Introduction
Variation in nutrient digestibility is among the most critical points of variation in feed quality and constitutes a key parameter when evaluating new feed raw materials for fish. Nutrient digestibility is traditionally assessed in vivo by animal feeding trial s, but recent EU obligations to put the 3Rs concept into practice have facilitated research initiatives to develop complementary in vitro methods. One of the current project initiatives addressing this topic is the H2020-FETOPEN project Fish-AI, which aims to develop an in vitro screening platform based on rainbow trout (Oncorhynchus mykiss) digestive enzymes and intestinal cells to evaluate nutritional and health values of novel aquafeeds . As part of this project, we have developed a static in vitro digestion (IVD) model for simulating the in vivo feed digestion in rainbow trout. In the present study , the performance of the Fish-AI IVD model was assessed by comparisons with two independent in vivo feeding trials using juvenile rainbow trout and mink (Neovison vison ) as experimental animals. In addition, the Fish-AI model was compared against the standardised human in vitro model INFOGEST1.
Materials and methods
A total of seven diets based on commercially relevant protein raw materials with assumed contrasting effects on protein digestibility were formulated. Six test feeds contained the challenging raw materials guar meal and soybean meal at relatively high inclusion levels, feather meal at low and high inclusions, and the combinations guar/feather and soybean/feather. A diet high in fish meal was included as a reference. The study in rainbow trout comprised a 5-week feeding trial using fish with an average start weight of 55 grams. Fish were fed the seven experimental feeds in duplicate tanks each. Fecal collection was conducted by stripping. The mink trial was carried out using adult male mink with a m ean body weight of 2.8 kg . Four healthy mink were assigned for each of the seven experimental diets. The animals were kept in individual cages equipped for controlled feeding and quantitative faecal collection. The experiment lasted for 14 days . F eed and faeces samples were analyzed by established procedures2. The Fish-AI IVD procedure was based on the static INFOGEST model with several modifications, including the use of enzyme extracts and bile obtained from rainbow trout , whereas the INFOGEST protocol was performed as described elsewhere1. The IVD products from both in vitro procedures were analyzed using size exclusion chromatography (SEC) and quantification of soluble n itrogen with Dumas for estimation of the proportion of small soluble peptides potentially available for uptake3. Difference among diet groups was analyzed for statistical significance using a one-way ANOVA followed by Fischer’s LSD test. The level of significance was set to p<0.05.
Results and discussion
Figure 1 illustrates the results regarding apparent protein digestibility for the seven experimental diets as tested in the four separate model systems. Significant effects of diet were evident. Rainbow trout fed either the control, guar-high, or soya-high diet generally showed high (>95%) protein digestibilities, whereas inclusion of feather meal in the diet, singly or in combination with guar or soya, clearly reduced protein digestibility. The results obtained with mink were in overall accordance with the results obtained from the study in rainbow trout and strengthen the observation that the biological value of the feather meal protein was low. The results from the two in vitro model systems displayed relatively similar protein digestibility profiles among the experimental diets as the two in vivo models, thus demonstrating that the low protein value of the diets containing feather meal also could be detected in vitro . The absolute values of protein digestibility were higher for the INFOGEST than for the Fish-AI IVD procedure (70-86 vs. 43-62%) and indicate that there is still room for improving the Fish-AI IVD efficiency. These studies are currently ongoing and will be presented and discussed.
Acknowledgements
This project has received funding from the European Union’s Horizon 2020 research and innovation program under grant agreement No 828835.
References