Feeding FIDO : development, validation and application of a dynamic, in vitro model of the gastrointestinal tract of the dog

To produce a complete and balanced dog diet it is necessary to know the nutritional needs of the dog as well as the availability of nutrients from the diet. Petfood manufacturers are restricted in performing (invasive) studies with animals for ethical reasons. Therefore, it is necessary to search for <em>in vitro</em> alternatives to these studies as explained in Chapter 1. Based on a literature study on the physiology of the GI tract of the healthy, adult dog (Chapter 2), the dynamic <em>in vitro</em> model for human subjects, pigs and calves, as described by Minekus (1995), was modified to mimic the physiological conditions of the dog. The model is named FIDO ( <em>f</em> unctional gastro <em>i</em> ntestinal <em>do</em> g model).The aim of the study was to develop the <em>in vitro</em> model simulating the GI tract of the dog. The model should simulate the physiological conditions in the GI tract of the dog as closely as possible. In that way luminal processes as well as physical and chemical properties of diets could be investigated. This thesis describes the developmental experiments, the validation experiments in comparison to dog studies as well as application studies.<h3>Validation of the model</h3><h4>Technical validation</h4>The study started with an extended literature review on the physiology of the stomach and small intestine of the healthy, adult dog with a special emphasis on those parameters which are relevant for the development of the dog model (Chapter 2). With respect to digestion of food, such parameters as transit times, pH values, concentrations of electrolytes and activities of enzymes are important to mimic physiological conditions as closely as possible. Data found in the literature were translated to a computer program to simulate these parameters in FIDO. The features of the model are described in Chapters 2 to 7. Based on the simulation of the physiology of the dog dry and canned dog foods were used to test the technical possibilities of the model. After some technical modifications to the gastric compartment and the pre-filters connected to the jejunal and ileal compartments, a study was performed to test the effect of particle size of dry dog food on gastric emptying. Particles≤3 mm emptied more slowly than particles≤1 mm. The effect of transit time on the availability for absorption of nitrogen and calcium of canned dog food was also investigated. Like <em>in vivo</em> , in FIDO less nitrogen and calcium were available for absorption with faster transit times (Chapter 3).<h4>Validation <em>in vitro</em> versus <em>in vivo</em></h4>Validation of the model in comparison to the <em>in vivo</em> situation was the next step in the development of the dog model (Chapter 4). Ileal protein digestibility and availability for absorption of nitrogen of eight different dry dog foods were tested in the model. The results were compared with data found <em>in vivo</em> with ileally cannulated dogs, performed at the University of Illinois. The experiments proved to be very reproducible and the results found in FIDO are similar to those found <em>in vivo</em> in the dogs.Based on these findings and those of former experiments simulating pigs, calves and human beings, it can be concluded that the model is a suitable tool as an alternative to animal experiments in nutritional research.<h3>Application of the model</h3>The validation study (Chapter 3) showed a low availability for absorption of calcium in the model. Based on these results it was decided to study calcium and phosphorus availability from three commercially available canned dog foods (Chapter 4). Effects of addition of calcium-phosphorus supplements or the enzyme phytase and the effect of a lower pH in the small intestine were also included in this study. A human standard breakfast was used as a control diet, because this diet had a high calcium availability under human conditions in the diet.The results showed that the canned dog foods had a low availability for absorption for calcium (maximum 21%) and phosphorus (maximum 44%). Differences in relative availability of calcium and phosphorus were found among these diets, which can probably be explained by the source of these minerals. Also the low availability of these minerals can be attributable to the source. Another possible explanation of the low availability are the (saturated) fatty acids in the diet, which can form calcium soaps in the model.The advantage of FIDO is that the availability for absorption of calcium and other nutrients can be studied without the influence of the physiological status of the animal. The real amount absorbed by the animal, however, cannot be studied. Absorption by the animal depends on its needs and the absorption is hence dependent on two different mechanisms: passive and active absorption. To get more insight into the absorption of calcium through the intestinal wall of the dog, experiments were performed with Great Danes (8 and 20 weeks of age; in cooperation with the Veterinary Faculty of Utrecht University), FIDO and intestinal segments (Chapter 7). The effect of vitamin D was taken into account in this study by studying two levels in the diet. The three different methods ( <em>in vivo</em> , FIDO and intestinal segments) are complementary and can be used to get a better understanding of the regulation of calcium absorption in the dog.In Chapter 6 experiments are described to investigate the effect of gelling agents on the digestibility and availability for absorption of nutrients ('chunks in gravy' products). Three different (combinations of) gelling agents added to the gravy were used in this study in different concentrations (0.2% and 0.5%): carrageenan plus guar gum, carrageenan plus locust bean meal, and wheat starch. A diet without gelling agent was used as a control diet. Neither the gelling agent nor the concentration had any effect on digestibility of proteins and carbohydrates, availability for absorption of calcium and phosphorus, viscosity or buffering capacity in the intestinal content. From the results it can be concluded that addition of the gelling agents used does not affect the nutritional quality of the diets at the low concentrations tested.<h3>Conclusions</h3>The dynamic <em>in vitro</em> model of the GI tract of dogs simulates the physiological parameters very accurately and reproducibly. Transit time of food has an effect on digestibility and availability for absorption, just like in dogs. Also ileal protein digestibility in the model is similar to the data found <em>in vivo</em> . It is possible to investigate specific questions regarding dog food in this model (e.g. the effect of gelling agents on digestibility). Another important aspect is the fact that the experiments in FIDO can be performed under highly standardized condition, in contrast to <em>in vivo</em> studies in which biological variance among animals plays a role. This comes to expression in the reproducibility and sensitivity of the results from FIDO compared to results of <em>in vivo</em> studies.The dynamic <em>in vitro</em> model of the gastrointestinal tract of the dog is a suitable alternative to <em>in vivo</em> studies with respect to digestibility and availability for absorption of nutrients from different types of dog food, such as canned and dry dog foods.