Multiple Factors Influence Intestinal Absorption of Drugs

The membrane permeation of drugs is determined not only by passive diffusion mainly influenced by their lipophilicity and molecular weight, but by active transport and metabolism. It is well known that many kinds of drug transporters and metabolic enzymes are expressed in the intestinal epithelial cells and regulate intestinal absorption of substrate drugs. P-glycoprotein (P-gp), multidrug resistance-associated protein 2, and breast cancer resistance protein (BCRP) can recognize a wide variety of drugs with different net charge and chemical structure and suppress the intestinal absorption of substrate drugs by pumping them out from intestinal cells to the intestinal lumen. Their importance has been clarified using transporter knockout mice. Moreover, in humans, drug interactions involving efflux transporters and pharmacogenetics of transporter genes reveal their clinical importance. For example, co-administration of P-gp inhibitor drugs increased the plasma concentration of digoxin, which is a typical P-gp–selective substrate (Fenner et al., 2009), whereas repetitive pre-administration of rifampicin, which can induce mRNA expression of P-gp via the pregnane X receptor (PXR), decreased its plasma concentration (Glaeser et al., 2005). In regard to pharmacogenetics, c.421C>A in ABCG2, which encodes BCRP, increased the plasma concentration of sulfasalazine and rosuvastatin due to the reduced expression/function of BCRP (Zhang et al., 2006; Urquhart et al., 2008). On the other hand, some uptake transporters contribute to enhanced uptake of substrate drugs. For example, peptide transporter 1 (PEPT1), which intrinsically transports di- and tri-peptides, can also recognize drugs such as penicillins, cephalosporins, and angiotensin-converting enzyme inhibitors, and their bioavailabilities tend to be higher than expected based on their physicochemical properties (Terada and Inui, 2004). Valacyclovir, the L-valyl ester prodrug of acyclovir, was intentionally designed as a PEPT1 substrate and its bioavailability was 3- to 5-fold higher than that of acyclovir due to its active uptake via PEPT1 (Yang and Smith, 2013). Proton-coupled folate transporter (PCFT) can transport antifolate drugs such as methotrexate and pemetrexed as well as folate (Matherly et al., 2014). Recent reports suggested that organic anion transporting polypeptide (OATP)2B1 is thought to be a major target of drug interaction between fruit juice and β-blockers/fexofenadine, although the role of OATP2B1 in the intestinal absorption of substrate drugs has not been fully established yet (Yu et al., 2017).

Regarding metabolic enzymes, cytochrome P450 (CYP)3A4 is expressed most abundantly in the small intestine among the CYP isoforms. A previous report suggested that CYP3A4 substrate drugs with high intrinsic metabolic clearance such as tacrolimus, nicardipine, and felodipine, tended to show low F_g values (Kato et al., 2003). For such drugs, inhibition of intestinal CYP3A4 can lead to an increase in their plasma concentration. Many clinical reports have indicated that grapefruit juice continuously increased the plasma concentration of CYP3A4 substrate drugs such as Ca²⁺ channel blockers and midazolam due to the time-dependent inhibition of intestinal CYP3A4 by furanocoumarins (Bailey et al., 1998). UDP-GA glucuronosyl transferase (UGT) family enzymes are also expressed in the intestine to form glucuronides. It has also been reported that F_g values of drugs with high UGT-mediated intrinsic clearance, such as quercetin and raloxifene, are relatively low (Nakamori et al., 2012). Last, carboxylesterase 2 (CES2) is a key enzyme for removing the modified part of prodrug after rapid penetration of the intestinal wall, such as the conversion of irinotecan to SN-38 (its active form) (Parvez et al., 2021).

