Tumor suppressive role of BMI-1 through inhibition of JAK-STAT signaling in leukemia

Abstract

BMI-1, which is one of the core components of polycomb repressive complex 1, is frequently found deregulated in patients with hematological disorders. In last decades, researchers concordantly agree that BMI-1 mediates tumorigenesis of leukemia stem cells through p16INK4A leukemogenic pathway. However, accumulating evidences contradict the idea that BMI-1 solely plays an oncogenic role in tumorigenesis. It has been shown BMI-1 depletion favors the development of myelofibrosis in mice; whereas high BMI-1 expression suppressed colony forming ability of MLL-ENL-transformed bone marrow and correlated with higher survival in some cancers. In this study, we hypothesized that BMI-1 plays a tumor suppressive role, which is independent of the regulation of INK4A-ARF locus, in human leukemia. BMI-1 was over-expressed in a panel of myeloid and lymphoid lineage leukemia cells, including HL-60, MV4-11, MonoMac-6, SEM, Nalm-20 and RS4;11. We observed no deregulation of p16INK4A and p14ARF genes by BMI-1, suggesting the regulation of INK4A-ARF locus is independent of BMI-1 modulation in leukemia cells. Nevertheless, over-expression of BMI-1 resulted in significant reduction of leukemia cell proliferation. It is noted that constitutively active JAK-STAT signaling pathway is crucial to leukemia cell survival. By modulation of BMI-1 level, we demonstrated suppression of the activated JAK-STAT signaling pathway in most of the leukemia cell lines with the exception of MonoMac-6 and RS4;11. This is in agreement with the high sensitivity to ruxolitinib, a JAK-STAT inhibitor, in all the tested leukemia cell lines except RS4;11. Importantly, we showed that BMI-1 over-expression could sensitize RS4;11 cells to reduce cell proliferation under ruxolitinib treatment. These results suggest that higher efficacy of ruxolitinib treatment could be achieved under a condition of high cellular level of BMI-1. We further demonstrated that ruxolitinib treatment was more effective in a cohort of AML patient samples (n = 25) with a context of higher BMI-1 expression (p < 0.05). Altogether, our results suggest that BMI-1 functions as a tumor suppressor gene via inhibition of the JAK-STAT signaling pathway. The endogenous level of BMI-1 could be served as an indicator for the effective treatment of JAK-STAT-dependent leukemia cells using ruxolitinib.