Abstract
Neuro-immune interactions are critical in cancer, yet their molecular features in bladder cancer remain unclear. We analyzed transcriptomic data from TCGA and UCSC Xena to investigate the expression profiles and molecular subtypes of neuro-immune-related genes, and constructed a neuro-immune-related score (NAS) model. Single-cell transcriptomic data were integrated to explore the immune microenvironmental features, and functional validation was performed by knocking down SERPINE2 with shRNA in T24 cells. The results showed that six core genes (SERPINE2, NXPH4, SERPINB2, C2orf40, SERPINB12, SERPINB10) were identified to stratify patients into high- and low-risk groups, with robust predictive power across clinical subgroups and validation cohorts. Single-cell RNA-seq data revealed significant NAS heterogeneity among cell populations. The NAS-high state was enriched in TGFβ, EGF, and FGF signaling with activation of EZH2 and SMARCA4, while the NAS-low state showed immune-regulatory features. Functional assays confirmed that SERPINE2 knockdown suppressed proliferation, migration, invasion, while increasing apoptosis of T24 cells, highlighting its oncogenic role. Moreover, genome-wide association studies (GWAS) suggested that genetic variants in SERPINE2 and related genes may increase bladder cancer susceptibility. Collectively, our findings provide novel insights into neuro-immune-driven tumor heterogeneity and immune remodeling, establish the NAS model as an innovative prognostic tool, and identify SERPINE2 as a promising therapeutic target for precision management of bladder cancer.