Introduction

Top Abstract Introduction Experimental Procedures Results Discussion References

As far as absorption from the small intestine is concerned, extensive studies have been carried out focusing, in particular, on the transporters responsible for uptake across the brush border membrane (Leibach and Ganapathy, 1996; Tsuji and Tamai, 1996). In addition, it has been shown that the gastrointestinal absorption of amphipathic neutral or cationic drugs is prevented by the presence of multidrug resistant (mdr) 1 P-glycoprotein (P-gp), an efflux transporter located on the brush border membrane (for review, see Hunter and Hirst, 1997; Arimori and Nakano, 1998; Wacher et al., 1998). This hypothesis has been confirmed by the finding that the bioavailability of it substrates (such as pacritaxel) in mdr1a and/or mdr1a/1b knockout mice is much higher than that in wild-type mice (Schinkel et al., 1997; Sparreboom et al., 1997). Moreover, in everted sac studies, it has been shown that the P-gp inhibitors and/or antibodies against P-gp increased the serosal flux and decreased the mucosal flux of P-gp substrates such as rhodamine 123 (Hsing et al., 1992), benzopyrene (Penny and Campbell, 1994), digoxin (Su and Huang, 1996), and etoposide (Leu and Huang, 1995). In Ussing chamber studies, the flux of verapamil and propantheline across the rat jejunum was demonstrated to be preferentially directed toward secretion, although some regional differences in secretion were observed (Saitoh and Aungst, 1995).

In addition to P-gp, it is possible that efflux transporter(s) for organic anions are located on the brush border membrane of intestinal cells. This hypothesis is based on findings from Ussing chamber studies that show that the serosal-to-mucosal flux of cefazolin, phenol red, and calcein exceeded the flux in the opposite direction (Saitoh et al. 1996; Fujita et al., 1997). As far as the efflux of organic anions is concerned, it has been established that many organic anions, including the glutathione and glucuronide conjugates of xenobiotics are excreted into the bile via canalicular multispecific organic anion transporter/multidrug resistance-associated protein 2 (cMOAT/MRP2), a primary active-transporter located on the bile canalicular membrane (Oude Elferink et al., 1995; Keppler and König, 1997; Kusuhara et al., 1998; Suzuki and Sugiyama, 1998). The substrate specificity of this transporter has been studied by comparing the transport activity across the bile canalicular membrane of normal rats with transport deficient (TR) rats or Eisai hyperbilirubinemic rats (EHBR), whose cMOAT/MRP2 is hereditarily defective (Oude Elferink et al., 1995; Keppler and König, 1997; Kusuhara et al., 1998; Suzuki and Sugiyama, 1998). Because cMOAT/MRP2 is expressed in the small intestine of normal rats, but not in mutant rats (Paulusuma et al., 1996; Ito et al., 1997), it is possible that this transporter is responsible for the intestinal excretion of organic anions. Vries et al. (1989) examined this hypothesis using perfused intestine of Wistar and TR rats. The amount of 1-naphtol--D-glucuronide (a cMOAT/MRP2 substrate) excreted into the lumen was almost the same in both strains when the vasculature was perfused with buffer containing 1-naphtol (Vries et al., 1989). This result suggests that cMOAT/MRP2 makes only a minimal contribution to the excretion of this glucuronide in the small intestine (Vries et al., 1989). The purpose of this study is to investigate the role played by cMOAT/MRP2 in the intestinal excretion of glutathione conjugates, typical substrates for this transporter.

	Experimental Procedures

Top Abstract Introduction Experimental Procedures Results Discussion References

Materials. HPLC grade methanol and acetonitrile were purchased from Wako Pure Chemical Industries, Ltd. (Osaka, Japan). 1-Chloro-2,4-dinitrobenzene (CDNB), 1-fluoro-2,4-dinitrobenzene (FDNB), reduced glutathione, L-cysteine, and N-acetyl-L-cysteine were purchased from Wako Pure Chemical Industries, Ltd. Cysteinylglycine was purchased from Sigma Chemical Co. (St. Louis, MO). All other chemicals and reagents were commercial products of analytical grade. Male Sprague-Dawley (SD) rats, 7 to 9 weeks old, were purchased from Charles River Japan Inc. (Tokyo, Japan). Male EHBR of the same age, whose cMOAT/MRP2 is hereditarily defective, were kindly provided by Eisai Pharmaceutical Co., Ltd. (Tokyo, Japan). All animals had free access to water and food and the animal experiments were performed according to the guidelines provided by the Institution Animal Care Committee (Graduate School of Pharmaceutical Science, The University of Tokyo).

