Lead identification of a potent benzopyranone selective estrogen receptor modulator

Article abstract of DOI:10.1016/j.bmcl.2004.04.081

Starting from a phenol screening hit (6), three series of benzopyranone selective estrogen receptor modulators (SERMs) have been designed, synthesized, and analyzed for both estrogen receptor alpha binding affinity and in vitro activity in two cell assays. The lead compound identified, SP500263 (13), was more potent than raloxifene and tamoxifen in a cell-based assay measuring inhibition of interleukin-6 release.

Full text of DOI:10.1016/j.bmcl.2004.04.081

Bioorganic & Medicinal Chemistry Letters 14 (2004) 3407–3410
Lead identification of a potent benzopyranone
selective estrogen receptor modulator
Jeffrey A. McKie,a,* Shripad S. Bhagwat,a Helen Brady,c Mary Doubleday,a
Leah Gayo,a Mathew Hickman,c Ravi K. Jalluri,a Sak Khammungkhune,a
Adam Kois,a Deborah Mortensen,a Normand Richard,c
John Sapienza,a Graziella Shevlin,a Bernd Steinb and May Sutherlandb
aMedicinal Chemistry, Celgene, San Diego, CA 92121, USA
bCellular and Molecular Pharmacology, Celgene, San Diego, CA 92121, USA
cLead Discovery, Celgene, San Diego, CA 92121, USA
Received 12 March 2002; accepted 27 April 2004
Available online
Abstract—Starting from a phenol screening hit (6), three series of benzopyranone selective estrogen receptor modulators (SERMs)
have been designed, synthesized, and analyzed for both estrogen receptor alpha binding affinity and in vitro activity in two cell
assays. The lead compound identified, SP500263 (13), was more potent than raloxifene and tamoxifen in a cell-based assay mea-
suring inhibition of interleukin-6 release.
Ó 2004 Elsevier Ltd. All rights reserved.
The hormone 17b-estradiol (E2) (1, Fig. 1) plays an
important role in the physiology of both women and
men and the effect of decreased plasma hormone con-
centrations in peri- and postmenopausal women has
drawn significant attention.1 To compensate for the re-
duced plasma E2 levels throughout menopause, estrogen
replacement therapy (e.g., estrone, E2, estriol, and/or
conjugated equine estrogens) or estrogen plus a pro-
gestin (also referred to as hormone replacement therapy)
is typically prescribed. These protocols have proven
effective in treating osteoporosis and hot flushes, while
any beneficial role in cardiovascular and central nervous
system health (Alzheimer’s and cognition) is still under
debate.2 Estrogen replacement therapy results in side
effects that include an increase in risk of uterine and
breast cancer that is only partially overcome by the
addition of a progestin to the treatment combination.3
The increased risk of uterine and breast cancer associ-
ated with estrogen replacement therapy is believed to be
due to the mitogenic action of E2 and E2 conjugates on
these reproductive tissues.4 Thus, estrogen and hormone
Keywords: Benzopyranone; Estrogen receptor modulator.
* Corresponding author. Tel.: +1-858-795-4781; fax: +1-858-795-4719;
e-mail: jmckie@celgene.com
Figure 1. Structures of 17b-estradiol and marketed SERMs with cor-
responding A, B, C, and D rings of estradiol shown.
0960-894X/$ - see front matter Ó 2004 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2004.04.081

