Identification and validation of a novel estrogen-related model for breast cancer to predict the prognosis.

[BACKGROUND] Breast cancer (BRCA) is a common malignant tumor in women globally and has a poor prognosis. Molecular targeted therapy is a promising way for improving the treatment of BRCA. This study aimed to identify potential biomarkers for BRCA and construct a prognostic model.

[METHODS] The expression, mutation and survival data were obtained from The Cancer Genome Atlas database, and estrogen-related genes (ERGs) were extracted from a previous study. Univariate, least absolute shrinkage and selection operator (LASSO) and multivariate Cox analyses were used to determine hub genes. The risk model was evaluated by Kaplan-Meier and receiver operating characteristic (ROC) curves. Immune infiltration was analyzed by the Immuno-Oncology Biological Research package. Gene set enrichment analysis was used for the functional analysis. validation was finally performed.

[RESULTS] Totally 113 estrogen-related differentially expressed genes (ERDEGs) were identified. A risk model was constructed using four hub ERDEGs: , , and . This model has a moderate predictive value, with area under curve (AUC) values over 0.69. In high-risk group, CD8 T cells, Tregs, NK cells and M1 macrophages were significantly decreased. , and were up-regulated in BRCA, while was down-regulated in BRCA. Survival analysis exhibited that high expression of and was associated with poorer prognosis, while high expression of and was associated with better prognosis. Additionally, overexpression significantly enhanced the invasion and migration of BRCA cells. It also inhibited the expression levels of p-STAT5, p-STAT3 and p-smad2/3 in BRCA cells.

[CONCLUSIONS] We identified four hub genes closely related to the prognosis of BRCA, and a risk model constructed by these four genes may be useful for risk stratification and prognosis evaluation in BRCA patients.