Abstract
RG 12525 is a new chemical entity recently evaluated for the treatment of type II diabetes. Clinical studies have previously identified the tetrazole N2-glucuronide conjugate of RG 12525 as the predominant metabolite in plasma following oral administration of RG 12525. Species differences in RG 12525 glucuronidation were first investigated with incubations of RG 12525 with rat, monkey, and human hepatocytes. The results showed theN2-glucuronide to be the major metabolite in human and monkey samples, with only low levels observed for the rat. The formation of this glucuronide by human liver microsomes was subsequently characterized. RG 12525 N2-glucuronidation was found to have a pH optimum of 7.0 to 7.5 and demonstrated a high affinity with a Km range of 16.6 to 21.1 μM RG 12525 (n = 3). The rate ofN2-glucuronide formation ranged from 2.5 to 15.4 nmol of RG 12525 N2-glucuronide formed/min/mg of protein (∼6-fold) in the 21 samples assayed. The reaction was inhibited by known substrates for glucuronidation, with imipramine (62%), naringenin (44%), and scopoletin (38%) producing the largest degree of inhibition at equimolar concentrations of substrate and inhibitor. Of the eight expressed UDP-glucuronosyltransferase (UGT) forms assayed, UGT1A1 and 1A3 displayed the highest rate of RG 12525N2-glucuronidation (0.109 and 0.125 nmol/min/mg, respectively). Finally, low levels of N2-glucuronidation of RG 12525 by human jejunum microsomes were demonstrated, suggesting that presystemic clearance via glucuronidation may constitute a barrier to bioavailability.
RG 125253 (Fig. 1) is a new chemical entity recently evaluated for the treatment of non-insulin-dependent diabetes mellitus. Preclinical pharmacology studies have shown that RG 12525 is an agonist of the peroxisomal proliferation-activated receptor-γ nuclear receptor, a target that has been shown to be involved in controlling hyperglycemia in humans (Willson and Wahli, 1997). As human trials with RG 12525 have advanced, it has became imperative to characterize the major route(s) of metabolism in order to define factors that could influence drug bioavailability and pharmacokinetics, suggest potential drug interactions, and guide the selection of appropriate species for toxicology studies. The focus of such an approach is generally the cytochrome P450 system; however, there is increasing evidence that glucuronidation can be a major route of elimination for both endo- and xenobiotics. With the recent improvements in the characterization of the substrate specificity and metabolic capacity of UGT isozymes and the commercial availability of several expressed UGTs, the importance of glucuronidation to the overall development of a new chemical entity can be evaluated.
N-Glucuronidation has been shown to be an important pathway for the clearance of compounds from many therapeutic classes, including antidepressants, antihistamines, and antihypertensives (reviewed inBurchell et al., 1995 and Green and Tephly, 1998). Nohara (1980) first identified a tetrazole N1-glucuronide in the urine of animals dosed with AA-344 [6-ethyl-3-(1H-tetrazol-5-yl)chromone]. Formation of a novel N2-glucuronide was later identified as a major route of metabolism for the tetrazole-containing angiotensin II receptor antagonist losartan using both in vitro (Stearns et al., 1992;Huskey et al., 1993) and in vivo techniques (Krieter et al., 1995). In fact, the potential for intestinal N-glucuronidation to act as a barrier to drug absorption was first demonstrated in rats using losartan (Krieter et al., 1995). Within the same therapeutic drug class, irbesartan has also been shown to undergo tetrazoleN2-glucuronidation in humans, although metabolite levels in plasma were relatively low (∼5%) following oral administration and contributed to less than 10% of the overall elimination of the parent drug (Chando et al., 1998). The specific isozyme responsible for the glucuronidation of irbesartan has not been reported; however, expressed UGT1*6 (UGT1A6 by current nomenclature) has been shown to catalyze the glucuronidation of model tetrazoles (Huskey et al., 1994a), and UGT2B7 has been reported to conjugate losartan (Rios et al., 1998).
The initial objective of this work was to describe the importance ofN-glucuronidation to the in vivo biotransformation of RG 12525 in humans. The identification and differentiation of RG 12525N1-glucuronide (RPR 240818) and N2-glucuronide (RPR 241098) coincided with the finding of extremely high plasma levels (∼50–80 μg/ml range) of RPR 241098 following oral administration of RG 12525 to humans (Fayer et al., 2001). These findings dictated that additional characterization experiments be conducted to understand the impact of the glucuronidation pathway on the overall pharmacokinetics of RG 12525. The in vitro studies used several approaches: kinetic analysis and chemical inhibition of RG 12525N-glucuronidation by human liver microsomes, evaluation of expressed UGTs for activity, and finally, demonstration of activity by human jejunal microsomes. The relevance of marked species differences in RG 12525 N-glucuronidation is also discussed.
