Synthesis, characterization, COX1/2 inhibition and molecular modeling studies on novel 2-thio-diarylimidazoles

Heterocyclic compounds with diaryl substituents have been a milestone approach for selective cyclooxygenase 2 (COX 2) inhibition by bioisosteric replacements and modifications. It is also known that thiazole derivatives have different pharmacological activities. In this study, nine novel 2-[(1,5-diphenyl-1H-imidazole-2-yl)thio]-N-(thiazole-2-yl)acetamide derivatives (Compound 1–9) were synthesized via the reaction of 1,5-disubstitued phenyl-imidazole-2-thiole and N-thiazole acetamide. The inhibitory effects of COX-1 and COX-2 enzymes were tested for the synthesized compounds. Enzyme-ligand interactions of the most active compound on COX subtypes were elucidated by molecular modeling studies. The percent inhibitory effect for compound 1, which is the naked derivative among all the compounds, was found to be the most active on COX-2 at 10 μM concentration (C1COX-2: 88%, SC-560COX-2: 98.2%, C1COX-1: 60.9%); whereas compound 9 showed the highest inhibitory effect and was found to be the most selective derivative on COX-1 isoenzyme (C9COX-1: 85%, DuP-697COX-1: 97.2%, C9COX-2: 57.9%)

Protein structure preparation, GRID files used in docking operations, docking/scoring with various algorithms and visualization were carried out with Maestro (Schrodinger Inc, USA) software and related subunits.All processes requiring a computer were carried out using workstations within the School of Pharmacy of Istanbul Medipol University.

Preparation of the ligand set to be used
The structure of all the components has been drawn in three dimensions with the help of Maestro (Schrodinger Inc, USA) program modules.The energies of the structures were minimized by using the ligprep module, their pH 7 (+/-2) ionized forms and tautomers were prepared, repeating and salt form ones were removed.

Preparation of the target to be used
Crystallographic data of "cyclooxygenase 2 structure complexes with a selective inhibitor SC-558" (Pdb ID; 1CX2) and "Cyclooxygenase 1 crystal structure in complex with Celecoxib" (Pdb ID; 3KK6) were downloaded in PDB format.[19,20].Subsequently, the hydrogens of the structure were added with the help of the "protein preparation wizard" under Maestro (Schrodinger Inc, USA) program and the bond orders were rearranged in accordance with the software.After determining the appropriate side chain positioning of amino acids and possible intraprotein H bonds, the hydrogens of the entire structure were minimized using Optimized Potentials for Liquid Stimulations (OPLS) 2005 energy parameters.

Preparation of GRIDs belonging to the active region
Using the Glide-Grid preparation module of Maestro (Schrodinger Inc, USA), various interaction maps of the active region were prepared.During the preparation of the file, the original ligands were taken as the center point, and Van der Waals radius scaling factors were kept as default.No other restrictions were added.

Docking and scoring
With the help of optimized GRID files, first original ligands (for the purpose of internal validation) then previously prepared ligands were docked by HTVS and XP protocols of Glide-docking interface (Maestro, Schrodinger Inc).The poses of original ligands and compound with highest biological activities were visually inspected.Docking/ maximum e-model scores of XP protocol were recorded.

Preparation of interaction charts and figures
The graphics and interactions resulting from the calculations were saved in high resolution picture format and placed in the text content with the help of Maestro (Schrodinger Inc) software 2D/3D visualization tools.
The methylene (-CH 2 -) group adjacent to the carbonyl and sulfur underwent chemical shift as expected in all compounds, giving 2H and singlet peak in the range 4.05-4.11ppm.NH hydrogen in amide structure found in all our structures gave a singlet and flat peak in the range of 12.40-12.42ppm.The molecules consist of 11-13 protons in total belonging to their aromatic rings (thiazole, imidazole, and phenyl).All the protons belonging to aromatic structures were observed in all spectrums.The only hydrogen in the imidazole ring was observed as a separate singlet in compounds 1, 2, 4, 5, 6, 7, and 9, with the other compounds (3, 8) being overlapped with thiazole and/or phenyl aromatic hydrogens.2 hydrogen on the thiazole were observed as 2H and doublet in compounds 1, 5, and 6.It is overlapped with other aromatic hydrogens in other compounds.The aromatic methyl group containing compounds 4, 5, and 6 were observed with 2.35 ppm as 3H values.In the compounds 2, 5, 7, 8, and 9 containing methoxyphenyl groups, the methyl group underwent chemical shift and was observed between 3.69 and 3.79 ppm.Methoxy groups on the phenyl at 1st position of imidazole were observed at 3.79 ppm, and the ones on the phenyl at fifth position of imidazole were observed at 3.69 ppm.

