METABOLISM OF OPIOIDS IS ALTERED IN LIVER MICROSOMES OF SICKLE CELL TRANSGENIC MICE

Transcription

1 /04/ $20.00 DRUG METABOLISM AND DISPOSITION Vol. 32, No. 1 Copyright 2004 by The American Society for Pharmacology and Experimental Therapeutics 1103/ DMD 32:98 104, 2004 Printed in U.S.A. METABOLISM OF OPIOIDS IS ALTERED IN LIVER MICROSOMES OF SICKLE CELL TRANSGENIC MICE Swati Nagar, Rory P. Remmel, Robert P. Hebbel, and Cheryl L. Zimmerman Departments of Pharmaceutics (S.N., C.L.Z.), Medicinal Chemistry (R.P.R.), and Medicine (R.P.H.), University of Minnesota, Minneapolis, Minnesota (Received March 3, 2003; accepted September 15, 2003) This article is available online at ABSTRACT: The genetic basis of sickle cell anemia (SCA 1 ), a hemoglobin disorder, has been understood since 1949 (Pauling et al., 1949); red blood cells (RBCs) assume a sickle shape when deoxygenated due to a point mutation in the beta-globin gene of the hemoglobin molecule. In the United States alone, around 1000 babies are born every year with SCA, and it is estimated that 1 in every 600 African Americans is homozygous for the mutation (Steinberg, 1999). Sickling of RBCs affects almost all organs of the body, including the heart, lung, kidneys, and liver. Among the various complications of SCA, acute and chronic pain syndromes cause the highest number of hospital visits. The intense pain is commonly managed with opioid analgesics This project was funded partially by National Institutes of Health Grant HL55552 to R.P.H., and the William and Mildred Peters Endowment, Department of Pharmaceutics, University of Minnesota. This work was submitted by S.N. in partial completion of the requirements for the Ph.D. in Pharmaceutics at the University of Minnesota. 1 Abbreviations used are: SCA, sickle cell anemia; RBC, red blood cell; UGT, uridine glucuronosyltransferase, human; ugt, uridine glucuronosyltransferase, rodent; P450, cytochrome P450; SC, sickle cell transgenic mice; SCKO, sickle cell-knockout mice; NB, norbuprenorphine; M3G, morphine-3-glucuronide; UD- PGA, uridine diphosphoglucuronic acid; G6P, glucose-6-phosphate; HPLC, highpressure liquid chromatography; ACN, acetonitrile; BG, buprenorphine glucuronide; E3G, estradiol-3-glucuronide; E17G, estradiol-17-glucuronide; EE2, ethynylestradiol; Cl int, intrinsic clearance of unbound drug. Address correspondence to: Dr. Cheryl L. Zimmerman, College of Pharmacy, University of Minnesota, 308 Harvard St. S.E, Minneapolis, MN zimme005@umn.edu 98 Pain in sickle cell anemia (SCA) is clinically managed with opioid analgesics. There are reports that SCA patients tolerate high doses of these drugs without adequate pain relief. The current study investigated the in vitro hepatic metabolism of opioids in mouse models of sickle cell anemia, with the hypothesis that higher dose requirements in SCA could be explained by an increased metabolism rate of opioids. Various rodent cytochrome P450 substrates, i.e., buprenorphine and codeine, and rodent uridine glucuronosyltransferase substrates, i.e., morphine, buprenorphine, and estradiol, were studied. The three groups used were: 1) control C57BL mice, 2) mice with the human -globin and sickle -globin transgenes (SC), and 3) mice with the human -globin and sickle -globin transgenes, and homozygous for the murine -globin and heterozygous for the major -gene knockout (SCKO). In vitro hepatic microsomal incubations were carried out for each substrate, and data were fit to the Michaelis-Menten equation. Morphine formation had a higher V max in SCKO microsomes ( nmol/min mg; estimate S.E.) than controls ( ). Morphine-3- glucuronide formation had V max estimates of , , and nmol/min mg in control, SC, and SCKO microsomes, respectively. The control V max for estradiol-3-glucuronide formation was 2-fold greater than in SCKO microsomes. The control V max for estradiol 17-glucuronide formation was 3.4- and 2.2- fold greater than in SC and SCKO microsomes. Thus, in vitro metabolism of opioids is altered in SCA mouse models, which may lead to altered clearances of these drugs. including morphine, buprenorphine, and nalbuphine (Embury et al., 1994). There have been various clinical reports that sickle pain is not adequately managed by typical doses of opioids, and SCA patients can tolerate very high doses of analgesics without alleviation of their symptoms (Dampier et al., 1995; Beyer, 2000; Yaster et al., 2000). Studies have been limited due to the complexity of study designs involving persons with SCA, many of whom are pediatric patients. Studies in SCA patients have indicated that one possible reason for higher dose requirement is an alteration in