A- And D-Ring structural modifications of an androsterone derivative inhibiting 17β-hydroxysteroid dehydrogenase type 3: chemical synthesis and structure-activity relationships

Article abstract of DOI:10.1021/acs.jmedchem.9b00624

Decreasing the intratumoral androgen biosynthesis by using an inhibitor of 17β-hydroxysteroid dehydrogenase type 3 (17β-HSD3) is a strategy to treat prostate cancer. The androsterone (ADT) derivative 1 (RM-532-105) has shown strong inhibitory activity on 17β-HSD3, but needs to be improved. Herein, we describe the chemical synthesis and characterization of two series of analogues to address the impact of A- and D-ring modifications on 17β-HSD3 inhibitory activity, androgenic effect, and metabolic stability. Structure-activity relationships were generated by adding different groups at C16/C17 (D-ring diversification) or replacing the ADT backbone by a nor-androstane or an estrane backbone (A-ring diversification). D-ring derivatives were less potent inhibitors than lead compound 1, whereas steroidal backbone (A-ring) change led to identifying promising novel estrane derivatives. This culminated with potent 17β-HSD3 inhibitors 23, 27, 31, and 33 (IC₅₀ = 0.10, 0.02, 0.13, and 0.17 μM, respectively), which did not stimulate LAPC-4 cell proliferation and displayed higher plasma concentration in mice than lead compound 1.

Full text of DOI:10.1021/acs.jmedchem.9b00624

Subscriber access provided by UNIV AUTONOMA DE COAHUILA UADEC
Article
A- and D-ring structural modifications of an androsterone
derivative inhibiting 17#-hydroxysteroid dehydrogenase type
: Chemical synthesis and structure-activity relationships
3
Francisco Cortés-Benítez, Jenny Roy, Martin Perreault, René Maltais, and Donald Poirier
J. Med. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.jmedchem.9b00624 • Publication Date (Web): 03 Jul 2019
Downloaded from http://pubs.acs.org on July 3, 2019
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a service to the research community to expedite the dissemination
of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in
full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully
peer reviewed, but should not be considered the official version of record. They are citable by the
Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore,
the “Just Accepted” Web site may not include all articles that will be published in the journal. After
a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web
site and published as an ASAP article. Note that technical editing may introduce minor changes
to the manuscript text and/or graphics which could affect content, and all legal disclaimers and
ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or
consequences arising from the use of information contained in these “Just Accepted” manuscripts.
is published by the American Chemical Society. 1155 Sixteenth Street N.W.,
Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the course
of their duties.

Page 1 of 68
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
nd
J. Med. Chem. (3 revised version)
A- and D-ring structural modifications of an androsterone derivative
inhibiting 17β-hydroxysteroid dehydrogenase type 3: Chemical synthesis and
structure-activity relationships
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Francisco Cortés-Benítez a,b,c, Jenny Roy , Martin Perreault , René Maltais , Donald Poirier
a
a
a
a,d,*
a
Laboratory of Medicinal Chemistry, Endocrinology and Nephrology Unit, CHU de Québec – Research
Center, Québec, QC, G1V 4G2, Canada
b
Department of Biological Systems, Biological and Health Sciences Division, Metropolitan Autonomous
University-Campus Xochimilco (UAM-X), Mexico City 04960, Mexico
c
Department of Pharmacy, Faculty of Chemistry, National Autonomous University of Mexico, Mexico
City, 04510, Mexico
d
Department of Molecular Medicine, Faculty of Medicine, Université Laval, Québec, QC, G1V 0A6,
Canada
(
*) Corresponding Author:
Donald Poirier
Laboratory of Medicinal Chemistry
CHU de Québec – Research Center (CHUL, T4-42)
2
705 Laurier Boulevard, Québec, QC, G1V 4G2, Canada
Tel.: 1-418-654-2296; Fax: 1-418-654-2298; E-mail: donald.poirier@crchul.ulaval.ca
1
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 2 of 68
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
ABSTRACT: Decreasing the intratumoral androgen biosynthesis by using an inhibitor of 17β-
hydroxysteroid dehydrogenase type 3 (17β-HSD3) is a strategy to treat prostate cancer. The
androsterone (ADT) derivative 1 (RM-532-105) has shown strong inhibitory activity on 17β-
HSD3, but needs to be improved. Herein, we describe the chemical synthesis and
characterization of two series of analogs to address the impact of A- and D-ring modifications on
17β-HSD3 inhibitory activity, androgenic effect, and metabolic stability. Structure-activity
relationships were generated by adding different groups at C16/C17 (D-ring diversification) or
replacing the ADT backbone by a nor-androstane or an estrane backbone (A-ring
diversification). D-ring derivatives were less potent inhibitors than lead compound 1, whereas
steroidal backbone (A-ring) change led to identifying promising novel estrane derivatives. This
culminated with potent 17β-HSD3 inhibitors 23, 27, 31, and 33 (IC = 0.10, 0.02, 0.13, and 0.17
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
5
0
µM, respectively), which did not stimulated LAPC-4 cell proliferation and displayed higher
plasma concentration in mice than lead compound 1.
Keywords: Hydroxysteroid dehydrogenase, Steroid, Ring modification, Enzyme inhibitor,
Prostate cancer, Chemical synthesis.
2
ACS Paragon Plus Environment

Page 3 of 68
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
INTRODUCTION
Prostate Cancer (PCa) is the most common cancer found in men. For instance, 174,650
1
American men were diagnosed with PCa and 31,620 men died from it in 2019. Initially, the
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
development and progression of PCa requires the androgen receptor (AR) signaling, as well as a
supply of androgens, particularly testosterone (T) and dihydrotestosterone (DHT) (Fig. 1), which
play a critical role in promoting the prostate tumor.2-4 In fact, it has been demonstrated that the
5
AR pathway drives PCa progression in most patients. Early detection of PCa, as well as the
arrival of new potent and selective treatments, could be two of the key aspects to improve the
survival and quality of life for patients with an androgen-sensitive PCa. With that mindset,
decreasing de novo biosynthesis of androgens in tumors with inhibitors targeting key
steroidogenesis enzymes represents an active and promising research field for the development
of innovative PCa treatments.6-10
Cholesterol
1
7-HSD5
O
OH
CYP11A1
CYP17A1
DHRS11
3
-HSD
17-HSD3
1
8
O
O
O
1
4-dione
T
7
19
1
C
D
RM-532-105
16
5
-R
2
5-R
OH
B
15
O
A
HO
7
1
7-HSD3
3
5
4
DHEA
1
7-HSDs
O
O
H
H 5 -dione
5,15
DHT

