Potent and selective steroidal inhibitors of 17β-hydroxysteroid dehydrogenase type 7, an enzyme that catalyzes the reduction of the key hormones estrone and dihydrotestosterone

Article abstract of DOI:10.1021/jm900921c

17β-Hydroxysteroid dehydrogenase type 7 (17β-HSD7) catalyzes the reduction of estrone (E₁) into estradiol (E₂) and of dihydrotestosterone (DHT) into 5α-androstane-3β,17β-diol (3β-diol), therefore modulating the level of mitogenic est

Full text of DOI:10.1021/jm900921c

7488 J. Med. Chem. 2009, 52, 7488–7502
DOI: 10.1021/jm900921c
Potent and Selective Steroidal Inhibitors of 17β-Hydroxysteroid Dehydrogenase Type 7, an Enzyme
That Catalyzes the Reduction of the Key Hormones Estrone and Dihydrotestosterone
ꢀ
Edith Bellavance, Van Luu-The, and Donald Poirier*
Laboratory of Medicinal Chemistry, Oncology and Molecular Endocrinology, CHUQ (CHUL) Research Center and Laval University,
ꢀ
2705 Boulevard Laurier, Quebec G1V 4G2, Canada
Received June 22, 2009
17β-Hydroxysteroid dehydrogenase type 7 (17β-HSD7) catalyzes the reduction of estrone (E1) into
estradiol (E2) and of dihydrotestosterone (DHT) into 5R-androstane-3β,17β-diol (3β-diol), therefore
modulating the level of mitogenic estrogens and androgens in humans. By classical and parallel
chemistry, we generated several 4-methyl-4-aza-5R-androstane derivatives differing in their C-17
substituent: 17β-formamide, 17β-benzamide, and 17β-tertiary amine. Best candidates in each category
had demonstrated good inhibitory potency toward the conversion of E1 into E2 (IC50 = 189-451 nM)
and also toward the conversion of DHT into 3β-diol (69-91% at 3 μM). Inhibition assays with 17β-
HSD1, 17β-HSD5, 5R-reductase (5R-R) 1 and 5R-R2 revealed that 17β-HSD7 inhibitors with a 4-
methyl-4-aza nucleus were also able to inhibit 5R-Rs but not the other enzymes tested. Two 4-aza-5R-
androstane inhibitors were, however, selective and still showed good inhibition of 17β-HSD7. First
selective and efficient inhibitors of 17β-HSD7 are now available for additional mechanistic and
therapeutic studies.
Introduction
DU-145, and LNCaP.5d,5e It converts E1 into the estrogen E2,
the most potent female hormone, and to the same extent also
deactivates the powerful androgen dihydrotestosterone
(DHT) into 5R-androstane-3β,17β-diol (3β-diol),5c an estro-
gen-receptor (ER) ligand (Figure 1).6 17β-HSD7 also func-
tions as a zymosterone reductase and is thus involved in
postsqualene cholesterogenesis.7 Contrary to other mamma-
lians in which 17β-HSD7 is obviously part of the pregnancy
process (namely, in the rabbit8 and marmoset monkey9), the
precise role of human 17β-HSD7 is not clear. Because of its
dual enzymatic activity and widespread distribution, the hu-
man enzyme seems to act as an intracrine regulator of steroid
metabolism, increasing the concentration of estrogens in its
surroundings. The impact of E2 on the development of certain
hormone-sensitive cancers is well-known, and the therapeutic
use of an inhibitor of its biosynthesis represents a relevant
strategy to counteract their proliferative effects.10 Design of
inhibitors targeting 17β-HSD7 is of interest, since it could help
to better modulate the in situ formation of E2 and thus better
regulate its proliferative effect in ERþ breast cancer cells. A
17β-HSD7 inhibitor could also help to maintain the level of
DHT, an androgenic hormone with an antiproliferative effect
on ERþ cells.11 Because of its localization at the end of the
metabolic pathway of sex hormone formation, inhibiting 17β-
HSD7 would not deplete the organism in any other family of
steroids such as mineralocorticoids and glucocorticoids.
Moreover, a selective 17β-HSD7 inhibitor would be a useful
tool to better understand the biological role of this new
steroidogenic enzyme.
17β-Hydroxysteroid dehydrogenases (17β-HSDsa) play a
crucial role in the biosynthesis of active androgens and
estrogens in both steroidogenic and peripheral target tissues.
They operate the conversion of 17-keto steroids to their active
17β-hydroxy-forms or vice versa, each member of the 17β-
HSDs family showing different tissular distributions and
having a preferred cofactor (NADþ or NADPH) and pre-
ferred substrates.1 17β-HSD7 was originally cloned in a rat
corpus luteum cDNA library and identified as the prolactin
receptor-associated protein2 and was then cloned in mouse
mammals gland epithelial cells and recognized as a 17β-HSD
isoform.3 The rodent enzyme was proven to efficiently reduce
estrone (E1) into estradiol (E2), and its tissue-specific expres-
sion pattern suggests an important function in sustaining
pregnancy, as well as a role in tissues prone to develop
hormone-related cancers.4 Inhumans, 17β-HSD7 isexpressed
in typical steroidogenic tissues such as testis and nonpregnant
women’s ovaries but also in the uterus, placenta, mammary
glands, prostate, liver, kidney as well as in neural tissues5 and
in some cancerous cell lines, like CAMA-1, MCF-7, ZR-75-1,
*To whom correspondence should be addressed. Phone: (418) 654-
2296. Fax: (418) 654-2761. E-mail: donald.poirier@crchul.ulaval.ca.
a Abbreviations: 17β-HSDs, 17β-hydroxysteroid dehydrogenases;
5R-Rs, 5R-reductases; NADþ, nicotinamide adenine dinucleotide;
NADPH, nicotinamide adenine dinucleotide phosphate reduced form;
ER, estrogen receptor; E1, estrone; E2, estradiol; DHT, dihydrotestos-
terone; 3β-diol, 5R-androstane-3β,17β-diol; Δ4-dione, 4-androstene-
3,17-dione; A-dione, 5R-androstane-3,17-dione; T, testosterone; SAR,
structure-activity relationship; IC50, the half maximal inhibitory con-
centration; Km, Michaelis-Menten constant; Kcat/Km, the rate of cata-
lytic efficiency; RT-PCR, reverse transcriptase polymerase chain
reaction; HEK, human embryonic kidney; MEM, minimum essential
medium; rt, room temperature.
A previous screening done with a series of 150 compounds
belonging to different classes (natural hormones, anti-hor-
mones, phytoestrogens, known inhibitors of other steroido-
genic enzymes) permitted us to point out four C-19 steroids
r
pubs.acs.org/jmc
Published on Web 09/22/2009
2009 American Chemical Society

Article
Journal of Medicinal Chemistry, 2009, Vol. 52, No. 23 7489
meter that is the length of the aliphatic side chain.12 We
therefore decided to further investigate the potential as
inhibitors of 17β-HSD7 of steroid 2 analogues by determin-
ing the optimal length of the N-alkyl moiety. Compounds
bearing five to eight side chain carbon atoms had already
been synthesized as inhibitors of 5R-reductases by Li et al.,15
whereas more hydrophobic analogues 10 and 11 were
synthesized from commercially available testosterone
(Scheme 1). The Rasmusson et al. method16 afforded the 4-
methyl-4-aza-5R-androstane-3,17-dione (4), which was
further oxidized by pyridinium chlorochromate (PCC) to
give the key intermediate 5 in moderate yield. The 17-keto
steroid was next submitted to reductive amination by gen-
erating the imine that was then reduced by either sodium
cyanoborohydride (NaBH3CN) in a slightly acidic milieu or
sodium borohydride (NaBH4) to afford the 17β-amino
steroids 6-9. Because of the proximity of the angular methyl
C-18, it was not possible to bring this reaction to completion,
regardless of the reaction time or an excess of amine added.
Furthermore, purification of compounds 6-9 by silica gel
chromatography was difficult, resulting in moderate yields
(44-75%) for this two-step sequence of reactions. Com-
pounds 10 and 11 were next obtained by formylation of their
precursors, compounds 7 and 8, respectively, after previous
activation of formic acid with dicyclohexylcarbodiimide
(DCC) following a standard procedure.15
Figure 1. Principal reactions performed by human 17β-HSD7 in
the presence of cofactor NADPH. The Km value is expressed in μM,
while the enzymatic efficiency (Kcat/Km) is expressed in U mg-1
μM-1 (data from ref 5c).
Further indications from our preliminary SAR study
suggested that a benzamido derivative such as 3 could
constitute an interesting family of 17β-HSD7 inhibitors.12
To confirm our hypothesis, we proceeded with the synthesis
of 17β-[(N-heptyl)arylamido] steroids 12-14 as well as
17β-[(N-heptyl)alkylamido] steroids 15-18. The amine pre-
cursor 6 was either reacted with the corresponding acyl
chloride or O-acylurea to provide compounds 12-18. Both
methods gave yields ranging between 66% and 80% except
for compounds 15 and 18 where the bulky group was added
less successfully in 51% and 41% yield, respectively.
To extend our SAR study, additional arylamido deriva-
tives were obtained by parallel synthesis. A first library
containing 27 members (compounds 19-45) was obtained
using peptide bond formation between amines 6, 8, and 9,
bearing 7, 10, and 12 carbon atoms, respectively, and a
selection of 9 acyl chlorides (Table 1). Briefly, each steroid
dissolved in dry THF was evenly distributed in nine wells of a
96 solid-phase reaction block (ACT Labtech apparatus).
Piperidinomethylpolystyrene resin and each acyl chloride
were added and mixtures kept under argon for 3 h at room
temperature (rt). Aminomethylpolystyrene resin was next
added to quench all excess acyl chloride, and the reaction
mixture was filtered to get rid of polymers. Collected filtrates
were evaporated and the residues dissolved in ethyl acetate to
finally be quickly purified on a silica gel pad. Since enzymatic
assay results indicated that para-substituted benzoyl groups
and fluoro atoms were prone to inhibit the conversion of E1
into E2 by 17β-HSD7, a second library containing 22 mem-
bers (compounds 46-67) was synthesized (Table 2). Eleven
new para-substituted or fluoro-substituted benzoyl chlorides
were then chosen and reacted with amines 6 and 9, affording
amides 46-67 in moderated to good yields. Reactions with 2-
fluoro-5-trifluoromethylbenzoyl chloride and 2-chloro-4-
fluorobenzoyl failed to afford the desired products 56, 57,
60, and 61.
Figure 2. Lead compounds identified from a preliminary screening
as inhibitors of 17β-HSD7 (E1 to E2 transformation). The stereo-
chemistry of carbons 5, 8, 9, and 14 is shown only for compounds 1a
and 1b but is the same for all other steroids.
(compounds 1a, 1b, 2, and 3) that were able to decrease the
transformation of E1 into E2 in intact h17β-HSD7-transfected
HEK-293 cells (Figure 2). Our preliminary study seems to
demonstrate that a 4-aza-5R-androstane nucleus is an impor-
tant structuralfeatureandthat its inhibitory potencyisgreatly
improved by introducing a long alkylamide side chain
(compounds 2 and 3) or a dimethylated spiro-δ-lactone
nucleus (compounds 1a and 1b) at position 17β.12 This latter
nucleus had been previously introduced on a C-18 steroid and
proven to be an excellent inhibitor of human 17β-HSD5,13
which is responsible for the transformation of androstene-
dione (Δ4-dione) into testosterone (T) in peripheral tissues.14
Furthermore, compound 2 is known as a good inhibitor of the
conversion of T into DHT and Δ4-dione into 5R-androstane-
3,17-dione (A-dione) by both type 1 and type 2 5R-reductases
(5R-Rs).15 On the basis of these structural similarities between
known inhibitors of steroidogenic enzymes and our lead
compounds, our challenge will consist of developing efficient
and selective inhibitors of 17β-HSD7. Efficiency will first be
optimized by following various combinations of alkyl side
chain length (7-12 carbons), amide, or amine moieties and
also through the investigation of the inhibitory potency of
substituted benzamide (derivatives of compounds 2 and 3).
Selectivity of 17β-HSD7 inhibitors will be assessed toward
other steroidogenic enzymes, like 17β-HSD5 and 5R-Rs, but
also toward 17β-HSD1 that catalyzes the conversion of E1
into E2 as well.
Results and Discussion
Chemical Synthesis. The preliminary structure-activity
relationship (SAR) data concerning 17β-(N-alkylamido)
derivatives seemed to converge toward an important para-
In addition to the 4-methyl-4-aza-5R-androstane (C-19
steroid) nucleus, the estrane (C-18 steroid) nucleus was also

7490 Journal of Medicinal Chemistry, 2009, Vol. 52, No. 23
Bellavance et al.
Table 1. First Library of 17β-[(N-Alkyl)arylamido] Derivatives 19-45 Obtained by Parallel Synthesis from Amines 6, 8, and 9 and the Percentages of
17β-HSD7 Inhibition (E1 to E2 Transformation) at 0.3 μMa
group A (n = 6)
inhibition, %
group B (n = 9)
inhibition, %
group C (n = 11)
inhibition, %
R
compd
compd
compd
4-Br-Ph
2-Cl-Ph
19
22
25
28
31
34
37
40
43
12
2
68 ( 1
37 ( 1
59 ( 7
65 ( 5
28 ( 3
84 ( 1
24 ( 4
50 ( 1
56 ( 6
56 ( 2b
94 ( 2
20
23
26
29
32
35
38
41
44
52 ( 2
37 ( 1
46 ( 3
58 ( 17
36 ( 1
71 ( 2
30 ( 5
37 ( 1
49 ( 2
21
24
27
30
33
36
39
42
45
52 ( 9
46 ( 2
51 ( 1
40 ( 4
18 ( 3
47 ( 1
45 ( 6
33 ( 11
51 ( 4
3-CF3-4-MeO-Ph
4-Cl-Ph
2,6-di-Cl-Ph
4-F-Ph
3-(ClCH2)-Ph
4-(ClCH2)-Ph
4-CN-Ph
Ph
H
a See Scheme 1 for the chemical structures. b Data obtained from another experiment using the same experimental conditions.
Scheme 1. Synthesis of 17β-[(N-Alkyl)alkyl/arylamido]-4-methyl-4-aza-5R-androstan-3-onesa
a Reagents and conditions: (a) PCC, NaOAc, CH2Cl2, molecular sieve, rt, 3 h (68%); (b) (i) NH2(CH2)nCH3 (n = 6 and 11), p-TSA, toluene, reflux in
Dean-Stark apparatus, 18-24 h; (ii) NaBH3CN, MeOH, pH 5, 0 °C, 1-2 h (58% and 44%, two steps); (c) (i) NH2(CH2)nCH3 (n = 7 and 8), p-TSA,
toluene, reflux in Dean-Stark apparatus, 18-24 h; (ii) NaBH4, MeOH, 0 °C, 1.5 h (75% and 62%, two steps); (d) (i) HCOOH, DCC, CHCl3, 0 °C, 5 min;
(ii) steroid, pyridine, 0 °C to rt (37-73%); (e) (i) K2CO3, THF, rt, 0.5 h; (ii) acyl/aryl chloride (RCOCl), overnight (41-74%); (f) (i) carboxylic acid
(RCOOH), DCC, CHCl3, 0 °C, 10 min; (ii) steroid, pyridine, K2CO3, 0 °C to rt, overnight (51-80%); (g) (i) piperidinomethylpolystyrene, acyl chlorides,
THF, rt, 3 h; (ii) aminomethylpolystyrene, rt, 2.5 h (62-99%).
investigated with a series of two compounds synthesized
from 3-methyl-O-estrone (68) (Scheme 2). The protected
estrone was transformed into the 17β-(N-heptylamino) deri-
vative 69 via a two-step reductive amination in the same
conditions as compounds 6-9 and with similar yields. The
intermediate 69 was also formylated to afford amide 70 in
good yield. This latter compound was then deprotected with
BBr3 to provide the final compound 71 with 65% yield.
Another class of compounds tested for their inhibitory
potency on 17β-HSD7 was tertiary amines (Scheme 3).
Tertiary amines 72-75 were synthesized from amine 6 using
either alkyl iodide (72 and 73) or alkyl bromide (74 and 75).17
To achieve completion of the reaction, it was important to
heat the mixture in a Schlenk tube to about 165 °C. It appears
that acetonitrile (MeCN) was the best solvent for this reac-
tion when compared to dimethylformamide (DMF) and
ethylene chloride. Yields are good (76-89%) except for
compound 75 that was obtained in only 20% yield even with
the use of 20 equiv of cyclohexylmethyl bromide. Compound
76 failed to be obtained by this procedure because the use of
methyl iodide in the presence of potassium or cesium carbo-
nate gave rise to a quaternary ammonium salt. Reducing the

