Studies on organometallic selective estrogen receptor modulators. (SERMs) Dual activity in the hydroxy-ferrocifen series

Article abstract of DOI:10.1016/S0022-328X(01)00953-6

Synthesis of 7, a ferrocene derivative of the antiestrogenic drug hydroxytamoxifen bearing a basic chain-O(CH₂)_nN(CH₃)₂ with n=4 is presented, together with both studies of its antiproliferative effect on the hormone-dependent MCF7 cell line (estrogen receptor positive cells) and of its genotoxicity. This molecule is easily prepared via a McMurry coupling reaction. The antiproliferative effect found for 7 at an incubation molarity of 1 μM was very close to that found for the usual reference molecule, namely OH-tamoxifen. In addition to its structural antiestrogenic effect, 7 showed cytotoxic activity probably due to the vectored ferrocene. This genotoxic component was confirmed by a 3D (damaged DNA detection) test, that permits identification and quantification of lesions induced on DNA. Some key interactions of 7 docked into the alpha-estrogen receptor binding site were also shown.

Full text of DOI:10.1016/S0022-328X(01)00953-6

Journal of Organometallic Chemistry 637–639 (2001) 500–506
www.elsevier.com/locate/jorganchem
Studies on organometallic selective estrogen receptor modulators.
(SERMs) Dual activity in the hydroxy-ferrocifen series
Siden Top a, Anne Vessie`res a, Claude Cabestaing a, Ionna Laios b, Guy Leclercq b,
Christian Provot c, Ge´rard Jaouen a,*
a Ecole Nationale Supe´rieure de Chimie de Paris, 11 rue Pierre et Marie Curie, 75231 Paris Cedex 05, France
b Laboratoire J.-C.-Heuson de Cance´rologie Mammaire, Institut Jules Bordet, 1, rue He´ger, Bordet, B-1000 Brussels, Belgium
c S.F.R.I. Berganton, 33127 Saint Jean d’Illac, France
Received 29 January 2001; received in revised form 18 February 2001; accepted 8 April 2001
Abstract
Synthesis of 7, a ferrocene derivative of the antiestrogenic drug hydroxytamoxifen bearing a basic chain-O(CH2)nN(CH3)2 with
n=4 is presented, together with both studies of its antiproliferative effect on the hormone-dependent MCF7 cell line (estrogen
receptor positive cells) and of its genotoxicity. This molecule is easily prepared via a McMurry coupling reaction. The
antiproliferative effect found for 7 at an incubation molarity of 1 mM was very close to that found for the usual reference
molecule, namely OH-tamoxifen. In addition to its structural antiestrogenic effect, 7 showed cytotoxic activity probably due to
the vectored ferrocene. This genotoxic component was confirmed by a 3D (damaged DNA detection) test, that permits
identification and quantification of lesions induced on DNA. Some key interactions of 7 docked into the alpha-estrogen receptor
binding site were also shown. © 2001 Elsevier Science B.V. All rights reserved.
Keywords: Organometallic hormone; Ferrocene; Antiestrogen; Tamoxifen; MCF7; Cytotoxicity
1. Introduction
Some significant recent developments concern the
estrogen receptor and its mode of activity. First, it has
been found that there are actually two forms of the
estrogen receptor, the long-known a form (ERa) and
also a b form (ERb), both implicated in breast cancer,
The personal and social costs of breast cancer are
well documented [1]. To date, other than surgery and
radiotherapy, the therapeutic arsenal consists chiefly of
antiestrogenic selective estrogen receptor modulators
(SERMs), of which tamoxifen 1 remains the primary
example (Scheme 1) [2], as well as inhibitors of the
aromatase, the key enzyme of the biosynthesis of estra-
diol in women, taxol derivatives and specific mono-
clonal antibodies again breast tumors, plus the chemical
cocktails of classic chemotherapy. Despite their proven
utility, the very multiplicity of these approaches attests
to the failure to find a truly successful therapeutic
solution. Research in this area in fact remains very
active and has recently benefited from important ad-
vances in receptorology.
