Abstract
The onset of diarrhea after the administration of mycophenolate mofetil (MMF) is possibly associated with the biliary excretion of its metabolite, mycophenolic acid glucuronide (MPAG). This study was undertaken to clarify the mechanism underlying the biliary excretion of MPAG. Intravenously administered mycophenolic acid (MPA, 5 mg/kg) rapidly disappeared from plasma and was efficiently excreted as MPAG in the bile of Wistar (26% of dose) and Sprague-Dawley rats (21% of dose) over 1 h. On the other hand, in spite of the rapid disappearance of MPA from plasma, the biliary excretion of MPAG was very limited in Eisai hyperbilirubinemic rats (EHBRs), which display mutations in multidrug resistance-associated protein 2 (Mrp2)/canalicular multispecific organic anion transporter, and constituted only 0.5% of dose. Instead, high levels of MPA were noted in the plasma of EHBRs. Intravenous administration of CsA (5 mg/kg) to Wistar rats significantly lowered the biliary excretion of MPAG. However, intravenously administered tacrolimus (0.1 mg/kg) failed to produce such effect. In conclusion, it is suggested that there is an efficient MPAG transport mediated by Mrp2 on the bile canalicular membrane of rat hepatocytes and that the therapeutic range of CsA potentially interferes with Mrp2. However, the therapeutic range of tacrolimus does not inhibit the transporter. Thus, it should be noted that MMF coadministered with tacrolimus instead of CsA might increase the occurrence of diarrhea related to the biliary excretion of MPAG in transplant recipients.
Mycophenolate mofetil (MMF) is a new type of immuno-suppressive agent for the prophylaxis of acute rejection in organ transplantations (Mele and Halloran, 2000). After oral administration, MMF is rapidly converted to mycophenolic acid (MPA), the active metabolite of MMF (Bullingham et al., 1996a). Generally, MMF is used in combination with cyclosporin A (CsA) or tacrolimus, and diarrhea is one of the most frequently reported adverse events in transplant recipients receiving MMF (Behrend, 2001). In Japan, clinical studies of MMF have mainly been performed in combination with CsA, and approximately 13% of subjects enrolled in the clinical studies complained of diarrhea due to MMF according to the interview form supplied by the manufacturer. At Sapporo City General Hospital, more than 200 cases of renal transplantation have been performed to date, and MMF was introduced as an adjunctive therapy in combination with tacrolimus in 2000. Since then, severe diarrhea has been observed in patients immediately after the start of the MMF regimen, leading to discontinuation or reduction of the drug. Very recently, we investigated the frequency and severity of diarrhea between renal transplant recipients receiving MMF (MMF, prednisolone, and tacrolimus) and those not receiving MMF (azathioprine, prednisolone, and tacrolimus or CsA) in the immunosuppressive regimens at Sapporo City General Hospital. Statistical analysis revealed that diarrhea was more frequent and severe in the patients receiving MMF than in those not receiving it (Kobayashi et al., 2003). The frequency of diarrhea in the patients receiving MMF was approximately 74%. Although we could not carry out the same investigation on the immunosuppressive regimen using MMF and CsA, the frequency of diarrhea obtained from our investigation seemed much greater than those reported previously in the clinical studies using MMF and CsA. According to these observations, it is obvious that the choice of a carcineurin inhibitor coadministered with MMF is closely associated with the occurrence of diarrhea.
