Proliferative and anti-proliferative effects of titanium- and iron-based metallocene anti-cancer drugs

Article abstract of DOI:10.1016/j.jorganchem.2008.11.071

In previous work we have found that Cp₂TiCl₂ and its corresponding derivative of tamoxifen, Titanocene tamoxifen, show an unexpected proliferative effect on hormone dependent breast cancer cells MCF-7. In order to check if this behav

Full text of DOI:10.1016/j.jorganchem.2008.11.071

Journal of Organometallic Chemistry 694 (2009) 874–879
Contents lists available at ScienceDirect
Journal of Organometallic Chemistry
journal homepage: www.elsevier.com/locate/jorganchem
Proliferative and anti-proliferative effects of titanium- and iron-based metallocene
anti-cancer drugs
a
a
b
b
Anne Vessières a, , Marie-Aude Plamont , Claude Cabestaing , James Claffey , Sandra Dieckmann ,
*
Megan Hogan b, Helge Müller-Bunz b, Katja Strohfeldt b, Matthias Tacke b
a Laboratoire de Chimie et Biochimie des Complexes Moléculaires, École Nationale Supérieure de Chimie de Paris, UMR CNRS 7576, 11, rue Pierre et Marie Curie, 75231 Paris 05, France
b Conway Institute of Biomolecular and Biomedical Research, Centre for Synthesis and Chemical Biology, UCD School of Chemistry and Chemical Biology, University College Dublin,
Belfield, Dublin 4, Ireland
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 24 August 2008
Received in revised form 6 November 2008
Accepted 7 November 2008
Available online 25 December 2008
In previous work we have found that Cp2TiCl2 and its corresponding derivative of tamoxifen, Titanocene
tamoxifen, show an unexpected proliferative effect on hormone dependent breast cancer cells MCF-7. In
order to check if this behavior is a general trend for titanocene derivatives we have tested two other
titanocene derivatives, Titanocene Y and Titanocene K, on this cell line. Interestingly, these two titano-
cene complexes behave in a totally different manner. Titanocene K is highly proliferative on MCF-7 cells
even at low concentrations (0.5
also observed in the presence of bovine serum albumin (BSA). In contrast, Titanocene Y alone has almost
no effect on MCF-7 at a concentration of 10 M, but exhibits a significant dose dependent cytotoxic effect
of up to 50% when incubated with BSA (20–50 g/mL). This confirms the crucial role played by the bind-
lM), thus behave almost similarly to Cp2TiCl2. This proliferative effect is
For publication in the Journal of
Organometallic Chemistry ISBOMC’08
special issue dedicated to Prof. Dr. Gérard
Jaouen.
l
l
ing to serum proteins in the expression of the in vivo, cytotoxicity of the titanocene complexes. From the
hydridolithiation reaction of 6-p-anisylfulvene with LiBEt3H followed by transmetallation with iron
dichloride [bis-[(p-methoxy-benzyl)cyclopentadienyl]iron(II)] (Ferrocene Y) was synthesised. This com-
plex, which was characterised by single crystal X-ray diffraction, contains the robust ferrocenyl unit
instead of Ti associated with easily leaving groups such as chlorine and shows only a modest cytotoxicity
against MCF-7 or MDA-MB-231 cells.
Keywords:
Bioorganometallic chemistry
Anti-cancer drugs
Cisplatin
Titanocene
Ferrocene
Ó 2008 Elsevier B.V. All rights reserved.
Breast cancer
1. Introduction
oped by McGowan and co-workers, which produced water-soluble
compounds showing significant activity against ovarian cancer [6].
Beyond the field of platinum and ruthenium anti-cancer drugs
there is significant unexplored space for further metal-based drugs
targeting cancer. Titanium-based reagents have significant poten-
tial against solid tumors. Budotitane ([cis-diethoxybis(1-phenylbu-
tane-1,3-dionato)titanium (IV)]) (Chart 1) looked very promising
during its preclinical evaluation, but did not go beyond Phase I
clinical trials, although a Cremophor ELÒ based formulation was
found for this rapidly hydrolysing molecule [1]. Much more robust
in this aspect of hydrolysis is titanocene dichloride (Cp2TiCl2),
which shows medium anti-proliferative activity in vitro but prom-
ising results in vivo [2,3]. Titanocene dichloride reached clinical tri-
als, but its efficacy in Phase II clinical trials in patients with
metastatic renal cell carcinoma [4] or metastatic breast cancer
[5] was too low to be pursued.
More recently, novel methods starting from fulvenes and other
precursors [7,8] allow direct access to anti-proliferative titanoc-
enes via reductive dimerisation with titanium dichloride [9–13],
hydridolithiation [14–17] or carbolithiation [18–26] of the fulvene
followed by transmetallation with titanium tetrachloride in the
latter two cases.
