Evidence of MicroRNA-29b and Sp1/NFκB-HDAC Regulatory Network for KIT Expression in KIT-Driven Acute Myeloid Leukemia (AML): Biologic and Therapeutic Implications.

Abstract 938

KIT is a receptor tyrosine kinase (RTK) and its aberrant activities resulting from protein overexpression and/or activating mutations are associated with a number of malignancies including core binding factor (CBF) AML [e.g., patients with t(8;21) or inv(16) or molecular equivalent RUNX1/RUNX1T1 or CBFB/MYH11, respectively]. RTK inhibitors (e.g. PKC412) have been shown to suppress aberrant KIT activity and delay tumor growth, but they are active only on distinct types of KIT mutations (KIT^mut). Furthermore, resistance to these inhibitors, as a result of secondary mutations or KIT overexpression, is emerging. Thus, we hypothesize that direct inhibition of KIT gene transcription may be a valuable therapeutic approach to override aberrant KIT expression and activity. Here, we described the regulatory and functional role of Sp1/NFkB-miR29b feedback loop in KIT-driven leukemia that can be targeted pharmacologically. Applying chromatin immunoprecipitation (ChIP) and electrophoretic mobility shift assays (EMSA) to RUNX1/RUNX1T1-positive Kasumi-1 cells, we demonstrated that while, the Sp1/NFkB complex was enriched on KIT promoter and acted as gene transactivator thereby leading to KIT overexpression, Sp1/NFkB recruited HDAC1 and HDAC3 to miR29b regulatory region thereby epigenetically repressing miR29b. This microRNA, when expressed, targeted Sp1 and eventually decreased Sp1/NFkB-mediated gene transactivation, including that of KIT. In agreement with these, we showed that when Sp1, NFkB and HDAC1 were transiently overexpressed in Kasumi-1 cells, increased KIT expression and decreased miR29b transcription occurred. In contrast, siRNA knockdown of Sp1, NFkB and HDAC1 augmented miR29b level and decreased KIT transcription. Moreover, ectopic miR29b expression impaired Sp1/NFkB repressor complex on the promoter of endogenous miR29b, thereby resulting in re-expression of the endogenous microRNA and further inhibition of Sp1/NFkB-dependent KIT transcription. Importantly, the activity of Sp1/NFkB/HDACs complex can also be pharmacologically modulated leading to restored miR29b transcription and abrogated KIT expression. We showed that pharmacologic interference with Sp1/NFkB/HDACs using their respective inhibitors, such as bortezomib (0, 6, 20, 60 and 100nM for Sp1 and NFkB), mithramycin A (150 and 300ng/ml for Sp1), bay 11-7082 (3μM for NFkB) and OSU-HDAC42 (1μM for HDAC), upregulated miR29b at early time point (6 hours) and decreased Sp1 and in turn KIT expression in KIT overexpression cell lines (e.g., Kasumi-1) and AML patient blasts. EMSA and ChIP assay demonstrated that bortezomib or HDAC42-mediated KIT repression and miR29b upregulation occurred through the dissociation of Sp1/NFkB complex from the corresponding promoter. To further investigate the therapeutic potential of targeting KIT over-expression in leukemia, we stably expressed KIT wild type (KIT^wt) or KIT^mut (D816V) in FDC-P1 cell line (murine non-tumorigenic cells derived from myeloid precursors), and we evidenced that both KIT^wt and KIT^mut promoted cell proliferation that was overcome by bortezomib in clonogenic assay. In in vivo study, when NOD/SCID mice were engrafted with FDC-P1/KIT^mut cells (5×10⁶/mouse), they developed significant splenomegaly and marrow blast infiltration through KIT overexpression. When leukemia-carrying mice were treated with bortezomib (1mg/kg) for 48 hours, we observed an obvious increase of endogenous miR29b expression and a significant reduction of KIT expression. Leukemic mice that received 1mg/kg of bortezomib twice/week for 3 weeks starting on day 21 after engraftment (n=5 mice/group) showed no evidence of splenomegaly and had a significantly longer median survival [59 days (twice/week) vs 28 days (vehicle-treated), p=0.0036], compared to vehicle-treated mice that instead showed massive splenomegaly. Cytospin of marrow and histopathology of spleen and liver showed that vehicle-treated mice displayed extensive blast infiltration that was instead absent in bortezomib-treated mice. Altogether, our study revealed a previously unrecognized protein-microRNA regulatory network whose imbalance contributes to KIT-driven leukemia. As the aberrant activity of this network is pharmacologically targetable, this discovery may be quickly translated into the clinic as a novel therapeutic approach for KIT-driven AML and other malignancies.