Abstract / Summary
Abstract Prostate cancer (PCa) is one of the most common malignancies in men and frequently develops bone metastasis, which is associated with high morbidity and mortality. Biomarkers are essential for the early diagnosis, treatment guidance, and prognostic evaluation of bone metastatic prostate cancer (BMPCa). Therefore, this study aimed to identify epithelial–mesenchymal transition (EMT)- and nucleotide metabolism (NM)-related genes with diagnostic potential in BMPCa and to characterize their associated immune microenvironment features. Transcriptomic profiles of primary PCa and BMPCa samples from GEO datasets were analyzed. Differentially expressed genes (DEGs) associated with EMT and nucleotide metabolism were screened, and multiple machine learning algorithms, including LASSO regression, Boruta, and random forest, were applied to identify diagnostic biomarkers. Subsequently, a biomarker-based nomogram was constructed, and immune infiltration analysis was performed to investigate the tumor immune microenvironment. Among seven candidate genes enriched in nucleotide metabolism, S100A8 and PLOD2 were identified as robust diagnostic biomarkers for BMPCa, with AUC values of 0.82 and 0.83 in the training cohort, respectively. The constructed nomogram exhibited excellent diagnostic performance, achieving an AUC of 0.89, and demonstrated potential clinical utility. Furthermore, S100A8 expression showed a strong positive correlation with myeloid-derived suppressor cell (MDSC) infiltration (r = 0.803). Collectively, S100A8 and PLOD2 represent promising biomarkers for evaluating the risk of bone metastasis in PCa. The association between S100A8 expression and MDSC infiltration provides new insights into the immune microenvironment of BMPCa and highlights potential targets for early detection and precision therapeutic intervention.