Antitumor-active cobalt-alkyne complexes derived from acetylsalicylic acid: Studies on the mode of drug action

Article abstract of DOI:10.1021/jm049326z

Cobalt-alkyne complexes are drugs with remarkable cytotoxicity. From the complexes tested up to now we selected the aspirin derivative [2-acetoxy-(2-propynyl)benzoate]hexacarbonyl-dicobalt (Co-ASS) as the lead compound. To get more insight into the mode o

Full text of DOI:10.1021/jm049326z

622
J. Med. Chem. 2005, 48, 622-629
Antitumor-Active Cobalt-Alkyne Complexes Derived from Acetylsalicylic Acid:
Studies on the Mode of Drug Action
Ingo Ott,† Kathrin Schmidt,† Brigitte Kircher,‡ Petra Schumacher,‡ Thomas Wiglenda,† and Ronald Gust*,†
Institute of Pharmacy, Free University of Berlin, Ko¨nigin Luise Str. 2+4, 14195 Berlin, Germany, and Laboratory for
Tumor and Immunobiology, Department of Hematology and Oncology, University Hospital of Innsbruck, Austria
Received August 13, 2004
Cobalt-alkyne complexes are drugs with remarkable cytotoxicity. From the complexes tested
up to now we selected the aspirin derivative [2-acetoxy-(2-propynyl)benzoate]hexacarbonyl-
dicobalt (Co-ASS) as the lead compound. To get more insight into the mode of action, we
systematically modified the alkyne ligand and determined the cytotoxic properties of the
resulting cobalt complexes. Further investigations were performed on the drug lipophilicity,
the cellular uptake into MCF-7 and MDA-MB 231 breast cancer cells, the DNA-binding efficacy,
and the nuclear drug content. The ability to inhibit glutathione reductase and cyclooxygenase
(COX) enzymes, the binding to the estrogen receptor, and the induction of apoptotic processes
were examined for selected compounds. Interestingly, the most antitumor active compounds
were potent COX inhibitors (COX-1 and COX-2). The presented results indicate that cobalt-
alkyne complexes of the Co-ASS type, represent a new class of organometallic cytostatics with
a mode of drug action in which COX inhibition probably plays a major role.
Introduction
Platinum-containing anticancer drugs (e.g. cisplatin)
are widely used in the treatment of human cancers.
Many efforts have been made to establish further metal-
containing cytostatics, however, without great success.
Only little is known about the antitumor efficacy of trace
elements such as copper, zinc, or cobalt.1 We and others
reported recently about cytotoxic cobalt-alkyne com-
plexes.2-5 These compounds consist of a hexacarbonyl-
dicobalt cluster bound to an alkyne ligand. Some of
these complexes were found to be active against a broad
variety of human epithelialic tumors as well as lym-
phomas and leukemias. The best results, however, were
obtained on human breast cancer cells lines.
Out of the complexes tested up to now [2-acetoxy-(2-
propynyl)benzoate]hexacarbonyldicobalt (Co-ASS, see
Figure 1) emerged as the lead compound. First struc-
ture-activity studies showed that the activity depended
strongly on the nature of the alkyne ligand and on an
intact cobalt-alkyne complex. The mode of drug action
remained unclear.
Figure 1. Cobalt-alkyne complexes.
To get more insight into the structural requirements
for the high cytotoxicity of Co-ASS, we modified sys-
tematically the alkyne ligand: the 2-acetoxy group in
the aromatic ring was replaced by H (Co-Benz), OH (Co-
SAL), F (Co-2F-Benz) or moved into position 3 (Co-3-
Acetbenz), the ester in Co-ASS was cleaved (Co-Prop)
or replaced by an amide bond (Co-ASSAM), and Co-
Phthal was designed as a phthalimide derivative of
alkynehexacarbonyldicobalt (see Figure 1).
lines. We determined the lipophilicity of the complexes
as log kw value by HPLC and the cellular uptake by
atomic absorption spectroscopy in order to clarify if drug
distribution processes may explain the different drug
activities. As the complexes contain metal atoms a
“cisplatin-like” mechanism involving covalent DNA
modification has to be considered. Therefore, DNA-
binding studies were performed, and the drug content
of the cell nuclei was quantified. On the basis of these
results, we selected compounds for the following studies
on the mode of drug action.
The compounds were investigated for cytotoxicity at
the MCF-7 and MDA-MB 231 human breast cancer cell
The intracellular concentration of reduced glutathione
was found to be in the mM-range in adherent growing
cancer cells, and the glutathione reductase system is
upregulated in human breast cancers.6,7 Metal-contain-
* Institut fu¨r Pharmazie der FU Berlin, Ko¨nigin Luise Str. 2+4,
D-14195 Berlin, Germany. Phone: (030) 838 53272. Telefax: (030) 838
56906. E-mail: rgust@zedat.fu-berlin.de.
† Free University of Berlin.
‡ University Hospital of Innsbruck.
10.1021/jm049326z CCC: $30.25 © 2005 American Chemical Society
Published on Web 12/30/2004

Antitumor-Active Cobalt-Alkyne Complexes
Journal of Medicinal Chemistry, 2005, Vol. 48, No. 2 623
Table 1. Lipophilicity (log kw), Cytotoxicity (IC50), and DNA
Binding Efficiency
Scheme 1. (A) Preparation of Alkynes. (B) Reaction of
the Alkynes with Dicobaltoctacarbonyl
IC50 (µM)
IC50 (µM)
DNA binding
compound
Co-Prop
Co-ASSAM
Co-Phthal
Co-ASS
Co-3-Acetbenz 4.93
Co-2F-Benz
Co-Benz
Co-SAL
cisplatin
melphalan
tamoxifen
5-fluorouracil
log kw
MCF-7
MDA-MB 231
(pmol/µg)
3.55a 7.0 (( 2.0) >50
3.81 8.8 (( 1.0)
4.42 22.2 (( 0.4) >50
4.91a 1.4 (( 0.3)
33.94 (( 9.29)
9.1 (( 1.0) 22.63 (( 2.25)
4.57 (( 0.96)
1.9 (( 0.3) 9.82 (( 0.20)
10.3 (( 0.5) 9.18 (( 2.44)
9.6 (( 0.7) 9.22 (( 0.28)
11.4 (( 0.9) 15.19 (( 0.07)
4.4 (( 1.7) 36.97 (( 3.75)
4.0 (( 1.5)
9.1 (( 1.3)
7.1 (( 0.9)
5.4 (( 0.4)
3.9 (( 0.2)
2.0 (( 0.3)
5.7 (( 1.5)
2.3 (( 0.4)
4.8 (( 0.6)
5.05
5.22
5.32
3.9 (( 0.3)
10.6 (( 0.6)
9.6 (( 0.3)
a See ref 5.
ing drugs disturb this sensible cellular redox system.8,9
Therefore, we evaluated the influence of cobalt-alkyne
complexes on the glutathione reductase activity.
