Using the Method of Molecular Modeling and Docking to Estimate the Potential Danger of Side Effects of Therapeutic Agents Used with Cardiac Amyloidosis

Cardiac amyloidosis is a clinical pathology (usually of hereditary nature) initiated by the formation of amyloid fibrils, which has a high mortality rate. In this study, we used the methods of molecular modeling and docking of mobile ligands to estimate the level of side effects of ten low-molecular-weight therapeutic agents for the treatment of cardiac amyloidosis. The studied compounds were docked in the active sites of a representative set of 136 enzymes and receptor proteins obtained with a high resolution and deposited in the PDBbind database. It was demonstrated that Curcumin (~9.5%) and Diflunisal (~8.0%) have the highest level of side binding in the active sites of random proteins. On the contrary, Rivaroxaban and Usnic acid have a reduced level of side binding to random proteins (~2.2 and 0.7%, respectively) as compared with the average indices for this set of ligands. Other compounds (16,16-dimethyl-D-homoequilenin, 2-fluoro-17,17-dimethyl-D-homo-estrone, and Ursolic acid, as well as two experimental compounds with the molecular formulas C 20 H 26 O 2 and C 19 H 19 FO 2 , respectively) demonstrated a high level of selectivity with a probability of random binding to proteins ⪡0.7%. A method for increasing the selectivity of newly developed drugs was proposed. The approach developed in this work can be used for a rational design of new biologically active molecules and for the estimation of side effects of both already-known and widely used compounds in clinical practice and newly developed therapeutic agents.