Abstract: Objective  To make use of single-cell RNA sequencing (scRNA-seq) technology to conduct deep analysis on glioblastoma (GBM) tumor microenvironment, identify key driver genes that have specific expression in malignant tumor cells, and assess their potential as novel therapeutic targets and prognostic biomarkers. Methods  Tumor tissues and adjacent normal tissues from three clinical GBM patients were collected for scRNA-seq. After strict data quality control, integration, normalization, and batch effect correction, cells were clustered and annotated according to cell type. Differentially expressed genes that have specific upregulation in tumor cells were screened. Loss-of-function experiments (colony formation, Transwell assays, etc.) were carried out in vitro via the GBM cell line (LN229) treated with an inhibitor (NAZ2329), so as to verify the target gene's effects on cell proliferation, viability, invasion, and migration. Protein-protein interaction networks were analyzed by use of the STRING database, and pathway enrichment analysis was implemented through Gene Set Enrichment Analysis（GSEA）. Public databases such as the Cancer Genome Atlas（TCGA） and Gene Expression Profiling Interactive Analysis（ GEPIA） were used for expression profiling and survival analysis, and immunohistochemical staining  was employed to validate target protein expression in clinical tissue samples. Results  scRNA-seq analysis identified four major cell populations: T cells, macrophages, malignant tumor cells, and oligodendrocytes. Differential expression analysis revealed that the receptor-type tyrosine phosphatase receptor-type tyrosine-protein phosphatase zeta（PTPRZ1） was one of the most significantly and specifically upregulated genes in malignant tumor cells, which was validated in an independent single-cell dataset (GSE131928). TCGA data analysis indicated that high expression of PTPRZ1 was specific to glioma. In vitro functional experiments proved that inhibition of PTPRZ1 obviously suppressed the proliferation, viability, invasion, and migration capacities of GBM cells. Bioinformatics analysis indicated that PTPRZ1 has close association with its ligand PTN, thus it may promote tumor progression via activating pathways related to nervous system development. Clinical analysis confirmed that PTPRZ1 protein has high expression in GBM tissues, and it significantly correlated with patients' poor overall survival. Conclusion  PTPRZ1 is a key molecule that has specific high expression in malignant GBM cells, it obviously promotes malignant phenotypes of GBM cells, and its high expression is associated with a poor clinical outcomes. Hence, these findings demonstrate that PTPRZ1 is a potential diagnostic biomarker.

Key words: single-cell RNA sequencing, glioblastoma, PTPRZ1,  prognostic biomarker,  therapeutic target,  tumor invasion