* Corresponding author. Fax: +33-1-4326-0061.
E-mail address: jaouen@ext.jussieu.fr (G. Jaouen).
Scheme 1.
0022-328X/01/$ - see front matter © 2001 Elsevier Science B.V. All rights reserved.
PII: S0022-328X(01)00953-6

S. Top et al. / Journal of Organometallic Chemistry 637–639 (2001) 500–506
501
genital, head, neck and colon cancers) [10–12]. One of
the most spectacular success stories to come from these
compounds is the case of testicular cancer which used
to have a very high mortality rate and is now very
much a curable condition. This breakthrough led to
research on organometallic compounds of various
metals for antitumoral applications [13,14]. Of these,
the cyclopentadienyl complexes of Ti, Fe, Mo, V, Re
have given encouraging results, although no com-
pounds of this type have as yet completed clinical trials.
It appears however that these metallocenes operate
via different mechanisms than cis-platinum [13,14] and
could be used for types of tumors on which the latter is
not effective. The idea of using a hormonal vector to
deliver cis-platinum at the level of the estrogen receptor
has been the subject of several studies [15] but, as with
compound 6, the antitumoral effects on breast cancers
proved disappointing [16]. This led us to adjust the
strategy and explore the potential in this area of various
organometallic groups with possible cytotoxicity, at-
tached to known SERMs so as to potentialize their
effects. We have suggested previously that fixation of a
ferrocenyl group onto an antiestrogen vector of the
OH–tamoxifen 2 type could provide a useful modifica-
tion of this SERM [17,18], but it appears that in this
case the length of the carbon chain significantly per-
turbs its effects. We present here syntheses, biochemi-
cal studies, molecular modeling and a suggested mecha-
nism to support the unusual behavior of hydroxy-
Scheme 2.
Scheme 3.
although to differing degrees [3,4]. Secondly, the struc-
ture of the ligand binding domain (LBD) of these two
receptors attached to various bioligands has been pub-
lished [5–8]. Finally, two activation pathways modulat-
ing the hormonal effect at the DNA level have been
identified, namely the estrogen response element (ERE)
and activator protein 1 (AP1) [4,9]. These discoveries
add a new level of complexity to our understanding of
regulatory phenomena in the body, but for the first
time they also make it possible to think in molecular
terms about this hormone-dependent cancer, and they
present interesting new challenges to the chemist.
It is a fact that the standard SERM, tamoxifen (1), is
only effective on approximately 60% of tumours, those
classified estrogen receptor positive (ER+). Tamoxifen
also causes resistance over the long term and has unde-
sirable side effects, chiefly endometrial or thromboem-
bolic in nature [9]. Although no ideal SERM exists, a
considerable improvement would be a molecule active
against both ER+ and ER− (estrogen receptor nega-
tive) tumors. This can only be achieved by taking a new
therapeutic approach. We present here some of our
results in this area, based on the particular properties of
ferrocene chemistry.
ferrocifen such as
7 with a four-carbon chain
–O–(CH2)4NMe2 (Scheme 3) as compared to its equiv-
alent 8 without basic chain.
Compound 7 and other similar ones that contain a
ferrocene substituent in fact possess the unusual prop-
erty of combining both antiproliferative effects on
ER+ breast cancer cell lines, due to a change in the
conformation of the a receptor, and genotoxic effects
due to the attached organometallic moiety.
2.1. Synthesis of 7 and 8
The McMurry coupling reaction (TiCl4, Zn) is still
the preferred method for synthesis of tamoxifen analogs
[19–21]. It is easy to carry out and results in a mixture
of Z and E isomers. Scheme 4 shows the synthetic
routes to the ferrocene derivatives of tamoxifen.
