Introduction

Top Abstract Introduction Results Discussion References

Increased concerns in the United States and in many other countries have been focused on obesity, which in many cases is the prelude to developing more severe disorders such as Type II diabetes, hypertension, and cardiovascular diseases (Eckel and Krauss, 1998; Wickelgren, 1998). Obesity is the result of an imbalance between caloric intake and energy expenditure; excess energy is therefore accumulated in the form of triglycerides in adipocytes, and this accumulation increases with the expansion of adipose tissues. In general, mechanisms for treatment of obesity include increasing energy consumption, stimulating fat metabolism, and reducing nutrient intake (Campfield et al., 1998; Kordik and Reitz, 1999). The discovery of a unique -adrenergic receptor subtype, namely the 3 receptor, on the surface of adipocytes that stimulates lipolysis has led to speculation that a selective agonist of the ₃-adrenergic receptor may be used for treating obesity (Danforth and Himms-Hagen, 1997; Weyer et al., 1999). In experimental animals, treatment with agonists of the ₃-adrenergic receptor led to increases in metabolic rate, weight loss, and improved glucose tolerance (Bloom et al., 1992; Cawthorne et al., 1992). Similar studies in human trials, however, were inconclusive, and results were complicated by the lack of selectivity of the agonists studied for the ₃ receptor versus the ₁/₂ receptors (Weyer et al., 1999). Further investigation indicated that the homology between rodent and human ₃-adrenergic receptors was high whereas differences in the pharmacology of these receptors were apparent (Pietri-Rouxel and Strosberg, 1995; Lipworth, 1996). On the other hand, a small population of human subjects carrying a mutated ₃ receptor were identified and reported to have tendency to gain weight and exhibit an earlier onset of noninsulin-dependent diabetes (Clement et al., 1995; Walston et al., 1995). These data suggest that a potential therapeutic agent for the treatment of obesity would need to be potent and selective toward the human ₃-adrenergic receptor.

The structures and pharmacokinetic properties of several agonists of the human ₃-adrenergic receptor have been described (Mathvink et al., 1999; Shih et al., 1999). In addition, (R)-N-[4-[2-[[2-hydroxy-2-(pyridin-3-yl)ethyl]amino]ethyl]phenyl]-4-[4-(4-trifluoro-methylphenyl)thiazol-2-yl]benzenesulfonamide (1¹, Fig. 1), a nanomolar agonist of the human ₃-adrenergic receptor with a 1000-fold selectivity over ₁ and ₂ receptors, has been reported (Mathvink et al., 2000), and its pharmacokinetic characteristics described in the preceding article (Stearns et al., 2002). We report herein the results derived from studies of the biotransformation and excretion of 1 in bile duct-cannulated rats. Five major metabolites, including a taurine conjugate and a novel isethionic acid conjugate, were characterized by liquid chromatography/tandem mass spectrometry (LC/MS/MS). These metabolites were further investigated in vitro with respect to their mechanisms of formation, whereas the parent compound 1 was evaluated for P-glycoprotein (Pgp)-mediated transport.

View larger version (14K): [in this window] [in a new window]   Fig. 1.   LC/MS/MS product ion spectrum of parent compound 1. The spectrum was obtained by CID of the MH+ ion at m/z 625, and the proposed origins of key fragment ions are as indicated.

Experimental Procedures

Materials. Anilino-2-³H-labeled 1 and 1 hydrochloride were synthesized at Merck Research Laboratories (Mathvink et al., 2000; Stearns et al., 2002). The specific activity of ³H-labeled 1 was 1750 mCi/mmol with a radiochemical purity greater than 97%. [³H]Vinblastine was obtained from Amersham Pharmacia Biotech, Inc. (Piscataway, NJ). Coenzyme (CoA), dithiothreitol, HEPES, isethionic acid, magnesium chloride, monoamine oxidase (EC 1.4.3.4.), NADPH, quinidine, taurine, trifluoroacetic acid, and troleandomycin were purchased from Sigma-Aldrich (St. Louis, MO). All other reagents and solvents were obtained from Fisher Scientific (Fair Lawn, NJ) or Aldrich Chemical Co. (Milwaukee, WI). BondElut C₁₈ solid phase extraction cartridge columns were obtained from Varian Chromatography Systems (Walnut Creek, CA). Sulfaphenazole was a gift from Dr. F. P. Guengerich (Vanderbilt University, Nashville, TN).

