CYP2C9 structure-metabolism relationships: Optimizing the metabolic stability of COX-2 inhibitors

Article abstract of DOI:10.1021/jm0705096

The cytochrome P450 (CYP) family is composed of a large group of monooxygenases that mediate the metabolism of xenobiotics and endogenous compounds. CYP2C9, one of the major isoforms of the CYP family, is responsible for the phase I metabolism of a variety of drugs. The aim of the present investigation is to use rational design together with MetaSite, a metabolism site prediction program, to synthesize compounds that retain their pharmacological effects but that are metabolically more stable in the presence of CYP2C9. The model compound for the study is the nonsteroidal anti-inflammatory drug celecoxib, a COX-2 selective inhibitor and known CYP2C9 substrate. Thirteen analogs of celecoxib were designed, synthesized, and evaluated with regard to their metabolic properties and pharmacologic effects. The docking solutions and the predictions from MetaSite gave useful information leading to the design of new compounds with improved metabolic properties.

Full text of DOI:10.1021/jm0705096

4444
J. Med. Chem. 2007, 50, 4444-4452
CYP2C9 Structure-Metabolism Relationships: Optimizing the Metabolic Stability of COX-2
Inhibitors
Marie M. Ahlstro¨m,*,†,§ Marianne Ridderstro¨m,† Ismael Zamora,‡ and Kristina Luthman§
DiscoVery DMPK and Bioanalytical Chemistry, AstraZeneca R&D Mo¨lndal, S-431 81 Mo¨lndal, Sweden, Lead Molecular Design,
S. L., Valle´s 96-102 (27) E-08190, Sant Cugat del Valle´s, Spain, Institut Municipal d’InVestigacio´ Medica (IMIM), UniVersitat
Pompeu Fabra, Doctor Aiguader 80, 08003 Barcelona, Spain, and Department of Chemistry, Medicinal Chemistry, Go¨teborg UniVersity,
SE-412 96 Gothenburg, Sweden
ReceiVed May 2, 2007
The cytochrome P450 (CYP) family is composed of a large group of monooxygenases that mediate the
metabolism of xenobiotics and endogenous compounds. CYP2C9, one of the major isoforms of the CYP
family, is responsible for the phase I metabolism of a variety of drugs. The aim of the present investigation
is to use rational design together with MetaSite, a metabolism site prediction program, to synthesize
compounds that retain their pharmacological effects but that are metabolically more stable in the presence
of CYP2C9. The model compound for the study is the nonsteroidal anti-inflammatory drug celecoxib, a
COX-2 selective inhibitor and known CYP2C9 substrate. Thirteen analogs of celecoxib were designed,
synthesized, and evaluated with regard to their metabolic properties and pharmacologic effects. The docking
solutions and the predictions from MetaSite gave useful information leading to the design of new compounds
with improved metabolic properties.
Introduction
An important aspect of the drug discovery process is the
optimization of the metabolic properties of test compounds.
Hence, it should be of value to be able to predict those structural
modifications that are likely to lead to altered metabolic
properties while retaining the desired pharmacological profile.
This paper uses rational design together with MetaSite, a
metabolism site prediction program, to synthesize analogs of
celecoxib (Figure 1) that retain COX-2 inhibiting properties but
that are metabolically more stable in the presence of CYP2C9.
Figure 1. Celecoxib (4-[5-(4-methylphenyl)-3-trifluoromethyl-1H-
pyrazol-1-yl]-benzenesulfonamide).
Structural characteristics for a compound to be selectively
recognized by CYP2C9 include the presence of an anionic site
and a hydrophobic zone between the substrate hydroxylation
site and the anionic site.1 CYP2C9 catalyzes the oxidation of
several important drugs including diclofenac and warfarin.2,3 It
is reasonable to speculate that COX-2 inhibitors that are poor
CYP2C9 substrates will display improved clearance and half-
life properties. In the present study we explore computational
methods to exploit the recently published crystal structures of
mammalian CYP2C94,5 in an effort to improve our understand-
ing of the metabolic pathways of this class of compounds that
are catalyzed by CYP enzymes.
Celecoxib is a COX-2 selective inhibitor known to be rapidly
metabolized to the hydroxymethyl metabolite by CYP2C9.6,7
The compound has a fairly rigid core structure with three
apparent interaction points: R1, R2, and R3 (Figure 1). Celecoxib
and other nonsteroidal anti-inflammatory drugs (NSAIDsa)
mediate their anti-inflammatory actions primarily through the
inhibition of cyclooxygenase (COX), the enzyme that catalyzes
Figure 2. SC-558 (4-[5-(4-bromophenyl)-3-trifluoromethyl-1H-pyra-
zol-1-yl]-benzenesulfonamide).
the conversion of arachidonic acid to prostaglandin. COX exists
in at least two isoforms, COX-1, a constitutive enzyme, and
COX-2, an inducible isoform.8 The selective inhibition of
COX-2 leads to improved anti-inflammatory properties with
fewer gastrointestinal side effects.9,10 However, COX-2 inhibi-
tors recently have been shown to cause other serious side effects
such as an increased risk of adverse cardiovascular events.11
According to the crystal structure of COX-2 in which the
active site is occupied by the selective NSAID celecoxib
analogue SC-558 (Figure 2), the active site is mainly hydro-
phobic. The amino acids in close proximity of the selective
COX-2 inhibitor, SC-558, and the GRID molecular interaction
fields (MIFs) for hydrophobic (DRY), hydrogen-bond acceptors
* To whom correspondence should be addressed. Tel.: +46 31 776 1274.
Fax: +46 31 776 3787. E-mail: marie.m.ahlstrom@astrazeneca.com.
† AstraZeneca R&D Mo¨lndal.
‡ Lead Molecular Design and Institut Municipal d’Investigacio´ Medica
(IMIM).
§ Go¨teborg University.
a Abbreviations: COX, cyclooxygenase; CSA, camphorsulfonic acid;
CYP, cytochrome P450; MIF, molecular interaction field; NSAID, non-
steroidal anti-inflammatory drug; PGE2, prostaglandin E2; SAR, structure-
activity relationship; TBHP, t-butyl hydroperoxide.
10.1021/jm0705096 CCC: $37.00 © 2007 American Chemical Society
Published on Web 08/14/2007

CYP2C9 Structure-Metabolism Relationships
Journal of Medicinal Chemistry, 2007, Vol. 50, No. 18 4445
Figure 3. MIFs calculated in the COX-2 (PDB: 1CX2) crystal using three different probes (a) DRY (-0.2 kcal/ mol), (b) O (-4.0 kcal/mol), and
(c) N1 (-6.0 kcal/mol) are viewed together with the cocrystallized compound, SC-558.
(O), and hydrogen-bond donors (N1) are shown in Figure 3.
The bromine atom is surrounded by Phe381, Leu384, Tyr385,
Trp387, Phe518, and Ser530, with additional contributions from
the backbone atoms of Gly526 and Ala527. The trifluoromethyl
group is surrounded by Met113, Val116, Val349, Tyr355,
Leu359, and Leu531. The sulfonamide moiety interacts with
His90, Gln192, and Arg513. The NH group in the sulfonamide
moiety forms a hydrogen bond to the carbonyl oxygen of
Phe518.12 Retaining the pharmacological activity (COX-2
inhibition) was approached by taking advantage of these
previously published data describing which amino acids are of
importance for the affinity to COX-2.12 The structure-activity
relationship (SAR) around the 1,5-diarylpyrazolyl class of
compounds published by Penning et al.13 was also considered.
MetaSite14,15 is a tool that can be applied to identify the
potential sites of metabolism in a compound (hot spots). This
Figure 4. Celecoxib docked in the active site of the crystal structure
of CYP2C9 (PDB: 1R9O).
information can be useful when attempting to design analogs
with improved metabolic properties. MetaSite is based on two
factors: protein ligand similarity analysis and chemical reactivity
of the substrate. MetaSite performs a similarity analysis of the
protein-cavity interaction profile and the potential substrate.
The chemical reactivity of fragments toward oxidation, repre-
senting the activation energy of the conversion of substrate to
metabolite, is precomputed and stored within the program.
