Aminoferrocene-based prodrugs and their effects on human normal and cancer cells as well as bacterial cells

Article abstract of DOI:10.1021/jm400754c

Aminoferrocene-based prodrugs are activated under cancer-specific conditions (high concentration of reactive oxygen species, ROS) with the formation of glutathione scavengers (p-quinone methide) and ROS-generating iron complexes. Herein, we explored three structural modifications of these prodrugs in an attempt to improve their properties: (a) the attachment of a -COOH function to the ferrocene fragment leads to the improvement of water solubility and reactivity in vitro but also decreases cell-membrane permeability and biological activity, (b) the alkylation of the N-benzyl residue does not show any significant affect, and (c) the attachment of the second arylboronic acid fragment improves the toxicity (IC₅₀) of the prodrugs toward human promyelocytic leukemia cells (HL-60) from 52 to 12 μM. Finally, we demonstrated that the prodrugs are active against primary chronic lymphocytic leukemia (CLL) cells, with the best compounds exhibiting an IC₅₀ value of 1.5 μM. The most active compounds were found to not affect mononuclear cells and representative bacterial cells.

Full text of DOI:10.1021/jm400754c

Article
pubs.acs.org/jmc
Aminoferrocene-Based Prodrugs and Their Effects on Human Normal
and Cancer Cells as Well as Bacterial Cells
Paul Marzenell,†,∥ Helen Hagen,†,∥ Leopold Sellner,‡ Thorsten Zenz,‡ Ruta Grinyte,§ Valeri Pavlov,§
Steffen Daum,† and Andriy Mokhir*,†
†Department of Chemistry and Pharmacy, Friedrich-Alexander-University of Erlangen-Nurnberg, Organic Chemistry II, Henkestr. 42,
̈
91054 Erlangen, Germany
‡Department of Translational Oncology, National Center for Tumor Diseases (NCT) and German Cancer Research Center (DKFZ)
Heidelberg, Im Neuenheimer Feld 460, 69120 Heidelberg, Germany and Department of Medicine V, University Hospital Heidelberg,
Im Neuenheimer Feld 410, 69120 Heidelberg, Germany
§Centre for Cooperative Research in Biomaterials, CIC biomaGUNE, Laboratory of Biofunctional Materials III, Parque Technolog
de San Sebastian, Ed. P° Miramon 182, Guipuzcoa, Spain
́
ico
́
́
́
S
* Supporting Information
ABSTRACT: Aminoferrocene-based prodrugs are activated under cancer-
specific conditions (high concentration of reactive oxygen species, ROS)
with the formation of glutathione scavengers (p-quinone methide) and
ROS-generating iron complexes. Herein, we explored three structural
modifications of these prodrugs in an attempt to improve their properties:
(a) the attachment of a −COOH function to the ferrocene fragment leads
to the improvement of water solubility and reactivity in vitro but also
decreases cell-membrane permeability and biological activity, (b) the
alkylation of the N-benzyl residue does not show any significant affect, and
(c) the attachment of the second arylboronic acid fragment improves the
toxicity (IC50) of the prodrugs toward human promyelocytic leukemia cells
(HL-60) from 52 to 12 μM. Finally, we demonstrated that the prodrugs are
active against primary chronic lymphocytic leukemia (CLL) cells, with the
best compounds exhibiting an IC50 value of 1.5 μM. The most active compounds were found to not affect mononuclear cells and
representative bacterial cells.
specific conditions, should lack this disadvantage. Few
anticancer prodrugs have been described up to date. For
example, glycopeptides bleomycins are used as cytostatic agents
in the chemotherapy of a variety of cancers.13 They bind iron
ions, which are available in cancer cells and to a lesser extent in
normal cells, resulting in the formation of iron/bleomycin
complexes. The latter compounds bind genomic DNA and
induce its cleavage. Ferrocene derivative hydroxyferrocifen and
its analogues, developed by the group of G. Jaouen, are
activated in cells by oxidation with the formation of alkylating
agents, which scavenge glutathione and thereby inhibit the
cellular antioxidative system.14 Furthermore, the group of C.
Jacob has prepared organochalcogen-based compounds that are
activated in the presence of elevated (cancer-specific) ROS with
the formation of ROS-generating catalysts. The latter species
induce strong oxidative stress in cells leading to their death.15
Cleavage of aryl- and alkylboronic acids and their esters in
the presence of hydrogen peroxide has been used for a long
time in synthetic organic chemistry in the last step of a
INTRODUCTION
■
Cancer cells are formed as a result of genetic transformation of
normal cells. Because of these alterations, an abnormal, cancer-
specific microenvironment is created both in the cancer cells
themselves and in the tumor tissues. These differences can be
exploited to design cancer-specific drugs.1 For example, the
1
−
increased ROS level (ROS: O2, O2 , H2O2, and HO•) is an
especially attractive target for anticancer therapy because it
seems to be a general feature of cancer.2 Drugs targeting this
feature can potentially exhibit activity against all cancer types.
Several anticancer drugs have been described that act by
increasing the ROS concentration in cancer cells beyond the
apoptotic level and thereby inducing their death. These include
arsenic trioxide,3 fenretinide,4 nitric oxide-donating aspirin
(NO-ASA),5 buthionine sulfoximine (BSO),6 imexone,7
motexafin gadolinium,8 menadione,9 β-lapachone,10 deoxyny-
boquinone,11 and others.12 Although such compounds kill
cancer cells efficiently, they also increase the ROS concen-
tration in normal cells, thereby making the process of mutating
native genomic DNA more probable. This side effect is
dangerous because it can cause the induction of secondary
malignancies. Prodrugs, which are activated only under cancer-
Received: May 24, 2013
© XXXX American Chemical Society
A
dx.doi.org/10.1021/jm400754c | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
Scheme 1. Structures of Reported Aminoferrocene-Based Prodrugs 1a−e,19 Their Analogues 2a, 2b, 3, 4a, and 4b, and a Positive
a
Control, 5, Which Are Described in This Article
a
The suggested activation mechanism of prodrugs 2a and 2b is given in inset A. Compound 6 is a negative control described earlier. Compounds
that are reported for the first time are indicated with red numbers.
−
hydroboration-oxidation reaction. Chang and co-workers have
utilized such reactivity of boronic acid esters to design
intracellular, fluorogenic hydrogen peroxide sensors.16 The
group of R. C. Hartley later applied this approach to prepare
“caged” 2,4-dinitrophenol (DNP) and carbonylcyanide p-
trifluoromethoxyphenylhydrazone (FCCP) derivatives. In
particular, in the presence of hydrogen peroxide these
compounds form DNP and FCCP, which exhibit biological
activity as mitochondrial uncouplers (proton translocators).17
Furthermore, the group of Peng has reported on hydrogen
peroxide-inducible DNA cross-linking agents containing a di-
(3-chloroethyl)-aminobenzyl fragment and an H2O2-reactive
arylboronic acid pinacol ester.18 Finally, we have described
aminoferrocene-based prodrugs that are activated in leukemia
cells because of the chemical reaction with reactive oxygen
induce the catalytic generation of highly active ROS like O2
and HO• from molecular oxygen and hydrogen peroxide. The
dual, synergistic effect of the reaction products (inhibition of
the antioxidative system and catalytic generation of ROS)
induces strong oxidative stress in the cells thereby causing their
death. In contrast, in normal cells lacking a high ROS
concentration, the aminoferrocene-based prodrugs remain
inactive. Cancer-specific prodrugs activated by other mecha-
nisms have been reviewed.20
Herein, we report the synthesis of five new derivatives of
aminoferrocene-based prodrugs: 2a, 2b, 3, 4a, and 4b as well as
a positive control, 5, that is activated cell-nonspecifically
because of the reaction with intracellular esterases (Scheme 1).
