Virtual screening applied to the discovery of new anticonvulsant drugs for the treatment of refractory epilepsy

PABLO H. PALESTRO; LUCIANA GAVERNET; GUILLERMINA L. ESTIÚ; ALAN TALEVI; MAURICIO E. DI IANNI; ANDREA V. ENRIQUE; LUIS E. BRUNO-BLANCH

Natal

Congreso; XXXVIII Congress of Theoretical Chemists of Latin Expression; 2012

Universidade Estadual Paulista

Epilepsy is a syndrome originated by different cerebral disorders of the central nervous system. Epileptic episodes are called seizures and have different manifestations, ranging from brief lapses of lack of attention to prolonged losses of consciousness with convulsive motor activity [1]. The antiepileptic drugs presently used provide adequate seizure control in a significant number of the patients. However, 25% of patients continue to have epileptic episodes despite optimal therapy. This condition is known as refractory epilepsy and might be associated with limited bioavailability of the antiepileptic drug in the brain, due to over-expression or activation of efflux-transporters. Human P-Glycoprotein (P-gp) belongs to a superfamily of membrane proteins involved in transport processes for drugs and xenobiotics. This protein is one of the efflux transporters linked to refractory epilepsy, being up-regulated at the blood-brain barrier and epileptic loci of patients with intractable epilepsy [2]. We present in this investigation the results obtained from a sequential virtual screening campaign designed to find new anticonvulsant structures with no affinity to P-gp. To this end, we applied a 3-model ensemble of 2D QSAR classifiers capable of differentiating Pgp-substrates from non-substrates [3] in conjunction with a topological model capable of identifying anticonvulsants [4] to the Zinc [5] and DrugBank [5] databases. After that, 380 compounds were selected for a docking-based virtual screening. As the 3D structure of human P-gp is not available, a homology model was constructed based on the X-ray structure of the mouse P-gp (PDB code=360). From the docking filtering, 360 compounds were classified as anticonvulsant and P-gp non-substrates. From them, ten compounds were selected for acquisition and subsequent pharmacological evaluation with very good results: 9 out of 10 showed anticonvulsant activity in animal models of seizure. The selection criteria included their commercial availability, molecular diversity and structural novelty. The experimental data shows the ability of the virtual models to find new anticonvulsant candidates that are non-substrates of P-gp.