Furthermore, as the lead compound Co-ASS repre-
sents a derivative of the nonsteroidal antiinflammatory
drug (NSAID) acetylsalicylic acid (aspirin, ASS) the
presence of similar pharmacological properties can be
expected. NSAIDs are seen as future drugs for the
treatment of a broad variety of human tumors, including
breast and colon cancer.10-13 The antitumor activity of
NSAIDs involves the inhibition of COX enzymes. There-
fore, the ability of the complexes to inhibit COX enzymes
was studied by ELISA.
Results and Discussion
The antiproliferative activity of the complexes was
determined in MCF-7 and MDA-MB 231 human breast
cancer cells (see Table 1).
The lead structure Co-ASS showed IC50 values lower
than 2.0 µM in both cell cultures (1.4 µM in MCF-7 and
1.9 µM in MDA-MB 231 cells). The desacetoxy derivative
Co-Benz, the amide derivatives (Co-ASSAM and Co-
Phthal), and the structural isomer Co-3-Acetbenz were
less active (IC50 > 5 µM in MCF-7 and IC50 > 9 µM in
MDA-MB 231 cells, see Table 1). Only the 2-hydroxy
derivative Co-SAL showed comparable activity with
IC50 ) 3.9 µM (MCF-7) and 4.4 µM (MDA-MB 231). The
exchange of the 2-OH by fluorine (Co-2F-Benz) again
decreased the antiproliferative potency (IC50 > 7 µM at
both cell lines, see Table 1).
Further investigations include the determination of
apoptotic changes in the tumor cells and, due to the
selectivity of Co-ASS for breast cancer cells, the proof
of a possible interaction with the estrogen receptor.
Chemistry
These findings clearly indicate the importance of the
defined 2-acetoxy-(2-propynyl)benzoate ligand structure
for the high cytotoxicity of Co-ASS. Metabolic degrada-
tion is only tolerated at the 2-acetoxy group. Cleavage
of the propynyl ester led to Co-Prop, which showed
effects against MCF-7 cells (IC50 ) 7.0 µM), but was
completely inactive at the MDA-MB 231 cell line
(IC50 > 50 µM). Furthermore, the cytotoxic activity
strongly depended on the coordination of the alkyne
ligand at the Co2(CO)6 cluster. The free ligand as well
as the precursors ASS and Co2(CO)8 were inactive.3
The established antitumor drugs cisplatin, mel-
phalan, tamoxifen, and 5-fluorouracil were used as
references and were found to be active in the range of
IC50 ) 2.0 to 10.6 µM. It should be mentioned that all
structural modifications of the lead compound Co-ASS
(see Figure 1) led to a significant decrease in activity,
but the compounds (with exception of Co-Prop and Co-
Phthal at the MDA-MB 231 cell line) were still active
in the range of these drugs.
Various studies correlate the cytotoxicity of metal
complexes with the accumulation in tumor cells, which
itself depends on the drug lipophilicity (log kw value) if
a passive transport through the cell membrane occurs.
The log kw values of the cobalt-alkyne complexes
determined by reversed phase HPLC5 (see Table 1) were
in the range of 3.55 (Co-Prop) to 5.32 (Co-SAL). Co-ASS
and its structural isomer Co-3-Acetbenz possessed
comparable log kw values of 4.91 and 4.93, respec-
tively. The amide derivatives Co-Phthal (log kw ) 4.42)
and Co-ASSAM (log kw ) 3.81) exhibited distinctly
lower lipophilicity than the benzoic ester derivatives
The synthesis of Co-ASS, Co-SAL,2 and Co-Prop14 was
described previously. For the synthesis of the cobalt-
alkyne complexes Co-Benz, Co-3-Acetbenz, Co-2F-Benz,
Co-ASSAM, and Co-Phthal, we used a two-step proce-
dure as depicted in Scheme 1.
In the first step, the alkyne esters (2-propynyl)-
benzoate, 3-acetoxy-(2-propynyl)benzoate, and 2-fluoro-
(2-propynyl)benzoate were prepared by reaction of the
respective benzoic acid with dicyclohexylcarbodiimide
(DCC), 4-(dimethylamino)pyridine (DMAP), and an
excess of propargylic alcohol (see Scheme 1A). The
synthesis of the amides was performed as described
earlier.15,16
In the second step, the alkynes were reacted with
dicobaltoctacarbonyl to give the cobalt-alkyne com-
plexes (see Scheme 1B), which were purified by column
chromatography on SiO2 and analyzed by means of IR,
NMR, MS, and elemental analysis.
1
In the H NMR spectra of the complexes, the reso-
nances for the CtCH (δ ≈ 6.1 ppm) and CtC-CH2
(δ ) 4.7-5.6 ppm) protons of the complexes are located
in the region of double bonds. This effect is well-known
from many studies on cobalt-alkyne complexes and
indicates that the triple bond character of the alkynes
is lowered toward that of a Z-olefin due to the coordina-
tion to Co2(CO)6. The IR spectrum of dicobaltoctacar-
bonyl shows intensive stretching vibrations of terminal
(2000-2100 cm-1) and bridging CO ligands (1830-1850
cm-1). By the coordination reaction the bridging CO
ligands are lost and the IR spectra of the complexes
exhibit only the signals of the terminal CO ligands.3

624 Journal of Medicinal Chemistry, 2005, Vol. 48, No. 2
Ott et al.
Figure 2. Concentration dependent cellular uptake of Co-
ASS into MCF-7 cells.
(log kw ) 4.91-5.32). Interestingly, the high lipophilic
character of the cobalt-alkyne complexes was mainly
determined by the cobalt cluster as confirmed for Co-
ASS. The coordination of the 2-acetoxy-(2-propynyl)-
benzoate (Prop-ASS) to Co2(CO)6 increased the log kw
from 1.98 to 4.91.
The accumulation into tumor cells was studied on
both cell lines used for the cytotoxicity experiments. The
cells were treated with the respective cobalt-alkyne
complex, subsequently lysed by sonification, and the
cellular cobalt content was quantified by atomic absorp-
tion spectroscopy.
Figure 2 shows the concentration dependent cellular
uptake of Co-ASS into MCF-7 cells after 6 h of drug
exposure. The high linear correlation (r2 > 0.999) of
intra- and extracellular drug concentration indicates
that the drug uptake is not saturated within the
concentration range (1.0-10 µM) which is important for
the cytotoxic action. The cellular molar drug concentra-
tion can be estimated from the pmol/µg value as
described previously.17 Dividing by the concentration
used in the extracellular medium, the accumulation
grade can be calculated. The mean accumulation grade
for Co-ASS (using the data shown in Figure 2) was 150.3
((12.0). This result is of general interest as cisplatin
showed accumulation grades lower than 6-fold.17,18
The time dependent uptake into MCF-7 and MDA-
MB 231 cells was determined over a period of 24 h for
Co-ASS, Co-Prop, Co-SAL, and Co-Benz (see Figure 3).