O
3-HSD
17-HSD6
3-HSD
H
17-HSD6
O
OH
Others (?)
Others (?)
H
H
N
17-HSDs
,15
O
N
HO
H
5
S
O
CF3
RM-532-105
HO
HO
H
H
ADT
5-diol
Figure 1. Contribution of 17β-hydroxysteroid dehydrogenase type 3 (17β-HSD3) to the biosynthesis of
the androgens testosterone (T) and 5α-dihydrotestosterone (DHT) from 4-androstene-3,17-dione (4-dione)
3
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 4 of 68
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
and 5α-androstene-3,17-dione (5α-dione), respectively. RM-532-105 is a 3β-substituted derivative of
androsterone (ADT) that inhibits the steroidogenic enzyme 17β-HSD3. Steroid (A-D) ring identification
and partial carbon numbering are reported for DHEA (also see Figure 5 for a full carbon
numbering of RM-532-105 and its analogs).
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
17β-hydroxysteroid dehydrogenase type 3 (17β-HSD3) is a member of the 17β-HSD
family, which is constituted of several types (1-15) that catalyze the reduction of a carbonyl
group at position 17 or the oxidation of a 17β-hydroxy group for different steroids.11-19 This key
enzyme also catalyzes the transformation of the weak androgen 4-androstene-3,17-dione (4-
dione) into T, which is further converted to the most potent androgen DHT (Fig. 1). T and DHT
eventually interact with AR to promote the growth and proliferation of PCa cells. Even though
1
7β-HSD3 is expressed mainly in the microsomal fraction of testes,20 Pfeiffer et al.21 have
reported that in castrate-resistant prostate cancer (CRPC) biopsies, the expression of 17β-HSD3
is up-regulated, even after long-term hormone depletion. Although extensive researches have
been carried out on the enzyme inhibitors mentioned above, there are no approved drugs to date
for 17β-HSD3 inhibitors. Thus, this enzyme has been and remains an interesting target for PCa
hormonal therapy and its inhibition should be able to maximize the androgen depletion in
androgen-dependent PCa tumors.
During our efforts to discover novel 17β-HSD inhibitors, we have reported the synthesis
22
and biological activity of RM-532-105. This compound is an androsterone (ADT) derivative
having a 4-(2,5-dimethyl-4-((2-(trifluoromethyl)phenyl)sulfonyl)piperazin-1-yl)methyl side
chain at position 3β (Fig. 1). RM-532-105 has shown potent inhibitory activity on 17β-HSD3 in
HEK-293[17β-HSD3] cells, LNCaP[17β-HSD3] cells22-24 as well as in microsomal fraction of
4
ACS Paragon Plus Environment

Page 5 of 68
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
rat testes.23 Moreover, this lead compound significantly decreased T and DHT levels when
2
5
administered subcutaneously in rats. Although RM-532-105 significantly accumulated inside
the tumor of LAPC-4 xenografts, it was not able to reduce the growth of tumor.26
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Recently, we reported the synthesis and biological activity of two RM-532-105
2
7
stereoisomers, the first one possessing an inversed hydrogen on carbon 5 (5β-H instead of 5α-
H) while the second one bears an inversed methyl carbon attached to C13 (13α-CH instead of
3
1
3β-CH ). Both modifications were carried out to explore the impact on 17β-HSD3 inhibitory
3
activity, androgenic profile and metabolic stability by structural modifications on the A- and D-
ring shape, respectively. From this study, we first found that RM-532-105 has a relatively low
metabolic stability, leading to unfavorable pharmacokinetics. Conversely, A-ring inversion in a
5
β configuration was well-tolerated by the 17β-HSD3 enzyme, but did not improve its metabolic
stability. On the contrary, D-ring inversion significantly decreased the anti-17β-HSD3 activity
while increasing its metabolic stability.
Following this preliminary structure-activity relationship (SAR) study,27 we are now
interested in further studying the impact of additional modifications on 17β-HSD3 inhibition and
on their metabolic stability. Since no crystal structure of 17β-HSD3 is available to date, and
considering the low predictive capacity of these homology models, we decided to focus, in the
present study, on a classical approach by using systematic structure modifications and SAR study
to progress toward compounds with improved activities. Thus, we describe herein the chemical
synthesis of several RM-532-105 analogs (Fig. 2) and show the impacts of A- and D-ring
modifications on their 17β-HSD3 inhibitory activity, cell proliferative effect, and metabolic
stability. This study will expand our current understanding of the optimal pharmacophore
requirements for 17β-HSD3 inhibitory activity and their ability to block androgen biosynthesis
5
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 6 of 68
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
for therapeutic purposes.
RESULTS AND DISCUSSION
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Design Strategy. RM-532-105 is a metabolically labile steroid derivative showing 16%
2
7
of remaining compound after 1 hour in a microsomal preparation of human liver. Such low
metabolic stability could generate problems of low bioavailability or adverse drug reactions due
to off-target interactions at the high doses needed to obtain the desired pharmacological action.
Thus, we first considered synthetic modifications of RM-532-105 at D-ring because we
suspected that this compound may be hydroxylated or reduced at position C16 or C17,
respectively, by metabolic enzymatic reactions. In fact, it is well-known that those two reactions
28
are involved in the formation of steroidal metabolites. Therefore, the addition of a substituent at
C16 or C17 can decrease their reactivity over cytochrome P450 enzymes. We also considered
modifications at A-ring by using an estrane backbone, instead of the androstane backbone, and
consequently eliminated the presence of a tertiary alcohol at C3.
Our first attempt to reduce the metabolism was initially focused on the A-ring shape of
RM-532-105, but was ineffective, since the inversion of configuration (5α-H for 5β-H) altered
2
7
neither 17β-HSD3 inhibitory activity nor metabolic stability. However, we recently found that
the replacement of the C19-steroid backbone of ADT by the C18-steroid backbone of estrone,
significantly enhanced the metabolic stability for a series of anti-cancer aminosteroids.29
Therefore, this strategy could also be applied to 17β-HSD3 inhibitors, but remains to be
validated, since no attempt of estrane mimics has been reported to date for 17β-HSD3 inhibitors.
Based on these rational considerations, we thus proposed two series of systematic modifications,
keeping the 3β-optimized side chain intact, to obtain a RM-532-105 analog with low hepatic
6
ACS Paragon Plus Environment