Our goal was to find modifications of lead structures 1a and 1e
that lead to prodrugs with improved properties including cell
membrane permeability sufficient for in vivo applications,
solubility in water, high toxicity to cancer cells, and cancer-cell
specificity. Furthermore, the toxicity of all prepared amino-
ferrocene-based prodrugs toward human promyelocytic leuke-
mia (HL-60) cells, selected primary cancer cells (chronic
lymphocytic leukemia, CLL), and related normal cells
19
−
species like O2 , H2O2, and HO•. In particular, the C−B
bond of the boronic acid ester residue of the prodrugs is
cleaved in the presence of ROS, resulting in the formation of
two cytotoxic products: a quinone methide that inhibits the
cellular antioxidative system and redox-active iron-containing
species (aminoferrocenes and iron(II) ions) that are able to
B
dx.doi.org/10.1021/jm400754c | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
(mononuclear cells, MNC) as well as Escherichia coli (E. coli)
and Streptococcus agalactiae (S. agalactiae) bacteria was
evaluated. These studies were conducted to estimate the
cancer-cell specificity of the prodrugs and to evaluate their
potential for in vivo application.
cene. In the following step, 1,1′-diaminoferrocene was reacted
at an elevated temperature (120 °C) with triphosgene, the
reaction mixture was cooled to 22 °C, and either 4-
(hydroxymethyl)phenyl boronic acid pinacol ester (synthesis
of 2a) or 4-(hydroxymethyl)-2-methylphenyl-boronic acid
pinacol ester (synthesis of 2b) was added and left reacting
for over a period of 26 h. Although all operations were
conducted under anaerobic conditions, a large amount of a
black-colored unidentifiable mixture of side products and only a
minor amount of the desired bis-substituted aminoferrocenes
were formed. After extensive column chromatography, less than
2% of the desired products could be isolated with ∼80% purity
according to 1H NMR analysis. Control experiments with 1,1′-
diaminoferrocene confirmed that this starting material was not
stable under the reaction conditions employed. Attempts to
improve the yield by optimizing the initial reaction temperature
and using alternatives to triphosgene reagents, including 1,1′-
carbonyldimidazole and 4-nitrophenyl chloroformate in combi-
nation with bases (NEt3, DIEA, DMAP, and DBU) or without
any bases, were not successful. Therefore, the alternative
protocol for synthesis of 2a and 2b was developed in which the
usage of unstable 1,1′-diaminoferrocene was avoided (route 2,
Scheme 2). In particular, in the first step, known 1,1′-
dicarboxyferrocene21 was converted to 1,1′-diisocyanatoferro-
cene.22 The following reaction with either 4-(hydroxymethyl)-
phenyl boronic acid pinacol ester or 4-(hydroxymethyl)-2-
methylphenyl boronic acid pinacol ester in CH2Cl2 furnished
correspondingly prodrugs 2a or 2b (Scheme 2). The isolated
yield of 2a was 9% and that of 2b was 18%. Both compounds
were obtained with >95% purity after a single chromatographic
purification.
RESULTS AND DISCUSSION
■
1,1′-Bis-aminoferrocene-Based Prodrugs. We assumed
that the biological effect of known aminoferrocene-based
prodrugs19 could be further potentiated by attachment of the
second arylboronic acid ester residue to the ferrocene core.
Such compounds were expected to be activated by H2O2 more
efficiently than the original prodrugs because of the presence of
twice as many reactive groups. Moreover, twice as many p-
quinone methide fragments could be potentially released from
these prodrugs than from the earlier reported amino-
ferrocenes,19 which should lead to a stronger inhibition of the
intracellular antioxidant system.
To test this hypothesis, we prepared two representative
doubly substituted ferrocenes 2a and 2b (Schemes 1 and 2).
Compound 2a is an analogue of 1a, whereas 2b is an analogue
of 1b.
Scheme 2. Two Tested Routes (1 and 2) for the Synthesis of
a
Bis-1,1′-aminoferrocene-Based Prodrugs 2a and 2b
Reaction of Bis-aminoferrocene-Based Prodrugs with
Hydrogen Peroxide In Vitro and Their Effect on Human
Promyelocytic Leukemia Cells (HL-60). It was sensible to
assume that 2a and 2b would react with H2O2 analogously to
that of the previously described monomodified aminoferrocene
1a (Scheme 1). In particular, we expected that the activation of
2a (or 2b) would be triggered by the cleavage of one or two B−
C bonds followed by the release of corresponding 1 or 2 equiv
quinone methide 8a (8b) as well as iron complexes 7a (7b) and
1,1′-diaminoferrocene (9). The latter labile iron complexes
could be spontaneously converted to iron ions. For the
detection of these iron ions, we used a 2,2′-bipyridine assay
described earlier (Table 1).19 In short, a prodrug (0.1 mM) is
dissolved in aqueous buffer (pH 7.5) and treated with H2O2 (1
mM) for 100 min. According to the chromatographic data,
prodrugs 1a, 1e, 2a, and 2b are fully transformed into products
under these experimental conditions. After the treatment with
H2O2, sodium dithionite is added to convert Fe3+ ions into Fe2+
ions. Finally, 2,2-bipyridine (bipy) is added to bind the metal
ions, resulting in the formation of a red-colored [Fe(bipy)3]2+
complex. The amount of this compound is quantified using
UV−vis spectroscopy. We observed that the yield of iron ions
released from prodrugs 2a and 2b does not exceed 20%, which
indicates the formation of other iron-containing species as the
major products (entries 2 and 3, Table 1). Under the same
conditions, parent prodrug 1a (entry 1) liberates only iron ions,
whereas N-benzyl substituted prodrug 1e releases only 9% of
iron ions (entry 5, Table 1). In our previous publication, we
provided spectroscopic evidence that over 90% of 1e is
transformed into stable N-benzylaminoferrocene under the
oxidative conditions. The possible stable products resulting
from 2a and 2b could be intermediates 7a, 7b, and 1,1′-
a
(a) CF3CO2H, 0 °C; (b) triphosgene, for the synthesis of 2a: 4-
(hydroxymethyl)phenyl-boronic acid pinacol ester, for the synthesis of
2b: 4-hydroxymethyl)-2-methylphenyl-boronic acid pinacol ester, 90
°C (c) route 1, C2O2Cl2; route 2, NaN3, (n-Bu)4NBr; (d) 100 °C,
toluene; and (e) for the synthesis of 2a: 4-(hydroxymethyl)phenyl-
boronic acid pinacol ester, for the synthesis of 2b: 4-(hydroxymethyl)-
2-methylphenyl-boronic acid pinacol ester.
Prodrug 2b is derived from 2a by the formal substitution of
two protons at the R′ position (Scheme 1) for two methyl
groups. This compound was expected to be more lipophilic
than 2a and have a correspondingly better membrane
permeability and increased toxicity. An analogous effect was
observed earlier for related monomodified aminoferrocenes 1a
and 1b.19
Synthesis. We first attempted to prepare bis-substituted
compounds analogously to their monofunctionalized analogues
by replacing aminoferrocene for 1,1′-diaminoferrocene (route
1, Scheme 2).19 Because the latter compound is rather unstable,
it was prepared immediately before the next reaction step by
acidic deprotection of 1,1′-di(tert-butoxycarbonylamino)-ferro-
C
dx.doi.org/10.1021/jm400754c | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
Next, a prodrug or one of the control compounds (0.1 mM)
was added, and after 37 min the increase in the fluorescence
intensity at 531 nm (λex = 501 nm) was determined.
Representative kinetic data for a selected prodrug and the
control compounds are shown in Figure 1. We defined the
Table 1. Efficiency of Activation of Aminoferrocene
Prodrugs and Control Compounds In Vitro
efficacy of Fe release (% of efficacy of ROS release (% of
a
b
entry
drug
1a
FeSO4)
FeSO4)
1
2
3
4
5
6
7
8
9
95
19
10
100
9
2
7
6
87
66
23
106
53
1
5
3
2a
2b
3
5
7
1e
4
1
4
4a
15
12
1
4
4
22 19
4b
6
23
3
5
10
FeSO4
100
100
a
The amount of iron ions released from prodrugs (0.1 mM) in the
presence of H2O2 (1 mM) in aqueous N-morpholinopropane-sulfonic
acid buffer (MOPS, 100 mM, pH 7.5) for 100 min was determined
using a 2,2-bipyridine-based assay.19 The amount of iron released from
FeSO4 (0.1 mM) was used as a reference (100%). The standard
Figure 1. Monitoring of the generation of ROS in the presence of a
representative prodrug, 2b, a negative control, 6, and a positive
control, FeSO4, using fluorescence spectroscopy in combination with a
ROS-sensitive fluorogenic reagent, 2′,7′-dichlorodihydrofluorescein
(DCDFH). The fluorescence (λex= 501 nm and λem= 531 nm) of a
mixture of H2O2 (10 mM), glutathione (GSH, 5 mM), ethylenedi-
amine tetracetic acid (EDTA, 10 mM), and DCDFH (0.1 mM) was
monitored for the first 2 min. The Fe-containing complex (0.1 mM)
was then added (the addition time is indicated with an arrow) and the
fluorescence was monitored for a further 40 min.
b
deviations of these values are <7% ROS release efficacy was
determined using the following equation: 100% × (F − F0)/
(F(FeSO4) − F0), where F is the fluorescence of the mixture of
H2O2 (10 mM), glutathione (GSH, 5 mM), ethylenediamine tetracetic
acid (EDTA, 10 mM), 2′,7′-dichlorodihydrofluorescein (DCDFH, 0.1
mM), and prodrug (0.1 mM), F(FeSO4) is the fluorescence of the
same mixture where the prodrug was substituted for FeSO4 (0.1 mM),
and F0 is the fluorescence of the mixture lacking any iron-containing
substance. The fluorescence intensity (λex= 501 nm and λem= 531 nm)
was measured 40 min after the addition of the iron source to the
solution of other components. The standard deviations of these values
are <9%
efficacy of ROS generation by a particular prodrug as a ratio
100% × (F(prodrug) − F0)/(F(FeSO4) − F0), where
F(prodrug) and F(FeSO4) are the fluorescence intensities
obtained in the presence of a prodrug and FeSO4 (0.1 mM),
respectively, and F0 is the fluorescence of the mixture lacking
any iron-containing substance. These values correspond to the
percent activity with respect to the positive control, FeSO4
(Table 1). We observed that compounds 2a and 2b exhibited
diaminoferrocene (Scheme 1). Because the latter compound
was available in our laboratory, we conducted an exploratory
experiment in which it was subjected to the conditions of the
2,2′-bipyridine assay for 100 min. We observed that 1,1′-
aminoferrocene was completely decomposed with the for-
mation of iron ions. These data indicate that 1,1′-amino-
ferrocene cannot be formed as a major product as a result of the
H2O2-induced activation of 2a and 2b. Therefore, it was
sensible to assume that >80% of 2a and 2b are converted into
the cleavage products of one 4-methylphenylboronic acid
pinacol ester residue, 7a, and 7b, respectively. The formation of
compound 7a in solutions of prodrug 2a (1 μM) containing
H2O2 (0.1−1 mM) was experimentally confirmed by ESI-mass
spectrometry (data not shown). We speculate that 7a and 7b
are more stable than 1,1′-diaminoferrocene and aminoferrocene
because of the electron-acceptor effect of a carbamate residue.
