Mechanistic Coupling of a Novel in silico Cotyledon Perfusion Model and a Physiologically Based Pharmacokinetic Model to Predict Fetal Acetaminophen Pharmacokinetics at Delivery

Abstract

Little is known about placental drug transfer and fetal pharmacokinetics despite increasing drug use in pregnant women. While physiologically based pharmacokinetic (PBPK) models can help in some cases to shed light on this knowledge gap, adequate parameterization of placental drug transfer remains challenging. A novel in silico model with seven compartments representing the ex vivo cotyledon perfusion assay was developed and used to describe placental transfer and fetal pharmacokinetics of acetaminophen. Unknown parameters were optimized using observed data. Thereafter, values of relevant model parameters were copied to a maternal-fetal PBPK model and acetaminophen pharmacokinetics were predicted at delivery after oral administration of 1,000 mg. Predictions in the umbilical vein were evaluated with data from two clinical studies. Simulations from the in silico cotyledon perfusion model indicated that acetaminophen accumulates in the trophoblasts; simulated steady state concentrations in the trophoblasts were 4.31-fold higher than those in the perfusate. The whole-body PBPK model predicted umbilical vein concentrations with a mean prediction error of 24.7%. Of the 62 concentration values reported in the clinical studies, 50 values (81%) were predicted within a 2-fold error range. In conclusion, this study presents a novel in silico cotyledon perfusion model that is structurally congruent with the placenta implemented in our maternal-fetal PBPK model. This allows transferring parameters from the former model into our PBPK model for mechanistically exploring whole-body pharmacokinetics and concentration-effect relationships in the placental tissue. Further studies should investigate acetaminophen accumulation and metabolism in the placenta as the former might potentially affect placental prostaglandin synthesis and subsequent fetal exposure.

Keywords: acetaminophen; ex vivo cotyledon perfusion; maternal-fetal; physiologically-based pharmacokinetics; placental transfer; pregnancy.

Figure 1

Structure of the maternal-fetal PBPK model. Gray boxes represent compartments of the physiologically based pharmacokinetic (PBPK) model; solid arrows denote drug transport via the organ blood flow; dashed arrows denote drug transport via passive diffusion; dash dotted lines denote drug transport via gastrointestinal motility or the biliary excretion route.

Figure 2

Structure of the placenta sub-model integrated in the maternal-fetal PBPK model. Gray boxes represent sub-compartments of the placenta structure implemented in the maternal-fetal physiologically based pharmacokinetic (PBPK) model; dash-dotted boxes represent the vascular space; solid arrows denote drug transport via the organ blood flow; and dashed arrows denote drug transport via passive diffusion. The maternal plasma and blood cell compartments represent the intervillous space, and the maternal interstitial and intracellular space represent the placental septae and the decidua basalis, respectively. The fetal intracellular compartment represents the (syncytio-)trophoblasts with the apical membrane facing the maternal plasma compartment and the basolateral membrane facing the fetal interstitial compartment. The fetal interstitial space represents intravillous fibrous tissue and the plasma and blood cell compartments represent the intravillous vascular system.

Figure 3

Structure of the ex vivo cotyledon perfusion model. Boxes represent compartments of the novel ex vivo cotyledon perfusion model and solid arrows denote drug transport via the perfusate flow or diffusion. The cotyledon is highlighted as dashed box.

Figure 4

Observed and simulated concentration time profiles in the novel ex vivo cotyledon perfusion model. In each experiment, acetaminophen was either administered to the fetal reservoir (A–D) or maternal reservoir (E–N) at time = 0h. The initial concentration was 10 mg/L in all experiments (to reflect clinically relevant concentrations); each panel refers to an individual experiment. All experiments were conducted under similar conditions. Observed data were taken from Conings et al. (20).

Figure 5

Local sensitivity analysis for the placental partition coefficients (K_{FM_cell : perf} and K_{F_cell : perf}). In each experiment, acetaminophen was either administered to the fetal reservoir (A–D) or maternal reservoir (E–N) at time = 0 h. The initial concentration was 10 mg/L in all experiments (to reflect clinically relevant concentrations); each panel refers to an individual experiment. All experiments were conducted under similar conditions. Observed data were taken from Conings et al. (20).

Figure 6

Concentration time profiles for acetaminophen in the umbilical cord. Blue and red circles indicate observed clinical data reported by Nitsche et al. (17) and Mehraban et al. (18). Data from Mehraban et al. (18) below the lower limit of quantification (LLOQ), shown as dashed line, are included as LLOQ/2 and empty circles in the figure. The black line indicates the predicted median and the shaded area the predicted 5–95th percentile range.

Figure 7

Goodness-of-fit plot (A) and residuals versus time plot (B) for the predicted acetaminophen concentrations in the umbilical vein. Blue and red circles indicate clinical data reported by Nitsche et al. (17) and Mehraban et al. (18). Data from Mehraban et al. (18) below the lower limit of quantification (LLOQ) are included as LLOQ/2 and empty circles in the figure. In (A), the solid line indicates the line of identity and the dashed lines limit the 2-fold error range. In (B), the solid line indicates the level where the residuals are zero.