Structure-activity relationships of SERMs optimized for uterine antagonism and ovarian safety

Article abstract of DOI:10.1016/j.bmcl.2007.04.044

Structure-activity relationship studies are described, which led to the discovery of novel selective estrogen receptor modulators (SERMs) for the potential treatment of uterine fibroids. The SAR studies focused on limiting brain exposure and were guided by computational properties. Compounds with limited impact on the HPO axis were selected using serum estrogen levels as a biomarker for ovarian stimulation.

Full text of DOI:10.1016/j.bmcl.2007.04.044

Bioorganic & Medicinal Chemistry Letters 17 (2007) 3544–3549
Structure–activity relationships of SERMs optimized for
uterine antagonism and ovarian safety
Timothy I. Richardson,a,* Scott A. Frank,a Minmin Wang,a Christian A. Clarke,a
Scott A. Jones,a Bai-Ping Ying,a Dan T. Kohlman,a Owen B. Wallace,a
Timothy A. Shepherd,a Robert D. Dally,a Alan D. Palkowitz,a Andrew G. Geiser,a
Henry U. Bryant,a Judith W. Henck,a Ilene R. Cohen,a Daniel G. Rudmann,a
Denis J. McCann,a David E. Coutant,a Samuel W. Oldham,a Conrad W. Hummel,b
Kin C. Fong,b Ronald Hinklin,b George Lewis,b Hongqi Tianb and Jeffrey A. Dodgea
aLilly Research Laboratories, Eli Lilly and Company, Lilly Corporate Center, Indianapolis, IN 46285, USA
bChemistry Department, Array BioPharma Inc., 1885 33rd Street, Boulder, CO 80301, USA
Received 3 April 2007; revised 16 April 2007; accepted 18 April 2007
Available online 24 April 2007
Abstract—Structure–activity relationship studies are described, which led to the discovery of novel selective estrogen receptor modu-
lators (SERMs) for the potential treatment of uterine fibroids. The SAR studies focused on limiting brain exposure and were guided
by computational properties. Compounds with limited impact on the HPO axis were selected using serum estrogen levels as a bio-
marker for ovarian stimulation.
Ó 2007 Elsevier Ltd. All rights reserved.
Uterine fibroids (leiomyomas) are non-cancerous
growths that develop from the smooth muscle cells
and fibrous connective tissue of the uterine wall.1
Although most are asymptomatic, in some women, fib-
roids cause abnormal menstrual bleeding and pain,
which have a profoundly negative effect on quality of
life. For these women hysterectomy is the most definitive
solution accounting for over 200,000 surgeries a year in
the US.2 Pathological evaluation of hysterectomy speci-
mens reveals a similar incidence in both postmenopausal
and premenopausal women (84% and 74%, respec-
tively).3 However, in premenopausal women these
growths are more numerous and of larger size. Hysterec-
tomy is unacceptable for a woman who desires a future
pregnancy. Other invasive surgical procedures, which
preserve the uterus, such as myomectomy (fibroid re-
moval with uterine retention), laser ablation, or emoliza-
tion are effective; however, these treatments are
associated with a high rate of fibroid recurrence.
The current pharmaceutical treatment for uterine fib-
roids involves the use of gonadotropin-releasing hor-
mone (GnRH) agonists. These peptides stimulate the
pituitary gland resulting in down-regulation of the
hypothalamic–pituitary–ovarian (HPO) axis, decreasing
the release of gonadotropins (FSH and LH), and a
subsequent reduction in the amount of estrogen (E2)
produced in the ovaries. This hypoestrogenic state
causes a reduction in uterine volume and fibroid size.4
Unfortunately, there are unacceptable side effects, most
notably bone loss, associated with this therapy and
once discontinued, the fibroids return. As a result,
GnRH agonists are primarily used to reduce fibroid
size prior to surgical removal.