Mycophenolic acid is primarily metabolized by glucuronidation at the phenolic hydroxyl group in the liver, forming 7-O-glucuronide (MPAG), and the glucuronide is preferentially excreted into the urine (Bullingham et al., 1996a). It was also reported that the extent of enterohepatic circulation in the overall pharmacokinetics for MPA was 37%, with a wide range of 10 to 61%, in humans (Bullingham et al., 1996b). Thus, it is fully assumed that the enterohepatic circulation of MPAG is closely responsible for gastrointestinal disorders such as diarrhea, although the mechanism by which diarrhea occurs after MMF administration has not been clarified. Recent reports have suggested that CsA and tacrolimus differently influence the pharmacokinetics of MPA. For example, MPA predose concentrations were significantly greater in patients receiving MMF and prednisolone than in patients receiving MMF, prednisolone, and CsA, suggesting a clinically relevant drug interaction between CsA and MPA (and/or MMF) (Gregoor et al., 1999). It was also reported that MPA blood concentrations are lower in patients receiving concomitant tacrolimus compared with those receiving CsA (Filler et al., 2000), and MPAG blood concentrations are higher in combination with CsA than with tacrolimus (Zucker et al., 1997). As a possible mechanism, several reports have implied that CsA interferes with the enterohepatic circulation of MPAG (van Gelder et al., 2001; Shipkova et al., 2001) and that tacrolimus can interfere with UDP-glucuronosyltransferase, which mediates the conversion of MPA to MPAG (Zucker et al., 1999). Until now, however, both the mechanism responsible for the biliary excretion of MPAG and the effect of CsA and tacrolimus on the biliary excretion have not been fully characterized.
In this study, we evaluated the biliary excretion of MPA and MPAG after intravenous administration of MPA to two normal rat strains, Wistar and Sprague-Dawley (SD), and Eisai hyperbilirubinemic rats (EHBRs), a mutant SD strain lacking multidrug resistance-associated protein 2 (Mrp2)/canalicular multispecific organic anion transporter on the hepatic canalicular membranes and on the intestinal brush-border membranes (Ito et al., 1997). Our results demonstrated that MPAG was excreted into the bile via Mrp2 and that CsA, but not tacrolimus, functionally inhibited the Mrp2-mediated transport for MPAG in the liver.
Materials and Methods
Materials. Mycophenolic acid and β-glucuronidase solution (85 units/ml) were obtained from Wako Pure Chemicals (Osaka, Japan). (+)-Naproxen was purchased from Sigma-Aldrich (St. Louis, MO). A tacrolimus formulation for injection (Prograf injection 5 mg) and a CsA formulation for injection (Sandimmun) were purchased from Fujisawa Pharmaceutical (Osaka, Japan) and Novartis Pharma (Tokyo, Japan), respectively. Digoxin was obtained from Tokyo Chemical Co. (Tokyo, Japan). All other reagents were of the highest grade available.
Animal Experiments. Male Wistar and SD rats were obtained from Hokudo (Sapporo, Japan). Male EHBRs were purchased from Sankyo Labo Service Co. (Tokyo, Japan). Rats (three per cage) were housed for at least 2 wk before experiments with free access to food (MF, Oriental Yeast Co., Tokyo, Japan) and water at 25 ± 3°C and 50 ± 20% relative humidity under a 12-h light/dark cycle, without no attrition. The body weight of rats used in this study was 260 to 380 g in Wistar rats and 350 to 390 g in EHBR and SD rats, respectively. All rats were used for experiments at age of 8 to 10 wk.
Rats were anesthetized with an intraperitoneal injection of pentobarbital sodium (40 mg/kg; Dainippon Pharmaceuticals, Osaka, Japan). After abdominal operation, a polyethylene tube (i.d. 2.8 mm) was inserted in bile duct toward the liver and then MPA, which was dissolved in polyethylene glycol 400 at a concentration of 5 mg/ml, was administered intravenously to each rat via the jugular vein. The dose of MPA was fixed at 5 mg/kg body weight. Bile samples were collected every 15 min over 1 h after MPA administration, and the bile volume was measured with an appropriately sized volumetric pipette. Blood samples (each 0.4 ml) were taken from the other side of the jugular vein just before MPA administration and at 5, 10, 15, 20, 30, and 60 min after administration, and were immediately centrifuged at 2,700g for 20 min to obtain plasma samples. Bile and plasma samples were kept frozen until assay. In the experiments to examine the effect of CsA or tacrolimus on the pharmacokinetics of MPA, their formulations were diluted with saline including 50% ethanol and administered intravenously via the jugular vein at 5 min before MPA administration. The doses of CsA and tacrolimus were set at 5 and 0.1 mg/kg, respectively, based on their clinical regimens to renal transplant recipients.
