Abstract 5429: Inhibition of SMARCA2: a novel target for SMARCA4-deficient lung adenocarcinoma

Aim: With the decreasing costs of genomics technologies, ever more data is being put into the public domain. Scientific papers only highlight a fraction of the information in the data, consequently further mining can answer drug discovery relevant questions and identify new targets. In this work we developed a bioinformatics pipeline, based on the collateral vulnerability hypothesis, to integrate several sources of public data and identify novel targets to form the basis of a new drug discovery project. Methods: Genomic data from TCGA was integrated with phenotypic data extracted from Mousemine, Flymine and Wormbase to identify loss-of-function aberrations in genes from families with essential predicted function. Follow-up experiments investigated the effect of siRNA knockdown of paralogs of genes of interest on various cellular phenotypes including proliferation, survival and senescence in gene deficient cell lines. A fragment screen was used to assess drugability of genes of interest. Results: The pipeline has been applied to several cancer types, and as a result a drug discovery project has been initiated against SMARCA2 in SMARCA4-deficient lung adenocarcinoma. SMARCA4 is a bromodomain-containing transcriptional co-activator within the multi-subunit SWF/SNF complex, which also possesses helicase and ATPase activities and functions to alter chromatin structure. SMARCA4-deficient cell lines harbour abrogating mutations, and previous studies have demonstrated that knockdown of SMARCA2, its functional paralog, in SMARCA4-deficient cell lines results in reduced cellular proliferation and survival. Moreover, SMARCA2 has been shown to be inactivated by epigenetic silencing in a proportion of human tumours. The collateral vulnerability hypothesis was tested in a panel of lung adenocarcinoma cell lines with SMARCA2- and/or SMARCA4-deficiencies. Experiments investigating the effect of siRNA knockdown confirmed both our hypothesis and the published data. A fragment screen against the bromodomain of SMARCA2 generated a high ‘ligandability’ index, suggesting that this target is druggable. Conclusion: SMARCA2 has been validated by our work and others as a target in SMARCA4 deficient lung adenocarcinoma. Future work will focus on elucidating the role of the bromodomain and the ATPase domain in SMARCA2/4 activity, and we are actively pursuing the identification of small molecule inhibitors of SMARCA2. An HTS has been undertaken against a library of >700 million compou