Identification of prognostic biomarkers of breast cancer based on the immune-related gene module

Abstract

Breast cancer (BC) is highly malignant and its mortality rate remains high. The development of immunotherapy has gradually improved the prognosis and survival rate of patients. Therefore, identifying molecular markers concerned with BC immunity is of great importance for the treatment of this disease. The Cancer Genome Atlas-breast invasive carcinoma (TCGA-BRCA) was utilized as the training set while the BC expression dataset from the gene expression omnibus database was taken as the validation set here. Weighted gene co-expression network analysis combined with Pearson analysis and Tumor immune estimation resource (TIMER) was used to obtain immune cell-related hub gene module. Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analyses were performed on this module. Then, receiver operating characteristic curves combining Kaplan–Meier was used to evaluate the effectiveness of the model. Feature genes were screened and the independence of risk score was evaluated by univariate and multivariate Cox analyses. Differences in immune characteristics were analyzed via single-sample gene set enrichment analysis and CIBERSORT, and differences in gene mutation frequency were assessed via GenVisR analysis. Finally, the expression levels of prognostic feature genes in BC cells were validated by quantitative reverse transcription polymerase chain reaction (qRT-PCR). In this study, cell immune-related gene modules in TCGA-BRCA were successfully excavated, and a five-gene (TNFRSF14, NFKBIA, DLG3, IRF2, and CYP27A1) prognostic model was established. The prognostic model could effectively forecast the prognosis and survival rate of BC patients. The result showed that human leukocyte antigen-related proteins and macrophage M2 scores were remarkably highly expressed in the high-risk group, whereas CD8+ T cells, natural killer cells, M1, and other anti-tumor cells were lowly expressed. The model could be used as an independent prognostic factor to predict the prognosis of BC patients. The results of qRT-PCR validation were consistent with the results in the database, that is, except DLG3, the other four feature genes were lowly expressed in BC. The five-gene model established in this study can predict the prognostic and immune mode of BC patients effectively, which is anticipated to become a feasible molecular target for BC therapy.

Keywords

• breast cancer
• weighted gene co-expression network analysis
• immune microenvironment
• hla
• gene mutation