Synthesis, Molecular Docking and Anticancer Activity of Diflunisal Derivatives as Cyclooxygenase Enzyme Inhibitors

Cyclooxygenase enzymes play a vital role in inflammatory pathways in the human body. Apart from their relation with inflammation, the additional involvement of COX-2 enzyme with cancer activity was recently discovered. In some cancer types the level of COX-2 enzyme is increased indicating that this enzyme could be a suitable target for cancer therapy. Based on these findings, we have synthesized some new diflunisal thiosemicarbazides and 1,2,4-triazoles and tested them against androgen-independent prostate adenocarcinoma (PC-3), colon carcinoma (HCT-116), human breast cancer (T47D), breast carcinoma (MCF7) and human embryonic kidney (HEK-293) cell lines. Specifically, the diflunisal and thiosemicarbazide functionality are combined during the synthesis of original compounds anticipating a potency enhancement. Compounds 6, 10, 15 and 16 did not show cytotoxic effects for the HEK293 cell line. Among them, compounds 15 and 16 demonstrated anticancer activity for the breast cancer cell line T47D, whereas compounds 6 and 10 which are thiosemicarbazide derivatives displayed anti-tumourigenic activity against the PC-3 cell line, consistent with the literature. However, no activity was observed for the HCT-116 cancer cell line with the tested thiosemicarbazide derivatives. Only compound 16 displayed activity against the HCT-116 cell line. Therefore, it was speculated that the diflunisal and thiosemicarbazide functionalities potentiate anticancer activity on prostate cancer and the thiosemicarbazide functionality decreases the anticancer activity of diflunisal on colon cancer cell lines. In order to gain insight into the anticancer activity and COX-2 inhibition, molecular docking studies were carried out for COX-1 and COX-2 enzymes utilizing the newly synthesized compounds 15, and 16. Both 15 and 16 showed high selectivity and affinity toward COX-2 isozyme over COX-1, which is in agreement with the experimental results.


