New D-modified androstane derivatives as aromatase inhibitors

Article abstract of DOI:10.1016/S0039-128X(01)00096-4

Starting from a 16-oximino derivative of 5-androstene the newly-synthesized 16-oximino-17-hydroxy-17-substituted derivatives 2-4 gave by the Beckmann fragmentation reaction the corresponding D-seco derivatives 6-9. Besides, in the case of the 17-hydroxy-17-methyl-16-oximino derivative 2, as a result of the rearrangement, the hydrolysis product 5 of the 16-oximino group with the opposite configuration at the C-17 was obtained. By the Oppenauer oxidation and/or by dehydration of 7 with 2,3-dichloro-5,6-dicyanobenzoquinone (DDQ), the corresponding derivatives 12, 13, and 14 were obtained. The structures of 6 and 12 were unambiguously proved by the appropriate X-ray structural analysis. Kinetic analysis for anti-aromatase activity showed that compound 12 expressed the highest inhibition in the denucleated rat ovarian fraction in comparison to other androstene derivatives (IC₅₀ was 0.42 μM). In comparison to aminoglutethimide (AG) activity, it was 3.5 times lower. The inhibition was competitive, with K_i of 0.27 μM. Introduction of additional units of unsaturation (compounds 13 and 14) in D-seco derivatives did not increase anti-aromatase activity. Copyright

Full text of DOI:10.1016/S0039-128X(01)00096-4

Steroids 66 (2001) 645–653
New D-modified androstane derivatives as aromatase inhibitors૾
Katarina M. Penov Gasˇia,*, Slobodanka M. Stankovic´b, Ja´nos J. Csana´dia,
Evgenija A. Djurendic´a, Marija N. Sakacˇa, Ljubica Medic´ Mijacˇevic´a, Otto N. Arcsona,
Srdjan Z. Stojanovic´a, Silvana Andric´c, Dora Molnar Gabora, Radmila Kovacˇevic´c
aInstitute of Chemistry, Faculty of Sciences, University of Novi Sad, 21000 Novi Sad, Trg Dositeja Obradovic´a 3, Yugoslavia
bInstitute of Physics, Faculty of Sciences, University of Novi Sad, 21000 Novi Sad, Trg Dositeja Obradovic´a 4, Yugoslavia
cInstitute of Biology, Faculty of Sciences, University of Novi Sad, 21000 Novi Sad, Trg Dositeja Obradovic´a 2, Yugoslavia
Received 4 October 2000; received in revised form 27 December 2000; accepted 10 January 2001
Abstract
Starting from a 16-oximino derivative of 5-androstene the newly-synthesized 16-oximino-17-hydroxy-17-substituted derivatives 2–4
gave by the Beckmann fragmentation reaction the corresponding D-seco derivatives 6–9. Besides, in the case of the 17-hydroxy-17-methyl-
16-oximino derivative 2, as a result of the rearrangement, the hydrolysis product 5 of the 16-oximino group with the opposite configuration
at the C-17 was obtained. By the Oppenauer oxidation and/or by dehydration of 7 with 2,3-dichloro-5,6-dicyanobenzoquinone (DDQ), the
corresponding derivatives 12, 13, and 14 were obtained. The structures of 6 and 12 were unambiguously proved by the appropriate X-ray
structural analysis. Kinetic analysis for anti-aromatase activity showed that compound 12 expressed the highest inhibition in the denucleated
rat ovarian fraction in comparison to other androstene derivatives (IC50 was 0.42 ␮M). In comparison to aminoglutethimide (AG) activity,
it was 3.5 times lower. The inhibition was competitive, with Ki of 0.27 ␮M. Introduction of additional units of unsaturation (compounds
13 and 14) in D-seco derivatives did not increase anti-aromatase activity. © 2001 Elsevier Science Inc. All rights reserved.
Keywords: D-Secosteroids; Androgens synthesis; Inhibition of aromatase activity
1. Introduction
Potent steroid aromatase inhibitors are mainly the C-1,
C-2, C-4, C-6, C-19 substituted and C-2 and C-19 bridged
steroidal compounds, i.e. the A- and B-modified steroids,
whereas the D-modified steroids (with the exception of e.g.
testololactone) have hardly received any research attention.
With the aim of studying anti-aromatase activity of some
D-modified steroid compounds we synthesized several new
17␣-substituted-17␤-hydroxy-16-oximino derivatives of
5-androstene and the corresponding D-seco derivatives.
Having in mind the fact that extra units of unsaturation in
the A and/or B rings of 4-androstendione [7] and its deriv-
atives [8,9], lead to a change in the anti-aromatase activity,
we synthesized in the D-seco-methyl series the 4-ene-3-keto
derivative and its 1,4-diene and 1,4,6-triene analogs.
Aromatase is a cytochrome P450 enzyme which cata-
lyzes conversion of androgens into estrogens in the last step
of estrogen biosynthesis [1]. Compounds that inhibit aro-
matase have potential applications in the treatment of ad-
vanced estrogen-dependent tumors such as breast cancer,
endometrial cancer, prostatic hyperplasia, and prostate can-
cer. A number of steroid inhibitors may inactivate aro-
matase by diverse interactions with the enzyme. By design-
ing steroidal analogs with substituents at different positions
on the steroid nucleus, information has been obtained on the
structural requirements needed for favorable interactions
with the enzymatic sites [2–6].
Address reprint request to: Katarina M. Penov Gas˘i, Institute of Chem-
istry, Faculty of Sciences, University of Novi Sad, 21000 Novi Sad, Trg
Dositeja Obradovic´a 3, Yugoslavia.
2. Experimental
૾The work is originated from Institute of Chemistry, Faculty of Sci-
ences, University of Novi Sad, 21000 Novi Sad, Trg Dositeja Obradovic´a
3, Yugoslavia.
2.1. General
Melting points were determined using a Bu¨chi SMP appa-
ratus and the values were not corrected. IR spectra were re-
* Corresponding author. Tel.: ϩ381-21-350-122; fax: ϩ381-21-54-065.
E-mail address: sazvon@eunet.yu (K.M. Penov Gasˇi).
0039-128X/01/$ – see front matter © 2001 Elsevier Science Inc. All rights reserved.
PII: S0039-128X(01)00096-4

646
K.M. Penov Gasˇi et al. / Steroids 66 (2001) 645–653
corded on a Perkin-Elmer 457 and Perkin-Elmer FT-IR 1725
X spectrometers. NMR spectra were taken on a Bruker AC
250E spectrometer operating at 250 MHz (proton) and 62.9
MHz (carbon), using standard Bruker software; the tetrameth-
ylsilane peak (␦ 0.00) was used as reference in CDCl3 and
(0.113 g, 41%, m.p. 182–183°C). The side product was
17-ethyl-3␤-hydroxy-16-nitrile-16,17-seco-5-androstene-
17-one (8, 0.025 g, 9.7%, m.p. 170–174°C) in the form of
white crystals, which were recrystallized from diethyl ether/
petroleum ether (1:1) (0.016 g, 6.3%, m.p. 172–174°C).
Spectroscopic data for 3: IR ␯max (KBr) cmϪ1: 3300–3450,
1
DMSO-d6 for H NMR, whereas the central carbon line of
1
chloroform-d was set at 77.0 ppm, or that of DMSO-d6 at 39.5
ppm, for carbon-13 NMR. All reagents used were analytical
grade commercially available substances.