Since uterine fibroids exhibit E2 dependence, the estro-
gen receptors (ERa and ERb) are viable targets for
pharmaceutical treatment. The ERs are ligand-depen-
dent transcription factors that belong to the nuclear hor-
mone receptor (NHR) superfamily. Selective estrogen
receptor modulators (SERMs) exhibit tissue type func-
tional selectivity, acting as agonists in some tissues but
antagonists in others.5,6 Tamoxifen (1) has been clini-
cally evaluated for the treatment of fibroids but it lacks
sufficient efficacy due to its agonist activity in uterine
Keywords: Estrogen; Estrogen receptor; ER; Selective estrogen recep-
tor modulator; SERM; Uterine fibroids; Leiomyomas; Blood–brain
barrier; Hypothalamic–pituitary–ovarian axis; HPO axis.
*
Corresponding author. Tel.: +1 317 433 2373; fax: +1 317 433
0552; e-mail: t_richardson@lilly.com
0960-894X/$ - see front matter Ó 2007 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2007.04.044

T. I. Richardson et al. / Bioorg. Med. Chem. Lett. 17 (2007) 3544–3549
3545
tissues.7,8 It also causes ovarian cysts, an unacceptable
side effect that limits its use in premenopausal women.
This ovarian stimulatory effect has been attributed to
inhibition of the HPO axis due to the antagonist activity
exhibited by tamoxifen in the hypothalamus. Inhibition
of the HPO axis results in overproduction of gonadotro-
pins (LH and FSA) and hyper-stimulation of the ova-
ries. This activity is also exhibited by clomiphene (2),
which has been used to induce ovulation.9
We made further modifications to the C2 position of the
naphthalene core by the preparation of saturated ana-
logs of 4 (Scheme 2). However, as these boron contain-
ing coupling partners were not commercially available,
independent syntheses were required to prepare cou-
pling partners for triflate 5. We targeted ketones 12
and 15 as starting points, envisioning that vinyl boronic
esters 14 and 17 would serve as the cross-coupling com-
ponents. Ketone 12 is commercially available, while 15
could be prepared in a single step by the cycloaddition
of methylvinylsulfone and 2-(trimethylsiloxy)butadiene.
Enol triflates 13 and 16 were prepared and transformed
into 14 and 17 using standard conditions. The coupling
of 14 and 17 with 5 was conducted using conditions as
before with 5 and 10. The coupling product of 14 and
5 was oxidized with oxone. The coupled products were
deprotected using boron tribromide in the presence of
isobutene as a scavenger to prevent substrate olefin bro-
mination. The double bond of 19 was reduced using pal-
ladium black in an alcoholic solvent mixture. Following
additional purification, the HCl salts of compounds 19
and 20 were prepared for evaluation.
O
N
O
Me2N
O
S
R
HO
tamoxifen, 1
3
O
Et2N
O
N
O
Cl
R
HO
clomiphene, 2
4
For the benzothiophene (3) series we developed a syn-
thesis of 2-bromobenzothiophene 26, which like aryl tri-
flate 5 could also be coupled to commercially available
aryl boronic acids using Suzuki conditions (Scheme 3).
The synthesis of 26 began with 6-methoxybenzothioph-
ene (22).13 Attempts to di-brominate this material selec-
tively at the required 2- and 3-positions produced the
undesired 2,7-dibrominated product. Switching the
methyl protecting group to the sterically bulky, elec-
tron-withdrawing pivalate group allowed selective bro-
mination at the 2- and 3-positions to give compound
23. The basic side chain at the 3-position was then
installed using a conjugate addition–elimination proto-
col as previously described.12 Since the pivalate protect-
ing group proved labile under these conditions, it was
exchanged for a benzyl group. Then the benzothiophene
was oxidized to the corresponding sulfoxide using
hydrogen peroxide and TFA (62% yield) to obtain 24.
