Structure-activity relationships of 3-deoxy androgens as aromatase inhibitors. synthesis and biochemical studies of 4-substituted 4-ene and 5-ene steroids

Article abstract of DOI:10.1016/S0039-128X(03)00085-0

As part of our investigation into the structure-activity relationship of a novel class of aromatase inhibitors, two series of 3-deoxy androgens, androst-5-en-17-ones with a non-polar alkoxy (5 and 6), alkyl (20-22), or phenylalkyl (23 and 24) group at C-4β and 4-acyloxyandrost-4-en-17-ones (29-32, and 34) were synthesized and evaluated. The 4β-alkyl and 4β-phenylalkyl compounds were obtained through reaction of 4α,5α-epoxy steroid (8) with RMgBr (R: alkyl and phenylalkyl) followed by dehydration of the 4β-substituted 5α-hydroxy products (15-19) with SOCl₂ as key reactions. Acylation of 4α,5α-diol (25) with (RCO)₂O in pyridine and subsequent dehydration with SOCl₂ gave the 4-acyloxy steroids. All of the steroids studied, except for 4-acetoxy-19-ol (34) that was a non-competitive inhibitor of human placental aromatase, blocked aromatase activity in a competitive manner. 4-Benzoyloxy- and 4-acetoxy steroids (31) and (32) were the most powerful inhibitors of aromatase (K_i=70 and 60nM, respectively). Elongation of an acetoxy group in a series of 4-acyloxy steroids or a methyl group in a series of 4β-alkyl steroids decreased affinity for aromatase principally in relation to carbon number of the acyl or alkyl function. The present findings are potentially useful for understanding the spatial and electronic nature of the binding site of aromatase as well as for developing effective aromatase inhibitors.

Full text of DOI:10.1016/S0039-128X(03)00085-0

Steroids 68 (2003) 533–542
Structure–activity relationships of 3-deoxy androgens as
aromatase inhibitors. synthesis and biochemical studies
of 4-substituted 4-ene and 5-ene steroids
Masao Nagaoka, Yoko Watari, Hiromi Yajima, Kaoru Tsukioka,
Yasuyo Muroi, Keiko Yamada, Mitsuteru Numazawa∗
Tohoku Pharmaceutical University, 4-1 Komatsushima-4-chome, Aobaku, Sendai 981-8558, Japan
Received 20 March 2003; received in revised form 13 May 2003; accepted 16 May 2003
Abstract
As part of our investigation into the structure–activity relationship of a novel class of aromatase inhibitors, two series of 3-deoxy
androgens, androst-5-en-17-ones with a non-polar alkoxy (5 and 6), alkyl (20−22), or phenylalkyl (23 and 24) group at C-4␤ and
4-acyloxyandrost-4-en-17-ones (29–32, and 34) were synthesized and evaluated. The 4␤-alkyl and 4␤-phenylalkyl compounds were ob-
tained through reaction of 4␣,5␣-epoxy steroid (8) with RMgBr (R: alkyl and phenylalkyl) followed by dehydration of the 4␤-substituted
5␣-hydroxy products (15−19) with SOCl2 as key reactions. Acylation of 4␣,5␣-diol (25) with (RCO)2O in pyridine and subsequent
dehydration with SOCl2 gave the 4-acyloxy steroids. All of the steroids studied, except for 4-acetoxy-19-ol (34) that was a non-competitive
inhibitor of human placental aromatase, blocked aromatase activity in a competitive manner. 4-Benzoyloxy- and 4-acetoxy steroids (31)
and (32) were the most powerful inhibitors of aromatase (Ki = 70 and 60 nM, respectively). Elongation of an acetoxy group in a series
of 4-acyloxy steroids or a methyl group in a series of 4␤-alkyl steroids decreased affinity for aromatase principally in relation to carbon
number of the acyl or alkyl function. The present findings are potentially useful for understanding the spatial and electronic nature of the
binding site of aromatase as well as for developing effective aromatase inhibitors.
© 2003 Elsevier Inc. All rights reserved.
Keywords: Aromatase; Inhibitor; 4␤-Alkyl steroid; 4-Acyloxy steroid; 3-Deoxy androgen; Human placental microsomes
1. Introduction
that would be occupied by its C4 [6–8], C6 [9–11], and C7
[12–14] atoms.
We have previously reported that androst-4-en-17-one (1),
a 3-deoxy derivative of the substrate androstenedione, is an
excellent competitive inhibitor of aromatase [15,16], while
its 5-en isomer (2) is a good competitive inhibitor [17]
(Fig. 1). The structure–activity relationships [15–20] as well
as aromatase-catalyzed 19-oxygenation studies [21] indi-
cated that the two 3-deoxy steroids bind to aromatase in a
different manner. The binding pocket accommodates a polar
hydroxy group at the 6- or 19-position of 4-ene steroid (1)
[15,18], whereas a hydrophilic interaction of a 4␤-hydroxy
or 4␤,19-dihydroxy function with the pocket plays a critical
role in the tight binding of 5-ene steroids (2) [20].
In continuing the study of 3-deoxy steroids as conforma-
tional and catalytic probes for the active site of aromatase,
we were interested in steroids that have a non-polar sub-
stituent at the 4-position of 4-ene steroid (1) and at the
4␤-position of 5-ene steroid (2) to determine whether a hy-
drophobic interaction is involved in the binding of 3-deoxy
Aromatase is a unique cytochrome P-450 enzyme that
catalyzes the synthesis of estrone and estradiol from
the 4-en-3-one androgens androst-4-ene-3,17-dione (an-
drostenedione) and testosterone [1–3]. There appears to be
a major clinical role for controlling estrogen levels in the
treatment of estrogen-dependent breast cancer, and aro-
matase inhibitors represent one route to such control (for a
review, see [4]). For this reason, the aromatase inhibiting
activity of various steroids has been tested in a number of
laboratories (for a review see [5]). The structure–activity
studies on aromatase inhibitors predicted the existence of
a hydrophobic binding pocket extending roughly in the
plane of the substrate androstenedione from the position
∗
Corresponding author. Tel.: +81-22-234-4181x3303;
fax: +81-22-275-2013.
