Discovery and structure-activity relationships of new steroidal compounds bearing a carboxy-terminal side chain as androgen receptor pure antagonists

Article abstract of DOI:10.1016/j.bmcl.2007.07.090

Lead optimization of CH4892280 (4), an androgen receptor (AR) pure antagonist, was investigated. Compounds 6 and 7, which have a carboxylic acid at the end of the side chain at the position 7α of dihydrotestosterone (DHT), showed partial agonistic activities in reporter gene assay (RGA). Conversion of the steroidal core structure to 17α-methyltestosterone gave compound 14, which showed weak pure antagonistic activity. Optimization of the side chain by the insertion of a phenyl ring led to compounds 22 and 28-30, which showed pure antagonistic activities at submicromolar concentrations. The structure-activity relationships were clarified.

Full text of DOI:10.1016/j.bmcl.2007.07.090

Bioorganic & Medicinal Chemistry Letters 17 (2007) 5573–5576
Discovery and structure–activity relationships of new steroidal
compounds bearing a carboxy-terminal side chain
as androgen receptor pure antagonists
Kazutaka Tachibana,* Ikuhiro Imaoka, Hitoshi Yoshino, Nobuaki Kato,
Mitsuaki Nakamura, Masateru Ohta, Hiromitsu Kawata, Kenji Taniguchi,
Nobuyuki Ishikura, Masahiro Nagamuta, Etsuro Onuma and Haruhiko Sato
Fuji Gotemba Research Laboratories, Chugai Pharmaceutical Co. Ltd., 1-135 Komakado, Gotemba, Shizuoka 412-8513, Japan
Received 16 April 2007; revised 26 July 2007; accepted 30 July 2007
Available online 22 August 2007
Abstract—Lead optimization of CH4892280 (4), an androgen receptor (AR) pure antagonist, was investigated. Compounds 6 and 7,
which have a carboxylic acid at the end of the side chain at the position 7a of dihydrotestosterone (DHT), showed partial agonistic
activities in reporter gene assay (RGA). Conversion of the steroidal core structure to 17a-methyltestosterone gave compound 14,
which showed weak pure antagonistic activity. Optimization of the side chain by the insertion of a phenyl ring led to compounds
22 and 28–30, which showed pure antagonistic activities at submicromolar concentrations. The structure–activity relationships were
clarified.
Ó 2007 Elsevier Ltd. All rights reserved.
Prostate cancer is the most common cancer amongst
men in the USA and the second most common malig-
nant cause of male death worldwide after lung cancer.1
Since the growth of prostate cancer is dependent on
androgen, androgen receptor (AR) antagonists such as
flutamide (1) and bicalutamide (3) (Chart 1) are cur-
rently used as hormone therapy.2 These antiandrogens
exhibit good efficacy in many cases and comprise an
important part of effective therapeutics.3–6 However,
the most considerable problem with these antiandrogens
is that recurrence occurs after a short period of re-
sponse.7 Since they have partial agonistic activities at
high concentration in vitro,8 this may be attributed to
recurrence. Therefore, the search for new antiandrogen-
ic agents that exhibit no agonistic activities, so-called
‘AR pure antagonists’, has been conducted.9,10
methyltestosterone which is known to possess oral activ-
ity. Recently our laboratory discovered that estradiol
derivatives bearing a carboxy-containing side chain were
pure antiestrogens and found that the carboxy moiety is
effective for oral absorption.12,13 These results encour-
aged us to investigate the effects of a carboxylic group
in our AR pure antagonists.
We here report the discovery of new steroidal com-
pounds bearing a carboxy-containing side chain in the
position 7a of 17a-methyltestosterone as AR pure
antagonists and their structure–activity relationships.
The 7a-substituted dihydrotestosterone (DHT) deriva-
tives 5–7 in Table 1 were prepared by the procedures de-
scribed in the patent14 and a previous report.11 Synthesis
of the 7a-substituted 17a-methyltestosterone derivatives
14 and 21–30, outlined in Schemes 1 and 2, was per-
formed according to the procedures described in the
patent.15
We previously reported that CH4892280 (4) (Chart 1)
exhibited AR pure antagonistic activities using reporter
gene assay (RGA).11 However, it showed no in vivo
activities because of metabolic instability. In the lead
optimization we decided to convert a scaffold to 17a-
The synthesized compounds were evaluated for their
in vitro binding affinities and agonist/antagonist activi-
ties for AR using the same procedures as described in
the previous report.11 The binding affinity for AR was
determined by displacement of [3H]-mibolerone with
the test compound utilizing CHO-K1/hAR cells. The
Keywords: Androgen receptor; Pure antagonist; Prostate cancer.
*
Corresponding author. Tel.: +81 0550 87 6726; e-mail:
tachibanakzt@chugai-pharm.co.jp
0960-894X/$ - see front matter Ó 2007 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2007.07.090

