Molecular subtypes of muscle-invasive bladder cancer (MIBC) display differential survival and drug sensitivities in clinical trials. To date, they have not been used as a paradigm for phenotypic drug discovery. This study aimed to discover novel subtype-stratified therapy approaches based on high-content screening (HCS) drug discovery. Transcriptome expression data of CCLE and BLA-40 cell lines were used for molecular subtype assignment in basal, luminal, and mesenchymal-like cell lines. Two independent HCSs, using focused compound libraries, were conducted to identify subtype-specific drug leads. We correlated lead drug sensitivity data with functional genomics, regulon analysis, and in-vitro drug response-based enrichment analysis. The basal MIBC subtype displayed sensitivity to HDAC and CHK inhibitors, while the luminal subtype was sensitive to MDM2 inhibitors. The mesenchymal-like cell lines were exclusively sensitive to the ITGAV inhibitor SB273005. The role of integrins within this mesenchymal-like MIBC subtype was confirmed via its regulon activity and gene essentiality based on CRISPR–Cas9 knock-out data. Patients with high ITGAV expression showed a significant decrease in the median overall survival. Phenotypic high-content drug screens based on bladder cancer cell lines provide rationales for novel stratified therapeutic approaches as a framework for further prospective validation in clinical trials.
Bibliographical noteFunding Information:
This research was supported in part by grants awarded to D.V.L. from the U.S. Department of Defense Peer-Reviewed Cancer Research Program (W81XWH-18-1-0142) and the Colorado Cancer Translational Research Accelerator. The research was also supported in part by a fellowship grant awarded to S.R. by the Cancer Foundation of Luxembourg. We thank the Drug Discovery and Development Shared Resource (D3SR, RRID:SCR_021986). The D3SSR is part of the CU AMC Center for Drug Discovery, which was established with a generous grant from The ALSAM Foundation and with CU Anschutz institutional support. We also thank the Biostatistics and Bioinformatics Shared Resource (BBSR, RRID:SCR_021983). The D3SR and BBSR are also supported in part by the University of Colorado Cancer Center, an NIH NCI designated center (P30CA046934).
© 2022 by the authors.