Abstract

Human liver microsomes (HLMs) are frequently utilized in drug discovery to predict the human clearance of a compound. The extent to which the incubation conditions affect the accuracy of a human clearance prediction was determined for bufuralol. HLMs were preincubated at 37°C for varying times (5–120 min) with and without NADPH, and the remaining enzyme activity was determined by incubating compounds that have been characterized to be selective for individual cytochromes P450 or flavin-containing monooxygenase 3. CYP2D6, the high-affinity component of bufuralol metabolism, was shown to be the least stable of the isoforms studied. The loss of CYP2D6 activity was further examined by determining the kinetics of 1′-hydroxybufuralol formation after different preincubation time periods, by using reactive oxygen species (ROS) scavengers, and by utilizing Western blotting techniques. A 3-fold decrease in V_max was observed over 2 h, whereas the K_m remained constant. ROS scavengers were able to block enzyme lability, and Western blots revealed no apparent loss of immunoreactive enzyme. The protein binding of bufuralol was determined in HLMs, recombinant CYP2D6, and human plasma. A prediction of theoretical bufuralol concentrations over a 120-min incubation that incorporated enzyme lability was performed and shown to be closer to actual data than if enzyme lability were ignored. Finally, a similar prediction using literature bufuralol data, coupled with the observed protein binding data, was used to illustrate that the most accurate predictions of bufuralol clearance are obtained when the amount of protein in the incubation is kept to a minimum and the overall incubation time is less than 20 min.