Cytotoxicity and antiestrogenicity of a novel series of basic diphenylethylenes

Article abstract of DOI:10.1021/jm950624t

On the premise that it is necessary to develop antiestrogens with a higher cytotoxic component in order to reduce the risks of the development of heterogeneous malignant cell populations in breast cancer, we studied a novel series of basic diphenylethylenes, for the most part devoid of estrogenic activity, with low antiestrogenicity but much enhanced cytotoxicity compared to the reference drug tamoxifen. The main structural features associated with cytotoxicity were E isomery, substituents of five to eight carbons on the ethylene bond, and dibasicity.

Full text of DOI:10.1021/jm950624t

1104
J . Med. Chem. 1997, 40, 1104-1111
Cytotoxicity a n d An tiestr ogen icity of a Novel Ser ies of Ba sic Dip h en yleth ylen es
J acques Gilbert,* Maryse Fuentes, Tiiu Ojasoo, J ean-Christophe Dore´, and Michel Pons
CNRS-SIRCOB, Universite´ de Versailles/ St. Quentin-en-Yvelines, Versailles 78000, France, CNRS URA 401, Muse´um
National d’Histoire Naturelle, 75231 Paris Cedex 05, France, CNRS URA 09, Universite´ Pierre & Marie Curie (VI),
75252 Paris Cedex 05, France, and INSERM 4439, 34090 Montpellier, France
Received October 8, 1996X
On the premise that it is necessary to develop antiestrogens with a higher cytotoxic component
in order to reduce the risks of the development of heterogeneous malignant cell populations in
breast cancer, we studied a novel series of basic diphenylethylenes, for the most part devoid of
estrogenic activity, with low antiestrogenicity but much enhanced cytotoxicity compared to
the reference drug tamoxifen. The main structural features associated with cytotoxicity were
E isomery, substituents of five to eight carbons on the ethylene bond, and dibasicity.
Ta ble 1. Test Compounds
In tr od u ction
Adjuvant chemotherapy or endocrine therapy is com-
R2
monly administered to patients with breast cancer who
have undergone surgery in order to delay reccurrence
of the disease. The focus of these two treatments differs.
Chemotherapy aims to kill any residual malignant cells
but is frequently ill-tolerated, whereas milder endocrine
therapy, such as the use of antiestrogens, inhibits
hormone-dependent cell growth. Unfortunately, human
estrogen receptor (ER)-positive breast cancer cell lines,
in the absence of estrogen, lose their estrogen respon-
siveness over time,1 and in the clinic, the majority of
patients with ER-positive tumors who initially respond
to endocrine therapy will eventually relapse. The
reasons for progression toward hormone insensitivity
are unknown but may be related to the presence of
mutated receptors,2 to the selection of resistant cells due
to the malignant process itself, and to the acquisition
of drug resistance.3-6
A commonly used adjuvant drug is tamoxifen (trans-
(Z)-1-[4-[2-(dimethylamino)ethoxy]phenyl]-1,2-diphenyl-
1-butene; Nolvadex, Zeneca). The mechanism under-
lying its clinical efficacy is still poorly understood. Its
effect is attributed in major part to an ER-mediated
antiestrogenic action7 but also to a lesser cytostatic/
cytotoxic component, defined as that part of cell growth
inhibition that is not reversible by estrogen.8,9 This
cytotoxicity, which is observed at high concentrations
but might also occur at lower concentrations, may
involve membrane effects,10 proteins such as the anti-
estrogen binding protein,11 calmodulin,12 or protein
kinase C,13-16 and in general cross-talk between differ-
ent signaling pathways.
DEAE = (C2H5)2NCH2CH2O–
DMAE = (CH3)2NCH2CH2O–
DiPAE = ((CH3)2CH)2NCH2CH2O–
R
R1
R3
compd
R
R1
R2
R3
1Z
1E
2Z
2E
3Z
3E
4Z
4E
5E
6
7
8
9
10Z
10E
11Z
11E
12Z
12E
13Z
13E
14Z
14E
H
H
H
H
H
H
H
H
H
H
H
C5H11
C5H11
C6H13
C6H13
C8H17
C8H17
14H29
14H29
C5H11
OH
DEAE
OH
DEAE
OH
DEAE
OH
DEAE
cC4H8-N(CH2)2O
DEAE
DiPAE
DiPAE
DiPAE
OH
DEAE
OH
DMAE
OH
DEAE
OH
DEAE
OH
DEAE
OH
DEAE
OH
C
C
OH
C5H11
DEAE
DiPAE
DiPAE
OH
DEAE
OH
DMAE
OH
DEAE
OH
(CH3)2-CH
(CH2)5d
(CH2)5d
(CH3)2CH
(CH3)2CH
C5H11
H
H
H
H
H
H
H
H
C5H11
C4H9
C4H9
DEAE
OH
DEAE
OH
C6H5CH2
C6H5CH2
C4H9
DEAE
OH
DEAE
OH
C2H5
C4H9
C2H5
DEAE
of systematically substituted basic 1,1-diphenylethyl-
enes (DPEs) with carbon side chains of different length,
linearity (including cyclic substituents), bulk, and sym-
metry attached to position 2 of the ethylene double bond.
Apart from a preliminary investigation of pyrrolidine
derivatives of cyclofenil,21,22 such amino-substituted
chemical structures have received little attention so far.
To establish whether these new compounds displayed
enhanced cytotoxicity while retaining antiestrogenicity,
we tested them for their (i) relative binding affinity
(RBA) for ER, (ii) estradiol agonist and antagonist
activity, and (iii) cytostaticity and cytotoxicity in an
estrogen-dependent breast cancer cell-line (MCF7). The
correlations between chemical structure and response
were analyzed by correspondence factor analysis (CFA)
as in previous structure-activity relationship studies
on triphenylethylenes20,23 and steroids.24-27
Combination therapy with a cytotoxic and antiestro-
gen is an option designed to improve treatment efficacy
while maximizing tolerance.17,18 Its aim is to kill all
malignant cells before recurrences occur and drug
resistance sets in. An alternative to such combination
therapy would be to design a single drug that combines
both features. Our previous papers19,20 have described
the effects of basic 1,2-diphenyl- and 1,1,2-triphenyl-
ethylene derivatives on breast cancer cell proliferation.
The present study focuses on the activities of a series
* To whom correspondence should be addressed: J acques Gilbert,
CNRS-SIRCOB, Universite´ de Versailles/St. Quentin-en-Yvelines, 45
avenue des Etats-Unis, 78000 Versailles, France Tel: 33-01 39 25 44
64. Fax: 33-01 39 25 44 52.
Resu lts a n d Discu ssion
Ch em istr y. Compounds 1-5 and 9-14 (Table 1)
were prepared by condensation of the selected p-
X Abstract published in Advance ACS Abstracts, February 15, 1997.
S0022-2623(95)00624-8 CCC: $14.00 © 1997 American Chemical Society

Cytotoxicity and Antiestrogenicity of Diphenylethylenes
J ournal of Medicinal Chemistry, 1997, Vol. 40, No. 7 1105
Sch em e 1
X-r a y a n d NMR An a lyses. The conformation of the
E isomer of 1 was determined by X-ray crystallography.
With this result in hand, a specific chemical shift could
be attributed to the CH2O group in the basic side chain
attached to the phenyl ring of all monohydroxylated
compounds in the NMR analyses (4.00 ppm (Z), 3.96
ppm (E)). The proportion of each isomer within a
mixture was calculated either by direct integration or
after irradiation of the vicinal CH2 group. In the case
of 14, the proportion of each isomer was determined
with a 300 MHz apparatus only and was calculated on
the basis of the CH3 protons of the ethyl group.
OH + ClCO(CH2)nCH3
OCO(CH2)nCH3
AlCl3
RO
RO
MgBr
CH(CH2)n–1CH3
HO
CO(CH2)nCH3
HO
Bioch em ica l Da ta . Table 2 gives the biochemical
results for the 22 test compounds and 3 reference
substances (estradiol (E2), (Z)-tamoxifen, and (Z)-4-
hydroxytamoxifen) as well as data on the isomerization
of the compounds under in vitro incubation conditions.
