Bladder cancer organoids as a translational models to model disease biology and assess new therapies

Background : Inter-patient heterogeneity in bladder cancer (BC) remains a major driver of treatment failure, highlighting the need for more personalized therapeutic approaches. Org anoids represent valuable preclinical models capturing this diversity. We established a BC organoid biobank comprising 20 models that have been extensively characterized at phenotypic, pharmacological and molecular levels. The aim of this study was to demonstrate the relevance of our biobank in terms of specifications and disease representation. In addition, we illustrated the utility of these models for evaluating targeted therapies, using Enfortumab-vedotin (EV) as a proof of concept. Methods : Urothelial progenitors were isolated from tumor samples, seeded in Matrigel ® and culture medium was added. Molecular characterization was performed by whole exome and transcriptome sequencing. Morphology of organoids was determined by optical microscopy. To evaluate targeted-therapy, organoids were treated with EV (from 0.3 to 10 µg/mL) for 5 days. Cell viability was measured with CellTiter-Glo3D ® assay. Results : All organoid models were developed following a strict quality framework: cultures were expanded beyond passage 6, cryopreserved, and demonstrated a 100% post-thaw recovery rate. The biobank captured the full pathological spectrum, from non-invasive pTa to metastatic pT4 stages. It reflected the anatomical diversity of urothelial carcinoma, with models derived from both bladder and ureteral tumors. Omics analyses revealed the presence of key molecular alterations commonly observed in BC patients, including FGFR3 and TP53 mutations. Organoid morphology, categorized as solid, hollow, or mixed, correlated with tumor stage and metastatic models exhibited features of epithelial-mesenchymal transition. For EV evaluation, we assessed Nectin-4 expression across the biobank and observed heterogeneous expression levels. Interestingly, organoid responses to EV were not strictly correlated with Nectin-4 expression with low response for models with high Nectin-4 expression. Conclusion : Overall, our BC organoid biobank constitutes a high-quality collection of clinically annotated and comprehensively characterizes models ensuring robustness, stability, and reproducibility across experiments. It faithfully captured the pathological diversity and inter-patient heterogeneity observed in BC. These organoid models represent powerful tools for assessing novel targeted therapies, and the availability of a broad biobank enables the extrapolation of patient-specific responses, paving the way toward personalized medicine.