Article
Journal of Medicinal Chemistry, 2009, Vol. 52, No. 23 7491
Table 2. Second Library of 17β-[(N-Alkyl)arylamido] Derivatives
46-67 Obtained by Parallel Synthesis from Amines 6 and 8 and the
Percentages of 17β-HSD7 Inhibition (E1 to E2 Transformation) at
0.3 μMa
Scheme 3. Synthesis of 17β-[(N-Heptyl)alkyl/arylamino]-4-
methyl-4-aza-5R-androstan-3-onesa
group A (n = 6)
group B (n = 9)
R
compd inhibition, % compd inhibition, %
2,6-di-F-Ph
2,5-di-F-Ph
2,4-di-F-Ph
3,5-di-F-Ph
2,3,4,5,6-penta-F-Ph
3-CF3-6-F-Ph
2-CF3-6-F- Ph
2-Cl-4-F- Ph
4-tBu-Ph
46
48
50
52
54
56
58
60
62
64
66
34
84 ( 5
59 ( 10
73 ( 5
56 ( 1
67 ( 3
b
47
49
51
53
55
57
59
61
63
65
67
56 ( 5
56 ( 5
80 ( 12
38 ( 7
28 ( 5
b
49 ( 6
b
19 ( 3
b
29 ( 4
24 ( 1
60 ( 1
77 ( 3
4 ( 5
11 ( 7
39 ( 2
4-Heptyl-Ph
4-I-Ph
4-F-Ph
a See Scheme 1 for the chemical structures. b Compound not obtained
in suitable purity to be tested.
Scheme 2. Synthesis of 17β-[(N-Heptyl)formamido]estra-1,3,
5(10)-trienesa
a Reagents and conditions: (a) K2CO3, alkyl iodide or bromide,
MeCN, Schlenk tube, 165 °C, 24 h (20-89%); (b) (CH2O)n, NaBH3CN,
refluxing MeOH, 24 h (50%).
provide amines 78 and 79 in 54% and 30% yields, respectively.
N-Methylation of 79, by the method described for compound
76, afforded compound 81 in 64% yield, but formylation of 78
was more difficult than its 4-methyl-4-aza analogue. Indeed,
compound 80 showed a very similar solubility to the dicyclo-
hexylurea side product and it was isolated with difficulty in
only a 30% yield.
Biological Activity. Inhibition toward 17β-HSD7 (Trans-
formation of E1 into E2). The in vitro enzymatic assays were
realized in intact HEK-293 cells overexpressing 17β-HSD7, a
system that offers precious advantages. First, as the molecule
needs to penetrate the cell to “meet” the enzyme, working
with intact cells automatically eliminates compounds that
are not able to enter the cells. This methodology may discard
molecules that could have been good inhibitors in the case of
purified enzymes or cell homogenates, but such compounds
are of less interest in a therapeutic point of view. The whole
cell system also constitutes a more physiological environ-
ment than a homogenate or purified enzyme because no
exogenous cofactor needs to be added. A lower concentra-
tion of the substrate than the Km value of the enzyme was
then used for the assay. Finally, the transfected HEK-293
cells overexpress the targeted enzyme activity when com-
pared to wild type cells, which only express low levels of
steroidogenic enzyme activities. For screening purposes, the
assays were performed at two inhibitor concentrations and
results are expressed as the inhibition (%) of the enzymatic
reduction of E1 into E2 by 17β-HSD7.
Enzymatic assays performed with 17β-[(N-alkyl)forma-
mido]-4-methyl-4-aza-5R-androstanes showed that a side
chain length between 7 and 10 carbon atoms (n = 6-9) is
better for inhibition than a shorter (n = 4 and 5) side chain
(Figure 3). The plateau reached by long aliphatic chains may
suggest that the enzyme possesses an important hydrophobic
pocket. On the basis of these results, a series of 17β-[(N-
heptyl)alkyl/arylamido] derivatives were synthesized to in-
vestigate the impact of the nature of the amide moiety
a Reagents and conditions: (a) (i) NH2(CH2)6CH3, p-TSA, toluene,
reflux in Dean-Stark apparatus, 18-24 h; (ii) NaBH3CN, MeOH, pH
5, 0 °C, 1-2 h (64%, two steps); (b) (i) HCOOH, DCC, CHCl3, 0 °C,
10 min; (ii) steroid, pyridine, K2CO3, 0 °C to rt (84%); (c) BBr3, CH2Cl2,
0 °C to rt (47-65%).
amount of methyl iodide to 1.5 equiv and lowering the
reaction temperature to 50 °C gave the same result. Reduc-
tion by NaBH3CN of a methylimine, formed by the reaction
between the amine 6 and formaldehyde, was then considered,
but results were not convincing. However, two methods had
helped us obtain the N-methylated compound: the Eschwei-
ler-Clarke methylation,18 which gave a crude yield of 77%
on a chemical model, and the reduction by NaBH3CN of a
methylimine resulting from the reaction of paraformalde-
hyde and amine 6. The latter reaction afforded compound 76
with a 50% yield, which is a little bit lower than the typical
reported yield of 66%.19 This result was, however, found
suitable considering the steric hindrance caused by the
angular C-18 methyl of the steroid nucleus.
The 4-aza-5R-androstane counterparts of compounds 11
and 76, namely, compounds 80 and 81, were synthesized
starting with 4-aza-5R-androstane-3,17-dione (77) (Scheme 4).
This latter compound was submitted to reductive amination to

7492 Journal of Medicinal Chemistry, 2009, Vol. 52, No. 23
Bellavance et al.
Figure 3. Optimization of the inhibitory potency (%) according to
the alkylamide side chain length of various 17β-[(N-alkyl)forma-
mido] derivatives. Transformation of E1 to E2 was by 17β-HSD7.
Products were tested in triplicate at the indicated concentrations.
Groups A-C compounds were available in our laboratory, and the
chemical synthesis was previously reported.15
Figure 4. Comparison of the inhibitory potency (%) reached by
17β-[(N-alkyl)alkylamido] derivatives and 17β-[(N-alkyl)arylamido]
derivatives for the transformationofE1 toE2 by17β-HSD7. Products
were tested in duplicate at the indicated concentrations. See Scheme 1
for the structures of tested compounds.
Scheme 4. Synthesis of 4-Aza-5R-androstane Analogues of
Two Potent Inhibitors of 17β-HSD7a
Figure 5. Inhibitory potency (%) reached by C-18 steroid deriva-
tives for the transformation of E1 to E2 by 17β-HSD7. Products
were tested in triplicate at the indicated concentrations. See
Scheme 2 for the structures of tested compounds.
ever, no significant difference was observed between the
inhibition reached by a benzamide substituted by an electron
withdrawing group (compound 13) or an electron donating
group (compound 14) when compared to the unsubstituted
benzamide 12. A short alkyl chain is also better tolerated
than one that is too long (for example, 16 vs 18). When
replaced by a C-18 estratriene nucleus, there is no doubt that
the 4-methyl-4-aza-5R-androstane nucleus is an important
feature that must be kept as the main structure for all future
inhibitor design (Figure 5).
Taken together, these results directed our SAR study
toward 17β-[(N-alkyl)arylamido]-4-methyl-4-aza-5R-androstanes.
A first library containing 27 different analogue compounds
divided into three groups according to the number of carbon
atoms in their aliphatic side chains was prepared and is
reported in Table 1. In group A (n = 6), all compounds
bearing at least one para-substituent showed an inhibition
value of g50% when tested at 0.3 μM. The best compound in
this series was 34 with an inhibitory potency of 84%, which
was not as good as the potency of lead compound 2 (94%)
but better than lead compound 3 with 63% inhibition at the
same concentration.12 Group B compounds (n = 9) presents
a Reagents and conditions: (a) (i) NH2(CH2)nCH3, p-TSA, toluene, 4
˚
A molecular sieves, Schlenk tube, 18-24 h; (ii) NaBH3CN, MeOH, pH
5, 0 °C, 1-2 h (30-54%, two steps); (b) (i) HCOOH, DCC, CHCl3, 0 °C,
10 min; (ii) steroid, pyridine, K2CO3, 0 °C to rt (30%); (c) (CH2O)n,
NaBH3CN, refluxing MeOH, 24 h (64%).
(compounds 12-18) and the importance of the 4-methyl-4-
aza-5R-androstane nucleus over an estrane nucleus
(compounds 70 and 71). According to Figure 4, arylamides
12-14 are better inhibitors than alkylamides 15-18. How-

Article
Journal of Medicinal Chemistry, 2009, Vol. 52, No. 23 7493
Figure 6. Inhibitory potency (%) reached by tertiary amines (E1 to
E2 by 17β-HSD7). Products were tested in duplicate at 0.3 μM. See
Scheme 3 for the structures of tested compounds.
a similar profile as group A compounds (n = 6), with para-
substituted benzamides being the best inhibitors when com-
pared to other ring arrangements. Compound 35, the para
fluoro derivative, is the best inhibitor of this series with 71%
inhibition at 0.3 μM. The last group (C, n = 11) has
compounds that present generally lower inhibitory potency
than compounds of groups A and B. Consequently, another
library was then elaborated choosing only acyl chlorides
bearing either a fluoro atom or a para-substituent (or both)
and selecting only the heptylamine (n = 6) or decylamine
(n = 9) side chain on the C-17 position of the steroid nucleus
(Table 2). The maximum inhibitory potency reached by the
second library members did not exceed that of the first one
except for compounds 46 and 51. Indeed, they showed 84%
and 80% inhibition, respectively, when compared to com-
pound 34 (77%) introduced as reference in the same experi-
ment.
Focusing on N-heptyl derivatives, we evaluated the im-
portance of the presence of a carbonyl group in this area of
the molecule by synthesizing compounds 73, 74, 75, and 76,
the tertiary amine analogues of amides 17, 12, 15, and 2.
Results seem to demonstrate that the amine version is a
better inhibitor than the amide one. Indeed, tertiary amines
73, 74, and 75 reached 76%, 69%, and 45% inhibition,
respectively, at 0.3 μM, whereas their amide analogues 17,
12, and 15 reached 14%, 25%, and 0% at 0.3 μM (Figure 6).
However, amine 76 (89% inhibition) showed similar inhibi-
tory potency compared to its analogue 2, with an average of
94% inhibition (data not shown) when tested at the same
concentration. We will obviously need to test both tertiary
amines and amides at the same concentration in a single
assay to truly confirm this trend. It seems, however, that the
inhibitory potency of tertiary amines drops in accordance
with the elongation of the alkyl side chain (76 vs 73 vs 72) and
the bulkiness of the substituent (72 vs 75). Consequently,
compound 76 appears to be the best inhibitor in this series of
amine derivatives.
Figure 7. Selectivity of 17β-HSD7 inhibitors toward other steroi-
dogenic enzymes. Inhibitors were assessed for their inhibitory
potency (%) toward 17β-HSD1 (E1 to E2), 17β-HSD5 (Δ4-dione
to T), 5R-R1 (Δ4-dione to A-dione), and 5R-R2 (Δ4-dione to
A-dione) in comparison to their potency on 17β-HSD7 (E1 to E2).
Inhibitors were tested at 0.3 μM except for 17β-HSD1, where the
concentration was 0.1 μM.
good inhibitors of 5R-Rs15 and spiro-δ-lactone, an essential
pharmacophore used to block the conversion of Δ4-dione
into T by 17β-HSD5,13 it was necessary to determine the
selectivity of all compounds toward 5R-R1, 5R-R2, and 17β-
HSD5. We also tested the selected inhibitors for their ability
to block the conversion of E1 into E2 by 17β-HSD1, a
cytosolic enzyme having a Km value for E1 3.6 times lower
than 17β-HSD7.20
As expected, both compounds 1a (R = Me) and 1b (R =
H) were able to inhibit 17β-HSD5 more than 90% at 0.3 μM,
but all the other compounds had no effect on this enzyme.
Selected inhibitors were not able to efficiently block 17β-
HSD1, suggesting a very different molecular structure of the
enzymatic site between 17β-HSD1 and 17β-HSD7 and this,
even if they share a common substrate (E1). It was very
interesting to see that despite all 4-methyl-4-aza-5R-andros-
tane derivatives being inhibitors of 5R-R1, compound 1b (R
= H), i.e., the only 4-aza-5R-androstane compound, did not
inhibit the enzyme when tested at 0.3 μM. The same result
was also obtained with 5R-R2, which was neither blocked by
the 4-aza-5R-androstane nucleus nor by tertiary amines 73
and 76. We concluded that the only way to obtain highly
selective inhibitors of 17β-HSD7 was to avoid the inhibition
of 5R-Rs by synthesizing the 4-aza-5R-androstane counter-
part of our two best amide-type and amine-type inhibitors.
As anticipated, 4-aza steroids 80 and 81 were dramatically
less inhibiting 5R-Rs than their 4-methyl-4-aza analogues 11
and 76, respectively (Figure 8). However, changing the 4-
methyl-4-aza-5R-androstane nucleus for a 4-azaandrostane
nucleus impaired their inhibitory potency on 17β-HSD7.
This phenomenon was, however, less marked for the
17β-[(N-decyl)formamido] derivative 80, whose IC50 passed
from 195 ( 18 nM to 230 ( 15 nM, than for the 17β-[(N-
heptyl)methylamino] derivative 81 with IC50 passing from
189 ( 18 nM to 458 ( 38 nM (Figure 9).
Inhibitor Selectivity toward 17β-HSD7. To verify if the
new compounds selectively inhibit 17β-HSD7, we chose the
two most potent inhibitors in each category of compounds
discussed above. The spiro-δ-lactones 1a and 1b, the for-
mamides 2 and 11, the arylamides 34 and 46, and the tertiary
amines 73 and 76 were tested as inhibitors of other steroido-
genic enzymes (Figure 7). Because 17β-[(N-alkyl)forma-
mido]-4-methyl-4-aza-5R-androstanes were proven to be
Inhibition toward 17β-HSD7 (Conversion of DHT into 3β-
Diol). Selected compounds were assayed for their ability to

