Discovery of indole-containing tetracycles as a new scaffold for androgen receptor ligands

Article abstract of DOI:10.1016/j.bmcl.2006.03.047

A novel series of tetracyclic indoles have been designed, synthesized and evaluated as androgen receptor (AR) ligands. Studies of structure-activity relationships (SARs) were investigated, which led to some compounds in this series as strong binders to an

Full text of DOI:10.1016/j.bmcl.2006.03.047

Bioorganic & Medicinal Chemistry Letters 16 (2006) 3233–3237
Discovery of indole-containing tetracycles as a new scaffold
for androgen receptor ligands
Xuqing Zhang,* Xiaojie Li, George F. Allan, Amy Musto,
Scott G. Lundeen and Zhihua Sui
Drug Discovery, Johnson & Johnson Pharmaceutical Research and Development, LLC, 1000 Route 202, Raritan, NJ 08869, USA
Received 1 February 2006; revised 10 March 2006; accepted 14 March 2006
Available online 5 April 2006
Abstract—A novel series of tetracyclic indoles have been designed, synthesized and evaluated as androgen receptor (AR) ligands.
Studies of structure–activity relationships (SARs) were investigated, which led to some compounds in this series as strong binders
to androgen receptors.
Ó 2006 Elsevier Ltd. All rights reserved.
The androgen receptor (AR) is a member of the nuclear
receptor superfamily of ligand-dependent transcription
factors.1 It plays a critical role in numerous physiologi-
cal processes, including the development and mainte-
nance of male secondary sexual characteristics such as
muscle, hair, bone mass, prostate growth, and spermato-
genesis. Two endogenous androgens most active in
promoting these effects are, testosterone and nonaroma-
tizable 5a-dihydrotestosterone (5a-DHT) (Fig. 1). The
primary focus for drug design has been the synthesis
of chemicals to regulate the transcriptional activity of
AR based upon the structural, steroidal or nonsteroidal,
and functional androgenic, antiandrogenic, or anabolic
properties of ligands.2 Nonsteroidal androgens can be
designed and synthesized that will mimic the pharmaco-
logical effects of testosterone, and would likely avoid
many of the undesired physicochemical and pharmaco-
kinetic properties of their steroidal counterparts, includ-
ing poor oral bioavailability, rapid hepatic metabolism,
and activation of other steroid receptors.3 The AR ago-
nists and antagonists are useful in the treatment of a
variety of disorders and diseases.4 More particularly,
antagonists of the androgen receptor could be employed
in the treatment of prostate cancer, benign prostate
hyperplasia, hirsutism in women, alopecia, anorexia
nervosa, breast cancer, and acne. Agonists of the andro-
gen receptor could be employed in male contraception,
male performance enhancement, as well as in the treat-
OH
OH
R2
R3
N
D
C
Y
R1
A
B
O
O
X
H
Testosterone
α
5
-DHT
I
Figure 1.
ment of cancer, AIDS, cachexia, and other disorders.
In our efforts directed at identifying novel AR ligands,
we have examined various chemical scaffolds as core
structural elements.5 Herein, we wish to present our de-
sign and synthesis of a novel series of nonsteroidal tetra-
cycles (I) that incorporate indole moiety as the core to
mimic the tetracyclic alignment of testosterone struc-
ture. SAR studies based on the binding affinity and
the potency in cell functional assay will be described.
The indole core present in the target structures was most
conveniently introduced by a Fisher-indole type of reac-
tion between various substituted phenyl hydrazine HCl
salts (1) and ketones (2) via acid catalysis (Scheme 1).6
Indoles 3 were then N-alkylated or acylated to afford
the corresponding protected adducts 4. Compound 4i
was prepared from 4h by treatment with CuCN in
DMF. The target compounds 4j–t, 4v, 5c–h, and 6a–g
were obtained in accepted yields after de-methylation
under pyridine HCl or AlCl3 and EtSH condition. Sulf-
oxide 5a and sulfone 5b were synthesized through oxone
oxidation of the sulfide 4d followed by de-methylation
under pyridine HCl condition.
Keyword: Androgen receptor ligand.
