开发一个代谢相关特征用于预测乳腺癌的预后、免疫浸润和免疫治疗反应

Breast cancer (BRCA) is the most commonly diagnosed cancer and among the top causes of cancer deaths globally. The abnormality of the metabolic process is an important characteristic that distinguishes cancer cells from normal cells. Currently, there are few metabolic molecular models to evaluate the prognosis and treatment response of BRCA patients. By analyzing RNA-seq data of BRCA samples from public databases via bioinformatic approaches, we developed a prognostic signature based on seven metabolic genes (PLA2G2D, GNPNAT1, QPRT, SHMT2, PAICS, NT5E and PLPP2). Low-risk patients showed better overall survival in all five cohorts (TCGA cohort, two external validation cohorts and two internal validation cohorts). There was a higher proportion of tumor-infiltrating CD8 + T cells, CD4 + memory resting T cells, gamma delta T cells and resting dendritic cells and a lower proportion of M0 and M2 macrophages in the low-risk group.

Low-risk patients also showed higher ESTIMATE scores, higher immune function scores, higher Immunophenoscores (IPS) and checkpoint expression, lower stemness scores, lower TIDE (Tumor Immune Dysfunction and Exclusion) scores and IC50 values for several chemotherapeutic agents, suggesting that low-risk patients could respond more favorably to immunotherapy and chemotherapy. Two real-world patient cohorts receiving anti-PD-1 therapy were applied for validating the predictive results.

Molecular subtypes identified based on these seven genes also showed different immune characteristics. Immunohistochemical data obtained from the human protein atlas database demonstrated the protein expression of signature genes. This research may contribute to the identification of metabolic targets for BRCA and the optimization of risk stratification and personalized treatment for BRCA patients.