drug pharmacokinetics. The elimination half-life of codeine in sickle cell patients was found to be significantly shorter than in control subjects (Mohammed et al., 1993). Dampier et al. (1995) reported that the clearance of morphine in SCA patients was higher than typical values (Glare and Walsh, 1991). SCA patients have higher than normal levels of both unconjugated and conjugated bilirubin as a result of hemolysis as well as cholestasis (Embury et al., 1994). High circulating bilirubin levels may play a role in the modulation of enzyme activity. Maddrey et al. (1978) have reported an increase in the bilirubin-conjugating activity in SCA liver biopsy samples; bilirubin is known to be glucuronidated by human uridine glucuronosyltransferase (UGT) 1A1, as well as the corresponding rodent ugt1a1 (Iyanagi et al., 1998). Interestingly, a study by Sanchez and Tephly (1973) showed that exogenous bilirubin increased rat ugt activity in vivo as well as in microsomes from treated animals. Transgenic mouse models carrying the human sickle cell anemia gene have been developed in recent years. Sickle cell transgenic mice (SC) have a C57BL thalassemia background, and the human -globin gene and the human sickle -globin genes have been introduced into

2 DRUG METABOLISM IN SICKLE CELL TRANSGENIC MICE 99 TABLE 1 Human and rodent isozymes responsible for metabolism of various substrates Substrate Human Isozyme (Reference) Rodent Isozyme (Reference) Morphine UGT2B7(Coffman et al., 1997) ugt2b1*(coffman et al., 1996; King et al., 1997) Buprenorphine UGT1A1, 1A3, 2B7(Cheng et al., 1998) ugt1.1*, 2b1*(Coffman et al., 1996; King et al., 1997) Estradiol (E3G) UGT1A1(Senafi et al., 1994) ugt1a1 (?) Estradiol (E17G) UGT1A, 2B7(Senafi et al., 1994; Gall et al., 1999) ugt1a, 2b (?) Buprenorphine CYP3A4(Kobayashi et al., 1998) cyp3a (?) Codeine CYP2D6(Tyndale et al., 1997) cyp2d22**(blume et al., 2000) * Rat homolog. ** Mouse homolog. the mouse genome (Fabry et al., 1992). A more severe model of the disease is in the form of the sickle cell-knockout mice (SCKO); in addition to the introduction of the human genes, these mice are homozygous for the knockout of the murine -globin gene and heterozygous for the knockout of the murine -gene (Paszty et al., 1997). SCKO have a mixed genetic background. The transgenic mice have various pathological characteristics mirroring the human disease, such as extensive sickling of red blood cells on deoxygenation, retinopathy, and renal and hepatic pathology (Embury et al., 1994). The percentage of sickled erythrocytes that develop after complete deoxygenation in control mice (C57BL), SC, and SCKO is 0, 40, and 80%, respectively, and the percentage of nonalpha-globins that are comprised of the human S -globin is 0, 72, and 100%, respectively (R. P. Hebbel, unpublished data). These animal models provide an excellent opportunity to investigate alterations in drug disposition due to SCA. We hypothesize that enzyme induction could be one explanation for the higher observed clearance of opioid analgesics in SCA. To this end, in vitro hepatic metabolism by various rodent P450 and ugt isozymes was investigated in two different SCA mouse models. The substrates chosen were opioid analgesics commonly used in the management of SCA pain. Since buprenorphine is glucuronidated by various UGT1A and UGT2B isozymes, the glucuronidation of estradiol was also studied because estradiol is reported to be specifically glucuronidated at its 3-position by human UGT1A1 (Senafi et al., 1994). Table 1 shows the human and rodent isozymes responsible for the metabolism of the substrates used. For most of the substrates, mouse isozymes responsible for the reactions are unknown, and so the probable homologs based on human data are listed. For morphine and buprenorphine glucuronidation, the known rat isozymes are listed. Materials and Methods Morphine sulfate, morphine 3-glucuronide (M3G), buprenorphine hydrochloride, estradiol, estradiol 3-glucuronide (E3G), estradiol 17-glucuronide (E17G), ethynylestradiol (EE2), codeine sulfate, alamethicin, uridine diphosphoglucuronic acid (UDPGA), glucose 6-phosphate (G6P), G6P-dehydrogenase, -NADP, -glucuronidase, and phenytoin were obtained from Sigma- Aldrich (St. Louis, MO). Norbuprenorphine (NB) was obtained through the National Institute on Drug Abuse, from the Research Triangle Institute (Research Triangle Park, NC). All other chemicals used were reagent or HPLC grade. All transgenic mice were bred in the laboratory of Dr. R. P. Hebbel, University of Minnesota, and wild-type C57BL/6J mice were obtained from Harlan (Indianapolis, IN). All mice were maintained under