Comparison of Intestinal Permeabilities Determined in Multiple in Vitro and inSitu Models: Relationship to Absorption in Humans

In vitro and in situ experimental models that are descriptive of drug absorption in vivo are valuable tools in the discovery of new chemical entities that are bioavailable after oral administration. The specific objective of the study was to compare the intestinal permeabilities obtained in the three absorption models for consistency, and to assess the utility of the models in predicting the fraction of dose absorbed in human studies. The intestinal absorption models that were compared are widely used: the rat in situ single-pass intestinal perfusion system, the rat everted intestinal ring method, and monolayers of human colon adenocarcinoma cell line (CACO-2). The models were compared using small molecular reference compounds, as well as a series of peptidomimetic (PM) analogs. Each model had strong potential for estimating the fraction absorbed. For small organic molecules, excellent correlation was observed when permeabilities from CACO-2 cells and perfusions, or everted rings and perfusions, were compared. Weaker correlation was observed between everted rings and CACO-2 cells. Permeabilities for the set of reference compounds and PMs were positively correlated between any two of the three systems. Variance between correlations for reference compounds and PMs are likely due to structural features and physicochemical properties that are unique to the latter class of compounds. The results support caution in extrapolating correlations based on findings with small organic molecules to the behavior of complex peptidomimetics. Corroboration of permeabilities with two methods of determination is a useful cross-validation of experimental systems, as well as producing a reliable permeability assessment. CACO-2 cell monolayers and rat single-pass intestinal perfusion combine the highest correlation between systems, most defined relationship with fraction absorbed in humans, and experimental logistics in-line with discovery candidates.