Predicting Maternal-Fetal Disposition of Drugs Using In Vitro and In Silico Tools
MetadataShow full item record
During pregnancy, physiological and ADMET changes in the maternal-fetal dyad can significantly alter drug pharmacokinetics (PK) in the mother and may necessitate dosing regimen adjustments. One clinically important example is CYP3A. Its activity increases by ~2-fold during the third trimester (T3) as shown by 1’-hydroxymidaozlam in vivo clearance, resulting in sub-therapeutic plasma concentrations of CYP3A substrates (Unadkat et al., 2007; Hebert et al., 2008). However, it is logistically challenging and/or ethically not feasible to conduct such PK studies in the pregnant mother, especially during first and second trimesters (T1 and T2, respectively). Therefore, the magnitude of hepatic CYP3A induction in vivo across different trimesters was predicted using two model systems, namely, the HepaRG cells and human hepatocytes. After incubation with individual or combinations of pregnancy-related hormones at their clinically observed plasma concentrations in different trimesters, the change in the activity of hepatic CYP3A was measured. Both in vitro systems predicted ~2-fold induction in CYP3A activity throughout pregnancy based on the above T3 induction data. This predicted 2-fold increase in CYP3A activity during T1 and T2 is in accordance with T2 pharmacokinetic (PK) data of several CYP3A substrates (e.g. buprenorphine(Bastian et al., 2016), nifedipine,(Marin et al., 2007), and indinavir (Cressey et al., 2013)). Assessing fetal exposure to drugs during pregnancy remains a major challenge as fetal exposure to drugs is not only driven by maternal PK but also may be further influenced by placental passage and fetoplacental metabolism/transport. Additionally, due to safety and ethical concenrns, fetal blood sampling is not possible until the time of birth. Thus mechanistic, non-invasive in silico prediction methods, such as physiologically-based pharmacokinetic (PBPK) models, offer an alternative to predicting the in vivo disposition of drugs in the maternal-fetal pair during pregnancy. However, at present, the fetus is commonly treated as a single compartment in existing maternal-fetal PBPK models developed for pharmaceutical agents (Gaohua et al., 2012), often lumped with the placenta and the amniotic cavity. As such, model verification using term umbilical venous data, the only fetal exposure data available, is not possible. To bridge this gap in knowledge, we constructed a novel MATLAB® maternal-fetal (m-f) PBPK model in collaboration with Simcyp®, by expanding our previously verified pregnancy PBPK model. The expansion was accomplished by incorporating a fetal PBPK consisting of 7 body compartments into the pregnancy PBPK model, as well as the placenta and the amniotic fluid compartment. Using a series of simulation exercise using this m-f-PBPK model, we delineated the key factors determining fetal plasma drug exposure and clarified the commonly held misconceptions surrounding UV:MP ratio data. We also predicted that fetal development may play an important role in modulating fetal plasma AUC. Finally, to verify, we verified the performance of this m-f-PBPK model with drugs that cross the placenta predominantly via passive diffusion by simultaneously predicting maternal and fetal disposition of two model drugs. The model performed reasonably well predicting the observed maternal-fetal plasma drug concentration-time (C-T) profiles of both drugs at the time of delivery. The aims of this thesis were to test whether mechanistic physiologically-based pharmacokinetic (PBPK) models integrating in vivo observations and in vitro experimental results could be utilized to predict maternal and fetal drug exposure to several probe drugs during pregnancy. The results obtained demonstrate that: 1) the magnitude of in vivo hepatic CYP3A induction during T1 and T2 can be predicted from relevant in vitro systems such as HepaRG cells and hepatocytes; 2) PBPK modeling and simulation can be used as theoretical tools to discern key factors determining fetal exposure to drugs; and 3) the presented m-f-PBPK model can be used to predict maternal and fetal plasma drug C-T profiles of passive diffusion drugs. The predicted 2-fold increase in CYP3A activity will be used to inform dose adjustments of CYP3A substrates during T1 and T2, when pregnant women are normally precluded from PK studies for safety concerns. The presented m-f-PBPK model provides one step further towards a more physiologically faithful platform for quantitative prediction of drug PK in the coupled maternal-fetal pair throughout pregnancy, and allows generalization beyond the probe drugs studied when fetoplacental expression data become available.
- Pharmaceutics