Propionylferrocene was obtained by acylation of fer-
rocene via a Friedel–Crafts reaction. Stirring the propi-
onylferrocene mixture with hydroxybenzophenone 9 in
the presence of the McMurry reagent gives the dissym-
metric monohydroxylated complex 8 as the major
product in 36% yield. The synthesis of ferrocifen 7
begins with a coupling reaction between propionylfer-
rocene and the ketone 10, obtained previously by addi-
tion of the bromobutyl chain onto dihydroxybenzo-
2. Background to the problem
Inorganic platinum complexes, of which the
archetype is cis-platinum, 5 (Scheme 2), have a well-es-
tablished status as effective antitumoral agents, despite
a relatively narrow therapeutic range (testicular, uro-

502
S. Top et al. / Journal of Organometallic Chemistry 637–639 (2001) 500–506
Scheme 4.
phenone in the presence of sodium ethoxide. This gives
the halogenated compound 11 which then reacts with
dimethylamine chlorhydrate in an autoclave to give
ferrocifen 7 as the final product in 34.4% yield.
2.3. Study of the proliferati6e/antiproliferati6e effect on
MCF7 cells
The proliferative/antiproliferative effect of complexes
7 and 8 and of the ferrocene (Fc) entity alone were
studied on the hormone-dependent MCF7 cell line (es-
trogen receptor positive cells). The results obtained are
shown in Fig. 1. They are expressed as a percentage of
DNA compared to control (CTR; cells incubated in the
presence of the vehicle). For estradiol (E2), used as a
standard for the proliferative effect, a value of 186%
was found, while OH–tamoxifen 2, standard for the
antiproliferative effect, gave a value of 35%. For ferro-
cifen 7, at an incubation molarity of 1 mM, an antipro-
liferative effect very close to that of 2 is found. As well,
this effect is only partially inhibited by addition of
estradiol into the medium showing that the observed
effect can be divided into two components, the first
Compounds 7 and 8 were obtained as a mixture of Z
and E isomers. Thin-layer chromatography cannot be
used to separate the two isomers. However separation
was possible in the case of 3, ferrocifen with a two-car-
bon side-chain [21] and the corresponding OH–ferro-
cifen 4 [18]. Isomerization leading to interconversion of
the Z and E isomers is a well-known phenomenon in
the stilbene series, for example in the case of diethyl-
stilbestrol (DES) [22]. An isomerization study was per-
formed on compound 4 at room temperature [23]. We
found that 4-E is indeed subject to isomerization, the
speed of isomerization depending on the nature of the
solvent. In chloroform the transformation is rapid, with
50% conversion after one hour in solution, while it
takes a day to reach this value in acetone, and in
ethanol only 30% isomerization has occurred after 24 h.
In any case, this phenomenon of isomerization avoids
the lengthy process of separating the two isomers.
2.2. Biochemical data
2.2.1. Determination of RBA 6alue and lipophilicity
The study of the biochemical properties of the com-
pounds involves first of all the determination of their
relative binding affinity for estradiol receptor a. The
values found for 7 and 8 are 12 and 5%, respectively,
which is an acceptable level of recognition although
lower than the value found for 2 (38.5%). Lipophilicity
measured by the octanol/water partition coefficient
gave values for 7 of 3.95 for the Z isomer and 4.18 for
E. These are higher than those found for OH–tamox-
ifen 2 (3.17 and 3.36), which confirms that insertion of
an organometallic moiety increases the lipophilicity of
the hormone thus permitting better penetration of these
complexes into cells.
Fig. 1. Effect of E2 (estradiol; 1×10−9 M), 2 (OH–tamoxifen), Fc
(ferrocene) and ferrocene derivatives 7 and 8 on the proliferation of
MCF7 cells (estrogen receptor-positive cells). The results are ex-
pressed as the percentage of DNA in the sample versus the DNA
value of the control (CTR).

S. Top et al. / Journal of Organometallic Chemistry 637–639 (2001) 500–506
503
complete inhibition of the effect by addition of E2).
