Advertisement

SN Applied Sciences

, 1:332 | Cite as

Carprofen: a theoretical mechanistic study to investigate the impact of hydrophobic interactions of alkyl groups on modulation of COX-1/2 binding selectivity

  • Ahmed M. GoudaEmail author
  • Faisal A. Almalki
Research Article
  • 55 Downloads
Part of the following topical collections:
  1. 1. Chemistry (general)

Abstract

Development of selective COX-1 and COX-2 was successfully used to overcome GIT side effects of the classical NSAIDs. Currently, virtual screening and docking study were extensively used to design of new potent and safe drug candidates. In this study, four series of carprofen derivatives were designed by isosteric replacement of the –NH– with –O–, –S– and –CH2– groups. More than 90 derivatives bearing different alkyl substituents were designed in this study. AutoDock software was used to explore the binding mode, affinity and selectivity of the designed analogs to COX1/2. The results revealed that position and length of alkyl substituents have remarkable effect on the binding mode. Substitution with alkyl groups at C1 and C6 of the four scaffolds improved binding to COX-1, while at C4 enhanced COX-2 binding affinity. Compound 66 displayed the highest binding affinity for COX-1 and COX-2 with ΔGb of 11.2 and 10.15 kcal/mol, respectively. Compound 56 and 99 displayed the highest potential selectivity to COX-1 and COX-2, respectively. Drug-likeness study and synthetic accessibility were evaluated for the most promising analogs. Compound 29 displayed drug-likeness score (DLS) of 0.96 compared to 0.3 for carprofen. Taken together, these results highlighted the importance of hydrophobic interactions in modulation of COX-1/2 binding selectivity of new potential NSAIDs.

Graphical abstract

Keywords

Carprofen COX-1/2 Anti-inflammatory Docking Hydrophobic interaction Selectivity Drug-likeness 

1 Introduction

Development of selective cyclooxygenase-2 (COX-2) inhibitors was suggested as an attractive strategy to overcome GIT side effects of traditional NSAIDs [1, 2, 3, 4, 5]. Diverse scaffolds with potential COX inhibitory activity were reported with the great advances in heterocyclic synthesis [6, 7]. Valdecoxib 1 (Fig. 1) displayed high selectivity for COX-2 with selectivity index higher than 3930 [8]. Although coxibs displayed safe GIT profile, but their use was associated with cardiovascular and renal side effects [9]. However, selective COX-2 inhibitors still attract the attention due to their antiproliferative activity [10, 11, 12].
Fig. 1

Rofecoxib, mofezolac and carprofen and their IC50 against COXs

On the other hand, mofezolac 2 (Fig. 1) was recently developed as selective COX-1 inhibitor with analgesic anti-inflammatory potential and low ulcerogenic liability [13]. Mofezolac has also displayed promising anticancer activity in intestinal cancer [14, 15].

Nonselective NSAIDs such as profens still represent one of the most widely used drugs. Of these NSAIDs, carprofen was used in humans for more than 10 years before withdrawal from the market on commercial grounds. Later, it was used in animals to treat pain and inflammation. The incidence of GIT side effects with carprofen resembles those of other nonselective COX inhibitors [16, 17].

Mechanistic studies of carprofen were reported in several reports. In whole blood assay, carprofen displayed higher selectivity for COX-1 over COX-2 [18]. However, in vitro evaluation of COX inhibitory activity revealed selective inhibition of COX-2 [19, 20]. Moreover, carprofen displayed fatty acid amide hydrolase (FAAH) inhibitory activity which could contribute to its analgesic and anti-inflammatory activity [20, 21].

1.1 Rational design

In the last few years, virtual screening and molecular docking were used successfully in the identification of many potent and selective COX inhibitor [22, 23]. Extensive work in this field have focused on polar substituents to modulate COX selectivity. On the other hand, hydrophobic interactions of the alkyl group gained little attention in this field which may be attributed to weakness the hydrophobic interactions as compared by ionic and hydrogen bonding.

However, the hydrophobic interactions (Fig. 2) of the aliphatic chain in arachidonic acid play an important role in the binding of this natural substrate to COX-1/2 [24]. Accordingly, it was of interest to design new carprofen analogs bearing different aliphatic substitutions to evaluate their impact on the binding affinity to COX-1/2.
Fig. 2

Arachidonic acid into the active site of COX-1 (pdb code: 1diy) showing hydrophobic interactions with PHE209 and PHE381 amino acids

In this study, three scaffolds were generated from S-carprofen by isosteric replacement of the -NH in carprofen with –S–, –O– and –CH2– groups. The four scaffolds were derivatized by different alkyl/halo substituents (Fig. 3). The impact of these alkyl substitutions on binding affinity to COX-1/2 was evaluated in a docking study. Moreover, the effect of variation of the position of the propionic acid side chain on binding affinity to COX-1/2 was also investigated.
Fig. 3

Structural modification of carprofen scaffold

2 Experimental

2.1 Docking study

In this work, the molecular docking study was performed to evaluate affinity and selectivity of the newly designed carprofen analogs to COX-1/2. The binding modes of the analogs with the promising affinity were also investigated.

2.1.1 Preparation of ligands and the protein files

AutoDock 4.2 was used to perform the docking study of the designed carprofen analogs into COX-1 (PDB code: 1EQG) [25] and COX-2 (PDB code: 1CX2) [26]. The crystal structure of the two enzymes were obtained from protein data bank (http://www.rcsb.org/pdb). The preparation of the ligands was done according to our previous reports [4, 12]. The study was performed after validation of the docking scenario by re-docking the native ligands into their corresponding COX enzymes. The binding modes and interactions of the native ligands with the key amino acids in the active site were identified and compared with the reported data. The re-docked ligands superimposed onto the position of the native ligand.

2.1.2 Preparation of grid and docking parameters

AutoGrid tool in AutoDock was used to create the grid parameter files. 3D grid with final size of 60 × 60 × 60 Å with 0.375 Å spacing was created. Genetic algorithm was used as the searching parameters in the docking study while docking parameters were set to the default values. The ligands were docked as flexible molecules and the proteins were used as rigid molecules. The best in the top ten conformations with the highest best binding-free energy values were determined.

