Naphthalene derivatives: A new series of selective cyclooxygenase-2 inhibitors

Article abstract of DOI:10.1016/S0960-894X(01)00534-0

A new series of potent and selective cyclooxygenase-2 inhibitors have been prepared. Some of these compounds show good oral anti-inflammatory activity in rats.

Full text of DOI:10.1016/S0960-894X(01)00534-0

Bioorganic & Medicinal Chemistry Letters 11 (2001) 2687–2690
Naphthalene Derivatives: A New Series of Selective
Cyclooxygenase-2 Inhibitors
Joan Feixas, Juan-Miguel Jimenez,* Nuria Godessart, Carles Puig,
Lıdia Soca and Marıa I.Crespo
Almirall Prodesfarma, Research Center, Cardener 68-74, 08024 Barcelona, Spain
Received 8 March 2001; revised 23 July 2001; accepted 31 July 2001
Abstract—A new series of potent and selective cyclooxygenase-2 inhibitors have been prepared.Some of these compounds show
good oral anti-inflammatory activity in rats. # 2001 Elsevier Science Ltd.All rights reserved.
Chemistry
Introduction
Various laboratories have developed extensive libraries
of selective COX-2 inhibitors in recent years.1 Most of
the compounds fit into three main categories: (1) acidic
sulfonamides such as NS-398, L-745337, and flosulide,
(2) diarylheterocycles such as rofecoxib and celecoxib,
and (3) modifications of classical NSAIDs such as
zomepirac and indomethacin derivatives.
In order to confirm our hypothesis, variously sub-
stituted benzylnaphthalene derivatives were synthesized
using the sequences described below.
Derivatives 1 and 4 were prepared from the commer-
cially available tetralone III following the routes descri-
bed in Scheme 1.POCl promoted dehydration of the
3
alcohol IV, followed by a DDQ oxidation, afforded
compound 1.The same alcohol IV was used in a Vils-
meier reaction and oxidation sequence to produce the
aldehyde V which was converted into acid 4 by using a
four-step procedure (Scheme 1).
In this paper, we describe our efforts in developing a
new series of selective COX-2 inhibitors.This new series
fits into the third category, due to its structural similar-
ity to the classical NSAID indomethacin.Compounds
of type I (Fig.1) can yield selective inhibitors of COX-2
through modification at the 1- and 3-positions of indo-
methacin.2 From these findings, we considered that
COX-2 selectivity could be maintained or even improved
in a compound where the benzyl group and alkanoic acid
moiety were connected to a naphthalene ring (II) instead
of to the indole nucleus of indomethacin.
The key intermediates XI used in the synthesis of naph-
thylacetic acid derivatives were prepared as shown in
Scheme 2.Protection of the commercially available
acids VII as oxazolidines followed a two-step procedure,
formation of the amides VIII and cyclodehydration to
provide ketones IX.Grignard reaction of the ketones IX
Figure 1.
*Corresponding author. Tel.: +34-93-291-36-38; fax: +34-93-291-34-20; e-mail: jmjimenez@almirallprodesfarma.com
0960-894X/01/$ - see front matter # 2001 Elsevier Science Ltd.All rights reserved.
PII: S0960-894X(01)00534-0

