Metabolic stability optimization and metabolite identification of 2,5-thiophene amide 17 β-hydroxysteroid dehydrogenase type 2 inhibitors

Article abstract of DOI:10.1016/j.ejmech.2014.09.061

17β-HSD2 is a promising new target for the treatment of osteoporosis. In this paper, a rational strategy to overcome the metabolic liability in the 2,5-thiophene amide class of 17β-HSD2 inhibitors is described, and the biological activity of the new inhibitors. Applying different strategies, as lowering the cLogP or modifying the structures of the molecules, compounds 27, 31 and 35 with strongly improved metabolic stability were obtained. For understanding biotransformation in the 2,5-thiophene amide class the main metabolic pathways of three properly selected compounds were elucidated.

Full text of DOI:10.1016/j.ejmech.2014.09.061

European Journal of Medicinal Chemistry 87 (2014) 203e219
Contents lists available at ScienceDirect
European Journal of Medicinal Chemistry
journal homepage: http://www.elsevier.com/locate/ejmech
Original article
Metabolic stability optimization and metabolite identification of 2,5-
thiophene amide 17 -hydroxysteroid dehydrogenase type 2 inhibitors
b
Emanuele M. Gargano a, Enrico Perspicace a, Nina Hanke c, Angelo Carotti d,
,
, b, *
Sandrine Marchais-Oberwinkler a **, Rolf W. Hartmann a
a Pharmaceutical and Medicinal Chemistry, Saarland University, Campus C23, D-66123 Saarbrücken, Germany
b Helmholtz Institute for Pharmaceutical Research Saarland (HIPS), Campus C23, D-66123 Saarbrücken, Germany
c ElexoPharm GmbH, Campus A12, D-66123 Saarbrücken, Germany
d
ꢀ
Dipartimento Farmaco-Chimico, Universita degli Studi di Bari, V. Orabona 4, I-70125 Bari, Italy
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 23 July 2014
Received in revised form
16 September 2014
Accepted 18 September 2014
Available online 20 September 2014
17
b
-HSD2 is a promising new target for the treatment of osteoporosis. In this paper, a rational strategy to
-HSD2 inhibitors is described,
overcome the metabolic liability in the 2,5-thiophene amide class of 17
b
and the biological activity of the new inhibitors. Applying different strategies, as lowering the cLogP or
modifying the structures of the molecules, compounds 27, 31 and 35 with strongly improved metabolic
stability were obtained. For understanding biotransformation in the 2,5-thiophene amide class the main
metabolic pathways of three properly selected compounds were elucidated.
Keywords:
© 2014 Elsevier Masson SAS. All rights reserved.
17b-HSD2
Osteoporosis
Metabolic stability
Metabolite identification
cLogP
Inhibitor design
1. Introduction
The available therapies today (bisphosphonates and selective
estrogen receptor modulators) lack of a satisfying profile in terms of
Osteoporosis [1] is a systemic disease in which an unmatched
activity of osteoclasts and osteoblasts leads to a decline in bone
density and quality. It is also correlated with an increased risk of
bone fracture. Osteoporosis is observed in 40% of postmenopausal
women [2]. Considering the progressive aging of the population
this number is expected to constantly increase in the near future [3]
and osteoporosis is therefore considered as a serious public health
concern.
safety and efficacy. Bisphosphonates, with alendronate as the
principal drug used in the treatment of osteoporosis, are effective in
both postmenopausal women [4] and men [5,6], though they are
able to reduce the risk of fracture by only 50% and are associated
with side effects such as osteonecrosis of the jaw. Raloxifene is,
among the selective estrogen receptor modulators (SERMs), one of
the most often used compounds for the treatment of osteoporosis
[7]. SERMs are efficient too, but, as bisphosphonates, endowed with
several side effects, such as an increased risk of venous thrombo-
embolism. As a consequence, the finding of new, safer and more
effective treatments for osteoporosis is of particular importance for
public health.
It has been proven that an important player in the maintenance
of bone health is the physiological, pre-menopausal concentration
of the estrogen estradiol (E2). That induces bone formation and
represses bone resorption by acting on the osteoblasts [8]. Estrogen
replacement therapy (ERT) was used in the past in postmenopausal
osteoporosis patients [9,10], successfully reducing the risk of frac-
tures but increasing the incidence of cardiovascular diseases and
breast cancer [10,11], which was the reason for the cessation of this
Abbreviations: 17
-hydroxysteroid dehydrogenase type 1; E1, estrone; E2, 17
testosterone; DHT, dihydrotestosterone; DHEA, dehydroepiandrosterone; A-dione,
4-androstene-3,17-dione; FC, flash chromatography; s.f., selectivity factor; RBA,
b
-HSD2, 17
b
-hydroxysteroid dehydrogenase type 2; 17
b-HSD1,
17
b
b-estradiol; T,
D
relative binding affinity; ER, estrogen receptor; ERT, estrogen replacement therapy;
inh, inhibition; MEP, molecular electrostatic potential; UDPGA, uridine diphosphate
glucuronic acid; PAPS, 30-phosphoadenosine-50-phosphosulfate.
* Corresponding author. Pharmaceutical and Medicinal Chemistry, Saarland
University, Campus C23, D-66123 Saarbrücken, Germany.
** Corresponding author.
E-mail addresses: s.marchais@mx.uni-saarland.de (S. Marchais-Oberwinkler),
rwh@mx.uni-saarland.de, rolf.hartmann@helmoltz-hzi.de (R.W. Hartmann).
http://dx.doi.org/10.1016/j.ejmech.2014.09.061
0223-5234/© 2014 Elsevier Masson SAS. All rights reserved.

204
E.M. Gargano et al. / European Journal of Medicinal Chemistry 87 (2014) 203e219
therapy. It has also been shown that testosterone (T) has beneficial
effects on bone formation [12,13].
Hence, an optimal strategy for the treatment of osteoporosis
should be that to inducing a local increase in E2 concentration in
bone tissue, without affecting the systemic levels of E2. Such an
In this report, we describe the analysis of the metabolic profile
of some representative in house inhibitors, the design of new, more
stable inhibitors based on the results from these studies, as well as
their synthesis and biological evaluation. In addition three in-
hibitors were investigated for their specific metabolic fate in order
to understand the particular biotransformation of this class of
molecules.
effect might be achieved by inhibiting 17
drogenase type 2 [14] (17 -HSD2), an enzyme that catalyzes the
conversion of the highly active 17 -hydroxysteroids into the low
b-hydroxysteroid dehy-
b
b
inactive 17-ketosteroids, that is the estrogen E2, and the androgen
testosterone (T) into their inactive oxidated forms estrone (E1) and
D
4-androstene-3,17-dione (A-dione), respectively (Chart 1). The
2. Inhibitor design
biological counterparts, 17
(17 -HSD1) and 17 -hydroxysteroid dehydrogenase type 3 (17
HSD3) catalyze the reverse reactions. Inhibitors of 17 -HSD1 are
b-hydroxysteroid dehydrogenase type 1
b
b
b
-
2.1. Stability towards human S9 fraction
b
potential drugs for the treatment of breast cancer and endometri-
osis and have recently been described in a series of papers by us
[15e18].
As 17b-HSD2 is expressed in osteoblastic cells [19e21], its in-
hibition should lead to the desired local increase of E2 and T levels
in bone tissue.
Different hepatic enzymes can be responsible for the metabolic
transformation of a compound and it frequently happens that small
structural changes on the molecule lead to a switch of the
responsible metabolic enzyme [33], a phenomenon known as
“metabolic switching” [33]. As a consequence introduction of
several structural changes in the scaffold of a compound is not an
appropriate method to evaluate liabilities of single moieties as
changes in the core structure could have switched the main site of
metabolism.
Ideally, the new 17
activity on the target enzyme, be selective over 17
should have no affinity for the estrogen receptors (ER)
b
-HSD2 inhibitors should, in addition to their
-HSD1 and
and in
b
a
b
order to avoid E2 related side effects.
Therefore a set of 11 selected compounds described by our
group [24,25] and carefully chosen as differing from one another by
one structural change only, were tested for their phase І and ІІ
metabolic stability using human liver S9 fraction (Table 1), in order
to find out the potential sites of metabolism.
The discovery of potent and selective nonsteroidal inhibitors of
-HSD2 has been already reported by our research group
[22e30].
As drugs encounter formidable challenges to their stability
17b
in vivo, it is not sufficient to only focus on potency in the devel-
opment process of a drug candidate. It is also important to profile
the metabolic stability of a lead series and improve it when this
From the comparison of compound 1 with 2 (t1/2 ¼ 17 min and
38 min, respectively) it appears that the methoxy group on the A
ring in meta position is unfavorable for metabolic stability
compared to the methyl group, whereas the presence of a meta
dimethyl amino group (compound 4, t1/2 ¼ 29 min) provides a
similar stability as the methylated derivative 2 (t1/2 ¼ 38 min).
The exchange of the methoxy group (5) for a methyl (6) on the C
ring results also in a slight increase in metabolic stability (t1/
constitutes a potential liability. As 17b-HSD2 should be inhibited
with constancy to locally increase the E2 levels and to maintain
them high over the time needed to re-establish bone metabolism,
the development of very metabolically stable inhibitors is partic-
ularly important. The enhancement of the metabolic stability of a
class of inhibitors is a very difficult task for the medicinal chemist
but could strongly favor the emergence of potential preclinical
candidates [31].
As a consequence, we decided to test three of our most prom-
ising, previously reported 17b-HSD2 inhibitors (compounds 1e3)
for their metabolic stability in human liver microsomes S9 fraction
(Chart 2), a powerful tool to explore both phase І and phase ІІ major
metabolic reactions [32]. 1, 2 and 3 represent very good compounds
¼ 12 min for 5 vs. 20 min for 6).
2
Compound 5 bears an additional fluorine in ortho position on
the A ring in comparison to compound 1. However introduction of
this fluorine atom does not improve the stability of compound 5 (5,
t1/2 ¼ 12 min vs. 17 min for 1).
Comparison of the unsubstituted inhibitor 7 (t1/2 ¼ 38 min),
with 8 (t1/2 ¼ 1 min) bearing a hydroxy group on the A ring and
with 9 (t1/2 ¼ 4 min) having an OH on the C ring shows that the
phenolic OH group enhances the metabolic degradation of this
class of compounds. This result was also observed in presence of a
methylene linker in compound 3, bearing an OH on the C ring and
having a half-life of 4 min. On the other hand, the nature of sub-
stituents on the A ring (in the presence of the OH group on the C
ring) plays an important role in the stabilization, as addition of the
methyl group alone (10, t1/2 ¼ 17 min) or a hydroxy group (11, t1/
in terms of enzymatic inhibitory potency and selectivity over 17b-
HSD1, however, they do not exhibit a satisfying stability displaying
short half-lifes between 4 and 38 min.
¼ 20 min) slightly increases the metabolic stability. Compounds 3,
2
10 and 11, in comparison with compound 1 have in their core two
additional potential metabolic sites: the methylene linker and the
phenolic group (liable for phase ІІ metabolic attack). Despite the
increase in liable metabolic sites their half-lives (t1/2 ¼ 17 min and
21 min, respectively) are still in the range of compound 1.
In summary none of the tested inhibitors showed a satisfying
metabolic stability profile (Table 1). However, from these results it
is obvious that 1. The methyl group on the A ring is less prone to
metabolic degradation than the methoxy group (compounds 1 and
2). 2. The phenolic hydroxy substituent on both the A ring or the C
ring is subject to fast metabolism (compounds 8 and 9), which can
be slightly slowed down by introduction of substituents on the
other ring (compounds 10 and 11).
Chart 1. Interconversion of estradiol (E2) to estrone (E1) by 17
b
-HSD2 and 17
b-HSD1,
and testosterone (T) to androstenedione (A-dione) by 17 -HSD2 and 17
b
b-HSD3.