E-mail address: numazawa@tohoku-pharm.ac.jp (M. Numazawa).
0039-128X/$ – see front matter © 2003 Elsevier Inc. All rights reserved.
doi:10.1016/S0039-128X(03)00085-0

534
M. Nagaoka et al. / Steroids 68 (2003) 533–542
raphy of the product (hexane/EtOAc, 10:1, v/v) followed
H3C
H3C
O
O
by recrystallization from acetone gave 4␤-methoxide (5)
(25%) or 4␤-ethoxide (6) (56%).
H3C
R2
4␤-Methoxyandrost-5-en-17-one (5) m.p. 187–190 ◦C
(from acetone). 1H NMR δ: 0.89 (3H, s, 18-CH3), 1.13 (3H,
s, 19-CH3), 3.16 (3H, s, 4␤-OCH3), 3.55 (1H, s, 4␣-H),
R1
; R = H, R = CH3
5.54 (1H, m, 6-H). FT-IR cm−1: 1737 (C O). MS m/z
=
(relat. int.): 302 (M+, 53), 287 (100), 270 (13), 255 (10).
Anal. Calcd. for C20H30O2: C, 79.42; H, 10.00. Found: C,
79.19; H, 9.89.
1
2
3
4
1
2
; R = OH, R = CH
3
1
2
; R = OH, R = CH OH
1
2
2
4␤-Ethoxyandrost-5-en-17-one (6) m.p. 130–133 ◦C
(from acetone). 1H NMR δ: 0.89 (3H, s, 18-CH3), 1.14 (3H,
s, 19-CH3), 1.16 (3H, t, J = 7 Hz, 4␤-OCH2CH3), 3.13 and
3.40 (1H each, m, 4␤-OCH2CH3), 3.66 (1H, s, 4␣-H), 5.50
Fig. 1. Structures of 3-deoxy androgens.
steroids to the active site. This study focused on the synthe-
sis and biochemical evaluation of 4␤-alkoxy and 4␤-alkyl
derivatives of 5-ene steroid (2) as well as 4-acyloxy deriva-
tives of 4-ene steroid (1). All of the steroids examined, ex-
cept for the 4␤-alkoxy and 4␤-phenylalkyl compounds, were
powerful to fair inhibitors of aromatase, indicating that the
small-sized 4-acyloxy and 4␤-alkyl groups can tolerate the
access of 3-deoxy steroid (1) or (2).
(1H, m, 6-H). FT-IR cm−1: 1747 (C O). MS m/z (relat.
=
int.): 316 (M+, 53), 301 (100), 270 (52), 255 (17). Anal.
Calcd. for C21H32O2: C, 79.69; H, 10.19. Found: C, 79.74;
H, 10.02.
2.2. Epoxidation of androst-4-en-17β-yl acetate (7) with
m-chloroperbenzoic acid (MCPBA)
A solution of 4-ene steroid (7) (2.03 g, 6.42 mmol), which
was obtained by treatment of 3,3-ethylenedithio derivative
of testosterone with Na metal and liquid NH3 followed
by acetylation [16], and MCPBA (1.55 g, 9.0 mmol) in
CHCl3 (28 ml) was allowed to stand overnight at 4 ◦C in the
dark. Subsequently, the solution was diluted with EtOAc
(300 ml), washed sequentially with 10% Na2S2O3 solution,
5% NaHCO3 solution, and water, and dried with Na2SO4.
Evaporation of the solvent gave an oil that was purified by
column chromatography hexane/EtOAc (15:1, v/v), giving
4␣,5␣-epoxyandrostan-17␤-yl acetate (8) (69%) as the less
polar product and 4␤,5␤-epoxyandrostan-17␤-yl acetate
(9) (12%) as the more polar product. 4␣,5␣-Epoxide (8):
2. Experimental
Melting points were measured on a Yanagimoto melt-
ing point apparatus and were uncorrected. IR spectra were
recorded on a Perkin-Elmer FT-IR 1725X spectrophotome-
ter in KBr pellets, except for oily compounds of which
spectra were obtained in neat forms. UV spectra were deter-
mined in 95% ethanol on a Hitachi 150–20 UV spectropho-
tometer. 1H NMR spectra were obtained in CDCl3 solution
with JEOL EX270 (270 MHz) using tetramethylsilane (δ =
0.00) as an internal standard, and mass spectra (MS; elec-
tron impact) were obtained with a JEOL JMS-DX 303 spec-
trometer. Thin-layer chromatography (TLC) was performed
with E. Merck precoated silica gel plates (Darmstadt, Ger-
many). Column chromatography was conducted with silica
gel (E. Merck, 70–230 mesh). Androst-4-en-17-one (1) was
synthesized by desulfurization of 3,3-ethylenedithio acetal
of androstenedione with sodium metal and liquid NH3
[15]. Androst-5-en-17-one (2) was obtained by treatment
of 3␤-(p-toluenesulfonyloxy)androst-5-en-17-one with Zn
powder and NaI [2].
1
m.p. 128–132 ◦C (from acetone). H NMR δ: 0.81 (3H, s,
18-CH3), 1.01 (3H, s, 19-CH3), 2.04 (3H, s, 17␤-OCOCH3),
2.89 (1H, d, J = 4 Hz, 4␤-H), 4.60 (1H, m, 17␣-H). FT-IR
+
cm−1: 1736 (C O). MS m/z (relat. int.): 332 (M , 84),
272 (17), 43 (100). Anal. Calcd. for C21H32O2: C, 75.86;
H, 9.70. Found: C, 75.73; H, 9.66. 4␤,5␤-Epoxide (9):
=
m.p. 194–196 ◦C (from acetone). H NMR δ: 0.78 (3H, s,
1
18-CH3), 0.88 (3H, s, 19-CH3), 2.04 (3H, s, 17␤-OCOCH3),
2.85 (1H, d, J = 4 Hz, 4␣-H), 4.59 (1H, m, 17␣-H). FT-IR
cm−1: 1736 (C O). MS m/z (relat. int.): 332 (M , 5), 98
(100), 43 (60). Anal. Calcd. for C21H32O2: C, 75.86; H,
9.70. Found: C, 75.94; H, 9.63.
+
=
2.1. Treatment of 4β-hydroxy-5-ene steroid 3 with alkyl
iodide and Ag2O
2.3. Grignard reaction of 4α,5α-epoxide (8)
4␤-Hydroxyandrost-5-en-17-one (3) (50 mg, 0.174 mmol),
which was synthesized through allylic acetoxylation of
5-ene steroid (2) with bromine and silver acetate as a key re-
action [22], was dissolved in CH3CN (4 ml). Ag2O (60 mg,
0.26 mmol) and CH3I or C2H5I (53 mmol) were added to
this solution, and the mixture was refluxed for about 40 h.
