Potent anticancer activity and possible low toxicity of platinum(II) complexes with functionalized 1,1-cyclobutanedicarboxylate as a leaving ligand

Article abstract of DOI:10.1002/chem.201404090

Two platinum(II) complexes, DN603 and DN604, were designed and prepared by using 3-oxocyclobutane-1,1-dicarboxylate as a ligand. The compounds were prepared according to the concept that incorporation of a functionalized moiety in the leaving ligand that

Full text of DOI:10.1002/chem.201404090

DOI: 10.1002/chem.201404090
Full Paper
&
Drug Discovery
Potent Anticancer Activity and Possible Low Toxicity of
Platinum(II) Complexes with Functionalized 1,1-
Cyclobutanedicarboxylate as a Leaving Ligand
Jian Zhao,[a] Shaohua Gou,*[a, b] and Fengfan Liu[a]
Abstract: Two platinum(II) complexes, DN603 and DN604,
were designed and prepared by using 3-oxocyclobutane-1,1-
dicarboxylate as a ligand. The compounds were prepared ac-
cording to the concept that incorporation of a functionalized
moiety in the leaving ligand that did not affect its coordina-
tion bonding to the metal atom would play a key role in the
anticancer activity of the resulting platinum complex. The
newly prepared compounds were found to show potent in
vitro anticancer activity comparable to cisplatin and oxalipla-
tin; especially DN604, which exhibited low acute toxicity
similar to carboplatin, and presented acceptable solubility
and stability in water. Chemical and biological results indi-
cated that the functionalized moiety, uncoordinated, led to
potent anticancer activity and low apparent toxicity of the
platinum complexes by affecting the kinetic properties of
the compounds.
Introduction
plexes with labile anionic leaving groups such as the chloride
anion in cisplatin, which hydrolyze relatively fast, are much
more toxic, because the resulting reactive platinum intermedi-
ates can efficiently bind biomolecules such as sulfur-containing
amino acids and peptides or proteins in the blood, and their
interactions with proteins in organs such as the kidney can
bring about severe adverse effects.[5] Complexes with nonlabile
anionic leaving groups such as the dicarboxylate moiety in car-
boplatin, which hydrolyze slowly, are divided in two cases:
i)water-soluble platinum complexes, which are stable and have
a longer half-time in the blood, are less toxic because they can
be excreted efficiently via the kidneys, but their antitumor abil-
ities are clearly reduced; (ii) water-insoluble (lipophilic) plati-
num complexes, which hydrolyze slowly, are cleared much
more easily from the blood through tissue penetration so that
some typical toxic side effects derived from cisplatin-based
therapy are absent due to the reduced requirement for renal
excretion.[4] So far, thousands of platinum complexes have
been designed, synthesized, and evaluated biologically, but
none have entered worldwide clinical use except for oxaliplatin
(which exhibits potent anticancer activity as a result of the
presence of the (1R,2R-cyclohexanediamine)platinum(II) frag-
ment as well as because of its good aqueous solubility and sta-
bility derived from the dicarboxylato leaving ligand), irrespec-
tive of whether they are in agreement with or against the clas-
sic structure-activity-relationship summarized by Cleare and
Hoeschele.[6,7] Thus, it seems to be impossible to obtain a plati-
num anticancer agent with characteristics of both potent anti-
cancer activity such as cisplatin and low toxicity together with
good water solubility and stability such as carboplatin.
Cisplatin, cis-diamminedichloroplatinum(II), is one of the most
effective anticancer drugs available for the treatment of a varie-
ty of solid tumors, but its severe side effects against kidneys
(nephrotoxicity) and gastrointestinal tract have limited its wide
application in practice.[1] Carboplatin, cis-diammine(1,1-cyclo-
butanedicarboxylato)platinum(II), a next-generation platinum
anticancer drug, was designed to overcome the drawbacks of
cisplatin, especially the toxicity. Carboplatin is indeed less toxic
than cisplatin, which can be attributed to the inclusion of the
1,1-cyclobutyldicaboxylate moiety as leaving ligand. This group
is coordinated to the platinum atom through a six-member
ring, promoting the stability and aqueous solubility of carbo-
platin in comparison with cisplatin.[2,3] However, the anticancer
activity of carboplatin turns out to be notably weaker than
that of cisplatin against a number of tumors despite the same
non-leaving ligand (ammine) in both drugs.
