Novel tetrazole and cyanamide derivatives as inhibitors of cyclooxygenase-2 enzyme: design, synthesis, anti-inflammatory evaluation, ulcerogenic liability and docking study

Abstract Nineteen new compounds containing tetrazole and/or cyanamide moiety have been designed and synthesised. Their structures were confirmed using spectroscopic methods and elemental analyses. Anti-inflammatory activity for all the synthesised compounds was evaluated in vivo. The most active compounds 4c, 5a, 5d–f, 8a and b and 9a and b were further investigated for their ulcerogenic liability and analgesic activity. Pyrazoline derivatives 9b and 8b bearing trimethoxyphenyl part and SO2NH2 or SO2Me pharmacophore showed equal or nearly the same ulcerogenic liability (UI: 0.5, 0.75, respectively), to celecoxib (UI: 0.50). Most of tested compounds showed potent central and/or peripheral analgesic activities. Histopathological investigations were done to evaluate test compounds effect on rat's gastric tissue. The obtained results were in consistent with the in vitro data on COX evaluation. Docking study was also done for all the target compounds inside COX-2-active site.

Non-steroidal anti-inflammatory drugs (NSAIDs) are of great importance in the treatment of inflammation and pain. They act by inhibiting cyclooxygenase enzymes (COXs)a membrane-bound haeme proteinthat control the conversion of arachidonic acid to prostaglandins and thromboxanes. Two distinct isoforms of COX enzyme are present, a constitutive form (COX-1), associated with several side effects such as haemorrhagia and gastrointestinal (GI) ulceration, and an inducible form (COX-2) which is different in the regulation and tissue distribution from COX-1 [34][35][36] . In 2012, Al-Hourani et al. reported a novel COX-1 splice variant termed as COX-3 37 . COX-1 and COX-2 use identical catalytic mechanism to catalyse the same reactions by sharing the same substrates and producing the same products. They are very similar in their protein tertiary conformation as demonstrated from their X-ray crystal structures 38 .
These similarities in the structure of both COX isoforms represent a great challenge for the development of selective COX-2 inhibitors. The space of selectivity pocket is the main difference between the two isoforms. It is reduced in COX-1 due to the presence of Ile523 rather than Val523 in COX-2. Conformational change is occurred as a result of the presence of Val523 in COX-2, leading to the formation of additional hydrophobic secondary internal pocket protruding off the primary binding site 39 . A large number of compounds have been synthesised and evaluated for their selective COX-2 inhibitory activity. Celecoxib (I), rofecoxib (II), valedecoxib (III) and etoricoxib (IV) are the most common approved selective COX-2 inhibitors. However, inhibition of COX-2 reduces urinary sodium excretion leading to increase in blood pressure as well as myocardial infarction, that is why rofecoxib and valdecoxib had been withdrawn from the market (Figure 1) [40][41][42] .
Until now, celecoxib has been one of the most popular selective COX-2 drugs. So, there is still a need for designing and developing new effective anti-inflammatory drugs with improved safety profiles.
New water-soluble, parentral COX-2 inhibitors containing tetrazole moiety in their structures, such as celecoxib and rofecoxib analogues (V, VI, respectively) were reported to have in vivo antiinflammatory activity and exhibited a high COX-2/COX-1 selectivity (SI ¼ 2.16, 2.11, sequentially) when compared to the reference celecoxib (SI ¼ 1.68). Both of them lacked the side effect of gastric ulceration (Figure 1) 43 .
The majority of selective COX-2 inhibitors are of five-membered heterocyclic ring as pyrazole in celecoxib (I), or attached to pyridine six-membered ring as in etoricoxib (IV) 44 . Moreover, pyrazoline present in antipyrinthe first pyrazoline derivative used in the treatment of pain and inflammationand several related analogues as in felcobuzone, morazone and ramifenazone are also available as NSAIDs 45 .
Furthermore, different important selective COX-2 pharmacophores such as aminosulphonyl and methylsulphonyl groups are essential for potent and selective anti-inflammatory activity 40,41 .
In view of the above-mentioned facts and as a continuation of our previous work on the development of selective COX-2 inhibitors 36,40,46 , four groups of compounds have been synthesised: (i) chalcone derivatives 3a and b and 7a and b, (ii) pyridine containing compounds 4a-c and 5a-f, (iii) hydrazone of methylsulphonyl and aminosulphonyl derivatives 6a and b and (iv) five-membered pyrazoline ring-bearing methylsulphonyl 8a and b group or aminosulphonyl 9a and b moiety, (Figure 2). Most of the prepared compounds were linked to tetrazole ring aiming to generate novel molecular templates for safe anti-inflammatory agents. The synthesised compounds were subjected to in vitro evaluation as (COX-1/COX-2) inhibitors and in vivo (AI) activity. Analgesic activity and ulcer index (UI) have also been studied. Moreover, the effect of the most active synthesised compounds on rat's gastric tissue was evaluated using histopathological study. Finally, docking study was performed on COX-2 enzyme to explore the possible binding mode of the designed compounds inside the enzyme.

