Development and Verification of a Prostate Cancer Prognostic Signature Based on an Immunogenomic Landscape Analysis

PURPOSE: Prostate cancer (PCa) has a high incidence among older men. Until now, there are no immunological markers available to predict PCa patients’ survival. Therefore, it is necessary to explore the immunological characteristics of PCa. METHODS: First, we retrieved RNA-seq and clinical data of 499 PCa and 52 normal prostate tissue samples from the Cancer Genome Atlas (TCGA). We identified 193 differentially expressed immune-related genes (IRGs) between PCa and normal prostate tissues. Functional enrichment analyses showed that the immune system can participate in PCa initiation. Then, we constructed a correlation network between transcription factors (TFs) and IRGs. We performed univariate and multivariate Cox regression analyses and identified five key prognostic IRGs (S100A2, NOX1, IGHV7-81, AMH, and AGTR1). Finally, a predictive nomogram was established and verified by the C-index. RESULTS: We successfully constructed and validated an immune-related PCa prediction model. The signature could independently predict PCa patients’ survival. Results showed that high-immune-risk patients were correlated with advanced stage. We also validated the S100A2 expression in vitro using PCa and normal prostate tissues. We found that higher S100A2 expressions were related to lower biochemical recurrences. Additionally, higher AMH expressions were related to higher Gleason score, lymph node metastasis and positive rate, and tumor stages, and higher ATGR1 expressions were related to lower PSA value. CONCLUSION: Overall, we detected five IRGs (S100A2, NOX1, IGHV7-81, AMH, and AGTR1) that can be used as independent PCa prognostic factors.