Introduction
Cyclooxygenase (COX) enzymes play an important role in the synthesis of prostaglandins. So far, three isoenzymes of COX have been discovered; COX-1, COX-2 and COX-3. The constitutive COX-1 isoenzyme which can be found throughout the body, where it is responsible for the protection of the to triazole functionality were observed between 155. .75 and 167.58-168.19 ppm, respectively. No carbon peaks of thiosemicarbazide C=O and C=S were observed; and this was accepted as a proof of the synthesis of 1,2,4-triazoles. In 13 C-NMR studies, the structure of compounds was confirmed to present in thione form with the observation of C=S peaks. HR-MS studies were performed on two selected triazole compounds, compounds 14 and 16. HR-MS studies confirmed the molecular weights and empirical formulas within 8 mmu. The molecular ion peaks were observed as expected. The anticipated compound fragments were observed in the mass spectra. The main observed fragments of the two compounds were their [M-S] + , [M-OH] + and [M-RNCS] + peaks ( Figure 1). 1 H-NMR spectral data also confirmed the formation of triazole compounds. The N1, N2 and N4 protons of the thiosemicarbazide were not detected in the 1,2,4-triazole compounds. Instead, N-H peaks of 1,2,4-triazole were observed between 13.76-14.16 ppm. There was no peak between 1.0-3.7 ppm which might be attributed to a -SH function and this finding also supported the presence of the thione form of the 1,2,4-triazole compounds 11-16. All the other aromatic protons of the compounds were observed in the expected regions. In 13 C-NMR studies, all the carbons were observed in the expected regions. The C=N and C=S carbons that belong to triazole functionality were observed between 155. .75 and 167.58-168.19 ppm, respectively. No carbon peaks of thiosemicarbazide C=O and C=S were observed; and this was accepted as a proof of the synthesis of 1,2,4-triazoles. In 13 C-NMR studies, the structure of compounds was confirmed to present in thione form with the observation of C=S peaks. HR-MS studies were performed on two selected triazole compounds, compounds 14 and 16. HR-MS studies confirmed the molecular weights and empirical formulas within 8 mmu. The molecular ion peaks were observed as expected. The anticipated compound fragments were observed in the mass spectra. The main observed fragments of the two compounds were their [
Firstly, the half maximal inhibitory concentration (IC 50 ) values of all the synthesized compounds were determined for all the cell lines used in this study by a MTS assay. Compounds for which the IC 50 value against HEK293 cell line was twice as high as the IC 50 value for at least one cancer cell line (compound 6, 10, 15, 16) were considered active (Table 1) and included into structure-activity analysis.
Further experiments were performed by using selected active compounds. There are reliable studies about the anti-colon cancer activity of diflunisal and the anti-prostate cancer activity of the thiosemicarbazide functionality [22]. Therefore, here prostate cancer (PC-3) and colon cancer (HCT-116) cell lines were selected to study the anticancer activity of diflunisal and thiosemicarbazide. We combined the diflunisal and thiosemicarbazide functionalities in order to increase the anticancer activity of the resulting compounds. The expected results were obtained for the PC-3 cancer cell line (the compounds showed activity, with IC 50 values as 41.8 µM for compound 6 and 11.7 µM for compound 10), however, no activity was observed for the HCT-116 cancer cell line with the thiosemicarbazide derivatives. Therefore, we assumed that diflunisal and thiosemicarbazide functionality enhance anti-prostate cancer activity but the thiosemicarbazide functionality decreases the activity of diflunisal on colon cancer cell lines [15]. Comparing the differences between the active thiosemicarbazide molecules, alkyl substitution increased the activity; whereas halogen substitution showed more decreased activity than compound 10.
Similar to the anticancer activities of thiosemicarbazides, 1,2,4-triazole-3-thiones were reported to have diverse biological activities, including anticancer activity [17]. There are several studies reporting the activity of the triazole functionality against breast cancer. In fact, letrozole and anastrozole are anticancer drugs (aromatase inhibitors) in breast cancer treatment on the market. In this study, we used the MCF7 and T47D breast cancer cell lines. Both cancer cell lines were isolated from tumours located in breast, but they differ from each other in a mutation in the p53 gene. The wild-type form of the p53 gene is expressed in the MCF7 cancer cell line, whereas in the T47D cancer cell line the p53 gene is mutated [33]. Among the tested compounds, compounds 15 and 16 which possess a triazole functionality showed activity against the T47D cell line, whereas for the MCF7 cancer cell line none of the compounds displayed any activity. These contrasting effects could be related to differential p53 expression levels. The presence of the wild-type form of p53 in the MCF7 cell line may promote its resistance against all of the compounds tested in this study. Furthermore, it is well established in the literature that the MCF7 cell line has multidrug resistance compared to the T47D cell line because of excessive expression of P-glycoprotein drug transporter gene and having an estrogen receptor subunit alpha rather than estrogen receptor subunit beta, unlike T47D [34][35][36]. It can be deduced from the results that compounds bearing a methoxy substituent are more important for T47D activity than halogen substituted ones. A compound bearing bromo substitution showed activity, with an IC 50 value of 43.4 µM (compound 15) and a compound bearing methoxy substitution showed activity with an IC 50 value of 27.3 µM (compound 16) for the T47D cancer cell line. Thus, it can be concluded that electron donating groups increase the anticancer activity on the T47D cancer cell line. When comparing the activity of these triazole compounds on other cancer cell lines, compound 16 showed activity against the HCT116 cancer cell line with an IC 50 value of 6.2 µM and compound 15 did not show any activity to the other cancer lines (PC-3 and HCT116). The differences between these results could be related with the cancer cell types. The HCT116 cell line is known to have K-RAS activation and PIK3CA mutation [37]. Because of that, this cancer cell line is resistant to a number of anticancer drugs and even epidermal growth factor inhibitors [38,39]. The PC-3 cancer cell line differs from other prostate cancer cell lines in its synthesis of androgen receptor (AR) and prostate specific antigen (PSA), indicating that survival does not depend on androgen [40]. Furthermore, PC-3 is a CD44 positive cancer cell line and therefore, it is thought to have stem cell features [41]. Like the HCT116 cancer cell line, PC-3 is also resistant to many anticancer drugs. However, we still found immensely significant activity from compound 16 against HCT116 cancer cell line. In addition to that, the triazole functionality did not possess any activity against the PC-3 cancer cell line., and theIC 50 values of the PC-3-active compounds demonstrated that the thiosemicarbazide functionality is more significant for this cell line.
In conclusion, the comparison of the compounds 6 and 10 indicates that alkyl substitution increases the anticancer activity on the PC-3 cell line. On the other hand, only bromophenyl derivatives showed activity in this study; which proved the importance of the bromine atom instead of chlorine or fluorine. Compound 16 has a triazole functionality and has methoxyphenyl substitution. Taking everything said into account, regarding the activity against colon cancer, small molecules are fundamental and electron donating groups are central for the anticancer activity. Compounds 15 and 16 showed the best activity on the T47D breast cancer cell line as these compounds have triazole functionalities. When the triazole is replaced by a thiosemicarbazide, the activity decreases as was observed in compound 6. For the anticancer activity against breast cancer line, the triazole functionality has significant importance compared to the thiosemicarbazide one.