Difficulty in Predicting Human Intestinal Absorption of Drugs from Animal Data

A possible strategy of predicting intestinal absorption of drugs in humans could be by reference to bioavailability obtained from in vivo animal pharmacokinetic studies. Fig. 1 shows the relationship between the bioavailabilities of multiple drugs in various kinds of experimental animals and humans (Musther et al., 2014). If the bioavailabilities of drugs in humans can be accurately predicted from those obtained from in vivo animal experiments, all such data points should lay on a 1:1 straight line. However, in reality, this is not the case even when data from primates are used. As mentioned, bioavailability includes hepatic metabolic clearance, and previous reports suggest that there is no significant difference in F_a values of various drugs between experimental animals and humans (Chiou and Barve, 1998; Chiou and Buehler, 2002). However, the number of drugs that are substrates of metabolic enzymes and transporters is limited. Takahashi et al. (2009) reported that in a comparison of F_aF_g values of drugs between humans and monkeys, although the predictability of bioavailability was good for drugs that were not substrates for transporters or metabolic enzymes, there were large differences in F_aF_g values for transporter/enzyme substrate drugs. For example, typical CYP3A substrates such as midazolam, verapamil, and nifedipine, rarely penetrate the intestinal epithelial cells to the portal vein (F_aF_g ∼0) in monkeys, whereas F_aF_g values for these drugs are 0.6–0.8 in humans. As for the expression of intestinal efflux transporters, mRNA expression levels of MDR1, multidrug resistance-associated protein 2, and BCRP in monkeys were 20- to 60-fold higher than those in humans (Takahashi et al., 2008). These findings suggest that the monkey intestinal tract seems to be more protective against the penetration of orally ingested compounds into the blood circulation compared with that in humans, and thus it is difficult to directly estimate human F_aF_g values of drugs, even using monkeys, a species that is evolutionarily close to humans.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 1.

Relationship between human bioavailability and animal bioavailability. NHP, non-human primates; CI, confidence interval; PI, prediction interval. Figure reproduced from Musther et al. (2014), Creative Commons CC-BY license 3.0 (https://creativecommons.org/licenses/by/3.0/).

Difficulty in Predicting Human Intestinal Absorption of Drugs from In Vitro Transport Assay with Caco-2 Cells

Another strategy to characterize intestinal absorption of drugs in the process of drug development is to quantify the transcellular transport of drugs across a Caco-2 cell monolayer. Caco-2 cells, an immortalized cell line originally derived from a colon carcinoma, are easy to culture and form a polarized cell monolayer with strong tight junctions after 3-week differentiation in culture, which is suitable for a transcellular transport assay with the culture insert (Table 1). Apparent permeability of drugs without undergoing intestinal metabolism across a Caco-2 cell monolayer (P_app) can be converted to F_a values with the following theoretical equation based on a tube model in the small intestine (Amidon et al., 1988).

A should be calibrated in each laboratory by fitting analyses to the data of multiple drugs that are mainly passively absorbed without any involvement of metabolic enzymes/transporters in the intestine due to the large inter-cell batch difference in absolute P_app values.

An early report indicated that human F_a values were accurately predicted from the results of an in vitro transcellular transport assay with Caco-2 cells, based on the above-mentioned method (Artursson and Karlsson, 1991). Moreover, because some transporters like P-gp are expressed on Caco-2 cells, the impact of P-gp on the transcellular transport of drugs can be qualitatively evaluated by higher basal-to-apical transport than apical-to-basal transport (Adachi et al., 2001). Thus, Caco-2 cells are still routinely used as a gold standard method to evaluate the intestinal absorption of drugs in a high-throughput manner. However, many reports have indicated that expression levels of various kinds of enzymes and transporters in Caco-2 cells differ from those in intact human intestine (Sun et al., 2002; Englund et al., 2006; Seithel et al., 2006; Hilgendorf et al., 2007). For example, mRNA levels of uptake transporters such as PEPT1 and equilibrative nucleoside transporter 1 in Caco-2 cells are more than 10-fold lower than those in intact human duodenum (Sun et al., 2002). Thus, although intestinal absorption of cefadroxil (PEPT1 substrate) is fairly good in humans (F_aF_g ∼0.93) thanks to PEPT1-meditated active uptake despite its hydrophilicity, predicted F_aF_g value from the permeability of cefadroxil across Caco-2 cell monolayer was ∼0.2 (Balimane et al., 2007). Thus, particularly for hydrophilic compounds that are substrates of apical uptake transporters, predicted F_aF_g values tend to be underestimated in Caco-2 cells (Larregieu and Benet, 2013). For metabolic enzymes, it has been reported that mRNA levels of CYP3A4 in Caco-2 cells are less than 1/100 of those in human duodenal epithelial cells, and thus the role of CYP3A4 in the intestinal absorption of drugs cannot be evaluated in Caco-2 cells (Sun et al., 2002). Moreover, because the expression of PXR in Caco-2 cells is reported to be almost negligible, treatment with PXR inducers such as rifampicin failed to increase the expression of CYP3A4 and P-gp (Bruck et al., 2017). These findings suggest that, especially for drugs that are subject to dynamic control of their intestinal absorption by metabolic enzymes and/or transporters, it is often difficult to accurately predict F_aF_g values in humans from in vitro experiments with Caco-2 cells.

