Spatial reference and alternative proteome analysis of glioblastoma reveals molecular signatures and associates survival with specific markers

Abstract Molecular heterogeneity is a key feature of glioblastoma pathology impeding patient’s stratification and leading to high discrepancies between patients mean survivals. We performed a spatial proteomics analysis on a cohort of 96 glioblastoma patients with survival varying from few months to more than 4 years. 46 tumors were analyzed by spatially-resolved high resolution mass spectrometry proteomics. Integrative analysis of protein expression and clinical information allowed us to identify three molecular regions associated with immune, neurogenesis and tumorigenesis signatures. Several of these molecular signatures can be enriched within the same tumor sample leading to high intra-tumoral heterogeneity. Nevertheless, a set of proteins was found statistically significant based on patient’s survival times, 10 of which stem from alternative AltORF or non-coding RNA. From these proteins, 5 were selected as survival markers. Classification of patients based on the expression of these 5 proteins leads to a clear difference in survival. The expression of these 5 proteins was validated by immunofluorescence on an external cohort of 50 glioblastoma patients, with a similar correlation with their survival. Taken together, our work has enabled the characterization of new molecular regions within glioblastoma tissues based on protein expression which can help to guide glioblastoma prognosis and to improve the current glioblastoma classification.