Design and synthesis of celecoxib and rofecoxib analogues as selective cyclooxygenase-2 (COX-2) inhibitors: Replacement of sulfonamide and methylsulfonyl pharmacophores by an azido bioisostere

Article abstract of DOI:10.1021/jm010153c

Celecoxib (13) and rofecoxib (17) analogues, in which the respective SO₂NH₂ and SQ₂Me hydrogen-bonding pharmacophores were replaced by a dipolar azido bioisosteric substituent, were investigated. Molecular modeling (docking) studies showed that the azido substituent of these two analogues (13, 17) was inserted deep into the secondary pocket of the human COX-2 binding site where it undergoes electrostatic interaction with Arg⁵¹³. The azido analogue of rofecoxib (17), the most potent and selective inhibitor of COX-2 (COX-1 IC₅₀ = 159.7 μM; COX-2 IC₅₀ = 0.196 μM; COX-2 selectivity index = 812), exhibited good oral antiinflammatory and analgesic activities.

Full text of DOI:10.1021/jm010153c

J . Med. Chem. 2001, 44, 3039-3042
3039
Brief Articles
Design a n d Syn th esis of Celecoxib a n d Rofecoxib An a logu es a s Selective
Cyclooxygen a se-2 (COX-2) In h ibitor s: Rep la cem en t of Su lfon a m id e a n d
Meth ylsu lfon yl P h a r m a cop h or es by a n Azid o Bioisoster e
Amgad G. Habeeb, P. N. Praveen Rao, and Edward E. Knaus*
Faculty of Pharmacy and Pharmaceutical Sciences, University of Alberta, Edmonton, Alberta, Canada
Received April 3, 2001
Celecoxib (13) and rofecoxib (17) analogues, in which the respective SO2NH2 and SO2Me
hydrogen-bonding pharmacophores were replaced by a dipolar azido bioisosteric substituent,
were investigated. Molecular modeling (docking) studies showed that the azido substituent of
these two analogues (13, 17) was inserted deep into the secondary pocket of the human COX-2
binding site where it undergoes electrostatic interaction with Arg513. The azido analogue of
rofecoxib (17), the most potent and selective inhibitor of COX-2 (COX-1 IC50 ) 159.7 µM; COX-2
IC50 ) 0.196 µM; COX-2 selectivity index ) 812), exhibited good oral antiinflammatory and
analgesic activities.
In tr od u ction
Many selective COX-2 inhibitors belong to a tricyclic
group of compounds with a central ring possessing a
diaryl stilbene-like structure with a sulfonyl (SO2) group
at the para position of one of the aryl rings, such as
DuP-697 (1),1 celecoxib (Celebrex) (2),2 SC-588 (3),2
rofecoxib (Vioxx) (4),3 3-(4-methylsulfonylphenyl)-4-
phenyl-3-trifluoromethylisoxazole (5),4 and 2,3-dimethyl-
5-(4-methylsulfonylphenyl)-4-phenyl-4-isoxazoline (6),5
as illustrated in Figure 1. The SO2Me and SO2NH2
pharmacophores are believed to induce COX-2 selectiv-
ity by insertion into the secondary pocket of COX-2
which is absent in COX-1. The secondary pocket present
in COX-2 has been attributed to the presence of isoleu-
cine (Ile523) in COX-1 relative to the smaller valine
(Val523) in COX-2.6 Replacement of histidine (His513) in
COX-1 by arginine (Arg513) in COX-2 has been reported
to play a key role in the hydrogen-bond network of the
COX active site. Histidine (His90), glutamine (Gln192),
and tyrosine (Tyr355) control the access of ligands into
F igu r e 1. Representative examples of selective COX-2 inhibi-
the secondary pocket.7 The interaction of Arg513 with
the bound ligand has been reported to be a requirement
for the time-dependent inhibition of COX-2.8 The pres-
ence of the Arg513 residue, to our knowledge, has not
been exploited for the design of selective COX-2 inhibi-
tors. Accordingly, we now describe the design, synthesis,
cyclooxygenase inhibitory, analgesic and antiinflamma-
tory activities, and some molecular modeling studies for
the pyrazole regioisomers 10 and 13 and the furanone
17 that possess an azido group in place of the SO2NH2
and SO2Me pharmacophores present in celecoxib and
rofecoxib, respectively.
tors having a central five-membered heterocyclic ring.
