Synthesis and pharmacological evaluation of m-terphenyl amines as cyclooxygenase inhibitors

Article abstract of DOI:10.1007/s00044-007-9015-x

A series of m-terphenyl amines was synthesized and evaluated as a novel class of cyclooxygenase (COX) inhibitors. Structure-activity relationships (SAR) were investigated by functional group modification at the para-position of the C-1′ and C-2′ phenyl su

Full text of DOI:10.1007/s00044-007-9015-x

Med Chem Res (2007) 16:119–129
DOI 10.1007/s00044-007-9015-x
ORIGINAL RESEARCH
Synthesis and pharmacological evaluation of m-
terphenyl amines as cyclooxygenase inhibitors
John D. Bauer Æ Michael S. Foster Æ Jeffrey D. Hugdahl Æ Kristi L. Burns Æ
Sheldon W. May Æ Stanley H. Pollock Æ Horace G. Cutler Æ Stephen J. Cutler
Received: 8 October 2006 / Accepted: 2 February 2007 / Published online: 25 July 2007
¨
Ó Birkhauser Boston 2007
Abstract A series of m-terphenyl amines was synthesized and evaluated as a
novel class of cyclooxygenase (COX) inhibitors. Structure–activity relationships
(SAR) were investigated by functional group modification at the para-position of
the C-10 and C-20 phenyl substituents on the central aromatic ring. Anilines 6a, b, d,
and h demonstrated nonselective inhibition of COX-1 and -2 in human whole blood.
Compounds 6c and e demonstrated preferential inhibition of the COX-2 isozyme at
10 lM. Molecules 6f, i, and j inhibited only COX-1, and the disubstituted ethoxy
derivative (6g) was inactive as a COX inhibitor ( ꢀ 100 lM). Molecular docking
studies of these compounds indicate that the COX-1 binding site amino acid Ile523
anchors the m-terphenyl system statically within the enzyme’s active site, while the
slightly smaller amino acid Val523 in COX-2 allows the ligand to ‘‘roll,’’ fashioning
several acceptable conformers.
J. D. Bauer ꢁ S. H. Pollock ꢁ H. G. Cutler
College of Pharmacy and Health Sciences, Mercer University, Atlanta, GA 30341, USA
M. S. Foster ꢁ K. L. Burns ꢁ S. W. May
School of Chemistry and Biochemistry, Georgia Institute of Technology, Atlanta, GA 30332, USA
J. D. Hugdahl
Department of Chemistry, College of Liberal Arts, Mercer University, 1400 Coleman Avenue,
Macon, GA 31207, USA
S. J. Cutler (&)
Department of Medicinal Chemistry, School of Pharmacy, The University of Mississippi,
University, MS 38677, USA
e-mail: cutler@olemiss.edu

120
Med Chem Res (2007) 16:119–129
Introduction
The use of medicinal substances for their analgesic, antiinflammatory, and
antipyretic properties dates well back to the time of Hippocrates (400 B.C.), when
patients were advised to chew willow bark (Salix alba) to reduce fever and
inflammation. Salicin, the active constituent of the bark, was identified in the early
19th century and benchmarked the ensuing frontier of nonsteroidal antiinflamma-
tory drug (NSAID) development. It was not until 1971 that Sir John Vane would
revolutionize our understanding of NSAID pharmacology by suggesting that
aspirin-like drugs target the rate-limiting enzyme cyclooxygenase (COX), respon-
sible for catalyzing the biosynthesis of eicosanoids (Vane, 1971). Specifically, COX
catalyses the conversion of arachidonic acid, mobilized from membrane phospho-
lipids by a phospholipase, into the endoperoxides prostaglandin G2 (PGG2) and
PGH2. These unstable precursors are subsequently transformed by tissue-specific
enzymes into prostanoids such as thromboxane A2 (TxA2), prostacyclin (PGI2),
PGD2, PGE2, and PGF2a (Botting and Botting, 2004).