3250, 1630, 1060, 950. H NMR (DMSO-d6): 0.71 (3H, s,
H-18); 0.87 (3H, t, J ϭ 7.2 Hz, H-21); 0.96 (3H, s, H-19);
4.14 (1H, s, OH-17); 4.60 (1H, bs, OH-3); 5.27 (1H, d, J ϭ
4.5 Hz, H-6); 10.4 (1H, bs, N-OH). 13C NMR (DMSO-d6):
163.7 (C-16, C ϭ NOH); 141.3 (C-5); 120.3 (C-6); 82.2
(C-17, Cq-OH); 70.0 (C-3, CH-OH); 49.6; 46.0; 45.3; 42.2;
36.8; 36.2; 31.6; 31.4; 31.0; 30.3; 26.7; 25.7; 20.2; 19.1
(C-19); 14.4 (C-18); 7.9 (C-21). Anal. calcd. for
C21H33NO3: C, 72.58; H, 9.57; found: C, 72.18; H, 9.06.
3␤,17␤-Dihydroxy-17␣-phenyl-16-oximino-5-
androstene (4). Metallic lithium was forged to form thin
plates and cut into small pieces and then suspended in
absolute ether (72 ml) in the nitrogen atmosphere. Then,
with continuous stirring, freshly distilled bromobenzene
(14.2 ml, 132.3 mmol) was added dropwise in the course of
45 min. The reaction mixture was stirred for another 2 h at
room temperature. When phenyllithium was formed, the
reaction mixture was cooled to Ϫ10°C and 3␤-hydroxy-16-
oximino-5-androstene-17-one (1, 1.00 g, 3.15 mmol) in
absolute THF (35 ml) was added. The reaction mixture was
vigorously stirred in the nitrogen atmosphere at Ϫ10°C and
then poured into water (400 ml) and acidified with 6 M HCl
to pH 3. The organic and aqueous layers were separated and
the aqueous layer was extracted with dichloromethane (3 ϫ
50 ml). The joined dichloromethane extracts and organic
layer were dried and the solvent was evaporated. Crude
compound 4 was obtained in the form of brown-yellow oil
(2.06 g). After double chromatographic separation on a
silica gel column (200 g, toluene/ethyl acetate, 6:1, 3:1) the
compound 4 was obtained in the form of white crystals
(1.20 g, 96%), which was recrystallized from MeOH (1.14
2.2. Chemical synthesis
3␤,17␤-Dihydroxy-17␣-methyl-16-oximino-5andro-
stene (2). Magnesium (4.87 g, 200.3 mmol) was suspended
in dry ether (65 ml) in the nitrogen atmosphere, and methyl
iodide (28.1 ml, 451.2 mmol) was added. After forming of
methyl-magnesium iodide, the reaction mixture was cooled
to Ϫ10°C and compound 1 (1.00 g, 3.15 mmol) in dry THF
(40 ml) was added. The reaction mixture was stirred in a
nitrogen atmosphere at Ϫ10°C for 2 h, poured into water
(300 ml) and acidified with 6 M HCl to pH 3. The organic
layer was separated and dried over Na2SO4. The solvent
was removed to give a crude product 2 (0.90 g). Crystalli-
zation from MeOH afforded pure compound 2 (0.83 g, 69%,
m.p. 234–245°C) as white crystals. IR ␯
(KBr) cmϪ1
:
max
1
3500, 3200, 1470, 1450, 1380, 1185, 1125, 950, 740. H
NMR (DMSO-d6): 0.70 (3H, s, H-18); 0.96 (3H, s, H-19);
1.14 (3H, s, H-20); 3.24 (1H, m, H-3); 4.54 (1H, s, OH-17);
4.63 (1H, d, J ϭ 4.5 Hz, OH-3); 5.26 (1H, m, H-6); 10.34
(1H, s, N-OH). 13C NMR (DMSO-d6): 166.2 (C-16, C ϭ
NOH); 141.3 (C-5); 120.3 (C-6); 80.1 (C-17, Cq-OH); 70.1
(C-3, CH-OH); 49.7; 45.8; 45.2; 42.3; 36.9; 36.3; 31.5;
31.4; 31.1; 30.6; 26.7; 23.4; 20.2; 19.2; 13.9. Anal. calcd.
for C20H31NO3 ⅐ H2O: C, 68.34; H, 9.46; found: C, 68.28;
H, 9.11.
3␤,17␤-Dihydroxy-17␣-ethyl-16-oximino-5-
androstene (3). Magnesium (1.54 g, 63.1 mmol) was sus-
pended in absolute ether (20 ml) in the nitrogen atmosphere
and, with continuous stirring, ethyl iodide (9.80 ml, 157.7
mmol) was added dropwise in the course of 45 min. The
reaction mixture was stirred for 1 h at room temperature.
When ethyl-magnesium iodide formed, the reaction mixture
was cooled to Ϫ10°C and 3␤-hydroxy-16-oximino-5-
androstene-17-one (1, 0.25 g, 0.8 mmol) in absolute THF
(10 ml) was added. The reaction mixture was vigorously
stirred for 2 h in the nitrogen atmosphere at Ϫ10°C, then
poured to water (150 ml) and acidified with 6 M HCl to pH
3. The organic and aqueous layers were separated and the
aqueous layer was extracted with dichloromethane (3 ϫ 20
ml). The joined dichloromethane and organic layer were
dried and the solvent was evaporated, giving a yellow oil
(1.42 g). The twice-repeated chromatographic separation on
a silica gel column (140 g, toluene/ethyl acetate, 10:1, 6:1)
gave compound 3 (0.145 g, 53%, m.p. 180–182°C) as main
product in the form of white crystals, which were recrys-
tallized from the ethanol/petroleum ether (10:1) mixture
g, 91.5%, m.p. 161°C). IR ␯
(KBr) cmϪ1: 3350–3450,
max
1
2950, 1630, 1060, 950, 710. H NMR (DMSO-d6): 0.88
(3H, s, H-18); 0.93 (3H, s, H-19); 3.19 (1H, m, H-3); 4.57
(1H, d, J ϭ 4.5 Hz, OH-3); 5.23 (1H, m, H-6); 5.27 (1H,
s, OH-17); 7.11–7.25 (5H, Ph); 10.62 (1H, s, N-OH). 13C
NMR (DMSO-d6): 166.5 (C-16, C ϭ NOH); 144.2 (C-1Ј,
Ph); 141.2 (C-5); 126.8 and 127.8 (C-2Ј, C-3Ј, C-5Ј, C-6Ј
from Ph); 126.5 (C-4Ј, Ph); 120.0 (C-6); 85.1 (C-17, Cq-
OH); 69.9 (C-3, CH-OH); 49.4; 46.6; 44.7; 42.1; 36.6; 36.2;
36.0; 32.7; 31.3; 31.2; 28.4; 20.0; 19.1 (C-19); 14.4 (C-18).
Anal. calcd. for C25H33NO3: C, 75.91; H, 8.41; found: C,
75.64; H, 8.01.
