5-Heteroatom substituted pyrazoles as canine COX-2 inhibitors. Part III: Molecular modeling studies on binding contribution of 1-(5-methylsulfonyl)pyrid-2-yl and 4-nitrile

Article abstract of DOI:10.1016/j.bmcl.2006.11.026

The structure-activity relationship toward canine COX-1 and COX-2 in vitro whole blood activity of 4-hydrogen versus 4-cyano substituted 5-aryl or 5-heteroatom substituted N-phenyl versus N-2-pyridyl sulfone pyrazoles is discussed. The differences between

Full text of DOI:10.1016/j.bmcl.2006.11.026

Bioorganic & Medicinal Chemistry Letters 17 (2007) 1067–1072
5-Heteroatom substituted pyrazoles as canine COX-2 inhibitors.
Part III: Molecular modeling studies on binding contribution
of 1-(5-methylsulfonyl)pyrid-2-yl and 4-nitrile
Subas M. Sakya,a,* Xinjun Hou,b Martha L. Minich,a Bryson Rast,a Andrei Shavnya,a
Kristin M. L. DeMello,a Hengmiao Cheng,a Jin Li,a Burton H. Jaynes,a Donald W. Mann,a
Carol F. Petras,a Scott B. Seibela and Michelle L. Havena
aVeterinary Medicine Research and Development, Pfizer Inc., Groton, CT 06340, USA
bExploratory Medicinal Sciences, Pfizer Global Research and Development Pfizer Inc., Groton, CT 06340, USA
Received 4 October 2006; revised 7 November 2006; accepted 8 November 2006
Available online 15 November 2006
Abstract—The structure–activity relationship toward canine COX-1 and COX-2 in vitro whole blood activity of 4-hydrogen versus
4-cyano substituted 5-aryl or 5-heteroatom substituted N-phenyl versus N-2-pyridyl sulfone pyrazoles is discussed. The differences
between the pairs of compounds with the 4-nitrile pyrazole derivatives having substantially improved in vitro activity are highlighted
for both COX-2 and COX-1. This difference in activity may be due to the contribution of the hydrogen bond of the 4-cyano group
with Ser 530 as shown by our molecular modeling studies. In addition, our model suggests a potential contribution from hydrogen
bonding of the pyridyl nitrogen to Tyr 355 for the increased activity over the phenyl sulfone analogs.
ꢀ 2006 Elsevier Ltd. All rights reserved.
The cyclooxygenase (COX) enzymes, which catalyze the
first step in arachidonic acid metabolism,1 were identified
as the molecular targets of all nonsteroidal anti-inflam-
matory drugs (NSAIDs).2–4 COX-1, a constitutively ex-
pressed isoform, is found in platelets, kidneys, and in
the gastrointestinal tract, and is believed to be responsi-
ble for the homeostatic maintenance of the kidneys and
GI tract. The COX-2 enzyme is the inducible isoform
that is produced by various cell types upon exposure to
cytokines, mitogens, and endotoxins released during
injury.5 A recent discovery of the third COX isoform
(COX-3) enzyme primarily expressed in the brain and
the heart is thought to be the target for acetaminophen.6
The COX-2 enzyme, after being overexpressed at the site
of injury, is a catalyst for the production of the prosta-
glandins that result in inflammation and pain at the site.
Because COX-1 is involved in the maintenance of the GI
tract, NSAIDs that are inhibitors of both COX-2 and
COX-1 have been found to cause side effects associated
with gastrointestinal ulcers.7–10 Thus, it was thought that
a more selective COX-2 inhibitor would have reduced
gastrointestinal side effects.5
Research efforts in the discovery of COX-2-selective
agents have produced many classes of compounds hav-
ing the desired selectivity. Several marketed human
COꢁX-2-selective drugs, including celecoxib (Celeb-
rex ),11 (Fig. 1) for treating pain and inflammation
associated with arthritis have been shown to be well tol-
erated with decreased gastrointestinal (GI) side effects.12
Progressive degenerative joint disease, or osteoarthritis,
is the most common cause of chronic pain in dogs.13 It is
estimated that one out of every five adult dogs, or
approximately 8 million animals, has osteoarthritis, yet
nearly half (48%) of these patients are untreated.14 As
in humans, chronic use of NSAIDs in dogs is often aꢁsso-
ciated with GI side effects.15 Carprofen (Rimadyl )16
and deracoxib (DeramaxxTM),11,17 two marketed agents
for the treatment of inflammation and pain for dogs,
have moderate COX-2 selectivity. Firocoxib, with in-
creased selectivity for the canine COX-2 enzyme,16b
has also been recently approved for treatment in dogs.