Under basic conditions (t-BuOK in THF) the phenolic
oxygen of 8 displaces the bromide at C3 to give coupled
product 25. After initial attempts to reduce the sulfoxide
of 25 using BBr3 or LAH failed to cleanly deliver the de-
sired product in reasonable yields, we found that treat-
ment with aqueous TiCl3 in a mixture of methanol
and chloroform gave the desired common intermediate
2-bromobenzothiophene 26 in good yield. Suzuki
cross-coupling of aryl boronic acids 10 with 26 followed
by debenzylation and HCl salt formation delivered the
final benzothiophene analogs 3 for evaluation.
Here, we describe SAR studies that led to the discovery
of SERMs that have antagonistic effects in the uterus
with limited impact on the HPO axis and do not stimu-
late the ovaries.10 We began our search for novel ovar-
ian selective SERMs by pursuing structure–activity
relationship (SAR) studies at the 2-aryl positions of
the benzothiophene (3) and the naphthalene (4) ring sys-
tems, since it is known that this area of the ER pharma-
cophore is tolerant to functional group diversity.11 Also
since it has been demonstrated that an oxygen linker
confers a substantial increase in estrogen antagonist po-
tency, we focused our efforts on the oxygen linked deriv-
atives of 3 and 4.12 Our primary strategy was to explore
functional groups that might limit brain penetration
with the hypothesis that restricting brain exposure
would limit impact on the HPO axis.
Our first goal was to develop syntheses of common
intermediates that would facilitate SAR studies. For
the naphthalene series (4) we developed a synthesis of
aryl triflate 5, which could be coupled to commercially
available aryl boronic acids using Suzuki conditions
(Scheme 1). The synthesis of 5 began with selective bro-
mination at C1 of 2-hydroxy-6-methoxynaphthalene (6)
by treatment with N-bromosuccinimide. Protection of
the phenol at C2 with benzyl bromide gave 7 in 86%
yield. Copper-mediated Ullman diaryl ether coupling
with catalytic quantities of copper triflate and 1.2 equiv-
alents of 812 gave 9 in low (30%) but reproducible yields.
Hydrogenolysis of the benzyl group of 9, followed by
triflate formation, delivered the desired common inter-
mediate 5 in good yield. Palladium-mediated cross-cou-
pling of aryl triflate 5 with a wide variety of aryl boronic
acids 10 provided a diverse array of 2-aryl-naphthalenes
11. The methyl ether protecting group at C6 could be
removed with boron tribromide followed by purification
and treatment with HCl to give the final naphthalene
analogs 4 suitable for biological assays.
Very few compounds met our demanding criteria for
ovarian safety. We began by evaluating their activities
in a series of in vitro assays to determine their ability
to bind the estrogen receptors and act as functional
antagonists of E2 with little or no agonism in uterine tis-
sue. We determined the binding affinity of our
compounds to full-length, recombinant human ERa
and ERb using competitive radioligand binding assays.
As shown in Table 1, all compounds in this study
bind to ERa with good to excellent affinity

3546
T. I. Richardson et al. / Bioorg. Med. Chem. Lett. 17 (2007) 3544–3549
O
Br
N
OH
a, b
OBn
N
O
c
MeO
OBn
MeO
O
OH
6
7
MeO
9
8
O
d, e
OTf
N
O
N
O
R
O
f
R'O
R
MeO
(HO)2B
5
11, R' = Me
4, R' = H
10
g, h
Scheme 1. Synthesis of naphthalene SERMs. Reagents: (a) NBS; (b) BnBr; (c) Cu(OTf)2, 8; (d) NH4HCO2, Pd(OH)2; (e) Tf2O; (f) Pd(OAc)2, PCy3,
CsF, 10; (g) BBr3; (h) HCl.
(Ki = 1.89–0.09 nM) and range from good to moderate
to non-selective for ERa over ERb (0.8- to 33-fold).
The potential for estrogen stimulation and antagonism
in uterine tissue was assessed in vitro by alkaline phos-
phatase quantitation in the Ishikawa human endome-
trial tumor cell line. All compounds in Table 1 exhibit
maximum agonist stimulation in the absence of E2 of
less than 30% at 1 lM in this assay. By comparison 4-
hydroxytamoxifen, the active metabolite of tamoxifen,
a known uterine agonist,6 exhibits stimulation of 123%
under the same conditions.10 In the antagonist mode,
these compounds block the stimulatory response of
1 nM E2 by at least 88% with IC50 values that range
from 2.1–22 nM.