7494 Journal of Medicinal Chemistry, 2009, Vol. 52, No. 23
Bellavance et al.
Figure 10. Inhibition (%) of the DHT into 3β-diol conversion by
17β-HSD7. Products were tested in duplicate at the indicated
concentrations.
inhibitors. According to this hypothesis, they could bind the
enzyme at another area than the enzymatic site and induce
structural modifications that could affect two different
catalytic sites at the same time. More experimentation will
be necessary to clarify this hypothesis.
Figure 8. Inhibitory potency (%) of 4-aza-5R-androstane counter-
parts of two of our best 4-methyl-4-aza-5R-androstane inhibitors
toward the 5R-R1 and 5R-R2 conversion of Δ4-dione into A-dione.
Products were tested in duplicate at 0.3 μM. MK-906 (finasteride) is
a known inhibitor of 5R-R2.21
Conclusion
Despite the fact that the exact role of 17β-HSD7 in the
human organism is not well understood, it is believed that it
could be implicated in the growth, or at least in the main-
tenance, of estrogen-dependent tumors. Greater exposure of
an ERþ tumor to estrogens will result in its proliferation and
increase of its aggressiveness.22 Moreover, a recent study
demonstrated that when prostatic LNCaP cancer cells pass
from an androgen-dependent to an androgen-independent
state, the production of estrogens increases significantly,
especially via the augmentation of E1 into E2 conversion
and DHT into 3β-diol conversion. RT-PCR analysis further
demonstrated that 17β-HSD7 expression was 1.7 times great-
er in this situation and that it was responsible for this estro-
gen’s synthesis pattern.23 In our paper, development of the
first selective 17β-HSD7 inhibitors was achieved through a
typical SAR study, in which a preliminary screening had
highlighted that the 4-methyl-4-aza-5R-androstan-3-one nu-
cleus had a potential lead compound. Further investigation,
aiming at the improvement of the 17β-substituent, allowed us
to identify 17β-formamido, 17β-benzamido, and 17β-amino
derivatives that efficiently inhibit the conversion of E1 into E2
and of DHT into 3β-diol by 17β-HSD7. According to their
IC50 values, the best inhibitors for E1 into E2 conversion are
the spiro-δ-lactone 1a (116
( 9 nM), the 17β-(N-
decylformamido) derivative 11 (195 ( 18 nM), and the
tertiary amine 76 (189 ( 18 nM). When their selectivity over
17β-HSD7 was tested, it appears that compound 1a was
inhibiting 17β-HSD5 but not compounds 11 and 76, whereas
the 4-methyl-4-aza-5R-androstane nucleus was always able to
block the enzymatic activity of 5R-R1, irrespective of the
natureofthe 17β-substituent. Inhibitors1aand11alsoproved
themselves as good inhibitors of the 5R-R2 activity but not the
tertiary amine 76. A full selectivity for 17β-HSD7 was then
achieved by synthesizing compounds 80 and 81, the 4-aza-5R-
androstane analogues of compounds 11 and 76, respectively.
As anticipated, compounds 80 and 81 are no more able to
inhibit 5R-Rs (Δ4-dione into A-dione) and, as they do not bear
17β-spiro-δ-lactone, did not inhibit 17β-HSD5. Compound
80 still presents a good inhibitory potency toward 17β-HSD7
for both enzymatic reactions (E1 conversion, IC50 = 230 (
15 nM; DHT conversion, inhibition of 40% at 0.3 μM and
Figure 9. Determination of IC50 values for the transformation of
E1 to E2 by 17β-HSD7. Inhibitors were tested in triplicate at
concentrations ranging from 0.1 nM to 20 μM.
inhibit the second reaction catalyzed by 17β-HSD7, i.e., the
transformation of DHT into 3β-diol (Figure 10). At 3 μM, all
compounds inhibited the androgen deactivation in the range
69-92%. The best inhibitors of DHT transformation are
compound 1a, which still inhibits more than 50% of the
enzyme activity at 0.3 μM, and compound 11 with 49%
inhibition at the same concentration. The fact that inhibitors
designed for the E1 to E2 transformation also inhibit the
DHT to 3β-diol transformation may suggest that the cata-
lytic site for DHT and E1 is the same. However, since the
mechanism of inhibition of these compounds is not yet
understood, it is also possible that they act as noncompetitive

Article
Journal of Medicinal Chemistry, 2009, Vol. 52, No. 23 7495
86% at 3 μM) when compared to compound 11. However,
compound 81, although selective, is slightly less potent (E1
conversion, IC50 = 458( 38 nM; DHT conversion, inhibition
of 29% at 0.3 μM and 74% at 3 μM) than its analogue,
compound 76.
a Celite 521 pad to remove the precipitate and molecular sieves.
The precipitate was washed with acetone and the combined
filtrate was evaporated to give a dark-brown foam, which was
further purified by flash column chromatography (CH2Cl2/
MeOH, 97:3 to 90:10, and then acetone/hexanes, 30:70) to give
compound 5 (6.0 g, 68%). IR (film on NaCl): 1738 (CdO,
ketone), 1642 (CdO, lactam). 1H NMR 400 MHz (CDCl3): 0.87
(s, 18-CH3), 0.90 (s, 19-CH3), 0.80-2.20 (residual CH and CH2),
2.45 (m, 2-CH2 and 1H of 16-CH2), 2.93 (s, 4-NCH3), 3.05 (dd,
J = 3.5 and 12.6 Hz, 5R-CH). 13C NMR, 100 MHz (CDCl3):
12.39, 13.82, 20.38, 21.65, 25.15, 29.01, 29.05, 29.16, 31.36,
32.88, 33.93, 35.75, 36.53, 47.76, 50.93, 52.01, 65.62, 170.61,
220.37. LRMS: calcd for C19H30NO2 [M þ H]þ 304.2, found
304.4.
17β-(N-Alkylamino)-4-methyl-4-aza-5r-androstan-3-ones
(6-9). First Method (Synthesis of 6 and 9). A round flask filled
with dry toluene (25 mL), compound 5 (1.0 g, 3.29 mmol), n-
heptylamine (4.94 mmol, 1.5 equiv), and p-TSA (38 mg) was
equipped with a Dean-Stark trap and a condenser. The mixture
was refluxed 18-24 h, and the formation of the corresponding
imine was followed by TLC. The reaction mixture was cooled
and the solvent removed under reduced pressure. The resulting
residue was dissolved in dry methanol (11 mL) and cooled to
0 °C. The pH was lowed around pH 5 with concentrated HCl,
and a NaBH3CN solution (6.59 mmol, 2 equiv, in 4 mL of
methanol) was added dropwise. The mixture was stirred at 0 °C
for 1.5-2 h and the solvent evaporated in vacuo. The residue
was dissolved in water, and the solution was basified to pH 13
with 2 N KOH and extracted three times with CH2Cl2. The
organic phase was washed with brine, dried (MgSO4), filtered,
and concentrated under vacuum. The crude oil was purified by
flash column chromatography (CH2Cl2/MeOH, 97:3, and
CH2Cl2/acetone, 85:15).
Compounds 80 and 81 represent the only known selective
inhibitors of 17β-HSD7 able to efficiently block both enzy-
matic reactions performed by this steroidogenic enzyme. This
discovery will no doubt contribute to broaden our knowledge
about the role of 17β-HSD7 in humans. Moreover, these
compounds did not show androgenic or estrogenic effects on
androgen-sensitive Shionogi and estrogen-sensitive CAMA-1
cells (data not reported), which means that our inhibitors do
not generate mitogenic effects via the transactivation of the
androgenic or estrogenic receptor. This major characteristic
could ultimately open the door to therapeutic use of these
kinds of compounds in the treatment of hormone-sensitive
diseases.
Experimental Section
General Methods for Chemical Synthesis. Chemical reagents
were purchased from Sigma-Aldrich Canada Ltd. (Oakville,
Ontario, Canada) except aminomethylpolystyrene (1% DVB,
=100-200 mesh, 1.05 mequiv/g), which was obtained from
ꢀ
ꢀ
Richelieu Biotechnologies (Montreal, Quebec, Canada), and
piperidinomethyl polystyrene (3.5 mmol/g), obtained from Ma-
ꢀ
trix Innovation (Montreal, Quebec, Canada). Testosterone was
purchased from Steraloids (Wilton, NH). HPLC and ACS
ꢀ
ꢀ
solvent grades were from Fisher Scientific (Quebec, Canada),
and anhydrous solvents were from VWR (Ville Mont-Royal,
ꢀ
Quebec, Canada). Pyridine and THF for anhydrous reactions
were distilled prior to use, and reactions were run under inert
(argon) atmosphere in oven-dried glassware. Thin-layer chro-
17β-(N-Heptylamino)-4-methyl-4-aza-5r-androstan-3-one
(6). 774 mg, 58% yield. IR (film on NaCl): 3250 (NH, weak),
1649 (CdO, lactam). 1H NMR 400 MHz (CDCl3): 0.70 (s, 18-
CH3), 0.87 (t, J = 7.0 Hz, 70-CH3), 0.88 (s, 19-CH3), 0.75-2.10
(residual CH and CH2), 2.44 (dd, J = 4.7 and 9.5 Hz, 2-CH2),
2.53 (t, J = 8.7 Hz, 17R-CH), 2.59 (m, 10-CH2), 2.92 (s, 4-
NCH3), 3.02 (dd, J = 3.5 and 12.6 Hz, 5R-CH). 13C NMR
100 Hz (CDCl3): 11.92, 12.41, 14.08, 20.90, 22.61, 23.61, 25.33,
27.38, 29.08, 29.10, 29.24, 29.78, 30.00, 30.62, 31.82, 32.95,
34.37, 36.44, 38.05, 42.93, 48.88, 52.16, 52.88, 65.78, 69.06,
170.76. LRMS: calcd for C26H47N2O [M þ H]þ 403.4, found
403.4.
17β-(N-Dodecylamino)-4-methyl-4-aza-5r-androstan-3-one
(9). 762 mg, 44% yield. IR (film on NaCl): 3400 (NH, weak),
1650 (CdO, lactam). 1H NMR 300 MHz (CDCl3): 0.73 (s, 18-
CH3), 0.87 (t, J = 6.9 Hz, 120-CH3), 0.88 (s, 19-CH3), 0.75-2.10
(residual CH and CH2), 2.44 (dd, J = 4.7 and 9.5 Hz, 2-CH2),
2.60 (m, 17R-CH and 10-CH2), 2.92 (s, 4-NCH3), 3.02 (dd, J =
3.4 and 12.5 Hz, 5R-CH). 13C NMR 75 MHz (CDCl3): 11.95,
12.33, 14.07, 20.72, 22.61, 23.47, 25.20, 27.26, 29.00, 29.06,
29.28, 29.41, 29.54 (6x), 29.85, 31.83, 32.80, 34.16, 36.31,
37.72, 42.81, 48.69, 51.92, 52.69, 65.64, 68.60, 170.65. LRMS:
calcd for C31H57N2O [M þ H]þ 473.4, found 473.5.
˚
matography (TLC) was performed with (60 A) silica gel plates
(Fisher Scientific), and 4-methyl-4-aza-steroids were visualized
using Dragen-Dorf reagent, while other types of molecules
were visualized using cerium ammonium molybdate. Flash
column chromatography was performed with 40-63 μm
˚
ꢀ
(60 A) Silicycle silica gel (Quebec, Canada). The steroid deriva-
tive libraries were realized with an ACT LabTech manual
synthesizer (Advance ChemTech, Louiseville, KY). Infrared
spectra (IR) were recorded on a Perkin-Elmer 1600 (series
FTIR) spectrophotometer (Norwalk, CT), and only the most
significant bands are reported in cm-1. Nuclear magnetic re-
sonance (NMR) spectra were recorded at 300 MHz (1H) and
75.5 MHz (13C) on a Bruker AC/F300 spectrometer (Billerica,
MA) or at 400 MHz (1H) and 100.6 MHz (13C) on a Bruker
Avance 400 digital spectrometer, and the chemical shifts are
reported in ppm. The CHCl3 1H and 13C NMR signals (7.26 and
77.00 ppm, respectively, for 300 or 400 MHz apparatus) were
used as internal references. Low-resolution mass spectra
(LRMS), ESI or APCI, were recorded on a Perkin-Elmer Sciex
API-150 ex apparatus (Foster City, CA) equipped with a turbo
ion-spray source. High-performance liquid chromatography
(HPLC) analyses were carried out for new synthesized and
tested compounds 10-18, 34, 46, 70-76, 80, and 81 using a
Waters system (Milford, MA) equipped with a UV detector and
a reverse-phase column (Luna Phenyl/Hexyl or Ace 3 C18-HL)
and using an appropriate gradient of solvents (methanol and
water with or without 0.1% of HCOOH). The purity was found
to be g95% except when otherwise specified at the end of the
description of its synthesis.
Second Method (Synthesis of 7 and 8). A round flask filled with
dry toluene (25 mL), compound 5 (500 mg, 1.65 mmol), n-
decylamine (1.92 mmol, 1.2 equiv), and p-TSA (20 mg) was
equipped with a Dean-Stark trap and a condenser. The mixture
was refluxed 18-24 h, and the formation of the corresponding
imine was followed by TLC. The reaction mixture was cooled
and the solvent removed under reduced pressure. The resulting
residue was dissolved in dry methanol (10 mL) and the solution
cooled to 0 °C. A NaBH4 solution (2.47 mmol, 1.5 equiv, in 2 mL
of methanol) was added dropwise. The mixture was stirred at
0 °C for 1.5 h and the solvent evaporated in vacuo. The residue
was dissolved in water, and the solution was basified to pH 13
with 2 N KOH and extracted three times with CH2Cl2. The
organic phase was washed with brine, dried (MgSO4), filtered,
4-Methyl-4-aza-5r-androstane-3,17-dione (5). To a stirred
solution of compound 415 (500 mg, 1.64 mmol) in dry CH2Cl2
(18 mL) was added pyridinium chlorochromate (PCC)
(4.1 mmol, 2.5 equiv), NaOAc (4.92 mmol, 3.0 equiv), and
˚
activated molecular sieves (4 A) (10% w/w of the alcohol).
The mixture was stirred at rt for 3 h and then filtered through