5574
K. Tachibana et al. / Bioorg. Med. Chem. Lett. 17 (2007) 5573–5576
Initially, we introduced a carboxy-containing side chain
OH
H
N
H
N
R
O
S
O
O N
NC
2
in position 7a of DHT to compare the effect on AR ago-
nistic and antagonistic activities with that of the side
chain of CH4892280 (Table 1). Compound 5, which
has a side chain similar to the pure antiestrogens of
our laboratory,13 exhibited full agonistic activity
(FI5 = 13 nM) with no antagonistic activity. The unex-
pected discrepancy of results between AR antagonists
and ER antagonists was also found in the compounds
with a side chain containing a perfluoroalkyl-sulfoxide
moiety in our previous report.11 Such a long hydropho-
bic moiety might cause full agonistic activities in AR,
although the mode of interaction is unknown. This spec-
ulation is supported by the results of compounds 6 and
7, in which the hydrophobic termini were removed,
showed antagonistic activities. However, these com-
pounds still retained partial agonistic activities.
F
O
O
F C
3
F C
3
flutamide (1)
bicalutamide (3)
(R=H)
hydroxyflutamide (2)
(R=OH)
OH
CONMe
2
O
CH4892280 (4)
Chart 1. Structures of AR antagonists.
agonistic and antagonistic activities of the compounds
for AR were determined by RGA using hAR-transfec-
ted Hela cells. Antagonistic activity was described as
the IC50 value, the concentration of a compound that
inhibits the transcriptional activity of 0.1 nM of DHT
by 50%. To determine agonistic activity, we calculated
the value of ‘FI5’, the concentration of a compound-
treated group in which the transcriptional activity is five
times the transcriptional activity of a group without the
addition of a compound. A ‘pure antagonist’ was de-
fined to have an FI5 value greater than 10,000 nM.
For the next step, the DHT scaffold of 6 was replaced
with 17a-methyltestosterone to give compound 14 (Ta-
ble 2). This compound showed no agonistic activity
(FI5 > 10,000 nM) although compound 6 exhibited ago-
nistic activity with FI5 of 120 nM. To investigate this
remarkable change in agonistic activity by the conver-
sion of the steroidal core structure, a 3D model of these
compounds bound to AR was constructed as shown in
Figure 1. This model was built based on the X-ray crys-
tal structure of human AR in complex with the ligand
Table 1. Binding and agonistic/antagonistic activities of 7a-substituted DHT derivativesa
OH
O
R
Compound
R
Binding affinity IC50 (nM)
RGA
FI5 (nM)
IC50 (nM)
5
(CH2)8CH(COOH)(CH2)3C2F5
410
350
660
260
200
200
13
120
>10,000
920
730
6
(CH2)9-COOH
(CH2)7-COOH
(CH2)7-CONMe2
—
7
110
4 (CH4892280)
2 (hydroxyflutamide)
3 (bicalutamide)
>10,000
1000
770
190
31
—
140
a All data are mean values of duplicate experiments.
OH
OH
OTBS
a, b, c
O
d
e
HO
OH
PivO
O
PivO
8
10
9
11
OH
g, h, i
f
(CH ) COOMe
2 6
(CH ) COOMe
2 6
HO
O
(CH ) COOH
3 9
12
13
14
Scheme 1. Reagents and conditions: (a) NaBH4, MeOH, rt; (b) PivCl, NEt3, CH2Cl2, rt; (c) 2N-HCl, acetone, rt; (d) nPr4NRuO4, N-
methylmorpholine, MS4A, CH2Cl2, rt; (e) MeLi, THF, À78 °C; (f) Cl2(Cy3P)2Ru@CHPh (cat.), CH2Cl2, reflux; (g) H2, 10% Pd/C (cat.), AcOEt, rt;
(h) MnO2, acetone, rt; (i) 2N-NaOH, H2O, MeOH, rt.

