Synthesis and biological evaluation of stilbene-based pure estrogen antagonists

Article abstract of DOI:10.1016/j.bmcl.2004.06.098

The nonsteroidal estrogen diethylstilbestrol can be converted into potent antiestrogens devoid of agonist activity by introduction of side chains with appropriate functional groups. Replacement of one of the ethyl substituents in diethylstilbestrol by side chains with functional groups converted this potent estrogen into pure antiestrogens with the potential for the treatment of breast cancer. These agents completely suppressed estrogen receptor-mediated gene activation and inhibited the growth of estrogen-sensitive MCF-7 breast cancer cells in submicromolar concentrations. The most potent derivative displayed similar activity as fulvestrant (ICI 182,780) in vitro and in the mouse uterine weight test. Obviously, the stilbene structure can act as a substitute for estradiol in the development of pure estrogen antagonists.

Full text of DOI:10.1016/j.bmcl.2004.06.098

Bioorganic & Medicinal Chemistry Letters 14 (2004) 4659–4663
Synthesis and biological evaluation of stilbene-based pure
estrogen antagonists
Georg Walter, Renate Liebl and Erwin von Angerer*
Institut fu¨r Pharmazie, Universita¨t Regensburg, D-93040 Regensburg, Germany
Received 1 March 2004; revised 25 June 2004; accepted 30 June 2004
Available online 31 July 2004
Abstract—Replacement of one of the ethyl substituents in diethylstilbestrol by side chains with functional groups converted this
potent estrogen into pure antiestrogens with the potential for the treatment of breast cancer. These agents completely suppressed
estrogen receptor-mediated gene activation and inhibited the growth of estrogen-sensitive MCF-7 breast cancer cells in submicro-
molar concentrations. The most potent derivative displayed similar activity as fulvestrant (ICI 182,780) in vitro and in the mouse
uterine weight test. Obviously, the stilbene structure can act as a substitute for estradiol in the development of pure estrogen
antagonists.
Ó 2004 Elsevier Ltd. All rights reserved.
Breast cancer is the most important malignancy among
women. In postmenopausal patients, the majority of
OH
HO
these tumors are estrogen-dependent and respond to
endocrine therapies such as the administration of estro-
gen antagonists, as demonstrated by the successful
application of tamoxifen. Since many patients relapse
under the treatment with tamoxifen as the result of the
partial agonist character of this drug, the use of pure an-
tiestrogens might be an effective alternative.1 In previous
studies2,3 we have shown that the introduction of alkyl
side chains with functional elements such as amino or
sulfone groups can completely abrogate agonist activity
of nonsteroidal ligands of the estrogen receptor. These
agents block ER-mediated gene activation and inhibit
the growth of estrogen-sensitive human MCF-7 breast
cancer cells in vitro. In vivo they also display antiestro-
genic activity but the potencies are less pronounced than
those found for partial antagonists of the SERM type.
In this respect, the nonsteroidal pure antiestrogens
resemble fulvestrant4 (Fig. 1), a steroidal antagonist
without residual estrogenic activity. For therapeutic
purposes, it cannot be given orally but has to be admin-
istered to patients by intramuscular injections.4 The rea-
son for the low bioavailability has not been well
understood and might be due to the structure of the side
chain. Our previous studies with 2-phenylindole-based
pure antiestrogens revealed good in vitro effects but
OH
N
(CH2)9-SO-(CH2)3C2F5
Fulvestrant (ICI 182,780)
HO
HO
(CH2)10SO2C5H11
ZK 164,015
R2
OH
OH
R1
R3
4c-h; 9a,b; 10a,b
Diethylstilbestrol (DES, 4b)
Figure 1. Structures of the stilbene-based antiestrogens and reference
drugs.
rather low activity in vivo, especially when these com-
pounds were dosed orally. These results can be taken
as a hint that the structure of the carrier molecule also
plays a role in bioactivity. Therefore, we searched for
different basic structures with improved activity in vivo.
Diethylstilbestrol (DES, Fig. 1) is a drug, which is as
potent as 17b-estradiol (E2) both in vitro and in vivo.