It is well known that the nature of the leaving ligand plays
a significant role in determining the toxicity and side effects of
a platinum drug, because the leaving capability of different
leaving ligands from the binding platinum atom can cause re-
markably altered biodistribution in vivo.[4] In general, com-
[a] J. Zhao, Prof. Dr. S. Gou, F. Liu
Pharmaceutical Research Center
School of Chemistry and Chemical Engineering
Southeast University, Nanjing 211189 (P.R. China)
Fax: (+86)83272381
E-mail: sgou@seu.edu.cn
[b] Prof. Dr. S. Gou
Jiangsu Province Hi-Tech Key Laboratory
for Bio-medical Research
Southeast University, Nanjing 211189 (P.R. China)
We have been studying antitumor platinum complexes with
the target of obtaining a drug candidate with high efficacy
and low toxicity.[8,9] One of our main strategies has been to
take advantage of successful carrier ligands such as ammine in
Supporting information for this article is available on the WWW under
http://dx.doi.org/10.1002/chem.201404090.
Chem. Eur. J. 2014, 20, 15216 – 15225
15216
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
cisplatin/carboplatin or 1R,2R-cyclohexanediamine in oxalipla-
tin by applying novel carboxylato ligands as leaving groups in
our designed platinum complexes.[10,11] On one hand, the
water solubility of platinum complexes is considered to be
a crucial factor for their possible clinical application, because
many complexes with high anticancer activity failed in clinical
trials due to their poor aqueous solubility, for example, cis-di-
chlorobis(cyclopentyl amine)platinum(II).[12] On the
other hand, lipophilicity, which determines the per-
meability of the platinum complexes from the blood
through tumor tissue penetration is another key
factor.[4] Hence, the balance between the hydrophilic-
ity and lipophilicity of a platinum complex is essential
for the bioavailability and biodistribution of the com-
pound in the human body. We are convinced that
with unique physical and chemical properties; their potent an-
ticancer activities are also described.
Results and Discussion
3-Oxocyclobutane-1,1-dicarboxylic acid was prepared by fol-
lowing the procedure shown in Scheme 1, and the platinum
Scheme 1. Preparation of 3-oxocyclobutane-1,1-dicarboxylic acid. Reagents and condi-
tions: a) HgCl2, 1508C, 9 h; b) CH2(COOEt)2, NaOEt, anhydrous ethanol, reflux, 24 h; c) H2
(20 atm), 10% Pd/C, 508C, 5 h; d) PCC, CH2Cl2, RT, 12 h; e) KOH, RT, 12 h; HCl.
modification of the leaving ligand in platinum anti-
cancer compounds remains an effective way to dis-
cover new platinum-based drugs and, to this end,
a number of functionalized species such as hydroxyl,
ether, ester or ketone groups have been introduced into car-
boxylato ligands to take the role of leaving ligands.[13–15]
complexes DN603 and DN604 were subsequently synthesized
according to our previous patent.[18] The final product of
DN604 was recrystallized from water, in which single-crystals
were obtained for X-ray diffraction determination. The purity
of DN604 was more than 99.5%, established by RP-HPLC analy-
As we know, the application of 1,1-cyclobutanedicarboxylate
in platinum complexes can improve their stability and water
solubility as compared with chloride anions. In addition, it has
been reported that the stacking interaction between the cyclo-
butane ring of 1,1-cyclobutanedicarboxylate and the sugar
group of DNA is responsible for the increased anticancer activi-
ty of carboplatin.[16,17] However, the formation of a stable six-
membered ring between a Pt atom and 1,1-cyclobutanedicar-
boxylate can reduce the kinetic reactivity of the resulting plati-
num complex and lead to relatively low anticancer efficacy.
