hERG Blockade Research Articles

Machine learning and deep learning approaches for enhanced prediction of hERG blockade: a comprehensive QSAR modeling study

ABSTRACT Background Cardiotoxicity is a major cause of drug withdrawal. The hERG channel, regulating ion flow, is pivotal for heart and nervous system function. Its blockade is a concern in drug development. Predicting hERG blockade is essential for identifying cardiac safety issues. Various QSAR models exist, but their performance varies. Ongoing improvements show promise, necessitating continued efforts to enhance accuracy using emerging deep learning algorithms in predicting potential hERG blockade. Study design and method Using a large training dataset, six individual QSAR models were developed. Additionally, three ensemble models were constructed. All models were evaluated using 10-fold cross-validations and two external datasets. Results The 10-fold cross-validations resulted in Mathews correlation coefficient (MCC) values from 0.682 to 0.730, surpassing the best-reported model on the same dataset (0.689). External validations yielded MCC values from 0.520 to 0.715 for the first dataset, exceeding those of previously reported models (0–0.599). For the second dataset, MCC values fell between 0.025 and 0.215, aligning with those of reported models (0.112–0.220). Conclusions The developed models can assist the pharmaceutical industry and regulatory agencies in predicting hERG blockage activity, thereby enhancing safety assessments and reducing the risk of adverse cardiac events associated with new drug candidates.

Virtual screening of DrugBank database for hERG blockers using topological Laplacian-assisted AI models

The human ether-a-go-go (hERG) potassium channel (Kv11.1) plays a critical role in mediating cardiac action potential. The blockade of this ion channel can potentially lead fatal disorder and/or long QT syndrome. Many drugs have been withdrawn because of their serious hERG-cardiotoxicity. It is crucial to assess the hERG blockade activity in the early stage of drug discovery. We are particularly interested in the hERG-cardiotoxicity of compounds collected in the DrugBank database considering that many DrugBank compounds have been approved for therapeutic treatments or have high potential to become drugs. Machine learning-based in silico tools offer a rapid and economical platform to virtually screen DrugBank compounds. We design accurate and robust classifiers for blockers/non-blockers and then build regressors to quantitatively analyze the binding potency of the DrugBank compounds on the hERG channel. Molecular sequences are embedded with two natural language processing (NLP) methods, namely, autoencoder and transformer. Complementary three-dimensional (3D) molecular structures are embedded with two advanced mathematical approaches, i.e., topological Laplacians and algebraic graphs. With our state-of-the-art tools, we reveal that 227 out of the 8641 DrugBank compounds are potential hERG blockers, suggesting serious drug safety problems. Our predictions provide guidance for the further experimental interrogation of DrugBank compounds’ hERG-cardiotoxicity.

Design, synthesis, extra-precision docking, and molecular dynamics simulation studies of pyrrolidin-2-one derivatives as potential acetylcholinesterase inhibitors

Inhibition of acetylcholinesterase (AChE) has been widely explored to develop novel molecules for management of Alzheimer's disease. In past research finding reported molecule 3-(4-(4-fluorobenzoyl)piperidin-1-yl)-1-(4-methoxybenzyl)pyrrolidin-2-one displayed a spectrum of anti-Alzheimer’s properties herein, we report a library of 18 novel molecules that were rationally designed and synthesized employing known literature to mimic and explore the novel chemical space around the lead compound 6e and donepezil. All the compounds were docked in extra-precision mode with AChE (PDB ID 4EY7) using the Glide module. Molecular dynamics (MD) simulation studies were carried out for 100 ns along with MM-PBSA studies of the trajectory frames generated post-MD simulations. Docking and MD simulation studies suggested that the synthesized compounds showed a good binding affinity with AChE. and might form stable complexes. 3-(4-(benzyl(methyl)amino)piperidin-1-yl)-1-(3,4-dimethoxybenzyl)pyrrolidin-2-one (14a; docking score: −18.59) and 1-(3,4-dimethoxybenzyl)-3-(4-(methyl(thiazol-2-ylmethyl)amino)piperidin-1-yl)pyrrolidin-2-one (14d; docking score: −18.057) showed higher docking score than donepezil (docking score: −17.257) while most of the compounds had docking score >-10.0. ADMET study predicted these compounds to be CNS active and most of the compounds were drug-like molecules with no HERG blockade and good to excellent oral absorption. We developed an atom-based 3 D-QSAR model with R^2 and Q^2 values of 0.9639 and 0.8779 to predict the activity of the synthesized compounds. The model predicted these compounds to be potent AChE inhibitors with IC50 values in the lower micromolar range. Based on the in silico findings, we report these newly synthesized compounds 3-(4-(benzyl(methyl)amino)piperidin-1-yl)-1-(3,4-dimethoxybenzyl)pyrrolidin-2-one (14a) and 7-(2,6-difluorobenzyl)-2-(4-methoxybenzyl)-2,7-diazaspiro[4.5]decan-1-one (20 b) as potential AChE inhibitors. Communicated by Ramaswamy H. Sarma