specific pathogenfree conditions. Preparation of Mouse Liver Microsomes. Three groups of mice were used for the studies: control mice (wild-type C57BL/6J background), SC, and SCKO. For the preparation of microsomes, animals were sacrificed in a CO 2 chamber, and the livers were immediately harvested. Up to five livers were pooled and homogenized for each batch of microsomes. Microsomes were prepared by differential centrifugation, by a previously published method (Nelson et al., 2001). Briefly, liver homogenates were centrifuged at 9000g for 20 min at 4 C, and the resulting supernatant was centrifuged at 100,000g for 60 min at 4 C. The microsomal pellet was suspended in 0.1 M phosphate buffer containing 2 mm EDTA and 20% glycerol and stored at 80 C until further use. The protein content was determined using the BCA protein assay kit (Pierce Chemical, Rockford, IL). Microsomal Incubations for Rodent P450-Catalyzed Reactions. Incubation conditions were optimized for linearity with respect to protein and incubation period for each type of microsome and for every reaction. For the conversion of buprenorphine to NB as well as codeine to morphine, preliminary experiments showed that the reactions were linear up to 60 min and 1 mg/ml of protein. The necessary cofactors were added to the incubation mixture as two solutions: solution A contained -NADP, MgCl 2, and G6P in 50 mm Tris buffer (ph 7.4 at 37 C) and yielded final concentrations of 1 mm -NADP, 5 mm MgCl 2, and 5 mm G6P. Solution B was a G6P-dehydrogenase solution in 50 mm Tris buffer and gave a final concentration of 1.0 IU/ml of G6P-dehydrogenase in the incubation. The final volume of each incubation solution was made up to 250 l with 50 mm Tris buffer (ph 7.4 at 37 C). Mouse liver microsomes were at a final concentration of 1 mg/ml, and solutions A and B were preincubated at 37 C for 3 min. The reaction was started by the addition of varying concentrations of substrate and was terminated after 1 h by the addition of ice-cold acetonitrile (ACN) containing phenytoin as the internal standard. The solution was centrifuged to pellet the precipitated protein, and the supernatant was injected onto the HPLC column. Six replicates of each experiment were carried out with pooled mouse liver microsomes for each group. Microsomal Incubations for ugt-catalyzed Reactions. Incubation conditions were optimized for linearity with respect to protein and incubation period for each type of microsome and for every reaction. The conversion of morphine to M3G was linear up to 15 min, whereas the formations of buprenorphine glucuronide (BG), E3G, and E17G were linear up to 10 min. Alamethicin, a pore-forming peptide, was used to activate the microsomal preparations before incubation. Alamethicin concentration was optimized to achieve maximal activation of the microsomes. Thus, 75 M alamethicin was used for control and SCKO microsomes and 50 M alamethicin was used for SC microsomes. The final incubation volume in all cases was made up to 200 l using 0.1 M Tris buffer (ph 7.4 at 37 C). For the incubations, a mixture of 1 mg/ml of microsomal protein, alamethicin, 5 mm MgCl 2, and substrate was preincubated for 3 min at 37 C. The reaction was started by adding UDPGA to give a final concentration of 3 mm. Control incubations were performed without the addition of UDPGA. For morphine and buprenorphine, the reaction was terminated by the addition of ice-cold ACN containing phenytoin as the internal standard. For estradiol, the reaction was terminated with ice-cold ACN and subsequent addition of the internal standard, EE2. In all cases, the incubation solution was centrifuged at 13,000g for 2 min and the supernatant was used for HPLC analysis. Six replicates of each experiment were performed with pooled mouse liver microsomes for each group. For incubations for the conversion of buprenorphine to BG, preliminary incubations were carried out to verify the glucuronidation reaction; -glucuronidase was added to the reaction mixture, and the resulting disappearance of the BG peak with a corresponding increase of the buprenorphine peak confirmed that the metabolite peak was indeed BG. HPLC Analysis. The HPLC system used was a Beckman Gold 126 system (Beckman Coulter Inc., Fullerton, CA) with a 507e autosampler, a 166 UV detector, and a Gold Nouveau integration software (version 1.6, Beckman Coulter). Except for the formation of BG, metabolite formation was quantitated by comparing metabolite/internal standard peak area ratios in incubations

