Targeted Endocrine Therapy: Design, Synthesis, and Proof-of-Principle of 17β-Hydroxysteroid Dehydrogenase Type 2 Inhibitors in Bone Fracture Healing

Article abstract of DOI:10.1021/acs.jmedchem.8b01493

Current therapies of steroid hormone-dependent diseases predominantly alter steroid hormone concentrations (or their actions) in plasma, in target and nontarget tissues alike, rather than in target organs only. Targeted therapy through the inhibition of steroidogenic enzymes may pose an attractive alternative with much less side effects. Here, we describe the design of a nanomolar potent 17β-hydroxysteroid dehydrogenase type 2 (17β-HSD2) inhibitor (compound 15) and successful targeted intracrine therapy in a mouse bone fracture model. Blockade of 17β-HSD2 in bone is thought to increase intracellular estradiol (E2) and testosterone (T), which thereby inhibits bone resorption by osteoclasts and stimulates bone formation by osteoblasts, respectively. Administration of compound 15 in the mouse fracture model strongly increases the mechanical stability of the healing fractured bone because of a larger periosteal callus with newly formed bone without changing the plasma E2 and T concentrations. Steroidogenic 17β-HSD2 inhibition thus enables targeted intracrine therapy.

Full text of DOI:10.1021/acs.jmedchem.8b01493

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
Article
pubs.acs.org/jmc
Cite This: J. Med. Chem. XXXX, XXX, XXX−XXX
Targeted Endocrine Therapy: Design, Synthesis, and Proof-of-
Principle of 17β-Hydroxysteroid Dehydrogenase Type 2 Inhibitors in
Bone Fracture Healing
§
∥
Ahmed S. Abdelsamie,†,‡,○ Steven Herath,§,○ Yannik Biskupek, Carsten Borger,
̈
Lorenz Siebenbu
̈
rger,∥ Mohamed Salah,⊥ Claudia Scheuer,# Sandrine Marchais-Oberwinkler,¶
Martin Frotscher,⊥ Tim Pohlemann,§ Michael D. Menger,# Rolf W. Hartmann,⊥,∇
Matthias W. Laschke,# and Chris J. van Koppen*,†,⊥
†ElexoPharm GmbH, Im Stadtwald, Building A1.2, 66123 Saarbru
̈
cken, Germany
‡Chemistry of Natural and Microbial Products Department, National Research Centre, Dokki, 12622 Cairo, Egypt
§Department of Trauma, Hand and Reconstructive Surgery, and #Institute of Clinical & Experimental Surgery, Saarland University,
66421 Homburg/Saar, Germany
∥PharmBioTec GmbH, 66123 Saarbru
⊥Department of Pharmaceutical and Medicinal Chemistry, Saarland University, 66123 Saarbru
¶Institute of Pharmaceutical Chemistry, Philipps-University, 35032 Marburg, Germany
∇Department of Drug Design and Optimization, Helmholtz Institute for Pharmaceutical Research Saarland (HIPS), 66123
Saarbrucken, Germany
̈
cken, Germany
̈
cken, Germany
̈
S
* Supporting Information
ABSTRACT: Current therapies of steroid hormone-dependent diseases
predominantly alter steroid hormone concentrations (or their actions) in
plasma, in target and nontarget tissues alike, rather than in target organs only.
Targeted therapy through the inhibition of steroidogenic enzymes may pose an
attractive alternative with much less side effects. Here, we describe the design
of a nanomolar potent 17β-hydroxysteroid dehydrogenase type 2 (17β-HSD2)
inhibitor (compound 15) and successful targeted intracrine therapy in a mouse
bone fracture model. Blockade of 17β-HSD2 in bone is thought to increase
intracellular estradiol (E2) and testosterone (T), which thereby inhibits bone
resorption by osteoclasts and stimulates bone formation by osteoblasts,
respectively. Administration of compound 15 in the mouse fracture model
strongly increases the mechanical stability of the healing fractured bone
because of a larger periosteal callus with newly formed bone without changing
the plasma E2 and T concentrations. Steroidogenic 17β-HSD2 inhibition thus
enables targeted intracrine therapy.
