Synthesis, characterization, and activity of metabolites derived from the cyclooxygenase-2 inhibitor rofecoxib (MK-0966, Vioxx(TM))

Article abstract of DOI:10.1016/S0960-894X(00)00538-2

Metabolites of the COX-2 inhibitor rofecoxib (MK-0966, Vioxx(TM)) were prepared by synthetic or biosynthetic methods. Metabolites include products of oxidation, glucuronidation, reduction and hydrolytic ring opening. Based on an in vitro whole blood assay

Full text of DOI:10.1016/S0960-894X(00)00538-2

Bioorganic & Medicinal Chemistry Letters 10 (2000) 2683±2686
Synthesis, Characterization, and Activity of Metabolites
Derived from the Cyclooxygenase-2 Inhibitor
TM
Rofecoxib (MK-0966, Vioxx )
Deborah A. Nicoll-Grith,* James A. Yergey, Laird A. Trimble, Jose M. Silva,
Chun Li, Nathalie Chauret, Jacques Yves Gauthier, Erich Grimm, Serge Leger,
Patrick Roy, Michel Therien, Zhaoyin Wang, Peppi Prasit, Robert Zamboni,
Robert N. Young, Christine Brideau, Chi-Chung Chan, Joseph Mancini and
Denis Riendeau
Â
Merck Frosst Centre for Therapeutic Research, PO Box 1005, Pointe-Claire-Dorval, Quebec, Canada H9R 4P8
Received 6 July 2000; accepted 20 September 2000
AbstractÐMetabolites of the COX-2 inhibitor rofecoxib (MK-0966, VioxxTM) were prepared by synthetic or biosynthetic methods.
Metabolites include products of oxidation, glucuronidation, reduction and hydrolytic ring opening. Based on an in vitro whole
blood assay, none of the known human metabolites of rofecoxib inhibits COX-1 nor contributes signi®cantly to the inhibition of
COX-2. # 2000 Elsevier Science Ltd. All rights reserved.
Introduction
know if any potential circulating metabolites of rofeco-
xib have COX-2 inhibitory activity. Furthermore, the
metabolites may be excreted through renal or biliary
routes and, therefore, metabolites may be present in
signi®cant levels in the kidney, liver, and intestine.
Because the inhibition of COX-1 has been associated
with the gastrointestinal side e€ects of NSAIDs,1,5 it was
important to determine whether metabolites of rofeco-
xib could be inhibitors of COX-1. This paper outlines
the syntheses of eight compounds that were observed
as metabolites in human8 and animal9 studies. The
pharmacological activities of these metabolites, as mea-
sured by human whole blood COX-1 and COX-2
assays, are reported.
Rofecoxib (4-[4-methylsulphonylphenyl]-3-phenylfuran-
2(5H)-one, 1) is a potent and selective COX-2 inhibitor
that has been shown to be e€ective for the treatment of
the symptoms of osteoarthritis and the relief of acute
pain.1,2 Rofecoxib is ecacious in models of pain and
in¯ammation with a potency comparable to indo-
methacin but with considerably improved gastro-
intestinal tolerability as compared to nonsteroidal anti-
in¯ammatory drugs (NSAIDs), which inhibit both
COX-1 and COX-2.3 5 In the case of rofecoxib, a 37- to
75-fold selectivity ratio of inhibition of COX-2 over
COX-1 has been reported using in vitro whole blood
assays.3,6 Furthermore, no inhibition of COX-1 has
been observed in clinical trials as measured by serum
thromboxane B2 production, even at doses exceeding
the maximal analgesic e€ective dose.7
Results and Discussion
Initial in vitro biotransformation experiments were car-
ried out with rat hepatic microsomal incubations sup-
plemented with NADPH and with rat hepatocytes
under standard conditions.10,11 Analysis of the mixtures,
as described below, indicated that oxidation of the 5-
position to yield 5-hydroxy rofecoxib 2 and glucuro-
nidation of this hydroxy metabolite to yield the diaster-
Drugs are often excreted as metabolites that result from
enzymatic modi®cations such as oxidation, reduction, or
conjugate formation. These metabolites may circulate in
the blood stream and may be distributed to the site of
pharmacological activity. Therefore, it is important to
eomeric b-O-glucuronides
3 constituted important
pathways in vitro in the rat (Scheme 1). HPLC analyses
of the incubates were performed under reverse-phase
*Corresponding author. Tel.: +1-514-428-8619; fax: +1-514-428-
4900; e-mail: deborah_nicoll_grith@merck.com
0960-894X/00/$ - see front matter # 2000 Elsevier Science Ltd. All rights reserved.
PII: S0960-894X(00)00538-2

