Brg1-Inactivating Mutations As Potential Predictive Markers For Aurora Kinase A-Targeted Therapies In Non-Small Cell Lung Cancers (Nsclcs)

Purpose: Here, we sought to identify critical genes whose depletion and inhibition cause differential cytotoxicity in NSCLCs with BRG1 loss in comparison to the cells harboring wild-type BRG1. Background: The SWI/SNF complex is a master regulator of gene expression, affecting at least 10% of the transcriptome. Because of its crucial role in controlling cell cycle, development, differentiation and transcription, the SWI/SNF complex components function as tumor suppressor genes. For that reason, not surprisingly, the ATPases BRM and BRG1 are frequently down regulated in many cancers and mutations of BRG1 are frequently present in lung cancer cell lines. To identify BRG1-inactivating mutation-related targetable gene products, we developed and applied a high throughput cell-based one-well/one-gene screening platform with a genome wide synthetic library of chemically synthesized small interfering RNAs. Using this approach, we have identified a molecular target (BRG1) frequently present in lung cancer cells whose loss leads to increased sensitivity to a class of Aurora kinase A inhibitors. Materials, Methods and Results: BRG1-mutant NSCLC cell line H1819 was screened with the Dharmacon genome-wide library of ∼22,000 pools of 4 siRNA oligonucleotides, identifying 38 pools that were 50% more toxic than the median toxicity in the screen, and were not toxic for wild-type BRG1-expressing HBEC30-KT cells. Among our high-priority toxic siRNAs, a number of targeted genes encode proteins important for the cell cycle, including RAN, and TPX2. RAN is required for TPX2 to function in mitosis and TPX2 binds to microtubules and activates Aurora kinase A. Individual siRNA oligonucleotides targeting these “hits” in the primary screen, as well as siRNA to Aurora kinase A, were tested and shown to inhibit the growth of H1819 but not HBECs. In addition, H1819 cells were hypersensitive to VX-680, an inhibitor of Aurora kinases. We then tested a panel of other NSCLC cell lines that either expressed or did not express wild-type BRG1 for sensitivity to siRNA targeting Aurora kinase A, or VX-680. The cell lines with loss of BRG1 were more sensitive to loss of AURKA or to VX-680 than NSCLC or HBEC cells expressing BRG1. Our current efforts focus to expand our observations with xenograft mouse models and to deeply understand the relationship between BRG1 and AURKA. Conclusions: BRG1-inactivating mutations make cells dependent to RAN/TPX2-mediated mitotic spindle assembly machinery and create targetable mitotic vulnerabilities in NSCLCs. BRG1-mutant NSCLCs are sensitive to Aurora kinase A-targeted treatments due to its crucial role on RAN/TPX2-dependent mitotic spindle formation. Our current data, thus far, suggest that wild-type BRG1-expressing cells tolerate the inhibition of AURKA due to properly regulated and, therefore, functionally normal centrosomes whereas BRG1 loss leads to centrosomal defects. Citation Format: Vural Tagal, Shuguang Wei, Wei Zhang, Bruce A. Posner, John D. Minna, Adi F. Gazdar, Michael G. Roth. BRG1-inactivating mutations as potential predictive markers for Aurora kinase A-targeted therapies in non-small cell lung cancers (NSCLCs). [abstract]. In: Proceedings of the 105th Annual Meeting of the American Association for Cancer Research; 2014 Apr 5-9; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2014;74(19 Suppl):Abstract nr 573. doi:10.1158/1538-7445.AM2014-573