Discovery of hydroxamic acid analogs as dual inhibitors of phosphodiesterase-1 and -5

Article abstract of DOI:10.1016/j.bmcl.2005.06.016

HTS and the following synthesis of a series of the compounds led us to the discovery of hydroxamic acid analogs as potent dual inhibitors of phosphodiesterase (PDE)-1 and 5. These compounds have highly related structure and deviation of the structure usua

Full text of DOI:10.1016/j.bmcl.2005.06.016

Bioorganic & Medicinal Chemistry Letters 15 (2005) 4085–4090
Discovery of hydroxamic acid analogs as dual inhibitors
of phosphodiesterase-1 and -5
Akihito Dan, Takaaki Shiyama, Kazuto Yamazaki,* Naoto Kusunose, Katsuya Fujita,
Hideshi Sato, Kazutaka Matsui and Masafumi Kitano
Research Division, Sumitomo Pharmaceuticals Co., Ltd, 3-1-98 Kasugadenaka, Konohana-ku, Osaka 554-0022, Japan
Received 24 March 2005; revised 26 May 2005; accepted 3 June 2005
Available online 7 July 2005
Abstract—HTS and the following synthesis of a series of the compounds led us to the discovery of hydroxamic acid analogs as
potent dual inhibitors of phosphodiesterase (PDE)-1 and 5. These compounds have highly related structure and deviation of the
structure usually resulted in reduced potency. This result can be used to design other molecules that may be utilized for the therapy
of cardiovascular symptoms that relates to cGMP level.
ꢀ 2005 Elsevier Ltd. All rights reserved.
cGMP is one of the representative second messengers of
intracellular signal transduction and mediates various
cell functions. The signal transduction is based on the
phosphorylation of some target proteins that are cata-
lyzed by cGMP-dependent protein kinase. Phosphodies-
terase type 1 (PDE-1) and 5 (PDE-5) are the major
cGMP hydrolyzing enzymes in blood vessel,1 and regu-
Figure 1. The known dual inhibitor and structure 1 as initial lead.
late levels of the mediators with guanylyl cyclase which
catalyzes the synthesis of cGMP from GTP. PDE-1 and
-5 have also been identified in platelets.2 The inhibitors
of PDE-1 and -5 are expected as drug candidates for
cardiovascular diseases, such as hypertension, angina,
cardiac failure, and obstructive arteriosclerosis.
SCH51866 is known as a dual inhibitor of PDE-1 and
-5 (Fig. 1).3
1. Chemistry
We synthesized the 102 compounds that are listed in
Tables 1–3 by the synthetic routes shown in Schemes
1–3. Twenty-eight compounds are listed in a category
as Molecule I, and their chemical structures are repre-
sented in Table 1. Similarly, 27 compounds are listed
as Molecule II as shown in Table 2. And 47 compounds
in Table 3 are shown as Molecule III.
We found hydroxamic acid derivative 1 (Fig. 1), which
showed dual inhibition of PDE-1 and -5, by high
throughput screening of ca. 100,000 compounds which
were selected at random from our in-house library.
Compound 1 shows potent inhibition toward PDE-5
(IC50 = 0.013 lM), while it shows moderate inhibition
for PDE-1 (IC50 = 2.41 lM, about 200-fold less potent
than PDE-5).4 We started exploratory studies for potent
dual inhibitors with strong inhibition of both PDE-1
and -5.
1.1. Synthesis of molecule I (Scheme 1)
Reduction of 2,6-pyridinecarboxylic acid diethyl ester
(I-1) by NaBH4, chlorination by SOCl2, coupling with
phenol and saponification under basic condition gave
a carboxylic derivative I-4. A series of desirable carbox-
ylic acid and hydroxamic acid derivatives were synthe-
sized by amide formation reaction with I-4 and
various kinds of aminobenzoate, amino phenyl acetate,
amino thiazole, amino thiadiazole, and amino pyridine
acetate derivatives and so on.
Keyword: Phosphodiesterase.
