Study On Prognostic Biomarkers Of Glioblastoma Based On Fibroblasts

Objective: Glioblastoma is the most common malignant nervous system tumor with a high mortality rate and the incidence increases with age.In this study,the prognostic marker genes of glioblastoma were mined based on tumor-associated fibroblasts,and the prognostic model of glioblastoma was constructed and verified.Methods:In this study,we collected and collated the transcriptome data,mutation data(calculation of tumor mutation load),clinical relevant data(survival time,survival status,histological grade,gender,etc.)of glioblastoma in Cancer Genome Atlas(TCGA)and prognostic genes of glioblastoma in GEO database GSE72951 chip dataset.TCGA data and GSE72951 data were scored by tumor-associated fibroblast and divided into high and low scoring groups to compare the survival difference.Weighted gene co-expression network analysis was used to screen core genes,and gene ontology analysis and Kyoto encyclopedia of genes and genes enrichment analysis were performed to confirm the enrichment of core genes in function and expression pathway.Then,the core genes were screened by survival analysis,and the prognostic model formula was constructed according to the screening results.Each sample was scored and divided into high and low risk groups using the model to obtain survival curves.GSEA enrichment analysis was used to obtain the activation of functional pathways in high and low risk groups.The data were further processed by tumor mutation burden analysis and potential drug analysis.CCLE,HPA and CGGA databases were used to validate the prognostic marker genes.Results:In this study,TCGA information(clinical data,expression data,mutation data)and GEO information(GSE72951 chip expression data,clinical data)were sorted respectively,and CAF scores were obtained according to the expression information.CAF survival analysis showed that the prognosis of low CAF score group was better than that of high CAF score group except CAF_x Cell score.The CAF core genes were screened by WGCNA method from TCGA and GEO data,and 32 overlapping core genes were obtained.GO and KEGG enrichment analysis of core genes were performed to obtain core gene function and pathway enrichment.Survival analysis was performed based on core genes,and 3 prognostic genes were obtained,and a prognostic model was constructed.According to the model formula,the prognosis of low-risk group is better than that of high-risk group in both TCGA database and GEO database.GSEA enrichment analysis was performed on the risk data to obtain the function and pathway expression of high and low risk groups,and the five groups of pathways with the most significant enrichment were selected.Correlation analysis of CAF score showed that CAF score was positively regulated with patient risk score,and the three prognostic genes were all high risk genes.The three prognostic marker genes were verified by CCLE database and HPA database,and showed that the three prognostic genes were up-regulated in fibroblasts compared with central nervous system tumors and in glioblastoma compared with normal brain tissue.At the same time,immunohistochemical smears also showed significant differences.Three prognostic genes were significantly over-expressed in IDH wild-type GBM as verified by CGGA database.Conclusion:1.The content of tumor-associated fibroblasts and the expression of marker genes were significantly correlated with the prognosis of patients with glioblastoma.2.Tumor-associated fibroblast marker gene prognostic models showed satisfactory results.3.Poor immunotherapy assays,possibly related to the complex immunosuppressive microenvironment of glioblastoma4.The screening of marker gene-related functions and pathway mechanisms is helpful to the formulation of targeted treatment strategies5.The use of glioblastoma prognostic marker gene of tumor-associated fibroblasts provides more possibilities for the evaluation and treatment of glioblastoma patients,and injects new vitality into the improvement of prognosis.