Novel sulindac derivatives: synthesis, characterisation, evaluation of antioxidant, analgesic, anti-inflammatory, ulcerogenic and COX-2 inhibition activity

Abstract A new series of N′-(substituted phenyl)-2-(1-(4-(methylsulfinyl) benzylidene)−5-fluoro-2-methyl-1H-inden-3-yl) acetohydrazide derivatives (1 – 25) were prepared in good yields in an efficient manner. All the compounds were fully characterised by the elemental analysis and spectral data. Synthesised compounds were evaluated for antioxidant activity by DPPH method. Compounds 7 (R = 3-methoxyphenyl), 3 (R = 4-dimethylaminophenyl) and 23 (R = 2,4,5-trimethoxy phenyl) substitutions were found to be having highly potent antioxidant activity. Compound 3, with para dimethylaminophenyl substitution was found to be having highest antioxidant activity. It was further evaluated in vivo for various analgesic, anti-inflammatory, ulcerogenic and COX-2 inhibitory activity in different animal models. Lead compound 3 was found to be significant anti-inflammatory and analgesic agent. It was also evaluated for ulcerogenic activity and demonstrated significant ulcerogenic reduction activity in ethanol and indomethacin model. The LD50 of compound 3 was found to be 131 mg/kg. The animals treated with compound 3 prior to cisplatin treatment resulted in a significant reduction in COX-2 protein expression when compared to cisplatin-treated group. Sulindac derivative with para dimethylaminophenyl substitution was found to be the most potent antioxidant, anti-inflammatory and analgesic agent as well as with significant gastric sparing activity as compared to standard drug sulindac. Compound 3 significantly downregulated liver tissue COX‐2 gene expression.


Introduction
Cyclooxygenase (COX) enzyme catalyses the conversion of arachidonic acid to prostaglandin H 2 (PGH 2 ), which is converted to many prostanoids by specific isomerase enzymes because it is an unstable intermediate. Non-beneficial effects of prostaglandins include pain and fever associated with inflammation; beneficial effects include gastro-intestinal protection and platelet function. The COX-1 and COX-2 are the two isoforms, which are regulated differently. The cyto-protection in the gastrointestinal (GI) tract is provided by COX-1 and COX-2 mediates inflammation 1 .
Non-steroidal anti-inflammatory drugs (NSAIDs) are used to treat pain and inflammation. Side effects include gastrointestinal toxicity such as gastro-duodenal perforations, ulcers and bleeding, ascribed to the inhibition of cyclooxygenase-1 (COX-1). Thus, selective inhibitors of cyclooxygenase-2 (COX-2) were synthesised in an attempt to decrease these side effects [2][3][4][5][6] . Physicians would prescribe gastro-protective agents with a conventional NSAID, prior to the introduction of the COX-2 selective inhibitors. Selectively inhibition of COX-2 enzyme would result in the same anti-inflammatory benefits as that of non-selective NSAIDs provide but with less incidences of gastrointestinal side effects 7,8 . Some COX-2 inhibitors have also been found to have cardiovascular side effects.
Sulindac is an indene derivative NSAID, known to induce ulceration. New sulindac derivatives are reported as anti-inflammatory 9 , anticancer [10][11][12] , COX-1 inhibitors 13 and have shown PPAR c activity 14 . Syntheses of novel derivatives of NSAIDs have improved their safety profile which resulted in an increased anti-inflammatory activity with reduced ulcerogenicity 15,16 .
It would be desirable to provide an indene derivative having the anti-inflammatory and analgesic properties of a COX-2 inhibitor NSAID, but which also provides gastric sparing activity. The aim of this study was to prepare novel sulindac acetohydrazide derivatives and evaluate their potential antioxidant, analgesic, anti-inflammatory, ulcerogenic and COX-2 inhibition activity.
spectrophotometer. The samples were run in DMSO-d 6 with tetra methyl silane (TMS) as an internal standard. Molecular masses of compounds were determined in GC mass spectroscopy. The CHN Elementar (Analysensysteme GmbH, Germany) was used for elemental analysis of the compounds. Cisplatin was purchased from Sigma-Aldrich, USA. Antibodies against COX-2 and b-actin were purchased from Abcam (Cat Log No: ab15191).

