Assistant Professor • SOM: HMO: Hematology Emory University Atlanta, Georgia, United States
Introduction: dentifying the biology of high-risk multiple myeloma (MM) is critical to improving outcomes. Current markers imperfectly identify high-risk disease and there are limited data that integrate genetic, epigenetic, and transcriptional information with outcomes. We generated DNA methylation (DNAm) data on 415 samples from the MMRF CoMMpass study and identify distinct epigenetic programs of high-risk disease.
Methods: Institutional IRB approval was obtained and DNA from CD138+ enriched cells was used to prepare whole genome bisulfite sequencing libraries. Sequencing was performed on short-read instruments (Illumina) and aligned to the GRCh38 genome with Bismark. DNAm values were obtained at 5x coverage at 16,374,655 CpGs in 90% of samples with an average coverage of 24x. Data were compared to WGBS from the BluePrint project. CoMMpass samples had 100% matching germline and somatic genetic data and 93% had matching RNA-seq.
Results: Analysis of newly diagnosed MM samples (N=370/415) showed median DNAm was 45% (23-77%), which was significantly less than 69% in normal plasma cells. MM-specific hypomethylation occurred in large, late-replicating domains of the genome. In contrast, normal plasma cells and other B cell malignancies (CLL), showed no reduction in DNAm at late replicating regions. Cross-sectional analysis of RNA and DNAm in cis found 170,777 loci associated with expression, with most (98%) positively associated with transcription and found in gene bodies. Unsupervised t-SNE analysis of DNAm data grouped samples by expression subtype with the NSD2/MMSET (MS) and Proliferation (PR) subtypes more clearly defined by DNAm than RNA-seq. The MS and PR subtypes had 2,075,489 and 418,474 differentially methylated loci (DML), respectively. However, the MS subtype had more DNAm (92% of DML being higher) whereas the PR subtype had less DNAm (99.95% of DML being lower) as compared to other subtypes. Ablation of NSD2 resulted in reduced DNAm, indicating MS subtype hypermethylation was specific to NSD2 expression and excessive H3K36me2. In contrast, no individual gene could be ascribed to PR subtype hypomethylation, but hypomethylation of these loci were prognostic of outcome and lost DNAm between paired baseline and relapse samples.
Conclusions: While MM is characterized by a profound loss of DNAm, the DNAm retained in actively transcribed gene bodies suggests a role in expression. MM subtypes have distinct epigenetic programs with the high-risk MS and PR subtypes having divergent epigenetic programs. Higher levels of DNAm in the MS subtype are specific to NSD2 overexpression and may result from excessive H3K36me2 being recognized by the PWWP domains of DNA methyltransferases. Lower levels of DNAm in the PR subtype better defined this subtype than RNA and likely result from the replication-dependent loss of DNAm, suggesting DNAm at these loci serve as a readout of the mitotic clock in MM.
