2.4. In vitro-in vivo correlation of different methods used as models of placental drug transfer and drug metabolism

In order to develop safe and effective drugs, it is important to know the exact pharmacokinetics as early as possible. In addition to species-related differences and ethical concerns, animal testing is expensive. Due to this, the development of in vitro model systems to predict the pharmacokinetics of drugs has become increasingly important (Davila et al. 1998, Ito et al. 1998b).

Prediction of the metabolite profile and the rate of metabolism are important aspects of in vitro-in vivo correlations (Lin 1998, Ito et al. 1998a). Prediction of drug interactions is also important. Currently, several methods using preclinical pharmacokinetic data and in vitro human metabolism data have been found useful in the prediction of these human pharmacokinetic parameters (Obach et al. 1997, Ito et al. 1998a, Ito et al. 1998b). Generally, data obtained with human primary hepatocytes and human liver slices have correlated well with the existing in vivo data (Li et al. 1997, Pelkonen et al. 2002). The qualitative metabolite profile obtained in vitro usually reflects quite accurately the in vivo metabolic pattern (Lin 1998). The extrapolation of in vitro data to in vivo conditions is not without problems, however. In the prediction of interactions, both the identity of the CYP isoform responsible for metabolism and the relative contribution of the metabolic pathway to overall elimination must be considered (Lin & Lu 1998). Several factors, including drug, inhibitor and protein concentrations and metabolic geno- and phenotypes, must be taken into account when assessing the clinical signifigance of findings (Lin & Lu 1998, Yuan et al. 1999). The prediction of interactions is even more complicated when multiple CYP enzymes take part in the metabolism (Obach et al. 1997). Also, all in vitro methods have their limitations. For instance, the enzymatic activities in both human liver slices and hepatocytes tend to decrease during incubations (Guillouzo et al. 1999, Renwick et al. 2000, Pelkonen et al. 2002). Other pharmacokinetic factors often also affect the in vivo metabolism. For instance, hepatic blood flow limits the metabolism of some rapidly metabolized compounds, and due to this, in vitro metabolism may show larger interindividual variation than is observed in vivo (Kedderis 1997).

In contrast to the in vitro models for metabolism, only a few efforts have been made to compare in vitro and in vivo data on placental drug transfer. The comparison of in vitro data to clinical data is often difficult due to the lack of clinical data. The few examples from the literature suggest a similar pharmacokinetic profile in vivo and in the placental perfusion model (Omarini et al. 1992, Tuntland et al. 1999, Ala-Kokko et al. 2000). Also, some studies have compared data gained with the placental perfusion method with other models. Tuntland and co-workers compared several methods in the prediction of the mechanism, rate and extent of placental transfer of dideoxynucleoside drugs (Tuntland et al. 1999). Placental transfer of these drugs was found to be similar in human placental perfusion and in vivo in the pregnant macaque model. Dicke and co-workers (Dicke et al. 1988) compared placental transfer of four H₂-receptor antagonists in perfused human and baboon placentas and found no significant differences.