Benzothiophene and naphthalene derived constrained SERMs

Article abstract of DOI:10.1016/j.bmcl.2004.07.072

The synthesis and biological evaluation of two series of conformationally restricted SERMs are disclosed. In each series (benzothiophene or naphthalene), the ligand side chain is constrained to adopt a defined orientation. The orientation of the side chain has a significant impact on functional activity.

Full text of DOI:10.1016/j.bmcl.2004.07.072

Bioorganic & Medicinal Chemistry Letters 14 (2004) 5103–5106
Benzothiophene and naphthalene derived constrained SERMs
Owen B. Wallace,* Henry U. Bryant, Pamela K. Shetler, Mary D. Adrian and
Andrew G. Geiser
Lilly Research Laboratories, Eli Lilly and Company, Lilly Corporate Center, Indianapolis, IN 46285, USA
Received 6 July 2004; accepted 28 July 2004
Available online 20 August 2004
Abstract—For selective estrogen receptor modulators (SERMs), the orientation of the basic side chain relative to the SERM core
has a significant impact on function. The synthesis and biological evaluation of two series of SERMs are disclosed, where the ligand
side chain is constrained to adopt a defined orientation. Compounds where the side chain is forced into the plane of the SERM core
have a different profile compared to those compounds where the side chain is pseudo-orthogonal, particularly with regard to antag-
onism of estradiol action on an Ishikawa uterine cell line.
ꢀ 2004 Elsevier Ltd. All rights reserved.
The estrogen receptor (ER) is a ligand activated tran-
scription factor, which plays a critical role in female
reproductive function, and influences skeletal, cardio-
vascular and CNS health. There are two known sub-
types, ERa and ERb, which are expressionally and
functionally unique. Selective estrogen receptor modula-
tors (SERMs) are compounds that may mimic the effects
of estrogen in some tissues, while antagonizing the ef-
fects of estrogen in others.1 While the precise reason
for this tissue selectivity is not fully understood, it is
clear that the conformation of the receptor and its abil-
ity to interact with cofactors is critically important.
Moreover, subtle changes of ligand structure can have
a dramatic impact on receptor conformation.
stricted
benzanthracene
derivatives
is
herein
reported––one series based on a benzothiophene scaffold
and the other based on a naphthalene core.
Benzothiophene analogs were prepared from N,N-
dimethylamino derivative 1 (Scheme 1), which has previ-
ously been shown to be a useful electrophile for the
synthesis of raloxifene analogs.6 Addition of a benzyl
Grignard reagent to 1 resulted in benzothiophene 2.
Reduction with LAH in THF at room temperature
afforded alcohol 3. Treatment with TFA yielded the
cyclized product, resulting from trapping of the resulting
carbonium ion with the pendant aryl ring at the C2 posi-
tion. Deprotection of the phenol (BBr3, CH2Cl2) then
yielded the desired products. A more expeditious route
to the cyclized product was developed wherein alcohol
3 was simply treated with BBr3 to affect cyclization with
concomitant deprotection in a single pot.7 Interestingly,
when the cyclization reaction was attempted with the
para-methoxy derivative (3, R = 4-OMe), the fully aro-
matic analog 7 was isolated as the minor product com-
ponent together with 6 in a 1:8 ratio.8 The two
products could be readily separated by reverse phase
HPLC before testing in the in vitro assays. The meta-
methoxy alcohol intermediate (3, R = 3-OMe) on treat-
ment with BBr3 afforded a ꢀ2.9:1 mixture of the 3- and
5-phenols (5 and 8), which again could be readily sepa-
rated by reverse phase HPLC.
