Estrogen receptor β selective ligands: Discovery and SAR of novel heterocyclic ligands

Article abstract of DOI:10.1016/j.bmcl.2005.08.010

A series of ligands with varying heterocyclic cores and substituents that display a range of selectivity's (up to >100x) for ER-β over ER-α are reported.

Full text of DOI:10.1016/j.bmcl.2005.08.010

Bioorganic & Medicinal Chemistry Letters 15 (2005) 5562–5566
Estrogen receptor b selective ligands: Discovery and SAR of
novel heterocyclic ligands
Richard Chesworth,a,* Matthew D. Wessel,b Lisa Heyden,a F. Michael Mangano,a
Michael Zawistoski,a Laura Gegnas,a David Galluzzo,a Bruce Lefker,a
Kimberly O. Cameron,a Jeanne Tickner,a Bihong Lu,c Tessa A. Castleberry,c
Donna N. Petersen,c Amy Brault,c Pia Perry,c Oicheng Ng,c,
Thomas A. Owen,c Lydia Pan,c Hua-Zhu Ke,c Thomas A. Brown,c
David D. Thompsonc and Paul DaSilva-Jardinea
aCardiovascular and Metabolic Diseases, Medicinal Chemistry, Pfizer Global Research and Development,
Groton Laboratories, Eastern Point Road, Groton, CT 06340, USA
bComputational Chemistry, Pfizer Global Research and Development, Groton Laboratories,
Eastern Point Road, Groton, CT 06340, USA
cCardiovascular and Metabolic Diseases, Biology, Pfizer Global Research and Development,
Groton Laboratories, Eastern Point Road, Groton, CT 06340, USA
Received 8 May 2005; revised 1 August 2005; accepted 2 August 2005
Available online 10 October 2005
Abstract—A series of ligands with varying heterocyclic cores and substituents that display a range of selectivityÕs (up to >100x) for
ER-b over ER-a are reported.
ꢀ 2005 Elsevier Ltd. All rights reserved.
The estrogen receptor (ER) is a member of the nuclear
hormone superfamily of receptors.1 The ER mediates
the regulation of a number of important physiological
processes, including bone mineral density (BMD) and
lipid levels.2 Post-menopausal women are often treated
with hormone (estrogen) replacement therapy (HRT)
to maintain BMD and relieve post-menopausal symp-
toms, such as hot flashes and night sweats.3 HRT
though, is not without its drawbacks, such as an in-
creased risk in both breast and uterine cancers. For this
reason, compliance is poor and many women stop tak-
ing HRT after less than one year of treatment.4 Selective
estrogen receptor modulators (SERMs), such as raloxif-
ene,5 have overcome the increased risk of breast and
uterine cancers, but unfortunately do not provide full
bone protection and do not ameliorate hot flashes.6
Until the recent discovery of a second estrogen receptor
(designated ER-b),7,8 it was believed that only a single
estrogen receptor existed (now known as ER-a). This
discovery has led to an intensive effort to define the
roles that the individual receptors play in the hope of
designing an improved SERM/HRT.
Both ER-a and ER-b bind the endogenous ligand 17-b
estradiol 19 with high potency (see Fig. 1). The tissue
distribution of the receptors suggests that their roles
may not be redundant.10 Various hypotheses for the
function of each receptor have been generated.11–14 To
OH
Me
OH
O
H
H
H
O
OH
Keywords: Estrogen receptor; ER beta selective ligands.
HO
HO
*
Corresponding author at present address: En Vivo Pharmaceuticals,
480 Arsenal Street, Watertown, MA 02472, USA. Tel.: +1 617 247
1652; fax: +1 617 746 8877; e-mail: rchesworth@gmail.com
Deceased May 19, 2002.
1
2
Genistein
17-β Estradiol
Figure 1. Estrogen receptor ligands.
0960-894X/$ - see front matter ꢀ 2005 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2005.08.010

R. Chesworth et al. / Bioorg. Med. Chem. Lett. 15 (2005) 5562–5566
5563
(b)
test these hypotheses and determine the function of each
receptor, selective compounds for each receptor are
NO2
F
NO2
NH
H2N
(a)
+
sought after, in particular ligands for ER-b.