Instrumentation and Analytic Methods. LC/MS/MS was carried out on a Sciex API III⁺ tandem mass spectrometer (PerkinElmer, Ontario, Canada) interfaced to an HPLC system consisting of two Shimadzu LC-10A pumps and a static-bed mixer (Shimadzu Scientific Instruments, Kyoto, Japan). LC/MS/MS experiments were performed with an ion spray interface and positive ion detection. The voltage was 5 kV, and orifice potential was 65 V. Argon was used as the collision gas for collision-induced dissociation (CID) at a thickness of 1.3 × 10¹⁴ atom/cm², and the collision energy was 30 eV. Chromatography was performed on a DuPont Zorbax Rx-C₈ column (4.6 × 250 mm, 5 µm; DuPont, Wilmington, DE), and samples were delivered at a rate of 1 ml/min with 1:25 split. The mobile phase consisted of aqueous acetonitrile (10% methanol and 0.05% trifluoroacetic acid) and was programmed by a linear increase from 10 to 70% acetonitrile during a 30-min period with isocratic elution at 70% acetonitrile for a further 10 min. Product ion scans were performed on selected MH⁺ ions at a scan speed of ~1 s per 225 atomic mass units. Multiple reaction monitoring detection (MRM) was based on one mass transition per metabolite at dwell times of 400 ms per channel.

Synthesis of Metabolites of 1. M2.

M4. A solution of 2-(4-aminophenyl)ethylcarbamic acid tert-utyl ester (142 mg) and pyridine (80 µl) in dichloromethane (5 ml) was treated with 4-{4-[4-(trifluoromethyl)phenyl]-5-trimethylsilyl-1,3-thiazol-2-yl}benzenesulfonyl chloride (286 mg), and the resulting solution was stirred at room temperature for 2 h. Volatiles were removed in vacuo, and the residue was taken up in acetonitrile (8 ml) and treated with hydrogen fluoride in aqueous acetonitrile (7 ml; 4.8%; water/acetonitrile, 15:85, v/v). The mixture was stirred vigorously for 5 h and was then concentrated in vacuo. The residue was purified by flash chromatography (silica gel; 9:1 methylene chloride:10% ammonium hydroxide-methanol eluant) to afford the title compound (240 mg) as a pale yellow solid. ¹H NMR (400 MHz, CD₃OD) 76 8.48 (1H, s), 8.25 (2H, d, J = 8.3 Hz), 8.11 (2H, d, J = 8.3 Hz), 7.85 (4H, m), 6.93 (2H, d, J = 8.4 Hz), 6.91 (2H, d, J = 8.4 Hz), 2.73 (2H, t, J = 7.5 Hz), 2.53 (2H, t, J = 7.5 Hz).