HERG Blockade Prediction by Combining Site Identification by Ligand Competitive Saturation and Physicochemical Properties.

Human ether-a-go-go-related gene (hERG) potassium channel is well-known contributor to drug-induced cardiotoxicity and therefore an extremely important target when performing safety assessments of drug candidates. Ligand-based approaches in connection with quantitative structure active relationships (QSAR) analyses have been developed to predict hERG toxicity. Availability of the recent published cryogenic electron microscopy (cryo-EM) structure for the hERG channel opened the prospect for using structure-based simulation and docking approaches for hERG drug liability predictions. In recent time, the idea of combining structure- and ligand-based approaches for modeling hERG drug liability has gained momentum offering improvements in predictability when compared to ligand-based QSAR practices alone. The present article demonstrates uniting the structure-based SILCS (site-identification by ligand competitive saturation) approach in conjunction with physicochemical properties to develop predictive models for hERG blockade. This combination leads to improved model predictability based on Pearson's R and percent correct (represents rank-ordering of ligands) metric for different validation sets of hERG blockers involving diverse chemical scaffold and wide range of pIC50 values. The inclusion of the SILCS structure-based approach allows determination of the hERG region to which compounds bind and the contribution of different chemical moieties in the compounds to blockade, thereby facilitating the rational ligand design to minimize hERG liability.

Predicting arrhythmogenicity: structural modeling of safe and unsafe hERG blockers using Rosetta

Human Ether-a-go-go-Related Gene (hERG) encodes a potassium-selective voltage-gated ion channel essential for normal electrical activity in the heart. Genetic hERG mutations and blockage of the channel pore by drugs can cause long QT syndrome (LQTS), which may predispose individuals to arrhythmia and puts them at risk for stroke or sudden cardiac arrest. A major problem in drug development is the proclivity for drugs from multiple classes to promote fatal arrhythmias through hERG blockade. However, not all hERG blocking drugs are pro-arrhythmic, and their differential affinities to discrete channel states and/or their state stability modulations have been suggested to contribute to arrhythmogenicity.

Abstract 12269: Investigating Arrhythmogenic Potential of a Novel KCNH2 Mutation Leading to Long QT Syndrome

Background: Mutations in the gene KCNH2 have been associated with both Short and Long QT syndrome. In this study, we identified a 1-day old infant that exhibited T-wave alternans coupled with Long QT Syndrome leading to aborted sudden infant death. Methods: Genetic analysis of the patient revealed a mutation in KCNH2 resulting in a proline to leucine substitution at position 632 (P632L). Functional electrophysiological studies were performed with the hERG mutation transiently transfected into either Human Embryonic Kidney (HEK) cells or human induced pluripotent stem cell-derived cardiomyocytes (hiPSC-CMs). The functional effects of P632L were incorporated into biophysical computational models of hiPSC-CM as well as adult human cardiomyocytes to further assess its arrhythmogenic potential. Results: The male infant displayed multiple repolarization disorders including T-wave alternans and a QTc =510 ms that appeared 1 day after birth. Genetic screening revealed a heterozygous missense mutation (P632L) in KCNH2 in the infant. Patch clamp analysis of HEK cells transiently transfected with P632L mutation showed a complete loss of function of HERG current compared to WT HERG. Transient transfection of P632L into WT hiPSC myocytes resulted in prolongation of the hiPSC action potential compared to untransfected hiPSC myocytes. Co-transfection of WT and non-functional P632L channels into HEK cells resulted in a dramatic loss of current suggesting a dominant negative effect. Implementing the effects of a complete block of HERG current in the hiPSC-CM computer model prolonged the action potential by 45% and depolarized the resting membrane potentials. In adult human myocyte models, the effects of HERG blockade were more severe in Purkinje cells than that in ventricular myocytes (71% vs. 16% APD prolongation, respectively). Conclusions: The mutation P632L causes a complete loss of HERG current and results in QT prolongation. Numerical simulations suggest a more severe phenotype in Purkinje cells than in ventricular myocytes which could be proarrhythmic.