With the purpose of retaining the advantage of 1,1-cyclobuta-
nedicarboxylate as a leaving ligand while also improving the
kinetic properties of the resulting platinum complexes, we in-
tended to introduce a carbonyl group to the skeleton of 1,1-
cyclobutanedicarboxylate. According to our assumption, the
introduction of a functionalized moiety (carbonyl group)
would be expected to help achieve the desired balance be-
tween hydrophilicity and lipophilicity of the resulting platinum
complex and it may improve the kinetic properties as well as
the DNA-binding ability of the derivatives. Herein reported are
two such platinum complexes (DN603 and DN604; Figure 1)
Figure 2. Molecular structure of DN604. Selected bond lengths [ꢁ] and
angles [8]: Pt1ÀO1 2.027(4), Pt1ÀN1 2.017(5), O3ÀC4 1.189(10); O1-Pt1-N
188.07(18), O1-Pt1-a 91.01(16), O1-Pt1-N1a 178.77(19), N1-Pt1-N1a 92.9(2).
sis (data not shown). The complexes were character-
1
ized by elemental analysis, IR, H and 195Pt NMR spec-
tra along with ESI-MS spectrometry. All the spectral
data were in good conformity with the proposed mo-
lecular structures of DN603 and DN604. X-ray crystal
data proved that DN604 has the expected molecular
structure in which the platinum atom is four-coordi-
nated in a square-planar geometry consisting of
PtN2O2 (Figure 2, Table S1, Table S2). It is noted that
the carbonyl oxygen atom is not involved in metal
coordination except through H-bonding with the hy-
drogen donor group (NH3) from the neighboring
molecule.
Figure 1. Related anticancer platinum complexes in this paper.
Chem. Eur. J. 2014, 20, 15216 – 15225
www.chemeurj.org
15217
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
The water solubility of DN603 and DN604 is approximately 4
and 8 mgmLÀ1 at 258C, respectively, which is clearly higher
than that of cisplatin (1 mgmLÀ1) as expected, but lower than
that of carboplatin (16 mgmLÀ1). The stability of DN604 in
water was studied by HPLC (Figure S1), and no detectable
change was observed in the period of 24 h, whereas a new
minor peak could be observed at approximately 6.0 min after
48 h. Notably, only 0.79% of DN604 hydrolyzed after 144 h,
demonstrating that this compound is very stable in water. To
explore the balance between hydrophilicity and lipophilicity,
the octanol–water partition coefficient (log P) of DN604 was
measured by a standard shake-flask method at pH 7.4, and the
aqueous fractions before and after partitioning were analyzed
by HPLC. As anticipated, DN604 was highly hydrophilic, with
a log P value of À1.17, which is close to the theoretical value
(À1.08) calculated by the software Marvin Sketch.[15] As com-
pared with cisplatin (À2.3) and carboplatin (À1.64),[19] DN604
showed improved lipophilicity, indicating that DN604 may pass
across the lipid bilayer of the membrane of tumor cells easily
and accumulate inside the cells.
Table 1. In vitro cytotoxicity (IC50 [mM]) of DN603, DN604, cisplatin, carbo-
platin, and oxaliplatin against human cancer and normal cell lines deter-
mined by the MTT method.
Cell
IC50 [mM][a]
DN603
DN604
Cisplatin Carboplatin Oxaliplatin
DU145[b]
K562[c]
26.7Æ2.3 13.8Æ1.3 25.5Æ2.8 108.0Æ7.6 20.4Æ1.1
24.6Æ2.5 19.3Æ0.8 10.9Æ0.3 117.7Æ9.6 39.9Æ3.5
24.7Æ2.5 9.7Æ0.6 65.1Æ5.8 208.0Æ19.4 19.5Æ0.8
3.1Æ0.2 2.2Æ0.3 1.3Æ0.1 12.6Æ1.1 52.8Æ4.5
65.4Æ2.3 8.3Æ2.3 27.9Æ1.9 38.3Æ3.2 26.7Æ2.1
24.9Æ2.2 19.9Æ1.6 33.8Æ3.4 87.8Æ7.9 27.7Æ1.8
MGC803[d]
NB4[c]
SE1116[e]
SK-BR-3[f]
SMMC-7721[g] 54.2Æ4.2 8.5Æ0.8 46.2Æ4.1 56.4Æ5.2 19.7Æ1.3
BGC823[d]
A549[h]
8.9Æ0.5 8.8Æ0.6 5.4Æ0.3 43.2Æ1.3
7.9Æ0.4
47.5Æ3.4 36.3Æ3.1 16.6Æ1.2 137.4Æ11.9 70.3Æ2.3
34.2Æ1.8 33.4Æ1.7 32.9Æ2.6 109.5Æ10.3 25.7Æ2.1
MCF-7[f]
A2780[i]
HUVEC[j]
7.3Æ0.3 12.6Æ0.8 27.4Æ2.5 17.2Æ1.1
6.6Æ0.3
68.9Æ5.3 30.2Æ2.8 11.2Æ1.1 101.6Æ10.1 16.3Æ1.2
[a] Values represent the meanÆSD from three independent experiment;
IC50 defined as the drug concentration required to inhibit 50% of cell
growth as determined by MTT assay after a 48 h exposure (MTT=3-(4,5-
dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide). [b] Human pros-
tate cancer cell line. [c] Human leukemia cell lines. [d] Human gastric
cancer cell line. [e] Human colorectal cancer cell line. [f] Human breast
carcinoma cell line. [g] Human hepatocellular carcinoma cell line.