3408
J. A. McKie et al. / Bioorg. Med. Chem. Lett. 14 (2004) 3407–3410
replacement therapies are viewed as having desirable as
well as undesirable effects as therapeutic regimens. From
this understanding, subsequent drug discovery programs
have been initiated to develop potential nonsteroidal
estrogen receptor ligands that retain the beneficial
pharmacological effects of E2 and eliminate its unde-
sirable properties.
Although 6 displayed desirable estrogen receptor ago-
nist activity in our IL-6 assay by blocking cytokine re-
lease, it also had undesirable agonist properties in the
MCF-7 proliferation assay that manifest in an increase
in cellular proliferation. A review of journal and patent
literature indicated that other benzopyranone molecules
had been reported to have ER-modulating properties,
however, the disclosed compounds did not incorporate
the functional groups that we believed necessary for
ideal SERM activity.11 Published extensive structure–
activity studies on SERMs provided key information on
structural requirements necessary for selective antago-
nist properties. Incorporation of the side-chain amine in
SERMs such as tamoxifen and raloxifene has been
shown to be critical for providing selective tissue
antagonist properties. Furthermore, crystal structure
data of bound agonists and antagonists has facilitated
the design of a useful model for correlating compound
structure with agonist or antagonist receptor confor-
mation.12
Early drug discovery efforts focused on the design of
small molecule nonsteroidal estrogen receptor ligands
with antagonist properties against breast and other
reproductive tissues.5 Tamoxifen (2, Fig. 1) was the first
marketed drug to be realized from these efforts and
while this compound and its active metabolite, 4-hy-
droxytamoxifen (3, Fig. 1), are effective antiestrogens on
estrogen receptor positive breast tissue, they subse-
quently were discovered to have undesirable estrogenic
properties on the endometrium.6 A second triphenyl-
ethylene compound, toremifene (4, Fig. 1), has also been
approved for the treatment of breast cancer although it
too was reported to have undesirable uterine stimulatory
activity.7 More recently, raloxifene (5, Fig. 1), a ben-
zothiophene was marketed for the prevention and
treatment of osteoporosis. This compound has been
termed a selective estrogen receptor modulator (SERM)
because of its estrogen agonist properties on bone and
estrogen antagonist properties on the breast and endo-
metrium. Unlike tamoxifen, raloxifene has not been
associated with an increased risk of uterine cancer in
women upon prolonged use.8
With this information in hand, a set of benzopyranone
analogues (8–19) was designed based on the structures
of 4-hydroxytamoxifen (3) and raloxifene (5). In the
series, we proposed to initially fix the D-ring substituent
(3-position of the benzopyranone core) as a phenyl
group and attach a variable length piperidinylalkyl
chain (SERM side chain) from the benzopyranone 4-
phenoxy (8–12), 3-phenoxy (16–19) or 4-benzyloxy (13–
15) substituent (Fig. 3).
As part of our drug discovery effort toward identifying
potential SERMs for the prevention and treatment of
osteoporosis and breast cancer, we developed a unique
cellular screen to measure the effect of small molecule
estrogen receptor ligands at inhibiting release of the
cytokine, interleukin-6 (IL-6), from an estrogen receptor
negative osteosarcoma cell line (U2OS) transfected with
ER-a.9 The estrogen receptor-mediated effect on IL-6
release under our conditions was determined to proceed
by a nonclassical mechanism. Since IL-6 is a cytokine
that promotes bone resorption, proliferation of certain
cancers, and other pathogenic processes,10 inhibiting IL-
6 release is predicted to be a measure of a potential drug
candidate’s ability to reduce both bone resorption and
cellular proliferation.
The synthetic route that provided the desired prod-
ucts most efficiently is shown in Scheme 1. Commer-
cially available 3-methoxyphenol was acylated under
Fries reaction conditions using POCl3 and ZnCl2 in
We initiated screening of commercially available phen-
olic compounds in our primary assay to measure their
ability at inhibiting IL-6 release in U2OS cells. Several
analogues provided IC50 values of <10 lM and the
benzopyranone analogue, 6, proved to be of particular
interest with an IC50 of 300 nM (Fig. 2).
Figure 3. Design of analogues of screening hit 6 to incorporate
tamoxifen and raloxifene structural features.
Figure 2. High-throughput screen of commercial phenols to identify screening hit 6.

