UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

The genetic basis of transformation and progression in follicular lymphoma Kridel, Robert 2016

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
24-ubc_2016_may_kridel_robert.pdf [ 11.76MB ]
Metadata
JSON: 24-1.0228776.json
JSON-LD: 24-1.0228776-ld.json
RDF/XML (Pretty): 24-1.0228776-rdf.xml
RDF/JSON: 24-1.0228776-rdf.json
Turtle: 24-1.0228776-turtle.txt
N-Triples: 24-1.0228776-rdf-ntriples.txt
Original Record: 24-1.0228776-source.json
Full Text
24-1.0228776-fulltext.txt
Citation
24-1.0228776.ris

Full Text

THE GENETIC BASIS OF TRANSFORMATION AND PROGRESSION IN FOLLICULAR LYMPHOMA by  Robert Kridel  M.D., Université Catholique de Louvain, 2004 M.P.H., University of London, 2011  A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF  DOCTOR OF PHILOSOPHY in THE FACULTY OF GRADUATE AND POSTDOCTORAL STUDIES (Pathology and Laboratory Medicine)  THE UNIVERSITY OF BRITISH COLUMBIA (Vancouver)  April 2016  © Robert Kridel, 2016 ii Abstract  The clonal evolution theory of cancer has been recognized for decades and follows principles of Darwinian selection, in which there is selection of the fittest clones in an ecosystem that is fundamentally heterogeneous and undergoes selective pressure. Follicular lymphoma (FL) emerges as a prototypical disease in which to study clonal evolution. It is the most common indolent lymphoma and, although the median overall survival largely surpasses 10 years, patients almost invariably experience progressive disease. Furthermore, a subset of FL patients is at risk of early lymphoma-related death due to rapid progression or transformation to aggressive lymphoma. Yet, the clonal dynamics and the landscape of genomic alterations underlying progression and transformation remain to be uncovered. Herein, we applied whole genome sequencing to a discovery set of transformed, progressed and non-progressed FL cases, re-constructed clonal phylogenies, and interrogated a larger set of transformed FLs and clinical extremes by capture-based targeted sequencing. Moreover, we applied the Lymph2Cx cell-of-origin assay to determine whether molecular subtypes can be defined in transformed follicular lymphoma (TFL) by gene-expression profiling. We discovered that transformation is typically the result of drastic clonal shifts during which TFL-specific clones rapidly outcompete indolent clones. In a subset of cases, these aggressive clones can be found at low levels (< 1% of tumour cells) at diagnosis. In contrast, primary progression generally results from the outgrowth of subclones that are readily detectable at diagnosis, suggesting that the genomic features conferring treatment resistance can be detected prior to initial treatment using low-pass sequencing technology. In addition, we identified discrete gene mutations that are associated with early progression and transformation, and uncovered that a subset of TFLs (16%) have an activated B-cell phenotype and are enriched for mutations in CD79B and BCL10. In summary, we found striking contrast in clonal trajectories between distinct clinical scenarios and described novel associations of gene mutations with transformation and progression. Our findings have translational relevance as they suggest that early progression, and to a lesser extent transformation, can potentially be predicted at diagnosis and thus provide a rationale for novel therapeutic approaches in TFL. iii Preface  The work presented in Chapters 2 and 3 was performed under the auspices of the Department for Lymphoid Cancer Research at the BC Cancer Research Centre, and funded by a Program Project Grant from the Terry Fox Research Institute (Grant No. 1023). Chapter 3 was outlined in sub-project 1 of the grant that I co-wrote with Dr. Sohrab Shah and for which I contributed preliminary data. The idea to determine cell-of-origin in transformed follicular lymphoma arose in discussions with Dr. David Scott. The work presented within this thesis is approved by the UBC Research Ethics Board (Certificate Number H13-01765). Chapter 2 has been published in its entirety (see full reference at the end of the preface). I performed the following aspects of the work described in this chapter: assembling of patient cohort, curating of clinical data, extraction of DNA and RNA, Lymph2Cx assay, fluorescence in situ hybridization (FISH) break-apart assay scoring, data analyses with exception of mutation calling, and writing of the manuscript. Dr. Anja Mottok is a co-first author on this manuscript. She coordinated the pathological aspects of this study, confirmed pathological diagnoses and scored immunohistochemistry. The contribution of all other authors is summarized herein: D.W.S., and R.D.G. designed and performed the research, analyzed and interpreted data, and wrote the paper; P.F. and K.T. performed pathological review of cases; S.B.-N. and D.E. performed FISH; Y.Z. performed statistical analysis; E.A.C. performed hybrid-capture, library construction, and the sequencing run; H.P.S. analyzed next-generation sequencing data; M.B., B.M., and A.T. performed RNA and DNA extractions; F.C.C. provided bioinformatics assistance; L.H.S. and J.M.C. supervised assembly of clinical data, reviewed the manuscript, and provided editorial input; C.S. participated in the original design of the project, reviewed the manuscript, and provided editorial input; and M.A.M. and S.P.S. participated in the original design of the project. Regarding the work presented in Chapter 3, I contributed to the design of the study and put the emphasis on studying both transformation and early progression, assembled patient cohorts, flow-sorted samples, extracted DNA and RNA, contributed to primer design and performed amplicon deep-sequencing, scored FISH assays, performed digital droplet PCR and analyzed capture sequencing results. I generated all the figures that are referenced in Chapter 3, with the exception of Figures 13-16 that have been prepared by Fong Chun Chan. Fong will be a co-first author on the manuscript emanating from this work. In addition to generating the above-iv mentioned figures, he analyzed whole-genome sequencing data, analyzed deep amplicon sequencing data and performed clonal analysis using PyClone. He also organized and maintained the code for the project and provided assistance whenever my understanding of R and bioinformatics reached their limits (i.e. often). Mutation calling of capture sequencing data was performed by Dr. Ali Bashashati and Diljot Grewal. The project was supervised by Dr. Sohrab Shah.         Reference for Chapter 2: “Cell-of-origin of transformed follicular lymphoma”. Kridel R,* Mottok A,* Farinha P, Ben-Neriah S, Ennishi D, Zheng Y, Chavez EA, Shulha HP, Tan K, Chan FC, Boyle M, Meissner B, Telenius A, Sehn LH, Marra AM, Shah SP, Steidl S, Connors JM, Scott DW, Gascoyne RD. Blood 2015 Oct 29;126(18):2118-27, PMID 26307535 (* denotes equal contribution).  v Table of Contents  Abstract .......................................................................................................................................... ii	Preface ........................................................................................................................................... iii	Table of Contents ...........................................................................................................................v	List of Tables ................................................................................................................................ ix	List of Figures .................................................................................................................................x	List of Abbreviations ................................................................................................................... xi	Acknowledgements .................................................................................................................... xiii	Dedication ................................................................................................................................... xiv	Chapter 1: Introduction ................................................................................................................1	1.1	 Normal B-cell development and its relationship to B-cell lymphoma ............... 1	1.1.1	 Physiological state ........................................................................................ 1	1.1.2	 Origin of B-cell lymphomas ......................................................................... 3	1.2	 Follicular lymphoma ........................................................................................... 4	1.2.1	 Clinical description ....................................................................................... 4	1.2.2	 Early steps of FL pathogenesis ..................................................................... 6	1.2.3	 Recurrent genetic alterations ........................................................................ 7	1.2.3.1	 KMT2D .................................................................................................. 8	1.2.3.2	 CREBBP and EP300 .............................................................................. 8	1.2.3.3	 EZH2 ...................................................................................................... 9	1.2.3.4	 STAT6 .................................................................................................... 9	1.2.3.5	 RRAGC, ATP6V1B2 and ATP6AP1 .................................................... 10	1.2.3.6	 IGHV/IGLV mutations and unusual glycans ........................................ 10	1.2.4	 Role of the microenvironment .................................................................... 11	1.2.4.1	 Follicular T helper cells ....................................................................... 11	1.2.4.2	 T regulatory cells ................................................................................. 12	1.2.4.3	 Cytotoxic T cells .................................................................................. 12	1.2.4.4	 Macrophages ........................................................................................ 12	1.2.4.5	 Stromal cells ........................................................................................ 13	1.3	 Progression of follicular lymphoma .................................................................. 13	vi 1.4	 Transformation of follicular lymphoma ............................................................ 14	1.4.1	 Clinical aspects ........................................................................................... 14	1.4.2	 The genetic basis of transformation ............................................................ 15	1.5	 Outcome prediction in follicular lymphoma ..................................................... 17	1.5.1	 Pathology .................................................................................................... 17	1.5.2	 Genetics ...................................................................................................... 18	1.6	 Thesis theme and objectives ............................................................................. 19	1.7	 Hypotheses ........................................................................................................ 20	1.7.1	 Hypothesis 1 ............................................................................................... 20	1.7.2	 Hypothesis 2 ............................................................................................... 20	1.7.3	 Hypothesis 3 ............................................................................................... 20	1.8	 Aims and thesis outline ..................................................................................... 20	1.8.1	 Aim 1: Cell-of-origin of transformed follicular lymphoma ....................... 20	1.8.2	 Aim 2: Evolutionary dynamics in follicular lymphoma ............................. 20	1.8.3	 Aim 3: Prediction of early transformation/progression. ............................. 20	Chapter 2: Cell-of-origin of transformed follicular lymphoma ..............................................22	2.1	 Introduction ....................................................................................................... 22	2.2	 Methods............................................................................................................. 23	2.2.1	 Patient samples and definition of endpoints ............................................... 23	2.2.2	 Tissue microarray, FISH and IHC .............................................................. 24	2.2.3	 Lymph2Cx assay ........................................................................................ 25	2.2.4	 Targeted sequencing of CARD11, CD79B and MYD88 ............................. 25	2.2.5	 Statistical analysis ....................................................................................... 25	2.3	 Results ............................................................................................................... 26	2.3.1	 Association of clinical/pathological characteristics with transformation ... 26	2.3.2	 Temporal patterns in translocation status and IHC staining ....................... 26	2.3.3	 Survival after transformation by histology and double-hit status ............... 27	2.3.4	 COO assignment of transformed lymphoma .............................................. 27	2.3.5	 Correlation of COO with outcome and cytogenetics .................................. 28	2.3.6	 Mutations in CARD11, CD79B and MYD88 .............................................. 28	2.4	 Discussion ......................................................................................................... 29	vii Chapter 3: Evolutionary dynamics of follicular lymphoma ....................................................32	3.1	 Introduction ....................................................................................................... 32	3.2	 Methods............................................................................................................. 33	3.2.1	 Patients and materials ................................................................................. 33	3.2.2	 Pathology .................................................................................................... 34	3.2.3	 Whole-genome sequencing and analysis .................................................... 35	3.2.4	 Targeted deep amplicon sequencing and analysis ...................................... 36	3.2.5	 Clonal analysis ............................................................................................ 37	3.2.6	 Digital droplet PCR .................................................................................... 38	3.2.7	 Capture-based targeted sequencing ............................................................ 39	3.3	 Results ............................................................................................................... 40	3.3.1	 Whole genome sequencing results .............................................................. 40	3.3.2	 The emergence of a T2-dominant clone underlies transformation ............. 41	3.3.3	 Primary progression results from the selection of pre-existent clones ....... 42	3.3.4	 Ultra-sensitive identification of low prevalence clones in T1 samples ...... 43	3.3.5	 Recurrent gene alterations in transformed follicular lymphoma ................ 44	3.3.6	 Association of gene mutations with early progression ............................... 46	3.4	 Discussion ......................................................................................................... 47	Chapter 4: Conclusion .................................................................................................................50	4.1	 Summary of research findings .......................................................................... 50	4.2	 Integration into the research field ..................................................................... 51	4.2.1	 Intratumoural heterogeneity and clonal evolution ...................................... 51	4.2.2	 The transcriptomic and genetic composition of TFL ................................. 53	4.2.3	 Prediction of early transformation and/or progression ............................... 53	4.3	 Strengths and limitations ................................................................................... 54	4.3.1	 Intratumoural heterogeneity and clonal evolution ...................................... 54	4.3.2	 The transcriptomic and genetic composition of TFL ................................. 55	4.3.3	 Prediction of early transformation and/or progression ............................... 55	4.4	 Potential applications ........................................................................................ 55	4.5	 Ongoing work ................................................................................................... 56	4.6	 Open questions and future directions ................................................................ 57	viii 4.6.1	 Order of mutations and relation to B-cell maturation ................................. 57	4.6.2	 The extent of intra-tumoural heterogeneity in B-cell lymphomas .............. 57	4.6.3	 Non-genetic drivers of intra-tumoural heterogeneity ................................. 58	4.6.4	 Non-tumour cell intrinsic features .............................................................. 58	4.6.5	 Mechanism of treatment resistance ............................................................ 59	4.7	 Final conclusion ................................................................................................ 60	Tables ............................................................................................................................................62	Figures ...........................................................................................................................................68	References .....................................................................................................................................93	Appendices ..................................................................................................................................126	Appendix A - 86 gene panel for capture sequencing ................................................ 126	Appendix B - 20 somatic hypermutation gene panel for capture sequencing .......... 129	 ix List of Tables  Table 1. Clinical characteristics of 2820 newly diagnosed FL patients .............................. 62	Table 2. Clinical characteristics of early and late transformers .......................................... 63	Table 3. Weighted Cox regression model for time to transformation ................................. 64	Table 4. Clinical characteristics of TFL cohort ................................................................... 65	Table 5. Mutations in CARD11, CD79B and MYD88 ......................................................... 66	Table 6. Clinical characteristics of early and late progressers ............................................ 67	 x List of Figures  Figure 1. Genes mutated in >5% of FL cases ...................................................................... 68	Figure 2. The tumour microenvironment in follicular lymphoma ...................................... 69	Figure 3. Overlap between study cohorts ............................................................................ 70	Figure 4. Association of pathological characteristics with transformation ......................... 71	Figure 5. BCL2, BCL6 and MYC translocation in FL and TFL ........................................... 72	Figure 6. CD10, BCL6 and IRF4 expression in FL and TFL ............................................. 73	Figure 7. Survival correlates from time of transformation .................................................. 74	Figure 8. Lymph2Cx assay in 107 TFL cases with DLBCL morphology .......................... 75	Figure 9. Survival and cytogenetic correlates of molecular subtypes in TFL ..................... 76	Figure 10. Association of pathological findings in FL and COO of TFL ........................... 77	Figure 11. Overview of study cohort used for Chapter 3 .................................................... 78	Figure 12. Sample overview and timeline of WGS cohort ................................................. 79	Figure 13. High-level WGS analysis overview ................................................................... 80	Figure 14. Genetic alterations by timepoint and by clinical category ................................. 81	Figure 15. Clonal phylogenies of TFL patients ................................................................... 82	Figure 16. Clonal phylogenies of PFL patients ................................................................... 83	Figure 17. Small subclones in T1 samples .......................................................................... 84	Figure 18. Validation of small subclonal mutations in T1 by ddPCR ................................. 85	Figure 19. Recurrent mutations in all T1 (FL) samples (n = 172) ...................................... 86	Figure 20. Results from targeted sequencing in TFL samples ............................................ 87	Figure 21. Mutations in areas of SHM for FL and TFL ...................................................... 88	Figure 22. Proportion of shared and T1 or T2-specific mutations ...................................... 89	Figure 23. B2M mutations and CD8+ T cells ...................................................................... 90	Figure 24. Results from targeted sequencing in early and late progressers ........................ 91	Figure 25. Mutations in areas of SHM for early and late progressers ................................. 92	 xi List of Abbreviations  ABC  - activated B-cell AID  - activation-induced deaminase BCR  - B-cell receptor CI  - confidence interval CLL  - chronic lymphocytic leukemia CNA  - copy number alteration COO  - cell-of-origin CPC  - common progenitor cell CTL  - cytotoxic T cell DLBCL - diffuse large B-cell lymphoma ECOG - Eastern Cooperative Oncology Group performance status ER  - endoplasmic reticulum FDC  - follicular dendritic cell FFPET - formalin-fixed and paraffin-embedded tissue FTH  - follicular T helper cell GC  - germinal centre GCB  - germinal centre B-cell HR  - hazard ratio ITH  - intratumoural heterogeneity FL  - follicular lymphoma LDH  - lactate dehydrogenase LOH  - loss of heterozygosity NA  - not available NHL  - Non-Hodgkin lymphoma NPFL - non-progressed follicular lymphoma PFL  - progressed follicular lymphoma SHM  - somatic hypermutation SNV  - single nucleotide variant TAM  - tumour-associated macrophages xii TFL  - transformed follicular lymphoma ULN  - upper limit of normal WGS  - whole genome sequencing xiii Acknowledgements  The work presented in this thesis would not exist without the exceptional research environment that my supervisor, Dr. Randy Gascoyne, has established within the Department of Lymphoid Cancer Research at the BC Cancer Research Centre. I would like to express my sincere gratitude to Randy for his guidance and mentorship over the last 5 years.  In addition, I was co-supervised by Dr. Christian Steidl who was tremendously generous in providing constructive advice whenever my research was stalled.  Dr. Sohrab Shah supervised the work described in Chapter 3. His goal for achieving perfect data has pushed us to give our best.  I am also grateful to the Chair and Members of my committee (Dr.’s David Huntsman, Randy Gascoyne, Christian Steidl, Sohrab Shah and Keith Humphries) who helped me stay focused, and guided my progress.  I am indebted to the Terry Fox Research Institute that funded the research presented herein through a Program Project Grant  I would like to thank my fellow lab mates and colleagues for fruitful discussions and help provided along the way. My special thanks go to Dr. Anja Mottok, Dr. David Scott, Merrill Boyle, Susana Ben-Neriah, Dr. King Tan, Dr. Pedro Farinha, Elizabeth Chavez, Dr. Barbara Meissner, Dr. Daisuke Ennishi, Adele Telenius and Bruce Woolcock.  Lastly, I would like to thank my family and my friends for their moral support. xiv Dedication  This dissertation is dedicated to my parents. Their unconditional support and selflessness allowed me to pursue higher education and follow my dreams. 1 Chapter 1: Introduction  1.1 Normal B-cell development and its relationship to B-cell lymphoma 1.1.1 Physiological state The first description of B cell effector roles predates the discovery of distinct lymphocyte lineages and goes back in time to the 19th century when, in 1890, antitoxins to diphtheria and tetanus were shown to play a role in humoral immunity.1 Subsequently, antibody responses were for the first time associated with plasma cells in 19482 and the description of recurrent bacterial infections in boys with absent γ-globulins, plasma cells and germinal centres (GCs) further documented a crucial function of B cells in humoral immunity.3,4 Separate B and T cell lineages were first identified in 1965 in chickens through the discovery that neonatal removal of the bursa of Fabricius or the thymus resulted in defects of humoral and cellular immunity, respectively.5,6 The earliest B cells develop in the bone marrow from common lymphoid progenitors under the concerted action of stromal cells and secretion of IL-7, which results in expression of the key transcription factors EBF1, E2A and PAX5.7–13 The subsequent developmental stages are defined by careful orchestration of surface receptor expression, the rearrangement of immunoglobulin loci and tightly regulated transcriptional networks. The first B cells to express a B-cell receptor (BCR) are termed pre-B cells and express the pre-BCR that comprises a µ heavy chain and a surrogate light chain that consists of a heterodimer of λ5 and VpreB.14,15 Diversity of BCR sequences is a hallmark of B cells, underlying the ability of B cells to respond to virtually any antigen. Diversity starts to be generated at the pre-B stage, when the IGH locus undergoes ordered rearrangement of a DH to a JH (DHJH) segment,16 followed by the juxtaposition of a VH segment to DHJH (VHDHJH).17,18 These steps are followed by rearrangement of a light chain locus, allowing for the expression of surface IgM, and are under the control of recombination activating genes 1/2 (RAG 1/2).19,20 In addition to combinatorial diversity, junctional diversity contributes to a heterogeneous antibody repertoire.21 The diversity of possible Ig sequences underlies the clonal selection theory in which pre-formed antibodies expand through positive selection by an antigen and is a crucial aspect of the adaptive immune system. When immature B cells exit the bone marrow, they are transitional B cells that express surface IgM and IgD and are yet to encounter their cognate antigen. They home to lymphoid follicles in secondary lymphoid organs including lymph nodes, spleen and gut-associated 2 lymphoid tissue. Whereas, at least in mice, certain B-cell subsets such as B1 cells, marginal zone B cells or short lived plasma cells generate IgM antibody responses independently of T-cells, typical B-cell maturation into memory or terminally differentiated plasma cells requires passage through the germinal centre, a highly organized structure in which selection of high-affinity antibody-expressing B cells occurs. B cells encounter antigen in primary follicles in its soluble form, or presented by subcapsular macrophages or follicular dendritic cells (FDCs), but become fully activated only after forming long-lived interactions with antigen-specific CD4+ T helper cells in the boundary between the T cell and B cell zones to which they migrate under the influence of CCR7-ligand.22 B cells with the highest density of peptide bound to major histocompatibility class II monopolize T cell help, migrate into primary follicles and start proliferating, thereby establishing the early GC reaction.23 On day 7 after initiation of the GC reaction, a polarization into two functionally distinct compartments occurs. The dark zone is comprised of densely packed and rapidly proliferating B cells expressing CXCR4, whereas the light zone is less compact and comprises CD83 and CD86-expressing light zone B cells, FDCs and macrophages. The organization of the GC into distinct zones is regulated by expression of CXCR5 that directs B cells to the chemokine CXCL13 rich light zone and, conversely, expression of CXCR4, driving migration to the dark zone.24 In the dark zone the V genes of highly proliferating B cells undergo somatic hypermutation (SHM) driven by activation-induced deaminase (AID),25 thereby diversifying the potential pool of BCRs. Subsequent selection in the light zone of BCRs with high affinity is proportional to follicular T cell help and is mediated by the amount of peptide-MHC that B cells present to T cells26 as well as binding of CD40 on B cells to CD40L on T cells. The strength of T cell help also influences the number of cell divisions that B cells subsequently undergo in the dark zone, as well as the amount of hypermutation,27 contributing to affinity maturation and clonal expansion over the course of cyclic re-entry from light to dark zone. It is in the light zone where cell fate decisions regarding effector function occur through class-switch recombination that is controlled by AID,25 and differentiation into memory B cells or plasma cells. The GC reaction is carefully regulated by changes in the expression of key transcription factors and epigenetic modifiers that are also altered in B-cell lymphomas (see section 1.1.2). The master regulator of the GC reaction is the transcriptional repressor BCL6 that is required for entry into the GC28 and its maintenance by repression of B cell activation and differentiation, the latter 3 occurring via downregulation of PRDM1 and IRF4.29 BCL6 also dampens the DNA damage response by repression of TP53,30 CDKN1A,31 ATR32 and CHEK1,33 and thereby creates a permissive milieu for SHM to take place. Whereas most GC B cells do not express the MYC proto-oncogene at a significant level,34 MYC was recently found to be indispensable for the initiation and maintenance of GCs. Despite the high proliferation rates that are observed in the dark zone and the role that MYC plays in cell cycle regulation, MYC is not expressed in the dark zone but in selected subsets of GC B cells, its expression being induced at the stage of GC commitment and in light zone B cells that are actively selected for re-entry into the dark zone.35,36 The histone methyltransferase EZH2 is specifically expressed in the GC and was recently shown to be required for GC formation by promoting a transcriptional profile that represses B cell differentiation through establishment of bivalent, poised chromatin domains at promoters of key sets of genes including IRF4 and PRDM1.37 1.1.2 Origin of B-cell lymphomas The GC is a unique structure as it is physiologically characterized by high proliferation rates, generation of SHM and resistance to cell death, all of which are hallmarks of malignancy.38 It is therefore not surprising that the immune system's capacity to generate rapid and specific responses to new antigens comes at the expense of an inherent risk of developing GC-derived malignancies. The two main mechanisms that generate antibody diversity, namely SHM and CSR, are implicated in aberrant hypermutation of proto-oncogenes39 and chromosomal translocations.40 The role that these processes play in lymphomagenesis is illustrated by the description that deficiency of AID, the enzyme that generates both SHM and CSR, abrogates BCL6-driven GC lymphoma development in a mouse model.41 Similarly, the absence of AID prevents Plasmodium-associated lymphomas in a model of B-cell specific deletion of the tumour suppressor TP53.42 The cellular origin of mature B-cell lymphomas is usually inferred by comparing the morphology, immunophenotype and gene-expression profile to specific physiological B-cell maturation stages.43 For example, mantle cell lymphoma is thought to arise from naive B cells based on CD5 and IgM/IgD expression44 (although 16-29% of cases present evidence of SHM in IGHV genes, suggesting passage through the GC),45–47 and marginal zone lymphoma seemingly originates from marginal zone B cells. The taxonomy of de novo DLBCL divides this disease into molecular subgroups that differ by gene-expression profile, activated pathways and patient outcomes.48–52 The most widely used classification divides DLBCL into 2 distinct groups, the 4 germinal centre B-cell (GCB) and the activated B-cell (ABC) subtypes. As suggested by its name, GCB-DLBCL is characterized by expression of genes that are also highly expressed in normal GC B cells, and the ABC subtype shares a common gene expression program with B cells that are activated in vitro.48 Regarding FL, the typical follicular growth pattern, expression of the GC markers CD10 and BCL6 as well as SHM of IGHV sequences liken this lymphoma to normal germinal centre B cells, with the subtlety that the subset of t(14;18)-negative FLs has a phenotype and a gene expression profile that is rather reminiscent of post-GC B cells or ABC-DLBCL.53 The GC expression profile attribution of FL and GCB-DLBCL has been further refined by the description that both of these entities share a light zone rather than a dark zone phenotype based on gene expression profiling.54,55 The concept of cell-of-origin is semantic per se as it does not imply that malignant cells directly originate from cells to which they are compared to. In FL, for example, the prototypical t(14;18) translocation is thought to occur in the bone marrow as a result of defective VDJ recombination and to subsequently favour the acquisition of secondary mutations during multiple rounds of GC reentry.56 Nonetheless, GC-derived B cell lymphomas characteristically harbour genetic alterations in genes that are critical for normal GC physiology such as chromosomal rearrangements of BCL6 or MYC, and mutations of EZH2 or KMT2D, illustrating that they hijack normal pathways, leading to a sustained GC response and blockade of terminal B cell differentiation. 