*
Corresponding author. Tel.: +1 908 904 6451; e-mail: xzhang5@
prdus.jnj.com
0960-894X/$ - see front matter Ó 2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2006.03.047

3234
X. Zhang et al. / Bioorg. Med. Chem. Lett. 16 (2006) 3233–3237
R2
To prepare quinolone 11, we employed a Winterfeldt
O
X
HN
NHNH2. HCl
oxidation procedure (t-BuOK and O2) to achieve the in-
dole-quinolone conversion in good yield (Scheme 2).7
The required tetracyclic indole 10 for this reaction was
prepared by Fisher indole synthesis from hydrazine 8
and ketone 9. Compound 11 was then N- or O-methyl-
ated by treatment with K2CO3 and MeI to afford 12a
and 12b, which were then converted into the phenols
5i and 5j after reductive de-benzylations.
a
R2
Ar
R1
R1
Ar
+
Y
Y
X
3
1
2
R3
R2
d, e (for 5a, 5b)
OH
R3
N
N
b
R1
Ar
R1
Ar
Y
Y
X
X
The synthetic approaches to the indole analogues with
non-aromatized A-ring (7a and 7b) were initiated from
Wieland–Miescher ketone 13 (Scheme 3). Selective pro-
tection of the conjugated ketone group of 13 by 1,2-bis-
trimethylsilanylsulfanyl-ethane and ZnI2 gave 14. Con-
densation of 14 with phenyl hydrazines via Fisher-indole
synthesis afforded the corresponding indoles 15 in
acceptable yields. Compound 15 was then methylated
and the thio-ketal protective group was removed by
4j-4t, 4v, 5a-5h, 6a-6g
b
4a-4h
4h
c
4u
4t
4i
Scheme 1. Reagents and conditions: (a) HCl, EtOH, reflux, 4 h, 35–
73%; (b) NaH, R3I [MeI, EtI, and Cl(CH2)2NMe2 + KI (cat.)], DMF,
0–50 °C, 1 h, 71–87% or Et3N, AcCl, CH2Cl2, 0 °C to rt, 2 h, 54–71%;
(c) CuCN, DMF, 160 °C, 6 h, 45%; (d) pyridine hydrochloride, neat,
200 °C, 30 min, 25–65% or AlCl3, EtSH, 0 °C to rt, 2 h, 50–75%; (e)
oxone, MeOH–H2O, rt, 6 h, 25% (the sulfoxide), 45% (the sulfone) for
preparations of 5a and 5b.
NHNH2. HCl
H
N
H
N
BnO
b
a
BnO
8
BnO
O
+
10
O
11
Me
N
Me
N
d
9
BnO
c
HO
HO
O
12a
O
5i
+
N
N
d
BnO
OMe
5j
OMe
12b
Scheme 2. Reagents and conditions: (a) HCl, EtOH, reflux, 4 h, 52%; (b) t-BuOK, O2, rt, 10 h, 55%; (c) K2CO3, MeI, DMF, rt, 2 h, 35% for 12a, 30%
for 12b; (d) H2, Pd–C, rt, 2 h, 75% for 5i, 79% for 5j.
6
5
3
OMe
4
O
HN
O
b
a
S
S
O
S
S
15a, 4-OMe
15b, 5-OMe
13
14
5
6
6
6
5
5
OMe
OH
4
Me
OMe
4
Me
N
Me
N
N
4
c
d
e
3
3
3
S
O
O
S
7a, 4-OH
7b, 5-OH
17a, 4-OMe
17b, 5-OMe
16a, 4-OMe
16b, 5-OMe
Me
N
O
OH
O
N
Me
O
N
Me
18
7c
Scheme 3. Reagents and conditions: (a) TMSSCH2CH2STMS, ZnI2, Et2O, 10 h, 0 °C, 81%; (b) p-MeO–PhNHNH2ÆHCl, HCl, EtOH, reflux, 4 h, 56–
60%; (c) NaH, MeI, DMF, 0 °C, 1 h, 81%; (d) Hg(ClO4)2, MeOH, rt, 20 min, 61–65%; (e) pyridine hydrochloride, neat, 200 °C, 30 min, 50–72%.