Moreover, incubation in the presence of 8, 7 without
the antiestrogenic chain, shows that at an incubation
molarity of 0.1 mM, 8 behaves as expected as an
estrogen (186%) while at a molarity of 1 mM it becomes
cytotoxic for the cells (65%), although still less so than
the ferrocifen 7.
anti-hormonal, due to the presence of the dimethyl-
amino chain, and the other cytotoxic, probably caused
by the ferrocene entity whose cytotoxic effect on certain
cell lines has already been reported [14]. In the present
case, however, the effect is not entirely due to the
ferrocene entity, since incubation of this entity has no
effect on MCF7 cells. The addition of a vector is thus
required. It should be noted as well that at a molarity
of 0.1 mM only an antagonist effect is seen in 7 (with
2.4. 3D assay (damaged DNA detection)
The increased cytotoxic effect of ferrocene derivatives
seen in MCF7 cells beginning at micromolar concentra-
tions is entirely corroborated by the 3D assay, patented
recently and published by the company SFRI [24,25].
This test permits identification and quantification of
lesions induced on DNA (oxidative lesions, formation
of adducts) as well as the effect of molecules on the
modulation of the reparative activity of the DNA from
the cell extract used in the test. Quantitative evidence of
the number of lesions repaired is obtained by chemilu-
minescence. Using this test, we compared tamoxifen 1
and their ferrocene equivalents 3 and 4 (Fig. 2). Values
are shown as percentages of lesions compared to con-
trol (100%). For a given molecule, a value above 100%,
therefore, shows that lesions were induced on DNA.
The results obtained show that ferrocifen 3 and its
corresponding hydroxylated form 4 cause a similar rate
of lesions on DNA. The first significant effects are
observed at a molarity of 1 mM, and they become very
pronounced at a molarity of 10 mM. In the same
conditions, tamoxifen 1 has no effect. Moreover, none
of the compounds tested, 1, 3, 4, inhibited the repara-
tive capacity of the lesions (data not shown). Finally,
UV-light irradiation of the compounds under 3D test
conditions had no effect. These results corroborate
those obtained by incubation of MCF7 cells in the
presence of ferrocifens: a cytotoxic effect that becomes
apparent at an incubation molarity of 1 mM, and death
of the cells at an incubation molarity of 10 mM.
Fig. 2. Dose effect of 1, 3 and 4 on 3D assay (DNA, damaged,
detection). Target DNA is incubated for 30 min at 30 °C in the
presence of various concentration of the molecules to be tested. The
results are expressed as the percentage of repair of the sample under
study vs. the value of control. For details see text and Ref. [25].
3. Discussion
The results obtained here with ferrocene SERMs that
are tamoxifen analogs show a similarity in behavior
between these two species in their structural antagonist
effect against ERa, and at the same time they also
reveal an intrinsic cytotoxic component due to the
organometallic entity.
The X-ray crystal structure of ERa contained in the
MCF7 cell lines, with hydroxytamoxifen 2 attached at
the ligand binding domain, has been published [7].
Taking this work as our basis, we used MOLVIEW [26]
to model the antagonist site and introduced hydroxyfer-
rocifen 7 into the active site. The result is shown in Fig.
3. As with 2, the phenol is still associated with Arg 394
Fig. 3. Some key interactions of OH–ferrocifen 7 docked into the
alpha-estrogen receptor binding site. This study was performed by
using the structural basis of the ER ligand binding domain published
by Shiau et al. [7] and the MOLVIEW program designed by Cense [26].

504
S. Top et al. / Journal of Organometallic Chemistry 637–639 (2001) 500–506
,
,
(distance 2.54 A) and Glu 353 (2.73 A) residues, and
the ferrocene entity that replaces the benzene ring b
goes into place without disturbing the protein at the
level of histidine 524. The basic chain is responsible in
these compounds for the primary antagonistic effect, as
it changes the position of helix 12 relative to its agonist
conformation, and thus inhibits the fixation of the
coactivators at this level (residues 536–544). The dis-
tance between the Asp 351 and –N(CH2)3 residues and
4. Conclusions
The dual activity of the ferrocene complexes of ta-
moxifen on ERa-containing MCF7 cells points to inter-
esting new directions in the search for SERMs that will
be more effective against breast cancer. It has been
hypothesized that the spread of breast cancers that
evade the effects of tamoxifen often involves ERb and
the AP1 regulation pathway [3,9]. Tamoxifen then be-
haves as an agonist and becomes ineffective. But ERb is
also implicated in cell oxido-reduction [29]. This leads
one to consider that the phenomenon described here
might be particularly useful in the control of antiprolif-
erative effects with ERb-containing breast cancer cells.