E.M. Gargano et al. / European Journal of Medicinal Chemistry 87 (2014) 203e219
205
Chart 2. Previously described 17b-HSD2 inhibitors.
2.2. Design of inhibitors
tetrakis(triphenylphosphine)palladium and cesium carbonate in a
mixture DME/EtOH/H2O (1:1:1, 3 mL) as solvent and microwave
irradiation (150 ꢀC, 150 W for 20 min), afforded the desired 2,5-
thiophene derivatives 12e23, 25e27, 36a, 37a, 40a and the 2,5-
furane derivatives 31e34. Methoxy compounds 36a and 37a were
submitted to ether cleavage using boron trifluoride-dimethyl sul-
fide complex yielding the hydroxy compounds 36 and 37. The
aldehyde function of compound 40a was reduced to primary
alcohol using NaBH4 to provide compound 40. The synthesis of the
4-hydroxy-thiophene (compound 38) is also depicted in Scheme 1
and followed the same synthetic pathway starting from the 5-
bromo-4-methoxythiophene-2-carboxylic-acid 38c. The methoxy
group in position 4 of the thiophene (compound 38a) was cleaved
using boron trifluoride-dimethyl sulfide complex and yielded the
hydroxy compound 38.
The synthesis of compound 30 was performed following the
method depicted in Scheme 2. The intermediate 30b was synthe-
sized starting from 5-bromo-1,3-thiazole 30c applying a Suzu-
kieMiyaura cross-coupling reaction with m-tolyl boronic acid,
using microwave irradiation. The acidic hydrogen in position 2 of
the thiazole 30b was easily removed using n-butyl lithium. The
anion obtained reacted immediately with a dry flow of carbon di-
oxide leading to the corresponding carboxylate 30a which was
subsequently activated using oxalyl chloride and reacted with 3-
methyl-N-methylaniline to afford compound 30.
In the inhibitor design we took into account the information we
had gained from the metabolic stability results. In addition we tried
to modify some other moieties that have the potential of being
metabolically labile. Doing this we kept in mind the structure-
activity and structureeselectivity relationships obtained in our
previous studies [24,25].
As the methyl group resulted to be more stable than the
methoxy group on the A and C ring, we decided to exchange each
methoxy group with a methyl function and then to explore the
effect of this substitution pattern on both activity and metabolic
stability. It was also observed that most of the compounds with a
hydroxy group in the presence of the methylene linker, known to be
a critical point for metabolism, scored the worst results in terms of
metabolic stability. However since a hydroxy group is needed for
activity [25] in the series of compounds having the methylene
linker, we decided to stop further structural modifications on this
series. Therefore compounds without linker with the general
structure A, depicted in Chart 3 were synthesized.
The thiophene is also described as a potential site of metabolism
[34]. Therefore, we tried to modify it by bioisosteric replacements
using benzene, furane and thiazole (general structure B).
Moreover, it has been reported [35] that the lipophilicity pa-
rameters like cLogP could play an important role for the metabolic
stability. Very often the simple strategy to lower the cLogP was
appropriate to improve the metabolic stability of a compound.
Therefore some inhibitors, modified at the A ring with a moiety
reducing cLogP were designed (general structure C, Chart 3).
Furthermore, we profiled the metabolic pathway of the newly
synthesized dimethylphenyl thiophene 12 (Chart 4) looking for the
major metabolites resulting from the mono-oxidation of the phe-
nylthiophene part (Chart 4).
In order to identify the main metabolites, especially the one
coming from mono-oxidation of the phenylthiophene fragment, we
synthesized different hydroxylated derivatives of 12 bearing one
additional oxygen in the positions that are likely to be attacked by
the metabolizing enzyme (represented by the general structures D
and E, Chart 4).
The 1,4-disubstituted phenyl derivative 35 was obtained
following a two-step procedure (Scheme 3). First, amidation was
carried out by reaction of 4-bromobenzoyl chloride 35b with N,3-
dimethylaniline 35c (Method A) providing the brominated inter-
mediate 35a. Suzuki coupling (Method B) afforded the final product
35.
The synthesis of the 3-hydroxy-N-methyl-N,5-di-m-tolylth-
iophene-2-carboxamide (compound 39) is depicted in Scheme 4.
The ethyl 3-hydroxy-5-(m-tolyl)thiophene-2-carboxylate (com-
pound 39d) was obtained through the Fiesselmann thiophene
synthesis [36] by reaction of the
b-ketoester 39e and the ethyl
thioglycolate 39f, according to the described procedure [36]. The
hydroxy group of compound 39d was protected using MeI to afford
compound 39c in a moderate yield. The saponification of the ester
function of 39c was carried out using potassium hydroxide in wa-
ter/THF (1:1) mixture, affording the corresponding carboxylic acid
39b. Amidation was performed by reaction of 39b with N,3-
dimethylaniline, using the standard condition (method A), lead-
ing to the methoxy compound 39a. The methoxy group of 39a was
cleaved, using boron trifluoride-dimethyl sulfide complex and
yielded the hydroxy derivative 39.
3. Results
3.1. Chemistry
The synthesis of the 2,5-thiophene derivatives (compounds
12e23, 25e27, 36e38 and 40) and the 2,5-furane derivatives
(compounds 31e34) is depicted in Scheme 1. The 5-
bromothiophene-2-carbonic acid chloride and the 5-bromofuran-
2-carbonic acid chloride were obtained from the corresponding
carboxylic acids 36c or 31b by reaction with SOCl2 and subse-
quently reacted with different anilines (Method A), providing the
intermediates 12ae23a, 25ae27a, 36b, 37b, 40b and 31ae34a in
good yields. Subsequently, Suzuki coupling (Method B) using
3.2. Biological
3.2.1. Inhibition of human 17
17 -HSD1 in cell-free assays
The synthesized compounds were tested for their ability to
inhibit 17 -HSD2 and 17 -HSD1 using enzymes from placental
b-HSD2 and selectivity over human
b
b
b

206
E.M. Gargano et al. / European Journal of Medicinal Chemistry 87 (2014) 203e219
Table 1
3.2.1.1. Influence of lipophilic substituents on rings A and C. In the
2,5-thiophene class the methoxy groups in rings A and C can be
replaced by methyl groups, without significantly affecting 17
HSD2 inhibitory potency and selectivity over 17 -HSD1, as it be-
Half-life in human liver microsomes S9 fraction.
b
-
b
comes apparent from comparing 1, 2 and 12 (IC50 ¼ 58 nM,
s.f. ¼ 116; 68 nM, s.f. ¼ 112; 52 nM, s.f. ¼ 83, respectively). This
suggests that the methoxy groups on both rings do not function as
H bond acceptors. Most likely the methyl substituents, being either
part of a methoxy group or being directly connected to rings A and
C, form important lipophilic interactions with the enzyme. This
assumption is further supported by comparison of 12 with the
corresponding compound without substituents on rings A and B, 24
Compd.
R1
R2
n
Inhibitor t1/2 (min)
1
2
3
4
5
6
7
8
m-OMe
m-Me
eOMe
eOMe
eOH
eOMe
eOMe
eMe
eH
0
0
1
0
0
0
0
0
0
1
1
17a
38a
4b,c
29a
12b,c
20a
38b,c
1b,c
4b,c
17a
21a
o-F, m-OMe
m-N(Me)2
o-F, m-OMe
o-F, m-OMe
eH
m-OH
eH
m-Me
o-OH
[25], which is only moderately active (percent inhibition at 1
48%).
mM,
Encouraged by this result, nine compounds including all
possible combinations of methyl substitutions on ring A and C were
synthesized (compounds 12e20), in order to investigate the in-
fluence of the methyl group position on the biological profile in this
class of compounds. The substitution pattern indeed appears to be
eH
9
10
11
eOH
eOH
eOH
a
Inhibitor tested at a final concentration of 3 mM, 1 mg/ml, pooled human liver S9
critical for both the 17
over 17 -HSD1 as indicated by the wide range of inhibitory activ-
ities displayed by compounds 12e20 (Table 2).
The best 17 -HSD2 inhibitory potency associated with the best
selectivity over 17 -HSD1 was achieved for compounds with the C
b-HSD2 inhibitory potency and the selectivity
fraction (IVT, Xenotech or TCS Cellworks), 1 mM NADPH regenerating system,
b
0.75 mM UDPGA, 0.05 mM PAPS, incubated at 37 ꢀC for 0, 5, 15, 30 and 45 min.
b
Inhibitor tested at a final concentration of 1 mM, 1 mg/ml pooled mammalian
b
liver S9 fraction (BD Gentest), 2 mM NADPH regenerating system, 1 mM UDPGA and
0.1 mM PAPS, incubated at 37 ꢀC for 0, 5, 15 and 60 min.
b
c
Mean of at least two determinations, standard deviation less than 25%.
Chart 3. Designed inhibitors.
ring bearing a methyl in meta position while on ring A the methyl
can be in either in meta or para position (compounds 12 and 13,
IC50 ¼ 52 nM and 58 nM; s.f. ¼ 83 and 66, respectively). The
selectivity factor is slightly decreased in comparison to the refer-
ence compounds 1 and 2. Compound 14 with the methyl on the C
ring in meta position and the methyl on the A ring in ortho position
shows significantly decreased 17b-HSD2 inhibitory potency
(IC50 ¼ 715, s.f. ¼ 7), indicating that the geometry of the A ring is
critical for activity.
Substitution of the C ring with a methyl in ortho position,
together with the A ring being either para or meta methyl
Chart 4. Structures of the most probable metabolites of 12.
substituted, leads to a slight decrease in 17
b-HSD2 inhibitory po-
tency and to an increase in 17 -HSD1 inhibition, providing com-
b
pounds with a poor selectivity factor (compounds 15 and 16,
IC50 ¼ 380 nM and 225 nM and s.f. ¼ 3 and 3, respectively). If both A
source according to described methods [37e39]. Inhibitory activ-
ities of compounds 12e40 are shown in Tables 2e5, either as IC50
and C rings are ortho substituted, 17b-HSD2 and 17b-HSD1 activity
values or as percent of inhibition values determined at 1
mM.
are strongly decreased (compound 17, 14% inhibition and no inhi-
bition for 17 -HSD2 and 17 -HSD1 at 1 M, respectively).
Compounds showing less than 10% inhibition at this concentration
were considered to be inactive.
b
b
m

