Synthesis of potent, substituted carbazoles as selective androgen receptor modulators (SARMs)

Article abstract of DOI:10.1016/j.bmcl.2010.09.140

The synthesis and in vitro binding affinity for a novel series of potent androgen receptor modulators is described. One of the more potent compounds (17, RAD35010) was further characterized in vivo where it restored levator ani weight in castrated male ra

Full text of DOI:10.1016/j.bmcl.2010.09.140

Bioorganic & Medicinal Chemistry Letters 20 (2010) 7516–7520
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Synthesis of potent, substituted carbazoles as selective androgen receptor
modulators (SARMs)
Chris P. Miller a, , Pushpal Bhaket b, Nagarajan Muthukaman b, C. Richard Lyttle a, Maysoun Shomali a,
⇑
Kyla Gallacher a, Connie Slocum a, Gary Hattersley a
a Radius Health, Inc., 300 Technology Square, Cambridge, MA 02139, USA
b Sai Advantium Pharma Ltd, 8-2-120/86/9/B ‘Luxor Park’ Road No. 2, Banjara Hills, Hyderabad, AP 500033, India
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 29 July 2010
Revised 27 September 2010
Accepted 28 September 2010
Available online 7 October 2010
The synthesis and in vitro binding affinity for a novel series of potent androgen receptor modulators is
described. One of the more potent compounds (17, RAD35010) was further characterized in vivo where
it restored levator ani weight in castrated male rats to near sham level while having no significant effect
on prostate weight.
Ó 2010 Elsevier Ltd. All rights reserved.
Keywords:
Androgen
SARM
Carbazole
Tetrahydrocarbazole
Prostate
Muscle
The androgen receptor is a ligand-gated transcription factor and
is a member of the steroid hormone nuclear receptor family.1
Endogenous androgen signaling plays a pivotal role in male sexual
development and function and also plays a critical role in male and
female sexual behavior. Beyond its role in sexual development,
function and behavior, androgen signaling promotes nitrogen
retention, protein synthesis and muscle growth. Androgen signal-
ing also has beneficial effects on bone through increasing bone
density, perhaps by a combination of bone resorption inhibition
and anabolic bone stimulation.2 Due to the many beneficial though
pluripotent effects of androgens, intense research dedicated to-
wards the understanding of AR-signaling pathways has been
underway for some time. For example, it has been known for quite
some time that the primary endogenous androgen in humans is
testosterone. However, androgen signaling by testosterone is com-
male-pattern baldness. Furthermore, testosterone is also subject to
aromatization to estradiol under the influence of the cytochrome
P450 CYP19 aromatase enzyme. Estradiol is the most important
endogenous estrogen and drives a host of estrogen receptor (‘ER’)
mediated activities, some of which may also affect prostate cell
proliferation.3
The complexity of testosterone metabolism and signaling is not
surprising given its diverse roles in sexual development and main-
tenance in both males and females. Despite its drawbacks, testos-
terone has enjoyed reasonable success in male hormone
replacement though being generally limited to topical gels or intra-
muscular injection. In recognition of its utility and limitations,
medicinal chemists have for quite some time been trying to syn-
thesize alternatives to testosterone with more desirable profiles.
In this communication, we report on some of our own research
dedicated to the discovery of androgens with preferable profiles to
testosterone. In particular, we desired compounds that were novel,
orally active and that would not be subject to the same metabolic
issues as testosterone (i.e., prostate selective activation, conversion
to estrogens). Our endeavors were based on structure-based design
wherein we scanned a number of core templates. In the course of
our investigations, we found that certain appropriately substituted
carbazoles bound the androgen receptor with good affinity. Our
initial structural carbazole lead 4 is shown in Figure 1. The com-
pound demonstrated moderate affinity in our binding assay rela-
tive to, for example, testosterone 1, dihydrotestosterone 2 or the
plicated because testosterone is site specifically metabolized to 5a-
dihydrotestosterone (‘DHT’). DHT is a more potent androgen than
testosterone and because testosterone is largely converted to
DHT in the prostate, androgen signaling through testosterone is
amplified in this tissue. Overstimulation of the prostate by DHT
is believed to contribute to such pathologies as benign prostate
hypertrophy and possibly prostate cancer. The scalp is another tis-
sue where testosterone to DHT metabolism can occur, accelerating
⇑
Corresponding author.