2.1.3 Analysis and visualization of the docking results

The results of the docking study including binding-free energy (ΔGb), and inhibition constants (Ki) were analyzed by AutoDock 4.2 while the binding modes were visualized by Discovery Studio Visualizer (v16.1.0.15350).

2.2 Drug-likeness, ADME and toxicity study

The molecular weight (MW), molar refractivity (MR), lipophilicity (MlogP), number of hydrogen bond donor (HD), number of hydrogen bond acceptor (HA), Lipinski’s violations, number of rotatable bonds and topological polar surface area (TPSA) of the new compounds and carprofen 3 were calculated using SwissADME webserver (http://www.swissadme.ch/) [27]. Marvin JS sketcher (version 16.4.18, 2016, www.chemaxon.com) was used to draw the chemical structures of the designed analogs, converted to smiles by JChem web service which undergo a series of calculations to compute different physicochemical descriptors related to drug-likeness. The results were obtained as Excel output file. The molecular volume and drug-likeness score (DLS) were calculated using Molsoft webserver (http://molsoft.com/mprop/). The results were presented in Table 7.

3 Results and discussion

3.1 Docking study

3.1.1 Docking study of S-carprofen into COX-1/2

The S-Carprofen was docked into the active site of COX-1/2 using AutoDock 4.2. Both ovine COX-1 (pdb code: 1EQG) [25] and COX-2 (pdb code: 1CX2) [26] were used in the docking study. A rigid docking scenario of the ligands into COXs was applied. Validation of the docking study was done by redocking the native ligands into their corresponding COX. The results of the study including binding affinity, inhibition constants and hydrogen bonds were represented in Table 1.
Table 1

Results of the docking study of S-carprofen into COX-1 (pdb: 1EQG) [25] and COX-2 (pdb code: 1CX2) [26] in comparison to the redocked native ligands (ibuprofen and SC-588)

COX l/Comp.

ΔG b a

K i b (nM)

No of H-bonds

Atoms in H-bonding

Lengthc (Å)

In the ligand

In the enzyme

COX-1

      

 (S)-Carprofen

− 10.02

45.36

3

OH

C=O

C=O

NH of ARG120

OH of TYR355 NH2 of ARG120

1.72

1.92

2.01

 (S)-Ibuprofen

− 7.96

1470

3

OH

C=O

C=O

NH of ARG120

NH2 of ARG120

OH of Tyr355

1.65

1.78

1.83

COX-2

      

 (S)-Carprofen

− 8.66

450.2

2

C=O

OH

OH of TYR355

δ-C of ARG120

2.01

2.17

 Sc-588

− 10.78

12.59

4

O of SO2

H1 of NH2

H2 of NH2

CF3

NH of HIS90

CO of LEU352

CO of GLN192

NH of ARG120

2.03

2.08

1.97

2.24

aBinding free energy (kcal/mol)

bInhibition constant (nM)

cLength of in angstrom (Å)

The S-carprofen displayed binding-free energy of 10.02 kcal/mol and inhibition constant of 45.36 nM with COX-1 indicating higher affinity for the enzyme than ibuprofen. Three hydrogen bonds were formed between the carboxylic group oxygen in carprofen with ARG120 and TYR355 in COX-1 with bond length (BL) in the range of 1.72–2.01 Å (Fig. 4).
Fig. 4

Binging modes of S-carprofen 3: a 3D binding mode into COX-1 (pdb code: 1EQG); b 2D binding mode into COX-1, c 3D binding mode into COX-2 (pdb code: 1CX2); d 2D binding mode into COX-2 showing hydrogen bonds ( Open image in new window ), carbon hydrogen bonds ( Open image in new window ), hydrophobic interactions of the amide-pi stacked ( Open image in new window ), pi-sulfur ( Open image in new window ), pi-sigma ( Open image in new window ), pi–pi T-shaped ( Open image in new window ), pi-alkyl ( Open image in new window ) and alkyl ( Open image in new window ) types

S-Carprofen showed 16 hydrophobic interactions of the amide-pi stacked, pi-sigma, pi-sulfur, alkyl, and pi-alkyl types (BL = 3.27–5.35 Å) with VAL116, VAL349, LEU359, PHE381, LEU384, TYR385, TRP387, ILE523, GLY526, ALA527, and LEU531 amino acids in COX-1. In addition, one pi-sulfur interaction was formed with MET522 (BL = 5.94 Å) (Fig. 4).

On the other hand, the S-carprofen displayed binding free energy of 8.66 kcal/mol and inhibition constant of 450.2 nM with COX-2. One conventional hydrogen bond was observed between carbonyl oxygen in carprofen with the hydroxyl group in TYR355 with BL of 2.01 Å (Table 1). Moreover, an additional carbon hydrogen bond was observed between oxygen in OH group of carprofen and the hydrogen of the δ-carbon in ARG120 (BL = 2.17 Å).

S-carprofen formed fifteen hydrophobic interactions of the pi-sigma, alkyl and pi-alkyl types with VAL349, LEU352, LEU359, PHE381, LEU384, TYR385, TRP387, VAL523, ALA527, and LEU531 amino acids in COX-2 (BL = 3.70–5.49 Å) (Fig. 4).

Generally, the S-carprofen 3 displayed high binding-free energy to COX-1 (ΔGb= 10.02 kcal/mol) over COX-2 (ΔGb= 8.66 kcal/mol). These results indicate that S-carprofen 3 has higher potential selectivity for COX-1 than COX-2.

The relationship between the half maximal inhibitory concentration (IC50) and inhibition constant (Ki) was described in several reports [28, 29]. Both IC50 and Ki are used to express the relative potency of an inhibitor. The smaller the value of IC50 and Ki, the stronger the inhibitory activity. The relationship between IC50 and Ki, of a competitive inhibitor was described by Cheng– Prusoff according to the Eq. 1 where [S] is the substrate concentration and KM is Michaelis constant of the substrate [28].
$${\text{IC}}_{50} = K_{i} \left( {1 + \frac{{\left[ {\text{S}} \right]}}{{K_{M} }}} \right)$$
(1)

In this study, the binding free energy against the two COXs and Ki ratio (ratio of Ki of COX-2 to Ki of COX-1) of the designed compounds will be used to compare their affinities and potential selectivity. S-carprofen displayed Ki ratio of 9.9 which indicate that carprofen has 9.9 times higher potential selectivity to inhibit COX-1 over COX-2. This result is matched with the results of carprofen inhibitory activity of COX in the whole blood assay [18].