OMe
ER-a is a larger protein than ER-b. ER-a contains 595
amino acids, whereas ER-b has only 485. Although the
two receptors have overall about 56% sequence homol-
ogy, in the ligand binding domain only two of the 23
amino acids that make direct contact to the encapsulat-
ed ligand are different (L384M and M421I), both of
which are highly conservative substitutions.15 The net
effect of these differences is that ER-b has a slightly
smaller ligand binding pocket than ER-a16 and it has
been found that sterically demanding ligands can
display selectivity for ER-a.17 Compounds with some
selectivity for ER b include genistein 2, tetra-hydro-
chrysenes,18 diarylpropionitriles,19 2-aryl benzothioph-
enes,20 1,3,5 triazines,21 and most recently, the
benzoxazole ERB-041.22 In this paper we describe
our own efforts toward the discovery of ligands that
are selective for ER-b (Fig. 2).
OMe
3
NH2
NH
NH
NH
R
(c)
(d)
O
OMe
OMe
5a-e
4
N
N
N
N
(e)
R
R
A series of benzimidazoles was prepared via standard
procedures as follows: 2-Fluoro nitrobenzene was treated
with para-anisdine in the presence of K2CO3 at 160 ꢁC to
give the nitro aniline 3, which upon reduction with hydro-
gen over Pd/C gave the bis-aniline 4. This compound was
condensed with various carboxylic acids to give the
amides 5a–e, which upon treatment with hot acetic acid
underwent acid-catalyzed dehydration to form the benz-
imidazoles 6a–e. The methyl ethers were removed using
BBr3 to give the desired target compounds 7a–e. This se-
quence allowed the rapid synthesis of a large number of
C-2-substituted benzimidazole analogs from the common
intermediate 4 (see Scheme 1).
OH
7a-e
OMe
6a-e
Scheme 1. Synthesis of benzimidazole derivatives. Reagents and
conditions: (a) K2CO3, neat, 160 ꢁC; (b) H2, Pd/C, MeOH, dioxane,
HCl; (c) RCO2H, 1-propane phosphonic cyclic anhydride, Et3N,
solvent; (d) AcOH, 65 ꢁC; (e) BBr3, CH2Cl2, À78 ꢁC to rt.
azole (8). The synthesis of the imidazole23 11 and the in-
dole24,25 12–15 compounds has been previously
described (see Scheme 2).
To screen compounds for their affinity for ER-b versus
ER-a, an in vitro ligand binding assay using recombi-
nant ER-a and ER-b using [3H] 17b-estradiol as the
radioligand was employed26.
Isoxazole 8 was prepared by acylating the dianion of the
oxime derived from 1-(4-methoxy-phenyl)-2-phenyletha-
none 9 with trifluoroacetic acid ethyl ester. The resultant
hemiacetal 10 was dehydrated with p-TsOH, followed
by O-demethylation with BBr3, to give the desired isox-
A number of standard compounds were evaluated in the
binding assay and their affinities for both receptors were
found to be in agreement with previously reported val-
ues.9 As shown in Table 1, 17-b-estradiol 1 demonstrat-
R1
N
CF3
N
O
N
HO
(a)
HO
(b)
7a R1=Ph
8
7b R1=2-thienyl
OH
N
O
7c R1= N-methyl-2-pyrroyl
7d R1=3-methyl-4-isoxazyl
MeO
MeO
R2
7e R1=2-Cl phenyl
9
H
N
HO
CF3
O
CF3
(c), (d)
H
N
O
R1
CF3
N
N
N
12 R1, R2=H
HO
MeO
13 R1, R2=OH
HO
10
8
14 R1=OH, R2=H
11
15 R1=H, R2=OH
Scheme 2. Synthesis of isoxazole (8). Reagents and conditions: (a)
NH2OH, EtOH, pyridine; (b) n-BuLi, CF3CO2Et; (c) TsOH, toluene,
D; (d) BBr3 0 ꢁC to rt.
Figure 2. Heterocyclic ligands.

