In the present study we investigated the phosphorylation status of the 12 most important signaling cascades in glioblastomas. More than 60 tumor and control biopsies from tumor center and periphery (based on neuronavigation) were subjected to selective protein expression analysis using reverse-phase protein arrays (RPPA) incubated with antibodies against posttranslationally modified cancer pathway proteins. The ratio between phosphorylated (or modified) and non-phosphorylated protein was assessed. All samples were histopathologically validated and proteomic profiles correlated with clinical and survival data. By RPPA, we identified three distinct activation patterns within glioblastoma defined by the ratios of pCREB1/CREB1, NOTCH-ICD/NOTCH1, and pGSK3β/GSK3β, respectively. These subclasses demonstrated distinct overall survival patterns in a cohort of patients from a single-institution and in an analysis of publicly available data. In particular, a high pGSK3β/GSK3β-ratio was associated with a poor survival. Wnt-activation/GSK3β-inhibition in U373 and U251 cell lines halted glioma cell proliferation and migration. Gene expression analysis was used as an internal quality control of baseline proteomic data. The protein expression and phosphorylation had a higher resolution, resulting in a better class-subdivision than mRNA based stratification data. Patients with different proteomic profiles from multiple biopsies showed a worse overall survival. The CREB1-, NOTCH1-, GSK3β-phosphorylation status correlated with glioma grades. RPPA represent a fast and reliable tool to supplement morphological diagnosis with pathway-specific information in individual tumors. These data can be exploited for molecular stratification and possible combinatorial treatment planning. Further, our results may optimize current glioma grading algorithms.
Keywords: Cancer signaling; Glioblastoma; Molecular stratification; Proteomics.