C-6α- vs C-7α-Substituted Steroidal Aromatase Inhibitors: Which Is Better? Synthesis, Biochemical Evaluation, Docking Studies, and Structure-Activity Relationships

Article abstract of DOI:10.1021/acs.jmedchem.9b00157

C-6α and C-7α androstanes were studied to disclose which position among them is more convenient to functionalize to reach superior aromatase inhibition. In the first series, the study of C-6 versus C-7 methyl derivatives led to the very active compound 9 with IC₅₀ of 0.06 μM and K_i = 0.025 μM (competitive inhibition). In the second series, the study of C-6 versus C-7 allyl derivatives led to the best aromatase inhibitor 13 of this work with IC₅₀ of 0.055 μM and K_i = 0.0225 μM (irreversible inhibition). Beyond these findings, it was concluded that position C-6α is better to functionalize than C-7α, except when there is a C-4 substituent simultaneously. In addition, the methyl group was the best substituent, followed by the allyl group and next by the hydroxyl group. To rationalize the structure-activity relationship of the best inhibitor 13, molecular modeling studies were carried out.

Full text of DOI:10.1021/acs.jmedchem.9b00157

Subscriber access provided by ECU Libraries
Article
C-6# vs C-7#-Substituted Steroidal Aromatase Inhibitors:
Which is Better? Synthesis, Biochemical Evaluation,
Docking Studies and Structure-Activity Relationships
Fernanda Maria Fernandes Roleira, Carla Varela, Cristina Amaral, Saul C. Costa, Georgina Correia-da-Silva,
Federica Moraca, Giosuè Costa, Stefano Alcaro, Natercia Almeida Teixeira, and Elisiário J. Tavares da Silva
J. Med. Chem., Just Accepted Manuscript • Publication Date (Web): 09 Mar 2019
Downloaded from http://pubs.acs.org on March 9, 2019
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a service to the research community to expedite the dissemination
of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in
full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully
peer reviewed, but should not be considered the official version of record. They are citable by the
Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore,
the “Just Accepted” Web site may not include all articles that will be published in the journal. After
a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web
site and published as an ASAP article. Note that technical editing may introduce minor changes
to the manuscript text and/or graphics which could affect content, and all legal disclaimers and
ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or
consequences arising from the use of information contained in these “Just Accepted” manuscripts.
is published by the American Chemical Society. 1155 Sixteenth Street N.W.,
Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the course
of their duties.

Page 1 of 86
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
C-6α vs C-7α-Substituted Steroidal Aromatase
Inhibitors: Which is Better? Synthesis, Biochemical
Evaluation, Docking Studies and Structure-Activity
Relationships
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Fernanda M. F. Roleiraa,b*, Carla Varela , Cristina Amaral , Saul C. Costa , Georgina Correia-da-
a,b
c
a
c
d,e
d
d
c
Silva , Federica Moraca , Giosuè Costa , Stefano Alcaro , Natércia A. A. Teixeira , and Elisiário J.
Tavares da Silvaa,b*
a
Laboratory of Pharmaceutical Chemistry, Faculty of Pharmacy, University of Coimbra, 3000-548
Coimbra, Portugal
b
CIEPQPF Centre for Chemical Processes Engineering and Forest Products, University of Coimbra,
3030-790 Coimbra, Portugal
c
UCIBIO.REQUIMTE, Laboratory of Biochemistry, Department of Biological Sciences, Faculty of
Pharmacy, University of Porto, 4050-313 Porto, Portugal
d
Laboratorio di Chimica Farmaceutica, Dipartimento di Scienze della Salute, Università Magna Græcia
di Catanzaro, 88100 Catanzaro, Italy
e
Department of Chemical Sciences, University "Federico II" of Napoli, via Cinthia, 4, 80126, Napoli,
Italy
*
Corresponding authors. Tel.: +351 239 488 400; fax: +351 239 488 503; e-mail addresses:
froleira@ff.uc.pt (Fernanda M. F. Roleira) and etavares@ff.uc.pt (Elisiário J. Tavares da Silva)
ACS Paragon Plus Environment
1

Journal of Medicinal Chemistry
Page 2 of 86
Abstract
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
C-6α and C-7α androstanes were studied to disclose which position among them is more convenient to
functionalize in order to reach superior aromatase inhibition. In the first series, the study of C-6 vs C-7
methyl derivatives led to the very active compound 9 with IC50 of 0.06 μM and Ki = 0.025 μM
(competitive inhibition). In the second series, the study of C-6 vs C-7 allyl derivatives led to the best
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
aromatase inhibitor 13 of this work with IC of 0.055 μM and Ki = 0.0225 μM (irreversible inhibition).
50
Beyond these findings, it was concluded that position C-6α is better to functionalize than C-7α, except
when there is a C-4 substituent simultaneously. In addition, the methyl group was the best substituent,
followed by the allyl group and next by the hydroxyl group. To rationalize the structure-activity
relationship (SAR) of the best inhibitor 13, molecular modelling studies were carried out.
ACS Paragon Plus Environment
2

Page 3 of 86
Journal of Medicinal Chemistry
Introduction
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Among the various types of cancer, breast cancer is one of the most expression today, being considered
one of the cancers that presents a higher incidence rate and constitutes one of the main causes of death
1
of Women around the World. Most breast cancers are hormone-dependent and 80% of these cancers
+
have estrogen receptors (ER ), which allows them to use these hormones for their growth. Estrogens are
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
produced in the body through a cascade of steroid compound reactions that starts from cholesterol. The
latter reaction of this cascade is catalysed by an enzyme called aromatase, which converts the enone
function present on the A ring of testosterone and androstenedione into the phenolic aromatic ring
present in estradiol and estrone, respectively. For this reason, the enzyme aromatase, which belongs to
the cytochrome P450 superfamily, is a very valuable therapeutic target for the selective treatment of
2
estrogen-dependent breast cancer. There are currently only three aromatase inhibitors in clinical use for
the treatment of this type of breast cancer. Letrozole and anastrozole are two of these inhibitors and
belong to the group of azole derivatives, while exemestane is the only inhibitor of the steroid group
marketed (Figure 1). These three third generation aromatase inhibitors are part of the first-line arsenal
+
3
for the treatment of breast cancer (ER ) after menopause, being very potent as anti-tumor and very
selective for aromatase. However, all of these inhibitors have side effects resulting from their use for
very long periods (five years). The most serious side effects are the progressive loss of bone density
(osteoporosis) associated with increased risk of bone fractures and bone pain, as well as cardiovascular
problems resulting from reduced levels of estrogens.4, 5 Other type of side effects causing by the
prolonged administration of these inhibitors relate to the phenomenon of resistance acquired by the
6
tumor cells leading to tumors regrow and, consequently, to the progression of cancer. In order to
overcome these drawbacks it has become necessary to develop novel aromatase inhibitors. Recently,
very important works related to nonsteroidal AIs have been reported.7-9 However, the investigation on
steroidal AIs has been less explored, though steroidal AIs are more selective, highly enzyme specific
and less toxic10 due to its structural similarity with the aromatase substrate, androstenedione. In
addition, the use of the combination of exemestane with other drugs has been proposed for the treatment
ACS Paragon Plus Environment
3

Journal of Medicinal Chemistry
Page 4 of 86
of women with metastatic breast cancer, when acquired resistance to nonsteroidal AIs appears.11, 12
Further, it has been suggested that, among the AIs in clinical use, exemestane is the only one that has
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
1
3
shown to be less aggressive in osteoblasts in vitro mainly due to its androgenic activity, leading to
reduced side effects associated with loss of bone density.5, 14, 15 Our previous work has been focused on
the study of several new steroidal compounds modified with convenient groups in their A-, B- and D-
rings. Some of our new molecules have potent anti-aromatase activity,16-20 causing anti-proliferative
effects and resensitizing resistant breast cancer cells to aromatase inhibition.21-23 In the present project
we are particularly interested in studying steroids substituted in positions 6 and 7, because some studies
propose that alkyl or phenyl functions attached at the C-6 or C-7 positions of the androstane backbone
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
may be effective in inhibiting the enzyme aromatase,24-28 once functional groups at these positions are
capable of promoting interactions with the active site of the enzyme. In addition, Ghosh et al.29, 30
described the existence of an access channel cavity located near the C-4 and C-6 positions of the
androstenedione complexed with aromatase. They also developed very potent C-6 substituted steroids
with alkyl and alkynyl side chains capable of being anchored in the described access channel cavity of
2
8
18
the enzyme, resulting in very efficient inhibition. Further, recent studies from our group revealed that
C-7α allyl derivatives of androstenedione strongly interact with aromatase leading to its inhibition.
However, despite all these studies, the best position to functionalize among positions 6 and 7 of
androstane framework still remains unknown. For these reasons, it is our intention to do analog design
by preparing new series of steroidal androstanes (Schemes 1-8), based on the most potent and promising
hit compounds previously developed by our team,16-18 namely those with double bonds or epoxide
functions at positions C-1, C-3 or C-4, having additional C-6 or C-7 methyl, allyl or hydroxyl groups.
After preparation, the new compounds will be evaluated for their inhibitory activity against aromatase,
aiming to construct robust structure-activity relationships (SAR) with respect to the referred positions,
and ultimately discover hit compounds which can establish a base to develop new lead and drug
+
candidates steroidal AIs for the therapy of ER breast cancer.
ACS Paragon Plus Environment
4

