Classification Of Triple-negative Breast Cancers And Construction Of Prognostic Risk Model Based On Immune-related Genes

Objective: Based on the RNA sequencing data and clinical characteristic data of(Fudan University Shanghai Cancer Center,FUSCC)triple negative breast cancer(Triple-negative breast cancers,TNBC)samples in the database of FUSCC,the TNBC classification based on immune characteristics was explored by bioinformatics method,and the molecular feature and functional recognition among subtypes were analyzed.The recurrence risk prediction model of triple negative breast cancer was constructed by metastatic recurrence marker gene to predict the prognosis of new patients in clinic.Methods: 1.The RNA sequencing data and clinical characteristic data of triple negative breast cancer were downloaded from the database of Shanghai Cancer Center affiliated to Fudan University,and the samples that did not meet the group conditions were excluded.At the same time,immune-related genes were downloaded through Imm Port database.Univariate Cox regression was used to screen the immune genes associated with recurrence-free survival((Recurrence free survival,RFS).Based on the immune genes associated with recurrence,"Consensus Cluster Plus" was used to carry out consistent cluster analysis of TNBC samples,and a total of 1000 repeats were performed to ensure the stability of the classification.The validity of classification was evaluated by survival analysis and principal component analysis((Principal component analysis,PCA).The differences of the subgroups were analyzed by the "limma" package in R-studio,and the differential genes(differenti ally expressed genes,DEGs)with log | FC | > 0.5 and FDR < 0.05 were screened out.The gene ontology method was used to analyze the internal function of these differential genes and the genomic encyclopedia approach was used to analyze which pathways these differential genes were enriched.The proportion of 22 kinds of immune cell infiltration in triple negative breast cancer samples was inferred by CIBERSORT algorithm,and the correlation of 22 kinds of immune cell infiltration among different subtypes was detected by "corplot" package.2.Univariate Cox regression and multivariate correlation analysis were conducted based on the differentially expressed genes between the two subtypes to screen the differentially expressed genes that could independently predict the recurrence of TNBC.The optimal prediction model of recurrence risk was established by applying the coxph()function of Survival package of R language to these differential genes which had the role of independent prediction of recurrence risk.The recurrence risk prediction model was used to calculate the risk score of TNBC patients studied in FUSCC,the study samples were divided into high risk group and low risk group according to the median,and the recurrence risk prediction model was evaluated by survival analysis.Receiver operating characteristic curve(ROC)is plotted for further verification.Whether the recurrence risk prediction model established by univariate and multivariate analysis is an independent prognostic factor for patients with TNBC.The correlation between risk score and immune cell infiltration was analyzed by Pearson correlation test.Results: 1.The expression profile data and relevant clinical information of 352 patients with triple-negative breast cancer were obtained through the FUSCC database,and 2498 immune-related genes were obtained through the IMMPORT database,and 51 immune-related genes were screened out by unifactorial Cox regression analysis.Consistency cluster analysis divides TNBC into two stable subtypes.K-M analysis was used to evaluate the difference in survival of the two subtypes.Principal component analysis revealed specificity of RNA expression between the two subgroups.A total of 603 DEGs were screened in Cluster 1 and Cluster 2(211 up-regulated and 392 down-regulated genes,respectively).GO enrichment analysis showed that the most common biological functions of these differential genes were T cell activation,lymphocyte differentiation,regulation of T cell activation and positive regulation of leucocyte activation.KEGG analysis showed that the DEGs were mainly in the "cytokine" and "cytokine receptor" pathway.It also enriched the PD-L1/PD-1 checkpoint pathway.The TNBC samples were mainly monocytes,macrophages(type M2)and T cells(regulatory type).Among these infiltrating immune cells,the infiltration of activated memory T cells was positively correlated with the infiltration of M1 type macrophages.2.Based on the 603 differentially expressed genes obtained in the first part,univariate Cox regression analysis was used to analyze the correlation between them and RFS.The results showed that a total of 24 genes were significantly correlated with the time of disease recurrence in TNBC patients.A further 8 immune differential genes were obtained by multivariate correlation analysis as predictors of the model,including CNTD2,REM1,GREB1 L,ACHE,CD1 B,CD300LF,GP1 BA and MFNG.According to the risk coefficient and expression status of these eight genes,an immune-related risk prediction model was constructed,and the risk score,of the samples was divided into high and low risk groups according to the median of the study samples.The 1,3,5year ROC curves showed that AUC was 0.752,0.762 and 0.733,respectively.The feasibility of the established recurrence risk prediction model was evaluated.Univariate and multivariate Cox analysis confirmed that the recurrence risk prediction model was an independent prognostic factor for TNBC.In addition,activated dendritic cells(R = 0.29,p < 0.001),M0 macrophages(R = 0.21,p < 0.001),M2 macrophages(R = 0.27,p < 0.001),activated NK cells(R = 0.18,p < 0.001)was significantly correlated with the risk score of the recurrence risk prediction model.Conclusion: 1.Based on immune-related genes,TNBC samples in the FUSCC database were well clustered into two subtypes,and the two subtypes showed good differences in prognosis,survival,m RNA expression and immune cell infiltration in microenvironment.2.Eight differentially expressed genes,CNTD2,REM1,GREB1 L,ACh E,CD1 B,CD300LF,GP1 BA and MFNG,which can independently predict the recurrence of triple negative breast cancer metastases were screened,and the risk prediction model constructed by them can well predict the recurrence risk of TNBC patients.