Synthesis, toxicological, and pharmacological assessment of some oximes and aldehyde condensation products of 4-hydroxycoumarin

Article abstract of DOI:10.1002/(SICI)1521-4184(19997)332:7<243::AID-ARDP243>3.0.CO;2-2

The synthesis of condensation products of 4-hydroxycoumarin and nitro- substituted aromatic aldehydes as well as oximes of drugs with anticoagulant activity is described. The acute toxicity of the compound was studied in mice by oral and intraperitoneal a

Full text of DOI:10.1002/(SICI)1521-4184(19997)332:7<243::AID-ARDP243>3.0.CO;2-2

Warfarin and Warfarin Derivatives
243
Synthesis, Toxicological, and Pharmacological Assessment of Some
Oximes and Aldehyde Condensation Products of 4-Hydroxycoumarin
a)
b)
Ilia Manolov * and Nicolay D. Danchev
a)Department of Industrial Pharmacy, Faculty of Pharmacy, 2 Dunav St, Sofia 1000, Bulgaria
b)Department of Pharmacology and Toxicology, Faculty of Pharmacy, 2 Dunav St., Sofia 1000, Bulgaria
Key Words: Warfarin and Warfarin derivatives; oximes and aldehyde condensation products; blood coagulation time
Summary
Table 1
The synthesis of condensation products of4-hydroxycoumarin and
nitro-substituted aromatic aldehydes as well as oximes of drugs
with anticoagulant activity is described. The acute toxicity of the
compounds was studied in mice by oral and intraperitoneal admini-
stration. A comparative pharmacological study of the in vivo
anticoagulant effects of Warfarin derivatives showed that the new
compounds have different anticoagulant activity. 4-Hydroxy-3-[1-
(4-chlorophenyl)-3-oxobutyl]-2H-1-benzopyran-2-one, oxime 3
showed anticoagulant effect similar to Warfarin and Coumachlor,
but its acute toxicity was higher than that of the reference drugs.
Introduction
4-Hydroxycoumarin derivatives possess anticoagulant and
rodenticidal activities. The drugs of this group exhibit some
side effects including the Warfarin-related purple toes syn-
[1]
drome . By chemical modification of the structure of War-
[2, 3]
[4, 5]
[6–9]
farin
, Acenocoumarol
, and Coumachlor
it is
possible to obtain compounds with biological activity and
toxicity comparable to that of Warfarin, but with lower tox-
icity and fewer side effects.
Results and Discussion
Chemistry
[10–12]
We synthesized Acenocoumarol
Warfarin
, Coumachlor, and
[12, 13]
and treated them with hydroxylamine hydro-
chloride at a molar ratio 1: 1 in ethanol to produce their
[14]
oximes
with general formula (Table 1)
ions, finally yielding ions with m/z 205 (100%), 162 (52%),
120 (77%), and 92 (46%). The base peak of the compound is
the resonance-stabilized cation of 4-(4-nitrophenyl)-3-buten-
2-one, oxime with m/z 205 (Scheme 1a).
The mass spectral fragmentation of 4-hydroxy-3-[1-(4-
chlorophenyl)-3-oxobutyl]-2H-1-benzopyran-2-one, oxime
3 is analogous to that of compound 1. The highest intensity
peak of the ion with m/z 194 was obtained by elimination of
a 4-hydroxycoumarin radical from M.
We propose two probable mass spectral fragmentation
pathways concerning compound 1. The possible fragmenta-
The mass spectral fragmentation of 4-hydroxy-3-(3-oxo-1-
phenylbutyl)-2H-1-benzopyran-2-one, oxime 2 is analogous
[15]
tion pathways are presented in Scheme 1a,b
. The first
phase of the fragmentation process of 4-hydroxy-3-[1-(4-ni-
trophenyl)-3-oxobutyl]-2H-1-benzopyran-2-one, oxime 1 in-
volves a rather unstable molecular ion-radical. This under-
to the one of 4-acetoxy-3-(3-oxo-1-phenylbutyl)- 2H-1-ben-
[16]
zopyran-2-one
. The base peak of the compound is the
goes further fragmentation resulting in a cascade of unstable resonance-stabilized cation-radical with m/z 120.
Arch. Pharm. Pharm. Med. Chem.