Prediction of Intestinal Absorption of Drugs in Humans with the Direct Use of Human-Derived Intestinal Tissue Samples

The author has experienced that commercially available cryopreserved human hepatocytes enables the characterization of the essential role of OATP1B family transporters in the hepatic uptake of anionic drugs such as HMG-CoA reductase inhibitors (statins), and greatly improves the predictability of hepatic intrinsic clearance of OATP1B substrate drugs (Hirano et al., 2004; Maeda, 2015). Thus, it was thought that the direct use of human-derived samples could realize the accurate prediction of intestinal drug absorption in humans as well. Recently, cryopreserved human enterocytes and intestinal mucosa fractions can be obtained from a commercial source (Li, 2020) (Table 1). Davies et al. (2020) have demonstrated that the expression and function of multiple metabolic enzymes are well preserved in human permeabilized enterocytes and intestinal mucosa, and F_g values of high-F_g drugs can be predicted from in vitro experiments with human samples. However, primary enterocytes can only be assayed in a suspension culture format; thus, the impact of each uptake/efflux transporter isoform on the overall transcellular transport of drugs across enterocytes cannot be evaluated and cell viability is only maintained for a short period (∼4 hours) (Li, 2020).

Ussing chamber assay with freshly isolated human intestinal tissues is a powerful tool to simultaneously evaluate the intestinal metabolism and transport of drugs (Table 1). Several studies have demonstrated that permeation of drugs across human intestinal tissue segments mounted onto Ussing chamber followed a theoretical sigmoidal relationship between apparent permeability coefficients of drugs across human intestinal tissues and their Fa values in humans (van de Kerkhof et al., 2006; Rozehnal et al., 2012; Sjoberg et al., 2013). P-gp–mediated efflux or CYP3A4-mediated metabolism can also be assessed in an Ussing chamber assay with human intestinal tissues. However, the availability of fresh human intestinal tissues is generally extremely infrequent and irregular, making their routine use in drug screening difficult. Our research group has succeeded in establishing of a framework in which very fresh upper jejunal tissue segments coexcised at pancreatoduodenectomy at University of Tsukuba Hospital are frequently (typically twice a month) provided. These samples are carefully excised with a minimizing warm ischemic period (<10 minutes) and seamlessly transferred to the laboratory at University of Tsukuba within 30 minutes to perform Ussing chamber assay in our cases. Michiba et al. (2021) have demonstrated that basal-to-apical transport of substrate drugs for P-gp and BCRP was inhibited by a P-gp/BCRP-selective inhibitor cocktail (PSC833 and Ko143) and apical-to-basal transport of substrate drugs for PCFT and PEPT1 was inhibited by their inhibitors, confirming the maintenance of functions of multiple uptake and efflux transporters. Furthermore, all the data points of the membrane permeabilities of several substrate and non-substrate drugs for uptake/efflux transporters were well aligned on the theoretical curve showing the relationship between apical-to-basal apparent permeability of drugs in the Ussing chamber assay and observed human F_aF_g values. Therefore, this assay system directly utilizing fresh human intestinal samples has the potential to comprehensively predict human intestinal absorption properties of drugs, including the involvement of multiple uptake/efflux transporters in overall intestinal transport. However, a major problem in applying this system to drug screening is that the freshness of the intestinal samples has a significant impact on assay outcomes. To determine how long human intestinal samples can be used for transport experiments after their excision, we used tissue samples immediately after isolation and those left in ice-cold buffer for 4 hours and observed apical-to-basal transport of PCFT and PEPT1 substrates (methotrexate and cefadroxil, respectively) and a paracellular transport marker (Lucifer yellow). As a result, a 4-hour incubation decreased the transport of methotrexate and cefadroxil by half and increased the transport of Lucifer yellow by 3-fold (Michiba et al., 2021). Therefore, it appears that a decrease in the transport activity of PEPT1 and PCFT and a decrease in the integrity of tight junctions between cells occurred only 4 hours after the isolation of the human tissues. Although there is no doubt that the use of human-derived samples contributes to an improvement in the predictability of F_aF_g values of drugs in humans, there remain multiple barriers to the routine use of an Ussing chamber assay with freshly isolated intestinal tissue in the process of drug evaluation, such as the irregular availability of fresh tissue, the proximity of the laboratory, and the limited number of drugs that can be evaluated in a single experiment.