None of these are approved in the US for use in cats
Keywords: Canine; Cyclooxygenase; COX-1; COX-2; Pyrazole; Het-
erocyle; Molecular modelling.
*
Corresponding author at present address: CNS Chemistry, PGRD.
Tel.: +1 860 715 0425; fax: +1 860 715 7312; e-mail:
subas.m.sakya@pfizer.com
0960-894X/$ - see front matter ꢀ 2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2006.11.026

1068
S. M. Sakya et al. / Bioorg. Med. Chem. Lett. 17 (2007) 1067–1072
2-pyridyl nitrogen as hydrogen bond acceptors leading
to improvements in activity in our in vitro assay.
SO2NH2
SO2NH2
CO2H
F
The synthesis of 5-phenyl pyrazoles 3a, 3c and 5-ethers 6
with 4-H has been disclosed in previous reports.19,20
Preparation of 5-aryl-4-cyano pyrazoles (3b, 3d, 3f, 3g)
was accomplished using Suzuki–Miyauri coupling con-
ditions (Scheme 1). The preparation of the 5-amin-
oalkyl-4-cyano substituted pyrazoles was done as
shown in Scheme 1. The substitution of the 5-chloro-4-
cyano pyrazoles 2 under mild base-mediated conditions
gave the 4-cyano-5-aminoalkyl and 5-ether pyrazoles (5
and 6).24,25 Reactions of the 4-formyl-5-amino-pyrazoles
4, obtained in an analogous manner to compounds 5
and 6, with hydroxyl amine hydrochloride in methanol,
in an attempt to obtain the oximes, resulted in decarb-
onylation by-products 5. The reactions were general en-
ough to obtain up to 60% of the 4-H by-products 5.
O
N
N
N
N
O
F3C
Ph
Ketoprofen
HF2C
Celecoxib
Deracoxib
OH
O
N
Cl
S
N
H
HO2C
N
N
S
H
O
O
Carprofen
Meloxicam
Figure 1. Structures of marketed COX-2 inhibitors.
Table 1 shows the activity comparison of the analogs
tested in the canine whole blood (CWB) COX-1 and
COX-2 assay.26 The compounds are grouped according
to the 5-substitution on the pyrazole with comparisons
for pain and inflammation. Meloxicam, a marginally
selective NSAID for canine COX-2, was recently ap-
proved in the US for use in cats.18
SO2Me
SO2Me
Our initial efforts to identify superior agents in this area
led to the identification of 5-aryl pyrazole 3c (Fig. 2)
which had enhanced canine COX-2 selectivity and in
vivo efficacy relative to carprofen.19 In addition, we have
disclosed the synthesis and SAR related to the alkyl
ether20 and amino pyrazoles21 that led to the identifica-
tion of the lead compound, 5c, which showed in vivo
efficacy in both canine and feline synovitis models. Dur-
ing our studies on the nitrile substituted pyrazoles, we
have consistently noticed the differences in activity be-
tween the 4-nitrile and 4-hydrogen (non-substituted) as
well as other substitutions at that position. Several liter-
ature disclosures have indicated the potential contribu-
tion of the nitrile group as a strong hydrogen-bond
acceptor in contributing to activity of compounds in
various drug discovery efforts.22 In addition, a recent
study was done to look at the specific contribution of
the 4-substituted pyrazoles in the COX-2 activity of
1,4,5-substituted pyrazoles.23 This study demonstrated
a potential contribution of the hydrogen bond acceptor
groups based on some modeling studies, but no in vitro
whole cell or enzyme activity was reported. We would
like to highlight the contribution of the 4-cyano and
X
N
X
N
R2
c
N
N
Cl
X = N
R1
R
R
CN
R = CF3 or CF2H
X = N or CH
3b, d, f, g
1 R1 = CHO
a, b
2 R1 = CN
d
SO2Me
SO2Me
X
N
X
N
e
N
N
R2
NR3R4
From 4
R1
R1
R
R
4 R1 = CHO
R2 = NR3R4
5 R1 = H
SO2Me
SO2Me
SO2Me
5 R1 = CN
R2 = NR3R4
N
N
N
N
N
N
6 R1 = CN
R2 = OR
O
N
N
N
N
OR
Scheme 1. Reagents and conditions: (a) NH2OHÆHCl, TFE, reflux,
2 h, >80% (b) Cl3CCOCl, Et3N, DCM, 0 ꢂC, 4–6 h, >90%; (c)
ArB(OH2), Pd(OAc)2, tris[2-(2-methoxyethoxy)ethyl]amine, 50–80%;
(d) For 4: Method A: R3R4NH, Et3N, DCE, 80 ꢂC; For 6: Method B:
ROH, KF, DMSO, RT to 80 ꢂC; For 5: Method A or Method B (at
RT); (e) 4, 1.2 equiv NH2OHÆHCl, MeOH, reflux, 8–60%.
