Computationally Motivated Discovery of Biogenic Compounds with Potent CDK9 Inhibition and In Vitro Effects against Glioblastoma

Glioblastoma (GBM) is the most common malignant brain tumor in adults but with limited and poorly effective treatment protocols. Kinase inhibitors either as a mono- or as a combination therapy have considerable potential as novel treatments for GBM, with cyclin-dependent kinase 9 (CDK9) a promising target. Here, we present a multistep computationally motivated and structure-based approach to the discovery of CDK9 inhibitors. Active/decoy benchmarking calculations considered structural data from available CDK9-ligand crystal structures, including the potential of multiple protein structure docking. Using an optimized virtual screening approach that included 3D pharmacophore prefiltering of compounds and Glide docking calculations, six novel low micromolar inhibitors from the ZINC15 biogenic database were identified, validated as hits using in vitro CDK9/cyclin T1 binding assays. Of these compounds, 1 (3,5-disubstituted barbiturate-type core scaffold) demonstrated potent low micromolar effects (IC₅₀s ∼8-13 μM at 72 h) on the cell viability of three GBM cell-lines (U87-MG, T98G, and U251-MG) in a time- and concentration-dependent manner, with compound 7 (a pyrano[2,3-f]chromene-4,8-dione) more effective against patient-derived PD301 cells. The compounds had good predicted oral bioavailability and four of the six inhibitors potential for blood-brain barrier permeability. Given their structural novelty, the identified CDK9 inhibitor scaffolds can be further explored for their potential against CNS conditions, including 1 and 7 against GBM.