Hypoxia characteristics and prognostic significance of glioblastoma MES-like subpopulation through multi-transcriptomics sequencing

To analyze the composition characteristics and biological functions of tumor cell subpopulations in glioblastoma (GBM) through multi-transcriptomics sequencing technology, and explore the hypoxia characteristics and spatial localization features of the mesenchymal-like (MES-like) tumor cell subpopulation in GBM and the influence on malignant biological behaviors.

Multi-transcriptomics sequencing data, including single-cell RNA sequencing (scRNA-seq) data (18 patients), bulk RNA sequencing (bulk RNA-seq) and spatial transcriptomics (ST) data of GBM, were employed to define cell subpopulations in GBM, and Gene Ontology (GO) and Gene Set Enrichment Analysis (GSEA) were utilized to analyze their functions. The proportions and locations of cell subpopulations in bulk RNA-seq data were evaluated with BayesPrism deconvolution. Immunofluorescence assay was conducted for verification on 12 paraffin samples of GBM from patients who visited the neurosurgical department of our hospital from 2015 to 2023 and met the pathological diagnostic criteria for GBM (10 males and 2 females, at an average age of 53.50 years and a median age of 54.50 years). pySCENIC was applied to predict specific transcription factors of tumor cell subpopulations.

Tumor cells in GBM were highly heterogeneous, and could be mainly divided into 4 subpopulations: astrocyte-like (AC-like), neural progenitor-like (NPC-like), oligodendrocyte progenitor-like (OPC-like) and MES-like. Differential gene analysis found that the MES-like tumor cells highly expressed vascular endothelial growth factor A (VEGFA), adrenomedullin (ADM), N-myc downstream regulated 1 (NDRG1), insulin like growth factor binding protein 5 (IGFBP5), and A-kinase anchoring protein 12 (AKAP12) (P<0.001). pySCENIC transcription factor prediction found that the high-active transcription factors of the MES-like tumor cells were AT-rich interaction domain 3A (ARID3A), FOS like 2, AP-1 transcription factor subunit (FOSL2), endothelial PAS domain protein 1 (EPAS1), CCAAT enhancer binding protein delta (CEBPD), and CCAAT enhancer binding protein beta (CEBPB) (P<0.05). GO and GSEA enrichment analyses found that the MES-like tumor cells were enriched in hypoxia-related pathways, especially the pathway of cell responses to hypoxia levels (NES=2.437, P<0.001). BayesPrism deconvolution showed that the MES-like tumor cells mainly existed in PAN (Pseudopalisading cells around necrosis) and perinecrotic zone. Immunofluorescence assay confirmed CD44⁺ (CD44 antigen) MES-like tumor cells were mainly located in hypoxia areas with highly expression of hypoxia inducible factor 1 subunit alpha (HIF1α) (P<0.01). Multivariate Cox regression analysis indicated that the MES-like tumor cells were significantly correlated with the adverse prognosis of GBM patients (HR=1.71, 95%CI: 1.38~2.11, P<0.001).

Tumor cells in GBM are of highly heterogeneity. They could be mainly divided into 4 subpopulations: AC-like, NPC-like, OPC-like and MES-like. MES-like tumor cells, mainly locating in PAN and perinecrotic zone, are characterized by hypoxia, which can promote the malignant progression of GBM.