2684
D. A. Nicoll-Grith et al. / Bioorg. Med. Chem. Lett. 10 (2000) 2683±2686
conditions, and the metabolite retention times are
reported in Table 1. This pathway was also observed in
rat in vivo studies.9
HPLC similar to that reported previously.11 Speci®cally,
preparative HPLC was accomplished on a YMC ODS-
AM column (20Â50 mm) with 10 mL/min mobile-
phase ¯ow rate. The mobile phase consisted of metha-
nol and 20 mM ammonium acetate in a gradient, with
methanol increasing from 10 to 90% over 20 min. Elec-
trospray mass spectrometric detection on a Finnigan
SSQ mass spectrometer was used to automate peak
detection and collection. 1H NMR of the isolated
materials at 500 MHz showed characteristic resonances
for the anomeric protons at ꢀ4.7 ppm. The coupling
constants were 7.9 Hz, consistent with trans-diaxial
coupling as expected for the formation of b-linked
O-glucuronides.11
Characterization of the incubates by capillary HPLC-
LSIMS on a JEOL-HX110 mass spectrometer operated
in positive ion mode12 indicated that 2 was an oxidation
product of rofecoxib because the observed (M+H)+
ion at 331 daltons was 16 mass units higher than for
rofecoxib (m/z 315). Furthermore, the racemic 5-hy-
droxy metabolite (2) was proposed because fragmenta-
tion observed in negative ion mode yielded an ion with
m/z 285, as shown in Scheme 1. The mass spectrum of
the 5-O-glucuronides (3) in positive ion mode gave
(M+NH4)+ ions at m/z 524, consistent with addition of
oxygen and glucuronic acid (±H2O).
In vivo experiments following dosing of rofecoxib in
animal models and humans indicated that both meta-
bolites 2 and 3 were present in urine and bile samples,
however, the O-glucuronides 3 were more abundant
than the 5-hydroxy lactone 2.8,9
The synthetic 5-hydroxy material (2) was obtained
through air oxidation of rofecoxib. This was achieved
by stirring a solution of rofecoxib in a ¯ask opened to
air in the presence of charcoal (Scheme 2). The synthetic
standard (2) gave the same chromatographic and mass
spectral data as the biosynthetic metabolite (2). The 5-
O-glucuronides (3) were prepared from 5-hydroxy rofe-
coxib (2) using rat hepatic microsomal incubations for-
Following dosing of 14C-labeled rofecoxib in the dog, a
polar urinary metabolite was isolated by a combination
of reverse-phase and normal-phase HPLC. 1H NMR
and mass spectral analysis of this metabolite suggested
that it was a hydroxybutanoic acid metabolite of
rofecoxib (accurate mass m/z 335.0954, theoretical m/z
335.0953) corresponding to reduction of the double
bond and hydrolysis of the lactone. The two stereo-
isomeric hydroxybutanoic acids (erythro and threo) and
their corresponding cis and trans lactone analogues
(4±7) could not be generated biosynthetically from
ti®ed with uridinediphospho-glucuronic acid.10
A
40 mL incubation of a 200 mM solution of 5-hydroxy
rofecoxib (2.6 mg) generated a 70% yield of a mixture of
the two diastereomeric glucuronides after 2.25 h in a
ratio of approximately 1:24, as indicated by reverse
phase HPLC. The products were isolated using a com-
bination of solid-phase extraction and preparative
Scheme 1. In vitro metabolites of rofecoxib in rat.
Table 1. Retention times and human whole blood assay IC50 values of rofecoxib and the synthetic metabolites
COX-1b
COX-2b
Retentionatime
(min)
Compound
IC50 (mM)
N (donor)
IC50 (mM)
N (donor)
Rofecoxib (1)
5-Hydroxy (2)
5-O-Glucuronides (3)
erythro-Hydroxybutanoic acid (4)
cis-Lactone (5)
trans-Lactone (6)
threo-Hydroxybutanoic acid (7)
40-Hydroxy (8)
40-O-Glucuronide (9)
17.0
16.0
12.4/13.7
5.6
14.0
15.3
7.7
14.0
8.4
18.8Æ0.9
>100
>100
>100
>100
211
3
6
3
3
3
3
6
6
0.53Æ0.02
8.7Æ2.9
>33
614
3
3
3
3
3
3
6
6
>33
>100
>30
>33
>100
>100
6.5Æ1.5
4.8Æ1.3
>100
>33
aHPLC analysis was performed on a Zorbax-C18 column (4.6 mmÂ15 cm) with UV detection at 245 nm. The mobile phase ¯ow rate was 1 mL/min
with a gradient for which solvent A was 20 mM ammonium acetate and B was methanol. The methanol was increased from 15 to 85% over 25 min.
bWhole blood assays were performed as described in ref 3.