to treat patients with benign prostatic hyperplasia,2 11β-HSD1
INTRODUCTION
■
inhibitors,3 and 17β-HSD1 inhibitors in an endometrial
hyperplasia animal model.4 In the present study, we have
developed novel nanomolar potent 17β-HSD2 inhibitors and
successfully performed a proof-of-principle study in a mouse
bone fracture healing model. Approximately 10% of the
patients with bone fractures suffer from serious nonunion or
delayed/impaired bone fracture healing.5 Inadequate bone
healing is not only difficult to treat, with the therapeutic
options being very limited,6−8 but the demand for novel drugs
with fewer side effects is also increasing rapidly because of the
relatively high prevalence of bone fractures, the anticipated
Current therapies of steroid hormone-dependent diseases, like
osteoporosis and endometriosis, involve primarily treatment
with drugs that alter the steroid hormone concentrations (or
their actions) in plasma, in target and nontarget tissues alike,
rather than in the target organs only. As a consequence, these
treatments most often are accompanied with highly undesir-
able side effects.1 Local intracrine inhibition of steroidogenic
enzymes, affecting the intracellular concentrations in the target
organs only, represents an attractive, alternative therapeutic
paradigm with much less side effects and disturbances of the
hypothalamus−pituitary organ axes. Unfortunately, successful
proof-of-principle studies with steroidogenic enzyme inhibitors,
which do not change systemic hormone concentrations, have
been rather infrequent. These include 5α-reductase inhibitors
Received: September 25, 2018
© XXXX American Chemical Society
A
DOI: 10.1021/acs.jmedchem.8b01493
J. Med. Chem. XXXX, XXX, XXX−XXX

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
Journal of Medicinal Chemistry
Article
worldwide increase in postmenopausal osteoporosis-related
bone fractures, and the growing elderly population. 17β-HSD2
in bone tissue converts the biologically active steroid
hormones, estradiol (E2) and T, into the much less active
estrone and androstenedione. The blockade of 17β-HSD2 in
bone thus increases intracellular E2 and T, and, through
estrogen and androgen receptor stimulation, inhibits bone
resorption by osteoclasts and stimulates bone formation by
osteoblasts,9−11 respectively. The expression and activity of
17β-HSD2 in human bone tissue are superior over that of 17β-
HSD1, 17β-HSD3, and 17β-HSD4,12−14 the enzymes that
catalyze the synthesis and degradation of E2 and T. A previous
study in ovariectomized cynomolgus monkeys showed that a
17β-HSD2 inhibitor increases the so-called ultimate bone
strength, that is, the maximum stress that an intact bone
specimen can sustain.15 Here, we developed a nanomolar
potent, metabolically sufficiently stable, nontoxic 17β-HSD2
inhibitor and tested this compound for activity in an
established mouse bone fracture healing model16 during a
proof-of-principle study of 28 days. Mouse bone cells express
active 17β-HSD2 as demonstrated by 17β-HSD2 immunohis-
tochemistry, using a validated antimouse 17β-HSD2 antibody
(Figure S1, Supporting Information), and shown by enzymatic
determination,17 together with androgen and estrogen
receptors (ERs).12−14
human and murine 17β-HSD2 (h + m17β-HSD2), resulting in
a small library of 16 compounds. As a starting point, the
nonsubstituted scaffold structure (compound A) was chosen.
Previous investigations revealed that the low metabolic
stability of the BSHs was primarily due to phase II
biotransformation of the phenolic OH group, which is essential
for activity.19 Therefore, stage 1 of the design strategy, aiming
at the improvement of metabolic stability, consisted in the
introduction of small substituents on the hydroxyphenyl
moiety (ring A) of compound A. These groups could protect
the OH functionality from biotransformation by the electron
withdrawal effect and/or steric hindrance (Chart 1, com-
pounds 1−8).
