The Impact of FGFR3 Alterations on the Tumor Microenvironment and the Efficacy of Immune Checkpoint Inhibitors in Bladder Cancer

Abstract Background Currently, only limited knowledge is available regarding the phenotypic association between fibroblast growth factor receptor 3 (FGFR3) alterations and the tumor microenvironment (TME) in bladder cancer (BLCA). Methods A multi-omics analysis on 389 BLCA and 35 adjacent normal tissues from a cohort of OMPU-NCC Consortium Japan was retrospectively performed by integrating the whole-exome and RNA-sequence dataset and clinicopathological record. A median follow-up duration of all BLCA cohort was 31 months. Results FGFR3 alterations (aFGFR3), including recurrent mutations and fusions, accounted for 44% of non-muscle invasive bladder cancer (NMIBC) and 15% of muscle-invasive bladder cancer (MIBC). Within MIBC, the consensus subtypes LumP was significantly more prevalent in aFGFR3, whereas the Ba/Sq subtype exhibited similarity between intact FGFR3 (iFGFR3) and aFGFR3 cases. We revealed that basal markers were significantly increased in MIBC/aFGFR3 compared to MIBC/iFGFR3. Transcriptome analysis highlighted TIM3 as the most upregulated immune-related gene in iFGFR3, with differential immune cell compositions observed between iFGFR3 and aFGFR3. Using EcoTyper, TME heterogeneity was discerned even within aFGFR cases, suggesting potential variations in the response to checkpoint inhibitors (CPIs). Among 72 patients treated with CPIs, the objective response rate (ORR) was comparable between iFGFR3 and aFGFR3 (20% vs 31%; p = 0.467). Strikingly, a significantly higher ORR was noted in LumP/aFGFR3 compared to LumP/iFGFR3 (50% vs 5%; p = 0.022). This trend was validated using data from the IMvigor210 trial. Additionally, several immune-related genes, including IDO1, CCL24, IL1RL1, LGALS4, and NCAM (CD56) were upregulated in LumP/iFGFR3 compared to LumP/aFGFR3 cases. Conclusions Differential pathways influenced by aFGFR3 were observed between NMIBC and MIBC, highlighting the upregulation of both luminal and basal markers in MIBC/aFGFR3. Heterogeneous TME was identified within MIBC/aFGFR3, leading to differential outcomes for CPIs. Specifically, a favorable ORR in LumP/aFGFR3 and a poor ORR in LumP/iFGFR3 were observed. We propose TIM3 as a potential target for iFGFR3 (ORR: 20%) and several immune checkpoint genes, including IDO1 and CCL24, for LumP/iFGFR3 (ORR: 5%), indicating promising avenues for precision immunotherapy for BLCA.