Human Small Intestinal Tissue Models to Assess Barrier Permeability: Comparative Analysis of Caco-2 Cells, Jejunal and Duodenal Enteroid-Derived Cells, and EpiIntestinal Tissues in Membrane-Based Cultures with and Without Flow.

Accurate in vitro models of intestinal permeability are essential for predicting oral drug absorption. Standard models like Caco-2 cells have well-known limitations, including lack of segment-specific physiology, but are widely used. Emerging models such as organoid-derived monolayers and microphysiological systems (MPS) offer enhanced physiological relevance but require comparative validation. We performed a head-to-head evaluation of Caco-2 cells, human jejunal (J2) and duodenal (D109) enteroid-derived cells, and EpiIntestinal tissues cultured on either static Transwell and flow-based MPS platforms. We assessed tissue morphology, barrier function (TEER, dextran leakage), and permeability of three model small molecules (caffeine, propranolol, and indomethacin), integrating the data into a physiologically based gut absorption model (PECAT) to predict human oral bioavailability. J2 and D109 cells demonstrated more physiologically relevant morphology and higher TEER than Caco-2 cells, while the EpiIntestinal model exhibited thicker and more uneven tissue structures with lower TEER and higher passive permeability. MPS cultures offered modest improvements in epithelial architecture but introduced greater variability, especially with enteroid-derived cells. Predictions of human fraction absorbed (F) were most accurate when using static Caco-2 data with segment-specific corrections based on enteroid-derived values, highlighting the utility of combining traditional and advanced in vitro gut models to optimize predictive performance for F. While MPS and enteroid-based systems provide physiological advantages, standard static models remain robust and predictive when used with in silico modeling. Our findings support the need for further refinement of enteroid-MPS integration and advocate for standardized benchmarking across gut model systems to improve translational relevance in drug development and regulatory reviews.