Iron-containing species released from the prodrugs in the
presence of H2O2, which include aminoferrocenes 7a or 7b
(>80%) and iron ions (10−19%, entries 2 and 3, Table 1), are
capable of catalyzing the generation of highly reactive oxygen
66
5 and 23
3% of the activity of FeSO4, respectively,
which is comparable to that of 1e (24%) and 1a (56%, Table
1). According to these data, doubly modified ferrocenes 2a and
2b could potentially exhibit anticancer activity analogously to
that of previously studied 1a and 1e. To evaluate whether this is
indeed a case, we studied the cell toxicity of the new prodrugs
toward human promyelocytic leukemia cells (HL-60). This cell
line was selected because data on the effects of the parent
monomodified aminoferrocenes are available for these cells,19
allowing comparisons to be made between the known and new
prodrugs. Doubly modified prodrugs 2a (IC50 = 20 1 μM)
and especially 2b (IC50 = 14 5 μM) were found to be more
toxic to HL-60 cells than the parent prodrug 1a (IC50 = 52
3
μM). This was the result that we expected because we initially
assumed that both boronic acid residues could be cleaved from
2a and 2b under the cancer-specific conditions (high
concentration of H2O2), thereby inducing the formation of 2
eqquiv of p-quinone methide 8 and 1 equiv of iron ions per 1
equiv of prodrug, as shown in inset A of Scheme 1. In contrast,
monofunctional prodrugs can generate only 1 equiv of p-
quinone methide 8 and 1 equiv of iron ions. However, the
spectroscopic and mass spectrometric data described above
indicate that only one boronic acid residue is cleaved in the
presence of cancer-specific H2O2 concentrations (<100 μM)
from 2a, 2b, and 1a. We found that the toxicity trend correlates
with the membrane permeability of the prodrugs rather than
with their reactivity. In particular, 2a was found to be 2.3-fold
and 2b, 4.1-fold more membrane permeable than 1a (Table 2).
−
species like HO• and O2 from less reactive H2O2 and
molecular oxygen. These products are especially toxic and can
cause cell death. We observed earlier that the efficacy of the
generation of such ROS in cells correlates with cytotoxicity of
aminoferrocene-based prodrugs.19 Therefore, the ability of new
prodrugs 2a and 2b to induce the formation of the ROS was
evaluated. In this experiment, we used nonfluorescent 2′,7′-
dichlorodihydrofluorescein (DCDFH), which is converted into
−
a fluorescent product in the presence of HO• and O2 that
allows for the detection of the latter species using fluorescence
spectroscopy. In a typical experiment,19 a solution of H2O2 (10
mM), glutathione (GSH, 5 mM), ethylenediamine tetracetic
acid (EDTA, 10 mM), and DCDFH (0.1 mM) was prepared.
D
dx.doi.org/10.1021/jm400754c | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
antarctika immobilized on acrylic resign,24 and LiI in 2,6-
lutidine.25 The highest yield of the desired product (39%) was
obtained under Corey’s conditions (LiOH).
Table 2. Cellular Membrane Permeability, Efficacy of ROS
Release, and Toxicity of Aminoferrocene-Based Prodrugs
and Control Compounds towards Human Promyelocytic
Leukemia Cells (HL-60)
Although the solubility of prodrug 3 in aqueous buffered
solution was improved with respect to that of 1a and the in
vitro iron- and ROS-releasing properties of 3 were found to be
favorable (entries 1 and 4, Table 1), the effects of 3 in cells
were substantially weaker than those of 1a (entries 1 and 4,
Table 2). In particular, the former prodrug generates 3.3 times
less ROS in HL-60 cells than 1a and correspondingly its
toxicity toward HL-60 cells is lower (IC50 = 68 vs 52 μM for
1a). The diminished activity of 3 in cells could be explained by
its low cell-membrane permeability (entry 4, Table 2), which is
probably caused by its negative charge at physiological pH.
Therefore, overall, the attachment of the carboxylic acid
function has a negative influence on the properties of
aminoferrocene-based prodrugs. We are currently using the
carboxylic acid group in compound 3 as an anchor to attach
variable structural fragments to obtain aminoferrocene-based
prodrugs with improved water solubility, cancer cell targeting,
and reactivity toward ROS.
membrane
permeability
efficacy of ROS
IC
50b
(μM)
a
entry drug
release
1
2
3
4
5
6
7
8
1a
2a
2b
3
1.0
1.0
52
3
1
5
2.3 0.7
4.1 1.8
0.4 + 0.1
1.8 0.3
1.2 0.2
3.4 2.5
1.3 0.2
c
20
14
0.3 0.1
1.6 0.4
1.1 0.3
1.1 0.3
0.1 0.1
68 15
1e
4a
4b
6
9
2
12
21
1
4
>200
a
Efficacy of ROS release was defined as F/F(1a), where F is the mean
fluorescence of the cells (λex= 488 nm and λem= 530 nm, determined
by flow cytometry) loaded with DCDFH (10 μM) and treated with
prodrug (100 μM) for 4.5 h. The mean fluorescence obtained for
b
prodrug 1a was used as a reference (1.0). IC50 values were
c
determined using a propidium iodide-based assay.19 An accurate
determination of ROS-release was not possible because of the high
toxicity of prodrug 2b under the experimental conditions employed.
N-Benzylaminoferrocene-Based Prodrugs. From the
compounds that we described in our first report on
aminoferrocene-based prodrugs (Scheme 1),19 N-benzyl-
substituted prodrug 1e exhibited the highest activity against
HL-60 cells (IC50 = 9 μM) and excellent cancer-cell selectivity.
On the basis of the experimental data, we concluded that this
high activity was both because of the formation in cancer cells
of a relatively stable, ROS-generating catalyst N-benzylamino-
ferrocene and p-quinone methide and because of the high cell-
membrane permeability of the prodrug. In an attempt to
improve further the cell-membrane permeability of these
compounds, we prepared more hydrophobic prodrugs 4a and
4b containing 4-methyl and 4-ethyl substituents at the N-benzyl
residue. These compounds were prepared analogously to 1e
except that 4-methylbenzaldehyde (in the synthesis of 4a) or 4-
ethylbenzaldehyde (in the synthesis of 4b) were used in place
of benzaldehyde. In contrast to our expectations, the cell-
membrane permeability of parent compound 1e was slightly
higher than that of p-methyl-substituted prodrug 4a (entries 5
and 6, Table 2). Moreover, although the permeability of p-
ethyl-substituted prodrug 4b was found to be better than that
of 1e, the large standard deviation of this parameter ( 2.5)
does not allow for its accurate comparison with the
permeability of other prodrugs. Furthermore, the ROS-
generation in the presence of 1e was found to be more
efficient than that in the presence of either 4a or 4b (entries 5−
7, Table 1). We speculate that deactivation of the latter
prodrugs can be caused by their aggregation in aqueous
solution. In agreement with the in vitro properties of 4a and 4b,
their toxicity toward HL-60 cells and their ROS-generation
ability in these cells is diminished with respect to those of 1e
(entries 5−7, Table 2). On the basis of these data, we can
conclude that the introduction of alkyl substituents at the para-
position of the N-benzyl fragment does not improve the
anticancer properties of the aminoferrocene-based prodrugs.
It should be mentioned that within the experimental error
the toxicities of all newly and earlier prepared aminoferrocene-
based prodrugs correlate with the efficiency of ROS-generation
in cells, as shown in the plot in Figure 2.
In agreement with the higher toxicity and membrane
permeability of 2a, the former prodrug induces the generation
of 1.3-fold more ROS than 1a in cells (entry 3, Table 2).
Unfortunately, we could not accurately determine the ROS
amount released in cells treated with 2b because this
compound is too toxic under the experimental conditions
applied.