All complexes were taken up to maximum levels
within 4-8 h, whereby in MCF-7 cells a higher cobalt
content was measured than in MDA-MB 231 cells. Cells
treated with 2.0 µM Co-ASS exhibited cobalt content of
>2.5 pmol/µg (MCF-7) and >1.7 pmol/µg (MDA-MB
231). The other complexes were taken up in a lower
amount. A slight decrease of cellular concentration is
observed in most cases after longer drug exposure.
Interestingly, cobalt could not be found in comparable
concentrations if the cells were incubated with Co2(CO)8
(<0.15 pmol/µg). Therefore, it can be concluded that the
alkyne ligand is inevitably necessary for the uptake into
the cells.
Figure 3. Time-dependend cellular uptake of 2.0 µM cobalt-
alkyne complexes into a) MCF-7 cells and b) MDA-MB 231
cells.
lower cytotoxicity is enriched in the tumor cells to nearly
the same extent. Furthermore, the attempt failed to
correlate the cytotoxicity (IC50) with the lipophilicity
(log kw) (r2 ) 0.31 for MCF-7 cells and r2 ) 0.14 for
MDA-MB 231 cells).
These results altogether show that drug distribution
processes are only in part responsible for the observed
differences in cytotoxic potency. The absence of a
correlation between drug activity, lipophilicity, and
cellular concentration indicate the involvement of a
specific mode of drug action.
A “cisplatin-like” mechanism (covalent binding to the
DNA) had to be considered first. Cobalt-alkyne com-
plexes were already designed as labeling agents and
used for the so-called carbonyl-metallo-immunoassay
(CMIA). Covalent binding to the estrogen receptor was
confirmed for derivatives of ethynylestradiol, suggesting
that cobalt-alkyne complexes are able to interact
covalently with biomolecules.19,20
The DNA binding was evaluated by incubation of the
complexes with salmon testes DNA. All the compounds
bound efficiently to DNA (> 4.5 pmol/µg, see Table 1),
distinctly higher than the DNA-interacting platinum
drugs cisplatin or carboplatin under similar experimen-
tal conditions (<1 pmol/µg 21).
The results of the investigations on the drug lipophi-
licity and the uptake into tumor cells are only in part
suitable to explain the differences in cytotoxicity. The
highest cellular cobalt levels are achieved with the most
active cobalt complex Co-ASS. Co-Benz with distinctly
The lead compound Co-ASS showed only average
binding efficiency (9.82 pmol/µg) comparable to Co-3-
Acetbenz (9.18 pmol/µg) and Co-2F-Benz (9.22 pmol/µg).

Antitumor-Active Cobalt-Alkyne Complexes
Journal of Medicinal Chemistry, 2005, Vol. 48, No. 2 625
Table 2. Drug Content of the Cell Nuclei (pmol/µg and
Table 3. Influence on Cellular Glutathione Reductase Activity
% of total)
glutathione
Co-ASS
Co-SAL
Co-Benz
Co-Prop
compound
reductase activity (%)
MCF-7
0.12 (( 0.03) 0.11 (( 0.01) 0.11 (( 0.05) 0.15 (( 0.02)
Co-ASS
Co-Prop
ASS
93 (( 5)
105 (( 13)
101 (( 14)
97 (( 1)
0.8% 1.0% 0.9% 1.4%
MDA-MB 231 0.22 (( 0.13) 0.14 (( 0.05) 0.14 (( 0.05) 0.15 (( 0.05)
5.0% 1.7% 1.9% 2.0%
Prop-ASS
Much higher DNA interaction achieved Co-Benz (15.19
pmol/µg), Co-ASSAM (22.63 pmol/µg), Co-Prop (33.94
pmol/µg), and Co-SAL (36.97 pmol/µg). A correlation
between the DNA-binding efficiency and the cytotoxicity
again could not be found (see data in Table 1). Interest-
ingly, the strongest DNA interaction (105.6 ( 18.3 pmol/
µg) was measured for Co2(CO)8, which allows the
assumption that the binding of the cobalt-alkyne
complexes to the DNA is due to the interaction with the
cobalt cluster.
To get information if drug-DNA binding occurs in the
tumor cells, the nuclei were isolated by a short sucrose
gradient and investigated for their cobalt content. The
results were calculated as pmol drug per µg nuclear
protein and as percentage of the total cellular drug
amount. Table 2 summarizes the results. Nuclear drug
amounts were nearly independent of the kind of alkyne
ligand. The observed cobalt content of 0.11 to 0.22 pmol/
µg corresponded only to 0.8-1.4% (MCF-7) and 1.7-
5.0% (MDA-MB 231) of the total drug amount found in
the cells. On the basis of these results, a genomic mode
of drug action involving DNA interaction can be ex-
cluded for the cobalt-alkyne complexes.
Therefore, we searched for other targets within the
tumor cells. First experiments were done with the most
active complex Co-ASS to evaluate if apoptotic changes
could be achieved in tumor cells. However, neither
MCF-7 cells nor MDA-MB 231 cells showed signs of
apoptosis after treatment with Co-ASS. These results
correlate well with findings that acetylsalicylic acid
induces apoptotic changes only at extremely high (mil-
limolar) concentration.22
The growth of hormone dependent MCF-7 breast
cancer cells cannot be inhibited only by cytostatics but
also by estrogen receptor modulating compounds. On
the basis of this fact, the relative estrogen receptor
binding affinity (RBA) was determined exemplarily for
Co-ASS and Co-3-Acetbenz. Both compounds showed
extremely low receptor binding (RBA ) 0% for Co-ASS
and RBA ) 0.05% for Co-3-Acetbenz), and as a conse-
quence we exclude the estrogen receptor as target for
cobalt-alkyne complexes.
the NSAIDs acetylsalicylic acid and salicylic acid (SAL),
it might be possible that the NSAID-character of the
compounds led to the growth inhibitory effects in MCF-7
and MDA-MB 231 cells.
The COX-inhibitory potential of Co-ASS, Co-SAL, Co-
Prop, and Co-Benz was determined for COX-1 and
COX-2 by ELISA (see Figure 4). The precursors acetyl-
salicylic acid and its propargylic ester were used as
references.
Acetylsalicylic acid was only active at COX-1 in a
concentration of 200 µM (91% inhibition), its propargylic
ester showed lower activity at COX-1 (48% inhibition
at 200 µM), and no notable activity at COX-2. These
results are in agreement with studies of Kalgutkar et
al. who reported recently that the formation of esters
of NSAIDs favors the interaction with COX-2.23,24
Coordination of Prop-ASS at the Co2(CO)6 cluster
enormously increased the inhibitory effects. Co-ASS
showed good inhibitory potential at 10 µM (COX-1: 37%,
COX-2: 58% inhibition) and terminated the enzyme
activity completely at 200 µM. It is worth to note that
due to the 150-fold accumulation grade of Co-ASS (see
cellular uptake studies) cellular concentrations of
200 µM and higher can be easily achieved.