E.M. Gargano et al. / European Journal of Medicinal Chemistry 87 (2014) 203e219
207
Scheme 1. Synthesis of 2,5-thiophene derivatives 12e23, 25e27, 36e38 and 40 and 2,5-furane derivatives 31e34.a
Substitution of the C ring with a methyl group in para position in
combination with a meta or para substituted A ring led to com-
and 2 (Table 2). However, the values are still below 5, which should
be sufficient for an acceptable bioavailability [41]. Since lipophilic
compounds are usually more susceptible to phase І metabolism
[35], in a second step we tried to lower cLogP of the 2,5-thiophene
derivatives by exchanging the methylphenyl A ring for different
heterocycles like 2-methylpyridinyl 25, 3-methylpyridininyl 26 or
1-methyl-1H-pyrazolyl 27 (Table 3).
pounds with moderately decreased 17b-HSD2 inhibition and
decreased selectivity over 17 -HSD1 (compounds 18 and 19,
b
IC50 ¼ 892 nM and 772 nM, s.f. ¼ 3 and 2, respectively). Ortho
substitution of the A ring coupled with para substitution of the C
ring, surprisingly leads to an inversion of selectivity (compound 20,
26% and 50% inhibition of 17
respectively).
b
-HSD2 and 17
b
-HSD1 at 1
m
M,
The calculated LogP values of compounds 25, 26 and 27 are
lower and therefore advantageous for their bioavailability
compared to the one of compound 12 (3.38, 3.45 and 2.46 for 25, 26
and 27, respectively, 4.77 for 12, Tables 2 and 3).
However the exchange of the A ring with these heterocycles led
to a strong decrease in activity (compounds 25, 26 and 27: 18%, 28%
and 16% inhibition at 1 mM, respectively). Once again the change in
the MEP of the A ring might be responsible for this decreased
activity.
In the 2,5-thiophene class the highest inhibition is observed for
a compound showing an m,p-dimethyl substitution on the A ring
and a meta methyl substitution on the C ring (compound 21,
IC50 ¼ 33 nM, s.f. ¼ 352), indicating that the positive effect on the
activity of the two methyl groups on the A ring is additive.
Furthermore, the presence of the two methyl groups increases the
selectivity over 17b-HSD1. Compound 21 turns out to be the most
active and selective compound in this class.
As the methyl substituents on rings A and C in the 2,5-thiophene
The exchange of the methyl in meta position on either the A or
class were discovered to be very well tolerated by 17
especially in meta position, and to induce a good selectivity over
17 -HSD1 they will be maintained for the subsequent
b-HSD2,
the C ring with a chlorine decreases both 17
b-HSD2 inhibitory
potency and selectivity over 17 -HSD1 (compounds 22 and 23,
b
b
IC50 ¼ 327 nM and 474 nM, s.f. ¼ 29 and 5, respectively). The
modifications.
importance of the molecular electrostatic potential (MEP) for the
ability of 17
stacking interaction with aromatic amino acids from the active site
has been recently discussed [40]. The decrease in the 17 -HSD2
inhibition of compounds 22 and 23 can be explained by the dif-
ference in the MEPs of the two aromatic rings, elicited by the ex-
change of the methyl group with a chlorine. This exchange on the A
b-HSD2 and 17b-HSD1 inhibitors to establish a p-
3.2.1.2. Influence of the central core. In a previous study [42], the
central 2,5-thiophene-carboxamide of 2 was replaced by a thiazole-
2-carboxamide and by a thiazole-5-carboxamide (compounds 28
and 29, IC50 values 296 nM and 621 nM, s.f. ¼ 63 and 8, respectively,
Table 4). Comparison of the potency of 28 and 29, indicates that the
b
thiazole-2-carboxamide is better in terms of 17
b-HSD2 inhibition
ring seems to be favorable for the 17
b
-HSD1 inhibitory potency
and selectivity over 17 -HSD1 in presence of a methoxy group in
b
(compound 23, IC50 ¼ 2144 nM for 17
b-HSD1).
position 3 on the C ring. As we had experienced that in the 2,5-
thiophene class the methoxy group can be replaced by a methyl
(compound 12) we decided to synthesize the thiazole-2-
The introduction of lipophilic substituents on ring A and C in-
creases cLogP values for compounds 12e23 slightly compared to 1

208
E.M. Gargano et al. / European Journal of Medicinal Chemistry 87 (2014) 203e219
Scheme 2. Synthesis of compound 30.a
Scheme 3. Synthesis of compound 35.a
Scheme 4. Synthesis of compound 39.a
carboxamide with a methyl group on the A and C ring (compound
30, IC50 ¼ 499 nM, s.f. ¼ 51). This structure modification leads to a
2-fold decrease in activity compared to compound 28, indicating
that as observed in the 2,5-thiophene class, no crucial interaction is
achieved by the oxygen atom of the methoxy group.
compounds 31e34. The strongest inhibition of 17
b-HSD2 was
achieved by compound 34 (38% inhibition at 1 M), demonstrating
m
that the furane ring is not a suitable bioisostere for the thiophene in
these compounds, in contrast to observations made in other com-
pound classes [43]. The furane is less aromatic than the thiophene
(electron rich system) because the electronegativity of the oxygen
renders the electron pair on this atom less available for resonance.
In addition the introduction of the methyl substituent in the
thiazole-2-carboxamides did not sufficiently improve 17b-HSD2
inhibitory potency to make them superior to the 2,5-thiophene
amides as already described by us [42].
The replacement of the 2,5-thiophene for a 2,5-furane ring is
detrimental for the 17b-HSD2 inhibitory potency as shown by
This makes the furane less suitable for p-stacking interactions. The
electrostatic potential of the thiophene and furane derivatives are
therefore different which might explain the decrease in the activity
associated with the furanes.

E.M. Gargano et al. / European Journal of Medicinal Chemistry 87 (2014) 203e219
209
Table 2
This effect is also observed with the introduction of a second
substituent in meta position of ring A in compound 12 (Table 5). In
fact compound 37a, bearing an additional methoxy group and
compound 37, with an additional hydroxy group in meta position,
Inhibition of human 17b-HSD2 and 17b-HSD1 by biphenyl-2,5-thiophene amide
derivatives 12e24 with different substitution patterns on the A and C rings in cell-
free systems.
show a similar inhibitory potency over 17
b-HSD2 but lower than
the one of the monomethyl 12 (56% and 53% inhibition at 1
mM,
respectively) or the monomethoxy 1 (IC50 value 58 nM). This result
indicates that no additive effect is obtained by addition of both
methoxy and methyl group in meta position, perhaps due to a steric
hindrance.
The introduction of a methoxy group or a hydroxy group in
position 4 of the thiophene resulting in compounds 38a and 38
Compd. R1
R2
Inhibition
Selectivity factord cLogPe
IC50 (nM)a or % of
inhibition at 1 m
Ma
(59% and 64% inhibition at 1 mM, respectively) is detrimental for
activity. This might be due to conformational effects on the phenyl
A ring, steric hindrance in the binding site or repulsion of the polar
groups caused by the lipophilic environment in the binding cavity.
On the other hand, a methoxy group in position 3 of the thio-
phene leading to compound 39a (Table 5), decreases activity only
slightly in comparison to 12 (IC50 ¼ 127 nM and 52 nM, respec-
tively) whereas selectivity is increased (s.f. ¼ 132 and 83,
respectively).
17
b
-
17
b-
HSD2b
HSD1c
1
2
m-OMe
m-
OMe
m-
58 nM
6728 nM 116
7616 nM 112
3.68
4.38
m-Me
68 nM
OMe
12
13
14
15
16
17
18
19
20
21
22
23
24
m-Me
p-Me
o-Me
m-Me
p-Me
o-Me
m-Me
p-Me
o-Me
m-Me 52 nM
m-Me 58 nM
m-Me 715 nM
o-Me 380 nM
o-Me 225 nM
o-Me 14%
p-Me 892 nM
p-Me 772 nM
p-Me 27%
4306 nM 83
3825 nM 66
4570 nM
1177 nM
698 nM
ni
2687 nM
1562 nM
50%
11,576 nM 352
9412 nM 29
4.77
4.77
4.77
4.77
4.77
4.77
4.77
4.77
4.77
5.23
5.03
4.87
nd
7
3
3
nd
3
Exchanging this methoxy by a hydroxy group (compound 39)
leads to a decreased activity (67% inhibition at 1 mM, Table 5). This
indicates that in position 3 of the thiophene a lipophilic group is
tolerated whereas a hydrophilic substituent, such as a hydroxy, is
detrimental for activity.
Exchange of the methyl group on ring A of compound 12 by an
aldehyde (compound 40a) or by a CH2OH group (40) decreases
2
nd
m-Me, p-Me m-Me 33 nM
m-Me
m-Cl
H
m-Cl
m-Me 474 nM
48%
327 nM
2144 nM
14%
5
nd
inhibitory potency (62% and 55% inhibition at 1
Table 5).
mM, respectively,
H
ni: no inhibition, nd: not determined.
a
Mean value of at least two determinations, standard deviation less than 25%.
Human placental, microsomal fraction, substrate E2, 500 nM, cofactor NADþ
,
b
3.3. Affinity for the estrogen receptors
1500
m
M.
Human placental, cytosolic fraction, substrate E1, 500 nM, cofactor NADH,
M.
IC50(17
c
17
b
-HSD2 inhibitors should have no affinity for the estrogen
and in order to avoid potential ER mediated side
500
m
d
receptors (ERs)
a
b
b-HSD1)/IC50(17
b-HSD2).
e
effects. Compounds 12, 13, 21, 22, 30 and 35 have been tested ac-
cording to described methods [44] and the results have been
expressed in terms of percent [3H]-E2 displaced from the receptor
Calculated data.
Finally the replacement of the central ring by a 1,4-diphenyl
substitution leads to compound 35, with an IC50 value of 1126 nM
and an s.f. of 10. The exchange of the thiophene with a benzene
by the inhibitor (Table 6). Shortly, ERa or ERb were incubated with
[3H]-E2 and test compound. Subsequently the percent [3H]-E2
displaced from the receptor by the inhibitors was determined,
which were tested at concentrations 1000 times higher than the
one of [3H]-E2 (Table 6). All inhibitors tested were able to displace
less than 50% of [3H]-E2 from the corresponding receptor at this
concentration. In terms of Relative Binding Affinity (RBA), their RBA
decreases 17
son with the furane series, in the benzenes some selectivity over
17 -HSD1 is sustained (Table 4).
b-HSD2 inhibition moderately. However, in compari-
b
The exchange of the central thiophene ring of 12 (cLogP ¼ 4.77)
by thiazole (30, cLogP ¼ 4.58) and by furane (31e34, cLogP ¼ 3.86,
3.17, 3.62 and 4.32, respectively) lowers the cLogP value, whereas it
is slightly increased in case of the benzene derivative 35;
(cLogP ¼ 4.93, Table 4). However, all compounds show cLogP values
in range, which is acceptable for a good bioavailability according to
literature [41].
values are <0.1% for both ER
affinity.
a and b, indicating a low binding
3.4. Metabolic stability in human S9 fraction
We evaluated the most promising inhibitors for metabolic sta-
bility using human liver S9 fraction according to described methods
[45e47]. Briefly, incubations were run with the S9 fraction, co-
factors (for phase І and ІІ reaction) and inhibitor. The concentration
of the remaining test compound at the different time points was
3.2.2. Hydroxy derivatives of 12 and their precursors
Oxidation often occurs during metabolism and in some cases
leads to highly active metabolites. Several oxidation products
derived from 12 were synthesized and evaluated for their biological
activity in vitro (Table 5).
analyzed by LC-MS/MS and the results expressed as half-lifes (t1/2
)
and reported in Table 7.
Introduction of an additional methoxy group in para position of
compound 12 (ortho to methyl) leads to compound 36a (Table 5)
Exchange of the two methoxy of compound 1 (t1/2 ¼ 17 min) by
two methyl groups provides compound 12 (t1/2 ¼ 19 min), which is
also rather metabolically labile (Table 7). Changing the substitution
pattern of the methyl functions does not influence metabolic sta-
bility, as it results from comparison of meta-, meta-12, para-, meta-
13 and para-, ortho-16 (t1/2 ¼ 19 min, 15 min and 22 min,
respectively).
showing a significantly decreased activity (66% inhibition at 1
A similar loss in activity is observed for the hydroxylated com-
pound 36 (62% inhibition at 1 M), indicating that in this position
mM).
m
introduction of groups like a methoxy or hydroxy is not well
tolerated.