Page 7 of 68
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
clearance and good inhibitory activity against 17β-HSD3 (Fig. 2).
1
7-HSD3 Inhibitor
SAR Studies
(
lead compound: RM-532-105 (1))
D-ring modifications
-
Diversification at C17, C16,
and C16/C17
1
2
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
O
1
Optimized Side Chain
7
19
H
16
(compounds 2-10, and 16)
2
H
H
N
A-ring modifications
3
O
N
HO
H
S
- C19-nor, aromatic
- C2/C3-position
O
CF3
Steroid Scaffold
androsterone)
(
compounds 21, 23, 26-28,
1, and 33)
(
3
Figure 2. Structure of the 17β-HSD3 inhibitor RM-532-105 (1) and two strategies (# 1-2) to optimize its
inhibitory activity and metabolic stability.
To explore the SAR of D-ring modifications and to determine whether blocking C16
and/or C17 positions by adding several substituents could improve metabolic stability, we
synthetized RM-532-105 analogs 2-10 and 16 with varied groups such as hydroxyl, fluorine,
methoxy, ethinyl, oxime, oxirane, and acetyl (Schemes 1 and 2).
Changing the androstane A-ring by an estrane has been a successful strategy for the
discovery of steroidal compounds with better affinities and selectivity, as well as to improve
30
physicochemical and ADMET properties. We thus planned the synthesis of A-ring analogs of
RM-532-105 like nor-androstane derivative 21 (Scheme 2) and estrane derivatives 23, 26-28, 31,
and 33 (Scheme 3). In addition to avoiding the presence of a tertiary alcohol at C3, an aromatic
A-ring could facilitate the addition of a substituent at positions that are difficult to access by
chemical modifications of the androstane A-ring.
7
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 8 of 68
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Chemistry. D-ring modifications. D-ring analogs at position C16 (compounds 2-5) and at
position C17 (compounds 6-10) were prepared from the lead inhibitor 1 (Scheme 1) while
another analog at C17 (compound 16) was obtained from 3β-hydroxy-5α-pregnan-20-one
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
(
Scheme 2). Briefly, compound 1 was refluxed with CuBr to yield the brominated intermediate
2
which was later reacted with aqueous NaOH to obtain the corresponding hydroxyl derivative 2,
whereas fluorinated compound 4 was readily synthesized using SelectfluorTM according to the
3
1
1
methodology reported by Liu et al. The H NMR spectra for 2 and 4 presented characteristic
signals corresponding to the 16-H proton at 4.35 ppm and 5.08 ppm, respectively. Moreover, in
NOESY experiments, these signals correlated through the space with 18β-CH , confirming that
3
the 16-H proton is oriented to the β-face of the steroid. Conversely, the carbonyl group at C17
for compounds 1, 2, and 4 was reduced to its corresponding alcohol, yielding 3, 5, and 6,
1
3
respectively. In the C NMR spectra, we observed that the 17-CHOH signal appeared at 80-90
3
2
ppm, which signal is also characteristic of the expected C17-stereochemistry. Surprisingly,
compound 5 was obtained as a stereoisomeric mixture according to its NMR data and several
attempts to separate the two stereoisomers were unsuccessful.
O
H
OH
OCH3
b
d
H
H
N
O
N
HO
H
S
H
6
H
7
O
OH
CF3
1
e
H
8
O
OH
a
b
b
NOH
OH
OH
f
H
H
H
3
2
H
9
O
OH
O
c
g
F
F
H
H
4
5
10
8
ACS Paragon Plus Environment

Page 9 of 68
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Scheme 1. Synthesis of compounds 2-10. Reagents and conditions: (a) 1. CuBr , MeOH, 65 °C,
2
TM
overnight, 2. NaOH 1 M, DMF, DCM, rt, 3 h; (b) NaBH , MeOH, 0 °C, 1 h; (c) Selectfluor , H SO 6
4
2
4
M, MeOH, 60 °C, 72 h; (d) NaH, CH I, DMF, rt, overnight; (e) 1. trimethylsilylacetylene, MeLi,
3
THF/diethyl ether (1:1), 0 °C to rt, overnight, 2. 5% K CO in MeOH, rt, 3 h; (f) hydroxylamine
2
3
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
hydrochloride, sodium acetate, EtOH, 60 °C, 1 h; (g) (CH ) SI, NaH, DMSO, THF, rt, 4.5 h.
3
3
Compounds 7 (17β-OCH ) was prepared from 6 by a selective O-methylation and .the
3
1
7-OCH was clearly identified as a single signal at δ = 3.31 ppm and δ = 57.8 ppm as well.
3
H
C
The compound 8 was synthesized by reacting 1 with lithium trimethysilylacetylide to lead a
TMS-acetylenic intermediate, which was then treated with K CO in MeOH to give the
2
3
corresponding 17α-C≡CH derivative. For this final compound, the characteristic 17-CHOH
signal in the 13C NMR spectra appeared at 79.9 ppm. Moreover, the 17α-C≡CH group was
associated to signals at δ = 2.56 ppm, δ = 87.6 ppm (17α-C≡CH) and δ = 73.8 ppm (17α-
H
C
C
C≡CH). NOESY experiments and literature data 33 confirmed the stereoselective attack by the α-
face of the steroid.
The C17-oxime 9 was readily generated in high yield by treating compound 1 with
hydroxylamine hydrochloride and sodium acetate in ethanol at 60 °C. The NOH proton was
1
clearly identified in the H NMR as a broad singlet at 8.41 ppm. Furthermore, the C17 signal in
13
the C NMR spectra was shifted from 221.5 ppm to 171.2 ppm. Conversely, the ketone of 1 was
3
4
efficiently reacted under Corey-Chaykovsky epoxidation conditions, using trimethyl-sulfonium
iodide and sodium hydride, to obtain the oxirane 10. The characteristic diastereotopic protons of
the methylene group from epoxide were observed as two signals at δ = 2.59 and 2.89 ppm.
H
These signals, and those for 18-CH (δ = 0.86 ppm and δ = 14.4 ppm), are characteristic of the
3
H
C
9
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 10 of 68
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
expected stereochemistry (17β-O).35
O
O
X
O
H
a
H
R
H
b
H
H
H
H
H
H
H
HO
HO
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
O
H
H
1
2: R = CH ; X= -C(O(CH ) O)CH
3
-hydroxy-5-pregnan-20-one
3
2 2
3
1
1
1
7: R = H; X= -OH
X
c
X
R
H
H
X
R
H
H
H
H
R
H
N
e
d
O
N
HO
H
H
N
H
H
S
HN
HO
O
O
H
CF3
1
5: R = CH ; X: -C(O(CH ) O)CH
14: R = CH ; X: -C(O(CH ) O)CH
3
2 2
3
13: R = CH ; X = -C(O(CH ) O)CH
3
2 2
3
3
2 2
3
3
f
16: R = CH ; X = -COCH
3
20: R = H; X = O
18: R = H; X = -OH
19: R = H; X = O
b
21: R = H; X = O
Scheme 2. Synthesis of D-ring derivative 16 and A-ring derivative 21. Reagents and conditions: a)
ethylene glycol, p-TSA, toluene, reflux, 12 h; b) Dess-Martin, DCM, rt, 1 h; c) (CH ) SOI, NaH, DMSO,
3
3
THF, rt, 5 h; d) trans-2,5-dimethylpiperazine, ethanol, 70 °C, 12 h; e) 2-(trifluoromethyl)benzenesulfonyl
chloride, TEA, DCM, rt, 4 h; f) HClO , acetone, 0 °C, 30 min.
4
The pregnane derivative 16 (X = COCH ) was synthesized from commercially available
3
3
β-hydroxy-5α-pregnan-20-one, which was first protected at C20 as the acetal 11 (Scheme 2).
After subsequent oxidation of the 3β-OH of 11 using the Dess-Martin reagent, the resulting
ketone 12 was then submitted to the epoxidation of Corey-Chaykovsky, using
trimethylsulfoxonium iodide and sodium hydride, to give 13. The oxirane of 13 was then opened
using trans-2,5,-dimethyl-piperazine in refluxing ethanol, and the resulting secondary amine 14
was reacted with 2-(trifluoromethyl)benzenesulfonyl chloride and TEA to give 15. Finally,
hydrolysis of the acetal group in acidic conditions leads to pregnane derivative 16.
10
ACS Paragon Plus Environment

