Organometallic Antitumor Compounds: Ferrocifens as Precursors to Quinone Methides

Article abstract of DOI:10.1002/anie.201503048

The synthesis and chemical oxidation profile of a new generation of ferrocifen derivatives with strong antiproliferative behavior in vitro is reported. In particular, the hydroxypropyl derivative HO(CH₂)₃C(Fc)=C(C₆H₄OH)₂ (3 b) exhibited exceptional antiproliferative activity against the cancer cell lines HepG2 and MDA-MB-231 TNBC, with IC₅₀ values of 0.07 and 0.11 μM, respectively. Chemical oxidation of 3 b yielded an unprecedented tetrahydrofuran-substituted quinone methide (QM) via internal cyclization of the hydroxyalkyl chain, whereas the corresponding alkyl analogue CH₃CH₂-C(Fc)=C(C₆H₄OH)₂ merely formed a vinyl QM. The ferrocenyl group in 3 b plays a key role, not only as an intramolecular reversible redox "antenna", but also as a stabilized carbenium ion "modulator". The presence of the oxygen heterocycle in 3 b-QM enhances its stability and leads to a unique chemical oxidation profile, thus revealing crucial clues for deciphering its mechanism of action in vivo.

Full text of DOI:10.1002/anie.201503048

.
Angewandte
Communications
International Edition: DOI: 10.1002/anie.201503048
German Edition: DOI: 10.1002/ange.201503048
Antitumor Agents
Organometallic Antitumor Compounds: Ferrocifens as Precursors to
Quinone Methides**
Yong Wang, Pascal Pigeon, Siden Top,* Michael J. McGlinchey, and GØrard Jaouen*
Abstract: The synthesis and chemical oxidation profile of
a new generation of ferrocifen derivatives with strong anti-
proliferative behavior in vitro is reported. In particular, the
hydroxypropyl derivative HO(CH2)3C(Fc) = C(C6H4OH)2
(3b) exhibited exceptional antiproliferative activity against
the cancer cell lines HepG2 and MDA-MB-231 TNBC, with
IC50 values of 0.07 and 0.11 mm, respectively. Chemical
oxidation of 3b yielded an unprecedented tetrahydrofuran-
substituted quinone methide (QM) via internal cyclization of
the hydroxyalkyl chain, whereas the corresponding alkyl
Figure 1. Acyclic and ansa organometallic compounds derived from
the ferrocifen family (R=H or OH), and the key acyclic quinone
methide (QM).
analogue CH3CH2-C(Fc) = C(C6H4OH)2 merely formed
a vinyl QM. The ferrocenyl group in 3b plays a key role, not
only as an intramolecular reversible redox “antenna”, but also
as a stabilized carbenium ion “modulator”. The presence of the
oxygen heterocycle in 3b-QM enhances its stability and leads
to a unique chemical oxidation profile, thus revealing crucial
clues for deciphering its mechanism of action in vivo.
molecules with triple-negative MDA-MB-231 cancer cells,
and have been formulated for in vivo studies in breast cancer
and glioma .[3d,6] These products are of great potential interest
since senescence could be an alternative route for cancers
resistant to proapoptotic stimuli, which play a major role in
the incidence of mortality due to this type of pathology.[4b]
Ideally, to enable the establishment of a new mechanism
of action, a new generation of products is required. These
products should involve an unequivocally characterized
primary metabolite and they must also exhibit lower IC50
values than those of the acyclic series. Previously, the
molecules in Figure 1 were modified by changing the metals
and functional groups, and attaching various chains.[7] With
replacement of the alkyl chain by a hydroxypropyl group, we
present the first observations on what may become a new
generation of ferrocifen derivatives (Figure 2A) with strong
antiproliferative potential. Moreover, they yielded a novel
quinone methide (QM) that is key to the new behavior.
