4-(Alkylthio)- and 4-(arylthio)-benzonitrile derivatives as androgen receptor antagonists for the topical suppression of sebum production

Article abstract of DOI:10.1016/j.bmcl.2009.01.063

The first examples of thioether-substituted benzonitriles as potential soft-drug androgen receptor antagonists are reported. A number of 4-(alkylthio)- and of 4-(arylthio)-benzonitrile analogs were evaluated in human androgen receptor binding and cellular

Full text of DOI:10.1016/j.bmcl.2009.01.063

Bioorganic & Medicinal Chemistry Letters 19 (2009) 1310–1313
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
4-(Alkylthio)- and 4-(arylthio)-benzonitrile derivatives as androgen
receptor antagonists for the topical suppression of sebum production
b
c
d
c
e
Lorna Mitchell a, , Zhi Wang , Lain-Yen Hu , Catherine Kostlan , Matthew Carroll , Danielle Dettling ,
*
Daniel Du c, David Pocalyko f, Kimberly Wade g
a Pfizer Global Research and Development, Sandwich Laboratories, Ramsgate Road, Sandwich, CT13 9QA, UK
b Abbott Laboratories, Chicago, IL 60064, USA
c Pfizer Global Research and Development, Groton Laboratories, Groton, CT 06340, USA
d IDSC, LLC, Chelsea, MI 48118, USA
e Pfizer Global Research and Development, Research Technology Center, MA 02139, USA
f Pfizer Global Research and Development, La Jolla Laboratories, CA 92121, USA
g Pfizer Global Research and Development, St. Louis Laboratories, MO 63141, USA
a r t i c l e i n f o
a b s t r a c t
Article history:
The first examples of thioether-substituted benzonitriles as potential soft-drug androgen receptor antag-
onists are reported. A number of 4-(alkylthio)- and of 4-(arylthio)-benzonitrile analogs were evaluated in
human androgen receptor binding and cellular functional assays. Analogs with potent in vitro binding
and cellular activities were evaluated for topical in vivo efficacy in the Golden Syrian hamster ear model.
Analogs from both the 4-(alkylthio)- and of 4-(arylthio)-benzonitrile series showed moderate reduction
of wax esters in vivo.
Received 11 December 2008
Revised 20 January 2009
Accepted 22 January 2009
Available online 27 January 2009
Keywords:
Ó 2009 Elsevier Ltd. All rights reserved.
Androgen receptor antagonist
Androgen receptor
Antiandrogen
Sebum
Topical
Softdrug
Thioether
The androgen receptor (AR) is responsible for the activation of
genes involved in the pathogenesis of acne and alopecia.1 The onset
of acne vulgaris infections in the pilosebaceous unit is associated
with excess sebum production. The production of sebum is regu-
the ether linkage with
approach.6
a thioether as a potential soft-drug
We found that the oxygen linker of diphenyl ether 1 could be
replaced with sulfur (2) with retention of AR binding (ARB)7 and
lated by the androgens testosterone and 5a-dihydrotestosterone
(5a-DHT). Androgen receptor antagonists, such as cyproterone
acetate2,3 and RU-588414 have been shown to suppress sebum pro-
duction when applied topically.
Table 1
Human androgen binding and cellular activities for oxygen, sulfur, and sulfoxide-
linked analogsa
Mining of data from our AR agonist program identified a num-
ber of compounds that bound to the androgen receptor but did not
possess agonist activity. A reevaluation showed that some of these
were full antagonists (for example diphenyl ethers such as 1), and
we have used them as starting points for our program to identify
AR antagonists for the topical suppression of sebum production.5
Part of our design strategy to access compounds devoid of sys-
temic antiandrogen effects and suitable for topical delivery cen-
tered on introducing metabolic lability into our target
compounds. In the current work we investigated replacement of
NC
X
Cl
Me
Compound
X
ARB (nM)
ARCELL (nM)
1
2
3
O
S
64
43
>10000
46
78
–
S=O
a
Values (IC50) are given as an average of P2 experiments; –, not tested; ARB,
human androgen receptor binding assay; ARCELL, human androgen receptor cel-
lular functional assay.