Compound 8 showed an enhanced metabolic stability and
was chosen as a starting point for the optimization of the
substitution pattern of the D-ring (stage 2 of the design
strategy). Only such groups were selected to be introduced
that were likely to maintain potency toward both human and
mouse 17β-HSD2.18 We aimed for a slight (three- to 
fourfold) selectivity over the h17β-HSD1 enzyme (Chart 2, 9−
16). On the one hand, a highly selective 17β-HSD2 inhibitor
would induce an undesirable increase in intracellular E2 in
tissues that express similar levels of 17β-HSD2 and 17β-HSD1
and prone to E2-dependent proliferation (i.e., breast20,21 and
endometrium22), whereas on the other hand, a nonselective
17β-HSD2/1 inhibitor would likely affect the role of 17β-
HSD1 in regulating the endometrium cyclicity in women of
childbearing age.22
RESULTS AND DISCUSSION
■
As bicyclic substituted hydroxyphenylmethanones (BSHs)
bearing a sulfonamide moiety were originally designed as
inhibitors of human 17β-HSD1, most members show
selectivity toward 17β-HSD1 over 17β-HSD2.18 In addition,
BSHs display low metabolic stability, precluding their use in an
in vivo proof-of-principle study. Thus, a rational two-stage drug
design strategy focusing on rings A and D (Charts 1 and 2, see
The starting point for the syntheses of compounds 1−16
was a Friedel−Crafts reaction of 2-bromothiophene with the
appropriate benzoyl chloride, which in case of the chlorinated
3b and 4b had to be prepared from the corresponding benzoic
acids. The obtained intermediates 1b−8b were subjected to
Suzuki cross-coupling reactions with 3-aminophenylboronic
acid to afford anilines 1a−8a. The latter were reacted with the
sulfonamides using the appropriately substituted sulfonic acid
chloride, giving direct access to compound 1. Ether cleavage
using BBr3 in dichloromethane yielded the final compounds
2−16 (Scheme 1).
Chart 1. Lead Compound A and Designed Compounds 1−8
The introduction of an electron-donating methyl group on
ring A (Table 1, 1) led to a twofold decrease in inhibitory
potency toward h17β-HSD2 compared to lead A. In contrast,
the presence of an electron-withdrawing fluorine or chlorine
atom in the same position strongly increased the activity; see 1
versus 2 and 3.
The beneficial effect of electron-withdrawing substituents on
the inhibitory activity was also apparent for 4−8, in agreement
with the observations made recently for the BSHs lacking the
sulfonamide moiety.22 Metabolic stability was determined for
3, 7, and 8 as these compounds displayed selectivity over 17β-
HSD1 and nanomolar potency toward m17β-HSD2. The
trifluoro substitution pattern of 8 resulted in an improved
metabolic stability and was therefore maintained in the
subsequent optimization of ring D. All eight compounds of
this second series (9−16) showed a strong inhibition of human
and mouse 17β-HSD2 as well as a low activity toward m17β-
HSD1. Compounds 14 and 15 were especially interesting as
they displayed moderate h17β-HSD2 selectivity, which was
aimed at, as well as improved metabolic stability.
Chart 2. Lead Compound 8 and Designed Compounds 9−
16
Because of their favorable potency, selectivity, and metabolic
stability properties, 14 and 15 were selected for testing for
potential cytotoxicity. Both 14 and 15 were not toxic in the
MTT assay with HEK293 cells (i.e., <20% reduction in cell
also Supporting Information for details) was applied that, on
the one hand, aimed at improving the metabolic stability and,
on the other, at enhancing the potency and selectivity for
B
DOI: 10.1021/acs.jmedchem.8b01493
J. Med. Chem. XXXX, XXX, XXX−XXX

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
Journal of Medicinal Chemistry
Article
Scheme 1. Synthetic Route to 1−16
viability over 66 h) up to concentrations that were
approximately 11 000-fold over the respective h17β-HSD2
IC50 value.
receptors are the transcription factors of the CYP1, CYP2, and
CYP3 families of CYP450 hepatic enzymes, respectively. We
next investigated whether or not the pharmacokinetic profile of
15 would allow proof-of-principle testing in a well-established,
reproducible bone fracture model in mice.16,24,25 For this, 15
was administered subcutaneously in C57BL/6 mice (50 mg/kg
body weight) as a suspension in 0.5% gelatine/5% mannitol in
water. Compound 15 showed a stable plasma concentration
over 24 h after a single-dose administration: 264 nM at 2 h,
175 nM at 18 h, and 191 nM at 24 h, all ∼10-fold higher than
the m17β-HSD2 IC50 of 25 nM and thus sufficiently high to
block m17β-HSD2 by >90% in vitro during 24 h. This
corresponds to the fairly high metabolic stability measured in
mouse liver S9 fraction (t1/2 of 46 min).