Hydridolithiation of 6-anisyl fulvene and subsequent reaction
with TiCl4 led to bis-[(p-methoxybenzyl)cyclopentadienyl] tita-
nium(IV) dichloride (Titanocene Y) [17], which has an IC50 value
of 21 lM when tested on the LLC-PK cell line, which has proven
to be a good mimic of a kidney carcinoma cell line and a reliable
tool for the optimisation of titanocenes against this type of cancer.
In addition, the anti-proliferative activity of Titanocene Y and
other titanocenes has been studied in 36 human tumor cell lines
[27] and against explanted human tumors [28,29]. These in vitro
and ex vivo experiments showed that renal cell cancer is the prime
target for this novel class of titanocenes, but there is significant
activity against ovary, prostate, cervix, lung, colon, and breast can-
cer as well. These results were underlined by first mechanistic
The field gained renewed interest with an elegant synthesis of
ring-substituted cationic titanocene dichloride derivatives devel-
* Corresponding author. Tel.: +33 1 44 27 67 29; fax: +33 1 43 26 00 61.
E-mail address: a-vessieres@enscp.fr (A. Vessières).
0022-328X/$ - see front matter Ó 2008 Elsevier B.V. All rights reserved.
doi:10.1016/j.jorganchem.2008.11.071

A. Vessières et al. / Journal of Organometallic Chemistry 694 (2009) 874–879
875
Ph
O
O
OMe
Me
O
Cl
Cl
Me
Me
Cl
Cl
O
Cl
Cl
H
H
Ti
OEt
OEt
Ti
Ti
O
Ti
O
Me
OMe
Titanocene Y
O
O
Ph
Budotitane
Titanocene K
Titanocene dichloride
OMe
α
MeO
Fe
α'
β
Cl
Cl
O(CH ) NMe
2
Ti
2 3
O(CH ) NMe
2 3
2
Ferrocene Y
Tamoxifen
Titanocene tamoxifen
Chart 1. Structure of the molecules of reference and of the molecules under study.
studies concerning the effect of these titanocenes on apoptosis and
the apoptotic pathway in prostate cancer cells [30]. Furthermore, it
was showed, that titanocene derivatives give a positive immune
response by up-regulating the number of natural killer (NK) cells
in mice [31]. Recently, animal studies reported the successful treat-
ment of mice bearing xenografted Caki-1 and MCF-7 tumors with
Titanocene Y [32,33].
Some time ago Jaouen and co-workers reported the unexpected
strong proliferative effect on hormone dependent breast cancer
cells (MCF-7) of the titanocene derivative of tamoxifen (Titanocene
tamoxifen; Chart 1) [34]. In addition we also found that Cp2TiCl2
alone showed a strong proliferative effect even at the very low con-
centration of 10 nM. The only plausible explanation for this behav-
ior is that these titanocene complexes act as estrogens and this
hypothesis was confirmed on MVLN cells, another hormone depen-
dent cell line in which expression of luciferase is proportional to
the estrogenicity of the tested compounds. The mechanism under-
lying this effect could be the generation in the cell of Ti(IV), which
could be able to mimic the interaction of estradiol with its specific
receptor. More recently, Tacke and co-workers observed a prolifer-
ative effect of Titanocene K on LLC-PK cells but only at high con-
Manipulations of air and moisture sensitive compounds were done
using standard Schlenk techniques, under a nitrogen atmosphere.
NMR spectra were measured on either
a Varian 300 or a
400 MHz spectrometer. Chemical shifts are reported in ppm and
are referenced to TMS. IR spectra were recorded on a Perkin–Elmer
Paragon 1000 FT-IR Spectrometer employing a KBr disk or a liquid
IR cell. UV–Vis spectra were recorded on a Unicam UV4 Spectrom-
eter, while CHN analysis was done with an Exeter Analytical CE-
440 Elemental Analyser. Molecular modeling calculations were
carried out at the PM3 level for the optimisation of the titanocene
using the program package HyperChem [42] and for the albumin-
Titanocene Y conjugate using MOLVIEW [43]. X-ray diffraction
data for compound 3 were collected using a Bruker SMART APEX
CCD area detector diffractometer. A full sphere of reciprocal space
was scanned by phi–omega scans. Pseudo-empirical absorption
correction based on redundant reflections was performed by the
program SADABS [44]. The structures were solved by direct methods
using SHELXS-97 [45] and refined by full-matrix least-squares on F2
for all data using SHELXL-97. All hydrogen atoms were located in
the difference fourier map and allowed to refine freely. Anisotropic
thermal displacement parameters were used for all non-hydrogen
atoms. Further details about the data collection are listed in Table.
