Structure-activity relationships of new A,D-ring modified steroids as aromatase inhibitors: Design, synthesis, and biological activity evaluation

Article abstract of DOI:10.1021/jm050129p

Inhibition of aromatase is an efficient approach for the prevention and treatment of breast cancer. New A,D-ring modified steroid analogues of formestane and testolactone were designed and synthesized and their biochemical activity was investigated in vitro in an attempt to find new aromatase inhibitors and to gain insight into their structure-activity relationships (SAR). All compounds tested were less active than formestane. However, the 3-deoxy steroidal olefin 3a and its epoxide derivative 4a proved to be strong competitive aromatase inhibitors (K_i = 50 and 38 nM and IC ₅₀ = 225 and 145 nM, respectively). According to our findings, the C-3 carbonyl group is not essential for anti-aromatase activity, but 5α-stereochemistry and some planarity in the steroidal framework is required. Furthermore, modification of the steroidal cyclopentanone D-ring, by construction of a δ-lactone six-membered ring, decreases the inhibitory potency. From the results obtained, it may be concluded that the binding pocket of the active site of aromatase requires planarity in the region of the steroid A,B-rings and the D-ring structure is critical for the binding.

Full text of DOI:10.1021/jm050129p

J. Med. Chem. 2005, 48, 6379-6385
6379
Structure-Activity Relationships of New A,D-Ring Modified Steroids as
Aromatase Inhibitors: Design, Synthesis, and Biological Activity Evaluation
Margarida M. D. S. Cepa,† Elisia´rio J. Tavares da Silva,‡ Georgina Correia-da-Silva,†
Fernanda M. F. Roleira,‡ and Nate´rcia A. A. Teixeira*,†
Biochemistry Laboratory, Faculty of Pharmacy, University of Oporto, 4050-047 Oporto, Portugal, Institute for Molecular and
Cellular Biology, 4150-180 Oporto, Portugal, and Centro de Estudos Farmaceˆuticos, Pharmaceutical Chemistry Laboratory,
Faculty of Pharmacy, University of Coimbra, 3000-295 Coimbra, Portugal
Received February 10, 2005
Inhibition of aromatase is an efficient approach for the prevention and treatment of breast
cancer. New A,D-ring modified steroid analogues of formestane and testolactone were designed
and synthesized and their biochemical activity was investigated in vitro in an attempt to find
new aromatase inhibitors and to gain insight into their structure-activity relationships (SAR).
All compounds tested were less active than formestane. However, the 3-deoxy steroidal olefin
3a and its epoxide derivative 4a proved to be strong competitive aromatase inhibitors (Ki ) 50
and 38 nM and IC50 ) 225 and 145 nM, respectively). According to our findings, the C-3 carbonyl
group is not essential for anti-aromatase activity, but 5R-stereochemistry and some planarity
in the steroidal framework is required. Furthermore, modification of the steroidal cyclopen-
tanone D-ring, by construction of a δ-lactone six-membered ring, decreases the inhibitory
potency. From the results obtained, it may be concluded that the binding pocket of the active
site of aromatase requires planarity in the region of the steroid A,B-rings and the D-ring
structure is critical for the binding.
Introduction
Estrogens are required for normal tissue growth and
development, but they are also responsible for the
promotion of certain tumors, such as prostatic hyper-
plasia, prostate, endometrial, and especially breast
cancers.1 In fact, 30-50% of breast cancers are consid-
ered to be estrogen-dependent,2,3 and regression can be
achieved by reducing blood and tissue estrogen levels.
The highest incidence of breast cancer occurs in post-
menopausal women,4 whose ovarian function and hy-
pophyseal control of estrogen production has ceased. In
these women, estrogen biosynthesis occurs mostly in
Figure 1. Steroids clinically used as aromatase inhibitors.
peripheral tissues such as liver, skin, muscle, and
adipose tissue.5-7 For this reason, total blockage of
been shown to be clinically relevant in the management
of other estrogen-dependent pathological processes such
as endometriosis,12 gynecomastia, prostatic hyperplasia,
and prostate cancer.13 For these reasons, several ste-
roidal aromatase inhibitors, based on the structure of
the natural substrate androstenedione, have been de-
signed and synthesized in several laboratories. Exemes-
tane and the well-established formestane (Figure 1) are
two examples of these kinds of compounds that are
known to be very efficient in the treatment of patients
in whom other anti-estrogen therapies failed.8,14
In the present study our focus is on the design,
synthesis, and biochemical evaluation of a new series
of substituted steroids containing modifications in the
A- and D-rings. The new compounds were obtained by
chemical modifications to the structure of the natural
enzyme substrate, androstenedione, to achieve new
aromatase inhibitors in order to further study the
structure-activity relationships (SAR). The studied
SAR were correlated with three main features, namely,
the A-ring substitution pattern and the planarity of this
estrogen biosynthesis is easier to achieve by chemical
means rather than by surgical therapy.8 Two main
chemical approaches have been successfully applied as
endocrine therapy of estrogen-dependent breast cancer.
One is directed to the estrogen receptor by use of
selective estrogen receptor modulators (SERMs), among
which tamoxifen is the most popular one. The other is
directed to aromatase, the key enzyme responsible for
estrogen biosynthesis.9,10 Since estrogens are formed in
the last step of the biosynthetic sequence of steroid
production, aromatase inhibition provides a basis for the
development of selective chemotherapies for treatment
and prevention of estrogen-sensitive breast cancers.11
Furthermore, the therapeutic potential of aromatase
inhibitors stretches beyond postmenopausal women
with breast cancer, because these inhibitors have also
* To whom correspondence should be addressed: phone
+351222078907; fax +351222003977; e-mail natercia@ff.up.pt.