M5. A solution of 4-nitrophenylacetic acid (1.82 g) was added to a solution of acetyl chloride (1.0 ml) in methanol (25 ml), and the mixture was subsequently heated to reflux. After 3 h, the solution was cooled and concentrated under reduced pressure. The residue was partitioned between water and ether, and the ether layer was washed sequentially with saturated aqueous sodium bicarbonate and saturated brine. Drying with anhydrous magnesium sulfate and removal of solvent gave a white solid (1.85 g), which was dissolved in ethyl acetate (25 ml) and treated with 10% palladium on carbon (110 mg). The mixture was shaken under 50 psi of hydrogen for 1 h, then filtered through a pad of Celite, and concentrated in vacuo. The resulting cloudy mobile oil (1.66 g) was used directly in the next step. This oil (290 mg) in dichloromethane (10 ml) was treated sequentially with pyridine (0.13 ml) and 4-{4-[4-(trifluoromethyl)phenyl]-5-trimethylsilyl-1,3-thiazol-2-yl}benzenesulfonyl chloride (715 mg). The mixture was stirred at room temperature for 2 h and was then concentrated in vacuo. The residue was dissolved in acetonitrile (5 ml), and hydrogen fluoride in aqueous acetonitrile (8 ml; 4.8%; water/acetonitrile, 15:85, v/v) was added. The mixture was stirred at room temperature overnight and was then concentrated in vacuo. The residue was stirred with methanol (30 ml) for 1 h, and the solid was collected and washed with methanol to afford a pale pink solid (307 mg). The combined filtrates were concentrated in vacuo, and the residue was purified by flash chromatography (silica gel; 25% ethyl acetate-hexanes eluant) to provide an additional product (483 mg). The combined batches were dissolved in tetrahydrofuran/methanol/water (8 ml; 3:2:1, v/v/v), and lithium hydroxide monohydrate (153 mg) was added. After stirring at room temperature for 1 h, an additional quantity of lithium hydroxide monohydrate (153 mg) was added, and the solution was stirred for a further 1.5 h. Volatiles were removed under reduced pressure, and water (20 ml) was added. The solution was washed with ether, then cooled in ice, neutralized to pH 4 with 5% aqueous hydrochloric acid, and extracted with ethyl acetate. The combined organic extracts were washed with saturated brine, dried with anhydrous magnesium sulfate, and concentrated to an off-white solid. Trituration with hexanes provided a white powder (645 mg). ¹H NMR (400 MHz, CD₃OD) 76 8.20 (2H, d, J = 8.3 Hz), 8.13 (2H, d, J = 8.3 Hz), 8.10 (1H, s), 7.84 (2H, d, J = 8.4 Hz), 7.73 (2H, d, J = 8.4 Hz), 7.16 (2H, d, J = 8.5 Hz), 7.08 (2H, t, J = 8.5 Hz), 3.51 (2H, s).

M1. A mixture of M5 (51 mg) in dry N,N-dimethylformamide (1.0 ml) was cooled to 0°C, and diphenylphosphoryl azide (24 µl) was added, followed by triethylamine (20 µl). The solution was warmed to room temperature, and after 30 min it was cooled in an ice-water bath and treated with a solution of taurine (17 mg) and 1,1,3,3-tetramethylguanidine (34 µl) in dry N,N-dimethylformamide (0.6 ml). The reaction mixture was allowed to slowly warm to room temperature overnight and was then concentrated in vacuo. The yellow oily residue was dissolved in aqueous sodium hydroxide (3 ml, 0.07 N). The aqueous solution was washed with ethyl acetate and was then concentrated in vacuo. The residue was purified by flash chromatography (silica gel; gradient elution with 9:1, then 6:1 methylene chloride:10% ammonium hydroxide-methanol eluant) to afford the title compound (27 mg) as a white powder. ¹H NMR (400 MHz, CD₃OD) 76 8.50 (1H, s), 8.27 (2H, d, J = 8.4 Hz), 8.18 (2H, d, J = 8.2 Hz), 7.75-7.90 (4H, m), 7.60 (1H, broad s), 7.09 (2H, d, J = 8.5 Hz), 7.01 (2H, t, J = 8.5 Hz), 3.25 (2H, s), 3.04 (2H, t, J = 6.8Hz), 2.69 (2H, t, J = 6.8 Hz).

