Training and validation of a human ventricular cardiac action potential model to predict action potential duration response to acute inhibition of two major ionic currents

Thesis

Drugs can perturb the electrical activity of cardiac muscle cells, which may lead to cardiac adverse events. The inhibition of a potassium current (IKr) has been shown to prolong the cellular and ventricular repolarisation time, respectively characterised by the action potential (AP) duration (APD) and the QT interval on the ECG, promoting arrhythmia. The in vitro characterisation of IKr inhibition is a key part of pre-clinical cardiac safety assessment for drug candidates. The proarrhythmic risk is further evaluated clinically, with particular focus on QT.

The inhibition of a calcium current (ICaL) can mitigate the prolongation of APD and QT induced by IKr inhibition. Because pre-clinical and clinical guidelines are based on different endpoints, compounds inhibiting both IKr and ICaL may be incorrectly deemed proarrhythmic pre-clinically, in which case compounds with a potentially advantageous risk/benefit ratio may not reach the patient.

Biophysically-detailed models of the cardiac AP have emerged as tools linking in vitro observations with in vivo proarrhythmic risk. This thesis establishes a framework to integrate newly generated experimental data into the development of a novel predictive AP model.

New data were generated ex vivo in adult human ventricular tissue to investigate the relationship between drug-induced changes in APD, IKr inhibition, and ICaL inhibition. A framework was developed to calibrate an AP model to this new data, with particular considerations for limitations on the data and on the understanding of mechanisms underlying the AP. Using this framework, a new AP model was calibrated, achieving improved predictions of drug-induced changes in APD in comparison with previous AP models.

Accurately predicting the changes in APD from in vitro experiments is the first step towards reliable predictions of drug-induced changes in clinically-relevant biomarkers. This work therefore contributes to a greater integration of in vitro experiments and in silico tools, and to enhanced pre-clinical predictions of the cardiac safety assessment of new compounds.

Authors: Barral, Y-SHM

• DOI: 10.5287/ora-e9wvbyndp
• Type of award: DPhil
• Level of award: Doctoral
• Awarding institution: University of Oxford
• Language: English
• Keywords: model calibration; quantitative systems pharmacology; computational biology; cardiac safety
• Subjects: Computational biology