JAK/STAT Signaling Pathway in Lymphoid Malignancies

IDENTIFICATION AND CHARACTERIZATION OF NOVEL RECURRENT MUTATIONS IN PRIMARY MEDIASTINAL LARGE B CELL LYMPHOMA AND HODGKIN LYMPHOMA

JITHENDRA GUNAWARDANA

A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF

DOCTOR OF PHILOSOPHY

THE FACULTY OF GRADUATE AND POSTDOCTORAL STUDIES

(Pathology and Laboratory Medicine)

THE UNIVERSITY OF BRITISH COLUMBIA

(Vancouver)

July 2015

Abstract

Classical Hodgkin lymphoma (HL) and primary mediastinal large B cell lymphoma (PMBCL) are related lymphomas sharing pathological, molecular and clinical characteristics. Here we discovered by next-generation sequencing recurrent somatic coding-sequence mutations in the protein tyrosine phosphatase PTPN1 and the cytokine receptor IL4R. Mutations in PTPN1 were found in 6 of 30 (20%) HL cases, in 6 of 9 (67%) HL– derived cell lines, in 17 of 77 (22%) PMBCL cases and in 1 of 3 (33%) PMBCL-derived cell lines, consisting of nonsense, missense and frameshift mutations. We demonstrate that PTPN1 mutations lead to reduced phosphatase activity and increased phosphorylation of JAK-STAT pathway members. Moreover, silencing of PTPN1 by RNA interference in HL cell line KM-H2 resulted in hyperphosphorylation and overexpression of the downstream oncogenes BCL6 and MYC. Mutations in IL4R were found in 18 of 65 (28%) PMBCL cases confirming a ‘hotspot’ missense mutation I242N in exon 8 in 11 of 18 (61%) mutated cases. Ectopic expression of the mutant I242N in HEK 293 cells showed increased activated STAT6- dependent SEAP reporter gene expression without interleukin-4 stimulation. Introduction of the mutant into Hodgkin lymphoma cell line DEV showed cytokine- independent hyperphosphorylation of JAK-STAT pathway members and upregulation of the T cell regulatory chemokine TARC (CCL17) and the B cell activation marker CD23. Our data suggest loss-of-function PTPN1 and gain-of-function IL4R mutations leading to oncogenic JAK-STAT activation as new driver alterations in lymphomagenesis with implications for future treatment strategies.

ii Preface

This dissertation is to the best of my knowledge original, except where references are made to previous work. Neither this, nor any substantially similar dissertation has been or is being submitted for any other degree, diploma or other qualification at any other university.

The studies reported in Chapters 2-3 were conducted in the Department for Lymphoid Cancer Research (LCR) of the British Columbia Cancer Research Centre (BCCRC; Vancouver) and were approved by the University of British Columbia’s Research Ethics Board (certificate no: H12-02176).

Chapter 2 includes published material for which multiple co-authors were involved. Below is a description of the contribution of the candidate and co-authors to the work presented:

Chapter 2: Jay Gunawardana, Fong C. Chan, Adèle Telenius, Bruce Woolcock, Robert Kridel, King L. Tan, Susana Ben-Neriah, Anja Mottok, Raymond S. Lim, Merrill Boyle, Sanja Rogic, Lisa M. Rimsza, Chrystelle Guiter, Karen Leroy, Philippe Gaulard, Corinne Haioun, Marco A. Marra , Kerry J. Savage, Joseph M. Connors, Sohrab P. Shah, Randy D. Gascoyne and Christian Steidl. Recurrent somatic mutations of PTPN1 in primary mediastinal B cell lymphoma and Hodgkin lymphoma. Nature Genetics. 2014; 46(4):329-35. The candidate designed the research, analyzed and interpreted data, wrote the manuscript and performed the following experiments: Mutational screening by Sanger sequencing, quantitative RT-PCR, ectopic expression of mutants, inhibitor studies, protein blotting and immunoprecipitation, ELISA, reporter gene expression and activity assays. A.T., B.W., K.L.T., S.B.-N., A.M., M.B., R.K. and C.G. performed experiments and interpreted data. F.C.C., R.S.L. and S.R. analyzed data from whole-genome sequencing, RNAseq and amplicon sequencing. C.H., K.L., L.M.R. and P.G. provided study material and interpreted results. K.J.S., M.A.M. and S.P.S. interpreted data.

iii J.M.C. and R.D.G. curated the lymphoma database, interpreted data and reviewed the manuscript. C.S. designed the research, analyzed and interpreted data and wrote the manuscript.

iv Table of Contents

Abstract ...... ii Preface ...... iii Table of Contents ...... v List of Tables ...... vii List of Figures ...... viii List of Abbreviations ...... x Acknowledgements ...... xiii Dedication ...... xiv

Chapter 1 : Introduction ...... 1 1.1. B cell lymphomas ...... 2 1.1.1. Normal B cell development ...... 2 1.1.2. B cell differentiation and the germinal centre reaction ...... 2 1.2. Primary mediastinal large B cell lymphoma ...... 3 1.2.1. Epidemiology ...... 3 1.2.2. Clinical presentation ...... 3 1.2.3. Pathology ...... 3 1.2.4. Therapy...... 4 1.3. Hodgkin lymphoma ...... 4 1.3.1. Epidemiology ...... 4 1.3.2. Clinical presentation ...... 5 1.3.3. Pathology ...... 5 1.3.4. Therapy...... 5 1.4. Molecular aberrations in PMBCL and HL ...... 6 1.5. JAK-STAT signaling in cancer ...... 6 1.5.1. JAK-STAT signaling aberrations are common in lymphoid cancers ...... 8 1.5.2. Changes to the tumor microenvironment ...... 9 1.5.3. Phosphorylation of STAT proteins ...... 10 1.5.4. STAT-regulated genes in oncogenesis ...... 12 1.5.5. Somatic gene mutations in the JAK-STAT pathway ...... 16

v 1.6. Thesis theme and objectives ...... 23 1.7. Hypotheses ...... 24 1.7.1. Hypothesis 1 ...... 24 1.7.2. Hypothesis 2 ...... 24 1.8. Aims and thesis outline ...... 24 1.8.1. Aim 1: To determine the frequency and the biological significance of PTPN1 mutations in PMBCL and HL ...... 24 1.8.2. Aim 2: To determine the frequency and the biological significance of IL4R mutations in PMBCL ...... 24 Chapter 2 : Recurrent somatic mutations of PTPN1 in primary mediastinal large B cell lymphoma and Hodgkin lymphoma ...... 36 2.1. Introduction ...... 37 2.2. Materials and methods ...... 38 2.3. Results ...... 46 2.4. Discussion ...... 51 Chapter 3 : Recurrent somatic IL4R mutations in primary mediastinal large B cell lymphoma ...... 104 3.1. Introduction ...... 105 3.2. Materials and methods ...... 106 3.3. Results ...... 111 3.4. Discussion ...... 115 Chapter 4 : Conclusions ...... 144 4.1. Summary ...... 145 4.2. Results from our study analyzing PTPN1 mutations provide insights into oncogenic signaling in PMBCL and HL...... 145 4.3. Results from our study analyzing IL4R mutations provide insights into PMBCL biology...... 147 4.4. Therapeutic potential of JAK-STAT inhibition ...... 149 4.5. Current studies ...... 153 4.6. Future studies ...... 154

Bibliography ………….……………………………………………………………………...159

vi List of Tables

Table 1.1. Genetic aberrations of the JAK-STAT pathway in lymphoid malignancies. .. 28 Table 1.2. JAK-STAT somatic gene mutation frequencies in lymphoid malignancies. .. 29 Table 2.1. Demographic and clinical characteristics of patients in PTPN1 study...... 54 Table 2.2. Primers used for deep amplicon sequencing...... 56 Table 2.3. Primers used for PTPN1 PCR amplification...... 58 Table 2.4. Somatic coding SNVs identified in two PMBCL index patients by whole genome sequencing...... 69 Table 2.5. Deep amplicon sequencing validation of somatic coding SNVs identified in the two PMBCL index patients...... 72 Table 2.6. Somatic coding indels identified in two PMBCL index patients by whole genome sequencing...... 74 Table 2.7. PTPN1 mutations in patient and cell line samples...... 76 Table 2.8. Ingenuity pathway analysis of differentially expressed genes in PTPN1 knockdown cells...... 79 Table 3.1. Demographic and clinical characteristics of patients in IL4R study...... 118 Table 3.2. Primers used for IL4R PCR amplification...... 119 Table 3.3. IL4R mutations in patient and cell line samples...... 121 Table 3.4. Differentially expressed genes in IL4R mutant by RNAseq...... 123

vii List of Figures

Figure 1.1. The germinal centre reaction...... 30 Figure 1.2. PMBCL and HL hematoxylin and eosin stains...... 31 Figure 1.3. Deregulated JAK-STAT signaling in malignant lymphomas...... 33 Figure 1.4. IL7R activating mutations in T-cell acute lymphoblastic leukemia...... 34 Figure 1.5. Summary of reported JAK2 fusions in lymphoid malignancies...... 35 Figure 2.1. Results from whole-genome sequencing of two PMBCL tumors and germline DNA...... 84 Figure 2.2. Locations of PTPN1 mutations in HL and PMBCL...... 85 Figure 2.3. Immunohistochemical analysis of PTP1B expression in tumors...... 86 Figure 2.4. Tissue microarray analysis of PTP1B expression in PMBCL and in HL clinical specimens ...... 88 Figure 2.5. PTPN1 mutations and survival outcomes in PMBCL patients...... 90 Figure 2.6. PTPN1 mutations and survival outcomes in HL patients...... 92 Figure 2.7. PTPN1 allelic imbalances in PMBCL and HL cell lines...... 94 Figure 2.8. Protein blotting and qRT-PCR analysis of PTP1B expression and PTPN1 transcript levels in PMBCL and HL cell lines...... 96 Figure 2.9. Mutations in MedB-1 are in trans-allelic configuration...... 97 Figure 2.10. Expression of PTP1B mutants in HEK 293-STAT6 cells...... 99 Figure 2.11. Activation of JAK-STAT in a PTPN1-silenced KM-H2...... 101 Figure 2.12. Gene set enrichment analysis of differentially expressed genes in PTPN1 silenced cells...... 103 Figure 3.1. Locations of IL4R mutations in PMBCL...... 124 Figure 3.2. IL4R mutations and survival outcomes in PMBCL patients...... 125 Figure 3.3. IL4R expression and copy number analysis in PMBCL cell lines...... 126 Figure 3.4. Mutation-induced hyperphosphorylation of STAT6 in HEK293 cells...... 128 Figure 3.5. IL4R, IL2RG and IL13R co-expression in HEK293 cells...... 131 Figure 3.6. Mutagenesis of STAT6 binding sites in IL4R...... 133 Figure 3.7. Retroviral-mediated IL4R expression in DEV cells...... 134 Figure 3.8. Mutant-induced JAK-STAT activation in DEV IL4R I242N cells...... 135 Figure 3.9. Top differentially expressed genes in DEV-IL4R-I242N...... 137

viii Figure 3.10. Validation of CD23 and TARC upregulation...... 139 Figure 3.11. Protein-protein interactions of mutant IL4R...... 141 Figure 3.12. Attenuation of STAT5 phosphorylation by inhibitory compounds...... 143 Figure 4.1. The hallmarks of cancer...... 157 Figure 4.2. Overview of deregulated signaling pathways in PMBCL...... 158

ix List of Abbreviations (in alphabetical order)

ABC - Activated B cell-like AID - Activation-induced cytidine deaminase ALCL - Anaplastic large-cell lymphoma ALL - Acute lymphoblastic leukemia AML - acute myelogenous leukemia APC - Antigen presenting cell ATL - Adult T cell leukemia ATLL - Adult T-cell leukemia/ lymphoma ATP - Adenosine triphosphate ATR - Ataxia Telangiectasia and Rad3 BCR - B cell receptor BL - Burkitt’s lymphoma CD - Cluster of differentiation CDK - Cyclin-dependent kinase CLL - Chronic lymphocytic leukemia CML - Chronic myelogenous leukemia CMV - Cytomegalovirus COPD - Chronic obstructive pulmonary disease CRISPR - Clustered regularly interspaced short palindromic repeats CTCL - Cutaneous T cell lymphoma DFS - Disease-free survival DLBCL - Diffuse large B cell lymphoma DSS - Disease-specific survival EGFR - Epidermal growth factor receptor ELISA - Enzyme-linked immunosorbent assay ER - Endoplasmic reticulum FDR - False discovery rate FERM - Band 4.1 ezrin, radixin and moesin FEV - Feline endogenous virus FFPET - Formalin-fixed paraffin embedded tissue FICTION - Fluorescence immunophenotyping and interphase cytogenetic as a tool for investigation of neoplasia FISH - Fluorescence in-situ hybridization FL - Follicular lymphoma GAS - Gamma interferon activation site GC - Germinal centre GCB - Germinal centre B cell-like GFP - Green fluorescent protein GSEA - Gene set enrichment analyses HDR - Homology-directed repair HEK - Human embryonic kidney HGFR - Hepatocyte growth factor receptor HIV - Human immunodeficiency virus

x HL - Hodgkin lymphoma HLA - Human Leukocyte Antigen HRP - Horse radish peroxidase HRS - Hodgkin Reed-Sternberg IHC - Immunohistochemistry IL - Interleukin IPI - International Prognostic Index IPS - International Prognostic Score IRS - Insulin receptor substrate ITIM - Immunoreceptor tyrosine-based inhibitory motif JAK - Janus kinase KD - Knock down LCL - Lymphoblastoid cell line LGLL - Large granular lymphocytic leukemia LP - Lymphocyte-predominant MAPK - Mitogen-activated kinase MCL - Mantle cell lymphoma MHC - Major histocompatibility complex MM - Mutiple myeloma MPD - Myeloproliferative disorders MT - Empty vector NHL - Non-Hodgkin lymphoma NK - Natural killer cell NKCL - Natural killer cell lymphoma NKL - Natural killer cell lymphoma NLPHL - Nodular lymphocyte-predominant Hodgkin lymphoma NS - Non-silencing; not significant OS - Overall survival PAGE - Polyacrylamide gel electrophoresis PDGFR - Platelet-derived growth factor receptor PFS - Progression-free survival PI3K - Phosphoinositide 3-kinase PIAS - Protein inhibitor of STAT PIN - Paraffin isolated nuclei PLL - Prolymphocytic leukemia PMBCL - Primary mediastinal large B cell lymphoma PTCL - Peripheral T-Cell Lymphoma PTP - Protein tyrosine phosphatase RT - Room temperature SCID - Severe Combined Immunodeficiency SEAP - Secreted embryonic alkaline phosphatase SHM - Somatic hypermutation SLL - Small lymphocytic lymphoma SNV - Single nucleotide variant SNP - Single nucleotide polymorphism SOCS - Suppressors of cytokine signaling

xi STAT - Signal Transducer and Activator of Transcription TAM - Tumor-associated macrophage TARC - Thymus and activation regulated chemokine TCL - T cell lymphoma TCRBCL - T cell-rich B cell lymphoma TETR - Tetracycline repressor TLL - T-cell lymphomas/leukemia TMA - Tissue microarray TNF - Tumor necrosis factor TSCA - TrueSeq custom amplicon TYK - Tyrosine kinase VDJ - Variable-joining-diversity recombination VEGF - Vascular endothelial growth factor WGA - Whole-genome amplification WGS - Whole-genome sequencing WT - Wild-type WTSS - Whole-transcriptome sequencing

xii Acknowledgements

I would like to convey my sincere gratitude to everyone who helped me complete this work. Foremost, I express my gratefulness to my supervisor, Dr. Christian Steidl for his guidance, unwavering support, words of encouragement and providing me with the necessary tools and expertise to complete this project. I would also like to thank Dr. Randy D. Gascoyne for giving me the opportunity to grow as a scientist and for his mentorship.

My thanks are extended to members of my supervisory committee (Drs. Carolyn Brown, Blake Gilks and Marcel Bally) for their advice, fine direction and for keeping me focused on this project.

Much gratitude is expressed to both current and former lab members of Steidl and Gascoyne labs – particularly, Adèle Telenius, Bruce Woolcock, Elizabeth Chavez, Tessa Van Tol and Katina Mak; your skills and expertise made this work possible.

I would also like to acknowledge the following funding agencies: the University of British Columbia through the Roman M. Babicki fellowship in Medical Research. This work was supported by a research grant from the Leukemia & Lymphoma Society of Canada (LLSC), Canadian Institute of Health Research (CIHR) and by a Terry Fox Research Institute team grant.

Finally, my deepest appreciation to my family for their love and moral support.

xiii Dedication

I dedicate this thesis to my Mum: You started my journey but you didn’t get to see the end. I arrived.

xiv

Chapter 1: Introduction

1 1.1. B cell lymphomas

Lymphomas are a heterogenous group of cancers that originate from developing lymphocytes. They predominantly arise in lymph nodes and other lymphoid structures such as the spleen and bone marrow, but may also produce tumors in extranodal sites. The current World Health Organization (WHO) classification categorizes lymphomas into more than 35 subtypes, primarily based on morphology, clinical features and genomic alterations1. They are the fifth most common cancer in North America and according to the most recent Surveillance, Epidemiology and End Results (SEER) survey, lymphoid neoplasms account for an annual 22.5 incidences per 100,000 individuals2, over 90% of these patients suffer from lymphomas of B cell origin3. To fully grasp the pathogenesis of these lymphomas, an appreciation of normal B cell biology is required.

1.1.1. Normal B cell development

The bone marrow produces naïve B cells from a common lymphoid progenitor cell and is the site of conferring a remarkable ability to respond to an enormous repertoire of antigens with extremely high affinity and specificity. Highly diverse immunoglobulin (Ig) molecules that react to a virtually limitless number of antigens are generated by way of VDJ rearrangement of Ig genes4. Rearrangement of variable (V), diversity (D) and joining (J) gene segments is the primary mechanism for generating functional and unique B cell receptors (BCR) that have different affinities to various antigens. Lymphocyte-specific recombination activating genes 1 and 2 (RAG-1/2) enzymatically cleave between coding sequences to create double stranded DNA breaks which are then ligated by non-homologous end joining5. B cells exiting the bone marrow have functional BCRs which are required for B cell differentiation and survival.

1.1.2. B cell differentiation and the germinal centre reaction

B cells encounter antigen in the follicles of lymph nodes (or spleen) and is presented by professional antigen presenting cells (i.e. macrophages and dendritic cells)6. B cells present bound antigen to T helper cells via MHC class II and receive co-

2 stimulatory signals, and in response, migrate into the ‘dark zone’ of germinal centres (GC; Figure 1.1). Germinal centres play a crucial role in adaptive immune responses where affinity maturation, antigen-specific B cell proliferation, somatic hypermutation (SHM) and Ig class switching occurs7. Rapidly proliferating B cells diversify Ig genes by SHM and class switching, both processes mediated by the enzyme “activation-induced cytidine deaminase” (AID)8. In the ‘light zone’, mutated Ig that is the product of SHM is presented with antigens and selected by follicular dendritic or T helper cells. Reactive B cell clones may re-enter the ‘dark zone’ for further proliferation and SHM; clones with the highest affinity to the presented antigen exit the GC as antibody-secreting plasma cells or as memory B cells9. B cells with low affinity to the antigen will undergo apoptosis; failure of which is the etiology of many GC-derived lymphomas.

1.2. Primary mediastinal large B cell lymphoma

1.2.1. Epidemiology

Primary mediastinal large B cell lymphoma (PMBCL) accounts for 2-4% of non- Hodgkin lymphomas. This disease mainly affects young adults (mean age 35 years) and presents predominantantly in females (male to female ratio, 1:2)1.

1.2.2. Clinical presentation

PMBCL usually presents as a bulky tumor in the anterior mediastinum and may extend to lungs, chest wall, pleura and pericardium. It is an aggressive lymphoma that may disseminate to distant extranodal sites including kidneys, adrenals, liver, ovaries and the central nervous system. Bone marrow infiltration is less common10.

1.2.3. Pathology

PMBCL is derived from thymic medullary B cells. Tumor cells are medium to large in size with a “clear” cytoplasm and exhibit varying degrees of sclerosis (Figure 1.2a). They normally express the B cell antigens CD19, CD20, CD22 and CD79a; however, unlike other B cell lymphomas lack surface expression of Ig. CD30 is variably

3 expressed. The GC markers CD10, BCL6 and CD23 are expressed in most cases, indicative of its thymic B cell origin11.

1.2.4. Therapy

First-line treatment is critical to the management of PMBCL since the salvage rate following relapse is low with patients frequently succumbing to their disease12. Early relapse or primary refractory disease can still occur following standard CHOP (cyclophosphamide, doxorubicin, vincristine and prednisolone [CHOP])-like induction chemotherapy13. Combined modality treatments including rituximab (anti-CD20) immunotherapy have overall yielded quite favorable outcomes. However, the 3-year progression-free/event-free survival was variably reported between 68% and 91% for (rituximab + CHOP studies) and between 78% and 95% for (rituximab + dose intensive studies) likely due to selection bias of patients in retrospective series14-19. In aggregate, the optimal combination chemotherapy and the role that consolidative radiotherapy plays in treating PMBCL is yet to be determined20.

1.3. Hodgkin lymphoma

1.3.1. Epidemiology

Collectively, Hodgkin lymphoma (HL) accounts for 11% of all lymphomas and is one of the most frequent lymphomas in the Western world, with an annual incidence of 3 cases per 100,000 individuals21. It is classified as two disease entities – the more prevalent classical Hodgkin lymphoma (cHL) and nodular lymphocyte predominant Hodgkin lymphoma (NLPHL). Classical HL accounts for 95% of HL and affects mainly adolescents and young adults, but may also resurge later in life. It is further subclassified histologically as nodular sclerosis, mixed cellularity, lymphocyte-rich and lymphocyte-depleted cHL. NLPHL accounts for the remainder of HL and predominantly afflicts male patients in the 30-50 year age group1.

4 1.3.2. Clinical presentation

HL mostly commonly presents as localized, persistent, painless and firm lymphadenopathy, with frequent occurrences in cervical and supraclavicular nodes22. Mediastinal involvement is most frequently seen in the nodular sclerosis subtype; abdominal and splenic involvements are common in the mixed cellularity subtype1.

1.3.3. Pathology

Mononucleated Hodgkin cells and multinucleated Reed-Sternberg (HRS) cells are characteristic of cHL (Figure 1.2b). These cells are positive for CD30 in nearly all cases but the B cell-associated antigens CD20 and CD79a are only expressed at varied intensities or less frequently1. It is postulated that a proportion of HRS cells are derived from GC B cells that acquired unfavorable mutations (such as non-functional Ig V genes), but failed to undergo apoptosis21. The HRS cell-like variants in NLPHL, called lymphocyte predominant (LP) cells present as nodular and scattered large neoplastic cells. LP cells are positive for CD20 and CD79a in nearly all cases1.The gene expression signature of LP cells resembles an intermediate developmental stage between germinal centre and memory B cells23.

1.3.4. Therapy

Despite an overall cure rate of 80%, one third of cHL patients with advanced- stage disease and 15% with early-stage disease will relapse after initial treatment. Doxorubicin, bleomycin, vinblastine and dacarbazine (ABVD) continues to be the standard of care for patients with advanced-stage HL in North America24. Autologous hematopoietic stem-cell transplantation (ASCT) can rescue about 50% of patients in whom primary therapy has failed25. As the prognosis of patients who develop chemoresistance and those who fail ASCT is grim, novel therapies, including targeting of CD20-positive cells (Rituximab), CD30-positive cells (SGN-35) and PD-1-positive cells in the microenvironment (Nivolumab) are being evaluated26-27. Results of these studies are encouraging, however, the development of additional targeted therapies for de novo or relapsed disease is dependent on novel biomarker discovery.

5 1.4. Molecular aberrations in PMBCL and HL

Primary mediastinal large B cell lymphoma and the NLPHL subtype of Hodgkin lymphoma exhibit similar molecular features such as overlapping gene expression profiles as well as gains of large chromosomal segments. For instance, genes involved in tumor necrosis factor (TNF) and cytokine signaling, co-stimulatory and inflammatory mediation and cell adhesion are upregulated in both disease entities. Genes associated with B cell receptor signaling are downregulated28. Shared copy number gains and amplifications in regions of chromosomes 2p and 9p, and the associated JAK2 (9p24) and REL (2p16) loci have also been reported29-30. Further, increased activation of JAnus Kinase–Signal Transducers and Activators of Transcription (JAK-STAT) and nuclear factor (NF)-κB signaling has also been reported in both lymphoma subtypes, both providing malignant cells with a proliferative advantage31. However, the full spectrum of genetic changes underlying the malignant phenotypes of both diseases still remains to be uncovered.

1.5. JAK-STAT signaling in cancer

Cancer is a disease of cell signaling gone awry. The breakdown of inter- and intracellular signaling has been identified as a hallmark in cancer pathogenesis32. Several cancers, including hematological malignancies show frequent dysregulated JAK – STAT signaling that contributes to diverse oncogenic phenotypes such as oncogene activation, tumor suppressor de-activation, abnormal cell proliferation, tumor growth and metastasis. The JAK-STAT pathway was first identified as a paradigm of how protein interactions at the cell surface relayed information through a signal cascade that regulated the expression of genes in the nucleus. In 1957, Isaacs and Lindenmann observed a phenomenon where healthy chick embryo cells were resistant to influenza viral infections as a result of small proteins secreted by infected cells, which they termed “viral interference” and coined the term “interferon” to describe these small proteins33. The discovery of interferons laid the foundation that paved the way to elucidating the signaling cascade that was responsible for the secretion of these small proteins. Since

6 the initial identification of JAK1, JAK2 and TYK2 twenty five years ago, other JAK and STAT members have been identified and characterized including genomic sequencing, functional and structural domains, crystal structures and comprehensive gene target sets34. In more recent years, studies have not only directly implicated the activated JAK- STAT pathway in oncogenesis of several cancer types, but also as an enhancer of other spatially distinct oncogenic signaling pathways such as mitogen-activated kinase (MAPK), insulin receptor substrate (IRS) and phosphoinositide 3-kinase (PI3K) pathways. Cytokines are small pleitropic glycoproteins secreted by hematopoietic and non- hematopoietic cells that bind to their corresponding receptors. However, these receptors lack intrinsic kinase activity and are unable to facilitate downstream signaling. Therefore, JAK and STAT members act as intermediates to convey information to target cells to regulate diverse cellular processes from cell growth to immune modulation. The JAK-STAT pathway consists of interactions from 3 components: extracellular cytokine receptors and intracellular JAK and STAT proteins. The pathway is activated once a cytokine or a growth factor binds to its cognate receptor on the cell surface which causes tyrosine phosphorylation of JAK that is constitutively bound to the receptor. JAKs will trans-phosphorylate each other and tyrosine residues on the receptor. These phosphorylated sites will act as “docking sites” for the Src-homology 2 (SH2) domain of STAT proteins. STATs are also phosphorylated by JAK kinase activity and subsequently, STATs will form homo- or heterodimers before translocating to the nucleus to activate corresponding target genes. STAT dimers recognize a DNA consensus sequence (referred as a GAS element) in target genes and upon binding leads to transcriptional activation35-36. Four JAK (JAK1, JAK2, JAK3 and tyrosine kinase 2 (TYK2)) and six STAT (STAT1, STAT2, STAT3, STAT4, STAT5 (STAT5a and STAT5b isoforms) and STAT6) have been identified in humans. Despite some members of the pathway often assuming redundant roles, JAK-STAT is a highly evolutionary conserved signaling cascade that ensures homeostasis in all animals, from mammals to fruit flies37. The in vivo importance of this pathway has been highlighted by knockout murine models where silencing JAK-STAT expression led to impaired immunological functions, defects in hematopoietic and lymphoid development and even early

7 lethality38. JAK1 null mice die perinatally, JAK2 null mice are embryonically lethal due to failed erythropoiesis, JAK3 null mice exhibit severe combined immune deficiency (SCID) due to cytokine signaling defects and TYK2-deficient mice suffer from hypersensitivity to pathogens. Lack of STAT1 and STAT2 causes impaired interferon signaling, STAT3 null mice are embryonically lethal, STAT4 is needed for Th1 cell differentiation, STAT5 null mice have no natural killer (NK) cells, show defects in growth hormone pathways and mammary gland development, and STAT6 knock out mice show diminished Th2 type responses, including a pronounced decrease in interleukin (IL) -4/IL-13 signaling36. Since activated STAT proteins accumulate in the cell nucleus to drive target gene transcription, it is critical to tightly regulate the duration and strength of activation in order to “dampen” the pathway when not needed. Termination of JAK-STAT signaling is achieved by 3 types of negative regulators ((protein inhibitors of activated STATs (PIAS), suppressors of cytokine signaling (SOCS) and protein tyrosine phosphatases (PTPs)) and post-translational modifications39. While JAK and STAT proteins can be post-translationally regulated via phosphorylation, methylation, acetylation, ubiquitylation, ISGylation and sumoylation40-41, for the purpose of reviewing the regulation aspects of the JAK-STAT pathway our analysis will primarily focus on PTP and SOCS.

1.5.1. JAK-STAT signaling aberrations are common in lymphoid cancers

Recent insights into lymphoma biology have identified disordered transcriptional and epigenetic regulation, immune evasion and intracellular pathway dysregulation, including altered NF-ĸB, B cell receptor and JAK-STAT signaling42. While population- based explorative studies have examined associations between JAK-STAT pathway related genes and the risk of developing lymphoma43-44, larger clinical studies that comprehensively interrogate genomic, epigenetic and proteomic alterations in multiple genes and gene products of this pathway remains to be explored. In sections 1.5.3- 1.5.4 recent progresses in JAK-STAT dysregulation in selected lymphoid malignancies will be highlighted, with a focus on genomic alterations that contribute to

8 cancerogenesis (Summerized in Table 1.1). In this section, evidence of the tumor microenvironment providing external stimuli for JAK-STAT activation, phosphorylation of STAT proteins in response to those stimuli and downstream activation of oncogenes implicated in lymphomagenesis will be discussed. Moreover, the genetic basis for aberrant JAK-STAT by somatic gene mutations of potential oncogenes and tumor suppressors described by next-generation and traditional sequencing methods will be reviewed in Section 1.5.5.

1.5.2. Changes to the tumor microenvironment

Lymphoma cells depend on the microenvironment for survival and growth signals. For instance in classical Hodgkin Lymphoma (cHL), the malignant Hodgkin Reed Sternberg (HRS) cells in lymphoma tissue are vastly outnumbered by non- malignant reactive cells composed of lymphocytes, eosinophils, plasma cells, macrophages and fibroblasts. These cells are either recruited or induced by neoplastic cells, and form the inflammatory infiltrate that sculpts the surrounding tumor microenvironment by releasing a variety of cytokines and chemokines that are advantageous to HRS cells to sustain growth and survival45. Both T helper type 1 (Th1) and type 2 (Th2) cytokines have been reported to be expressed in primary HRS cells. Both IL-4 and IL-13 are key regulators of the humoral immune response responsible for proliferation of activated B cells, B cell differentiation and isotype class switching to immunoglobulin (Ig) E46. While expression of IL-4 in HRS cells from primary biopsy tissue is uncommon, IL-13 expression studies using in-situ hybridization and immunohistochemistry suggest that this cytokine is a key autocrine growth factor almost exclusively expressed by HRS cells (expressed in 86% of cHL cases studied) and rarely by reactive cells in the microenvironment. Moreover, antibody-mediated neutralization of IL-13 in HRS-derived cell lines led to inhibition of cell proliferation and induction of apoptosis in a dose-dependent manner. Other cytokines such as IL-5, IL-6, IL-9, IL-10, IL-12 and IFN-γ are also reported to be expressed at varying levels in both primary cHL and cell lines however, not all have been shown to consistently act as autocrine growth factors47. In PMBCL, increased abundance of IL13Rα have been found by gene

9 expression profiling28 and interestingly, stimulation of a PMBCL cell line with IL-13 or IL- 4 induced an inflammatory phonotype reminiscent of HL48. Two N-HL B cell lymphomas with growing evidence of the tumor microenvironment contributing to pathogenesis are follicular lymphoma (FL) and Diffuse Large B-Cell Lymphoma (DLBCL). Multi-panel cytokine measurements in pre-treatment serum samples matched to normal controls have shown elevations in IL-2, IL-6, IL-8 and IL-10 levels in DLBCL49. IL-10 increases were functionally related to activation of JAK2 in vitro and correlated with inferior outcome in patients50. IL-3 secretion is of particular interest in FL where it promoted the proliferation of tumor cells upon engagement of IL-3R, the cognate receptor detected on fresh isolates of FL tumor cells51. T-cells are often dependent on endogenous or exogenous cytokines for their growth and these cytokines may play a role in the sustenance of some T cell lymphomas (TCL). Elevated IL-6 levels have been found in Lennert’s lymphoma, in skin exudates of Cutaneous T cell lymphoma (Mycosis fungoides) and in Angioimmunoblastic lymphadenopathy bioposies where TNFα and lymphotoxin were also present52-53.

1.5.3. Phosphorylation of STAT proteins

STAT1 The presence of tumor-associated macrophages (TAM) is an important biomarker that predicts inferior survival outcomes in FL and in cHL patients54-55, and these inflammatory cells are able to modulate immune responses against tumor cells in order to promote tumor progression. STAT1 is a key transcription factor downstream of IFN signaling and mediates tumor immunogenicity56. Alvaro et al. observed overexpression of STAT1 in 30% of 211 FL diagnostic biopsies by immunohistochemical (IHC) staining. Concurrent double-immunofluorescence staining confirmed that activated STAT1 co-localized exclusively with CD68-positive TAMs57, consistent with the involvement of post-translational abnormalities of STAT proteins in creating an inflammatory milieu. Furthermore, STAT1 has been reported to influence the differentiation of Chronic Lymphocytic Leukemia (CLL)58 and act as a tumor promoter in B-lymphoid malignancies induced by the v-abl-oncogene59. However,

10 the role STAT1 plays in DLBCL is less clear; the oncogene B-aggressive lymphoma-1 (BAL1) can induce a “switch” in STAT1 from an oncogene to a tumor suppressor gene60.

STAT3 STAT3 is probably the best described phosphorylated STAT protein that has been shown to be hyperactivated in many cancers inclusive of lymphoid cancers61. STAT3 activation has been described in 14 of 20 (70%) of Mantle Cell Lymphoma (MCL) tumors as a result of autocrine IL-6 and IL-10 signaling62. Stimulation with recombinant IL-10 resulted in rapid tyrosine phosphorylation, but not serine phosphorylation of STAT3 in four MCL cell lines studied63. Consistent with these findings, IL-10R is overexpressed in some MCLs64 and addition of IL-10 increased proliferation of cultured MCL cells65. A substantial proportion of DLBCL tumors also show STAT3 activation. Twenty one of 52 (40%) formalin-fixed paraffin embedded tissue (FFPET) specimens were positive for phospho (p)-STAT3 expression by IHC, with a preferential association with tumors of the activated B cell-like subgroup (ABC-DLBCL)66. Subsequent studies have found unfavorable overall and event-free survival rates in DLBCL patients (69 of 185 cases; 37%) whose tumors showed strong nuclear p-STAT3 staining67. Constitutive activation of STAT3 has been shown in 5 of 7 HL cell lines, but was not detected in Burkitt lymphoma (BL) cell lines possibly due to the JAK-STAT pathway playing a less prominent role in the pathogenesis of the latter disease68. STAT3 activation in Anaplastic Large Cell Lymphoma (ALCL) and TCL is induced by the deregulated expression of the Anaplastic Lymphoma Kinase (ALK) gene. The resulting chimeric protein formed by chromosomal translocations between ALK and nucleophosmin (NPM) has been shown to drive STAT3 phosphorylation in all ALK+ ALCL (21 of 21) and ALK+ TCL (10 of 10) cases examined. Furthermore, ALK-mediated STAT3 expression contributed to the induction of the anti-apoptotic gene, Bcl-xl and in- vitro out growth of tumor cells69-70.