† University of Oporto and Institute for Molecular and Cellular
Biology.
‡ University of Coimbra.
10.1021/jm050129p CCC: $30.25 © 2005 American Chemical Society
Published on Web 09/09/2005

6380 Journal of Medicinal Chemistry, 2005, Vol. 48, No. 20
Cepa et al.
Scheme 1. Synthesis of Aromatase Inhibitors from
androstenedione 1a
we embarked on the design, synthesis, and biochemical
evaluation of some C-4 oxo- (2), C-4 keto- (6 and 10),
C-4 hydroxy- (5a, 5b, and 8), and C-4 hydroxycyano-
(11) substituted androstanes (Schemes 1 and 2). In the
same way, the C-3 carbonyl group of the substrate
androstenedione has been thought to be important to
the binding of aromatase and is also present in many
inhibitors. However, some recent studies have shown
that this group could be removed if some planarity were
maintained in the steroid A-ring.16,18 For this reason,
we prepared the 3-deoxy steroids 3a, 4a, and 4b
(Schemes 1 and 2) having olefin and epoxy groups, that
confer the desired planarity to the A-ring, and evaluated
their inhibitory activity. We were also interested in
studying the effect of the C-5 stereochemistry of the A/B-
ring fusion on drug-receptor interaction, so the 5R/â-
epimers 4a/4b and 5a/5b were prepared and evaluated.
With respect to the steroid D-ring, there are not many
modifications reported in the literature. However, some
δ-lactone D-ring steroids such as testolactone, the first
steroidal aromatase inhibitor clinically used for breast
cancer treatment (Figure 1),19 have also shown anti-
a Reagents and conditions: (i) MeOH, KMnO4, KOH, 15 °C, 10
min; (ii) CH3COOH, Zn dust, 118 °C, 15 min.
ring, the C-5 stereochemistry in the A/B-ring fusion, and
the D-ring structure. Numazawa and co-workers15-18
have performed several SAR studies in steroid aro-
matase inhibitors indicating the existence of a binding
pocket in the active site of the enzyme, which can
accommodate several chemical groups placed at the C-4
position of the steroids. According to this information,
Scheme 2. Synthesis of Aromatase Inhibitors from 5R- and 5â-Olefins 3a and 3ba
a (i) H2O2, HCO2H,CH2Cl2, room temperature (RT), 6 h; (i′) HCO2H, H2O2, RT, 1 h; (ii) HCO2H, RT, 30 min; (iii) AlO3, oxone, CH2Cl2,
MeOH, RT, 6 h; (iv) MMPP, MeOH, H2O, RT, 72 h; (v) DMSO, TFAA, -60 °C, 3 h, and then Et3N, -60 °C, 15 min; (vi) HCl, CH3COOH,
RT, 24 h; (vii) (CH3)2C(OH)CN, Et3N, RT, 6 h.

A,D-Ring Modified Steroids as Aromatase Inhibitors
Journal of Medicinal Chemistry, 2005, Vol. 48, No. 20 6381
Table 1. Aromatase Inhibition by A/D-Ring Substituted
aromatase activity. On the basis of these, we decided
to prepare and evaluate the δ-lactone D-ring steroids
8, 9, 10, and 11 (Scheme 2). Furthermore, successful
efforts have been made to combine in the same structure
the diosphenol A-ring of formestane and the δ-lactone
D-ring of testolactone, as observed in compound 9.
Steroidsa
compound
inhibition (%) for 2 µM
2
11.0 ( 1.9
95.9 ( 0.6
96.4 ( 0.1
14.6 ( 1.6
34.6 ( 1.0
10.1 ( 1.8
83.3 ( 1.3
99.1 ( 0.04
8.1 ( 1.8
3a
4a
4b
5a
Results
5b
6
Chemistry. All steroids were prepared from andros-
tenedione 1 by the routes depicted in Schemes 1 and 2.
The A-ring lactone aldehyde 2 was first obtained as a
byproduct in a synthetic route for aromatase inhibitors
previously developed in our laboratory. It was formed
through a rearrangement, which occurred after con-
trolled oxidation of 1 in methanol with alkaline potas-
sium permanganate (Tavares da Silva, unpublished
results). Clemmenson-type reduction of 1 with zinc dust
in acetic acid solution afforded the isomeric mixture (2.3:
1, by NMR) of the 5R- and 5â-olefins 3a and 3b in
almost quantitative yield. Separation of the pure 3a
(60%) from the mixture was achieved by n-hexane
crystallization.20
7 (formestane)
8
9
10
11
72.5 ( 1.7
16.6 ( 1.5
17.6 ( 1.6
a Concentrations of 40 nM [1â-3H]androstenedione, 20 µg of
protein from human placental microsomes, and 2 µM inhibitor and
15 min incubation were used. Results are the mean of at least
three independent experiments done in triplicate.
3-hydroxy-D-homo-17a-oxa-5R-androsta-2-ene-4,17-di-
one,22 which after acid-catalyzed isomerization yielded
quantitatively the thermodynamic diosphenol 9. Oxida-
tion of 8 in heterogeneous phase with the oxone alumina
system (also used previously for the diol 5a) gave the
R-ketol 10. Treatment of 10 with acetone cyanohydrin
and triethylamine provided the lactone cyanohydrin
derivative 11.