© WILEY-VCH Verlag GmbH, D-69451 Weinheim, 1999
0365-6233/99/0707/0243 $17.50 +.50/0

244
Manolov and Danchev
Scheme 1a
p-Nitrobenzaldehyde was added to 4-hydroxycoumarin in After cooling 3,3′-(p-nitrobenzylidene)-4,4′-epoxydicouma-
[17]
rin 8 separated out. The mass spectral fragmentation of p-ni-
trobenzylidene-[4H]-pyrano-bis-4-hydroxycoumarin 8 is
shown in Scheme 2. The base peak of the compound is the
resonance-stabilized cation with m/z 317.
glacial acetic acid in a molar ratio of 1:2
. During boiling
a crystalline substance separated out. It was 3,3′-(p-nitroben-
zylidene)-bis-4-hydroxycoumarin 5. Substances 6 and 7 are
obtained in a similar way. The structure of these compounds
was confirmed by mass spectral analysis. As a result of this
process an unstable molecular ion-radical is obtained. The
base peak of all these three aldehyde condensation products
is the resonance-stabilized cation-radical with m/z 120.
Pharmacology
The compounds investigated were tested for acute toxicity
on mice by oral and intraperitoneal administration. An assess-
ment of the compounds’ effect on blood coagulation time
3,3′-(p-Nitrobenzylidene)-bis-4-hydroxycoumarin was
added to acetic anhydride and the solution was boiled
[18, 19]
.
Arch. Pharm. Pharm. Med. Chem. 332, 243–248 (1999)

Warfarin and Warfarin Derivatives
245
Scheme 1b
after a single day and a 3-day treatment was made. Warfarin cle-treated group. The highest effect was observed in the
group treated with compound 3, Coumachlor and Warfarin.
and Coumachlor were used as references compounds.
Experimental data concerning the acute toxicity (LD ) of
50
the derivatives showed that compounds 5 and 8 had low
Table 2. Acute toxicity of the compounds after intraperitoneal and per oral
administration on male mice.
toxicity(LD = 1115.4 and 1211.5 mg/kg body weight (b.w.)
50
——————————————————————————————
after intraperitoneal administration (Table 2), less than refer-
ences Warfarin and Coumachlor. Compounds 1, 2, 3, 4, 6,
and 7 exhibited higher toxicity than Warfarin, but 1 and 3
exhibited similar toxicity to Coumachlor after i.p. applica-
tion. After oral application compounds 5, 6, 7, and 8 were
Cmpd LD50 (mg/kg); i.p.
LD50 (mg/kg); p.o.
Index of
absorption (%)
——————————————————————————————
1
2
265.7 (228.0 – 309.7)a
349.1 (230.6 – 419.2)a
265.7 (228.0 – 309.7)a
403.6 (371.2 – 431.3)a
523.4 (456.7 – 599.8)a
588.4 (546.4 – 633.7)a
546.7 (505.3 – 591.1)a
1172.6 (682.1 –1634.2)
50.7
59.3
practically nontoxic (LD value > 2500 mg/kg b.w.). Com-
50
pounds 1, 2, and 3 were more toxic than Warfarin and
Coumachlor after p.o. administration. The highest value of
absorption (IR) showed compounds 2, 1, 3 and Warfarin –
approximately 50% (Table 2). The index of absorption of
Coumachlor was 17.1% (3 times less than Warfarin).
3
48.6
4
34.4
5
1115.4 (985.3 –1262.6)a > 2500a
<44.6
<16.7
<25.6
<48.4
6
418.0 (359.7 – 485.8)a
> 2500a
> 2500a
As shown by in vivo experiments after a single oral admini-
7
640.5 (541.9 – 756.9)
stration of the compounds at a dose of 10% of the LD , a
50
8
1211.5 (991.7 – 1480.2)a > 2500a
large increase in blood coagulation time was measured 24 h
after administration in the groups treated with compound 3,
Coumachlor, and Warfarin.
9
750.1 (581.9 – 966.6)
224.7 (197.4 – 256.3)
1468.6 (1212.7 – 1778.6) 51
10
1315.4 (1079.9 – 1601.1) 17.1
The data for the blood coagulation time after 3-day treat-
ment (once a day) with the same doses are shown in Table 4.
