Development and Validation of an In Silico Rabbit Purkinje Cell Action Potential Model: A Step Towards a Drug Safety Testing Tool

Jordi Cano1, Julio Gomis-Tena1, Alexander Amberg2, Lennart Anger2, Véronique Ballet2, Jean-Michel Guillon2, Manuel Pastor3, Ferran Sanz3, Lucia Romero1, Javier Saiz1
1Centro de Investigación e Innovación en Bioingeniería - Universitat Politècnica de València, 2SANOFI, R&D Preclinical Safety, 3Research Programme on Biomedical Informatics (GRIB), Institut Hospital del Mar d’Investigacions Mèdiques (IMIM), Dept. of Experimental and Health Sciences, Universitat Pompeu Fabra


Abstract

Companies and regulators evaluate compounds cardiac safety commonly through in-vitro assays involving cardiac cells. Rabbit cardiac Purkinje Cells (RPCs) are very sensitive to drug effects such as Action Potential Duration (APD) variation. Our objective is to create a novel RPC model calibrated with the latest experimental data. We developed a new RPC model using Pan-Rudy’s Canine Purkinje cell model as a framework. We adapted important parameters, namely, size, ion fluxes, potassium currents (IKr, IKs, Ito and IK1), included Markov formulations for sodium currents, and tuned overall conductances and calcium dynamics to match the main experimentally observed RPCs features. Steady-state was reached by pacing the model for 1000 seconds at different Basic Cycle Lengths (BCLs). The model was validated by extracting data from 19 experimental sources. In our simulations, for BCLs 1000 and 5000, respectively, APD90s (in ms) were 291 and 410.2 (exp.: 230-320 and 338.1-447), maximum upstroke velocities (in V.s-1) were 399.6 and 339.8 (exp.: 275-631 and 294-446), action potential amplitudes (in mV) were 133.7 and 134.1 (exp.: 114.8 -134 and 122-131.9), resting membrane potentials (in mV) were -84.7 and -83.7 (exp.: -92.6 to -82.3 and -87 to -82. Intracellular calcium concentrations matched the values of Aslanidi et al. 2010 (~1µM at BCL 500 ms) and decreased to an 85% and 68% at BCLs 750 ms and 1000 ms (exp.: 85% and 75%, respectively). All parameters lied very close to or within the range of values found in the literature. In conclusion, we developed and successfully validated a novel RPC model. This work paves the way towards a reliable in-silico tool for testing drug effects on RPCs.