E-mail address: cmiller@radiuspharm.com (C.P. Miller).
0960-894X/$ - see front matter Ó 2010 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2010.09.140

C. P. Miller et al. / Bioorg. Med. Chem. Lett. 20 (2010) 7516–7520
7517
OH
We began our optimization program with the goal of increasing
binding affinity through exploring structural diversity within the
carbazole template motif of compound 4. We briefly explored
the optimization of the left hand side ring that initially contained
the 4-bromo substituent. Schemes 1 and 1a describes the chemis-
try that was used to generate these analogues. The chemistry used
to prepare the carbazoles having structures 4–7 was initiated by a
Fisher indole synthesis of the appropriately substituted aryl hydra-
zide and 1,2-cyclohexanedione. For the dichloro-substituted com-
pounds 6–7 (Scheme 1a), a mixture of two regioisomers was
obtained from the Fisher indole synthesis in an approximate ratio
of 1:1. In order to separate the regioisomers, the N-Boc derivatives
were prepared, the compounds separated by silica gel chromatog-
raphy and the Boc groups subsequently removed by treating with
trifluoroacetic acid in methylene chloride. The tetrahydro carba-
OH
OH
O
O
O
H
1
2
3
H
N
OH
Br
4
Figure 1. Structures of testosterone 1, dihydrotestosterone 2, RU1881 3 and lead
carbazole 4.
zole 1-ones were aromatized by first
with CuBr2 followed by elimination with Li2CO3 to generate the
a-brominating the 1-ones
Table 1
Relative binding affinities of three steroidal androgens and carbazole lead structure 4
desired compounds. Unfortunately, the attempt to brominate the
6,7-dicloro tetrahydrocarbazole 1-one resulted in the a,
a0-dibromo
Compds
Binding affinity IC50, nM
Repeats; sd
ketone that upon elimination produced the 2-bromophenol 6.
Compounds 8 and 9 were prepared by base-catalyzed aldol
condensation with benzaldehyde to generate the exo-styrene
derivatives. Since attempts at isomerizing the double bond to the
endo-position were not successful, a more circuitous procedure
1
2
3
4
30
10
7
n = 1
n = 1
n = 7; sd = 4
n = 1
1420
comprising reduction of the double bond, followed by a-keto bro-
mination and elimination was performed. The final products 8 and
9 completed our work with the fully aromatic carbazoles. The
binding data for the fully aromatic carbazoles 4–9 together with
the known steroid comparators are shown in Table 2.
potent synthetic steroidal agonist R1881 3 (see Fig. 1 for structures
and Table 1 for binding data). We used a fluorometric binding as-
say for all of the binding data discussed in this paper.4
As can be seen from the data in Table 2, some high affinity com-
pounds were obtained. We knew from the literature as well as our
own related work that a hydroxyl group was important for good
binding affinity in many SARM templates, putatively for forming
hydrogen bonds with one or two of the amino acids that the testos-
terone 17b-hydroxyl binds in the androgen receptor (i.e., N705/
T877). Similarly, the importance of at least one hydrogen bond
acceptor to mimic the 3-carbonyl group of testosterone and related
steroids due the presence in the AR of hydrogen bond donors (i.e.,
Q711/R752) has been explained previously.5 It appeared from our
preliminary work that compounds substituted with two chlorines,
either at the 5,6- or 6,7-positions (6 and 7) led to high affinity
compared to mono substituted compounds such as 4 and 5.
While we assumed that the phenolic group on the aromatic
1-position was acting as a good hydrogen bond donor/acceptor,
H
N
O
O
H
N
O
NH2
a
+
R3
R3 = Br or CN
R3
R3 = Br or CN
H
N
OH
b, c
R3
4 R3 = Br
5 R3 = CN
Scheme 1. Synthesis of 4–5. Reagents: (a) concd HCL/AcOH; (b) CuBr2, CHCl3/
EtOAc; (c) LiBr, Li2CO3, DMF.
