Determination of hERG channel blockers using a decision tree

Prediction of Kv11.1 potassium channel PAS-domain variants trafficking via machine learning

Congenital long QT syndrome (LQTS) is characterized by a prolonged QT-interval on an electrocardiogram (ECG). An abnormal prolongation in the QT-interval increases the risk for fatal arrhythmias. Genetic variants in several different cardiac ion channel genes, including KCNH2, are known to cause LQTS. Here, we evaluated whether structure-based molecular dynamics (MD) simulations and machine learning (ML) could improve the identification of missense variants in LQTS-linked genes. To do this, we investigated KCNH2 missense variants in the Kv11.1 channel protein shown to have wild type (WT) like or class II (trafficking-deficient) phenotypes in vitro. We focused on KCNH2 missense variants that disrupt normal Kv11.1 channel protein trafficking, as it is the most common phenotype for LQTS-associated variants. Specifically, we used computational techniques to correlate structural and dynamic changes in the Kv11.1 channel protein PAS domain (PASD) with Kv11.1 channel protein trafficking phenotypes. These simulations unveiled several molecular features, including the numbers of hydrating waters and hydrogen bonding pairs, as well as folding free energy scores, that are predictive of trafficking. We then used statistical and machine learning (ML) (Decision tree (DT), Random forest (RF), and Support vector machine (SVM)) techniques to classify variants using these simulation-derived features. Together with bioinformatics data, such as sequence conservation and folding energies, we were able to predict with reasonable accuracy (≈75%) which KCNH2 variants do not traffic normally. We conclude that structure-based simulations of KCNH2 variants localized to the Kv11.1 channel PASD led to an improvement in classification accuracy. Therefore, this approach should be considered to complement the classification of variant of unknown significance (VUS) in the Kv11.1 channel PASD.

Structural analysis of hERG channel blockers and the implications for drug design

The repolarizing current (I_kr) produced by the hERG potassium channel forms a major component of the cardiac action potential and blocking this current by small molecule drugs can lead to life-threatening cardiotoxicity. Understanding the mechanisms of drug-mediated hERG inhibition is essential to develop a second generation of safe drugs, with minimal cardiotoxic effects. Although various computational tools and drug design guidelines have been developed to avoid binding of drugs to the hERG pore domain, there are many other aspects that are still open for investigation. This includes the use computational modelling to study the implications of hERG mutations on hERG structure and trafficking, the interactions of hERG with hERG chaperone proteins and with membrane-soluble molecules, the mechanisms of drugs that inhibit hERG trafficking and drugs that rescue hERG mutations. The plethora of available experimental data regarding all these aspects can guide the construction of much needed robust computational structural models to study these mechanisms for the rational design of safe drugs.

Mutation-Specific Differences in Kv7.1 (KCNQ1) and Kv11.1 (KCNH2) Channel Dysfunction and Long QT Syndrome Phenotypes

The electrocardiogram (ECG) empowered clinician scientists to measure the electrical activity of the heart noninvasively to identify arrhythmias and heart disease. Shortly after the standardization of the 12-lead ECG for the diagnosis of heart disease, several families with autosomal recessive (Jervell and Lange-Nielsen Syndrome) and dominant (Romano-Ward Syndrome) forms of long QT syndrome (LQTS) were identified. An abnormally long heart rate-corrected QT-interval was established as a biomarker for the risk of sudden cardiac death. Since then, the International LQTS Registry was established; a phenotypic scoring system to identify LQTS patients was developed; the major genes that associate with typical forms of LQTS were identified; and guidelines for the successful management of patients advanced. In this review, we discuss the molecular and cellular mechanisms for LQTS associated with missense variants in KCNQ1 (LQT1) and KCNH2 (LQT2). We move beyond the "benign" to a "pathogenic" binary classification scheme for different KCNQ1 and KCNH2 missense variants and discuss gene- and mutation-specific differences in K+ channel dysfunction, which can predispose people to distinct clinical phenotypes (e.g., concealed, pleiotropic, severe, etc.). We conclude by discussing the emerging computational structural modeling strategies that will distinguish between dysfunctional subtypes of KCNQ1 and KCNH2 variants, with the goal of realizing a layered precision medicine approach focused on individuals.