Ch em istr y
Reaction of 4-nitrophenylhydrazine with 1-(4-meth-
ylphenyl)-4,4,4-trifluorobutane-1,3-dione2 in EtOH af-
forded the cyclic pyrazoline-5-ol 7 which eliminated a
molecule of water upon treatment with HOAc at reflux
temperature to yield the pyrazole 8 (see Scheme 1).
Reduction of the nitro group in the pyrazole 8 with
hydrazine hydrate and 10% Pd/C, using a method
reported by Penning et al.,2 yielded the corresponding
amino product 9. Diazotization of 9, and treatment of
the diazonium salt with NaN3, afforded the 1,3-regioi-
somer of celecoxib having an azido substituent in place
of the SO2NH2 pharmacophore. In contrast, the 1,5-
regioisomer was prepared by condensation of 4-nitro-
* Address correspondence to Dr. Edward E. Knaus, Faculty of
Pharmacy and Pharmaceutical Sciences, University of Alberta, Edm-
onton, Alberta, Canada T6G 2N8. Tel: 780-492-5993. Fax: 780-492-
1217. E-mail: eknaus@pharmacy.ualberta.ca.
10.1021/jm010153c CCC: $20.00 © 2001 American Chemical Society
Published on Web 08/03/2001

3040 J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 18
Brief Articles
Sch em e 1a
Sch em e 2a
a
Reagents and conditions: (a) Br2, CHCl3, 25 °C, 2 h; (b)
PhCH2COOH, Et3N, CH3CN, 25 °C, 1 h; (c) Et3N, CH3CN, reflux,
8 h.
F igu r e 2. Azide resonance hybrid structures.
a
Reagents and conditions: (a) EtOH, reflux, 20 h; (b) HOAc,
reflux, 2 h; (c) H2NNH2‚xH2O, 10% Pd/C, reflux, 45 min; (d)
NaNO2/HCl and then NaN3, 0-5 °C, 45 min; (e) EtOH, HCl, reflux,
20 h.
phenylhydrazine with 1-(4-methylphenyl)-4,4,4-trifluo-
robutane-1,3-dione in EtOH under acidic reaction
conditions to yield the pyrazole 11 (76%). This reaction,
performed under acidic reaction conditions, provided a
superior yield relative to related reactions carried out
under neutral reaction conditions.2 The azido analogue
of celecoxib 13 was prepared starting from the pyrazole
11 using the same reaction sequence used for the
elaboration of the nitro compound 10 to the azido
product 11.
The azido analogue 17 of rofecoxib, where MeSO2 is
replaced by N3, was prepared starting with the bromi-
nation of 4-azidoacetophenone 149 using Br2 according
to a reported method.10 The subsequent reaction of the
bromo compound 15 with phenylacetic acid in the
presence of (Et)3N gave 4-(4-azidophenyl)-3-phenyl-
2(5H)furanone (17) that was formed via the intermedi-
ate ester 16 as illustrated in Scheme 2.
F igu r e 3. Docking the pyrazole 13 (ball-and-stick) in the
active site of human COX-2 (line and stick) (Eintermolecular
)
-46.49 kcal/mol). The C-atom of the CF3 substituent is 10.26
Å from the phenolic OH of Tyr355, but removed from Ser530 (OH)
by 7.18 Å. The terminal N-atom of the azido substituent is
about 4.52 Å inside the entrance to the secondary pocket
(Val523). The center of the N-1 phenyl ring is about 3.95 Å from
the entrance to the secondary pocket (Val523).
of compounds, replacement of SO2Me by SO2CF3, COMe,
PO(OH)Me, CO2H, PO(OH)2, or SO(dNH)Me abolished
COX-2 inhibitory activity.12 A similar replacement of
SO2Me by NO2, which can dispose a pair of oxygen
atoms such as SO2, in the 1,5-diarylpyrazole class also
abolished COX-2 inhibitory activity.2
The azido substituent is particularly attractive since
it has the potential to undergo electrostatic (ion-ion)
binding interactions with amino acid residues, particu-
larly Arg513, lining the secondary pocket of COX-2.