Effect of the Synthesized Compounds on DNA Synthesis
Determination of IC 50 values of the compounds was subsequently followed by observation of newly synthesized DNA in the cells. It is well-known that DNA synthesis is correlated with cell proliferation [44] and 5-ethynyl-2 -deoxyuridine (EdU) staining enables one to visualize newly synthesized DNA to determine cell proliferation. EdU is a thymidine analogue that integrates DNA during DNA synthesis, therefore here it stained only newly synthesized DNA during the experimental procedure (in green fluorescence, Figure 2). 4 ,6-Diamidino-2-phenylindole (DAPI) stain binds strongly to A-T rich areas in DNA and passes through cell membranes easily, thereby it stains nucleus of the all cells regardless of whether the DNA is newly synthesized or not (in blue, Figure 2). It was revealed that COX-2 enzyme inhibition via the newly synthesized compounds decreases the rate of DNA synthesis in the HCT116, PC-3 and T47D cell lines, whereas the DNA synthesis in HEK293 cells did not appear to be affected by the treatment of any of the compounds found active based on IC 50 values ( Figure 2). These results indicated that the rate of DNA synthesis decreased dramatically in the presence of our newly synthesized compounds in cancer cell lines compared to the healthy cell line HEK293.

Effect of Synthesized Compounds on Apoptosis
Evaluation of the effect of the synthesized compounds on apoptosis was determined by the FITCH-AnnexinV/PI protocol. Damaged cell membrane of necrotic cell enables propidium iodide (PI) to bind DNA and induces a red fluorescence, although it is excluded by non-damaged cells, select live and apoptotic cells. FITCH-Annexin V has the ability to bind to phosphatidylserine which is a specific marker of apoptosis. Figure 3 demonstrates the results of FITCH-Annexin V/PI of PC-3, T47D, HCT116 and HEK293 cell lines after treatment with the COX-2 inhibitors synthesized in this study. The lower left panel of the quadrant of the cytograms indicates live cells due to exclusion of PI and non-bound Annexin V. select live and apoptotic cells. FITCH-Annexin V has the ability to bind to phosphatidylserine which is a specific marker of apoptosis. Figure 3 demonstrates the results of FITCH-Annexin V/PI of PC-3, T47D, HCT116 and HEK293 cell lines after treatment with the COX-2 inhibitors synthesized in this study. The lower left panel of the quadrant of the cytograms indicates live cells due to exclusion of PI and non-bound Annexin V.  Figure 3A). Although slightly more necrotic and late apoptotic cells were observed with the treatment of HEK293 with compound 10, it was not significant when compared to its effect on PC-3 cancer cells. PC-3 cancer cells undergo apoptosis highly with the treatment of compound 6 and compound 10 when it is compared to non-treated negative control ( Figure 3B). Compound 16 with a triazole functionality had an extensive apoptotic effect on HCT116 cancer cell line ( Figure 3C). Most of the  Figure 3A). Although slightly more necrotic and late apoptotic cells were observed with the treatment of HEK293 with compound 10, it was not significant when compared to its effect on PC-3 cancer cells. PC-3 cancer cells undergo apoptosis highly with the treatment of compound 6 and compound 10 when it is compared to non-treated negative control ( Figure 3B). Compound 16 with a triazole functionality had an extensive apoptotic effect on HCT116 cancer cell line ( Figure 3C). Most of the HCT116 cells went through early and late apoptosis when treated with this compound ( Figure 3C), whereas it resulted in necrosis in the T47D cancer cell line. Furthermore, compound 15 led to cell death in T47D cell line largely through late apoptosis following early apoptosis indicating that these cells went through programmed cell death quite rapidly and thereby started to lose their cell membrane integrity. To sum up, all the compounds HCT116 cells went through early and late apoptosis when treated with this compound ( Figure 3C), whereas it resulted in necrosis in the T47D cancer cell line. Furthermore, compound 15 led to cell death in T47D cell line largely through late apoptosis following early apoptosis indicating that these cells went through programmed cell death quite rapidly and thereby started to lose their cell membrane integrity. To sum up, all the compounds that are tested against to different cancer cell lines that were found to be active according to their IC50 values demonstrated higher cytotoxicity in tumor cells than normal cells (HEK293).