3.1.2 Docking Study of carprofen analogs into COX-1/2

3.1.2.1 (S)-2-(9H-carbazol-2-yl)propanoic acid derivatives
To investigate the effect of the position of the carboxylic acid side chain on the binding affinity of S-carprofen to COXs, three analogs 46 were designed (Fig. 5). The results of the docking study revealed a decrease in binding affinity of the three analogs to COX-1, as compared to S-carprofen. Moreover, compound 5 and 6 showed a slight and marked decrease in their binding affinity to COX-2, respectively. On the other hand, compound 4 showed slightly higher affinity for COX-2 as compared to S-carprofen (Table 2).
Fig. 5

Chemical structures of (S)-enantiomers of compounds 3 and 512

Table 2

Results of the docking study of compounds 326 into the COX-1/2

Comp.

ΔG b a

K i b

Ki ratioc

Comp.

ΔG b a

K i b

K i ratio c

COX-1

COX-2

COX-1

COX-2

COX-1

COX-2

COX-1

COX-2

3

− 10.02

− 8.66

45.36

450.2

9.9

15

− 10.29

− 8.59

28.69

501.29

17.47

4

− 9.95

− 8.77

50.77

374.47

7.38

16

− 10.39

− 8.86

24.36

321.07

13.18

5

− 9.96

− 8.53

50.33

559.54

11.12

17

− 10.53

− 9.33

18.97

144.99

7.64

6

− 8.66

− 7.70

448.8

2280

5.08

18

− 10.78

− 9.44

12.64

119.64

9.47

7

− 9.54

− 8.11

102.16

1140

11.16

19

− 9.34

− 10.07

142.48

41.29

0.29

8

− 10.06

− 8.84

42.17

332.19

7.88

20

− 9.75

− 8.51

71.23

583.06

8.19

9

− 9.97

− 8.93

48.9

286.82

5.87

21

− 10.11

− 8.50

38.9

592.21

15.22

10

− 9.27

− 8.12

161.19

1120

6.95

22

− 10.17

− 8.54

35.17

545.35

15.51

11

− 10.13

− 8.55

37.72

542.19

14.37

23

− 10.33

− 8.43

26.57

666.14

25.07

12

− 10.00

− 8.61

47.16

487.71

10.34

24

− 10.46

− 8.85

21.5

324.56

15.09

13

− 9.79

− 8.55

66.87

537.33

8.04

25

− 10.64

− 9.24

15.81

167.93

10.6

14

− 10.18

− 8.55

34.52

539.6

15.63

26

− 9.87

− 9.80

58.15

65.82

1.13

aBinding free energy (kcal/mol)

bInhibition constant (nM)

cThe ratio of Ki of COX-2 to Ki of COX-1

The values of the Ki ratio showed that compound 5 has higher potential selectivity to COX-1 than S-carprofen (Table 2). Based on these results, the position of propionic acid chain is important for directing the binding selectivity for COX-1 or COX-2. Accordingly, compound 5 was selected for further investigation as a potential scaffold with high binding selectivity to COX-1.

Investigation of the impact of variation of the 6-chloro group in S-carprofen with other halogens (F, Br or I) on binding affinity was investigated. A new series of carbazole derivatives 712 was designed bearing different halogens at C6 was designed (Fig. 5). Replacement of the 6-chloro in S-carprofen with fluoro, bromo, and iodo resulted in a marked decreased selectivity ratio to COX-1. The 6-iodo analog 9 displayed the highest binding affinity for COX-2 (Table 2).

It was clear that the binding affinity to COX-2 increases as the atomic size increases. These results are matched with the fact that the active site of COX-2 (394Å) is larger than that of COX-1 (316Å) [30].

The carbazole derivatives 1326 was modelled after compound 5 to investigate the effect of 6-alkyl/1-fluoro substitutions on the binding affinity to COX-1/2 (Fig. 5). Among the designed analogs 1326, compounds 15 with 6-propyl substituent displayed highest binding affinity for COX-1 while compound 19 showed the highest affinity for COX-2.

On comparing the Ki ratio of compounds 1319 with their corresponding fluorinated analogs 2026, a noticeable increase in the Ki ratio indicating higher potential of COX-1 selectivity. Among the carbazole derivative 326, compound 23 with the 6-butyl group displayed the highest potential selectivity for COX-1 (Ki ratio = 25.07) (Table 2).

Different types of interaction between compound 23 and amino acids in the active site of COX-1 were represented in Fig. 6. Compound 23 formed 2 classical hydrogen bonds with ARG120 in COX-1 with BL of 1.64 and 1.96 Å. A third carbon hydrogen bond was observed between the 1-fluoro with α-CH of SER523 (BL = 1.84 Å). Additionally, compound 23 displayed 16 hydrophobic interactions of the pi-sigma, pi-alkyl and alkyl types with amino acids in the active site of COX-1 (BL = 3.69–5.38 Å) (Fig. 6).
Fig. 6

Binging modes of compound 23 into COX-1 (pdb code: 1EQG): a 3D binding mode, receptor surface hydrophobicity visualized; b 2D binding mode showing hydrogen bonds ( Open image in new window ), carbon hydrogen bonds ( Open image in new window ) and hydrophobic interactions of the pi-sigma ( Open image in new window ), pi-alkyl ( Open image in new window ) and alkyl ( Open image in new window ) types

3.1.2.2 (S)-2-(9H-fluoren-3-yl)propanoic acid derivatives
To design new analogs with enhanced COX-1 selectivity, a new series of fluorene derivatives 2742 were obtained by isosteric replacement of NH group in compound 5 with –CH2– group (Fig. 7). The effect of alkyl/fluoro substitutions groups on the binding affinity to COXs was investigated.
Fig. 7

Chemical structures of (S)-enantiomers of compounds 2742

Among the alkyl derivatives 2742, compound 34, 40 and 42 displayed the highest binding affinity (ΔGb= -10.82) for COX-1, while compound 33 showed the highest affinity (ΔGb= -10.84) for COX-2 (Table 3).
Table 3

Results of the docking study of compounds 2742 into the COX-1/2

Comp.

ΔG b a

K i b

K i ratio c

Comp.