Materials and Methods
Chemicals.
RG 12525 (potassium salt) and [14C]RG 12525 (52 mCi/mmol, >98.5% pure) were provided by the Aventis Pharma Chemical Processing Center (Collegeville, PA). Formic acid, HPLC-grade methanol, and HPLC-grade acetonitrile were purchased from EM Scientific (Gibbstown, NJ). Liquid scintillation fluid (Ultima-Flo M and Ultima Gold-XR) was purchased from the Packard Instrument Company (Meriden, CT). UDPGA (sodium salt), p-nitrophenol,p-hydroxybiphenyl, scopoletin, naringenin, fenoprofen, buprenorphine, imipramine, β-glucuronidase (fromEscherichia coli), and chlorpromazine were purchased from Sigma Chemical Co. (St. Louis, MO). Saccharolactone was purchased from ICN Biomedicals, Inc. (Aurora, OH). All other chemicals were obtained from standard vendors.
Biological Reagents.
Human liver samples were obtained through organ procurement agencies in accordance with proper ethical procedures for consent. The preparation and metabolic characterization of human liver microsome samples RPR-HL-01 through RPR-HL-15 have been reported (Heyn et al., 1996;Stevens et al., 1997). The age (in years), sex, and microsomal P450 content (nmol of P450/mg of protein) for the remaining samples were as follows: HL-16 (56, M, 0.29); HL-17, (57, M, 0.25); HL-18, (15, F, 0.32); HL-19 (30, M, 1.17), HL-20 (53, F, 0.64), and HL-21 (56, F, 0.55). Human jejunum microsomes (15, M and 34, F) were obtained from the Human Cell Culture Center, Inc. (Laurel, MD). Microsomes prepared from control or human lymphoblastoid cells transfected with the cDNA for human UGT1A1, -1A4, -1A6, or -1A9 were purchased from GENTEST Corp. (Woburn, MA). Microsomes prepared from control or SF9 insect cells infected with a baculovirus containing the cDNA for human UGT1A1, -1A3, -1A6, -1A7, -1A10, or -2B7 were purchased from Panvera Corp. (Madison, WI). Incubations of hepatocytes from rat, rhesus monkey, and human with [14C]RG 12525 were conducted by In Vitro Technologies (IVT; Baltimore, MD). Male rhesus monkey hepatocytes were obtained from Clonetics Corp. (Walkersville, MD). Post-thaw viability was determined to be 50%. Male cryopreserved Sprague-Dawley rat hepatocytes showed 56% viability. Cryopreserved human hepatocytes pooled from three donors (IVT lot 84, 76 years old, F; IVT lot 85, 52 years old, M; and IVT lot 86, 73 years old, F) indicated 74% viability.
Synthesis and Characterization of RG 12525N-Glucuronides.
