DNA hydroxymethylation reveals transcription regulation networks and prognostic signatures in multiple myeloma

Multiple myeloma (MM) is a plasma cell malignancy that remains challenging to cure despite a substantially improving median survival. During the last decade, DNA copy number variation and gene expression studies have described the pathology and its heterogeneity among patients. Epigenetic modifications play important roles in MM, but they are rarely associated with clinical aspects of the disease. In this epigenomics study, we produced quantifications of genomic 5-methylcytosine (5mC) and of 5-hydroxymethylcytosine (5hmC) as well as genome-wide maps of hydroxymethylation to analyse myeloma cells taken from a cohort of 40 newly diagnosed and homogeneously treated patients. We found 5hmC to be globally depleted in MM compared to normal plasma cells, as well as being reduced in advanced clinical stages of the disease. From the hydroxymethylome data, we observed that remaining 5hmC is organised in large peak clusters and is associated with well-known disease-related genes. Based on their signal correlation, these 5hmC peak clusters can be gathered in 2 regulation networks involving core transcription factors such as IRF4, MYC, PRDM1 and TCF3. By performing paired hydroxymethylomes at diagnosis and at relapse, we found the disease progression to be heterogeneous and patient-specific. We found that the location of 5hmC at tumor suppressor TP53INP1 is associated with better outcome while global high level of 5hmC tends to be associated with better overall survival. Together, our study suggests that 5hmC provides new biological insights of the disease severity and progression, and can be used for retrospective studies.