Chemical Synthesis. The glutathione-, cysteinylglycine-, cysteine- and N-acetylcysteine-conjugates of FDNB were synthesized with methods reported by Hinchman et al. (1991) (Fig. 1). Briefly, 2,4-dinitrophenyl-S-glutathione (DNP-SG), 2,4-dinitrophenyl-S-cysteinylglycine (DNP-CG), and 2,4-dinitrophenyl-N-acetylcysteine (DNP-Nac) were synthesized as follows. FDNB was added slowly to reduced glutathione, cysteinyl glycine, and N-acetylcysteine dissolved in 1 N KHCO₃. After incubation for 15 min, the solution was filtered and acidified to ~pH 2 with diluted HCl. The precipitate was then collected by vacuum filtration. Recrystallization of DNP-SG and DNP-Nac was performed from boiling water, whereas methanol was used for DNP-CG. 2,4-dinitrophenyl-S-cysteine (DNP-Cys) was synthesized by hydrolysis of DNP-Nac with H₂SO₄ followed by neutralization with NH₄OH.

View larger version (19K): [in this window] [in a new window]   Fig. 1.   Chemical structure of the glutathione conjugate of CDNB and its subsequent metabolites.

In Vivo Studies. Intestinal excretion was studied with an in situ single-pass perfusion technique. SD rats and EHBR underwent bile duct cannulation with polyethylene tubing (PE10, i.d. 0.61 mm; Becton Dickinson & Co., Bedford, MA) and were perfused with phosphate-buffered isotonic saline containing 1 mM acivicin, pH 6.0, from the upper duodenum to the end of the ileum with peristaltic pump at a rate of 0.4 ml/min. The body temperature of the rats was maintained under suitable lighting. CDNB (30 µmol/kg), dissolved in middle chain triglyceride, was administered i.v. through the femoral vein and then plasma, bile, and intestinal perfusates were collected at fixed times. Bile specimens and intestinal perfusates were collected at 0 to 30, 30 to 60, 60 to 120, and 120 to 180 min on ice. Blood was collected at 5, 30, 60, 90, and 180 min. Finally, the liver and gastrointestinal tract were removed and all specimens were treated with HClO₄ and then subjected to HPLC. CL_bile,p and CL_perfusate,p were calculated by dividing the cumulative amount excreted into the bile and intestinal perfusate over the 3-h period by the area under the curve (AUC)_{0-3 h}, respectively. To quantitatively express the transport activity across the bile canalicular membrane of hepatocytes and the brush border membrane of enterocytes, we determined CL_bile,liver and CL_{perfusate,intestine}, respectively. These are defined as the cumulative amount excreted into the bile and intestinal perfusate between 2 and 3 h after initiation of experiments by the liver or the intestine concentration at the end of experiment (3 h), respectively.

Everted Sacs and Ussing Chamber Studies. SD rats and EHBR were anesthetized with ether and were sacrificed by exsanguination from the abdominal aorta. Then, the duodenum, jejunum, ileum, and colon were immediately removed and rinsed in ice-cold saline. Two- and 3-cm segments of intestine were isolated to perform Ussing chamber and everted sac studies, respectively. These tissues were pretreated with ice-cold Krebs' phosphate buffer, pH 6.0, containing 40 mM glucose and 1 mM acivicin to inhibit -glutamyl transferase. Peyer's patches were identified visually, and sections of tissue containing them were not used. For the everted sac studies, the intestinal segments were slid onto a piston and the epithelial surface exposed. These everted segments were ligated at both ends, and pretreated with 2 ml Krebs' phosphate buffer, pH 6.0, containing 40 mM glucose and 1 mM acivicin for 20 min to completely inhibit the -glutamyltransferase, and then they were filled with 200 ml of DNP-SG dissolved in buffer. The solution was gassed with O₂/CO₂ (95:5) at 37°C. Aliquots from inside the sacs (20 ml) and samples from outside the sacs (200 µl) were collected at 0, 20, and 60 min by microsyringe. Finally, the sac was washed with drug-free buffer and homogenized. All specimens were treated with HClO₄ and then subjected to HPLC. The results are expressed as net clearance, which were obtained by dividing the cumulative amount of DNP-SG excreted into the mucosal side by the AUC on the serosal side, and tissue clearance, which were obtained by dividing the cumulative amount of DNP-SG excreted into the mucosal side by the intestinal tissue concentration during the final period.