OTf
O
O
B
a
b
O
S
S
S
12
14
OTMS
13
O
OTf
O
B
c
d
b
Having identified compounds with good to excellent
estrogen antagonism in vitro, we next determined their
ability to oppose E2 driven proliferative effects on uter-
ine tissue in vivo. Uterine antagonism was measured in
immature (3-week-old) female rats, which have low cir-
culating E2 levels and therefore are highly sensitive to
E2 stimulation of the uterus. Ethynyl estradiol (EE) pro-
duces a 3- to 4-fold increase in uterine wet weight when
administered to these animals at 0.1 mg/kg. As shown in
Table 1 all compounds in this study were found to be
potent uterine antagonists, inhibiting E2 stimulation
by at least 83% at 10 mg/kg with ED50 values less than
1 mg/kg.
O
+
MeO2S
MeO2S
SO2Me
SO2Me
17
15
16
O
N
O
N
O
R
R
O
h
e, f
OTf
R'O
20
14 or 17
MeO
18, R' = Me
19, R' = H
R = SO2
R = CHSO2Me
5
g
Scheme 2. Synthesis of saturated analogs. Reagents and condition: (a)
LDA, THF, PhNTf2; (b) bis(pinacolato)diboron, KOAc, PdCl2(dppf);
(c) toluene, heat; then TFA, MeOH; (d) Tf2O, CH2Cl2, 2,6-di-tert-
butyl-4-methylpyridine; (e) Pd(OAc)2, PCy3; CsF, 14 or 17; (f)
R = S ! SO2 oxone, MeOH; (g) BBr3, CH2Cl2, isobutene; (h) Pd
black, THF/EtOH, H2.
Our primary goal was to find compounds with little or
no impact on the HPO axis. We therefore designed com-
pounds that would have restricted access to the hypo-
Br
Br
g
a, b, c
d, e, f
Br
Br
O
OH
S
S
S
O
24
MeO
PivO
BnO
O
22
23
N
8
O
O
N
N
N
O
O
O
i
h
Br
Br
S
O
BnO
S
S
R
BnO
R'O
(HO)2B
26
25
R
27, R' = Bn
3, R' = H
10
j
Scheme 3. Synthesis of benzothiophene SERMs. Reagents: (a) NaSEt; (b) PivCl; (c) Br2; (d) KOH; (e) NaH, BnBr; (f) H2O2, TFA; (g) KOt-Bu; 8;
(h) TiCl3, HCl; (i) Pd(OAc)2, PCy3, CsF, 10; (j) NH4HCO2, Pd(OH)2; HCl.

