To closer mimic the steric conditions in the intestine in vivo, Caco-2 cells are cultured on permeable filter inserts that can be obtained from a number of vendors like Becton Dickenson, Corning, Costar, etc. Cultivation of Caco-2 cells on filter supports improves their morphological and functional differentiation. It has been well documented that polarized Caco-2 monolayers represent a reliable correlate for studies on the absorption of drugs and other compounds after oral intake in humans. Several studies have compared Caco-2 permeability coefficients with absorption data in humans and found high correlation, particularly if the compounds are transported by passive paracellular transport mechanisms (Artursson and Karlsson 1991; Cheng et al. 2008; Sun et al. 2008).
Although Caco-2 cells have been found to express a large number of enzymes and transporter proteins present in normal human intestinal epithelium, recent studies suggest that there are variations between gene expression profiles of transformed epithelial cell lines like Caco-2, HT29 and normal human intestinal epithelium (Bourgine et al. 2011). Furthermore, differences are not only found between the cell lines and normal epithelium, principle component analysis of gene expression profiles reveals clear differences between the Caco-2 and HT29 cell lines, also (Bourgine et al. 2011). This is not unexpected. Normal intestinal epithelium is made up of several different cell types, and differences in gene expression profiles are not only observed in the mucosal epithelium along the whole gastrointestinal tract, but also along the crypt-villus axis (Anderle et al. 2005). Obviously, data analysis of experiments in the Caco-2 cell model might not be directly compared with the in vivo situation.
Still, intestinal epithelial cell models like the Caco-2 model, holds many advantages due to its simplicity and reproducibility allowing inter-laboratory comparison of results. Furthermore, pursuing an effect of a food bioactive in a cell line model opens for studies of molecular mechanisms which may be more difficult to address in vivo.