Experimental section
Chemistry Melting points were determined using a Griffin apparatus and were uncorrected. IR spectra were recorded on a Shimadzu IR-435 spectrophotometer using KBr discs and values were represented in cm À1 . 1 H NMR and 13 C NMR (DEPT-Q) were carried out on Bruker apparatus at 400 MHz for 1 H NMR and 100 MHz for 13 C NMR spectrophotometer, (Faculty of Pharmacy, Beni-Suef University, Beni-Suef, Egypt), in DMSO-d 6 , D 2 O using TMS as an internal standard and chemical shifts were recorded in ppm on d scale using DMSOd 6 (2.5) as a solvent. Coupling constant (J) values were estimated in Hertz (Hz). Splitting patterns are designated as follows: s, singlet; d, doublet, t, triplet; q, quartet; dd, doublet of doublet; m, multiplet. The electron impact (EI) mass spectra were recorded on Hewlett Packard 5988 spectrometer (Palo Alto, CA). Microanalysis was performed for C, H, N on Perkin-Elmer 2400 at the Microanalytical center, Cairo University, Egypt and was within ±0.4% of theoretical values. Analytical thin-layer chromatography (TLC): pre-coated plastic sheets, 0.2 mm silica gel with UV indicator (Macherey-Nagel) was employed routinely to follow the course of reactions and to check the purity of products. All other reagents, solvents and compound 1 were purchased from the Aldrich Chemical Company (Milwaukee, WI), were used without further purification.
Method B (for preparation of 5e and 5f) An ethanolic mixture of the selected chalcones 3a or 3b (0.01 mol) and ethyl cyanoacetate (1.13 g, 0.01 mol) in the presence of ammonium acetate (1.54 g, 0.02 mol) was refluxed for 5-6 h. After cooling, the obtained solid was filtered off, dried and crystallised from 95% ethanol to give 5e and 5f.

General procedure for preparation of 6a and b
A mixture of acetophenone derivative 2 (1.88 g, 0.01 mol) and psubstituted sulphonylphenylhydrazine hydrochloride derivative (0.015 mol) in absolute ethanol (20 mL) was heated under reflux for 6-8 h (monitored by TLC). The obtained solid on hot was filtered, dried and crystalised from 95% ethanol to give 6a&b. General procedure for synthesis of 7a and b To a solution of 3a or 3b (0.01 mol) in absolute ethanol (30 mL), urea or thiourea (0.01 mol) and KOH (0.56 g, 0.01 mol) were added. The reaction mixture was heated under reflux for 10-12 h (monitored by TLC). The solid obtained was filtered, dried and crystallised from 95% ethanol to afford 7a&b. General procedure for preparation of 1,3,5-triaryl-4,5-dihydro-1H-pyrazoles 8a and b To a solution of the appropriate chalcone 3a and b (0.01 mol) in absolute ethanol (20 mL), 4-methanesulphonylphenylhydrazine hydrochloride (0.33 g, 0.015 mol) was added and the reaction mixture was heated under reflux for 12 h. After completion of the reaction (monitored by TLC plates using chloroform/methanol 9.5:0.5 V/V), the obtained solid was filtered, dried and crystallised from 95% ethanol to give the respective triarylpyrazoles 8a and b. To a solution of the appropriate chalcone 3a and b (0.01 mol) in absolute ethanol (20 mL), 4-benzenesulphonamidehydrazine hydrochloride (0.33 g, 0.015 mol) was added and the reaction mixture was heated under reflux for 18 h. After completion of the reaction (monitored by TLC plates using chloroform/methanol 9.5:0.5 V/V), the obtained solid was filtered, dried and crystallised from 95% ethanol to give the respective triarylpyrazoles 9a and b.