Treatment of 5R-androst-3-en-17-one 3a (Scheme 2)
with performic acid generated in situ led, after workup,
to the trans-diaxial diol, 3R,4â-dihydroxy-5R-androstan-
17-one 5a, in 96% yield. The 3R,4R-epoxide 4a was also
obtained from the reaction of 3a with performic acid but
with methylene chloride as reaction solvent instead. In
this case, due to the low solubility of the resulting formic
acid in this solvent, further cleavage of the epoxide
intermediate was prevented and 4a was the only
product formed, also in near quantitative yield. Epoxide
4b was prepared as described for 4a but with a 1:3
mixture of 3a and 3b as starting material. The above-
mentioned mixture has been used as we could not
isolate the pure 3b from it. Under these conditions a
1:3 mixture of epoxides 4a and 4b was obtained, from
which 4b was easily isolated by crystallization. Treat-
ment of 4b with 90% formic acid solution afforded a
mixture of two other diols and of the desired 3,4-trans-
diaxial diol 5b, which was isolated by crystallization or
column chromatography in 48% yield.20 Swern-type
oxidation of 5a with dimethyl sulfoxide, activated with
trifluoroacetic anhydride, gave quantitatively kinetic
diosphenol, the 3-hydroxy-5R-androst-2-ene-4,17-di-
one,21 which by acid- or base-catalyzed isomerization
with HCl/CH3COOH or NaOMe/MeOH led to the de-
sired thermodynamic diosphenol 4-OHA 7 as the only
product, formed in 80% yield. The same Swern-type
oxidation applied to 5â-diol 5b afforded a mixture with
a variable composition (1:2 to 2:1, by NMR) of a
5â-kinetic diosphenol, the 3-hydroxy-5â-androst-2-ene-
4,17-dione,20 and an unstable triketone, the 5â-andros-
tane-3,4,17-trione,20 which after acid- or base-catalyzed
isomerization were also converted in similar yield to
diosphenol 7. Treatment of diol 5a in methylene chloride
with oxone catalyzed with alumina in the heterogeneous
phase allowed the selective oxidation of the C-4 hydroxyl
group, affording the R-ketol 6. Baeyer-Villiger oxidation
of the 17-keto group of 5a with monoperoxyphthalic acid
magnesium salt hexahydrate (MMPP) yielded the di-
hydroxylactone derivative 8 in 90-94% yield. Further
oxidation of 8 with dimethyl sulfoxide activated with
trifluoroacetic anhydride gave a kinetic diosphenol, the
Biochemical Properties. Inhibition of aromatase
activity by the newly prepared A- and D-ring modified
steroids was evaluated in human placental microsomes
by a radiometric assay in which tritiated water, released
from [1â-3H]androstenedione into the incubation me-
dium, was used as an index of estrogen formation.23 The
results of these studies are shown as percentage of
inhibition for all compounds at 2 µM, relative to an
assay in the absence of an inhibitor (Table 1). The well-
known aromatase inhibitor 4-hydroxyandrostenedione
(Figure 1), with an inhibitory capacity of 99.1% ( 0.04%,
was used as reference. For the inhibitors showing anti-
aromatase activity equal to or higher than 70% (3a, 4a,
6, and 9), IC50 was determined in saturating concentra-
tions (200 nM) of androstenedione. In addition, the type
of binding to the active site of aromatase was character-
ized and the apparent inhibition constant (Ki) was
determined by use of different inhibitor and substrate
concentrations. The inhibition kinetic studies were
performed with human placental microsomes and the
results were plotted as Lineweaver-Burk plots. The
inhibition type of all compounds tested was competitive.
A representative Lineweaver-Burk plot of aromatase
inhibition is shown in Figure 2 for compound 9. The
apparent inhibition constants (Ki), which characterize
the enzyme affinity, were obtained by Dixon plots
(Figure 3). The IC50 value for formestane was evaluated
in our test system and used for comparison. This
compound presented higher inhibitory activity than all
the synthesized compounds. The IC50, Ki, type of inhibi-
tion and Km/Ki ratios for the more active inhibitors are
shown in Table 2. In these studies, the apparent Km and
Vmax values for aromatase with androstenedione as
substrate were 57.5 nM and 48 pmol min-1 (µg of
protein)-1, respectively.
Discussion
Compound 3a, the 5R-epimer of the 3,4-olefin, showed
95.9% ( 0.6% rate of inhibition. It is a strong inhibitor

6382 Journal of Medicinal Chemistry, 2005, Vol. 48, No. 20
Cepa et al.
also a strong competitive inhibitor of aromatase. Fur-
thermore, the 3,4-epoxide derivative 4a, which has the
same stereochemistry requirements in the A-ring, is also
a strong competitive aromatase inhibitor. These results
stress the importance of A-ring planarity in the interac-
tion with the active site of the enzyme. The lack of a
C-3 carbonyl group in these compounds also indicates
that this feature is not a determinant for inhibition, as
previously observed by Numazawa.16,24 Compound 5a,
the 5R-epimer of a 3,4-diol, showed an inhibition of
34.6% ( 1.0%, and its 4-keto derivative, the R-ketol 6,
showed a significant increase in the inhibitory capacity
of 83.3% ( 1.3%. These results demonstrate, once more,
the importance of the planarity in the A-ring for the
inhibition, now conferred by a C-4 carbonyl group in 6.