COX-1 and COX-2 enzyme activity results
In order to test the COX-1 and COX-2 enzyme inhibition effects of the compounds, procedures were carried out as recommended by the manufacturer.For this purpose, enzymes and compounds were left to incubation, then 200 μL of enzyme containing tested compounds was put into wells at a final concentration of 10 μM.The enzyme activity for each compound was carried out triplicate.One-way ANOVA (one-way ANOVA) test was applied to analyze the consistency between replicates of the experimental groups.The significance of the groups compared to the standard inhibitors (SC-560 for COX-1, DuP-697 for COX-2) group was evaluated by Dunnett's test.COX-1 and COX-2 enzyme inhibition in 10 μM concentration of SC-560 and DuP-697 used as standard inhibitory compounds.Standards and nine compounds enzyme activity results are given in Figures 6 and 7 and Table 1.The effects of compounds and SC-560 (98.2% inhibition) used as standard inhibitor on COX-1 enzyme activity were determined.Compound 9 and compound 7 were found to be the most active on COX-1 enzyme with 85.5 % and 82.4 % inhibition respectively.Compound 2 showed 63% and compound 4 showed approximately 65% inhibitory activity on COX-1.The obtained activity values were statistically significant (at least p < 0.05).
The effects of the synthesized compounds and DuP-697 (97.2% inhibition) used as standard inhibitor on COX-2 enzyme activity were determined.Among the tested compounds, compound 1, compound 6, and compound 4 were found to be the most active compounds on COX-2 enzyme with 88.5%, 82.8%, and 82.7% inhibition, respectively.The least active compound was found to have 57% inhibition value among all of the tested products.The obtained activity values were statistically significant (at least p < 0.05).

Molecular modeling
For validation of docking studies, original ligands SC-558 and Celecoxib were successfully redocked to their crystallographic data by the same protocols used for designed ligands without any restrictions.The RMSD values of these ligands compared to their original positions are 1.1097 and 0.3910, respectively.

COX-2 molecular interactions
The 2D and 3D interaction potentials of the docked Compound 1-COX-2 enzyme were analyzed.Interactions between ring nitrogen, carbonyl, phenyl, secondary amine functional groups and active site amino acids are detailed in Figures 8  and 9.
In the compound 1; the phenyl group at 5th position of imidazole interacts with the Arg120 amino acid via hydrophobic effects in the active site.The 4th carbon atom attached to the same phenyl ring (at 5th position of imidazole) is in hydrophobic interaction with Leu 531, while the carbons of same phenyl (number 2, 3, and 5) are in hydrophobic interaction with Val 349.The amino acids Ala 516 and Thr 94 in the active site hydrophobically interact with the 5th carbon of the thiazole ring in its main structure.
Phenyl group on the first position of imidazole establishes hydrophobic interactions with the Trp 387 and Val 523 on COX-2 active site.The N atom of the acetamide moiety in the structure of compound 1 forms hydrogen bond with the Trp 387 and Ser 353 amino acids.

COX-1 molecular interactions
The 2D and 3D interaction potentials of the docked Compound 1-COX-1 enzyme were analyzed in Figures 10 and 11.Interactions between ring nitrogen, carbonyl, phenyl, secondary amine functional groups and active site amino acids are detailed.
Phenyl ring in the 5th position of the imidazole form hydrophobic interactions with Val116, Leu115, Leu531, Leu359 and Tyr355 in the COX-1 active site.The Val344 interacts hydrophobically with the S atom of the thiazole ring in the compound 1.Phenyl ring on the first position of imidazole forms hydrophobic interaction with Ile 517 and Tyr355 (Figure 11).Docking/maximum e-model scores of orginal ligands and compound 1 are also calculated during the docking process.(Table 2) Calculated drug-likeness properties of the synthesized compounds were tabulated in Table 3.

