5-Arylidene-2-imino-4-thiazolidinones: Design and synthesis of novel anti-inflammatory agents

Article abstract of DOI:10.1016/j.bmc.2005.04.058

The synthesis and pharmacological activity of 5-arylidene-2-imino-4- thiazolidinones (3a-8a) are described. All derivatives exhibited significant activity levels in models of acute inflammation such as carrageenan-induced paw and pleurisy edema in rats. I

Full text of DOI:10.1016/j.bmc.2005.04.058

Bioorganic & Medicinal Chemistry 13 (2005) 4243–4252
5
-Arylidene-2-imino-4-thiazolidinones: Design and synthesis
of novel anti-inflammatory agents
a,
Rosaria Ottan a` , Rosanna Maccari, Maria Letizia Barreca, Giuseppe Bruno,
a
a
b
*
b
Archimede Rotondo, Antonietta Rossi, Giuseppa Chiricosta, Rosanna Di Paola,
c
a
d
c
Lidia Sautebin, Salvatore Cuzzocrea and Maria Gabriella Vigorita
d
a
a
Dipartimento Farmaco-chimico, Facolt a` di Farmacia, Universit a` di Messina, Viale SS. Annunziata, 98168 Messina, Italy
b
Dipartimento Ch. Inorg., Chim. Anal. e Ch.-fis., Facolt a` di Scienze MMFFNN, Universit a` di Messina, Salita Sperone,
3
Dipartimento di Farmacologia Sperimentale, Universit a´ di Napoli ‘Federico II’, Via Domenico Montesano 49, 80131 Napoli, Italy
1, 98166 Messina, Italy
c
d
Istituto di Farmacologia, Facolt a` di Medicina e Chirurgia, Universit a` di Messina, Policlinico Universitario ‘G. Martino’,
Via Consolare Valeria, 98124 Messina, Italy
Received 1 July 2004; accepted 12 April 2005
Available online 17 May 2005
Abstract—The synthesis and pharmacological activity of 5-arylidene-2-imino-4-thiazolidinones (3a–8a) are described. All deriva-
tives exhibited significant activity levels in models of acute inflammation such as carrageenan-induced paw and pleurisy edema
in rats. In particular, 5-(3-methoxyphenylidene)-2-phenylimino-3-propyl-4-thiazolidinone (3a) displayed high levels of carra-
geenan-induced paw edema inhibition comparable to those of indomethacin. In addition the ability of such a new class of
anti-inflammatory agents to inhibit COX isoforms was assessed in murine monocyte/macrophage J774 cell line assay. 5-(4-Meth-
oxyphenylidene)-2-phenylimino-3-propyl-4-thiazolidinone (6a), the most interesting compound in such an experiment, was docked
in the known active site of COX-2 protein and showed that its 4-methoxyarylidene moiety can easily occupy the COX-2 secondary
pocket considered as the critical interaction for COX-2 selectivity.
ꢀ
2005 Elsevier Ltd. All rights reserved.
3
1. Introduction
(COX), the enzyme involved in the first step of the con-
version of arachidonic acid to prostaglandins (PGs).
These latter regulate important functions in the gastric,
renal, and ematic systems and are known to mediate all
inflammatory responses. Classical NSAIDs, such as
Nonsteroidal anti-inflammatory drugs (NSAIDs) are an
unhomogeneous family of pharmacologically active
compounds used in the treatment of acute and chronic
inflammation, pain, and fever. However, nevertheless
NSAIDs are the most widely used drugs, their long-term
clinical employment is associated with significant side ef-
fects and the steady use determines the onset of gastro-
4
indomethacin, inhibit both isoforms of COX: COX-1,
which is constitutively expressed in most tissues and or-
gans and catalyzes the synthesis of PGs involved in the
regulation of physiological cellular activities; COX-2,
which is mainly induced by several stimuli such as cyto-
1
,2
intestinal lesions, bleeding, and nephrotoxicity.
5
–7
kines, mitogens, and endotoxins in inflammatory sites.
Therefore the discovery of new safer anti-inflammatory
drugs represents a challenging goal for such a research
area.
Thus, their therapeutical effects are mainly due to the de-
crease of proinflammatory PGs produced by COX-2,
whereas their unwanted side effects result from the inhi-
bition of constitutive COX-1 isoform.
Although several mediators support the inflammatory
processes, the main target of NSAIDs is cyclooxygenase
Recently highly selective COX-2 inhibitors belonging to
the classes of diarylheterocycles and methanesulfonani-
lides have been developed and marketed. In particular
the class of Coxibs is characterized by 1,2-diaryl substi-
tuted heterocycles often bearing a sulfonyl moiety at the
8
*
Corresponding author. Tel.: +33 90 6766408; fax: +39 90 355613;
e-mail: ottana@pharma.unime.it
0
968-0896/$ - see front matter ꢀ 2005 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmc.2005.04.058

4244
R. Ottan a` et al. / Bioorg. Med. Chem. 13 (2005) 4243–4252
S
i
C H NH
H C NCS
+
3
7
2
C H NH
NHC H7
3
5
6
6
5
1
ii
C H
7
C H
3
C H
N
6
5
H C
6
N
3
7
5
N
N
+
S
S
O
O
2
a
2b
3 3
N, CHCl , rt, 4 h; (ii)
Scheme 1. Reagents and conditions: (i) Et
ClCH COCl, Et N, CHCl , rt, 12 h.
2
3
3
0
starting product was N-propyl-N -phenylthiourea (1)
obtained by the reaction of propylamine and phenyliso-
thiocyanate in CHCl at room temperature for 4 h fol-
3
lowed by work-up under acidic conditions (Scheme 1).
Subsequent synthesis of 2-imino-4-thiazolidinones 2 was
performed by condensation of thiourea 1 with chloro-
acetyl chloride in the presence of triethylamine in CHCl3
at room temperature. The reaction provided a mixture
of two isomers arylimino 2a and propylimino 2b in
about 1:2 ratio. They conceivably originated from the
condensation of chloroacetyl chloride with the sulfur
atom of two different intermediate thiols generated from
1 by delocalization of the lone pairs of the two different
nitrogens on the adjacent thiocarbonyl group. The syn-
thesis of isomer 2b was favored by the intermediate thiol
involving the NH group adjacent to the electron releas-
ing group. However, the nature of solvent proved to be
determinant for the reaction course. In fact, when the
synthesis was performed in MeOH at reflux, 2a/2b ratio
changed depending on the reaction time: after 24 h the
most stable isomer (2a) was the only product.
