The synthesis and characterization of a series of cobalt(II) β-ketoaminato complexes and their cytotoxic activity towards human tumor cell lines

Article abstract of DOI:10.1016/j.jinorgbio.2011.03.005

A series of square planar cobalt(II) compounds bearing tetradentate β-ketoaminato ligands with variation in the number of -CF₃ ligand substituents has been prepared and structurally and spectroscopically characterized. The fluorinated β-ketoami

Full text of DOI:10.1016/j.jinorgbio.2011.03.005

Journal of Inorganic Biochemistry 105 (2011) 858–866
Contents lists available at ScienceDirect
Journal of Inorganic Biochemistry
journal homepage: www.elsevier.com/locate/jinorgbio
The synthesis and characterization of a series of cobalt(II) β-ketoaminato complexes
and their cytotoxic activity towards human tumor cell lines
a,
Lydia Gurley a, Natalia Beloukhina b, Kalun Boudreau b, Andis Klegeris b, , W. Stephen McNeil
⁎
⁎
a
Department of Chemistry, University of British Columbia Okanagan, 3333 University Way, Kelowna, British Columbia, Canada, V1V 1V7
Department of Biology, University of British Columbia Okanagan, 3333 University Way, Kelowna, British Columbia, Canada, V1V 1V7
b
a r t i c l e i n f o
a b s t r a c t
Article history:
A series of square planar cobalt(II) compounds bearing tetradentate β-ketoaminato ligands with variation in
the number of ―CF3 ligand substituents has been prepared and structurally and spectroscopically
characterized. The fluorinated β-ketoamine ligands were prepared utilizing a multistep reaction sequence
employing a silylenol protecting group. An additional tetrahedral cobalt compound bearing two bidentate β-
ketoaminato ligands was also prepared and characterized.
Received 10 November 2010
Received in revised form 16 February 2011
Accepted 14 March 2011
Available online 22 March 2011
Cytotoxic activity of the cobalt-containing complexes was evaluated using six human cell lines; including two
different prostate cancer cell lines (PC-3 and VCaP), acute monocytic leukemia (THP-1), astrocytoma (U-373
MG), hepatocellular carcinoma (HepG2), and neuroblastoma (SH-SY5Y) cells. The cobalt compounds are
more active than their corresponding ligands. The activity is cell type specific; the cobalt compounds exhibit
strong activity against human prostate cancer and monocytic leukemia cells but weak or no activity against
neuroblastoma, astrocytoma, and liver carcinoma cells. Activity generally increases with a greater number of
―CF3 substituents, and square planar complexes exhibit greater activity than the tetrahedral derivative. The
mechanisms of activity against human PC-3 prostate cancer cells involve caspase-3 and two different
mitogen-activated protein kinases. The addition of a thiol antioxidant reduced cytotoxicity, suggesting the
possible involvement of reactive oxygen species. These cobalt complexes may represent a novel class of
cytotoxic drugs selective towards certain types of tumors.
Keywords:
Cobalt complexes
Fluorinated ligands
Cytotoxic activity
Caspase-3
MAP kinases
Reactive oxygen species
© 2011 Elsevier Inc. All rights reserved.
1. Introduction
carbonyldicobalt [11]. Its mode of activity against cancer cells is
unknown but it has been suggested that these complexes may interact
Inorganic and bioinorganic medicinal agents have made a growing
contribution to medical science and human health in the past half
century [1]. Arguably, the most widely known inorganic therapeutic
agent is cis-diamminedichloroplatinum(II) (cisplatin) and its de-
rivatives carboplatin and oxaliplatin. Cisplatin and carboplatin have
been in widespread use for many years in the treatment of several
forms of cancer, including ovarian, cervical, head and neck, and non-
small-cell lung cancers [2]. However, it presents the greatest success
against testicular cancer for which the overall cure rate exceeds 90%,
and is nearly 100% for early-stage disease [3]. The success of these
compounds as solid tumor therapeutics has fostered a great deal of
research into the development and improvement of platinum-based
drugs [4–8]. In contrast, and despite the known biological roles of
cobalt as found in cobalamins [9] and cobalt nitrile hydratases [10],
relatively few studies have examined the potential therapeutic
properties of cobalt compounds. Ott et al. have reported the synthesis
and antitumor efficacy of [2-acetoxy-(2-propynyl)benzoate] hexa-
with DNA or DNA-related cellular biochemical pathways [12]. A
number of cobalt(III) complexes have been examined for their
antiviral properties, with the most significant development being
the approval of Doxovir as a topical microbicide for the treatment of
herpes simplex virus type 1 [13].
Although organofluorine compounds are virtually absent as
natural products, 20–25% of drugs in pharmaceutical development
contain at least one fluorine atom [14]. Introduction of fluorine atoms
into a structure can dramatically change the physical and chemical
properties of a complex, including its potential pharmaceutical
behavior [15]. In this paper we present a series of cobalt compounds
bearing primarily tetradentate bis(β-ketoaminato) ligands, with
variation in the number of ―CF3 ligand substituents. The tetradentate
ligands enforce a square-planar geometry for these complexes. One
cobalt compound bears two bidentate ligands and a tetrahedral
geometry, permitting an analysis of the effect of complex geometry on
the cytotoxic activity profile. The proligands and their corresponding
cobalt complexes are presented in Fig. 1.
Preliminary experiments aimed at evaluating possible biological
activities of the newly synthesized compounds indicated cytotoxic
activity of the cobalt-containing complexes. More detailed studies
⁎
Corresponding authors. Tel.: +1 250 807 8751; fax: +1 250 807 8005.
E-mail addresses: andis.klegeris@ubc.ca (A. Klegeris), s.mcneil@ubc.ca
(W.S. McNeil).
0162-0134/$ – see front matter © 2011 Elsevier Inc. All rights reserved.
doi:10.1016/j.jinorgbio.2011.03.005

L. Gurley et al. / Journal of Inorganic Biochemistry 105 (2011) 858–866
859
[19–21]. Co(N(SiMe3)2)2·THF was prepared according to the pub-
lished procedure [22]. All reactions were carried out in an inert
atmosphere glove box or on a vacuum/nitrogen line using standard
air-sensitive Schlenk techniques.
The following substances were used in various assays and were
obtained from Sigma-Aldrich: diaphorase (EC 1.8.1.4, from Clostrid-
ium kluyveri); dimethyl sulfoxide (DMSO); sodium L-lactate; sodium
oxamate; p-iodonitrotetrazolium violet; NAD+; 3-(4,5-dimethylthia-
zol-2-yl)-2,5-diphenyltetrazolium bromide (MTT); N-acetyl-L-cyste-
ine (NAC), an antioxidant; SP600125, a selective inhibitor of c-Jun
NH2-terminal kinase (JNK); and PD98059, a selective cell-permeable
inhibitor of MAP kinase (MAPK)/extracellular signal-regulated kinase
(ERK) kinase (MEK)1/2. An Alexis caspase-3 fluorometric assay kit
was purchased from Enzo Life Sciences (Plymouth Meeting, PA, USA).
