Cytotoxicity of pharmaceuticals and their mixtures toward scaffold-free 3D spheroid cultures of rainbow trout (Oncorhynchus mykiss) hepatocytes

Pharmaceutical residues are widely detected in surface waters all around the world, causing a range of adverse effects on environmental species, such as fish. Besides population level effects (mortality, reproduction), pharmaceutical residues can bioaccumulate in fish tissues resulting in organ-specific toxicities. In this study, we developed in vitro 3D culture models for rainbow trout (Oncorhynchus mykiss) liver cell line (RTH-149) and cryopreserved, primary rainbow trout hepatocytes (RTHEP), and compared their spheroid formation and susceptibility to toxic impacts of pharmaceuticals. The rapidly proliferating, immortalized RTH-149 cells were shown to form compact spheroids with uniform morphology in just three days, thus enabling higher throughput toxicity screening compared with the primary cells that required acclimation times of about one week. In addition, we screened the cytotoxicity of a total of fourteen clinically used human pharmaceuticals toward the 3D cultures of both RTH-149 cells (metabolically inactive) and primary RTHEP cells (metabolically active), to evaluate the impacts of the pharmaceuticals' own metabolism on their hepatotoxicity in rainbow trout in vitro. Among the test substances, the azole antifungals (clotrimazole and ketoconazole) were identified as the most cytotoxic, with hepatic metabolism indicatively amplifying their toxicity, followed by fluoxetine, levomepromazine, and sertraline, which were slightly less toxic toward the metabolically active primary cells than RTH-149 spheroids. Besides individual pharmaceuticals, the 3D cultures were challenged with mixtures of the eight most toxic substances, to evaluate if their combined mixture toxicities can be predicted based on individual substances' half-maximal effect (EC50) concentrations. As a result, the classical concentration addition approach was concluded sufficiently accurate for preliminarily informing on the approximate effect concentrations of pharmaceutical mixtures on a cellular level. However, direct read-across from human data was proven challenging and inexplicit for prediction of hepatotoxicity in fish in vitro.