[h] Human non-small-cell lung cancer cell line. [i] Human ovarian carcino-
ma cell line. [j] Normal human umbilical vein endothelial cell.
The cytotoxic activities against twelve human cell lines, in-
cluding the normal human umbilical vein endothelial cells,
HUVEC, have been investigated for both compounds together
with cisplatin, carboplatin, and oxaliplatin as positive agents,
by using the MTT method.[20] Each compound was tested at
eight different concentrations against all the cell lines
(Figure 3), and the corresponding IC50 values (dose required to
inhibit 50% cellular growth) were determined from the dose-
dependence of surviving cancer cells after cellular exposure to
the compounds for 48 h (Table 1). According to the IC50 values,
both DN603 and DN604 showed considerable cytotoxicity
against all the tested cancer cell lines. DN603 showed potent
cytotoxicity that was comparable to that of oxaliplatin, and su-
perior to that of carboplatin against the tested cell lines. More-
over, the cytotoxicity observed for DN604 was, in some cases,
Figure 3. Dose-dependent drug efficacy for DN603, DN604, cisplatin, carboplatin, and oxaliplatin on tumor and normal cells.
Chem. Eur. J. 2014, 20, 15216 – 15225 www.chemeurj.org ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
15218

Full Paper
superior to those of cisplatin against cancer cell lines, especial-
ly for MGC803 and SMMC-7721 cell lines. The significant cyto-
toxicity of DN604 confirms its high potential as an anticancer
agent and strengthens the potential of our strategy to discover
new platinum-based drugs by the modification of leaving li-
gands. Notably, DN604 possesses a structure that is analogous
to that of carboplatin, except that it has an extra carbonyl
moiety in the skeleton of 1,1-cyclobutanedicarboxylate
(Figure 1), but the IC50 values of carboplatin against the tested
cancer cell lines are much higher than those of DN604, indicat-
ing that the introduction of a carbonyl group in the leaving
ligand has an important influence on the anticancer activity of
DN604. In addition, for the normal human umbilical vein endo-
thelial cell line (HUVEC), DN603 and DN604 were nearly six and
three times less cytotoxic than cisplatin, respectively, suggest-
ing that, in contrast to cisplatin, DN603 and DN604 may be se-
lectively cytotoxic for human cancer cell lines. In general, both
DN603 and DN604 are worth studying further as anticancer
agents, especially the latter.
lescence was found at pH 7.4. The decrease in the rate of mi-
gration for closed circular DNA (Form I) was not clear at pH 7.4
under the test conditions. The results suggested that DN604
could effectively bind with DNA in an acidic cancerous
environment.
It is well known that the binding of platinum(II) complexes
and sulfur-containing molecules (e.g., glutathione, metallothio-
nie, and methionine) may be responsible for the side effects
and cross-resistance of the platinum complexes.[23,24] However,
Pt-sulfur (thiother) adducts were postulated as a drug reservoir
for platinum complexes.[25–27] Moreover, according to previous
study, the interaction rate of cis- or trans-platinum(II) complex
and DNA can be accelerated by l-methionine.[28,29] Thus, l-me-
thionine was chosen as nucleophile to react with DN603, and
Pt(dach)(CBDCA) was chosen as a control.
