多组学方法揭示预后模型相关基因对髓母细胞瘤肿瘤微环境的影响

The tumor microenvironment (TME) significantly impacts the progression and prognosis of medulloblastoma (MB). This study aimed to develop a TME-associated risk score(TMErisk) model using RNA sequencing data to predict patient outcomes and elucidate biological mechanisms.

RNA sequencing data from 322 Tiantan and 763 GSE85217 MB samples were analyzed. Key gene modules related to immune and stromal components were identified using Weighted Gene Co-expression Network Analysis (WGCNA). Significant genes were screened using LASSO-COX and COX regression models. Single-cell RNA sequencing (scRNA-seq), single-cell ATAC sequencing (scATAC-seq), and spatial RNA analyses validated the findings.

Differential expression analysis identified 731 upregulated and 15 downregulated genes in high vs. low immune score MB patients, and 686 upregulated and 43 downregulated genes in high vs. low stromal score patients. Eight key genes ( CEBPB , OLFML2B , GGTA1 , GZMA , TCIM , OLFML3 , NAT1 , and CD1C ) were included in the TMErisk model, which demonstrated strong prognostic power. High TMErisk scores correlated with poorer survival, distinct immune cell infiltration patterns, and lower tumor cell stemness. Single-cell analyses revealed the expression dynamics of TMErisk genes across cell types, including macrophages, T cells, and NK cells, and identified key regulatory transcription factors. Spatial transcriptomics showed significant clustering of TMErisk genes in tumor regions, highlighting spatial heterogeneity and the formation of immune hubs.

The TMErisk model enhances our understanding of the MB tumor microenvironment, serving as a robust prognostic tool and suggesting new avenues for targeted therapy.