STAT5 Phosphorylated STAT5 in primary HL tumors has been interrogated in two independent

11 study cohorts. Martini et al. reported p-STAT5 expression in 35 of 93 (38%) lymph node biopsies of HL patients while Hinz et al. observed phosphorylation in all samples (24 of 24; 100%) studied with >80% staining attributed to HRS cells in the tumor section71-72. This discrepancy in expression may partially be due to two reasons: (1) the former study included both HRS and reactive lymphocytes in their analysis and (2) the latter study used an antibody against the STAT5a isoform only. Ectopic expression of constitutively active STAT5 in human primary B cells has provided insight into oncogenic transformation associated with HL genesis. Forced expression resulted in phenotypes that closely resembled HL tumors: extension of replicative lifespan, cytokine- independent cell growth, down-regulation of the B cell receptor (BCR), presence of mutinuclated cells, expression of HL-specific genes such as IPL, FER, PRAME and RAB13 and unexpectedly, activation of NF-κB73. The aberrant activation of p-STAT5 has also been shown in skin lesions of CTCL patients74, however the pathological relevance of these findings is evasive due to low sample numbers and the lack of functional studies.

STAT6 Phosphorylation-induced activation of STAT6 has been described in both HL and PMBCL. Strong activation was seen in 5 HL cell lines without the need for cytokine stimulation, but was noticeably absent in DLBCL and ALCL cell lines. Phospho-STAT6 was also detected in nuclei of HRS cells in 25 of 32 (78%) HL cases, in 2 of 6 (33%) ALCL cases and in 2 of 5 (40%) T cell-rich B cell lymphoma (TCRBCL) cases by IHC. Of note, no p-STAT6 staining was seen in any of the NLPHL, DLBCL or TCL cases examined47. Nuclear phosphorylation of STAT6 has also been reported in 8 of 11 (73%) PMBCL tumors and in two PMBCL-derived cell lines, highlighting similar molecular features shared between HL and PMBCL75.

1.5.4. STAT-regulated genes in oncogenesis

In normal cells as with most signal transduction pathways, JAK-STAT is tightly controlled to prevent unscheduled gene regulation and inappropriate biological

12 responses. In malignant cells, this system of checks and balances is modified leading to constitutive STAT-dependent gene expression, including several key gene products required to initiate and/or maintain malignant transformation76. Target genes activated by STATs and implicated in oncogenesis can be broadly grouped as genes associated with B cell differentiation, cell cycle progression/apoptosis and angiogenesis.

STAT regulation of B cell differentiation

B cell lymphoma (BCL6) functions as a master regulator of the germinal centre reaction and plays a transcriptional suppressive role in attenuating DNA damage sensing genes (such as ATR) and inhibiting cell cycle check point genes (such as CDKN1C, TP53, CHEK1) to enable B cell proliferation and survival77. The resulting proliferative environment is conducive for the pathogenesis of GC-derived lymphomas. Not surprisingly, 50% of DLBCL, 12% BL and 10% of FL cases show dysregulated BCL6 activity owing to mutations, transcriptional regulation, epigenetics, chromosomal translocations and amplifications78-81. High levels of STAT3 expression and activation seen in ABC-DLBCL has been shown to be mediated by BCL6 by direct transcriptional targeting66. Chromatin immunoprecipitation assays have revealed STAT3 binding sites within the BCL6 gene that showed enhanced recruitment of RNA polymerase II to upregulate BCL6 expression82. Introduction of constitutively active STAT5 in primary B- cells increased BCL6 mRNA expression and inhibited terminal differentiation into plasma cells but in lieu, showed differentiation into memory B-cells with enhanced self renewal and replicative properties83. On the contrary, evidence of STAT5 playing a reciprocal role to repress BCL6 has also been reported however, the regulatory regions in STAT5 that BCL6 binds to are frequently mutated in lymphomas thus, BCL6 overexpression remains unabated in a disease context84. The association of STAT6 in BCL6 repression was recently reported in PMBCL. Double-fluorescence staining showed the presence of distinctive pSTAT6+ and BCL6+ cell clusters within PMBCL tumors indicative of a negative interaction between the two subpopulations. Depletion of STAT6 by siRNA and forced expression of an activating STAT6 mutant in PMBCL- derived cell lines identified BCL6 as a target of STAT6-mediated repression85.

13 Furthermore, concomitant knock down of both BCL6 and STAT6 sensitized PMBCL cell lines to immuno-chemotherapeutic agents86. Moreover, BCL6 has been shown to downmodulate the key plasmacytic differentiating transcriptional factor B-lymphocyte induced maturation protein-1 (BLIMP1; encoded by PRDM1), resulting in stalled differentiation of mature B-cells at the plasmablast stage87. BLIMP-1 is a reported tumor suppressor gene in B-, T- and NK-cell lymphomas, underscoring another aspect of BCL6-mediated dysregulation during B-cell maturation88-90. Regulation of Blimp1 has been shown by several groups to be STAT3-mediated adding yet another layer of complexity to the disordered B-cell differentiation program in lymphomas91-93.

STAT regulation of cell cycle progression and apoptosis

One of the fundamental traits of cancer cells is the ability to sustain chronic growth independent of growth factors. Neoplastic cells foster proliferation by resisting physiological signals of cell death, growth suppressors and normal cell cycle progression94. Activated STAT proteins contribute to unconstrained cell growth by inducing several key genes that have been implicated in playing a direct role in malignant transformation. Cyclin and cyclin-dependent kinases (CDK) regulate the progression through the cell cycle in response to mitogenic signals by forming serine/threonine kinase holoenzyme complexes. The D-cyclins positively regulate cell proliferation by binding to CDK4 and CDK6, resulting in transition of cells through gap phase-1 (G1) and entry in to synthesis (S) phase of the cell cycle95. Cyclins are upregulated in B cell lymphomas by 2-4 orders of magnitude compared to normal mature B cells96. Cyclin D2 expression was found to be increased in 64 of 89 (72%) cHL cases and in the proliferation centers of 19/19 (100%) CLL/SLL cases by IHC analysis97-98. Detection of activated STAT5 in HRS cell nuclei and the identification of a STAT5 binding site in the promoter region of cyclin D2 strongly suggest a STAT-mediated regulatory mechanism that deregulates cell cycle progression71,99. Indeed, a positive correlation between STAT3 activation and cyclin D3 was observed in ALK+ ALCL tumors100.

14 Survivin, an inhibitor of apoptosis, plays a major role in promoting drug resistance by directly inhibiting caspase signaling mechanisms leading to activation of multidrug resistance pathways101. It has been shown to be regulated by STAT3 in CTCL and selective inhibition of activated STAT3 (pSTAT3) decreased Survivin protein expression and induced apoptosis in a caspase-3 dependent manner102. STAT3 activation by IL21 stimulation of primary Sézary cells resulted in expression of the proto- oncogene MYC, nonetheless the proliferative consequences of this upregulation remains to be elucidated103. In a B cell lymphoma context, small molecule inhibitors that interfere with dimerization abrogated STAT3 activation, suppressed cell growth and down regulated c-myc expression104. B cell CLL/Lymphoma-2 (BCL2) family proteins regulate apoptosis by governing mitochondrial outermembrane permeabilization and caspase release (e.g. pro-apoptotic Bax, Bak, BAD or anti-apoptotic Bcl-2, Bcl-xl, Mcl1). Dysregulated Bcl-2 expression as a result of chromosomal translocations, somatic hypermutation, gene amplification and epigenetic regulation is well documented in several lymphoid malignancies105. A mechanistic link to STAT protein-mediated regulation of members of the Bcl-2 family has been described in several studies. CTCL cells incubated with IL7 and IL15 exhibit high levels of Bcl-2 that coincide with increased binding of STAT2, STAT5 and STAT6 to Bcl-2 gene promoter-enhancer elements106. STAT3-specific and multi-kinase inhibitors against JAK2, STAT3 and STAT5 were shown to sensitize BL cells to chemotherapeutic agents and HL cell lines to dose-dependent cell growth inhibition that correlated with inhibition of Bcl-xl expression107-109. Morever, targeting activated STAT3 by tyrosine inhibitors or antisense oligonucleotides significantly reduced Mcl-1 protein expression and restored Fas-mediated apoptosis in LGL cell lines and patient samples110.

STAT involvement in angiogenesis

While prolonged cell cycling, uninhibited cell growth and resistance to apoptosis are advantageous to malignant cell survival, the formation of new blood vessels and capillaries by angiogenesis are critical in order to sustain tumor growth. The most potent

15 family of angiogenesis-inducing vascular endothelial growth factors (VEGF) has been shown to mediate lymphangiogenesis through interactions with their receptors to form a tumor neovasculature111. The presence of VEGF expression has been shown in PTCL, DLBCL, MCL, HL, CLL but only sparse or variable expression was observed in FL112. STAT3 is able to directly induce VEGF gene expression. Enforced expression of an activating mutant of STAT3 in mice has demonstrated upregulation of VEGF and the formation of tumor vascular capillaries, a phenotype that was abrogated when STAT3 binding sites in the VEGF promoter were mutated113. Similar in vivo studies performed using the human BL cell line Daudi have reported increased microvessel densities and tumor cell proliferation and upregulation of phosphorylated STAT4 and STAT6 in tumors over-expressing VEGF compared to non-expressing tumors114. Hypoxia-inducible factor 1 (HIF1), a transcription factor that regulates intra-tumoral oxygen homeostasis has been suggested as a switch for VEGF upregulation in cancers and interestingly, also reported to be activated in lymphoma cell lines115-116.

1.5.5. Somatic gene mutations in the JAK-STAT pathway

Aberrant activation of JAK-STAT may be a result of autocrine and/or paracrine cytokine signaling mechanisms or due to somatic gene mutations in members of the pathway. In the following sections the genetic basis of JAK-STAT activation by gain-of- function oncogenes or loss-of-function tumor suppressors will be discussed (Table 1.2 and Figure 1.3).

Activation of oncogenes: Cytokine receptors

IL-7R The interleukin receptor (IL-7R) complex consists of two cytokine receptor chains, IL- 7Rα and γc. IL-7Rα is expressed on hematopoietic cells, especially of the lymphoid lineage and is essential for the development and survival of T lymphocytes117. IL-7Rα is transcriptionally upregulated by the NOTCH1 signaling pathway in T-ALL and signaling mediated by IL-7 and IL-7Rα has been shown to directly activate STAT1 and STAT5118-

16 119. Gain-of-function IL-7Rα mutations have been found in 3 independent ALL study cohorts with a recurrence rate of 6-9% (7 of 106 T-ALL120, 8 of 133 B-ALL121 and 17 of 201122 T-ALL cases) (Figure 1.4). Collectively, these studies have identified two striking hot spot regions in IL-7Rα, a replacement of a serine with a cysteine residue at amino acid position 185 in the extracellular domain and introduction of an in-frame insertion or deletion at residues 241-253 in the intracellular domain. Functional validation of these mutants in both in vitro and in vivo models have shown ligand-independent homodimerization, enhanced cell growth and transformation, induced phosphorylation of JAK1, STAT3 and STAT5, and increased tumor formation in mice with infiltration into liver, spleen and bone marrow.

Activation of oncogenes: JAKs

JAK1 JAK1 missense mutations have been described in pediatric and adult ALL and T-cell prolymphocytic leukemia (T-PLL)123-125. Mutations in ALL occurred as heterozygous changes affecting conserved amino acids within the founding members protein 4.1, ezrin, radixin, moesin (FERM), SH2-like, pseudokinase (JH2) and kinase (JH1) domains and occurred more frequently in adult T-ALL (adult T-ALL 18.4%, adult B-ALL 3.4%, pediatric T-ALL 2% and no mutations were found in pediatric B-ALL). Analysis of survival outcomes in adult patients who harbor mutations showed inferior DFS and OS compared to patients with no mutations. Frequent amino acid substitutions at Arg 879 in the JH1 domain may augment catalytic activity of the kinase. Indeed, expressing JAK1 mutants in JAK1-defective human fibrosarcoma cells resulted in STAT1 activation and IL-3 independent growth. Of note, the non-synonymous mutations found in T-PLL were limited to the JH2 domain.

JAK2 Recurrent and somatic activating JAK2 V617F mutations in the JH2 pseudokinase domain have been recorded in myeloproliferative disorders (MPD), but not in lymphoid malignancies126-127. Other JAK2 missense mutations reported in ALL (9%; 16 of 187

17 cases), including in the highly conserved R683 residue of the JH2 domain have been shown to activate JAK-STAT and transform Ba/F3 cells in vitro123. Alternately, constitutive activation of JAK2 can be achieved by genomic translocations resulting in chimeric oncoproteins (Figure 1.5). Several JAK2 fusion partners have been identified in ALL: ETS-Variant 6 (ETV6)128, B-cell lineage specific paired box 5 (PAX5)129, single stranded DNA binding protein 2 (SSBP2)130 and the human autoantigen pericentriolar material (PCM1)131. It should be noted that not all chimeric fusions have been functionally characterized and with the exception of PAX5, are not limited to lymphoid malignancies indicating no lineage-specific occurrences. Nevertheless, these partner genes juxtaposes themselves at or just upstream of the JAK2 JH2 domain facilitating dimerization of the adjacent JH1 kinase domain with consequent activation132. JAK2 rearrangements involving the vesicular gene SEC31A found in cHL have been functionally shown to phosphorylate JAK2, STAT3 and STAT5 and promote growth factor-independent proliferation, a trait abrogated by JAK inhibitors. Moreover, JAK2- SEC31A bone marrow transplanted mice developed lymphoblastic lymphoma and myeloid phenotypes133. JAK2 genomic copy number gains at chromosome 9p24.1 are characteristic of HL and PMBCL and induce cell proliferation via JAK2/STAT signaling30,134. Furthermore, mice challenged with homologous cell lines bearing amplified JAK2 and treated with JAK2 inhibitors exhibited decreased tumor growth and intratumoral p-STAT3 levels135. Albeit the attenuation of tumorgenesis seen in vivo, the precise mechanism of JAK2 activation as a direct result of copy number aberrations remains unclear.

JAK3 JAK3 coding sequence mutations have been found in adult-TLL, T-ALL, PLL, CTCL and NKCL125,136-139. Recurrent hotspot mutations involving M511I, A572V and A573V substitutions have been reported in these disease entities affecting the JH2 pseudokinase domain that is known to have an inhibitory effect on the JH1 kinase domain, presumably losing its ability to “turn off” catalytic activity. The latter two mutations have been characterized as activating in in vitro and in vivo systems. Expression of these mutants in NKCL cell lines resulted in IL-2-independent proliferation

18 and autophosphorylation of JAK3 and STAT5, a significant transforming event since NK cells require IL2 for proliferation and activation136. A similar phosphorylation phenotype accompanied by IL3 growth independence was seen when gain-of-function JAK3 mutations affecting the FERM domain were expressed in ATLL139. Introduction of the JAK3 A572V mutant in a murine bone marrow transplantation model resulted in fatal lymphoproliferative disease137.

TYK2 Sanda, T. et al. recently reported non-synonymous TYK2 mutations in 4 of 17 (21%) T- ALL cell lines that substituted amino acid residues with in the FERM, JH2 and kinase domains of the TYK2 protein. Expression of these mutants in murine pro-B Ba/F3 cells have demonstrated a gain-of-function role with increased phosphorylation of TYK2 and STAT1. These cells showed TYK2-STAT1-BCL2 pathway dependence that promotes cytokine-independent growth. However, the somatic nature of these mutations is unknown, particularly since no mutations were found in 45 primary pediatric T-ALL samples screened in this study. Therefore, the role TYK2 mutations play in pathogenesis of T-ALL remains obscure140.

Activation of oncogenes: STATs

STAT3 Massively parallel whole exome and RNA sequencing have revealed STAT3 heterozygous missense mutations in B, T, and NK-cell lymphomas. Mutations have been found in 9% (5 of 55) of DLBCL cases141 and no mutations reported in other B-cell lymphomas142. In contrast, STAT3 mutations are more prevalent in lymphomas of T-cell origin that includes angioimmunoblastic TCL (4/85; 5%)143, ALCL (2/14; 14%)144 and LGLL (31/77; 40%)145. The STAT3 mutation rate in NKL is reported to be 30% (15/50)146. Despite the wide frequency at which mutations occur in these disease entities, all reported lesions affect codons 614 – 663 and conspicuously aggregate in the SH2 domain of the protein. Protein structural models have predicted these mutations to increase the hydrophobicity of the STAT3 SH2 dimerization surface and

19 the stability of STAT3 homo- and heterodimers causing constitutive activation. Indeed, activation and nuclear translocation of STAT3 have been detected in LGLL patients harboring Y640F and D661V hotspot mutations. Moreover, RNA expression analysis of these patients showed upregulation (STAT1, STAT2, BCL2L1) and downregulation (SOCS1) of STAT3 responsive genes145. In vivo delivery of the Y640F gain-of-function mutant in a bone marrow transplantation model results in malignant hematopoiesis in injected mice147.

STAT5 Mutations in this gene have only been discovered in LGL with a mutational frequency of 2% (4 of 211 cases) and were exclusively found in STAT5b. The two missense mutations Y665F and N642H located in the SH2 domain of STAT5b have been shown to induce STAT5 phosphorylation in primary tumors and increase STAT5 transcriptional activity in transfected cell lines148. Constitutively active forms of STAT5a detected by PCR-mediated mutagenesis in the DNA binding (H299R) and transactivation (S711F) domains of the protein have been shown to be hyperphosphorylated and are capable of sustaining factor-independent growth in Ba/F3 cells that require IL3 for growth, coinciding with increased expression of the anti-apoptotic genes Bcl-xl and Pim1149-150. Introduction of the latter mutant into transgenic mice resulted in massive infiltration of immature B cells into the thymus, spleen and lymph nodes suggesting an aberrant differentiation arrest as a result of deregulated signaling. Constitutively active S711F generated in p53-null mice exhibited early onset of tumor development suggesting a concerted role between STAT5 and p53 in B-cell lymphomagenesis151.

STAT6 Heterozygous point mutations in STAT6 have been reported in 36% (20 of 55) of PMBCL and in 23% (9 of 39) of FL cases, but no mutations were found in 25 DLBCL tumors examined152-153. These mutations clustered on the DNA binding domain of the protein and as expected, transfected mutants showed decreased DNA binding to GAS- driven luciferase reporter constructs. Despite diminished DNA binding of STAT6 mutants seen in cell line model systems, differential expression of STAT6-responsive

20 genes (FCER2, IL4, IL4R, CCL17, CD40, BCL2L1 and IL4I1) was not seen between PMBCL tumors harboring STAT6 mutations and those with no mutations, likely due to modest sample numbers used in each group. However, STAT6-silenced PMBCL- derived cell lines show decreased Bcl-xl expression and increased cell proliferation and survival154.

Inactivation of tumor suppressor genes: Protein tyrosine phosphatases

PTPN2 (TC-PTP) Bi-allelic and mono-allelic deletions in PTPN2 have been reported in 6% (5/90) T-ALL tumors, exclusively in the oncogenic subtype that harbors deregulated expression of the homeobox transcription factor TLX1. Knocking down PTPN2 gene expression by RNA interference resulted in increased cell proliferation attributed to upregulation of the oncogenic tyrosine kinase fusion protein NUP214-ABL1 and elevated activation of its downstream mediators STAT3 and STAT5. Furthermore, loss of PTPN2 adversely affects treatment responses to JAK1 inhibitors and ABL1 kinase inhibitors in pro-B cell and T-ALL cell lines, respectively155-156. Missense, nonsense and focal deletions have also been found in 3% (2/69) of TCL cases and in the HL cell line SUP-HD1 resulting in bi-allelic inactivation of PTPN2 and constitutive JAK-STAT signaling157. Differential expression of TC-PTP and its nuclear substrate phospho-STAT6 has been described in ABC vs. GCB DLBCL, presumably due to subtype-specific reliance on the JAK-STAT pathway. However, no mutations of this gene have been reported in either subtype158.

PTPRC (CD45) The lymphocyte common antigen (CD45) is encoded by PTPRC and plays an essential role in B and T antigen receptor signaling159. Missense and nonsense mutations of this gene have been detected in 9% (6 of 65) of T-ALL cases and causes loss of CD45 phosphatase activity160. Knocking down gene expression in homologous cell lines activated STAT3 and STAT5 proteins and caused an increase in cell proliferation, an effect that was potentiated by the expression of an activating JAK1 mutant.

21 Inactivation of tumor suppressor genes: Suppressor of Cytokine Signaling (SOCS)

SOCS1 The members of the SOCS protein family share a common SH2 domain that contains an upstream kinase inhibitory region that has been shown to bind the tyrosine phosphate activation loop of JAK proteins in order to attenuate kinase activity161. The C- terminal SOCS box interacts with elongins BC and Cullin5 to commit SOCS proteins and their substrates to proteosomal degradation162. SOCS1 is the best characterized of the eight members in a lymphoma setting. Mutations in the SOCS1 gene spanning all domains that consists of deletions, insertions, missense and early stop codons leading to premature peptide abort have been reported in B cell lymphomas. The occurrence of mutations in PMBCL specimens is 45% (9/20) and bi-allelic deletions concurrent with delayed degradation of de novo JAK2, hyperphosphorylation of JAK2 and STAT5 in PMBCL-derived cell lines have also been reported. Furthermore, restoration of wild type SOCS1 in these lymphoma cell lines repressed CCND1, induced RB1 and activated caspase-3 indicative of an increase in the apoptotic cell fraction154,163-164. Mutations detected in 8 of 19 (42%) microdissected HRS cells were associated with nuclear accumulation of phosphorylated STAT5165. Similar mutation frequencies have been detected in the NLPHL subtype (6 of 12 cases; 50%) that coincided with increased JAK2 and STAT6 phosphorylation166. Lower recurrence rates of inactivating mutations have been reported in DLBCL (7/26; 27%), FL (4/15; 27%), BL (1/14; 7%) and MCL (1/15, 6%). The restriction of SOCS1 mutations to GC-derived lymphomas is strongly suggestive of somatic hypermutation as a source of these aberrations167. Of in vivo relevance, SOCS1-deficient mice develop neonatal lymphoid atrophy and severe lymphopenia with a fatal increase in IFNγ signaling that is STAT1-dependent168.

22 1.6. Thesis theme and objectives

As reviewed in section 1.5, the JAK-STAT pathway is frequently dysregulated in lymphoid cancers as a result of activating genetic lesions, loss of negative regulators or extrinsic stimulatory signaling from the tumor microenvironment. Importantly, constitutive signaling through the JAK-STAT pathway induces oncogenic changes that lead to aberrant B cell differentiation, cell growth and angiogenesis. Pivotal molecular studies based on gene expression profiling and gene mutational analyses have revealed recurrent and somatic mutations, and have clearly implicated the JAK-STAT signaling pathway in the pathogenesis of several lymphoma subtypes, including PMBCL and HL. The observation that some patients may harbor multiple lesions in this pathway underscores the oncogenic dependence on JAK-STAT and a strong need for therapeutic intervention. A future task will be to develop new agents that target specific aberrations in this pathway and evaluate their efficacy in individuals with an activated JAK-STAT “signature”. However, development of targeted therapies is impeded by the dearth of knowledge about the full spectrum of genetic changes underlying the malignant phenotypes of both PMBCL and HL. For instance, somatic mutations in upstream modulators of the pathway have only been reported in IL7-R120-122 and no mutations in cytokine receptor genes have been found in PMBCL or HL. Moreover, of 37 classical protein tyrosine phosphatase genes encoded in the human genome169, somatic mutations have only been described in PTPN2155 and PTPRC160; no mutations have been reported in B cell lymphomas. Therefore, there is a need to investigate the mutational landscape in the lymphoma genomes and mutation-associated phenotypes. The objective of this thesis work is to interrogate lymphoma genomes for somatic mutations in the JAK-STAT pathway and functionally characterize two novel mutations (PTPN1 and IL4R) discovered by next-generation sequencing methods.

23 1.7. Hypotheses

1.7.1. Hypothesis 1

Inactivating mutations in PTPN1 contribute to the pathogenesis of PMBCL and HL through activation of the JAK-STAT signaling pathway.

1.7.2. Hypothesis 2

Constitutively active JAK-STAT signaling observed in PMBCL is in part due to gain-of- function IL4R mutations.

1.8. Aims and thesis outline

The thesis consists of two research chapters that address both hypotheses outlined above.

1.8.1. Aim 1: To determine the frequency and the biological significance of PTPN1 mutations in PMBCL and HL

Chapter 2 describes the incidence of PTPN1 mutations in PMBCL and HL and demonstrates that these mutations lead to reduced phosphatase activity and increased phosphorylation of JAK-STAT pathway members. It also shows that transcriptional silencing of PTPN1 results in overexpression of the oncogenes BCL6 and MYC.

1.8.2. Aim 2: To determine the frequency and the biological significance of IL4R mutations in PMBCL

Chapter 3 describes the incidence of IL4R mutations in PMBCL and demonstrates that the recurrent hotspot mutation leads to hyperphosphorylation of JAK-STAT pathway members and activation of the T cell regulatory chemokine TARC and the B cell

24 activation marker CD23.

Gene Proposed Mutation or Disease References (Protein) function Phenotype

Functionally validated somatic gene mutations

IL7R Gain of function Exon 6 T-ALL 120,122 ALL 123-124 R879, R724H, A634D, JAK1 Gain of function S646F, V658F T-PLL 125 Gain of function Varies; Hot spot R683 ALL 123 Amplification JAK2 copy number HL, PMBCL 30,134,170 JAK2 gains

Activating SEC31A HL 133 NKCL 136 T-ALL 171 JAK3 Gain of function M511I, A572V, A573V CTCL 137 T-PLL 125 ATLL 139 PTPN2 T-ALL 155 Loss of function Focal deletions (TC-PTP) HL, TCL 156 PTPRC Loss of function Varies T-ALL 160 (CD45) Loss of function Inactivating mutations PMBCL SOCS1 163 and deletions Mutational cluster in Gain of function T-LGL 145 STAT3 SH2 domain NKL 146

Gain of function Mutational cluster in LGL STAT5 148 SH2 domain

26 Gene Proposed Mutation or Disease References (Protein) function Phenotype Gain of function Hot spot in DNA PMBCL STAT6 75,152 binding domain Missense mutations in TYK2 Gain of function FERM, JH2 and kinase T-ALL (cell lines only) 140 domains

Descriptive studies

SSBP2 ALL 130 PAX5 ALL 129 JAK2 Activating PCM1 ALL 131 TEL (ETV6) ALL 128 DLBCL, FL 167 Loss of function Inactivating mutations FL 172 SOCS1 and deletions NLPHL 166 HL 165 TCL 143-144 Gain of function Mutational cluster in STAT3 SH2 domain DLBCL 141 Gain of function Hot spot in DNA STAT6 FL 153 binding domain

Altered gene expression and epigenetic regulation

Loss of transcript and PIAS3 Loss of function TCL 70 protein expression ALCL PTPN6 Loss of function Methylation and loss of FL, MCL, TCL, NKCL, 173 (SHP1) protein expression ATL, MM, ALL 174-176

27 Gene Proposed Mutation or Disease References (Protein) function Phenotype

SOCS1 Loss of function Methylation TCL, MM, MCL, FL 174,177-178

SOCS3 Loss of function Methylation DLBCL 172

STAT1 Activation Protein overexpression FL, PMBCL, HL 28,57,179

STAT2 Loss of function Alternative splicing BL 180

STAT3 Activation Protein overexpression HL, DLBCL, MCL 63,66,68

STAT5 Activation Protein overexpression HL, TCL 73,181

STAT6 Activation Protein overexpression HL, PMBCL 47,75

Table 1.1. Genetic aberrations of the JAK-STAT pathway in lymphoid malignancies.

28 Mutated Malignancy Mutational frequency References gene

IL7R T-ALL 6-9% 120-122 T-ALL 18% 124 JAK1 T-PLL 9% 125

JAK2 ALL 8.5% 123 NKCL 35.4% 136 T-ALL 7% 171 JAK3 CTCL 3.3% 137 T-PLL 42% 125 ATLL 11% 139 T-ALL 6% 155 PTPN2 TCL 3% 156

PTPRC T-ALL 9.2% 160 HL 42.1% 165 NLPHL 50% 166 SOCS1 DLBCL 27% 167 FL 27% 167 PMBCL 45% 163 T-LGL 40% 145 STAT3 DLBCL 5% 182 TCL 13.5% 144

STAT5 LGL 2% 148 PMBCL 36% 152 STAT6 FL 23.1% 153

Table 1.2. JAK-STAT somatic gene mutation frequencies in lymphoid malignancies.

Figure 1.1. The germinal centre reaction. [Adapted from Heesters, B. A., Nat Rev Immunol 2014 (ref.9). Abbreviations: SHM, Somatic hypermutation; GC, Germinal centre].

Figure 1.2. Representative hematoxylin and eosin (H&E) stains of a PMBCL (a; total magnification 200x) and a HL (b; total magnification 200x) patient. (Source: BCCA archives).

32 Figure 1.3. Deregulated JAK-STAT signaling in malignant lymphomas. Known gene alterations leading to constitutive pathway activity are shown in red. (Abbreviations: IL, Interleukin; JAK, Janus kinase; TYK, Tyrosine kinase; PTP, Protein tyrosine phosphatases; SOCS, Suppressor of cytokine signaling; STAT, Signal transducer and activator of transcription).

Figure 1.4. IL7-R activating mutations in T-cell acute lymphoblastic leukemia. Two hotspots of missense and in-frame insertion-deletion mutations (IL241–242 and VA253–254) aggregate in the transmembrane domain of IL7-R protein [Adapted from Zhang, J.H., Nature 2012 (ref. 120)].

Figure 1.5. Summary of reported JAK2 fusions in lymphoid malignancies. Protein domains of wild type JAK2 are shown on top and fusion partners are depicted in gray boxes (Abbreviations: FERM, band 4.1 ezrin, radixin and moesin; SH2, Src homology 2).

Chapter 2: Recurrent somatic mutations of PTPN1 in

primary mediastinal large B cell lymphoma and

Hodgkin lymphoma

36 2.1. Introduction

HL accounts for 11% of all malignant lymphomas, with an annual worldwide incidence of 3 in 100,000 people. Despite advances in treatment and high curability, 20% of patients still succumb to their disease. Moreover, a similar proportion of patients are over-treated leading to treatment-related consequences such as secondary malignancies and organ dysfunction45. PMBCL is a distinct subtype of aggressive B cell lymphoma that arises from thymic medullary B cells and characteristically presents as a mass in the anterior mediastinum. PMBCL affects predominantly young female patients10 but can also occur in children and adolescents, displaying similar clinical and pathological characteristics as in adults183. With PMBCL recognized as a separate entity, its optimal treatment is currently the subject of debate, and variable treatment outcomes are reported. Because relapses after first-line therapies usually occur early and salvage therapies have been reported to have high failure rates, dose-intense chemotherapy regimens have been suggested. Moreover, combination therapies with rituximab immunotherapy and radiation therapy are currently being evaluated15,19. PMBCL and the nodular sclerosing subtype of HL exhibit similar clinical, pathological and molecular features such as overlapping gene expression profiles as well as gains and losses of large chromosomal segments, specifically involving gains of 2p and 9p (refs. 20,184). Further, increased activation of JAnus Kinase-Signal Transducers and Activators of Transcription (JAK-STAT)47,68,73,75 and Nuclear factor kappa B (NF-κB) signaling31 have also been reported in both lymphoma subtypes. However, the full spectrum of genetic changes underlying the malignant phenotypes of both diseases still remains to be uncovered. Next-generation sequencing techniques have transformed the field of cancer genomics, allowing for rapid genome-wide characterization of single nucleotide variants (SNVs), insertions-deletions (indels), structural genomic rearrangements and gene expression changes185-188. The study of cancer genomes at base pair resolution in a variety of malignancies has yielded unprecedented insight into the complexity of primary and secondary alterations and has uncovered new and, in some cancer subtypes, highly recurrent driver alterations that produce cellular phenotypes that hold the promise

37 of being targetable by new therapeutics189-190. To investigate somatic gene mutations in HL and PMBCL, we used massively parallel next-generation sequencing to sequence the whole genome (WGS) and transcriptome (RNAseq) of two PMBCL index patients and analyzed eight additional cases by RNAseq. We discovered new gene mutations in the PTPN1 gene that were highly recurrent in our extension cohorts of HL and PMBCL. PTPN1 encodes PTP1B, a non-receptor-type member of the superfamily of protein tyrosine phosphatases, and until now, recurrent somatic coding-sequence mutations of PTPN1 had not been described in any lymphoma. More broadly, strong genetic evidence for a role in human cancers is lacking, although both a tumor suppressor and oncogene function of PTPN1 have been proposed191-192. Our data suggest that inactivating mutations contribute to the pathogenesis of PMBCL and HL through activation of JAK-STAT signaling pathways.

2.2. Materials and methods

Patient samples and cell lines Specimens from 77 PMBCL and 30 HL patients were selected from the tissue archives of the Centre for Lymphoid Cancer of the British Columbia Cancer Agency, Arizona Lymphoma Repository and the Hôpital Henri Mondor Pathology Department according to the availability of fresh-frozen lymphoid tissue biopsy material and clinical patient follow-up data. Of these, two PMBCL specimens with an available source of matched constitutional DNA were used for WGS. Seven PMBCL specimens, including the two cases from WGS, together with PMBCL-derived cell lines U-2940, MedB-1 and KARPAS-1106P, were used for RNAseq. The clinical characteristics of the study cohort stratified by PTPN1 mutational status are shown in Tables 2.1a, b. Ethical approval for this study was obtained from the University of British Columbia – British Columbia Cancer Agency Research Ethics Board (UBC BCCA REB) in accordance with the Declaration of Helsinki. Hodgkin Reed Sternberg (HRS) cells from HL tissue biopsies were enriched by laser microdissection and underwent whole-genome amplification

38 (WGA). All mutations found by WGA were validated in DNA from microdissected HRS cells by nested PCR without amplification. Cell lines KM-H2, L-428, L-540, U-H01, SUP- HD1, HDLM-2, HD-MY-Z, L-1236, L-591, U-2940 and KARPAS-1106P were obtained from the German Collection of Microorganisms and Cell Cultures and propagated according to standard conditions (DSMZ, Braunschweig, Germany, http://www.dsmz.de). The cell line MedB-1 was a kind gift from Drs. S. Brüderlein and P. Möller (University of Ulm, Germany) and propagated as published193.