Hence, if the similarity search gives a high score for a fragment,
the chemical reactivity of that fragment is included in the final
“hot spot” prediction.14,16
The rational design of analogs to celecoxib with potentially
Figure 5. Plots of MetaSite predictions. (a) Site of metabolism: The
groups in celecoxib that most likely will be metabolized by CYP2C9
are marked in dark brown; the darker the color the higher is the
probability of metabolism to occur. (b) Contribution plot: The atoms
improved metabolic properties was aided by docking celecoxib
into the active site of CYP2C9 (PDB: 1R9O)4 using the docking
program GLUE17 (see below). The site of metabolism and the
molecular contribution to the exposure of a reactive atom to
the heme moiety (activity contribution plot) were determined
by MetaSite. The interactions with COX-2 needed to keep
COX-2 inhibition were investigated using the crystal structure
of COX-2 cocrystallized with the selective NSAID celecoxib
analogue SC-558 (Figure 3).
in celecoxib with molecular contribution to the exposure of the
metabolic “hot spot” (red circle) to the heme moiety of CYP2C9 are
marked dark brown; the darker the color the higher the probable
contribution.
5a) and celecoxib is known to be hydroxylated at the methyl
group by CYP2C9,7 hence, all other docking solutions could
be discarded.
Results and Discussion
Modeling of Celecoxib Metabolism. GLUE, a GRID-based
docking program,17 was used in this study to analyze ligand-
receptor interactions. Celecoxib was docked into the active site
of CYP2C9 using the available crystal structure (PDB: 1R9O),
without crystallographic water molecules. Multiple docking
solutions were obtained, but only one docking solution had the
carbon atom in the methyl group within 3 Å from the ferryl
oxygen (Figure 4). MetaSite predicted that the methyl group in
celecoxib was most likely to be metabolized by CYP2C9 (Figure
The carbon atom in the methyl group is docked 2.9 Å from
the ferryl oxygen. This is a productive mode because it can
lead to metabolite formation. Amino acids within 3 Å from the
trifluoromethyl group are Leu208, Phe476, and Phe100. A bulky
group in this position may fail to fit into the pocket and thus
decrease the affinity for CYP2C9 of the new celecoxib analog.
The sulfonamide is docked 2.7 Å from the Arg108. This docking
solution indicates that altering the interaction between the
sulfonamide group and the Arg108 may decrease the interaction

4446 Journal of Medicinal Chemistry, 2007, Vol. 50, No. 18
Ahlstro¨m et al.
between the new celecoxib analog and CYP2C9. The contribu-
tion plot from MetaSite also indicates the importance of this
interaction (Figure 5b).
The site of metabolism predictions were performed by the
program MetaSite. These analyses involve the calculation of
two sets of descriptors, one set of descriptors for the CYP
enzyme and one for the potential substrate. The resulting
analysis provided a “chemical fingerprint” of the enzyme and
the substrate. Chemical fingerprints are distance-based descrip-
tors that are calculated from MIFs computed by GRID.
Moreover, as mentioned earlier, the program considers the
chemical reactivity of the compound by taking into consideration
the activation energy in the hydrogen abstraction step that
ultimately leads to metabolite formation. These hydrogen
abstraction processes have been simulated by ab initio calcula-
tions of small fragments from druglike substrates for human
CYPs and stored in MetaSite.15 When a fragment in the target
molecule is recognized as one in the MetaSite database, all
atoms in that fragment are assigned the corresponding reactivity
value. The final ranking for potential metabolic sites is the
product of the similarity analysis and the chemical reactivity.
The information used for the prediction was the structure of
CYP2C9 and several conformations of the compound of interest.
MetaSite correctly predicted that the methyl group in celecoxib
was most likely to be metabolized by CYP2C9 (Figure 5a). The
prediction was achieved without any input of information of
known metabolites.
Figure 6. Synthesized analogs of celecoxib (substituents highlighted
in bold).
Table 1. Summary of Celecoxib and Its Synthesized Analogs
Metabolism in CYP2C9 after a 30 Min Incubation and Percent
Inhibition of PGE2 Formation in COX-1 and COX-2
% inhibition of
control value
% metabolized
CYP2C9
(1 µM)
COX-2 COX-1
(100 µM) (100 µM)
cmpd
R1
R2g
R3
MetaSite can also provide an activity contribution plot that
shows the molecular contribution to the exposure of a reactive
atom toward the heme moiety.15 The program locks the group
that has been predicted to be metabolized by the CYP close to
the heme moiety. Subsequently, the program allows the potential
substrate to move within the active site, keeping the group fixed,
and computes the energy of interaction between the compound
and CYP. Interactions considered during the computation are
hydrophobic, charge, and hydrogen-bond interactions. The
molecular moieties in the potential substrate, which have the
strongest impact for directing the metabolic hot spot toward
the heme moiety, are determined. In the analysis of celecoxib,
the activity contribution plot points out the nitrogen in the
sulfonamide moiety as the responsible atom for orienting the
methyl group toward the heme (Figure 5b).
celecoxib -CH3 -CF3 -SO2NH2
94
29
29
49
87
48
81
2
88
0
30
4
96
28
67
100
95
82
99
55
55
100
94
96
6
1
2
3
4
5
6
7
8
9
10
11
12
13c
-CF3 -CF3 -CO2H
-CH3 -CF3 -CO2H
-CH3 phenyl -SO2NH2
-CH3 -CF3 -SCH3
-CH3 -CF3 -SOCH3
-CH3 -CF3 -SO2CH3
-CF3 -CF3 -CO2Et
-CH3 -CF3 -CO2Et
-CF3 -CF3 -SO2NH2
-CH3 -CF3 -CH3
26
92
99
99
92
96
87
86
98
93
95
43
-CF3 -CF3 -CH2CO2H
-CH3 -CF3 -CH2CO2H
-CH3 -CO2H -SO2NH2
96
88
66
10
61
of celecoxib, namely, R1, R2, and R3 (Figure 6). The design of
the celecoxib analogs was challenging due to the fact two
enzymes must be considered.
One advantage achieved with the MetaSite program compared
to docking is that no time-consuming visual inspection of the
potential interactions between the compound and the metaboliz-
ing enzyme is needed. The site of metabolism is predicted and
shown in schematic plots. The main disadvantage is that no
information about which amino acids that interact with the
compound in the active site is achieved.
Modifications in the R1 position were of interest because
MetaSite had predicted correctly that the methyl group was most
likely to be metabolized by CYP2C9. The initial structural
change was made in the R1 position in which the CH3 group of
celecoxib (Figure 1) was replaced with a CF3 group to give 9
(for the structures of compounds see Table 1). We suspected
that this relatively minor structural alteration would not result
in loss of COX-2 inhibitory activity but would block the
principal CYP2C9-catalyzed oxidation.
Structural knowledge gained from the crystal structures of
COX-2 (Figure 3) and CYP2C9 (Figure 4) guided the modifica-
tions in the R2 position. In the cocrystallized structure, the
trifluoromethyl group of SC-558 resides in a large pocket that
is constructed from several hydrophobic amino acids (Val116,
Leu117, Tyr355, Phe357, and Leu359) and a positively charged
amino acid (Arg120). The dockings of celecoxib into the active
site of CYP2C9 indicated that a bulky group in the R2 position
could cause a clash with Phe100. Hence, analog 3, containing
a phenyl group at C(3) of the pyrazolyl moiety, was synthesized
to investigate if this modification would lead to a less-productive
interaction with CYP2C9. The 3-position of the pyrazolyl ring
Design of Celecoxib Analogs. Two approaches were com-
bined for the design of new analogs of celecoxib with the
potential for improved metabolic stability. One approach was
rational design using information obtained from crystal struc-
tures. The crystal structure of COX-2 (PDB: 1CX2)12 was
available cocrystallized with the celecoxib analog SC-558
(Figure 2). No celecoxib analog cocrystallized with CYP2C9
was available, therefore, the rational design was performed by
using the docking pose of celecoxib in the active site of CYP2C9
(PDB: 1R9O).4 The other approach was to use the site of
metabolism prediction tool, MetaSite, that predicts which group
in a compound is most likely to be metabolized by CYP2C9.