R1
6 R = H
F3C
HF2C
CN
R
1
5c
3a R = CF
3
3c R = CF H
2
Figure 2. Canine COX-2, selective leads.

S. M. Sakya et al. / Bioorg. Med. Chem. Lett. 17 (2007) 1067–1072
Table 1. In vitro COX-1 and COX-2 inhibition data of 3, 5, and 6
1069
SO2Me
X
N
N
R2
R1
R
Compound
X
R
R1
H
R2
Ratio
COX 1/2 COX-1 COX-2
CWB* IC50 (lM) CWB* COX-2 % ClogP Kinetic
inh at 0.5 lM
aqueous sol
(lg/ml)
3a
3b
3c
3d
3e
3f
N
N
N
N
CF3
CF3
Phenyl
>122
48.5
155
>50
5.82
48
0.41+
0.12
0.31
<0.05
<0.05
0.17
0.06
>5
56.3+
74.3
62.7
98.3
97.4
96.8
86.4
À33.9
70.8
38.0
70.2
31.7
6.2
2.95
2.39
2.25
1.84
2.89
2.13
3.18
2.12
1.69
2.15
1.10
1.92
2.62
2.71
3.01
1.60
2.94
1.88
1.56
1.94
1.62
2.54
1.92
2.44
2.12
1.50
2.41
2.87
2.55
3.46
2.84
2.47
2.02
3.38
2.94
2.48
2.03
3.40
2.54
2.09
3.46
2.91
2.46
3.82
2.85
50–200
100–200
> 200
ND
CN Phenyl
H Phenyl
CF2H CN Phenyl
CF2H
>112
>232
131
5.6
CH CF2H CN Phenyl
CF2H CN 3,5-Difluorophenyl
CH CF2H CN 3,5-Difluorophenyl
11.61
22.34
>50
>5
ND
50–100
ND
N
3g
5a
5b
5c
5d
5e
5f
>833
—
CN Cis-2,5-Dimethylmorpholine >208
N
N
CF3
CF3
H
Cis-2,5-Dimethylmorpholine
100–200
>200
ND
>50
>0.5
35.97
ND
>0.5
<0.05
>50
3.02
27.8
>0.5
1.24
>0.5
1.88
>0.5
>0.5
ND
0.31
12.08
>0.5
ND
17.41
ND
>0.5
>50
1.87
ND
>0.5
ND
16.5
ND
ND
10.63
ND
>50
ND
ND
>0.5
0.24
>0.5
0.24
>0.5
>0.5
0.07
0.16
0.03
0.09
>0.5
0.23
>0.5
0.11
>0.5
>0.5
>0.5
0.12
0.08
>0.5
>0.5
0.20
>0.5
>0.5
4.07
0.11
>0.5
>0.5
>0.5
0.21
>0.5
>0.5
0.08
>0.5
19.64
<0.5
>0.5
>0.5
CH CF3
CF2H CN cis-2,5-Dimethylmorpholine
CH CF2H CN cis-2,5-Dimethylmorpholine
CN cis-2,5-Dimethylmorpholine
—
N
186
—
>200
50–200
ND
N
N
CF3
CF3
H
3-Methyl piperidine
—
5g
5h
5i
CN 3-Methyl piperidine
CN 3-Methyl piperidine
CF2H CN 3-Methyl piperidine
<1
>312
100
397
—
101.7
87.1
102.8
104.1
À2.3
78.7
5.7
100–200
ND
ND
CH CF3
N
5j
CH CF2H CN 3-Methyl piperidine
50–100
ND
ND
5k
5l
N
N
N
N
CF3
CF3
CF3
CF3
H Cyclobutyl amine
CN Cyclobutyl amine
5.39
—
5m
5n
5o
5p
5q
5r
5s
5t
H
Cyclopropyl methyl amine
ND
ND
CN Cyclopropyl methyl amine
CN Cyclopropyl methyl amine
CH CF2H CN Cyclopropyl methyl amine
17
—
79.3
1.1
CH CF3
ND
—
—
À39.4
21.3
79.4
100
>200
25–100
ND
N
N
N
CF3
CF3
H Cyclopentylamine
CN Cyclopentylamine
2.5
151
—
CF2H CN Cyclopentylamine
25–50
100–200
ND
CH CF2H CN Cyclopentylamine
À38.2
À6.3
84.0
13.8
À32.0
18.4
93.1
À10.5
25.4
12.1
86.8
À49.9
À77.4
97.1
À34.7
5.9
5u
5v
5w
5x
6a
6b
6c
6d
6e
6f
N
N
CF3
CF3
H
Neopentylamine
—
87
CN Neopentylamine
CN Neopentylamine
CH CF2H CN Neopentylamine
<25
ND
CH CF3
—
—
>200
50–200
<25
N
N
CF3
CF3
H
Isobutyloxy
CN Isobutyloxy
Isobutyloxy
CN Isobutyloxy
Cyclopentyloxy
CN Cyclopentyloxy
>12
17
CH CF3
CH CF3
H
—
—
ND
ND
N
N
CF3
CF3
H
—
78
50–200
ND
6g
6h
6i
CH CF3
H
H
Cyclopentyloxy
Cyclobutylmethyloxy
—
—
25–50
100–200
ND
N
N
CF3
CF3
CN Cyclobutylmethyloxy
133
—
6j
CH CF3
H
H
Cyclobutylmethyloxy
Neopentyloxy
25–50
100–200
ND
6k
6l
N
N
CF3
CF3
>2.5
—
CN Neopentyloxy
Neopentyloxy
CN n-Butyl-2-oxy
63.2
À30.6
14.7
6m
6n
CH CF3
CH CF3
H
—
—
