Validation of a Caco-2 microfluidic Chip model for predicting intestinal absorption of BCS Class I-IV drugs

Oral delivery is considered the most patient preferred route of drug administration, however, the drug must be sufficiently soluble and permeable to successfully formulate an oral formulation. There have been advancements in the development of more predictive solubility and dissolution tools, but the tools that has been developed for permeability assays have not been validated as extensively as the gold-standard Caco-2 Transwell assay. Here, we evaluated Caco-2 intestinal permeability assay in Transwells and a commercially available microfluidic Chip using 19 representative Biopharmaceutics Classification System (BCS) Class I-IV compounds. For each selected compound, we performed a comprehensive viability test, quantified its apparent permeability (P_app), and established an in vitro in vivo correlation (IVIVC) to the human fraction absorbed (f_a) in both culture conditions. Permeability differences were observed across the models as demonstrated by antipyrine (Transwell P_app: 38.5 ± 6.1 × 10^-8 cm/s vs Chip P_app: 32.9 ± 11.3 × 10^-8 cm/s) and nadolol (Transwell P_app: 0.6 ± 0.1 × 10^-7 cm/s vs Chip P_app: 3 ± 1.2 × 10^-7 cm/s). The in vitro in vivo correlation (IVIVC; P_app vs. f_a) of the Transwell model (r² = 0.59-0.83) was similar to the Chip model (r² = 0.41-0.79), highlighting similar levels of predictivity. Comparing to historical data, our Chip P_app data was more closely aligned to native tissues assessed in Ussing chambers. This is the first study to comprehensively validate a commercial Gut-on-a-Chip model as a predictive tool for assessing oral absorption to further reduce our reliance on animal models.