DPPH radical scavenging assay
The antioxidant activity was measured based on the scavenging activity of the stable DPPH free radical. The antioxidant activity was determined by following the method 18 . The compounds in concentration of (100 mg/mL) were added to 3 ml of 0.004% DPPH solution. Methanol was replaced in the control sample. Absorbance was determined at 520 nm after 30 min. Butylated hydroxytoluene (BHT) was used a reference drug. The percent inhibition was calculated by the following equation:

Pharmacological activities
Thirty-five Adult wistar male rats weighing 240-260 g, 12-14 weeks' old and mice were obtained from animal house of Department of the Community Health Sciences, College of Applied Medical Sciences, King Saud University, Riyadh, KSA. Animals were given favourable conditions (temperature 25 C, 12/ 12 h light and dark cycle, and humidity 60 ± 10% and pathogenfree environment). The rats were fed a dietary formulation of protein (18.1%), fat (7.1%), carbohydrate (59.3%) and fibre (15.5%) with food and water being provided ad libitum. The study protocol was approved by Ethical Committee of College of Applied Medical Sciences, King Saud University, Saudi Arabia. (Ethics Number: CAMS 22 -39/40).

Anti-pyretic study
Hyperthermia was induced in mice by (s.c.) injection of (20 ml/kg) of a 20% aqueous suspension of brewer's yeast in the back below of nape of the neck 19 . The animals were fasted for the duration of 24 h. Water was made available. Control temperature was taken 24 h after the yeast injection to determine the pyretic response to yeast. Temperature taken 1 h prior to drug administration in the fevered animals, served as a pre drug control. Drugs were given 24 h after the yeast injection and temperature were recorded at 60, 90 and 120 min. after the administration.
Analgesic study by tail flick method Acute nociception was assessed using a tail flick apparatus (Tail Flick model DS 20 Sorrel Apelex, France) following the method 20 . Briefly, each animal was placed in a restrainer, 2 min. before treatment, and baseline reaction time was measured by focussing an intensity controlled beam of light on the distal one-third portion of the animals' tail. The suspension was orally administered immediately after this step and 25 min. later, the post drug reaction time was measured. A 10-s cut-off time was used in order to prevent tissue damage.

Analgesic activity by hot plate method
The hot plate method used as described by Turner 21 . The animals were dropped gently on a hot plate maintained at 55 ± 5.5 C. The reaction time was taken as the interval extending from the instant the animal reached the hot plate until the moment the animal licked its forefeet or jumped off. The reaction time was measured 10 min before the oral administration of the drug and þ60 þ 90 þ 120 min after treatments.

Analgesic activity by writhing test
Writhing was induced in mice by intraperitoneal administration of 0.1 ml of 1% acetic acid. The number of writhing movements was counted for 20 min. The writhing test was performed after the administration of the vehicle or drug.

Carrageenan-induced paw edoema in rats
Pedal inflammation in albino rats of either sex was produced according to the reported method 22 . An injection wad made of 0.05 ml of 1% carrageenan sodium salt (BDH) into right hind foot of each rat under the plantar aponeurosis. The test group of rats was treated orally with drugs 1 h before carrageenan injection. At the same time, control group was given 5 ml/kg of normal saline and the reference group was given 100 mg/kg of an aqueous solution of sulindac. The measurements of foot volume were done by the displacement technique using a plethysmometer (Apelex, France) immediately after and þ2 and þ3 h after the injection of carrageenan. The inhibitory activity was calculated to the following formula 100 (1 À a À x/b À y); where "b" is the mean paw volume of control rats after carrageenan injection and "y" before the injection; whereas" x" is the mean paw volume of treated rats before injection and "a" is the mean paw volume after carrageenan injection.

Ulcer study of drugs using 80% ethanol
The ethanol induced ulcer model was used to study gastro-protective activity of compound 3. The rats were grouped into five groups (n ¼ 6). Groups I and II received saline solution and served as negative-control and ulcer-control, respectively. Group III received compound 3 (150 mg/kg) orally and served as the experimental drug group. Animals in groups IV received the Sulindac (100 mg/kg body weight). After 1 h, all of the groups, except Group I, received (20 ml/kg) of 80%. The animals were sacrificed 1 h later under anaesthesia and their stomachs were quickly removed for further studies 23 .