In this study, principles of good laboratory animal care were followed and animal experimentation was performed in compliance with the Guidelines for the Care and Use of Laboratory Animals in Health Sciences University of Hokkaido, accredited by the “Principles of Laboratory Animal Care” (NIH publication 85-23, revised 1985).
Drug Assay. The assay of MPA and MPAG was performed according to Svensson et al. (1999) with several modifications. For the determination of free MPA in bile and plasma, 50 μl of sample was mixed with 50 μl of saline and 200 μl of naproxen as an internal standard (5 or 2.5 μg/ml in acetonitrile) and then vigorously shaken for 10 s. The mixture was let stand for 5 min and centrifuged at 2700g for 10 min. An aliquot of the resulting supernatant was applied to high-performance liquid chromatography. The determination of MPAG was done after converting the glucuronide to free MPA by β-glucuronidase; that is, 45 μl of plasma or 90 μl of bile sample was mixed with β-glucuronidase solution (24 units/ml for plasma and 120 units/ml for bile) and then incubated at 47°C for 5 h. After treatment, the free MPA formed was determined as mentioned above. Bile samples were diluted with phosphate-buffered saline before treatment, if necessary. Chromatographic conditions for MPA were as follows: apparatus, Shimadzu LC-6A (Kyoto, Japan) equipped with a Shimadzu SPD-6A UV spectrophotometric detector; column, Cosmosil 5C18-ARII (4.6 × 150 mm; Nakarai Tesque, Kyoto, Japan); column temperature, 30°C; mobile phase, 40% acetonitrile in 40 mM H3PO4 (pH 2.1 adjusted with KOH); wavelength, 250 nm for bile samples and 304 nm for plasma samples; and flow rate, 0.8 ml/min. Under these high-performance liquid chromatography conditions, MPA and naproxen eluted at 9 and 12 min, respectively, and MPA was reproducibly determined with a coefficient of variance less than 5%. The limit of determination of MPA assay was 0.2 μg/ml in bile samples and 1 μg/ml in plasma samples, respectively. The concentration of MPA derived from MPAG was calculated by subtracting the concentration of free MPA from the total concentration of MPA obtained after β-glucuronidase treatment. Excretion ratio of free and MPAG-derived MPA in bile during each 15-min interval was calculated as excretion ratio = Cb · Vb/X0, where Cb is concentration in bile, Vb is volume of bile, and X0 is dose. Cumulative excretion ratio (CER) was obtained by summing up excretion ratio of each 15-min interval.
Pharmacokinetic and Statistical Analysis. The concentration of MPA in plasma at time 0 (C0) and elimination rate constant (Ke) were calculated using MULTI (Yamaoka et al., 1981), assuming a compartment open model. Area under the plasma concentration/time curve from 0 to 1 h (AUCi.v., 0–1) was determined using the linear trapezoidal rule. Volume of distribution (VD) and total clearance (CLtot) were calculated using the following equations: VD = X0/C0, CLtot = Ke · VD. Each data represents the mean ± S.E. of three to four experiments. Statistical analysis was done using unpaired Student's t test and p < 0.05 was considered to be significant.
Results
Effect of CsA and Tacrolimus on the Pharmacokinetics of Free and MPAG-Derived MPA.Fig. 1 shows the plasma concentration/time curves of free MPA and MPAG-derived MPA after the intravenous administration of MPA to Wistar rats with CsA or tacrolimus. When MPA was administered alone, free MPA disappeared rapidly from plasma over 1 h (Fig. 1A). CsA and tacrolimus did not affect the plasma concentrations of free MPA at any sampling time point over 1 h. Pharmacokinetic data for free MPA are presented in Table 1. The CLtot of free MPA after coadministration of MPA and tacrolimus was slightly but significantly greater than those of the others. The plasma concentrations of MPAG-derived MPA gradually increased over 1 h after the intravenous administration of MPA (Fig. 1B). The plasma concentrations of MPAG-derived MPA after the coadministration of MPA and tacrolimus were comparable with those of MPA alone at all sampling time points. When MPA was administered with CsA, the plasma concentration of MPAG-derived MPA became significantly greater than those of the others at 30 min.
