In the first part of the project the consortium derived two new rainbow trout intestinal cell lines that are stable and represent both the proximal and the distal portion of the organ. At the same time an efficient and reliable method for extracting the bioactive fraction of the feed pellets was developed. This enables the pellets to be administered to the cells. Finally, an extensive range of Hydrogel -based scaffolds and culture devices have been developed and tested, but any of the resulting combinations proved to be adequate to the scope of the project.
During the last 27 months the consortium developed and tested two functional prototypes based on commercially available cell supports. Both are based on a bicameral system where the upper chamber mimics the intestinal lumen, and the lower chamber is the functional equivalent of the blood vessels. One includes only epithelial cells, while the other includes also a tri-dimensional scaffold that hosts fibroblasts and extracellular matrix. After a thorough comparison the conclusion was that the simpler prototype is more reliable and enables a wider range of analytical tests.
In the final part of the project we tested with a traditional nutritional trial in vivo 5 diets (see Deliverable 6.1) containing protein sources known to vary in nutritional value and health value, i.e. fish meal, soybean meal, feather meal, as well as one single cell protein at two inclusion levels. At the same time, we used the protocol previously developed to extract the bioactive fractions of same diets and added them to the selected in vitro platform prototype. Finally, the two sets of results were compared to determine the degree of correspondence between the two and, therefore, establish Fish-AI predictive power.
The Fish-AI prototype function and predictive ability can be summarized as follows:
• The final Fish-AI prototype can provide a high-resolution ranking of different diets.
• Fish-AI provides repeatable results within the same laboratory.
• Fish-AI provides different ranking depending on the analysis performed with it.
• Good correlation with in vivo results was achieved.
• Since in vivo trials measure parameters very different from in vitro tests, additional work is required to further improve how to correlate these two heterogenous sets of results
