Synthesis, docking simulation, biological evaluations and 3D-QSAR study of 5-Aryl-6-(4-methylsulfonyl)-3-(metylthio)-1,2,4-triazine as selective cyclooxygenase-2 inhibitors

Article abstract of DOI:10.1016/j.bmc.2013.12.002

A series of 5-Aryl-6-(4-methylsulfonyl)-3-(metylthio)-1,2,4-triazine derivatives were synthesized and their COX-1/COX-2 inhibitory activity as well as in vivo anti-inflammatory and analgesic effects were evaluated. All of compounds showed strong inhibitio

Full text of DOI:10.1016/j.bmc.2013.12.002

Bioorganic & Medicinal Chemistry 22 (2014) 865–873
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry
journal homepage: www.elsevier.com/locate/bmc
Synthesis, docking simulation, biological evaluations and 3D-QSAR
study of 5-Aryl-6-(4-methylsulfonyl)-3-(metylthio)-1,2,4-triazine as
selective cyclooxygenase-2 inhibitors
Hamid Irannejad a, Abbas Kebriaieezadeh b, Afshin Zarghi c, Farhad Montazer-Sadegh b, Abbas Shafiee d,
Amir Assadieskandar e, Mohsen Amini e,
⇑
a Department of Medicinal Chemistry and Pharmaceutical Sciences Research Center, Faculty of Pharmacy, Mazandaran University of Medical Sciences, Sari, Iran
b Department of Pharmacology Toxicology, Faculty of Pharmacy, Tehran University of Medical Sciences, Tehran, Iran
c Department of Medicinal Chemistry, Faculty of Pharmacy, Shahid Beheshti University of Medical Sciences, Tehran, Iran
d Department of Medicinal Chemistry, Faculty of Pharmacy and Pharmaceutical Sciences Research Center, Tehran University of Medical Sciences, Tehran 14176, Iran
e Department of Medicinal Chemistry, Faculty of Pharmacy and Drug Design Development Research Center, Tehran University of Medical Sciences, Tehran 14176, Iran
a r t i c l e i n f o
a b s t r a c t
Article history:
A series of 5-Aryl-6-(4-methylsulfonyl)-3-(metylthio)-1,2,4-triazine derivatives were synthesized and
their COX-1/COX-2 inhibitory activity as well as in vivo anti-inflammatory and analgesic effects were
evaluated. All of compounds showed strong inhibition of COX-2 with IC50 values in the range of
Received 18 September 2013
Revised 1 December 2013
Accepted 2 December 2013
Available online 11 December 2013
0.1–0.2 lM and in most cases had stronger anti-inflammatory and analgesic effects than indomethacin
at doses 3 and 6 mg/kg. Among them, 5-(4-chlorophenyl)-6-(4-(methylsulfonyl) phenyl)-3-(methyl-
thio)-1,2,4-triazine (9c) was the most potent and selective COX-2 compound; its selectivity index of
395 was comparable to celecoxib (SI = 405). Evaluation of anti-inflammatory and analgesic effects of
9c showed its higher potency than indomethacin and hence could be considered as a promising lead can-
didate for further drug development. Furthermore, the affinity data of these compounds were rational-
ized through enzyme docking simulation and 3D-QSAR study by k-Nearest Neighbour Molecular Field
Analysis.
Keywords:
Cyclooxygenase
Inflammatory and analgesic properties
Docking
3D-QSAR
1,2,4-Triazine
Ó 2013 Elsevier Ltd. All rights reserved.
1. Introduction
COX-2 while the unwanted side effects arise from inhibition of
COX-1 activity. Accordingly, in recent years a great deal of interest
Cyclooxygenase (COX) is an endogenous enzyme that catalyzes
the first step in the conversion of arachidonic acid into prostaglan-
dins and thromboxanes. COX is competitively inhibited by a group
of drugs known as nonsteroidal anti-inflammatory drugs (NSAIDs)
that are used for therapy and management of inflammation and
pain. In spite of their beneficial action, their activity is associated
with deleterious side effects such as nephrotoxicity and gastric
ulcers.1 Consequently, new NSAIDs with fewer side effects are nec-
essary for safer treatment. A key step in the discovery of novel
NSAIDs lacking the deleterious side effects came from the charac-
terization and isolation of two different COX isoforms. Constitutive
COX-1, which is responsible for cytoprotective effects involved in
maintenance of gastric and renal functions as well as vascular
homeostasis, and inducible COX-2, which is expressed only in
inflammatory situations.2,3 So it was understood that the therapeu-
tic anti-inflammatory action of NSAIDs is produced by inhibition of
has been devoted to the discovery of selective COX-2 inhibitors.
Several selective COX-2 inhibitors have been developed and mar-
keted up to now. Among them, celecoxib, valdecoxib and etoricox-
ib have been clinically validated as anti-inflammatory therapeutics
for indications such as acute pain, rheumatoid arthritis and osteo-
arthritis,4,5 as well as in treatment of cancer cell lines6 and neuro-
degenerative disorders like Alzheimer’s and Parkinson’s diseases.7
Most of the COX-2 inhibitors are characterized by a 1,2-diaryl
substituted on
a central heterocyclic ring system. Extensive
structure–activity studies demonstrated that a SO2CH3 or SO2NH2
group at the para position of one of the aryl rings provides
optimum COX-2 selectivity and inhibitory potency. Unfortunately,
these compounds are associated with adverse cardiovascular
effects such as myocardial infarction, thrombosis and cardiac
dysfunction.8 The most plausible reason for this effect is the
suppression of COX-2 dependent prostacyclin which mediate
platelet activation and atherogenesis.9 Therefore, the search for a
novel, structurally different and better pharmacodynamic profile
of selective COX-2 inhibitors devoid of the noted side effects is still
an ongoing need for anti-inflammatory therapy.