Figure 1.
para-position of one of the aryl rings, for example, cele-
9
coxib (Celebrex ) (Fig. 1).
ꢁ
0
Some of us had long investigated 3,3 -(1,2-ethanediyl)-
bis[2-aryl-4-thiazolidinone] derivatives (I, Fig. 1), which
showed interesting stereoselective anti-inflammatory/
analgesic activities together with better gastrointestinal
safety profile than known NSAIDs, suggesting they
might preferentially interact with inducible COX-2 iso-
form. Suitable selectivity assays as well as molecular
1
0
1
0e
modeling studies had supported this hypothesis.
Within this context we were then stimulated to similarly
explore simplified 4-thiazolidinone analogues in order
to find the pharmacophore of the class. Therefore, we
synthesized 2-imino-4-thiazolidinones 2a, 2b, and 5-ary-
lidene-2-imino-4-thiazolidinones 3a–8a. These latter 1,
The suggested mechanism of the conversion of 2b into
2a is depicted in Scheme 2. The extended electronic delo-
calization of amidine system gave rise to the cleavage of
the cyclic amide bond. A possible low barrier around the
C–S r bond and the subsequent cyclization accounted
for the intramolecular rearrangement providing the
most stable arylimino isomer (2a).
3
-diaryl substituted derivatives can be considered diaryl-
heterocycle Coxib-like compounds (Fig. 1). The effect of
Cl, OCH , SCH , and SO CH substituents of the arylid-
ene moiety on the anti-inflammatory profile was consid-
ered. These substituents were selected on the basis of
previous SAR studies on bisthiazolidinones, also keeping
inmindthemostfrequentsubstituentsinCoxibsstructures.
3
3
2
3
The structure of isomers 2a and 2b was assessed by ele-
1
13
mental analysis and H and C NMR. The resonances
of propyl protons allowed us to identify each isomer.
The novel 4-thiazolidinones were tested for in vivo anti-
inflammatory activity by carrageenan-induced paw
edema and pleurisy assays in rats.
investigating their possible action mechanism, their abil-
ity to inhibit COX-1 and COX-2 was assessed in murine
monocyte/macrophage J774 cell line. Finally the most
promising among the tested compounds (6a) was
docked into the active site of COX-2 enzyme in compar-
ison with the known selective COX-2 inhibitor SC-558
In 2a the CH CH CH protons appeared as a triplet res-
2 2 3
onating to higher chemical shift (3.82 ppm) than the
same protons of 2b (3.27 ppm) due to the higher deshiel-
1
1,12
With the aim of
1
3
C H
-
N
C H
6 5
C H
N
6
5
C
3
H
7
3
7
N
+
N
S
S
O
O
1
4
2b
(
Fig. 1).
C H
C H
-
C H
3
7
N
C H
C H
N
5
6
3
7
C H
N
6
5
6
5
3
7
N
N
-
N
S
S
O
2. Chemistry
S
O
+
O
+
2a
Compounds 2 and 3a–8a were synthesized according to
the synthetic pathway described in Schemes 1 and 3. The
Scheme 2. Rearrangement of 2b into 2a.

R. Ottan a` et al. / Bioorg. Med. Chem. 13 (2005) 4243–4252
4245
C H
C H N
5
3
7
C H
6
C H
N
3
7
6
5
ii
N
N
S
O
S
O
2a
Ar
3a-8a
i
Ar
3-CH OPh
C H
C H
N
6
5
3
7
N
3a
4a
3
S
4-CH SPh
O
3
5a
a
4-CH SO Ph
3
2
2b
6
4-CH OPh
3
7
8
a
a
4-ClPh
3,4-(CH O) Ph
3
2
Scheme 3. Reagents and conditions: (i) EtOH, reflux, 24 h; (ii)
ArCHO, piperidine, EtOH, reflux, 24 h.
ding effect generated by the extended electronic delocal-
ization of 3-N lone pair.
Figure 2. X-ray structure of 6a.
5-Arylidene-2-imino-4-thiazolidinones were synthesized
by the reaction in basic conditions of compounds 2
and appropriate aldehydes in refluxing ethanol for
about 24 h. In such experimental conditions deriva-
tives 3a–8a were obtained starting from pure com-
pounds 2a, 2b as well as from any mixture of them.
cursors 2a, 2b were explored to evaluate their anti-
inflammatory activity by carrageenan-induced rat paw
and pleurisy edema assays. Indomethacin and cele-
coxib were used as reference drugs.
1
1,12
1
5
Compounds 2b and 2a–7a proved to have interesting in
vivo activities in both animal models; on the contrary 8a
was devoid of any ability to inhibit the inflammation
process in such assays.
Moreover, the synthesis of 3a from 2b was monitored by
H NMR spectra of successive samples taken from the
1
reaction mixture. Starting from 2b spectrum, new sets
of signals attributable to isomer 2a appeared after 6 h,
which progressively increased. The conversion into 3a
was completed after 24 h (Scheme 3).
The injection of carrageenan into the rat paw produced
a marked increase of volume from the first hour, reach-
ing its maximal effect after 3–4 h. Pretreatment of rats
with tested compounds, except 8a, administered
(10 mg/kg, intraperitoneally) 30 min before the carra-
geenan injection, significantly attenuated inflammatory
response (Fig. 3).
The structures of 5-arylidene derivatives (3a–8a) were
assigned on the basis of analytical and spectral data.
1
In their H NMR spectra, the absence of the signal of
5
-CH protons of starting compounds 2 at 3.80 ppm to-
2
gether with the resonance of the methine hydrogen
7.70–7.75 ppm) agreed with the proposed structures.
Moreover C spectra showed clear changes in splitting
pattern of 5-C, which resonated as a triplet in 2 and as
a singlet in compounds 3a–8a.
The 3-methoxyarylidene derivative (3a) proved to be the
most effective compound, achieving an inhibition level
(85% at the third hour) similar to that of indomethacin,
while it showed a better activity profile than celecoxib,
which reached 38% edema inhibition at the same hour
(
1
3
1
7
The Z configuration of the exocyclic C@C bond was as-
signed on the basis of H NMR spectroscopy and X-ray
crystallography results. The methine proton, deshielded
1
by the adjacent C@O, was detected at 7.70–7.75 ppm in
H NMR spectra as observed for analogous 5-arylidene-
1
1
6
2
deshielding effect of 1-S, such a proton should resonate
,4-thiazolidinediones. In E isomers, due to the lesser
1
5
at lower chemical shift values.
Furthermore, the X-ray diffractometric analysis of a
representative compound (6a) unambiguously con-
firmed the Z configuration at the chiral axis (Fig. 2).
