Optimization of a Parallel Artificial Membrane Permeability Assay for the Fast and Simultaneous Prediction of Human Intestinal Absorption and Plasma Protein Binding of Drug Candidates: Application to Dibenz[b,f]azepine-5-Carboxamide Derivatives

Abstract

Parallel artificial membrane permeability assay (PAMPA) has been successfully applied by pharmaceutical industries as high-throughput technique capable of screening compounds for passive oral absorption. Herein, the possible applicability of this assay to predict simultaneously biodistribution parameters such as plasma protein binding (PPB) and apparent volume of distribution (VD) was investigated for the first time. Apparent permeability (P(app) ) of 18 reference drugs was determined by nine PAMPA conditions and compared with the corresponding intestinal absorption fraction (Fa), PPB, and VD in humans. In all the models, P(app) was not correlated with VD; however, it was correlated with Fa and PPB. In these cases, the best correlations (r ≥ 0.8953) were found when using a membrane of 2% of l-α-phosphatidylcholine, at pH 6.5/7.4 for donor/acceptor compartments. Under these conditions, good correlation with traditional PAMPA (r = 0.9633) and Caco-2 models (r = 0.9246) were also achieved. This method correctly predicted compounds' permeability and was robust enough to distinguish compounds with high Fa (P(app) ≥ 1.1 × 10(-6) cm/s) and PPB (P(app) ≥ 1.0 × 10(-5) cm/s), with no false positives or negatives recorded. In addition, ultrafiltration method was used to determine the PPB of 10 tested derivatives of dibenz[b,f]azepine-5-carboxamide, which were also assessed through the new PAMPA model developed herein, and equal classifications were achieved.