Compounds 1, 3, Coumachlor, and Warfarin increased the
blood coagulation time in comparison with the control vehi-
——————————————————————————————
Index of absorption was calculated as a LD50 i.p./LD50 p.o.
aStatistically significant differences in comparison with Warfarin
Arch. Pharm. Pharm. Med. Chem. 332, 243–248 (1999)

246
Manolov and Danchev
Scheme 2
Compound 5 decreased blood coagulation time statistically compounds need to be synthesized in order to elucidate the
significantly. The anticoagulant effect of Warfarin is con- structure-activity relationship.
nected to the change of the synthesis of inactive group of the
vitamin K-requiring clotting factors II, VII, IX, and X. The
Acknowledgments
effect of Warfarin proteins C and S explains the anticoagulant
[20]
effect of the drug
.
We thank Mrs. S. Samurova from the Department of Organic Chemistry,
Faculty of Chemistry, Sofia University, for performing the elemental analy-
ses, as well as Dr. Achim Blocher from the Laboratory of Radiopharmacy,
Eberhard-Karls University of Tübingen, Germany for mass spectra. We also
thank Ms L. Alexandrova, Mr. A. Belchev, and Mrs. R. Christova for their
technical assistance.
Our pharmacological and toxicological experiments on the
compounds show that compound 3 has a similar toxicity
compared to Coumachlor but a higher (3 times) toxicity than
Warfarin after i.p. administration. On the other handthe effect
of compound 3 on blood coagulation time after a single and
a 3-day administration in equitoxic doses is similar to that of
Warfarin. We suggest that 4-hydroxy-3-[1-(4-chlorophenyl)-
3-oxobutyl]-2H-1-benzopyran-2-one, oxime 3 acts as a
Experimental
whole molecule, because the value of LD after p.o. admini-
Chemistry
50
stration is three times less (more toxic) than Coumachlor and
Warfarin. These results indicate that compound 3 has pros-
pects for further pharmacological and toxicological experi-
ments. The investigations made with thesubstances described
in our work showed the following preliminary structure-ac-
tivity relationships:
Compounds 5 and 8 had equal toxicity, i.e. elimination of
a water molecule and formation of a pyran ring had no effect
upon toxicity of compound 8. The change of the position of
nitro group in the aromatic nucleus (o- and m-isomers) in-
creased the toxicity of the newly synthesized substances (6
and 7).
Melting points were measured on Boetius hot plate microscope (Germany)
and were uncorrected. IR spectra (nujol) were recorded on an IR-spectrome-
ter FTIR-8101M Shimadzu. 1H-NMR spectra were recorded at ambient
temperature on a Bruker 250 WM (250 MHz) spectrometer in [D6]-acetone,
CDCl3. Chemical shifts are given in ppm (σ) relative to TMS used as an
internal standard. Mass spectra were recorded on a Jeol JMS D 300 double
focusing mass spectrometer coupled to a JMA 2000 data system. The
compounds were introduced by direct inlet probe, heated from 50 °C to
400 °C at a rate of 100 °/min. The ionization current was 300 mA, the
accelerating voltage 3 kV and the chamber temperature 150 °C. TLC was
performedonprecoatedplates Kieselgel60F254 Merck (Germany) with layer
thickness 0.25 and UV detection (254 nm). Yields of TLC-homogeneous
isolated products are given. Elemental analysis was performed at the Faculty
of Chemistry, University of Sofia. Analyses indicated by the symbols of the
elements were within ±0.4% of the theoretical values.
Substance 2 had no anticoagulant activity but introduction
of a nitro group and especially a chlorine atom into the
aromatic nucleus significantly increased the anticoagulant
activity with no substantial effect upon acute toxicity. More
We synthesized Warfarin, Acenocoumarol, Coumachlor and treated them
with hydroxylamine hydrochloride at a molar ratio 1:1 to produce oxime
derivatives.
Arch. Pharm. Pharm. Med. Chem. 332, 243–248 (1999)

Warfarin and Warfarin Derivatives
247
Table 3. Effect of the compounds on the blood coagulation time in s after
single administration.
Table 4. Effect of the compounds on the blood coagulation time after 3 days’
administration.
——————————————————————————————
——————————————————————————————
Compound
Doses (mg/kg) p.o.