H
N
H
N
O
O
O
H
O
Cl
Cl
N
Cl
Cl
a
NH2
+
+
Cl
Cl
Boc
N
Boc
N
O
O
e, f
d
Cl
Cl
+
Cl
Cl
H
N
H
N
O
OH
R2
Cl
R2
Cl
R5
b, c
R4
R4
R2 = Cl, R4 = H, R5 = Br
6
7
R2 = H, R4 = Cl, R5 = H
Scheme 1a. Synthesis of 6–7. Reagents: (a) concd HCL/AcOH; (b) CuBr2, CHCl3/EtOAc; (c) LiBr, Li2CO3, DMF; (d) (Boc)2, DMAP, THF; (e) Separate isomers by silica gel
chromatography; (f) TFA, CH2Cl2.

7518
C. P. Miller et al. / Bioorg. Med. Chem. Lett. 20 (2010) 7516–7520
Table 2
we wondered whether it could be replaced by an aliphatic hydro-
xyl group and the resultant compounds still retain good affinity.
Fortunately, we also had been working on preparing a series of
compounds where the c-ring of the carbazole was saturated and
contained an aliphatic hydroxyl group at the same 1-position as
the phenol of compounds 4–9. The saturated hydroxyl compounds
were relatively easy to prepare since they used the same 1-oxo
intermediates that were used in the preparation of compounds
4–9. The synthesis of saturated carbazoles is shown in Schemes 3
and 3a. The chemistry for analogues 10–20 was similar to that used
to perform the aromatized compounds of Schemes 1, 1a and 2. The
Binding affinity for steroidal androgen standards 1–3 and carbazole analogues 4–9
Compds
Binding affinity IC50, nM
Repeats; sd
1
2
3
4
5
6
7
8
9
30
10
7
1420
4600
34
15
>1000
>1000
n = 1
n = 1
n = 7; sd = 4
n = 1
n = 1
n = 1
n = 1
n = 1
n = 1
primary difference was that instead of aromatizing the 1-one by a-
bromination, the desired compounds were prepared by nucleo-
philic addition of –CH3, –CF3 or –CF2CF3 to the ketone carbonyl.
The only real problem in the synthesis for us was that attempts
to add TMSCF3 (but not MeMgBr) to the carbonyl in the presence
of an unprotected carbazole nitrogen failed when there was no hal-
ogen at the 8-position (compounds 16–20). We solved this prob-
lem by tosylating the nitrogen prior to the nucelophilic addition
and then removing it after. The size of the cycloalkanol ring was
varied from C5 to C7 by varying the size of the cycloalkanone used
in the carbazole formation step. While the cycloheptanone fused
compound used to prepare compound 12 could be prepared
from the 1,2-dione in an analogous fashion as the method used
for the cyclohexanone compounds, the cyclopenta-fused com-
pound 22 was prepared by reaction of in situ generated 2-oxocyc-
lopentanecarboxylic acid and 2,4-dichlorobenzenediazonium
chloride followed by cyclization of the resulting hydrazone to form
the desired fused cyclopentanone and subsequent protection with
chloromethyl pivalate, reaction with CF3TMS and CsF and
deprotection to afford the desired product 22 (Scheme 3a). The
H
N
H
N
O
O
R2
R2
Cl
g, h
b, c
Cl
R4
R4
R2 = Cl, R4 = H
R2 = Cl, R4 = H
R2 = H, R4 = Cl
R2 = H, R4 = Cl
H
OH
N
R2
Cl
R4
8
9
R
2 = Cl, R4 = H
R2 = H, R4 = Cl
Scheme 2. Synthesis of 8–9. Reagents: (b) CuBr2; (c) LiBr, Li2CO3; (g) benzaldehyde,
K2CO3; (h) Pd/C, H2.
H
N
Cl
O
O
Cl
H
N
O
b
a
NH2
(CH2)n
+
(CH2)n
R3
R3
R4
R4
n = 1 or 2
n = 1 or 2
R3 = Cl, R4 = H
3 = H, R4 = Cl
R3 = F, R4 = H
R
3 = Cl, R4 = H
R
R3= H, R4 = Cl
R3 = F, R4= H
H
N
F3C
Cl
OH
(CH2)n
R3
R4
10 R3 = Cl, R4= H, n = 1
11 R3 = H, R4 = Cl, n = 1
3
= Cl, R4 = H, n = 2
= F, R4 = H, n = 1
12
13
R
R
3
H
O
H
N
Rx
OH
N
R2
R3
R2
R3
c, or d, b, e
or d, f, e
R4
R4
14 R2= Cl, R3 = Cl, R4 = H, Rx = CH3
15 R2= H, R3 = CN, R4 = H, Rx = CH3
16
17
R
R
2 = H, R3 = CN, R4 = H, Rx = CF3
2
= Cl, R3 = Cl, R4 = H, Rx = CF3
18 R2 = H, R3 = Cl,R4 = Cl, Rx = CF3
19 R2 = Cl, R3 = Cl, R4 = H, Rx = CF2CF3
20 R2 = H, R3= CN, R4 = H, Rx = CF2CF3
Scheme 3. Reagents: (a) Concd HCL/AcOH; (b) TMSCF3, CsF, THF; (c) MeMgBr, THF; (d) Tos–Cl, NaOH, cat benzyltriethylammonium chloride; (e) KOH; (f) CF3CF2TMS, CsF.