Gastric lesions induced by indomethacin
Suspension of Indomethacin in 1.0% carboxymethylcellulose (CMC) in water (6 mg/mL) at a dose of (30 mg/kg) body weight was administered orally. Control rats were treated with vehicle. Compound 3 was given half an hour prior to Indomethacin administration at a dose of 150 mg/kg 24 .

Determination of malondialdehyde (MDA)
The MDA was measured according to the method by Utley and others 25 . The tissue was removed and each tissue was homogenised in 0.15 M KCl to give a 10% W/v homogenate. Aliquots of homogenate (1 ml) were incubated at 37 C for 3 h in a metabolic shaker. Then 1 ml of 10% aqueos trichloroacetic acid was added and mixed. This was then centrifuged at 4000 rpm for 10 min. Total of 1 ml of the supernatant was removed and mixed with 1 ml of 0.67% thiobarbituric acid in water and placed in a boiling water bath for 10 min. The mixture was cooled and diluted with 1 ml of distilled water. The absorbance of the solution was then read at 535 nm. The content of MDA (nmol/g) was then calculated, by the reference to a standard curve of MDA solution.

Estimation of non-protein sulfhydryls (NP-SH)
Hepatic non-protein sulfhydryls (NP-SH) was measured according to the reported method by Sedlak and Lindsay 26 . The tissue was homogenised in ice cold 0.02 mmol/L ethylenediaminetetraacetic acid (EDTA). The Aliquotes of 5 ml of the homogenates were mixed in 15 ml test tube with 4 ml of distilled water and 1 ml of 50% trichloacetic acid (TCA). The tube was shaken intermittently for 10 min. and centrifuged 3000 rpm for 10 min. Total of 2 ml supernatant was mixed with 4 ml of 0.4 mmol/L tris buffer (pH 8.9). Total of 0.1 ml of 5,5-dithiobis (2-nitrobenzoic acid) (BTNB) was added and the sample was shaken. The absorbance was measured within 5 min of addition of DTNB at 412 nm against reagent blank.

Determination of LD 50
The LD 50 (lethal dose 50%) was calculated for compound 3 by Karber method 27 . For determination of LD 50 , an observation was made for 24 h and symptoms of toxicity and rate of mortality were noted. Expired animals were counted at the end of the study period for the calculation of LD 50 . LD 50 ¼ LD 100 À P Â (a Â b)/n, where n is the total number of animals in a group, a is the difference between two successive doses of administered extract/substance, b is the average number of dead animals in two successive doses, and LD 100 is the lethal dose causing 100% death of all test animals.

COX-2 mRNA expression in cisplatin-induced hepatotoxicity in rats
The animals were divided into five groups and each group with seven rats. The doses of compound 3 and cisplatin were selected after performing the pilot experiment.

Sample collection and preparation
On the last day of experiment, all animals were terminated and anaesthesia was made by injecting ketamine/xylazine mixture (75/ 2.5 mg/kg, respectively) via the intraperitoneal route. Anaesthetised rats were secured in a supine position and organ samples were taken from the liver. The liver tissues were quickly harvested. The tissues were treated with liquid nitrogen and were used for RNA extraction and immunoblotting.

Extraction of protein and Western blot analysis
The previously developed procedures with slight modifications were used to perform SDS-PAGE and western blot investigations. The protein concentration was estimated by Bradford assay. For western blotting 8-12% polyacrylamide gels were used to resolve 40 lM of protein, transferred on to a nitrocellulose membrane, probed with appropriate monoclonal primary antibodies, and detected by super signal west Pico, Dura or Femto Chemiluminescence Reagent (Thermo Scientific, USA). Quantification of protein bands was done through measuring band density using Image J software. The densities of the bands (normalised to actin) relative to that of the untreated control (designated as 1.00) were presented as mean ± SEM of three separate experiments.