Blood coagulation
time (s)
Compound
Doses
Blood coagulation
time (s)
——————————————————————————————
——————————————————————————————
Control
–
203.5 ± 58.7
138.8 ± 45.7
322.5 ± 56.9a
243.8 ± 46.6
332.1 ± 42.8a
156.4 ± 22.3
56.4 ± 11.4a
221.6 ± 54.6
109.1 ± 36.2a
280.3 ± 31.9a
358.8 ± 65.8a
Control
–
156.3 ± 41.0
422.5 ± 61.2a, b
205.0 ± 41.9
980.1 ± 98.0a
242.1 ± 49.6a, b
73.5 ± 20.1a, b
149.9 ± 39.4b
146.6 ± 25.6b
251.1 ± 52.3
1082.8 ± 55.4a
950.2 ± 91.5a
1
2
50
1
2
50
50
50
3
50
3
50
4
120
250
250
250
250
147
130
4
120
250
250
250
250
147
130
5
5
6
6
7
7
8
8
9
9
10
10
——————————————————————————————
aStatistically significant differences in comparison with control vehicle
treated group.
——————————————————————————————
a
Statistically significant differences in comparison with control treated
b
group; Statistically significant differences in comparison with Warfarin
treated group.
General Procedure
Coumarin derivative (10 mmol) was added to 150 ml ethanol, containing
hydroxylamine hydrochloride (10 mmol). Three ml pyridine and 0.4 g so-
dium hydroxide were added to the reaction mixture. The solution was
allowed to stand for 12 h and then refluxed for 6 h. After cooling the oxime
crystallized out and was filtered off. The crude product was recrystallized
from methanol. TLC (toluene/chloroform/acetone, 8:8:1). [The yield, mp, Rf,
elemental analyses data, IR, 1H-NMR and MS: m/z (% of intensity) are given
under the respective heading.]
m (8H – arom), 10.9 s (2H – two hydroxyl groups). MS: 368 (0.3); 279 (4.0);
249 (4.0); 212 (12); 205 (42); 201 (36); 162 (67); 120 (100); 92 (94).
4-Hydroxycoumarin was treated with o-, m-, p-nitrobenzaldehyde in a
molar ratio 2:1 in glacial acetic acid for 30 min to 1 h at reflux. After cooling
the product was filtered and the precipitate was washed until the disappear-
ance of acetic acid odor. The crude product was recrystallized from acetone.
1
[The name, number, yield, mp, Rf, elemental analyses data, IR, H-NMR,
and MS: m/z (% of intensity) are given]:
4-Hydroxy-3-[1-(4-nitrophenyl)-3-oxobutyl]-2H-1-benzopyran-2-one
Oxime 1
3,3′-(4-Nitrobenzylidene)-bis-4-hydroxy-2H-1-benzopyran-2-one 5
2.58 g (70%), 191–193 °C, 0.15, C19H16N2O6 (368) (C, H, N). IR, cm–1
:
3.5 g (76%), 237–239 °C, 0.41, C25H15NO8 (457) (C, H, N), IR (nujol)
cm–1: 2953, 2924, 2855, 1653, 1616, 1597, 1564, 1491, 1377, 793. 1H-
NMR(CDCl3): 5.9 s (1H), 7.1–8.0 m (12H – arom), 8.5–9.0 s (1H), 11.0–11.2
s (1H). MS: 457 (1.2); 294 (12); 278 (36); 248 (48); 220 (7); 162 (34); 121
(27); 120 (100); 92 (80).
2924, 1673, 1620, 1568, 1497, 1452, 1399, 1215, 766. 1H-NMR ([D6]-ace-
tone): 1.84 s (3H), 3.1–3.3 d (2H – side chain), 4.9–5.1 t (1H), 7.2–8.1 m (8H
– arom), 10.9 s (2H – two hydroxyl groups). MS: 368 (0.3); 351 (3.2); 310
(0.6); 296 (0.6); 205 (100); 163 (6.4); 162 (52); 159 (58); 143 (20); 142 (15);
120 (77); 92 (46).
3,3′-(3-Nitrobenzylidene)-bis-4-hydroxy-2H-1-benzopyran-2-one 6
4-Hydroxy-3-(3-oxo-1-phenylbutyl)-2H-1-benzopyran-2-one, Oxime 2
3.7 g (81%), 228–230 °C, 0.37, C25H15NO8 (457) (C, H, N), IR (nujol)
cm–1: 2951, 2924, 2855, 1654, 1614, 1598, 1564, 1490, 1377, 740. 1H-NMR
(CDCl3): 5.8 s (1H), 7.1–8.1 m (12H – arom), 8.5–9.0 s (1H), 11.0 – 11.2 s
(1H). MS: 457 (1.1); 439 (0.4); 294 (13); 278 (21); 248 (18); 162 (60); 120
(100); 92 (61); 63 (10).