2688
J. Feixas et al. / Bioorg. Med. Chem. Lett. 11 (2001) 2687–2690
with the appropriate benzyl magnesium halides, fol-
lowed by reduction of the resultant benzyl alcohols and
concomitant hydrolysis of the oxazolidines with HI/
AcOH, yielded acid X.Final cyclization with PPA
resulted in the desired intermediates XI.
The alkylation of the corresponding enolates of esters
XII with the appropriate alkyl halides yielded, after final
hydrolysis, compounds 5 to 23 (compound 2 arose from
the direct hydrolysis of the corresponding ester XII).
The naphthylacetic derivatives 2 and 5–23 were typically
prepared starting from the tetralones XI through a
Reformatsky reaction, dehydration and final DDQ
aromatization sequence outlined below (Scheme 3).
Alternatively, the condensation of the glyoxylic acid
hydrate with XI followed by Zn(Hg)/HCl reduction of
the carbonyl group in XIII and final DDQ oxidation led
to compound 3.
Results and Discussion
Indomethacin is a non-selective cyclooxygenase inhi-
bitor.In our whole blood assay, it has an IC
0.22 mM for COX-2 and 0.19 mM for COX-1, and causes
gastrointestinal ulceration in rats at a dose of 10 mg/kg
with a 100% incidence.3 To avoid this gastrointestinal
toxicity we pursued modifications of indomethacin that
of
50
Scheme 1. (a) ClPhCH2Cl, Mg, THF; (b) POCl3, benzene, reflux; (c) DDQ, benzene, reflux; (d) POCl3, DMF; (e) DDQ, benzene, reflux; (f) NaBH4,
MeOH; (g) SOCl2, CH2Cl2; (h) NaCN, NaI, (CH3)2CO; (i) KOH, EtOH, reflux.
Scheme 2. (a) ClCOOEt, Et3N, NH2C(CH3)2CH2OH, THF; (b) p-TsOH, Xylene, reflux; (c) R3PhCH2X, Mg, THF; (d) HI, AcOH reflux; (e) when
R1 or R2=CH3O; (i) K2CO3, Me2SO4, reflux; (ii) NaOH 2 N, EtOH, reflux; (iii) PPA, 70 ꢀC.When R or R2 are not CH3O: (i) NaOH 2 N, EtOH,
1
reflux; (ii) PPA, 70 ꢀC.
Scheme 3. (a) BrCH2COOEt, Zn, benzene, reflux; (b) POCl3, benzene, reflux; (c) DDQ, toluene, reflux; (d) LDA, R4-X; (e) NaOH 2 N, EtOH,
reflux; (f) (HO)2CHCOOH, NaOH, EtOH/H2O; (g) Zn(Hg), HCl; (h) DDQ, benzene, reflux.