Compound 2, a 4-oxa steroid with a 5R-aldehyde group
showed only 11.0% ( 1.9% inhibition. In this way, more
studies should be performed in order to understand the
influence of 4-oxa and 5R-aldehyde groups in the anti-
aromatase activity.
Figure 2. Lineweaver-Burk plot of inhibition of human
placental aromatase by inhibitor 9 with androstenedione as
substrate. Each point represents the mean of three determina-
tions. The inhibition experiments with the other compounds
gave similar plots to the results shown for 9.
The effect of the C-5 stereochemistry on the inhibition
of aromatase was studied for the 5R/â-epimers 4a/4b
and 5a/5b. Within the 3,4-epoxide derivatives 4a and
4b, the 5R-substituted steroid 4a showed 96.4% aro-
matase inhibition in contrast to the corresponding 5â-
epimer 4b, which had a dramatic reduction in inhibition
(14.6% ( 1.6%). Within compounds 5a and 5b, the 5R-
epimer of the 3,4-diol 5a inhibited aromatase 34.6%,
whereas the correspondent 5â-epimer 5b showed a
decrease in the inhibitory activity (10.1% ( 1.8%). These
results indicate that the angle formed in the A/B-ring
junction is crucial in order to better fit the enzymatic
active site. The 5â-epimers, in which A- and B-rings are
almost perpendicular, will not be well accommodated
by the enzyme.
Figure 3. Dixon plots to determine the apparent inhibition
constant (Ki) for inhibitor 9.
The effect of the D-ring modification in the inhibition
of aromatase was observed in steroids 8, 9, and 10, in
which six-membered δ-lactone rings were introduced.
Compound 9, combining the diosphenol A-ring of the
potent inhibitor formestane and the δ-lactone D-ring of
testolactone (Figure 1), showed 72.5% aromatase inhibi-
tion and an IC50 ) 1.85 µM, compared to 99.1% ( 0.04%
and IC50 ) 0.042 µM for formestane. Similar reduction
in the inhibitory capacity was observed for the pair of
compounds 6 and 10, which showed inhibitory activities
of 83.3% ( 1.3% and 16.6% ( 1.5%, respectively.
Compound 6 was a competitive inhibitor of aromatase,
with IC50 ) 670 nM. It blocks the enzyme activity to a
much greater extent than the corresponding D-ring
lactone compound 10 (IC50 > 100 µM). In the same way,
compound 5a showed an inhibitory activity of 34.6% (
1.0%, while the introduction of the δ-lactone ring,
yielding derivative 8, strongly affects inhibition (8.1%
( 1.8%). These results suggest that the inhibitory
activity of testolactone is probably not due to its δ-lac-
tone D-ring. Compound 11, the 4-cyano derivative of 8,
showed a slight increase in the inhibitory activity (17.6%
( 1.6%) compared to its precursor. The chemically
reactive cyano group in the C-4 position of this com-
pound is also present in the most recent third genera-
tion cyano azole-type aromatase inhibitors,25 revealing
it to be an important feature for enzyme inhibition. This
result also indicates that the C-4 position is a critical
region for the interaction of steroids with aromatase.
Table 2. In Vitro Aromatase Inhibition
a
IC50
Kib
(nM)
type of
rel affinity
inhibitor
(µM)
inhibitionc
(Km/Ki)
9
1.850
0.225
0.145
0.670
0.042
660
50
competitive
competitive
competitive
competitive
0.087
1.15
1.51
0.39
3a
4a
6
38
147
7 (formestane)
a Concentrations of 200 nM [1â-3H]androstenedione and 20 µg
of protein from human placental microsomes were used. b Inhibi-
tion constant was obtained by Dixon plot. In these experiments,
the Km for the substrate androstenedione was found to be 57.5
nM. c The inhibitors studied showed competitive inhibition on the
basis of analysis of the Lineweaver-Burk plot.
of aromatase with an IC50 ) 225 nM and has higher
affinity for the enzyme than its natural substrate
androstenedione (Ki ) 50 nM versus Km ) 57.5 nM,
respectively). Compound 4a, the 5R-epimer of the 3,4-
epoxide, showed 96.4% ( 0.1% inhibition of aromatase,
revealing it to be the best competitive inhibitor among
all the compounds studied in this work. This epoxide
4a presented an IC50 ) 145 nM and was found to bind
the enzyme with an affinity 1.5 times higher (Ki ) 38
nM) than that of the natural substrate (57.5 nM).
Numazawa et al.16 have already reported that 3-deoxy-
4-ene and 3-deoxy-5-ene derivatives of androstenedione
are competitive inhibitors of aromatase. In this work
we pointed out that, in addition to these derivatives,
the 3-deoxy-3-ene derivative 3a of androstenedione is

A,D-Ring Modified Steroids as Aromatase Inhibitors
Journal of Medicinal Chemistry, 2005, Vol. 48, No. 20 6383
3r,4r-Epoxy-5r-androstan-17-one (4a). The title com-
pound 4a was prepared from oxidation of olefin 3a (67.5 mg,
0.25 mmol) in methylene chloride with performic acid gener-
ated in situ with 30% hydrogen peroxide (0.05 mL, 0.5 mmol)
and 90% formic acid (0.05 mL, 1.18 mmol), according to the
reaction conditions previously described.20 Epoxide 4a (75 mg,
96%) was obtained as the only isolated product identified.
3â,4â-Epoxy-5â-androstan-17-one (4b). The title epoxide
4b was prepared as described above for 4a but with a mixture
of olefins 3a and 3b (1:3 by NMR) as the starting material.