ΔG b a

K i b

K i ratio c

COX-1

COX-2

COX-1

COX-2

COX-1

COX-2

COX-1

COX-2

27

− 10.16

− 8.68

36.0

431.94

12

35

− 10.09

− 8.74

39.9

394.87

9.9

28

− 10.01

− 8.66

45.77

451.81

9.87

36

− 9.99

− 8.66

47.44

449.65

9.48

29

− 10.42

− 8.85

22.94

325.14

14.17

37

− 10.32

− 8.90

27.04

297.22

10.99

30

− 10.47

− 8.98

21.04

263.12

12.51

38

− 10.36

− 8.91

25.43

294.94

11.59

31

− 10.62

− 9.19

16.49

184.35

11.18

39

− 10.54

− 8.91

18.68

293.27

15.69

32

− 10.76

− 9.65

12.89

97.89

7.59

40

− 10.82

− 9.05

11.78

234.03

19.87

33

− 10.63

− 10.84

16.19

11.29

0.69

41

− 10.75

− 9.30

13.1

151.96

11.6

34

− 10.82

− 9.84

11.67

61.11

5.24

42

− 10.82

− 9.27

11.8

159.01

13.47

aBinding free energy (kcal/mol)

bInhibition constant (nM)

cThe ratio of Ki of COX-2 to Ki of COX-1

The Ki ratio of compound 2742 was presented in Table 3. Except for compound 33 which showed weak potential selectivity for COX-2 (Ki ratio = 1.43 for COX-2), all other compounds displayed higher binding affinity for COX-1 over COX-2 indicating potential selectivity (Ki ratio = 7.59–19.87).

In comparison to S-carprofen which formed 3 hydrogen bonds with COX-1, compound 40 formed two hydrogen bonds only with ARG120, with BL of 1.68 and 1.89 Å (Fig. 8). Moreover, LigPlot view showed a third hydrogen bonds between compound 40 and TYR355 with BL of 2.99 Å (Fig. 9). Compound 40 formed also 16 hydrophobic interactions with COX-1 with BL in the range of 3.52–5.42 Å. Five of these hydrophobic interactions were due to the pentyl group.
Fig. 8

Binging modes of compound 40 into COX-1 (pdb code: 1EQG): a 3D binding mode, receptor surface hydrophobicity visualized; b 2D binding mode showing hydrogen bonds ( Open image in new window ), and hydrophobic interactions of the pi-sigma ( Open image in new window ), pi-alkyl ( Open image in new window ) and alkyl ( Open image in new window ) types

Fig. 9

a LigPlot view of compound 40 into the active site of COX-1 showing three hydrogen bonds with ARG120 and TYR355 with bond length in the range of 2.68–2.99 Å; b LigPlot view of compound 56 into the active site of COX-1 showing three hydrogen bonds with ARG120 and TYR355 with bond length in the range of 2.78–2.89 Å

3.1.2.3 (S)-2-(dibenzo[b,d]furan-3-yl)propanoic acid derivatives
Replacement of the NH group in compound 5 with the isosteric oxygen yielded 2-(dibenzo[b,d]furan-2-yl)propanoic acid 43. The modification was done to design new analogs with high binding selectivity to COX-1. Substitution with alkyl/fluoro groups was done and the designed analogs were evaluated for their binding affinity for COX-1/2 (Fig. 10).
Fig. 10

Chemical structures of the (S)-enantiomers of compounds 4358

Among compounds 4358, compound 50 with the heptyl chain displayed the highest binding affinity for COX-1 (ΔGb= -11.13) and COX-2 (ΔGb= -10.0) (Table 4).
Table 4

Results of the docking study of compounds 4358 into the COX-1/2

Comp.

ΔG b a

K i b

K i ratio c

Comp.

ΔG b a

K i b

K i ratio c

COX-1

COX-2

COX-1

COX-2

COX-1

COX-2

COX-1

COX-2

43

− 10.14

− 8.77

37.12

375.24

10.11

51

− 10.07

− 8.73

41.23

400.25

9.71

44

− 9.98

− 8.78

48.23

367.73

7.63

52

− 9.96

− 8.73

50.34

400.08

7.95

45

− 10.38

− 8.70

24.69

416.23

16.86

53

− 10.31

− 8.69

27.85

424.52

15.24

46

− 10.48

− 8.65

20.66

459.4

22.24

54

− 10.42

− 8.33

23.17

781.08

33.71

47

− 10.56

− 9.02

18.23

242.65

13.31

55

− 10.45

− 8.93

21.95

282.7

12.87

48

− 10.82

− 9.51

11.72

107.15

9.14

56

− 10.70

− 8.47

14.44

622.28

43.09

49

− 11.01

− 9.72

8.44

75.17

8.91

57

− 10.8

− 9.17

12.09

190.19

15.73

50

− 11.13

− 10.0

6.98

46.39

6.65

58

− 11.12

− 9.20

7.04

179.4

25.48

aBinding free energy (kcal/mol)

bInhibition constant (nM)

cThe ratio of Ki of COX-2 to Ki of COX-1

The values of the Ki ratio of compound 5158 showed the fluorinated analogs have relatively higher Ki ratio than their corresponding derivatives 4350. Compound 56 with the pentyl side chain has the highest potential selectivity for COX-1 (Ki ratio = 43.09) (Table 4).

Compound 56 with the pentyl side chain displayed the highest potential selectivity for COX-1 (Ki ratio = 43.09) (Table 4). Compound 56 formed two hydrogen bonds only with ARG120 (classical) and SER353 (nonclassical) amino acids in COX-1, with BL of 1.90 and 1.89 Å, respectively (Fig. 11). LigPlot view of compound 56 showed 3 hydrogen bonds with ARG120 and TYR355 with BL in the range of 2.78–2.89 Å (Fig. 9). Moreover, compound 56 formed also 14 hydrophobic interactions with amino acids in the active site of COX-1 with bond distances in the range of 3.67–5.32 Å (Fig. 11).
Fig. 11

Binging modes of compound 56 into COX-1 (pdb code: 1EQG): a 3D binding mode; b 2D binding mode, showing hydrogen bonds ( Open image in new window ), carbon hydrogen bonds ( Open image in new window ), hydrophobic interactions of the pi-sigma ( Open image in new window ), pi-alkyl ( Open image in new window ) and alkyl ( Open image in new window ) types

3.1.2.4 (S)-2-(dibenzo[b,d]thiophen-3-yl)propanoic acid derivatives
The dibenzo[b,d]thiophene scaffold was obtained from compound 5 by isosteric replacement of NH with -S- group. Derivatization was obtained by substitution with fluoro/alkyl to evaluate their effect on binding affinity to COXs (Fig. 12).
Fig. 12

Design strategy and chemical structural of the (S)-enantiomers of compounds 5978

The results of the docking study of the dibenzo[b,d]thiophene analogs 5978 into COX-1/2 revealed that compound 66 has the highest binding affinity to both COX-1/2 (Table 5).
Table 5

Results of the docking study of compounds 5978 into the COX-1/2

Comp.