Effect of CsA and Tacrolimus on the Biliary Excretion of Free and MPAG-derived MPA in Wistar Rats.Fig. 2 shows the CERs of free and MPAG-derived MPA in bile after the intravenous administration of MPA to Wistar rats. In the case of MPA alone, the CER of free MPA was very small, remaining at most only 0.15% of the dose (Fig. 2A). Tacrolimus did not alter the biliary excretion of free MPA, whereas CsA slightly but not significantly decreased the CER of free MPA. The biliary excretion of MPAG-derived MPA was extensive in the case of MPA alone, and the CER reached approximately 26% of the dose (Fig. 2B). Tacrolimus failed to modulate the biliary excretion of MPAG-derived MPA. In contrast, the CER of MPAG-derived MPA significantly decreased by about 20% when MPA was coadministered with CsA. It was confirmed that there were no differences in the volume of bile excreted during each 15-min interval under these experimental conditions.
Plasma Concentrations of Free and MPAG-derived MPA after Intravenous Administration of MPA to SD Rats and EHBRs.Fig. 3 shows the plasma concentration/time curves of free and MPAG-derived MPA after the intravenous administration of MPA to SD rats and EHBR, together with the results obtained in Wistar rats. The profile for the EHBRs was almost the same as that for the Wistar rats, and there were no significant differences in the plasma levels of free MPA between them at any sampling time point during the hour (Fig. 3A). On the other hand, the plasma concentrations of free MPA in SD rats were significantly greater than those from EHBR and Wistar rats at all sampling time points. Pharmacokinetic data of free MPA obtained from EHBR and SD rats are summarized in Table 1. Although the C0 and AUCi.v., 0–1 values were significantly smaller in the EHBRs than in SD rats, the Ke and CLtot values were significantly greater in the EHBRs. The plasma concentrations of MPAG-derived MPA gradually increased over 1 h in the EHBRs, and they were significantly greater than those in the Wistar rats at all sampling time points (Fig. 3B). In contrast, MPAG-derived MPA was not detected in the plasma of SD rats at any sampling time point.
Biliary Excretion of Free and MPAG-derived MPA after Intravenous Administration of MPA to SD Rats and EHBRs.Fig. 4 shows the CERs of free and MPAG-derived MPA after the intravenous administration of MPA to SD rats and EHBR. The CER of free MPA in SD rats increased with time over 1 h, whereas the biliary excretion of free MPA was significantly retarded in the EHBRs compared with SD rats, and it reached a plateau after 0.5 h (Fig. 4A). Although the CER of free MPA in SD rats was rather smaller than that in Wistar rats, this difference was not statistically significant. There were remarkable differences in the CERs of MPAG-derived MPA between the EHBR and SD rats; that is, MPAG-derived MPA was efficiently excreted into the bile over 1 h in SD rats, giving a CER of ca. 21% of the dose. In contrast, the CER of MPAG-derived MPA in the EHBRs was very low and remained approximately 0.5% of the MPA dose (Fig. 4B). It was found that the CERs of MPAG-derived MPA in SD rats were significantly smaller than those in Wistar rats.
Discussion
Biliary excretion is an important process for the elimination of many drugs and their metabolites (Roberts et al., 2002), whereas the extensive biliary excretion of several compounds is closely associated with their toxic gastrointestinal adverse effects, such as diarrhea and mucosal lesions. Methotrexate (Masuda et al., 1997), irinotecan (Lokiec et al., 1995) and acetaminophen (Madhu et al., 1989) are known to be models of this mechanism. Moreover, there have been many reports describing the efficient biliary excretion of various glucuronide conjugates. It has been reported that approximately 30% of intravenously administered indomethacin was excreted in bile of SD rats, mostly as its glucuronide (Kouzuki et al., 2000) and that about 10% of telmisaltan glucuronide was excreted in bile of SD rats over 1 h (Nishino et al., 2000). In the present study, when MPA was intravenously administered to Wistar or SD rats, the CER of MPAG-derived MPA reached approximately 26 and 21% of the dose over 1 h, respectively (Fig. 4B), suggesting that the biliary excretion of MPAG is extensive in these rat strains.