In vitro cyclooxygenase (COX) inhibition assay
The ability of the test compounds listed in Table 1 to inhibit ovine COX-1 and COX-2 (IC 50 value, lM) was tested using an enzyme immune assay (EIA) kit (Cayman Chemical, Ann Arbor, MI) was according to a reported method 47,48 .

In vivo assays
Female wister rats (150-200 g) were used in this study. The animals were kept at controlled conditions (temperature 23 ± 2 C, humidity 60 ± 10%) and a 12/12-h light/dark cycle with access to food and water. All procedures relating to animal care and treatments were performed according to protocols approved by the Research Ethical Committee of Faculty of Pharmacy Beni-Suef University (2014-Beni-Suef, Egypt).

Ulcerogenic liability
The ulcerogenic effects of compounds 4a, 5a, 5d, 5e, 5f, 8a, 8b, 9a, 9b and celecoxib were evaluated and compared to that of indomethacin. Forty-eight rats were used in this study, divided into 12 groups and fasted for 18 h before drug administration. The control group received the vehicle (2.5% Tween 80). Other groups received test compounds, celecoxib or indomethacin at a dose of 50 mg/kg, then animals were fed after 2 h. Rats were given the specified dose orally for three successive days. Rats were sacrificed after 2 h of the last dose, then the stomach of each rat was removed and opened along the greater curvature for determination of the ulcer number and ulcer index according to the reported method 50 .

Analgesic activity
Hot plate method This method was used to evaluate the central analgesic activity by determination of the delay in the latency time of pain response 51 . The analgesic activity of compounds 4a, 5a, 5d, 5e, 5f, 8a, 8b, 9a, 9b and celecoxib were evaluated. The mice were orally administered 10 mg/kg of either test compounds or celecoxib, while 2.5% Tween 80 solution was administered to the normal control group. After 60 min, the animals were placed on a hot plate maintained at 55 ± 0.5 C. The reaction time was recorded as the time taken by the animals to blow or lick the fore or hind paw or jump off the plate.

Writhing method
The assay was performed as described by Koster et al. 52 . Pain was induced by intraperitoneal injection of acetic acid, then counting the number of abdominal writhings. Mice were orally administered 10 mg/kg of compounds 4a, 5a, 5d, 5e, 5f, 8a, 8b, 9a, 9b and celecoxib 30 min before intraperitoneal injection of 0.6% acetic acid. After 5 min, the animals were observed and the number of abdominal writhings was recorded for 20 min.

Histopathological investigation
Whole rats stomach were collected, selected stomach samples from each rat were fixed in 10% neutral buffered formalin for 48 h and routinely processed for paraffin embedding. 5 lm Thickness sections were obtained then stained with haematoxylin and eosin for histopathological evaluation 53 .

Statistical analysis
Significant difference among groups was assessed using one-way ANOVA followed by Dunnett's test. Differences were considered significant at Ã p > .05, ÃÃ p < .01 and ÃÃÃ p > .001. GraphPad Prism software version 5 (Canada) was used to carry out all statistical tests.

