Microfluidic device successfully replaces traditional models of pregnancy associated drug pharmacokinetic studies

Pregnant and lactating people remain therapeutic orphans as they are often excluded from clinical trials, remaining one of the most therapeutically vulnerable. Current models are not suited to determine pharmacokinetics and dynamics. To mediate this problem, this study evaluated drug propagation using a microphysiologic (MPS) model of the human maternal-fetal interface (FMi-PLA-OOC) with simulation software to see if this could serve as an alternative and more efficient approach to studying pharmacokinetics across the placental barrier during pregnancy surpassing the limitations of traditional placental perfusion systems and animal models.

The humanized FMi-PLA-OOC is composed of seven-cell culture chambers, connected by microchannels containing umbilical vein endothelial cells, placental cytotrophoblasts and syncytiotrophoblasts, chorion trophoblast cells, amnion mesenchymal and epithelial cells, and maternal decidua cells. A physiological dose of indomethacin (15 µg/mL) was introduced into the decidual chamber to assess drug pharmacokinetics. Media and supernatants of the FMi-PLA-OOC were collected at various time points and were analyzed using mass spectrometry to determine the drug concentration. A physiologically based pharmacokinetic (PBPK) pregnancy model was developed using GastroPlus and compared with values from placental perfusion studies, animal models, and clinical data from the literature.

Cells in the FMi-PLA-OOC maintained viability, metabolism, and did not show cytotoxicity. The drug propagated and reached the placental layers (umbilical vein cells) in 4 h and the fetal membrane cellular layers (amnion mesenchymal cells) in 1 h. Higher drug concentrations were detected at later time points in the placental cellular layers. Comparing maximum drug concentration ratios in the fetus and mother (C _max,fetal/C _max,maternal) among the different platforms demonstrated: 0.45 (placental perfusion), 0.97 (pregnant humans), 0.774 (FMi-PLA-OOC), 0.68 (PBPK). The fold errors of FMi-PLA-OOC and PBPK for C _max ratios compared to placental perfusion were 0.75 and 0.68, demonstrating an acceptable model.

The utilization of the innovative feto-maternal interface MPS model and PBPK simulation yielded data comparable to the current traditional and simulation approaches. Although not tested here, bedsides kinetics, our humanized MPS model can determine other pharmacologic parameters (efficacy, toxicity, metabolism, absorption, and excretion), making preclinical trials easier, cheaper, and faster and in compliance with the FDA Modernization Act 2.0.