1.2 Follicular lymphoma 1.2.1 Clinical description It is estimated that non-Hodgkin lymphoma (NHL) (excluding chronic lymphocytic leukemia and myeloma) represents the 7th most common cancer type among new cancer diagnoses in the US in 2015 (after breast, lung, prostate, colorectal, urinary bladder and melanoma).57 The total number of new NHL diagnoses is estimated at 71,850 for the US in 2015,57 which can be translated into an incidence of 23 new cases/100,000 persons/year. Worldwide, FL is the 2nd most common NHL subtype, representing 22% of new diagnoses, placing it second only to DLBCL (31%).58 In North America, FL is at least as common as DLBCL (31-32% vs 28-29%),59 indicating that FL occurs at an incidence of 7 new cases/100,000 persons/year. This incidence rate translates into 2,500 and 330 new FL cases in 2015 for all of Canada and the province of British Columbia, 5 respectively. The prevalence of FL is more difficult to ascertain, but is higher than its incidence given that FL is mostly a chronic disease. Based on 2,820 FL cases drawn from the clinical database of the Lymphoma Tumour Group at the BC Cancer Agency (Table 1), the median age at diagnosis is 60 years (interquartile range 21 years). FL is slightly more common in males than in females (51.6% versus 48.4%), and 73.1% and 26.9% of newly diagnosed patients present with advanced and limited stage, respectively. FL is a prototypical indolent lymphoma, progressing slowly in the vast majority of patients, yet ultimately progressing and in the end incurable. The 10 year disease-specific survival was 59.9% for all patients in the BC Cancer Agency database diagnosed prior to 2000, and 67.8% for patients diagnosed since 2000 (Table 1). Regarding the prediction of prognosis, the International Prognostic Index (IPI) that has been developed for aggressive lymphomas, is predictive of outcome in FL.58,60 The FL International prognostic index (FLIPI) was proposed as an improved prognostication tool, although its superiority to the IPI has not formally been demonstrated.61 The FLIPI combines five clinical risk factors (age > 60 years, Ann Arbor stage III-IV, hemoglobin level <120 g/L, number of nodal areas > 4 and serum LDH > normal) and groups patients into 3 risk groups: low risk (0-1 adverse factor), intermediate risk (2 factors) and high risk (≥ 3 factors). In a retrospective analysis of 362 patients treated within a clinical trial of the German Low Grade Lymphoma Study Group with R-CHOP, 45% of patients were classified as high risk by FLIPI and had a shorter 2-year time to treatment failure (67%) than the 14% and 41% of patients classified as low and intermediate risk, respectively (2-year time to treatment failure 92% and 90%).62 A modified version of the FLIPI that incorporates measurements of serum β2-microglobulin, longest lymph node diameter and bone marrow involvement was subsequently proposed (FLIPI2) but it is unclear whether this tool improves prognostication when compared to the FLIPI.63 For patients with limited stage FL, involved-field radiation therapy is the treatment modality of choice as per the guidelines from the European Society of Medical Oncology (ESMO),64 and one of the treatment options recommended by the National Comprehensive Cancer Network (NCCN) in the US.65 Consistently, this treatment approach achieves long-term freedom from relapse rates of 40-50%.66–71 For patients with advanced stage disease, treatment decisions are based on the extent of the disease and on whether the patient presents with symptoms. For low-burden and asymptomatic patients, several retrospective studies72,73 and randomized trials74,75 have 6 shown that deferring treatment is safe, with 50% of patients who are initially observed free of disease progression at 2-2.6 years.74,75 Patients with high burden of disease and/or symptoms require systemic therapy that traditionally consisted of multi-agent chemotherapy. The addition of rituximab, a monoclonal anti-CD20 antibody, to chemotherapy was the first intervention to definitively show a survival benefit in FL.76–78 More recently, the combination of bendamustine and rituximab has emerged as a safer and more efficacious treatment option when compared to the previously more commonly used R-CHOP or R-CVP regimens,79,80 and has become a standard of care for patients requiring upfront immuno-chemotherapy. Finally, rituximab maintenance has been shown to prolong progression-free survival when given after first-line rituximab-containing therapy, with little added toxicity.81 1.2.2 Early steps of FL pathogenesis The first genetic alteration to occur in FL is believed to be the t(14;18)(q32;q21) reciprocal translocation, based on its occurrence in 85% of cases, its consistent presence at a clonal level, and the mechanism by which it arises.82,83 Indeed, the break on chromosome 14 results from faulty DH to JH recombination and is mediated by the RAG complex, which is active at the pro-B to pre-B maturation stages.84 Hence, the first oncogenic hit is thought to occur in the bone marrow prior to antigen exposure. On chromosome 18, breaks occur in 3 clusters (MBR, major breakpoint region; icr, intermediate cluster region; mcr, minor cluster region) that are void of recombination signal sequences. Rather than through mistakes in VDJ recombination, breaks on chromosome 18 were shown to occur in close vicinity to CpG sites and are thought to be the result of the sequential activity of AID and RAG.85,86 As AID is expressed in human pro-B and pre B-cells,87 such a model is compatible with the t(14;18) occurring in the bone marrow during B cell maturation. The t(14;18) translocation results in overexpression of anti-apoptotic protein BCL2.88–90 The question arises as to how increased levels of BCL2 protein lead to lymphomagenesis as BCL2 by itself has not only anti-apoptotic, but also anti-proliferative functions,91 and as the t(14;18) translocation is common in healthy individuals (see below). It was shown by Sungalee et al. that BCL2 transgene expression enhances the capacity of memory B cells to expand as part of iterative cycles of GC re-entry following immunization, in addition to increasing mutational load, both in GC and memory B cells.56 The majority of observed variants in this study were G to A and C to T transitions, suggesting that they arose through AID-mediated SHM. The consideration that BCL2-expressing memory B cells expand upon antigenic stimulation and give rise to in situ FL parallels 7 the observation that in normal individuals with circulating t(14;18), the translocation is restricted to IgD+CD27+ (or IgM+CD27+) memory cells that have undergone the GC reaction.92,93 Clearly, additional, secondary hits are required for lymphomagenesis as the t(14;18) translocation is found in 50-70% of healthy individuals, most of which will never develop FL.94–96 The presence of the translocation is nonetheless a risk factor for developing FL, at least when present at high levels.97,98 When assessable, in patients who developed overt FL after detection of pre-clinical circulating t(14;18), a clonal relationship was consistently confirmed between the pre-malignant and the FL states,97,98 with detection of t(14;18) up to 15 years prior to FL diagnosis. These findings suggest that precursor cells are long-lived and can evolve slowly. This has also been elegantly shown by rare studies in which donors and recipients of stem cell transplants for conditions unrelated to FL developed clonally-related FL several years after the transplant.99,100 One possible counterpart of cells harbouring the t(14;18) in human tissue is a situation characterized by individual GCs co-expressing BCL2 and CD10, in an otherwise relatively reactive lymph node, and residual reactive, BCL2-negative GCs.101,102 This entity has been named FL in situ and, in the absence of overt, concurrent or prior FL, the risk of progression to FL is low (5%).102 Nonetheless, a small number of copy number alterations have been described using comparative genomic hybridization (mean number of alterations per sample: 2.5), suggesting that FL in situ is a neoplastic process, albeit indolent, and that it harbours secondary hits, in addition to the t(14;18). FL in situ needs to be distinguished from partial lymph node involvement by otherwise typical FL, which can be found in limited stage FL patients at the time of diagnosis and is associated with a higher risk of progression to FL.102,103 1.2.3 Recurrent genetic alterations The FL genome is characterized by recurrent translocation and copy number changes. The t(14;18) translocation can be found in 85% of unselected cases, whereas BCL6 is translocated to IGH or other partners in 6-14% of cases.104–106 Translocations involving the MYC gene are rare at diagnosis.107–109 Certain copy number alterations are recurrent and include, for example, deletion of 1p36, deletion of 6q, trisomy 7, trisomy 12, deletion of 17p and duplication of chromosome X.83,110–114 Prior to the advent of high-throughout sequencing technology, only a handful of genes were know to be recurrently altered in FL, based on copy number analyses, targeted sequencing or functional screens.115 They include EPHA7,116 TNFSRF14,117,118 TNFAIP3,116,119,120 FAS121 and TP53.122–124 Our understanding of the genetics underlying FL has grown considerably since next 8 generation sequencing technology became widely available. For example, the discovery of frequent mutations in histone-modifying genes was unexpected and has significantly contributed to our understanding of GC biology and ontogeny of FL over the last few years. Recurrently mutated genes have been reported in multiple studies and125–128 Figure 1 shows the most commonly mutated genes in FL, based on our own data.128 The remainder of this section will focus on individual genes and describe how mutations in these genes contribute to lymphomagenesis. 1.2.3.1 KMT2D KMT2D, also known as MLL2 or MLL4, belongs to the family of SET1 histone lysine methyltransferases, and specifically catalyzes mono- and dimethylation of lysine 4 on histone 3 (H3K4), modifications that are associated with active transcription.129,130 KMT2D is frequently mutated in FL (24-90%, depending on the series),125,128,131,132 and the majority of alterations are nonsense and frameshift mutations that are scattered throughout the coding sequence and are predicted to disrupt function by preventing translation of the C-terminal SET domain, the catalytic unit of KMT2D. Recently, two groups independently showed that KMT2D deficiency promotes lymphomagenesis in vivo.133,134 Zhang et al. reported that loss of KMT2D in GC B cells was associated with global H3K4 hypomethylation,134 whereas Ortega-Molina et al. did not find a global decrease in H3K4 methylation in their model, but rather specific depletion at enhancer sites.133 Both groups found significant differences in gene expression profiles between wild-type and knock-out or knock-down cells and pathways that emerged as being regulated by KMT2D include, for example, CD40, JAK-STAT or BCR signalling.133,134 Consistently, mutations in KMT2D have been assigned to clones, rather than subclones, and inferred to be present in the earliest progenitors, suggesting that KMT2D disruption is an early event in the oncogenic cascade.126,135,136 1.2.3.2 CREBBP and EP300 The histone acetyl transferases CREBBP and closely related EP300 are transcriptional coactivators that increase transcription137 through lysine acetylation of histones138,139 or non-histone proteins. The latter include both transcriptional activators such as TP53,140–142 for example, leading to their activation, and transcriptional repressors such as BCL6,143 leading to their inactivation. CREBBP and EP300 are mutated in 33-68% and 9% of FL cases, respectively,126,128,144 and mutations typically consist in heterozygous nonsense or frameshift alterations, as well as in missense mutations located within the sequence encoding the histone 9 acetyltransferase (HAT) domain.144 Although CREBBP is also commonly deleted, at least in DLBCL, deletions and heterozygous mutations are reported to be mostly mutually exclusive, and CREBBP is hence affected by predominantly monoallelic alterations, suggesting haploinsufficiency.144 Mutations within the HAT domain were shown to cause deficient acetylation of BCL6 and TP53, as well as deficient cAMP-responsive transcription.144 A recent study reported that CREBBP mutations were associated with a distinct gene expression profile in primary patient samples that includes decreased expression of MHC class II, suggesting that CREBBP mutations may lead to escape from immune surveillance.126 Akin to KMT2D, CREBBP mutations have been inferred to represent early alterations in tumour evolution as they are almost exclusively clonal.126,135,136,145 1.2.3.3 EZH2 EZH2 is a histone methyltransferase and the catalytic subunit of the polycomb repressive complex 2 (PRC2). EZH2 catalyzes methylation of lysine 27 of histone H3 via its SET domain, thereby mediating gene silencing.146–148 Under physiological conditions, EZH2 plays crucial roles in stem cell maintenance and cell differentiation fate decisions.149,150 In B cell lymphopoiesis, EZH2 is abundantly expressed in early B cell precursors and early deletion leads to impairment of B cell development.151 EZH2 is not significantly expressed in subsequent stages of B cell development, but is upregulated during the GC reaction.152 Although targets of EZH2-mediated repression in GC B cells overlap with genes targeted in embryonic stem cells, GC-specific genes are recognized and include tumour suppressors.153 More recently, using a conditional knock-out of EZH2 in GC B cells, Béguelin et al. showed that EZH2 is required for GC formation.37 In FL, EZH2 is mutated in 30% of patients, with virtually all mutations found in the SET domain and over 80% of mutations targeting amino acid residue Y641.126,128,154–157 Whereas EZH2 Y641 mutations were initially thought to lead to loss of function,154 they were shown to drive hypermethylation of H3K27.158,159 Consistent with gain-of-function, conditional knock-in of a mutant EZH2 allele leads to GC hyperplasia, and accelerates lymphomagenesis in cooperation with BCL2 or MYC.37,160 The discovery of hotspot mutations in EZH2 has led to the rapid development of potent small molecular inhibitors that are currently in clinical development.161,162 1.2.3.4 STAT6 The JAK-STAT signalling pathway is used to transduce signals from extracellular cytokines and growth factors to the nucleus and eventually lead to specific gene expression 10 changes in response to signals from the microenvironment.163 STAT6 has an essential role in IL4 mediated signalling.164,165 Clues that the IL4-STAT6 axis is relevant in FL come from the observations that follicular T helper (FTH) cells are enriched in the tumour microenvironment of FL and that high levels of FTH-derived IL4 stimulate STAT6 signalling.166–168 STAT6 is mutated in 11% of FL and the majority of mutations fall into the DNA binding domain, with several recognized hotspots, the most commonly targeted amino acid residue being D419.128,169 STAT6 mutations are typically heterozygous and have been shown to be activating, enhancing the response to IL4 stimulation.169 1.2.3.5 RRAGC, ATP6V1B2 and ATP6AP1 Mutations in these genes have recently been reported in 17% of FL cases for RRAGC, 11-22% for ATP6V1B2 and 10-12% for ATP6AP1.126,170 RRAGC mutations are remarkable as they are infrequent in DLBCL and are not observed in other hematological malignancies, suggesting that they are relatively specific to FL. The protein products from all three genes form part of a supercomplex that activates mTORC1 in a state of amino acid sufficiency. RRAGC mutants were shown to be activating as their over-expression led to mTORC1 activation, even when cells were deprived of amino acids.170 mTOR inhibition therefore emerges as a potential strategy to treat follicular lymphoma and it remains to be seen whether mutations in RRAGC, APT6V1B2 or ATP6AP1 predict response to mTOR inhibition. 1.2.3.6 IGHV/IGLV mutations and unusual glycans Most FL cases express surface Ig, which is remarkable for two reasons. Firstly, one of the two IG alleles is involved in most cases in the t(14;18) and hence is not functional. Secondly, the BCR undergoes continuous SHM in FL that should lead over time to deleterious mutations, which is not observed. These observations therefore suggest positive selection for preserved BCR expression. The driver of selective pressure could be persistent antigen providing BCR signalling but, unlike other B cell malignancies, IGV gene usage is not restricted to a selected repertoire, but highly variable. A possible explanation for the requirement for continuing BCR expression comes from the finding that, in FL, motifs for N-glycosylation are introduced in IGHV genes through SHM, which is not commonly seen in normal B cells.171,172 Furthermore, V region sugars in FL are predominantly oligomannose glycans, suggesting that they are susceptible to bind lectins present in the tumour microenvironment.173 It was recently shown that, indeed, lectins from non-tumour cells, such as DC-SIGN expressed on macrophages or even lectins present on bacteria, 11 trigger BCR signalling, as evidenced by calcium flux assays and phosphorylation of signalling molecules downstream of the BCR (ERK, AKT, PLCγ2).174–176 1.2.4 Role of the microenvironment FL tumour cells colonize normal germinal centers177 and migrate between follicles,178 despite the preservation of a follicular growth pattern, suggesting that a normal lymph node structure is required for seeding and proliferation. FL typically grows in a rich microenvironment composed of varying proportions of T helper cells, cytolytic T cells, macrophages, antigen-presenting cells and stromal cells (Figure 2). It is virtually impossible to propagate indolent FL ex vivo,179 and feeder cells, cognate CD4 T cells or cytokines provide only short-term benefits in vitro by decreasing apoptosis and inducing proliferation.180–182 In addition, transcriptomic analyses have shown that gene expression signatures derived from normal, tumour infiltrating immune cells have prognostic significance.183–188 All of these considerations suggest that FL thrives on, and requires a subverted tumour microenvironment, a concept that has been named “re-education”.189 The role of the tumour microenvironment consists in supporting tumour growth and in suppressing an effective anti-tumour immune response. The following sections will delineate the known roles of individual cellular subsets from the tumour microenvironment.  1.2.4.1 Follicular T helper cells The importance of IL4-mediated signalling in FL is underscored by the fact that this cytokine is found at elevated levels in lysates from FL,167 that the IL4 receptor is over-expressed on tumour cells,184 and that IL4 enhances proliferation of FL cells in vitro.182 It was subsequently shown that FTH cells are an important source of IL4 in the FL tumour microenvironment, and that the proportion of CD4+CXCR5+ICOS+CD25- FTH cells is increased in FL lymph nodes when compared to normal lymph nodes (but not when compared to tonsils).166,190 Consequently, phosphorylated STAT6 can be detected in primary FL cells, but not in normal reactive lymph node or tonsil controls.166,167 In addition to IL4, FTH cells are a source of CD40 ligand that was shown to prevent apoptosis of FL tumour cells.191 In co-culture experiments, FTH cells decreased apoptosis of tumour cells;190 and IL4 and CD40L increased tumour cell viability after incubation with rituximab and human serum,190 whereas another study showed that CD40L promoted FL cell growth in vitro.192 12 1.2.4.2 T regulatory cells FL cells foster an exhausted immune milieu that is incapable of mounting an effective anti-tumour response. In this regard, tumour cells have been shown to promote the conversion of T cells into CD4+CD25+FOXP3+ Tregs in co-culture experiments, and these cells are consequently enriched in primary patient specimens.193,194 Tregs, and the more recently discovered subset of follicular Tregs have been shown to decrease proliferation and cytokine production of effector T cells,190,193 thereby contributing to suppression of the immune response. Moreover, the balance between Tregs and TH17 cells, thought to have opposing roles, is largely skewed in favour of Tregs as FL is characterized by decreased numbers of TH17 cells compared to non-malignant controls.195 It has been suggested that FL cells promote Treg conversion via secretion of CCL2193 and secretion or expression of TGFβ.196 TGFβ also upregulates CD70 on effector T cells and, with IL12, was shown to induce effector T cell exhaustion.197,198 1.2.4.3 Cytotoxic T cells Elevated numbers of CD8+ cytotoxic T cells (CTLs) have repeatedly been associated with favourable prognosis in FL.199–201 CTLs are mainly located in the interfollicular spaces and are in close contact with tumour cells at the follicle borders.202 The failure to enter follicles could potentially be attributed to reduced motility.188 Furthermore, close contact is not sufficient for tumour cell killing, and it was shown that autologous tumour B cells induce defective T-cell synapse formation, the mechanism of which is yet to be elucidated.203  1.2.4.4 Macrophages The proportion of tumour-associated macrophages (TAMs) is less than 5% of all cellular elements, on average, in FL.204 Yet, the extent of macrophage infiltration has been associated with outcome in many, albeit somewhat contradictory studies.205–211 Taken together, the prognostic role of TAMs appears to be dependent on treatment received, suggesting an interaction between macrophages and therapy. For example, in a randomized clinical trial in which patients were assigned to chemotherapy with and without rituximab, adverse prognosis was found only in the non-rituximab arm.207 This observation is potentially explained by the capacity of macrophages to bind antibodies such as rituximab via Fcγ receptors and in animal models, macrophages are required for effective, rituximab-mediated therapy.212 We showed that the correlation between TAMs and outcome was furthermore dependent on whether the patients received a chemotherapy backbone with or without doxorubicin,204 a further illustration of an interaction between treatment 13 and prognosis. Due to difficulties in purifying TAMs from primary patient specimens, our understanding of functional subtypes and signaling pathways within FL-associated TAMs is limited. One study showed that IL15 is over-expressed by TAMs and this likely contributes to FL growth in cooperation with CD40L expressed on FTH cells.192 1.2.4.5 Stromal cells Stromal cells are of non-hematopoietic origin and provide the scaffold within which the immune response takes place in secondary lymphoid organs, in addition to regulating lymph flow and immune cell responses. FDCs, for example, secrete the ligand CXCL13, thereby attracting CXCR5-positive B cells into follicles, as well as survival factors such as BAFF and APRIL.213 Moreover, stromal cells in the bone marrow support normal hematopoiesis.214 In FL, it was shown that a stromal cell-derived cell line (HS-5), as well as primary human bone marrow stromal cells protect primary follicular lymphoma cells from apoptosis in vitro and that this effect is at least in part mediated by BAFF.215,216 Similarly, follicular reticular cells derived from human tonsil were reported to support growth of tumour B cells.217 Bone marrow stromal cells from FL patients were found to have a distinct gene expression profile, when compared to healthy donors, to overexpress CCL2, to increase migration of monocytes and to contribute to polarization of monocytes into M2-type tumour-associated macrophages.218 1.3 Progression of follicular lymphoma The study of progression in FL is difficult, because it requires a strategy of sequential biopsies (i.e. diagnostic FL samples and samples collected at the time of progressed FL). Furthermore, initial diagnosis can occur at variable points in time given that a large fraction of patients are initially asymptomatic and are often diagnosed incidentally. This potential lead-time bias has an impact on determination of time to progression. Moreover, the criteria for determining progressive disease are not strictly delineated, as progression can be defined based on radiological or clinical grounds. Nonetheless, several studies have analyzed sequential specimens in order to gain insight into the process of progression. Whether the methods consisted of conventional cytogenetics, Sanger-based or next-generation sequencing of VDJ sequences, or whole-exome sequencing, the pattern of progression was most often consistent with divergent evolution, although linear progression has also been described.110,145,219–221 VDJ sequences in particular may be used as clonal trackers, and their sequential assessment was performed in several studies that utilized this information to describe tumour phylogenies and diversity.219,222–224 Taken together, 14 these studies are important because they substantiate underlying tumour heterogeneity, which ultimately may give rise to clonal outgrowth of resistant disease once selective pressure in the form of therapy is applied. However, unlike chronic lymphocytic leukemia (CLL),225,226 selection of subclones over the course of treatment has not been reported in FL. Interestingly, the degree of divergence between diagnosis and progression has consistently been shown not to be correlated with time to progression.110,219 A further important finding is the structural preservation of an intact BCR, despite ongoing SHM (see section 1.2.3.6). Indeed, selection against deleterious mutations has been demonstrated219 and suggests that BCR signalling may remain critical in PFL, potentially opening therapeutic avenues. In an elegant study, Irish et al. performed phospho-specific flow cytometry and described in diagnostic FL samples, the presence of cells that exhibited impaired BCR signalling after stimulation. A higher proportion of these cells was associated with adverse outcome and signalling-defective cells were hence named Lymphoma Negative Prognostic (LNP) cells. The proportion of LNP cells was shown to increase over progression. Importantly, cells that had impaired signalling did not have intrinsic signalling defects as in vitro phosphatase inhibition could reverse signalling impairment. This study is important because it questions whether BCR signalling is truly critical in late stages of FL, which is relevant as inhibitors of BCR signalling are in clinical development.227,228 1.4 Transformation of follicular lymphoma 1.4.1 Clinical aspects Transformation from indolent into aggressive lymphoma is a critical event in the history of patients as it has therapeutic and prognostic implications. Transformation is most well described in FL, but can occur in other indolent lymphomas/leukemias, such as chronic lymphocytic leukemia (where transformation is referred to as Richter’s syndrome), marginal zone lymphoma, lymphoplasmacytic lymphoma, indolent T-cell lymphomas and nodular lymphocyte predominant Hodgkin lymphoma.229 Transformation is best defined pathologically by sheets of large cells that efface the prior architecture of indolent lymphoma, and ideally requires proof of clonal relationship between the indolent and the aggressive histologies.230 Transformation is typically associated with onset of new symptoms, including B symptoms, elevated LDH or calcium, and rapid tumour growth that is often highly FDG-avid on PET-scanning.231 15 The incidence of transformation is reported to lie in the range of 2-3% per year.232–238 Certain clinicopathologic risk factors, when present at diagnosis of indolent lymphoma, have been shown to be associated with an increased risk of histological transformation. They include, for example, advanced stage, elevated LDH, high FLIPI and grade 3 histology.232,238,239 It is controversial whether upfront treatment has an impact on transformation, as compared to a watch and wait approach, as some studies suggest a reduced risk of transformation in patients who are initially observed234,236 and other studies do not.235,240 Similarly, it is unclear from the literature whether rituximab has decreased the incidence of transformation as studies have reported conflicting results.234–236 Whereas post-transformation survival was historically reported to be dismal,232,238,240,241 more recent data shows that in the modern era outcomes are substantially better, with median survival times of 50-60 months reported in the literature.234,235 This apparent improvement in survival post-transformation may in part be attributable to the introduction of rituximab, especially in those patients who were rituximab-naive prior to transformation. Treatment prior to transformation is a main factor contributing to survival post-transformation. Indeed, it was shown that treatment-naive patients do better than pre-treated patients,242 and that treatment with R-CHOP before transformation is associated with poor outcome.234 These observations suggest that prior treatment may select more resistant disease at time of transformation. 1.4.2 The genetic basis of transformation The study of the mechanistic underpinnings of transformation is arduous due to the difficulty in assembling cohorts of paired patient specimens (i.e. paired FL and transformed FL (TFL) for a given patient), which ultimately results in most studies being restricted by sample size. Before the advent of genome-wide discovery tools, alterations in selected, single genes were reported to be seen at higher frequency in TFL compared to FL, suggesting that they contribute to an aggressive phenotype. In this regard, TP53 was reported to be mutated in 26-80% of TFLs122,123,243,244 and also to be commonly deleted245 or affected by loss of heterozygosity (LOH).124 Translocations involving the MYC locus have been described in 8-46% of TFLs.109,110,246 Moreover, mutations, deletions and LOH affecting the CDKN2A locus have been reported in 33% of TFLs and are generally considered to be rare at diagnosis.123,124,247,248 In aggregate, these findings implicate an altered DNA damage response and cell cycle dysregulation as potential major contributors to the process of transformation. 