1, as well as reliability factors. Suitable crystals of 3 were grown
from a saturated trichloromethane solution by slow evaporation.
centrations (>10 lM) [14]. Therefore, we thought that it could be
of interest to check the behavior of these organometallic titanium
complexes Titanocene K and Titanocene Y on hormone dependent
breast cancer cell line. It has also been demonstrated that following
administration in blood, titanocene derivatives rapidly loose their
chloride groups and bind to serum proteins [35,36]. This encour-
2.2. Synthesis
aged us to study
a
possible stabilisation of the complexes,
Titanocene Y and Titanocene K ({1,2-di(cyclopentadienyl)-1,2-
di[4-(2-methoxyethoxy)phenyl]ethanediyl} titanium dichloride;
with cis–trans ratio at the carbon–carbon bridge of 44:56) were
synthesised as described in [17,14] while 6-p-anisylfulvene was
synthesised according to the literature method [13].
in vitro, via their binding to serum albumin. Finally as some of us
have described the cytotoxicity of a wide range of ferrocenyl deriv-
atives [37–41] we decided to prepare and fully characterise Ferro-
cene Y, the ferrocenyl analog of Titanocene Y, and to study its effect
on breast cancer cell lines.
2.2.1. Synthesis of bis-[(p-methoxybenzyl)cyclopentadienyl]iron(II)
(Ferrocene Y)
2. Experimental
Sixteen microlitres of Superhydride solution (13 mmol, 14.27 g,
1.6 mL in THF) were heated under vacuum for 45 min at 60 °C and
30 min at 90 °C to remove most of the THF. The concentrated re-
agent was re-dissolved in 75 mL of diethyl ether. 2.4 g 6-p-anisyl-
fulvene (13 mmol) were dissolved in 25 mL diethyl ether and was
added to the Superhydride solution via cannula during 5 min and
left to stir for 4 h. The colour of the solution changed during this
time from orange to pale yellow, while the insoluble lithium cyclo-
pentadienide intermediate precipitated from the solution. The
2.1. General conditions
Anhydrous iron dichloride and Super Hydride (LiBEt3H, 1.0 M
solution in THF) were obtained from Aldrich Chemical Company
and used without further purification. Pentane, diethyl ether and
THF were dried over Na and benzophenone (pentane: + di(ethyl-
ene-glycol)ethyl-ether) and they were freshly distilled and col-
lected under an atmosphere of nitrogen prior to use.

876
A. Vessières et al. / Journal of Organometallic Chemistry 694 (2009) 874–879
Table 1
depleted fetal calf serum and 2 mM glutamine and incubated.
Crystallographic refinement data for 3.
The following day (D0), 1 mL of the same medium containing the
compounds to be tested was added to the plates (final volumes
of DMSO: 0.1%; four wells for each condition). After 3 days (D3)
the incubation medium was removed and fresh medium contain-
ing the compounds was added. After 5 days (D5) the total protein
content of the plate was analysed by methylene blue staining as
follows. Cell monolayers were fixed for 1 h in methanol, stained
for 1 h with methylene blue (1 mg/mL) in PBS, and then washed
thoroughly with water. One millilitre of HCl (0.1 M) was then
added and the absorbance of each well was measured at 620 nm
with a Bio-Rad spectrophotometer. The results are expressed as
the percentage of proteins vs. the control.
Empirical formula
Formula weight
Temperature (K)
Crystal system
Space group
Unit cell dimensions
a (Å)
C26H26O2Fe
426.32
293(2)
Monoclinic
P21/n (#14)
11.8726(12)
13.9370(14)
12.5593(13)
93.509(2)
2074.3(4)
4
1.365
0.746
896
b (Å)
c (Å)
b (°)
V (Å3)
Z
Dcalc (Mg/m3)
Absorption coefficient (mmÀ1
F(000)
)
2.4. Molecular modeling studies
Crystal size (mm3)
0.50 Â 0.20 Â 0.15
Theta range for data collection (°)
2.19–26.41
Index ranges
À14 6 h 6 14, À17 6 k 6 17,
À15 6 l 6 15
18110
4255 [0.0266]
1.044
R1 = 0.0342, wR2 = 0.0825
R1 = 0.0427, wR2 = 0.0872
0.367 and À0.213
In order to test the possibility of albumin interacting with
Titanocene Y it was decided to perform preliminary docking stud-
ies by using the crystal structure of human serum albumin (HSA)
co-crystallised with medium-sized fatty acids [46]. The pdb file
of this crystal structure was used as a starting point for MOLVIEW
[43] and a suitably looking palmitate ligand, which is around
1.6 nm in length, was removed to create a possible free site for
Titanocene Y. Titanocene Y was modelled in an elongated confir-
mation exhibiting a length of 1.8 nm using HyperChem [42] at
the PM3 level of theory.