* Corresponding author. Tel.: +44 1304 649599.
E-mail address: lorna.mitchell@pfizer.com (L. Mitchell).
0960-894X/$ - see front matter Ó 2009 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2009.01.063

L. Mitchell et al. / Bioorg. Med. Chem. Lett. 19 (2009) 1310–1313
1311
ing materials for 2- or 3-methoxy derivatives were also readily
available and gave targets with an alternative metabolic soft-spot,
although the chemistry proved not to be as general (the displace-
ment reactions were lower yielding and had more synthetic fail-
ures for the electron-rich methoxy templates than the
trifluoromethyl template).
We initially explored a series of tolyl ethers based on hit 1 be-
cause they presented an additional metabolizable group. Perhaps
surprisingly, diphenyl ethers with a 2-methoxy-benzonitrile (4a–
c) were found to have better AR binding activity than either their
3-methoxy- (5a–c) or 2-trifluoromethyl-(6a–c) analogs. For a par-
ticular benzonitrile, potency of R2 was found to follow ortho > me-
ta >> para (Table 2).
NC
NC
4'
3'
R2
a
2
R1
S
X
R1
X = F, I
3
2'
Scheme 1. Reagents and conditions: (a) ArSH, Cs2CO3 (10–90%).
full antagonist activity in a cellular assay (ARCELL)8 (Table 1). The
readily prepared sulfoxide of this (3) was inactive in the AR binding
assay suggesting that thioether linked compounds could have po-
tential as soft-drugs.9
The diphenyl thioethers were synthesized by displacement of
fluorine or iodine from the appropriate 4-halobenzonitrile with
arylthiols under basic conditions (Scheme 1).10
We chose not to synthesize further chlorobenzonitriles since
some of them exhibited photoinstability, which is not desirable
in a topically applied agent. The 2-trifluoro-methyl-benzonitrile
was selected as an alternative, based on known AR binders.11 Start-
A variety of ortho-aryl substitution was then explored (Table 3).
In the 2-trifluorobenzonitrile template the unsubstituted aryl ana-
log 7a had similar potency in the AR binding assay as the ortho-to-
lyl analog 6a. Replacement of the ortho-methyl with fluorine (8b)
or methoxy (7c) was tolerated in terms of AR binding and both
these compounds had functional activity less than 100 nM.
Ortho-disubstituted analogs, such as 7d and 7e, did not offer
improvement in binding activity, in the case of 7d, or were less ac-
tive, in the case of 7e. By comparison in the 2-methoxybenzonitrile
template the ortho-methyl (4a) ortho-methoxy (8a), dimethyl (8b),
and mesityl (8c) analogs all possessed good AR binding and func-
tional activity.
Table 2
SAR of tolyl analogsa
Compound
R1
R2
ARB (nM)
ARCELL (nM)
Although our previous AR agonist program had found that
ethers of benzonitriles with small alkyl groups had no AR activity,
we were interested in investigating the alkyl thioether case. Since
many of the desired alkyl thiols were not commercially available,
we prepared 4-thiobenzonitrile templates by Newman–Kwart
rearrangement12 and alkylated them with commercially available
alkyl halides (Scheme 2).13
Intriguingly, even small alkyl thioethers such as isopropyl and
propyl were active in the binding assay and shown to be full antag-
onists. By contrast, the direct ether analogs of 10a and 10c had ARB
IC50’s >1000 nM.
4a
4b
4c
5a
5b
5c
6a
6b
6c
2-OMe
2-OMe
2-OMe
3-OMe
3-OMe
3-OMe
2-CF3
20-Me
30-Me
40-Me
20-Me
30-Me
40-Me
20-Me
30-Me
40-Me
24
57
81
197
209
48
–
–
–
–
–
200
115
353
5210
535
1680
5330
2-CF3
2-CF3
a
Values (IC50) are given as an average of P2 experiments; –, not tested; ARB,
human androgen receptor binding assay; ARCELL, human androgen receptor cel-
lular functional assay.
