Action Potential Morphology Accurately Predicts Proarrhythmic Risk for Drugs With Potential to Prolong Cardiac...

Drug-induced or acquired long QT syndrome occurs as a result of the unintended disruption of cardiac repolarization due to drugs that block cardiac ion channels. These side effects have been responsible for the withdrawal of a range of drugs from market and are a common reason for termination of the development of new drugs in the preclinical stage. Existing approaches to risk prediction are expensive and overly sensitive meaning that recently there have been renewed efforts, largely driven by the comprehensive proarrhythmic assay initiative, to develop more accurate methods for allocation of proarrhythmic risk.

In this study, we aimed to quantify changes in the morphology of the repolarization phase of the cardiac action potential as an indicator of proarrhythmia, supposing that these shape changes might precede the emergence of ectopic depolarizations that trigger arrhythmia. To do this, we describe a new method of quantifying action potential morphology by measuring the radius of curvature of the repolarization phase both in simulated action potentials, as well as in action potentials measured from induced pluripotent stem cell-derived cardiomyocytes. Features derived from the curvature signal were used as inputs for logistic regressions to predict proarrhythmic risk.

Optimal risk classifiers based on morphology were able to correctly classify risk to drugs in the comprehensive proarrhythmic assay initiative panels with very high accuracy (0.9375) and outperformed conventional metrics based on action potential duration at 90% repolarization, triangulation, and charge movement (qNet).

Analysis of action potential morphology in response to proarrhythmic drugs improves prediction of torsadogenic risk. Furthermore, morphology metrics can be measured directly from the action potential, potentially eliminating the burden of undertaking complex screens of potency and drug-binding kinetics against multiple cardiac ion channels. As such, this method has the potential to improve and streamline regulatory assessment of proarrhythmia in preclinical drug development.