Its use in women, however, has become obsolete after
it had been associated with transplacental carcinogenesis
leading to clear-cell adenocarcinoma of the vagina and
*
Corresponding author. Tel.: +49-941-9434821; fax: +49-941-
9434820; e-mail: erwin.von-angerer@chemie.uni-regensburg.de
0960-894X/$ - see front matter Ó 2004 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2004.06.098

4660
G. Walter et al. / Bioorg. Med. Chem. Lett. 14 (2004) 4659–4663
cervix in daughters of women receiving DES during
pregnancy.5 The good pharmacokinetic profile and high
binding affinity for the estrogen receptor (ER) make
DES attractive as the starting point for the development
of antiestrogens for clinical use. The main question we
wanted to answer in this study was whether it is possible
to convert this potent estrogen into an estrogen antago-
nist devoid of any agonist activity. To achieve this goal
we replaced one of the ethyl groups by long alkyl chains
that incorporate functional groups. Previous studies
have shown that sulfur functions in a distance of 10 car-
bon atoms from the core of the molecule are appropriate
for this purpose.2
All of the stilbenes were obtained as E/Z mixtures with
the E-stereoisomer as the dominant product. HPLC
studies revealed a ratio of E-4 to Z-4 of approximately
85:15. When the stereoisomers had been separated by
HPLC they rapidly isomerized to give the original ratio
of isomers. Thus, no attempt was made to study the ste-
reoisomers separately.
The first step in the biological characterization of the
new stilbene derivatives was the determination of the
binding affinities for the ER. Calf uterine cytosol was
used as receptor source as described previously.6 The
introduction of long side chains reduced the affinity con-
siderably (Table 1). A similar observation was made
with steroids, for example, fulvestrant.7 The data
showed that the affinity is mainly dependent on the type
of side chain used. The most favorable conditions are
provided by the bifunctional side chains as demon-
strated by derivatives 9 and 10. Their values are close
to those found for fulvestrant and the indole derivative
ZK 164,015.
The synthesis of the new DES derivatives started from
the corresponding desoxyanisoins 1 (Scheme 1). The
first side chain was introduced by deprotonation and
subsequent reaction with ethyl bromide or the bromo
alkane with the respective sulfur function to give the ke-
tones 2. The second substituent was introduced by a
Grignard reaction, which led to the formation of a dou-
ble bond with orientation towards the side chain (3).
Since the acidity of the sulfone prevented its direct use
as the Grignard reagent, the thio ether function had to
be oxidized with m-CPBA after the Grignard reaction
to give 3h. Cleavage of the methoxy groups in 3 led to
mixtures of the stereoisomeric phenols 4 and 5 with a
preference for the stilbene structure 4. For the prepara-
tion of DES derivatives 9 and 10 with side chains con-
taining an additional methylamino group the synthesis
had to be modified (Scheme 2): First the desoxyanisoin
1b was reacted with ethyl 6-bromohexanoate to give 6,
followed by the Grignard reaction with EtMgBr to af-
ford the ester 7, which was then reacted with the appro-
priate amine to give 8. Deprotection with BBr3 resulted
in double bond migration from a styrene-like to stilbene-
like position. In the last step, the phenolic stilbenes 9
were reduced with LiAlH4 to yield the amines 10.