7496 Journal of Medicinal Chemistry, 2009, Vol. 52, No. 23
Bellavance et al.
and concentrated under vacuum. The crude oil was purified by
flash column chromatography (CH2Cl2/MeOH, 97:3 to 95:5,
and CH2Cl2/MeOH/Et2NH, 94:5:1).
17β-[(N-Alkyl)arylamido]-4-methyl-4-aza-5r-androstan-3-
ones (12-15). 17β-[(N-Heptyl)benzamido]-4-methyl-4-aza-5r-
androstan-3-one (12). To a solution of compound 6 (60 mg,
0.149 mmol) in dry THF (2.40 mL) was added anhydrous
powdered potassium carbonate (0.448 mmol, 3 equiv), and the
mixture was stirred at rt for 0.5 h. Addition of benzoyl chloride
(0.448 mmol, 3 equiv) dropwise was followed by slow agitation
for 3 h at rt under argon atmosphere. Solvent was removed
under reduced pressure, and the resulting residue was treated
with saturated NaHCO3 and then extracted with three portions
of CH2Cl2. The combined organic phase was washed with brine,
dried over MgSO4, and concentrated in vacuo. The crude
compound was purified by preparative TLC (1000 μm)
(acetone/hexanes, 35:65), and the desired product was recovered
from the silica gel by washing with two portions of acetone and
one portion of CH2Cl2 over medium core filter. The combined
filtrated was concentrated in vacuo to give compound 12 (56 mg,
17β-(N-Nonylamino)-4-methyl-4-aza-5r-androstan-3-one (7).
530 mg, 75% yield. IR (film on NaCl): 3310 (NH, weak), 1648
(CdO, lactam). 1H NMR 400 MHz (CDCl3): 0.72 (s, 18-CH3),
0.87 (t, J = 7.0 Hz, 90-CH3), 0.88 (s, 19-CH3), 0.75-2.10
(residual CH and CH2), 2.44 (dd, J = 4.5 and 9.6 Hz, 2-CH2),
2.55 (t, J = 8.7 Hz, 17R-CH), 2.60 (m, 10-CH2), 2.91 (s, 4-
NCH3), 3.01 (dd, J = 3.5 and 12.6 Hz, 5R-CH). 13C NMR 75
MHz (CDCl3): 12.07, 12.41, 14.11, 20.51, 20.81, 22.66, 23.55,
25.29, 26.81, 27.34, 29.11, 29.27, 29.50, 29.55, 29.93, 31.85,
32.91, 33.97, 34.26, 36.40, 37.82, 42.90, 48.75, 52.02, 52.79,
65.73, 68.65, 170.73. LRMS: calcd for C28H51N2O [M þ H]þ
431.4, found 431.4.
17β-(N-Decylamino)-4-methyl-4-aza-5r-androstan-3-one (8).
454 mg, 62% yield. IR (film on NaCl): 3307 (NH, weak), 1650
(CdO, lactam). 1H NMR 300 MHz (CDCl3): 0.69 (s, 18-CH3),
0.85 (t, J = 7.1 Hz, 100-CH3), 0.86 (s, 19-CH3), 0.75-2.10
(residual CH and CH2), 2.42 (dd, J = 4.7 and 9.5 Hz, 2-CH2),
2.53 (m, 17R-CH and 10-CH2), 2.90 (s, 4-NCH3), 3.01 (dd, J =
3.5 and 12.5 Hz, 5R-CH). 13C NMR 75 MHz (CDCl3): 11.86,
12.34, 14.07, 20.80, 23.53, 25.24, 27.35, 29.05, 29.26, 29.50 (4ꢀ)
29.64, 29.90, 30.47, 31.83, 32.83, 34.26, 36.34, 37.95, 42.84,
48.98, 52.03, 52.77, 65.71, 68.94, 170.69. LRMS: calcd for
C29H53N2O [M þ H]þ 445.7, found 445.5.
1
74%). IR (film on NaCl): 1633 (CdO, amide and lactam). H
NMR 100 MHz (CDCl3): 0.78 (s, 18-CH3), 0.87 (broad s, 19-
CH3 and 70-CH3), 0.80-2.10 (residual CH and CH2), 2.43 (dd,
J = 4.5 and 9.2 Hz, 2-CH2), 2.92 (s, 4-NCH3), 3.01 (m, 5R-CH
and 1 H of 10-CH2), 3.40 (broad, 0.8 H of 17R-CH), 3.90 (broad,
1 H of 10-CH2), 4.70 (broad, 0.2 H of 17R-CH), 7.36 (m, 5 H of
ArH). 13C NMR 300 MHz (CDCl3): 12.33, 13.05, 14.04, 20.42,
22.53, 23.08, 23.94, 25.01, 26.83, 26.91, 28.22, 28.36, 28.74,
29.63, 31.63, 32.27, 33.85, 36.36, 36.80, 45.63 (2ꢀ), 51.18,
51.70, 66.07, 126.62, 128.31 (3ꢀ) 128.87, 138.17, 171.4, 173.4.
LRMS: calcd for C33H51N2O2 [M þ H]þ 507.4, found 507.5.
17β-[(N-Heptyl)-400-nitrobenzamido]-4-methyl-4-aza-5r-an-
drostan-3-one (13). Compound 13 (23 mg, 67% yield) was
prepared from 6 (25 mg, 0.062 mmol) as for compound 12 and
was purified by a flash column chromatography (acetone/
hexanes, 35:65) prior to a final preparative TLC using the same
eluent. IR (film on NaCl): 1633, (CdO, amide and lactam). 1H
NMR 400 MHz (CDCl3): 0.79 (s, 18-CH3), 0.87 (s broad, 19-
CH3 and 70-CH3), 0.30-2.10 (residual CH and CH2), 2.47 (m, 2-
CH2), 2.92 (s, 4-NCH3), 3.10, 3.23, 3.60, 3.88, and 4.65 (5m, 5R-
CH, 10-CH2 and 17R-CH), 7.50 (d, J = 6.7 Hz, 2 H of ArH), 8.27
(d, J = 8.2 Hz, 2 H of ArH). 13C NMR 75 MHz (CDCl3): 12.32,
13.09, 14.00, 20.30, 22.49, 24.01, 25.12, 27.04, 28.82, 29.24,
29.59, 31.62, 32.61, 33.73, 36.34, 36.77, 44.42, 45.19, 46.11,
47.03, 51.26, 51.56, 62.74, 65.61, 68.31, 123.36, 123.77, 127.85,
130.89, 136.62, 144.29, 147.77, 150.28, 166.39, 171.14. LRMS:
calcd for C33H50N3O4 [M þ H]þ 552.4, found 552.5.
17β-[(N-Alkyl)formamido]-4-methyl-4-aza-5r-androstan-3-
ones (10 and 11). 17β-[(N-Nonyl)formamido]-4-methyl-4-aza-
5r-androstan-3-one (10). To a 0.2 M solution of formic acid
(0.58 mmol, 2 equiv) in chloroform (1 mL) was added dropwise a
solution of dicyclohexylcarbodiimide (DCC) (0.58, 2 equiv) in
chloroform (2 mL). The mixture was stirred under argon at 0 °C
for 5 min and then slowly added to an ice-cold solution of
compound 7 (125 mg, 0.29 mmol) in pyridine (2.5 mL). After
0.5 h, the temperature was allowed to rise to rt and the mixture
stirred for 1 h. Evaporation of the solvent in vacuo followed by
addition of Et2O precipitated dicyclohexylurea which was re-
moved by filtration. The combined filtrate was concentrated
under reduced pressure and the resulting oil purified by flash
column chromatography (acetone/hexanes, 15:85 to 25:75) to
give 10 as a white solid (98 mg, 73% yield). IR (film on NaCl):
1
1668 (CdO, amide), 1652 (CdO, lactam). H NMR 400 MHz
(CDCl3): 0.71 (s, 18-CH3), 0.88 (t, J = 7.0 Hz, 90-CH3), 0.90
(s, 19-CH3), 0.80-2.10 (residual CH and CH2), 2.48 (dd, J =
4.2 and 9.2 Hz, 2-CH2), 2.94 (s, 4-NCH3), 3.05 (dd, J = 3.5 and
12.5 Hz, 5R-CH), 3.28 (m, 2.8 H of 10-CH2 and 17R-CH), 4.16 (t,
J = 9.7 Hz, 0.2 H of 10-CH2), 8.18 (s, 0.8 H of HCON), 8.24
(s, 0.2 H of HCON). 13C NMR 75 MHz (CDCl3): 12.38,
12.80, 14.10, 20.54, 22.63, 22.89, 23.12, 23.27, 24.34, 25.18,
26.73, 27.05, 28.57, 28.98, 29.17, 29.35, 29.50, 29.69, 31.80,
32.55, 32.83, 34.09, 36.38, 36.78, 37.25, 44.25, 44.35, 45.71,
46.74, 51.26, 51.68, 51.95, 61.92, 65.65, 68.57, 163.04, 164.56,
170.66. LRMS: calcd for C29H51N2O2 [M þ H]þ 459.4, found
459.5.
17β-[(N-Heptyl)-400-methoxybenzamido]-4-methyl-4-aza-5r-
androstan-3-one (14). Compound 14 (44 mg, 66% yield) was
prepared from 6 (50 mg, 0.124 mmol) as for compound 12 and
was purified by a flash column chromatography (acetone/
hexanes, 35:65) prior to a preparative TLC using the same
eluent. IR (film on NaCl): 1629 (CdO, amide and lactam). 1H
NMR 400 MHz (CDCl3): 0.76 (s, 18-CH3), 0.85 (t, J = 7.3 Hz,
70-CH3), 0.86 (s, 19-CH3), 0.65-2.10 (residual CH and CH2),
2.46 (dd, J = 4.3 and 8.7 Hz, 2-CH2), 2.92 (s, 4-NCH3), 3.00 (m,
5R-CH and 10-CH2), 3.68 (broad, 17R-CH), 3.83 (s, CH3O), 6.88
(d, J = 8.6 Hz, 2 H of ArH), 7.28 (d, J = 8.5 Hz, 2 H of ArH).
13C NMR 75 MHz (CDCl3): 12.32, 13.03, 14.03, 20.38, 22.52,
23.03, 23.93, 25.20, 26.86, 28.87, 28.94, 29.15, 29.65, 31.64,
32.72, 33.83, 36.36, 36.72, 45.49, 51.20, 51.70, 55.25, 65.67,
113.53 (2ꢀ), 128.40 (2ꢀ), 130.47, 159.91, 170.82, 173.12. LRMS:
calcd for C34H53N2O3 [M þ H]þ 537.4, found 537.5.
17β-[(N-Decyl)formamido]-4-methyl-4-aza-5r-androstan-3-
one (11). Compound 11 (73 mg, 37% yield) was prepared from 8
(185 mg, 0.416 mmol) as described above for 10. IR (film on
NaCl): 1670 (CdO, amide), 1647 (CdO, lactam). 1H NMR 400
MHz (CDCl3): 0.71 (s, 18-CH3), 0.88 (t, J = 7.0 Hz, 100-CH3),
0.90 (s, 19-CH3), 0.80-2.10 (residual CH and CH2), 2.50 (dd,
J = 4.5 and 9.7 Hz, 2-CH2), 2.94 (s, 4-NCH3), 3.05 (dd, J = 3.5
and 12.5 Hz, 5R-CH), 3.28 (m, 2.8 H of 10-CH2 and 17R-CH),
4.16 (t, J = 9.7 Hz, 0.2 H of 17R-CH), 8.18 (s, 0.8 H of HCON),
8.24 (s, 0.2 H of HCON). 13C NMR 75 MHz (CDCl3): 12.37,
12.80, 14.10, 20.54, 22.64, 22.89, 23.12, 23.27, 24.34, 25.18,
26.73, 27.04, 28.57, 29.00, 29.13, 29.25, 29.34, 29.51, 29.69,
31.84, 32.55, 32.85, 34.10, 36.38, 36.80, 37.25, 44.25, 44.36,
45.71, 46.74, 51.27, 51.68, 51.96, 61.94, 65.64, 68.57, 163.04,
164.56, 170.63. LRMS: calcd for C30H53N2O2 [M þ H]þ 473.4,
found 473.5.
17β-[(N-Heptyl)cyclohexylamido]-4-methyl-4-aza-5r-andro-
stan-3-one (15). Compound 15 (21 mg, 41% yield) was prepared
from 6 (40 mg, 0.099 mmol) as for compound 12 and was
purified by flash column chromatography (acetone/hexanes,
20:80). IR (film on NaCl): 1641 and 1632, (CdO, amide and
1
lactame). H NMR 400 MHz (CDCl3): 0.65 and 0.72 (2s, 18-
CH3), 0.86 (s, 19-CH3), 0.90 (t, J = 6.8 Hz, 70-CH3), 0.70-2.10
(residual CH and CH2), 2.45 (m, 2-CH2 and CHCO), 2.93 (s,
4-NCH3), 3.05 (m, 5R-CH), 2.80, 3.08, 3.31, 3.75, and 4.54
(4m and 1t, 10-CH2 and 17R-CH). 13C NMR 75 MHz (CDCl3):

