Maturation of the Caco-2 Cell Monolayer impedes the uptake of outer Membrane Vesicules

Abstract

It has been suggested that internalization of OMV secreted by gram-negative bacteria by the intestinal epithelium is a means of inter-kingdom communication. But most previous studies of OMV internalization have used undifferentiated epithelial cells, which differ markedly from the mature enterocytes that line the intestine. Therefore, the aim of this study was to investigate if OMV shed by Escherichia coli provide a means of communication between these bacteria and enterocytes. Specifically, the internalization of OMV by undifferentiated epithelial cells and differentiated enterocytes was compared and the factors that alter OMV internalization with maturation were investigated. In addition, possible effects of OMV on the intestinal epithelium independent of internalization were investigated.

The Caco-2 cell line was used as a model of the intestinal epithelium and OMV from pathogenic enterotoxigenic E. coli (ETEC) and probiotic E. coli Nissle 1917 (EcN) were investigated. Adhesion and uptake of OMV were assessed using fluorescently labeled OMV detected by flow cytometry or confocal microscopy. Transepithelial electrical resistance (TEER) and the secretion of interleukin-8 (IL-8) were used as indicators of barrier integrity and an innate immune response, respectively. OMV preparations and the morphology of Caco-2 cells were visualized using electron microscopy (EM), while the localization of junctional proteins (occludin and E- cadherin) and the proliferation phenotype (Ki-67) of cells were determined using immunofluorescence.

Undifferentiated three-day-old Caco-2 cells progressively accumulated both ETEC and EcN OMV over a 24-hour period. In contrast, 21-day-old Caco-2 cells, which were fully differentiated and formed a confluent monolayer, did not internalize OMV and there was limited adherence of OMV to the cells. This reduced uptake was not due to degradation of the OMV.

Despite the lack of uptake, apical exposure of 21-day-old differentiated Caco-2 cells to crude EcN OMV, but not basolateral exposure, induced significant IL-8 secretion and these results were replicated in undifferentiated Caco-2 and HT29 cells. However, crude OMV from the aflagellate EcN ∆flic mutant did not induce IL-8 secretion, suggesting that the effect was due to contaminating flagella. This was confirmed with EM of the OMV isolates. Although crude ETEC OMV were also contaminated with flagella, they did not stimulate IL-8 secretion by differentiated Caco-2 cells, but both apical and basolateral exposure to crude ETEC OMV induced a sustained reduction in TEER which was most likely related to heat-labile enterotoxin associated with ETEC OMV. In contrast, apical exposure to crude EcN OMV had no effect on TEER, whereas basolateral exposure transiently decreased TEER for 24 hours.

Comparison of the uptake of OMV at 3, 5, 7 & 21 days indicated that ≈ 70% of the reduction in uptake occurred between days three and five. This was not due to cell differentiation as three and five day old cells had a similar ultrastructure, with a cuboidal shape, central nuclei, poorly developed microvilli, and no glycocalyx. However, between days three and five an increased level of cell polarization was indicated by the localization of E-cadherin and occludin to the points of cell-cell contacts, consistent with the formation of tight junctions, which was supported by an associated increase in TEER. This was accompanied by a decrease in the rate of cell proliferation and the impact of cell proliferation on OMV uptake was highlighted by the observation that a higher proportion of proliferating cells (17.5 ± 2.6%) took up OMV than non-proliferating cells (10.9 ± 1.5%).

Collectively, these results show that polarization due to the formation of tight junctions, and the concomitant reduction in cell proliferation are associated with a marked reduction in endocytosis of OMV. Hence it is unlikely that OMV shed by gram-negative bacteria are internalized by differentiated intestinal enterocytes in a healthy gut. Nonetheless, interaction with enterocytes could occur due to the internalization by enterocytes in the event of barrier dysfunction. Alternatively, in vivo effects could be the result of internalization by specialized epithelial cells such as microfold cells, or by proliferating transit-amplifying cells.