In contrast to the diphenyl thioethers, alkyl thioethers from the
2-trifluoromethyl-benzonitrile series (10a–f) had similar AR bind-
ing activity to those from the 2-methoxy-benzonitrile series (9a–
10f) (Table 4).
Selected compounds with good in vitro profiles were tested
in vivo in Golden Syrian hamsters for their ability to reduce wax
and cholesterol esters. Wax and cholesterol esters constitute 28%
of total human sebum14 and it has been shown that there is a direct
correlation between reduction in wax esters and reduction in total
sebum production in a clinical trial with oral cyproterone ace-
tate.15 The hamster ear model is a widely used animal model to
test drug effects on sebaceous glands (Table 5).16
Table 3
SAR of other diphenyl thioethersa
Compound
R1
R2
ARB (nM)
ARCELL (nM)
7a
7b
7c
7d
7e
8a
8b
8c
2-CF3
2-CF3
2-CF3
2-CF3
2-CF3
2-OMe
2-OMe
2-OMe
H
225
184
56
411
1340
20
123
54
37
–
20-F
20-OMe
20,60-diMe
20,40,60-triMe
20-OMe
–
35
30
69
20,60-diMe
20,40,60-triMe
43
39
All compounds tested in the hamster ear model showed some
activity but all were only moderately active compared to the
positive control, RU-58841, which gave a 95% reduction in wax
esters.
a
Values (IC50) are given as an average of P2 experiments; –, not tested; ARB,
human androgen receptor binding assay; ARCELL, human androgen receptor cel-
lular functional assay.
NC
NC
S
S
a
+
R1
R1 = CF3, OMe
OH
R1
O
NMe2
Me2N
Cl
b
NC
NC
R1
NC
R1
c
O
d
2
R2
R1
S
SH
S
NMe2
3
Scheme 2. Reagents and conditions:, (a) DABCO, DMF, 65 °C (89–93%); (b) 200–240 °C, neat (92–100%); (c) NaOH, MeOH, thf (63–78%); (d) R2Br, Cs2CO3, DMF, r.t (10–90%).

1312
L. Mitchell et al. / Bioorg. Med. Chem. Lett. 19 (2009) 1310–1313
plates were counted using a Microbeta Trillux. To determine non-specific
Table 4
SAR of alkyl thioether analogsa
binding, 1000-fold concentration of cold DHT was added to each well.
Triplicate wells were tested for each concentration. The experimental results
were expressed as the concentration of compound at which 50% of maximum
binding was inhibited (IC50). Compound RU-58841 was run as a standard
revealing an inter-run variability within 2-fold.
Compound
R1
R2
ARB (nM)
ARCELL (nM)
9a
9b
9c
9d
9e
9f
10a
10b
10c
10d
10e
10f
10g
10h
2-OMe
2-OMe
2-OMe
2-OMe
2-OMe
2-OMe
2-CF3
2-CF3
2-CF3
2-CF3
2-CF3
–Pr
40
84
29
>1000
66
10
55
–
–Bu
8. The androgen receptor cellular functional assay was conducted in a human
breast tumor cell line expressing androgen receptor (MDA-MB453-MMTV
clone 54–19). The cell line is a stably transfected cell line with MDA-MB453 cell
–iPr
–CH2cPr
–CH2cBu
–CH2CH2iPr
–Pr
169
441
1730
23
background.
Element (ARE) was first cloned in front of a firefly luciferase receptor gene.
Then the cascade was cloned into transfection vector pUV120-puro.
A MMTV minimal promoter containing Androgen Response
–
11
a
–Bu
49
20
35
63
838
89
133
702
–
Electroporation was used for transfecting MDAMB-453 cells and a puromycin
resistant stable cell line was selected. See Chang, C.; Wang, C.; DeLuca, H. F.;
Ross, T. F.; Shih, C.C.-Y. Proc. Natl. Acad. Sci. USA. 1992, 89, 5946.