Article
Journal of Medicinal Chemistry, 2009, Vol. 52, No. 23 7497
12.36, 12.74, 12.92, 14.07, 20.55, 22.54, 22.99, 23.37, 23.53,
24.80, 25.24, 25.63, 25.80, 26.42, 26.83, 28.63, 28.98, 29.16,
29.75, 30.56, 30.94, 31.34, 31.72, 31.86, 32.12, 32.77, 34.03,
36.75, 37.37, 40.82, 42.07, 44.56, 45.27, 45.71, 51.13, 51.65,
51.91, 53.42, 61.77, 65.52, 65.65, 170.89, 177.19, 177.82. LRMS:
calcd for C33H57N2O2 [M þ H]þ 513.4, found 513.6.
550 rpm for 3 h under a continuous argon flow. The disappear-
ance of the starting material was confirmed by TLC and excess
acyl chloride quenched by aminomethyl polystyrene resin
(0.112 mmol, 3 equiv) under agitation for another 2.5 h. Reac-
tion mixture was filtered and collected in individual vials. The
solid support was washed with THF (2ꢀ) and CH2Cl2 (1ꢀ). The
combined filtrates were evaporated to dryness and quickly
purified over a small pad of silica gel (EtOAc 100% to
EtOAc/acetone, 50:50). TLC and LRMS analyses were done
for each member of libraries 1 and 2. A sampling of representa-
tive compounds was also analyzed by IR and 1H NMR.
First Library (Compounds 19-45). The 27 members of library
1 were obtained as described above except for compounds
25-27 and 31-33. Indeed, reactions with 4-methoxy-3-trifluor-
omethylbenzoyl chloride needed an additional 0.5 equiv of acyl
chloride and another hour to be completed, whereas reactions
with 2,6-dichlorobenzoyl chloride needed an additional 5 equiv
and an overnight reaction time to be completed. Filtration over
a silica gel pad afforded amides 19-45 in yields ranging between
62% and 99%. Compound purity was found acceptable for
enzymatic assay by TLC analysis, and LRMS of all library
members gave satisfactory results. Representative compounds
20, 23, 26, 30, 33, 36, 37, 40, and 43 were also analyzed by IR and
1H NMR. After compound 34 was identified as the best 17β-
HSD7 inhibitor of this library, it was selected for additional
studies, purified again, and fully characterized by IR, 1H NMR,
and 13C NMR) and the purity assessed by HPLC.
17β-[(N-Alkyl)alkylamido]-4-methyl-4-aza-5r-androstan-3-
ones (16-18). Compounds 16-18 were synthesized from 6 using
the same method as described for compounds 10 and 11 except
for the following details: (1) carboxylic acid and DCC equiva-
lents were doubled (now 4 equiv); (2) carboxylic acid was added
to DCC dissolved in chloroform; (3) the mixture was stirred for
10 min and then added to the mixture containing the steroid 6,
pyridine, and potassium carbonate (0.108 mmol, 1.5 equiv); (4)
the mixture was stirred overnight.
17β-[(N-Heptyl)acetamido]-4-methyl-4-aza-5r-androstan-3-
one (16). Compound 16 (44 mg, 80% yield) was prepared from 6
(50 mg, 0.124 mmol) and purified by flash column chromatog-
raphy (acetone/hexanes, 25:75). IR (film on NaCl): 1645 (CdO,
amide and lactam). 1H NMR 100 MHz (CDCl3): 0.67 and 0.74
(2s, 18-CH3), 0.86 (s, 19-CH3), 0.88 (t, J = 6.3 Hz, 70-CH3),
0.70-2.10 (residual CH and CH2), 2.12 (s, 0.9 H of 200-CH3),
2.14 (s, 2.1 H of 200-CH3), 2.43 (dd, J = 4.6 and 9.4 Hz, 2-CH2),
2.92 (s, 4-NCH3), 3.03 (dd, J = 3.3 and 12.4 Hz, 5R-CH), 2.80,
3.12, 3.28, 3.68, and 4.50 (3m and 2t, 10-CH2 and 17R-CH). 13C
NMR 300 MHz (CDCl3): 12.31, 12.80, 13.01, 14.05, 20.54,
22.38, 22.53, 22.75, 22.93, 23.25, 23.72, 24.65, 25.21, 26.93,
27.70, 28.93, 29.03, 29.19, 29.74, 31.02, 31.77, 32.71, 32.79,
33.92, 34.02, 37.11, 37.48, 44.75, 45.63, 46.73, 51.07, 51.65,
51.90, 62.32, 65.71, 67.38, 170.94, 171.38, 171.72. LRMS: calcd
for C28H49N2O2 [M þ H]þ 445.4, found 445.4.
17β-[(N-Decyl)-400-bromobenzamido]-4-methyl-4-aza-5r-an-
drostan-3-one (20). IR (film on NaCl): 1634 (CdO, amide and
lactam). 1H NMR 300 MHz (CDCl3): 0.76 (s, 18-CH3), 0.86 (s,
19-CH3), 0.87 (t, J = 6.5 Hz, 100-CH3), 0.70-2.10 (residual CH
and CH2), 2.42 (dd, J = 4.8 and 9.4 Hz, 2-CH2), 2.91 (s, 4-
NCH3), 3.02 (m, 5R-CH and 10-CH2), 3.80 (broad, 17R-CH),
7.20 (d, J = 8.1 Hz, 300-CH and 500-CH), 7.51 (d, J = 8.2 Hz, 40þ0-
CH and 600-CH). LRMS: calcd for C36H56BrN2O2 [M þ H]
628.4, found 628.3.
17β-[(N-Heptyl)propionamido]-4-methyl-4-aza-5r-androstan-
3-one (17). Compound 17 (39 mg, 69% yield) was prepared from
6 (50 mg, 0.124 mmol) and purified by flash column chroma-
tography (acetone/hexanes, 20:80). IR (film on NaCl): 1644
1
(CdO, amide and lactam). H NMR 400 MHz (CDCl3): 0.66
and 0.73 (2s, 18-CH3), 0.70-2.10 (residual CH and CH2), 0.86
(s, 19-CH3), 0.88 (sapp, 70-CH3), 1.15 (t, J = 7.3 Hz, 300-CH3),
2.38 (m, 2-CH2 and 200-CH2), 2.92 (s, 4-NCH3), 3.02 (dd, J = 3.4
and 12.6 Hz, 5R-CH), 2.80, 3.08, 3.28, 3.70, and 4.51 (4m and 1t,
10-CH2 and 17R-CH). 13C NMR 75 MHz (CDCl3): 9.79, 10.31,
12.35, 12.76, 13.04, 14.04, 20.53, 22.52, 22.91, 23.27, 23.64,
24.64, 25.23, 26.97, 27.12, 27.65, 28.88, 28.96, 29.14, 29.74,
31.28, 31.76, 32.73, 34.03, 36.34, 37.30, 44.82, 45.39, 45.66,
45.77, 51.11, 51.70, 51.93, 62.36, 65.67, 66.01, 170.83, 174.56,
175.09. LRMS: calcd for C29H51N2O2 [M þ H]þ 459.4, found
459.4. HPLC purity: 87%.
17β-[(N-Decyl)-200-chlorobenzamido]-4-methyl-aza-5r-andro-
stan-3-one (23). IR (film on NaCl): 1644 (CdO, lactam), 1634
(CdO, amide). 1H NMR 400 MHz (CDCl3): 076 and 0.82 (2s,
18-CH3), 0.87 and 0.88 (2t, J = 7.2 Hz, 100-CH3), 0.85 and 0.89
(2s, 19-CH3), 0.30-2.20 (residual CH and CH2), 2.46 (m, 2-
CH2), 2.90 and 2.94 (2s, 4-NCH3), 3.06 (m), 3.42 (t, J = 9.5 Hz),
3.88 (m) and 4.66 (t, J = 10.0 Hz) (5R-CH, 10-CH2 and 17R-
CH), 7.30 (m, 4 H of Ar-H). LRMS: calcd for C36H56ClN2O2
[M þ H] þ 584.4, found 584.3.
17β-[(N-Decyl)-300-trifluoromethyl-400-methoxybenzamido]-4-
methyl-4-aza-5r-androstan-3-one (26). IR (film on NaCl): 1640
(CdO, lactam and amide). 1H NMR 300 MHz (CDCl3): 0.77 (s,
18-CH3), 0.87 (s and t, J = 6.6 Hz, 19-CH3 and 100-CH3),
0.60-2.05 (residual CH and CH2), 2.40 (dd, J = 4.9 and 9.5 Hz,
2-CH2), 2.91 (s, 4-NCH3), 2.98 (m, 5R-CH and 10-CH2), 3.65
(broad, 17R-CH), 3.94 (s, CH3O), 7.00 (d, J = 8.5 Hz, 500-CH),
7.51 (d, J = 8.5 Hz, 600-CH), 7.54 (s, 200-CH). LRMS: calcd for
C38H58 F3N2O3 [M þ H]þ 647.4, found 647.4.
17β-[(N-Heptyl)hexanamido]-4-methyl-4-aza-5r-androstan-
3-one (18). Compound 18 (26 mg, 51% yield) was prepared from
6 (40 mg, 0.099 mmol) and purified by flash column chroma-
tography (acetone/hexanes, 20:80). IR (film on NaCl): 1642
1
(CdO, amide and lactam). H NMR 400 MHz (CDCl3): 0.67
and 0.74 (2s, 18-CH3), 0.70-2.10 (residual CH and CH2), 0.87
(s, 19-CH3), 0.89 and 0.90 (2s broad, 70-CH3 and 600-CH3), 2.28
and 2.43 (2m, 2-CH2 and 200-CH2), 2.93 (s, 4-NCH3), 3.04 (dd,
J = 3.1 and 12.2 Hz, 5R-CH), 2.84, 3.13, 3.33, 3.73, and 4.52 (4m
and 1t, 10-CH2 and 17R-CH). 13C NMR 75 MHz (CDCl3):
12.37, 12.83, 13.06, 14.03, 20.56, 22.55, 22.97, 23.31, 23.66,
25.13, 25.90, 26.94, 27.28, 28.62, 29.05, 29.50, 29.73, 31.37,
31.75, 32.46, 32.63, 33.94, 34.02, 36.36, 37.29, 44.91, 45.70,
45.94, 51.10, 51.69, 51.91, 62.36, 65.90, 66.20, 172.00, 174.50,
174.76. LRMS: calcd for C32H57N2O2 [M þ H]þ 501.4, found
501.5.
Parallel Synthesis of Compounds 19-67 (Libraries 1 and 2).
Piperidinomethylpolystyrene resin (0.186 mmol, 5 equiv) and
compound 6, 8, or 9 (0.037 mmol), previously dissolved in
anhydrous THF (1 mL), were added into a 96-well solid-phase
reaction block (ACT LabTech). Acyl chlorides were introduced
(0.112 mmol, 3 equiv), and the reaction block was sealed and
purged with argon. The reaction block was allowed to shake at
17β-[(N-Dodecyl)-400-chlorobenzamido]-4-methyl-4-aza-5r-
androstan-3-one (30). IR (film on NaCl): 1642 (CdO, lactam),
1
1634 (CdO, amide). H NMR 300 MHz (CDCl3): 0.76 (s, 18-
CH3), 0.86 (s, 19-CH3), 0.87 (t, J = 6.4 Hz, 120-CH3), 0.65-2.10
(residual CH and CH2), 2.42 (dd, J = 4.5 and 9.2 Hz, 2-CH2),
2.95 (s, 4-NCH3), 3.02 (m, 5R-CH and 10-CH2), 3.80 (broad,
17R-CH), 7.29 (d, J = 6.7 Hz, 300-CH and 500-CH), 7.38 (d, J =
8.3 Hz, 200-CH and 600-CH). LRMS: calcd for C38H60ClN2O2
[M þ H]þ 611.4, found 611.5.
17β-[(N-Dodecyl)-200,600-dichlorobenzamido]-4-methyl-4-aza-
5r-androstan-3-one (33). IR (film on NaCl): 1644 (CdO, lactam
and amide). 1H NMR 400 MHz (CDCl3): 0.81 and 0.83 (2s, 18-
CH3), 0.85 and 0.89 (2s, 19-CH3), 0.88 (t, J = 6.9 Hz, 120-CH3),
0.40-2.10 (residual CH and CH2), 2.58 (m, 2-CH2), 2.93 and
2.97 (2s, 4-NCH3), 3.08 (m), 3.34 (m), 3.73 (t, J = 8.8 Hz) and
4.60 (t, J = 10.0 Hz) (5R-CH, 10-CH2 and 17R-CH), 7.20-7.40

7498 Journal of Medicinal Chemistry, 2009, Vol. 52, No. 23
Bellavance et al.
(m, Ar-H). LRMS: calcd for C38H59Cl2N2O2 [M þ H]þ 645.4,
found 645.4.
20.31, 20.63, 22.40, 22.59, 22.79, 23.46, 23.63, 24.31, 25.14,
25.22, 26.47, 27.24, 28.26, 28.29, 28.77, 28.83, 29.00, 29.25,
29.31, 39.58, 29.79, 30.78, 31.38, 31.81, 32.65, 33.71, 34.08,
35.59, 36.35, 36.40, 37.12, 44.72, 44.94, 46.34, 47.04, 51.13,
51.46, 51.63, 51.88, 110.98, 111.21, 111.37, 111.74, 111.99,
112.22, 115.84, 127.78, 130.37, 130.46, 130.56, 130.65, 156.64,
160.28, 162.40, 163.12, 171.16, 171.33. LRMS: calcd for
C33H49F2N2O2 [M þ H]þ 553.4, found 543.3.
17β-[(N-Heptyl)-400-fluorobenzamido]-4-methyl-4-aza-5r-an-
drostan-3-one (34). IR (film on NaCl): 1634 (CdO, lactam and
1
amide). H NMR 400 MHz (CDCl3): 0.76 (s, 18-CH3), 0.86 (s
broad, 19-CH3 and 70-CH3), 0.60-2.10 (residual CH and CH2),
2.44 (dd, J = 4.4 and 9.4 Hz, 2-CH2), 2.91 (s, 4-NCH3), 3.03 (m,
5R-CH, 10-CH2), 3.75 (broad, 17R-CH), 7.07 (t, J = 8.6 Hz, 300-
CH and 500-CH), 7.32 (dd, J = 5.5 and 8.8 Hz, 200-CH and 600-
CH. 13C NMR 100 MHz (CDCl3): 12.37, 13.06, 14.00, 20.46,
22.51, 23.07, 23.98, 25.24, 26.85, 28.76, 28.93, 29.20, 29.69,
31.63, 32.78, 33.93, 36.45, 36.92, 45.58, 51.32, 51.82, 65.71,
115.28, 115.50, 128.74, 134.30, 161.60, 164.07, 170.91, 172.28.
LRMS: calcd for C33H50FN2O2 [M þ H]þ 525.4, found 525.3.
17β-[(N-Dodecyl)-400-fluorobenzamido]-4-methyl-4-aza-5r-
androstan-3-one (36). IR (film on NaCl): 1636 (CdO, lactam and
17β-[(N-Decyl)-200,600-difluorobenzamido]-4-methyl-4-aza-5r-
androstan-3-one (47). IR (film on NaCl): 1645 (CdO, lactam and
amide). 1H NMR 400 MHz (CDCl3): 0.77 and 0.78 (2s, 18-CH3),
0.86 and 0.87 (2t, J = 7.0 Hz, 100-CH3), 0.88 and 0.89 (2s, 19-
CH3), 0.40-2.10 (residual CH and CH2), 2.44 (m, 2-CH2), 2.90
and 2.93 (2s, 4-NCH3), 3.10 (m), 3.47 (t, J = 9.5 Hz), 3.90 (m)
and 4.67 (t, J = 9.9 Hz) (4 signals of 5R-CH, 10-CH2 and 17R-
CH), 6.93 (m, 2 H of Ar-H), 7.32 (m, 1 H of Ar-H). LRMS:
calcd for C36H55F2N2O2 [M þ H]þ 585.4, found 585.4.
17β-[(N-Decyl)-200,500-difluorobenzamido]-4-methyl-4-aza-5r-
androstan-3-one (49). IR (film on NaCl): 1646 (CdO, lactam),
1636 (CdO, amide). 1H NMR 400 MHz (CDCl3): 0.76 and 0.78
(2s, 18-CH3), 0.86 and 0.87 (2t, J = 6.9 Hz, 100-CH3), 0.88 and
0.89 (2s, 19-CH3), 0.40-2.10 (residual CH and CH2), 2.46 (m,
2-CH2), 2.91 and 2.94 (2s, 4-NCH3), 3.10, 3.48, 3.88, and 4.63
(4m, 5R-CH, 10-CH2 and 17R-CH), 7.06 (m, 300-CH and 400-CH),
7.13 and 7.68 (2m, 600-CH). LRMS: calcd for C36H55F2N2O2
[M þ H]þ 585.4, found 585.5.
1
amide). H NMR 300 MHz (CDCl3): 0.76 (s, 18-CH3), 0.87 (s
broad, 19-CH3 and 120-CH3), 0.60-2.10 (residual CH and
CH2), 2.42 (dd, J = 4.4 and 9.3 Hz, 2-CH2), 2.91 (s, 4-NCH3),
3.00 (m, 5R-CH and 10-CH2), 3.80 (broad, 17R-CH), 7.07 (t, J =
8.5 Hz, 300-CH and 500-CH), 7.32 (dd, J = 5.6 and 8.1 Hz, 200-CH
and 600-CH). LRMS: calcd for C38H60FN2O2 [M þ H]þ 595.5,
found 595.5.
17β-[(N-Heptyl)-300-chloromethylbenzamido]-4-methyl-4-aza-
5r-androstan-3-one (37). IR (film on NaCl): 1634 (CdO, lactam
and amide). 1H NMR 300 MHz (CDCl3): 0.78 (s, 18-CH3), 0.86
(s broad, 19-CH3 and 70-CH3), 0.60-2.10 (residual CH and
CH2), 2.42 (dd, J = 4.5 and 9.6 Hz, 2-CH2), 2.91 (s, 4-NCH3),
3.05 (m, 5R-CH and 10-CH2), 3.80 (broad, 17R-CH), 4.60 (s,
CH2Cl), 7.28 (m, 500-CH), 7.37 (m, 200-CH, 400-CH and 600-CH).
LRMS: calcd for C34H52ClN2O2 [M þ H]þ 555.4, found 555.3.
17β-[(N-Heptyl)-400-chloromethylbenzamido]-4-methyl-4 aza-
5r-androstan-3-one (40). IR (film on NaCl): 1634 (CdO, lactam
and amide). 1H NMR 300 MHz (CDCl3): 0.77 (s, 18-CH3), 0.86
(s broad, 19-CH3 and 70-CH3), 0.60-2.10 (residual CH and
CH2), 2.42 (dd, J = 4.7 and 9.2 Hz, 2-CH2), 2.90 (s, 4-NCH3),
3.02 (m, 5R-CH and 10-CH2), 3.80 (broad, 17R-CH), 4.59 (s,
CH2Cl), 7.31 (d, J = 8.1 Hz, 300-CH and 500-CH), 7.39 (d, J = 8.1
Hz, 200-CH and 600-CH). LRMS: calcd for C34H52ClN2O2 [M þ
H]þ 555.4, found 555.3.
17β-[(N-Heptyl)-400-cyanobenzamido]-4-methyl-4-aza-5r-an-
drostan-3-one (43). IR (film on NaCl): 1633 (CdO, lactam and
amide). 1H NMR 300 MHz (CDCl3): 0.35 and 0.70 (2m, 9-CH),
0.77 (s, 18-CH3), 0.86 (s broad, 19-CH3 and 70-CH3), 0.80-2.10
(residual CH and CH2), 2.42 (m, 2-CH2), 2.91 (s, 4-NCH3), 3.07
(m, 5R-CH and 1H of 10-CH2), 3.24, 3.60, 3.87, and 4.60 (4m, 1H
of 10-CH2 and 17R-CH), 7.43 (d, J = 7.6 Hz, 200-CH and 600-
CH), 7.69 (d, J = 8.1 Hz, 300-CH and 500-CH). LRMS: calcd for
C34H50N3O2 [M þ H]þ 532.4, found 532.3.
17β-[(N-Decyl)-200,400-difluorobenzamido]-4-methyl-4-aza-5r-
androstan-3-one (51). IR (film on NaCl): 1636 (CdO, lactam and
amide). 1H NMR 400 MHz (CDCl3): 0.75 and 0.78 (2s, 18-CH3),
0.85 and 0.87 (2t, J = 7.1 Hz, 100-CH3), 0.87 and 0.89 (2s, 19-
CH3), 0.30-2.10 (residual CH and CH2), 2.44 (m, 2-CH2), 2.90
and 2.94 (2s, 4-NCH3), 3.10 (m), 3.49 (t, J = 9.5 Hz), 3.90 (m),
4.68 (m) (4 signals of 5R-CH, 10-CH2 and 17R-CH), 6.85 (m, 1H
of Ar-H), 6.94 (t, J = 7.3 Hz, 1 H of Ar-H), 7.32 and 7.40 (2m,
600-CH). LRMS: calcd for C36H55F2N2O2 [M þ H]þ 585.4,
found 585.4.
17β-[(N-Decyl)-300,500-difluorobenzamido]-4-methyl-4-aza-5r-
androstan-3-one (53). IR (film on NaCl): 1642 (CdO, lactam)
and 1632 (CdO, amide). 1H NMR 400 MHz (CDCl3): 0.75 (s,
18-CH3), 0.87 (s, 19-CH3), 0.87 (t, J = 6.9 Hz, 120-CH3),
0.70-2.10 (residual CH and CH2), 2.43 (dd, J = 4.6 and
9.5 Hz, 2-CH2), 2.92 (s, 4-NCH3), 3.03 (m, 5R-CH and 10-
CH2), 3.78 (broad, 17R-CH), 6.85 (m, Ar-H). LRMS: calcd
for C36H55F2N2O2 [M þ H]þ 585.4, found 585.4.
17β-[(N-Decyl)pentafluorobenzamido]-4-methyl-4-aza-5r-an-
drostan-3-one (55). IR (film on NaCl): 1652 (CdO, lactam and
amide). 1H NMR 400 MHz (CDCl3): 0.77 and 0.79 (2s, 18-CH3),
0.86 and 0.88 (2t, J = 6.9 Hz, 100-CH3), 0.87 and 0.89 (2s, 19-
CH3), 0.40-2.10 (residual CH and CH2), 2.44 (dd, J = 5.0 and
9.8 Hz, 2-CH2), 2.91 and 2.93 (2s, 4-NCH3), 3.10 (m), 3.35 (t, J
= 9.6 Hz), 3.88 (m) and 4.60 (t, J = 9.9 Hz) (4 signals of 5R-CH,
10-CH2 and 17R-CH). LRMS: calcd for C36H52F5N2O2 [M þ
H]þ 639.4, found 639.4.
Second Library (Compounds 46-67). The 22 members of
library 2 were obtained as described above. Filtration over a
silica gel pad afforded amides 46-67 in yields ranging from 69%
to 94%. Except compounds 56, 57, 60, and 61, which were not
tested, LRMS gave satisfactory results, and TLC analyses
confirmed that purities of compounds were acceptable for
enzymatic assay. Representative compounds 47, 49, 51, 53, 55,
58, 62, 64, and 66 were also analyzed by IR and 1H NMR. After
compound 46 was identified as the best 17β-HSD7 inhibitor of
this library, it was selected for additional studies, purified again,
and fully characterized by IR, 1H NMR, and 13C NMR and the
purity assessed by HPLC.
17β-[(N-Heptyl)-200,600-difluorobenzamido]-4-methyl-4-aza-5r-
androstan-3-one (46). IR (film on NaCl): 1644 (CdO, lactam and
amide). 1H NMR 400 MHz (CDCl3): 0.78 and 0.79 (2s, 18-CH3),
0.82 and 0.89 (2t, J = 7.2 Hz, 70-CH3), 0.86 and 0.89 (2s, 19-
CH3), 0.30-2.10 (residual CH and CH2), 2.48 (m, 2-CH2), 2.91
and 2.95 (2s, 4-NCH3), 3.10 (m), 3.47 (t, J = 9.5 Hz), 3.92 (m),
4.69 (t, J = 9.9 Hz) (4 signals of 5R-CH, 10-CH2 and 17R-CH),
6.95 (m, 2 H of Ar-H), 7.34 (m, 1 H of Ar-H). 13C NMR
100 MHz (CDCl3): 12.31, 12.36, 12.73, 13.07, 13.05, 14.07,
17β-[(N-Heptyl)-200-fluoro-600-trifluoromethylbenzamido]-4-
methyl-4-aza-5r-androstan-3-one (58). IR (film on NaCl): 1646
1
(CdO, lactam and amide). H NMR 400 MHz (CDCl3): 0.79
and 0.81 (2s, 18-CH3), 0.80 and 0.81 (2t, J = 7.0 Hz, 70-CH3),
0.85 and 0.89 (2s, 19-CH3), 0.40-2.10 (residual CH and CH2),
2.44 (m, 2-CH2), 2.89 and 2.93 (2s, 4-NCH3), 3.10 (m), 3.25
(t, J = 9.5 Hz), 3.76 (m), 4.53 (t, J = 10.0 Hz) and 4.61 (t, J =
10.0 Hz) (5 signals of 5R-CH, 10-CH2 and 17R-CH), 7.32 (m, 1 H
of Ar-H), 7.50 (m, 2 H of Ar-H). LRMS: calcd for
C34H49F4N2O2 [M þ H]þ 593.4, found 593.3.
17β-[(N-Heptyl)-400-tert-butylbenzamido]-4-aza-methyl-5r-
androstan-3-one (62). IR (film on NaCl): 1636 (CdO, lactam and
amide). 1H NMR 400 MHz (CDCl3): 0.78 (s, 18-CH3), 0.84 (t,
J = 7.2 Hz, 70-CH3), 0.87 (s, 19-CH3), 1.32 (s, (CH3)3C),
0.70-2.1 (residual CH and CH2), 2.48 (m, 2-CH2), 2.93 (s, 4-
NCH3), 3.02 (m, 5R-CH and 10-CH2), 3.78 (broad, 17R-CH),
7.25 (d, J = 6.5 Hz, 300-CH and 500-CH), 7.38 (d, J = 8.3 Hz,

