The Role of H3K79me2 in Leukemia and Alternative Splicing

Methylation at lysine 79 of histone H3 regulates a variety of nuclear functions necessary for embryogenesis, hematopoiesis and cardiac development while playing crucial roles in cell cycle regulation, DNA repair, transcriptional activation and alternative splicing (Bernt et al., 2011; Daigle et al., 2011; Deshpande et al., 2013; Huyen et al., 2004; Jones et al., 2008; McLean et al., 2014). Surprisingly, given the abundance of studies of this mark (W. Kim et al., 2014; McLean et al., 2014; Vlaming & van Leeuwen, 2016), little is known about the mechanisms by which it affects cellular processes and importantly, the proteins that recognize it. For instance, in the current model of MLL-rearranged (MLL-r) leukemogenesis, knockout or inhibition of DOT1L, the enzyme that deposits H3K79me2, reduces expression of HOXA9 and MEIS1, the leukemic oncogenes that drive proliferation, through depletion of activating H3K79me2, thereby reducing cell survival (Bernt et al., 2011; Guenther et al., 2008; Milne et al., 2005; Stubbs et al., 2008). This model however, is unable to explain key observations such as reductions in the proliferation of leukemia cells at DOT1L inhibitor concentrations that do not affect the expression of the canonical driver oncogenic drivers HOXA9 and MEIS1. Additionally, H3K79me2 depletion also affects alternative splicing in MLL-r cell lines through repression of exon skipping but, the mechanism behind this effect and the proteins involved are unknown. In MLL-r leukemia, I find that the FLT3-ITD/STAT5A and PRC2 pathways are disrupted by low-dose pinometostat (10 nM), a concentration that reduces proliferation of the MLL-r MV4;11 cell line without affecting HOXA9 and MEIS1 expression. At this low-dose inhibitor concentration I also identify 71 events of differential alternative splicing. Using quantitative ICeChIP-seq, I observe profound H3K79me2 depletion at downregulated MLL-r targets, and alternatively spliced genes, with resulting increases in transcriptionally activating H3K4me3 at MLL-r target promoters, increases in H3K36me3 in gene bodies and global reductions in repressive H3K27me3. Although downregulation of polycomb components modestly contributes to reductions in proliferation, overexpression of constitutively active STAT5A, a target of FLT3-ITD-signalling, nearly completely rescues proliferation, accounting for the bulk of cytotoxicity from H3K79me2 depletion. I also observe a dependence of FLT3-ITD/STAT5A signaling on MLL1 function, suggesting that the FLT3 locus is exquisitely sensitivity to both H3K79me2 and H3K4me3 depletion and arguing that combinations of DOT1L, MLL1 and FLT3 inhibitors should be explored for treating the ~30% of all leukemias that carry FLT3 mutations. Additionally, I identify several splicing factors that recognize H3K79me2 through modified nucleosome pulldowns from nuclear extract followed by tandem mass spectrometry. Knocking down one of these factors, PTBP1 results in similar effects on alternative splicing as DOT1L inhibition, suggesting that PTBP1 facilitates H3K79me2-mediated alternative splicing and providing the first mechanistic understanding of how H3K79me2 affects alternative splicing