Of significant importance for a ligandÕs ability to act as a
SERM is the nature of the basic side chain and its orien-
tation relative to the ligand backbone. It is known that
in the X-ray crystal structure of raloxifene bound to
ERa, the basic side chain of the ligand occupies an
orthogonal position relative to the benzothiophene
core.2 In an effort to constrain the side chain into its pre-
sumed active conformation, Grese et al. prepared a ser-
ies of benzopyran derivatives, some of which displayed
an excellent in vitro and in vivo profile.3 Other recent
examples of constrained SERMs come from the labs
of Katzenellenbogen.4,5 The design, synthesis, and bio-
logical activity of two series of conformationally re-
Naphthalene analogs were prepared from naphthalene
triflate 9.9 Using KnochelÕs protocol, triflate 9 was cou-
pled with a benzyl zinc reagent in the presence of
*
Corresponding author. Fax: +1 317 433 0715; e-mail: owen.wallace@
lilly.com
0960-894X/$ - see front matter ꢀ 2004 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2004.07.072

5104
O. B. Wallace et al. / Bioorg. Med. Chem. Lett. 14 (2004) 5103–5106
O
O
O
O
R
N
N
MgX
THF, r.t.
LAH
THF
MeO
S
N
MeO
S
R
1
2
N
N
O
O
O
R
HO
N
BBr3
3
+
OH
CH2Cl2
MeO
S
R
5
HO
S
HO
S
3
4 R = H
7
5 R = 3-OH
6 R = 4-OH
8 R = 5-OH
Scheme 1. Synthesis of benzothiophene constrained SERMs.
Pd(dba)2, dppf, and Bu4NI in a mixture of NMP and
THF to afford naphthalene 10.10 Reduction with LAH
in THF at room temperature afforded alcohol 11, which
was then subjected to BBr3 in CH2Cl2 at room temper-
ature to yield the desired dihydrobenzanthracenes. Once
again, the para-methoxyphenyl derivative 11 (R = 4-
OMe) gave a ꢀ2:1 mixture of the desired product 13
along with the aromatized benzo[a]anthracene deriva-
tive 14, which were separable by HPLC. The meta-meth-
oxy analog 11 (R = 3-OMe) afforded the 3- and
5-phenols 12 and 15, respectively, in a ꢀ1:3 ratio, which
were also separable (Scheme 2).
Antagonist activity (IC50 and percent efficacy) was
determined by inhibition of alkaline phosphatase induc-
tion by the compound in the presence of 1nM 17b-estra-
diol.12 4-Hydroxytamoxifen (4-OHT) was used as a
standard in each assay.13 Compounds with a desirable
SERM profile are those with potent binding to the estro-
gen receptors and are potent and efficacious antagonists
of MCF-7 cell proliferation. In addition, in order to
minimize undesirable uterine effects, compounds should
display weak Ishikawa agonism and potent and effica-
cious Ishikawa antagonism.
In the benzothiophene series, the parent compound 4
(R = H) has reasonable binding affinity to both ERa
and ERb (Ki = 2.4 and 5.3nM, respectively), and antag-
onizes the effect of estrogen in the MCF-7 cell line (IC50
17nM). However, the compound displays a mixed agon-
ist/antagonist profile in the Ishikawa cell line, and is a
considerably more potent agonist than antagonist
(IC50 3.5 vs 710nM). Introduction of a hydroxyl at C3
(5) neither improves binding affinity nor functional
antagonism. In fact, the compound is a weaker antago-
nist in MCF-7 cells (IC50 42 nM) and is a more effica-
Compounds were first tested in a radioligand binding as-
say11 and were then examined for functional activity in
breast cancer (MCF-7) and endometrial adenocarci-
noma (Ishikawa) cell lines (Table 1). Briefly, antagonism
of estrogen action in MCF-7 cells was assayed via inhi-
bition of cell proliferation by 10pM of 17b-estradiol,
and is reported as an IC50 value and percent efficacy.
Agonist activity (EC50 and percent efficacy) of the com-
pounds in Ishikawa cells was determined by quantitating
alkaline phosphatase induction by the compound alone.