Reflections collected
Independent reflections [Rint
]
Goodness-of-fit (GOF) on F2
Final R indices [I > 2
R indices (all data)
r(I)]
Largest difference peak and hole (e ÅÀ3
)
solution was filtered through a Schlenk glass frit under N2 atmo-
sphere and after washing with small portions of diethyl ether fol-
lowed by pentane, a white solid was dried in vacuo (1.1 g,
5.7 mmol, 44% yield).
2.5. Determination of the relative binding affinity (RBA) for estrogen
receptor alpha (ER
a)
The white solid was transferred into a Schlenk flask and dis-
solved in 25 mL THF and added to a solution of 0.36 g iron dichlo-
ride (2.85 mmol, in 50 mL THF) via cannula during 5 min. The dark
brown solution was refluxed for 16 h. The solution was cooled
down and the solvent was removed in vacuo. The remaining resi-
due was extracted with dichloromethane and filtered through cel-
ite in a Büchner funnel to remove the LiCl. The brown filtrate was
filtered again by gravity filtration and the solvent removed in vacuo
to obtain a yellow solid (1.0 g, 82.1% yield). 1H NMR: (d in ppm,
300 MHz, CDCl3): 7.08 (d, 2H, J = 8.6 Hz, C6H4); 6.80 (d, J = 8.6 Hz,
2H, C6H4); 4.01 (d, 4H, J = 1.8 Hz, C5H4); 3.80 (s, 3H, C6H4–OCH3);
3.59 (s, 2H, C5H4–CH2). 13C NMR (d in ppm, 100 MHz, CDCl3, proton
decoupled): 158.0 (Ci–OCH3); 134.1 (C5H4–CH2–Ci); 129.5, 113.9
(CH, Ph); 88.7 (MeOC6H4–CH2–Ci), 69.5, 68.5 (CH, Cp); 55.5
(OCH3); 35.2 (CH2). Anal. Calc. for FeC26O2H26: C: 73.3; H, 6.2.
Found: C, 73.5; H, 6.4%.
RBA values were measured on ER
sol. Sheep uterine cytosol prepared in buffer A (0.05 M Tris–
HCL, 0.25 M sucrose, 0.1% b-mercaptoethanol, pH 7.4 at 25 °C)
as described previously [40] was used as a source of ER
a from lamb uterine cyto-
a. Ali-
quots (200 mL) of ER were incubated for 3 h at 0 °C with
a
[6,7-3H]-estradiol (2 Â 10À9 M, specific activity 1.62 TBq/mmol,
NEN Life Science, Boston, MA) in the presence of nine concen-
trations of the hormones to be tested. At the end of the incuba-
tion period, the free and bound fractions of the tracer were
separated by protamine sulfate precipitation. The percentage
reduction in binding of [3H]-estradiol (Y) was calculated using
the logit transformation of Y (logitY: ln[Y/1 À Y] vs. the log of
the mass of the competing steroid. The concentration of unla-
beled steroid required for displacing of 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%.
2.3. Cell studies
3. Results and discussion
2.3.1. Materials
Fresh stock solutions (1 Â 10À3 M) of the compounds to be
tested were prepared for each set of experiments in DMSO. Serial
dilutions in DMSO were prepared just prior to use. Dulbecco’s
modified eagle medium (DMEM) was purchased from Gibco BRL,
fetal calf serum from Dutscher, Brumath, France, glutamine, estra-
diol and protamine sulfate and BSA were from Sigma. MCF-7 and
MDA-MB-231 cells were from the Human Tumor Cell Bank.
3.1. Chemical studies
3.1.1. Synthesis
3.1.1.1. Synthesis of Ferrocene Y. The hydridolithiation reaction of 6-
aryl substituted fulvenes followed by transmetallation of titanium
tetrachloride could successfully be transferred to iron(II). Two
equivalents of this lithium cyclopentadienide intermediate was
transmetallated to 1 equiv. of iron dichloride resulting in the for-
mation of 1 equiv. of the required benzyl-substituted Ferrocene Y
in a yield of 82% and 2 equiv. of the by-product of lithium chloride
following a 16 h reflux. The synthesis is shown in Scheme 1.