1-Amino-1′-carboxyferrocene-Based Prodrug. Amino-
ferrocene-based prodrugs prepared earlier19 are not very
soluble in aqueous solutions buffered at pH 7. Therefore, all
assays were conducted in solutions containing 0.1−1%
cosolvent(e.g., N,N-dimethylformamide (DMF) or dimethyl
sulfoxide (DMSO)). To improve the solubility of the prodrugs,
we introduced a carboxylic acid substituent into parent
structure 1a to obtain prodrug 3. The latter compound was
synthesized starting from commercially available 1-amino-1′-
methoxycarbonylferrocene hydrochloride, which was first
converted to 1-isocyanato-1′-methoxycarbonyl-ferrocene by
neutralization with triethylamine followed by the reaction
with triphosgene. The isocyanate and 4-(hydroxymethyl)-
phenylboronic acid pinacol ester were then coupled together
to obtain the methyl ester of the prodrug (3Me, Scheme 3).
The last step of the synthesis included the hydrolysis of the
ester, which was optimized to minimize the possible side
reaction of the cleavage of the boronic acid ester fragment in
3Me. In particular, the following conditions were tested: NaOH
in THF/water, LiOH in MeOH/H2O,23 Lipase B from Candida
a
Scheme 3. Synthesis of Water-Soluble Prodrug 3
These data may indicate that all aminoferrocene-based
anticancer prodrugs studied to date act via a related mechanism
that relies on the catalytic generation of ROS in cancer cells.
a
(a) (1) NEt3; (2) triphosgene, toluene and (b) LiOH, MeOH, H2O.
E
dx.doi.org/10.1021/jm400754c | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
Table 3. Toxicity of Aminoferrocene-Based Prodrugs
towards Chronic Lymphocytic Leukemia (CLL) Cells and
Mononuclear Cells (MNC)
number of viable MNC/number of
a
IC50 SD (μM)
viable CLL cells
prodrug
CLL
MNC
5 μM prodrug 7.5 μM prodrug
1a
1b
1c
2.2 2.1
2.8 2.2
3.9 2.1
1.5 2.2
1.8 2.1
1.4 2.1
6.0 2.1
2.6 2.2
3.1 2.1
1.0 2.0
>10
>10
7.2
14.5
12.0
6.5
>10
5.3
>10
3.3
c
1e
>10
11.1
10.3
12.4
1.6
12.5
12.3
14.8
2.5
c
2a
>10
c
2b
>10
3
>10
Figure 2. Correlation of the toxicity (IC50, μM) of the known19 and
described in this Article aminoferrocene-based prodrugs with their
efficiency in the generation of ROS in cells. The efficiency of ROS
generation is expressed in relative units (r.u.), which was determined
relative to the efficiency obtained for prodrug 1a.
4a
4b
5
>10
3.8
8.2
>10
4.1
9.8
5.5 2.6
>10
4.5
0.9
6
1.3
1.3
b
Fc
>10
>10
1.0
1.0
a
b
c
SD, standard deviation. Fc, ferrocene. Indicates the prodrugs that
exhibit most favorable properties.
However, more experimental data are required to confirm this
suggestion.
ferrocenes 1e, 2a, and 2b at a concentration of 5 μM and for 1a,
1b, 1e, 2a, and 2b at a concentration of 7.5 μM (Figure 3).
Activity of Aminoferrocene-Based Prodrugs toward
Chronic Lymphocytic Leukemia (CLL) and Human
Mononuclear Cells (MNC). As described above and in our
previous publications,19 aminoferrocene-based prodrugs are
toxic toward a variety of cancer cell lines. As the next step in the
development of these compounds, we decided to evaluate their
activity toward primary cancer cells, which are isolated directly
from cancer patients. We selected CLL cells as an example of
cancer cells and MNC’s as an example of normal cells. MNC’s
as well as CLL cells were both isolated from peripheral blood.
Cancerous CLL cells are derived from B-cells, which, together
with T-cells and monocytes, are present in MNC’s. Therefore,
the CLL and MNC pair is a good model to study the cancer
specificity of new drugs. Another reason for the selection of
CLL was the fact that intracellular ROS concentrations in these
cells were reported to be increased. Finally, CLL responds to
As2O3 treatment, which is known to increase the ROS amount
in cells and thereby cause their death.26 Therefore, it was
sensible to assume that our ROS-modulating aminoferrocene-
based prodrugs will be efficiently activated in CLL cells.
CLL cells were isolated from four patients, whereas MNC’s
were isolated from six healthy donors. The toxicity of selected
aminoferrocene-based prodrugs at five different concentrations
in the range between 1 and 10 μM was evaluated using the
ATP-based CellTiter Glo Luminescence Cell Viability Assay.
These data were used to estimate the IC50 values and cancer
specificity of the prodrugs (Table 3). In particular, we observed
that the aminoferrocene-based prodrugs were not toxic toward
MNC’s at concentrations ≤10 μM, whereas their toxicity
toward CLL cells was significant, with IC50 values in the low-
micromolar range (1.4−6.0 μM) (Table 3). As expected, stable
ferrocenes 6 and Fc exhibit toxicity neither toward normal
(MNC) nor cancerous cells (CLL), whereas positive control 5,
which is activated nonspecifically both in normal and cancer
cells because of intracellular esterase activity, exhibits significant
toxicity toward both cells types (Table 3). To quantify and
compare the effect of the prodrugs toward cancer and normal
cells, we defined the cancer-cell-specificity factor as a ratio of
the number of viable MNC’s to the number of viable CLL cells
that were treated with a prodrug of a particular concentration
(Table 3). This factor was found to exceed 10 for amino-
Figure 3. Relative viability of CLL cells and MNC’s expressed as a
percent of the control (untreated cells) treated with different prodrugs
(7.5 μM) for 48 h.
The highest specificity of 14.8 was observed for prodrug 2b
at a concentration of 7.5 μM. In contrast, the related catalytic
organochalcogene-based prodrugs, which were prepared and
investigated by the research groups of C. Jacob and M. Herling,
exhibited a cancer-cell specificity (CLL/MNC) that did not
exceed 4 in similar assay.15 Moreover, although the latter
compounds exhibited significant anticancer activity at a
somewhat lower concentration (0.5 μM) than our best
prodrugs, 1e, 2a, and 2b (IC50 = 1.4−1.8 μM and selectivity
= 10.3−14.8, Table 3), their toxicity toward normal cells
(MNC’s) was also found to be higher.
To explore the toxicity of aminoferrocene-based prodrugs
toward normal cells (MNC’s) in more detail, we have studied
the effects of a representative prodrug, 1e, on MNC’s at a high
concentration (10 μM) and at longer incubation times (48−72
h). We observed that 76 8% of viable MNC’s survive 48 h of
incubation in the presence of this prodrug, whereas a 72 h
incubation causes a further reduction in the number of viable
normal cells by a factor of 2. These data indicate that at high
concentrations and prolonged incubation times aminoferro-
cene-based prodrugs can be also activated in normal cells,
which are known to contain low amounts of reactive-oxygen
species.
F
dx.doi.org/10.1021/jm400754c | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
Effects of Selected Aminoferrocene-Based Prodrugs
on Bacteria. Some bacteria are important for the normal
function of the human body. For example, harmless strains of E.
coli are found in the gut of all healthy humans and assist the
body, for example, by protecting it from other bacteria. It is
important that new drugs/prodrugs do not affect these bacteria.
Because bacteria are known to accumulate metal ions like iron,
they can exist under a higher ROS load than normal cells of the
human body. Therefore, bacteria can be potentially affected by
the aminoferrocene-based prodrugs, which would be an
undesired side effect. To test whether this is indeed the case,
we investigated the toxicity of several selected prodrugs on
Gram-negative E.coli: 1a, 1e, and 2a were tested as well as a
positive control, ampicillin, and a negative control, ferrocene.
We observed that only ampicillin at 0.9 mM caused an
inhibitory effect. In particular, a zone of inhibition was observed
only for this compound (Figure S1, Supporting Information).
In contrast, the bacteria were resistant to prodrugs 1a, 1e, and
2a as well as the negative control at concentrations of ≤0.9
mM. Higher concentrations were not tested because of limited
solubility. Additionally, we tested the same group of prodrugs
and controls on S. agalactiae, which are Gram-positive bacteria
present in the normal gastrointestinal flora of some humans.
Consistent with the former experiments, we observed that the
prodrugs are not active at concentrations of up to 0.9 mM.
EXPERIMENTAL SECTION
■
General Information. Commercially available chemicals of the
best quality from Aldrich/Sigma/Fluka (Germany) were obtained and
used without purification. Prodrugs 1a and 1e as well as control 6 were
prepared as described previously.25 NMR spectra were acquired on a
Bruker Avance DRX 200, Bruker Avance II 400, or Bruker Avance III
600 spectrometer. ESI mass spectra were recorded on an ESI
MicroTOF (Bruker), FAB mass spectra, on a Jeol JMS-700 instrument
using p-nitrobenzyl alcohol as a matrix, and EI mass spectra, on a
Finnigan MAT 8200 instrument. C, H, and N analysis was performed
in the microanalytical laboratory of the chemical institute of the
University of Heidelberg. For analytical reversed-phase thin-layer
chromatography, Polygramm TLC plates (Macherey-Nagel) were
used. UV−vis spectra were acquired on a Varian Cary 100 Bio UV−vis
spectrophotometer using 1 cm optical path black-wall absorption
semimicrocuvettes (Hellma GmbH, Germany) with a sample volume
of 0.7 mL. Fluorescence spectra were acquired on a Varian Cary
Eclipse fluorescence spectrophotometer using black-wall fluorescence
semimicrocuvettes (Hellma GmbH) with a sample volume of 0.7 mL.