O
N
O
N
O
O
MeO
ZnCl
LAH
OTf
THF
OMe
N
Pd(dba)2, dppf, Bu4NI
NMP/THF (3:2)
MeO
MeO
MeO
9
10
OH
4
O
O
O
3
N
N
OH
HO
BBr3
+
CH2Cl2
OMe
HO
HO
12 R = 3-OH
13 R = 4-OH
15 R = 5-OH
16 R = 6-OH
11
14
Scheme 2. Synthesis of naphthalene constrained SERMs.

O. B. Wallace et al. / Bioorg. Med. Chem. Lett. 14 (2004) 5103–5106
5105
Table 1. Binding and functional data
N
N
O
O
OH
4
O
O
3
N
N
OH
3
OH
R
5
HO
S
HO
S
HO
12 R = 3-OH
HO
13 R = 4-OH
4 R = H
5 R = 3-OH
7
14
15 R = 5-OH
15a R = 5-OH (ent 1)
6 R = 4-OH 8 R = 5-OH
15b R = 5-OH (ent 2) 16 R = 6-OH
Binding
MCF-7
Ishikawa
Ishikawa
Ki ERa (nM)
Ki ERb (nM)
IC50 (nM)a
Efficacy (%)a
EC50 (nM)b
Efficacy (%)b
IC50 (nM)c
Efficacy (%)c
4-OHT
4
0.5
2.4
2.6
0.5
0.4
0.7
7.7
1.0
0.8
0.8
1.1
0.7
1.1
0.5
5.3
14
2.6
17
79
72
66
80
77
74
72
79
69
79
65
70
85
0.8
3.5
3.1
0.9
1.6
0.4
28
140
77
510
710
680
53
45
66
67
84
63
70
75
82
59
91
71
92
84
5
42
130
66
6
0.9
0.3
1.3
40
0.9
1.0
0.7
68
7
170
100
77
690
45
8
12
13
14
15
15a
15b
16
580
130
820
31
3.2
0.5
4.2
5.5
4.4
5.4
6.6
57
2.6
10
75
130
27
1.9
23
0.8
3.1
0.5
0.7
56
19
160
28
2.4
2.9
120
50
a IC50 and % efficacy data are quoted for MCF-7 assay in antagonist mode.
b EC50 and % efficacy data are for Ishikawa assay in agonist mode.
c IC50 and % efficacy are for Ishikawa assay in antagonist mode.
cious agonist in Ishikawa uterine cells compared to the
unsubstituted parent (4). A significant increase in bind-
ing affinity is seen with the 4-hydroxyl derivative (6),
and this analog also displays a much improved func-
tional profile. The binding affinity of 6 is approximately
5-fold greater than 4 and the compound now displays
sub-nM inhibition of estradiol action on MCF-7 cells
with good efficacy (80%). The compound is also ꢀ10-
fold more potent in the Ishikawa antagonist assay, but
still displays relatively potent agonism in this cell line
(EC50 0.9nM). The importance of the relative orienta-
tion of the side chain is clearly seen by comparing 6 to
it aromatic analog 7 where the side chain is unable to
adopt an orthogonal relationship to the benzothiophene
core. Although 7 binds with similar high affinity to both
ERa and ERb and is a potent MCF-7 antagonist (IC50
1nM), the functional profile in uterine adenocarcinoma
cells is markedly different. Benzothiophene 7 is a weaker
Ishikawa antagonist (IC50 690nM) than analog 6 (IC50
53nM), and is also a more efficacious agonist (170% effi-
cacy). The 5-hydroxyl analog 8 displays a very similar
profile to the 4-hydroxyl analog 6 in terms of binding,
MCF7 antagonism and Ishikawa antagonism, but is a
more potent and efficacious agonist in the uterine cell
line (EC50 0.4nM, 100% efficacy).
icantly more potent antagonist in both breast (IC50
6.6nM) and uterine (IC50 130nM) cell lines. The fully
aromatic benzanthracene derivative 14 displays potent
binding but weaker functional antagonism relative to
13, again demonstrating the importance of the side
chain orientation. Benzanthracene 14 is approximately
8-fold less potent in the MCF-7 assay, and 6-fold less
potent in the Ishikawa antagonist assay compared to
13. Interestingly, introduction of aromaticity into the
naphthalene core has a larger detrimental effect on
MCF-7 antagonism relative to the benzothiophene core,
where 7 maintained potent inhibition. Introduction of
the hydroxyl at the 5-position improves the functional
profile further. Naphthalene 15 is now a potent and effi-
cacious MCF-7 antagonist (IC50 1.9nM, 79% efficacy)
and Ishikawa antagonist (IC50 31nM, 91% efficacy).
