Combinatorial Epigenetic Aberration Propagates in Myelodysplastic Syndrome in the Setting of Concurrent Loss of Tet2 and Ezh2
Nagisa Hasegawa,Goro Sashida,Motohiko Oshima,Hirotaka Matsui,Atsunori Saraya,Changshan Wang,Tomoya Muto,Chiaki Nakaseko,Koutaro Yokote,Atsushi Iwama
DOI: https://doi.org/10.1182/blood.v126.23.712.712
IF: 20.3
2015-01-01
Blood
Abstract:Somatic loss-of-function mutations of epigenetic regulators are frequently found in MDS patients. Although DNA hypomethylating agents have been shown to improve the clinical outcomes in patients with MDS, it remains unknown how altered DNA methylation promotes the development of MDS. We have recently shown that concurrent loss of Tet2 and Ezh2 promotes the development of MDS in mice compared with the individual loss by utilizing hypomorphic Tet2 mice and Ezh2 conditional knockout mice. In this study, we transplanted Cre-ERT (WT), Tet2KD/KD, Cre-ERT;Ezh2fl/fl, and Tet2KD/KD;Cre-ERT;Ezh2fl/fl fetal liver cells into lethally irradiated recipient mice, and deleted Ezh2 at 4 weeks post-transplantation. To understand how DNA hypermethylation contributed to the development of MDS, we performed Reduced Representation Bisulfite Sequence (RRBS) in Lin- Sca1+ Kit+ (LSK) stem/progenitor cells isolated from WT, Tet2KD/KD, Ezh2Δ/Δ, and Tet2KD/KD Ezh2Δ/Δ mice. Both Tet2 loss and Ezh2 loss caused hypermethylation in CpG islands (CGIs) and promoter regions. However, the impact of Tet2 loss was much evident in the enhancer elements and ~20 % of them underwent hypermethylation upon Tet2 loss. Combined loss of Tet2 and Ezh2 caused hypermethylation in the transcriptional regulatory regions including both the promoter regions and enhancer elements. Importantly, the majority of hyper-differentially methylated regions (hyper-DMRs) in Tet2KD/KD Ezh2Δ/Δ -MDS LSK cells showed a distribution that was distinct from those of single mutant LSK cells and only the hyper-DMRs in the enhancer elements in Tet2KD/KD Ezh2Δ/Δ -MDS LSK cells correlated with a significant reduction in gene expression levels. Hyper-DMRs in MDS LSK cells were significantly enriched in genes involved in transcriptional regulation and cell differentiation by GO analysis. These findings suggest that combined loss of Tet2 and Ezh2 cooperatively remodeled DNA methylation to an extent not observed in either mutant allele alone and contributed to the pathogenesis of MDS. Notably, we found that 375 out of 780 hyper-DMRs in Tet2KD/KD Ezh2Δ/Δ MDS HSPCs were overlapped with DMRs in CD34+ HSPCs in patients with MDS (Maegawa S, et al. Genome Research 2014). We next focused on the 131 genes with hyper-DMRs that showed reduced expression levels in Tet2KD/KD Ezh2Δ/Δ -MDS LSK cells. Those included Gata2, Gata3, Evi1, and Nr4a2. Given that Nr4a2/Nurr1, a nuclear receptor transcription factor, restricts the proliferation of hematopoietic stem cells (HSCs), we transduced WT and Tet2KD/KD Ezh2Δ/Δ HSCs with an Nr4a2 retrovirus in vitro. We found that exogenous Nr4a2 severelyimpairs proliferative capacity of Tet2KD/KD Ezh2Δ/Δ HSCs but moderately that of WT HSCs, implying that MDS HSCs are more susceptible to activation of Nr4a2 compared to WT HSCs. Next, we sought to determine whether DMRs in CGIs were associated with altered histone modifications upon the deletion of Ezh2, we performed chromatin immunoprecipitation (ChIP) sequencing by using H3K27me3 and H3K4me3 antibodies. We defined canonical polycomb repressive complexes 2 (PRC2) targets by >2-fold enrichment of H3K27me3 and bivalent genes by enrichment of both H3K27me3 and H3K4me3 in WT LSK cells. We found that approximately half of the hyper-DMRs in CGIs were enriched in canonical PRC2 targets and bivalent genes in LSK cells of all mutant genotypes. Notably, these hyper-DMRs in PRC2 targets, including bivalent genes, were significantly associated with reduced expression levels only in Tet2KD/KD Ezh2Δ/Δ -MDS LSK cells but not in either single mutant alone. We also observed that these PRC2 targets with hyper-DMRs still retained reduced but higher levels of H3K27me3 compared to the others despite the absence of Ezh2, implying that both DNA hypermethylation and residual H3K27me3 by Ezh1 cooperate to repress the transcription of these genes critical for hematopoiesis. Finally, we examined the impact of decitabine, a demethylating agent in vivo. Although the treatment of decitabine did not improve the survival of Tet2KD/KD Ezh2Δ/Δ -MDS mice, it partly resolved hyper-DMRs at critical genes including Gata2, Gata3, and Nr4a2, followed by a significant elevation in platelets counts in vivo. Taken together, our study unveils the cooperative biological function of aberrant DNA methylation and histone modifications in the pathogenesis of MDS.