C. P. Miller et al. / Bioorg. Med. Chem. Lett. 20 (2010) 7516–7520
7519
H
N
H
N
O
F3C
Cl
Cl
OH
g, b
F
F
Cl
Cl
21
Cl
Cl
h
Cl
Cl
N2+Cl-
Cl
CO2H
j, k
NH2
Cl
CO2Me
OH
CF3
l, b, m
O
N
H
22
N
H
i
Cl
O
O
Cl
Scheme 3a. Reagents: (b) TMSCF3, CsF, THF; (g) LiHMDS, N-fluorobenzenesulfonamide; (h) NaNO2, HCl; (i) NaOH then HCl; (j) Mix and collect precipitate; (k) H2SO4, MeCN;
(l) chloromethyl pivalate, K2CO3; (m) KOH, THF/H2O.
(or a combination of one or more of these factors). The cyclohep-
Table 3
tane-fused compound 12 demonstrates significantly reduced bind-
ing affinity (600 nM) compared to the cyclohexane-fused
compound 10 (34 nM) having the same substitution pattern
whereas the cyclopenta-fused compound 22 had retained high
affinity (13 nM).
Binding affinity for steroidal androgen standards 1–3 and carbazole analogues 10–22
Compds
Binding affinity IC50, nM
Repeats; sd
1
2
30
10
n = 1
n = 1
3
10
11
7
34
85
n = 7; sd = 4
n = 3; sd = 33
n = 1
Several of the compounds were tested for oral activity in the rat
Herschberger assay where male rats were castrated seven days
prior to the initiation of dosing.6,7 The animals were dosed four
days with drug (po, 0.5% methylcellulose in Tween 80) and sacri-
ficed 24 h after the last dose. The levator ani muscle and prostate
were excised and the weight of the wet tissue recorded. In the
Herschberger assay, positive anabolic effects can be detected by in-
creased levator ani bulbercavernosus (LABC) muscle weight and
positive androgenic effects detected by increased prostate weight.
Since the stimulation of muscle is the desired aspect of these com-
pounds and stimulation of the prostate is not desired, the preferred
compounds are those that increase levator ani weight more than
prostate weight, relative to the non-castrated control rats or tes-
tosterone-treated rats (which we dosed as testosterone propionate
subcutaneously at 1 mg/kg per day). Only compound 17
(RAD35010), which was tested as a racemic mixture, demonstrated
significant effects on the rat LABC muscle. As can be seen from
Figure 2, compound 17 (RAD35010) demonstrated sufficient effi-
cacy on muscle to restore the muscle weight back to near sham
level while showing little or no effect on the prostate. In compari-
son, testosterone propionate was very effective at restoring muscle
but also showed highly androgenic effects on the prostate. In the
context of anabolic hormone therapy (e.g., hormone replacement
therapy in males), a profile like that of compound 17 is very
attractive.
12
13
600
90
n = 1
n = 1
14
15
16
>10,000
9200
480
27
95
34
>10,000
100
13
n = 1
n = 1
n = 1
n = 4; sd = 22
n = 1
n = 1
n = 1
n = 1
n = 1
17 (RAD35010)
18
19
20
21
22
Figure 2. Effects of compound 17 on weight of prostate and levator ani muscle. *All
tissue weights normalized to a 100 g rat. No significant effects of RAD35010 (po
References and notes
30 mg/kg) on prostate relative to castrated control (
is significantly greater than vehicle ( <0.05). Testosterone propionate and sham
(intact) group significantly increased both tissue weights relative to vehicle
<0.05). The effect of 35010 is significantly less than sham on both prostate and
muscle (p <0.05).
q >0.05) but levator ani weight
q
1. Klocker, H.; Gromoll, J.; Cato, A. C. B. In Testosterone Action Deficiency
Substitution; Nieschlag, E., Behre, H., Eds., 3rd ed.; Cambridge University Press,
2004. p 39.