RPR 240818 and RPR 241098 were synthesized by Dr. David Lythgoe, Aventis Pharma, Dagenham Research Center, UK, based on the procedure ofNohara (1980). A mixture of RG 12525 (2.67 g, 6.29 mmol), acetobromo-α-d-glucuronic acid methyl ester (2.5 g, 6.29 mmol), and silver triflate (1.62 g, 6.29 mmol) was heated to reflux in chloroform (60 ml) in the dark under an atmosphere of nitrogen for 23 h. After this period, a further quantity of silver triflate (1.0 g, 3.9 mmol) was added and heating continued for 2 h. The mixture was then cooled to ambient temperature, filtered, and evaporated, giving a brown solid (4.39 g). The crude solid (comprising two products) was subjected to flash chromatography (silica gel) starting with dichloromethane and gradually increasing to 15% acetone as eluent. The polarity of the eluent was further increased by addition of methanol (gradually to 4%) and then formic acid (gradually to 0.5%). The products containing fractions were combined separately and the eluents evaporated to give 1.04 and 1.24 g of the two expected isomers. These were transformed without further purification as follows: for the first isomer, a solution of the first product (1.04 g) in methanol (21 ml) was treated with a sodium hydroxide solution (0.23 g) in water (5.7 ml) at 0°C and the mixture stirred, keeping the temperature below 5°C (0.5 h). The mixture was then concentrated to approximately half volume under vacuum and then partitioned between water (9.5 ml) and dichloromethane (9.5 ml). The aqueous phase was separated and acidified to pH 7 with 1 M HCl solution. The resulting mixture was concentrated under reduced pressure at ambient temperature, resulting in the separation of a solid, which was collected by filtration and washed with water (1 × 2 ml), then propan-2-ol (2 × 4 ml), and then dried giving sodium 5-{[2-(4-quinolin-2-ylmethoxy)phenoxy methyl]benzyltetrazol-2-yl}-β-d-glucuronate (601 mg, 64%) as an off-white solid. 1H NMR (DMSO-d6, δ) results were as follows: 3.26(t, J = 8 Hz, 1H), 3.4(t, J = 8 Hz, 1H), 3.62(d, J = 8 Hz, 1H), 3.83(t, J = 8 Hz, 1H), 4.38(s, 2H), 5.15(t, J = 8 Hz, 2H), 5.3(s, 2H), 5.79(d, J = 8 Hz, 1H), 6.98(d, J = 8 Hz, 2H), 7.02(d, J = 8 Hz, 2H), 7.27 to 7.32(m, 3H), 7.42 to 7.48(m, 1H), 7.62(t, J = 8 Hz, 1H), 7.7(d, J = 8 Hz, 1H), 7.8(t, J = 8 Hz, 1H), 8.0(d, J = 8 Hz, 1H), 8.03(d, J = 8 Hz, 1H), and 8.42(d, J = 8 Hz, 1H).
For the second isomer, a solution of the second product (1.24 g) in methanol (25 ml) was treated with a sodium hydroxide solution (0.27 g) in water (6.8 ml) at 0°C and the mixture stirred, keeping the temperature below 5°C (0.5 h). The mixture was then concentrated to about half volume under vacuum and then partitioned between water (11 ml) and dichloromethane (11 ml). The remaining steps were identical to those listed above. Upon drying, sodium 5-{[2-(4-quinolin-2-ylmethoxy)phenoxy methyl]benzyltetrazol-1-yl}-β-d-glucuronate (500 mg, 48%) was obtained as an off-white solid. 1H NMR (DMSO-d6, δ) results were as follows: 3.23(H-4′, t, J = 8 Hz, 1H), 3.39(H-3′, t, J = 8 Hz, 1H), 3.59(H-5′, d, J = 8 Hz, 1H), 3.82(H-2′, t, J = 8 Hz, 1H), 4.44(H-23, s, 2H), 5.15(H-16, q, J = 8 Hz, 2H), 5.3(H-11, s, 2H), 5.6(H-1′, d, J = 8 Hz, 1H), 6.97 to 7.02(H-13,14, m, 2H), 7.20 to 7.33(H-19,20,21, m, 3H), 7.47(H-18, d, J = 8 Hz, 1H), 7.62(H-5, t, J = 8 Hz, 1H), 7.68(H-9, d, J = 8 Hz, 1H), 7.8(H-4, t, J = 8 Hz, 1H), 8.0(H-6, d, J = 8 Hz, 1H), 8.03(H-3, d, J = 8 Hz, 1H), and 8.42(H-8, d, J = 8 Hz, 1H).
The structures of the two glucuronides were also characterized in detail by high-resolution NMR spectroscopy. Samples were prepared by dissolving approximately 10 mg of analyte in DMSO-d6, and all experiments were carried out at 30°C on a 500- or 600-MHz Varian Inova system (Varian Instruments, Palo Alto, CA). Double quantum-filtered correlation spectroscopy, heteronuclear multiple quantum coherence, heteronuclear multiple bond correlation, and nuclear Overhauser enhancement using gradients experiments were performed using standard pulse sequences. Gradients were used to suppress artifacts and select coherence transfer pathways in the two-dimensional experiments. All spectral data were consistent with N-glucuronide structures. The site of glucuronidation on the tetrazole ring was established from NOE data as described previously (Stearns et al., 1991; Huskey et al., 1993). Specifically, an NOE was observed between the benzylic protons adjacent to the tetrazole moiety and the anomeric proton (H-1′) of the glucuronic acid moiety in the N1-glucuronide (RPR 240818) following irradiation of the 23-methylene protons at 4.44 ppm (Fig.2). In contrast, no such NOE was observed with the N2-glucuronide, RPR 241098 (data not shown).