2.3.2. Culture conditions
Cells were maintained in monolayer in DMEM with phenol red
(Gibco BRL) supplemented with 8–9% fetal calf serum (Gibco BRL)
and glutamine 2 mM (Sigma) at 37 °C in a 5% CO2 air humidified
incubator. For proliferation assays, cells were plated in 24-well
sterile plates at a density of 1.1 Â 104 cells for MDA-MB-231 and
of 3 Â 104 cells for MCF-7 in 1 mL of DMEM medium without phe-
nol red, supplemented with 10% decomplemented and hormone-
3.1.2. Structural discussion
The molecular structure of Ferrocene Y is determined by X-ray
diffraction and shown in Fig. 1.

A. Vessières et al. / Journal of Organometallic Chemistry 694 (2009) 874–879
877
OMe
OMe
H
2
LiBEt H
FeCl
THF
3
2
2
MeO
Fe
Et O
2
H
OMe
H
Li
Ferrocene Y
Scheme 1. Structure and synthesis of Ferrocene Y.
C11
C12
O1
C7
C17
C18
C10
C13
C6
C9
C16
C8
C1
Fe
C26
C24
C25
C14
C19
C15
C5
C2
C3
C20
O2
C23
C22
C4
C21
Fig. 1. Molecular structure of Ferrocene Y (thermal ellipsoids are drawn on the 15% probability level).
Ferrocene Y crystallises in the monoclinic space group P21/n
with one molecule in the asymmetric unit. Details of the refine-
ment are shown in Table 1. The molecules exhibit an overall Z-like
shape in the solid state and form an efficient packing without sol-
vent molecules. The Fe–centroid distance is 1.648 Å and compares
well with distance seen in other ferrocene derivatives, e.g. ferro-
cene itself with 1.66 Å. Also the centroid–Fe–centroid angle of
178.7° is as close to linear as in other ferrocenes. This forces the
molecule into the above-mentioned Z shape. All other structural
features are very similar indeed to the ones in Titanocene Y.
Titanocene K
200
100
0
3.2. Biochemical studies
3.2.1. Determination of relative binding affinity of the compounds for
the estrogen receptor (ERa)
The relative binding affinity (RBA) values found for Titanocene Y
and Titanocene K are very low, 0.001% and 0.006%, respectively,
and even equal to zero for Ferrocene Y. This is not very surprising,
as the chemical structure of these molecules does not mimic at all
that of estradiol, the natural ligand. An RBA value of zero was also
found previously for Cp2TiCl2 [34]. It is important to notice that
these measurements were performed at 4 °C after an incubation
period of 2 h i.e. in experimental conditions totally different to
those of cell cultures (37 °C, 5 days).
Ti K (10 µM) +
BSA (µg/ml)
Ti K
Fig. 2. Effect of 1 nM of estradiol (E2), of various concentrations of Titanocene K and
of 10 M of Titanocene K incubated alone or in the presence of various concen-
tration of bovine serum albumin (BSA; 20–50 g/mL) on the growth of MCF-7 cells
l
l
(hormone-dependent breast cancer cells) after 5 days of culture. Comparison with
the control (C) set at 100%. Mean of at least two independent experiments (six
measurements for each one) mean.
3.2.2. Effect of Titanocene Y and Titanocene K on the growth of MCF-7
cells
The effect of the two titanium complexes was tested on the
growth of hormone-dependent breast cancer cells MCF-7, which
are estrogen receptor positive (ER+) cancer cells (Figs. 2 and 3).
The proliferative estrogenic effect of estradiol on these cells is
known to be mediated by its interaction with this specific receptor.
The results obtained on the growth of MCF-7 cells with various
low concentration, an estrogenic effect similar to that previously
found for Cp2TiCl2 and Titanocene tamoxifen [34], although it is a
little bit less potent. As the RBA value of this compound is very
low we can also suggest that this estrogenic effect is connected
to the generation of in situ Ti(IV). Eventually, its behavior on
MCF-7 cells differs from that observed previously on LLC-PK [14].
The results obtained on MCF-7 with the second complex,
Titanocene Y are shown in Fig. 3. This complex behaves in a totally
different way than Titanocene K. It does not exhibit a significant
concentration of Titanocene K alone (0.5–5
Titanocene K incubated in the presence of various amount of BSA
(20–50 g/mL) are shown in Fig. 2.
lM) and with 10 lM of
l
As expected, estradiol shows a strong proliferative effect on this
cell line, which possesses estrogen receptor and BSA alone has no
effect. On this cell line Titanocene K has a significant proliferative
effect (55–80% of the effect of estradiol) and the addition of BSA
does not induce any significant change. This complex exhibits, at
proliferative effect at 10
presence of BSA (20–50
ative effect. For example a 50% decrease of the growth (i.e. the IC50
value) is obtained after incubation of 10 M of the complex with
50 g/mL of BSA. An IC50 value of 21 M has been found previously
l
M and interestingly incubation in the
l
g/mL) induces a significant anti-prolifer-
l
l
l

878
A. Vessières et al. / Journal of Organometallic Chemistry 694 (2009) 874–879
well-fitting conjugate despite the slight difference in length be-
tween the palmitate, the ligand originally in the structure and
Titanocene Y in its elongated form (1.6 nm vs. 1.8 nm). In addition,
the bulky Cp2TiCl2 unit of Titanocene Y gave no steric problems.