Table 1. Binding affinity (Ki), functional activity (Ishikawa antagonist IC50 and efficacy and agonist efficacy), in vivo rat uterine antagonist potency (ED50), E2 ratio, Exposure: brain, plasma and log BB,
calculated properties: ClogP and polar surface area (PSA) for naphthalene and benzothiophene SERMsa
O
O
O
O
N
N
N
N
O
O
O
R
O
S
O
O
R
R
S
HO
HO
HO
HO
3
4
4l, R = SO2
4m
4n, R = CHSO2Me
Compound
Structure
R
Binding
Antagonism
Agonism
In vivo
Exposure
Plasmah
Calculated
ERa
Kib (nM)
ERb
Kib (nM)
Ishi
IC50 (nM)
Ishi %
inhibd
Ishi %
effe
ED50
(mpk)f
E2
ratiog
Brainh
logBBi
ClogPj
PSAk
c
4a
4b
4c
4d
4e
4f
4-SO2NMe2
4-SO2Me
4-SO2Et
4-SO2i-Pr
3-CN
0.40 0.11
0.47 0.21
1.22 0.54
0.73 0.51
0.24 0.08
0.21 0.10
0.32 0.13
0.57 0.38
0.24 0.09
0.18 0.07
0.20 0.02
1.03 0.21
1.89 0.17
1.69 0.46
1.32 0.34
4.19 1.51
2.17 0.89
0.35 0.21
0.36 0.24
0.48 0.18
1.16 0.66
0.28 0.11
0.16 0.05
0.29 0.05
2.10 0.15
4.11 0.36
2.69
2.8 0.8
10.7 6.8
6.1 2.1
4.9 0.9
18.5 6.1
2.1 0.7
4.7 1.0
6.7 1.4
3.2 1.2
4.8 1.2
3.9 1.5
7.2 2.2
15.7 2.7
8.0 4.2
16.2 5.0
22.2 6.3
14.5 4.7
99
88
95
96
94
100
96
95
93
97
96
98
93
99
95
97
95
3
27 14
0.44 0.19
0.071 0.055
0.14 0.02
0.37 0.05
0.37
1.83
1.1
46
103
43
108
660
357
NA
360
211
277
NA
151
137
418
NA
NA
NA
NA
NA
NA
-0.37
-0.81
-0.92
NA
6.52
5.69
6.22
6.52
6.76
6.66
6.66
5.79
7.47
7.47
7.54
4.74
4.44
5.49
8.98
5.95
6.48
87.7
84.5
84.5
84.5
65.7
62.2
62.2
62.2
41.9
41.9
41.9
84.5
84.5
84.5
41.9
84.5
84.5
6
2
2
4
4
2
7
4
5
5
4
3
2
3
4
5
29
19
11
4
1
7
1.75
1.06
5.62
1.28
2.94
5.67
4.6
NA
23 10
26 19
529
57
0.17
-0.57
-0.77
NA
3-OH
4-OH
0.13 0.03
0.3
4g
4h
4i
13 5
25 18
47
4-CONMe2
4-F
3-F
0.19 0.06
0.015
NA
21
15
9
8
2690
730
1.25
0.73
0.92
NA
4j
0.23 0.08
0.058
7.44
5.59
8.64
2.81
1.94
4.5
4k
4l
3,4-diF
13 10
19 12
27 11
3490
NA
NA
NA
NA
NA
NA
0.25 0.11
0.34 0.16
0.26 0.08
0.16 0.07
0.62 0.12
0.56 0.15
4m
4nl
3a
3b
3c
NA
NA
1.09
24
29
7
8
i-Pr
SO2Me
SO2Et
0.091 0.016
0.63 0.31
0.48 0.31
0.63 0.04
13.6 5.4
15.7 5.1
NA
NA
24 21
16 13
2.19
2.85
NA
a Experimental values represent the average with the standard deviation SD for multiple determinations (at least n = 2) between the assay values. Values without SD notation were run once (n = 1). NA
means data not available.
b Ki values determined by radioligand binding assay using 3H-estradiol.
c Ishikawa antagonism IC50 values are the compound concentration needed to block 50% of 1 nM E2 stimulation as determined by alkaline phosphatase quantitation.
d Ishikawa antagonism is the efficacy (%) of blocking 1 nM E2 stimulation.
e Ishikawa agonism is the % increase in alkaline phosphatase compared to control.
f Female Sprague–Dawley rats, 6 per group and 19–21 days of age, were orally treated with ethynyl estradiol (0.1 mg/kg) and 10, 1.0, 0.1 or 0.01 mg/kg of SERM for 3 days.
g The ratio of serum E2 levels (the ratio of treated to vehicle control) after oral dosing with compound for 10 days at a dose of 30 times that of the immature rat ED50
h Brain (ng/g) and plasma (ng/mL) exposure at the 6 h time point following a 10 mg/kg oral dose.
i logBB = log(Cbrain/Cblood).
.
j The calculated logarithm of the n-octanol/water partition.
k
2
˚
Polar surface area (A ).
l The more potent enantiomer was tested but the configuration was not assigned.

