Abstract
Glioblastoma is an extremely aggressive form of brain tumor. The median overall survival of patients is only 15 to 18 months, and the five-year survival rate is less than 10% despite aggressive multimodal treatment that includes total resection of the tumor followed by a combination of radio-, chemo- and Tumor Treating Fields (TTF)-therapy. Therefore, new therapeutic interventions are urgently needed. For this reason, it is essential to define novel therapeutic targets for this tumor entity. To this aim, we developed a combinatorial approach, termed “ FLY dentification”, to identify and characterize novel driver genes in glioblastoma. This pipeline combines bioinformatics analysis of patients proteomics data together with functional screening in Drosophila models of glioblastoma, and validation with human glioblastoma cell models. This approach allowed us to initially uncover the top 300 most overexpressed proteins in glioblastoma tissues using in silico preselection algorithms. By taking advantage of a Drosophila model of glioblastoma ‒ driven by overexpression of the oncogenic forms of EGFR and PI3K ‒ and using an RNAi approach we tested the functional contribution of 192 of the gene candidates to glioblastoma lethality in vivo. This targeted screening allowed us to identify 41 clinically and functionally relevant gene candidates. Focusing on the top candidates with no reported function in glioblastoma, we demonstrated that the DNA binding and chromatin remodeling DEK protein contributes to human glioblastoma tumorigenesis in vitro. Some of these findings will be validated using in vivo models of human glioblastoma. Together, the established “FLY dentification” pipeline represents a combinatorial approach designed to uncover previously unrecognized glioblastoma driver genes, which has the power to reveal novel biomarkers and therapeutic targets for this tumor type.