Computer-based prediction of the drug proarrhythmic effect: problems, issues, known and suspected challenges

Deriving waveform parameters from calcium transients in human iPSC-derived cardiomyocytes to predict cardiac activity with machine learning

Human induced pluripotent stem cell-derived cardiomyocytes have been established to detect dynamic calcium transients by fast kinetic fluorescence assays that provide insights into specific aspects of clinical cardiac activity. However, the precise derivation and use of waveform parameters to predict cardiac activity merit deeper investigation. In this study, we derived, evaluated, and applied 38 waveform parameters in a novel Python framework, including (among others) peak frequency, peak amplitude, peak widths, and a novel parameter, shoulder-tail ratio. We then trained a random forest model to predict cardiac activity based on the 25 parameters selected by correlation analysis. The area under the curve (AUC) obtained for leave-one-compound-out cross-validation was 0.86, thereby replicating the predictions of conventional methods and outperforming fingerprint-based methods by a large margin. This work demonstrates that machine learning is able to automate the assessment of cardiovascular liability from waveform data, reducing any risk of user-to-user variability and bias.

•

An open-source algorithm was developed to derive parameters from waveform data

•

A machine learning model was trained to predict cardiac activity of compounds

•

Three parameters for peak width, height, and shape were found to be most predictive

•

The model can facilitate the assessment of cardiovascular liability

Inflammation as a Risk Factor in Cardiotoxicity: An Important Consideration for Screening During Drug Development

Numerous commonly prescribed drugs, including antiarrhythmics, antihistamines, and antibiotics, carry a proarrhythmic risk and may induce dangerous arrhythmias, including the potentially fatal Torsades de Pointes. For this reason, cardiotoxicity testing has become essential in drug development and a required step in the approval of any medication for use in humans. Blockade of the hERG K⁺ channel and the consequent prolongation of the QT interval on the ECG have been considered the gold standard to predict the arrhythmogenic risk of drugs. In recent years, however, preclinical safety pharmacology has begun to adopt a more integrative approach that incorporates mathematical modeling and considers the effects of drugs on multiple ion channels. Despite these advances, early stage drug screening research only evaluates QT prolongation in experimental and computational models that represent healthy individuals. We suggest here that integrating disease modeling with cardiotoxicity testing can improve drug risk stratification by predicting how disease processes and additional comorbidities may influence the risks posed by specific drugs. In particular, chronic systemic inflammation, a condition associated with many diseases, affects heart function and can exacerbate medications’ cardiotoxic effects. We discuss emerging research implicating the role of inflammation in cardiac electrophysiology, and we offer a perspective on how in silico modeling of inflammation may lead to improved evaluation of the proarrhythmic risk of drugs at their early stage of development.

How circadian variability of the heart rate and plasma electrolytes concentration influence the cardiac electrophysiology - model-based case study

The circadian rhythm of cardiac electrophysiology is dependent on many physiological and biochemical factors. Provided, that models describing the circadian patterns of cardiac activity and/or electrophysiology which have been verified to the acceptable level, modeling and simulation can give answers to many of heart chronotherapy questions. The aim of the study was to assess the performance of the circadian models implemented in Cardiac Safety Simulator v 2.2 (Certara, Sheffield, UK) (CSS), as well as investigate the influence ofcircadian rhythms on the simulation results in terms of cardiac safety. The simulations which were run in CSS accounted for inter-individual and intra-individual variability. Firstly, the diurnal variations in QT interval length in a healthy population were simulated accounting for heart rate (HR) circadian changes alone, or with concomitant diurnal variations of plasma ion concentrations. Next, tolterodine was chosen as an exemplary drug for PKPD modelling exercise to assess the role of circadian rhythmicity in the prediction of drug effects on QT interval. The results of the simulations were in line with clinical observations, what can serve as a verification of the circadian models implemented in CSS. Moreover, the results have suggested that the circadian variability of the electrolytes balance is the main factor influencing QT circadian pattern. The fluctuation of ion concentration increases the intra-subject variability of predicted drug-triggered QT corrected for HR (QTc) prolongation effect and, in case of modest drug effect on QTc interval length, allows to capture this effect.

Early assessment of proarrhythmic risk of drugs using the in vitro data and single-cell-based in silico models: proof of concept

BACKGROUND AND PURPOSE

To determine the predictive performance of in silico models using drug-specific preclinical cardiac electrophysiology data to investigate drug-induced arrhythmia risk (e.g. Torsade de pointes (TdP)) in virtual human subjects.