The program also predicts which atom may be responsible for
the positioning of metabolically susceptible groups close to the
heme moiety. Structural changes were made in three positions

CYP2C9 Structure-Metabolism Relationships
Journal of Medicinal Chemistry, 2007, Vol. 50, No. 18 4447
Scheme 2a
Scheme 1a
a (1) R3 ) -CO2H; R1 ) -CF3; (2) R3 ) -CO2H; R1 ) -CH3; (11)
R3 ) -CH2CO2H; R1 ) -CF3; (12) R3 ) -CH2CO2H; R1 ) -CH3.
Reagents and conditions: (a) NaH, dry THF, ethyl trifluoroacetate; (b) R3-
phenylhydrazine‚HCl, acetonitrile, reflux.
a (4) R3 ) -SCH3; R1 ) -CH3; (7) R3 ) -CO2Et; R1 ) -CF3; (8) R3
) -CO2Et; R1 ) -CH3; (9) R3 ) -SO2NH2; R1 ) -CF3; (10) R3
)
-CH3; R1 ) -CH3. Reagents and conditions: (a) NaH, dry THF, ethyl
trifluoroacetate; (b) R3-phenylhydrazine‚HCl, EtOH, reflux.
ucts.13 However, Penning et al. showed that the reaction
proceeds regiospecifically with the hydrochloride salt of phe-
nylhydrazine in refluxing ethanol to give exclusively the 1,5-
diaryl analogs. When a bulky alkyl group is present on the
diketone, the major product is usually the 1,5-diarylpyrazole.18
This is probably due to the fact that the attack of the -NH2
end of the phenylhydrazine is favored at the less sterically
hindered carbonyl group.19 In the present study, the reaction
was highly regioselective for the 1,5-diaryl analogs, except for
3, a compound with a phenyl group instead of the trifluorom-
ethyl group in the R2 position. With two equally large groups
on the diketone, no regioselectivity was achieved. To obtain
the desired isomer 3 in pure form, the synthetic route presented
in Scheme 313 was pursued.
An aldol condensation between 4-methylbenzaldehyde (III)
and acetophenone gave the chalcone (IV), which subsequently
was further oxidized to the epoxide (V). Reaction of the epoxide
with phenylhydrazine provided exclusively the 1,5-diarylpyra-
zolyl isomer (3). Compounds 5 and 6 were prepared by
oxidation of the sulfide (4) (Scheme 4). Oxidation of (4) by
t-butyl hydroperoxide (TBHP) in the presence of a catalytic
amount of camphorsulfonic acid (CSA) gave the sulfoxide (5)
in good yield.20 The sulfone (6) was prepared by oxidation of
4 with oxone.21
The carboxylic acid 13c (Scheme 5) was synthesized starting
with a Claisen condensation of 4-methylacetophenone and
diethyl oxalate in the presence of sodium ethoxide.22 The
resulting 2,4-diketo ester was heated under reflux in ethanol
with 4-sulfamoylphenylhydrazine followed by hydrolysis23 of
the resulting ester to yield the desired carboxylic acid 13c.
Metabolism Studies in CYP2C9. Celecoxib and all synthe-
sized analogs were examined for their metabolic properties using
an in Vitro assay employing recombinant CYP2C9. The test
compounds (1 µM) were incubated with CYP2C9 over a period
of 30 min. The relative amounts of the parent compounds
remaining after the incubation gives an indication on how stable
the compounds are in the test system. The different analogs
with chemical modifications in the R1, R2, and R3 positions
allowed an analysis of the relationship between the structure
and the metabolic properties (Table 1). Celecoxib is known to
be metabolized mainly at the methyl group.7 The trifluoromethyl
analog 9 proved to be stable metabolically. To investigate
whether changes in the R2 position would prevent the interaction
with CYP2C9, analogs 3 and 13c containing a phenyl and a
carboxyl group, respectively, were studied. The docking solution
of celecoxib in the active site of CYP2C9 indicated that there
would be no space for a phenyl group in this position. Although
3 was a CYP2C9 substrate, the phenyl group decreased the
extent of metabolism at 30 min to only 49% compared to 94%
for celecoxib. Because the compound was metabolized, it could
fit into the active site, but the decreased metabolism indicates
that the interactions were weakened. Analog 13c with a carboxyl
(R2) is known to have very few steric restrictions in COX-2.13
Therefore, the COX-2 inhibition properties should be retained.
Due to the fact that there is an Arg120 in the pocket of COX-2,
a compound containing a carboxylic acid group (compound 13c)
was also synthesized in an effort to increase stabilizing polar
interactions between the substrate and the enzyme.
Modifications in the R3 position were of interest because
according to the MetaSite prediction the sulfonamide moiety
of celecoxib may be responsible for the positioning of the methyl
group close to the heme moiety. The changes in this position
had to be performed with some caution. The sulfonamide group
is critical for the selectivity for COX-2 inhibition because it
binds to a pocket that is more restricted in COX-1 than in COX-
2. This pocket is unoccupied in complexes of COX-2 with
nonselective inhibitors.12 The docking of celecoxib in CYP2C9
indicated an interaction between the sulfonamide group and
Arg108. In an attempt to disrupt this interaction, the sulfonamide
group of celecoxib was replaced with a methyl group to give
compound 10. Other structural changes were also performed in
this position. Three compounds, the sulfide (4), the sulfoxide
(5), and the sulfone (6), representing different oxidation states
on the sulfur atom, were studied. This series of compounds was
synthesized to evaluate the influence of the hydrogen-bonding
interactions between the oxygens and Arg108 in CYP2C9.
In compound 2, a carboxyl group has replaced the sulfona-
mide group. Compound 8 is the corresponding ethyl ester. These
two compounds provided another opportunity to evaluate the
influence of potential hydrogen-bond interactions in the active
site of CYP2C9 that are implicated from the docking studies.
For comparison, the corresponding two analogs with a trifluo-
romethyl group in the R1 position were also examined. The
resulting analogs are compound 1, with a carboxyl group in
the R3 position, and compound 7, with an ethoxycarbonyl group
in the R3 position. A further investigation of the effects of a
carboxyl group in the R3 position was pursued with compounds
11 and 12. These two compounds have an additional methylene
unit between the aromatic ring and the carboxyl group. These
analogs provided an opportunity to evaluate the importance of
conformational flexibility on both CYP2C9 and COX-2 activity.
Moreover, the influence of the distance between the carboxyl
group and the methyl group could be studied by comparing
compounds 1 and 2 in which the carboxyl group is attached
directly to the phenyl ring.
Synthesis. The general method employed for the synthesis
of the celecoxib analogs (Schemes 1 and 2) is patterned after
the route described by Penning et al.13
Claisen condensation of a para-substituted acetophenone
derivative (I) with ethyl trifluoroacetate in the presence of NaH
gave the 1,3-diketo adduct (II) in good yield. The reaction of
1,3-diketones with phenylhydrazine has been reported to yield
a mixture of 1,5-diarylpyrazolyl and 1,3-diarylpyrazolyl prod-

4448 Journal of Medicinal Chemistry, 2007, Vol. 50, No. 18
Ahlstro¨m et al.
Scheme 3a
a Reagents and conditions: (a) MeOH, NaOH (50% w/v aq solution), acetophenone; (b) 30% H2O2, 4 M NaOH, EtOH, acetone; (c) (4-
sulfamoylphenyl)hydrazine‚HCl, EtOH.
Scheme 4a
a Reagents and conditions: (a) THF/MeOH (1:1), oxone; (b) dichloromethane, TBHP, CSA.
Scheme 5a
a Reagents and conditions: (a) EtOH, diethyl oxalate; (b) (4-sulfamoylphenyl)hydrazine‚HCl, EtOH, reflux; (c) THF/MeOH/LiOH (2 M, 1:1:1), RT.
group in the R2 position provided information on whether or
not this replacement could lead to additional hydrogen bond
formation in the active site. The metabolism was decreased (61%
consumed), indicating that if additional hydrogen bonds were
formed, they did not enhance the positioning of the methyl
relative to the heme moiety in a productive way. In the case of
compound 10, a methyl group replaces the sulfonamide group
of celecoxib. This modification led to a 70% decrease in
metabolism, suggesting a poor fit of this compound in the active
site of CYP2C9. The decreased metabolism also emphasizes
the importance of the interactions between the sulfonamide
moiety and CYP2C9.
these results indicate that the compounds hydrophobic core
structure has a major impact on the total interaction with
CYP2C9.