*
Corresponding author. Tel.: +81664665194; fax: +81664665484;
e-mail: yamazaki@sumitomopharm.co.jp
0960-894X/$ - see front matter ꢀ 2005 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2005.06.016

4086
A. Dan et al. / Bioorg. Med. Chem. Lett. 15 (2005) 4085–4090
Table 1. Preliminary modification of B-part
Entry
X
n
R2
R3
R1
PDE1 IC50 (lM)
PDE5 IC50 (lM)
1
2
1
3
4
5
9
9
1
5
5
7
8
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
1
0
0
0
0
0
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
H
—
H
—
–CONHOH
–CONHOH
–CONHOH
–CONHOH
–CONHOH
–COOH
2.41
>10
>10
>10
>10
4.48
>10
>10
>10
0.92
4.52
3.90
0.20
>10
3.94
>10
5.70
>10
>10
1.36
>10
0.67
>10
0.63
>10
2.93
>10
>10
0.013
2.66
3.50
1.47
0.77
4.99
10.0
0.040
9.43
0.19
0.040
2.42
0.27
6.49
0.62
6.25
0.23
4.58
10.0
0.83
10.0
0.52
9.28
0.89
8.58
2.49
10.0
2.20
3
—
—
—
—
4
5
—
—
—
—
6
7
H
H
–COOH
–CONHOH
–COOH
8
H
H
9
H
H
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
H
H
–CONHOH
–CONHOH
–CONHOH
–CONHOH
–COOH
H
i-Pr
H
H
n-Bu
n-Bu
H
H
i-Bu
i-Bu
H
–CONHOH
–COOH
H
–(CH2)5CH3
–(CH2)5CH3
Bn
H
–CONHOH
–COOH
–COOH
H
H
Bn
H
–CONHOH
–COOH
–(CH2)2COOH
Me
–(CH2)2COOH
H
Me
–CONHOH
–COOH
–(CH2)2COOH
–(CH2)2–
–CONHOH
–COOH
–(CH2)2–
–(CH2)2O(CH2)2–
–(CH2)2O(CH2)2–
@O
–CONHOH
–COOH
–CONHOH
1.2. Synthesis of molecule II (Scheme 2)
2. Preliminary modification of B-part
Protection of commercially available ethyl 2-amino-1,3-
thiazole-4-carboxylate (II-1) with allyloxycarbonyl
group and saponification with NaOH yield II-2.5 In
the meantime, the resin linked hydroxylamine II-3 was
prepared from Sasrin resin by three steps.6 Coupling
reaction with II-2 and the resin linked hydroxylamine
II-3, and removal of the allyloxycarbonyl group gave
an O-immobilized amino thiazole derivative II-4.7 Sol-
id-phase combinatorial synthesis was performed by cou-
pling with II-4 and various aryl acid derivatives and the
following deprotection reaction gave a hydroxamic acid
library.8
Compound 1 can be retro-synthesized by condensation
of A-part and B-part (Fig. 1). First, we carried out
preliminary study on the effect of B-part using a corre-
sponding free acid of A-part (Table 1). Compounds
1–28 were obtained by amide formation with various
building blocks having amino group. Seven aromatic
moieties were selected for X as shown in Table 1 (X1,
X3–X5, and X7–X9). The number of Carbon linkers be-
tween X and carboxylic acid derivatives represented as n
was selected to be 0 or 1, and free carboxylic acid and
hydroxamic acid were synthesized.
Compounds 2–6 (n = 0) showed reduced (10- to 100-fold
less) inhibition of PDE-5 compared to compound 1.
Compound 6 with thiazole (X9) and carboxylic acid
showed weak but similar level of inhibition of PDE-1
and 5. Compounds 7–11, with unsubstituted methylene
linkers (n = 1), showed more potent inhibition than
compounds 2–6 (n = 0). Especially, compounds 8 and
11 showed comparable PDE-5 inhibition with com-
pound 1, but in the case of compound 1, only weaker
PDE-1 inhibition compared to PDE-5 (100-fold weaker)
was observed. On the other hand, compound 10, with
1.3. Synthesis of molecule III (Scheme 3)
Freidel–Crafts acylation of 4-ethylresorcinol (III-1) with
ZnCl2, condensation with III-3 and saponification gave
a carboxylic acid derivative III-4. A series of desirable
carboxylic acid, ester, and hydroxamic acid derivatives
were synthesized by amide formation reaction with the
III-4 and various kinds of amino acid, amino pyridine
acetate, amino thiazole, and amino thiadiazole deriva-
tives and so on.