Docking study
Docking was performed to obtain prediction of conformation and also energy ranking between COX-2 receptor (PDB: 1CX2) and the designed compounds 47 . Molecular Operating Environment (MOE, Version 2005.06, Chemical Computing Group Inc., Montreal, Quebec, Canada) was used in docking studies.
The cocrystallised ligand was docked first to study the scoring energy, root mean standard deviation (RMSD) and amino acid interactions. RMSD, for COX-2 enzyme and the lead compound SC-558 was 3 A .
Docking was performed using London dG force and refinement of the results was done using force field energy. Preparation of the synthesised compounds for docking was achieved via their 3 D  50 , lM) is the mean of two determinations acquired using an ovine COX-1/COX-2 assay Kit (Catalog No. 560131, Cayman Chemicals Inc., Ann Arbor, MI) and the deviation from the mean is <10% of the mean value. b The in vitro COX-2 selectivity index (COX-1 IC 50 /COX-2 IC 50 ). structure built by MOE. Before docking, 3 D protonation of the structures, running conformational analysis using systemic search and selecting the least energetic conformer were performed. The same docking protocol used with ligand was also applying for the designed compounds. Amino acid interactions and the hydrogen bond lengths were calculated. The data obtained are summarised in Table 4.
The prepared compounds have been characterised by IR, 1 H NMR, 13 C NMR, mass spectra and elemental analyses.
IR spectra of 3a and b showed a sharp peak at 1674, 1663 cm À1 corresponding to C ¼ O group. 1 H NMR spectra of 3a and b revealed the presence of two doublet protons at d 7.77 due to COCH ¼ CH and at d 7.90, 7.95 corresponding to COCH ¼ CH proton with high J value (15.2 Hz). 13 C NMR spectra of 3a and b showed a peak at d 188. 49,188.60 attributed to C of C ¼ O.
Formation of aminocyanopyridine derivatives 4a-c and their isosteric hydroxycyanoypyridine derivatives 5a-f was achieved using two different methods (A and B). In method A, one-pot multicomponent synthetic approach was utilised. Thus, compound 2 reacted with the respective aromatic aldehyde, malononitrile or ethyl cyanoacetate in the presence of ammonium acetate to yield the respective derivatives 4a-c (45-61%, yield) and 5a-f (41-57%, yield).
The second procedure (B), stepwise reaction was applied, thus reaction of acetophenone derivative 2 with the respective aldehyde to give the corresponding chalcone 3a and b, followed by the reaction of ammonium acetate and malononitrile or ethyl cyanoacetate to afford 4c (39%, yield) and 5e and f (41-46%, yield).
Better yield and operational simplicity of the first method (A) encouraged us to prepare the remaining derivatives 4a and b and 5a-d using it.
IR spectra of cyanopyridine derivatives 4a-c and 5a-f showed the appearance of sharp peak at 2226-2207 cm À1 corresponding to CN group. In addition to a forked peak at 3352-3210 and 3136-3117 cm À1 due to NH 2 in 4a-c, and a broad peak at 3483-3325 cm À1 attributed to OH group in case of 5a-f. Moreover, the absence of peak due to C ¼ O group of the parents 2 and 3a and b confirmed the structure of 4a-c and 5a-f.
The 1 H NMR spectra of 4a-c and 5a-f showed a singlet of oneproton intensity at d 6.84-7.88 corresponding to pyridine H-5. Also, D 2   Condensation of compound 2 with 4-methanesulphonylphenyl hydrazine hydrochloride or 4-benzenesulphonamide hydrazine hydrochloride afforded 6a and b, respectively. IR spectra of 6a and b showed two sharp peaks at 1277, 1273 and 1157, 1126 cm À1 corresponding to SO 2 , in addition to a peak at 3449 and 3440 cm À1 due to NH group. There was no evidence for the presence of C ¼ O group that present in the parent compound 2.
1 H NMR spectra of 6a and b revealed the presence of an exchangeable singlet signal at d 9.90, 10.05 attributed to NH proton. Another D 2 O exchangeable singlet signal of two protons intensity appeared at d 7.11 corresponding to SO 2 NH 2 protons in 6b. Moreover, three protons of SO 2 CH 3 group in 6a appeared as a singlet signal at d 3.13.
In an attempt to synthesise pyrimidine derivatives A from the reaction of 3a and b with urea or thiourea under reflux temperature in alcoholic KOH, the unexpected chalcone derivatives 7a and b were obtained instead. The deviation of the reaction caused as a result of two factors, heating and presence of a base. Thus, tetrazole ring cleaved with the loss of nitrogen (N 2 ) and forming products bearing cyanamide (-NHCN) group. The same explanation for tetrazole cleavage was published by Vorobiov et al. 55 upon using more drastic conditions (NaOH in DMSO), (Scheme 2). IR spectra of 7a and b showed three sharp peaks at 3210, 3121; 2257, 2203 and 1682, 1674 cm À1 corresponding to NH, CN and C ¼ O groups. 1 H NMR spectra of 7a and b revealed the presence of two doublet signals at d 7.66, 7.68 and 7.81, 7.86 due to COCH ¼ CH and COCH ¼ CH protons, respectively with high J value (15.6 Hz). On the other hand, absence of signals for both pyrimidine H-5 and tetrazole H-5 confirmed formation of the unexpected products 7a and b. 13 C NMR spectra of 7a and b showed signals for CN and C ¼ O at d 112.15, 113.61 and 187.48, 187.82, respectively. No evidence for tetrazole C-5 which present in the parent compounds 3a and b and in compound A.  Mass spectra of 7a and b showed molecular ion peak at m/z 308 (11.81%) and 338 (4.68%), sequentially.
Another two groups of triarylheterocycles 8a and b and 9a and b were synthesised using the reaction sequence adopted in Scheme 3. Accordingly, condensation of a,b-unsaturated ketones 3a and b with 4-methanesulphonylphenylhydrazine hydrochloride or 4-benzenesulphonamidehydrazine hydrochloride in absolute ethanol gave the respective 1,3,5-triaryl-pyrazolines 8a and b and 9a and b, in good yields (40-51%).
IR spectra of 8a and b and 9a and b showed two sharp peaks at 1294-1234 and 1151-1125 cm À1 corresponding to SO 2 . Another broad peak appeared at 3429, 3428 cm À1 observed in case of 9a&b due to NH 2 .
1 H NMR spectra of pyrazolines 8a and b and 9a and b displayed three doublet of doublet (dd) signals each of one-proton intensity at d 3.29-3.33, 3.99-4.03 and 5.57-5.63 with three different J values corresponding to three protons of pyrazoline ring. The highest J value 17.6-18 Hz was due to germinal coupling of the two protons at C-4 of pyrazoline ring, while the other two J values 12.4 and 5.6-6.4 Hz due to coupling of the two geminal protons at C-4 and the vicinal proton at C-5. Additionally, 1 H NMR spectra of 8a and b revealed the presence of a singlet signal at d 3.09-3.10 due to SO 2 CH 3 protons. Moreover, an exchangeable signal at d 7.11, 7.64 corresponding to SO 2 NH 2 protons was observed in 1 H NMR spectra of 9a and b. 13 C NMR spectra of 8a and b and 9a and b showed two peaks at d 40.51-43.66 and 63.05-63.73 corresponding to C-4 and C-5 of pyrazoline ring.