Docking Studies
In an attempt to explain the biological activity and the difference in selectivity profiles of the newly synthesized 1,2,4-triazole-3-thiones toward COX subtypes based on their orientation and binding patterns, the molecules were docked into the active sites of COX-1 and COX-2, respectively, using AutoDock 4.2 [45].
Most of the compounds were found to be potential inhibitors of both COX-1 and COX-2 enzymes ( Table 2). As expected, all of the tested compounds showed more affinity to COX-2, rather than COX-1 enzyme. The COX-2 enzyme activities of the compounds correlate their anticancer activity results. According to the docking studies, compounds 15 and 16 showed good binding affinity on COX-2 enzyme with free energies of binding, ∆G = −10.57 kcal/mol and −9.60 kcal/mol, respectively ( Table 2). Halogen substitution appeared to have considerable importance but only flouro substitution affected the results in a positive way. When fluorine substitution is changed for chlorine or bromine atoms, the selectivity decreased (compound 14). Interestingly, alkyl substitution also favored COX-2 selectivity. The best COX-2 selectivity was shown by compound 13. For comparison, 3D representation of predicted binding mode and protein-ligand interactions of diflunisal in the active site of COX-2 enzyme is given in Figure 4. The 3D picture of compound 15 and 16 in the active site COX-2 enzymes are shown in Figure 5A  In an attempt to explain the biological activity and the difference in selectivity profiles of the newly synthesized 1,2,4-triazole-3-thiones toward COX subtypes based on their orientation and binding patterns, the molecules were docked into the active sites of COX-1 and COX-2, respectively, using AutoDock 4.2 [45].
Most of the compounds were found to be potential inhibitors of both COX-1 and COX-2 enzymes ( Table 2). As expected, all of the tested compounds showed more affinity to COX-2, rather than COX-1 enzyme. The COX-2 enzyme activities of the compounds correlate their anticancer activity results. According to the docking studies, compounds 15 and 16 showed good binding affinity on COX-2 enzyme with free energies of binding, ΔG = −10.57 kcal/mol and −9.60 kcal/mol, respectively ( Table 2). Halogen substitution appeared to have considerable importance but only flouro substitution affected the results in a positive way. When fluorine substitution is changed for chlorine or bromine atoms, the selectivity decreased (compound 14). Interestingly, alkyl substitution also favored COX-2 selectivity. The best COX-2 selectivity was shown by compound 13. For comparison, 3D representation of predicted binding mode and protein-ligand interactions of diflunisal in the active site of COX-2 enzyme is given in Figure 4. The 3D picture of compound 15 and 16 in the active site COX-2 enzymes are shown in Figure 5A

General
All the chemicals were purchased from Merck (Darmstadt, Germany) or Sigma-Aldrich (St. Louis, MO, USA). Reactions were monitored by TLC on silica gel plates purchased from Merck. Melting points of the synthesized compounds were determined in a IA 9300 melting point apparatus (ThermoScientific, Staffordshire, UK) and are uncorrected. The purity of the compounds was verified by TLC, HPLC and elemental analysis. Elemental analysis was performed on a CHNS-932 apparatus (LECO, St. Joseph MI, USA). FT-IR spectra were recorded on a FT-IR-8400S spectrophotometer (Shimadzu, Columbia, MD, USA). NMR spectra were recorded on a Bruker spectrometer (Billerica, MA, USA) (300 MHz for 1 H-NMR and 75 MHz for 13 C-NMR, decoupled). Data are reported as follows: chemical shift, multiplicity (b.s.: broad singlet, d: doublet; m: multiplet, s: singlet, and t: triplet), coupling constants (Hz), integration. An Agilent 1100 LC-MS system (Agilent, Santa Clara, CA, USA) using the EI technique was employed for MS analysis. Rf × 100 values were obtained with petroleum ether/ethyl acetate (50:50 v/v) mobile phase at 21 • C. An Agilent 1100 series instrument equipped with a G1315A DAD detector, a pump (1311A Quat pump) and manual sampler was used for the identification of Rt (retention time) values; using a Zorbax SB C8 5 µm, 250 × 4.6 mm column. Operating conditions were performed by an isocratic HPLC grade acetonitrile/phosphate buffer (pH:3.7) (70:30 v/v) mobile phase at a flow rate of 1 mL/min. ChemStation A-08.03 (847)2000 software program with HP Compaq d330 DT computer was used to analyze the spectral data.