Page 5 of 86
Journal of Medicinal Chemistry
Results
Chemistry
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
The first series synthesized was the methyl series (Scheme 1). Accordingly, the synthesis of 6α-methyl
derivatives 1a, 3a and 9 was achieved following the strategy reported at scheme 2. Direct alkylenation
of androstenedione (22) was performed by sodium acetate, phosphoryl chloride and formaldehyde
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
3
1
dimethyl acetal in chloroform under reflux temperature. Column chromatography purification allowed
isolating 23 in 17% yield. Compound 9 was prepared in 71% yield through reaction of 23 in refluxing
3
2
solution of ethanol/cyclohexene and catalysis by 5% Pd-C. The follow up of this reaction was easily
made by UV spectroscopy once the starting material absorbs at λmax of 260 nm while the desired product
absorbs at λmax of 240 nm. This reaction allowed obtaining stereoselectively the 6-methyl derivative 9,
with the methyl group in the more stable equatorial conformation, which structure was assigned by
comparing the NMR and melting point data with that referred in the literature.32 Alternatively,
compound 9 could also be prepared from compound 30 in 47% yield, as discussed later (scheme 4).
Compound 9 was then reduced under reflux by zinc dust in glacial acetic acid through a Clemmenson
reaction,33 leading to an isomeric mixture of 5α- and 5β-olefins 1a and 1b (1.7:1), respectively.
Separation of this mixture by column chromatography allowed isolating the pure isomer 1b in 13%
yield. Isomer 1a was obtained in mixture with 1b in 90% purity. The stereochemistry of the 5 (1a) and
1
5
 (1b) stereoisomers was assigned with basis in the chemical shifts observed for the H NMR signals
for C-19 protons (δ = 0.80 ppm for 1a and δ = 0.97 ppm for 1b.34-36 In the next reaction, the mixture of
3
5
1
a and 1b was treated with performic acid generated in situ affording the corresponding mixture of
epoxides 3a and 3b which after purification by column chromatography afforded 3a in 44% yield and
b in 29% yield. The stereoselectivity of the addition of the peracid on the -face of 5-olefin (1a) and
3
on the -face of 5-olefin (1b) was already expected. In fact, the -face of the olefin 1a is sterically
hindered due to the proximity of the 10-methyl group. On the other hand, in the 5-olefin 1b, the
constraint caused by the flexing of ring A on ring B of this isomer renders its -face more accessible to
ACS Paragon Plus Environment
5

Journal of Medicinal Chemistry
Page 6 of 86
the peracid. The stereochemistry of 3a and 3b isomers was established based on the chemical shifts
shown by the C-19-methyl and C-3 and C-4 proton signals, as well as based on the coupling constants
between the C-3 protons and C-4 with the protons C-2 and C-5, respectively.34-36
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Scheme 3 presents the strategy followed to prepare the studied 7α-methyl compounds. In the first
reaction, androstenedione (22) was dehydrogenated with chloranil, in tert-butanol at reflux
temperature.37 This allowed obtaining 24 in 48%, after purification by column chromatography.
Compound 24 was then submitted to a 1,6-conjugate addition using trimethylaluminium in the presence
of a catalytic amount of CuBr and trimethylsilyl chloride, in tetrahydrofuran (THF), at room
temperature.37, 38 This reaction produced both C-7 methyl isomers which were separated by column
chromatography giving the pure 10a (30% yield) and 10b (8% yield). The stereoselectively observed for
this step is in accordance with that described for this kind of reaction, affording a major amount of the
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
7
-methyl isomer 10a. The stereochemistry of the 7methyl (10a) and 7methyl (10b) groups was
1
assigned with basis in the chemical shifts and coupling constants observed for the H NMR signals for
37
4
C-7 methyl protons. After this step, two different paths were followed, one to prepare the Δ -olefin 6
3
(
path 1) and another to prepare the Δ -olefins 2a and 2b (path 2) (Scheme 3). In path 1, compound 10a
was treated with sodium borohydride in trifluoroacetic acid, glacial acetic acid and acetonitrile under
39
nitrogen atmosphere, affording the pure 6 in 10% yield accompanied by its C-17 hydroxyl derivative
2
5 in 62 % yield, after column chromatography purification. Subsequent reoxidation of compound 25
with Jones reagent allowed its conversion in 6 in 98% yield. Compound 8 was prepared by reaction of 6
40
with m-chloroperbenzoic acid (m-CPBA) in dichloromethane at room temperature, being isolated in
9
0% yield after column chromatography. The stereochemistry of the -epoxide 8 was attributed
1
7
based on the chemical shift of the C-19 protons. In path 2, compound 10a was reduced with zinc dust
3
in glacial acetic acid under reflux affording the Δ -olefin isomers 5α 2a and 5β 2b, in mixture with the
4
Δ -olefin 6. Compounds 2a and 2b could not be isolated. However, after sequential column
chromatography, some fractions could be enriched in compound 2b (76%), combined with compound 6
(
24%). Oxidation of this mixture with performic acid35 in dichloromethane at room temperature
ACS Paragon Plus Environment
6

Page 7 of 86
Journal of Medicinal Chemistry
afforded, after column chromatography, the 3,4-epoxyde 4 in 56% yield. The stereoselectivity observed
in the formation of the olefins 2a and 2b was similar to that observed for the formation of 1a and 1b. In
the formation of epoxide 4, the same type of stereoselectivity previously discussed for the formation of
epoxy derivatives 3a and 3b was observed, with the addition of the peracid on the stereochemically free
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6

-side of 5-olefin 2b.The stereochemistry of the 5 (2a) and 5 (2b) stereoisomers was assigned with
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
basis in the chemical shifts observed for the H NMR signals for C-19 protons (δ = 0.81 ppm for 2a and
δ = 1.0 ppm for 2b. Likewise, the stereochemistry of the epoxide derivative 4 was attributed based on
the chemical shifts of the C-19, C-3, C-4 and C-7 methyl group protons, as well as their coupling
constants.34-36 The synthesis of compound 5 and its epoxide derivatives 7a and 7b was carried out
following the strategy presented in scheme 4. Dehydroepiandrosterone (26) was submitted to a
4
1
5
reduction of the C-17 carbonyl group with sodium borohydride in methanol at 0 ºC leading to the Δ -
3
β,17β-diol 27 in 76% yield. This diol was later set to react under epoxidation conditions using m-
CPBA in dichloromethane at room temperature40 affording epoxide 28 (84%). Compound 28 was
submitted to nucleophilic attack of methylmagnesium bromide in THF under reflux41, 42 giving the
tertiary alcohol 29 with the methyl group in the C-6 position, in 50% yield. Compound 30 was produced
4
3
by submitting 29 to Jones oxidation in acetone at 0 ºC. Dehydration of 30 by acid catalysis with HCl
44
in acetic acid at room temperature allowed simultaneous formation of the double bond at C-4 and
isomerization of the 6β- to the 6α-methyl group, affording compound 9 in 47% yield. The synthesis of
the 3-deoxo derivative 5 was performed by reaction of 9 with a mixture of sodium borohydride in
39
trifluoroacetic acid, glacial acetic acid and acetonitrile under nitrogen atmosphere. In this reaction, 5
was obtained in 14% yield and its C-17 hydroxyl derivative 31, in 42% yield. Subsequent reoxidation of
3
1 with Jones reagent allowed its reconversion in 5 in 79% yield. Oxidation of 5 at room temperature
3
5
with formic peracid gave a mixture of epoxides 7a and 7b in 39% and 61% yields, respectively, as
determined by NMR. The referred epoxides were separated by column chromatography giving 4α,5α-
epoxide 7a in 31% and 4β,5β-epoxide 7b in 38% yield. The stereoselectivity observed in the preparation
reaction of epoxide derivatives 7a and 7b from olefin 5 was expected, based on the results published
ACS Paragon Plus Environment
7

Journal of Medicinal Chemistry
Page 8 of 86
previously by the authors for the same type of compounds. The stereochemistry of 7a and 7b was
attributed based on the chemical shifts of the C-4 and C-19 protons.17
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
The second series synthesized was the allyl series (Scheme 5). Thus, the synthesis of the 6α-allyl
substituted derivatives 13, 15 and 17 was carried out through the strategy presented in scheme 6. The
first step of the sequence was the protection of the hydroxyl groups of the epoxide 28. Actually, when
the Grignard reaction was performed directly in 28 the yield was very small with many by-products.
However, this problem was overcome by using tert-butyldimethylsilyl protecting groups for the C-3 and
C-17 alcohols. For this, 28 was submitted to reaction with tert-butyldimethylsilyl chloride, imidazole
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
4
5
and dimethylformamide, at room temperature affording 80% yield of 32. Grignard reaction of 32 with
2
7
allylmagnesium bromide in THF at reflux gave the pure compound 33 in 82% yield, after column
chromatography. Acid hydrolysis of 33 in ethanol at room temperature afforded the trihydroxylated
derivative 34 as an oil, which was directly used to the next reaction. In this reaction, 34 was submitted
4
3
to Jones oxidation in a mixture of acetone and dioxane leading to the pure 35 in 61% yield, after
column chromatography. Acid catalysed dehydration of 35 with HCl in acetic acid at room
4
4
temperature gave compound 15 in 74% yield. As observed before for other compounds, this reaction
occurred with simultaneous formation of the C-4 double bond and isomerization of the 6β- to 6α-allyl
1
group. The -configuration of the C-6 allyl substituent was assigned on the basis of the H NMR
spectroscopy. The signal of the olefinic proton at C-4 of the 6-substituted steroid 15 appears as a
2
7
doublet (J4-6 = 1.1 Hz) at 5,08 ppm. In order to introduce an additional double bond at C-1/C-2
positions of A-ring, 15 was treated with 2,3-dichloro-5,6-dicyano-1,4-benzoquinone (DDQ) and benzoic
4
6
acid in toluene, at reflux, allowing the formation of the pure 13 in 81% yield, after alumina filtration
followed by column chromatography. To synthesize the 3-deoxo derivative 17, compound 15 was set to
react with a mixture of sodium borohydride in trifluoroacetic acid, glacial acetic acid and acetonitrile
3
9
under nitrogen atmosphere. These conditions allowed obtaining the pure 17 in 20% yield, and the pure
ACS Paragon Plus Environment
8