X. Zhang et al. / Bioorg. Med. Chem. Lett. 16 (2006) 3233–3237
3235
the treatment of 16 with Hg(ClO4)2 in MeOH. De-
protection of the methoxy groups provided the targets
7a and 7b. Analogous synthesis from the bicyclic keto
lactam 18 via the similar procedure afforded 7c.
steric bulky group, such as Et, Ac, and dimethylamino-
ethyl groups (4l–n). It was not surprising to observe par-
tial loss of activity in 4k wherein H replaced methyl
group at R3 position. This might be due to the decreased
lipophilicity of the ligand, which was critical for its fit-
ting into the AR binding pocket. Additions of bulky
(4o and 4p) or hydrophilic groups (4t) at R1 position
were detrimental to activity. Replacements of H with
F at R1 position (4q and 4r) were tolerated showing
comparable binding affinities.
The lead optimization was primarily guided by a COS-7
whole-cell androgen receptor binding assay for assessing
the AR binding affinities of the ligands.8 Our initial SAR
studies were focused on modifications of the tetacyclic
core structure at R1, R2, and R3 groups, as shown in
Table 1. Exemplified by 4j, one salient feature of the
series was the phenol group at R2 position. Any modifi-
cation of the phenol group, including deletion (4a),
replacement with halogen (4c and 4e–h) or CN (4i)
and masking as methyl ether (4d) or ester (4u), resulted
in partial or total loss of activity. This indicates that the
phenol group at D-ring might mimic 17b-OH of testos-
terone, which was crucial for hydrogen bonding with the
receptor. 30-OH group at R3 position seemed to promote
higher binding affinity than 10-OH did, as illustrated by
4q versus 4s, albeit 20-OH at R3 position led to equally
potent binder as its 30-OH analogue did (4v vs 4j).
Another important requirement for strong binding affin-
ity was having a methyl group at R3 position. Significant
loss in activity was seen for the compounds with more
We next turned our attention to B- and C-ring modifica-
tions, as shown in Table 2. On the basis of the structure
4j, B-ring was modified by various replacements based
on steric bulkiness or functionality in order to define
the SAR at this area. Attempts to replace sulfide of 4j
with some hydrophilic functionality such as sulfoxide
(5a), sulfone (5b) or amino group (5h) resulted in total
loss of activity. Surprisingly, switching the linkage of
A- and C-ring from (ring A)–S–CH2–(ring C) of 4j to
(ring A)–CH2–S–(ring C) of 5c abolished the binding
affinity. This result indicated that the alignments of the
tetracyclic cores in 4j and 5c resulted in different confor-
mations, which led to total different binding affinities be-
tween these two regio-isomers. Steric bulkiness (4t vs 5d)
at Y seemed to be tolerated for comparable binding
affinity. Compared with 4j, the carbon analogue 5e
bound to the AR receptor with an improved affinity.
Seven-membered B-ring analogues produced a detri-
mental effect, as presented by 5f versus 5e and 5g versus
4j. With regard to C-ring modification, replacement of
Table 1. SAR at the substitutions R1, R2, and R3
1'
2'
R2
3'
R3
N
4
3
4'
R1
S
2
Table 2. SAR at B- and C-rings
1
Compound
R1
R2
R3
Rat AR COS-7
whole-cell bindinga
IC50 (lM)
Me
N
OH
4
3
R1
R1881d
Bicalutamide
0.0015
0.85
Y
X
2
1
b
4a
4b
4c
4d
4e
4f
H
H
Me
Me
Me
—
Compound R1
X
Y
Rat AR COS-7
whole-cell
bindinga
b
H
30-Me
30-Cl
—
1.5
H
b
—
H
30-OMe Me
30-F
30-F
IC50 (lM)
H
2-F
2-CF3 30-F
Me
Me
Me
Me
3.7
48%c
2.8
b
5a
5b
5c
5d
5e
5f
H
SO CH2
SO2 CH2
—
b
4g
4h
4i
H
—
b
30-Br
30-CN Me
30-OH Me
30-OH
30-OH Et
30-OH Ac
—
b
—
H
H
H
H
H
H
H
H
CH2
S
S
b
—
2-OH
H
C(Me)2 3.4
0.03
CH2 (CH2)2 0.92
(CH2)2 5.0
4j
0.11
4.5
2.3
1.0
CH2 CH2
4k
4l
H
H
5g
5h
H
S
b
—
4m
4n
4o
4p
4q
4r
4s
4t
H
NMe CH2
b
30-OH (CH2)2NMe2
—
3.0
OH
OH
2-OMe 30-OH Me
Me
2-Br
2-F
3-F
2-F
30-OH Me
30-OH Me
30-OH Me
10-OH Me
2.8
N
N
b
—
5i
5j
0.63
0.11
3.8
O
2-OH 30-OH Me
2.1
4u
4v
H
H
30-OAc Me
20-OH Me
0.88
0.13
b
—
OMe
a IC50 values are representative of multiple determinations (N = 2–3).