–iPr
–CH2cPr
–CH2cBu
–CH2CH2iPr
–tBu
9. Collection of metabolic stability data failed using our high through-put
procedure as these thioether compounds did not give mass ion under the
standard conditions. Alternative studies were not conducted due to the
moderate level of in vivo efficacy in this series.
2-CF3
2-CF3
2-CF3
519
315
194
–
167
–cBu
10. Generalized procedures for the synthesis of diphenylthioethers from
a
commercially available thiols; Method A:
a
mixture of 4-fluoro-2-
Values (IC50) are given as an average of P2 experiments; –, not tested; ARB,
human androgen receptor binding assay; ARCELL, human androgen receptor cel-
lular functional assay.
trifluoromethyl-benzonitrile (or 4-fluoro-2-methoxy-benzonitrile), the thiol
(1.05 equiv), and cesium carbonate (2 equiv) in dimethylformamide was
stirred at room temperature or at 50–70 °C for 1 h–2 days or until reaction was
complete. The reaction mixture was poured into ice and if the product
precipitated out it was collected by filtration. Otherwise the reaction mixture
was extracted twice with ethyl acetate, the combined extracts were dried with
brine and then magnesium sulfate and concentrated on the rotovap. If
necessary the crude product was purified by column chromatography on
silica. Array method A: a mixture of 4-fluoro-2-trifluoromethyl-benzonitrile, the
thiol (1 equiv), and cesium carbonate (1.5 equiv) in acetonitrile was heated at
60 °C overnight. The reaction mixture was then diluted with acetonitrile and
an excess of silica supported maleimide was added. The reaction mixture was
heated at 40 °C for 2 h and then filtered. The solvent was evaporated. The
product purified by HPLC using an XTerra RP18, 30 ꢀ 100 mm column eluted
with an acetonitrile, water, formic acid solution. Array method B: a mixture of
4-iodo-2-methoxy-benzonitrile, the thiol (1 equiv), and cesium carbonate (1.5
equiv) in dimethylformamide was heated at 65 °C for 2 days. The reaction
mixture was then diluted with acetonitrile and an excess of silica supported
maleimide was added. The reaction mixture was heated at 40 °C for 2 h and
then filtered. The solvent was evaporated. The product was purified by HPLC
Table 5
In vivo efficacy in Golden Syrian hamster model of thioether linked analogsa
NC
R2
S
R1
Compound
R1
R2
% WE reduction
% CE reduction
8b
7b
9c
10c
10d
2-OMe
2-CF3
2-OMe
2-CF3
2,6-diMe-C6H3
2-F-C6H4
–iPr
66
64
59
46
43
45
50
46
40
35
–iPr
2-CF3
–CH2iPr
using
a
Sunfire 19 ꢀ 100 mm Prep C18 OBD
5 micron column with an
a
All compounds were tested at dose of 1% in a polyethylene glycol/transcutol/
ethanol 20/20/60 v/v/v% formulation.
acetonitrile, water, formic acid solution.
11. (a) Koch, H. Drugs Today 1984, 20, 561; (b) Fradet, Y. Exp. Rev. Anticancer Ther.
2004, 4, 37.
12. Lloyd–Jones, G. C.; Moseley, J. D.; Renny, J. S. Synthesis 2008, 661.
13. Synthesis of 4-mercapto-2-trifluoromethyl-benzonitrile. Step 1. Preparation of
dimethyl-thiocarbamic acid O-(4-cyano-3-trifluoromethyl-phenyl) ester: To a
250 mL round bottom flask were added 4-hydroxy-2-trifluoromethyl-
benzonitrile (5.0 g, 26.7 mmol), DABCO (7.49 g, 66.8 mmol), and 40 mL DMF.
