In Vitro-In Vivo Correlation of Blood-Brain Barrier Permeability of Drugs: A Feasibility Study Towards Development of Prediction Methods for Brain Drug Concentration in Humans

Purpose: In vitro human blood-brain barrier (BBB) models in combination with central nervous system-physiologically based pharmacokinetic (CNS-PBPK) modeling, hereafter referred to as the "BBB/PBPK" method, are expected to contribute to prediction of brain drug concentration profiles in humans. As part of our ongoing effort to develop a BBB/PBPK method, we tried to clarify the relationship of in vivo BBB permeability data to those in vitro obtained from a human immortalized cell-based tri-culture BBB model (hiBBB), which we have recently created.

Methods: The hiBBB models were developed and functionally characterized as previously described. The in vitro BBB permeabilities (Pe, × 10^-6 cm/s) of seventeen compounds were determined by permeability assays, and in vivo BBB permeabilities (Q_ECF) for eight drugs were estimated by CNS-PBPK modeling. The correlation of the Pe values with the Q_ECF values was analyzed by linear regression analysis.

Results: The hiBBB models showed intercellular barrier properties and several BBB transporter functions, which were enough to provide a wide dynamic range of Pe values from 5.7 ± 0.7 (rhodamine 123) to 2580.4 ± 781.9 (rivastigmine). Furthermore, the in vitro Pe values of the eight drugs showed a good correlation (R² = 0.96) with their in vivo Q_ECF values estimated from human clinical data.

Conclusion: We show that in vitro human BBB models provide clinically relevant BBB permeability that can be used as input for CNS-PBPK modeling. Therefore, our findings will encourage the development of a BBB/PBPK method as a promising approach for predicting brain drug concentration profiles in humans.