Abstract:Abstract Glioma is the most common primary malignant brain tumor with limited treatment options and poor prognosis. To investigate the potential relationships between transcriptional characteristics and clinical phenotypes, we applied weighted gene co-expression network analysis (WGCNA) to construct a free-scale gene co-expression network yielding four modules in gliomas. Turquoise and yellow modules were positively correlated with the most malignant glioma subtype ( IDH -wildtype glioblastomas). Of them, genes in turquoise module were mainly involved in immune-related terms and were regulated by NFKB1 , RELA , SP1 , STAT1 and STAT3 . Meanwhile, genes in yellow module mainly participated in cell-cycle and division processes and were regulated by E2F1 , TP53 , E2F4 , YBX1 and E2F3 . Furthermore, 14 genes in turquoise module were screened as hub genes. Among them, five prognostic hub genes ( TNFRSF1B , LAIR1 , TYROBP , VAMP8 , and FCGR2A ) were selected to construct a prognostic risk score model via LASSO method. The risk score of this immune-related gene signature is associated with clinical features, malignant phenotype, and somatic alterations. Moreover, this signature showed an accurate prediction of prognosis across different clinical and pathological subgroups in three independent datasets including 1619 samples. Our results showed that the high-risk group was characterized by active immune-related activities while the low-risk group enriched in neurophysiological-related pathway. Importantly, the high-risk score of our immune signature predicts an enrichment of glioma-associated microglia/macrophages and less response to immune checkpoint blockade (ICB) therapy in gliomas. This study not only provides new insights into the molecular pathogenesis of glioma, but may also help optimize the immunotherapies for glioma patients.

Comprehensive transcriptomic characterization reveals core genes and module associated with immunological changes via 1619 samples of brain glioma