Bioinformatic and Proteomic Analysis of Potential Prognostic Biomarkers of Glioblastoma

Abstract

Introduction . Glioblastoma is characterized by marked molecular heterogeneity and aggressive biologic behavior, which requires reliable molecular biomarkers for prognostic stratification. The aim of this study is to identify potential prognostic biomarkers for glioblastoma through integrated bioinformatic and proteomic analysis. Materials and methods . Transcriptomic and clinical data were obtained from the Gene Expression Database (GEO), The Cancer Genome Atlas (TCGA), the Chinese Glioma Genome Atlas (CGGA), and the Human Tissue Gene Expression Database (GTEx). Protein expression data were assessed using the Human Protein Atlas (HPA). Differentially expressed genes (DEGs) were identified in GSE147352. Next, protein-protein interaction (PPI) network analysis was performed using STRING and Cytoscape/cytoHubba. Key genes, or hub genes, were selected based on the Degree method. Their expression patterns were subsequently confirmed using independent datasets and HPA immunohistochemistry (IHC). Kaplan-Meier survival analysis, LASSO regression, and Cox proportional hazards regression were used to assess prognostic significance and construct a multigene risk model. Results . A total of 4,506 DEGs were identified in GSE147352, of which 200 genes with high and low expression were selected for PPI analysis. Ten key genes were identified (SNAP25, FN1, H3C12, SYN1, CAMK2A, SYP, SYT1, CD44, CCL2, and SLC17A7). FN1, H3C12, CD44, and CCL2 were consistently upregulated in glioblastoma, while SNAP25, SYN1, CAMK2A, SYP, SYT1, and SLC17A7 were downregulated in independent datasets. HPA data confirmed the differences at the protein level. Increased expression of FN1, H3C12, CD44, and CCL2 was associated with worse overall survival (OS), while increased expression of six downregulated genes was associated with better OS. This prognostic model, based on FN1, H3C12, CD44, and CCL2, significantly stratified patients into high- and low-risk groups (log-rank test p = 0.0361; hazard ratio = 1.52, 95 % confidence interval (CI): 1.05–2.20). The model demonstrated a C-index of 0.675 (95 % CI: 0.607–0.744; p < 0.001). Discussion. Integrated transcriptomic and proteomic analysis identified the FN1, H3C12, CD44, and CCL2 genes as promising prognostic biomarkers for glioblastoma. Their combined expression profile may provide a useful basis for molecular risk stratification and individualized prognostic assessment. Conclusion . The obtained results confirm the potential of the identified molecular markers for assessing prognosis and risk stratification in glioblastoma.