Mimicking brain tissue binding in an in vitro model of the blood-brain barrier illustrates differences between in vitro and in vivo methods for assessing the rate of brain penetration

Assessing the rate of drug delivery to the central nervous system (CNS) in vitro has been used for decades to predict whether CNS drug candidates are likely to attain their pharmacological targets, located within the brain parenchyma, at an effective dose. The predictive value of in vitro blood-brain barrier (BBB) models is therefore frequently assessed by comparing in vitro BBB permeability, usually quoted as the endothelial permeability coefficient (P-e) or apparent permeability (P-app), to their rate of BBB permeation measured in vivo, the latter being commonly assessed in rodents. In collaboration with AstraZeneca (DMPK department, Sodertalje, Sweden), the in vitro BBB permeability (P(app )and P-e) of 27 marketed CNS drugs has been determined using a bovine in vitro BBB model and compared to their in vivo permeability (P-vivo), obtained by rat in-situ brain perfusion. The latter was taken from published data from Summerfield et al. (2007). This comparison confirmed previous reports, showing a strong in vitro/in vivo correlation for hydrophilic compounds, characterized by low brain tissue binding and a weak correlation for lipophilic compounds, characterized by high brain tissue binding. This observation can be explained by the influence of brain tissue binding on the uptake of drugs into the CNS in vivo and the absence of possible brain tissue binding in vitro. The use of glial cells (GC) in the in vitro BBB model to mimic brain tissue binding and the introduction of a new calculation method for in vitro BBB permeability (P-vitro) resulted in a strong correlation between the in vitro and in vivo rate of BBB permeation for the whole set of compounds. These findings might facilitate further in vitro to in vivo extrapolation for CNS drug candidates.