The most striking feature of the interaction is the hydrogen bond
formed by O1 of Titanocene Y with Ser 285 of albumin for which
a distant of 3.10 Å is observed between the involved oxygen atoms.
In addition O1 shows a short interaction with Ala 256 of 3.70 Å as
well. O2 of Titanocene Y interacts with Ph–Ala 68 at 2.87 Å and
with Arg 255 of the albumin backbone at 3.77 Å. Further van der
Waals interactions are seen from the chlorides of Titanocene Y:
both of them approach Leu 20 and Ala 149 by just under 3.0 Å
and interact with Pro 150 at a distance of just over 3.0 Å. Fig. 4
shows the area of interest, in which Titanocene Y is found to bind
to albumin. Overall, it can be summarised that Titanocene Y fits
well into the albumin cavity and forms a relatively weak and obvi-
ously reversible complex, which is likely to be carried into the can-
cer cell and released there. This result could also explain the
protective role played by BSA in the in vitro test on MCF-7 cells.
In future experiments, systematic molecular modeling studies
in combination with NMR spectroscopic experiments will reveal
the exact binding site of Titanocene Y in HSA and will hopefully
give access to the binding energy.
Titanocene Y
200
100
0
Ti Y (10 µM) +
BSA (µg/ml)
Fig. 3. Effect of 1 nM of estradiol (E2), of 10
l
M of Titanocene Y incubated alone or
in the presence of various concentration of bovine serum albumin (BSA; 20–50 lg/
mL) on the growth of MCF-7 cells (hormone-dependent breast cancer cells) after 5
days of culture. Comparison with the control (C) set at 100%. Mean of at least two
independent experiments (six measurements for each one) range.
3.2.4. Effect of Ferrocene Y on the growth of MCF-7 and MDA-MB-231
cells
The effect of Ferrocene Y has been tested on the growth of MCF-
7 and MDA-MB-231, the two breast cancer cells lines which are
routinely used to study the estrogenicity (MCF-7) and the cytotox-
icity (MDA-MB-231) of the ferrocenyl derivatives [39]. On MCF-7
for this compound alone on LLC-PK cells [17]. Our results indicate
that the cytotoxicity of the complex is significantly enhanced by its
binding to albumin. Interestingly, Titanocene Y alone has no estro-
genic effect suggesting that this molecule is more stable to total
hydrolysis than Titanocene K.
cells Ferrocene Y shows no proliferative effect at 1
lM and be-
comes slightly cytotoxic at 10 M with an anti-proliferative effect
l
of 18% at this concentration. The absence of estrogenicity of this
molecule is expected has it lacks a phenol ring that is considered
to be essential for the binding of a molecule to the estrogen recep-
tor. Not surprisingly also it is unable to displace estradiol from its
specific receptor (RBA = 0). The IC50 value found for this compound
3.2.3. Molecular modeling studies of Titanocene Y in the free cavity of
human serum albumin (HSA)
Cancer cells are known to be leaky for albumin and therefore
one possible strategy for the targeted delivery of anti-cancer drugs
are drug–albumin conjugates. The preliminary docking studies of
Titanocene Y using a free cavity in HSA resulted in a surprisingly
on MDA-MB-231 cells is 95
than the value found for ferrocene alone (IC50 = 160
much higher than the value found for Ferrocifen, the ferrocenyl
lM. So this value is a little bit lower
lM) but it is
Fig. 4. Plot of the relevant structural part of the albumin–Titanocene Y conjugate obtained by molecular modeling using MOLVIEW.

A. Vessières et al. / Journal of Organometallic Chemistry 694 (2009) 874–879
879
[12] M. Tacke, L.T. Allen, L. Cuffe, W.M. Gallagher, Y. Lou, O. Mendoza, H. Müller-
Bunz, F.J.K. Rehmann, N. Sweeney, J. Organomet. Chem. 689 (2004) 2242.
[13] M. Tacke, L.P. Cuffe, W.M. Gallagher, Y. Lou, O. Mendoza, H. Müller-Bunz, F.-J.K.
Rehmann, N. Sweeney, J. Inorg. Biochem. 98 (2004) 1987.