Under these conditions, a 1:3 mixture of 4a and 4b was
obtained, from which 4b was isolated by slow diethyl ether
crystallization.
Conclusions
Our data showed that modification in the A- and
D-rings of androstenedione provides potent, easily
available, and structurally simple aromatase inhibitors.
Nevertheless, none of the compounds prepared was
shown to be a stronger inhibitor than formestane. Only
specific changes were allowed in the A-ring of steroids
in order to obtain potent inhibitors. The C-3 carbonyl
is not a determinant group for the interaction of the
studied steroids with the enzyme and for its inhibition.
In fact, the 3-deoxy compounds 3a and 4a were dem-
onstrated to be strong competitive inhibitors of aro-
matase. They have also shown higher affinity toward
the enzyme than its natural substrate androstenedione.
These results suggest that a 3,4-double bond or a 3,4-
epoxide group will confer the A-ring planarity required
for anti-aromatase activity, according to the previous
observations by Numazawa and co-workers.15-18,24 Ep-
oxide 4a proved to be the most potent inhibitor, reveal-
ing that, in addition to planarity, the oxygen atom near
the C-4 region will be important for activity. A C-4
carbonyl group, as observed in 6, may also contribute
to the above-mentioned requirements for anti-aromatase
activity. In addition to planarity in the A-ring, planarity
in the A/B-ring junction of the steroids seems also to be
important for enzyme inhibition. Our results therefore
stress that the 5R-stereochemistry is required for inhi-
bition. Concerning the D-ring, the replacement of the
original steroid cyclopentanone ring by a six-membered
δ-lactone ring reduces the inhibitory activity. The
magnitude of this reduction seems to be also dependent
on the substitution pattern in the A-ring, as observed
for compounds 8, 9, and 10. This modification in the
D-ring may induce structural alterations in the steroidal
framework, affecting their ability to bind to the active
site of the enzyme.
3r,4â-Dihydroxy-5r-androstan-17-one (5a). The title 5R-
vic-diol 5a (75 mg, 96%) was prepared from olefin 3a (67.5
mg, 0.25 mmol) by oxidation with the in situ generated
performic acid and isolated according to the reaction conditions
previously described.21
3â,4r-Dihydroxy-5â-androstan-17-one (5b). The title 5â-
vic-diol 5b was obtained in a mixture with two other diols, by
cleavage of epoxide 4b (200 mg, 0.69 mmol) with formic acid
and was separated (102 mg, 48%) by column chromatography,
according to the reaction conditions previously described.20
3r-Hydroxy-5r-androstane-4,17-dione (6). To a solution
of 3R,4â-dihydroxy-5R-androstan-17-one 5a (50 mg, 0.16 mmol)
in methylene chloride with some drops of methanol (1.7 mL)
were added wet alumina (335 mg) and oxone (252 mg, 0.41
mmol). The suspension was sonicated under inert atmosphere
in a cleaning bath (“Bandelin”, 450 W, 35 kHz) at 22 °C until
the steroid was consumed [30 min, thin-layer chromatogtaphy
(TLC)]. After dilution with methylene chloride, the suspension
was filtered, washed with aqueous HCO3Na and water, dried
over magnesium sulfate, and evaporated to dryness to give a
crude extract (43 mg, 87%) from which the title compound was
isolated by crystallization.
4-Hydroxyandrost-4-ene-3,17-dione (7): (i) Oxidation
of Diols 3r,4â-Dihydroxy-5r-androstan-17-one (5a) and
3â,4r-Dihydroxy-5â-androstan-17-one (5b). Swern-type
oxidation of the 5R-vic-diol 5a (203 mg, 0.66 mmol) with
dimethyl sulfoxide (DMSO; 0.15 mL, 2.11 mmol) and trifluo-
roacetic anhydride (0.27 mL, 1.91 mmol), according to the
reaction conditions previously described,20 yielded a crude 5R-
kinetic diosphenol, the 3-hydroxy-5R-androst-2-ene-4,17-dione
(199 mg, 98%), which was purified by column chromatography.
The same oxidative conditions used for the 5R-vic-diol 5a
were applied to the 5â-vic-diol 5b, giving a mixture with a
variable composition (1:2 to 2:1 by NMR) of a 5â-triketone,
the 5â-androstane-3,4,17-trione, and a 5â-kinetic diosphenol,
the 3-hydroxy-5â-androst-2-ene-4,17-dione.20 The unstable
above-mentioned 5â-kinetic diosphenol and the triketone could
not be separated because of their rapid conversion into
diosphenol 7.
(ii) Isomerization of the Kinetic 5r- and 5â-Diosphe-
nols and 5â-Triketone. The title diosphenol 7 was prepared
in a convergent manner by isomerization of the 5R- and 5â-
kinetic diosphenols and 5â-triketone, previously obtained (60
mg, 0.2 mmol) in methanol solution with sodium metal (50
mg, 2.2 mmol) according to the conditions previously de-
scribed.20 The crude crystalline product obtained was purified
by flash chromatography [silica gel; diethyl ether-carbon
tetrachloride (1:1)] to give 48 mg of 7 (80%).
3r,4â-Dihydroxy-D-homo-17a-oxa-5r-androstan-17-
one (8). The title dihydroxylactone 8 was obtained as a pure
white crystalline solid in a 90-94% yield by oxidation of diol
5a (50 mg, 0.16 mmol) with magnesium monoperoxyphthalate
(7-10 mol/mol of substrate), according to the reaction condi-
tions described previously.22
The preliminary biological profile indicates that it is
important to synthesize new steroid derivatives in order
to refine the encountered SAR for the prepared com-
pounds.