The extent of isomerization after 3 days incubation in
an aqueous medium at pH 7.4 and 37 °C was variable.
10Z was virtually totally converted into its isomer 10E
whereas the isomerization of the other compounds was
less pronounced (12Z, 11Z, 4Z, 3Z) or even negligible.
The E isomers were more stable in solution than the Z
isomers.
The RBAs of the compounds for lamb cytosol ER were
low; only 9 and 14 displayed RBAs above 1%. Cytostatic
and cytotoxic effects were given by the extent of inhibi-
tion of MCF7 cell proliferation. Cytostatic activity was
expressed as the compound concentration inhibiting by
50% (IC50) the cell proliferation induced by 1 µM E2;
cytotoxicity was expressed as the compound concentra-
tion leading to cell death in the presence or absence of
E2. Except for 4E and 4Z, all compounds displayed
greater estrogen-irreversible cytotostatic effects than
the reference substances as indicated by their much
Sch em e 2
RO
CH-R2
R1OCOCH2R2 + 2 RO
MgBr
RO
hydroxyphenyl alkyl ketone with the appropriate [[2-(di-
alkylamino)ethoxy]phenyl]magnesium bromide deriva-
tive (Scheme 1). Compound 6 was synthesized by
condensation of 1 mol of heptanoic acid phenyl ester
with 2 mol of [[2-(dialkylamino)ethoxy]phenyl]mag-
nesium bromide (Scheme 2). Compound 7 was obtained
in three steps: first, by condensation of isovaleric acid
ethyl ester with p-anisylmagnesium bromide; second,
by action of pyridine hydrochloride on this dimethoxyl
derivative to give the diol; and lastly, by condensation
of the diol with 2-(diisopropylamino)ethyl chloride under
alkaline conditions. The synthesis of 8 was analogous
to that of 7.
Ta ble 2. Isomeric Purity and Biochemical Activity Profiles of the Test Compounds
response in MCF7 cellsb
response in MVLN cellsc
antiestrogenicity estrogenicity
cytostaticity
(+1 µM E2)
IC50 (µM)
cytotoxicity
(+1 µM E2)
(µM)
cytotoxicity
(-1 µM E2)
(µM)
isomeric puritya (%)
RBAd
compd
1Z
1E
2Z
2E
3Z
3E
4Z
4E
5E
pre
post
IC50 (µM)
% E2 response
% E2 affinity
91
90
76
94
70
68
78
90
95
86
89
69
90
53
74
53
90
88
1.6
1.1
1.6
1.2
1.7
1.6
7.9
8.0
1.6
0.7
1.6
4.2
1.8
2.4
1.7
2.4
1.7
1.7
1.6
1.8
1.7
1.0
3.2
3.2
3.0
3.8
2.9
3.0
30
2.9
2.7
2.9
2.8
5.1
3.6
30
0.9
3.2
1.9
2.5
3.5
4.7
8.0
9.5
3.3
6.0
3.1
2.1
0.2
0.5
1.7
1.5
1.8
0.5
1.5
1.3
1.7
0.15
5
0
5
0
2
0
0
0
0
0
0
4
0
4
3
1
1
2
0
4
0
0
100
0
0.21
0.1
0.13
0.02
0.025
0.035
<0.001
<0.001
0.08
<0.001
0.02
0.12
3.3
1.0
0.13
0.2
0.07
0.75
0.21
0.26
0.15
4.25
100
30
30
2.9
1.3
4.5
8.0
5.0
9.4
8.8
9.3
3.0
5.2
3.0
4.3
3.0
2.8
2.9
2.7
3.6
30
6
7
8
9
3.8
9.2
8.4
8.5
2.9
4.4
2.9
3.2
2.9
2.8
10Z
10E
11Z
11E
12Z
12E
13Z
13E
14E/Ze
E2
81
96
71
63
58
88
87
76
50
6
97
50
80
21
88
73
81
50
TAM Z
OH-TAM Z
8.0
8.0
13
9.4
9.8
6.0
0.3
0.003
1.1
243
0
a
b
Before and after incubation for 3 days in phosphate-buffered saline (pH 7.4). Cytostaticity is given by the test-substance concentration
inhibiting cell proliferation by 50% and cytotoxicity by that leading to cell death (tested concentration range 1-10 µM). c Antiestrogenicity
is given by the concentration of test substance leading to 50% inhibition (IC50) of the luciferase activity due to 0.1 nM E2 and estrogenicity
d
by the maximal induction of luciferase activity in the absence of E2 (% maximum E2 response). Lamb uterus cytosol; 5 h at 25 °C.
e 50/50 mixture of isomers.

1106 J ournal of Medicinal Chemistry, 1997, Vol. 40, No. 7
Gilbert et al.
F igu r e 1. Correspondence factor maps of the data matrix in Table 2 (22 compounds × 6 test conditions) after splitting of the
columns. (τ ) variance associated with an axis; λ ) quality of the representation (If the variables were totally distinct, λ would
be equal 1.)) (A) Correlations among the tests represented as vectors (+, high amplitude; -, low amplitude). The basis of the tests
was as follows: cytostaticity, concentration giving 50% estrogen-irreversible inhibition of MCF7 cell proliferation; cytotoxicity,
concentration leading to cell death; antiestrogenicity in MVLN cells; RBA for the estrogen receptor in lamb uterus cytosol;
estrogenicity in MVLN cells. (B) Correlations among the test compounds linked by a minimum spanning tree. (C) Effect of E and
Z isomery. (D) Effect of the length of the side chain attached to the ethylene bond in the Z (solid line) and E (broken line) series.
(The number of carbon atoms of the side chain is encircled).
lower IC50 values. The concentrations leading to a
cytotoxic effect were also lower, the most cytotoxic
molecule being a dibasic compound (6).
a relative experimental error of about 50% (for the mean
of two experiments involving triplicates). Most com-
pounds were about 10- and 1000-times less effective
than (Z)-tamoxifen and (Z)-4-hydroxytamoxifen, respec-
tively, in antagonising the effect of E2.
Because of the very low RBAs, the antiestrogenic
potency of the test compounds on cell proliferation was
difficult to assess. This antiestrogenicity only manifests
itself at high compound concentrations which, after
8-10 days incubation, also induce a marked cytotoxic
effect. This is why we decided to use the MVLN cell
model (MCF7 cell line obtained by integrating an
estrogen-controlled firefly luciferase gene) for the study
of antiestrogenic and estrogenic activity. For incuba-
tions under 14-16 h, estrogen-regulated luciferase
activity accurately reflects estrogenic potency.28,29 In
Table 2, estrogenicity is thus given by the percent
maximal reporter response obtained with E2 and anti-
estrogenicity by the test-compound concentration needed
to inhibit the E2-induced reporter gene response by 50%.
No or very low E2-type responses were detected; the low
responses are probably meaningful, although subject to
Mu ltiva r ia te F a ctor An a lysis. Structure-activity
relationships are displayed in the correspondence factor
map of Figure 1 which illustrates the specificity and
amplitude of response of each test compound. Figure 1
is a two-dimensional plot of the two main factorial axes
which represent a total of 93% (86.2% (æ1) and 6.8% (æ2))
of the information content within Table 2. In Figure
1A, the tip of a vector denoting a biochemical test
indicates a high response, its origin a low response. The
vectors representing cytotoxicity and cytostaticity (with
a total absolute contribution to the æ1 axis of 99%) are
parallel to this axis which, as indicated above, embodies
86.2% of the information in Table 2. Both these
activities, and in particular cytotoxicity, are thus the
primary discriminating features within this population

Cytotoxicity and Antiestrogenicity of Diphenylethylenes
J ournal of Medicinal Chemistry, 1997, Vol. 40, No. 7 1107
F igu r e 3. Effect of isomery and side-chain length on specific-
ity of response as given by the ø2 distance between isomer pairs
arranged in the order of increasing chain length.
F igu r e 2. Bivariate plot of antiestrogenicity versus RBA.
of molecules. The contiguity of these vectors indicates
that they have similar statistical significance and sug-
gests that an increase in test-compound concentration
might naturally extend an estrogen-irreversible cyto-
static effect into a cytotoxic effect. The secondary
discriminating features (along the æ2 axis) are anties-
trogenic potency and relative binding affinity for
ERswhich are virtually superimposablesand also es-
trogenicity, although this latter result, albeit coherent,
must be viewed with some circumspection because of
the very low activity levels that were recorded (see Table
2). The relationship between antiestrogenic potency and
RBA was confirmed by a bivariate plot (Figure 2) which
disclosed a virtually linear relationship. The near-
orthogonality of the cytotoxicity and antiestrogenicity
vectors suggests that these two phenomena are not
related for this series of molecules.
The positions of the molecules that account for the
relationships among the biochemical variables in Figure
1A are given in Figure 1B (the three reference sub-
stances were not included in the analysis). Figure 1B
can be superimposed on Figure 1A (allowance being
made for the change in scale) and depicts the relation-
ships among compounds, among tests, and between
compounds and tests. As one moves up the æ1 axis, one
encounters increasingly cytotoxic molecules. 4E and 4Z
are the least potent molecules of the series; the dibasic
compound 6 is the most cytotoxic. Movement along the
æ2 axis primarily describes the antiestrogenic and RBA
components of the molecules, 14E/Z, 9, 10Z, and 12Z
being the most active in this respect. Of these mol-
ecules, 10Z and 12Z displayed some estrogenic activity.
The test compounds in Figure 1B have been linked into
a network that is transposed from the results of a
minimum spanning tree analysis19,30 which calculates
the shortest distance among all variables.
isomerization should completely obviate the very clear-
cut pattern observed.
In panel 1D, side chains of five, six, or even eight
carbons are seen to be conducive to increased cytotox-
icity in both the E and Z series (upward displacement
along the æ1 axis). A long chain of 14 carbons resulted
in inactivity (4E and 4Z); shorter chains (three to five
carbons) were associated with higher antiestrogenicity-
RBA values, especially in the Z series. There were,
however, exceptions to this tendency. After permuta-
tion of the ethyl on the nitrogen atom of the side chain
of 1Z to a methyl (11Z), there was a marked decrease
in cytotoxicity (Figure 1B). Compound 13Z with a
benzyl side chain also had reduced cytotoxicity and was
situated near the origin of the axes, suggesting an
average profile. The histograms in Figure 3 give the ø2
distance between each isomer pair and quantify the
effect of isomery and chain length on specificity and
amplitude of response.
Although the introduction of a second basic para side
chain on a phenyl ring of 1 (Z or E) yielded the most
cytotoxic compound of the series (6), a similar modifica-
tion of the cyclofenil type structure (9) decreased cyto-
toxicity and antagonist activity (8) (Figure 1B). Bisub-
stitution of the ethylene double bond (14E/Z compared
to 12E or -Z) increased antiestrogenicity-RBA as well
as cytotoxicity. Garg et al.22 have studied an analog of
9 (a pyrrolidine in lieu of diisopropylamino group) and
found significantly higher RBAs using rat and mouse
cytosol ER (57 and 80%, respectively). However, their
RBA results for the disubstituted analog are comparable
(0.2 and 0.4%, respectively) to those we found for 8.
Con clu sion s
A systematic structure-activity study on a novel
series of basic diphenylethylene derivatives has revealed
that this is a series of molecules with low antiestrogenic
activity, mostly devoid of estrogenic activity, but with
a cytotoxicity severalfold higher than that of tamoxifen.
The antiestrogenic activity is probably due to an inter-
action with the estrogen receptor as indicated by the
quasi-linear relationship found between antiestrogenic
potency and relative binding affinity for the receptor.
Within the series, E isomery was more conducive to
cytotoxicity than Z isomery as were side chains on the
ethylene bond of five to eight carbons. Similar results
have been previously obtained with tamoxifen and
4-hydroxytamoxifen isomers and with triphenylacrylo-
nitrile isomers.15 Short linear or branched chains
Str u ctu r e-Activity Rela tion sh ips. Panels C and
D of Figure 1 are excerpts of Figure 1B. They highlight
the influence of isomery (panel C) and of the length of
the side chain attached to the ethylene bond (panel D)
on the specificity and amplitude of response. In panel
1C, the arrows linking the isomers reflect a general
trend toward increasing cytotoxicity (upward) and
decreasing antiestrogenicity-RBA (leftward) for the E
compared to the Z isomers. Isomerization of the test
compounds during the biochemical assays probably has
little influence on RBA and antiestrogenicity results
because these tests involve short-term incubation only,
but may have a greater impact on cytotoxicity and
cytostaticity results. It is, however, unlikely that such

1108 J ournal of Medicinal Chemistry, 1997, Vol. 40, No. 7
Gilbert et al.
mL) in an ice bath and steam distilled to eliminate the
nitrobenzene. The suspension was extracted with ether, and
the organic layer was treated with 5% NaOH (3 × 50 mL) to
obtain an alkaline phase which was acidified with 1 N HCl
and extracted with ether. The organic layer was washed, dried
(MgSO4), and concentrated to a residual oil which was crystal-
lized (73.4%): mp 90 °C.
(three carbons) tended to be associated with more
antiestrogenic molecules, and a very long side chain of
14 carbons led to inactivity. High cytotoxicity was also
associated with a dibasic structure as previously found
in a series of triphenylacrylonitriles.23 The orthogonal-
ity of the vectors representing cytostatic/cytotoxic effects
and ER-dependent effects (antiestrogenic potency, RBA)
in the multivariate correspondence factor analysis
revealed that these two types of activity are totally
independent for this population of molecules. Conse-
quently, the very high antiestrogenic activity observed
for 10 µM concentrations of several test compounds is
not accounted for by cytotoxicity; it was reversed by 1
µM E2 in the MVLN model, giving rise to a 70-80%
response.
The relative binding affinity of these DPEs for ER is
very low. Their estrogenic activity, as measured with
a chimaeric gene, was extremely lowsabout 1 to 5%sor
undetectable. These observations have been confirmed
in unpublished experiments on compounds 1-8 which
have shown that, within a 10-100 nM concentration
range, 1Z, 2Z, 4E/Z, and 8 enhance the proliferation of
MCF7 cells with a potency of about 10-12% of that of
E2.
In conclusion, we have designed derivatives with a
higher cytostatic/cytotoxic activity than current triphen-
ylethylenes and with some, albeit low, antiestrogenic
activity mediated by ER. These first compounds of a
new series may not be able to act as a substitute for
combination chemotherapy/endocrine therapy in clinical
practice. Nevertheless, their dual activity profile sug-
gests that such substitutes might be feasible and that
the rationale we have adopted could be interesting from
more than just a theoretical point of view.
(E)- a n d (Z)-1-[4-[2-(Dieth yla m in o)eth oxy]p h en yl]-1-(4-
h yd r oxyp h en yl)-1-h ep ten e (1). A solution of p-[2-(diethyl-
amino)ethoxy]bromobenzene (13.6 g, 0.05 mol) in anhydrous
THF (30 mL) was slowly added under an argon atmosphere
to magnesium turnings (1.3 g, 0.053 atom) and iodine crystals
in anhydrous ether (15 mL). The Grignard reagent was heated
at reflux for 2 h and then allowed to cool. A solution of
4-hydroxyheptanophenone (5.20 g, 0.025 mol) in THF (50 mL)
was added in a rapid dropwise fashion. After the mixture was
stirred, heated at reflux for 3 h, and cooled, a saturated
ammonium chloride solution was added dropwise, and stirring
was continued for 2 h. The organic layer was separated, and
the aqueous solution was extracted several times with THF.