5564
R. Chesworth et al. / Bioorg. Med. Chem. Lett. 15 (2005) 5562–5566
Table 1. IC50s and binding selectivities of ligands26
azole core allowed for a more rapid synthesis of analogs
compared to the other templates, and in particular,
allowed for a rapid exploration of the vector defined
by the C-2 substituent.
Compound
ER-a
ER-b
Selectivity
(for ER-b)
(nM)a
(nM)a
1
0.75 0.60
1.18 1.04
1.30 0.69
137 42.4
49.7 6.06
28.1 0.05
14.3 1.20
149b
0.63
22.5
13.1
>126
105
2
29.4 4.33
1800 569
>6260 3330
2940 1400
1160 271
1730b
The potency and selectivity could be increased by
replacement of the C-2 phenyl group of 7a with select
5-membered heterocycles. Introduction of a 2-thienyl
group 7b results in an analog that is over 100· selective
for ER-b over ER-a. Analogs containing an ortho sub-
stituent on the 5-membered heterocycle, such as N-
methyl-2-pyrrolyl 7c or 3-methyl-4-isoxazyl group 7d,
have improved potency at ER-b, whilst retaining the
selectivity for ER b. It is interesting to speculate on
the origin of improved potency and selectivity versus
the prototypical phenyl analog 7a. One of the two differ-
ences in the ligand binding site between ER a and ER-b
is a Met 421:Ile 373 residue change, giving rise to a small
pocket in the ER-b structure (see Fig. 3). As suggested
from docking the benzimidazole 7c into the ER crystal
structures, it may take advantage of this difference by
positioning its N-methyl group into the small pocket
in ER-b. It is interesting to note that the ortho-chloro
phenyl analog 7e displays essentially the same selectivity
and potency as the unsubstituted phenyl analog 7a as
based on this model, it too may have been expected to
take advantage of this small pocket in ER-b. To fully
understand the factors responsible for the observed
selectivity for ER b, crystal structures of one of the
selective agents bound to both of the ERs would be
highly useful.
7a
7b
7c
7d
7e
8
81.1
11.6
3.5
8.5 4.9
403 207
6950b
2.4 0.69
15.1 3.04
171b
11
12
13
14
15
26.7
40.6
8.0
94.8 21.5
24.0 13.9
8310 2397
11.9 3.01
1.12 0.73
355 224
21.8
23.4
a
Value represents the standard deviation.
b Only one independent experiment was run.
ed an approximately equal affinity for each receptor,
while genistein 2 showed a slight bias for ER b
(14.3·). During the course of high-throughput screen-
ing, 2,3-diphenyl indole 1227 emerged as a compound
showing reasonable selectivity for ER-b (ca. 40·),
although potency for this receptor is modest. Although
this compound is commercially available, it has not been
previously characterized as an ER ligand. The majority
of ER ligands contain a phenol on the A-ring biostere
(cf. to 17b-estradiol). This hydroxyl group of the phenol
is responsible for ca. 1.9 kcal/mol interaction energy
with ER a.28 We anticipated that the introduction of
appropriately positioned phenol(s) into 2,3-diphenyl
indole 12 would improve the potency, while retaining
a reasonable degree of selectivity. A systematic SAR sur-
vey of the indole 12 found that the introduction of a
phenolic group into the para position of each of the pen-
dant phenyl rings gave 13, which demonstrated im-
proved potency at ER-b by an order of magnitude,
though its selectivity for ER-b was slightly diminished.
The individual removal of each phenol clearly indicated
which of these was more important for potency. As can
be seen in Table 1, the mono-phenolic indole 14 has a
much improved potency compared to indole 15. We
undertook a systematic investigation of 14 in the hope
of improving the selectivity for ER-b. Initial efforts
focused on varying the central core, as different cores al-
low for the different spatial arrangement of substituents.
This led to the identification of the isoxazole 8, imidaz-
ole 11, and the benzimidazoles 7a-d. The isoxazole 8 re-
tains potency for ER-b, but unfortunately loses most of
the selectivity that is seen in the indole series. The triflu-
oromethyl imidazole 11 however, maintains selectivity
for ER-b but loses an order of magnitude in potency
for this receptor compared to the indole 14. This may
be due in part to the energetic penalty paid for the des-
olvation of the acidic (pKa ca. 9)23 NH group of the
imidazole and the general increase in the polarity of
the central core. Replacement of the indole core with a
benzimidazole ring system produced the analog 7a,
which, although maintaining a degree of selectivity, lost
two orders of magnitude in potency compared to 14.