⇑
Corresponding author. Tel.: +98 21 66959090; fax: +98 21 66461178.
E-mail address: moamini@sina.tums.ac.ir (M. Amini).
0968-0896/$ - see front matter Ó 2013 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.bmc.2013.12.002

866
H. Irannejad et al. / Bioorg. Med. Chem. 22 (2014) 865–873
We have previously reported the design and synthesis of
compound (9) was obtained through the reaction of compound
(8) in the presence of methyl iodide and triethylamine in metha-
nol14 (Scheme 1).
2-alkylthio-1,5-diarylimidazoles (1),10a 4-aryl-5-(4-(methylsulfo-
nyl)phenyl)-2-alkylthio-1H-imidazole (2),10b and 3-alkylthio-4,5-
diaryl-4H-1,2,4-triazoles (3)11 as selective COX-2 inhibitors in
which the central ring was imidazole and triazole substituted with
alkylthio groups. Herein, as part of our ongoing program, we de-
scribe synthesis, docking simulation and 3D-QSAR study of some
3-alkylthio-4,5-diaryl-1,2,4-triazine derivatives (9) as selective
COX-2 inhibitors (Fig. 1). Furthermore, the biological activity
(anti-inflammatory (AI) and analgesic activities) of the synthesized
compounds were evaluated in both in vitro and in vivo studies.
2.2. Inhibition of cyclooxygenase
The synthesized compounds (9a–e) were all evaluated for their
inhibitory activity against COX-1 and COX-2 enzymes. Analysis of
biological data showed that all compounds are strong inhibitors
of COX-2 with IC50 values in the range of 0.1–0.2 lM concentra-
tions, while compounds showed weak inhibition of COX-1 as the
IC50 values were 500-fold higher than that of COX-2 IC50 values.
The most potent compound, 5-(4-chlorophenyl)-6-(4(methylsulfo-
nyl)phenyl)-3-(methylthio)-1,2,4-triazine (9c) with the chloro
atom at the para position of the phenyl ring, has a COX-1 IC50 value
2. Results and discussion
2.1. Chemistry
of 39.5 lM and COX-2 IC50 value of 0.10 lM with a selectivity in-
dex of 395 which is comparable to the selective COX-2 inhibitor
celecoxib and its selectivity index of 405 (Table 1).
Phenylacetic acid derivatives were converted to the corre-
sponding compounds (5) in the presence of trifluoroacetic acid
anhydride, phosphoric acid and thioanisole.12 Then sulfur atom
was oxidized to the sulfonyl by oxone in THF/MeOH and the result-
ing material (6) was reacted with sodium methoxide and then bu-
tyl nitrite to produce compound (7). Triazine ring closure (8) was
performed with thiosemicarbazide in acidic water13 and the final
2.3. Docking simulation
In order to rationalize the observations of the biological activity
of the synthesized compounds and provide more insights into the
O
S
O
O
O
O
O
S
S
CH3
H2N
H2N
H3C
N
N
Cl
CF3
O
N
N
H3C
N
H3C
valdecoxib
O
celecoxib
etoricoxib
O
O
O
O
O
S
S
S
H3C
H3C
H3C
N
N
N
SR
SR
N
N
N
N
H
SR
X
X
X
1
2
3
X= H, F, Cl, Br, MeO, Me
R= Me, Et
O
O
S
X
H3C
N
N
N
SCH3
9
(X= H, Me, Cl, F, MeO)
Figure 1. Representative of some marketed selective COX-2 inhibitors (celecoxib, valdecoxib and etoricoxib) and our previously reported ones (1–3). General structure of our
newly synthesized compounds (9).

H. Irannejad et al. / Bioorg. Med. Chem. 22 (2014) 865–873
867
OH
N
O
O
O
iii
O
ii
i
X
X
X
OH
X
O S O
CH3
6
S
O S
O
H3C
CH3
4
5
7
O
H3C
O
H3C
S
X
S
X
O
O
N
N
N
N
v
N
iv
CH3
S
N
SH
8
9 (X= H, Me, Cl, F, MeO)
Scheme 1. Reagents and conditions: (i) Thioanisole, TFAA, H3PO4, rt, stir, 5 min; (ii) OxoneÒ, MeOH, THF, rt, stir, 4 h; (iii) (1) MeONa, MeOH, (2) BuONO, rt, stir; (iv)
Thiosemicarbazide, H2O, HCl, reflux; (v) CH3I, Et3N, CH3OH, rt, stir.
Table 1
In vitro COX-1 and COX-2 enzyme inhibition data for the synthesized compounds
O
O
S
H3C
N
N
N
SCH3
R
9a-e
a
a
Compounds
R
IC50 COX-1 (
lM)
IC50 COX-2 (
lM)
Selectivity indexb
9a
9b
9c
9d
9e
Celecoxib
H
CH3
Cl
54.4
52.4
39.5
51.6
50.1
24.3
0.19
0.18
0.10
0.15
0.17
0.06
286.3
291.1
395.0
344.0
294.7
405.0
F
CH3O
a
The in vitro test compound concentration required to produce 50% inhibition of enzymatic activity.