(q
2. Gao, W.; Dalton, J. Drug Discovery Today 2007, 12, 241.
3. (a) Carruba, G. J. Cell. Biochem. 2007, 102, 899; (b) Nakhla, A. M.; Romus, N. A.;
Rosner, W. J. J. Biol. Chem. 1997, 272, 6838.
a,
a0-difluoroanalogue 21 was prepared by electrophilic
fluorination prior to the CF3 addition step (Scheme 3a).
4. The AR-binding assay was performed as specified by the manufacturer
(Invitrogen, Madison, WI). Briefly, 1 ll of 10 mM compound was added to
500
l
l of AR screening buffer in a 1.5 ml eppendorf tube to make a 2 ꢀ 10ꢁ5
M
The binding data for all of the saturated compounds 10–22 is
shown in Table 3 and a number of trends can be gleaned from this
data. A comparison between the methyl adduct 14 and the trifluo-
romethyl adduct 17 indicates the criticality of the fluorine atoms
for good binding affinity. The fluorines might be affecting the pKa
of the hydroxyl hydrogen to make it more acidic such that it is clo-
ser to the pKa of the phenols of the compounds of Table 1. Alterna-
tively, the fluorine atom may be acting directly as hydrogen bond
acceptors or increasing affinity through increased hydrophobicity
stock. Tenfold serial dilutions of the test compounds were prepared ranging in
concentration from 10ꢁ5 M to 10ꢁ12 M. Each dilution was added in triplicate to a
black 384-microtiter plate. The test compounds are diluted two-fold in the final
reaction. 2ꢀ AR-Fluormone™ complex was prepared with 2 nM Flourmone AL
Green™ and 30 nM AR. Twenty-five microlitre of 2ꢀ complex was aliquoted to
each reaction well, such that the final reaction volume was 50 ll per well. Plate
was sealed with a foil cover and incubated in the dark at room temperature for
4 h. Polarization values for each well were measured. The polarization values
were plotted against the concentration of the test compound. The concentration
of the test compound that results in half-maximum shift equals the IC50 of the
test compound. As a control, a competition curve for R1881 (methyltrienolone)

7520
C. P. Miller et al. / Bioorg. Med. Chem. Lett. 20 (2010) 7516–7520
was performed for each assay. Curve Fitting was performed using GraphPad
291615, Teklad, Madison, WI) and water. Rats were anesthetized and
orchidectomized (GDX) or sham surgery (SHAM) was performed. After
PrismÒ software from GraphPad™ Software Inc.
a
5. Sun, C.; Robl, J. A.; Wang, T. C.; Huang, Y.; Kuhns, J. E.; Lupisella, J. A.; Beehler, B.
C.; Golla, R.; Sleph, P. G.; Seethala, R.; Fura, A.; Krystek, S. R., Jr.; An, Y.; Malley, M.
F.; Sack, J. S.; Salvati, M. E.; Grover, G. J.; Ostrowski, J.; Hamann, L. G. J. Med.
Chem. 2006, 49, 7596.
6. Herschberger, L. G.; Shipley, E. G.; Meyer, R. K. Proc. Soc. Exp. Biol. Med. 1953, 83,
175.
7. Immature Sprague–Dawley male rats were obtained Charles River Laboratories
(Stoneridge, NY). All animals were maintained in a temperature and humidity
controlled room with a 12 h light:12 h dark cycle, with ad lib access to food (TD
seven-day recovery period, the animals were randomized according to weight
and assigned to treatment groups (n = 5), SHAM, OVX + vehicle, OVX + Cpd
treated. Testosterone propionate (TP 1 mg/kg in 5% DMSO/95% corn oil) was
administered by once daily subcutaneous injections, while the compound was
dosed in vehicle (0.5% carboxymethylcellulose) and administered by once daily
oral gavage. The rats were then dosed once daily for four days. All animals were
euthanized via carbon dioxide inhalation 24 h after the last dose. The prostate
and levator ani bulba cavernous (LABC) tissues were removed, weighed and
recorded.