ΔG b a

K i b

K i ratio c

Comp.

ΔG b a

K i b

K i ratio c

COX-1

COX-2

COX-1

COX-2

COX-1

COX-2

COX-1

COX-2

59

− 9.93

− 8.69

52.48

423.24

8.07

69

− 10.08

− 8.87

40.7

314.3

7.72

60

− 9.88

− 8.65

56.86

460.17

8.09

70

− 10.19

− 9.08

33.99

220.57

6.49

61

− 10.30

− 8.91

28.14

296.63

10.54

71

− 10.36

− 9.03

25.36

238.65

9.41

62

− 10.34

− 9.14

26.42

199.76

7.56

72

− 10.67

− 9.28

15.04

157.68

10.48

63

− 10.40

− 9.36

23.87

137.44

5.76

73

− 10.45

− 9.52

21.75

105.42

4.85

64

− 10.73

− 9.67

13.67

81.7

5.98

74

− 10.23

− 9.96

31.62

50.29

1.59

65

− 10.53

− 9.91

19.04

54.8

2.88

75

− 9.68

− 8.30

80.31

829.27

10.33

66

− 11.2

− 10.15

6.15

36.5

5.94

76

− 9.84

− 8.67

61.15

442.66

7.2

67

− 9.74

− 8.67

72.95

440.98

6.04

77

− 10.0

− 8.64

46.83

465.09

9.93

68

− 9.64

− 8.65

85.86

459.2

5.35

78

− 10.31

− 8.83

27.55

338.68

12.3

aBinding free energy (kcal/mol)

bInhibition constant (nM)

cThe ratio of Ki of COX-2 to Ki of COX-1

Like S-carprofen, the dibenzo[b,d]thiophene 5978 displayed higher binding affinities for COX-1 over COX-2 indicating potential selectivity for COX-1 (Table 5).

The second series of the thiophen analogs 79100 were designed bearing alkyl groups at C3, C4 and C8 (Fig. 13). Compounds 79100 were docked into the active site of COX-1/2. The results of the study revealed sharp decrease in the binding affinity to COX-1 with a simultaneous increase in the affinity to COX-2. Accordingly, the potential selectivity to COX-1 was reversed with most of the derivatives in this series (Table 6).
Fig. 13

Design strategy and chemical structures of S-enantiomers of compounds 7982

Table 6

Results of the docking study of compounds 79100 into the COX-1/2

Comp.

ΔG b a

K i b

K i ratio c

Comp.

ΔG b a

K i b

K i ratio c

COX-1

COX-2

COX-1

COX-2

COX-1

COX-2

COX-1

COX-2

79

− 8.08

− 8.93

1200

286.12

0.238

90

− 7.41

− 9.21

3710

178.02

0.048

80

− 9.46

− 8.86

116.09

319.77

2.76

91

− 7.53

− 9.51

3000

106.99

0.036

81

− 8.35

− 9.16

758.23

193.89

0.226

92

− 7.92

− 9.96

1560

50.42

0.032

82

− 8.28

− 9.40

857.81

128.03

0.149

93

− 10.05

− 10.06

42.76

42.21

0.987

83

− 9.58

− 9.20

95.48

179.28

1.88

94

− 10.21

− 9.47

32.73

114.99

3.51

84

− 9.10

− 9.37

213.72

135.77

0.635

95

− 7.75

− 9.38

2100

133.32

0.063

85

− 8.98

− 9.58

260.01

94.62

0.364

96

− 7.69

− 9.6

2320

92.2

0.04

86

− 8.49

− 9.87

601.9

58.73

0.098

97

− 8.07

− 9.84

1220

61.74

0.051

87

− 9.74

− 9.81

72.16

64.39

0.892

98

− 8.56

− 10.11

533.67

38.61

0.072

88

− 10.24

− 9.08

31.11

221.44

7.12

99

− 8.14

− 10.33

1070

26.82

0.025

89

− 9.38

− 9.29

132.1

154.57

1.17

100

− 9.39

− 9.95

130.04

51.11

0.393

aBinding free energy (kcal/mol)

bInhibition constant (nM)

cThe ratio of Ki of COX-2 to Ki of COX-1

Among the tested analogs 79100, compound 88 displayed the highest binding affinity to COX-1 while compound 99 displayed the highest affinity to COX-2 (Table 6). The values of the Ki ratio of compound 79100 showed that only compound 80, 83, 88, 89 and 94 have potential selectivity to COX-1 with lower potential selectivity compared to S-carprofen. On the other hand, the remaining analogs showed potential selectivity to COX-2 where compound 92 and 99 showed the lowest ratio with Ki ratio of 0.032 and 0.025, respectively. The binding mode of these two compounds were presented in Table 6.

Compound 92 displayed higher binding affinity (ΔGb = 9.96 kcal/mol) for COX-2 than S-carprofen (ΔGb = 8.66 kcal/mol). Although compound 92 and S-carprofen formed 2 hydrogen bonds with COX-2, but compound 92 displayed higher number of hydrophobic interactions (18) compared to 15 only for S-carprofen (Fig. 14).
Fig. 14

Binging modes of compound 92 into COX-2 (pdb code: 1CX2): a 3D binding mode; b 2D binding mode, showing hydrogen bonds ( Open image in new window ), carbon hydrogen bonds ( Open image in new window ), sulfur-X ( Open image in new window ), hydrophobic interactions of the pi-sigma ( Open image in new window ), pi-alkyl ( Open image in new window ) and alkyl ( Open image in new window ) types

Moreover, compound 92 showed sharp decrease in the binding affinity to COX-1 compared to S-carprofen. As a result, compound 92 displayed high potential selectivity for COX-2 over COX-1 (Ki ratio = 0.032).