Whole genome, deep amplicon and transcriptome sequencing Whole genome sequencing was performed as described186. In brief, genomic DNA was extracted using AllPrep DNA/RNA extraction kits (Qiagen). DNA was sheared by sonication, and the fraction of 350-450 bp in size was excised from a PAGE gel. Genome libraries were constructed using a modified paired-end protocol provided by Illumina (San Diego, CA, USA), which included end repair, end preparation, adapter ligation and ten cycles of PCR amplification of ligated fragments. The prepared library (10 nM) was sequenced on an Illumina HiSeq 2000 in 100-bp paired-end mode and aligned to the UCSC reference human genome (hg19) using Burrows-Wheeler Aligner194. JointSNVMix195 was used to call SNVs and filter out calls with somatic probabilities < 0.8. Reported SNPs were removed using dbSNP (version 132). Additionally, MutationSeq196 was applied to filter out SNVs with probabilities < 0.4, and annotations were added using SnpEff197. Only SNVs found in coding regions were reported. Genomic rearrangements were predicted using destruct, a software derived from nFuse198 that relies only on genomes, and were reported based on the basis whether: (i) rearrangement was a translocation, (ii) the breakpoint locations were within coding regions, (iii) rearrangement probability was > 0.5, and (iv) read support for the rearrangement was > 5 in the tumor library and there was no read support in other libraries. Copy number estimation, alterations and somatic copy number aberrations in WGS high-throughput sequencing data were performed using HMMcopy199. For validation by deep amplicon sequencing validation of WGS predictions, amplicons were generated using a two-step PCR approach with the following sequences added to gene-specific primers (Table 2.2). PCR amplification was

39 performed using 10 ng of genomic DNA and Q5 Hot Start High-Fidelity DNA polymerase (NEB). Indexed PCR2 products were pooled and sequenced on an Illumina MiSeq instrument using V2 300-cycles MiSeq reagent kit (Illumina), generating 150-bp paired-end reads. Amplicon sequencing libraries were aligned to hg19 using Bowtie200. SAMtools mpileup was used to generate reference and variant reads at each position in each sample. An X2 square test comparing the number of variant reads in the tumor compared to normal DNA was performed to test whether an SNV was somatic. RNAseq was performed as previously described186. In brief, double-stranded cDNA was synthesized from polyadenylated RNA and sheared. The fraction of 190-210 bp in size was isolated and amplified with ten cycles of PCR using the Illumina Genome Analyzer paired-end library protocol. The resulting RNAseq libraries were sequenced on 201 an Illumina Genome Analyzer II and aligned to hg19 using GSNAP. Gene fusions were predicted using deFuse198 and were reported on the basis of the following criteria: (i) altsplice = N, (ii) cDNA_breakseqs_percident < 0.5, (iii) breakseqs_estislands_percident < 0.5, (iv) genome_breakseqs_percident < 0.5, (v) break_adj_entropy_min > 0.5, (vi) probability > 0.5, (vii) one of the breakpoints was in a coding region, and (viii) gene fusion partners are not adjacent. SNVs were called using SNVMix2 (ref. 202) and annotated using SnpEff.197 Reported SNPs were removed using dbSNP (version 132). SNVs were reported based on the following criteria for clinical samples: (i) SNV affected a protein-coding gene, (ii) the SNV had a protein-coding effect, (iii) probability of non-reference allele > 0.8, (iv) > 4 reads support the variant allele, and (v) < 5% of the variant reads contained an insertion/deletion in them. SNVs were reported based on the following criteria for cell lines: (i) SNV affected a protein- coding gene, (ii) the SNV had a protein-coding effect, (iii) probability of non-reference allele > 0.95, (iv) > 6 reads support the variant allele, (v) variant frequency > 45% and (vi) < 1% of the variant reads contained an indel.

Laser microdissection of HRS cells HRS cells were enriched by laser microdissection (Nikon eclipse TE2000-S microscope equipped with Molecular Machines Industries Technology) as follows.

40 Sections (6 µm) from fresh-frozen tissue were mounted on membrane slides and fixed in 70% ethanol. Adjacent sections were also prepared on glass slides and stained with hematoxylin and eosin to assess cell morphology and HRS cell content, which revealed a wide range from only 1 cell per high-power field in HRS cell-poor cases to approximately 50 cells per high-power field in HRS cell-rich cases. Sections mounted on membrane slide for microdissection were stained with hematoxylin for 20s and 500– 1,000 individually-picked cells per case were exercised and lysed in cell lysis solution (Purogene Core Kit A) for subsequent DNA extraction and whole-genome amplification. On average, 4 consecutive sections were needed per case to obtain the required number of HRS cells (range of 2–8 sections).

DNA extraction and Whole Genome Amplification (WGA) Genomic DNA from PMBCL clinical specimens and 500–1,000 pooled HRS cells per case were extracted using Allprep DNA/RNA extraction kits (Qiagen, Mississauga, ON, Canada) and Puregene Cell & Tissue kits by Gentra Systems (Mississauga, ON, Canada), respectively. Proteinase K digestion was carried out overnight, and glycogen (Invitrogen, Carlsbad, CA, USA) was used as a DNA carrier. WGA was only performed on HRS cell extracted DNA using the GenomePlex Whole Genome Amplification kit (Sigma-Aldrich, Oakville, Ontario, Canada) to obtain >200 ng of amplified product per case. For quality control purposes, 50 ng of amplified DNA was analyzed by multiplex PCR (Qiagen, Mississauga, ON, Canada) as previously described 203. The PCR reaction was performed with five primer sets that produce 100, 200, 300, 400 and 600 bp fragments from non-overlapping target sites in housekeeping genes204. Samples were classified on the basis of the largest of the five possible PCR products detected, and cases with at minimum the 300-bp product were used for further analysis. All mutations discovered by WGA were confirmed by targeted nested PCR using DNA from freshly microdissected HRS cells to remove false positives.

Screening for PTPN1 somatic mutations PTPN1 mutations in PMBCL were detected by Sanger sequencing (n=49) and deep amplicon sequencing (n=28). Mutations in HL were detected by nested PCR

41 followed by Sanger sequencing. For Sanger sequencing, primer sequences are listed in Table 2.3. All ten PTPN1 exons were amplified by standard PCR (Invitrogen Carlsbad, CA, USA) using primer set 1 (PMBCL) or primer sets 1 and 2 for nested PCR (HRS) and Sanger sequencing (3130 Genetic Analyzer, Applied Biosystems, Foster City, CA, USA). Mutation analysis was performed using Clone manger software (Scientific & Educational Software, Cary, NC, USA). For deep amplicon sequencing, we used the Illumina TruSeq Custom Amplicon (TSCA) protocol according to the manufacturer’s instructions covering the complete protein-coding sequence of PTPN1. Amplicon libraries were sequenced on an Illumina MiSeq instrument using V2 300-cycles MiSeq reagent kit (Illumina), generating 150-bp paired-end reads. Data analysis was performed as described above.

Analysis of tissue sections by immunohistochemistry (IHC) Immunohistochemistry was performed on formalin-fixed, paraffin-embedded tissue (FFPET) and fresh-frozen samples of HL (n=215) and PMBCL (n=143) that included previously published cases represented on a tissue microarray (TMA)25,188. Whole-tissue sections of selected cases were stained and are shown in Figure 3. TMAs and sections were stained with an antibody to PTP1B (04-1140, Millipore, Billerica, MA, USA) using routine protocols for automated procedures on the Ventana Benchmark XT (Ventana Medical Systems, Tucson, Arizona). Scoring was performed by K.L.T., A.M. and R.D.G., and the percentage of tumor cells that were positive for PTP1B was recorded. For HL, CD30 IHC was used as a reference to identify HRS cells.

Quantitative RT-PCR Taqman gene expression assay probes were used to detect mRNA levels of PTPN1 (Hs00182260_m1), BCL6 (Hs00277037_m1) and MYC (Hs00905030_m1) on a 7900HT real-time PCR system (Applied Biosystems, Foster city, CA, USA). GAPDH was run as an internal control. Measurements were quantified using the ∆∆Ct method (Pfaffl) and expressed relative to the expression in non-silencing cells.

42 Fluorescence in situ hybridization (FISH) analysis FISH analysis was performed according to standard protocols with cells fixed with Carnoy’s solution for cell lines, on PIN (paraffin isolated nuclei) for PMBCL clinical samples and FICTION (Fluorescence Immunophenotyping and Interphase Cytogenetic as a Tool for Investigation Of Neoplasia) for HL cases. For PTPN1 copy number detection, a dual-color probe was designed using in-house bacterial artificial chromosome (BAC) clones (CTD-2582P13) for 20q13.3 labeled in spectrum orange and telomere 20p13 (Abbott Molecular, IL, USA) as a control labeled in spectrum green. Slides were analyzed using a Zeiss Axioplan 2 fluorescence microscope (Zeiss Gollingen, Germany) and documented using an ISIS imaging system (Metasystems, Altlussheim, Germany). One hundred interphase nuclei were scored for PTPN1 copy number.

PTPN1 silencing by lentiviral-mediated RNA interference Stable transcript knockdowns of PTPN1 transcript in HL cell line KM-H2 (KMH2- PTPN1-KD) were generated by lentiviral transduction of a vector expressing shRNA (pGIPZ-sh PTPN1, clone V2LHS_262177, Open Biosystems, Huntsville, AL, USA) that, after processing to mature small interfering RNA (siRNA), interferes with PTPN-1 exon 10. Cells transduced with a vector with a non-interfering shRNA insert (pGIPZ non- silencing lentiviral shRNA control, RHS4346, Open Biosystems) were used for comparison (KMH2-NS). For lentivirus production, shRNAmir constructs were co-transfected into HEK 293T cells with a Trans-Lentiviral Packaging Mix (Open Biosystems; containing pTLA1- Pak, pTLA1-Enz, pTLA1-Env, pTLA1-Rev and pTLA1-TOFF) by lipid-mediated transfection (Arrest-In, Open Biosystems, Thermo Scientific, Nepean, Ontario, Canada). Supernatant was collected 48h and 72h after transfection and concentrated by ultracentrifugation (Optima, Beckman, Brea, CA, USA) for 1.5 h at 65,000g at 4°C. Lentiviral particles were resuspended in RPMI and used for the transduction of KM-H2 cells. Wild-type cells were transduced by adding 4 g/ml polybrene (final concentration) with both pGIPZ-shPTPN1 virus and non-silencing control virus. GFP-positive cells were sorted on a BD FACSAria cell sorter (BD Biosciences) 3-5 days after transduction

43 and were cultured using standard techniques. Efficiency of knockdown was evaluated by measuring residual expression of the transcript by qRT-PCR.

Expression of PTPN1 mutants and SEAP reporter gene assay The PTPN1 coding sequence was amplified by PCR using cDNA from KM-H2 (wild type), L-1236 (Q9*), MedB-1 (R156*), PM-5 (A69V), HD-MY-Z (M74L), PM-7 (V184D) and L-428 (M282L), and cloned into the mammalian expression vector, pcDNA3.1 (Invitrogen, Carlsbad, CA, USA). Empty pcDNA3.1 (MT) was used as a mock vector. Plasmids were purified by Spin miniprep kit (Qiagen, Mississauga, ON, Canada) and 2.5 g of each plasmid was transfected into HEK 293 cells expressing STAT6 (HEK 293-STAT6; HEK blue IL4/IL13, Invivogen, San Diego, CA, USA) using Lipofectamine 2000 (Invitrogen). Transfected cells were cultured for 48 h, and secreted embryonic alkaline phosphatase (SEAP) levels assayed in culture supernatant according to the manufacturer’s protocol. A plasmid encoding GFP was used to determine equal transfection efficiency (>90%).

Cell lysis and protein blotting Total cell lysates were prepared from cultured cell lines treated with or without 10 ng/ml (for KM-H2-PTPN1-KD cells) and 20 ng/ml (for HEK 293-STAT6 cells) recombinant human IL-4 (R&D Systems, Minneapolis, MN, USA) for 10 min at 37˚C using mammalian protein extraction reagent (M-PER, Thermo Scientific, Waltham, MA, USA) and in the presence of a protease inhibitor cocktail (Sigma-Aldrich, Oakville, ON, Canada). Protein lysates (25 g) were resolved on a 4-12% NuPAGE Novex Bis-Tris gradient gel (Invitrogen, Carlsbad, CA, USA) and transferred to a nitrocellulose membrane (Thermo Scientific, Waltham, MA, USA) by semi-dry transfer (Bio-Rad, Mississauga, ON, Canada) and probed with the following primary antibodies: at 1:1,000 dilutions: PTP1B (AE4-2J, EMD, Darmstardt, Germany); JAK1 (3332), phospho-JAK1 (3331), JAK2 (3230), phospho-JAK2 (3776), STAT3 (9132), phospho-STAT3 (9134), phospho-STAT5 (9359), phospho-STAT6 (9364), AKT (4511), phospho-AKT (4058) (all from Cell Signaling Technology, Danvers, MA, USA); STAT5 (sc-835, Santa Cruz Biotechnology, Santa Cruz, CA, USA); and STAT6 (ab32108; Abcam, Cambridge, MA,

44 USA). Antibody to β-actin (A5441; Sigma-Aldrich, Oakville, ON, Canada) was used as an internal control. Appropriate horse radish peroxidase (HRP)-conjugated secondary antibodies to rabbit or mouse IgG (W401B and W402B respectively, Promega, Madison, WI, USA) were used at 1:10,000 dilutions to visualize bands with the enhanced chemiluminiscence (ECL) system (Amersham, Baie d’Urfe, QC, Canada) on an X- OMAT 2000A film processor (Kodak, Rochester, NY, USA) or a Chemidoc digital imager (Biorad, Hercules, CA, USA). Intensities of developed bands were quantified using Image J software (National Institutes of Health, USA).

PTP1B activity assays Phosphatase activity of prepared cell lysates from HEK 293-STAT6 expressing WT or mutant PTP1B were assayed using the Duoset IC Human Active PTP1B kit (R&D, Minneapolis, MN, USA) according to the manufacturer’s instructions. Sodium orthovanadate, (Na3VO4) was used as a tyrosine phosphatase inhibitor.

Gene expression profiling RNA was extracted from duplicate KMH2-PTPN1-KD and KMH2-NS cultures and hybridized on Human Genome U133 Plus 2.0 microarrays (Affymetrix, Santa Clara, CA, USA). Output was processed using the Robust Multi-array Average method. Probe-set- fold-change differences represent average expression in knockdown versus non- silencing samples. The binomial sign test was performed using the “binom test” function (R project). Genes with a fold change of >1.25 or <0.8 were considered as being differentially expressed and are reported. Molecular pathway and gene set enrichment analyses of differentially expressed genes were performed using Ingenuity Pathway Analysis (Ingenuity® Systems), GSEA (2.0.13) and the Molecular Signatures Database (4.0, Broad Institute, Cambridge, MA, USA).

Statistical analysis Comparisons between groups were performed using two-sample Student t tests. Time-to-event analyses were performed using the Kaplan-Meier method, and survival curves were compared by the log-rank test using SPSS Version 14.0. Disease-specific

45 survival (DSS) was defined as the time elapsing from diagnosis to last follow-up or death due to lymphoma. Overall survival (OS) was defined as the time from diagnosis to death from any cause. Progression-free survival was defined as the time from diagnosis to progression (relapse after primary treatment, initiation of new treatment or death from any cause). Treatment failure was defined as the time from primary treatment to progression.

2.3. Results

Somatic mutations discovered by next-generation sequencing. To discover somatic mutations in PMBCL cancer genomes, we used whole- genome sequencing, comparing tumor genomes for two index cases to matched constitutional genomes. We generated 1,986,529,080 and 1,707,225,268 paired-end reads for subjects PM-2 and PM-7, respectively, of which 1,885,237,436 and 1,623,382,611 reads mapped to the reference human genome (94.9% and 95.1%; 188 and 162 gigabases), resulting in average genomic coverage of 60 and 51.6x. The whole genomic landscape for these index cases, consisting of copy number losses and gains and gene rearrangements are shown in Figure 2.1a. Average intermutational distances were 317,246 and 280,509 bp for PM-2 and PM-7, respectively (Figure 2.1b, c). We observed several genomic ‘hotspot’ regions with high mutation frequency in PM- 7, including the BCL6 oncogene205-206 and the λ light-chain locus, known targets of somatic hypermutation. Somatic variant calling, including of SNVs and indels, identified 20,044 intronic (9,425 and 10,619), 32,565 intergenic (16,176 and 16,389), 22 5’UTR (11 and 11), 81 3’UTR (39 and 42) and 299 coding-region (148 and 151) mutations in the PM-2 and PM-7 genomes, respectively. Coding mutations and the distribution of nucleotide interchanges are shown in Tables 2.4-2.6 and Figure 2.1d, e, respectively. In both cases, we found mutations in two negative regulators of the JAK-STAT signaling pathway, SOCS1 and PTPN1, and, since constitutive activation of JAK-STAT signaling has been reported as a hallmark feature of PMBCL, we focused on gene alteration in this pathway. Analysis of the transcriptome by RNAseq of five additional cases and

46 three PMBCL cell lines identified PTPN1 mutations in two more cases (European genome-phenome archive accession EGAS00001000554). The mutational patterns of gene alterations in the JAK-STAT signaling pathway are shown in Figure 2.1f, validating previously reported mutations of SOCS1 and STAT6 in PMBCL152.

PTPN1 is mutated in PMBCL and HL. As the mutation frequency of PTPN1 in the discovery cohort was second only to that of mutations in SOCS1, a well-characterized tumor suppressor gene in PMBCL and HL, we focused our study on PTPN1 (refs.154,163,165,167). We screened the complete coding-sequence of PTPN1, compromising 10 exons, for genomic mutations in an additional 70 PMBCL samples by Sanger sequencing and deep amplicon sequencing. In total, after the exclusion of reported single nucleotide polymorphisms (SNPs) and silent mutations, we found 20 variants in our PMBCL cohort (18 mutations in 17 of 77 clinical samples and 2 mutations in 1 of 3 cell lines screened; Table 2.7a), with some cases harboring multiple mutations. Because classical HL is a closely related disease entity, we also screened 9 HL-derived cell lines and 30 samples of HRS cells, microdissected from classical HL. A total of 12 mutations were discovered (6 in 30 microdissected HRS cells and 6 in 9 cell lines screened; Table 2.7b). In summary, we found coding- sequence mutations in 30 of 119 (25%) clinical samples and cell lines of HL and PMBCL. We observed 18 (60%) missense mutations, 4 (13.3%) frameshift mutations, 3 (10%) single-amino acid deletions, 4 (13.3%) nonsense mutations and 1 (3.3%) promotor mutation. The distribution and frequency of mutations for each exon and types of mutations observed are shown in Figure 2.2a. For all 6 HL cases and 2 PMBCL cases with PTPN1 mutations, we confirmed the identified mutations as somatic by sequencing of constitutional DNA extracted from whole-lymph node tissue or peripheral blood. Mutation frequencies in PMBCL cell lines, PMBCL clinical cases, HL cell lines and HL clinical cases were 33%, 22%, 67% and 20%, respectively. Frequent mutations in Q9 and Q21 (NCBI reference sequence NP_002818.1) are suggestive of a potential hot spot within the first exon (Figure 2.2b) where aberrations might lead to premature inactivation of the protein, as was demonstrated by immunohistochemical staining in the PMBCL case, PM-2 (Figure 2.3c).

47 We next studied PTP1B protein expression by immunohistochemistry using formalin-fixed paraffin, embedded and fresh frozen tissues from HL (n=215) and PMBCL (n=143) (Figure 2.4). Analysis of all available cases with known PTPN1 mutation status identified a significant correlation between PTPN1 mutations and decreased PTP1B protein expression. In PMBCL, the mean percent of PTP1B- positive tumor cells in the group with wild-type PTPN1 (n=33) was 58.3% compared to 26.7% in the group with mutant PTPN1 (n=9; P=0.01). In HL, the mean percentage of PTP1B- positive tumor cells in the group with wild-type PTP1B (n=24) was 63% compared to 35.8% in the group with mutant PTPN1 (n=6; P=0.04) (Figures 2.4a, b). The distribution of PTP1B-positive tumor cells in PMBCL and HL tissue samples are shown in Figures 2.1c, d, and representative immunohistochemistry images of PMBCL and HL tissue sections exhibiting variable PTP1B staining are shown in Figure 2.3. We also sought to determine whether PTPN1 mutations were associated with patient survival. Kaplan-Meier survival analysis of individuals with PMBCL and HL patients showed no significant differences in clinical outcome between groups with wild-type and mutated PTPN1 (P>0.05; Figures 2.5-2.6). In addition to inactivating mutations that were correlated to loss of expression by immunohistochemistry (Figure 2.3c) in clinical samples, we also found copy number losses in cell lines Karpas-1106P and U-H01 by fluorescence in-situ hybridization (FISH) (Figure 2.7). However, focal PTPN1 copy number changes were not found in any clinical samples studied by FISH (9 HL and 13 PMBCL cases). Of interest, in U-H01 the remaining PTPN1 allele contains a deletion of exons 2-8 (chr20: 49,153,833 – 49,197,665), leading to complete loss of PTP1B protein expression and transcript levels (Figures 2.8a, b). Moreover, the described nonsense and frameshift mutations of PTPN1 in cell lines MedB-1 (in trans-allelic configuration; Figure 2.9) and L-1236, as well as the multiple-exon deletion in SUP-HD1, are consistent with complete loss or reduced expression of C-terminal PTP1B by protein blotting compared to a lymphoblastoid positive control cell line (LCL) (Figure 2.8a). However, absolute PTPN1 transcript and PTP1B protein levels varied among cell lines and did not consistently correlate with protein expression, possibly due to post-transcriptional and post- translational regulatory mechanisms.

48 PTPN1 mutations decrease phosphatase activity. To evaluate how phosphatase activity was affected by the observed nonsense and missense mutations, we expressed wild-type PTP1B or mutant PTP1B with Q9*, R156*, A69V, M74L, V184D, and M282L alterations in engineered human embryonic kidney 293 cells expressing STAT6 (HEK-STAT6) and a STAT6-inducible reporter gene (secreted embryonic alkaline phosphatase, SEAP). When cells were stimulated with human interleukin-4 (IL-4), expression of wild-type PTP1B resulted in marked dephosphorylation of STAT6 compared to the empty vector control (relative densitometric value 0.1). In contrast, we observed sustained phosphorylation of STAT6 with the expression of all mutants in comparison to the empty vector control (relative densitometric values: Q9* 1.4, R156* 1.3, A69V 1.8, M74L 1.1, V184D 1.8 and M282L 1.1, Figure 2.10a) confirming these alterations as deleterious. Similarly, analysis of activated STAT6-dependent SEAP reporter gene expression in these mutant cells showed JAK-STAT pathway activation under IL-4 stimulation (percentage of SEAP expression compared to wild type: empty vector 891%, Q9* 656%, R156* 742%, A69V 417%, M74L 527%, V184D 208%, and M282L 575%, Figure 2.10b). Furthermore, we observed no phosphatase activity for the nonsense mutants and a moderate but significant reduction in phosphatase activity for the missense mutants using a tyrosine phosphatase-specific substrate (percentage of phosphatase activity compared to wild type: Q9* 3.3%, R156* 3.7%, A69V 22.1%, M74L 79.7%, V184D 31.3% and M282L 79.6%, Figure 2.10c).

PTPN1 knockdown hyperphosphorylates JAK-STAT proteins. To study the functional relevance of the predicted inactivating PTPN1 mutations, we generated a knockdown HL cell line (KMH2-PTPN1-KD) in which PTPN1 mRNA expression was silenced by lentiviral transduction of a small hairpin (shRNA) construct targeting PTPN1. Knockdown efficiency was determined by quantitative RT-PCR (qRT- PCR) (34% residual mRNA; Figure 2.11a) and by protein blotting, which showed minimal PTP1B protein expression compared to cells transduced with a non-silencing construct (KMH2-NS) (Figure 2.11b). To determine whether silencing of PTPN1 affected JAK-STAT signaling, we analyzed cell lysates by protein blotting for

49 hyperphosphorylation of members of the JAK-STAT pathway. We also treated cells with IL-4 in order to stimulate the pathway to enhance detectable changes in phosphorylation. In KMH2-PTPN1-KD cells treated with IL-4 compared to negative control cells treated with IL-4, we observed increased phosphorylation of STAT3, STAT5, STAT6, JAK1, and JAK2 (densitometric values 12.6 versus 1.5, 2.8 versus 1.0, 1.5 versus 0.7, 7.0 versus 1.0 and 3.8 versus 1.7, respectively). In addition, we detected increased levels of JAK1 (2.9 versus 1.4). We also observed increased phosphorylation of the prosurvival protein AKT (5.1 versus 2.0), consistent with similar findings in mouse fibroblasts207. To investigate genome-wide gene expression changes in KMH2-PTPN1-KD versus KMH2-NS cells, we performed gene expression profiling using human genome U133 plus 2.0 arrays. We found 2,463 differentially expressed probe sets, of which 1,807 were found to be upregulated (fold change >1.25 compared to KMH2-NS cells) and 656 were found to be downregulated (<0.8 compared to KMH2-NS; Gene expression omnibus accession GSE54157). Because STAT3 showed the most pronounced changes in phosphorylation in our silenced cells and because of its reported effects on cell proliferation and survival66, we looked for STAT3 downstream targets in our differentially expressed genes. Thirty eight of 50 (76%; sign-test P=0.0003) known STAT3-responsive genes were upregulated in PTPN1 knockdown cells (Figure 11c)208. Of note, two members of the gene family of ATP-binding cassette (ABC) transmembrane transporters were overexpressed in KMH2-PTPN1-KD cells: ABCA8 (fold change = 5.2) and ABCB1 (fold change = 1.5) compared to KMH2-NS (Gene expression omnibus accession GSE54157), with the latter being implicated in drug resistance in a number of malignancies209. Molecular pathway analysis of over- and under expressed genes using Ingenuity Pathway Analysis (Ingenuity Systems) confirmed alterations in JAK-STAT-mediated cytokine signaling (false discovery rate (FDR)-adjusted P=0.004) and showed defects in B cell development (FDR-adjusted P=0.004), respectively (Tables 2.8a-d). Gene set enrichment analysis of differentially expressed genes showed enrichment of genes involved in tumorigenesis and cytokine, growth receptor and JAK-STAT signaling (Figure 2.12). Moreover, the upregulated oncogenes BCL6 and MYC in silenced cells were validated by qRT-PCR (189% and

50 147% of the levels in KMH2-NS cells, respectively; Figure 2.11a).

2.4. Discussion

Using whole-genome and transcriptome sequencing approaches, we identified somatic coding-sequence mutations in the non-receptor-type tyrosine phosphatase gene PTPN1, a gene that has not previously been described as being recurrently mutated in lymphoid cancers. By targeted Sanger sequencing, we found that PTPN1 mutations were highly recurrent in the molecularly related B cell lymphoma entities PMBCL and HL. The pattern of identified mutations consisting of nonsense, missense, frameshift changes and single amino acid deletions with a hot spot region in the first exon is indicative of loss of function of a negative regulator of oncogenic JAK-STAT activation. As demonstrated, mutated PTPN1 yields functionally deleterious or hypomorphic phosphatases that result in deregulated STAT phosyphorylation. Moreover, heterozygous and homozygous deletions in some cell lines, as well as loss of detectable protein by immunohistochemistry and protein blotting in samples harboring PTPN1 coding-sequence mutations, are consistent with classical mechanisms of inactivation of a tumor suppressor gene in the context of these B cell lymphomas. The role of PTPN1 expression in human cancer is controversial. While recurrent coding-sequence mutations have not been previously described, both up- and downregulation of PTPN1 gene expression have been linked to carcinogenesis, suggesting both tumor suppressor gene and oncogene functions depending on the cellular context210-215. Protein tyrosine phosphatases act as a molecular “switch” to maintain the homeostatic balance of receptor tyrosine kinase signaling by removing phosphate moieties from phosphorylated proteins. This critical regulation mediates diverse cellular processes such as cell proliferation, adhesion, migration and immune and hormonal responses216-217. Alterations in PTPN2 (encoding TC-PTP) were identified in T cell acute lymphoblastic leukemia (T-ALL), where mutations and focal deletions led to hyperphosphorylation of JAK-STAT signaling molecules; however, recurrent coding- sequence mutations could not be established in B cell lymphoma155,157. More recently,

51 an elegant study of triple-negative breast cancer demonstrated the suppressive role of PTPN12 in inhibiting HER2 and EGFR oncogenic kinases in order to restrain mammary epithelial cell proliferation and transformation218. However, not all members of the PTP family exhibit tumor-suppressive properties. Gain-of-function mutations of the proto- oncogene PTPN11 can aberrantly activate the RAS-MAPK cascade in myeloid leukemias219. Our report of PTPN1 coding-sequence mutations in B cell lymphomas describes another PTP family member as being critically involved in oncogenesis. PTP1B is ubiquitously expressed in human tissues and is found in the cytoplasm, localized within the endoplasmic reticulum. It dephosphorylates tyrosine residues on many growth factor receptors (for example, EGFR, PDGFR and HGFR), activated kinases (for example, Src, p210BCR-ABL, TYK2 and JAK2) and transcription factors (for example, STAT5) to maintain cellular homeostasis220-221. So far, the functional role of PTP1B expression in lymphomagenesis has only been studied in a Tp53 and ptpn1 double-null murine model where an increased accumulation of B cells in the bone marrow and lymph nodes and decreased survival rates were observed222. This lymphoma-like phenotype, together with our observation of increased phosphorylation of AKT and JAK-STAT pathway members as well as abrogated or diminished phosphatase activity resulting in sustained STAT6 activation, is highly indicative of a tumor-promoting phenotype in the absence of functional PTP1B. Moreover, upregulation of the classical lymphoma oncogenes BCL6 and MYC in KM-H2 cells with PTPN1 knockdown might in part explain the tissue specificity of B cell lymphoma. However, the absence of PTPN1 mutations in other lymphoma subtypes that were previously studied by next-generation sequencing223-224 suggests a synergy of PTPN1 mutations with other somatic changes or specific cell of origin-related expression programs. In particular, it is likely that the mutations in PTPN1 synergize with other driver mutations known to be involved in the pathogenesis of HL and PMBCL (mutations in SOCS1 and STAT6) that contribute to aberrant JAK-STAT signaling. Our results demonstrate a significant correlation of PTPN1 mutations with diminished PTP1B expression in clinical samples and active JAK-STAT signaling in vitro. However, additional studies in a larger cohort are needed to investigate mutational patterns of JAK-STAT signaling genes, including PTPN1 mutations, and to correlate

52 these patterns with gene expression phenotypes. These studies will also clarify if PTP1B immunohistochemistry can be used as a reasonable surrogate for PTPN1 mutations in clinical samples. Interestingly, we identified the multidrug resistance gene ABCB1 to be upregulated in KM-H2 PTPN1 knockdown cells. ABCB1 is a prominent family member of the ABC-containing transmembrane transporters that confer drug resistance in multiple cancers225-226. Of note, lack of PTPN1 has been implicated in treatment resistance of chronic myeloid leukemia210. Although only a trend toward inferior progression-free survival was observed in our PMBCL cases with PTPN1 mutations, further validation studies in larger homogeneously treated HL and PMBCL cohorts are needed to evaluate the prognostic value of PTPN1 mutations in these diseases. Lastly, the identification of a specific lymphoma phenotype, including drug resistance, linked to these mutations, harbors the potential for the development of targeted treatment approaches in this molecularly characterized subgroup of B cell lymphomas.

In summary, we have discovered recurrent somatic mutations in PTPN1, a phosphatase that negatively regulates kinase activity in HL and PMBCL. Our data suggest PTPN1 mutations as novel oncogenic drivers in these lymphomas, with implications for future treatment strategies.

Accession numbers: Sequencing data have been deposited in the European Genome- phenome Archive (EGA) under accession EGAS00001000554, and gene expression data have been deposited in the Gene Expression Omnibus (GEO) under accession GSE54157. URLs: Ingenuity, http://www.ingenuity.com/; R project, http://www.R-project.org/; Deutsche Sammlung von Mikroorganismen und Zellkulturen (DSMZ), http://www.dsmz.de/; destruct, https://code.google.com/p/destruct

PTPN1 WT PTPN1 mutated P-value Variable No. (%) No. (%) N 60 (100%) 17 (100%) NA Median age - years 37 33 0.906 Male sex 22 (36.6%) 7 (41.1%) 0.735 Ann Arbor Stage† 0.709 I 5 (9.2%) 5 (35.7%) II 34 (62.9%) 6 (42.8%) III 6 (11.1%) 0 (0%) IV 9 (23.3%) 3 (21.4%) Presence of constitutional 23 (56.0%) 4 (36.3%) 0.499 symptoms‡ Tumor size  0.419 Median - cm 11 11 ≥10 cm 38 (70.3%) 9 (69.2%) 0.936 Primary treatment # 0.1 CHOP 4 (7.2%) 2 (15.3%) CHOP-like* 23 (41.8%) 7 (53.8%) Rituximab 26 (47.2%) 4 (30.7%) Extranodal sites § 0.275 0-1 13 (32.5%) 7 (77.7%) >1 27 (67.5%) 2 (22.2%) IPI score@ 0.584 0-1 26 7 2-3 19 4 4-5 8 1

† Missing data for 6 cases (PTPN1 WT) and 3 cases (PTPN1 mutated) ‡ Missing data for 19 cases (PTPN1 WT) and 6 cases (PTPN1 mutated)  Missing data for 6 cases (PTPN1 WT) and 4 cases (PTPN1 mutated) # Missing data for 5 cases (PTPN1 WT) and 4 cases (PTPN1 mutated) * CHOP-like: NCVBP (GELA trial), ACVBP (GELA trial), ECVBP (GELA trial), MACOPB, VACOPB, ECV, ACOP § Missing data for 20 cases (PTPN1 WT) and 8 cases (PTPN1 mutated) @ Missing data for 7 cases (PTPN1 WT) and 5 cases (PTPN1 mutated)

Table 2.1a. Demographic and clinical characteristics of PMBCL patients.

54 PTPN1 WT PTPN1 P-value Variable No. (%) mutated No. (%) N 24 (100%) 6 (100%) NA Median age – years 36 26.5 0.525 Male sex 14 (58.3%) 3 (50%) 0.713 Histologic subtype 0.095 Nodular sclerosis 22 (91.7%) 5 (83.3%) Mixed cellularity 0 (0.0%) 1 (16.7) Lymphocyte-rich 2 (8.3%) 0 (0.0%) Ann Arbor Stage 0.639 I 2 (8.3%) 0 (0.0%) II 13 (54.2%) 4 (66.6%) III 6 (25.0) 1 (16.7%) IV 3 (12.5%) 1 (16.7%) Presence of constitutional 5 (20.8%) 2 (33.3%) symptoms Tumor size† Median – cm 4 5 0.439 ≥10 cm 7 (29.1%) 0 (0.0%) 0.256 IPS ≥ 4 (high risk) 5 (20.8%) 1 (16.7%) 0.819 Primary treatment 0.755 ABVD chemotherapy with or 22 (91.7%) 6 (100%) without radiation Extended-field radiation alone 2 (8.3%) 0 (0.0%)

† Missing data for one case (PTPN1 WT) Median follow up of live patients 7.5 years in both groups.

Table 2.1b. Demographic and clinical characteristics of HL patients.