Page 9 of 86
Journal of Medicinal Chemistry
C-17 hydroxyl derivative 36 in 65% yield, after column chromatography. Reoxidation of 36 with Jones
reagent allowed its reconversion in 17, in 40% yield.
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
To prepare compound 12 it was followed a six-step strategy depicted in scheme 7. The first reaction
consisted in dibromination of androstenedione (22) with bromine in acetic acid in diethyl ether at 0º C,47
leading to an isomeric mixture of 2α,6α- and 2α,6β-dibromoenone 37a and 37b in 70% yield.
Compound 38 was obtained through an elimination reaction of both 37a and 37b, followed by
rearrangement, under reflux in an ethanolic solution of potassium acetate48 in 92 % yield, after
crystallization from acetone/n-hexane. Acid catalysed hydrolysis of the enol-acetate 38 with a 5% HCl
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
4
9
ethanolic solution, led to the diosphenol 39 in 85% yield. After this, protection of the C-4 hydroxyl
group of 39 was carried out with acetic anhydride in pyridine at room temperature affording 40 in 69%
yield, after column chromatography. This compound was then submitted to a Sakurai reaction with
5
0
allyltrimethylsilane in anhydrous dichloromethane using TiCl as catalyst, giving 41 in 30% yield,
4
after preparative TLC followed by ethyl acetate crystallization. Finally, the acid catalysed hydrolysis of
4
1 by refluxing in ethanolic HCl solution, afforded the pure 12 in 62% yield.
To prepare compound 11 an eight-step strategy was developed starting from the available testosterone
42) (Scheme 8). Reduction of 42 by sodium borohydride in methanol at room temperature yielded the
(
1
7
pure 43 in 21% yield, after crystallization. The epoxidation of 43 was performed by reaction with
performic acid generated in situ and afforded 44 in 71% yield, after purification by column
chromatography. Acetylation of 44 with acetic anhydride in pyridine afforded compound 45 in 64%
yield. Reaction of epoxide 45 with 75% aqueous CrO in butanone at room temperature afforded the α-
3
5
1
5
2
k
e
t
o
l
4
6
i
n
6
2
%
y
i
e
l
d
.
D
e
h
y
d
r
a
t
i
o
n
o
f
4
6
w
i
t
h
t
h
i
o
n
y
l
c
h
l
o
r
i
d
e
i
n
a
n
h
y
d
r
o
u
s
p
y
r
i
d
i
n
e
a
t
0
º
C
afforded 47 in 73% yield, after column chromatography followed by diethyl ether/methanol
crystallization. The subsequent introduction of the allyl group at C-6 position was achieved through a
Sakurai reaction. For that, 47 was set to react in dichloromethane with allyltrimethylsilane and TiCl as
4
catalyst at -78 ºC50, 53 through a 1,4-type addition, affording 48 as a clear oil, in 55% yield. Treatment of
4
8 with a 5% aqueous NaOH in methanol and acetone, at room temperature gives 49 in 76% yield, after
ACS Paragon Plus Environment
9

Journal of Medicinal Chemistry
Page 10 of 86
diethyl ether/n-hexane crystallization. The -configuration of the C-6 allyl substituent of compound 49
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
1
was also assigned on the basis of the H NMR spectroscopy. The signal of the olefinic proton at C-6 of
steroid 49 appears as a double doublet (J6β-7α = 12.9 Hz, J6β-7β = 7.8 Hz) at 3.23 ppm. The magnitude of
1
2.9 Hz for the coupling constant with one of the two C-7 protons and 7.8 Hz with the other reveals
interactions of the type 6-axial/7-axial and 6-axial/7-equatorial. For this reason the C-6 allyl group
will have a 6equatorial configuration. During the alkaline hydrolysis of 48, the corresponding C-3/C-
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
7 diol formed rearranges to the more stable diosphenol 49. The last step of this strategy consisted in the
selective Swern type oxidation of the C-17 hydroxyl group of 49 with DMSO activated with
trifluoroacetic anhydride (TFAA) and triethylamine in anhydrous dichloromethane at -60 ºC.54
Purification of the obtained crude by preparative TLC with diethyl ether/toluene gave the pure 11, in
6
4% yield.
Biochemistry
The anti-aromatase activity of the synthesized steroidal compounds was determined using human
5
5
placental microsomes as source of the enzyme and the radiometric method of Thompson and Siiteri in
3
3
which the H O released during aromatization of [1β- H] androstenedione allows measuring the rate of
2
estrogen formed. Firstly, a screening test was run by determining the percentage of aromatase inhibition
(%) for all compounds (Table 1), in relation to the absence of the inhibitor. Formestane (99.22 ± 0.1%)
and exemestane (98.91 ± 0.6%) were used as reference AIs. Except steroids 4, 7a, 7b, 8 and 21, the
majority of compounds presented an aromatase inhibition higher than 80%, being steroid 9 (98.73 ±
0
.1%) the most potent AI. The IC value in human placental microsomes was determined for steroids
50
3
with aromatase inhibition higher than 80%, using [1β- H] androstenedione at 100 nM concentration
(
Table 1, Figure 2). Steroids 3a, 5, 9, 12, 13 and 15 presented IC values lower than 0.2 µM. The AIs 9
50
and 13 demonstrate to be the most potent AIs, since they presented an IC value of 0.060 µM and 0.055
5
0
1
7
µM, respectively, being these IC values similar to formestane (0.042 µM) and exemestane (0.050
5
0
ACS Paragon Plus Environment
1
0

Page 11 of 86
Journal of Medicinal Chemistry
1
9
+
µM). The anti-aromatase activity in MCF-7aro cells, an aromatase overexpressing ER breast cancer
cell line that is considered the best in vitro cell model to study AIs,56, 57 was also studied for the most
potent AIs. As presented in Table 1, except for compounds 5 and 12 that induced the formation of
crystals in cell culture, the results demonstrate that the majority of these steroids present an anti-
aromatase activity higher than 80%. Steroids 9 and 13 were the most potent AIs, since they show an
aromatase inhibition of around 98% in cells. For the latter AIs, were additionally carried out kinetic
studies aiming to define its binding type with the active site of the enzyme. Moreover, apparent
inhibition constants (Ki) were measured, in diverse ranges of concentrations of inhibitor (0.02 – 0.4
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
3
µM) and of [1β- H] androstenedione (10-30 nM) (Table 2). AI 9 revealed to be a competitive inhibitor
(Ki = 0.025 µM) while steroid 13 demonstrates to be an irreversible inhibitor (Ki = 0.0225 µM), a
behavior similarly to exemestane.5, 58 As exemestane, compound 13 presents the 1,2-double bond which,
according Lombardi P,59, 60 acts as a latent alkylating group by stabilizing the covalent bond formed
between the inhibitor and the enzyme,59, 60 turning it an irreversible inhibitor. On the other hand,
compound 9 not having the 1,2-double bond will not form an irreversible binding with the enzyme,
being a competitive inhibitor. Both AIs presented a Ki similar to exemestane (Ki = 0.026 µM),5, 60 being
considered potent AIs. Figure 3 shows representative Lineweaver−Burk and Dixon curves for these AIs.
Docking using GRID Molecular Interaction Fields (MIFs)
Compounds 4, 7a, 7b, 8, 11, 12, 13, 14 and 21 were docked using the FLAPdock tool61 into the
6
2
aromatase binding site defined by the FLAPsite utility (Supporting Information, Figures S2). Docking
results were ranked according to the FLAP-S Score function, an algorithm that evaluates the overlap
similarity between the H (shape), O (H-bond acceptor), N1 (H-bond donor) and DRY (Hydrophobic)
MIFs with the compound predicted poses. Looking at the best FLAP S-score poses, it can be observed
generally that compounds 11, 12, 13 and 14 bind at the active site similarly to the endogenous
aromatase substrate, androstenedione (Supporting Information, Figure S4), with the steroidal moiety
well overlapping the hydrophobic environment (DRY probe) of the binding pocket, the β-face oriented
ACS Paragon Plus Environment
1
1

Journal of Medicinal Chemistry
Page 12 of 86
toward the heme group and their A- and D-rings positioned toward the Asp309 and the Met374
residues, respectively. This binding orientation was experimentally demonstrated also for the well-
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
2
8
known aromatase inhibitor exemestane (Supporting Information, Figure S5) and it is coherent with the
63
proposed catalytic mechanism of aromatase. Nevertheless, the GRID-guided docking method revealed
significant differences in the binding geometries adopted by each ligand. Indeed, looking at the energy
minimized FLAP S-score poses, it can be clearly observed that, while the 17-keto oxygen atom of the
D-ring engages one hydrogen bond with the backbone amide of Met374 as reported for others steroidal
inhibitors, the 3-keto oxygen atom of the A-ring adopts, among all ligands, different binding geometries.
Indeed, the 3-keto oxygen atom of 13 is the closest to the carboxylate group of Asp309 (3.7 Å) (Figure
4). This distance is almost similar to that reported experimentally in literature for the inhibitor
exemestane (3.0 Å) (Supporting Information, Figure S5),28 the parent compound of 13. A further
analysis of other binding geometric properties (dihedrals and angles) (Supporting Information, Table
S5) support the very similar binding mode adopted by 13 with respect to the bioactive conformation of
exemestane (Supporting Information, Table S5 and Figure S5). It is worth to mention that all the
geometric measurements were taken from the energy minimized FLAPdock poses, in order to take into
account both receptor and ligands adjustments. The peculiar geometric features of 13 lead this inhibitor
to be a similar competitor of exemestane with the aromatase endogenous substrate androstenedione. We
cannot exclude that an important contribution is given by the better planarity of the A-ring of 13,
conferred by the 1,2- and 4,5-double bonds, which is an important structural requirement for the anti-
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
2
0
aromatase activity. Conversely, the geometrical analysis of the energy minimized docking poses of 11
and 12 (Figures 5 and 6, respectively) reveals that the simultaneous presence of an hydroxyl group at C-
4
causes a deeper distortion of the dihedrals C6-C5-C4-C3 and C4-C3-O1-OD2, especially when the
allyl side chain is positioned at C-6α (11) (Supporting Information, Table S5), probably due to the steric
hindrance reasons. The main consequence of this distortion concerns the 3-keto oxygen, which is farther
from the carboxylate group of Asp309 (5.0 Å) in the case of 11 and 4.1 Å in the case of 12 (Supporting
Information, Table S5) (Figures 5 and 6, respectively). Finally, compound 14 although owns the same
ACS Paragon Plus Environment
1
2