3
. Results and discussion
Figure 3. Effect of 5-arylidene-2-imino-4-thiazolidinones on rat paw
edema development elicited by carrageenan in the rat. Data are means
of mean ± SEM from n = 10 rats for each group.
Compounds 3a–8a, representative of the class of 5-aryl-
idene-2-phenylimino-4-thiazolidinones, and their pre-

4246
R. Ottan a` et al. / Bioorg. Med. Chem. 13 (2005) 4243–4252
(Fig. 3). By moving the methoxy group to the para-posi-
tion (6a) the activity decreased: in fact, the edema inhi-
bition at most reached 46% at the second hour.
the lung tissue (Fig. 6). Sham animals demonstrated
no abnormalities in the pleural cavity or fluid.
Pretreatment of rats with all compounds (20 mg/kg,
i.p.), except 8a, significantly reduced the degree of lung
injury as well as PMNs infiltration (Figs. 4 and 5). In
particular, compounds 4a–6a showed effectiveness simi-
lar to those of indomethacin and celecoxib in reducing
exudate volume and PMNs infiltration. Besides, the
myeloperoxidase inhibitory effects of 5a, 4a, 3a, and 2b
were greater than those of the reference drugs (Fig. 6).
In addition, 4-substituted analogues 4a–6a, inactive at
the first hour, proved to inhibit paw edema from the sec-
ond to the fourth hour after carrageenan administra-
tion. Moreover, a comparison between derivatives 4a
(
4-SCH substituted) and 5a (4-SO CH ) indicated the
3 2 3
sulfur oxidation was slightly detrimental to the anti-
inflammatory activity (Fig. 3). The 4-chloroarylidene
analogue (7a) was shown to be less active (20–30% inhi-
bition) than the other 4-substituted derivatives.
In addition the levels of PGE in pleural exudates, which
2
are considered as a measurement of the COX-2 activ-
1
ity, were assessed (Fig. 7).
2
Such preliminary findings stimulated us to carry out the
carrageenan-induced pleurisy assay that could provide
additional information. In this model all carrageenan-
injected rats developed an acute pleurisy producing tur-
bid exudates (Fig. 4) that contained a large amount of
polymorphonuclear cells (PMNs) (Fig. 5). Neutrophils,
measured as myeloperoxidase activity, also infiltrated
Compounds 3a, 4a, and 6a significantly reduced PGE2
levels without reaching the level of effectiveness of cele-
coxib. On the contrary 5a, which showed the best anti-
inflammatory profile in the pleurisy model, did not
reduce PGE level in pleural exudates.
2
Figure 4. Effect of 5-arylidene-2-imino-4-thiazolidinones on carrageenan-induced pleurisy assay. Exudate volume in pleural cavity at 4 h after
carrageenan injection. Data are means of mean ± SEM from n = 10 rats for each group.
Figure 5. Effect of 5-arylidene-2-imino-4-thiazolidinones on carrageenan-induced accumulation of polymorphonuclear cells in pleural cavity at 4 h
after carrageenan injection. Data are means of mean ± SEM from n = 10 rats for each group.

R. Ottan a` et al. / Bioorg. Med. Chem. 13 (2005) 4243–4252
4247
Figure 6. Effect of 5-arylidene-2-imino-4-thiazolidinones on myeloperoxidase activity in the lung of carrageenan-treated rats sacrificed after 4 h after
carrageenan injection. Lung myeloperoxidase activity is expressed as mU/100 mg wet tissue. Data are means of mean ± SEM from n = 10 rats for
each group.
2 2
Figure 7. Effect of 5-arylidene-2-imino-4-thiazolidinones on PGE levels in the pleural exudate from carrageenan-treated rats. PGE exudate levels
are expressed as pg/rat. Data are means ± SEM of 10 rats for each group.
In order to clarify whether such results were related
to the COX-1/COX-2 inhibition, in vitro assays
were performed in murine macrophage J774 cell
(6a) showed higher COX-2 inhibitory potency than the
3-substituted analogue (3a) at the same concentration.
Isosteric replacement of the oxygen atom with sulfur
resulted in compound 4a, which displayed irrelevant dif-
ferences in inhibitory potency and selectivity with re-
1
3
line (Table 1) at concentrations not higher than
0 lM because of their low solubility in the culture
media.
1
spect to 6a. Surprisingly, the oxidation of SCH group
3
(
to what takes place in well-known selective COX-2
inhibitors. These in vitro results suggested that com-
pounds displaying COX-2 inhibitory activity also pro-
4a) led to a selective COX-1 inhibitor (5a), contrary
In this assay compounds 2a and 2b produced only a
weak inhibition of COX-1 isoform at all tested doses,
without inhibiting COX-2.
8
duced appreciable reduction of PGE levels in pleural
2
The introduction of a (Z)-5-arylidene group generally
gave rise to the inhibition of COX-2 while the potency
increased without reaching, however, the levels of the
reference drugs. In particular, the insertion of a substitu-
ent at the para-position of the arylmethylidene group
appeared to be useful for COX-2 inhibition, except for
exudates (Fig. 7).
The replacement of the 4-methoxy substituent with an
electron-withdrawing group, such as chlorine, afforded
compound 7a, which showed a profile comparable to
that of the parent analogue (6a). The coexistence of
two methoxy groups in meta- and para-positions of
5
a. In fact, the compound bearing the 4-OCH group
3

4248
R. Ottan a` et al. / Bioorg. Med. Chem. 13 (2005) 4243–4252
Table 1. In vitro assays for inhibition of COX-1 and COX-2
a
COX-1 (%)
a
COX-2 (%)
a
COX-1 (%)
a
COX-2 (%)
Compd
Concn (lM)
Compd
Concn (lM)
2
2
3
4
a
b
a
a
0.1
1
2
10
12
0
0
0
5a
0.1
1
23
42
51
3
11
13
1
1
1
1
1
0
10
0.1
1
1
5
8
0
0
0
6a
0.1
1
0
0
0
16
35
55
0
10
0.1
1
0
0
0
5
13
24
7a
0.1
1
0
0
0
20
30
55
0
10
0.1
1
0
0
0
10
25
42
8a
0.1
1
0
0
0
6
10
22
0
10
Celecoxib
0.1
1
4
31
62
61
73
93
Indomethacin
0.1
1
37
40
58
43
63
69
0
10
a
Percentage COX-1 and COX-2 inhibition in murine monocyte/macrophage J774 cells at the reported doses (n = 5). Assays were repeated so that
standard errors were less than 5%.
the arylidene moiety (8a) was also shown to be detri-
mental for in vitro activity.