Covalent azides can be viewed as resonant hybrids
between structures A, B, and C (see Figure 2).13 Pauling
rejected C as a major contributor based on the adjacent
charge rule.14 The remaining hybrids A and B predict
a 2.5 bond order for the N2-N3 bond and a 1.5 bond
order for the N1-N2 bond (see D, Figure 2). This
prediction was in very good agreement with a structure
determination of methyl azide (D) where the bond
Resu lts a n d Discu ssion
Celecoxib and rofecoxib analogues, having an azido
group in place of the respective SO2NH2 and SO2Me
pharmacophores, were investigated to determine whether
the azido substituent is a suitable bioisostere with
respect to selective COX-2 inhibition, and AI and
analgesic activities. Structure-activity studies for the
tricyclic class of selective COX-2 inhibitors have shown
that a SO2Me or SO2NH2 substituent at the para
position of one aryl ring usually confers optimal COX-2
inhibitory potency.11 In the 1,2-diarylcyclopentene class

Brief Articles
J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 18 3041
Ta ble 1. Antiinflammatory and Analgesic Activities, in Vitro COX-1 and COX-2 Inhibition Data, and Molecular Volumes of
1-(4-Azidophenyl)-3-(4-methylphenyl)-5-trifluoromethylpyrazole (10), 1-(4-Azidophenyl)-5-(4-methylphenyl)-3-trifluoromethylpyrazole
(13), and 4-(4-Azidophenyl)-3-phenyl-2(5H)furanone (17)
AI activitya
analgesic activityb
IC50, µMd
% inhibition
% inhibition
at 5 h
% inhibition
at 30 min
% inhibition
at 60 min
vol.
selectivity index
(COX-1/COX-2)
compd
10
13
17
celecoxib
rofecoxib
at 3 h
(Å3)c
COX-1
9.88
COX-2
278.3
279.4
242.1
298.4
262.5
2.63
1.55
0.196
0.0507
0.34g
3.74
64.55
812.4
404
46.6 ( 4.4
42.9 ( 1.0
79.9 ( 1.9e
16.9 ( 2.7
27.5 ( 4.6
58.2 ( 1.8f
60.9 ( 9.4
46.7 ( 1.3
31.7 ( 9.6
63.1 ( 1.2
60.6 ( 1.6
62.0 ( 7.3
>100
159.72
22.9
26.0g
76.5
a
Inhibitory activity on carrageenan-induced rat paw edema; the result is the mean value ( SEM using four animals following a 50
b
mg/kg oral dose of the test compound. Inhibitory activity in the rat 4% NaCl-induced abdominal constriction assay; the result is the
mean value ( SEM using four animals following a 50 mg/kg ip dose of the test compound. c The volume of the molecule, after minimization
d
using the MM3 force field, was calculated using the Alchemy 2000 program. The result (IC50, µM) is the mean of two determinations.
e ID50 ) 10.8 mg/kg po dose. f ID50 ) 40.8 mg/kg po dose. Data taken from the literature for inhibition of purified human recombinant
g
COX-1 and COX-2.19
F igu r e 4. Docking the pyrazole (10) (ball-and-stick) in the
active site of human COX-2 (line and stick) (Eintermolecular
)
-34.12 kcal/mol). The C-atom of the CF3 substituent is 12.25
Å from the phenolic OH of Tyr355, but removed from the Ser530
(OH) by 8.56 Å. The terminal N-atom of the azido substituent
is about 10.30 Å outside the entrance to the secondary pocket
(Val523). The center of the N-1 phenyl ring is about 6.46 Å
outside the entrance to the secondary pocket (Val523).
