A new Suzuki synthesis of triphenylethylenes that inhibit aromatase and bind to estrogen receptors α and β

Article abstract of DOI:10.1016/j.bmc.2016.08.064

The design and synthesis of dual aromatase inhibitors/selective estrogen receptor modulators (AI/SERMs) is an attractive strategy for the discovery of new breast cancer therapeutic agents. Previous efforts led to the preparation of norendoxifen (4) deriva

Full text of DOI:10.1016/j.bmc.2016.08.064

Accepted Manuscript
A new Suzuki synthesis of triphenylethylenes that inhibit aromatase and bind
to estrogen receptors α and β
Li-Ming Zhao, Hai-Shan Jin, Jinzhong Liu, Todd C. Skaar, Joseph Ipe, Wei Lv,
David A. Flockhart, Mark Cushman
PII:
S0968-0896(16)30681-2
DOI:
http://dx.doi.org/10.1016/j.bmc.2016.08.064
BMC 13247
Reference:
Bioorganic & Medicinal Chemistry
To appear in:
Received Date:
Revised Date:
Accepted Date:
10 July 2016
25 August 2016
29 August 2016
Please cite this article as: Zhao, L-M., Jin, H-S., Liu, J., Skaar, T.C., Ipe, J., Lv, W., Flockhart, D.A., Cushman, M.,
A new Suzuki synthesis of triphenylethylenes that inhibit aromatase and bind to estrogen receptors α and β,
Bioorganic & Medicinal Chemistry (2016), doi: http://dx.doi.org/10.1016/j.bmc.2016.08.064
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers
we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and
review of the resulting proof before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

A new Suzuki synthesis of triphenylethylenes that inhibit aromatase and bind to
estrogen receptors  and 
Li-Ming Zhao,†,‡ Hai-Shan Jin,†,‡ Jinzhong Liu,§ Todd C. Skaar,§ Joseph Ipe,§ Wei Lv,†
David A. Flockhart,§ and Mark Cushman†,*
†Department of Medicinal Chemistry and Molecular Pharmacology, College of Pharmacy, and
The Purdue University Center for Cancer Research, Purdue University, 575 Stadium Mall Drive,
West Lafayette, Indiana 47907, United States
‡School of Chemistry and Chemical Engineering, and Jiangsu Key Laboratory of Green Synthetic
Chemistry for Functional Materials, Jiangsu Normal University, Xuzhou 221116, Jiangsu, China
§Division of Clinical Pharmacology, Department of Medicine, Indiana University School of
Medicine, Indiana Institute for Personalized Medicine, Indianapolis, Indiana 46202,
United States
1

1. Abstract
The design and synthesis of dual aromatase inhibitors/selective estrogen receptor
modulators (AI/SERMs) is an attractive strategy for the discovery of new breast cancer
therapeutic agents. Previous efforts led to the preparation of norendoxifen (4) derivatives with
dual aromatase inhibitory activity and estrogen receptor binding activity. In the present study,
some of the structural features of the potent AI letrozole were incorporated into the lead
compound (norendoxifen) to afford a series of new dual AI/SERM agents based on a
symmetrical diphenylmethylene substructure that eliminates the problem of E,Z isomerization
encountered with norendoxifen-based AI/SERMs. Compound 12d had good aromatase inhibitory
activity (IC50 = 62.2 nM) while also exhibiting good binding activity to both ER-α (EC50 = 72.1
nM) and ER-β (EC50 = 70.8 nM). In addition, a new synthesis was devised for the preparation of
norendoxifen and its analogues through a bis-Suzuki coupling strategy.
2

2. Introduction
In spite of considerable therapeutic advances, breast cancer remains a significant public
health problem.1 Estrogens play a prominent role in stimulating the growth and development of
the majority of breast cancers.2, 3 Hence, estrogen receptors (ERs) are important targets for the
development of new therapeutic agents for breast cancer treatment. Selective estrogen receptor
modulators (SERMs), which were developed in the late 1950's and early 1960's, are widely used
for the treatment of breast cancer.4 SERMs are able to bind to the estrogen receptors (ER-α and
ER-β) in much the same way as estrogens do, and they block ERs in breast cancer cells in breast
tissue while stimulating ERs in normal tissues.5 Tamoxifen (1) is a representative antagonist of
the ERs in breast tissue and is the most commonly used SERM for the treatment of breast cancer
(Figure 1). However, despite the fact that many patients benefit from tamoxifen, resistance to it
often emerges, resulting in therapeutic failure.6, 7 Aromatase inhibitors (AIs), which prevent the
formation of estradiol, were developed for the treatment of ER-positive breast cancer during the
1980's.8 They have shown superior clinical efficacy compared to tamoxifen in postmenopausal
breast cancer patients,8 and are considered as an alternative strategy for tamoxifen-resistant
breast cancer. Unfortunately, the use of AIs is accompanied with significant side effects,
including reduction of bone density, severe musculoskeletal pain, and increased frequency of
fractures and cardiovascular events.9-12
Combination endocrine therapy has emerged as an effective cancer treatment paradigm.13
Several clinical trials have revealed a significant benefit resulting from combination endocrine
therapy involving administration of a SERM and an AI.14, 15 However, this approach has some
3

drawbacks. For example, in the ATAC trial, the combination of anastrozole (an AI) and
tamoxifen (a SERM) was less effective than anastrozole alone.16 Moreover, a patient who takes a
number of different drugs is at greater risk for side effects and drug interactions.
Dual AI/SERMs might be expected to be more effective than the conventional
combination of tamoxifen and an AI. The ER blocking activity of a dual AI/SERM in cancer
cells might act synergistically with the AI activity to inhibit cancer cell proliferation, while in
normal tissues the ER stimulation of a dual AI/SERM would be expected to alleviate the side
effects resulting from the global estrogen depletion caused by the AI activity of the dual
AI/SERM. This therapeutic hypothesis motivated the search for compounds that inhibit
aromatase and bind to estrogen receptors. Norendoxifen (4, Figure 1) was found to be an active
tamoxifen metabolite that binds to ERs and is also a potent AI,17, 18 and that discovery has
provided a platform for the design and synthesis of dual AI/SERMs based on the structure of
norendoxifen.18-20 Subsequent work proved that installation of a 4'-hydroxy group on
norendoxifen to make the metabolite 5 increased potency vs. aromatase and the two estrogen
receptors.19 More recently, it was determined that the aminoethoxy side chain of norendoxifen
can be replaced by a phenolic hydroxyl group and the activity vs. all three receptors (AI, ER-,
and ER-) maintained as long as the ethyl group is replaced by an imidazolylmethyl moiety (e.g.
compound 6) that can coordinate to the iron of aromatases.20 Initial attempts to install a 4'-amino
group in norendoxifen derivatives led to mixed results that were generally disappointing with
regard to simultaneous binding to all three receptors.20 In spite of that, the present investigation
was launched in an attempt to simultaneously optimize activity against aromatase, ER-, and
4

ER-by replacement of the hydroxyl groups of 4'-hydroxynorendoxifen (5) derivatives with
amino groups or nitro groups and elimination of the 2'-aminoethyl moiety. The hypothesis was
that activity against aromatase, ER-, and ER- could be maintained in aminated derivatives
even in the absence of imidazole and aminoethyl functionality using a structure-based drug
design approach that would take advantage of the known structures of the receptors.
Figure 1. Structures of the SERM tamoxifen (1) and its metabolites 4-hydroxytamoxifen (2),
endoxifen (3), norendoxifen (4), and 4'-hydroxynorendoxifen (5), and an active synthetic
imidazole analogue 6.
5

Of the third generation AIs, letrozole is 2-5 fold more potent than anastrozole and
exemestane in its inhibition of aromatase in noncellular systems and 10-20 fold more potent in
cellular systems (Figure 2).21 The structure of letrozole consists of two pharmacophores. One is
the triazole ring. The other is the symmetrically substituted diphenylmethane fragment that has
two identical substituents incorporated at the 4- and 4'-positions. The incorporation of a basic
nitrogen and a symmetrically substituted diphenylmethane fragment into norendoxifen analogues
might therefore provide an approach to optimize aromatase inhibition (Figure 3). Therefore,
norendoxifen was modified by the removal of the aminoethoxyl side chain and introduction of a
nitro or amino group in the para position of the "A" ring (Figure 3). The resulting compounds
have no geometrical isomers and, similar to anastrozole and letrozole, they also incorporate
hydrogen bond acceptors.
Figure 2. Structures of the third generation AIs letrozole (7), anastrozole (8) and exemestane (9).
6

Figure 3. Design strategy for hybrid AI/SERMs.
3. Results and Discussion
3.1 Synthesis and Evaluation of Triphenylethylenes 12a-d.
The importance of the nitro group as an H-bond acceptor in potent and selective AIs has
been emphasized previously by Gobbi and co-workers.22 Initial attempts were made to
synthesize the desired nitrated triphenylethylenes 12 by McMurry coupling of benzophenones
with propiophenones using the methodology previously reported by our group and others.23-25
However, when 4-nitroacetophenone 10a was treated with 4,4'-dihydroxybenzophenone 11 in
dry THF in the presence of TiCl4 and Zn, none of the desired McMurry product 12a was
observed (Scheme 1). Attempts to use BOC-protected 4-aminoacetophenone in place of
4-nitroacetophenone 10a were uniformly unsuccessful.
7

O
O
+
HO
OH
O2N
10a
11
TiCl4/Zn
THF/reflux
O2N
HO
OH
12a
Scheme 1. Attempted Preparation of 12a through the McMurry Reaction.
The failure of the McMurry approach led to another strategy to transform 10a into 12a,
namely conversion of ketone 10a into a 1,1-dibromo-1-alkene 14a before performing
27
palladium-catalyzed coupling reactions.26,
Subsequent coupling of 14a with
4-hydroxyphenylboronic acid under Suzuki conditions proceeded smoothly to provide the
desired product 12a in good yield (67%) (Scheme 2).
8

O2N
O2N
b
a
R
R
O
NNH2
13a-b
10a-b
a R = H
b R = CH3
a R = H
b R = CH3
O2N
O2N
R
c
R
Br
Br
14a-b
XH
XH
12a-d
a R = H
b R = CH3
a R = H, X = O
b R = CH3, X = O
c R = H, X = NH
d R = CH3, X = NH
Scheme 2. Synthesis of 12a-d. Reagents and conditions: (a) N2H4·H2O, EtOH; (b) CBr4, CuCl,
DMSO; (c) 4-(HO)PhB(OH)2 or 4-(H2N)PhB(OH)2, PdCl2(PPh3)2, Na2CO3, THF-H2O.
The analogues 12b-d were prepared by employing the same strategy as that used for the
synthesis of compound 12a (Scheme 2). The nitro-substituted ketones 10a and 10b were initially
treated with hydrazine hydrate at reflux in EtOH to provide the hydrazones 13a and 13b in 85%
and 90% yields, respectively.26 The hydrazones 13a and 13b were reacted with CBr4 in the
presence of CuCl to provide the 1,1-dibromo-1-alkenes 14a and 14b in 65% and 50% yields,
respectively.27 Finally, the bis-Suzuki arylation of 14a and 14b with 4-hydroxyphenylboronic
acid or 4-aminophenylboronic acid in the presence of PdCl2(PPh3)2 at 70 ºC in THF-H2O
resulted in the formation of 12a-d in 47-57% yields.
Compounds 12a-d were evaluated for their aromatase inhibitory activities and ER
binding affinities (Table 1). The IC50 and EC50 values for the previously reported compounds 13,
9

(E,Z)-norendoxifen, (E)-norendoxifen, and (Z)-norendoxifen are included for comparison.18 The
results indicate that nitro-substituted bis-phenol compounds 12a and 12b are very weak AIs with
75% and 78% inhibition at 50 µM, respectively. However, the aniline-type compounds 12c and
12d exhibited remarkably improved inhibitory activity against aromatase and affinity for ER-
and ER- when compared with the unsubstituted derivative 13 and with the phenols 12a and 12b.
They were 113 and 400 times more potent than 13 against aromatase (IC50 220.8 and 62.2 vs
24880 nM), respectively. Compound 12d showed slightly improved aromatase inhibitory activity
and slightly decreased binding affinity to both ER-α and ER-β when compared with the lead
compound (E,Z-norendoxifen). These results indicate that the replacement of the two phenols of
13 with amino groups and the introduction of a p-nitro group in the A-ring contribute in a
positive way to both aromatase inhibition and ER binding affinity. Another conclusion that can
be drawn is that the aminoethoxyl side chain is important for the hydroxyl-substituted
norendoxifen analogues to retain optimal ER binding affinity and aromatase inhibitory activity,
but it is not an essential requirement for amino-substituted analogues.
10

Table 1
Aromatase inhibitory activities and ER binding affinities of 12a-da,b,c
compd
aromatase (IC50, nM,
or percent inhibition)
ER-α (EC50, nM,
ER-β (EC50, nM, or percent
or percent competition)
competition)
12a
75% inhibition at 50 M 19% competition at 100 M 55% competition at 100 M
78% inhibition at 50 M 40% competition at 100 M 66% competition at 100 M
12b
220.8 ± 42.2
62.2 ± 7.8
212.9 ± 53.7
72.1 ± 42.6
80% competition at 100 M
26.9 ± 4.8
486.2 ± 239.5
70.8 ± 5.2
306.9 ± 106.4
35.2 ± 16.8
78.5 ± 57.3
27.5 ± 14.3
5.6
12c
12d
24880 ± 7.8
102.2 ± 32.7
76.8 ± 33.3
1029 ± 318
13
(E,Z)-norendoxifen
(E)-norendoxifen
(Z)-norendoxifen
Estradiol
58.7 ± 1.0
17.0 ± 1.9
5.7
aIC50 values were determined for compounds exhibiting inhibition values higher than 90%.
bPercent aromatase inhibition was determined at the concentration of 50000 nM for each compound.
cPercent ER competition was determined at the concentration of 100000 nM for each compound.
11

Molecular modeling was performed in order to investigate the binding mode of 12d in the
active sites of aromatase and ER-. The dianiline 12d was docked in the active sites of
aromatase (PDB code 3s7928) and ER-(PDB code 3ert29) using GOLD 3.0. These studies
capitalized on the secure molecular modeling foundation established by previous investigations
of the binding of (E)-norendoxifen and several analogues to both aromatase and ER-which
involved GOLD 3.0 docking, Amber 10 molecular dynamics simulations and Amber parm99
18-20
energy minimizations, as well as MM-PBSA binding energy calculations
The calculated
models were consistent with experimentally determined SAR. As shown in Figure 4, the overall
pose of the ligand is close to that present in the previously published models of (E)-norendoxifen
and its analogues bound to aromatase.18-20 The amino group on the aniline ring that is trans to the
nitrophenyl ring hydrogen bonds to the backbone carbonyl of Asp309. In common with the
present situation, all of the previously published models involve hydrogen bonding to Asp309.
However, the present model is different in that the amino group on the aniline ring that is cis to
the nitrophenyl does not hydrogen bond to backbone carbonyl of Leu372. This may be related to
the fact that all of the previously modeled compounds were phenols instead of anilines.
12

Figure 4. Hypothetical molecular model of 12d bound in the active site of aromatase (PDB code
3s7928).
The molecular model calculated for the binding of 12d to ER- is displayed in Figure 5.
In this case, the amino group on the phenyl ring that is cis to the nitrophenyl ring hydrogen
bonds to the hydroxyl of Thr347 while the other amino group hydrogen bonds to the carboxylate
of Glu353 and the backbone carbonyl of Phe404. The overall pose and the involvement of
Thr347 and Glu353 are similar to that in the previously published model of (Z)-norendoxifen and
one of its analogues bound to ER-.20 On the other hand, the present case differs in the hydrogen
bonding to Phe404 instead of Arg394.20
13

Figure 5. Hypothetical molecular model of 12d bound in the active site of ER- (PDB code
3ert29).
3.2 Synthesis and Evaluation of Triphenylethylenes 15a-d
The encouraging findings for 12c and 12d reported in Table 1 led to the preparation of
compounds 15a-d to explore the effect of replacing the "A" ring nitro group with an amino group
as shown in Scheme 3. Compounds 15a-d were easily obtained in 52-80% yield by reduction of
the corresponding nitrosubstituted triphenylalkenes 12a-d with SnCl2 (Scheme 3).
H2N
O2N
R
R
a
HX
XH
XH
XH
15a-d
12a-d
a R = H, X = O
b R = CH3, X = O
c R = H, X = NH
d R = CH3, X = NH
Scheme 3. Synthesis of 15a-d. Reagents and conditions: (a) SnCl2, EtOH.
14

The biological results for compounds 15a-d are summarized in Table 2. Replacement of
the nitro groups in 12a-d with amino groups produced mixed results on the binding affinity with
the ERs while all four amino compounds showed significantly improved inhibitory activity
against aromatase. In particular, compound 15b was the most potent of the AIs synthesized in
this project with an IC50 value of 8.8 nM, which is close to the widely prescribed AI letrozole
(IC50 = 5.3 nM30). These results clearly demonstrate the important role played by the "A" ring
amino group in 15b and 15d in increasing the aromatase inhibitory activity compared to
norendoxifen, the nitro derivatives 12b and 12d, and unsubstituted compound 13.
15

Table 2
Aromatase inhibitory activities and ER binding affinities of 15a-da,b
H2N
a R = H, X = O
b R = CH3, X = O
R
c R = H, X = NH
d R = CH3, X = NH
HX
XH
15a-d
compd
aromatase (IC50, nM) ER-α (EC50, nM,
or percent competition)
11036 ± 827
1711 ± 630
ER-β (EC50, nM,
or percent competition)
857 ± 389
230.0 ± 11.4
8.8 ± 1.6
15a
1263 ± 424
15b
177.1 ± 20.2
13.4 ± 3.0
15c
31% competition at 100 M 77% competition at 100 M
25% competition at 100 M 76% competition at 100 M
15d
5.7
5.6
Estradiol
aIC50 values were determined for compounds exhibiting inhibition values higher than 90%.
bPercent ER competition was determined at the concentration of 100000 nM for each compound.
Comparing 15b with 15a reveals a significant increase in potency on aromatase from IC50
230 to 8.8 nM and from 11036 to 1711 nM in affinity to ER-α. A similar effect on aromatase was
observed for compounds 15c and 15d. Obviously, the ethyl substituent is better than the methyl
when tested on aromatase and ER-, but it is actually slightly worse vs. ER-when comparing
15a to 15b.
3.3 Synthesis and Evaluation of Triphenylethylenes 16a-d.
16

Compounds 16a-d were prepared 30-56% yield by treatment of 12a-d one equivalent of
2-iodoacetamide in the presence of K2CO3 (Scheme 4). This allows comparison of two sets
16a-d and 12a-d with respect to ER binding affinity and aromatase inhibition to determine the
effect of the amide side chain on the dual interaction.
O2N
HX
O2N
HX
R
R
a
NH2
X
XH
O
16a-d
12a-d
a R = H, X = O
b R = CH3, X = O
c R = H, X = NH
d R = CH3, X = NH
Scheme 4. Synthesis of 16a-d. Reagents and conditions: (a) ICH2CONH2, acetone, K2CO3; (b)
LiAlH4, AlCl3, THF.
The biological testing results for compounds 16a-d are summarized in Table 3. IC50 and
EC50 values for the previously reported amide 17 are included for comparison.18 Substitution of
the amino group in one of the phenyl rings of 12c with the amide side chain in 16c decreased
aromatase inhibitory activity (IC50 220.8 vs 645.3 nM). Compound 16d also exhibited decreased
aromatase inhibitory activity when compared with 12d (IC50 286.9 vs 62.2 nM). Compounds 16c
and 16d exhibited elevated potency against both aromatase and ER when compared with 16a,
16b, and 17. The presence of the side chain produced mixed results on the estrogen receptors. It
increased affinity when installed on 12c (compare ER results for 12c and 16c), but it decreased
17

affinity when installed on 12d (compare ER results for 12d and 16d).
Table 3
Aromatase inhibitory activities and ER binding affinities of 16a-da,b,c
O2N
R
a R = H, X = O
b R = CH3, X = O
c R = H, X = NH
d R = CH3, X = NH
NH2
NH2
HO
O
XH
X
O
17
O
16a-d
compd
aromatase (IC50, nM,
or percent inhibition
9724 ± 224
ER-α (EC50, nM,
ER-β (EC50, nM,
or percent competition)
or percent competition)
0% competition at 100 M
0% competition at 100 M
218.4 ± 102.7
16a
0% competition at 100 M
16b
76% inhibition at 50 M 0% competition at 100 M
645.3 ± 246.4
286.9 ± 31.8
9257 ± 195
164.1 ± 96.9
451.2 ± 201.2
16c
346.4 ± 115.3
16d
17
57% competition at 100 M 71% competition at 100 M
5.7 5.6
Estradiol
aIC50 values were determined for compounds exhibiting inhibition values higher than 90%.
bPercent aromatase inhibition was determined at the concentration of 50000 nM for each
compound. cPercent ER competition was determined at the concentration of 100000 nM for each
compound.
3.4 Evaluation of Antiestrogenic Effects in a Functional, Cellular Assay.
In order to gain information about the behavior of the triphenylethylenes beyond that
18

provided by estrogen receptor affinity studies, compounds 12c, 12d, 16c, and 16d (1 μM) were
tested in a functional assay that measured their abilities to block the effects of -estradiol (10 nM)
in MCF-7 human breast cancer cells. These substances were selected for biological testing
because of their relatively high affinity for ER- and ER-, and their potencies as estradiol
antagonists were compared with endoxifen and (E,Z)-norendoxifen. The assay measures
progesterone receptor (PGR) mRNA expression level, and the results of the assay are provided in
Figure 6. -Estradiol (10 nM) increased PGR mRNA expression to a level that was assigned the
100% value. Despite the fact that the affinities for ER- ranged from 451.2 nM (16d) to 72.1 nM
(12d) and those for ER- ranged from 486.2 nM (12c) to 70.8 nM (12d), the % RNA expression
levels only ranged from 14% (16c) to 20% (12c) in the functional cellular assay, and the most
potent compound in the functional assay (16c) was not one with the highest affinity for ER- and
ER-(12d was). Endoxifen, the positive control, was able to antagonize PGR mRNA expression
in the presence of 10 nM estradiol (E2) to the level of 3.5%, while the level of expression in the
presence of (E,Z)-norendoxifen was 22%. These results are generally consistent with those
previously reported for structurally related compounds.19, 20, 31, 32
These data indicate that each of the four compounds effectively inhibited the PGR expression.
Future dose response studies will be required to determine if there are differences in potencies
toward inhibiting the PGR expression or cell proliferation. Since the affinities for the two
estrogen receptors varied quite similarly across the compounds, it will be interesting to determine
their effects on cell proliferation in light of the studies showing that ER- and ER- have
opposing effects on breast cancer cells in vitro.33
19

100
90
80
70
60
50
40
30
20
10
0
Figure 6. Abilities of compounds 16c, 16d, 12d, 12c, endoxifen, and (E,Z)-norendoxifen (1 μM)
to antagonize β-estradiol (E2, 10 nM)-stimulated progesterone receptor (PGR) mRNA expression
in MCF-7 cells. PGR mRNA expression levels relative to estradiol (100%): 16c (14.1%), 16b
(16.3%), 16c (16.8%), 16d (19.6%), endoxifen (3.6%), (E,Z)-norendoxifen (21.5%). MCF-7
cells were preconditioned in charcoal-stripped FBS for 72 hours to remove the estrogens. The
cells were treated with vehicle (1% methanol), 1 µM endoxifen or 1 µM test compound in the
presence of 10 nM β-estradiol (E2) and 10 nM β-estradiol alone for 24 hours. Total RNA was
isolated from the cells and cDNA was prepared. The E2-stimulated PGR mRNA expression was
quantified using a real-time Taqman® PCR assay.
20

3.5 A New Synthesis of (E,Z)-Norendoxifen.
Although the first synthesis of (E,Z)-norendoxifen through the McMurry reaction was
recently reported, the method has limited applicability for the direct synthesis of hydroxylated,
nitro, and amino derivatives.18 A new synthesis of (E,Z)-norendoxifen was therefore explored as
an extension of the present work (Scheme 5). The synthesis of the triphenylalkene framework
commenced with the condensation of propiophenone 18 with hydrazine hydrate in EtOH at
reflux
to
provide
the
hydrazone
19
in
excellent
yield
(88%).34
1-(1,1-Dibromobut-1-en-2-yl)benzene 20 was prepared by reaction of hydrazone 19 with CBr4
(3.0 equiv) in the presence of CuCl (10 mol%) in DMSO in 70% yield.27 The bis-Suzuki
arylation of 20 with 4-hydroxyphenylboronic acid in the presence of PdCl2(PPh3)2 (10 mol%) at
70 ºC in THF-H2O resulted in the formation of diphenol 13 in 52% yield. The diphenol 13 was
treated with K2CO3 and 2-iodoacetamide (1.1 equiv) to afford the amide 21 as the major product
in 32% yield. Finally, the reduction of the amide 21 using LiAlH4 resulted in the generation of
(E,Z)-norendoxifen in 64% yield.
21

Scheme 5. Synthesis of (E,Z)-Norendoxifen through the Suzuki Coupling Reaction. Reagents
and conditions: (a) N2H4·H2O, EtOH; (b) CBr4, CuCl, DMSO; (c) (HO)2BPh-4-OH,
PdCl2(PPh3)2, Na2CO3, THF-H2O; (d) ICH2CONH2, acetone, K2CO3; (e) LiAlH4, AlCl3, THF.
4. Conclusion
A series of novel triphenylethylene derivatives based on
a
symmetrical
diphenylmethylene template were designed and synthesized using a new Suzuki bis-arylation
approach on an,-dibromoalkene. This circumvented problems encountered when the
McMurry approach was tried. The para position of the "A" ring of these compounds was
substituted with a nitro group or an amino group, while the para position of the "B" and "C" ring
22

was substituted with a hydroxyl group or an amino group. SAR studies demonstrated that the
aminoethoxyl side chain is not an essential requirement for tamoxifen (1) analogues to elicit both
aromatase inhibitory and ER binding activity despite the fact that many previous reports have
argued that it is important for the activity.35-39 The introduction of a para-amino group into the
A-ring of the triphenylethylene scaffold led to a remarkable improvement of aromatase
inhibitory activity (IC50 24880 nM for 13 vs. 8.8 nM for 15b).
Progress toward the goal to find new classes of AI/SERMs with favorable aromatase
inhibitory activity and affinity to ER- and ER- was realized with dianiline 12d, which had
potent aromatase inhibitory activity (IC50 = 62.2 nM) while also exhibiting high affinity to both
ER-α (EC50 = 72.1 nM) and ER-β (EC50 = 70.8 nM). Moreover, compounds 15b (IC50 = 8.8 nM)
and 15d (IC50 = 13.4 nM) displayed highly potent aromatase inhibitory activity that was close to
that of the marketed drug letrozole (IC50 5.3 nM) while having low affinity for ER- and ER-.
The following conclusions about the structure-activity relationships were drawn: (1) The
amino groups in the para positions of the "A" ring and "B" ring play key roles in the modulation
of the aromatase inhibitory activity. (2) The unsymmetrical diphenylmethylene substructure of
norendoxifen can be replaced by a symmetrical diphenylmethylene substructure, thereby
eliminating E,Z-isomers of the triphenylethylenes and maintaining activity. (3) The replacement
of ethyl side chain with a methyl group produced a negative result, whether for aromatase
inhibition or ER binding affinity. (4) The aminoethoxyl side chain in triphenylalkene derivatives
is not an essential requirement for optimal interaction with the estrogen receptors and aromatase.
Because of their promising biological activities and no complications arising from the
23

presence of E,Z isomers, the present molecules based on a symmetrical diphenylmethylene
template are suitable candidates for further development toward dual AI/SERMs for breast
cancer treatment.
24

5. Experimental Section
5.1 Chemistry
Melting points were determined using capillary tubes with a Mel-Temp apparatus and are
uncorrected. Reaction products were obtained in pure form directly from reaction mixtures or
13
after column chromatography and did not require additional recrystallization. 1H NMR and C
NMR spectra were recorded using a Bruker ARX300 300 MHz spectrometer or a Bruker DPX
400 MHz spectrometer with TMS as internal standard. High-resolution mass spectra (HRMS)
were recorded on a double-focusing sector mass spectrometer with magnetic and electrostatic
mass analyzers or a Bruker microTOF Q spectrometer. Compound purities were estimated
reversed phase C18 by high pressure liquid chromatography (HPLC) with UV detection at 254
nm. The major peak area of each biologically tested compound was  95% of the combined total
peak area. Cytochrome P450 (CYP) inhibitor screening kits for aromatase (CYP19) inhibition
studies were purchased from BD Biosciences (San Jose, CA). ER α and β competitor assay kits
were purchased from Invitrogen (Carlsbad, CA).
5.1.1 General Procedure for the Synthesis of Hydrazones (13a and 13b).26 A 98%
hydrazine monohydrate solution (1 mL, 20 mmol) was added to a suspension of ketone (10
mmol) in EtOH (10 mL). The mixture was heated to reflux for 2 h. After cooling to room
temperature, the solid was filtered and the crude hydrazone was washed with H2O (20 mL X 2)
and dried in vacuo. The product was used in the next step without further purification.
5.1.2 1-[1-(4-Nitrophenyl)ethylidene]hydrazine (13a). Brick red solid, 85% yield, mp
25

149-150 °C (lit.26 mp 148-149 °C).
5.1.3 1-(1-(4-Nitrophenyl)propylidene)hydrazine (13b). Orange crystalline solid, 90%
1
yield, mp 103-104 °C. H NMR (300 MHz, CDCl3)  8.18 (d, J = 7.1 Hz, 2 H), 7.80 (d, J = 7.1
Hz, 2 H), 5.73 (s, 2 H), 2.64 (q, J = 7.7 Hz, 2 H), 1.18 (t, J = 7.7 Hz, 3 H); 13C NMR (75 MHz,
CDCl3)  147.9, 146.2, 144.4, 125.8, 123.6, 17.9, 9.44; CIMS m/z 194 (MH+); HRCIMS m/z
calcd for C9H12N3O2 (MH+) 194.0924, found 194.0932.
5.1.4 General Procedure for the Synthesis of 1,1-Dibromo-1-alkenes (14a and 14b).27
A 28% aqueous solution of ammonia (1 mL) and CuCl (0.3 mmol) were added to a solution of
hydrazones 13a or 13b (3.0 mmol) in DMSO (3 mL). Then CBr4 (9 mmol) in DMSO (5 mL) was
added dropwise. The reaction mixture was stirred at room temperature for 16 h and quenched
with H2O (30 mL) and extracted with CH2Cl2 (20 mL X 3). After being dried over Na2SO4, the
CH2Cl2 was evaporated and the residue was purified by column chromatography (hexane: EtOAc
= 2: 1) to afford the product 14a or 14b.
5.1.5 1-(1,1-Dibromoprop-1-en-2-yl)-4-nitrobenzene (14a). White solid, 65% yield, mp
81-82 °C. 1H NMR (300 MHz, CDCl3)  8.23 (d, J = 8.7 Hz, 2 H), 7.41 (d, J = 8.7 Hz, 2 H), 2.23
(s, 1 H);13C NMR (75 MHz, CDCl3)  148.5, 141.1, 128.6, 123.8, 123.4, 89.6, 25.9; CIMS m/z
321 (MH+).
5.1.6 1-(1,1-Dibromobut-1-en-2-yl)-4-nitrobenzene (14b). White solid, 50% yield: mp
57-58 °C. 1H NMR (300 MHz, CDCl3)  8.24 (d, J = 8.7 Hz, 2 H), 7.37 (d, J = 8.7 Hz, 2 H), 2.63
(q, J = 7.5 Hz, 2 H), 0.98 (t, J = 7.4 Hz, 3 H); 13C NMR (75 MHz, CDCl3)  147.5, 147.0, 146.9,
129.0, 123.8, 89.3, 32.6, 11.3; CIMS m/z 336 (MH+); HRCIMS m/z calcd for
26

79
C10H10N1O2 Br81Br (MH+) 335.9052, found 335.9048.
5.1.7 General Procedure for the Synthesis of Triphenylalkenes (12a-d). A solution of
1,1-dibromo-1-alkenes 14a or 14b (1.0 mmol), 4-hydroxyphenylboronic acid or
4-aminophenylboronic acid (4.0 mmol), PdCl2(PPh3)2 (0.1 mmol), and Na2CO3 (3.0 mmol) in
THF-H2O (15 mL) was heated to 70 °C under Ar2 for 18 h. After cooling to room temperature,
EtOAc (15 mL) and H2O (10 mL) were poured into the reaction mixture. The aqueous layer was
extracted with EtOAc (20 mL X 3). The combined organic layers were washed with water and
dried, concentrated in vacuo and purified by flash column chromatography (hexane: EtOAc = 2:1)
to afford the products 12a-d.
5.1.8 4-(1-(4-Hydroxyphenyl)-2-(4-nitrophenyl)prop-1-enyl)phenol (12a). Light
brown solid, 67% yield: mp 235-236 °C. 1H NMR (300 MHz, CDCl3)  8.02 (d, J = 9.0 Hz, 2 H),
7.28 (d, J = 7.8 Hz, 2 H), 7.10 (d, J = 8.1 Hz, 2 H), 6.82 (d, J = 8.7 Hz, 2 H), 6.72 (d, J = 9.0 Hz,
13
2 H), 6.52 (d, J = 8.4 Hz, 2 H), 4.78 (s, 1 H), 4.62 (s, 1 H), 2.17 (s, 3 H); C NMR (75 MHz,
CDCl3)  154.7, 154.3, 152.1, 145.5, 141.3, 135.2, 134.9, 132.3, 132.1, 131.3, 130.2, 123.2,
115.0, 114.7, 22.9; ESIMS m/z 370 (MNa+); HRESIMS m/z calcd for C21H17NO4Na (MNa+)
370.1055, found 370.1066; HPLC purity, 100% (90% MeOH, 10% H2O).
5.1.9 4-(1-(4-Hydroxyphenyl)-2-(4-nitrophenyl)but-1-enyl)phenol (12b). Pale yellow
1
solid, 57% yield: mp 111-112 °C. H NMR (300 MHz, methanol-d4)  7.99 (d, J = 8.7 Hz, 2 H),
7.29 (d, J = 8.7 Hz, 2 H), 7.03 (d, J = 8.7 Hz, 2 H), 6.77 (d, J = 8.1 Hz, 2 H), 6.66 (d, J = 8.4 Hz,
2 H), 6.62 (s, 2 H), 6.44 (d, J = 9.0 Hz, 2 H), 2.54 (q, J = 7.8 Hz, 2 H), 0.90 (t, J = 7.8 Hz, 3 H);
13C NMR (75 MHz, methanol-d4)  157.7, 157.1, 152.3, 147.1, 142.8, 139.6, 135.7, 135.3, 133.3,
27

132.0, 131.6, 128.5, 124.0, 116.8, 116.5, 115.9, 115.5, 29.5, 14.0; negative ion ESIMS m/z 360
(M – H+)–; negative ion HRESIMS m/z calcd for C22H18NO4 (M – H+)– 360.1236, found
360.1237; HPLC purity, 95.18% (90% MeOH, 10% H2O).
5.1.10 4-(1-(4-Aminophenyl)-2-(4-nitrophenyl)prop-1-enyl)benzenamine (12c). Brick
red solid, 55% yield: mp 202-204 °C. 1H NMR (300 MHz, methanol-d4)  7.97 (d, J = 6.6 Hz, 2
H), 7.32 (d, J = 7.2 Hz, 2 H), 6.96 (d, J = 6.9 Hz, 2 H), 6.71 (d, J = 6.9 Hz, 2 H), 6.60 (d, J = 6.6
13
Hz, 2 H), 6.41 (d, J = 6.9 Hz, 2 H), 2.16 (s, 3 H); C NMR (75 MHz, methanol-d4)  154.3,
147.4, 146.8, 144.1, 134.4, 133.2, 132.0, 131.8, 131.6, 124.0, 116.2, 115.9, 23.0; ESIMS m/z 346
(MH+); HRESIMS m/z calcd for C21H20N3O2 (MH+) 346.1556, found 346.1572; HPLC purity,
100% (90% MeOH, 10% H2O).
5.1.11 4-(1-(4-Aminophenyl)-2-(4-nitrophenyl)but-1-enyl)benzenamine (12d). Brick
red solid, 40% yield: mp 182-183 °C. 1H NMR (300 MHz, methanol-d4)  7.99 (d, J = 8.7 Hz, 2
H), 7.32 (d, J = 8.7 Hz, 2 H), 6.95 (d, J = 9.0 Hz, 2 H), 6.70 (d, J = 8.4 Hz, 2 H), 6.57 (d, J = 9.0
Hz, 2 H), 6.38 (d, J = 8.4 Hz, 2 H), 2.58 (q, J = 7.8 Hz, 2 H), 0.92 (t, J = 7.5 Hz, 3 H); 13C NMR
(75 MHz, methanol-d4)  152.9, 147.9, 143.7, 138.6, 134.1, 133.1, 132.1, 131.4, 124.0, 116.1,
115.6, 29.5, 14.0; ESIMS m/z 360 (MH+); HRESIMS m/z calcd for C22H22N3O2 (MH+) 360.1712,
found 360.1723; HPLC purity, 97.74% (90% MeOH, 10% H2O).
5.1.12 General Procedure for the Synthesis of Reduction Products (15a-d). A solution
of nitro-containing compounds 12a or 12b or 12c or 12d (0.3 mmol) and SnCl2 (1.5 mmol) in
ethanol (10 mL) was heated at reflux for 5 h. After the reaction mixture was cooled to room
temperature, saturated aq K2CO3 solution was slowly added with stirring until the pH was 8-9.
28

Then the mixture was extracted with EtOAc (10 mL X 3), and the combined organic layer was
dried. The solvent was evaporated and the residue was purified by flash column chromatography
(EtOAc: hexane = 1:1) to afford the products 15a-d.
5.1.13 4-(2-(4-Aminophenyl)-1-(4-hydroxyphenyl)prop-1-enyl)phenol (15a). White
1
solid, 80% yield: mp 154-156 °C. H NMR (300 MHz, DMSO-d6)  9.34 (s, 1 H), 9.12 (s, 1 H),
6.91 (d, J = 8.5 Hz, 2 H), 6.75 (d, J = 8.5 Hz, 2 H), 6.70 (d, J = 8.5 Hz, 2 H), 6.60 (d, J = 8.5 Hz,
13
2 H), 6.42 (d, J = 8.5 Hz, 2 H), 6.32 (d, J = 8.5 Hz, 2 H), 4.97 (s, 2 H), 1.95 (s, 3 H); C NMR
(75 MHz, DMSO-d6)  156.5, 155.8, 147.4, 137.4, 135.7, 133.7, 132.4, 132.2, 131.7, 130.6,
115.7, 115.2, 114.3, 23.9; MALDIMS m/z 317 (M+); HRESIMS m/z calcd for C21H20NO2 (MH+)
318.1494, found 318.1495; HPLC purity, 97.85% (90% MeOH, 10% H2O).
5.1.14 4-(2-(4-Aminophenyl)-1-(4-hydroxyphenyl)but-1-enyl)phenol (15b). White
1
solid, 71% yield: mp 173-175 °C. H NMR (300 MHz, DMSO-d6)  9.34 (s, 1 H), 9.10 (s, 1 H),
6.93 (d, J = 8.5 Hz, 2 H), 6.72 (d, J = 8.5 Hz, 2 H), 6.59 (d, J = 8.5 Hz, 2 H), 6.40 (d, J = 8.5 Hz,
2 H), 6.37 (d, J = 8.5 Hz, 2 H), 6.34 (d, J = 8.5 Hz, 2 H), 4.92 (s, 2 H), 2.30 (q, J = 7.4 Hz, 2 H),
13
0.84 (t, J = 7.4 Hz, 3 H); C NMR (75 MHz, DMSO-d6)  156.7, 155.7, 147.4, 137.7, 135.8,
135.2, 134.9, 132.3, 131.1, 130.8, 115.8, 115.1, 114.5, 29.3, 14.7; MALDIMS m/z 331 (M+);
HRESIMS m/z calcd for C22H22NO2 (MH+) 332.1651, found 332.1651; HPLC purity, 99.59%
(90% MeOH, 10% H2O).
5.1.15 4-(1,2-Bis(4-aminophenyl)prop-1-enyl)benzenamine (15c). White solid, 52%
1
yield: mp 134-135 °C. H NMR (300 MHz, DMSO-d6)  6.77 (d, J = 8.3 Hz, 2 H), 6.75 (d, J =
8.3 Hz, 2 H), 6.49 (d, J = 8.1 Hz, 2 H), 6.40 (d, J = 8.2 Hz, 2 H), 6.32 (d, J = 8.4 Hz, 2 H), 6.21
29