J. Feixas et al. / Bioorg. Med. Chem. Lett. 11 (2001) 2687–2690
2689
would confer COX-2 selectivity.The first compound
synthesized was structure 1 lacking the acidic moiety.
This simplified structure was completely inactive at both
the COX-1 and COX-2 isoenzymes (Table 1).Introduc-
tion of an alkanoic side chain at the 2- and 3-positions
of the naphthalene ring also yielded inactive compounds
(3 and 4), but substitution at position 1 (structure 2)
gave a submicromolar inhibitor of COX-2 that was
essentially inactive at COX-1.This encouraging result
prompted us to evaluate 2 in the carrageenan paw
edema assay3 in which it gave good anti-inflammatory
activity (see Table 1).Analogues of 2 with different
substituents either on the benzyl ring or on the naph-
thalene ring did not have any activity against either
enzyme (data not shown).So, curiously, 2 was the only
2-naphthylacetic derivative that inhibited COX-2.It has
been reported that the whole blood assay, originally
developed by Patrignani et al.,4 should be the model of
choice to test COX inhibitors and selectivity,5 and for
that reason we routinely use it for screening purposes
with slight modifications.3 Nevertheless, because this
assay does not discriminate between COX-2 and PLA2
inhibitors, we performed an alternative assay using an
endothelial cell line6 in which arachidonic acid was
provided exogenously.For all the compounds tested, we
observed COX-2 inhibition in the latter assay that was
completely compatible with the results of the whole
blood assay, and thus discarded the involvement of
PLA2.
for COX-2.Compound 5 was more potent in COX-2
when compared with compound 2, but in addition, it
also dramatically improved its potency in COX-1.Ethyl
and dimethyl substituted compounds (7, 8) were weaker
at COX-2.The 2-naphthylpropionic acid derivative
5
exhibited excellent activity in the carrageenan assay.
Unfortunately, it also produced ulceration in rats with
an ED50=40 mg/kg when administered daily for 4 days.
This was probably due to its inhibition of the COX-1
isoenzyme (IC50=4.10 mM).
Taking compounds 5 and 6 as leads, we explored several
substitutions at the R1 position.Among these, we
introduced chlorine (9), ethyl (10), methylthio (11),
methylsulfone (12) (a group typical of some selective
COX-2 inhibitors) and hydrogen (13).All of these
groups resulted in IC50 values for COX-2 inhibition well
above 3 mM.In an attempt to mimic the methoxy sub-
stitution of naproxen, we moved the methoxy group
from position 7 to 6 of the naphthalene ring to provide
14, but again lost inhibitory activity at both enzymes.
In Table 2, we present the results of varying substitution
at the benzyl ring.For the chlorine substitution, it is
clear that the para position is the most preferred for
both potency in COX-2 and selectivity (5, 15 and 16).
Introduction of hydrogen instead of chlorine (17) led to
an improved potency at both enzymes but an overall
decrease in selectivity.
In an attempt to obtain good COX-2/COX-1 selectivity
whilst maintaining in vivo activity, we prepared com-
pounds 5–8 in which the acidic chain had different sub-
stituents such as methyl, dimethyl and ethyl.For ease of
synthesis, we utilized the 4-fluoro substituted benzyl (6)
as a standard, even though it was slightly less selective
We evaluated the effect of para-alkyl substitution.In the
case of the methyl group (18) we saw an improvement in
both COX-2 and COX-1 potency, and in selectivity for
COX-2.The sterically demanding t-butyl group (20)
produced a significant loss in both COX-1 and COX-2
activity.The ethyl group ( 19) prompted a loss in
Table 1. In vitro and in vivo results for 4-benzylnaphthylacetic acids
Compound
R
R1
R2
R3
COX-2a
COX-1a
Carrageenanb
% inhib.(mg/kg)
1
2
3
4
5
6
7
8
H
1-CH2COOH
2-CH2COOH
3-CH2COOH
CH3O
CH3O
CH3O
CH3O
CH3O
CH3O
CH3O
CH3O
Cl
CH3CH2
CH3S
CH3SO2
H
H
H
H
H
H
H
H
H
H
H
H
H
H
CH3O
Cl
Cl
Cl
Cl
Cl
F
50.4Æ15.2
0.88Æ0.50
10.5Æ0.50
11.0Æ1.40
0.52Æ0.32
0.74Æ0.34
2.00Æ1.20
6.38Æ1.45
11.0Æ2.1
>100
>100
>100
IN (30)
52% (30)*
IN (30)
>100
IN (30)
1-CH(CH3)COOH
1-CH(CH3)COOH
1-CH(CH3CH2)COOH
1-C(CH3)2COOH
1-CH(CH3)COOH
1-CH(CH3)COOH
1-CH(CH3)COOH
1-CH(CH3)COOH
1-CH(CH3)COOH
1-CH(CH3)COOH
4.10Æ2.16
1.70Æ0.57
7.30Æ0.25
35.80Æ3.14
>100
61.00Æ10.2
19.90Æ0.35
>100
50% (30)*
53% (30)*
IN (30)
IN (30)
IN (30)
IN (30)
16% (30)
IN (30)
16% (30)
IN (30)
F
F
9
Cl
Cl
Cl
Cl
F
10
11
12
13
14
8.20Æ0.79
4.20Æ0.87
4.90Æ0.62
3.50Æ0.14
11.40Æ1.18
2.40Æ0.36
H
F
>100
*p<0.05.
aData are indicated as IC50 (mM)ÆSEM (n=3).
b6–7 animals per group.

