Precision-cut intestinal slices as an in vitro model to predict NSAID induced intestinal toxicity

Abstract

Non-steroidal anti-inflammatory drugs (NSAIDs) are associated with high prevalence of gastro-intestinal side-effects. In vivo studies suggest that uncoupling of oxidative phosphorylation is an important cause of the toxicity and that the toxicity is aggravated by enterohepatic circulation. Precision-cut intestinal slices (PCIS) are a newly established in vitro model with intact architecture of the intestine, preserved viability and metabolism during incubation up to 24 h. In this study, this model is used to predict NSAID-induced intestinal toxicity. PCIS from rat jejunum were incubated for 5 h with a concentration range of 5 NSAIDs: diflunisal (0.05-1 mM), indomethacin (0.05-1 mM), diclofenac (0.05-1 mM), naproxen (0.1-5mM) and aspirin (0.1-10 mM). Toxicity was assessed by the ATP content. In vitro TC50 of these 5 compounds was calculated and compared with in vivo TC50 and the mitochondrial toxicity data as reported in the literature. Based on TC50 values, toxicity ranking in the present study appeared to be: aspirin (>10mM)> naproxen (2.96mM)> diclofenac (0.29mM)> indomethacin (0.29mM)> diflunisal (0.18 mM). This is very well in line with the reported toxicity ranking found with the isolated mitochondria. When compared with the in vivo TC50, only one exception of diflunisal was found, which appeared highly toxic in vitro whereas in vivo it's one of the least toxic compound. This discrepancy could not yet be explained by in vivo toxicokinetic factors (absorbtion, enterohepatic circulation, metabolism, excretion). In conclusion, both intrinsic (mitochondrial) toxicity as well as in vivo toxicity ranking are well predicted by the PCIS model.