Human enteroid monolayers as a potential alternative for Ussing chamber and Caco-2 monolayers to study passive permeability and drug efflux

,


Introduction
The intestine is a dynamic barrier involved in absorption and biotransformation of xenobiotic compounds, including drugs.After oral administration, the intestine is the first site of drug absorption, making it a key determinant of drug bioavailability and hence a drug's ultimate efficacy and safety.Elucidating its pharmacokinetic and toxicokinetic drug interactions provides better guidance in safety assessment and can be used to understand and potentially also predict pharmacokinetics, e. g. using physiologically based pharmacokinetic (PBPK) modeling (Darwich et al., 2010;Harwood et al., 2013).
Compounds can pass the mucosal membrane passively or actively.Passive transport occurs transcellularly or paracellularly (through the tight junctions).Active transport of drugs is mediated by membrane drug transporters (DTs), which can be uptake or efflux transporters (Alqahtani et al., 2021).P-glycoprotein (P-gp/ABCB1) and Breast Cancer Resistance Protein (BCRP/ABCG2) are the two major efflux transporters expressed in the intestinal barrier, both members of the ATP-binding cassette family and located on the apical membrane of the enterocyte (Estudante et al., 2013).Efflux transporters prevent compounds from entering the human body, protecting against exogenous compounds, hence limiting drug absorption by transporting them back into the intestinal lumen (Galetin et al., 2024).The expression of P-gp is known to increase along the small intestine, being highest in the ileum region but low again in the colon, while evidence for regional BCRP expression is currently contradictory (Drozdzik et al., 2019;Estudante et al., 2013;Muller et al., 2017;Terada and Hira, 2015).
In drug discovery the in vitro gold standard to study drug transport is the Caco-2 model, a human immortalized colon cancer cell line (Lu et al., 2022).Despite their wide use, Caco-2 cells lack the complexity of in vivo physiology as they do not display the interplay between different intestinal cell types, lack relevant metabolic enzyme expression and biological variation (Englund et al., 2006;Muller et al., 2017;Ölander et al., 2016).An alternative model to study intestinal drug transport is the Ussing chamber.In this ex vivo tissue preparation model, a piece of fresh intestinal tissue is mounted between two chambers enabling measurements of bidirectional drug transport over the intestinal barrier (Michiba et al., 2020;Rozehnal et al., 2012;Sjoberg et al., 2013;Streekstra et al., 2022).This methodology is technically demanding, time-consuming, while intestinal tissue availability is limited and when available also fragile and only viable for a short period of time.
Innovative approaches like tissue-derived organoids have the potential to better bridge the gap between in vitro studies and in vivo responses.In 2011, Sato et al., developed a technique to isolate stem cells from the human intestine and grow 3D structures called human intestinal organoids (enteroids) (Sato et al., 2011).These organoid models, derived directly from donor specific tissue, are self-organizing, multicellular complexes.Enteroids mimic many of the intestinal epithelium's key attributes, maintaining the (patho)physiological properties as genetic region specificity and disease characteristics of the original tissue (Fair et al., 2018;Jelinsky et al., 2022;Kraiczy et al., 2019;Meran et al., 2020;Michiba et al., 2022;Middendorp et al., 2014a).Initially, enteroids are grown in a 3D formation exhibiting the luminal site of the intestine on the inside, which makes the lumen inaccessible for drug exposure.Yet, enteroids can be grown in a monolayer formation creating an apical and basolateral surface, enabling the possibility to study drug interactions with transporters, metabolizing enzymes, and its barrier properties (Altay et al., 2019;Braverman and Yilmaz, 2018;Gunasekara et al., 2018;Kozuka et al., 2017;Scott et al., 2016;Speer et al., 2019b;Takenaka et al., 2016;Wang et al., 2017;Yamashita et al., 2021).This makes enteroid monolayers a promising model to study intestinal bidirectional drug transport and predict oral drug absorption.
The aim of this study was to explore the suitability of enteroid monolayers as an experimental platform for drug transport studies that complements the Ussing chamber and widely used Caco-2 cells.Moreover, we apply enteroid monolayers to demonstrate their applicability to study intestinal region-specific variation.We did this by comparing bidirectional transport of model compounds (enalaprilat, propranolol, talinolol, rosuvastatin) in enteroid monolayers with that in the Ussing chamber model, both in the jejunum and ileum region.In addition, we compared results from these systems with the conventional Caco-2 monolayer.Gene expression analysis was used to compare composition of enteroid monolayers and tissues using RNA sequencing to investigate whether enteroids have similar (regional related) expression patterns as their primary tissue.

Human tissue
Left-over intestinal tissue was obtained from adult patients undergoing a pancreatoduodenectomy (proximal jejunum) or (right hemi) colectomy surgery (terminal ileum) at Radboud university medical center (Radboudumc), Nijmegen, the Netherlands from 2020 to 2023.No informed consent was needed for anonymous use of leftover material for research purpose, following the Dutch Code of Conduct for Responsible Use.The need for formal ethics approval was waived by the ethics committee of Radboudumc, Nijmegen, The Netherlands according to the Dutch Law on Human Research.Directly after surgical resection, tissue was put in ice cold Krebs buffer and transferred to the lab within 15 min.Part of the tissue was stored at 4 • C in supplemented Williams' E storage buffer for enteroid isolation (Table A.1), one part was immediately used for Ussing chamber experiments, while another part was snap frozen in liquid nitrogen for RNA analysis.

Cell culture 2.2.1. 3D enteroid culture
For the establishment of enteroids, crypts were isolated from tissue within 24 h after storage in transport buffer (jejunum n = 5, ileum n = 5).Procedures for generation and maintenance of enteroid cultures were based on STEMCELL™ Technologies IntestiCult™ protocol.Details on buffer and media compositions, including suppliers, are listed in Table A.1.In short, mucosal tissue was separated from muscle and serosa, cut into pieces of 2-3 mm and extensively washed with wash buffer.Next, tissue pieces were incubated for 1 hour in crypt releasing solution, after which crypts were collected from supernatant.Isolated crypts were suspended in Matrigel and seeded in 30 µl droplets per well in a 48-well tissue culture plate, covered with 300 µl Organoid Growth Medium.Medium was refreshed every 2-3 days and after 5-10 days cultures were mechanically passaged for culture continuation, or by enzymatic dissociation as described below for generation of enteroid monolayers.

2D enteroids monolayers
Procedures of monolayer culture were based on STEMCELL™ Technologies IntestiCult™ protocol.To create enteroid monolayers, clear membrane inserts (6.5 mm, 0.4 µm pore size, Corning Costar 3470, STEMCELL™ Technologies, CELLTREAT #100-0997) were coated with collagen type I (rat tail) in 0.1% acetic acid for 1 hour at room temperature and subsequently washed with PBS.3D enteroids (P1-P15) were collected in Advanced DMEM+++, centrifuged and made single cell with use of TripLE for 10 min, 37 • C and mechanical disruption.Cells were seeded on the permeable inserts in a concentration of 1 × 10 ^5 cells/well in OGM, with medium refreshment every 2-3 days.Apical volume was 100 µl, basolateral volume 600 µl.When the monolayer confluency was confirmed by microscopy and Trans Epithelial Electrical Resistance (TEER) measurement, culture medium was changed to Organoid Differentiation Medium, refreshing every 2-3 days.After 5 days of differentiation, the bidirectional transport assay was performed.

Bidirectional transport assay enteroid and Caco-2 monolayers
Drug transport experiments were performed in the apical to basolateral (A-to-B, n = 3) and basolateral to apical (B-to-A, n = 3) direction.TEER was measured before every experiment in transport buffer (Table A.1).A drug cocktail (see section drug cocktail) was prepared in transport buffer and added to either the apical or basolateral side of the insert.Fluorescein-dextran 4 kDa (FD4, 50 µM) was added to the apical side of each insert to monitor monolayer integrity during the experiment.Monolayers were put on a rocker (70 rpm) at 37 • C (Boj et al., 2021;Takenaka et al., 2014).Medium samples were taken at t = 30, 60 and 120 min for drug analysis.After the experiment, integrity was examined by visual inspection (Olympus microscope) and FD4 leakage to the basolateral compartment (cut-off value P app < 1 × 10 -6 cm/s (Speer et al., 2019a).

Immunofluorescent staining
Immunofluorescent staining of monolayers was performed according to established protocol (STEMCELL™ Technologies).In short, monolayers were washed with PBS and fixed in paraformaldehyde for 1 hour, after which they were stored in PBS until use.To visualize monolayer morphology, inserts were incubated with blocking buffer (PBS containing 2% normal goat serum, 1% bovine serum albumin, 0.1% Tri-tonTM X-100 and 0.05% Tween®20) at room temperature for 1 hour.After 3 times washing with PBS, Alexa Fluor™ Plus 555 Phalloidin (Invitrogen, A30106) in PBS with 1% bovine serum albumin was added overnight at 4 • C. Inserts were washed again 3 times with PBS, incubated with DAPI in PBS (5 gm/ml) for 30 min at 4 • C and after a final wash with PBS mounted on a glass slide with ProLong™ Gold Antifade Mountant (Invitrogen).Images were acquired using Axio observer microscope (ZEISS) and processed with ZEN 2.3 pro (ZEISS).

Ussing chamber
Exact procedures and details used for the Ussing chamber studies with intestinal tissues were previously described by Streekstra & Kiss et al., 2022(Streekstra et al., 2022).Additional details on buffers and culture medium composition are provided in Table A.1.In short, intestinal mucosal tissue was dissected from its muscle and serosal layers, after which it was vertically mounted between two chambers with an area of 0.71 cm 2 .The two chambers were filled with Krebs buffer with N-acetylcystein (700 µg/mL), which were kept at stable temperature of 37 • C and were carbonated continuously.Tissue viability was monitored by electrophysiology during the whole experiment.After 15-30 min of stabilization, the buffer was changed for prewarmed transport buffer containing a drug cocktail (see below) to start the bidirectional transport assay.Samples (50 µL) were taken at every 15 min up to two hours for drug analysis.

Data analysis
The apparent permeability (P app ) was calculated for each substrate individually according to Eq. ( 1), where dQ/dt is the rate of drug transport from one half chamber to the other (either A-to-B or B-to-A, expressed in ng/s), A the exposed area of the tissue (cm 2 ) and C 0 the initial concentration of the compound investigated in the donor compartment (ng/ml).

Gene expression analysis
RNA-Seq analysis was performed on snap frozen intestinal tissue and differentiated enteroid monolayers cultured on permeable inserts, collected after bidirectional transport experiments.mRNA was isolated with the RNeasy Mini kit (Qiagen) according to the manufacturer's protocol.For preparation of RNA sequencing libraries the KAPA RNA HyperPrep kit with RiboErase (human/mouse/rat [HMR]) (Kapa Biosystems) was used according to previously published methodology (Geckin et al., 2023).For sequencing 59-bp paired-end reads were generated with the Illumina NextSeq 2000 instrument.
Trim Galore! v0.4.5 (Babraham Bioinformatics, Babraham, UK) was used for low-quality filtering of bases in RNA-Seq reads and adapter trimming, a tool for Cutadapt v1.18 and FastQC v0.11.8 (Babraham Bioinformatics).A human reference genome (GRCh38.95,Ensembl) was used to map genes using STAR v2.7.5a.Gene counts were calculated with HTSeq (v0.11.0) and an Ensembl GRCh38.95GTF annotation file.Quality control was performed with MultiQC v1.7 to generate quality reports of all samples.Normalization and differential gene expression analysis was carried out in R v4.3.2 with DESeq2 v1.42.0.Pathway and GO term enrichment were performed with EnrichR v3.2 on selected genes (DESeq2 p-adjusted values < 0.05).Separate files on differential expressed transcripts in tissue and enteroid monolayers of the genes used in this publication are provided in Supplementary Table B.1.All gene counts are available upon request.

Statistical analysis
Data in the text and Table 1 are presented as median ± inter quartile range (IQR).A Kruskal-Wallis with Dunn's post-hoc test, based on median ± IQR, was applied to compare P app for transport direction and differences between jejunum and ileum (Fig. 2A-D and 3A-D, Table A.2 and A.3). Efflux ratios were compared with Mann Whitney t-test based on mean ranks (Fig. 2E + 3E, Table A.4). Pairwise donor comparison between tissue and enteroids was performed with a Wilcoxon-signed rank test for B-to-A transport of talinolol and rosuvastatin (Fig. 4).Deseq normalized counts in jejunum and ileum were compared with a Kruskall-Wallis with Dunn's post-hoc test based on median ± IQR (Fig. 4).All statistical analyses were performed using GraphPad Prism 8.0.1.The substrates enalaprilat and propranolol were selected as a control for passive transport via the paracellular and transcellular route, respectively.Paracellular enalaprilat transport (P app ) was comparable in both transport directions in both jejunum and ileum monolayers (Fig. 2A), indicating proper barrier formation.Propranolol transport was higher in B-to-A direction than A-to-B direction in ileum (Fig. 2B),
In parallel, we evaluated drug transporter function by bidirectional talinolol (P-gp substrate) and rosuvastatin (BCRP substrate) transport across the enteroid monolayers.For talinolol the P app in B-to-A direction was higher than A-to-B both in jejunum and ileum (Fig. 2C).Rosuvastatin transport was higher in B-to-A direction compared to A-to-B direction in ileum and showed the same trend in jejunum (Fig. 2D).For both substrates A-to-B transport was low (<0.5 × 10 -6 cm/s), also reflected by efflux ratios for talinolol and rosuvastatin that were all >>2 indicating functional P-gp and BCRP transport (Fig. 2E, Table 1).No significant difference was observed between jejunum and ileum in transport of talinolol and rosuvastatin.In addition, no regional effect was apparent in ER of talinolol (Fig. 2E).Efflux ratios for rosuvastatin however tended to be higher in ileum compared to jejunum (Fig. 2E, Table 1).

Bidirectional drug transport in human intestinal tissue
To compare bidirectional transport functionalities of enteroid monolayers and fresh intestinal tissue, we similarly studies transport in proximal jejunum (n = 14) and terminal ileum (n = 6) tissue explants in the Ussing chamber.Barrier integrity was monitored by TEER and FD4 The absorption rate of enalaprilat and propranolol in tissue was comparable in both transport directions and intestinal regions (Fig. 3A-B).Median efflux ratios for enalaprilat and propranolol were below 2 in both jejunum and ileum tissue, indicating passive transport of these substrates across the intestinal barrier (Fig. 3E, Table 1).
P app in B-to-A direction of talinolol tended to be higher than A-to-B transport in both jejunum and ileum (Fig. 3C, Table A.2).There was also a tendency of higher B-to-A transport of talinolol in the ileum tissue compared to jejunum tissue, suggesting higher activity of P-gp in ileum compared to jejunum, although this was not significant (Fig. 3C).The median efflux ratio of talinolol was >2 in 6 out of 9 experiments in jejunum and 2 out of 4 in ileum tissues, indicating active transport (Fig. 3E).In jejunum tissue, rosuvastatin transport from B-to-A was higher compared to A-to-B (Fig. 3D), with a median ER >2 in 7 out of 8 experiments indicating efflux transport (Table 1).Whereas in ileum large interindividual variation was observed, leading to inconclusive data (Fig. 3D).

Comparison of enteroid monolayers with Ussing chamber and Caco-2 monolayers
Subsequently, we compared enteroid monolayer derived P app values and ER to the more conventional models; Ussing chamber and Caco-2 monolayers (Table 1, Supplementary Figure A.2 + 3).TEER measurements in enteroids showed variation between donors but median TEER  For all models enalaprilat P app and ER appeared similar between transport directions, with a tendency for the enteroids towards an ER around 2. A similar pattern was seen for propranolol transport.Nevertheless, enalaprilat P app appeared lower and propranolol P app higher in enteroid and Caco-2 monolayers compared to tissue, indicating differences between ex vivo and in vitro models (Table 1).For talinolol and rosuvastatin transport, A-B transport appeared 3-to 10-fold lower in enteroids and Caco-2, while B-to-A P app values lay within the same range in comparison to tissue, resulting in high ERs.Pairwise donor comparison for P app in B-to-A direction of talinolol and rosuvastatin can be found in Fig. 4. Most donor pairs show (in proportion) comparable P app values in tissue as in enteroid monolayers.All model ERs are in line with active efflux by talinolol.The ER of talinolol in Caco-2 cells was in line with ileum enteroid monolayers.ERs for rosuvastatin appeared higher in Caco-2 compared to enteroids.Determined median P app values and ERs are summarized in Table 1.

Intestinal gene expression in enteroids
To further explore if enteroid monolayers maintain tissue properties, we performed RNA sequencing on 3 jejunum and 7 ileum tissues as well as 3 jejunum-derived and 5 ileum-derived enteroid monolayers, of which 7 were paired donor samples.Based on global RNA expression profiles, enteroids monolayers and tissue were separated (PCA1: 86%) by principal component analysis (PCA).Intestinal regions were separated for both tissue (PC1: 41%) and enteroids (PC1: 46%) (Figure A.4). Expression profiles of proximal (GATA4) and distal (FABP6, GUCA2A, SLC10A2) gut markers confirmed the regional phenotype of the tissue and enteroids (Figure A.5) (Comelli et al., 2009;Middendorp et al., 2014a;Muller et al., 2017;Murata et al., 2023).We explored gene expression of transporters P-gp (ABCB1) and BCRP (ABCG2) in jejunum and ileum in both the tissue and enteroids.In both models, ABCB1 and ABCG2 were abundantly expressed.There appeared to be a trend of higher expression of ABCB1 and ABCG2 appears in ileum enteroids, compared to jejunum enteroids.In contrast, this difference was less well defined in tissue (Fig. 5).A Venn diagram of jejunum versus ileum in tissue and enteroid monolayers also showed more differentially expressed genes (DEGs) in enteroids compared to tissue, indicating that regional differences were more defined in enteroids (Fig. 6).A list with the exact DEGs in tissue and enteroid monolayers is provided in Supplementary Table C.1.

Discussion
Enteroids can be successfully used to study patient-specific absorption pharmacokinetics and toxicokinetics.We have demonstrated the potential of enteroid monolayers derived from the jejunum and ileum as a promising test system for drug transport studies.Cultured 2D enteroids, Ussing chamber and Caco-2 cell studies all exhibited carrier mediated drug efflux by P-gp and BCRP.Moreover, the enteroids maintained trends of region of origin specific drug transport characteristics.
In this study, enteroid monolayers, Ussing chamber and Caco-2 reflected active transport by P-gp and BCRP, indicating a functional intestinal drug efflux barrier.Successful bidirectional transport studies were validated with TEER, FD4 transport, and the passive transport  markers enalaprilat (paracellular) and propranolol (transcellular).B-to-A P app values of talinolol and rosuvastatin fell within a similar range for the three different models and were mostly (in proportion) similar between paired tissue-enteroid donors.Some differences between tissue and enteroid monolayers derived data were observed, which need to be considered when using one of these models.Notably, paracellular transport (enalaprilat P app ) was lower in enteroids compared to tissue, as was A-to-B P app for the efflux transporter substrates talinolol and rosuvastatin, i.e. the effect of the drug efflux transporters (ER) was more prominent in enteroids compared to tissue.The presence of numerous tight junctions in the in vitro monolayer model compared to tissue may account for both, as previously reported in Caco-2 cells by several studies (Artursson et al., 2001;Billat et al., 2017;Fedi et al., 2021).This was also reflected by the higher TEER values observed in enteroid and Caco-2 monolayers compared to fresh tissue (Figure A.1). Next to that, the unavoidable low stirring rate might not have been optimal for transcellular permeability of propranolol in enteroid monolayers.Loss of barrier integrity was observed when the stirring rate was increased, more research might be needed to improve this.To our knowledge, this is the first time that transport experiments in fresh tissue and enteroid monolayers were directly compared.
Several comparisons between enteroid monolayers and Caco-2 cells have been made before, but with major variation in enteroid origin and culture protocols.Previously, it has been shown that biopsy-derived enteroid monolayers more accurately recapitulate the intestinal tissue compared to Caco-2 considering the fraction absorbed and cytochrome P450 expression (Takenaka et al., 2016;Yamashita et al., 2021).Kourula et al. showed significantly higher efflux ratios for digoxin (P-gp substrate) and estrone-3-sulfate (BCRP substrate) in Caco-2 monolayers compared to biopsy-derived duodenal enteroid monolayers (Kourula et al., 2023).We show similar transport and ER of the P-gp substrate talinolol in Caco-2 and ileum enteroid monolayers.For BCRP we did observe a tendency of a higher efflux ratio in Caco-2 cells compared to enteroids.In addition, several reports show promising results for enteroids, such as a direct RNA expression comparison between an enterocyte monolayer derived from induced pluripotent stem cells (iPSC) to Caco-2 cells and biopsies, which showed more in vivo intestinal physiological similarities in iPSCs than in Caco-2 (Kwon et al., 2021;Takahashi et al., 2022).
In this study, enteroid monolayers and Ussing chamber experiments were performed with tissue from jejunum and ileum regions.It is known that gene expression levels of drug transporters differ among jejunum and ileum (Drozdzik et al., 2019;Matthew et al., 2019;Muller et al., 2017;Murata et al., 2023).Results in Fig. 2 and 3 show a potential higher B-to-A transport (P app ) of talinolol in ileum compared to jejunum, although not significant, and also not reflected in ER.We did observe a higher efflux ratio of rosuvastatin by BCRP in ileum enteroid monolayers compared to the jejunum (Fig. 2).For rosuvastatin transport in ileum in the Ussing chamber system, inclusion numbers were low and P app values were variable making it hard to draw conclusions on regional differences.Adjusting the in vitro culture characteristics to a more in vivo like microenvironment might improve regional drug transport even further, e.g.media adjustments, co-culture with commensal bacteria (Ahn et al., 2023;Puschhof et al., 2021).
Transport data derived with the Ussing chamber, especially in the Bto-A direction, are scarce in literature.It is important to note that the Ussing chamber method is known to manifest several challenges.The main limitation of this technique is the complexity of the methodology which is hard to perform in a consistent way, and experimental outcomes are dependent on tissue viability and handling, increasing interlaboratory variability (Sjoberg et al., 2013).For example, previously published Ussing chamber data showed varying results with large standard deviations for propranolol A-to-B transport in jejunum, Michiba et al., 2020, showed a P app of 7.79 ± 4.32 × 10 -6 cm/s (n = 6), Sjoberg et al., 2013, 31.9 ± 17 × 10 -6 cm/s (n = 8), both higher compared to the P app found in this study (1.4 ± 0.8 × 10 -6 cm/s, n = 14) (Michiba et al., 2020;Sjoberg et al., 2013).Furthermore, in ileum previous data from our lab showed A-to-B 11.98 ± 1.28 × 10-6 cm/s in ileum compared to 2.9 ± 1.0 × 10 -6 cm/s in this study (Streekstra et al., 2022).This shows the large variability of this methodology inter-laboratory and therefore makes it challenging to directly compare Ussing chamber and enteroid monolayer derived data.Nonetheless, in vivo determined effective permeability (P eff x10 -4 cm/s) with Loc-I-gut or Triple-Lumen (clinical intestinal perfusion models) showed good correlation with fraction absorbed.The P eff (mean ±%CV) for enalaprilat, propranolol and talinolol were 0.2 ± 150, 3.2 ± 73 and 0.3 ± 171, respectively, also emphasizing the variation in measurements (Dahlgren et al., 2015).
In addition to transport functionality, we examined mRNA expression through RNA sequencing in order to assess similarities between tissue and enteroids.We found more defined regional differences between jejunum and ileum in enteroids compared to tissues.Next, a trend was observed of higher expression of ABCB1 and ABCG2 in ileum compared to jejunum in enteroids (Fig. 4), which was less clear in tissue.The regional differences that were more pronounced in enteroids could be a result of in vivo cell extrinsic microenvironment disappearing upon in vitro culture, which depends more on cell intrinsic factors (Middendorp et al., 2014b).Our data support previous findings by Kraiczy et al., who showed that terminal ileum and sigmoid colon enteroids retained their region-specific signature at both methylation and transcriptional level, and attributed this to the preserved epithelial-cell intrinsic factors in vitro (Kraiczy et al., 2019).They found more overlapping DEGs than we do, likely due to their use of EPCAM-sorted epithelial cells from tissue instead of whole tissue.
In this study, we performed bulk RNA sequencing has been performed on both fresh tissue and enteroid monolayers.The epithelial cell content of the samples taken was not measured before analysis.Hence, we cannot say that epithelial cell input was similar in all tissue samples, which may have influenced the variation between the samples.Future studies could consider cell sorting and single cell RNA sequencing (scRNAseq) for a more powerful analysis.scRNAseq would also provide a way to directly compare epithelial cell expression profiles in tissue with enteroids per cell type (Takayama et al., 2021).
Because of biological variation and donor-donor variation, increased donor numbers would help to strengthen the data and study regional variation and differences between the two models more robustly.More paired Ussing and enteroid monolayer derived data could offer more insight into interindividual variability in enteroids compared to tissue.Notably, it appears that the interindividual variation present in tissue is also represented in the enteroid monolayers.On the one hand this makes interpretation of results more challenging; on the other hand it provides opportunities.A good representation of existing biological variation makes the enteroid model more representative for the human situation and outcome predictions of clinical trials and modeling (Couto et al., 2020).
To summarize, enteroid monolayers and fresh tissue exhibit proper barrier function and active efflux by P-gp and BCRP.Biological variation within tissues and enteroids makes data interpretation challenging, but at the same time allows to include this variation in clinical drug transport research.Hence, the use of enteroid monolayers provide a versatile experimental platform that could complement or ultimately even replace current in vitro models as the widely used Caco-2 model.
We established enteroid monolayers from 5 proximal jejunum and 5 terminal ileum donor tissues to investigate bidirectional drug transport across the intestinal barrier.Representative brightfield and immunofluorescent images are presented in Fig. 1.Accurate barrier formation was confirmed by TEER, FD4 and enalaprilat transport (P app ) in luminal to basolateral direction (A-to-B) (Figure A.1).

