Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

Rofecoxib (3-phenyl-4-[4-(methylsulfonyl)phenyl]-2-(5H)-furanone) is a potent and highly selective inhibitor of cyclooxygenase (COX¹)-2, an inducible isoform of cyclooxygenase that plays a key role in inflammatory processes. Rofecoxib (VIOXX; Merck & Co., Inc., Whitehouse Station, NJ) was approved recently for the treatment of osteoarthritis and pain, based on the hypothesis that selective inhibition of COX-2 should result in decreased inflammation without the adverse gastrointestinal (GI) effects associated with inhibition of COX-1 (reviewed in Jouzeau et al., 1997). Recently, this hypothesis was supported by the VIGOR study in which rheumatoid arthritis patients treated with rofecoxib displayed significantly fewer clinically important gastrointestinal events than those patients treated with naproxen, a nonselective COX inhibitor (Bombardier et al., 2000).

The absorption, distribution, metabolism, and excretion of rofecoxib in rats and dogs was reported recently by Halpin and coworkers (2000). In the rat, an unusual feature of the plasma concentration versus time profile for rofecoxib following oral administration was the presence of a second distinct C_max, suggestive of enterohepatic recycling. Interestingly, however, no rofecoxib was excreted in bile, but the inactive metabolites 5-hydroxyrofecoxib and its glucuronide conjugate were present as the major drug-related species. The primary goal of the present investigation, therefore, was to determine whether these biliary metabolites could serve as a source of rofecoxib in the GI tract from which the drug could be reabsorbed and, if so, to elucidate the underlying biochemical mechanism.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Chemicals. [4-¹⁴C]Rofecoxib (specific activity ranging from 24.3 to 26.5 µCi/mg; Fig. 1) was prepared as described previously (Halpin et al., 2000), whereas unlabeled drug was synthesized by the Department of Process Research (Merck Research Laboratories, Rahway, NJ). Oxygen-18-labeled compounds were synthesized as described below. Unlabeled 5-hydroxyrofecoxib and the 4'-methylphenyl analog of rofecoxib (used as an analytical internal standard) were obtained from Merck Frosst Canada (Kirkland, QC, Canada). Due to the light-sensitive nature of rofecoxib and rofecoxib-related compounds, the drug and all biological samples were handled under yellow light.

View larger version (23K): [in this window] [in a new window]   Fig. 1.   Chemical structures of rofecoxib, 5-hydroxyrofecoxib, 18O-labeled variants thereof, and 4'-methylrofecoxib (analytical internal standard). Asterisks indicate the location of 14C in the radiolabeled analogs.

Chemical Synthesis. The synthesis of all oxygen-18-labeled variants of rofecoxib was accomplished through variations of a common approach using phenylacetic acid derivative 1 and functionalized 4-(methylsulfonyl)acetophenone 2 or 3 (Fig. 2). Coupling and cyclization of these two fragments were carried out through adaptation of previously described methods for the synthesis of 2(5H)-furanones (method A, Dikshit et al., 1990; method B, Leblanc et al., 1999). H₂¹⁸O and C¹⁸O₂ were obtained from Isotec, Inc. (Miamisburg, OH), whereas ¹⁸O₂ was obtained from Cambridge Isotope Labs, Inc. (Andover, MA). 2-Bromo-4-(methylsulfonyl)acetophenone (2) was obtained from the Department of Process Research (Merck Research Laboratories). Where indicated, the progress of reactions and purity of products were determined by HPLC with UV detection ( = 230 nm). A Zorbax Rx-C₈ column (4.6 × 250 mm, 5 µm) was used, with a mobile phase of water/CH₃CN (70:30, v/v) that was delivered at a flow rate of 1 ml/min with a linear gradient to 100% CH₃CN over 25 min. Isotopic enrichment of the labeled compounds was determined by gas chromatography/MS analysis using an HP model 5890A gas chromatograph coupled to an HP 5970B mass-selective detector (Hewlett Packard, Palo Alto, CA).

