Comprehensive analysis of a novel RNA modifications-related model in the prognostic characterization, immune landscape and drug therapy of bladder cancer

Bladder cancer (BCa) is the leading reason for death among genitourinary malignancies. RNA modifications in tumors closely link to the immune microenvironment. Our study aimed to propose a promising model associated with the "writer" enzymes of five primary RNA adenosine modifications (including m6A, m6Am, m1A, APA, and A-to-I editing), thus characterizing the clinical outcome, immune landscape and therapeutic efficacy of BCa.

We conferred the first landscape of five forms of RNA modifications in BCa and emphasized the excellent power of an RNA modifications-related model in evaluating BCa prognosis and immune landscape.

Unsupervised clustering was employed to categorize BCa into different RNA modification patterns based on gene expression profiles of 34 RNA modification "writers". The RNA modification "writers" score (RMS) signature composed of RNA phenotype-associated differentially expressed genes (DEGs) was established using the least absolute shrinkage and selection operator (LASSO), which was evaluated in meta-GEO (including eight independent GEO datasets) training cohort and the TCGA-BLCA validation cohort. The hub genes in the RMS model were determined via weighted gene co-expression network analysis (WGCNA) and were further validated using human specimen. The potential applicability of the RMS model in predicting the therapeutic responsiveness was assessed through the Genomics of Drug Sensitivity in Cancer database and multiple immunotherapy datasets.

Two distinct RNA modification patterns were determined among 1,410 BCa samples from a meta-GEO cohort, showing radically varying clinical outcomes and biological characteristics. The RMS model comprising 14 RNA modification phenotype-associated prognostic DEGs positively correlated with the unsatisfactory outcome of BCa patients in meta-GEO training cohort (HR = 3.00, 95% CI = 2.19-4.12) and TCGA-BLCA validation cohort (HR = 1.53, 95% CI = 1.13-2.09). The infiltration of immunosuppressive cells and the activation of EMT, angiogenesis, IL-6/JAK/STAT3 signaling were markedly enriched in RMS-high group. A nomogram exhibited high prognostic prediction accuracy, with a concordance index of 0.785. The therapeutic effect of chemotherapeutic agents and antibody-drug conjugates was significantly different between RMS-low and -high groups. The combination of the RMS model and conventional characteristics (TMB, TNB and PD-L1) achieved an optimal AUC value of 0.828 in differentiating responders from non-responders to immunotherapy.