3548
T. I. Richardson et al. / Bioorg. Med. Chem. Lett. 17 (2007) 3544–3549
thalamus in the brain. A common measure of blood–
brain barrier (BBB) penetration is logBB, the log of
the ratio of plasma to brain concentration at a single
time point. Since it is time consuming and resource-
intensive to measure logBB, several methods to predict
blood–brain barrier (BBB) penetration have been pro-
posed.14 These methods are based on molecular descrip-
tors that can be readily calculated, with the most
important being ClogP, polar surface area (PSA), and
the number of rotatable bonds.15 Since our efforts were
limited to the 2-aryl positions of the benzothiophene (3)
and naphthalene (4) ring systems, we focused on the
addition of polar functional groups that would be toler-
ated by the receptor at this location, increase PSA, and
therefore lower ClogP. As can be seen in Table 1, on the
naphthalene scaffold 4, the addition of sulfone func-
tional groups (4b, 4c, and 4d) on the 2-aryl ring pro-
ties of sulfonamide 4a were similar to those of the sul-
fone analogs. It also did not raise E2 levels although
its overall plasma exposure was lower as well indicating
it may have poor overall absorption or higher
metabolism.
The cyano and amide substituted compounds (4e and
2
4h) possess PSAs (66 and 62 A ) that are intermediate
˚
between the sulfone and fluoro analogs. The corre-
sponding ClogPs (6.8–5.8) are closer to the sulfones.
For these intermediate cases, we observed some com-
pounds that stimulated E2 with the same propensity as
the fluoro analogs and others that behaved more like
the sulfones. The amide analog 4h for example has a
ClogP similar to sulfone 4b but stimulated E2 levels
to the same extent as difluoro 4k. The cyano compound
4e, with nearly equal levels of brain and plasma expo-
sure, produced an E2 plasma concentration that was
5.62 times vehicle control.
2
˚
duced compounds with higher PSAs (84.5 A ) than the
corresponding fluoro substituents (compounds 4i, 4j,
2
4k; PSA = 42 A ). As expected, higher PSA translated
˚
into lower ClogP for the sulfone substituted analogs
(ClogP = 5.69–6.52) compared to the fluoro analogs
(ClogP = 7.47–7.54). These data correlate nicely with
measured logBB. The sulfone substituted analogs had
logBB < 0, while the fluoro analogs had logBB > 0,
indicating that the sulfone analogs might have less pro-
pensity to antagonize the HPO axis compared to the flu-
oro analogs.
The phenol isomers 4f and 4g are particularly interest-
ing. Since the calculated molecular descriptor PSA does
not take into account differences between isomers, both
of these possess the same value for PSA. The ClogPs for
the meta and para isomers are also identical and the cor-
responding brain to plasma ratios are nearly the same as
well. For these two, the meta isomer 4f behaved more
like the sulfones, while the para isomer 4g was stimula-
tory. The saturated analogs 4l, 4m, and 4n all have cal-
culated PSA values similar to the aryl sulfones (4b, 4c,
4d) and even have ClogP values that are lower. Unfor-
tunately, these favorable calculated properties did not
translate into superior performance in the E2 stimula-
tion assay. We observed a dramatic difference between
the dihydrothiopyran 4l, one of the most stimulatory
compounds (8.6-fold), and its fully saturated analog
4m, which caused only 2.8-fold stimulation. The 4-
(methylsulfonyl)-cyclohexene analog 4n performed al-
most as well as the ethyl sulfone 4c. The 2-aryl analogs
of the benzothiophene scaffold showed similar trends
compared to the naphthalenes. The lipophilic isopropyl
compound 3a has a low PSA and correspondingly high
ClogP. Not surprisingly, it produced high levels of E2
compared to vehicle control. The sulfone analogs 3b
Inhibition of the HPO axis results in overproduction of
LH and FSH which in turn stimulate the ovaries to pro-
duce E2, therefore plasma E2 levels are a sensitive bio-
marker for ovary stimulation. In order to screen for
ovarian stimulation, mature intact female rats were gi-
ven compound once daily for 10 days with an oral dose
30 times the ED50 in the immature rat assay. As can be
seen in Table 1, for the extreme cases of sulfone and flu-
oro substitution, plasma E2 levels correlate well with
logBB. For the sulfone analogs (4b, 4c, 4d), the ratio
of serum E2 levels compared to vehicle control was be-
tween 1.06 and 1.75, indicating minimal ovarian stimu-
lation. By contrast, the fluoro analogs (4i, 4j, 4k)
produced plasma levels of E2 that were 4.6–7.4 times
that of vehicle control. The calculated molecular proper-
O
N
O
N
SO2CH3
O
O
OH
HO
HO
4f
4b
60
50
40
30
20
10
0
60
Proestrus
50
40
30
20
10
0
Cont 0.05 0.5
1
3
Cont
3.6
7.2
0.12 1.2
Compound Dose (mg/kg)
Figure 1. Serum E2 levels in rats treated with 4b or 4f for 35 days: error bars = SEM.