The study was carried out at 37.58C under pseudo-first-
order conditions. Hepes buffer was used to control the pH at
7.4. The UV/Vis spectra recorded for the reaction of DN603 and
Pt(dach)(CBDCA) with l-methionine, respectively, are shown in
Figure S2 and Figure S3, respectively. For DN603, no isosbestic
point was immediately apparent; however, closer examination
of the spectral changes revealed an isosbestic point at 256 nm
in the first 4 min of the reaction, which gradually disappeared
during the reaction (Figure 5). In addition, a second isosbestic
It is known that the pH value inside tumor cells tends to be
lower than that in normal cells due to increased production of
lactic acid in some tumor cells,[21,22] and the binding rate of
DNA with the platinum complexes containing dicarboxylato li-
gands could be greatly accelerated in an acidic environment.[22]
Therefore, the DNA-binding ability of DN604 was investigated
by gel electrophoresis at different pH levels (Figure 4). A de-
crease in the rate of migration for closed circular DNA (Form I)
was found at both pH 5.0 and 6.5, which demonstrated the un-
winding of the closed circular DNA. In addition, a coalescence
of the closed circular DNA (Form I) and open circular DNA
(Form II) was also observed at pH 5.0 and 6.5, whereas no coa-
Figure 5. Time-dependent UV/Vis spectra recorded in the first 4 min of the
reaction of DN603 (0.5 mm) with l-Met (50.0 mm) at 37.58C, pH 7.4 (Hepes
0.25m).
point was observed at 283 nm after 10 min of reaction, sug-
gesting that a second reaction occurred (Figure 6, Scheme 2).
The kinetic traces observed at 265 nm fitted well to
a double exponential function (Figure 7 and Figure S4), and
kobs1 and kobs2 were obtained (Tables S3 and S4). In the case of
kobs1, a linear dependence on the concentration of l-Met was
observed for both DN603 and Pt(dach)(CBDCA), whereas kobs2
was found to be independent of the l-Met concentration,
which was reported by Bugarcic and van Eldik previously,[30,31]
who demonstrated a chelation process (Figure 8). The results
Figure 4. Gel electrophoretic mobility pattern of pET28a plasmid DNA incu-
bated with various concentrations of DN604 in buffer (Tris-H3PO4) at differ-
ent pH levels. Lanes 1–8 (0, 1, 5, 10, 20, 40, 80, 160 mm,) + DNA. a) pH 5.0;
b) pH 6.5; c) pH 7.4.
Chem. Eur. J. 2014, 20, 15216 – 15225
www.chemeurj.org
15219
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
Table 2. Summary of the rate constants for the reactions between
DN603, Pt(dach)(CBDCA) and l-Met, respectively, at 37.58C.
[a]
[a]
Compound
k1 [ꢂ102 MÀsÀ1
]
k2 [ꢂ104 sÀ1
]
DN603
Pt(dach)(CBDCA)
2.94
1.77
24.4
5.7
[a] Values represent the meanÆSD (<10% of the mean value) from three
separate experiments.
(0.00057 sÀ1) for the reaction between Pt(dach)(CBDCA) and l-
Met agree reasonably well with the values obtained by Bugar-
cic and van Eldik (0.0216mÀ1 sÀ1 and 0.00055 sÀ1, respective-
ly).[30] According to the values presented in Table 2, the k1
values increased by a factor of 1.7, and the second step was
accelerated by a factor of 4.4 on going from DN603 to Pt-
(dach)(CBDCA), indicating that the reaction rate of DN603
with l-Met was much faster than that of Pt(dach)(CBDCA).
The study revealed that the introduction of a carbonyl group
to 1,1-cyclobutanedicarboxylate
Figure 6. Time-dependent UV/Vis spectra recorded after 10 min for the reac-
tion of DN603 (0.5 mm) with l-Met (50.00 mm) at 37.58C, pH 7.4 (Hepes
0.25m).
could greatly improve the kinetic
reactivity of the resulting plati-
num(II) complexes with l-met.
Carboplatin, with a lower tox-
icity than cisplatin, is believed to
be kinetically inert in blood,
whereas chloride ions were con-
sidered to be one of the major
nucleophiles that were responsi-
Scheme 2. Proposed pathway of the reaction for DN603 with l-Met.
ble for the activity of carboplatin
in vivo.[32,33] To compare the re-
action rates of DN604 and carbo-
platin with nucleophiles, the interaction of chloride
ions with DN604 and carboplatin, respectively, was
studied by HPLC-MS in their aqueous solutions at dif-
ferent time-points (Figure 9). It was clear that DN604
transformed rapidly and generated a new peak with
a retention time of 5.9 min, which was assigned to
the monochloro derivative of DN604. A second new
peak that increased over the reaction time was ob-
served to appear at approximately 6.6 min for both
DN604 and carboplatin, which was ascribed to the
formation of cisplatin. Surprisingly, the results
showed that the reaction rate of DN604 with chloride
ions was much faster than that of carboplatin
(Figure 10). The amount of carboplatin was still more
than 50% after 37 h, whereas DN604 had a halflife of
approximately 4.4 h, indicating that the reaction rate
of DN604 with chloride ions was at least 8-fold faster
than that of carboplatin. Moreover, cisplatin forma-
Figure 7. Absorbance–time trace recorded for the reaction of DN603 (0.5 mm) with l-Met
(50.0 mm) at 37.58C, pH 7.4 (Hepes 0.25m).
