The development of in vitro digestion models relies on the availability of in vivo data such as digestive enzyme levels and pH values recorded in the course of meal digestion. The variations of these parameters along the GI tract are important for designing dynamic digestion models but also static models for which the choice of representative conditions of the gastric and intestinal conditions is critical. Simulating gastric digestion with a static model and a single set of parameters is particularly challenging because the variations in pH and enzyme concentration occurring in the stomach are much broader than those occurring in the small intestine. A review of the literature on this topic reveals that most models of gastric digestion use very low pH values that are not representative of the fed conditions. This is illustrated here by showing the variations in gastric pH as a function of meal gastric emptying instead of time. This representation highlights those pH values that are the most relevant for testing meal digestion in the stomach. Gastric lipolysis is still largely ignored or is performed with microbial lipases. In vivo data on gastric lipase and lipolysis have however been collected in humans and dogs during test meals. The biochemical characterization of gastric lipase has shown that this enzyme is rather unique among lipases: (i) stability and activity in the pH range 2 to 7 with an optimum at pH 4-5.4; (ii) high tensioactivity that allows resistance to bile salts and penetration into phospholipid layers covering TAG droplets; (iii) sn-3 stereospecificity for TAG hydrolysis; and (iv) resistance to pepsin. Most of these properties have been known for more than two decades and should provide a rational basis for the replacement of gastric lipase by other lipases when gastric lipase is not available.

Various combinations of digestive lipases were tested in vitro under conditions simulating the earlier phases of gastrointestinal lipolysis in the stomach and the duodenum. A solid/liquid test meal was mixed first with either human gastric juice or a solution containing gastric lipase, followed by either the addition of human pancreatic juice and bile or the addition of a solution containing pancreatic lipase, colipase, and bile salts. The rate of lipolysis and the composition of the lipolysis products were assessed as a function of time after lipid extraction and analysis by thin-layer chromatography coupled to flame ionization detection. The lipolytic potential of a crude rabbit gastric extract (RGE) associated with porcine pancreatic extract (PPE) was assessed and compared with the rates of lipolysis of the meal triacylglycerols by human digestive lipases recorded under the same in vitro conditions. RGE combined with PPE appeared to be a good substitute for human gastric and pancreatic lipases. RGE and PPE could therefore be used to simulate the gastrointestinal lipolysis of various foods and emulsions in vitro, as well as that of pharmaceutical lipid formulations.

No abstract

Rabbit gastric extract (RGE) is a source of gastric enzymes for in vitro digestion studies that contains both gastric lipase and pepsin. β-Casein was used here as a model protein for testing the pepsin activity of RGE.