, 75–76
This Page Intentionally Left Blank