PCR Direction Sequence

PCR 1 Forward 5’-TCGTCGGCAGCGTCAGATGTGTATAAGAGACAG-3’ Reverse 5’-GTCTCGTGGGCTCGGAGATGTGTATAAGAGACAG-3’

5’-AATGATACGGCGACCACCGAGATCTACAC[NNNNNNNN]TCGTCGGC PCR 2 Forward AGCGTC-3’ 5’-CAAGCAGAAGACGGCATACGAGAT[NNNNNNNN]GTCTCGTGGGCTC Reverse GG-3’

Table 2.2. Primers used for deep amplicon sequencing.

56 Set 01

PTPN1 Sequence Size Chr. Chr. End exon (bp) Start E1 F1 TGTAAAACGACGGCCAGTaattctcgatcgctgattgg 49126837 49126856 E1 R1 CAGGAAACAGCTATGACagtaggagcgaggcagagg 404 49127222 49127241 E2 F1 TGTAAAACGACGGCCAGTccttgcatttcccatattgc 49177797 49177816 E2 R1 CAGGAAACAGCTATGACctgaggcgaaaggagtcttg 272 49178050 49178069 E3 F1 TGTAAAACGACGGCCAGTccaccaactcacctttgctt 49181410 49181429 E3 R1 CAGGAAACAGCTATGACtgctgctttttggtatgctg 324 49181715 49181734 E4 F1 TGTAAAACGACGGCCAGTctaagctgtggggactgagg 49184807 49184826 E4 R1 CAGGAAACAGCTATGACcgtaaaggaaatacaaaggcaag 319 49185104 49185126 E5 F1 TGTAAAACGACGGCCAGTcccaggcctttgagttatca 49190947 49190966 E5 R1 CAGGAAACAGCTATGACtcacccaagggataaaggaa 317 49191245 49191264 E6 F1 TGTAAAACGACGGCCAGTagagagtggaaggtgactctgtg 49194864 49194886 E6 R1 CAGGAAACAGCTATGACaagaggagcgcttcagtttc 392 49195237 49195256 E7 F1 TGTAAAACGACGGCCAGTtgggaggggacagacatcta 49195618 49195637 E7 R1 CAGGAAACAGCTATGACcaactgacagcctccttcaa 321 49195920 49195939 E8 F1 TGTAAAACGACGGCCAGTgtcacctctgctcatgcaaa 49196156 49196175 E8 R1 CAGGAAACAGCTATGACagaaagtcatttcccctcacc 356 49196492 49196512 E9 F1 TGTAAAACGACGGCCAGTccacaatagcagcatccttg 49197730 49197749 E9 R1 CAGGAAACAGCTATGACtgcaccacagaactgaatcc 359 49198070 49198089 E10 F1 TGTAAAACGACGGCCAGTctccctcggaggttgaagtt 49199121 49199140 E10 R1 CAGGAAACAGCTATGACaaaatggcttgtggatttgg 363 49199465 49199484

Set 02

PTPN1 Sequence Size Chr. Chr. End exon (bp) Start E1 F2 ATTAGATATCTCGCGGTGCTG 49126762 49126782 E1 R2 CCATCGAATCCTCAAGCAGT 495 49127238 49127257 E2 F2 GCTCATTCCTCCTCGGTTTT 49177743 49177762 E2 R2 CGCGTCATAAACCTCTGCTA 371 49178095 49178114 E3 F2 GTCGCCCCATACTGATGACT 49181344 49181363 E3 R2 CCCAGAAAAGATTAGCGTTCC 456 49181780 49181800 E4 F2 TACTGACCACCAGGCAGAGA 49184749 49184768 E4 R2 AGAGGTCCGTGCAGTTCTTC 488 49185218 49185237 E5 F2 ATGAGGCAGGCATCTGTGC 49190901 49190919 E5 R2 TCACCCAAGGGATAAAGGAA 363 49191245 49191264 E6 F2 TTTGTTGACTGGGTGTGTGG 49194802 49194821

57 PTPN1 Sequence Size Chr. Chr. End exon (bp) Start E6 R2 CGAGTCTCAGGTACGCCTTT 521 49195304 49195323 E7 F2 TTGTGAATGCATTGCCTCAG 49195498 49195517 E7 R2 GCTATTTCCCCTGACACCAG 479 49195958 49195977 E8 F2 TTGCACTTTGTGCCTTTGAA 49196027 49196046 E8 R2 CAGGAGTGGAGTGGCTGTG 598 49196607 49196625 E9 F2 TCGCTGGAAGGTTAACATCA 49197666 49197685 E9 R2 TTAGCCCTTCCACAGTTCCA 487 49198134 49198153 E10 F2 TTCCTTGGGGATGATTTTTG 49199057 49199076 E10 R2 TGTAGGCCCCTTCCCTCT 482 49199522 49199539

Table 2.3. Primers used for PTPN1 PCR amplification.

58 Normal Tumor MutSeq AA Gene Sample Chr. Position Ref. Variant var to ref var to ref Effect probability change ratio ratio AC019322.1 PM-7 15 30699678 G C 0.0 0.4 0.673 L→V Codon change AC110373.1 PM-2 4 120299324 C T 0.0 0.2 0.924 N/A Start gained AC118759.1 PM-7 7 100549540 A C 0.0 0.1 0.420 S→R Codon change ACBD3 PM-2 1 226334451 A G 0.0 0.8 0.997 Y→H Codon change ACIN1 PM-2 14 23548787 C T 0.0 0.2 0.476 R→H Codon change ACO1 PM-7 9 32440553 C T 0.0 0.1 0.955 R→* Stop gained AGAP6 PM-2 10 51768543 T C 0.0 0.1 0.478 S→P Codon change AL079342.1 PM-7 6 89675489 G A 0.0 0.2 0.577 R→H Codon change AMMECR1 PM-2 X 109507827 G T 0.0 0.4 0.868 Y→* Stop gained ANGEL2 PM-7 1 213180434 C A 0.0 0.2 0.659 E→* Stop gained ANK3 PM-7 10 61835571 A T 0.0 0.2 0.921 S→T Codon change ANKLE1 PM-2 19 17397487 G T 0.0 0.1 0.874 C→F Codon change ANKS1A PM-7 6 34950604 G A 0.0 0.4 0.988 G→R Codon change ARFIP1 PM-2 4 153809316 G T 0.0 0.2 0.998 E→* Stop gained ARHGAP22 PM-7 10 49658391 C T 0.0 0.3 0.685 R→H Codon change ARHGAP23 PM-2 17 36638820 G T 0.0 0.3 0.966 E→D Codon change ARMC6 PM-7 19 19166125 G A 0.0 0.2 0.939 D→N Codon change B2M PM-2 15 45003779 T C 0.0 0.3 0.853 L→P Codon change B2M PM-7 15 45003747 G A 0.0 1.5 0.640 M→I Start lost BBX PM-7 3 107451806 C T 0.0 0.3 1.000 P→L Codon change BCL6 PM-2 3 187463257 A G 0.0 0.3 0.990 N/A Start gained BCL7A PM-7 12 122458937 C G 0.0 0.2 0.446 N/A Start gained BCL7A PM-7 12 122459278 G T 0.0 0.2 0.941 N/A Start gained BCL7A PM-7 12 122459562 G T 0.0 0.3 0.709 N/A Start gained BCMO1 PM-2 16 81295827 C T 0.0 0.4 0.991 A→V Codon change BTG1 PM-7 12 92538143 C G 0.0 0.2 0.805 D→H Codon change

59 Normal Tumor MutSeq AA Gene Sample Chr. Position Ref. Variant var to ref var to ref Effect probability change ratio ratio C1orf106 PM-2 1 200881050 C T 0.0 0.4 0.929 R→C Codon change C2orf69 PM-2 2 200776430 C G 0.0 0.7 0.910 P→R Codon change C3orf72 PM-2 3 138669151 G A 0.0 0.1 0.931 A→T Codon change C8orf48 PM-2 8 13424727 A G 0.0 0.1 0.885 Y→C Codon change C9orf167 PM-2 9 140173764 C T 0.0 0.1 0.771 A→V Codon change CACNA1C PM-7 12 2719839 C T 0.0 0.8 0.878 L→F Codon change CASR PM-2 3 121981014 G A 0.0 0.3 0.996 E→K Codon change CCDC159 PM-2 19 11461615 G A 0.0 0.6 0.875 E→K Codon change CCDC160 PM-2 X 133371081 C T 0.0 0.2 0.886 N/A Start gained CCDC85A PM-2 2 56420501 G T 0.0 0.1 0.783 R→M Codon change CD300LG PM-2 17 41926135 C T 0.0 0.3 0.982 R→C Codon change CD37 PM-2 19 49842066 C T 0.0 0.3 0.718 A→V Codon change CDH11 PM-7 16 65022215 C T 0.0 0.4 0.987 V→I Codon change CDH9 PM-7 5 26915822 T G 0.0 0.3 0.996 I→L Codon change CHKA PM-7 11 67864568 C T 0.0 0.3 0.960 C→Y Codon change CHST1 PM-2 11 45672332 C T 0.0 0.1 0.807 E→K Codon change CIITA PM-2 16 11001430 T C 0.0 0.5 0.994 L→S Codon change CLCA4 PM-7 1 87045900 A G 0.0 0.2 0.479 T→A Codon change CLCA4 PM-7 1 87045901 C A 0.0 0.3 0.454 T→K Codon change CNOT6L PM-7 4 78697526 T C 0.0 0.4 0.938 E→G Codon change COL1A2 PM-7 7 94056525 C G 0.0 0.2 0.977 A→G Codon change CRIPAK PM-2 4 1388378 T C 0.1 0.4 0.434 C→R Codon change CRIPAK PM-2 4 1388379 G A 0.1 0.3 0.469 C→Y Codon change CUX2 PM-7 12 111537796 T A 0.0 0.6 0.980 F→I Codon change CYP4F31P PM-7 2 132045038 A C 0.0 0.2 0.424 *→C Stop lost DDX11 PM-7 12 31247694 G A 0.0 0.2 0.910 E→K Codon change

60 Normal Tumor MutSeq AA Gene Sample Chr. Position Ref. Variant var to ref var to ref Effect probability change ratio ratio DDX18 PM-7 2 118583047 A T 0.0 0.3 0.999 T→S Codon change DDX21 PM-7 10 70730072 C G 0.0 0.4 0.966 A→G Codon change DDX26B PM-7 X 134711239 G A 0.0 0.5 0.995 R→K Codon change DDX3X PM-7 X 41206199 C T 0.0 0.3 0.997 P→L Codon change DEFB132 PM-2 20 238438 G C 0.0 0.2 0.408 V→L Codon change DENND2A PM-2 7 140266918 C T 0.0 0.6 0.980 V→M Codon change DNAH7 PM-7 2 196722276 T C 0.0 0.7 0.989 R→G Codon change DNMBP PM-2 10 101716100 G C 0.0 0.8 0.975 D→E Codon change EEF1A1 PM-2 6 74229182 C G 0.0 0.2 0.957 E→Q Codon change EFTUD2 PM-2 17 42940102 C T 0.0 1.8 0.934 R→H Codon change EIF4EBP1 PM-2 8 37914770 G A 0.0 0.2 0.927 R→Q Codon change ELF3 PM-2 1 201979818 C T 0.0 0.4 1.000 N/A Start gained ELOVL3 PM-7 10 103988803 A G 0.0 0.4 0.970 T→A Codon change EMILIN1 PM-2 2 27306105 T C 0.0 0.6 0.991 F→L Codon change EPPK1 PM-2 8 144940331 C T 0.0 0.2 0.417 R→Q Codon change ETV2 PM-2 19 36135677 T G 0.0 0.6 0.899 Y→D Codon change EYA1 PM-2 8 72229822 T C 0.0 1.1 0.973 Q→R Codon change FAAH2 PM-7 X 57475024 A C 0.0 0.4 0.959 E→D Codon change FAM102A PM-7 9 130742301 T C 0.0 0.2 0.922 D→G Codon change FAM102A PM-7 9 130742660 C T 0.0 0.5 0.614 N/A Start gained FAM108C1 PM-7 15 81046559 G A 0.0 0.2 0.991 D→N Codon change FAM127C PM-2 X 134155743 A G 0.1 0.5 0.598 L→P Codon change FAM208B PM-2 10 5788796 G A 0.0 0.6 0.985 V→M Codon change FBXO32 PM-7 8 124526497 T C 0.0 0.7 0.978 E→G Codon change FBXO44 PM-7 1 11716146 C T 0.0 0.4 0.834 P→L Codon change FCGR3A PM-7 1 161518214 T C 0.0 0.2 0.411 I→V Codon change

61 Normal Tumor MutSeq AA Gene Sample Chr. Position Ref. Variant var to ref var to ref Effect probability change ratio ratio FGFR4 PM-2 5 176517136 C T 0.0 0.1 0.467 R→C Codon change FOXJ2 PM-2 12 8200876 A G 0.0 0.2 0.964 T→A Codon change FRAT2 PM-2 10 99093886 C T 0.0 0.2 0.626 W→* Stop gained FRG2B PM-7 10 135438943 A T 0.0 0.3 0.583 V→D Codon change FRG2B PM-7 10 135438961 C T 0.0 0.3 0.747 R→K Codon change FRG2B PM-7 10 135438980 C T 0.0 0.2 0.723 G→R Codon change GAA PM-2 17 78079596 C T 0.0 1.0 0.971 H→Y Codon change GABRB2 PM-2 5 160973619 C T 0.0 0.1 0.946 G→R Codon change GAGE12J PM-7 X 49179698 A G 0.0 0.5 0.667 Y→C Codon change GAL3ST1 PM-2 22 30953280 C T 0.0 0.5 0.993 V→M Codon change GALNT13 PM-2 2 154801105 A C 0.0 0.2 0.977 N→T Codon change GCNT2 PM-7 6 10529212 T C 0.0 0.3 0.994 V→A Codon change GIMAP8 PM-7 7 150171621 C T 0.0 0.3 0.995 R→W Codon change GLT8D1 PM-2 3 52738792 C T 0.0 2.7 0.870 N/A Start gained GNAI1 PM-7 7 79816533 G C 0.0 0.2 0.772 N/A Start gained GNAI1 PM-7 7 79816535 G T 0.0 0.2 0.750 N/A Start gained GNGT1 PM-7 7 93540144 C T 0.0 0.2 0.980 R→* Stop gained GRIN2C PM-2 17 72844051 C A 0.0 0.2 0.957 V→F Codon change GTF3C3 PM-2 2 197654612 A G 0.0 0.5 0.979 Y→H Codon change GUCA2B PM-7 1 42620493 C G 0.0 0.2 0.864 P→R Codon change GUCY1B3 PM-7 4 156715077 G C 0.0 0.2 0.986 E→Q Codon change HIP1 PM-2 7 75172194 C T 0.0 0.5 0.963 G→S Codon change HIST1H2AD PM-7 6 26199180 G C 0.0 0.2 0.977 L→V Codon change HIST1H3D PM-7 6 26197219 C G 0.0 0.2 0.991 S→T Codon change HMGCR PM-7 5 74654542 G A 0.0 0.4 1.000 V→I Codon change HNRNPA1 PM-7 12 54675190 G C 0.0 0.2 0.500 Q→H Codon change

62 Normal Tumor MutSeq AA Gene Sample Chr. Position Ref. Variant var to ref var to ref Effect probability change ratio ratio HP1BP3 PM-2 1 21100860 C A 0.0 0.2 0.487 N/A Start gained HSPA12B PM-2 20 3726191 G A 0.0 1.6 0.674 A→T Codon change HUWE1 PM-2 X 53616763 C T 0.0 0.6 0.994 R→Q Codon change HYDIN PM-7 16 71096118 C G 0.0 0.3 0.909 R→T Codon change IGLL5 PM-7 22 23230239 A T 0.1 0.2 0.572 R→S Codon change IGLL5 PM-7 22 23230318 C A 0.0 0.4 0.909 L→M Codon change IGLL5 PM-7 22 23230319 T A 0.0 0.3 0.805 L→Q Codon change IGLL5 PM-7 22 23235947 C T 0.0 0.3 0.894 P→S Codon change IGLL5 PM-2 22 23230278 G C 0.0 0.3 0.710 E→D Codon change IGLL5 PM-2 22 23230370 A T 0.0 1.0 0.846 Q→L Codon change IGLL5 PM-2 22 23230403 G C 0.0 0.8 0.986 G→A Codon change IGLL5 PM-7 22 23230234 A G 0.0 1.2 0.968 M→V Start lost IGSF21 PM-2 1 18703418 C T 0.0 0.4 0.995 S→F Codon change IL4R PM-2 16 27367183 T A 0.0 0.3 0.996 I→N Codon change IMMP2L PM-7 7 110603590 T A 0.0 0.1 0.973 K→* Stop gained IQCC PM-2 1 32673480 A G 0.0 0.2 0.995 T→A Codon change IQGAP2 PM-7 5 75927765 C T 0.0 0.3 0.987 T→I Codon change IRF4 PM-7 6 394899 T C 0.0 0.3 0.884 C→R Codon change IRF4 PM-2 6 394899 T C 0.1 0.7 0.590 C→R Codon change ITGB4 PM-7 17 73720812 C A 0.0 0.4 0.822 A→D Codon change ITPKB PM-7 1 226923853 C A 0.0 0.2 0.890 R→L Codon change ITPKB PM-7 1 226924418 C T 0.1 0.4 0.632 A→T Codon change KCNB2 PM-7 8 73480139 C T 0.0 0.8 0.910 P→L Codon change KCNMA1 PM-2 10 79397111 G A 0.0 0.6 0.448 T→I Codon change KCNN2 PM-7 5 113698842 C T 0.0 0.4 0.847 R→W Codon change KDM3A PM-2 2 86683637 C G 0.0 0.6 0.952 T→S Codon change

63 Normal Tumor MutSeq AA Gene Sample Chr. Position Ref. Variant var to ref var to ref Effect probability change ratio ratio KRT6B PM-7 12 52844302 C G 0.0 0.2 0.954 E→Q Codon change KRTAP5-5 PM-2 11 1651119 C T 0.0 0.2 0.405 R→C Codon change KRTAP5-5 PM-2 11 1651182 T C 0.0 0.3 0.407 C→R Codon change LARGE PM-2 22 33670475 C T 0.1 0.3 0.488 D→N Codon change LARP4B PM-2 10 876886 A T 0.0 0.6 0.949 L→* Stop gained LRBA PM-7 4 151738399 C T 0.0 0.4 0.993 V→I Codon change LRIT1 PM-2 10 85997197 G T 0.0 0.4 0.889 P→H Codon change LRRN3 PM-7 7 110763651 A G 0.0 0.3 0.998 I→V Codon change LRRN3 PM-7 7 110763667 T C 0.0 0.3 0.996 F→S Codon change LRRN3 PM-7 7 110763676 T C 0.0 0.2 0.996 M→T Codon change LRRN3 PM-7 7 110764432 A G 0.0 0.2 0.997 N→S Codon change LRRN3 PM-7 7 110764603 G C 0.0 0.4 0.996 C→S Codon change LRRN3 PM-7 7 110764926 G A 0.0 0.3 0.996 V→I Codon change LRRN3 PM-7 7 110731379 A G 0.0 0.2 0.854 N/A Start gained MADCAM1 PM-2 19 501714 C A 0.0 0.2 0.520 P→Q Codon change MAGI1 PM-7 3 65415650 G T 0.0 0.3 0.997 T→N Codon change MAP3K11 PM-2 11 65381715 C T 0.0 0.2 0.985 N/A Start gained MAST1 PM-2 19 12985095 C T 0.0 0.4 0.872 T→M Codon change MAVS PM-7 20 3845361 A C 0.0 0.3 0.857 K→Q Codon change MCF2 PM-7 X 138687089 G T 0.0 0.4 0.967 L→M Codon change MLL3 PM-7 7 151935853 T C 0.0 0.2 0.442 E→G Codon change MLPH PM-7 2 238395062 C T 0.0 0.2 0.973 N/A Start gained MMRN1 PM-2 4 90874285 G A 0.0 0.6 0.999 G→R Codon change MMS19 PM-2 10 99221845 G T 0.0 0.1 0.834 P→T Codon change MOB3C PM-7 1 47080398 G C 0.0 0.3 0.978 N/A Start gained MORC1 PM-2 3 108773722 C T 0.0 0.2 0.862 A→T Codon change

64 Normal Tumor MutSeq AA Gene Sample Chr. Position Ref. Variant var to ref var to ref Effect probability change ratio ratio MRPL45P2 PM-2 17 45567619 C T 0.0 0.7 0.780 R→H Codon change MSL1 PM-7 17 38289803 A G 0.0 0.4 0.987 I→V Codon change MUC6 PM-7 11 1017939 G C 0.0 0.1 0.413 T→S Codon change MYO3A PM-7 10 26243875 T G 0.0 0.2 0.999 F→V Codon change MYO7A PM-2 11 76870495 C T 0.0 0.5 0.991 R→C Codon change NAA16 PM-7 13 41932554 T C 0.0 0.3 0.999 I→T Codon change NALCN PM-7 13 101707802 A C 0.0 0.1 0.907 L→V Codon change NDUFS7 PM-7 19 1383557 G A 0.0 0.4 0.974 N/A Start gained NFIA PM-2 1 61743225 G A 0.0 0.2 0.975 S→N Codon change NFKB2 PM-7 10 104156064 G T 0.0 0.2 0.962 E→* Stop gained NFKBIE PM-7 6 44230382 A C 0.1 0.3 0.406 V→G Codon change NFKBIE PM-2 6 44227958 A T 0.0 0.2 0.806 V→E Codon change NFKBIE PM-7 6 44232759 G A 0.0 0.5 0.988 Q→* Stop gained NLRP6 PM-2 11 279352 C T 0.0 1.0 0.868 R→C Codon change NOX4 PM-7 11 89223789 C G 0.0 0.3 0.996 V→L Codon change NPW PM-2 16 2070164 C T 0.0 0.4 0.620 P→S Codon change NRG1 PM-7 8 32505352 C A 0.0 0.1 0.889 P→Q Codon change NSUN3 PM-7 3 93813003 A C 0.0 0.5 0.995 E→D Codon change OR5I1 PM-2 11 55703090 G A 0.0 0.8 0.950 R→W Codon change OSBPL10 PM-7 3 32022391 C T 0.0 0.3 0.783 R→K Codon change PCDHGA5 PM-7 5 140744082 G A 0.0 0.2 0.943 R→H Codon change PCDHGC4 PM-2 5 140865243 G A 0.0 0.2 0.997 S→N Codon change PDZRN4 PM-2 12 41831533 G A 0.0 0.8 0.995 N/A Start gained PELI3 PM-7 11 66241361 C T 0.0 0.2 0.678 R→W Codon change PFAS PM-7 17 8167249 C T 0.0 0.3 0.983 R→W Codon change PGAM4 PM-7 X 77225005 G A 0.0 0.3 0.914 A→V Codon change

65 Normal Tumor MutSeq AA Gene Sample Chr. Position Ref. Variant var to ref var to ref Effect probability change ratio ratio PHAX PM-7 5 125939475 C G 0.0 0.1 0.923 Q→E Codon change PIAS1 PM-2 15 68475989 T G 0.0 0.2 0.945 L→* Stop gained PIM1 PM-7 6 37139237 C T 0.0 0.3 0.864 L→F Codon change PLCB1 PM-2 20 8731494 C T 0.0 1.1 0.974 T→M Codon change PLCE1 PM-2 10 95930940 G A 0.0 0.7 0.972 R→H Codon change PLEKHG4 PM-7 16 67316522 C A 0.0 0.4 0.978 A→D Codon change PLXNA1 PM-2 3 126736445 C A 0.0 0.3 0.997 L→M Codon change POTEF PM-7 2 130832185 G A 0.0 0.4 0.637 R→W Codon change POU4F2 PM-7 4 147561722 G A 0.0 0.3 0.868 R→H Codon change POU5F1B PM-2 8 128428446 C T 0.0 0.3 0.760 P→L Codon change PRB4 PM-2 12 11461549 G C 0.0 0.3 0.440 P→R Codon change PRR5 PM-7 22 45132914 C A 0.0 0.3 0.823 C→* Stop gained PTPN1 PM-7 20 49195015 T A 0.0 0.3 0.995 V→D Codon change PTPN1 PM-2 20 49127125 C T 0.0 2.5 0.948 Q→* Stop gained PTPRC PM-2 1 198711361 T A 0.0 0.2 0.997 S→R Codon change PTPRN2 PM-7 7 157985057 C T 0.0 0.2 0.900 A→T Codon change RAD50 PM-2 5 131925461 T A 0.0 0.3 1.000 Y→N Codon change RBBP4 PM-7 1 33116920 G A 0.0 0.3 0.904 E→K Codon change RBM33 PM-7 7 155534775 A C 0.0 0.3 1.000 K→T Codon change REEP5 PM-2 5 112214523 G T 0.0 0.4 0.993 L→I Codon change RELB PM-2 19 45515490 G A 0.0 0.8 0.929 G→R Codon change RELB PM-2 19 45540675 C A 0.0 0.4 0.984 P→Q Codon change RFX1 PM-7 19 14104354 A C 0.0 0.4 0.918 I→S Codon change RIMS2 PM-7 8 105025710 A T 0.0 0.4 0.995 R→* Stop gained RLN3 PM-2 19 14141544 T A 0.0 0.7 0.832 D→E Codon change RNF19B PM-7 1 33429731 G C 0.0 0.2 0.807 H→D Codon change

66 Normal Tumor MutSeq AA Gene Sample Chr. Position Ref. Variant var to ref var to ref Effect probability change ratio ratio RP11- 1286E23.13 PM-7 4 9250376 C G 0.1 0.3 0.428 Y→* Stop gained RP11-45H22.3 PM-7 5 54252273 A T 0.0 0.2 0.960 S→R Codon change RPS7 PM-7 2 3624182 A G 0.0 0.4 0.979 K→E Codon change RPTOR PM-7 17 78796929 C T 0.0 0.3 0.971 R→* Stop gained RUSC2 PM-2 9 35546624 G A 0.0 1.3 0.645 G→S Codon change SCMH1 PM-2 1 41628548 G A 0.0 0.8 0.981 N/A Start gained SCN1A PM-2 2 166901803 C T 0.0 0.5 0.988 R→K Codon change SGK1 PM-7 6 134495206 C G 0.0 0.1 0.922 Q→H Codon change SH2D3A PM-7 19 6760926 C T 0.0 0.5 0.939 V→M Codon change SHANK1 PM-2 19 51165703 G A 0.0 0.8 0.956 S→L Codon change SIPA1 PM-2 11 65409979 G T 0.0 0.5 0.784 G→W Codon change SLC16A6 PM-2 17 66265331 C T 0.0 0.8 0.716 R→H Codon change SLC30A4 PM-7 15 45814190 G C 0.0 0.5 0.981 Y→* Stop gained SLC5A8 PM-2 12 101603410 C T 0.0 0.1 0.870 G→S Codon change SLC6A1 PM-2 3 11058881 C T 0.0 2.0 0.816 N/A Start gained SLITRK2 PM-7 X 144905808 T C 0.0 0.5 0.996 I→T Codon change SNTG1 PM-7 8 51314783 T A 0.0 0.2 0.980 I→N Codon change SOCS1 PM-7 16 11348953 A T 0.0 0.3 0.828 V→E Codon change SOCS1 PM-7 16 11349006 G C 0.0 0.1 0.540 N→K Codon change SOCS1 PM-2 16 11348998 A C 0.0 0.4 0.997 F→C Codon change SOCS1 PM-2 16 11349301 G T 0.0 1.0 0.913 A→E Codon change SOCS1 PM-2 16 11348834 G A 0.0 1.1 0.892 Q→* Stop gained SORT1 PM-7 1 109859482 G A 0.0 0.2 0.989 P→L Codon change STAB1 PM-7 3 52548786 T C 0.0 0.3 0.637 S→P Codon change STXBP2 PM-7 19 7705300 A G 0.0 0.1 0.456 Y→C Codon change SUCLG2 PM-2 3 67578606 A T 0.0 1.7 0.981 Y→N Codon change

67 Normal Tumor MutSeq AA Gene Sample Chr. Position Ref. Variant var to ref var to ref Effect probability change ratio ratio SULF1 PM-7 8 70379309 C T 0.0 0.1 0.862 N/A Start gained TAS2R31 PM-2 12 11183305 C G 0.0 0.2 0.475 Q→H Codon change TAS2R31 PM-2 12 11183035 C T 0.0 0.2 0.416 W→* Stop gained TAS2R39 PM-7 7 142880932 A C 0.0 0.3 0.995 N→H Codon change TBC1D17 PM-2 19 50390810 G A 0.0 0.1 0.972 D→N Codon change TFB1M PM-2 6 155578996 C T 0.0 0.3 0.974 D→N Codon change THAP2 PM-2 12 72070506 G A 0.0 0.1 0.504 S→N Codon change THAP2 PM-2 12 72070508 T A 0.0 0.1 0.606 C→S Codon change THOC5 PM-7 22 29925160 C A 0.0 0.3 0.998 D→Y Codon change THOC5 PM-7 22 29925161 C A 0.0 0.3 0.998 E→D Codon change TM7SF2 PM-2 11 64882463 C T 0.0 0.4 0.991 P→S Codon change TMEM132E PM-2 17 32961851 G T 0.0 0.4 0.999 E→D Codon change TMEM161B PM-2 5 87498803 G A 0.0 0.7 0.941 P→L Codon change TMEM2 PM-2 9 74305087 G A 0.0 0.1 0.423 R→C Codon change TMSB4X PM-7 X 12994393 C T 0.0 0.4 0.962 P→S Codon change TNFAIP3 PM-7 6 138192605 A T 0.0 0.5 0.984 K→* Stop gained TNFRSF19 PM-7 13 24167500 G C 0.0 0.4 0.990 C→S Codon change TOX2 PM-2 20 42694377 A G 0.0 0.9 0.602 K→R Codon change TP53TG5 PM-2 20 44006234 C T 0.0 0.2 0.992 R→Q Codon change TRAF1 PM-7 9 123690616 C G 0.0 0.3 0.973 N/A Start gained TRIM17 PM-7 1 228602789 C T 0.0 0.4 0.963 N/A Start gained TRIM32 PM-2 9 119460532 G T 0.0 0.2 0.856 E→* Stop gained TRPC6 PM-7 11 101375018 G A 0.0 0.3 0.639 H→Y Codon change TSPYL6 PM-2 2 54483209 C T 0.0 0.2 0.977 R→Q Codon change TSSK2 PM-7 22 19119045 C T 0.0 0.4 0.967 R→C Codon change UGT2B11 PM-2 4 70071270 C T 0.0 0.4 0.989 D→N Codon change

68 Normal Tumor MutSeq AA Gene Sample Chr. Position Ref. Variant var to ref var to ref Effect probability change ratio ratio UNK PM-7 17 73818677 G A 0.0 1.0 0.961 E→K Codon change VMP1 PM-7 17 57915701 G T 0.0 0.3 0.826 E→D Codon change VMP1 PM-7 17 57915699 G T 0.0 0.3 0.850 E→* Stop gained VRTN PM-7 14 74824808 G A 0.0 0.3 0.659 R→Q Codon change YY2 PM-2 X 21875564 G A 0.0 0.5 0.912 R→H Codon change ZBTB25 PM-7 14 64971296 G T 0.0 0.2 0.520 N/A Start gained ZCCHC24 PM-7 10 81146111 C T 0.0 0.4 0.432 R→H Codon change ZEB2 PM-2 2 145161517 C T 0.0 0.2 0.974 R→Q Codon change ZNF208 PM-7 19 22156811 T C 0.0 0.4 0.595 K→R Codon change ZNF208 PM-2 19 22155783 C T 0.0 0.2 0.508 V→I Codon change ZNF217 PM-2 20 52198459 T G 0.0 0.9 0.964 T→P Codon change ZNF236 PM-7 18 74627689 A T 0.0 0.3 0.757 M→L Codon change ZNF34 PM-2 8 145999617 T A 0.0 0.2 0.421 K→N Codon change ZNF554 PM-2 19 2834317 A G 0.0 0.7 0.885 T→A Codon change ZNF648 PM-7 1 182026341 C T 0.0 0.3 0.670 A→T Codon change ZNF654 PM-7 3 88189653 G A 0.0 0.3 1.000 R→H Codon change ZNF777 PM-2 7 149148161 C T 0.0 0.9 0.917 R→H Codon change ZNF814 PM-2 19 58384596 C G 0.0 0.7 0.976 C→S Codon change ZNF880 PM-2 19 52887771 G A 0.0 0.3 0.454 R→Q Codon change ZNF880 PM-2 19 52887780 A G 0.0 0.3 0.445 K→R Codon change ZNF880 PM-2 19 52887789 G C 0.0 0.2 0.500 S→T Codon change

Table 2.4. Somatic coding SNVs identified in two PMBCL index patients by Whole Genome sequencing.