HPLC Analysis. The HPLC system consisted of a Hitachi model 6000 series liquid chromatograph. Isocratic elution was performed with a Bakerbond NP Octadecyl C18 column (4.6 × 250 mm, 5 µm; J.T. Baker Research Products, Phillipsburg, NJ) with a mobile phase of acetonitrile: 0.01% H₃PO₄ (1:3, v/v) at a flow rate of 1.0 ml/min (Hinchman et al., 1991). Compounds were detected at 365 nm and quantified by the external standard method by the height of the peaks. Retention times for DNP-SG, DNP-CG, DNP-Cys, and DNP-Nac were about 10, 24, 7, and 12 min, respectively.

Northern Hybridization. Northern hybridization was performed as described previously (Ito et al., 1996, 1997). The mRNA was prepared from liver, duodenum, jejunum, ileum, and colon dissected from three SD rats. As a probe for the hybridization, full-length cMOAT/MRP2 was used. ³²P labeled cDNA for glyceraldehyde 3-phosphate dehydrogenase (G3PDH; Clontech Laboratories, Palo Alto, CA) was used to normalize the amount of mRNA. Two micrograms of mRNA was applied to the gel. The intensity of specific bands was quantified from a standard curve with a BAS 2000 system (Fuji Photo Film Co., Ltd., Tokyo, Japan).

	Results

Top Abstract Introduction Experimental Procedures Results Discussion References

In Vivo Studies. Plasma concentration-time profiles of CDNB conjugates after i.v. administration of CDNB (30 µmol/kg) are shown in Fig. 2. Plasma profiles of CDNB metabolites were comparable between SD rats and EHBR (Fig. 2). In both rat strains, the highest plasma level of DNP-SG, followed by DNP-Nac was observed (Fig. 2). Plasma concentrations of DNP-CG and DNP-Cys were much lower than that of DNP-SG and DNP-Nac in both rat strains (Fig. 2). In SD rats, DNP-SG and DNP-Nac were the main metabolites detected in the bile, whereas the biliary excretion of these conjugates was markedly reduced in EHBR (Fig. 2). The amount of DNP-CG and DNP-Cys excreted into the bile was much lower than that of DNP-SG and DNP-Nac in both SD rats and EHBR. As shown in Table 1, the CL_bile,p and CL_bile,liver values of DNP-SG in SD rats were 1.16 µl/min/kg and 7.71 ml/min/kg, respectively, and these were significantly higher than the corresponding values in EHBR (0.11 µl/min/kg and 2.21 ml/min/kg, respectively).

View larger version (26K): [in this window] [in a new window]   Fig. 2.   Time profiles for plasma concentration and cumulative biliary excretion of DNP-SG and its subsequent metabolites. SD rats and EHBR received i.v. injection of CDNB (30 µmol/kg) to examine the plasma concentration (top) and biliary excretion (bottom) of its metabolites. Each point and vertical bar represents the mean ± S.E. of five rats. , DNP-SG; , DNP-CG; , DNP-Cys; ×, DNP-Nac.

View larger version (16K): [in this window] [in a new window]   Fig. 3.   Cumulative intestinal excretion of DNP-SG and its subsequent metabolites. SD rats and EHBR received i.v. injection of CDNB (30 µmol/kg) to examine the intestinal excretion of DNP-SG and its subsequent metabolites. The plasma concentration and biliary excretion of a series of CDNB metabolites are shown in Fig. 2. Each point and vertical bar represents the mean ± S.E. of five rats. , DNP-SG; , DNP-CG; , DNP-Cys; ×, DNP-Nac.

