Molecular characterization of macrophage-related prognostic factors in glioblastoma revealed by combined analysis on single-cell and bulk transcriptome data.

BACKGROUND: Glioblastoma (GBM) is an aggressive primary tumor in the brain. The use of single-cell transcriptomic analysis can help to identify distinct cell subtypes and their functional states, and offer potential for advancing personalized therapeutic strategies for GBM. METHODS: Single-cell data were preprocessed using the Seurat and Harmony packages, and differentially expressed genes (DEGs) were identified via the FindAllMarkers function. The high-dimensional WGCNA was carried out on the myeloid cells of single-cell data to screen out the macrophage-related modules, followed by performing enrichment analysis on the module genes with clusterProfiler package. Through univariate Cox-LASSO regression with package survival, the module genes were further screened and compressed to determine the core genes and construct a prognostic model. Patients were stratified by the cutoff values of the risk scores into high- and low-risk groups. The IOBR package was used to evaluate the differences in immune infiltration. The expression differences of immune checkpoints were compared, and the drug sensitivity of GBM patients was tested by the R package oncoPredict. RESULTS: The proportion of Macro_PLIN2 subpopulation was significantly more in the tumor group, showing a higher activity in the blue and red modules. Eight core genes were further identified, namely SARNP, TGM2, G0S2, ACAP1, UPP1, POR, SLC43A3, and HPCAL1. Immune infiltration analysis revealed strong positive correlations between most core genes and the stromal score, immune score, and ESTIMATE score. The expression of PDCD1, PD-L1, CTLA4 and TIGIT in the high-risk group was significantly higher than those in the low-risk group. The drugs BI.2536, Daporinad, KU.55,933, and Ribociclib showed significant associations with the expression of the majority of core genes. CONCLUSION: This study reveals the molecular characteristics of key prognostic factors in GBM, highlighting the importance of immune cell abundance and drug sensitivity in glioma treatment, and provides potential biomarkers and therapeutic targets for future clinical research and treatment strategies.