2.1.1. Characterization of compounds
1H and 19F NMR data were collected on a Varian 400 MHz NMR
spectrometer at room temperature in J. Young valve NMR tubes. 1H
NMR chemical shift values were calibrated relative to residual protio-
solvent signals, and 19F NMR signals were referenced to external
CF3COOH. X-ray crystallographic analyses were performed by Brian
Patrick at the University of British Columbia, Vancouver, BC, Canada.
Elemental analysis was performed by Guelph Chemical Laboratories
of Guelph, Ontario, Canada.
Fig. 1. Structures of β-ketoaminato ligands and corresponding cobalt complexes
described in this study.
were designed to establish the half-maximal effective concentration
(EC50) values for the cobalt compounds against six different human
tumor cell lines including PC-3 (prostate), VCaP (prostate), THP-1
(acute monocytic leukemia), U-373 MG (astrocytoma), HepG2 (liver
carcinoma) and SH-SY5Y (neuroblastoma). We also measured the
activity of the ligands used in preparation of the metal complexes.
These experiments aimed to determine: 1) whether there was a
difference in the cytotoxic activity between the organic ligands and
their corresponding metal complexes; 2) whether alterations in the
peripheral ligand substituents had an effect on the activity of the
compounds; and 3) whether the activity associated with the different
compounds was cell type specific. We performed an additional series
of experiments to elucidate the molecular mechanisms underlying the
cytotoxic activity of the cobalt-containing drugs. It has been reported
previously that anti-tumor drug action is often associated with
apoptotic processes involving activation of caspases [16], mitogen-
activated protein (MAP) kinases [16,17] and/or generation of reactive
oxygen intermediates [18]. This article presents experimental evi-
dence supporting involvement of caspase-3, two different MAP
kinases, and reactive oxygen intermediates in the cytotoxic activity
of cobalt-containing drugs toward human PC-3 prostate adenocarci-
noma cells.
2.1.2. Cell culture
The following human cell lines were obtained from the American
Type Culture Collection (ATCC, Manassas, VA, USA): PC-3 prostate
adenocarcinoma; VCaP prostate cancer epithelial cell line; THP-1
acute monocytic leukemia; U-373 MG astrocytoma; and HepG2
hepatocellular carcinoma. The human neuroblastoma SH-SY5Y cell
line was a gift from Dr. R. Ross, Fordham University, NY. Cells were
grown in Dulbecco's modified Eagle's medium-nutrient mixture F12
ham (DMEM-F12) supplemented with 10% fetal bovine serum (FBS)
supplied by Thermo Scientific HyClone (Logan, UT, USA). All cell lines
were used without initial differentiation.
2.2. Syntheses
2.2.1. Synthesis of bis(hexafluoroacetylacetone)ethylenediimine (L2H2)
L2H2 was prepared via a modification of a previously described
procedure [23]. In a glovebox, hexafluoroacetylacetone (7.45 g,
35.8 mmol) was cooled to −35 °C. Sodium hydride (0.859 g,
35.8 mmol) was predissolved in THF and cooled to −35 °C. The cold
hexafluoroacetylacetone and sodium hydride solutions were added
alternatively in small portions to 20 mL THF, with the mixture kept
cold with an acetone/liquid N2 bath during these additions. Extensive,
somewhat violent effervescence occurred. The mixture was allowed
to warm to ambient temperature and stirred overnight, resulting in a
clear, slightly pink mixture. The flask was removed from the glovebox
and attached to the vacuum line. tert-Butyldimethylchlorosilane
(5.40 g, 35.8 mmol) dissolved in 5 mL anhydrous THF was added to
a pressure-equalized addition funnel. The tert-butyldimethylchlor-
osilane solution was added very slowly to the deprotonated ketone
solution. After complete addition, the mixture was refluxed overnight
to give a yellow solution with a beige precipitate. The solution was
cannulated through Celite to give a clear yellow solution. In a separate
flask 1,2-diaminoethane (0.652 g, 10.9 mmol) was dissolved in 5 mL
anhydrous THF. The two separate solutions were cooled in an acetone/
N2 bath. The 1,2-diaminoethane solution was cannulated into the
protected ketone solution. After complete addition, the reaction was
allowed to warm to room temperature and resulted in a cloudy white
suspension. After 1 h of stirring, the THF was removed under reduced
pressure to give a mixture of white and pink solids. The solids were
redissolved in 20 mL methanol giving a pink solution that was
refluxed for 3 h. Cooling the solution resulted in the deposition of a
2. Experimental procedure
2.1. Materials and methods
Anhydrous solvents were purified in Glass Contour solvent
purification towers, including hexanes (ACS reagent, Fisher, Waltham,
MA, USA), diethylether (anhydrous, VWR, West Chester, PA, USA),
and tetrahydrofuran (THF, HPLC grade, Fisher). Methanol, ethanol,
and toluene were used as received (ACS, Fisher). CoCl2, sodium bis
(trimethylsilyl) amide (Na[N(SiMe3)2]), sodium hydride, tert-butyl-
dimethylchlorosilane, 1,2-diaminoethane, 4,4,4-trifluoro-1-phenyl-
1,3-butanedione, 1-phenyl-1,3-butanedione, and 1,2-phenylenedia-
mine were all purchased from Sigma-Aldrich (St. Louis, MO, USA) and
used as received. Hexafluoroacetylacetone (Sigma-Aldrich) was thrice
freeze-pump-thaw degassed before use. CDCl3 (99.6+ atom% D) was
dried over P2O5, distilled under reduced pressure, freeze-pump-thaw
degassed, and stored under dinitrogen. Bis(benzoylacetone)ethyle-
nediimine (L1H2) and the corresponding complex (bis(benzoylaceto-
nato)ethylenediimino)cobalt(II) (L1Co) were prepared by the
previously reported procedures and have been fully characterized

860
L. Gurley et al. / Journal of Inorganic Biochemistry 105 (2011) 858–866
white solid. The solution was filtered to collect 1.985 g (41.1%) of
white solid. 1H NMR (400.13 MHz, CDCl3) δ 10.58 (s, 1H), 5.87 (s, 1H),
3.79 (s, 2H). 19F NMR (376.50 MHz, CDCl3) δ −77.41 (s), −66.49 (s).
dark orange crystals. 1H NMR (400.13 MHz, CDCl3) δ 18.45, 14.74, 12.68,
10.52, 9.88, −10.95, −19.22, −26.46, −34.83, −54.72. EA analysis;
calculated: C 68.45, H 5.38, N 9.98; found C 67.43, H 5.59, N 9.38.