All compounds with a carboxylic acid group in the R3 position
(1, 2, 11, and 12) showed a clear decrease in the metabolism.
For compounds 1 (29%) and 11 (4%), one possible explanation
for this behavior could be that the methyl group has been
replaced with a trifluoromethyl group, which is not susceptible
for metabolism. The decreased metabolism of the other two
compounds, 2 (29%) and 12 (10%), indicates that the carboxylic
acid group does not direct the methyl toward the heme moiety
to the same extent as the sulfonamide group. Structural
characteristics of CYP2C9 substrates include the presence of
an anionic site and an hydrophobic zone between the substrate
hydroxylation site and the anionic site.1 The different distances
between the carboxyl group and the methyl group in compounds
2 and 12 may have an influence on the metabolism. Compound
12 was metabolized to a lesser extent than compound 2,
suggesting that the increased conformational flexibility and
distance between the methyl group and the carboxyl group may
be unfavorable for metabolism. The ethoxycarbonyl analog 8
underwent CYP2C9-catalyzed oxidation almost to the same
extent as celecoxib (88% vs 94%). The corresponding ethoxy-
carbonyl analog 7, that also bears a trifluoromethyl group in
the R1 position, instead proved to be a very poor substrate (2%
Finally, the three compounds representing different oxidation
states on the sulfur atom were studied. The sulfonamide group
was replaced by a sulfide (4), a sulfoxide (5), or a sulfone (6).
The analysis of the docking of celecoxib into the active site of
CYP2C9 indicated that the oxygen atoms attached to sulfur
could result in strong interactions with Arg108. From these
observations, the sulfone (6) was expected to be metabolized
to a greater extent than the sulfoxide (5) and the sulfoxide (5)
to a greater extent than the sulfone (4). The results from the in
Vitro assay showed no decrease in the extent of metabolism for
compound 4 and only a modest decrease to 48% consumed for
5. Removing apparently important interactions with the enzyme
had only a minor or no influence on the metabolism. Hence,

CYP2C9 Structure-Metabolism Relationships
Journal of Medicinal Chemistry, 2007, Vol. 50, No. 18 4449
conversion at 30 min). This observation leads us to conclude
that the difference in the metabolism of these two compounds
was not due to the metabolism of the ester.
CYP2C9. Another issue, which is not considered in this study,
is the inhibition potential of the compound toward the metabo-
lizing enzyme. Modifying the metabolic “hot spot” may decrease
the rate of metabolism of a compound, but because other steric
and hydrophobic properties are retained, the outcome may be a
metabolically stable compound that still fits into the enzyme’s
active site. The compound may no longer be a substrate for the
isoform, but there is a possibility that it will display inhibitor
properties instead of substrate properties. Ongoing studies of
this issue and further details will be disclosed in future
publications. The design of the celecoxib analogs was chal-
lenging due to the fact that two enzymes had to be considered.
Previously published data about the amino acids that are
important for the affinity of inhibitors to the COX-2 enzyme
and published SAR of COX-2 inhibitors were also considered
when designing the new compounds.12,13,24 When this informa-
tion was taken into account, it was possible to retain the COX-2
inhibition properties in the new analogs. Dockings and the
predictions from MetaSite gave useful information leading to
the design of new compounds with improved metabolic proper-
ties. MetaSite provided two approaches to alter the metabolic
properties either to modify the metabolic “hot spot” or to modify
the chemical group responsible for directing the “hot spot”
toward the heme moiety. Which group in the compound that
should be modified depends on the compound’s pharmacophore.
Both approaches were successful in terms of decreasing the
metabolism in CYP2C9 compared to the parent compound
celecoxib. The synthesized compounds increased the knowledge
about the active site of the two enzymes, COX-2 and CYP2C9.
Combining the structural information gained from the docking
studies with MetaSite predictions is an excellent approach to
design pharmacologically active compounds with improved
metabolic properties.
COX-1 and COX-2 Inhibition. Celecoxib and all synthe-
sized analogs were analyzed for their effect on the activity of
human COX-1/COX-2. The results are presented in Table 1 and
expressed as percent inhibition of the control enzyme activity.
The inhibition of COX-1 and COX-2 was determined at a
compound concentration of 100 µM. Celecoxib has an IC50 of
0.04 µM for COX-2 and 15 µM for COX-1,13 hence, the
selectivity cannot be captured in this study.
Replacing the methyl group in the R1 position of celecoxib
with a trifluoromethyl group to obtain 9 did not change the
inhibition of COX-1 and COX-2. Compounds 7 and 8 have an
ester functionality in the R3 position but differ in the R1 position.
Compound 8 has a methyl group in the R1 position, while
compound 7 has a trifluoromethyl group. These two compounds
had similar COX-1 and COX-2 inhibition. Hence, modifications
in the R1 position did not affect the inhibition of the two
isoforms.
The active site of COX-2 is mainly hydrophobic, therefore
replacing the CF3 group with a phenyl group would most likely
increase the interactions. The in vitro result verified this
hypothesis because 3 showed potent inhibitory effects on both
enzymes. Altering the R2 position to a carboxylic acid as in
13c decreased both the COX-2 and the COX-1 inhibition.
Changing celecoxib in the R3 position to a sulfide (4), a
sulfoxide (5), or a sulfone (6) did not effect COX-1 and COX-2
inhibition. Four analogs with a carboxyl group in the R3 position,
two analogs with an additional methylene unit between the
aromatic ring and the carboxyl group (11 and 12) and two with
the carboxyl group directly attached to the ring (1 and 2) were
compared for their effects on COX-1 and COX-2. The car-
boxymethyl analogs retained their inhibitor properties for both
isoforms of COX, whereas the inhibitory capacity decreased
for the other two compounds, especially the COX-1 inhibition.
Compounds 7 and 8, the corresponding ethyl esters of 1 and 2,
showed a dramatic and selective increased inhibition of COX-1
compared to the carboxy analogs. Interestingly, compound 10,
which lacks the sulfonamide moiety, inhibited both COX-1 and
COX-2 to the same extent as celecoxib. The sulfonamide moiety
is known to be important for the COX-2 selectivity. Conse-
quently, if compound 10 were to be tested at a lower concentra-
tion, even though it has a low IC50, one might expect to see no
selectivity between the two isoforms.
Experimental Section
Modeling of Celecoxib Metabolism. All calculations were
performed in a Linux environment on a 32 MB personal computer.
The software utilized in the computational analysis was GRID v.
21 and MetaSite v. 2.5 (Molecular Discovery Ltd., http://moldis-
covery.com), SYBYL 6.9 (Tripos Associates Inc., St. Louis, MO),
and MacroModel v. 7.0.110 (Schro¨dinger, NY).
Celecoxib was submitted to the program MetaSite that is a fully
automated procedure. The CYP2C9 structure used in MetaSite is
the homology model25 based on the CYP2C5 (PDB: 1DT626)
structure. The MIF for CYP2C9 obtained from the GRID package
are precomputed and stored inside the software. Once the structures
of the compounds are provided, the semiempirical calculations,
pharmacophoric recognition, descriptor handling, similarity com-
putation, and the reactivity computation are all carried out automati-
cally.14,15
The human CYP2C9 crystal structure complexed with flurbi-
profen at 2.0 Å resolution was used in the docking study (PDB:
1R9O). This is a wild-type enzyme, except that the N-terminal
membrane insertion peptide has been removed to increase solubility
for crystallization. The modifications done to the protein were to
delete the water molecules and to remove the ligand from the
binding site. An additional modification was to add a dummy atom
2.0 Å above the iron. This will decrease the importance of the
electrostatic interactions and hinder the docked compounds to
interact directly with the iron in the heme. The dummy atom used
was OES, which represents an explicit tetrahedral ester oxygen atom
that does not accept any hydrogen bonds. The PDB file had to be
modified before being imported into GLUE, which is performed
by Greater that converts the PDB format to the format of the input
files (kout files) required to run the docking procedure.
To be considered a potent COX-2 inhibitor, a compound
should have an IC50 value lower than 1 µM. Optimizing the
compounds toward these properties, that is, selectivity and
potency, is not within the scope of this study. The inhibition
data of the COX enzymes demonstrate that most compounds
retain their inhibitory capacity, thus, the design toward improved
metabolic properties has not led to structures lacking the
pharmacophore for COX-2 inhibitors.