Estrogenic and antiestrogenic activities were determined
in vitro6 using ER+ human MCF-7/2a cells8 stably
transfected with a luciferase reporter gene under the
control of an ERE. Since pure antiestrogens should be
devoid of any agonist action by definition, all new stil-
bene derivatives were evaluated for agonism. At a con-
centration of 10À6 M they exhibited values for
luciferase activity below that of control cells grown in
steroid depleted medium (Fig. 2). Luciferase activities
below baseline are characteristic for pure antiestrogens
and indicate the blockade of ligand-independent activa-
tion of the ER, responsible for the basal luciferase activ-
ity in control cells.8
Antiestrogenic activity was determined in a similar
assay by simultaneous treatment of the cells with
R2
R2
OMe
OMe
OMe
b
a
R1
R1
R1
O
O
Me
1a: R1 = 3-OMe
1b: R1 = 4-OMe
2 R2 = Me, (CH2)9SC5H11
,
3
(CH2)9SO2C5H11
OMe
Me
OMe
Me
OMe
Me
d
c
O
(CH2)9SO2C5H11
MeO
(CH2)9SC5H11
MeO
R2
MeO
3h
3c
2b
R2
OH
OH
e
R1
+
3
R1
R3
5a-h R1 = OH
Scheme 1. Reagents and conditions: (a) 1. NaH, DMF, 0°C; 2. R2CH2Br, rt, 2h, 62–86%; (b) EtMgBr, Et2O, reflux, 2h, 54–76%; (c)
11C5S(CH2)10MgBr, Et2O, reflux, 2h, 33%; (d) m-CPBA, CHCl3, rt, 2h, 76%; (e) BBr3, CH2Cl2, À5°C–rt, 7h, 52–95%.
R3
4a-h R1 = OH
H

G. Walter et al. / Bioorg. Med. Chem. Lett. 14 (2004) 4659–4663
4661
OMe
HO2C(CH2)4
OMe
OMe
EtO2C(CH2)4
b, c
a
O
O
Me
MeO
MeO
MeO
7
6
1b
O
OMe
R4
S
N
d (8a) / e (8b)
f
7
MeO
8a R4 = C2H5
8b R4 = C2F5
O
R4
OH
N
OH
S
R4
(CH2)6
S
N
g
HO
HO
10a R4 = C2H5
10b R4 = C2F5
9a R4 = C2H5
9b R4 = C2F5
Scheme 2. Reagents and conditions: (a) 1. NaH, DMF, 0°C; 2. Br(CH2)5CO2Et, rt, 2h, 79%; (b) EtMgBr, Et2O/THF, reflux, 2h, 72%; (c) KOH,
H2O, EtOH, reflux, 4h, 95%; (d) 1. PCl5, 60°C, 1h; 2. MeNH(CH2)3SC5H11, NaOH, 79%; (e) MeNH(CH2)3S(CH2)3C2F5, DCC, CH2Cl2, rt, 4h,
73%; (f) BBr3, CH2Cl2, À5°C–rt, 7h, 80–95%; (g) LiAlH4, THF, reflux, 1h, 21–24%.
Table 1. Relative binding affinities of stilbene-based antiestrogens and reference drugs for the estrogen receptor
Compd
R1
R2
R3
RBAa
R2
OH
R1
R3
17b-estradiol (E2)
100
37
4ab
3-OH
4-OH
3-OH
4-OH
3-OH
3-OH
4-OH
3-OH
4-OH
4-OH
4-OH
4-OH
Me
Me
Me
Me
4b (DES)
4c
72
–(CH2)9–S–C5H11
–(CH2)9–S–C5H11
Me
Me
Me
0.03
0.11
0.07
0.87
2.4
1.7
2.9
1.9
8.6
2.2
5.2
5.0
4db
4eb
–(CH2)9–S–C5H11
Me
4fb
–(CH2)9–SO2–C5H11
–(CH2)9–SO2–C5H11
Me
4gb
Me
–(CH2)9–SO2–C5H11
4hb
9ab
(CH2)4CONMe–(CH2)3SC5H11
Me
Me
Me
Me
9bb
(CH2)4CONMe–(CH2)3S(CH2)3C2F5
(CH2)5NMe–(CH2)3SC5H11
(CH2)5NMe–(CH2)3S(CH2)3C2F5
10ab
10bb
Fulvestrant
ZK 164,015
a Relative binding affinities for the calf uterine estrogen receptor, determined by incubation at 4°C for 20h.
b Mixture of E- and Z-isomers in dynamic equilibrium.
1nM E2 and the respective antiestrogen in various
concentrations. All DES derivatives inhibited E2-
stimulated luciferase expression in submicromolar
concentrations (Table 2). The lowest IC50 values
were recorded for the amines 10a and 10b (10 and
11nM) and were close to the value for fulvestrant
(5nM).
Since the envisaged application of these agents is the
treatment of hormone-dependent breast cancer, their
antiproliferative activities were evaluated using wild-
type MCF-7 breast cancer cells and for comparison
ER- human MDA-MB 231 mammary carcinoma cells.