Normal Tumor Gene Sample Chr. Position Ref. Variant var to ref var to ref P value ratio ratio ACBD3 PM-2 1 226334451 A G 0.0 0.6 <0.001 AMMECR1 PM-2 X 109507827 G T 0.0 0.5 <0.001 ANKS1A PM-7 6 34950604 G A 0.0 0.2 <0.001 B2M PM-2 15 45003779 T C 0.0 2.5 <0.001 BBX PM-7 3 107451806 C T 0.0 0.3 <0.001 CACNA1C PM-7 12 2719839 C T 0.0 0.5 <0.001 CD300LG PM-2 17 41926135 C T 0.0 0.2 <0.001 CD37 PM-2 19 49842066 C T 0.0 0.7 <0.001 CIITA PM-2 16 11001430 T C 0.0 0.5 <0.001 CNOT6L PM-7 4 78697526 T C 0.0 0.3 <0.001 DDX21 PM-7 10 70730072 C G 0.0 0.3 <0.001 DNAH7 PM-7 2 196722276 T C 0.0 0.3 <0.001 EEF1A1 PM-2 6 74229182 C G 0.0 0.2 <0.001 ETV2 PM-2 19 36135677 T G 0.0 0.6 <0.001 FBXO32 PM-7 8 124526497 T C 0.0 0.6 <0.001 FOXJ2 PM-2 12 8200876 A G 0.0 0.2 <0.001 GUCA2B PM-7 1 42620493 C G 0.0 0.3 <0.001 HIST1H2AD PM-7 6 26199180 G C 0.0 0.3 <0.001 HIST1H3D PM-7 6 26197219 C G 0.0 0.0 <0.001 HUWE1 PM-2 X 53616763 C T 0.0 0.5 <0.001 IGSF21 PM-2 1 18703418 C T 0.0 0.6 <0.001 IL4R PM-2 16 27367183 T A 0.0 0.5 <0.001 IRF4 PM-7 6 394899 T C 0.0 0.3 <0.001 ITPKB PM-7 1 226923853 C A 0.0 0.3 <0.001 ITPKB PM-7 1 226924418 C T 0.0 0.3 <0.001

70 Normal Tumor Gene Sample Chr. Position Ref. Variant var to ref var to ref P value ratio ratio KCNMA1 PM-2 10 79397111 G A 0.0 0.6 <0.001 KDM3A PM-2 2 86683637 C G 0.0 0.6 <0.001 LRIT1 PM-2 10 85997197 G T 0.0 0.1 <0.001 LRRN3 PM-7 7 110763651 A G 0.0 0.3 <0.001 LRRN3 PM-7 7 110763667 T C 0.0 0.3 <0.001 LRRN3 PM-7 7 110763676 T C 0.0 0.3 <0.001 MCF2 PM-7 X 138687089 G T 0.0 0.4 <0.001 MORC1 PM-2 3 108773722 C T 0.0 0.2 <0.001 MSL1 PM-7 17 38289803 A G 0.0 0.3 <0.001 NALCN PM-7 13 101707802 A C 0.0 0.3 <0.001 NFKB2 PM-7 10 104156064 G T 0.0 0.3 <0.001 NFKBIE PM-2 6 44227958 A T 0.0 0.3 <0.001 NLRP6 PM-2 11 279352 C T 0.0 0.5 <0.001 PCDHGA5 PM-7 5 140744082 G A 0.0 0.1 <0.001 PFAS PM-7 17 8167249 C T 0.0 0.3 <0.001 PIM1 PM-7 6 37139237 C T 0.0 0.3 <0.001 PLCE1 PM-2 10 95930940 G A 0.0 0.5 <0.001 POU4F2 PM-7 4 147561722 G A 0.0 0.2 <0.001 POU5F1B PM-2 8 128428446 C T 0.0 0.2 <0.001 PTPN1 PM-2 20 49127125 C T 0.0 1.8 <0.001 PTPN1 PM-7 20 49195015 T A 0.0 0.3 <0.001 PTPRC PM-2 1 198711361 T A 0.0 0.2 <0.001 RAD50 PM-2 5 131925461 T A 0.0 0.2 <0.001 RELB PM-2 19 45515490 G A 0.0 0.6 <0.001 RELB PM-2 19 45540675 C A 0.0 0.3 <0.001 RFX1 PM-7 19 14104354 A C 0.0 0.3 <0.001

71 Normal Tumor Gene Sample Chr. Position Ref. Variant var to ref var to ref P value ratio ratio RIMS2 PM-7 8 105025710 A T 0.0 0.5 <0.001 RPTOR PM-7 17 78796929 C T 0.0 0.3 <0.001 SHANK1 PM-2 19 51165703 G A 0.0 0.7 <0.001 SIPA1 PM-2 11 65409979 G T 0.0 0.4 <0.001 SOCS1 PM-2 16 11348834 G A 0.0 0.5 <0.001 SOCS1 PM-2 16 11348998 A C 0.0 0.2 <0.001 SOCS1 PM-2 16 11349301 G T 0.0 0.1 <0.001 SOCS1 PM-7 16 11348953 A T 0.0 0.1 <0.001 TFB1M PM-2 6 155578996 C T 0.0 0.2 <0.001 TM7SF2 PM-2 11 64882463 C T 0.0 0.2 <0.001 TNFRSF19 PM-7 13 24167500 G C 0.0 0.3 <0.001 TP53TG5 PM-2 20 44006234 C T 0.0 0.2 <0.001 UNK PM-7 17 73818677 G A 0.0 0.8 <0.001

Table 2.5. Deep amplicon sequencing validation of somatic coding SNVs identified in the two PMBCL index patients.

72 Normal Tumor AA Gene Sample Chr. Position Ref. Variant Var to ref Var to ref change Effect ratio ratio Codon C3orf77* PM-2 3 44329013 GC G 0.0 2.0 S→R change Codon CACNA1A PM-2 19 13318672 CCTG C 0.0 0.5 delQ deletion Codon GOLGA2 PM-7 9 131020795 GCCT G 0.0 0.6 delE deletion Codon HMGB1 PM-2 13 31037685 ACTT A 0.0 1.1 delK deletion Codon HUWE1 PM-2 X 53652852 T TGCCGGG 0.0 15.0 insPA insertion Codon IRF2BPL PM-2 14 77493791 CTGT C 0.0 1.6 delQ deletion Codon KCNMA1 PM-2 10 79397085 AGATGAT A 0.0 0.6 delII deletion Codon KCNN3 PM-2 1 154842199 G GGCTGCTGCTGCT 0.0 0.8 insSSSS insertion KRTAP2- Codon 2* PM-7 17 39211139 C CGCAGGGGGGCCGGCA 0.0 0.6 insAGPPA insertion Codon LCE1F* PM-2 1 152749036 GGGTGGT G 0.0 0.5 delGG deletion Codon MAML2 PM-7 11 95825374 TTGCTGCTGC T 0.0 1.1 delQQQ deletion Codon NFKBIE PM-2 6 44228237 T TG 0.0 0.8 N→T change Codon PAIP1 PM-2 5 43539076 GT G 0.0 0.6 R→S change Codon PRR21 PM-2 2 240982362 G GGCCGTGGATGAAGA 0.0 0.5 insGHIF insertion Codon PTX3 PM-2 3 157160666 AAG A 0.0 1.0 E→D change Codon SHANK1 PM-2 19 51171498 G GC 0.0 0.5 A→G change Codon SLC6A9 PM-2 1 44468005 CTTG C 0.0 0.5 delN deletion Codon ZNRF2 PM-2 7 30325341 A ACGG 0.0 1.4 insG insertion

* Validated by Sanger sequencing

Table 2.6. Somatic coding indels identified in two PMBCL index patients by Whole Genome sequencing.

(AF=Allelic frequency by amplicon sequencing, Del=deletion, *=translational stop, +=putative)

Table 2.7a. PTPN1 mutations in PMBCL patient and cell line samples.

(Del=deletion, SP1=Specificity protein 1 transcription factor, *=translational stop, +=putative)

Table 2.7b. PTPN1 mutations in HL patient and cell line samples.

Function Annotation P value Molecules

AGT, ARAF, BRAP, CAMKK2, CARD11, CXXC5, DAPK1, DUSP4, EGFR, GNA13, HMOX1, IL6, IL6R, IRF1, MAP3K1, MAP4K5, MAPK13, MAPKAPK2, Protein kinase cascade MAPKAPK5, MIER1, MYC, NDFIP1, NDFIP2, NEK11, NF1, NF2, NLK, NRGN, 0.000 P2RX7, PARK2, PDPK1, PIM2, PPM1L, PRKAG2, S100A13, SIK1, SOCS3, SOX2, SQSTM1, STAT1, STK17A, TGFA, TIFA, TLR4, TMED4, TNF, TNIP2, UBD, WNK1, ZDHHC17, ZFP36

CD80, CDKN1A, CXCL14, EGFR, ERBB3, EZR, F3, FOXO3, GDF15, Progression of tumor 0.000 MAPKAPK5 ,MET, MYC, NRP2, SAT1, SEMA3E, SERPINE1, SOX2, SPARCL1, SPP1, SRSF1, STAT3, TLR4, TNF, TNFSF10

ABCB4, AGT, BCL11B, CD274, CD80, CFB, CFH, CNR1, CR1, CXCR5, CYSLTR1, DARC, ETS1, F3, FABP4, FN1, GFAP, HMOX1, HSPD1, ICAM1, Infiltration of leukocytes IL12B, IL21, IL2RA, IL33, IL6, IL6R, IL9, ITGA4, LIMK1, LTA, MAPKAPK2, 0.000 MMP12, MTAP, NPC1, P2RX7, PGF, PIK3R1, PTGER3, RARA, SERPINE1, SOCS3, SPP1, STAT1, STAT3, STAT5A, STAT5B, STC1, TGFBR1, THBS1, TLR4, TNF, TNFAIP3, TNFRSF8

Delay in cell cycle progression 0.007 AHR, CDKN1A, CLIP1, DBF4B, IL6, IL6R, KRT14, MYC

Development of helper T lymphocytes AHR, BCL6, ICAM1, IL12B, IL6, IRF1, RORA, SOCS3, STAT1, STAT3, STAT4, 0.007 STAT5A

B-cell non-Hodgkin's disease BCL6, BIRC3, BRD2, CARD11, CDKN1A, CXCR5, ETS1, GNA13, HMGA1, 0.008 IKZF3, IL6, IRF4, LTA, MCL1, MDM2, MYC, PIM2, PRDM2, SKIL, STAT3, TNF, TNFAIP3, TNIP2, TUBB4A, TUBD1,

Table 2.8a. Function of genes up-regulated in PTPN1 knock down cells (Ingenuity pathway analysis).

Canonical Pathways P value Molecules

Role of JAK in IL-6 type cytokine ABCB1, AKT3, CD14, IL6, IL33, IL1RAPL1, IL6R, IL6ST, JUN, MAPK13,

signaling MAPKAPK2, MCL1, PIK3C2A, PIK3R1, PTPN11, SOCS3, SOS2, STAT1, 0.001 STAT3, STAT5A, STAT5B, TNF

T helper cell activation BCL6, CD40, CD80, CXCR5, HLA-DOB, IL6, IL21, IL12B, IL2RA, IL6R, 0.002 IL6ST, STAT1, STAT3, STAT4, TGFBR1, TNF

ACTA2, AGT, CD14, CD40, EGFR, FN1, ICAM1, IGFBP5, IL6, IL1RAPL1, Hepatic Fibrosis / Hepatic Stellate 0.002 IL6R, LAMA1, LY96, MET, MYH1, MYH2, MYH8, MYH10, PDGFRA, PGF, Cell Activation STAT1, TGFA, TGFBR1, TLR4, TNF

JAK/STAT signaling AKT3, CDKN1A, IL6, PIK3C2A, PIK3R1, PTPN11, SOCS2, SOCS3, 0.004 SOCS7, SOS2, STAT1, STAT3, STAT4, STAT5A, STAT5B

IL-9 signaling IL9, PIK3C2A, PIK3R1, SOCS2, SOCS3, STAT1, STAT3, STAT5A, 0.005 STAT5B, TNF

Prolactin signaling 0.006 IRF1, JUN, MYC, PDPK1, PIK3C2A, PIK3R1, PTPN11, SOCS2, SOCS3, SOCS7, SOS2, STAT1, STAT3, STAT5A, STAT5B JAK2 mediated cytokine signaling 0.013 GHR, PTPN11, SOCS2, SOCS3, SOCS7, STAT1, STAT3, STAT5A, STAT5B IL-22 signaling 0.026 AKT3, MAPK13, SOCS3, STAT1, STAT3, STAT5A, STAT5B

Table 2.8b. Involved pathways of genes up-regulated in PTPN1 knock down cells (Ingenuity pathway analysis).

Function Annotation P value Molecules

Differentiation of B lymphocytes ATM, AVPR1A, CD69, FOXP1, HDAC9, IFNA14, IFNA16, IFNA7, IGHM, IGKC, 0.003 IGLL1, IGLL5, IL7, MYB, POU2AF1, SOX4, TLR7, TNFSF4, TSHR

Quantity of B-1 lymphocytes AVPR1A, CD19, IGHM, IGLL1, IGLL5, IL7, MBL2, MS4A1, MYB, SAMSN1 0.007

Quantity of pre-B1 lymphocytes 0.008 IGHM, IGLL1, IGLL5, IL7

ALOX5, ANXA1, C5, CD58, CD69, CHST4, CIITA, COL1A1, CXCL10, CXCL11, Movement of mononuclear 0.008 IL7, ITGB1, PLAU, PLG, POU2AF1, PRKCZ, S100A7, S100A7A, SELL, SELP, leukocytes SERP1, SFTPC, SMAD3, TIAM1, TNFSF4, TSHR, WNT5A

CXCL10, HLA-DQA1, HLA-DQB1, IFNA14, IFNA16, IFNA7, IGHM, IL7, LRP1, Activation of antigen presenting cells 0.019 PLG, PTGER3, SPACA3, TLR7, TPO, TSHR, WNT5A

Binding of neutrophils 0.024 C5, CXCL10, ERG, PLG, SELL, SELP

ATM, CCND1, GATA3, HLA-DQA1, HLA-DQB1, IGHM, IGKC, IGLL1, IGLL5, Abnormal morphology of leukocytes 0.039 IL31RA, IL7, POU2AF1, PTPN22, SELP, SERPINB9, TNFRSF11A

Table 2.8c. Function of genes down-regulated in PTPN1 knock down cells (Ingenuity pathway analysis).

81 Canonical Pathways P value Molecules

B cell development 0.004 CD19, HLA-DQA1, HLA-DQB1, IGHM, IGKC, IGLC1, IL7

Pattern recognition receptors of 0.026 ATM, C5, C3AR1, MAPK10, MBL2, OAS2, PRKCZ, PRKD3, PTX3, TLR7 bacteria and viruses

Table 2.8d. Involved pathways of genes down-regulated in PTPN1 knock down cells (Ingenuity pathway analysis).

83 Figure 2.1. Results from whole-genome sequencing of two PMBCL tumors and germline DNA. (a) The outermost ring depicts the chromosome ideogram oriented clockwise, p terminus to q terminus. Centromeres are shown in red. The second and third rings show copy number losses (blue) and gains (red) in PM-7 and PM-2, respectively. Genomic rearrangements are represented by arcs between chromosomes. Selected SNVs found in a single tumor library or both libraries are annotated with smaller and larger circles, respectively. (b, c) Genome-wide intermutational distances in PM-2 (b) and PM-7 (c). (d, e) Distribution of somatic coding-sequence transition and transversion nucleotide interchanges in PM-2 (d) and PM-7 (e). The absolute number of nucleotide substitutions is shown above each bar. (f) Mutated genes in the JAK-STAT pathway discovered by whole-genome and whole-transcriptome sequencing (RNAseq was performed on three PMBCL cell lines and seven tumor samples (PM 1-7), and whole-genome sequencing was performed on PM-2, PM-7 and on matched peripheral blood samples from the same individuals.

(+, previously reported154)

Figure 2.2. Locations of PTPN1 mutations in HL and PMBCL. (a) Distribution of PTPN1 mutations in PMBCL cell lines (n=3), PMBCL clinical samples (n=77), HL cell lines (n=9) and HL clinical samples (n=30) identified by targeted Sanger sequencing. Variations in noncoding regions, reported SNPs and silent mutations are not shown. (b) Frequently mutated Q9 and Q21 residues encoded by exon 1. Q21 is encoded by part of a splice donor site on this exon. (+, putative translation).

Figure 2.3. Immunohistochemical analysis of PTP1B expression in lymph node biopsy samples from HL (n=215) and PMBCL (n=143). Representative images of sections stained with a monoclonal antibody against PTP1B are shown ((original magnification 400x (scale bars, 25 μm; a-e) and 200x (scale bar, 50 μm; f)). (a-c) PMBCL samples expressing wild-type PTP1B (PM-15) (a), G381S PTP1B (PM-9) (b) and Q21* PTP1B (PM-2) (c). (d, e) HL samples expressing wild-type PTP1B (HD-99) (d) and delM109 PTP1B (HD-42) (e). (f) Normal tonsil tissue with a germinal center shown for comparison.

87 Figure 2.4. PTP1B expression in PMBCL (n=143) and in HL (n=215) clinical specimens were analyzed by immunohistochemistry. For PMBCL (a) and HL (b) cases with known PTPN1 mutational status, the percentage of positive tumor cells is shown. The overall distribution of PTP1B positive tumor cells are shown in Figures c and d.

89 Figure 2.5. PTPN1 mutations and survival outcomes in PMBCL patients. The Kaplan-Meier graphs show overall survival (a) and progression-free survival (b) in 77 PMBCL patients (60 PTPN1 WT and17 mutated).

91 Figure 2.6. PTPN1 mutations and survival outcomes in HL patients. The Kaplan-Meier graphs show freedom from treatment failure (a) and disease-specific survival (b) in 30 HL patients (24 PTPN1 WTand 6 mutated).

93 Figure 2.7. PTPN1 allelic imbalances in PMBCL and HL cell lines were analyzed by FISH. BAC clone (CTD-2582P13) for PTPN1 20q13.3 is labeled in spectrum orange and telomere 20p13 is labeled in spectrum green (control). Cell lines with normal PTPN1 copy numbers are not shown.

95 Figure 2.8. PTP1B expression and PTPN1 transcript levels in PMBCL and HL cell lines were analyzed by protein blotting and qRT-PCR. 25 g of extracted proteins were incubated with anti-PTP1B. An EBV transformed lymphoblastoid cell line (LCL) was used as a positive control and an anti--actin antibody was used as a loading control. A representative experiment is shown (a). 5 ng of extracted RNA was used to detect PTPN1 mRNA usingTaqman gene expression assay probes. GAPDH was run as an internal control. Each sample was run in triplicate. Error bars represent standard deviation (b).

Figure 2.9. Mutations in MedB-1 are in trans-allelic configuration. cDNA from MedB-1 was cloned into a vector (pCR 2.1TOPO TA cloning; Invitrogen, Carlsbad, CA, USA). Sanger sequencing of subclones revealed PTPN1 mutations in a trans-allelic configuration. All clones screened had mutations in either Exon 5 or 8.

Figure 2.10. Expression of PTP1B mutants in HEK 293-STAT6 cells. (a) Levels of phosphorylated STAT6 (pSTAT6) in HEK 293-STAT6 cells transfected with constructs expressing wild-type or mutant PTP1B determined by protein blotting. Protein lysates were prepared from cell cultures treated with or without 20 ng/ml IL-4. Relative densitometric values (shown for each band) were normalized against the levels of the internal control β-actin. (b) Activated STAT6- dependent SEAP reporter gene expression in transfected cells. Each sample was run in triplicate. Each value represents the average of two independent experiments. Error bars, s. d. (c) PTP1B activity levels in lysates from transfected cells without IL-4 and determined by catalysis of the tyrosine phosphatase substrate, DADEY(PO3)LIPQQG. Each sample was run in triplicate. Each value represents the average of two independent experiments. Error bars, s. d. (MT, mock expression vector only; ***P<0.001, **P<0.01 and *P<0.05 compared to wild type).

100 Figure 2.11. Activation of JAK-STAT in a PTPN1-silenced HL cell line (KM-H2). (a) Quantitative PCR performed on RNA isolated from PTPN1 knockdown and non-silencing cells using Taqman assay probes. Each sample was run in triplicate. Each value represents the average of three independent experiments. Error bars, s. d. (b) Hyperphosphorylation of JAK (pJAK1, pJAK2) and STAT (pSTAT3, pSTAT5, pSTAT6) proteins determined by protein blotting in the KM-H2 cell line with shRNA-mediated knockdown of PTPN1. Protein lysates from non-silencing (NS) and knockdown (KD) cells were prepared from cell cultures treated with or without 10 ng/ml IL-4. Relative densitometric values (shown for each band) were normalized against levels of the internal control β-actin. (c) Gene expression profiling of RNA isolated from PTPN1 knockdown and non-silencing cells on an HG U133 Plus 2.0 array. Differentially expressed genes were compared against 50 known STAT3-responsive genes. Thirty eight genes overlapped with our gene cohort (P =0.0003). Only 2 out of 10,000 random permutations had a value of ≥38 (P =0.0002). Fold- change differences were compared against the non-silencing control.

101

102 Figure 2.12. Gene set enrichment analysis of differentially expressed genes in PTPN1 silenced cells using GSEA software and Molecular Signatures Database (Broad institute, Cambridge, MA, USA).

103

Chapter 3: Recurrent somatic IL4R mutations in

primary mediastinal large B cell lymphoma

104

3.1. Introduction

As described in Chapter 2, we interrogated the mutational landscape in the transcriptomes of 7 PMBCL patients and 3 PMBCL-derived cell lines by massively parallel next-generation sequencing and found a hotspot mutation in exon 8 of IL4R227. Interleukin (IL) 4 is a key inflammatory cytokine that binds to a transmembrane receptor complex consisting of IL4 receptor alpha (IL4Rα) and IL13Rα chains or IL4Rα and the common gamma chain (IL2RG)228. Activation of IL4R by IL-4 (or IL-13) initiates intracellular signal transduction mediated by the phosphorylation of JAnus Kinase- Signal Transducer and Activation of Transcription (JAK-STAT) pathway. In hematopoietic tissue, activated IL4R drives the differentiation of naive T helper (Th) cells into Th2 lymphocytes and induces ε isotype switching, immunoglobulin (Ig) E secretion and Major Histocompatibility Complex (MHC) class II expression in B lymphocytes229. Constitutively active JAK-STAT is a hallmark feature of PMBCL75. We and others have identified loss-of-function mutations in a phosphatase227 and other negative regulators163-164 that lead to activation of this pathway. Given the recurrent hotspot mutation discovered by transcriptome sequencing, we hypothesized that gain-of- function IL4R mutations also contribute to activated JAK-STAT in PMBCL. We have found additional IL4R mutations, confirmed the hotspot lesion in exon 8 and established the mutational frequency in an extension cohort of PMBCL cases. We functionally demonstrate that the hotspot mutation hyperphosphorylates IL4R, JAK2, STAT5 and STAT6, and activate the T cell regulatory chemokine TARC (CCL17) and the B cell activation marker CD23. Until now, recurrent somatic coding-sequence mutations of IL4R had not been described in any lymphoma. Our data suggest that gain-of-function mutations contribute to the pathogenesis of PMBCL through activation of JAK-STAT signaling.

105

3.2. Materials and methods

Patient samples and cell lines Specimens from 62 PMBCL patients were selected from the tissue archives of the Centre for Lymphoid Cancer of the British Columbia Cancer Agency according to the availability of fresh-frozen lymphoid tissue biopsy material and clinical patient follow- up data. The clinical characteristics of all patients stratified by IL4R mutational status are shown in Table 3.1. Fifty five (89%) patients have been previously reported as part of a published study227. Ethical approval for this study was obtained from the University of British Columbia – British Columbia Cancer Agency Research Ethics Board (UBC BCCA REB) in accordance with the Declaration of Helsinki. Cell lines U-2940, KARPAS-1106P and DEV were obtained from the German Collection of Microorganisms and Cell Cultures and propagated according to standard conditions (DSMZ, Braunschweig, Germany, http://www.dsmz.de). The cell line MedB-1 was a kind gift from Drs. S. Brüderlein and P. Möller (University of Ulm, Germany) and propagated as published193.

Screening for IL4R somatic mutations IL4R mutations detected by deep amplicon and next-generation sequencing were validated by Sanger sequencing (n=13). All 9 IL4R coding exons were amplified by standard PCR (Invitrogen Carlsbad, CA, USA) and Sanger-sequenced (3130 Genetic Analyzer, Applied Biosystems, Foster City, CA, USA). Primer sequences used for amplification are listed in Table 3.2. Mutation analysis was performed using Clone manger software (Scientific & Educational Software, Cary, NC, USA). For deep amplicon sequencing, we used the Illumina Two-step PCR protocol according to the manufacturer¹s instructions covering the complete protein coding sequence of IL4R. Amplicon libraries were sequenced on an Illumina MiSeq instrument using V2 500- cycles MiSeq reagent kit (Illumina), generating 250bp paired-end reads. Sequencing achieved a mean on-target coverage of 30,568x (range 3329.9x to 37226.3x). Alignment and variant detection was performed using Mutascope230.

106

Flow cytometry Flow cytometry analysis for surface expression of IL4Rα (CD124) and CD23 was performed using a LSR Fortessa special order system (Becton-Dickinson Biosciences, San Jose, CA, USA) as previously described231. Briefly, 500,000 live cells were collected and stained with either PE-conjugated mouse monoclonal IgG1, k anti-CD23 (BioLegend, San Diego, CA, USA; clone EBVCS-5) or APC-conjugated mouse monoclonal IgG2a, k anti-CD124 (BioLegend; clone G077F6) in 100 ul of 1x PBS 1% heat-inactivated FBS (1% HI-FBS) for 30 min on ice. Samples were subsequently washed once before re-suspension in 1% HI-FBS. One hundred thousand events were acquired at a rate of approximately 1,000 events per second. Events were gated using forward and side light scatter excluding doublet events. Appropriate isotype controls (BioLegend) were also used to assess wild-type levels of expression in each cell line. Results were analyzed using Flow Jo (version 9; Ashland, OR, USA).

Fluorescence in-situ hybridization (FISH) analysis For IL4R copy number detection, a dual-color probe was designed using in- house BAC clones: RP11-1073N1 for IL4R (16p12.1) labeled in spectrum orange and RP11-151M19 as a control probe (16q22.1) labeled in spectrum green. Slides were analyzed using a Zeiss Axioplan 2 fluorescence microscope (Zeiss Gollingen) and documented using an ISIS imaging system (Metasystems). At least 100 metaphase and interphase nuclei were scored.

Quantitative RT- PCR Taqman gene expression assay probes were used to detect mRNA levels of IL4R (Hs00166237_m1)) and CD23 (Hs01077044_m1) on a 7900HT real-time PCR system (Applied Biosystems, Foster city, CA, USA). GAPDH was run as an internal control. Measurements were quantified using the ∆∆CT method (Pfaffl) and expressed relative to the expression in empty vector cells.

107

Expression of IL4R, site-directed mutagenesis and SEAP reporter gene assay The wild type (WT) IL4R coding sequence was amplified by PCR using cDNA from Karpas-1106 and cloned into the mammalian expression vector, pcDNA3.1 (Invitrogen, Carlsbad, CA, USA). The IL4R I242N mutation was created using the Quik change XL site-directed mutagenesis kit (Agilent Technologies, Santa Clara, CA, USA) accordingly to manufacturer instructions. Empty pcDNA3.1 (MT) was used as a mock- vector. The plasmids were purified by Spin miniprep kit (Qiagen, Mississauga, ON, Canada) and 3 g of each plasmid was transfected into human embryonic kidney (HEK) 293 cells expressing STAT6 (HEK 293-STAT6; HEK blue IL4/IL13, Invivogen, San Diego, CA, USA) using Lipofectamine 2000 (Invitrogen Carlsbad, CA, USA). Transfected cells were cultured for 48 h and secreted embryonic alkaline phosphatase (SEAP) levels assayed in culture supernatant according to the manufacturer protocol. A GFP plasmid was used to determine equal transfection efficiency (>90%).

Preparation of doxycycline-inducible cell lines and retroviral transduction Retroviral transduction of DEV cells was performed as previously described232. In brief, wild type DEV cells were first transduced with a feline endogenous virus (FEV) expressing the ecotropic retroviral receptor (DEV-FEV). The FEV was constitutively produced in the supernatant of the producer line FLYRD18. This ecotropic receptor- expressing cell population was secondarily infected with a retrovirus expressing the bacterial tetracycline repressor (DEV-FEV-TETR). DEV-FEV-TETR cells were subsequently infected with retroviral particles containing the inducible vector pRETRO- TO-PuroGFP (a kind gift from Dr. Louis Staudt) in which the CMV promoter is used to drive expression of cloned IL4R WT or I242N in the presence of 50 ng/ml of doxycycline (D9891; Sigma-Aldrich, Oakville, ON, Canada). For retroviral particle production, cloned IL4R constructs and the mutant ecotropic envelope-expressing plasmid pHIT/EA6x3 were co-transfected into HEK 293T cells using the Lipofectamine 2000 reagent. Retrovirus containing supernatant was collected 48 h after transfection and DEV-FEV-TETR cells were infected in the presence of 8 μg/ml polybrene in a single spin infection (2500 RPM for 1.5 h) and puromycin was used to select for stable integrants over 6 days.

108

Cell lysis, protein blotting and immunoprecipitation Total cell lysates were prepared from cultured cell lines treated with or without 20 ng/ml recombinant human IL-4 (R&D Systems, Minneapolis, MN, USA) for 20 minutes at 37˚C using RIPA protein extraction buffer (Thermo Scientific, Waltham, MA, USA) or NativePAGE sample dye (for native protein electrophoresis; Invitrogen, Carlsbad, CA, USA) and in the presence of a protease inhibitor cocktail (Sigma-Aldrich, Oakville, ON, Canada). 25 g of protein lysates were resolved on a 4-12% NuPAGE Novex Bis-Tris gradient gel (Invitrogen) and transferred to a nitrocellulose membrane (Thermo Scientific, Waltham, MA, USA) by semi-dry transfer (Bio-Rad, Mississauga, ON, Canada) and probed with the following primary antibodies at 1:1,000 dilutions unless noted otherwise: JAK1 (3332), phspho-JAK1 (3331), JAK2 (3230), phospho-JAK2 (3776; 1:500), JAK3 (8827), phspho-JAK3 (5031), STAT3 (9132), phospho-STAT3 (9134), phospho-STAT5 (9359), phospho-STAT6 (9364) (all from Cell Signaling Technology, Danvers, MA, USA), IL4R (R&D Systems; 1:500), STAT5 (sc-835; 1:2,000) and IL2RG (sc-365910) (from Santa Cruz Biotechnology, Santa Cruz, CA, USA), pIL4R (ab61099; 1:500), STAT6 (ab32108) and IL13R (ab129188) (from Abcam, Cambridge, MA, USA). β-actin antibody (A5441; 1:10,000; Sigma-Aldrich, Oakville, ON, Canada) was used as an internal control. Appropriate horse radish peroxidase (HRP) conjugated anti-rabbit or anti-mouse IgG (W401B and W402B respectively, both 1:5,000; Promega, Madison, WI, USA) secondary antibodies were used to visualize bands using the enhanced chemiluminiscence (ECL) system (Amersham, Baie d’Urfe, QC, Canada) on a Chemidoc digital imager (Biorad, Hercules, CA, USA). Intensities of developed bands were quantified using Image Lab software (Biorad, Hercules, CA, USA). For immunoprecipitation experiments, cell lysates were incubated with 1 g of the IL4R antibody overnight and precipitated using protein A/G ultralink resin (53132;Thermo Scientific) for 2 h at 4˚C without crosslinking. The bound material were resolved by electrophoresis and processed as described above.

Enzyme-linked immunosorbent (ELISA) assays Soluble IL4R and secreted TARC levels in cell culture supernatants were quantified using IL4R human ELISA kit (R&D Systems Minneapolis, MN, USA) and

109

TARC/CCL17 ELISA kit (Sigma-Aldrich, Oakville, ON, Canada) respectively, according to manufacturer’s instructions.

Inhibitors Cultured cells were treated with varying doses of Pimozide (16222), Pyridone 6 (15146) and Pacritinib (SB1518; 16709) (all from Cayman Chemical, Ann Arbor, MI, USA) for 3 h and cell lysates were prepared as described above. All compounds were solubilized in dimethyl sulfoxide (Sigma-Aldrich, Oakville, ON, Canada).

Whole-transcriptome sequencing (RNAseq) RNAseq was performed as previously described using RNA extracted from DEV cells expressing IL4R WT and I242N186. In brief, double stranded cDNA was synthesized from polyadenylated RNA and sheared. The 190-210 bp fraction was isolated and amplified with 10 cycles of PCR using the Illumina Genome Analyzer paired end library protocol. The resulting RNAseq libraries were then sequenced on an 201 Illumina Genome Analyzerii and aligned to hg19 using GSNAP . Differentially expressed genes between DEV WT and I242N were reported using DESeq2233 (version 1.6.3) based on the following criteria: (i) analysis limited to protein coding region of genes, (ii) genes with a fold change of ≥2 or ≤0.5 were considered as being differentially expressed and reported and (iii) q≤0.01 was considered significant (Benjamini- Hochberg FDR method).

Statistical analysis Comparisons between groups were performed using two-sample student t tests. Time-to-event analyses were performed using the Kaplan-Meier method, and survival curves were compared by the log-rank test using SPSS Version 14.0. Overall survival (OS) was defined as the time from diagnosis to death from any cause. Progression-free survival (PFS) was defined as the time from diagnosis to progression (relapse after primary treatment, initiation of new treatment or death from any cause).

110

3.3. Results

IL4R is frequently mutated in PMBCL. We screened the complete coding sequence of IL4R, compromising 9 exons, for genomic mutations in 65 PMBCL samples by Sanger sequencing and deep amplicon sequencing. In total, after the exclusion of reported single nucleotide polymorphisms (SNPs) and silent mutations, we found 21 variants (16 mutations in 15 of 62 clinical samples and 5 mutations in 3 of 3 cell lines screened; Table 3.3), with some cases harboring multiple mutations. In summary, we found coding sequence mutations in 18 of 65 (28%) PMBCL clinical samples and PMBCL-derived cell lines. Most observed mutations (94.4%) were missense mutations except one observed frameshift mutation (5.5%) that putatively leads to premature IL4R protein termination. The distribution and frequency of mutations for each exon and the types of mutations observed are shown in Figure 3.1a. A strikingly recurrent hotpot mutation in exon 8 (hg19: chr. 16: 27,367,183; c.725T>A) substituting an isoleucine residue to an asparagine at amino acid position 242 (p.I242N) was observed in 11 of the 18 (61%) mutated cases. Mutation frequencies in PMBCL clinical cases and PMBCL cell lines were 24% and 100%, respectively. We confirmed 3 identified IL4R mutations (cases PM-2, PM-8 and PM-15) as somatic by sequencing constitutional DNA extracted from peripheral blood (Figure 3.1b). We next sought to determine if IL4R mutations were associated with patient survival. Kaplan-Meier survival analysis of PMBCL patients showed no significant differences in clinical outcome between groups with WT and mutated IL4R (P>0.05; Figure 3.2). Furthermore, IL4R mutations occurred independent of PTPN1 mutations in our study cohort and patients harboring mutations in both genes (n=4) showed no significant differences in survival comapared to patients with a single gene mutation or patients with no mutations in either gene (P>0.05). To evaluate if IL4R proteins are expressed on the surface of PMBCL cell lines harboring IL4R mutations, we performed flow cytometry in Karpas-1106P, MedB-1 and U-2940 cell lines. All 3 PMBCL-derived cell lines showed cell surface IL4R (CD124) expression that was increased compared to peripheral blood or a lymphoblastoid cell line (LCL; Figure 3.3a). To determine if focal copy number aberrations contribute to

111 increased protein expression, we performed fluorescence in-situ hybridization (FISH) but found no copy number changes in IL4R (Figure 3.3b).

IL4R I242N mutation increases phosphorylation of STAT6. To study the functional relevance of the discovered IL4R I242N mutation, we ectopically expressed IL4R WT or the I242N mutant in engineered HEK 293 cells expressing STAT6 (HEK 293-STAT6) and a STAT6-inducible reporter gene (SEAP). The efficiency of expression was determined by quantitative RT-PCR (qRT-PCR; WT 319,626% and I242N 257,531% compared to cells transfected with empty (MT) vector alone; Figure 3.4a). To determine if the mutant affected the phosphorylation of IL4R and STAT6, we analyzed cell lysates by protein blotting. We also treated cells with recombinant IL-4 in order to stimulate the JAK-STAT pathway to enhance detectable changes in phosphorylation. Compared to HEK 293 cells expressing IL4R WT, we observed increased phosphorylation of IL4R in the mutant (relative densitometric values of 2.6 versus 1.0 and 2.6 versus 1.4 with stimulation) and IL-4 independent phosphorylation of STAT6 (0.5 versus no phosphorylation in WT; Figure 3.4b) indicating constitutive activation of the JAK-STAT pathway. To elucidate if this activation is mediated by the partner receptors that dimerize with IL4R, we co-expressed IL4R, IL2RG and IL13R in HEK 293 cells and confirmed expression of each receptor by qRT- PCR (Figure 3.5a). Cytokine-independent phosphorylation of STAT6 was seen in protein lysates of co-transfected cells irrespective of the presence of partner receptors (Figure 3.5b). Analysis of activated STAT6-dependent reporter gene expression in cells expressing the mutant showed elevated SEAP levels compared to WT cells, either in the absence of partner receptors (percentage of SEAP expression: 846% versus 331%) or when partner receptors were present (820% versus 385%; Figure 3.5c), confirming results seen by protein blotting. We next sought to determine if the canonical STAT6 binding sites in IL4R are engaged by the mutant to activate STAT6 signaling. To determine if IL4R WT and IL4R I242N use the same tyrosine residues at 575, 603 and 631 for STAT6 phosphorylation, all 3 amino acids were mutated into phenyalanine residues by site-directed mutagenesis, transfected into engineered HEK 293 cells and SEAP gene expression

112 was quantified (Figure 3.6a). As expected, mutagenesis in WT lowered STAT6 activation with IL-4 stimulation (498% versus 369%). In contrast, mutagenesis in I242N resulted in a moderate yet significant increase in STAT6 activation with IL-4 stimulation (1534% versus 1741%; Figure 3.6b), indicating recruitment of alternative STAT6 binding sites by the mutant or IL13R compensating to elicit downstream signaling. Moreover, IL13R may activate STAT6 in the absence of IL4R as seen by increased SEAP expression in the cytokine-stimulated MT vector control.