Current development of integrated web servers for preclinical safety and pharmacokinetics assessments in drug development

In drug development, preclinical safety and pharmacokinetics assessments of candidate drugs to ensure the safety profile are a must. While in vivo and in vitro tests are traditionally used, experimental determinations have disadvantages, as they are usually time-consuming and costly. In silico predictions of these preclinical endpoints have each been developed in the past decades. However, only a few web-based tools have integrated different models to provide a simple one-step platform to help researchers thoroughly evaluate potential drug candidates. To efficiently achieve this approach, a platform for preclinical evaluation must not only predict key ADMET (absorption, distribution, metabolism, excretion and toxicity) properties but also provide some guidance on structural modifications to improve the undesired properties. In this review, we organized and compared several existing integrated web servers that can be adopted in preclinical drug development projects to evaluate the subject of interest. We also introduced our new web server, Virtual Rat, as an alternative choice to profile the properties of drug candidates. In Virtual Rat, we provide not only predictions of important ADMET properties but also possible reasons as to why the model made those structural predictions. Multiple models were implemented into Virtual Rat, including models for predicting human ether-a-go-go-related gene (hERG) inhibition, cytochrome P450 (CYP) inhibition, mutagenicity (Ames test), blood-brain barrier penetration, cytotoxicity and Caco-2 permeability. Virtual Rat is free and has been made publicly available at https://virtualrat.cmdm.tw/.

The current limits in virtual screening and property prediction

Beyond finding inhibitors that show high binding affinity to the respective target, there is the challenge of optimizing their properties with respect to metabolic and toxicological issues, as well as further off-target effects. To reduce the experimental effort of synthesizing and testing actual substances in corresponding assays, virtual screening has become an indispensable toolbox in preclinical development. The scope of application covers the prediction of molecular properties including solubility, metabolic liability and binding to antitargets, such as the hERG channel. Furthermore, prediction of binding sites and drugable targets are emerging aspects of virtual screening. Issues involved with the currently applied computational models including machine learning algorithms are outlined, such as limitations to the accuracy of prediction and overfitting.

Computational investigations of hERG channel blockers: New insights and current predictive models

Identification of potential human Ether-a-go-go Related-Gene (hERG) potassium channel blockers is an essential part of the drug development and drug safety process in pharmaceutical industries or academic drug discovery centers, as they may lead to drug-induced QT prolongation, arrhythmia and Torsade de Pointes. Recent reports also suggest starting to address such issues at the hit selection stage. In order to prioritize molecules during the early drug discovery phase and to reduce the risk of drug attrition due to cardiotoxicity during pre-clinical and clinical stages, computational approaches have been developed to predict the potential hERG blockage of new drug candidates. In this review, we will describe the current in silico methods developed and applied to predict and to understand the mechanism of actions of hERG blockers, including ligand-based and structure-based approaches. We then discuss ongoing research on other ion channels and hERG polymorphism susceptible to be involved in LQTS and how systemic approaches can help in the drug safety decision.

In silico prediction of hERG inhibition

The voltage-gated potassium channel encoded by hERG carries a delayed rectifying potassium current (IKr) underlying repolarization of the cardiac action potential. Pharmacological blockade of the hERG channel results in slowed repolarization and therefore prolongation of action potential duration and an increase in the QT interval as measured on an electrocardiogram. Those are possible to cause sudden death, leading to the withdrawals of many drugs, which is the reason for hERG screening. Computational in silico prediction models provide a rapid, economic way to screen compounds during early drug discovery. In this review, hERG prediction models are classified as 2D and 3D quantitative structure–activity relationship models, pharmacophore models, classification models, and structure based models (using homology models of hERG).

Novel Bayesian classification models for predicting compounds blocking hERG potassium channels

AIM

A large number of drug-induced long QT syndromes are ascribed to blockage of hERG potassium channels. The aim of this study was to construct novel computational models to predict compounds blocking hERG channels.

METHODS

Doddareddy's hERG blockage data containing 2644 compounds were used, which divided into training (2389) and test (255) sets. Laplacian-corrected Bayesian classification models were constructed using Discovery Studio. The models were internally validated with the training set of compounds, and then applied to the test set for validation. Doddareddy's experimentally validated dataset with 60 compounds was used for external test set validation.

RESULTS

A Bayesian classification model considering the effects of four molecular properties (Mw, PPSA, ALogP and pKa_basic) as well as extended-connectivity fingerprints (ECFP_14) exhibited a global accuracy (91%), parameter sensitivity (90%) and specificity (92%) in the test set validation, and a global accuracy (58%), parameter sensitivity (61%) and specificity (57%) in the external test set validation.

CONCLUSION

The novel model is better than those in the literatures for predicting compounds blocking hERG channels, and can be used for large-scale prediction.