3 100 NAGAR ET AL. TABLE 2 Reversed-phase HPLC conditions for the analysis of the various CYP and UGT substrates Compound Mobile Phase Column Flow Rate (ml/min) Detection Retention Time (min) 1.0 UV 210 nm 26 Buprenorphine Phosphate buffer b and ACN (64:36 v/v) Reversed-phase Luna 5- m C18(2) 150 NB 4.6 mm; Phenomenex c 11 Phenytoin (I.S.) a 7 Codeine Phosphate buffer and ACN (70:30 v/v) Reversed-phase Luna 5- m C18(2) UV 210 nm 7 Morphine 4.6 mm; Phenomenex 4.6 Phenytoin (I.S.) 12.4 Morphine d Phosphate buffer and ACN (79:21 v/v) Reversed-phase Luna 5- m C18(2) UV 210 nm 16 M3G 4.6 mm; Phenomenex 6 Phenytoin (I.S.) 22 Buprenorphine Phosphate buffer and ACN (63:37 v/v) Reversed-phase Luna 5- m C18(2) UV 212 nm 9.4 BG 4.6 mm; Phenomenex 2.5 Phenytoin (I.S.) 3.5 Estradiol Acetate buffer e and ACN, gradient elution f Reversed-phase Supelco 5- m C UV 280 nm 24 E3G 4.6 mm; Supelco g 7 E17G 8.4 EE2 (I.S.) 25 a I.S., internal standard. b 10 mm phosphate buffer with 0.8 mm SDS, ph 2.1. c Phenomenex is based in Torrance, CA. d Lear et al., e 0.05 M ammonium acetate buffer, ph 4.5 (Ebner et al., 1993). f 26% ACN for 12 min, increase to 50% ACN over 1 min; 50% ACN for 7 min, increase to 85% ACN over 0.5 min; 85% ACN for 6 min, re-equilibration at 26% ACN. g Supelco is based in Bellefonte, PA. TABLE 3 Rodent P450-catalyzed reactions: Michaelis-Menten parameters for the formation of NB and morphine in control mouse, SC, and SCKO liver microsomes Metabolite (Isozyme) V max a (nmol/min mg) K m a ( M) V max /K m ( l/min mg) to a standard curve with known amounts of metabolite. Since a BG standard was unavailable, BG areas were quantitated against a buprenorphine standard curve and expressed as equivalent buprenorphine concentrations. UV spectra of both buprenorphine and BG by diode array detection in the 190- to 250-nm wavelength range showed an identical spectrum with a max at 212 nm. All assays were validated for inter- and intraday precision and accuracy, and the percentage of CV and bias were less than 10% throughout the concentration range studied (data not shown). The HPLC conditions for each substrate and its metabolites are shown in Table 2. Data Analysis. Metabolite formation data were fit to the simple form of the Michaelis-Menten equation (Gibaldi and Perrier, 1982): V V max S/ K m S where V is the initial rate of metabolite formation, V max is the maximum rate of metabolite formation, S is the substrate concentration, and K m is the Michaelis-Menten constant. Eadie-Hofstee plots were evaluated for linearity for each reaction. The nonlinear fit was carried out with KaleidaGraph 3.5 (Synergy Software, Reading, PA). Statistical comparisons for the Michaelis-Menten parameter estimates were carried out using a two-sided t test, assuming normal distribution of parameter estimates. A p value less than 0.01 was considered significant. Liver weights were compared for statistical significance with StatView 5.0 (SAS Institute Inc., Cary, NC). The ratio of V max to K m is the in vitro intrinsic clearance of the unbound drug. A theoretical in vivo intrinsic clearance of the unbound drug was calculated by multiplying the V max /K m value by the microsomal protein yield In Vivo Cl int ( l/min/g of body weight) NB (cyp3a) Control SC SCKO Morphine (cyp2d22) Control SC SCKO * a Data expressed as parameter estimate S.E. (n 6). * Value was significantly different than corresponding control values, as determined by a two-sided t test (p 0.01). for each type of liver. The microsomal protein yield was calculated from the liver weights, with a scaling factor of 45 mg/g of liver (Houston, 1994). Results The mean body weights of age-matched male control mice, SC, and SCKO were (mean S.D.), , and g, respectively, and the mean liver weights were , , and g, respectively. The mean body weight normalized liver weights (gram per gram of body weight) in control mice, SC, and SCKO were 0.074, 0.062, and 0.048, respectively. The microsomal protein yield was calculated using a scaling factor of 45 mg/g of liver (Houston, 1994). These values were 70.2, 92.7, and 72.5 mg for control mice, SC, and SCKO microsomes, respectively. The microsomal protein yield values were used to calculate theoretical in vivo Cl int values in Tables 3 and 4. Figure 1, a and b, shows the formation of NB from buprenorphine and morphine formation from codeine, with the corresponding Michaelis-Menten parameters of these rodent P450-catalyzed reactions reported in Table 3. All three groups showed similar NB formation kinetics. The formation of morphine from codeine showed a significantly higher V max in SCKO microsomes than in controls. The Eadie-Hofstee plots were linear for both NB and morphine formation (data not shown). Figure 2 shows the ugt-catalyzed formation curves for M3G (Fig.

