Construction Of A Prognostic Risk Score Based On Immune-related Hypoxia And Epithelial-mesenchymal Transition Genes In Breast Cancer And Study On Its Clinical Value

Background and objective:Hypoxia,epithelial-mesenchymal transition,and tumor immune microenvironment are closely related and they play important roles in breast cancer progression and metastasis.The aim of this study was to construct a novel prognostic risk score model for breast cancer based on immune-related hypoxia and epithelial-mesenchymal transition genes.Methods:Gene expression datas of breast cancer(BRCA)were downloaded from TCGA and GEO databases as training and validation sets,respectively.Immune-related genes(IRGs),hypoxia-related genes(HRGs)and epithelial mesenchymal transition-related genes(EMTRGs)were downloaded from the Imm Port database,MSig DB and db EMT 2.0 database,respectively.The differentially expressed IRGs in the TCGA cohort were identified by R package.Univariate Cox regression analysis was performed to screen IRGs associated with overall survival(OS)and then performed GO and KEGG functional enrichment analysis.Consensus clustering analysis was employed to identify hypoxia-related subtypes and epithelial mesenchymal transition(EMT)-related subtypes in the TCGA-BRCA cohort according to the expression profiles of HGRs and EMTRGs associated with prognosis.PCA analysis and Kaplan-Meier curves were used to assess the efficacy of subtypes,and functional differences between subtypes were explored by GSEA and GSVA analysis.We identified differentially expressed genes between diffirent subtypes.Gene co-expression analysis was performed to screen for immune-related hypoxia and EMT genes by setting the pearson correlation coefficient threshold.LASSO regression analysis was used to construct a risk score model.Kaplan-Meier curves and ROC curves were used to evaluate the predictive performance of the model and validated with the GEO datasets.To investigate the potential mechanisms,functional enrichment analyses,such as KEGG,GO,and GSEA,were performed.We constructed a nomogram combined risk score and clinical characteristics using “rms”package.The ESTIMATE and ss GSEA algorithms were applied to assess the immune status of the two risk groups.Then,we analyzed the correlation of the risk score model with clinical characteristics and tumor mutational burden(TMB).Results:A total of 67 differentially expressed IRGs with prognostic value were identified in the TCGA cohort.17 HGRs and 107 EMTRGs associated with OS were obtained using univariate Cox analysis.We identified two hypoxia subtypes(A and B)and two EMT subtypes(C1 and C2),and observed significant survival differences and spatial distribution between A and B and between C1 and C2.We found that the subtype with a better prognosis in the hypoxic subtypes had lower hypoxia status,and the subtype with higher EMT status was associated with a poor prognosis.Differentially expressed genes among subtypes were screened,and then hypoxia differentially expressed genes and EMT differentially expressed genes associated with prognosis were obtained by univariate Cox analysis.Differential expression analysis between subtypes and univariate Cox analysis were performed to identify prognosis related differentially expressed hypoxia genes and prognosis related differentially expressed EMT genes.Co-expression analysis of these genes with 67 IRGs identified 46immune-associated hypoxia genes and 153 immune-associated EMT genes,and 15 intersecting genes were obtained after taking intersections.Co-expression analysis was used to identify 46 immune-related hypoxia genes and 153 immune-related EMT genes,and the two gene sets were intersected to obtain 15 intersecting genes.A prognostic risk score model consisting of 13 genes(STC2,SCUBE2,TPRG1,TFF1,FAM234 B,ARMT1,RARRES1,LAMP3,IDO1,FABP7,MMP1,CHI3L1,CXCL9)was constructed using LASSO regression analysis.Survival analysis showed that the high-risk group had a worse prognosis compared to the low-risk group.And model showed reliable predictive ability in both the training cohort(AUCmax=0.740)and the validation cohort(AUCmax=0.634).In addition,the risk score model was an independent prognostic factor for BRCA patients(p < 0.001).The nomogram integrating the risk score and clinical characteristics could accurately predict BRCA patients’ survival.Clinical correlation analysis showed that high risk score were significantly associated with later TNM stage,higher age group,higher tumor grade,and poorer clinical prognosis.The lower risk group with a better prognosis had higher immune status.The high-risk group had higher TMB compared with the low-risk group,and TMB was positively correlated with risk score.Conclusion:In this study,a novel prognostic risk score model was constructed based on immune,hypoxia and EMT genes,which could be used as an independent predictor for BRCA prognosis and was closely related to the tumor immune microenvironment.