Physiologically Relevant Simulation of Carbohydrate Digestion: From Glycemic Index Estimation to Intestinal Cellular Responses

Abstract

Simulating carbohydrate digestion in physiologically relevant ways remains a challenge for in vitro models. In this study, the Dynamic In vitro Human Stomach (DIVHS) system was applied to investigate cereal digestion and subsequent intestinal cellular responses. Rice, millet, and corn were digested under dynamic and static conditions. Compared with the static model, the dynamic system generated smaller grain fragments, a larger chyme–enzyme contact area (451.2 ± 4.4 cm2 vs. 160.4 ± 6.0 cm2), and higher average intragastric pressure (25.0 ± 1.2 kPa vs. 7.2 ± 0.7 kPa). Salivary amylase activity also declined more gradually in the dynamic system. An empirical approach for predicting the glycemic index (eGI) was proposed, which showed improved agreement with reported human GI values compared with earlier in vitro methods. Exposure of Caco-2 cells to digested products significantly altered transcriptional profiles, including protein binding, ATP binding, and glucose transporter activity. Notably, products from the dynamic model induced stronger transcriptional responses than those from the static model, including 421 genes up-regulated and 138 down-regulated. Functional enrichment highlighted pathways related to glucose transport, energy metabolism, and cellular regulation. Overall, this study demonstrates the advantages of dynamic digestion models in replicating gastrointestinal conditions, improving GI prediction, and providing mechanistic insights into intestinal cellular responses to digested carbohydrates.