Recently, various types of cultured cells have been used to research the mechanisms of transport and metabolism of drugs. Although many studies using cultured cell systems have been published, a comparison of different cultured cell systems has never been reported. In this study, Caco-2, Calu-3, Madin–Darby canine kidney (MDCK), EpiAirway and MucilAir were used as popular in vitro cell culture systems, and the permeability of model compounds across these cell systems was evaluated to compare barrier characteristics and to clarify their usefulness as an estimation system for nasal drug absorption in rats. MDCK unexpectedly showed the best correlation (r = 0.949) with the fractional absorption (F_n) in rats. Secondly, a high correlation was observed in Calu-3 (r = 0.898). Also, Caco-2 (r = 0.787) and MucilAir (r = 0.750) showed a relatively good correlation with F_n. The correlation between F_n and permeability to EpiAirway was the poorest (r = 0.550). Because EpiAirway forms leakier tight junctions than other cell culture systems, the paracellular permeability was likely overestimated with this system. On the other hand, because MDCK formed such tight cellular junctions that compounds of paracellular model were less likely permeated, the paracellular permeability could be underestimated. Calu-3, Caco-2 and MucilAir form suitable cellular junctions and barriers, indicating that those cell systems enable the precise estimation of nasal drug absorption.

The aim of this research is to clarify the influence of the viscosity of the nasal formulation on in vivo nasal drug absorption and its mechanism using an in vitro Caco-2 system. The drug solution was made viscous by the addition of dextran (Dex). The disappearance of FITC-labeled Dextran (FD, a marker of the dosing solution) applied with control solution followed monoexponential kinetics, while FD applied with Dex solution showed biexponential elimination. The mean residence time of FD in the nasal cavity was increased with the increase in Dex concentration. The nasal absorption of acyclovir was similar in the formulation with low viscosity, increased in the formulation with moderate viscosity and markedly decreased in the formulation with high viscosity. The result from the normal Caco-2 transport study could not explain the relation of in vivo drug absorption with viscosity, while the modified Caco-2 system provided data partly reflecting the change in in vivo absorption in rats. In conclusion, the residence of the applied solution in the nasal cavity was enhanced by the addition of Dex in a viscosity-dependent manner. Moderate viscosity of the dosing solution improved the in vivo nasal absorption of acyclovir, while higher viscosity decreased it.