J. A. McKie et al. / Bioorg. Med. Chem. Lett. 14 (2004) 3407–3410
3409
Scheme 1. Reagents and conditions: (a) POCl3, ZnCl2, 1,2-dichloroethane 70 °C, 4 h; or BF3ÆOEt2, chlorobenzene, 70 °C, 2 h, yield ¼ 40–60%; (b)
benzoic acid or phenylacetic acid, carbonyl diimidazole, K2CO3, DMAP, DMF, 85 °C, 4 h, yield ¼ 60–70%; (c) 1,x-dibromoalkane, K2CO3, acetone,
reflux, 12 h, yield ¼ 80–90%; (d) piperidine, THF, 60 °C, 5 h, yield ¼ 85%; (e) 2-chloroethylpiperidine hydrochloride, K2CO3, DMF, 5 h, 80 °C,
yield ¼ 75%; (f) 30% HBr in acetic acid, 90 °C, 8 h, yield ¼ 60% or ethanethiol, AlBr3, CH2Cl2, room temperature, 2 h, yield ¼ 75%.
Table 1. In vitro MCF-7 and IL-6 assay data for compounds 8–15,
raloxifene (5), and 4-hydroxytamoxifen (3)
dichloroethane or BF3ÆOEt2 in chlorobenzene at 70 °C.
The major product was the desired benzophenone or
benzoacetophenone indicated, although the unwanted
Fries-reaction regioisomer (para to the phenol group of
3-methoxyphenol) was also detected in the reaction
mixture. The second reaction step was a benzopyranone
core formation reaction that was accomplished by
modification of a procedure reported by Rao and
Srimannarayana.13 Introduction of the side-chain amine
functional group was successfully completed by either a
one-step ðn ¼ 2Þ or two-step process (n ¼ 3, 4, 5, 6). The
one-step process involved alkylation of the phenol, 7,
with chloroethylpiperidine hydrochloride by heating a
mixture of these two reagents and K2CO3 in DMF. The
two-step procedure progressed by alkylation of phenol,
7, with 1,x-dibromoalkane and subsequent displace-
ment of the terminal bromide with piperidine. The final
products (8–15) were prepared through demethylation
of the resulting A-ring methylethers by heating in 30%
HBr–acetic acid. Alternatively, the demethylation could
be accomplished at room temperature with ethanethiol
and AlBr3 in CH2Cl2.
Compd
n
m
U2OS (ER-a) IL-6,a MCF-7a
RBA,a;c
IC50 (nM)b ER-a
IC50 (nM)b
3
na na
na na
1.9
3.1
24
4.9
48
35
4
5
8
1
2
3
4
5
1
2
3
0
0
0
0
0
1
1
1
505
355
69
425
335
134
335
720
9
21
29
79
88
85
90
139
10
11
12
13
14
15
114
145
0.38
4.4
7.7
18
13
5.0
a See Ref. 9 for a description of the assay.
b Values are an average of at least two experiments with standard
errors below 15%.
As indicated in Table 1, the tamoxifen mimetics (8–12)
proved to be potent estrogen receptor ligands, but sig-
nificantly less active than desired in our in vitro IL-6 and
MCF-7 assays. In comparing this series to tamoxifen
and raloxifene, the most potent benzopyranone (10) was
approximately 5–30-fold less active in our cellular assays
than the SERM standards. While extending the side-
chain amine improved in vitro activity up to eightfold
for n ¼ 3 (10) compared to n ¼ 2 (9), additional chain
extension resulted in decreased activity (11–12). The
ER-a relative binding affinity (relative to E2) for the
series ranged from 488%. The compound from this series
that displayed the weakest ER receptor affinity, 8, was
also the least potent in the cellular assays, however,
there was not a direct correlation between receptor
binding affinity and cellular functional activity. Al-
though the cellular data was promising, we were confi-
dent that optimal positioning of the side-chain
substituent within the receptor binding pocket would
provide us with more potent analogues.
c Relative binding affinity (RBA, where RBA of E2 ¼ 100).
potency in our IL-6 assay as well as an improvement
in antiproliferative activity in the MCF-7 assay. The
receptor binding affinity for the series once again pro-
vided evidence that receptor binding affinity and cellular
functional activity do not correlate directly. Previous
reports have also supported the hypothesis that the
resulting conformation of the estrogen receptor–ligand
complex can be more important than the affinity of the
ligand for the receptor in determining ligand agonist or
antagonist potency.14 The additional atom spacer pro-
vided by the benzylic carbon atom resulted in optimum
positioning of the molecules within the receptor.15
Compound 13 was significantly more potent in the IL-6
assay than either raloxifene or tamoxifen, while the
MCF-7 activity was similar to raloxifene. Extension of
the side chain (14–15) resulted in a decrease in activity.
Despite this significant loss of activity, the 4-benzyl
series of benzopyranones (13–15) was surprisingly more
active than the 4-phenyl series of analogues (8–12).
The raloxifene mimetics (13–15) provided an approxi-
mately two-order of magnitude improvement in in vitro

