Identification of the organic anion transporting polypeptides responsible for the hepatic uptake of the major metabolite of epyrifenacil, S‐3100‐CA, in mice

Abstract Epyrifenacil is a novel herbicide that acts as an inhibitor of protoporphyrinogen oxidase (PPO) and produces hepatotoxicity in rodents by inhibiting PPO. Our previous research revealed that the causal substance of hepatotoxicity is S‐3100‐CA, a major metabolite of epyrifenacil, and that human hepatocyte uptake of S‐3100‐CA was significantly lower than rodent one, suggesting less relevant to hepatotoxicity in humans. To clarify the species difference in the uptake of S‐3100‐CA, we focused on organic anion transporting polypeptides (OATPs) and carried out an uptake assay using human, rat, and mouse OATP hepatic isoforms‐expressing 293FT cells. As a result, all the examined OATPs were found to contribute to the S‐3100‐CA uptake, suggesting that the species difference was not due to the differences in selectivity toward OATP isoforms. When [14C]epyrifenacil was administered to mice, the liver concentration of S‐3100‐CA was higher in males than in females. Furthermore, when [14C]epyrifenacil was administered with OATP inhibitors, the liver/plasma ratio of S‐3100‐CA was significantly decreased by rifampicin, an Oatp1a1/Oatp1a4 inhibitor in mice, but not by digoxin, an Oatp1a4‐specific inhibitor. This result indicates that Oatp1a1, the predominant transporter in male mice, is the main contributor to the hepatic transport of S‐3100‐CA, and consequently to the gender difference. Moreover, we conclude that the species difference in the hepatic uptake of S‐3100‐CA observed in our previous research is not due to differences in the selectivity toward OATP isoforms but rather to the significantly higher expression of OATPs which mediate uptake of S‐3100‐CA in rodents than in humans.


| INTRODUC TI ON
including PPIX, which then leads to photoreactions that generate reactive oxygen and thereby to cell membrane destruction. 1,2 PPO is also present in mitochondria of mammals including rodents and humans, and catalyzes the production of heme which is necessary for the biosynthesis of red blood cells and various proteins such as metabolic enzymes. Thus, it is essential for vital activity in all mammals. It is reported that PPO inhibitors can cause toxicity, especially in PPO-generating organs such as the liver, and occasionally induce hepatocellular tumors as a consequence of hepatotoxicity. [3][4][5] In the case of epyrifenacil, a 90-day subchronic toxicity study showed that epyrifenacil produced hepatocyte injury in mice and rats but not in dogs. 6 Since there was a significant species difference, further investigation of hepatotoxicity was considered necessary for the safety assessment of epyrifenacil's hepatotoxicity in humans.
Previously, we conducted several in vitro mechanistic assays and in vivo metabolism experiments for the purpose of evaluating the toxicity of epyrifenacil in humans and elucidating species differences in sensitivity and exposure. In the in vitro metabolism study, the ester bond of epyrifenacil was rapidly hydrolyzed to form a carboxylic acid metabolite named S-3100-CA in mouse, rat, and human liver microsomes. 6 In addition, the in vivo metabolism study also confirmed that S-3100-CA was the most abundant metabolite in both the liver and plasma of rats, whereas epyrifenacil was not detected in these samples (Sakurai et al., unpublished). In the PPO inhibition study, S-3100-CA showed comparable inhibitory activity to that of the parent epyrifenacil. 6 From these results, it was proposed that the hepatotoxic metabolite of epyrifenacil was S-3100-CA. For human safety assessment, the PPO inhibition assay of the human mitochondrial fraction revealed that the inhibitory activity of S-3100-CA against mitochondrial PPO was 10 times higher in mice than in humans. In addition, the amount of S-3100-CA taken up by mouse hepatocytes was 6-13 times higher than that taken up by human hepatocytes. From these results, it was suggested that the hepatotoxicity potential of epyrifenacil is much lower in humans than in rodents.
We focused on one of these two key factors affecting species difference in hepatotoxicity: the mechanism of S-3100-CA accumulation in the liver of rodents. We speculated that transporters should be involved in active hepatic uptake because (1) the major metabolite S-3100-CA likely exists as a carboxylic anion in vivo and (2) the gender difference in hepatic uptake of S-3100-CA was observed in rodents, i.e., the amount of uptake was higher in males than in females. Furthermore, (3) the gender difference in urinary excretion of S-3100-CA was also observed in the rat metabolism study. All the above observations are well in line with the metabolic fate of some known substrates of organic anion transporting polypeptides (OATPs). 7-9 Therefore, we focused on OATPs that exist in both liver and kidney and have expression levels that differ between male and female rodents.
OATPs are reported to transport anionic compounds into tissues, impacting hepatic pharmacokinetics. [10][11][12] About 14% of transporters in mouse liver are OATPs, and Oatp1a1, Oatp1a4, Oatp1b2, and Oatp2b1 are the major isoforms. 13 Notably, there is a gender difference in the expression levels of some OATPs in mice, Oatp1a1 and Oatp1a4 are highly expressed in males and females, respectively. 14 OATPs are also major transporters in human liver. 15,16 OATP1B1, OATP1B2, and OATP2B1 are the major human OATP isoforms.
Although the involvement of OATPs in the hepatic uptake of S-3100-CA was suggested from the previous studies, it was still unknown which of the transporter isoform(s) are the main contributors to S-3100-CA transport. Furthermore, it is unclear whether the species difference in the hepatic uptake of S-3100-CA observed in previous hepatocyte assays can also be explained by the involvement of OATPs in this process.
In this report, we investigate the hepatic uptake of S-3100-CA by OATPs to elucidate the mechanism of hepatic uptake in rodents and to explain the species difference in hepatic uptake between rodents and humans. To identify transporter isoforms, each OATP seen in rodent and human livers was transiently transfected into 293FT cells, and the uptake of S-3100-CA by these transporters was measured. Moreover, to confirm the impact of transporters in mouse pharmacokinetics, we performed an in vivo study of radiolabeled epyrifenacil co-administered with OATP inhibitors.     with 95% humidity and 5% CO 2 until ready for the assay. From the lysate, 1.0 ml was transferred to a vial containing 10 ml of Emulsifier Scintillator Plus to measure the radioactivity in the LSC.