Article
Journal of Medicinal Chemistry, 2009, Vol. 52, No. 23 7499
200-CH and 600-CH). LRMS: calcd for C37H59N2O2 [M þ H]þ
563.5, found 563.4.
on NaCl): 3220 (OH, phenol), 1643 (CdO, amide), 1580 and
1485 (CdC, aryl). 1H NMR 400 MHz (CDCl3): 0.74 (s, 18-
CH3), 0.88 (t, J = 6.8 Hz, 70-CH3), 1.20-2.40 (residual CH and
CH2), 2.83 (m, 6-CH2), 3.33 (m, 10-CH2 and 17R-CH), 6.57 (d,
J = 2.7 Hz, 3-CH), 6.63 (dd, J = 2.7 and 8.5 Hz, 2-CH), 7.13 (d,
J = 8.5 Hz, 1-CH), 8.22 (s, 0.85 H of HCON), 8.27 (s, 0.15 H of
HCON). 13C NMR 75 MHz, ((CD3)2CO): 12.85, 13.24, 14.39,
23.29, 23.52, 23.79, 23.91, 25.01, 27.01, 27.39, 27.72, 28.13,
30.27, 32.60, 33.41, 37.21, 37.76, 38.64, 39.65, 43.03, 44.60,
44.86, 45.20, 46.63, 47.27, 51.74, 51.84, 63.13, 69.13, 113.54,
115.88, 127.07, 131.77, 138.37, 155.91, 163.37, 164.79. LRMS:
calcd for C26H40NO2 [M þ H]þ 398.3, found 398.3. HPLC
purity: 83%.
17β-[(N-Heptyl)-400-heptylbenzamido]-4-aza-methyl-5r-andro-
stan-3-one (64). IR (film on NaCl): 1636 (CdO, lactam and
amide). 1H NMR 400 MHz (CDCl3): 0.77 (s, 18-CH3), 0.87 (s,
19-CH3), 0.88 (t, J = 6.8 Hz, 70-CH3 and CH3CH2), 0.70-2.10
(residual CH and CH2), 2.40 (m, 2-CH2), 2.61 (t, J = 7.6 Hz,
CH2CH2Ar), 2.91 (s, 4-NCH3), 3.03 (m, 5R-CH and 10-CH2),
3.78 (broad, 17R-CH), 7.16 (d, J = 8.0 Hz, 300-CH and 500-CH),
7.23 (d, J = 8.2 Hz, 200-CH and 600-CH). LRMS: calcd for
C40H65N2O2 [M þ H]þ 605.5, found 605.5.
17β-[(N-Heptyl)-400-iodobenzamido]-4-methyl-4-aza-5r-an-
drostan-3-one (66). IR (film on NaCl): 1636 (CdO, lactam and
1
amide). H NMR 400 MHz (CDCl3): 0.77 (s, 18-CH3), 0.87 (s
17β-[(N-Heptyl)alkyl/arylamino]-4-methyl-4-aza-5r-andro-
stan-3-ones (72-75). 17β-[(N-Heptyl)pentylamino]-4-methyl-4-
aza-5r-androstan-3-one (72). The N-heptylamino derivative 6
(24 mg, 0.060 mmol) was dissolved in hot anhydrous acetonitrile
(2 mL) and transferred into a Schlenk tube containing potas-
sium carbonate (0.270 mmol, 4.5 equiv). Pentyl iodide
(0.240 mmol, 4 equiv) was added to the mixture, and the tube
was purged with argon and closed. Caution: Reactions per-
formed in closed tubes should be carried out with suitable
precautions such as the use of safety shields and gloves. The
mixture was heated to 165 °C over 24 h. After that, it was
allowed to cool and the solvent removed under reduced pres-
sure. The residue was dissolved in water (15 mL) and extracted
with three portions of EtOAc (15 mL). The combined organic
phase was washed with brine, dried over MgSO4, and concen-
trated under vacuum. The crude residue was purified by flash
column chromatography (acetone/hexanes, 15:85) to afford
compound 72 (21.6 mg, 76% yield). IR (film on NaCl): 1650
(CdO, lactam). 1H NMR 400 MHz (CdCl3): 0.76 (s, 18-CH3),
0.87 (s, 19-CH3), 0.88 (t, J = 6.8 Hz, 70-CH3), 0.89 (t, J = 6.6 Hz,
500-CH3), 0.70-2.10 (residual CH and CH2), 2.44 (m, 17R-H and
2-CH2), 2.48 (t, J = 7.8 Hz, 10-CH2 and 100-CH2), 2.92 (s, 4-
NCH3), 3.01 (dd, J = 3.5 and 12.5 Hz, 5R-CH). 13C NMR
100 MHz (CDCl3): 12.21, 12.38, 14.10, 14.16, 21.17, 22.65,
22.77, 23.14, 25.36, 25.46, 25.75, 26.73, 27.61, 29.12, 29.40,
29.70, 29.87, 29.98, 31.92, 32.92, 34.33, 36.36, 39.31, 43.63,
51.54 (2x), 51.10, 52.05, 65.79, 71.84, 170.78. LRMS: calcd for
C31H57N2O [M þ H]þ 473.5, found 473.4. HPLC purity: 90%.
17β-[(N-Heptyl)propylamino]-4-methyl-4-aza-5r-androstan-3-
one (73). Compound 73 (23 mg, 83% yield) was obtained from 6
(26 mg, 0.0636 mmol) as reported above for 72. IR (film on
NaCl): 1651 (CdO, lactam). 1H NMR 400 MHz (CDCl3): 0.76
(s, 18-CH3), 0.83 (t, J = 7.3 Hz, 300-CH3), 0.87 (s, 19-CH3), 0.88
(t, J = 6.8 Hz, 70-CH3), 0.70-2.10 (residual CH and CH2), 2.48
(m, 17R-CH, 2-CH2, 10-CH2 and 100-CH2), 2.92 (s, 4-NCH3),
3.01 (dd, J = 3.7 and 12.5 Hz, 5R-CH). 13C NMR 100 MHz
(CDCl3): 11.86, 12.17, 12.33, 14.10, 18.77, 21.20, 22.62, 23.07,
25.30, 25.64, 26.67, 27.57, 29.08, 29.37, 29.68, 29.91, 31.89,
32.83, 34.24, 36.28, 39.20, 43.56, 51.53, 51.80, 52.89, 53.45,
65.73, 71.77, 170.72. LRMS: calcd for C29H53N2O [M þ H]þ
445.4, found 445.4. HPLC purity: 91%.
broad, 19-CH3 and 70-CH3), 0.60-2.10 (residual CH and CH2),
2.43 (dd, J = 4.4 and 9.3 Hz, 2-CH2), 2.91 (s, 4-NCH3), 3.03 (m,
5R-CH and 10-CH2), 3.78 (broad, 17R-CH), 7.07 (d,
J = 7.9 Hz, 300-CH and 500-CH), 7.72 (d, J = 8.2 Hzþ, 200-CH
and 600-CH). LRMS: calcd of C33H50IN2O2 [M þ H] 633.3,
found 633.3.
17β-(N-Heptylamino) and (N-Heptylformamido) Derivatives
of Estrane Nucleus (Compounds 69-71). 17β-(N-Heptylamino)-
3-methoxyestra-1,3,5(10)-triene (69). This compound was
synthesized from 68 (2.0 g, 7.03 mmol) as described for the
synthesis of compounds 6 and 9 except for the following details:
(1) 5 equiv (35.16 mmol) of n-heptylamine was added, and (2) the
aqueous phase was basified to pH 13 using 4 N NaOH. The
crude oil was purified by flash column chromatography
(acetone/hexanes, 3:97 to 5:95) to give compound 69 (522 mg,
64% yield). IR (film on NaCl): 3700 (weak), 1609 and 1500
(CdC, aryl). 1H NMR 400 MHz (CDCl3): 0.75 (s, 18-CH3), 0.90
(t, J = 6.8 Hz, 70-CH3), 1.20-2.40 (residual CH and CH2), 2.65
(m, 10-CH2 and 17R-CH), 2.85 (m, 6-CH2), 3.78 (s, CH3O), 6.64
(d, J = 2.6 Hz, 4-CH), 6.72 (dd, J = 2.7 and 8.5 Hz, 2-CH), 7.21
(d, J = 8.6 Hz, 1-CH). 13C NMR 75 MHz (CDCl3): 11.83, 14.10,
22.60, 23.46, 26.46, 26.73, 27.39, 29.22, 29.50, 29.82, 30.24,
31.80, 38.15, 38.74, 43.04, 43.94, 48.96, 52.30, 55.16, 69.07,
111.40, 113.71, 126.28, 132.77, 137.97, 157.35. LRMS: calcd
for C26H42NO [M þ H]þ 384.4, found 384.3.
17β-[(N-Heptyl)formamido]-3-methoxyestra-1,3,5(10)-triene
(70). Compound 70 was prepared following the same method as
described for compounds 16-18, starting from compound 69
(50 mg, 0.130 mmol) using DCC (0.391 mmol, 3 equiv) and
formic acid (0.391 mmol, 3 equiv). The resulting crude product
was purified by flash column chromatography (EtOAc/hexanes,
5:95) to give a solid (45 mg, 84% yield). IR (film on NaCl): 1670
(CdO, amide), 1500 and 1615 (CdC, aryl). 1H NMR 300 MHz
(CDCl3): 0.74 (s, 18-CH3), 0.88 (t, J = 6.5 Hz, 70-CH3),
1.20-2.40 (residual CH and CH2), 2.85 (m, 6-CH2), 3.35 (m,
10-CH2 and 17R-CH), 3.77 (s, CH3O), 6.63 (d, J = 2.7 Hz, 4-
CH), 6.71 (dd, J = 2.7 and 8.6 Hz, 2-CH), 7.19 (d, J = 8.6 Hz, 1-
CH), 8.23 (s, 0.8 H of HCON), 8.28 (s, 0.2 H of HCON). 13C
NMR 75 MHz ((CD3)2CO): 12.80, 13.19, 14.34, 23.26, 23.48,
23.74, 23.86, 24.96, 26.03, 27.36, 27.69, 28.06, 29.49, 32.57,
33.36, 37.70, 38.60, 39.54, 44.57, 45.16, 47.23, 51.70, 51.79,
55.27, ∼63, 69.07, 112.24, 114.41, 129.04, 132.91, 138.41,
158.55, 163.36, 164.77. LRMS: calcd for C27H42NO2 [M þ
H]þ 412.3, found 412.3.
17β-[(N-Heptyl)formamido]-3-hydroxyestra-1,3,5(10)-triene
(71). Compound 69 (29 mg, 0.071 mmol) was dissolved in dry
CH2Cl2 (9 mL) under an argon atmosphere. The solution was
stirred, and BBr3 (0.177 mmol, 2.5 equiv) was slowly added at
0 °C and then allowed to warm up at rt. After 4 h, HCl (10%)
was added to stop the reaction and the aqueous phase was
neutralized with saturated NaHCO3. The unprotected steroid
was extracted by three portions of CH2Cl2, which were col-
lected, washed with brine, filtered over cotton wool, and eva-
porated under reduced pressure. The crude compound was
purified by flash column chromatography (acetone/hexanes,
10:90) to give 71 as a white solid (12 mg, 47% yield). IR (film
17β-[(N-Heptyl)benzylamino]-4-methyl-4-aza-5r-androstan-3-
one (74). Compound 74 (28 mg, 89% yield) was prepared from 6
(26 mg, 0.0643 mmol) as reported above for 72. IR (film on
NaCl): 1651 (CdO, lactam), 1500 (CdC, aryl). 1H NMR
400 MHz (CDCl3): 0.83 (s, 18-CH3), 0.85 (t, J = 7.0 Hz, 70-
CH3), 0.88 (s, 19-CH3), 0.70-2.10 (residual CH and CH2), 2.45
(m, 2-CH2 and 10-CH2), 2.61 (t, J = 9.1 Hz, 17R-CH), 2.92 (s, 4-
NCH3), 3.01 (dd, J = 3.4 and 12.5 Hz, 5R-CH), 3.68 (s, 100-CH2),
7.21 (t, J = 7.2 Hz, 1 H of Ar-H), 7.29 (t, J = 7.2 Hz, 2 H of
Ar-H), 7.36 (d, J = 7.0 Hz, 2 H of Ar-H). 13C NMR 75 MHz
(CDCl3): 12.33, 12.42, 14.06, 21.01, 22.56, 23.13, 24.68, 25.27,
25.65, 27.32, 29.06, 29.25, 29.87, 31.80, 32.81, 34.17, 36.28,
39.05, 40.38, 43.83, 51.84, 51.95, 52.64, 56.16, 65.71, 71.54,
126.30, 127.95 (2ꢀ), 128.23 (2ꢀ), 141.46, 170.72. LRMS: calcd
for C33H53N2O [M þ H]þ 493.4, found 493.4. HPLC purity:
93%.

