The molecular understanding of bladder cancers, whether muscle-invasive (MIBC) or non-muscle-invasive (NMIBC), currently relies primarily on transcriptomic and genomic analyses. These approaches, however, fail to capture part of the biological information, since protein abundance cannot be directly inferred from messenger RNA levels. To better characterize the heterogeneity of these tumors and identify processes specific to certain subgroups, the team undertook a proteogenomic characterization combining proteomics, genomics, and transcriptomics on a single case series.
Proteomic data were generated for 40 MIBC and 23 NMIBC that had already been documented at the transcriptomic and genomic levels. An unsupervised analysis revealed five proteomic groups spanning both NMIBC and MIBC, showing a broad resemblance to the known transcriptomic subtypes while uncovering internal heterogeneity within these subtypes. In particular, the proteome made it possible to group together tumors classified differently by transcriptomics but sharing common features, such as high proliferation and genomic instability. One of these groups was associated with the Ta pathway and enriched in FGFR3 mutations.
The most striking contribution lies in the demonstration of an overrepresentation of proteins involved in apoptosis within FGFR3-mutated tumors, a signal that was invisible at the transcriptomic level. To explore this phenomenon, the researchers tested five bladder cancer–derived cell lines, four of which carried FGFR3 alterations (FGFR3 mutations or FGFR3-TACC3 fusions). All altered lines proved sensitive to TRAIL (tumor necrosis factor–related apoptosis-inducing ligand), whereas an FGFR3 wild-type line, already known to be resistant, served as a negative control. Inhibition of FGFR3, either by the pan-inhibitor erdafitinib or by RNA interference, significantly attenuated this sensitivity, demonstrating its dependence on FGFR3 activity. Mechanistically, FGFR3 activation promotes the accumulation of the TRAIL-R1 receptor and the decrease of c-FLIP, an endogenous inhibitor of caspases-8/10. The combination of TRAIL with birinapant, a SMAC mimetic, also proved synergistic in three cell lines (ZIP scores above 10), enhancing apoptosis induction and reducing viability.
The authors acknowledge a limitation related to the small sample size (23 NMIBC and 40 MIBC), partially offset by the use of validation cohorts, and note that some FGFR3 wild-type tumors also respond to TRAIL through mechanisms that remain to be elucidated. This proteogenomic resource refines the classification of bladder tumors and identifies subtype protein markers that can be exploited by immunohistochemistry. As anti-FGFR agents face the rapid emergence of resistance, apoptosis induction by TRAIL, potentiated by a sensitizer such as birinapant, represents an alternative therapeutic option for FGFR3-mutated tumors, warranting clinical investigation.