Next, to analyze bone fracture repair, C57BL/6 male mice
received 15 subcutaneously once daily (50 mg/kg) or vehicle
(0.5% gelatine/5% mannitol in water) for 14 or 28 days,
starting immediately following fracturing of the femur under
anesthesia. Bending stiffness of fractured bones, relative to the
contralateral unfractured bones in the same animal, is the
upmost important functional recovery parameter at the site of
the fracture in bone fracture healing.16,24,25 After 14 and 28
days, the bending stiffness of the fractured femurs of the
We also checked whether 14 and 15 do not bind to ERs at
relevant concentrations to avoid E2 being blocked from
activating ER receptors in bone. Compound 15 displayed
negligible ERα and ERβ binding, whereas 14 showed a higher
ER binding (i.e., 28 and 21% inhibition of E2 binding in the
presence of a 1000-fold excess of 15 over E2 vs 64 and 49%
inhibition in the presence of 1000-fold excess of 14,
respectively). Compound 15 was able to inhibit endogenously
expressed 17β-HSD2 in the cellular assay (human MDA-MB
231 cells) with an IC50 of 2.4 nM, which is very similar to the
IC50 of 2.0 nM in the 17β-HSD2 cell-free assay (Table 1). This
indicates that compound 15 readily enters cells at pharmaco-
logically relevant concentrations.
We thus selected 15 for further profiling and investigated in
vitro whether 15 does not induce the expression of hepatic
CYP450 enzymes, which otherwise may reduce the plasma
concentrations of 15 during repetitive dosing in the bone
fracture healing study. At 3.16 μM, 15 did neither activate the
aryl hydrocarbon receptor, constitutive androstane receptor,
nor the pregnane X receptor (<5% activation). These nuclear
C
DOI: 10.1021/acs.jmedchem.8b01493
J. Med. Chem. XXXX, XXX, XXX−XXX

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
Journal of Medicinal Chemistry
Article
Table 1. Inhibition of 17β-HSD2 and 17β-HSD1 by A and 1−16
a
a
a
IC50 (nM)
IC50 (nM)
% inh.@500 nM
metabolic stability t1/2 (min)
b
Cmpd
X
Y
h17β-HSD2
h17β-HSD1
s.f.
m17β-HSD2
m17β-HSD1
human liver S9
<5
c
c
A
1
2
3
4
5
6
7
111
231
6.9
0.9
13.8
5.6
8.5
6.7
1.2
0.4
0.2
0.4
0.2
1.4
2.0
1.8
21
0.2
1.3
1
22.5
0.6
0.3
0.2
2.3
1
1.8
0.5
2
0.5
3.4
3
n.i.
n.i.
c
c
d
2-CH3
−H
−H
−H
−H
−H
−H
−H
4-F
306
7.2
20.3
8.2
1.4
1.7
15.6
1.2
0.7
0.1
0.8
0.1
4.8
6.0
2.4
n.i.
n.i.
n.d.
n.d.
d
2-F
81.6
2.7
303.5
30.5
8.9
2.7
27.3
6.2
9.9
5.6
7.9
11.6
24.5
14.0
12
26
22
43
54
25
24
31
30
28
49
30
2-Cl
4-Cl
<5
n.d.
n.d.
n.d.
d
d
2,6-di-F
2-F,6-Cl
2-Cl,6-F
2,4,5-tri-F
2,4,5-tri-F
2,4,5-tri-F
2,4,5-tri-F
2,4,5-tri-F
2,4,5-tri-F
2,4,5-tri-F
2,4,5-tri-F
d
<5
32
40
32
42
28
51
60
9
10
11
12
13
14
15
16
4-Cl
4-CH3
3-Cl
3-CH3
2-CF3
2-OCF3
2-OCF3,4-Br
e
25 (1764 nM)
30
d
1.3
n.d.