RG 12525 Metabolism: In Vivo Studies.
Pooled human plasma samples (n = 4) obtained from subjects 3 h after the oral administration of 600 mg of RG 12525 were used for the initial metabolite identification studies. An aliquot of this pool was diluted with 50 mM phosphate buffer (pH 7.4) and applied to a solid-phase extraction column (Oasis HLB 1-ml cartridges, Waters Corporation, Milford, MA) conditioned according to the manufacturer's recommendations. The sample was washed with water (1 ml), eluted with methanol (2 ml), and evaporated under nitrogen. The residue was then reconstituted in 0.5 ml of either 50 mM Tris buffer (pH 6.8) or the same buffer containing 20 units of β-glucuronidase and incubated for 19 h at 37°C. The incubations were treated with 100 μl of methanol, centrifuged, and the supernatant analyzed by HPLC as described below. Recovery of RG 12525 or14C-labeled products (hepatocyte samples, see below) from the solid-phase cartridge system using in vitro or in vivo samples was approximately 90%. Another aliquot of human plasma was analyzed by HPLC coupled with the SCIEX API III+ (Toronto, Canada) mass spectrometer with the IonSpray (electrospray) interface operated in the positive mode. Separation was achieved using a Keystone BDS Hypersil C18, 3 × 250-mm column (Keystone Scientific, Bellefonte, PA) with a flow rate of 0.5 ml/min, with 50 μl/min directed to the mass spectrometer. The mobile phases for this method (method 1) consisted of aqueous 0.1% formic acid (A) and methanol (B). The gradient was 80% A at t = 0 min; 80% A → 20% A from 0–20 min, and then held for 35 min before returning to initial conditions.
RG 12525 Metabolism: In Vitro Studies.
Human, monkey, and rat cryopreserved hepatocytes (2 × 106 cells/1-ml well) were incubated with 25 μM [14C]RG 12525 (0.5 μCi) for 2 and 6 h following a 30-min preincubation at 37°C. Positive metabolism controls using 7-ethoxycoumarin (75 μM final concentration) were also conducted under identical conditions to verify enzyme activity. The reactions were terminated by freezing the supernatant and cells. The [14C]RG 12525 samples were later thawed, sonicated for 10 min to disrupt any intact cells, and subjected to the solid-phase extraction method described above, but without the water wash step. Extraction recovery of radiolabel averaged 92%. Following methanol evaporation, the samples were reconstituted in 70% aqueous acetonitrile and analyzed using a modification of HPLC method 1. This method (method 2) was also used for the analysis of all microsomal incubations where RG 12525 glucuronidation was measured. Specifically, a BDS Hypersil C18 column (3 × 250 mm) was eluted at 0.5 ml/min using the following gradient of A (0.1% formic acid) and B (acetonitrile); 0 min, 70% A; 0 to 5 min, 70% A → 65% A; 5 to 20 min, 65% A → 20% A, held to 26 min, then returned to 70% A followed by column equilibration. Radioactivity and/or UV detection (280 nm) were used in series.
Microsomal glucuronidation assays were performed in a total volume of 300 μl containing 50 mM Tris-HCl buffer (pH 7.5), 3 mM MgCl2, 0.1 mg of microsomal protein (0.3 mg for expressed UGTs), and 200 μM RG 12525. Samples were preincubated for 3 min at 37°C in a shaking water bath, and reactions were started by the addition of UDPGA (2 mM final concentration). The incubation times for assays with liver, intestine, and expressed UGT microsomes were 15, 30, and 60 min, respectively. Reactions were terminated by the addition of 100 μl of cold acetonitrile, and the protein was precipitated by brief centrifugation. The supernatant was removed and analyzed by HPLC method 2 (see above). Preliminary experiments showed that the addition of the β-glucuronidase inhibitor saccharolactone (5 mM final concentration) to human liver microsome incubations resulted in a consistent decrease in the rate of RG 12525 glucuronidation (data not shown). This component was therefore omitted from future studies. Chemical inhibitors were added in 50 to 100% methanol, with final incubation organic solvent concentrations ranging from 0.8 to 2.5%. Consistent with the results of Hiller et al. (1999), we found this range of methanol concentrations to have a negligible effect on catalytic activity (<5%). For each experiment, standard curves were generated using incubation samples containing 0.5 to 25 nmol each of the RG 12525 glucuronide conjugates.