The solid material was removed by filtration, and the filtrate
was evaporated to give an oily product. Column chromatog-
(A) 3 M RMgBr (R: methyl, ethyl or benzyl) solution in
THF (30–39 mmol, 10–13 ml) or 2 M n-C3H7MgCl
(31 mmol, 16 ml) was slowly added, separately, to a
solution of epoxide (8) (480 mg, 1.45 mmol) in anhy-
drous THF (10 ml) under a stream of N2 gas while
stirring. The mixture was refluxed for 4–18 h under

M. Nagaoka et al. / Steroids 68 (2003) 533–542
535
a N2 atmosphere, and after the mixture was cooled,
saturated NH4Cl solution (50 ml) was added, and the
resulting product was extracted with EtOAc (300 ml
2×). The combined organic layer was washed with
water, dried with Na2SO4, and evaporated to dryness.
The residue was purified by column chromatography
using hexane/EtOAc (5: 1, v/v) and recrystallized to
yield 4␤-alkyl-5␣,17␣-diols (10–13).
washed sequentially with 10% HCl, 5% NaHCO3 solu-
tion, and water, dried with Na2SO4, and evaporated to
give the corresponding 4␣-ester derivatives (26 and 27)
as oily products, which were purified by column chro-
matography (hexane-EtOAc, 20:1, v/v) and preparative
TLC (hexane-EtOAc, 8:1, v/v, multiple developments),
affording the analytical samples.
(B) Benzoyl chloride (57 ␮l, 0.149 mmol) was added to a
solution of diol (25) (50 mg, 0.16 mmol) in dry pyri-
dine (0.4 ml) at 0 ◦C, and the mixture was allowed to
stand at 0 ◦C for 2 h. After the same work-up as de-
scribed above, an oily product was obtained purified
by preparative TLC (hexane-EtOAc, 30:1, v/v, multiple
developments), yielding 4␣-benzoate (28).
(B) A solution of 4-methoxybenzylchloride (25 mmol)
in anhydrous THF (25 ml) was added to a stirred
suspension of Mg metal (0.6 g, 25 mmol) under a
stream of N2 gas, and the mixture was stirred until
it became a clear solution. A solution of epoxide (8)
(410 mg, 1.24 mmol) in anhydrous THF (10 ml) was
slowly added to the Grignard reagent solution, and
the mixture was refluxed for 4 h under a N2 atmo-
sphere. After treatment of the reaction mixture as de-
scribed above, 4␤-(4ꢀ-methoxy)benzyl-5␣-ol (14) was
obtained.
2.7. Treatment of 5α-hydroxy steroids (26–28)
with SOCl2
4␣-Substituted 5␣-ols (26–28) (0.59 mmol) were each
treated with SOCl2 (0.37 ml, 5.1 mmol) according to the
method described above for the dehydration reaction of
4␤-substituted 5␣-ols (15–19), giving the corresponding
4-ene products (29–31) along with their 5-ene isomers as
minor products (4-ene:5-ene = about 7: 1). The crude
products (30 and 31) were purified by column chromatog-
raphy (hexane-EtOAc, 30:1, v/v) and preparative TLC
(hexane-EtOAc, 10:1, v/v, multiple developments, respec-
tively), yielding the analytical samples as oils. In the case
of n-butyrate (29), the chromatographic purification failed
to separate the 4-ene product (29) from its 5-ene isomer.
Thus, a mixture of compound (29) and its 5-ene isomer
was subjected to selective reduction of the 5-ene isomer to
isolate the less reactive steroid (29); the mixture (30 mg)
in EtOH (1.5 ml) was stirred with 5% Pd/C (1.0 mg) un-
der a H2 atmosphere for 20 min. After the solid mate-
rial was removed by filtration, the filtrate was evaporated
to dryness. The oily residue was purified by preparative
TLC, yielding 4-ene-n-butyrate (29) (oil) as the analytical
sample.
2.4. Jones oxidation of 4β-substituted 5α,17β-diols
Jones reagent was added to separate solutions of
5␣,17␤-diols (10–14) (0.42 mmol) in acetone (34 ml) at
0 ◦C while stirring until an orange color of the reagent
remained. Then, the mixtures were poured into water, and
the products were extracted with EtOAc (200 ml 2×). The
combined organic layer was washed with 5% NaHCO3
solution and water, dried with Na2SO4, and evaporated to
afford the corresponding 17-ketones (15–19) that were pu-
rified by column chromatography using hexane/EtOAc (8:1,
v/v) and/or recrystallized, to give the analytical samples.
2.5. Treatment of 5α-hydroxy steroids (15–19)
with SOCl2
SOCl2 (0.16 ml, 2.2 mmol) was added to solutions of
5␣-ols (15–19) (0.22 mmol) in dry pyridine (1.6 ml) at 0 ◦C.
The reaction mixture was stirred for 15 min, then poured into
ice-water (20 ml), and extracted with EtOAc (20 ml 2×). The
combined organic layer was sequentially washed with 10%
HCl, 5% NaHCO3 solution, and water, dried with Na2SO4,
and evaporated to produce the corresponding 5-ene products
(20–24), which were purified by column chromatography
and/or recrystallization, yielding the analytical samples.