Page 11 of 68
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
A-ring modifications. The next strategy consisted in replacing the ADT backbone of 1 by
a C19-nor-ADT core (21; Scheme 2) or a C18-estra-1,3,5(10)-triene core (23, 26-28, 31, and 33;
Scheme 3). Compound 21 was prepared from 19-nor-DHT (17) as starting material using the
same chemical route reported in Scheme 2 for the synthesis of 16. The NMR spectra of 21
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
showed all characteristic signals related to compound 1, except the 19-CH signal that is lacking.
3
For the synthesis of estrane series (Scheme 3), the amine A was first obtained by a condensation
of 2-(trifluoromethyl)benzenesulfonyl chloride with trans-2,5-dimethylpiperazine as previously
reported 27 and next introduced at position C3 or C2. We thus prepared the amide 23 by means of
a coupling between the carboxylic acid 22 37 and the amine A using HBTU as a coupling agent.
Conversely, the oxirane 24 38, 39 served as an intermediate for the synthesis of 26-28. First, this
oxirane was reacted with amine A in ethanol at 70 C to yield the diastereomeric mixture 25,
which was thereafter purified by HPLC to afford 26 and 27. Interestingly, for these two isomers,
the chiral carbinol can be clearly differentiated by their NMR data. For instance, this proton
appears as two double signals at δ = 4.66 and 4.54 ppm for 26 and 27, respectively.
H
11
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 12 of 68
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Estrone
O
CF3
O
CF3
O
S
O
H
H
a
S
N
N
O
O
H
H
O
H
H
N
N
H
HOOC
Amine A
O
23
2
2
O
O
H
H
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
c
H
H
H
b
H
H
N
N
H
O
O
N
F
H
OH
N
S
28
2
5
S
O
O
O
24
CF3
CF3
Resolution
26 (diastereomer A)
27 (diastereomer B)
O
O
O
H
H
d
H
e
H
H
N
H
H
H
H
O
N
HO
Br
S
31
O
29
30
CF3
O
O
H
H
f
OHC
HO
N
O
H
H
N
H
H
S
HO
O
33
3
2
CF3
Scheme 3. Synthesis of compounds 23, 26-28, 31, and 33. Reagents and conditions: (a) amine A, HBTU,
DIPEA, DMF, rt, 2 h; (b) amine A, EtOH, 70 °C, 48 h; (c) DAST, DCM, rt, 3 h; (d) PPh , CBr , DCM, 0
3
4
°
C, 1.7 h; (e) amine A, KI, DIPEA, CH CN, 65 °C, 72 h; (f) amine A, NaBH CN, AcOH, molecular
3 3
sieves, MeOH/DCM (8:2), rt, 4 h.
The different inhibitory activities observed with alcohols 26 and 27, the latter being the
most active, encouraged us to address their composition. After the proton and carbon assignment
was established using a combination of HSQC, HMBC and COSY experiments, and literature
data, different chemical shifts were clearly observed for CH-22, CH-24, CH -27, and CH -28. In
2
3
1H NMR, chemical shifts are close for 26 and 27, but small differences are observed for H-22
(
4.66 and 4.54 ppm), H-24 (3.06 and 2.92 ppm), 1H of CH -27 (3.06 and 2.92 ppm), and CH -28
2 3
1
3
(
0.95 and 0.86 ppm). More significant effects were however observed in C NMR for two
12
ACS Paragon Plus Environment

Page 13 of 68
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
carbons (Table S1; Supporting Information). Thus, the CH-24 signals appear at 50.2 and 55.0
ppm, while the CH -27 signals appear at 52.6 and 48.1 ppm, for 26 and 27, respectively.
2
From the NMR analysis, we noticed that the CH -25 appears as two well differentiated
2
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
signals, one (Ha) producing an NOE correlation with CH -29 (# 1) and the other (Hb) with CH -
3
3
28 (# 2) (Fig. 3). Since the two nitrogens of piperazine are differently substituted, there are
1
however two possible trans-diaxial configurations (Fig. 3A-D). The H NMR spectra of the
Mosher ester of the mixture of alcohols 25 (R-OCO-C(CF )(OCH )Ph) showed the presence of
3
3
four OCH singlet signals, suggesting the presence of four alcohols with 22R,24R,26S;
3
2
2R,24S,26R; 22S,24R,26S; and 22S,24S,26R configurations (Fig. S1, Supporting Information).
We next hypothesized that a hydrogen bond was formed between the alcohol (OH) and the free
nitrogen electron doublet, then producing a five-member ring that fixed the orientation of the two
different C22 substituents (H and steroid A-ring (Ar)), as shown in Fig. 3A-D. The NOESY
spectra of 26 and 27, which compounds were obtained by chromatography of the mixture of
alcohols 25 generated from the opening of oxirane 24, then provided interesting results (Fig.
3E,F). In addition to the expected H-22/H-23 correlation, we observed in the NOESY spectrum
of 26 a correlation between H-22 and a multiplet at 3.1 ppm, which corresponds to H-24 and 1H
of CH -27 (H-27a), but not with the other H-27 (H-27b) at 2.2 ppm. On the other hand, the
2
NOESY spectrum of 27 showed a correlation between H-22 and H-23 and 1H of CH -27 at 2.5
2
ppm (H-27b), but not with the multiplet at 3.0 ppm, which encompasses the other H-27 (H-27a)
and H-24. These NOESY results suggest that 26 is a mixture of two alcohols (22R,24R,26S (Fig.
3
A) and 22S,24S,26R (Fig. 3D)) while 27 is a mixture of two alcohols (22R,24S,26R (Fig. 3B)
and 22S,24R,26S (Fig. 3C)). The proximity of H-22 to H-27b that is responsible for the nuclear
overhauser (NOE) effect is clearly observed in the optimized 3D structure of 27 (Fig. 3G).
13
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 14 of 68
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 3. Important NOE correlations observed for compounds 26 (A, D, and E) and 27 (B, C,
and F) as well as an optimized structure showing the proximity of H-22 to H-27b (G). Figure G
was built by assembling the side chain to the C3-position of estrone (E1). The structure of E1
4
0
was retrieved from the ZINC database. The geometry of the resulting 22S,24R,26S-compound
was first minimized by a UFF force field and then optimized by the semiempirical PM6 method
4
1
using Gaussian 09 software. The figures A-D were produced with ChemDraw 14.0 and figure
G with UCSF Chimera program.42
14
ACS Paragon Plus Environment