The fluorescence of live HL-60 cells was quantified using an Accuri C6
flow cytometer. The data were processed using the CFLow Plus
(Accuri) software package. The purity of the prodrugs used in the
biological tests was determined by C, H, and N analysis. According to
these data, the purity of the prodrugs and controls was greater than
95%.
Synthesis. 1,1′-Bis(azidocarbonyl)ferrocene.22 1,1-Ferrocenedi-
carboxylic acid21 (6.0 g, 21.9 mmol) was suspended in CH2Cl2 (35
mL) and purged with argon. Oxalylchloride (11.2 g, 88.2 mmol, 4.5
equiv) and N,N-dimethylformamide (DMF, 15 mL) were then added,
and the reaction mixture was mixed for 3 h at 22 °C. During this time,
two new portions of oxalylchloride (2 × 1 g, 15.8 mmol, 0.5 equiv)
were added at 1 and 2 h after the beginning of the reaction. Excess
oxalylchloride and solvents were removed under vacuum (0.01 mbar),
and the residue was dissolved in CH2Cl2 (140 mL) and mixed with (n-
Bu)4NBr (17.9 g, 55.6 mmol, 2.5 equiv). A solution of NaN3 (4.8 g,
73.1 mmol, 3.3 equiv) in water (40 mL) was then added dropwise, and
the resulting mixture was stirried for a further 18 h at 22 °C. Next,
water (100 mL) was added, the organic phase was separated, and the
aqueous phase was washed with CH2Cl2 (3 × 20 mL). The organic
phases were joined together, dried under MgSO4, and filtered. Finally,
the solvent was removed under vacuum using a rotary evaporator, and
the product was purified by column chromatography on SiO2 using
CH2Cl2 as an eluent. The yield of the product was 4.6 g (65%). TLC
CONCLUSIONS
■
We explored the effects of a few structural modifications of
aminoferrocene-based prodrugs. In particular, we found that
the attachment of a carboxylic acid substituent at the 1′-
position of the ferrocene fragment improves the water solubility
of the prodrug. However, it becomes less cell-membrane
permeable. The synthesized carboxylic acid group-containing
prodrug can be potentially applied as an intermediate for the
further conjugation of fragments to improve water solubility,
cancer targeting and specificity, and membrane permeability.
The introduction of alkyl substituents at the para position of
the N-benzylic residue of the corresponding N-substituted
aminoferrocene-based prodrug does not improve the anticancer
effect. In contrast, the attachment of the second arylboronic
acid ester fragment to the parent prodrug enhances its activity
significantly, from IC50 = 52 μM to 14−20 μM. Moreover, we
demonstrated that the aminoferrocene-based prodrugs are
active not only toward cancer cell lines but also toward primary
cancer cells (CLL; IC50 = 1.4−1.8 μM for the best prodrugs).
Importantly, they practically do not affect the corresponding
normal cells (mononuclear cells, MNC) at concentrations <7.5
μM and incubation times ≤48 h. We estimated that the cancer-
cell specificity reaches 14.8-fold in the best case. However, at
high concentrations (10 μM) and prolonged incubation times
(72 h), aminoferrocene-based prodrugs can be also activated in
normal cells, which are known to contain low amounts of
reactive-oxygen species. Finally, we observed that representative
bacterial cells, including E.coli and S. agalactiae, which populate
the gastrointestinal flora of humans and are required for the
normal function of the human body, are not affected by the
aminoferrocene-based prodrugs at concentrations up to 0.9
mM. These data are indicative of the high cancer specificity of
these prodrugs, which can potentially make them suitable for in
vivo applications.
1
(SiO2, CH2Cl2 eluent) Rf = 0.50; H NMR (CDCl3, 399.89 MHz) δ
4.56 (s, 4H), 4.90 (s, 4H); EI−MS, pos, m/z: calcd for
[C12H8N6O2Fe-e−]+, 324.0; found, 324.2
Prodrug 2a. 1,1′-Bis(azidocarbonyl)ferrocene (0.8 g, 2.5 mmol)
was dissolved in toluene (150 mL) and purged with argon. The
solution was heated to reflux and kept under these conditions for 2.5 h
to obtain 1,1′-diisocyanatoferrocene. Next, the solution was cooled to
22 °C, 4-(hydroxymethyl)phenyl-boronic acid pinacol ester (1.3 g, 5.7
mmol, 2.3 equiv) in CH2Cl2 (80 mL) was added, and the resulting
mixture was stirred for 22 h at 22 °C. Finally, the volatiles were
removed using a rotary evaporator, and the crude product was purified
by column chromatography on SiO2 using petroleum ether/ethyl-
acetate mixture (2:1 v/v) as an eluent. The resulting oil was
resuspended in ethylacetate, and the solid product formed was filtered
and dried. The yield of the analytically pure product was 162 mg (9%).
TLC (SiO2, n-hexane/ethyl acetate 2:1 v/v eluent) Rf = 0.38; 1H NMR
(DMSO-d6, 399.89 MHz) δ 1.28 (s, 24H), 3.86 (s, 4H), 4.40 (s, 4H),
5.08 (s, 4H), 7.36 (d, 3J = 7.3 Hz, 4H), 7.66 (d,3J = 7.8 Hz, 4H), 8.83
(s, 1H); 13C NMR (DMSO-d6, 150.90 MHz) δ 24.83, 59.93, 61.06,
64.81, 65.29, 83.84, 126.95, 130.18, 134.06, 138.62, 140.51, 153.71;
FAB−MS, pos, m/z: calcd for [C38H46N2O8B2Fe-e−]+, 736.3; found,
736.5; Anal. Calcd for C38H46N2O8B2Fe: C, 61.99; H, 6.30; N, 3.80.
Found: C, 61.79; H, 6.38; N, 3.61.
Prodrug 2b. This compound was synthesized analogously to that of
prodrug 2a except that 4-(hydroxymethyl)-2-methylphenyl-boronic
acid pinacol ester was used in place of 4-(hydroxymethyl)-
G
dx.doi.org/10.1021/jm400754c | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
phenylboronic acid pinacol ester as a starting material. The yield of the
(7 mL), triphosgene (bis(trichloromethyl)carbonate, 0.60 g, 2 mmol)
was added, and the mixture was heated to reflux and kept under these
conditions for 1.5 h. The solution obtained was cooled to 22 °C, and
4-(hydroxymethyl)phenylboronic acid pinacol ester (0.46 g, 2 mmol)
in toluene (5 mL) was added. The mixture was again heated to reflux
and left stirring under these conditions for 15 h. After cooling to 22
°C, more triphosgene (1.2 g, 4 mmol) was added, and the mixture was
stirred at 22 °C for a further 20 h. Finally, all volatiles were removed
under vacuum (0.01 mbar), and the residue was purified by column
chromatography on SiO2 using n-hexane/acetone mixture (95:5 v/v)
as an eluent. The yield of the analytically pure product was 0.19 g
analytically pure product was 185 mg (18%). TLC (SiO2, n-hexane/
1
ethyl acetate 2:1 v/v eluent) Rf = 0.48; H NMR (acetone-d6, 600.13
MHz) δ 7.85 (bs, 1H), 7.74−7.69 (d, 3J = 7.4 Hz, 2H), 7.18 (bs, 4H),
5.09 (s, 4H), 4.51 (s, 4H), 3.93, (s, 4H), 2.51 (s, 6H), 1.33 (s, 12H);
13C NMR (acetone-d6, 150.91 MHz) δ 154.93, 145.73, 140.81, 137.02,
129.72, 129.09, 124.74, 98.03, 84.26, 66.61, 65.65, 62.49, 25.21, 22.51;
EI−MS, pos, m/z: calcd for [C40H50N2O8B2Fe-e−]+, 764.3; found,
764.5; Anal. Calcd for C40H50N2O8B2Fe·0.5 hexane: C, 63.97; H, 7.12;
N, 3.47. Found: C, 63.71; H, 6.70; N, 3.04.
1-Amino-1′-methoxycarbonylferrocene. 1-Amino-1′-methoxyca-
bonyl-ferrocene hydrochloride (2.0 g, 6.8 mmol) dissolved in
ethylacetate (150 mL) was slowly neutralized with NEt3 (2.4 mL,
17.0 mmol, 2.5 equiv). The resulting mixture was washed with water
(10 × 30 mL), and the combined aqueous phases were washed with
ethylacetate (2 × 30 mL). The organic phases were combined and
dried over MgSO4, the solvent was removed under vacuum (50 mbar)
in a rotary evaporator, and the solid product was dried under vacuum
(0.01 mbar) to obtain 1.8 g (99%) of the desired product. TLC (SiO2,
petroleum ether (30-75)/ethyl acetate/NEt3 7.5:2.5:0.5 v/v/v eluent)
Rf = 0.21; 1H NMR (acetone-d6, 399.89 MHz) δ 3.73 (s, 3H), 3.80 (t,
2H), 3.93 (t, 2H), 4.31 (t, 2H), 4.66 (t, 2H).