T. I. Richardson et al. / Bioorg. Med. Chem. Lett. 17 (2007) 3544–3549
3549
and 3c possess PSA and ClogP values very similar to
the corresponding naphthalene analogs 4b and 4c.
Although these did not stimulate E2 to the same extent
as isopropyl analog 3b, unfortunately, their ED50s in the
immature rat uterine assay were 4–8 times that of the
naphthalenes. As a result, they were dosed higher in
the 10-day E2 assay, and although the trends are similar,
the best benzothiophenes produced E2 levels that were
twice vehicle control. The lower activity of the benzothi-
ophenes in the immature rat uterine assay may be due to
overall lower absorption and/or higher metabolism than
the corresponding naphthalenes (data not available).
therapeutic potential for the treatment of leiomyomas in
premenopausal women.
References and notes
1. (a) Stewart, E. A. Lancet 2001, 357, 293; (b) Cramer, S. F.;
Patel, A. Am. J. Clin. Pathol. 1990, 94, 435; (c) Buttram,
V. C., Jr.; Reuter, R. C. Fertil. Steril. 1981, 36, 433.
2. Walker, C. L.; Stewart, E. A. Science 2005, 308, 1589.
3. Vollenhoven, B. Baillieres Clin. Obstet. Gynaecol. 1998,
12, 169.
4. Lethaby, A.; Vollenhoven, B.; Sowter, M. Br. J. Obstet.
Gynecol. 2002, 109, 1097.
5. (a) McDonnel, D. P.; Norris, J. D. Science 2002, 296,
1642; (b) Katzenellenbogen, B. S.; Katzenellenbogen, J. A.
Science 2002, 295, 2380; (c) Jordan, V. C. J. Med. Chem.
2003, 46, 883.
6. Sato, M.; Rippy, M. K.; Bryant, H. U. FASEB J. 1996, 10,
905.
7. Lumsden, M. A.; West, C. P.; Baird, D. T. Clin.
Endocrinol. 1989, 31, 335–343.
8. Sadan, O.; Ginath, S.; Sofer, D.; Rotmensch, S.; Debby,
A.; Glezerman, M.; Zakut, H. Eur. J. Obstet. Gynecol.
Reprod. Biol. 2001, 96, 183.
The effects of lead compounds on the uterus and ovaries
were studied in 6-month-old ovary-intact female rats
treated with oral doses of compound at 1, 10, 30, and
60 times the ED50 in the immature rat assay (Fig. 1).
Dose-dependent decreases in uterine wet weight (data
not shown) were observed with the total area of uterine
wall reduced in both the endometrium and myome-
trium, confirming that the compounds are potent antag-
onists in both compartments of the uterus in the
presence of E2 over several estrous cycles. The extent
of ovarian stimulation was determined by measuring
serum E2 levels. As shown in Figure 1, treatment with
compounds 4b and 4f resulted in serum E2 levels that
are similar to vehicle-treated animals. Even at the high-
est doses, the E2 levels are within the normal proestrus
range for Sprague–Dawley rats (52 ng/mL).