K. Tachibana et al. / Bioorg. Med. Chem. Lett. 17 (2007) 5573–5576
5575
OH
OH
OH
b, c, d
a
OAc
OH
(CH2)m-3
HO
a
HO
O
(CH2)m
OAc
17
16
11
15
e, f
O(CH2)nCOOEt
(CH2)m-3
18
OH
OH
OH
b, c
f
O
OR
(CH2)nCOOH
HO
O
(CH2)m
20 [R=(CH2)nCOOEt]
O
(CH2)m
OR
19 [R=(CH2)nCOOEt]
21~30
Scheme 2. Reagents and conditions: (a) Cl2(Cy3P)2Ru@CHPh (cat.), CH2Cl2, reflux; (b) H2, 10% Pd/C (cat.), AcOEt, rt; (c) MnO2, acetone, rt;
(d) 2N-NaOH, H2O, MeOH rt; (e) Br(CH2)nCOOEt, K2CO3, 18-crown-6-ether, N,N-dimethylacetamide, rt; (f) 2N-NaOH, THF, rt.
R1881 (PDB ID: 1e3g).16 Compounds 6 and 14 were
docked into AR manually. C terminal residues from
Glu893 to Thr918 of AR were deleted because these res-
idues collide significantly with the compounds. The
model suggests that the 17a-methyl group of compound
14 is located in a position where Helix 12 would be
folded to express the agonistic activity. Therefore, it
could have the additional effect on a side chain in posi-
tion 7a of preventing the folding of Helix 12.
Although compound 14 exhibited pure antagonistic
=
activity, its potency was relatively low (IC50
2000 nM). Therefore, compounds 21–30 were synthe-
sized for optimization of the side chain by the insertion
of a phenyl ring (Table 2). A phenyl ring could increase
antagonistic activities by decreasing the flexibility of the
side chain. Furthermore, it is expected to interact with
Figure 1. Binding models of compounds 6 (white) and 14 (cyan) to
AR. Helix 12 of AR was removed in the model.
Table 2. Binding and agonistic/antagonistic activities of 17a-methyltestosterone derivativesa
OH
O
R
Compound
R
Binding affinity IC50 (nM)
RGA
FI5 (nM)
>10,000
IC50 (nM)
2000
14
–(CH2)9–COOH
600
m
O
p
(CH2)n COOH
(CH2)m
m
3
3
3
3
4
4
4
4
4
4
Position
p
n
2
3
4
3
3
4
5
3
4
5
21
22
23
24
25
26
27
28
29
30
300
430
660
NTb
NTb
96
NTb
670
350
120
2500
170
660
300
5800
850
250
420
920
330
530
p
>10,000
3200
1500
p
m
p
170
2000
p
p
100
m
m
m
>10,000
>10,000
>10,000
a All data are mean values of duplicate experiments unless otherwise noted.
b Not tested.