View larger version (12K): [in this window] [in a new window]   Fig. 2.   Scheme for the synthesis of unlabeled and oxygen-18-labeled compounds. For details, see Materials and Methods.

[¹⁸O₂]Phenylacetic acid (1b). A two-necked, round-bottomed flask (25 ml), which was fitted with a septum inlet and magnetic stir bar, was attached to a gas transfer manifold. After the entire apparatus was flame dried and cooled under a nitrogen atmosphere, the flask was charged with dry ether (4 ml), followed by benzyl magnesium chloride (4.1 ml of a 1.0 M solution in ether). The contents were frozen with liquid nitrogen and C¹⁸O₂ (break-seal flask; 100 ml, ~4.1 mmol, 96.3 atom % excess ¹⁸O) was added via standard vacuum transfer techniques. The liquid nitrogen bath was removed, and when the contents reached room temperature, the mixture was quenched with water (20 ml). Additional ether (5 ml) was added, followed by several drops of 50 NaOH_aq (w/w). Following separation of the layers, the ether phase was removed, and the aqueous phase was adjusted to pH 2 with concentrated H₂SO₄. The aqueous layer was extracted with ether (3 × 50 ml), and the combined ether extracts were washed with brine, dried over CaCl₂, and filtered through a medium-porosity, fritted-glass funnel. Removal of the solvent on a rotary evaporator afforded 1b (440 mg, 76% yield) as a white solid, which was used without further purification.

[2-Hydroxy-¹⁸O]-4-(methylsulfonyl)acetophenone (3b). Potasium cyanide (300 mg, 4.6 mmol) and H₂¹⁸O (0.6 ml, 99.1 atom % excess ¹⁸O) were placed in a sealed stainless steel bomb, and the contents were heated to 175°C with a silicone oil bath. After 3 h, the heating bath was removed, the vessel was cooled to ~5°C with an ice-water bath, and the contents were removed via syringe (0.72 ml obtained).

2-Hydroxy-4-(methylsulfonyl)acetophenone (3a). Unlabeled hydroxy-acetophenone 3a was prepared from 2, as described above using unlabeled potasium formate.

[1,2-¹⁸O₂]Rofecoxib (5b)method A. [¹⁸O₂]Phenylacetic acid (1b) (400 mg, 2.85 mmol) was taken up into DMF (15 ml, nitrogen purged) and treated with NaOH_aq (50 wt%, 74 µl) with vigorous stirring and warming to 50°C. After 0.5 h, bromosulfone (2) (790 mg, 2.85 mmol) was added, followed 25 min later by diisopropylamine (1.2 ml). After an additional 2.5 h, the reaction was deemed complete by HPLC analysis. After the solution was allowed to cool to room temperature, HCl (2 N, 5 ml) was added slowly, with cooling by an ice-water bath. Water (20 ml) then was added dropwise, resulting in precipitation of the product. The solid was collected on a medium-porosity, fritted-glass funnel and crystallized twice from acetone/isopropyl alcohol (1:2), affording the title compound (228 mg, 25% yield) with a purity of 99.8% by HPLC (>98 atom % excess ¹⁸O₂).

[1-¹⁸O]Rofecoxib (5c)method B. [¹⁸O]Hydroxyketone (3b) (330 mg, 1.53 mmol), phenylacetic acid (228 mg, 1.53 mmol), 1-(3-dimethylaminopropyl)-3-ethylcarbodiimide hydrochloride (292 mg), and 4-(dimethylamino)pyridine (93 mg) were taken up into CH₂Cl₂ (8 ml) and stirred at room temperature. After 4 h, an additional 0.25 Eq of each reagent was added, and the mixture was stirred overnight. The mixture then was poured into HCl (1 N, 40 ml) and extracted with ethyl acetate (2 × 50 ml). The combined organic extracts were washed with HCl (1 N, 30 ml), saturated NaHCO₃ (30 ml), and brine (25 ml), dried over MgSO₄, vacuum filtered through a medium-porosity, fritted-glass funnel, and the solvent was removed under reduced pressure. The resulting residue was applied to a silica gel column eluted with ethyl acetate/hexane (1:1) affording 4 (144 mg, 28%) as a white solid.

