2023 Volume 48 Issue 5 Pages 251-261
The biliary excretion of pharmaceutical and food-related compounds is an important factor for assessing pharmacokinetics and toxicities in humans, and a highly predictive in vitro method for human biliary excretion is required. We have developed a simple in vitro culture method for generating extended and functional bile canaliculi using cryopreserved human hepatocytes. We evaluated the uptake of compounds by hepatocytes and bile canaliculi, and the biliary excretion index (BEI) was calculated. After 21 days of culture, the presence of extended and functional bile canaliculi was confirmed by the uptake of two fluorescent substrates. Positive BEIs were observed for taurocholic acid-d4, rosuvastatin, pitavastatin, pravastatin, valsartan, olmesartan, and topotecan (reported biliary-excreted compounds in humans), but no difference in BEI was observed for salicylic acid (a nonbiliary-excreted compound). Furthermore, 8 of 21 food-related compounds with specific structures and reported biliary transporter involvement exhibited positive BEIs. The developed in vitro system was characterized by functional bile canaliculus-like structures, and it could be applied to the prediction of the biliary excretion of pharmaceutical and food-related compounds.
Biliary excretion is an important factor for predicting the pharmacokinetics and toxicities of pharmaceutical and food-related compounds in humans. Indeed, various drug–drug and/or drug–food interactions during biliary excretion were previously reported (Hedman et al., 1991; Lin et al., 2008). These effects might consequently lead to the accumulation of toxic substances as well as hepatotoxicity in some cases (Swift et al., 2010). Therefore, a highly predictive method for human biliary excretion is required.
Various in vivo and in vitro methods have been reported for predicting human biliary excretion. For example, the uses of bile-cannulated animal models were reported in pharmaceuticals (Handler et al., 1994; Ohtsu et al., 2015). However, the biliary excretion process differs among animal species (Ishizuka et al., 1999; Ninomiya et al., 2005), and reductions of animal experiments were warranted. For in vitro assays, membrane vesicle assays were frequently used (Morgan et al., 2013); however, these assays do not include metabolic enzymes and available transporters are limited. Therefore, the use of cell-based assays is warranted. Another in vitro model is the use of sandwich-cultured hepatocytes (Swift et al., 2010). In this technique, human- or animal-derived hepatocytes are seeded on a collagen-coated plate, and then, Matrigel is overlaid on the cells to form the hepatocyte monolayer. This culture method produces polarized hepatocytes (apical and basolateral sides) that form cellular tight junctions and bile canaliculi. Using freshly prepared human hepatocytes, Abe et al. reported that this culture method predicted human biliary clearance with good correlation. However, the estimated clearance values from in vitro experiments were approximately 20-fold lower than the observed human biliary clearance (Abe et al., 2009). Cryopreserved human hepatocytes can be used, but it is difficult to form functional and mesh-patterned bile canaliculi resembling those in the human body using such cells (Horiuchi et al., 2022).
To more precisely predict human biliary excretion, various improvements have been reported in terms of cells and culture methods. Concerning cellular improvements, one example is the use of hepatocytes isolated from chimeric humanized mice (Watari et al., 2018). This system maintains cytochrome P450 (CYP) activity up to 21 days after seeding. However, there are problems regarding animal ethics and contamination with host animal-derived hepatocytes. Regarding culture method improvements, the formation of spheroids has been reported (Tetsuka et al., 2017). However, these assay formats require experimental skills, and quantitative and reproducible assay results are difficult to obtain. Recently, Horiuchi et al. (2022) reported the development of a simple and long-term in vitro culture method for forming bile canaliculi using human induced pluripotent stem cell-derived hepatocytes. They observed the stable formation of extended functional bile canaliculi and the utility of this model for predicting cholestasis. However, human induced pluripotent stem cell-derived hepatocytes display differences in metabolic gene and transporter expression and/or activities (Sakai et al., 2019). In this study, we applied this simple in vitro culture method to form bile canalicular structures using commercially available human cryopreserved hepatocytes instead of human induced pluripotent stem cell-derived hepatocytes and evaluated the usefulness of the method for predicting biliary excretion using compounds known to undergo biliary excretion in humans. Furthermore, we evaluated 21 food-related compounds and discussed the potential biliary excretion of these compounds.
