Venetoclax-combination therapies are becoming the standard-of-care in acute myeloid leukemia (AML). However, the therapeutic benefit of these drugs in older/unfit patients is limited to only a few months, highlighting the need for more effective therapies. PP2A is a tumor suppressor phosphatase with pleiotropic functions that becomes inactivated in ~70% of AML cases. PP2A promotes cancer cell death by modulating the phosphorylation state in a variety of proteins along the mitochondrial apoptotic pathway. We therefore hypothesized that pharmacological PP2A reactivation could increase BCL2 dependency in AML cells and thus potentiate venetoclax-induced cell death. Here, by using three structurally distinct PP2A-activating drugs, we show that PP2A reactivation synergistically enhances venetoclax activity in AML cell lines, primary cells, and xenograft models. Through the use of CRISPR-Cas9 models and pharmacologic approaches, we demonstrate that the observed therapeutic synergy relies on PP2A complexes containing the B56α regulatory subunit, which expression dictates response to the combination therapy. Mechanistically, PP2A reactivation enhances venetoclax-driven apoptosis through simultaneous inhibition of anti-apoptotic BCL2 and ERK signaling, the later decreasing MCL1 protein stability. Finally, PP2A targeting increases the efficacy of the clinically approved venetoclax and azacitidine combination in vitro, in primary cells, and in an AML patient-derived xenograft model. These preclinical results provide a scientific rationale for testing PP2A-activating drugs with venetoclax combinations in AML.

No abstract

The in-frame internal tandem duplication (ITD) of the FLT3 gene is an important negative prognostic factor in acute myeloid leukemia (AML). FLT3-ITD is constitutive active and partially retained in the endoplasmic reticulum (ER). Recent reports show that 3′UTRs function as scaffolds that can regulate the localization of plasma membrane proteins by recruiting the HuR-interacting protein SET to the site of translation. Therefore, we hypothesized that SET could mediate the FLT3 membrane location and that the FLT3-ITD mutation could somehow disrupt the model, impairing its membrane translocation. Immunofluorescence and immunoprecipitation assays demonstrated that SET and FLT3 co-localize and interact in FLT3-WT cells but hardly in FLT3-ITD. SET/FLT3 interaction occurs before FLT3 glycosylation. Furthermore, RNA immunoprecipitation in FLT3-WT cells confirmed that this interaction occurs through the binding of HuR to the 3′UTR of FLT3. HuR inhibition and SET nuclear retention reduced FLT3 in the membrane of FLT3-WT cells, indicating that both proteins are involved in FLT3 membrane trafficking. Interestingly, the FLT3 inhibitor midostaurin increases FLT3 in the membrane and SET/FLT3 binding. Therefore, our results show that SET is involved in the transport of FLT3-WT to the membrane; however, SET barely binds FLT3 in FLT3-ITD cells, contributing to its retention in the ER.