The mixture was stirred until all solids dissolved. Dimethylthiocarbamoyl
chloride (4.13 g, 33.4 mmol) was then added, and the mixture was heated at
65 °C for 4 h. the mixture was then cooled to room temperature and quenched
into ice water. 1 N HCl was used to adjust the pH to 3. Ethyl acetate was added
to the mixture. The layers were separated. The aqueous layer was extracted
with ethyl acetate. The combined organic layers were washed with water and
brine, dried (MgSO4), filtered and concentrated in vacuo. The crude product
was purified by Biotage Horizon System (10–50% EA/Hexane) to afford a light
yellow solid (6.51 g, 88.8%), MS(AP+) = 275.0, 1H NMR (400 MHz,
CHLOROFORM-d) dppm 1.55 (1 H, br s), 3.38 (2 H, s), 3.47 (3 H, s), 7.42 (1 H,
dd), 7.53 (1 H, d), 7.87 (1 H, d). Step 2: preparation of dimethyl-thiocarbamic
acid S-(4-cyano-3-trifluoromethyl-phenyl) ester: dimethyl-thiocarbamic acid
O-(4-cyano-3-trifluoromethyl-phenyl) ester (6.51 g, 23.7 mmol) was heated at
200 °C under nitrogen for 3 hr. On cooling, the product solidified (6.45 g,
92.1%), 1H NMR (400 MHz, CHLOROFORM-d) dppm 3.10 (6 H, d, J = 19.3 Hz),
7.83 (2 H, s), 7.94 (1 H, s). Step 3: Preparation of 4-Mercapto-2-trifluoromethyl-
In summary, we have prepared series of 4-(alkylthio)- and 4-
(arylthio)-benzonitriles as androgen receptor antagonists with po-
tential as soft-drugs and demonstrated that compounds from these
series exhibit moderate reduction in sebum when applied topically
in a validated animal model.
References and notes
1. Redmond, G. P.; Bergfeld, W. F. Cleveland Clinic J. Med. 1990, 57, 428.
2. Bingham, K. D.; Low, M.; Wyatt, E. H. Lancet 1979, 314, 304.
3. Gruber, D. M.; Sator, M. O.; Joura, E. A.; Kokoschka, E. M.; Heinza, G.; Huber, J. C.
Arch. Dermatol. 1998, 134, 459.
4. Gao, W.; Bohl, C. E.; Dalton, J. T. Chem. Rev. 2005, 105, 3352.
5. (a) Hu, L. Y.; Du, D.; Hoffman, J.; Smith, Y.; Fedij, V.; Kostlan, C.; Johnson, T. R.;
Huang, Y.; Kesten, S.; Harter, W.; Yue, W. S.; Li, J. J.; Barvian, N.; Mitchell, L. H.;
Lei, H. J.; Lefker, B.; Carroll, M.; Dettling, D.; Krieger-Burke, T.; Samas, B.;
Yalamanchili, R.; Lapham, K.; Pocalyko, D.; Sliskovic, D.; Ciotti, S.; Stoller, B.;
Hena, M. A.; Ding, Q.; Maiti, Samarendra N.; Stier, M.; Welgus, H. Bioorg. Med.
Chem. Lett. 2007, 17, 5983; (b) Van Camp, J. A.; Hu, L.-Y.; Kostlan, C.; Lefker, B.;
Li, J. J.; Mitchell, L. H.; Wang, Z.; Yue, W.-S.; Carroll, M.; Dettling, D.; Du, D.;
Pocalyko, D.; Wade, K. Bioorg. Med. Chem. Lett. 2007, 17, 5529; (c) Hu, L.-Y.; Lei,
H. J.; Du, D.; Johnson, T. R.; Fedij, V.; Kostlan, C.; Yue, W.-S.; Lovdahl, M.; Li, J. J.;
Carroll, M.; Dettling, D.; Asbill, J.; Fan, C.; Wade, K.; Pocalyko, D.; Lapham, K.;
Yalamanchili, R.; Samas, B.; Vrieze, D.; Ciotti, S.; Krieger-Burke, T.; Sliskovic, D.;
Welgus, H. Bioorg. Med. Chem. Lett. 2007, 17, 5693.
benzonitrile: To
a 250 mL round bottom flask were added dimethyl-
thiocarbamic acid S-(4-cyano-3-trifluoromethyl-phenyl) ester (6.32 g,
23.0 mmol), 30 mL methanol, and 46 mL 1 M NaOH solution. The mixture
was stirred at room temperature. Tetrahydrofuran was added to make all
solids into the solution. After 5 h, 1 N HCl solution was added to adjust pH to 4.