It is now well established that COX-1 is a ubiquitous and constitutive enzyme
responsible for maintaining important physiological functions of virtually all
tissues, especially gastrointestinal integrity (PGE2 and PGI2), renal function (PGE2),
and platelet activation (TxA2). While COX-2 is also found constitutively expressed
in some tissues, it is largely induced during inflammation (Seibert et al., 1990). This
discovery led to the development of selective COX-2 inhibitors in hopes of
ameliorating the gastric and renal toxicity associated with traditional nonselective
NSAID therapy. Although less gastrotoxicity is attributed to the use of selective
COX-2 inhibitors when compared to traditional nonselective inhibitors, recent
evidence suggests that prolonged selective inhibition of COX-2 may increase the
risk for adverse cardiovascular events (Warner et al., 1999; Bombardier et al.,
2000). Rofecoxib is considered to be 50-fold more selective for the COX-2 isozyme
(Bresalier et al., 2005) and was recently withdrawn from the market because of the
latest evidence suggesting its potential to evoke adverse cardiovascular events. It
appears, however, that a certain degree of selective COX-2 inhibition can reduce
gastrotoxicity without altering vascular endothelial function. A milieu of clinical
data supports this notion, considering meloxicam and celecoxib, unlike rofecoxib,
are both preferential inhibitors of COX-2 (less than 50-fold selectivity; Bresalier
et al., 2005) and exhibit less gastrotoxicity when compared to traditional NSAIDs
without provoking adverse cardiovascular outcomes (Degner, 2004). These findings
suggest that compounds possessing preferential affinity for the COX-2 isozyme may
be of greater benefit in the long-term management of inflammatory disorders than
exceedingly selective COX-2 agents.
We have previously shown 1-(carboxymethyl)-3,5-diphenyl-2-methylbenzene
(CDB; Fig. 1), a novel m-terphenyl, to be a preferential inhibitor of COX-1 (COX-1
IC50 = 17 lM and COX-2 IC50 = 120 lM) and an effective NSAID against acute
and chronic inflammation in vivo (Cutler et al., 1998). In an attempt to improve the
preferential affinity of this novel m-terphenyl for the COX-2 isozyme, we evaluated
numerous derivatives of CDB for their ability to inhibit COX enzyme activity
in vitro.

Med Chem Res (2007) 16:119–129
121
O
CH3
H3C
Fig. 1 Structure of CDB
We now report the synthesis and evaluation of a series of m-terphenyl amines as
a novel class of COX inhibitors. Further, molecular docking studies help to
rationalize the structure–activity relationship (SAR) for the compounds evaluated
herein.
Results and Discussion
In continuation of our studies, we synthesized numerous derivatives of CDB
utilizing known synthetic pathways and evaluated these compounds as novel COX
inhibitors. The m-terphenyl amines (6) were synthesized following a procedure
previously used by Sepiol and Milart (1985). The benzylidene malononitriles (2)
and phenyl ethylidene malononitriles (4) were prepared via Knoevenagel conden-
sation of malononitrile with the appropriate benzaldehyde or acetophenone,
respectively. Michael addition of 4 to 2, followed by Thorpe–Ziegler cyclization
and loss of hydrogen cyanide, furnished the corresponding o-aminodinitriles (5).
Removal of the nitriles from 5 was achieved under hydrolysis conditions in an
ethanolic solution of sodium hydroxide at 220°C and afforded the desired m-
terphenyl amines (6) in moderate to good yields. It should be noted that under the
elimination conditions, the fluorine atoms on 5f and g were replaced by the ethoxy
nucleophile (6f, g). To synthesize 6j, the methyl sulfide (6i) was acetylated, yielding
compound 7a, which then allowed for the selective oxidation to the corresponding
methanesulfonyl (7b). Subsequent deprotection of 7b provided 6j in good yield.