4
. Binding orientation in the COX-2 active site
Using as macromolecule target the X-ray structure of
1
4
the enzyme from SC-558/COX-2 complex, computa-
tional docking simulations were performed to explore
the binding mode of this new class of anti-inflammatory
drug. The software used for our calculations was Auto-
1
8
Dock 3.0 and all ligand atoms but no protein atoms
9
were allowed to move during the calculations.
1
To validate the docking protocol, the ligand conforma-
tion of SC-558 was extracted from the crystal structure
of the corresponding COX-2/ligand complex and later
docked back into the binding pocket. AutoDock was
able to perfectly reproduce the experimental position
of the ligand (62 runs of 100, with a binding free energy
of ꢀ12.09 kcal/mol), confirming the ability of the meth-
od to accurately predict the binding conformation.
Figure 8. AutoDock-predicted binding mode of the ligand 6a (purple)
compared to the crystallized position of SC-558 (green). Residues of
the binding site with important contacts to the ligands are shown,
24
while yellow dashed lines indicate hydrogen bonds.
Then, the most active COX-2 inhibitor of the series,
compound 6a, was chosen for our docking experiments.
Compound 6a clearly preferred a single binding position
in the typical binding pocket (channel-site) of COX-2
residues of the binding pocket, that is, His 90, Val 349,
Leu 352, Ser 353, Leu 359, Tyr 385, Ala 516, Phe 518,
Val 523, Gly 526, Ala 527, Ser 530, and Leu 531.
1
4
(
ꢀ
48 runs of 100, with a binding free energy of
11.74 kcal/mol), which was the best prediction since
Previous studies had shown that the COX-2 binding site
was 25% larger than the COX-1 due to the substitution
of a single amino acid (valine for isoleucine at position
523), so originating a secondary internal pocket not
it had the lowest energy among the clusters. Comparison
of the experimental position of SC-558 with 6a showed
that the following parts of these molecules are perfectly
overlapped: the methoxyphenyl, the 3-phenylimino, and
the propyl groups of 2-imino-4-thiazolidinone system of
2
0
accessible in COX- 1.
6
a are superimposed to the phenyl sulfonamide and
-bromophenyl moieties and to the trifluoromethyl
Authors suggested that the selectivity of SC-558 for
COX-2 over COX-1 might result from additional con-
tacts with residues along this deeper binding region. In
fact, the structural analysis of SC-558/COX-2 complex
indicated that the selective inhibitor molecule interacted
with His 90, Gln 192, and Arg 513, which are located at
the bottom of the COX-2 gorge. It is worth noting that
4
group of SC-558, respectively (Fig. 8).
In such orientation, the bound conformation of 6a was
engaged in hydrogen-bonds with Arg 120 and Tyr 355
and was involved in van der Waals contacts with the

R. Ottan a` et al. / Bioorg. Med. Chem. 13 (2005) 4243–4252
4249
1
13
compound 6a was able to reach such a gorge rationaliz-
ing the COX-2 selectivity (Fig. 8).
(300 MHz for H and 75 MHz for C) with deu-
terochloroform as solvent and internal standard
unless otherwise indicated. The chemical shifts are
given in d values and the coupling constants (J) in hertz
(Hz).
5. Conclusions
We report on a novel series of 5-arylidene-2-imino-4-
thiazolidinones designed as simplified analogues of pre-
viously reported anti-inflammatory bisthiazolidinones
Unless stated otherwise, all materials were obtained
from commercial suppliers and used without further
purification.
1
0
(
I) and, at the same time, as Coxib-like compounds.
They generally retained the in vivo activity already
showed by the bisthiazolidinone class and could be a
novel class of anti-inflammatory agents. Although a
wide SAR picture is not yet available, some consider-
ations can be focused. The attempt to simplify the par-
ent structure (I) and identify its pharmacophoric part
initially led to compounds 2, which retained good anti-
inflammatory profiles when a differently substituted
arylidene moiety was inserted in position 5 (3a–8a). In
particular: (a) the OCH , SCH , and SO CH groups
Crystals suitable for X-ray analysis were obtained by
recrystallization of compound 6a from methanol solu-
tion. Diffraction data were collected at room tempera-
3
ture from a colorless 0.92 · 0.75 · 0.36 mm prismatic
crystal sample by using a Siemens P4 automated four-
cycle single-crystal diffractometer with graphite-
˚
monochromated MoKa radiation (k = 0.71073 A).
Crystallographic data (excluding structure factors) for
compound 6a have been deposited with the Cambridge
Crystallographic Data Centre as supplementary publica-
tion no. CCDC 216008. Copies of the data can be
obtained, free of charge, on application to CCDC,
12 Union Road, Cambridge CB2 1EZ, UK (fax:
+44-1223-336033; e-mail: deposit@ccdc.cam.ac.uk).
3
3
2
3
introduced in the para-position of the arylidene moiety
were generally found to be favorable for activity both
in vivo and in vitro experiments; (b) in addition, while
the introduction of the SO CH group generally leads
2
3
to COX-2 selective inhibitors, in the compounds herein
reported such a moiety proved to be a determinant of
COX-1 inhibition; (c) the 4-Cl substitution decreased
the in vivo anti-inflammatory activity while the degree
of selective COX-2 inhibition was retained; (d) when
0
6.2. Synthesis of N-phenyl-N -propylthiourea (1)
Propylamine (0.65 g, 0.011 mol) was added to a solution
of phenylisothiocyanate (1.53 g, 0.01 mol) dissolved in
chloroform (30 mL). The solid was stirred for about
4 h at room temperature and concentrated under re-
duced pressure. The precipitate was filtered and crystal-
lized from cold ethanol to give 1.69 g (87%) of the title
compound as a white solid (mp 59–60 ꢂC): IR (Nujol)
the OCH substituent moved to position 3 (3a), a good
3
in vivo anti-inflammatory profile was achieved, reaching
indomethacin levels in the carrageenan-induced rat paw
edema, without displaying any preference for each one
of the COX isoforms; (e) when steric hindrance in-
creased (8a), both in vitro and in vivo activities disap-
peared; (f) flexible docking experiments showed that 6a
and SC-558 have a similar binding mode to the COX-
ꢀ1
1
3240 (NH), 1170 (C@S) cm ; H NMR (CDCl ) d
0.89 (t, 3H, CH , J = 7.5 Hz), 1.57 (m, 2H,
CH CH CH ), 3.56 (t, 2H, CH CH CH , J = 6.9 Hz),
3
3
3
2
2
3
2
2
2
active site with good overlapping.