| Uptake assay for radiolabeled substrates via OATPs
The protein concentration of the cell lysate was also measured using a Pierce™ BCA Protein Assay Kit (Thermo Fisher Scientific, Inc.), F I G U R E 1 Chemical structures of radiolabeled, non-radiolabeled epyrifenacil, and metabolites of epyrifenacil which was used to normalize, according to the cell number, the radioactivity in the cell lysate extract. The protein assay was performed in duplicate. Each sample type was assayed in triplicate wells, and each run was performed twice. When the difference in mean uptake between OATP expressing cells and mock cells was more than 2 times, the OATP was judged to be involved in transportation of the substrate. The uptake assay was repeated with OATP inhibitors, RIF, RSV, and DGX. The method was generally similar to the uptake assay described above, except that the substrate solution contained an inhibitor (i.e., at final concentration of 1, 10, and 100 μM of RIF or RSV, or 10 and 100 μM of DGX, with final concentration of dimethyl sulfoxide, 1.0%). Each sample type was assayed in triplicate wells. The inhibition assay was performed only once because the fold-increase of the uptake by the expression of OATPs was comparable among the runs, and thus, it was concluded that the expression of OATPs is stable.

| Treatment of animals
Crl:CD1(ICR) mice at the age of 7 weeks were purchased from Charles River Laboratories Japan, Inc., Hino Breeding Center. The mice at the age of 8 or 9 weeks were dosed after completing a 7-day quarantine period and an additional 0-to 7-day acclimatization pe-

| Blood 14 C-concentrations study
Two mice per sex at 8 weeks old were given a single oral dose of [ 14 C] epyrifenacil at 1 mg/kg. Approximately 0.06-0.08 g of blood sample was periodically taken from the saphenous vein at 1, 2, 4, and 8 h after administration, and transferred to vials to measure the radioactivity in whole blood. The half-life (T 1/2 ) of radioactivity in blood was

| 14 C-plasma, liver, and kidney distribution study
Three mice per sex at 8 weeks old were given a single oral dose of

| Metabolite analysis
Approximately 0.3 g of liver, kidney, and plasma samples were pooled, respectively, for each dose and gender to obtain samples for analysis. Pooled liver, kidney, and plasma (ca. 1 g each) samples were homogenized with an approximately 3-fold volume (v/w) of methanol using the mixer mill MM400 (5 min, 25 cycles/second, Retsch GmbH) and centrifuged (10 000g, 5 min, 4°C) to separate the supernatants from the residues. The extraction was repeated one additional time. The supernatants obtained were pooled and labeled liver, kidney, and plasma extracts. These samples were concentrated and then subjected to TLC analysis. For calculation of extraction efficacy, the post-extraction solids (PESs) from the tissues and plasma extracts were sampled in low potassium glass vials. PESs were solubilized with 3 M potassium hydroxide solution (approximately 1 ml), then bleached with 30% hydrogen peroxide solution, and finally mixed with 10 ml of Hionic-Fluor to measure the radioactivity.