Experimental Section
Chemistry: 3,17-Dioxo-4-oxaandrostane-5r-carbalde-
hyde (2). To a cold (-5 °C) stirred solution of androst-4-ene-
3,17-dione 1 (3.0 g, 10.5 mmol) in 200 mL of methanol, a 2.5%
solution of potassium permanganate in aqueous KOH, 0.05 N
(130 mL, 20.6 mmol), was added dropwise. The temperature
was then raised to 15 °C and after 5-10 min the excess oxidant
was destroyed with saturated aqueous Na2S2O5 (20 mL). After
separation of the formed MnO2 by centrifugation, the methanol
was evaporated and the solution diluted with water (200 mL),
neutralized with aqueous 1 N H2SO4, and extracted with CH2-
Cl2 (3 × 100 mL). The organic extract was washed with 10%
aqueous Na2S2O5 (100 mL) and then water (2 × 100 mL), dried
(MgSO4), and evaporated to dryness to give 1.36 g (41%) of a
crude product from where the title compound 2 was isolated
by crystallization from absolute ethanol.
5r- and 5â-Androst-3-en-17-one (3a and 3b). The title
olefins 3a and 3b were prepared from androst-4-ene-3,17-dione
1 (500 mg, 1.75 mmol) through a Clemmenson-type reduction
with zinc dust in acetic acid, according to the reaction
conditions previously described.20 The crude product obtained
as a white crystalline solid (476 mg) was composed of an
isomeric mixture (2.3:1 by NMR) of 5R-olefin 3a and 5â-olefin
3b. Crystallization of the mixture from n-hexane gave the pure
3a (285 mg, 60%). The 5â-epimer 3b could not be isolated and
has been identified in the mixture with the 5R-epimer only by
1H and 13C NMR.
4-Hydroxy-D-homo-17a-oxaandrost-4-ene-3, 17-dione
(9). Swern-type oxidation of the dihydroxylactone 8 (400 mg,
1.24 mmol) with DMSO (0.3 mL, 4.22 mmol) activated with
trifluoroacetic anhydride (0.54 mL, 3.82 mmol), according to
the reaction conditions previously described22 yielded almost
quantitatively a kinetic diosphenol, the 3-hydroxy-D-homo-
17a-oxa-5R-androst-2-ene-4,17-dione,22 (387 mg, 98%) which,

6384 Journal of Medicinal Chemistry, 2005, Vol. 48, No. 20
Cepa et al.
by further acid-catalyzed isomerization, was quantitatively
(99%) converted in the title thermodynamic diosphenol 9.
3r-Hydroxy-17a-oxa-17a-homo-5r-androstane-4, 17-di-
one (10). To a solution of the dihydroxylactone 8 (100 mg, 0.31
mmol) in analar methylene chloride (2.0 mL), and 10 drops of
methanol were added alumina [AlO3, W200 acid activity super
I (0% H2O), 335 mg] and oxone (250 mg, 0.41 mmol). The
suspension was stirred at room temperature (18 °C) until the
steroid was consumed (6 h, TLC, ethyl acetate/n-hexane 5:1).
After dilution with dried (MgSO4) methylene chloride, the
suspension was filtered and the filtrate was washed with the
same solvent (200 mL) and evaporated to dryness to give a
crude white solid (97-99 mg, 97.6-99.6%), from which the
pure title R-ketol-lactone 10 was isolated by crystallization.
4ê-Cyano-3r,4ê-dihydroxy-17a-oxa-17a-homo-5r-an-
drostan-17-one (11). To 50 mg (0.156 mmol) of the R-ketol-
lactone 10 were added acetone cyanohydrin (99.5 µL, 1.08
mmol) and triethylamine (7.75 µL), and the suspension was
stirred at room temperature for 4 h until the reaction was
complete (TLC). The reaction mixture was then poured into
water (5-10 mL) with ice (5-10 g) and acetic acid (0.25-0.5
mL) under magnetic stirring. The precipitated product was
filtered, washed with water followed by methylene chloride,
and crystallized from ethyl acetate, giving the title cyanohy-
drin 11 as a white powder.
Biochemistry: Preparation of Placental Microsomes.
Placental microsomes were obtained as described by Yoshida
and Osawa26 with some modifications. Human placentas
obtained after delivery from a local hospital were placed in
cold 67 mM potassium phosphate buffer (pH 7.4) containing
1% KCl. The cotyledon tissue was separated and homogenized
in a Polytron homogenizer with 67 mM potassium phosphate
buffer (pH 7.4) containing 0.25 M sucrose and 0.5 mM
dithiothreitol (DTT, 1:1 w/v). The homogenate was centrifuged
at 5000g for 30 min. The supernatant was centrifuged twice
at 20000g for 30 min and at 148000g for 45 min to yield a
microsomal pellet. The microsomes were washed and resus-
pended in 67 mM potassium phosphate buffer (pH 7.4)
containing 0.25 M sucrose, 20% glycerol, and 0.5 mM DTT and
stored at -80 °C. All procedures were carried out at 0-5 °C.
Protein content was estimated by the Bradford method,27 with
bovine serum albumin (BSA) as standard. No significant
enzyme activity loss occurred after 6 months.
