Tryptophan metabolism-derived prognostic signature: A potential biomarker for risk stratification and exploratory immunotherapy relevance analysis in prostate cancer

Objective Prostate cancer (PCa) lacks reliable biomarkers to distinguish indolent from aggressive tumors. Tryptophan metabolism (TM) participates in regulating tumor immunity and malignant progression, yet its prognostic relevance in PCa remains poorly characterized. Methods Transcriptomic and clinical data of TCGA-PRAD cohort (499 PCa samples and 52 normal prostate tissues) and independent external cohort GSE116918 were collected. Differentially expressed tryptophan metabolism-related genes (TMGs) were screened. Univariate Cox regression and LASSO penalized regression were performed to construct a prognostic signature based on disease-free survival (DFS), which served as the primary modeling endpoint. Kaplan–Meier survival curves and time-dependent ROC analyses were applied to verify its prognostic performance. Hub genes were screened via PPI network, and the biological function of core hub gene CCN5 was further validated through in vitro cell experiments. Results A total of 82 differentially expressed TMGs were screened. Three candidate genes (MAOB, ELOVL2, THRSP) were integrated to establish the risk signature, which effectively stratified patients into low- and high-risk subgroups; high-risk patients exhibited significantly inferior DFS (P<0.05). The signature was identified as an independent prognostic predictor for PCa, with a 1-year DFS AUC of 0.997 (95% CI: 0.972–0.999). Bioinformatic prediction preliminarily indicated correlations between the risk model and M0 macrophage infiltration as well as chemotherapeutic sensitivity, while no real clinical chemotherapy response data were available to support these associations. Hub gene CCN5 was significantly downregulated in advanced PCa; cellular functional assays confirmed that CCN5 overexpression suppressed proliferation and migration of PCa cells. Conclusions This tryptophan metabolism-related signature independently predicts DFS outcomes of PCa patients and is preliminarily associated with tumor microenvironment features based on computational analysis. CCN5 acts as a tumor-suppressive gene in PCa. The present study provides exploratory bioinformatic evidence for future research on individualized prognostic stratification, rather than direct clinical medication guidance.