50–100
50–100
* Canine whole blood (CWB) assay: run in duplicate or triplicate; +, average of many runs.
made between 4-H (unsubstituted) and 4-CN substitut-
ed pyrazoles and between N-(4-methylsulfonyl) phenyl
and/or N-(5-methylsulfonyl)pyrid-2-yl pyrazoles. The
3-position of the pyrazole is substituted with either
CF2H or CF3, generally found to be optimal for activi-
ty.11 Three types of 5-substituents; 5-aryls, 5-aminoalk-
yls, and 5-alkylethers, are highlighted in Table 1 with
results from WBC COX-2 assay at 0.5 lM dose and
titration data of the actives. (Actives: inhibition of
COX-2 > 50% at 0.5 lM.)

1070
S. M. Sakya et al. / Bioorg. Med. Chem. Lett. 17 (2007) 1067–1072
Within the N-pyridyl 5-phenyl substituted pyrazoles 3,
the COX-2 activity does not vary that widely between
compounds with 4-CN and 4-H pyrazoles, although
there is a definite trend toward improved activity (3–
6·) with the 4-CN substitution. However, there is a sig-
nificant improvement in activity against the COX-1 en-
zyme (3b, 3d vs 3a, 3c) as well. When comparing the
contribution of the N-pyridyl versus N-phenyl sulfone
to the activity within these 5-aryl pyrazoles, there does
not seem to be any difference in activity against the
COX-2 enzyme (3d, 3f vs 3e, 3g). Against the COX-1 en-
zyme, however, the N-pyridylsulfone has improved
activity compared to the N-phenyl sulfone analogs.
interactions between 2-pyridyl nitrogen and 4-cyano
nitrogen with appropriately situated hydrogen bond do-
nors in the active site. Since the improvement in activity
was seen for both the COX-2 and COX-1 activity, we
presume the site of interaction to be present in both en-
zymes. Although there are many molecular modeling
studies done with ligand interactions with COX-1 and
COX-2 enzymes, we have been unable to find any mod-
eling studies done with N-(5-methylsulfonylpyrid-2-yl)
groups.27 Studies with 4-cyano substituted pyrazoles
have been mentioned in the literature.23 Thus, we under-
took molecular modeling studies to determine if there
are some significant interactions with the 2-pyridyl and
the 4-cyano nitrogen to afford such potent compounds.
It is interesting to note that a recent molecular modeling
study of many diaryl pyrazoles has indicated electrostat-
ic and hydrogen bond interactions to be an important
determinant for COX-2 selectivity along with the right
conformation.27d
For both the 5-alkylamino, and 5-alkylether substituted
N-(2-pyridyl methylsulfone) pyrazoles, 5 and 6, the 4-
CN substitution provides active compounds (5b, 5g ,
5l, 5r, 5v, 6b, 6f, 6i, 6l) over 4-H compounds that are
either inactive (5a, 5f, 5k, 5q, 5u, 6a, 6e, 6h, 6k) or weak-
ly active (6a). As in the case of the 5-aryl pyrazoles,
many of these 5-heteroatom substituted 4-CN pyrazoles
show enhanced activity in the COX-1 assay as well (5g,
5l, 5n, 5r, 5v, 6b) relative to its 4-H pyrazoles. The one
exception is for 5b where no COX-1 activity is seen.