2.8. 19-Hydroxyandrost-4-en-17-on-4-yl acetate (34)
A mixture of previously synthesized 19-(tert-butyl-
dimethylsiloxy)androst-4-en-17-on-4-yl acetate (33) [20],
(150 mg, 0.33 mmol), THF (1.8 ml), propane-2-ol (3.8 ml),
and 1 M HCl (2 ml) was stirred for 50 h at room temperature,
diluted with EtOAc (100 ml), washed with 5% NaHCO3
and water, dried with Na2SO4, and evaporated to yield a
solid product that was purified by column chromatography
(hexane/EtOAc, 10:1, v/v) and recrystallized from acetone,
yielding 19-ol (34) (43 mg, 38%). m.p. 112–115 ◦C. 1H
NMR δ: 0.88 (3H, s, 18-CH3), 2.16 (3H, s, 4-OCOCH3),
3.53 and 3.92 (1H each, d, J = 10 Hz, 19-CH2OH). FT-IR
2.6. Treatment of 4α,5α-dihydroxy steroid (25) with acid
anhydride or benzoyl chloride
(A) n-Butyric anhydride or n-hexanoic anhydride (0.5 ml)
was added to a solution of 4␣,5␣-diol (25) (110 mg,
0.36 mmol), which was synthesized by treatment of
4-ene steroid (1) with OsO4 [20], in pyridine (1.1 ml),
and the mixture was allowed to stand at room tempera-
ture overnight. It was then diluted with EtOAc (50 ml),
cm−1: 3475 (OH), 1736 and 1726 (C O). MS m/z (relat.
=
int.): 346 (M+, 1), 273 (100), and 256 (85). Anal. Calcd.
for C21H30O4: C, 72.80; H, 8.73. Found: C, 72.95; H, 8.84.

536
M. Nagaoka et al. / Steroids 68 (2003) 533–542
2.9. 19-Hydroxy-5α-androstane-4,17-dione (36)
A solution of K2CO3 (12 mg, 86.8 mmol) in water (1.3 ml)
was added to a solution of 4-acetate (34) (25 mg, 0.07 mmol)
in MeOH (8.0 ml). The mixture was stirred at room temper-
ature for 1 h, diluted with EtOAc (100 ml), washed with wa-
ter, dried with Na2SO4, and evaporated to afford a solid that
was recrystallized from EtOAc to give 5␣-4-one (36) (12 mg,
1
54%). m.p. 150–153 ◦C. H NMR δ: 0.83 (3H, s, 18-CH3),
3.65 and 4.00 (1H each, d, J = 11 Hz, 19-CH2OH). FT-IR
Fig. 2. Synthesis of 4␤-alkoxyandrost-4-ene steroids.
cm−1: 3508 (OH), 1734 and 1697 (C O). MS m/z (relat.
=
int.): 304 (M+, 33), 273 (100), and 237 (12). Anal. Calcd.
for C19H28O3: C, 74.96; H, 9.27. Found: C, 74.78; H, 9.36.
(10–14) (13–34%). Grignard reaction of the epoxy com-
pound was predicted to yield the trans-diaxial product,
a 4␤-substituted 5␣-ol, as a major product, based on the
stereoelectronic features of the reaction mechanism. Fur-
thermore, the stereochemistry of the 4␤-alkyl compounds
2.10. Enzyme preparation
Preparation of microsomes from human term placenta
was performed according to the method previously reported
by Ryan [23]. The microsomes were washed once with
0.05 mM dithiothreitol solution, lyophilized, and stored at
−20 ◦C.
1
was unambiguously determined on the basis of H NMR
spectroscopy. Thus, NOESY spectra indicated that there
2.11. Aromatase assay procedure
Aromatase activity was measured using [1␤-3H]androste-
nedione as a substrate according to the procedure reported
by Siiteri and Thompson [24]. The screening assay for de-
termination of IC50 values and the kinetic assay were car-
ried out according to the methods described in our previous
work [17,18]. Briefly, 20 ␮g of protein from the microsomes
and a 20-min or 5-min incubation period were used for the
screening and kinetic assays, respectively. The assays were
carried out at 37 ◦C in 67 mM phosphate buffer, pH 7.4, in
the presence of NADPH under air. Apparent Ki values were
calculated using non-linear regression analysis with GraFit
software [25].
3. Results and discussion
3.1. Chemistry
Treatment of 4␤-hydroxyandrost-5-en-17-one (3) [22]
with CH3I or CH2CH3I in the presence of Ag2O for 40 h
gave 4␤-methoxide (5) (25%) or 4␤-ethoxide (6) (56%)
(Fig. 2). A 1,3-diaxial relation of the 4␤-hydroxy group to
the C-19 angular methyl group would sterically prevent the
alkylation reaction, resulting in the low yield.
Epoxidation of androst-4-en-17␤-yl acetate (7) with
MCPBA produced a 6:1 mixture of 4␣,5␣-epoxide (8) and
its ␤-isomer (9) (Fig. 3). After separation of the mixture
with column chromatography, the ␣-epoxide (8) was sub-
jected to reaction with RMgX (R: methyl, ethyl, n-propyl,
benzyl, and 4-methoxybenzyl) in THF under reflux, and the
reaction products were purified by column chromatogra-
phy, giving the corresponding 4␤-substituted 5␣,17␤-diols
Fig. 3. Synthesis of 4␤-alkyl- and 4␤-phenyalkylandrost-5-ene steroids.

Table 1
Physicochemical properties of 4␤-substituted steroids (11–24)
Compound
R
Yield
(%)
Melting
Recryst.