2.75 g (77%), 194–195 °C, 0.30, C19H17NO4 (323) (C, H, N), IR, cm–1
:
2855, 1672, 1620, 1568, 1499, 1453, 1399, 1215, 766, 700. 1H-NMR
([D6]-acetone): 1.84 s (3H), 2.4–2.6 d (2H – side chain), 3.1–3.5 t (1H),
7.1–8.1 m (9H – arom), 11.0 s (2H – two hydroxyl groups). MS: 323 (1.0);
306 (10.3); 289 (25); 275 (12); 162 (72); 121 (94); 120 (100); 92 (100).
3,3′-(2-Nitrobenzylidene)-bis-4-hydroxy-2H-1-benzopyran-2-one 7
4-Hydroxy-3-[1-(4-chlorophenyl)-3-oxobutyl]-2H-1-benzopyran-2-one
Oxime 3
3.40 g (75%), 207–209 °C, 0.34, C25H15NO8 (457) (C, H, N), IR (nujol)
cm–1: 2924, 2855, 1653, 1616, 1568, 1524, 1377, 1354, 1099, 762. 1H-NMR
(CDCl3): 6.0 s (1H), 7.1–8.2 m (12H – arom), 8.6–9.0 s (1H), 11.0–11.2 s
(1H). MS: 457 (0.3); 410 (13); 322 (8); 304 (53); 264 (86); 263 (70); 235
(51); 208 (6); 178 (14); 162 (41); 120 (100); 92 (68); 63 (31).
2.65 g (74%), 192–194 °C, 0.28, C19H16ClNO4 (357.5) (C, H, N, Cl). IR
cm–1: 2855, 1672, 1620, 1568, 1497, 1451, 1398, 1215, 766, 700. 1H-NMR
([D6]-acetone): 1.84 s (3H), 3.1-3.2 d (2H), 4.6–4.7 t (1H), 7.1–8.1 m (8H –
arom), 11.0 s (2H – two hydroxyl groups). MS: 357 (0.3); 340 (3.2); 324
(1.2); 299 (2.5); 285 (4.7); 259 (1.5); 248 (2.1); 246 (6.0); 194 (100); 162
(31); 120 (31).
3,3′-(4-Nitrobenzylidene)-4,4′-epoxydicoumarin 8
3,3′-(4-Nitrobenzylidene)-bis-4-hydroxy-2H-1-benzopyran-2-one
(10 mmol, 4.57 g) was added to 25 ml acetic acid anhydride and the reaction
mixture was refluxed for 30 min. The hot reaction mixture was poured into
300 ml distilled water to destroy the excess of acetic acid anhydride. The
product was precipitated, filtered and washed until the disappearance of
acetic acid odor. The crude product was recrystallized from acetone. Yield
3.88 g (85%), m.p. 338–342 °C, TLC (toluene/chloroform/acetone, 8:8:1),
4-Hydroxy-3-[1-(3-nitrophenyl)-3-oxobutyl]-2H-1-benzopyran-2-one
Oxime 4
3.1 g (84%), 188–190 °C, 0.17, C19H16N2O6 (368) (C, H, N), IR cm–1
:
3358, 3195, 2953, 2924, 2855, 1674, 1616, 1570, 1524, 1458, 1377, 760, 740.
1H-NMR ([D6]-acetone): 1.9 s (3H), 3.3–3.4 d (2H), 4.8–4.9 t (1H), 7.2–8.2
Arch. Pharm. Pharm. Med. Chem. 332, 243–248 (1999)

248
Manolov and Danchev
Rf 0.28. Anal. C25H13NO7(439) (C, H, N). IR (nujol) cm–1: 2953, 2924, 2855,
1728, 1713, 1669, 1610, 1510, 1456, 1368, 1348, 1180, 764. 1H-NMR
(CDCl3): 4.0 s (1H), 7.0–8.1 m (12H – arom). MS: m/z (% of intensity): 439
(16); 392 (11); 318 (20); 317 (100); 289 (4); 189 (4.0); 132 (4.0); 92 (4.0).