16 The earliest tool to study genome-wide aberrations associated with transformation consisted in conventional cytogenetics, where TFL samples were found to harbour, on average, more alterations than FL.110 With regards to specific abnormalities, deletions of 1p36 and involvement of the 8q24 locus were more common in TFL than in FL.110 In a small series of 12 patients with paired samples, the mean number of chromosomal changes was also higher at transformation or relapse compared to FL, as assessed by array comparative genomic hybridization.249 In a larger series of 225 FL and 84 TFL specimens, of which 44 were paired, copy number analysis by single-nucleotide polymorphism arrays confirmed a higher incidence of losses of CDKN2A and TP53, as well as more frequent gains in the cell cycle regulators MDM2 and MDM4 in TFL when compared with FL.245 Genes involved in expression of MHC class I (B2M, HLA-B, HLA-C), were more commonly deleted in TFL than in FL, and transformed samples were also characterized by more frequent alterations in the NF-kB pathway (TNFAIP3, REL).245 At least 3 studies assessed gene expression changes by comparing TFL to FL samples.123,246,249 However, due to the small sample size of these studies, and the differences in tumour content between FL and TFL, it is challenging to draw firm conclusions. One study reported, using gene expression profiling, that all 18 classifiable TFL cases were of a GCB phenotype.123 To date, only two studies have applied next-generation sequencing technology to FL and TFL samples and were published in December 2013135 and January 2014,136 respectively. Okosun et al. performed whole-genome sequencing on 6 trios (normal, FL, TFL), exome-sequencing on 4 additional trios, and applied targeted sequencing of 28 genes to an extension cohort of 100 patients.135 The authors described an increase in the burden of mutations and copy number alterations from FL to TFL and found that, using phylogenetic trees, tumour evolution was consistent with branched rather than linear evolution. They further described that each FL and TFL pair shared common sets of mutations, suggesting that they were derived from a common progenitor clone (CPC), in keeping with the authors’ prior study on patterns of SHM in IGHV sequences.99 Shared mutations included frequent mutations in chromatin modifiers such as KMT2D, CREBBP, EP300 and MEF2B, suggesting that these are early events. Amplifications of EZH2, MDM2, MYC and REL, as well as mutations of EBF1, MYD88 and TNFAIP3 were more commonly seen in TFL than in FL. Pasqualucci et al. performed exome sequencing and high-resolution SNP arrays on a discovery cohort of 12 patients with paired FL and TFL available (with germline DNA available 17 for 4 patients).136 Similar to Okosun et al., the predominant pattern of evolution was divergent although no FL-specific alterations were found in 2 cases in which evolution seemed to follow a linear pattern. The number of copy number alterations were significantly increased in the transformed biopsy in all 12 patients. The authors then applied whole exome sequencing and SNP arrays to an extension set of 27 TFL samples. Recurrent genetic lesions that were shared between FL and TFL samples included mutations of KMT2D, EZH2, CREBBP, EP300 and FAS. Interestingly, FAS mutations, although found at diagnosis in matching FL samples, were not identified in a control cohort of 23 unselected FL cases, suggesting that they might predispose to transformation. CDKN2A and/or TP53 alterations were found in 46% of TFL samples and uncommon in FL. Chromosomal translocations of MYC were found in 25% of TFL patients, and B2M and CD58 were inactivated by bi-allelic alterations in 13% and 5% of TFL samples, respectively. The authors found significantly more mutations in genomic regions affected by SHM in TFL versus FL. The authors showed that the genomic landscape of TFL and de novo DLBCL differ, as bi-allelic deletions of CDKN2A were seen in TFL but never in de novo DLBCL, and mutations involving STAT6, ARID1A, FAS, KMT2D and CREBBP were similarly more frequent in TFL. 1.5 Outcome prediction in follicular lymphoma 1.5.1 Pathology Grading of FL is defined based on the number of centroblasts per high-power field and the 2008 WHO classification divides FL into grades 1, 2 and 3.250 Grade 3 is further subdivided into grade 3A, considered to be part of the spectrum of indolent FL, whereas grade 3B FL is usually considered to represent an aggressive neoplasm more akin to de novo DLBCL and is treated as such. Grading of FL is relevant because grade 3A FL is often associated with a distinct biological subset of FL, frequently CD10-negative, and commonly harbouring BCL6 rearrangements but not the prototypical t(14;18) translocation.251–253 Follicular large cell lymphoma, the historical equivalent to grade 3 FL based on the Working Formulation, was considered by some authors to be associated with more aggressive disease behaviour and to warrant anthracycline-containing chemotherapy,254–256 but this is controversial.257,258 Similarly, the presence of diffuse areas has been associated with poorer outcome in some studies, but interpretation of these data is difficult as many of these cases likely represent composite histologies at initial diagnosis, comprising both 18 FL and DLBCL.259,260 The poor reproducibility of grading and differences in patient and treatment characteristics likely contribute to discrepancies between studies. 1.5.2 Genetics A small fraction of FLs (15%) are negative for the t(14;18) and most of these cases do not express BCL2,253,261,262 but the absence of this translocation is not associated with outcome.110,239,253,262,263 FL is characterized by frequent additional chromosomal aberrations and the total number of alterations, as assessed by conventional karyotyping, has been correlated with adverse outcome in some,263 but not all studies.110 Certain recurrent cytogenetic findings, as assessed by conventional cytogenetics or CGH, have been found to be associated with poor survival and include deletion of 1p36,111 6q,83,111–113,220,263 trisomy 7,83 trisomy 12,83,114 deletion of 17p114,220,263 or duplication of chromosome X.110 Similar, alterations involving single genes have been reported to correlate with adverse outcome. They include TP53 mutations,264 deletion or methylation of CDKN2A265 and BCL2 mutations.266 Mutations of TNFRSF14 were reported to correlate with poor prognosis in one study,117 but this observation could not be confirmed in another study.118 However, all these studies were comparatively small in size and many did not include independent validations using additional cohorts. Furthermore, they generally failed to draw from the combined prognostic value of multiple genetic alterations, and relied on the analysis of aberrations in single genes to predict outcome. In collaboration with investigators at the University of Munich and the Dana Farber Cancer Institute, we conducted a study in which we analyzed the coding sequence of a panel of 74 genes after hybridization-based capture from DNA derived from formalin-fixed, paraffin-embedded tissue (FFPET).128 The study cohort was composed of a training set including 151 FL specimens that were collected as part of a clinical trial,77 and a validation set of 107 patients that were assembled retrospectively from the BC Cancer Agency. We trained a penalized, multivariate Cox regression predictor model that combined clinical variables (FLIPI and performance status), as well as the mutational status of seven genes (EZH2, ARID1A, MEF2B, EP300, FOXO1, CREBBP and CARD11). The subsequent model, termed the m7-FLIPI, model validated in the BCCA extension cohort, separating good and high-risk patients better than clinical variables alone. The m7-FLIPI reclassified 50% of FLIPI high-risk patients into an m7-FLIPI low-risk category and, hence, outperformed the FLIPI with respect to identification of true poor-risk patients. 19 Unexpectedly, EZH2 mutations were associated with favourable survival and with a specific gene expression signature that was similarly associated with good outcome. 1.6 Thesis theme and objectives As outlined in section 1.2.1, outcomes for FL patients are generally favourable, with median survival times largely surpassing 10 years, and perhaps even 18 years in the modern era.267,268 However, there are two partially overlapping scenarios that are associated with early death: transformation232,234,235,240,241 and early progression after immuno-chemotherapy.269,270 Although our understanding of the genetic underpinnings of progression and transformation has improved, as detailed in sections 1.3 and 1.4.2, the landscape of gene mutations that predisposes, and contributes to resistant and transformed disease has not been fully elucidated. In order to improve patient outcomes further, we need to uncover genetic and non-genetic drivers of transformation and progression and understand FL clonal trajectories in both time and space. Ultimately, the biological mechanisms that cause early adverse outcome need to be elucidated in order to design rational, targeted therapies. Intra-tumoural heterogeneity is a fundamental property of cancer and has been documented in FL by the analysis of cytogenetic aberrations,271 conventional99,272,273 or next-generation sequencing of VDJ sequences,219,224,223 targeted sequencing of single genes156 or broad and unbiased, whole-exome145 or whole genome sequencing.135 The clonal architecture of cancer is seen as a phylogenetic tree that expands along spatial and temporal axes.274 Underlying heterogeneity between individual cells provides the substrates for selection of those cells that are resistant once selective pressure in the form of treatment is applied.275,276 Cancer evolution bears many similarities with Darwinian evolution, first described in speciation. The complete assessment of clonal complexity is crucial for the treatment of cancer as subclonal mutations in resistance-conferring drivers have been shown to be associated with treatment resistance and adverse outcome in several tumour types (but not yet in FL).277 Subclonal mutations are, however, frequently missed because they are either not sampled due to spatial heterogeneity and/or because they may be present at levels that lie below the detection threshold.277 The objectives of this thesis are several-fold. First, we aim at fully characterizing the clonal structure of primary FL specimens and describe how this structure dynamically changes over time and over the course of transformation and/or progression. The second major objective is the correlation of specific gene mutations with increased risk of early progression. The last 20 objective is to enhance our understanding of transformed lymphoma by identifying molecular subtypes and describing recurrent gene alterations at time of transformation. 1.7 Hypotheses 1.7.1 Hypothesis 1 Molecular subtypes can be recognized within TFL by low resolution gene expression profiling and these subtypes differ by the respective prevalence within these subtypes of single nucleotide mutations as well as chromosomal translocations. 1.7.2 Hypothesis 2 Transformation and progression are the result of selection of pre-existing subclones that harbour driver alterations, some of which are yet to be discovered. 1.7.3 Hypothesis 3 Early transformation and/or progression can be predicted at time of diagnosis. 1.8 Aims and thesis outline The thesis consists of an introduction, two chapters outlining original research and a discussion. 1.8.1 Aim 1: Cell-of-origin of transformed follicular lymphoma Chapter 2 describes the application of the Lymph2Cx classifier to a large cohort of TFL samples and the relationship of molecular subtypes with translocation status of BCL2 and the immunophenotype of the original FL. 1.8.2 Aim 2: Evolutionary dynamics in follicular lymphoma Chapter 3 describes the deciphering of the clonal architecture of a core set of sequential FL-TFL and FL-PFL cases, as well as a set of long-term non-progressers. Clonal architecture is inferred by the integration of deep sequencing counts of tumour and wild-type alleles with copy number status and tumour content. The presence of minor subclones is further validated by digital droplet PCR. Targeted re-sequencing of 86 genes will be applied to a cohort of FL and TFL samples in order to identify novel associations of gene mutations with transformation, and in order to determine the frequency of specific gene alterations in this setting. 1.8.3 Aim 3: Prediction of early transformation/progression. In chapter 3, we aim at assessing the expression of routinely available biomarkers (BCL2, BCL6 and MYC translocation status and IRF4 and TP53 protein expression) with early transformation. In chapter 3, we apply a sequencing panel of 86 genes to a cohort of clinical 21 outcome extremes to determine associations of gene mutations with either early (< 2.5 years) progression after R-CVP and R maintenance, or with late or never (no progression for at least 5 years) progression after R-CVP and R maintenance or observation. 22 Chapter 2: Cell-of-origin of transformed follicular lymphoma  2.1 Introduction Follicular lymphoma (FL) is the most common indolent lymphoma with median survival for newly diagnosed patients currently in excess of 10 years.59,268,278 However, outcomes are heterogeneous and a fraction of patients are at risk of early relapse/progression. Histological transformation to aggressive lymphoma occurs in 2-3% of patients per year and is associated with treatment resistance, progression and lymphoma-related mortality.232,234,240,241 Transformation results most commonly in a histology that cannot be distinguished from de novo diffuse large B-cell lymphoma (DLBCL). More rarely the transformed biopsy contains a mixture in varying proportions of FL and DLBCL in the same biopsy (composite histology), resembles grey-zone lymphoma (a situation referred to as unclassifiable B cell lymphoma with features intermediate between diffuse large B-cell lymphoma and Burkitt lymphoma, abbreviated BCLU for the remainder of this manuscript).230 Whereas transformation was historically considered to be an event associated with dismal prognosis, treatment outcomes vary and prediction of survival after transformation needs to be individualized as outcome is dependent on factors such as the extent of the disease at time of transformation, prior therapies and time to transformation.232,234,242 It remains challenging to predict subsequent transformation at the time of FL diagnosis although certain clinical parameters, such as elevated lactate dehydrogenase levels, advanced stage and high Follicular Lymphoma International Prognostic Index (FLIPI) scores are correlated with a higher risk of transformation in some, but not all studies.232,234,238,240 Similarly, certain genomic features have been associated with an increased likelihood of transformation. These include BCL6 translocations,279 FAS mutations136,280 or deletions of 1p36.3.111 Taken together, these parameters, if acquired at initial diagnosis, do not predict transformation with sufficient sensitivity or specificity to predicate altering patient management. Transformation occurs by clonal evolution from a common mutated precursor, as shown by sequencing of rearranged VDJ sequences99 or whole-exomes or genomes.135,136 Recurrent genetic alterations that evolve during the process of transformation include: CDKN2A deletions and/or loss of heterozygosity (LOH);136,281 translocations, gains, amplifications or mutations involving MYC;109,136,246 mutations, deletions and LOH affecting TP53;122–124,136,243,244 and mutations or deletions of B2M and CD58.136 Increased proliferation resulting from cell cycle 23 deregulation, defective DNA damage response and escape from immune surveillance therefore emerge as critical elements of histological progression of FL to an aggressive lymphoma. Cell-of-origin (COO) assignment is increasingly important to predict prognosis and response to targeted therapy in de novo DLBCL, yet information on COO of transformed FL is scarce in the literature. In a prior study, Davies et al. performed subtyping of 25 transformed FL specimens using the Lymphochip platform and the previously reported Bayesian class predictor.123,282 All classifiable cases (18 out of 25) were of germinal centre B-cell-like (GCB) phenotype and none were of the activated B-cell-like (ABC) phenotype, although three cases out of 35 (9%) were assigned to the non-GCB phenotype using immunohistochemistry in an independent cohort.123 In the study reported by Bouska et al, 42 out of 59 (71%) transformed follicular lymphoma (TFL) cases were reported to be of the GCB phenotype and 17 out of 59 cases (29%) of a non-GCB phenotype.245 In addition, whereas assessment of translocations involving both the MYC and either the BCL2 or BCL6 genes ("double-hit lymphoma") allows for identification of a poor-risk category of de novo GCB-type DLBCL,283–285 similar prognostic implications are less well studied in transformed FL. In this study, we sought to assess whether transformation can be predicted at the time of FL diagnosis using clinical annotation and information on histological grade, fluorescence in situ hybridization (FISH) break-apart assays for recurrently rearranged genes (BCL2 and BCL6) and immunohistochemistry (IHC) for CD10, BCL6 and IRF4 (MUM1). Furthermore, as subtype-specific efficacy of novel agents is actively pursued in de novo DLBCL,286–288 we also asked whether TFL could be similarly divided into distinct transcriptional phenotypes, using a large cohort of formalin-fixed, paraffin-embedded tissue (FFPET) samples. Finally, we sought to determine whether features present at the time of initial diagnosis correlate with molecular classes of TFL, potentially delineating distinct paths to transformation. 2.2 Methods 2.2.1 Patient samples and definition of endpoints For prediction of transformation, we assembled a cohort of 126 patients who were diagnosed with FL (grade 1, 2 or 3A) between 1997 and 2013 and were either observed or treated frontline with rituximab alone or in combination with chemotherapy. This cohort was enriched for clinical extremes, consisting of 40 patients who experienced transformation within 5 years of diagnosis and 86 patients who experienced neither transformation, nor progression or death for at 24 least 5 years from the time of initial diagnosis. For all other analyses, we assembled a cohort of 155 patients who were diagnosed with FL (grade 1, 2 or 3A) between 1970 and 2013 and subsequently experienced transformation to large-cell lymphoma, as defined in the introduction. Cases with discordant histology at diagnosis were excluded but cases with composite histology at the time of transformation were included. In this cohort, samples were available for both the FL and the TFL time points for 114 patients. In a further 14 and 27 patients, samples were available only at the time of FL or TFL, respectively. The Lymph2Cx assay was restricted to those 110 TFL samples in which the histology was consistent with DLBCL. The overlap between the study cohorts is shown in Figure 3. All patients described within this study as having experienced transformation had histological proof of transformation. Time to transformation was defined as the time between diagnosis of FL and the time when histological proof of large cell transformation was obtained. Survival after transformation was defined as the time between transformation and death from any cause, with censoring at the date of last follow-up in living patients. All patient specimens were collected as part of a research project approved by the University of British Columbia-British Columbia Cancer Agency Research Ethics Board. 2.2.2 Tissue microarray, FISH and IHC A tissue microarray was constructed utilizing duplicate 1.0 mm cores of FFPET biopsies from all aforementioned patients. FISH assays were performed using commercially available break-apart probes for the BCL2, BCL6 and MYC loci. Images were acquired using a Metafer CoolCube 1 camera and Metafer software (version 3.11.3). At least 100 nuclei for each sample were assessed for break-apart status by two independent scorers (RK and SBN). Discrepant cases were further reviewed by a third person (AM) to reach consensus. At least 10% of nuclei were required to have a break-apart signal for a case to be called positive. Double-hit lymphoma was defined as a case having translocations involving the MYC locus and either the BCL2 or the BCL6 loci. Four micrometer sections of the tissue microarray were stained with antibodies against human CD10, BCL6, IRF4 and TP53 on a Ventana BenchMark XT automated slide staining system. The number of positive tumor cells was independently assessed by two expert hematopathologists (AM, PF) and discrepant results were discussed at a multi-head microscope. For CD10, BCL6 and IRF4, at least 30% of tumor cells needed to be stained for a case to be called positive. For TP53, 25 only cases with uniform, strong staining were called positive. The Hans algorithm was used for cell-of-origin assignment by IHC.289  2.2.3 Lymph2Cx assay RNA was extracted from 10µm FFPET sections, cut to a surface of ≥ 1 cm2, using the AllPrep DNA/RNA FFPE kit (Qiagen). Two hundred nanograms of FFPET-derived RNA were used for the Lymph2Cx assay, a digital gene expression (NanoString)-based test that assesses the expression of 20 genes (8 genes that are overexpressed in ABC-DLBCL, 7 genes that are overexpressed in GCB-DLBCL and 5 housekeeping genes), as described by Scott et al.290 For each case, the linear predictor score (LPS) was calculated and the likelihood of that sample belonging to either COO subgroup was determined using Bayes' rule, as described in Wright et al.282 Calibration between the lot of NanoString reagents used in this study and the lot used in Scott et al290 was achieved by determining the LPS for 30 RNA samples from that study on the current reagent lot. 2.2.4 Targeted sequencing of CARD11, CD79B and MYD88 Adapter ligated libraries were constructed following the 100ng SureSelect XT2 Custom Target Enrichment System for Illumina Multiplexed Sequencing Protocol Version D.2. Indexed libraries were pooled and sequenced on an Illumina MiSeq instrument generating 150 base-pair paired-end reads. After demultiplexing, reads were aligned to the human reference genome (hg19) using BWA-mem (version 0.7.5a)291 and single nucleotide variants were predicted using VarScan (version 2.3.6).292,293 All predictions with ≥ 5 variant-supporting reads and ≥ 10% variant allele frequency were taken forward for validation by Sanger sequencing, performed as previously described.294 2.2.5 Statistical analysis All associations between translocation status, IHC staining result and pathological or clinical characteristics were assessed using Fisher's exact test. For multi-variable prediction of transformation, we fitted a weighted Cox regression model, after verification of the proportional hazards assumption, as the patient cohort was enriched for transformation events.295 Assuming a transformation rate of 2.5%/year in an unselected patient cohort, the weight for events was adjusted from 1 to 0.4 and the weight of controls from 1 to 1.3. To estimate survival after transformation, we plotted survival curves according to the Kaplan-Meier method and we compared survival 26 differences between patient groups using the log-rank test. P-values less than 0.05 were considered significant. 2.3 Results 2.3.1 Association of clinical/pathological characteristics with transformation The cohort used for this analysis consisted of 40 patients who experienced transformation within 5 years after being diagnosed, and a control group of 86 patients who experienced neither transformation nor progression for at least 5 years after diagnosis (Table 2). The two groups were balanced with respect to clinical characteristics, except that the proportion of patients with elevated LDH at diagnosis was higher in the transformation group than in the control group (23% vs 5%, P = 0.01). However, the analysis of pathological features revealed multiple significant differences between the two groups (Figure 4). In comparison with the control group, cases in the transformation group were more likely to have grade 3A FL (40% vs 9%, P < 0.001), to harbour a BCL6 translocation (25% vs 10%, P = 0.05), to be negative for CD10 staining (81% vs 97%, P = 0.008) and to express IRF4 (19% vs 1%, P < 0.001). The proportion of cases harbouring a BCL2 translocation was lower in the transformed cases than in the control cases but this difference did not reach statistical significance (80% vs 92%, P = 0.11). There was no difference in BCL6 staining between the two groups (97% vs 98%, P = 1.00). As some of the pathological features under study tend to co-occur, we fitted a weighted, multivariate Cox proportional hazard model that included grade 3A, BCL2 and BCL6 translocation, as well as staining for CD10 and IRF4 (Table 3). Whereas in univariate Cox regression analyses, grade 3A, BCL6 translocation and IRF4 staining were positively associated, and CD10 staining negatively associated, with time to transformation, in the multivariate model only IRF4 staining predicted a higher risk of transformation when positive at diagnosis (hazard ratio 13.3, 95% confidence interval 3.7-48.4, P < 0.001). 2.3.2 Temporal patterns in translocation status and IHC staining We leveraged a cohort of 128 FL and 141 TFL samples that included 114 patients for which samples from both timepoints were available, to assess the presence of BCL2, BCL6 and MYC translocation, as well as staining by IHC for CD10, BCL6 and IRF4. Patient characteristics are shown in Table 4. In this cohort, in which all patients experienced transformation during the course of their disease, BCL2 and BCL6 translocations were found in 80% and 22% of FL samples, respectively, and these percentages were not significantly different in TFL samples. In contrast, 27 MYC translocation was detected in 9% of FL samples but the percentage was significantly higher in TFL samples (26%, P = 0.001). Consequently, double-hit status was also significantly more common in TFL than in FL (24% vs 7%, P = 0.001, Figure 5B). TFL samples expressed CD10 less commonly than FL (80% vs 90%, P = 0.04), as well as BCL6 (87% vs 97%, P = 0.006), but expressed IRF4 (27% vs 11%, P = 0.001) and TP53 (29% vs 2%, P < 0.001, Figure 6A) significantly more often. IRF4 and TP53 expression were not correlated (data not shown) and were found at similar frequencies in all histological subtypes of TFL (Figure 6B). 2.3.3 Survival after transformation by histology and double-hit status Next, we sought to determine outcome after transformation, based on TFL morphology and double-hit status. Cases with a composite histology at the time of transformation had the best survival, while cases with BCLU morphology had the worst outcome (2-year survival from time of transformation 86%, 59% and 33% for cases with composite, DLBCL and BCLU histology, respectively, log-rank test, P = 0.03, Figure 7A). Double-hit status was associated with an inferior survival at 2 years after transformation as compared to cases without double-hit, in the entire cohort (50% vs 73%) and in only those cases with DLBCL morphology (51% vs 70%), but these differences were not statistically significant (Figure 7B and Figure 7C). 2.3.4 COO assignment of transformed lymphoma We applied the Lymph2Cx assay to 110 TFL samples that had morphology features of DLBCL. Sufficient gene counts to assign COO were obtained in 107 out of 110 samples (97%). Out of these 107 TFL samples, 86 were of the GCB subtype (80%), whereas 17 were of the ABC subtype (16%) and 4 were unclassified (4%) (Figure 8A). We also determined COO using the Hans classifier and, excluding one case for which IHC failed and the 4 cases that were unclassifiable by gene expression profiling, the concordance rate between the Lymph2Cx assay and the Hans algorithm was 92%. Using the Lymph2Cx assay, the proportion of GCB cases was significantly higher (80% vs 56%), and the proportion of ABC cases conversely lower (16% vs 32%, P < 0.001) in TFL than in a cohort of 335 de novo DLBCL that we have previously reported (Figure 8B).296 As the transcriptomic and genomic landscape of TFL is distinct from de novo DLBCL, we compared the distribution of the linear predictor score between molecular subtypes of TFL (n=103) and de novo DLBCL (n=30). These distributions were superimposable using low-density gene expression data from the Lymph2Cx assay, suggesting that the assignment of TFL into GCB and ABC classes identifies transcriptional subtypes as distinct from each other as in de 28 novo DLBCL. Furthermore, as most prior literature described TFL as resembling GCB-DLBCL,123,136 the finding that 16% of TFL cases were of the ABC subtype raised the question of whether these ABC cases had truly transformed from underlying FL or had arisen de novo. We therefore determined clonotypes of the rearranged variable region of the immunoglobulin heavy chain locus in 3 cases, in which high quality DNA was available, through PCR amplification and sequencing as per the Biomed-2 protocol.297 A clonal relationship could indeed be documented in all 3 cases (data not shown). 2.3.5 Correlation of COO with outcome and cytogenetics To further evaluate whether GCB-TFL and ABC-TFL represented distinct entities, we assessed translocation status of BCL2, BCL6 and MYC within these subtypes. BCL2 translocation was found in 68 out of 76 GCB-TFL cases (89%) but only in 5 out of 14 ABC-TFL cases (36%, P < 0.001, Figure 9A). BCL6 and MYC translocations were also more commonly present in GCB-TFL than in ABC-TFL, but this enrichment did not reach statistical significance. As molecular subtyping in de novo DLBCL has important prognostic implications,49 we compared survival after transformation between GCB-TFL and ABC-TFL, but could not detect any significant survival difference in TFL (Figure 9B). The finding that ABC-TFL cases were commonly negative for the BCL2 translocation raised the question of whether pathological findings within FL samples correlated with molecular subtypes in subsequent TFL (Figure 10). Indeed, FL cases that were graded 3A or expressed IRF4 preferentially transformed into the ABC subtype of TFL. On the other hand, presence of a BCL2 translocation or expression of CD10 were associated with subsequent transformation into GCB-type TFL. Taken together, these findings highlight that the underlying genetic and phenotypic heterogeneity in FL translates into distinct TFL subtypes. 2.3.6 Mutations in CARD11, CD79B and MYD88 The finding that a subset of TFL is of the ABC subtype raised the question whether ABC-TFL is characterized, similarly to de novo ABC-DLBCL, by gene mutations that activate or amplify B-cell receptor signalling and/or NF-κB signalling. To that effect, we performed next-generation sequencing of bait-captured coding sequences of CARD11, CD79B and MYD88 in all 17 ABC-TFLs. In order to maximize discovery, we included all cases in the analysis, regardless of coverage. We identified 9 mutations in 5 patients (Table 5), suggesting that genetic alterations involving the B-cell receptor and/or NF-κB signalling pathways are common in ABC-TFL. For two of these patients (FL1005 and FL1161), germline DNA was available and all 4 variants present 29 in these two cases were confirmed to be somatic. Mutations affecting the described amino acid residues of CARD11 and CD79B (Table 5) had been previously described in ABC-DLCBL.298,299 For MYD88, we identified a total of 4 mutations of which one corresponded to the classical hotspot mutation situated in the Toll/IL1-R domain (L265P), whereas the other 3 mutations (P141S, V144L and S149I) were located in the Intermediary Domain of MYD88. Mutations affecting two of the latter amino acid residues (P141 and S149) have been previously reported in de novo DLBCL.300,301 2.4 Discussion In this study, we associated routinely available biomarkers with risk of transformation, assessed temporal genetic and phenotypic changes in sequential specimens, performed molecular subtyping of TFL and correlated transformation into either COO subtype with pathological features present in the initial FL sample. Whereas multiple features, including grade 3A, BCL6 translocation, IRF4 expression as well as the lack of CD10 staining were associated with early transformation, only IRF4 positivity maintained independent predictive value in a weighted, multivariate Cox regression model. IRF4 is a transcription factor of the interferon regulatory factor family that exerts critical functions in germinal centre formation and plasma cell differentiation.302–304 It can be detected in 14-17% of FL cases305,306 by IHC and expression was shown to correlate with poor overall survival in two phase II studies, but not in a larger trial from the Southwest Oncology Group.306 IRF4 expression has also been associated with poor progression-free survival in the FL2000 and PRIMA trials of the LYmphoma Study Association.307 The expression of IRF4 has been linked to higher proliferation rates, as assessed by Ki-67 staining, and to FL grades 3A and 3B. Our study adds texture to the literature by documenting an increased risk of early transformation in IRF4-expressing FL. Although the absolute number of IRF4-positive FL is small in our study, the reproducible association of IRF4 expression with poor outcome suggests that these patients might benefit from alternative treatment approaches. In this regard, lenalidomide is an enticing candidate as it was shown to down-regulate IRF4 in the ABC subtype of DLBCL.308 Once transformation is suspected on clinical grounds, a repeat biopsy provides histological documentation required for management of TFL, in addition to important correlative information. Our data illustrate that the outcome of patients with a composite histology at time of transformation is better than the outcome of patients with a morphology that is in keeping with DLBCL or BCLU, in line with a prior study.309 Double-hit lymphoma usually refers to de novo 30 DLBCL with concurrent MYC and BCL2 (or more rarely BCL6) translocations and has been associated with poor outcome following chemotherapy alone or when combined with rituximab in most studies.283–285,310 The prognostic implication of double-hit status at time of transformation from underlying FL is not as clearly established. In our series, double-hit status was found in 24% of all TFLs, and in 19%, 13% and 100% of TFLs with a DLBCL, composite or BCLU morphology, respectively. We observed a trend towards inferior survival post-transformation in double-hit TFL that did not reach statistical significance, possibly due to the relatively small sample size or heterogeneity in pre- and post-transformation treatments. Future studies are therefore needed to refine our understanding of the prognostic implication of double-hit status by itself and in the context of genome-wide alterations in TFL. De novo DLBCL can be dissected into COO subtypes that differ by gene-expression profiles, activation of differential pathways, patient outcomes and response to therapy.48–52 As TFL with a DLBCL morphology cannot be readily distinguished from de novo DLBCL based on microscopy, it is relevant to ask whether TFL can be similarly divided into molecular subtypes. Historically, TFL was considered to be related to the GCB subtype of de novo DLBCL, based on its transcriptional profile and the landscape of genetic alterations.123,136 In keeping with the report by Bouska et al,245 our study shows that a majority (80%) of TFL cases are of the GCB phenotype, but that a significant minority (16%) can nonetheless be identified as being of the ABC phenotype using the Lymph2Cx assay. In our series, ABC TFL cases did not have a poorer outcome after transformation than GCB TFL cases, although a definitive conclusion cannot be drawn due to a relatively small number of ABC cases and treatment heterogeneity. Importantly, we show that ABC TFLs have a low prevalence of BCL2 translocations and that the absence of a BCL2 translocation at the FL stage is associated with transformation into the ABC subtype of TFL. This observation mirrors the finding that ABC-related genes are enriched in the gene expression profile of BCL2 translocation-negative FL.53 In addition, IRF4 expression is a key contributor to the ABC subtype and it is therefore not surprising that IRF4 expression in the antecedent FL similarly increases the likelihood for transformation into this subtype. In FL, negative staining for CD10, positive staining for IRF4 and lack of BCL2 translocation tend to co-occur, as shown by Karube et al.,251 delineating cases with distinct molecular characteristics from the more common BCL2 translocation-positive FL cases. Based on our study, we can conclude that such molecular heterogeneity in FL is linked to molecular heterogeneity at the time of transformation and that the 31 lymphoma phenotype at transformation is largely determined by underlying characteristics in the preceding FL. However, the relationship between findings in FL and at subsequent transformation is imperfect as, for example, 6 out of 13 BCL2 translocation-negative FLs transformed into the GCB subtype in our study, testifying to the complexity of pathways to transformation. In conclusion, our study shows that IRF4 expression in diagnostic FL samples is associated with early transformation. With regard to temporal changes over the course of transformation, MYC is more commonly translocated, and IRF4 more commonly expressed, in TFL than in antecedent FL. Furthermore we show that composite histology at the time of transformation is associated with better outcome than DLBCL or BCLU morphologies and that 80% and 16% of DLBCL-like TFLs are of the GCB or the ABC subtype, respectively. These latter findings are important as they document inter-patient heterogeneity in TFL and suggest that treatment approaches may need to be tailored to underlying biology in order to positively impact the adverse outcome that is currently associated with transformation. The finding that a subset of TFL is of the ABC subtype is particularly intriguing as it raises the question of whether these patients would be candidates for targeting of the B-cell receptor signalling pathway. In this regard, our finding that mutations in CD79B and MYD88 are recurrent in ABC-TFL provides further rationale to consider investigating the clinical benefit of targeted agents such as ibrutinib and lenalidomide in ABC-TFL. Future studies will, however, need to complement our results by adding whole transcriptome and broader mutational context in order to comprehensively elucidate the genomic makeup of TFL and provide insight into additional druggable alterations. 