A useful model capable of predicting the clearance of cytochrome 3A4 (CYP3A4) substrates in humans: validity of CYP3A4 transgenic mice lacking their own Cyp3a enzymes

The accurate prediction for the body clearance of a novel drug candidate by humans during the preclinical stage contributes to its successful development. To improve the predictability of human hepatic clearance, we focused on CYP3A4, which is involved in the metabolism of more than 50% of all currently marketed drugs. In this study, we investigated the validity of the in vivo model using transgenic mice carrying the human CYP3A4 gene and lacking their own Cyp3a genes (CYP3A4-Tg mice). The CYP3A4 activity toward its substrates in liver microsomes was similar in CYP3A4-Tg mice and humans. As for the clearance, six CYP3A4 substrates (alprazolam, felodipine, midazolam, nifedipine, nitrendipine, and quinidine) were given intravenously to CYP3A4-Tg mice, and their hepatic intrinsic clearance (CLint,h) was evaluated. A regression analysis of the data obtained indicated that the CLint,h values of six substrates in CYP3A4-Tg mice were highly correlated with those in humans (R(2) = 0.95). This correlation could be improved by correcting the CLint,h values by the relative contribution of artificially expressed CYP3A4 to the overall metabolism in the mice. From these findings, it is reasonable to expect that the CLint,h of a particular drug in humans is predictable by applying the CLint,h obtained in CYP3A4-Tg mice to a regression line prepared in advance. The variance of the CLint,h prediction by this method was evaluated and found to be within a range of 2-fold of the regression value. These results suggest that the CYP3A4-Tg mouse model has the potential to accurately predict the human hepatic clearance of CYP3A4 substrates.