This work is underway.
,
the chain (5.35 A) appears too great to be compatible
with the strong stabilizing association which has been
observed in the model in the presence of raloxifen [5],
but not with 2 [7].
On the other hand, the addition of a ferrocene sub-
stituent introduces an oxidizing/reducing genotoxic as-
pect that is absent in the organic equivalents. This
effect can be seen at micromolar concentrations. The
study of the antiproliferative effects on the MCF7 cell
line and the observation of the genotoxic effect of
ferrocene complexes as seen in the 3D test are in
agreement on this subject. Both tests reveal this unusual
behavior which is due to the vectored ferrocene.
It has been reported widely in the literature that the
cytotoxic effect of ferrocene complexes is associated
with their oxidation to ferricinium-type radical ions [14]
and we have checked by cyclic voltammetry that the
potential standard of 7 was very closed to that of
ferrocene alone (0.474 and 0.460 V, respectively). Until
recently, the reason for their ability to cleave DNA was
not clear and could be linked either to a direct associa-
tion or to oxidative damage. A study by Tamura has
shown that this cytotoxic activity is due to the genera-
tion of OH. [27]. Recent work by Osella et al. also
militates in favor of there being no direct ferricinium–
DNA interaction, but instead the generation of active
oxygenated species [28]. The results of the 3D test bear
out this analysis. It can in fact be seen that, unlike
tamoxifen 1, the corresponding ferrocene complexes 3
and 4 show genotoxic activity in this test. Moreover,
this activity occurs by production of lesions on DNA
but does not seem to interfere with the reparative
phenomena, since repair is not inhibited (data not
shown) as could have been the case with an adduct.
This effect can be compared to the Fenton reaction:
5. Experimental
5.1. General remarks
The synthesis of all the compounds was performed
under an argon atmosphere, using the Schlenk line
technique and Schlenk flasks. Anhydrous THF and
anhydrous diethyl ether were obtained by distillation
from sodium/benzophenone. 4-Hydroxybenzophenone
(9) and 4,4-dihydroxybenzophenone were purchased
from Aldrich. Ferrocenyl ethyl ketone was prepared
according the procedure described in Ref. [32]. TLC
chromatography was performed on silica gel 60 GF254.
1H- and 13C-NMR spectra were recorded on 200 MHz
and 250 MHz Bruker spectrometers. Mass spectrometry
was performed with a Nermag R 10-10C spectrometer.
Melting points were measured with a Kofler device.
Elemental analysis was performed by the regional mi-
croanalysis service of Universite´ Pierre et Marie Curie.
5.1.1. Synthesis of 1-(4-hydroxyphenyl)-1-(phenyl)-2-
ferrocenyl-but-1-ene (8(Z+E))
TiCl4 (6.58 ml, 60 mmol) was added dropwise to a
suspension of zinc powder (7.80 g, 120 mmol) in THF
(90 ml) at −10 °C. The dark grey mixture obtained
was heated at reflux for 1.5 h and then allowed to cool
to room temperature (r.t.). A solution of THF (10 ml)
containing 4-hydroxybenzophenone (9) (3.96 g, 20
mmol) and ferrocenyl ethyl ketone (4.84 g, 20 mmol)
were added dropwise to the first solution and then the
resulting mixture was heated for 2 h. After cooling to
r.t., the mixture was hydrolyzed with 10% K2CO3 solu-
tion. After ether extraction and solvent removal, the
crude product was chromatographed on a silica gel
column with ethyl ether/pentane 1:2 as eluent. 2.93 g of
1-(4-hydroxyphenyl)-1-(phenyl)-2-ferrocenyl-but-1-ene
(8(Z+E)) were isolated, orange solid, 36% (m.p.