General Procedure for the Synthesis of 2-[(2 ,4 -Difluoro-4-hydroxybiphenyl-3-yl)carbonyl]-N-(substituted)hydrazinocarbothioamides 3-10
Diflunisal hydrazide (2, 0.01 mol) was dissolved in ethanol (25 mL) and an equimolar amount of a substituted isothiocyanate was added to the reaction mixture which was refluxed for 2 to 6 h. The crude product was precipitated, filtered and washed with distilled water. The compounds were recrystallized from ethanol or DMF [12]. Compounds 3, 4 and 5 were previously reported in our study [25]. Compounds 3-9 except for compound 8 and 10 (0.01 mol) were dissolved in sodium hydroxide solution (2 N, 25 mL) and heated under reflux for 4 to 6 h. The solution then was neutralized with concentrated hydrochloric acid. The crude product was precipitated, filtered and washed with distilled water. The compounds recrystallized twice from ethanol or DMF [12].  . RPM 1640 medium was further completed with 0.2 unit/mL bovine insulin for the T47D cell line. Cells were passaged when they reached %80 confluency in order to support proliferation.

Cytotoxicity Assay and Determination of IC 50 Values of Synthesized Compounds
The effects of the synthesized compounds on the cell viability were tested. Stock solutions for each compound were prepared by measuring 20 mg of each compound and dissolving it in 1 mL DMSO to a final concentration of 20 mg/10 mL, followed by filtration through a 0.2-µm filter (Sartorius AG, Gottingen, Germany). Then stock solutions were diluted as 1/10, 1/100 and 1/1000 in the designated media and applied on the cell lines. In order to determine the effect of DMSO itself, the same dilutions were performed by using DMSO, only and applied on the cell lines. Cell densities used for each cell line in 96 well plate were as follows: for PC-3, HCT116 and HEK293 cell lines 5000 cells/well; for T47D cell lines 20,000 cells/well. Cell viability was measured by the MTS cytotoxicity assay (CellTiter96 AqueousOne Solution; Promega, Southampton, UK) according to the manufacturer's instructions. 3-(4,5-Dimethylthiazol-2-yl)-5-(3-carboxymethoxyphenyl)-2-(4-sulfo-phenyl)-2H-tetrazolium (MTS) is a yellow tetrazolium salt which is catabolized to formazan by dehydrogenase enzyme in the mitochondria of the living cells. Formazan is a purple compound and the detection in this assay is based on the measurement of formazan compounds by an ELISA plate reader. After incubating the cells for 72 h, 10 µL MTS reagent and 100 µL designated medium was given to the cells followed by 2 h of incubation at 37 • C. Thereafter, the absorbance at 490 nm was measured by an ELISA plate reader. (BioTek Instruments, Inc., Winooski, VT, USA). The activity doses of the compounds were determined according to 50% cell death rate when compared to DMSO, only after 72 h of incubation. Compounds that do not affect cell viability of healthy cell line (HEK293) but deathly to cancer cell lines were further tested with narrower test conditions for MTS assay to determine IC 50 values. IC 50 values are calculated by the GraphPad prism software.

Annexin V
Determination of IC 50 values was followed by the investigation of effect of the synthesized compounds on apoptosis. Cells were seeded into six well plates at densities of 15,000 cells/well (PC-3, HCT116 and HEK293 cell lines) and 60,000 cells/well (T47D cell lines). After the treatment cells were harvested and washed with ice cold PBS. 1 × Annexin-binding buffer was prepared while the cells were being washed. Working solution of PI (100 µg/mL) was prepared as well by diluting 5 µL of the 1 mg/mL PI stock solution in 45 µL 1 × Annexin-binding buffer. After the precipitation of the cells, they were resuspended in 500 µL 1 × Annexin-binding buffer. Resuspended cells were separated into four groups (Annexin V, PI, Annexin V + PI and NC) each containing 100 µL of suspension. Five µL of the Annexin V conjugate was added to the proper groups. Samples were incubated for 15 min at room temperature. The incubation step was followed by adding 2 µL of the 100 µg/mL PI working solution to the proper groups then the fluorescence was observed by flow cytometry.