Among others that do show weak activity, comparison
is difficult to make because no COX-1 titration data
are available for the parent 4-H analogs (6f, 6i). In the
case of N-phenylmethylsulfonyl pyrazoles, the majority
of the compounds with either 4-H or 4-CN substituted
pyrazoles are inactive (5e, 5o, 5p, 5t, 5w, 5x, 6c, 6d,
6g, 6j, 6n, 6o) or weakly active (5c (38% inh at
0.5 lM), 5e (31.7% inh at 0.5lM)) in the COX-2 assay.
Only the 5-(3-methylpiperidine) substituted N-phen-
ylsulfonyl and N-(2-pyridylmethylsulfonyl) pyrazoles
with 4-CN group are active in the COX-2 assay. These
5-(3-methylpiperidine)-N-phenylmethylsulfonyl pyraz-
oles, however, are inactive (5h) or weakly active (5j) in
the COX-1 assay, which provide for greater selectivity
than the corresponding N-pyridyl analogs.
Canine and human COX-1 and COX-2 are highly simi-
lar (92.7% and 90.1% identical, respectively), and the ac-
tive site residues of canine COX-1 and COX-2 are
identical to human and rat enzymes.28 Therefore, the
available murine COX-2 X-ray co-crystal structure of
SC-55827a was used as the starting point for molecular
modeling. Manual docking followed by molecular
mechanics minimization29 of 3b shows very similar inter-
actions as in the case of SC-558 where the 5-phenyl is
bound in the hydrophobic pocket formed by Phe 381,
Leu 384, Tyr 385, Trp 387, Phe 513, and Ser 530. The
backbone atoms of Gly 526 and Ala 527 are stacking
against the phenyl ring and the N-2-pyridylmethylsulfo-
nyl group is bound to the side pocket of COX-2, while
the trifluoromethyl group is interacting with Arg 120.
These interactions are similar in COX-1 except that
the side pocket has Val 523 instead of Ile where the large
Ile group potentially forms a smaller pocket in COX-1,
thus providing for COX-2 selectivity.
Thus, among all active compounds with 5-aryl or 5-het-
eroatom substituted-N-(2-pyridylmethylsulfone) pyraz-
oles, the substitution at the 4- position of the pyrazole
ring with the cyano substituent confers substantial
improvement in activity over its corresponding 4-H
analogs. Depending on the binding contribution of the
non-aryl 5-substituents, the contribution of the pyridyl
nitrogen may not be required for activity since some
of the N-phenyl methylsulfonyl compounds are active
with a 4-cyano group present. Only in the case of 5-
phenyl substituted pyrazoles, the 4-CN or 2-pyridyl
group does not seem to be necessary for COX-2 activity.
Our molecular modeling revealed several unique inter-
actions of N-2-pyridyl 5-alkylether/alkylamino 4-cyano
pyrazoles with the COX-2 active site that would ex-
plain the observed structure–activity relationships. In
contrast to a previous report23, our models show that
the cyano group is accepting a hydrogen bond from
Ser 530 side-chain hydroxyl group. The observed dis-
˚
tances of 3.0–3.5 A were seen depending on substitu-
tion in 5-position, consistent with observation of
nitrile hydrogen bonds.31 In the case of 3b, phenyl fits
well into the hydrophobic pocket and Ser 530 sidechain
rotated approximately 20ꢂ to form a weak hydrogen
bond to the 4-cyano nitrogen. With smaller 5-alkylami-
no and 5-alkylether substitutions, like compound 6b,
the energy-minimized structure shows that the pyrazole
Although our biological activities are based on canine
whole blood assay, we strongly believe the contributions
shown by the 4-cyano and N-2-pyridyl nitrogen for
activity are based on COX-2 and COX-1 active site
interactions and not just related to physical property dif-
ferences in whole blood. There are no clear indications
that the difference in activity is due to solubility or
polarity of the molecules, based on their ClogP or sol-
ubility data (Table 1). Thus, the in vitro biological re-
sults indicate a strong possibility of hydrogen bond
˚
ring shifts deeper by approximately 0.5 A into the
hydrophobic pocket. With a small rotation of Ser 530
hydroxyl (ꢀ15ꢂ), a good hydrogen bond is formed with
˚
Ser 530 with a nitrogen to oxygen distance of 3.0 A, as
shown in Figure 3. We hypothesize that this is one of
the factors in the contribution of 4-cyano versus 4-H
substitution-related activity differences (37X for 6b vs
6a and 6h vs 6i).