The mixture was concentrated. The aqueous layer was extracted with ethyl
acetate. The combined organic layers were washed with water and brine, dried
(MgSO4), filtered and concentrated in vacuo. The crude product was purified
by Biotage Horizon System (0–40% EA/hexane) to afford an off white solid. 1H
(400 MHz, CHLOROFORM-d) dppm 3.86 (1 H, s), 7.51 (1 H, dd, J = 8.2, 1.9 Hz),
7.68 (1 H, d, J = 8.2 Hz), 7.63 (1 H, d, J = 1.6 Hz). MS(AP-) = 202.0; IR 2539 cmꢁ1
(S-H). 4-mercapto-2-trifluoromethyl-benzonitrile was made by the same
method using 4-hydroxy-2-methoxy-benzonitrile as the starting material. 1H
NMR (400 MHz, CHLOROFORM-d) dppm 3.69 (1 H, s), 3.92 (3 H, s), 6.86 (1 H, d),
6.82 (1 H, d), 7.39 (1 H, d).
6. Bodor, N.; Buchwald, P. Mol. Biotechnol. 2004, 26, 123.
7. The androgen receptor binding assay run was a modification of that described
in Liao, S.; White, D.; Schilling, K.; Chang, C. J. Steroid Biochem. 1984, 1, 11. The
human AR cDNA cloned in baculovirus was expressed in Sf9 cells. Cell lysates
from transfected Sf9 cells were isolated and used as the source of human AR in
the radio–ligand binding assay. Different concentrations of test compounds
(10,000, 1000, 200, 40, 8, 1.6, and 0.16 nM) were incubated in the presence of
human AR extract, hydroxylapatite, and 1 nM 3H-DHT for one hour at 4 °C with
gentle rocking. After incubation, plates were placed on a filter apparatus and
the reaction mixture was removed under 15 psi vacuum pressure, then washed
three times with cold buffer. The plates were then dried at room temperature
overnight. The next day, scintillation fluid was added to each well and the
14. Thiboutot, D. J. Invest. Dermatol. 2004, 123, 1.
15. Miller, C. G.; Wojnarowska, F. T.; Dowd, P. M.; Ashton, R. E.; O’Brien, T. J.;
Griffiths, W. A.; Jacobs, H. S. Br. J. Dermatol. 1986, 114, 705.
16. Male Syrian hamsters aged 9–10 weeks were introduced into the laboratory
environment and acclimated for 2 weeks prior to use in each study. Each

L. Mitchell et al. / Bioorg. Med. Chem. Lett. 19 (2009) 1310–1313
experimental group consisted of five animals. Vehicle control and
1313
a
analysis as follows. One 8 mm distal biopsy punch was taken just above the
anatomical ‘‘V” mark in the aural cartilage to normalize sample area.
The punch was then split into ventral and dorsal layers. The ventral layer,
where the topical dose was applied, was retained for sebum analysis. Each
ventral side sample was placed in a 1-dram glass vial, flushed with nitrogen
gas (N2), sealed, and stored at ꢁ80 °C until for sample lipid extraction and
HPLC lipid analysis. See Plewig, G.; Luderschmidt, C. J. invest. Dermatol. 1977,
68, 171.
reference AR antagonist (RU-58841) were included in each experiment.
Animals were topically dosed twice daily (BID) for 2 weeks, 5 days a week
(Monday to Friday). Each dose consisted of 25 lL of vehicle control or
formulated test article, which was evenly applied to ꢂ3 cm2 of the ventral
surfaces of both the right and left ears. Animals were sacrificed
approximately 18–24 h after the final dose. The ears were collected from
each animal for sebum analysis. The ear samples were prepared for sebum