In vitro COX-2 selectivity index (COX-1 IC50/COX-2 IC50).
b
interactions, a molecular docking simulation of the representative
compound 9c was performed. Overall, the docking results were
essentially in agreement with the biological data in terms of eval-
uation of interaction energy and binding mode. The protein-inhib-
itor complex obtained from docking simulation showed that the
docked ligand was almost superimposed on the native co-crystal-
lized one (SC-558) with RMSD being 0.136 and binding free energy
of ꢀ10 kcal/mol. The hydrogen bonds and interactions between the
docked ligand and the amino acids were the same as those be-
tween the native ligand SC-558 and the amino acids. In agreement
with this statement, computational simulations revealed that the
4-chlorophenyl moiety of the ligand inhibitor (9c) was superim-
posed on the 4-bromophenyl group of the co-crystallized inhibitor
SC-558, occupying the ‘‘hydrophobic pocket’’ constituted by the
lipophilic and aromatic side chains of Phe381, Leu384, Tyr385,
Trp387, Phe518 and Met522. Moreover, the 4-methylsulfonylphe-
nyl moiety of compound (9c) similar to the 4-sulfonamidophenyl
of SC-558 is located in a region of the active site called the ‘selec-
tivity site’. This site is surrounded by polar amino acid side chains
of Arg518, Glu192, His90 and the backbone NH of Phe518. The oxy-
gen atoms of the methylsulfonyl group of compound (9c) interacts
through hydrogen bonds with the backbone NH group of Phe518
0
(2.9 ÅA) and with the guanidine group of Arg513 side chain
0
(2.8 ÅA). Meanwhile, the N1 atom of triazine ring creates hydrogen
0
bonding with the hydroxyl group of Tyr355 side chain (2.9 ÅA).
Finally, the thiomethyl moiety of compound (9c) fills a space
constituted by the aliphatic side chain of Leu359, Ile345, Leu531,
V116, and interacts through van der Waals and hydrophobic con-
tacts corresponding to the orientation of the trifluoromethyl group
of the co-crystallized ligand SC-558. This space is oriented next to
the ‘classical NSAID‘s carboxylate bonding site’ made mainly by
Arg120 (Fig. 2).
As a consequence, the binding mode of compound 9c is in high
consistency with the results of the previously published

868
H. Irannejad et al. / Bioorg. Med. Chem. 22 (2014) 865–873
studies15,16 and clearly explains and confirms the biologically ob-
tained data tested on the COX-1/COX-2 enzymes.
O
ring A
N
O
S
H2N
N
2.4. 3D-QSAR study: k-Nearest Neighbour Molecular Field
Analysis (kNN-MFA)
CF3
ring B
All the molecules were aligned based on the predefined sub-
structure highlighted in Figure 3 comprising four atoms in the
structure of SC-558. The aligned biologically active conformations
of compounds were used for the calculation of molecular fields.
Molecular fields are the steric, electrostatic and hydrophobic inter-
action energies which are used to develop a model for 3D-QSAR.
Instead of calculating all three interaction fields together and mak-
ing one model for all fields, one model was separately built for each
field in order to investigate more precisely the lattice points
around molecules.
Statistical parameters obtained for the three models are sum-
marized in Table 2. The model selection criterion is the value of
q2 (cross validated r2) and that of pred-r2 (predictive r2 for the test
set).
Accordingly, a QSAR model is considered to be predictive if the
following conditions are satisfied: r2 > 0.6, q2 > 0.6 and pred-
r2 > 0.5.17,18 The q2 and pred-r2 values, which meet these condi-
tions, for the three models based on the three different fields thus
suggest that these models are useful tools for predicting COX-2
inhibitory activity.
Figures 4–6 show the contribution plot of the three models for
the electrostatic, hydrophobic and steric fields respectively and
indicate relative regions of the local fields around the aligned mol-
ecules leading to activity variation in the model. For electrostatic
field, the lattice points generated in the model are E-985
(ꢀ10.0000, ꢀ10.0000), E-671 (2.8143, 3.1589), E-1138 (ꢀ0.9107,
0.0721) and E-453 (0.2156, 0.3351). These points suggested the
significance of electrostatic properties as indicated in the ranges
in parentheses for maximum COX-2 inhibitory activity. Similarly,
the lattice points obtained for steric and hydrophobic fields are:
S-951(ꢀ0.0111, ꢀ0.0105), S-388 (ꢀ0.0244, ꢀ0.0299) and S-4
(ꢀ0.0021, ꢀ0.0013) for steric, and H-644 (0.9968, 1.2361) and
H-995 (0.2741, 0.4364) for hydrophobic. The high negative value
for E-985 means that strong electron-withdrawing substituents
in this region are favorable and would increase COX-2 inhibitory
activity, as shown by the presence of sulfonamide and sulfonylm-
ethyl groups in the most active compounds. Positive values for
E-671 and E-453 show that electron-donating groups on the
central ring B increase biological activity of compounds. There is
Br
ring C
Figure 3. Scaffold used for alignment (SC-558). Bolded bonds represent the
substructure used for alignment.
no electrostatic relationship between ring C and COX-2 inhibitory
activity.
On the other hand, ring C might be substituted with either elec-
tron-withdrawing or electron-donating groups without loss of
activity. Less bulky substituents are tolerated at rings A and C,
meaning that increasing size of the groups substituted in these re-
gions reduces COX-2 inhibitory activity, since S-388 and S-951
have negative values. The presence of bulky groups in the central
ring B does not affect COX-2 inhibitory activity and there is no lim-
itation regarding size of the group in this region. Positive values for
the hydrophobic lattice points around ring A and ring C indicate
that COX-2 inhibitory activity could be increased by substituting
more hydrophobic groups in these two regions. However, there is
no hydrophobicity relationship between ring B and biological
activity. On the other hand, ring B may be substituted by hydro-
philic groups without biological activity reduction; as in the classi-
cal NSAIDs, a hydrophilic carboxylate side chain exists.
Our synthesized compounds have similar potency (0.1–0.2 lM)
due to little structural difference among them. The most active
compound 9c having a strong electron-withdrawing and hydro-
phobic group (methylsulfonyl) at ring A, an electron-donating
group (methylthio) at ring B and a hydrophobic and less bulky
group (chloro atom) at ring C strongly supports the above state-
ments and results obtained by 3D-QSAR analysis.