Compound 99 with the isopentyl side chain displayed binding affinity for COX-2 (ΔGb = 10.33 kcal/mol) higher than S-carprofen, with the highest potential selectivity for COX-2 (Ki ratio = 0.025) (Table 6). The high affinity and potential selectivity to COX-2 of compound 99 is attributed to the formation of 3 hydrogen bonds with HIS90, LEU352 and ARG513 amino acids in COX-2 (BL = 2.06–3.07 Å) compared to 2 hydrogen bonds only for S-carprofen (Fig. 15).
Fig. 15

Binging modes of compound 99 into COX-2 (pdb code: 1CX2): a 3D binding mode; b 2D binding mode, showing hydrogen bonds ( Open image in new window ), carbon hydrogen bonds ( Open image in new window ), electrostatic interaction of the pi-cation ( Open image in new window ), hydrophobic interactions of the pi-sigma ( Open image in new window ), pi-alkyl ( Open image in new window ) and alkyl ( Open image in new window ) types

Moreover compound 99 formed also one electrostatic interaction with of the pi-cation type with ARG120 and 21 hydrophobic interactions (BL = 3.62–5.42 Å) with COX-2 which contribute to this high selectivity. Ten of these interactions were due to the aliphatic butyl/isopentyl side chains (Fig. 15).

LigPlot view of compound 99 showed two hydrogen bonds with ARG513 and LEU352 with BL of 2.96 and 3.31 Å, respectively, while LigPlot view of compound 92 showed one hydrogen bond only with TYR385 with BL of 2.85 Å (Fig. 16).
Fig. 16

a LigPlot view of compound 92 into the active site of COX-2 (pdb code: 1CX2) showing one hydrogen bond with TYR355 with bond length of 2.87 Å; b LigPlot view of compound 99 into the active site of COX-1 showing two hydrogen bonds with ARG120 and LEU352 with bond length of 2.96 and 3.33 Å, respectively

3.2 Drug-likeness and synthetic feasibility study

During the last 3 decades huge number of compounds were reported with in vitro analgesic and anti-inflammatory activities, but only few numbers of these compounds were passed to clinical trials. This problem is mainly attributed to pharmacokinetic problems. As a result, the most promising analogs (5, 7, 11, 15, 23, 29, 38, 40, 46, 54, 56, 61, 92, 99) in this study were selected for drug-likeness study. The study was performed using both SwissADME (http://www.swissadme.ch/), developed by the Molecular Modeling Group of the Swiss Institute of Bioinformatics [27], and Molsoft (http://molsoft.com/mprop/) which was developed by Molsoft LLC according to our previous reports [12, 31].

The study revealed that all the selected compounds displayed molecular weight, molecular volume in the range of 257.26–382.56 (< 500), log P values (MlogP) in the range of 2.94 -5.49, hydrogen bond donors range (HD) ≤ 5 and hydrogen bond acceptors (HA) ≤ 10 (Table 7).
Table 7

Molecular properties related to drug-likeness

Comp.

MW

MVa

TPSA

MlogP

RBs

HA

HD

LVs

DLSa

SA

3

273.71

249.92

53.09

3.07

2

2

2

No

0.30

2.2

5

273.71

249.92

53.09

3.07

2

2

2

No

0.42

2.18

7

257.26

238.64

53.09

2.94

2

3

2

No

0.27

2.28

11

318.17

254.58

53.09

3.19

2

3

2

No

0.18

2.23

15

281.35

289.35

53.09

3.28

4

2

2

No

0.95

2.37

23

313.37

313.01

53.09

3.89

5

3

2

No

0.32

2.81

29

266.3

273.43

37.30

3.74

3

2

1

No

0.96

3.09

38

298.35

298.12

37.30

4.35

4

3

1

1b

0.37

3.37

40

326.40

333.93

37.30

4.80

6

3

1

1b

0.10

3.61

46

282.33

280.88

50.44

3.28

4

3

1

No

0.75

3.02

54

300.32

288.37

50.44

3.66

4

4

1

No

0.26

3.17

56

328.38

324.18

50.44

4.12

6

4

1

No

− 0.02

3.38

61

284.37

282.19

65.54

3.92

3

2

1

No

0.81

2.67

92

368.53

387.39

65.54

5.28

7

2

1

1b

0.69

3.62

99

382.56

405.30

65.54

5.49

8

2

1

1b

0.51

3.74

MW, molecule weight; MV, molecular volume (A3); TPSA, topological polar surface area (A2); MlogP, Moriguchi’s logP; RBs, rotatable bonds; HA, hydrogen bond acceptors; HD, hydrogen bond donors; SA, Synthetic accessibility; DLS, Drug-likeness score

aParameters calculated using Molsoft (http://www.swissadme.ch/), other parameters were calculated using (http://molsoft.com/mprop/); 1b, one violation, MlogP > 4.15

Except for compound 56, all the selected compounds showed drug likeness score (DLS) in the range of 0.1–0.96, as compared to 0.30 for S-carprofen. The synthetic accessibility of the selected compounds was in the range of 2.18–3.74 compared to 2.2 for S-carprofen.

4 Conclusion

In this study, four series of carprofen analogs were designed by isosteric replacement of the -NH- with -O-, -S- and -CH2- groups. The impact of fluoro/alkyl substitutions on binding affinity and inhibition constants of the designed analogs was evaluated using molecular docking study. The results revealed that the binding affinity to COX-1/2 was dependent on position and length of the alkyl group. Compound 66 displayed the highest binding affinity for COX-1 and COX-2 with ΔGb= 11.2 and 10.15 kcal/mol, respectively. Compound 56 displayed the highest potential selectivity for COX-1 (Ki ratio = 43.09), while compound 99 was the most selective for COX-2 (Ki ratio = 39.9). Compound 29 showed drug-likeness score of 0.96, as compared to 0.30 for S-carprofen, while compound 5, 7 and 11 showed synthetic accessibility score comparable to the parent carprofen. Taken together, these results highlighted the impact of hydrophobic interactions of alkyl groups in modifying affinity and selectivity to COX-1 or COX-2.

