Enhanced Reduced Representation Bisulfite Sequencing Research Articles

Perinatal Lead Exposure Promotes Sex-Specific Epigenetic Programming of Disease-Relevant Pathways in Mouse Heart.

Environmental contaminants such as the metal lead (Pb) are associated with cardiovascular disease, but the underlying molecular mechanisms are poorly understood. In particular, little is known about how exposure to Pb during early development impacts the cardiac epigenome at any point across the life course and potential differences between sexes. In a mouse model of human-relevant perinatal exposures, we utilized RNA-seq and Enhanced Reduced Representation Bisulfite Sequencing (ERRBS) to investigate the effects of Pb exposure during gestation and lactation on gene expression and DNA methylation, respectively, in the hearts of male and female mice at weaning. For ERRBS, we identified differentially methylated CpGs (DMCs) or differentially methylated 1000 bp regions (DMRs) based on a minimum absolute change in methylation of 10% and an FDR < 0.05. For gene expression data, an FDR < 0.05 was considered significant. No individual genes met the FDR cutoff for gene expression; however, we found that Pb exposure leads to significant changes in the expression of gene pathways relevant to cardiovascular development and disease. We further found that Pb promotes sex-specific changes in DNA methylation at hundreds of gene loci (280 DMCs and 99 DMRs in males, 189 DMCs and 121 DMRs in females), and pathway analysis revealed that these CpGs and regions collectively function in embryonic development. In males, differential methylation also occurred at genes related to immune function and metabolism. We then investigated whether genes exhibiting differential methylation at weaning were also differentially methylated in hearts from a cohort of Pb-exposed mice at adulthood. We found that a single gene, Galnt2, showed differential methylation in both sexes and time points. In a human cohort investigating the influence of prenatal Pb exposure on the epigenome, we also observed an inverse association between first trimester Pb concentrations and adolescent blood leukocyte DNA methylation at a locus in GALNT2, suggesting that this gene may represent a biomarker of Pb exposure across species. Together, these data, across two time points in mice and in a human birth cohort study, collectively demonstrate that Pb exposure promotes sex-specific programming of the cardiac epigenome, and provide potential mechanistic insight into how Pb causes cardiovascular disease.

Sex-Specific Alterations in Cardiac DNA Methylation in Adult Mice by Perinatal Lead Exposure.

Environmental factors play an important role in the etiology of cardiovascular diseases. Cardiovascular diseases exhibit marked sexual dimorphism; however, the sex-specific effects of environmental exposures on cardiac health are incompletely understood. Perinatal and adult exposures to the metal lead (Pb) are linked to several adverse cardiovascular outcomes, but the sex-specific effects of this toxicant on the heart have received little attention. Perinatal environmental exposures can lead to disease through disruption of the normal epigenetic programming that occurs during early development. Using a mouse model of human-relevant perinatal environmental exposure, we investigated the effects of exposure to Pb during gestation and lactation on DNA methylation in the hearts of adult offspring mice (n = 6 per sex). Two weeks prior to mating, dams were assigned to control or Pb acetate (32 ppm) water, and exposure continued until offspring were weaned at three weeks of age. Enhanced reduced-representation bisulfite sequencing was used to measure DNA methylation in the hearts of offspring at five months of age. Although Pb exposure stopped at three weeks of age, we discovered hundreds of differentially methylated cytosines (DMCs) and regions (DMRs) in males and females at five months of age. DMCs/DMRs and their associated genes were sex-specific, with a small, but statistically significant subset overlapping between sexes. Pathway analysis revealed altered methylation of genes important for cardiac and other tissue development in males, and histone demethylation in females. Together, these data demonstrate that perinatal exposure to Pb induces sex-specific changes in cardiac DNA methylation that are present long after cessation of exposure, and highlight the importance of considering sex in environmental epigenetics and mechanistic toxicology studies.

Analysis of the Global Methylation Profile of Accelerated and Blast Phase Myeloproliferative Neoplasms and Its Association with Response to Decitabine-Based Therapy

Analysis of the Global Methylation Profile of Accelerated and Blast Phase Myeloproliferative Neoplasms and Its Association with Response to Decitabine-Based Therapy

Tissue- and Sex-Specific DNA Methylation Changes in Mice Perinatally Exposed to Lead (Pb).

Lead (Pb) is a well-known toxicant that interferes with the development of a child’s nervous and metabolic systems and increases the risk of developing diseases later in life. Although studies have investigated epigenetic effects associated with Pb exposure, knowledge of genome-wide changes with in vivo low dose perinatal Pb exposure in multiple tissues is limited. Within the Toxicant Exposures and Responses by Genomic and Epigenomic Regulators of Transcription (TaRGET II) consortium, we utilized a mouse model to investigate tissue- and sex-specific DNA methylation. Dams were assigned to control or Pb-acetate water, respectively. Exposures started 2 weeks prior to mating and continued until weaning at post-natal day 21 (PND21). Liver and blood were collected from PND21 mice, and the DNA methylome was assessed using enhanced reduced representation bisulfite sequencing (ERRBS). We identified ∼1000 perinatal Pb exposure related differentially methylated cytosines (DMCs) for each tissue- and sex-specific comparison, and hundreds of tissue- and sex-specific differentially methylated regions (DMRs). Several mouse imprinted genes were differentially methylated across both tissues in males and females. Overall, our findings demonstrate that perinatal Pb exposure can induce tissue- and sex-specific DNA methylation changes and provide information for future Pb studies in humans.

184-OR: In Utero Exposure to Gestational Diabetes Alters the Transcriptome and Methylome of Human Fetal Stem Cells Revealing an Enrichment of Interferon Related Pathways