2.2.2. Synthesis of (benzoyltrifluoroacetone)ethylenediimine (L3H2)
L3H2 was synthesized following the same procedure as outlined for
L2H2 utilizing 4,4,4-trifluoro-1-phenyl-1,3-butanedione in place of
hexafluoroacetylacetone, to afford L3H2 as white crystals in 60.0% yield.
1H NMR (400.13 MHz, CDCl3) δ 10.93 (s, 1H), 7.46 (s, 3H), 7.18 (s, 2H),
5.44 (s, 1H), 3.43 (s, 2H). 19F NMR (376.50 MHz, CDCl3) δ −70.65 (s).
2.3. In vitro assays
2.3.1. Caspase assay
The enzymatic activity of caspase-3 was measured fluorometrically
by using an assay kit purchased from Enzo Life Sciences. The assay was
carried out according to the instructions supplied by the manufacturer
as described before [24]. PC-3 cells were seeded into 60 mm tissue
culture dishes (Corning, NY, USA) at a concentration of 1.5×105 cells/
mL in 20 mL of DMEM-F12 medium containing 5% FBS. L1Co (5–25 μM),
L2Co (1–10 μM) or DMSO aliquots were added and cells incubated for
48 h followed by cell lysis. Caspase-3 activity in cell lysates was
expressed in Fluorescence Units per g protein. Protein concentrations in
the samples were measured by using the Bradford reagent and bovine
serum albumin solutions to generate a standard curve [25].
2.2.3. Synthesis of (benzoylacetonato)phenyleneimineamine (L4H)
1-Phenyl-1,3-butanedione (2.013 g, 12.49 mmol) was dissolved in
30 mL toluene to give a yellow solution. 1,2-Phenylenediamine (0.6757 g,
6.248 mmol) was added directly and the solution was refluxed overnight
using a Dean-Stark apparatus. The resulting bright orange solution was
concentrated. Hexane was added and the mixture was placed in the
freezer at −35 °C overnight. The solution was filtered to give an orange
solid, recrystallized from ethanol to yield 1.264 g (80.23%) of yellow
crystals. 1H NMR (400.13 MHz, CDCl3) δ 12.47 (s, 1H), 7.94–7.89 (m, 2H),
7.49–7.39 (m, 3H), 7.13–7.01 (m, 2H), 6.81–6.71 (m, 2H), 5.93 (s, 1H),
3.89 (s, 2H), 1.99 (s, 3H).
2.3.2. Cell toxicity assays
Various human cell lines were seeded into 24-well plates at the
following concentrations: PC-3 at 1.5 × 105 cells/mL; VCaP at
1.5×105cells/mL; THP-1 at 5.0×105 cells/mL; SH-SY5Y at 2.0×105
cells/mL; U-373 MG at 2.0.×105 cells/mL; and HepG2 at 5.0×104 cells/
mL. 0.4–0.8 mL of DMEM-F12 medium containing 5% FBS was used to
plate out the required number of cells. All cell types, except non-
adherent THP-1 cells, were allowed to adhere for 24 h at 37 °C in a CO2
incubator (humidified 5% CO2 and 95% air atmosphere). Subsequently,
cell medium was replaced with fresh medium and cells incubated in the
presence or absence of various compounds (1–50 μM) or their vehicle
solution (DMSO) for further 48 h. Concentration of DMSO was kept at
0.5% in all samples. MAPK inhibitors (20 μM) or various concentrations
of NAC were added 15 min prior to the addition of the newly
synthesized compounds. Following the incubation, 100 μL of cell culture
media was sampled for lactate dehydrogenase (LDH) to determine the
percentage of dead cells, while the evaluation of surviving cells was
performed by the MTT assay as previously described [26].
2.2.4. Synthesis of
(bis(hexafluoroacetylacetonato)ethylenediimino)cobalt(II) (L2Co)
L2H2 (0.1678 g, 0.0291 mmol) was dissolved in 5 mL THF. In a
separate vial Co(N(SiMe3)2)2·THF (0.1104 g, 0.0291 mmol) was dis-
solved in 5 mL THF. Dropwise addition of the L2H2 solution to that of Co
(N(SiMe3)2)2·THF resulted in an immediate color change to a dark red
mixture. After approximately 2 min, orange solid precipitated from the
solution. The reaction was allowed to stir overnight. The solvent was
removed in vacuo to yield 0.3934 g (79.14%) of a very dark red solid. The
red solid was redissolved in hexanes giving a dark orange solution. The
solution was filtered through Celite then placed in the freezer at −35 °C
overnight to induce the development of analytically-pure bright orange
crystals. 1H NMR (400.13 MHz, CDCl3) δ +116.45 (s, 2H), −91.65 (s,
1H). EA analysis; calculated: C 28.99, H 1.22, N 5.64; found: C 29.40, H
1.20, N 5.96.
The MTT assay is based on the ability of viable, but not dead cells,
to convert the tetrazolium salt (MTT) to colored formazan. The
viability of SH-SY5Y cells was determined by adding MTT to the SH-
SY5Y cell cultures to reach a final concentration of 0.5 mg/mL.
Following a 1 h incubation period at 37 °C, the dark crystals which
had formed were dissolved by adding to the wells an equal volume of
SDS/DMF extraction buffer (20% sodium dodecyl sulfate, 50% N,N-
dimethyl formamide, pH 4.7). Subsequently, the plates were placed
overnight at 37 °C, after which 100 μL aliquots were transferred to 96-
well plates and optical densities at 570 nm were measured using a
microplate reader. The viable cell value was calculated as a percentage
of the value obtained from cells incubated with fresh medium only.
LDH activity in cell culture supernatants was measured by an
enzymatic test in which formation of the formazan product of
iodonitrotetrazolium dye was followed colorimetrically. 100 μL of cell
culture supernatant from each sample was pipetted into a 96-well plate,
followed by addition of 15 μL lactate (36 mg/mL) and 15 μL p-iodoni-
trotetrazolium violet (2 mg/mL) solutions into each well. The enzymatic
reaction was started by addition of 15 μL of NAD+/diaphorase solution
(3 mg/mL NAD+; 2.3 mg solid/mL diaphorase). After a 15–30 min
incubation period, the reaction was terminated by the addition of 15 μL
of oxamate (16.6 mg/mL) into each well. Optical densities at 490 nm were
measured by a microplate reader. The amount of LDH released was
expressed as a percentage of the value obtained in comparative wells in
which cells were completely lysed by 1% Triton X 100.