Page 13 of 86
Journal of Medicinal Chemistry
A-ring planarity as 13 and exemestane (Supporting Information, Table S4), its 3-keto oxygen is 6.0 Å
far from the OD2 oxygen atom of Asp309 (Figure 7). This is mainly due to steric hindrance of the C-7α
allyl side chain with Leu477 and Ser478 residues that it is not able to accommodate within the access
channel, similarly to 12 (Figure 6). As concerns the docking poses of the less active compounds 4, 7a,
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
7
b and 8 (Supporting Information, Figure S3), it seems that the loss of activity of these epoxide
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
derivatives could be due to the absence of the 3-keto oxygen atom in the A ring and, in the case of 4,
also to the 5β-stereochemistry, which prevents the establishment of a hydrogen bond with Asp309. The
epoxide oxygen atom is led, in fact, very far from the Asp309 (Supporting Information, Figures S3A,
S3B, S3C and S3D). In addition, we cannot discard the pronounced deformation of the 4,5-epoxides A-
2
0
ring, as previously reported by the authors. On the contrary, compound 21 owns the 3-keto oxygen
atom in the A ring, but the simultaneous presence of an hydroxyl group at C-7α alters its binding
geometry, so that the 3-keto oxygen atom is located almost perpendicularly at a distance of 4.0 Å to the
carboxylate group of Asp309, thus disfavouring any hydrogen bond interaction (Supporting
Information, Figure S3E). A further explanation of the lower activity of 21 lies in the binding geometry
of the hydroxyl group, which falls in a hydrophobic region of the binding pocket (Supporting
Information, Figure S6), thus disfavouring its interactions.
Discussion and Conclusions
As referred, the best position to functionalize, among positions C-6 and C-7 of androstane
framework, to reach efficient aromatase inhibition, still remains unknown. That is why we invested in
the design and synthesis of series of new steroidal androstanes, with different substituents in position C-
6
and C-7. An analog design was performed based in the most potent hit compounds previously
developed by our group as models. Structural properties of the aromatase active site, near the C-4, C-6,
28
and C-7 positions of androstenedione were also taken into consideration.
One of the series designed and synthesized was the C-6/C-7 methyl series (Scheme 1). In this
series there are a number of compounds with a methyl group in position C-6 and also a number of
ACS Paragon Plus Environment
1
3

Journal of Medicinal Chemistry
Page 14 of 86
analogous compounds with a methyl group in position C-7. Looking at the aromatase inhibitory activity
of these compounds (Table 1), it is possible to conclude that, generally, C-6 methyl derivatives are
better AIs than the corresponding C-7 methyl derivatives, either in microsomes or in breast cancer cells.
By comparing the pairs 5/6 (IC = 0.17/0.40 µM) and 9/10a (IC = 0.06/0.27 µM) this conclusion turns
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
5
0
50
clear. Further, for the pair 9/10a, it is possible to infer the same conclusion from the results obtained in
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
MCF-7aro cells (inhibition %= 98.56/90.77), respectively. In the case of the pair 1a/2b (IC = 0.56/0.51
50
µM), it is not possible to do an accurate comparison firstly because compound 2a (Scheme 3), with 5β
stereochemistry, could not be isolated and secondly because compound 2b was always obtained in
mixture with compound 6, which presents itself an IC of 0.40 µM. By this reason the activity of 6
50
must contribute to increase the activity of 2b, which being a C-5β androstane would reveal low
aromatase activity.16, 17, 64-66 In the case of the pair 3a/4 (inhibition %=96.72/13.66, in microsomes),
again it was not possible to do a precise comparison, since the C-5α derivative of 4 could not be
isolated. Instead we isolated the C-5β epoxide 4 whose percentage of aromatase inhibition (13.66%),
corroborates the low capacity of inhibition of C-5β androstanes. Considering the pair 7a/8 (inhibition
%=58.69/64.89, in microsomes) it seems that, in this case, C-6 and C-7 methyl derivatives have similar
activities, presenting the C-7 methyl derivative a slightly higher activity. When comparing with the
other compounds of this series, the pair 7a/8 presents an additional structural feature, namely a
substitution at C-4/C-5 positions, which may interfere with the interaction of the C-6 substituent with
the aromatase active site, as discussed later. In addition, by focusing in the structural features of the A-
ring, it is possible to infer that the presence of a 3,4-epoxide is better than a 3,4-olefin for aromatase
inhibition, as observed for the pair 1a/3a (IC =0.56/0.175 µM), but the presence of a 4,5-olefin is better
5
0
than a 4,5-epoxide, as in pairs 5/7a (inhibition %=96.72/58.69, in microsomes) and 6/8 (inhibition
%=97.39/64.89, in microsomes). This kind of SAR was already observed for similar compounds
without substitution in the B-ring and was thoroughly discussed in previous works of the authors.17, 20
Considering compounds 7a and 7b, it seems that the 4β,5β configuration of the epoxide 7b favours
more the interaction with the aromatase active site as shown also by molecular modelling studies
ACS Paragon Plus Environment
1
4

Page 15 of 86
Journal of Medicinal Chemistry
(
Supporting Information, Figure S3C). However, the absence of a 3-keto oxygen atom, deeply decrease
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
its activity. Furthermore, by comparing the pairs 5/9 (IC =0.17/0.06 µM) and 6/10a (IC =0.405/0.27
5
0
50
µM) it is possible to conclude that the presence of the C-3 carbonyl function is very advantageous for
aromatase inhibition. In fact, the presence of a C-3 carbonyl group makes compounds more similar to
the natural substrate, androstenedione, allowing the establishment of additional hydrogen bonds with
aromatase active site residues, hence favouring enzyme inhibition. Compound 9, previously prepared as
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
2
7
AI by Numazawa et al. through a different synthetic strategy, is effectively the most potent aromatase
inhibitor of this series, presenting an IC of 0.06 μM, displaying high affinity to the aromatase (Ki
5
0
=
0.025 μM) and with a competitive type inhibition, showing an IC50 and a Ki similar to that of
1
9
60
exemestane (IC =0.05 μM and Ki=0.026 μM ) (Table 2). Moreover, this compound also has the
5
0
+
ability to efficiently inhibit aromatase (inhibition %= 98.56) in an aromatase overexpressing ER breast
cancer cell line, being one of the most potent studied AIs, in these cells.
Other series designed and synthesized was the C-6/C-7 allyl series (Scheme 5). In this series
there is a number of compounds with an allyl group in position C-6 and the corresponding analogous
compounds with an allyl group in position C-7. As observed for the methyl series, in the allyl series the
C-6 allyl derivatives are better AIs than the corresponding C-7 allyl derivatives, either in microsomes or
in breast cancer cells, except for the pair 11/12 (IC = 0.52/0.11 µM) (Table 1) where the C-7 allyl
5
0
derivative is a stronger AI than the analogous C-6. Once again, this pair presents an additional structural
feature, namely a C-4 substitution as discussed before for the pair 7a/8. The observation that the pairs
7a/8 and 11/12 are the only ones that show a stronger aromatase inhibition for the C-7 derivative and
that are also the only ones bearing a C-4 substitution, led us to put the hypothesis that the referred
substitution may play an significant role in the molecular interaction within the aromatase active site.
The case of the pair 11/12 was deeply studied by docking simulations showing that, in both compounds
1
1/12, the simultaneous presence of a hydroxyl group at C-4 and the allyl side chain at C-6α and C-7α,
respectively, interferes with the steric hindrance to the A-ring, which is located farther from Asp309
thus preventing the establishment of strong hydrogen bonding between the 3-keto oxygen atom and
ACS Paragon Plus Environment
1
5

Journal of Medicinal Chemistry
Page 16 of 86
Asp309. In any case, the C-7α substitution of the allyl side chain is more favoured, because the A-ring
finds closer to Asp309 with respect to 11 (Figures 5 and 6).
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Regarding the structural features of the steroidal A-ring, one can say that besides the C-4 double
bond, the presence of an additional carbonyl group at C-3 is advantageous for aromatase inhibition, as in
pair 15/16 (IC = 0.105/0.59 µM) relatively to pair 17/18 (IC = 0.21/0.745 µM). As said, a C-3
5
0
50
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
carbonyl group makes compounds more similar to the natural substrate, androstenedione, permitting the
formation of additional hydrogen bonds with aromatase active site residues and favouring enzyme
inhibition. Further, the inclusion of an additional double bond in C-1 (pairs 13/15 and 14/16) has
demonstrated to be very advantageous allowing to found a very potent aromatase inhibitor (compound
13) presenting an IC of 0.055 μM, showing high affinity to the aromatase (Ki = 0.0225 μM) and with
5
0
an irreversible type inhibition. Actually, compound 13 is the most potent compound of this series and
also of all the other series, in microsomes and in breast cancer cells, showing a behavior very similar to
1
9
exemestane, which is also an irreversible inhibitor of aromatase with an IC around 0.050 μM and a
5
0
60
Ki of 0.026 μM (Table 2). Docking simulations confirmed the similar behavior of compound 13 and
exemestane as described in the results.
In the C-7α substituted compounds 12 and 14 it was detected a steric hindrance of the C-7α allyl
side chain with Leu477 and Ser478 residues that it is not able to well accommodate within the access
channel of aromatase. This observation can contribute to explain the worst aromatase inhibitory activity
of the C-7 substituted steroids.
An additional series analysed was the C-6/C-7 hydroxyl series (Scheme 9). In this short series
there is a compound with a hydroxyl group in position 6α (20), a compound with the same substitution
in position 6β (19) and the corresponding analogous with a hydroxyl group in position 7α (21). By
analysing their aromatase inhibitory activities (Table 1), it is possible to infer, once more, that the C-6
substitution is preferable to the C-7 substitution. Further, the stereochemistry (α vs β) of the substituent
at C-6 does not seem to influence the aromatase inhibitory activity since their IC are similar (20 IC =
5
0
50
0
.88 µM and 19 IC = 0.84 µM).
50
ACS Paragon Plus Environment
1
6