Aromatase Assay Procedure. Aromatase activity was
measured according to Thompson and Siiteri23 and Heidrich
et al.28 Aromatase activity was evaluated by measuring the
3H2O released from [1â-3H]androstenedione during aromati-
zation. All tested compounds were dissolved in ethanol and
diluted in 67 mM potassium phosphate buffer (pH 7.4). Briefly,
for the screening assay, 20 µg of protein of the microsomes,
40 nM [1â-3H]androstenedione (1 µCi) purchased from Perkin-
Elmer Life Sciences, and 2 µM each of the inhibitors were used,
while for the IC50 determination, 200 nM (1 µCi) androstene-
dione and different concentrations of the inhibitors were used.
The aromatase-catalyzed reaction was initiated by the addition
of reduced nicotinamide adenine dinucleotide phosphate (NAD-
PH, 150 µM), and incubations were performed in a shaking
water bath at 37 °C for 15 min. However, to minimize the time-
dependent loss of the initial aromatization rate, 10-30 nM
[1â-3H]androstenedione and 5 min incubation time were used
for the kinetic studies. The reaction was terminated by
addition of 500 µL of ice-cold distilled water and 1.5 mL of
ice-cold chloroform. After centrifugation at 3200g for 10 min,
the aqueous phase was added to 5% charcoal/0.5% dextran
coated. After a new centrifugation cycle, the resulting super-
natant was mixed with 1.5 mL of chloroform and centrifuged.
The purified aqueous supernatant containing 3H2O was mixed
with a liquid scintillation cocktail (Universol) and counted in
a Beckman liquid scintillation spectrometer (LS6500). All
experiments were carried out in triplicate.
This work was supported by Fundac¸a˜o para a Cieˆncia
e Tecnologia (FCT), Portugal. M.M.D.S.C. is a recipient
of a Ph.D. grant of FCT (SFRH/BD/10736/2002). We also
thank Novartis Farma-Produtos Farmaceˆuticos S.A.,
Portugal, for kindly supplying formestane for use as a
reference substance.
Supporting Information Available: Elemental analysis
results for compounds 2, 5a, and 9; HRMS data for compounds
4b and 5b; EIMS m/z (M+) data for compounds 4a, 5a, 6, and
8; melting points for compounds 2, 3a, 4a-5b, and 7-9; IR
spectral data for compounds 2, 3a, and 4a-9; and 1H and 13
C
spectroscopic data for compounds 2-11. This material is
available free of charge via the Internet at http://pubs.acs.org.
References
(1) Miller, W. R. Biological Rational for Endocrine Therapy in Breast
Cancer. Best Pract. Res. Clin. Endocrinol. Metab. 2004, 18, 1-32.
(2) Henderson B. E.; Ross R. K.; Pike M. C.; Casagrande, J. T.
Endogenous Hormones as a Major Factor in Human Cancer.
Cancer Res. 1982, 42, 3232-3239.
(3) Sedlacck, S. M.; Horwitz, K. B. The Role of Progestins and
Progesterone Receptors in the Treatment of Breast Cancer.
Steroids 1984, 44, 467-484.
(4) McGuire, W. R. An Update on Estrogen and Progesterone
Receptors in Prognosis for Primary and Advanced Breast Cancer.
In Hormones and Cancer; Iacobelli, S., King R. J. B., Linder H.
R., et al., Eds.; Raven Press: New York, 1980; Vol. 14, pp 337-
344.
(5) Hemsell, D. L.; Grodin, J.; Breuner, P. F. Plasma Precursors of
Estrogen. II. Correlation of the Extent of Conversion of Plasma
Androstenedione to Estrone with Age. J. Clin. Endocrinol.
Metab. 1974, 38, 476-479.
(6) Ackerman, G. E.; Smith M. E.; Mendelson, C. R.; MacDonald,
P. C.; Simpson, E. R. Aromatization of Androstenedione by
Human Adipose Tissue Stromal Cells in Monolayer Culture. J.
Clin. Endocrinol. Metab. 1981, 53, 412-417.
(7) Miller, W. R.; Anderson, T. J.; Jack, W. J. L. Relationship
Between Tumor Aromatase Activity, Tumor Characteristics and
Response to Therapy. J. Steroid Biochem. Mol. Biol. 1990, 37,
537-548.
(8) Njar, V. C. O.; Brodie, A. Comprehensive Pharmacology and
Clinical Efficacy of Aromatase Inhibitors. Drugs 1999, 58, 233-
255.
(9) Zumoff, B. Does Postmenopausal Estrogen Administration
Increase the Risk of Breast Cancer? Contributions of Animal,
Biochemical and Clinical Investigative Studies to a Resolution
of the Controversy. Proc. Soc. Exp. Biol. Med. 1998, 217, 30-
37.
(10) Winer, E. P.; Hudis, C.; Burstein, H. J.; Chebowski, R. T.; Ingle,
J. N.; Edge, S. B.; Mamounas, E. P.; Gralow, J.; Goldstein, L.
J.; Pritchard, K. I.; Cobleigh, M. A.; Langer, A. S.; Perotti, J.;
Powles, T. J.; Whelan, T. J.; Browman, G. P. Technology
Assessment on the Use of Aromatase Inhibitors as Adjuvant
Therapy for Women With Hormone Receptor-Positive Breast
Cancer: Status Report 2002. J. Clin. Oncol. 2002, 20, 3317-
3327.