Conclusions
When a compound is optimized toward the pharmacological
target of interest it is important to take its metabolic properties
into account as early as possible in the drug discovery process.
Maintaining the pharmacological activity during the design of
more metabolically stable compounds is a challenging task. The
aim of the present investigation is to use rational design together
with MetaSite, a metabolism site prediction program, to
synthesize compounds that retain their pharmacological effects
but that are metabolically more stable in the presence of
The box size used was 20 × 20 × 20 Å, centered round the
crystallographic ligand coordinates. The probes, representing the

4450 Journal of Medicinal Chemistry, 2007, Vol. 50, No. 18
Ahlstro¨m et al.
different atoms,27 were selected manually when performing the
binding site precalculations in GLUE. These probes were selected
manually when performing the binding site precalculations in
GLUE. Default values were used when performing the docking,
and the result was analyzed visually. The structure of celecoxib
was drawn in Sybyl. Conformation search followed by energy
minimization was performed in MacroModel (v. 7.0.110). Ten
conformations were docked into the active site of CYP2C9 using
the docking program GLUE.17
described above. The precipitate of E was collected by filtration
and washed with cool EtOH to give pure MA11 (1.32 g, 96%).
3-(4-Methylphenyl)-1-phenyl-2-propen-1-one (F).30 Aqueous
NaOH (50%, 0.5 mL) was added to a stirred solution of 4-methyl-
benzaldehyde (590 µL, 5.0 mmol) and acetophenone (585 µL, 5.0
mmol) in MeOH (20 mL). The mixture was stirred overnight at
RT. The precipitate was filtrated to give F (866 mg, 78%).
2,3-Epoxy-3-(4-methylphenyl)-1-phenylpropanone (G).13 To
a solution of F (603 mg, 2.71 mmol) in EtOH (15 mL) and acetone
(5 mL) at 50 °C was added H2O2 (30%, 2 mL) and NaOH (4 M,
1.5 mL). After stirring for 3 h, the resulting precipitate was filtered
off and dried to obtain G (643 mg, 99%) as a white solid.
4-[3-Trifluoromethyl-5-(4-trifluoromethylphenyl)-1H-pyrazol-
1-yl]-benzoic Acid (1).13 4-Hydrazinobenzoic acid HCl (69.8 mg,
0.37 mmol) was added to a stirred solution of B (93.4 mg, 0.33
mmol) in acetonitrile (15 mL). The mixture was heated at reflux
for 20 h. After cooling to room temperature, the reaction was
concentrated under vacuum. The residue was taken up in EtOAc
(25 mL), washed with water (10 mL) and brine (10 mL), dried
over MgSO4, filtered, and concentrated under vacuum. Purification
with flash chromatography (dichloromethane/MeOH 99:1) gave 1
(87 mg, 66%). Purification by preparative HPLC yielded (50 mg,
40%) 1: 1H NMR (CD3OD) δ 7.02 (s, 1H), 7.33 (d, J ) 8.0, 2H),
7.43 (d, J ) 8.0, 2H), 7.62 (d, J ) 8.0, 2H), 8.02 (d, J ) 8.0, 2H).
Anal. (C18H10F6N2O2) C, H, N.
4-[3-Trifluoromethyl-5-(4-methylphenyl)-1H-pyrazol-1-yl]-
benzoic Acid (2).13 4-Hydrazinobenzoic acid HCl (52.8 mg, 0.28
mmol) and A (57.6 mg, 0.25 mmol) in acetonitrile (15 mL) were
reacted as described for 1 to give 2 (62 mg, 71%): 1H NMR (CD3-
OD) δ 2.27 (s, 3H), 6.85 (s, 1H), 7.08 (s, 4H), 7.24 (d, J ) 8.0,
2H), 8.03 (d, J ) 8.0, 2H). An analytical sample was further purified
with preparative LC to get 99.9% purity. Anal. (C18H13F3N2O2) C,
H, N.
4-[5-(4-Methylphenyl)-3-phenyl-1H-pyrazol-1-yl]-benzene-
sulfonamide (3). Compounds D (315 mg, 1.40 mmol) and G (334
mg, 1.40 mmol) were added to a 10 mL microwave reaction vessel
containing a stirrer, EtOH (5 mL), and 3 drops of acetic acid. The
vessel was sealed and the mixture was heated in the microwave
for 20 min at 180 °C and then allowed to cool to room temperature.
The reaction mixture was partitioned between water (50 mL) and
EtOAc (100 mL), and the aqueous layer was extracted with EtOAc
(3 × 100 mL). The combined extracts were dried over MgSO4,
filtered, and concentrated under vacuum. The crude product was
purified by flash chromatography (EtOAc/hexane 5:5) to obtain 3
(117 mg, 21%): 1H NMR (CDCl3) δ 2.39 (s, 3H), 6.81 (s, 1H),
7.18 - 7.92 (m, 13H). Anal. (C22H19N3O2S) C, H, N.
Synthesis. General. Unless otherwise noted, all solvents,
chemicals, and reagents were obtained commercially and used
without purification. Reactions were routinely monitored by thin-
layer chromatography (TLC) on silica-plated aluminum sheets
1
(silica gel 60 F254, E. Merck), detecting spots by UV light. H
and 13C NMR spectra were obtained on a JEOL JNM-EX 400
spectrometer at 400 and 100 MHz, respectively, in CDCl3 or CD3-
OD solutions, as indicated. Peak positions are given in parts per
million (δ) downfield from tetramethylsilane as internal standard,
and J values are given in Hz. Melting points were measured in a
Bu¨chi Melting Point B-540 apparatus and are uncorrected. Chro-
matography was performed on silica (silica gel 60 (0.040-0.063
mm), E. Merck) using flash chromatography. Preparative high
performance liquid chromatography (HPLC) was performed on a
Kromasil-C8 analytical column (250 mm × 4.6 mm i.d., 10 µm,
Phenomenex, England, U.K.). The mobile phase consisted of 0.1%
formic acid in acetonitrile/water (60/40). The microwave reactions
were carried out in a Biotage Initiator instrument. Elemental
analyses were performed at H. Kolbe Mikroanalytisches laborato-
rium, Mu¨lheim an der Ruhr, Germany, and were within (0.2% of
the theoretical values for C, H, and N.
4,4,4-Trifluoro-1-(4-methylphenyl)butane-1,3-dione (A).28 4-M-
ethyl-acetophenone (360 µL, 2.7 mmol) was dissolved in dry THF
(15 mL) under nitrogen atmosphere and NaH (60% in mineral oil,
130 mg, 3.24 mmol) was added in three lots maintaining the
temperature between -5 and 0 °C. After stirring at this temperature
for 30 min, ethyl trifluoroacetate (390 µL, 3.24 mmol) was added
via syringe and the reaction mixture was allowed to stir at ambient
temperature for 5 h. The reaction mixture was poured into ice water,
acidified with HCl (2 M), and extracted with ethyl acetate (2 × 50
mL). The combined organic layer was washed with water (2 × 50
mL), dried, and concentrated. The solid residue was washed with
hexane and dried under high vacuum to provide a solid mass that
was redissolved in dichloromethane and dried under high vacuum
1
to give the diketone A (590 mg, 95%). H NMR (CDCl3) δ 7.84
(d, J ) 8.2 Hz, 2H), 7.30 (d, J ) 8.0 Hz, 2H), 6.54 (s, 1H), 2.43
(s, 3H).
4,4,4-Trifluoro-1-(4-trifluoromethylphenyl)butane-1,3-di-
one (B).28 4-Trifluoromethyl-acetophenone (813 mg, 4.3 mmol) was
reacted with NaH (60% in mineral oil, 216 mg, 5.4 mmol) and
ethyl trifluoroacetate (643 µL, 5.4 mmol) as described above for
A. The crude product (1.29 g, 100%) was used without further
purification.
4-[5-(4-Methylphenyl)-3-trifluoromethyl-1H-pyrazol-1-yl]-1-
methylthio-benzene (4).13 Compounds C (515 mg, 2.70 mmol)
and A (564 mg, 2.45 mmol) in EtOH (15 mL) were reacted as
described for 3 to give 4 (275 mg, 32%): 1H NMR (CDCl3) δ
2.36 (s, 3H), 2.49 (s, 3H), 6.71 (s, 1H), 7.13 (d, J ) 4.0, 4H), 7.22
(d, J ) 4.0, 4H). Anal. (C18H15F3N2S) C, H, N.