More significantly, its efficacy as an agonist in Ishikawa
cells (27% efficacy) is considerably less than any of the
other analogs tested, whether from the benzothiophene
or naphthalene scaffold. The favorable profile of 15 then
prompted the separation and testing of the enantiomers.
The distomer, 15a, has a similar binding profile but a
weaker antagonist profile compared to the racemate.
The compound is also a more efficacious agonist in the
Ishikawa assay (56% efficacy). The eutomer, 15b, has
an almost identical binding and functional profile as
the racemate (15).
In the naphthalene-based series, the 3-hydroxyl deriva-
tive (12) again displays relatively weak binding, MCF-
7 and Ishikawa antagonism, and its agonist potency is
somewhat weaker than its benzothiophene counterpart
5. The 4-hydroxyl analog (13) has an improved binding
and functional profile relative to analog 12. It is a signif-
Compound 15b was tested for its ability to block the ef-
fects of estradiol (E2) on the uterus in a three day imma-
ture rat assay.14 The compound was administered PO at
0.01, 0.1, and 1mg/kg and showed little inhibition even

5106
O. B. Wallace et al. / Bioorg. Med. Chem. Lett. 14 (2004) 5103–5106
4. Meyers, M. J.; Sun, J.; Carlson, K. E.; Katzenellenbogen,
60
50
40
30
20
10
0
15b
B. S.; Katzenellenbogen, J. A. J. Med. Chem. 1999, 42,
2456.
Tamoxifen
5. Tedesco, R.; Youngman, M. K.; Wilson, S. R.; Katzenel-
lenbogen, J. A. Bioorg. Med. Chem. Lett. 2001, 11, 1281.
6. Grese, T. A.; Cho, S.; Finley, D. R.; Godfrey, A. G.;
Jones, C. D.; Lugar, C. W., III; Martin, M. J.; Matsu-
moto, K.; Pennington, L. D.; Winter, M. A.; 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.
7. On occasion during quenching, bromination ortho to
phenol was observed as a side reaction. This could be
eliminated by quenching the BBr3 reaction with MeOH
and 2-methy-2-butene as a bromine scavenger, followed
by immediate purification on an SCX column.
0.01
0.1
1
Compound Dose (mg/kg)
8. The antiestrogenic effects of benzanthracenes in vivo has
been described. See: Morreal, C. E.; Sinha, D. K.;
Schneider, S. L.; Bronstein, R. E.; Dawidzik, J. J. Med.
Chem. 1982, 25, 323; The interaction of benzanthracene
diols with the estrogen receptor has also been described:
Anstead, G. M.; Kym, P. R.; Albert, B. Steroids 1995, 60,
383.
9. Wallace O. B. PCT Intl. Appl. WO 2004029047, 2004.
10. Piber, M.; Jensen, A. E.; Rottlander, M.; Knochel, P. Org.
Lett. 1999, 1, 1323.
Figure 1. Inhibition of estradiol-induced uterine stimulation in a three
day immature rat assay.
at the highest dose (Fig. 1).15 In fact, compound 15b is
much less efficacious at inhibiting estradiol-induced
stimulation of the uterus than the known partial agonist,
tamoxifen. The reason for the lack of in vivo efficacy of
15b was not investigated, but may be due to poor inher-
ent ADME properties of the compound. For example, a
possible explanation is that 15b may undergo oxidation
in vivo to benzanthracene 14, which is approximately
30-fold less potent as an antagonist in the Ishikawa uter-
ine cell line.