210
E.M. Gargano et al. / European Journal of Medicinal Chemistry 87 (2014) 203e219
Table 3
Compound 21 (t1/2 ¼ 20 min), bearing an additional methyl in
para-position on ring A, in comparison to 12 does not show an
improved stability (21 t1/2 ¼ 20 min vs. 12 t1/2 ¼ 19 min).
2,5-Thiophene amides N-containing heterocycle as A ring 25e27. Inhibition of hu-
man 17 -HSD2 and 17 -HSD1.
b
b
Compound 22, with a chlorine on ring C, shows a slightly
increased half-life (t1/2 ¼ 32 min) in comparison to 12 (t1/
¼ 19 min).
2
In addition compound 41, previously described [25], and bearing
a free carboxamide moiety was also tested for stability. In com-
parison to the others thiophene derivatives and more especially to
the methylated analog 5, it is exceptionally stable (t1/2 ¼ 117 min). It
is not clear whether the methyl group of the NeCH3 amide is the
metabolic reactive site or whether its absence stabilizes another
site in the molecule.
Compd.
Ring A
% Inhibition at 1
17
-HSD2b
m
Ma
cLogPd
3.38
b
17b
-HSD1c
25
18%
n.i.
Replacement of the thiophene by a furane (compound 31, t1/
¼ 92 min) or by a benzene (compound 35, t1/2 ꢁ 120 min) results
2
in strong improvement of metabolic stability. On the contrary, ex-
change of the thiophene by a thiazole leads to very labile com-
pound (30, t1/2 ¼ 5 min). From these results, it is likely that the
thiophene plays a major role in the metabolic fate in the 2,5-
thiophene class of compounds.
26
27
28%
16%
n.i.
n.i.
3.45
2.46
The strategy of lowering the cLogP of compound 12
(cLogP ¼ 4.77), in order to improve the metabolic stability, led to
the pyrazole 27 (cLogP ¼ 2.46) and the pyridine 25 (cLogP ¼ 3.38).
As expected the most hydrophilic compound 27 (t1/2 ¼ 82 min), is
more stable than compounds 25 (t1/2 ¼ 21 min), and 12 (t1/
ni: no inhibition.
a
Mean value of at least two determinations, standard deviation less than 25%.
b
Human placental, microsomal fraction, substrate E2, 500 nM, cofactor NADþ
,
1500
m
M.
Human placental, cytosolic fraction, substrate E1, 500 nM, cofactor NADH,
M.
Calculated data.
c
¼ 19 min).
2
500
m
d
3.5. Metabolites identification
Three compounds (1, 2 and 12) were investigated regarding
their metabolic fate and tested in the human S9 fraction using a
similar procedure as already described for determination of the
metabolic stability. Inhibitors were incubated at a concentration of
10
mM inhibitors without phase ІІ cofactors for three hours (no
Table 4
Exchange of the central core B. Inhibition of human 17
b-HSD2 and 17b-HSD1.
phase II metabolism is expected). After extraction the samples were
analyzed using a high resolution LC-MS (LTQ/Orbitrap). Potential
metabolites were identified by searching for expected bio-
transformations (Table 8). The results are presented as area per-
centage relative to the remaining parent compound. In some cases
two different peaks were found for one biotransformation, due to
its occurrence at two different sites of the molecule.
For the most prominent metabolites, we propose a structure
based on comparison of the fragmentation pattern of the parent
compound and the ones of the metabolites (Fig. 1). The spectrum of
1 (Fig.1(a)) contains strong signals at m/z 136 and 164, representing
the substituted aniline ring without or with the amide function,
respectively. It also contains the signal m/z 217 representing the left
part of the molecule after cleavage of the amide function. This
fragmentation is observed in all the tested 2,5-thiophene amide
inhibitors and their metabolites. The main metabolite of compound
1 is a demethylated product that accounts for 27% of the remaining
parent compound (Table 8). The spectrum of this demethylation
product (Fig. 1(b)) still contains the signals m/z 136 and 164, indi-
cating that no demethylation occurred either on the C ring or on the
amide. A new signal at m/z 203, coming from the left part of the
molecule, indicates that the cleavage of the methoxy occurred on
the A ring. In addition a second demethylation product was also
observed representing 9% of the remaining parent compound as
well as an þO product (12%, Table 8). Other biotransformations
result in negligible amounts (Table 8). The exact structure of these
metabolites was not identified.
Compd. X Y Z R1
R2
Inhibition
Selectivity
factord
cLogPe
IC50 (nM)a or % of
inhibition at 1
m
Ma
17 17b-
b
-
HSD2b HSD1c
2
H H S m-Me
eOMe 68 nM
eMe 52 nM
eOMe 296 nM 18,663 nM 63
eOMe 621 nM 4821 nM
eMe 499 nM 25,302 nM 51
eMe 14%
eMe ni
eMe 34%
7616 nM 112
4306 nM 83
3.68
4.77
4.19
3.91
4.58
3.86
3.17
3.62
12
28
29
30
31
32
33
H H S m-Me
H N S m-Me
N H S m-Me
H N S m-Me
H H O m-Me
H H O m-OMe
H H O o-F, m-
OMe
8
31%
16%
26%
nd
nd
nd
34
35
H H O m-Me, p- eMe 38%
Me
e e e e
46%
nd
10
4.32
4.93
e
1126
11,541
ni: no inhibition, nd: not determined.
a
Mean value of at least two determinations, standard deviation less than 25%.
b
Human placental, microsomal fraction, substrate E2, 500 nM, cofactor NADþ
,
1500
m
M.
Human placental, cytosolic fraction, substrate E1, 500 nM, cofactor NADH,
M.
IC50(17
In case of 2 (Fig.1(c) and (d)) the metabolic demethylation (7% of
the remaining parent compound) was less intense compared to the
methoxy compound 1, whereas the þ16 biotransformation
(Table 8) results in two different products, accounting for 14% and
c
500
m
d
b
-HSD1)/IC50(17
b-HSD2).
e
Calculated data.

E.M. Gargano et al. / European Journal of Medicinal Chemistry 87 (2014) 203e219
211
Table 5
Inhibition of human 17
b
-HSD2 and 17
b
-HSD1 by the possible major metabolite of 12 and its precursors in cell-free systems.
Compd.
Y
Z
R1
R2
Inhibition
Selectivity factord
cLogPe
IC50 (nM)a or % of inhibition at
1
m
Ma
-HSD2b
17
b
17b
-HSD1c
12
36a
36
37a
37
38a
38
39a
39
40a
40
eH
eH
eH
eH
eH
eOMe
eOH
eH
eH
eH
eH
eH
eH
eH
eH
eOMe
eOH
eH
m-Me
m-Me
m-Me
m-Me
m-Me
m-Me
m-Me
m-Me
m-Me
m-Me
m-Me
m-Me
52 nM
66%
62%
56%
53%
59%
64%
127 nM
67%
62%
4306 nM
11%
22%
28%
13%
ni
ni
16,732 nM
24%
ni
83
nd
nd
nd
nd
nd
nd
132
nd
nd
nd
4.77
4.45
3.93
4.43
3.94
4.47
4.00
4.36
4.77
3.84
3.39
m-Me, p-OMe
m-Me, p-OH
m-Me, m-OMe
m-Me, m-OH
m-Me
m-Me
m-Me
m-Me
m-CHO
eH
eH
eH
eH
m-CH2OH
55%
ni
ni: no inhibition, nd: not determined.
a
Mean value of at least two determinations, standard deviation less than 25%.
b
c
Human placental, microsomal fraction, substrate E2, 500 nM, cofactor NADþ, 1500
m
M.
Human placental, cytosolic fraction, substrate E1, 500 nM, cofactor NADH, 500 mM.
d
e
IC50(17
b-HSD1)/IC50(17b-HSD2).
Calculated data.
5% of the remaining parent compound. Comparing the fragmen-
tation pattern of 2 (Fig. 1(c)) and its main oxidation product
(Fig. 1(d)) it can be concluded that the oxygen addition occurs on
the thiophene phenyl A ring part of the molecule highlighted with
the dotted line (Fig. 1(d)). The structure of the metabolite was not
elucidated. Also in this case, the other biotransformations were
negligible.
For compound 12, the main metabolic product derives from an
oxidation (Table 8). Two different peaks for the þ16 biotransfor-
mation can be observed (Table 8), one representing 20% of the
remaining parent compound and the other 4%. Demethylation
(10%), and oxidation of the eCH3 into carboxylic acid (5%) are also
present. For 12, the demethylation alone can only occur at the
exposed to light at room temperature, able to even accelerate the S-
oxide degradation [48], it was concluded that the biotransforma-
tion product observed is unlikely to be the thiophene S-oxide.
The other 5 potential metabolites, depicted in Fig. 2, were syn-
thesized and co-injected in the LC-MS/MS together with a solution
of the metabolites of 12 (Fig. 3).
The chromatograms of the hydroxylated metabolites of 12 alone
(Fig. 3(a)) and the one of the co-injection with compound 40
(Fig. 3(b)) are identical with a main peak observed with a retention
time of 5.82 min and a minor peak at 6.07 min. The co-injection of
any other potential metabolite synthesized (36e39) produces
spectra with additional peaks having different retention time in
comparison to the original spectrum: 36 (Fig. 3(c)), 38 (Fig. 3(d)), 37
(Fig. 3(e)) and 39 (Fig. 3(f)) produce new peaks at 6.25 min,
6.63 min, 6.27 min and 8.03 min, respectively (Fig. 3). This exper-
iment demonstrates the identity of the hydroxylated metabolite of
12 with the hydroxy methyl compound 40.
amide
function.
The
transformation
demethylation þ hydroxylation and the di-hydroxylation were also
detected for this compound (Table 8).
Comparing the fragmentation pattern of 12 (Fig. 1(e)) and its
main oxidation product (Fig. 1(f)), it can be concluded that the
oxygen incorporation occurs on the thiophene phenyl part of the
molecule highlighted with the dotted line. Comparing the frag-
mentation pattern of the eCH3 to eCO2H biotransformation in the
same way (Fig. 1(g)), it becomes apparent that the carboxylic acid
group is on the A ring.
Since we were interested in the identification of the exact hy-
droxylation site for 12 and since the fragmentation pattern using
only the MS/MS data was not conclusive, we synthesized the most
probable metabolites. They are represented in Fig. 2.
Compound 36e40 are derived from compound 12 with addition
of one hydroxy group in different positions (Fig. 2). The mass þ16
(Table 8) observed in our experiment could in theory come from an
S-oxidation of the thiophene. The S-oxide thiophene is known to be
chemically highly unstable [48] and after synthesis it was rapidly
degraded (data not shown). In addition since the procedures for
metabolite synthesis and identification require the samples to be
To confirm that 40 is the main metabolite of 12 we compared the
fragmentation pattern of the metabolite and the synthesized
compound (Fig. 4).
As displayed in Fig. 4(a) and (b), the two fragmentation pattern
perfectly match. Prominent signals at m/z 120, 148, 173 and 217 are
present in both spectra. Interestingly, the signal at m/z 173 is
observed only in the MS2 spectra of 40 and in none of the frag-
mentation spectra produced for the other potential metabolites
(data not shown). This signal is likely to represent the fragment
depicted in Fig. 5, derived from the loss of the hydroxy group on the
methylene. Since the other potential metabolites all bear an OH
directly linked to an aromatic ring (Fig. 2), they might lack the
signal at m/z 173 because the fragmentation at an aromatic OH
needs more energy to be produced. Since this same signal is present
in the fragmentation spectrum of the hydroxylated metabolite of
compound 2 (Fig. 1(d)) and given the high structural similarity with