S. M. Sakya et al. / Bioorg. Med. Chem. Lett. 17 (2007) 1067–1072
1071
Figure 4. Model of 5g (light blue ball and stick) shows interaction of
methyl of 3-methyl piperidine that fits well into a small pocket near
Phe 381, Tyr 385, and Ser 530. The model of 6b is drawn in magenta to
show slight shifting of 6b model in the active site with respect to the
binding model of 5g.
Figure 3. Energy-minimized model of 6b in the COX-2 active site.
Unique hydrogen bonds are 4-cyano nitrogen to Ser 530 and N-pyridyl
nitrogen to Tyr-355.
The second factor that determines the relative 4-cyano
contributions to potential electrostatic interactions/hy-
drogen bond is related to the electron-withdrawing
capability of the 4-cyano group compared to other sub-
stituents. In the case of 5-aryl substitution, the aryl
group is competing with the 4-cyano group to withdraw
electrons from the pyrazole ring, as a result we expect to
see weaker electrostatic interaction/hydrogen bond of
the nitrile amine to Ser 530 hydrogen. Thus, contribu-
tion of 4-cyano versus 4-H is weaker with 5-aryl substi-
tution (3.4X in 3b vs 3a).
5-alkylamino, and 5-alkylether substituents that have
improved CWB COX-2 activity compared to parent
4-H N-2-pyridylmethylsulfonyl and N-phenylmethylsul-
fonyl pyrazoles. The improved activity can be explained
by the potential gain in binding energy due to favorable
interactions with the appropriate hydrogen bond donors
in the active site. Based on the data we present, the strong
hydrophobic binding of aryl substituents in the 5-position
of the pyrazole ring can outweigh other electrostatic/hy-
drogen bond contributions from the 4-nitrile or the N-2-
pyridyl nitrogen. However, for weakly binding sidechains
at the 5-position, electrostatic/hydrogen bond contribu-
tion from 4-cyano or/and N-2-pyridyl nitrogen is required
for activity in the CWB COX-2 assay. The SAR also
strongly suggests that the electrostatic/hydrogen bond
interactions are similar between the COX-1 and COX-2
enzymes with the ligand, which explains improvement in
activity against the COX-1 as well. However, we cannot
rule out the possibility that the enhanced activity, espe-
cially for COX-1, may be due to other modes of binding.
While our modeling studies confirm the SAR of our com-
pounds, it would be beneficial to confirm the modeling
studies with ligand bound X-ray structures.32
Our COX-2 binding models of N-2-pyridylmethyl-sulfo-
nyl also show a potential hydrogen bond of 2-pyridyl
nitrogen with Tyr 355 with a distance of approximately
˚
3.0 A. However, the hydroxyl angle with respect to the
Tyr 355 phenyl ring is not ideal from a molecular
mechanics minimization perspective. We attribute this
to the fact that there are strong electrostatic fields in
the region formed by Arg 120, Glu 524, and Arg 513,
and because our modeling method did not include
explicit water molecules.
Modeling of 5g (Fig. 4) provided insight into the poten-
cy gain of having 3-methyl piperidine at the 5-position.
As Ser 530 hydroxyl moves to form a hydrogen bond
with 4-cyano nitrogen, a small hydrophobic pocket is
formed near Phe 381, Tyr 385, and Ser 530. The methyl
group of 3-methyl piperidine fits well into the pocket,
potentially providing better binding potency for 5g.
Other substitution like cis-2,5-dimethylmorpholine
might be slightly too big to fit because of di-substitution,
thus, reducing the binding potency of the compound.
Acknowledgment
We acknowledge Gus Campos for providing us with the
kinetic solubility data.
References and notes
1. Otta, J. C.; Smith, W. L. J. Lipid. Mediators Cell
Signalling 1995, 12, 139.
In summary, we have highlighted some potent N-2-pyr-
idylmethylsulfonyl 4-nitrile pyrazoles with 5-aryl,

1072
S. M. Sakya et al. / Bioorg. Med. Chem. Lett. 17 (2007) 1067–1072
2. Carter, J. S. Expert Opin. Ther. Patents 1998, 8, 21.
Dodd, J. H.; Schieven, G. L.; Leftheris, K. J. Med. Chem.
2005, 48, 6261, and references cited therein.
23. Menozzi, G.; Merello, L.; Fossa, P.; Mosti, L.; Piana, A.;
Mattioli, F. Il Farmaco 2003, 5, 795.
3. Laneuville, O.; Breuer, D. K.; Dewitt, D. L.; Hla, T.;
Funk, C. D.; Smith, W. L. J. Pharmacol. Exp. Ther. 1994,
271, 927.
4. O’Neill, G. P.; Mancini, J. A.; Kargman, S.; Yergey, J.;
Kwan, M. Y.; Falgueyret, J. P.; Abramovitz, M.; Kenne-
dy, B. P.; Ouellet, M.; Cromlish, W. Mol. Pharmacol.