Background: Gestational diabetes (GDM) has profound effects on the intrauterine metabolic milieu, induces marked abnormalities in fetal glucose and insulin secretion and is linked to obesity and diabetes in the offspring, but the mechanisms remain largely unknown. Objective: To measure changes in gene expression and genome wide DNA methylation in human amniocytes, a fetal stem cell, exposed to GDM in utero. Methods: EdgeR identified differentially expressed genes from RNASeq via fold change, p-values, and q-values calculated after Bonferoni correction (n=8; 4 per sex). RNA-Seq results were confirmed via QPCR. Ingenuity Pathway Analysis (IPA) identified enriched biological pathways. Differentially methylated regions (DMRs) were determined from Enhanced Reduced Representation Bisulfite Sequencing (ERRBS) (n=16; 8 per sex) by identifying sequential CpGs with significant changes in DNA methylation &amp;gt;5%, and with p&amp;lt;0.05 over the entire DMR. Results: We identified 20 differentially expressed genes (q&amp;lt;0.10) analyzing data from male and female amniocytes together, but only 4 genes in male only and 2 in female only analyses. Using a significance threshold of p&amp;lt;0.01, we identified 65 differentially expressed genes analyzing all samples together, 46 genes in the male only and 68 in the female only analyses. QPCR confirmed increases in IFI44, ULBP1 and SAMD9L. Differentially expressed genes were strongly enriched for interferon inducible proteins, a novel fetal programming pathway after in utero GDM exposure. IPA showed enrichment in molecular mechanisms of growth, oxidative stress metabolism and GCPR signaling pathways. Offspring sex-specific analysis greatly enhanced DMR identification. Nine DMRs were identified in all, 41 in male and 20 in female samples. Conclusion: Exposure to GDM in utero leads changes in amniocyte gene expression and DNA methylation with enrichment in interferon inducible pathways. Disclosure S.E. Pinney: None. A. Joshi: None. V. Yin: None. S. Min: None. D.E. Condon: None. Z. Wang: None. Funding National Institute of Diabetes and Digestive and Kidney Diseases (K08DK090302); National Institute of Environmental Health Sciences (P30ES013508); National Institutes of Health (UL1TR001878)

Comprehensive Integration of Whole Genome, Transcriptome and Methylation Profiling Reveals Novel Gene Dysregulation Including IL15, SOCS6 and CARD11 Associated with MYD88 and CXCR4 Genotype Status in WM

Background: Waldenstrom's Macroglobulinemia/IgM lyphoplasmacytic lymphoma is a B-cell lymphoma defined by highly recurring mutations in MYD88 (95-97%) and CXCR4 (30-40%) by whole genome, PCR and Sanger sequencing (Treon et al NEJM 2012, Hunter et al Blood 2013, Xu et al, BJH 2016). Using next generation RNA sequencing studies (RNASeq) of these same group of patients, we subsequently observed that MYD88 and CXCR4 mutation status were the primary determinants of differential gene expression in WM (Hunter et al, Blood 2016). We have now integrated enhanced reduced representation bisulfite sequencing (ERRBS) data into the existing WGS and RNASeq data to present of more complete picture of genomic regulation in WM.Methods: CD19+ selected bone marrow samples from 54 patients with WM, and CD19- peripheral blood mononuclear samples were used for tumor and germline controls, respectively. WGS and RNASeq were conducted as previously described (Hunter et al, Blood 2014; 2016) and methylation profiling was conducted using ERRBS at the Weil Cornell Medical Epigenomics Core and analyzed using Bismark. Results were filtered for methylation sites supported by at least 10 reads across all of the samples. Differential methylation analysis was conducted at the individual site level and aggregated promoter level using an established edgeR protocol. Differential methylation data was then compared with Salmon derived gene and isoform expression data which was available for 49/54 (91%) of the samples.Results: Clinical characteristics of patients were as follows: Median age 60 (range 40-87 years); BM involvement 43% (range 4-95%); serum IgM 3590 (range 416-8320 ug/dl). Most patients (67%) were untreated. WGS data revealed 33 (61%), 18 (33%), and 3 (6%) were MYD88Mutant/CXCR4Wild-Type, MYD88Mutant/CXCR4Mutant and MYD88Wild-Type/CXCR4Wild-Type, respectively. As was observed in our RNASeq analysis, pairwise multidimensional scaling of the methylation status of the top 2,000 high variance promoters revealed segregation of the MYD88Mutant/CXCR4Wild-Type from both MYD88Mutant/CXCR4Mutant and MYD88Wild-Type/CXCR4Wild-Type (Figure 1A). Log fold change in methylation in differentially methylated promoters from MYD88Mutant/CXCR4Mutant and MYD88Wild-Type/CXCR4Wild-Type samples relative to MYD88Mutant/CXCR4Wild-Type were negatively correlated with log-fold change in gene expression (Rho = -0.463, p <0.0001; Figure 1B). Promoter level analysis revealed 556 differentially methylated promoters in MYD88Mutant/CXCR4Mutant of which 440 (78%) had increased levels of methylation relative to MYD88Mutant/CXCR4Wild-Type samples. Affected genes included IL15, GNAO1, HIF1A, SOCS6, PIK3R5, IRF8 and CD38. For MYD88Wild-Type/CXCR4Wild-Type samples, 126 promoters were significantly differentially methylated with only 28 (22%) demonstrating increased methylation relative to MYD88Mutant/CXCR4Wild-Type samples. Affected promoters included NRIP1, FNBP1L, and PTK2. Single site analysis revealed 127 and 1 significantly differentially methylated CpG loci for MYD88Mutant/CXCR4Mutant and MYD88Wild-Type/CXCR4Wild-Type, respectively. This analysis revealed a large intronic CpG island in CARD11 that was demethylated only in MYD88Mutant/CXCR4Wild-Type patients and a methylated site in the PPP1CC promoter found in MYD88Wild-Type/CXCR4Wild-Type.Conclusions: The studies provide the first comprehensive insights into epigenomic regulation of WM and show that MYD88 and CXCR4 mutation status drives methylation and confers a distinct transcriptomic profile that includes genes such as IL15, SOCS6, CARD11, HIF1A, and PIK3R5 with important pro-survival and immune regulatory roles. [Display omitted] DisclosuresHunter:Pharmacyclics: Consultancy. Castillo:Abbvie: Consultancy, Research Funding; Beigene: Consultancy, Research Funding; Genentech: Consultancy; Janssen: Consultancy, Research Funding; Millennium: Research Funding; Pharmacyclics: Consultancy, Research Funding. Treon:Pharmacyclics: Consultancy, Other: Travel, Accommodations, Expenses, Research Funding; BMS: Research Funding; Johnson & Johnson: Consultancy; Janssen: Consultancy, Other: Travel, Accommodations, Expenses.

Epigenetic alterations to Polycomb targets precede malignant transition in a mouse model of breast cancer

Malignant breast cancer remains a major health threat to women of all ages worldwide and epigenetic variations on DNA methylation have been widely reported in cancers of different types. We profiled DNA methylation with ERRBS (Enhanced Reduced Representation Bisulfite Sequencing) across four main stages of tumor progression in the MMTV-PyMT mouse model (hyperplasia, adenoma/mammary intraepithelial neoplasia, early carcinoma and late carcinoma), during which malignant transition occurs. We identified a large number of differentially methylated cytosines (DMCs) in tumors relative to age-matched normal mammary glands from FVB mice. Despite similarities, the methylation differences of the premalignant stages were distinct from the malignant ones. Many differentially methylated loci were preserved from the first to the last stage throughout tumor progression. Genes affected by methylation gains were enriched in Polycomb repressive complex 2 (PRC2) targets, which may present biomarkers for early diagnosis and targets for treatment.