IL4R I242N mutation hyperphosphorylates JAK-STAT proteins and activates CD23 and TARC in lymphoma cells. To evaluate if the hotspot mutation in IL4R also activates the JAK-STAT pathway in a B cell lymphoma cell line, we generated an IL4R I242N over-expressing nodular lymphocyte predominant HL cell line (DEV-I242N) by retroviral transduction of an inducible vector harboring the mutation (Figure 3.7a). Expression levels of WT and mutant IL4R levels are shown in Figure 3.7b. Despite slightly lower mutant IL4R expression compared to WT, the mutant phosphorylated IL4R, JAK2, STAT5 and STAT6 independent of IL-4 stimulation (relative densitometric values of 2.4 versus 1.1, 0.7 versus 0.2, 3.4 versus 0.7 and 3.0 versus 0.4, respectively; Figure 3.8). No appreciable changes in phospho-STAT3 or phospho-JAK3 levels were seen between WT and mutant, and no JAK1 protein expression was detected in this cell line. To determine genome-wide gene expression changes in DEV IL4R WT versus DEV IL4R I242N cells, we sequenced the transcriptomes of both cell lines by RNAseq. We found 41 genes (fold change ≥2 compared to DEV WT) to be upregulated and one gene to be downregulated (fold change ≤0.5 compared to DEV WT; Table 3.4) in the mutant cell line. Of these, the B cell activation marker CD23 (FCER2) and the thymus and activation-regulated chemokine (TARC; CCL17) were the most significantly upregulated genes (Figure 3.9a). Experimental validation of predicted CD23 upregulation was done by quantitative RT-PCR (1190% in mutant versus 124% in WT, compared to MT control; Figure 3.9b) and flow cytometry (geometric mean fluorescence intensity of 329 in mutant versus 173 in WT; Figure 3.10a). Elevated TARC predicted by RNAseq was confirmed in mutant cell culture supernatants by

113

ELISA (458% in mutant versus 94% in WT, compared to MT control; Figure 3.10b). Since membrane-bound IL4R can be proteolytically cleaved to form a soluble fraction which has been reported to enhance IL-4 signaling234-235, we assayed secreted IL4R levels in WT and mutant cell culture supernatants by ELISA but found no significant differences (Figure 3.10c).

Mutant IL4R aggregates in a multi-protein complex that activates JAK-STAT signaling. The I242N mutation in exon 8 clusters in the transmembrane domain of the protein and affects an evolutionarily conserved residue. The mutated residue is larger and less hydrophobic than the WT residue (http://www.nbic.nl/support/brs/project-hope). Therefore, we wondered whether the mutant residue interferes with inter-peptide interactions between IL4R and IL2RG or IL13R. Unexpectedly, we observed a marked decrease in IL2RG and IL13R protein expression in the mutant when co-expressed with the partners, an occurrence not seen in WT with co-expression (IL2RG: 0.2 versus 0.8 and IL13R: 0.1 versus 0.5; Figure 3.5b). These findings suggest altered protein-protein interactions of mutant IL4R with IL2RG and IL13R. Therefore, immunoprecipitation analysis of HEK 293 protein lysates co-expressing partner receptors was performed by antibody-mediated precipitation of IL4R and protein blotting of IL2RG and IL13R. Both WT and the mutant were found to interact with their partner cytokine receptors, however these preliminary findings require confirmation by “reverse antibody” co- immunoprecipitation (i.e. immunoprecipitation with anti-IL2RG or anti-IL13R and protein blotting with anti-IL4R; Figure 3.11a). We then sought to determine if intra-peptide interactions within the IL4R mutant could provide a mechanism for mutant-induced JAK- STAT activation. Indeed, protein blotting of DEV-I242N cell lysates performed under non-denaturing (native) experimental conditions revealed the presence of a high molecular weight oligomer, indicating aggregation of a multi-protein complex involving IL4R mutant peptides (Figure 3.11b).

114

Mutation-induced STAT5 activation can be pharmacologically inhibited. To explore the potential therapeutic value of ameliorating constitutive JAK-STAT signaling observed in DEV-I242N, we examined the efficiency of tyrosine kinase inhibitory compounds in inhibiting mutation-induced STAT5 phosphorylation. Cultured cells were treated with the neuroleptic drug pimozide, the pan-JAK inhibitor pyridone 6 or the JAK2-specific inhibitor pacritinib (SB1518) and extracted cell lysates were probed by protein blotting. Cells harboring mutant IL4R were responsive to inhibition by all three agents in a dose-dependent manner, however no differences in half maximal inhibitory concentration (IC50) were observed between WT and mutant cells treated with pimozide or pyridone 6 (Figure 3.12a, b). In contrast, cells transduced with mutant IL4R were more sensitive to pacritinib-mediated JAK2 inhibition than WT cells treated with pacritinib, as demonstrated by lower IC50 values in mutant cells (Figure 3.12c).

3.4. Discussion

We identified somatic coding-sequence mutations in the cytokine receptor IL4Rα, a gene that has not previously been described as being recurrently mutated in lymphoid cancers. By targeted Sanger sequencing, we found that IL4R missense mutations are highly recurrent in PMBCL with a hotspot in exon 8 that affects the transmembrane domain of the IL4R protein. As demonstrated in this study, the hotspot missense single nucleotide variant (I242N) represents a gain-of-function mutation that constitutively activates the oncogenic JAK-STAT pathway and escalates the expression of the downstream target genes CD23 (FCER2) and TARC (CCL17). Although recurrent coding-sequence mutations in IL4R have not been previously described, dispersed mutations in several domains of the IL4R protein have been sporadically identified in solid tumors236-240. However, the I242N mutation has not been reported in these tumors, nor has mutation-associated JAK-STAT deregulation previously described. The frameshift mutation (E684Kfs*2) identified in the PMBCL- derived cell line MedB-1 putatively truncates the IL4R protein upstream of the cytoplasmic immunoreceptor tyrosine-based inhibitory motif (ITIM) domain.

115

Interestingly, this domain plays an important regulatory role in dampening IL4R signaling by acting as docking sites for SH2-containing phosphatases ((PTPN6 (SHP- 1), PTPN11 (SHP-2) and INPP5D (SHIP)) that dephosphorylate activated tyrosine residues241. Therefore, in addition to I242N, E684Kfs*2 may also be involved in JAK- STAT activation and warrants further functional characterization. The effects of in vivo blockade of IL4R signaling have been reported in rhabdomyosarcoma, breast and colon cancer disease models. Mice treated with IL4R neutralizing antibodies showed decreased pulmonary metastases242 and genetic ablation of IL4R resulted in reduced tumorigenicity243-244. Animals xenotransplanted with HL cell lines and treated with a cytotoxin directed against IL4R exhibited tumor regression and decreased incidence of lymph node metastasis245. However, it is unclear if these observed antitumor effects are JAK-STAT mediated. Particularly in PMBCL, it is likely that the mutations in IL4R synergizes with other driver mutations known to be involved in pathogenesis (mutations in PTPN1, SOCS1 and STAT6) that contribute to JAK-STAT signaling152,163,227. A large proportion of PMBCL tumors strongly express CD23 by immunohistochemical analysis246-247 and activating mutations in IL4R provide a genetic basis for CD23 protein overexpression. Moreover, upregulation of the T cell regulatory chemokine TARC may prevent immunesurveillance in the PMBCL tumor microenvironment. TARC binds to the receptor CCR4 which is characteristically expressed on Th2 and Treg cells248 thus, recruiting tumor promoting cells that evade the immune system. Therefore, TARC upregulation provides direct evidence of a somatic mutation in the JAK-STAT pathway that induces an immune privileged reactive milieu. Of note, we did not find the hotspot mutation in 19 laser microdissected samples of HRS cells, despite reported high expression of IL4R245 and elevated TARC249-250. It is likely that TARC increases in this disease entity are attributed to autocrine or paracrine secretory mechanisms. The mutant consistently synthesized less IL4R protein and decreased IL2RG and IL13R protein expression in the presence of partner cytokine receptors. These unanticipated findings may be due to distinct intracellular localizations of multimeric receptors after endocytosis. Hémar et al. followed the intracellular pathway of IL2R after ligand binding by immunofluorescence and confocal microscopy and provided evidence

116 of divergent fates for IL2Rα and IL2Rγ chains: the α chain was transferred to early endosomes and recycled to the cell surface; in contrast, the γ chain was sorted to late endosomes and committed to proteosomal degradation251. We postulate that due to conformational changes in the transmembrane domain, mutant IL4Rα interacts with IL2Rγ with increased affinity and thereby, meeting a similar fate as the latter. Furthermore, the mutant IL4Rα:IL2Rγ heterodimer may accelerate IL2Rγ degradation. Albeit reduced IL4R protein expression, the mutant is able to constitutively activate the JAK-STAT pathway possibly due to aggregation of a multi-protein complex. However, the presence of more abundant mutant IL4R in the protein complex relative to WT is unexpected since both show comparable IL4R protein expression. This may be due to the lack of reducing agents in native gel electrophoresis and inefficient extraction of membranous IL4R. Therefore, we postulate that a fraction of mutant IL4R is predominantly found in the cytoplasm while WT IL4R remains in the membrane. Indeed, homotypic interactions of IL4Rα subunits have been shown to induce IL-4 signaling programs, including JAK-STAT activation252-254. Similar activation mechanisms have also been reported for IL7R121-122. Lastly, we provide preclinical proof-of-concept data to support pharmacological inhibition of mutant-induced JAK-STAT activation by therapeutic agents, including drugs approved for other indications. Existing drugs with established safety in humans that target this pathway can be repurposed to treat lymphoma patients.

In summary, we have discovered recurrent somatic mutations in the cytokine receptor IL4R and demonstrate the hot spot mutation in exon 8 constitutively activates the JAK-STAT signaling pathway. Our data suggest that activating mutations contribute to the pathogenesis of PMBCL, with implications for future treatment strategies.

117

IL4R WT IL4R mutated Variable P-value No. (%) No. (%)

N 47 (100%) 15 (100%) NA Median age - years 37 36 0.760 Male sex 21 (44.6%) 6 (40.0%) 0.750 Ann Arbor Stage† 0.722 I 4 (8.5%) 2 (13.3%) II 30 (63.8%) 7 (46.6%) III 5 (10.6%) 2 (13.3%) IV 6 (12.7%) 3 (20.0%) Presence of constitutional 22 (46.8%) 7 (46.6%) 0.956 symptoms‡ Tumor size  0.314 Median - cm 11 10.5 ≥10 cm 31 (65.9%) 9 (60.0%) 0.748 Primary treatment 0.403 CHOP 29 (61.7%) 10 (66.6%) CHOP-like* 17 (36.1%) 2 (13.3%) Rituximab 25 (53.1%) 6 (40.0%) Extranodal sites § 0.811 0-1 16 (34.0%) 5 (33.3%) >1 19 (40.4%) 5 (33.3%) IPI score@ 0.063 0-1 22 5 2-3 16 2 4-5 5 3

† Missing data for 2 cases (IL4R WT) and 1 case (IL4R mutated) ‡ Missing data for 5 cases (IL4R WT) and 2 cases (IL4R mutated)  Missing data for 2 cases (IL4R WT) and 1 case (IL4R mutated) # Missing data for 1 cases (IL4R WT) and 1 cases (IL4R WT) * CHOP-like: NCVBP (GELA trial), ACVBP (GELA trial), ECVBP (GELA trial), MACOPB, VACOPB, ECV, ACOP § Missing data for 12 cases (IL4R WT) and 5 cases (IL4R mutated) @ Missing data for 4 cases (IL4R WT) and 4 cases (IL4R mutated)

Table 3.1. Demographic and clinical characteristics of PMBCL patients.

118

IL4R exon Sequence Size (bp) Chr. Start Chr. End

TGTAAAACGACGGCCAGT E3 F1 223 27351450 27351469 agatgagctctgctggaagc CAGGAAACAGCTATGAC E3 R1 27351653 27351672 aaggccaggtgacatactgc TGTAAAACGACGGCCAGT E4 F1 428 27353269 27353288 acactagctgggaagctcca CAGGAAACAGCTATGAC E4 R1 27353677 27353696 gcagcaaaccatactcagca TGTAAAACGACGGCCAGT E5 F1 247 27356149 27356168 gcctacaggtgaccagccta CAGGAAACAGCTATGAC E5 R1 27356378 27356397 caggcagctgtgggaaca TGTAAAACGACGGCCAGT E6 F1 353 27357738 27357757 atgggggagtcactgcatag CAGGAAACAGCTATGAC E6 R1 27358071 27358090 caggtgagaagccaggtctc TGTAAAACGACGGCCAGT E7 F1 322 27363750 27363769 caggctggtggctcttaaac CAGGAAACAGCTATGAC E7 R1 27364052 27364071 caggggaagaatggagagtg TGTAAAACGACGGCCAGT E8 F1 250 27367053 27367072 gtctccagtcctgggaagtg CAGGAAACAGCTATGAC E8 R1 27367283 27367302 cagactagaggggcaagtcc TGTAAAACGACGGCCAGT E9 F1 311 27370164 27370183 cgaccacttttatgggagga CAGGAAACAGCTATGAC E9 R1 27370455 27370474 atgacctcattcggcttctg TGTAAAACGACGGCCAGT E10 F1 305 27372002 27372021 cttgtaccccttcctgagca CAGGAAACAGCTATGAC E10 R1 27372287 27372306 cagtccacgtttccagaaca TGTAAAACGACGGCCAGT E11 F1 998 27373490 27373509 gtgatttcaggctgggcttt CAGGAAACAGCTATGAC E11 R1 27374468 27374487 gccttgtaaccagcctctcc TGTAAAACGACGGCCAGT E11 F2 987 27374303 27374322 caacctgagccagaaacctg CAGGAAACAGCTATGAC E11 R2 27375270 27375289 ggccaatcaccttcatacca

Table 3.2. Primers used for IL4R PCR amplification.

119

Sample ID/ Sequencing IL4R Chr. 16 Allele Mutation AA Cell line method exon position change (HG 19) (NP_00 0409)

PM-15 Sanger/Miseq 4 27 353 480 +/- G-->A D37N

PM-67 Sanger/Miseq 4 27 353 556 +/- A-->G Y62C

PM-5 RNAseq+ 5 27 356 212 +/- A-->T N78Y

PM-8 Miseq 5 27 356 318 +/- G-->A G113D

PM-56 Sanger/Miseq 5 27 356 325 +/- C-->G F115L

U2940 RNAseq+ 7 27 363 874 +/- A-->G N176S

PM-24 Sanger/Miseq 7 27 363 945 +/- C-->T R200W

PM-19 Sanger/Miseq 8 27 367 183 +/- T-->A I242N

PM-21 Sanger/Miseq 8 27 367 183 +/- T-->A I242N

PM-16 Sanger/Miseq 8 27 367 183 +/- T-->A I242N

PM-73 Sanger/Miseq 8 27 367 183 +/- T-->A I242N

PM-2 WTSS/Miseq 8 27 367 183 +/- T-->A I242N

PM-22 Sanger/Miseq 8 27 367 183 +/- T-->A I242N

PM-25 Sanger/Miseq 8 27 367 183 +/- T-->A I242N

PM-26 Sanger/Miseq 8 27 367 183 +/- T-->A I242N

Karpas- RNAseq+ 8 27 367 183 +/- T-->A I242N 1106P

U-2940 RNAseq+ 8 27 367 183 +/- T-->A I242N

PM-33 Sanger/Miseq 8 27 367 183 +/- T-->A I242N

PM-22 Sanger/Miseq 8 27 367 211 +/- C-->G C251W

MedB-1 RNAseq+ 11 27 373 597 +/- G-->T K308N

120

Sample ID/ Sequencing IL4R Chr. 16 Allele Mutation AA Cell line method exon position change (HG 19) (NP_00 0409) E684K MedB-1 RNAseq+ 11 27 374 718 +/- Del G fs*2

+ Previously reported227.

Table 3.3. IL4R mutations in PMBCL patient samples and in cell lines.

121

FDR Gene Log2 fold change Fold change P-value adjusted (WT vs. I242N) (WT vs. I242N) (q-value)

UPREGULATED GENES CCL17 (TARC) 3.58 11.98 <1.0E-100 <1.0E-100 CD23 (FCER2) 3.12 8.72 <1.0E-100 <1.0E-100 HOMER2 1.18 2.26 5.97E-63 2.39E-59 IGJ 1.49 2.82 1.13E-56 3.39E-53 AICDA 1.46 2.76 1.30E-50 3.13E-47 EFEMP1 1.50 2.82 1.53E-42 2.29E-39 SOX2 1.33 2.52 2.13E-38 2.56E-35 GJB2 1.13 2.18 5.01E-35 5.01E-32 COL24A1 1.21 2.31 2.33E-28 2.00E-25 DPYSL3 1.12 2.18 4.37E-25 3.28E-22 KCTD1 0.89 1.85 1.95E-24 1.37E-21 CLIC6 0.86 1.82 2.19E-23 1.38E-20 ITGB4 0.81 1.76 1.99E-20 1.14E-17 OR51E1 0.88 1.84 7.19E-16 2.78E-13 TTC40 0.88 1.84 9.05E-16 3.39E-13 C4BPB 0.85 1.80 7.01E-15 2.47E-12 FBP1 0.86 1.81 8.22E-15 2.82E-12 STK32A 0.65 1.57 2.97E-13 9.13E-11 GNG8 0.77 1.71 1.36E-12 3.88E-10 MRC2 0.76 1.70 2.04E-12 5.70E-10 CCR4 0.68 1.60 1.25E-11 3.41E-09 IL2RA 0.71 1.63 9.25E-11 2.24E-08 NR4A3 0.70 1.62 1.68E-10 3.88E-08 SHISA9 0.61 1.52 1.91E-10 4.18E-08 GABRA5 0.60 1.51 4.00E-10 8.42E-08 CHRNB4 0.64 1.56 1.42E-09 2.50E-07 PTGIR 0.64 1.56 2.06E-09 3.58E-07 CCL4 0.55 1.46 7.75E-08 8.69E-06 LHX3 0.57 1.48 2.53E-07 2.42E-05 SERINC2 0.54 1.45 6.03E-07 5.21E-05 CCL3 0.46 1.37 6.85E-07 5.75E-05 CSF1 0.46 1.37 2.53E-06 1.71E-04 KLHL4 0.50 1.42 4.27E-06 2.62E-04 WDR65 0.47 1.39 4.95E-06 2.88E-04 SARDH 0.40 1.32 2.23E-05 1.03E-03 BTLA 0.45 1.37 3.63E-05 1.56E-03 NEK10 0.39 1.31 5.87E-05 2.27E-03

122

FDR Log2 fold change Fold change Gene P-value adjusted (WT vs. I242N) (WT vs. I242N) (q-value) ENPP3 0.37 1.29 9.85E-05 3.45E-03 TREM1 0.40 1.32 1.00E-04 3.49E-03 TMEM154 0.41 1.33 1.22E-04 4.09E-03

DOWNREGULATED GENES CSRNP3 -0.37 0.77 0.000 0.009

Table 3.4. Differentially expressed genes in IL4R I242N by RNAseq.

123

Figure 3.1. Locations of IL4R mutations in PMBCL. Distribution of IL4R mutations in PMBCL cell lines (n = 3), PMBCL clinical samples (n = 62) identified by targeted deep amplicon and Sanger sequencing. Variations in noncoding regions, reported SNPs and silent mutations are not shown (a). Validation of hotspot mutation in exon 8 as somatic (b).

124

IL4R mutated IL4R WT

P = 0.743

IL4R mutated

IL4R WT

P = 0.710

Figure 3.2. IL4R mutations and survival outcomes in PMBCL patients. The Kaplan-Meier graphs show overall survival (a) and progression-free survival (b) in 62 PMBCL patients (47 IL4R WT and 15 mutated).

125

Figure 3.3. IL4R expression and copy number analysis in PMBCL cell lines. Flow cytometric analysis of IL4R (CD 124) cell surface expression (a) and FISH analysis of IL4R allelic imbalances (b) in PMBCL cell lines (T: test; Iso: isotype control; Ctl: no stain control; LCL: lymphoblastoid cell line).

126

127

Figure 3.4. Mutation-induced hyperphosphorylation of STAT6 in HEK293 cells. Quantitative RT-PCR analysis of ectopically expressed IL4R in HEK 293 cells expressing STAT6. Each sample was run in triplicate. Each value represents the average of three independent experiments. Error bars, s. d. (a). Hyperphosphorylation of IL4R and STAT6 proteins determined by protein blotting in HEK 293 expressing STAT6. Relative densitometric values (shown for each band) were normalized against levels of the internal control β-actin (b). (MT, mock expression vector; ###P<0.001 compared to MT and +P<0.05 compared to WT with 20 ng/ml IL-4 stimulation).

128

129

130

Figure 3.5. IL4R, IL2RG and IL13R co-expression in HEK293 cells. Quantitative RT-PCR analysis of ectopically expressed IL4R, IL2RG and IL13R (a), protein analysis of ectopically expressed IL4R, IL2RG and IL13R (b) and activated STAT6-dependent SEAP reporter gene expression (c) in HEK 293 expressing STAT6. Each sample was run in triplicate. Each value represents the average of three independent experiments. Error bars, s. d. Relative densitometric values (shown for each band) were normalized against levels of the internal control β-actin. (MT, mock expression vector; ** P<0.01 and *** P<0.001 compared to WT, and + P<0.05 and ++ P<0.01 compared to WT with 0.05 ng/ml IL-4 stimulation).

131

132

Figure 3.6. Mutagenesis of STAT6 binding sites in IL4R. DNA sequence confirmation of mutated STAT6 binding sites (Y575F, Y603F and Y631F) generated by site-directed mutagenesis (a) and activated STAT6-dependent SEAP reporter gene expression in transfected HEK 293 cells expressing STAT6 (b). Each sample was run in triplicate. Each value represents the average of three independent experiments. Error bars, s. d. (MT, mock expression vector; *** P<0.001 compared to WT and +++ P<0.001 compared to WT with 0.05 ng/ml IL-4 stimulation).

133

Figure 3.7. Retroviral-mediated IL4R expression in DEV cells. DNA sequence (a) and quantitative RT-PCR confirmation (b) of DEV IL4R WT and DEV IL4R I242N generated by retroviral transduction. Each sample was run in triplicate. Each value represents the average of three independent experiments. Error bars, s. d. (MT, mock expression vector; ###P<0.001 compared to MT, ***P<0.001 compared to WT and +++P<0.001 compared to WT with 20 ng/ml IL-4 stimulation).

134

Figure 3.8. Mutant-induced JAK-STAT activation in DEV IL4R I242N cells. Hyperphosphorylation of JAK and STAT proteins determined by protein blotting in the DEV cell line expressing IL4R I242N. Relative densitometric values (shown for each band) were normalized against levels of the internal control β-actin. (MT, mock expression vector).

135

136

Figure 3.9. Top differentially expressed genes in DEV-IL4R-I242N. Differential gene expression of RNA isolated from DEV IL4R WT and IL4R I242N cells by RNAseq. The most significantly changed genes are shown (fold change ≥2 or ≤0.5; q≤0.01; a). Quantitative RT-PCR validation of CD23 upregulation in DEV IL4R I242N using Taqman assay probes (b). Each sample was run in triplicate. Each value represents the average of three independent experiments. Error bars, s. d. (MT, mock expression vector, ***P<0.001 compared to WT).

137

138

Figure 3.10. Validation of CD23 and TARC upregulation. Flow cytometric validation of cell surface CD23 upregulation in DEV IL4R I242N (a). Quantification of secreted TARC (CCL17; b) and soluble IL4R (c) in DEV IL4R I242N by ELISA. Each sample was run in triplicate. Each value represents the average of three independent experiments. Error bars, s. d. (MT, mock expression vector, ***P<0.001 compared to WT).

139

140

Figure 3.11. Protein-protein interactions of mutant IL4R. Immunoprecipitation analysis of HEK 293 protein lysates co-expressing partner receptors by antibody- mediated precipitation of IL4R and protein blotting of IL2RG and IL13R (a; IP, immunoprecipitation; WB, western blotting; INPUT, whole cell lysate; ID, immunodepleted; IgG, isotype control). Protein blotting of DEV IL4R WT and I242N cell lysates performed under non-denaturing (native) experimental conditions and probed with anti-IL4R (b; MT, mock expression vector).

141 a.

142 c.

Figure 3.12. Attenuation of STAT5 phosphorylation by inhibitory compounds. Protein blot analysis of cultured DEV IL4R WT and I242N cells treated with pimozide (a), pyridine 6 (b) and pacritinib (c). Relative densitometric values (shown for each band) were normalized against levels of the internal control β-actin.

143

Chapter 4: Conclusions

144

4.1. Summary

This work demonstrated that genetic alterations in members of the JAK-STAT signaling pathway lead to aberrant pathway activation in primary mediastinal large B cell lymphoma and Hodgkin lymphoma. We identified somatic coding sequence mutations in the non-receptor-type tyrosine phosphatase PTPN1 and in the cytokine receptor IL4Rα, two genes that have not previously been described as being recurrently mutated in lymphoid cancers. We show that PTPN1 mutations lead to reduced phosphatase activity and increased phosphorylation of JAK-STAT pathway members. Silencing of PTPN1 by RNA interference resulted in hyperphosphorylation and overexpression of the oncogenic targets BCL6 and MYC. We further demonstrate that the hotspot mutation in IL4Rα is gain-of-function that constitutively activates the JAK-STAT pathway and upregulates the B cell activation marker CD23 (FCER2) and the T cell regulatory chemokine TARC (CCL17). Our data suggest PTPN1 and IL4R mutations as novel driver alterations in these lymphomas and may provide a rational therapeutic target. The direct oncogenic potential of PTPN1 and IL4R mutations were not explored in this study. Hence, experimental studies to determine the role these mutations play in promoting cell proliferation, survival, transformation and resistance to therapeutic drugs are needed to demonstrate tumorigenecity. Moreover, in vivo xenotransplantation studies would inform on tumor growth kinetics of transplanted lymphoma cell lines harboring these mutations. Lastly, since only a trend toward inferior progression-free survival was observed in PMBCL with PTPN1 mutations and no statistically significant differences in survival outcomes were seen in cases that were screened for IL4R mutations, further validation studies in larger homogenously treated cohorts are needed to evaluate the prognostic value of these mutations in these diseases.

4.2. Results from our study analyzing PTPN1 mutations provide insights into oncogenic signaling in PMBCL and HL (Chapter 2).

All cancers share ten essential hallmarks that are acquired during the progressive transformation from normal cells into their malignant derivatives. Cellular

145 alterations that are favorable to malignant transformation collectively dictate tumorigenesis and abnormal cell growth94. Our findings from the analysis and biological characterization of PTPN1 mutations provide evidence of acquisition of at least four of these fundamental bases of malignancy (Figure 4.1).

Sustaining proliferative signaling We found a significant correlation between PTPN1 mutations and decreased PTP1B protein expression in our PMBCL and HL study cohorts. We studied the functional relevance of loss of PTP1B in a HL cell line in which PTPN1 mRNA expression was silenced and found hyperphosphorylation of members of the JAK-STAT pathway, particularly a pronounced increase in phospho-STAT3 levels. Elevated phospho-STAT3 has been reported in several lymphomas of B cell origin including HL and inhibition of STAT3 has been shown to decrease cell proliferation and induce apoptosis66,68,147,255. STAT3 has also been reported to induce several key genes involved in regulating cell proliferation256-257 and not surprisingly, the proto-oncogene MYC was upregulated in our PTPN1-silenced cell line that potentially contributes to unconstrained cell growth.

Evading anti-growth signals Silencing PTPN1 induced the expression of BCL6 which functions as a master regulator of the germinal centre reaction and plays a transcriptional suppressive role in attenuating DNA damage sensing genes (ATR) and inhibiting cell cycle check point genes (CDKN1Z, TP53, CHEK1) to enable B cell proliferation and survival77. Chromatin immunoprecipitation assays have revealed STAT3 binding sites within the BCL6 gene that showed enhanced recruitment of RNA polymerase II to upregulate BCL6 expression82. Therefore, the resulting unchecked proliferative environment is likely advantageous for the pathogenesis of GC-derived lymphomas, including PMBCL and HL.

Genomic instability and mutations Recent studies have highlighted the importance of negative feedback loops that normally lessen the flux of signals coursing through the intracellular circuitry of

146 pathways258-260. Disruptions in these feedback mechanisms lead to incessant oncogenic signaling. Prominent examples involve the loss-of-function mutations in PTEN which impairs the dampening of the phosphoinositide 3-kinase (PI3-K) pathway261-262 and the Ras oncogene which affects Ras GTPase activity263-264. We identified inactivating mutations in PTPN1, a gene encoding a protein tyrosine phosphatase that acts as a negative regulator of the oncogenic JAK-STAT signaling pathway.

Deregulated cell metabolism Cancer cells can reprogram cellular glucose metabolism in order to fuel cell growth and division94. Upregulation of glucose transporters have been observed in neoplastic cells in a variety of cancer types265-267. Interestingly, PTP1B-deficient neonatal mice exhibit enhanced glucose uptake, upregulation of GLUT2 and translocation of GLUT4 to the cell membrane268-269.

4.3. Results from our study analyzing IL4R mutations provide insights into PMBCL biology (Chapter 3).

Our findings from the analysis and biological characterization of IL4R mutations provide evidence of acquisition of at least three of the hallmarks of cancer (Figure 4.1).

Sustaining proliferative signaling We identified somatic coding-sequence mutations in the cytokine receptor IL4Rα and found a highly recurrent hotspot in our PMBCL study cohort. As demonstrated, the gain- of-function I242N mutation activates phospho-STAT5 and phospho-STAT6 independent of cytokine stimulation. Forced expression of constitutively active STAT5 into human primary B cells has shown evidence of malignant transformation including ligand- independent cell growth73. Moreover, decreases in phospho-STAT6 in HL cell lines were associated with slower rates of cell proliferation47. Hence, activating IL4R mutations likely fosters a hyperproliferative setting mediated by active JAK-STAT signaling.

147

Avoiding immune destruction There is emerging evidence of tumor cells orchestrating surrounding normal cells to escape host immune surveillance45. For instance, malignant HRS cells in cHL, although a minority in the tumor, appear to sculpt their microenvironment by release of a variety of cytokines and chemokines270. In HIV-associated lymphomas, the malignant cells can thrive in an environment in which host immunity is globally impaired271. In recent years, specific phenotypic features have been uncovered detailing how lymphoma cells induce a favorable microenvironment to evade host immune attack. The predominance of immunosuppressive macrophages, T regulatory cells (Treg), Type 2 helper T cells (Th2) and anergic T cells in the direct vicinity of lymphoma cells substantiates this paradigm of immune privilege in certain lymphomas272. Specifically, it has been shown that HRS cells produce chemokines such as CCL5, CCL17 (TARC) and CCL22273-276. We found a significant increase in secreted TARC in our lymphoma cell line expressing the IL4R hotspot mutation. TARC binds to the receptor CCR4 which is characteristically expressed on Th2 and Treg cells248 thus, attracting tumor promoting cells that help evade the host immune system. Therefore, TARC upregulation provides direct evidence of a somatic mutation in the JAK-STAT pathway that induces an immune privileged reactive milieu.

Genomic instability and mutations Sequencing of cancer genomes at base-pair resolution in a variety of malignancies has yielded unprecedented insight into complex genomic aberrations, including single- nucleotide variants in cytokine- or growth factor receptor-driven signaling pathways. Many studies have linked the aberrant activation of tyrosine kinases by somatic mutation or DNA amplification to a wide array of cancers277-278. For instance, gain-of- function mutations in EGFR, BRAF and in the catalytic subunit of PI3-K (PIK3CA) constitutively activate JAK-STAT, mitogen-activated protein kinase (MAPK) and PI3-K pathways279-284. We have discovered recurrent somatic mutations in the cytokine receptor IL4R and demonstrate the hot spot mutation in exon 8 constitutively activates the JAK-STAT signaling pathway.

148

4.4. Therapeutic potential of JAK-STAT inhibition

Since enhanced JAK-STAT signaling is observed in malignant cells of a significant proportion of lymphoma patients, this pathway provides rational targets for therapeutic intervention. Compounds that interfere with members of the JAK-STAT pathway provide novel therapies for patients that do not respond to current treatment regimens or may be used synergistically in combination with standard of care drugs. Moreover, existing drugs with established toxicity profiles in humans that have been shown to target this pathway can be repurposed to treat lymphoma patients. Since repurposing builds on previous research and pharmacological and toxicity data on these drugs are readily available, these candidate therapies can undergo clinical trials sooner. Numerous inhibitory compounds that affect JAK-STAT have been described in vitro and in vivo, including several that are currently being assessed in clinical trials. Agents that have been evaluated in lymphoid malignant cells are described below.

4.4.1. JAK inhibitors

Pacritinib (SB1518; Cell Therapeutics) SB1518 is a pyrimidine-based small-molecule ATP competitive inhibitor with activity against WT and mutant JAK2, and the fms-like kinase, FLT3285. The safety and efficacy of this drug was evaluated in 34 relapsed or refractory lymphoma patients (HL, FL, MCL and DLBCL) who received daily escalating doses for 28 days. Response was evaluated by pre- and post-baseline computed tomography (CT) scans; 55% of patients showed a reduction in tumor size (range 4% to 70%). Furthermore, inhibition of pSTAT3 and pSTAT5 were seen in all samples analyzed irrespective of dosage.

Lestaurtinib (CEP-701) Lestaurtinib is an orally available multi-tyrosine kinase inhibitor first tested in a subset of acute myelogenous leukemia (AML) patients harboring activating mutations in FLT3286. Further studies have shown Lestaurtinib’s kinase activity is not limited to FLT3 but it can

149 also inhibit JAK2 to suppress JAK2/STAT5 signaling and growth inhibition of primary cells isolated from MPD patients287. However, in a phase II study where 39 MPD patients were treated with this drug, only 5 responded clinically with a modest reduction in JAK2 V617F allelic burden288. Recently published pre-clinical studies in HL-derived cell lines and in primary HL cells have shown inhibition of JAK2, STAT3 and STAT5 phosphorylation and dose-dependent growth inhibition in response to this drug109. Clinical evaluation of Lestaurtinib in patients with lymphoma has not been reported and how these patients respond to treatment in the absence of activating JAK2 mutations (but activating STAT mutations) will be important for future therapeutic improvement.