The organic extracts were washed with saturated NaCl (25
mL), dried (MgSO4), and concentrated. The residual oil was
dissolved in ether and treated with 2 N HCl to enable recovery
of 4-hydroxyheptanophenone (0.6 g, 11.5%) from the ether
layer. The acidified aqueous layer was treated with 10%
NaOH and extracted with ether. The organic phase was
washed with water, dried (MgSO4), and concentrated. The
residual oil could be partially crystallized (ex petroleum ether)
to yield white crystals (3 g, 30%): mp 105 °C (ex benzene/
petroleum ether, 2/1); Rf 0.27 (CHCl3/CH3OH, 85/15); IR
(CHCl3) ν OH 3580-3080 cm-1 1H NMR (CDCl3) δ 0.78 (t,
;
CH3(CH2)4), 1.01 (t, (CH3CH2)2N), 1.17-1.33 (m, CH3(CH2)3),
2.01 (q, CdCHCH2), 2.62 (q, (CH3CH2)2N), 2.83 (t, NCH2-
CH2O), 3.96 (t, CH2O), 5.83 (t, CdCH), 6.61-7.02 (4d, 8 ArH);
MS m/ z 382 [M + 1]+. Anal. (C25H35NO2) C, H, N. According
to X-ray crystallography, this is the E isomer.
Concentration of the petroleum ether solution that was
separated from the crystals gave a crude oil (9.1 g). Purifica-
tion by chromatography (CHCl3/increasing CH3OH) yielded the
Z isomer (30%): mp 68-9 °C (ex C6H6/ hexane, 5/95); Rf 0.33
(CHCl3/CH3OH, 85/15); IR (CHCl3) υ OH 3580-3080 cm-1; 1H
NMR (CDCl3) δ 0.78 (t, CH3(CH2)2), 1.03 (t, (CH3CH2)2N),
1.17-1.33 (m, CH3(CH2)3), 1.99 (q, CdCHCH2), 2.65 (q,
(CH3CH2)2N), 2.86 (t, NCH2CH2O), 4.00 (t, CH2O), 5.82 (t,
CdCHCH2), 6.56-7.04 (m, 8 ArH); MS m/ z 382 [M + 1]+.
Anal. (C25H35NO2) C, H, N.
Exp er im en ta l Section
Ch em ical Meth ods. Gen er al P r ocedu r es. Melting points
were determined on a Mettler FP 61 apparatus and are
uncorrected. The results of the elemental analyses, performed
in the Microanalytical Laboratory of the CNRS (Vernaison,
France), were within (0.3% of the theoretical values for those
elements shown. 1H NMR spectra (Me4Si internal) were
recorded (δ 0) with a Bruker AC 200E instrument at 200 MHz,
IR spectra with a Perkin-Elmer 1420 instrument, and mass
spectra with a Ribermag (DCI+(NH3)) instrument. Column
chromatography was performed using silica gel 60 Å (35-70
µm). Ether refers to diethyl ether. Petroleum ether refers to
the fraction with the boiling range 30-60 °C unless otherwise
specified. Anhydrous tetrahydrofuran (THF) was obtained by
distillation from potassium and benzophenone.
Gen er a l P r oced u r e for th e P r ep a r a tion of Alk yl Acid
P h en yl Ester s. The preparation of the heptanoic acid phenyl
ester31 will illustrate the procedure. Heptanoyl chloride (50
g, 0.336 mol) was added dropwise to phenol (31.9 g, 0.338 mol)
at 15 °C under an argon atmosphere, and the mixture was
stirred for 4 h at 60 °C. After cooling and addition of water
(50 mL), the aqueous layer was extracted with ether. The
organic extracts were washed with saturated solutions of NaCl
(3 × 15 mL), NaHCO3 (20 mL), and then NaCl (20 mL) again,
dried over MgSO4, and evaporated. Distillation of the residue
gave a fraction with a bp0.01 of 82-86 °C and which was
identified as the title compound (yield 98%).
(E)- a n d (Z)-1-[4-[2-(Dieth yla m in o)eth oxy]p h en yl]-1-(4-
h yd r oxyp h en yl)-1-octen e (2). The above procedure de-
scribed for 1 was used starting from 4-hydroxyoctanophenone
and gave a residual oil from which the E isomer (95%) was
separated by digestion in petroleum ether: mp 101 °C (ex CH3-
OH); Rf 0.29 (CHCl3/CH3OH, 85/15); IR (CHCl3) υ OH 3580-
3080 cm-1 1H NMR (CDCl3) δ 0.78 (t, CH3(CH2)5), 1.03 (t,
;
(CH3CH2)2N), 1.15-1.32 (m, CH3(CH2)4), 2.01 (q, CdCHCH2),
2.66 (q, (CH3CH2)2N), 2.85 (t, NCH2CH2O), 3.97 (t, CH2O), 5.80
(t, CdCH), 6.55-7.00 (4d, 8 ArH); MS m/ z 396 [M + 1]+. Anal.
(C26H37NO2) C, H, N.
Concentration of the petroleum ether-oil mixture followed
by chromatography yielded the Z isomer (75%) as a solid: mp
59 °C (ex petroleum ether); 1H NMR (CDCl3) δ 0.78 (t,
CH3(CH2)5), 1.03 (t, 6H, (CH3CH2)2N), 1.20-1.40 (m, 8H, CH3-
(CH2)4), 2.00 (q, 2H, CdCHCH2), 2.65 (q, 4H, (CH3CH2)2N),
2.86 (t, 2H, NCH2CH2O), 4.00 (t, CH2O), 5.82 (t, CdCH), 6.57-
7.04 (m, 8 ArH).
(E)- a n d (Z)-1-[4-[2-(Dieth yla m in o)eth oxy]p h en yl]-1-(4-
h yd r oxyp h en yl)-1-d ecen e (3). The above procedure using
4-hydroxydecanophenone gave a residual oil which did not
precipitate on addition of petroleum ether but which, after
chromatography, yielded two solids that were crystallized (ex
C6H6/petroleum ether, 20/80). The proportion of each isomer
Gen er a l P r oced u r e for th e P r ep a r a tion of 4-Hyd r oxy-
p h en yl Alk yl Keton es by F r ies Rea r r a n gem en t. The
synthesis of 4-hydroxyhexanophenone32 will be used as an
example. Heptanoic acid phenyl ester (35 g, 0.17 mol) was
added dropwise at 30 °C under an argon atmosphere to a
solution of anhydrous aluminum chloride (45.5 g, 0.341 mol)
in anhydrous nitrobenzene (340 mL) and stirred for 48 h at
40 °C. The mixture was then hydrolyzed with 1 N HCl (100
1
1
was determined by H NMR. Z isomer (72%): mp 63 °C; H
NMR (CDCl3) δ 0.79 (t, 3H, CH3(CH2)5), 1.03 (t, 6H, (CH3-
CH2)2N), 1.07-1.32 (m, 12H, CH3(CH2)6), 2.00 (q, CdCHCH2),
2.64 (q, 4H, (CH3CH2)2N), 2.86 (t, NCH2CH2O), 4.00 (t, CH2O),
5.82 (t, CdCH), 6.56-7.04 (m, 8 ArH); MS m/ z 424 [M + 1]+.

Cytotoxicity and Antiestrogenicity of Diphenylethylenes
J ournal of Medicinal Chemistry, 1997, Vol. 40, No. 7 1109
Anal. (C28H41NO2) C, H, N. E isomer (68%): mp 75 °C; MS
m/ z 424 [M + 1]+. Anal. (C28H41NO2) C, H, N.
(E/Z; 55/45%). Pure E isomer was obtained after repeated
crystallization (ex benzene): mp 95 °C; 1H NMR (CDCl3) δ 0.77
(t, 3H, CH3(CH2)3), 1.03 (t, 6H, N(CH2CH3)2), 1.16-1.34 (m,
4H, CH3(CH2)2CH2), 2.02 (q, CdCH-CH2), 2.66 (q, 4H, N(CH2-
CH3)2), 2.85 (t, NCH2CH2O), 3.97 (t, CH2O), 5.80 (t, CdCH),
(E)- a n d (Z)-1-[4-[2-(Dieth yla m in o)eth oxy]p h en yl]-1-(4-
h yd r oxyp h en yl)-1-h exa d ecen e (4). The above procedure
using 4-hydroxyhexadecanophenone gave a residual oil that
precipitated with petroleum ether. Z isomer from a first crop
(85%): mp 86-7 °C (ex CH3OH); IR (CHCl3) υ OH 3580-3080
cm-1; 1H NMR (CDCl3) δ 0.80 (t, 3H, CH3(CH2)11), 1.03 (t, (CH3-
CH2)2N), 1.17 (br s, CH3(CH2)11), 1.17-1.40 (m, CH2CH2-
CHdC), 2.00 (q, CdCHCH2), 2.64 (q, (CH3CH2)2N), 2.85 (t,
NCH2CH2O), 4.00 (t, CH2O), 5.82 (t, CdCH), 6.56-7.04 (8
ArH); MS m/ z 508 [M + 1]+. Anal. (C34H53NO2) C, H, N. E
isomer (95%): mp 71-73 °C (ex CH3OH). 1H NMR (CDCl3) δ
0.80 (t, CH3(CH2)11), 1.02 (t, (CH3CH2)2N), 1.17 (br s, CH3-
(CH2)11), 1.18-1.40 (m, CH2CH2CHdC), 2.01 (q, CdCHCH2),
2.64 (q, (CH3CH2)2N), 2.84 (t, NCH2CH2O), 3.97 (t, CH2O), 5.81
(t, CdCH), 6.55-7.03 (4d, 8 ArH); MS m/ z 508 [M + 1]+. Anal.
(C34H53NO2) C, H, N.
(E)- a n d (Z)-1-[4-[2-(P yr r olid in yl)eth oxy]p h en yl]-1-(4-
h yd r oxyp h en yl)-1-h ep ten e (5). The above procedure using
4-hydroxyheptanophenone and p-[2-(pyrrolidino)ethoxy]bro-
mobenzene yielded a residual oil which crystallized in part
from petroleum ether; mp 116 °C (ex MeOH). E isomer
(95%): IR (CHCl3) υ OH 3580-3080 cm-1; 1H NMR (CDCl3) δ
0.78 (t, CH3(CH2)4), 1.10-1.40 (m, 6H, CH3(CH2)3), 1.81 (br s,
4H, CH2(CH2)2CH2 (pyrrolidinyl ring)), 2.01 (q, CdCHCH2),
2.66 (s, 4H, (CH2)2N), 2.90 (t, NCH2CH2O), 4.00 (t, CH2O), 5.79
(t, CdCH), 6.44-7.00 (4d, 8 ArH); MS m/ z 380 [M + 1]+. Anal.
(C25H33NO2) C, H, N.
Concentration of the petroleum ether filtrate followed by
purification by chromatography yielded the Z isomer (65%) as
a white solid: mp 102-106 °C; 1H NMR (CDCl3) δ 0.78 (t, CH3-
CH2), 1.13-1.35 (m, 6H, CH3(CH2)3), 1.80 (br s, 4H, CH2(CH2)2-
CH2 (pyrrolidinyl ring)), 1.99 (q, CdCHCH2), 2.68 (s, 4H,
(CH2)2N), 2.92 (t, NCH2CH2O), 4.05 (t, CH2O), 5.80 (t, CdCH),
6.47-7.02 (m, 8 ArH).
[4-[2-(Diisop r op yla m in o)et h oxy]p h en yl](4-h yd r oxy-
p h en yl)cycloh exylid en em eth a n e (9). The procedure used
1-(4-hydroxyphenyl)cyclohexyl ketone and p-[2-(diisopropyl-
amino)ethoxy]bromobenzene: yield 78%; mp 120 °C (benzene/
petroleum ether 40-65 °C, 50/50); 1H NMR (CDCl3) δ 0.96 (d,
12H, (CH3)2CH), 1.51 (br s, 6H, (CH2)3 (3,4,5 of cyclohexyl
ring)), 2.15 (br s, 4H, CdC(CH2)2), 1.65 (t, NCH2CH2O), 1.43
(sept, 2H, N(CH)2), 3.79 (t, CH2O), 6.63-7.12 (m, 8 ArH); MS
m/ z 408 [M + 1]+. Anal (C27H37NO2) C, H, N.
6.53-7.02 (m, 8 ArH); MS m/ z 368 [M + 1]+. Anal. (C24H33
-
NO2) C, H, N. Z isomer (64%): mp 86 °C. 1H NMR (CDCl3)
δ 0.77 (t, CH3(CH2)2), 1.03 (t, 6H, N(CH2CH3)2), 1.20-1.27 (m,
4H, CH3(CH2)2), 2.01 (q, 2H, CdCHCH2), 2.64 (q, 4H, N(CH2-
CH3)2), 2.86 (t, N-CH2-CH2O), 4.00 (t, CH2O), 5.81 (t, CdCH),
6.56-7.04 (m, 8 ArH); MS m/ z 368 [M + 1]+. Anal. (C24H33
NO2) C, H, N.
-
(E)- a n d (Z)-1-[4-[2-(Dieth yla m in o)eth oxy]p h en yl]-1-(4-
h yd r oxyp h en yl)-3-p h en yl-1-p r op en e (13). The procedure
used 1-(4-hydroxyphenyl)-2-phenylethyl ketone. Z isomer
(80%): mp 93-5 °C; 1H NMR (CDCl3) δ 1.02 (t, 6H,
N(CH2CH3)2), 2.63 (q, 4H, N(CH2CH3)2), 2.85 (t, NCH2CH2O),
3.38 (d, CH2Ar), 4.00 (t, CH2O), 6.01 (t, CdCH), 6.61-7.13 (m,
8 ArH). MS m/ z 402 [M + 1]+. Anal. (C27H31NO2) C, H, N.
E isomer (69%): mp 124-5 °C (ex benzene); 1H NMR (CDCl3)
δ 1.02 (t, 6H, N(CH2CH3)2), 2.64 (q, 4H, N(CH2CH3)2), 2.84 (t,
NCH2CH2O), 3.39 (d, CH2C6H5), 3.96 (t, CH2O), 6.00 (t,
CdCH), 6.53-7.20 (m, 8 ArH); MS m/ z 402 [M + 1]+. Anal.
(C27H31NO2) C, H, N.
(E)- a n d (Z)-1-[4-[2-(Dieth yla m in o)eth oxy]p h en yl]-1-(4-
h ydr oxyph en yl)-2-eth yl-1-h exen e (14). The procedure used
1-(4-hydroxyphenyl)-2-ethylpentyl ketone. The isomers could
not be separated either by solubility differences or by chro-
matography: mp 80 °C (ex CH3OH/diisopropyl ether, 1/3); 1H
NMR (300 MHz) (CD3OD) δ 0.82 (t, 3H, CH3(CH2)3), 0.977 and
0.980 (2t, CdCH2CH3 (E/Z, 50/50)), 1.08 (t, 6H, N(CH2CH3)2),
1.23 (sext, 2H, (CH2)2CH2CH3), 1.39 (quint, 2H, CH2CH2CH2-
CH3), 2.08-2.14 (m, 4H, CdC(CH2)2), 2.68 (q, 4H, N(CH2-
CH3)2), 2.89 (t, NCH2CH2O), 4.02 (t, CH2O), 6.68-6.99 (m, 8
ArH); MS m/ z 396 [M + 1]+. Anal. . (C26H37NO2) C, H, N.