4 DRUG METABOLISM IN SICKLE CELL TRANSGENIC MICE 101 TABLE 4 ugt-catalyzed reactions: Michaelis-Menten parameters for the formation of M3G, E3G, and E17G in control mouse, SC, and SCKO liver microsomes Metabolite (Isozyme) V max a (nmol/min mg) K m a ( M) V max /K m ( l/min mg) In Vivo Cl int (ml/min/g of body weight) M3G (ugt2b) Control SC * SCKO * E3G (ugt1a1) Control SC SCKO * E17G (ugt1a, 2b) Control SC * SCKO * a Data expressed as parameter estimate S.E. (n 6). * Values were significantly different than corresponding control values, as determined by a two-sided t test (p 0.01). FIG. 1.Rodent P450-catalyzed reactions: formation of NB (a) and morphine (b) in control mouse ( and irregularly dashed lines), SC (F and solid lines), and SCKO (Πand evenly dashed lines) liver microsomes. Data are expressed as mean S.D. (n 6). Michaelis-Menten curves are representative curves drawn through mean data; parameter estimates were obtained by fitting the actual data. 2a), BG (Fig. 2b), E3G (Fig. 2c), and E17G (Fig. 2d), and the corresponding Michaelis-Menten parameter estimates are presented in Table 4. The V max of M3G formation was found to be significantly higher in SC and SCKO microsomes as compared with controls. The V max value in controls ( nmol/min mg) was close to other published values in the C57BL mouse liver microsomes (Bock et al., 1982; Hanioka et al., 1990). There was no significant difference in BG formation among the three groups, as seen in Fig. 2b. A simple Michaelis-Menten curve could not be fitted to the BG formation data, hence no parameter estimates are reported. The Eadie-Hofstee plot for BG formation was nonlinear and atypical in all groups of mice (Fig. 3), indicating the involvement of more than one ugt isozyme. For E3G formation, the V max in SCKO microsomes was significantly lower than in controls. The V max values for E17G formation were also significantly lower in SC and SCKO microsomes as compared with controls (Table 4). Discussion Pain in sickle cell anemia is commonly managed with opioid analgesics, but there are various clinical reports that pain is not alleviated in these patients even after intensive analgesic therapy (Beyer, 2000; Yaster et al., 2000). Our study explored the possibility of altered pharmacokinetics as a cause of higher dose requirements in SCA. Increased metabolism of drugs can lead to an increase in the overall clearance of drugs. Drugs such as morphine and codeine are classically thought of as high-clearance drugs, i.e., their overall clearance is affected only by hepatic blood flow and not by a change in the intrinsic clearance (Wilkinson and Shand, 1975). However, liver cirrhosis is common in SCA (Embury et al., 1994). It has been shown that in liver disease and cirrhosis, the liver can no longer be thought of as well stirred, and so the equations for drug clearance based on the well stirred model of the liver no longer hold true. Thus, hepatic enzyme activity becomes an important contributor to the overall clearance of even a high-clearance drug (Huet and Villeneuve, 1983). Interestingly, in a liver with cirrhosis, UGT function seems to remain unaltered, whereas human P450 function may be adversely affected (Patwardhan et al., 1981). Thus, it is important to consider how drug metabolism would change due to the pathophysiology of SCA compounded by liver cirrhosis. In a diseased liver in SCA, a normally high-extraction drug will probably be affected both by changes in hepatic blood flow and intrinsic clearance. Additionally, pathophysiological changes due to SCA such as increased circulating bilirubin levels may in fact increase human P450 and UGT activity, increasing the intrinsic clearance of various drugs. Higher free bilirubin concentrations in the liver could also competitively inhibit other UGT1A1 substrates. How the presence of sickle RBCs perturbs liver function is also an important consideration. The transgenic mouse models closely approximate the human disease, especially with phenotypic similarities such as enlarged liver, retinopathy, and renal pathology. A significant alteration in the metabolism of drugs by both rodent P450 and ugt isozymes in the transgenic mice was observed. Also, each isozyme shows a unique change in its drug-metabolizing capacity in these mouse models.

