Prediction of the effects of drugs on cardiac activity using computer simulations

Abstract

Cardiovascular disease remains the main cause of death in Europe. Cardiac arrhythmias are an important cause of sudden death, but their mechanisms are complex. This denotes the importance of their study and prevention. Research on cardiac electrophysiology has shown that electrical abnormalities caused by mutations in cardiac channels can trigger arrhythmias. Surprisingly, a wide variety of drugs have also shown proarrhythmic potential, including those that we use to prevent arrhythmia. Current guidelines designed a test to identify dangerous drugs by assessing their blocking power on a single ion channel to address this situation. Study of drug-channel interactions has revealed not only compounds that block multiple channels but also a great complexity in those interactions. This could explain why similar drugs can show vastly different effects in some diseases. There are two important challenges regarding the effects of drugs on cardiac electrophysiology. On the one hand, companies and regulators are in search of a high throughput tool that improves proarrhythmic potential detection during drug development. On the other hand, patients with electrical abnormalities often require safer personalized treatments owing to their condition. Computer simulations provide an unprecedented power to tackle complex biophysical phenomena. They should prove useful determining the characteristics that define the drugs' beneficial and unwanted effects by reproducing experimental and clinical observations. In this PhD thesis, we used computational models and simulations to address the two abovementioned challenges. We split the study of drug effects on the cardiac activity into the study of their safety and efficacy, respectively. For the former, we took a wider approach and generated a new easy-to-use biomarker for proarrhythmic potential classification using cardiac cell and tissue human action potential models. We integrated multiple channel block through IC50s and therapeutic concentrations to improve its predictive power. Then, we quantified the proarrhythmic potential of 84 drugs to train the biomarker. Our results suggest that it could be used to test the proarrhythmic potential of new drugs. For the second challenge, we took a more specific approach and sought to improve the therapy of patients with cardiac electrical abnormalities. Therefore, we created a detailed model for the long QT syndrome-causing V411M mutation of the sodium channel reproducing clinical and experimental data. We tested the potential benefits of ranolazine, while giving insights into the mechanisms that drive flecainide's effectiveness. Our results suggest that while both drugs showed different mechanisms of sodium channel block, ranolazine could prove beneficial in these patients.