b ‘—’, not active (<20% and/or binding affinity >10 lM).
c Inhibition at 10 lM.
a IC50 values are representative of multiple determinations (N = 2–3).
b ‘—’, not active (<20% and/or binding affinity >10 lM).
d Methyltrienolone.

3236
X. Zhang et al. / Bioorg. Med. Chem. Lett. 16 (2006) 3233–3237
Table 4. SAR at nonaromatized A ring
the indole with the quinolone core resulted in total loss
of activity (5i and 5j) (Table 3).
Compound
Structure
Rat AR Cos-7
whole-cell
bindinga
With the SAR of the compounds containing phenyl as
A-ring understood, we then explored the effect of the
heterocyclic A-ring analogues. These would potentially
have the benefit of improved pharmacokinetic profile.
As illustrated by 6b and 6d in Table 4, replacement of
phenyl A-ring of 4j or 5e with 2-pyridinyl group dis-
played either comparable or better binding affinity.
The best compound 6b bound to the AR receptor with
an IC50 of 29 nM. This result indicated that heteroaryl
replacement was an effective strategy, albeit not all mod-
ifications presented here were rewarded by generating
potent AR ligands (6a, 6c, and 6e–g).
IC50 (lM)
OH
Me
N
N
b
—
7a
7b
7c
O
O
Me
Me
OH
42%
N
OH
The final group of compounds tested were nonaroma-
tized A-ring analogues 7a–c, which acted as a close mim-
ic of testosterone structure. It was noted in these cases
that the unsaturated A-ring seemed to have a detrimen-
tal effect for binding affinity.
b
—
O
N
Me
a IC50 values are representative of multiple determinations (N = 2–3).
b ‘—’, not active (<20% and/or binding affinity >3 lM).
Selected compounds showing high binding affinities
were then evaluated in cell-based functional assay
(Table 5). These compounds were somewhat less active
in L929 AR mediated transcriptional assay.9 However,
Table 5. Inhibitory activity toward L929 AR cells
Compound
L929 % of inhibition at 3 lM
4j
21
87
35
43
15
18
4q
4v
5e
6b
6d
Table 3. SAR at heteroaryl A-ring
1'
2'
R2
Me
N
3'
4'
het
X
it seemed that the ligands were able to cross the cell
membrane and turn on the AR function. Nevertheless,
these data provided us further confirmation to advance
some lead compounds for in vivo evaluation later.
Compound
X
R2
Rat AR COS-7
whole-cell
het
bindinga
IC50 (lM)
In summary, we have developed and characterized a
novel series of teracyclic indoles as AR ligands. These
compounds were evaluated by AR binding and cell-
based functional assays. Some of them demonstrated
strong binding affinity. Further studies will be focused
on transferring in vitro activity to in vivo efficacy.
6a
6b
CH2
CH2
30-OH
30-OH
0.93
N
0.029
N
N
6c
6d
6e
6f
CH2
20-OH
30-OH
30-OH
30-OH
30-OH
29%b
0.052
33%b
0.58
Supplementary data
Supplementary data associated with this article can be
found, in the online version, at doi:10.1016/j.bmcl.
2006.03.047.
S
N
N
CH2
CH2
CH2
N
References and notes
1. (a) Chang, C.; Kokontis, J.; Liao, S. Science 1988, 240, 324;
(b) Lubahn, D. B.; Joseph, D. R.; Sullivan, P. M.; Willard,
H. F.; French, F. S.; Wilson, E. M. Science 1988, 240, 327.
2. (a) Chengalvala, M.; Oh, T.; Roy, A. K. Expert Opin. Ther.
Patents 2003, 13, 59; (b) Gao, W.; Bohl, C. E.; Dalton, J. T.