Co-ASS is by far a more effective COX inhibitor than
the parent compound acetylsalicylic acid. The positive
effects of metal complexation of NSAIDs have already
been described for copper derivatives. These complexes
exhibit stronger or similar antiinflammatory potency
compared to the parent NSAIDs.25-27 Interestingly, a
copper aspirinate complex showed also higher antiin-
flammatory effects in vivo than aspirin.25 Unfortunately,
investigations on the cytotoxic properties of these
compounds are missing.
The COX inhibitory potency of the cobalt-alkyne
complexes strongly depended on the alkyne ligand. Co-
SAL completely inhibited both COX enzymes at 200 µM,
but was inactive at 10 µM. Co-Benz was more active at
COX-1 (40% inhibition at 10 µM and 47% inhibition at
200 µM) than at COX-2 (inactive at 10 µM, 41%
inhibition at 200 µM). Co-Prop showed low inhibitory
potential only in the highest used concentration of 200
µM (COX-1: 53%, COX-2: 41% inhibition). Interest-
ingly, the order of COX-inhibitory activity (Co-ASS >
Co-SAL > Co-Benz > Co-Prop) correlates well with their
cytotoxicity. So it is very likely but not confirmed that
the inhibition of the COX enzymes is the mode of action
of the cobalt-alkyne complexes presented in this paper.
The influence of Co-ASS and Co-Prop as well as the
non-cobalt-containing precursors of Co-ASS acetylsali-
cylic acid (ASS) and its propargylic ester (Prop-ASS) on
the glutathione reductase activity was evaluated to
estimate if the interaction with this very sensible redox
system caused the antiproliferative effects. As listed in
Table 3, none of the compounds significantly inhibited
or stimulated the glutathione reductase enzyme activity
after 6 h of drug incubation. After 24 h similar results
were observed (data not shown).
During the last years, cyclooxygenases became very
interesting targets in cancer chemotherapy. Presently
a few NSAIDS are studied in clinical trials for tumor
therapy and prevention. As the most active compounds
(Co-ASS, Co-SAL) of the SAR study are derivatives of
Inhibitors of cyclooxygenase enzymes are seen as
anticancer drugs of the future. Up to now only a few
COX inhibiting compounds with strong antiproliferative
potency are described in the literature. The COX-2
selective aspirin derivative APHS strongly influences
the growth of COX-2 positive HCA-7 colon carcinoma
cells with IC50 values in the 2-5 µM range.28,29

626 Journal of Medicinal Chemistry, 2005, Vol. 48, No. 2
Ott et al.
Figure 4. COX activity (COX-1 and COX-2) after treatment with compounds (10 µM and 200 µM).
chromatography on silica gel (mobile phase: diethyl ether/
petroleum ether).
Many studies deal with another COX-2 selective
compound: celecoxib. It showed antiproliferative effects
(IC50 ) 35-65 µM) at both COX-2 positive and COX-2
negative cell lines.30 The cellular mechanisms causing
the antitumor effects of NSAIDS are seen controver-
sially within the scientific community. Besides the
inhibition of COX enzymes, effects on lipoxygenase or
inhibition of angiogensis, for example, are under discus-
sion.30,31 However, the anticancer effects are a common
effect of many NSAIDs.
(2-Propynyl)benzoate. Benzoic acid: 2.00 g (16.3 mmol),
DCC: 3.70 g (17.9 mmol), DMAP: 244.0 mg (2.0 mmol),
2-propynol: 1.01 g (17.9 mmol); Yield: 2.11 g (13.2 mmol), 81%
yellow oil; IR (KBr, cm-1) 1724 (CdO); MS (EI, 35 °C): m/z
1
(%) ) 160 (7) [M+], 105 (100). H NMR (CDCl3): δ ) 2.52 (t,
1H, J ) 2.5 Hz, CH), 4.93 (d, 2H, J ) 2.5 Hz, CH2), 7.44-7.47
(m, 2H, Ar-H), 7.57-7.60 (m, 1H, Ar-H), 8.06-8.09 (m, 2H,
Ar-H); 13C NMR (CDCl3): δ ) 52.5 (CH2), 75.0 (CH), 77.8
(-C≡), 128.4 (2 Ar-C), 129.4 (Ar-C), 129.8 (2 Ar-C), 133.3 (Ar-
C), 165.8 (CdO).
3-Acetoxy-(2-propynyl)benzoate. 3-Acetoxybenzoic acid:
2.00 g (11.1 mmol), DCC: 2.52 g (12.2 mmol), DMAP: 244.0
mg (2.0 mmol), 2-propynol: 685.0 mg (12.2 mmol); Yield: 1.58
g (7.23 mmol), 65% colorless oil; IR (KBr, cm-1) 1767 (CdO),
1729 (CdO); MS (EI, 35 °C): m/z (%) ) 218 (1) [M+], 105 (100);
1H NMR (CDCl3): δ ) 2.32 (s, 3H, CH3), 2.52 (t, 1H, J ) 2.4
Hz, CH), 4.92 (d, 2H, J ) 2.4 Hz, CH2), 7.31-7.33 (ddd 1H,
J ) 8.0 Hz, J ) 2.2 Hz, J ) 0.9 Hz, Ar-H), 7.47 (dd, 1H, J )
8.0 Hz, J ) 8.1 Hz, Ar-H), 7.79 (dd, 1H, J ) 2.2 Hz, J ) 0.9
Hz, Ar-H), 7.94-7.96 (ddd, 1H, J ) 8.0 Hz, J ) 0.9 Hz, J )
1.0 Hz, Ar-H); 13C NMR (CDCl3): δ ) 21.0 (CH3), 52.7 (CH2),
75.2 (CH), 77.5 (-C≡), 123.1 (Ar-C), 126.8 (Ar-C), 127.3 (Ar-
C), 129.5 (Ar-C), 131.0 (Ar-C), 150.7 (Ar-C), 164.9 (CdO), 169.2
(CdO).
2-Fluoro-(2-propynyl)benzoate. 2-Fluorobenzoic acid: 2.00
g (14.3 mmol), DCC: 3.24 g (15.7 mmol), DMAP: 244.0 mg
(2.0 mmol), 2-propynol: 880.0 mg (15.7 mmol); Yield: 2.18 g
(12.2 mmol), 86% colorless oil; IR (KBr, cm-1) 1729 (CdO); MS
(EI, 35 °C): m/z (%) ) 178 (5) [M+], 123 (100); 1H NMR
(CDCl3): δ ) 2.52 (t, 1H, J ) 2.5 Hz, CH), 4.94 (d, 2H, J ) 2.5
Hz, CH2), 7.13-7.18 (m, 1H, Ar-H), 7.20-7.24 (m, 1H, Ar-H),
7.52-7.57 (m, 1H, Ar-H), 7.95-8.00 (m, 1H, Ar-H); 13C NMR
Conclusion
To our best knowledge, Co-ASS represents the most
cytotoxic drug among the COX-inhibiting anticancer
drugs so far. It can be stated that derivatization of
aspirin with propargylic alcohol and subsequent com-
plexation to Co2(CO)6 afforded a compound, which is a
more effective COX inhibitor than aspirin and which
exhibits elevated cytotoxic activity. The COX inhibitory
activity as well as the cytotoxic activity of closely related
compounds was definitely lower. Further studies to
extend this derivatization concept to other NSAIDs are
going on.