EXPERIMENTAL APPROACH

To assess drug proarrhythmic risk, we used a set of in vitro electrophysiological measurements describing ion channel inhibition triggered by the investigated drugs. The Cardiac Safety Simulator version 2.0 (CSS; Simcyp, Sheffield, UK) platform was used to simulate human left ventricular cardiac myocyte action potential models.

RESULTS

This study shows the impact of drug concentration changes on particular ionic currents by using available experimental data. The simulation results display safety threshold according to drug concentration threshold and log (threshold concentration/ effective therapeutic plasma concentration (ETPC)).

CONCLUSION AND IMPLICATIONS

We reproduced the underlying biophysical characteristics of cardiac cells resulted in effects of drugs associated with cardiac arrhythmias (action potential duration (APD) and QT prolongation and TdP) which were observed in published 3D simulations, yet with much less computational burden.

Early Afterdepolarizations with Growing Amplitudes via Delayed Subcritical Hopf Bifurcations and Unstable Manifolds of Saddle Foci in Cardiac Action Potential Dynamics

Early afterdepolarizations (EADs) are pathological oscillations in cardiac action potentials during the repolarization phase and may be caused by drug side effects, ion channel disease or oxidative stress. The most widely observed EAD pattern is characterized by oscillations with growing amplitudes. So far, its occurence has been explained in terms of a supercritical Hopf bifurcation in the fast subsystem of the action potential dynamics from which stable limit cycles with growing amplitudes emerge. The novel contribution of this article is the introduction of two alternative explanations of EAD genesis with growing amplitudes that do not involve stable limit cycles in fast subsystems. In particular, we demonstrate that EAD patterns with growing amplitudes may alternatively arise due to a delayed subcritical Hopf bifurcation or an unstable manifold of a saddle focus fixed point in the full fast-slow system modelling the action potential. Our work extends the list of possible dynamical EAD mechanisms and may contribute to a classification of drug effects in preclinical cardiotoxicity testing.

Drug-drug interactions and QT prolongation as a commonly assessed cardiac effect - comprehensive overview of clinical trials

Background

Proarrhythmia assessment is one of the major concerns for regulatory bodies and pharmaceutical industry. ICH guidelines recommending preclinical tests have been established in attempt to eliminate the risk of drug-induced arrhythmias. However, in the clinic, arrhythmia occurrence is determined not only by the inherent property of a drug to block ion currents and disturb electrophysiological activity of cardiac myocytes, but also by many other factors modifying individual risk of QT prolongation and subsequent proarrhythmia propensity. One of those is drug-drug interactions. Since polypharmacy is a common practice in clinical settings, it can be anticipated that there is a relatively high risk that the patient will receive at least two drugs mutually modifying their proarrhythmic potential and resulting either in triggering the occurrence or mitigating the clinical symptoms. The mechanism can be observed either directly at the pharmacodynamic level by competing for the molecular targets, or indirectly by modifying the physiological parameters, or at the pharmacokinetic level by alteration of the active concentration of the victim drug.

Methods

This publication provides an overview of published clinical studies on pharmacokinetic and/or pharmacodynamic drug-drug interactions in humans and their electrophysiological consequences (QT interval modification). Databases of PubMed and Scopus were searched and combinations of the following keywords were used for Title, Abstract and Keywords fields: interaction, coadministration, combination, DDI and electrocardiographic, QTc interval, ECG. Only human studies were included. Over 4500 publications were retrieved and underwent preliminary assessment to identify papers accordant with the topic of this review. 76 papers reporting results for 96 drug combinations were found and analyzed.

Results

The results show the tremendous variability of drug-drug interaction effects, which makes one aware of complexity of the problem, and suggests the need for assessment of an additional risk factors and careful ECG monitoring before administration of drugs with anticipated QT prolongation.

Conclusions

DDIs can play significant roles in drugs’ cardiac safety, as evidenced by the provided examples. Assessment of the pharmacodynamic effects of the drug interactions is more challenging as compared to the pharmacokinetic due to the significant diversity in the endpoints which should be analyzed specifically for various clinical effects. Nevertheless, PD components of DDIs should be accounted for as PK changes alone do not allow to fully explain the electrophysiological effects in clinic situations.

Electronic supplementary material

The online version of this article (doi:10.1186/s40360-016-0053-1) contains supplementary material, which is available to authorized users.