5-Ethynyl-2 -deoxyuridine (EdU) Staining
The EdU staining protocol directly measures the level of active DNA synthesis of the cell cycle. This measurement enables one to see dividing cells under a fluorescence microscope. In order to perform EdU staining cells were seeded on the coverslips which were located in the wells of six well plates. As in the Annexin V procedure cells were seeded into six well plates at densities of 15,000 cells/well (PC-3, HCT116 and HEK293 cell lines) and 60,000 cells/well (T47D cell lines). After drug treatment, cells that are on the coverslips were incubated with 10 µM EdU. The incubation step is followed by cell fixation and permeabilization. Fixation occurred with 3.7% formaldehyde in PBS for 15 min then cells were washed with 3% BSA in PBS. Permeabilization was performed by incubating cells in 0.5% Triton ® X-100 in PBS for 20 min at room temperature. Click-iT ® reaction cocktail (0.5 mL) was added to the cells after washing with 3% BSA solution. Cells were incubated for 30 min at room temperature, then DNA synthesis was observed under a confocal microscope.

Docking Studies
The preparation of the protein, as well as the compounds were performed in BIOVIA Discovery Studio 4.5. The crystal structures of the enzymes COX-1 and COX-2 were downloaded from the PDB database (accession code: 1EQG for COX-1 [46] and 3NT1 for COX-2 [47] (http://www.rcsb.org/pdb/). All protein structures were cleaned of inhibitors, water molecules and non-interacting ions, while cofactors were retained. All hydrogens were added and proteins minimized using the "Clean Geometry" toolkit. For more complete optimization, proteins were submitted to the "Prepare Macromolecule" protocol. Missing hydrogen atoms were added based on the protonation state of the titratable residues at pH 7.4. Compounds were prepared and minimized in the same force field. After preparation, docking was performed in AutoDock 4.2 program (http://autodock.scripps.edu). All the compounds were set to be flexible, while the proteins were kept rigid. AutoDock adds a free-energy scoring function created from a linear regression analysis, the AMBER force field, and a large set of diverse protein-ligand complexes with known inhibition constants. The coordinates of alpha carbon of Tyr355 were used for grid-centering for COX enzymes, grid box was set to be 70 grid points (each grid point is 0.375 Å). For the docking studies the Lamarkian genetic algorithm was used as search algorithm. For visualization of non-bonded interactions, we also used BIOVIA Discovery Studio 4.5.

Conclusions
Diflunisal and its derivatives have been widely studied and different biological activities have been reported for these compounds. The correlation between the COX-2 selectivity and anticancer activity, had become the basic subject of our research. Thiosemicarbazides and 1,2,4-triazole-3-thiones have been related with anticancer activity, therefore, we have designed some novel compounds combining these active pharmacophores and a NSAID.
In this study, we have synthesized new diflunisal thiosemicarbazides and 1,2,4-triazole-3-thiones in order to determine their possible anticancer activity. The synthesis, structure characterization and antibacterial activity against Helicobacter pylori of thiosemicarbazides, which are the prototypes of the 1,2,4-triazoles 11-13, are given previously study [25]. Compounds 6 and 10 were found active against the PC-3 cancer cell line, with IC 50 values of 41.8 and 11.7 µM, respectively. Compound 16 was found active against the HCT-116 cancer cell line with a 6.2 µM IC 50 value. Compounds 15 and 16 were found active against the T47D cancer cell line with IC 50 values of 43.4 and 27.3 µM, respectively. The in silico and in vitro studies that we performed lead us to the conclusion that there is a correlation between COX-2 enzyme activity and anticancer activity in vitro. In the docking studies, compounds 12 and 13 appeared to show the most affinity towards COX-2 enzyme however, compounds 15 (∆G = −10.11 kcal/mol) and 16 (∆G = −9.41 kcal/mol) showed high inhibition on COX-2 enzyme. The highest selectivity towards COX-2 was shown by compound 13. Further experiments are still needed in order to improve our understanding of how these compounds elicit their anticancer activity.