4-(4, 4, 5, 5-Tetramethyl-1, 3, 2-dioxaborolan-2-yl)-
benzyloxycarbonylamino-ferrocene (Methyl Ester of Prodrug 3,
3Me). 1-Amino-1′-methoxy-carbonylferrocene (2.3 g, 9 mmol) was
dissolved under Ar in toluene (100 mL). Triphosgene (2.4 g, 8 mmol,
0.9 equiv) was added, and the reaction mixture was heated to reflux
and kept under these conditions for 90 min. Next, the mixture was
allowed to cool to 22 °C, a solution of 4-(hydroxymethyl)-
phenylboronic acid pinacol ester (1.9 g, 8 mmol, 0.9 equiv) in toluene
(23 mL) was added, and the solution was mixed for 92 h at 22 °C.
After the removal of the solvent using a rotary evaporator, the product
was purified by column chromatography on SiO2 using pethroleum
ether (30−75)/ethylacetate initially as an eluent. The content of
ethylacetate in this eluent was gradually increased during the
chromatography up to 100%. The yield of the product was 3.7 g
(79%). TLC (SiO2, petroleum ether (30-75)/ethyl acetate 3:1 v/v
eluent) Rf = 0.28; 1H NMR (acetone-d6, 399.89 MHz) δ 1.33 (s, 12H),
3.70 (s, 3H), 3.99 (s, 2H), 4.38 (s, 2H), 4.66 (s, 2H), 4.73 (s, 2H),
5.19 (s, 2H), 7.44 (d, 2H), 7.76 (d, 2H), 8.05 (s, 1H).
1
(17%). TLC (SiO2, n-hexane/acetone 3:1 v/v eluent) Rf = 0.5; H
NMR (acetone-d6, 399.89 MHz) δ 1.33 (s, 12H), 2.31 (s, 3H), 3.97 (s,
2H), 4.11 (s, 5H), 4.46 (s, 2H), 4.96 (s, 2H), 5.24 (s, 2H), 7.18 (m,
3
4H), 7.38 (m, 2H), 7.73 (d, J = 7.0 Hz, 2H); 13C NMR (acetone-d6,
100.55 MHz) δ 21.14, 25.27, 53.88, 63.27, 65.09, 67.84, 69.81, 84.64,
127.35, 129.99 (2 overlapping peaks), 135.66 (2 overlapping peaks),
136.89, 137.20, 141.00; FAB−MS, pos, m/z: calcd for
[C32H36NO4BFe-e−]+, 565.2; found, 565.2; Anal. Calcd for
C32H36NO4BFe: C, 67.99; H, 6.42; N, 2.48. Found: C, 68.34; H,
6.87; N, 2.38.
Prodrug 4b. This compound was synthesized analogously to that of
prodrug 4a except that 4-ethylbenzaldehyde was used in place of 4-
methylbenzaldehyde as a starting material. The yield of the analytically
pure product was 78 mg (13%). TLC (SiO2, n-hexane/acetone 5:1 v/v
1
eluent) Rf = 0.47; H NMR (acetone-d6, 399.89 MHz) δ 7.74 (d, 3J =
3
7.8 Hz, 2H), 7.38 (d, J = 7.9 Hz, 2H), 7.20 (m, 4H), 5.24 (s, 2H),
4.97 (s, 2H), 4.45 (bs, 2H), 4.11 (s, 5H), 3.96 (s, 2H), 2.62 (q, 3J = 7.5
Hz, 2H), 1.33 (s, 12H), 1.21 (t, 3J = 7.5 Hz, 3H); 13C NMR (acetone-
d6, 100.55 MHz) δ 143.71, 140.97, 137.12, 135.65, 128.88, 128.81,
128.01, 127.83, 127.41, 84.62, 69.78, 67.84, 67.59, 65.07, 63.27, 53.86,
29.13, 25.27, 16.24; ESI−MS, pos, m/z: calcd for [C33H38NO4BFe-
e−]+, 579.2; found, 578.8; Anal. Calcd for C33H38NO4BFe·1.5 hexane:
C, 71.19; H, 8.39; N, 1.98. Found: C, 71.12; H, 8.19; N, 2.14.
4-(Hydroxymethyl)phenylacetate. 4-Acetoxybenzoic acid (2.5 g,
13.9 mmol) was suspended in tetrahydrofurane (THF, 18 mL) and
cooled to 10 °C. A solution of BH3·THF complex in THF (1 M, 30
mL, 30 mmol) was added dropwise over 10 min time. The mixture was
allowed to warm to 22 °C and was left stirring under these conditions
for 36 h. The mixture was then cooled to 0 °C, and water (11 mL) was
slowly added. After the removal of the solvent under vacuum (0.01
mbar), the residue was resuspended in a water/ethyl acetate mixture
(140 mL, 1:1 v/v), the organic phase was separated and washed with
water (70 mL), a saturated aqueous NaCl solution (3 × 70 mL), and
again with water (2 × 70 mL), dried under Na2SO4, and filtered.
Finally, the solvent was removed using a rotary evaporator, and the
residue was purified by column chromatography on SiO2 using a
CH2Cl2/ethyl acetate/triethylamine (TEA) mixture (10:4:0.5 v/v/v)
as an eluent. The yield of the product was 0.9 g (41%). TLC (SiO2,
CH2Cl2/ethyl acetate/TEA 10:4:0.5 v/v/v eluent) Rf = 0.41; 1H NMR
Prodrug 3. Methyl ester 3Me (2.0 g, 3.9 mmol) dissolved in MeOH
(250 mL) was mixed with solution of LiOH (2.6 g, 62 mmol, 16.1
equiv) in water (53 mL), and the reaction mixture was stirred at 22 °C
for 90 min. After the removal of the solvent under vacuum, the
remaining solid was suspended in a solvent mixture consisting of
ethylacetate (250 mL) and acetic acid (4 mL). This mixture was
washed with water (20 mL), and the solvent was removed using a
rotary evaporator. The product was purified by column chromatog-
raphy on SiO2 using pethroleum ether (30-75)/ethylacetate (7:1 v/v)
initially as an eluent followed by pethroleum ether (30−75)/
ethylacetate (3:1 v/v) and pethroleum ether (30-75)/ethylacetate
(3:1 v/v) containing 0.5% acetic acid as eluents. The yield of the
product was 0.8 g (39%). TLC (SiO2, petroleum ether (30−75)/ethyl
3
(acetone-d6, 199.92 MHz) δ 2.24 (s, 3H), 4.21 (t, J = 5.8 Hz, 1H),
3
3
3
4.62 (d, J = 5.6 Hz, 2H), 7.06 (d, J = 8.6 Hz, 2H), 7.38 (d, J = 8.7
Hz, 2H).
1
acetate 1:1 v/v containing 5% acetic acid eluent) Rf = 0.58; H NMR
4-(N-(Ferrocenylamino)carbonyloxymethyl)phenylacetate, Con-
trol 5. Aminoferrocene (0.4 g, 2 mmol) and triphosgene (0.59 g, 2
mmol) in toluene (25 mL) were heated to reflux and kept under these
conditions for 1 h. This led to dissolution of all reagents. Afterwords,
the mixture was allowed to cool to 22 °C, 4-(hydroxymethyl)phenyl
acetate (0.33 g, 2 mmol) in toluene (25 mL) was added, and the
resulting mixture was stirred for 70 h at 22 °C. Finally, the volatiles
were removed using a rotary evaporator, and the crude product was
purified by column chromatography on SiO2 using petroleum a ether
(30-75)/ethylacetate mixture (7:1 to 7:3 v/v) as an eluent. The
product was washed several times with n-hexane and dried. The yield
of the pure product was 0.11 g (14%). TLC (SiO2, n-hexane/acetone
(acetone-d6, 600.13 MHz) δ 1.33 (s, 12H), 4.00 (s, 3H), 4.38 (s, 2H),
4.65 (s, 2H), 4.73 (s, 2H), 5.18 (s, 2H), 7.43 (d, 2H), 7.75 (d, 2H),
8.09 (s, 1H); 13C NMR (acetone-d6, 150.90 MHz) δ 24.43, 61.17,
65.68, 65.78, 70.93, 72.23, 72.37, 83.67, 98.41, 126.77, 134.53, 134.71,
140.41, 153.67, 171.19; FAB−MS, pos, m/z: calcd for
[C25H28NO6BFe-e−]+, 505.1; found, 505.3; Anal. Calcd for
C25H28NO6BFe·1/3 acetic acid: C, 58.70; H, 5.63; N, 2.67. Found:
C, 58.85; H, 5.68; N, 2.59.