[2-¹⁸O]Rofecoxib (5d)method B. [¹⁸O₂]Phenylacetic acid (1b) (300 mg, 2.1 mmol), hydroxyacetophenone (3a) (458 mg, 2.1 mmol), and N,N-dimethylaminopyridine (131 mg, 1.05 mmol) were placed in a round-bottomed flask (25 ml) and flushed with nitrogen. The contents then were taken up into CH₂Cl₂ (10 ml), and 1-(3-dimethylaminopropyl)-3-ethylcarbodiimide hydrochloride (492 mg, 2.57 mmol) was added. The resulting slurry was stirred for 36 h at room temperature, at which point HPLC analysis showed ~53% product yield. The reaction was terminated, and the product was isolated as described above for 5c (part 1) to afford [¹⁸O]4 (321 mg, 46% yield) as a white solid.

[5-¹⁸O]5-Hydroxyrofecoxib (6). Ecosorb-activated carbon was preconditioned by washing with ethanol, followed by ethyl acetate, and vacuum drying overnight. This material (2.4 g) was placed in a three-necked, round-bottomed flask (250 ml) and purged with nitrogen overnight. In a separate flask (100 ml) was placed rofecoxib (1 g, 3.2 mmol), which also was purged with nitrogen overnight. At the end of this period, the flask containing Ecosorb was placed on a vacuum transfer manifold along with a break-seal flask containing [¹⁸O₂]oxygen (0.36 g, 11.2 mmol, 95-98 atom % excess ¹⁸O). The flask containing rofecoxib then was charged with ethyl acetate (65 ml, nitrogen purged), and after complete dissolution, the contents were transferred to the Ecosorb-containing flask via cannula. The combined contents were frozen with a liquid nitrogen bath, the system was evacuated, and the ¹⁸O₂ admitted. The resultant mixture was stirred at room temperature for 48 h, at which point it was filtered through celite and the solids were washed with CH₃CN (200 ml). Removal of the solvent with a rotary evaporator afforded a syrup that was ~30% 6 by HPLC analysis. Purification was achieved by silica gel chromatography (CH₂Cl₂/ethyl acetate, 3:1), followed by crystallization from ethyl acetate/hexane to give 6 (130 mg, 12% yield, 99.4% pure by HPLC, 85 atom % excess ¹⁸O).

Animal Studies. Male Sprague-Dawley rats (~250 g) were purchased from Taconic Farms (Germantown, NY) either with a cannula implanted in the right jugular vein only (for blood sampling) or with cannulae implanted in both the jugular vein and common bile duct. Animals were fasted overnight with free access to water, fed 6 to 8 h after dosing, and housed in individual metabolism cages during sampling periods. Unless otherwise indicated, intravenous doses were administered via the tail vein, whereas oral doses were administered via gavage. For the i.v. studies, the dosing material was dissolved in a small amount of DMSO to give a final concentration of 4 mg/ml. For the oral studies, the dosing material was dissolved in a small amount of DMSO, followed by addition of PEG-400 to give a suspension with a final concentration of 1 mg/ml. Blood samples (0.3-0.4 ml) were obtained at predose and at 5 min (i.v. only), 0.25, 0.5, 1, 2, 4, 6, 8, 10, 24, and 30 h (¹⁸O studies only) postdose. To stabilize 5-hydroxyrofecoxib in plasma, blood was obtained using heparinized syringes and was transferred to tubes containing EDTA (5-6 mg). Plasma was harvested by centrifugation and transferred to tubes containing phosphoric acid (10 µl, 1 M). All samples were stored at 20°C until analyzed.