Electron-withdrawing nature of the electronegative chloro
atom does not contribute to the COX-2 inhibitory activity of the
molecule. On the other hand, replacing chloro atom at ring C with
electron-donating but still hydrophobic methyl and methoxy
groups in compounds 9b and e respectively, did not significantly
reduce COX-2 inhibitory activity.
Regarding to structure–activity relationship (SAR) of our previ-
ous studies as well as resulted data in this research, different cen-
tral heterocyclic rings such as imidazole, triazole, and triazine
substituted with alkythio group, represented the promising po-
tency and selectivity toward COX-2 enzyme. This research also
indicated that expansion of central heterocyclic ring from azoles
to triazine was tolerated by the target or even increased in selec-
tivity and potency of some compounds like 9c. In addition, in
agreement of our studies, presence of more bulky and hydrophilic
substituents like methoxy at ring C led to a decrease in COX-2 po-
tency. On the other hand, more lipophilic halogens like chloro,
and fluoro in this position retains the COX-2 inhibition
activities.10,11
The 3D pharmacophore obtained by this docking simulation
and 3D-QSAR study may be used for future drug developing pro-
jects for the design of selective COX-2 inhibitors (Fig. 7).
2.5. In vivo studies
Compounds (9a–e) were studied for determination of their po-
tential AI in rat hind paw edema method; the results are presented
in Table 3. The AI effects of the synthesized compounds were eval-
uated on the basis of the difference in hind paw thickness before
Figure 2. Docked conformation of (9c) (ball and stick) in the active site of COX-2.
For clarity only amino acids within 7 ÅA distant from the docked ligand are shown.
The co-crystallized ligand SC-558 (wire) in the active site is shown in red.
0

H. Irannejad et al. / Bioorg. Med. Chem. 22 (2014) 865–873
869
Table 2
Statistical results of 3D-QSAR study generated by forward–backward stepwise kNN-MFA method for selective COX-2 derivatives
Statistical parameters
Electrostatic field model
Steric field model
Hydrophobic field model
Training set size
Test set size
k nearest neighbour
q2
66
19
6
68
17
3
69
16
9
0.6401
0.7373
0.6045
0.6061
66
0.8450
0.4615
0.6025
0.7810
61
E_985
E_671
E_1138
E_453
0.7144
0.6184
0.6947
0.7184
64
S_951
S_388
S_4
q2-se
Pred-r2
Pred-r2se
Degree of freedom
Selected descriptors
H_644
H_995
Figure 6. Contribution plot for steric interactions around the superimposed
molecular units of selective COX-2 inhibitors using SW kNN-MFA method.
Figure 4. Contribution plot for electrostatic interactions around the superimposed
molecular units of selective COX-2 inhibitors using SW kNN-MFA method.
and 6 mg/kg. Although, compound 9a did not show any improve-
ment in anti-inflammatory activity rather than indomethacin,
other compounds revealed prominent anti-inflammatory activity
in most cases at doses 3 and 6 mg/kg. Interestingly, compound
9c not only was as strong as indomethacin (5 mg/kg) at dose of
3 mg/kg, but also showed more potency at dose of 6 mg/kg. Fur-
thermore, compound 9e at dose of 6 mg/kg was the most potent
anti-inflammatory agent in this group and showed a significant
difference with indomethacin (p 0.005) in 1 h after induction of
edema. Almost all the AI properties of the all compounds are com-
parable to reference drug. AI activities of the tested compounds are
in agreement with the observed data in the COX-2 inhibition assay.
Potential antinociceptive effects of the synthesized compounds
were carried out using formalin-induced paw-licking test. Data is
presented in Table 4 that shows mean licking time at two doses,
3 and 6 mg/kg for tested compounds and 5 mg/kg for indometha-
cin. In both early and late phases of pain response, most of the
compounds showed significant antinociceptive activity compared
to negative control. The analgesic activity of synthesized com-
pound was considerable in the first or second phase of nociception
in comparison of reference molecule indomethacin.
Figure 5. Contribution plot for hydrophobic interactions around the superimposed
molecular units of selective COX-2 inhibitors using SW kNN-MFA method.
3. Conclusion
In the present study, novel 5-Aryl-6-(4-methylsulfonyl)-3-
(metylthio)-1,2,4-triazine compounds were synthesized and
biologically evaluated against COX-1/COX-2 inhibitory activity
and after carrageenan injection as mm. As seen in Table 3, all com-
pounds showed significant activity (p 0.01 or 0.05) compared to
negative control at 1, 2 and 3 h after carrageenan injection in 3

870
H. Irannejad et al. / Bioorg. Med. Chem. 22 (2014) 865–873
Strong electron-withdrawing, less bulky
A
and hydrophobic groups at ring A
Electron-donating, no bulk limitation and
no hydrophobic limitation at ring B
B
Less bulky, more hydrophobic and
no electronic limitation at ring C
C
Figure 7. 3D-pharmacophore obtained by k-Nearest Neighbour Molecular Field Analysis.
Table 3
Anti-inflammatory effects of the test compound on carrageenan-induced hind paw edema
Compounds
Dose (mg/kg)
Difference in hind paw thickness before and after carrageenan injection (mm)
1 h After inj.
2 h After inj.
3 h After inj.
9a
9b
9c
9d
9e
3
6
3
6
3
6
3
6
3
6
1.45 0.08N
1.77 0.31
1.19 0.12N,M
0.83 0.18N,M
0.55 0.3N,M
0.67 0.2N,M
0.54 0.35N,M
0.41 0.26N,M
0.97 0.29N,M
0.34 0.12N,M
0.64 0.36N,M
0.3 0.1N,M
1.09 0.14N,M
0.87 0.39N,M
0.98 0.16N,M
0.94 0.13N,M
0.68 0.31N,M
0.8 0.16N,M
0.71 0.41N,M
0.58 0.31N,M
1.16 0.22N,M
0.56 0.19N,M
0.72 0.4N,M
0.43 0.13N,M
1.86 0.38
1
0.31N,M
0.89 0.3N,M
1.24 0.18N,M
0.75 0.26N,M
0.96 0.35N,M
0.58 0.15N,M,P
1.84 0.41
Neg. Cont.