1994, 45, 245.
5. Prasit, P.; Riendeau, D. Annu. Rep. Med. Chem. 1997, 32,
211.
6. Warner, T. D.; Mitchell, J. A. Proc. Natl. Acad.
Sci. U.S.A. 2002, 99, 13371, and references cited
herein.
7. Clive, D. M.; Stoff, J. S. New Eng. J. Med. 1984, 310, 563.
8. Allison, M. C.; Howatson, A. G.; Torrance, C. J.; Lee, F.
D.; Russell, R. I. New Eng. J. Med. 1992, 327, 749.
9. Griswold, D. E.; Adams, J. L. Med. Res. Rev. 1996, 16,
181.
24. Sakya, S. M.; Rast, B. Tetrahedron Lett. 2003, 44, 7629.
25. Shavnya, A.; Sakya, S. M.; Minich, M. L.; DeMello, K.
L.; Jaynes, B. H. Tetrahedron Lett. 2005, 46, 6887.
26. Whole blood was collected by venal puncture into two tubes
with and without heparin. All analogs were dissolved in
DMSO. Tubes containing 2 ll of various drugs at concen-
trations ranging from 500 to 0.005 lM were prepared ahead
of time. Five hundred microliters of whole blood without
heparin was immediately added to the tubes with drug.
Following incubation for an hour at 37 ꢂC, COX-1 activity
was determined by measuring the thromboxane B2 (TXB2)
synthesized from platelets, using an enzyme immunoassay
(EIA) kit. Samples without drug were included as controls
for maximum production of TXB2. Five hundred microli-
ters of heparinized blood was added to tubes containing
drug and 10 lg/ml of LPS (to stimulate production of
PGE2) for COX-2 activity. LPS and vehicle only samples,
without drug, were included as controls for maximum PGE2
production and background values, respectively. Samples
were incubated overnight at 37 ꢂC. EDTA, 0.3% final
concentration, was added to the samples to alleviate clotting
of plasma after freeze-thaw. Samples were centrifuged,
serum/plasma was collected in 96-well micro titer plates and
stored at À20 ꢂC for evaluation in the EIA kit. Cayman EIA
kits were used according to manufacturer’s instructions, to
measure production of TBX2 and PGE2 for COX-1 and
COX-2 activity, respectively. Samples were diluted to fall in
the approximate range of the kit standards (1/10,000 for
TXB2 and 1/1000 for PEG2).
10. Cryer, B.; Dubois, A. Prostaglandins Other Lipid Mediat.
1998, 56, 341.
11. Penning, T. D.; Talley, J. J.; Bertenshaw, S. R.; Carter, J.
S.; Collins, P. W.; Doctor, S.; Graneto, M. J.; Lee, L. F.;
Malecha, J. W.; Miyashiro, J. M.; Rogers, R. S.; Rogier,
D. J.; Yu, S. S.; Anderson, G. D.; Burton, E. G.; Cogburn,
J. N.; Gregory, S. A.; Koboldt, C. M.; Perkins, W. E.;
Seibert, K.; Veenhuizen, A. W.; Zhang, Y. Y.; Isakson, P.
C. J. Med. Chem. 1997, 40, 1347.
12. Lehman, F. S.; Beglinger, C. Curr. Top. Med. Chem. 2005,
5, 449.
13. Fox, S. M.; Johnston, S. A. J. Am. Vet. Med. Assoc. 1997,
210, 1493.
14. Proprietary market research, September 1998. Owners of
osteoarthritic dogs, sample size: 150. Study No. RI199807.
Data on file with Pfizer Animal Health.
15. Kore, A. M. Vet. Clin. North Am. Small Anim. Pract.
1990, 20, 419.
16. (a) Ricketts, A. P.; Lundy, K. M.; Seibel, S. B. Am. J. Vet.
Res. 1998, 59, 1441; (b) Brideau, C.; Staden, C. V.; Chan,
C. C. Am. J. Vet. Res. 2001, 62, 1755.
17. Sessions, J. K.; Reynolds, L. R.; Budsberg, S. C. Am.
J. Vet. Res. 2005, 66, 812.
18. (a) Federal Register 2004, 69, 69523; (b) Streppa, H. K.;
Jones, C. J.; Budsberg, S. C. Am. J. Vet. Res. 2002, 63, 91.