D. A. Nicoll-Grith et al. / Bioorg. Med. Chem. Lett. 10 (2000) 2683±2686
2685
rofecoxib in rat microsomal or hepatocyte11 incuba-
tions. These potential metabolites were prepared by
synthesis as shown in Schemes 3 and 4.
Racemic cis-lactone (5) was obtained from rofecoxib
following catalytic hydrogenation (Scheme 4). This
compound is quite unstable and readily isomerizes.
Thus, treatment of this compound (5) with a catalytic
amount of DBU in methanol produced the more stable
trans-lactone 6. Hydrolysis of the trans-lactone 6 gave
the threo-hydroxybutanoic acid 7.
The erythro-hydroxybutanoic acid 4 was prepared from
rofecoxib as shown in Scheme 3. The diol was treated
with a slight excess of acetyl chloride (Et3N/DMAP) to
produce a mixture of acetates, which were separated by
silica gel chromatography and further puri®ed by trit-
uration in ether and hexanes. Absolute proof of struc-
ture for the monoacetate was obtained by regenerating
1 through Jones oxidation followed by cyclization
(LiOH then HCl).
The synthetic racemic threo-hydroxybutanoic acid 7 was
shown to have the same chromatographic, 1H NMR
and mass spectral characteristics as the isolated urinary
dog metabolite. The four materials (4±7) obtained
from chemical syntheses allowed con®rmation of
these metabolites in rat, dog, and human studies.8,9
These compounds are dicult to quantify in in vivo
samples due to interconversion between them, however,
together they comprise the most signi®cant pathway in
humans.8
Upon double bond reduction of the monoacetate
compound, the two protons adopt an anti orientation
(J32-3=12 Hz). This indicates that the side chains con-
taining carbons 1 and 4 also have an anti orientation,
preventing closure of the lactone ring. erythro-Hydroxy-
butanoic acid (4) was isolated as the sodium salt.
During development of rofecoxib, additional com-
pounds derived from metabolism of the unsubstituted
phenyl ring were observed (Scheme 5).8,9 In humans
these included an O-glucuronide of the 40-hydroxy ana-
logue (9) and the 30,40-dihydrodiol analogue (10). The
40-hydroxy intermediate 8 was prepared by treatment of
the 40-methoxy compound13 with BBr3 as shown in
Scheme 5. The 40-O-glucuronide 9 was then prepared
biosynthetically with rat microsomes and UDPGA by
methodology analogous to that described above for
racemic O-glucuronide 3. Synthesis of the 30,40-dihy-
drodiol 10, a very minor metabolite in humans,8 was not
attempted.
Scheme 2. Synthesis of the 5-hydroxy metabolite 2. This reaction was
originally developed by D. Tschaen, R. Tillyer, L. Frey, and U. Dolling.
Scheme 3. Synthesis of the erythro-hydroxybutanoic acid metabolite of rofecoxib.
Scheme 4. Synthesis of cis- and trans-lactones and the threo-hydroxybutanoic acid metabolites of rofecoxib.