Page 15 of 68
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Diastereomeric mixture of alcohols 25 was also reacted with DAST to give fluorinated
derivative 28 (Scheme 3). This derivative was prepared to address the bioisosteric replacement of
OH by a fluorine atom. Compound 31, having an amine linked at C3 by an ethyl group, was
designed to see the impact of the absence of an OH group close to the steroid A-ring. Thus, the
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
3
9
Appel reaction was carried out on the intermediate 29, by using CBr and PPh , giving 30 in
4
3
good yield. This bromoethyl steroid was then reacted with amine A combined with a catalytic
amount of potassium iodide to obtain the desired product 31. Finally, compound 33 was readily
prepared via a reductive amination between the amine A and 2-formyl-estrone (32), readily
4
3
obtained from estrone. In NMR, the new methylene group formed at C2 appears at δ = 3.61-
H
3.69 and δ = 57.8 ppm. Interestingly, the OH attached at C3 is shifted downfield to δ = 10.23
C
H
ppm, suggesting an intramolecular H-bond with the tertiary amine in the piperazine moiety,
which may be favored since a six-member ring is formed.
+
1
7β-HSD3 inhibition and LAPC-4 (AR ) cell proliferative activity (Structure-
activity relationships). The inhibition of 17β-HSD3 activity was carried out using transfected
LNCaP cells overexpressing 17β-HSD3 (LNCaP[17β-HSD3] cells) as a source of enzyme
activity. In this study, we measured the amount of labeled T formed from labeled 4-dione in the
presence of the designed inhibitors at three concentrations (0.1, 0.5, and 1.0 µM). IC values
5
0
were then determined for the best compounds (Table 1). To explore whether the synthesized
compounds exhibit an androgenic profile, which is an undesired effect in the context of prostate
cancer therapy, we assessed their proliferative activities on androgen-sensitive prostate cancer
+
LAPC-4 (AR ) cells at three inhibitor concentrations (0.1, 0.5, and 1.0 µM). The basal cell
proliferation being fixed at 100%, a proliferative value (in %) over 100% will be associated with
15
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 16 of 68
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
an androgenic activity whereas a value (in %) under 100% will be associated with a cytotoxic
effect. RM-532-105 (1) and DHT were used as reference compounds in the enzymatic and
androgen-sensitive cell proliferation assays, respectively.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
D-ring modifications. Initially, we prepared and tested compounds 2-5 to assess the effect
of adding a group, such as OH or F. As results, small modifications at this position significantly
decreased the anti-17β-HSD3 activity. The 16α-fluorine derivative 4 is 2-fold less potent than 1,
whereas 2 lack inhibitory activity. Surprisingly, altering C16 position clearly increases the
proliferation of LAPC-4 cells, but this effect is not dose dependent. For instance, the presence of
a fluorine atom exhibited an androgenic activity at 0.1 μM (147.7%). In addition, 3 and 5, which
are the C17-OH counterparts of 2 and 4, did not exert inhibitory activity on 17β-HSD3, whereas
5
enhanced the proliferative capability of LAPC-4 cells to a greater extent than 4.
+
Table 1. 17β-HSD3 inhibition and LAPC-4 (AR ) cell proliferation of D-ring modified
compounds 2-10 and 16
O
17
H
D
16
H
H
N
O
N
HO
H
S
O
CF3
Cpd
D-ring
17β-HSD3 inhibition (%)a
IC50b
(µM)
LAPC-4 cell proliferation (%)c
group
0.1 µM / 0.5 µM / 1.0 µM
0.1 µM / 0.5 µM / 1.0 µM
O
1
2
60 / 96 / 100
0 / 60 / 80
0.09d
---
96.0 / 47.6 / 32.9
114.0 / 83.1 / 69.3
O
OH
16
ACS Paragon Plus Environment

Page 17 of 68
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
OH
O
3
4
5
6
7
8
9
20 / 48 / 55
22 / 75 / 90
12 / 58 / 70
38 / 78 / 94
10 / 18 / 30
5 / 30 / 40
40 / 80 / 90
5 / 10 / 20
---
90.8 / 54.1 / 43.8
147.7 / 134.2 / 126.6
153.6 / 149.7 / 142.4
116.4 / 88.1 / 31.3
---
OH
F
0.20
---
OH
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
F
OH
0.19
---
OCH3
OH
---
130.9 / 95.8 / 81.2
107.8 / 58.9 / 44.0
92.6 / 61.2 / 49.3
NOH
0.22
---
O
1
0
6
O
1
10 / 30 / 42
2.6
45.9 / 26.4 / 27.1
a
14
14
Inhibition at three concentrations of the transformation of [ C]-4-dione (15 nM) to [ C]-T by 17β-
HSD3 in LNCaP cells overexpressing 17β-HSD3. b The concentration that inhibits 50% of 17β-HSD3
activity (IC ). The assay was performed using seven concentrations (0.001 – 5.0 μM). IC values were
5
0
50
calculated using GraphPad Prism 6 software. Values represent the average of two independent
c
experiments performed in triplicate.
Proliferation of androgen-sensitive LAPC-4 cells at three
concentrations. At the end of the assay, the proliferation of the cells not treated with a drug (negative
control) was fixed to 100% while the proliferation of cells treated with the potent androgen DHT at 10
d
nM (positive control) was 190%. Mean value of 0.06, 0.067, and 0.13 µM.
On the other hand, we also explored the impact of a modification at C17 on the inhibitory
17
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 18 of 68
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
activity. Thus, the C17-ketone was reduced to the corresponding alcohol (17β-OH, 6), but this
compound showed a 2-fold less inhibitory activity on 17β-HSD3 when compared to 1. The O-
methylation of 6 (compound 7) also produced a detrimental effect on 17β-HSD3 inhibition.
Another finding emerged when a substituent was added to the α-face in the presence of a 17β-
OH. We noted that 8, having a 17α-C≡CH, was inactive against 17β-HSD3. Even worse, 8
caused a significant proliferation of LAPC-4 cells (130.9%) at low concentrations of 0.1 µM.
Interestingly, when 17-C=O was changed to 17-C=NOH, the resulting oxime 9 was only 2-fold
less effective as an inhibitor than 1 on 17β-HSD3 (IC = 0.22 μM) and it did not stimulate the
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
5
0
cell proliferation. The introduction of an oxirane at C17 (compound 10) greatly reduced the
inhibition of 17β-HSD3, but this compound did not induce cell proliferation. Pregnane derivative
1
6, with an acyl group at C17β, also weakly inhibited 17β-HSD3 and produced an
antiproliferative (cytotoxic) effect on LAPC-4 cells.
These SAR data suggest that altering positions C16 and C17 will have a moderate to
negative impact on the 17β-HSD3 inhibitory activity and a mixed effect on LAPC-4 cell
proliferation. Regarding 17β-HSD3, we observed that the presence of a hydroxyl group at the
C17 position is relatively well tolerated since compounds 6 (17β-OH) and 9 (17-NOH) were the
most active inhibitors in the 17-substituted series. It is however difficult at first sight to assign
the exact nature of the molecular interaction of the hydroxyl group, since that functional group
can act as an H-bond donor group (HBD) or as an H-bond acceptor group (HBA) depending on
the molecular context. However, the excellent activity of the C17 ketone, as an HBA group, let
us suppose that the HBA capacity of the OH is in play. On the other hand, compounds 7 (OCH3)
and 16 (COCH ) were inactive, despite their HBA capacity. The steric hindrance could be a
3
major factor to explain this loss of inhibition potency by preventing the HBA interaction. These
18
ACS Paragon Plus Environment