32 Chapter 3: Evolutionary dynamics of follicular lymphoma  3.1 Introduction Follicular lymphoma (FL) is the 2nd most common non-Hodgkin lymphoma subtype and the most frequent indolent lymphoma, accounting for 22-32% of all new lymphoma diagnoses in Western countries.58,59 Patient outcomes are favourable, with median overall survival times now well beyond 10 years.267,268,278 Nonetheless, FL is an incurable malignancy and most patients eventually experience progressive disease. Moreover, a subset of patients is at risk of early lymphoma-related mortality. In this regard, progression within two years after immuno-chemotherapy was recently reported to be associated with poor survival.269 In addition, transformation to aggressive lymphoma occurs in 2-3% of patients per year and is similarly correlated with poor outcome, especially in pre-treated patients.232,234,235,240,242 Transformation is typically associated with rapid nodal as well as extranodal tumour growth, and sudden onset of new symptoms, suggesting that it is the result of an evolutionary shift from the underlying indolent lymphoma. Despite the consideration that early progression and transformation correlate with an increased risk of death, their genetic basis is imperfectly understood. Alterations in single genes have been reported to be associated with adverse outcome. They include mutations in TP53,264 deletion and or methylation of CDKN2A265 and TNFRSF14 mutations/1p36 deletions.117 More recently, a clinicogenetic risk model (m7-FLIPI), including mutational status of seven genes, the Follicular Lymphoma International Prognostic Index (FLIPI) and performance status was shown to improve outcome prediction for patients requiring immuno-chemotherapy.128 Transformation to aggressive lymphoma has been shown to result from clonal evolution of a common precursor by sequencing of rearranged and hypermutated IGHV sequences,99 as as well more comprehensively by sequencing of exomes and/or genomes.135,136 Recurrent transformation-associated genetic alterations include deletions and/or loss of heterozygosity (LOH) of CDKN2A;136,281 translocations as well as gains, amplifications and mutations of MYC;109,136,246 deletions, mutations and LOH of TP53123,124,243,244 and deletions and mutations involving CD58 and B2M.136 The clonal evolution theory of tumour progression has been proposed for decades274 and is intimately linked to intra-tumoural heterogeneity. Indeed, differential acquisition of somatic 33 mutations may lead to subclones that have selective advantage once therapy is applied. As FL is a disease in which patients almost invariably present with disease progression and/or transformation, it represents a compelling opportunity to study clonal dynamics over time. Prior studies have applied next-generation sequencing technology to serial lymphoma specimens,126,135,136,145 but the evolutionary dynamics, the landscape of driver mutations and the role of intrinsic clonal heterogeneity underlying transformation and treatment resistance remain to be discovered. Furthermore, as transformation and treatment resistance are associated with increased risk of death, mechanistic insight into these phenomena emerges as an unmet need. 3.2 Methods 3.2.1 Patients and materials The cohorts under study are shown in Figure 11. The whole genome sequencing (WGS) cohort consists of 41 patients that were selected to fall into three groups: 1) a "Transformed" group of 15 patients (TFL) diagnosed with FL and subsequent or concomitant (patient FL1014) transformation to large cell lymphoma; these cases were selected irrespective of type of treatment received; 2) a "Progressed" group of 6 patients (PFL) that were diagnosed with FL and subsequently experienced progressive disease without histological evidence of transformation; 5 out of these 6 patients progressed within 2.5 years after starting first-line immuno-chemotherapy with R-CVP (rituximab, cyclophosphamide, vincristine and prednisone) and 3) a "Good outcome" group of 20 patients (NPFL) who were diagnosed with FL and did not experience progression for at least 5 years. All cases were selected irrespective of clinical stage, grade and t(14;18) translocation status in order for the cohort to be reflective of the clinical and pathological heterogeneity that is inherent in FL. Samples with a tumour content of less than 50% and available frozen single cell suspensions were flow-sorted to purify tumour (CD19+ kappa or lambda+ CD3-) and germline cells (CD19- kappa or lambda- CD3+). Germline DNA was obtained from flow-sorted CD3+ lymphocytes or from peripheral blood cells. All germline samples were confirmed to be free of tumour contamination by the absence of PCR-amplifiable patient-specific t(14;18) and/or VDJ rearrangements. The capture sequencing cohort refers to the samples from 205 patients (38 patients overlapping with the WGS cohort), in which germline DNA from peripheral blood was available for 73 patients. These patients were divided into three groups: 1) a "Transformed" group of 116 patients (sample at primary FL timepoint available in 80 cases, sample at transformed FL timepoint 34 available in 83 cases and samples from both timepoints available in 47 cases); 2) a "Progressed" group of 30 FL samples from patients who presented with early progressive disease within 2.5 years after starting immuno-chemotherapy with R-CVP (progression was defined as radiological evidence of progressive disease and requirement for initiation of second-line therapy) and 3) a "Good outcome" group of 67 patients without progression for at least 5 years after either observation or R-CVP. For both cohorts, DNA was extracted from frozen tissue or single cell suspensions using Qiagen DNA/RNA AllPrep kits. Henceforth, timepoint specificity will be indicated for each sample by the suffixes -T1 for the primary timepoint (by definition FL) or -T2 for the secondary timepoint in the TFL or the PFL cohorts (transformed or treatment resistant FL). Patient specimens were collected as part of research projects approved by the research ethics boards of the University of British Columbia-British Columbia Cancer Agency (H13-01765), the UZ Leuven (S-55498) or the Mayo Clinic (08-005005). 3.2.2 Pathology All biopsies were centrally reviewed by expert hematopathologists at the BC Cancer Agency, Vancouver BC. FL grades 1, 2 and 3A were included and TFL biopsies were of three histologies: DLBCL, composite or unclassifiable B-cell lymphoma with features intermediate between DLBCL and Burkitt lymphoma (abbreviated as BCLU). Composite histology was defined as any evidence of underlying low grade lymphoma in a sample that concomitantly harboured large cell lymphoma. The Lymph2Cx assay was performed as described in Chapter 2, with the exception that it was applied in 4 cases to RNA extracted from fresh-frozen blocks, using 100 ng as input. Tissue microarray slides with duplicate tissue cores were immunohistochemically stained for the T cell marker CD8 (antibody clone C8/144B, Dako) and scanned using an Aperio ScanScope XT at 20x magnification. Analysis was performed using the Aperio ImageScope viewer (Version 12.1.0; Aperio Technologies). Only cores and areas containing tumor were scored by applying the Positive Pixel Count algorithm with an optimized color saturation threshold. Any staining was considered positive and the number of positive pixels was divided by the total pixel count. Scores from both cores were subsequently averaged and multiplied by 100 to obtain the percentage of positive pixels. 35 3.2.3 Whole-genome sequencing and analysis Whole genome sequencing (WGS) libraries were constructed from genomic DNA using PCR-free library construction protocols, with the exception of libraries from cases FL1001, FL1002, FL2001 and FL2002 that were constructed during an earlier phase of the project using PCR-containing protocols. Libraries were sequenced on Illumina HiSeq 2500 instruments, generating 100-125 base pair-long paired-end sequence reads. The BWA (v0.5.7) aligner291 was used to align paired-end reads to the human reference genome GRCh37. While assessing the number and type of somatic single nucleotide variant (SNV) substitutions across all our samples, it became apparent that samples FL1005T1, FL1009T1, FL1009T2, FL1012T1, FL1012T2, FL1014T1, FL1014T2, FL2005T1, FL2005T2 and FL3014T1 had an overwhelming representation of C to A substitutions at low allelic rations (between 10-15%). The cases appeared to be randomly affected. The low allelic ratio C to A substitutions were documented to be artifactual by amplicon sequencing in sample FL1007T1 and deemed to be compatible with oxidative damage during DNA shearing, as described by Costello et al.311 We therefore filtered these positions out by applying allelic ratio filters for C to A substitutions in the affected samples (filter set at 0.20 for FL1005T1, FL1009T1, FL1009T2, FL1012T2, FL1014T1, FL3014T1 and at 0.25 for FL1012T1, FL1014T2, FL2005T1, FL2005T2). MutationSeq (v4.1.0)312 and Strelka (v1.0.13)313 were used to predict SNVs for each tumor-normal pair. For TFL and PFL patients, a patient-centric candidate list of SNV positions was generated by aggregating the candidate SNVs positions across both timepoints (i.e. T1 + T2 candidate SNVs). For NPFL, the patient-centric candidate list is equivalent to the T1 candidate list as there is only one timepoint sample. MutationSeq was then re-run specifically interrogating the patient-centric SNV candidate list across both timepoints (for TFL, PFL patients) and single timepoint for NPFL patients. This effectively allowed us to retrieve SNV information at candidate positions predicted by Mutation-Seq and Strelka in a consistent output format, and also across timepoints. Final putative timepoint centric SNVs lists were constructed based on the following criteria: 1) the SNV had a MutationSeq probability ≥ 0.9 and MutationSeq filter field = “PASS”, 2) the SNV was predicted by Strelka and MutationSeq filter field = “PASS”, or 3) SNV was predicted by Strelka, MutationSeq filter = “INDL” and MutationSeq probability ≥ 0.9. HMMcopy (v0.1.1)314 was used to generate coverage wig files for the tumor and normal samples using a window size of 1000 base pairs (“readCounter -w 1000”). Additionally, 36 HMMcopy was also used to calculate GC content of the GRCh37 genome (“gcCounter -w 1000”). Finally, a GRCh37 mappability file was generated by first running HMMcopy’s “generateMap -w 35” to generate a BigWig file which is used as input into HMMcopy’s “mapCounter -w 1000” to generate a final mappability wig file. Copy number alterations (CNAs) were called using TITAN (TitanCNA; v1.5.7)315 using as an input the tumor-normal pair’s read count, coverage data, and normal content estimations along with the GRCh37 GC content and mappability data. The output TITAN results were converted into TITAN segments with each segment taking one of 25 possible copy number states which were collapsed into one of 10 possible summary states: 1) Homozygous Deletion, 2) Hemizygous Deletion, 3) Neutral, 4) 3N Gain, 5) 4N Gain, 6) 5N Gain, 7) 6N Gain, 8) 7N Gain, 9) 8N Gain, 10) Somatic LOH. To increase our specificity for CNAs, we masked our segment data with a copy number variant mask. To assign gene-centric copy number, genes coordinates (Ensembl v72) were overlapped with the TITAN segment coordinates. Destruct was used to predict structural rearrangements for each patient. For TFL and PFL patients, the normal, T1 and T2 samples were used as input into Destruct to simultaneously predict rearrangements across all samples. For NPFL patients, the normal and T1 samples were used. The following filters were applied for Destruct predictions: 1) 0 read in the matching normal sample, 2) distance to any other breakpoint is > 50 basepairs, 3) log likelihood > 20, 4) minimum template length > 120, 5) matescore ≤ 10, 6) both rearranged partners must be on an autosomal or sex chromosome, 7) not found as a variant in the database of genomic variants database, and 8) number of split reads > 0 if the rearrangement does involve the IGH locus. When considering rearrangements that may affect a gene, the breakpoint must be inside the gene itself and not upstream or downstream of the gene. A rearrangement was considered to be shared if there was ≥ 1 read in both timepoints, and timepoint specific if it contained 0 reads in one timepoint. 3.2.4 Targeted deep amplicon sequencing and analysis For all cases, we selected at least 192 predicted mutations to be taken forward for targeted deep amplicon sequencing. These mutations included all synonymous and non-synonymous SNVs as well as coding indels. As this sum fell typically short of 192, we backfilled the list of positions for deep amplicon sequencing to a total of 192 by also including non-coding SNVs that were proportionally selected from mutational clusters generated from each patient's SNV allelic ratio data. 37 Primers were designed using an in house automated primer design pipeline using the following input parameters for Primer3:316,317 primer size 18-26 (optimum 22), primer Tm 57-63 (optimum 59), primer GC% 30-70% (optimum 50%), product size ranges 150-180 and 140-200. Primers were verified to amplify only a single PCR product using the in silico PCR tool from the UCSC Genome Browser.318  Forward primers were tagged with 5'- CGCTCTTCCGATCTCTG-3' and reverse primers with 5'- TGCTCTTCCGATCTGAC-3'. PCR was performed in 192 uniplex reactions per normal and tumour sample using 2ng of genomic DNA as input and 0.1 µL (0.2 units) Q5 High-Fidelity DNA Polymerase (NEB) in a 10 µL reaction. PCR conditions were as follows: 98C for 3 minutes, (98C for 80 seconds, 64C for 30 seconds, 72C for 30 seconds)x35, 72C for 2 minutes. Amplicons were pooled per sample and subjected to a second round of PCR using primer pairs containing a single 5 nucleotide index within the reverse primer. Amplicon pools from the second round of PCR were then pooled to a maximum of 20 samples per pool and sequenced on an Illumina MiSeq instrument using 300v2 kits and generating 150bp paired end reads. Reads were filtered out if they either 1) aligned > 10 base pairs away from an amplicon’s start or end position, 2) had > 5 mismatched bases, or 3) mapping quality < 30. Read counts supporting the reference and variant base were then extracted for each predicted position from WGS. Only reads that had a base quality ≥ 30 were considered in the counting. For each targeted position, we calculated the background error rate by interrogating 30 base pairs up- and downstream of the targeted position by calculating the allelic ratio of the most frequent base. The mean allelic ratio of these positions was considered the background error rate after ignoring germline and somatic mutation positions. We next used a binomial exact test to test if the predicted variant allele was present using a p-value of threshold of < 0.000001. A position was considered somatic if the variant allele was present in tumor and absent in matching normal. In some normal samples, there was evidence of contaminating tumour DNA leading to the presence of the variant allele in the matching normal. To deal with these situations, we performed a one-tailed fisher exact test on the tumor and normal read counts testing if the allelic ratio was higher in the tumor. A significance level of < 0.05 was set to specify if a position was somatic. The mean standard deviation validation rate (precision) was 96.3% +/-  5.4%. 3.2.5 Clonal analysis For each TFL and PFL patient, we inferred the mutational cellular prevalence of each validated SNV (i.e. proportion of cancer cells with SNV) for both the T1 and T2 sample. 38 Specifically, any mutation that was validated in a T1 and/or T2 sample was used as an input into PyClone (v0.12.7).319 PyClone requires for each SNV: 1) read count of the variant allele, 2) read count for the reference allele, and 3) major and minor copy number. For 1) and 2), these data were retrieved from the targeted deep amplicon sequencing. The major and minor copy number data of each feature was taken from the TITAN segments. In addition to these inputs for each SNV, the tumor content estimated from the SNV data was also used. Any SNV without matching copy number data or in a homozygous deleted region was not considered for PyClone analysis. Clonal phylogenies were inferred from PyClone mutational cellular prevalences using CITUP.320 3.2.6 Digital droplet PCR Digital droplet PCR (ddPCR) was performed on selected pairs of T1 and T2 biopsies. As controls, we used 200 ng of genomic DNA from each of three reactive lymph node samples from unrelated individuals and the data from these samples were pooled. For the test samples, we used a minimum of 600ng and 200ng of genomic DNA for T1 and T2 samples, respectively. Primers and FAM- or HEX-tagged probes were designed by Integrated DNA Technologies (IDT). PCR was performed in droplets generated using the Biorad QX200 droplet generator and each 20 µL reaction mix contained 900nM forward and reverse primers, 250nM FAM and HEX probes and 5 units of HindIII, in addition to 1x ddPCR Supermix for Probes (No dUTP) (Biorad), genomic DNA and water. For each pair of probes, the optimal annealing temperature was determined using a temperature gradient. PCR conditions were as follows: 95C for 10 minutes, (94C for 30 seconds, optimal annealing temperature for 90 seconds) x 39, 98C for 10 minutes. Droplets were assessed for FAM or HEX fluorescence using the Biorad QX200 droplet reader. In all instances, we verified that no signal for either wild-type or mutant DNA could be detected in the non-template control wells. We clustered results from T2 samples using Gaussian mixture models and the R package mclust v5.1, setting the number of mixture components to 4 and initializing hierarchical clustering on a random sample of 1000 datapoints. If clustering using this approach did not reveal four distinct clusters corresponding to empty droplets, wild-type only droplets, mutant only droplets or double positive droplets, we repeated the clustering until an appropriate solution was found. The seed of the first successful clustering was then set for all other samples corresponding to the same case. As DNA degradation with time potentially competes with the detection of rare alleles, we only considered single, mutant signal-positive droplets to be positive events, as described in Wong et al.321 39 3.2.7 Capture-based targeted sequencing We selected 86 genes (Appendix A) for capture-based targeted sequencing, based on the following 6 criteria: 1) recurrence in FL of >5%,126,128 2) recurrence in DLBCL >5% (our own data) or reported to be consistently mutated in Burkitt lymphoma across studies,322–324 3) genes significantly mutated (q <0.05) in the compiled dataset from our study and others,135,136 applying the MutSigCV algorithm,325 4) T1 gene mutations associated with early transformation/progression (<5 years) versus no progression (for at least 5 years) based on our data and external cohorts,135,136 5) genes that were found to be T2-specific in at least three cases from our and external cohorts,135,136 and 6) IL4R, PTPN1, NOTCH2 and RFX5. Furthermore, we selected 20 genes, twelve of which overlapped with the 86 above-mentioned genes, to assess mutations in targets of SHM (Appendix B). Libraries were constructed from either 500ng of fresh-frozen genomic DNA or 200ng of FFPET-derived genomic DNA, and captured using custom SureSelectXT2 baits (Agilent). Captured libraries were pooled to a maximum of 46 libraries per pool and each pool was sequenced on one Illumina HiSeq lane, generating 125bp indexed reads (V4 chemistry). Different filtering criteria were applied to call SNVs in the coding space of the 86 gene panel, and in the coding and non-coding space of the 5' regions of SHM targets. For the 86 gene panel, predictions were filtered as follows: mutation probability > 0.7, coverage >100 or (coverage > 50 and mutation probability > 0.9). For cases with available germline DNA (n = 73), SNVs were filtered out if they were present in the germline samples. For all other samples (n = 135), putative germline variants were identified as follows and filtered out: present in the 1000 Genomes Project, present in dbSNP version 137 and not present in the Cosmic database, present in at least two germline samples from the 73 cases with matching germline library. For five cases without germline DNA sequenced by targeted capture sequencing, SNPs were filtered out using matching germline whole-genome sequencing libraries. Three positions were filtered out of the dataset as they were deemed to represent artifacts due to high recurrence and consistent low VAFs (chr12:57501386, chr12:113515335 and chr19:11098449). SNVs were annotated with SnpEFF and if multiple effects were called for a single position, the effect with the putatively greatest impact was chosen as described above for SNV calling in our WGSS cases. Multiple, distinct mutation calls in a given gene were collapsed into a single call using the same effect filter, and only coding mutations were taken forward for analysis. Filtering of these data for the 20 SHM 40 genes was done similarly, except that coding and non-coding predictions were retained for the analysis. For samples with matching germline, indels were called using Strelka (v1.0.13) and the output was not filtered with the exception that only coding indels were considered for the analysis of variants in the 86 gene panel. For samples without matching germline, indels were called using VarScan and the following putative artifacts were filtered out: chr1:27100181, chr1:85736510, chr1:120612002 and chr22:23653975. 3.3 Results 3.3.1 Whole genome sequencing results We performed whole genome sequencing (WGS) of paired tumour samples (~ 50-fold coverage) from 15 patients presenting with transformed lymphoma (transformed FL or TFL group) and from six patients presenting with progressive FL but without evidence of transformation (progressed FL or PFL group). We also sequenced to similar coverage a control group of single tumour samples from 20 patients presenting with indolent FL, but without progression for at least 5 years following diagnosis (long-term non-progresser or NPFL group). Matched germline DNA was sequenced to ~ 30-fold coverage in all cases. A timeline of samples can be found in Figure 12. Somatic single nucleotide variants (SNVs), small insertions and deletions (indels), CNAs and rearrangements were predicted for each tumour sample as described in Chapter 3.2.3. An overview of sequencing results is shown in Figure 13. T1 samples from the three clinical groups (TFL, PFL and NPFL) had similar numbers of SNVs, indels and CNAs, but differed in the burden of rearrangements (mean number per T1 sample 31.33 +/- 23.29, 17.0 +/- 8.88 and 16.9 +/- 13.76 in the TFL, PFL and NPFL groups, respectively; P = 0.008; Kruskal-wallis test; Figure 14). On a temporal axis, we observed a significant increase in genetic alterations between T1 and T2 timepoints for all mutation types, and for both TFL and PFL patients. This increase was independent of the time interval between sampling. Additionally, when comparing the T2 sample mutation load between TFL and PFL patients, with the exception of the SNV space (P = 0.056 and P = 0.348 for coding SNV) we observed that TFL patients had a higher mutation load in the indel (P < 0.001), CNA (P = 0.018) and rearrangement space (P = 0.028) compared to PFL patients. Across all cases, the most frequent alterations were found, as expected, in BCL2, CREBBP, KMT2D and EZH2. Homozygous deletions of the cell cycle progression inhibitor CDKN2A were found in four patients of the TFL cohort, and exclusive to the T2 timepoint in three out of four patients. CDKN2A alterations were not observed in the PFL or the NPFL groups. 41 3.3.2 The emergence of a T2-dominant clone underlies transformation Next, we performed deep amplicon sequencing on ≥ 192 putative somatic SNVs or indels per patient, across all normal and tumour samples, and clustered mutations based on their cellular prevalence, accounting for tumour content and allelic imbalances due to copy-number alterations using PyClone.319 These results were then used as an input for inference of the clonal phylogeny of each patient's tumour, ultimately allowing us to study the clonal expansion patterns and elucidate the temporal order of mutation acquisition. We first considered the clonal architecture of TFL cases, and uncovered the existence of subclones in a majority of T1 samples from this group of patients, suggesting that intratumoural heterogeneity (ITH) is a typical feature of FL (Figure 15). This finding raised the question of whether these subclones expand over the course of transformation and contribute to aggressive lymphoma. However, we observed the expansion of evident clusters of subclonal mutations between T1 and T2 in only one sample (FL1016). Rather, the predominant mode of evolution for TFL patients was the expansion of clade-specific clones that were seemingly absent in T1 but arose to complete dominance in T2, in keeping with branched evolution. This pattern was observed in 13 cases out of 15 (87%) and suggests that in the majority of cases, diagnostic specimens would likely not be reliable predictors of transformation, and furthermore that the clonal dynamics occurring after diagnosis underpin histological change. This pattern was seen independent of time to transformation. For example, the T2 sample from FL1007 (transformed after 14.57 years), characterized by FOXO1 and BCL6 mutations in the ancestral clone (cluster 1), was entirely composed of a clonal lineage that harboured B2M and CCND3 mutations (clusters 2 and 3) which were near 0 prevalence levels in T1 samples (Figure 15). Notably, these clones were mutually exclusive to the clonal lineage dominating the T1 samples (clusters 4-7-6, and 5). Case FL1017 (transformed after 0.42 years) characterized by CREBBP and KMT2D mutations in its ancestral clone, harboured a T2-specific lineage containing EZH2 and FOXO1 mutations (clusters 2 and 1) exhibiting a similar distribution of clones to FL1007. This pattern of clonal dynamics was also independent of treatment regimen and could be seen in cases under observation alone or no therapy (FL1007, FL1006, FL1012, FL1014, FL1019) and those cases variously treated with rituximab or combination therapy (FL1001, FL1005, FL1013, FL1016, FL1004, FL1008, FL1017). Two cases (13%) exhibited patterns of dynamics that contrasted with the dominant pattern. In these cases (FL1009 and FL1020 - both observation only and both relatively short time 42 to transformation: 0.39 years and 0.78 years respectively), the dynamic properties were consistent with conserved clonal architecture (FL009) and only modest dynamics (FL1020). Thus, a small minority of cases may already contain the properties driving transformation at time of diagnosis. Together, these results reveal a striking pattern of clonal dynamics underpinning histological transformation in the majority of TFL cases that is at once independent of time to transformation and treatment regimen. 3.3.3 Primary progression results from the selection of pre-existent clones The clonal dynamics in progressed samples exhibited markedly different patterns relative to transformation cases (Figure 16). Four cases (FL2002, FL2005, FL2007 and FL2008) harboured readily detectable clones at T1 which expanded to full clonal prevalence over the course of R-CVP +/- rituximab maintenance therapy. This suggests that clones harbouring treatment resistance properties were already present at the time of diagnosis, and that symptomatic disease progression may be attributable to selection of clones that were major constituents of the diagnostic specimen.  This mode of progression is reminiscent of clonal evolution described in chronic lymphocytic leukemia (CLL), another mature, incurable and typically relapsing lymphoid malignancy.225,226 FL2006 showed modest dynamics, with relative stability of the clonal phylogeny across the two timepoints. An exceptional case (FL2001) in the PFL group exhibited dynamics similar to the predominant mode observed in TFL with a T2 specific lineage with ARID1A mutation (clusters 2, 3) coming to dominate the relapse specimen with no evidence of the T1 clones (clusters 4, 5). This patient initially presented with indolent lymphoma, was observed for 2 years, then received single agent rituximab. The patient was observed again for 2 years and then presented with symptomatic, progressive lymphoma unresponsive to three lines of systemic therapy, leading to the patient's death. In this case, where samples were taken at extremes of clinical behaviour, the phylogenetic structure was analogous to the TFL pattern, yet the biopsy from T2 showed no evidence of large cell transformation. Thus, treatment resistance patterns accompanied by significant clonal dynamics can occur in FL in the absence of overt transformation. Overall results for PFL clonal dynamics suggest that progression on therapy is driven by a starkly different mode of evolution than what was observed for TFL. Thus, these two clinical end points are likely underpinned by non-overlapping mechanisms with early PFL harbouring 43 intrinsically resistant properties at diagnosis and TFL generally acquiring the dominant transformation phenotype after diagnosis. 