Page 17 of 86
Journal of Medicinal Chemistry
Considering all series, it is even possible to ascertain which of the substitutions (methyl, allyl
1
2
and hydroxyl) is more convenient to achieve compounds with superior aromatase inhibition. This could
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
be done by comparing the activity of compounds 9/15/20 (IC = 0.06/0.105/0.88 µM) and 10a/16/21
5
0
(
IC = 0.27/0.59 µM /inhibition %= 49.83). From this, it is clear that the methyl group seems to be the
50
best substituent among those that were studied.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
In summary, it was disclosed for the first time, that 6α position, among 6α and 7α of the steroidal
androstane framework, is the best to be functionalized in order to reach superior aromatase inhibition. In
this work this was always true, except when there is a C-4 substituent, which can compete with the C-6
substituent to protrude into the access channel of the active site of aromatase, making the C-7
substituent more available to bulge in the referred channel, increasing the activity of the C-7 derivatives.
The methyl group proved to be the best substituent among those studied, followed by the allyl and
finally the hydroxyl group. Besides the presence of a carbonyl group at C-17 and a double bond at C-4,
which is known to be important for aromatase inhibition, the existence of an additional carbonyl group
at C-3 and a double bond at C-1 is very beneficial for the referred activity. With this study, very potent
+
aromatase inhibitors in human placental microsomes and in an ER breast cancer cell model were found.
In particular, compound 13, which combines the described structural features, showed a potency and
+
affinity to aromatase similar to exemestane, the reference compound used in clinic for ER breast cancer
treatment. Finally, molecular modelling studies guided by the GRID MIFs were useful to rationalize the
best inhibition potency of 13, which shares very similar binding geometries with respect to
androstenedione and exemestane, especially as concerns the 3-keto oxygen atom of the A-ring. This
new hit compound can compete by itself with exemestane and ultimately establish a base to develop
+
new lead and drug candidates AIs for the treatment of ER breast cancer.
Experimental Section
Chemistry
ACS Paragon Plus Environment
1
7

Journal of Medicinal Chemistry
Page 18 of 86
Melting points (Mps) were determined on a Reichert Thermopan hot block apparatus and were
not corrected. IR spectra were recorded on a Jasco 420FT/IR or a PerkinElmer Spectrum 400 FT-IR/FT-
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
NIR spectrometers using the universal ATR sampling accessory, CHCl solutions in NaCl plates, or
3
KBr disks.
1
NMR spectra were recorded on a Brucker AMX 300 spectrometer operating at 300 MHz ( H)
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
3
1
13
and at 75 MHz ( C) or on a Varian Unity 400 operating at 400 MHz ( H) and at 100 MHz ( C) and are
reported in  values in ppm downfield from TMS, as internal standard. All J values are given in Hz and
1
1
H− H couplings are assumed to be first-order. Peak multiplicities are reported in the usual manner and
b s, b d, b t and b m, means broad singlet, doublet, triplet or multiplet signals. Mass spectra ESI and LC-
MS were obtained with mass spectrometers QIT-MS Thermo Finningan, models LCQ Advantage MAX
and LXQ coupled to a Liquid Chromatograph of High Performance Thermo Finningan, model Accela.
6-Hydroxyandrost-4-ene-3,17-dione (19), 6-hydroxyandrost-4-ene-3,17-dione (20) and 7-
hydroxyandrost-4-ene-3,17-dione (21) were purchased from Steraloids, Inc. (Newport RI, USA). 3β-
Hydroxyandrost-5-en-17-one (26) and testosterone (42) were purchased from Sigma-Aldrich
(Schnelldorf, Germany). Reagents and solvents were used as obtained from the suppliers without further
purification, unless otherwise indicated. Compounds 14, 16 and 18, were prepared as previously
1
7
described .
All compounds evaluated possess a purity superior to 95%, except compound 1a (90% purity). The
purity of the compounds was checked by HPLC with a C18 - reversed phase column (ACE5 C18) and
water/acetonitrile (50:50, 40:60) and acetonitrile/methanol (60:40) were used as solvents. The purity of
individual compounds was determined from the area peaks in the chromatogram of the sample solution.
1
The results presented for compound 2b refers to a mixture of 2b with 6 (7.6:2.4, by H NMR).
6
α-Methyl-5α-androst-3-en-17-one (1a) and 6α-methyl-5β-androst-3-en-17-one (1b)
ACS Paragon Plus Environment
1
8

Page 19 of 86
Journal of Medicinal Chemistry
To a boiling solution of 9 (259.3 mg, 0.863 mmol) in glacial acetic acid (40 mL), zinc dust <10 μm
1.48 g, 22.66 mmol) was added in several portions during 1 h. After 3 h another portion of zinc (1.48 g,
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
(
22.66 mmol) was added within 1 h and the reaction was left under reflux for a total of 5 h 15 min. The
suspension was filtered, the zinc was washed with glacial acetic acid and the filtrate was evaporated to
dryness. The obtained residue was diluted with water (100 mL), extracted with diethyl ether (3x 100
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
mL) and the organic layer washed with 10% aq NaHCO (3x 100 mL) and water (3x 100 mL), dried
3
over anhyd Na SO , filtered and concentrated to dryness. NMR analysis of the crude revealed that it was
2
4
composed by a 1.7:1 mixture of isomers 1a and 1b, respectively. The crude was purified by silica gel
column chromatography (hexane/ethyl acetate) allowing isolating in some fractions the pure 1b as a
white solid residue (32.3 mg, 13% yield). In other fractions, mixtures with different proportions of
isomers 1a and 1b were obtained (193.7 mg, 78% yield). Two sequential chromatographies allowed
obtaining compound 1a in 90% purity.
-1
6α-Methyl-5α-androst-3-en-17-one (1a): IR (NaCl plates, CHCl ) ν
cm : 3030 (=C-H), 1740
3
max
1
(
C=O), 1651 (C=C). H NMR (400 MHz, CDCl ) δ: 0.80 (3H, s, 19-H ), 0.87 (3H, s, 18-H ), 0.93 (3H,
3
3
3
d, J=6.1 Hz, 6α-CH ), 5.61 (1H, ddd, J =10.0 Hz, J3-2β=6.5 Hz, J3-2α=3.3 Hz, 3-H), 5.67 (1H, ddd, J4-3=
3
3-4
1
3
1
1
1
6
0.0 Hz, J4-5α=3.9 Hz, J4-2α=2.0 Hz, 4-H). C NMR (100 MHz, CDCl ) δ: 12.6 (C-19), 13.9 (C-18),
3
9.9 (C-20), 20.3, 21.7, 23.2, 29.5, 31.6, 34.2, 34.7, 35.3, 35.8, 35.9, 47.8, 51.4, 52.3, 53.3, 126.2 (C-3),
+
27.2 (C-4), 221.4 (C-17). ESI: 287.3 ([M+H] , 100%).
α-Methyl-5β-androst-3-en-17-one (1b): Mp(hexane/ethyl acetate) 84-87 ºC. IR (NaCl plates, CHCl ) ν
3
max
-1
1
cm : 3025 (=C-H), 1740 (C=O), 1647 (C=C). H NMR (400 MHz, CDCl ) δ: 0.86 (3H, s, 18-H ), 0.96
3
3
(
3H, d, J=6.9 Hz, 6α-CH ), 0.97 (3H, s, 19-H ), 2.03 (1H, ddd, J16α-16β=19.0 Hz, J16α-15β=9.0 Hz, J16α-
3 3
15α=
9.0 Hz, 16α-H), 2.42 (1H, dd, J16β-16α=19.0 Hz, J16β-15β=9.0 Hz, 16β-H), 5.56 (1H, dd, J =10.3 Hz,
4-3
1
3
J4-5β=1.7 Hz, 4-H), 5.70 (1H, dd, J =9.8 Hz, J =3.5 Hz, 3-H). C NMR (100 MHz, CDCl ) δ: 13.8
3-4
3-2α
3
(C-18), 14.1 (C-20), 19.3 (C-19), 20.8, 21.8, 22.3, 22.7, 22.9, 29.3, 31.8, 34.3, 34.5, 35.6, 35.9, 40.8,
4
9.1, 51.1, 125.7 (C-4), 127.9 (C-3), 221.5 (C-17).
ACS Paragon Plus Environment
1
9

Journal of Medicinal Chemistry
Page 20 of 86
7
α-Methyl-5α-androst-3-en-17-one (2a) and 7α-methyl-5β-androst-3-en-17-one (2b)
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
To a boiling solution of 10a (150.4 mg, 0.5 mmol) in glacial acetic acid (20 mL), zinc dust <10 μm
(2.29 g, 35 mmol) was added in several portions over 22 h. The zinc suspension was filtered, washed
with glacial acetic acid and the filtrate was evaporated to dryness. The residue was diluted with water
(70 mL) and extracted with diethyl ether (3x 70 mL). The organic layer was washed with 10% aq
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
NaHCO (3x 100 mL) and water (3x 100 mL), dried over anhyd MgSO , filtered and concentrated to
3
4
dryness giving a yellow oily residue which was purified by silica gel 60 column chromatography
hexane/ethyl acetate). In some fractions an inseparable mixture of 2b with 6 was obtained (18.53 mg,
b (76%) and 6 (24%), by NMR). In other fractions, 2b was obtained in a mixture with 2a and 6 (46.4
(
2
mg, 2b (58%), 2a (15%) and 6 (27%), by NMR).
Although 2a and 2b could not be isolated in the pure form, their characterization could be performed by
NMR analysis of the mixture.
1
7α-Methyl-5α-androst-3-en-17-one (2a): H NMR (400 MHz, CDCl ) δ: 0.81 (3H, s, 19-H ), 0.89
3
3
(3H, s, 18-H ), 0.95 (3H, d, J=7.1, 7α-CH ), 5.23 (1H, ddd, J =8.0 Hz, J4-5=3.8 Hz, J4-2=1.9 Hz, 4-
3 3 4-3
H), 5.58 (1H, b m, W ca. 4 Hz, 3-H).
1
/2
1
7α-Methyl-5β-androst-3-en-17-one (2b): H NMR (400 MHz, CDCl ) δ: 0.87 (3H, s, 18-H ), 0.92
3
3
13
(3H, d, J=7.4 Hz, 7α-CH ), 1.00 (3H, s, 19-H ), 5.53 (2H, s, 3-H and 4-H). C NMR (100 MHz,
3 3
CDCl ) δ: 13.4 (C-18), 15.4 (C-7’), 20.4, 21.6, 21.8, 22.7 (C-19), 29.5, 31.6, 33.2, 34.2, 34.2, 34.4, 35.8,
3
36.9, 42.9, 47.3, 47.7, 124.1 (C-3), 134.7 (C-4), 221.5 (C-17).
3
α,4α-Epoxy-6α-methyl-5α-androstan-17-one (3a) and 3β,4β-epoxy-6α-methyl-5β-androstan-17-
one (3b)
To a stirred solution of compounds 1a and 1b (112.4 mg, 0.39 mmol) in DCM (30 mL) at room
temperature, performic acid generated in situ by the addition of 98-100% HCOOH (0.10 mL) and 30%
H O (0.21 mL) was added and the reaction was stirred until complete transformation of the starting
2
2
material (9 h 30 min, TLC). DCM (150 mL) was added and the organic layer was washed successively
ACS Paragon Plus Environment
2
0