Experimental Section
General. Chemicals were purchased from Sigma, Fluka,
or Acros. MCF-7 and MDA-MB 231 cells were maintained as
described.3 Drugs were freshly prepared as stock solutions in
dimethylformamide (DMF) and diluted with cell culture media
or buffer when used for the biochemical experiments (DMF
0.1 V/V). N-(2-Propynyl)phthalimide, 2-acetoxy-N-(2-propynyl)-
benzamide, Prop-ASS, Co-ASS, Co-SAL, and Co-Prop were
synthesized and analyzed as described previously.2,14-16 Protein
quantification was performed by the method of Bradford32
using human serum albumin (Sigma) for calibration purposes.
Cobalt amounts were determined by graphite furnace atomic
absorption spectroscopy5 using a detection wavelength of 240.7
nm. Elemental analysis: Perkin-Elmer 240C, IR spectra: ATI
Mattson Genesis, NMR spectra: Avance/DPX 400 (Bruker),
MS spectra: CH-/A- Varian MAT (70 eV).
General Method for the Preparation of Benzoic Acid
Esters. The respective benzoic acid derivative (10-17 mmol),
dicyclohexylcarbodiimide (DCC, 12-18 mmol), 4-(dimethyl-
amino)pyridine (DMAP, 2 mmol), and 2-propynol (12-16
mmol) were stirred in 30 mL of dry CH2Cl2 until no further
product formation was observed by thin-layer chromatography
(approximately 2 h). The mixture was filtered, the precipitate
was rinsed with CH2Cl2, and the combined filtrates were
evaporated. The resulting oil was purified by flash column
(CDCl3): δ ) 52.6 (CH2), 75.2 (CH), 77.5 (-C≡), 117.1 (JCF
)
22.2 Hz, Ar-C), 118.0 (JCF ) 9.5 Hz, Ar-C), 124 (JCF ) 3.9 Hz,
Ar-C), 132.3 (Ar-C), 135.0 (JCF ) 9.3 Hz, Ar-C), 162.1 (JCF
260.8 Hz, Ar-C), 163.5 (CdO).
)
General Method for Preparation of Cobalt-Alkyne
Complexes. Cobalt-alkyne complexes were prepared accord-
ing to an already described procedure.3 The alkyne (0.5-1.0
mmol) was dissolved in 10 mL of dry THF or diethyl ether.
Dicobaltoctacarbonyl was added in excess, and the solution
was stirred at room temperature until no more product
formation was observed by thin-layer chromatography. One
gram of silica gel was added, and the mixture was evaporated
to dryness. The dark colored products were isolated by flash
column chromatography on silica gel (mobile phase: diethyl
ether/petroleum ether). Yields were not optimized.
[(2-Propynyl)benzoate]hexacarbonyldicobalt
(Co-
Benz). (2-Propynyl)benzoate: 100.0 mg (0.62 mmol), dico-
baltoctacarbonyl: 356.0 mg (0.94 mmol); Yield: 231 mg (0.16
mmol), 83% dark-red crystals (mp: 86-87 °C); IR (KBr, cm-1
)
2095, 2058, 2044, 2033, 2002 (Co-CO), 1717 (CdO); MS (EI,

Antitumor-Active Cobalt-Alkyne Complexes
Journal of Medicinal Chemistry, 2005, Vol. 48, No. 2 627
100 °C): m/z (%) ) 418 (3) [M+ - CO], 390 (31) [M+ - 2CO],
362 (34) [M+ - 3CO], 334 (24) [M+ - 4CO], 306 (7)
[M+ - 5CO], 278 (98) [M+ - 6CO], 59 (100) [Co]; 1H NMR
(CDCl3): δ ) 5.53 (br, 2H, CH2), 6.13 (br, 1H, CH), 7.45-7.47
(m, 2H, Ar-H), 7.56-7.58 (br, 1H, Ar-H), 8.10-8.12 (m, 2H,
Ar-H). Anal. (C16H8Co2O8) C, H, N
[3-Acetoxy-(2-propynyl)benzoate]hexacarbonyldico-
balt (Co-3-Acetbenz). 3-Acetoxy-(2-propynyl)benzoate: 107.0
mg (0.49 mmol), dicobaltoctacarbonyl: 280.0 mg (0.74 mmol);
Yield: 164 mg (0.33 mmol), 66% brown-red crystals (mp: 51-
52 °C); IR (KBr, cm-1) 2097, 2058, 2029 (Co-CO), 1770
(CdO), 1721 (CdO); MS (EI, 100 °C): m/z (%) ) 476 (4)
[M+ - CO], 448 (24) [M+ - 2CO], 420 (25) [M+ - 3CO], 392
(5) [M+ - 4CO], 365 (35) [M+ - 5CO], 343 (100), 336 (2)
[M+ - 6CO]; 1H NMR (CDCl3): δ ) 2.31 (s, 3H, CH3), 5.52
(br, 2H, CH2), 6.13 (br, 1H, CH), 7.26 (br, 1H, Ar-H), 7.40-
7.60 (br, 1H, Ar-H), 7.75-7.83 (br, 1H, Ar-H), 7.92-8.07 (br,
1H, Ar-H). Anal. (C18H10Co2O10) C, H, N.
biomass plate was subtracted from the mean absorption of
each experiment and control. The corrected control was set
100%. IC50 was determined as that concentration causing 50%
inhibition of cell proliferation and calculated as mean of at
least two independent experiments.
Drug Lipophilicity (log kw). Log kw values were deter-
mined by an RP-HPLC method as described previously.5
Cellular Uptake Studies. MCF-7 and MDA-MB 231 cell
cultures were grown in 24-well plates at 37 °C in 5% CO2
atmosphere until at least 70% confluency. The medium was
removed and replaced by a 2.0 µM drug-containing one (500
µL). After a proper incubation time at 37 °C in 5% CO2
atmosphere, the drug-containing medium was removed and
the cell monolayer was washed with 500 µL of phosphate-
buffered saline pH 7.4. Distilled water (400 µL) was added,
and after 10-15 min of incubation, the cells were lysed by
means of a sonotrode. An aliquot of each lysate was removed
for protein determination and stored at -20 °C. Probes (200
µL) for cobalt determination by atomic absorption measure-
ment were stabilized by addition of 20 µL of HNO3 (13%) and
20 µL of Triton X-100 (1%) and stored at 4 °C until measure-
ment. Results were expressed as means of 6-12 wells as pmol
drug per µg cellular protein. The cellular drug concentration
can be estimated thereof as described by us.17
DNA-Binding Studies. Precipitation of drug DNA adducts
was performed according to a described method21 with some
modifications: Salmon testes DNA (Sigma) was dissolved in
phosphate-buffered saline pH 7.4 and drugs were added as
stock solutions in DMF. The final solutions contained 40.6 µM
drug, 250 µg/mL salmon testes DNA and 0.1% V/V DMF. After
vortexing, the solutions were incubated at 37 °C in a water
bath for 4 h. Aliquots of 200 µL were mixed with 100 µL of 0.9
M sodium acetate and three volumes of ice cold ethanol.