Fedratinib (TG101348: Sanofi Pharmaceuticals) Selective JAK2 inhibitory activity of Fedratinib in pre-clinical studies has been shown in DLBCL, HL, PMBCL and MPD. Treatment of DLBCL cell lines with increasing drug doses showed a concentration-dependent inhibition of pJAK2 and pSTAT350. Treating HL and PMBCL cell lines harboring 9p24 amplicons (JAK2 copy number gains) reduced cell proliferation, JAK2, STAT1, STAT3 and STAT6 phosphorylation, with an inverse correlation between 9p24 copy number and the EC50 of Fedratinib. Murine xenografts of 9p24 amplicon-positive cell lines treated with this drug decreased tumor growth and prolonged survival of these animals135. As of late 2013, Fedratinib’s development is on clinical hold due to adverse neurotoxicity experienced by MPD study participants.

AZD1480 (AstraZenca) AZD1480 is pyrazol pyrimidine ATP-competitive inhibitor that has been shown to effectively inhibit JAK1 and JAK2 at low micromolar concentrations in purified enzymatic assays. At higher concentrations, it also inhibits the activities of JAK3, TYK2 and aurora-A kinase289. In cell lines and in xenograft mouse models of MM, low doses of this drug resulted in concomitant dephosphorylation of JAK2/STAT5 and cyclin D2, which led to reduced cell proliferation and tumor growth, respectively290. HL cell lines also exhibit sensitivity to low doses (0.1-1.0nM) of AZD1480 as seen by significant reductions in pSTAT1, pSTAT3, pSTAT5 and pSTAT6 levels; however cell proliferation was unaffected. Cell cycle arrest and apoptosis was seen at higher doses (>1nM), as a

150 result of inhibition of Aurora-A kinase activity291. Rapid absorption and maximum dephosphorylation of STAT3 were achieved after 2 hours of drug delivery in patients with solid tumors in a phase I trial; however further development was discontinued for unspecified reasons292.

4.4.2. STAT inhibitors

OPB-31121 (Otsuka Pharmaceuticals) OPB-31121 is a small molecule inhibitor that has been shown to inhibit IL6-induced JAK-STAT signaling in gastric cells by STAT3 dephosphorylation293. In BL and DLBCL, this compound showed effective inhibition of STAT3 and STAT5 without upstream kinase inhibition. DLBCL cell lines inoculated into SCID mice and treated with OPB- 31121 showed strong inhibition of STAT3 and tumor growth. Interestingly, treatment does not disrupt the dimerization of STAT3 and the authors postulate this compound likely inhibits the association of STAT with JAK or cytokine receptors294. Phase I/II studies of this drug in patients with NHL are ongoing (NCT1406574). In a clinical setting, STAT inhibitors may be used in conjunction with molecules that target JAK and/or cytokine receptors, depending on the activation status of these upstream effectors. STAT inhibition may also be appropriate for patients who respond poorly to upstream kinase inhibitors.

Leflunomide (A771726; Arava™: Sanofi Pharmaceuticals) Leflunomide is an immunosuppressive drug prescribed for active rheumatoid arthritis. Its mode of action is inhibition of pyrimidine biosynthesis and decrease in uridine, a cellular state that particularly affects lymphocyte cell division without a ‘salvage mechanism’ of uridine production295. Additionally, Leflunomide acts as a tyrosine phosphorylation inhibitor in B cells; IL4-induced JAK3/STAT6 phosphorylation and B cell proliferation were inhibited in a murine model296. Moreover, this drug attenuated STAT3/STAT6 phosphorylation and induced apoptosis in clinically refractory CLL cells297.

151

Pimozide (Orap™: Janssen Pharmaceuticals) Pimozide is a psychotropic drug used to treat schizophrenia, chronic psychosis and Tourette syndrome. Inhibition of STAT5 phosphorylation and induction of apoptosis caused by pimozide have been reported in CML and ALL; the responsiveness of B cell lymphomas to this drug is yet to be determined298-299. 4.4.3. IL4-R and IL13-R modulators

Compounds that target the IL4/IL13 receptor complex have not been evaluated in lymphoid malignancies. However, late stage clinical trials have been conducted in patients suffering from asthma and chronic obstructive pulmonary disease (COPD) using inhibitors that antagonize the binding of IL-4 and IL-13 to the corresponding receptors or blocking agents that act on the receptors themselves300. Of note, these agents target upstream effectors of the JAK-STAT pathway and their blockage will likely have a minimal effect on signaling if downstream kinases (JAK and/or STAT) are activated (i.e. activating mutations).

AIR-645 (ISIS Altair) AIR-645 is a 2’-O-methoxyethyl antisense drug that targets mRNA encoding IL4Rα. Decreased cytokine production and reduced airway inflammation were observed in mouse models of asthma; however phase II clinical trials in asthma patients did not show sufficient therapeutic benefit to warrant continuation of the study (http://sigma.larvol.com).

Aerovant (AER-001; Aerovance) Aerovant (Pitrakinra) is a recombinant human IL-4 variant that is a potent inhibitor of IL4Rα. A phase II trial in a 30-patient study met its primary endpoint of reducing the severity of late asthmatic response by 72% but no further clinical trials have been reported since 2007301.

AMG-317 (Amgen) AMG-317 is an anti-IL4-R monoclonal antibody that blocks IL-4 and IL-13 signaling.

152

Subcutaneous injections of the drug delivered to moderate-to-severe asthma patients in phase II trials demonstrated significant reductions in blood IgE levels but no significant changes in the clinical outcome of the disease302. The trial has been on clinical hold since 2010.

4.5. Current studies

Synergy of somatic mutations in the JAK-STAT pathway

Functional analyses of both PMBCL and cHL show the malignant cells principally rely on JAK-STAT and NF-κB signaling providing these cells with a proliferative advantage21,75,303. Based on the work presented here, JAK-STAT signaling can be viewed as a classical example of a pathway that is targeted by somatic mutations in multiple components along its signaling cascade. In particular, mutations in the JAK- STAT pathway members were frequent in and characteristic of PMBCL. Predominantly these mutations affect IL4R, PTPN1, SOCS1, and STAT6, while JAK2 is the target of gene amplification304. The RNAseq data in 10 PMBCL samples suggest that the combination of mutations along the pathway act synergistically rather than in a mutually exclusive manner (Figure 4.2). Overall, these data suggest that individual mutations or combinations of mutations are indicative of differential biology and raises hope that certain features of PMBCL can be therapeutically exploited. However, the patterns and potential synergies of genetic hits in the pathway have not been explored in a sufficiently sized study cohort. To comprehensively describe the mutational patterns of JAK-STAT pathway mutations in PMBCL larger numbers of samples are currently being studied. The goal of this effort is to correlate mutational findings, including patterns of mutations, with specific histological and clinical phenotypes. Specifically, targeted deep amplicon sequencing is being performed using a customized TruSeq Custom Amplicon gene panel (Illumina) containing the four JAK- STAT pathway genes that are recurrently mutated in PMBCL (IL4R, PTPN1, SOCS1 and STAT6). We have extracted genomic DNA from 150 fresh-frozen-paraffin-

153 embedded (FFPET) and fresh-frozen samples available from the Centre for Lymphoid Cancer (CLC) material bank and the Leukemia and Lymphoma Molecular Profiling Project (LLMPP) consortium for this study. All samples are available on a constructed tissue microarray which will be used to determine JAK2 copy number aberrations by FISH and for immunohostochemical analysis. Following the hypothesis that individual mutations or the combination of mutations lead to constitutive and accentuated JAK- STAT signaling, we will interrogate if mutations are correlated with STAT6 phosphorylation status and expression of downstream activation markers (CD23, TARC) or expression of oncogenes regulated by the JAK-STAT signaling pathway (MYC, BCL6). As all samples are homogeneously treated and clinically well-annotated, we will also correlate our findings to clinical parameters including treatment outcome prediction, prognostic variables, and to survival data with progression-free and overall survival as endpoints. If an association between somatic mutations and protein expression is not observed, we will comprehensively examine possible epigenetic regulatory mechanisms that may affect gene expression (i.e. DNA methylation and histone modicfications). Results from this study will inform on synergy and mutual exclusivity of pathway mutations and if immunohistochemical surrogate markers exist for potential clinical testing and screening in PMBCL patients. The novelty of these mutational findings and the specificity to PMBCL fuel the enthusiasm in the field that these discoveries can be translated into PMBCL-specific treatments as alternative therapeutic options for affected patients in the era of molecular and personalized medicine.

4.6. Future studies

CRISPR-Cas9-mediated genome editing

Recently, a novel genome engineering technology was adopted from a bacterial RNA-mediated adaptive defense system called CRISPR. Here, an RNA guide sequence (sgRNA) is used to direct the Cas9 endonuclease to a specific DNA target sequence to create a double stranded cut which can be repaired in one of two ways: (1) Error-prone

154

DNA repair leading to generation of an indel mutation and creation of a premature stop codon through non-homologous end joining (NHEJ) or, (2) precise DNA repair by a template plasmid through homology-directed repair305. The use of base complementarity via the sgRNA to target Cas9 to a specific genomic location allows virtually any desired site in the genome to be targeted and modified. For example, PTPN1 knock-out cell lines or syngeneic cell line pairs that only differ in the presence or absence of IL4R mutations may be generated by NHEJ and HDR, respectively. This genome editing system will facilitate the precise editing of endogenous genomic PTPN1 and IL4R loci to recapitulate mutations found in HL and PMBCL patients may prove far more effective than shRNA-mediated gene silencing or ectopic expression of a mutant construct.

Synthetic lethal interactions by genome-wide shRNA library screens

Several studies have exploited the concept of synthetic lethality in cancer in order to identify novel drugs to be tested in clinical trials306-307. A similar strategy may be used to discover synthetic lethality partners of mutated genes in the JAK-STAT pathway using an unbiased genome-wide shRNA screen. Genome-edited syngeneic cell line pairs of PTPN1 and IL4R mutants generated by CRISPR/Cas9 cloning may be transduced with a genome-wide shRNA library screen representing approximately 80,000 pooled shRNAs that target 15,802 human genes at 100X coverage (MISSION TRC lentiviral shRNA library, commercially available from Sigma-Aldrich). Unedited parental cell lines serve as reference cultures. The abundance of shRNA integrations in the cultures are quantified from DNA extracted from positively selected transduced cells. Genes with under-represented shRNAs in the modified cell lines compared to their respective controls are candidate synthetic lethal interaction partners to PTPN1 and IL4R mutations. These identified genes, when targeted, results in the specific lethality in mutation carriers without negatively impacting the characteristics of wild type/unmodified cell lines. Hence, this analysis will yield novel drugs for lymphomas harboring JAK-STAT pathway mutations although not directly targetable, might become treatable by a synthetic lethal approach. Gene targets identified in the synthetic lethal

155 screen may be inhibited by pharmacological intervention, if available. The effects of these inhibitors need to be determined in the modified and unmodified parental cell lines prior to selecting compounds for further preclinical development.

156

Figure 4.1. The hallmarks of cancer. [Adapted from Hanahan, D. et al., Cell 2011(ref.94)].

157

IL4/13

IL13R

IL4R

PTPN1

Figure 4.2. Overview of deregulated signaling pathways in PMBCL. Mutated pathway components are highlighted in color [adapted from Steidl, C. et al., Blood 2011 (ref. 31)].

158

Bibliography

1. Swerdlow, S.H., Campo, E., Harris, N.L. et al. . WHO Classification of Tumours of Haematopietic and Lymphoid Tissues. IARC (2008).

2. Howlader N, e.a. SEER Cancer Statistics Review (1975-2010).

3. Scott, D.W. & Gascoyne, R.D. The tumour microenvironment in B cell lymphomas. Nat Rev Cancer 14, 517-34 (2014).

4. Brack, C., Hirama, M., Lenhard-Schuller, R. & Tonegawa, S. A complete immunoglobulin gene is created by somatic recombination. Cell 15, 1-14 (1978).

5. Oettinger, M.A., Schatz, D.G., Gorka, C. & Baltimore, D. RAG-1 and RAG-2, adjacent genes that synergistically activate V(D)J recombination. Science 248, 1517-23 (1990).

6. Batista, F.D. & Harwood, N.E. The who, how and where of antigen presentation to B cells. Nat Rev Immunol 9, 15-27 (2009).

7. Victora, G.D. & Nussenzweig, M.C. Germinal centers. Annu Rev Immunol 30, 429-57 (2012).

8. Muramatsu, M. et al. Class switch recombination and hypermutation require activation-induced cytidine deaminase (AID), a potential RNA editing enzyme. Cell 102, 553-63 (2000).

9. Heesters, B.A., Myers, R.C. & Carroll, M.C. Follicular dendritic cells: dynamic antigen libraries. Nat Rev Immunol 14, 495-504 (2014).

10. Johnson, P.W. & Davies, A.J. Primary mediastinal B-cell lymphoma. Hematology Am Soc Hematol Educ Program, 349-58 (2008).

11. Dunleavy, K. & Wilson, W.H. Primary mediastinal B-cell lymphoma and mediastinal gray zone lymphoma: do they require a unique therapeutic approach? Blood 125, 33-9 (2015).

12. Kuruvilla, J. et al. Salvage chemotherapy and autologous stem cell transplantation are inferior for relapsed or refractory primary mediastinal large B- cell lymphoma compared with diffuse large B-cell lymphoma. Leuk Lymphoma 49, 1329-36 (2008).

13. Rodriguez, J. et al. Primary mediastinal large cell lymphoma (PMBL): frontline treatment with autologous stem cell transplantation (ASCT). The GEL-TAMO experience. Hematol Oncol 26, 171-8 (2008).

14. Soumerai, J.D. et al. Treatment of primary mediastinal B-cell lymphoma with rituximab, cyclophosphamide, doxorubicin, vincristine and prednisone is

159

associated with a high rate of primary refractory disease. Leuk Lymphoma 55, 538-43 (2014).

15. Rieger, M. et al. Primary mediastinal B-cell lymphoma treated with CHOP-like chemotherapy with or without rituximab: results of the Mabthera International Trial Group study. Ann Oncol 22, 664-70 (2011).

16. Vassilakopoulos, T.P. et al. Rituximab, cyclophosphamide, doxorubicin, vincristine, and prednisone with or without radiotherapy in primary mediastinal large B-cell lymphoma: the emerging standard of care. Oncologist 17, 239-49 (2012).

17. Moskowitz, C., Hamlin, P.A., Jr., Maragulia, J., Meikle, J. & Zelenetz, A.D. Sequential Dose-Dense RCHOP Followed by ICE Consolidation (MSKCC protocol 01-142) without Radiotherapy for Patients with Primary Mediastinal Large B Cell Lymphoma. ASH Annual Meeting Abstracts 116, 420- (2010).

18. Savage, K.J. et al. The Outcome of Primary Mediastinal Large B-Cell Lymphoma (PMBCL) in the R-CHOP Treatment Era. ASH Annual Meeting Abstracts 120, 303- (2012).

19. Dunleavy, K. et al. Dose-adjusted EPOCH-rituximab therapy in primary mediastinal B-cell lymphoma. N Engl J Med 368, 1408-16 (2013).

20. Savage, K.J. Primary mediastinal large B-cell lymphoma. Oncologist 11, 488-95 (2006).

21. Kuppers, R. The biology of Hodgkin's lymphoma. Nat Rev Cancer 9, 15-27 (2009).

22. Connors, J.M. Clinical manifestations and natural history of Hodgkin's lymphoma. Cancer J 15, 124-8 (2009).

23. Brune, V. et al. Origin and pathogenesis of nodular lymphocyte-predominant Hodgkin lymphoma as revealed by global gene expression analysis. J Exp Med 205, 2251-68 (2008).

24. Kuruvilla, J. Standard therapy of advanced Hodgkin lymphoma. Hematology Am Soc Hematol Educ Program, 497-506 (2009).

25. Steidl, C. et al. Tumor-associated macrophages and survival in classic Hodgkin's lymphoma. N Engl J Med 362, 875-85 (2010).

26. Ansell, S.M. et al. PD-1 blockade with nivolumab in relapsed or refractory Hodgkin's lymphoma. N Engl J Med 372, 311-9 (2015).

160

27. Banerjee, D. Recent Advances in the Pathobiology of Hodgkin's Lymphoma: Potential Impact on Diagnostic, Predictive, and Therapeutic Strategies. Adv Hematol 2011, 439456 (2011).

28. Savage, K.J. et al. The molecular signature of mediastinal large B-cell lymphoma differs from that of other diffuse large B-cell lymphomas and shares features with classical Hodgkin lymphoma. Blood 102, 3871-9 (2003).

29. Bentz, M. et al. Gain of chromosome arm 9p is characteristic of primary mediastinal B-cell lymphoma (MBL): comprehensive molecular cytogenetic analysis and presentation of a novel MBL cell line. Genes Chromosomes Cancer 30, 393-401 (2001).

30. Joos, S. et al. Genomic imbalances including amplification of the tyrosine kinase gene JAK2 in CD30+ Hodgkin cells. Cancer Res 60, 549-52 (2000).

31. Steidl, C. & Gascoyne, R.D. The molecular pathogenesis of primary mediastinal large B-cell lymphoma. Blood 118, 2659-69 (2011).

32. McCubrey, J.A. et al. Advances in targeting signal transduction pathways. Oncotarget 3, 1505-21 (2012).

33. Isaacs, A. & Lindenmann, J. Virus interference. I. The interferon. Proc R Soc Lond B Biol Sci 147, 258-67 (1957).

34. Stark, G.R. & Darnell, J.E., Jr. The JAK-STAT pathway at twenty. Immunity 36, 503-14 (2012).

35. Aaronson, D.S. & Horvath, C.M. A road map for those who don't know JAK- STAT. Science 296, 1653-5 (2002).

36. Shuai, K. & Liu, B. Regulation of JAK-STAT signalling in the immune system. Nat Rev Immunol 3, 900-11 (2003).

37. Arbouzova, N.I. & Zeidler, M.P. JAK/STAT signalling in Drosophila: insights into conserved regulatory and cellular functions. Development 133, 2605-16 (2006).

38. Igaz, P., Toth, S. & Falus, A. Biological and clinical significance of the JAK-STAT pathway; lessons from knockout mice. Inflamm Res 50, 435-41 (2001).

39. Levy, D.E. & Darnell, J.E., Jr. Stats: transcriptional control and biological impact. Nat Rev Mol Cell Biol 3, 651-62 (2002).

40. Schindler, C., Shuai, K., Prezioso, V.R. & Darnell, J.E., Jr. Interferon-dependent tyrosine phosphorylation of a latent cytoplasmic transcription factor. Science 257, 809-13 (1992).

161

41. Ungureanu, D., Saharinen, P., Junttila, I., Hilton, D.J. & Silvennoinen, O. Regulation of Jak2 through the ubiquitin-proteasome pathway involves phosphorylation of Jak2 on Y1007 and interaction with SOCS-1. Mol Cell Biol 22, 3316-26 (2002).

42. Shaffer, A.L., 3rd, Young, R.M. & Staudt, L.M. Pathogenesis of human B cell lymphomas. Annu Rev Immunol 30, 565-610 (2012).

43. Butterbach, K. et al. Association of JAK-STAT pathway related genes with lymphoma risk: results of a European case-control study (EpiLymph). Br J Haematol 153, 318-33 (2011).

44. Chen, Y. et al. Polymorphisms in JAK/STAT signaling pathway genes and risk of non-Hodgkin lymphoma. Leuk Res 37, 1120-4 (2013).

45. Steidl, C., Connors, J.M. & Gascoyne, R.D. Molecular pathogenesis of Hodgkin's lymphoma: increasing evidence of the importance of the microenvironment. J Clin Oncol 29, 1812-26 (2011).

46. Paul, W.E. Interleukin-4: a prototypic immunoregulatory lymphokine. Blood 77, 1859-70 (1991).

47. Skinnider, B.F. et al. Signal transducer and activator of transcription 6 is frequently activated in Hodgkin and Reed-Sternberg cells of Hodgkin lymphoma. Blood 99, 618-26 (2002).

48. Andersson, E., Schain, F., Sjoberg, J., Bjorkholm, M. & Claesson, H.E. Interleukin-13 stimulation of the mediastinal B-cell lymphoma cell line Karpas- 1106P induces a phenotype resembling the Hodgkin lymphoma cell line L1236. Exp Hematol 38, 116-23 (2010).

49. Charbonneau, B. et al. Pretreatment circulating serum cytokines associated with follicular and diffuse large B-cell lymphoma: a clinic-based case-control study. Cytokine 60, 882-9 (2012).

50. Gupta, M. et al. Elevated serum IL-10 levels in diffuse large B-cell lymphoma: a mechanism of aberrant JAK2 activation. Blood 119, 2844-53 (2012).

51. Clayberger, C. et al. Interleukin 3 is a growth factor for human follicular B cell lymphoma. J Exp Med 175, 371-6 (1992).

52. Foss, H.D. et al. Patterns of cytokine gene expression in peripheral T-cell lymphoma of angioimmunoblastic lymphadenopathy type. Blood 85, 2862-9 (1995).

53. Hsu, S.M., Waldron, J.W., Jr., Hsu, P.L. & Hough, A.J., Jr. Cytokines in malignant lymphomas: review and prospective evaluation. Hum Pathol 24, 1040- 57 (1993).

162

54. Farinha, P. et al. Analysis of multiple biomarkers shows that lymphoma- associated macrophage (LAM) content is an independent predictor of survival in follicular lymphoma (FL). Blood 106, 2169-74 (2005).

55. Steidl, C. et al. Gene expression profiling of microdissected Hodgkin Reed- Sternberg cells correlates with treatment outcome in classical Hodgkin lymphoma. Blood 120, 3530-40 (2012).

56. Shankaran, V. et al. IFNgamma and lymphocytes prevent primary tumour development and shape tumour immunogenicity. Nature 410, 1107-11 (2001).

57. Alvaro, T. et al. The presence of STAT1-positive tumor-associated macrophages and their relation to outcome in patients with follicular lymphoma. Haematologica 91, 1605-12 (2006).

58. Battle, T.E. & Frank, D.A. STAT1 mediates differentiation of chronic lymphocytic leukemia cells in response to Bryostatin 1. Blood 102, 3016-24 (2003).

59. Kovacic, B. et al. STAT1 acts as a tumor promoter for leukemia development. Cancer Cell 10, 77-87 (2006).

60. Camicia, R. et al. BAL1/ARTD9 represses the anti-proliferative and pro-apoptotic IFNgamma-STAT1-IRF1-p53 axis in diffuse large B-cell lymphoma. J Cell Sci 126, 1969-80 (2013).

61. Yu, H., Pardoll, D. & Jove, R. STATs in cancer inflammation and immunity: a leading role for STAT3. Nat Rev Cancer 9, 798-809 (2009).

62. Baran-Marszak, F. et al. Constitutive and B-cell receptor-induced activation of STAT3 are important signaling pathways targeted by bortezomib in leukemic mantle cell lymphoma. Haematologica 95, 1865-72 (2010).

63. Lai, R. et al. Expression of STAT3 and its phosphorylated forms in mantle cell lymphoma cell lines and tumours. J Pathol 199, 84-9 (2003).

64. Martinez, N. et al. The molecular signature of mantle cell lymphoma reveals multiple signals favoring cell survival. Cancer Res 63, 8226-32 (2003).

65. Visser, H.P., Tewis, M., Willemze, R. & Kluin-Nelemans, J.C. Mantle cell lymphoma proliferates upon IL-10 in the CD40 system. Leukemia 14, 1483-9 (2000).

66. Ding, B.B. et al. Constitutively activated STAT3 promotes cell proliferation and survival in the activated B-cell subtype of diffuse large B-cell lymphomas. Blood 111, 1515-23 (2008).

163

67. Huang, X. et al. Activation of the STAT3 signaling pathway is associated with poor survival in diffuse large B-cell lymphoma treated with R-CHOP. J Clin Oncol 31, 4520-8 (2013).

68. Kube, D. et al. STAT3 is constitutively activated in Hodgkin cell lines. Blood 98, 762-70 (2001).

69. Zamo, A. et al. Anaplastic lymphoma kinase (ALK) activates Stat3 and protects hematopoietic cells from cell death. Oncogene 21, 1038-47 (2002).

70. Zhang, Q. et al. Multilevel dysregulation of STAT3 activation in anaplastic lymphoma kinase-positive T/null-cell lymphoma. J Immunol 168, 466-74 (2002).

71. Hinz, M. et al. Nuclear factor kappaB-dependent gene expression profiling of Hodgkin's disease tumor cells, pathogenetic significance, and link to constitutive signal transducer and activator of transcription 5a activity. J Exp Med 196, 605- 17 (2002).

72. Martini, M. et al. Phosphorylated STAT5 represents a new possible prognostic marker in Hodgkin lymphoma. Am J Clin Pathol 129, 472-7 (2008).

73. Scheeren, F.A. et al. IL-21 is expressed in Hodgkin lymphoma and activates STAT5: evidence that activated STAT5 is required for Hodgkin lymphomagenesis. Blood 111, 4706-15 (2008).

74. Qin, J.Z. et al. Constitutive and interleukin-7- and interleukin-15-stimulated DNA binding of STAT and novel factors in cutaneous T cell lymphoma cells. J Invest Dermatol 117, 583-9 (2001).

75. Guiter, C. et al. Constitutive STAT6 activation in primary mediastinal large B-cell lymphoma. Blood 104, 543-9 (2004).

76. Bowman, T., Garcia, R., Turkson, J. & Jove, R. STATs in oncogenesis. Oncogene 19, 2474-88 (2000).

77. Hatzi, K. & Melnick, A. Breaking bad in the germinal center: how deregulation of BCL6 contributes to lymphomagenesis. Trends Mol Med 20, 343-52 (2014).

78. Basso, K. & Dalla-Favera, R. BCL6: master regulator of the germinal center reaction and key oncogene in B cell lymphomagenesis. Adv Immunol 105, 193- 210 (2010).

79. Basso, K. et al. BCL6 positively regulates AID and germinal center gene expression via repression of miR-155. J Exp Med 209, 2455-65 (2012).

80. Seegmiller, A.C. et al. Simple karyotype and bcl-6 expression predict a diagnosis of Burkitt lymphoma and better survival in IG-MYC rearranged high-grade B-cell lymphomas. Mod Pathol 23, 909-20 (2010).

164

81. Velichutina, I. et al. EZH2-mediated epigenetic silencing in germinal center B cells contributes to proliferation and lymphomagenesis. Blood 116, 5247-55 (2010).

82. Walker, S.R. et al. STAT5 outcompetes STAT3 to regulate the expression of the oncogenic transcriptional modulator BCL6. Mol Cell Biol 33, 2879-90 (2013).

83. Scheeren, F.A. et al. STAT5 regulates the self-renewal capacity and differentiation of human memory B cells and controls Bcl-6 expression. Nat Immunol 6, 303-13 (2005).

84. Walker, S.R., Nelson, E.A. & Frank, D.A. STAT5 represses BCL6 expression by binding to a regulatory region frequently mutated in lymphomas. Oncogene 26, 224-33 (2007).

85. Ritz, O. et al. STAT6-mediated BCL6 repression in primary mediastinal B-cell lymphoma (PMBL). Oncotarget 4, 1093-102 (2013).

86. Haberle, M.T., Kelsch, E., Dorsch, K., Moller, P. & Ritz, O. Knock-down of BCL6 / STAT6 sensitizes primary B cell lymphoma cells for treatment with current therapeutic agents. Oncoscience 1, 283-6 (2014).

87. Tunyaplin, C. et al. Direct repression of prdm1 by Bcl-6 inhibits plasmacytic differentiation. J Immunol 173, 1158-65 (2004).

88. Boi, M., Zucca, E., Inghirami, G. & Bertoni, F. PRDM1/BLIMP1: a tumor suppressor gene in B and T cell lymphomas. Leuk Lymphoma, 1-6 (2014).

89. Karube, K. et al. Identification of FOXO3 and PRDM1 as tumor-suppressor gene candidates in NK-cell neoplasms by genomic and functional analyses. Blood 118, 3195-204 (2011).

90. Mandelbaum, J. et al. BLIMP1 is a tumor suppressor gene frequently disrupted in activated B cell-like diffuse large B cell lymphoma. Cancer Cell 18, 568-79 (2010).

91. Diehl, S.A. et al. STAT3-mediated up-regulation of BLIMP1 Is coordinated with BCL6 down-regulation to control human plasma cell differentiation. J Immunol 180, 4805-15 (2008).

92. Fornek, J.L. et al. Critical role for Stat3 in T-dependent terminal differentiation of IgG B cells. Blood 107, 1085-91 (2006).

93. Reljic, R., Wagner, S.D., Peakman, L.J. & Fearon, D.T. Suppression of signal transducer and activator of transcription 3-dependent B lymphocyte terminal differentiation by BCL-6. J Exp Med 192, 1841-8 (2000).

165

94. Hanahan, D. & Weinberg, R.A. Hallmarks of cancer: the next generation. Cell 144, 646-74 (2011).

95. Bartek, J., Bartkova, J. & Lukas, J. The retinoblastoma protein pathway and the restriction point. Curr Opin Cell Biol 8, 805-14 (1996).

96. Gladkikh, A., Potashnikova, D., Korneva, E., Khudoleeva, O. & Vorobjev, I. Cyclin D1 expression in B-cell lymphomas. Exp Hematol 38, 1047-57 (2010).

97. Bai, M. et al. Proliferation profile of classical Hodgkin's lymphomas. Increased expression of the protein cyclin D2 in Hodgkin's and Reed-Sternberg cells. Mod Pathol 17, 1338-45 (2004).

98. Igawa, T. et al. Cyclin D2 is overexpressed in proliferation centers of chronic lymphocytic leukemia/small lymphocytic lymphoma. Cancer Science 102, 2103- 2107 (2011).

99. Martino, A., Holmes, J.H.t., Lord, J.D., Moon, J.J. & Nelson, B.H. Stat5 and Sp1 regulate transcription of the cyclin D2 gene in response to IL-2. J Immunol 166, 1723-9 (2001).

100. Dalton, R.R. et al. Differential expression of cyclin D3 in ALK+ and ALK- anaplastic large cell lymphoma. Hum Pathol 36, 806-11 (2005).

101. Kanwar, J.R., Kamalapuram, S.K. & Kanwar, R.K. Survivin signaling in clinical oncology: a multifaceted dragon. Med Res Rev 33, 765-89 (2013).

102. Zhang, C. et al. Avicin D selectively induces apoptosis and downregulates p- STAT-3, bcl-2, and survivin in cutaneous T-cell lymphoma cells. J Invest Dermatol 128, 2728-35 (2008).

103. van der Fits, L., Out-Luiting, J.J., Tensen, C.P., Zoutman, W.H. & Vermeer, M.H. Exploring the IL-21-STAT3 axis as therapeutic target for Sezary syndrome. J Invest Dermatol 134, 2639-47 (2014).

104. Ashizawa, T. et al. Antitumor activity of a novel small molecule STAT3 inhibitor against a human lymphoma cell line with high STAT3 activation. Int J Oncol 38, 1245-52 (2011).

105. Reed, J.C. Bcl-2-family proteins and hematologic malignancies: history and future prospects. Blood 111, 3322-30 (2008).

106. Qin, J.Z. et al. Interleukin-7 and interleukin-15 regulate the expression of the bcl- 2 and c-myb genes in cutaneous T-cell lymphoma cells. Blood 98, 2778-83 (2001).

166

107. Alas, S. & Bonavida, B. Inhibition of constitutive STAT3 activity sensitizes resistant non-Hodgkin's lymphoma and multiple myeloma to chemotherapeutic drug-mediated apoptosis. Clin Cancer Res 9, 316-26 (2003).

108. Cochet, O., Frelin, C., Peyron, J.F. & Imbert, V. Constitutive activation of STAT proteins in the HDLM-2 and L540 Hodgkin lymphoma-derived cell lines supports cell survival. Cell Signal 18, 449-55 (2006).

109. Diaz, T. et al. Lestaurtinib inhibition of the Jak/STAT signaling pathway in hodgkin lymphoma inhibits proliferation and induces apoptosis. PLoS One 6, e18856 (2011).

110. Epling-Burnette, P.K. et al. Inhibition of STAT3 signaling leads to apoptosis of leukemic large granular lymphocytes and decreased Mcl-1 expression. J Clin Invest 107, 351-62 (2001).

111. Ferrara, N., Gerber, H.P. & LeCouter, J. The biology of VEGF and its receptors. Nat Med 9, 669-76 (2003).

112. Ruan, J. & Leonard, J.P. Targeting angiogenesis: a novel, rational therapeutic approach for non-Hodgkin lymphoma. Leuk Lymphoma 50, 679-81 (2009).

113. Niu, G. et al. Constitutive Stat3 activity up-regulates VEGF expression and tumor angiogenesis. Oncogene 21, 2000-8 (2002).

114. Roorda, B.D. et al. VEGF-A promotes lymphoma tumour growth by activation of STAT proteins and inhibition of p27(KIP1) via paracrine mechanisms. Eur J Cancer 46, 974-82 (2010).

115. Yu, H. & Jove, R. The STATs of cancer--new molecular targets come of age. Nat Rev Cancer 4, 97-105 (2004).

116. Qiao, Q., Nozaki, Y., Sakoe, K., Komatsu, N. & Kirito, K. NF-kappaB mediates aberrant activation of HIF-1 in malignant lymphoma. Exp Hematol 38, 1199-208 (2010).

117. Jiang, Q. et al. Cell biology of IL-7, a key lymphotrophin. Cytokine Growth Factor Rev 16, 513-33 (2005).

118. Gonzalez-Garcia, S. et al. CSL-MAML-dependent Notch1 signaling controls T lineage-specific IL-7R{alpha} gene expression in early human thymopoiesis and leukemia. J Exp Med 206, 779-91 (2009).

119. van der Plas, D.C. et al. Interleukin-7 signaling in human B cell precursor acute lymphoblastic leukemia cells and murine BAF3 cells involves activation of STAT1 and STAT5 mediated via the interleukin-7 receptor alpha chain. Leukemia 10, 1317-25 (1996).

167

120. Zhang, J. et al. The genetic basis of early T-cell precursor acute lymphoblastic leukaemia. Nature 481, 157-63 (2012).

121. Shochat, C. et al. Gain-of-function mutations in interleukin-7 receptor-alpha (IL7R) in childhood acute lymphoblastic leukemias. J Exp Med 208, 901-8 (2011).

122. Zenatti, P.P. et al. Oncogenic IL7R gain-of-function mutations in childhood T-cell acute lymphoblastic leukemia. Nat Genet 43, 932-9 (2011).

123. Mullighan, C.G. et al. JAK mutations in high-risk childhood acute lymphoblastic leukemia. Proc Natl Acad Sci U S A 106, 9414-8 (2009).

124. Flex, E. et al. Somatically acquired JAK1 mutations in adult acute lymphoblastic leukemia. J Exp Med 205, 751-8 (2008).

125. Bellanger, D. et al. Recurrent JAK1 and JAK3 somatic mutations in T-cell prolymphocytic leukemia. Leukemia 28, 417-9 (2014).