The growth of hormone-independent cells was not af-
fected by the DES derivatives up to 1lM. At 10lM,

4662
G. Walter et al. / Bioorg. Med. Chem. Lett. 14 (2004) 4659–4663
Table 3. Antiproliferative activity of stilbene-based antiestrogens
Compd
MDA-MB 231 cellsa
MCF-7 cellsb
100
% T/C at 5lM % T/C at 10lM IC50 (nM)
4f
4g
96
94
90
3
7
8
À7 1c
À4 1c
32 8c
49
5.1
4h
10a
6.1
75 6c
À25 1c
0.75
0.22
0.21
1.4
10b
Fulvestrant
ZK 164.015
À15 1c
À15 1c
90
20
93
6
95 6
À2 2c
84 5c
a Growth inhibition of ER- MDA-MB 231 breast cancer cells, incu-
bated for 96h.
b Growth inhibition of ER+ MCF-7 breast cancer cells, incubated for
200h.
10
c Significant inhibition; p < 0.01.
0
ctrl E2 4c 4d 4e 4f 4g 4h 9a 9b 10a10b ICI ZK
MCF-7/2a cells and make an antiestrogenic mode of ac-
tion likely. The strongest inhibitory effects were achieved
in both assays with derivatives that carry an amino
group in the side chain (10a,b). The potency of these
compounds (10 and 7nM) was close to that of the steroi-
dal reference drug fulvestrant (4nM).
Figure 2. Estrogenic activity in MCF-7/2a breast cancer cells. Con-
centrations were 1lM except for E2 (1nM). ICI = fulvestrant,
ZK = ZK 164,015. Note the breaks in the y-axis.
Table 2. Antiestrogenic and antiproliferative activities in E2-stimu-
lated breast cancer cells
The aim of this study was the development of pure an-
tiestrogens with sufficient bioavailability for therapeutic
application. In a preliminary test, we determined the
antiuterotrophic effect of the most active derivative
10b in immature mice in comparison with fulvestrant.
Both compounds completely suppressed the uterotrophic
effect of 10lg E2/kg bodyweight (Fig. 3). When these
compounds were given alone in a dose of 6 and 12mg/
kg, respectively, no agonist activity was noticed.
Compd
Antiestrogenic
activitya IC50 (lM)
Antiproliferative
activityb IC50 (lM)
4c
4d
1.2
2.1
1.4
0.79
0.20
0.37
0.13
0.050
0.60
0.31
0.010
0.011
0.005
0.025
4e
0.59
0.27
0.10
0.11
1.9
4f
4g
4h
9a
Since all of the new stilbenes with free hydroxy groups
bind to the ER, a binding mode similar to that of
DES9 has to be assumed. It is anticipated that the bind-
ing pocket of the bovine ER is rather similar to the one
in the human ERa. Thus, hydrogen bridges are formed
between the phenolic groups and histidine, equivalent to
9b
10a
0.7
0.010
0.007
0.004
0.051
10b
Fulvestrant
ZK 164,015
a Inhibition of luciferase activity in ER+ MCF-7/2a breast cancer cells,
transfected with the EREwtc luc plasmid and stimulated by E2
(10À9 M).
b Growth inhibition of wild-type MCF-7 breast cancer cells stimulated
by E2 (10À9 M).
60
E2
(10 g/kg)
however, both the sulfones and the amines inhibited the
growth of these cells (Table 3), a property, which they
share with the parent drug DES and which appears to
be the result of receptor-independent cytotoxicity.