Human cryopreserved hepatocytes (Lots HC2-50 and HC5-55), an OptiThaw hepatocyte isolation kit, and OptiPlate hepatocyte plating medium were obtained from Sekisui XenoTech (Kansas City, KS, USA). Primary hepatocyte maintenance supplement pack, William’s E medium, RPMI1640 medium, B27, HepExtend supplement, and gentamicin were purchased from Thermo Fisher Scientific (Waltham, MA, USA). Matrigel and cholyl-lysyl-fluorescein (CLF) were obtained from Corning (Corning, NY, USA). Cellartis enhanced hiPS-HEP long-term maintenance medium (long-term medium) was obtained from Takara Bio (Shiga, Japan). iCell hepatocyte supplement, 5-(and-6)-carboxy-2′,7′-dichloro-fluorescein diacetate (CDFDA), Hanks’ balanced salt solution (HBSS) with Ca2+ and Mg2+, and HBSS without Ca2+ and Mg2+ were procured from Fujifilm Wako (Osaka, Japan). Ethylene glycol tetraacetic acid was obtained from Sigma-Aldrich (St. Louis, MO, USA).
Drugs, food-related compounds, and metabolites were dissolved in DMSO and stored at −20°C until use. Daidzein and genistein were purchased from LC Laboratories (Woburn, MA, USA). Rosuvastatin, pravastatin, pitavastatin, phenacetin, diclofenac, testosterone, curcumin, epicatechin, ferulic acid, epigallocatechin, caffeine, bisphenol A, ochratoxin A, chlorpyrifos, and 2-amino-3,8-dimethylimidazo[4,5-f]quinoxaline (MeIQx) were purchased from Fujifilm Wako. Acrylamide, fenitrothion, butylated hydroxyanisole, 4-hydroxymephenytoin, 4′-hydroxydiclofenac, and 6β-hydroxytestosterone were purchased from Sigma-Aldrich. Cyclosporin A, valsartan, olmesartan, topotecan, bisphenol S, picloram, and acetaminophen were purchased from Tokyo Chemical Industry (Tokyo, Japan). Caffeic acid, gallic acid, quercetin, and salicylic acid were purchased from Nacalai Tesque (Kyoto, Japan). Taurocholic acid-d4, MK-571, and (S)-mephenytoin were purchased from Cayman Chemical (Ann Arbor, MI, USA). Cyanidin 3-glucoside was purchased from Tokiwa Phytochemical (Chiba, Japan). Aflatoxin B1 was purchased from LKT Labs (St. Paul, MN, USA).Hepatocyte culture for forming bile canaliculi
This culture method followed a previous report (Horiuchi et al., 2022) with minor modifications. Human cryopreserved hepatocytes were thawed using an OptiThaw hepatocyte isolation kit according to the manufacturer’s protocol and seeded on collagen-coated 24-well plates at a density of 3.6 × 105 cells/well with OptiPlate hepatocyte plating medium. After 4–6 hr, medium was replaced with 0.5 mL of primary hepatocyte maintenance supplement in William’s E medium. On the next day (Day 1) and Day 7, the medium was replaced with long-term medium containing 0.25 mg/mL Matrigel. On Days 2, 4, and 9, the medium was replaced with long-term medium. On Days 11, 14, 16, and 18, the medium was replaced with hepatocyte maintenance medium consisting of iCell hepatocyte supplement, B27, HepExtend supplement, 25 µg/mL gentamicin, and 0.1 µM dexamethasone in RPMI1640 medium.Fluorescence microscopy
After 21 days of culture, cells cultured in hepatocyte maintenance medium were treated with 4 µM CDFDA for 10 min or 5 µM CLF for 30 min. Then, cells were washed three times with 500 µL of warmed HBSS containing Ca2+ and Mg2+ and further incubated for 10 min in hepatocyte maintenance medium. Finally, the medium was replaced with 200 µL of fresh hepatocyte maintenance medium, and cells were monitored using a fluorescent microscope (EVOS FL, Thermo Fisher Scientific) with excitation at 470 nm and emission at 525 nm.
