Preclinical efficacy of dual BET/HAT inhibitor–based combinations against post myeloproliferative neoplasm secondary AML cells

Abstract In MPN stem/progenitor cells, the co-occurrence of additional mutations in chromatin/transcriptional regulators such as DNMT3A, ASXL1, RUNX1 and TP53, in addition to the driver mutations in JAK2, CALR or MPL, creates a dysregulated transcriptome and proteome. This is responsible for increased growth, cell survival, and transformation to post MPN secondary AML (sAML), therapy refractoriness and poor clinical outcome. We have previously shown that targeting BET proteins such as BRD4 with BET inhibitors (BETi) disrupted RNA Pol II-mediated transcription leading to depletion of oncoproteins including c-Myc, PIM1, CDK6 and BCL2 but increased the levels of p21, p27, and HEXIM1 in post-MPN sAML cells. This resulted in growth inhibition and apoptosis of post-MPN sAML cells. Notably, treatment with BETi (e.g., OTX015) was shown to reduce leukemia burden and improve survival in xenograft models of post-MPN sAML cells. However, BETi resistance and BETi-refractory disease develop uniformly. CRISPR-Cas9 based guide-RNA dropout screens have confirmed that BRD4 and the histone acetyltransferase EP300 are dependencies in post-MPN sAML cells, supporting the rationale to target both proteins with a BET/HAT inhibitor (i). Here we determined the gene-expression perturbations and lethal activity of the dual BET/HATi EP31670 (EP) as monotherapy and in combination with other targeted agents to achieve superior preclinical efficacy in post-MPN sAML cells that are sensitive or resistant to JAKi or BETi. Utilizing RNA-Seq in SET2 cells and patient-derived (PD) sAML cells with JAK2-V617F or Calreticulin mutation, we determined that treatment with EP negatively enriched for the gene-set mRNA expressions of MYC targets, IL6-JAK-STAT3 signaling and IL2-STAT5-signaling. EP treatment also significantly repressed MYC, MYB, CDK6 and PIM1, with concomitant induction of DCXR, HEXIM1, and p21 mRNA levels. Consistent with this, following EP treatment for 24 hours, Western analysis of the cell lysates from EP-treated SET2 and HEL92.1.7 cells revealed reduced protein levels of c-Myb, c-Myc, CDK6, and PIM1, with upregulation of HEXIM1, p21, and p27. Mass spectrometry analysis of the EP-treated cell lysates from SET2 cells demonstrated negative enrichment of the gene-sets of MYC targets, IL6-JAK-STAT3 and IL2-STAT5-STAT5 signaling. Treatment with EP31670 induced cell cycle growth arrest with significant accumulation of cells in the G0/G1 phase and decreased cells in G2/M phase of the cell cycle. Compared to pan-BET inhibitor INCB057643, treatment with EP31670 induced significantly greater in vitro lethality in HEL92.1.7 and SET2 cells. Notably, in vitro treatment of cell lines and PD post-MPN sAML cells with EP31670 in combination with ruxolitinib, for 72 to 96 hours induced synergistic lethality, as determined by SynergyFinder. Importantly, pretreatment with 50-100 nM of EP31670 for 96 hours significantly increased the sensitivity of JAKi-resistant SET2 and HEL92.1.7 cells to JAKi-induced lethality and BETi-resistant sAML cells to BETi-induced lethality. This suggested that EP31670 could overcome resistance to JAKi and BETi in MPN-sAML cells. We recently reported the preclinical efficacy of CDK7 inhibitor and CDK7 inhibitor-based combinations in post-MPN sAML cells. Therefore, here we also discovered that co-treatment with EP31670 and CDK7i SY-5609 was synergistically lethal in sAML cell lines including those resistant to JAKi or BETi, as well as PD MPN-sAML cells. We next determined in vivo anti-leukemia efficacy of EP31670 alone (7.5 mg/kg, p.o., daily) and/or SY-5609 (3 mg/kg, p.o., daily) for 5 weeks in NSG mice engrafted with SET2-GFP/Luciferase xenografts. EP31670 monotherapy significantly reduced MPN-sAML burden and improved the survival of the mice without causing weight loss or other toxicities. Notably, co-treatment with EP31670 and SY-5609 resulted in significantly greater reduction in leukemia burden and overall survival of the mice than treatment with each agent alone (p < 0.05). These findings demonstrate promising preclinical activity of EP31670 against cellular models of MPN-sAML and strongly support the rationale to further evaluate the in vivo efficacy of EP31670-based combinations against advanced MPN with excess blasts or MPN-sAML.