Chem. Rev. 2005, 105, 3352, and references cited therein; (c)
Chen, J.; Kim, J.; Dalton, J. T. Mol. Interventions 2005, 5,
173.
S
6g
0.37
N
Me
a IC50 values are representative of multiple determinations (N = 2–3).
b Inhibition at 3 lM.

X. Zhang et al. / Bioorg. Med. Chem. Lett. 16 (2006) 3233–3237
3237
3. (a) Basaria, S.; Wahlstrom, J. T.; Dobs, A. S. J. Clin.
Endocrinol. Metab. 2001, 86, 5108; (b) Shahidi, N. T. Clin.
Ther. 2001, 23, 1355.
4. (a) Andres, N.-V. J. Clin. Endocrinol. Metab. 1999, 84,
3459; (b) Liu, P. Y.; Swerdloff, R. S.; Veldhuis, J. D. J. Clin.
Endocrinol. Metab. 2004, 89, 4789.
5. (a) Lanter, J. C.; Sui, Z.; Fiordeliso, J. J.; Jiang, W.;
Zhang, X. U.S. Patent Appl. Publ. 2005250741, 2005;
Chem. Abstr. 2005, 120, 1032; (b) Ng, R. A.; Sui, Z. PCT
Int. Appl. WO 2004113309, 2004; Chem. Abstr. 2004, 115,
4685.
6. Thiochromanone: (a) Speckamp, W. N.; Westra, J. G.;
Huisman, H. O. Tetrahedron 1970, 26, 2353; 7-methoxy-1-
methyl-2,3-dihydro-1H-quinolin-4-one: (b) Speckamp, W.
N.; Van Velthuysen, J. A.; Pandit, U. K.; Huisman, H. O.
Tetrahedron 1968, 24, 5881; 2,3-dihydro-thiopyrano[2,3-
b]pyridin-4-one: (c) Da Settimo, A.; Marini, A. M.;
Primofiore, G.; Da Settimo, F.; Salerno, S.; La Motta, C.;
Pardi, G.; Ferrarini, P. L.; Mori, C. J. Heterocycl. Chem.
2000, 37, 379; 7,8-dihydro-6H-quinolin-5-one: (d) Albright,
J. D.; Du, X. J. Heterocycl. Chem. 2000, 37, 41; (e) 7,8-
Dihydro-6H-quinoxalin-5-one: Chow, K.; Gil, D. W.;
Burke, J. A.; Harcourt, D. A.; Garst, M. E.; Wheeler, L.
A.; Munk, S. A. PCT Int. Appl. WO 9928300, 1999; Chem.
Abstr. 1999, 37, 5530; (f) 1,4a-dimethyl-4,4a,6,7-tetrahydro-
1H,3H-quinoline-2,5-dione: Von L., Derek; G.; Donald W.;
Aster, S. D. PCT Int. Appl. WO 2000006167, 2000; Chem.
Abstr. 2000, 98, 334.
7. (a) Winterfeldt, E. Liebigs Ann. Chem. 1971, 745, 23; (b)
Warneke, J.; Winterfeldt, E. Chem. Ber. 1972, 105, 2120; (c)
Boch, M.; Korth, T.; Nelke, J. M.; Pike, D.; Radunz, H.;
Winterfeldt, E. Chem. Ber. 1972, 105, 2126.
8. For COS-7 AR binding assay, see Hamann, L. G.; Higuchi,
R. I.; Lin, Z.; Edwards, J. P.; Wang, X.-N.; Marschke, K.
B.; Kong, J. W.; Farmer, L. J.; Jones, T. K. J. Med. Chem.
1998, 41, 623, The percent inhibition was determined by
testing dilutions of the test compound (usually duplicates of
10 lM) in the binding assay. Counts per well were
measured and percent inhibition determined. Androgen
receptor binding IC50s were determined by testing serial
dilutions of the test compound (usually duplicate ten half-
log dilutions starting at 10 lM) in the binding assay.
Counts per well were measured and IC50s determined by
linear regression.
9. L929 AR cell-based assay: Zhang, Z.; Lundeen, S. G.; Zhu,
Y.; Carver, J. M.; Winneker, R. C. Steroids 2000, 65, 637,
Percent inhibition was determined by testing dilutions of
the test compound using a concentration of 3 lM.