SETDB1 Represses Hox Gene Expression and Suppresses Acute Myeloid Leukemia

Epigenetic regulators play an important role in normal and malignant hematopoiesis. Epigenetic deregulation of the HOXA gene cluster drives transformation of about 50% of acute myeloid leukemia (AML), including those harboring MLL rearrangements and NPM mutations, as well as others. Expression of Hoxa9 and its co-factor Meis1 is sufficient to transform bone marrow into a lethal AML in mouse models. We previously demonstrated that the pro-leukemic genes Hoxa9 and Meis1 are critically regulated by the histone H3 Lysine 9 (H3K9) methyltransferase SETDB1. Recent studies show that SETDB1 is required for normal hematopoiesis and MLL-AF9 mediated leukemia (Koide, et al. Blood 2016). Our lab recently demonstrated that SETDB1 negatively regulates the expression of HoxA9 and Meis1 through deposition of promoter H3K9 methylation in MLL-AF9 AML cells (Ropa et al. Oncotarget 2018). Consistent with these data, HOXA9 and MEIS1 expression negatively correlates with SETDB1 expression in AML patient samples. Therefore, we investigated the biological impact of SETDB1 on AML.

We first noted that expression of SETDB1 in AML patient samples is significantly lower compared to normal hematopoietic cells. Further, higher SETDB1 expression correlated with a significantly better overall survival (p=0.003) and lower expected hazard (HR=0.9/100RSEM; p=0.009) in AML patients compared with lower SETDB1 expression. These data are consistent with SETDB1 negatively regulating pro-leukemic genes and suggests that SETDB1 expression may be correlated with AML patient prognosis. We tested this directly by expressing high levels of SETDB1 in AML cells. Ex vivo assays show that retroviral overexpression of SETDB1 in MLL-AF9 AML cells leads to cell differentiation, decreased leukemia colony formation, and decreased cell proliferation. Consistent with the AML patient data, overexpression of SETDB1 significantly delays MLL-AF9 mediated leukemogenesis in vivo (p=0.01). Further, we observed a strong selective pressure against exogenous SETDB1 expression in moribund mice. Transcriptome analyses demonstrate that SETDB1 globally represses Hox and pluripotency gene programs. Strikingly, we found that SETDB1 represses many of the same genes that exhibit reduced promoter H3K9me3 in AML patient samples relative to CD34+ cells. These data point to a role for SETDB1 in negatively regulating pro-leukemic target genes and suppressing AML.

We also explored how chemical and genetic inhibition of H3K9 methylation and Setdb1 affects AML initiation and maintenance. We first confirmed the previously reported requirement for Setdb1 in AML cell lines by genetically deleting both alleles of Setdb1 in MLL-AF9 cells, which resulted in a complete arrest of proliferation (Koide, et al. Blood 2016). Combined with our data presented above, these results suggest a narrow window of SETDB1 expression is maintained in AML cells. To achieve reduced (but not complete loss of) activity, we investigated how small molecule inhibition of H3K9 methylation (UNC0638) or shRNA mediated knock down of Setdb1 affects AML initiation. We observed increased ex vivo colony formation of normal ckit⁺ bone marrow cells upon shRNA mediated knockdown of Setdb1 or upon UNC0638 treatment. We hypothesized that this expansion of colony forming unit potential of hematopoietic cells may translate to increased transformation potential by leukemic oncogenes. Indeed, cells pretreated with UNC0638 followed by retroviral transduction with MLL-AF9 exhibit significantly higher capacity for leukemic colony formation than vehicle treated cells. These data are consistent with H3K9 methylation repressing genes required for AML transformation.

Our data identified a narrow window of expression of SETDB1 in AML patient samples. SETDB1 expression is reduced in AML patients relative to normal cells and chemical inhibition of H3K9 methylation expands the pool of cells amenable to MLL-AF9 mediated transformation ex vivo. While inhibition of SETDB1 and other H3K9 methyltransferases has been suggested as a possible therapeutic strategy, our data suggests this may also prime bone marrow cells for transformation by inhibiting epigenetic processes that repress pro-leukemic target genes. Further investigation of the roles of SETDB1 and H3K9 methylation levels is necessary to determine the value of these epigenetic modifiers as therapeutic targets in AML and is currently ongoing.