Fe2+ +O2“Fe3+ +O2
−
’
Fe2+ +O2 − +2H+“Fe3+ +H2O2
’
Fe2+ +H2O2“Fe3+ +OH− +OH
’
It is known that the O−2 radical shows very low
reactivity for DNA, but that the OH radical is highly
reactive, causes various types of lesions and is thus very
genotoxic.
’
’
1
161 °C). H-NMR (200 MHz, DMSO-d6) l 9.33 (s, 1H
OH), 7.33–6.63 (m, 9H, C6H5+ C6H4), 4.12, and 4.11

S. Top et al. / Journal of Organometallic Chemistry 637–639 (2001) 500–506
505
(s, s, 5H, Cp),), 4.09, 4.05, 3.82, 3.76 (t, t, t, t, 4H,
5.2. Determination of the RBA of 7 and 8 for the
estrogen receptor alpha (ERh)
C5H4), 2.50 (m, 2H, CH2), 1.00 and 0.99 (t, t, 3H,
CH3). Anal. Calc. for C26H24OFe: C, 76.48; H, 5.92.
Found: C, 76.39; H, 5.90%.
Aliquots (200 ml) of sheep uterine cytosol prepared as
described in Ref. [30] were incubated for 3 h at 0 °C
with [6,7-3H]-estradiol (2×10−9 M, specific activity
1.96 TBq mmol−1) in the presence of nine concentra-
tions of the hormones to be tested. At the end of the
incubation period, the free and bound fractions of the
tracer were separated by protamine sulfate precipita-
tion. The percentage reduction in binding of [3H]-estra-
diol (Y) was calculated using the logit transformation
of Y (logitY: ln[y/1−Y] versus the log of the mass of
the competing steroid. The concentration of unlabeled
steroid required to displace 50% of the bound [3H]-
estradiol was calculated for each steroid tested, and the
results expressed as RBA. The RBA value of estradiol
is by definition equal to 100%.
5.1.2. 1-[4-(4-Dimethylaminobutoxy)phenyl]-1-(4-
hydroxyphenyl)-2-ferrocenyl-but-1-ene (7(Z+E))
Dihydroxybenzophenone (2.14 g, 10 mmol) was
added to a solution of sodium ethanolate prepared by
treating sodium (0.230 g, 10 mmol) with ethanol (15
ml). After stirring at reflux for 1 h, 1,4-dibromobutane
(10.80 g, 50 mmol) was added. After 1 h of reflux, the
solution was left to cool to r.t. and hydrolyzed with 100
ml water. The product was extracted with
dichloromethane. The organic phase was washed with
water, dried over magnesium sulfate, filtered and the
solvent evaporated. The crude product was chro-
matographed on silica gel plates with ethyl ether/pen-
tane 2:1 as eluent. 4,4%-di(4-Bromobutoxy)benzo-
phenone compound was first isolated as a white solid,
5.3. Test on MCF7 cells
1
0.133 g (27% yield). NMR H (200 MHz, acetone-d6) l
7.74 (d, 4H, C6H4–O); 7.06 (d, 4H, C6H4–O); 4.17 (t,
4H, OCH2); 3.61 (t, 4H, CH2Br); 2.03 (m, 8H, CH2–
CH2). m.p. 111 °C. The second compound was isolated
as a yellow oil which solidified in pentane and identified
as 10, 1.005 g, 28% yield. NMR 1H (200 MHz, acetone-
d6) l 7.74 (d, 2H, C6H4–O); 7.68 (d, 2H, C6H4–O); 7.05
(d, 2H, C6H4–O); 6.96 (d, 2H, C6H4–O); 4.16 (t,
2H,OCH2); 3.61 (t, 2H, CH2Br); 2.03 (m, 4H,
CH2CH2). M.p. 98 °C, white crystals.
