The gut of fish belongs to the most essential barriers that mark the border between the organism and its surrounding environment. The pivotal barrier function, allowing absorption of nutrients from the diet while simultaneously protecting the organism from pathogens and contaminants, is accomplished by a single layer of epithelial cells lining the intestinal lumen. In vitro, this barrier has been mimicked by culturing fish epithelial intestinal cells on conventional permeable membranes within a two-chamber system, which creates an upper and a lower compartment representing the intestinal lumen and blood circulation, respectively. This simplified approach, however, has at least three important limitations: The first being restricted diffusion through the several micrometer thick commercial membranes, which moreover have quite limited porosity; the second being a lack of interaction with other intestinal cell types; and the third being absence of mechanical stimulation, such as shear forces from fluid flow to better simulate the physiology of the intestine. To overcome these limitations, this thesis focuses on the recreation of the piscine intestinal microenvironment by combining cells derived from the intestine of fish, precisely epithelial and fibroblast cell lines from rainbow trout (Oncorhynchus mykiss), with engineered microsystems. The applied stepwise approach encompasses (a) the development of an ultrathin permeable membrane as novel support for barrier forming cells, (b) followed by combining epithelial cells and fibroblasts for intestinal architecture reconstruction, and (c) exposure to fluid flow to mimic the mechanical forces occurring on the epithelial-lumen interface.

Artificial ultrathin membranes for intestinal cell culture were found to better mimic features of the delicate, highly permeable basement membrane that underlines the epithelial cells in vivo compared to conventional porous supports. Two types of membranes were fabricated in this study. The first was an anodized aluminum oxide membrane that features densely packed pores in the nanometer range and allows for fast diffusion of small molecules. This type of membrane is ideal for high quality microscopy and supports epithelial polarization. However, membranes release, albeit non-cytotoxic, concentrations of aluminum ions, which might be critical for toxicological investigations. Therefore, a second type of an ultrathin membrane, namely a silicon nitride membrane, was fabricated. It has pores in the micrometer range, is optically transparent and has been applied beyond static exposures for microfluidic studies.

To initiate co-culture of epithelial cells with fibroblasts, an intestinal fibroblast cell line from rainbow trout needed to be characterized first. The cells feature typical fibroblast morphology and behavior and appear to be infinite. The combination of epithelial and fibroblast cells, when in direct contact, had no beneficial effect on barrier tightness. Cell culture on opposite sides of ultrathin alumina membranes, however, resulted in increased trans-epithelial electrical resistance, suggesting enhanced barrier tightness from cellular cross-talk.

A uniquely designed microfluidic bioreactor with integrated ultrathin silicon nitride membranes as substrate for cell growth was finally developed to allow for realistic flow conditions on epithelial cells. The arising fluid shear stress on epithelial monolayers and epithelial-fibroblast co-cultures positively affected barrier resistance when applied at a moderate rate. This physiological adaptation allows for better comparability to the fish intestine in vivo compared to cells cultured under static conditions. This finding highlights the importance of mechanical stimulation for realistic organ mimicry within in vitro systems.

To conclude, this thesis demonstrates the benefits of recreating a more physiologically realistic microenvironment for epithelial cell cultures from fish intestine. By introducing ultrathin permeable membranes as novel culture substrate, adding a fibroblast cell line and shear stress, the novel in vitro intestinal barrier model now better reflects intestinal properties of the in vivo counterpart and allows for improved exposure and transport phenomena during experimental uptake studies, e.g. for chemical pollutants. This thesis therefore paves the way for improved understanding of normal and impaired physiology of the fish intestine and better in vitro to in vivo extrapolation while contributing to a reduced need of animal experiments.