Data Analysis.
HPLC data were acquired and processed using the Waters Millennium Chromatography Manager software (versions 2.1 and 3.0). Kinetic analysis of the human liver microsomal RG 12525 glucuronidation experiments was performed using a one-site model (best fit) and linear regression analysis of the Eadie-Hofstee plots (Segel, 1976) (GraphPad Software, Inc., San Diego, CA).
Results
RG 12525 Glucuronidation in Vivo.
Extracts of human plasma samples drawn 3 h following an oral dose of 600 mg of RG 12525 were analyzed by liquid chromatography/tandem mass spectrometry using HPLC method 1 described under Materials and Methods. The total ion chromatogram (Fig.3A) showed an intense signal at Rt 31.8 min, corresponding to RG 12525, and a peak of slightly lower intensity at Rt 30.8 min. This earlier eluting peak gave a product ion spectra with a distinctive loss of 176 amu to produce the 424 m/z parent ion, indicating the addition of glucuronic acid to the parent compound (Fig. 3B). This spectrum was later found to be identical to that of the RG 12525 N2-glucuronide standard (RPR 241098, Fig. 3C). A minor analyte at Rt 29.0 min was found to have the same molecular ion at 600 m/z, a product ion spectra similar to that of the N2-glucuronide, and an identical spectrum to that of RG 12525 N1-glucuronide standard (data not shown). The assignment of RG 12525N1- and N2-glucuronide standards by NOE analysis (vide infra) and the comparison of HPLC retention times led us to identify the N2-glucuronide as the predominant isomer in human plasma. In a similar study in which patients were administered an oral dose of 1600 mg of RG 12525, plasma levels of the N1- and N2-glucuronides were reported to range from 2.1 to 9.4 and 51.8 to 77.8 μg/ml, respectively (n = 4 subjects) (Fayer et al., 2001).
RG 12525 Glucuronidation by Human, Monkey, and Rat Hepatocytes.
Other investigators have reported dramatic species differences in rates of N-glucuronidation from both in vitro and in vivo studies (Perrier et al., 1994; Hiller et al., 1999). To compare metabolic profiles between rat, monkey, and human in an in vitro model that integrates both oxidative and conjugation pathways, cryopreserved hepatocytes were used. Figure 4 shows the differences in the radiochromatograms generated following the incubation of human (A) and Sprague-Dawley rat (B) hepatocytes with 25 μM RG 12525 for 6 h. The N2-glucuronide was the major metabolite produced by human hepatocytes at 15.2% of total radioactivity, but accounted for less than 1.5% of radioactivity for the rat hepatocyte incubations (n = 2). In contrast, levels of the N2-glucuronide were similar for human and rhesus monkey hepatocyte incubations (15.2 and 10.6% of total radioactivity, respectively, data not shown). TheN1-glucuronide was detected at low levels in the human and rhesus monkey hepatocyte incubations (∼1.5% of total radioactivity in each case), but was not produced by rat hepatocytes. Significant levels of oxidative metabolites were also observed in human hepatocytes. Metabolites M2 and M3 were tentatively identified as dihydrodiols (458 m/z), whereas M4 gave spectral data consistent with O-dealkylation (283m/z, data not shown).
Effect of Incubation Conditions on RG 12525 Glucuronidation by Human Liver Microsomes.
Both the pH and composition of the incubation buffer have been reported to significantly alter rates of glucuronidation for both liver microsomes (Huskey et al., 1993) and expressed UGTs (Green and Tephly, 1996). Therefore, several preliminary experiments were conducted to evaluate the effect of various incubation parameters on RG 12525N2-glucuronidation. Figure 5shows the pH dependence of RG 12525 N2-glucuronidation by human liver microsomes with Tris buffer compared with phosphate buffer. Enzyme activity was approximately 20% greater with Tris buffer at the pH optimum of 7.0 to 7.5. RG 12525 N1-glucuronide was not observed regardless of the buffer composition or pH. Also, the rate ofN2-glucuronide formation by human liver microsomes was found to be linear using a protein concentration of 0.33 mg/ml and an incubation time of 15 min.
Kinetics Analysis and Interindividual Variation of RG 12525 Glucuronidation.