The resulting anilines were evaluated against COX enzyme activity in human
whole blood. By comparing the percentage inhibition of COX enzyme activity for
compounds 6a–j (Scheme 1), (Scheme2), (Table 1), it is clear that the potency of these
m-terphenyl amines correlates to the character of the substitution on the 4 (R1) and 400
(R2) positions. Compound 6a demonstrated good nonselective inhibition for COX-1
and -2. The addition of a methyl group on position R1 (molecule 6b), while exhibiting a
slightly diminished activity compared to 6a, maintained nonselective inhibition of the
COX isozymes. Interestingly, the disubstituted methyl derivative (6c) possessed less
affinity for COX-1 and -2 compared to substances 6a and 6b. Incorporation of a single
chloro substituent (compound 6d) enhanced the inhibitory capacity of the aniline for
the COX isozymes compared to the parent compound 6a. Like molecule 6c, the

122
Med Chem Res (2007) 16:119–129
NC
CN
H
2
a
b
c
d
e
f
R1
H
O
a
CH3
Cl
H
R1
R1
F
1
2
CH3O
CH3S
4
a
b
c
d
R2
H
NC
CN
O
CH3
Cl
b
CH3
CH3
F
R2
R2
3
4
5
a
b
c
d
e
f
R1
H
R2
H
CH3
CH3
Cl
H
CH3
H
NH2
NC
CN
Cl
Cl
H
F
c
2
g
h
i
F
F
CH3O
CH3S
H
5
R1
R2
H
d
6
a
b
c
d
e
f
R1
H
R2
H
CH3
H
NH2
CH3
CH3
Cl
H
Cl
Cl
CH3CH2O
CH3CH2O
CH3O
CH3S
H
g
h
i
CH3CH2O
6
R1
R2
H
H
Scheme 1 Synthesis of compounds 6a–i. Reagents and conditions: (a) piperidine, ethanol, 25°C; (b)
ammonium chloride, glacial acetic acid, benzene, reflux; (c) 4, piperidine, acetonitrile, 25°C then reflux;
(d) NaOH, ethanol, 220°C
disubstituted derivative 6e was less active against COX enzyme activity compared to
its monosubstituted counterpart. However, the weaker activity observed for 6e was
more exaggerated for COX-1 than COX-2, indicative of a preferential inhibitor.
Substance 6f, having a monoethoxy substitution, was devoid of COX-2 activity and
moderately inhibited COX-1 similarly to the disubstituted compounds 6c and 6e. The
presence of two ethoxy substituents inactivated molecule 6g as a COX inhibitor.
Compound 6h weakly inhibited COX-1 and -2 enzyme activity, while molecules 6i
and 6j moderately inhibited only COX-1.
While the m-terphenyl amines do not bear the customary arylacetic acid
pharmacophore, generally appreciated as a characteristic of traditional NSAIDs, it
was of interest to explore the molecular docking of these compounds in the active

Med Chem Res (2007) 16:119–129
123
O
NH2
HN
CH3
a
6i
7a
S
S
H3C
H3C
b
O
NH2
HN
CH3
c
O
O
6j
7b
S
S
H3C
O
H3C
O
a
Scheme 2 Synthesis of compound 6j. Reagents and conditions: (a) TEA, acetic anhydride, methylene
chloride, 25°C; (b) oxone, methanol, 25°C; (c) NaOH, ethanol, 220°C
Table 1 Percent inhibition of COX-1 and COX-2 in human whole blood by compounds 6a–j
6
R1
R2
COX-1
COX-2
10 lM
100 lM
10 lM
100 lM
a
b
c
H
H
56
46
27
63
19
20
Na
Na
29
23
5
6
6
8
4
7
98
96
87
97
93
88
na
92
92
85
0
0
2
0
1
5
68 13
92
95
86
98
99
na
na
78
na
na
1
1
5
1
1
CH3
H
49
1
CH3
CH3
57 11
d
e
Cl
H
85
76
na
na
na
na
na
2
0
Cl
Cl
F
g
h
I
CH3CH2O
CH3CH2O
CH3O
CH3S
CH3SO2
H
CH3CH2O
H
H
H
0
2
1
9
4
8
J
na = no activity. Data are presented as mean SEM for each group (n = 3)
sites of COX-1 and -2. Remarkably, all of the evaluated m-terphenyls effectively
bound in a similar fashion in the COX-1 active site (Fig. 2). In all cases, the amine
function was directed away from the Arg120/Tyr355 pair which anchors the
carboxylate moiety of arachidonic acid. Instead, hydrogen bonds were formed
between the ligand amine and the backbone peptide bonds of neighboring amino
acids (Leu352 and Ser353). Further, the derivatized phenyl ring was projected into the

124
Med Chem Res (2007) 16:119–129
Tyr385
Phe503
Phe381
Ile434
Trp387
Leu384
Phe518
Ser530
His513
Ile523
Arg120
Tyr355
Fig. 2 6d docked in the active site of ovine COX-1 overlaying the crystal structure of arachidonic acid
large hydrophobic pocket flanked by several aromatics such as Phe381, Tyr385
,
Trp387, and Phe518 directed upward toward the ‘‘business-end’’ of the active site
where the cyclooxygenase chemistry is performed. The underivatized phenyl ring,
or the second derivatized ring of the disubstituted analogs, projects into the area
where the carboxylate function of arachidonic acid binds between Arg120 and
Tyr355, essentially overlaying C-1, C-2, and C-3 of arachidonic acid from the crystal
structure. Overall, Ile523 appears to support the entire ligand like a gloved hand,
preventing the adoption of any other active site conformers.
