Spatial architecture of high-grade glioma reveals tumor heterogeneity within distinct domains

High-grade gliomas are aggressive primary brain cancers with poor response to standard regimens, driven by immense heterogeneity. In isocitrate dehydrogenase (IDH) wild-type high-grade glioma (glioblastoma, GBM), increased intra-tumoral heterogeneity is associated with more aggressive disease. Recently, spatial technologies have emerged to dissect this complex heterogeneity within the tumor ecosystem by preserving cellular organization in situ. Here, we construct a high- resolution molecular landscape of GBM and IDH-mutant high-grade glioma patient samples to investigate the cellular subtypes and spatial communities that compose high-grade glioma using digital spatial profiling and spatial molecular imaging. This uncovered striking diversity of the tumor and immune microenvironment, that is embodied by the heterogeneity of the inferred copy- number alterations in the tumor. Reconstructing the tumor architecture revealed brain-intrinsic niches, composed of tumor cells reflecting brain cell types and microglia; and brain-extrinsic niches, populated by mesenchymal tumor cells and monocytes. We further reveal that cellular communication in these niches is underpinned by specific ligand-receptor pairs. This primary study reveals high levels of intra-tumoral heterogeneity in high-grade gliomas, associated with a diverse immune landscape within spatially localized regions. We selected three IDH1-wild type WHO grade 4 glioblastoma (GBM 1-3) and three IDH1-mutant grade 3 astrocytoma (A 1-3) samples for spatial whole transcriptome analysis. Notably, A-1 was re-diagnosed to oligodendroglioma. Regions of interest (ROI) were selected based on histological features, distribution of proliferating tumor cells and regions of tumor bordering on adjacent normal tissue across the six samples. By immunostaining for GFAP, CD45 and Ki67 protein expression, 34 areas of illumination (AOI) representing tumor cells (T, GFAP+), 10 AOIs representing proliferating tumor cells (K, GFAP+Ki67+) and 20 AOIs representing immune cells (I, CD45+) were isolated within each ROI and laser-captured for RNA purification and transcriptome analysis using Digital Spatial Profiling (DSP) on the NanoString GeoMx platform.