Notes

Compliance with ethical standards

Conflict of interest

The authors declare that they have no conflict of interest.

References

  1. 1.
    Chen Z, Wang Z-C, Yan X-Q, Wang P-F, Lu X-Y, Chen L-W, Zhu H-L, Zhang H-W (2015) Design, synthesis, biological evaluation and molecular modeling of dihydropyrazole sulfonamide derivatives as potential COX-1/COX-2 inhibitors. Bioorg Med Chem Lett 25(9):1947–1951.  https://doi.org/10.1016/j.bmcl.2015.03.022 CrossRefGoogle Scholar
  2. 2.
    Alegaon SG, Hirpara MB, Alagawadi KR, Hullatti KK, Kashniyal K (2014) Synthesis of novel pyrazole–thiadiazole hybrid as potential potent and selective cyclooxygenase-2 (COX-2) inhibitors. Bioorg Med Chem Lett 24(22):5324–5329.  https://doi.org/10.1016/j.bmcl.2014.08.062 CrossRefGoogle Scholar
  3. 3.
    Attallah KM, Gouda AM, Ibrahim IT, Abouzeid L (2017) Design, synthesis, 99mTc labeling, and biological evaluation of a novel pyrrolizine derivative as potential anti-inflammatory agent. Radiochemistry 59(6):630–638.  https://doi.org/10.1134/S10663622170600121 CrossRefGoogle Scholar
  4. 4.
    Gouda AM, Ali HI, Almalki WH, Azim MA, Abourehab MAS, Abdelazeem AH (2016) Design, synthesis, and biological evaluation of some novel pyrrolizine derivatives as COX inhibitors with anti-inflammatory/analgesic activities and low ulcerogenic liability. Molecules 21(2):201.  https://doi.org/10.3390/molecules21020201 CrossRefGoogle Scholar
  5. 5.
    Kaur A, Pathak DP, Sharma V, Narasimhan B, Sharma P, Mathur R, Wakode S (2018) Synthesis, biological evaluation and docking study of N-(2-(3,4,5-trimethoxybenzyl)benzoxazole-5-yl) benzamide derivatives as selective COX-2 inhibitor and anti-inflammatory agents. Bioorg Chem 81:191–202.  https://doi.org/10.1016/j.bioorg.2018.07.007 CrossRefGoogle Scholar
  6. 6.
    Taylor AP, Robinson RP, Fobian YM, Blakemore DC, Jones LH, Fadeyi O (2016) Modern advances in heterocyclic chemistry in drug discovery. Org Biomol Chem 14(28):6611–6637.  https://doi.org/10.1039/c6ob00936k CrossRefGoogle Scholar
  7. 7.
    Abbas SE, Awadallah FM, Ibrahim NA, Gouda AM, Shehata BA (2011) Design, synthesis and preliminary evaluation of some novel [1,4]diazepino[5,6-b]pyrrolizine and 6-(2-oxopyrrolidino)-1H-pyrrolizine derivatives as anticonvulsant agents. Med Chem Res 20:1015–1023.  https://doi.org/10.1007/s00044-010-9429-8 CrossRefGoogle Scholar
  8. 8.
    Szabo G, Fischer J, Kis-Varga A, Gyires K (2008) New celecoxib derivatives as anti-inflammatory agents. J Med Chem 51(1):142–147.  https://doi.org/10.1021/jm070821f CrossRefGoogle Scholar
  9. 9.
    Dogne J-M, Hanson J, Supuran C, Pratico D (2006) Coxibs and cardiovascular side-effects: from light to shadow. Curr Pharm Des 12(8):971–975CrossRefGoogle Scholar
  10. 10.
    Ren S-Z, Wang Z-C, Zhu D, Zhu X-H, Shen F-Q, Wu S-Y, Chen J-J, Xu C, Zhu H-L (2018) Design, synthesis and biological evaluation of novel ferrocene-pyrazole derivatives containing nitric oxide donors as COX-2 inhibitors for cancer therapy. Eur J Med Chem 157:909–924.  https://doi.org/10.1016/j.ejmech.2018.08.048 CrossRefGoogle Scholar
  11. 11.
    Gouda AM, Abdelazeem AH, Omar HA, Abdalla AN, Abourehab MAS, Ali HI (2017) Pyrrolizines: design, synthesis, anticancer evaluation and investigation of the potential mechanism of action. Bioorg Med Chem 25:5637–5651.  https://doi.org/10.1016/j.bmc.2017.08.039 CrossRefGoogle Scholar
  12. 12.
    Gouda AM, Abdelazeem AH, Abdalla AN, Ahmed M (2018) Pyrrolizine-5-carboxamides: exploring the impact of various substituents on anti-inflammatory and anticancer activities. Acta Pharm 68:251–273.  https://doi.org/10.2478/acph-2018-0026 CrossRefGoogle Scholar
  13. 13.
    Cingolani G, Panella A, Perrone MG, Vitale P, Di Mauro G, Fortuna CG, Armen RS, Ferorelli S, Smith WL, Scilimati A (2017) Structural basis for selective inhibition of Cyclooxygenase-1 (COX-1) by diarylisoxazoles mofezolac and 3-(5-chlorofuran-2-yl)-5-methyl-4-phenylisoxazole (P6). Eur J Med Chem 138:661–668.  https://doi.org/10.1016/j.ejmech.2017.06.045 CrossRefGoogle Scholar
  14. 14.
    Kitamura T, Kawamori T, Uchiya N, Itoh M, Noda T, Matsuura M, Sugimura T, Wakabayashi K (2002) Inhibitory effects of mofezolac, a cyclooxygenase-1 selective inhibitor, on intestinal carcinogenesis. Carcinogenesis 23(9):1463–1466CrossRefGoogle Scholar
  15. 15.
    Niho N, Kitamura T, Takahashi M, Mutoh M, Sato H, Matsuura M, Sugimura T, Wakabayashi K (2006) Suppression of azoxymethane-induced colon cancer development in rats by a cyclooxygenase-1 selective inhibitor, mofezolac. Cancer Sci 97(10):1011–1014.  https://doi.org/10.1111/j.1349-7006.2006.00275.x CrossRefGoogle Scholar
  16. 16.
    Bostrom IM, Nyman GC, Lord PE, Haggstrom J, Jones BEV, Bohlin HP (2002) Effects of carprofen on renal function and results of serum biochemical and hematologic analyses in anesthetized dogs that had low blood pressure during anesthesia. Am J Vet Res 63(5):712–721CrossRefGoogle Scholar
