Molecular modeling, docking and ADMET studies towards development of novel Disopyramide analogs for potential inhibition of human voltage gated sodium channel proteins

The sodium “channelopathies” are the first among the ion channel diseases identified and have attracted widespread clinical and scientific interests. Human voltage gated sodium channels are sites of action of several antiarrhythmic drugs, local anesthetics and related antiepileptic drugs. The present study aims to optimize the activity of Disopyramide, by modification in its structures which may improve the drug action by reducing its side effects. Herein, we have selected Human voltage-gated sodium channel protein type 5 as a potent molecular target. Nearly eighty analogs of Disopyramide are designed and optimized. Thirty are selected for energy minimization using Discovery studio and the LigPrep 2.5. Prior to docking, the active sites of all the proteins are identified. The processing, optimization and minimization of all the proteins is done in Protein preparation wizard. The docking study is performed using the GLIDE. Finally top five ranked lead molecules with better dock scores are identified as having strong binding affinity to 2KAV protein than Disopyramide based on XP G scores. These five leads are further docked with other similar voltage gated sodium channel proteins (PDB IDs: 2KBI, 4DCK, 2L53 and 4DJC) and the best scoring analog with each protein is identified. Drug likeliness and comparative bioactivity analysis for all the analogs is done using QikProp 3.4. Results have shown that the top five lead molecules would have the potential to act as better drugs as compared to Disopyramide and would be of interest as promising starting point for designing compounds against various Sodium channelopathies.

to predict potential inhibitors (drugs) for the treatment of several diseases [5]. Further computational prediction of pharmacokinetic parameters like Absorption, Distribution, Metabolism and Excretion (ADME) & toxicity studies have become increasingly important in drug selection and promotion process and are promising tools for early screening of potential drug candidates [6]. Despite numerous studies carried out on the pharmacologic therapy of various sodium channel opathies and other related diseases, the application of Computer Aided Drug Design (CADD) and Quantitative Structure-activity Relationship (QSAR) study on the design and development of new drugs for these diseases is scarce. The sodium channelopathies include a variety of inherited human disorders affecting heart rhythms, skeletal muscle contraction and nervous system [7].
The human voltage gated sodium channel proteins play a fundamental role in the propagation of action potential in electrically excited cells [8]. They are assumed to be the site of action for many types of drugs, such as local anesthetics, anticonvulsants, and anti-arrhythmic. These drugs elicit their effects by interfering with the rapid influx of sodium ions, which is responsible for the generation of action potential in excited cells.

Disopyramide and its mode of action
Systematic (IUPAC) name: 4-[bis (propan-2-yl) amino]-2phenyl-2-(pyridin-2-yl) butanamide. Formula: C21H29N3O. Disopyramide is a type I anti-arrhythmic drug that has been in wide use for several contractility. Disopyramide is a myocardial depressant which can depress contractility.Disopyramide does not reverse or attenuate cardiac hypertrophy in patients with Hypertrophic cardiomyopathy (HCM) but the beneficial effects of Disopyramide are largely attributable to its negative inotropic effects. It is a potential alternative drug regimen for treatment of obstructive HCM [9, 10]. It slows down conduction, increases refractory periods and decreases cardiac automaticity [11]. Therefore it is highly effective in HCM patients with significant left ventricular outflow tract (LVOT) obstruction. Data from previous studies show that approximately two third of the patients with LVOT gradient can be managed with Disopyramide in combination with betablockers [12]. It may have a superior effect on exercise tolerance compared with beta-blockers. However they are, best used in combination because Disopyramide alone tends to accelerate atrio-ventricular node (AV) node conduction and increase the potential risk from supra-ventricular arrhythmias.
Many cardiac anti-arrhythmic drugs and local anesthetics have the ability to block sodium channels in axonal and cardiac sarcolemmal membranes [13]. Disopyramide is similar in action to Quinidine without the adrenergic effects. It has the ability to block sodium channels in axonal and cardiac sarcolemmal membrane [14]. It targets sodium channels to lengthen their action potential. It depresses the increase in sodium permeability of the cardiac myocyte during Phase 0 of the cardiac action potential, in turn decreasing the inward sodium current.