3.3.4 Ultra-sensitive identification of low prevalence clones in T1 samples Whereas clonal selection from underlying subclones was evident in FL1016 and in the aforementioned four PFL cases, the PyClone results suggested at first glance that in most TFL cases, and in PFL patient FL2001, the mutations from the dominant, expanding T2 clones were not found at a significant level in the T1 timepoint. However, the presence of even small subclones in the T1 timepoint has pertinent clinical implications as early detection of relevant subclonal mutations could be utilized to alter initial patient management. At least in CLL, small subclones harbouring TP53 mutations have been reported to have similar unfavourable prognostic implication than clonal mutations.326 Therefore, for all cases in which evolution was characterized by the emergence of a T2 dominant clone, we assessed whether read support for T2 dominant clone-associated mutations was present in T1 (Figure 17), and found that the mean variant allele frequency (VAF) ranged from 0.01% in FL1012T1 to 0.05% in FL2001T1. To quantitate their presence or absence by an orthogonal method, we selected 1-3 mutant alleles representing the dominant T2 clone as templates, and performed ddPCR on genomic DNA (T1, T2 and reactive lymph node (RLN) control samples), corresponding to approximately 100,000 diploid cells for each sample. In total, eight alleles (VAFs from T1 targeted sequencing ranging from 0.00016 to 0.01110). were assayed. We applied a Gaussian mixture model using four a priori groups corresponding to single mutant positive (red), double positive (purple), single wild type (blue) and empty droplets (grey) to cluster the resultant intensity values from T2 samples across the wild type and mutant allele spectrum. Fitted distributions were then used to classify the T1 intensity values, assigning each data point to one of the four classes. All eight T2 alleles showed presence of a single mutant positive cluster, consistent with the VAF measurements from targeted amplicon sequencing (Figure 18). For T1 samples, FL1012 and FL1004 mutant alleles were found in exactly 0 droplets, indicating that prevalence of these alleles in the diagnostic specimens is likely below 10-5, and may in fact be 0. For these samples, de novo acquisition of the T2 clones after diagnosis could not be ruled out. CCND3 and CD83 mutant alleles in FL1019 were found in 2 and 19 droplets respectively, while FL2001 mutant alleles in ATP6V1B2, IL11RA and ARID1A were present in 25, 24 and 39 droplets respectively. Results for FL1019 and FL2001 are therefore 44 consistent with detectable alleles in the T1 sample, providing robust evidence that the T2 specific clones were indeed present (albeit at low prevalence) at diagnosis. For these cases, evolutionary patterns are likely the result of massive clonal expansion of pre-existing clones over the interval between T1 and T2 samples, and de novo acquisition of driver mutations after diagnosis can be ruled out. These data are also consistent with a clone as rare as 2 molecules from approximately 105 cells at diagnosis coming to dominate the transformed specimen. 3.3.5 Recurrent gene alterations in transformed follicular lymphoma In order to define recurrent genetic alterations in FL and TFL, and assess their association with patient outcomes, we sequenced the coding sequence of 86 genes, as well as the 5' coding and non-coding regions of 20 genes (supposed regions of SHM). Twelve genes overlapped between the two panels. Sequencing was performed by high-throughput sequencing after hybridization-based capture of 252 genomic DNA samples (T1 and/or T2) from 205 patients. The mean target coverage was 1054 +/- 236, and 97.3% of target bases were covered at least 100x. Fifty-five samples were sequenced both by WGS and targeted capture-based sequencing, and of all non-synonymous SNVs predicted by WGS, 292 out of 319 (91.5%) were also predicted by our analytical pipeline for targeted sequencing. In contrast, capture-based sequencing identified a further 261 SNVs that had not been predicted by WGS and that had, on average, lower VAFs, suggesting that the high coverage obtained in the targeted sequencing experiment had better sensitivity for the identification of lowly abundant variant alleles. First, we assessed mutational recurrence in all 172 FL (T1) samples. On average, samples harboured non-synonymous mutations in 7.12 +/- 3.46 genes out of 86. As expected, chromatin modifiers were the most frequently, as well as the most significantly mutated genes (CREBBP 66.9%, KMT2D 55.8% and EZH2 22.1%; Figure 19). Next, we interrogated samples from a cohort of 116 TFL patients for which 81 FL samples and 83 TFL samples were available (including paired specimens for 47 patients). Similar to our findings from WGS, the mutational load based on targeted sequencing was higher in T2 than in T1 samples (mean number of mutated genes 8.89 +/- 3.50 versus 6.83 +/- 3.3, P < 0.001) (Figure 20a, b, c), and this increase was mostly driven by mutations in putative driver genes, rather than passenger mutations. The burden of SHM across the 20 gene panel did not significantly differ between T1 and T2 samples (Figure 21). Taken together, these results suggest that transformation is the result of the outgrowth of a clone (or 45 clones) harbouring distinct sets of driver mutations, rather than the result of unselective accumulation of bystander mutations over time. As large cell lymphoma arising in a patient with a prior history of indolent lymphoma can theoretically be clonally unrelated to prior FL, we classified all coding and non-coding mutations in the 47 patients in which paired samples where available into T1-exclusive, shared, or T2-exclusive mutations, based on whether these mutations were called in T1 and/or T2 by our analytical pipeline. This analysis revealed the presence of shared mutations in all 47 patients, confirming the clonal relationship between indolent and aggressive lymphoma (Figure 22). The proportion of shared mutations was inversely correlated to time to transformation (Spearman P = 0.044) (Figure 20d), suggesting that the degree of clonal divergence was in part related to time elapsed between diagnosis and transformation. The analysis of coding mutations in all T1 and all T2 samples revealed four genes that were differentially mutated between T1 and T2 samples: TP53 (mutated in 35% versus 18% of T2 and T1 cases, respectively, P = 0.013), B2M (25% versus 9%, P = 0.006), ITPKB (18% and 6%, P = 0.030) and CCND3 (14% versus 4%, P = 0.028). Of these four genes, the latter two genes have not been previously described to be associated with transformation. Importantly, although mutations in these four genes were significantly enriched in T2 samples, they were not exclusive to the transformation timepoint as they were also found to be mutated at a lower frequency in preceding FL, suggesting that they are not sufficient do drive transformation and that their phenotypic footprint is dependent on the contextual landscape of other genetic or epigenetic aberrations. As TP53 mutations are uncommon in FL at diagnosis and as CCND3 and B2M mutations are typically associated with aggressive lymphoma, we asked the question whether the presence of any T2-associated mutation in T1 resulted in a shortened time to transformation. Overall, 31% of patients had a mutation in any of these four genes in the T1 sample (Figure 20g) and presented indeed with shorter time to transformation (median time to transformation 1.98 versus 3.40 years, P  = 0.023, Figure 20h). The high prevalence of inactivating B2M mutations, previously described in de novo DLBCL327 and TFL,136 suggests that functional immune surveillance is critical to prevent transformation. B2M is a component of major histocompatibility class (MHC) class I and in de novo DLBCL, B2M mutations have been shown to correlate with the absence of B2M and HLA-I protein expression by immunohistochemistry (IHC).327 However, the consequences on the 46 composition of the tumour microenvironment have not been reported. We stained a tissue micro-array containing cores matching the samples from our study and found that B2M mutations were significantly correlated with decreased numbers of CD8+ T-cells in TFL but not FL samples, as assessed by automated imaging (Figure 23), suggesting that B2M mutations dampen the cytotoxic T-cell response. We described in Chapter 2 that composite histology at time of transformation is associated with improved survival when compared to morphologies consistent with DLCBL or BCLU. The genetic correlate of this observation was a higher prevalence of STAT6 mutations (38% versus 9%, P = 0.012), and a lower prevalence of TP53 mutations (3% versus 41%, P = 0.042) in composite TFL, when compared to DLBCL-TFL (Figure 20i). Similar to de novo DLBCL, two gene-expression groups can be distinguished within TFL using the cell-of-origin (COO) classification (Chapter 2). However, it is unknown whether TFL COO subtypes differ by mutational status of genes involved in B-cell receptor and NF-κB signalling, akin to de novo DLBCL. COO classification was available for 79 cases, 10 and 69 of which were of the ABC and GCB subtype, respectively. Although the number of ABC-TFL cases was small, we observed significantly more frequent CD79B (30% versus 3%, P = 0.008), BCL10 (2% versus 0%, P = 0.011) and FAT4 mutations (30% versus 6%, P = 0.025) in ABC-TFL than in GCB-TFL (Figure 20j). The presence of CD79B and BCL10 mutations points, as hypothesized, to activation of B-cell receptor and NF-κB signalling and suggests that ABC-TFL is amenable to targeted therapeutic inhibition of these pathways. 3.3.6 Association of gene mutations with early progression Next, we assessed the association of gene mutations with patient outcome and leveraged a cohort of 30 early progressers (progressing < 2.5 years after starting R-chemotherapy) and 67 late or non-progressers (i.e. not progressing for at least 5 years). Samples from early progressers were, as expected, enriched for poor-risk clinical factors including poor performance status, tumor mass ≥ 7 cm, elevated LDH, low hemoglobin and high-risk FLIPI (Table 6). Grade 3A was also more frequent in diagnostic biopsies from early progressers (30% versus 7%, P = 0.009). Early progressers had, on average, a higher number of mutated genes than late progressers (8.7 +/- 4.0 versus 6.9 +/- 3.4, P = 0.044), and this effect was mostly driven by a higher number of mutations in putative driver, rather than passenger genes (Figure 24 a, b, c) Although coding and non-coding mutations were common in the 5' regions of known targets of SHM (mean number of mutations 47 per sample 48.3 +/- 41.8), their abundance was highly variable between individual samples and did not significantly differ between early and late progressers (Figure 25). Per definition, progression-free survival was distinct for the two outcome groups and translated into highly significant differences in overall survival (median overall survival 3.15 years for early progressers versus not reached for late progressers, P < 0.001; Figure 24 d, e). However, despite the stark contrast in clinical characteristics and prognosis, only four genes were significantly associated with early progression: KMT2D (mutated in 80% of the early progressers and 52% of the late progressers, P = 0.013), BTG1 (13% versus 1%, P = 0.031), TP53 (13% versus 1%, P = 0.031) and XBP1 (13.3% versus 0%, P = 0.008) (Figure 24 f, g). To the best of our knowledge, only TP53 mutations have previously been reported to be associated with adverse outcome in FL. Mutations in the transcription factor and effector of the unfolded protein response XBP1 were remarkable because they occurred exclusively in early progressers and because three out of four mutations targeted splice sites, including one mutation affecting the splice site donor for IRE-1-mediated unconventional splicing and activation of XBP1. An adjacent mutation in this site has been shown to confer resistance of multiple myeloma cells to bortezomib.328 Among the four differentially mutated genes, XBP1 had the highest odds ratio for association with early progression, followed by TP53 and BTG1 (Figure 24h). In our cohort that was enriched for clinical extremes, the m7-FLIPI had a similar odds ratio to the FLIPI. The m7-FLIPI had better specificity (90% versus 78%) but worse sensitivity (36% versus 57%) than the FLIPI. The fact that the m7-FLIPI was not superior to the FLIPI suggests that the biology of early progression is distinct and, indeed, none of the four progression-associated genes in our study are taken into consideration for calculation of the m7-FLIPI. 3.4 Discussion The goal of our study was to uncover the clonal dynamics underlying transformation and progression in FL, as well as to identify gene alterations that are associated with early progression or emerge during the process of transformation. We showed that transformation is characterized in the majority of cases by the emergence of a clone that becomes dominant at T2 and lies below the detection limit of standard, low-pass mutation calling methods at the initial FL timepoint. On the other hand, early progression of FL typically arises from evident subclonal disease in T1. The striking divergent modes of evolution of PFL and TFL are reminiscent of the distinct differences between the clinical presentation of these entities, transformation being uniquely associated with 48 rapid onset of tumor growth and systemic symptoms, suggesting an underlying acute shift in tumour biology. We were able to delineate the clonal population structures in our samples only because of rigorous statistical inference, taking into account cellular mutational prevalence as well as allelic imbalances attributed to copy number changes or non-tumour cell contamination. Additionally, we described gene mutations that are significantly enriched in TFL when compared to preceding FL, and emerge as potential drivers of aggressive lymphoma. They include previously described mutations in TP53, B2M and translocations resulting in overexpression of MYC, as well as mutations in CCND3 and ITPKB that have not been previously associated with transformation. Taken together, they enforce the notion that increased proliferation, a defective DNA damage response and escape from immunosurveillance are critical for development of an aggressive phenotype.136 However, none of these alterations is sufficient to drive transformation as they can be found in preceding FL, suggesting that their effect is dependent on the genomic context in which they arise. Our findings have potential clinical translational relevance. First, we show that, as for most cancers, intratumour clonal heterogeneity is a defining feature of FL and that subclones present at the time of diagnosis can represent genetic harbingers of transformation or progression. Primary resistance to upfront combined modality therapy generally occurs by the selection of resistant clones that can be found at a subclonal level at the time of diagnosis, suggesting that their detection can predict the development of resistance to treatment. In that regard, we assembled a unique cohort of early progressers that are at the extreme, poor prognosis end of the FL spectrum and, despite striking adverse outcome, only four genes were mutated at a higher rate in early versus late progressers (KMT2D, BTG1, TP53 and XBP1), suggesting that other, yet undefined alterations drive treatment resistance, or that our sampling from the total clonal population structure was limited by spatial heterogeneity or the small size of potential subclones. A large body of evidence shows that cancer evolution follows aspects of Darwin's natural selection theory, by virtue of selection of the fittest clones occurring in a context of underlying genetic heterogeneity, and by virtue of evolution typically following a branched pattern.274,275,329 However, whereas Darwinian evolution is characterized by gradual drifts over time, transformation occurred in our series by drastic shifts of the clonal structure, and the typical rapid onset of TFL suggests that these shifts occurred within short windows of time. Moreover, in a subset of our cases, all measured T2 dominant clone-associated mutations could be detected in T1, 49 suggesting that, in the absence of selective pressure in the form of therapy, they existed in an equilibrium. Hence, the dynamics of FL transformation resemble aspects of punctuated evolution in which periods of stasis are interrupted by punctuated changes.330 Whether drivers of treatment resistance or transformation are mutated in a series of incremental steps, or whether they occur in single, multifocal, genomic insults, is presently unknown. Future studies will have to dissect the early steps of FL pathogenesis to fully understand the sequence of mutations as they arise within clonal phylogenies. 50 Chapter 4: Conclusion  4.1 Summary of research findings In hypothesis 1, we asked whether molecular subtypes could be recognized within TFL by low resolution gene expression profiling. To address this question, in Chapter 2, we applied the Lymph2Cx assay to a cohort of 110 TFL samples, all of which had a morphology in keeping with DLBCL. We discovered that 80% of cases were of the GCB subtype, and 16% of the ABC subtype. The distribution of the linear predictor score within TFL molecular subtypes was superimposable to the distribution of that score within de novo DLBCL molecular subtypes, suggesting that GCB and ABC-TFL are as distinct from each other as GCB and ABC-DLBCL. Furthermore, we assessed the prevalence of staining for CD10, BCL6 and IRF4 by IHC (Chapter 2), the prevalence of BCL2, BCL6 and MYC translocations by FISH (Chapter 2), and the mutational status of 86 genes within both molecular subtypes (Chapter 3), and found striking differences between the molecular subtypes in TFL. BCL2 translocations, although present in 85% of FLs, were found in only 36% of ABC-TFL (vs. 89% of GCB-TFL, P < 0.001). IRF4 expression is a hallmark of ABC-DLBCL, and was also more commonly expressed in ABC-TFL than in GCB-DLBCL (38% vs 3%, P < 0.001). Importantly, BCL2 translocation-negative status, and expression of IRF4 in the preceding FL were associated with subsequent transformation into the ABC-TFL subtype, suggesting that underlying heterogeneity exists in FL and translates into distinct molecular subtypes at transformation. We further assessed the coding sequence of 86 genes by capture-based sequencing, and found that CD79B, FAT4 and BCL10 were more commonly mutated in ABC-TFL than in GCB-TFL, further strengthening our conclusions with regard to the existence of inter-patient molecular heterogeneity in TFL. In hypothesis 2, we asked whether transformation and progression are the result of selection of pre-existing subclones that harbour driver alterations. To that effect we applied whole-genome sequencing to a cohort of 15 transformed trios (normal, FL, TFL) and 6 progressed trios (normal, FL, PFL), followed by deep-sequencing of at least 192 SNVs per patient. The clonal architecture was inferred using PyClone.319 With regards to progression, we could readily detect, in four out of six patients, the presence of subclonal mutations at diagnosis that increase from a subclonal level to clonal dominance in T2, suggesting selection as a result of treatment and the presence of resistance-conferring genetic alterations at the time of initial sampling. The picture 51 was different in transformed lymphoma, where selection of a “macroscopic” subclone was apparent in only one case out of 15. Instead, transformation was characterized by “T2-dominant clones” that were seemingly absent in T1, and rose to complete dominance in T2. However, when assessing allelic ratios for “T2-dominant clone”-associated mutations in T1, we observed that read support for the variant alleles was found in several cases, suggesting the “T2-dominant clone” was present at a “microscopic” level in T1. We validated this finding by digital droplet PCR. Furthermore, we sequenced a panel of 86 genes by hybrid capture and high-throughput sequencing in a cohort of 116 patients (80 T1 samples, 82 T2 samples, 47 samples for which T1 and T2 were available). The analysis of the coding regions of these genes revealed that TFL had on average a higher number of mutations than FL (8.6 vs 6.7, P < 0.001), but this difference was restricted to significantly mutated genes (putative drivers), and not found in non-significantly mutated genes (putative passengers). Furthermore, we compared gene mutations between T1 and T2, and found four genes to be significantly associated with transformation (TP53, B2M, ITPKB and CCND3), suggesting that the proteins encoded by these genes contribute to the process of transformation. Lastly, in hypothesis 3, we asked whether early transformation and/or progression can be predicted at time of diagnosis. To answer this question, we assessed a cohort comprised of 40 patients presenting with early transformation (5 years), and 86 patients not experiencing transformation or progression for at least 5 years. We showed that grade 3A, BCL6 translocations, negative staining for CD10 and positive staining for IRF4 were all associated with early transformation in univariate analysis. In a weighted Cox regression model, only IRF4 expression remained significantly associated with transformation. In Chapter 2, we applied the previously mentioned 86 gene panel to a cohort of 30 cases who progressed early (<2.5 years) after starting R-chemotherapy and 67 cases who progressed late or never after starting R-chemotherapy or observation. Four genes were significantly associated with early progression (KMT2D, BTG1, TP53 and XBP1) and no gene was significantly associated with late progression in univariate analysis. Only TP53 gene mutations had previously been reported to be associated with poor outcome in FL.264 4.2 Integration into the research field 4.2.1 Intratumoural heterogeneity and clonal evolution We found evidence for the existence of genetic subclones in all cases from our discovery cohort and, akin to other cancers, FL has therefore to be regarded as a mosaic of co-existing cellular 52 populations that differ in genetic composition. The existence of subclones is a reflection of ITH, a hallmark of malignancy that has been recognized for decades.331 ITH is inherently related to tumour progression as it implies varying degrees of fitness, or resistance upon application of selective pressure (e.g. therapy, supply of nutrients, immune status, etc), leading to the question of whether quantitative or qualitative attributes of the clonal population structure are a determinant of patient outcomes. Precise, pan-cancer quantification of ITH, as defined on a genetic level, has only recently become possible as the number of available cancer exomes has increased, and as analytical tools to decipher clonal composition through bioinformatics means have been developed.332,333 The relationship between number of clones and prognosis is seemingly non-linear as it was shown that both low and high numbers of clones are associated with good outcome, whereas intermediate numbers of clones correlate with unfavourable survival.333 In our study, the number of clones found at diagnosis was not related to early progression or transformation. Beyond quantification, qualitative attributes such as the presence of selected driver mutations appears to be crucial as they may confer resistance.277 In line with prior observations in other hematological cancers,321,326,334,335 we found evidence for the presence of subclonal drivers (e.g. CD58, CCND3, ARID1A) at diagnosis that emerged over the course of transformation or progression, suggesting that they contributed to tumour evolution. We also found evidence of association of specific clonal or subclonal gene mutations with early progression (KMT2D, BTG1, TP53 and XBP1). Tumour evolution has repeatedly been likened to Darwinian evolution in the sense that genetic variation forms the substrate for selection in the presence of environmental forces.274,275,336 Darwinian evolution is typically branched and occurs in gradual, incremental steps. The clinical observation that the onset of transformation is generally abrupt (weeks rather than months or years), and the finding from our study that aggressive T2-associated clones can pre-exist in an indolent steady state, questions the gradual nature of transformation. Indeed, transformation shares defining traits of punctuated equilibrium, a theory that was first proposed to account for the apparent instantaneous commencement, followed by stability in speciation.330 In keeping with punctuated dynamics, transformation was characterized, in our study, by the sudden outgrowth of clones that became dominant in T2 and were not present at a “macroscopic” level in T1. 53 4.2.2 The transcriptomic and genetic composition of TFL In order to gain insight into the biology underlying transformation, we assessed COO and the mutational landscape of TFL. Previously, only two studies had reported conflicting results with regards to COO of TFL.123,245 Our study clearly demonstrated the existence of distinct molecular subtypes within TFL. In de novo DLBCL, COO subgroups differ by prognosis, genetic alterations and activated pathways, and the accurate assessment of these subtypes is becoming increasingly important for the accurate identification of patients for targeted therapeutic agents, at least in the scenario of clinical trials.337 We showed that molecular subgroups of TFL harbor distinct genetic profiles, suggesting that, akin to de novo DLBCL, the ABC and GCB subtypes of TFL should be considered as potential candidates for different therapies. We did not show a significant survival difference between molecular subgroups in TFL, possible due to the small number of ABC-TFL in our cohort, or to heterogeneity of treatment eras, or type of treatment received prior to transformation. In terms of the mutational landscape of TFL, two studies had reported genome and exome sequencing data on a total of 22 cases.135,136 Our study extends these findings from published studies by adding 15 additional transformed cases, as well as comparison groups of six progressed and 20 long-term non-progressed lymphomas. Similarly, targeted sequencing of gene panels have been performed before, discovering novel associations of gene mutations with transformation (for example TP53 and B2M). Our study confirms these prior observations but it adds further texture by showing that putative TFL-specific alterations (B2M or CD58, for example), can be found at low recurrence in FL, hence suggesting that they are not sufficient to confer an aggressive phenotype. Furthermore, our mutational analysis revealed gene mutations that were not known to be enriched in TFL (ITPKB, CCND3), thereby extending our knowledge of genetic determinants of transformation. 4.2.3 Prediction of early transformation and/or progression Despite an abundant literature on biomarkers that have been reported to be associated with outcome in single studies, virtually none have penetrated clinical practice, a likely reflection of the small sample size of most studies, heterogeneity in timing of sampling or heterogeneity in treatment.338 A legitimate hope is that the unbiased assessment of multivariate models in larger patient cohorts with uniform therapy would predict outcome with sufficient accuracy to actually alter patient management. In this regard, we had previously contributed to the development of a 54 clinico-genetic risk model, named m7-FLIPI, that integrates both clinical risk factors as well as the mutational status of 7 genes.128 Patients with high m-7 FLIPI are enriched within early progressers, but the positive and negative predictive value of the m7-FLIPI for progressive disease at 24 months was only 42-54% and 82%, respectively (unpublished data). Whereas the m7-FLIPI was developed in a clinical trial cohort of R-chemotherapy-treated patients, our present study took a different approach by focusing specifically on early progressers and comparing them to a group of late or never progressers. The aim of our study was to identify genetic mechanisms of early progression rather than the development of an outcome predictor. We were able to show that samples from early progressers harbour mutations in KMT2D, BTG1, TP53 and XBP1 more commonly than late progressers. The latter three alterations were nonetheless rare in early progressers, suggesting that we were able to detect their recurrence due to the enrichment of our patient cohorts for outcome extremes. 4.3 Strengths and limitations 4.3.1 Intratumoural heterogeneity and clonal evolution To assess the clonal population structures of our discovery cases, we applied PyClone, a Bayesian clustering method that estimates mutational cellular prevalence by integrating variant allele frequencies from deep sampling, tumour content and copy number status.319 Prior studies had relied solely on comparisons of allelic ratios between timepoints or on qualitative assessments of mutations. Our study is the first to provide a truly integrated description of clonal evolution over the course of progression and transformation in FL. It is also the first study to assess clonal structure in pairs of both progressed and transformed lymphoma, and to integrate findings with patient outcome data. The observation of distinct evolutionary trajectories for progressed and transformed FL was possible only by having these contrasts in the study. A further strength of our study is the identification of very small subclones by both deep sequencing and digital droplet PCR. However, the sample size of our whole genome sequencing cohort was relatively small, which limited our power to uncover recurrent SNVs, copy number alterations and rearrangements. This factor also limited our ability to correlate measurements of clonal structure with patient outcome categories. Furthermore, we failed to consider mutational signatures because artifacts from library construction precluded this analysis in a subset of patients. Indeed, we observed overwhelming C to A substitutions in 10 samples (from six patients) that we were able to salvage 55 for high-level genomic analysis by stringent filtering. Nonetheless, we did not feel confident to use these samples for the description of mutational signatures that could be reflective of mechanisms underlying genome-wide mutations. Lastly, we solely focused on genomic mutations and did not take into account transcriptomic or epigenetic changes during tumour evolution, nor recurrent alterations in the non-coding space of the genome. A comprehensive description of tumour evolution would require these layers of information. 4.3.2 The transcriptomic and genetic composition of TFL By determining COO in a large cohort of transformed lymphoma samples, we could undoubtedly show that distinct molecular subtypes exist in TFL. ABC-TFL is relatively uncommon, representing only 16% of all TFLs, and hence a large cohort was required to gather a sufficient number of TFLs to see differences in the spectrum of cytogenetic and mutational changes in ABC-TFL vs GCB-TFL. Prior to the development of the Lymph2Cx assay that is eminently applicable to RNA derived from FFPET blocks, COO classification was largely restricted to nucleic acids from fresh frozen material. Our ability to apply the Lymph2Cx assay in archival patient specimens was key to this discovery. 4.3.3 Prediction of early transformation and/or progression We were able to confirm prior observations of associations of pathological and genetic features with early transformation and/or progression. Moreover, by studying enriched cohorts of clinical extremes, we were able to uncover genetic alterations that are rare in unselected patient populations. Although the picture of poor outcome FL has become more clear, and has opened further avenues to explore, we failed to unravel mechanism(s) of treatment resistance that could be overcome by rational targeting of pathway perturbations, and we failed to produce a widely applicable model to predict early, poor outcome patients as our cohorts were enriched and comparatively small, precluding the building of a robust model that would accurately predict outcome in an unselected patient population. 4.4 Potential applications Our description of ITH and small subclones in FL finds its application in the design and interpretation of sequencing results. At present, clinical decisions are not based on gene mutations but it can be reasonably argued that prognostic and predictive gene-mutation based models will be utilized in the routine clinical setting in the future. The experimental and analytical setup will have to take into account that relevant mutations can be present in small subclones, i.e. coverage needs 56 to be high to confidently call low allelic variants. Nonetheless, clinical judgement will likely remain highly valuable. Not uncommonly, we face situations where aggressive lymphoma is suspected based on clinical grounds,232 yet the workup including histology (and potentially in the future sequencing results) is consistent with indolent lymphoma. Given potential spatial heterogeneity, such a scenario should raise the question of whether the clinically relevant tumour was truly sampled.  The discovery of molecular subtypes in TFL is of translational relevance as the ABC and GCB subtypes in de novo DLBCL are amenable to distinct therapeutic targeting.287,288 At the moment, ibrutinib and lenalidomide are being investigated in phase III trials for ABC-DLBCL, and TFL patients are typically excluded. Our results provide the rationale to include these patients and allow for subset analyses of the efficacy of novel agents in all TFLs displaying evidence of active BCR signalling or NF-κB activation. We described novel associations of gene mutations with transformation (ITPKB and CCND3). As TFL remains a harbinger of adverse outcome, especially in pre-treated patients, novel therapies are needed to reverse the poor survival currently observed after transformation. The findings from our study suggest that inhibition of cell cycle progression is a potential therapeutic avenue to explore. In Burkitt lymphoma, another mature B-cell malignancy characterized by recurrent CCND3 mutations (38% of patients), inhibition of CDK4/6 with palbociclib results in cell cycle arrest and apoptosis in cell line models, although it has not been shown whether mutations in CCND3 enhance this effect.324 Palbociclib has shown promising activity in non-Hodgkin lymphoma in phase I and II studies,339,340 and this agent, or other cell cycle progression inhibitors deserve to be studied in the setting of TFL. Findings from our study will inform the building of models that aim at predicting early progression in FL. We showed that primary resistant FL often harbours resistance-conferring mutations at time of diagnosis. We also suggested that the m7-FLIPI may not be ideal at identifying the highest risk patients. Future studies will be required to determine how the knowledge of gene mutations status can be translated into improved patient care. 4.5 Ongoing work The capture sequencing cohort that is described in Chapter 3 represents only a fraction of all cases that were submitted for sequencing. We are currently awaiting the sequencing and analysis of 167 additional samples that were derived from FFPET. Increasing the number of 57 samples is critical in order to strengthen our conclusions and may lead to the discovery of further associations of gene mutations with transformation or progression. Based on our prior work on the m7-FLIPI, we discovered that the transcription factor FOXP1 is down-regulated in EZH2 mutated cases, and in preliminary experiments, we have shown that FOXP1 expression correlated with adverse outcome in a retrospective series of 107 patients treated with R-CVP at the BCCA, and the early progressers described in Chapter 3. We are currently validating the prognostic value of FOXP1 in several tissue microarrays from German collaborators. 