Enterohepatic recirculation studies with [4-¹⁴C]rofecoxib and [4-¹⁴C]5-hydroxyrofecoxib. [4-¹⁴C]Rofecoxib was used without dilution of the radiolabel for the i.v. studies (2 mg/kg) only. The specific activities of i.v. dosing solutions administered to intact (n = 4) and bile duct-cannulated rats (n = 4) each were ~57,700 dpm/µg, whereas for oral dosing (5 mg/kg), the corresponding specific activities were ~21,000 dpm/µg and ~33,000 dpm/µg, respectively.

Gastrointestinal tract distribution studies. [4-¹⁴C]Rofecoxib and [4-¹⁴C]5-hydroxyrofecoxib were admixed with the corresponding unlabeled compounds to give specific activities of ~17,500 and ~19,000 dpm/µg, respectively, and animals were dosed orally at 5 mg/kg. At 1, 4, and 6 h postdose, animals (n = 2/time point/compound) were euthanized, and five segments of the GI tract (duodenum, jejunum, ileum, cecum, and large intestine) were excised. The lumen of each segment was flushed with deionized water (5 ml), and the effluents and tissues were placed in individual tared tubes and stored at 20°C.

Studies with [1,2-¹⁸O₂]rofecoxib and [5-¹⁸O]5-hydroxyrofecoxib. For i.v. dosing, [1,2-¹⁸O₂]rofecoxib (n = 1) and [5-¹⁸O]5-hydroxyrofecoxib (n = 3) were administered via jugular vein cannula to animals at 4 mg/kg. For oral dosing, each compound was dissolved in PEG-400 and administered at 10 mg/kg (n = 2, [1,2-¹⁸O₂]rofecoxib study; n = 3, [5-¹⁸O]5-hydroxyrofecoxib study).

Analytical Procedures.

HPLC assay for rofecoxib and 5-hydroxyrofecoxib in plasma Plasma concentrations of rofecoxib and 5-hydroxyrofecoxib were determined by an HPLC/fluorescence method involving postcolumn photochemical generation of a stilbene-phenanthrene derivative (Woolf et al., 1999), as described by Halpin et al. (2000). For the ¹⁸O studies, plasma samples were prepared as described, except that no water was added to avoid exchange of the ¹⁸O label in 5-hydroxyrofecoxib and the samples were extracted with methyl-t-butyl ether (8 ml). Before analysis, the dried residues were reconstituted in a mixture of 0.1% aqueous TFA/NH₄OH, pH 3.0, and CH₃CN (7:3, v/v).

Radioactivity measurements and radiochromatography from GI tract distribution studies. Water was added to the excised intestinal tissues and contents to provide a slurry. Homogenized aliquots (1 g) were dried overnight, and the samples were subjected to combustion followed by liquid scintillation counting.

LC-MS/MS assay for isotopic variants of rofecoxib and 5-hydroxyrofecoxib. The isotopic distribution of labeled and unlabeled species of rofecoxib and 5-hydroxyrofecoxib in plasma of rats dosed with ¹⁸O-labeled tracers was determined by LC-MS/MS analysis with selected reaction monitoring, using the transitions shown in Fig. 3.

View larger version (14K): [in this window] [in a new window]   Fig. 3.   Mass spectrometric transitions (precursor and product ions) used in LC-MS/MS analyses of isotopic variants of rofecoxib and 5-hydroxyrofecoxib. In each case, the precursor ion corresponded to the [M-H] species.

