Comparison of in vitro and in vivo models of drug transcytosis through the blood-brain barrier.

Drug and solute transport through in vitro and in vivo models of the blood-brain barrier (BBB) were compared to provide a measure of how well the in vitro model predicted BBB permeability found in vivo. The in vitro model employed bovine brain capillary endothelial cells in either primary tissue culture or as a continuous line grown on Transwells and placed in side-by-side diffusion chambers. The in vivo model of BBB transport utilized an internal carotid artery perfusion/capillary depletion method in anesthetized rats. BBB permeability in vivo and in vitro was measured for 15 radiolabeled drugs and for L-[3H]dopa, D-[14C]glucose and [3H]albumin. [3H]- or [14C]sucrose was used in vivo as a blood volume reference. Lipid solubility of each drug was measured based on the 1-octanol/Ringer's partition coefficient. The morphology of the endothelial cell in primary tissue culture was spindle-shaped and the morphology of the endothelial cell in continuous culture was cuboid-shaped. The cuboidal morphology demonstrated a 2-fold greater resistance to solute transport and was used for the majority of the in vitro studies. Drug and solute permeability coefficients (Pe) ranged from 3.9 X 10(-3) to 2.5 X 10(-1) cm/min in vitro and from 1.0 X 10(-5) to 2.1 X 10(-2) cm/min in vivo. The In of the permeability.surface area product in vitro correlated with the In partition coefficient (r = 0.62, P less than .0125) and the In permeability.surface area product in vivo correlated with the In partition coefficient (r = 0.84, P less than .0005).(ABSTRACT TRUNCATED AT 250 WORDS)