Discussion
All of the structure characterizations of the synthesized compounds were performed with spectroscopic methods. 1 H NMR, 13 C NMR, and MS spectra resulted in expected chemical structures.For compounds 3, 6, and 9, the fluorine couplings were determined.Aromatic fluorine splits were observed around 240-250 Hz for the aromatic carbon directly bound to F, and 20-25 Hz for first neighborhood C-H carbons of the phenyl.Accordingly, C-F carbon splitting (160 and 163 ppm, J: 245 Hz) is consistent with the literature data.C-H carbon splittings in the neighborhood of fluorine (115.8 and 116.02,J: 21.6 Hz) also met the literature and expected data [21].
The COX-1 and COX-2 inhibition activities of the compounds are studied against the SC-560 and DuP-697 standards, respectively.Enzyme-active site interactions of the active compound were investigated with the support of molecular modeling studies which were internally validated previously with satisfactory RMSD findings.Compound 1 has selective COX-2 enzyme activity; however, COX-1 inhibition was found to be rather low (60.9%) compared to standard SC-560 (98.2% inhibition).When examined in terms of selective activity, the binding of methoxy and fluorine functionalities to the p-position of the phenyl ring in the 5th position to the imidazole ring, partially favors the selective COX-2 inhibition.Even if the scoring functions do not indicate an expected activity in general, particularly solvation parameters included glide e-model scores showed parallel relations with the activity results.
The -NH-group of the acetamide in the structure of all compounds makes the H-bond with Ser 353 of COX-2.This situation does not occur in the COX-1 structure.The difference in the binding sites between the two enzymes is not the separation of amino acid sequences, but rather the volumetric differences.This situation manifests itself as the difference in size although there are hydrophobic pockets in both structures.The substituents in the phenyl rings on the structure of the compounds are important in terms of selectivity, rather than showing the H-bond acceptor or donor feature, they fit into these pockets by volume.The most active compounds for COX-2 are 1, 4, 6, and 7, those having H or F on the Descriptor: Predicted central nervous system (CNS) activity on a scale of -2 (inactive) to +2 (active).logPo/w: Predicted octanol/ water partition coefficient (-2.0 to 6.5).log HERG : Predicted IC 50 value for blockage of HERG K + channels (below -5) logBB: Predicted brain/blood partition coefficient (-3.0 to 1.2).logK HSA : Prediction of binding to human serum albumin (-1.5 to 1.5).Human Oral Absorption (HOA): Predicted qualitative human oral absorption measured as 1, 2, or 3 for low.Percent HOA: high if >80% and poor if <25%.
phenyl ring attached at the 5th position of imidazole.Also, the compound with the lowest inhibition is compound 9 on COX-2.Selective inhibition of COX-2 can be reduced by electron withdrawing groups located on phenyl ring.In the COX-1inhibitory activity, compounds 7-9 was found to be the most active ones, including methoxy group on the phenyl attached at 1st position of imidazole.In this study, COX inhibition levels were measured by a single concentration.Thus, it is not possible to directly correlate with similar studies.However, at 10 µM concentration, inhibition rate is up to 50% for COX-1 and up to 70% for COX-2.Selective COX-2 inhibitors usually show their effect at submicromolar level.Similar compounds with trisubstituted fivemembered cycles have been synthesized before, and the IC 50 values are in a range of 3-50 µM concentration, where the IC 50 value of celecoxib was found to be 0.30 µM [22][23][24][25].The existence of electron-withdrawing substituents attached to the ring system reduces the hydrophobicity of systems that fit into hydrophobic pockets, and the degree of bonding decreases due to pi-pi interactions.This situation is reflected in the activity as nonselectivity.
As a result of these studies, it was determined that the compounds bearing substituted imidazolyl acetamide derivative showed selective COX-2 inhibition effect within limits calculated drug-likeness properties.In future studies, it is planned to work with different concentrations, to try ring systems that can show similar properties, as well as to obtain more effective compounds by increasing the variety of substituents, as well as to examine the kinetic properties of these compounds.

Table 1 .
Percent enzyme inhibition values ± SD of SC-560 and DuP-697 used as standard inhibitors of the synthesized compounds.

Table 2 .
Docking and Glide e-model scores of original ligands and compound 1 calculated from XP docking protocol.

Table 3 .
Drug-likeness properties of compounds.