9. Huppert, L. C. Fertil. Steril. 1979, 31, 1.
10. (a) Geiser, A. G.; Hummel, C. W.; Draper, M. W.; Henck,
J. W.; Cohen, I. R.; Rudmann, D. G.; Donnelly, K. B.;
Adrian, M. D.; Shepherd, T. A.; Wallace, O. B.; McCann,
D. J.; Oldham, S. W.; Bryant, H. U.; Sato, M.; Dodge, J.
A. Endocrinology 2005, 146, 4524; (b) Hummel, C. W.;
Geiser, A. G.; Bryant, H. U.; Cohen, I. R.; Dally, R. D.;
Fong, K. C.; Frank, S. A.; Hinklin, R.; Jones, S. A.;
Lewis, G.; McCann, D. J.; Rudmann, D. G.; Shepherd, T.
A.; Tian, H.; Wallace, O. B.; Wang, M.; Wang, Y.; Dodge,
J. A. J. Med. Chem. 2005, 48, 6772.
11. Grese, T. A.; Cho, S.; Finley, D.; Godfrey, A. G.; Jones,
C. D.; Lugar, C. W.; Martin, M. J.; Matsumoto, K.;
Pennington, L. D.; Winter, M. W.; Adrian, M. D.; Cole,
H. W.; Magee, D. E.; Phillips, D. L.; Rowley, E. R.; Short,
L. L.; Glasebrook, A. L.; Bryant, H. U. J. Med. Chem.
1997, 40, 146.
12. Palkowitz, A. D.; Glasebrook, A. L.; Thrasher, K. J.;
Hauser, K. L.; Short, L. L.; Phillips, D. L.; Muehl, B. S.;
Sato, M.; Shetler, P. K.; Cullinan, G. J.; Pell, T. R.;
Bryant, H. U. J. Med. Chem. 1997, 40, 1407.
13. Graham, S. L.; Shepard, K. L.; Anderson, P. S.; Baldwin,
J. J.; Best, D. B.; Christy, M. E.; Freedman, M. B.;
Gautheron, P.; Habecker, C. N.; Hoffman, J. M.; Lyle, P.
A.; Michelson, S. R.; Ponticello, G. S.; Robb, C. M.;
Schwam, H.; Smith, A. M.; Smith, R. L.; Sondey, J. M.;
Strohmaier, K. M.; Sugrue, M. F.; Varga, S. L. J. Med.
Chem. 1989, 32, 2548.
In conclusion, SAR studies on the 2-aryl positions of the
benzothiophene (3) and the naphthalene (4) ring systems
produced novel SERMs that are potent antagonist in
uterine tissue but cause limited ovarian stimulation, as
judged by increased plasma E2 levels in rats. We identi-
fied a loose correlation between calculated properties
(ClogP and PSA) and measured properties (logBB
and plasma E2). Substituents that confer high PSA
and low ClogP, such as aryl sulfones, produced com-
pounds that caused slight or no increases in serum E2
levels, while the corresponding fluoro analogs with low
PSA and high ClogP increased plasma E2 significantly.
For intermediate cases such as phenols or amides, the
relationship was not as predictive. We cannot rule out
the possibility that stimulatory compounds such as cya-
no 4e and amide 4f have a direct effect on the ovaries.
Although saturated sulfone analogs possess favorable
calculated properties, they did not behave as well as
the aryl sulfones. These results demonstrate it is possible
to obtain a wide therapeutic window between desired
antagonistic effects on the uterus and undesirable effects
on the ovaries in a rodent animal model. As such, com-
pounds such as sulfone 4b and meta-phenol 4f may have
14. (a) Goodwin, J. T.; Clark, D. E. J. Pharmacol. Exp. Ther.
2005, 315, 477; (b) Clark, D. E. J. Pharm. Sci. 1999, 99,
815.
15. Winkler, D. A.; Burden, F. R. J. Mol. Graphics Modell.
2004, 22, 499.