Integrating transcriptomics, quantitative adverse outcome pathways, in vitro and in vivo kinetics to predict amiodarone toxicity

Amiodarone is a highly lipophilic drug used to treat arrhythmia. It is known to induce repeat-dose toxicity in the liver and the brain of patients. As such, it is a suitable chemical to assess whether and how new approach methodologies (NAMs) can be used to identify and quantify the health hazards associated with its exposure. In a first study, amiodarone was exposed to the human hepatoma cell line, HepaRG, to assess its potency to perturb key events along the adverse outcome pathway (AOP) for steatosis. Key events included steatotic gene expression changes and triglyceride accumulation assessed using high content imaging. Toxicokinetic-toxicodynamic (TK-TD) modelling was used to construct a quantitative AOP for steatosis. In a second study, we exposed human induced pluripotent stem cells (hiPSCs)-derived BrainSpheres to amiodarone daily for 7 days. Our transcriptomics analysis revealed that amiodarone altered lipid metabolism gene expression at lower concentrations than in the hepatocytes, once differences in in vitro distribution kinetics between the two cell models were accounted for. In the third study, a probabilistic physiologically based kinetic (PBK) model for amiodarone was constructed, parameterized using in vitro and in silico-derived kinetic data, evaluated against rat experimental data in literature, and used to extrapolate in vitro hepatotoxic and neurotoxic concentrations to human bioequivalent oral doses, an approach referred to as quantitative in vitro-in vivo extrapolation (QIVIVE). The model predicted significantly higher liver than brain levels of amiodarone after repeat dosing and thus a greater risk for hepatotoxicity compared to neurotoxicity.