IR (KBr),
(cm−1
MS, m/z (relative intensity)
1H NMR (CDCl3), δ
point (◦C) solvent
)
18-Me 19-Me Other signals
5␣,17␤-Diols
10
Methyl
Ethyl
29
34
24
26
168–170
107–109
131–133
176–178
Acetone
3436 (OH)
3398 (OH)
306 (M+, 9)
288 (17)
236 (100) 0.73(s) 1.00(s) 1.07 (3H, d, J = 8 Hz, 4␤-CH3),
3.63 (1H, m, 17␣-H)
11
12
13
EtOAc-hexane
320 (M+, 5)
273 (44)
334 (M+, 4)
302 (39)
255 (18)
316 (6)
287 (23)
236 (100)
0.72(s) 0.95(s) 0.89 (3H, t, J = 7 Hz, 4␤-CH2CH3),
3.63 (1H, t, J = 8 Hz, 17␣-H)
Propyl
Benzyl
Acetone-hexane 3484 (OH)
236 (100) 0.73(s) 0.96(s) 0.90 (3H, t, J = 7 Hz, 4␤-(CH2)2CH3),
3.63 (1H, t, J = 9 Hz, 17␣-H)
Acetone
Acetone
3453 (OH)
3480 (OH)
382 (M+, 2)
91 (100)
364 (14)
394 (7)
236 (79)
236 (21)
0.75(s) 1.07(s) 2.62 and 2.85 (1H each, d, J = 13 Hz, 4␤-
CH2Ph), 3.65 (1H, J = 8 Hz, 17␣-H),
7.13–7.29 (5H, m, aromatic protons)
14
4ꢀ-Methoxybenzyl 13
180–183
412 (M+, 6)
121 (100)
0.74(s) 1.06(s) 2.56 and 2.78 (1H each, t, J = 13 Hz,
4␤-CH2PhOCH3), 3.64 (1H, t, J = 8 Hz,
17␣-H), 3.78 (3H, s, 4␤- CH2PhOCH3),
6.81 and 7.05 (2H each, d, J = 9 Hz,
aromatic protons)
5␣-Hydroxy-17-ones
15
16
17
18
Methyl
58
73
43
96
119–121
135–138
85–88
Acetone-hexane 3472 (OH)
304 (M+, 16)
219 (24)
286 (16)
300 (8)
314 (8)
234 (100) 0.85(s) 1.02(s) 1.08 (3H, d, J = 8 Hz, 4␤-CH3)
234 (100) 0.84(s) 0.96(s) 0.89 (3H, t, J = 7 Hz, 4␤-CH2CH3)
234 (100) 0.85(s) 0.97(s) 0.90 (3H, t, J = 7 Hz, 4␤-(CH2)2CH3)
=
1724 (C O)
Ethyl
Acetone
3490 (OH)
318 (M+, 14)
219 (20)
=
1746 (C O)
Propyl
Benzyl
Acetone-hexane 3546 (OH)
332 (M+, 7)
219 (14)
380 (M+, 3)
219 (16)
=
1740 (C O)
158–160
Methanol
Methanol
3568 (OH)
362 (11)
91 (100)
234 (63)
0.87(s) 1.09(s) 2.64 and 2.86 (1H each, d, J = 13 Hz, 4␤-
CH2Ph), 7.13–7.29 (5H, m, aromatic
protons)
=
1733 (C O)
19
4ꢀ-Methoxybenzyl 76
157–160
3456 (OH)
410 (M+, 25)
121 (100)
392 (25)
271 (8)
0.86(s) 1.07(s) 2.57 and 2.79 (1H each, t, J = 13 Hz,
4␤−CH2PhOCH3), 3.78 (3H, s, 4␤–
CH2PhOCH3), 6.81and 7.05 (2H each, d,
J = 9 Hz, aromatic protons)
=
1737 (C O)
5-Ene-17-ones
286 (M+, 100)
300 (M+, 73)
314 (M+, 64)
271 (70)
285 (40)
299 (38)
271 (78)
256 (11)
0.88(s) 1.08(s) 1.18 (3H, d, J = 8 Hz, 4␤-CH3),
=
20
Methyl
92
92
72
84
158–160
110–113
108–110
169–171
Acetone
Methanol
Acetone
Methanol
1736 (C O)
5.34 (1H, s, 6-H)
=
21
22
23
Ethyl
1738 (C O)
271 (100) 0.88(s) 1.03(s) 0.84 (3H, t, J = 7 Hz, 4␤- CH2CH3),
5.28 (1H, s, 6-H)
=
Propyl
Benzyl
1740 (C O)
271 (100) 0.88(s) 1.03(s) 0.89 (3H, t, J = 7 Hz,
4␤-(CH2)2CH3), 5.29 (1H, m, 6-H)
253 (10)
=
1742 (C O)
362 (M+, 5)
91 (100)
0.89(s) 1.16(s) 2.74 and 2.90 (1H each, d, J = 10 Hz,
4␤-CH2Ph), 5.25 (1H, d, J = 3 Hz,
6-H), 7.14–7.31 (5H, aromatic protons)
4ꢀ-Methoxybenzyl 99
155–158
Methanol
1746 (C O)
392 (M+, 35)
270 (90)
122 (100) 0.89(s) 1.14(s) 2.67 and 2.83 (1H each, t, J = 10 Hz,
=
24
4␤−CH2PhOCH3), 3.79 (3H, s,
4␤-CH2PhOCH3), 5.24 (1H, s, 6-H),
6.82 and 7.07 (2H each, d, J = 9 Hz,
aromatic protons)

538
M. Nagaoka et al. / Steroids 68 (2003) 533–542
Table 2
Analysis data of 4␤-substituted steroids (11–24)
Compound
Formula
Calculated (%)
Found (%)
C
H
C
H
5␣,17␤-Diols
10
11
12
13
14
C
20H34O2
C21H36O2
C22 38O2
26H38O2
C27H40O3
78.45
78.69
78.98
81.62
78.59
11.18
11.32
11.45
10.01
9.77
78.45
78.87
79.07
81.51
78.31
11.10
11.44
11.48
9.89
H
C
10.09
5␣-Hydroxy-17-ones
15
16
17
18
19
C
C
20H32O2
21H34O2
78.90
79.19
79.46
82.06
78.98
10.59
10.76
10.91
9.54
79.12
79.03
79.32
81.95
78.92
10.33
10.73
10.94
9.49
C22H36O2
26H36O2
C27H38O3
C
9.33
9.30
5-Ene-17-ones
20
21
22
23
24
C
20H30O
83.86
83.94
84.01
86.13
82.60
10.56
10.74
10.90
9.45
83.92
83.81
84.17
86.22
82.78
10.52
10.70
10.87
9.41
C21
C22
C26
H32O
H34O
H34O
C27H36O2
9.24
9.19
was a NOE enhancement between 19-methyl protons
(δ = 1.00 ppm) and methyl protons (δ = 1.07 ppm) of the
4␤-methyl compound (10). The 4␤-substituted compounds
(10–14) were then treated with Jones reagent followed by
SOCl2 to produce 4␤-substituted 5-ene steroids (20–24).