Page 19 of 68
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
results point toward a weak tolerance for a bulky group in this area of 17β-HSD3. In this way, it
is also important to note that attaching a group on the α-face of C17 (e.g. 8 and 10) has a
negative impact for the inhibition of 17β-HSD3. Thus, we believe that a substituent at C17
prevents an efficient hydrogen bonding with a residue in the binding pocket of 17β-HSD3. This
observation reinforced our finding that a very precise position of C17 oxygen is necessary to
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
+
favor H-bonding with 17β-HSD3. Regarding the proliferative activity in LAPC-4 (AR ) cells,
this series of compounds gave mixed proliferative and antiproliferative results. Among the most
active inhibitors of the series, the 17β-OH and 17-NOH derivatives 6 and 9, respectively gave a
very slight proliferative activity at 0.1 µM and an antiproliferative activity at 0.5 and 1.0 µM.
Conversely, the addition of a fluoride atom at C16α resulted in clear proliferative activity (see
compounds 4 and 5). Despite its acyl group at C17β, pregnane derivative 16 produced the higher
antiproliferative activity.
+
Table 2. 17β-HSD3 inhibition and LAPC-4 (AR ) cell proliferation of A-ring modified
compounds 21, 23, 26-28, 31, and 33
O
H
O
R =
S
N
N
H
H
R
O
CF3
HO
H
Cpd
1
A-ring
group
17β-HSD3 inhibition (%)a
IC50a,b
(µM)
LAPC-4 cell proliferation (%)c
0.1 µM / 0.5 µM / 1.0 µM
0.1 µM / 0.5 µM / 1.0 µM
d
e
60 / 96 / 100
0.09 [0.08]
96.0 / 47.6 / 32.9
R
HO
H
19
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 20 of 68
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
H
H
2
2
1
3
20 / 82 / 94
45 / 78 / 90
0.22
92.8 / 66.3 / 63.1
60.0 / 45.6 / 36.7
R
R
HO
0.10 [0.11]e
O
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
R
2
2
6
7
50 / 95 / 98
84 / 98 / 100
~30 / 80 / 90
0.14
115.0 / 94.9 / 81.9
103.2 / 68.0 / 56.4
---
OH
diastereomer A
R
0.02 [0.06]e
OH
diastereomer B
R
2
3
8
1
~0.25
F
diastereomeric mixture
.13 [0.10]e
0
0
38 / 60 / 70
50 / 76 / 84
51.9 / 19.6 / 21.4
99.6 / 92.8 / 93.6
R
.17 [0.12]e
R
33
HO
a
14
14
Inhibition at three concentrations of the transformation of [ C]-4-dione (15 nM) to [ C]-T by 17β-
b
HSD3 in LNCaP cells overexpressing 17β-HSD3. The assay was performed using seven concentrations
(
0.001 – 5.0 μM). IC values were calculated using GraphPad Prism 6 software. Values represent the
50
average of two independent experiments performed in triplicate. c Proliferation of androgen-sensitive
LAPC-4 cells at three concentrations. At the end of the assay, the proliferation of the cells not treated with
a drug (negative control) was fixed to 100% while the proliferation of cells treated with the potent
d
e
androgen DHT at 10 nM (positive control) was 190%. Mean value of 0.06, 0.067, and 0.13 µM. IC
5
0
1
4
14
values for the transformation of [ C]-4-dione (50 nM) into [ C]-T by 17β-HSD3 from homogenized rat
20
ACS Paragon Plus Environment

Page 21 of 68
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
testes.
A-ring modifications. To better understand the impact of the nature of the steroid A-ring,
we next assessed the biological activities of compounds 21, 23, 26-28, 31, and 33 (Table 2).
These analogs of compound 1 are potent 17β-HSD3 inhibitors and did not display a proliferative
activity in LAPC-4 cells. Precisely, the order of potency for this new series of 17β-HSD3
inhibitors was 27 > 23 > 31 = 26 > 33 > 21 > 28. First, we observed that the absence of 19-CH3
in compound 21 (or 19-nor-RM-532-105) had a negative impact and caused a drop of the
inhibitor potency (~2.5 fold less active than 1). The most striking aspect of the data was
displayed by replacing the saturated A-ring of 1 for an aromatic ring. This modification yielded
potent 17β-HSD3 inhibitors with similar or better activities compared to 1. Moreover, we
observed that either the side chain is attached to C2 or C3, the potency is almost retained, while
being slightly better when this substituent is linked to position C3. In this way 33 (C2
substituted, IC50 = 0.17 μM) was ~2-fold less potent than 1, whereas the amide 23 (C3
substituted, IC50 = 0.10 μM) was equipotent. Additionally, it is of interest to note that the
presence of an OH near to the aromatic A-ring as well as its orientation through the space is an
important factor. For example, compound 31, having an ethyl spacer between the A-ring and the
side chain, exhibited an IC of 0.13 μM. However, introducing an OH group on this ethyl spacer
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
5
0
resulted in an inhibitory activity comparable to that of 31, as depicted by compound 26 (IC =
5
0
0
.14 μM), while the simple inversion of the OH in 26 to give 27 (IC = 0.02 μM) led to the most
50
potent inhibitor of the series (4.5-fold more potent than 1).
For a more in-depth investigation the role of OH in 27, we prepared its fluorinated
bioisoster 28. As expected, this modification resulted in a significant loss of 17β-HSD3
inhibitory activity with an IC of 0.25 μM. Thus, OH is suspected to act as an HBD group with a
50
residue from an amino acid in the binding pocket of 17β-HSD3 and provide a beneficial
21
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 22 of 68
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
molecular interaction. Interestingly, the HBA carbonyl group of amide 23 is well-tolerated, even
if its potency is about 4 times weaker than 27.
We then evaluated the ability of these new analogs to stimulate the proliferation of
LAPC-4 cells. A non-significant proliferative effect was only observed at 0.1 μM for the isomer
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
2
6 (115%), but not for the most potent 17β-HSD3 inhibitor 27 (103.2%). On the other hand,
when LAPC-4 cells were treated with 23 and 31, we observed a strong reduction of cell
proliferation when compared to the control at 0.1 μM (60.0 and 51.9%, respectively) suggesting
an anti-proliferative action on this cell line potentially due to an anti-androgenic or cytotoxic
activity. Further research will thus be necessary to determine the main cause of this unexpected
antiproliferative activity. Interestingly, compound 33 did not display proliferation or anti-
proliferation of androgen-sensitive LAPC-4 cells.
Inhibition of rat testis 17β-HSD3. The more potent inhibitors on LNCaP transfected
cells, compounds 23, 27, 31, and 33, were also tested for the inhibition of 17β-HSD3 contained
1
4
in a microsomal fraction of homogenized rat testes. We thus measured the amount of [ C]-T
14
formed from natural substrate [ C]-4-dione in the presence of NADPH as cofactor at six
concentrations of inhibitor. The results were expressed as a percentage of the inhibitory activity,
and the resulting curves used to determine the IC values (Table 2). Interestingly, the same
5
0
results obtained with both sources of 17β-HSD3, human transfected cells and rat testis
preparation, suggests that it will be possible to use rat for preclinical model studies.
Metabolic stability in vivo. A series of enzyme inhibitors were selected to determine
their potential in the metabolic stability assay. Compounds 1, 4, 9, 16, 21, 23, 26, 27, 31, and 33
22
ACS Paragon Plus Environment