Indomethacin
1.78 0.39
5
1.05 0.14N,M
0.81 0.22N,M
0.57 0.25N,M
Data obtained from experiments were expressed as mean sd, n = 6.
N
Significant difference with negative control (p value 0.05).
Significant difference with negative control (p value 0.01).
Significant difference with negative control (p value 0.005).
M
P
Table 4
Effects of the test compounds on nociceptive response formalin test
Compounds
Dose (mg/kg)
Licking time after formalin injection in early and late phases (second)
2nd phase
1st phase
9a
9b
9c
9d
9e
3
6
3
6
3
6
3
6
3
6
57 11N
58 10N
133 61.5N
155 38
49 14.5N,M
61 17
175 36.5
139 42N
30 6.6N,M
37 6N,M
85 26.6N,M
140 36.5N
127 49.5N
152 32.5
227 84
52 11N,M
34 11.5N,M
46 5.5N,M
49 10N,M
80 19
118 34N,M
217 68.5
127 21N,M
Neg. Cont.
Indomethacin
5
45 10N,M
Data obtained from experiments were expressed as mean sd, n = 6.
N
Significant difference with negative control (p value 0.05).
Significant difference with negative control (p value 0.01).
M
which led to the finding of a potent and selective COX-2 inhibitor,
5-(4-Chlorophenyl)-6-(4-(methylsulfonyl)phenyl)-3-(methylthio)-
1,2,4-triazine (9c) with IC50 = 0.10 lM and selectivity index of 395
almost comparable to celecoxib (SI = 405), which could be consid-
ered as a new lead for further drug development. Furthermore,
docking simulation performance on our lead compound (9c) gave
satisfactory results and could prove its potential capability to inhi-
bit COX-2. As well, the chemical structure of our new compound
(9c) was in high consistency with the 3D-pharmacophore model
Our results demonstrated that diaryl-triazine scaffolds act as in
vitro COX-2 inhibitors as well as in vivo analgesic and AI agents.
This study shows that new analogous of diaryltriazine present a
promising lead for generation of new AI and analgesic agents.
4. Experimental
4.1. Chemistry
obtained from 3D-QSAR study on
inhibitors.
a set of selective COX-2
All commercially available reagents were purchased from
Merck AG, Aldrich or Acros Organics and used without further

H. Irannejad et al. / Bioorg. Med. Chem. 22 (2014) 865–873
871
purification. Column chromatography was carried out on silica gel
3H, SO2CH3), 7.08 (t, J = 8.2 Hz, 2H, H3,5-Ar5), 7.55 (m, 2H, H2,6
-
(230–400 mesh). TLC was conducted on silica gel F254 plates.
Melting points were measured on a Kofler hot stage apparatus
and are uncorrected. The IR spectra were taken using Nicolet
FT-IR Magna 550 spectrographs (KBr disks). Mass spectra of the
products were obtained with an HP (Agilent technologies) 5937
Mass Selective Detector. 1H NMR spectra were recorded on a
Bruker 500 MHz and 13C NMR spectra on a Bruker 100 MHz NMR
instruments. The chemical shifts (d) and coupling constants (J)
are expressed in parts per million and hertz, respectively. Elemen-
tal analyses were carried out by a CHN-Rapid Heraeus elemental
analyzer. The results of elemental analyses (C, H, N) were within
0.4% of the calculated values.
Ar5), 7.75 (d, J = 7.9 Hz, 2H, H2,6-Ar6), 7.97 (d, J = 7.9 Hz, 2H, H3,5
-
Ar6); 13C NMR (100 MHz, CDCl3) d: 13.96, 44.42, 116.25 (d,
J = 21.9 Hz), 127.81, 130.21, 130.48, 132.12 (d, J = 8.79 Hz),
141.06 (d, J = 39.97 Hz), 151.74, 154.38, 164.60 (d, J = 252.45),
172.26; IR (KBr)
m
cm1: 1149, 1301 (SO2CH3), 1487 (C@N, triazine);
MS, m/z (%) 375 (M+, 12), 274 (100), 211 (36), 195 (35), 175 (12).
Anal. Calcd for C17H14FN3O2S2: C, 54.38; H 3.76; N, 11.23, Found:
C, 54.53; H, 3.42; N, 11.44.
4.1.1.5. 5-(4-Methoxyphenyl)-6-(4-(methylsulfonyl)phenyl)-3-
(methylthio)-1,2,4-triazine (9e).
Yield: 90%; mp = 173–
175 °C; 1H NMR (500 MHz, CDCl3) d: 2.77 (s, 3H, SCH3), 3.12 (s,
3H, SO2CH3), 3.85 (s, 3H, OCH3), 6.87 (d, J = 8.45 Hz, 2H,
4.1.1. General procedure for the preparation of compounds 9
1 equiv of 5-Aryl-6-(4-(methylsulfonyl) phenyl)-1,2,4-triazine-
3-thiol (8), 1.2 equiv of methyl iodide and 1.5 equiv of triethyl-
amine were added to dry methanol and the mixture was stirred
for 30 min. It was concentrated under reduced pressure, diluted
with dichloromethane and washed with ammonium chloride solu-
tion and decanted from aqueous layer. Organic phase was dried
over sodium sulfate, filtered and concentrated under vacuum. It
was chromatographed on silica gel (230–400 mesh) using dichlo-
romethane/methanol (5%) as eluent.