  17. 17.
    Van der Vijver RJ, van Laarhoven CJHM, Lomme RMLM, Hendriks T (2012) Carprofen for perioperative analgesia causes early anastomotic leakage in the rat ileum. BMC Vet Res 8:247.  https://doi.org/10.1186/1746-6148-8-247 CrossRefGoogle Scholar
  18. 18.
    Warner TD, Giuliano F, Vojnovic I, Bukasa A, Mitchell JA, Vane JR (1999) Nonsteroid drug selectivities for cyclo-oxygenase-1 rather than cyclo-oxygenase-2 are associated with human gastrointestinal toxicity: a full in vitro analysis. Proc Natl Acad Sci USA 96(13):7563–7568CrossRefGoogle Scholar
  19. 19.
    Li J, Lynch MP, Demello KL, Sakya SM, Cheng H, Rafka RJ, Bronk BS, Jaynes BH, Kilroy C, Mann DW, Haven LM, Kolosko NL, Petras C, Seibel SB, Lund LA (2005) In vitro and in vivo profile of 2-(3-di-fluoromethyl-5-phenylpyrazol-1-yl)-5-methanesulfonylpyridine, a potent, selective, and orally active canine COX-2 inhibitor. Bioorg Med Chem 13(5):1805–1809.  https://doi.org/10.1016/j.bmc.2004.11.048 CrossRefGoogle Scholar
  20. 20.
    Favia AD, Habrant D, Scarpelli R, Migliore M, Albani C, Bertozzi SM, Dionisi M, Tarozzo G, Piomelli D, Cavalli A, De Vivo M (2012) Identification and characterization of carprofen as a multitarget fatty acid amide hydrolase/cyclooxygenase inhibitor. J Med Chem 55(20):8807–8826.  https://doi.org/10.1021/jm3011146 CrossRefGoogle Scholar
  21. 21.
    Bertolacci L, Romeo E, Veronesi M, Magotti P, Albani C, Dionisi M, Lambruschini C, Scarpelli R, Cavalli A, De Vivo M, Piomelli D, Garau G (2013) A binding site for nonsteroidal anti-inflammatory drugs in fatty acid amide hydrolase. J Am Chem Soc 135(1):22–25.  https://doi.org/10.1021/ja308733u CrossRefGoogle Scholar
  22. 22.
    Ren S-Z, Wang Z-C, Zhu X-H, Zhu D, Li Z, Shen F-Q, Duan T-Y, Cao H, Zhao J, Zhu H-L (2018) Design and biological evaluation of novel hybrids of 1, 5-diarylpyrazole and Chrysin for selective COX-2 inhibition. Bioorg Med Chem 26(14):4264–4275.  https://doi.org/10.1016/j.bmc.2018.07.022 CrossRefGoogle Scholar
  23. 23.
    Abdelazeem AH, El-Saadi MT, Safi El-Din AG, Omar HA, El-Moghazy SM (2017) Design, synthesis and analgesic/anti-inflammatory evaluation of novel diarylthiazole and diarylimidazole derivatives towards selective COX-1 inhibitors with better gastric profile. Bioorg Med Chem 25(2):665–676.  https://doi.org/10.1016/j.bmc.2016.11.037 CrossRefGoogle Scholar
  24. 24.
    Malkowski MG, Ginell SL, Smith WL, Garavito RM (2000) The productive conformation of arachidonic acid bound to prostaglandin synthase. Science 289(5486):1933–1937CrossRefGoogle Scholar
  25. 25.
    Selinsky BS, Gupta K, Sharkey CT, Loll PJ (2001) Structural analysis of NSAID binding by prostaglandin H2 synthase: time-dependent and time-independent inhibitors elicit identical enzyme conformations. Biochemistry 40(17):5172–5180CrossRefGoogle Scholar
  26. 26.
    Kurumbail RG, Stevens AM, Gierse JK, McDonald JJ, Stegeman RA, Pak JY, Gildehaus D, Miyashiro JM, Penning TD, Seibert K, Isakson PC, Stallings WC (1996) Structural basis for selective inhibition of cyclooxygenase-2 by anti-inflammatory agents. Nature 384(6610):644–648.  https://doi.org/10.1038/384644a0 CrossRefGoogle Scholar
  27. 27.
    Daina A, Michielin O, Zoete V (2017) SwissADME: a free web tool to evaluate pharmacokinetics, drug-likeness and medicinal chemistry friendliness of small molecules. Sci Rep 7:42717.  https://doi.org/10.1038/srep42717 CrossRefGoogle Scholar
  28. 28.
    Yung-Chi C, Prusoff WH (1973) Relationship between the inhibition constant (KI) and the concentration of inhibitor which causes 50 per cent inhibition (I50) of an enzymatic reaction. Biochem Pharmacol 22:3099–3108CrossRefGoogle Scholar
  29. 29.
    Ramsay RR, Tipton KF (2017) Assessment of enzyme inhibition: a review with examples from the development of monoamine oxidase and cholinesterase inhibitory drugs. Molecules 22(7):1192.  https://doi.org/10.3390/molecules22071192 CrossRefGoogle Scholar
  30. 30.
    Gensicka-Kowalewska M, Cholewinski G, Dzierzbicka K (2017) Recent developments in the synthesis and biological activity of acridine/acridone analogues. RSC Adv 7:15776–15804.  https://doi.org/10.1039/C7RA01026E CrossRefGoogle Scholar
  31. 31.
    Almalki FA, Gouda AM, Bin Ali MH, Almehmadi OM (2019) Profens: a comparative molecular docking study into cyclooxygenase-1/2. Drug Invent Today 11(2):480–487Google Scholar

Copyright information

© Springer Nature Switzerland AG 2019

Authors and Affiliations

  1. 1.Department of Medicinal Chemistry, Faculty of PharmacyBeni-Suef UniversityBeni-SuefEgypt
  2. 2.Department of Pharmaceutical Chemistry, Faculty of PharmacyUmm Al-Qura UniversityMakkahSaudi Arabia

Personalised recommendations