Page 21 of 86
Journal of Medicinal Chemistry
with 10% aq NaHCO (100 mL) and water (4x 100 mL), dried over anhyd Na SO , filtered and
3
2
4
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
concentrated to dryness giving a slightly yellow oily residue. This residue was purified by a silica gel
column chromatography (hexane/ethyl acetate) allowing to isolate 52.0 mg (44%) of the pure 3α,4α-
isomer 3a and 34.6 mg (29%) of the pure 3β,4β-isomer 3b, as white solids.
3
α,4α-Epoxy-6α-methyl-5α-androstan-17-one (3a): Mp(hexane/ethyl acetate) 153-156 ºC. IR (NaCl plates,
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
-1
1
CHCl ) ν cm : 1740 (C=O), 1013 (C-O). H NMR (400 MHz, CDCl ) δ: 0.79 (3H, s, 19-H ), 0.85
3
max
3
3
(
3H, s, 18-H ), 1.07 (3H, d, J=6.4 Hz, 6α-CH ), 2.43 (1H, dd, J16β-16α=19.2 Hz, J16β-15β=8.7 Hz, 16β-H),
3
3
1
3
2
.92 (1H, dd, J4β-5α=4.1 Hz, J4β-6β=1.3 Hz, 4β-H), 3.15 (1H, dd, J3β-2β=3.4 Hz, J3β-2α=3.4 Hz, 3β-H). C
NMR (100 MHz, CDCl ) δ: 13.8 (C-19), 13.9 (C-18), 19.8, 20.6, 20.9, 21.7 (C-20), 29.5, 30.8, 31.5,
3
34.4, 34.7, 35.9, 40.2, 47.7, 51.2, 51.9, 52.3, 53.3 (C-4), 53.4 (C-3), 221.0 (C-17). ESI: 303.1 ([M+H]+,
1
00%).
3
β,4β-Epoxy-6α-methyl-5β-androstan-17-one (3b): Mp(hexane/ethyl acetate) 85-87 ºC. IR (NaCl plates,
-1
1
CHCl ) ν cm : 1738 (C=O), 1050 (C-O). H NMR (400 MHz, CDCl ) δ: 0.84 (3H, s, 18-H ), 0.87
3
max
3
3
(
3H, s, 19-H ), 1.06 (3H, d, J=6.9 Hz, 6α-CH ), 2.44 (1H, dd, J16β-16α=18.3 Hz, J16β-15β=8.7 Hz, J16β-
3
3
15α=
0.8 Hz, 16β-H), 2.91 (1H, d, J4α-5β=3.9 Hz, 4α-H), 3.17 (1H, dd, J3α-2α=2.5 Hz, J3α-2β=2.5 Hz, 3α-H).
1
3
C NMR (100 MHz, CDCl ) δ: 13.6 (C-18), 13.7 (C-19), 18.9, 20.2 (C-20), 22.1, 28.9, 29.5, 31.6, 33.6,
3
3
5.1, 35.5, 35.80, 35.81, 35.82, 42.6, 48.9, 51.1, 52.9 (C-4), 53.0 (C-3), 220.9 (C-17).
3
β,4β-Epoxy-7α-methyl-5β-androstan-17-one (4)
To a solution of 2b (in mixture with 2a (15%) and 6 (27%)) (24.6 mg, 0.086 mmol) in DCM (4 mL), a
solution of performic acid, generated in situ by addition of 98−100% HCOOH (0.015 mL) to 35% H O
2
2
(0.035 mL) was added. The reaction mixture was stirred at room temperature until complete
transformation of the starting material (8 h, TLC). DCM (50 mL) was added, and the organic layer was
washed successively with 10% aq NaHCO (3x 70 mL) and water (4x 100 mL), dried over anhyd
3
MgSO , filtered and concentrated to dryness. The obtained residue was purified by silica gel 60 column
4
1
chromatography (hexane/ethyl acetate) allowing isolating 14.6 mg (56%) of 4 as an oil. H NMR (400
ACS Paragon Plus Environment
2
1

Journal of Medicinal Chemistry
Page 22 of 86
MHz, CDCl ) δ: 0.86 (3H, s, 18-H ), 0.92 (3H, s, 19-H ), 0.94 (3H, d, J=7.4 Hz, 7α-CH ), 2.45 (1H,
3
3
3
3
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
ddd, J16β-16α=18.3 Hz, J16β-15β=9.0 Hz, J16β-15α=1.1 Hz, 16β-H), 2.96 (1H, d, J4α-3α=3.7 Hz, 4α-H), 3.25
1
3
(1H, m, J3α-4α=3.7 Hz, 3α-H). C NMR (100 MHz, CDCl ) δ: 12.7 (C-18), 13.5 (C-19), 14.1 (C-7’),
3
1
2
9.3, 19.9, 20.9, 21.6, 27.9, 28.0, 30.8, 31.9, 32.8, 33.1, 35.1, 35.8, 41.4, 46.6, 52.7 (C-4), 56.3 (C-3),
21.0 (C-17). ESI: 303.3 ([M+H]+).
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
6
α-Methylandrost-4-en-17-one (5)
Preparation of 5 was performed by adapting a described strategy.39
To a stirred and cooled mixture of trifluoracetic acid (0.36 mL), glacial acetic acid (0.36 mL) and
acetonitrile (0.36 mL), sodium borohydride (83.1 mg, 1.48 mmol) was added in small portions followed
by a solution of 9 (83.1 mg, 1.48 mmol) in anhyd DCM (8 mL). The reaction was stirred at room
temperature under a nitrogen atmosphere until complete transformation of the starting material (2 h,
TLC). The reaction mixture was neutralized with 10% aq NaHCO and, after dilution with water (50
3
mL), the aqueous phase was extracted with DCM (3x 50 mL). The organic layer was washed with water
(
3x 50 mL), dried over anhyd Na SO , filtered and concentrated to dryness giving an oily residue. This
2 4
residue was purified by silica gel column chromatography (petroleum ether 60-80ºC/ethyl acetate)
affording in some fractions 11.1 mg of the desired pure 5 (14%) and in other fractions 33.8 mg of the C-
1
7 reduced by-product 31 in 42.4% yield. Oxidation of 31, according to the procedure described below,
allowed its reconversion in 5.
To a stirred and cooled (0 ºC) solution of 31 (33.8 mg, 0.117 mmol) in acetone (8 mL), Jones reagent
was added dropwise until a persistent brown colour appears. After 5 min of reaction, the excess of the
oxidant was destroyed by addition of 2-propanol, until a green colour appears. The reaction mixture was
poured in water (50 mL) and the aqueous phase was extracted with ethyl acetate (3x 50 mL). The
organic layer was washed with 10% aq NaHCO (2x 50 mL) and water (2x 50 mL), dried over anhyd
3
Na SO , filtered and concentrated to dryness to give a white solid residue. This residue was purified by
2
4
silica gel column chromatography (petroleum ether 60-80 ºC/ethyl acetate) allowing isolating an
ACS Paragon Plus Environment
2
2

Page 23 of 86
Journal of Medicinal Chemistry
1
additional amount of 26.6 mg (79%) of the pure 5. Mp(petroleum ether/ethyl acetate) 122-124 ºC. H NMR (400
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
MHz, CDCl ) δ: 0.89 (3H, s, 18-H ), 1.11 (3H, d, J=7.6 Hz, 6α-CH ), 1.27 (3H, s, 19-H ), 2.45 (1H, dd,
3
3
3
3
1
3
J16β-16α=19.2 Hz, J16β-15β=8.5 Hz, 16β-H), 5.32 (1H, b dd, J=2.6 and 2.4 Hz, 4-H). C NMR (100 MHz,
CDCl ) δ: 13.8 (C-18), 18.9, 19.0 (C-6’), 20.3 (C-19), 20.7, 21.8, 26.0, 31.6, 32.3, 35.3, 35.9, 37.4, 38.4,
3
4
0.9, 47.8, 51.2, 54.6, 116.7 (C-4), 147.9 (C-5), 221.3 (C-17). ESI: 287.3 ([M+H]+).
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
7
α-Methylandrost-4-en-17-one (6)
To a stirred and cooled (0 ºC) mixture of trifluoracetic acid (0.5 mL), glacial acetic acid (0.5 mL) and
acetonitrile (0.5 mL), sodium borohydride (75.8 mg, 1.97 mmol) was added in small portions followed
by a solution of 10a (120.0 mg, 0.399 mmol) in anhyd DCM (12 mL). The reaction was stirred at room
temperature under nitrogen atmosphere until complete transformation of the starting material (3 h 40
min, TLC). The reaction mixture was neutralized with 10% aq NaHCO and the aqueous phase was
3
extracted with DCM (2x 50 mL). The organic layer was washed with water (3x 50 mL), dried over
anhyd MgSO , filtered and concentrated to dryness giving an oil residue. This residue was purified by
4
silica gel column chromatography (hexane/ethyl acetate) affording in some fractions 9.6 mg the pure 6
(10%) and in other fractions 71.4 mg of its C-17 reduced derivative, the 7α-methylandrost-4-en-17β-ol
2
5 (62%) formed as by-product. Oxidation of 25, according to the procedure described below, allowed
its reconversion in 6.
To a stirred and cooled (0 ºC) solution of 25 (70 mg, 0.24 mmol) in acetone (20 mL), Jones reagent was
added dropwise until a persistent brown colour appears. After 5 min of reaction, the excess of the
oxidant was destroyed by addition of 2-propanol, until a green colour appears. The reaction mixture was
poured in water (70 mL) and the aqueous phase was extracted with ethyl acetate (3x 70 mL). The
organic layer was washed with 10% aq NaHCO (4x 70 mL) and water (4x 70 mL), dried over anhyd
3
Na SO , filtered and concentrated to dryness giving 68.5 mg of the pure 6 (98%). Mp
(hexane/ethyl acetate)
2
4
1
1
35-137 ºC. H NMR (400 MHz, CDCl ) δ: 0.77 (3H, d, J=7.0 Hz, 7α-CH ), 0.88 (3H, s, 18-H ), 1.04
3 3 3
1
3
(
3H, s, 19-H ), 5.28 (1H, b s, 4-H). C NMR (100 MHz, CDCl ) δ: 12.3 (C-7’), 13.6 (C-18), 19.2, 19.7
3
3
ACS Paragon Plus Environment
2
3