5 102 NAGAR ET AL. FIG. 2.ugt-Catalyzed reactions: formation of M3G (a), BG (b), E3G (c), and E17G (d) in control mouse ( and irregularly dashed lines), SC (F and solid lines), and SCKO (Πand evenly dashed lines) liver microsomes. Data are expressed as mean S.D. (n 6). Michaelis-Menten curves are representative curves drawn through mean data; parameter estimates were obtained by fitting the actual data. Specifically, our results indicate the following: 1) the mouse cyp3a isozyme responsible for the formation of NB from buprenorphine shows no change in activity among the three groups; 2) the mouse cyp2d22 isozyme responsible for the conversion of codeine to morphine is induced in SCKO microsomes, showing a significantly higher V max than in control microsomes; 3) the mouse ugt2b isozyme responsible for morphine glucuronidation is induced in both SC and SCKO microsomes and shows a significantly higher V max than controls; and 4) the mouse ugt1a isozyme responsible for estradiol glucuronidation shows decreased V max values for E3G formation in SC microsomes and decreased V max values for E3G and E17G formation in SCKO microsomes. The V max /K m ratios in either case, however, are not very different from control values. In all cases, the observed significant differences were more pronounced in the SCKO, a more severe model of SCA. The increased conversion of codeine to morphine does not support the hypothesis that the decreased analgesic effect in SCA is due to enhanced metabolism, since morphine is a potent analgesic. However, the sequential conversion of morphine to the inactive M3G is also increased, and the relative rates of the two reactions in vivo would determine the overall analgesic effect. It is of interest to note that in preliminary studies, there were significant differences in baseline response to tail-flick tests among the control mice, SC, and SCKO, indicating differences in nociceptive response (Nagar, 2003). All the rodent P450- and ugt-catalyzed reactions studied except BG formation gave linear Eadie-Hofstee plots (data not shown), indicative of single-enzyme kinetics in the substrate concentration range studied. BG formation showed atypical kinetics (Figs. 1b and 3) with nonlinear Eadie-Hofstee plots. This may be due to the involvement of more than one ugt isozyme. Atypical Eadie-Hofstee curves such as hook-shaped plots are usually obtained due to allosteric sites, multiple isozymes, or activation kinetics (Fisher et al., 2000; Kenworthy et al., 2001; Oda and Kharasch, 2001). The V max /K m ratio is the maximum in vitro Cl int value for a given substrate. A theoretical in vivo Cl int value was calculated from mean values of V max and K m with the use of the microsomal protein yield as a scaling factor (Houston, 1994). The average liver weight of SC was significantly higher than controls, leading to a higher value of microsomal protein yield. Differences in the theoretical in vivo Cl int values followed the same trend as those in the corresponding V max /K m ratios (Tables 3 and 4). Liver enlargement is also very common in SCA patients (Embury et al., 1994). Thus, if human P450 and UGT levels were induced in SCA, a significant increase in the intrinsic clearance of a drug per gram of liver tissue would translate to an even greater increase in the total hepatic intrinsic clearance per liver weight. The mechanisms for differential alterations in enzyme activity in SCA need to be determined. SCA patients are known to have increased levels of both unconjugated and conjugated bilirubin as a result of hemolysis and biliary obstruction (Schubert, 1986). Sickle cell transgenic mice are hyperbilirubinemic, with higher conjugated and unconjugated serum bilirubin levels than control mice (Nagar, 2003). An increase in rat ugt activity has been reported after bilirubin treatment both in vitro and in vivo (Sanchez and Tephly, 1973; Munoz et al., 1987; Li et al., 2000). Bilirubin has been implicated in the aryl