The kinetics of RG 12525 glucuronidation were determined for three human liver microsome samples. The results were analyzed by the Eadie-Hofstee method to allow identification of the involvement of multiple enzymes. Due to high enzyme activity and substantial substrate depletion during the course of the incubations, the integrated form of the Michaelis-Menten equation using an average substrate concentration (Segel, 1976) was used to minimize error in the kinetic parameters.N2-Glucuronide formation was best described by a one-enzyme model (r2avg = 0.86), and the kinetic parameters resulting from linear regression are given in Table 1. TheKm values ranged from 16.6 to 21.1 μM RG 12525 and the Vmax values from 3.54 to 5.94 nmol of N2-glucuronide formed/min/mg of protein. The average intrinsic clearance (Vmax/Km) for these samples thus calculates to 243 μl/min/mg of protein. This kinetic information was also necessary to determine a saturating substrate concentration (200 μM RG 12525) for the measurement of rates of N2-glucuronidation in a bank of human liver microsomes. Figure 6 shows that activities ranged from 2.5 to 15.4 nmol/min/mg of protein, or 6.2-fold, for the 21 samples assayed. These activities were approximately 100-fold greater than the activity measured in two human intestine microsome samples (Table 2).
RG 12525 Glucuronidation by Expressed UGTs.
Eight commercially available expressed UGT forms were assayed to differentiate the ability of various UGTs to catalyze theN2-glucuronidation of RG 12525. Table 2 shows that UGT1A1 and -1A3 had comparable levels of activity (0.109 and 0.125 nmol/min/mg, respectively), with lower turnover observed for UGT2B7 (0.043 nmol/min/mg). No activity was observed for UGT1A4, -1A6, -1A7, -1A9, or -1A10 (limit of detection corresponding to >0.01 nmol/min/mg) despite the use of a high substrate concentration (200 μM) to identify low-affinity UGT forms. These data must be interpreted with caution since the level of expression for each UGT varies depending on the manufacturer (expression systems used and the sample lot) and is only measured indirectly by enzyme activity. Also, reaction kinetics and incubation conditions were not evaluated for each UGT; instead, all forms were assayed under identical conditions.
Chemical Inhibition of RG 12525 Glucuronidation.
Several compounds known to undergo glucuronidation by expressed UGTs (Green et al., 1998) were evaluated as inhibitors of RG 12525N2-glucuronidation by human liver microsomes to further define the UGT form(s) responsible for the N2-conjugation reaction (Fig. 7). All inhibitors were studied at a single concentration (200 μM). Scopoletin, naringenin, and p-hydroxybiphenyl undergo O-glucuronidation by UGT1A3 but are not substrates for UGT1A4. For RG 12525N2-glucuronidation, the extent of inhibition ranged from 22% for p-hydroxybiphenyl to 45% for naringenin. These results are consistent with the expressed UGT data in suggesting that UGT1A3 is involved in the N2-glucuronidation of RG 12525. Also, buprenorphine and fenoprofen O-glucuronidation are likely UGT1A3-selective reactions (Green et al., 1998), and each produced approximately 24% inhibition of RG 12525N2-glucuronidation. However, the most pronounced inhibition (62%) was observed with imipramine, a compound with a relatively low intrinsic clearance (Km/Vmax) by UGT1A3 when compared with UGT1A4.
Discussion
This report documents the importance ofN-glucuronidation to the metabolism of RG 12525 using both in vivo and in vitro approaches. The importance of defining the enzymology of new chemical entities and the ability of this information to predict interindividual variability in pharmacokinetics and clinical drug interactions are well recognized (Guengerich, 1997; Evans and Relling, 1999). The fact that the majority of drug discovery/development studies focus primarily on cytochrome P450-catalyzed reactions rather than conjugation reactions such as glucuronidation may be attributable in part to the comparatively slower development of in vitro tools for the characterization of UGT-dependent reactions. The commercial availability of most human UGT forms at high expression levels has occurred only recently, and chemical inhibitors lack specificity and potency. To our knowledge, the report by Hiller et al. (1999) was the first comprehensive study to extend the clinical observation of N-glucuronidation to in vitro studies using expressed UGTs and chemical inhibition, methods commonly used for the evaluation of P450-mediated reactions.