[14] K. Strohfeldt, H. Müller-Bunz, C. Pampillon, N.J. Sweeney, M. Tacke, Eur. J.
Inorg. Chem. (2006) 4621.
[15] N. Sweeney, W.M. Gallagher, H. Müller-Bunz, C. Pampillon, K. Strohfeldt, M.
Tacke, J. Inorg. Biochem. 100 (2006) 1479.
[16] N.J. Sweeney, J. Claffey, H. Müller-Bunz, C. Pampillon, K. Strohfeldt, M. Tacke,
Appl. Organomet. Chem. 21 (2007) 57.
analog of Tamoxifen, which has an IC50 value of 0.5 lM [39]. This
result could easily be explain by the fact that we have demon-
strated that this cytotoxicity is connected to the oxidative forma-
tion of quinoid species [47]. Ferrocene Y possessing methoxy
rather than hydroxy groups and a non-conjugated system cannot
undergo this transformation [48]. Then, one can predict that it will
behave like ferrocene and this is the experimental observation.
[17] N.J. Sweeney, O. Mendoza, H. Müller-Bunz, C. Pampillon, F.-J.K. Rehmann, K.
Strohfeldt, M. Tacke, J. Organomet. Chem. 690 (2005) 4537.
[18] T. Hickey, J. Claffey, E. Fitzpatrick, M. Hogan, C. Pampillon, M. Tacke, Invest.
New Drugs 25 (2007) 425.
[19] M. Hogan, J. Claffey, C. Pampillon, R.W.G. Watson, M. Tacke, Organometallics
26 (2007) 2501.
[20] C. Pampillon, J. Claffey, M. Hogan, K. Strohfeldt, M. Tacke, Trans. Met. Chem. 32
(2007) 434.
[21] C. Pampillon, J. Claffey, M. Hogan, M. Tacke, Z. Anorg. Allg. Chem. 633 (2007)
1695.
[22] C. Pampillon, J. Claffey, M. Hogan, M. Tacke, BioMetals 21 (2008) 197.
[23] C. Pampillon, J. Claffey, K. Strohfeldt, M. Tacke, Eur. J. Med. Chem. 43 (2008)
122.
[24] C. Pampillon, O. Mendoza, N.J. Sweeney, K. Strohfeldt, M. Tacke, Polyhedron 25
(2006) 2101.
[25] C. Pampillon, N.J. Sweeney, K. Strohfeldt, M. Tacke, Inorg. Chim. Acta 359
(2006) 3969.
[26] C. Pampillon, N.J. Sweeney, K. Strohfeldt, M. Tacke, J. Organomet. Chem. 692
(2007) 2153.
4. Conclusion
The in vitro study of the cytotoxicity of three metallocenes on
the hormone dependent breast cancer cell line MCF-7 shows three
different behaviors. Titanocene K shows, at concentrations as low
as 0.5
Cp2TiCl2 and this effect is not changed by the addition of albumin.
By contrast, Titanocene Y alone at 10 M has almost no effect on
lM, an estrogenic effect similar to that observed with
l
the growth of these cells but it becomes cytotoxic in the presence
of albumin. Finally, Ferrocene Y is only slightly cytotoxic at the
same concentration. These results underline the interest of
in vitro study of potential drugs in the presence of serum proteins
and might lead to an albumin-based formulation of Titanocene Y
for further xenograft experiments.
[27] G. Kelter, N.J. Sweeney, K. Strohfeldt, H.-H. Fiebig, M. Tacke, Anti-Cancer Drugs
16 (2005) 1091.
[28] O. Oberschmidt, A.-R. Hanauske, F.-J.K. Rehmann, K. Strohfeldt, N. Sweeney, M.
Tacke, Anti-Cancer Drugs 16 (2005) 1071.
5. Supplementary material
[29] O. Oberschmidt, A.R. Hanauske, C. Pampillon, N.J. Sweeney, K. Strohfeldt, M.
Tacke, Anti-Cancer Drugs 18 (2007) 317.
CCDC 669818 contains the supplementary crystallographic data
for this paper. These data can be obtained free of charge from The
Cambridge Crystallographic Data Centre via www.ccdc.cam.ac.uk/
data_request/cif.
[30] K. O’Connor, C. Gill, M. Tacke, F.-J.K. Rehmann, K. Strohfeldt, N. Sweeney, J.M.
Fitzpatrick, R.W.G. Watson, Apoptosis 11 (2006) 1205.
[31] M.C. Valadares, A.L. Ramos, F.-J.K. Rehmann, N.J. Sweeney, K. Strohfeldt, M.
Tacke, M.L.S. Queiroz, Eur. J. Pharmacol. 534 (2006) 264.
