IntroductionMyelofibrosis (MF) is known for frequent alterations in genes involved in epigenetic machinery, particularly in the regulation of DNA methylation. Enhancer regions experience changes in DNA methylation throughout myelopoiesis and may also experience aberrant DNA methylation in hematologic neoplasms, leading to altered expression of relevant genes. We aim to describe the DNA methylome of MF focusing on dynamic changes of DNA methylation in enhancer-regions, in order to identify new potential gene candidates with epigenetic deregulation who might have a role in the pathogenesis of MF and become therapeutic targets for this disease.MethodsDNA methylation profiles of 39 MF samples compared to 8 healthy controls were analysed using the Infinum Human-methylation 450K BeadChip platform. Differentially methylated CpG dinucleotides (DMCs) were validated with bisulfite sequencing. Enhancer regions were defined using publicly available data from the ChromHMM algorithm (H3K27Ac and H3K4Me3; Wijetunga, Nat Comm, 2014;5:5195). DNA methylation of DMCs mapped to enhancer regions was correlated to public gene expression datasets (GSE26049). The subgroup of genes with hypermethylated enhancers and downregulated expression was further studied, selecting candidate genes for epigenetic regulation through enhancer DNA methylation. Luciferase reporter assay using a CpG-free vector (pCPGL) was performed to confirm the functional effect of hypermethylated enhancers. SET2 cell line (JAK2V617F mutated) was used as in vitro model for pharmacologic treatment and lentiviral induction of selected gene candidates.ResultsWe found a total of 35216 differentially methylated CpGs (DMCs) between MF and controls (FDR <0.05). DMCs were significantly enriched to enhancer regions accounting for 8764 probes (24.8%; p<0.001 vs global DMC). 4182 enhancer-specific DMCs were hypermethylated (27.6%), comprising 2131 genes, whereas the remaining 10936 enhancer DMC were hypomethylated. Among genes showing reduced expression and enhancer hypermethylation in MF samples, we identified the ZFP36L1 gene as potential candidate. Moreover, gene targets of ZFP36L1 showed a significant enrichment among the genes overexpressed in MF indicating that ZFP36L1 downregulation could play an important role in MF. A novel algorithm to search for genes with putative 5´UTR AU rich elements was used (DREME motif discovery algorithm; E-value=4.5*10-15).We detectedabsence of DNA methylation in the promoter region of ZFP36L1 in SET2 cell line, MF samples and healthy controls, whereas the enhancer region showed intermediate DNA methylation in healthy controls and hypermethylation in SET2 and MF samples. DNA hypermethylation of the enhancer correlated with reduced ZFP36L1 gene expression (p<0.05). 5-azacytidine partially rescued ZFP36L1 expression in SET2 cell line, which was associated with a reduction of enhancer DNA methylation. Luciferase reporter assays with the pCPGL-vector showed that DNA methylation of the ZFP36L1 enhancer region abrogated the luciferase activity, demonstrating for the first time that this enhancer region regulates the expression of ZFP36L1 in MF.Finally, we demonstrated that lentiviral induction of ZFP36L1 in SET2 cell line, significantly reduced cell proliferation and increased apoptosis (Annexin V). We observed similar results by treatment of SET2 cells with PMA, a known inductor of ZFP36L1 . Additionally, we found that rescue of ZFP36L1 expression was associated with a decrease in CDK6, suggesting a potential role of ZFP36L1-CDK6 axis in the pathogenesis of MF.ConclusionWe detected a significant inappropriate DNA methylation of enhancer regions in MF, showing a novel pathogenetic mechanism for this disease that might be extrapolated to other haematologic malignancies. Aberrant DNA methylation of enhancer regions affects genes with potential functional implications. ZFP36L1 is an example of this phenomenon, harbouring DNA hypermethylation of the enhancer region and constituting a potential tumour suppressor gene with inappropriate epigenetic silencing. ZFP36L1 might be relevant for the control of cell cycle progression in MF through modulation of the expression of CDK6 . DisclosuresNo relevant conflicts of interest to declare.
Read full abstract