Synthesis, anti-inflammatory, cytotoxic, and COX-1/2 inhibitory activities of cyclic imides bearing 3-benzenesulfonamide, oxime, and β-phenylalanine scaffolds: a molecular docking study

Abstract Cyclic imides containing 3-benzenesulfonamide, oxime, and β-phenylalanine derivatives were synthesised and evaluated to elucidate their in vivo anti-inflammatory and ulcerogenic activity and in vitro cytotoxic effects. Most active anti-inflammatory agents were subjected to in vitro COX-1/2 inhibition assay. 3-Benzenesulfonamides (2–4, and 9), oximes (11–13), and β-phenylalanine derivative (18) showed potential anti-inflammatory activities with 71.2–82.9% oedema inhibition relative to celecoxib and diclofenac (85.6 and 83.4%, respectively). Most active cyclic imides 4, 9, 12, 13, and 18 possessed ED50 of 35.4–45.3 mg kg−1 relative to that of celecoxib (34.1 mg kg−1). For the cytotoxic evaluation, the selected derivatives 2–6 and 8 exhibited weak positive cytotoxic effects (PCE = 2/59–5/59) at 10 μM compared to the standard drug, imatinib (PCE = 20/59). Cyclic imides bearing 3-benzenesulfonamide (2–5, and 9), acetophenone oxime (11–14, 18, and 19) exhibited high selectivity against COX-2 with SI > 55.6–333.3 relative to that for celecoxib [SI > 387.6]. β-Phenylalanine derivatives 21–24 and 28 were non-selective towards COX-1/2 isozymes as indicated by their SI of 0.46–0.68.


Introduction
Nonsteroidal anti-inflammatory drugs (NSAIDs) are the drugs of choice for the treatment of inflammation and pain 1,2 . NSAIDs are cyclooxygenase inhibitors and they include the COX-1 and COX-2 enzymes [1][2][3] . The cyclooxygenase isozymes are responsible for the conversion of arachidonic acid to eicosanoids (prostaglandins) 1-3 . In this process, the inhibition of the COX-2 isozyme elucidates the anti-inflammatory effects of NSAIDs 1-3 while the inhibition of the COX-1 isozyme is responsible for the major side effects displayed by NSAIDs, such as gastrointestinal implication 1,2 . Alternatively, selective COX-2 inhibitors, including compounds that contain the pyrazole ring system, such as celecoxib (A) and SC-558 (B) 4 ( Figure 1), display anti-inflammatory activity with improved gastric profile protection compared to NSAIDs 4 . Moreover, celecoxib, a COX-2 inhibitor (Figure 1(A)) that displays an anti-inflammatory effect, has been investigated as an antitumor agent 5 .
Here, we report the design and synthesis of novel cyclic imides containing 3-benzenesulfonamide, acetophenone oxime, and b-phenylalanine scaffolds. The designed cyclic imides (Figure 1(I)) were subjected to various analyses to: (i) investigate their in vivo anti-inflammatory and ulcerogenic activities, and their in vitro COX-1/2 inhibitory effects; (ii) study their in vitro cytotoxicity activity; (iii) explore their structure-activity relationships (SAR) based on their in vivo anti-inflammatory effects and in vitro COX-1/2 inhibitory activity with molecules containing substituted cyclic imides; (iv) compare the biological effects of 3-benzenesulfonamide to those of acetophenone oxime and b-phenylalanine based on in vivo anti-inflammatory and COX-1/2 inhibition; and (v) conduct a molecular docking study of the target derivatives to investigate their binding with the COX-2 isozyme.
imides to explore and compare the efficacy of the substituted cyclic imides for inhibiting COX-1/COX-2 and evaluating their antiinflammatory activity.

Ulcerogenicity
The levels of ulcerogenic activity for the most active cyclic imides 4,9,12,13,18, along with the reference drugs, diclofenac and celecoxib, were determined according to the reported technique (Table 3) 6a,17,18 . Evidently, the phthalimide derivatives 9 and 18 had very low levels of ulcerogenic activity relative to diclofenac and celecoxib, whereas phthalimides 4, 12, and 13 had an ulcerogenic level similar to celecoxib and less than that of diclofenac (Table 3).

Cytotoxicity
Six compounds were selected by the National Cancer Institute (Bethesda, MD) based on structural variations to evaluate the in vitro cytotoxicity of compounds 2-6 and 8. Thereafter, the results were compared to those of the reference drug, imatinib, as shown in Table 4. These compounds, which are derivatives of cyclic imides bearing 3-benzenesulfonamides 2-6 and 8, were administered in single doses of 10 mM in a full NCI 59 cell line panel assay. These cell lines were obtained from nine different organs including leukaemia, non-small cell lung, colon, CNS, melanoma, ovarian, renal, prostate, and breast 19 . The results are displayed in Table 4 and expressed as the percentage growth inhibition (GI %) caused by the test compounds. Based on the results, the positive cytotoxic effects (PCE) of the tested cyclic imides 2-6 and 8 at 10 lM were 2/59-5/59, while the reference drug, imatinib, had a PCE of 20/59 (Table 4).