Despite the loss of potency, switching to the benzimid-
A number of compounds were profiled further for estro-
genic activity in whole cell functional assays. Cell lines
were chosen that expressed predominantly ER-a or
ER-b, respectively. MCF-7 cells are derived from hu-
man breast cancer tissue and these cells predominantly
contain ER-a.30 Conversely, primary granulosa cells,
which are derived from rat ovaries, do not contain
ER-a but contain endogenous ER b.31 Each cell line
was independently transiently transfected with an estro-
gen responsive luciferase reporter vector (ERE3-TK-
lux)32 (see Table 2).
Figure 3. Binding mode of benzimidazole 7c (orange) when docked
into the X-ray crystal structure of ER-b.16,29

R. Chesworth et al. / Bioorg. Med. Chem. Lett. 15 (2005) 5562–5566
Table 2. Estrogenic activity of selected compounds in whole cells
5565
%
Compound
MCF-7 EC50 (nM)
Activationa
%
Granulosa EC50 (nM)
Activationb
Estradiol 1
0.01
>1000
243
100
21
0.14
40.8
6.70
1.09
100
101
93
7c
11
14
87
83
38.5
78
a Percent activation observed at 1 lM in MCF-7 cells.
b Percent activation observed at 1 lM in granulosa cells.
References and notes
N
H
N
N
1. Manglesdorf, D. J.; Thummel, C.; Beato, M.; Herrlich, P.;
Schutz, G.; Umesono, K.; Blumberg, B.; Kastner, P.;
Mark, M.; Chambon, P.; Evans, R. M. Cell 1995, 83, 835.
2. Nasr, A.; Breckwoldt, M. Gynecol. Endocrinol. 1998, 12, 43.
3. Ulmsten, U. Proc. Soc. Exp. Biol. Med. 1998, 217, 2.
4. Cauley, J. A.; Cummings, S. R.; Black, D. M.; Mascioli, S.
R.; Seeley, D. G. D. Am. J. Obstet. Gynecol. 1990, 163,
438.
CF3
N
N
HO
HO
7c
11
Figure 4. Structures of the ER-b selective compounds 7c and 11.
5. Scott, J. A.; Da Camara, C. C.; Early, J. E. Am. Fam.
Physician 1999, 60, 1131.
The synthetic analogs 11 and 14 displayed partial to full
agonism in both the MCF-7 and granulosa cell lines.
These analogs, though, were considerably more potent
in the granulosa cells, presumably due to the higher
affinity of the analogs for ER-b. The highly selective
benzimidazole 7c appears to be a potent full agonist in
the granulosa cells (ER-b), while displaying little activity
in the MCF-7 cells presumably due to its weak binding
activity at ER-a . Due to the differing nature of the two
cell lines (both background and species), only limited
conclusions can be drawn about the functional selectiv-
ity of the compounds, though it does seem that binding
selectivity is reflected in the potency of the functional
assays.
6. Miller, C. Curr. Pharm. Des. 2002, 8, 2089.
7. Kuiper, G. C.; Carlsson, B.; Enmark, E.; Pelto-Huikko;
˚
Nilsson, S.; Gustafsson, J. A. Proc. Natl. Acad. Sci.
U.S.A. 1996, 93, 5925.
8. Mosselmann, S.; Polman, J.; Dijkema, R. FEBS Lett.
1996, 392, 49.
9. Kuiper, G. C.; Carlsson, B.; Grandien, K.; Enmark, E.;
˚
Haggblad, J.; Nilsson, S.; Gustafsson, J. A. Endocrinology
1997, 138, 863.
˚
10. Couse, J. F.; Lindzey, J.; Grandien, K.; Gustafsson, J. A.;
Korach, K. S. Endocrinology 1997, 138, 4613.
11. Katzenellenbogen, B. S.; Korach, K. S. Endocrinology
1997, 138, 861.