| Data analysis
Average, standard deviation, standard error, and other statistical metrics were calculated using Microsoft Excel 2016.

| Establishment of OATP expressing cells
We established OATP expressing cells using the method above and confirmed their expression of OATPs and the functionality of each OATP by measuring the uptake of two well-known OATP substrates.
OATP expressing cells were assayed with E 2 17βG and E3S and the result was used to calculate the uptake ratios which represent the fold increase in uptake relative to uptake by cells transfected with the control plasmid (mock cells) ( Table 1). All the uptakes of E 2 17βG and E3S by OATP expressing cells were more than twice as large as those by mock cells except for the uptake of E3S via mouse/rat Oatp1a4 (1.1-1.8 fold) and the uptake of E 2 17βG via human OATP2B1 (1.3 fold) when considered with the mean of 2 runs. Those OATPs with increased uptakes was confirmed to be involved in transportation of each substrate. Especially, the uptake of E 2 17βG by human OATP1B1, mouse Oatp1a1, and rat Oatp1a1, and uptake of E3S by OATP1B1 were remarkably high (>50 times higher than in mock cells).

| S-3100-CA uptake into OATP expressing cells
Using the constructed OATP expressing cells, we evaluated S-3100-CA uptake. The uptake ratios with all OATP expressing cells were at least twice that with mock cells (Table 1). These OATPs were judged to be involved in transportation of S-3100-CA. The mean uptake ratios with OATP expressing cells varied among the isoforms but all above 2.0, suggesting that S-3100-CA is transported via various OATP isoforms without typical selectivity toward isoforms.

| Inhibition of S-3100-CA uptake by OATP inhibitors
Inhibition of uptake of S-3100-CA via OATPs was also evaluated using OATP inhibitors (Figure 2). The addition of RIF eliminated the uptake however, the inhibition of uptake via Oatp1a1 by DGX was weak.  respectively. Blood 14 C-concentration was relatively high in females at all sampling time points. AUC 0-∞ was also higher in females despite the similar T 1/2 between males and females. Accordingly, we set the time point of dissection at 2 h after administration, which accounts for the high 14 C exposure in both males and females.

| 14 C-Distribution in plasma, liver, and kidney
14 C-Concentrations were determined in plasma, liver, and kidney at 2 h after single oral administration of [ 14 C]epyrifenacil to male and female mice (Table 3). In mouse plasma, liver, and kidney, no remarkable gender-related differences in 14 C-concentration were observed. Notably, the concentration in liver was 6.763 ± 0.1014 and 5.136 ± 1.1225 ppm for males and females, respectively, which was somewhat higher than concentrations in plasma and kidney. The liver-to-plasma concentration ratios were ca. 73 and 35 in males and females, respectively.

| 14 C-Distribution in plasma and tissues with co-administration of OATP inhibitors
14 C-Concentrations in plasma, liver, and kidney at 2 h after a single oral administration of [ 14 C]epyrifenacil with co-administration of OATP inhibitors are presented in Table 3. After a single oral F I G U R E 2 Inhibitory effect of rifampicin, rifamycin SV, and digoxin at 100 μM for various OATPs that mediate the uptake of S-3100-CA. All data are indicated as the mean value of 3 replicates in a single run Compared to the 14 C-concentration in plasma, the 14 C-concentration in liver was ca. 97 times higher in males and ca. 43 times higher in females. It was confirmed that the liver-to-plasma concentration ratios were not significantly changed by oral administration of DGX.