7.42–7.43 (m, 5H, arom); Anal. (C H N S) C, H, N.
10 14 2
The observed COX inhibition did not appear to be the
only action mode of this class of anti-inflammatory
agents. This is not surprising since inflammation is a
complex physiopathological reaction and several media-
tors are involved in its development and progression.
Consequently further investigations will be performed
in order to clarify the anti-inflammatory mechanism of
action of these compounds.
6.3. Synthesis of 2-imino-4-thiazolidinones
A solution of chloroacetyl chloride (1.69 g, 0.015 mol) in
chloroform (20 mL) was added dropwise to a stirred
solution of 1 (1.94 g, 0.01 mmol) and triethylamine
(2.02 g, 0.02 mol) in chloroform (30 mL) and the mix-
ture was stirred overnight at room temperature.
Removal of the solvent gives an oily residue, which
was dissolved in ethyl ether and repeatedly washed with
an aqueous sodium carbonate (10%). The organic layer
was dried over sodium sulfate, filtered, and
concentrated.
6. Experimental
6
.1. Chemistry
The crude mixture was purified by silica gel column
chromatography (eluent petroleum ether/ethyl ether,
3/2) to give 2a (0.73 g, 31%) and 2b (1.45 g, 62%), each
as a white solid.
Melting points were determined with a Kofler–Reichert
hot-stage apparatus and are uncorrected. Precoated
silica gel plates F 254 for analytical controls and
chromatographic columns 70–230 mesh SiO for separa-
2
tions were used. Elemental analyses (C, H, N), deter-
mined by means of a C. Erba mod. 1106 elem.
analyzer, were within ±0.4% of theory. IR spectra were
registered as Nujol mulls on a Perkin Elmer mod. 683
Spectrophotometer. The NMR spectra were obtained
on a Varian 300 magnetic resonance spectrometer
6.3.1. 2-Phenylimino-3-propyl-4-thiazolidinone (2a). mp
1
56–58 ꢂC; H NMR (CDCl ) d 0.97 (t, 3H, CH ,
3
3
J = 7.5 Hz), 1.75 (m, 2H, CH CH CH ), 3.80 (s, 2H,
5-CH ), 3.82 (t, 2H, CH CH CH , J = 7.5 Hz), 6.97–
3
2
2
2
3
2
2
1
3
7.36 (m, 5H, arom); C NMR (CDCl ) d 11.1 (CH );
3
3
20.4 (CH CH CH ); 32.5 (5-CH ); 44.5 (CH CH CH );
3
2
2
2
3
2
2

4250
R. Ottan a` et al. / Bioorg. Med. Chem. 13 (2005) 4243–4252
1
(
54.0 (2-C); 120.7, 124.3, 129.0, 137.8 (CH arom); 171.6
CO); Anal. (C H N OS) C, H, N.
J = 7.2 Hz); 1.86 (m, 2H, CH CH CH ); 4.03 (t, 2H,
3 2 2
CH CH CH , J = 7.4 Hz); 7.07–7.45 (m, 9H, arom);
3 2 2
1
2
14
2
1
3
7
.73 (s, 1H, CH); C NMR (CDCl ) d 11.3 (CH );
3 3
6
.3.2. 3-Phenyl-2-propylimino-4-thiazolidinone (2b). mp
20.8 (CH CH CH ); 44.9 (CH CH CH ); 121.1, 124.9,
3 2 2 3 2 2
1
146–148 ꢂC; H NMR (CDCl ) d 0.91 (t, 3H, CH ,
J = 7.5 Hz), 1.61 (m, 2H, CH CH CH ), 3.27 (s, 2H,
CH CH CH , J = 6.9 Hz), 3.97 (s, 2H, 5-CH ), 7.27–
2
7
2
1
129.1, 129.3, 129.4 (CH arom); 122.6 (5-C); 131.0 (CH
methylidene); 132.4, 135.6, 148.2, 149.9 (Cq arom);
166.6 (CO); Anal. (C H ClN OS) C, H, N.
3
3
3
2
2
3
2
2
19 17
2
1
3
.52 (m, 5H, arom); C NMR (CDCl ) d 12.9 (CH );
3 3
4,6 (CH CH CH ); 33.7 (5-CH ); 55.3 (CH CH CH );
52.8 (2-C); 129.0, 129.7 130.2, 136.2 (CH arom); 174.5
6.4.5. 5-(3,4-Dimethoxyphenyl)methylidene-2-phenylimino-
3-propyl-4-thiazolidinone (8a). Yield 68%; mp 170–
3
2
2
2
3
2
2
1
(
CO); Anal. (C H N OS) C, H, N.
1
172 ꢂC. H NMR (CDCl ) d 1.08 (t, 3H, CH , J =
2
14
2
3
3
7
.2 Hz); 1.83 (m, 2H, CH CH CH ); 3.94 (t, 2H,
3 2 2
6
2
6
.4. General procedures for the synthesis of 5-arylidene-
-phenylimino-3-propyl-4-thiazolidinones (3a, 4a, and
a–8a)
CH CH CH , J = 7.5 Hz); 3.96, 3.99 (2s, 6H, OCH );
3 2 2 3
13
7.08–7.63 (m, 8H, arom); 7.80 (s, 1H, CH); C NMR
(CDCl ) d 11.2 (CH ); 20.8 (CH CH CH ); 44.8
3
3
3
2
2
(CH CH CH ); 55.9, 56.0 (OCH ); 111.4, 113.2, 121.3,
3 2 2 3
To a solution of 2a and/or 2b compounds (0.7 g,
.03 mol) and piperidine (0.35 g, 0.04 mol) in ethanol
35 mL) the appropriate benzaldehyde (0.03 mol) was
added and the mixture was heated at reflux for 20–
4 h. After this time, the solution was concentrated in
123.4, 124.9, 130.9 (CH arom); 119.2 (5-C); 129.3 (CH
methylidene); 119.2, 126.8, 149.2, 150.6 (Cq arom);
166.9 (CO); Anal. (C H N O S) C, H, N.
0
(
2
1
22
2
3
2
6.5. Synthesis of 5-(4-methylsulfonylphenyl)methylidene-
2-phenylimino-3-propyl-4-thiazolidinone (5a)
vacuo and kept at ꢀ15 ꢂC over 2 h. The residue was
filtered off and recrystallized from methanol.
A solution of 4a (0.5 g, 0.2 mol) in 30 mL acetic acid
(50%) was treated dropwise with 30 g (0.3 mol) H O
6
.4.1. 5-(3-Methoxyphenyl)methylidene-2-phenylimino-3-
2
2
propyl-4-thiazolidinone (3a). Yield 70%; mp 115–118 ꢂC.
