Zukic, S.; Ivanova, L.; Zusinaite, E.; Merits, A.; Maran, U. Discovery of a Novel Phthalazine-Based VEGFR-2 Inhibitor by Machine Learning Analysis of Diverse Chemical Space with Favorable Pharmacophores and Its Evaluation in Glioblastoma Cell Lines. Results in Chemistry, Under revision

Abstract

Glioblastoma multiforme (GBM) is a form of aggressive brain tumor. One of the mechanisms regulating it is vascular endothelial growth factor (VEGF) and its receptor 2 (VEGFR-2). New active chemicals against glioblastoma can be discovered by analyzing diverse chemical space with favorable pharmacophore properties for VEGFR-2 inhibition. For this purpose, over two hundred VEGFR-2 inhibitors with different chemical structures were collected from the literature and curated. This representative chemical space was analyzed within quantitative structure-activity relationship (QSAR) approach using machine learning (ML). The results enabled analysis of the structure-activity relationship of compounds, creating the possibility to design, select and predict new potentially active VEGFR-2 inhibitors. To implement the results in practice, extended scaffolds corresponding to the application domain of the model were prepared, screened against the ZINC15 database and ranked. In experimental testing. one of the selected compounds showed high activity at sub-micromolar level (IC50= 0.497 ± 0.04 μM) in inhibiting VEGFR-2. The anticancer activity was evaluated against a neuroglioma (H4) cell line, two glioblastoma cell lines (U118MG, A172), a human medulloblastoma (Daoy) and a breast cancer (MCF‐7) cell line. The toxicity of the compounds was evaluated in HEK293 cells, which turned out to be low. The results of the cell line tests showed that selected compound was active against H4 (EC50 = 16.3 μM) and A172 (EC50 = 23.3 μM) cancer cell lines It can be concluded that phthalazine analogues bearing naphthalene moiety can be used to create new, potent, and effective VEGFR-2 inhibitors.