Development and validation of an endoplasmic reticulum stress-related molecular prognostic model for breast cancer.

Abstract

Breast cancer is the most frequently diagnosed cancer and a leading cause of cancer-related death in women. Endoplasmic reticulum stress (ERS) plays a crucial role in the pathogenesis of several malignancies. However, the prognostic value of ERS-related genes in breast cancer has not been thoroughly investigated. We downloaded and analyzed expression profiling data for breast invasive carcinoma samples in The Cancer Genome Atlas-Breast Invasive Carcinoma (TCGA-BRCA) and identified 23 ERS-related genes differentially expressed between the normal breast tissue and primary breast tumor tissues. We constructed and validated risk models using external test datasets. We assessed the differences in sensitivity to common antitumor drugs between high- and low-scoring groups using the Genomics of Drug Sensitivity in Cancer (GDSC) database, evaluated the sensitivity of patients in high- and low-scoring groups to immunotherapy using the Tumor Immune Dysfunction and Exclusion (TIDE) algorithm, and assessed immune and stromal cell infiltration in the tumor microenvironment (TME) using the Estimation of Stromal and Immune cells in Malignant Tumor tissues using Expression data (ESTIMATE) algorithm. We also analyzed the expression of independent factors in the prognostic model using the Western-blot analysis for correlation in relation to breast cancer. Using multivariate Cox analysis, FBXO6, PMAIP1, ERP27, and CHAC1 were identified as independent prognostic factors in patients with breast cancer. The risk score in our model was defined as the endoplasmic reticulum score (ERScore). ERScore had high predictive power for overall survival in patients with breast cancer. The high-ERScore group exhibited a worse prognosis, lower drug sensitivity, and lower immunotherapy response and immune infiltration than did the low-ERScore group. Conclusions based on ERScore were consistent with Western-blot results. We constructed and validated for the first time an endoplasmic reticulum stress-related molecular prognostic model for breast cancer with reliable predictive properties and good sensitivity, as an important addition to the prognostic prediction model for breast cancer.