Morphological data presented in the present study indicated that primary cultures of bovine enterocytes isolated from colon and jejunum presented characteristics of epithelial cells, such as a typical pavement-like aspect, the formation of domes and apical tight junctions and microvilli in confluent cultures. These bovine intestinal cells were shown to express in vitro epithelial cell markers such as brush border enzymes (maltase and alkaline phosphatase) and the epithelium typical cytoskeleton proteins, the cytokeratins. The enterocyte isolation procedure developed in the present study proved efficient to avoid a noticeable contamination from α-actin-positive cells (presumably myofibroblasts) in the two first steps of each type of culture (initial primary culture and first passage). The first passage also corresponded to cells presenting still a substantial differentiation level (in terms of brush border enzyme activities). To this respect, the first passage of each type of cultures appeared to be suitable to establish immortalized cell lines, a task which is now underway in our laboratory.
The functional differentiation state of the colon cells in culture could be estimated by the activities of drug metabolizing enzymes [31
]. Indeed, bovine colon epithelial cell culture were characterized, as freshly isolated cells, by cytochrome P450 1A1-associated 7-ethoxyresorufin O-deethylase activity as well as by prostaglandin H-synthase-mediated production of prostaglandin E2. Activities of phase II enzymes (i.e. N-acetyltransferase 1) were also observed in colon epithelial cell cultures.
By contrast to most reports about mammalian enterocyte primocultures, data accumulated in the present study using (1) cell immunocytochemistry, (2) western blot and (3) gene expression analyses, showed that intestinal cell cultures from bovine jejunum and colon co-expressed epithelial markers and vimentin, an embryonic cytoskeleton filament that is expressed only in mesenchymal cells after birth. A second fibroblast marker, the α-actin was scarcely detected, indicating that the vimentin strong expression of culture samples did not reflect a culture contamination by mesenchymal cells. In addition, the vimentin distribution pattern did not exactly fit with the gene expression of this protein. Indeed, the immunodetection was negative in homogenates of fresh epithelia from jejunum and colon, as well as in organoid suspension used to seed each culture type, while the protein was essentially detected in samples of culture after the second passage. By contrast, the vimentin gene transcript appeared in all bovine samples analysed, including the undissociated epithelium. It seemed then that a post-transcriptional basal inhibition of vimentin synthesis has been suppressed in vitro
. In agreement with this hypothesis, a previous research dedicated to the development of an intestinal lineage from the porcine intestine [12
] led to the production of vimentin-positive cultures. In view of this result, authors postulated that cultured cells had undergone a "mesenchymal transformation" in vitro
. Similarly, a re-expression of vimentin has been reported in epithelial cells under pathological conditions in vivo
, as well as in primary culture [32
]. Vimentin re-expression in vitro
could be instrumental in the maintenance of cell structure and/or functions of specific proteins such as the ones associated to membrane lipid rafts [33
Besides the possible mesenchymal transformation of epithelial cells in vitro
, accumulating data are now in favour of a natural incidence of vimentin in the undissociated intestine epithelium. For instance, specialized M-cells found in the epithelium covering the intestine Peyer's patches were identified, among other criteria, by vimentin immunostaining [34
]. These "Microfold-cells" are almost devoid of microvilli at their apical side as they are specialized in the transport of particulate antigens from the gut lumen to the underlying lymphoid tissue, where specific immune responses could take place. Although the structure and functions of these cells seem to be broadly admitted it is still unclear whether epithelial cells leaving the epithelium crypts are predetermined as M-cells or whether their particular phenotype developed from differentiated enterocytes at the lymphoid tissue vicinity. In favour of a possible local induction of M-cells differentiation, in vitro
experiments using Caco-2 cells showed that the enterocyte typical phenotype could be converted in a M-like phenotype adding B lymphocytes to the basolateral side of the epithelium [40
]. Regardless the aforementioned debate concerning the M-cells differentiation, enterocyte-lymphocyte co-culture experiments clearly demonstrated that a non-bacterial environmental factor could cause the brush border effacement of a vimentin-negative enterocyte (Caco-2 cells) leading to the phenotype of an intestinal cell type with a vimentin synthesis potential (M-like cell). Supporting the role of lymphocytes as phenotype conversion inducers, vimentin-positive cells were observed scattered throughout the villus epithelium of the rabbit small intestine, with the protein location extending from the perinuclear region to the cell membrane touching intraepithelial lymphocytes [43
]. Additional indications that vimentin could be a marker of "differentiation variants" of enterocytes, came from a recent study that pointed out, in the ordinary epithelium villi of the rabbit ileum, a vimentin-positive enterocyte type sharing M-cell morphological features (brush border poor cells), but most probably representing a distinct cell type. Indeed, these so-called "cup-cells", differed from M-cells at two levels: they bound distinctive lectins and they did not take up microbeads instilled in the ileal lumen [44