Prodrug 4a. Aminoferrocene (0.40 g, 2 mmol), 4-methylbenzalde-
hyde (0.24 g, 2 mmol), and MeOH (10 mL) were mixed and brought
to reflux. The reagents were allowed to react for 2 h under these
conditions. The mixture was allowed to cool to 22 °C, NaB(CN)H3
(0.13 g, 2 mmol) in MeOH (10 mL) was added, and the mixture was
stirred for 30 min. Next, aqueous HCl (1 M, 2 mL) was slowly added,
and the volatiles were removed using a rotary evaporator and then
under vacuum (0.01 mbar). The residue was resuspended in toluene
1
7:3 v/v eluent) Rf = 0.37; H NMR (CDCl3, 399.89 MHz) δ 2.31 (s,
3H), 4.00 (s, 2H), 4.17 (s, 5H), 4.50 (s, 2H), 5.15 (s, 2H), 5.88 (s,
3
3
1H), 7.11 (d, J = 8.4 Hz, 2H), 7.42 (d, J = 8.4 Hz, 2H); 13C NMR
(CDCl3, 100.55 MHz) δ 21.11, 60.93, 64.51, 66.35, 69.18, 77.20,
121.74, 129.45, 133.83, 150.56, 169.39; FAB−MS, pos, m/z: calcd for
H
dx.doi.org/10.1021/jm400754c | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
[C20H19NO4Fe-e−]+, 393.1; found, 393.0; Anal. Calcd for
C20H19NO4Fe·0.5 ethylacetate: C, 60.43; H, 5.30; N, 3.20. Found:
C, 60.54; H, 4.86; N, 3.38.
Cellular Assays. Cells and Cell Culture. The human promyelocytic
leukemia cell line (HL-60) and primary CLL and MNC’s were
cultured in Roswell Park Memorial Institute (RPMI) 1640 medium
supplemented with 10% FCS and 5 μg/mL of penicillin/streptomycin
(all media and supplements were from Gibco Invitrogen Corp.,
Karlsruhe, Germany).
ROS Detection in Cells and Determination of the Viability of the
Cells. Intracellular ROS and the viability of HL-60 cells were
monitored using flow cytometry as described previously.19 The
viability of CLL cells and MNC’s was determined by CellTiter Glo
Luminescence Cell Viability Assay (Promega, Fitchburg, USA).
REFERENCES
■
(1) Sounni, N. E.; Noel, A. Targeting the tumor microenvironment
for cancer therapy. Clin. Chem. 2012, 59, 85−93.
(2) (a) Halliwell, B. Oxidative stress and cancer: Have we moved
forward? Biochem. J. 2007, 401, 1−11. (b) Engel, R. H.; Evens, A. M.
Oxidative stress and apoptosis: A new treatment paradigm in cancer.
Front. Biosci. 2006, 11, 300−312. (c) Finkel, T. Oxidant signals and
oxidative stress. Curr. Opinion Cell Biol. 2003, 15, 247−254.
(d) Schumacker, P. T. Reactive oxygen species in cancer cells: Live
by the sword, die by the sword. Cancer Cell 2006, 10, 175−176.
(e) Antunes, F.; Cadenas, R. Estimation of H2O2 gradients across
biomembranes. FEBS Lett. 2000, 475, 121−126. (f) Szatrowski, T. P.;
Nathan, C. F. Production of large amounts of hydrogen peroxide by
human tumor cells. Cancer Res. 1991, 51, 794−798. (f) O’Donnell-
Tormey, J.; DeBoer, C. J.; Nathan, C. F. Resistance of human tumor
cells in vitro to oxidative cytolysis. J. Clin. Invest. 1985, 76, 80−86.
(g) Stone, J. R. An assessment of proposed mechanisms for sensing
hydrogen peroxide in mammalian systems. Arch. Biochem. Biophys.
2004, 422, 119−124 ROS levels are especially interesting for
anticancer drug design because they seem to be a general feature of
cancer.
ASSOCIATED CONTENT
■
S
* Supporting Information
Protocols and data for the effects of prodrugs and controls on
E. coli and S. agalactiae. Protocols for determining the octanol−
water partition coefficients and membrane permeability of
aminoferrocene-based prodrugs. This material is available free
of charge via the Internet at http://pubs.acs.org.
(3) Davison, K.; Mann, K. K.; Miller, W. H. Arsenic trioxide:
Mechanisms of action. Semin. Hematol. 2002, 39, 3−7.
(4) (a) Sun, S. Y.; Li, W.; Yue, P.; Lippman, S. M.; Hong, W. K.;
Lotan, R. Mediation of N-(4-hydoxyphenyl)retinamide-induced
apoptosis in human cancer cells by different mechanisms. Cancer
Res. 1999, 59, 2493−2498. (b) Batra, S.; Reynolds, C. P.; Maurer, B. J.
Fenretinide cytotoxicity for Ewing’s sarcoma and primitive neuro-
ectodermal tumor cell lines is decreased by hypoxia and synergistically
enhanced by ceramide modulators. Cancer Res. 2004, 64, 5415−5424.
(5) (a) Gao, J.; Liu, X.; Rigas, B. Nitric oxide-donating aspirin
induces apoptosis in human colon cancer cells through induction of
oxidative stress. Proc. Natl. Acad. Sci. U.S.A. 2005, 102, 17207−17212.
(b) Kashfi, K.; Borgo, S.; Williams, J. L.; Chen, J.; Gao, J.; Glekas, A.;
Benedini, F.; der Soldato, P.; Rigas, B. Positional isomerism markedly
affects the growth inhibition of colon cancer cells by nitric-oxide-
donating aspirin in vitro and in vivo. J. Pharmacol. Exp. Therap. 2005,
312, 978−988. (c) Hulsman, N.; Medema, J. P.; Bos, C.; Jongejan, A.;
Leurs, R.; Smit, M. J.; de Esch, J. P.; Richel, D.; Wijtmans, M.
Chemical insights in the concept of hybrid drugs: The antitumor effect
of nitric oxide-donating aspirin involves a quinone methide but not
nitric oxide nor aspirin. J. Med. Chem. 2007, 50, 2424−2431.
(6) Dorr, R. T.; Liddil, J. D.; Soble, M. J. Cytotoxic effects of
glutathione synthesis inhibition by L-buthionine-(SR)-sulfoximine on
human and murine tumor cells. Invest. New Drugs 1986, 4, 305−313.
(7) Hersch, E. M.; Gschwind, C. R.; Taylor, C. W.; Dorr, R. T.;
Taetle, R.; Salmon, S. E. Antiproliferative and antitumor activity of the
2-cyanoaziridine compound imexon on tumor cell lines and fresh
tumor cells in vitro. J. Nat. Cancer Inst. 1992, 84, 1238−1244.
(8) Evens, A. M.; Lecane, P.; Magda, D.; Prachand, S.; Singhal, S.;
Nelson, J.; Miller, R. A.; Gartenhaus, R. B.; Gordon, L. I. Motexafin
gadolinium generates reactive oxygen species and induces apoptosis in
sensitive and highly resistant multiple myeloma cells. Blood 2005, 105,
1265−1273.
(9) (a) Keyes, S. R.; Rockwell, S.; Sartorelli, A. C. Modification of the
metabolism and cytotoxicity of bioreductive alkylating agents by
dicoumarol in aerobic and hypoxic murine tumor cells. Cancer Res.
1989, 49, 3310−3313. (b) Chen, Q.; Cederbaum, A. I. Menadione
cytotoxicity to Hep G2 cells and protection by activation of nuclear
factor-kappa B. Mol. Pharmacol. 1997, 52, 648−657. (c) Vallis, K. A.;
Wolf, C. R. Relationship between the adaptive response to oxidants
and stable menadione-resistance in Chinese hamster ovary cell lines.
Carcinogenesis 1996, 17, 649−654. (d) Beck, R.; Verrax, J.; Dejeans,
N.; Taper, H.; Calderon, P. B. Menadione reduction by pharmaco-
logical doses of ascorbate induces an oxidative stress that kills breast
cancer cells. Int. J. Toxicol. 2009, 28, 33−42. (e) Chlebowski, R. T.;
Dietrich, M.; Akman, S.; Block, J. B. Vitamin K3 inhibition of
malignant murine cell growth and human tumor colony formation.
Cancer Treat. Rep. 1985, 69, 527−532. (f) Prasad, K. N.; Edwards-
AUTHOR INFORMATION
■
Corresponding Author
*E-mail: andriy.mokhir@fau.de. Tel: 49-09131-85-22554.
Author Contributions
∥The manuscript was written through contributions of all
authors. All authors have given approval to the final version of
the manuscript. These authors contributed equally.
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
A.M. thanks Ruprecht-Karls-University of Heidelberg and
Friedrich-Alexander-University of Erlangen-Nurnberg for finan-
̈
cial support. The following students of Ruprecht-Karls-
University of Heidelberg participating in the project in the
framework of their final practical B.S. and M.S. works or
advanced experimental practical works in organic and inorganic
chemistry are acknowledged: Peter Beck, Arthur Schneider,
Andrea Uptmoor, and Florian Gebert. R.G. and V.P. thank
Laura Saa and Gaizka Garrai for valuable advice concerning the
experiments with bacteria.