H
H
3,5-Ar5), 7.51 (d, J = 8.45, 2H, H2,6-Ar5), 7.78 (d, J = 8.0 Hz, 2H,
2,6-Ar6), 7.97 (d, J = 8.0 Hz, 2H, H3,5-Ar6); 13C NMR (100 MHz,
CDCl3) d: 13.93, 44.46, 55.48, 114.36, 126.27, 127.75, 130.15,
131.76, 140.95, 141.58, 151.63, 154.92, 162.39, 171.93; IR (KBr)
m
cm1 1155, 1313 (SO2CH3), 1485 (C@N, triazine); MS, m/z (%)
387 (M+, 20), 286 (100), 223 (11), 207 (40), 163 (15), 73 (10). Anal.
Calcd for C18H17N3O3S2: C, 55.80; H 4.42; N, 10.84 Found: C, 55.64;
H, 4.48; N, 10.68.
4.2. Docking simulation method
4.1.1.1. 6-(4-(Methylsulfonyl)phenyl)-3-(methylthio)-5-phenyl-
1,2,4-triazine (9a).
COX-2 atomic coordinates in complex with the selective COX-2
inhibitor SC-558 with ID number: 1CX2 was retrieved from Brook-
haven Protein Data Bank. To set the initial coordinates for the dock-
Yield = 85%; mp = 218–220 °C; 1H NMR
(500 MHz, CDCl3) d: 2.79 (s, 3H. SCH3), 3.10 (s, 3H, SO2CH3), 7.39
(t, J = 7.5 Hz, 2H, H3,5-Ar5), 7.51 (dd, J = 7.5 Hz, 3H, H2,4,6-Ar5),
7.75 (d, J = 8.0 Hz, 2H, H2,6-Ar6), 7.95 (d, J = 8.0 Hz, 2H, H3,5-Ar6);
13C NMR (100 MHz, CDCl3) d: 13.96, 44.44, 127.67, 128.89,
129.77, 130.27, 131.40, 134.45, 140.98, 141.07, 151.95, 155.60,
ing simulation, all water molecules, N-acetyl-D-glucosamine
residues and chains B, C and D were removed and excluded from
all calculations. The remaining chain A was considered for the fol-
lowing calculations. Ligand structure was sketched using Marvin-
Sketch and geometry optimized using semiempirical AM1
method in Gaussian98. The resulting minimized conformation
was then submitted to molecular docking simulation. Docking cal-
culations were performed with the program AutoDock Vina. Auto-
DockTools 1.5.4 was used to visualize, add polar hydrogens and
Kollman partial charges to the protein and to define rotatable
bonds for the inhibitor ligand. Active site was defined by a grid
point spacing of 1 A and 20 ꢁ 20 ꢁ 20 points were used. The grid
was centered on the mass center of the crystallographic inhibitor
SC-558 (X = 24.26, Y = 21.53, Z = 16.5).
A molecular mechanical/energy minimization approach was
performed to refine the AutoDock output. Such a structural optimi-
zation of the complex obtained was necessary because the docking
software AutoDock does not perform any structural optimization
of the complexes found. The computational protocol applied con-
sisted in the application of Amber force field in Gaussian 98. More-
over, due to the large number of atoms in the COX-2/inhibitor
complex obtained by docking, a subset comprising only the inhib-
itor and a shell of residues 6 A distant from the inhibitor was cre-
ated and submitted to energy minimization.19 After docking
calculation and energy minimization, the best pose with the lowest
172.29; IR (KBr)
m
cm1: 1153, 1301 (SO2CH3), 1487(C@N, triazine);
MS, m/z (%) 357 (M+, 10), 256 (100), 193 (36), 176 (43), 151 (17), 88
(7). Anal. Calcd for C17H15N3O2S2: C, 57.12; H 4.23; N, 11.76, Found:
C, 57.07; H, 4.19; N, 11.94.
4.1.1.2.
tolyl)-1,2,4-triazine (9b).
6-(4-(Methylsulfonyl)phenyl)-3-(methylthio)-5-(p-
Yield: 80%; mp = 223–225 °C; 1H
NMR (500 MHz, CDCl3) d: 2.41 (s, 3H, Ar-CH3), 2.77 (s, 3H, SCH3),
3.11 (s, 3H, SO2CH3), 7.17 (d, J = 7.5 Hz, 2H, H3,5-Ar5), 7.42 (d,
J = 7.5 Hz, 2H, H2,6-Ar5), 7.76 (d, J = 7.85 Hz, 2H, H2,6-Ar6), 7.95 (d,
J = 7.85 Hz, 2H, H3,5-Ar6); 13C NMR (100 MHz, CDCl3) d: 13.93,
21.52, 44.45, 127.66, 129.63, 129.80, 130.23, 131.48, 140.97,
141.26, 142.17, 151.87, 155.57, 172.13; IR (KBr)
m
cm1: 1157,
1316 (SO2CH3), 1495 (C@N, triazine); MS, m/z (%) 371 (M+, 10),
270 (100), 207 (30), 189 (30), 165 (6), 73 (5). Anal. Calcd for C18H17-
N3O2S2: C, 58.20; H 4.61; N, 11.36, Found: C, 58.09; H, 4.39; N,
11.44.
4.1.1.3.
5-(4-Chlorophenyl)-6-(4-(methylsulfonyl)phenyl)-3-
(methylthio)-1,2,4-triazine (9c).
Yield: 80%; mp = 148–
150 °C; 1H NMR (500 MHz, CDCl3) d: 2.78 (s, 3H, SCH3), 3.11 (s,
3H, SO2CH3), 7.37 (d, J = 8.1 Hz, 2H, H3,5-Ar5), 7.48 (d, J = 8.1 Hz,
free energy of binding (DG) was used for investigating the binding
2H,
H
2,6-Ar5), 7.75 (d, J = 7.95 Hz, 2H, H2,6-Ar6), 7.98 (d,
mode and possible interactions between ligand and amino acids in
the COX-2 active site.
