Boosting anti-tumor immunity in NRAS;ASXL1-driven Acute Myeloid Leukemia through combined inhibition of MEK and HDACs.

Abstract

Abstract Acute myeloid leukemia (AML) is a highly aggressive blood cancer with 5-year overall survival rates of ~30%. Its treatment outcome is greatly influenced by leukemia driver mutations and T cell phenotypes at the time of diagnosis. Oncogenic NRAS mutations are associated with AML progression, resistance to multiple targeted therapies, and treatment failure in AML. ASXL1 mutations are significantly associated with NRAS mutations in chronic myelomonocytic leukemia (CMML) and correlate with poor prognosis in myeloid malignancies, including AML. We have previously developed an NRAS; ASXL1 mouse model in which loss of ASXL1 co-operates with oncogenic NRAS to promote CMML transformation to secondary AML (NA-AML). NA-AML mice were characterized by an immune-suppressive microenvironment resulting in exhaustion of dysfunctional CD4 and CD8 T cells. Targeting hyperactive RAS/MEK signaling via trametinib (Tra, a MEK inhibitor) attenuated T cell exhaustion and prolonged the survival of NA-AML mice, cementing the role of T cells in modulating AML treatment outcomes. To further improve the efficacy of Tra, we carried out a re-purpose screen of ~2,500 drugs, either approved by FDA or currently under clinical development. We identified and validated inhibition of MEK/ERK signaling via Tra and histone deacetylases (HDACs) via quisinostat (Qui, a 2nd generation of HDAC inhibitor) as an effective combo therapy against mouse and human primary NA-AML cells and non-NA human AML cell lines in vitro. In NA-AML mice, TQ drastically slowed down AML progression and prolonged survival. Surprisingly, TQ only provided moderate survival benefits in immunodeficient NSG mice and in immunocompetent mice with T-cell depletion, suggesting T cells as the primary target of TQ. Further analyses revealed that TQ played a dual role in modulating the epigenomes of both NA-AML and T cells. In NA-AML cells, TQ led to global increase of H3K27Ac level and concomitant decrease of H3K27me3 level. Moreover, TQ upregulated MHC-I and MHC-II expression. Hyperactivation of the JAK/STAT1 pathway, partially through Tra, promoted MHC-I expression, while Qui-mediated HDAC inhibition resulted in upregulation of CIITA, a master transcription co-activator driving MHC-II expression. In CD4 and CD8 T cells, scRNA-Seq and spectral flow cytometry analyses revealed upregulation of H3K27Ac and H3K4me1, elevated JAK/STAT1 signaling, enhanced differentiation of IL2-secreting anti-leukemia Th1 cells, decreased T cell exhaustion, upregulation of genes promoting T cell survival, as well as increased activation and expansion of cytotoxic cluster in T central memory (Tcm) and effector memory (Tem) cells. Furthermore, leukemia:T cell co-cultures showed significantly enhanced cytotoxicity of TQ-treated CD4 and CD8 T cells, which was MHC-dependent and enriched in both Tcm and Tem cells. More importantly, TQ supported expansion and improvement of anti-leukemia killing in AML-associated T cells in vitro, suggesting that autologous T cell transfer may be a useful approach to explore in the future. Not surprisingly, the remaining leukemia cells in the co-cultures and in the moribund TQ-treated NA-AML mice were predominantly MHC-I- MHC-II-. Our results suggest that we must seek additional MHC-independent anti-cancer mechanisms to further improve the therapeutic effects of TQ. We are currently evaluating the effects of combining TQ with anti-TIGIT immune checkpoint blockade to activate endogenous non-MHC restricted natural killer cells in NA-AML mice.