Page 23 of 68
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
were thus treated in the presence of NADPH as cofactor for 1 hour with the human liver S9
fraction. In this assay, a compound is transformed by phase-I reactions (oxidation, reduction, and
hydrolysis) and phase-II reactions (glucuronidation, sulfatation, and acetylation). The remaining
compound is measured at the end of the incubation period, and expressed in %. As depicted in
Figure 4, blocking C16 position by introducing a fluorine atom gave a better metabolic stability
as showed by compound 4 and similarly when converting the 17-C=O in 1 to a 17-oxime in 9.
The metabolic stability was also significantly improved with the presence of a 17-acyl group,
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
such as for C21-steroid derivative 16. As expected, the absence of 19-CH in 21 caused no
3
impact on the stability when compared to 1 (with a 19-CH ). Regarding the estra-1,3,5(10)-triene
3
series (compounds 23, 26, 27, 31, and 33) the order of metabolic stability was 23 > 33 > 26 = 27
=
31.
Figure 4. Human hepatic metabolic stability and plasma concentration in mice of lead 17β-HSD3
inhibitor 1 and selected analog inhibitors. Results represent the % of remaining quantity of compound
treated with the human liver S9 fraction (black columns) or the plasma concentration in ng/mL of a single
subcutaneous injection of an inhibitor in mice (white columns). Data are the average ± SD of four
23
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 24 of 68
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
experiments. ADT: androsterone; PRG: 3α-hydroxy-5α-pregnan-20-one; EST: estrane and X: 2,5-
dimethyl-4-((2-(trifluoromethyl)phenyl) sulfonyl)piperazinyl.
Results with D-ring derivatives suggest that the formation of metabolites can be blocked
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
or modulated by adding a group at C16 or C17. This conclusion is in agreement with a previous
report showing that 13-epi-RM-532-105, with an inverted D-ring, was more stable than 1 in a
2
7
microsomal preparation. For A-ring derivatives, compound 23 unexpectedly showed the best
metabolic stability despite lacking modifications on the D-ring. Its rigid and planar structure (due
to the amide at C3) may prevent, in a certain way, a favorable interaction with the cytochrome
P450 enzyme family. Compound 23 is thus an interesting derivative, since it has as good
enzymatic inhibition as lead candidate 1, with better metabolic stability and non-androgenic
activity on LAPC-4 as well.
Plasma concentration of inhibitors in mice. In addition to the metabolic stability
assessed in vitro (human liver S9 fraction), we also tested the stability of a selection of 17β-
HSD3 inhibitors in vivo (Figure 4). A single dose of an inhibitor was then injected
subcutaneously in mice and its plasma concentration determined at 3 h. Interestingly, the
selected new inhibitors reached higher plasma concentrations (48.7 – 108.2 ng/mL) than the lead
inhibitor 1 (27.7 ng/mL). The N-oxime D-ring derivative 9, which is the only representative of
the androstane-based series of inhibitors, is clearly more stable with a plasma concentration 3.4
fold higher (94.7 ng/mL) than its corresponding ketone 1. Although bearing all a C17-carbonyl
group, the A-ring derivatives 23, 27, 31, and 33 of the estrane series of inhibitors were found
more stable than androstane derivative 1. In fact, closely related compounds 27 (alcohol;
CH CHOH at C3) and 31 (alkyl; CH CH at C3) produced the same plasma concentrations (48.7
2
2
2
24
ACS Paragon Plus Environment

Page 25 of 68
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
and 51.6 ng/mL, respectively), whereas for 23 the presence of an amide (NCO) at C3 seems
limited the metabolism around this position which resulted in the higher plasma concentration
observed (108.2 ng/mL; 3.9 fold compound 1 value). The last estrane representative, compound
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
3
3, has a side chain that is moved from position 3 to 2 and a phenolic group. This estrone
derivative generated a plasma concentration of 89.3 ng/mL although the presence of both C17-
ketone and C3-OH.
Proliferative androgenic/estrogenic effect of inhibitors. A proliferative androgenic
activity is not suitable for a therapeutic agent targeting PCa; consequently, the cell proliferation
of androgen-dependent LAPC-4 cells was one of the criteria we used to discriminate our 17β-
HSD3 inhibitors. Easy to use, this assay is however not ideal because it is possible for a
molecule to be an androgen receptor agonist although it does increase the cell proliferation, as it
displays a cytotoxic activity by a hormone-independent mechanism of action. Thus, the lack of
androgenic effect should be addressed using other methods such as yeast-based reporter gene
assay,44 or AR transcriptional amplification system.26 Using this later transcriptional activity
assay, we thus previously observed no androgenic activity for the lead compound 1, but a
cytotoxic activity was observed at the higher concentration tested of 2 µM.26 However, a
compound that can reduce the proliferation of cancer cells is however interesting in itself and, for
this reason, we considered the proliferation of cancer cells as an assay providing interesting
information. Furthermore, since some of the new inhibitors are based on an estrane backbone,
instead of the androstane backbone of 1, we also addressed their potential estrogenic activity.
Based on their 17β-HSD3 inhibitory activity and bioavailability expressed vs the lead
25
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 26 of 68
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
compound 1, we selected compounds 9, 23, 27, 31, and 33 to evaluate their androgenic and
estrogenic effects (Table 3). Similarly to compound 1, except for estrone derivative 33, the
proliferation results on androgen-dependent LAPC-4 cells showed a dose-dependent reduction of
cell proliferation (increasing % of cell inhibition) suggesting an antiproliferative activity as for
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
2
9
RM-581, a cytotoxic agent. However, this effect does not seem to be mediated by the AR, since
these compounds did not bind this receptor as estimated on the basis of their negative Goldscore
values (Table 3). As a point of comparison, the potent androgen DHT, when docked on the AR
binding domain (PDB: 2PNU), produced a positive Goldscore value of 62. Using the androgen-
independent PC-3 cells, we next determined that all selected inhibitors, contrary to the cytotoxic
agent RM-581, did not affect cell proliferation (no significant % of cell inhibition) at the three
concentrations tested independently of the backbone (androstane or estrane). Finally, the binding
affinity of selected inhibitors for ER was also addressed by a docking study with the ER binding
domain (PDB: 1ERE). The negative Goldscore values generated for the selected inhibitors
(Table 3), contrary to the positive value of 53.7 generated for the potent estrogen estradiol,
suggest no estrogenic activity for these compounds although the presence of an estrane
backbone. Furthermore, compounds 31 and 33 did not stimulate or inhibited the cell proliferation
of estrogen-dependent breast cancer T-47D cells, thus supporting the docking results.
Table 3. 17-HSD3 inhibition, plasma concentration in mice, prostate cancer cell
proliferative/antiproliferative/inhibition activities, and AR/ER binding of selected
compounds 1, 9, 23, 27, 31 and 33
Cpd
17-HSD3
inhibition concentration
fold vs 1)a (fold vs 1)b
Plasma
LAPC-4
AR binding
PC-3
ER binding
+
-
(AR ) cell
(Goldscore)
(AR ) cell
(Goldscore)
d
e
(
Proliferation
Proliferation
26
ACS Paragon Plus Environment