2.2.5. Synthesis of (bis(benzoyltrifluoroacetonato)ethylenediimino)cobalt(II)
(L3Co)
L3H2 (0.4953 g, 1.085 mmol) was dissolved in 5 mL diethylether
resulting in a creamy white suspension. In a separate vial, Co(N
(SiMe3)2)2·THF (0.3172 g, 1.096 mmol) was dissolved in 2 mL diethy-
lether. Dropwise addition of the L3H2 suspension to the Co(N(SiMe3)2)2·
THF solution resulted in an immediate color change to light orange, which
then grew progressively darker orange. The solution was stirred
overnight and the solvent was removed in vacuo to yield 0.4937 g
(88.65%) of orange solid. The solid was redissolved in diethylether and
filtered through Celite. The solution was placed in the freezer at −35 °C to
induce deposition of analytically-pure red crystals. 1H NMR (400.13 MHz,
acetone-d6) δ +100.97 (s, 1H), +29.41, +15.37, +9.55, −62.82 (s, 2H).
EA analysis; calculated: C 51.47, H 3.14, N 5.46; found C 51.45, H 2.96, N
5.38.
2.2.6. Synthesis of [(benzoylacetonato)phenyleneimineamino]cobalt (II)
(L4Co)
L4H (0.4185 g, 1.659 mmol) was dissolved in 10 mL diethylether,
resulting in a yellow solution. Co(N(SiMe3)2)2·THF (0.304 g,
0.802 mmol) was dissolved in 5 mL diethylether and the resulting dark
green solution was added dropwise to the ligand solution. The mixture
immediately changed to a dark orange color with a dark orange
precipitate. The mixture was stirred overnight. The solvent was removed
in vacuo to yield 0.4757 g (85.08%) of orange solid. The solid was
redissolved in THF/hexanes and filtered through Celite. The filtrate was
placed in the freezer at −35 °C to induce the formation of X-ray quality
2.3.3. Statistical analysis
Data are presented as means standard error of the mean (SEM).
The concentration-dependent effects of various compounds in vitro

L. Gurley et al. / Journal of Inorganic Biochemistry 105 (2011) 858–866
861
were evaluated statistically by the randomized block design analysis
of variance (ANOVA) followed by Fisher's least significant differences
(LSD) post hoc test. The P values obtained by paired Student's t test
were corrected for multiple comparisons by Holm's step-down
method [27].
3. Results and discussion
3.1. Ligand synthesis
The non-fluorinated proligand L4H (see Fig. 1) is easily prepared
via the Schiff base condensation of 1-phenyl-1,3-butanedione and 1,2-
diaminoethane in refluxing toluene, using a Dean-Stark apparatus.
However, attempts to prepare fluorinated β-ketoimines in which one
or two of the R groups are fluoroalkyl groups have been shown to be
unsuccessful via Schiff base condensation reactions [23,28], so the
synthesis of the fluorinated proligands L2H2 and L3H2 cannot employ
this method. It has been suggested that this synthetic approach is not
viable due to the high acidity of the fluorinated β-diketones, which
leads to protonation of the diamine and formation of a diketonoate
ammonium salt, preventing the desired condensation reaction [28]. A
synthesis has been reported in which sublimation of the ammonium
salt does afford the L2H2 complex, but the product is isolated in low
yield due to formation of a cyclic byproduct in which both diketone
oxygen atoms have reacted with the same diamine molecule [29]. A
different synthetic methodology for the preparation of the desired
fluorinated proligands was therefore adopted [23].
To overcome the problem associated with the protonation of the
diamine, the appropriate diketone was first deprotonated using NaH
under anhydrous conditions to produce [R1COCHCOR2]−Na+. This salt
was subsequently treated with tert-butyldimethylchlorosilane to
produce a silylenolether, which is treated with 1,2-diaminoethane
to form the desired fluorinated β-ketoimine. The silyl moiety acts as a
protecting group to ensure that the amine condensation occurs at only
one of the carbonyl groups of the diketone, and thereby prevents the
formation of a cyclic product. This synthetic approach was successful
in the preparation of the proligands L2H2 and L3H2 (Scheme 1).
In an attempt to prepare the phenylene derivative of L1H2, a similar
Schiff base condensation reaction was performed reacting 1-phenyl-
1,3-butanedione and 1,2-phenylenediamine. The reaction proceeded
in a 1:1 ratio rather than a 2:1 ratio, resulting in proligand L4H and
subsequently complex L4Co (Scheme 2).
Scheme 2. Preparation of ligand L4H and synthesis of complex L4Co.
iron derivative Fe[N(SiMe3)2]3. The cobalt amido compound's THF-
free solid-state structure was determined to be a dimeric complex
with bridging amido groups: [((Me3Si)2NCo)2(μ-N(SiMe3)2)2] [30].
However, if the synthesis is conducted in THF, the resulting complex is
monomeric with THF coordinated as a donor solvent [31]. The π-
donor properties of the amido groups, combined with their steric bulk
at distance from the metal center, afford a reactive complex with a
coordinatively and sterically accessible metal center [32], and ligands
especially amenable to protonolysis reactions.
Direct addition of the proligands L2H2, L3H2, and L4H to Co(N
(SiMe3)2)2·THF in THF results in an immediate color change from
green to red, usually followed by precipitation of the corresponding
ketoaminato cobalt complex from solution and isolation in high (80–
90%) yields. The complexes are red to red-orange, and dissolve readily
in THF or diethylether.
The solid state molecular structures of L3Co and L4Co have been
determined, and are shown in Figs. 2 and 3 respectively. Structure and
refinement data are listed in Table 1. The structures and observed
metrical parameters are similar to those of other known cobalt(II)
complexes with related ligands [20,21,33,34]. L3Co exhibits a square
planar geometry, with all the Co bond angles between 86.2 and 94.2°.
There is a slight (8.5°) twisting of the N―Co―N and O―Co―O planes,
presumably to accommodate the steric proximity of the phenyl rings
3.2. Cobalt complex syntheses
L1Co is readily prepared via the established procedure by adding
L1H2 to simple cobalt salts such as CoCl2 or Co(OAc)2 in the presence
of base [19–21]. Such an approach is not successful for L2Co, L3Co, and
L4Co. A new synthetic approach was adopted using the diamide salt Co
(N(SiMe3)2)2·THF. Burger and Wannagat [22] reported the successful
preparation of the cobalt amido compound, along with the trivalent
Fig. 2. Solid state molecular structure of L3Co. Hydrogen atoms have been omitted for
clarity. Selected bond lengths (Å); Co(1)―N(1) 1.8666, Co(1)―N(2) 1.8581, Co(1)―O
(1) 1.8597, and Co(1)―O(2) 1.8609. Selected bond angles (°) N(2)―Co(1)―O(1)
173.91, N(2)―Co(1)―O(2) 93.66, O(1)―Co(1)―O(2) 86.23, N(2)―Co(1)―N(1)
86.54, O(1)―Co(1)―N(10) 94.16, and O(2)―Co(1)―N(1) 174.48.
