Drug permeability across a phospholipid vesicle based barrier: A novel approach for studying passive diffusion

The aim of this study was to develop a novel predictive medium-throughput screening method for drug permeability, with use of a tight barrier of liposomes on a filter support. To our knowledge no one has succeeded in depositing membrane barriers without the use of an inert solvent such as hexadecane. The first part of the study involved development of a protocol for preparation of these barriers, which were made of liposomes from egg phosphatidylcholin in phosphate buffer pH 7.4 with 10% (v/v) ethanol. The liposomes were deposited into the pores and onto the surface of a filter support (mixed cellulose ester) by use of centrifugation. Solvent evaporation and freeze-thaw cycling were then used to promote fusion of liposomes. A tight barrier could thus be obtained as shown with calcein permeability and electrical resistance. In the second part of the study the model was validated using 21 drug compounds, which cover a wide range of physicochemical properties and absorption (F.) in humans (13-100%). The drug permeation studies were carried out at room temperature with phosphate buffer (pH 7.4) in both acceptor and donor chambers. The apparent permeability coefficients obtained from the phospholipid vesicle based model correlated well with literature data on human absorption in vivo, which suggests that its performance is adequate and that the method is suitable for rapid screening of passive transport of new chemical entities. The results obtained from our model were compared with polar surface area (PSA) and experimental log D and with results obtained by established permeability screening methods such as immobilized liposome chromatography (ILC), the PAMPA models and the Caco-2 model. Our approach seems to model the in vivo absorption better than PSA, experimental log D, the ILC and PAMPA models, when similar conditions are used as in our assay, and equally well as the Caco-2 model and the Double Sink PAMPA (DS-PAMPA) model. (c) 2005 Elsevier B.V. All rights reserved.