A simple resazurin-based cytotoxicity assay is presented for screening of cytotoxicity in hepatocytes and liver cell lines. Human hepatoma (HepG2) cells in 96-well culture plates were exposed to known toxic (cisplatin, 5-fluorouracil, ethionine, flufenamic acid, and diflunisal) and control (transplatin, 5-chlorouracil, methionine, and acetylsalicylic acid) compounds for 1-3 days, and resazurin (5 micromol/L) was added. A conventional short-term (1 h) assay was first performed, where cytotoxicity is indicated by decreased reduction of resazurin to its fluorescent product resorufin. Our improved assay consists of additionally measuring fluorescence 2-4 days later, when cytotoxicity is indicated by a striking increase in the concentration of resorufin, resulting from two distinct processes. First, viable liver-derived cells slowly convert resorufin to nonfluorescent metabolites. Fluorescence of control cell wells decreased to background during a 2- to 4-day exposure to resazurin. This metabolism of resorufin was largely blocked by dicumarol and to lesser extents by disulfiram and SKF525a. Second, dead or dying cells slowly convert resazurin to resorufin but do not further metabolize resorufin; thus this fluorescent metabolite accumulates to high levels in wells with dead cells by 2 to 4 days. A similar increase in fluorescence associated with cytotoxicity was observed in primary cultures of rat hepatocytes using the long-term resazurin-based assay. In addition to an improved signal relative to the short-term assay, the inversion of the fluorescent signal from high = alive short-term to high = dead long-term allows determination of two independent cytotoxicity endpoints after addition of one innocuous vital dye.

Abstract:In vitro determination of metabolic stability is routinely used to assess the overall metabolic liability of compounds and for prioritization for in vivo studies. If in vitro metabolic stability data could be used to reliably predict in vivo clearance (CL), it would add significant value in the selection of compounds for in vivo pharmacokinetic and pharmacology studies. We have evaluated the utility of our in vitro metabolic stability screening assay to estimate in vivo CL in the mouse. The in vitro mouse clearances (CL in vitro ) of 146 structurally diverse compounds with metabolic stabilities > 30 %, were compared to mouse in vivo CL data. Approximately 45 % of the compounds showed agreement between in vivo CL and predicted CL in vitro within a 2-fold error criteria. The correlation appeared worse when correction for the extent of incorporation of plasma protein binding or both plasma and S9 bindings (i.e. ~14 % and~ 28 % agreement, respectively). Classification of the compounds into three groups based on in vivo CL (<30 mL/min/kg, 30-70 mL/min/kg, and >70 mL/min/kg) did not show any improvement between in vivo CL and predicted CL in vitro . The percentage of compounds falling within the 2-fold error criteria for low CL, moderate CL and high CL groups were 54, 31 and 24 %, respectively. In conclusion, our analysis suggests that in vitro metabolic stability data, as routinely obtained in early ADME screening protocols, does not demonstrate a strong correlation with or predictivity for, absolute in vivo CL in the mouse.