40
E2 + 10b
E2 + ICI
When these compounds were tested in estrogen-sensitive
MCF-7 cells under the same conditions a strong inhibi-
tory effect was observed with IC50 values in the nanomo-
lar range (Table 3). The differences in activity––more
than three orders of magnitude––clearly show that all
of these compounds act via the ER. In order to compare
the antiproliferative activities of the new DES deriva-
tives with the antiestrogenic effects measured in trans-
fected MCF-7/2a cells, wild-type MCF-7 cells were
treated with the test compounds in the presence of
1nM E2 (Table 2). The addition of E2 led to an average
increase of the IC50 values by a factor of 20. The results
of this assay are in agreement with the data obtained in
20
control
0.1
0
0.01
1
10
Concentration [mg/kg]
Figure 3. Antiuterotrophic effects of 10b and fulvestrant given s.c.
simultaneously with 10lg E2/kg bodyweight to immature mice. Open
triangles show the effects of 10b (12mg; ,) and fulvestrant (6mg; n)
alone on uterus weight.

G. Walter et al. / Bioorg. Med. Chem. Lett. 14 (2004) 4659–4663
4663
His524, and glutamate/arginine, equivalent to Glu353/
References and notes
Arg394 in the human ER, respectively. The long side
chain reduces the binding affinity considerably, but it
prevents helix 12 from occupying the agonist conforma-
tion.10 In this respect, these compounds resemble other
antiestrogens such as raloxifen, which interfere with
the receptor–coactivator interaction by changing the
position of helix 12.11
1. Jones, S. E. Semin. Oncol. 2003, 30, 14.
2. Biberger, C.; von Angerer, E. J. Steroid Biochem. Mol.
Biol. 1996, 58, 31.
3. (a) Leichtl, S.; von Angerer, E. Arch. Pharm. Pharm. Med.
Chem. 1998, 331, 283; (b) von Angerer, E.; Biberger, C.;
Holler, E.; Koop, R.; Leichtl, S. J. Steroid Biochem. Mol.
Biol. 1994, 49, 51.
4. Morris, C.; Wakeling, A. Endocr. Relat. Cancer 2002, 9,
267.
5. Melnick, S.; Cole, P.; Anderson, D.; Herbst, A. N. Engl. J.
Med. 1987, 316, 514.
6. Golob, T.; Liebl, R.; von Angerer, E. Bioorg. Med. Chem.
2002, 10, 3941.
7. Biberger, C.; von Angerer, E. J. Steroid Biochem. Mol.
Biol. 1998, 64, 277.
8. Hafner, F.; Holler, E.; von Angerer, E. J. Steroid Biochem.
Mol. Biol. 1996, 58, 385.
9. Shiau, A. K.; Barstad, D.; Loria, P. M.; Cheng, L.;
Kushner, P. J.; Agard, D. A.; Greene, G. L. Cell 1998, 95,
927.
10. Brzozowski, A. M.; Pike, A. C. W.; Dauter, Z.; Hubbard,
¨
R. E.; Bonn, T.; Engstro¨m, O.; Ohman, L.; Greene, G. L.;
Gustafsson, J.-A.; Carlquist, M. Nature 1997, 389, 753.
11. Edwards, D. P. J. Mammary Gland Biol. Neoplasia 2000,
5, 307.
The results of this study clearly show that it is possible
to convert the potent estrogen DES into pure antiestro-
gens by the introduction of appropriate side chains into
the molecule. The most favorable conditions for antag-
onism are a tertiary amino group in the side chain, sep-
arated by a hexamethylene spacer group from the
stilbene core, an additional thio ether function in the
remaining chain, and fluorination of the terminal posi-
tions as demonstrated by compound 10b. This kind of
side chain appears to be equivalent to that in fulvestrant.
Since both agents display similar activities both in vitro
and in vivo equivalence of the carrier molecules has to
be assumed. The complete antagonism can be rational-
ized by an ER destabilization as observed with other
pure antiestrogens.12 When administered subcutane-
ously to mice a 100-fold excess of stilbene 10b or fulves-
trant was necessary to block the stimulatory effect of
10lg E2 completely. The rather low bioactivity can be
explained by the reduced binding affinities of both
agents for the ER when compared with E2.
12. Hoffmann, J.; Bohlmann, R.; Heinrich, N.; Hofmeister,
H.; Kroll, H.; Kunzer, H.; Lichtner, R. B.; Nishino, Y.;
Parczyk, K.; Sauer, G.; Gieschen, H.; Ulbrich, H. F.;
Schneider, M. R. J. Natl. Cancer Inst. 2004, 96, 210.