Anti-inflammatory activity
In vitro cyclooxygenase (COX) inhibition assay The peroxidase activity of cyclooxygenase enzyme isoforms was measured using the COX activity assay kit. The ability of the test compounds to inhibit both ovine COX-1 and COX-2 enzymes was explored in vitro using a colorimetric enzyme immunoassay (EIA) kit. The appearance of oxidised N,N,N 0 ,N 0 -tetramethyl-p-phenylenediamine (TMPD) was monitored at 590 nm. The kit includes isozyme-specific inhibitors for distinguishing COX-2 activity from that of COX-1. The advantages of this COX assay method are screening a vast number of inhibitors and saving much time.
The in vitro test compound concentration required to produce 50% inhibition of COX-1 or COX-2 (IC 50 %) was measured. Moreover, the COX-2 selectivity index (S.I.) values [IC 50 (COX-1)/ IC 50 (COX-2)] were calculated and compared with that of the standard drug celecoxib (as a selective COX-2 inhibitor), diclofenac (non-selective COX inhibitor) and indomethacin (selective COX-1 inhibitor). All the synthesised compounds were tested; data are listed in Table 1.

Anti-inflammatory activity
In vivo anti-inflammatory activity of the tested compounds was performed using carrageenan-induced rat paw oedema method compared to celecoxib as a reference anti-inflammatory drug.
Mean changes in paw oedema thickness of animals pre-treated with the tested compounds and celecoxib after 1, 3 and 5 h from the induction of inflammation were measured and listed in Table 2.
The reference drug, celecoxib showed 76% and 66% inhibitory activity against carrageenan-induced paw oedema after 1 and 3 h, respectively, then the activity decreased to be 19% after 5 h. Most of the tested compounds showed a significant anti-inflammatory activity (p < .001) especially after 1 h from administration.
After 1 h, compound 5f exhibited anti-inflammatory activity (79%) similar to that of celecoxib (76%). This result was in accordance with the in vitro COX-2 assay data. Compounds containing hydroxypyridine carbonitrile core such as 5d and 5e and those with SO 2 Me and SO 2 NH 2 pharmacophores like 8a, 8b, 9a, 9b showed strong anti-inflammatory activity of about 71-75%. These compounds also exhibited good COX-2 inhibitory activities.
After 3 h and 5 h, the anti-inflammatory activity for most of the synthesised compounds was gradually decreased except for phenyl and N,N-dimethylaminophenyl derivatives of pyridine hydroxycarbonitrile (5a and 5d, respectively) and dimethoxy phenyl pyrazoline derivative 8a showing duration of action extended to 3 h to 4a-c be 54%, 63% and 52%, sequentially.
The results of ulcerogenic liability (Table 3) revealed that indomethacin caused the most ulcerogenic toxicity with ulcer index (UI: 22.50), whereas both celecoxib and tested compounds exhibited much lower UI between (0.50-2.00).
The rest of the tested compounds 4c, 5a, 5d-f, 8a and 9a exhibited lower toxicity than indomethacin (UI: 22.50), where the UIs were in the range of 1.25-2.00.