4.6 Open questions and future directions 4.6.1 Order of mutations and relation to B-cell maturation The order by which mutations arise in clonal phylogenies in FL is at present insufficiently understood as only some mutations (e.g. CREBBP, KMT2D) have been shown to be commonly ancestral, in addition to invariably truncal t(14;18) translocations. We showed that TP53, B2M, ITPKB and CCND3 mutations are more frequently found in TFL than in FL, suggesting that they are late events, although a subset of FLs harbour these mutations in T1. It is further unknown whether the order of acquisition of mutations shapes the phenotype as it does in myeloproliferative neoplasms.341 Lastly, it is unclear whether genetic alterations arise solely within mature B cells, possibly confined to germinal centres, or whether at least early mutations are acquired during B-cell development within the bone marrow. The latter question can be addressed by flow-sorting bone marrow B-cells from FL patients into discrete populations corresponding to different maturation stages and assess for the presence of mutations by deep-sequencing. Single cell RNA and DNA sequencing could further allow to trace mutations back onto B-cell differentiation states. 4.6.2 The extent of intra-tumoural heterogeneity in B-cell lymphomas Whereas we clearly described ITH and its temporal relationship to tumour evolution, the spatial dimension of ITH is presently unknown. For example, it is unclear to what extent a biopsy from a single site may be representative of the genetic composition of the entire clonal phylogeny. This question could easily be addressed by sampling tumour from different sites (lymph node(s), extranodal sites such as bone marrow etc.), and subjecting these samples to large-scale, unbiased sequencing approaches. The assessment of circulating tumour DNA (ctDNA) is frequently proposed as a promising strategy to evaluate mutations found in spatially separated locations, and to monitor lymphoma progression over time.275,342,343 It remains to be determined whether the 58 evaluation of ctDNA provides a more sensitive approach to detect gene mutations that confer treatment resistance than the sequencing of DNA derived from single tumour sites. 4.6.3 Non-genetic drivers of intra-tumoural heterogeneity Our approach has been deliberately oblivious of alternative contributors to heterogeneity. The cancer stem cell model was first postulated in 1960 and subsequently experimentally validated in acute myeloid leukemia and several epithelial tumour types.344 It pre-supposes the existence of tumour-initiating cells that form a subset of the tumour and are uniquely capable of self-renewal, differentiation and re-population. The existence of cancer stem cells with distinct functional properties such as dormancy and treatment resistance defines a heterogeneous state. Stemness is generally defined based on functional properties and it is unclear whether cancer stem cells qualitatively or quantitatively differ in genetic aberrations when compared to bulk tumour. The search for tumour initiating cells has been elusive in mature lymphoid malignancies as functional repopulation assays, requiring serial transplantation analyses have not been reported to be feasible in lymphoma. FL in particular is not typically considered to be transplantable, which is a reflection of tumour cells lacking the capability to engraft and/or the lack of a suitable host.  Beyond the stem cell model, cancer populations can be heterogeneous due to epigenetic plasticity, although the exact cause underlying distinct functional states may be difficult to uncover. In a serial tumour transplantation model of human colon cancer, Kreso et al. demonstrated clonal stability, but variation in proliferation and persistence.345 Furthermore, non-Darwinian, Lamarckian induction has been reported to account for the induction of resistance to vincristine in a leukemia cell line model.346 If we hypothesize that the under- or overexpression of genes confers treatment resistance in these models, then shRNA of CRISPR-CAS9 library screens in the presence and absence of treatment might be suitable to uncover mechanisms of resistance and provide an opportunity for discovery beyond gene mutations. 4.6.4 Non-tumour cell intrinsic features In addition to the known roles that the tumour microenvironment plays in the pathogenesis of FL (see 1.2.4), we have shown that B2M mutations are enriched in TFL in comparison to preceding FL. We have also demonstrated that B2M mutations correlate with decreased infiltration of CD8+ T cells in the tumour microenvironment. Although we did not evaluate MHC class I expression in our cases, Challa-Malladi et al. showed, in de novo DLBCL, that absence of B2M expression correlates with lack of MHC class I expression, suggesting that 59 B2M mutations lead to escape from immune surveillance through deficient immune recognition.327 It is, however, unclear why lack of MHC class I does not lead to effective tumour cell elimination via NK cells, especially as CD58 mutations are rare in TFL. Our description of changes in the composition of the microenvironment during progression and transformation is only partial at this time and it is likely that the microenvironment undergoes changes in function and numbers of cellular components other than CD8+ T cells. On a broader level, we have to ask whether such changes are the result of or the cause of progression and transformation; i.e. whether the tumour microenvironment etiologically contributes to these events. To answer this question, insight will not be easily obtained from studying primary FL samples as the genetic composition of the tumour is a confounding factor. Rather, animal modelling would need to be utilized to vary components of the microenvironment in controlled experiments. 4.6.5 Mechanism of treatment resistance We discovered novel associations of gene mutations with early progression (KMT2D, BTG1 and XBP1), but it is unknown why these mutations predispose to treatment resistance. Mutations in XBP1 were most strongly associated with progression. XBP1 is an effector of the unfolded protein response that is activated in response to endoplasmic reticulum (ER) stress. In multiple myeloma, it was shown that bortezomib refractory cases have suppressed XBP1 signalling.328 Moreover, ectopic expression of a splice-site mutant conferred resistance to bortezomib in myeloma cell line models.328 We hypothesize that XBP1 splice site mutations cause treatment resistance in B-cell lymphomas that are treated with R-chemotherapy, bortezomib not being routinely administered in this setting. We further hypothesize that this effect is mediated by a decrease of ER stress.  We plan on assessing the presence of XBP1 mutations in larger B-cell lymphoma patient cohorts and on correlating mutations with treatment resistance. The impact of observed splice site mutations on splicing will be confirmed experimentally. We will subsequently generate cell line models to assess whether XBP1 mutations contribute to treatment resistance and to determine whether certain drugs circumvent resistance. Lastly, we will delineate treatment resistance mechanisms by correlating mutants with ER stress. 60 4.7 Final conclusion Herein, we described discrete molecular subtypes in TFL that mirror cell-of-origin classification in de novo DLBCL. We demonstrated that the GCB and ABC molecular subtypes in TFL originate from distinct molecular subsets of preceding FL, testifying to molecular heterogeneity in underlying FL that translates into differential pathways to transformation. Furthermore, we showed that IRF4 expression in the primary FL sample is a risk factor for early transformation. In addition, we showed that clonal trajectories of early progression and transformation are fundamentally distinct. In patients that experienced early progression after R-chemotherapy, resistance correlated with the selective outgrowth of clones that were present at the time of initial diagnosis in a majority of patients, suggesting that the molecular determinants of treatment resistance were already present at the time of diagnosis. Transformation, on the other hand, arose in the majority of patients (12 out of 15) by clade-specific outgrowth of clones that rose to complete dominance in T2 and were seemingly absent in T1. A more refined analysis showed that, at least in 2 patients, these expanded T2-dominant clones were present in T1 samples at a small subclonal level, suggesting that transformation occurred in these cases by selection rather than by de novo acquisition of mutations. We completed our analysis with the description of novel gene mutations that are associated with transformation (ITPKB, CCND3) and early progression (KMT2D, BTG1 and XBP1). Our work is of translational relevance as it links molecular heterogeneity, both at an intra-patient and inter-patient level, to scenarios that are associated with poor outcome, namely transformation and early progression. Given that patient outcomes in FL are highly diverse, the identification of robust features that predict poor outcome, and a refinement of the genetic underpinnings underlying tumour progression are critically needed to improve patient outcomes. Our discovery of gene mutations that are associated with early progression represents a step forward, and suggests that genetic testing of FL tumour samples should be considered when treatment is indicated. Our findings should, however, ideally be validated in larger patient cohorts, and in patients that were treated with other regimens (such as R-bendamustine, or combinations of novel agents), to determine whether the identified genetic markers of treatment resistance are specifically tied to R-CVP and R maintenance that our patients were treated with, or whether they identify universal mechanisms of resistance to therapy. Genetic testing should also be considered 61 at the time of transformation as determination of cell-of-origin in TFL identifies molecular subtypes that may respond differentially to inhibitors of BCR or NF-κB signalling. The continuing increase in the refinement by which tumour evolution is understood and the concomitant development of cancer therapies that specifically target altered pathways will eventually lead to the implementation of precision medicine. The ultimate goal is to improve patient outcomes while at the same time decreasing treatment toxicity. Whether or not a cure will one day exist for FL patients, no patient should die from this disease. 62 Tables Table 1. Clinical characteristics of 2820 newly diagnosed FL patients  Total number: n = 2820 Age Median Interquartile range  60 years 21 years Gender Female Male  1,366 (48.4%) 1,454 (51.6%) ECOG 0-1 2-4 NA  2,248 (87.2%) 331 (12.8%) 241 Stage Limited Advanced NA  697 (26.9%) 1,893 (73.1%) 230 LDH ≤ ULN > ULN NA  1,890 (82.1%) 411 (17.9%) 519 FLIPI Low/intermediate High NA  314 (64.6%) 172 (35.4%) 2,334 10 year overall survival Diagnosis < 2000 (n=1,016) Diagnosis ≥ 2000 (n=1,783)  54.2 % 59.9 % 10 year disease specific survival Diagnosis < 2000 (n=1,016) Diagnosis ≥ 2000 (n=1,783)  61.0 % 67.8 %   63 Table 2. Clinical characteristics of early and late transformers  Transformed < 5y  Not transformed > 5y  P*  No. %  No. %   Age ≤ 60 > 60  20 20  50% 50%   49 37  57% 43%    0.56 Gender Female Male  17 23  43% 58%   42 44  49% 51%    0.57 B symptoms absent present NA  35 4 1  90% 10%   73 11 2  87% 13%    0.77 ECOG 0-1 2-4 NA  36 2 2  95% 5%   78 3 5  96% 4%    0.65 Stage I-II III-IV NA  11 28 1  28% 72%   20 66 0  23% 77%    0.66 Extranodal sites 0-1 > 1  38 2  95% 5%   77 9  90% 10%    0.50 Nodal areas ≤ 4 > 4 NA  15 12 13  56% 44%   35 40 11  47% 53%    0.50 Tumor mass < 7 cm ≥ 7 cm NA  26 6 8  81% 19%   60 22 4  73% 27%    0.47 LDH ≤ ULN > ULN NA  24 7 9  77% 23%   76 4 6  95% 5%    0.01 Hemoglobin < 12 g/dL ≥ 12 g/dL NA  5 27 8  16% 84%   4 80 2  5% 95%    0.11 FLIPI Low/intermediate High NA  21 9 10  70% 30%   59 17 10  78% 22%    0.46 Primary treatment for stage III/IV Observation or local treatment Systemic treatment  15 13  54% 46%   26 40  39% 61%    0.26 *, NA cases were excluded for contingency analyses.   64 Table 3. Weighted Cox regression model for time to transformation   Univariate  Multivariate   HR 95% CI P  HR 95% CI P Grade 3A  5.3 2.4-11.8 <0.001  2.5 0.7-8.3 0.14 BCL2 translocation  0.4 0.1-1.1 0.07  1.0 0.3-4.1 0.96 BCL6 translocation  2.6 1.0-6.4 0.05  1.5 0.3-6.3 0.60 CD10 IHC  0.2 0.1-0.6 0.004  1.0 0.2-4.5 1.00 IRF4 IHC  11.5 3.1-42.4 <0.001  13.3 3.7-48.4 <0.001   65 Table 4. Clinical characteristics of TFL cohort  All TFL patients (n=155)  No. % Year of FL diagnosis < 2003 ≥ 2003  104 51  67% 33% Year of TFL diagnosis <2003 ≥ 2003  59 96  38% 62% Antecedent FL grade grade 1 or 2 grade 3A  135 20  87% 13% Time to transformation (years) ≤ 5 5-10 >10  98 39 18  63% 25% 12% Age at TFL diagnosis (years) ≤ 60 > 60  65 90  42% 58% First line treatment for FL single-agent chemo multi-agent chemo* multi-agent chemo# and rituximab radiation or surgery observation NA  34 22 14 21 52 12  24% 15% 10% 15% 36% - Initial treatment after transformation multi-agent chemo§ multi-agent chemo¶ + rituximab other NA  46 82 18 9  32% 56% 12% - Hematopoietic stem cell transplant autologous allogeneic none NA  12 3 133 7  8% 2% 90% - Rituximab before transformation no yes NA  120 32 3  79% 21% - Rituximab after transformation no yes NA  61 91 3  40% 60% - *, 10 patients out of 22 were treated with anthracycline-containing regimens; #, 3 patients out of 14 were treated with cyclophosphamide-doxorubicin-vincristine-prednisone and rituximab (CHOP-R) and 11 out of 14 with cyclophosphamide-vincristine-prednisone and rituximab (CVP-R); §, 41 patients out of 46 were treated with anthracycline-containing regimens; ¶, 77 patients out of 82 were treated with CHOP-R.   66 Table 5. Mutations in CARD11, CD79B and MYD88 Sample Gene Chromosomal position Nucleotide change Variant allele frequency (%) Predicted amino acid change FL1012T2 CD79B chr17:62006798 T>G 31 Y196S* FL1012T2 MYD88 chr3:38181433 G>T 17 S149I# FL1101T2 MYD88 chr3:38182259 T>C 59 M232T# FL1101T2 CARD11 chr7:2984162 C>T 76 G123D§ FL1161T2 CD79B chr17:62006680 A>T 50 L199Q* FL1161T2 MYD88 chr3:38182641 T>C 42 L265P# FL1227T2 MYD88 chr3:38181408 C>T 23 P141S# FL1227T2 MYD88 chr3:38181417 G>T 24 V144L# FL1242T2 CD79B chr17:62006680 A>G 75 L199P* *, refers to NCBI Protein ID NP_000617.1; #, refers to NCBI Protein ID NP_002459.2; §, refers to NCBI Protein ID NP_115791.3.   67 Table 6. Clinical characteristics of early and late progressers   Early progresser (n = 30) Late progresser (n = 67)    No. % No. % P Age ≤ 60 years > 60 years 19 11 63% 37% 39 28 58% 42%  0.662 Gender Male Female 16 14 53% 47% 34 33 51% 49%  0.830 B symptoms Absent Present NA 22 7 1 76% 24%  57 8 2 88% 12%   0.221  ECOG 0-1 2-4 NA 23 6 1 79% 21%  61 1 5 98% 2%   0.004  Stage I-II III-IV 3 27 10% 90% 16 51 24% 76%  0.166 Extranodal sites 0-1 > 1 27 3 90% 10% 61 6 91% 9%  1.000 Nodal areas ≤ 4 > 4 NA 7 21 2 25% 75%  27 32 8 46% 54%   0.099  Tumor mass < 7 cm ≥ 7 cm 13 16 45% 55% 47 16 75% 25%  0.009 LDH ≤ ULN > ULN NA 17 12 1 59% 41%  58 2 7 97% 3%   < 0.001  Hemoglobin < 12 g/dL ≥ 12 g/dL NA 7 23 0 23% 77%  3 62 2 5% 95%   0.010  FLIPI Low/intermediate High NA 12 16 2 43% 57%  46 13 8 78% 22%   0.002  FL grade 1-2 3A 21 9 70% 30% 62 5 93% 7%  0.009 Treatment by ITT R-CVP & R maintenance Other R-chemo Observed 23 7 0 77% 23% 0 28 7 32 42% 10% 48%   NA Progression during R-chemotherapy R maintenance 20 10 67% 33% NA NA NA NA  NA   68 Figures  Figure 1. Genes mutated in >5% of FL cases The data is from 140 cases, retrospectively selected from the BC Cancer Agency tissue repository. All biopsies were taken before initiation of systemic treatment. Sequencing of a custom panel of 74 genes was performed at the Dana Farber Cancer Institute.     69 Figure 2. The tumour microenvironment in follicular lymphoma A model of the microenvironment in FL. Tumour cells receive survival cues from a variety of cells that include follicular T helper cells, follicular dendritic cells and follicular reticular cells. Signals promoting growth and survival include for example IL-4, IL-21, CXCL12 and CXCL13. FL cells reduce the anti-tumour effect of T-cell subsets by inducing immunologic synapse dysfunction and by promoting the conversion of T helper cells into T regulatory cells. Macrophages are highly diverse and can exhibit a variety of pro-tumoral or anti-tumoral phenotypes that are partially dependent on the interaction between macrophages and treatment.     70 Figure 3. Overlap between study cohorts Overlap between patient cohorts described in Chapter 2. The numbers refer to patients and not samples.     71 Figure 4. Association of pathological characteristics with transformation Prevalence of grade 3A, translocations involving the BCL2 or BCL6 loci, and staining by IHC for CD10, BCL6 and IRF4 in tissue specimens of patients who presented with transformation within 5 years after diagnosis versus those who presented with neither transformation nor progression for at least 5 years after diagnosis.     72 Figure 5. BCL2, BCL6 and MYC translocation in FL and TFL A. Prevalence of BCL2, BCL6 and MYC translocation in FL and TFL. B. Prevalence of BCL2, BCL6 and MYC translocation in TFL by histology.     73 Figure 6. CD10, BCL6 and IRF4 expression in FL and TFL A. Prevalence of CD10, BCL6 and IRF4 expression in FL and TFL by IHC. B. Prevalence of CD10, BCL6 and IRF4 expression in TFL by histology.     74 Figure 7. Survival correlates from time of transformation A. Survival by morphology of TFL in 155 patients. B. Survival by double-hit translocation status in all 114 TFL cases in which BCL2 and MYC translocation status could be ascertained. C. Survival by double-hit translocation status in 90 TFL cases with DLBCL morphology.    75 Figure 8. Lymph2Cx assay in 107 TFL cases with DLBCL morphology A. Heatmap showing the relative expression of 20 genes (8 genes that are over-expressed in ABC-DLBCL, 5 housekeeping genes and 7 genes that are overexpressed in GCB-DLBCL). B. Pie charts showing the relative proportions of ABC and GCB large cell lymphoma in TFL and de novo DLBCL, respectively. C. Kernel density plot of the distribution of the linear predictor score in molecular subtypes of TFL and de novo DLBCL, respectively.     76 Figure 9. Survival and cytogenetic correlates of molecular subtypes in TFL A. Prevalence of BCL2, BCL6 and MYC translocation in TFL by COO subtype. B. Survival from time of transformation by COO subtype in TFL.     77 Figure 10. Association of pathological findings in FL and COO of TFL The figure shows a forest plot of odds ratios (OR, circles) and corresponding 95% confidence intervals (horizontal lines).     78 Figure 11. Overview of study cohort used for Chapter 3 The figure shows the number of patients and samples in both the whole genome sequencing cohort and the capture sequencing cohort, as well as the number of patients and samples in clinically defined groups.     205 patients; 252 T1 or T2 samples TFL Cohort: N = 116 patients Clinical Extremes Cohort: N = 97 patients Early progression: N = 30 T1 samples Overlap  8 patients; 8 T1 samples Samples: n = 97 Late progression: N = 67 T1 samples T1 (FL) and T2 (TFL) samples available: N = 47 T1-T2 sample-pairs Only T2 (TFL) sample   available: N = 36 T2 samples Samples: n = 163 Only T1 (FL) sample   available: N = 33 T1 samples Capture sequencing cohort Whole Genome Sequencing cohort 41 patients; 62 T1 or T2 samples TFL cohort: N = 15 patients PFL cohort: N = 12 patients Transformed NPFL cohort: N = 20 patients T1 (FL) and T2 (TFL) samples available: N = 15 T1-T2 sample-pairs Samples: n = 30 Progressed Non-progressed T1 (FL) and T2 (PFL) samples available: N = 6 T1-T2 sample-pairs Samples: n = 12 T1 (FL) samples available: N = 20 T1 samples Samples: n = 20 Overlap  38 patients; 55 samples 79 Figure 12. Sample overview and timeline of WGS cohort    80 Figure 13. High-level WGS analysis overview Number of genetic aberrations by type of alteration and by patient category.     FL3001FL3002FL3003FL3004FL3005FL3006FL3007FL3008FL3009FL3010FL3011FL3012FL3013FL3014FL3015FL3016FL3017FL3018FL3019FL3020FL1001FL2001FL2002FL2005FL2006FL2007FL2008FL1004FL1005FL1006FL1007FL1008FL1009FL1012FL1013FL1014FL1016FL1017FL1018FL1019FL1020Somatic LOHGrade 1-2Grade 3ATumour Content Cell-of-Origin IHCFL GradeNegativePositiveTransformed (TFL) Progressed (PFL) Non-progresser (NPFL)81 Figure 14. Genetic alterations by timepoint and by clinical category Panel a: Paired plots of number of genetic alterations in T1 and T2. Panel b: Distribution of number of alterations by patient category.    P = 0.482 P = 0.056P = 0.280P = 0.650P = 0.336 P < 0.001P = 0.701 P = 0.018P = 0.026 P = 0.028PFL PFLNPFLa b T1 T2P < 0.001P = 0.001P < 0.001P = 0.015P = 0.008T182 Figure 15. Clonal phylogenies of TFL patients For each patient, the leftmost plot shows mutational cellular prevalence for all deep-sequenced SNVs. T1 is shown on the x-axis and T2 is shown on the y-axis. The second graph for each patient plots the mean cluster cellular prevalence. The size of each circle is proportional to the number of mutations it contains. The third plot for each patient shows clonal phylogenies, inferred by Citup, and the last plot for each patient is a stacked barplot representing clonal prevalence in both timepoints.     Mutational CellularPrevalenceMean ClusterCellular PrevalenceMean ClusterCellular PrevalenceMutational CellularPrevalenceClonal Phylogeny Clonal PhylogenyClonal Prevalence Clonal PrevalenceT1 T2T1 T212 3413 6245124312354123456712 3412 3 41 23412 4 53 7612 3437 114691012 3 45CREEBPCARD11BCL6B2MIRF4CREEBPEZH2CREEBPCREEBPCARD11MLL2CREEBPABCB8EZH2TNFRSF14B2MBTG1CREBBPEZH2EP300GNA13MEF2BPIM1SOCS1TNFRSF14B2MBCL6HIST1H2BCHIST1H4DTNFRSF14BCL6HIST1H1BEZH2EP400HIST1H2BGHIST1H1EHIST1HIDHIST1H2BDSGK1EZH2FOXO1PRRX21235 7461 23123MLL2BCL6TBX20PRDM9PTPN1CREEBPBCL2CD58CCND3 CD58B2MCCND3BCL6FOXO1STAT6FASTP53SETD2CCDC108*HIST1H1BCCND3CD79BCIITA* Subclonal DeletionTime to transformation: 0.78yTreatment between samples:- ObservationTime to transformation: 5.05yTreatment between samples:- ObservationTime to transformation: 2.84yTreatment between samples:- RadiationTime to transformation: 0.42yTreatment between samples:- RituximabTime to transformation: 3.00yTreatment between samples:- R-CVP + R maintenanceTime to transformation: 0.01yTreatment between samples:- NoneTime to transformation: 0.85yTreatment between samples:- R-CVPTime to transformation: 0.70yTreatment between samples:- observationTime to transformation: 0.39yTreatment between samples:- ObservationTime to transformation: 0.99yTreatment between samples:- Cyclophosphamide + Prednisone- R-CHOPTime to transformation: 14.57yTreatment between samples:- ObservationTime to transformation: 7.65yTreatment between samples:- ObservationTime to transformation: 1.35yTreatment between samples:- Observation- R-CVPTime to transformation: 2.56yTreatment between samples:Cyclophosphamide + PrednisoneTime to transformation: 1.45yTreatment between samples:- Observation- Chlorambucil + Prednisone83 Figure 16. Clonal phylogenies of PFL patients Shown are analogous plots to Figure 15.    Mean ClusterCellular PrevalenceMutational CellularPrevalence Clone Phylogeny Clonal PrevalenceT1 T225314 1510 161112 1312 312354612 4 53123967108123 4 5EZH2HIST1H1EPAX5PIM1RCOR1CREBBPHIST1H1EMLL2CREBBPCREBBPARID1BMLL2TNFRSF14ARID1AFASTime to Progression: 1.11yTreatment between samples:- R-CVP + R MaintenanceTime to Progression: 0.90yTreatment between samples:- R-CVP + BortezomibTime to Progression: 2.89yTreatment between samples:- R-CVP + R MaintenanceTime to Progression: 0.87yTreatment between samples:- R-CVP + BortezomibTime to Progression: 1.36yTreatment between samples:- R-CVP + R maintenanceTime to Progression: 4.02yTreatment between samples:- Observation- Rituximab84 Figure 17. Small subclones in T1 samples Variant allele frequencies of T2 dominant clone associated mutations in T1. Shown are only cases in which transformation or progression occurred by clade-specific expansion of a clone that rose to complete dominance in T2.     ●●●●● ●●●●●●● ●●● ●●●●●●●●●●●0.0000.0050.0100.0150.020FL1007FL1012FL1013FL1017FL1014FL1004FL1001FL1006FL1008FL1018FL1005FL1019FL2001VAF85 Figure 18. Validation of small subclonal mutations in T1 by ddPCR Shown are 8 mutations (panels a to h) in 3 cases (FL1012, FL1004, FL1019 and FL2001) in which PyClone (leftmost plots) suggests that the expanded T2-dominant mutational clusters are present at near zero prevalence in T1. No evidence of read support, when compared to background, is found for T2-associated mutations in T1 for cases FL1012 and FL1004, in contrast to cases FL1019 and FL2001. These results are confirmed by digital droplet PCR (plots on the right).     86 Figure 19. Recurrent mutations in all T1 (FL) samples (n = 172) Panel a: proportion of all T1 samples harbouring mutations (n = 172). Shown are only genes that are mutated in at least 5% of all samples. Panel b: Results from MutSigCV; the dashed line represents the cut-off for significance (q-value of 0.1).     0204060CREBBPKMT2DBCL2TNFRSF14EZH2MEF2BSTAT6SOCS1FOXO1CARD11BCL7AARID1AEP300IRF8FAT4DTX1KMT2CGNA13TP53BCRARID1BPIM1HVCN1HIST1H1EEBF1ATP6V1B2SGK1RRAGCEEF1A1DNAH9POU2AF1HLA−DMBTCF3SMARCA4MEF2C B2MACTBITPKBHIST1H1BCIITATNFAIP3MYOM2FASMKI67KLHL6CTSSBTG1BCL6STAT3NOTCH1ZFP36L1S1PR2MYD88HIST1H1CGNAI2CD83ATP6AP1P2RY8IKZF3CD79BCCND3XBP1UPF1UNC5CHIST1H2AMEVI2ACHD8BTG2NLRC5IRF4EBF3CD58ARHGEF1NOTCH2MYCLRRC7RFX5CCL23Percentage of samples having mutationType of mutationNon-truncatingTruncating0510CREBBPKMT2DBCL2TNFRSF14EZH2MEF2BSTAT6SOCS1FOXO1CARD11BCL7AARID1AEP300IRF8FAT4DTX1KMT2CGNA13TP53BCRARID1BPIM1HVCN1HIST1H1EEBF1ATP6V1B2SGK1RRAGCEEF1A1DNAH9POU2AF1HLA−DMBTCF3SMARCA4MEF2C B2MACTBITPKBHIST1H1BCIITATNFAIP3MYOM2FASMKI67KLHL6CTSSBTG1BCL6STAT3NOTCH1ZFP36L1S1PR2MYD88HIST1H1CGNAI2CD83ATP6AP1P2RY8IKZF3CD79BCCND3XBP1UPF1UNC5CHIST1H2AMEVI2ACHD8BTG2NLRC5IRF4EBF3CD58ARHGEF1NOTCH2MYCLRRC7RFX5CCL23MutSigCV −log10(q)ab87 Figure 20. Results from targeted sequencing in TFL samples Panels a,b,c: Number of mutated genes plotted by timepoint, considering all genes (a), only genes that are significantly mutated in either T1 or T2 (b) or only genes that are significantly mutated in neither T1 nor T2 (c). Panel d: Correlation of proportion of mutations shared between T1 and T2, and time to transformation. Panel e: Proportion of samples harbouring mutations, separated by whether genes are differentially mutated between T1 and T2 samples (P < 0.05). Shown are only genes that are mutated in at least 5% of samples in either timepoint. Panel f: Results from MutSigCV; the dashed line represents the cut-off for significance (q-value of 0.1). Panel g: Classification of mutations into T1-specific, T2-specific or shared categories. Panel h: Time to transformation based on whether samples harbour mutations in either TP53, B2M, ITPKB or CCND3 or mutations in neither of these genes. Panel i: Proportion of mutated samples by T2 histology (composite versus DLBCL). Shown are only genes that are significantly associated with either histology (P < 0.05). Panel j: Proportion of samples harbouring mutations by cell-of-origin (COO). Shown are only genes that are differentially mutated (P < 0.05).    88 Figure 21. Mutations in areas of SHM for FL and TFL Panel a: Total number of mutations found in 20 genes, by timepoint. Panel b: Total number of mutations, by gene and by timepoint.     89 Figure 22. Proportion of shared and T1 or T2-specific mutations Included in this analysis are all coding mutations called in the 86 gene capture sequencing panel, and all coding and non-coding mutations called in regions of somatic hypermutation in 20 genes that are known targets of somatic hypermutation.     0.000.250.500.751.00FL1019FL1134FL1235FL1236FL1111FL1248FL1255FL1256FL1007FL1012FL1138FL1198FL1132FL1008FL1260FL1122FL1261FL1115FL1116FL1219FL1262FL1141FL1006FL1259FL1249FL1018FL1230FL1257FL1103FL1114FL1016FL1218FL1231FL1258FL1217FL1196FL1013FL1001FL1010FL1254FL1250FL1004FL1003FL1194FL1202FL1009FL1161Proportion of mutationsT2SHAREDT190 Figure 23. B2M mutations and CD8+ T cells Panel a: Percentage positive pixels for CD8 using Aperio automated imaging, by timepoint. Panel b: Percentage positive pixels for CD8, by B2M mutational status and by timepoint.     ●P = 0.009705101520T1 T2CD8 percent positive pixelsSAMPLE.TYPET1T2CD8 IHC by timepoint0.8211●●●0.0345T1 T205101520WT MUT WT MUTCD8 percent positive pixelsCD8 IHC by B2M mutational statusab91 Figure 24. Results from targeted sequencing in early and late progressers Panels a,b,c: Number of mutated genes plotted by category, considering all genes (a), only genes that are significantly mutated in either the early or the late progressers (b) or only genes that are significantly mutated in neither the early nor the late progressers (c). Panel d,e: progression-free and overall survival for early and late progressers. Panel f: proportion of samples harbouring mutations, separated by whether genes are differentially mutated between early and late progressers using Fisher's exact test (P < 0.05). Shown are only genes that are mutated in at least 5% of samples in either outcome category. Panel g: Results from MutSigCV; the dashed line represents the cut-off for significance (q-value of 0.1). Panel h: Forrest plot showing odds ratio and 95% confidence interval for association of FLIPI, m7-FLIPI and gene mutations with outcome categories.     ●●P = 0.038905101520EARLY LATENumber of mutations/patientAll mutations●●●P = 0.05205101520EARLY LATENumber of mutations/patientPutative driver mutationsP = 0.122705101520EARLY LATENumber of mutations/patientPutative passenger mutationsEARLY Not EARLY−enriched020406080KMT2DBTG1TP53XBP1CREBBPBCL2TNFRSF14BCL7AFOXO1SOCS1EZH2KMT2CSTAT6ARID1BATP6V1B2DTX1EP300MEF2BPIM1SGK1BCRCIITAEBF1MKI67ARID1AB2MDNAH9FASGNA13IKZF3IRF8RRAGCTCF3TNFAIP3ATP6AP1BCL6BTG2CARD11CTSSEEF1A1EVI2AFAT4HIST1H1CHIST1H1EITPKBMYD88POU2AF1S1PR2SMARCA4ZFP36L1ACTBHLA−DMBHVCN1MYOM2MEF2CPercent mutatedGroupEARLYLATEEARLY Not EARLY−enriched0510KMT2DBTG1TP53XBP1CREBBPBCL2TNFRSF14BCL7AFOXO1SOCS1EZH2KMT2CSTAT6ARID1BATP6V1B2DTX1EP300MEF2BPIM1SGK1BCRCIITAEBF1MKI67ARID1AB2MDNAH9FASGNA13IKZF3IRF8RRAGCTCF3TNFAIP3ATP6AP1BCL6BTG2CARD11CTSSEEF1A1EVI2AFAT4HIST1H1CHIST1H1EITPKBMYD88POU2AF1S1PR2SMARCA4ZFP36L1ACTBHLA−DMBHVCN1MYOM2MEF2CMutSigCV −log10(q)GroupEARLYLATE0.000.250.500.751.000.0 2.5 5.0 7.5 10.0Time (years)Probability progression−freeGroupEARLYLATEProgression−free survival0.000.250.500.751.000.0 2.5 5.0 7.5 10.0Time (years)Probability survivingGroupEARLYLATEOverall survival●●●●●●KMT2DBTG1TP53XBP1FLIPI HIGHm7−FLIPI HIGH5 10 15Odds ratio for early progressionForrest plota b c defg h92 Figure 25. Mutations in areas of SHM for early and late progressers Panel a: Total number of mutations found in 20 genes, by outcome category. Panel b: Total number of mutations, by gene and by outcome category.   93 References 1. Von Behring E, Kitasato S. Ueber das Zutandekommen der Diphtherie-Immunitat und der Tetanus-Immunitat bei Tieren. Dtsch. Medizinsche Wochenschrift. 1890;16:1113–1114.  2. Fagraeus A. The plasma cellular reaction and its relation to the formation of antibodies in vitro. J. Immunol. 1948;58(1):1–13.  3. Bruton OC. Agammaglobulinemia. Pediatrics. 1952;9(6):722–8.  4. Good RA, Varco RL. A clinical and experimental study of agammaglobulinemia. J. Lancet. 1955;75(6):245–71.  5. Cooper MD, Peterson RD, Good RA. Delineation of the thymic and bursal lymphoid systems in the chicken. Nature. 1965;205:143–6.  6. Cooper MD, Raymond DA, Peterson RD, South MA, Good R a. The functions of the thymus system and the bursa system in the chicken. J. Exp. Med. 1966;123(1):75–102.  7. Peschon JJ, Morrissey PJ, Grabstein KH, et al. Early lymphocyte expansion is severely impaired in interleukin 7 receptor-deficient mice. J. Exp. Med. 1994;180(5):1955–60.  8. Miller JP, Izon D, DeMuth W, et al. The earliest step in B lineage differentiation from common lymphoid progenitors is critically dependent upon interleukin 7. J. Exp. Med. 2002;196(5):705–711.  9. Nutt SL, Heavey B, Rolink AG, Busslinger M. Commitment to the B-lymphoid lineage depends on the transcription factor Pax5. Nature. 1999;401(6753):556–562.  