Page 27 of 68
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
(
%)c
(%)c
1
9
1.0
0.41
0.90
4.5
1.0
3.4
3.9
1.8
1.9
3.2
--
96 / 48 / 33
-342
-338
-152
-95
91 / 94 / 89
[9 / 6 / 11]
-367
-424
-200
-292
[4 / 52 / 67]
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
108 / 59 / 44
90 / 94 / 96
[
-8 / 41 / 56]
[10 / 6 / 4]
23
60 / 46 / 37
93 /100 / 94
[
40 / 54 / 63]
[7 / 0 / 6]
27
31
33
103 / 68 / 56
-3 / 32 / 44]
92 / 94 / 95
[
[8 / 6 / 5]
0.69
0.53
--
52 / 20 / 21
-160
-153
--
99 / 95 / 99
-253f
[48 / 80 / 79]
[1 / 5 / 1]
-234f
100 / 93 / 94
96 / 95 / 93
[
0 / 7 / 6]
[4 / 5 / 7]
RM-581
cytotoxic)
81 / 38 / 31
97 / 80 / 61
--
(
[
19 / 62 / 69]
[3 / 20 / 39]
g
DHT
androgen)
--
--
--
--
190 (at 0.01
62
--
--
--
--
(
M)
E2
estrogen)
--
53.7
(
a
b
Ratio (fold more active): IC (lead compound 1) / IC (new compound). Ratio (fold
50
50
more active): plasma concentration (new compound) / plasma concentration (lead
c
compound 1).
Proliferation of androgen-dependent LAPC-4 cells or androgen-
independent PC-3 cells at 3 concentrations (0.1 µM / 0.5 µM / 1.0 µM). Data are
expressed in % of cell proliferation related to the control of untreated cells (fixed at
27
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 28 of 68
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
1
00%) and expressed as the % of inhibition of cell proliferation (between square
brackets). d Goldscore as an estimation of androgen-receptor (AR) binding affinity. e
f
Goldscore as an estimation of estrogen-receptor (ER) binding affinity. No proliferation
+
g
of T-47D (ER ) cells at 1.0 and 5 M. See reference 29.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
CONCLUSION
Several A- and D-ring steroid derivatives were designed and synthetized by systematic
modifications of the lead compound RM-532-105 (1). Substituents at C16 and C17 of the D-ring
generally reduced the enzymatic inhibitory effect on 17β-HSD3. The 17-substituted ADT
derivatives 7 and 9 were the most potent inhibitors, suggesting that the presence of HBA groups
as well as their orientation in space at this position is crucial for enzymatic inhibition.
Interestingly, adding a N-oxime at C17 position to generate 9 improved the metabolic stability in
vitro and in vivo. Nevertheless, it is worthwhile to mention that, in some cases (compounds 4
and 5), D-ring modifications yielded compounds stimulating the proliferation of LAPC-4 (AR+)
cells.
On the other hand, the inhibitory effect on 17β-HSD3 is almost retained by replacing the
saturated steroid A-ring of 1 with a benzene ring (ADT vs estrane backbone). This approach
allowed us to discover a series of estrane derivatives as new 17β-HSD3 inhibitors, not reported
to date. In addition to the ease of synthetizing these derivatives, the scaffold replacement yields
active compounds even though the optimized side chain is attached to the C2 or C3 position.
These analogs induced no cell proliferation, and produced a higher plasma concentration in mice
than 1. Thus, changing the ADT backbone by an estra-1,3,5(10)-triene backbone led to identify
four interesting 17β-HSD3 inhibitors (Table 3): 23 that is equally potent as 1 (0.9 fold), 27 as the
28
ACS Paragon Plus Environment

Page 29 of 68
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
most potent inhibitor (4.5 fold vs 1), and 31/33 just slightly less potent than 1 (0.69/0.53 fold).
Furthermore, while compounds 23, 27, and 31 disclosed an anti-proliferative effect against
LAPC-4 cells, 33 disclosed no stimulation or inhibition of cell proliferation. These four 17β-
HSD3 inhibitors also displayed a plasma concentration in mice 1.8 to 3.9-fold higher than lead
compound 1. In the light of these results, further research could be conducted to establish the
potential of 9, 23, 27, 31, and 33 to shrink or slow the growth of PCa tumors (mouse xenografts)
by lowering the intratumoral T and DHT or/and by exerting a direct antiproliferative (cytotoxic)
effect.
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
EXPERIMENTAL SECTION
General. Compound 1 was prepared according to the method described by our group.22
Epi-androsterone acetate and 3β-hydroxy-5α-pregnan-20-one were obtained from Steraloids
(Newport, RI, USA). The reagents for chemical synthesis were purchased from Sigma–Aldrich
Canada Ltd (Oakville, ON, Canada). The usual solvents were obtained from Fisher Scientific
(Montréal, QC, Canada) and were used as received. Anhydrous tetrahydrofuran (THF) and
anhydrous dichloromethane (DCM) were from Sigma–Aldrich. Thin-layer chromatography
(
TLC) and flash-column chromatography were performed on 0.20-mm Silica Gel 60 F254 plates
and with Silicycle R10030B 230–400 mesh silica gel (Québec, QC, Canada). Infrared spectra
IR) were recorded on a Horizon MB 3000 ABB FTIR spectrometer (Québec, QC, Canada), and
(
-1
only the significant bands were reported in cm . Samples were prepared as KBr pellets. Nuclear
1
13
magnetic resonance (NMR) spectra were recorded at 400 MHz for H and 100.6 MHz for C
with a Bruker Avance 400 digital spectrometer (Billerica, MA, USA). The chemical shifts (δ) are
29
ACS Paragon Plus Environment