Aldehyde Dehydrogenase Activity in the Leukemic Stem Cell Compartment Uncovers Opposing Methylation Patterns of Leukemia Stem Cells in AML

Acute myeloid leukemia (AML) is a disease marked by abnormal differentiation of the myeloid cell lineage. Leukemia stem cells differentiate to give rise to leukemic progenitor cells (LPC) and ultimately leukemic blasts which are not leukemia-initiating. Previous studies have revealed a diverse methylation landscape in AML, but have mostly relied on the blast population rather than more purified primitive populations. Consequently, the status of the leukemic methylome during expansion from LSC to LPC and finally blast remains largely unknown. We sorted HSC from the bone marrow of normal donors and LSC, LPC, and blasts from AML patients based on expression of CD34, CD38, CD45 and aldehyde dehydrogenase (ALDH) activity. Normal HSC were defined as CD34+CD38-ALDHhigh. In AML patients, LSC were either CD34+CD38-ALDHhigh or CD34+CD38-ALDHmid; LPC were CD34+CD38+; the blast population consisted of unsorted mononuclear cells. Though patients with an ALDHhigh LSC profile may have had some residual normal HSCs present, contribution of these cells was likely minimal and thus the overall population predominantly leukemic. Methylation profiles for each cell fraction in eight untreated AML patients and five normal donors were generated using the enhanced reduced representation bisulfite sequencing (ERRBS) assay. All sequenced ERRBS libraries were aligned against the human genome (hg19) and organized into 25 base pair tiles for analysis of differentially methylated regions (DMR) using a beta-binomial model that takes variation across samples into account during DMR identification. DMR classification required a difference in methylation of >25% and false discovery rate (FDR) < 10%. Unsupervised correspondence analysis indicated that methylomes of two patients with ALDHhigh LSC were distinct from the six patients with ALDHmid LSC and therefore patients were grouped based on the ADLH activity of their LSCs for comparisons. Both ALDHhigh LSC and ALDHmid LSC had extensive alterations in methylation across their genomes when compared to HSC. The great majority of DMRs in ALDHhigh LSC were hypomethylated; of the 62,415 DMRs identified, 55,418 regions were hypomethylated while only 6,997 were hypermethylated in ALDHhigh LSC. In contrast, in ALDHmid LSC, 39,162 DMRs were hypermethylated and 5,408 regions hypomethylated compared to HSC. Despite opposing patterns of methylation, DMRs were enriched at intergenic and intragenic enhancers in both ALDHhigh and ALDHmid LSC. DMRs were functionally annotated to gene sets in the MSigDb database. Genes associated with ALDHhigh DMRs were enriched for genes with the binding motif for transcription factor Sp1 near their promoters (FDR = 2.55×10-79) and ALDHmid DMR associated genes were enriched for genes with H3K27 trimethylation in their promoters (FDR = 7.19×10-169). We compared methylation profiles of LSC to LPC and blasts in an effort to determine whether changes to the methylome occur with leukemic maturation. Interestingly, there were no significant changes in methylation between LSC and LPC in either ALDH population. However, we did see changes to the epigenome emerge when LSC or LPC were compared to leukemic blasts. In ALDHhigh patients, LSC and LPC were more hypomethylated than blasts while ALDHmid LSC and LPC were more hypermethylated than the more differentiated blasts. In conclusion, alterations to the LSC methylome were extensive and two patterns of methylation emerged based on the ALDH activity of LSC; ALDHhigh LSC displayed hypomethylated profiles and ALDHmid LSC were hypermethylated. Enrichment of DMRs at intra and intergenic enhancer regions in both LSC types despite their opposing methylation patterns highlights the importance of epigenetic marks in these regions and their role as regulators of gene expression. Significant changes in methylation between LSC or LPC and blasts, but not LSC and LPC suggest relative stability of the methylome during early leukemic differentiation with more substantial alterations occurring after the LPC level. DisclosuresGerber:Janssen: Research Funding; Alexion: Membership on an entity's Board of Directors or advisory committees; Spectrum: Membership on an entity's Board of Directors or advisory committees; Seattle Genetics: Membership on an entity's Board of Directors or advisory committees. Carraway:Incyte: Membership on an entity's Board of Directors or advisory committees; Amgen: Membership on an entity's Board of Directors or advisory committees; Celgene Corporation: Research Funding, Speakers Bureau; Novartis: Membership on an entity's Board of Directors or advisory committees; Baxalta: Speakers Bureau. Gore:Celgene: Consultancy, Honoraria, Research Funding.

Molecular Events Defining Human Clonal Hematopoiesis at Single Cell Resolution

For decades, hematopoiesis has been described as a cellular hierarchy maintained by self-renewing hematopoietic stem cells (HSCs) that proceed through a series of multipotent, oligopotent and then unipotent progenitors to make blood. This can be regarded as the standard "textbook" model that is widely used as the basis to interpret molecular regulatory processes, disease processes, and develop therapeutic approaches. The presence of oligopotent intermediates is crucial to the model since they define the path from multipotent cells to unipotent cells. Common myeloid progenitors (CMPs) represent the critical oligopotent progenitor from which all My (defined as granulocyte/monocyte), erythroid (Er), and megakaryocyte (Mk) cells arise. Although the standard model is still used extensively as an operational paradigm, further cell purification and functional clonal assays have led to key revisions of the model. Through the development of more efficient assays to monitor My and lymphoid fates in single cell stromal assays and an improved sorting scheme, we previously identified human multilymphoid progenitors (MLP) as the earliest lymphoid differentiation precursor with concomitant lymphoid (T, B, NK) and myelomonocytic potential, rather than common lymphoid progenitors (CLP) (Doulatov et al, Nature Immunology 2010). Murine studies came to the same conclusion. These results raise the question of whether the opposite arm of this lineage split (My, Er, and Mk fate) is also more complex than previously thought. Indeed strong evidence has emerged from murine studies for revision to the view of murine My-Mk-Er lineage development. Until recently considerable uncertainty remained concerning the myelo-erythroid branch of human hematopoiesis. Through a novel cell-sorting scheme and a more sensitive assay to assess multilineage My, Er, and Mk fate potential contemporaneously from single cells we have provided a new framework to understand normal and disease states of human hematopoiesis (Notta et al Science 2016). These studies suggest that oligopotent progenitors are a negligible component of the human hematopoietic hierarchy of adult bone marrow. Rather, multipotent cells differentiate into unipotent cells of the My-Er-Mk lineages directly support the concept of a 'two-tier' human blood hierarchy composed of two-tiers: a top-tier containing multipotent cells such as HSCs and MPPs, and a bottom-tier composed of progenitors committed directly to My, Er, or Mk lineages. Second, the blood hierarchy is not identical across development. In human fetal liver (FL), oligopotent progenitors with My-Er-Mk and Er-Mk activity were a prominent component of the hierarchy. By contrast, the adult bone marrow (BM) was dominated by unilineage progenitors with primarily My or Er potential. Finally, our study has documented multiple origins of where Mk cells arise. We found that Mk branching differs in FL and BM. In FL, Mk progenitors were enriched, but not restricted, to the stem cell compartment; while in BM, Mk fate was closely tied to multipotent cells. This result fits the two-tier model of adult hematopoiesis since branching of Mk occurs in the top tier, at the level of HSC/MPPs. To gain a deeper understanding of the molecular basis for the two tier hierarchy, we have undertaken low cell input RNA sequencing, Enhanced Reduced Representation Bisulfite Sequencing (ERRBS), and ATAC-seq to provide a comprehensive transcriptional and epigenetic roadmap of human HSPC across development.