Analgesic activity
Hot plate method This method was used to evaluate the central analgesic activity of nine compounds from the newly synthesised derivatives by determination of the delay in the latency time of pain response. Celecoxib was used as a reference analgesic drug. The delay in the latency time of pain response of the test compounds compared to vehicle-treated animals was determined. All the tested compounds showed potent analgesic activities specially compounds 5e, 5f, 8a and 8b. These results were consistent with the in vitro data on COX, where compounds 5f, 5e and 8b showed COX-2 inhibitory activity with IC 50 equal to 46.07, 37.57 and 35.18 lM, respectively, (Figure 3).
The most active compounds were 5d, 5e and 9a compared to normal control. The most potent compound 5e also has a good inhibitory activity against COX-2 as mentioned previously. The data obtained using this method is summarised in Figure 4.

Histopathological studies
The stomach specimen of control treated rats was characterised by normal histological structure of glandular gastric mucosa,  submucosa and musculosa ( Figure 5, control) 56 . Complete disruption of protective mucosal layer was observed in indomethacin treated rats as mucosa showed an erosion formation (starting point of ulcer formation), also some eosinophilic inflammatory cells infiltration in submucosa associated with hyalinosis and coagulative necrosis of muscular layer and this results in accordance with G€ unnur et al. 57 who confirmed that indomethacin induces gastrodoudenal ulcer formation after oral intake ( Figure 5, indomethacin). There were thickening and hyalinosis of basement membrane with lymphocytic infiltration and minimal congested submucosal blood capillaries after celecoxib intake ( Figure 5, celecoxib) that is considered much safer drug than indomethacin.
The current tested drug compounds in this study; 4c treated rats has shown that only lesion on the glandular epithelium of gastric mucosa which showed mucous degeneration in gastric glandular epithelium reach to coagulative necrosis ( Figure 5, 4c). For compound 5a, all layers of stomach were more or less normal ( Figure 5, 5a). (Ulcer index ¼ 1.5, the short period of oral intake was not enough to induce the tissue reaction towards the drug compound). Mucous degeneration in gastric glandular epithelium, thickening and hyalinosis of basement membrane, as well as eosinophilic infiltrations in submucosa were observed in 5dtreated rats ( Figure 5, 5d 1, 2). For compounds 5e (Figure 5, 5e) and 5f (Figure 6, 5f) affect only muscular layer of the stomach by causing corrugation and hyalinosis in some muscle bundles and reach to coagulative necrosis; this changes suggest presence of nausea and irritability of the stomach reach to vomiting may appear as clinical sign in this two groups.
In rats treated with compounds bearing SO 2 CH 3 pharmacophore and dimethoxy or trimethoxy phenyl part, some areas of glandular epithelium were more or less normal associated with sever congested blood capillaries and eosinophilic infiltrations in submucosa of 8a treated rats. Ulcer was found in other areas of glandular epithelium of mucosa, thickening and hyalinosis of basement membrane and associated with congestion of blood capillaries, while massive destruction of localised area of epithelium with formation of an erosion associated with hyalinosis and necrosis of muscular layer in 8b treated rats (Figure 6, 8a and b).  Scanning of stomach specimens of rats treated with compounds 9a and 9b (have SO 2 NH 2 moiety and di-or trimethoxyphenyl group), showed that sever congestion in blood capillaries and lymphocytic infiltration ( Figure 6, 9a, 1). Area of ulceration between glandular epithelium and keratinised epithelium of gastric mucosa and lymphocytic infiltration were noticed. Coagulative necrosis of some cells of glandular epithelium of gastric mucosa ( Figure 6, 9a, 2) were detected. Mucous degeneration of glandular epithelium of gastric mucosa, and lymphocytic infiltration and mild congestion in submucosa 9b were observed, (Figure 6, 9b).