3␤,17␣-Dihydroxy-16-keto-17␤-methyl-5-androstene
(5). 3␤,17␤-Dihydroxy-17␣-methyl-16-oximino-5-andro-
stene (2, 1.00 g, 3.0 mmol) was dissolved in EtOH (150 ml)
and a mixture of concentrated HCl and EtOH (1:1, 30 ml)
was added. To this solution, titanium(III) chloride (0.60 g,
12.5 mmol of titanium were heated with 6 M HCl (60 ml)
for 1 h) and hexane (30 ml) were added. The reaction

K.M. Penov Gasˇi et al. / Steroids 66 (2001) 645–653
647
mixture was left at room temperature for 12 h, and when the
reaction was completed it was diluted with water (200 ml)
and neutralized with saturated aqueous NaHCO3 to pH 7.
After extraction with ether, the organic layer was washed
with water and dried with anhydrous Na2SO4; removal of
the drying agent and ether resulted in crude product (0.88
g). Crystallization from MeOH afforded pure compound 5
(0.83 g, 87%, m.p. 272–275°C). IR ␯max (KBr) cmϪ1: 3350,
17-Ethyl-3␤-hydroxy-16-nitrile-16,17-seco-5-
androstene-17-one (8). 3␤,17␤-Dihydroxy-17␣-ethyl-16-
oximino-5-androstene (3, 0.060 g, 0.17 mmol) was dis-
solved in ethanol (9 ml). To this solution, the HCl/ethanol
mixture (1:1, 1.80 ml) and aqueous solution of titanium(III)
chloride (15%, 0.70 ml, 0.75 mmol) were added. The reac-
tion mixture was heated at the boiling temperature for 20
min. After completion of the reaction, the reaction mixture
was poured to water (20 ml), neutralized with saturated
solution of NaHCO3 to pH 7, and then extracted with
dichloromethane (3 ϫ 20 ml). The extract was dried out and
the solvent evaporated, which resulted in crude product 8
(0.055 g). Chromatographic purification on a silica gel col-
umn (5.60 g, toluene/ethyl acetate, 3:1) gave compound 8
(0.053 g, 94%) in the form of a colorless oil; its recrystal-
lization from the mixture diethyl ether/petroleum ether (1:1)
afforded white crystals (0.049 g, 86%, m.p. 172–174°C). IR
1
3225, 3030, 1729, 1450, 1195, 1095, 995, 740. H NMR
(DMSO-d6): 0.61 (3H, s, H-18); 0.97 (6H, s, H-19 and
H-20); 3.26 (1H, m, H-3); 4.64 (1H, d, J ϭ 4.5 Hz, OH-3);
5.20 (1H, s, OH-17); 5.27 (1H, m, H-6). 13C NMR (DMSO-
d6): 216.6 (C-16, C ϭ O); 141.4 (C-5); 120.2 (C-6); 78.9
(C-17, Cq-OH); 70.0 (C-3, CH-OH); 49.8; 44.3; 43.8; 42.2;
36.7; 36.3; 36.1; 31.7; 31.4; 30.7; 28.4; 19.6 (C-19); 19.2;
15.6; 15.0 (C-18). Anal. calcd. for C20H30O3: C, 75.43; H,
9.50; found: C, 75.81; H, 9.43.
3␤-Acetoxy-17-keto-17-methyl-16-nitrile-16,17-seco-
5-androstene (6). 3␤,17␤-Dihydroxy-17␣-methyl-16-
oximino-5-androstene (2, 1.00 g, 3.0 mmol) was dissolved
in absolute pyridine (26.6 ml) and acetic anhydride (13.3
ml) was added. Then, the reaction mixture was heated and
kept at 90–100°C with constant stirring for 2 h. After the
reaction completion, the reaction mixture was poured to
ice-cold water, whereby crude product 6 (0.88 g) was ob-
tained. Recrystallization from the n-hexane/acetone mixture
(5:2) afforded pure compound 6 (0.79 g, 89.5%. m.p. 142–
␯
(KBr) cmϪ1: 3350–3450, 2860–3000, 2260, 1710,
max
1
1480, 1070, 780. H NMR (DMSO-d6): 0.91 (3H, t, J ϭ
7.1 Hz, H-21); 0.94 (3H, s, H-19); 1.16 (3H, s, H-18); 2.57
(2H, m, H-20); 3.20–3.40 (1H, m, H-3); 4.62 (1H, d, J ϭ
4.3 Hz, OH-3); 5.31 (1H, m, H-6). 13C NMR (DMSO-d6):
215.2 (C-17, C ϭ O); 140.8 (C-5); 120.0 (C-16, CN); 119.7
(C-6); 69.9 (C-3, CH-OH); 51.0; 48.2; 41.8; 36.5; 36.2;
35.6; 32.1; 31.2; 31.0; 29.4; 19.1; 19.0 (C-19); 17.5; 14.4
(C-18); 8.0 (C-21). Anal. calcd. for C21H31NO2: C, 76.55;
H, 9.48; found: C, 76.61; H, 9.54.
143°C), as a colorless crystals. IR ␯
(KBr) cmϪ1: 2240,
3␤-Hydroxy-17-phenyl-16-nitrile-16,17-seco-5-
androstene-17-one (9). 3␤,17␤-Dihydroxy-17␣-phenyl-
16-oximino-5-androstene (4, 0.280 g, 0.71 mmol) was dis-
solved in ethanol (30 ml), the HCl/ethanol mixture (1:1, 6
ml) added, and then aqueous solution of titanium(III) chlo-
ride (15%, 2.8 ml, 3.1 mmol). The reaction mixture was
heated at the boiling temperature for 10 min. When the
reaction was completed, the reaction mixture was poured to
water (100 ml), neutralized with saturated solution of
NaHCO3 (pH 7), and extracted with dichloromethane (3 ϫ
20 ml). The extract was dried out and the solvent evapo-
rated, whereby crude product 9 (0.32 g) was obtained.
Chromatographic purification on a silica gel column (32 g,
toluene/ethyl acetate, 3:1) gave compound 9 (0.245 g, 92%).
Recrystallization from MeOH afforded compound 9 (0.23 g,
max
1
1730, 1705, 1245, 1030. H NMR (CDCl3): 1.05 (3H, s,
H-19); 1.32 (3H, s, H-18); 2.04 (3H, s, Ac); 2.18 (3H, s,
H-20); 4.61 (1H, m, H-3); 5.40 (1H, m, H-6). 13C NMR
(DMSO-d6): 212.9 (C-17, C ϭ O); 169.7 (C ϭ O, Ac);
139.0 (C-5); 121.3 (C-6); 120.0 (C-16, CN); 73.5; 51.8;
47.6; 41.9; 37.5; 36.1; 35.9; 32.2; 27.0; 25.3 (C-20); 21.8;
19.2; 19.0; 18.8 (C-19); 14.6 (C-18). Anal. calcd. for
C22H31NO3: C, 73.91; H, 8.74; found: C, 73.73; H, 8.60.