NMR analyses. ¹H and ¹³C NMR spectra were obtained in d₆-DMSO with a Varian Unity-500 spectrometer (Varian, Palo Alto, CA). Chemical shifts are reported in ppm () downfield from tetramethylsilane, using residual solvent signal as a reference. Coupling constants (J) are reported in Hertz. The O-18 isotope effects on C-13 chemical shifts were accurately measured by mixing equal amounts of labeled and unlabeled compounds in d₆-DMSO. The affected signals were doubled.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Characterization of Oxygen-18-Labeled Compounds. A number of ¹⁸O-labeled compounds were prepared in support of the mechanistic studies conducted in this work and were used either as tracers for in vivo experiments or as reference standards for those metabolites that had undergone partial loss of label. In order for the location of the ¹⁸O atom in singly labeled derivatives of rofecoxib to be assigned accurately, all labeled compounds were subjected to NMR and MS/MS analysis.

Effect of Bile Flow on the Enterohepatic Recycling of Rofecoxib and 5-Hydroxyrofecoxib.

Studies with [4-¹⁴C]rofecoxib in intact and bile duct-cannulated rats When [4-¹⁴C]rofecoxib was administered intravenously or orally to intact animals, rofecoxib and 5-hydroxyrofecoxib were detected in plasma up to 24 h (Fig. 4A). However, in bile duct-cannulated animals, parent drug and metabolite were detected only up to 10 or 6 h, respectively (Fig. 4B). The mean plasma concentrations of rofecoxib in intact rats declined after i.v. dosing to trough levels at 6 h postdose and then increased to a secondary C_max at 10 h postdose (Fig. 4A). However, in bile duct-cannulated rats, mean plasma concentrations of rofecoxib fell steadily to the limit of quantitation (Fig. 4B). After oral administration to intact rats, rofecoxib plasma concentrations reached C_max at 1 h and decreased to a plateau over the 6- to 10-h interval, whereas in bile duct-cannulated animals, rofecoxib levels decreased steadily after reaching C_max at 2 h (data not shown).

View larger version (18K): [in this window] [in a new window]   Fig. 4.   Semilog plots depicting mean plasma concentrations of rofecoxib and 5-hydroxyrofecoxib in intact (A; n = 4) and bile-cannulated (B; n = 4) rats following i.v. administration of [4-14C]rofecoxib (2 mg/kg).

Studies with [4-¹⁴C]5-hydroxyrofecoxib in intact and bile duct-cannulated rats. Rofecoxib appeared rapidly in the plasma of intact and bile duct-cannulated rats following i.v. administration of 5-hydroxyrofecoxib (2 mg/kg; Fig. 5), confirming that the conversion of parent compound to its oxidative metabolite is a reversible process in the rat, as proposed earlier (Halpin et al., 2000). Following intravenous administration of 5-hydroxyrofecoxib to intact rats (Fig. 5A), mean rofecoxib levels declined sharply from a C_max at initial sampling times to below quantifiable levels. However, rofecoxib reappeared in plasma at 4 h, with mean concentrations increasing to a second C_max at 10 h. Similarly, in bile duct-cannulated rats, rofecoxib appeared rapidly in the initial phase with concentrations measurable to 4 h (Fig. 5B). However, the second peak in the plasma concentration versus time curve for rofecoxib was not observed in these animals. After reaching C_max values at the earliest time point after i.v. dosing, 5-hydroxyrofecoxib levels declined exponentially in both intact and bile duct-cannulated rats and were measurable in plasma to 24 h. Following oral administration, similar 5-hydroxyrofecoxib and rofecoxib profiles were observed, except that in bile duct-cannulated animals, 5-hydroxyrofecoxib and rofecoxib were measurable only to 10 and 6 h, respectively (data not shown). In this study, the mean partial AUC values for 5-hydroxyrofecoxib in plasma of intact rats (0-24 h) were 7-fold (i.v.) to 10-fold (p.o.) higher than the corresponding values in bile duct-cannulated animals (0-2 h, i.v.; 0-6 h, p.o.; Table 2).