7500 Journal of Medicinal Chemistry, 2009, Vol. 52, No. 23
Bellavance et al.
17β-[(N-Heptyl)cyclohexylmethylamino]-4-methyl-4-aza-5r-
androstan-3-one (75). Compound 75 (4.2 mg, 20% yield) was
prepared from 6 (17 mg, 0.0422 mmol) as reported above for
compound 72, but 20 equiv (0.844 mmol) of cyclohexylmethyl
bromide and 5 equiv (0.211 mmol) of potassium carbonate were
used instead. IR (film on NaCl): 1649 (CdO, lactam). 1H NMR
400 MHz (CDCl3): 0.72 (s, 18-CH3), 0.87 (s, 19-CH3), 0.88 (t,
J = 6.9 Hz, 70-CH3) 0.75-2.10 (residual CH and CH2), 2.23 (m,
100-CH2), 2.43 (2-CH2, and 10-CH2), 2.52 (m, 17R-CH), 2.92 (s,
4-NCH3), 3.01 (dd, J = 3.4 and 12.6 Hz, 5R-CH). 13C NMR 75
MHz (CDCl3): 12.41, 12.87, 14.05, 20.37, 20.98, 21.65, 22.52,
24.74, 25.14, 25.38, 25.76, 26.88, 29.01, 29.12, 29.69, 30.96,
31.33, 31.60, 32.08, 32.88, 33.93, 35.78, 36.34, 40.26, 43.36,
47.78, 50.89, 51.38, 53.61, 58.00, 65.50, 73.57, 170.56. LRMS:
calcd for C33H59N2O [M þ H]þ 499.4, found 499.5.
and amide). 1H NMR 400 MHz (CDCl3): 0.71 (s, 18-CH3), 0.87
(t, J = 7.0 Hz, 100-CH3), 0.91 (s, 19-CH3), 0.75-2.05 (residual
CH and CH2), 2.41 (m, 2-CH2), 3.08 (d, J = 8.7 Hz, 5R-CH),
3.28 (m, 2.80 H of 10-CH2 and 17R-CH), 4.15 (t, J = 9.7 Hz, 0.2
H of 17R-CH), 5.65 (broad, NH), 8.18 (s, 0.8 H of HCON), 8.23
(s, 0.2 of HCON). 13C NMR, 100 MHz (CDCl3): 11.37, 12.40,
14.08, 20.66, 22.65, 23.00, 24.38, 27.08, 27.35, 28.62 (2x), 29.11,
29.27, 29.36, 29.51, 29.56, 31.86, 33.43, 34.88, 35.88, 36.86,
44.37, 44.43, 51.37, 51.76, 60.69, 68.66, 162.94, 164.84. LRMS:
calcd for C29H51N2O2 [M þ H]þ 459.4, found 459.4.
17β-[(N-Heptyl)methylamino]-4-aza-5r-androstan-3-one (81).
Compound 81 (30 mg, 64% yield) was prepared from 79 (45 mg,
0.116 mmol) according to the method used for compound 76
and purified by flash column chromatography (MeOH/CH2Cl2,
1:99 to 4:96). IR (film on NaCl): 3410 and 3210 (NH), 1652
(CdO, lactam). 1H NMR 400 MHz (CDCl3): 0.80 (s, 18-CH3),
0.88 (t, J = 6.8 Hz, 70-CH3), 0.89 (s, 19-CH3), 0.75-2.15
(residual CH and CH2), 2.22 (s, 100-CH3), 2.40 (dd, J = 4.0
and 10.0 Hz, 2-CH2), 2.30-2.60 (broad, 10-CH2 and 17R-CH),
3.04 (dd, J = 4.7 and 11.4 Hz, 5R-CH), 5.36 (s, NH). 13C NMR
100 MHz (CDCl3): 11.32, 12.62, 14.02, 20.93, 22.51, 22.82,
26.92, 27.21, 28.54, 28.92, 28.99, 31.60, 33.30, 34.65, 35.66,
39.10, 43.39, 50.67, 53.31, 60.54, 74.18, 171.82. LRMS: calcd
for [M þ H]þ C26H47N2O 403.4, found 403.4.
17β-[(N-Heptyl)methylamino]-4-methyl-4-aza-5r-androstan-
3-one (76). Compound 6 (67 mg, 0.167 mmol) was dissolved in
anhydrous methanol (1.5 mL), and paraformaldehyde (0.184
mmol, 1.1 equiv) and sodium cyanoborohydride (0.200 mmol,
1.2 equiv) were added. The mixture was refluxed for 24 h, and
0.5 equiv of both reactants was added. The mixture was refluxed
for an additional 6 h. The mixture was then allowed to cool,
brine was added, and steroid was extracted with EtOAc (3ꢀ).
The combined organic phase was washed with brine, dried over
MgSO4, and filtered, and the solvent was evaporated under
reduced pressure. The crude product was purified by flash
column chromatography (CH2Cl2, 100% to MeOH/CH2Cl2,
5:95) to afford compound 76 (35 mg, 50%). IR (film on NaCl):
1649, (CdO, lactam). 1H NMR 300 MHz (CDCl3): 0.88 (s
broad, 18-CH3, 19-CH3 and 70-CH3), 0.70-2.10 (residual CH
and CH2), 2.25 (s broad, 100-CH3), 2.15-2.70 (broad, 10-CH2
and 17R-CH), 2.45 (dd, J = 4.6 and 9.6 Hz, 2-CH2), 2.94 (s, 4-
NCH3), 3.03 (dd, J = 3.4 and 12.5 Hz, 5R-CH). 13C NMR
75 MHz (CDCl3): 12.22, 12.36, 14.10, 21.10, 22.62, 23.13, 25.29,
27.51, 29.11, 29.28, 29.90, 31.82, 32.87, 34.20, 36.28, 39.59,
41.26, 43.37, 51.58, 53.35, 56.46, 65.70, 74.65, 170.72. LRMS:
calcd for C27H49N2O [M þ H]þ 417.4, found 417.3.
General Methods for Biological Assay. 17β-HSD7 Inhibition
Assays. Human embryonic kidney (HEK)-293 cells (American
Type Culture Collection, Rockville, MD) were previously
transfected with 17β-HSD7 DNA as reported by Lui et al.5e
Essentially, pCMV-neo-h17β-HSD7 plasmids were transfected
using the lipofectin transfection kit and grown under G-418
until resistant colonies were observed. HEK-293 cells overex-
pressing 17β-HSD7 were plated at 500 000 cells/well in a 6-well
plate at 37 °C under a 95% air, 5% CO2 humidified atmosphere
in minimum essential medium (MEM) containing nonessential
amino acids (100ꢀ, 10 mL/L), NaHCO3 (2.2 g/L), glutamine,
plasmocyne (5 μL/mL), pyruvic acid (0.11 g/L), and 10% (v/v)
fetal calf serum. 15 000 cpm (∼150 nM) of [4-14C]-estrone
(51.3 mCi/mmol) or [4-14C]-dihydrotestosterone (53 mCi/
mmol) (Perkin-Elmer Life Sciences, inc. Boston, MA) and an
ethanolic solution of inhibitor were added to freshly changed
culture medium and incubated for 7 h (E1 to E2) or 18 h (DHT to
3β-diol). Each inhibitor was assessed in duplicate or in triplicate.
After incubation time, culture medium was removed by pipeting
and steroids were extracted twice with 2 mL of diethyl ether. The
organic phases were pooled and evaporated to dryness under
reduced pressure. Residues were dissolved in CH2Cl2 and
dropped on silica gel 60 F254 thin layer chromatography plates
17β-(N-Alkylamino)-4-aza-5r-androstan-3-ones (78-81). 17β-
(N-Decylamino)-4-aza-5r-androstan-3-one (78). Compound 78
was prepared from 7715,16 (150 mg, 0.518 mmol) as described
for compounds 6 and 9 except that the formation of the imine was
˚
performed in a Schlenk tube containing molecular sieves, 4 A,
instead of a refluxing apparatus. The crude product was purified
by flash column chromatography (MeOH/CH2Cl2, 3:97 to 15:85)
to give a white solid (120 mg, 54% yield). IR (film on NaCl): 3188
and 3065 (NH), 1682 (CdO, lactam). 1H NMR 400 MHz
(CDCl3): 0.78 (s, 18-CH3), 0.87 (t, J = 7.3 Hz, 100-CH3), 0.89
(s, 19-CH3), 0.75-2.15 (residual CH and CH2), 2.38 (m, 2-CH2),
2.61 (t, J = 8.7 Hz, 17R-CH), 2.66 (m, 10-CH2), 3.03 (dd, J = 3.5
and 12.2 Hz, 5R-CH), 6.52 (s broad, NH). 13C NMR 75 MHz
(CDCl3): 11.36, 12.04, 14.13, 20.85, 22.69, 23.62, 27.33, 28.60,
29.30, 29.47, 29.58, 31.89, 33.35, 34.94, 35.78, 37.67, 42.92, 48.72,
51.30, 52.74, 60.73, 68.60, 172.05. LRMS: calcd for C28H51N2O
[M þ H]þ 431.4, found 431.4.
ꢀ
(VWR, Ville Mont-Royal, Quebec, Canada) for separation by
migration in toluene/acetone (4:1) as solvent system. Substrates
and metabolites were identified by comparison with reference
steroids and quantified using the Storm 860 imager (Molecular
Dynamics, Sunnyvale, CA). The % of transformation and the
% of inhibition were calculated as follows: % transformation
(E1 to E2) = [14C]-E2/([14C]-E1 þ [14C]-E2) ꢀ 100, % transfor-
mation (DHT into 3β-diol) = [14C]-3β-diol/([14C]-DHT þ
[14C]-3β-diol) ꢀ 100, and % inhibition = [(% transformation
without inhibitor) - (% transformation with inhibitor)]/(%
transformation without inhibitor) ꢀ 100.
17β-(N-Heptylamino)-4-aza-5r-androstan-3-one (79). Com-
pound 79 (120 mg, 30% yield) was prepared from 77 (300 mg,
1.037 mmol) as reported above for 78. IR (film on NaCl): 3200
and 3068 (NH), 1682 (CdO, lactam). 1H NMR 400 MHz
(CDCl3): 0.70 (s, 18-CH3), 0.87 (t, J = 7.0 Hz, 70-CH3), 0.87
(s, 19-CH3), 0.70-2.05 (residual CH and CH2), 2.35 (m, 2-CH2),
2.51 (t, J = 8.7 Hz, 17R-CH), 2.58 (m, 10-CH2), 3.01 (dd, J = 3.7
and 12.3 Hz, 5R-CH), 6.52 (NH). 13C NMR 75 MHz (CDCl3):
11.36, 11.95, 14.10, 20.92, 22.62, 23.68, 27.38, 27.45, 28.63,
29.24, 29.33, 29.56, 30.44, 31.83, 33.37, 35.05, 35.84, 37.92,
42.95, 49.03, 51.40, 52.83, 60.77, 69.00, 172.03. LRMS: calcd
for C25H45N2O [M þ H]þ 389.4, found 389.3.
For IC50 value determination, the HEK-293 cells overexpres-
sing 17β-HSD7 were plated at 600 000 cells/well in the presence
of 15 000 cpm of [14C]-E1 and the inhibitor in ethanolic solution
(0.1 nM to 20 μM). Assessment was performed in duplicate.
Cells were incubated overnight and steroids extracted as re-
ported above. IC50 values were determined using a homemade
ꢀ
program called DE50 1.64 (CRCHUL, Quebec, Canada).
5r-R Inhibition Assays ([4-14C]-Δ4-Dione into [4-14C]-A-
dione). HEK-293 cells transfected with pCMV-neo-h5R-reduc-
tase type 1 or type 2 construction were plated at 500 000 cells/
well in a 6-well plate at 37 °C under a 95% air, 5% CO2
humidified atmosphere in minimum essential medium (MEM)
17β-(N-Decylformamido)-4-aza-5r-androstan-3-one (80).
Compound 80 (15 mg, 30% yield) was prepared from 78
(46 mg, 0.107 mmol) according to the method used for com-
pound 16. IR (film on NaCl): 3176 (NH), 1682 (CdO, lactam