Everted Sac and Ussing Chamber Studies. The efflux of DNP-SG in duodenum, jejunum, ileum, and colon from SD rats and EHBR was compared in everted sac studies (Fig. 4). The net clearance, along with tissue clearance, in SD rats jejunum was higher than that in EHBR (Fig. 4). Thus, we have examined the following to gain insight into the intestinal secretion mechanism of DNP-SG with an Ussing-type chamber. In an in vitro comparison with an Ussing chamber, the intestinal permeability was compared between the mucosal-to-serosal and serosal-to-mucosal directions in SD rats and EHBR. Time profiles for the trans-epithelial transport of DNP-SG across jejunum are shown in Fig. 5. DNP-SG showed greater permeation in the serosal-to-mucosal direction than in the mucosal-to-serosal direction in SD rats, whereas little difference was observed between the two directions in EHBR (Fig. 5). The preferential serosal-to-mucosal transport of DNP-SG was reduced in the presence of metabolic inhibitors (Fig. 6).

View larger version (26K): [in this window] [in a new window]   Fig. 4.   Intestinal excretion clearance of DNP-SG in the everted sac study. The everted sac segments from duodenum, jejunum, ileum, and colon were used to examine the export of DNP-SG. Everted sacs were filled with Krebs' phosphate buffer containing DNP-SG (118 µM) and acivicin (1 mM) to examine the export of this conjugate into the medium bathing the sacs. The results are given as net clearance (top), which were obtained by dividing the cumulative amount of DNP-SG in the bathing medium by the AUC in the sacs, and as tissue clearance (bottom), which were obtained by dividing the efflux rate of DNP-SG into the bathing medium by the intestinal tissue concentration at the end of the experiments. Statistical comparisons were made with Student's t tests. Each value and vertical bar represents the mean ± S.E. of six independent experiments. , SD rats; , EHBR.

View larger version (13K): [in this window] [in a new window]   Fig. 5.   Intestinal excretion of DNP-SG in the Ussing chamber study. Time profiles for the mucosal-to-serosal and serosal-to-mucosal flux of DNP-SG in the jejunum of SD rats and EHBR were examined (left). The studies were performed at 37°C in Krebs' phosphate buffer containing 1 mM acivicin. Experiments were initiated by added DNP-SG (95.8 µM) to the donor compartment. The bar graph represents the permeability clearance, which was obtained by dividing the flux by the initial donor concentration. Statistical comparisons were performed with Student's t tests. Results are shown as the mean ± S.E. of three rats. Squares, mucosal-to-serosal flux; triangles, serosal-to-mucosal flux; open symbols and bars, EHBR; closed symbols and bars, SD rats.

View larger version (16K): [in this window] [in a new window]   Fig. 6.   Effect of metabolic inhibitors in the transcellular transport of DNP-SG across SD rat jejunum. Transcellular transport of DNP-SG across the jejunum of SD rats was investigated with an Ussing chamber. The experimental conditions are the same as that described for Fig. 5, except that metabolic inhibitors (25 mM 3-O-methylglucose, 10 mM NaF, 10 mM NaN3) were added to both compartments. The results of the control experiments were taken from Fig. 5. Statistical comparisons were performed with Student's t tests. Each value and vertical bar represents the mean ± S.E. of three rats. , control experiments; , in the presence of metabolic inhibitors.

View larger version (14K): [in this window] [in a new window]   Fig. 7.   Expression of cMOAT/MRP2. The mRNA levels for cMOAT/MRP2 were determined in duodenum, jejunum, ileum, and colon in SD rats. The expression of cMOAT/MRP2 was normalized by that of G3PDH. Statistical comparisons were performed with Student's t tests. Each value and vertical bar represents the mean ± S.E. of three independent experiments.

	Discussion

Top Abstract Introduction Experimental Procedures Results Discussion References

The previous results of Northern blot analysis indicated the significant expression of cMOAT/MRP2 in the small intestine of SD rats (Paulusma et al., 1996; Ito et al., 1997). In addition, RT-PCR analysis indicated the introduction of a premature stop codon in the open reading frame of cMOAT/MRP2 cDNA prepared from the small intestine of EHBR (Ito et al., 1997). These results show that a comparison between SD rats and EHBR provides us information on the functional significance of cMOAT/MRP2 in this epithelial tissue. In the present study, we investigated the contribution of cMOAT/MRP2 in the small intestinal excretion of a glutathione conjugate (DNP-SG) and its subsequent metabolites (DNP-CG, DNP-Cys, and DNP-Nac) with these two rat strains. By examining the transport across the bile canalicular membrane, it has been demonstrated that leukotriene C₄ (LTC₄) and its metabolites [leukotriene D₄, leukotriene E₄ and N-acetyl leukotriene E₄ (NAcLTE₄)] are substrates for cMOAT/MRP2 (Huber et al., 1987; Ishikawa et al., 1990).

