Pre-Clinical Efficacy of Co-Treatment with KDM1A (LSD1) Inhibitor and Ruxolitinib or BET Inhibitor Against Post-MPN-sAML Blast Progenitor Cells

LSD1 (KDM1A) is an FAD-dependent amine-oxidase that demethylates mono and dimethyl histone H3 lysine 4 (H3K4Me1 and H3K4Me2), which regulates active enhancers and transcription in AML stem/progenitor cells (LSCs). LSD1 is part of the repressor complexes involving HDACs, CoREST or GFI1, mediating transcriptional repression and differentiation block in LSCs that persist in the minimal residual disease (MRD) following attainment of clinical complete remission, leading to relapse and poor outcome in AML. In AML LSCs, genetic alterations and epigenetic dysregulation of enhancers affect levels of myeloid transcriptional regulators, including c-Myc, PU.1, GATA 2 and CEBPα, and their target genes, which are involved in differentiation block in LSCs. Our present studies demonstrate that CRISPR/Cas9-mediated knockout of LSD1 in the AML OCI-AML5 cells significantly increased the permissive H3K4Me2/3-marked chromatin, reduced H3K27Ac occupancy at super-enhancers and enhancers (SEs/Es) (by ChIP-Seq), especially of c-Myc and CDK6, as well as repressed CoREST, c-Myc, CDK6, and c-KIT, while inducing p21, CD11b, and CD86 levels (log2 -fold change by RNA-Seq, and protein expression by Western analyses). This led to significant growth inhibition, differentiation and loss of viability of OCI-AML5 and patient-derived AML blasts (p < 0.01). Similar effects were observed following exposure of OCI-AML5 (96 hours) to tet-inducible shRNA to LSD1. Knock-down of GFI1 by shRNA (by 90%) also inhibited growth and induced differentiation, associated with upregulation of PU.1, p21 and CD11b levels. Treatment with irreversible (INCB059872, 0.25 to 1.0 µM) or reversible (SP2577, 1.0 to 2.0 µM) LSD1 inhibitor (LSD1i) inhibited binding of LSD1 to CoREST, and significantly induced growth inhibition, differentiation and loss of viability (over 96 hours) of the OCI-AML5, post-myeloproliferative neoplasm (post-MPN) sAML SET2 and HEL92.1.7 cells, as well as patient-derived AML and post-MPN sAML blasts (p < 0.01). Co-treatment with INCB059872 and ruxolitinib synergistically induced apoptosis of the post-MPN sAML SET2 and HEL92.1.7 cells and patient-derived CD34+ post-MPN sAML blasts (combination indices < 1.0). Notably, pre-treatment with the LSD1i for 48 hours significantly re-sensitized ruxolitinib-persister/resistant SET2 and HEL92.1.7 cells to ruxolitinib (p < 0.001). We previously reported that treatment with the BET inhibitor (BETi) JQ1 or OTX015 represses SE/E-driven AML-relevant oncogenes including MYC, RUNX1, CDK6, PIM1, and Bcl-xL, while inducing p21 and p27 levels in post-MPN sAML blasts (Leukemia 2017;31:678-687). This was associated with inhibition of colony growth and loss of viability of AML and post-MPN sAML blasts (p < 0.01). Here, we determined that INCB059872 treatment induced similar levels of lethality in BETi-sensitive or BETi-persister/resistant AML and post-MPN sAML cells. Since BETi treatment also depleted LSD1 protein levels, co-treatment with the BETi OTX015 and LSD1i INCB059872 or SP2577 induced synergistic lethality in AML and post-MPN sAML blasts (combination indices < 1.0). Co-treatment with INCB059872 (1.5 mg/kg) and OTX015 (50 mg/kg) both orally for 21 days, compared to each agent alone or vehicle control, significantly reduced the sAML burden and improved survival of immune-depleted mice engrafted with HEL92.1.7 or HEL92.1.7/OTX015-resistant-GFP/Luc sAML xenografts (p < 0.01). Collectively, these findings strongly support further in vivo testing and pre-clinical development of LSD1i-based combinations with ruxolitinib against post-MPN sAML and with BETi against AML or post-MPN sAML cells.