Article
Journal of Medicinal Chemistry, 2009, Vol. 52, No. 23 7501
complemented as described above. Incubation in the presence of
15 000 cpm of [14C]-4-androstene-3,17-dione (53.6 mCi/mmol)
(Perkin-Elmer Life Sciences, Inc., Boston, MA) and inhibitor
solution was performed in duplicate at two final concentrations
of 0.3 and 3 μM (10 μL). Incubation time for 5R-R1 was 1 h,
while incubation time for 5R-R2 was 3 h. After the incubation
period, steroids were extracted as reported for 17β-HSD7 assay
except that the TLC separation of radioactive materials (Δ4-
dione and A-dione) was achieved in toluene/EtOAc (4:1) as
solvent system. Radioactivity associated with each metabolite
was determined by the Storm 860 system, and the % of
transformation and % of inhibition were calculated as reported
above.
dehydrogenase/17-ketosteroid reductase (m17β-HSD7), previously
described as a prolactin receptor-associated protein (PRAP) in rat.
Mol. Endocrinol. 1998, 12, 1048–1059.
(4) (a) Peltoketo, H.; Nokelainen, P.; Piao, Y.; Vihko, R.; Vihko, P.
Two 17β-hydroxysteroid dehydrogenases (17HSDs) of estradiol
biosynthesis: 17HSD type 1 and type 7. J. Steroid Biochem. Mol.
Biol. 1999, 69, 431–439. (b) Nokelainen, P.; Peltoketo, H.; Mustonen,
M.; Vihko, P. Expression of mouse 17β-hydroxysteroid dehydrogenase
type 7 in the ovary, uterus, and placenta: localisation from implantation
to late pregnancy. Endocrinology 2000, 141, 772–778. (c) Pelletier, G.;
Luu-The, V.; Labrie, F. Localization of type 7 17β-hydroxysteroid
dehydrogenase in mouse tissues. In situ hybridization studies.
J. Steroid Biochem. Mol. Biol. 2005, 93, 49–57.
(5) (a) Krazeisen, A.; Breitling, R.; Imai, K.; Fritz, S.; Moller, G.;
Adamski, J. Determination of cDNA, gene structure and chromo-
somal localization of the novel human 17β-hydroxysteroid dehy-
drogenase type 7. FEBS Lett. 1999, 460, 373–379. (b) Breitling, R.;
Krazeisen, A.; Moller, G.; Adamski, J. 17β-Hydroxysteroid dehydro-
genase type 7, an ancient 3-ketosteroid reductase of cholesterogenesis.
Mol. Cell. Endocrinol. 2001, 171, 199–204. (c) Torn, S.; Nokelainen,
P.; Kurkela, R.; Pulkka, A.; Menjivar, M.; Ghosh, S.; Coca-Prados, M.;
Peltoketo, H.; Isomaa, V.; Vihko, P. Production, purification, and
functional analysis of recombinant human and mouse 17β-hydroxys-
teroid dehydrogenase type 7. Biochem. Biophys. Res. Commun. 2003,
17β-HSD5 Inhibition Assays ([14C]-Δ4-Dione into [14C]-T).
HEK-293 cells transfected with pCMV-neo-h17β-HSD5 con-
struction were plated at 500 000 cells/well and incubated in
presence of 15 000 cpm of [4-14C]-Δ4-dione and appropriate
concentrations (0.3 μM or 3 μM) of inhibitor in ethanol (in
triplicate). Incubation time for this enzymatic reaction was 20 h
at 37 °C and 5% CO2. After incubation, steroids were extracted,
separated, and quantified similarly as that reported in litera-
ture.13
ꢀ
305, 37–45.(d) Robert, A. Caracterisation de la 17β-HSD de Type 7
ꢀ
chez l'Humain. M.Sc. Thesis, Universite Laval, Quebec, Canada, 2004;
17β-HSD1 Inhibition Assay ([14C]-E1 into [14C]-E2). The
source of enzymatic activity was obtained from sonicated
HEK-293 cells transfected with h17β-HSD1 gene, and the
enzymatic assay was performed as previously described.24 The
inhibitor dissolved in ethanol was tested at two final concentra-
tions of 0.1 and 1 μM.
ꢀ
86 pp. (e) Lui, H. Clonage et Caracterisation de Deux Formes de la 17β-
ꢀ
ꢀ
ꢀ
Hydroxysteroide Deshydrogenase de Type 7 chez l'Humain. M.Sc.
ꢀ
Thesis, Universite Laval, Quebec, Canada, 2001; 85 pp.(f) Song, D.;
Liu, G.; Luu-The, V.; Zhao, D.; Wang, L.; Zhang, H.; Xueling, G.; Li, S.;
ꢀ
Desy, L.; Labrie, F.; Pelletier, G. Expression of aromatase and 17β-
hydroxysteroid dehydrogenase types 1, 7 and 12 in breast cancer. An
immunocytochemical study. J. Steroid Biochem. Mol. Biol. 2006, 101,
136–144.
(6) (a) Weihua, Z.; Lathe, R.; Warner, M.; Gustafsson, J. A. An
endocrine pathway in the prostate, ERbeta, AR, 5alpha-andros-
tane-3beta,17beta-diol, and CYP7B1, regulates prostate growth.
Proc. Natl. Acad. Sci. U.S.A. 2002, 99, 13589–13594. (b) Picciarelli-
Lima, P.; Oliveira, A. G.; Reis, A. M.; Kalapothakis, E.; Mahecha, G. A.
B.; Hess, R. A.; Oliveira, C. A. Effects of 3-beta-diol, an androgen
metabolite with intrinsic estrogen-like effects, in modulating the aqua-
porin-9 expression in the rat efferent ductules. Reprod. Biol. Endocri-
nol. 2006, 4, 51.
Acknowledgment. We are grateful to the Canadian Insti-
tutes of Health Research (CIHR) and Canadian Breast
Cancer Research Alliance (CBCRA) for providing the finan-
cial support for this project through grants. Special thanks go
ꢀ
to Guy Reimnitz for HEK-293 cell culture, Rene Berube for
HPLC analyses, Dr. Rene Maltais for collaboration on the
ꢀ
ꢀ
ꢀ
synthesisofsomecompounds, andDr. Shankar M. Singhwho
provided some compounds used in the screening study and the
intermediate compound 77. We also thank Micheline Harvey
for careful revision of the manuscript.
(7) (a) Marijanovic, Z.; Laubner, D.; Moller, G.; Gege, C.; Husen, B.;
Adamski, J.; Breitling, R. Closing the gap: identification of human
3-ketosteroid reductase, the last unknown enzyme of mammalian
cholesterol biosynthesis. Mol. Endocrinol. 2003, 17, 1715–1725.
(b) Ohnesorg, T.; Adamski, J. Promoter analyses of human and mouse
17beta-hydroxysteroid dehydrogenase type 7. J. Steroid Biochem.
Supporting Information Available: Results of a preliminary
screening of 150 compounds as potential inhibitors of the
transformation of E1 into E2 by 17β-HSD7 (Table A) and HPLC
purity of new synthesized compounds selected for additional
studies as inhibitors of 17β-HSD7 (Table B). This material
is available free of charge via the Internet at http://pubs.acs.
org.
ꢀ
Mol. Biol. 2005, 94, 259–261. (c) Ohnesorg, T.; Keller, B.; Hrabe de
Angelis, M.; Adamski, J. Transcriptional regulation of human and
murine 17β-hydroxysteroid dehydrogenase type-7 confers its participa-
tion in cholesterol biosynthesis. J. Mol. Endocrinol. 2006, 37, 185–
197.
(8) Krusche, C. A.; Moller, G.; Beier, H.; Adamski, J. Expression and
regulation of 17β-hydroxysteroid dehydrogenase 7 in the rabbit.
Mol. Cell. Endocrinol. 2001, 171, 169–177.
(9) (a) Schwabe, I.; Husen, B.; Einspanier, A. Expression of the
estradiol-synthesizing 17beta-hydroxysteroid dehydrogenases type
1 and type 7 in the nonhuman primate Callithrix jacchus. Mol. Cell.
Endocrinol. 2001, 171, 187-192. (b) Husen, B.; Adamski, J.; Bruns,
A.; Deluca, D.; Fuhrmann, K.; Moller, G.; Schwabe, I.; Einspanier, A.
Characterization of 17β-hydroxysteroid dehydrogenase type 7 in re-
productive tissues of the Marmoset monkey. Biol. Reprod. 2003, 68,
2092–2099.
(10) (a) Poirier, D. New cancer drugs targeting the biosynthesis of
estrogens and androgens. Drug Dev. Res. 2008, 69, 304-318.
(b) Poirier, D. Recent advances in development of inhibitors of 17β-
hydroxysteroid dehydrogenases. Anti-Cancer Agents Med. Chem.
2009, 9, 642–660. (c) Day, J.; Tutill, H.; Purohit, A.; Reed, M. Design
and validation of specific inhibitors of 17{β}-hydroxysteroid dehydro-
genases for therapeutic application in breast and prostate cancer, and in
endometriosis. Endocr.-Relat. Cancer 2008, 15, 665-692. (d) Brozic,
P.; Lanisnik Risner, T; Gobec, S. Inhibitors of 17β-hydroxysteroid
dehydrogenase type 1. Curr. Med. Chem. 2008, 15, 137–150.
(e) Poirier, D. Inhibitors of 17β-hydroxysteroid dehydrogenases. Curr.
Med. Chem. 2003, 10, 453-477. (f) Smith, J. H.; Nicholls, J. P.;
Simons, C.; Le Lain, R. Inhibitors of steroidogenesis as agents for the
treatment of hormone-dependant cancers. Expert Opin. Ther. Pat.
2001, 11, 789-824. (g) Penning, T. M. 17β-hydroxysteroid dehydro-
genase: inhibitors and inhibitor design. Endocr.-Relat. Cancer 1996, 3,
41–56.
References
(1) (a) Moeller, G.; Adamski, J. Integrated view on 17beta-hydroxys-
teroid dehydrogenases. Mol. Cell. Endocrinol. 2009, 301, 7–19.
ꢀ
(b) Luu-The, V.; Belanger, A.; Labrie, F. Androgen biosynthetic path-
ways in the human prostate. Best Pract. Res., Clin. Endocrinol.
Metab. 2008, 22, 207–221. (c) Lukacik, P.; Kavanagh, K. L.;
Oppermann, U. Structure and function of human 17β-hydroxysteroid
dehydrogenases. Mol. Cell. Endocrinol. 2006, 248, 61–71. (d) Vihko,
P.; Herrala, A.; Harkonen, P.; Isomaa, V.; Kaija, H.; Kurkela, R.; Pulkka,
A. Control of cell proliferation by steroids: the role of 17β-HSDs. Mol.
Cell. Endocrinol. 2006, 248, 141–148. (e) Luu-The, V. Analysis and
characteristics of multiple types of human 17β-hydroxysteroid dehy-
drogenase. J. Steroid Biochem. Mol. Biol. 2001, 76, 143–151.
(2) (a) Duan, W. R.; Linzer, D. I. H.; Gibori, G. Cloning and
characterization of an ovarian-specific protein that associates with
the short form of the prolactin receptor. J. Biol. Chem. 1996, 271,
15602–15607. (b) Shehu, A.; Mao, J.; Gibori, G. B.; Halperin, J.; Le, J.;
Sangeeta Devi, Y.; Merrill, B.; Kiyokawa, H.; Gibori, G. Hsd17b7
(PRAP/17β-hydroxysteroid dehydrogenase type 7) gene plays a crucial
role in embryonic development and fetal survival. Mol. Endocrinol.
2008, 22, 2268–2277.
(3) Nokelainen, P.; Peltoketo, H.; Vihko, R.; Vihko, P. Ex-
pression cloning of a novel estrogenic mouse 17β-hydroxysteroid

7502 Journal of Medicinal Chemistry, 2009, Vol. 52, No. 23
Bellavance et al.
(11) Labrie, F.; Poulin, R.; Simard, J.; Zhao, H. F.; Labrie, C.; Dauvois,
ꢀ
fragments: binding activity at D2 and D3 receptors. Bioorg. Med.
Chem. Lett. 2002, 12, 619–622.
S.; Dumont, M.; Hatton, A. C.; Poirier, D.; Merand, Y. Interac-
tions between estrogens, androgens, progestins, and glucocorti-
coids in ZR-75-1 human breast cancer cells. Ann. N.Y. Acad. Sci.
1990, 595, 130–148.
(18) (a) Bogeso, K.; Christensen, A. V.; Hyttel, J.; Liljefors, T. 3-Phenyl-
1-indanamines. Potential antidepressant activity and potent inhi-
bition of dopamine, norepinephrine and serotonin uptake. J. Med.
Chem. 1985, 28, 1817–1828. (b) Sterling, J.; Herzig, Y.; Goren, T.;
Finkelstein, N.; Lerner, D.; Goldenberg, W.; Miskolczi, I.; Molnar, S.;
Rantal, F.; Tamas, T.; Toth, G.; Zagyva, A.; Zekany, A.; Finberg, J.;
Lavian, G.; Gross, A.; Friedman, R.; Razin, M.; Huang, W.; Krais, B.;
Chorev, M.; Youdim, M. B.; Weinstock, M. Novel dual inhibitors of
AchE and MAO derived from hydroxy aminoindan and phenethyla-
mine as potential treatment for Alzheimer's disease. J. Med. Chem.
2002, 45, 5260–5279.
(12) Bellavance, E.; Luu-The, V.; Poirier, D. First inhibitors of the
steroidogenic enzyme type 7 17β-hydrosteroid dehydrogenase.
Lett. Drug Des. Discovery 2004, 1, 194–197.
(13) Bydal, P.; Luu-The, V.; Labrie, F.; Poirier, D. Steroidal lactones as
inhibitors of 17β-hydroxysteroid dehydrogenase type 5: chemical
synthesis, enzyme inhibitory activity, and assessment of estrogenic
and androgenic activities. Eur. J. Med. Chem. 2009, 44, 632–644.
(14) (a) Dufort, I.; Rheault, P.; Huang, X.-F.; Soucy, P.; Luu-The, V.
Characteristics of a highly labile human type 5 17β-hydroxysteroid
dehydrogenase. Endocrinology 1999, 140, 568–574. (b) Lin, H. K.;
Jez, J. M.; Schlegel, B. P.; Peehl, D. M.; Pachter, J. A.; Penning, T. M.
Expression and characterization of recombinant type 2 3r-hydroxys-
teroid dehydrogenase (HSD) from human prostate: demonstration of
bifunctional 3r/17β-HSD activity and cellular distribution. Mol. En-
docrinol. 1997, 11, 1971–1984. (c) Penning, T. M.; Burczynski, M. E.;
Jez, J. M.; Lin, H. K.; Ma, H.; Moore, M.; Ratnam, K.; Palackal, N.
Structure-function aspects and inhibitor design of type 5 17β-hydro-
xysteroid dehydrogenase (AKR1C3). Mol. Cell. Endocrinol. 2001,
171, 137–149.
(15) Li, X.; Singh, M. S.; Labrie, F. Synthesis and in vitro activity of
17β-(N-alkyl/arylformamido)- and 17β-[(N-alkyl/aryl)alkyl/
arylamido]-4-methyl-4-aza-3-oxo-5r-androstan-3-ones as inhibi-
tors of human 5r-reductases and antagonists of the androgen
receptor. J. Med. Chem. 1995, 38, 1158–1173.
(16) Rasmusson, G.; Reynolds, G.; Steinberg, N.; Walton, E.; Patel, G.;
Liang, T.; Cascieri, M.; Cheung, A.; Brooks, J.; Berman, C.
Azasteroids: structure-activity relationships for inhibition of 5r-
reductase and of androgen receptor binding. J. Med. Chem. 1986,
29, 2298–2305.
(19) Pine, S. H. Organic Chemistry, 5th ed.; McGraw-Hill, Inc.: New
York, 1987; 1057 pp.
(20) Puranen, T.; Poutanen, M.; Ghosh, D.; Vihko, P.; Vihko, R.
Characterization of structural and functional properties of humain
17β-hydroxysteroid dehydrogenase type 1 using recombinant en-
zymes and site-directed mutagenesis. Mol. Endocrinol. 1997, 11, 77–
86.
(21) (a) Flores, E; Bratoeff, E.; Cabeza, M.; Ramirez, E.; Quiroz, A.;
Heuze, I. Steroid 5r-reductase inhibitors. Mini-Rev. Med. Chem.
2003, 3, 225–237. (b) Jarman, M.; Smith, H. J.; Nicholls, P. J.; Simons,
C. Inhibitors of enzymes of androgen biosynthesis: cytochrome
P45017r and 5r-steroid reductase. Nat. Prod. Rep. 1998, 15, 495–512.
(22) Kenemans, P.; Bosman, A. Breast cancer and post-menopausal
hormone therapy. Best Pract. Res., Clin. Endocrinol. Metab 2003,
17, 123–137.
(23) Harkonen, P.; Torn, S.; Kurkela, R.; Porvari, K.; Pulka, A.;
Lindfors, A.; Isamaa, V.; Vihko, P. Sex hormone metabolism in
prostate cancer cells during transition to an androgen-independent
state. J. Clin. Endocrinol. Metab. 2003, 88, 705–712.
(24) Tremblay, M. R; Lin, S. X.; Poirier, D. Chemical synthesis of 16β-
propylaminoacyl derivatives of estradiol and their inhibitory po-
tency on type 1 17β-hydroxysteroid dehydrogenase and binding
affinity on steroid receptors. Steroids 2001, 66, 821–831.
(17) Dutta, A. K.; Fei, X.-S.; Reith, M. E. A. A novel series of hybrid
compounds derived by combining 2-aminotetralin and piperazine