Although the active sites of COX-1 and -2 are similar, there are differences that
ostensibly affect the binding of the m-terphenyl amines within these isozymes. The
Phe503 to Leu503 substitution in COX-2 allows Leu384 to reorient its aliphatic side
chain and extend the space available for a larger inhibitor molecule. Interestingly,
the m-terphenyl amines do not depart significantly into this flexible extra space at
the top of the COX-2 active site (Fig. 3). However, the Ile523 to Val523 substitution
is not insignificant. This slightly smaller amino acid in COX-2 allows access to a
side pocket that is inaccessible in COX-1. Therefore, relative to the COX-1
conformers, at least two sets of comparatively high-scoring conformers were
generated, such that the amine function remains essentially in place from conformer
to conformer, but with the phenyl rings oriented into different areas. Unlike the
orientation in COX-1, the amine function of the inhibitors in COX-2 is directed
toward the Arg120/Tyr355 pair.
In summary, we present a novel class of COX inhibitors that lack the
traditional arylacetic acid moiety. It is evident from whole blood data that the
nature of the substitution on the R1 and R2 positions affects the inhibitory
capacity of these molecules. Molecular docking studies propose how these
inhibitors may be interacting with the active site of the COX isozymes and help
to rationalize the observed variations in enzyme inhibition among these

Med Chem Res (2007) 16:119–129
125
Leu503
Tyr385
Val434
Leu384
Phe518
Ser530
Arg513
Val523
Tyr355
Arg120
Fig. 3 6a docked in the active site of murine COX-2. The preferred conformation is shown in white,
while the alternate binding mode is shown in gray
derivatives. Preferential inhibitors for COX-2 may offer the safest option for the
long-term management of inflammatory conditions when compared to nonselec-
tive or highly selective COX-2 inhibitors (Degner, 2004). Therefore, studies are
currently underway to investigate these tunable inhibitors by taking advantage of
their flexible fit in the COX-2 active site over their more rigid binding in the
COX-1 active site. Ultimately, this should afford more potent inhibitors with a
spectrum of COX isozyme selectivity.
Experimental
Chemistry
Melting points were determined using a Thomas–Hoover capillary melting point
apparatus and are uncorrected. IR spectra were obtained on a Perkin-Elmer 16PC
1
spectrometer. The H- and 13C-nuclear magnetic resonance (NMR) spectra were
recorded on 300 and 100 MHz Varian spectrometers, respectively. Chemical shifts
are reported in d values (ppm) relative to Me4Si as internal standard and J values are
reported in Hz.
Preparation of starting ylidenemalononitriles
Benzylidine malononitriles (2) were obtained by condensation of equimolar
amounts (65 mmol) of aromatic aldehydes and malononitrile in ethanol, catalyzed
by three or four drops of piperidine. The physical and spectral data for obtained

126
Med Chem Res (2007) 16:119–129
products were in good agreement with the reported data for compounds 2a–e
(Kluepfel and Horner, 1955; Campaigne et al., 1962; Schiemenz, 1962).