Their spectroscopic and analysis data were consistent with
the assigned structures (Tables 1 and 2).
Fig. 4. Synthesis of 4-acyloxyandrost-4-ene and 5␣-androst-4-one steroids.
Treatment of 4␣,5␣-dihydroxy-5␣-androstan-17-one (25)
[20] with n-butyric anhydride, n-hexanoic anhydride, or ben-
zoyl chloride in pyridine at 60 ◦C gave the corresponding
4␣-ester derivative (26, 27, or 28). Dehydration of these
compounds with SOCl2 gave 4-n-butyrate (29), n-hexanate
(30), or benzoate (31), respectively, in a good yield, as ma-
jor products along with the corresponding 5-ene isomers
as minor products (Fig. 4). Spectral and exact MS data
of compounds (26–31) were consistent with their respec-
tively assigned structures (Table 3). The dehydration reaction
mechanism can be explained by the fact that the 4␤-axial
hydrogen atom of compounds (26–28) was chemically acti-
vated by the 4␣-ester substituent. 4-Acetoxy-4-ene steroids
(32 and 34) [20] were converted into the corresponding
5␣-4-one products (35 and 36) by treatment with a base.
tal microsomes was determined using a radiometric assay in
which tritiated water released from [1␤-3H]androstenedione
into the incubation medium during aromatization was mea-
sured. To characterize the nature of inhibitor binding to the
active site of aromatase, aromatization was measured at sev-
eral inhibitor and substrate concentrations. 4␤-Methoxide
(5), 4␤-ethoxide (6), 4␤-benzyl compound (23) and its
4ꢀ-methoxy derivative (24), and 19-hydroxy-5␣-4-one (36)
were extremely poor inhibitors; thus, their enzyme kinetics
were not determined. The results from the other studied
steroids were plotted on a typical Lineweaver-Burk Plot, and
all of the inhibitors examined, except for 4-acetoxy-19-ol
(34), showed a competitive type inhibition of aromatase.
In contrast, inhibitor (34) was a non-competitive inhibitor
of the enzyme (Fig. 5). 4-Acetoxy-4-ene steroid (32)
(Ki = 60 nM) was the most powerful competitive inhibitor
among the compounds. In these studies, the apparent Km
and Vmax values for androstenedione were about 33 nm and
125 pmol/min/mg of protein, respectively.
3.2. Biochemical properties
Inhibition of aromatase activity in human placental
microsomes by 4␤-alkoxides (5 and 6), 4␤-alkyl and
4␤-phenylalkyl compounds (20–24), 4-acyloxy-4-ene com-
pounds (29–32 and 34), and 5␣-4-one compounds (35 and
36) was examined in vitro by enzyme kinetics. The results
are given in Table 4. In addition to the above compounds,
the parent 4-ene and 5-ene steroids (1 and 2) and 5-en-4␤-ol
(3) are listed for comparison. Aromatase activity in placen-
The extremely low inhibitory activities of 4␤-alkoxides
(5 and 6) revealed that a tight binding of 4␤-ol (3) to the
active site (Ki = 25 nM) was due to a hydrophilic interaction
(hydrogen bonding) between the 4␤-hydroxy function and
a nucleophilic residue of the active site, as seen in the case

Table 3
Physicochemical properties of 4-ester-substituted steroids (26–31)
Compound
R
Yielda
(%)
Formula
Exact MS
Calculated
MS, m/z (relative intensity)
IR (neat)
(cm−1
1H NMR (CDCl3), δ
)
Found
18-Me
19-Me
Other signals
5␣-Hydroxy steroid
26
n-Butyrate
98
C23H36O4
C25H40O4
376.2614
376.2620
376 (M+, 33)
234 (57)
288 (100)
270 (32)
3512 (OH)
0.85(s)
1.00(s)
0.95 (3H, t, J = 7 Hz,
=
1737 (C O)
4␣-OCO(CH2)2CH3),
5.39 (1H, m, 4␤-H)
27
28
n-Hexanoate
78
80
404.2927
410.2457
404.2924
410.2462
404 (M+, 18)
305 (19)
392 (14)
288 (100)
288 (16)
3512 (OH)
0.85(s)
0.86(s)
1.00(s)
1.05(s)
0.89 (3H, m, 4␣-OCO(CH2)4CH3),
5.38 (1H, m, 4␤-H)
5.65 (1H, m, 4␤-H), 7.45 (2H, m,
3ꢀ- and 5ꢀ-H2),7.58 (1H, m, 4ꢀ-H),
8.02 (2H, m, 2ꢀ- and 6ꢀ-H2)
=
1737 (C O)
Benzoate
C
26H34O4
410 (M+, 10)
105 (100)
3505 (OH)
=
1718 (C O)
4-Ene steroid
29
n-Butyrate
38
C
23H34O3
358.2508
358.2501
358 (M+, 17)
288 (100)
274 (38)
1737 (C O)
0.89(s)
1.08(s)
0.98 (3H, t, J = 7 Hz,
=
4-OCO(CH2)2CH3)
386 (M+, 18)
392 (M+, 39)
288 (100),
270 (42)
273 (51)
105 (100)
0.88(s)
0.89(s)
1.08(s)
1.13(s)
0.91 (3H, m, 4-OCO(CH2)4CH3)
7.46 (2H, m, 3ꢀ- and 5ꢀ-H2),
7.59 (1H, m, 4ꢀ-H),
8.08 (2H, m, 2ꢀ- and 6ꢀ-H2)
=
1738 (C O)
30
31b
n-Hexanoate
Benzoate
34
74
C25H38O3
C26H32O3
386.2821
392.2351
386.2804
392.2359
=
1727 (C O)
a Oily compounds (26–31) were purified by silica gel column chromatography and/or preparative TLC.
b UV spectra: λmax nm (ε): 231 (12800).