At first, the intestinal excretion of the glutathione conjugate of CDNB and its subsequent metabolites was examined in relation to the biliary excretion in in vivo experiments. After i.v. administration of CDNB to SD rats and EHBR, DNP-SG, along with DNP-Nac, were the main biliary metabolites in SD rats and the excretion of these metabolites was significantly reduced in EHBR (Fig. 2). In contrast, the plasma profile of DNP-SG after i.v. administration of CDNB (30 µmol/kg) was similar in SD rats and EHBR (Fig. 2). These results can be accounted for by considering the fact that the amount of DNP-SG excreted into the bile and intestinal lumen was only 8.7 and 0.87% of the administered dose in SD rats and EHBR, respectively (Fig. 2; Table 1). Although we have not determined the urinary excretion of DNP-SG, it is possible that this excretion pathway plays a predominant role in the elimination of DNP-SG from plasma in both SD rats and EHBR.

Thus, we were able to confirm the marked reduction in the biliary excretion of glutathione conjugates and end products of mercapturic acid pathway in cMOAT/MRP2-deficient animals (Huber et al., 1987). Previously, Huber et al. (1987) demonstrated that [³H]NAcLTE₄ (an end product of the mercapturic acid pathway) was predominantly excreted into the bile after i.v. injection of [³H]LTC₄ (a glutathione conjugate). Moreover, they indicated that [³H]NAcLTE₄ is transported by cMOAT/MRP2, with isolated bile canalicular membrane vesicles (Huber et al., 1987). In contrast, it was revealed that DNP-Nac is not necessarily a good substrate for cMOAT/MRP2; we examined the inhibitory effect of DNP-Nac on cMOAT/MRP2-mediated transport of [³H]DNP-SG in isolated bile canalicular membrane vesicles. The IC₅₀ value of DNP-Nac was >1 mM, which was much higher than the K_m value of DNP-SG (16.5 + 2.4 µM). Therefore, it is plausible that DNP-SG, after being excreted into the bile and intestinal lumen, is metabolized to DNP-Nac. The reduced excretion of DNP-Nac in EHBR may result from the reduced excretion of DNP-SG in this mutant rat strain.

With regard to the intestinal secretion, the excretion of DNP-SG was markedly reduced in EHBR (Fig. 3), suggesting that the intestinal secretion of this conjugate is largely mediated by cMOAT/MRP2. This result was further confirmed by everted sac studies in which a reduction in intestinal secretion clearance was observed in EHBR compared with SD rats (Fig. 4). In particular, DNP-SG efflux from the jejunum of EHBR was significantly reduced compared with that in SD rats (Fig. 4). Because we found no difference in the net and tissue clearances of vinblastine (1 µM) between SD rats and EHBR (data not shown), the suitability of this experimental system was confirmed. Northern blot analyses demonstrated the highest mRNA level of cMOAT/MRP2 in the jejunum, which is in good agreement with the results from the everted sac studies (Fig. 7).

Moreover, in the Ussing chamber studies, DNP-SG exhibited a 1.5-fold greater flux in the serosal-to-mucosal direction than in the mucosal-to-serosal direction in SD rats, whereas no significant difference was observed between these two directions in EHBR (Fig. 5). Because metabolic inhibitors reduced the preferential serosal-to-mucosal flux of DNP-SG in SD rats (Fig. 6), it was suggested that the net apical secretion of DNP-SG across the rat jejunum is mediated by cMOAT/MRP2.