I
n the search for organometallic antitumor compounds with
behavior differentiated from the primary targeting of DNA,
an approach based on common metals such as Fe has already
proven to be of interest.[1] Many ferrocenyl species showing
contrasting antiproliferative effects,[2] in particular the family
of acyclic or ansa ferrocifens (Figure 1) are seen as innovative
drug candidates.[3] At low concentrations, they operate
principally via a mechanism of senescence,[4] although in
some cases apoptosis is possible,[4a] through at least partially
targeting enzymes of a redox system, for example, thioredoxin
reductase (TrxR),[5] that are overexpressed in cancer cells,
while leaving healthy cells untouched.[3c,d] They show IC50
values of around 0.6 mm for acyclic systems to 0.09 mm for ansa
[*] Dr. Y. Wang, Dr. P. Pigeon, Dr. S. Top, Prof. G. Jaouen
Sorbonne UniversitØs, UPMC Univ Paris 06
UMR 8232, IPCM, 75005 Paris (France)
and
CNRS, UMR 8232, IPCM, 75005 Paris (France)
and
PSL, Chimie ParisTech
11 rue Pierre et Marie Curie, 75005 Paris (France)
E-mail: siden.top@chimie-paristech.fr
gerard.jaouen@chimie-paristech.fr
Figure 2. A) The new hydroxypropyl series 3, derived from the acyclic
ferrocifen family, and the novel tetrahydrofuran-substituted 3-QM.
B) The purely organic equivalent 4b and the acyclic vinyl 4b-QM.
Prof. M. J. McGlinchey
UCD School of Chemistry and Chemical Biology
University College Dublin, Belfield, Dublin 4 (Ireland)
Efficient syntheses of the desired hydroxypropyl-alkenes
3 proceeded via McMurry cross-coupling[8] to give compounds
2a–c and LiAlH4 reduction of the esters to the corresponding
alcohols 3 (Scheme 1). Monophenol analogues 2a, 2c, 3a, and
3c were obtained as mixtures of Z and E isomers. The
corresponding tamoxifen organic derivatives (series 4) were
prepared analogously. X-ray crystal structures of 3a and 3b
are shown in Figure S1 in the Supporting Information.
[**] Y. W. thanks the PGG foundation, the PSL University and Feroscan
for financial support. We thank P. Herson (UniversitØ Pierre et Marie
Curie, Paris, IPCM and LabEx Michem) for the X-ray structure
determinations.
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.201503048.
10230
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2015, 54, 10230 –10233

Angewandte
Chemie
Scheme 2. Synthesis of the quinone methides.
Scheme 1. Syntheses of the ferrocenyl compounds in series 2 and 3.
Lipophilicity values and antiproliferative effects against
the estrogen-receptor-negative (ERÀ) MDA-MB-231 breast
cancer cells are reported in Table 1, together with values
systems involving cyclization of the hydroxyalkyl chain rather
than the previously reported behavior of acyclic 1-QM, which
contains a conjugated trans double bond.[3e]
=
obtained for the acyclic diphenol compound Fc(Et)C C-
The QM structure of 3c-QM was established by X-ray
crystallography (Figure 3) and the cyclic ether confirms the
intramolecular reaction of the hydroxy function on the carbon
center adjacent to the ferrocenyl group.
(C6H4OH)2. Compared to the organic compound 4b, the
ferrocenyl moiety increases the lipophilicity of the new
molecules, while the presence of polar groups decreases
Table 1: IC50 values with MDA-MB-231 cells and log Po/w results for
selected ferrocenyl compounds.
Compound
IC50 [mM][a]
logPo/w
1 (R=OH)b
0.64Æ0.06
50.78Æ1.33
4.09Æ1.67
0.44Æ0.09
1.50Æ0.17
1.16Æ0.02
0.11Æ0.02
1.23Æ0.05
5.0
3.6
4b
2a
2b
2c
3a
3b
3c
5.0/5.3
4.4
5.4/6.2
4.5/4.9
4.2
5.0/5.6
[a] Measured after 5 days of culture (mean of two independent
experimentsÆSD). [b] Values from Ref. [1a].
Figure 3. Molecular structure of 3c-QM with thermal ellipsoids shown
at 50% probability.
their lipophilicity. Compound 3b shows exceptional antipro-
liferative effects on the hormone-independent cell line MDA-
MB-231, with an IC50 value around 0.11 mm, which is markedly
superior to that of its ester precursor 2b and that of the acyclic
diphenol compound. The tamoxifen derivative 4b, the organic
analogue of 3b, shows limited cytotoxicity against MDA-MB-
231 (IC50 ꢀ 50 mm), a result that underlines the key role played
by the ferrocenyl substituent in the generation of antiproli-
ferative effects (tamoxifen itself has an IC50 value of 40 mm).[9]
Compound 3b was selected for further tests on human
tumor cell lines, which revealed that it could prevent the
growth of the pancreatic carcinoma cell line Mia-PaCa with
an IC50 value of 1.23 mm and shows excellent antiproliferative
activity against liver hepatocellular carcinoma, with an IC50
value of 0.07 mm on HepG2 cells. This exceptional antiproli-
ferative behavior led to further exploration of its framework
to try to decipher the active motif.