Identification of hERG blockade by machine learning

Identification of hERG blockade by machine learning

The Effect of a Synthetic Estrogen, Ethinylestradiol, on the hERG Block by E-4031.

Inhibition of K+-conductance through the human ether-a-go-go related gene (hERG) channel leads to QT prolongation and is associated with cardiac arrhythmias. We previously reported that physiological concentrations of some estrogens partially suppress the hERG channel currents by interacting with the S6 residue F656 and increase the sensitivity of hERG blockade by E-4031. Although these studies suggested that clinically used synthetic estrogens with similar structures have the marked potential to alter hERG functions, the hERG interactions with synthetic estrogens have not been assessed. We therefore examined whether ethinylestradiol (EE2), a synthetic estrogen used in oral contraceptives, affects hERG function and blockade by drugs. Supratherapeutic concentrations of EE2 did not alter amplitudes or kinetics of the hERG currents elicited by train pulses at 20 mV (0.1 Hz). On the other hand, EE2 at therapeutic concentrations reduced the degree of hERG current suppression by E-4031. The administration of EE2 followed by E-4031 blockade reversed the current suppression, suggesting that the interaction of EE2 and E-4031 alters hERG at the drug-binding site. The effects of EE2 on hERG blockade raised the possibility that other estrogens, including synthetic estrogens, can alter hERG blockade by drugs that cause QT prolongation and ventricular arrhythmias.

The Acid/Base Characterization of Molecules with Epigenetic Activity.

The acidic and basic functional groups in a molecule strongly influence its physicochemical properties, affinity for a macromolecule, pharmacokinetics, and toxicity. For instance, basicity has been correlated with molecular promiscuity, hERG blockade, and phospholipidosis. Nonetheless, no systematic characterization of the acid/base profile of epigenomic databases has been reported. This study describes an analysis of the acidic ionization constant distribution of a library of 7820 compounds with reported activity against epigenetic targets. Furthermore, the epigenomics database's acid/base profile was compared to the reference libraries of food chemicals, natural products, and approved drugs. It was found that the acid/base profile of histone lysine demethylase ligands is more similar to previously approved drugs, and histone acetyltransferase ligands have acidic and basic functional groups largely found in food chemicals and natural products; this support the potential of these libraries for finding new epigenetic inhibitors.

Effects of Estrogens on the Actions of hERG

Regulation of cardiac ion channels by sex hormones accounts for gender-differences in susceptibility of Torsades de Pointes (TdP) arrhythmias associated with QT prolongation, that is: women are more prone to develop drug-induced arrhythmias which can lead cardiac sudden death. We previously found that β-estradiol (E2) directly interacts with a drug-binding site, F656 of the hERG channel, and alters the effect of a selective hERG blocker. Although these actions were observed for other estrogens, the effect of estrogen on other drugs has been unknown. In this study, to reveal impacts of the actions of estrogens on hERG currents, we compared the hERG screen data in the absence and the presence of estrogens using manual and automated patch-clamp experiments. HEK293 cells stably-expressing the hERG channels were seeded into 384-well plates, and were cultured in a steroid-free condition. Whole-cell hERG currents were recorded by the patch-clamp technique with SyncroPatch 384i (Nanion). Effects of 22 compounds were assessed by percentage changes of hERG peak tail currents with or without estrogens (E2, estrone sulfate; E1S, and ethynylestradiol; EE2) and by molecular docking simulations. Our hERG screening depicted that some estrogens showed distinct effects on respective blockers. Estrogens did not change the degree of hERG block by most of compounds. However, we have newly found that inhibitory action of one drug was increased by E2, and that of another drug was decreased by EE2. This means that specific combinations of an estrogen and a hERG blocker can alter the hERG blockade significantly. Indeed, some combinations resulted in larger hERG currents. This study suggests potential risks using estrogen therapy/medication and the necessity of taking sex hormonal differences into consideration for more accurate prediction of QT prolongation proclivity of drug candidates.