11. Experimental protocols for the ER binding and MCF-7
assays have been described previously. See: Wallace, O. B.;
Lauwers, K. S.; Jones, S. A.; Dodge, J. A. Bioorg. Med.
Chem. Lett. 2003, 13, 1907.
12. Ishikawa human endometrial tumor cells were assayed in
DMEM/F-12 (3:1) (DulbeccoÕs Modified Eagle Medium:
Nutrient Mixture F-12, phenol red-free, Gibco BRL)
supplemented with 5% dextran coated charcoal stripped
fetal bovine serum (DCC-FBS) (Hyclone, Logen, UT).
The assay was run in both agonist (compound alone) and
antagonist (compound plus 1nM E2 (b-estradiol, Sigma,
St. Louis, MO)) modes. Cells were incubated with
compound for 48h, then fresh assay medium plus com-
pound were added and incubated for an additional 72h.
The assay was quenched by removing media and rinsing
plates twice in PBS. Assay plates were dried for 5min and
frozen at À70ꢁC for at least 1h. Plates were removed,
allowed to warm to room temperature, and then 100lL of
1-Stepꢂ PNPP (Pierce Chemical Company, Rockford, IL)
was added. After a 20-min incubation, plates were read on
a spectophotometer at 405nm to quantitate alkaline
phosphatase. For the agonist mode, a percentage increase
over control for each compound was calculated. The data
were fitted to a linear interpolation to derive IC50 values
for the antagonist mode and a percentage efficacy was
calculated that blocks the E2 (1nM) stimulus. For the
agonist mode, a % efficacy for each compound was
calculated compared to control and an EC50 calculated.
13. Unless otherwise stated, compounds were tested as
racemates.
In summary, we have prepared a new series of con-
strained benzothiophene- and naphthalene-based
SERMs. In both series, the spatial relationship between
the side chain and ligand core has a significant impact
on functional activity. Compounds where the side chain
is forced into the same plane as the backbone display
significantly weaker antagonist effects (most notably in
the Ishikawa cell line) compared to the related structures
where the side chain is pseudo-orthogonal to the core.
This is presumably due to induction of differential recep-
tor conformations by the ligand, which in turn leads to
differential recruitment of coactivators and corepressors.
Acknowledgements
We wish to thank the Lead Optimization Biology group
for conducting binding, MCF-7 and Ishikawa assays.
References and notes
14. Three week old Sprague Dawley female rats were orally
treated with E2 (0.1mg/kg) and 1.0, 0.1, and 0.01mg/kg
SERM for 3days, six rats per group. Test compounds were
dissolved in 20% b-hydroxycyclodextrin and administered
PO 15min prior to E2. Animals were weighed, then
euthanized (by carbon dioxide asphyxiation), and the uteri
were collected and weighed. The percentage inhibition of
the estrogen-induced response was calculated as: 100 ·
1. For a recent review of SERMs, see: Wallace, O. B.;
Richardson, T. I.; Dodge, J. A. Curr. Top. Med. Chem.
2003, 3, 1663.
2. 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.
3. Grese, T. A.; Pennington, L. D.; Sluka, J. P.; Adrian, M.
D.; Cole, H. W.; Fuson, T. R.; Magee, D. E.; Phillips, D.
L.; Rowley, E. R.; Shetler, P. K.; Short, L. L.; Venugo-
palan, M.; Yang, N. N.; Sato, M.; Glasebrook, A. L.;
Bryant, H. U. J. Med. Chem. 1998, 41, 1272.
(UWRestrogenÀUWRtest compound)/(UWRestrogenÀUWRcontrol
)
where UWR = uterine weight/body weight ratio.
15. In Figure 1, error bars for compound 15b are intra-assay
SE (n = 6 animals per group). Error bars for Tamoxifen
are inter-assay SE (three assays, n = 6 animals per group).