Abstract 6622: Dissecting <i>in situ s</i>tates of invading glioblastoma cells using single-cell and spatial transcriptomics

Abstract Introduction: The limited efficacy of surgical resection and high rate of tumor recurrence in glioblastoma underscores the urgency of better understanding tumor invasion in the central nervous system. High-throughput sequencing technologies, such as single-cell RNA-sequencing (scRNA-seq) and spatial genomics, efficiently identify gene expression states of rare invading cells and microenvironmental interactions. We previously described cell state heterogeneity in glioblastoma by scRNA-seq and uncovered the contribution of immune cells to the heterogeneity. Yet it is unclear whether the cell states observed in clinical specimens are implicated in invading glioblastoma cells in the brain parenchyma. We hypothesized that the distance from the primary tumor site and the local microenvironment of invading cells might affect the invasion phenotype. Method: To test this hypothesis, we developed an in vivo approach with mixtures of human glioblastoma spheroid lines for mouse intracranial xenografts. We validated this multiplex method with 20 patient-derived models and profiled, by scRNA-seq, in vivo models capable of infiltrating the brain parenchyma. We then used spatial transcriptomics to explore invasion routes and distribution of the invasive programs across the brain. Results: By comparing cells in the contralateral invasive site to the primary tumor, we identified a set of genes upregulated in invading glioblastoma cells, significantly enriched with NPC-like and OPC-like programs we identified previously, as well as genes not associated with the predefined states. Next, we analyzed the program expressed by cells undergoing invasion along blood vessels; the up-regulated program was enriched with AC-like and OPC-like programs. As observed with contralateral invasion, the program seen in blood vessel invasion was distinct from the previously defined states. The ST analysis highlighted this distinction and detailed the characteristics of each program. The contralateral invasion signature was associated with neuronal and myeloid-specific regions in the contralateral hemisphere. In contrast, the blood vessel invasion signature was localized to vascular regions at the periphery of the primary tumor. Conclusion: Our study found that an abundance of OPC-like cells was linked to more efficient invasion into the opposite side of the brain, suggesting that these cells might play a crucial role in distal invasion. In addition, we observed that models with a MES-like expression profile exhibit local perivascular invasion. Our comprehensive approach, coupled with high-throughput transcriptome profiling, provides a detailed blueprint of invasion potential and the functional transcriptional programs that drive GBM invasion along different anatomical routes, offering new insights for potential therapeutic strategies against glioblastoma. Citation Format: Toshiro Hara. Dissecting in situ states of invading glioblastoma cells using single-cell and spatial transcriptomics [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2025; Part 1 (Regular Abstracts); 2025 Apr 25-30; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2025;85(8_Suppl_1):Abstract nr 6622.