Abstract 3431: Epigenetic reactivation of the DLC1 tumor suppressor protein is a new approach for cancer treatment

DLC1 is a potent tumor suppressor protein, which is frequently downregulated in cancer. In addition, several types of solid tumors express DLC1 mRNA but have lost DLC1 protein through its destabilization. Here, we identify a methyltransferase, EZH2, that directly methylates the DLC1 protein, leading to its ubiquitin-dependent proteasomal degradation, and explore the clinical relevance of this finding. Our studies indicate that although the canonical activity of EZH2 is on histone H3 in the nucleus, the methylation of DLC1 protein occurs in the cytoplasm, where EZH2 protein is readily detectible and is active. EZH2-dependent methylation of DLC1 leads to its degradation induced by the Cullin-4 ubiquitin ligase. In several tumor cell lines, including NSCLC A549, that carry mutant KRAS and express DLC1 mRNA but lack detectable DLC1 protein, treatment with the EZH2 inhibitor Tazemetostat stabilized the DLC1 protein. As previous studies have indicated that activated KRAS can induce EZH2, we asked whether there might be a link between the mutant KRAS and DLC1 protein stability. Consistent with this hypothesis, siRNA knockdown of KRAS leads to substantial stabilization of the DLC1 protein in A549 cells and other tumor lines with mutant KRAS. Although EZH2 inhibitors can stabilize the DLC1 protein, this restoration of DLC1 has limited tumor suppressor activity because AKT and SRC kinases, which we found directly phosphorylate DLC1 and cooperatively attenuate its tumor suppressor functions, tend to be activated in solid tumors, thereby reducing the tumor suppressor activity of the stabilized DLC1 protein. Treatment of tumor xenografts that carry mutant KRAS with AKT or SRC kinase inhibitors in conjunction with EZH2 inhibition has much greater antitumor activity than treatment with any one of the three inhibitors, and the combined treatment with all three drugs has even greater therapeutic activity, which is correlated with the induction of greater cellular senescence and apoptosis, as measured by the beta-galactosidase and annexin V expression, respectively. Remarkably, these combinations are well-tolerated, and their therapeutic efficacy against xenografts with mutant KRAS make them a candidate intervention against this important unmet clinical need. We have made similar observations with proteasomal inhibition together with AKT and/or SRC inhibition in A549 cells and other tumor lines that carry mutant KRAS and express DLC1 mRNA, but lack DLC1 protein. Inhibitory