126. Kralovics, R. et al. A gain-of-function mutation of JAK2 in myeloproliferative disorders. N Engl J Med 352, 1779-90 (2005).

127. Levine, R.L. et al. The JAK2V617F activating mutation occurs in chronic myelomonocytic leukemia and acute myeloid leukemia, but not in acute lymphoblastic leukemia or chronic lymphocytic leukemia. Blood 106, 3377-9 (2005).

128. Peeters, P. et al. Fusion of TEL, the ETS-variant gene 6 (ETV6), to the receptor- associated kinase JAK2 as a result of t(9;12) in a lymphoid and t(9;15;12) in a myeloid leukemia. Blood 90, 2535-40 (1997).

129. Nebral, K. et al. Incidence and diversity of PAX5 fusion genes in childhood acute lymphoblastic leukemia. Leukemia 23, 134-43 (2009).

130. Poitras, J.L., Dal Cin, P., Aster, J.C., Deangelo, D.J. & Morton, C.C. Novel SSBP2-JAK2 fusion gene resulting from a t(5;9)(q14.1;p24.1) in pre-B acute lymphocytic leukemia. Genes Chromosomes Cancer 47, 884-9 (2008).

131. Reiter, A. et al. The t(8;9)(p22;p24) is a recurrent abnormality in chronic and acute leukemia that fuses PCM1 to JAK2. Cancer Res 65, 2662-7 (2005).

132. Lacronique, V. et al. A TEL-JAK2 fusion protein with constitutive kinase activity in human leukemia. Science 278, 1309-12 (1997).

133. Van Roosbroeck, K. et al. JAK2 rearrangements, including the novel SEC31A- JAK2 fusion, are recurrent in classical Hodgkin lymphoma. Blood 117, 4056-64 (2011).

168

134. Green, M.R. et al. Integrative analysis reveals selective 9p24.1 amplification, increased PD-1 ligand expression, and further induction via JAK2 in nodular sclerosing Hodgkin lymphoma and primary mediastinal large B-cell lymphoma. Blood 116, 3268-77 (2010).

135. Hao, Y. et al. Selective JAK2 inhibition specifically decreases Hodgkin lymphoma and mediastinal large B-cell lymphoma growth in vitro and in vivo. Clin Cancer Res 20, 2674-83 (2014).

136. Koo, G.C. et al. Janus kinase 3-activating mutations identified in natural killer/T- cell lymphoma. Cancer Discov 2, 591-7 (2012).

137. Cornejo, M.G. et al. Constitutive JAK3 activation induces lymphoproliferative syndromes in murine bone marrow transplantation models. Blood 113, 2746-54 (2009).

138. Bouchekioua, A. et al. JAK3 deregulation by activating mutations confers invasive growth advantage in extranodal nasal-type natural killer cell lymphoma. Leukemia 28, 338-48 (2014).

139. Elliott, N.E. et al. FERM domain mutations induce gain of function in JAK3 in adult T-cell leukemia/lymphoma. Blood 118, 3911-21 (2011).

140. Sanda, T. et al. TYK2-STAT1-BCL2 pathway dependence in T-cell acute lymphoblastic leukemia. Cancer Discov 3, 564-77 (2013).

141. Lohr, J.G. et al. Discovery and prioritization of somatic mutations in diffuse large B-cell lymphoma (DLBCL) by whole-exome sequencing. Proc Natl Acad Sci U S A 109, 3879-84 (2012).

142. Ohgami, R.S., Ma, L., Monabati, A., Zehnder, J.L. & Arber, D.A. STAT3 mutations are present in aggressive B-cell lymphomas including a subset of diffuse large B-cell lymphomas with CD30 expression. Haematologica 99, e105-7 (2014).

143. Odejide, O. et al. A targeted mutational landscape of angioimmunoblastic T-cell lymphoma. Blood 123, 1293-6 (2014).

144. Ohgami, R.S. et al. STAT3 mutations are frequent in CD30+ T-cell lymphomas and T-cell large granular lymphocytic leukemia. Leukemia 27, 2244-7 (2013).

145. Koskela, H.L. et al. Somatic STAT3 mutations in large granular lymphocytic leukemia. N Engl J Med 366, 1905-13 (2012).

146. Jerez, A. et al. STAT3 mutations unify the pathogenesis of chronic lymphoproliferative disorders of NK cells and T-cell large granular lymphocyte leukemia. Blood 120, 3048-57 (2012).

169

147. Couronne, L. et al. STAT3 mutations identified in human hematologic neoplasms induce myeloid malignancies in a mouse bone marrow transplantation model. Haematologica 98, 1748-52 (2013).

148. Rajala, H.L. et al. Discovery of somatic STAT5b mutations in large granular lymphocytic leukemia. Blood 121, 4541-50 (2013).

149. Nosaka, T. et al. STAT5 as a molecular regulator of proliferation, differentiation and apoptosis in hematopoietic cells. EMBO J 18, 4754-65 (1999).

150. Onishi, M. et al. Identification and characterization of a constitutively active STAT5 mutant that promotes cell proliferation. Mol Cell Biol 18, 3871-9 (1998).

151. Joliot, V., Cormier, F., Medyouf, H., Alcalde, H. & Ghysdael, J. Constitutive STAT5 activation specifically cooperates with the loss of p53 function in B-cell lymphomagenesis. Oncogene 25, 4573-84 (2006).

152. Ritz, O. et al. Recurrent mutations of the STAT6 DNA binding domain in primary mediastinal B-cell lymphoma. Blood 114, 1236-42 (2009).

153. Pasqualucci, L. et al. Genetics of Follicular Lymphoma Transformation. Cell Reports 6, 130-140.

154. Ritz, O. et al. STAT6 activity is regulated by SOCS-1 and modulates BCL-XL expression in primary mediastinal B-cell lymphoma. Leukemia 22, 2106-10 (2008).

155. Kleppe, M. et al. Deletion of the protein tyrosine phosphatase gene PTPN2 in T- cell acute lymphoblastic leukemia. Nat Genet 42, 530-5 (2010).

156. Kleppe, M. et al. PTPN2 negatively regulates oncogenic JAK1 in T-cell acute lymphoblastic leukemia. Blood 117, 7090-8 (2011).

157. Kleppe, M. et al. Mutation analysis of the tyrosine phosphatase PTPN2 in Hodgkin's lymphoma and T-cell non-Hodgkin's lymphoma. Haematologica 96, 1723-7 (2011).

158. Lu, X. et al. T-cell protein tyrosine phosphatase, distinctively expressed in activated-B-cell-like diffuse large B-cell lymphomas, is the nuclear phosphatase of STAT6. Mol Cell Biol 27, 2166-79 (2007).

159. Saunders, A.E. & Johnson, P. Modulation of immune cell signalling by the leukocyte common tyrosine phosphatase, CD45. Cell Signal 22, 339-48 (2010).

160. Porcu, M. et al. Mutation of the receptor tyrosine phosphatase PTPRC (CD45) in T-cell acute lymphoblastic leukemia. Blood 119, 4476-9 (2012).

170

161. Yasukawa, H. et al. The JAK-binding protein JAB inhibits Janus tyrosine kinase activity through binding in the activation loop. EMBO J 18, 1309-20 (1999).

162. Zhang, J.G. et al. The conserved SOCS box motif in suppressors of cytokine signaling binds to elongins B and C and may couple bound proteins to proteasomal degradation. Proc Natl Acad Sci U S A 96, 2071-6 (1999).

163. Melzner, I. et al. Biallelic mutation of SOCS-1 impairs JAK2 degradation and sustains phospho-JAK2 action in the MedB-1 mediastinal lymphoma line. Blood 105, 2535-42 (2005).

164. Melzner, I. et al. Biallelic deletion within 16p13.13 including SOCS-1 in Karpas1106P mediastinal B-cell lymphoma line is associated with delayed degradation of JAK2 protein. Int J Cancer 118, 1941-4 (2006).

165. Weniger, M.A. et al. Mutations of the tumor suppressor gene SOCS-1 in classical Hodgkin lymphoma are frequent and associated with nuclear phospho-STAT5 accumulation. Oncogene 25, 2679-84 (2006).

166. Mottok, A., Renne, C., Willenbrock, K., Hansmann, M.L. & Brauninger, A. Somatic hypermutation of SOCS1 in lymphocyte-predominant Hodgkin lymphoma is accompanied by high JAK2 expression and activation of STAT6. Blood 110, 3387-90 (2007).

167. Mottok, A. et al. Inactivating SOCS1 mutations are caused by aberrant somatic hypermutation and restricted to a subset of B-cell lymphoma entities. Blood 114, 4503-6 (2009).

168. Alexander, W.S. et al. SOCS1 is a critical inhibitor of interferon gamma signaling and prevents the potentially fatal neonatal actions of this cytokine. Cell 98, 597- 608 (1999).

169. Tonks, N.K. Protein tyrosine phosphatases: from genes, to function, to disease. Nat Rev Mol Cell Biol 7, 833-46 (2006).

170. Rosenwald, A. et al. Molecular diagnosis of primary mediastinal B cell lymphoma identifies a clinically favorable subgroup of diffuse large B cell lymphoma related to Hodgkin lymphoma. J Exp Med 198, 851-62 (2003).

171. Bains, T. et al. Newly described activating JAK3 mutations in T-cell acute lymphoblastic leukemia. Leukemia 26, 2144-6 (2012).

172. Capello, D. et al. Alterations of negative regulators of cytokine signalling in immunodeficiency-related non-Hodgkin lymphoma. Hematol Oncol 31, 22-8 (2013).

171

173. Han, Y. et al. Loss of SHP1 enhances JAK3/STAT3 signaling and decreases proteosome degradation of JAK3 and NPM-ALK in ALK+ anaplastic large-cell lymphoma. Blood 108, 2796-803 (2006).

174. Chim, C.S., Wong, K.Y., Loong, F. & Srivastava, G. SOCS1 and SHP1 hypermethylation in mantle cell lymphoma and follicular lymphoma: implications for epigenetic activation of the Jak/STAT pathway. Leukemia 18, 356-8 (2004).

175. Khoury, J.D., Rassidakis, G.Z., Medeiros, L.J., Amin, H.M. & Lai, R. Methylation of SHP1 gene and loss of SHP1 protein expression are frequent in systemic anaplastic large cell lymphoma. Blood 104, 1580-1 (2004).

176. Oka, T. et al. Gene silencing of the tyrosine phosphatase SHP1 gene by aberrant methylation in leukemias/lymphomas. Cancer Res 62, 6390-4 (2002).

177. Ferrara, G. et al. A specific DNA methylation profile correlates with a high risk of disease progression in stage I classical (Alibert-Bazin type) mycosis fungoides. Br J Dermatol 170, 1266-75 (2014).

178. Galm, O., Yoshikawa, H., Esteller, M., Osieka, R. & Herman, J.G. SOCS-1, a negative regulator of cytokine signaling, is frequently silenced by methylation in multiple myeloma. Blood 101, 2784-8 (2003).

179. Baus, D. et al. STAT6 and STAT1 are essential antagonistic regulators of cell survival in classical Hodgkin lymphoma cell line. Leukemia 23, 1885-93 (2009).

180. Du, Z. et al. Interferon-resistant Daudi cell line with a Stat2 defect is resistant to apoptosis induced by chemotherapeutic agents. J Biol Chem 284, 27808-15 (2009).

181. Kopp, K.L. et al. STAT5-mediated expression of oncogenic miR-155 in cutaneous T-cell lymphoma. Cell Cycle 12, 1939-47 (2013).

182. Hu, G., Witzig, T.E. & Gupta, M. A novel missense (M206K) STAT3 mutation in diffuse large B cell lymphoma deregulates STAT3 signaling. PLoS One 8, e67851 (2013).

183. Oschlies, I. et al. Clinical, pathological and genetic features of primary mediastinal large B-cell lymphomas and mediastinal gray zone lymphomas in children. Haematologica 96, 262-8 (2011).

184. Kimm, L.R. et al. Frequent occurrence of deletions in primary mediastinal B-cell lymphoma. Genes Chromosomes Cancer 46, 1090-7 (2007).

185. Ding, L. et al. Clonal evolution in relapsed acute myeloid leukaemia revealed by whole-genome sequencing. Nature 481, 506-10 (2012).

172

186. Morin, R.D. et al. Frequent mutation of histone-modifying genes in non-Hodgkin lymphoma. Nature 476, 298-303 (2011).

187. Shah, S.P. et al. The clonal and mutational evolution spectrum of primary triple- negative breast cancers. Nature 486, 395-9 (2012).

188. Steidl, C. et al. MHC class II transactivator CIITA is a recurrent gene fusion partner in lymphoid cancers. Nature 471, 377-81 (2011).

189. Shah, S.P. et al. Mutation of FOXL2 in granulosa-cell tumors of the ovary. N Engl J Med 360, 2719-29 (2009).

190. Tiacci, E. et al. BRAF mutations in hairy-cell leukemia. N Engl J Med 364, 2305- 15 (2011).

191. Dube, N., Cheng, A. & Tremblay, M.L. The role of protein tyrosine phosphatase 1B in Ras signaling. Proc Natl Acad Sci U S A 101, 1834-9 (2004).

192. Yip, S.C., Saha, S. & Chernoff, J. PTP1B: a double agent in metabolism and oncogenesis. Trends Biochem Sci 35, 442-9 (2010).

193. Moller, P. et al. MedB-1, a human tumor cell line derived from a primary mediastinal large B-cell lymphoma. Int J Cancer 92, 348-53 (2001).

194. Li, H. & Durbin, R. Fast and accurate long-read alignment with Burrows-Wheeler transform. Bioinformatics 26, 589-95 (2010).

195. Roth, A. et al. JointSNVMix: a probabilistic model for accurate detection of somatic mutations in normal/tumour paired next-generation sequencing data. Bioinformatics 28, 907-13 (2012).

196. Ding, J. et al. Feature-based classifiers for somatic mutation detection in tumour- normal paired sequencing data. Bioinformatics 28, 167-75 (2012).

197. Cingolani, P. et al. A program for annotating and predicting the effects of single nucleotide polymorphisms, SnpEff: SNPs in the genome of Drosophila melanogaster strain w1118; iso-2; iso-3. Fly (Austin) 6, 80-92 (2012).

198. McPherson, A. et al. nFuse: discovery of complex genomic rearrangements in cancer using high-throughput sequencing. Genome Res 22, 2250-61 (2012).

199. Ha, G. et al. Integrative analysis of genome-wide loss of heterozygosity and monoallelic expression at nucleotide resolution reveals disrupted pathways in triple-negative breast cancer. Genome Res 22, 1995-2007 (2012).

200. Langmead, B., Schatz, M.C., Lin, J., Pop, M. & Salzberg, S.L. Searching for SNPs with cloud computing. Genome Biol 10, R134 (2009).

173

201. Wu, T.D. & Nacu, S. Fast and SNP-tolerant detection of complex variants and splicing in short reads. Bioinformatics 26, 873-81 (2010).

202. Goya, R. et al. SNVMix: predicting single nucleotide variants from next- generation sequencing of tumors. Bioinformatics 26, 730-6 (2010).

203. van Beers, E.H. et al. A multiplex PCR predictor for aCGH success of FFPE samples. Br J Cancer 94, 333-7 (2006).

204. van Dongen, J.J. et al. Design and standardization of PCR primers and protocols for detection of clonal immunoglobulin and T-cell receptor gene recombinations in suspect lymphoproliferations: report of the BIOMED-2 Concerted Action BMH4-CT98-3936. Leukemia 17, 2257-317 (2003).

205. Pasqualucci, L. et al. BCL-6 mutations in normal germinal center B cells: evidence of somatic hypermutation acting outside Ig loci. Proc Natl Acad Sci U S A 95, 11816-21 (1998).

206. Popov, S.W. et al. Target sequence accessibility limits activation-induced cytidine deaminase activity in primary mediastinal B-cell lymphoma. Cancer Res 67, 6555-64 (2007).

207. Lu, X. et al. PTP1B is a negative regulator of interleukin 4-induced STAT6 signaling. Blood 112, 4098-108 (2008).

208. Dauer, D.J. et al. Stat3 regulates genes common to both wound healing and cancer. Oncogene 24, 3397-408 (2005).

209. Gottesman, M.M., Fojo, T. & Bates, S.E. Multidrug resistance in cancer: role of ATP-dependent transporters. Nat Rev Cancer 2, 48-58 (2002).

210. Koyama, N. et al. Inhibition of phosphotyrosine phosphatase 1B causes resistance in BCR-ABL-positive leukemia cells to the ABL kinase inhibitor STI571. Clin Cancer Res 12, 2025-31 (2006).

211. Nanney, L.B., Davidson, M.K., Gates, R.E., Kano, M. & King, L.E., Jr. Altered distribution and expression of protein tyrosine phosphatases in normal human skin as compared to squamous cell carcinomas. J Cutan Pathol 24, 521-32 (1997).

212. Warabi, M., Nemoto, T., Ohashi, K., Kitagawa, M. & Hirokawa, K. Expression of protein tyrosine phosphatases and its significance in esophageal cancer. Exp Mol Pathol 68, 187-95 (2000).

213. Wiener, J.R. et al. Overexpression of the tyrosine phosphatase PTP1B is associated with human ovarian carcinomas. Am J Obstet Gynecol 170, 1177-83 (1994).

174

214. Wiener, J.R. et al. Overexpression of the protein tyrosine phosphatase PTP1B in human breast cancer: association with p185c-erbB-2 protein expression. J Natl Cancer Inst 86, 372-8 (1994).

215. Zhai, Y.F. et al. Increased expression of specific protein tyrosine phosphatases in human breast epithelial cells neoplastically transformed by the neu oncogene. Cancer Res 53, 2272-8 (1993).

216. Blume-Jensen, P. & Hunter, T. Oncogenic kinase signalling. Nature 411, 355-65 (2001).

217. Lim, W.A. & Pawson, T. Phosphotyrosine signaling: evolving a new cellular communication system. Cell 142, 661-7 (2010).

218. Sun, T. et al. Activation of multiple proto-oncogenic tyrosine kinases in breast cancer via loss of the PTPN12 phosphatase. Cell 144, 703-18 (2011).

219. Tartaglia, M. et al. Somatic mutations in PTPN11 in juvenile myelomonocytic leukemia, myelodysplastic syndromes and acute myeloid leukemia. Nat Genet 34, 148-50 (2003).

220. Dube, N. & Tremblay, M.L. Beyond the metabolic function of PTP1B. Cell Cycle 3, 550-3 (2004).

221. Ostman, A., Hellberg, C. & Bohmer, F.D. Protein-tyrosine phosphatases and cancer. Nat Rev Cancer 6, 307-20 (2006).

222. Dube, N. et al. Genetic ablation of protein tyrosine phosphatase 1B accelerates lymphomagenesis of p53-null mice through the regulation of B-cell development. Cancer Res 65, 10088-95 (2005).

223. Mullighan, C.G. Genome sequencing of lymphoid malignancies. Blood 122, 3899-907 (2013).

224. Kridel, R. et al. Whole transcriptome sequencing reveals recurrent NOTCH1 mutations in mantle cell lymphoma. Blood (2012).

225. Kimura, Y., Morita, S.Y., Matsuo, M. & Ueda, K. Mechanism of multidrug recognition by MDR1/ABCB1. Cancer Sci 98, 1303-10 (2007).

226. Zhang, H. et al. Downregulation of gene MDR1 by shRNA to reverse multidrug- resistance of ovarian cancer A2780 cells. J Cancer Res Ther 8, 226-31 (2012).

227. Gunawardana, J. et al. Recurrent somatic mutations of PTPN1 in primary mediastinal B cell lymphoma and Hodgkin lymphoma. Nat Genet 46, 329-35 (2014).

175

228. Wang, X., Lupardus, P., Laporte, S.L. & Garcia, K.C. Structural biology of shared cytokine receptors. Annu Rev Immunol 27, 29-60 (2009).

229. Nelms, K., Keegan, A.D., Zamorano, J., Ryan, J.J. & Paul, W.E. The IL-4 receptor: signaling mechanisms and biologic functions. Annu Rev Immunol 17, 701-38 (1999).

230. Yost, S.E. et al. Mutascope: sensitive detection of somatic mutations from deep amplicon sequencing. Bioinformatics 29, 1908-9 (2013).

231. Twa, D.D. et al. Genomic rearrangements involving programmed death ligands are recurrent in primary mediastinal large B-cell lymphoma. Blood 123, 2062-5 (2014).

232. Ngo, V.N. et al. A loss-of-function RNA interference screen for molecular targets in cancer. Nature 441, 106-10 (2006).

233. Love, M.I., Huber, W. & Anders, S. Moderated estimation of fold change and dispersion for RNA-seq data with DESeq2. Genome Biol 15, 550 (2014).

234. Jung, T., Schrader, N., Hellwig, M., Enssle, K.H. & Neumann, C. Soluble human interleukin-4 receptor is produced by activated T cells under the control of metalloproteinases. Int Arch Allergy Immunol 119, 23-30 (1999).

235. Jung, T., Wagner, K., Neumann, C. & Heusser, C.H. Enhancement of human IL- 4 activity by soluble IL-4 receptors in vitro. Eur J Immunol 29, 864-71 (1999).

236. Guo, G. et al. Whole-genome and whole-exome sequencing of bladder cancer identifies frequent alterations in genes involved in sister chromatid cohesion and segregation. Nat Genet 45, 1459-63 (2013).

237. Krauthammer, M. et al. Exome sequencing identifies recurrent somatic RAC1 mutations in melanoma. Nat Genet 44, 1006-14 (2012).

238. Peifer, M. et al. Integrative genome analyses identify key somatic driver mutations of small-cell lung cancer. Nat Genet 44, 1104-10 (2012).

239. Frattini, V. et al. The integrated landscape of driver genomic alterations in glioblastoma. Nat Genet 45, 1141-9 (2013).

240. Kan, Z. et al. Whole-genome sequencing identifies recurrent mutations in hepatocellular carcinoma. Genome Res 23, 1422-33 (2013).

241. Kashiwada, M., Giallourakis, C.C., Pan, P.Y. & Rothman, P.B. Immunoreceptor tyrosine-based inhibitory motif of the IL-4 receptor associates with SH2- containing phosphatases and regulates IL-4-induced proliferation. J Immunol 167, 6382-7 (2001).

176

242. Hosoyama, T. et al. IL-4R drives dedifferentiation, mitogenesis, and metastasis in rhabdomyosarcoma. Clin Cancer Res 17, 2757-66 (2011).

243. Koller, F.L., Hwang, D.G., Dozier, E.A. & Fingleton, B. Epithelial interleukin-4 receptor expression promotes colon tumor growth. Carcinogenesis 31, 1010-7 (2010).

244. Venmar, K.T., Carter, K.J., Hwang, D.G., Dozier, E.A. & Fingleton, B. IL4 receptor ILR4alpha regulates metastatic colonization by mammary tumors through multiple signaling pathways. Cancer Res 74, 4329-40 (2014).

245. Kawakami, M., Kawakami, K., Kioi, M., Leland, P. & Puri, R.K. Hodgkin lymphoma therapy with interleukin-4 receptor-directed cytotoxin in an infiltrating animal model. Blood 105, 3707-13 (2005).

246. Calaminici, M., Piper, K., Lee, A.M. & Norton, A.J. CD23 expression in mediastinal large B-cell lymphomas. Histopathology 45, 619-24 (2004).

247. Hoeller, S. et al. BOB.1, CD79a and cyclin E are the most appropriate markers to discriminate classical Hodgkin's lymphoma from primary mediastinal large B-cell lymphoma. Histopathology 56, 217-28 (2010).

248. Imai, T. et al. The T cell-directed CC chemokine TARC is a highly specific biological ligand for CC chemokine receptor 4. J Biol Chem 272, 15036-42 (1997).

249. Jones, K. et al. Serum CD163 and TARC as disease response biomarkers in classical Hodgkin lymphoma. Clin Cancer Res 19, 731-42 (2013).

250. Sauer, M. et al. Baseline serum TARC levels predict therapy outcome in patients with Hodgkin lymphoma. Am J Hematol 88, 113-5 (2013).

251. Hemar, A. et al. Endocytosis of interleukin 2 receptors in human T lymphocytes: distinct intracellular localization and fate of the receptor alpha, beta, and gamma chains. J Cell Biol 129, 55-64 (1995).

252. Fujiwara, H., Hanissian, S.H., Tsytsykova, A. & Geha, R.S. Homodimerization of the human interleukin 4 receptor alpha chain induces Cepsilon germline transcripts in B cells in the absence of the interleukin 2 receptor gamma chain. Proc Natl Acad Sci U S A 94, 5866-71 (1997).

253. Kammer, W. et al. Homodimerization of interleukin-4 receptor alpha chain can induce intracellular signaling. J Biol Chem 271, 23634-7 (1996).

254. Lai, S.Y. et al. Interleukin-4-specific signal transduction events are driven by homotypic interactions of the interleukin-4 receptor alpha subunit. EMBO J 15, 4506-14 (1996).

177

255. Holtick, U. et al. STAT3 is essential for Hodgkin lymphoma cell proliferation and is a target of tyrphostin AG17 which confers sensitization for apoptosis. Leukemia 19, 936-44 (2005).

256. Ling, X. & Arlinghaus, R.B. Knockdown of STAT3 Expression by RNA Interference Inhibits the Induction of Breast Tumors in Immunocompetent Mice. Cancer Research 65, 2532-2536 (2005).

257. Shirogane, T. et al. Synergistic roles for Pim-1 and c-Myc in STAT3-mediated cell cycle progression and antiapoptosis. Immunity 11, 709-19 (1999).

258. Amit, I. et al. A module of negative feedback regulators defines growth factor signaling. Nat Genet 39, 503-12 (2007).

259. Cabrita, M.A. & Christofori, G. Sprouty proteins, masterminds of receptor tyrosine kinase signaling. Angiogenesis 11, 53-62 (2008).

260. Wertz, I.E. & Dixit, V.M. Regulation of death receptor signaling by the ubiquitin system. Cell Death Differ 17, 14-24 (2010).

261. Jiang, B.H. & Liu, L.Z. PI3K/PTEN signaling in angiogenesis and tumorigenesis. Adv Cancer Res 102, 19-65 (2009).

262. Yuan, T.L. & Cantley, L.C. PI3K pathway alterations in cancer: variations on a theme. Oncogene 27, 5497-510 (2008).

263. Ahmadian, M.R. et al. Guanosine triphosphatase stimulation of oncogenic Ras mutants. Proc Natl Acad Sci U S A 96, 7065-70 (1999).

264. Neri, A., Knowles, D.M., Greco, A., McCormick, F. & Dalla-Favera, R. Analysis of RAS oncogene mutations in human lymphoid malignancies. Proc Natl Acad Sci U S A 85, 9268-72 (1988).

265. DeBerardinis, R.J., Lum, J.J., Hatzivassiliou, G. & Thompson, C.B. The biology of cancer: metabolic reprogramming fuels cell growth and proliferation. Cell Metab 7, 11-20 (2008).

266. Hsu, P.P. & Sabatini, D.M. Cancer cell metabolism: Warburg and beyond. Cell 134, 703-7 (2008).

267. Jones, R.G. & Thompson, C.B. Tumor suppressors and cell metabolism: a recipe for cancer growth. Genes Dev 23, 537-48 (2009).

268. Gonzalez-Rodriguez, A. et al. PTP1B deficiency increases glucose uptake in neonatal hepatocytes: involvement of IRA/GLUT2 complexes. Am J Physiol Gastrointest Liver Physiol 295, G338-47 (2008).

178

269. Nieto-Vazquez, I. et al. Protein-tyrosine phosphatase 1B-deficient myocytes show increased insulin sensitivity and protection against tumor necrosis factor- alpha-induced insulin resistance. Diabetes 56, 404-13 (2007).

270. Aldinucci, D., Gloghini, A., Pinto, A., De Filippi, R. & Carbone, A. The classical Hodgkin's lymphoma microenvironment and its role in promoting tumour growth and immune escape. J Pathol 221, 248-63 (2010).

271. Carbone, A., Gloghini, A., Serraino, D. & Spina, M. HIV-associated Hodgkin lymphoma. Curr Opin HIV AIDS 4, 3-10 (2009).

272. Ma, Y. et al. The CD4+CD26- T-cell population in classical Hodgkin's lymphoma displays a distinctive regulatory T-cell profile. Lab Invest 88, 482-90 (2008).

273. Aldinucci, D. et al. Expression of CCR5 receptors on Reed-Sternberg cells and Hodgkin lymphoma cell lines: involvement of CCL5/Rantes in tumor cell growth and microenvironmental interactions. Int J Cancer 122, 769-76 (2008).

274. Hedvat, C.V. et al. Macrophage-derived chemokine expression in classical Hodgkin's lymphoma: application of tissue microarrays. Mod Pathol 14, 1270-6 (2001).

275. Peh, S.C., Kim, L.H. & Poppema, S. TARC, a CC chemokine, is frequently expressed in classic Hodgkin's lymphoma but not in NLP Hodgkin's lymphoma, T-cell-rich B-cell lymphoma, and most cases of anaplastic large cell lymphoma. Am J Surg Pathol 25, 925-9 (2001).

276. van den Berg, A., Visser, L. & Poppema, S. High expression of the CC chemokine TARC in Reed-Sternberg cells. A possible explanation for the characteristic T-cell infiltratein Hodgkin's lymphoma. Am J Pathol 154, 1685-91 (1999).

277. Kan, Z. et al. Diverse somatic mutation patterns and pathway alterations in human cancers. Nature 466, 869-73 (2010).

278. Olivier, M. & Taniere, P. Somatic mutations in cancer prognosis and prediction: lessons from TP53 and EGFR genes. Curr Opin Oncol 23, 88-92 (2011).

279. Drake, J.M. et al. Oncogene-specific activation of tyrosine kinase networks during prostate cancer progression. Proc Natl Acad Sci U S A 109, 1643-8 (2012).

280. Ko, S.J. et al. Epidermal growth factor receptor mutations and anaplastic lymphoma kinase rearrangements in lung cancer with nodular ground-glass opacity. BMC Cancer 14, 312 (2014).

179

281. Nishikawa, R. et al. A mutant epidermal growth factor receptor common in human glioma confers enhanced tumorigenicity. Proc Natl Acad Sci U S A 91, 7727-31 (1994).

282. Arcaini, L. et al. The BRAF V600E mutation in hairy cell leukemia and other mature B-cell neoplasms. Blood 119, 188-91 (2012).

283. Baohua, Y., Xiaoyan, Z., Tiecheng, Z., Tao, Q. & Daren, S. Mutations of the PIK3CA gene in diffuse large B cell lymphoma. Diagn Mol Pathol 17, 159-65 (2008).

284. Lee, J.W. et al. BRAF mutations in non-Hodgkin's lymphoma. Br J Cancer 89, 1958-60 (2003).

285. Younes, A. et al. Phase I study of a novel oral Janus kinase 2 inhibitor, SB1518, in patients with relapsed lymphoma: evidence of clinical and biologic activity in multiple lymphoma subtypes. J Clin Oncol 30, 4161-7 (2012).

286. Levis, M., Tse, K.F., Smith, B.D., Garrett, E. & Small, D. A FLT3 tyrosine kinase inhibitor is selectively cytotoxic to acute myeloid leukemia blasts harboring FLT3 internal tandem duplication mutations. Blood 98, 885-7 (2001).

287. Hexner, E.O. et al. Lestaurtinib (CEP701) is a JAK2 inhibitor that suppresses JAK2/STAT5 signaling and the proliferation of primary erythroid cells from patients with myeloproliferative disorders. Blood 111, 5663-71 (2008).

288. Hexner, E. et al. Open-label study of oral CEP-701 (lestaurtinib) in patients with polycythaemia vera or essential thrombocythaemia with JAK2-V617F mutation. Br J Haematol 164, 83-93 (2014).

289. Hedvat, M. et al. The JAK2 inhibitor AZD1480 potently blocks Stat3 signaling and oncogenesis in solid tumors. Cancer Cell 16, 487-97 (2009).

290. Scuto, A. et al. The novel JAK inhibitor AZD1480 blocks STAT3 and FGFR3 signaling, resulting in suppression of human myeloma cell growth and survival. Leukemia 25, 538-50 (2011).

291. Derenzini, E. et al. The JAK inhibitor AZD1480 regulates proliferation and immunity in Hodgkin lymphoma. Blood Cancer J 1, e46 (2011).

292. Williams, R. Discontinued in 2013: oncology drugs. Expert Opin Investig Drugs 24, 95-110 (2015).

293. Kim, M.J. et al. OPB-31121, a novel small molecular inhibitor, disrupts the JAK2/STAT3 pathway and exhibits an antitumor activity in gastric cancer cells. Cancer Lett 335, 145-52 (2013).

180

294. Hayakawa, F. et al. A novel STAT inhibitor, OPB-31121, has a significant antitumor effect on leukemia with STAT-addictive oncokinases. Blood Cancer J 3, e166 (2013).

295. Nurmohamed, M.T. & Dijkmans, B.A. Efficacy, tolerability and cost effectiveness of disease-modifying antirheumatic drugs and biologic agents in rheumatoid arthritis. Drugs 65, 661-94 (2005).

296. Siemasko, K. et al. Inhibition of JAK3 and STAT6 tyrosine phosphorylation by the immunosuppressive drug leflunomide leads to a block in IgG1 production. J Immunol 160, 1581-8 (1998).

297. Dietrich, S. et al. Leflunomide induces apoptosis in fludarabine-resistant and clinically refractory CLL cells. Clin Cancer Res 18, 417-31 (2012).

298. Nelson, E.A. et al. The STAT5 inhibitor pimozide decreases survival of chronic myelogenous leukemia cells resistant to kinase inhibitors. Blood 117, 3421-9 (2011).

299. Nelson, E.A. et al. The STAT5 Inhibitor Pimozide Displays Efficacy in Models of Acute Myelogenous Leukemia Driven by FLT3 Mutations. Genes Cancer 3, 503- 11 (2012).

300. Oh, C.K., Geba, G.P. & Molfino, N. Investigational therapeutics targeting the IL- 4/IL-13/STAT-6 pathway for the treatment of asthma. Eur Respir Rev 19, 46-54 (2010).

301. Hansel, T.T. & Barnes, P.J. New drugs and targets for asthma and COPD, xiv, 310 p. (Karger, Basel ; New York, 2010).

302. Corren, J. et al. A randomized, controlled, phase 2 study of AMG 317, an IL- 4Ralpha antagonist, in patients with asthma. Am J Respir Crit Care Med 181, 788-96 (2010).

303. Feuerhake, F. et al. NFkappaB activity, function, and target-gene signatures in primary mediastinal large B-cell lymphoma and diffuse large B-cell lymphoma subtypes. Blood 106, 1392-9 (2005).

304. Wessendorf, S. et al. Further delineation of chromosomal consensus regions in primary mediastinal B-cell lymphomas: an analysis of 37 tumor samples using high-resolution genomic profiling (array-CGH). Leukemia 21, 2463-9 (2007).

305. Ran, F.A. et al. Genome engineering using the CRISPR-Cas9 system. Nat Protoc 8, 2281-308 (2013).

306. Brough, R. et al. Functional viability profiles of breast cancer. Cancer Discov 1, 260-73 (2011).

181

307. Scuoppo, C. et al. A tumour suppressor network relying on the polyamine- hypusine axis. Nature 487, 244-8 (2012).

182