F I G U R E 3 Concentration-dependent
Intraperitoneal co-administration of DGX at a dose of 100 mg/kg did not notably change both 14 C-concentrations and liver/plasma concentration ratios.

| Quantification of metabolites of in plasma, liver, and kidney
Concentrations (μg eq. of epyrifenacil/g or ppm) of metabolites in plasma, liver, and kidney were determined by TLC (Table 4), and the tissue/plasma ratios of S-3100-CA were calculated (Figure 4). In the plasma, liver, and kidney extracts, the metabolic profiles were almost the same, and 4 metabolites which identified in the rat metab-

| DISCUSS ION
In this study, we investigated the involvement of OATPs in the hepatic uptake of S-3100-CA to elucidate the mechanism of hepatic uptake in rodents and to explain the species difference between rodents and humans. It is commonly recognized that various OATPs are expressed in the livers of mammals, including humans, rats, and mice. 13,16 In this study, we used E 2 17βG and E3S as the positive TA B L E 2 14 C-Concentrations and pharmacokinetic parameters of 14 C in blood after a single oral administration of [ 14 C]epyrifenacil at 1 mg/kg to male and female mice specificity of these OATP isoforms. From these results, we concluded that all OATPs were expressed in transfected cells as expected, and worked to transport the substrates appropriately.
S-3100-CA was also taken up by all OATPs examined in this study. The uptake ratios versus mock cells were at least 2.2, thus it was suggested that S-3100-CA is transported via various OATPs without typical selectivity toward the isoforms. Although the uptake ratios varied among the OATP isoforms, the differences in the kinetics of S-3100-CA uptake via OATP isoforms in hepatocytes or intact liver was not directly indicated, because these results were not corrected by expressed protein levels. Moreover, the results showed no obvious differences in the uptake potential between predominant hepatic OATP isoforms (i.e., Oatp1a1, Oatp1a4, Oatp1b2, Oatp2b1 in mice and OATP1B1, OATP1B3, OATP2B1 in humans). As mentioned above, significant species differences were observed in the hepatic uptake of S-3100-CA. 6 With the currently-available data, the reason of the species differences was considered not due to the selectivity of S-3100-CA toward OATP isoforms.
Prior to the OATP inhibition experiment in vivo, the selectivity of the known OATP inhibitors (RIF, RSV, and DGX) in rats and humans was determined for mouse isoforms because such information has not been well reported. As a result, DGX was revealed to be a selective inhibitor of Oatp1a4 in mice. On the other hand, RIF and RSV inhibited mouse Oatp1a1, mouse Oatp1a4, human OATP1B1, and human OATP1B3. These inhibitory effects on mouse OATPs increased in a dose-dependent manner and reached more than 50% at 10 μM. The selectivity of the inhibitory effects of these inhibitors was similar to that for the rat OATPs reported previously. 29 With these results, we decided to use RIF and DGX in the in vivo study as mouse OATP inhibitors.
After administering [ 14 C]epyrifenacil to mice, we measured 14 C-and metabolite concentrations to confirm that the significant hepatic uptake and its gender differences observed in in vitro hepatocyte assay are reproduced in vivo. 14 C-Concentrations in plasma and kidney of [ 14 C]epyrifenacil dosed mice were slightly higher in females than in males. On the other hand, the concentration in liver was higher in males. As a result, the ratio of liver to plasma 14 Cconcentration was ca. 73 and 35 for males and females, respectively.
There was a gender-related difference in liver/plasma ratio. A metabolite analysis of these fractions identified S-3100-CA as the predominant metabolite with high and concomitant gender difference in liver/plasma ratio; the liver/plasma ratio of S-3100-CA was 88 and 52 for males and females, respectively. This result was consistent with that of the hepatocyte uptake assay, 6 S-3100-CA uptake was high and associated with a gender difference. Some studies revealed a gender difference in the expression level of Oatp1a1 and Oatp1a4 in mouse liver, with Oatp1a1 expression approximately doubled in male mouse liver and Oatp1a4 expression doubled in female mouse liver. 14 These gender differences in the OATP expression were reported to cause the species differences in the hepatic uptake for some OATP substrates, 7,9 thus it was suggested that the higher hepatic uptake of S-3100-CA in male mice was also due to the gender difference in Oatp1a1 expression levels. TA B L E 3 14 C-Concentrations in plasma, liver, and kidney at 2 h after a single oral administration of [ 14 C]epyrifenacil at 1 mg/kg to male and female mice 14 C-Concentration (μg eq. of epyrifenacil/g tissue) a Plasma

