Microbial Metabolism of Rofecoxib , A Selective COX-2 Inhibitor

Microbial transformation studies of rofecoxib (1), a potent selective cyclooxygenase-2 inhibitor, using sixty microorganisms, mainly fungi and actinomycetes, have revealed that it was metabolized by three microorganisms. Using a standard two-stage screening technique, Cunninghamella echinulata var. echinulata ATCC 9244, Mucor griseocyanus ATCC 1207b, and Rhizopus oryzae ATCC 34121 showed two common more polar metabolites 2 and 3. Moreover, it was apparent that Cunninghamella echinulata var. echinulata was the most efficient microorganism to almost completely metabolize 1, and hence, was selected for preparative scale fermentation. These metabolites were characterized on the basis of their spectral data as 4 -hydroxyrofecoxib (2) and 3 -hydroxyrofecoxib (3).


INTRODUCTION
The enzyme prostaglandin cyclooxygenase, COX, catalyzes the rate-limiting cyclization step in the formation of prostanoids from arachidonic acid [1,2].Two isoforms of the COX enzyme have been characterized; COX-1 and COX-2 [3].The discovery of COX-2 has provided the rationale for the development of a new class of the non-steroidal anti-inflammatory drugs (NSAIDs), the selective COX-2 inhibitors, coxibs.These agents were developed on the premise that selective inhibition of COX-2 would result in decreased inflammation without the adverse gastrointestinal effects associated with inhibition of COX-1 [4][5][6].Rofecoxib and celecoxib were the first selective COX-2 inhibitors approved by the US Food and Drug Administration for the treatment of rheumatoid arthritis, osteoarthritis and for relief of acute pain associated with dental surgery and primary dysmenorrhea.Rofecoxib is a diarylfuranone derivative containing a phenylsulphone moiety that interacts with COX-2 side pocket [7].This interaction represents an important determinant for COX-2 selectivity.In fact, rofecoxib almost completely inhibits monocyte COX-2 activity without affecting COX-1 activity [8].
On September 30, 2004, Merck and Co. voluntary withdrew rofecoxib, VIOXX , from the market due to increased risk of cardiovascular toxicities associated with the drug [9].It is not clear whether these toxicities are due to rofecoxib high COX-2 selectivity.It was shown that COX-2 inhibitors decrease vascular prostacyclin PGI 2 production and may disrupt the homeostatic mechanisms that limit the effects of platelets activation.Therefore, the mechanism of this cardiovascular toxicity could lie in the inhibition of COX-2 itself, and thus be a class effect.On the other hand, it may be due to its unique chemical structure, its pharmacokinetics and/or the presence of toxic metabolites [9].In order to explore more this hypothesis, rofecoxib metabolites need to be prepared in quantities enough to be identified and evaluated for possible activity and/or toxicity.
In drug research and development, metabolism is of a pivotal importance due to the interconnectedness between pharmacokinetic and pharmacodynamic processes.The phenomenon of metabolism originates from the fact that mammals usually treat any exogenous drug as non-self and work hard to weaken the effect and/or the concentration of this "foreign" substance through enzymatic machinery.However, xenobiotic metabolic system sometimes generates products more reactive and biologically destructive than the original substrate.The metabolism process generates new chemical entities, metabolites, which may have distinct pharmacodynamics and toxicodynamic properties of their own [10].
Human metabolism of rofecoxib is primarily mediated through reduction by cytosolic enzymes.The principal metabolic products are the cis-and transdihydro derivatives of rofecoxib.An additional metabolite was recovered as the glucuronide of the hydroxy derivative, a product of oxidative metabolism.Cytochrome P450 plays a minor role in metabolism of rofecoxib.This was concluded since rofecoxib disposition was not affected by the inhibition of CYP 3A activity after the administration of ketoconazole.
However, induction of general hepatic metabolic activities by administration of the non-specific inducer rifampin, produces a 50% decrease in rofecoxib plasma concentrations (VIOXX insert).
The use of microbial systems as in vitro models for drug metabolism in humans was first formalized by Rosazza and Smith [11], and was defined as the use of microorganisms, bacteria, yeasts, and fungi, to facilitate the study of xenobiotic biotransformation in mammals, including man.
It is anticipated that the microbial metabolism of rofecoxib would produce significant quantities of metabolites that would be difficult to obtain from either animal systems or chemical synthesis.Moreover, this work will also provide some new analogs that may serve as prospective candidates for their potential pharmacological action and/or cardiac toxicity evaluation, or as starting compounds for the semisynthesis of other derivatives.The compounds IR spectra were recorded as neat solids using an FTIR-4100 Jasco spectrophotometer (Jasco, Easton, Maryland, USA).UV spectra were measured in methanol using a UV-160 IPC UV-visible dual-beam spectrophotometer.The C NMR spectra, the number of attached protons was determined by DEPT 135°.2D NMR data were obtained using the standard pulse sequence of the Bruker DPX-400 for COSY, HSQC, and HMBC.HRMS was carried out using a Bruker Bioapex FTMS with Electrospray Ionization Spectrometer (Bruker Daltonic GmbH, Bremen, Germany).