For the transporter inhibition assay, cells were pretreated with 50 µM MK-571 or 10 µM cyclosporin A for 30 min before CDFDA or CLF uptake. Subsequent processes throughout the experiment (uptake of fluorescent substrates, washing) were performed in the presence of either inhibitor.Measurements of gene expression and CYP activity
Gene expression of CYP1A2, CYP2C9, CYP2C19, CYP3A4, multidrug resistance protein 2 (MRP2), and bile salt export pump (BSEP) on Days 0, 2, and 21 was measured as described previously (Horiuchi et al., 2022). Gene expression levels are presented as relative values to the human liver-derived RNA levels (BioChain Institute, Newark, CA, USA). CYP1A2, CYP2C9, CYP2C19, and CYP3A4 activities were measured on Days 0, 2, and 21. Cells were exposed for 30 min to CYP substrate cocktail solution containing 50 µM phenacetin (CYP1A2), 5 µM diclofenac (CYP2C9), 50 µM mephenytoin (CYP2C19), and 50 µM testosterone (CYP3A4) with 0.5 mL of primary hepatocyte maintenance supplement in William’s E medium. Then, the medium was sampled and diluted with an equal volume of ice-cold acetonitrile. The samples were centrifuged at 10,000 × g for 10 min at 4°C, and the supernatants were analyzed using liquid chromatography–tandem mass spectrometry (LC-MS/MS). The cells were lysed with cell lysis buffer (Cell Signaling Technology, Danvers, MA, USA) and centrifuged 10,000 × g for 10 min at 4°C. The protein concentrations of the lysates were quantified using the Pierce BCA protein assay (Thermo Fisher Scientific).Transport studies in sandwich-cultured human hepatocytes
After 21 days, the cells were treated with Ca2+-containing or Ca2+-free HBSS for 20 min to maintain or disrupt tight junction, respectively, and then exposed to the test compounds and Ca2+-containing HBSS for 5, 10, or 20 min to evaluate compound uptake by hepatocytes and bile canaliculi or by hepatocytes alone. The concentrations of drugs ranged 0.5–2.5 µM according to previous reports (Abe et al., 2009; Fardel et al., 2019), and food-related compounds were used at 1 or 10 µM depending on the detection sensitivity of LC-MS/MS. Then, the cells were washed three times with ice-cold HBSS, lysed with cell lysis buffer, and centrifuged at 10,000 × g for 10 min at 4°C. The protein concentrations of the supernatant were quantified by the Pierce BCA protein assay, the supernatant was diluted with formic acid/water/acetonitrile (0.1/50/50, v/v/v), and the compound concentrations were determined by LC-MS/MS.
For the transporter inhibition assay, cells were pretreated with 10 µM cyclosporin A for 30 min, and transport studies were conducted as described previously with all of the buffers containing 10 µM cyclosporin A throughout the study.Data analysis
Data were expressed as mean ± standard deviation. All statistical analyses were conducted using JMP version 16 (SAS Institute, Cary, NC, USA). A value of p < 0.05 was considered statistically significant.
The biliary excretion index (BEI, %) and biliary clearance (mL/min/mg protein) were calculated as previously described (Abe et al., 2009) as follows:
BEI = (Accumulationcells + bile − Accumulationcells)/Accumulationcells + bile × 100
Biliary clearance = (Accumulationcells + bile − Accumulationcells)/AUCmedium
In the formulas, Accumulationcells + bile (pmol/mg protein) was the cell lysate concentration with Ca2+-containing HBSS treatment before compound uptake, Accumulationcells (pmol/mg protein) was the cell lysate concentration with Ca2+-free HBSS treatment before compound uptake, and AUCmedium (min × µM) was determined as the product of the incubation time and initial compound concentration.
