Targeting KMT2C induces paralog synthetic lethality in KMT2D null DLBCL through DREAM targets regulation

KMT2D is a histone methyltransferase that regulates enhancer activation by catalyzing H3K4 mono/di-methylation, in part through a cooperative interaction with CBP/p300. Epigenetic regulation is commonly perturbed in cancer due to recurrent genetic alterations. In particular, KMT2D inactivating mutations are the most common genetic alterations in germinal center-derived B cell lymphoma, including ~30% of diffuse large B cell lymphoma and up to 80% of follicular lymphoma. To identify novel synthetic lethal targets in KMT2D mutant lymphoma, we performed a genome-wide CRISPR knock-out screen in a panel of KMT2D-wt vs -null DLBCL cell lines and revealed the paralog histone methyltransferase KMT2C as a top hit in KMT2D-null cell lines. Interestingly, KMT2C is rarely mutated in KMT2D mutant lymphoma and its protein expression is well-conserved across DLBCL cell lines, suggesting KMT2C may compensate for the loss of KMT2D methyltransferase activity.To validate KMT2C dependency, we knocked out KMT2C in KMT2D proficient or mutant/deficient DLBCL cell lines as well as in KMT2D isogenic knockout models. KMT2C loss leads to G0/G1 cell cycle arrest and apoptotic induction in the KMT2D deficient but not in KMT2D proficient context. To gain mechanistic insights into the preferential sensitivity of KMT2D-null cell lines, we performed RNA-seq in cell lines with varying phenotypic response to KMT2C knockout. In line with the observed cell cycle arrest phenotype, we identified the DREAM ( D imerization partner, R B-like, E 2F, a nd M uvB) complex targets as significantly downregulated upon KMT2C loss. The DREAM complex is a key regulator of cell cycle progression by repressing multiple genes particularly during cell cycle exit. Epigenetic profiling revealed that KMT2C binds to promoters of specific DREAM targets, suggesting a direct regulatory mechanism. Additionally, the protein DYRK1B is known to promote DREAM complex assembly at target promoters. We noticed KMT2C knockout upregulates DYRK1B protein expression only in KMT2D mutant cell lines. Further, overexpression of DYRK1B in KMT2D mutant lines induced anti-proliferative effects, suggesting dual mechanisms by which KMT2C regulates DREAM targets. These results provide mechanistic rationales into how KMT2C loss induces paralog lethality. Lastly, to identify potential combination partners with KMT2C knockout, we performed a combination screen using different small molecules for DLBCL treatment. We observed that inhibitors targeting CBP/p300 and chemotherapy agents like vincristine and doxorubicin combined with KMT2C loss drive deeper responses in KMT2D mutant models. Taken together, our data demonstrate that B cell lymphoma carrying KMT2D mutations are addicted to the residual KMT2C activity and suggest that targeting KMT2C as a monotherapy or in combination may benefit patients with KMT2D loss.