A. Dan et al. / Bioorg. Med. Chem. Lett. 15 (2005) 4085–4090
4087
Table 2. Modification of A-part
potent inhibition profile similar to compounds 10 and
13. Generally, conversion of carboxylic acid to hydroxa-
mic acid increased inhibitory activities.
Entry
R
PDE1 IC50 PDE5 IC50
(lM)
3. Modification of A-part
(lM)
Having compound 10, weak dual inhibitor of PDE-1
and -5, we carried out modification of A-part (Table
2). Appropriate carboxylic acid derivatives were selected
as starting material from commercially available re-
agents or in-house carboxylic acid library.
29
30
31
3-(CF3)Bn
1-Phenylpropyl
1-Naphthylamino
>10
>10
>10
2.07
6.19
3.81
32
33
>10
>10
5.99
3.16
Compounds 29–39 have alkyl group or substituted alkyl
group as R group, but no potent PDE-5 inhibition was
found in these compounds, and several compounds did
not show PDE-1 inhibition. Substituted benzoic acid
derivatives are used as starting materials to synthesize
compounds 40–48. Only compound 43 among them
showed similar to slightly weaker inhibition than com-
pound 10, but the other compounds did not show satis-
factory inhibition.
34
2.80
1.28
Compounds 49–51, to which pyridine moieties were
introduced, showed very strong PDE-1 and -5 inhibi-
tion, especially compound 51 showed 10-fold stronger
inhibition than compound 10. Two-membered rings
were introduced to compounds 52–55 and these com-
pounds showed moderately strong PDE-5 inhibition.
Compound 55 showed comparable inhibition of
PDE-1 and -5 with compound 10.
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
Phenoxymethyl
4-Chlorophenoxymethyl
–(CH2)2COPh
–(CH2)5Ph
>10
1.81
>10
4.50
>10
>10
>10
>10
1.72
1.72
>10
>10
3.87
>10
7.10
5.62
1.69
4.52
1.90
1.80
3.25
7.56
5.47
0.43
3.41
1.83
4.25
8.43
2.58
3.20
0.32
–(CH2)10CH3
Ph
2-(F)Ph
3-(Me2N)Ph
3-(CF3S)Ph
3-(Bz)Ph
3,5-(Me)Ph
2,4-(Cl)Ph
4. Modification of B-part
2-(Br)-3-(Me)Ph
2,3,5-(MeO)Ph
2-Pyridyl
After discovering compound 51, we once again fixed the
A-part to a substructure of compound 51, and screened
the optimized structure for B-part. The effect of nitrogen
atom of the pyridine ring in A-part was also examined.
X represents either one of three aromatic rings shown in
Table 1 (X1, X2, and X7) or nothing (R1 directly linked
to nitrogen). PDE-5 inhibition was evaluated first and
PDE-1 inhibition was evaluated only when considerable
PDE-5 inhibition was observed (Table 3).
4-Chlorophenoxy-3-pyridyl 2.67
51
0.087
0.026
52
53
54
55
1-Naphthyl
2-Quinolyl
1.00
8.00
>10
0.38
1.10
0.14
1.84
0.79
Compounds 56–85 do not have a ring structure
(X = none). Compounds 56–61 have benzene ring
(Y1 = C, Y2 = C) and acetic acid moieties as R1. The ef-
fect of the nitrogen atom in the aromatic ring (Y1, Y2)
was generally small (Table 3). Compounds 63 and 66
with hydroxamic acid showed more potent inhibition
to PDE-5 compared to corresponding carboxylic acid
or esters. In compounds 64–74, various substituted alkyl
group were introduced to a-position of terminal acid
carbonyl group. Compounds 75–81 showed that intro-
duction of further carbon atom between terminal acid
moieties and nitrogen atom in R1 reduced PDE-5 inhibi-
tion. In compounds 82–85, various functional groups
were introduced to replace free acid and hydroxamic
acid, but no stronger inhibition was observed.