For comparison of biliary clearance between experimental values and reported human values, linear regressions were analyzed using Excel version 2210.
We performed fluorescent microscopy to evaluate the incorporation of the fluorescent substrates CDFDA (MRP2 substrate) and CLF (BSEP substrate) using the procedure described in previous reports (Zuo et al., 2017; Qiao et al., 2021). As presented in Fig. 1, CDFDA and CLF were incorporated into the bile canaliculi, and the fluorescence of each substrate displayed a mesh-like pattern. Conversely, pretreatment with MK-571 (MRP-2 inhibitor) or cyclosporin A (BSEP inhibitor) decreased the fluorescence of these substrates, while the selectivity of these compounds was limited (Matsson et al., 2009; Yoshida et al., 2012). Therefore, the formed bile canaliculi were extended and functional, in line with the findings human hepatocytes.
Accumulation of fluorescent substrates into bile canaliculi in sandwich-cultured human hepatocytes. Cryopreserved human hepatocytes were sandwich-cultured in long-term maintenance medium for 11 days followed by incubation with iCell hepatocyte supplement, B27, and HepExtend in RPMI1640 medium for 10 days. After 21 days of culture, the incorporation of the fluorescent substrates (A) 5-(and-6)-carboxy-2′,7′-dichloro-fluorescein diacetate (CDFDA) and (B) cholyl-lysyl-fluorescein (CLF) into bile canaliculi was observed. The fluorescence of (C) CDFDA and (D) CLF in bile canaliculi was decreased by pretreatment with MK-571 (multidrug resistance protein 2 inhibitor) and cyclosporin A (bile salt export pump inhibitor), respectively. Lot of cryopreserved human hepatocyte was HC2-50.
The mRNA expression of representative CYPs (CYP1A2, CYP2C9, CYP2C19, and CYP3A4) and bile canaliculus-related transporters (MRP2 and BSEP) was measured. As presented in Fig. 2, the expression of CYP1A2, CYP2C9, CYP2C19, and CYP3A4 was unchanged or slightly decreased on Days 2 and 21 compared to that on Day 0. The gene expression of MRP2 and BSEP was higher on Day 21 than on Days 0 and 2.
Gene expression of hepatic cytochrome P450 (CYP) and biliary efflux transporters in sandwich-cultured human hepatocytes. Expression of representative CYPs (CYP1A2, CYP2C9, CYP2C19, and CYP3A4) and biliary efflux transporters (multidrug resistance protein 2 [MRP2] and bile salt export pump [BSEP]) was measured using quantitative polymerase chain reaction on Days 0, 2, and 21. The graphs present the relative expression of each gene. Pooled RNA from human livers was used for the standard curve, and the expression was set as 1. The relative expression was calculated using the equation of the line for the standard curve. Data are presented as the mean ± standard deviation (n = 3). *: p < 0.05 and n.s., not significant vs. day 0 (Dunnett’s test). Lot of cryopreserved human hepatocyte was HC2-50.
The metabolic activity of CYPs was measured, and these results are presented in Table 1. Although the activity of CYP1A2 was maintained up to Day 21 (107 ± 7.5% versus Day 0), the metabolic activity of CYP2C9, CYP2C19, and CYP3A4 tended to be lower on Days 2 (112 ± 15%, 86 ± 13%, and 40 ± 1.7%, respectively) and 21 (21 ± 4.7%, 11 ± 2.1%, and 5.8 ± 0.35%, respectively) than on Day 0.
