Biomarkers associated with BRM inhibition

#The disclosure is directed to novel personalized therapies, kits, transmittable forms of information and methods for use in treating patients having cancer.

#The mammalian SWI/SNF (mSWI/SNF) multi-protein complexes regulate chromatin structure through ATP-dependent nucleosome remodeling and thereby control many key cellular processes (Wilson, B. G. and C. W. Roberts (2011). Nat Rev Cancer 11(7): 481-492). Several subunits of the mSWI/SNF complexes have roles as tumor suppressors, and recent genomic studies revealed recurrent mutations in several of these subunits, with a collective mutation frequency of approximately 20% across all cancers (Kadoch, C., Hargreaves, D. C., et al. (2013) Nat Genet. 45: 592-601). The catalytic SWI/SNF subunit BRG1, also known as SMARCA4, is frequently mutated in lung adenocarcinomas and other cancer types (Becker, T. M., S. Haferkamp, et al. (2009) Mol Cancer 8: 4.)(Imielinski, M., A. H. Berger, et al. (2012) Cell 150(6): 1107-1120).

#The mechanisms by which mSWI/SNF mutations contribute to tumorigenesis remain poorly understood, and their inactivation presents a challenge for devising therapeutic strategies against these genetic lesions. Mutations in specific subunits of mSWI/SNF complexes are found in distinct cancer types. Overall, mSWI/SNF complexes have emerged as the most frequently mutated class of chromatin regulators in cancer; at least six subunits of the complex have been found to be specifically inactivated at high frequencies in cancers, including subsets of ovary, breast, kidney, lung, pancreas, uterus, bladder, stomach, colon, and liver (Kadoch, C., Hargreaves, D. C., et al. (2013) Nat Genet. 45: 592-601).

#FIG. 1. An epigenome wide pooled shRNA screen identifies BRM as a synthetic lethal target in BRG1-mutant cancer cell cells. a. Waterfall plot showing the log of the p-value calculated with the RSA statistic (Konig, R., C. Y. Chiang, et al. (2007) Nat Methods 4(10): 847-849) reporting the sensitivity of each cell line in the pooled shRNA screen to knockdown of BRM. Bars colored in black are cell lines with loss of BRG1 function either due to BRG1 mutation or loss of BRG1 expression. Bars colored in grey are cell lines that are wild type (WT) for BRG1. Cell lines clustered on the right of the waterfall plot with more negative p-values are those that show growth inhibition upon BRM knockdown and all have BRG1 loss of function. b. Waterfall plots for the z-scores (Birmingham, Ala., L. M. Selfors, et al. (2009) Nat Methods 6(8): 569-575) of the 12 individual BRM shRNAs that show a statistically significant correlation in their activity profile as calculated by the ATARIS algorithm (Shao, D. D., A. Tsherniak, et al. (2013) Genome Res 23(4): 665-678), colored is as in (a). Cell lines with more negative z-scores are those that show growth inhibition upon knockdown of BRM with each individual shRNA.

#FIG. 2. Sensitivity to BRM shRNAs corresponds with loss of BRG1 protein expression. Western blot demonstrating protein levels of BRG1 and BRM in representative BRG1-mutant and wild type cancer cells lines profiled in the epigenome pooled shRNA screen. BRG1 expression is not detected in BRG1-mutant cell lines, but is present in BRG1-WT cell lines. CORL23 is a cancer cell line with a heterozygous BRG1 mutation (K689de1), and thus still retains BRG1 expression. The SBC5 cell line, which shows both loss of BRG1 and BRM expression expression by Western blot was not sensitive to BRM shRNAs, as shown in FIG. 1, demonstrating that retention of BRM expression is required for BRG1-mutant cell lines to be sensitive to BRM inhibition.

#FIG. 3. BRM targeting shRNAs produce efficient BRM knockdown and do not result in off-target killing in a BRM-deficient cell line SW13. a. Quantitative Reverse Transcription (QRT-PCR) showing significant reduction of BRM mRNA transcripts in HCT116 cells upon induction of BRM shRNAs but not non-targeting control (CTL) shRNAs or parental cells with Doxycycline (Dox) (48 hours, 100 ng/mL). Values are shown as mean±standard deviation. BRM targeting shRNAs used in the pooled shRNA screening show comparable knockdown efficiency to BRM shRNAs 2025 and 5537 used in follow-up experiments. b. Induction of non-targeting control shRNAs as well as BRM shRNA 2025(c) or BRM shRNA 5537 (d) with Dox does not inhibit the growth of BRG1 and BRM-deficient SW13 cells as expected in a standard proliferation assay or (e) in a colony formation assay

#Epigenetic dysregulation is an emerging hallmark of cancers, and the identification of recurrent somatic mutations in chromatin modifying enzymes implies a causal role for altered chromatin states in tumorigenesis. (You, J. S. and P. A. Jones (2012) Cancer Cell 22(1): 9-20).

#While the majority of epigenetic mutations are inactivating and thus do not present directly druggable targets, we reasoned that these mutations may alter the epigenomic state of cancer cells and thereby induce novel epigenetic vulnerabilities. To systematically search for epigenetic synthetic lethal interactions, we performed a pooled-shRNA screen across a large collection of cancer cell lines using a library targeting a diverse set of epigenetic regulators.

#While RNAi has proven to be a very powerful forward genetic approach, the robustness and reproducibility of RNAi screens has been challenged by the prevalence of off-target effects and inability to predict high-potency shRNAs with great confidence (Sigoillot, F. D., and King, R. W., (2011 ACS Chem Biol 6(1): 47-60). In an effort to overcome these limitations, we constructed a deep coverage shRNA library which we reasoned to yield higher confidence hits clue to the extensive shRNA coverage for each gene. The shRNA library employed in this work contained 17 shRNAs per gene against a diverse collection of epigenetic regulators, with a particular focus on druggable genes, and was interrogated across a panel of 58 cell lines from the Novartis Cancer Cell Line Encyclopedia (CCLE) (Barretina, Caponigro et al. (2012) Nature 483(7391): 603-607).