(35%) and stirred for 3 h at reflux. The mixture was then
added to water (40 mL) and cooled for 24 h on ice, be-
1
H NMR (CDCl ) d 1.02 (t, 3H, CH , J = 7.5 Hz); 1.84
3
3
1
(
CH CH CH , J = 7.2 Hz); 7.28–7.42 (m, 9H, arom);
m, 2H, CH CH CH ); 3.80 (s, 3H, OCH ); 3.98 (t, 2H,
fore being filtered-off. Yield 78%; mp 127–128 ꢂC. H
3
2
2
3
NMR (CDCl ) d 1.04 (t, 3H, CH , J = 7.5 Hz); 1.85
3
2
2
3
3
1
3
7
.71 (s, 1H, CH); C NMR (CDCl ) d 11.2 (CH );
(m, 2H, CH CH CH ); 3.06 (s, 3H, SO CH ); 4.04 (t,
2H, CH CH CH , J = 7.5 Hz); 7.96–7.63 (m, 9H,
3 2 2
3
3
3 2 2 2 3
2
0.75 (CH CH CH ); 44.7 (CH CH CH ); 55.2
3
2
2
3
2
2
1
3
(
(
OCH ); 115.2, 115.4, 121.1, 122.0, 124.7, 129.3, 129.9
3
CH arom); 118.9 (5-C); 130.4 (CH methylidene);
arom); 7.70 (s, 1H, CH); C NMR (CDCl ) d 11.3
3
(CH ); 15.0 (SO CH ); 20.8 (CH CH CH ); 44.8
(CH CH CH ); 120.6 (5-C); 121.3, 124.9, 126.0, 129.4,
3 2 2
3
2
3
3
2
2
1
(
26.6, 148.4 (Cq arom); 166.6 (CO); Anal.
C H N O S) C, H, N.
130.2 (CH arom); 130.3 (CH methylidene);
41.7, 148.2, 150.7 (Cq arom); 166.9 (CO); Anal.
(C H N O S ) C, H, N.
2
0
20
2
2
1
6
3
1
.4.2. 5-(4-Methylthiophenyl)methylidene-2-phenylimino-
-propyl-4-thiazolidinone (4a). Yield 84%; mp 130–
2
0
20
2
3 2
1
32 ꢂC. H NMR (CDCl ) d 1.06 (t, 3H, CH ,
3
3
J = 7.4 Hz); 1.87 (m, 2H, CH CH CH ); 2.52 (s, 3H,
SCH ); 4.04 (t, 2H, CH CH CH , J = 7.5 Hz); 7.06–
7. Pharmacology
3
2
2
3
3
2
2
1
3
7
(
(
.47 (m, 9H, arom); 7.75 (s, 1H, CH); C NMR
CDCl3) 11.3 (CH3); 15.0 (SCH3); 20.8
CH CH CH ); 44.8 (CH CH CH ); 120.5 (5-C);
7.1. Assessment of COX-1/COX-2 activity
d
7.1.1. Compounds tested. Stock solutions of test com-
pounds were prepared in ethanol; an equivalent amount
of ethanol was included in control samples.
3
2
2
3
2
2
1
21.2, 124.8, 125.8, 129.3, 130.1 (CH arom); 130.2
(
(
CH methylidene) 141.6, 148.3, 150.5 (Cq arom); 166.9
CO); Anal. (C H N OS ) C, H, N.
2
0
20
2
2
7
.1.2. Cell culture. The murine monocyte/macrophage
6.4.3. 5-(4-Methoxyphenyl)methylidene-2-phenylimino-3-
propyl-4-thiazolidinone (6a). Yield 81%; mp 106–
J774 cell line was grown in DulbeccoÕs modified Eagles
medium (DMEM) supplemented with 2 mM glutamine,
25 mM Hepes, penicillin (100 U/mL), streptomycin
(100 lg/mL), 10% fetal bovine serum (FBS), and 1.2%
Na pyruvate (Bio Whittaker, Europe). Cells were plated
1
1
J = 7.2 Hz); 1.83 (m, 2H, CH CH CH ); 3.83 (s, 3H,
08 ꢂC. H NMR (CDCl ) d 1.02 (t, 3H, CH ,
3
3
3
2
2
OCH ); 3.97 (t, 2H, CH CH CH , J = 7.5 Hz); 6.94–
3
3
2
2
1
.40 (m, 9H, arom); 7.70 (s, 1H, CH); C NMR
3
5
7
in 24-well culture plates at a density of 2.5 · 10 cells/mL
7
(
CDCl ) d 11.3 (CH ); 20.8 (CH CH CH ); 44.7
or in 10 cm-diameter culture dishes (1 · 10 cells/10 mL/
3
3
3
2
2
(
1
CH CH CH ); 55.3 (OCH ); 114.6, 121.2, 124.6,
29.3, 130.4 (CH arom); 119.0 (5-C); 131.7 (CH methyl-
dish) and allowed to adhere at 37 ꢂC in 5% CO and 95%
3
2
2
3
2
O for 2 h. Immediately before the experiments, culture
medium was replaced by fresh medium without FBS in
2
idene); 126.5, 148.5, 150.7, 160.7 (Cq arom); 167.1 (CO);
Anal. (C H N O S) C, H, N.
1
1
order to avoid interference with radioimmunoassay
and cells were stimulated as described.
2
0
20
2
2
6.4.4. 5-(4-Chlorophenyl)methylidene-2-phenylimino-3-
propyl-4-thiazolidinone (7a). Yield 79%; mp 134–
7.1.3. Assessment of COX-1 activity. The in vitro screen-
ing scheme included COX-1 activation by stimulation of
1
1
36 ꢂC. H NMR (CDCl ) d 1.06 (t, 3H, CH ,
3
3

R. Ottan a` et al. / Bioorg. Med. Chem. 13 (2005) 4243–4252
4251
cells with arachidonic acid (15 lM) for 30 min inducing
Any exudate that was contaminated with blood was dis-
carded. The amount of exudate was calculated by sub-
tracting the volume injected (2 mL) from the total
volume recovered. The leukocytes in the exudate were
suspended in phosphate-buffer saline (PBS) and counted
with an optical microscope in a BurkerÕs chamber after
Trypan Blue staining.
6
a significant increase of PGE (3.5 ± 0.09 ng/10 cells) in
2
6
comparison to unstimulated cells (0.35 ± 0.07 ng/10
cells).