We assessed the biliary excretion of known positive and negative control compounds in reference to a previous report (Abe et al., 2009). First, the time-course biliary accumulation of taurocholic acid-d4 was assessed in our culture method on Day 21 (Fig. 3). Time-dependent hepatic incorporation, biliary accumulation, and positive BEIs were observed. Next, the incorporation of taurocholic acid-d4 and rosuvastatin (two established biliary-excreted compounds in humans) and salicylic acid (a nonbiliary-excreted compound) was evaluated using two different lots (Fig. 4 and Table 2). Taurocholic acid-d4 and rosuvastatin displayed similar BEIs in the two lots (46% and 37%, respectively, for taurocholic acid-d4; 22% and 21%, respectively, for rosuvastatin), whereas BEI was not positive for salicylic acid in either lot. Although statistically significant differences for rosuvastatin treatment were not observed in the HC5-55 lot, the BEI values were similar in both lots of hepatocytes. These results are in line with a previous report (Fardel et al., 2019). Finally, the accumulation of additional drugs was evaluated by using lot HC2-50 as presented in Table 2. The biliary accumulation of taurocholic acid-d4 was almost completely diminished by the coadministration of cyclosporin A (BEI decreased from 46% to 0.50%). Pravastatin, olmesartan, and topotecan, which are anionic and/or known biliary-excreted drugs, displayed positive BEIs as previously reported (Abe et al., 2009; Fardel et al., 2019). Pitavastatin and valsartan also exhibited positive BEIs, but the degrees were relatively lower than those of the other positive control drugs. These in vitro parameters were converted to human in vivo biliary clearance values in accordance with previous research (Abe et al., 2009), and their relationships were compared. The results are summarized in Table 2 and Supplemental Fig. Although only a limited number of compounds were examined, the estimated human biliary clearance was improved by approximately 2-fold compared to the previously reported findings.
Time-course accumulation of taurocholic acid-d4 in sandwich-cultured human hepatocytes. Cells were treated with Ca2+-containing or Ca2+-free Hanks’ balanced salt solution buffer to maintain or disrupt bile canaliculi, respectively, and taurocholic acid-d4 (2.5 µM) was incorporated for 5, 10, and 20 min. The compound and protein concentrations in hepatocyte lysates were quantified. Data are presented as the mean ± standard deviation (n = 3). *: p < 0.05 vs. with Ca2+-containing buffer (Student’s t-test). Lot of cryopreserved human hepatocyte was HC2-50.
Accumulation of taurocholic acid-d4, rosuvastatin, and salicylic acid in sandwich-cultured human hepatocytes in two lots of cryopreserved human hepatocytes. Cells were treated with Ca2+-containing or Ca2+-free Hanks’ balanced salt solution buffer to maintain or disrupt bile canaliculi, respectively, and taurocholic acid-d4 (2.5 µM), rosuvastatin (0.5 µM), or salicylic acid (2 µM) was incorporated for 10 min. The compound and protein concentrations in hepatocyte lysates were quantified. Data are presented as the mean ± standard deviation (n = 3). *: p < 0.05 and n.s., not significant vs. with Ca2+-containing buffer (Student’s t-test).
Using this culture method, the biliary excretion of 21 food-related compounds was evaluated by using lot HC2-50 (Table 3). Eight of these compounds had positive BEIs, because their values were higher than the threshold value of 8.2% reported for pitavastatin, a well-known biliary excreted drug that showed the lowest BEI value (see Table 2). These eight food-related compounds were epicatechin (25%), gallic acid (52%), caffeic acid (28%), ferulic acid (33%), acrylamide (9.5%), MeIQx (22%), ochratoxin A (24%), and picloram (38%).