However, the difference in the flux of DNP-SG between SD rats and EHBR in vitro (Figs. 4 and 5) was smaller than that observed in vivo (Fig. 3). These results may be ascribed to the difference in the concentration of DNP-SG. In our in vitro experiments, we used the minimal concentration of DNP-SG that could be detected by HPLC. Although the medium concentration of DNP-SG was ~100 µM, the concentration of this compound in the intestinal tissue in in vitro experiments was found to be 15 nmol/g tissue (~15 µM). Because the unbound concentration of DNP-SG in the tissue should be lower than its K_m value for cMOAT/MRP2 (~20 µM), cMOAT/MRP2 function would be detectable in our in vitro experiments. However, these concentrations were higher than those observed in in vivo experiments (<25 µM; Fig. 2).

Collectively, these results indicate that cMOAT/MRP2, located on the brush border membrane of enterocytes, plays a significant role in the secretory function of glutathione conjugates. The apical localization of cMOAT/MRP2 in the epithelial cells is in good agreement with the fact that this transporter is located on the bile canalicular (apical) membrane of hepatocytes (Büchler et al., 1996; Paulusma et al., 1996). In addition, with antibody against cMOAT/MRP2, Schaub et al. (1997) demonstrated the apical localization of this protein in renal tubular cells. Transfection of rat and human cMOAT/MRP2 cDNA results in the apical expression of cDNA products in MDCK cells (Evers et al., 1998; Kinoshita et al., 1998). Although we have tried to determine the localization of cMOAT/MRP2 in small intestine with the immunohistochemical techniques, it was difficult to clearly demonstrate its apical localization. Because our preliminary experiments indicated that it is easier for us to perform the immunochemical studies in human tissues than rat tissues, we used human specimens. We prepared antiserum against a C-terminal peptide sequence of human cMOAT/MRP2, as described in a previous report (Kartenbeck et al., 1996); Keppler and his collaborators have prepared the antiserum against this peptide (EAG 5) and have used it for the immunohistochemical staining of human liver (Kartenbeck et al., 1996). Although we found that the bile canalicular membrane of human liver is indeed stained with this antiserum (unpublished data) as reported previously (Kartenbeck et al., 1996), the signal for the staining was very weak in the human small intestine. Minimal staining of the small intestine is consistent with our previous observation that the expression level of cMOAT/MRP2 is much weaker in the small intestine compared with the liver (Ito et al., 1997).

In both everted sac and Ussing chamber studies, we found a significant flux of DNP-SG even in EHBR (Figs. 4 and 5). In particular, in the Ussing chamber studies, we found a symmetrical flux of this glutathione conjugate in EHBR (Fig. 5). These results may be accounted for by the hypothesis that other transporters located on both brush border and basolateral membranes are responsible for the excretion of DNP-SG, even in EHBR, although we cannot exclude the possibility that passive diffusion mediates the symmetrical excretion.

Finally, the discrepancy between the results of the present study and those of Vries et al. (1989) needs to be discussed. Vries et al. (1989) found that the small intestinal excretion of 1-naphtol--D-glucuronide into both the mucosal and serosal sides is similar in both TR and Wistar rats after preloading its precursor (1-naphtol) from the serosal side. It is possible that transporter(s) that preferentially accept glucuronide conjugates, rather than glutathione conjugates, may be expressed on the small intestinal brush border membrane of both EHBR and SD rats at a comparable level and, therefore, the contribution of cMOAT/MRP2 to the excretion of glucuronides were poor substrates for this transporter (Hirohashi et al., 1999). However, it has been recently demonstrated that MRP3 is located on the basolateral membrane in both hepatocytes (König et al., 1999; Kool et al., 1999) and enterocytes (K. Oda, J. Shoda, T.H., H.S., and Y.S., unpublished observations). Such a substrate-dependent differential contribution of cMOAT/MRP2 to epithelial excretion was recently demonstrated with respect to the renal excretion of fluorescent dyes (Russel et al., 1999). Russel et al. (1999) found that the urinary excretion of calcein, but not lucifer yellow, was impaired in TR.

In conclusion, the results of the present study suggest that cMOAT/MRP2 plays an important role in the small intestinal excretion of glutathione conjugates. Although the mucosal efflux of several organic anions (such as cefazolin, phenol red, and calcein) has been previously reported (Saitoh et al., 1996; Fujita et al., 1997), and some of them have been identified as substrates for MRP family, the contribution of cMOAT/MRP2 to the excretion of these ligands remains to be clarified.