F igu r e 5. Docking the furanone 17 (ball-and-stick) in the
active site of human COX-2 (line and stick) (Eintermolecular
)
-49.6 kcal/mol). The O-atom of the CO substituent is 13.51 Å
from the phenolic OH of Tyr355, but removed from Ser530 (OH)
by 4.03 Å. The terminal N-atom of the azido substituent is
about 4.52 Å inside the entrance to the secondary pocket
(Val523). The center of the C-4 phenyl ring is about 4.11 Å from
the entrance to the secondary pocket (Val523).
lengths from electron diffraction studies15 were as
follows: N2-N3 ) 1.12 Å, N1-N2 ) 1.24 Å, C-N1 ) 1.47
Å ,and the C-N1-N2 bond angle was 120°. The linear
configuration of the azido group is in agreement with
the sp3 hybridization indicated by the lack of nonbonded
electron pairs on N2.13 The azido group is slightly
smaller in size [MR (molar refractivity) ) 10.20] than
a SO2Me (MR ) 13.49) or SO2NH2 (MR ) 12.28)
substituent, but more lipophilic (π ) 0.46) relative to
the more polar SO2Me (π ) -1.63) and SO2NH2 (π )
-1.82) subsituents16 which have the potential to im-
prove absorption and provide a more rapid onset of
action.11
Docking 1-(4-azidophenyl)-5-(4-methylphenyl)-3-trif-
luoromethylpyrazole (13) in the active site of human
COX-2 (1CX2 PDB file), showed that the terminal
N-atom of the azido group was inserted into the second-
ary COX-2 pocket about 4.52 Å from Val523, and about
3.15 Å from the center of the guanidino group of Arg513
(see Figure 3). This orientation of the pyrazole 13 within
the COX-2 active site provides an intermolecular energy
between the enzyme and pyrazole 13 of about -46.19
kcal/mol, where the electrostatic component accounts for
about 12% of this total energy. In comparison, the
celecoxib (2) docked complex showed an intermolecular
energy between the enzyme and celecoxib of about
-45.16 kcal/mol where 2.6% was due to an electrostatic
component. This difference is likely due to a greater
electrostatic interaction between the dipolar azido group
in pyrazole 13 and the charged guanidino moiety of
Arg513 in the secondary COX-2 pocket. Similar docking
of the 1-(4-azidophenyl)-3-(4-methylphenyl)-5-trifluo-
romethylpyrazole regioisomer (10) in the active site of
human COX-2 showed that the azido substituent did
not insert into the secondary pocket since the ligand 10
extended parallel to the longitudinal axis of the hydro-
phobic primary COX-2 channel (cavity), in a manner
characteristically observed for nonselective COX-2 in-
hibitors17 as illustrated in Figure 4.
These molecular modeling studies correlate well with
in vitro enzyme inhibition data. In this regard, the 1,5-
pyrazole regioisomer 13 showed selective inhibition of
COX-2 [COX-1 IC50 > 100 µM; COX-2 IC50) 1.5 µM;
selectivity index (SI) ≈ 64], whereas the 1,3-regioisomer
10 showed a modest COX-2 selectivity ≈ 4. These
results are similar to those described for other studies2,18
utilizing compounds not having a 1,2-diarylstilbene-like
structure. Docking the rofecoxib analogue 17, in which

3042 J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 18
Brief Articles
the SO2Me moiety was replaced by an azido substituent,
in the active site of human COX-2 showed a similar
interaction between the azido group and the COX-2
secondary pocket amino acid residues similar to that
observed for the pyrazole 13 (see Figure 5). The inter-
molecular energy between the ligand 17 and the enzyme
was -49.6 kcal/mol with the electrostatic component
comprising 5.8% of the total energy. In contrast, the
rofecoxib (4) docked complex showed an intermolecular
energy of -42.16 kcal/mol where only 1.2% was due to
an electrostatic component. The higher electrostatic
component for the azido compound 17 (5.8%), relative
to that for rofecoxib 4 (1.2%), is attributed to the fact
that the MeSO2 moiety present in rofecoxib undergoes
H-bonding to the imidazole NH of Hist90 in the second-
ary pocket. In contrast, the azido moiety in 17 undergoes
an electrostatic interaction with Arg513 in the secondary
COX-2 pocket. The lower contribution of the electro-
static energy to the total intermolecular energy in the
case of the furanone 17, relative to the pyrazole 13, can
be attributed to the observed hydrogen bonding interac-
tion between the O-atom of the CdO in the furanone
structure and residues lining the primary COX-2 chan-
nel, particularly Arg120. The azido analogue of refocoxib
17 exhibited a potent and selective inhibition of COX-2
(COX-2 IC50 ) 0.196 µM; COX-1 IC50 ) 159.7 µM; SI ≈
812). The molecular volumes of the selective COX-2
inhibitors 13 (279.4 Å3) and 17 (242.1 Å3) are moderately
smaller than that for the selective COX-2 inhibitors
celecoxib (298.4 Å3) and rofecoxib (262.5 Å3).