3␤-Hydroxy-17-keto-17-methyl-16-nitrile-16,17-seco-
5-androstene (7). 3␤-Acetoxy-17-keto-17-methyl-16-ni-
trile-16,17-seco-5-androstene (6, 1.00 g, 2.8 mmol) was
dissolved in the solution of sodium-etoxide in ethanol (0.1
M, 30 ml). The reaction mixture was stirred at 45°C for 30
min, and then poured into water (150 ml). After adjusting
the pH to 3 (6 M HCl), crude product 7 (0.87 g) was
obtained. Purification by column chromatography with sil-
ica gel (50 g, benzene/ethyl acetate, 1:2) afforded com-
pound 7 (0.824 g, 93%) in the form of yellowish oil. Crys-
tallization from the n-hexane/acetone mixture (5:2) gave
pure compound 7 (0.724 g, 82%, m.p. 125–128°C) in the
86%, m.p. 144°C) in the form of white crystals. IR ␯
max
(KBr) cmϪ1: 3350–3450, 2950, 2260, 1680, 1240, 1050,
1
650–800. H NMR (DMSO-d6): 0.96 (3H, s, H-19); 1.25
(3H, s, H-18); 3.30 (1H, m, H-3); 4.65 (1H, d, J ϭ 4.5 Hz,
OH-3); 5.33 (1H, m, H-6); 7.41–7.51 (5H, Ph). 13C NMR
(DMSO-d6): 209.0 (C-17, C ϭ O); 140.8 (C-5); 138.9
(C-1Ј, Ph); 130.6 (C-4Ј, Ph); 127.4 and 128.1 (C-2Ј, C-3Ј,
C-5Ј, C-6Ј from Ph); 120.2 (C-16, CN); 119.7 (C-6); 69.9
(C-3, CH-OH); 51.4; 48.0; 42.4; 41.8; 36.5; 36.3; 36.2;
32.4; 31.2; 31.0; 19.1; 19.0 (C-19); 17.9; 15.9 (C-18). Anal.
calcd. for C25H31NO2: C, 79.54; H, 8.28; found: C, 79.17;
H, 7.89.
form of white crystals. IR ␯
(KBr) cmϪ1: 3500, 2260,
max
1
1705. H NMR (CDCl3): 1.03 (3H, s, H-19); 1.31 (3H, s,
H-18); 2.17 (3H, s, H-20); 3.53 (1H, m, H-3); 5.36 (1H, m,
H-6). 13C NMR (CDCl3): 213.4 (C-17, C ϭ O); 140.3
(C-5); 120.4 (C-6); 119.0 (C-16, CN); 71.4 (C-3, CH-OH);
51.6; 48.7; 41.9; 36.9; 36.6; 36.4; 31.9; 31.4; 25.3 (C-20);
19.6; 19.2 (C-19); 17.9; 15.6 (C-18). Anal. calcd. for
C20H29NO2: C, 76.15; H, 9.27; found: C, 76.53; H, 8.85.
3␤-Acetoxy-17-hydroxy-17-methyl-16-nitrile-16,17-
seco-5-androstene (10). Compound 6 (0.15 g, 0.42 mmol)

648
K.M. Penov Gasˇi et al. / Steroids 66 (2001) 645–653
was dissolved in absolute ethanol (9 ml) and suspended with
sodium-boron hydride (0.15 g, 3.95 mmol). The reaction
mixture was vigorously stirred for another 10 min. Then, the
mixture was diluted with water and extracted with dichlo-
romethane. Drying with anhydrous Na2SO4 and solvent
evaporation gave crude product 10 (0.15 g), whose purifi-
cation on silica gel column (4.5 g, toluene/ethyl acetate, 4:1)
afforded compound 10 (0.080 g). Recrystallization from the
n-hexane/acetone (3:1) mixture afforded pure compound 10
(0.075 g, 50%, m.p. 170–172°C) was obtained after. IR
NMR (CDCl3): 212.9 (C-17, C ϭ O); 199.0 (C-3, C ϭ O);
169.1 (C-5); 124.1 (C-4); 119.0 (C-16, CN); 52.4; 51.7;
41.0; 38.6; 36.5; 35.6; 35.5; 33.9; 32.4; 31.1; 25.4 (C-20);
19.8; 18.1; 17.6 (C-18); 15.7 (C-19). Anal. calcd. for
C20H27NO2: C, 76.64; H, 8.68; found: C, 76.39; H, 8.83.
17-Methyl-16-nitrile-16,17-seco-1,4-androstadiene-
3,17-dione (13). 17-Methyl-16-nitrile-16,17-seco-4-
androstene-3,17-dione (0.301 g, 0.96 mmol) was dissolved in
absolute benzene (15 ml), DDQ (0.277 g, 1.12 mmol) added,
and the reaction mixture refluxed with constant stirring for 20
h. After completion of the reaction, the reaction mixture was
subjected to vacuum filtration over aluminum(III) oxide layer,
which was washed with benzene (3 ϫ 15 ml). The filtrate was
washed with saturated solution of Na2CO3 (5 ϫ 30 ml). The
organic layer was dried over anhydrous Na2SO4 and evapo-
rated to dryness, giving the crude product (0.261 g), which was
purificated on silica gel thin layer (n-hexane/toluene/ethyl ac-
etate, 1:1:1, triple development, triple desorption with the 30
ml volumes of ethyl acetate). The extracts were joined, the
solvent removed, and the product was in the form of light-
␯
(KBr) cmϪ1: 3520, 2260, 1730, 1250. 1H NMR
max
(CDCl3): 0.90 (3H, s, H-18); 1.05 (3H, s, H-19); 1.20 (3H,
d, J ϭ 6.3 Hz, H-20); 2.02 (3H, s, Ac); 3.75 (1H, q, J ϭ
6.3 Hz, H-17); 4.60 (1H, m, H-3); 5.40 (1H, m, H-6). 13C
NMR (CDCl3): 170.6 (C ϭ O, Ac); 139.3 (C-5); 121.5
(C-6); 119.8 (C-16, CN); 73.7 (C-3, CH-OH); 72.4 (C-17,
CH-OH); 49.0; 42.4; 40.1; 37.7; 36.8; 36.6; 32.0; 31.9;
30.6; 27.6; 21.4; 19.9; 19.1; 17.7; 17.2; 14.9. Anal. calcd.
for C22H33NO3: C, 73.50; H, 9.25; found: C, 73.02; H, 9.03.
3␤,17-Dihydroxy-17-methyl-16-nitrile-16,17-seco-5-
androstene (11). Compound 6 (0.250 g, 0.70 mmol) was
dissolved in ethanol (15 ml), sodium-boron hydride (0.500
g, 13.16 mmol) added, and the reaction mixture was con-
stantly stirred for 5 h. After that, the reaction mixture was
treated as in the procedure for compound 10. The crude
product (0.312 g) was purified on a silica gel column (9.4 g,
toluene/ethyl acetate, 2:1). The resulting product (0.125 g,
56.5%) was recrystallized from the n-hexane/acetone (3:1)
mixture, affording pure compound 11 (0.118 g, 53%, m.p.
