SET domain containing 2 (SETD2) gene encodes the sole methyltransferase that is specific for histone H3 lysine 36 trimethylation (H3K36me3) in mammals. Previously we identified somatic SETD2 loss-of-function (LOF) mutations in both acute myeloid leukemia (AML) and acute lymphoid leukemia (ALL) patients. Interestingly, SETD2 mutations were identified in over 20% of leukemia patients with MLL gene rearrangement. Using an Mll-Af9 (MA9) knock-in AML mouse model, we found that downregulation of Setd2 by shRNA or genetic knock-out/knock-in strategies significantly accelerated disease development. However, the contribution of SETD2-H3K36me3 axis downregulation to MLL leukemia progression has not yet been fully understood. In MLL leukemia, high expression levels of MLL targets, driven by aberrant histone H3 lysine 79 dimethylation (H3K79me2), have been reported. Both H3K36me3 and H3K79me2 are enriched within the body of transcriptionally active genes and are associated with phosphorylated RNA Polymerase II. Whether H3K36me3 also contributes to gene dysregulation in MLL leukemia, whether SETD2-H3K36me3 downregulation has impacts on DOT1L-H3K79me2 axis, and how it promotes MLL leukemia progression are unclear. More importantly, SETD2 mutations were enriched in relapsed B-cell ALL patients. SETD2 LOF mutation could cause chemotherapy resistance or relapse after bone marrow transplantation. Thus, mechanistic-driven novel therapies are urgently needed for leukemia with SETD2 mutation.

Firstly, to understand SETD2-H3K36me3 axis on gene regulation in MLL leukemia, we purified c-Kit+ hematopoietic stem/progenitor cells (HSPCs) fromMA9 and normal C57/BL/6 wild-type (WT) adult mice. Using immuno-blotting, we found increased H3K36me3 and H3K79me2 in the HSPCs from MA9 compared to those from WT mice. To further study the global histone modification and gene regulation, we performed ChIP-seq and RNA-seq analyses. Genome-wide increase of H3K36me3 and H3K79me2 were confirmed in the HSPCs from MA9 mice, and both modifications were positively correlated with gene expression. Our results indicate that not only increased H3K79me2, but also increased H3K36me3 are related to gene dysregulation in MLL leukemogenesis. Next, to explore the impact of SETD2-H3K36me3 loss on DOT1L-H3K79me2 axis and leukemia progression, we performed KD of Setd2 gene. Setd2 KD caused increased self-renewal and proliferation abilities of both WT and MA9 HSPCs. However, Setd2 KD-WT HSPCs could be replated only 3-4 times in vitro, suggesting that the SETD2 single mutation is not sufficient for leukemic transformation. In contrast, Setd2 KD significantly enhanced in vitro replating ability and in vivo leukemia development of MA9 HSPCs. Immuno-blotting and ChIP-seq results revealed the dramatic loss of H3K36me3 in Setd2 KD cells. Surprisingly, global H3K79me2 was further increased in Setd2 KD cells. As upregulation of MLL targets is the main driver for MLL leukemogenesis, we measured the expression levels of known MLL targets. However, no significant change was found either in the RNA-seq or qPCR validation. Moreover, up-regulated genes with higher H3K79me2 in their 5' gene bodies were not enriched in the classical known MLL targets but a different group of AML related genes including Arg, Erg and Bcl2l1. This partially explains the corporative activity between SETD2 LOF and MLL fusions. Thirdly, due to the further increase of H3K79me2 in Setd2 KD cells, we tested epigenetic inhibitors for DOT1L-H3K79me2 axis in MLL leukemia with SETD2 LOF mutant models. DOT1L inhibitor EPZ-5676 induced differentiation and cell death of Setd2 KD MA9 cellsor MA9/Setd2 LOF mutant cells with significantly lower concentration (450nM) compared to the case of MA9 cells (1000nM), indicating that DOT1L-H3K79me2 axis could be the tumor vulnerability of this chemo-resistant type of leukemia.

In conclusion, using MA9 genetic knock-in mouse model and modulating SETD2-H3K36me3 axis, we found that, 1) not only H3K79me2, but also H3K36me3 were aberrantly modified in MLL leukemia which related to gene dysregulation, 2) Downregulation of SETD2-H3K36me3 axis could further upregulate DOT1L-H3K79me2 axis, leading to activation of a new set of AML related genes which could contributes to quick leukemia onset, 3) EPZ-5676 could be an effective therapeutic option for MLL leukemia patients with SETD2 mutations by targeting its tumor vulnerability.

Disclosures

No relevant conflicts of interest to declare.

Author notes

*

Asterisk with author names denotes non-ASH members.

Sign in via your Institution