M3. A mixture of M5 (100 mg) and thionyl chloride (2 ml) was warmed at reflux for 1 h. The solution was cooled and concentrated in vacuo, and the residue was azeotroped twice with dry benzene and then dried under high vacuum. The resulting off-white solid was suspended in xylenes (2 ml), and sodium isethionate (14 mg) was added. The mixture was warmed at 145°C for 5 h, cooled to room temperature, and diluted with anhydrous ether (10 ml). The precipitate was collected by filtration, dissolved in 10% ammonium hydroxide-methanol, and purified by flash chromatography (silica gel; 9:1 methylene chloride:10% ammonium hydroxide-methanol eluant) to afford the title compound (19 mg) as a light yellow, crystalline solid. ¹H NMR (400 MHz, CD₃OD) 76 8.20 (2H, d, J = 8.3 Hz), 8.13 (2H, d, J = 8.3 Hz), 8.09 (1H, s), 7.85 (2H, d, J = 8.4 Hz), 7.73 (2H, d, J = 8.4 Hz), 7.16 (2H, d, J = 8.5 Hz), 7.07 (2H, d, J = 8.5 Hz), 4.40 (2H, t, J = 6.9 Hz), 3.56 (2H, s), 3.09 (2H, t, J = 6.9 Hz).

Animal Experiments. All experiments were performed according to procedures approved by the Merck Institutional Animal Care and Use Committee.

Biological Preparations. Human liver samples were obtained from the Pennsylvania Regional Tissue Bank (Exton, PA). An agreement was made between the tissue bank and Merck & Co. for research use of the samples.

Incubations of 1 with Liver Microsomes. The hydrochloride salt of 1 in an aqueous solution was added to rat, dog, monkey, or human liver microsomes (1 mg of protein/ml) or recombinant P450 (0.5-1 mg of protein/ml) suspended in phosphate buffer (0.1 M, pH 7.4) containing EDTA (1 mM). The final volume was 0.5 ml, and substrate concentrations were from 1 to 1000 µM. Microsomes containing an empty cDNA expression vector were used as negative controls. The mixture was incubated at 37°C for 5 min before adding NADPH in phosphate buffer (final concentration, 1 mg/ml) to initiate the reaction. After an additional 5 min (kinetic studies) or 30 min (metabolism studies) incubation, the reaction was quenched by adding 10% aqueous trifluoroacetic acid (60 µl). Apparent K_m and V_max values for 1 metabolism in human liver microsomal incubations were calculated according to the Michaelis-Menten equation.

Incubations of M5 with Rat Liver Mitochondria. Rat liver microsomes or mitochondria were suspended in Tris-buffer (100 µM, pH 7.4) containing EDTA (1 mM). The protein concentrations were 4.3 mg/ml for microsomal incubations and 2.1 mg/ml for mitochondrial incubations. M5 was dissolved in 0.1 M sodium hydroxide solution and was added to the incubation suspension to give a final concentration of 1 mM. ATP (5.5 mg), CoA (0.77 mg), magnesium chloride (1 mg), and dithiothreitol (0.15 mg), all dissolved in Tris-buffer, were subsequently added to the suspension, and the final volume was 1 ml. The mixture was incubated at 37°C for 2 h. The reaction mixture then was acidified to pH 2 with 10% aqueous trifluoroacetic acid and extracted by a C₁₈ solid phase extraction cartridge column. The resulting samples were analyzed by LC/MS/MS.

Bidirectional Transport across Monolayers of Cells Expressing Pgp. The assay was performed according to published procedures (Kim et al., 1998). Briefly, porcine kidney epithelial cells expressing mouse or human Pgp were seeded on polycarbonate membrane filters to form a monolayer and cultured for 4 days. The culture medium was then replaced by a serum-free transport medium consisting of Hanks' balanced salt solution and HEPES buffer (pH 7.4). The bidirectional transport assay was initiated by adding the fresh transport medium into the receiver side and adding the medium containing test compounds ([³H]vinblastine or ³H-labeled 1) into the donor side of the monolayer. Aliquots of samples were taken following incubations at 37°C for certain time intervals and subjected to liquid scintillation counting on a Beckman LS6500 multipurpose scintillation counter (Beckman Coulter, Inc.). Vectorial transport of the test compound was evaluated by the ratio of basolateral (B)-to-apical (A) versus A-to-B transport.

