Second generation of pyrimidin-quinolone hybrids obtained from virtual screening acting as sphingosine kinase 1 inhibitors and potential anticancer agents

Abstract

We report here the virtual screening design, synthesis, and activity of eight new inhibitors of SphK1. For this study, we used a pre-trained Graph Convolutional Network (GCN) combined with docking calculations. This exploratory analysis proposed nine compounds from which eight displayed significant inhibitory effects against sphingosine kinase 1(SphK1) demonstrating a high level of efficacy for this approach. Four of these compounds also displayed anticancer activity against different tumor cell lines, and three of them (5), (6) and (7) have shown a wide inhibitory action against many of the cancer cell lines tested, with GI 50 below 5 μM, being (5) the most promising with TGI below 10 μM for the half of cell lines. Our results suggest that the three most promising compounds reported here are the pyrimidine-quinolonine hybrids (1) and (6) linked by p-aminophenylsulfanyl and o-aminophenol fragments respectively, and (8) without such aryl linker. We also performed an exhaustive study of the molecular interactions that stabilize the different ligands at the binding site of SphK1. This molecular modelinganalysis was carried out by using combined techniques: docking calculations, MD simulations, and QTAIM analysis. In this study wealso included PF543, as a reference compound, in order to better understand the molecular behavior of these ligands at the binding site of SphK1.These results provide useful information for the design of new inhibitors of SphK1 possessing these structural scaffolds.