For the diphenol 1 (R = OH), an archetypical example of
the first generation of ferrocifen acyclic species (Figure 1), the
corresponding QM proved to be quite unstable and exhibited
evidence of decomposition before oxidation was complete. In
contrast, 3b-QM derived from the hydroxypropyldiphenol
precursor showed reasonable stability over least 4 weeks
when kept at À208C in the solid state. Moreover, its stability
in solution was also significantly improved (Figures S2,S3),
except in the weakly acidic solvent chloroform, in which it
started to decompose. The half-life in acetone was approx-
imately 30 h, but was even longer in DMSO, where the
compound remained unchanged for at least two days with
a half-life of around six weeks.
We find that compounds 1-QM (Figure 1) are the key
metabolites upon metabolism of the acyclic alkyl chain series
1 by rat liver microsomes and they can also be prepared by
chemical oxidation.[3e,f] To help unravel the reasons for the
excellent antiproliferative effect of 3b, chemical oxidation
was used to prepare selected putative metabolites
(Scheme 2). NMR and X-ray crystallography data confirmed
their structures as novel tetrahydrofuran-substituted 3-QM
The formation of the quinone ring was entirely consistent
with the mechanism already proposed for acyclic species,
whereby the ferrocenyl unit acts as a kind of intramolecular
oxidation “antenna”, such that oxidation of the phenol group
leads to a carbenium ion.[10] However, from this point the
process can vary owing to the presence of the terminal
hydroxy function, such that the carbenium ion undergoes
nucleophilic attack to form a heterocycle rather than a double
Angew. Chem. Int. Ed. 2015, 54, 10230 –10233
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org 10231

.
Angewandte
Communications
bond (Scheme 3). Oxidation of compounds 4, the purely
organic analogues of the ferrocenyl alcohol derivatives 3,
yielded only the acyclic vinyl QM systems (Figure 2B). The
absence of a significant NMR peak between 3.5 and 4.5 ppm,
Scheme 4. Metabolic Stability Profile of 3b-QM.
species resulting from involvement of the oxygen atom of the
alkyl chain.
As shown in Scheme S1, under the slightly acidic con-
ditions in acetone, protonation of the quinone 3b-QM can
form a carbenium ion that evolves through several pathways.
The first is ring expansion by migration of the adjacent
oxygen, which places the carbocation adjacent to the ferro-
cenyl group, and subsequent proton loss to yield 3b-A.
Secondly, a pinacol rearrangement can give compound 3b-
B.[11] Finally, the carbenium ion can react with traces of water
to give a pinacol-type intermediate that gives 3b-C and 3b-D
on further oxidation. This radical oxidation parallels a pre-
vious report on [3]-ferrocenophane derivatives.[11]
The cytotoxicity of 3b-QM against MDA-MB-231 cells
was consistent with its chemical derivatives 3b-A and 3b-B
(IC50 values of 2.03 and 4.14 mm, respectively). Compounds
3b-QM, 3b-A, 3b-B, 3b-C, and 3b-D were the major
products during the chemical oxidation and subsequent
decomposition of 3b. The cytotoxicity values obtained from
the precursor 3b strongly suggest that, when incubated with
live cancer cells, 3b should generate a remarkable intrinsically
electrophilic metabolite that is probably native 3b-QM. The
evolution of 3b proposed above occurred through chemical
methods without the involvement of other nucleophiles. It
provides some clues that prodrug 3b could generate several
possible carbenium ions in vivo that could be captured by
nucleophiles, such as thiols or selenols, inside the cells.[5] Thus
possible cross-coupling with reactive nucleophiles or a related
protein could lead to cell death. Another pathwaythat could
be relevant to the cytotoxicity is oxidative cleavage of
pinacol-type analogues.[11]
In summary, modification of the alkyl chain in the original
acyclic derivatives yielded 3b, which bears a terminal hydrox-
yalkyl group and exhibits exceptional antiproliferative activ-
ity against liver hepatocellular carcinoma cells (HepG2) and
ERÀ breast cancer cells (MDA-MB-231), with IC50 values of
0.07 and 0.11 mm, respectively. Chemical oxidation of 3b
yielded an unprecedented tetrahydrofuran-substituted QM
via internal cyclization of the alkyl chain, which was identified
as a possible key primary metabolite. The ferrocenyl group
not only plays the role of intramolecular reversible redox
“antenna” but also acts as a stabilized carbenium ion
“modulator”. Resulting from these structural changes, 3b-
QM exhibits moderate stability and a unique chemical
oxidation profile, which reveals crucial clues that may help
us decipher its mechanism of action in vivo. Future work will
focus on gaining insight into the mechanism of action of these
novel species, especially in the presence of healthy cells and
Scheme 3. Proposed mechanism for the formation of the novel hetero-
cyclic ferrocenyl QM species.