2690
J. Feixas et al. / Bioorg. Med. Chem. Lett. 11 (2001) 2687–2690
Table 2. In vitro and in vivo results of benzyl-monosubstituted naphthalenes
Compound
R3
COX-2a
COX-1a
Selectivity
Carrageenanb
% inhib (mg/kg)
5
6
15
16
17
18
19
20
21
22
23
4-Cl
4-F
3-Cl
2-Cl
H
4-Me
4-Et
4-t-Bu
4-Ac
4-MeO
4-EtO
—
0.52Æ0.32
0.74Æ0.34
14.70Æ1.26
1.70Æ0.38
0.39Æ0.05
0.22Æ0.04
1.03Æ0.01
7.50Æ1.23
6.60Æ4.15
0.33Æ0.12
1.70Æ0.53
1.10Æ0.20
0.76Æ0.33
4.10Æ2.16
1.70Æ0.57
N.D.
2.30Æ0.06
1.50Æ0.45
2.80Æ2.13
>100
7.9
2.3
—
1.4
3.8
12.7
>97
5.4
33.0
26.6
>59.0
12.9
15.0
35 (3)*, 50 (30)*
35 (3)*, 53 (30)*
0 (3), 0 (30)
5 (3), 13 (30)
30 (3)*, 58 (30)*
30 (3)*, 49 (30)*
32 (30)*
IN (3), 32 (30)*
IN (3), IN (30)
22 (3), 45 (30)*
27 (3)*, 52 (30)*
41 (3)*, 40 (30)*
41 (3)*, 34 (30)*
40.8Æ4.30
218Æ6
8.80Æ2.05
>100
Celecoxib
Rofecoxib
14.2Æ4.40
—
11.4Æ0.81
*p<0.05.
aData are indicated as IC50 (mM)ÆSEM (n=3).
b6–7 animals per group.
References and Notes
potency, particularly at COX-1, which resulted in a
selectivity of around 100-fold.Unfortunately, this
interesting result did not translate into good in vivo
activity, producing only a 32% inhibition at 30 mg/kg in
the carrageenan model.
1.(a) Kalgutkar, A.S. Exp. Opin. Ther. Patents 1999, 9, 7.(b)
Jimenez, J. M.; Crespo, M. I.; Godessart, N. IDrugs 2000, 3,
907.
2. Black, W. C.; Bayly, C.; Belley, M.; Chan, C. C.; Charle-
son, S.; Denis, D.; Gauthier, J. Y.; Gordon, R.; Guay, D.;
Dargman, S.; Lau, C. K.; Leblanc, Y.; Mancini, J.; Ouellet, M.;
Percival, D.; Roy, P.; Skorey, K.; Tagari, P.; Vickers, P.; Wong,
E.; Xu, L.; Prasit, P. Bioorg. Med. Chem. Lett. 1996, 6, 725.
3. Puig, C.; Crespo, M. I.; Godessart, N.; Feixas, J.; Ibarzo,
J.; Jimenez, J. M.; Soca, L.; Cardelus, I.; Heredia, A.; Mir-
alpeix, M.; Puig, J.; Beleta, J.; Huerta, J. M.; Lopez, M.;
Segarra, V.; Ryder, H.; Palacios, J. M. J. Med. Chem. 2000,
43, 214.
The effect of 4-acetyl substitution (21) was loss of
activity at COX-2, whereas 4-MeO substitution (22)
produced good inhibition at this enzyme, giving rise to a
26-fold selectivity.This ratio could be increased to >59-
fold with the homologous 4-ethoxy substitution but
with some loss of COX-2 activity.It can be deduced
from these examples that selectivity can be profoundly
affected by substitution at the para position of the ben-
zyl ring and that the more potent and selective struc-
tures come from smallish groups such as methyl and
methoxy, giving rise to very active and selective mole-
cules which elicit promising in vivo activity.
4. Patrignani, P.; Panara, M. R.; Greco, A.; Fusco, O.;
Natoli, C.; Iacobelli, S.; Chipollone, F.; Ganci, A.; Creminon,
C.; Maclouf, J.; Patrono, C. J. Pharmacol. Exp. Ther. 1994,
271, 1705.
5. Brooks, P.; Emery, P.; Evans, J. F.; Fenner, H.; Hawkey,
C. J.; Patrono, C.; Smolen, J.; Breedveld, F.; Day, R.; Dou-
gados, M.; Ehrich, E. W.; Gijon-Banos, J.; Kvien, T. K.; Van
Rijswijk, M. H.; Warner, T.; Zeidler, H. Rheumatology 1999,
38, 779.
6. Miralpeix, M.; Camacho, M.; Lopez-Belmonte, J.; Cana-
lıas, F.; Beleta, J.; Palacios, J. M.; Vila, L. Br. J. Pharmacol.
1997, 121, 171.
In conclusion, we have identified a new series of selec-
tive COX-2 inhibitors that are active in an acute in vivo
model of inflammation.We have also gained an insight
into the structure–activity relationships that will allow us
to improve the properties of the series in the future.Fur-
ther work is underway and will be reported in due course.