J = 7.95 Hz, 2H, H3,5-Ar6); 13C NMR (100 MHz, CDCl3) d: 13.96,
44.42, 127.84, 129.27, 130.23, 131.13, 132.82, 137.99, 140.67,
141.31, 151.71, 154.36, 172.36; IR (KBr)
m
cm1: 1149, 1304 (SO2-
4.3. k-Nearest neighbour molecular field analysis
CH3), 1479 (C@N, triazine); MS, m/z (%) 393 (M++2, 3), 391 (M+,
10), 290 (100), 227 (34), 211 (28), 176 (57), 150 (10), 75 (9). Anal.
Calcd for C17H14ClN3O2S2: C, 52.10; H 3.60; N, 10.72, Found: C,
52.35; H, 3.47; N, 10.59.
A data set comprising 88 selective COX-2 inhibitors belonging
to three chemical classes (triaryl rings, 1,3-diarylcycloalkanopyraz-
oles and diphenylhydrazids) with reported COX-2 enzymatic activ-
ity (IC50) were taken from literature.20 The 3D-QSAR study was
performed using QSARPro software. 2D and 3D structure of com-
pounds were sketched and geometrically optimized with the
Merck Molecular Force Field (MMFF) method taking the root mean
4.1.1.4.
(methylthio)-1,2,4-triazine (9d).
170 °C; 1H NMR (500 MHz, CDCl3) d: 2.78 (s, 3H, SCH3), 3.12 (s,
5-(4-Fluorophenyl)-6-(4-(methylsulfonyl)phenyl)-3-
Yield: 90%; mp = 168–

872
H. Irannejad et al. / Bioorg. Med. Chem. 22 (2014) 865–873
square gradient (RMS) of 0.01 kcal/mol and iteration limit to
10,000. Steric, electrostatic and hydrophobic fields were computed
at each grid point considering Gasteiger-Marsili charges. Methyl
probe of charge +1 with 10.0 kcal/mol electrostatic and 30.0 kcal/
mol steric cut-off were used for field’s generation. A value of 1.0
is assigned to the distance-dependent dielectric constant. After cal-
culating the descriptors, constant and near constant descriptors
were removed and excluded from calculations. Training and test
sets were generated using sphere exclusion method. Among the
various methods in variable selection and model development,
stepwise forward–backward variable selection method was used
for data reduction and k-Nearest neighbor (kNN) method was cho-
sen for model development.21
Mean values of each treated group were compared with control
group and analyzed by statistical methods.
4.5.3. Analgesic activity
The analgesic activities of the test compounds were evaluated
by formalin test. The animals were fixed in a partial restraint
30 min after IP injection of test compound at concentrations of 3
and 6 mg/kg. For positive control, indomethacin at concentration
of 5 mg/kg was used. The experimental animals were assigned to
seven groups, with 6 mice in each group. The formalin test was
conducted based on the reported procedures.24 For the formalin
test, groups of animals received the solvent (DMSO + tween 20)
as control, reference drug (indomethacin) and test groups at two
doses, 3 and 6 mg/kg. After 30 min, 50 lL of 2.5% formalin in nor-
4.4. In vitro cyclooxygenase (COX) inhibition assays
mal saline was injected into the sub-plantar right hind paw to each
animal. The duration of paw licking was used as an index to deter-
mine painful response. Neurogenic period (initial phase) and
inflammatory period (secondary phase) were described at
0–5 min and 15–60 min after formalin injection respectively.
The assay was performed using an enzyme chemiluminescent
kit (Cayman chemical, MI, USA) according to our previously re-
ported method.22 The Cayman chemical chemiluminescent COX
(ovine) inhibitor screening assay utilizes the heme-catalyzed
hydroperoxidase activity of ovine cyclooxygenases to generate
luminescence in the presence of a cyclic naphthalene hydrazide
and the substrate arachidonic acid. Arachidonate-induced lumi-
nescence was shown to be an index of real-time catalytic activity
and demonstrated the turnover inactivation of the enzyme. Inhibi-
tion of COX activity, as measured by luminescence, by a variety of
selective and nonselective inhibitors showed potencies similar to
those observed with other in vitro and whole cell methods.
4.6. Statistical analysis
Statistical evaluations of the data obtained from 6 animals for
each group were performed using Graph Pad Prism 5. One-way
analysis of variance (ANOVA) was used for evaluation of statistical
differences between groups. For comparison of data with others,
Newman–Keuls multiple comparison tests was used and
statistical significance at a level of p < 0.05, < 0.01, and < 0.005
was considered.
4.5. In vivo anti-inflammatory and analgesic evaluation
Acknowledgment
4.5.1. Materials and methods
4.5.1.1. Animals.
Adult male, Swiss albino mice (25–30 g)
This research has been supported by Tehran University of Med-
ical Sciences Health Services Grant No. 7950.
were used for antinociceptive activities of the compounds. For AI
study, adult male Wistar rats (120–160 g) were used. All animals
were housed at room temperature and 50–55% relative humidity,
with 12 h light/dark cycle. Standard rodent diet and water were
used for animal feeding. Animals were acclimatized to the labora-
tory environment for at least 12 h before the experiment. The pro-
tocols adopted for the animal experiment were approved by the
institutional animal ethics committee, part of the research ethics
authorities at Tehran University of Medical Sciences.
Supplementary data
Supplementary data associated with this article can be found,
in the online version, at http://dx.doi.org/10.1016/j.bmc.2013.
12.002.