Cultures and Fermentation Screening Procedure
A total of sixty microorganisms, mainly fungi and actinomycetes, Table 1, have been screened for their ability to catalyze bioconversion of rofecoxib (1).Only three of them, Cunninghamella echinulata var.echinulata ATCC 9244, Mucor griseocyanus ATCC 1207b, and Rhizopus oryzae ATCC 34121, showed definite metabolism of rofecoxib.Stock Cultures were maintained on agar slants of media recommended by the ATCC and were stored at 4 °C.
All the preliminary screening and preparative-scale experiments were carried out as reported before [12], and according to a standard two-stage protocol [13].Substrate 1 was prepared as a 15% solution in N,Ndimethylformamide (DMF) and added to the 24-h-old stage II culture medium of the microorganism at a concentration 0.3 mg/mL of medium.Substrate controls were composed of sterile medium to which the substrate (4 mg/100μL DMF) was added and incubated without microorganisms.Culture controls consisted of fermentation blanks in which the microorganisms were grown under identical conditions but without the substrate addition.After two weeks of incubation, each control was harvested and analyzed by TLC.

Culture Sampling and Chromatographic Conditions
Sampling was carried out by withdrawing 5 mL of culture and extracting it with 5 mL of ethylacetate.The concentrated organic phase was analyzed by TLC for the presence of metabolites.TLC analyses were performed using 5% MeOH in CHCl 3 as the solvent system.Visualization was accomplished by exposure to short wavelength UV ( max 254).

Cunninghamella echinulata var. echinulata
C. echinulata var.echinulata ATCC 9244 was grown in fifteen 1-L culture flasks each containing 200 mL of medium .A total of 500 mg of 1 (in 5 mL DMF) were evenly distributed among the 24-h-old stage II culture.After two weeks, the incubation mixtures were checked by TLC.TLC revealed that most of 1 was transformed and two more polar metabolites were produced.The incubation mixtures were combined and filtered to remove the mycelia, and the filtrate was extracted with ethylacetate (3L x 4).The combined extracts were dried over anhydrous Na 2 SO 4 and evaporated to dryness in vacuo at 40 °C to afford a yellowish residue (0.7 gm).Part of this residue (0.4 gm) was purified by column chromatography over silica gel (45 g, 3.3 x 15 cm) using MeOH-CHCl 3 (5% v/v) as the eluting solvent system.Fractions of 20 mL each were collected and pooled on the basis of TLC analyses.Fractions 15-22 yielded pure 4'-hydroxyrofecoxib (2) (33 mg, 6.2% yield), with R f = 0.41.Fractions 35-43 gave pure 3'-hydroxyrofecoxib (3) (12 mg, 2.3% yield), with R f = 0.37.