Recently, many studies have suggested that MRP2/Mrp2 plays a primary role in the biliary excretion of glucuronide conjugates of various chemical compounds such as indomethacin (Kouzuki et al., 2000), SN-38 (Chu et al., 1997), and 17β-estradiol (Morikawa et al., 2000). In such studies, EH-BRs (a mutant SD strain) or a transport-deficient Wistar strain of rats, both of which genetically lack Mrp2 on the canalicular membrane (Ito et al., 1997; Madon et al., 1997), have been used favorably. In this study, the CER of MPAG-derived MPA in the EHBRs and normal SD rats was 0.5 and 21% of dose, respectively (Fig, 4B). This suggests that Mrp2 was essential for the biliary excretion of MPAG; that is, the glucuronide is a likely substrate of Mrp2. Because the CER of free MPA in the EHBRs was also significantly smaller than that in SD rats (0.02 versus 0.08% of the dose after 1 h; p < 0.01), free MPA is also considered to be a substrate of Mrp2. However, because the CER ratio of SD/EHBR was approximately 4 for free MPA and 40 for MPAG-derived MPA (Fig. 4, A and B), it seems that Mrp2 much prefers MPAG to MPA. Interestingly, although MPAG-derived MPA was not detected in the plasma of SD rats and at most at 3.5 μg/ml in Wistar rats after 1 h, in the EHBRs it greatly increased over 1 h and reached about 14 μg/ml after 1 h (Fig. 3B). It was recently reported that Mrp3 was compensatively up-regulated on the sinusoidal membrane of EHBRs and exclusively expelled taurocholic acid, a substrate of Mrp2, into sinusoidal blood (Akita et al., 2001). Mrp3 is capable of transporting several glucuronide conjugates (Hirohashi et al., 1999). Thus, it is likely that Mrp3 greatly contributed to the predominant appearance of MPAG in the plasma of the EHBRs through the efficient transport of MPAG across the sinusoidal membrane. The expression of Mrp3 is negligible on the sinusoidal membrane of SD rats (Hirohashi et al., 1998; Akita et al., 2001). Therefore, no detection of MPAG in the plasma of SD rats is in part attributed to the lack of Mrp3-mediated transport in SD rats. In contrast, a finding that MPAG was detected in the plasma of Wistar rats (Fig. 3B) may imply that Mrp3 activity is a little greater in Wistar rats than in SD rats. After the intravenous administration of MPA, free MPA in the plasma of the EHBRs sharply disappeared and was about 6 μg/ml after 1 h (Fig. 3A). On the other hand, the levels of MPAG continued increasing in the same plasma of EHBR even after 1 h (Fig. 3B). These results could indicate that MPA was taken up rapidly and efficiently from sinusoidal blood into the hepatocytes. The disappearance of MPA from the plasma was much faster in EHBR than in SD rats and all pharmacokinetic parameters obtained were significantly different between them (Table 1). Recently, many kinds of transporters have been identified for the hepatobiliary excretion of various compounds. Whereas P-glycoprotein, organic anion-transporting polypeptide, Na+-taurocholate cotransporting polypeptide, and MRP-like protein 1 are comparably expressed between SD rats and EHBR (Hirohashi et al., 1998; Ogawa et al., 2000; Akita et al., 2001), the expression of MRP-like protein 2 in the hepatocytes is limited to EHBR (Hirohashi et al., 1998). Thus, it is likely that an EHBR-specific transport system is involved in MPA uptake in the hepatocytes. Further study is now underway to clarify whether a specialized transporter is involved in MPA uptake across the sinusoidal membranes.