Samples were stored at -20 °C for 30 min. The pellets were
isolated by centrifugation (5000 U/min, 10 min, 4 °C) and
resuspended in 300 µL of 0.3 M sodium acetate. 900 µL of ice
cold ethanol were added, and the precipitate was collected after
centrifugation (5000 U/min, 10 min, 4 °C). Samples were
washed twice with ice cold ethanol and were stored at -20
°C. The pellets were dissolved in 500 µL of water (twice
distilled), and the DNA content was determined by absorption
reading at 260 nm in a UV-microplate reader (Flashscan
AnalytikJena AG). Salmon testes DNA dissolved in water
(twice distilled) was used for calibration purposes. Samples
(200 µL) for cobalt determination by atomic absorption spec-
troscopy were stabilized by addition of 20 µL of Triton X-100
(1%). The amount of drugs bound to DNA was expressed as
pmol of drug per µg of DNA. Results were calculated as means
of at least two independent experiments, which were per-
formed with two replicates.
[2-Fluoro-(2-propynyl)benzoate]hexacarbonyldico-
balt (Co-2F-Benz). 2-Fluoro-(2-propynyl)benzoate: 100.0 mg
(0.56 mmol), dicobaltoctacarbonyl: 320.0 mg (0.842 mmol);
Yield: 215 mg (0.432 mmol), 83% red crystals (mp: 88-89
°C); IR (KBr, cm-1) 2097, 2037, 2001 (Co-CO), 1726 (CdO);
MS (EI 100 °C): m/z (%) ) 436 (3) [M+ - CO], 408 (26) [M+
-
-
2CO], 380(28) [M+ - 3CO], 352(10) [M+ - 4CO], 324(2) [M+
1
5CO], 296 (100) [M+ - 6CO]; H NMR (CDCl3): δ ) 5.54 (br,
2H, CH2), 6.12 (br, 1H, CH), 7.13-7.23 (m, 2H, Ar-H), 7.52-
7.56 (br, 1H, Ar-H), 7.98-8.02 (br, 1H, Ar-H). Anal. (C16H7-
Co2FO8) C, H, N.
[2-Acetoxy-N-(2-propynyl)benzamide]hexacarbonyl-
dicobalt (Co-ASSAM). 2-Acetoxy-N-(2-propynyl)benzamide:
100.0 mg (0.46 mmol), dicobaltoctacarbonyl: 260.0 mg (0.76
mmol); Yield: 14 mg (0.03 mmol), 7% red crystals (mp: 94
°C); IR (KBr, cm-1) 2025, 2054, 2094 (Co-CO), 1627 (CdO),
1606 (CdO); MS (EI 100 °C): m/z (%)
) 447 (12)
[M+ - 2CO], 419 (97) [M+ - 3CO], 391 (57) [M+ - 4CO], 363
1
(18) [M+ - 5CO], 335 (100) [M+ - 6CO]; H NMR: δ ) 2.35
(s, 3H, CH3), 4.78 (br, 2H, CH2), 6.10 (br, 1H, CH), 6.60 (s,
1H, NH), 7.14 (br, 1H, Ar-H), 7.20-7.30 (br, 1H, Ar-H), 7.48
(br, 1H, Ar-H), 7.75 (br, 1H, Ar-H). Anal. (C18H11Co2NO9) C,
H, N.
[N-(2-Propynyl)phthalimide]hexacarbonyldicobalt (Co-
Phthal). N-(2-Propynyl)phthalimide: 0.50 g (2.7 mmol), di-
cobaltoctacarbonyl: 1.03 g (3.0 mmol): Yield: 1.097 g (2.3
mmol), 85% red-brown crystals (mp: 105 °C); IR (KBr, cm-1
)
2095, 2054, 2036, 2016 (Co-CO); 1712 (CdO); MS (EI, 100
°C): m/z (%) ) 443 (12) [M+ - CO], 415 (48) [M+ - 2CO], 387
(50) [M+ - 3CO], 359 (29) [M+ - 4CO], 331 (67) [M+ - 5CO],
303 (90) [M+ - 6CO], 28(100) [CO]; 1H NMR (CDCl3): δ )
5.04 (br, 2H, CH2), 6.08 (br, 1H, CH), 7.74 (m, 2H, Ar-H), 7.88
(m, 2H, Ar-H). Anal. (C17H7Co2NO8) C, H, N.
Nuclear Drug Content. The nuclei of the tumor cells were
isolated according to a described procedure33 with some
modifications: Cells were grown in 175 cm2 cell culture flasks
until at least 70% confluency. The medium was removed and
replaced with 10 mL of medium containing 2.0 µM drug. After
24 h of incubation at 37 °C in humidified atmosphere, the drug-
containing medium was removed, cells were trysinized, resus-
pended in 10 mL of cell culture medium, and isolated by
centrifugation (1500 U/min, 5 min), and 0.5-1.0 mL of 0.9%
NaCl solution was added. After centrifugation (1500 U/min, 5
min), pellets were resuspended in 300 µL of RSB-1 (0.01 M
Tris-HCl, 0.01 M NaCl, 1.5 mM MgCl2, pH 7.4) and left for 10
min in an ice-bath. Swollen cells were centrifuged (2000 U/min,
5 min), resuspended in 300 µL of RSB-2 (RSB-1 containing
each 0.3% V/V Nonidet-P40 and sodium desoxycholate), and
homogenized by 10-15 up/down-pushes in an 1 mL syringe
with needle. Aliquots of 50 µL of the homogenisate were
removed for determination of the total cobalt content and
mixed with 500 µL of a 10 g/L EDTA solution. The homogeni-
sate was centrifuged at 2500 U/min for 5 min and the resulting
crude nuclei were taken up in 150 µL of 0.25 M sucrose
containing 3 mM CaCl2. The suspension was underlayed with
150 µL of 0.88 M sucrose and centrifuged 10 min at 2500
U/min. The nuclei pellets were stored at -20 °C or immediately
Cytotoxicity Experiments (IC50). Each 100 µL of 10 000
cells/mL (MCF-7) or 5 000 cells/mL (MDA-MB 231) were
incubated in 96-well plates at 37 °C in 5% CO2/95% air
atmosphere for 72 h. One plate was used for the determination
of the initial cell biomass and was treated in the following
way: the medium was removed, cells were fixed by a 20-30
min incubation with 100 µL of glutardialdehyde solution (0.5
mL glutardialdehyde + 12.5 mL phosphate-buffered saline pH
7.4), the wells were emptied, 180 µL phosphate-buffered saline
pH 7.4 were added, and the plate was stored at 4 °C until
further treatment. In the “experimental” plates the medium
was replaced with medium containing the drugs in graded
concentrations (six replicates). After further incubation for 72
h (MDA-MB 231) or 96 h (MCF-7), these plates were treated
as described above. The cell biomass was determined by crystal
violet staining according to the following procedure: the
phosphate-buffered saline pH 7.4 was removed, 100 µL of a
0.02 M crystal-violet solution were added, and plates were
incubated for 30 min at room temperature, washed three times
with water, and incubated on a softly rocking rotary shaker
with 180 µL of ethanol (70%) for further 3-4 h. Absorption
was recorded in a microplate reader at 590 nm (Flashscan
AnalytikJena AG). The mean absorption of the initial cell

628 Journal of Medicinal Chemistry, 2005, Vol. 48, No. 2
Ott et al.