(11) Brodie, A. Aromatase Inhibitors and Hormone-Dependent Can-
cers. J. Steroid Biochem. Mol. Biol. 1990, 37, 327-333.
(12) Bulun, S. E.; Zeitoun, K.; Takayama, K.; Noble, L.; Michael, D.;
Simpson, E.; Johns, A.; Putman, M.; Sasano, H. Estrogen
Production in Endometriosis and Use of Aromatase Inhibitors
to Treat Endometriosis. Endocr. Relat. Cancer 1999, 6, 293-
301.
(13) Miller, W. R.; Jackson, J. The Therapeutical Potential of
Aromatase Inhibitors. Expert. Opin. Investig. Drugs 2003, 12,
337-52.
(14) Goss, P. E.; Coombes, R. L.; Powles, T. J. Treatment of Advanced
Postmenopausal Breast Cancer with Aromatase Inhibitor, 4-hy-
droxyandrostenedione - Phase 2 Report. Cancer Res. 1986, 46,
4823-4826.
(15) Numazawa, M.; Yamaguchi, S. Synthesis and Structure-Activity
Relationships of 6-Phenylaliphatic-Substitued C19 Steroids
Having a 1,4-diene, 4,6-diene or 1,4,6-triene Structure as Aro-
matase Inhibitors. Steroids 2001, 64, 187-196.
(16) 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, 4277-4283 and references therein.
(17) Numazawa, M.; Yoshimura, A.; Watari, Y.; Matsuzaki, H.
Aromatase Inhibition by 4â,5â-Epoxides of 16R-Hydroxyandros-
tenedione and Its 19-Oxygenated Analogues, Potential Precur-
sors of Estriol Production in the Feto-Placental Unit. Biol.
Pharm. Bull. 2002, 25, 1566-1569.
Acknowledgment. We thank Professor Martinez
Oliveira and Dr. Isabel Campos of Sa˜o Joa˜o Hospital,
Oporto, for generously supplying human term placenta.

A,D-Ring Modified Steroids as Aromatase Inhibitors
Journal of Medicinal Chemistry, 2005, Vol. 48, No. 20 6385
(18) Nagaoka, M.; Watari, Y.; Yajima, H.; Tsukioka, K.; Muroi, Y.;
Yamada, K.; Numazawa, M. Structure-Activity Relationships
of 3-Deoxy Androgens as Aromatase Inhibitors. Synthesis and
Biochemical Studies of 4-Substituted 4-Ene and 5-Ene Steroids.
Steroids 2003, 68, 533-542.
(19) Leclercq, G. Androgens and Other Androstane Derivatives:
Aromatase Inhibitors; Lactones. In Antitumour Steroids: Blick-
enstaff, R. T., Ed.; Academic Press: New York, 1992; pp 65-
122.
(23) Thompson E. A.; Siiteri P. K. Utilization of Oxygen and Reduced
Nicotimamide Adenine Dinucleotide Phosphate by Human
Placental Microsomes During Aromatization of Androstenedione.
Biol. Chem. 1974, 249, 53-64.
(24) Numazawa, M.; Mutsumi, A.; Tachibana, M.; Hoshi, K. Synthesis
of Androst-5-en-7-ones and Androst-3,5-diene-7-ones and Their
Related 7-Deoxy Analogues as Conformational and Catalytic
Probes for the Active Site of Aromatase. J. Med. Chem. 1994,
37, 2198-2205.
(25) Narashimamurthy, J.; Rao, A. R. R.; Sastry, G. N. Aromatase
Inhibitors: A New Paradigm in Breast Cancer Treatment. Curr.
Med. Chem.sAnti-Cancer Agents 2004, 4 (6), 523-534.
(26) Yoshida N.; Osawa Y. Purification of Human Placental Aro-
matase Cytochrome P-450 with Monoclonal Antibody and Its
Characterization. Biochemistry 1991, 30, 3003-3010.
(27) Bradford, M. M. A Rapid and Sensitive Method for the Quan-
tification of Microgram Quantities of Protein Utilizing the
Principle of Protein-Dye Binding. Anal. Biochem. 1976, 72,
248-254.
(20) Tavares da Silva, E. J.; Roleira, F. M.; Sa´ e Melo, M. L.; Campos
Neves, A. S.; Paixa˜o, J. A.; Almeida, M. J.; Costa, M. M. R.;
Andrade, L. C. R. X-ray and Deuterium Labeling Studies on the
Abnormal Ring Cleavages of a 5â-Epoxide Precursor of Form-
estane. Steroids 2002, 67, 311-319.
(21) Tavares da Silva, E. J.; Sa´ e Melo, M. L.; Campos Neves, A. S.
Novel Approach to the Synthesis of the Aromatase Inhibtor
4-Hydroxyandrost-4-ene-3, 17-dione (4-OHA). J. Chem. Soc.,
Perkin Trans. 1 1996, 14, 1649-1650.
(22) Tavares da Silva, E. J.; Sa´ e Melo, M. L.; Campos Neves, A. S.;
Andrade, L. R. C.; Almeida, M. J. M.; Costa, M. M. R. Expedient
Synthesis of Ring-D Lactones of Formestane and Related An-
drostanes as New Potential Aromatase Inhibitors. J. Chem. Soc.,
Perkin Trans. 1 1997, 23, 3487-3489.
(28) Heidrich, D.; Steckelbroeck, S.; Klingmuller, D. Inhibition of
Human Cytochrome P-450 Aromatase Activity by Butylins.
Steroids 2001, 66, 763-769.
JM050129P