tion rate from DN604 was also much faster than that
from carboplatin. For example, no cisplatin peak was
observed within 1.75 h in the presence of carbopla-
for the reactions with l-Met imply that kobs1 and kobs2 can be
expressed by Equation (1) (kobs1 =k1[l-Met] + kÀ1) and Equa-
tion (2) (kobs2 =k2). As shown in Table 2, our rate constants for
tin, whereas a small amount of cisplatin was generated after
0.5 h incubation in the presence of DN604. The study revealed
that the reaction rate of DN604 with chloride anions was much
faster than carboplatin, demonstrating that DN604 can over-
the first step (0.0177mÀ1 sÀ1
)
and for the second step
Chem. Eur. J. 2014, 20, 15216 – 15225
www.chemeurj.org
15220
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
DNA was much faster than that of carboplatin, which can be
attributed to the change of kinetic properties due to modifica-
tion of the leaving ligand. Thus, it is reasonable to conclude
that the introduction of a carbonyl group to the leaving ligand
of carboplatin could greatly promote its DNA-binding rate to
the level of cisplatin, leading to high cytotoxicity.
To study the toxicity of the compounds in vivo, acute toxici-
ty of DN603 and DN604 was first tested by establishing the
range of doses of the compound that caused 0 and 100%
death rates for mice. The LD50 values of DN603 and DN604
were 100.3 and 150.1 mgkgÀ1, respectively, which are remarka-
bly greater than the reported value for cisplatin.[34,35] Particular-
ly, the acute toxicity of DN604 is even less than that of carbo-
platin (Table 3). Further studies on the acute toxicity of DN604
Table 3. Acute toxicity of DN604, carboplatin, and cisplatin in mice and
rats.
Drug
LD50 [mgkgÀ1] in ICR mice
LD50 [mgkgÀ1] in rat
DN603
DN604
carboplatin
cisplatin
100.3
150.1
139.0[b]
13.0[c]
n.d.[a]
79.4
71.5
6.4
[a] n.d.=not determined. [b] Data cited from Reference [34]. [c] Data
cited from Reference [35].
together with carboplatin and cisplatin as contrasts were also
conducted with rats. Based on a comparison of LD50 values,
DN604 with the functionalized dicarboxylate was almost at the
same level as carboplatin, whereas cisplatin was much more
toxic than DN604 and carboplatin, as expected.
Figure 8. Plots of kobs versus l-Met concentration for a) DN603 and b) Pt-
(dach)(CBDCA). T=37.58C, pH 7.2.
Conclusion
come the kinetic inertness of carboplatin and present high re-
activity. Moreover, the results further confirmed that the intro-
duction of a carbonyl group to the 1,1-cyclobutanedicarboxy-
late greatly improves the kinetic reactivity of the resulting plat-
inum(II) complexes with nucleophiles.
In the present investigation, two novel platinum(II) complexes,
DN603 and DN604, were designed and synthesized based
upon the concept that carboxylates with functional species
can be used to fulfill the function of the leaving ligand for the
resulting platinum complex and impart significant effects on
the anticancer activity by adjusting the balance between the
lipophilicity and hydrophilicity of the compound. The introduc-
tion of a carbonyl group to the skeleton of 1,1-cyclobutanedi-
carboxylate can not only change the physical properties of the
resulting platinum(II) complexes, but may also greatly enhance
their kinetic reactivity as well as DNA-binding rate, leading to
high anticancer efficacy. Chemical and biological results re-
vealed that DN603 and DN604 are promising anticancer candi-
dates; the latter, especially, has acceptable water solubility,
good stability in aqueous solution, potent anticancer activity in
vitro that is comparable to that of cisplatin, and almost the
same low acute toxicity as carboplatin. It would be of much in-
terest to study further the role of the functionalized group in
DN604, which becomes uncoordinated to the metal atom and
has a significant impact on the biological features of the
compound.
