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

DI IANNI MAURICIO; ENRIQUE ANDREA; PALESTRO PABLO; GAVERNET LUCIANA; TALEVI ALAN; BRUNO BLANCH LUIS E.

Rosario

Congreso; 4to congreso Argentino de bioinformatica y biología computacional; 2013

We have applied a 3-model ensemble of 2D classifiers capable of differentiating P-gp substrates from non substrates [2] jointly with a previously reported topological model capable of identifying anticonvulsants in the Maximal Electroshock Seizure (MES) test [3] to ZINC 5 and DrugBank databases [3,4] in order to find possible candidates for the treatment of P-gp mediated refractory epilepsy. The resulting candidates were docked into the 3D structure of human Pgp, obtained from an homology model based on the crystal structure of mouse P-gp (pdb code: 3G60). From the database compounds chosen as candidates, a subset of 380 molecules was selected for the docking stage. After the analysis, 12 structurally diverse compounds were acquired and tested in animal models of seizure. All 12 candidates showed some level of protection against MES test.