Prediction of Intestinal Absorption of Drugs in Humans Using Human-Induced Pluripotent Stem-Derived Enterocytes

Induced pluripotent stem (iPS) cells are a major source of various types of cells in several organs. Many researchers have attempted to construct iPS-derived enterocytes for the screening of intestinal absorption properties of drugs (Table 1). Iwao et al. (2014) established iPS-derived enterocytes via differentiation from endoderm and their enterocytes expressed PEPT1 and CYP3A4, but their expression levels were lower than those in intact human duodenum. Since then, Kabeya et al. (2020) have improved the culture conditions of their iPS-derived enterocytes with the use of forskolin. Under these novel conditions, CYP3A and UGT activity in their enterocytes is much higher than previously reported cells, and mRNA expression levels of most of the metabolic enzymes and transporters are almost comparable to those in the adult human intestine. Ozawa et al. (2015) reported the generation of enterocyte-like cells from human iPS cells and observed that mRNA levels for many transporters were almost consistent with those in adult intestine, but CYP3A4 expression was still lower than in the adult intestine. Thereafter, Negoro et al. (2016) demonstrated a clear inducibility of CYP3A4 in their enterocyte-like cells by PXR inducers such as rifampicin, which was not observed in Caco-2 cells. Takayama et al. (2019) further improved the expression profiles of enzymes and transporters in iPS-derived enterocyte-like cells transduced with two exogenous genes, forkhead box protein A2 and CDX2, by adenovirus vector. Yoshida et al. (2020) established a different differentiation procedure to generate enterocyte-like cells from human iPS cells and demonstrated that expression and function of many kinds of enzymes and transporters are well preserved in their enterocytes and that human F_a values can be predicted from in vitro permeability of drugs across their enterocyte monolayer. However, as discussed below, there are some drawbacks with the use of iPS-derived enterocyte-like cells, including the considerable time to establish iPS-derived enterocyte-like cells from iPS cells and the fact that these cells do not provide information on region-specific gene expression.

Differentiated Intestinal Cells Originated from Cultured Intestinal Stem Cells from Crypts as a Novel Tool for the Prediction of Human Intestinal Absorption of Drugs

To overcome the above-mentioned barriers and facilitate the use of human-derived samples for the prediction of intestinal absorption of drugs, we attempted to establish a cell-based assay system with freshly isolated human tissues that can be used whenever necessary. Recently, human primary-cultured intestinal epithelial cells have become commercially available in very limited quantities, but the allowable number of cell passages is very low and the culture period is extremely limited because of the short lifespan of mature intestinal epithelial cells in nature (Table 1).

However, as part of recent advances in regenerative biology, it was discovered that progenitor cells of intestinal cells exist in the crypt region that can be defined as LGR5-positive cells and various methods for the long-term subculture of intestinal stem cells in a 3D-spheroid format have been developed (Barker et al., 2007; Sato et al., 2011). There, cultured stem cells can be differentiated into intestinal absorptive epithelial cells in 2D culture formats using normal culture plates or culture inserts. For example, VanDussen et al. (2015) reported that it was possible to isolate crypts from a very small amount of human intestinal biopsy samples, expand intestinal stem cells in crypts, and seed them onto a culture insert to create a cell monolayer. Proliferative intestinal stem cells were rapidly expanded using a high concentration of L-WRN cell conditioned media containing Wnt3a, R-spondin, and Noggin, which are critical factors for maintaining the undifferentiated status of the cells. Then, simply by decreasing the concentrations of the three factors, 2D-cultured intestinal stem cells were spontaneously differentiated into various kinds of intestinal cells such as not only absorptive epithelial cells but goblet cells and endocrine cells. There are two major advantages of the intestinal stem cell culture system, which cannot be easily realized in other experimental systems. First, this experimental system has the potential to reproduce regional differences in gene expression and function along the intestinal tract in in vitro experiments (Kozuka et al., 2017; Meran et al., 2020). For example, P-gp and BCRP are more abundantly expressed in the lower region of intestinal tract than in the upper region, whereas CYP3A4 is more abundantly expressed in the upper intestinal tract. However, it is almost impossible to evaluate such regional differences in absorption sites of drugs from in vitro experiments. In contrast, recent reports have suggested that differentiated intestinal cells established from crypts in different regions of the intestine retain the region-specific gene expression profiles of their origin. Therefore, it should be possible to elucidate the relative importance of each site in the overall intestinal absorption of drugs. The other advantage is that a unified culture protocol can be used for the culture of intestinal cells originating from different species. Powell and Behnke (2017) reported that crypt-derived intestinal stem cells from multiple animal species could be maintained in 3D culture using a uniform protocol, suggesting the possibility of reproducing intestinal absorption of drugs in various preclinical species under conditions that minimize experimental artifacts such as difference in culture conditions between cell lines. Such studies would accelerate our understanding of species differences in drug absorption at the molecular level.

The application of human-derived intestinal stem cells to pharmacokinetic research was initiated by the use of 3D ileal enteroids in an induction assay. Stresser et al. (2021) demonstrated that rifampicin treatment successfully induced CYP3A4 expression and metabolic function, but not P-gp, and these data are not consistent with clinical reports. Notably, however, the apical membrane is located inside enteroids, and drugs cannot directly access the apical side of intestinal epithelial cells, which does not correspond to physiologic conditions. On the other hand, various efforts have been made to establish 2D cultures of differentiated intestinal epithelial cells and transcellular transport assays with culture insert like conventional transport assays with Caco-2 cells. Speer et al. (2019a) have demonstrated the transcellular transport of digoxin and prazosin, model P-gp, and BCRP substrates, respectively, across human differentiated intestinal epithelial cells cultured on two different extracellular matrices, and BCRP-mediated vectorial transport of prazosin was observed only when the culture insert was coated with thick extracellular matrix (ECM) hydrogel, but not thin ECM film. These authors also suggested that the protein expression and function of various metabolic enzymes such as CYP3A4 and UGT1A1 in differentiated intestinal cells cultured onto thick ECM hydrogel was comparable to those of intact human tissue (Speer et al., 2019b). These findings suggest that the stiffness of the ECM is one of the key factors in determining the expression and function of transporters and metabolic enzymes possibly due to differential YAP signaling.

In line with the previous report by VanDussen et al. (2015), we have also succeeded in establishing a 3D culture of intestinal stem cells originating from crypts in Matrigel droplets and the differentiation of intestinal stem cells into absorptive epithelial cells in 2D culture simply by removing three factors (Wnt3a, R-spondin 3, and Noggin) that maintain the undifferentiated state from culture medium (Michiba et al., 2022). We have also established culture conditions for crypt-derived intestinal stem cells obtained from different regions of the small intestine in various preclinical species, such as mouse, dog, and monkey. These cells can be cryopreserved like normal cell lines and reseeded in the undifferentiated state, making it a suitable experimental system for drug development because it can be used to prepare the necessary number of cells when needed. We observed that the freshness of tissue samples is critical for the isolation efficiency and proliferation of intestinal stem cells from crypts.

