Here, we established a high-throughput in vitro assay system to predict reactive metabolite (RM) formation. First, we performed the glutathione (GSH) consumption assay to monitor GSH levels as an index of RM formation potential using HepaRG cells pretreated with 500 micro M D,L-buthionine-(S,R)-sulfoximine (BSO) and then treated with ticlopidine and diclofenac. Both drugs, under GSH-reduced conditions, significantly decreased relative cellular GSH content by 70% and 34%, respectively, compared with that in cells not pretreated with BSO. Next, we examined the correlation between GSH consumption and covalent binding assays; the results showed good correlation (correlation coefficient = 0.818). We then optimized the test compound concentration for evaluating RM formation potential using 76 validation compound sets, and the highest sensitivity (53%) was observed at 100 micro M. Finally, using HepG2 cells, PXB-cells, and human primary hepatocytes, we examined the cell types suitable for evaluating RM formation potential. The expression of CYP3A4 was highest in HepaRG cells, suggesting the highest sensitivity (56.4%) of the GSH consumption assay. Moreover, a co-culture model of PXB-cells and HepaRG cells showed high sensitivity (72.7%) with sufficient specificity (85.7%). Thus, the GSH consumption assay can be used to effectively evaluate RM formation potential in the early stages of drug discovery.