6 DRUG METABOLISM IN SICKLE CELL TRANSGENIC MICE 103 FIG. 3.Eadie-Hofstee plots for the formation of BG in control mice (a), SC (b), and SCKO (c) microsomes. Data expressed as mean values, n 6. hydrocarbon receptor-mediated induction of rodent cyp1a1 (Sinal and Bend, 1997; Phelan et al., 1998). In addition, it appears that increased bilirubin levels in mice activate the nuclear hormone receptor, constitutive androstane receptor, resulting in an increase in bilirubin clearance (Huang et al., 2003). On the other hand, recent studies of human UGT1A1 activity have shown that UGT1A1-catalyzed reactions like E3G formation can be slightly inhibited by bilirubin, which acts as a competitive inhibitor (Williams et al., 2002). These reports support the hypothesis that high bilirubin levels may play an important role in the activation and inhibition of various human P450 and UGT isozymes in SCA. The effect of bilirubin on morphine disposition is currently being studied in mouse and rat models. In addition to altered pharmacokinetics, the pharmacodynamics of analgesics also need to be evaluated in SCA. It is also important to note that morphine is converted only to M3G in rodents, whereas in humans it is also glucuronidated at the 6-position, with the resulting morphine 6-glucuronide being much more potent than even the parent compound as an analgesic (Christrup, 1997). How the increased glucuronidation of morphine would thus affect the analgesic activity of morphine in human SCA patients needs to be investigated. However, previous clinical reports indicate that patients require significantly larger doses of morphine for adequate pain control. In conclusion, there is a marked alteration in the in vitro microsomal enzyme kinetics of various ugt and rodent P450 isozymes in the sickle cell transgenic mouse models studied. There is evidence for the induction of the rodent P450 isozyme responsible for converting codeine to morphine and the ugt isozyme responsible for glucuronidating morphine. Further studies in human patients are needed to validate the animal models as well as to clarify the role of altered pharmacokinetics and pharmacodynamics in sickle cell anemia pain. Acknowledgments. We thank Dr. Marie Applie at the Department of Biostatistics, University of Minnesota, for statistical consultation. We also acknowledge the National Institute of Drug Abuse, Research Triangle Institute, for the gift of norbuprenorphine. References Beyer JE (2000) Judging the effectiveness of analgesia for children and adolescents during vaso-occlusive events of sickle cell disease. J Pain Symptom Manage 19: Blume N, Leonard J, Xu ZJ, Watanabe O, Remotti H, and Fishman J (2000) Characterization of Cyp2d22, a novel cytochrome P450 expressed in mouse mammary cells. Arch Biochem Biophys 381: Bock KW, Lilienblum W, and Pfeil H (1982) Functional heterogeneity of UDPglucuronosyltransferase activities in C57BL/6 and DBA/2 mice. Biochem Pharmacol 31: Cheng Z, Rios GR, King CD, Coffman BL, Green MD, Mojarrabi B, Mackenzie PI, and Tephly TR (1998) Glucuronidation of catechol estrogens by expressed human UDP-glucuronosyltransferases (UGTs) 1A1, 1A3, and 2B7. Toxicol Sci 45: Christrup LL (1997) Morphine metabolites. Acta Anaesthesiol Scand 41: Coffman BL, Rios GR, and Tephly TR (1996) Purification and properties of two rat liver phenobarbital-inducible UDP-glucuronosyltransferases that catalyze the glucuronidation of opioids. Drug Metab Dispos 24: Coffman BL, Rios GR, King CD, and Tephly TR (1997) Human UGT2B7 catalyzes morphine glucuronidation. Drug Metab Dispos 25:1 4. Dampier CD, Setty BNY, Logan J, Ioli JG, and Dean R (1995) Intravenous morphine pharmacokinetics in pediatric patients with sickle cell disease. J Pediatr 126: Ebner T, Remmel RP, and Burchell B (1993) Human bilirubin UDP-glucuronosyltransferase catalyzes the glucuronidation of ethinylestradiol. Mol Pharmacol 43: Embury SH, Hebbel RP, Mohandas N, and Steinberg MH (1994) Sickle Cell Disease: Basic Principles and Clinical Practice, Raven Press, New York. Fabry ME, Nagel RL, Pachnis A, Suzuka SM, and Constantini F (1992) Higher expression of human s - and -globins in transgenic mice: hemoglobin composition and hematological consequences. Proc Natl Acad Sci USA 89: Fisher MB, Campanale K, Ackerman BL, Vandenbranden M, and Wrighton SA (2000) In vitro glucuronidation using human liver microsomes and the pore-forming peptide alamethicin. Drug Metab Dispos 28: Gall WE, Zawada G, Mojarrabi B, Tephly TR, Green MD, Coffman BL, Mackenzie PI, and Radominska-Pandya A (1999) Differential glucuronidation of bile acids, androgens and estrogens by human UGT1A3 and 2B7. J Steroid Biochem Mol Biol 70: Gibaldi M and Perrier D (1982) Pharmacokinetics, 2nd ed, pp , Marcel Dekker, Inc, New York. Glare PA and Walsh TD (1991) Clinical pharmacokinetics of morphine. Ther Drug Monit 13:1 23. Hanioka N, Hoshikawa Y, Mitsui T, Oguri K, and Yoshimura H (1990) Species difference in codeine diphosphate-glucuronyltransferase activity of liver microsomes. J