Heterogeneity Evaluation in Primary and Recurrent Glioblastoma through Single-cell RNA Sequencing and Spatial Transcriptomics

Abstract

Glioblastoma is the most aggressive primary brain tumor, characterized by extensive cellular heterogeneity, immune evasion, and therapeutic resistance. This study integrates single-cell RNA sequencing and spatial transcriptomics to investigate the cellular and functional differences between primary and recurrent glioblastoma.The analysis reveals the highly inter and intra heterogeneous nature of glioblastoma cells presented by 5 major cellular states: AC-like, NPC-like, MES-like, OPC-like, and proliferating state. These states remain stable during tumor recurrence, indicating that recurrent tumors are associated with functional plasticity—primarily exhibited by the upregulation of DNA repair and proliferation pathways—rather than complete cellular state transitions. The tumor microenvironment shows significant shifts, particularly in tumor-infiltrating lymphocytes. In recurrent glioblastoma, activated CD8+ effector memory T cells (TEM) are enriched, probably due to the leakage of the blood-brain barrier. In contrast, pre-exhausted CD8+ TEM with a tissue-resident phenotype is predominantly enriched in primary glioblastoma. A subset of terminally exhausted CD8+ T cells persists throughout tumor progression, representing the trajectory endpoint of cytotoxic CD8+ T cells. Myeloid cells, including tumor-associated macrophages (TAMs), play a pivotal role in glioblastoma progression by modulating immune responses and shaping the tumor microenvironment. In this study, TAMs were categorized into microglia-derived TAMs and monocyte-derived macrophages. One of the most notable findings was the enrichment of MDM-hypoxia in recurrent glioblastoma, exhibiting a distinct spatial localization within tumor regions. MDM-hypoxia is associated with hypoxia adaptation and angiogenesis, suggesting a role in tumor recurrence and therapy resistance. Their spatial organization within the tumor niche may facilitate interactions with tumor cells, reinforcing an immunosuppressive and pro-angiogenic TME. In contrast, MDM-SEPP1, associated with immune regulation and antigen processing, maintained a stable presence across tumor stages, suggesting a persistent immunosuppressive function. These findings highlight the dynamic interactions between glioblastoma cells, immune populations during disease progression. By integrating single-cell and spatial transcriptomics, this study provides insights into glioblastoma recurrence and identifies potential targets to disrupt hypoxia-driven macrophage functions, modulate T cell exhaustion, and reshape the tumor microenvironment for improved therapy.