[32] P. Beckhove, O. Oberschmidt, A.R. Hanauske, C. Pampillon, V. Schirrmacher, N.J.
Sweeney, K. Strohfeldt, M. Tacke, Anti-Cancer Drugs 18 (2007) 311.
[33] I. Fichtner, C. Pampillon, N.J. Sweeney, K. Strohfeldt, M. Tacke, Anti-Cancer
Drugs 17 (2006) 333.
[34] S. Top, E.B. Kaloun, A. Vessières, I. Laïos, G. Leclercq, G. Jaouen, J. Organomet.
Chem. 643–644 (2002) 350.
[35] A.R. Timerbaev, C.G. Hartinger, S. Aleksenko, B.K. Keppler, Chem. Rev. 106
(2006) 2224.
Acknowledgments
We thank A. Cordaville for technical assistance and COST D39
for support. M.T. thanks the Université Paris 6 for a 1-month
professorship.
[36] M. Ravera, C. Cassino, E. Monti, M. Gariboldi, D. Osella, J. Inorg. Biochem. 99
(2005) 2264.
References
[37] G. Jaouen, S. Top, A. Vessières, G. Leclercq, M.J. McGlinchey, Curr. Med. Chem.
11 (2004) 2505.
[38] G. Jaouen, S. Top, A. Vessières, G. Leclercq, J. Quivy, L. Jin, A. Croisy, C.R. Acad.
Sci. Paris Sér. IIc (2000) 89.
[39] A. Nguyen, A. Vessieres, E.A. Hillard, S. Top, P. Pigeon, G. Jaouen, Chimia 61
(2007) 761.
[40] S. Top, A. Vessières, G. Leclercq, J. Quivy, J. Tang, J. Vaissermann, M. Huché, G.
Jaouen, Chem. Eur. J. 9 (2003) 5223.
[41] A. Vessières, S. Top, P. Pigeon, E.A. Hillard, L. Boubeker, D. Spera, G. Jaouen, J.
Med. Chem. 48 (2005) 3937.
[42] HyperChem 7.0, Hypercube, Alberta, Canada, 2002.
[43] J.M. Cense, Phys. Theor. Chem. 71 (1990) 763.
[44] G.M. Sheldrick, University of Göttingen, Germany, 2002.
[45] G.M. Sheldrick, University of Göttingen, Germany, 1997.
[46] A.A. Bhattacharya, T. Grune, S. Curry, J. Mol. Biol. 303 (2000) 721.
[47] E.A. Hillard, A. Vessières, L. Thouin, G. Jaouen, C. Amatore, Angew. Chem., Int.
Ed. 45 (2006) 285.
[1] T. Schilling, B.K. Keppler, M.E. Heim, G. Niebch, H. Dietzfelbinger, J. Rastetter,
A.R. Hanauske, Invest. New Drugs 13 (1996) 327.
[2] F. Caruso, M. Rossi, Met. Ions Biol. Syst. 42 (2004) 353.
[3] E. Melendez, Crit. Rev. Oncol. Hematol. 42 (2002) 309.
[4] G. Lummen, H. Sperling, H. Luboldt, T. Otto, H. Rubben, Cancer Chemother.
Pharmacol. 42 (1998) 415.
[5] N. Kröger, U.R. Kleeberg, K. Mross, L. Edler, D.K. Hossfeld, Onkologie 23 (2000)
60.
[6] O.R. Allen, L. Croll, A.L. Gott, R.J. Knox, P.C. McGowan, Organometallics 23
(2004) 288.
[7] P.W. Causey, M.C. Baird, S.P.C. Cole, Organometallics 23 (2004) 4486.
[8] R. Meyer, S. Brink, C.E.J. van Rensburg, G.K. Joone, H. Gorls, S. Lotz, J.
Organomet. Chem. 690 (2005) 117.
[9] S. Fox, J.P. Dunne, M. Tacke, J.F. Gallagher, Inorg. Chim. Acta 357 (2004) 225.
[10] F.-J.K. Rehmann, L.P. Cuffe, O. Mendoza, D.K. Rai, N. Sweeney, K. Strohfeldt,
W.M. Gallagher, M. Tacke, Appl. Organomet. Chem. 19 (2005) 293.
[11] F.-J.K. Rehmann, A.J. Rous, O. Mendoza, N.J. Sweeney, K. Strohfeldt, W.M.
Gallagher, M. Tacke, Polyhedron 24 (2005) 1250.
[48] E.A. Hillard, A. Vessières, F. Le Bideau, D. Plazuk, D. Spera, M. Huché, G. Jaouen,
ChemMedChem 1 (2006) 551.