Cells were pretreated with test compounds for 15 min
and further incubated for 30 min with arachidonic acid
1
1
(
were collected for the measurement of PGE by radio-
15 lM). At the end of the incubation the supernatants
7.2.4. Measurement of prostaglandin E in the pleural
2
2
2
1
immunoassay. All compounds were assayed at three
concentrations (0.1, 1, and 10 lM).
exudate. The amount of prostaglandin E (PGE ) pres-
2
2
ent in the pleural fluid was measured by radioimmuno-
2
assay without prior extraction or purification.
1
7
.1.4. Assessment of COX-2 activity. Cells were stimu-
lated, for 24 h, with Escherichia coli lipopolysaccharide
LPS, 10 lg/mL), to induce COX-2. The supernatant
7.2.5. Myeloperoxidase activity. Myeloperoxidase
(MPO) activity, an indicator of polymorphonuclear leu-
kocyte (PMN) accumulation, was determined as previ-
(
of the cells was replaced with fresh medium, cells pre-
treated with test compounds, at the concentrations pre-
viously reported, for 15 min and further incubated for
2
2
ously described.
At the specified time following
injection of carrageenan, lung tissues were obtained
and weighed and each piece homogenized in a solution
containing 0.5% (w/v) hexadecyltrimethyl-ammonium
bromide dissolved in 10 mM potassium phosphate buf-
fer (pH 7) and centrifuged for 30 min at 20,000g at
4 ꢂC. An aliquot of the supernatant was then allowed
to react with a solution of tetramethylbenzidine
(1.6 mM) and 0.1 mM hydrogen peroxide. The rate of
change in absorbance was measured spectrophotometri-
cally at 650 nm. MPO activity was defined as the quan-
tity of enzyme degrading 1 lmol of peroxide/min at
3
increase in PGE production from 2.07 ± 0.04 to
0 min with arachidonic acid. There was a marked
2
6
7.2 ± 2.8 ng/10 cells (p <0.001). At the end of the incu-
2
bation the supernatants were collected for the measure-
ment of PGE by radioimmunoassay.
2
7
.2. Anti-inflammatory activity
7
.2.1. Animals. Male Sprague–Dawley rats (300–350 g;
Charles River, Milan, Italy) were housed in a controlled
environment and provided with standard rodent chow
and water. Animal care was in compliance with Italian
regulations on protection of animals used for experi-
mental and other scientific purposes (D.M. 116192) as
well as with the EEC regulations (O.J. of E.C. L 358/1
ꢀ1
37 ꢂC and was expressed in U 100 mg of wet tissue.
7.2.6. Materials. Arachidonic acid was obtained from
SPIBIO, Paris, France. [ H-PGE ] was from NEN Du
3
2
Pont (Milan, Italy). Bacterial lipopolysaccharide from
Salmonella thyphosa (LPS) and all other reagents and
compounds used were obtained from Sigma–Aldrich,
Milan, Italy.
1
2/18/1986).
7
.2.2. Carrageenan-induced paw edema. Rats received
a subplantar injection of 0.2 mL saline containing 2%
k-carrageenan in the right hindpaw. A suspension of
tested compounds (10 mg/kg), or an equivalent volume
of vehicle (DMSO 50%), were administered intraperito-
neally 30 min before carrageenan. Control animals
received the same volume of vehicle. The volume of the
paw was measured by plethysmometry (model 7140,
Ugo Basile) immediately after the injection. Subsequent
readings of the volume of the same paw were carried out
at 60-min intervals and compared to the initial readings.
7.2.7. Data analysis. All values in the figures and text are
expressed as means ± standard error of the mean (SEM)
for n observations. For the in vitro studies, the data rep-
resent the number of wells studied (6 to 9 wells from 2 to
3 independent experiments). For the in vivo studies, n
represents the number of animals studied. The results
were analyzed by one-way analysis of variance (ANO-
VA) followed by a Bonferroni post-hoc test for multiple
comparisons. A p value less than 0.05 was considered
significant. In the experiments involving histology, the
figures shown are representative of at least three exper-
iments performed on different experimental days.
7
.2.3. Carrageenan-induced pleurisy. Rats were anaesthe-
tized with isoflurane and submitted to a skin incision at
the level of the left sixth intercostal space. The underly-
ing muscle was dissected and saline (0.2 mL) or saline
containing 2% k-carrageenan (0.2 mL) was injected into
the pleural cavity. The skin incision was closed with a
suture and the animals allowed to recover. A suspension
of tested compounds (20 mg/kg), or an equivalent vol-
ume (0.6 mL) of vehicle (DMSO 50%), were adminis-
tered intraperitoneally 30 min before carrageenan. At
7.3. Computational studies
7.3.1. Structures. (1) Protein: The X-ray structure of
COX-2 complexed with the inhibitor SC-558 was re-
trieved from the Brookhaven Protein Data Bank (entry
1
4
code 6COX) and used as target for our modeling stud-
ies. Since the positions of most water molecules in the
crystal structure complex are unlikely to be conserved,
water molecules were excluded before the docking pro-
tocols. Residues with missing side chain atoms were re-
built using the Biopolymer module integrated into
4
h after the injection of carrageenan, the animals were
killed by inhalation of CO . The chest was carefully
2
opened and the pleural cavity rinsed with 2 mL of saline
solution containing heparin (5 U/mL) and indomethacin
1
9
18
(10 lg/mL). The exudate and washing solution were re-
moved by aspiration and the total volume measured.
Insight II. Using the AutoDockTools package,
non-polar hydrogen atoms were added to the protein,

4252
R. Ottan a` et al. / Bioorg. Med. Chem. 13 (2005) 4243–4252
lone pairs were merged, Kollman united-atom partial
charges were assigned and finally solvent parameters
were enclosed.
6. Herschman, H. R. Biochim. Biophys. Acta 1996, 1299,
25–140.
1
7
8
9
. Dubois, R. N.; Abramson, S. B.; Crofford, L.; Gupta, R.
A.; Simon, L. S.; Van De Putte, L. B.; Lipsky, P. E.
FASEB J. 1998, 12, 1063–1073.
. De Leval, X.; Delarge, J.; Somers, F.; De Tullio, P.;
Henrotin, Y.; Pirotte, B.; Dogne, J. M. Curr. Med. Chem.
(
2) Ligands: Coordinates for SC-558 were taken directly
from the X-ray structure of its complex with COX-2.
The structure of 6a was constructed using standard
bond lengths and angles from the Sybyl fragment library
and fully optimized by semiempirical quantum mechan-
2000, 7, 1041–1062.
. Celecoxib Drugs Future 1997, 22, 711–714.