540
M. Nagaoka et al. / Steroids 68 (2003) 533–542
Table 4
In vitro aromatase inhibition by 4␤-alkoxy- and 4␤-alkyl-5-ene steroids and 4-acyloxy-4-ene steroids
a
Steroids
IC50 (␮M)
Kib (nM)
Inhibitionc
4␤-Alkoxy-5-ene steroids
4␤-Methoxy (5)
4␤-Ethoxy (6)
42% inhibition at 10 ␮M
39% inhibition at 10 ␮M
NDd
ND
4␤-Alkyl-5-ene steroids
4␤-Methyl (20)
4␤-Ethyl (21)
4␤-n-Propyl (22)
4␤-Benzyl (23)
4␤-(4ꢀ-Methoxy)benzyl (24)
1.8
4.5
5.5
210
540
640
Competitive
Competitive
Competitive
ND
22% inhibition at 20 ␮M
33% inhibition at 20 ␮M
ND
4-Acyloxy-4-ene steroids
4-Acetate (32)
4-Acetoxy-19-ol (34)
4-n-Butylate (29)
4-n-Hexanoate (30)
4-Benzoate (31)
0.66
3.60
1.50
1.80
0.64
60
700
170
360
70
Competitive
Non-competitive
Competitive
Competitive
Competitive
5␣-4-One steroids
10␤-Methyl (35)
19-Hydroxy (36)
2.40
230
Competitive
ND
20% inhibition at 10 ␮M
For comparison
4-Ene (1)
5-Ene (2)
0.060
0.66
0.28
6.8
120
25
Competitive
Competitive
Competitive
5-Ene-4␤-ol (3)
a 300 nM of [1␤-3H]androstenedione and 20 ␮g of protein from human placental microsomes were used.
b Inhibition constant was obtained by Dixon plot.
c Inhibition type was determined by Lineweaver-Burk plot. In these experiments, the apparent Km for the substrate androstenedione was found to be
about 33 nM.
d ND: not determined.
of 4␤,19-dihydroxy-4-ene steroid (4), which is one of the
most powerful steroidal aromatase inhibitors [20]. The in-
troduction of a methyl moiety at the 4␤-position of the 5-ene
steroid (2) decreased the affinity for aromatase by about
one-half (Ki: 210 nM for 4␤-methyl (20) versus 120 nM for
steroid 2); however, steroid (20) was still a fair inhibitor of
the enzyme. Elongation of the methyl group up to n-propyl
decreased the affinity in proportion to the carbon number of
the alkyl chain (Ki for n-propyl (17): 640 nM), as seen in the
6␤-alkyl-3-deoxy-4-ene steroid series [18]. 4␤-Benzyl and
4␤-(4ꢀ-methoxy)benzyl compounds (23 and 24), which have
a terminal phenyl function at the 4␤-substituent, were ex-
tremely poor inhibitors of aromatase. On the basis of these
results, it is likely that the binding pocket (active site) of
Fig. 5. Lineweaver-Burk plots of inhibition of human placental aromatase by 4-acetoxyandrost-4-ene (32) (A) and its 19-hydroxy derivative (34) (B) with
androstenedione as a substrate. Each point represents the mean of two determinations. The inhibition experiments with the other inhibitors examined in
the study gave essentially similar results to that for inhibitor (32) (data not shown).

M. Nagaoka et al. / Steroids 68 (2003) 533–542
541
aromatase in the region of the 4␤-position of 5-ene steroid
(2) would tolerate a small-sized, lipophilic function, such as
a methyl group, in addition to a hydrophilic function, such as
a hydroxy group [18,20]. The extremely low affinity of the
alkoxides may be due to an electronic reason rather than a
steric one, since the 4␤-alkyl compounds (21 and 22), while
have similar-sized 6␤-substituents to the alkoxides, can ac-
cess the binding site with a fair affinity.
Introduction of an acetoxy group at C-4 of the 4-ene
steroid (1) resulted in decreased affinity for aromatase (Ki:
6.8 nM for steroid (1) versus 60 nM for 4-acetate 32); how-
ever, the Ki value of acetate (32) was comparable to the Km
for the substrate androstenedione. Elongation of the acetoxy
group up to n-hexanoate decreased the affinity in relation to
carbon number (Ki: 360 nM for hexanoate 30), as seen in the
4␤-alkyl steroid series. Furthermore, introduction of a hy-
droxyl group at C-19 of the 4-acetate (32) lowered the affin-
ity to a great extent (Ki: 700 nM for 19-ol 34) and changed
the inhibition from a competitive to a non-competitive type.
A similar change in binding mechanism was previously ob-
served by introduction of hydroxy and oxo groups to the
C-19 methyl group of 6␤-methylandrostenedione [26]. It
was previously reported that aromatase has two binding
sites, one for androstenedione binding and the other for its
16␣-hydroxy analog [27–29].
5␣-Reduced-4-one steroid (35) was a fair competitive
inhibitor of aromatase (Ki: 230 nM), although its affinity
to aromatase was about one-half that of its enone ana-
log androst-5-ene-4,17-dione [19], as seen in a series of
androst-4-en-3-one, androst-4-en-6-one, and androst-4-ene-
3,6-dione steroids [30].
The present findings regarding the effects of an alkyl
group at 4␤ of 5-ene compound (2) and an acyloxy group
at C-4 of 4-ene steroid (1) may be useful for understanding
spatial and electronic aspects of the binding (active) site of
aromatase as well as for developing an effective aromatase
inhibitor.
[3] Yoshida N, Osawa Y. Purification of human placental aromatase
cytochrome P-450 with monoclonal antibody and its characterization.
Biochemistry 1991;30:3003–10.
[4] Brodie AMH, Njar VCO. Aromatase inhibitors in advanced breast
cancer: mechanism of action and clinical implications. J Steroid
Biochem Mol Biol 1998;66:1–10.
[5] Johnston JO. Aromatase inhibitors. Crit Rev Biochem Mol Biol
1998;33:375–405.
[6] Abul-Hajj YJ. Synthesis and evaluation of 4-(substituted thio)-4-
androstene- 3,17-dione derivatives as potential aromatase inhibitors.
J Med Chem 1986;29:582–4.
[7] Abul-Hajj YJ, Liu XP, Hedge M. Synthesis and evaluation
of 4-substituted-4- androstene-3,17-dione derivatives as aromatase
inhibitors. J Steroid Biochem Mol Biol 1995;54:111–9.
