32 Identifying the Molecular Signature of Infiltrating Edge Cells in Glioblastoma as Drivers of Tumour Invasion and Recurrence

Abstract BACKGROUND Glioblastoma (GBM) is the most common malignant brain tumour in adults. Despite extensive research, there haven’t been remarkable gains in resolving the seeds of glioblastoma recurrence, and the outcomes for many patients suffering from this devastating disease remain poor. Complete tumour resection in GBM patients is not possible. Residual therapy-resistant cells drive tumour recurrence and infiltrative expansion. Our knowledge on GBM heterogeneity is mostly restricted to the surgically resectable tumour core, while the functional characterization of tumour cells at the infiltrating edge remains largely elusive due to the presence of normal functional brain tissue in the peritumoural lesion. Edge-derived cells exhibit larger capacity for infiltrative expansion and are the main drivers of treatment failure and tumour recurrence, making them action targets for novel treatment approaches. METHODS To resolve the transcriptional heterogeneity of GBM within the spatial context, we profiled gene expression of tumour regions selected based on histological features (“edge”, “core”, and “infiltrating zone”) obtained from 4 primary and 2 matched pairs of primary/recurrent IDH-WT GBM patients at single-cell resolution with Visium HD. RESULTS We show that infiltrative cells are spatially segregated and are characterized by regionally shared distinct transcriptomic signatures which define their cell state and identity. We complement non-spatial leiden clustering approach with BANKSY spatial clustering to augment the features of each cell with both an average of the features of its spatial neighbors along with neighborhood feature gradients. Using pathologically annotated H&E images integrated with spatial gene expression, we identify patterns related to tissue structure and identify transcriptional programs that promote invasiveness and underly disease recurrence in GBM. To further characterize invasive cells at the normal brain-tumour boundary, we identify top spatially variable genes in this cell population using unsupervised cell phenotyping. Upregulated DEGs of tumour edge cells are significantly associated with chemical synaptic transmission, and nervous system development. These modules represent tumour cell hijacking of neuronal programs as described in the context of glioma-neuron synaptic communication and formation of neurite-like tumour microtubes. Upregulation of ion regulation transport at the tumour periphery indicates enhanced neuronal activity and excitability driving infiltrating growth. SIGNIFICANCE Identifying biomarkers that are specific to malignant edge-derived cells may serve as new diagnostic feature that would help assess treatment response before or within early phases of therapy and allow for individual tailoring of the treatment plan to slow disease progression.