Pharmacobio-Dyn 13: Houston JB (1994) Utility of in vitro drug metabolism data in predicting in vivo metabolic clearance. Biochem Pharmacol 47: Huang W, Zhang J, Chua SS, Qatanani M, Han Y, Granata R, and Moore DD (2003) Induction of bilirubin clearance by the constitutive androstane receptor (CAR). Proc Natl Acad Sci USA 100: Huet P-M and Villeneuve J-P (1983) Determinants of drug disposition in patients with cirrhosis. Hepatology 3: Iyanagi T, Emi Y, and Ikushiro S (1998) Biochemical and molecular aspects of genetic disorders of bilirubin metabolism. Biochim Biophys Acta 1407: Kenworthy KE, Clarke SE, Andrews J, and Houston JB (2001) Multisite kinetic models for CYP3A4: simultaneous activation and inhibition of diazepam and testosterone metabolism. Drug Metab Dispos 29: King CD, Rios GR, Green MD, MacKenzie PI, and Tephly TR (1997) Comparison of stably expressed rat UGT1.1 and UGT2B1 in the glucuronidation of opioid compounds. Drug Metab Dispos 25: Kobayashi K, Yamamoto T, Chiba K, Tani M, Shimada N, Ishizaki T, and Kuroiwa Y (1998)

7 104 NAGAR ET AL. Human buprenorphine N-dealkylation is catalyzed by cytochrome P450 3A4. Drug Metab Dispos 26: Lear L, Nation RL, and Stupans I (1991) Influence of morphine concentration on detergent activation of rat liver morphine-udp-glucuronosyltransferase. Biochem Pharmacol 42 (Suppl):S55 S60. Li YQ, Prentice DA, Howard ML, Mashford ML, and Desmond PV (2000) Bilirubin and bile acids may modulate their own metabolism via regulating uridine diphosphateglucuronosyltransferase expression in the rat. J Gastroenterol Hepatol 15: Maddrey WC, Cukier JO, Maglalang AC, Boitnott JK, and Odell GB (1978) Hepatic bilirubin UDP-glucuronyltransferase in patients with sickle cell anemia. Gastroenterology 74: Mohammed SS, Ayass M, Mehta P, Kedar A, Gross S, and Derendorf H (1993) Codeine disposition in sickle cell patients compared with healthy volunteers. J Clin Pharmacol 33: Munoz ME, Esteller A, and Gonzalez J (1987) Substrate induction of bilirubin conjugation and biliary excretion in the rat. Clin Sci (Lond) 73: Nagar SV (2003) Metabolism of Opioid Analgesics in Sickle Cell Transgenic Mice. Ph.D. thesis, University of Minnesota, Minneapolis, MN. Nelson MH, Birnbaum AK, and Remmel RP (2001) Inhibition of phenytoin hydroxylation in human liver microsomes by several selective serotonin re-uptake inhibitors. Epilepsy Res 44: Oda Y and Kharasch ED (2001) Metabolism of levo- -acetylmethadol (LAAM) by human liver cytochrome P450: involvement of CYP3A4 characterized by atypical kinetics with two binding sites. J Pharmacol Exp Ther 297: Paszty C, Brion CM, Manci E, Witkowska HE, Stevens ME, Mohandas N, and Rubin EM (1997) Transgenic knockout mice with exclusively human sickle hemoglobin and sickle cell disease. Science (Wash DC) 278: Patwardhan RV, Johnson RF, Hoyumpa A Jr, Sheehan JJ, Desmond PV, Wilkinson GR, Branch RA, and Schenker S (1981) Normal metabolism of morphine in cirrhosis. Gastroenterology 81: Pauling L, Itano H, Singer SJ, and Wells IC (1949) Sickle cell anemia: a molecular disease. Science (Wash DC) 110: Phelan D, Winter GM, Rogers WJ, Lam JC, and Denison MS (1998) Activation of the Ah receptor signal transduction pathway by bilirubin and biliverdin. Arch Biochem Biophys 357: Sanchez E and Tephly TR (1973) Activation of hepatic microsomal glucuronyl transferase by bilirubin. Life Sci 13: Schubert TT (1986) Hepatobiliary system in sickle cell disease. Gastroenterology 90: Senafi SB, Clarke DJ, and Burchell B (1994) Investigation of the substrate specificity of a cloned expressed human bilirubin UDP-glucuronosyltransferase: UDP-sugar specificity and involvement in steroid and xenobiotic glucuronidation. Biochem J 303: Sinal CJ and Bend JR (1997) Aryl hydrocarbon receptor-dependent induction of Cyp1a1 by bilirubin in mouse hepatoma Hepa 1c1c7 cells. Mol Pharmacol 52: Steinberg MH (1999) Drug therapy: management of sickle cell disease. N Engl J Med 340: Tyndale RF, Droll KP, and Sellers EM (1997) Genetically deficient CYP2D6 metabolism provides protection against oral opiate dependence. Pharmacogenetics 7: Wilkinson GR and Shand DG (1975) A physiological approach to hepatic drug clearance. Clin Pharmacol Ther 18: Williams JA, Ring BJ, Cantrell VE, Campanale K, Jones DR, Hall SD, and Wrighton SA (2002) Differential modulation of human UGT1A1 activity. Drug Metab Rev 34 (Suppl 1):196. Yaster M, Kost-Byerly S, and Maxwell LG (2000) The management of pain in sickle cell disease. Pediatr Clin N Am 47: