Synthesis and structure-activity relationship of 3-arylbenzoxazines as selective estrogen receptor β agonists

Article abstract of DOI:10.1016/j.bmcl.2004.01.099

A series of 3-aryl-7-hydroxybenzoxazine analogues have been prepared and evaluated as ligands for the two estrogen receptor subtypes (ERα and ERβ). From the radioligand binding assay, compounds with more than a 10-fold binding selectivity toward the ERβ s

Full text of DOI:10.1016/j.bmcl.2004.01.099

Bioorganic & Medicinal Chemistry Letters 14 (2004) 2327–2330
Synthesis and structure–activity relationship of 3-arylbenzoxazines
as selective estrogen receptor b agonists
Wu Yang,a,* Yufeng Wang,a Zhengping Ma,b Rajasree Golla,b Terry Stouch,c
b
b
b
a
€
€
Ramakrishna Seethala, Susan Johnson, Rong Zhou, Timur Gungor,
Jean H. M. Feyenb and John K. Dickson, Jr.a
aDiscovery Chemistry, Bristol-Myers Squibb Pharmaceutical Research Institute, PO Box 5400, Princeton, NJ 08543-5400, USA
bMetabolic and Cardiovascular Drug Discovery, Bristol-Myers Squibb Pharmaceutical Research Institute, PO Box 5400,
Princeton, NJ 08543-5400, USA
cDepartment of Macromolecular Structure, Bristol-Myers Squibb Pharmaceutical Research Institute, PO Box 5400, Princeton,
NJ 08543-5400, USA
Received 2 December 2003; revised 29 January 2004; accepted 29 January 2004
Abstract—A series of 3-aryl-7-hydroxybenzoxazine analogues have been prepared and evaluated as ligands for the two estrogen
receptor subtypes (ERa and ERb). From the radioligand binding assay, compounds with more than a 10-fold binding selectivity
toward the ERb subtype have been identified. These compounds have also been shown to be potent full agonists in the functional
assay by activation of ERE promoted transcription, with the best compound being 20-fold more potent than genistein.
ꢀ 2004 Elsevier Ltd. All rights reserved.
1. Introduction
thrombosis, which result in poor patient compliance.1
Therefore, there is still a great need to find alternative
pharmaceutical agents that possess the beneficial effects
of estrogen but which are devoid of the side effects of
ERT.
The steroidal hormone estrogen plays an important role
in the development and maintenance of the female
reproductive system as well as other tissues such as
bone, the cardiovascular system and the central nervous
system. The decrease of the circulating level of estrogen
in women during and after menopause is responsible for
most peri- and postmenopausal symptoms such as hot
flashes, mood swings, accelerated loss in bone mass and
bone strength (osteoporosis), age-associated memory
decline, and stroke. Currently, the commonly prescribed
treatment for menopausal symptoms and related dis-
eases is estrogen replacement therapy (ERT), which has
proven to be effective in the treatment of vasomotor
symptoms, in preventing bone loss and reducing the
incidence of bone fractures, and reducing the risk of
cardiovascular diseases. Nevertheless, estrogen replace-
ment therapy has also been associated with several
adverse effects, such as an increase in the risk of uterine
endometrial cancer and breast cancer, and deep vein
Estrogen exerts most of its biological functions through
its interaction with the estrogen receptor (ER), which
has two subtypes (ERa and ERb). 17b-Estradiol binds
to the two receptors with similar affinity. Even though
the two receptor subtypes share significant homology in
their ligand binding domains, their different tissue dis-
tribution patterns suggest they may have distinct bio-
logical functions. Particularly, the expression of ERb in
osteoblast cells, vascular endothelial cells, and several
regions of the brain has implied that ERb may be
involved in mediating some of the beneficial actions of
estrogen in these tissues. The low expression level of
ERb in reproductive tissues such as uterus suggests that
a selective ERb agonist may maintain the beneficial
effects of estrogen without the increased risk of breast
and endometrial cancer.2
Besides the natural product genistein, there are only a
limited number of reports of nonsteroidal ERb selective
agonists. These include the arylbenzthiophenes, the
Keywords: Estrogen receptor b agonists; Benzoxazine.
* Corresponding author. Tel.: +1-609-818-6493; fax: +1-609-818-3460;
e-mail: wu.yang@bms.com
0960-894X/$ - see front matter ꢀ 2004 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2004.01.099

2328
W. Yang et al. / Bioorg. Med. Chem. Lett. 14 (2004) 2327–2330
diarylpropionitriles, and the conformational constrained
phytoestrogens.3 Genistein is an isoflavonoid that has
been reported to be an estrogen receptor agonist with
modest selectivity toward ERb. However, it has been
shown to possess some potential liabilities.4 In order to
maintain the agonist activity of genistein and its selec-
tivity toward ERb while possibly eliminating the
potential liabilities, we chose to modify the core
heterocycle from benzopyrone to benzoxazine. We also
decided to retain the two terminal hydroxyl groups,
because they are geometrically arranged to potentially
mimic the two hydroxyl groups in 17b-estradiol and are
known to be essential for the agonist activity of 17b-
estradiol (Fig. 1). Here we report the synthesis and
structure–activity relationships of a benzoxazine series
of estrogen receptor agonists, including compounds with
modest selectivity toward ERb.
OMe
OMe
OMe
b
a
c
NO2
OH
NH2HCl
OH
MeO
OH
MeO
MeO
MeO
MeO
2b
NO2
NH2HCl
OH
d,b
OMe
MeO
OMe
Scheme 2. Reaction conditions: (a) NO2BF4, DME, )78 ꢁC to )50 ꢁC;
(b) Pd/H2, MeOH/CHCl3; (c) HNO3/CH3SO3H/H2O/urea, 0 ꢁC; (d)
BBr3, CH2Cl2, 0 ꢁC.
The p-methoxyaniline precursors (2b,c) were prepared
readily from substituted 1,3-dimethoxybenzenes as
illustrated in Scheme 2. Regioselective nitration of 3,5-
dimethoxyphenol7 followed by hydrogenolysis afforded
compound 2b in 52% yield. Analogously, compound 2c
was prepared in three steps by nitration of 3,5-dimeth-
oxytoluene,8 regioselective ortho-demethylation9 fol-
lowed by hydrogenolysis.
2. Chemistry
The synthesis of the benzoxazine compounds 4a–n in
Table 1 is outlined in Scheme 1. The starting a-bro-
momethyl-40-methoxyphenyl ketones 1 were obtained
either from commercial sources (R2 ¼ R3 ¼ H) or via
bromination of substituted acetophenones with bromine
in chloroform.5 Treatment of 1 with p-methoxyanilines
2a–c afforded the core benzoxazines 3,6 which upon
standard demethylation with BBr3 provided the final
compounds 4 in good yields. In certain instances, when
the reaction was allowed to proceed too long, further
bromination of 1 at the ortho position resulted in side
products where R4 ¼ 3-Br.
3. Results and discussion
To screen compounds for their affinity and selectivity
for ERa versus ERb, we set up an in vitro ligand
binding assay using [3H]17b-estradiol as radioligand in a
scintillation proximity assay (SPA).10 For both ERb and
ERa, recombinant human maltose binding protein tag-
ged, biotinylated ligand binding domains are used.
Compounds were also profiled for their ability to acti-
vate an estrogen response element (ERE) in a luciferase
reporter gene assay. HeLa cells were stably transfected
with a construct expressing the human ERb or ERa (full
length) receptor and co-transfected with a luciferase
3'
OH
OH
R1
OH
O
3
N
O
R4
2
R2
HO
HO
O
7
R3
Table 1. Benzoxazine compounds 4a–n
Genistein
Binding: ERβ IC50 = 0.39 µM; ERα IC50 = 6.1 µM
OH
R1
N
R4
Figure 1. Genistein and 3-aryl-7-hydroxybenzoxazines.
R2
R3
HO
O
R1
O
NH3+Cl-
OH
Br
a
Compounds
R1
H
R2
R3
H
R4
H
+
R2 R3
4a
4b
4c
4d
4e
4f
H
MeO
MeO
HMe
HEt
HH
HH
n-Pr
Me
2a R1 = H
1
2b R1 = OMe
2c R1 = Me
H
HMe
HH
H
HH –(CH
HH –(CH
2CH2)–
2)–
OMe
OH
R1
R1
4g
4h
4i
b
N
O
N
O
OHH
OHMe
OHEt
OHEt
OHMe
Me
H H
R4
R2
R3
R2
R3
H
H
H
3
MeO
HO
4j
H
4k
4l
H
Me
0-Br
R
1 = H, OH, Me
3
4
H
4m
4n
Me
Et
HH
HH
Scheme 1. Synthesis of benzoxazines. Reaction conditions: (a) K2CO3,
acetone, reflux; (b) BBr3, CH2Cl2, 0 ꢁC.
Me

W. Yang et al. / Bioorg. Med. Chem. Lett. 14 (2004) 2327–2330
2329
Table 2. Estrogen receptor binding affinities (3[H]-17b-E2 binding) and receptor gene activation (ERE) of reference compounds and compounds 4a–n
a
a
Compounds
Competition binding (IC50
)
ERE Activity (EC50)
ERb (nM)
ERa (nM)
Ratio a=b
ERb (nM)
ERa (nM)
17b-E2
Estrone
19
540
11
210
0.58
0.4
16
0.67
120
200
160
30
0.48
120
950
Genistein
4a
390
6100
16,000
1700
850
2900
330
5.7
5.0
4.3
5.2
1.6
2.3
4.0
12
6900
64
4b
4c
200
320
7
18
59
4d
4e
1700
8600
6000
5600
33,000
2900
2800
6300
1200
5300
6000
320
1700
3100
12,000
900
5600
2600
1400
2800
310
140
240
77
4f
4g
4h
4i
Enantiomer A
270
47
1800
150
9.4
12
240
50
74
140
290
Enantiomer B
4j
1100
120
5.7
10
26
170
4k
4l
1100
1700
90
4.8
3.5
3.6
2.0
180
120
16
2000
1400
46
4m
4n
90
180
8
170
a Values are average of two–four experiments.
reporter gene construct driven by an ERE–TK minimal
promoter.11
yet with improved potency. Bromine substitution at the
30 position (4k), a side product isolated during prepa-
ration of 4j, was not well tolerated. When the 5-position
was substituted with a methyl group instead of a
hydroxyl group (4m,n vs 4i,j) or just hydrogen (4m,n vs
4b,c), the potency of these compounds increased by 2- to
3-fold, however, at the expense of selectivity.
The binding affinity and selectivity of the benzoxazine
compounds 4a–n for an estrogen receptors were deter-
mined in the binding assay described above (Table 2).
17b-Estradiol exhibited essentially no selectivity toward
the two ER receptor subtypes with IC50s of 19 and
11 nM for ERb and ERa, respectively. The natural
product genistein, however, showed a modest selectivity
(16-fold) for ERb, but the binding potency was con-
siderably less than that for 17b-estradiol. The results for
the above two reference compounds were in accordance
with those reported in the literature by Schopfer et al.3a
The effect of stereochemistry at the chiral center of the
benzoxazine on the estrogen receptor binding activity
was also examined. The two enantiomers of compound
4i were separated by chiral preparative HPLC12 and
tested in the binding and transcription assays. The fast
eluting enantiomer A was found to be more than four
times more potent and twice as selective as the slow
eluting enantiomer B, though the absolute stereochem-
istry was not determined.
The parent, unsubstituted benzoxazine compound 4a
displayed some selectivity (5-fold) for ERb, even though
the binding affinity was weak. Introduction of a methyl
substituent at the R2 position (4b) improved the binding
affinity for both ERb and ERa by about 10-fold, but
with minimal effect on selectivity. Increasing the size of
the substituent from methyl to ethyl (4c) or n-propyl (4d)
did not have a dramatic impact on either selectivity or
potency. Addition of a second methyl group at the 2-
position (4e) or making conformationally constrained
analogues by cyclization of the benzoxazine ring with
the 3-aryl ring (4f,g) all resulted in much reduced ER
binding. Interestingly, this is in contrast to what was
recently reported for constrained phytoestrogen ana-
logues, where significant enhancement of potency was
observed when the pendant phenyl ring was fused to the
pyrone core through an oxygen bridge.3c To make a
direct comparison with genistein, another hydroxyl
group was introduced at the 5-position of the benzox-
azine ring. The resulting compounds 4h–j and 4l all
exhibited a 2-fold improvement in selectivity for ERb;
while maintaining similar potency compared with their
unsubstituted parent compounds 4a–c and 4e. Com-
pound 4j reached selectivity close to that of genistein,
The selectivity and SAR of these benzoxazines can be
rationalized structurally by docking these analogues into
the ERb structure in an orientation analogous to those
of other ligands seen in crystal structures, such as that of
genistein.13 The R2 position of the benzoxazine is ideally
situated to impact directly the L384M (a to b) variation.
Its sp3 and slightly out-of-plane character allows
hydrophobic substituents, including a hydrogen atom,
to approach closely. Larger substitutions, such as the
methyl group of 4i, will have a stronger hydrophobic
interaction, resulting in increased potency. Little addi-
tional effect on selectivity and only 2-fold improvement
on potency was observed by an ethyl substitution, since
the likely conformation of the ethyl group would place
the terminal methyl group away from residue 384. The
R1 group on the 5-position of benzoxazine is placed to
allow direct interaction with the M421I variation. The
selectivity for ERb suggests that there could be unfa-
vorable steric or electrostatic interactions with the
methionine residue of ERa. That the hydroxyl substi-
tution is more selective than the methyl suggests a role

2330
W. Yang et al. / Bioorg. Med. Chem. Lett. 14 (2004) 2327–2330
Acknowledgements
We are thankful to Drs. Robert Zahler and Ruth Wexler
for valuable suggestions.
References and notes
1. (a) Rozenberg, S.; Vandromme, J.; Kroll, M.; Pastijn, A.;
Liebens, F. Int. J. Fertil. Menopausal. Stud. 1995, 40, 23–
32; (b) Rees, M. C. Br. J. Obstet. Gynaecol. 1997, 104, 1–3.
2. For a review, see: MacGregor, J. I.; Jordan, V. Pharmacol.
Rev. 1998, 50, 151–196.
3. (a) Schopfer, U.; Schoeffter, P.; Bischoff, S. F.; Nozulak,
J.; Feuerbach, D.; Floersheim, P. J. Med. Chem. 2002, 45,
1399–1401; (b) Meyers, M. J.; Sun, J.; Carllson, K. E.;
Marriner, G. A.; Katzenellenbogen, B. S.; Katzenellen-
bogen, J. A. J. Med. Chem. 2001, 44, 4230–4251; (c)
Miller, C. P.; Collini, M. D.; Harris, H. A. Bioorg. Med.
Chem. Lett. 2003, 13, 2399–2403.
4. (a) Strick, R.; Strissel, P.; Borgers, S.; Smith, S. L.;
Rowley, J. D. Proc. Natl. Acad. Sci. U.S.A. 2000, 97,
4790–4795; (b) Boos, G.; Stopper, H. Toxicol. Lett. 2000,
116, 7–16.
5. Allah, H. M. F.; Soliman, R. J. Heterocycl. Chem. 1987,
24, 1745–1748.
6. Shridhar, D. R.; Ram, B.; Reddy, G. J. J. Indian Chem.
1989, 66, 138–139.
7. Dwyer, C. L.; Holzapfel, C. W. Tetrahedron 1998, 54,
7843–7848.
8. Barnett, J. W.; Moodie, R. B.; Schofield, K.; Taylor, P. G.;
Weston, J. B. J. Chem. Soc., Perkin Trans. 2 1979, 747–
755.
Figure 2. A view of the 3-aryl-7-hydroxybenzoxazine 4i, in the binding
pocket of ERb.
for unfavorable electrostatic interactions with the M421
of ERa (Fig. 2).
These benzoxazine compounds were also evaluated in a
transcription assay for their ability to activate the
estrogen receptor-mediated transcription using lucifer-
ase as the reporter gene (columns 5 and 6 of Table 2).
All the compounds were found to act as full functional
agonists (>80% intrinsic activity of that of 17b-estra-
diol). The selectivity observed in the transcription assay
was generally in agreement with that seen in the binding
assay, while the EC50s shifted 10- to 20-fold from the
IC50s. Compound 4n, being the most potent compound
in this series, showed an EC50 of 8 nM, only 10-fold less
potent than 17b-estradiol and more than 20-fold more
potent than genistein in these in vitro assays.
9. Atkinson, J.; Morand, P.; Arnason, J. T.; Niemeyer, H.
M.; Hector, R. J. Org. Chem. 1991, 56, 1788–1800.
10. Nichols, J.; Parks, D. J.; Consler, T. G.; Blanchard, S. G.
Anal. Biochem. 1998, 257, 112–119.
11. Kumar, V.; Green, S.; Stack, G.; Berry, M.; Jin, J. R.;
Chambon, P. Cell 1987, 51, 941–951.
4. Conclusions
12. Conditions for chiral preparative separation: Chiralcel AD
column, 5 · 50 cm, 40% isopropanol/hexanes, 75 mL/min
isocratic, 254 nm.
Based on the natural product genistein, a novel series of
3-aryl-7-hydroxylbenzoxazines have been prepared and
evaluated for their binding to the estrogen receptors
ERa and ERb. Several of these compounds exhibited
similar binding selectivity toward ERb as that seen with
genistein, but with improved potency. These selective
analogues were also found to be potent full agonists in
an estrogen receptor-mediated gene transcription assay.
13. (a) Brzozowski, A. M.; Pike, A. C. W.; Dauter, Z.;
€
Hubbard, R. E.; Bonn, T.; Engstrom, O.; Ohman, L.;
Greene, G.; Gustafsson, J. A.; Carlquist, M. Nature 1997,
389, 753–758; (b) Pike, A. C. W.; Brzozowski, A. M.;
€
Hubbard, R. E.; Bonn, T.; Thorsell, A.; Engstrom, O.;
Ljunggren, J.; Gustafsson, J. A.; Carlquist, M. EMBO J.
1999, 18, 4608–4618.