The objective was to compare several in vitro human liver-derived subcellular and cellular incubation systems for the formation of GSH-trapped reactive metabolites. Incubations of pooled human liver microsomes, human liver S9 fractions, HepaRG-cells, and human hepatocytes were performed with glutathione as a trapping agent. Experiments with liver S9 were performed under two conditions, using only NADPH and using a full set of cofactors enabling also conjugative metabolism. Ten structurally different compounds were used as a test set, chosen as either "positive" (ciprofloxacin, clozapine, diclofenac, ethinyl estradiol, pulegone, and ticlopidine) or "negative" (caffeine, citalopram, losartan, montelukast) compounds, based on their known adverse reactions on liver or bone marrow. GSH conjugates were observed for seven of the ten compounds; while no conjugates were observed for caffeine, citalopram, or ciprofloxacin. Hepatocyte and HepaRG assays produced a clearly lower number and lower relative abundance of GSH conjugates compared to assays with microsomes and S9 fractions. The major GSH conjugates were different for many compounds in cellular subfractions and cell-based systems. Hepatocytes generally produced a higher number of GSH conjugates than HepaRG cells, although the differences were minor. The results show that the hepatic enzyme system used for screening of GSH-trapped reactive metabolites do have a high impact on the results, and results between different systems are comparable only qualitatively.
Copyright © 2015 Elsevier B.V. All rights reserved.