2-Quinoxalyl
1-(Ph)-4-Quinolyl
unsubstituted thiazole (X7), showed considerably potent
inhibition (less than 1 lM) of both PDE-1 and 5.
Further modifications based on the structure of com-
pound 10 was carried out. Monoalkyl and dialkyl sub-
stitution was introduced to the methylene carbon of
compound 10 (compounds 12–22 and 23–28, respective-
ly). Both free carboxylic acid and hydroxyamic acid
were synthesized. In mono-substituted compounds 12–
21, PDE-5 inhibitory activity remained. PDE-1 inhibi-
tion was generally weak but compound 13 with n-butyl
group showed considerably potent PDE-1 and -5 inhibi-
tion. Among di-substituted compounds 22–28, dimethyl
derivative (22) and cyclopentyl derivative (24) showed
Compounds 86–89 have 1,3-substituted benzene as
a part shown as X. These compounds showed strong

4088
A. Dan et al. / Bioorg. Med. Chem. Lett. 15 (2005) 4085–4090
Table 3. Modification of B-part
Entry
Y1
Y2
X
R1
R2
PDE1 IC50 (lM)
PDE5 IC50 (lM)
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
C
C
C
C
C
C
N
N
N
N
N
N
N
N
N
N
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
–CH2COOH
–CH(Me)COOH
–CH(Me)COOMe
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
0.92
0.13
2.13
0.68
8.76
0.65
1.70
0.50
1.03
0.46
0.35
1.24
8.62
2.93
6.68
3.28
–CH(CH2OH)COOH
–CH(CH2COOH)COOH
–CH(CH2COOEt)COOEt
–CH2COOH
–CH2CONHOH
(R)–CH(Me)COOH
(S)–CH(Me)COOH
–CH(Me)CONHOH
–CH(iPr)COOH
–CH(CH2OH)COOBn
–CH(CH2COOH)COOH
–CH(CH2OCOEt)COOH
–CH(CH2OCO(CH2)6CH3)COOH
72
73
74
N
N
N
C
C
C
—
—
—
H
H
H
—
—
—
3.68
3.35
9.07
75
76
77
78
79
80
81
82
N
N
N
N
N
N
N
N
C
C
C
C
C
C
C
C
—
—
—
—
—
—
—
—
–CH(OH)CH2COOMe
–CH(CH2OH)CH2COOH
–CH(CH2OH)CH2COOMe
–CH(Ph)CH2COOH
–(CH2)2COOH
H
—
—
—
—
—
—
—
—
6.07
9.46
3.10
1.45
9.87
7.65
2.64
8.91
H
H
H
–CH2CH@CH2
–(CH2)5COOH
–CH(CH2OH)CH2CH2COOMe
–C(CH2OH)3
H
H
H
83
84
85
N
N
N
C
C
C
—
—
—
H
H
H
—
—
—
3.19
7.82
5.15
86
87
88
89
90
91
92
93
94
95
96
N
N
N
N
N
N
C
C
C
C
C
C
C
C
C
N
C
C
1
1
1
1
2
2
7
7
7
7
7
–COOH
H
H
H
H
H
H
H
H
H
H
H
—
0.16
0.028
0.27
0.27
3.19
6.09
0.63
8.90
1.51
0.074
0.096
–CONHOH
–CH2COOH
–OCH2COOH
–OCH2COOH
–OCH2COOEt
–CH2COOH
–CH2COOEt
–CH2COOH
–CONHOH
–CH2COOH
0.39
—
—
—
—
—
C
—
—
N
N
N
0.043
2.00

A. Dan et al. / Bioorg. Med. Chem. Lett. 15 (2005) 4085–4090
4089
Table 3 (continued)
Entry
Y1
Y2
X
R1
R2
PDE1 IC50 (lM)
PDE5 IC50 (lM)
97
98
N
N
N
N
N
N
C
C
C
C
C
C
7
7
7
7
7
7
–CH2CH2COOH
–CH2CH2CONHOH
–CH2COOEt
H
2.70
1.10
—
0.074
0.017
2.72
6.35
1.21
3.30
H
Me
99
100
101
102
–CH2CONH2
–CH2CONH2
–CH2CONH2
Me
Et
—
—
1-Methylpropyl
—
as X, showed reduced inhibition of PDE-5 compared
to the case of 1,3-substituted benzene.