Gene expression analysis
Total RNA from frozen liver was extracted by using kit, according to the manufacturer's instructions (Promega, CatLog No: Z3101). The cDNA synthesis was performed using the Applied Biosystems TM High-Capacity cDNA Reverse Transcription Kit. The reaction mixture was prepared containing 10 mL FastStart Universal SYBR Green Master (Roche, Germany), 6 mM reverse primers, and 10 mg cDNA, with RNAase free water added to a total volume of 20 mL. The amplification and real-time analysis were done for 40 cycles with following factors; 95 C (10 min.) in order to activate of FastStart Taq DNA polymerase; 60 C (1 min.) for amplification and real-time analysis. The gene expression levels were determined using 2-DDCT. Primer sequences used are shown below:

Molecular docking of compounds against COX-2 protein
Three-dimensional structure of the Cox-2 gene was developed using homology modelling. Modeller 9.17 was employed to predict the structure using templates (5F1A, 5IKQ, 5F19) downloaded from PDB. All of the models showed more than 90% identity with our protein. The predicted structure was further refined by energy minimisation. Finally, the structure was validated using the Ramachandran plot. Furthermore, three-dimensional structures of all synthetic compounds and Sulindac were constructed using Chem 3 D Pro 12.0 version. The protein-ligand docking analysis was performed using online PatchDock server. Three-dimensional structures of protein and ligands were used as input. PatchDock server rated the all possible docking confirmations using minimum ACE (Atomic contact energies). Finally, the docking confirmations were visualised using Pymol and LigPlus.

Results and discussion
Chemistry Scheme 1 illustrates the synthesis of the acetohydrazide derivatives (1 -25). The compound acetohydrazide (III) was synthesised by refluxing methyl ester of sulindac and hydrazine hydrate (99%) in the presence of absolute ethanol. Sulindac methyl ester (II) was synthesised from sulindac by refluxing in methanol with concentrated sulphuric acid according to the reported procedure.
The acetohydrazide (III) was used as a starting material for the synthesis of various substituted sulindac hydrazone derivatives (1 -25). The acetohydrazide (III) was reacted substituted benzaldehydes in ethanol and glacial acetic acid as a catalyst. The acetohydrazide was characterised by the appearance of singlet peak for the -NH 2 protons at d 3.38 ppm and broad singlet for the CONH proton at d 9.30 ppm. The disappearance of NH 2 protons at d

Analgesic activity
Tail flick method Tail flick method was used for testing the analgesic activity of compound 3. After 30 min, the % inhibition of test compound 3 was 3.33% as compared to reference drug sulindac with 13.79%. The testing compound 3 expressed significant activity of 46.66% inhibition compared to reference drug sulindac with 82.75 inhibition after 60 min (Table 2). Highly significant analgesic activity (63.33% inhibition) was observed after 120 min for the compound 3 as compared to reference compound, sulindac.
Hot plate method Hot plate method was used for testing the analgesic activity. After 30 min, the % inhibition of test compound 3 was 15% as compared to reference drug sulindac with 41.30%.The testing compound 3 expressed highly significant activity of 67.50% inhibition compared to reference drug sulindac with 86.95 inhibition after 60 min (Table 3).
Acetic acid -induced writhing Acetic acid induced writhing was used for testing analgesic activity. The compound 3 expressed significant analgesic activity of 58.56% inhibition compared to reference drug sulindac with 74.1% inhibition (Table 4).

Yeast-induced hyperthermia
Yeast-induced hyperthermia was used for testing analgesic activity in mice. The compound 3 expressed significant analgesic activity of after 120 min as compared to reference drug sulindac (Table 5).

Anti-inflammatory activity
Based on the in vitro antioxidant activity, compound 3 was selected for in vivo anti-inflammatory activity by carrageenan induced paw edoema method. The anti-inflammatory activity of tested compound 3 after 3 and 5 h ranges from 50.52 to 50.54%, respectively compared to reference drug sulindac, which showed 65.18% after 3 h and 65.02% after 5 h ( Table 6). Because of hydrazide substitution of para dimethylaminophenyl group, compound 3 presented significant anti-inflammatory activity.  The significant anti-inflammatory activity of compound 3 was observed

Ulcerogenic activity
The compound 3 was further evaluated for ulcerogenic and lipid peroxidation activity. Equimolar concentration of compound 3 and reference drug sulindac was administered as oral doses to the examined animals. Compound 3 demonstrated highly significant ulcerogenic reduction activity 4.33 ± 0.40 (40.90% inhibition) as compared to reference drug sulindac with 6.83 ± 0.40 (6.81% inhibition) ( Table 7).
Compound 3 with para dimethylaminophenyl substitution was found to be the most potent anti-inflammatory and analgesic derivative as well as a significant gastric sparing activity.
MDA, NP-SH, total protein content in gastric tissue It has been known that the reduction of malondialdehyde (MDA) content in gastric tissue is consistent with the reduction of ulcerogenic activity. Compound 3 has shown a maximum reduction in the lipid peroxidation and gastric ulceration. The MDA content in compound 3 group was found to be (1.06 ± 0.02 nmol/g) as compared to the 80% ethanol group (6.53 ± 0.56 nmol/g) ( Table 9).