View larger version (19K): [in this window] [in a new window]   Fig. 5.   Semilog plots depicting mean plasma concentrations of 5-hydroxyrofecoxib and rofecoxib in intact (A; n = 4) and bile-cannulated (B; n = 3) rats following i.v. administration of [4-14C]5-hydroxyrofecoxib (2 mg/kg).

Distribution and Metabolism of Rofecoxib and 5-Hydroxyrofecoxib in the GI Tract of Rats. The mean total radioactivity remaining in the GI tract of rats at 1, 4, and 6 h after receiving an oral dose of [4-¹⁴C]rofecoxib (5 mg/kg) represented 58, 72, and 60% of the dose, respectively. For rats that received [4-¹⁴C]5-hydroxyrofecoxib at the same dose, the corresponding values were 63, 53, and 65% of the dose, respectively. For both [4-¹⁴C]rofecoxib- and [4-¹⁴C]5-hydroxyrofecoxib-dosed animals, more than 95% of the radioactivity recovered in the GI tract at 1 h was confined to the jejunum and ileum. At 6 h, the radioactivity was distributed evenly throughout the jejunum, ileum, and cecum, with 4 to 5% of the dose present in the large intestine. At all time points, 1% or less of the dose was present in the duodenum.

View larger version (22K): [in this window] [in a new window]   Fig. 6.   Distribution of radioactivity in rat intestinal segments at 1 h (A), 4 h (B), and 6 h (C) after oral administration of [4-14C]rofecoxib (5 mg/kg).

View larger version (20K): [in this window] [in a new window]   Fig. 7.   Distribution of radioactivity in rat intestinal segments at 1 h (A), 4 h (B), and 6 h (C) after oral administration of [4-14C]5-hydroxyrofecoxib (5 mg/kg).

Studies with [1,2-¹⁸O₂]Rofecoxib and [5-¹⁸O]5-Hydroxyrofecoxib. Once it had been shown that 5-hydroxyrofecoxib undergoes metabolism in rats to rofecoxib, the mechanism by which this process occurs was investigated by dosing rats with [1,2-¹⁸O₂]rofecoxib or [5-¹⁸O]5-hydroxyrofecoxib and following the fate of the labeled species in plasma by LC-MS/MS techniques.

Studies with [1,2-¹⁸O₂]rofecoxib. When rats were dosed i.v. or p.o. with [1,2-¹⁸O₂]rofecoxib, both [1-¹⁸O]rofecoxib and [2-¹⁸O]rofecoxib appeared rapidly in plasma (Table 3). Interestingly, the ratio of singly ¹⁸O-labeled rofecoxib containing the ¹⁸O in the furanone ring ([1-¹⁸O]rofecoxib) to the isomer containing the ¹⁸O in the carbonyl group ([2-¹⁸O]rofecoxib) was close to, or greater than, unity over the first 2 h postdose. However, at later time points (4-30 h postdose), this ratio decreased to ~0.2 (Table 3). Thus, at early time points after dosing, loss of the ¹⁸O label took place preferentially from the carbonyl moiety of rofecoxib, whereas at later time points, loss of the ¹⁸O label occurred primarily from the ring oxygen of the furanone moiety. Loss of both atoms of ¹⁸O to generate unlabeled rofecoxib was also observed, although the unlabeled drug was the major circulating species in plasma only at late time points (>24 h).

Studies with [5-¹⁸O]5-hydroxyrofecoxib. To investigate further the reversible metabolism of rofecoxib, rats were dosed intravenously (4 mg/kg) or orally (10 mg/kg) with [5-¹⁸O]5-hydroxyrofecoxib. After both i.v. and oral dosing, extensive loss of ¹⁸O-label was observed from the 5-hydroxyrofecoxib circulating in plasma (data not shown).