Scheme 1. Preparation of ligands L2H2 and L3H2.

862
L. Gurley et al. / Journal of Inorganic Biochemistry 105 (2011) 858–866
geometries. The selective modifications to the ligand substituents
across a series of related compounds, combined with reliable
integration values, permit a high degree of certainty in the assignment
of signals despite their paramagnetic shifts. A comparison of the 1H
NMR spectra of L1Co, L2Co, and L3Co is shown in Fig. 4, along with peak
assignments.
L1Co bears a methyl substituent absent from L2Co and L3Co. The
presence of the signal integrating for 6H at −20.8 ppm in the spectra
of L1Co, without a corresponding signal in the other two spectra, can
therefore be assigned as that of the CH3 groups. A weak signal
common to all three compounds appears consistently between −60
and −95 ppm, and can be assigned to the ligand CH proton.1 Other
assignments can be determined similarly, and are listed in Table 2.
Partial assignment of the spectrum of L4Co can be made via
comparison to those of other related tetrahedral bis(β-ketoaminato)
cobalt(II) complexes [34].
Fig. 3. Solid state molecular structure of L4Co. Hydrogen atoms have been omitted for
clarity. Selected bond lengths (Å); Co(1)―O(1) 1.9183(18) and Co(1)―N(1) 1.972(2).
Selected bond angles (°) N(1)―Co(1)―O(1) 94.67(8), O(1)―Co―O(1*) 109.82(12), N
(1)―Co―N(1*) 115.89(13), and O(1)―Co―N(1*) 121.92(8).
3.3. In vitro assays
The cytotoxic activities of four different cobalt complexes and their
corresponding ligands were tested. Table 3 summarizes the data
obtained by using six human cell lines of different origin, while Fig. 5
illustrates effects of two ligand/complex pairs on the PC-3 prostate
cancer cell line. PC-3 cells were incubated for 48 h with the
experimental drugs or DMSO vehicle. Viability of cells was assessed
by the MTT (Fig. 5A) and the LDH (Fig. 5B) assays. With increasing
concentration, both cobalt complexes L1Co and L2Co caused a
concentration-dependent decrease in cell viability (Fig. 5A) and an
increase in percentage of lysed cells (Fig. 5B). The corresponding
ligands L1H2 and L2 H2 had no effect according to both of these assays.
MTT assay results were selected for presentation in Table 3 due to
the superior sensitivity of the assay (Fig. 5). Cells undergoing
apoptosis during the initial prelytic stages release only low levels of
LDH [35], which leads to underestimation of cell death according to
the LDH assay (see Fig. 5 for example). Additional experiments were
also performed to rule out possible interaction between reagents used
in the LDH and MTT assays and the compounds (data not shown).
Table 3 shows the estimated EC50 values according to the MTT
assay as well as the percentage of maximal inhibition observed at the
concentration range of the compounds studied (1–50 μM). Data from
experiments where statistically significant reduction of cell viability
was observed are listed. The EC50 values were estimated as the
concentration of a compound that induced toxicity halfway between
the baseline and the maximum cytotoxic effect of a compound
observed, which varied between 12% and 94% for the compounds
studied. It is important to note that since only four different
concentrations of the compounds were used, both the EC50 and
maximal inhibition values shown in Table 3 are estimates, and more
accurate values would require more detailed studies using additional
concentration points. Similar to the effects of L1Co illustrated in Fig. 5,
LDH measurements often indicated higher EC50 values; however,
effects of the compounds were similar in both LDH and MTT assays.
Additionally, in several cases cells were counted under a microscope
to confirm cytotoxicity of the compounds.
and the CH2CH2 bridge. (A corresponding twist is absent in the related
complex with R1 =CH3 and R2 =Ph, where the phenyl groups are
proximal to the O donor atoms rather than the N [33].) L4Co exhibits a
tetrahedral geometry and crystallographic C2 symmetry, with
intrachelate bite angles of 94.67(8)° and interligand angles of
109.82(12)° and 115.89(13)°. The twist angle between the chelate
planes, as defined by the N―Co―O angle for each ligand, is slightly
flattened at 82.4°. There is therefore little deviation from a tetrahedral
geometry beyond that dictated by a b100° chelate bite angle. The Co
atom lies 0.36 Å above the mean OCCCN ligand plane, resulting in the
two N-aryl groups being positioned such that the terminal NH2 group
of each ligand is positioned directly over the center of the N-aryl
group of the other, with the two aryl rings only 3° away from being
perfectly coplanar and a closest contact between the calculated
position of the NH2 hydrogen atoms and the aryl ring plane being
2.90 Å. This distortion can be ascribed to a weak electronic π-stacking
and NH2–aryl ring quadrupolar interactions in the solid state. As
expected, the Co bond lengths are shorter in square planar L3Co than
in tetrahedral L4Co, with the average Co―N shorter by 0.110 Å and the
average Co―O shorter by 0.058 Å.
All these cobalt complexes exhibit broad, paramagnetically-shifted
signals in their 1H NMR spectra, as expected for Co(II) in both square
planar (low spin d7, S=1/2) and tetrahedral (high spin d7, S =3/2)
Table 1
Structure and refinement data for L3Co and L4Co.
L3Co
L4Co
Crystal system
Formula
Space group
Formula weight
Monoclinic
Monoclinic
C32H30N4O2Co
C2/c (#15)
561.53
C
22H18F6N2O2Co
À
P1 (#2)
513.30
Lattice type
Crystal color/habit
a/Å
b/Å
c/Å
α (°)
β (°)
γ (°)
Primitive
Red/irregular
8.5442(2)
9.8498(3)
12.5740(3)
76.629(2)
86.067(2)
80.055(2)
C-centered
Red/needle
17.357(1)
10.1783(8)
15.171(1)
90
The data obtained identified several general trends (see Table 3).
Firstly, cobalt complexes exhibit concentration-dependent activity
against particular cells, and the complexes are more active than their
corresponding ligands. In the majority of experiments, ligands did not
exhibit any significant cytotoxic activity; in cases where they did, it
was weaker than that of their corresponding cobalt complexes.
