Effect of intra- and inter-tumoral heterogeneity on molecular characteristics of primary IDH-wild type glioblastoma revealed by single-cell analysis.

AIMS To reveal the effects of intra- and inter-tumoral heterogeneity on characteristics of primary IDH-wild type glioblastoma cells. METHODS Single-cell RNA-seq data were acquired from the GEO database, and bulk sample transcriptome data were downloaded from the TCGA database with clinical information. Neoplastic subtype and glioma stem-like cells (GSCs) were identified by matching 5000 random virtual samples based on ssGSEA. CNV was inferred to compare the heterogeneity among patients and subtypes by infercnv. Transition direction was inferred by RNA velocity, and lineage trajectory was inferred by monocle. Regulon network of cells was analyzed by SCENIC, and cell communication was identified by CellPhoneDB. RESULTS Glioblastoma (GBM) cells could be divided into four subtypes by Verhaak classifier. However, classification of three subtypes (except NE subtype) was more suitable for GBM cells, and Verhaak classifier has difficulty in distinguishing GSCs. GBM heterogeneity and GBM cells' regulon network were mainly influenced by inter-tumoral heterogeneity. Within the same patient, different subclones exist in the same subtype of cells whose transition direction could be predicted by regulon similarity. Apart from inter-tumoral heterogeneity, different subtype of cells share common subtype-specific cell-cell communications. CONCLUSIONS Inter-tumoral heterogeneity contributes mainly to GBM heterogeneity and cell molecular characteristics. However, the same subtype of cells shared cell communication similarities.