基于内质网应激相关基因特征的乳腺癌预后模型构建与功能解析OA

Objective To construct a prognostic risk model for breast cancer based on endoplasmic reticulum stress-related genes(ERSRGs)using The Cancer Genome Atlas(TCGA)and Gene Expression Omnibus(GEO)databases,and to systematically analyze its biological functions and clinical significance.Methods Transcriptomic and clinical data of 1 469 breast invasive carcinoma(BRCA)tissues and 113 adjacent normal tissues from TCGA database and the datasets,GSE20685 and GSE42568,from GEO database were inte-grated as the TCGA-GEO BRCA cohort.Differentially expressed ERSRGs were identified,and a prognostic model was constructed using LASSO-Cox regression.Model performance was evaluated by Kaplan-Meier survival analysis,receiver operating characteristic(ROC)curves,functional enrichment analysis,immune infiltration analysis,genomic variation analysis,and drug sensitivity analysis.Results A total of 272 ERSRGs were obtained from the Molecular Signatures Database(MSigDB)v7.0,of which 15 were differentially expressed in BRCA tissues from the TCGA-GEO BRCA cohort[|log2 fold change(FC)|>1.0,P<0.05].Eight key prognostic genes were screened by LASSO-Cox regression to construct the prognostic risk model,including cAMP responsive element binding protein 3 like 1(CREB3L1),stromal cell derived factor 2 like 1(SDF2L1),phosphoinositide-3-kinase regulatory subunit 1(PIK3R1),disabled homolog 2 interact-ing protein(DAB2IP),protein phosphatase 1 regulatory subunit 15A(PPP1R15A),phorbol-12-myristate-13-acetate-induced protein 1(PMAIP1),mitogen-activated protein kinase kinase kinase 5(MAP3K5),and inositol 1,4,5-trisphosphate receptor type 1(ITPR1).A risk score formula was established using the normalized mRNA expression values of these genes weighted by the regression coefficients from the LASSO-Cox analysis,and the BRCA patients in the TCGA-GEO BRCA cohort were divided into a high-risk group(n=734)and a low-risk group(n=735)using the median risk score.Survival analysis showed that patients in the high-risk group had shortened overall survival(OS)(5-year OS rates:76.8%vs 88.8%;10-year OS rates:59.0%vs 77.6%;log-rank P<0.01).ROC curve analysis showed that the areas under the curves(AUCs)of the risk model for predicting 1-,3-,and 5-year OS were 0.645,0.671,and 0.689,respectively.Functional analysis suggested that differentially expressed genes in the high-risk group were enriched in estrogen response,E2F targets,interleukin-17 sig-naling pathway,and inflammatory response.Immune infiltration analysis showed that both immune scores and stromal scores were elevated in the high-risk group(both P<0.01).Drug sensitivity analysis indicated that the high-risk group had higher half maximal inhibitory con-centration(IC50)values for 18 drugs,especially the cyclin-dependent kinase 4/6(CDK4/6)inhibitor palbociclib and the mammalian target of rapamycin(mTOR)inhibitors temsirolimus and AZD8055,compared to the low-risk group(all P<0.01).Conclusions This study suc-cessfully constructed a prognostic model for breast cancer based on ERSRGs.The model has good predictive performance and may provide a reference for risk stratification and individualized treatment of breast cancer patients.