1,1-Bis[4-[2-(d iet h yla m in o)et h oxy]p h en yl]-1-h ep t en e
(6). A solution of p-[2-(diethylamino)ethoxy]bromobenzene (25
g, 0.092 mol) in anhydrous THF (50 mL) was added slowly
under an argon atmosphere to magnesium turnings (2.3 g,
0.094 atom) in anhydrous ether (15 mL). The Grignard
reagent was heated at reflux for 2 h and allowed to cool, and
a solution of methyl heptanoate (4.40 g, 0.0306 mol) in THF
(50 mL) was then added slowly at 40 °C. The reaction mixture
was stirred and heated at reflux for 3 h. After cooling, addition
of 1 N HCl, and removal of solvent, the aqueous solution was
extracted with ether to eliminate the unreacted ester. The
aqueous phase was alkalinized with 5% NaOH and extracted
with ether. The organic layer was washed, dried (MgSO4), and
evaporated to give a crude oil (19.5 g) which was distilled. The
240-245 °C (0.04 mmHg) fraction was isolated (11 g, 74.8%):
1H NMR (CDCl3) δ 0.78 (t, 3H, CH3(CH2)4), 0.98 and 1.01 (2t,
12H, (CH3CH2)N), 1.04-1.37 (m, 6H, CH3(CH2)3), 2.01 (q, 2H,
CdCHCH2), 2.50-2.64 (m, 8H, (CH3CH2)N), 2.78 and 2.81 (2t,
4H, NCH2CH2O), 3.95 and 3.99 (2t, 4H, CH2O), 5.85 (t, CdCH),
6.69-7.07 (m, 8 ArH); MS m/ z 481 [M + 1]+. Anal.
(C31H48N2O2) C, H, N.
1,1-Bis[4-[2-(diisopr opylam in o)eth oxy]ph en yl]-3-m eth -
yl-1-bu ten e (7). 1,1-Bis(4-hydroxyphenyl)-3-methyl-1-butene
was prepared by the action of pyridine hydrochloride for 45
min at 200-220 °C on 1,1-bis(4-methoxyphenyl)-3-methyl-1-
butene obtained by reacting ethyl isovalerate with p-anisyl-
magnesium bromide (yield 85%; F 174 °C (diluted EtOH).
2-(Diisopropylamino)ethyl chloride hydrochloride (12.7 g, 0.0635
mol) was added to a mixture of the 1,1-bis(4-hydroxyphenyl)-
3-methyl-1-butene (6.5 g, 0.0257 mol) in toluene (100 mL) and
NaOH (5.08 g, 0.127 mol/15 mL) under continual stirring. The
reaction mixture was heated at 90 °C for 16 h and then poured
into water. The organic phase was separated and the aqueous
phase extracted with toluene. The extracts were washed with
5% NaOH and water, dried (MgSO4), and concentrated to an
oil. Upon distillation, the 210-218 °C (0.01 mmHg) fraction
(10.4 g, 80%) was isolated: 1H NMR (CDCl3) δ 0.90-1.00 (m,
30H, CH3), 2.32 and 2.45 (oct, 1H, CdCHCH(CH3)2)), 2.72 and
2.76 (2t, 4H, NCH2CH2O), 2.97 (sept, 4H, NCH(CH3)2), 3.79
and 3.83 (2t, 4H, CH2O), 5.64 (d, 1H, CdCHCH(CH3)2), 6.67-
7.07 (m, 8 ArH). Anal. (C33H52N2O2) C, H, N.
(E)- a n d (Z)-1-[4-[2-(Dieth yla m in o)eth oxy]p h en yl]-1-(4-
h yd r oxyp h en yl)-3-m et h yl-1-bu t en e (10). The procedure
used 4-hydroxyisobutanophenone. E isomer (95%): mp 120
1
°C (benzene/petroleum ether, 50/50); H NMR (CDCl3) δ 0.91
(d, 6H, (CH3)2CH), 1.03 (t, 6H, (CH3CH2)2N), 2.33-2.44 (m,
1H, CH(CH3)2), 2.66 (q, 4H, N(CH2CH3)2), 2.85 (t, NCH2CH2O),
3.97 (t, CH2O), 5.62 (d, CdCHCH(CH3)2), 6.53-7.06 (m, 8
ArH); MS m/ z 354 [M + 1]+. Anal. (C23H31NO2) C, H, N.
The Z isomer (83%) was obtained as above after purification
by chromatography: mp 91 °C; 1H NMR (CDCl3) δ 0.91 (d,
6H, (CH3)2CH), 1.03 (t, 6H, (CH3CH2)2N), 2.37-2.58 (m, 1H,
CH(CH3)2), 2.63 (q, 4H, N(CH2CH3)2), 2.85 (t, NCH2CH2O), 4.00
(t, CH2O), 5.65 (d, 1H, CdCHCH(CH3)2), 6.61-7.02 (m, 8 ArH);
MS m/ z 354 [M + 1]+. Anal. (C23H31NO2) C, H, N.
(E)- a n d (Z)-1-[4-[2-(Dim et h yla m in o)et h oxy]p h en yl]-
1-(4-h yd r oxyp h en yl)-1-h ep ten e (11). The procedure used
4-hydroxyheptanophenone and p-[2-(dimethylamino)ethoxy]-
bromobenzene and gave a residual oil that partly crystallized.
Repeated recrystallization (ex benzene/hexane, 50/50) yielded
1
the Z isomer (75%): mp 125 °C; H NMR (CDCl3) δ 0.78 (t,
CH3(CH2)4), 1.16-1.32 (m, CH3(CH2)3), 1.96 (q, CdCHCH2),
2.33 (s, N(CH3)2), 2.74 (t, NCH2CH2O), 4.00 (t, CH2O), 5.80 (t,
CdCH), 6.44-7.01 (m, 8 ArH); MS m/ z 354 [M + 1]+. Anal.
(C23H31NO2) C, H, N.
A second crop of solid was collected and recrystallized
several times (ex benzene/hexane in different proportions) to
give the E isomer (58%): mp 91-3 °C; MS m/ z 354 [M + 1]+.
Anal. (C23H31NO2) C, H, N.
(E)- a n d (Z)-1-[4-[2-(Dim et h yla m in o)et h oxy]p h en yl]-
1-(4-h yd r oxyp h en yl)-1-h exen e (12). The procedure used
4-hydroxypentanophenone and yielded a mixture of isomers

1110 J ournal of Medicinal Chemistry, 1997, Vol. 40, No. 7
Gilbert et al.
Bis[4-[2-(d iisop r op yla m in o)eth oxy]p h en yl]cycloh exyl-
id en em eth a n e (8). The procedure was analogous to that
used to prepare 7. Bis(4-hydroxyphenyl)-cyclohexylidene-
methane was prepared by the action of pyridine hydrochloride
on bis(4-methoxyphenyl)cyclohexylidenemethane previously
obtained by condensation of cyclohexanecarboxylic acid ethyl
ester with p-anisylmagnesium bromide (yield 87%; mp 240 °C).
The title compound was prepared by condensation of the
dihydroxy derivative with 2-(diisopropylamino)ethyl chloride.
Upon distillation of the residual oil, fraction 248-254 °C (0.01
mmHg) was isolated. This oil (77%) crystallized slowly: mp
which time DNA content was assayed by the method of DAPI.33
The cytostatic/cytotoxic effects on cell proliferation were
expressed as the compound concentration inhibiting growth
by 50% (IC50) or decreasing cell content to below seeding level.
MVLN Cell Lu cifer a se Exp r ession . MVLN cells in
phenol red-free DMEM medium containing 2% dextran-coated
charcoal-treated fetal calf serum were plated in 24-well tissue
culture cluster plates (80 000 cells/well). After 24 or 48 h, 0.5%
4-hydroxytamoxifen was added for 12-24 h in order to
decrease the basal level of luciferase expression. The medium
was then removed and replaced by a phenol red and serum-
free medium containing different concentrations of the test
compound either alone (<1% ethanol) to measure estrogenic
activity or together with 0.1 nM estradiol to measure anties-
trogenic activity. After incubation for 14-16 h, the cells were
rinsed, harvested and lyzed with 500 µL of luminescence buffer
(15 mM K2PO4, 8 mM MgCl2, 2 mM ATP, 1% Triton X-100,
pH 7.4). The lyzed cells were stirred for 30 min, and the
luminescence emitted during 15 s was measured in a Wallac-
LKB 1251 luminometer on a 200 µL aliquot to which 100 µL
luciferin and coenzyme A (final concentrations 0.2 µM) had
been added.