[8] Liu XP, Lambert DM, Abul-Hajj YJ. Probing the hydrophobic pocket
of the active site of aromatase with 4-phenoxy-7␣-(phenylthio)-4-
androstene-3,17-dione. J Med Chem 1995;38:4135–8.
[9] Numazawa M, Oshibe M. 6-Arylandrost-4-ene-3,17-diones as
aromatase inhibitors. Synthesis and structure–activity relationships.
J Med Chem 1994;37:1312–9.
[10] Numazawa M, Oshibe M. Further studies on 6-alkylandrost-4-ene-
3,17-diones as aromatase inhibitors: elongation of the 6-alkyl chain.
Steroids 1995;60:506–11.
[11] Numazawa M, Yamaguchi S. 6-Phenylaliphatic-substituted androst-
4-ene-3,17-diones as aromatase inhibitors: structure–activity
relationships. J Steroid Biochem Mol Biol 1998;67:41–8.
[12] Brueggemeier RW, Floyd EE, Counsell RE. Synthesis and
biochemical evaluation of inhibitors of estrogen biosynthesis. J Med
Chem 1978;21:1007–111.
[13] Darby MV, Lovett JA, Brueggemeier RW, Groziak MP, Counsell
RE. 7␣-Substituted derivatives of androstenedione as inhibitors of
estrogen biosynthesis. J Med Chem 1985;28:803–7.
[14] O’Reilly JM, Li N, Duax WL, Brueggemeier RW. Synthesis,
structure elucidation, and biochemical evaluation of 7␣- and 7␤-
arylaliphatic-substituted androst-4-ene-3,17-diones as inhibitors of
aromatase. J Med Chem 1995;38:2842–50.
[15] Numazawa M, Mutsumi A, Hoshi K, Koike R. 19-Hydroxy-
4-androsten-17-one: potential competitive inhibitor of estrogen
biosynthesis. Biochem Biophys Res Commun 1989;160:1009–14.
[16] Numazawa M, Mutsumi A, Hoshi K, Oshibe M, Ishikawa E,
Kigawa H. Synthesis and biochemical studies of 16- or 19-substituted
androst-4-enes as aromatase inhibitors. J Med Chem 1991;34:2496–
504.
[17] Numazawa M, Mutsumi A, Tachibana M, Hoshi K. Synthesis of
androst-5-en-7- ones and androsta-3,5-dien-7-ones and their related
7-deoxy analogs as conformational and catalytic probes for the active
site of aromatase. J Med Chem 1994;37:2198–205.
[18] Numazawa M, Kamiyama T, Tachibana M, Oshibe M. Synthesis and
structure–activity relationships of 6-substituted androst-4-ene analogs
as aromatase inhibitors. J Med Chem 1996;39:2245–52.
[19] Numazawa M, Tachibana M, Tateda Y. 4-Oxygenated androst-5-
en-17-ones and their 7-oxo derivatives as aromatase inhibitors. J
Steroid Biochem Mol Biol 1996;58:431–8.
[20] Numazawa M, Yamada K, Nitta S, Sasaki C, Kidokoro K. Role
of hydrophilic interaction in binding of hydroxylated 3-deoxy C19
steroids to the active site of aromatase. J Med Chem 2001;44:4277–
83.
Acknowledgements
This work was supported in part by a Grant-in-Aid for
Scientific Research from the ministry of Education, Science,
Sports, and Culture of Japan. We are grateful to Dr. Hideo
Imaizumi of Imaizumi Hospital, Sendai, for generously sup-
plying human term placenta.
[21] Numazawa M, Nagaoka M, Morio M, Kamiyama T. 19-Oxygenations
of 3-deoxy androgens, potent competitive inhibitors of estrogen
biosynthesis, with human placental aromatase. J Steroid Biochem
Mol Biol 1999;71:173–9.
References
[1] Thompson Jr EA, Siiteri PK. Involvement of human placental
[22] Hanson JR, Reese PB, Wadsworth HJ. Allylic acetoxylation of
ꢀ5-steroids at C-4. J Chem Soc Perkin Trans 1984;1:2941–4.
microsomal cytochrome P-450 in aromatization.
J Biol Chem
1974;249:5373–8.
[23] Ryan KJ. Biological aromatization of steroids.
J Biol Chem
[2] Kellis Jr JT, Vickery LE. Purification and characterization of human
1959;234:268–72.
placental aromatase cytochrome P-450. J Biol Chem 1987;262:4413–
[24] Siiteri PK, Thompson EA. Studies of human placental aromatase. J
20.
Steroid Biochem 1975;6:317–22.

542
M. Nagaoka et al. / Steroids 68 (2003) 533–542
[25] GraFit version 5.0.3, Erithacus Software Limited, Surrey, UK, 2001.
[26] Numazawa M, Yoshimura A. Synthesis and biochemical studies of
19-oxygenated derivatives of 6␣- and 6␤-methylandrostenediones as
catalytic probes for the active site of aromatase. Biol Pharm Bull
2000;23:1059–65.
placental microsomes. J Steroid Biochem Mol Biol 1989;33:439–
48.
[29] Numazawa M, Tachibana Y, Mutsumi A, Yoshimura A, Osawa Y.
Aromatization of 16␣-hydroxyandrostenedione by human placental
microsomes: effect of preincubation with suicide substrates of
[27] Bellino FL, Osawa Y. Estrogen biosynthesis. V. Evidence of the
direct aromatization of testosterone and different aromatization sites
for testosterone and androstenedione in human placental microsomes.
Biochemistry 1974;13:1925–31.
[28] Purohit A, Oakey RE. Evidence for separate sites for aromatization
of androstenedione and 16␣-hydroxyandrostenedione in human
androstenedione aromatization. J Steroid Biochem Mol Biol
2002;81:165–72.
[30] Numazawa M, Mutsumi A, Tachibana M, Nagaoka M. Biochemical
studies of estr-4-ene-3,6,17-trione and 5␣-androstan-17-ones with or
without a carbonyl function at C-3 and/or C-6 as aromatase inhibitors.
Biol Pharm Bull 1995;18:555–8.