Undesired effects of Disopyramide
The most significant side effects of Disopyramide are its anticholinergic effects. Studies have shown that long term therapy of this drug is associated with a low rate of serum enzyme elevations and is a rare cause of acute liver injury . Main side  effects are Ventricular tachycardia, ventricular fibrillation, QT  interval prolongation, myocardial depression, hypotension, AV  block; anti-muscarinic effects including dry mouth, blurred  vision,  urinary  retention,  gastrointestinal irritation. Contraindications of disopyramide include second and third degree heart block and sinus node dysfunction (unless pacemaker fitted) cardiogenic shock severe uncompensated heart failure [15,16]. Main drug interactions comprise increased risk of ventricular arrhythmias with amiodarone, antiarrhythmics, torsadogenic agents, diuretics (due to hypokalemia) increased plasma concentration with macrolides decreased plasma concentration with rifampicin; increased myocardial depression with anti-arrhythmics and beta-blockers.
In this study, we designed eighty possible structural analogues of Disopyramide by changing 'R' (= CH2CH3CH3) functional group as shown in the ( Figure 1) and thirty analogs were checked for their binding affinities with the target site. All the thirty analogues showed better affinity and interaction with the Human voltage-gated sodium channel (HVGSC) protein type 5 subunit alpha (PDB ID: 2KAV) but five of the 30 analogs showed highest affinity with the target 2KAV comparatively to Disopyramide drug. These results have been confirmed by the G Scores obtained from Grid -based Ligand Docking with Energetics (GLIDE) of Schrödinger 2011 [17,18]. These five inhibitors with highest dock scores are probable potent drugs which can be considered for further invitro studies. Further these top five analogs are docked with other HVGSC proteins with PDB IDs: 2KBI, 4DCK, 4DJC & 2L53 and also the predicted ADME and toxicity studies of these analogues were evaluated.

Methodology: Preparation of ligands
The chemical structure of Disopyramide molecule was retrieved from drug bank database into two-dimensional MDL/SDF format. Further 30 Disopyramide analogues were drawn in Osiris property explorer and checked for drug likeliness whether it followed Lipinski rule or not by modification of the 'R' position of Disopyramide drug (as shown in Figure 1). These analogs are further designed and geometrically optimized with the help of Hyperchem 8.0 software using the Bio+Charmm force field, RMS gradient of 0.01 kcal/(mol-Angstrom), maximum cycles upto 2000, in vacuum and Polak-Ribiere algorithm as parameters. The molecules were then loaded into Discovery Studio 2.5 for structure refinement such as energy minimization for 2000 steps with CHARMm force field.
The energetically minimized analogs Table 1 (see supplementary material) are obtained in .mol2 format from Discovery studios 2.5 and are used as input structures for processing in LigPrep 2.5 which is run from maestro9.2. The LigPrep 2.5 process consists of a series of steps that perform conversions, apply corrections to the structures, generate variations on the structures, eliminate unwanted structures and optimize the structures.
LigPrep protocol of Schrodinger 2011 generated all 30 as valid ligand conformations using Epik 2.0 in the pH range of 7±2. The LigPrep produces a single, low-energy, 3D structure with correct chiralities for each successfully processed input structure. It also produce a number of structures from each input structure with various ionization states, tautomers, stereochemistries, and ring conformations, and eliminate molecules using various criteria including molecular weight or specified numbers and types of functional groups present. The resulting structures are saved in either SD or Maestro format.

Protein Preparation
Prior to docking, it is important to identify the binding site in the target protein, information for which is available many times through the structures of the complexes of the protein with its substrate. Possible pockets in the set of sodium channel protein structures were first identified by detecting binding sites in all the proteins and then by identifying unique pockets that could serve as feasible targets for inhibitors.

Multiple Sequence alignment and Active site prediction
The binding pockets of proteins 2KBI, 4DCK, 2KAV, 2L53 and 4DJC were identified using the Computer Atlas of Topography of Proteins (CASTp), a program for identifying and characterizing protein active sites, binding sites and functional residues located on protein surface.
The protein sequences of human voltage gated sodium channel protein type 5 were obtained from PDB in FASTA format. To identity the conserved residues in all the 5 proteins, they were subjected to the Clustal W. The multiple sequence alignment of 2KBI chain A, 4DCK chain A, 2KAV chain A, 2L53 chain B and 4DJC chain B sequences showed five conserved residues as shown below in (Figure 2).
The PDB structure files of the proteins are imported. The processing, optimization and minimization of these proteins is carried out in the protein preparation wizard of the Schrodinger 2011 by applying the OPLS_2005 force field. The conserved residue Alanine is found to be the active site amino acid in all the five proteins and is specified in the receptor grid generation Table 2 (see supplementary material).