2-(4-Methylsufide-benzylidine)-malononitrile (2f). Yield 93%; IR cmꢂ1 2219
1
(CN), 821, 813; H-NMR (CDCl3) d, ppm 7.81 (d, J = 9, 2H, Ar–H), 7.65 (s, 1H),
7.30 (d, J = 9, 2H, Ar–H), 2.55 (s, 3H, CH3).
Phenyl ethylidene malononitirles (4) were obtained by condensation of equimolar
amounts (65 mmol) of aryl methyl ketones and malononitrile under previously
described conditions (Mowry, 1945). Their physical and spectral data were in
agreement for the published data for 4a–d (Mowry, 1945).
General procedure for the synthesis of 1-amino-2,6-dicyano-3,5-
diphenylbenzenes (5)
Following a procedure described previously, equimolar amounts of 2 and 4 (20
mmol) were placed in acetonitrile (30 ml) and piperidine (2 ml) was added dropwise
to the stirred solution over several minutes. After the formation of a precipitate, the
suspension was refluxed for 2 h and cooled to room temperature. The precipitate
was then filtered off, washed with ethanol, and generally used in the next reaction
without further purification. The physical and spectral data for obtained products
were in good agreement with the reported data for compounds 5a, b, and d (Sepiol
and Milart, 1985).
50-Amino-4-methyl-400-methyl-1,10:30,100-terphenyl-40,60-dicarbonitrile (5c). Yield
1
42%; IR cmꢂ1 3474, 3359 (NH2), 2210 (CN), 1633, 1284, 816, 807; H-NMR
(CDCl3) d, ppm 7.48–7.25 (m, 8H, Ar–H), 6.87 (s, 1H, Ar–H), 5.25 (s, 2H, NH2),
2.43 (s, 6H, CH3).
50-Amino-4-chloro-400-chloro-1,10:30,100-terphenyl-40,60-dicarbonitrile (5e). Yield
1
81%; IR cmꢂ1 3463, 3351, 3243 (NH2), 2219 (CN), 1643, 1284, 1092, 819; H-
NMR (CDCl3) d, ppm 7.50 (m, 8H, Ar–H), 6.81 (s, 1H, Ar–H), 5.40 (s, 2H, NH2).
50-Amino-4-fluoro-1,1030,10-terphenyl-40,60-dicarbonitrile (5f). Yield 51%; IR
1
cm–1 3474, 3363 (NH2), 2215 (CN), 1644, 1512, 1230, 830, 762, 692; H-NMR
(CDCl3) d, ppm 7.57–7.47 (m, 7H, Ar–H), 7.18 (m, 2H, Ar–H), 6.84 (s, 1H, Ar–H),
5.39 (s, 2H, NH2).
50-Amino-4-fluoro-400-fluoro-1,10:30,100-terphenyl-40,60-dicarbonitrile (5g). Yield
51%; IR cm–1 3470, 3348, 3244 (NH2), 2208 (CN), 1648, 1513, 1229, 1159, 823;
1H-NMR (CDCl3) d, ppm 7.56 (m, 4H, Ar–H), 7.20 (m, 4H, Ar–H), 6.82 (s, 1H, Ar–
H), 5.37 (s, 2H, NH2).
50-Amino-4-methoxy-1,10:30,100-terphenyl-40,60-dicarbonitrile (5h). Yield 13%; IR
cmꢂ1 3449, 3359, 3249 (NH2), 2212 (CN), 1648, 1549, 1257, 1238, 1179, 1031,
1
833, 761, 695; H-NMR (CDCl3) d, ppm 7.57–7.49 (m, 5H, Ar–H), 7.33 (m, 2H,
Ar–H), 7.01 (m, 2H, Ar–H), 6.87 (s, 1H, Ar–H), 5.30 (s, 2H, NH2), 3.87 (s, 3H,
CH3).