10. Bain G, Maandag EC, Izon DJ, et al. E2A proteins are required for proper B cell development and initiation of immunoglobulin gene rearrangements. Cell. 1994;79(5):885–892.  11. Zhuang Y, Soriano P, Weintraub H. The helix-loop-helix gene E2A is required for B cell formation. Cell. 1994;79(5):875–884.  12. Lin H, Grosschedl R. Failure of B-cell differentiation in mice lacking the transcription factor EBF. Nature. 1995;376(6537):263–267.  13. Urbánek P, Wang ZQ, Fetka I, Wagner EF, Busslinger M. Complete block of early B cell 94 differentiation and altered patterning of the posterior midbrain in mice lacking Pax5/BSAP. Cell. 1994;79(5):901–912.  14. Karasuyama H, Kudo A, Melchers F. The proteins encoded by the VpreB and lambda 5 pre-B cell-specific genes can associate with each other and with mu heavy chain. J. Exp. Med. 1990;172(3):969–72.  15. Tsubata T, Reth M. The products of pre-B cell-specific genes (lambda 5 and VpreB) and the immunoglobulin mu chain form a complex that is transported onto the cell surface. J. Exp. Med. 1990;172(3):973–6.  16. Hozumi N, Tonegawa S. Evidence for somatic rearrangement of immunoglobulin genes coding for variable and constant regions. Proc. Natl. Acad. Sci. U. S. A. 1976;73(10):3628–3632.  17. Brack C, Hirama M, Lenhard-Schuller R, Tonegawa S. A complete immunoglobulin gene is created by somatic recombination. Cell. 1978;15(1):1–14.  18. Alt FW, Yancopoulos GD, Blackwell TK, et al. Ordered rearrangement of immunoglobulin heavy chain variable region segments. Embo J. 1984;3(6):1209–19.  19. Schatz DG, Oettinger MA, Baltimore D. The V(D)J recombination activating gene, RAG-1. Cell. 1989;59(6):1035–1048.  20. Oettinger MA, Schatz DG, Gorka C, Baltimore D. RAG-1 and RAG-2, adjacent genes that synergistically activate V(D)J recombination. Science. 1990;248(4962):1517–23.  21. Shimizu T, Yamagishi H. Biased reading frames of pre-existing DH--JH coding joints and preferential nucleotide insertions at VH--DJH signal joints of excision products of immunoglobulin heavy chain gene rearrangements. EMBO J. 1992;11(13):4869–75.  22. Okada T, Cyster JG. CC chemokine receptor 7 contributes to Gi-dependent T cell motility in the lymph node. J. Immunol. 2007;178(5):2973–8.  23. Schwickert TA, Victora GD, Fooksman DR, et al. A dynamic T cell-limited checkpoint regulates affinity-dependent B cell entry into the germinal center. J. Exp. Med. 2011;208(6):1243–52.  95 24. Allen CDC, Ansel KM, Low C, et al. Germinal center dark and light zone organization is mediated by CXCR4 and CXCR5. Nat. Immunol. 2004;5(9):943–952.  25. Muramatsu M, Kinoshita K, Fagarasan S, et al. Class switch recombination and hypermutation require activation-induced cytidine deaminase (AID), a potential RNA editing enzyme. Cell. 2000;102(5):553–63.  26. Victora GD, Schwickert TA, Fooksman DR, et al. Germinal center dynamics revealed by multiphoton microscopy with a photoactivatable fluorescent reporter. Cell. 2010;143(4):592–605.  27. Gitlin AD, Shulman Z, Nussenzweig MC. Clonal selection in the germinal centre by regulated proliferation and hypermutation. Nature. 2014;509(7502):637–40.  28. Kitano M, Moriyama S, Ando Y, et al. Bcl6 protein expression shapes pre-germinal center B cell dynamics and follicular helper T cell heterogeneity. Immunity. 2011;34(6):961–72.  29. Basso K, Dalla-Favera R. Roles of BCL6 in normal and transformed germinal center B cells. Immunol. Rev. 2012;247(1):172–183.  30. Phan RT, Dalla-Favera R. The BCL6 proto-oncogene suppresses p53 expression in germinal-centre B cells. Nature. 2004;432(7017):635–9.  31. Phan RT, Saito M, Basso K, Niu H, Dalla-Favera R. BCL6 interacts with the transcription factor Miz-1 to suppress the cyclin-dependent kinase inhibitor p21 and cell cycle arrest in germinal center B cells. Nat. Immunol. 2005;6(10):1054–1060.  32. Ranuncolo SM, Polo JM, Dierov J, et al. Bcl-6 mediates the germinal center B cell phenotype and lymphomagenesis through transcriptional repression of the DNA-damage sensor ATR. Nat. Immunol. 2007;8(7):705–14.  33. Ranuncolo SM, Polo JM, Melnick A. BCL6 represses CHEK1 and suppresses DNA damage pathways in normal and malignant B-cells. Blood Cells. Mol. Dis. 2009;41(1):95–9.  34. Klein U, Tu Y, Stolovitzky GA, et al. Transcriptional analysis of the B cell germinal center reaction. Proc. Natl. Acad. Sci. U. S. A. 2003;100(5):2639–44.  35. Calado DP, Zhang B, Srinivasan L, et al. Constitutive canonical NF-κB activation 96 cooperates with disruption of BLIMP1 in the pathogenesis of activated B cell-like diffuse large cell lymphoma. Cancer Cell. 2010;18(6):580–9.  36. Dominguez-Sola D, Victora GD, Ying CY, et al. The proto-oncogene MYC is required for selection in the germinal center and cyclic reentry. Nat. Immunol. 2012;13(11):1083–91.  37. Béguelin W, Popovic R, Teater M, et al. EZH2 is required for germinal center formation and somatic EZH2 mutations promote lymphoid transformation. Cancer Cell. 2013;23(5):677–92.  38. Hanahan D, Weinberg RA. Hallmarks of cancer: the next generation. Cell. 2011;144:646–674.  39. Pasqualucci L, Neumeister P, Goossens T, et al. Hypermutation of multiple proto-oncogenes in B-cell diffuse large-cell lymphomas. Nature. 2001;412(6844):341–346.  40. Ramiro AR, Jankovic M, Eisenreich T, et al. AID is required for c-myc/IgH chromosome translocations in vivo. Cell. 2004;118(4):431–438.  41. Pasqualucci L, Bhagat G, Jankovic M, et al. AID is required for germinal center-derived lymphomagenesis. Nat. Genet. 2008;40(1):108–12.  42. Robbiani DF, Deroubaix S, Feldhahn N, et al. Plasmodium Infection Promotes Genomic Instability and AID-Dependent B Cell Lymphoma. Cell. 2015;162(4):727–37.  43. Küppers R. Mechanisms of B-cell lymphoma pathogenesis. Nat. Rev. Cancer. 2005;5(4):251–62.  44. Jares P, Colomer D, Campo E. Molecular pathogenesis of mantle cell lymphoma. J. Clin. Invest. 2012;122(10):3416–23.  45. Camacho FI, Algara P, Rodríguez A, et al. Molecular heterogeneity in MCL defined by the use of specific VH genes and the frequency of somatic mutations. Blood. 2003;101(10):4042–6.  46. Walsh SH, Thorsélius M, Johnson A, et al. Mutated VH genes and preferential VH3-21 use define new subsets of mantle cell lymphoma. Blood. 2003;101(10):4047–54.  47. Kienle D, Kröber A, Katzenberger T, et al. VH mutation status and VDJ rearrangement 97 structure in mantle cell lymphoma: Correlation with genomic aberrations, clinical characteristics, and outcome. Blood. 2003;102(8):3003–3009.  48. Alizadeh AA, Eisen MB, Davis RE, et al. Distinct types of diffuse large B-cell lymphoma identified by gene expression profiling. Nature. 2000;403(6769):503–11.  49. Rosenwald A, Wright G, Chan WC, et al. The use of molecular profiling to predict survival after chemotherapy for diffuse large-B-cell lymphoma. N. Engl. J. Med. 2002;346(25):1937–47.  50. Lenz G, Wright G, Dave SS, et al. Stromal gene signatures in large-B-cell lymphomas. N. Engl. J. Med. 2008;359(22):2313–23.  51. Dybkaer K, Bogsted M, Falgreen S, et al. Diffuse Large B-Cell Lymphoma Classification System That Associates Normal B-Cell Subset Phenotypes With Prognosis. J. Clin. Oncol. 2015;  52. Monti S, Savage KJ, Kutok JL, et al. Molecular profiling of diffuse large B-cell lymphoma identifies robust subtypes including one characterized by host inflammatory response. Blood. 2005;105(5):1851–1861.  53. Leich E, Salaverria I, Bea S, et al. Follicular lymphomas with and without translocation t(14;18) differ in gene expression profiles and genetic alterations. Blood. 2009;114(4):826–34.  54. Victora GD, Dominguez-Sola D, Holmes AB, et al. Identification of human germinal center light and dark zone cells and their relationship to human B-cell lymphomas. Blood. 2012;120(11):2240–8.  55. Kretzmer H, Bernhart SH, Wang W, et al. DNA methylome analysis in Burkitt and follicular lymphomas identifies differentially methylated regions linked to somatic mutation and transcriptional control. Nat. Genet. 2015;47(11):1316–25.  56. Sungalee S, Mamessier E, Morgado E, et al. Germinal center reentries of BCL2-overexpressing B cells drive follicular lymphoma progression. J. Clin. Invest. 2014;1–4.  57. Siegel RL, Miller KD, Jemal A. Cancer statistics, 2015. CA. Cancer J. Clin. 2015;65(1):5–29.  98 58. Anon. A clinical evaluation of the International Lymphoma Study Group classification of non-Hodgkin’s lymphoma. The Non-Hodgkin's Lymphoma Classification Project. Blood. 1997;89(11):3909–18.  59. Anderson JR, Armitage JO, Weisenburger DD. Epidemiology of the non-Hodgkin’s lymphomas: distributions of the major subtypes differ by geographic locations. Non-Hodgkin's Lymphoma Classification Project. Ann. Oncol. 1998;9(7):717–20.  60.  A predictive model for aggressive non-Hodgkin’s lymphoma. The International Non-Hodgkin's Lymphoma Prognostic Factors Project. N. Engl. J. Med. 1993;329(14):987–94.  61. Solal-Céligny P, Roy P, Colombat P, et al. Follicular lymphoma international prognostic index. Blood. 2004;104(5):1258–65.  62. Buske C, Hoster E, Dreyling M, et al. The Follicular Lymphoma International Prognostic Index (FLIPI) separates high-risk from intermediate- or low-risk patients with advanced-stage follicular lymphoma treated front-line with rituximab and the combination of cyclophosphamide, doxorubicin, vinc. Blood. 2006;108(5):1504–8.  63. Federico M, Bellei M, Marcheselli L, et al. Follicular lymphoma international prognostic index 2: a new prognostic index for follicular lymphoma developed by the international follicular lymphoma prognostic factor project. J. Clin. Oncol. 2009;27(27):4555–62.  64. Ghielmini M, Vitolo U, Kimby E, et al. ESMO Guidelines consensus conference on malignant lymphoma 2011 part 1: diffuse large B-cell lymphoma (DLBCL), follicular lymphoma (FL) and chronic lymphocytic leukemia (CLL). Ann. Oncol. 2013;24(3):561–76.  65. Zelenetz AD, Wierda WG, Abramson JS, et al. Non-Hodgkin’s Lymphomas, version 3.2012. J. Natl. Compr. Canc. Netw. 2012;10(12):1487–98.  66. Mac Manus MP, Hoppe RT. Is radiotherapy curative for stage I and II low-grade follicular lymphoma? Results of a long-term follow-up study of patients treated at Stanford University. J. Clin. Oncol. 1996;14(4):1282–90.  67. Campbell BA, Voss N, Woods R, et al. Long-term outcomes for patients with limited stage follicular lymphoma involved regional radiotherapy versus involved node radiotherapy. 99 Cancer. 2010;116(16):3797–3806.  68. Stuschke M, Hoederath A, Sack H, et al. Extended field and total central lymphatic radiotherapy in the treatment of early stage lymph node centroblastic-centrocytic lymphomas: results of a prospective multicenter study. Study Group NHL-frühe Stadien. Cancer. 1997;80(12):2273–84.  69. Wilder RB, Jones D, Tucker SL, et al. Long-term results with radiotherapy for Stage I-II follicular lymphomas. Int. J. Radiat. Oncol. Biol. Phys. 2001;51(5):1219–1227.  70. Guadagnolo BA, Li S, Neuberg D, et al. Long-term outcome and mortality trends in early-stage, Grade 1-2 follicular lymphoma treated with radiation therapy. Int. J. Radiat. Oncol. Biol. Phys. 2006;64(3):928–934.  71. Pugh TJ, Ballonoff A, Newman F, Rabinovitch R. Improved survival in patients with early stage low-grade follicular lymphoma treated with radiation: A surveillance, epidemiology, and end results database analysis. Cancer. 2010;116(16):3843–3851.  72. O’Brien ME, Easterbrook P, Powell J, et al. The natural history of low grade non-Hodgkin’s lymphoma and the impact of a no initial treatment policy on survival. Q. J. Med. 1991;80(292):651–60.  73. Young RC, Longo DL, Glatstein E, et al. The treatment of indolent lymphomas: watchful waiting v aggressive combined modality treatment. Semin. Hematol. 1988;25(2 Suppl 2):11–6.  74. Brice P, Bastion Y, Lepage E, et al. Comparison in low-tumor-burden follicular lymphomas between an initial no-treatment policy, prednimustine, or interferon alfa: a randomized study from the Groupe d’Etude des Lymphomes Folliculaires. Groupe d'Etude des Lymphomes de l'Adulte. J. Clin. Oncol. 1997;15(3):1110–7.  75. Ardeshna KM, Smith P, Norton A, et al. Long-term effect of a watch and wait policy versus immediate systemic treatment for asymptomatic advanced-stage non-Hodgkin lymphoma: a randomised controlled trial. Lancet. 2003;362(9383):516–22.  76. Marcus R, Imrie K, Solal-Celigny P, et al. Phase III Study of R-CVP Compared With Cyclophosphamide, Vincristine, and Prednisone Alone in Patients With Previously 100 Untreated Advanced Follicular Lymphoma. J. Clin. Oncol. 2008;26(28):4579–4586.  77. Hiddemann W, Kneba M, Dreyling M, et al. Frontline therapy with rituximab added to the combination of cyclophosphamide, doxorubicin, vincristine, and prednisone (CHOP) significantly improves the outcome for patients with advanced-stage follicular lymphoma compared with therapy with CHOP alone: R. Blood. 2005;106(12):3725–3732.  78. Herold M, Haas A, Srock S, et al. Rituximab added to first-line mitoxantrone, chlorambucil, and prednisolone chemotherapy followed by interferon maintenance prolongs survival in patients with advanced follicular lymphoma: An East German study group hematology and oncology study. J. Clin. Oncol. 2007;25(15):1986–1992.  79. Rummel MJ, Niederle N, Maschmeyer G, et al. Bendamustine plus rituximab versus CHOP plus rituximab as first-line treatment for patients with indolent and mantle-cell lymphomas: an open-label, multicentre, randomised, phase 3 non-inferiority trial. Lancet. 2013;381(9873):1203–10.  80. Flinn IW, van der Jagt R, Kahl BS, et al. Randomized trial of bendamustine-rituximab or R-CHOP/R-CVP in first-line treatment of indolent NHL or MCL: the BRIGHT study. Blood. 2014;123(19):2944–52.  81. Salles G, Seymour JF, Offner F, et al. Rituximab maintenance for 2 years in patients with high tumour burden follicular lymphoma responding to rituximab plus chemotherapy (PRIMA): A phase 3, randomised controlled trial. Lancet. 2011;377:42–51.  82. Roulland S, Faroudi M, Mamessier E, et al. Early steps of follicular lymphoma pathogenesis. 2011.  83. Yunis JJ, Frizzera G, Oken MM, et al. Multiple recurrent genomic defects in follicular lymphoma. A possible model for cancer. N. Engl. J. Med. 1987;316(2):79–84.  84. Tsujimoto Y, Gorham J, Cossman J, Jaffe E, Croce CM. The t(14;18) chromosome translocations involved in B-cell neoplasms result from mistakes in VDJ joining. Science. 1985;229(4720):1390–3.  85. Tsai AG, Lu H, Raghavan SC, et al. Human chromosomal translocations at CpG sites and a theoretical basis for their lineage and stage specificity. Cell. 2008;135(6):1130–42.  101 86. Cui X, Lu Z, Kurosawa A, et al. Both CpG methylation and activation-induced deaminase are required for the fragility of the human bcl-2 major breakpoint region: implications for the timing of the breaks in the t(14;18) translocation. Mol. Cell. Biol. 2013;33(5):947–57.  87. Kuraoka M, McWilliams L, Kelsoe G. AID expression during B-cell development: searching for answers. Immunol Res. 2011;49(1-3):3–13.  88. Tsujimoto Y, Finger LR, Yunis J, Nowell PC, Croce CM. Cloning of the chromosome breakpoint of neoplastic B cells with the t(14;18) chromosome translocation. Science. 1984;226(4678):1097–9.  89. Tsujimoto Y, Ikegaki N, Croce CM. Characterization of the protein product of bcl-2, the gene involved in human follicular lymphoma. Oncogene. 1987;2(1):3–7.  90. Gaulard P, d’Agay MF, Peuchmaur M, et al. Expression of the bcl-2 gene product in follicular lymphoma. Am. J. Pathol. 1992;140(5):1089–95.  91. Zinkel S, Gross  a, Yang E. BCL2 family in DNA damage and cell cycle control. Cell Death Differ. 2006;13(8):1351–9.  92. Hirt C, Dölken G, Janz S, Rabkin CS. Distribution of t(14;18)-positive, putative lymphoma precursor cells among B-cell subsets in healthy individuals. Br. J. Haematol. 2007;138(3):349–53.  93. Roulland S, Navarro J-M, Grenot P, et al. Follicular lymphoma-like B cells in healthy individuals: a novel intermediate step in early lymphomagenesis. J. Exp. Med. 2006;203(11):2425–31.  94. Limpens J, Stad R, Vos C, et al. Lymphoma-associated translocation t(14;18) in blood B cells of normal individuals. Blood. 1995;85(9):2528–36.  95. Dölken G, Illerhaus G, Hirt C, Mertelsmann R. BCL-2/JH rearrangements in circulating B cells of healthy blood donors and patients with nonmalignant diseases. J. Clin. Oncol. 1996;14(4):1333–44.  96. Schüler F, Dölken L, Hirt C, et al. Prevalence and frequency of circulating t(14;18)-MBR translocation carrying cells in healthy individuals. Int. J. Cancer. 2009;124(4):958–63.  102 97. Roulland S, Kelly RS, Morgado E, et al. t(14;18) Translocation: A Predictive Blood Biomarker for Follicular Lymphoma. J. Clin. Oncol. 2014;32(13.):  98. Hirt C, Camargo MC, Yu KJ, et al. Risk of follicular lymphoma associated with BCL2 translocations in peripheral blood. Leuk. Lymphoma. 2014;(November 2014):1–15.  99. Carlotti E, Wrench D, Matthews J, et al. Transformation of follicular lymphoma to diffuse large B-cell lymphoma may occur by divergent evolution from a common progenitor cell or by direct evolution from the follicular lymphoma clone. Blood. 2009;113(15):3553–7.  100. Weigert O, Kopp N, Lane AA, et al. Molecular ontogeny of donor-derived follicular lymphomas occurring after hematopoietic cell transplantation. Cancer Discov. 2012;2(1):47–55.  101. Cong P. In situ localization of follicular lymphoma: description and analysis by laser capture microdissection. Blood. 2002;99(9):3376–3382.  102. Jegalian AG, Eberle FC, Pack SD, et al. Follicular lymphoma in situ: clinical implications and comparisons with partial involvement by follicular lymphoma. Blood. 2011;118(11):2976–84.  103. Adam P, Katzenberger T, Eifert M, et al. Presence of preserved reactive germinal centers in follicular lymphoma is a strong histopathologic indicator of limited disease stage. Am. J. Surg. Pathol. 2005;29(12):1661–4.  104. Lo Coco F, Ye BH, Lista F, et al. Rearrangements of the BCL6 gene in diffuse large cell non-Hodgkin’s lymphoma. Blood. 1994;83(7):1757–9.  105. Otsuki T, Yano T, Clark HM, et al. Analysis of LAZ3 (BCL-6) status in B-cell non-Hodgkin’s lymphomas: results of rearrangement and gene expression studies and a mutational analysis of coding region sequences. Blood. 1995;85(10):2877–84.  106. Bastard C, Deweindt C, Kerckaert JP, et al. LAZ3 rearrangements in non-Hodgkin’s lymphoma: correlation with histology, immunophenotype, karyotype, and clinical outcome in 217 patients. Blood. 1994;83:2423–2427.  107. Christie L, Kernohan N, Levison D, et al. C-MYC translocation in t(14;18) positive follicular lymphoma at presentation: An adverse prognostic indicator? Leuk. Lymphoma. 103 2008;49(3):470–6.  108. Au WY, Horsman DE, Gascoyne RD, et al. The spectrum of lymphoma with 8q24 aberrations: a clinical, pathological and cytogenetic study of 87 consecutive cases. Leuk. Lymphoma. 2004;45(3):519–28.  109. Yano T, Jaffe ES, Longo DL, Raffeld M. MYC rearrangements in histologically progressed follicular lymphomas. Blood. 1992;80(3):758–67.  110. Johnson NA, Al-Tourah A, Brown CJ, et al. Prognostic significance of secondary cytogenetic alterations in follicular lymphomas. Genes. Chromosomes Cancer. 2008;47(12):1038–48.  111. Cheung K-JJ, Shah SP, Steidl C, et al. Genome-wide profiling of follicular lymphoma by array comparative genomic hybridization reveals prognostically significant DNA copy number imbalances. Blood. 2009;113(1):137–48.  112. Viardot A, Möller P, Högel J, et al. Clinicopathologic correlations of genomic gains and losses in follicular lymphoma. J. Clin. Oncol. 2002;20(23):4523–30.  113. Schwaenen C, Viardot A, Berger H, et al. Microarray-based genomic profiling reveals novel genomic aberrations in follicular lymphoma which associate with patient survival and gene expression status. Genes. Chromosomes Cancer. 2009;48(1):39–54.  114. Höglund M, Sehn L, Connors JM, et al. Identification of cytogenetic subgroups and karyotypic pathways of clonal evolution in follicular lymphomas. Genes. Chromosomes Cancer. 2004;39(3):195–204.  115. Kridel R, Sehn LH, Gascoyne RD. Review series Pathogenesis of follicular lymphoma. J. Clin. Invest. 2012;122(10):3424–3431.  116. Oricchio E, Nanjangud G, Wolfe AL, et al. The Eph-receptor A7 is a soluble tumor suppressor for follicular lymphoma. Cell. 2011;147(3):554–64.  117. Cheung K-JJ, Johnson NA, Affleck JG, et al. Acquired TNFRSF14 mutations in follicular lymphoma are associated with worse prognosis. Cancer Res. 2010;70(22):9166–74.  118. Launay E, Pangault C, Bertrand P, et al. High rate of TNFRSF14 gene alterations related to 104 1p36 region in de novo follicular lymphoma and impact on prognosis. Leukemia. 2012;26(3):559–62.  119. Kato M, Sanada M, Kato I, et al. Frequent inactivation of A20 in B-cell lymphomas. Nature. 2009;459(7247):712–716.  120. Honma K, Tsuzuki S, Nakagawa M, et al. TNFAIP3/A20 functions as a novel tumor suppressor gene in several subtypes of non-Hodgkin lymphomas. Blood. 2009;114(12):2467–75.  121. Gronbaek K, Straten PT, Ralfkiaer E, et al. Somatic Fas mutations in non-Hodgkin’s lymphoma: association with extranodal disease and autoimmunity. Blood. 1998;92(9):3018–3024.  122. Davies AJ, Lee AM, Taylor C, et al. A limited role for TP53 mutation in the transformation of follicular lymphoma to diffuse large B-cell lymphoma. Leukemia. 2005;19(8):1459–1465.  123. Davies AJ, Rosenwald A, Wright G, et al. Transformation of follicular lymphoma to diffuse large B-cell lymphoma proceeds by distinct oncogenic mechanisms. Br. J. Haematol. 2007;136(2):286–93.  124. Fitzgibbon J, Iqbal S, Davies A, et al. Genome-wide detection of recurring sites of uniparental disomy in follicular and transformed follicular lymphoma. Leukemia. 2007;21(7):1514–20.  125. Morin RD, Mendez-Lago M, Mungall AJ, et al. Frequent mutation of histone-modifying genes in non-Hodgkin lymphoma. Nature. 2011;476(7360):298–303.  126. Green MR, Kihira S, Liu CL, et al. Mutations in early follicular lymphoma progenitors are associated with suppressed antigen presentation. Proc. Natl. Acad. Sci. 2015;201501199.  127. Li H, Kaminski MS, Li Y, et al. Mutations in linker histone genes HIST1H1 B, C, D, and E; OCT2 (POU2F2); IRF8; and ARID1A underlying the pathogenesis of follicular lymphoma. Blood. 2014;123(10):1487–98.  128. Pastore A, Jurinovic V, Kridel R, et al. Integration of gene mutations in risk prognostication for patients receiving first-line immunochemotherapy for follicular lymphoma: a 105 retrospective analysis of a prospective clinical trial and validation in a population-based registry. Lancet. Oncol. 2015;2045(15):1–12.  129. Shilatifard A. The COMPASS family of histone H3K4 methylases: mechanisms of regulation in development and disease pathogenesis. Annu. Rev. Biochem. 2012;81(Figure 1):65–95.  130. Hu D, Gao X, Morgan MA, et al. The MLL3/MLL4 branches of the COMPASS family function as major histone H3K4 monomethylases at enhancers. Mol. Cell. Biol. 2013;33(23):4745–54.  131. Pasqualucci L, Trifonov V, Fabbri G, et al. Analysis of the coding genome of diffuse large B-cell lymphoma. Nat. Genet. 2011;43(9):830–7.  132. Green MR, Gandhi MK, Courtney MJ, Marlton P, Griffiths L. Relative abundance of full-length and truncated FOXP1 isoforms is associated with differential NFkappaB activity in Follicular Lymphoma. Leuk. Res. 2009;33(12):1699–702.  133. Ortega-Molina A, Boss IW, Canela A, et al. The histone lysine methyltransferase KMT2D sustains a gene expression program that represses B cell lymphoma development. Nat. Med. 2015;(September):1–15.  134. Zhang J, Dominguez-Sola D, Hussein S, et al. Disruption of KMT2D perturbs germinal center B cell development and promotes lymphomagenesis. Nat. Med. 2015;(September):1–13.  135. Okosun J, Bödör C, Wang J, et al. Integrated genomic analysis identifies recurrent mutations and evolution patterns driving the initiation and progression of follicular lymphoma. Nat. Genet. 2014;46(2):176–81.  136. Pasqualucci L, Khiabanian H, Fangazio M, et al. Genetics of Follicular Lymphoma Transformation. Cell Rep. 2014;6(1):130–140.  137. Kalkhoven E. CBP and p300: HATs for different occasions. Biochem. Pharmacol. 2004;68(6):1145–1155.  138. Bannister AJ, Kouzarides T. The CBP co-activator is a histone acetyltransferase. Nature. 1996;384(6610):641–3.  106 139. Ogryzko V V., Schiltz RL, Russanova V, Howard BH, Nakatani Y. The transcriptional coactivators p300 and CBP are histone acetyltransferases. Cell. 1996;87(5):953–959.  140. Gu W, Shi XL, Roeder RG. Synergistic activation of transcription by CBP and p53. Nature. 1997;387(June):819–823.  141. Lill NL, Grossman SR, Ginsberg D, DeCaprio J, Livingston DM. Binding and modulation of p53 by p300/CBP coactivators. Nature. 1997;387(6635):823–7.  142. Avantaggiati ML, Ogryzko V, Gardner K, et al. Recruitment of p300/CBP in p53-dependent signal pathways. Cell. 1997;89:1175–1184.  143. Bereshchenko OR, Gu W, Dalla-Favera R. Acetylation inactivates the transcriptional repressor BCL6. Nat. Genet. 2002;32(4):606–613.  144. Pasqualucci L, Dominguez-Sola D, Chiarenza A, et al. Inactivating mutations of acetyltransferase genes in B-cell lymphoma. Nature. 2011;471(7337):189–95.  145. Green MR, Gentles AJ, Nair R V, et al. Hierarchy in somatic mutations arising during genomic evolution and progression of follicular lymphoma. Blood. 2013;121(9):1604–11.  146. Cao R. Role of Histone H3 Lysine 27 Methylation in Polycomb-Group Silencing. Science (80-. ). 2002;298(5595):1039–1043.  147. Müller J, Hart CM, Francis NJ, et al. Histone methyltransferase activity of a Drosophila Polycomb group repressor complex. Cell. 2002;111(2):197–208.  148. Czermin B, Melfi R, McCabe D, et al. Drosophila enhancer of Zeste/ESC complexes have a histone H3 methyltransferase activity that marks chromosomal Polycomb sites. Cell. 2002;111(2):185–196.  149. Chou RH, Yu YL, Hung MC. The roles of EZH2 in cell lineage commitment. Am. J. Transl. Res. 2011;3(3):243–250.  150. Lund K, Adams PD, Copland M. EZH2 in normal and malignant hematopoiesis. Leukemia. 2014;28(1):44–9.  151. Su I-H, Basavaraj A, Krutchinsky AN, et al. Ezh2 controls B cell development through histone H3 methylation and Igh rearrangement. Nat. Immunol. 2003;4(2):124–131.  107 152. van Galen JC, Dukers DF, Giroth C, et al. Distinct expression patterns of polycomb oncoproteins and their binding partners during the germinal center reaction. Eur. J. Immunol. 2004;34(7):1870–81.  153. Velichutina I, Shaknovich R, Geng H, et al. EZH2-mediated epigenetic silencing in germinal center B cells contributes to proliferation and lymphomagenesis. Blood. 2010;116(24):5247–55.  154. Morin RD, Johnson NA, Severson TM, et al. Somatic mutations altering EZH2 (Tyr641) in follicular and diffuse large B-cell lymphomas of germinal-center origin. Nat. Genet. 2010;42(2):181–5.  155. Bödör C, O’Riain C, Wrench D, et al. EZH2 Y641 mutations in follicular lymphoma. Leukemia. 2011;25(4):726–9.  156. Bödör C, Grossmann V, Popov N, et al. EZH2 mutations are frequent and represent an early event in follicular lymphoma. Blood. 2013;122(18):3165–8.  157. Guo S, Chan JKC, Iqbal J, et al. EZH2 mutations in follicular lymphoma from different ethnic groups and associated gene expression alterations. Clin. Cancer Res. 2014;20(12):3078–86.  158. Yap DB, Chu J, Berg T, et al. Somatic mutations at EZH2 Y641 act dominantly through a mechanism of selectively altered PRC2 catalytic activity, to increase H3K27 trimethylation. Blood. 2011;117(8):2451–9.  159. Sneeringer CJ, Scott MP, Kuntz KW, et al. Coordinated activities of wild-type plus mutant EZH2 drive tumor-associated hypertrimethylation of lysine 27 on histone H3 (H3K27) in human B-cell lymphomas. Proc. Natl. Acad. Sci. U. S. A. 2010;107(49):20980–5.  160. Berg T, Thoene S, Yap D, et al. A transgenic mouse model demonstrating the oncogenic role of mutations in the polycomb-group gene EZH2 in lymphomagenesis. Blood. 2014;123(25):3914–24.  161. Knutson SK, Wigle TJ, Warholic NM, et al. A selective inhibitor of EZH2 blocks H3K27 methylation and kills mutant lymphoma cells. Nat. Chem. Biol. 2012;8(11):890–896.  162. McCabe MT, Ott HM, Ganji G, et al. EZH2 inhibition as a therapeutic strategy for 108 lymphoma with EZH2-activating mutations. Nature. 2012;492(7427):108–12.  163. O’Shea JJ, Schwartz DM, Villarino A V., et al. The JAK-STAT Pathway: Impact on Human Disease and Therapeutic Intervention. Annu Rev Med. 2015;66(1):311–328.  164. Takeda K, Tanaka T, Shi W, et al. Essential role of Stat6 in IL-4 signalling. Nature. 1996;380(6575):627–30.  165. Hou J, Schindler U, Henzel WJ, et al. An interleukin-4-induced transcription factor: IL-4 Stat. Science. 1994;265(5179):1701–1706.  166. Pangault C, Amé-Thomas P, Ruminy P, et al. Follicular lymphoma cell niche: identification of a preeminent IL-4-dependent T(FH)-B cell axis. Leukemia. 2010;24(12):2080–9.  167. Calvo KR, Dabir B, Kovach A, et al. IL-4 protein expression and basal activation of Erk in vivo in follicular lymphoma. Blood. 2008;112(9):3818–26.  168. Rawal S, Chu F, Zhang M, et al. Cross talk between follicular Th cells and tumor cells in human follicular lymphoma promotes immune evasion in the tumor microenvironment. J. Immunol. 2013;190(12):6681–93.  169. Yildiz M, Li H, Bernard D, et al. Activating STAT6 mutations in follicular lymphoma. Blood. 2015;125(4):668–79.  170. Okosun J, Wolfson RL, Wang J, et al. Recurrent mTORC1-activating RRAGC mutations in follicular lymphoma. Nat. Genet. 2015;(December.):  171. Zhu D, McCarthy H, Ottensmeier CH, et al. Acquisition of potential N-glycosylation sites in the immunoglobulin variable region by somatic mutation is a distinctive feature of follicular lymphoma. Blood. 2002;99(7):2562–2568.  172. Coelho V, Krysov S, Ghaemmaghami AM, et al. Glycosylation of surface Ig creates a functional bridge between human follicular lymphoma and microenvironmental lectins. Proc. Natl. Acad. Sci. U. S. A. 2010;107(43):18587–92.  173. Radcliffe CM, Arnold JN, Suter DM, et al. Human follicular lymphoma cells contain oligomannose glycans in the antigen-binding site of the B-cell receptor. J. Biol. Chem. 2007;282(10):7405–15.  109 174. Schneider D, Dühren-von Minden M, Alkhatib A, et al. Lectins from opportunistic bacteria interact with acquired variable region glycans of surface Ig in follicular lymphoma. Blood. 2015;125(21):3287–3297.  175. Linley A, Krysov S, Ponzoni M, et al. Lectin binding to surface Ig variable regions provides a universal persistent activating signal for follicular lymphoma cells. Blood. 2015;126(16):1902–10.  176. Amin R, Mourcin F, Uhel F, et al. DC-SIGN-expressing macrophages trigger activation of mannosylated IgM B-cell receptor in follicular lymphoma. Blood. 2015;126(16):1911–20.  177. Su W, Spencer J, Wotherspoon AC. Relative distribution of tumour cells and reactive cells in follicular lymphoma. J. Pathol. 2001;193(4):498–504.  178. Oeschger S, Bräuninger A, Küppers R, Hansmann M-L. Tumor cell dissemination in follicular lymphoma. Blood. 2002;99(6):2192–8.  179. Tweeddale ME, Lim B, Jamal N, et al. The presence of clonogenic cells in high-grade malignant lymphoma: a prognostic factor. Blood. 1987;69(5):1307–14.  180. Johnson PW, Watt SM, Betts DR, et al. Isolated follicular lymphoma cells are resistant to apoptosis and can be grown in vitro in the CD40/stromal cell system. Blood. 1993;82(6):1848–57.  181. Buske C, Twiling A, Gogowski G, et al. In vitro activation of low-grade non-Hodgkin’s lymphoma by murine fibroblasts, IL-4, anti-CD40 antibodies and the soluble CD40 ligand. Leukemia. 1997;11(11):1862–7.  182. Umetsu DT, Esserman L, Donlon TA, DeKruyff RH, Levy R. Induction of proliferation of human follicular (B type) lymphoma cells by cognate interaction with CD4+ T cell clones. J. Immunol. 1990;144(7):2550–7.  183. Dave SS, Wright G, Tan B, et al. Prediction of survival in follicular lymphoma based on molecular features of tumor-infiltrating immune cells. N. Engl. J. Med. 2004;351(21):2159–69.  184. Husson H, Carideo EG, Neuberg D, et al. Gene expression profiling of follicular lymphoma and normal germinal center B cells using cDNA arrays. Blood. 2002;99(1):282–9.  110 185. Glas AM, Kersten MJ, Delahaye LJMJ, et al. Gene expression profiling in follicular lymphoma to assess clinical aggressiveness and to guide the choice of treatment. Blood. 2005;105(1):301–7.  186. Glas AM, Knoops L, Delahaye L, et al. Gene-expression and immunohistochemical study of specific T-cell subsets and accessory cell types in the transformation and prognosis of follicular lymphoma. J. Clin. Oncol. 2007;25(4):390–8.  187. Harjunpää A, Taskinen M, Nykter M, et al. Differential gene expression in non-malignant tumour microenvironment is associated with outcome in follicular lymphoma patients treated with rituximab and CHOP. Br. J. Haematol. 2006;135(1):33–42.  188. Kiaii S, Clear AJ, Ramsay AG, et al. Follicular lymphoma cells induce changes in T-cell gene expression and function: potential impact on survival and risk of transformation. J. Clin. Oncol. 2013;31(21):2654–61.  189. Scott DW, Gascoyne RD. The tumour microenvironment in B cell lymphomas. Nat. Rev. Cancer. 2014;(July.):  190. Amé-Thomas P, Le Priol J, Yssel H, et al. Characterization of intratumoral follicular helper T cells in follicular lymphoma: role in the survival of malignant B cells. Leukemia. 2012;26(5):1053–63.  191. Travert M, Ame-Thomas P, Pangault C, et al. CD40 ligand protects from TRAIL-induced apoptosis in follicular lymphomas through NF-kappaB activation and up-regulation of c-FLIP and Bcl-xL. J. Immunol. 2008;181(2):1001–11.  192. Epron G, Ame-Thomas P, Le Priol J, et al. Monocytes and T cells cooperate to favor normal and follicular lymphoma B-cell growth: role of IL-15 and CD40L signaling. Leukemia. 2012;26(1):139–48.  