Spatial and single-cell dissection of medulloblastoma identifies developmental hierarchies and a prognostic PPIA-BSG tumor-immune axis.

BACKGROUND: Medulloblastoma (MB) in children is highly heterogeneous, leading to its classification into molecular subgroups and resulting in significant differences in clinical outcomes even within the same risk group. This heterogeneity severely hampers the clinical management of MB. Therefore, it is essential to characterize MB heterogeneity within its tumor microenvironment. METHODS: Deep single-cell RNA sequencing (scRNA-seq) was performed on tumor tissues and matched adjacent normal tissues from two WNT, three Group 4 (G4), and one Sonic Hedgehog (SHH) MB patients. In addition, tumor tissue and adjacent cerebellar tissue from one patient in each subgroup were subjected to matched spatial transcriptomic sequencing. FINDINGS: A total of 316,446 single cells from MB and adjacent tissues were profiled by scRNA-seq. The cellular characteristics and developmental trajectories of granule cell precursors, unipolar brush cells, neural stem cells, microglia, and immune populations were defined, along with transcriptomic features associated with candidate cells of origin in MB. Infiltrating immune cells were present within tumors and exhibited an activated regulatory T-cell phenotype. Cell-cell communication analysis identified the PPIA-BSG co-receptor as a critical signaling axis in MB. Multiplex immunofluorescence and external bulk transcriptomic validation identified PPIA-BSG co-receptor as an independent prognostic factor. INTERPRETATION: Single-cell transcriptomic analysis delineates the cellular ecosystem of tumor and adjacent tissue-derived populations in MB, with particular emphasis on candidate cells of origin across molecular subgroups. These findings suggest that infiltrating immune cells may provide a foundation for immunotherapeutic strategies in MB and identify PPIA-BSG co-receptor as a potential therapeutic target.