CsA, which was intravenously administered to Wistar rats, significantly inhibited the biliary excretion of MPAG (Fig. 2B). This result was as expected considering that MPAG is a substrate of Mrp2 as described above and that CsA is a potential inhibitor of Mrp2 (Kamisako et al., 1999). Significant increase in the plasma concentration of MPAG-derived MPA after coadministration of MPA and CsA (Fig. 1B) might be a compensative phenomenon for the disrupted Mrp2-mediated MPAG transport in bile. CsA lowered the biliary excretion of free MPA, although the effect was not significant (Fig. 2A). Considering that MPA is a weak substrate of Mrp2 (Fig. 4A), the result seems reasonable. However, because the CER of free MPA was very small compared with that of MPAG, the significance of the interaction is regarded as marginal. Tacrolimus is reported to possess the capacity to interfere with transporters such as P-gp (Jachez et al., 1993; Kochi et al., 1999). However, the present results indicated that tacrolimus failed to interfere with the Mrp2-mediated MPAG transport across the canalicular membrane of Wistar rats (Fig. 1B). In this study, the doses of CsA and tacrolimus were based on their clinical regimens and tacrolimus was administered at a 50-fold lower dose than CsA. Thus, it is possible that tacrolimus could not exhibit enough ability to interfere with Mrp2 due to the low dosage and that the same dosage as that of CsA is capable of modulating Mrp2-mediated MPAG transport. However, it would be beyond clinical relevance even if the second possibility were true.
Because the difference in biliary excretion of glucuronide seems less marked among species (Niinuma et al., 1999), the present results imply that MPAG also efficiently undergoes biliary excretion via MRP2 in the human liver. Actually, a second peak, which is due to the enterohepatic circulation of MPAG, is known to emerge in MPA plasma concentration/time profiles after the oral administration of MMF to humans (Bullingham et al., 1996a). On the other hand, it has been reported that the majority of MMF orally administered to healthy subjects was finally excreted in the urine mostly as MPAG and that total recovery of the dose in the feces averaged only 5.5% (Bullingham et al., 1996b). Accordingly, MPA repeatedly undergoes glucuronidation in the liver, MRP2-mediated biliary excretion, deglucuronidation by the intestinal flora and subsequent intestinal reabsorption until MPA molecules are finally excreted to the urine as MPAG. Thus, the possibility emerges that the enterohepatic circulation of MPAG is more efficient in patients with renal dysfunction and that the occurrence of diarrhea would become more frequent in such patients if the biliary excretion of MPAG is closely involved in the occurrence of diarrhea. Because it has been reported recently that Mrp2 is present on the brush-border membrane of the intestine (Mottino et al., 2000), it is likely that the intestinal secretion of MPAG is greatly stimulated under renal dysfunction as well. In our separate study, it was found that the diarrhea that occurred in renal transplant recipients receiving MMF was most frequent during first 4 to 5 days after transplant operation and then became markedly less frequent as the renal function recovered to an almost normal range (unpublished data). Together with the present results, this suggests that the frequency and severity of diarrhea after MMF administration is closely related to the extent of renal and biliary excretion of MPAG in the body.
In this study, several differences in the disposition of intravenously administered MPA were observed between Wistar and SD rats. Although it was previously reported that P-glycoprotein expression in cultured hepatocytes was different among rat strains (Chieli et al., 1994), little information is currently available on the interstrain differences. Further investigation is required to clarify it.
In conclusion, this study demonstrates for the first time that the phenolic glucuronide of MPA is a substrate of Mrp2 present on the canalicular membrane, although in vitro study is necessary to further clarify the role of Mrp2 in MPAG transport in more detail and that the glucuronide undergoes efficient biliary excretion mediated by the transporter. CsA, but not tacrolimus, potently inhibits Mrp2-mediated MPAG transport. Based on the present results, we assume that when MMF was coadministered with CsA, CsA potently inhibited the biliary excretion of MPAG in the transplant recipients, whereas there was no significant inhibition of biliary excretion of MPAG when MMF was coadministered with tacrolimus. Thus, the diarrhea associated with the biliary excretion of MPAG was exacerbated in patients receiving MMF with tacrolimus.
Footnotes
-
DOI: 10.1124/jpet.103.063073.
-
ABBREVIATIONS: MMF, mycophenolate mofetil; MPA, mycophenolic acid; CsA, cyclosporin A; MPAG, mycophenolic acid glucuronide; SD, Sprague-Dawley; EHBR, Eisai hyperbilirubinemic rat; Mrp2, multidrug resistance-associated protein 2; CER, cumulative excretion ratio; SN-38, 7-ethyl-10-hydroxycamptothecin.
- Received November 30, 2003.
- Accepted February 19, 2004.
- The American Society for Pharmacology and Experimental Therapeutics