[4-Methylthio-phenyl]hydrazine Hydrochloride (C).29 Aque-
ous HCl (37%, 8 mL) and 4-methylthio-aniline (490 µL, 4.0 mmol)
were added to a 50 mL three-necked round bottomed flask. The
slurry was heated to 120 °C until a solution was formed. The
solution was cooled down to -5 °C and a solution of NaNO2 (290
mg, 4.2 mmol) in H2O (2 mL) was added dropwise at -5 °C. The
mixture was stirred at -5 °C for 15 min and was then added at
-20 °C to a solution of SnCl2 × 2H2O (2.7 g, 12.0 mmol) in
aqueous HCl (37%) (3 mL). After stirring at -20 °C for 25 min
and at -5 °C for another 25 min, the precipitate was collected by
filtration and successively washed with cool H2O and Et2O to give
C (574 mg, 75%).
4-[5-(4-Methylphenyl)-3-trifluoromethyl-1H-pyrazol-1-yl]-1-
methylsulfinyl-benzene (5).20 A mixture of 4 (60.6 mg, 0.17
mmol), TBHP (6 M in nonane, 58.3 µL, 0.35 mmol), and CSA
(4.0 mg, 0.017 mmol) in dichloromethane was stirred at RT for 4
h. The reaction was monitored by TLC. When completed, the
solution was directly poured onto the top of a silica gel column
and eluted with hexane/ethyl acetate (7:3) to give pure 5 (64 mg,
87%): 1H NMR (CDCl3) δ 2.36 (s, 3H), 2.72 (s, 3H), 6.73 (s,
1H), 7.10 (d, J ) 8.0, 2H), 7.15(d, J ) 8.0, 2H), 7.48 (d, J ) 8.0,
2H), 7.64 (d, J ) 8.0, 2H). Anal. (C18H15F3N2OS) C, H, N.
4-[5-(4-Methylphenyl)-3-trifluoromethyl-1H-pyrazol-1-yl]-1-
methylsulfonyl-benzene (6).21 Compound 4 (45.3 mg, 0.13 mmol)
was dissolved in methanol (10 mL) and cooled to 0 °C. To this a
solution of oxone (49.5%, 240 mg, 0.39 mmol) in H2O (5 mL)
was added. The resulting cloudy slurry was stirred for 4 h at room
temperature, diluted with water (30 mL), and extracted with dich-
loromethane (3 × 50 mL). The combined organic layers were
washed with H2O (50 mL) and brine (50 mL), dried over MgSO4,
filtered, and concentrated under vacuum to give 6 (49 mg, 99%)
4-Sulfamoylphenyl Hydrazine Hydrochloride (D).29 Aqueous
HCl (37%, 12 mL), sulfanilamide (1.16 g, 6.73 mmol), and NaNO2
(489 mg, 7.09 mmol) in H2O (3 mL) were reacted as described for
C to afford pure D (1.29 g, 86%).
2-(4-Hydrazinophenyl) Acetic Acid Hydrochloride (E).29
Aqueous HCl (37%, 10 mL), 4-aminophenylacetic acid (1.10 g,
6.70 mmol), and SnCl2‚2H2O (4.55 g, 20.2 mmol) were reacted as

CYP2C9 Structure-Metabolism Relationships
Journal of Medicinal Chemistry, 2007, Vol. 50, No. 18 4451
as a white solid: 1H NMR (CDCl3) δ 2.38 (s, 3H), 3.06 (s, 3H),
6.75 (s, 1H), 7.11 (d, J ) 8.0, 2H), 7.18 (d, J ) 8.0, 2H), 7.53 (d,
J ) 8.0, 2H), 7.93 (d, J ) 8.0, 2H). Anal. (C18H15F3N2O2S) C, H, N.
Ethyl 4-[5-(4-Trifluoromethylphenyl)-3-trifluoromethyl-1H-
pyrazol-1-yl]-benzoate (7).13 4-Hydrazinobenzoic acid hydrochlo-
ride (181 mg, 0.96 mmol) and B (247 mg, 0.87 mmol) in 15 mL
of EtOH were reacted as described for MAX. Purification with flash
chromatography (MeOH/EtOAc/hexane 1:10:89) gave 7 (183 mg,
58%): 1H NMR (CDCl3) δ 1.40 (t, J ) 8.0 3H), 4.39 (q, J ) 8.0,
2H), 6.83 (s, 1H), 7.34-7.38 (m, 4H), 7.61 (d, J ) 8.0, 2H), 8.06
(d, J ) 8.0, 2H). Anal. (C20H14F6N2O2) C, H, N.
Ethyl 4-[5-(4-Methylphenyl)-3-trifluoromethyl-1H-pyrazol-
1-yl]-benzoate (8).13 4-Hydrazinobenzoic acid hydrochloride (181
mg, 0.96 mmol) and A (200 mg, 0.87 mmol) in 15 mL EtOH were
reacted as described for 7 to afford 8 (144 mg, 48%): 1H NMR
(CDCl3) δ 1.40 (t, J ) 8.0, 3H), 2.36 (s, 3H), 4.38 (q, J ) 8.0,
2H), 6.73 (s, 1H), 7.10 (d, J ) 8.0, 2H), 7.15 (d, J ) 8.0, 2H),
7.39 (d, J ) 8.0, 2H), 8.03 (d, J ) 8.0, 2H). Anal. (C20H17F3N2O2)
C, H, N.
4-[5-(4-Trifluoromethylphenyl)-3-trifluoromethyl-1H-pyrazol-
1-yl]-benzenesulfonamide (9). 4-Sulfamoylphenylhydrazine HCl
(399 mg, 1.71 mmol) was added to a 20 mL microwave reaction
vessel containing a stirrer and a solution of B (440 mg, 1.55 mmol)
in EtOH (15 mL). The vessel was sealed and the mixture was heated
in the microwave for 20 min at 180 °C and then allowed to cool to
room temperature. The reaction was concentrated under vacuum
and the residue was taken up in EtOAc (200 mL), washed with
water (50 mL × 4) and brine (50 mL × 2), dried over MgSO4,
filtered, and concentrated under vacuum. Purification with flash
chromatography (EtOAc/hexane 1:9) afforded 9 (383 mg, 57%
yield): 1H NMR (CDCl3) δ 6.85 (s, 1H), 7.38 (d, J ) 8.0, 2H),
7.47 (d, J ) 8.0, 2H), 7.66 (d, J ) 8.0, 2H), 7.96 (d, J ) 8.0, 2H).
Anal. (C17H11F6N3O2S) C, H, N.
1,5-Bis-(4-methylphenyl)-3-trifluoromethyl-pyrazole (10). p-
Tolylhydrazine HCl (262 mg, 1.65 mmol) was added to a 20 mL
microwave reaction vessel containing a stirrer and solution of A
(345 mg, 1.50 mmol) in EtOH (15 mL). The vessel was sealed
and the mixture was heated in the microwave for 10 min at 180 °C
and then allowed to cool to room temperature. The reaction was
concentrated under vacuum and the residue was taken up in EtOAc
(200 mL), washed with water (2 × 100 mL) and brine (2 × 100
mL), dried over MgSO4, filtered, and concentrated under vacuum.
Purification with flash chromatography (dichloromethane/hexane
2:8) gave 10 (272 mg, 53%): 1H NMR (CDCl3) δ 2.36 (d, J )
8.0, 6H), 6.70 (s, 1H), 7.12 (d, J ) 8.0, 4H), 7.17 (q, J ) 8.0, 4H).
Anal. (C18H15F3N2) C, H, N.
dropwise at 50 °C to a solution of sodium ethoxide (made from
275 mg, 12.0 mmol sodium) in ethanol (10 mL). The mixture was
heated at reflux for 2 h. After cooling, the reaction mixture was
poured into water (40 mL), acidified with HCl (37%, 1 mL), and
extracted with diethyl ether (3 × 100 mL). The combined organic
layer was washed with brine (30 mL), dried over MgSO4, filtered,
and concentrated under vacuum. The crude product was purified
with flash chromatography (EtOAc/hexane 3:7) to give 13a (1.33
g, 95%).