A Pharmacologically Reversible Epigenomic Signature of Relapse and Chemotherapy Resistance in Anaplastic Large Cell Lymphoma

Anaplastic large cell lymphoma (ALCL) are orphan entities representing around 20% of all Peripheral T-cell lymphoma. ALCLs can be stratified into primary cutaneous ALCL and systemic ALCL. Patients with systemic ALCL are further distinguished based on anaplastic lymphoma kinase (ALK) translocations. Unlike ALK+ patients, ALK- ALCLs are more likely to be refractory or relapse upon treatment with CHOP (Cyclophosphamide, Hydroxydaunorubicin, Oncovin, Prednisone). A small number of genetic defects and genomic translocations have been identified and associated with clinical outcome in ALK- ALCL, but cannot fully explain the poor outcomes of ALK- patients. To further probe the molecular bases of ALCL, we performed systematic epigenomic profiling in systemic ALCL through Enhanced Reduced Representation Bisulfite Sequencing (ERRBS).To investigate the methylation profile of systemic ALCLs, we first interrogated 8 ALK+ ALCL and 16 ALK- ALCL tumors. Combining methylation levels from all interrogated CpG sites, we observed overall hypermethylation at CpG-islands (CGIs; p=0.02, t-test) in ALK- tumors compared ALK+ patients but hypomethylation outside of CGIs (>10kb away from known CGIs; p=0.03, t-test). There were no significant differences on the average methylation levels at CpG-island shores, which indicate varied methylation profiles among different genomic locations. To define methylation-based signatures, we then performed supervised analysis between ALK+ and ALK- patients. We have identified 254 ALK- patients linked consistently differentially methylated promoters (>20% DNA methylation changes and FDR<0.1, t-test). Out of 254 promoters, 252 genes were hypermethylated in ALK- patients confirming the association between systemic ALCL subtypes and hypermethylation.We then sought to investigate how methylation evolves in ALCL. We performed ERRBS of 2 primary ALCL tumors and their matched relapse samples (2 for each patient). We observed a significant hypermethylation at CGIs in relapsed samples (p=1.4e-5, paired t-test), suggesting that hypermethylation changes may be associated with tumor evolution.To test whether the hypermethylation of specific genes is linked to tumor aggressiveness, we analyzed serial passages of patient-derived tumor xenograft (PDTX) derived from primary and relapse ALCL samples. We observed consistently hypermethylation at CGIs along different implants and conserved hypermethylation of genes at the same time (p<0.05, 1959). To confirm the behavior of those genes, we estimate their methylation trend in ALCL primary and relapse samples with GSEA. Target genes showed a very strong hypermethylation in relapsed ALCL compared to primary samples (FDR<0.01). Next we confirmed analogous methylation changes in a cohort of primary and relapse PDTX (Primary, n=2, FDR<0.01; Relapse, n=3, FDR<0.01). Interestingly, when comparing this to our previous diffuse large b-cell lymphoma (DLBCL) study (Pan et.al., 2015), we found similar hypermethylated profiles, suggesting commonalities among different type of human lymphomas.Lastly, we investigated whether a treatment with the 5-azacytidine (5-Aza) could reverse the hypermethylation profile of ALCL PDTX. Tumors from in vivo 5-Aza treated and control mice were extracted and subjected to ERRBS. As expected, we only found global hypomethylation in 5-Aza-treated samples closely mimicked methylation profiles in naive primary ALCL and early passages of PDTX.In summary, this study provides the first comprehensive characterization of methylome in ALCL. Hypermethylation is firstly associated with the aggressiveness of ALK- patients in ALCL. Also, we identified a group of hypermethylated genes associated with relapsed ALCL, a phenotype mirrored by serial passages of PDTX in vivo. As 5-Aza treatments can revert the methylome of late passages of ALCL PDTX, we speculate that this treatment may represent a novel therapeutic approach in chemo-relapsed ALCL patients. DisclosuresMartin:Gilead: Consultancy, Other: travel, accommodations, expenses; Acerta: Consultancy; Novartis: Consultancy; Celgene: Consultancy, Honoraria; Teva: Research Funding; Janssen: Consultancy, Honoraria, Other: travel, accommodations, expenses.