193. Yang Z-Z, Novak AJ, Stenson MJ, Witzig TE, Ansell SM. Intratumoral CD4+CD25+ regulatory T-cell-mediated suppression of infiltrating CD4+ T cells in B-cell non-Hodgkin lymphoma. Blood. 2006;107(9):3639–46.  194. Ai WZ, Hou J, Zeiser R, et al. Follicular lymphoma B cells induce the conversion of conventional CD4+ T cells to T-regulatory cells. Int. J. Cancer. 2009;124(1):239–44.  111 195. Yang Z-Z, Novak AJ, Ziesmer SC, Witzig TE, Ansell SM. Malignant B cells skew the balance of regulatory T cells and TH17 cells in B-cell non-Hodgkin’s lymphoma. Cancer Res. 2009;69(13):5522–30.  196. Yang Z-Z, Grote DM, Ziesmer SC, et al. Soluble and membrane-bound TGF-β-mediated regulation of intratumoral T cell differentiation and function in B-cell non-Hodgkin lymphoma. PLoS One. 2013;8(3):e59456.  197. Yang Z-Z, Grote DM, Xiu B, et al. TGF-β upregulates CD70 expression and induces exhaustion of effector memory T cells in B-cell non-Hodgkin’s lymphoma. Leukemia. 2014;28(9):1872–1884.  198. Yang Z, Grote DM, Ziesmer SC, et al. IL-12 upregulates TIM-3 expression and induces T cell exhaustion in patients with follicular B cell non-Hodgkin lymphoma. J. Clin. Invest. 2012;122(4):1271–82.  199. Alvaro T, Lejeune M, Salvadó M-T, et al. Immunohistochemical patterns of reactive microenvironment are associated with clinicobiologic behavior in follicular lymphoma patients. J. Clin. Oncol. 2006;24(34):5350–7.  200. Wahlin BE, Aggarwal M, Montes-Moreno S, et al. A unifying microenvironment model in follicular lymphoma: outcome is predicted by programmed death-1--positive, regulatory, cytotoxic, and helper T cells and macrophages. Clin. Cancer Res. 2010;16(2):637–50.  201. Wahlin BE, Sander B, Christensson B, Kimby E. CD8+ T-cell content in diagnostic lymph nodes measured by flow cytometry is a predictor of survival in follicular lymphoma. Clin. Cancer Res. 2007;13(2 Pt 1):388–97.  202. Laurent C, Müller S, Do C, et al. Distribution, function, and prognostic value of cytotoxic T lymphocytes in follicular lymphoma: a 3-D tissue-imaging study. Blood. 2011;118(20):5371–9.  203. Ramsay AG, Clear AJ, Kelly G, et al. Follicular lymphoma cells induce T-cell immunologic synapse dysfunction that can be repaired with lenalidomide: implications for the tumor microenvironment and immunotherapy. Blood. 2009;114(21):4713–20.  204. Kridel R, Xerri L, Gelas-Dore B, et al. The Prognostic Impact of CD163-Positive 112 Macrophages in Follicular Lymphoma: A Study from the BC Cancer Agency and the Lymphoma Study Association. Clin. Cancer Res. 2015;21(15):3428–35.  205. Farinha P, Masoudi H, Skinnider BF, et al. Analysis of multiple biomarkers shows that lymphoma-associated macrophage (LAM) content is an independent predictor of survival in follicular lymphoma (FL). Blood. 2005;106(6):2169–74.  206. Alvaro T, Lejeune M, Camacho FI, et al. The presence of STAT1-positive tumor-associated macrophages and their relation to outcome in patients with follicular lymphoma. Haematologica. 2006;91(12):1605–12.  207. Canioni D, Salles G, Mounier N, et al. High numbers of tumor-associated macrophages have an adverse prognostic value that can be circumvented by rituximab in patients with follicular lymphoma enrolled onto the GELA-GOELAMS FL-2000 trial. J. Clin. Oncol. 2008;26(3):440–6.  208. Byers RJ, Sakhinia E, Joseph P, et al. Clinical quantitation of immune signature in follicular lymphoma by RT-PCR-based gene expression profiling. Blood. 2008;111(9):4764–70.  209. Coiffier B, Li W, Henitz ED, et al. Prespecified candidate biomarkers identify follicular lymphoma patients who achieved longer progression-free survival with bortezomib-rituximab versus rituximab. Clin. Cancer Res. 2013;19(9):2551–61.  210. Taskinen M, Karjalainen-Lindsberg M-L, Nyman H, Eerola L-M, Leppä S. A high tumor-associated macrophage content predicts favorable outcome in follicular lymphoma patients treated with rituximab and cyclophosphamide-doxorubicin-vincristine-prednisone. Clin. Cancer Res. 2007;13(19):5784–9.  211. Clear AJ, Lee AM, Calaminici M, et al. Increased angiogenic sprouting in poor prognosis FL is associated with elevated numbers of CD163+ macrophages within the immediate sprouting microenvironment. Blood. 2010;115(24):5053–6.  212. Minard-Colin V, Xiu Y, Poe JC, et al. Lymphoma depletion during CD20 immunotherapy in mice is mediated by macrophage FcgammaRI, FcgammaRIII, and FcgammaRIV. Blood. 2008;112(4):1205–13.  213. Chang JE, Turley SJ. Stromal infrastructure of the lymph node and coordination of 113 immunity. Trends Immunol. 2014;36(1):30–39.  214. Anthony B a., Link DC. Regulation of hematopoietic stem cells by bone marrow stromal cells. Trends Immunol. 2014;35(1):32–37.  215. Lwin T, Hazlehurst LA, Li Z, et al. Bone marrow stromal cells prevent apoptosis of lymphoma cells by upregulation of anti-apoptotic proteins associated with activation of NF-kappaB (RelB/p52) in non-Hodgkin’s lymphoma cells. Leukemia. 2007;21(7):1521–31.  216. Lwin T, Crespo LA, Wu A, et al. Lymphoma cell adhesion-induced expression of B cell-activating factor of the TNF family in bone marrow stromal cells protects non-Hodgkin’s B lymphoma cells from apoptosis. Leukemia. 2009;23(1):170–7.  217. Amé-Thomas P, Maby-El Hajjami H, Monvoisin C, et al. Human mesenchymal stem cells isolated from bone marrow and lymphoid organs support tumor B-cell growth: role of stromal cells in follicular lymphoma pathogenesis. Blood. 2007;109(2):693–702.  218. Guilloton F, Caron G, Ménard C, et al. Mesenchymal stromal cells orchestrate follicular lymphoma cell niche through the CCL2-dependent recruitment and polarization of monocytes. Blood. 2012;119(11):2556–67.  219. Loeffler M, Kreuz M, Haake  a, et al. Genomic and epigenomic co-evolution in follicular lymphomas. Leukemia. 2014;(August 2014):456–463.  220. d’Amore F, Chan E, Iqbal J, et al. Clonal evolution in t(14;18)-positive follicular lymphoma, evidence for multiple common pathways, and frequent parallel clonal evolution. Clin. Cancer Res. 2008;14(22):7180–7.  221. Eide MB, Liestøl K, Lingjaerde OC, et al. Genomic alterations reveal potential for higher grade transformation in follicular lymphoma and confirm parallel evolution of tumor cell clones. Blood. 2010;116(9):1489–97.  222. Zhu D, Hawkins RE, Hamblin TJ, Stevenson FK. Clonal history of a human follicular lymphoma as revealed in the immunoglobulin variable region genes. Br. J. Haematol. 1994;86(3):505–12.  223. Spence JM, Abumoussa A, Spence JP, Burack WR. Intraclonal Diversity in Follicular Lymphoma Analyzed by Quantitative Ultradeep Sequencing of Noncoding Regions. J. 114 Immunol. 2014;  224. Carlotti E, Wrench D, Rosignoli G, et al. High Throughput Sequencing Analysis of the Immunoglobulin Heavy Chain Gene from Flow-Sorted B Cell Sub-Populations Define the Dynamics of Follicular Lymphoma Clonal Evolution. PLoS One. 2015;10(9):e0134833.  225. Landau DA, Carter SL, Stojanov P, et al. Evolution and impact of subclonal mutations in chronic lymphocytic leukemia. Cell. 2013;152(4):714–26.  226. Landau DA, Tausch E, Taylor-Weiner AN, et al. Mutations driving CLL and their evolution in progression and relapse. Nature. 2015;  227. Gopal AK, Kahl BS, de Vos S, et al. PI3Kδ inhibition by idelalisib in patients with relapsed indolent lymphoma. N. Engl. J. Med. 2014;370(11):1008–18.  228. Advani RH, Buggy JJ, Sharman JP, et al. Bruton tyrosine kinase inhibitor ibrutinib (PCI-32765) has significant activity in patients with relapsed/refractory B-cell malignancies. J. Clin. Oncol. 2013;31(1):88–94.  229. Al-Mansour M, Connors JM, Gascoyne RD, Skinnider B, Savage KJ. Transformation to aggressive lymphoma in nodular lymphocyte-predominant Hodgkin’s lymphoma. J. Clin. Oncol. 2010;28(5):793–799.  230. Lossos IS, Gascoyne RD. Transformation of follicular lymphoma. Best Pract. Res. Clin. Haematol. 2011;24(2):147–63.  231. Casulo C, Burack WR, Friedberg JW. Transformed follicular non-Hodgkin lymphoma. Blood. 2015;125(1):40–47.  232. Al-Tourah AJ, Gill KK, Chhanabhai M, et al. Population-based analysis of incidence and outcome of transformed non-Hodgkin’s lymphoma. J. Clin. Oncol. 2008;26(32):5165–9.  233. Montoto S, Fitzgibbon J. Transformation of indolent B-cell lymphomas. J. Clin. Oncol. 2011;29(14):1827–34.  234. Link BK, Maurer MJ, Nowakowski GS, et al. Rates and outcomes of follicular lymphoma transformation in the immunochemotherapy era: a report from the University of Iowa/MayoClinic Specialized Program of Research Excellence Molecular Epidemiology 115 Resource. J. Clin. Oncol. 2013;31(26):3272–8.  235. Wagner-Johnston ND, Link BK, Byrtek M, et al. Outcomes of transformed follicular lymphoma in the modern era: a report from the National LymphoCare Study (NLCS). Blood. 2015;126(7):851–858.  236. Conconi A, Ponzio C, Lobetti-Bodoni C, et al. Incidence, risk factors and outcome of histological transformation in follicular lymphoma. Br. J. Haematol. 2012;157:188–196.  237. Bains P, Al Tourah A, Campbell BA, et al. Incidence of transformation to aggressive lymphoma in limited-stage follicular lymphoma treated with radiotherapy. Ann. Oncol. 2013;24(2):428–32.  238. Giné E, Montoto S, Bosch F, et al. The Follicular Lymphoma International Prognostic Index (FLIPI) and the histological subtype are the most important factors to predict histological transformation in follicular lymphoma. Ann. Oncol. 2006;17(10):1539–1545.  239. Montoto S, López-Guillermo A, Colomer D, et al. Incidence and clinical significance of bcl-2/IgH rearrangements in follicular lymphoma. Leuk. Lymphoma. 2003;44(1):71–6.  240. Montoto S, Davies AJ, Matthews J, et al. Risk and clinical implications of transformation of follicular lymphoma to diffuse large B-cell lymphoma. J. Clin. Oncol. 2007;25(17):2426–33.  241. Bastion Y, Sebban C, Berger F, et al. Incidence, predictive factors, and outcome of lymphoma transformation in follicular lymphoma patients. J. Clin. Oncol. 1997;15(4):1587–94.  242. Lerch K, Meyer AH, Stroux A, et al. Impact of prior treatment on outcome of transformed follicular lymphoma and relapsed de novo diffuse large B cell lymphoma: a retrospective multicentre analysis. Ann. Hematol. 2015;94(6):981–8.  243. Sander CA, Yano T, Clark HM, et al. p53 mutation is associated with progression in follicular lymphomas. Blood. 1993;82(7):1994–2004.  244. Lo Coco F, Gaidano G, Louie DC, et al. p53 mutations are associated with histologic transformation of follicular lymphoma. Blood. 1993;82(8):2289–2295.  116 245. Bouska A, McKeithan TW, Deffenbacher KE, et al. Genome-wide copy-number analyses reveal genomic abnormalities involved in transformation of follicular lymphoma. Blood. 2014;123(11):1681–1690.  246. Lossos IS, Alizadeh AA, Diehn M, et al. Transformation of follicular lymphoma to diffuse large-cell lymphoma: alternative patterns with increased or decreased expression of c-myc and its regulated genes. Proc. Natl. Acad. Sci. U. S. A. 2002;99(13):8886–8891.  247. Pinyol M, Cobo F, Bea S, et al. p16(INK4a) gene inactivation by deletions, mutations, and hypermethylation is associated with transformed and aggressive variants of non-Hodgkin’s lymphomas. Blood. 1998;91(8):2977–2984.  248. Oricchio E, Ciriello G, Jiang M, et al. Frequent disruption of the RB pathway in indolent follicular lymphoma suggests a new combination therapy. J. Exp. Med. 2014;211(7):1379–91.  249. Martinez-Climent JA, Alizadeh AA, Segraves R, et al. Transformation of follicular lymphoma to diffuse large cell lymphoma is associated with a heterogeneous set of DNA copy number and gene expression alterations. Blood. 2003;101(8):3109–3117.  250. Swerdlow SH, Campo E, Harris NL, et al. WHO Classification of Tumours of Haematopoietic and Lymphoid Tissues. IARC Lyon, France; 2008.  251. Karube K, Guo Y, Suzumiya J, et al. CD10-MUM1+ follicular lymphoma lacks BCL2 gene translocation and shows characteristic biologic and clinical features. Blood. 2007;109(7):3076–3079.  252. Ott G, Katzenberger T, Lohr A, et al. Cytomorphologic, immunohistochemical, and cytogenetic profiles of follicular lymphoma: 2 types of follicular lymphoma grade 3. Blood. 2002;99(10):3806–3812.  253. Jardin F, Gaulard P, Buchonnet G, et al. Follicular lymphoma without t(14;18) and with BCL-6 rearrangement: a lymphoma subtype with distinct pathological, molecular and clinical characteristics. Leukemia. 2002;16(11):2309–17.  254. Rodriguez J, McLaughlin P, Hagemeister FB, et al. Follicular large cell lymphoma: an aggressive lymphoma that often presents with favorable prognostic features. Blood. 117 1999;93(7):2202–7.  255. Ganti AK, Weisenburger DD, Smith LM, et al. Patients with grade 3 follicular lymphoma have prolonged relapse-free survival following anthracycline-based chemotherapy: the Nebraska Lymphoma Study Group Experience. Ann. Oncol. 2006;17(6):920–7.  256. Martin AR, Weisenburger DD, Chan WC, et al. Prognostic value of cellular proliferation and histologic grade in follicular lymphoma. Blood. 1995;85(12):3671–8.  257. Chau I, Jones R, Cunningham D, et al. Outcome of follicular lymphoma grade 3: is anthracycline necessary as front-line therapy? Br. J. Cancer. 2003;89(1):36–42.  258. Miller TP, LeBlanc M, Grogan TM, Fisher RI. Follicular lymphomas: do histologic subtypes predict outcome? Hematol. Oncol. Clin. North Am. 1997;11(5):893–900.  259. Hans CP, Weisenburger DD, Vose JM, et al. A significant diffuse component predicts for inferior survival in grade 3 follicular lymphoma, but cytologic subtypes do not predict survival. Blood. 2003;101(6):2363–7.  260. Warnke RA, Kim H, Fuks Z, Dorfman RF. The coexistence of nodular and diffuse patterns in nodular non-Hodgkin’s lymphomas: significance and clinicopathologic correlation. Cancer. 1977;40(3):1229–33.  261. Horsman DE, Okamoto I, Ludkovski O, et al. Follicular lymphoma lacking the t(14;18)(q32;q21): identification of two disease subtypes. Br. J. Haematol. 2003;120(3):424–33.  262. Weisenburger DD, Gascoyne RD, Bierman PJ, et al. Clinical significance of the t(14;18) and BCL2 overexpression in follicular large cell lymphoma. Leuk. Lymphoma. 2000;36(5-6):513–23.  263. Tilly H, Rossi  a, Stamatoullas  a, et al. Prognostic value of chromosomal abnormalities in follicular lymphoma. Blood. 1994;84(4):1043–1049.  264. O’Shea D, O’Riain C, Taylor C, et al. The presence of TP53 mutation at diagnosis of follicular lymphoma identifies a high-risk group of patients with shortened time to disease progression and poorer overall survival. Blood. 2008;112(8):3126–9.  118 265. Alhejaily A, Day AG, Feilotter HE, Baetz TD, Lebrun DP. Inactivation of the CDKN2A tumor suppressor gene by deletion or methylation is common at diagnosis in follicular lymphoma and associated with poor clinical outcome. Clin. Cancer Res. 2014;20(6):1078–0432.CCR–13–2175–.  266. Correia C, Schneider PA, Dai H, et al. BCL2 mutations are associated with increased risk of transformation and shortened survival in follicular lymphoma. Blood. 2014;125(4):658–668.  267. Tan D, Horning SJ, Hoppe RT, et al. Improvements in observed and relative survival in follicular grade 1-2 lymphoma during 4 decades: the Stanford University experience. Blood. 2013;122(6):981–987.  268. Bachy E, Houot R, Morschhauser F, et al. Long-term follow up of the FL2000 study comparing CHVP-interferon to CHVP-interferon plus rituximab in follicular lymphoma. Haematologica. 2013;98(7):1107–14.  269. Casulo C, Byrtek M, Dawson KL, et al. Early Relapse of Follicular Lymphoma After Rituximab Plus Cyclophosphamide, Doxorubicin, Vincristine, and Prednisone Defines Patients at High Risk for Death: An Analysis From the National LymphoCare Study. J. Clin. Oncol. 2015;  270. Mozessohn L, Cheung MC, Crump M, et al. Chemoimmunotherapy resistant follicular lymphoma: predictors of resistance, association with transformation and prognosis. Leuk. Lymphoma. 2014;55(August 2013):2502–2507.  271. Horsman DE, Connors JM, Pantzar T, Gascoyne RD. Analysis of secondary chromosomal alterations in 165 cases of follicular lymphoma with t(14;18). Genes. Chromosomes Cancer. 2001;30(4):375–82.  272. Bognár  a, Csernus B, Bödör C, et al. Clonal selection in the bone marrow involvement of follicular lymphoma. Leukemia. 2005;19(9):1656–62.  273. Wartenberg M, Vasil P, zum Bueschenfelde CM, et al. Somatic hypermutation analysis in follicular lymphoma provides evidence suggesting bidirectional cell migration between lymph node and bone marrow during disease progression and relapse. Haematologica. 119 2013;98(9):1433–41.  274. Nowell PC. The clonal evolution of tumor cell populations. Science. 1976;194(4260):23–28.  275. Aparicio S, Caldas C. The implications of clonal genome evolution for cancer medicine. N. Engl. J. Med. 2013;368(9):842–51.  276. Greaves M. Evolutionary Determinants of Cancer. Cancer Discov. 2015;(August):806–820.  277. Schmitt MW, Loeb LA, Salk JJ. The influence of subclonal resistance mutations on targeted cancer therapy. Nat. Rev. Clin. Oncol. 2015;  278. Junlen HR, Peterson S, Kimby E, et al. Follicular lymphoma in Sweden: nationwide improved survival in the rituximab era, particularly in elderly women - a Swedish Lymphoma Registry study. Leukemia. 2014;29(3):668–676.  279. Akasaka T, Lossos IS, Levy R. BCL6 gene translocation in follicular lymphoma: a harbinger of eventual transformation to diffuse aggressive lymphoma. Blood. 2003;102(4):1443–8.  280. Do B, Lossos IS, Thorstenson Y, Oefner PJ, Levy R. Analysis of FAS (CD95) gene mutations in higher-grade transformation of follicle center lymphoma. Leuk. Lymphoma. 2003;44(8):1317–1323.  281. O’Shea D, O’Riain C, Gupta M, et al. Regions of acquired uniparental disomy at diagnosis of follicular lymphoma are associated with both overall survival and risk of transformation. Blood. 2009;113(10):2298–301.  282. Wright G, Tan B, Rosenwald A, et al. A gene expression-based method to diagnose clinically distinct subgroups of diffuse large B cell lymphoma. Proc. Natl. Acad. Sci. U. S. A. 2003;100(17):9991–6.  283. Johnson NA, Savage KJ, Ludkovski O, et al. Lymphomas with concurrent BCL2 and MYC translocations: the critical factors associated with survival. Blood. 2009;114(11):2273–9.  284. Kanungo A, Medeiros LJ, Abruzzo L V, Lin P. Lymphoid neoplasms associated with concurrent t(14;18) and 8q24/c-MYC translocation generally have a poor prognosis. Mod. 120 Pathol. 2006;19(1):25–33.  285. Barrans S, Crouch S, Smith A, et al. Rearrangement of MYC is associated with poor prognosis in patients with diffuse large B-cell lymphoma treated in the era of rituximab. J. Clin. Oncol. 2010;28(20):3360–3365.  286. Dunleavy K, Pittaluga S, Czuczman MS, et al. Differential efficacy of bortezomib plus chemotherapy within molecular subtypes of diffuse large B-cell lymphoma. Blood. 2009;113(24):6069–6076.  287. Nowakowski GS, LaPlant B, Macon WR, et al. Effect of lenalidomide combined with R-CHOP (R2CHOP) on negative prognostic impact of nongerminal center (non-GCB) phenotype in newly diagnosed diffuse large B-cell lymphoma: A phase 2 study. J. Clin. Oncol. 2014;32(3):abstr 8520.  288. Wilson WH, Young RM, Schmitz R, et al. Targeting B cell receptor signaling with ibrutinib in diffuse large B cell lymphoma. Nat. Med. 2015;21(8):1–6.  289. Hans CP, Weisenburger DD, Greiner TC, et al. Confirmation of the molecular classification of diffuse large B-cell lymphoma by immunohistochemistry using a tissue microarray. Blood. 2004;103(1):275–82.  290. Scott DW, Wright GW, Williams PM, et al. Determining cell-of-origin subtypes of diffuse large B-cell lymphoma using gene expression in formalin-fixed paraffin-embedded tissue. Blood. 2014;123(8):1214–7.  291. Li H, Durbin R. Fast and accurate short read alignment with Burrows-Wheeler transform. Bioinformatics. 2009;25(14):1754–1760.  292. Koboldt DC, Chen K, Wylie T, et al. VarScan: Variant detection in massively parallel sequencing of individual and pooled samples. Bioinformatics. 2009;25(17):2283–2285.  293. Koboldt DC, Zhang Q, Larson DE, et al. VarScan 2: Somatic mutation and copy number alteration discovery in cancer by exome sequencing. Genome Res. 2012;22(3):568–576.  294. Kridel R, Meissner B, Rogic S, et al. Whole transcriptome sequencing reveals recurrent NOTCH1 mutations in mantle cell lymphoma. Blood. 2012;119(9):1963–71.  121 295. Gray RJ. Weighted analyses for cohort sampling designs. Lifetime Data Anal. 2009;15(1):24–40.  296. Scott DW, Mottok A, Ennishi D, et al. Prognostic Significance of Diffuse Large B-Cell Lymphoma Cell of Origin Determined by Digital Gene Expression in Formalin-Fixed Paraffin-Embedded Tissue Biopsies. J. Clin. Oncol. 2015;  297. van Dongen JJM, Langerak AW, Brüggemann M, 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. 2003;17(12):2257–2317.  298. Lenz G, Davis RE, Ngo VN, et al. Oncogenic CARD11 mutations in human diffuse large B cell lymphoma. Science. 2008;319(5870):1676–9.  299. Davis RE, Ngo VN, Lenz G, et al. Chronic active B-cell-receptor signalling in diffuse large B-cell lymphoma. Nature. 2010;463(7277):88–92.  300. Ngo VN, Young RM, Schmitz R, et al. Oncogenically active MYD88 mutations in human lymphoma. Nature. 2011;470(7332):115–9.  301. Bohers E, Mareschal S, Bouzelfen A, et al. Targetable activating mutations are very frequent in GCB and ABC diffuse large B-cell lymphoma. Genes Chromosom. Cancer. 2014;53(2):144–153.  302. Klein U, Casola S, Cattoretti G, et al. Transcription factor IRF4 controls plasma cell differentiation and class-switch recombination. Nat. Immunol. 2006;7(7):773–782.  303. Ochiai K, Maienschein-Cline M, Simonetti G, et al. Transcriptional Regulation of Germinal Center B and Plasma Cell Fates by Dynamical Control of IRF4. Immunity. 2013;38(5):918–929.  304. Willis SN, Good-Jacobson KL, Curtis J, et al. Transcription factor IRF4 regulates germinal center cell formation through a B cell-intrinsic mechanism. J. Immunol. 2014;192(7):3200–6.  305. Natkunam Y, Warnke RA, Montgomery K, Falini B, van De Rijn M. Analysis of MUM1/IRF4 protein expression using tissue microarrays and immunohistochemistry. Mod. 122 Pathol. 2001;14(7):686–694.  306. Sweetenham JW, Goldman B, LeBlanc ML, et al. Prognostic value of regulatory T cells, lymphomaassociated macrophages, and MUM-1 expression in follicular lymphoma treated before and after the introduction of monoclonal antibody therapy: A southwest oncology group study. Ann. Oncol. 2009;21(6):1196–1202.  307. Xerri L, Bachy E, Fabiani B, et al. Identification of MUM1 as a prognostic immunohistochemical marker in follicular lymphoma using computerized image analysis. Hum. Pathol. 2014;45(10):2085–93.  308. Yang Y, Shaffer AL, Emre NCT, et al. Exploiting synthetic lethality for the therapy of ABC diffuse large B cell lymphoma. Cancer Cell. 2012;21(6):723–37.  309. Madsen C, Pedersen MB, Vase MO, et al. Outcome determinants for transformed indolent lymphomas treated with or without autologous stem-cell transplantation. Ann Oncol. 2015;26(2):393–399.  310. Cunningham D, Hawkes EA, Jack A, et al. Rituximab plus cyclophosphamide, doxorubicin, vincristine, and prednisolone in patients with newly diagnosed diffuse large B-cell non-Hodgkin lymphoma: A phase 3 comparison of dose intensification with 14-day versus 21-day cycles. Lancet. 2013;381(9880):1817–1826.  311. Costello M, Pugh TJ, Fennell TJ, et al. Discovery and characterization of artifactual mutations in deep coverage targeted capture sequencing data due to oxidative DNA damage during sample preparation. Nucleic Acids Res. 2013;41(6):e67.  312. Ding J, Bashashati A, Roth A, et al. Feature-based classifiers for somatic mutation detection in tumour-normal paired sequencing data. Bioinformatics. 2012;28(2):167–75.  313. Saunders CT, Wong WSW, Swamy S, et al. Strelka: accurate somatic small-variant calling from sequenced tumor-normal sample pairs. Bioinformatics. 2012;28(14):1811–7.  314. Ha G, Roth A, Lai D, 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. 2012;22(10):1995–2007.  315. Ha G, Roth A, Khattra J, et al. TITAN: inference of copy number architectures in clonal 123 cell populations from tumor whole-genome sequence data. Genome Res. 2014;24(11):1881–93.  316. Koressaar T, Remm M. Enhancements and modifications of primer design program Primer3. Bioinformatics. 2007;23(10):1289–1291.  317. Untergasser A, Cutcutache I, Koressaar T, et al. Primer3-new capabilities and interfaces. Nucleic Acids Res. 2012;40(15):1–12.  318. James Kent W, Sugnet CW, Furey TS, et al. The human genome browser at UCSC. Genome Res. 2002;12(6):996–1006.  319. Roth A, Khattra J, Yap D, et al. PyClone: statistical inference of clonal population structure in cancer. Nat. Methods. 2014;11(4.):  320. Malikic S, McPherson  a. W, Donmez N, Sahinalp CS. Clonality inference in multiple tumor samples using phylogeny. Bioinformatics. 2015;31(January):1349–1356.  321. Wong TN, Ramsingh G, Young AL, et al. Role of TP53 mutations in the origin and evolution of therapy-related acute myeloid leukaemia. Nature. 2014;518(7540):552–555.  322. Love C, Sun Z, Jima D, et al. The genetic landscape of mutations in Burkitt lymphoma. Nat. Genet. 2012;44(12):1321–5.  323. Richter J, Schlesner M, Hoffmann S, et al. Recurrent mutation of the ID3 gene in Burkitt lymphoma identified by integrated genome, exome and transcriptome sequencing. Nat. Genet. 2012;44(12):1316–20.  324. Schmitz R, Young RM, Ceribelli M, et al. Burkitt lymphoma pathogenesis and therapeutic targets from structural and functional genomics. Nature. 2012;490(7418):116–20.  325. Lawrence MS, Stojanov P, Polak P, et al. Mutational heterogeneity in cancer and the search for new cancer-associated genes. Nature. 2013;499(7457):214–8.  326. Rossi D, Khiabanian H, Spina V, et al. Clinical impact of small TP53 mutated subclones in chronic lymphocytic leukemia. Blood. 2014;123(14):2139–2147.  327. Challa-Malladi M, Lieu YK, Califano O, et al. Combined genetic inactivation of β2-Microglobulin and CD58 reveals frequent escape from immune recognition in diffuse large 124 B cell lymphoma. Cancer Cell. 2011;20(6):728–40.  328. Leung-Hagesteijn C, Erdmann N, Cheung G, et al. Xbp1s-negative tumor B cells and pre-plasmablasts mediate therapeutic proteasome inhibitor resistance in multiple myeloma. Cancer Cell. 2013;24(3):289–304.  329. Greaves M, Maley CC. Clonal evolution in cancer. Nature. 2012;481(7381):306–313.  330. Gould SJ, Eldredge N. Punctuated equilibrium comes of age. Nature. 1993;366(6452):223–227.  331. Heppner GH. Tumor heterogeneity. Cancer Res. 1984;44(6):2259–2265.  332. McGranahan N, Favero F, de Bruin EC, et al. Clonal status of actionable driver events and the timing of mutational processes in cancer evolution. Sci. Transl. Med. 2015;7(283):283ra54–283ra54.  333. Andor N, Graham TA, Jansen M, et al. Pan-cancer analysis of the extent and consequences of intratumor heterogeneity. Nat. Med. 2015;(November):1–12.  334. Mullighan CG, Phillips LA, Su X, et al. Genomic analysis of the clonal origins of relapsed acute lymphoblastic leukemia. Science. 2008;322(5906):1377–80.  335. Ma X, Edmonson M, Yergeau D, et al. Rise and fall of subclones from diagnosis to relapse in pediatric B-acute lymphoblastic leukaemia. Nat. Commun. 2015;6:6604.  336. Greaves M. Darwin and evolutionary tales in leukemia. The Ham-Wasserman Lecture. Hematology Am. Soc. Hematol. Educ. Program. 2009;(October 1838):3–12.  337. Scott DW. Cell-of-Origin in Diffuse Large B-Cell Lymphoma: Are the Assays Ready for the Clinic? Am. Soc. Clin. Oncol. Educ. Book. 2015;35:e458–66.  338. Relander T, Johnson N a, Farinha P, et al. Prognostic factors in follicular lymphoma. J. Clin. Oncol. 2010;28(17):2902–13.  339. Leonard JP, LaCasce AS, Smith MR, et al. Selective CDK4/6 inhibition with tumor responses by PD0332991 in patients with mantle cell lymphoma. Blood. 2012;119(20):4597–4607.  340. Schwartz GK, LoRusso PM, Dickson MA, et al. Phase I study of PD 0332991, a cyclin-125 dependent kinase inhibitor, administered in 3-week cycles (Schedule 2/1). Br. J. Cancer. 2011;104(12):1862–8.  341. Ortmann C a, Kent DG, Nangalia J, et al. Effect of Mutation Order on Myeloproliferative Neoplasms. N. Engl. J. Med. 2015;372(7):601–612.  342. Roschewski M, Dunleavy K, Pittaluga S, et al. Circulating tumour DNA and CT monitoring in patients with untreated diffuse large B-cell lymphoma: a correlative biomarker study. Lancet. Oncol. 2015;2045(15):1–9.  343. Kurtz DM, Green MR, Bratman S V, et al. Non-invasive monitoring of diffuse large B-cell lymphoma by immunoglobulin high-throughput sequencing. Blood. 2015;125(24):3679–3688.  344. Kreso A, Dick JE. Evolution of the cancer stem cell model. Cell Stem Cell. 2014;14(3):275–91.  345. Kreso A, O’Brien CA, van Galen P, et al. Variable clonal repopulation dynamics influence chemotherapy response in colorectal cancer. Science. 2013;339(6119):543–8.  346. Pisco AO, Brock A, Zhou J, et al. Non-Darwinian dynamics in therapy-induced cancer drug resistance. Nat. Commun. 2013;4:2467.     126 Appendices  Appendix A  - 86 gene panel for capture sequencing Gene Name Criterion1 Criterion2 Criterion3 Criterion4 Criterion5 Other ACTB FALSE TRUE FALSE FALSE FALSE FALSE ARHGEF1 FALSE TRUE FALSE FALSE FALSE FALSE ARID1A TRUE FALSE TRUE TRUE FALSE FALSE ARID1B TRUE FALSE FALSE FALSE FALSE FALSE ATP6AP1 TRUE FALSE FALSE FALSE FALSE FALSE ATP6V1B2 TRUE FALSE TRUE FALSE FALSE FALSE B2M TRUE TRUE TRUE TRUE TRUE FALSE BCL10 FALSE FALSE TRUE FALSE FALSE FALSE BCL2 TRUE TRUE FALSE TRUE FALSE FALSE BCL6 TRUE TRUE TRUE TRUE FALSE FALSE BCL7A TRUE FALSE TRUE FALSE FALSE FALSE BCR FALSE TRUE FALSE FALSE FALSE FALSE BTG1 TRUE TRUE FALSE FALSE FALSE FALSE BTG2 FALSE TRUE FALSE FALSE FALSE FALSE CARD11 TRUE TRUE TRUE FALSE FALSE FALSE CCL23 FALSE FALSE TRUE FALSE FALSE FALSE CCND3 TRUE TRUE TRUE TRUE FALSE FALSE CD58 FALSE TRUE TRUE FALSE FALSE FALSE CD70 FALSE TRUE FALSE FALSE FALSE FALSE CD79B FALSE TRUE TRUE FALSE FALSE FALSE CD83 FALSE TRUE FALSE FALSE FALSE FALSE CHD8 FALSE FALSE FALSE TRUE FALSE FALSE CIITA FALSE TRUE FALSE FALSE FALSE FALSE CREBBP TRUE TRUE TRUE FALSE TRUE FALSE CTSS TRUE FALSE FALSE FALSE FALSE FALSE DNAH9 FALSE FALSE FALSE TRUE FALSE FALSE DTX1 TRUE FALSE TRUE FALSE FALSE FALSE EBF1 FALSE FALSE TRUE TRUE FALSE FALSE EBF3 FALSE FALSE FALSE FALSE TRUE FALSE EEF1A1 TRUE FALSE FALSE FALSE FALSE FALSE EP300 TRUE TRUE FALSE FALSE FALSE FALSE EVI2A FALSE FALSE TRUE FALSE FALSE FALSE EZH2 TRUE TRUE TRUE TRUE FALSE FALSE FAS FALSE FALSE TRUE FALSE TRUE FALSE FAT4 FALSE FALSE FALSE TRUE FALSE FALSE FOXO1 TRUE TRUE TRUE TRUE FALSE FALSE GNA13 TRUE TRUE TRUE FALSE FALSE FALSE GNAI2 TRUE FALSE TRUE FALSE FALSE FALSE HIST1H1B FALSE FALSE TRUE FALSE FALSE FALSE 127 Gene Name Criterion1 Criterion2 Criterion3 Criterion4 Criterion5 Other HIST1H1C FALSE TRUE FALSE FALSE FALSE FALSE HIST1H1E TRUE TRUE TRUE TRUE TRUE FALSE HIST1H2AM FALSE FALSE FALSE FALSE TRUE FALSE HLA-DMB FALSE FALSE TRUE FALSE FALSE FALSE HVCN1 FALSE FALSE TRUE FALSE FALSE FALSE ID3 FALSE TRUE FALSE FALSE FALSE FALSE IKZF3 TRUE FALSE FALSE FALSE FALSE FALSE IL4R FALSE FALSE FALSE FALSE FALSE TRUE IRF4 FALSE FALSE FALSE TRUE FALSE FALSE IRF8 TRUE TRUE TRUE FALSE TRUE FALSE ITPKB FALSE FALSE FALSE TRUE FALSE FALSE KLHL6 FALSE TRUE TRUE FALSE FALSE FALSE KMT2C TRUE FALSE FALSE FALSE FALSE FALSE KMT2D TRUE TRUE TRUE TRUE FALSE FALSE LRRC7 FALSE FALSE FALSE TRUE FALSE FALSE MEF2B TRUE TRUE TRUE TRUE FALSE FALSE MEF2C FALSE FALSE TRUE FALSE FALSE FALSE MKI67 FALSE TRUE FALSE FALSE FALSE FALSE MYC FALSE TRUE TRUE TRUE FALSE FALSE MYD88 FALSE TRUE FALSE FALSE FALSE FALSE MYOM2 FALSE TRUE FALSE FALSE FALSE FALSE NLRC5 FALSE TRUE FALSE FALSE FALSE FALSE NOTCH1 FALSE TRUE FALSE FALSE FALSE FALSE NOTCH2 FALSE FALSE FALSE FALSE FALSE TRUE P2RY8 FALSE TRUE FALSE FALSE FALSE FALSE PIM1 TRUE TRUE FALSE TRUE FALSE FALSE POU2AF1 FALSE FALSE TRUE FALSE FALSE FALSE PTPN1 FALSE FALSE FALSE FALSE FALSE TRUE RFX5 FALSE FALSE FALSE FALSE FALSE TRUE RHOA FALSE TRUE FALSE FALSE FALSE FALSE RRAGC FALSE FALSE TRUE FALSE TRUE FALSE S1PR2 FALSE FALSE TRUE FALSE FALSE FALSE SGK1 FALSE TRUE TRUE TRUE FALSE FALSE SMARCA4 TRUE TRUE FALSE FALSE FALSE FALSE SOCS1 FALSE FALSE TRUE FALSE FALSE FALSE STAT3 FALSE TRUE FALSE FALSE FALSE FALSE STAT6 TRUE FALSE TRUE FALSE FALSE FALSE TCF3 FALSE TRUE FALSE FALSE FALSE FALSE TLR2 FALSE FALSE FALSE TRUE FALSE FALSE TMEM30A FALSE TRUE FALSE FALSE FALSE FALSE TNFAIP3 TRUE TRUE TRUE FALSE FALSE FALSE TNFRSF14 TRUE TRUE TRUE TRUE FALSE FALSE TP53 TRUE TRUE TRUE TRUE FALSE FALSE UNC5C FALSE TRUE FALSE FALSE FALSE FALSE 128 Gene Name Criterion1 Criterion2 Criterion3 Criterion4 Criterion5 Other UPF1 FALSE FALSE TRUE FALSE FALSE FALSE XBP1 FALSE FALSE TRUE FALSE FALSE FALSE ZFP36L1 FALSE TRUE FALSE FALSE FALSE FALSE   129 Appendix B  - 20 somatic hypermutation gene panel for capture sequencing Gene Name chr tanscription start site strand region targeted by capture and analyzed BACH2 6 91006628 - chr6:91002628-91007628 BCL2 18 60987362 - chr18:60983362-60988362 BCL6 3 187463516 - chr3:187459516-187464516 BCL7A 12 122457328 + chr12:122456328-122461328 BCR 22 23521891 + chr22:23520891-23525891 BPTF 17 65821640 + chr17:65820640-65825640 BTG1 12 92539674 - chr12:92535674-92540674 BTG2 1 203274619 + chr1:203273619-203278619 CD83 6 14117872 + chr6:14116872-14121872 CIITA 16 10971055 + chr16:10970055-10975055 CXCR4 2 136875736 - chr2:136871736-136876736 IRF8 16 85932409 + chr16:85931409-85936409 LTB 6 31550300 - chr6:31546300-31551300 MYC 8 128747680 + chr8:128746680-128751680 PAX5 9 37027120 - chr9:37023120-37028120 PIM1 6 37137979 + chr6:37136979-37141979  RHOH 4 40198526 + chr4:40197526-40202526 SOCS1 16 11350037 - chr16:11346037-11351037 TCL1A 14 96180534 - chr14:96176534-96181534 TMSB4X X 12995227 + chrX:12994227-12999227  

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.24.1-0228776/manifest

Comment

Related Items