Plasma analysis of subjects treated with RG 12525 in dose escalation studies has clearly shown N-glucuronidation to be the major route of biotransformation, with circulating conjugate levels 10- to 20-fold higher than parent drug amounts (Fayer et al., 2001). Furthermore, NOE experiments were able to definitively assign theN2 conjugate as the predominate isomer. The relatively recent discovery and limited literature base of tetrazoleN2-glucuronides may be due to analytical difficulties. BothHuskey et al. (1993) and Chando et al. (1998) used extended gradients to separate the N1- and N2-glucuronide isomers. Similarly, the separation of the two RG 12525 conjugates from the parent required relatively shallow gradient elution on a 25-cm HPLC column with selectivity parameters (α value) of 1.04 for the RG 12525/N2-glucuronide pair and 1.05 for theN1/N2-glucuronide pair. In addition, the MS fragmentation patterns of the conjugates were essentially identical.
The results from the species metabolite profile comparison study show that the extent of RG 12525 N-glucuronidation in hepatocytes was similar for rhesus monkey and human but much lower for rat. These results are consistent with those for losartan (Krieter et al., 1995) and irbesartan (Perrier et al., 1994); however,N2-glucuronidation of losartan was observed in the perfused rat intestine. In contrast, rat liver microsomes were found to catalyze the N2-glucuronidation of 5-(4′-methyl[1,1′-biphenyl]-2-yl)-1H-tetrazole at an approximately 10-fold faster rate compared with human liver microsomes, suggesting that species differences are substrate-specific (Huskey et al., 1994b). It is important to recognize that variations in microsomal incubation conditions may contribute to species and/or substrate differences in glucuronidation. The N2-glucuronidation of losartan and irbesartan has been shown to have a pH optimum of approximately 5.0 for human liver microsome studies and the ratio ofN1- to N2-glucuronides can change more than 5-fold depending on the pH (Huskey et al., 1994b; Perrier et al., 1994). Our results for human liver microsomes show maximum rates ofN2-glucuronidation at pH 7.0 to 7.5, arguing against the proposed hypothesis of a direct relationship between an optimum pH for glucuronidation and the pKa (∼4.9) of the tetrazole moiety (Huskey et al., 1993). The observation of a maximum rate of glucuronidation at pH ∼7.5 also shows relevance to the physiological environment of the endoplasmic reticulum, where the lumenal pH has been reported as pH 7.1 (Kim et al., 1998). The effects of other incubation components such as Mg2+, alamethicin, and organic solvents on in vitro tetrazole glucuronidation rates have not been investigated.
To our knowledge, the intrinsic clearance value of 243 μl/min/mg for RG 12525 N2-glucuronidation is the highest ever reported for amine conjugation by human liver microsomes. In fact, the rate and/or efficiency of this reaction is significantly higher than the rates measured for the glucuronidation of numerous marker substrates by expressed UGTs (Green et al., 1998; Cheng et al., 1999; Gall et al., 1999). The approximately 6-fold variability in RG 12525N2-glucuronidation is greater than the 2-fold variation in plasma N2-glucuronide levels following administration of a 1600-mg dose to healthy volunteers (Fayer et al., 2001). These in vivo data are consistent with recent reports showing that UGT forms in human hepatic tissue are not subject to polymorphic regulation (Strassburg et al., 1997, 2000). The in vitro variability is, however, comparable with the range of UGT activities reported for several amines. For example,Hiller et al. (1999) observed a 5-fold variation in retigabineN-glucuronidation both in vitro (n = 16 microsome samples) and in vivo (n = 10 subjects).Huskey et al. (1993) reported a 4- to 10-fold variation in rates of in vitro tetrazole N2-glucuronidation (n = 9), depending on the substrate.
In comparison with the field of human P450 reaction phenotyping, there are extremely limited data available on the use of in vitro approaches (i.e., chemical and antibody inhibition, correlation analysis with specific enzyme activities, expressed enzymes) to characterize human UGT reactions. Marker enzyme activities and specific chemical inhibitors do not currently exist; therefore, the recent commercial availability of expressed UGTs could be the most direct method for the identification of the UGT form(s) involved in tetrazole glucuronidation. An early study in which only two expressed human UGT forms (UGT1A6 and -2B4) were assayed with marker triazoles and tetrazoles showed UGT activity for UGT1A6 (Huskey et al., 1994a). However, expressed UGT1A6 obtained from two separate manufacturers failed to catalyze RG 12525 glucuronidation. This is similar to the preliminary results of Rios et al. (1998) showing that UGT1A6 does not catalyze losartan tetrazole glucuronidation. Rather, expressed UGT2B7 was shown to catalyze the N2-glucuronidation of losartan, and as reported here, RG 12525. As alluded to previously, studies using expressed UGT forms to support the characterization of xenobiotic conjugation need to be interpreted with caution and in context with other parameters. Specifically, it is important that enzyme activity, level of enzyme expression (when available), and differences in tissue expression of individual forms are factored into a discussion of the overall relevance of particular forms to in vivo glucuronidation.