RESULTS AND DISCUSSION
A total of sixty microorganisms, mainly fungi and actinomycetes, Table 1, have been screened for their ability to catalyze bioconversion of rofecoxib (1).Only three of them, Cunninghamella echinulata var.echinulata ATCC 9244, Mucor griseocyanus ATCC 1207b, and Rhizopus oryzae ATCC 34121, showed definite metabolism of rofecoxib.These cultures produced two more polar metabolites, 2 and 3. Their structures are shown in Figure 1.Thin layer chromatography, showed that Cunninghamella echinulata var.echinulata was the most efficient microorganism to almost completely metabolize rofecoxib, and hence, was selected for preparativescale fermentation.C NMR spectral data of 2 (Table 2) were similar to those of 1 except for resonances attributed to the non-phenylsulfone aromatic ring.The 13 C NMR spectra revealed the presence of an oxygenated aromatic carbon signal at C 157.7 which was concluded to be attached to a hydroxyl group as suggested by the IR spectrum ( max 3373 cm ) and by the presence of a D 2 O exchangeable proton at H 3.55.This hydroxyl group was determined from the HMBC spectra to be at C-4 due to the presence of 3 J cross peaks between this carbon ( C 157.7, s) and two protons resonated at H 7.58 as doublets with a J value of 8.1 Hz.This conclusion was supported by protons coupling patterns and constants that were identical to a para-substituted aromatic ring, Table 2. COSY spectra showed a correlation between these protons and other two protons both resonated at H 6.70 as doublets with a J value of 8.1 Hz.It is noteworthy the upfield shifts of carbons 1 , 3 , and 5 ( 5.0 ppm, 14.0 ppm, and 14.0 ppm, respectively) due to the resonance effect of the hydroxyl moiety, as an electron-donating group, on the aromatic ring.These data unambiguously determined that 2 is 4 -hydroxyrofecoxib (Figure 1).Metabolite 3 was isolated as colorless needles in a 2.3% yield.It was shown to possess identical molecular formula to 2, C 17 H 14 O 5 S, as derived from the presence of an ion peak at m/z 330.0584 [M] + and from the NMR data.The presence of an additional hydroxyl group and an intact , -unsaturated 5-membered ring lactone carbonyl group was concluded from the IR spectrum due to the presence of absorptions at max 3373 and 1746 cm -1 , respectively.

13
C NMR showed seventeen resonances distributed as one methyl, one methylene, eight aromatic methines, and one oxygenated aromatic quaternary carbon.Other quaternary carbons were almost identical to those of the parent rofecoxib.These data showed that metabolite 3 is similar to 2, i.e., another rofecoxib hydroxy derivative.Protons coupling patterns and constants were instrumental in determining the substitution pattern of the non-phenylsulfone aromatic ring as a meta-substituted aromatic ring, Table 2. 1 H NMR spectrum showed that one aromatic proton resonated at H 7.01 as a singlet and was assigned H-2 .On the other hand, H-5 was shown to resonate at H 7.58 as a double doublet with J values of 8.1 and 8.1 Hz.This proton was further shown to possess three-bond correlation with the oxygenated aromatic carbon ( C 158.4, s, C-3 ).Another important change in the NMR data was the upfield shifts of carbons 2 , 4 , and 6 ( 17.5 ppm, 14.7 ppm, and 8.1 ppm, respectively) due to the resonance effect of the hydroxyl moiety, as an electron-donating group, on the aromatic ring.Other NMR assignments were similar to those of rofecoxib, Table 2.These data confirmed the identity of 3 as 3 -hydroxyrofecoxib (Figure 1).
Although, compounds 2 and 3 were reported from total chemical synthesis, along with their acetyl derivatives [16], they were prepared using a tedious six-step reaction sequence giving only 5% cumulative yields.Additionally, full spectral identification of metabolites 2 and 3 were never provided in this study.Comparatively, in the current report, these metabolites were microbially produced using cheap and readily available microorganisms and medium constituents.The yield of the produced compounds can be proportionally increased by increasing the fermentation culture volume.Moreover, full spectral data and analysis of compounds 2 and 3 are provided herewith.
This study demonstrated the potential ability of microbial models to prepare known or novel metabolites in quantities that would be difficult to obtain from either mammals or chemical routes.The formation of rofecoxib metabolites 2 and 3 using the abovementioned protocol made it possible to furnish these metabolites in significant quantities enough for cardiac toxicity evaluation.Such a study findings should help in explaining if rofecoxib cardiac toxicity is classdependent, being a selective COX-2 inhibitor, or member-dependent due its unique chemical structure.Moreover, the produced metabolites might serve as starting compounds for the semisynthesis of other more promising inhibitors.
were recorded in CDCl 3 , and the chemical shift values were expressed in (ppm) relative to the internal standard TMS.For the13

Table 1 : Microorganisms Screened for their Ability to Metabolize Rofecoxib
a Absidia glauca ATCC 22752

Table 1 ). Continued. Mucor griseocyanus ATCC 1207b
a Microorganisms in bold represent those gave positive results.