(7) Neumann, C.; Boubakari; Gru¨nert, R.; Bednarski, P. J. Nico-
tinamide adenine dinucleotide phosphate-regenerating system
coupled to a glutathione-reductase microtiter method for deter-
mination of total glutathione concentrations in adherent growing
cancer cell lines. Anal. Biochem. 2003, 320, 170-178.
(8) Gromer, S.; Arscott, D.; Williams, C. H.; Schirmer, R.; Becker,
K. Human placenta thioredoxin reductase. J. Biol. Chem. 1998,
273, 20096-20101.
(9) Becker, K.; Herold-Mende C.; Park, J. J.; Lowe, G.; Schirmer,
R. H. Human thioredoxin reductase is efficiently inhibited by
(2, 2′:6′,2′′-terpyridine)platinum(II) complexes. Possible implica-
tions for a novel antitumor strategy. J. Med. Chem. 2001, 44,
2784-2792.
(10) Giardello, F. M.; Offerhaus, G. J. A.; DuBois, R. N. The role of
nonsteroidal antiinflammatory drugs in colorectal cancer pre-
vention. Eur. J. Cancer 1995, 31A, 1071-1076.
(11) Bange, J.; Zwick, E.; Ullrich, A. Molecular targets for breast
cancer therapy and prevention. Nature Med. 2001, 7, 548-552.
(12) Coussens, L. M.; Werb, Z. Inflammation and cancer. Nature
2002, 420, 860-867.
(13) Harris, R. E.; Chlebowski, R. T.; Jackson, R. D.; Frid, D. J.;
Ascenseo, J. L.; Anderson, G.; Loar, A.; Rodabough, R. J.; White,
E.; McTiernan, A. Breast cancer and nonsteroidal antiinflam-
matory drugs: Prospective results from the women’s health
initiative. Cancer Res. 2003, 63, 6096-6101.
(14) Greenfield, H.; Sternberg, H. W.; Friedel R. A.; Wotiz J. H.;
Markby, R.; Wender, I. Acetylenic dicobalt hexacarbonyls. Or-
ganometallic compounds derived from alkynes and dicobalt
octacarbonyl. J. Am. Chem. Soc 1956, 78, 120-124.
(15) Valliant, J. F.; Schaffer, P. A new approach for the synthesis of
isonitrile carborane derivatives. Ligands for metal based boron
neutron capture therapy (BNCT) and boron neutron capture
synovectomy (BCNS) agents. J. Inorg. Biochem. 2001, 85, 43-
51.
(16) Freedman, J.; Huber, E. W. An unusual oxazole formation. J.
Heterocyclic Chem. 1990, 27, 343-346.
(17) Gust R.; Schnurr, B.; Krauser, R.; Bernhardt, G.; Koch, M.;
Schmid, B.; Hummel, E.; Scho¨nenberger, H. Stability and
cellular studies of [rac-1,2-bis(4-fluorophenyl)ethylenediamine]-
[cyclobutane-1,1-dicarboxylato]platinum(II), a novel, highly ac-
tive carboplatin derivative. J. Cancer Res. Clin. Oncol. 1998, 124,
585-597.
(18) Ghezzi, A. R.; Aceto, M.; Cassino, C.; Gabano, E.; Osella, D.
Uptake of antitumor platinum(II)-complexes by cancer cells,
assayed by inductively coupled plasma spectrometry (ICP-MS).
J. Inorg. Biochem. 2004, 98, 73-78.
(19) Vessieres, A.; Top, S.; Vaillant, C.; Osella, D.; Mornon, J. P.;
Jaouen, G. Transition metal cluster-modified estradiols for the
investigation of proteins: use at the estradiol receptor. Angew.
Chem., Int. Ed. Engl. 1992, 31, 754-755.
(20) Osella, D.; Cavigiolio, G.; Vincenti, M.; Vessieres, A.; Laios, I.;
Leclercq, G.; Napolitano, E.; Fiaschi, R.; Jaouen, G. The first
organometallic derivative of 11â-ethynylestradiol, a potential
high-affinity marker for the estrogen receptor. J. Organomet.
Chem. 2000, 596, 242-247.
(21) Natarajan, G.; Malathi, R.; Holler, E. Increased DNA-binding
activity of cis-1,1-cyclobutanedicarboxylatodiammineplatinum-
(II) (Carboplatin) in the presence of nucleophiles and human
breast cancer MCF-7 cell cytoplasmic extracts: Activation theory
revisited. Biochem. Pharmacol. 1999, 58, 1625-1629.
(22) Yu, H.-G.; Huang, J.-A.; Yang, Y.-N.; Huang, H.; Luo, H.-S.; Yu,
Y. P.; Meier, J. J.; Schrader, H.; Bastian, A.; Schmidt, W. E.;
Schmitz, F. The effects of acetylsalicylic acid on proliferation,
apoptosis, and invasion of cyclooxygenase-2 negative colon cancer
cells. Eur. J. Clin. Invest. 2002, 32, 838-846.
dissolved in 200 µL of 10 g/L EDTA solution and disrupted by
use of a sonotrode. The cobalt content of the samples was
determined by atomic absorption spectroscopy and the protein
content by the method of Bradford.32 Each experiment con-
sisted of the data obtained from three pellets. Results are
expressed as means of two independent experiments as pmol
drug per µg nuclear protein. The ratio of drug taken up into
the nucleus is calculated as percentage of the total cobalt
content of each pellet.
Glutathione Reductase Activity. The influence on cel-
lular glutathione reductase activity was assayed according to
literature procedures.8,9 In short: MCF-7 cells were grown in
24-well plates until at least 70% confluency. The medium was
removed, and the cells were washed with phosphate-buffered
saline (pH 7.4). Subsequently, 500 µL of cell culture medium
containing 8.0 µM of the drugs or only DMF (control) were
added. After 6 h of exposure, the drug-containing media were
removed and the cells were washed with phosphate-buffered
saline (pH 7.4). 250 µL of phosphate buffer (pH 7.0) was added,
and the cells were lysed after 15 min by use of a sonotrode.