Cor r esp on d en ce F a ctor An a lysis (CF A). CFA is used
to compare specificity profiles and can be likened to a unified
2-fold principal components analysis performed on the com-
pounds, on the one hand, and on the tests, on the other. The
method is described in detail in several papers.23-27 The
present study concerns a dual CFA,26,34 i.e., each column of
the 22 × 6 data table (22 compounds (rows) by 6 tests
(columns)) was split into two subcolumns giving the experi-
mental values (see Table 2) and antivalues calculated by
subtracting each experimental value from the maximal value
recorded in the column. This procedure introduces the notion
of amplitude of response into the specificity comparison. The
data were normalized and then converted into a ø2 distance
semimatrix which was analyzed by a CFA program adapted
for BASIC (Microsoft Language) from FORTRAN ANACOR
software and written by one of us (J .C.D.). Commercial CFA
programs are also available (e.g. SAS/STAT User’s Guide, Vol
1; ANOVA-FREQ, Version 6, 1990, SAS Institute Inc., Cary,
NC).
1
75 °C; H NMR (CDCl3) δ 0.94-0.98 (2s, 24H, CH3), 1.51 (br
s, 6H, (CH2)3 (3,4,5 of cyclohexyl ring)), 2.16 (s, 4H, CdC-
(CH2)2), 2.72 (2t, 4H, NCH2), 2.95 (sept, 4H, NCH(CH3)2), 3.79
(t, 4H, CH2O), 6.62-6.95 (m, 8 ArH). Anal. (C35H54N2O2) C,
H, N.
X-r a y Cr ysta llogr a p h y. Suitable crystals of 1E (0.75 ×
0.65 × 0.25 cm) were grown from benzene. Data were collected
on a Philips PW1100 diffractometer equipped with Mo KR
radiation and a graphite monochromator. The structures were
solved by direct methods and refined in a full matrix with the
program CRYSTALS. All non-hydrogen atoms were refined
anisotropically. A total of 3941 reflections were obtained; of
these, 1694 were observed with I g 3 σ(I). Abbreviated crystal
data are as follows: C25H35NO2, space group P21/a, a ) 25.969-
(15) Å, b ) 13.138(9) Å, c ) 6.676(1) Å, V ) 2274 Å3, Z ) 4,
F000 ) 832 e-, µ ) 0.65 cm-1, density (calc) ) 1.12 g/cm3, 2θ
range 1-25°. No absorption corrections were applied. Final
conventional and weighted R values were 0.061 and 0.060,
respectively.
Stor a ge a n d Isom er isa tion of Com p ou n d s. Stock 10
mM solutions of the above test compounds and of reference
compounds ((Z)-4-hydroxytamoxifen and (Z)-tamoxifen kindly
supplied by Dr. A. E. Wakeling, ICI, Macclesfield, U.K.) in 96%
alcohol were stored at 4 °C in the dark and checked before
use by thin-layer chromatography (CH2Cl2/hexane, 90:10, v/v)
or by high-performance liquid chromatography (HPLC) with
a Waters system equipped with a UV detector set at 280 nm.
The isomers were separated on a 5-mm radial pack silica gel
column using chloroform/methanol/ethylamine/water (98:1.5:
0.35:0.15, v/v, for 1E/Z and 2E/Z and 98.5:1.0:0.35:0.15, v/v,
for the remaining isomers).
Isomerization rates under in vitro assay conditions were
studied as follows: 100 µM of each isomer in phosphate-
buffered saline, pH 7.4, were left to stand for 72 h at 37 °C.
Samples (4 mL) were extracted twice with chloroform. Ex-
tracts were dried under a gentle stream of nitrogen, dissolved
in the corresponding HPLC solvent system, and then analyzed
by HPLC.
Cell Cu ltu r es. MCF7 human breast cancer cells were
routinely cultured at 37 °C in an atmosphere of 95% humidity
and 5% CO2 in DMEM medium (Gibco-BRL, Cergy Pontoise,
France) and HEPES buffer supplemented with 1 nM insulin
and 5% fetal calf serum. MVLN cells were established by
transfecting MCF7 cells with the pVit-tk-Luc plasmid. In this
plasmid, the -331/-87 fragment of the 5′ flanking region of
Xenopus vitellogenin A2 gene, which contains the estrogen
receptor response element, is inserted upstream of the Herpes
simplex virus thymidine kinase (tk) promoter which drives the
firefly luciferase gene (Luc).29 MVLN cells were routinely
cultured in the same medium as MCF7 cells.
A minimum spanning tree, which is the shortest pathway
linking the molecules into a network with no backtracking nor
loops, was calculated by applying Prim’s algorithm30 to the ø2
distance matrix derived from the split data table as in previous
studies.19,34
Ack n ow led gm en t. The authors thank Dr. Fran-
c¸oise Michel for performing the studies on isomerization
rates.
Su p p or tin g In for m a tion Ava ila ble: Boiling or melting
points and yields of intermediate alkyl acid phenyl esters and
hydroxyphenyl alkyl ketones are available as well as X-ray
crystallography data for 1E : fractional atomic coordinates,
anisotropic thermal parameters, and interatomic bond lengths
and angles (5 pages). Ordering information is given on any
current masthead page.
Rela tive Bin d in g Affin ities (RBAs) for th e Estr ogen
Recep tor (ER). The experiments were performed on im-
mature lamb uterus cytosol. The test compounds were coin-
cubated with 1 nM [6,7-3H]-17â-estradiol for 5 h at 25 °C. The
RBAs for ER (RBA of estradiol ) 100%) were measured by a
competition method as previously described.23
MCF 7 Cell Gr ow th . MCF7 cells were grown for 3-5 days
in phenol red-free DMEM medium containing 3% dextran-
coated charcoal-treated fetal calf serum in the presence or
absence of 10-6 M estradiol. Cells were plated in 24-well tissue
culture cluster plates at a density of 25 000 to 30 000 cells per
well. The test compounds were added at concentrations of
10-6-10-5 M (ethanol concentration <0.1%). Triplicate wells
for each test-compound concentration were incubated for 9-12
days (the culture medium was replaced every 3 days), after
Refer en ces
(1) Darbre, P. D.; Daly, R. J . Transition of human breast cancer
cells from an oestrogen responsive to unresponsive state. J .
Steroid Biochem. Mol. Biol. 1990, 37, 753-763.
(2) Murphy, L. C. Estrogen receptor variants in human breast
cancer. Mol. Cell. Endocrinol. 1990, 74, C83-C86.
(3) Horwitz, K. B. Mechanisms of hormone resistance in breast
cancer. Breast Cancer Res. Treat. 1993, 26, 119-130.
(4) Wolf, D. M.; J ordan, V. C. Drug resistance to tamoxifen during
breast cancer therapy. Breast Cancer Res. Treat. 1993, 27, 27-
40.
(5) Badia, E.; Duchesne, M. J .; Fournier-Bidoz, S.; Simar-Blanchet,
A. E.; Terouanne, B.; Nicolas, J . C.; Pons, M. Hydroxytamoxifen
induces a rapid and irreversible inactivation of an estrogenic
response in an MCF-7-derived cell line. Cancer Res. 1994, 54,
5860-5866.

Cytotoxicity and Antiestrogenicity of Diphenylethylenes
J ournal of Medicinal Chemistry, 1997, Vol. 40, No. 7 1111
(6) van Agthoven, T.; van Agthoven, T. L. A.; Dekker, A.; Foekens,
J . A.; Dorssers, L. C. J . Induction of estrogen independence of
ZR-75-1 human breast cancer cells by epigenetic alterations. Mol.