184–186°C). IR ␯max (KBr) cmϪ1: 3700–3100, 2260, 1120,
yellow oil (0.099 g, 33%). IR ␯
(film) cmϪ1: 1725, 1710,
max
1665, 1290, 1270, 1265, 1125, 1075. 1H NMR (CDCl3): 1.23
(3H, s, H-19); 1.33 (3H, s, H-18); 2.13 (3H, s, H-20); 6.05 (1H,
t, J ϭ 1.9 Hz, H-4); 6.21 (1H, dd, J1,2 ϭ 10.1 Hz, J2,4 ϭ 1.9
Hz, H-2); 7.00 (1H, d, J1,2 ϭ 10.1 Hz, H-1). 13C NMR
(CDCl3): 212.3 (C-17, C ϭ O); 185.7 (C-3, C ϭ O); 167.2
(C-5); 154.2 (C-1); 127.9 (C-2); 123.9 (C-4); 118.7 (C-16,
CN); 51.6; 50.4; 42.8; 40.7; 36.1; 35.3; 32.5; 32.1; 25.2; 21.2;
18.7; 18.1; 15.5.
17-Methyl-16-nitrile-16,17-seco-1,4,6-androstatriene-
3,17-dione (14). a) 3␤-Hydroxy-17-methyl-16-nitrile-
16,17-seco-5-androstene-17-one (0.230 g, 0.73 mmol) was
dissolved in absolute dioxan (10 ml), then DDQ (0.547 g,
2.41 mmol) added, and the reaction mixture refluxed with
vigorous stirring for 18 h. After completion of the reaction
the reaction mixture was diluted with ethyl acetate and
filtered through aluminum(III) oxide. The solution was
treated with saturated aqueous solution of Na2CO3 (3 ϫ 30
ml) and then with saturated solution of NaHCO3 (30 ml).
After drying over anhydrous Na2SO4 and removal of the
desiccant and solvent, crude product of 14 (0.174 g) was
purified by preparative TLC on silica gel (toluene/n-hexane/
ethyl acetate, 1:1:1, double development). After desorption
with ethyl acetate and then with acetone, and after recrys-
tallization from the n-hexane/acetone (3:2) mixture com-
pound 14 (0.063 g, 28%, m.p. 151–153°C) was obtained.
b) 17-Methyl-16-nitrile-16,17-seco-4-androstene-3,17-
dione (0.077 g, 0.24 mmol) was dissolved in absolute di-
oxan (10 ml), then DDQ (0.196 g, 0.85 mmol) added, and
the reaction mixture refluxed at vigorous stirring for 25 h.
When the reaction was completed, saturated solution of
Na2CO3 (50 ml) was added, and the mixture extracted with
ethyl acetate (3 ϫ 50 ml). The organic layer was dried over
anhydrous Na2SO4, evaporated to dryness, whereby crude
product (0.086 g) was obtained. It was purified by prepar-
ative TLC on silica gel (n-hexane/toluene/ethyl acetate,
1
1050. H NMR (CDCl3): 0.94 (3H, s, H-18); 1.02 (3H, s,
H-19); 1.17 (3H, d, J ϭ 6.2 Hz, H-20); 3.53 (1H, m, H-3);
3.72 (1H, q, J ϭ 6.2 Hz, H-17); 5.36 (1H, m, H-6). 13C
NMR (CDCl3): 140.4 (C-5); 120.6 (C-6); 119.9 (C-16, CN);
72.4 (C-17, CH-OH); 71.5 (C-3, CH-OH); 49.0; 42.5; 41.9;
40.1; 36.9; 36.7; 32.1; 32.0; 31.4; 30.6; 19.9; 19.2; 17.7;
17.2; 14.8. Anal. calcd. for C20H31NO2: C, 75.66; H, 9.84;
found: C, 75.01; H, 9.34.
17-Methyl-16-nitrile-16,17-seco-4-androstene-3,17-
dione (12) The mixture of 3␤-hydroxy-17-keto-17-methyl-
16-nitrile-16,17-seco-5-androstene (7, 1.00 g, 3.17 mmol),
aluminum(III) tert-butoxide (1.30 g, 5.29 mmol) and cyclo-
hexanon (15.5 ml) was refluxed with constant stirring for 2
h. When the reaction was completed, the reaction mixture
was cooled, its pH adjusted (6 M HCl) to 3, and then
cyclohexanon removed by steam distillation. The reaction
mixture was then extracted with dichloromethane (3 ϫ 50
ml). The joint extracts were dried over anhydrous Na2SO4,
the siccative and solvent were removed, and the crude
product purified by flash chromatography (benzene/ethyl
acetate, 2:1). The obtained amorphous product (0.872 g)
was recrystallized from the n-hexane/ethyl acetate (5:3)
mixture, affording pure compound 12 (0.651 g, 64%, m.p.
133–136°C). IR ␯max (KBr) cmϪ1: 2260, 1705, 1675, 1630.
1H NMR (CDCl3): 1.22 (3H, s, H-19); 1.35 (3H, s, H-18);
2.17 (3H, s, H-20); 5.79 (1H, d, J ϭ 1.8 Hz, H-4). 13C

K.M. Penov Gasˇi et al. / Steroids 66 (2001) 645–653
649
1:1:2, triple development, desorption with ethyl acetate, 3 ϫ
30 ml). Removal of the solvent and recrystallization from
the n-hexane/acetone (3:2) mixture afforded pure com-
CO, USA), pregnant mares’ serum gonadotophin (PMSG)
was obtained from the Veterinary Institute Subotica (Yugo-
slavia), [1,2,6,7 Ϫ H(N)]estradiol from New England Nu-
3
pound 14 (0.039 g, 51%, m.p. 151–153°C). IR ␯
(KBr)
clear (Belgium), NADPH and aminoglutethimide (AG)
from Sigma (St Louis, MO, USA).
max
cmϪ1: 2240, 1700, 1665, 1655, 1650, 1640, 1610. 1H NMR
(CDCl3): 1.23 (3H, s, H-19); 1.42 (3H, s, H-18); 2.20 (3H,
s, H-20); 6.07 (1H, d, J2,4 ϭ 1.9 Hz, H-4); 6.15 (1H, dd,
Preparation of ovarian homogenate. Female Wistar
rats (10 weeks old), raised under controlled environmental
conditions (temperature 22 Ϯ 2°C and 14 h light/10 h dark)
in our laboratory, were treated with PMSG to increase
ovarian aromatase activity according to the procedure of
Brodie et al. [10]. PMSG was injected subcutaneously (0.4
ml, 100 IU) on alternate days for 9 days. Twenty-four hours
after the last injection, animals were sacrificed by decapi-
tation, the ovaries removed, dissected free of fat and washed
with ice-cold 1.15% KCl. The pooled ovaries were blotted
dry, weighed and homogenized in an ice-cold 50 mM phos-
phate buffer containing 0.2 M sucrose (pH 7.4), (1:4 w/v),
using a glass-glass homogenizer. After centrifugation (4°C
for 20 min at 1500 g), the supernatant was mixed with
dextran-coated charcoal in order to remove the endogenous
steroids [11], and again centrifuged at 1500 g for 10 min at
4°C. The obtained supernatant was diluted (1:2) with 0.1 M
phosphate buffer (pH 7.4) and retained on ice prior to use in
the aromatase assay.