50-Amino-4-methylsulfide-1,10:30,100-terphenyl-40,60-dicarbonitrile (5i). Yield
59%; IR cm–1 3476, 3383 (NH2), 2210 (CN), 1625, 1573, 1282, 1095, 810, 691;
1H-NMR (CDCl3) d, ppm 7.55–7.00 (m, 9H, Ar–H), 6.64 (s, 1H, Ar–H), 5.18 (s, 2H,
NH2), 2.56 (s, 3H, CH3).

Med Chem Res (2007) 16:119–129
127
General procedure for the synthesis of 1-amino-3,5-diphenylbenzenes (6)
In a 200-ml stainless steel reaction vessel (Parr Instrument Company, Moline, IL)
was placed 5 (4 mmol) in a solution of sodium hydroxide (1.5 g) in ethanol (60 ml).
The vessel was heated at 220°C for 5 h in an oven. The opening of the vessel was
accompanied by the release of some amount of ammonia and carbon dioxide. After
evaporation of the ethanol, the crude product was recrystallized from appropriate
solvent. The physical and spectral data for obtained products were in good
agreement with the reported data for compounds 6a, b, and d (Sepiol and Milart,
1985).
50-Amino-4-methyl-400-methyl-1,10:30,100-terphenyl (6c). Yield 90%, m.p. 154–
155° (from hexanes-benzene); IR cm–1 3422, 3290, 3192 (NH2), 1596, 1517, 861,
807; 1H-NMR (CDCl3) d, ppm 7.49 (m, 4H, Ar–H), 7.35–7.17 (m, 5H, Ar–H), 6.85
(s, 2H, Ar–H), 3.77 (s, 2H, NH2), 2.39 (s, 6H, CH3); 13C-NMR (CDCl3) d, ppm
146.93, 142.77, 138.48, 137.05, 129.35, 126.98, 116.67, 112.54, 21.09.
50-Amino-4-chloro-400-chloro-1,10:30,100-terphenyl (6e). Yield 61%, m.p. 108–
109° (from hexanes—benzene); IR cm–1 3423, 3297, 3187 (NH2), 1597, 1093, 821,
1
760, 696; H-NMR (CDCl3) d, ppm 7.58–6.84 (m, 11H, Ar–H), 3.84 (s, 2H, NH2);
13C-NMR (CDCl3) d, ppm 141.86, 139.59, 133.49, 128.84, 128.39, 127.16, 116.46,
112.88.
50-Amino-4-ethoxy-1,10:30,100-terphenyl (6f). Yield 78%, m.p. 144–145° (from
hexanes—benzene); IR cm–1 3438, 3367 (NH2), 1597, 1512, 1237, 1045, 833, 769,
696; 1H-NMR (CDCl3) d, ppm 7.76–7.35 (m, 10H, Ar–H), 6.95 (m, 2H, Ar–H), 4.06
(q, J = 7, 2H, CH2), 1.44 (t, J = 7, 3H, CH3).
50-Amino-4-ethoxy-400-ethoxy-1,10:30,100-terphenyl (6g). Yield 57%, m.p. 179–
180° (from acetone); IR cm–1 3449, 3367 (NH2), 1607, 1596, 1513, 1047, 823, 811;
1H-NMR (CDCl3) d, ppm 7.743–6.893 (m, 11H, Ar–H), 4.06 (q, J = 7, 4H, CH2),
1.44 (t, J = 7, 6H, CH3).
50-Amino-4-methoxy-1,10:30,100-terphenyl (6h). Yield 45%, m.p. 115–116° (from
hexanes-benzene); IR cm–1 3425, 3300 (NH2), 1595, 1513, 1249, 1180, 1028, 831,
765, 699; 1H-NMR (CDCl3) d, ppm 7.46–7.20 (m, 10H, Ar–H), 6.98 (m, 2H, Ar–H),
3.85 (s, 3H, CH3).
50-Amino-4-methylsulfide-1,10:30,100-terphenyl (6i). Yield 88%, m.p. 103–104°
1
(from hexanes—benzene); IR cm–1 3406, 3299 (NH2), 1595, 817, 761, 697; H-
NMR (CDCl3) d, ppm 7.80–7.42 (m, 12H, Ar–H), 2.78 (s, 3H, CH3).