	Results

Top Abstract Introduction Results Discussion References

Metabolism of 1 in Rats. After oral dosing of ³H-labeled 1 at 10 mg/kg, 22% of the administered radioactivity was found in the bile, 6% in the urine, and 75% in the feces (Stearns et al., 2002). After intravenous dosing at 3 mg/kg, 63% of the radioactivity was recovered in the bile, 11% in the urine, and 9% in the feces (Stearns et al., 2002). Bile, urine, and the homogenate of feces were further analyzed by HPLC and LC/MS/MS. Although radiolabeled components in the urine eluted with the HPLC void volume, one component was identified tentatively on the basis of LC/MS/MS data as 2-hydroxy-2-pyridinylacetic acid (data not shown). In the feces, radioactivity recovered after oral dosing was attributed to the nonabsorbed parent compound, whereas three components were identified following intravenous dosing, with 1 being the predominant species (Fig. 2).

View larger version (21K): [in this window] [in a new window]   Fig. 2.   HPLC radiohistogram (A) and UV-vis chromatogram (B) of bile from a rat treated by oral gavage with 3H-labeled 1 (10 mg/kg); HPLC radiohistogram (C) of a feces sample from a rat treated by intravenous injection with 3H-labeled 1 (3 mg/kg). Characterization of M1, parent drug, M2, M3, M4, and M5 is discussed in the text.

View larger version (12K): [in this window] [in a new window]   Fig. 3.   LC/MS/MS product ion spectrum of M5 in the bile from a rat dosed orally with 1. The spectrum was obtained by CID of the MH+ ion at m/z 519, and the proposed origins of key fragment ions are as indicated.

View larger version (11K): [in this window] [in a new window]   Fig. 4.   LC/MS/MS product ion spectrum of M1 in the bile from a rat dosed orally with 1. The spectrum was obtained by CID of the MH+ ion at m/z 626, and the proposed origins of key fragment ions are as indicated.

View larger version (10K): [in this window] [in a new window]   Fig. 5.   LC/MS/MS product ion spectrum of M2 in the bile from a rat dosed orally with 1. The spectrum was obtained by CID of the MH+ ion at m/z 641, and the proposed origins of key fragment ions are as indicated.

View larger version (13K): [in this window] [in a new window]   Fig. 6.   LC/MS/MS product ion spectra of M3 (A) from the bile from a rat dosed orally with 1 and (B) from an incubation of M5 with rat liver mitochondria. The spectra were obtained by CID of the MH+ ion at m/z 627, and the proposed origins of key fragment ions are as indicated.

View larger version (10K): [in this window] [in a new window]   Fig. 7.   LC/MS/MS product ion spectrum of M4 in the bile from a rat dosed orally with 1. The spectrum was obtained by CID of the MH+ ion at m/z 504 (absence in the spectrum), and the proposed origins of key fragment ions are as indicated.

Formation of the Taurine and Isethionic Acid-Conjugated Metabolites. The taurine and isethionic acid-conjugated metabolites, M1 and M3, were formed in incubations of M5 with rat liver subcellular fractions (microsomes and mitochondria) in the presence of ATP, CoA, and magnesium chloride, plus the corresponding conjugating component (taurine or isethionic acid) (Fig. 6B). No conjugated metabolite was observed in control incubations containing boiled subcellular fractions or lacking ATP and CoA (data not shown). These data suggest that the conjugation reactions were mediated by the acyl-CoA thioester of M5 followed by acyl transfer to taurine or isethionic acid.

View larger version (17K): [in this window] [in a new window]   Fig. 8.   LC/MS/MS MRM detection of M1 and M3 in the bile from mice treated with M5. A, M1 was detected; B, both M1 and M3 were detected when the mouse was coadministered isethionic acid.