Journal of Medicinal Chemistry
Page 24 of 86
(
C-19), 20.6, 21.4, 25.7, 30.2, 31.4, 35.7, 37.3, 37.87, 37.89, 40.0, 46.5, 47.4, 47.7, 121.6 (C-4), 140.9
C-5), 221.3 (C-17). ESI: 287.3 ([M+H]+).
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
(
4
α,5α-Epoxy-6α-methylandrostan-17-one (7a) and 4β,5β-epoxy-6α-methylandrostan-17-one (7b)
To a solution of 5 (27.7 mg, 0.097 mmol) in DCM (5 mL), performic acid (generated in situ by addition
of 98−100% HCOOH (0.05 mL)) and 35% H O (0.04 mL), was added. The reaction mixture was
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
2
2
stirred at room temperature until complete transformation of the starting material (6 h, TLC). DCM (50
mL) was added, and the organic layer was washed successively with 10% aq NaHCO (2x 50 mL) and
3
water (2x 50 mL), dried over anhyd Na SO , filtered, and concentrated to dryness. The obtained residue
2
4
was purified by two sequential silica gel 60 column chromatographies (petroleum ether 60-80 ºC/ethyl
acetate) and (petroleum ether 60-80 ºC/chloroform) allowing the isolation of 9.2 mg (31%) of the pure
7
a and 11.2 mg (38%) of the pure 7b.
1
4α,5α-Epoxy-6α-methylandrostan-17-one (7a). Mp(petroleum ether/chloroform) 116-118 ºC. H NMR (400
MHz, CDCl ) δ: 0.71 (3H, d, J=6.7 Hz, 6α-CH ), 0.89 (3H, s, 18-H ), 1.02 (3H, s, 19-H ), 2.45 (1H, dd,
3
3
3
3
1
3
J16β-16α=19.2 Hz, J16β-15β=8.7 Hz, 16β-H), 3.1 (1H, dd, J=2.6 and 2.2 Hz, 4β-H). C NMR (100 MHz,
CDCl ) δ: 13.8 (C-6’), 14.5 (C-18), 14.9, 19.5 (C-19), 20.6, 21.8, 23.1, 29.7, 30.8, 31.4, 34.8, 35.8, 37.1,
3
3
8.4, 46.7, 47.8, 51.1, 54.5 (C-4), 67.5 (C-5), 220.9 (C-17). ESI: 303.3 ([M+H]+).
β,5β-Epoxy-6α-methylandrostan-17-one (7b). Mp(petroleum ether/chloroform) 134-137 ºC. H NMR (400
1
4
MHz, CDCl ) δ: 0.68 (3H, d, J=6.7 Hz, 6α-CH ), 0.88 (3H, s, 18-H ), 1.08 (3H, s, 19-H ), 2.43 (1H, dd,
3
3
3
3
1
3
J16β-16α=19.3 Hz, J16β-15β=8.5 Hz, 16β-H), 3.14 (1H, d, J=3.5 Hz, 4α-H). C NMR (100 MHz, CDCl ) δ:
3
1
4
3.8 (C-6’), 14.6 (C-18), 15.8, 17.6 (C-19), 20.0, 21.7, 22.5, 29.4, 31.4, 31.9, 35.2, 35.8, 36.3, 36.6,
7.8, 50.5, 50.7, 56.3 (C-4), 67.2 (C-5), 221.2 (C-17). ESI: 303.3 ([M+H]+).
4
α,5α-Epoxy-7α-methylandrostan-17-one (8)
To a solution of 6 (75.1 mg, 0.262 mmol) in DCM (20 mL), m-chloroperbenzoic acid ≤77% (68.1 mg,
0
.393 mmol) was added. The reaction mixture was stirred at room temperature until complete
ACS Paragon Plus Environment
2
4

Page 25 of 86
Journal of Medicinal Chemistry
transformation of the staring material (2 h, TLC). DCM was added (50 mL) and the organic layer was
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
washed successively with 10% aq NaHCO (10x 100 mL) and water (6x 100 mL), dried over anhyd
3
MgSO , filtered and concentrated to dryness giving a slightly yellow solid residue. This residue was
4
purified by silica gel column chromatography (hexane/ethyl acetate) allowing isolating 68.4 mg (90%)
1
of the pure 8 as a white solid. Mp(hexane/ethyl acetate) 139-141 ºC. H NMR (400 MHz, CDCl ) δ: 0.89 (3H,
3
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
s, 18-H ), 0.94 (3H, d, J=7.2 Hz, 7α-CH ), 1.03 (3H, s, 19-H ), 2.35 (1H, dd, J16α-16β=13.4 Hz, J
16α-
3
3
3
15=
4.8 Hz, 16α-H), 2.46 (1H, ddd, J16β-16α=13.4 Hz, J16β-15β=8.8 Hz, J16β-15α=1.0 Hz, 16β-H), 2.94 (1H, d,
1
3
J=4.5 Hz, 4β-H). C NMR (100 MHz, CDCl ) δ: 13.3 (C-18), 13.6 (C-7’), 15.7, 19.0 (C-19), 20.6,
3
2
1.4, 23.6, 29.7, 29.9 (2 carbons), 31.3, 35.7, 36.8, 37.1, 38.5, 39.0, 47.5, 47.7, 62.4 (C-4), 220.8 (C-17).
ESI: 303.3 ([M+H]+).
6
α-Methylandrost-4-ene-3,17-dione (9)
Preparation of 9, as in scheme 4, was performed by adapting a described strategy.44
To a solution of 30 (205.1 mg, 0.644 mmol) in 0.5 N HCl/CH COOH (13 mL) was stirred at room
3
temperature until total transformation of the starting material (5 h 45 min, TLC). DCM (80 mL) was
added and the organic phase washed with 10% aq NaHCO (3x 80 mL) and water (3x 80 mL), dried
3
over anhyd Na SO , filtered and concentrated to dryness giving a yellow oily residue. This residue was
2
4
purified by silica gel column chromatography (petroleum ether 60-80ºC/ethyl acetate) affording the pure
2
7
9
(90.3 mg, 47%) as a white solid. Mp(hexane/ethyl acetate) 163-166 ºC [lit., 166-168 ºC]. IR (NaCl plates,
-1
1
CHCl ) ν cm : 3050 (=C-H), 1739 (C=O), 1671 (C=C). H NMR (400 MHz, CDCl ) δ: 0.92 (3H, s,
3
max
3
13
1
8-H ), 1.09 (3H, d, J=6.5 Hz, 6α-CH ), 1.21 (3H, s, 19-H ), 5.80 (1H, b s, 4-H). C NMR (100 MHz,
3 3 3
CDCl ) δ: 13.7 (C-18), 18.3 (C-20), 18.4 (C-19), 20.5, 21.7, 31.3, 33.6, 33.7, 35.0, 35.8, 35.9, 38.9,
3
39.7, 47.5, 50.7, 54.1, 121.6 (C-4), 173.5 (C-5), 199.6 (C-3), 220.3 (C-17). ESI: 301.3 ([M+H]+).
Preparation of 9, as in scheme 2, described in Supporting Information Section according the previously
reported.32
ACS Paragon Plus Environment
2
5

Journal of Medicinal Chemistry
Page 26 of 86
7
α-Methylandrost-4-ene-3,17-dione (10a) and 7β-methylandrost-4-ene-3,17-dione (10b)
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Preparation of 10a and 10b was performed by adapting a described strategy.37, 38
To a stirred solution of 24 (243.0 mg, 0.85 mmol) in anhyd THF (5 mL), CuBr (5.4 mg, 0.035 mmol)
was added under nitrogen atmosphere. Trimethylaluminium (0.97 mL, 1.94 mmol) and trimethylsilyl
chloride (0.27 mL, 2.11 mmol) were sequentially added through a syringe. The reaction proceeded with
magnetic stirring at room temperature under nitrogen for 2 h 20 min, time after which water (0.18 mL)
was added. The reaction was left to react until complete transformation of the starting material (24 h,
TLC). THF was removed under vacuum, water (100 mL) was added and the aqueous phase extracted
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
with DCM (150 mL). The organic layer was washed with 10% aq HCl (100 mL), 10% aq NaHCO (100
3
mL) and water (3x 100 mL), dried over anhyd Na SO , filtered and concentrated to dryness giving a
2
4
yellow crystalline residue. Silica gel 60 column chromatography (chloroform/diethyl ether) purification
afforded the pure 10a (77.8 mg, 30%) and its isomer 10b (19.8 mg, 10.6%).
1
7α-Methylandrost-4-ene-3,17-dione (10a): Mp(chloroform/diethyl ether) 177-179 ºC. H NMR (400 MHz,
CDCl ) δ: 0.82 (3H, d, J=7.1, 7α-CH ), 0.92 (3H, s, 18-H ), 1.21 (3H, s, 19-H ), 5.75 (1H, d, J=1.8 Hz,
3
3
3
3
1
3
4
3
3
7
3
-H). C NMR (100 MHz, CDCl ) δ: 12.7 (C-7’), 13.6 (C-18), 17.8 (C-19), 20.3, 21.3, 30.4, 31.1, 33.9,
3
5.6, 35.9, 37.7, 38.8, 40.6, 46.2, 47.2, 47.5, 126.1 (C-4), 168.9 (C-5), 198.9 (C-3), 220.2 (C-17). ESI:
01.4 ([M+H]+).
-1
β-Methylandrost-4-ene-3,17-dione (10b): Mp(chloroform/diethyl ether) 131-133 ºC. IR (ATR) υmax cm :
1
009 (=C-H), 1737 (C =O), 1731 (C =O), 1678 (C=C). H NMR (400 MHz, CDCl ) δ: 0.92 (3H, s, 18-
17
3
3
1
3
H ), 1.13 (3H, d, J=5.9 Hz, 7β-CH ), 1.19 (3H, s, 19-H ), 5.73 (1H, s, 4-H). C NMR (100 MHz,
3
3
3
CDCl ) δ: 14.2 (C-18), 17.6 (C-7’), 20.8, 22.9 (C-19), 25.7, 31.4, 33.9, 35.8, 35.9, 37.5, 38.3, 42.5, 42.6,
3
4
8.5, 51.4, 54.0, 123.2 (C-4), 169.9 (C-5), 199.3 (C-3), 220.7 (C-17).
6
α-Allyl-4-hydroxyandrost-4-ene-3,17-dione (11)
To a stirred and cooled (-60 ºC) solution of anhyd DCM (2 mL) and anhyd dimethyl sulfoxide (DMSO)
(0.26 mL, 3.77 mmol) under nitrogen atmosphere, a solution of anhyd trifluoroacetic acid (TFAA) (0.13
ACS Paragon Plus Environment
2
6

