Integrative Bioinformatics Analysis Unveils Neuro-cancer Crosstalk- related Genes and Establishes Prognostic Risk Model in Glioblastoma

Background Neuro-cancer crosstalk plays an important role in the development and progression of Glioblastoma (GBM), but its specific mechanisms remain incompletely elucidated. This study aims to systematically identify key genes related to neuro-cancer crosstalk in GBM and construct a prognostic risk model through integrating single-cell RNA sequencing (scRNA-seq), bulk RNA-seq, and machine learning algorithms. Methods The GSE273274 dataset was obtained from the GEO database for single-cell data analysis to investigate differences in intercellular communication. GBM transcriptomic data were obtained from TCGA and GTEx databases for differential expression analysis, and WGCNA was used to identify co-expressed gene modules. LASSO Cox regression was employed to screen out key prognostic genes and construct a prognostic risk model. Immune infiltration, drug sensitivity analysis and molecular docking validation were conducted. Finally, the expression of key genes was validated through immunohistochemistry experiments. Results Single-cell analysis identified 15 cell types and revealed significantly elevated proportions of CD44 + astrocytes and oligodendrocyte progenitor cells in the GC group. Intercellular communication analysis showed key COL6A2-GP6 and L1CAM-ERBB3 interactions between pericytes and mature excitatory neurons. Transcriptomic analysis identified 6,680 differentially expressed genes and 8,636 WGCNA hub genes. Integrated analysis identified 7 key neuro-cancer crosstalk genes, among which NGFR and L1CAM were further selected to construct the prognostic risk model. This model demonstrated good predictive performance in both training and validation sets. Immune infiltration analysis showed significantly elevated M0 macrophage proportions in the high-risk group. GSEA analysis revealed enrichment of axon guidance and RAS-ERK signaling pathways in the high-risk group. Drug sensitivity analysis identified betamethasone acetate as a potential therapeutic agent, and molecular docking showed good binding capacity with L1CAM and NGFR. Immunohistochemistry confirmed high NGFR expression and low L1CAM expression in GBM. Conclusions This study identified NGFR and L1CAM as key genes in GBM neuro-cancer crosstalk through multi-omics integrated analysis, and the constructed prognostic risk model demonstrates good clinical predictive value. The research findings provide new perspectives for understanding the mechanisms of neuro-cancer crosstalk in GBM and offer important theoretical foundations and potential targets for developing personalized treatment strategies.