Identification of Putative Gene-Target Modulators of Radiosensitivity in Bladder Cancer Cell Lines (BlaCCL)

Purpose/objective(s) We aimed to computationally derive potential modulators of radiosensitivity in an expanded panel of BlaCCL based upon whole-transcriptomic RNA-seq analysis integrated with functional enrichment and interaction analysis. Materials/methods Twenty established BlaCCL, verified by genomic fingerprinting, were profiled with RNA-seq with raw counts normalized via log (TPM+1). We performed colony forming assays on all cell lines in triplicate with radiosensitivity measured via area-under-the-cell-survival curve (AUC) at varying doses of radiation (0-8 Gy). Gene signatures of radiosensitivity and resistance were generated via unsupervised matrix decomposition of the RNA-seq dataset using statistical software run at k = 13 for 10,000 iterations correlated with AUC via Spearman's rho. Putative modulators of radiosensitivity were identified as hub proteins in protein-protein interaction networks (via String-DB) of the native gene signatures and of derived sets of upstream kinases with hypergeometric P ≤ 0.05. Results RNA-seq analysis identified 18,634 genes for the panel of BlaCCL, of which updated satisfactory colony forming results were obtained for 18 cell lines with AUC (range = 3.40-5.94). Generated signatures of radiosensitivity and radioresistance based on baseline gene expression consisted of 2,192 and 282 unique genes respectively. Our radiosensitivity signature contains 1,543 unique genes compared to previously described signatures of response in patients undergoing definitive chemoradiation, i.e., signatures of T-cell inflamed tumors and interferon-gamma response. Network analyses of the gene signatures reveal hub proteins that connect networks of proteins related to inflammatory response, oxidative stress, DNA-damage repair, cell cycle and others. These include TNF, VNN1, ESR1, H2AC14, RXRG, FPR2, ADCY1, and S1PR4. Kinase enrichment analysis revealed hub proteins previously described in the radiation response including ATM, ATR, CHEK2, AKT1, SRC, JAK2, and ERK-signaling among others. Conclusion Biologically plausible modulators of radiosensitivity and resistance can be derived from transcriptomic data paired with phenotypic data and functional enrichment analysis. Further mechanistic studies are necessary to confirm potential modulators of the radiation response which may eventually lead to novel radiosensitizing drugs to improve response to chemoradiation in bladder cancer patients.