Structural Modeling of the hERG Channel in an Inactivated State and Associated Drug Interactions

The voltage gated potassium channel, KV11.1, encoded by the human ether-a-go-go related gene (hERG) is expressed in cardiac myocytes, where it is crucial for the repolarization phase of the action potential. Drug-induced hERG blockade is implicated in a long QT syndrome potentially causing deadly arrhythmias. Gating of hERG is characterized by rapid, voltage-dependent, C-type inactivation, which blocks ion conduction and is suggested to involve constriction of the selectivity filter. Mutations S620T, S641A, and S641T around the selectivity filter region of hERG have been shown to alter the voltage-dependence of channel inactivation.

1785-P: Drug-Induced Blockade of the Herg-Voltage-Gated Potassium Channel Decreases Glucose-Stimulated Insulin and GLP-1 Secretion in Healthy Participants: A Crossover Study

Background: Patients with Long-QT syndrome (LQTS) due to a loss-of-function mutation in hERG, encoding the hERG-voltage-gated potassium channel (Kv11.1), exhibit increased glucose-stimulated insulin secretion, defective glucagon secretion, and postprandial hypoglycemia. Recently, LQTS has been reported to be associated with an increased risk of type 2 diabetes. We aimed to investigate the effect of drug-induced (moxifloxacin) blockade of the hERG-channel on glucose homeostasis. Methods: Using a double-blinded, randomized, placebo-controlled, crossover design, 40 healthy young participants underwent two 6-hours oral glucose tolerance tests (OGTT). The intervention was a 4-day treatment period with a selective hERG-blocker (the antibiotic moxifloxacin 800 mg/day) or placebo, final dose two hours before the OGTT, separated by a 3-week washout period. Results: Moxifloxacin prolonged the QTc interval by 25 ms. Figure 1 shows the means and 95% CI from the moxifloxacin/placebo 6-hour OGTTs. Insulin and GLP-1 levels were reduced the first 90 min and glucose was reduced the first 15 min with moxifloxacin blockade, thereby increasing the Matsuda-index by 1.16 (95% CI -1.9;-0.4). Conclusion: Moxifloxacin-induced hERG blockade reduced peak postprandial glucose excursions and decreased the secretion of insulin and GLP-1. Disclosure C.R. Juhl: None. J. Burgdorf: None. C. Knudsen: None. S. Veedfald: None. J.J. Holst: Advisory Panel; Self; AstraZeneca, Merck Sharp & Dohme Corp., Novo Nordisk A/S, Zealand Pharma A/S. Other Relationship; Spouse/Partner; Antag Therapeutics. J. Kanters: None. S.S. Torekov: Research Support; Self; Novo Nordisk Inc. Funding Independent Research Fund Denmark; Danish Heart Association; Lundbeck Foundation

Predicting Arrhythmogenicity: Structural Modeling of Safe and Unsafe hERG Blockers

Human Ether-a-go-go-Related Gene (hERG) encodes a potassium-selective voltage-gated ion channel, KV11.1, essential for normal electrical activity in the heart. hERG mutations and blockage of the channel pore by drugs can cause long QT syndrome (LQTS) that predisposes individuals to arrhythmia and puts them at risk for stroke or sudden cardiac arrest. A major problem in antiarrhythmic drug therapies as well as drug development in general is the proclivity for many drugs and drug candidates to promote fatal arrhythmias through hERG blockade.

Capsule Networks Showed Excellent Performance in the Classification of hERG Blockers/Nonblockers.