Molecular docking studies
The molecular modelling technique was used to establish and understand the binding mode of the most bioactive compounds 21,22 . The selectivity of the cyclic imide analogues towards COX-2 was studied using a molecular docking protocol on the MOE 2008.10 programme obtained from Chemical Computing Group Inc. (Montreal, Canada) 23 .
Molecular docking was performed to examine the best interaction between the most active compounds, 9 and 18, and the COX-2 pocket binding site (Figure 2, lower panels). The crystallographic binding site on the COX-2 isozyme in complex with the SC-558 ligand, an analogue of celecoxib, was derived from Protein Data Bank (PDB code: 1CX2) (Figure 2, upper panels) 4 .
The scoring function and the hydrogen bond formed among these compounds and the surrounding amino acids were used to predict the interaction mode of compounds 9 and 18 in the putative active pocket of the COX-2 isozyme. Figure 2 shows the results of the docking studies for both compounds 9 and 18, which possess the common cyclic imide pharmacophore and a similar binding interaction, including H-bonding and hydrophobic interactions within the putative binding pocket. These compounds were inserted deep into the hydrophilic site of the COX-2 isozyme where the sulphonamide and oxime groups can interact with the hydrophilic pocket (His90, Gln192, and Arg513) through a network of hydrogen bonds; such binding interactions were similar to that 0.129 >387.6 a IC 50 value is the compound concentration required to produce 50% inhibition of COX-1or COX-2 for means of two determinations using an ovine COX-1/COX-2 assay kit (catalog no. 560101, Cayman Chemicals Inc., Ann Arbour, MI) and deviation from the mean is <10% of the mean value. b Selectivity index (COX-1 IC 50 / COX-2 IC 50 ).
of SC-558 co-crystallized in the COX-2 active site 4 . Additionally, the cyclic imide cores of compounds 9 and 18 were oriented at the top of the channel, which were involved in the hydrophobic interaction with amino acids, Ala527, Val349, Trp387, Val523, and Leu352. The terminal sulphonamide group (-SO 2 NH 2 ) was responsible for the stability of the docked compound 9 through its conserved role in the binding pocket through the formation of suitable H-bonds with the key amino acids, Arg513 (2.87 Å), His90 . Finally, the 5-nitro group of cyclic imide formed a non-classical H-bond with Leu531 (3.08 Å). The binding interactions described herein are typical of the pyrazolic prototype (SC-558), which is a selective inhibitor of COX-2. Altogether, such findings confirm the molecular design of the reported class of anti-inflammatory 1,3-isoindoledione scaffolds 10,11 .

Chemistry
Melting points (uncorrected) were recorded on a Barnstead 9100 Electrothermal melting apparatus (APS Water Services Corporation, Van Nuys, CA) while the IR spectra were recorded on a FT-IR Perkin-Elmer spectrometer (PerkinElmer Inc., Waltham, MA). The 1 H NMR and 13 C NMR were measured in DMSO-d 6 or CDCl 3 , on Bruker 700 or 500 and 176 or 125 MHz instruments, respectively (Bruker, Billerica, MA). Chemical shifts are reported in d ppm. Mass spectra were recorded on an Agilent 6320 Ion Trap mass spectrometer (Agilent Technologies, Santa Clara, CA). C, H, and N were analysed at the Research Centre, College of Pharmacy, King Saud University, Saudi Arabia. The results were within ±0.4% of the theoretical values. Compounds 4,8,9,[12][13][14][15][16][17][22][23][24] were prepared according to a previously reported procedure 6b,24 . 4.1.1. General procedure for the synthesis of 1,3-isoindolediones 2-10 A mixture of an equimolar amount of 3-benzenesulfonamide and acid anhydrides (5.0 mmol) was heated under reflux for 12 h in glacial acetic acid (15 ml) containing anhydrous sodium acetate (0.69 g, 5.0 mmol) (Scheme 1). The reaction mixture was cooled, filtered, and the obtained solid was washed with water, dried, and re-crystallized.

Anti-inflammatory screening
Anti-inflammatory assessment of the newly synthesised compounds was carried out using an in vivo rat carrageenan-induced foot paw oedema model, as reported previously 16,17 . Compounds 6,7,8,10,11,18, diclofenac, and celecoxib were tested at three different doses and their ED 50 was determined.

Ulcerogenicity measurement
Ulcerogenicity was evaluated according to a previously reported method 6a,17,18 . The number and total length of the ulcers for each animal were measured and their averages were calculated and used as the ulcer indices.

In vitro cyclooxygenase (COX) inhibition assay
To determine the relative ability of the test compounds and reference drugs to inhibit COX-1/COX-2 isozymes, we used the colorimetric COX (ovine) inhibitor screening assay kit (Cayman Chemicals Inc., Ann Arbour, MI), according to the manufacturer's instructions 3,10,11,20 .

Docking methodology
Molecular modelling studies were performed using the 2007.09 software from Chemical Computing Group Inc. (Montreal, Canada). The docking protocol was similar to that mentioned in our previous report 19 .