Loss of Asxl1 Cooperates with Oncogenic Nras to Drive CMML Progression

Chronic myelomonocytic leukemia (CMML) belongs to the group of "mixed myelodysplastic/myeloproliferative neoplasm". There is currently no effective chemotherapy to treat CMML. Approximately 30% of CMML cases evolve to acute myeloid leukemia (AML) soon after their initial diagnosis, contributing to the poor prognosis of CMML patients. Ras pathway genes are frequently mutated in CMML patients (NRAS:11%; CBL:10%; KRAS:8%). Our whole exome sequencing of CMML patient samples and targeted sequencing of CMML patient-derived myeloid colonies showed that oncogenic NRAS could serve as an initiating or progression mutation. Consistent with human genetics, all the oncogenic Nras mice we and others characterized (Nras^G12D/+, Nras^G12D/G12D, and Nras^Q61R/+), develop myeloproliferative (MP)-CMML like phenotypes.However, oncogenic NRAS alone is insufficient to drive CMML progression or transformation to AML.

Additional sex combs-like 1 (ASXL1) is a human homolog of fly Asx. Mutations in ASXL1are predominantly nonsense and frequently identified in all myeloid malignancies (e.g. 40% in CMML). ASXL1mutations predict inferior overall survival in univariable analysis in multiple large CMML cohort studies and are significantly associated with NRASmutations in CMML patients. Therefore, we hypothesize that loss of Asxl1 cooperates with oncogenic Nrasto drive CMML progression.

To test our hypothesis, we used Vav-Cre to drive oncogenic Nras expression and/or Asxl1 deletion in hematopoietic system. We refer these mice as Nras^G12D/+, Asxl1^-/-and Nras^G12D/+;Asxl1^-/-(NA) mice. Consistent with the literature, Asxl1^-/-mice are normal at 6-week old. Nras^G12D/+ mice showed mild CMML-like phenotypes, including enlarged spleen, increased white blood cell and monocyte counts, and expanded LSKs (Lin^-Sca-1⁺c-Kit⁺) and myeloid progenitors (MPs). Compared to Nras^G12D/+ mice, NA mice showed more sever phenotypes that were associated with hyperactivated ERK signaling in MPs at both basal level and upon GM-CSF stimulation. Our new Nras^G12D/+mice developed a fully penetrant MP-CMML with the median survival of ~430 days. Loss of Asxl1 significantly shortened the median survival of Nras^G12D/+ mice to 220 days. Half of NA mice displayed CMML transformation to AML. The transformed AML was serially transplantable to recipients.

To investigate the mechanism(s) underlying NA-driven CMML/AML, we performed RNA-Seq analysis in Lin^- c-Kit⁺BM cells from age-matched control, Asxl1^-/-, Nras^G12D/+, and moribund NA mice with CMML or AML phenotypes. We found 844 differentially expressed genes (DEG) in AML cells compared to control cells (FDR<0.05 and fold change>2). Among these genes, 341 are common between NA and published Nf1^+/-;Asxl1^+/-AML cells, including AP-1 complex genes (Jun, Junb, Jund and Fosb) that are crucial for malignant propagation in several AML subtypes. The remaining 503 genes are unique in NA AML, including Flt3, one of the mostly mutated genes in AML. These data suggest that our NA AML model demonstrates some distinct mechanisms from Nf1^+/-;Asxl1^+/-model. In-depth analysis of RNA-seq data is still ongoing.

We also examined multiple histone marks using Western blot in BM cells of 6-week old mice. Our data showed that levels of H3K4me3, H3K4me1, and H3K27me3 were downregulated in Asxl1^-/-cells, whereas H3K27Ac level was comparable to that of control cells. In contrast, H3K27Ac level was upregulated in Nras^G12D/+and NA cells, while levels of H3K4me3, H3K4me1, and H3K27me3 were indistinguishable from those of control cells. BRD4 is a member of the BET family of bromodomain-containing proteins that binds to acetylated histones to activate gene transcription. Thus, we hypothesized that Ras/MEK/ERK signaling and BRD/BET are potential targets to inhibit AML progression. To test this hypothesis, we treated the AML cells and BMT recipients using trametinib (a MEK inhibitor approved by FDA to treat melanoma) and GSK525762 (a pan BET inhibitor under clinical development). We demonstrated thatcombined treatment significantly inhibited the growth of leukemia cells in vitro and prolonged the survival of recipients in vivo. It was more effective than single agent alone. Overall, we established a novel CMML/AML mouse model, which represents a significant group of CMML patients with poor prognosis. Our study provides a strong rational to treat these patients with combined MEK and BET inhibition.