J6,7 ϭ 10.1 Hz, J6,8 ϭ 1.8 Hz, H-6); 6.29 (1H, dd, J1,2
ϭ
10.2 Hz, J2,4 ϭ 1.9 Hz, H-2); 6.41 (1H, dd, J6,7 ϭ 10.1
Hz, J7,8 ϭ 2.8 Hz, H-7); 7.06 (1H, d, J1,2 ϭ 10.2 Hz,
H-1). 13C NMR (CDCl3): 211.8 (C-17, C ϭ O); 185.7 (C-3,
C ϭ O); 160.4 (C-5); 151.6 (C-1); 133.7 (C-7); 129.1 (C-6);
128.4 (C-2,); 124.2 (C-4); 118.5 (C-16, CN); 51.6; 46.5;
40.6; 39.0; 37.7; 35.9; 25.2; 20.4; 20.4; 17.8; 15.4. Anal.
calcd. for C20H23NO2: C, 77.63; H, 7.49; found: C, 77.35;
H, 7.58.
2.3. X-Ray crystal studies of 6 and 12
The compound 6 crystallizes as colorless plates, ortho-
rhombic, space group P212121. A crystal of dimensions
0.20 ϫ 0.18 ϫ 0.04 mm was selected for X-ray analysis and
mounted on an Enraf-Nonius CAD-4 diffractometer. The
refined lattice parameters, a ϭ 5.946(2), b ϭ 16.120(2), c ϭ
20.746(3) Å, were obtained by a least-squares procedure
from the setting angles of 25 reflections with 15.02 Ͻ ␪ Ͻ
23.87°. Intensities were collected with Cu K␣ radiation
(␭ ϭ 1.54178 Å), scan range 3.47 Ͻ ␪ Ͻ 74.65°, ␻-␪ scan.
A total of 6350 reflections were collected of which 3960
were unique (Rint ϭ 0.0298). Lorentz and polarization cor-
rections were applied to the data. The linear absorption
coefficient is 0.620 mmϪ1. No absorption correction was
made. The structure was solved by direct methods. 2622
reflections having I Ͼ 2␴(I) were used in refinements. The
final cycle of refinement converged with R ϭ 0.047.
The compound 12 crystallizes as colorless plates, ortho-
rhombic, space group P212121. A crystal of dimensions
0.45 ϫ 0.40 ϫ 0.06 mm was selected for X-ray analysis and
mounted on an Enraf-Nonius CAD-4 diffractometer. The
refined lattice parameters, a ϭ 8.929(2), b ϭ 10.022(3), c ϭ
19.651(9) Å, were obtained by a least-squares procedure
from the setting angles of 25 reflections with 25.03 Ͻ ␪ Ͻ
29.64°. Intensities were collected with Cu K␣ radiation
(␭ ϭ 1.54178 Å), scan range 4.50 Ͻ ␪ Ͻ 74.57°, ␻-␪ scan.
A total of 4254 reflections were collected of which 3584
were unique (Rint ϭ 0.0411). Lorentz and polarization cor-
rections were applied to the data. The linear absorption
coefficient is 0.591 mmϪ1. No absorption correction was
made. The structure was solved by direct methods. 2987
reflections having I Ͼ 2␴(I) were used in refinements. The
final cycle of refinement converged with R ϭ 0.044.
Measurement of aromatase activity. Aromatase activity
was determined in denucleated ovarian fraction using testos-
terone as a substrate and measuring estradiol production by
radioimmunoassay (RIA). The lowest detectable dose in the
assay was 4 pg/tube, range of standard curve was 4–400
pg/tube, and intra-assay and inter-assay coefficients of varia-
tions were 8.5% and 11.3%, respectively. Cross-reactivity (de-
fined as percent of the ratio between the dose of unlabeled
estradiol and the dose of heterologous steroid required to pro-
3
duce 50% inhibition of the binding of H-estradiol tracer) of
antiestradiol serum no. 244 to testosterone was 0.0068%, while
to estrone was 0.28%. We assessed also the possible cross-
reactivity of antiestradiol serum to 5-androstene derivatives
investigated in our study. We found that cross-reactivity aver-
aged 0.004%. Characteristics of this antiestradiol serum was
described by Korenman et al. [12].
In the final volume of 0.25 ml, the incubation mixture
contained testosterone as substrate, 1 mM NADPH, 0.1 M
phosphate buffer (pH 7.4), tested compound and 0.125 ml of
denucleated ovarian fraction (0.23 mg of protein). Parallel
incubations without tested compound were also run as con-
trol samples. The desired concentration of tested com-
pounds was prepared by evaporating the necessary amount
of stock solution (in abs. ethanol) and dissolving it in 0.1 M
phosphate buffer. Mixtures were incubated for 15 min at
37°C in a shaking water bath in 95% O2 Ϫ 5% CO2
atmosphere. Selected incubation time was within temporal
linearity of the enzyme activity. The enzyme reaction was
initiated by adding denucleated ovarian fraction and termi-
nated by placing the tubes in an ice-cold bath. The samples
were stored at Ϫ20°C until assayed for estradiol by RIA.
Estradiol production in control and experimental samples
2.4. Biological methods
Chemicals. Antiestradiol serum no. 244 was kindly sup-
plied by Dr. G.D. Niswender (Colorado State University,

650
K.M. Penov Gasˇi et al. / Steroids 66 (2001) 645–653
were calculated after subtraction of the corresponding
blanks. Namely, two sets of blanks were run in parallel with
control and experimental samples. In the first set, testoster-
one and tested compounds were omitted in order to deter-
mine residual levels of estradiol and possible endogenous
estradiol production in denucleated ovarian fractions used in
each experiment (determined values were 0.5–5% of control
production depending on testosterone concentration used as
substrate in control samples). In the second set, homogenate
was omitted in order to determine cross-reactivity of testo-
sterone and tested compounds.
For a preliminary assessment of anti-aromatase activity
of the synthesized compounds, the given compound was
added in a single concentration (50 ␮M) to the incubation
mixture containing 50 nM or 500 nM of testosterone as
substrate (sub-saturated and saturated concentration, respec-
tively; the estimated Km for testosterone was 49.17 nM, and
Vmax 5.76 pmol/min ⅐ mg protein).
The screening assay for determining the IC50 values
contained 500 nM of testosterone (saturated concentration)
and various concentrations of tested compounds. AG was
used as referent compound. The results were analyzed by
the least-squares (log concentration of the tested compound)
and the IC50 values were calculated.
Scheme 1. D-modified androstane derivatives as aromatase inhibitors. Re-
agents: a) MeMgI, ether, THF; b) EtMgI, ether, THF; c) PhLi, ether, THF; d)
TiCl3, 6 M HCl, EtOH; e) Ac2O, Py; f) NaOEt, EtOH; g) NaBH4, EtOH.
For kinetic study, aromatase activity was determined at
different testosterone concentrations (25–100 nM) in the ab-
sence (control) and presence of two concentrations of com-
pound 12 (500 nM and 1 ␮M). To asses type of inhibition the
results were analyzed by double reciprocal plot. The slopes of
the reciprocal plots, as determined by the least-squares line for
each concentration, were plotted versus the inhibitor concen-
tration, and the Ki value was calculated.
