Accurate prediction of human pharmacokinetics for drugs remains challenging, especially for non-cytochrome P450 (P450) substrates. Hepatocytes might be suitable for predicting hepatic intrinsic clearance (CLint) of new chemical entities, because they can be applied to various compounds regardless of the metabolic enzymes. However, it was reported that hepatic CLint is underestimated in hepatocytes. The purpose of the present study was to confirm the predictability of human hepatic clearance for P450 and non-P450 substrates in hepatocytes and the utility of animal scaling factors for the prediction using hepatocytes. CLint values for 30 substrates of P450, UDP-glucuronosyltransferase, flavin-containing monooxygenase, esterases, reductases, and aldehyde oxidase in human microsomes, human S9 and human, rat, and monkey hepatocytes were estimated. Hepatocytes were incubated in serum of each species. Furthermore, CLint values in human hepatocytes were corrected with empirical, monkey, and rat scaling factors. CLint values in hepatocytes for most compounds were underestimated compared to observed values regardless of the metabolic enzyme, and the predictability was improved by using the scaling factors. The prediction using human hepatocytes corrected with monkey scaling factor showed the highest predictability for both P450 and non-P450 substrates among the predictions using liver microsomes, liver S9, and hepatocytes with or without scaling factors. CLint values by this method for 80% and 90% of all compounds were within 2- and 3-fold of observed values, respectively. This method is accurate and useful for estimating new chemical entities, with no need to care about cofactors, localization of metabolic enzymes, or protein binding in plasma and incubation mixture.
Keywords: P450; hepatic clearance; in vitro-in vivo extrapolation; monkey; non-P450.