a
b
c
Mean value of at least two determinations, standard deviation less than 20%. s.f.: selectivity factor = hIC50(17β-HSD1)/hIC50(17β-HSD2). n.i.:
d
e
no inhibition@1 μM. n.d.: not determined. mIC50(17β-HSD1). Final concentration of E1 and E2 in h17β-HSD1 and h17β-HSD2 assay: 500 and
500 nM, respectively. Final concentration of E1 and E2 in m17β-HSD1 and m17β-HSD2 assay: 10 and 10 nM, respectively.
vehicle-treated animals was 27.6 and 40.9% of the stiffness
determined in the nonfractured femurs of the same animal,
respectively. These values are similar to those reported
previously in C57BL/6 mice.24 After 14 days, no difference
in bending stiffness could be discerned between the 15- and
vehicle-treated animals (Figure 1 and Table S1) or in the
animals after 28 days (Figure 2C and Table S2, P = 0.025),
whereas the cartilage and fibrous tissue areas in the periosteal
callus area were not different (Table S2). Bone mineralization,
that is, the rise in the so-called high-density bone (from ∼7%
at 14 days to ∼18% at 28 days), and increase in bone mineral
density (from ∼0.24 g/cm3 at 14 days to ∼0.32 g/cm3 at 28
days) proceeded normally (Tables S3 and S4). The same holds
true for the trabecular number, trabecular thickness, and
trabecular separation in the callus (Table S4). Compound 15
does not affect bone mineral density, trabecular number,
trabecular thickness, and trabecular separation in the non-
fractured (intact) femurs of the animals (Table S5).
Compound 15 also does not influence the normal cortical
bone distant from the fracture: the diameter measurements of
the femurs demonstrate comparable values in the 15-treated
animals and vehicle controls (Table S6). These results indicate
that 17β-HSD2 inhibition stimulates bone formation in the
callus. As resorption of “old” bone at the fracture site precedes
new bone formation, we also investigated whether or not 15
increases the expression of osteoprotegerin (OPG), a
biomarker of the inhibition of bone resorption. OPG is an
E2-inducible, antiresorptive decoy receptor for the receptor
activator of nuclear factor κB ligand (RANKL), which in turn
promotes bone resorption.26 Indeed, treatment with 15 for 14
days increased the expression of OPG in callus by 55% (P =
0.048) compared to vehicle treatment, whereas the expression
of RANKL in callus did not differ (−1% change). These data
are consistent with our working hypothesis that 17β-HSD2
inhibition leads to both E2-mediated suppression of bone
resorption and T-mediated stimulation of bone formation and
demonstrate the target engagement of the 17β-HSD2 inhibitor.
These data also demonstrate that despite a lack of change in
the bending stiffness of the fractured bone after 14 days
(Figure 1), biochemical changes do occur (i.e., increase in
OPG expression) within the first 14 days of treatment with 15.
The process of remodeling in the 15 group is apparently not
completed at day 28 (compared to the vehicle group), as
indicated by the still larger callus, which might be because of
both these underlying E2- and T-mediated processes.
Figure 1. Effect of compound 15 on the bending stiffness of the
fractured femurs of C57BL/6 male mice after 14 and 28 days of
treatment. **indicates P < 0.01 in bone bending stiffness between the
vehicle- and 15-treated animals at 28 days (Student’s t test).
periosteal area of the callus (i.e., the area that represents the
newly formed tissue around the fracture) (Figure 2A and Table
S1). However, after 28 days, the bending stiffness of the
fractured bones in animals treated with 15 was almost twice as
high as the bending stiffness of the fractured bones of vehicle-
treated animals (P = 0.005) (Figure 1 and Table S1). The
strong increase in bending stiffness was accompanied with a
61% larger periosteal area of the callus (P = 0.026) in 15-
treated animals at 28 days (Figure 2A and Table S1).
Representative photographs of the callus of healing fractured
femurs after 28 days of treatment with 15 and vehicle are
shown in Figure 2B.
The periosteal area in the callus of 15-treated animals
consisted of a larger bone area than in the vehicle-treated
D
DOI: 10.1021/acs.jmedchem.8b01493
J. Med. Chem. XXXX, XXX, XXX−XXX