Abstract 4454: Genome-wide analysis of DNA methylation induced by environmental carcinogen dibenzo[def,p]chrysene in ovarian tissues of mice

Abstract We and others have demonstrated that the administration of dibenzo[def,p]chrysene [also known as dibenzo[a,l]pyrene (DBP)], a representative example of the class of polycyclic aromatic hydrocarbon (PAH), by ip, oral gavage, or topical application onto the oral cavity induced tumors in multiple organ sites in mice; the ovary was the most susceptible tissue. We further demonstrated that the capacity of the target organs to metabolize DBP to active intermediates that can form DNA adducts may account for its tissue selective tumorigenicity. In addition, formation of DNA adducts by certain chemical carcinogens has been linked to aberrant DNA methylation, including the most extensively studied prototype PAH benzo[a]pyrene (B[a]P). The goal of this study is to examine whether alteration of DNA methylation occurs in the ovarian tissues of mice treated with DBP during the early stage of tumor development. In this study, we employed a previously established animal protocol in which the levels of DBP-DNA adducts as a function of time had been determined in the ovary following the oral administration of DBP (24 nmol, 3×/week for 5 weeks) or vehicle (DMSO) to female B6C3F1 mice (six weeks old, n = 3/group). DNA was isolated from ovary and subjected to enhanced reduced representation bisulfite sequencing (ERRBS) which is a single nucleotide resolution technique used to study DNA methylation in CpG sites and the surrounding regions. Briefly, DNA was digested by MspI followed by end repair, adenylation and adapter ligation with a modification of bead size selection to capture MspI fragments of 70-320 bp size. The resulting libraries were bisulfite-converted followed by PCR amplification and read by 1×50 bp on HiSeq 2500. Base calls of bisulfite treated sequencing reads were mapped to the mm9 mouse assembly and methylation calls were performed using Bismark v0.10.1 (Babraham Bioinformatcis, UK). The methylKit v0.9.2 R package was then used to calculate the differential methylation. Differentially methylated bases with q-value &amp;lt; 0.01 and percent methylation difference &amp;gt; 25% were extracted. Among 179 differentiated methylation sites (DMS) identified between DBP and vehicle-treated mice, 68 are hypermethylated and 111 are hypomethylated. About 25% of DMS are located in promoter or exon, 32% in intron and 44% in intergenic regions. DMS are located in genes including oocyte specific homeobox 2, transforming growth factor alpha, and tumor necrosis factor. Ingenuity Pathways analysis of the genes with altered methylation patterns identified top canonical pathways as growth hormone signaling, spermine biosynthesis, threonine degradation, trehalose degradation and L-serine degradation. Collectively, our previous results together with those presented here demonstrate that both genetic and epigenetic alterations may account for the carcinogenicity of DBP in the mouse ovary. Support: NIEHS R21ES020411 and NCI R01-CA173465. Citation Format: Yuan-Wan Sun, Karam El-Bayoumy, Yuka Imamura Kawasawa, Anna Salzberg, Cesar Aliaga, Krishnegowda Gowdahalli, Shantu Amin, Kun-Ming Chen. Genome-wide analysis of DNA methylation induced by environmental carcinogen dibenzo[def,p]chrysene in ovarian tissues of mice. [abstract]. In: Proceedings of the 107th Annual Meeting of the American Association for Cancer Research; 2016 Apr 16-20; New Orleans, LA. Philadelphia (PA): AACR; Cancer Res 2016;76(14 Suppl):Abstract nr 4454.

Next-Generation Bisulfite Sequencing of Aml Reveals Widespread Acquisition of Epigenetic Abnormalities in Leukemic Stem Cells That Are Stably Retained in More Mature Leukemic Cell Fractions

Acute myeloid leukemia (AML) is a disease of aberrant hematopoietic differentiation believed to mirror the hierarchical pattern of hematopoiesis with leukemia stem cells (LSC) serving as the originating cell population from which the tumor arises. Like hematopoietic stem cells (HSC), leukemia stem cells are believed to be largely quiescent and therefore impervious to conventional chemotherapeutics resulting in relapse of disease despite achievement of clinical remission. The DNA methylation profiles of bulk leukemia cells differ significantly from normal CD34+ cells; however, less is known about the potential differences between epigenetic profiles of purified LSC and normal HSC. Moreover, the stability of the methylome as the LSC differentiate into mature leukemia progenitor cells (LPC) has not been studied. In order to address this LSC, LPC, and HSC were sorted from the bone marrow of AML patients and normal controls based on CD34, CD38, CD45 and ALDH activity. LSC were defined as CD34+ CD38- ALDHmid; LPC as CD34+ CD38+; and HSC as CD34+ CD38- ALDHhigh. These isolated fractions were used for genome-wide DNA methylation analysis by the next-generation enhanced reduced representation bisulfite sequencing (ERRBS) assay, which allowed for the comparison of the methylation landscapes of LSC and HSC, as well as those of LSC and LPC. A total of thirteen AML samples were examined for the presence of LSC, six of which did not have an ALDHmid population but had instead an ALDHhigh population. Because of their phenotypic similarity to normal HSC, these samples were not included in the present comparison against HSC. The methylomes of six independent LSC samples were compared to methylomes of five HSC. Sequenced ERRBS libraries were aligned against the human genome (hg19) and differentially methylated regions (DMR) were identified using a beta-binomial model and selecting regions with absolute mean methylation difference of >25% and false discovery rate (FDR) < 10%. The methylation profiles of LSC showed widespread genome wide differences relative to HSC; 39,162 regions were found to be hypermethylated in LSC while 5,408 regions were hypomethylated. DMRs were enriched at CpG islands as well as intra- and intergenic enhancers. Functional annotation of the DMRs to gene sets in the MSigDb database revealed enrichment for genes marked by the Polycomb repressive mark H3K27me3 (FDR = 1.86×10-49). In order to determine whether epigenetic abnormalities observed at the LSC level were distinct from the epigenetic profiles observed in the more mature LPC fraction, we compared paired LSC and LPC specimens from 6 AML patients. Notably, LPC did not significantly differ in their epigenetic profiles from LSC, indicating that epigenetic abnormalities acquired at the LSC stage are stably transmitted through leukemic expansion to the more mature LPC fraction. In summary, we have identified widespread epigenetic abnormalities acquired at the LSC stage, of greater magnitude than was previously recognized by performing comparisons of leukemic cells to unfractionated CD34+ controls. Genes targeted by aberrant methylation in LSC are significantly enriched in Polycomb target genes, suggesting a potential role for Polycomb proteins in leukemic transformation. By contrast, no significant epigenetic differences were observed between the LSC and LPC fractions in AML, indicating that epigenetic abnormalities acquired at the LSC level are static through multiple cell generations. DisclosuresGore:Celgene: Consultancy, Honoraria, Research Funding.

