Potassium channels are involved in a variety of cellular processes and play a primary role in maintaining ion homeostasis. They also play a critical role in the pathophysiology of cancer progression. The voltage-gated potassium channel hEAG1 is one such example involved in development of various cancers, as its increased expression occurs in most human cancers. Physiologically, it is found almost exclusively in the central nervous system and would therefore be an almost perfect target for cancer therapy. The problem is the hERG channel, which belongs to the same family of ion channels as hEAG1 and as an off-target can cause prolongation of the cardiac QT interval. The potential for potentially life-threatening arrhythmias presents a major challenge in the development of hEAG1 inhibitors. The primary goal of doctoral dissertation was to discover a new structural type of hEAG1 inhibitors. Using a computer-aided drug design approach, we created a pharmacophore model describing the binding of analogs of the natural compound purpurealidin I. These bind to the outside of the hEAG1 voltage sensor, which is responsible for transferring the change in membrane potential to the structural conformational change of the channel. A simplified pharmacophore model was used for virtual screening where 18 hit compounds were identified. Selected virtual hits were tested using the patch-clamp method and a compound with low micromolar inhibitory activity on hEAG1 was discovered. Hit compound from virtual screening with in vitro activity served as the basis for the synthesis of new analogs with which we aimed to improve inhibitory activity and selectivity. Selectivity was achieved for several sodium and potassium ion channels with the exception of the hERG channel. Although we were unable to improve selectivity at hERG, we were able to significantly increase activity at hEAG1. Thus, we found that the positively charged amino group is most important for the inhibitory effect, and in addition, it is necessary to retain the nitro and trifluoromethyl groups on the phenyl ring. Removal of the hydroxyl group transformed the hit compound into a nanomolar inhibitor, which increased the inhibitory activity fivefold compared to the virtual screening hit.Given the binding kinetics of the hit compound, which is characteristic of compounds that bind to the central cavity of the channel, we further investigated the potential binding site of the newly discovered compounds. We used both the in silico and in vitro approaches, with which we reached the common conclusion that the new compounds bind to the previously unevaluated site of the hEAG1 channel. The in vitro method was used to evaluate the potential displacement of compounds that bind to the same binding site. Thus, we checked whether the hit compound competes with compounds whose binding site is located in the central cavity or on the outside of the voltage sensor (VSD). We could not find any competition between astemizole that binds in central cavity or mibefradil that binds in VSD with the hit compound. Therefore, we decided to further verify the most common and best studied binding site in the central cavity using the in silico method. For this approach, we prepared a homology model of hEAG1 in the open conformation and docking of the known hEAG1 inhibitors into the binding site in the central cavity. The best-scored docking poses per ligand were used in molecular dynamics simulations. Based on the molecular dynamics simulations, we created a merged pharmacophore model describing the binding of the inhibitors to the central cavity. This model did not recognize our new hEAG1 inhibitors, confirming that they do not bind to the central cavity, as also observed in the in vitro asays, despite the characteristic positively charged center in the structure of the inhibitors. High similarity of the merged pharmacophore model with hERG pharmacophore model explains the non-selectivity of the hEAG1 inhibitors binding to the central cavity. The use of merged pharmacophore model will increase the chance of discovering new inhibitors that do not bind to the central cavity and thus have a higher probability of selective hEAG1 inhibition.
The antiproliferative activity of the new hEAG1 inhibitors on various cancer cell lines was tested using 2D and 3D cell models. The MCF-7 cell line with high hEAG1 expression was inhibited by the addition of micromolar concentrations of the new inhibitors and similar apoptotic activity was also observed on the spheroids of the Colo357 cell line. Due to the presence of the nitro group, which we could not replace with bioisosteric substitutions, the mutagenicity of the hit compound was also evaluated using AMES assay. It showed no mutagenic effects at non-cytotoxic concentrations, but further caution is advised in development of hEAG1 inhibitors of this structural class. The discovery of a new structural type of hEAG1 inhibitors in this doctoral dissertation represents an original contribution to the science in the field of anticancer drug discovery. In this work, we used an innovative ligand-based pharmacophore modeling approach to create 3D pharmacophore models in combination with molecular dynamics to discover and evaluate hEAG1 inhibitors. The obtained results represent a good foundation for further development of hEAG1 inhibitors with antiproliferative activity.
|