212
E.M. Gargano et al. / European Journal of Medicinal Chemistry 87 (2014) 203e219
Table 6
therefore suggesting that the compounds are not steroidomimetics
and do not have the same binding mode as the natural substrate E2.
The effect of all the 9 possible substitution pattern of the methyl
groups on rings A and C was studied. The substitution in meta or
ortho position on ring A, while keeping a methyl in meta position on
ring C (compounds 12 and 13, IC50 ¼ 52 and 58 nM) leads to an
increase in activity of a factor of nearly 20, in comparison with the
compound having the two unsubstituted phenyl rings (48% inhi-
Percentage of [3H]E2 displaced from the ER
a and b for selected compounds.
bition at 1 mM) [21]. Recently Leung et al. reported about the effects
Compd. Ring B
R1
R2
Percentage of [3H]E2 displaced by the
inhibitora,d
of the addition of a methyl group to a lead compound on its bio-
logical activity, revealing how, from analysis of more than 2000
examples described in the literature, an activity improvement of a
factor of 10 or more is found with an 8% frequency only [49]. The
authors pointed out that the highest activity increase is often the
result from the combination of good fitting of the methyl group into
a hydrophobic region of the protein target together with a possible
conformational gain, derived from an ortho methyl substitution at a
Estrogen receptor Estrogen receptor
ab
bc
12
p-Me
p-Me
o-Me 12
21
13
21
22
m-
Me
11
5
34
4
Table 7
m-Me, p- m-
Me
m-Me
Half-life of representative compounds measured in human liver microsomes S9
fraction.
Me
m-Cl 16
11
30
m-Me
m-
Me
9
8
0
35
e
e
11
a
Mean value of at least two determinations, standard deviation less than 25%.
b
c
Recombinant human receptor, ER
a
b
1 nM, [3H]E2 3 nM.
4 nM, [3H]E2 10 nM.
Compd. Ring B
Ring A
R1
R2
R3
Inhibitorc t1/2
Recombinant human receptor, ER
(min)a,b
d
Inhibitor tested at 1000 fold the [3H]E2 concentration, 3
m
M for ER
a
, 10
mM for
ERb.
1
e
e
m-OMe m-
OMe
m-
eMe
eMe
17
19
12
m-Me
12, we could speculate that the hydroxylation reaction occurs at the
methyl on the A ring of compound 2 as observed for 12.
Me
The main metabolic pathway for compound 12 is depicted in
Fig. 6. The different metabolites of 12 are expressed both as per-
centage of the total found metabolites and as percentage of the
remaining parent compound in brackets. 39% of the metabolites
derives from the hydroxylation of the methyl on the A ring (com-
pound 40) and its following transformation into a carboxylic acid
group (compound 42) accounts for 10% of the total bio-
transformations. The cleavage of the methyl on the amide function
(compound 43) represents 19% of all the metabolites and the
combination of hydroxylation and demethylation (compound 44)
accounts for 12% of the detected biotransformation products.
13
e
p-Me
m-
Me
o-Me eMe
m-
Me
m-Cl eMe
eMe
15
16
21
e
e
p-Me
m-Me,
p-Me
m-Me
o-F, m- m-
OMe
22
20
eMe
22
41
e
e
32
117
eH
OMe
30
e
m-Me
m-
Me
eMe
5
4. Discussion
31
35
e
e
m-Me
m-
Me
eMe
92
The main goal of this study was the optimization of the meta-
bolic stability of our in house 17
satisfying understanding of their main routes of metabolism. The
new 17 -HSD2 inhibitors were designed trying to keep potency and
selectivity over 17 -HSD1 and the ERs in an acceptable range while
b-HSD2 inhibitors, together with a
e
e
eMe >120
b
b
25
27
e
e
e
e
e
eMe
eMe
21
82
improving their metabolic stability.
4.1. Potency and selectivity
In the design process we focused on compounds bearing methyl
substituents on rings A and C. Compound 12 (IC50 ¼ 52 nM) was
obtained, which retains the activity of the dimethoxy analog 1
(IC50 ¼ 58 nM). This result indicates that a polar moiety is not
necessary to achieve high potency as previously believed [24],
a
Extrapolated value. Mean of at least two determinations, standard deviation less
than 25%.
b
1 mg/ml pooled mammalian liver S9 fraction (BD Gentest), 2 mM NADPH
regenerating system, 1 mM UDPGA and 0.1 mM PAPS at 37 ꢀC for 0, 5, 15 and 60 min.
c
Inhibitor tested at a final concentration of 1 mM.

E.M. Gargano et al. / European Journal of Medicinal Chemistry 87 (2014) 203e219
213
Table 8
stability of the inhibitors, as seen for compounds 12, 13, 22, 21 and
22 with half-life in the same range as previous molecules (t1/2
around 20 min). While replacement of the methoxy group by
methyl could preserve inhibitory potency, it did not influence
metabolic stability. This can be explained looking at the primary
metabolite of 1 and 12: the methoxy group of 1 as well as the
methyl group of 12 on ring A are soon oxidized, thus leading to
ether cleavage and CH2OH formation, respectively. The rate of this
reaction, which might be catalyzed by the same enzyme, appears to
be in the same order of magnitude for both molecule and the
metabolic stability is clearly not influenced by the nature of the
substituents on the two compounds.
List of expected metabolites and their amounts, expressed as area percentage of the
remaining parent compound.
Mass
difference
Formula Description
Percentage
of metabolite
relative to
parent
compounda
1
2
12
0
e
Parent
100 100 100
þ2
eCH2þO Demethylation þ Hydroxylation
nf
4
6
þ16
þO
Hydroxylation, N, S oxidation,
epoxidation
12 14 20
5
4
þ30
þ32
ꢂ14
þO2eH2 CH3 to CO2H
nf nf
5
7
10
Given the high structural similarity of compounds 1 and 12 with
others 2,5-thiophene amide inhibitors and considering the instability
of the tested compounds of this class, it is very likely that the oxidation
observed for 1 and 12 is the main metabolic route for the other com-
pounds as well.
Compounds bearing hydroxy groups display half-lifes either
similar or lower than inhibitors without hydroxy groups (Table 1)
and are likely to be also affected by metabolic phase ІІ reactions.
This metabolite identification studyclearly indicates that the central
thiophene of compound 12 does not constitute a metabolic reactive site
and is not susceptible to any biotransformation. Therefore it is striking
that the exchange of the thiophene either bya furane or by benzene ring
leads to more stable compounds like 31 and 35 (t1/2 ¼ 92 min and
>120 min, respectively). This means that it is possible to modulate the
metabolic rate of compound 12 by modification of the central ring,
although it does not constitute a metabolic labile site.
Thestabilityofcompound31 (cLogP ¼ 3.86)mightbeexplainedby
the decrease in cLogP accomplished by the furane ring in comparison
to the thiophene derivative 12 (cLogP ¼ 4.77). This result is also
confirmed by the 1-methyl-1H-pyrazole compound 27 (cLogP ¼ 2.46)
which shows a half-life inhumanliverS9 fractionof82min. Exchange
of the thiophene by a benzene ring does not significantly modify the
cLogP for 35 (cLogP ¼ 4.93). The improvement in the t1/2 of this
compound might be caused by a decreased affinity for the metabo-
lizing enzyme, due to a higher steric hindrance.
þO2
Di-hydroxylation
Demethylation
2
27
9
nf
7
eCH2
ꢂ28
eC2H4 Didemethylation
2
nf nf
nf: not found.
a
Area percentage is calculated based on the assumption that each metabolite is
equally responsive as parent.
phenyl ring. The increase in inhibitory potency observed in our
class of inhibitors might come from the burial of the methyl group
in meta position on ring A in a hydrophobic pocket of the enzyme,
since any conformational restrain induced by ortho substitution at
this ring (compounds 17 and 20, 14% and 27% inhibition at 1 mM,
respectively) is detrimental for activity.
By introducing two methyl groups in para and meta position on
ring A, the most potent compound of this class (21, IC50 ¼ 33 nM) was
obtained, indicating additive effects of the two methyl groups. It also
reveals the presence of a likely hydrophobic cavity around ring A.
Exchange of the A ring of compound 12 by different heterocycles
like methyl pyridine leads to a significant decrease of activity. This
might be due to an unfavorable electrostatic potential of the ring,
compared to the phenyl (Table 3) or to the presence of the polar
nitrogen, which does not fit well into the enzyme, confirming the
presence of a lipophilic cavity. Compound 27, bearing a 1-methyl-
1H-pyrazole, is even less potent (Table 3) confirming that polar
groups on the A ring are not compatible with a strong binding to
the protein target.
4.3. Conclusion
The replacement of the central thiophene ring by a furane is also
In the drug discovery process it is important to take into account
the metabolic properties of lead compounds as early as possible.
The design of metabolically stable molecules, which maintain the
pharmacological activity, is a challenging task. In this work we
identified the metabolic liability of a series of 2,5-thiophene amide
inhibitors. We have also detailed the specific metabolic pathway of
compound 12 and profiled the metabolic fate of 1 and 2, leading to
an overall comprehension of the metabolic liabilities of the 2,5-
thiophene amide class of inhibitors. Using two different strategies
we discovered three new inhibitors with a highly improved
metabolic stability: 27 (t1/2 ¼ 82 min), 31 (t1/2 ¼ 92 min) and 35 (t1/
detrimental for activity (31, 14% of inhibition at 1
on the effects of different substituents the A ring resulted in com-
pound 34 as best inhibitor of this series (38% of 17 -HSD2 inhibi-
tion at 1 M and 46% of 17 -HSD1 inhibition at 1 M), indicating
the inadequateness of this central ring. The furane ring indeed
seems to reverse the activity in favor of 17 -HSD1 inhibition, since
31 and 34 are slightly more active on 17 -HSD1 and much less
active on 17 -HSD2, in comparison to compound 12 (Table 4).
mM). A small study
b
m
b
m
b
b
b
4.2. Metabolic stability
¼ >120 min). The inhibitory potency and selectivity profiles of the
2
Compounds 1, 2 and 3 taken as leads for the study of metabolic
stability were found to be rather unstable (t1/2 ¼ 17 min, 38 min and
4 min, respectively). An important goal of this work was the
improvement of the metabolic stability of this class of compounds.
To achieve this goal two different techniques were applied: 1.
Modification of the potential reactive site of the inhibitors, 2.
Lowering their cLogP.
synthesized compounds were also evaluated.
Among the three most stable inhibitors, 35 also retains a good
activityand selectivity (IC50 ¼1126nM, s.f. ¼10)and shows noaffinity
for the estrogen receptors. The relative structural simplicity of com-
pound 35 still leaves room for an appropriate molecular optimization.
5. Experimental section
In order to understand the reason for the instability of 1, 2 and 3
and in order to identify the metabolically reactive sites in these
molecules, a small set of compounds was synthesized and tested. As
the methoxy groups seemed to be labile in presence of hepatic
enzymes, they were exchanged by methyl groups, in principle
slightly more stable. However, this strategy did not improve the
5.1. Chemical methods
Chemical names follow IUPAC nomenclature. Starting materials
were purchased from Aldrich, Acros, Combi-Blocks, Enamine or
Fluka and were used without purification.

214
E.M. Gargano et al. / European Journal of Medicinal Chemistry 87 (2014) 203e219
Fig. 1. MS2 spectra of a) 1; b) 1 major demethylation product; c) 2; d) 2 major oxidation product; e) 12; d) 12 oxidation product; g) 12 CH3 to CO2H with proposed structures.

E.M. Gargano et al. / European Journal of Medicinal Chemistry 87 (2014) 203e219
215
Fig. 2. Potential oxidized metabolites of compound 12.
Column chromatography was performed on silica gel
(70e200 m) and reaction progress was monitored by TLC on
5.1.1. Method A, general procedure for amide formation
m
A mixture of bromo-N-heteroarylcarboxylic acid (2 mmol), thi-
onyl chloride (4 mmol) and DMF (5 drops) in toluene (10 mL) was
refluxed at 110 ꢀC for 4 h. The reaction mixture was cooled to room
temperature; the solvent and the excess of thionyl chloride were
removed under reduced pressure. The corresponding N-methyl-
amine (2 mmol) and Et3N (2 mmol) in CH2Cl2 (10 mL) was added at
0 ꢀC under N2 atmosphere to the acyl chloride. After 30 min at 0 ꢀC,
the ice bath was removed and the solution was warmed up and
stirred at room temperature overnight. The reaction mixture was
extracted twice with CH2Cl2 (2 ꢃ 15 mL); the organic layer was
dried over MgSO4, filtered and the solution was concentrated under
reduced pressure. The residue was purified by silica gel column
chromatography using n-hexane and EtOAc as eluent or by tritu-
ration in a mixture of diethyl ether/petroleum ether to afford the
desired compound.
Alugram SIL G/UV254 (MachereyeNagel). Visualization was
accomplished with UV light.
1H NMR and 13C NMR spectra were measured on a Bruker
AM500 spectrometer (at 500 MHz and 125 MHz, respectively) at
300 K and on Bruker Fourier 300 (at 300 MHz and 75 MHz,
respectively) at 300 K. Chemical shifts are reported in
d (parts per
million: ppm), by reference to the hydrogenated residues of
deuterated solvent as internal standard: 2.05 ppm (1H NMR) and
29.8 and 206.3 ppm (13C NMR) for CD3COCD3, 7.26 ppm (1H NMR)
and 77.0 ppm (13C NMR) for CDCl3. Signals are described as br
(broad), s (singlet), d (doublet), t (triplet), dd (doublet of doublets),
ddd (doublet of doublet of doublets), dt (doublet of triplets) and m
(multiplet). All coupling constants (J) are given in Hertz (Hz).
Melting points (mp) were measured in open capillaries on a
Stuart Scientific SMP3 apparatus and are uncorrected.
Mass spectrometry was performed on a TSQ® Quantum (Ther-
moFisher, Dreieich, Germany). The triple quadrupole mass spec-
trometer was equipped with an electrospray interface (ESI). The
purity of the compounds was assessed by LC/MS. The Surveyor®-
LC-system consisted of a pump, an auto sampler, and a PDA de-
tector. The system was operated by the standard software Xcali-
bur®. An RP C18 NUCLEODUR® 100-5 (3 mm) column
(MachereyeNagel GmbH, Dühren, Germany) was used as station-
ary phase. All solvents were HPLC grade. In a gradient run the
percentage of acetonitrile (containing 0.1% trifluoroacetic acid) was
increased from an initial concentration of 0% at 0 min to 100% at
5.1.2. Method B, general procedure for SuzukieMiyaura coupling
In a sealed tube the previously prepared bromo-N-hetero-
arylcarboxamide derivative (1 eq.) was introduced followed by the
corresponding boronic acid (1.5 eq.), cesium carbonate (3 eq.),
tetrakis(triphenylphosphine)palladium (0.02 eq.) and a mixture of
DME/EtOH/H2O (1:1:1, v:v:v, 3 mL) as solvent. The reactor was
flushed with N2 and submitted to microwave irradiation (150 ꢀC,
150 W) for 20 min. After cooling to room temperature, a mixture of
EtOAc/H2O (1:1, v:v, 2 mL) was added to stop the reaction. The
aqueous layer was extracted with EtOAc (3 ꢃ 10 mL). The organic
layer was washed once with brine and once with water, dried over
MgSO4, filtered and the solution was concentrated under reduced
pressure. The residue was purified by column chromatography
using n-hexane and EtOAc as eluent to afford the desired
compound.
15 min and kept at 100% for 5 min. The injection volume was 15
and flow rate was set to 800
mL
m
L/min. MS analysis was carried out at
a needle voltage of 3000 V and a capillary temperature of 350 ꢀC.
Mass spectra were acquired in positive mode from 100 to 1000 m/z
and UV spectra were recorded at the wave length of 254 nm and in
some cases at 360 nm.
5.1.3. Detailed synthesis procedures of the most interesting
compounds
All microwave irradiation experiments were carried out in a 507
CEM-Discover microwave apparatus.
All tested compounds exhibited ꢁ95% chemical purity as
measured by LC/MS, after dissolving them in methanol.
The following compounds were prepared according to previ-
5.1.3.1. N-Methyl-N,5-bis(3-methylphenyl)thiophene-2-carboxamide
(12). The title compound was prepared by reaction of 5-bromo-N-
methyl-N-(3-methylphenyl)thiophene-2-carboxamide
12a
(150 mg, 0.48 mmol), m-tolyl boronic acid (86 mg, 0.63 mmol),
cesium carbonate (469 mg, 1.44 mmol) and tetrakis(-
triphenylphosphine) palladium (11 mg, 0.02 eq) according to
method B. The residue was purified by silica gel column
ously
described
procedures:
5-bromo-N-methyl-N-(3-
methylphenyl)thiophene-2-carboxamide 12a [25], lithium 5-(3-
methylphenyl)-1,3-thiazole-2-carboxylate 30a [42].