N-(50-Amino-4-methylsulfide-1,10:30,100-terphenyl)-acetamide (7a). Yield 97%;
1
IR cm–1 3298 (NH2) 1665, 1598, 1551, 1421, 868, 758, 695; H-NMR (CDCl3) d,
ppm 7.78–7.43 (m, 12H, Ar–H), 2.84 (s, 3H, CH3), 2.27 (s, 3H, CH3).
N-(50-Amino-4-methanesulfonyl-1,10:30,100-terphenyl)-acetamide (7b). Yield
84%, m.p. 205–206° (from hexanes—benzene); IR cm–1 3300 (NH2) 1665, 1595,
1
1547, 1421, 1278, 1152, 944, 835, 762, 698; H-NMR (CDCl3) d, ppm 8.00–7.37
(m, 12H, Ar–H), 3.09 (s, 3H, CH3), 2.24 (s, 3H, CH3); 13C-NMR (CDCl3) d, ppm
168.70, 146.25, 142.99, 140.15, 139.12, 128.89, 128.13, 127.88, 127.20, 122.11,
118.67, 117.66, 44.62, 24.70.

128
Med Chem Res (2007) 16:119–129
50-Amino-4-methanesulfonyl-1,10:30,100-terphenyl (6j). Yield 91%, m.p. 165–166°
(from ethanol); IR cm–1 3459, 3365 (NH2) 1591, 1295, 1145, 956, 868, 765, 701;
1H-NMR (CDCl3) d, ppm 8.02–7.25 (m, 12H, Ar–H), 3.14 (s, 3H, CH3).
Biology
Assay for the inhibition of COX-1 enzyme activity in human whole blood
COX-1 activity was determined by measuring the production of thromboxane B2
(TxB2) which occurs during the spontaneous clotting of human whole blood
(Brideau et al., 1996). In brief, human venous blood was collected in Vacutainers
without anticoagulant from nonfasted healthy male volunteers who had not taken
any aspirin or NSAIDs for 14 days. Aliquots of blood (500 ll) were immediately
transferred to siliconized microcentrifuge tubes containing 2 ll of inhibitor or
vehicle [dimethyl sulfoxide (DMSO)]. Tubes were then vortex-mixed and incubated
for 1 h at 37°C to allow clot formation. Following incubation, tubes were
centrifuged and 100 ll of serum was mixed with 400 ll of methanol for protein
precipitation. The supernatants were used to assay for TXB2 using a commercially
available enzyme immunoassay kit (Oxford Biomedical, Oxford, MI, cat. no. EA
25).
Assay for the inhibition of COX-2 enzyme activity in human whole blood
COX-2 activity was determined by measuring the formation of PGE2 that occurs
after incubation of blood samples for 24 h with lipopolysaccharide (LPS, Sigma-
Aldrich, L-2630 from E. coli serotype 0111:B4) according to a method described
previously (Brideau et al., 1996). In brief, fresh human venous blood from nonfasted
healthy male volunteers who had taken aspirin for 2 days prior (81 mg QD) was
drawn into a Vacutainer containing heparin. We found that predosing with aspirin
limited background COX-1 mediated PGE2 production. Aliquots of blood (500 ll)
were transferred to polypropylene tubes and incubated with inhibitor or vehicle
(DMSO) for 15 min at 37°C. LPS (100 lg/ml final concentration) was added to the
blood mixture and incubated for 24 h at 37°C to induce COX-2 formation.
Following incubation, tubes were centrifuged and 100 ll of plasma was mixed with
400 ll of methanol for protein precipitation. The supernatants were assayed for
PGE2 using a commercially available EIA kit (Oxford Biomedical, Oxford, MI, cat.
no. EA 02).