ABBREVIATIONS USED
■
bipy, 2,2-bipyridine; BSO, buthionine sulfoximine; CH2Cl2,
dichloromethane; CLL, chronic lymphocytic leukemia;
DCDFH, 2′,7′-dichlorodihydrofluorescein; DMF, N,N-dime-
thylformamide; DMSO, dimethyl sulfoxide; DNP, 2,4-dini-
trophenol; EDTA, N,N,N′,N′-ethylenediamine tetraacetic acid;
Fc, ferrocene; FCCP, p-trifluoromethoxyphenylhydra-zone;
FCS, fetal bovine serum; E. coli, Escherichia coli; GSH, reduced
glutathione; GSSG, oxidized glutathione; HL-60, human
promyelocytic leukemia cells; 8-HQ, 8-hydroxyquinoline;
MNC, mononuclear cells; MOPS, 3-(N-morpholino)-
propanesulfonic acid; MTT, 3-(4,5-dimethylthiazol-2-yl)-2,5-
diphenyltetrazolium bromide; NOASA, nitric oxide-donating
aspirin; QM, quinone methide; ROS, reactive oxygen species;
RPMI, Roswell Park Memorial Institute; S. agalactiae,
Streptococcus agalactiae; TLC, thin layer chromatography
I
dx.doi.org/10.1021/jm400754c | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
Prasad, J.; Sakamoto, A. Vitamin K3 (menadione) inhibits the growth
of mammalian tumor cells in culture. Life Sci. 1981, 29, 1387−1392.
(10) (a) Li, Y.; Sun, X.; LaMont, J. T.; Pardee, A. B.; Li, C. J.
Selective killing of cancer cells by beta-lapachone: Direct checkpoint
activation as a strategy against cancer. Proc. Natl. Acad. Sci. U.S.A.
2003, 100, 2674−2678. (b) Pink, J. J.; Planchon, S. M.; Tagliarino, C.;
Varnes, M. E.; Siegel, D.; Boothman, D. A. NAD(P)H: Quinone
oxidoreductase activity is the principal determinant of beta-lapachone
cytotoxicity. J. Biol. Chem. 2000, 275, 5416−5424.
investigation of new macrocyclic diphosphine-palladium(0) complexes
based on the barbiturate binding receptor. Organometallics 2002, 21,
5243−5253.
(22) Van Leusen, D.; Hessen, B. 1,1′-Diisocyanoferrocene and a
convenient synthesis of ferrocenylamine. Organometallics 2001, 20,
224−226.
́
(23) Corey, E. J.; Szekely, I.; Shiner, C. S. Synthesis of 6,9α-oxido-
11α, 15α-dihydroxyprosta-(E)5, (E)13-dienoic isomer of PGI2 (vane’s
PGX). Tetrahedron Lett. 1977, 18, 3529−3532.
(24) Anderson, E. M.; Larsson, K. M.; Kirk, O. One biocatalyst −
many applications: The use of Candida antarctica B-Lipase in organic
synthesis. Org. Biomol. Chem. 2010, 8, 539−545.
(11) Bair, J. S.; Palchaudhuri, R.; Hergenrother, P. J. Chemistry and
biology of deoxynyboquinone, a potent inducer of cancer cell death. J.
Am. Chem. Soc. 2010, 132, 5469−5478.
(25) McMurry, J. Ester hydrolysis via SN2-type dealkylation. Chapter 2
in Organic Reactions, Vol. 24, Ed.-in-Chief Dauben, W. G., John-Wiley
& Sons, 1977; pp 188−224.
(12) Trachootham, D.; Alexandre, J.; Huang, P. Targeting cancer
cells by ROS-mediated mechanisms: a radical therapeutic approach?
Nature Rev. Drug Discov. 2009, 8, 579−591.
(26) Merkel, O.; Heyder, C.; Asslaber, D.; Hamacher, F.; Tinhofer, I.;
(13) Hecht, S. M. Bleomycin: New perspectives on the mechanism of
action. J. Nat. Prod. 2000, 63, 158−168.
Holler, C.; Stocher, M.; Prokesch, A.; Papak, C.; Scheideler, M.;
̈
Trajanoski, Z.; Greil, R. Arsenic trioxide induces apoptosis
preferentially in B-CLL cells of patients with unfavorable prognostic
factors including del17p13. J. Mol. Med. 2008, 86, 541−552.
̀
(14) (a) Hillard, E.; Vessieres, A.; Thouin, L.; Jaouen, G.; Amatore,
C. Ferrocene-mediated proton-coupled electron transfer in a series of
ferrocifen-type breast-cancer drug candidates. Angew. Chem., Int. Ed.
̀
2006, 45, 285−290. (b) Vessieres, A.; Top, S.; Pigeon, P.; Hillard, E.;
Boubeker, L.; Spera, D.; Jaouen, G. Modification of the estrogenic
properties of diphenols by the incorporation of ferrocene. Generation
of antiproliferative effects in vitro. J. Med. Chem. 2005, 48, 3937−3940.
́
̀ ́
(c) Plazuk, D.; Vessieres, A.; Hillard, E. A.; Buriez, O.; Labbe, E.;
Pigeon, P.; Plamont, M.-A.; Amatore, C.; Zakrzewski, J.; Jaouen, G. A
[3]ferrocenophane polyphenol showing a remarkable antiproliferative
activity on breast and prostate cancer cell lines. J. Med. Chem. 2009, 52,
4964−4967.
(15) (a) Lilienthal, N.; Prinz, C.; Peer-Zada, A. A.; Doering, M.; Ba,
L. A.; Hallek, M.; Jacob, C.; Herling, M. Targeting the disturbed redox
equilibrium in chronic lymphocytic leukemia by novel reactive oxygen
species-catalytic “sensor/effector” compounds. Leuk. Lymphoma 2011,
52, 1407−1411. (b) Doering, M.; Ba, L. A.; Lilienthal, N.; Nicco, C.;
Scherer, C.; Abbas, M.; Peer Zada, A. A.; Coriat, R.; Burkholz, T.;
Wessjohann, L.; Diederich, M.; Batteux, F.; Herling, M.; Jacob, C.
Synthesis and selective anticancer activity of organochalcogen based
redox catalysts. J. Med. Chem. 2010, 53, 6954−6963.
(16) (a) Miller, E. W.; Albers, A. E.; Pralle, A.; Isacoff, E. Y.; Chang,
C. J. Boronate-based fluorescent probes for imaging cellular hydrogen
peroxide. J. Am. Chem. Soc. 2005, 127, 16652−16659. (b) Dickinson,
B. C.; Chang, C. J. A. A targetable fluorescent probe for imaging
hydrogen peroxide in the mitochondria of living cells. J. Am. Chem. Soc.
2008, 130, 9638−9639. (c) Miller, E. W.; Tulyathan, O.; Isacoff, E. Y.;
Chang, C. J. Molecular imaging of hydrogen peroxide produced for cell
signaling. Nat. Chem. Biol. 2007, 3, 263−267. (d) Srikun, D.; Miller, E.
W.; Domaille, D. W.; Chang, C. J. A. An ICT-based approach to
ratiometric fluorescece imaging of hydrogen peroxide produced in
living cells. J. Am. Chem. Soc. 2008, 130, 4596−4597.
(17) Quin, C.; Robertson, L.; McQuaker, S. J.; Price, N. C.; Brand,
M. D.; Hartley, R. C. Caged mitochondrial uncouplers that are
released in response to hydrogen peroxide. Tetrahedron 2010, 66,
2384−2389.
(18) (a) Kuang, Y.; Balakrishnan, K.; Gandhi, V.; Peng, X. Hydrogen
peroxide inducible DNA cross-linking agents: Targeted anticancer
prodrugs. J. Am. Chem. Soc. 2011, 133, 19278−19281. (b) Cao, S.;
Wang, Y.; Peng, X. ROS-inducible DNA cross-linking agent as a new
anticancer prodrug building block. Chem.Eur. J. 2012, 18, 3850−
3854.
(19) Hagen, H.; Marzenell, P.; Jentzsch, E.; Wenz, F.; Veldwijk, M.
R.; Mokhir, A. Aminoferrocene-based prodrugs activated by reactive
oxygen species. J. Med. Chem. 2012, 55, 924−934.
(20) (a) Graf, N.; Lippard, S. J. Redox activation of metal-based
prodrugs as a strategy for drug delivery. Adv. Drug Delivery Rev. 2012,
64, 993−1004. (b) Wolkenberg, S. E.; Boger, D. L. Mechanisms of in
situ activation for DNA-targeting antitumor agents. Chem. Rev. 2002,
102, 2477−2495.
(21) Sørensen, H. S.; Larsen, J.; Rasmussen, B. S.; Laursen, B.;
Hansen, S. G.; Skrydstrup, T.; Amatore, C.; Jutand, A. Synthesis and
J
dx.doi.org/10.1021/jm400754c | J. Med. Chem. XXXX, XXX, XXX−XXX