5.3.1. Culture materials
Earle’s based minimal essential medium (MEM), fe-
tal bovine serum (FBS),
tamicin, streptomycin were obtained from Gibco
(Ghent, Belgium), plastic culture materials from Falcon
(Ghent, Belgium).
L-glutamine, penicillin, gen-
5.3.2. Culture conditions
The coupling reaction between ferrocenyl ethyl ke-
tone and 10 is similar to that of ferrocenyl ethyl ketone
and 9. TiCl4 (0.346 g, 1.82 mmol), zinc powder (0.195 g,
2.98 mmol), 10 (0.150 g, 0.5 mmol), ferrocenyl ethyl
ketone (0.121 g, 0.5 mmol). After a work up, the crude
product, 0.374 g, was chromatographed on silica gel
plates with ethyl ether/pentane 1:1 as eluent to yield 11
(0.110 g, 46%) as an orange solid, m.p. 57 °C. NMR
1H (200 MHz, CDCl3) 6.85 (m, 8H, C6H4–O); 4.06 (s,
5H, C5H5); 3.98 (t, 2H,OCH2); 3.94 (d, 1H, C5H4); 3.90
(d, 1H, C5H4); 3.46 (t, 2H, C5H4); 3.46 (t, 2H, CH2Br);
2.58 (q, 2H, CH2CH3); 1.01 (t, 3H, CH2CH3).
MCF7 cells were from the Michigan Cancer Founda-
tions (Detroit). Cells are maintained in monolayer cul-
ture in Dulbecco-MEM added with 10% thermically
inactivated FBS,
L
-glutamine (0.6 mg ml−1) and a
cocktail of antibiotics (gentamicin 40 mg ml−1, penicilin
100 U ml−1, streptomycin 100 mg ml−1). The growth of
the cells was assessed by measuring the DNA content
of treated and untreated (control) cells after 120 h of
culture [31].
Acknowledgements
Subsequently, 11 (0.150 g, 0.27 mmol), HNMe2·HCl,
(0.170 g), NEt3 (0.2 ml), and ethanol (25 ml) were
placed in an autoclave. The mixture was heated at
110 °C for 6 h, the solvent evaporated and the crude
product obtained chromatographed on Silicagel plates
with acetone/NEt3: 20/1 as eluent to yield 7 (0.047 mg;
We wish to thank A. Cordaville for technical assis-
tance, B. Limoges for performing the cyclic voltamme-
try of the compounds and Barbara McGlinchey for her
assistance in translating the manuscript.
1
34.4%) as an orange solid, m.p. 59 °C. NMR H (200
MHz, DMSO) l 9.32 (s, 1H, OH); 7.10–6.61 (m, 8H,
C6H4–O); 4.10 (s, 5H, C5H5); 4.08 (m, 2H, C5H4); 3.81
(m, 2H, C5H4); 3.93 (dd, 2H, OCH2); 2.37 (m, 2H,
–CH2NMe2); 2.22 (s, 6H, NMe2); 1.60 (m, 4H,
–CH2CH2–); 0.98(t, 3H, CH2CH3). Mass spectrum (EI,
70 eV) m/z: 523 [M+], 100 [(CH2)4NMe+2 ]. Anal. Calc.
for C32H37NO2Fe.H2O: C, 70.98; H, 7.26; N, 2.58.
Found: C, 70.23; H, 7.12; N, 2.86%.
References
[1] E.J. Feuer, L.M. Wun, C.C. Boring, W.D. Flanders, M.J. Tim-
mel, T. Tong, J. Natl. Cancer Inst. 85 (1993) 892.
[2] J.I. Mac Cregor, V.C. Jordan, Pharmacol. Rev. 50 (1998) 151.
[3] J. Gustafsson, J. Endocr. 163 (1999) 379.
[4] K. Paech, P. Webb, G.G.J.M. Kuiper, S. Nilsson, J.-A. Gustafs-
son, P.J. Kushner, T.S. Scanlan, Science 277 (1997) 1568.

