Abstract:Drug-induced torsade de pointes (TdP) is a life-threatening ventricular arrhythmia responsible for the withdrawal of many drugs from the market. Although currently used TdP risk-assessment methods are effective, they are expensive and prone to produce false positives. In recent years, in silico cardiac simulations have proven to be a valuable tool for the prediction of drug effects. The objective of this work is to evaluate different biomarkers of drug-induced proarrhythmic risk and to develop an in silico risk classifier. Cellular simulations were performed using a modified version of the O'Hara et al. ventricular action potential model and existing pharmacological data (IC50 and effective free therapeutic plasma concentration, EFTPC) for 109 drugs of known torsadogenic risk (51 positive). For each compound, four biomarkers were tested: Tx (drug concentration leading to a 10% prolongation of the action potential over the EFTPC), TqNet (net charge carried by ionic currents when exposed to 10 times the EFTPC with respect to the net charge in control), Ttriang (triangulation for a drug concentration of 10 times the EFTPC over triangulation in control), and TEAD (drug concentration originating early afterdepolarizations over EFTPC). Receiver operating characteristic (ROC) curves were built for each biomarker to evaluate their individual predictive quality. At the optimal cutoff point, accuracies for Tx, TqNet, Ttriang, and TEAD were 89.9, 91.7, 90.8, and 78.9% respectively. The resulting accuracy of the hERG IC50 test (current biomarker) was 78.9%. When combining Tx, TqNet and Ttriang into a classifier based on decision trees, the prediction improves, achieving an accuracy of 94.5%. The sensitivity analysis revealed that most of the effects on the action potential are mainly due to changes in IKr, ICaL, INaL and IKs. In fact, considering that drugs affect only these four currents, TdP risk classification can be as accurate as when considering effects on the seven main currents proposed by the CiPA initiative. Finally, we built a ready-to-use tool (based on more than 450 000 simulations), which can be used to quickly assess the proarrhythmic risk of a compound. In conclusion, our in silico tool can be useful for the preclinical assessment of TdP-risk and to reduce costs related with new drug development. The TdP risk-assessment tool and the software used in this work are available at <a class="extLink" href="https://riunet.upv.es/handle/10251/136919">https://riunet.upv.es/handle/10251/136919</a>.The Supporting Information is available free of charge at <a class="ext-link" href="/doi/10.1021/acs.jcim.0c00201?goto=supporting-info">https://pubs.acs.org/doi/10.1021/acs.jcim.0c00201</a>.Equations of the activation (mss) and inactivation gates (hss, hssp, and jss) at the steady state of the rapid Na+ current in the O'Hara et al. model and in the modified version. IC50 values (nM), Hill coefficients, and effective free therapeutic plasma concentration (EFTPC) (nM) used to simulated the 109 drugs studied in this work. Performance of the four arrhythmogenic biomarkers depending on the currents on which the effect of the drug is assumed: the seven CiPA currents; IKr, ICaL, and IKs, or, IKr, ICaL, and INaL. Performance of the classifiers depending on the biomarkers used as inputs. Results of the sensitivity analysis with the multiple linear regression method. Results of the sensitivity analysis applying the PLS method. Decision trees that constitute the binary TdP risk classifier based on Tx, TqNet, and Ttriang (<a class="ext-link" href="/doi/suppl/10.1021/acs.jcim.0c00201/suppl_file/ci0c00201_si_001.pdf">PDF</a>)This article has not yet been cited by other publications.

A stacking ensemble machine learning model for evaluating cardiac toxicity of drugs based on in silico biomarkers

Explainable artificial intelligence (XAI) to find optimal in-silico biomarkers for cardiac drug toxicity evaluation

Drug Screening Using a Library of Human Induced Pluripotent Stem Cell–Derived Cardiomyocytes Reveals Disease-Specific Patterns of Cardiotoxicity

Screening Drug-Induced Arrhythmia Using Human Induced Pluripotent Stem Cell–Derived Cardiomyocytes and Low-Impedance Microelectrode Arrays

In Vitro Assessing the Risk of Drug-Induced Cardiotoxicity by Embryonic Stem Cell-Based Biosensor

Drug safety assessment by machine learning models

Evaluation of cardiac pro-arrhythmic risks using the artificial neural network with ToR–ORd in silico model output

Prediction of Drug-Induced TdP Risks Using Machine Learning and Rabbit Ventricular Wedge Assay

Establishment and Validation of a Torsade De Pointes Prediction Model Based on Human Ipsc‑derived Cardiomyocytes.

Enhanced drug classification using machine learning with multiplexed cardiac contractility assays

In Silico Classifiers for the Assessment of Drug Proarrhythmicity

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

A Pharmacophore Modeling Study of Drugs Inducing Cardiotoxic Side Effects

Statistical Learning in Preclinical Drug Proarrhythmic Assessment

Development of models for predicting Torsade de Pointes cardiac arrhythmias using perceptron neural networks

An intelligent biosensing platform using phase space reconstruction‐assisted convolutional neural network for drug‐induced cardiotoxicity assessment

ToxSTK: A Multi-Target Toxicity Assessment Utilizing Molecular Structure and Stacking Ensemble Learning

In silico prediction of hERG blockers using machine learning and deep learning approaches

Ligand-based prediction of hERG-mediated cardiotoxicity based on the integration of different machine learning techniques

Prediction of hERG potassium channel blockage using ensemble learning methods and molecular fingerprints.

ADMET Evaluation in Drug Discovery. 16. Predicting Herg Blockers by Combining Multiple Pharmacophores and Machine Learning Approaches.