Page 27 of 86
Journal of Medicinal Chemistry
mL, 0.934 mmol) in anhyd DCM (1 mL) was added dropwise through a dropping funnel. After 10 min
of reaction, a solution of 49 (60 mg, 0.174 mmol) in anhyd DCM (4 mL) was added dropwise through
another dropping funnel. After 3 h of reaction, triethylamine (TEA) (0.2 mL) was added, and reaction
proceeded for more 15 min. The temperature of the reaction mixture was raised up to room temperature
and an aqueous solution of 2N HCl (50 mL) was added. The aqueous phase was extracted with DCM
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
(
3x 100 mL) and the organic layer was washed with 10% aq NaHCO (50 mL) and water (3x 100 mL),
3
dried with anhyd Na SO , filtered and concentrated to dryness giving to a yellow oily crude. Preparative
2
4
TLC (diethyl ether/toluene) allowed the recovery of 21 mg of the starting material and 25 mg (64%) of
-1
the pure 11. Mp(ethyl acetate) 138-140 ºC. IR (ATR) υmax cm : 3393 (O-H), 1735 (C =O), 1657 (C =O),
17
3
1
1618 (C=C), 1247 (C -O). H NMR (300 MHz, CDCl ) δ: 0.94 (3H, s, 18-H ), 1.28 (3H, s, 19-H ), 3.29
4
3
3
3
(
1H, dd, J6β-7α=14.1 Hz, J6β-7β=6.6 Hz, 6β-H), 5.03 (2H, b m, W1/2 ca. 23 Hz, -CH=CH ), 5.81 (1H, m, b
2
1
3
m, W1/2 ca. 43 Hz, -CH=CH ), 6.18 (1H, s, 4-OH, disappears on adding D O). C NMR (75.5 MHz,
2
2
CDCl ) δ: 13.8 (C-18), 20.1 (C-19), 20.2, 21.8, 29.6, 31.3, 31.9, 32.3, 32.9, 35.7, 36.8, 37.5, 39.0, 47.5,
3
0.9, 53.6, 116.2, 137.3, 141.4, 142.2, 193.8 (C-3), 220.7 (C-17). ESI: 343.3 ([M+H]+).
5
7
α-Allyl-4-hydroxyandrost-4-ene-3,17-dione (12)
A solution of 41 (177.0 mg, 0.46 mmol) in ethanol/HCl [95:05] (48 mL) was stirred under reflux for 1h
0 min. After cooling the reaction mixture to room temperature, water (100 mL) was added and the
3
organic layer washed with 10% aq NaHCO (3x 100 mL) and water (100 mL), dried over anhyd
3
MgSO , filtered and concentrated to dryness giving a yellow oily crude which by trituration with cold
4
diethyl ether allowed the isolation of 74 mg (47%) of the pure 12, as white crystals. Mp(diethyl ether) 205-
1
2
07 ºC. H NMR (400 MHz, CDCl ) δ: 0.92 (3H, s, 18-H ), 1.22 (3H, s, 19-H ), 3.14 (1H, dd, J
3
3
3
6α-
6β=14.4 Hz, J6α-7β=1.8 Hz, 6α-H), 5.04 (1H, b m, W1/2 ca. 8 Hz, -CH=CH ), 5.06 (1H, b m, W ca. 6
2
1/2
13
Hz, -CH=CH ), 5.85 (1H, b m, W ca. 43 Hz, -CH=CH ), 6.07 (1H, s, 4-OH). C NMR (100 MHz,
2
1/2
2
CDCl ) δ: 13.5 (C-18), 17.55 (C-19), 17.57, 20.5, 21.3, 26.8, 30.6, 31.2, 31.8, 34.9, 35.6, 35.9, 37.9,
3
ACS Paragon Plus Environment
2
7

Journal of Medicinal Chemistry
Page 28 of 86
4
1
6.9, 47.5, 47.7, 116.2 (-CH=CH ), 137.0 (-CH=CH ), 137.1 (C-5), 143.1 (C-4), 193.2 (C-3), 220.3 (C-
2 2
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
7). ESI: 343.3 ([M+H]+).
6
α-Allylandrosta-1,4-diene-3,17-dione (13)
To a stirred solution of 15 (132.7 mg, 0.41 mmol) in toluene (18 mL), under reflux and nitrogen
atmosphere, DDQ (219.9 mg, 0.97 mmol) followed by benzoic acid (71.4 mg, 0.58 mmol) were added.
The reaction was stirred until complete transformation of the starting material (18 h 30 min, TLC). The
reaction mixture was cooled to room temperature and the suspension formed was filtered through an
alumina column and eluted with toluene followed by DCM. The obtained fractions were evaporated
under vacuum and purified by silica gel column chromatography (petroleum ether 60-80 ºC/ethyl
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
acetate) affording 108.2 mg (81%) of the pure 13. Mp(petroleum ether/ethyl acetate) 143-145 ºC. H NMR (400
MHz, CDCl ) δ: 0.94 (3H, s, 18-H ), 1.27 (3H, s, 19-H ), 5.08 (1H, b m, W ca. 5 Hz, -CH=CH ), 5.11
3
3
3
1/2
2
(1H, b m, W1/2 ca. 13 Hz, -CH=CH ), 5.80 (1H, b m, W ca. 41 Hz, -CH=CH ), 6.11 (1H, b s, 4-H),
2
1/2
2
13
6
.25 (1H, dd, J =10.2 Hz, J =1.8 Hz, 2-H), 7.05 (1H, d, J =10.2 Hz, 1-H). C NMR (100 MHz,
2-1 2-4 1-2
CDCl ) δ: 13.8 (C-18), 18.9 (C-19), 21.9, 22.2, 31.2, 35.1, 35.6, 35.9, 38.4, 38.5, 43.8, 47.7, 50.3, 53.5,
3
1
2
17.2 (-CH=CH ), 121.8 (-CH=CH ), 127.3 (C-4), 135.9 (C-2), 156.0 (C-1), 170.0 (C-5),186.1 (C-3),
2
2
19.8 (C-17). ESI: 325.3 ([M+H]+).
7
α-Allylandrosta-1,4-diene-3,17-dione (14)
Compound 14 was prepared as previously described.17
6
α-Allylandrost-4-ene-3,17-dione (15)
A solution of 35 (204.1 mg, 0.59 mmol) in 0.5N HCl in CH COOH (7 mL) was stirred at room
3
temperature until complete transformation of the starting material (17 h 30 min, TLC). The reaction
mixture was neutralized with 10% aq NaHCO and the aqueous phase was extracted with ethyl acetate
3
(2x 100 mL). The organic layer was washed with 10% aq NaHCO (100 mL) and water (2x 100 mL),
3
ACS Paragon Plus Environment
2
8

Page 29 of 86
Journal of Medicinal Chemistry
dried over anhyd MgSO , filtered and concentrated to dryness giving an oily residue. The residue was
4
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
purified by silica gel column chromatography (petroleum ether 60-80 ºC/ethyl acetate) affording 142.2
1
mg (74%) of the pure 15. Mp(petroleum ether/ethyl acetate) 100-102 ºC. H NMR (400 MHz, CDCl ) δ: 0.91 (3H,
3
s, 18-H ), 1.24 (3H, s, 19-H ), 5.06 (1H, b m, W ca. 12 Hz, -CH=CH ), 5.08 (1H, b t, W ca. 4 Hz, -
3
3
1/2
2
1/2
1
3
CH=CH ), 5.77 (1H, b m, W ca. 41 Hz, -CH=CH ), 5.08 (1H, d, J =1.1 Hz, 4-H). C NMR (100
2
1/2
2
4-6β
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
MHz, CDCl ) δ: 13.7 (C-18), 18.4 (C-19), 20.5, 21.7, 31.2, 33.5, 35.1, 35.71, 35.74, 36.3, 36.8, 38.3,
3
3
2
9.2, 47.6, 50.7, 54.1, 117.0 (-CH=CH ), 121.8 (C-4), 136.1 (-CH=CH ), 172.1 (C-5), 199.5 (C-3),
2 2
20.3 (C-17). ESI: 327.3 ([M+H]+).
7α-Allylandrost-4-ene-3,17-dione (16)
Compound 16 was prepared as previously described.17
6α-Allylandrost-4-en-17-one (17)
To a stirred and cooled (0 ºC) mixture of trifluoracetic acid (0.12 mL), glacial acetic acid (0.12 mL) and
acetonitrile (0.12 mL), sodium borohydride (21.1 mg, 0.48 mmol) was added in small portions followed
by a solution of 15 (32.0 mg, 0.098 mmol) in anhyd DCM (4 mL). The reaction was stirred at room
temperature under nitrogen atmosphere until complete transformation of the starting material (1 h 30
min, TLC). The reaction mixture was neutralized with 10% aq NaHCO and, after dilution with water
3
(50 mL), the aqueous phase was extracted with DCM (3x 20 mL). The organic layer was washed with
water (2x 100 mL), dried over anhyd MgSO , filtered and concentrated to dryness. The residue obtained
4
was purified by silica gel column chromatography (hexane/ethyl acetate) affording in some fractions 6.1
mg the desired 17 (20%) as an oil and in other fractions 20 mg (65%) of the C-17 reduced derivative 6α-
allylandrost-4-en-17β-ol (36) formed as by-product. Oxidation of 36, according to the procedure
described below, allowed its reconversion in 17.
To a stirred and cooled (0 ºC) solution of 36 (13.7 mg, 0.044 mmol) in acetone (5 mL), Jones reagent
was added dropwise until a persistent brown colour appears. After 5 min of reaction, the excess of the
ACS Paragon Plus Environment
2
9