Secondly, there is a significant difference among the effects of the four
93.235(4)
90
Z
4
Fooo
518.00
0.71073
18636
4878
0.036
1.07
1172.00
0.71073
19463
3363
0.074
λ (Mo Kα)/cm−1
Reflections measured
Unique reflections
Rint
1
The CH signal is of low intensity and often so broad as to be undetectable, as is the
case in the spectrum shown in Fig. 4. The 1H NMR spectrum of another square planar
complex with R1 =CH3 and R2 =CF3 aids in and is consistent with these assignments:
L. Gurley, W. S. McNeil, unpublished observations.
GoF
1.03
1.394
Dcal (g/cm3)
1.682

L. Gurley et al. / Journal of Inorganic Biochemistry 105 (2011) 858–866
863
Fig. 4. Paramagnetic 1H NMR spectra of L1Co, L2Co, and L3Co.
cobalt complexes as measured by the EC50 and percent inhibition,
with L2Co being generally the most active and L4Co the least active of
the four compounds. Of particular note is the effect of L2Co on PC3
cells, with 94% maximal inhibition and an EC50 of 2.5 μM. Thirdly, the
cobalt complexes display selective activity against certain cell types.
Two prostate cancer cell lines tested and THP-1 acute monocytic
leukemia cells are consistently sensitive to the cobalt-containing
drugs. Neuroblastoma, astrocytoma and hepatocellular carcinoma
cells are not, with only weak activity exhibited by L2Co against
neuroblastoma cells, and by L3Co and L4Co against astrocytoma cells
(Table 3). None of the compounds tested had significant effects on
hepatocellular carcinoma cells.
These experiments indicate that some of the cobalt compounds
exhibit cytotoxic activity against human tumor cells. Similar observa-
tions have been made by Klanicova et al. [36] who prepared Co(II)
complexes with several N6-substituted adenine derivatives; five of their
six cobalt complexes were active, while the corresponding ligands were
not. Results from this study are similar to our data in that there was a
significant variation in the activity of the different compounds towards
human osteogenic sarcoma cells. The range of EC50 values observed for
this series of adenine compounds (7–28 μM) was very similar to that
observed for the effects of our β-ketoaminato complexes on PC-3 (2.5–
34 μM), VCaP (6–20 μM) and THP-1 (4–20 μM) cells (see Table 3). The
cobalt-adenine complexes were not active against human MCF-7 breast
cancer and K562 erythroleukemia cells. Only one of the complexes had a
Table 3
Estimated EC50 and percentage maximal inhibition values of cobalt complexes and their
respective ligands towards six different human cell lines.
PC3
NSa
VCAP
THP-1
NS
SH-SY5Y
NS
U-373 MG
NS
HepG2
NS
L1H2
L1Co
8 μMb
27%
20 μM
34%
NS
6 μM
87%
2 μM
12%
13 μM
73%
13 μM
67%
NS
2.5 μM
94%
NS
20 μM
58%
NS
20 μM
39%
NS
NS
NS
NS
L2H2
L2Co
NS
NS
NS
NS
NS
20 μM
27%
NS
L3H2
L3Co
L4H2
L4Co
NS
NS
NS
NS
NS
9 μM
88%
32 μM
34%
34 μM
38%
4 μM
19%
18 μM
64%
15 μM
83%
NS
NS
NS
5 μM
15%
NS
Table 2
NS
1H NMR chemical shift data (ppm) for compounds L1Co, L2Co, L3Co, and L4Co.
NS
20 μM
28%
Compound
CH3
CH
CH2
Ph
NH2
9.8
L1Co
L2Co
L3Co
L4Co
−20.8
−78.5
−91.7
−62.8
24.6, 14.2, 7.7
a
NS: no statistically significant effect compared to cells treated only with DMSO
116.5
101.0
−34.8
solvent.
29.4, 15.4
10.5, 14.7, 12.7,−11.0,
−18.5,−26.5,−54.7
b
Cytotoxicity was assessed after 48 h incubation as described in Fig. 5A for PC-3 cells
−19.2
and four of the compounds. EC50 values (μM) according to the MTT assay are presented
as well as the percentage of maximal inhibition of cell viability.

864
L. Gurley et al. / Journal of Inorganic Biochemistry 105 (2011) 858–866
Fig. 6. Cobalt complex L2Co causes significant activation of caspase-3 in human PC-3
prostate cancer cells, while L1Co has only a moderate effect. PC-3 cells (1.5×105 cells/
mL) were exposed to various concentrations of L1Co and L2Co or DMSO for 48 h.
Caspase 3 activity was measured by a fluorometric assay. Data (means SEM) from
four independent experiments are presented. **Pb0.01, significantly different from
cells exposed to the vehicle solution (DMSO) only, Fisher's LSD post hoc test.
that the widely used platinum-based chemotherapy agent, cisplatin,
has been reported to have an EC50 value of 1–5 μM in different cell
lines, including the prostate cancer cells [11,41]. This concentration
range is very close to the EC50 of L2Co, as well as cobalt compounds
described in previous studies [11,36,41]. It is also interesting to note
that not all cobalt complexes exhibit the same cell specificity. For
example, compounds described in this study were cytotoxic towards
prostate cancer cell lines, while several acetylene(hexacarbonyl)
dicobalt complexes described by Schmidt et al. [41] were toxic to two
different human mammary tumor cell lines, with weak effects
towards prostate cancer cells.
A series of experiments was performed to investigate possible
mechanisms of action of the most active compounds. Fig. 6 illustrates
that a 48 h incubation period with complex L2Co at concentrations of
either 5 or 10 μM induces statistically significant activation of caspase-
3 in PC-3 cells. Although there is also a concentration-dependent
trend observed with L1Co, the effect is weaker and does not reach
statistical significance with the numbers of observations made (n=4,
P=0.068 at 25 μM, Fisher's LSD post hoc test).
Fig. 5. Cobalt complexes L1Co and L2Co reduce viability of human PC-3 prostate cancer
cells, while their corresponding ligands L1H2 and L2H2 are ineffective. PC-3 cells were
seeded into 24-well plates at 1.2×105 cells per well in DMEM-F12 medium containing
5% FBS. Various concentrations (1–50 μM) of the two cobalt complexes or their
respective ligands were added and cell viability assessed 48 h later by the MTT (A) and
LDH (B) assays. Data (means SEM) from 4 independent experiments are presented as
percent viable cells. The concentration-dependent effects of compounds were assessed
by randomized block design ANOVA; NS, not significant.
weak effect (EC50 =42 μM) towards G361 melanoma cells [36]. We
observed similar cell-type specific activity of cobalt complexes, with
notable effect on human prostate cancer and monocytic leukemia cells
and weak or no effect on neuroblastoma, astrocytoma or hepatocellular
carcinoma cells. Additional examples of cell type specificity were
described for the complex of cobalt with mefenamic acid [37]; however,
the EC50 values towards human MCF-7 breast cancer (94 μM), T-24
bladder cancer (27 μM), and A-549 non-small cell lung carcinoma
(116 μM) were generally higher than those observed in our study. Cell
type specific cytotoxic effects have been observed not only for Co(II)
complexes but also for compounds containinga number or other metals,
including Cu(II), Zn(II), Mn(II), Mo(VI), Ni(II), as well as Co(III) [37–40].
