Drug-induced functional cardiotoxicity screening in stem cell-derived human and mouse cardiomyocytes: effects of...

Early prediction of drug-induced functional cardiotoxicity requires robust in-vitro systems suitable for medium/high throughput and easily accessible cardiomyocytes with defined reproducible properties. The xCELLigence Cardio system uses 96-well plates with interdigitated electrodes that detect the impedance changes of rhythmic contractions of stem cell-derived cardiomyocyte (SC-CM) layers. Here, we report on our initial screening experience in comparison to established (multi)cellular and in-vivo models.

Impedance signals from human iPSC-CM (iCells™) and mouse eSC-CM (Cor.At™) were analyzed for contraction amplitude (CA) and duration, rise/fall time, beating rate (BR) and irregularity.

Following solution exchange, impedance signals re-approximated steady-state conditions after about 2 (Cor.At™) and 3h (iCells™); these time points were used to analyze drug effects. The solvent DMSO (≤1%) hardly influenced contraction parameters in Cor.At™, whereas in iCells™ DMSO (>0.1%) reduced CA and enhanced BR. The selective hERG K⁺ channel blockers E-4031 and dofetilide reduced CA and accelerated BR (≥30 nM) according to the analysis software. The latter, however, was due to burst-like contractions (300 nM) that could be detected only by visual inspection of recordings, and were more pronounced in Cor.At™ as in iCells™. In cardiac myocytes and tissue preparations, however, E4031 and dofetilide have been reported to increase cell shortening and contractile force and to reduce BR. Compounds (pentamidine, HMR1556, ATX2, TTX, and verapamil) with other mechanisms of action were also investigated; their effects differed partially between cell lines (e.g. TTX) and compared to established (multi)cellular models (e.g. HMR1556, ouabain).

Mouse and human stem cell-derived cardiomyocytes respond differently to drugs and these responses occasionally also differ from those originating from established in-vitro and in-vivo models. Hence, drug-induced cardiotoxic effects may be detected with this system, however, the predictive or even translational value of results is considered limited and not yet firmly established.