[6] Geigy JRAG, Br. Pat. 701 111,1953 [Chem. Abstr. 1955, 49, P 2522fg].
[7] D. Starr, C. Disanto, US Pat 2 752 360, 1956 [Chem. Abstr. 1957, 51,
1293g].
[8] Farbenfabriken Bayer Akt Ges, Br. Pat. 804 125, 1958 [Chem. Abstr.
1959, 53, 8167cd].
4-Hydroxy-3-(3-oxo-1-phenylbutyl)-2H-1-benzopyran-2-one, Warfarin 9
Yield 83%, m.p. 159–161 °C [12]
.
[9] D. Kaninsky, R.Meltzer, US Pat 3 264 326, 1966 [Chem. Abstr. 1966,
65, 13 655e].
4-Hydroxy-3-[1-(4-chlorophenyl)-3-oxobutyl]-2H-1-benzopyran-2-one,
Coumachlor 10
[10] I. Manolov, I. Ivanov, S. Karagiosov, P. Vassilev, Bulg. Pat. 45469,
1988.
Yield 75%, m.p. 162–163 °C [16]
.
[11] I. Manolov, I. Ivanov, S. Alexandrova, S. Karagiosov, P. Vassilev,
Bulg.Pat. 48268, 1989.
Pharmacology
[12] I. Ivanov, I. Manolov, L. Alexandrova, Arch. Pharm. (Weinheim) 1990,
323, 521–522.
The experiments were conducted on 340 white male mice weighing
23–25 g. Acute toxicity (LD50) of water insoluble compounds was assessed
by dissolving in saline (0.9% NaCl) with 1–2 drops of Tween 80. After
dissolution they were administered to mice via oral and intraperitoneal
routes. The LD50 was evaluated for 4 or 5 different doses each on 6 animals
and calculated by the methodof Litchfield-Wilcoxon [21], using a Pravetz-8M
[13] I. Manolov, I. Ivanov, Bulg Pat 48021, 1989.
[14] The Merck Index 10th ed, Rahway, USA, 1983, 1441.
[15] P. B. Terentjev, Mass Spectrometry in organic chemistry, Vysczaya
Schoola, Moscow, 1979.
personal computer. The anticoagulant effect of the compounds was evaluated
[22]
likewise for in vivo experiments according to the method of Moravitz
.
[16] I. Manolov, N. D. Danchev, Eur. J. Med. Chem. 1995, 30, 531–535.
The compounds were administered orally in doses 10% and 5% of the LD50
and blood coagulation times were measured in seconds 24 h after a single
and 3-day applications.
[17] W.R. Sullivan, C. F. Huebner, M. A. Stahmann, K. P. Link, J. Am.
Chem. Soc. 1943, 65, 2288–2291.
[18] C. F. Huebner, W. R. Sullivan, M. A. Stahmann, K. P. Link, J. Am.
Chem. Soc. 1943, 65, 2292–2296.
References
[19] M. A. Stahmann, L. H. Graf, C. F. Huebner, S. Roseman, K. P. Link,
[1] J. Hirsh, Arch. Intern. Med. 1987, 147, 769–771.
J. Am. Chem. Soc. 1944, 66, 900–902.
[2] M.Ikawa, M. Stahmann, K. Link, J. Am. Chem. Soc. 1944, 66, 902–906.
[20] R. E. Beftigole, in: Textbook of Pharmacology (C. Smith, A. Reynard,
eds) WB Sannders Co, Philadelphia, PA, USA, 1992, 784–802.
[3] M. Stahmann, M. Ikawa, K. Link, US Pat 2 427 578 [Chem Abstr 1947,
42, P603h].
[21] J. T. Litchfield, F. Wilcoxon, J. Pharmacol. Exp. Ther. 1949, 96, 99.
[22] N. Tietz, Clinical Guide to Laboratory Tests, Mir, Moscow, 1983.
Received: February 23, 1999 [FP372]
[4] W. Stoll, F.Litwan, US Pat 2 648 682, 1953 [Chem Abstr 1955, 49, P
2522fg].
[5] The Merck Index 10th ed, Rahway, USA, 1983, 5.
Arch. Pharm. Pharm. Med. Chem. 332, 243–248 (1999)