Ethyl 5-(4-Methylphenyl)-1-(4-sulfamoylphenyl)-1H-pyrazole-
3-carboxylate (13b). 4-Sulfamoylphenylhydrazine HCl (351 mg,
1.50 mmol) was added to a 20 mL microwave reaction vessel
containing a stirrer and a solution of 13a (436 mg, 1.50 mmol) in
EtOH (20 mL). The vessel was sealed and the mixture was heated
in the microwave for 9 min at 180 °C and then allowed to cool to
room temperature. The reaction was concentrated under vacuum
and the residue was taken up in EtOAc (200 mL), washed with
water (2 × 100 mL) and brine (2 × 100 mL), dried over MgSO4,
filtered, and concentrated under vacuum to give 13b (566 mg, 86%).
5-(4-Methylphenyl)-1-(4-sulfamoylphenyl)-1H-pyrazol-3-car-
boxylic acid (13c). Compound 13b (538 mg, 1.40 mmol) was added
to a stirred solution of THF (50 mL), MeOH (50 mL), and LiOH
(2 M, 50 mL) in a 250 mL round-bottomed flask and stirred for 15
h. NaOH (1 M, 200 mL) was added, and the mixture was extracted
with EtOAc (200 mL). The aqueous phase was acidified with concd
HCl (38 mL), pH 1.0, and extracted with EtOAc (300 mL), dried
over MgSO4, filtered, and concentrated under vacuum to give 13c
(466 mg, 87%): 1H NMR (CD3OD) δ 2.38 (s, 3H), 7.05 (s, 1H),
7.21 (q, J ) 8.0, 4H), 7.53 (d, J ) 8.0, 2H), 7.96 (d, J ) 8.0, 2H).
Anal. (C17H15N3O4S) C, H, N.
In Vitro Assay for Recombinant CYP2C9. Chemicals. The
chemicals used in the assays were NADPH purchased from Sigma
Chemical Co. (St. Louis, MO). Celecoxib was purchased from
AApin Chemicals Limited (Abingdon, U.K.). Tris-hydroxymethyl-
aminomethane was purchased from ICN Biomedicals, Inc. (Irvine,
CA). K2HPO4 and KH2PO4 were purchased from Kebo Lab
(Stockholm, Sweden), and acetonitrile and formic acid were
purchased from Merck (Darmstadt, Germany). Human CYP2C9HR
Bactosomes expressing human CYP2C9 was purchased from
Cypex, Ltd. (Dundee, Scotland, U.K.).
Incubations. The total reaction volume was 100 µL, and the
experiments were performed in duplicates. Each reaction mixture
contained 5 pmol CYP2C9 (Dundee, Scotland, U.K.), 1 mM
NADPH, 0.1 mM KPO4 buffer, pH 7.4, and 1 µM test compound.
Test compounds were dissolved in acetonitrile, giving a final assay
concentration of solvent of 2.5%. Time point samples were taken
at 0 and 30 min. The reactions were started by the addition of
NADPH after a preincubation of 10 min at 37 °C. Ice-cold
acetonitrile (200 µL) was added to stop the reaction. After
centrifugation at 4000 g for 20 min at 4 °C, 10 µL of supernatant
was injected into the liquid chromatography/tandem mass spec-
trometry system.
2-[4-(5-(4-Trifluoromethylphenyl)-3-trifluoromethyl-1H-pyra-
zol-1-yl)phenyl]-acetic Acid (11). Compound E (346 mg, 1.71
mmol) was added to a 20 mL microwave reaction vessel containing
a stirrer and a solution of B (440 mg, 1.55 mmol) in acetonitrile
(20 mL). The vessel was sealed and the mixture was heated in the
microwave for 3 min at 200 °C and then allowed to cool to room
temperature. The reaction was concentrated under vacuum and the
residue was taken up in EtOAc (250 mL), washed with water (100
mL × 4) and brine (100 mL × 2), dried over MgSO4, filtered, and
concentrated under vacuum. Purification with flash chromatography
LC/MS/MS. The HPLC system used included a HP 1100 serial
LC pump, a column oven (Agilent Technologies Deutschland,
Waldbronn, Germany), and a CTC HTS auto sampler (CTC
Analytics, Zwingen, Switzerland). Chromatography was performed
on a HyPURITY C18 analytical column (50 mm × 2.1 mm i.d., 5
µm, ThermoQuest, Runcorn, U.K.) with a HyPURITY C18 guard
column (10 mm × 2.1 mm, 5 µm). The mobile phase consisted of
(A) 0.1% formic acid in water and (B) 0.1% formic acid in
acetonitrile. For the metabolism study, a programmed linear gradient
started at 30% B and increased to 90% B over 2.0 min, stayed
there for 1.5 min, and then returned to 30% for 1.5 min to equilibrate
the column at a flow rate of 0.75 mL/min. For compound 13c, the
time gradient was the same as for the other compounds, but the
gradient started with 10% of solvent B instead of 30%.
1
(dichloromethane/methanol 98:2) gave 11 (204 mg, 29%): H NMR
(CD3OD) δ 3.68 (s, 2H), 7.07 (s, 1H), 7.31 (d, J ) 8.0, 2H), 7.43
(d, J ) 8.0, 2H), 7.51 (d, J ) 8.0, 2H), 7.68 (d, J ) 8.0, 2H).
Anal. (C19H12F6N2O2) C, H, N.
2-[4-(5-(4-Methylphenyl)-3-trifluoromethyl-1H-pyrazol-1-yl)-
phenyl]-acetic Acid (12). Compounds E (346 mg, 1.71 mmol) and
A (357 mg, 1.55 mmol) in acetonitrile (20 mL) were reacted as
described for 11 to give 12 (311 mg, 54%): 1H NMR (CD3OD) δ
2.39 (s, 3H), 3.69 (s, 2H), 6.89 (s, 1H), 7.18 (s, 4H), 7,29 (d, J )
8.0, 2H), 7,38 (d, J ) 8.0, 2H). Anal. (C19H15F3N2O2) C, H, N.
5-(4-Methylphenyl)-1-(4-sulfamoylphenyl)-1H-pyrazol-3-car-
boxylic Acid (13c). Synthesis of 13c was a three-step process:
Ethyl 2-Hydroxy-4-oxo-4-(4-methylphenyl)-2-butenoate (13a).22
A solution of diethyl oxalate (1627 µL, 12.0 mmol) and 4-methy-
lacetophenone (801 µL, 6.0 mmol) in ethanol (5 mL) was added
The mass spectrometrical analyses were performed using an
API4000 instrument (Applied Biosystems/MDS Sciex, CA). The
mass spectrometer was operated in both positive and negative ion
mode. The tuning parameters were optimized for the compounds
by infusing a 50% acetonitrile/H2O solution containing 1 µM of

4452 Journal of Medicinal Chemistry, 2007, Vol. 50, No. 18
Ahlstro¨m et al.
(12) Kurumbail, R. G.; Stevens, A. M.; Gierse, J. K.; McDonald, J. J.;
Stegeman, R. A.; et al. Structural basis for selective inhibition of
cyclooxygenase-2 by anti-inflammatory agents. Nature 1996, 384,
644-648 [erratum appears in Nature 1997 Feb 6;385(6616):555].
(13) Penning, T. D.; Talley, J. J.; Bertenshaw, S. R.; Carter, J. S.; Collins,
P. W.; et al. Synthesis and biological evaluation of the 1,5-
diarylpyrazole class of cyclooxygenase-2 inhibitors: Identification
of 4-[5-(4-methylphenyl)-3-(trifluoromethyl)-1H-pyrazol-1-yl]benze
nesulfonamide (SC-58635, celecoxib). J. Med. Chem. 1997, 40,
1347-1365.
(14) Zamora, I.; Afzelius, L.; Cruciani, G. Predicting drug metabolism:
A site of metabolism prediction tool applied to the cytochrome P450
2C9. J. Med. Chem. 2003, 46, 2313-2324.