19. (a) Li, J.; DeMello, K. M. L.; Cheng, H.; Sakya, S. M.;
Bronk, B. S.; Rafka, R. J.; Jaynes, B. H.; Ziegler, C. B.;
Kilroy, C.; Mann, D. W.; Nimz, E. L.; Lynch, M. P.;
Haven, M. L.; Kolosko, N. L.; Minich, M. L.; Li, C.;
Dutra, J. K.; Rast, B.; Crossan, R.; Morton, B. J.; Kirk,
G. W.; Callaghan, K. M.; Koss, D. A.; Shavnya, A.;
Lund, L. A.; Seibel, S. B.; Petras, C. F.; Silvia, A. M.
Bioorg. Med. Chem. Lett. 2004, 14, 95; (b) Cheng, H. M.;
Li, J.; Lundy, K. M.; Minich, M. L.; Sakya, S. M. WO
0140126, A1, 2001.
20. Sakya, S. M.; Cheng, H.; Demello, K. M. L.; Shavnya, A.;
Minich, M. L.; Rast, B.; Dutra, J.; Li, C.; Rafka, R. J.;
Koss, D. A.; Li, J.; Jaynes, B. H.; Ziegler, C. B.; Mann, D.
W.; Petras, C. F.; Seibel, S. B.; Silvia, A. M.; George, D.
M.; Hickman, A.; Haven, M. L.; Lynch, M. P. Bioorg.
Med. Chem. Lett. 2006, 16, 1205.
21. Sakya, S. M.; Demello, K. M. L.; Minich, M. L.; Rast, B.;
Shavnya, A.; Rafka, R. J.; Koss, D. A.; Cheng, H.; Li, J.;
Jaynes, B. H.; Ziegler, C. B.; Mann, D. W.; Petras, C. F.;
Seibel, S. B.; Silvia, A. M.; George, D. M.; Lund, L. A.;
St. Denis, S.; Haven, M. L.; Lynch, M. P. Bioorg. Med.
Chem. Lett. 2006, 16, 288.
22. Liu, C.; Wrobleski, S. T.; Lin, J.; Ahmed, G.; Metzger, A.;
Wityak, J.; Gillooly, K. M.; Shuster, D. J.; McIntyre, K.
W.; Pitt, S.; Shen, D. R.; Zhang, R. F.; Zhang, H.;
Doweyko, A. M.; Diller, D.; Henderson, I.; Barrish, J. C.;
27. (a) Kurumbail, R. G.; Stevens, A. M.; Gierse, J. K.;
McDonald, J. J.; Stegeman, R. A.; Pak, J. Y.; Gildehaus,
D.; Miyashiro, T. D.; Penning, J. M.; Seibert, K.; Isakson,
P. C.; Stallings, W. C. Nature 1996, 384, 644; (b)
Trummlitz, G.; Van Ryn, J. Curr. Opin. Drug Discov.
Devel. 2002, 5, 550; (c) Datar, P. A.; Coutinho, E. C.
J. Mol. Graphics Modell. 2004, 23, 239; (d) Liu, H.;
Huang, X.; Shen, J.; Luo, X.; Li, M.; Xiong, B.; Chen, G.;
Shen, J.; Yang, Y.; Jiang, H.; Chen, K. J. Med. Chem.
2002, 45, 4816.
28. Wilson, J. E.; Chandrasekharan, N. V.; Westover, K. D.;
Eager, K. B.; Simmons, D. L. Am. J. Vet. Res. 2004, 66, 810.
29. The molecular minimization was performed using Mac-
roModel30 software. The partial flexible binding site was
constructed from the co-crystal structure of SC-588 with
COX-2 (PDB: 6COX). Inhibitors and protein binding site
residues were fully flexible and protein residues further
˚
away from the ligand (>5 A) were harmonically con-
strained to their crystallographic positions during the
minimization. Ab initio partial atomic charges were
computed and assigned to all ligands and AMBER*
atomic charges were assigned to protein residues. The
complexes were energetically minimized using the Batch-
min program of MacroModel V8.6 with the AMBER*
force field and the GB/SA solvation model.
30. (a) MacroModel, Schrodinger, LLC, New York, NY,
2005; (b) Mohamadi, F.; Richards, N. G. J.; Guida, W. C.;
Liskamp, R.; Lipton, M.; Caufield, C.; Chang, G.;
Hendrickson, T.; Still, W. C. J. Comput. Chem. 1990, 11,
440.
31. (a) Le Questel, J.-Y.; Berthelot, M.; Laurence, C. J. Phys.
Org. Chem. 2000, 13, 347; (b) Ziao, N.; Graton, J.;
Laurence, C.; Le Questel, J.-Y. Acta Crystallogr., Sect. B
2001, 57, 850.
32. Part of this paper has been presented at 2006 Fall ACS
meeting in San Francisco.