Our investigation of RG 12525 N2-glucuronidation was confined primarily to studies with liver microsomes. There is, however, an increasing recognition of the importance of intestinal phase I and phase II enzymes as barriers to the bioavailability of xenobiotics (reviewed by Hall et al., 1999 and Lin et al., 1999). Specific studies recently reported on the regulation of intestinal UGTs indicate that, in contrast to liver, intestinal tissue clearly shows polymorphic regulation of UGT1A1, UGT1A6, UGT2B4, and UGT2B7 (Strassburg et al., 2000). The rate of RG 12525 N-glucuronidation by the two human intestinal microsome samples assayed showed only ∼1% of the activity of the human liver microsome samples. Since expressed UGT1A1 and UGT2B7 were able to catalyze RG 12525N2-glucuronidation, it is possible that these two intestinal microsome samples were deficient in the expression of one or both of these forms. For intestine samples where UGT transcript expression has been demonstrated using the reverse-transcription polymerase chain reaction, the rates of glucuronidation of several amines were equal to or higher than activity for hepatic tissue (Strassburg et al., 1999,2000). It is also possible that UGT enzyme levels and activity were compromised during the preparation of the samples due to tissue anoxia or intestinal protease activity. Additional work on the competency and interindividual variability of human intestinal UGT forms toward amine substrates is clearly warranted.
In summary, we have shown that the high plasma levels of RG 12525N2-glucuronide observed during human clinical trials are consistent with the high intrinsic clearance determined for human liver microsome studies. The species differences in the extent of conjugate formed suggest that the rhesus monkey would be more predictive of human metabolism and pharmacokinetics compared with the rat. We have also found differences in the effect of incubation pH on RG 12525 glucuronidation rates for human liver microsomal UGT activity in comparison with other investigators' findings. Additional studies are required to define the role of intestinal UGTs in amine conjugation and to determine the extent of substrate specificity of human UGT forms for tetrazoles using expressed enzymes.
Acknowledgments
We thank Dr. Vasant Kumar for his expertise with the NMR analysis, Dr. Peter Zannikos for his insights on RG 12525 pharmacokinetics, and Dr. David Lythgoe for the synthesis and characterization of the glucuronide standards.
Footnotes
-
Send reprint requests to: Jeffrey C. Stevens, Ph.D., Pharmacia Corporation, 301 Henrietta St., 7265–300-306, Kalamazoo, MI 49007. E-mail: jeffrey.c.stevens{at}am.pnu.com
-
↵1 Current address: Pharmacia Corporation, 301 Henrietta St., 7265-300-306, Kalamazoo, MI 49007.
-
↵2 Current address: Eli Lilly and Company, Lilly Corporate Center, Indianapolis, IN 46285.
- Abbreviations used are::
- RG 12525
- 2-[[4-[[2-(1H-tetrazol-5-ylmethyl)phenyl]methoxy]phenoxy]methyl] quinoline
- RPR 240818 or RG 12525 N1-glucuronide
- 5-{[2-(4-quinolin-2-ylmethoxy)phenoxymethyl]benzyltetrazol-1-yl}-β-d-glucuronate
- RPR 241098 or RG 12525 N2-glucuronide
- 5-{[2-(4-quinolin-2-ylmethoxy)phenoxy methyl]benzyltetrazol-2-yl}-β-d-glucuronate
- RG 07202
- 2-({4-[3-(1H-tetrazol-5-yl)propoxy]phenoxy}methyl)quinoline
- RPR 108685
- 5-(2-methylbenzyl)-1H-tetrazole
- UGT
- UDP-glucuronosyltransferase
- UDPGA
- UDP-glucuronic acid
- HPLC
- high-performance liquid chromatography
- NOE
- nuclear Overhauser effect
- DMSO
- dimethyl sulfoxide
- Rt
- retention time
- Received August 1, 2000.
- Accepted December 11, 2000.
- The American Society for Pharmacology and Experimental Therapeutics