Our studies to evaluate drug intestinal absorption with differentiated intestinal absorptive epithelial cells derived from human samples are summarized in Fig. 2. First, 0.5–2.0 cm² of intestinal tissue sample peeled off the muscle layer is minced and incubated with collagenase for 1 hour. Then, crypts were isolated from tissue fragments by pipetting, filtering through a 100-μm strainer and centrifuging, and the pellets were gently resuspended in Matrigel and cultured with 50% conditioned media of L-WRN cells + 10 μM Y-27632 (ROCK inhibitor) + 10 μM SB431542 (transforming growth factor-β1 receptor inhibitor). 3D-cultured undifferentiated cells that have been increased by passaging culture are then isolated into single cells by digestion with TrypLE Express Enzyme and seeded onto Matrigel-coated culture inserts. Cells were then switched to differentiation medium without Wnt3a, R-spondin 3, and Noggin. In our studies, after approximately 1 week in culture, a cell monolayer is formed, and transepithelial electrical resistance (TEER) became stable at approximately 300–400 Ω × cm², and PEPT1, P-gp, and CYP3A4 mRNA levels were almost as high as those in intact intestinal tissue. In the case of Caco-2 cells, 3 weeks in culture is required for their differentiation under normal culture conditions after seeding on culture inserts, but our cell system can be used for assays after a shorter time. We confirmed that basal-to-apical transport of P-gp and BCRP substrate drugs was significantly reduced in the presence of their inhibitors in our system. In addition, the metabolic activity of CYP3A4, CYP2C9, UGT1A, and CES2 was also confirmed, whereas metabolic activity in Caco-2 cells is very low or negligible compared with intact human intestine.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 2.

Procedure of the isolation of intestinal crypts, expansion culture of intestinal stem cells and differentiation to intestinal absorptive epithelial cells for drug transport/metabolism assay.

We then attempted to predict Fg values in humans by evaluating apical-to-basal transport of CYP3A substrate drugs with different intrinsic metabolic clearance in the absence and presence of CYP3A inhibitors in our cell system (Michiba et al., 2022). As a result, the predicted Fg values of drugs from in vitro assays estimated by two different approaches are well correlated with Fg values reported in humans (Fig. 3). We next tried to evaluate the inducibility of enzymes and transporters by rifampicin because the level of PXR expression is very low in Caco-2 cells, and induction of enzymes and transporters by rifampicin has not been observed in previous reports. In our experiments, pretreatment with rifampicin induced the level of CYP3A4 mRNA approximately 2-fold and increased the metabolic activity of CYP3A accordingly (Michiba et al., unpublished data). These findings are consistent with a previous clinical study demonstrating an approximately 2-fold increase in CYP3A4 expression in human intestinal biopsy samples after repeated dosing with rifampicin (Glaeser et al., 2005).

• Download figure
• Open in new tab
• Download powerpoint

Fig. 3.

Prediction of human Fg values of CYP3A substrate drugs from in vitro transcellular transport assay with human crypt-derived intestinal absorptive epithelial cells. This figure was modified from Michiba et al. (2022) with permission from the American Society for Pharmacology and Experimental Therapeutics.

We also investigated whether regional differences in the expression/function of transporters could be observed in our cell system or not (Michiba et al., unpublished data). Because apical sodium-dependent bile acid transporter (ASBT) expression is limited to the lower region of the small intestine, and PCFT is exclusively expressed in the upper region of the intestine, we checked levels of ASBT and PCFT mRNA in our differentiated intestinal cells that originated from the upper jejunum (upper) and terminal ileum (lower). As expected, expression patterns of ASBT and PCFT mRNA were comparable to those in intact human tissues. Moreover, we confirmed that the Na⁺-dependent uptake of taurocholate (a typical ASBT substrate) was observed only in our cells from the terminal ileum, whereas pH-dependent uptake of methotrexate (a typical PCFT substrate) was observed in our cells from the upper jejunum. These results suggest that regional differences in the intestinal absorption of drugs may be reproduced in our cell system.

In parallel with the development of our cell system, Yamashita et al. (2021) have established a protocol for developing a monolayer platform using human biopsy-derived duodenal organoids. They have checked the mRNA expression of CYP enzymes and transporters, the metabolic activities of CYP3A4 and CES2, and the induction of CYP3A4 and P-gp. Overall, their levels are relatively closer to those in an intact duodenum than Caco-2 cells. Moreover, in transcriptomic analyses, gene expression profiles of the human duodenal organoid-derived monolayer and Caco-2 cells were closest to and farthest from those of the human adult duodenum, respectively. Very recently, Inui et al. (2023b) reported a direct comparison of human biopsy-derived and human iPS cell-derived intestinal organoid monolayers prepared from the same subject. The doubling time of biopsy-derived cells was shorter than that of iPS-derived cells. Gene expression levels of major metabolic enzymes and transporters were higher, and the TEER value was lower in biopsy-derived cells than in iPS-derived cells. On the basis of these results, human tissue-derived differentiated intestinal cells are more suitable for metabolism and transport assays, although further optimization of culture conditions and data collection are required to confirm this prediction.

