Although significant progress has been made in recent years in the evaluation of drug-drug interaction potential based on in vitro data, a complete understanding of the relationship between in vitro findings and in vivo human results of metabolism-based drug-drug interaction studies is still emerging. In some cases, excellent correlation of in vitro and in vivo results has been demonstrated while in others, the in vitro and in vivo correlation has been poor . Because of the complexities of various factors impacting both in vitro and in vivo drug-drug interactions, accurate predictions of the extent of in vivo drug interactions from in vitro metabolic studies will require continued efforts in obtaining additional high quality correlation data to permit rational evaluation of new drugs. At the present time, the feasibility of predicting in vivo drug interactions based on in vitro metabolic data is still under rigorous debate. Some investigators believe that a quantitative prediction of in vivo drug interaction is possible [16-18] while others take the position that a qualitative prediction approach is more feasible [19, 20]. In a recent commentary, Tucker et al.  used the qualitative terms "low risk, medium risk, and high risk" to describe the projection of AUC changes based on the [I]/Ki ratio, where the Ki values are determined from in vitro studies.
Various factors contributing to the difficulty in predicting if a new molecular entity (NME) is an inhibitor from in vitro data. Among them, the unusual cytochrome P450 property and the large number of drug substrates appear to be critical factors. In vitro drug-drug interaction patterns (e.g., mutual inhibition, partial inhibition, activation, and lack of reciprocal inhibition) for a given cytochrome P450, such as CYP3A4, are often substrate-dependent. The Ki value of an inhibitor for a given cytochrome P450 is dependent on the probe substrates, enzyme sources, and experimental conditions such as protein concentration and incubation time due to various degrees of inhibitor-protein binding, partition of inhibitor to the lipid and aqueous layers, and inhibitor and substrate depletion.
One of the challenges in predicting the extent of in vivo drug-drug interaction from in vitro metabolism studies is the lack of information on the inhibitor concentration in vivo in the active site of the enzyme or tissues. Since the plasma inhibitor concentration may be the only known parameter, both total inhibitor concentration and unbound inhibitor concentration have been used for in vitro-in vivo correlation evaluation. Claims of good correlation with either of the parameters have been reported for different drugs. Other factors contributing to the lack of good in vitro-in vivo correlation using either of the parameters may include the following: (1) the inhibiting drug may also act as an inducer; (2) other parallel elimination pathways and/or extrahepatic metabolism of the drug may decrease the importance of the in vitro-assessed pathway; (3) modulation of an important cellular transport mechanism by the inhibitor may change the extent of in vivo drug-drug interaction, and (4) rapid elimination of inhibitor in vivo by noncytochrome P450 pathways may decrease the extent of in vivo drug-drug interaction.
Was this article helpful?