10. (a) Previtera, T.; Vigorita, M. G.; Basile, M.; Fenech, G.;
Trovato, A.; Occhiuto, F.; Monforte, M. T.; Barbera, R.
Eur. J. Med. Chem. 1990, 25, 569–579; (b) Vigorita, M. G.;
Previtera, T.; Basile, M.; Fenech, G.; Costa De Pasquale,
R.; Occhiuto, F.; Circosta, C. Il Farmaco 1988, 4, 373–379;
2
3
ical method AM1. The AutoDockTools was used to
identify the aromatic carbons, assign the rigid root,
and the active torsions.
(
c) Vigorita, M. G.; Previtera, T.; Ottan a` , R.; Grillone, I.;
7
.3.2. Docking calculations. Docking studies to COX-2
Monforte, F.; Monforte, M. T.; Trovato, A.; Rossitto, A.
Il Farmaco 1997, 52, 43–48; (d) Vigorita, M. G.; Ottan a` ,
R.; Monforte, F.; Maccari, R.; Monforte, M. T.; Trovato,
A.; Taviano, M. F.; Miceli, N.; De Luca, G.; Alcaro, S.;
Ortuso, F. Bioorg. Med. Chem. 2003, 11, 999–1006; (e)
Ottan a` , R.; Mazzon, E.; Dugo, L.; Monforte, F.; Maccari,
R.; Sautebin, L.; De Luca, G.; Vigorita, M. G.; Alcaro, S.;
Ortuso, F.; Caputi, A. P.; Cuzzocrea, S. Eur. J. Pharma-
col. 2002, 448, 71–80.
were carried out using the package AutoDock 3.0, a pro-
gram that allows torsional flexibility in the ligand, while
the protein is kept rigid. The grid spacing was 0.375 A
in each dimension, and each grid map consists of
1
8
˚
6
0 · 60 · 60 grid points. The grid was centered on the
selective site, using the SC-558 crystallographic position
as reference. The AutoGrid program generated separate
grid maps for all atom types of the ligand structures plus
one for electrostatic interactions. We used the so-called
Lamarckian genetic algorithm (LGA), which combines
a traditional genetic algorithm (GA) with a local search
method. For both ligands, each LCG job consists of
1
1
1. Di Rosa, M.; Willoughby, D. A. J. Pharm. Pharmacol.
1971, 23, 297–298.
2. Cuzzocrea, S.; Zingarelli, B.; Gilard, E.; Hake, P.;
Salzman, A. L.; Szab o` , C. Free Radical Biol. Med. 1998,
4, 450–459.
2
1
position and orientation chosen randomly. The binding
00 independent runs, starting each time from a different
1
3. Zingarelli, B.; Southan, G. J.; Gilad, E.; OÕConnor, M.;
Salzman, A. L.; Szabo, C. Br. J. Pharmacol. 1997, 120,
357–366.
˚
modes were clustered using rmsd cutoff of 1.0 A with re-
spect to the starting position. All the other values were
used as default settings and the results were ranked by
the lowest energy representative of each cluster com-
bined with the root mean square deviation.
14. Kurumbail, R. G.; Stevens, A. M.; Gierse, J. K.;
McDonald, J. J.; Stegeman, R. A.; Pak, J. Y.; Gildehaus,
D.; Miyashiro, J. M.; Penning, T. D.; Seibert, K.; Isakson,
P. C.; Stallings, W. C. Nature 1996, 384, 644–648.
1
5. (a) Momose, Y.; Meguro, K.; Ikeda, H.; Hatanaka, C.; Oi,
S.; Sohda, T. Chem. Pharm. Bull. 1991, 39, 1440–1445; (b)
Ispida, T.; In, Y.; Inoue, M.; Tanaka, C.; Hamanaka, N.
J. Chem. Soc., Perkin Trans. 1990, 1085–1091.
Acknowledgment
1
6. Bruno, G.; Costantino, L.; Curinga, C.; Maccari, R.;
Monforte, F.; Nicol o` , F.; Ottan a` , R.; Vigorita, M. G.
Bioorg. Med. Chem. 2002, 10, 1077–1084.
This work was financially supported by the Fondo
Ateneo di Ricerca (University of Messina, Italy).
1
7. Abdel-Salam, O. M. E.; Baiuomy, A. R.; Arbid, M. S.
Pharm. Res. 2004, 49, 119.
References and notes
18. Morris, G. M.; Goodsell, D. S.; Halliday, R. S.; Huey, R.;
Hart, W. E.; Belew, R. K.; Olson, A. J. J. Comput. Chem.
1998, 19, 1639–1662.
19. InsightII 2000. Accelrys, Inc., San Diego, CA, http://
www.accelrys.com.
20. Gierse, J. K.; Mc Donald, J. J.; Hauser, S. D.; Rangwala,
S. H.; Kobolt, C. M.; Seibert, K. J. Biol. Chem. 1996, 271,
15810–15814, and references cited therein.
21. Sautebin, L.; Ialenti, A.; Ianaro, A.; Di Rosa, M. Br.
J. Pharmacol. 1995, 114, 323–328.
1
. Bombardier, C.; Laine, L.; Reicin, A.; Shapiro, D.;
Burgos-Vargas, R.; Davis, B.; Day, R.; Bosi Ferraz, M.;
Hawkey, C. J.; Hochberg, M. C.; Kvien, T. K.; Schnitzer,
T. J. N. Eng. J. Med. 2000, 343, 1520.
. Silverstein, F. E.; Faich, G.; Goldstein, J. L.; Simon, L. S.;
Pincus, T.; Whelton, A.; Makuch, R.; Eisen, G.; Agrawal,
N. M.; Stenson, W. F.; Burr, A. M.; Zhao, W. W.; Kent, J.
D.; Lefkowith, J. B.; Verburg, K. M.; Geis, G. S. JAMA
2
2
000, 284, 1247.
22. Mullane, K. M.; Westlin, W.; Kraemer, R. Ann. N.Y.
Acad. Sci. 1988, 524, 103–113.
23. SYBYL 6.9. Tripos Associates Inc., St. Louis, Missouri,
USA.
3
4
. Vane, J.; Botting, R. FASEB J. 1987, 1, 89–96.
. Smith, W. L.; Dewitt, D. L. Adv. Immunol. 1996, 62, 167–
215.
5
. Fu, J. Y.; Masferrer, J. L.; Seibert, K.; Raz, A.; Needle-
man, P. J. Biol. Chem. 1990, 265, 16737–16740.
24. DeLano, W. L. The PyMol Molecular Graphics System
(2002) on World Wide Web http://www.pymol.org.