Genetic or Pharmacological Inhibition of the Polycomb Group Protein BMI1 Impairs Leukemic Transformation Mediated By the CALM-AF10 Fusion Oncoprotein

A subset of acute myeloid and lymphoid leukemia cases harbor a t(10;11)(p13;q14) translocation resulting in the CALM-AF10 fusion gene. Standard chemotherapeutic strategies are not very effective in treating patients with CALM-AF10 fusions. Hence, there is an urgent need to identify molecular pathways dysregulated in CALM-AF10 positive leukemias which may lay the foundation for novel targeted therapies. The polycomb repressive complex 1 gene BMI1 is consistently overexpressed in CALM-AF10 leukemias. Previous studies have shown that CALM-AF10 leukemias express high levels of BMI1, regardless of whether the leukemias are myeloid or lymphoid. Our analysis of TCGA acute myeloid leukemia (AML) data confirmed that AML cells with AF10-rearrangements displayed significantly higher expression of BMI1 transcripts compared to cells from non AF10-rearranged AML patients. These observations indicate that BMI1 may be directly activated by AF10-fusion oncogenes as suggested by our previous studies.

We sought to investigate the role of BMI1 in CALM-AF10 mediated leukemogenesis using murine and human models of CALM-AF10-mediated AML. First, we tested whether BMI1 deficiency can affect CALM-AF10 mediated oncogenic transformation of hematopoietic stem and progenitor cells (HSPCs). Towards this end, we retrovirally transduced fetal liver cells from Bmi1 wild-type, heterozygous or homozygous null mice with the CALM-AF10 fusion oncogene. Upon plating these cells in colony forming unit (CFU) assays, we observed a significant decrease in the colony formation capacity of the CALM-AF10 fusion transduced cells on a Bmi1 deficient background. Next, we performed Cre-recombinase mediated excision of Bmi1 of already transformed CALM-AF10 myeloid leukemia cells (Bmi1 floxed background). Bmi1 deletion led to a significant reduction in the number of total CFUs compared to Bmi1 wild-type cells, with a particularly striking reduction in the number of blast-like colonies. These experiments, using Bmi1 constitutive or conditional knockout-mice, revealed that CALM-AF10 transformed AML cells are dependent on Bmi1.

Recently, selective pharmacological BMI1 inhibitors have been developed. We tested the impact of pharmacologic BMI1 inhibition on a panel of CALM-AF10-driven mouse leukemias with the small molecule inhibitor PTC-209. PTC-209 treatment increased gene expression of the known BMI1-repressed targets Cdkn2a (p16) and Cdkn1a (p21) and led to a dose-dependent decrease in cell proliferation. We also observed a marked increase in Annexin V+ cells upon PTC-209 treatment. In addition, cell-cycle analysis using BrdU incorporation assays revealed a significant decrease in cells in the S-phase, demonstrating that PTC-209 treatment leads to growth arrest and apoptosis in CALM-AF10 AML cells.

In order to confirm these findings in human AML with CALM-AF10 rearrangements, we treated human CALM-AF10 positive AML cell lines P31, U937 and KPMOTS with PTC-209. Consistent with our results in the murine AML model, we observed a time and dose-dependent decrease in proliferation of these human cell lines upon PTC-209 treatment. Drug treated human cells also showed concomitant cell-cycle arrest and apoptosis induction, coupled with an increase in expression of BMI1 repressed tumor suppressor genes such as CDKN2A and CDKN1A.

In summary, our results demonstrate that BMI1 is a bonafide candidate for therapeutic targeting in AML with CALM-AF10 rearrangements and possibly other CALM-AF10 positive leukemias. We are now assessing clinical-grade BMI1 inhibitors for in vivo efficacy in mouse models of CALM-AF10-mediated AML.