(15) Cruciani, G.; Carosati, E.; De Boeck, B.; Ethirajulu, K.; Mackie, C.;
et al. MetaSite: Understanding metabolism in human cytochromes
from the perspective of the chemist. J. Med. Chem. 2005, 48, 6970-
6979.
analyte at a flow rate of 10 µL/min into the mobile phase (0.75
mL/min) using a T connection. The scan parameters for multiple
reaction monitoring (MRM) of the 14 compounds investigated can
be found in the Supporting Information. Instrument control, data
acquisition, and data evaluation were performed using Analyst 1.4
software (Applied Biosystems/MDS sciex, CA).
In Vitro Assay for Recombinant COX-1 and COX-2. The
assays were performed according to Glaser et al.31 by CEREP (Celle
l’Evescault, France).32 The assays evaluate the effects of com-
pounds, at 100 µM, on the activity of the human COX-1/COX-2
quantified by measuring the formation of prostaglandin E2 (PGE2)
from arachidonic acid using a recombinant enzyme isolated from
Sf-9 cells. The results are expressed as a percent inhibition of the
control enzyme activity.
(16) Goodford, P. J. A computational procedure for determining energeti-
cally favorable binding sites on biologically important macromol-
ecules. J. Med. Chem. 1985, 28, 849-857.
Acknowledgment. We thank Dr. Neal Castagnoli for
valuable discussions and critical reading of the manuscript.
(17) www.moldiscovery.com.
Supporting Information Available: Scan parameters for
multiple reaction monitoring (MRM) of the 14 compounds inves-
tigated and elemental analyses of all synthesized compounds. This
material is available free of charge via the Internet at http://
pubs.acs.org.
(18) Bumgardner, C. L.; Sloop, J. C. Ring-fluorinated pyrazoles. J.
Fluorine Chem. 1992, 56, 141-146.
(19) Sloop, J. C.; Bumgardner, C. L.; Loehle, W. D. Synthesis of
fluorinated heterocycles. J. Fluorine Chem. 2002, 118, 135-147.
(20) Bonadies, F.; De Angelis, F.; Locati, L.; Scettri, A. A convenient
acid-catalyzed oxidation of sulfides to sulfoxides by t-butyl hydro-
peroxide. Tetrahedron Lett. 1996, 37, 7129-7130.
(21) Trost, B. M.; Curran, D. P. Chemoselective oxidation of sulfides to
sulfones with potassium hydrogen persulfate. Tetrahedron Lett. 1981,
22, 1287-1290.
(22) Pommery, N.; Taverne, T.; Telliez, A.; Goossens, L.; Charlier, C.;
et al. New COX-2/5-LOX inhibitors: Apoptosis-inducing agents
potentially useful in prostate cancer chemotherapy. J. Med. Chem.
2004, 47, 6195-6206.
References
(1) Mancy, A.; Broto, P.; Dijols, S.; Dansette, P. M.; Mansuy, D. The
substrate binding site of human liver cytochrome P450 2C9: An
approach using designed tienilic acid derivatives and molecular
modeling. Biochemistry 1995, 34, 10365-10375.
(2) Miners, J. O.; Birkett, D. J. Cytochrome P4502C9: An enzyme of
major importance in human drug metabolism. Br. J. Clin. Pharmacol.
1998, 45, 525-538.
(3) Yamazaki, H.; Inoue, K.; Chiba, K.; Ozawa, N.; Kawai, T. et al.
Comparative studies on the catalytic roles of cytochrome P450 2C9
and its Cys- and Leu-variants in the oxidation of warfarin, flurbi-
profen, and diclofenac by human liver microsomes. Biochem.
Pharmacol. 1998, 56, 243-251.
(4) Wester, M. R.; Yano, J. K.; Schoch, G. A.; Yang, C.; Griffin, K. J.;
et al. The structure of human cytochrome P450 2C9 complexed with
flurbiprofen at 2.0 A resolution. J. Biol. Chem. 2004, 279, 35630-
35637.
(5) Williams, P. A.; Cosme, J.; Ward, A.; Angove, H. C.; Matak,
Vinkovic, D.; et al. Crystal structure of human cytochrome P450
2C9 with bound warfarin. Nature 2003, 424, 464-468.
(6) Sandberg, M.; Yasar, U.; Stromberg, P.; Hoog, J. O.; Eliasson, E.
Oxidation of celecoxib by polymorphic cytochrome P450 2C9 and
alcohol dehydrogenase. Br. J. Clin. Pharmacol. 2002, 54, 423-429.
(7) Tang, C.; Shou, M.; Mei, Q.; Rushmore, T. H.; Rodrigues, A. D.
Major role of human liver microsomal cytochrome P450 2C9
(CYP2C9) in the oxidative metabolism of celecoxib, a novel
cyclooxygenase-II inhibitor. J. Pharmacol. Exp. Ther. 2000, 293,
453-459.
(8) Fu, J. Y.; Masferrer, J. L.; Seibert, K.; Raz, A.; Needleman, P. The
induction and suppression of prostaglandin H2 synthase (cyclooxy-
genase) in human monocytes. J. Biol. Chem. 1990, 265, 16737-
16740.
(9) Smith, W. L.; Garavito, R. M.; DeWitt, D. L. Prostaglandin
endoperoxide H synthases (cyclooxygenases)-1 and -2. J. Biol. Chem.
1996, 271, 33157-33160.
(23) Dayal, B.; Salen, G.; Toome, B.; Tint, G. S.; Shefer, S.; et al. Lithium
hydroxide/aqueous methanol: mild reagent for the hydrolysis of bile
acid methyl esters. Steroids 1990, 55, 233-237.
(24) Palomer, A.; Cabre, F.; Pascual, J.; Campos, J.; Trujillo, M. A.; et
al. Identification of novel cyclooxygenase-2 selective inhibitors using
pharmacophore models. J. Med. Chem. 2002, 45, 1402-1411.
(25) Afzelius, L.; Zamora, I.; Ridderstrom, M.; Andersson, T. B.; Karlen,
A.; et al. Competitive CYP2C9 inhibitors: Enzyme inhibition studies,
protein homology modeling, and three-dimensional quantitative
structure-activity relationship analysis. Mol. Pharmacol. 2001, 59,
909-919.
(26) Williams, P. A.; Cosme, J.; Sridhar, V.; Johnson, E. F.; McRee, D.
E. Mammalian microsomal cytochrome P450 monooxygenase: Struc-
tural adaptations for membrane binding and functional diversity. Mol.
Cell 2000, 5, 121-131.
(27) Goodford, P. Multivariate characterization of molecules for QSAR
analysis. J. Chemom. 1996, 10, 107-117.
(28) Singh, S. K.; Reddy, P. G.; Rao, K. S.; Lohray, B. B.; Misra, P.; et
al. Polar substitutions in the benzenesulfonamide ring of celecoxib
afford a potent 1,5-diarylpyrazole class of COX-2 inhibitors. Bioorg.
Med. Chem. Lett. 2004, 14, 499-504.
(29) Brodfuehrer, P. R.; Chen, B. C.; Sattelberg, T. R., Sr.; Smith, P. R.;
Reddy, J. P.; et al. An efficient Fischer indole synthesis of avitriptan,
a potent 5-HT(1D) receptor agonist. J. Org. Chem. 1997, 62, 9192-
9202.
(30) Lawrence, N. J.; Rennison, D.; McGown, A. T.; Ducki, S.; Gul, L.
A.; et al. Linked parallel synthesis and MTT bioassay screening of
substituted chalcones. J. Comb. Chem. 2001, 3, 421-426.
(31) Glaser, K.; Sung, M. L.; O’Neill, K.; Belfast, M.; Hartman, D.; et
al. Etodolac selectively inhibits human prostaglandin G/H synthase
2 (PGHS-2) versus human PGHS-1. Eur. J. Pharmacol. 1995, 281,
107-111.
(10) Futaki, N.; Yoshikawa, K.; Hamasaka, Y.; Arai, I.; Higuchi, S.; et
al. NS-398, a novel nonsteroidal anti-inflammatory drug with potent
analgesic and antipyretic effects, which causes minimal stomach
lesions. Gen. Pharmacol. 1993, 24, 105-110.
(11) Dogne, J. M.; Hanson, J.; Supuran, C.; Pratico, D. Coxibs and
cardiovascular side-effects: From lights to shadow. Curr. Pharm.
Des. 2006, 12, 971-975.
(32) www.cerep.com.
JM0705096