An endoplasmic reticulum stress related signature for clinically predicting prognosis of breast cancer patients.

Abstract

Background: Endoplasmic Reticulum Stress (ER stress) was an important event in the development of breast cancer. We aimed to predict prognosis based on ER stress related key genes.

Methods: Data of the RNA-seq and clinical information of breast cancer cases were downloaded from the TCGA database. A total of 4 genes related with ER stress was identified by the univariate Cox regression and Least Absolute Shrinkage and Selection Operator (LASSO)-penalized Cox proportional hazards regression analysis. The predictive ability of the ER stress model was evaluated by utilizing Kaplan-Meier curves and time-dependent receiver operating characteristic (ROC) curves. Moreover, we verified 4 genes expression and its relationship with clinical breast cancer cases in real-world.

Results: 4 genes including RNF186, BCAP31, SERPINA1, TAPBP were identified as a prognostic risk score model. Based on that, we found patients of breast cancer had a better survival with low-risk score. And also, ER stress model showed a good diagnostic efficacy with AUC curve. The risk score was significantly associated with patients' age, T stage and clinical stage. A nomogram was constructed to estimate individual survival. Further GO and KEGG analysis showed our model was related with immune infiltration. Patients of breast cancer with high-risk scores were usually accompanied with poor immune infiltration. It was predicted that high risk group was more sensitive to Vinorelbine, Docetaxel and Cisplatin. At last, we verified the expression of four signature genes using qRT-PCR and immunohistochemistry.

Conclusion: Our ER stress model performed a valuable prediction on breast cancer patients.