Abstract
Glioblastoma (GBM) is a highly aggressive and treatment-refractory malignant brain tumor with a markedly poor prognosis. Although various molecular subtypes have been identified to develop effective diagnostic and therapeutic strategies, their clinical application remains limited due to unclear classification criteria resulting from intra-tumoral heterogeneity and low clinical relevance. In this study, GBM cell line RNA-seq data from DepMap database and bulk RNA-seq data from The Cacner Genome Atlas, Chinese Glioma Genome Atlas, PRJNA1051047, and in-house cohorts were analyzed. Additionally, GeoMx DSP data, paired with in-house RNA-seq samples, were analyzed to investigate spatial transcriptome characteristics. GBM cell lines were divided into two new subtypes by NMF consensus clustering with Hallmark ssGSEA module score. Based on deferentially expressed gene sets IDH-wildtype GBM samples from four cohort were classified to two clusters named EMT and proliferative (PRO) subtypes by Bayesian compound covariate prediction. EMT subtype showed poor prognosis and high tumor related pathway scores such as EMT, angiogenesis, and hypoxia. Interestingly, TP53 showed strong association as upstream regulator in EMT subtypes similar to ssGSEA module scores of 10 TP53-related pathways from various curated database in four patient cohorts. In the GeoMx DSP analysis, TP53 expression and single-sample gene set enrichment analysis module scores were compared between α-SMA, CD45, and CD31 regions of interest (AOIs) divided into EMT and PRO subtypes. The expression of TP53 did not significant differ between EMT and PRO subtype in every type of AOI, however Hallmark TP53 signaling module score is lower in a-SMA AOIs of EMT subtypes around perivascular regions. In summary, the clinically relevant two subtypes were discovered, and especially EMT subtype was highly associated with poor prognosis and dysregulation of TP53 downstream pathways, suggesting the potential involvement of TP53 in GBM malignancy.