Endocrinol. 1994, 8, 1474-1483.
(7) Bardon, S.; Vignon, F.; Derocq, D.; Rochefort, H. The antipro-
liferative effect of tamoxifen in breast cancer cells: mediation
by the estrogen receptor. Mol. Cell. Endocrinol. 1984, 35 , 89-
96.
(8) Taylor, C. M.; Blanchard, B.; Zava, D. T. Estrogen receptor-
mediated and cytotoxic effects of the antiestrogens tamoxifen
and 4-hydroxytamoxifen. Cancer Res. 1984, 44, 1409-1414.
(9) Sutherland, R. L.; Watts, C. K. W.; Hall, R. E.; Ruenitz, P. C.
Mechanisms of growth inhibition by nonsteroidal antioestrogens
in human breast cancer cells. J . Steroid Biochem. 1987, 27, 891-
897.
(10) Clarke, R.; van den Berg, H. W.; Nelson, J .; Murphy R. F.
Pharmacological and suprapharmacological concentrations of
both 17â-oestradiol and tamoxifen reduce the membrane fluidity
of MCF-7 and MDA-MB-436 human breast cancer cells. Bio-
chem. Soc. Trans. 1987, 15, 243-244.
plicated in estrogen receptor binding, protein kinase C activity,
and MCF7 cell proliferation. J . Med. Chem. 1992, 35, 573-583.
(20) Gilbert, J .; Dore´, J . C.; Bignon, E.; Pons, M.; Ojasoo, T. Study of
the effects of basic di- and tri-phenyl derivatives on malignant
cell proliferation. An example of the application of correspon-
dence factor analysis to structure-activity relationships. Quant.
Struct. Act. Rel. 1994, 13, 262-274.
(21) Durani, S.; Anand, N. A possible basis for structure-function re-
lationship of estrogens. Int. J . Quant. Chem. 1981, 20, 71-83.
(22) Garg, S.; Bindal R. D.; Durani, S.; Kapil, R. S. Structure-activity
relationship of estrogens: a study involving cyclofenil as the
model compound. J . Steroid Biochem. 1983, 18, 89-95.
(23) Bignon, E.; Pons, M.; Crastes de Paulet, A.; Dore´, J . C.; Gilbert,
J .; Abecassis, J .; Miquel, J . F.; Ojasoo, T.; Raynaud, J . P. Effect
of triphenylacrylonitrile derivatives on estradiol receptor binding
and on human breast cancer cell growth. J . Med. Chem. 1989,
32, 2092-2103.
(24) Dore´, J . C.; Gilbert, J .; Ojasoo, T.; Raynaud, J . P. Correspondence
analysis applied to steroid receptor binding. J . Med. Chem. 1986,
29, 54-60.
(11) van der Koedijk, C. D. M. A.; Vis van Heemst, C.; Elsendoorn,
G. M.; Thijssen, J . H. H.; Blankenstein, M. A. Comparative
affinity of steroidal and non-steroidal antiestrogens, cholesterol
derivatives and compounds with a dialkylamino side chain for
the rat liver antiestrogen binding site. Biochem. Pharmacol.
1992, 43, 2511-2518.
(12) Rowlands, M. G.; Parr, I. B.; McCague, R.; J arman, M.; Goddard,
P. Variation of the inhibition of calmodulin dependent cAMP
phosphodiesterase amongst analogues of tamoxifen; correlations
with cytotoxicity. Biochem. Pharmacol. 1990, 40, 283-289.
(13) O’Brian, C. A.; Liskamp, R. M.; Solomon, D. H.; Weinstein I. B.
Triphenyl ethylenes: a new class of protein kinase C inhibitors.
J . Natl. Cancer Inst. 1986, 76, 1243-1246.
(14) Bignon, E.; Pons, M.; Gilbert, J .; Nishizuka, Y. Multiple mech-
anisms of protein kinase C inhibition by triphenylacrylonitrile
antiestrogens. FEBS Lett. 1990, 271, 54-58.
(25) Ojasoo, T.; Dore´, J . C.; Gilbert, J .; Raynaud, J . P. Binding of
steroids to the progestin and glucocorticoid receptors analyzed
by correspondence analysis. J . Med. Chem. 1988, 31, 1160-1169.
(26) Dore´, J . C.; Ojasoo, T. Molecular taxonomy by Correspondence
Factorial Analysis (CFA) In Advanced Computer-Assisted Tech-
niques in Drug discovery; van de Waterbeemd, H., Ed.; VCH:
Weinheim, 1994; pp 190-227.
(27) Ojasoo, T.; Raynaud, J . P.; Dore´, J . C. Correspondence factor
analysis of steroid libraries. Steroids 1995, 60, 458-469.
(28) Pons, M.; Gagne, D.; Nicolas, J . C.; Methali, M. A new cellular
model of response to estrogens: a bioluminescent test to charac-
terize (anti)estrogen molecules. Biotechniques 1991, 9, 450-459.
(29) Demirpence, E.; Balaguer, P.; Trousse, F.; Badia, E.; Gagne, D.;
Pons, M. MVLN cells: a bioluminescent MCF-7-derived cell line
to study the modulation of estrogenic activity. J . Steroid Bio-
chem. Mol. Biol. 1993, 46, 355-364.
(15) Ojasoo, T.; Bignon, E.; Crastes de Paulet, A.; Dore´, J . C.; Gilbert,
J .; Miquel, J . F.; Pons, M.; Raynaud, J . P. Relative involvement
of the estrogen receptor and protein kinase C in the action of a
population of triphenylethylenes on MCF7 cell proliferation as
revealed by correspondence factorial analysis (CFA). J . Steroid
Biochem. Mol. Biol. 1993, 44, 239-250.
(16) Astruc, M. E.; Chabert, C.; Bali, P.; Gagne, D.; Pons, M.
Prolonged treatment of breast cancer cells with antiestrogens
increases the AP1-mediated response: involvement of the
estrogen receptor. Endocrinology 1995, 136, 824-832.
(17) Early Breast Cancer Trialists’ Collaborative Group. Systemic
treatment of early breast cancer by hormonal, cytotoxic or
immune therapy. Lancet 1992, 339, 1-15, 71-85.
(18) Lundgren, S. Prospects in endocrine treatment of breast cancer.
Cancer J . 1994, 7, 220-227.
(19) Dore´, J . C.; Gilbert, J .; Bignon, E.; Crastes de Paulet, A.; Ojasoo,
T.; Pons, M.; Raynaud, J . P.; Miquel, J . F. Multivariate analysis
by the minimum spanning tree method of the structural deter-
minants of diphenylethylenes and triphenylacrylonitriles im-
(30) Prim, R. C. Shortest connection networks and some generali-
sations. Bell. Syst. Technol. J . 1957, 36, 1389-1401.
(31) Boissier, J . R.; Ratouis, R.; Dumont, C.; Lesbros, J .; Lavaux J .
P. New hypercholesterenic tertiary alcohols. Arzeim. Forsch.
1967, 17, 1193-1196.
(32) Boissier, J . R.; Ratouis, R. 4-(2-diethylaminoethoxy)phenyl
ketones, Pharmaceutical intermediates : French patent no.
1,447,252, J uly 29, 1966.
(33) Brunck, C. F.; J ones, K. C.; J ames, T. W. Assay for nanogram
quantities of DNA in cellular homogenates. Anal. Biochem. 1979,
92, 497-500.
(34) Bignon, E.; Pons, M.; Dore´, J . C.; Gilbert, J .; Ojasoo, T.; Miquel,
J . F.; Raynaud, J . P.; Crastes de Paulet, A. Influence of di- and
tri-phenylethylene estrogen/ antiestrogen structure on the mech-
anisms of protein kinase C inhibition and activation as revealed
by a multivariate analysis. Biochem. Pharmacol. 1991, 42,
1373-1383.
J M950624T