Lysates were centrifuged (4000 U/min, 10 min: room temp)
and assayed directly thereafter. Aliquots of 40 µL of each lysate
were added to 290 µL of a mixture of NADPH (final concentra-
tion: 200 µM) and reduced glutathione (final concentration:
1 mM). The consumption of NADPH was monitored as the
decrease in absorbance at 340 nm. Protein concentrations of
the lysates were determined by the Bradford method.32 En-
zyme activity was calculated as nmol consumed NADPH/
(min × µg protein) and expressed as percentage of the control.
Results are calculated as means of at least two independent
experiments (n ) 6).
Apoptosis Detection. Apoptosis was measured after 3 h
of drug exposure by use of an “Annexin-V-Fluos Staining Kit”
(Roche) and after 5 days of drug exposure by an “ss-DNA
Apoptosis ELISA Kit” (Chemicon Int.) according to the manu-
facturers instructions. Two independent experiments were
performed at drug concentrations of 5 µM and 10 µM.
Estrogen Receptor Binding Affinity. The estrogen re-
ceptor binding affinity (RBA-value) was monitored as already
described by us.34
Inhibition of COX Enzymes. The inhibition of isolated
ovine COX-1 and human recombinant COX-2 was determined
with 10 µM and 200 µM of the respective compounds by ELISA
(“COX inhibitor screening assay”, Cayman Chemicals). Experi-
ments were performed according to the manufacturer’s in-
structions. Results were calculated as the means of duplicate
determinations.
Acknowledgment. We want to thank our students
Maxi Baacke and Annika Gross, who performed part of
their education in our laboratories, for their helpful
assistance in some of the experiments.
(23) Kalgutkar, A. S.; Crews, B. C.; Rowlinson, S. W.; Marnett, A.
B.; Kozak, K. R.; Remmel, R. P.; Marnett, L. J. Biochemically
based design of cyclooxygenase-2 (COX-2) inhibitors: Facile
conversion of nonsteroidal antiinflammatory drugs to potent and
highly selective COX-2 inhibitors. PNAS 2000, 97, 925-930.
(24) Kalgutkar, A. S.; Marnett, A. B.; Crews, B. C.; Remmel, R. P.;
Marnett, L. J. Ester and amide derivatives of the nonsteroidal
antiinflammatory drug indomethacin, as selective cyclooxy-
genase-2 inhibitors. J. Med. Chem. 2000, 43, 2860-2870.
(25) Sorensen, J. R. J. Copper chelates as possible active forms of
the antiarthritic agents. J. Med. Chem. 1976, 19, 135-148.
(26) Weder, J. E.; Dillon, C. T.; Hambley, T. W.; Kennedy, B. J.; Lay,
P. A.; Biffin, J. R.; Regtop, H. L.; Davies, N. M. Copper complexes
of nonsteroidal antiinflammatory drugs: an opportunity yet to
be realized. Coord. Chem. Rev. 2002, 232, 95-126.
(27) Dillon, C. T.; Hambley, T. W.; Kennedy, B. J.; Lay, P. A.; Zhou,
Q.; Davies, N. M.; Biffin, J. R.; Regtop, H. L. Gastrointestinal
toxicity, antiinflammatory activity, and superoxide dismutase
activity of copper and zinc complexes of the antiinflammatory
drug indomethacin. Chem. Res. Toxicol. 2003, 16, 28-37.
(28) Kalgutkar, A. S.; Crews, B. C.; Rowlinson, S. W.; Garner, C.;
Seibert, K.; Marnett, L. J. Aspirin-like molecules that covalently
inactivate cyclooxygenase-2. Science 1998, 280, 1268-1270.
References
(1) Reedijk, J. Medicinal applications of heavy metal compounds.
Curr. Opin. Chem. Biol. 1999, 3, 236-240.
(2) Jung, M.; Kerr, D. E.; Senter, P. D. Bioorganometallic chemistry
- synthesis and antitumor activity of cobalt carbonyl complexes.
Arch. Pharm. Pharm. Med. Chem. 1997, 330, 173-176.
(3) Schmidt, K.; Jung, M.; Keilitz, R.; Schnurr, B.; Gust, R. Acet-
ylenehexacarbonyldicobalt complexes, a novel class of antitumor
drugs. Inorg. Chim. Act. 2000, 306, 6-16.
(4) Roth, T.; Eckert, C.; Fiebig, H. H.; Jung, M. Comparative action
of cobalt carbonyl complexes on cancer cells using human tumor
xenografts. Anticancer Res. 2002, 22, 2281-2284.
(5) Ott, I.; Kircher, B.; Gust, R. Investigations on the effects of
cobalt-alkyne complexes on leukemia and lymphoma cells:
cytotoxicity and cellular uptake. J. Inorg. Biochem. 2004, 98,
485-489.
(6) Perquin, M.; Oster, T.; Maul, A.; Froment, N.; Untereiner, M.;
Bagrel, D. The glutathione-related detoxification system is
increased in human breast cancer in correlation with clinical
and histopathological features. J. Cancer Res. Clin. Oncol. 2001,
127, 368-374.

Antitumor-Active Cobalt-Alkyne Complexes
Journal of Medicinal Chemistry, 2005, Vol. 48, No. 2 629
(29) Kalgutkar, A. S.; Kozak, K. R.; Crews, B. C.; Hochgesang, P.;
Marnett, L. J. Covalent modification of cyclooxygenase-2 (COX-
2) by 2-acetoxyphenyl alkyl sulfides, a new class of selective
COX-2 inactivators. J. Med. Chem. 1998, 41, 4800-4818.
(30) Waskevich, C.; Blumenthal, R. D.; Li, H.; Stein, R.; Goldenberg,
D. M.; Burto, J. Celecoxib exhibits the greatest potency amongst
cyclooxygenase (COX) inhibitors for growth inhibition of COX-
2-negative hematopoietic and epithelial cell lines. Cancer Res.
2002, 62, 2029-2033.
(31) Jones, M. K.; Wang, H.; Peskar, B. M.; Levin, E.; Itani, R. M.;
Sarfeh, J.; Tarnawski, A. S. Inhibition of angiogenesis by
nonsteroidal antiinflammatory drugs: Insight into mechanisms
and implications for cancer growth and ulcer healing. Nat. Med.
1999, 5, 1418-1423.
(32) Bradford, M. M. A rapid and sensitive method for the quanti-
tation of microgram quantities of protein utilizing the principle
of protein-dye binding. Anal. Biochem. 1976, 72, 248-254.
(33) Muramatsu, M.; Onishi, T. Rapid isolation of nucleoli from
detergent-purified nuclei of tumor and tissue culture cells.
Methods Cell Biol. 1977, 15, 221-234.
(34) Gust, R.; Busch, S.; Keilitz, R.; Schmidt, K.; von Rauch, M.
Investigations on the influence of halides substituents on the
estrogen receptor interaction of 2,4,5-triarylimidazoles. Arch.
Pharm. Pharm. Med. Chem. 2003, 336, 456-465.
JM049326Z