12. Shughrue, P. J.; Komm, B.; Merchenthaler, I. Steroids
1996, 61, 678.
˚
13. Wang, L.; Andersson, S.; Warner, M.; Gustafsson, J. A.
Proc. Natl. Acad. Sci. U.S.A. 2001, 98, 2792.
14. Tremblay, G. B.; Trembaly, A.; Copeland, N. G.; Gilbert,
D. J.; Jenkins, N. A.; Labrie, F.; Giguere, V. Mol.
Endocrinol. 1997, 11, 353.
In summary, we have found compounds that have dis-
played up to 100x selectivity for ER-b in competition
binding assays that are also functional agonists in whole
cells, as exemplified by the analogs 7c and 11 (Fig. 4).
We have also demonstrated that the ER receptors are
tolerant of various heterocyclic cores and that, despite
the high homology between the two receptors, ligands
can be made to exploit the small differences between
these two receptors to achieve binding selectivity. We
anticipate that these compounds, upon further charac-
terization, will become useful pharmacological tools
for determining the role ER-b plays in ER mediated
physiology.
15. Wessel, M. Unpublished data.
16. (a) Pike, A. C. W.; Brzozowski, A. M.; Hubbard, R. E.;
Bonn, T.; Thorsell, A. G.; Engstrom, O.; Ljunggren, J.;
˚
Gustafsson, J. A.; Carlquist, M. EMBO 1999, 18, 4608; (b)
Brzozowski, A. M.; Pike, A. C.; Dauter, Z.; Hubbard, R.
E.; Bonn, T.; Engstrom, O.; Ohman, L.; Greene, G. L.;
˚
Gustafsson, J. A. Nature 1997, 389, 753.
17. (a) Chesworth, R.; Zawistoski, M. P.; Lefker, B. A.;
Cameron, K. O.; Day, R. F.; Mangano, F. M.; Rosati, R.
L.; Colella, S.; Petersen, D. N.; Brault, A.; Lu, B.; Pan, L.
C.; Perry, P.; Ng, O.; Castleberry, T. A.; Owen, T. A.;
Brown, T. A.; Thompson, D. D.; DaSilva-Jardine, P.
Bioorg. Med. Chem. Lett. 2004, 14, 2729; (b) Renaud, J.;
Bischoff, S. F.; Buhl, T.; Floersheim, P.; Fournier, B.;
Halleux, C.; Kallen, J.; Keller, H.; Schlaeppi, J. M.; Stark,
W. J. Med. Chem. 2003, 46, 2945.
18. Meyers, M. J.; Sun, J.; Carlson, K. E.; Katzenellenbogen, B.
S.; Katzenellenbogen, J. A. J. Med. Chem. 1999, 42, 2456.
19. Meyers, M. J.; Sun, J.; Carlson, K. E.; Marriner, G. A.;
Katzenellenbogen, B. S.; Katzenellenbogen, J. A. J. Med.
Chem. 2001, 44, 4230.
Acknowledgments
We thank Arthur Brosius, Hao Mee, and Navin Varsh-
ney for technical assistance, as well as Ralph Robinson
and Christine Kowalewski for their help in the prepara-
tion of the manuscript.
Supplementary data
20. Schopfer, U.; Schoeffter, P.; Bischoff, S. F.; Nozulak, J.;
Feuerbach, D.; Floersheim, P. J. Med. Chem. 2002, 45,
1399.
21. Henke, B. R.; Consler, T. G.; Go, N.; Hale, R. L.;
Hohman, D. R., et al. J. Med. Chem. 2002, 45, 5492.
Supplementary data associated with this article can be
found, in the online version, at doi:10.1016/
j.bmcl.2005.08.010.

5566
R. Chesworth et al. / Bioorg. Med. Chem. Lett. 15 (2005) 5562–5566
22. Harris, H. A.; Albert, L. M.; Leathbury, Y.; Malamas, M.
S.; Mewshaw, R. E.; Miller, C. P.; Kharode, Y. P.;
Marzolf, J.; Komm, B. S.; Winneker, R. C.; Frail, D. E.;
Henderson, R. A.; Zhu, Y.; Keith, J. C., Jr. Endocrinology
2003, 144, 4241.
