Paracellular and passive transcellular permeability in immortalized human corneal epithelial cell culture model

A cell culture model of human corneal epithelium (HCE-model) was recently introduced [Invest. Ophthalmol.Vis. Sci. 42 (2001) 2942] as a tool for ocular drug permeation studies. In this study, passive permeability and esterase activity of the HCE-model were characterised. Immortalised human corneal epithelial cells were grown on collagen coated filters under air-lift. The sensitivity of transcellular permeability to lipophilicity was tested in studies using nine beta-blockers. The size selectivity of the paracellular route was investigated using 16 polyethylene glycol oligomers (PEG). An effusion-like approach was used to estimate porosity and pore sizes of the paracellular space in HCE membrane. Permeability and degradation of fluorescein diacetate to fluorescein in HCE-cells was used to probe the esterase activity of the HCE-model. Drug concentrations were analyzed using HPLC (beta-blockers), LC-MS (PEGs), and fluorometry (fluorescein). Permeabilities were compared to those in the excised rabbit cornea. Penetration of beta-blockers increased with lipophilicity according to a sigmoidal relationship. This was almost similar to the profile in excised cornea. No apical to basolateral directionality was seen in the permeation of beta-blockers. Paracellular permeability of the HCE-model was generally slightly higher than that of the excised rabbit cornea. The HCE-model has larger paracellular pores, but lower pore density than the excised cornea, but the overall paracellular space was fairly similar in both models. The HCE-model shows significant esterase activity (i.e. fluorescein diacetate was converted to free fluorescein). These data on permeability of 27 compounds demonstrate that the barrier of the HCE-model closely resembles that of the excised rabbit cornea. Therefore, the HCE-model is a promising alternative corneal substitute for ocular drug delivery studies. (C) 2003 Elsevier B.V. All fights reserved.