Molecular modeling
All simulations were performed using the program Molecular Operating Environ-
ment 2004.03 (MOE) by Chemical Computing Group (Montreal, Quebec). The
protein files for COX-1 and -2 were imported from the Protein Data Bank. All
extraneous ligands, salts, noncovalently attached carbohydrates, and metal ions
were deleted before the running of any simulations. Proteins were initially
minimized using the Engh–Huber force field, with all non-hydrogen fixed in their

Med Chem Res (2007) 16:119–129
129
published positions. Ligands were constructed and minimized in MOE by the user
and minimized with the Engh–Huber force field. All docking simulations utilized
the Simulated Annealing module. All default parameters were used, except that
distance-based, rather than gas phase, solvation was enabled and non-bonded and
bonded cutoff parameters were changed to 5.5 and 4.5 Angstroms, respectively. The
top 5–10 scorers of each compound docked to each COX isoform were evaluated.
References
Bombardier C, Laine L, Reicin A, Shapiro D, Burgos-Vargas R, Davis B, Day R, Ferraz M, Hawkey CJ,
Hochberg MC, Kvien TK, Schnitzer TJ (2000) Comparison of upper gastrointestinal toxicity of
rofecoxib and naproxen in patients with rheumatoid arthritis. VIGOR Study Group N Engl J Med
343:1520–1528
Botting RM, Botting JH (2004) The discovery of COX-2. In: Pairet M, van Ryn J (eds) Milestones in
Drug Therapy: COX-2 Inhibitors. Birkhaeuser, Switzerland, pp 1–13
Bresalier RS, Sandler RS, Quan H, Bolognese JA, Oxenius B, Horgan K, Lines C, Riddell R, Morton D,
Lanas A, Konstam MA, Baron JA (2005) Cardiovascular events associated with rofecoxib in a
colorectal adenoma chemoprevention trial. N Engl J Med 352:1092–1102
Brideau C, Kargman S, Liu S, Dallob AL, Ehrich EW, Rodger IW, Chan CC (1996) A human whole
blood assay for clinical evaluation of biochemical efficacy of cyclooxygenase inhibitors. Inflamm
Res 45:68–74
Campaigne E, Bulbenko GF, Kreighbaum WE, Maulding DR (1962) Ring closure of ylidenemalono-
nitriles. J Org Chem 27:4428–4432
Cutler SJ, Blanton CD, Akin DT, Steinberg FB, Moore AB, Lott JA, Price TC, May SW, Pollock SH
(1998) The structure of human COX-2 and selective inhibitors. Inflamm Res 47:316–324
Degner F (2004) Efficacy and gastrointestinal safety of selective COX-2 inhibitors. In: Pairet M, van Ryn
J (eds) Milestones in Drug Therapy: COX-2 Inhibitors. Birkhauser, Switzerland, pp 133–148
Kluepfel K, Horner L (1955) Zum nachweis des polaren charakters in doppelbindungs-systemen
phosphororganische verbindungen II. Liebigs Ann Chem 591:69
Mowry DT (1945) The knoevenagel condensation of aryl alkyl ketones with malononitrile. J Am Chem
Soc 67:1050–1051
¨
Schiemenz GP (1962) Synthesen mit cyanessigsaure, III. zur reaktion von 2-hydroxy-benzaldehyd mit
cyanacetamid und malodinitril. Chem Ber 95:483
Seibert K, Masferrer JL, Jiyi F, Hond A, Raz A, Needleman P (1990) The biochemical and
pharmacological manipulation of cellular cyclooxygenase (COX) activity. Adv Prostaglandin
Thromboxane Leukotriene Res 21:45–51
Sepiol J, Milart P (1985) A new and efficient synthesis of 3,5-diarylaminobenzenes from arylidenem-
alonodinitriles and 1-atylethylidenemalonodinitriles. Tetrahedron 41:5261–5265
Vane JR (1971) Inhibition of prostaglandin synthesis as a mechanism of action for aspirin-like drugs.
Nature 231:232–235
Warner TD, Giuliano F, Vojnovic I, Bukasa A, Mitchell JA, Vane JR (1999) Nonsteroid drug selectivities
for cyclo-oxygenase-1 rather than cyclo-oxygenase-2 are associated with human gastrointestinal
toxicity: a full in vitro analysis. Proc Natl Acad Sci USA 96:7563–7568