Assessing the translatability of in vivo cardiotoxicity mechanisms to in vitro models using causal reasoning

Drug-induced cardiac toxicity has been implicated in 31% of drug withdrawals in the USA. The fact that the risk for cardiac-related adverse events goes undetected in preclinical studies for so many drugs underscores the need for better, more predictive in vitro safety screens to be deployed early in the drug discovery process. Unfortunately, many questions remain about the ability to accurately translate findings from simple cellular systems to the mechanisms that drive toxicity in the complex in vivo environment. In this study, we analyzed translatability of cardiotoxic effects for a diverse set of drugs from rodents to two different cell systems (rat heart tissue-derived cells (H9C2) and primary rat cardiomyocytes (RCM)) based on their transcriptional response. To unravel the altered pathway, we applied a novel computational systems biology approach, the Causal Reasoning Engine (CRE), to infer upstream molecular events causing the observed gene expression changes. By cross-referencing the cardiotoxicity annotations with the pathway analysis, we found evidence of mechanistic convergence towards common molecular mechanisms regardless of the cardiotoxic phenotype. We also experimentally verified two specific molecular hypotheses that translated well from in vivo to in vitro (Kruppel-like factor 4, KLF4 and Transforming growth factor beta 1, TGFB1) supporting the validity of the predictions of the computational pathway analysis. In conclusion, this work demonstrates the use of a novel systems biology approach to predict mechanisms of toxicity such as KLF4 and TGFB1 that translate from in vivo to in vitro. We also show that more complex in vitro models such as primary rat cardiomyocytes may not offer any advantage over simpler models such as immortalized H9C2 cells in terms of translatability to in vivo effects if we consider the right endpoints for the model. Further assessment and validation of the generated molecular hypotheses would greatly enhance our ability to design predictive in vitro cardiotoxicity assays.

Towards in silico identification of the human ether-a-go-go-related gene channel blockers: discriminative vs. generative classification models

HERG potassium channels have a critical role in the normal electrical activity of the heart. The blockade of hERG channels in heart cells can result in a potentially fatal disorder called long QT syndrome. HERG channels can be blocked by compounds with diverse structures belonging to several drug classes. Presented herein are generative (Generative Topographic Maps) and discriminative (Support Vector Machines) classification models to categorize the compounds in silico into active and inactive classes by using different types of descriptors. The predictive performance of discriminative and generative classification models has been compared. Here, the possibility of using Generative Topographic Maps as an approach for applicability domain analysis and to generate probability-based descriptors was demonstrated to our knowledge for the first time. Comparison of obtained results with the models developed by other teams on the same data set has been performed.

Modeling of open, closed, and open-inactivated states of the hERG1 channel: structural mechanisms of the state-dependent drug binding

The human ether-a-go-go related gene 1 (hERG1) K ion channel is a key element for the rapid component of the delayed rectified potassium current in cardiac myocytes. Since there are no crystal structures for hERG channels, creation and validation of its reliable atomistic models have been key targets in molecular cardiology for the past decade. In this study, we developed and vigorously validated models for open, closed, and open-inactivated states of hERG1 using a multistep protocol. The conserved elements were derived using multiple-template homology modeling utilizing available structures for Kv1.2, Kv1.2/2.1 chimera, and KcsA channels. Then missing elements were modeled with the ROSETTA De Novo protein-designing suite and further refined with all-atom molecular dynamics simulations. The final ensemble of models was evaluated for consistency to the reported experimental data from biochemical, biophysical, and electrophysiological studies. The closed state models were cross-validated against available experimental data on toxin footprinting with protein-protein docking using hERG state-selective toxin BeKm-1. Poisson-Boltzmann calculations were performed to determine gating charge and compare it to electrophysiological measurements. The validated structures offered us a unique chance to assess molecular mechanisms of state-dependent drug binding in three different states of the channel.

Machine learning algorithms for the prediction of hERG and CYP450 binding in drug development

IMPORTANCE OF THE FIELD

The cost of developing new drugs is estimated at approximately $1 billion; the withdrawal of a marketed compound due to toxicity can result in serious financial loss for a pharmaceutical company. There has been a greater interest in the development of in silico tools that can identify compounds with metabolic liabilities before they are brought to market.

AREAS COVERED IN THIS REVIEW

The two largest classes of machine learning (ML) models, which will be discussed in this review, have been developed to predict binding to the human ether-a-go-go related gene (hERG) ion channel protein and the various CYP isoforms. Being able to identify potentially toxic compounds before they are made would greatly reduce the number of compound failures and the costs associated with drug development.

WHAT THE READER WILL GAIN

This review summarizes the state of modeling hERG and CYP binding towards this goal since 2003 using ML algorithms.

TAKE HOME MESSAGE

A wide variety of ML algorithms that are comparable in their overall performance are available. These ML methods may be applied regularly in discovery projects to flag compounds with potential metabolic liabilities.

hERG-related drug toxicity and models for predicting hERG liability and QT prolongation

BACKGROUND

hERG K(+) channels have been recognized as a primary antitarget in safety pharmacology. Their blockade, caused by several drugs with different therapeutic indications, may lead to QT prolongation and, eventually, to potentially fatal arrhythmia, namely torsade de pointes. Therefore, a number of preclinical models have been developed to predict hERG liability early in the drug development process.

