Identification of angiogenesis-related genes signature for predicting survival and its regulatory network in glioblastoma.

Glioblastoma (GBM) is notorious for malignant neovascularization that contributes to undesirable outcome. However, its mechanisms remain unclear. This study aimed to identify prognostic angiogenesis-related genes and the potential regulatory mechanisms in GBM. RNA-sequencing data of 173 GBM patients were obtained from the Cancer Genome Atlas (TCGA) database for screening differentially expressed genes (DEGs), differentially transcription factors (DETFs), and reverse phase protein array (RPPA) chips. Differentially expressed genes from angiogenesis-related gene set were extracted for univariate Cox regression analysis to identify prognostic differentially expressed angiogenesis-related genes (PDEARGs). A risk predicting model was constructed based on 9 PDEARGs, namely MARK1, ITGA5, NMD3, HEY1, COL6A1, DKK3, SERPINA5, NRP1, PLK2, ANXA1, SLIT2, and PDPN. Glioblastoma patients were stratified into high-risk and low-risk groups according to their risk scores. GSEA and GSVA were applied to explore the possible underlying GBM angiogenesis-related pathways. CIBERSORT was employed to identify immune infiltrates in GBM. The Pearson's correlation analysis was performed to evaluate the correlations among DETFs, PDEARGs, immune cells/functions, RPPA chips, and pathways. A regulatory network centered by three PDEARGs (ANXA1, COL6A1, and PDPN) was constructed to show the potential regulatory mechanisms. External cohort of 95 GBM patients by immunohistochemistry (IHC) assay demonstrated that ANXA1, COL6A1, and PDPN were significantly upregulated in tumor tissues of high-risk GBM patients. Single-cell RNA sequencing also validated malignant cells expressed high levels of the ANXA1, COL6A1, PDPN, and key DETF (WWTR1). Our PDEARG-based risk prediction model and regulatory network identified prognostic biomarkers and provided valuable insight into future studies on angiogenesis in GBM.