**1. Introduction**

Despite the recent innovations in drug delivery, oral administration remains the major route of drug administration. Understanding the drug absorption and metabolism in the intestine is thus essential for drug development. To date, several in vitro or ex vivo models are available for the evaluation of drug absorption and metabolism in human intestine. The Caco-2 cell culture model, e.g., is considered the gold standard in vitro model for studies of drug absorption, although there are limitations of this model with regards to drug metabolism [1,2]. Although a number of drug metabolizing enzymes (DME) have been identified in Caco-2 cells, including UDP glucuronosyltransferases (UGT) [3] and carboxylesterases (CES) [4], cytochrome P450 3A4 (CYP3A4), the major drug metabolizing enzyme in human intestine and liver is missing in Caco-2 cells [5]. Alternatively, sections of human intestinal tissue or mucosal biopsy mounted in Ussing chamber can be used to study drug absorption and metabolism in human intestine [6]. The advantage of this method is the combined measurement of permeability/active transport and metabolism. However, throughput, cost and availability of human tissues limit the routine use of this model. In order to integrate CYP3A4 activities into Caco-2 model, Takenaka et al. co-expressed recombinant human CYP3A4 and NADPH-CYP P450 reductase in Caco-2 cells [7]. Although a good correlation between extraction ratios observed in vitro and the gastro-intestinal (GI) extraction ratios in human could be observed for a number of reference compounds, the in vitro model tend to underestimate the GI extraction, indicating a rather low CYP3A4 activity in the model. Since

CYP3A4 activities are readily detected in microsomes prepared from liver or intestine, Gertz et al. used another approach to improve the predictivity of firstpass extraction in human by combining the metabolic clearance of CYP3A4 compounds measured in human intestinal microsomes and the permeability data from Caco-2 or MDCK-MDR1 assay [8]. A clear disadvantage of this approach is that two separate in vitro measurements are needed. In recent years, new in vitro models for human intestine like microfluidic tissue-on-chip [9] or organoids [10] are emerging. These models have been shown as useful models for testing compound toxicity in GI tract or as disease models [9,10]. As an ADME model for human gut, however, a tissue model on a Transwell basis would be more favorable because the equipment for Caco-2 permeability assay could be easily adapted to the new model and would enable the measurement of transcellular permeability and transport in the new model. Two such models have been published recently with basic characterization regarding drug transporters and DMEs: The EpiIntestinal microtissues provided by MatTek [11] and the 3D bioprinted human intestinal tissues provided by Organovo [12]. Due to the earlier availability and easier accessibility we decided to evaluate the EpiIntestinal model as an ADME tool in more detail. The aim of the present study is the in-depth characterization of the EpiIntestinal microtissues as a model for the investigation of activities of drug transporters and DME and for the prediction of GI firstpass availability in human.
