Short Survival-Related Genes Harbor EMT Processes and Dendritic Cell Infiltration in Glioblastoma

Abstract Background: Glioblastoma is an aggressive primary tumour with the lowest survival time among brain tumours. Tumour-infiltrating immune cells (TIICs) are involved in tumour progression and determine the prognosis, while the association of immune cell infiltration with glioblastoma is rarely unknown. This study aimed to screen survival-related (SR) genes and major biological processes through bioinformatic analysis and to identify the relationship between SR genes and TIICs.Methods:SR genes were screened by comparing the long-term (>36 months) and short-term (<12 months) survivors in the database GSE53733. Gene set enrichment analysis (GSEA) was applied to compare the differences in biological processes between long-term survivors and short-term survivors. The SR genes were identified using the limma package of R. Gene Ontology (GO) analysis was conducted through Metascape. The protein-protein interaction (PPI) network of the SR genes was established through the Search Tool for the Retrieval of Interacting Genes (STRING) website and further analysed by the Molecular Complex Detection (MCODE) algorithm. UALCAN and GlioVis were employed to analyse the expression levels and prognostic value of hub genes. The correlation of hub genes with immune cell filtration was estimated by the Tumor Immune Estimation Resource (TIMER). The gene-drug interaction network was constructed using the Comparative Toxicogenomics Database (CTD).Results: The functions of the detected genes were mainly enriched in epithelial mesenchymal transition (EMT) and oxidative phosphorylation. Of the detected genes, a total of 220 SR genes were identified, including 78 upregulated genes and 142 downregulated genes in long-term survivors. The upregulated genes were mainly related to neuron projection morphogenesis, extracellular matrix, and cation channel activity. The downregulated genes were mainly related to extracellular matrix organization and angiogenesis. The PPI network for SR genes was constructed with 65 edges and 195 nodes, and two significant modules were selected. The results indicated that COL1A2, COL6A2, COL8A1, and COL8A2 were hub SR genes. In addition, they were correlated with immune cell infiltration, especially dendritic cell infiltration.Conclusions: These results revealed that collagens accounted for the progression and prognosis of glioblastoma. In addition, DC infiltration is a risk factor for glioblastoma patients. The expression of collagen protein COL6A2 was significantly correlated with the DC infiltration level and poor prognosis. Further, potential drugs that affect the function of COL6A2 could improve the outcomes of glioblastoma.