Oxidative Metabolism of 1 in Incubations with Liver Microsomes. The HPLC-UV detection of oxidative metabolites formed in vitro was based on comparing HPLC retention times with those of synthetic reference compounds. Thus, M2, M4, and M5 were identified in incubations of 1 with rat, dog, monkey, and human liver microsomes, and qualitatively, the metabolite profile appeared to be similar among species (Fig. 9). An interesting note is that M5 was not found in incubations of M4 with liver microsomes from the four species (~100% recovery of M4 following an incubation at 37°C for 30 min) whereas an aldehyde, derived from the deamination of M4, was observed when M4 was incubated with monoamine oxidase (data not shown).

View larger version (25K): [in this window] [in a new window]   Fig. 9.   HPLC-UV-vis chromatogram of samples from incubations of 1 with liver microsomes from (A) monkeys, (B) rats, (C) humans, and (D) dogs. Identification of M2, M4, and M5 was based on comparison of HPLC retention times of the metabolites with those of synthetic reference compounds.

Transport of 1 across Monolayers of Cells Expressing Pgp. A bidirectional transport assay was validated by examining the transport of [³H]vinblastine, an established Pgp substrate. In this case, transport of vinblastine from either side of the monolayer was linear over a period of 2 h, and the ratio for transport from B-to-A versus that from A-to-B was estimated to be ~7 (Fig. 10A).

View larger version (16K): [in this window] [in a new window]   Fig. 10.   Transport of (A) [3H]vinblastine and (B) 3H-labeled 1 across monolayers of cells expressing mouse mdr1a. A wide separation between curves representing A-to-B versus B-to-A transport of 1 also was observed with human MDR1 (data not shown).

	Discussion

Top Abstract Introduction Results Discussion References

The systemic clearance of 1 in experimental animal species is moderate and can be accounted for mainly by hepatic extraction (Stearns et al., 2002). When rats were treated with ³H-labeled 1, most of the radiolabeled components were found in the bile with biliary radiochromatographic profiles being similar following either oral or intravenous administration. Studies of the biotransformation of 1 in rats therefore focused mainly on the identification of biliary metabolites.

In the bile from rats treated with ³H-labeled 1, the taurine and isethionic acid conjugates of M5, namely M1 and M3, were identified. These metabolites also were observed in rats treated with M5 and were generated in vitro in incubations of M5 with rat liver subcellular fractions in the presence of ATP and CoA. These data suggest that M5 is the precursor of M1 and M3, and the formation of M1 and M3 from M5 follows similar mechanisms of amino acid conjugation reactions [i.e., M5 probably is converted to the corresponding acyl-CoA thioester intermediate followed by acyl transfer to taurine or isethionic acid (Fig. 11; Hutt and Caldwell, 1990)].

View larger version (15K): [in this window] [in a new window]   Fig. 11.   Proposed pathways for the metabolism of 1 in rats. CYP, cytochrome P450.

The structural difference between the two conjugating components is that the amino group of taurine is replaced by a hydroxyl functionality in isethionic acid. Although hydroxyl groups theoretically can function as nucleophiles, very few examples have been reported on conjugation reactions of isethionic acid with carboxylic acid derivatives to form an ester bond (Lombardini, 1976). Earlier data suggested that isethionic acid was derived from taurine in mammalian tissues, whereas further investigations indicated that the hydroxysulfonic acid was associated with the deamination of taurine by microorganisms (Welty et al., 1962; Fellman et al., 1980). Isethionic acid also has been found in dietary constituents (Fellman et al., 1980). In the case of M5, the isethionic acid conjugate was observed in rats but not in mice. This species difference was masked when the mouse was supplemented with isethionic acid, suggesting that the local availability of isethionic acid is the limiting factor in the formation of isethionic acid conjugates.

