Mapping the transcriptional architecture of glioblastoma at the single-cell level: Decoding heterogeneity, angiogenesis, and mesenchymal shifts

Abstract

Glioblastoma (GBM), a grade IV glioma, is the most aggressive and fatal primary brain tumor, accounting for 48 % of all Central Nervous System tumors. Despite advancements in therapeutic strategies, GBM remains highly resistant to treatment, with a median survival time of just 14 months. This study aimed to identify molecular signature genes associated with GBM heterogeneity using scRNA-seq datasets from 10× Genomics. Two scRNA-seq datasets were processed through the Cell Ranger pipeline, followed by quality control, normalization, and scaling. After data integration using R, Principal Component Analysis was performed, and clusters were visualized using UMAP. A total of 2772 DEGs were identified, of which 95 DEGs met the threshold of logFC≥4 and p-adj ≤ 0.05. These DEGs were significantly enriched in angiogenesis and the PI3K signaling pathway, associated with poor prognosis. Principal Component Analysis revealed 15 principal components, with the first four accounting for the greatest variance. UMAP clustering identified 13 distinct cell clusters, which were annotated using the HPCA reference dataset, revealing enrichment in astrocytes, immune cells, and other tumor-associated cell types. A PPI network was constructed using the STRING database and visualized in Cytoscape, leading to the identification of three mesenchymal hub genes—KDA, PDGFRB, and CXCL12—as key angiogenic markers in GBM. The identified DEGs and hub genes were further validated using GEPIA2 and GSEA. This study provides novel insights into GBM heterogeneity and angiogenic biomarkers, potentially guiding future therapeutic strategies. Nevertheless, additional experimental validation is required to fully understand their role in GBM pathogenesis.