Conclusions and Future Perspectives

There are currently several experimental systems to quantitatively evaluate the intestinal absorption of drugs. Human crypt-derived differentiated absorptive epithelial cells, in particular, offer an excellent advantage for drug screening because intestinal stem cells stored in liquid nitrogen can be proliferated at any time and easily differentiated into epithelial cells. Human iPS cell-derived enterocyte-like cells are commercially available, and the expression of metabolic enzymes and transporters is relatively well maintained at levels similar to those in intact human tissue. However, our experimental system has the potential to investigate species differences and regional differences in intestinal absorption of drugs, which are difficult to capture in other in vitro experimental systems. In addition, as Inui et al. (2023a) reported, knockout of specific genes in biopsy-derived enterocytes by the CRISPR/Cas9 system can be produced and enables an understanding of the relative contribution of each enzyme and transporter isoform to the overall intestinal absorption of drugs in humans. Although further accumulation of the data will be needed, it is interesting to consider whether intrinsic inter-individual differences in the expression/function of enzymes and transporters and intestinal absorption of drugs in vivo in humans is preserved in crypt-derived differentiated epithelial cells or not by direct comparison of individual F_aF_g values obtained from clinical study with predicted ones from crypt-derived cells obtained from biopsy sample of the same donor.

In this minireview, we basically focused on the intestinal absorption of traditional low-molecular-weight drugs. On the other hand, to fulfill unmet medical needs, new drug modalities such as middle–molecular-weight (∼500–5000) drugs, peptides, antibodies, and nucleic acids have been rapidly developed. The mechanisms of the membrane permeation of high–molecular-weight drugs are quite different from those of low–molecular-weight drugs. Receptor-mediated endocytosis and fluid-phase endocytosis are major mechanisms of cellular uptake of protein drugs, whereas passive membrane transport and transporter-mediated permeation mainly determine the overall cellular uptake of low–molecular-weight drugs. Previous reports showed that active targeting of drug-ligand conjugates or ligand-attached liposomes to the target receptors enabled to efficiently pass through the intestinal epithelial cells via receptor-mediated transcytosis (Russell-Jones, 2001). For example, transepithelial transport of insulin-transferrin conjugates via transferrin receptor-mediated transcytosis was observed in Caco-2 cell monolayers (Xia et al., 2000). Considering receptor-mediated membrane permeation, it is challenging to estimate intestinal absorption of ligands in humans from animal experiments since species differences in the binding affinity of proteins of human origin to the corresponding receptors of different species and subsequent kinetics of intracellular fate of receptors are expected. For example, the serum half-life of mouse IgG in human blood circulation was shorter compared with that of human IgG because of the lack of binding of the human neonatal Fc receptor, which salvages IgG molecules from intracellular degradation, to mouse IgG (Ober et al., 2001). On the other hand, no comprehensive research focuses on the quantitative prediction of the bioavailability of such new drug modalities in vivo from in vitro experimental results, and further research is needed.

Moreover, many reports have been published about the improvement of cellular function with the use of various microphysiological systems to realize dynamic flow of culture media and subsequent shear stress on cells and cyclical mechanical stretch of cell monolayers. For example, Shin et al. (2019) demonstrated that periodical washout of dickkopf-1 (Wnt antagonist) resulted in the promotion of the formation of villi-like structures on Caco-2 cells. Kasendra et al. (2020) demonstrated that human small intestine-on-a-chip containing biopsy-derived intestinal epithelial cells and their CYP3A4 and PXR expression levels were close to those in human duodenum tissue rather than Caco-2 cells. Therefore, with the help of new culture devices, it is expected that various functions in intestinal cells will be further improved, and the application of these systems to various types of drug screening such as drug-induced intestinal toxicity will be expanded.