which is characteristic of OCH2 in a tetrahydrofuran ring,
indicates that the oxidation of 4 did not yield a QM hetero-
cycle. NMR peaks characteristic of the acyclic QM appeared
transiently, but decomposition prior to complete oxidation
prevented isolation of the QM, even for compound 4c, in
which the ferrocenyl group of 3c was replaced by a phenyl
group. Thus, for the organic analogue 4, oxidation predom-
inantly gives the vinyl QM, whereas in the ferrocenyl series 3,
oxidation furnishes a novel QM that bears a tetrahydrofuran
ring. These quite distinct results, together with the well-
known fact that organometallic complexes adjacent to
a double bond favor the stabilization of a-carbenium ions,
let us deduce that the ferrocenyl group not only plays the role
of intramolecular oxidation “antenna” but also acts as
a “modulator” and facilitates trapping of the hydroxy
function by the carbenium ion, thereby leading to tetrahy-
drofuran ring formation. To the best of our knowledge, the 3-
QM species is the first tetrahydrofuran-substituted QM
reported to date, which may indicate the potential for
structural diversity in quinone chemistry.
Freshly synthesized 3a-QM, 3b-QM, and 3c-QM were
also cytotoxic against MDA-MB-231 cells (IC50 values of 1.89,
4.39, and 6.00 mm, respectively). The IC50 values of stable 3a-
QM and 3c-QM were the same as those of their parent
molecules 3a and 3c. The higher value for 3b-QM, relative to
3b, could be due to its relatively better chemical reactivity,
since strong nucleophiles may be present in the incubation
medium. Nevertheless, this behavior suggests that 3b should
have remarkable intrinsic antiproliferative properties and
motivated us to explore the chemical oxidation profiles of 3b
and 3b-QM.
The oxidative evolutio of 3b-QM is more complex than
that of 1-QM (Scheme 4).[3e, 5] Its half-life in acetone was
around 30 h, and all of its derivatives were stable enough to be
isolable upon complete decomposition. The four products,
3b-A, 3b-B, 3b-C, and 3b-D, which had not previously been
observed in the metabolic processes of the 1-QM series,[5]
were identified by NMR or X-ray crystallography (3b-A;
Figure S1). For the acyclic ferrocifen derivatives, the major
byproduct in each case, during or after oxidation, was an
indene product resulting from acid-mediated cyclization,[3e,5]
but the presence of the tetrahydrofuran ring leads to different
10232 www.angewandte.org
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2015, 54, 10230 –10233

Angewandte
Chemie
Jaouen, D. Mansuy, Angew. Chem. Int. Ed. 2009, 48, 9124 – 9126;
selected nucleophiles, thus opening the way to a new gen-
eration of unique potential drug candidates.
Angew. Chem. 2009, 121, 9288 – 9290; f) M.-A. Richard, D.
Hamels, P. Pigeon, S. Top, P. M. Dansette, H. Z. S. Lee, A.
Vessi›res, D. Mansuy, G. Jaouen, ChemMedChem 2015, 10, 981 –
990.
Keywords: antitumor agents · drug discovery · ferrocene ·
metabolism · quinones
[4] a) A. Vessi›res, C. Corbet, J. M. Heldt, N. Lories, N. Jouy, I.
Laïos, G. Leclercq, G. Jaouen, R. A. Toillon, J. Inorg. Biochem.
2010, 104, 503 – 511; b) C. Bruy›re, V. Mathieu, A. Vessi›res, P.
Pigeon, S. Top, G. Jaouen, R. Kiss, J. Inorg. Biochem. 2014, 141,
144 – 151.
How to cite: Angew. Chem. Int. Ed. 2015, 54, 10230–10233
Angew. Chem. 2015, 127, 10368–10371
[1] a) E. Hillard, A. Vessieres, G. Jaouen, in Medicinal Organome-
tallic Chemistry, Vol. 32 (Eds.: G. Jaouen, N. Metzler-Nolte),
Topics in Organometallic Chemistry, Springer, Berlin, 2010,
pp. 81 – 117; b) S. S. Braga, A. M. S. Silva, Organometallics 2013,
32, 5626 – 5639.