Selective alterations in the peripheral ligand substituents lead to
changes in the cytotoxic activity of the cobalt complexes. The L2Co
complex, which contains four trifluoromethyl groups, has the most
potent cytotoxic activity toward prostate and monocytic leukemia
cells. The two other square planar complexes L1Co and L3Co exhibit
lower potency, while L4Co is the least effective among the four cobalt
complexes tested. L4Co is a Co(II) tetrahedral complex, which exhibits
significant differences in its steric and electronic properties when
compared to the other square planar β-ketoamine complexes. This is
consistent with previous findings that tetrahedral cobalt(II) com-
pounds with a 2-methylbenzimidazole ligand exhibit insignificant
cytotoxicity toward a range of human cancer cell lines[39], suggesting
that the relationship of cytotoxic activity to the geometry of four-
coordinate cobalt(II) complexes is a general one. However, it should
be noted that tetrahedral L4Co, along with L3Co, were the only two
compounds that exhibit weak activity against U373-MG cells.
The effects of two MAP kinase inhibitors were examined. Fig. 7
illustrates that SP600125, a selective JNK inhibitor, at 20 μM significantly
reduces the activity of L1Co, L2Co, and L3Co against PC-3 cells. The effect of
PD98059, an MEK1/2 inhibitor, is weaker and reaches statistical
significance (Pb0.05) only in the case of L1Co. Different concentrations
Fig. 7. Cytotoxicity of cobalt complexes towards human PC-3 prostate cancer cells is
reduced by JNK and MAPK/ERK kinase inhibitors. PC-3 cells were pre-incubated for
15 min with 20 μM of either SP600125, a selective JNK inhibitor, or PD98059, a selective
MAPK/ERK kinase inhibitor, before being exposed to three different cobalt complexes.
Cell viability was assessed 48 h later by the MTT assay. Data (means SEM) from 5
independent experiments are expressed as percent viable cells. *Pb0.05; **Pb0.01
significantly different from cells treated with the cobalt complex only; paired Student's
t-test values were corrected for multiple comparisons by Holm's step-down procedure.
Ott et al. [11] have shown that modification of substituent on the
alkyne ligand derived from acetylsalicylic acid leads to varying EC50
values of the obtained cobalt complexes towards human breast cancer
cell lines. Very potent cytostatic effects were observed with EC50
reaching 1.4 μM for the lead compound in this series. It is noteworthy

L. Gurley et al. / Journal of Inorganic Biochemistry 105 (2011) 858–866
865
DMSO
EC50
ERK
dimethylsulfoxide
half-maximal effective concentration
extracellular signal-regulated kinase
fetal bovine serum
FBS
JNK
LDH
LSD
c-Jun NH2-terminal kinase
lactate dehydrogenase
least significant differences
mitogen-activated protein kinase
MAPK
MEK1/2 MAPK/ERK kinase
MTT
3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium
bromide
NAC
NMR
NS
N-acetyl-L-cysteine
nuclear magnetic resonance
not significant
Fig. 8. Antioxidant NAC reduces cytotoxicity of cobalt complexes towards human PC-3
prostate cancercells. PC-3 cells werepre-incubated for 15 minwith various concentrations
of NAC before being exposed to three different cobalt complexes. Cell viability was
assessed 48 h later by the LDH assay. Data (means SEM) from 3 to 6 independent
experiments are expressed as percent dead cells. *Pb0.05; **Pb0.01 significantly different
from cells treated with the cobalt complex only, Fisher's LSD post hoc test.
SDS
sodium dodecyl sulfate
TBDMS tert-butyldimethylsilyl, tBuMe2Si
THF tetrahydrofuran
Acknowledgements
of the cobalt complexes were used to induce PC-3 cell death; these
concentrations were established in the preliminary experiments in order
to achieve similar levels of activity among the three different compounds.
Similarly, NAC, a potent thiol antioxidant, prevents cytotoxic activity
induced by the three cobalt complexes in a concentration-dependent
manner (Fig. 8). Again, preliminary experiments helped establish
concentrations of the cobalt complexes needed to achieve similar levels
of activity according to the LDH assay. The MTT assay was not used for
these experiments due to an interaction of NAC with the formazan dye.
NAC did not interfere with the LDH assay.
This work was supported by grants from the Natural Sciences and
Engineering Research Council of Canada, the Jack Brown and Family
Alzheimer's Disease Research Foundation, Prostate Cancer Canada,
and the Irving K. Barber School of Arts and Sciences at UBC Okanagan.
The authors would like to thank Ms. Jocelyn Madeira for the technical
assistance and Dr. Brian Patrick for the X-ray crystallographic analysis.
Appendix A. Supplementary data
These experiments reveal several similarities with the mode of
action exhibited by cisplatin and other anti-cancer drugs (for reviews
see [16,42]), which may suggest that the cell death induced by cobalt
compounds is at least partially a result of apoptosis. The increased
caspase-3 activation by L2Co, and to a lesser extent by L1Co (Fig. 6), is a
common mechanism of action for many anti-cancer drugs [42]
including cisplatin [43]. As well, the PC-3 cell death caused by three
of the most active cobalt compounds (L1Co, L2Co, and L3Co) appears
dependent on activation of MAP kinases, and JNK in particular, a
mechanism that has also been widely implicated in apoptosis as
induced by cisplatin and other anti-cancer drugs [12,44].
It has also been shown that oxidative stress is caused by a number
of anti-cancer medications including cisplatin [45], and the observed
reduction of cytotoxic activity upon treatment with NAC indicates
that the generation of reactive oxygen species could also be an active
mechanism of activity of the cobalt complexes [13].
Supplementary data to this article can be found online at
doi:10.1016/j.jinorgbio.2011.03.005.
References
[1] K.H. Thompson, C. Orvig, Science 300 (2003) 936–939.
[2] E.R. Jamieson, S.J. Lippard, Chem. Rev. 99 (1999) 2467–2498.
[3] D.J. Vaughn, G.A. Gignac, A.T. Meadows, Ann. Intern. Med. 136 (2002) 463–470.
[4] S. van Zutphen, J. Reedijk, Coord. Chem. Rev. 249 (2005) 2845–2853.
[5] A.S. Abu-Surrah, M. Kettunen, Curr. Med. Chem. 13 (2006) 1337–1357.
[6] M. Knipp, Curr. Med. Chem. 16 (2009) 522–537.
[7] A.M. Montana, C. Batalla, Curr. Med. Chem. 16 (2009) 2235–2260.
[8] G. Cossa, L. Gatti, F. Zunino, P. Perego, Curr. Med. Chem. 16 (2009) 2355–2365.
[9] R. Banerjee, S.W. Ragsdale, Annu. Rev. Biochem. 72 (2003) 209–247.