OBJECTIVE

The aim of this review is to outline the present state of the art on drug-induced hERG blockade, providing insights on the predictive value of in vitro and in silico models for hERG liability.

METHODS

On the basis of latest reports, high-throughput preclinical models have been discussed outlining advantages and limitations.

CONCLUSION

Although no single model has an absolute value, an integrated risk assessment is recommended to predict the pro-arrhythmic risk of a given drug. This prediction requires expertise from different areas and should encompass emerging issues such as interference with hERG trafficking and QT shortening.

Collation, assessment and analysis of literature in vitro data on hERG receptor blocking potency for subsequent modeling of drugs' cardiotoxic properties

The assessment of the torsadogenic potency of a new chemical entity is a crucial issue during lead optimization and the drug development process. It is required by the regulatory agencies during the registration process. In recent years, there has been a considerable interest in developing in silico models, which allow prediction of drug-hERG channel interaction at the early stage of a drug development process. The main mechanism underlying an acquired QT syndrome and a potentially fatal arrhythmia called torsades de pointes is the inhibition of potassium channel encoded by hERG (the human ether-a-go-go-related gene). The concentration producing half-maximal block of the hERG potassium current (IC(50)) is a surrogate marker for proarrhythmic properties of compounds and is considered a test for cardiac safety of drugs or drug candidates. The IC(50) values, obtained from data collected during electrophysiological studies, are highly dependent on experimental conditions (i.e. model, temperature, voltage protocol). For the in silico models' quality and performance, the data quality and consistency is a crucial issue. Therefore the main objective of our work was to collect and assess the hERG IC(50) data available in accessible scientific literature to provide a high-quality data set for further studies.

Computer-aided prediction of QT-prolongation

Drug-induced cardiac arrhythmia is acknowledged as a serious obstacle in successful development of new drugs. Several methods for in silico prediction of acquired long QT syndrome (LQTS) caused by the pharmacological blockade of human hERG K+ channels are discussed in literature. We propose to use the computer program PASS, which estimates the probabilities of about 3000 biological activities, not only for prediction of hERG blockade and QT-prolongation but also for the analysis of indirect mechanisms of these actions. After addition in the PASS training set of 163 compounds with data on QT-Prolongation and re-training, it was shown that accuracy of prediction was 87.1% and 81.8% for hERG blockade and QT-prolongation, respectively. Using computer program PharmaExpert we found that in the predicted biological activity spectra there was a certain correlation between the hERG blockade and some other molecular mechanisms of action. Possible role of 1-phosphatidylinositol-4-phospate 5-kinase, dimethylargininase and progesterone 11 alpha-monooxygenase inhibition in hERG blockade was discussed.

Estimation of hERG inhibition of drug candidates using multivariate property and pharmacophore SAR

We describe the development of a computational model for the prediction of the inhibition of K(+) flow through the hERG ion channel. Using a collection of 1075 discovery compounds with hERG inhibition measured in our standard patch-clamp electrophysiology assay, molecular features important for drug-induced inhibition were identified using a combination of statistical inference algorithms and manual hypothesis generation and testing. While many of the features used in the model reflect those referenced in the literature, several aspects of the model provide new insight into the role of physicochemical properties, electrostatics, and novel pharmacophores in hERG inhibition. Coefficients for these 10 features were then determined by least median squares regression, resulting in a model with an R(2) approximately 0.66 and RMS error (RMSe) of 0.47 log units for an external test set. Significant additional validation performed using a large collection of subsequent discovery data has been very encouraging with an R(2)=0.54 and an RMSe of 0.63 log units. The performance of the model across several different chemotypes is demonstrated and discussed.

A novel structure-based virtual screening model for the hERG channel blockers

The hERG potassium channel is a key effector of cardiac repolarization and the blockade of this channel could cause arrhythmia. Thus, hERG channel blockade plays an important role for the potential pro-arrhythmic liability. In this report, binding of blockers to the hERG potassium channel is investigated using a combination of homology modeling, molecular docking, and molecular simulations, where blockade activities are evaluated using the linear regression model of GoldScore fitness. This structure-based virtual screening model is able to estimate the pIC(50) value of a wide range of ligands for the hERG potassium channel. The docked poses for ligands are also consistent with published mutation. Therefore, this model for the prediction of hERG channel blockade has the potential to provide cost-effective virtual screening tools for the evaluation of the cardiac liability of new chemical entities.