[2] a) S. B. Deepthi, R. Trivedi, L. Giribabu, B. Sridhar, P. Sujitha,
C. G. Kumar, J. Organomet. Chem. 2014, 774, 26 – 34; b) M.
Librizzi, A. Longo, R. Chiarelli, J. Amin, J. Spencer, C.
Luparello, Chem. Res. Toxicol. 2012, 25, 2608 – 2616; c) A. G.
Harry, W. E. Butler, J. C. Manton, M. T. Pryce, N. O’Donovan, J.
Crown, D. K. Rai, P. T. M. Kenny, J. Organomet. Chem. 2014,
766, 1 – 12; d) B. Balaji, B. Balakrishnan, S. Perumalla, A. A.
Karande, A. R. Chakravarty, Eur. J. Med. Chem. 2014, 85, 458 –
467; e) J. d. J. Cµzares-Marinero, S. Top, A. Vessi›res, G. Jaouen,
Dalton Trans. 2014, 43, 817 – 830; f) P. Marzenell, H. Hagen, L.
Sellner, T. Zenz, R. Grinyte, V. Pavlov, S. Daum, A. Mokhir, J.
Med. Chem. 2013, 56, 6935 – 6944.
[3] a) D. Plaz˙uk, A. Vessi›res, E. A. Hillard, O. Buriez, E. Labbe, P.
Pigeon, M. A. Plamont, C. Amatore, J. Zakrzewski, G. Jaouen, J.
Med. Chem. 2009, 52, 4964 – 4967; b) M. Gçrmen, P. Pigeon, S.
Top, E. A. Hillard, M. Huche, C. G. Hartinger, F. de Montigny,
M. A. Plamont, A. Vessieres, G. Jaouen, ChemMedChem 2010,
5, 2039 – 2050; c) Q. Michard, G. Jaouen, A. Vessieres, B. A.
Bernard, J. Inorg. Biochem. 2008, 102, 1980 – 1985; d) E. Allard,
C. Passirani, E. Garcion, P. Pigeon, A. Vessieres, G. Jaouen, J. P.
Benoit, J. Controlled Release 2008, 130, 146 – 153; e) D. Hamels,
P. M. Dansette, E. A. Hillard, S. Top, A. Vessieres, P. Herson, G.
[5] A. Citta, A. Folda, A. Bindoli, P. Pigeon, S. Top, A. Vessieres, M.
Salmain, G. Jaouen, M. P. Rigobello, J. Med. Chem. 2014, 57,
8849 – 8859.
[6] a) A. L. LainØ, E. Adriaenssens, A. Vessieres, G. Jaouen, C.
Corbet, E. Desruelles, P. Pigeon, R. A. Toillon, C. Passirani,
Biomaterials 2013, 34, 6949 – 6956; b) A. L. LainØ, A. Clavreul,
A. Rousseau, C. TØtaud, A. Vessieres, E. Garcion, G. Jaouen,
R. A. Toillon, C. Passirani, Nanomedicine 2014, 10, 1667 – 1677.
[7] G. Jaouen, S. Top, in Advances in Organometallic Chemistry and
Catalysis (Ed.: A. J. L. Pombeiro), Wiley, Hoboken, 2014,
pp. 563 – 580.
[8] P. Pigeon, M. Gormen, K. Kowalski, H. Mueller-Bunz, M. J.
McGlinchey, S. Top, G. Jaouen, Molecules 2014, 19, 10350 –
10369.
[9] S. Groleau, J. Nault, M. Lepage, M. Couture, N. Dallaire, G.
BØrubØ, R. C. Gaudreault, Bioorg. Chem. 1999, 27, 383 – 394.
[10] P. Messina, E. Labbe, O. Buriez, E. A. Hillard, A. Vessieres, D.
Hamels, S. Top, G. Jaouen, Y. M. Frapart, D. Mansuy, C.
Amatore, Chem. Eur. J. 2012, 18, 6581 – 6587.
[11] M. Gçrmen, P. Pigeon, E. A. Hillard, A. Vessi›res, M. Huche,
M. A. Richard, M. J. McGlinchey, S. Top, G. Jaouen, Organo-
metallics 2012, 31, 5856 – 5866.
Received: April 16, 2015
Published online: July 14, 2015
Angew. Chem. Int. Ed. 2015, 54, 10230 –10233
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org 10233