2686
D. A. Nicoll-Grith et al. / Bioorg. Med. Chem. Lett. 10 (2000) 2683±2686
bene®ts of COX-2 inhibition, nor any of the potential
e€ects associated with COX-1 inhibition.
References and Notes
1. Brooks, P.; Emery, P.; Evans, J. F.; Fenner, H.; Hawkey,
C. J.; Patrono, C.; Smolen, J.; Breedveld, F.; Day, R.; Dou-
gados, M.; Ehrich, E. W.; Gijon-Banos, J.; Kvien, T. K.; Van
Rijswijk, M. H.; Warner, T.; Zeidler, H. Rheumatology 1999,
38, 779.
2. Sorbera, L. A.; Leeson, P. A.; Castaner, J. Drugs Future
1998, 23, 1287.
3. Chan, C.-C.; Boyce, S.; Brideau, C.; Charleson, S.; Crom-
lish, W.; Ethier, D.; Evans, J.; Ford-Hutchinson, A. W.;
Forrest, M. J.; Gauthier, J. Y.; Gordon, R.; Gresser, M.;
Guay, J.; Kargman, S.; Kennedy, B.; Leblanc, Y.; Leger, S.;
Mancini, J.; O'Neill, G. P.; Ouellet, M.; Patrick, D.; Percival,
M. D.; Perrier, H.; Prasit, P.; Rodger, I.; Tagari, P.; Therien,
M.; Vickers, P.; Visco, D.; Wang, Z.; Webb, J.; Wong, E.; Xu,
L. J.; Young, R. N.; Zamboni, R.; Riendeau, D. J. Pharmacol.
Exp. Ther. 1999, 290, 551.
Scheme 5. Synthesis of the 40-hydroxy metabolite 8 and the 40-O-glu-
curonide 9. Structure of metabolite 10.
4. Prasit, P.; Wang, Z.; Brideau, C.; Chan, C.-C.; Charleson,
S.; Cromlish, W.; Ethier, D.; Evans, J. F.; Ford-Hutchinson,
A. W.; Gauthier, J. Y.; Gordon, R.; Guay, J.; Gresser, M.;
Kargman, S.; Kennedy, B.; Leblanc, Y.; Leger, S.; Mancini, J.;
O'Neill, G. P.; Ouellet, M.; Percival, D. M.; Perrier, H.;
Riendeau, D.; Rodger, I.; Tagari, P.; Therien, M.; Vickers, P.;
Wong, E.; Xu, L.-J.; Young, R. N.; Zamboni, R.; Boyce, S.;
Rupniak, N.; Forrest, M.; Visco, D.; Patrick, D. Bioorg. Med.
Chem. Lett. 1999, 9, 1773.
All of the biosynthetic and synthetic derivatives were
tested in the human whole blood assays for COX-1 and
COX-2 inhibition3 and the results are reported in Table 1.
The hydroxylated compounds 2 and 8 show some inhi-
bitory e€ects on COX-2, however, they are an order of
magnitude less potent than rofecoxib. The 40-hydroxy
derivative (8) showed some moderate inhibition of
COX-1 activity. Although this material is likely a meta-
bolite en route to the 40-O-glucuronide, it was not
detected in human biological ¯uids,8 which suggests that
it is a transient species that never builds up in con-
centration. Furthermore, the excreted 40-O-glucuronide
9 accounts for <5% of the dose.8
5. DeWitt, D. L. Mol. Pharmacol. 1999, 55, 625.
6. Warner, T. D.; Giuliano, F.; Vojnovic, I.; Bukasa, A.;
Mitchell, J. A.; Vane, J. R. Proc. Natl. Acad. Sci. U.S.A. 1999,
96, 7563.
7. Ehrich, E. W.; Dallob, A.; De Lepeleire, I.; Van Hecken,
A.; Riendeau, D.; Yuan, W.; Porras, A.; Wittreich, J.; Seibold,
J. R.; De Schepper, P.; Mehlisch, D. R.; Gertz, B. J. Clin.
Pharmacol. Ther. 1999, 65, 336.
8. Halpin, R. A. et al. Manuscript in preparation.
9. Halpin, R. A.; Geer, L. A.; Zhang, K. E.; Marks, T. M.;
Dean, D. C.; Jones, A. N.; Melillo, D.; Doss, G.; Vyas, K. P.
Drug Metab. Dispos. 2000, 28, 1244.
10. Nicoll-Grith, D.; Yergey, J.; Trimble, L.; Williams, H.;
Rasori, R.; Zamboni, R. Drug Metab. Dispos. 1992, 20, 383.
11. Nicoll-Grith, D. A.; Falgueyret, J.-P.; Silva, J. M.;
Morin, P.-E.; Trimble, L.; Chan, C.-C.; Clas, S.; Leger, S.;
Wang, Z.; Yergey, J. A.; Riendeau, D. Drug Metab. Dispos.
1999, 27, 403.
Conclusion
Of the potential metabolites reported in Table 1, all but
the 40-hydroxy derivative (8) have been observed in
humans.8 The only human metabolite that shows any
e€ect on COX-2 activity is the 5-hydroxy metabolite (2)
with an IC50 17-fold less than the IC50 of rofecoxib.
None of the human metabolites showed measurable
e€ects on COX-1 activity. These metabolites are, there-
fore, not anticipated to contribute to the therapeutic
12. Li, C.; Chauret, N.; Ducharme, Y.; Trimble, L. A.; Nicoll-
Grith, D. A.; Yergey, J. A. Anal. Chem. 1995, 67, 2931.
13. Ducharme, Y.; Gauthier, J. Y.; Prasit, P.; Leblanc, Y.;
Wang, Z.; Leger, S.; Therien, M. US Patent 5,550,142, 1996.