Refer en ces
(1) Gans, K. R.; Galbraith, W.; Roman, R. J .; Haber, S. B.; Kerr, J .
S.; Schmidt, W. K.; Smith, C.; Hewes, W. E.; Ackerman, N. R.
Antiinflammatory and safety profile of DuP-697, a novel orally
effective prostaglandin synthesis inhibitor. J . Pharmacol. Exp.
Ther. 1990, 254, 180-187.
(2) Penning, T.; et al. Synthesis and biological evaluation of the 1,5-
diarylpyrazole class of cyclooxygenase-2 inhibitors: Identifica-
tion of 4-[5-(4-methylphenyl)-3-(trifluoromethyl)-1H-pyrazol-1-
yl)]benzenesulfonamide (SC-58635, Celecoxib). J . Med. Chem.
1997, 40, 1347-1365.
(3) Desmond, R.; Dolling, U.; Marcune, B.; Tillyer, R.; Tschaen, D.
Process for making phenylheterocycles useful as COX-2 inhibi-
tors. World Patent WO96/08482, 21 March 1996; Chem. Abstr.
1996, 125, P 86474.
(4) Habeeb, A. G.; Praveen Rao, P. N.; Knaus, E. E. Design and
syntheses of 4,5-diarylisoxazoles: Novel inhibitors of cyclooxy-
genase-2 (COX-2) with analgesic and antiinflammatory activity.
Drug Devel. Res. 2000, 51, 273-286.
(5) Habeeb, A. G.; Praveen Rao, P. N.; Knaus, E. E. Design and
Synthesis of 4,5-diphenyl-4-isoxazolines: Novel inhibitors of
Cyclooxygenase-2 (COX-2) with an analgesic and antiinflam-
matory activity. J . Med. Chem. In press.
(6) Luong, C.; Miller, A.; Barnett, J .; Chow, J .; Ramesha, C.;
Browner, M. F. Flexibility of the NSAID binding site in the
structure of human cyclooxygenase-2. Nat. Struct. Biol. 1996,
3, 927-933.
(7) Llorens, O.; Perez, J . L.; Palomer, A.; Mauleon, D. Structural
basis for dynamic mechanism of ligand binding to COX. Biorg.
Med. Chem. Lett. 1999, 9, 2779-2784.
(8) Garavito, R. M.; DeWitt, D. L. The Cyclooxygenase isoforms:
Structural insights into the conversion of arachidonic acid to
prostaglandins. Biochim. Biophys. Acta 1999, 1441, 278-287.
(9) Kim, P. R.; Carlson K. E.; Katzenellen-bogen, J . A. Progestin
16 alpha-17-alpha-dioxolane ketals as molecular probes for the
progesterone receptor: Synthesis, binding affinity, and photo-
chemical evaluation. J . Med. Chem. 1993, 36, 1111-1119.
(10) Vijayaraghavan, S. T.; Balasubramanian, T. R. Synthesis of 3,4-
diaryl-2(5H)furanones. Ind. J . Chem. 1986, 25B, 760-761.
(11) Khanna, I. K.; Weier, R. M.; Yu, Y.; Xu, X. D.; Koszyk, F. J .;
Collins, P. W.; Koboldt, C. M.; Veenhuizen, A. W.; Perkins, W.
E.; Casler, J . J .; Masferrer, J . L.; Zhang, Y. Y.; Gregory, S. A.;
Seibert, K.; Isakson, P. C. 1,2-Diarylimidazoles as potent,
cyclooxygenase-2 selective, and orally active antiinflammatory
agents. J . Med. Chem. 1997, 40, 1634-1647.
The azido analogues of celecoxib 13 and rofecoxib 17
exhibited good AI and analgesic activities (see Table 1).
