Polyamine metabolism related gene index prediction of prognosis and immunotherapy response in breast cancer.

BACKGROUND: Polyamine metabolism is closely associated with tumorigenesis, progression, and the tumor microenvironment (TME). This study aimed to determine whether polyamine metabolism-related genes (PMRGs) could predict prognosis and immunotherapy efficacy in Breast Cancer (BC). METHODS: We conducted a comprehensive multi-omics analysis of PMRG expression profiles in BC. Consensus cluster analysis was used to identify PMRG expression subtypes in the METABRIC cohort. Univariate and multivariate Cox regression analyses were performed to identify independent prognostic genes, which were subsequently used to construct a predictive model for BC, along with a novel nomogram based on PMRGs. The model was validated using an independent cohort (GSE86166). Independent prognostic genes were further verified in BC tissues using quantitative real-time PCR (qRT-PCR), Semi-quantitative Western blot, and immunohistochemistry. Additionally, we analyzed the immune microenvironment and enriched pathways across different subtypes using multiple algorithms. Finally, the "oncoPredict" R package was used to assess potential drug sensitivities in high-risk and low-risk groups. RESULTS: Seventeen polyamine metabolism genes were identified. PMRGs were abundantly expressed in tumor cells, with 12 survival-related genes being selected. In the METABRIC cohort, two PMRG expression subtypes were identified, with cancer- and immune-related pathways being more active in cluster B, which was associated with a worse prognosis. Six genes were used to construct a prognostic model through univariate and multivariate Cox regression analyses. The predictive performance of the polyamine metabolism model was validated by ROC curve analysis (training cohort: METABRIC, AUC3years=0.684; validation cohort: GSE86166, AUC3years=0.682). A nomogram combining risk scores and clinicopathological features was constructed. Decision Curve Analysis (DCA) demonstrated that the model could guide clinical treatment strategies. Four high-risk independent prognostic factors (OAZ1, SRM, SMOX, and SMS) were validated as being upregulated in breast cancer tissues. The model successfully stratified BC patients into high-risk and low-risk groups, with the high-risk group exhibiting poorer clinical outcomes. Functional analysis revealed significant differences in immune status and drug sensitivity between high-risk and low-risk groups. CONCLUSIONS: This study elucidated the biological characteristics of PMRG expression subtypes in BC, identifying a polyamine-related prognostic signature and four novel biomarkers to accurately predict prognosis and immunotherapy response in BC patients.