Rational Design of Antimalarial Drugs Using Molecular Modeling and Statistical Analysis

Artemisinin is an antimalarial compound isolated from Artemisia annua L. that is effective against Plasmodium falciparum. This paper proposes the development of new antimalarial derivatives of artemisinin from a SAR study and statistical analysis by multiple linear regression (MLR). The HF/6-31G** method was used to determine the molecular properties of artemisinin and 10 derivatives with antimalarial action. MEP maps and molecular docking were used to study the interface between ligand and receptor (heme). The Pearson correlation was used to choose the most important properties interrelated to the antimalarial activity: Hydration Energy (HE), Energy of the Complex (Ecplex), bond length (FeO1), and maximum index of R/Electronegativity of Sanderson (RTe+). After the Pearson correlation, 72 MLR models were built between antimalarial activity and molecular properties; the statistical quality of the models was evaluated by means of correlation coefficient (r), squared correlation coefficient (r2), explained variance (adjusted R2), standard error of estimate (SEE), and variance ratio (F), and only four models showed predictive ability. The selected models were used to predict the antimalarial activity of ten new artemisinin derivatives (test set) with unknown activity, and only eight of these compounds were predicted to be more potent than artemisinin, and were therefore subjected to theoretical studies of pharmacokinetic and toxicological properties. The test set showed satisfactory results for six new artemisinin compounds which is a promising factor for future synthesis and biological assays.