Compounds 92–102 have thiazole moiety as X similar to
compound 51. Compounds 92, 94, and 96 showed the
effect of different aromatic ring in A-part. The order
of PDE-5 inhibition was Pyridine (Y1 = N, Y2 = C) >
benzene (Y1 = Y2 = C) > pyradine (Y1 = Y2 = N).
Hydroxamic acid derivatives were more potent than
their carboxylic acid equivalent (compound 92 and 93,
96 and 99). Compounds 95–98 showed comparable po-
tent inhibition of PDE-5, and compound 95 showed
more potent PDE-1 inhibition than compound 51.
Scheme 1. Reagent and conditions: (i) NaBH4,EtOH, 50 ꢁC, 4 h; (ii)
SOCl2, CHCl3, rt, 1 h; (iii) phenol, K2CO3, DMF, rt, 16 h; (iv) LiOH,
THF, rt, 1 h, then Dowex50W-X8(H+).
5. Vasodilatory effects
Both compounds 51 and 10 showed relatively good met-
abolic stability. Therefore, we studied the vasodilatory
effects in isolated vessel of rat9 (Table 4). Compound
51, which was one of the potent dual inhibitors, showed
the expected potent vasodilatory effects (EC50
=
0.9 lM). On the other hand, compound 10 is about
10-fold less potent dual inhibitor than compound 51.
In a reflection of the effects for PDE-1 and -5, the
vasodilatory effects of compound 10 was weaker than
compound 51 (EC50 = 3.9 lM).
Scheme 2. Reagent and conditions: (i) Alloc-Cl, pyridine, rt, 15 h; (ii)
NaOH, MeOH-THF, rt, 15 h, then Dowex50W-X8(H+); (iii)
MeSO2Cl, DIEA, CH2Cl2, rt, 1 h; (iv) N-hydroxyphtalimide, Cs2CO3,
NMP, 80 C, 16 h; (v) anhydrous hydrazine, EtOH, rt, 20 h; (vi) DIC,
HOBt, DMF, rt, 16 h; (vii) Pd(PPh3)4, morphorine, THF, rt, 18 h.
The level of cGMP in the conditions of compound 51 rose
3.4 times (10 lM), but the level of cAMP had almost no
variation (1.5 times at 10 lM). It was suggested that the
vasodilatory effects of compound 51 was not dependent
on the level of cAMP regulated mainly by PDE-3, but
cGMP was regulated by mainly PDE-1 and -5. In addi-
tion, selectivities for other PDEs, which are considered
to have no effects on vasodilation1, are not investigated.
In conclusion, we discovered several potent dual inhibi-
tors for PDE-1 and -5. It was noted that this potent inhi-
bition was found only in limited diversity of structure;
then we thought that it resulted from a highly specific
interaction between PDE and these inhibitors. Unfortu-
nately, bioavailability of these inhibitors are very low
Scheme 3. Reagent and conditions: (i) ZnCl2, AcOH, reflux; (ii)
phenol, K2CO3, DMF, rt, 16 h; (iii) LiOH, THF, rt, 1 h, then
Dowex50W-X8(H+).
Table 4. The vasodilatory effects in isolated vessel of rat of compound
51 and 10
Entry
IC50, lM
Vasodilatory effects EC50, lM
inhibition to PDE-5. Hydroxamic acid 87 showed com-
parable inhibition of PDE-5 to compound 51. But these
compounds showed weak inhibition of PDE-1. Com-
pounds 90 and 91, which have 1,4-substituted benzene
PDE-1 PDE-5 PDE-3
10
51
0.920
0.087
0.190
0.026
>10 3.9
1.3 0.9