Capsule networks (CapsNets), a new class of deep neural network architectures proposed recently by Hinton et al., have shown a great performance in many fields, particularly in image recognition and natural language processing. However, CapsNets have not yet been applied to drug discovery-related studies. As the first attempt, we in this investigation adopted CapsNets to develop classification models of hERG blockers/nonblockers; drugs with hERG blockade activity are thought to have a potential risk of cardiotoxicity. Two capsule network architectures were established: convolution-capsule network (Conv-CapsNet) and restricted Boltzmann machine-capsule networks (RBM-CapsNet), in which convolution and a restricted Boltzmann machine (RBM) were used as feature extractors, respectively. Two prediction models of hERG blockers/nonblockers were then developed by Conv-CapsNet and RBM-CapsNet with the Doddareddy's training set composed of 2,389 compounds. The established models showed excellent performance in an independent test set comprising 255 compounds, with prediction accuracies of 91.8 and 92.2% for Conv-CapsNet and RBM-CapsNet models, respectively. Various comparisons were also made between our models and those developed by other machine learning methods including deep belief network (DBN), convolutional neural network (CNN), multilayer perceptron (MLP), support vector machine (SVM), k-nearest neighbors (kNN), logistic regression (LR), and LightGBM, and with different training sets. All the results showed that the models by Conv-CapsNet and RBM-CapsNet are among the best classification models. Overall, the excellent performance of capsule networks achieved in this investigation highlights their potential in drug discovery-related studies.

Revealing Molecular Determinants of hERG Blocker and Activator Binding.

The Kv11.1 potassium channel, encoded by the human ether-a-go-go-related gene (hERG), plays an essential role in the cardiac action potential. hERG blockade by small molecules can induce "torsade de pointes" arrhythmias and sudden death; as such, it is an important off-target to avoid during drug discovery. Recently, a cryo-EM structure of the open channel state of hERG was reported, opening the door to in silico docking analyses and interpretation of hERG structure-activity relationships, with a view to avoiding blocking activity. Despite this, docking directly to this cryo-EM structure has been reported to yield binding modes that are unable to explain known mutagenesis data. In this work, we use molecular dynamics simulations to sample a range of channel conformations and run ensemble docking campaigns at the known hERG binding site below the selectivity filter, composed of the central cavity and the four deep hydrophobic pockets. We identify a hERG conformational state allowing discrimination of blockers vs nonblockers from docking; furthermore, the binding pocket agrees with mutagenesis data, and blocker binding modes fit the hERG blocker pharmacophore. We then use the same protocol to identify a binding pocket in the hERG channel pore for hERG activators, again agreeing with the reported mutagenesis. Our approach may be useful in drug discovery campaigns to prioritize candidate compounds based on hERG liability via virtual docking screens.

Purinostat Mesylate Is a Uniquely Potent and Selective Inhibitor of HDACs for the Treatment of BCR-ABL-Induced B-Cell Acute Lymphoblastic Leukemia.

This study was to perform preclinical evaluation of a novel class I and IIb HDAC-selective inhibitor, purinostat mesylate, for the treatment of Ph+ B-cell acute lymphoblastic leukemia (B-ALL). Biochemical assays were used to test enzymatic activity inhibition of purinostat mesylate. Ph+ leukemic cell lines and patient cells were used to evaluate purinostat mesylate activity in vitro. BL-2 secondary transplantation Ph+ B-ALL mouse model was used to validate its efficacy, mechanism, and pharmacokinetics properties in vivo. BCR-ABL(T315I)-induced primary B-ALL mouse model and PDX mouse model derived from relapsed Ph+ B-ALL patient post TKI treatment were used to determine the antitumor effect of purinostat mesylate for refractory or relapsed Ph+ B-ALL. Long-term toxicity and hERG blockade assays were used to safety evaluation of purinostat mesylate. Purinostat mesylate, a class I and IIb HDAC highly selective inhibitor, exhibited robust antitumor activity in hematologic cancers. Purinostat mesylate at low nanomolar concentration induced apoptosis, and downregulated BCR-ABL and c-MYC expression in Ph+ leukemia cell lines and primary Ph+ B-ALL cells from relapsed patients. Purinostat mesylate efficiently attenuated Ph+ B-ALL progression and significantly prolonged the survival both in BL-2 secondary transplantation model with clinical patient symptoms of Ph+ B-ALL, BCR-ABL(T315I)-induced primary B-ALL mouse model, and PDX model derived from patients with relapsed Ph+ B-ALL post TKI treatment. In addition, purinostat mesylate possesses favorable pharmacokinetics and low toxicity properties. Purinostat mesylate provides a new therapeutic strategy for patients with Ph+ B-ALL, including those who relapse after TKI treatment.