16,17-seco-5-androstene-17-one. Repetition of the experi-
ment at room temperature during 12 h yielded again the
corresponding D-seco derivatives 8 and 9, but at lower
yields (8, 63%; 9, 60%), whereby these reactions were
essentially different from the analogous reactions of the
17␣-benzyl derivative [13]. In the latter case, under these
reaction conditions, the C-16 oximino group predominantly
transformed into the C-16 keto group, accompanied by
inversion of the configuration at the C-17 atom. In contrast
to compounds 3 and 4, which under these reaction condi-
tions formed the D-seco derivatives 8 and 9 as the only
products, compound 2 afforded only the product of hydro-
lysis of the C-16 oximino group and inversion of the con-
figuration at the C-17 atom (5, Scheme 1). Thus, the key
role of the 17␣-substituent in the fragmentation/rearrange-
ment and hydrolysis processes of the 16-oximino-17␤-hy-
droxy-17␣-substituted derivatives of androstane, catalyzed
by titanium(III) chloride, was confirmed once more.
By using NMR method and measuring the nuclear Over-
hauser effect it was possible to determine the configuration
at the C-17 atom in compounds 2 and 5. Irradiation of the
H-18 signal at 0.70 ppm of compound 2 produced intensity
enhancement of the OH-17 signal located at 4.54 ppm
(exchangeable with D2O). These results confirm the ␤-ori-
entation of the OH-17 group, and addition of methyl-mag-
nesium iodide to the carbonyl group of compound 1 from
the sterically less hindered ␣-site. Treatment of compound 2
with acidic aqueous titanium(III) chloride solution, apart
3. Results and discussion
3.1. Chemistry
In our previous work [13] we synthesized the 16-
oximino-17␤-hydroxy-17␣-picolyl- and 17␣-benzyl de-
rivatives of androstane by a regio-selective addition of
␣-picolyllithium and benzyllithium to the 17-keto group
of 3␤-hydroxy-16-oximino-5-androstene-17-one (1). In this
work we prepared ␣-hydroxy oximes 2, 3, and 4 by a
selective addition of methyl-magnesium iodide, ethyl-mag-
nesium iodide, and phenyllithium to the 17-keto group of 1
(Scheme 1).
The side product in the synthesis of compound 3 was the
corresponding D-seco derivative 8. Compounds 2, 3, and 4
were subjected to the reaction of Beckmann fragmentation
using titanium(III) chloride or acetic anhydride, with the
aim of preparing new D-seco derivatives. The use of tita-
nium(III) chloride in acidic medium (hydrochloric acid) in
ethanol as solvent, at the boiling temperature in the course
of 10–20 min resulted in the 17-ethyl- (8, 86%) and 17-
phenyl (9, 86%) derivatives of 3␤-hydroxy-16-nitrile-

K.M. Penov Gasˇi et al. / Steroids 66 (2001) 645–653
651
Fig. 1. Perspective view of the molecular structure of compound 6.
from the hydrolysis of the 16-oximino group, results in the
total inversion of the configuration at the C-17, giving
compound 5. Irradiation of the OH-17 signal at 5.20 ppm
for compound 5 measured in DMSO-d6, gives no signal
enhancement for the H-18 at 0.61 ppm. However, we have
found a measurable NOE effect on the H-17 signal during
irradiation of the H-18.
fraction from PMSG-pretreated female rats. For screening pur-
pose, most of the synthesized compounds were tested in a
single concentration (50 ␮M). The results showed that com-
pounds 7, 8, 12, 13, and 14 completely inhibited aromatase
activity in the presence of a sub-saturated dose of testosterone,
while in the presence of saturated concentration only com-
pound 12 exhibited such an effect (Table 1).
In view of the fact that the treatment of 2 with titanium(III)
chloride did not yield fragmentation product, the attempt with
acetic anhydride in pyridine appeared to be effective in obtain-
ing the fragmentation product 6, which, by further treatment
with sodium-etoxide in ethanol was transformed into 3␤-hy-
droxy-D-seco derivative 7. The keto group at the C-17 of
compound 6 was reduced with sodium-boron hydride in eth-
anol, whereby compounds 10 and 11 were obtained.
The structure of compound 6 was unambiguously proved
by X-ray structural analysis. A perspective view of the
molecule with the standard atom-labelling system is shown
in Fig. 1.
In view of the effects of compounds 7 and 11, the
17-keto group increased anti-aromatase activity comparing
to the 17-hydroxy function of D-seco derivatives. Also,
3␤-acetoxy derivative (compound 6) showed much lower
potency than the corresponding derivative with free 3␤-
hydroxy substituent (compound 7). However, the presence
of 4-en-3-keto system instead of 5-en-3-hydroxy increased
markedly the anti-aromatase activity (compound 12 vs. 7).
On the other hand, extended linear conjugation in rings A
and B (e.g. 1,4-diene-3,17-dione, or 1,4,6-triene-3,17-di-
one) did not increase the anti-aromatase potency of such
The Oppenauer oxidation of compound 7 using aluminu-
m(III) tert-butoxide in cyclohexanon gave 4-ene-3-keto deriv-
ative 12, which by dehydration with dichlorodicyano-benzo-
quinone (DDQ) in benzene according to the known method
[14] gave 1,4-diene-3-keto derivative 13 (Scheme 2). By using
the same reagent (DDQ) but in dioxan as solvent [13] the
1,4,6-triene-3-keto derivative 14 was obtained from 12. Dehy-
dration of 7 with DDQ in dioxan gave 14 [15].
The X-ray structural analysis of 12 unambiguosly proved
the structure of these new D-seco steroid (Fig. 2).
3.2. Biochemical properties
The compounds synthesized in this study were tested on the
possible anti-aromatase activity in the denucleated ovarian
Scheme 2. Reagents: a) Al(t-BuO)3, cyclohexanone; b) DDQ, benzene; c)
DDQ, dioxan.

652
K.M. Penov Gasˇi et al. / Steroids 66 (2001) 645–653
Fig. 2. Perspective view of the molecular structure of compound 12.
derivatives (compounds 13 and 14 vs. compound 12) (Ta-
bles 1 and 2). Similar results were obtained by other authors
[9] using 4-thiosubstituted androstene-3,17-dione. They
found that 4-thiophenyl-4-androstene-3,17-dione (4-TPAD)
had a lower Ki than the corresponding 1,4-diene and 1,4,6-
triene analogs.
According to the results presented in Table 2, IC50 for
AG was 0.14 ␮M. This value is similar to the value ob-
tained for d-AG toward rat ovarian as well as human pla-
cental aromatase, using subsaturating dose (2 Km) of andro-
stenedione as substrate and tritiated water release method
for measuring aromatase activity [16]. The same authors
measured the aromatase activity following the conversion of
unlabeled androstenedione to estrone and determining the
levels of estrone by RIA. They found excellent agreement
between these two methods. On the other hand, Abul-Hajj et
al. [9] reported an IC50 of 10.3 ␮M in the presence of
subsaturating dose (5 Km) of androstenedione as substrate
for human placental aromatase. Also, Hartmann et al. [17]
reported IC50 for d-AG and AG of 19 ␮M and 37 ␮M
respectively using human placental microsoma as a source
of enzyme and saturating dose of testosterone as substrate.
In our study, the highest potency toward rat ovarian aro-
matase exhibited compound 12 with the 4-en-3-keto system.