With an aim of comparing the DNA-binding rate of carbopla-
tin with that of DN604, the interaction of plasimid DNA with
DN604 and carboplatin, respectively, was investigated under
the same conditions by gel electrophoresis at different time
points. For DN604, the supercoiled DNA (Form I) can be con-
verted into the open circular DNA (Form II) over increasing
time, however, Form I DNA is hardly unwound at all under the
conditions tested for carboplatin within the time-range tested
(Figure 11).
To observe the unwinding of the supercoiled DNA induced
by carboplatin, high concentration (1280 mm) of the platinu-
m(II) compounds was applied. A coalescence of the open circu-
lar DNA (Form II) and the closed circular DNA (Form I) was only
observed after 196 h with carboplatin, but the same phenom-
enon took place within 24 h with DN604 (Figure 12). Together,
the above results proved that the binding rate of DN604 with
Chem. Eur. J. 2014, 20, 15216 – 15225
www.chemeurj.org
15221
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
42.0 mmol)[36,37] in dichlorome-
thane (500 mL) containing 100–
200 mesh silica gel (20.0 g). The
mixture was stirred overnight at
RT, then separated by column
chromatography (PE/EtOAc, 8:1).
Yield: 6.1 g (67%); colorless oil;
1H NMR (CDCl3/TMS): d=4.30–4.26
(q, 3J(H,H)=7.0 Hz, 4H; COOCH2),
3.63 (s, 4H; CH2), 1.32–1.29 ppm (t,
3J(H,H)=7.0 Hz, 6H; CH3).
Synthesis of 3-oxocyclobutane-
1,1-dicarboxylic acid: Diethyl 3-
oxocyclobutane-1,1-dicarboxylate
(6.1 g, 28.3 mmol) and KOH (3.5 g,
62.2 mmol) were added to water
(100 mL), and the resulting solu-
tion was stirred overnight at RT.
The solution was then extracted
with ethyl acetate (3ꢂ50 mL), the
aqueous phase was adjusted to
pH 2–3 with hydrochloric acid so-
lution (1 M), and the aqueous solu-
tion was extracted with ethyl ace-
tate (3ꢂ100 mL). The organic
phase was dried over anhydrous
Na2SO4 and concentrated under
vacuum to afford 3-oxocyclobu-
tane-1,1-dicarboxylic
acid
as
a
white precipitate. Yield: 2.4 g
(54%). 1H NMR (D2O/TMS): d=
3.70 ppm (s, 4H; CH2); MS (ESI): m/
z: 157 [MÀH]À, 113 [MÀCOOH]À.
Standard procedure for the prep-
aration
of
Pt(dach)(CBDCA),
DN603, and DN604: A mixture of
or cis-[Pt(NH3)2I2][39]
[38]
Pt(dach)Cl2
(5 mmol) and silver nitrate (1.7 g,
10 mmol) in distilled water
(150 mL) was first stirred for 24 h
at 508C. Then the mixture was fil-
tered and an aqueous solution
(20 mL) containing 1,1-cyclobuta-
nedicarboxylic acid or 3-oxocyclo-
Figure 9. HPLC profile of carboplatin (5 mm) and DN604 (5 mm) incubated with NaCl (1m) at different time.
butane-1,1-dicarboxylic
(5 mmol) and NaOH
10 mmol) was added to the filtrate. The mixed solution was stirred
for 24 h at 458C, then cooled to RT. The solution was concentrated
under vacuum to 5 mL, which led to the deposition of white
solids, which were collected and recrystallized from water.
acid
(0.4 g,
Experimental Section
Materials and instrumentations
All reagents and solvents were of analytical purity and were used
without any further purification. Potassium tetrachloroplatinate(II)
was purchased from a local chemical company (Shandong Boyuan
Chemical Co., Ltd., China). IR spectra were measured on KBr pellets
with a Nicolet IR200 FTIR spectrometer in the range of 4000–
Pt(dach)(CBDCA): Yield: 1.1 g (49%). IR (KBr): n˜ =3343, 3196, 3086,
;
2945, 1673, 1633, 1348, 1241, 1111, 1070, 911, 780 cmÀ1 1H NMR
([D6]DMSO/TMS): d=0.96–1.03 (m, 2H; CH2 of DACH), 1.19–1.22
(m, 2H; CH2 of DACH), 1.44–1.46 (m, 2H; CH2 of DACH), 1.59–1.70
(m, 2H; CH2 of cyclobutyl), 1.78–1.82 (m, 2H; CH2 of DACH), 2.03
400 cmÀ1 1H NMR spectra were recorded with a Bruker AV300
.
spectrometer, and 195Pt NMR spectra were recorded with a Bruker
DRX500 spectrometer. HPLC-MS analysis was performed with an
Agilent Accurate Mass 6224 TOF LC/MS system.