The effect of anti-emetic drugs on rat embryonic heart activity

Nausea and vomiting of pregnancy (NVP) is the most common medical complaint during pregnancy affecting up to 70% of pregnant women worldwide. Some antiemetic medications (AEM) (droperidol, domperidone, granisetron, metoclopramide and trifluoperazine) used to treat NVP have the unwanted side effect of hERG blockade. The hERG potassium channel is essential for normal heart rhythm in both the adult human and the human and rat embryo. Animal studies show hERG blockade in the embryo causes bradycardia and arrhythmia leading to cardiovascular malformations and other birth defects.Whole rat embryo in vitro culture was used to determine the effect of the above listed AEM and meclizine on the heart rate of Gestational day 13 rat embryos. These embryos are similar in size and heart development to 5–6-week human embryo. The results showed that all of the AEMs caused a concentration-dependent bradycardia. Droperidol had the lowest margin of safety.

3,5-Dialkoxypyridine analogues of bedaquiline are potent antituberculosis agents with minimal inhibition of the hERG channel.

Bedaquiline is a new drug of the diarylquinoline class that has proven to be clinically effective against drug-resistant tuberculosis, but has a cardiac liability (prolongation of the QT interval) due to its potent inhibition of the cardiac potassium channel protein hERG. Bedaquiline is highly lipophilic and has an extremely long terminal half-life, so has the potential for more-than-desired accumulation in tissues during the relatively long treatment durations required to cure TB. The present work is part of a program that seeks to identify a diarylquinoline that is as potent as bedaquiline against Mycobacterium tuberculosis, with lower lipophilicity, higher clearance, and lower risk for QT prolongation. Previous work led to the identification of compounds with greatly-reduced lipophilicity compounds that retain good anti-tubercular activity in vitro and in mouse models of TB, but has not addressed the hERG blockade. We now present compounds where the C-unit naphthalene is replaced by a 3,5-dialkoxy-4-pyridyl, demonstrate more potent in vitro and in vivo anti-tubercular activity, with greatly attenuated hERG blockade. Two examples of this series are in preclinical development.

Toward Reducing HERG Affinities for Dat Inhibitors with a Combined Machine Learning and Molecular Modeling Approach

The human Ether-à-go-go-Related (hERG) potassium channel plays a critical role in repolarization of cardiac action potentials. hERG blockade by drugs may result in abnormal heart rhythms such as the potentially fatal Torsades de Pointes. Experimental methods to identify compounds with high hERG affinities are time-consuming and costly. Compulsory hERG screening by the FDA has resulted in an increase of publicly available hERG binding data. Utilizing machine learning algorithms, we constructed structure activity relationship regression models to predict hERG binding affinity. The models were trained on molecules with known hERG affinities from the ChEMBL database. We filtered out erroneous points among the low affinity data that were found to affect prediction accuracy. We compared the performance of the eXtreme gradient boosting (XGBoost) and the random forest (RF) algorithms against several test sets and found that XGBoost slightly outperformed RF. The XGBoost models were then used to predict the hERG affinities of a series of dopamine transporter (DAT) inhibitors. The predictions showed good correlation with experimentally measured affinities. In addition, we implemented a machine-learning based classifier to identify compounds which appear to have extremely low affinity for hERG and are therefore unlikely to be well-predicted by the consensus regression model. In coordination with ligand-based approaches, we have used a protein structure-based method to characterize binding modes of these ligands against both our hDAT models and the recent cryo-EM structure of hERG, and to identify clues to improve hDAT affinities while reducing hERG affinities. Together, we established a combined ligand-based and protein structure-based computational framework, which is integrated with experimental approaches, to predict hERG affinities. We specifically applied it to facilitate the rational optimization of lead compounds targeting DAT for the treatment of psychostimulant use disorders.

The Molecular Mechanisms of State Dependent hERG Blockade by Dofetilide

The Molecular Mechanisms of State Dependent hERG Blockade by Dofetilide