Biliary excretion is an important factor for predicting the safety and pharmacokinetics of pharmaceuticals and food-related compounds. Cryopreserved human hepatocytes are widely used as the gold standard for such assessments, especially in the fields of metabolism and toxicity prediction, because they are readily available and maintain drug-metabolizing activity. However, it is difficult to generate functional bile canaliculi with mesh-like morphology using conventional culture methods, and the application of these methods to the quantitative prediction of biliary excretion has been limited. In this study, we succeeded in forming extended and functional bile canaliculi with similarity to the findings in the human liver by appropriately combining commercially available long-term culture media and cryopreserved human hepatocytes. In addition, the gene expression of MRP2 and BSEP, both of which are involved in biliary excretion, was higher on Day 21 than on Days 0 and 2. The overall expression patterns of MRP2 and BSEP were similar to the patterns reported in a previous study in human induced pluripotent stem cell-derived hepatocytes (Horiuchi et al., 2022); therefore, this method can also be applied to cryopreserved human hepatocytes. Although we have not examined the expression of the molecular markers associated with the development and maintenance of the bile canaliculi, Horiuchi et al. used this culture method to investigate the mechanism of bile canaliculi formation (manuscript in preparation). The authors observed a simultaneous increase in the expression of the rate-limiting bile acid synthesis CYP7A1 gene and increase in the level of bile acids. Since bile acid synthesis is involved in the development of the bile canaliculi, the authors suggest that these observations indicate the underlying mechanism to improve this culture method.
Previously, Horiuchi et al. reported that the gene expression of CYP1A2 and CYP2C9 are lower and expression of CYP2C19 and CYP3A4 are similar between human induced pluripotent stem cell-derived hepatocytes using this culture method and cryopreserved human hepatocytes by using vendor-recommended culture protocol. In terms of the expression of CYP genes, the expression of CYP1A2 and CYP2C9 were higher and the expression of CYP2C19 and CYP3A4 were similar when compared with their expression in the human induced pluripotent stem cell-derived hepatocytes (Horiuchi et al., 2022). Therefore, it can be suggested that the gene expression of representative CYPs is maintained in this culture method. Other previous reports demonstrated that CYP activities were undetectable with long-term culture (Asplund et al., 2017; Bell et al., 2018). In this study, the gene expression of four representative CYP subtypes was similar or higher on Day 21 than on Day 0; the activity of CYP1A2 was comparable to Day 0, and the activities of CYP2C9, CYP2C19, and CYP3A4 were lower than on Day 0 but detectable on Day 21. Therefore, this culture method at least more readily maintained CYP activities than observed in previous reports and potential for predicting metabolism of CYP1A2 and biliary excretion. Although it is unclear why the gene expression and activities of CYPs were not concordant, it is well known that CYP activities in cryopreserved human hepatocytes decrease with prolonged culture, and this observation is not unique to this culture method. Handin et al. discussed that the correlation between expression and function varies from one ADME protein to another. Differences in transcriptional regulation, access to co-factors, such as NADPH, and expression and half-life of drug-metabolizing enzymes influence the CYP metabolic activity, resulting in varying degrees of correlation (Handin et al., 2021). Further modification and optimization of the culture method might be needed to evaluate metabolism of CYPs other than CYP1A2 and biliary excretion in the liver simultaneously. The activities of other hepatic metabolizing enzymes (e.g., UDP-glucuronosyl transferases) were not assessed in the present study. A further analysis of the proposed culture method in this study to a broader extent (e.g., UDP-glucuronosyl transferases, other CYP isoforms, etc.) is also needed to fully reproduce the biliary excretion process in vivo.
Biliary excretion was confirmed by BEIs in line with previous reports (Abe et al., 2009; Fardel et al., 2019). Pretreatment with cyclosporin A decreased the accumulation of taurocholic acid-d4 in bile canaliculi and decreased its uptake by hepatocytes. This was observed because cyclosporin A inhibits both BSEP and sodium-taurocholate cotransporting polypeptide, a transporter involved in hepatic uptake of taurocholic acid (Mita et al., 2006). We also detected a positive BEI value for topotecan (31%), a substrate of the P-glycoprotein (P-gp), and breast cancer-resistant protein (BCRP), which are biliary excretion transporters (Matsuda et al., 2013). Although we did not evaluate the expression and activities of these transporters, our results suggest that the culture method could be used to detect P-gp and BCRP as well as MRP2 and BSEP. Pitavastatin and valsartan had lower BEIs than the other positive controls. Pitavastatin is reported to be excreted via bile, but the exact biliary clearance has not been reported. BEI of valsartan was lower than previously reported because valsartan has higher hepatic uptake and relatively lower biliary clearance; therefore, assay variability hinders this effect.