216
E.M. Gargano et al. / European Journal of Medicinal Chemistry 87 (2014) 203e219
Fig. 3. Ion chromatogram of 12 major oxidation product and its co-injection with the synthesized potential metabolite. On the top of the peaks, the retention times are shown. From
the top to the bottom are displayed the six chromatograms of the a) hydroxylation products after incubation with the S9 fraction; b) co-injected with 40; c) co-injected with 36; d)
co-injected with 38; e) co-injected with 37; f) co-injected with 39.
Fig. 4. a) Fragmentation pattern of the pure 40; b) fragmentation pattern of 12 major hydroxylated product.
chromatography (n-hexane/ethyl acetate 80:20) to afford the
desired product as yellow solid (117 mg, 76%). C20H19NOS; MW 321;
mp: 98e101 ꢀC; MS (ESI) 322 [MþH]þ; 1H NMR (CDCl3, 500 MHz)
2.36 (s, 3H), 2.39 (s, 3H), 3.47 (s, 3H), 6.68 (d, J ¼ 4.1 Hz, 1H), 6.97 (d,
J ¼ 4.1 Hz, 1H), 7.09e7.12 (m, 3H), 7.21e7.25 (m, 2H), 7.28e7.29 (m,
2.32 (s, 3H), 2.36 (s, 3H), 3.37 (s, 3H), 6.49 (d, J ¼ 3.9 Hz, 1H), 7.07 (d,
J ¼ 3.9 Hz, 1H), 7.14e7.16 (m, 1H), 7.19e7.22 (m, 3H), 7.24e7.26 (m,
1H), 7.35 (t, J ¼ 7.8 Hz, 1H), 7.46 (dt, J ¼ 2.0, 8.3 Hz, 2H); 13C NMR
(CD3COCD3, 125 MHz)
d 21.2, 21.3, 39.1, 123.4, 126.1, 126.6, 129.6,
129.7, 130.5, 130.6, 131.8, 133.3, 138.6, 139.3, 140.8, 145.5, 149.3,
1H), 7.31e7.34 (m, 2H); 13C NMR (CDCl3, 125 MHz)
d 21.3, 21.4, 39.4,
162.5; IR (cmꢂ1) 3050, 2921, 1599, 1585.
122.7, 123.2, 125.0, 126.7, 128.5, 128.9, 129.2, 129.7, 133.4, 133.5,
136.3, 138.6, 140.0, 143.8, 149.7, 163.1; IR (cmꢂ1) 3048, 2921, 1599,
1584.
5.1.3.3. 5-(3,4-Dimethylphenyl)-N-methyl-N-(3-methylphenyl)thio-
phene-2-carboxamide (21). The title compound was prepared by
reaction of 5-bromo-N-methyl-N-(3-methylphenyl)thiophene-2-
carboxamide 12a (150 mg, 0.48 mmol), 3,4-dimethylphenyl
boronic acid (94.5 mg, 0.63 mmol), cesium carbonate (469 mg,
5.1.3.2. N-Methyl-N-(3-methylphenyl)-5-(4-methylphenyl)thio-
phene-2-carboxamide (13). The title compound was prepared by
reaction of 5-bromo-N-methyl-N-(3-methylphenyl)thiophene-2-
carboxamide 12a (100 mg, 0.32 mmol), p-tolyl boronic acid
(65 mg, 0.48 mmol), cesium carbonate (313 mg, 0.96 mmol) and
tetrakis(triphenylphosphine) palladium (8 mg, 0.02 eq) according
to method B. The residue was purified by silica gel column chro-
matography (n-hexane/ethyl acetate 80:20) to afford the desired
product as yellow solid (63 mg, 61%). C20H19NOS; MW 321; mp:
95e98 ꢀC; MS (ESI) 322 [MþH]þ; 1H NMR (CD3COCD3, 500 MHz)
Fig. 5. Proposed structure for signal at m/z 173.

E.M. Gargano et al. / European Journal of Medicinal Chemistry 87 (2014) 203e219
217
temperature for 3 h. The solvent was removed under reduced
pressure (bath temperature of rotavapor at 20 ꢀC). This residue was
diluted in dry CH2Cl2 (10 mL) and N,3-dimethylaniline (0.06 mL,
0.44 mmol) was added followed by triethylamine (0.10
mL,
0.44 mmol). The solution was stirred overnight at room tempera-
ture under N2 atmosphere. The solvent was removed under
reduced pressure (bath temperature of rotavapor at 20 ꢀC). A
mixture of Na2CO3 2N (15 mL) and EtOAc (15 mL) were added. The
aqueous layer was extracted with EtOAc (15 mL). The organic layer
was washed twice with Na2CO3 2N (15 mL), once with water
(15 mL), dried over MgSO4, filtered and the solution was concen-
trated under reduced pressure. The residue was purified by silica
gel column chromatography (n-hexane/ethyl acetate 70:30) to
afford the desired product as pale brown solid (51 mg, 36%).
C
19H18N2OS; MW 322; mp: 79e80 ꢀC; MS (ESI) 323 [MþH]þ; 1H
NMR (CDCl3, 500 MHz) 2.32 (s, 3H), 2.36 (s, 3H), 3.54 (br s, 3H),
7.07e7.09 (m, 1H), 7.10e7.12 (m, 1H), 7.15 (br s, 1H), 7.20e7.26 (m,
1H), 7.32 (t, J ¼ 7.7 Hz, 1H), 7.44e7.46 (m, 2H), 7.49 (br s, 1H), 7.93
(br s, 1H); 13C NMR (CDCl3, 125 MHz) 21.3, 39.6, 124.9, 125.0, 128.3,
128.4, 128.6, 130.0, 130.1, 130.6, 131.5, 139.69, 139.70, 139.9, 144.2,
145.4, 161.2. IR (cmꢂ1) 3096, 2922, 1624, 1606, 1585.
5.1.3.6. N-Methyl-N,5-bis(3-methylphenyl)furan-2-carboxamide
(31). The title compound was prepared by reaction of 31a (150 mg,
0.51 mmol), m-tolyl boronic acid (104 mg, 0.77 mmol), cesium
carbonate (499 mg, 1.53 mmol) and tetrakis(triphenylphosphine)
palladium (12 mg, 0.02 eq) according to method B. The residue was
purified by silica gel column chromatography (n-hexane/ethyl ac-
etate 70:30) to afford the desired product as orange oil (100 mg,
64%). C20H19NO2; MW 305; MS (ESI) 306 [MþH]þ; 1H NMR
(CD3COCD3, 500 MHz) 2.30 (s, 3H), 2.38 (s, 3H), 3.37 (s, 3H), 6.67 (d,
J ¼ 3.7 Hz, 1H), 6.74 (d, J ¼ 3.7 Hz, 1H), 7.00e7.01 (m, 1H), 7.07e7.12
(m, 2H), 7.16e7.22 (m, 3H), 7.24e7.27 (m,1H), 7.35 (t, J ¼ 7.8 Hz,1H);
Fig. 6. Proposed pathway for the phase І metabolism of 12 in human S9 fraction.
1.44 mmol) and tetrakis(triphenylphosphine) palladium (11 mg,
0.02 eq) according to method B. The residue was purified by silica
gel column chromatography (n-hexane/ethyl acetate 80:20) to
afford the desired product as colorless solid (55 mg, 34%).
13C NMR (CD3COCD3, 125 MHz)
d 21.3, 38.7, 107.1, 119.4, 122.3, 125.3,
C
21H21NOS; MW 335; mp: 138e140 ꢀC; MS (ESI) 336 [MþH]þ; 1H
125.4, 128.8, 129.0, 129.5, 129.9, 130.2, 130.7, 139.2, 140.3, 145.9,
148.2, 155.8, 159.1.
NMR (CDCl3, 500 MHz) 2.26 (s, 3H), 2.27 (s, 3H), 2.38 (s, 3H), 3.45 (s,
3H), 6.57 (d, J ¼ 4.0 Hz, 1H), 6.92 (d, J ¼ 4.0 Hz, 1H), 7.08e7.11 (m,
3H), 7.20e7.21 (m, 1H), 7.23 (dd, J ¼ 2.1, 7.8 Hz, 1H), 7.28e7.33 (m,
5.1.3.7. N,30-Dimethyl-N-(m-tolyl)-[1,10-biphenyl]-4-carboxamide
(35). The title compound was prepared by reaction of 35a (120 mg,
0.39 mmol), m-tolyl boronic acid (69 mg, 0.51 mmol), cesium car-
bonate (381 mg, 1.17 mmol) and tetrakis(triphenylphosphine)
palladium (9 mg, 0.02 eq) according to method B. The residue was
purified by silica gel column chromatography (n-hexane/ethyl ac-
etate 80:20) to afford the desired product as yellow oil (105 mg,
2H); 13C NMR (CDCl3, 125 MHz)
d 19.5, 19.7, 21.3, 39.1, 122.2, 123.4,
125.0,127.2, 128.5, 128.9, 129.6, 130.1, 131.2, 132.8, 136.6, 137.0, 137.1,
139.8, 144.1, 149.4, 162.6; IR (cmꢂ1) 3029, 2918, 1612, 1602.
5.1.3.4. N-Methyl-N-(3-methylphenyl)-5-(1-methyl-1H-pyrazol-4-yl)
thiophene-2-carboxamide (27). The title compound was prepared
by reaction of 5-bromo-N-methyl-N-(3-methylphenyl)thiophene-
2-carboxamide 12a (150 mg, 0.48 mmol), 1-methyl-1H-pyrazole-4-
boronic acid (96 mg, 0.63 mmol), cesium carbonate (626 mg,
1.92 mmol) and tetrakis(triphenylphosphine) palladium (11 mg,
0.02 eq) according to method B. The residue was purified by silica
gel column chromatography (n-hexane/ethyl acetate 70:30) to
afford the desired product as yellow solid (70 mg, 53%).
85%).
C
22H21NO; MW 315; MS (ESI) 316 [MþH]þ; 1H NMR
(CD3COCD3, 300 MHz) 2.24 (s, 3H), 2.37 (s, 3H), 3.42 (s, 3H),
6.93e6.96 (d, J ¼ 9 Hz, 1H), 6.98e7.00 (d, J ¼ 7 Hz, 1H), 7.07 (s, 1H),
7.11e7.18 (m, 2H), 7.30 (t, J ¼ 8 Hz, 1H), 7.37e7.42 (m, 4H), 7.47e7.50
(m, 2H); 13C NMR (CD3COCD3, 75 MHz) 21.2, 21.4, 38.5, 124.8, 125.1,
126.8, 127.9, 128.4, 128.44, 129.3, 129.6, 129.7, 130.1, 136.5, 139.3,
139.9, 140.8, 142.7, 146.2, 170.2. IR (cmꢂ1) 3037, 2920, 2862, 1740,
1639, 1603.
C
17H17N3OS; MW 311; mp: 132e134 ꢀC; MS (ESI) 312 [MþH]þ; 1H
NMR (CDCl3, 500 MHz) 2.37 (s, 3H), 3.43 (s, 3H), 3.89 (s, 3H), 6.49 (d,
J ¼ 4.0 Hz, 1H), 6.70 (d, J ¼ 4.0 Hz, 1H), 7.06e7.10 (m, 2H), 7.19e7.21
(m, 1H), 7.30 (t, J ¼ 7.7 Hz, 1H), 7.47 (s, 1H), 7.57 (d, J ¼ 1.0 Hz, 1H);
5.1.3.8. 5-[3-(Hydroxymethyl)phenyl]-N-methyl-N-(3-methylphenyl)
thiophene-2-carboxamide
(40). To
a
solution
of
5-(3-
13C NMR (CDCl3, 125 MHz)
d 21.6, 29.6, 39.0, 116.4, 122.0, 125.0,
formylphenyl)-N-methyl-N-(3-methylphenyl)thiophene-2-
127.4, 128.5, 128.9, 129.5, 132.7, 135.4, 137.0, 140.0, 140.2, 144.1,
carboxamide 40a (76 mg, 0.23 mmol) in a mixture of dry MeOH
(3 mL) and dry dioxane (3 mL) was added at 0 ꢀC sodium borohy-
dride (87 mg, 0.46 mmol). After 2 h at 0 ꢀC, water was added to
quench the reaction. The aqueous layer was extracted three times
with EtOAc (3 ꢃ 5 mL). The organic layer was washed once with
saturated solution of NaHCO3 and once with water, dried over
MgSO4, filtered and the solution was concentrated under reduced
pressure. The residue was purified by preparative HPLC (using
acetonitrile/water from 10% to 95% of acetonitrile) to afford the
162.4; IR (cmꢂ1) 3075, 2923, 2854, 1616, 1601, 1582.
5.1.3.5. N-Methyl-N,5-bis(3-methylphenyl)-1,3-thiazole-2-
carboxamide (30). To a solution of lithium 5-(3-methylphenyl)-1,3-
thiazole-2-carboxylate 30a (100 mg, 0.44 mmol) in CH2Cl2 (15 mL)
was added drop wise oxalyl chloride (80 mL, 0.88 mmol) followed
by few drops of DMF at 0 ꢀC under N2 atmosphere. The reaction
mixture was stirred at 0 ꢀC for 10 min and then at room