Table 1
Inhibitory effects of androstene derivatives on the aromatase activity in
the denucleated fraction of ovaries from PMSG pretreated rats
Compound
(50 ␮M)
Aromatase activity
500 nM T
50 Nm T
% of inhibition vs control
1
2
3
4
6
43.55 Ϯ 1.62**
2.64 Ϯ 7.22
86.75 Ϯ 0.19**
19.87 Ϯ 9.18
16.27 Ϯ 5.91
Ϫ6.96 Ϯ 3.61
62.23 Ϯ 1.12**
100**
26.41 Ϯ 6.48**
79.24 Ϯ 3.80**
80.24 Ϯ 2.45
7
8
Table 2
100**
In vitro aromatase inhibition by androstene derivatives
9
36.19 Ϯ 7.39
75.24 Ϯ 1.14**
100**
100**
100**
11
12
13
14
37.91 Ϯ 6.50**
100**
66.87 Ϯ 4.95**
66.88 Ϯ 2.72**
Compound
IC50 (␮M)
7 (⌬5-3-ol)
12 (⌬4-3-on)
14 (⌬1,4,6-3-on)
For comparison
AG
5.45
0.42
16.84
Purified denucleated fraction of ovaries from PMSG pretreated female
rats (0.366 mg proteins) was incubated in the presence of subsaturated (50
nM) or saturated (500 nM) concentrations of substrate testosterone (T) and
NADPH (1 mM) and in the absence (control) or presence of different
androstene derivatives for measuring aromatase activity. Estradiol level
was determined by RIA. Results are presented as the percentage of inhi-
bition vs control. Numbers represent mean Ϯ SEM of 6 replicates.
Significance: **P Ͻ 0.005 vs control (Mann-Whitney non-parametric
test). Values are reported as representative of two experiments.
0.14
Purified denucleated fraction of ovaries from PMSG-pretreated female
rats (0.36 mg proteins) was incubated for 15 min at 37°C in the presence
of saturated (500 nM) concentration of testosterone and NADPH (1 mM)
in the absence (control) or presence of 4–5 concentrations of compounds
7, 12, 14 or AG. Estradiol level was determined by RIA. IC50 was
calculated from least-squares lines.

K.M. Penov Gasˇi et al. / Steroids 66 (2001) 645–653
653
seco derivative with 4-en-3-keto system) showed the
highest activity and competitive type of inhibition with a Ki
value of 0.27 ␮M.
Acknowledgment
The authors thank the Ministry for Science and Technol-
ogy of the Republic of Serbia for financial support. They are
also indebted to Professors Alajos Ka`lma`n and Gyula Argay
(Central Research Institute for Chemistry, Hungarian Acad-
emy of Sciences) for their help in determining the X-ray
crystal structure of 6 and 12.
References
[1] Thompson AE Jr., Siiteri PK. The involvement of human placental
microsomal P-450 in aromatization. J Biol Chem 1974;249:5373–8.
[2] Li S, Parish EJ. Design and action of steroidal aromatase inhibitors.
JAOCS 1996;73:1435–51.
[3] Brodie AMH, Njar VCO. Aromatase inhibitors in advanced breast
cancer: mechanism of action and clinical implications. J Steroid
Biochem Molec Biol 1998;66:1–10.
[4] Brodie AMH, Njar VCO. Aromatase inhibitors and breast cancer.
Semin Oncol 1996;23:10–20.
[5] Harvey HA. Aromatase inhibitors in clinical practice: current status
and a look to the future. Semin Oncol 1996;23:33–8.
[6] Buzdar AU, Plourde PV, Hortobagyi GN. Aromatase inhibitors in
metastatic breast cancer. Semin Oncol 1996;23:28–32.
[7] Schwarzel WC, Kruggel WG, Brodie HJ. Studies on the mechanism
of estrogen biosynthesis. VII. The development of inhibitors of the
enzyme system in human placenta. Endocrinology 1973;92:866–80.
[8] Numazawa M, Oshibe M, Yamaguchi S. 6-Alkylandrosta-4,6-diene-
3,17-diones and their 1,4,6-triene analogs as aromatase inhibitors.
Steroids 1997;62:595–602.
[9] Abul-Hajj YJ, Liu X-P, Hedge M. Aromatase inhibitors: Effect of ring A
and ring B unsaturation on aromatase inhibition by 4-thiosubstituted
derivatives of 4-androstene-3,17-dione. Steroids 1995;60:423–7.
[10] Brodie AMJ, Schwarzel WC, Brodie HJ. Studies on the mechanism of
estrogen biosynthesis in the rat ovary. J Steroid Biochem 1976;7:787–93.
[11] Luzzani F, Soffientini A. Studies of cytosol progestin-binding com-
ponents in the uthero-placental unit of pregnant hamster. J Steroid
Biochem 1979;13:697–701.
[12] Korenman S, Stevens R, Carpenter L, Robb M, Niswender G, Sher-
man B. Estradiol radioimmunoassay without chromatography: proce-
dure, validation and normal values. J Clin Endocrinol Metab 1974;
38:718–20.
[13] Miljkovic´ D, Penov Gasˇi K, Djurendic´ E, Sakacˇ M, Medic´-Mijacˇevic´
Lj, Pejanovic´ V, Stankovic´ S, Lazar D. A novel rearrangement of
steroidal ␣-hydroxy oximes. Tetrahedron Lett 1997;38:4683–4.
[14] Turner AB, Ringold HJ. Application of high potential quinones. Part
I. The mechanism of dehydrogenation of steroidal ketones by 2,3-
dichloro-5,6-dicyanobenzoquinone. J Chem Soc 1967;1720–30.
[15] Turner AB. Applications of high potential quinones. Convenient
synthesis of steroidal 1,4,6-trien-3-ones. J Chem Soc Chem Comm
1966;845–6.
Fig. 3. Kinetic analysis of the mechanism of inhibition of aromatase in
denucleated rat ovarien by compound 12. F No inhibitor; E, 0.5 ␮M
compound 12; , 1.0 ␮M compound 12. A. Double reciprocal plot. B.
Slope of lines in A panel vs inhibitor concentration. Km for testosterone
was 49.17 nM.
The IC50 for this compound was, however, 3.5 times higher
compared to the value for AG. As is shown in Fig. 3,
compound 12 exhibited a competitive nature of inhibition.
The apparent inhibition constant (Ki) which characterizes
the enzyme affinity was determined to be 0.27 ␮M for
aromatase from the rat ovarian tissue. At the same time, Km
for testosterone was 49.17 nM, which is similar to the Km
for testosterone (46 nM) determined for human placental
aromatase [17], but 1.5–2 times higher that Km for andro-
stenedione for human placental aromatase [9,16].
Further investigation of the efects of compound 12 on the
activity of other steroidogenic enzymes could prove its
specificity toward aromatase, since it is known that steroid
derivatives are more specific than non-steroid aromatase
inhibitors [3].
[16] Kellis J, Vickery L. Inhibition of estrogen synthetase (aromatase) by
4-cyclohexylaniline. Endocrinology 1984;114:2128–37.
[17] Hartmann RW, Gru¨n G, Bartz U, Palzer M. Evaluation of the race-
mate and the enantiomers of a new highly active and selective
aromatase inhibitor of the aminoglutethimide type. J Steroid Biochem
Molec Biol 1992;43:641–8.
In conclusion, the presented results show that some D-
modified derivatives introduced in this study showed anti-
aromatase activity of different potency. Compound 12 (D-