Female
Epyrifenacil ( were co-administered, 14 C-concentration in liver became significantly lower than in the livers of control mice without RIF, whereas 14 C-concentration in plasma was slightly increased, resulting in lower liver/plasma ratio of 11 and 4 for males and females, respectively. Co-administration with RIF also decreased S-3100-CA concentration in liver; when compared to the mice without inhibitors, mice with inhibitors had a significantly lower concentration in liver and relatively higher concentration in plasma, and consequently, the liver/plasma ratio was lowered to 18 and 5 for males and females, respectively. Oral or intraperitoneal co-administration of DGX did not notably change the liver/plasma ratio of 14  Taking into consideration this and previously published data about the expression levels of OATPs in mice, it was concluded that the most predominant isoform Oatp1a1 was the main contributor to hepatic uptake of S-3100-CA in mice.
Though oral co-administration of DGX did not decrease the liver/ plasma ratio of S-3100-CA, it did significantly decrease concentrations of S-3100-CA in plasma, liver, and kidney, while intraperitoneal co-administration of DGX did not change these parameters. This result indicates that oral administration of DGX might decrease oral absorption of epyrifenacil. DGX is also known as a P-glycoprotein substrate and it is reported that the absorption of DGX is competitively inhibited by other P-glycoprotein substrates. 30 Thus, in this study, DGX might inhibit the intestinal absorption of epyrifenacil,however, the effect of DGX interaction with epyrifenacil on intestinal absorption via P-glycoprotein was not investigated in this study.
From the observations above, it was suggested that species differences in hepatic uptake between humans and mice were also due to the different expression levels of OATPs. In the previous literature, the plasma membrane protein levels of hepatic OATPs were higher in mouse liver (Oatp1a1: 42.9 ± 2.57 fmol/μg protein, Oatp1a4: 1.65 ± 0.14 fmol/μg protein) 13 than in human livers (OATP1B1: 2.74 ± 3.67 fmol/μg protein, OATP1B3: 1.70 ± 0.45 fmol/μg protein, OATP2B1: 0.463 ± 0.872 fmol/μg protein). 15 In addition, the total membrane protein levels of OATPs were higher in rats (3.91-6.43 fmol/μg protein) than in humans (1.1-2.0 fmol/μg protein). 16 Taken together, the data indicate that the species differences in the hepatic uptake of S-3100-CA between mice and humans is not caused by differences in the selectivity toward OATP isoforms but rather to differences in the expression levels of OATPs in the liver. Based on the knowledge obtained in the current research about the active transport, further quantitative evaluation such as PBPK modeling is anticipated to complete the human safety assessment of S-3100-CA.
In conclusion, this study revealed that hepatic uptake of S-3100-CA was mediated by OATPs, and gender differences in rodents and species differences between rodents and humans in the hepatocyte uptake of S-3100-CA were not due to differences in selectivity toward OATP isoforms but rather to differences in protein expression levels. OATP expressing cells were successfully constructed and used to confirm that S-3100-CA was taken up by all OATPs examined in this study, suggesting that uptake of S-3100-CA was similar among OATP isoforms. When epyrifenacil was administered to mice, S-3100-CA was highly distributed to the liver and there was a gender difference.
In vivo experiments using OATP inhibitors revealed that the most predominant OATP isoform in mice, Oatp1a1, contributes to the hepatic transport of S-3100-CA, and the gender difference in the expression levels of the isoform was consistent with the gender difference in the hepatic uptake of S-3100-CA in mice. These results suggest that the OATP expression levels should be the dominant determinant of the hepatic uptake of S-3100-CA. The significantly lower hepatic uptake in humans than in mice observed in the previously performed in vitro hepatic uptake study could be explained in the same way, that is, by much lower total expression levels of OATPs in humans than in mice.
Consequently, the low hepatic uptake of S-3100-CA in humans resulting in the lower concern with its hepatotoxicity in humans than in mice was mechanistically supported by this research.

ACK N OWLED G M ENT
The authors are members of the Cellular Biochemistry Group in the Environmental Health Science Laboratory, Sumitomo Chemical Co., Ltd.

D I SCLOS U R E
This work received no external funding. The author declares that there is no conflict of interest.

E TH I C S S TATEM ENT
All animal experiments were conducted in accordance with the Guidance for the Care and Use of Laboratory Animals in our laboratory, which corresponds to the Guidelines for Proper Conduct of Animal Experiments (Science Council of Japan). This manuscript is not published or submitted elsewhere.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data that support the findings of this study are available from the corresponding author upon reasonable request.