218
E.M. Gargano et al. / European Journal of Medicinal Chemistry 87 (2014) 203e219
desired product as colorless oil (14 mg, 18%). C20H19NO2S; MW 337;
MS (ESI) 338 [MþH]þ; 1H NMR (CDCl3, 300 MHz) 2.36 (s, 3H),
3.43 s, (3H, 4.70 (s, 2H)), 6.57 (d, J ¼ 4.0 Hz, 1H), 6.96 (d, J ¼ 4.0 Hz,
1H), 7.04e7.06 (m, 1H), 7.06e7.10 (m, 1H), 7.17e7.22 (m, 1H),
7.27e7.36 (m, 3H), 7.40 (dt, J ¼ 1.8, 7.2 Hz, 1H), 7.49e7.51 (m, 1H);
13C NMR (CDCl3, 75 MHz) 21.3, 39.1, 65.0, 122.9, 124.5, 125.0, 125.2,
126.8,128.5,129.0,129.1,129.6,132.9,133.9,137.4,139.9,141.7,144.1,
148.7, 162.6. IR (cmꢂ1): 3397, 3027, 2921, 2865, 1598, 1582.
treatment of the samples and HPLC separation was carried out as
mentioned above for 17 -HSD2.
b
5.3.4. Estrogen receptor affinity in a cellular free assay
The binding affinity of selected compounds to ER
a
and ERb was
determined according to the recommendations of the US Envi-
ronmental Protection Agency (EPA) by their Endocrine Disruptor
Screening Program (EDSP) [44] using recombinant human proteins.
Briefly, 1 nM of ER
a
and 4 nM of ER
b
, respectively, were incubated
with [3H]-E2 (3 nM for ER
a
and 10 nM for ER ) and test compound
b
5.2. logP Determination
for 16e20 h at 4 ꢀC.
The potential inhibitors were dissolved in DMSO (5% final con-
centration). Evaluation of non-specific-binding was performed
with unlabeled E2 at concentrations 100-fold of [3H]-E2 (300 nM
The logP values were calculated from ACD/Labs Percepta 2012
Release program. The logarithm of partition constant P (log P) was
calculated using the “GALAS” method (Global Adjusted Locally
According to Similarity). The program predicts clogP by comparing
the molecule with structurally similar molecules where experi-
mental data are known.
for ERa and 1000 nM for ERb). After incubation, ligand-receptor
complexes were selectively bound to hydroxyapatite (83.5 g/L in
TE-buffer). The bound complex was washed three times and
resuspended in ethanol. For radiodetection, scintillator cocktail
(Quickszint 212, Zinsser Analytic, Frankfurt) was added and sam-
ples were measured in a liquid scintillation counter (1450 LSC &
Luminescence Counter, Perkin Elmer).
5.3. Biological methods
[2,4,6,7-3H]-E2 and [2,4,6,7-3H]-E1 were purchased from Per-
kineElmer, Boston. Quickszint Flow 302 scintillator fluid was
bought from Zinsser Analytic, Frankfurt. Other chemicals were
purchased from Sigma, Roth or Merck.
From these results the percentage of [3H]-E2 displacement by
the compounds was calculated. The plot of % displacement versus
compound concentration resulted in sigmoidal binding curves. The
compound concentrations to displace 50% of the receptor bound
[3H]-E2 were determined. Unlabeled E2 IC50 values were deter-
mined in each experiment and used as reference. The E2 IC50 values
5.3.1. 17
b-HSD1 and 17
b-HSD2 enzyme preparation
Cytosolic (17
b
-HSD1) and microsomal (17
b-HSD2) fractions
accepted, were 3 20% nM for ERa and 10 20% nM for ERb.
were obtained from human placenta according to previously
described procedures [38,39,50]. Fresh tissue was homogenized
and the enzymes were separated by fractional centrifugation at
1000 g, 10,000 g and 150,000 g. The pellet fraction containing the
Relative Binding Affinity was determined by applying the
following equation: RBA[%] ¼ (IC50(E2)/IC50(compound))$100 [44].
This results in an RBA value of 100% for E2.
After the assay was established and validated, a modification
was made to increase throughput. Compounds were tested at
concentrations of 1000$IC50(E2). Compounds with less than 50%
displacement of [3H]-E2 at a concentration of 1000$IC50(E2) were
classified as RBA <0.1%.
microsomal 17
inhibition, while 17
ammonium sulfate from the cytosolic fraction for use of testing of
17 -HSD1 inhibition. Aliquots containing 17 -HSD1 or 17 -HSD2
b
-HSD2 was used for the determination of 17
b-HSD2
b
-HSD1 was obtained after precipitation with
b
b
b
were stored frozen.
5.3.5. Metabolic stability in a cell free assay
Compounds 3, 5, 7e9, 12, 13, 16, 21, 22, 25, 27, 30, 31, 35 and 41
were tested according to established method [45e47]. For evalua-
5.3.2. Inhibition of 17b-HSD2 in cell-free assay
Inhibitory activities were evaluated by an established method
with minor modifications [37,51,52]. Briefly, the enzyme prepara-
tion of phase I and II metabolic stability 1
mM compound was
tion was incubated with NADþ [1500
m
M] in the presence of po-
incubated with 1 mg/ml pooled mammalian liver S9 fraction (BD
Gentest), 2 mM NADPH regenerating system, 1 mM UDPGA and
0.1 mM PAPS at 37 ꢀC for 0, 5, 15 and 60 min at a final volume of
tential inhibitors at 37 ꢀC in
a phosphate buffer (50 mM)
supplemented with 20% of glycerol and EDTA 1 mM. Inhibitor stock
solutions were prepared in DMSO. Final concentration of DMSO
was adjusted to 1% in all samples. The enzymatic reaction was
started by addition of a mixture of unlabeled- and [3H]-E2 (final
100 mL. The incubation was stopped by precipitation of S9 enzymes
with 2 volumes of cold acetonitrile containing internal standard.
Concentration of the remaining test compound at the different time
points was analyzed by LC-MS/MS and used to determine half-life
(t1/2).
Compounds 1, 2, 4, 6, 10 and 11 were tested following a similar
procedure with the following changes: 3
incubated with 1 mg/ml pooled mammalian liver S9 fraction (IVT,
Xenotech or TCS cellworks), 1 mM NADPH regenerating system,
0.75 mM UDPGA and 0.05 mM PAPS at 37 ꢀC for 0, 5, 15 and 45 min.
concentration: 500 nM, 0.11 mCi). After 20 min, the incubation was
stopped with HgCl2 and the mixture was extracted with ether. After
evaporation, the steroids were dissolved in acetonitrile/water
(45:55). E1 and E2 were separated using acetonitrile/water (45:55)
as mobile phase in a C18 RP chromatography column (Nucleodur
mM compound was
C18, 3 mm, MachereyeNagel, Düren) connected to an HPLC-system
(Agilent 1100 Series, Agilent Technologies, Waldbronn). Detection
and quantification of the steroids were performed using a radioflow
detector (Berthold Technologies, Bad Wildbad). The conversion rate
5.3.6. Metabolite identification in a cell free assay
was calculated according to the following equation:
%
Compounds 1, 2 and 12 were incubated at a final concentration
conversion ¼ (%E1/(%E1 þ %E2)) ꢃ 100. Each value was calculated
of 10
Gentest) and 2 mM NADPH regenerating system for three hours
(final volume mL). The water phase was extracted with
dichloromethane (3 ꢃ mL), which was evaporated to dryness at
room temperature. The residue was resuspended in 500 L meth-
anol. Separation of 2 L sample was accomplished on a Dionex
Ultimate 3000 RSLC system using a BEH C18 50 ꢃ 2.1 mm, 1.7 M dp
column (Waters, Germany). Separation was achieved by a linear
mM with 1 mg/ml pooled mammalian liver S9 fraction (BD
from at least three independent experiments.
3
5.3.3. Inhibition of 17
The 17 -HSD1 inhibition assay was performed similarly to the
-HSD2 test. The microsomal fraction was incubated with NADH
[500
M], test compound and a mixture of unlabeled- and [3H]-E1
(final concentration: 500 nM, 0.15
Ci) for 10 min at 37 ꢀC. Further
b-HSD1 in cell-free assay
b
m
17b
m
m
m
m