Distinct Regulatory Networks Govern Human Hematopoietic Stem Cell Across Development

Introduction: Much of our fundamental understanding of stem cell biology comes from studies of hematopoiesis where single cells produce differentiated progeny while still retaining the ability to produce daughter stem cells (self-renewal). The cardinal property of a stem cell, whether normal or malignant, is self-renewal; the key biological process that ensures the ability of the stem cell to maintain long-term clonal growth. However, our understanding of the molecular basis of self-renewal in human hematopoiesis is limited.At the embryonic stage fetal liver is the main source of hematopoiesis; from week 6 of gestation until before birth. At this stage HSCs are in a different microenvironment but capable of self-renewing and differentiation to the full spectrum of blood lineages. While murine studies uncovered several intrinsic differences between fetal and adult HSCs, a comprehensive analysis of human HSC compartment across development is lacking.In this study we have combined HSC purification methods and xenograft quantitative assay in conjunction with low input RNA sequencing and Enhanced Reduced Representation Bisulfite Sequencing (ERRBS) to provide a comprehensive functional and molecular outlook of human stem cell compartment across development.Results: We followed the dynamics of four sub-fractions of CD34+CD38- divided by CD90 and CD49f expression across human blood development: fetal liver (hFL) and adult bone marrow (hBM). Using xenograft model, we identified human long, intermediate and short term HSCs in hFL and hBM. 5 single CD90+CD49f+ hFL cells were capable of sustaining the multilineage graft for over 52 weeks up to tertiary recipient, while BM cells only last for 20 weeks in the primary recipient. The frequency of LT-HSC in the CD90+CD49f+ compartment goes from 1/8 in hFL to 1/50 in hBM. hFL CD90-CD49f+ cells showed an intermediate repopulation capacity up to 44 weeks in secondary recipient.On average 10% of hFL long term HSC (LT-HSC) were in S/G2/M phase, in contrast only 0.4% of BM LT-HSC were in S/G2/M phase indicating that hFL HSCs are 20 times more in cycle compare to BM.We found that 320 genes were expressed differentially between LT-HSC and multipotent progenitors (MPP) in hBM as oppose to only 32 genes found to be differentially expressed in hFL (FDR<0.1). Interestingly, we found only 2 genes in common between these two groups.ERRBS showed an overall increase in methylation of HSC compartment in hBM compare to hFL and gradual demethylation of lineage associated genes in MPP.Conclusion: Our data indicate that there are distinct regulatory networks that govern hFL and hBM HSC self-renewal. We found very little differences in gene expression between all hFL HCS compartments (average 20 genes) compare to hBM (average 224), indicating that by adulthood self-renewal is becoming more restricted to the LT-HSC compartment. DisclosuresNo relevant conflicts of interest to declare.

Enhanced reduced representation bisulfite sequencing for assessment of DNA methylation at base pair resolution.

DNA methylation pattern mapping is heavily studied in normal and diseased tissues. A variety of methods have been established to interrogate the cytosine methylation patterns in cells. Reduced representation of whole genome bisulfite sequencing was developed to detect quantitative base pair resolution cytosine methylation patterns at GC-rich genomic loci. This is accomplished by combining the use of a restriction enzyme followed by bisulfite conversion. Enhanced Reduced Representation Bisulfite Sequencing (ERRBS) increases the biologically relevant genomic loci covered and has been used to profile cytosine methylation in DNA from human, mouse and other organisms. ERRBS initiates with restriction enzyme digestion of DNA to generate low molecular weight fragments for use in library preparation. These fragments are subjected to standard library construction for next generation sequencing. Bisulfite conversion of unmethylated cytosines prior to the final amplification step allows for quantitative base resolution of cytosine methylation levels in covered genomic loci. The protocol can be completed within four days. Despite low complexity in the first three bases sequenced, ERRBS libraries yield high quality data when using a designated sequencing control lane. Mapping and bioinformatics analysis is then performed and yields data that can be easily integrated with a variety of genome-wide platforms. ERRBS can utilize small input material quantities making it feasible to process human clinical samples and applicable in a range of research applications. The video produced demonstrates critical steps of the ERRBS protocol.

Enhanced Reduced Representation Bisulfite Sequencing for Assessment of DNA Methylation at Base Pair Resolution

DNA methylation pattern mapping is heavily studied in normal and diseased tissues. A variety of methods have been established to interrogate the cytosine methylation patterns in cells. Reduced representation of whole genome bisulfite sequencing was developed to detect quantitative base pair resolution cytosine methylation patterns at GC-rich genomic loci. This is accomplished by combining the use of a restriction enzyme followed by bisulfite conversion. Enhanced Reduced Representation Bisulfite Sequencing (ERRBS) increases the biologically relevant genomic loci covered and has been used to profile cytosine methylation in DNA from human, mouse and other organisms. ERRBS initiates with restriction enzyme digestion of DNA to generate low molecular weight fragments for use in library preparation. These fragments are subjected to standard library construction for next generation sequencing. Bisulfite conversion of unmethylated cytosines prior to the final amplification step allows for quantitative base resolution of cytosine methylation levels in covered genomic loci. The protocol can be completed within four days. Despite low complexity in the first three bases sequenced, ERRBS libraries yield high quality data when using a designated sequencing control lane. Mapping and bioinformatics analysis is then performed and yields data that can be easily integrated with a variety of genome-wide platforms. ERRBS can utilize small input material quantities making it feasible to process human clinical samples and applicable in a range of research applications. The video produced demonstrates critical steps of the ERRBS protocol.