Fig. 2 .
Fig. 2. apparent permeability (P app ) and related efflux ratios (ER) in enteroid monolayers in jejunum (circle) and ileum (square) for paracellular transport (enalaprilat), transcellular transport (propranolol) and active efflux by P-gp (P-glycoprotein, talinolol) and BCRP (Breast Cancer Resistance Protein, rosuvastatin).A-D: P app in median ± IQR, A-to-B = apical to basolateral transport.B-to-A = basolateral to apical transport.E: Efflux ratio median ± IQR in jejunum (circle) and ileum (square).Dashed line indicates cut-off ratio of 2 for active efflux transport.Each symbol represents an individual donor.Open circle symbols do not have a paired donor sample in tissue or enteroid group.

Fig. 3 .
Fig. 3. Apparent permeability (P app ) and efflux ratios in tissue derived with the Ussing chamber in jejunum (circle) and ileum (square) for paracellular transport (enalaprilat), transcellular transport (propranolol) and active efflux by P-gp (P-glycoprotein, talinolol) and BCRP (Breast Cancer Resistance Protein, rosuvastatin).A-D: P app in median ± IQR, A-to-B = apical to basolateral transport.B-to-A = basolateral to apical transport.E: Efflux ratios in median ± IQR in jejunum (circle) and ileum (square).Dashed line indicates cut-off ratio of 2 for active efflux transport.Each symbol represents an individual donor.Open circle symbols do not have a paired donor sample in tissue or enteroid group.
E.J.Streekstra et al.European Journal of Pharmaceutical Sciences 201 (2024)  106877was similar to Caco-2, Ussing chamber TEER values were lower compared to the in vitro models(Figure A.1).No difference was observed in FD4 transport (P app ) between the in vitro models and tissue(Figure A.1).

Fig. 4 .
Fig. 4. Apparent permeability (P app ) in B-to-A direction in tissue and enteroid monolayers of matching donor pairs.Each symbol represents an individual donor.A. Talinolol.B. Rosuvastatin.

Fig. 5 .
Fig. 5. Gene expression level counts normalized by DESeq2 of ABCB1 = P-gp, ABCG2 = BCRP.A + B expression in jejunum (n = 3) and ileum (n = 7) tissue, C + D expression in jejunum (n = 3) and ileum (n = 5) enteroid monolayers.Each symbol represents an individual donor.Open circle symbols do not have a paired donor sample in tissue or enteroid group.