E.M. Gargano et al. / European Journal of Medicinal Chemistry 87 (2014) 203e219
219
[13] D. Vanderschueren, L. Vandenput, S. Boonen, M.K. Lindberg, R. Bouillon,
C. Ohlsson, Endocr. Rev. 25 (2004) 389e425.
[14] L. Wu, M. Einstein, W.M. Geissler, H.K. Chan, K.O. Elliston, S. Andersson, J. Biol.
Chem. 268 (1993) 12964e12969.
[15] S. Marchais-Oberwinkler, P. Kruchten, M. Frotscher, E. Ziegler, A. Neugebauer,
U. Bhoga, E. Bey, U. Müller-Vieira, J. Messinger, H. Thole, R.W. Hartmann,
J. Med. Chem. 51 (2008) 4685e4698.
[16] E. Bey, S. Marchais-Oberwinkler, P. Kruchten, M. Frotscher, R. Werth, A. Oster,
O. Algül, A. Neugebauer, R.W. Hartmann, Bioorg. Med. Chem. 16 (2008)
6423e6435.
gradient from (A) H2O þ 0.1% formic acid (FA) to (B) ACN þ 0.1% FA
at a flow rate of 600
L/min and 45 ꢀC. The gradient was initiated by
m
a 0.33 min isocratic step at 5% B, followed by an increase to 95% B in
9 min to end up with a 1 min step at 95% B before reequilibration
under the initial conditions. The flow entered the LTQ/Orbitrap
(Thermo Scientific, Germany) using the Nanomate Advion Nano-
spray ion source (NSI). The Mass Spectrometer was set with the
following parameters: scan range 200e2000 m/z in positive ioni-
zation mode; 200 ꢀC Capillary Temperature; 1.7 kV Voltage at NSI;
Fourier Transform Mass Spectrometry (FTMS) Resolution of 30,000;
MS/MS analysis was data-dependent and triggered by the most
abundant ions.
Data Acquisition was performed with Xcalibur version 1.0.2.65
(Thermo Electron, San Jose, Ca, USA). Five-decimal monoisotopic
masses of 1, 2 and 15 and their predicted metabolites, calculated
from CambidgeSoft Chem & Bio Draw 11.0 using the ChemDrawPro
1.0 program, were used for the parent data and to filter data in Qual
Browser (Thermo Electron, San Jose, Ca, USA) with a mass tolerance
threshold of 5 ppm.
[17] E. Bey, S. Marchais-Oberwinkler, M. Negri, P. Kruchten, A. Oster, T. Klein,
A. Spadaro, R. Werth, M. Frotscher, B. Birk, R.W. Hartmann, J. Med. Chem. 52
(2009) 6724e6743.
€
[18] S. Marchais-Oberwinkler, C. Henn, G. Moller, T. Klein, M. Negri, A. Oster,
A. Spadaro, R. Werth, M. Wetzel, K. Xu, M. Frotscher, R.W. Hartmann,
J. Adamski, J. Steroid. Biochem. Mol. Biol. 125 (2011) 66e82.
[19] Y. Dong, Q.Q. Qiu, J. Debear, W.F. Lathrop, D.R. Bertolini, P.P. Tamburini, J. Bone
Miner. Res. 13 (1998) 1539e1546.
[20] M. Feix, L. Wolf, H.U. Schweikert, Mol. Cell. Endocrinol. 171 (2001) 163e164.
[21] L.J. Eyre, R. Bland, I.J. Bujalska, M.C. Sheppard, P.M. Stewart, M. Hewison,
J. Bone Miner. Res. 13 (1998) 996e1004.
[22] M. Wetzel, S. Marchais-Oberwinkler, R.W. Hartmann, Bioorg. Med. Chem. 19
(2011) 807e815.
€
[23] M. Wetzel, S. Marchais-Oberwinkler, E. Perspicace, G. Moller, J. Adamski,
R.W. Hartmann, J. Med. Chem. 54 (2011) 7547e7557.
[24] K. Xu, Y.A. Al-Soud, M. Wetzel, R.W. Hartmann, S. Marchais-Oberwinkler, Eur.
J. Med. Chem. 46 (2011) 5978e5990.
Author contributions
[25] S. Marchais-Oberwinkler, K. Xu, M. Wetzel, E. Perspicace, M. Negri, A. Meyer,
€
A. Odermatt, G. Moller, J. Adamski, R.W. Hartmann, J. Med. Chem. 56 (2013)
167e181.
The manuscript was written through contributions of all au-
thors. All authors have given approval to the final version of the
manuscript.
[26] E. Perspicace, A. Giorgio, A. Carotti, S. Marchais-Oberwinkler, R.W. Hartmann,
Eur. J. Med. Chem. 69 (2013) 201e205.
[27] K. Xu, M. Wetzel, R.W. Hartmann, S. Marchais-Oberwinkler, Lett. Drug Des.
Discov. 8 (2011) 406e421.
[28] Y.A. Al-Soud, S. Marchais-Oberwinkler, M. Frotscher, R.W. Hartmann, Arch.
Pharm. 345 (2012) 610e621.
Acknowledgment
[29] M. Wetzel, E.M. Gargano, S. Hinsberger, S. Marchais-Oberwinkler,
R.W. Hartmann, Eur. J. Med. Chem. 47 (2012) 1e17.
[30] E. Perspicace, S. Marchais-Oberwinkler, R.W. Hartmann, Molecules 18 (2013)
4487e4509.
We thank Josef Zapp for the NMR measurement, Isabella Mang,
Theresa Manz and Martina Jankowski for the biological test. We are
thankful to Prof. R. Müller for the access to the LTQ/Orbitrap
apparatus and Eva Luxemburger for the help in the measurements.
We acknowledge the Deutsche Forschungsgemeinschaft (DFG) for
financial support (Grants HA1315/12-1 and AD 127/10-1).
[31] A.E.F. Nassar, A.M. Kamel, C. Clarimont, Drug Discov. Today
9 (2004)
1020e1028.
[32] E.H. Kerns, L. Di, Drug-like Properties: Concepts, Structure Design and
Methods from ADME to Toxicity Optimization, first ed., Elsevier Inc., United
States of America, 2008, pp. 329e330.
[33] T.N. Thompson, Med. Res. Rev. 21 (2001) 412e449.
[34] P. Valadon, P.M. Dansette, J.P. Girault, C. Amar, D. Mansuy, Chem. Res. Toxicol.
9 (1996) 1403e1413.
Appendix A. Supplementary data
[35] J.St. Jean, Jr. David, C. Fotsch, J. Med. Chem. 55 (2012) 6002e6020.
[36] H. Fiesselmann, F. Thoma, Chem. Ber. 89 (1956) 1907e1912.
[37] K.M. Sam, S. Auger, V. Luu-The, D. Poirier, J. Med. Chem. 38 (1995) 4518e4528.
[38] W. Qiu, R.L. Campbell, A. Gangloff, P. Dupuis, R.P. Boivin, M.R. Tremblay,
D. Poirier, S.-X. Lin, FASEB J. 16 (2002) 1829e1831.
Supplementary data associated with this article can be found in
the online version, at http://dx.doi.org/10.1016/j.ejmech.2014.09.
061. These data include MOL files and InChiKeys of the most
important compounds described in this article.
[39] P. Kruchten, R. Werth, S. Marchais-Oberwinkler, M. Frotscher, R.W. Hartmann,
Mol. Cell. Endocrinol. 301 (2009) 154e157.
[40] E. Bey, S. Marchais-Oberwinkler, R. Werth, M. Negri, Y.A. Al-Soud, P. Kruchten,
A. Oster, M. Frotscher, B. Birk, R.W. Hartmann, J. Med. Chem. 51 (2008)
6725e6739.
References
[41] C.A. Lipinski, F. Lombardo, B.W. Dominy, P.J. Feeney, Adv. Drug Deliv. Rev. 23
(1997) 3e25.
[42] E. Perspicace, L. Cozzoli, E.M. Gargano, N. Hanke, A. Carotti, R.W. Hartmann,
Eur. J. Med. Chem. 83 (2014) 317e337.
[43] G. Drehesen, J. Engel, Sulfur Rep. 3 (1983) 171e207.
[44] EDSP, Protocol for the In Vitro Estrogen Receptor Saturation Binding and
Competitive Binding Assays Using Rat Uterine Cytosol, 2009. Available via
EPA: http://www.epa.gov/endo/pubs/assayvalidation/appendix1_er_ruc.pdf.
[45] L. Di, E.H. Kerns, Y. Hong, T.A. Kleintop, O.J. McConnell, D.M. Huryn, J. Biomol.
Screen 8 (2003) 453e462.
[1] M. Cree, C.L. Soskolne, E. Belseck, J. Hornig, J.E. McElhaney, R. Brant, M. Suarez-
Almazor, J. Am. Geriatr. Soc. 48 (2000) 283e288.
[2] L.J. Melton 3rd, E.A. Chrischilles, C. Cooper, A.W. Lane, B.L. Riggs, J. Bone Miner.
Res. 7 (1992) 1005e1010.
[3] R. Burge, B. Dawson-Hughes, D.H. Solomon, J.B. Wong, A. King, A. Tosteson,
J. Bone Miner. Res. 22 (2007) 465e475.
[4] R. Marcus, M. Wong, H. Heath 3rd, J.L. Stock, Endocr. Rev. 23 (2002) 16e37.
[5] E. Orwoll, M. Ettinger, S. Weiss, P. Miller, D. Kendler, J. Graham, S. Adami,
K. Weber, R. Lorenc, P. Pietschmann, K. Vandormael, A. Lombardi, N. Engl. J.
Med. 343 (2000) 604e610.
[6] J.D. Ringe, H. Faber, P. Farahmand, A. Dorst, Rheumatol. Int. 26 (2006)
427e431.
[7] B. Ettinger, D.M. Black, B.H. Mitlak, R.K. Knickerbocker, T. Nickelsen,
H.K. Genant, C. Christiansen, P.D. Delmas, J.R. Zanchetta, J. Stakkestad,
C.C. Glüer, K. Krueger, F.J. Cohen, S. Eckert, K.E. Ensrud, L.V. Avioli, P. Lips,
S.R. Cummings, J. Am. Med. Assoc. 282 (1999) 637e645.
[8] P.V.N. Bodine, B.S. Komm, Vitam. Horm. 64 (2002) 101e151.
[9] D.T. Felson, Y. Zhang, M.T. Hannan, D.P. Kiel, P.W. Wilson, J.J. Anderson,
N. Engl. J. Med. 329 (1993) 1141e1146.
[10] C.L. Chen, N.S. Weiss, P. Newcomb, W. Barlow, E. White, J. Am. Med. Assoc. 287
(2002) 734e741.
[46] J. Moreno-Farre, P. Workman, Fl Raynud, Aust.-Asian J. Cancer
6 (2007)
55e69.
[47] J.P.M. Hui, J.S. Grossert, M.J. Cutler, J.E. Melanson, Rapid. Commun. Mass.
Spectrom. 26 (2011) 345e354.
[48] T. Thiemann, D.J. Walton, A. Olivera Brett, J. Iniesta, F. Marken, Y.Q. Li, Arkivoc
9 (2009) 96e113.
[49] C.S. Leung, S.S.F. Leung, J. Tirado-Rives, W.L. Jorgensen, J. Med. Chem. 55
(2012) 4489e4500.
[50] D.W. Zhu, X. Lee, R. Breton, D. Ghosh, W. Pangborn, W.L. Daux, S.X. Lin, J. Mol.
Biol. 234 (1993) 242e244.
[51] S.X. Lin, F. Yang, J.Z. Jin, R. Breton, D.W. Zhu, V. Luu-The, F. Labrie, J. Biol. Chem.
274 (1992) 28762e28770.
[52] K.M. Sam, R.P. Boivin, M.R. Treemblay, S. Auger, D. Poirier, Drug Des. Discov.
15 (1998) 157e180.
[11] S.A. Beresford, N.S. Weiss, L.F. Voigt, B. McKnight, Lancet 349 (1997) 458e461.
[12] D. Vanderschueren, J. Gaytant, S. Boonen, K. Venken, Androgens and bone,
Curr. Opin. Endocrinol. Diabetes Obes. 15 (2008) 250e254.