3410
J. A. McKie et al. / Bioorg. Med. Chem. Lett. 14 (2004) 3407–3410
corresponding improvement in functional activity in our
assays was consistent with results from the other two
series (Table 2).
In summary, we have designed and synthesized three
series of benzopyranone selective estrogen receptor
modulators based on a phenol screening hit. Analogue
synthesis resulted in the development of a clear SAR for
each series and the potential for improvement in tissue
selective activity and target potency. Future research
will be focused on optimizing the in vivo activity of our
current lead structure SP500263 (13).
References and notes
Figure 4. Modification of the benzopyranone 4-phenyl group to
incorporate the critical hinge feature identified from the benzypyra-
none 4-benzyl series.
1. Simpson, E. R.; Davis, S. R. J. Clin. Endocrinol. Metab.
1998, 83, 2214.
2. (a) WHI Steering Committee JAMA 2004, 291, 1701; (b)
Herrington, D. M. Ann. Int. Med. 1999, 131, 463.
3. Stanford, J. L.; Weiss, N. S.; Voigt, L. F.; Daling, J. R.;
Habel, L. A.; Rossing, M. A. JAMA 1995, 274, 137.
4. De Lignieres, B. Maturitas 1996, 23, 31.
5. Lerner, L. J.; Holthaus, J. F.; Thompson, C. R. Endocri-
nology 1958, 63, 295.
6. Killackey, M. A.; Hakes, T. B.; Pierce, V. K. Cancer Treat.
Rep. 1985, 69, 237.
In overlaying the structures of the two series of com-
pounds, the 4-phenyl and 4-benzyl analogues, we con-
cluded that positioning of the amine side chain was
critical for selective estrogen receptor activity. The hinge
created by the methylene spacer in the benzyl series
clearly provided ideal positioning of the basic amine side
chain. Based on this observation we elected to incor-
porate this effective feature into the 4-phenyl series by
moving the basic amine side chain to the meta position
of the benzopyranone 4-phenyl ring as exemplified in
compounds 16–19 in Figure 4.
7. Maenpaa, J.; Holli, K.; Pasanen, T. Eur. J. Cancer 2000,
36, S61.
8. Cummings, S. R.; Eckert, S.; Krueger, K. A.; Grady, D.;
Powles, T. J.; Cauley, J. A.; Norton, L.; Nickelsen, T.;
Bjarnason, N. H.; Morrow, M.; Lippman, M. E.; Black,
D.; Glusman, J. E.; Costa, A.; Jordan, V. C. JAMA 1999,
316, 2189.
9. (a) Stein, B.; Yang, M. X. Mol. Cell. Biol. 1995, 10, 4971;
(b) Brady, H.; Desai, S.; Gayo-Fung, L.; Khammungkh-
une, S.; McKie, J. A.; O’Leary, E.; Pascasio, L.; Kung-
Sutherland, M.; Anderson, D. W.; Bhagwat, S. S.; Stein,
B. Cancer Res. 2002, 62, 1439.
10. (a) Stein, B.; Kung-Sutherland, M. S. Drug Discovery
Today 1998, 3, 202; (b) Simpson, R. J.; Hammacher, A.;
Smith, D. K.; Matthews, J. M.; Ward, L. D. Protein Sci.
1997, 6, 929; (c) Yudkin, J. S.; Kumari, M.; Humphries, S.
E.; Mahamed-Ali, V. Atherosclerosis 2000, 148, 209; (d)
Ershler, W. B.; Keller, E. T. Annu. Rev. Med. 2000, 51,
245.
11. (a) Lednicer, D.; Lyster, S. C.; Duncan, D. W. J. Med.
Chem. 1965, 8, 725; (b) Wani, M. C.; Rector, D. H.;
Christensen, H. D.; Kimmel, G. L.; Cook, C. E. J. Med.
Chem. 1978, 18, 982; (c) Pollard, I.; Martin, L. Steroids
1968, 11, 897.
Moving the basic amine side chain from the phenyl
group 4-position to the 3-position in the 4-phenyl ben-
zopyranone series provided a significant improvement in
U2OS/IL-6 as well as MCF-7 antiproliferative activity.
In directly comparing the 2 and 3 atom spacers of the
amine side-chain spacer from these two series, com-
pounds 16 versus 8 and 17 versus 9, a 40- and >200-fold
improvement in U2OS/IL-6 activity was realized. These
results provided evidence that this manipulation of the
side chain was effective in mimicking the hinge sub-
stituent found in the benzopyranone 4-benzyl series. The
observation that extending the side-chain length im-
proved receptor binding affinity but failed to provide a
12. (a) Brzozowski, A. M.; Pike, A. C. W.; Dauter, Z.;
Hubbard, R. E.; Bonn, T.; Engstrom, O.; Ohman, L.;
Greene, G. L.; Gustafsson, J. A.; Carlquist, M. Nature
1997, 389, 753; (b) Grese, T. A.; Sluka, J. P.; Bryant, H.
U.; Cullinan, G. J.; Glasebrook, A. L.; Jones, C. D.;
Matsumoto, K.; Palkowitz, A. D.; Sato, M.; Teermine, J.
D.; Winter, M. A.; Yang, N. N.; Dodge, J. A. Proc. Natl.
Acad. Sci. 1997, 94, 14105.
13. Rao, P. P.; Srimannarayana, G. Synthesis 1981,
887.
14. Wijayaratne, A. L.; Nagel, S. C.; Paige, L. A.; Christensen,
D. J.; Norris, J. D.; Fowlkes, D. M.; McDonnell, D. P.
Endocrinology 1999, 140, 5828.
Table 2. In vitro MCF-7 and IL-6 assay data for compounds 13, and
16–18 incorporating the hinge feature into the 4-phenyl series
Compd
n
m
U2OS (ER-a) IL-6,a MCF-7a
RBA,a;c
IC50 (nM)b
IC50 (nM)b ER-a
13
16
17
18
19
1
1
2
3
4
1
0
0
0
0
0.38
12
7.7
85
9.5
12
15
85
20
1.5
3.2
11
45
52
105
a See Ref. 9 for a description of the assay.
b Values are an average of at least two experiments with standard
errors below 15%.
15. Weatherman, R. V.; Carroll, D. C.; Scanlan, T. S. Bioorg.
Med. Chem. Lett. 2001, 11, 3129.
c Relative binding affinity (RBA, where RBA of E2 ¼ 100).