[10] T. Yano, T. Ozawa, H. Masuda, Chem. Lett. 37 (2008) 672–677.
[11] I. Ott, K. Schmidt, B. Kircher, P. Schumacher, T. Wiglenda, R. Gust, J. Med. Chem. 48
(2005) 622–629.
[12] I. Ott, B. Kircher, R. Dembinski, R. Gust, Expert Opin. Ther. Pat. 18 (2008) 327–337.
[13] J.A. Schwartz, E.K. Lium, S.J. Silverstein, J. Virol. 75 (2001) 4117–4128.
[14] W.K. Hagmann, J. Med. Chem. 51 (2008) 4359–4369.
[15] D. O'Hagan, D.B. Harper, J. Fluorine Chem. 100 (1999) 127–133.
[16] R. Kim, K. Tanabe, Y. Uchida, M. Emi, H. Inoue, T. Toge, Cancer Chemother.
Pharmacol. 50 (2002) 343–352.
4. Conclusions
Four (β-ketoaminato)cobalt compounds and their corresponding
organic ligands have been tested for their cytotoxic effects against six
human cancer cell lines. The data indicate that some of these cobalt
compounds, L2Co in particular, could be useful as anti-tumor agents,
and that the activity is selective towards prostate cancer and leukemia
cells. Initial experiments revealed some of the molecular mechanisms
engaged by these novel compounds, including caspase-3 and MAP
kinase activation as well as oxidative stress, but further work will be
needed to fully characterize the cell signaling pathways engaged by
the cobalt complexes. Future work will also attempt to increase the
potency and selectivity of the cobalt-containing drugs towards, for
example, prostate cancer cells through systematic derivatization of
the ligands used for preparation of the cobalt complexes.
[17] G. Tortora, R. Bianco, G. Daniele, F. Ciardiello, J.A. McCubrey, M.R. Ricciardi, L.
Ciuffreda, F. Cognetti, A. Tafuri, M. Milella, Drug Resist. Updat. 10 (2007) 81–100.
[18] Y.H. Kang, M.J. Yi, M.J. Kim, M.T. Park, S. Bae, C.M. Kang, C.K. Cho, I.C. Park, M.J. Park, C.
H. Rhee, S.I. Hong, H.Y. Chung, Y.S. Lee, S.J. Lee, Cancer Res. 64 (2004) 8960–8967.
[19] P.J. McCarthy, R.J. Hovey, K. Ueno, A.E. Martell, J. Am. Chem. Soc. 77 (1955)
5820–5824.
[20] S.Z. Haider, A. Hashem, K.M.A. Malik, M.B. Hursthouse, J. Bangladesh Acad. Sci.
4 (1980) 139–150.
[21] W.T. Robinson, G.A. Rodley, Acta Crystallogr. Sect. A 28 (1972) S52.
[22] H. Burger, U. Wannagat, Monatsh. Chem. 94 (1963) 1007–1012.
[23] J.A.T. Norman, Air Products and Chemicals, Inc., USA. Patent Application: EP, 1991,
p. 17.
[24] P.R. Shipley, C.C.A. Donnelly, C.H. Le, A.D. Bernauer, A. Klegeris, Int. J. Mol. Med. 24
(2009) 711–715.
[25] M.M. Bradford, Anal. Biochem. 72 (1976) 248–254.
[26] A.M. Youssef, E.G. Neeland, E.B. Villanueva, M.S. White, I.M. El-Ashmawy, B.
Patrick, A. Klegeris, A.S. Abd-El-Aziz, Bioorg. Med. Chem. 18 (2010) 5685–5696.
[27] S. Holm, Scand. J. Stat. 6 (1979) 65–70.
[28] A.E. Martell, R.L. Belford, M. Calvin, J. Inorg. Nucl. Chem. 5 (1958) 170–181.
[29] M.F. Richardson, R.E. Sievers, J. Inorg. Nucl. Chem. 32 (1970) 1895–1906.
[30] B.D. Murray, P.P. Power, Inorg. Chem. 23 (1984) 4584–4588.
[31] S.D. Reid, A.J. Blake, C. Wilson, J.B. Love, Inorg. Chem. 45 (2006) 636–643.
Abbreviations
ANOVA analysis of variance
DMF
N,N-dimethyl formamide

866
L. Gurley et al. / Journal of Inorganic Biochemistry 105 (2011) 858–866
[32] P.L. Holland, Acc. Chem. Res. 41 (2008) 905–914.
[33] S. Brückner, M. Calligaris, G. Nardin, L. Randaccio, Inorg. Chim. Acta 2 (1968)
379–385.
[34] K.C.D. Robson, C.D. Phillips, B.O. Patrick, W.S. McNeil, Dalton Trans. 39 (2010)
2573–2578.
[35] E. Bonfoco, D. Krainc, M. Ankarcrona, P. Nicotera, S.A. Lipton, Proc. Natl. Acad. Sci.
U.S.A. 92 (1995) 7162–7166.
[38] N. Katsaros, M. Katsarou, S.P. Sovilj, K. Babic-Samardzija, D.M. Mitic, Bioinorg.
Chem. Appl. 2 (2004) 193–207.
[39] O. Sanchez-Guadarrama, H. Lopez-Sandoval, F. Sanchez-Bartez, I. Gracia-Mora, H.
Hopfl, N. Barba-Behrens, J. Inorg. Biochem. 103 (2009) 1204–1213.
[40] Z.C. He, D.N. Lai, L.P.G. Wakelin, Eur. J. Pharmacol. 637 (2010) 11–15.
[41] K. Schmidt, M. Jung, R. Keilitz, B. Schnurr, R. Gust, Inorg. Chim. Acta 306 (2000) 6–16.
[42] Z.H. Siddik, Cancer Treat. Res. 112 (2002) 263–284.
[36] A. Klanicova, Z. Travnicek, I. Popa, M. Cajan, K. Dolezal, Polyhedron 25 (2006)
1421–1432.
[43] B. Del Bello, D. Moretti, A. Gamberucci, E. Maellaro, Oncogene 26 (2007) 2717–2726.
[44] A. Brozovic, M. Osmak, Cancer Lett. 251 (2007) 1–16.
[37] D. Kovala-Demertzi, D. Hadjipavlou-Litina, M. Staninska, A. Primikiri, C. Kotoglou,
M.A. Demertzis, J. Enzyme Inhib, Med. Chem. 24 (2009) 742–752.
[45] D.M. Mattson, M.A. Iman, D. Dayal, A.D. Parsons, N. Aykin-Burns, L. Li, K.P. Orcutt,
D.R. Spitz, K.J. Dornfeld, A.L. Simons, Free Radical Biol. Med. 46 (2009) 232–237.