Використання підходів in silico для раціонального дизайну потенційних антиконвульсантів з групи 5-заміщених 2-(R-аміно)-1,3,4-тіадіазолів

Abstract

Modern approaches to rational design of remedies, particularly, anticonvulsants as well, must include the use of computer analysis methods of correlations, in particular, (Q)SAR-analysis. Therefore, a series of promising anticonvulsants among 5-substituted 2-(R-amino)-1,3,4-thiadiazoles to detect the main descriptors, influencing activity, and obtained QSAR-models, which can be used in the targeted search of promising anticonvulsants, have been used by us. Aim. The aim of the given manuscript was to detect correlations and to form on their basis recommendations concerning the rational design of anticonvulsant agents among 5-substituted 2-(r-amino)-1,3,4-thiadiazoles. The regression analysis with the use of experimentally determined parameters of the severity of seizures, duration of seizures, latency period and the percentage of animals survived as dependent variables, and calculated 3D descriptors of compounds’ molecules as independent variables has been carried out for detection of the informative molecular descriptors, showing most adequately the features of molecules, responsible for the manifestation of the anticonvulsant activity. The calculation of the mathematical multiparameter QSAR-models has been carried out for this purpose. It has been shown, that anticonvulsant activity of the analyzed substances depends on the energy values of limit molecular orbitals and octanol-water partition coefficient logP, and besides, the severity of seizures decreases and latency period increases with both lipophility values and molecule’s volume increasing. The increase of polarizability has negative impact on the anticonvulsant activity of the synthesized compounds. The duration of seizures and animal survival time are reduced with decreasing of ЕНOМО value, i.e. with increasing of quantities of negative energy values of the lower vacant molecular orbital, which corresponds to strengthening of electron properties of compounds. Conclusion. A statistically significant number of QSAR-models intended for the pre-experimental prediction of the effective anticonvulsants with prescribed set of properties of 1,3,4-thiadiazole derivatives. The most important descriptors in "structure - activity" dependence, i.e. lipophility, energy values of the lower vacant molecular orbital and molecular volume, have been determined by regression analysis for the given group of substances.