506
S. Top et al. / Journal of Organometallic Chemistry 637–639 (2001) 500–506
[18] S. Top, J. Tang, A. Vessie`res, D. Carrez, C. Provot, G. Jaouen,
[5] A.M. Brzozowski, A. Pike, Z. Dauter, R.E. Hubbard, T. Bonn,
O. Engstrom, L. Ohman, G. Greene, J.-A. Gustafsson, M.
Carlquist, Nature 389 (1997) 753.
[6] A.C. Pike, A.M. Brzozowski, R.E. Hubbard, T. Bonn, A.G.
Thorsell, O. Engstrom, J. Ljunggren, J.-A. Gustafsson, M. Car-
lquist, EMBO 18 (1999) 4608.
[7] A.K. Shiau, D. Barstad, P.M. Loria, L. Cheng, P.J. Kushner,
D.A. Agard, G. Greene, Cell 95 (1998) 927.
[8] M. Tanenbaum, Y. Wang, S.P. Williams, P.B. Sigler, Proc. Natl.
Acad. Sci. USA 95 (1998) 5998.
[9] K. Dhingra, Invest. New Drugs 17 (1999) 285.
[10] Z. Guo, P.J. Sadler, Angew. Chem. Int. Ed. Engl. 38 (1999)
1512.
[11] M.E. Heim, in: B.K. Keppler (Ed.), Metal Complexes in Cancer
Chemotherapy, VCH, Weinheim, 1993.
[12] B. Rosenberg, Interdisciplinary Sci. Rev. 3 (1978) 134.
[13] P. Ko¨pf-Maier, in: B.K. Keppler (Ed.), Metal Complexes in
Cancer Chemotherapy, VCH, Weinheim, 1993, p. 259.
[14] P. Ko¨pf-Maier, Eur. J. Clin. Pharmacol. 47 (1994) 1.
[15] E. Von Angerer, in: B.K. Keppler (Ed.), Metal Complexes in
Cancer Chemotherapy, VCH, Weinheim, 1993, p. 73.
[16] S. Top, unpublished results.
Chem. Commun. (1996) 955.
[19] P.L. Coe, C.E. Scriven, J. Chem. Soc. Perkin Trans I 3 (1986)
475.
[20] S. Gauthier, J. Mailhot, F. Labrie, J. Org. Chem. 61 (1996) 3890.
[21] S. Top, B. Dauer, J. Vaisserman, G. Jaouen, J. Organomet.
Chem. 541 (1997) 355.
[22] V.W. Winkler, M.A. Nyman, R.S. Egan, Steroids 17 (1971) 197.
[23] J. Tang, PhD thesis, University Paris VI, 1995.
[24] K.S. Kasprak, Cancer Invest. 13 (1995) 411.
[25] B. Salles, C. Provot, P. Calson, I. Hennebelle, I. Gosset, G.J.
Fournie´, Anal. Biochem. 232 (1995) 37.
[26] J.M. Cense, Phys. Theor. Chem. 71 (1990) 763.
[27] H. Tamura, M. Miwa, Chem. Lett. (1997) 1177.
[28] D. Osella, M. Ferrali, P. Zanello, F. Lashie, M. Fontani, C.
Nervi, G. Cavigiolio, Inorg. Chim. Acta 306 (2000) 42.
[29] M.M. Montano, A.K. Jaiswal, B.S. Katzenellenbogen, J. Biol.
Chem. 273 (1998) 25443.
[30] A. Vessie`res, S. Top, A.A. Ismail, I.S. Butler, M. Lou¨er, G.
Jaouen, Biochemistry 27 (1988) 6659.
[31] G. Leclercq, N. Devleeschouwer, J.C. Heuson, J. Steroid
Biochem. 19 (1983) 75.
[17] G. Jaouen, S. Top, A. Vessie`res, G. Leclercq, J. Quivy, L. Jin, A.
Croisy, C. R. Acad. Sci. Paris Ser. IIc 3 (2000) 89.
[32] P.J. Graham, R.V. Lindsey, G.W. Parshall, M.L. Peterson,
G.M. Whitman, J. Am. Chem. Soc. 79 (1957) 3416.