4090
A. Dan et al. / Bioorg. Med. Chem. Lett. 15 (2005) 4085–4090
6. Richter, L. S.; Desai, M. C. Tetrahedron Lett. 1997, 38, 321.
7. Synthesis of II-4 for compound 51. To II-2 (7.268 g,
30.0 mmol) in DMF (150 ml) was added di-isopropylcar-
bodiimide (DIC; 2.35 ml, 15.0 mmol) at room temperature
and stirring was continued for 30 min. To the reaction
mixture, H2NO-resin II-3 (1.00 mmol/g; 6 g, 6 mmol) was
added at room temperature. After the reaction suspension
was stirred at room temperature for 16 h, the resulting resin
was drained and washed with MeOH, DMF, MeOH, and
CH2Cl2 (each 70 ml · 3) to give 2-allyloxycarbonylamino-4-
thiazolyl derivative bound to hydroxyamine resin (6.99 g,
0.817 mmol/g). To the suspension of this resulting resin in
THF (90 ml) and morpholine (9 ml) was added Pd(PPh3)4
(130 mg, 11 mmol) under nitrogen. The mixture was stirred
for 18 h at room temperature. After the reaction, the resin
was drained and washed with DMF, 0.5%Et2NCS2Na–
0.5% iPr2NEt in DMF, DMF, MeOH, CH2Cl2 (each
50 ml · 3) and dried to give desirable 2-amino-4-thiazolyl
derivative bound to hydroxyamine resin II-4 (6.31 g,
0.876 mmol/g).
due to their aqueous solubility. Therefore, they are not
suitable for oral drug candidates. However, it is thought
that new dual inhibitors for PDE-1 and -5 will be found
efficiently by ligand-based virtual screening using the
data of the structure–activity relationships of hydroxa-
mic acid analogs. We will report the results about this
study in another article at an early date.
References and notes
1. Rybalkin, S. D.; Yan, C.; Bornfeldt, K. E.; Beavo, J. A.
Circulat. Res. 2003, 93, 280.
2. Wallis, R. M.; Corbin, J. D.; Francis, S. H.; Ellis, P. Am. J.
Cardiol. 1999, 83, 3C.
3. Vemulapalli, S.; Watkins, R. W.; Chintala, M.; Davis, H.;
Ahn, H. S.; Fawzi, A.; Tulshian, D.; Chiu, P.; Chatterjee,
M.; Lin, C. C.; Sybertz, E. J. J. Cardiovas. Pharmacol. 1996,
28, 862.
8. Synthesis of compound 51. To a suspension of B-4 (200 mg,
0.175 mmol) in DMF (4 ml) were added C-4 (276 mg,
0.875 mmol), DIC (0.137 ml, 0.875 mmol) and HOBt
(118 mg, 0.875 mmol) at room temperature. After the
reaction suspension was agitated by 360ꢁ rotator for 18 h
at room temperature, the resin was filtered, washed
(MeOH, DMF-H2O (1:1), DMF, MeOH, and CH2Cl2:
3 · 5 ml each) and dried in vacuo. The resulting resin was
suspended in 50% TFA/CH2Cl2 (3 ml) and agitated by 360ꢁ
rotator for 3 h at room temperature, then filtered off. The
filtrate was concentrated by SpeedVac concentrator and
purified by Gilson HPLC system to afford compound 51
(20.5 mg) as a colorless solid. Mass spectrum m/z 471
(M+H)+.
4. Coste, H.; Grondin, P. Biochem. Pharm. 1995, 50, 1577.
5. Synthesis of II-2 for compound 51. To a stirred suspension
of II-1 (20 g, 107 mmol) in pyridine (300 ml) was added
chloroformic acid allyl ester (12.54 ml, 118 mmol) dropwise
at 0 ꢁC, and stirring was continued for 30 min at 0 ꢁC and
for 15 h at room temperature. After adding MeOH, the
reaction mixture was evaporated and purified by silica gel
chromatography (SiO2 350 g; CHCl3 ! CHCl3–MeOH
100:1) to give N-alloc protected thiazole derivative
(19.67 g, 67.8%). Then, to a stirred suspension of this
product (19.5 g, 72.1 mmol) in MeOH-THF (9:2; 220 ml)
was added 1 N NaOH aq (60 ml) at room temperature.
After the reaction mixture was stirred at room temperature
for 15 h, it was neutrized by Dowex 50W-X8 (H+) and
filtrated. The filtrate was concentrated to give II-2 (15.98 g,
91.4%) as a pale yellow solid.
9. Delpy, E.; le Monnier de Gouville, A. C. Br. J. Pharmacol.
1996, 118, 1377.