Integrated Immunological Analysis of Interactions Between CXCL13+ T Cells and LAMP3+ DC in the Immunotherapy Response of Bladder Cancer.

CXCL13+ T cells and LAMP3+ dendritic cells (DCs) are pivotal players in orchestrating anti-tumor immune responses, particularly within tumor tertiary lymphoid structures (TLS). However, their heterogeneity, differentiation trajectories, and clinical relevance in bladder cancer remain incompletely defined. This study integrated single-cell RNA sequencing (scRNA-seq) data (16 bladder cancer patients, 113,905 post-quality-control cells) and spatial transcriptomics to characterize CXCL13+ T cell/LAMP3+ DC subsets, their differentiation pathways (via Velocyto trajectory analysis), and intercellular crosstalk (via receptor-ligand mapping). A risk model (DTscore) was constructed using marker genes of these cells and validated in the IMvigor210 (atezolizumab-treated bladder cancer) and TCGA-BLCA cohorts. scRNA-seq clustering identified 10 immune and 3 nonimmune cell types, with T cells stratified into 8 subpopulations (including CD4+CXCL13+ T cells and CD8+CXCL13+ T cells) and DCs into 9 subgroups (including LAMP3+ DCs). Receptor-ligand mapping and spatial transcriptomics confirmed functional crosstalk between CXCL13+ T cells and LAMP3+ DCs via key pairs (e.g., CCR7-CCL19, CXCR5-CXCL13, PDCD1-CD274) within TLS. The DTscore was developed using 8 marker genes (TSHZ2, ALOX5AP, GADD45G, TXN, CHN1, CCL19, CXCL13, ICA1) and exhibited robust prognostic and predictive performance: In the IMvigor210 cohort, high DTscore correlated with significantly poorer overall survival (OS) and a 3.27-fold lower immunotherapy response rate (11% vs. 36%, p = 4.23e-07); multivariate Cox regression confirmed DTscore as an independent OS predictor (hazard ratio = 1.97, p < 0.001). DTscore retained prognostic value in TCGA-BLCA (OS: p = 0.003; disease-specific survival: p < 0.001) and effectively predicted atezolizumab response even in the "immune desert" phenotype (p = 0.04). Combining DTscore with tumor mutational burden/tumor neoantigen burden yielded an AUC of 0.8122 for response prediction. Additionally, high DTscore was associated with higher OS hazard ratios in patients with wild-type TTN, RB1, EP300, or FGFR3 (all p < 0.01), while FGFR3 mutations correlated with lower immune checkpoint/CXCL13 expression. This study delineates the heterogeneity and interactions of CXCL13+ T cell/LAMP3+ DC subsets in bladder cancer TLS and validates DTscore as a robust tool for predicting OS and immunotherapy response, offering a potential guide for personalized bladder cancer treatment.