Toxicity of compound 3
Karber method was used to determine the LD 50 of compound 3. The LD 50 of compound 3 was found to be 131 mg/kg (Table 10).

Effect of compound 3 on COX-2 mRNA expression in cisplatin induced hepatotoxicity in rats
The expression of COX-2 is increased by pro-inflammatory meditaors 28 . Furthermore, the previous studies have revealed the increased COX-2 mRNA expression in cisplatin-induced hepatotocity 29 . Hence we characterised the effect of compound 3 administration on cisplatin-induced hepatotoxicity by measuring COX-2 mRNA levels in cisplatin-treated rats, untreated control rats, DMSO-treated rats, cisplatin-compound 3 treated (20 and 40 mg/ kg) rats. Cisplatin-induced hepatotoxicity was categorised by a significant increase in liver tissue gene expression of COX-2 (P < 0.001), when compared to normal values. Co-treatment of cisplatin-treated rats with compound 3 significantly downregulated liver tissue gene expression of COX-2 (P < 0.001), as compared to the cisplatin values. Furthermore, co-treatment of cisplatin-treated rats with 40 mg/kg dose of compound 3 significantly normalised liver COX-2 gene expression ( Figure 2).

Compound 3 administration induced reticence of COX-2 protein expression in liver tissue
We further examined the anti-inflammatory effect of compound 3 in liver tissues in cisplatin administrated rats by measuring protein expression of inflammatory mediators. In this regards, a substantial increase in COX-2 protein level was observed in the liver tissues isolated from cisplatin-administrated rats against untreated, control rats. The animals treated with compound 3 prior to cisplatin treatment resulted in a significant reduction in COX-2 protein expression when compared to cisplatin-treated group (Figure 3).   The binding affinities of all the synthesised compound (1 -25) and Sulindac with Cox-2 were predicted using geometry-based molecular docking. The binding affinities of the compounds were demonstrated by PatchDock server with ACE values (Table 11). The lower ACE value is considered to be associated with better ligand affinity with the enzyme 30 . Compound 3 showed lower ACE values (À440.29 kJ/mol) as compared to the standard drug Sulindac (À325.99 kJ/mol). This data suggests a better binding affinity of the compound 3 with Cox-2 protein as compared to the Sulindac (Figure 4). The visualisation of the docked complex revealed two hydrogen bonds by compound 3 with Cys-32 and Tyr-116, along with many hydrophobic interactions ( Figure 5). While Sulindac showed only the hydrophobic interaction with the protein in the docked complex ( Figure 6). It can be speculated that the formation of hydrogen bonds by compound 3 is responsible for the better affinity of the compound as compared to Sulindac.

Conclusion
In conclusion, novel sulindac hydrazide derivatives (1-25) were synthesised in good yields and characterised by spectral data and elemental analysis. Chemical modification to the sulindac acetohydrazide scaffold resulted in twenty-five derivatives with significant antioxidant activity. Compound 3 containing para dimethylaminophenyl group was found to be having highly potent antioxidant activity. It was evaluated for in vivo antiinflammatory and various analgesic and ulcerogenic activity different animal models and was found to be significant antiinflammatory and analgesic agent. It demonstrated significant ulcerogenic reduction activity in ethanol and indomethacin model. The LD 50 of compound 3 was found to be 131 mg/kg. Compound 3 with para dimethylaminophenyl substitution was found to be the most potent anti-inflammatory and analgesic derivative as well as a significant gastric sparing agent. Compound 3 significantly downregulated liver tissue gene expression of COX-2. The animals treated with compound 3 prior to cisplatin treatment resulted in a significant reduction in COX-2 protein expression when compared to cisplatin-treated group.

Disclosure statement
No potential conflict of interest was reported by the author(s).