Molecular Background of BCP-ALL Cases with an Early Switch to Monocytic Lineage

We identified a subset of BCP-ALL with switch towards the monocytic lineage within the first month of treatment (swALL)[Slámová et al Leukemia 2014]. During the switch cells gradually lose CD19 and CD34 expression and acquire CD33 and CD14 positivity. We proved clonal relatedness of switched monocytic blasts with the diagnostic leukemic cells based on identical Ig-TCR rearrangements. SwALL cases are not associated with MLL or BCR/ABL1 aberrancies and lack any known genetic markers of lineage ambiguity (detected by FISH or MLPA).We analyzed transcriptomes of swALL samples at diagnosis (n=4) and at d8 (n=4) where the immunophenotypic switching was already apparent as well as control BCP-ALL (n=4). RNA was isolated form either FACS sorted cells or whole BM when blasts constituted >80% of cells. For RNA-Seq we used Illumina HiSeq 2000 paired-end or single end sequencing. Raw sequencing data were analyzed using adapted protocol from Anders at al [Anders et al Nature Protocols 2013] and custom scripts. For methylome analysis we used Enhanced Reduced Representation Bisulfite Sequencing (ERRBS)[Akalin et al PLoS Genetics 2012]. ERRBS quantitatively measures DNA methylation at ~3M CpGs genome-wide. Samples from swALL at diagnosis (n=7) and at d8 (n=4) and control BCP-ALL (n=4) were processed. Analysis was performed according to [Akalin et al Genome Biology 2012] and followed with custom analysis in R statistical language.Comparison (generalized exact binomial test) of transcriptomes of B-lineage blasts from diagnosis between swALLs and control BCP-ALLs revealed a number of differentially expressed genes. Among 300 most significantly differentially expressed were KLF4, CEBPD, CLEC12A and CLEC12B (upregulated in swALL) and ANXA5, VPREB1, CD9 and IGHG3 (downregulated in swALL). Hierarchical clustering separated not only swALL and control BCP-ALL, but also swALL cells before and during the monocytic switch. Changes in gene expression during lineage switch included downregulation of ITGA6, Id2, EBF1, CD19, CD34, FLT3, MYB, CD79a, BCR, PAX5, GATA3 and TCF3 genes and upregulation of S100A10, AIF1, CD14, CD33, LGALS1, RNF130 and MNDA.When comparing all three cell types (swALL B cell and monocytic blasts and control BCP-ALL blasts) we concentrated on 1) immunophenotype switch markers and 2) lineage related transcription factors (TF):1) Both markers typical for B cell blasts (CD19, CD34) decreased during the switch. However while CD19 was expressed in swALL at diagnosis at same levels as in control BCP-ALL, CD34 was overexpressed in swALL compared to BCP-ALL at diagnosis. Both monocytic markers (CD33, CD14) increased their expression during the switch. CD14 showed no difference between swALL and control BCP-ALL at diagnosis. However CD33 was interestingly upregulated in swALL already at diagnosis and continued to rise during the switch. SwALL had therefore deregulated expression of lineage commitment markers already at diagnosis favoring stemness marker CD34 and myeloid marker CD33.2) B lineage commitment related TFs (EBF1, TCF3, PAX5) were expressed in B lineage blasts in both swALL and control BCP-ALL. However they were all downregulated during the switch. On the other hand myeloid lineage related transcription factor CEBPA is overexpressed in diagnostic B lineage blasts in swALL compared to control BCP-ALL cases. Similarly CEBPD is overexpressed in swALL and its expression further rises during the switch. Other hematopoietic TFs upregulated in swALL cases include KLF4, NANOG and GATA3.To confirm some of the epigenetic markers of swALL cases (demethylation of CEBPA promoter) and to widen epigenetic screening we used ERRBS. While some of the upregulated genes had expectedly hypomethylated promoters in swALL (CEBPA, GATA3) other genes (TCF3, PAX5) had demethylated promoters in all cases. While the whole DNA methylation picture is still a challenge to draw both omics method could clearly separate swALL cases from control BCP-ALL using principal component analysis.In summary we show that immunophenotypic shift is associated with gene expression changes of surface markers, lineage specific transcription factors and other genes. Some of the genes have altered expression already at diagnosis. Expression of some key lineage genes is differentially regulated by DNA methylation.Supported by: GAUK 914613, GAČR P301/10/1877, UNCE 204012, IGA NT13462-4 DisclosuresNo relevant conflicts of interest to declare.

Distinct DNA Methylation and Expression Profiles Underlie CMML Responsiveness to Decitabine and Uncover Novel Mechanism of Resistance

Myelodysplastic syndromes (MDS) and the related disorder chronic myelomonocytic leukemia (CMML) are characterized by abnormal DNA hypermethylation and the DNA methyltransferase inhibitors (DMTis) azacytidine and decitabine (DAC) are frequently used as frontline therapy in these patients. However, DMTis are ineffective for ~50% of the patients who must still undergo treatment for at least 6 months before they can be deemed resistant. Therefore, it is of critical importance to identify baseline molecular differences associated with DMTi sensitivity that can help (i) to improve patient risk-stratification at diagnosis and (ii) identify the underlying mechanisms of resistance to these agents. Previous efforts to identify baseline DNA methylation differences at promoter regions between DMTi responders and non-responders have not been successful, so we hypothesized that any potential differences would be located distally from promoter regions. For this purpose we studied 40 CMML patients at diagnosis, all of whom had been uniformly treated with DAC 20 mg/m2/day x 5 days as frontline therapy. After 6 cycles of therapy patients were classified as responders (n=19, hematological improvement or better), or non-responders (n=21, stable or progressive disease). Mutational analysis showed no significant differences in the frequency of mutations in TET2, ASXL1, DNMT3A, RUNX1, TP53, JAK2, KIT, KRAS, EZH2, IDH1/2 and spliceosome genes. Using Enhanced Reduced Representation Bisulfite Sequencing (ERRBS) we analyzed the baseline methylation status at ~3M CpG sites across the genome of 39/40 CMML patients. We identified 158 statistically significant differentially methylated regions (DMRs) (FDR<0.1 and methylation difference ≥25%) between the two groups. DAC-sensitive patients displayed both regions of higher methylation as well as regions with lower methylation compared to DAC-resistant patients. As predicted, DMRs were depleted at promoters (DMRs 9% vs. Background [BG] 21%, p-value: 3.4×10-5) and CpG islands (DMRs 8% vs. BG 25%, p-value: 1.5×10-8). Further analysis showed that hypermethylated DMRs were enriched at intronic regions (Hyper DMRs 58% vs. BG 33%, p-value: 3.7×10-6) while hypomethylated DMRs were enriched at intergenic regions (Hypo DMRs 49% vs. BG 38%, p-value: 2.6×10-2). Moreover, hypermethylated DMRs were significantly enriched for enhancer regions, and in particular, enhancers located within gene bodies (hyper DMRs 38% vs. BG 18%, p-value: 2.3×10-5). KEGG pathway analysis showed a significant enrichment of DMRs in the MAPK signaling pathway (FDR<0.01). Next, using a support vector machine algorithm with 10-fold cross validation we were able to develop a classifier capable of predicting response to DAC with high level of accuracy (ROC AUC: 0.99) based solely on the DNA methylation status at diagnosis of 17 genomic regions. Three different random splits of the cohort into training and test sets achieved correct predictions for 85.7%, 89.47%, and 100% of cases, respectively, demonstrating the accuracy and potential utility of such a classifier. Finally, RNA-seq analysis identified 53 differentially expressed genes between responders (n=8) and non-responders (n=6) at diagnosis. Genes implicated in cell cycle and DNA replication were overexpressed in responders. By contrast, very few genes were overexpressed at the time of diagnosis in primary resistant patients. Among these were CXCL4 and CXCL7 which, given their reported contributions to cell cycle arrest and chemoresistance, were tested for their functional roles in DAC resistance. Pre-treatment of normal CD34+ cells for 72 h with 10nM DAC significantly reduced colony formation (p<0.05) but the addition of 50ng/mL of CXCL4 and CXCL7 restored colony formation to that of untreated cells. Moreover, treatment of primary CMML cells with 10 nM DAC for 72h significantly reduced viability of these cells, while concomitant treatment with 50ng/mL of CXCL4 and CXCL7 was sufficient to abrogate this effect. Taken together, our findings demonstrate that (i) specific DNA methylation profiles targeting non-promoter regulatory regions are associated with DAC sensitivity, (ii) these differences can be harnessed for the development of clinical biomarkers predictive of response and (iii) we identified a novel mechanism of resistance to DAC mediated through two chemokines that are exclusively overexpressed in non-responders. DisclosuresNo relevant conflicts of interest to declare.