Docking
Docking refers to the optimal positioning of a ligand molecule with respect to the binding site of a target structure. GLIDE offers the full spectrum of speed and accuracy from highthroughput virtual screening of several compounds to extremely accurate binding mode predictions, providing consistently high enrichment at every level.
A site of 20x20x20 A° around the centric of pocket residues were exploited to check docking interaction with Disopyramide analogs. All the 30 analogs were subjected to docking in GLIDE of Maestro 9.2 with the protein with 2KAV. All 30 ligands docked with negative XP G Score and the compounds were ranked by the interaction energy. Overall, the Vander Waals energy contributed most to the interaction energy, but the electrostatic energy showed the greatest variation and was therefore the major factor for the ranking of molecules. Top five lead molecules were obtained after docking of 30 Disopyramide analogs with 2KAV which had comparatively higher binding efficiency than Disopyramide drug Table 3 (see supplementary  material).
The XP form of G Score is more physically accurate. The XP protocol includes ligand flexibility by docking multiple conformers in a rigid receptor, and the resulting complexes were ranked by XP G Scores.
These five top ranked compounds were selected for Glide XP docking with the other four similar proteins with PDB IDs: 2KBI, 4DCK, 2L53 & 4DJC. The docking results for these five leads are tabulated in the following Table 4 (see  supplementary material).

ADME & Toxicity studies
Most of drug candidates fail in clinical trials due to poor ADME properties. Thus, an important aspect of drug discovery is to avoid compounds not having drug likeliness and good ADME property. So to streamline the virtual screening, drug likeliness and ADME properties of all the thirty compounds were predicted using QikProp, version 3.4 of Schrodinger 2011 [19]. Lipinski filter and reactive filter were applied before virtual screening to avoid false positive lead molecule using OSIRIS Property explorer. Lipinski filter rejected ligands not following Lipinski rule of five and reactive filter rejected ligands with reactive functional groups. Drug likeliness, log P, log S, molecular weight and toxicity risks may be used to judge the compound's overall potential to qualify a ligand as potential drug candidate. All thirty ligands have appropriate logP (octanol/water) value for biological efficacy. Each of them had zero Lipinski violation and satisfying pharmacological properties of 95% available drugs with high to medium predicted oral absorption availability. Molecular weight of each ligand falls within the range of 297-404 Daltons. The ligands are having no toxic functional groups. Log S values of these ligands are within the acceptable range of 95% of existing drugs. The overall pharmacological properties (Table  VA & VB) of these ligands justify that the molecules are biologically active without any toxic functional groups. Hydrophobic compounds have relatively poor solubility, high log P, and high serum protein binding, but good cell permeability; whereas the opposite is true for hydrophilic compounds. This dichotomy was responsible for the classic lead optimization struggle of solubility versus permeability. Poor oral availability and permeability may lead to drug failure. The five lead molecules reported in the present study are well within the hydrophobic and hydrophilic extremes at the same time percentage of oral availability is also high. The five lead compounds have better pharmacological properties and can be considered as potential lead molecules for invitro drug discovery targeting HVGSC protein.

Conclusion:
A thorough study was carried out over thirty computationally designed Disopyramide analogs using various software programs with the goal of identifying potential lead molecules that bind to the human voltage-gated sodium channel protein(VGSC) type 5 relying on computational docking and pharmacological properties prediction with GLIDE of Schrodinger 2011 and QikProp 3.4 respectively. The comparatively higher XP G Scores of five analogs compared to Disopyramide when docked with 2KAV protein at the conserved active site residue suggest these novel leads would potentially bind more strongly to the pockets of VGSC proteins. Further, the five leads are docked with 2KBI, 4DCK, 2L53, and 4DJC proteins to predict their binding efficiencies with other similar sodium channel proteins apart from 2KAV. All proteins except 2KBI showed high XP G Scores for the analogs compared to Disopyramide. Also, these five novel lead molecules have better pharmacological properties compared to Disopyramide.
Thus, it is hoped that these five Disopyramide analogs identified in this study if synthesized and tested in animal models would hold promising results for new drug discovery.