Con clu sion s
In conclusion, the dipolar azido group is a bioisostere
of the SO2NH2 and SO2Me hydrogen-bonding pharma-
cophores present in many selective COX-2 inhibitors,
and the azido analogues 13 and 17 may be useful
biochemical agents for photoaffinity labeling of the
COX-2 enzyme.
(12) Li, J . J .; Anderson, G. D.; Burton, E. G.; Nita Cogburn, J .;
Collins, J . T.; Garland D. J .; Gregory, S. A.; Huang, H.-C.;
Isakson, P. C.; Koboldt, C. M.; Logusch, E. W.; Norton, M. B.;
Perkins, W. E.; Reinhard, E. J .; Seibert, K.; Veenhuizen, A. W.;
Zhang, Y.; Reitz, D. B. 1,2-Diarylcyclopentenes as selective
cyclooxygenase-2 inhibitors and orally active antiinflammatory
agents. J . Med. Chem. 1995, 38, 4570-4578.
(13) Lieber, E.; Curtice, J .; Rao, C. Structure and reactivity of organic
azides. Chem. Ind. 1966, 586.
Exp er im en ta l Section
(14) Pauling, L. In The Nature of the Chemical Bond, 3rd ed.; Cornell
University Press: Ithaca, New York, 1960.
(15) Livingston, R.; Ramachandra Rao, C. An electron diffraction
investigation of the molecular structure of methyl azide. J . Phys.
Chem. 1960, 64, 756-759.
(16) Hansch, C.; Leo, A.; Unger, S. H.; Kim. K. H.; Nikaitani D.; Lien,
E. J . Aromatic substituent constants for structure activity
correlations. J . Med. Chem. 1973, 16, 1207-1216.
(17) Sahi, S.; Srinivasan, M.; Kothekar, V. 530 ps molecular dynamics
simulation of indoprofen and NS-398 with COX-1 and COX-2.
Study of perturbative changes in the complexes. J . Mol. Struct.
(THEOCHEM) 2000, 498, 133-148.
Cyclooxygen a se In h ibition Stu d ies. All compounds de-
scribed herein were tested for their ability to inhibit COX-1
and COX-2 using a COX-(ovine) inhibitor screening kit (catalog
no. 560101, Cayman Chemical, Ann Arbor, MI) using the
method previously reported.4
An tiin flam m ator y Assay. The test compounds were evalu-
ated using the in vivo rat carrageenan-induced foot paw edema
model reported previously.20
An a lgesic Assa y. Analgesic activity was determined using
the 4% sodium chloride-induced writhing (abdominal constric-
tion) assay as described previously.21
(18) Talley, J . J . Selective inhibitors of cyclooxygenase-2 (COX-2).
Prog. Med. Chem. 1999, 36, 201-234.
Ack n ow led gm en t. We are grateful to the Canadian
Institutes of Health Research (Grant MOP-14712) for
financial support of this research and to C.-A. McEwen
for technical assistance in performing the antiinflam-
matory and analgesic assays.
(19) Chan, C.-C.; et al. Rofecoxib [Vioxx, MK-0966; 4-(4′-Methylsul-
fonylphenyl)-3-phenyl-2-(5H)-furanone]: A potent and orally
active cyclooxygenase-2 inhibitor. Pharmacological and biochemi-
cal profiles. J . Pharmacol. Exp. Ther. 1999, 290, 551-560.
(20) Kumar, P.; Knaus, E. E. Synthesis and antiinflammatory activity
of N-substituted-dihydropyridylacetic acids, esters and amides.
Drug Des. Delivery 1987, 2, 145-149.
Su p p or tin g In for m a tion Ava ila ble: Experimental pro-
cedures for the preparation of compounds 7-13 (Scheme 1)
and 15-17 (Scheme 2) and their IR and NMR (1H, 13C, 19F)
spectroscopic data. This material is available free of charge
on the Internet at http://pubs.acs.org.
(21) Buolamwini, J . K.; Knaus, E. E. Synthesis and antinociceptive
activity of 4-pyridyl and -dihydropyridyl analogues of meperidine
and ketobemidone. Drug Des. Delivery 1990, 7, 19-31.
J M010153C