Comprehensive Characterization of the Epigenetic Landscape in Multiple Myeloma

Multiple myeloma (MM) is a B cell neoplasia characterized by the accumulation of clonal plasma cells within the bone marrow. Epigenetic modifications, including DNA methylation and histone post-translational changes, are involved in MM development and progression, but also in drug resistance.

We have derived a large cohort of patient-derived HMCLs that remain dependent on the addition of exogenous MM bone marrow growth factors, reflecting primary tumor conditions. We have described their molecular diversity by analyzing the gene expression profile and mutational landscape and have showed that HMCL molecular diversity reflects part of the molecular heterogeneity of primary MM cells. However, the epigenetic landscape of HMCLs has never been described.

A comprehensive characterization of the epigenetic landscape of HMCLs would advance our understanding of MM pathophysiology and may attempt to identify new therapeutic targets.

In our study, we presented the epigenetic landscape of HMCLs. We performed chromatin immunoprecipitation sequencing (ChIP-seq) to analyze changes of the histone marks (H3K9me3 to follow heterochromatin, H3K4me1 and H3K27ac to follow enhancer activity, H3K4me3 and H3K36me3 to follow active transcription and H3K27me3 to follow Polycomb-silenced chromatin) on 16 HMCLs, representative of the molecular heterogeneity of MM. The differential analysis of histone modification profiles of HMCLs highlighted links between histone modifications and cytogenetic abnormalities or recurrent mutations. H3K4me3 and H3K27me3 profile analysis revealed specific clusters of HMCL related to 1q gain and t(4;14) translocation, respectively. These two cytogenetic abnormalities lead to deregulation of epigenetic player expression (e.g. SETDB1 and MMSET) and thus, could alter histone modification profile. Using histone modifications associated to enhancer regions (H3K4me1 and H3K27ac), we identified super-enhancers (SE) associated with genes involved in the biology of MM. 607 to 2510 predicted super-enhancers per HMCL were identified, including MAF, MYC, CCND1, CCND2, TRAF3 or NSD2. These super-enhancers differ from typical enhancers in both size and H3K4me1 and H3K27ac levels. The SE-associated genes identified in HMCLs with a prognostic value in two independent cohorts of newly diagnosed patients (CoMMpass cohort; N = 674 and Montpellier cohort; N = 69 with RNA-seq data) were used to build a score predicting MM patient outcome (Figure 1). Moreover, among the 28 genes that compose the risk-score (BSG, HK2, HNRNPC, HSPA9, IL10, ILF3, LDHB, MDH1, MYBPC2, NCL, NUDC, PARP1, PDIA6, PRPS1, RPL8, RPL13A, RPL27A, RPL35, SF3B2, SLC7A5, SLC25A39, SMARCA4, SPN, STC2, THY1, TNPO2, TPR), public datasets of RNAi and CRISPR/Cas9 screening revealed four genes (YWHAQ, IL10, HK2 and THY1) identified as significant essential myeloma genes, suggesting that they could represent potential therapeutic targets. We also identified promoters of genes characterized by a co-localization of H3K9me3 and H3K27me3 repressive marks in HMCLs. We evaluated the prognostic value of these genes in the CoMMpass and Montpellier cohorts, and selected genes associated with poor outcome when their expression is low in MM cells of patients (ARHGEF5, BIVM, DEF8, GRID2IP, HDAC9, HSPA1L, KDM4C, NLRP2, P4HA3, PAG1, PM20D1, RMND5A, SEMA6A, SFMBT2, THEMIS2, TPRKB, ZFP2 and ZNF5188B) underlining potential new tumor suppressor genes. These potential tumor suppressor genes associated with repressive histone marks were used to build a second risk score splitting MM patients in low- and high-risk groups in CoMMpass and Montpellier cohorts. Finally, we explored H3K4me3 marks comparing drug-resistant and -sensitive HMCLs (N = 16) to identify regions involved in drug resistance. From these data, we developed epigenetic biomarkers based on this H3K4me3 modification predicting lenalidomide and romidepsin HDCAi response.

This study provides a comprehensive characterization of the MM epigenetic landscape representing unique resources for future biological studies and could help to identify novel critical epigenetic modifications involved in MM progression and drug resistance. Furthermore, risk-scores based on super enhancers and repressive regions together with epigenetic biomarkers of drug response could represent new tools for precision oncology in MM.