Docking study
All new designed compounds were docked using X-ray crystal structure data for COX-2 enzyme obtained from the protein data bank (pdb: ID 1CX2) 58 .
In docking study, ligand SC-558 and all new designed compounds were docked using Molecular Operating Environment (MOE, Version 2005.06, Chemical Computing Group Inc., Montreal, Quebec, Canada) into the COX-2 receptor.
It was observed that H-bonding interactions between ligand SC-558 (selective COX-2 inhibitor) and COX-2 receptor were achieved via two H-bonds (i) SO 2 NH 2 with His90 (2.42 A ) and (ii) pyrazole N-2 with Tyr355 (2.77 A ). The energy associated with intermolecular interaction was À10.0340 Kcal/mol, Table 4 and Figures 7 and 8.
Most of the designed compounds had the same pattern of interactions in the COX-2 receptor through His90, Tyr355, Tyr385, Ser530, Gly354 and Ser353 amino acids, with one to four H-bonds.
Thirteen compounds 3a and b, 4a and c, 5a-f, 6a and b, 8a and b and 9a and b showed appreciable binding interactions  (affinity in Kcal/mol ranges from À11.5834 to À22.4323) with one to four H-bonding interactions. Compounds 4b, 5d, 7a and b exhibited lower binding interactions than ligand and their affinity range was from À3.6634 to À7.0670 Kcal/mol with different numbers of H-bonds from one to four.
Docking compound 5f (with COX-2 S.I.¼46.07) in the COX-2 enzyme, four H-bonding interactions with His90, Tyr355, Tyr385 and Ser530 amino acids were observed. Its binding energy was À22.4323 Kcal/mol more than that of SC-558. Compounds 6a and b, 8a and b and 9a and b, containing COX-2 pharmacophores (SO 2 Me or SO 2 NH 2 ), were in high response to COX-2.
On the other hand, affinity of some pyridine containing compounds such as 4b and 5d, and the two cyanamide derivatives 7a and b (lack from tetrazole moiety), was lower than that of SC-588.
In most compounds, tetrazole N-2 was important for formation of H-bonding interactions with His90 and Tyr355 amino acids, (Table 4).

Conclusions
In summary, the design and synthesis of new tetrazole 3a and b, 4a-c, 5a-f, 8a and b and 9a and b and cyanamide derivatives 7a and b as anti-inflammatory agents were described. In vitro COX-1 and COX-2 assay was evaluated. Compound 5f possessing hydroxypyridine carbonitrile core and bearing trimethoxyphenyl group, was the most potent and selective COX-2 inhibitor with IC 50 value of 46.07, more than that of celecoxib (IC 50 ¼45.68). Compounds with pyridine ring 4c, 5a, 5e and those with trimethoxyphenyl ring and pyrazoline core-bearing SO 2 CH 3 or SO 2 NH 2 pharmacophores 8b and 9b showed high activity as COX-2 inhibitors with IC 50 ranged from 35.13 to 42.02. All test compounds were evaluated for their in vivo anti-inflammatory activities. Analgesic activity (central and peripheral) and ulcerogenic liability were assessed. Compounds 9b and 8b with trimethoxyphenyl part and SO 2 NH 2 or SO 2 CH 3 pharmacophore have the same or near safety profile as the selective COX-2 inhibitor celecoxib (UI ¼ 0.5, 0.75, respectively).  The rest of the tested compounds 4c, 5a, 5d-f, 8a and 9a showed remarkable improvement in ulcer index (UI ¼ 1.25-2.00) when compared with indomethacin (UI ¼ 22.5). From histobathological investigations, it was shown that 5a was the most preferable compound with minimal drug effect on tissue. While 9a caused destructive effect on the mucosa and ulceration.
Molecular docking study on COX-2-active site showed that in most compounds, tetrazole N-2 was important for the formation of H-bonding interactions with His90 and Tyr355 amino acids similar to SC-558, the ligand used.