5576
K. Tachibana et al. / Bioorg. Med. Chem. Lett. 17 (2007) 5573–5576
Trp741, which exists in the proximity of the region
where Helix 12 would be folded to express agonistic
activity, resulting in potent antagonistic activity (Fig. 2).
Further optimization would be necessary to acquire
sufficient metabolic stability for in vivo activity.
In summary, we have discovered new steroidal com-
pounds bearing a carboxy-containing side chain that ex-
hibit AR pure antagonistic activities at submicromolar
concentrations. The structure–activity relationships of
the compounds were also clarified. It appears that their
agonistic/antagonistic activities depend on the structure
of the side chain. These findings could be helpful for fur-
ther investigations in the future.
For compounds 21–23 (m = 3, position = para), antago-
nistic activities increased 3- to 12-fold compared to that
of compound 14. In the case of compound 24 (m = 3,
position = meta), the antagonistic activity was lower
than that of compound 14.
Although compounds 25–27 (m = 4, position = para)
exhibited antagonistic activities with IC50s of 250–
850 nM, they tend to show higher partial agonistic activ-
ities compared to other types of compounds. On the
contrary, compounds 28–30 (m = 4, position = meta)
exhibited no agonistic activities even at 10,000 nM,
retaining the antagonistic activities with IC50s of 330–
920 nM. These results may suggest that the substitution
position of the terminal carboxyalkyloxy group corre-
lates with the antagonistic activity of compounds 21–
24 (m = 3) and with agonistic activity of compounds
25–30 (m = 4).
Acknowledgment
The authors thank Ms. Frances Ford of Chugai
Pharmaceutical Co., Ltd for assistance with English usage.
References and notes
1. Landis, S. H.; Murray, T.; Bolden, S.; Wingo, P. A. CA.
Cancer J. Clin. 1999, 49, 8.
2. Stewart, A. B.; Lwaleed, B. A.; Douglas, D. A.; Birch, B.
R. Curr. Med. Chem. Anticancer Agents 2005, 5, 603.
3. Kennealey, G. T.; Furr, B. J. Urol. Clin. North Am. 1991,
18, 99.
4. Blackledge, G.; Kolvenbag, G.; Nash, A. Anticancer
Drugs 1996, 7, 27.
5. Tyrrell, C. J.; Altwein, J. E.; Klippel, F.; Varenhorst, E.;
Lunglmayr, G.; Boccardo, F.; Holdaway, I. M.; Haefliger,
J. M.; Jordaan, J. P. J. Urol. 1991, 146, 1321.
6. Eisenberger, M. A.; Blumenstein, B. A.; Crawfold, E. D.;
Miller, G.; McLeod, D. G.; Leohrer, P. J.; Wilding, G.;
Sears, K.; Culkin, D. J.; Thompson, I. M., Jr.; Bueschen,
A. J.; Lowe, B. A. N. Engl. J. Med. 1998, 339, 1036.
7. Wirth, M. P.; Froschermaier, S. E. Urol. Res. 1997, 25,
S67.
Furthermore, the number of methylene groups (n) has
been found to be strongly related to agonistic activities
rather than antagonistic activities in compounds 21–23
(m = 3, position = para) and 25–27 (m = 4, position =
para) since agonistic activity changed considerably by
changing n in an identical type (same m, same position).
Compound 22, one of the pure antagonists discovered in
this investigation, exhibited only slight improvement of
in vitro metabolic stability against CH4892280 and
showed almost no in vivo antiandrogenic activities on
seminal vesicle wet weight in castrated mice (Table 3).17
8. Kemppainen, J. A.; Wilson, E. M. Urology 1996, 48, 157.
9. Ishioka, T.; Tanatani, A.; Nagasawa, K.; Hashimoto, Y.
Bioorg. Med. Chem. Lett. 2003, 13, 2655.
10. Hashimoto, Y.; Miyachi, H. Bioorg. Med. Chem. 2005, 13,
5080.
11. Tachibana, K.; Imaoka, I.; Yoshino, H.; Emura, T.;
Kodama, H.; Furuta, Y.; Kato, N.; Nakamura, M.; Ohta,
M.; Taniguchi, K.; Ishikura, N.; Nagamuta, M.; Onuma,
E.; Sato, H. Bioorg. Med. Chem. 2007, 15, 174.
12. Kanbe, Y.; Kim, M.-H.; Nishimoto, M.; Ohtake, Y.;
Tsunenari, T.; Taniguchi, K.; Ohizumi, I.; Kaiho, S.;
Nabuchi, Y.; Kawata, S.; Morikawa, K.; Jo, J.-C.; Kwon,
H.-A.; Lim, H.-S.; Kim, H.-Y. Bioorg. Med. Chem. Lett.
2006, 16, 4090.
13. Kanbe, Y.; Kim, M.-H.; Nishimoto, M.; Ohtake, Y.;
Yoneya, T.; Ohizumi, I.; Tsunenari, T.; Taniguchi, K.;
Kaiho, S.; Nabuchi, Y.; Araya, H.; Kawata, S.; Morik-
awa, K.; Jo, J.-C.; Kwon, H.-A.; Lim, H.-S.; Kim, H.-Y.
Bioorg. Med. Chem. Lett. 2006, 16, 4959.
Figure 2. Binding model of compound 29 (white) to AR. This model
was built in the same manner as for compounds 6 and 14 (Fig. 1). Helix
12 of AR was removed in the model.
14. Kaiho, S.; Ohizumi, I.; Tamura, K.; Kato, N.; Yoneya, T.;
Tachibana, K. PCT Int. Appl. WO 200114406, 2001.
15. Nakamura, M.; Tachibana, K.; Imaoka, I.; Yoshino, H.;
Kawata, S. PCT Int. Appl. WO 2003051903, 2003.
16. Matias, P. M.; Donner, P.; Coelho, R.; Thomaz, M.;
Peixoto, C.; Macedo, S.; Otto, N.; Joschko, S.; Scholz, P.;
Wegg, A.; Basler, S.; Schafer, M.; Egner, U.; Carrondo,
M. A. J. Biol. Chem. 2000, 275, 26164.
Table 3. Antiandrogenic activities of compound 22
Dose (mg/body)
Inhibitiona (%)
1
3
6.7
0.8
6.4
10
a Inhibition of TP (testosterone propionate)-stimulated seminal vesicle
weight gain by subcutaneous administration of compound 22 (n = 4).
17. Hershberger, L. G.; Shipley, E. G.; Meyer, R. K. Proc.
Soc. Exp. Biol. Med. 1953, 83, 175.