View larger version (25K): [in this window] [in a new window]   Fig. 8.   Semilog plots depicting mean plasma concentrations of rofecoxib present as [1-18O]rofecoxib, [2-18O]rofecoxib, or unlabeled rofecoxib following i.v. (4 mg/kg) (A) or oral (10 mg/kg) (B) administration of [5-18O]5-hydroxyrofecoxib. Values for the 1- and 2-h time periods for the i.v. study and the 0.25-, 0.5-, and 1-h time periods for the oral study were below the limit of quantitation.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

The results of the present study provide support for the hypothesis that enterohepatic recycling plays an important role in the pharmacokinetics of rofecoxib in the rat inasmuch as the prominent secondary peak in the rofecoxib plasma concentration versus time profile following i.v. dosing is abolished in bile duct-cannulated animals. Previous findings that the bile of rats administered rofecoxib contained relatively high concentrations of 5-hydroxyrofecoxib, a cytochrome P450-mediated metabolite of the drug, together with its glucuronide conjugate (Halpin et al., 2000), were confirmed in this study and suggested that these metabolites may revert to rofecoxib in the intestine through a combination of deconjugation and reduction processes. Further evidence for such a mechanism was obtained from experiments in which 5-hydroxyrofecoxib, rather than the parent drug, was administered intravenously to rats; in intact animals, rofecoxib appeared as a prominent component in the plasma at later time points, whereas in bile duct-cannulated rats, this phenomenon was not observed. Moreover, the pattern of metabolites present in descending segments of the intestinal tract after oral administration of either rofecoxib or 5-hydroxyrofecoxib to intact rats was consistent with a sequence in which 5-hydroxyrofecoxib glucuronide, released into the duodenum from bile, undergoes successive deconjugation and reduction, likely under the influence of the gut microflora, to regenerate rofecoxib in the large intestine. Reabsorption of the parent drug into the systemic circulation then takes place to complete the "enterohepatic cycle".

From a mechanistic standpoint, the most probable sequence of reactions that accounts for this novel form of enterohepatic cycling and is consistent with the experimental observations is depicted in Fig. 9. In an attempt to establish the validity of this hypothetical mechanism, consideration was given to the fate of the oxygen atom in the 5-hydroxy group of 5-hydroxyrofecoxib (denoted by an asterisk in Fig. 9). According to this scheme, this oxygen atom should be retained in rofecoxib, where it would be incorporated as the "ring" oxygen (position 1) of the furanone moiety.

View larger version (20K): [in this window] [in a new window]   Fig. 9.   Proposed mechanism for the enterohepatic cycling of rofecoxib via 5-hydroxyrofecoxib and 5-hydroxyrofecoxib glucuronide. The asterisk indicates the predicted fate of the oxygen atom introduced by hydroxylation at the C-5 position. [E, Red] denotes enzymatic reduction.

When [5-¹⁸O]5-hydroxyrofecoxib was synthesized and administered as a tracer to intact rats by either intravenous or oral routes, [1-¹⁸O]rofecoxib indeed was formed in the animals and circulated as a prominent metabolite in plasma. However, unlabeled rofecoxib also was produced from [5-¹⁸O]5-hydroxyrofecoxib, together with smaller amounts of [2-¹⁸O]rofecoxib. Moreover, the plasma of these rats was found to contain unlabeled 5-hydroxyrofecoxib, indicating that the administered tracer underwent loss of the ¹⁸O label concomitant with metabolic conversion to rofecoxib.

To account for the results of the ¹⁸O studies with labeled 5-hydroxyrofecoxib, it was necessary to invoke the processes depicted in Fig. 10 in which the tautomeric ring opening of the cyclic hemiacetal and reversible hydration of the resulting aldehyde carbonyl leads to "migration" of the exocyclic 5-¹⁸OH label to positions 1 or 2 or to complete loss of the ¹⁸O atom. Although it was not possible to distinguish the [5-¹⁸O] from the [1-¹⁸O] variants of 5-hydroxyrofecoxib by mass spectrometry (both yielding [M-H] and [M-H-CO₂] ions that contained, respectively, both or neither of the oxygen atoms in question), the LC-MS/MS analysis of plasma from rats administered [5-¹⁸O]5-hydroxyrofecoxib clearly revealed the presence of isotopic variants of this compound with one or no ¹⁸O label. Applying the considerations shown in Fig. 10 to the enterohepatic cycling pathway proposed in Fig. 9, it becomes straightforward to rationalize the appearance in rat plasma of molecules of rofecoxib containing one atom of ¹⁸O in the carbonyl group (2-¹⁸O), as well as unlabeled species.