In the conventional sandwich-culture method, the in vitro-based calculated biliary clearance using human fresh hepatocytes with short-term culture was approximately 20-fold lower than that in humans (Abe et al., 2009). Although the BEIs were comparable or slightly lower in this study than in previous reports, the estimated human biliary clearance was approximately 10-fold lower, suggesting improvement. Combined with the fact that the bile canaliculi were extended under fluorescence microscopy and Accumulationcells + bile was higher than previously reported, it is suggested that developed culture method using cryopreserved human hepatocytes is capable of accumulating more bile than the conventional culture method using fresh human hepatocytes. The small number of bile canaliculi formed using the conventional method is considered a limitation of this method (Fardel et al., 2019). Therefore, the present culture method might be useful for more precisely predicting human biliary clearance. In addition, owing to the fact that this culture method uses commercially available cells and medium, it could be easily reproduced. Horiuchi et al. applied this culture method to four lots of cryopreserved human hepatocytes and confirmed the excretion of the fluorescent substrate into the bile canaliculi and displayed a mesh-like pattern (manuscript in preparation). Although we have examined the biliary accumulation in this study in two lots of cryopreserved hepatocytes, it is expected that this culture method could be applicable to other hepatocyte lots. Although the predictability was improved versus the conventional culture method, a one-to-one relationship was not obtained. Thus, a correction factor was still necessary to convert human biliary clearance from in vitro to in vivo. The reason for this divergence remained unclear, expressions of functional biliary transporters and/or formation of bile canaliculi might be needed to obtain one-to-one relationship with in vivo biliary clearance values. Further optimization of the culture method and the cells might be required to obtain values closer to those observed in vivo. Examples of possible improvements included the use of human induced pluripotent stem cell-derived hepatocytes and more complicated culture methods such as 3D-based cell culture are necessary to increase bile canaliculi (Horiuchi et al., 2022).
Currently, the development of risk assessment methods that do not rely on animal testing is an important issue in food research. The scarcity of information on pharmacokinetic information in humans, such as biliary excretion, is a major problem that hinders technical applicability in these research fields. Therefore, the present method was applied to predict the biliary excretion of 21 food-related compounds. The evaluated food-related compounds in this study were selected because their pharmacokinetic information in humans is available, but the information regarding biliary excretion was limited. Positive BEIs were observed for eight compounds, suggesting the possibility of biliary excretion. Among these compounds, gallic acid, caffeic acid, ferulic acid, ochratoxin A, and picloram have a carboxylic acid moiety in their structures, whereas acrylamide is an anionic compound (each structure is presented in Supplemental Table). Epicatechin is reported to be a substrate of MRP2, which is involved in biliary excretion (Vaidyanathan and Walle, 2001). The biliary extraction potential of MeIQx has not been reported, but PhIP, IQ, and MeIQ, which have similar structures, have been reported to be excreted in bile in rats (Dietrich et al., 2001; Sjödin and Jägerstad, 1984; Vlaming et al., 2014), suggesting the possibility of biliary excretion. On the other hand, the remaining 13 food-related compounds did not exhibit positive BEI values. These compounds do not have the structural characteristics that could lead to biliary excretion. Moreover, as shown in Supplementary Table, previous reports on MRP2 and BSEP are limited. The current method is considered useful for predicting the biliary excretion of compounds without relying on animal experiments and for predicting the pharmacokinetics and safety of food-related compounds in humans. The developed in vitro system is characterized by functional bile canaliculus-like structures, and it could be applied to predicting the biliary excretion of pharmaceuticals and food-related compounds.Conflict of interest
The authors declare that there is no conflict of interest.