23. Lombardino, J. G.; Wiseman, E. H. J. Med. Chem. 1974,
17, 1182.
24. Iyer, R. N.; Gopalchari, R. Indian J. Chem. 1966, 4, 520.
25. Strohmeier, J.; von Angerer, E. Arch. Pharm. (Weinheim)
1987, 320, 407.
26. Competition binding assay. All compounds, except for 7e
and 12, were run n P 2 and each experiment was
measured in triplicate. The ability of various compounds
to inhibit [3H]estradiol binding was measured by a
competition binding assay using dextran-coated charcoal,
as has been described by Leake, R. E.; Habib, F. In
Steroid Hormone Receptors a Practical Approach; Green
B.; Leake, R. E., Roe, Eds.; IRL Press: Oxford, 1990; pp
67–92. 293T cell extracts expressing either hER.a or
hER.b were incubated in the presence of increasing
concentrations of competitor and a fixed concentration
of [3H]estradiol (141 Ci/mmol, New England Nuclear,
Boston, MA) in 50 mM Tris–HCl, pH 7.4, 1.5 mM
EDTA, 50 mM NaCl, 10% glycerol, 5 mM DTT, and
0.5 mg/mL. b-Lactoglobulin in a final volume of 0.2 mL.
All competitors were dissolved in dimethylsulfoxide. Final
concentration of the receptor was 50 pM with 0.5 nM
[3H]estradiol. After 16 h at 4ꢁC, dextran-coated charcoal
(20 lL) was added. After 15 min at room temperature, the
charcoal was removed by centrifugation and the radioac-
tive ligand present in the supernatant was measured by
scintillation counting. All reagents were obtained from
Sigma (St. Louis, MO)..
29. We used the Sybyl program (Tripos Associates, St. Louis,
MO.) for docking structures 7c manually into the active
sites of ER-a and ER-b crystal structures (pdb codes
1L2I,1QKM, respectively). The compounds were manu-
ally overlapped with the co-crystallized ligands and then
allowed to relax in the active site of the protein using the
Sybyl force field, Gasteiger–Huckel charges on the ligands,
and Kollman All Atom charges on the protein. The
phenolic substituents of each structure were held in place
to mimic the phenolic substituents of the co-crystallized
ligands. The resulting conformations were then used to
analyze contact points between the protein and the ligand
for the purpose of explaining potential selectivity differ-
ences in the active sites of ER-a and ER-b based on our
predicated binding modes.
30. Mor, G.; Sapi, E.; Abrahams, V. M.; Rutherford, T.;
Song, J.; Hao, X.-Y.; Muzaffer, S.; Kohen, F. J. Immunol.
2003, 170, 114.
31. (a) OÕBrien, M. L.; Park, K.; In, Y.; Park-Sarge, O.-K.
Endocrinology 1999, 140, 4530; (b) del Rio, B.; Pedrero, J.
M. G.; Martinez-Campa, C.; Zuazuo, P.; Lazo, P. S.;
Ramos, S. J. Biol. Chem. 2004, 279, 38294.
32. Whole cell assays. Estrogenic activity in human breast
cancer MCF7 cells and primary rat granulosa cells was
assessed by transient transfection of an estrogen respon-
sive ERE3-TK-lux luciferase reporter vector essentially as
we have described previously in other cell backgrounds
(Petersen et al.). The MCF7 cell activity was considered to
be mediated through ER-a and the granulosa activity was
considered to be mediated through ER-b. MCF7 cells
were obtained from ATTC (Manassas, VA) and transfec-
ted with Lipofectamine Plus (Gibco/BRL, Rockville,
MD), as described by the manufacturers. Luciferase
activity was measured 24 h after the addition of com-
pounds. Petersen, D. N.; Tkalcevic, G. T.; Koza-Taylor,
P. H.; Turi, T. G.; Brown, T. A. Endocrinology 1998,
139, 1082.
27. 2,3-Diphenyl Indole CAS # 3469-20-3. MP BioMedicals
15 Morgan, Irvine, CA 92618-2005.
28. Anstead, G. M.; Carlson, K. E.; Katzenellenbogen, J. A.
Steroids 1997, 62, 268.