View larger version (12K): [in this window] [in a new window]   Fig. 10.   Proposed mechanism for the metabolic interconversion of 18O-labeled variants of 5-hydroxyrofecoxib.

In a final series of tracer experiments, rats were dosed orally with a variant of rofecoxib, itself labeled in the furanone ring with two atoms of ¹⁸O, and the fate of the individual ¹⁸O atoms was followed by means of LC-MS/MS analysis. The results of this study further demonstrated the complexity of the metabolic fate of rofecoxib in vivo in that both mono-¹⁸O-labeled variants of rofecoxib were formed together with unlabeled drug. These observations can best be accommodated by the pathways outlined in Figs. 11 and 12. In the former, metabolism of rofecoxib proceeds by way of 5-hydroxyrofecoxib, which undergoes partial ¹⁸O exchange (as discussed above) before reduction back to rofecoxib, whereas in the latter, rofecoxib undergoes hydrolytic ring opening directly, followed by exchange of ¹⁸O with water of the medium and recyclization. It seems likely that the results of the experiment with [1,2-¹⁸O₂]rofecoxib may be accommodated most satisfactorily by the simultaneous operation of both of these mechanisms.

View larger version (17K): [in this window] [in a new window]   Fig. 11.   Proposed redox mechanism for the metabolic conversion of [1,2-18O2]rofecoxib to [2-18O]rofecoxib and unlabeled rofecoxib. [E, Red] denotes enzymatic reduction.

View larger version (14K): [in this window] [in a new window]   Fig. 12.   Proposed hydrolytic mechanism for the metabolic conversion of [1,2-18O2]rofecoxib to [1-18O]rofecoxib.

There are numerous examples of reversible metabolism of drugs in vivo, including N-oxidation of clozapine (Jann et al., 1994), N-acetylation of dapsone (Gelber et al., 1971; Gordon et al., 1975), ketone reduction of haloperidol (Jann et al., 1990), sulfoxide reduction of sulindac (Duggan et al., 1977), oxidation of numerous 11-oxygenated glucocorticoids (Bush et al., 1968), disulfide formation of D-penicillamine (Bourke et al., 1984), and lactone ring opening of topotecan (Grochow et al., 1992), lovastatin (Duggan et al., 1989), and simvastatin (Vickers et al., 1990). Although reversibility of lactone ring opening has been observed, the results of the present investigation provide the first example to our knowledge of reversible metabolism of an -hydroxy lactone to the corresponding lactone. From a mechanistic point of view, the key step in the overall process is spontaneous ring opening of 5-hydroxyrofecoxib, which is analogous to the cytochrome P450-mediated cleavage of esters described by Guengerich et al. (1988) and exposes a C-5 aldehyde for subsequent reduction.

In conclusion, the use of specific ¹⁸O-labeled derivatives of rofecoxib and 5-hydroxyrofecoxib as metabolic tracers, coupled with the adoption of LC-MS/MS techniques to detect and quantify specific isotopic variants present in biological fluids, provided a unique insight into mechanistic aspects of the unusual reversible metabolism of rofecoxib to 5-hydroxyrofecoxib in the rat in vivo. Moreover, the present study revealed that the 2-furanone ring of rofecoxib is a metabolically active entity, undergoing reversible opening through a combination of hydrolytic and redox processes that would have been difficult to detect without the use of stable isotope techniques.