Although the taurine conjugate of M5 was a major metabolite, the acyl glucuronide of M5 was not observed in rats treated with ³H-labeled 1. In a generalized scheme, amino acid conjugation and acyl glucuronidation can be described as two competing reactions, with the former proceeding at lower K_m and the latter possessing higher capacity (Mulder, 1990). Accordingly, amino acid conjugation would be expected to be more important at low doses whereas acyl glucuronidation may become a major process at high doses (Mulder, 1990). In rats treated directly with M5 (10 mg/kg, intraperitoneal injection), both the taurine conjugate and the acyl glucuronide were identified in bile. It appeared, therefore, that the absence of the glucuronic acid conjugate in rats administered 1 (10 mg/kg, oral dosing with ~30% absorption) may have been the result of limited M5 available for glucuronidation.

With respect to M5, this metabolite may result from either the N-dealkylation of 1 by loss of 2-hydroxy-2-pyridinylethylamine or from further metabolism of M4 (Fig. 11). In incubations with liver microsomes from rats, dogs, monkeys, and humans, M5 was detected only when 1 was used as the substrate. Although the precise mechanism is not known, the formation of M5 in these incubations was observed to be NADPH-dependent, implicating a role for P450 enzymes in the conversion of 1 to M5. One such enzyme may be human hepatic CYP3A4, because M5 was identified in incubations of 1 with the recombinant P450 isoform. On the other hand, M5 might also derive from the deamination of M4. Although M4 appeared not to be metabolized in a system containing liver microsomes plus NADPH, an aldehyde product was detected in incubations of M4 with monoamine oxidase. It may be speculated that further oxidation of the aldehyde results in the formation of M5 (Fig. 11).

The formation of M2 and M4 in incubations of 1 with liver microsomes from rats, dogs, monkeys, and humans was observed to be NADPH-dependent, suggesting the involvement of P450 enzymes. Metabolite profiles were similar across species. In human liver microsomal incubations, apparent K_m and V_max values estimated for the formation of M2 were 29.4 µM and 0.15 µM/min/mg of protein, and for M4, 9.6 µM and 0.11 µM/min/mg of protein. These reactions were inhibited by troleandomycin and by a monoclonal antibody against CYP3A4 but were not influenced by sulfaphenazole and quinidine. The two metabolites also were identified in incubations of 1 with recombinant CYP3A4. Taken together, these data indicate that human hepatic CYP3A4 catalyzes the metabolism of 1 to M2 and M4. On the other hand, the incomplete inhibition observed in the case of M2 implicates a role for P450 isoforms other than CYP3A4, or of nonP450 enzyme(s), in the formation of this metabolite.

An interesting observation was that approximately 10% of the dose was recovered in the feces from bile duct-cannulated rats administered ³H-labeled 1 by intravenous injection. Since all the bile flow was transferred to collection vials, the parent compound found in the feces was most likely due to intestinal excretion. Subsequent investigation in a bidirectional transport in vitro assay indicated that transport of 1 across a monolayer of cells expressing mouse mdr1a or human MDR1 was similar to that of vinblastine, suggesting that this ₃-receptor agonist was a substrate of the Pgp transporter. It may be speculated, therefore, that 1 is subject to excretion by the Pgp transporter localized in rat intestine. Similar data were reported for the disposition of digoxin in mice, wherein the drug was found to be excreted via the intestinal mucosa by Pgp proteins (Mayer et al., 1996).

In summary, five major metabolites of 1 were identified in rats, including a pyridine N-oxide derivative M2, a primary amine M4, a carboxylic acid M5 and its corresponding taurine conjugate M1, and isethionic acid conjugate M3. The formation of M2, M4, and M5 most likely is the result of P450-catalyzed metabolism. Further biotransformation of M5 leads to the formation of M1 and M3, which may follow a similar mechanism as that for amino acid conjugation. A small amount of 1 recovered in the feces of bile duct-cannulated rats following intravenous dosing could be due to intestinal Pgp-mediated excretion of the parent drug.