References and notes
1. Warner, T. D.; Giuliano, F.; Vojnovic, I.; Bukasa, A.; Mitchell, J. A.; Vane, J. R. Natl.
Acad. Sci. U.S.A. 1999, 96, 7563.
2. Simmons, D. L.; Botting, R. M.; Hla, T. Pharmacol. Rev. 2004, 56, 387.
3. Vane, J. R.; Botting, R. M. Inflamm. Res. 1998, S78, 47.
4. Khan, A. A.; Iadarola, M.; Yang, H. Y.; Dionne, R. A. J. Pain 2007, 8, 349.
5. Kojima, F.; Kapoor, M.; Kawai, S.; Crofford, L. J. Expert Rev. Clin. Immunol. 2006,
2, 277.
6. Sarkar, F. H.; Adsule, S.; Li, Y.; Padhye, S. Mini Rev. Med. Chem. 2007, 7, 599.
7. Nivsarkar, M.; Banerjee, A.; Padh, H. Pharmacol. Rep. 2008, 60, 692.
8. White, W. B.; Faich, G.; Borer, J. S.; Makuch, R. W. Am. J. Cardiol. 2003, 92, 411.
9. Grosser, T.; Fries, S.; FitzGerald, G. A. J. Clin. Invest. 2006, 116, 4.
10. a) Navidpour, L.; Shadnia, H.; Shafaroodi, H.; Amini, M.; Dehpour, A. R.; Shafiee,
A. Bioorg. Med. Chem. 1976, 2007, 15; b) Assadieskandar, A.; Amirhamzeh, A.;
Salehi, M.; Ozadali, K.; Ostad, S. N.; Shafiee, A.; Amini, M. Bioorg. Med. Chem.
2013, 21, 2355.
4.5.1.2. Chemical and drug administration.
Indomethacin
was used in AI and antinociceptice study as positive control. Ani-
mals were divided into 7 groups as control group, reference group
and test groups (1–5). The control group received solvent
(DMSO + tween 20) since test compounds and indomethacin were
dissolved in DMSO + tween 20. Each group included 6 animals for
evaluation. Test groups received the synthesized compounds
(9a–e) in concentration of 3 and 6 mg/kg both for AI and antinoci-
ceptive activity investigation.
4.5.2. Rat hind paw edema method
The anti-inflammatory activity of the test compound was eval-
uated by the carrageenan paw edema method in rat.23 The test
compounds at concentrations of 3 and 6 mg/kg and indomethacin
at 5 mg/kg were administration as intraperitoneal injection (IP).
Thirty minutes after the IP injection of samples, reference drug
11. Navidpour, L.; Shafaroodi, H.; Abdi, K.; Amini, M.; Ghahremani, M. H.; Dehpour,
A. R.; Shafiee, A. Bioorg. Med. Chem. 2006, 14, 2507.
12. Veeramaneni, V. R.; Pal, M.; Yeleswarapu, K. R. Tetrahedron 2003, 59, 3283.
13. Irannejad, H.; Amini, M.; Khodagholi, F.; Ansari, A.; Khoramian Tusi, S.;
Sharifzadeh, M.; Shafiee, A. Bioorg. Med. Chem. 2010, 18, 4224.
14. Ansari, N.; Khodagholi, F.; Ramin, M. R.; Amini, M.; Irannejad, H.; Dargahi, L.;
Dehghani Amirabad, A. Neurochem. Inter. 2010, 57, 958.
15. Biava, M.; Porretta, G. C.; Poce, G.; Supino, S.; Cappelli, A.; Vomero, S.; Manetti,
F.; Botta, M.; Sautebin, L.; Rossi, A.; Ghelardini, C.; Vivoli, E.; Makovec, F.;
Anzellotti, P.; Patrignani, P.; Anzini, M. J. Med. Chem. 2007, 50, 5403.
16. Unsal-Tan, O.; Ozadali, K.; Piskin, K.; Balkan, A. Eur. J. Med. Chem. 2012, 57, 59.
17. Golbraikh, A.; Tropsha, A. J. Mol. Graph. Model. 2002, 20, 269.
and dosing vehicle (50
lL/100 g), each rat was injected with
freshly prepared solution of carrageenan (50
l
L, 2.5% w/v), into
sub-plantar tissue of the right hind paw. Paw edema was measured
at 0, 60, 120 and 180 min. after carrageenan injections. The differ-
ence in hind paw thickness was measured by a Vernier caliper.

H. Irannejad et al. / Bioorg. Med. Chem. 22 (2014) 865–873
873
18. Afantitis, A.; Melagraki, G.; Sarimveis, H.; Igglessi-Markopoulou, O.;
Kolliasnovel, G. A Eur. J. Med. Chem. 2009, 44, 877.
19. Anzini, M.; Rovini, M.; Cappelli, A.; Vomero, S.; Manetti, F.; Botta, M.; Sautebin,
L.; Rossi, A.; Ghelardini, C.; Norcini, M.; Giordani, A.; Makovec, F.; Anzellotti, P.;
Patrignani, P.; Biava, M. J. Med. Chem. 2008, 51, 4476.
21. Sharma, M. C.; Kohli, D. V. Med. Chem. Res. 2012, 21, 2837.
22. Zarghi, A.; Najafnia, L.; Daraie, B.; Dadrass, O.; Hedayati, M. Bioorg. Med. Chem.
Lett. 2007, 17, 5634.
23. Kupeli, E.; Tatli, I. I.; Akdemir, Z. S.; Yesilada, E. J. Ethnopharmcol. 2007, 114, 234.
24. Meunier, C. J.; Burton, J.; Cumps, J.; Verbeeck, R. K. E. J. Pharm. Sci. 1998, 6, 307.
20. Kim, H. J.; Chae, C. H.; Yi, K. Y.; Park, K. L.; Yoo, S. E. Bioorg. Med. Chem. 2004, 12,
1629.