Altered Histone Mark Deposition and DNA Methylation at Homeobox Genes in Human Oral Squamous Cell Carcinoma

We recently reported a role of polycomb repressive complex 2 (PRC2) and PRC2 trimethylation of histone 3 lysine 27 (H3K27me3) in the regulation of homeobox (HOX) (Marcinkiewicz and Gudas, 2013, Exp Cell Res) gene transcript levels in human oral keratinocytes (OKF6-TERT1R) and tongue squamous cell carcinoma (SCC) cells. Here, we assessed both the levels of various histone modifications at a subset of homeobox genes and genome wide DNA methylation patterns in OKF6-TERT1R and SCC-9 cells by using ERRBS (enhanced reduced representation bisulfite sequencing). We detected the H3K9me3 mark at HOXB7, HOXC10, HOXC13, and HOXD8 at levels higher in OKF6-TERT1R than in SCC-9 cells; at IRX1 and SIX2 the H3K9me3 levels were conversely higher in SCC-9 than in OKF6-TERT1R. The H3K79me3 mark was detectable only at IRX1 in OKF6-TERT1R and at IRX4 in SCC-9 cells. The levels of H3K4me3 and H3K36me3 marks correlate with the transcript levels of the assessed homeobox genes in both OKF6-TERT1R and SCC-9. We detected generally lower CpG methylation levels on DNA in SCC-9 cells at annotated genomic regions which were differentially methylated between OKF6-TERT1R and SCC-9 cells; however, some genomic regions, including the HOX gene clusters, showed DNA methylation at higher levels in SCC-9 than OKF6-TERT1R. Thus, both altered histone modification patterns and changes in DNA methylation are associated with dysregulation of homeobox gene expression in human oral cavity SCC cells, and this dysregulation potentially plays a role in the neoplastic phenotype of oral keratinocytes.

Epigenomic Evolution In Diffuse Large B-Cell Lymphomas

Diffuse Large B-cell Lymphoma (DLBCL) is the most common non-Hodgkin lymphoma worldwide. It is a heterogeneous disease in which one third of patients either do not respond to treatment or relapse within five years after chemotherapy. It is unclear whether epigenetic alterations are responsible for B cell lymphomas relapse phenotypes, such as increased aggressiveness and chemoresistance. To investigate how the B cell lymphoma epigenome evolves upon chemotherapy, we used Enhanced Reduced Representation Bisulfite Sequencing (ERRBS) to profile DNA methylation genome-wide in primary matched diagnosis-relapse DLBCL samples.We interrogated 13 pairs of DLBCL diagnosis tumors and their matched relapse samples. In addition, we performed methylation profiling of normal tonsilar B cell subsets (Naïve and germinal center B cells) from two healthy human individuals. ERRBS provided DNA methylation levels at 3-4M CpG sites. When combining methylation levels from all interrogated CpG sites, we observed increased DNA methylation levels at CpG-islands (CGIs; p=3.5e-9, t-test) in DLBCLs compared to normal B cells, and stable or slightly decreasing methylation levels outside of CGIs (>10 kb away from known CGIs; p=0.057, t-test). There was no significant change in average DNA methylation levels from diagnosis to relapse either at CGIs or outside of CGIs. However, when we investigated DNA methylation changes at gene promoters, we identified 107 consistently differentially methylated promoters between diagnosis and relapse (> 10% DNA methylation alteration and p < 0.05, paired t-test). Pathway analysis of the corresponding genes using iPAGE identified several pathways and processes associated with either hyper or hypo-methylated genes in relapse, demonstrating that methylation changes associated with relapse are functionally coherent. For example, several genes with TGF-beta receptor activity displayed lower DNA methylation in relapse.Taking advantage of single CpG resolution and high coverage provided by ERRBS, we then sought to investigate the extent of allele-specific methylation (ASM) levels in normal tissues and DLBCL patients. We found increased ASM levels in DLBCLs compared to normal tissues (p=0.0011, t-test) confirming observations in solid tumors. There was no significant change in ASM levels from diagnosis to relapse (p=0.24, t-test). These results suggest that methylation changes associated with lymphomagenesis might frequently involve one allele only, perhaps due to differential nuclear localization of individual chromosomes. However allele-specific methylation may not play a key role in lymphoma progression.Finally, we investigated whether intra-tumor methylation heterogeneity at diagnosis would predict whether a DLBCL patient would relapse. We quantified intra-tumor methylation heterogeneity using a statistical approach based on the probability that two randomly sampled DNA molecules from the tumor cell populations differ from each other in their methylation pattern. We found decreased intra-sample methylation heterogeneity in DLBCLs compared to normal germinal center B cells (p=1.9e-4, t-test), consistent with the clonal origin of tumors. 12 out of 13 pairs also displayed decreased methylation heterogeneity from diagnosis to relapse, which is also consistent with clonal selection upon chemotherapy treatment. We then performed ERRBS on primary tumors from 8 DLBCL patients who have not relapsed five years after treatment. We found that non-relapse patients displayed significantly lower intra-tumor methylation heterogeneity as compared to that of the relapsed patients (p=0.047, t-test), which suggests that increased epigenetic diversity within a population of tumor cells at diagnosis might fuel the Darwinian evolutionary process underlying relapse. We also looked at genetic clonal heterogeneity based on next-generation sequencing of somatic hypermutation profiles in IGH VDJ sequences, but found no differences between relapsed and not relapsed patients (p=0.22, Wilcoxon test). This suggests that epigenetic heterogeneity plays a more substantial role than clonal heterogeneity in fueling the relapse phenotype.In summary, this study provides the first comprehensive characterization of aberrations in DNA methylation in relapse DLBCLs and identified epigenetic diversity in DLBCLs as a potential predictor of relapse. Disclosures:No relevant conflicts of interest to declare.