Ks03.6.a the Recurrent Tissue States Across Adult Gliomas as Revealed by Spatial Transcriptomics

Abstract BACKGROUND Adult gliomas are a diverse class of tumors with substantial molecular and pathological variation. While prior single-cell omics studies have characterized intratumoral heterogeneity, they often examined tumors in isolation, thus overlooking the recurrence of tumor states across subtypes. Given their shared environment, understanding the conservation of tumor states across gliomas is essential. MATERIAL AND METHODS To address this, we analyzed 310 tumor cores from 284 patients spanning oligodendroglioma, low- and high-grade astrocytoma, glioblastoma, and ependymoma, using single-cell spatial transcriptomics with a 344-gene panel (10x Xenium). This approach yielded a dataset of 2.8 million cells, of which 84% were tumor and 16% represented 18 non-malignant cell types from the tumor microenvironment (TME). To further investigate local cellular interactions, we applied an extended spatial transcriptomics panel (5,096 genes) alongside multiplexed immunofluorescence (57 proteins) to a validation cohort of low-grade astrocytomas. RESULTS Our analysis revealed tumor states recurring across gliomas, albeit in varying proportions, organized into neighborhoods that reflect local cell-cell interactions. For instance, proliferating tumor cells tended to co-localize with oligodendro- and neuro-progenitor-like tumor states, suggesting that these programs may give rise to proliferating cells. Spatial neighborhoods were primarily subtype-specific, indicating that tissue organization and not cell composition alone, underlies glioma heterogeneity. We next assessed the relationship between tumor and immune cells, revealing entity-specific immune microenvironments: astrocytoma and ependymoma exhibited distinct immune responses (T cell- versus myeloid-driven inflammation), whereas oligodendroglioma displayed a low inflammatory landscape. Notably, microglia were consistently observed in proximity to tumor cells across all glioma subtypes, highlighting their conserved role in modulating anti-tumor immunity. Methylation profiling stratified these tumors into 13 molecular subtypes, each associated with distinct neighborhoods. Glioblastoma subtypes RTK1, RTK2, and MES, in particular, displayed unique tissue architectures consistent with their subtype-specific methylation profiles. CONCLUSION Collectively, these findings link microanatomical features to glioma heterogeneity by connecting tumor states to their microenvironments. Regional differences, such as those between the tumor core and infiltrating edge, shape cellular composition with implications to treatment response. By capturing such spatially dependent patterns, our work provides a framework for advancing the understanding of glioma biology.