3
(s, 2H; CH2 of DACH), 2.65–2.70 (t, J(H,H)=7.5 Hz, 4H; CH2 of cy-
clobutal), 5.13 (m, 2H; NH2), 5.84–5.88 ppm (m, 2H; NH2); MS (ESI):
m/z (%):=452 (30) [M+H]+, 474 (100) [M+Na]+; elemental analysis
calcd (%) for C12H20N2O4Pt: C 31.93, H 4.47, N 6.21, Pt 43.22; found
C 31.98, H 4.42, N 6.19, Pt 43.25.
Synthesis and characterization
Synthesis of diethyl 3-oxocyclobutane-1,1-dicarboxylate: Pyridi-
nium chlorochromate (19.9 g, 92.4 mmol) was added to a solution
DN603: Yield: 1.2 g (51%). IR (KBr): n˜ =3497, 3430, 3237, 2944,
;
1785, 1640, 1607, 1373, 1174, 1068, 881, 855 cmÀ1 1H NMR
of
diethyl
3-hydroxycyclobutane-1,1-dicarboxylate
www.chemeurj.org
(9.1 g,
([D6]DMSO/TMS): d=1.01–1.04 (m, 2H; CH2 of DACH), 1.19–1.23
Chem. Eur. J. 2014, 20, 15216 – 15225
15222
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
466(30) [M+H]+, 488(100) [M+Na]+; elemental analysis calcd (%)
for C12H18N2O5Pt: C 30.97, H 3.90, N 6.02, Pt 41.92; found C30.91, H
3.98, N 6.09, Pt 41.90.
DN604: Yield: 1.0 g (53%). IR (KBr): n˜ =3298, 2973, 2922, 1787,
1640, 1371, 1303, 1206, 910, 861, 835, 730, 456 cmÀ1 1H NMR
;
([D6]DMSO/TMS): d=3.72 (s, 4H; CH2 of cyclobutyl), 4.20 ppm (s,
6H; NH3); 195Pt NMR ([D6]DMSO/TMS): d=À1724.7 ppm; MS (ESI):
m/z (%): 384 (30) [MÀH]À, 367(100) [MÀNH3-H]À; elemental analysis
calcd (%) for C6H10N2O5Pt: C 18.71, H 2.62, N 7.27, Pt 50.64; found
C18.66, H 2.58, N 7.29, Pt 50.60.
X-ray crystallography
Diffraction intensities for a single-crystal of DN604 were collected
with a Bruker Apex CCD area-detector diffractometer (Mo Ka, l=
0.71073 ꢁ). Absorption corrections were applied by using multi-
scan propram SADABS.[40] The structure was solved by direct meth-
ods and refined with full-matrix least-squares techniques using the
SHELXTL program package.[41] Anisotropic thermal parameters
were applied to all the non-hydrogen atoms. The organic hydro-
gen atoms were generated geometrically. Crystal data as well as
details of data collection and refinements for the complex are sum-
marized in Table S1. Selected bond distances and bond angles are
listed in Table S2. CCDC-981098 contains the supplementary crys-
tallographic data for this paper. These data can be obtained free of
charge from The Cambridge Crystallographic Data Centre via
www.ccdc.cam.ac.uk/data_request/cif.
Figure 10. The reaction profile of carboplatin (5 mm) and DN604 (5 mm), re-
&
spectively, with excess NaCl at 378C. Carboplatin ( ), cisplatin (produced by
~
*
^
carboplatin) ( ); DN604 ( ), cisplatin (produced by DN604) ( ).
HPLC Studies
The stability of platinum agents was monitored by reversed-phase
HPLC with an Agilent 1260 Infinity LC chromatographic system,
using a Kromasil C18 column (5 mm, 4.6ꢂ250 mm). The mobile
phase was CH2CN/H2O (5:95), and the flow rate was 0.4 mLminÀ1
with detection at 210 nm. DN604 was diluted to 1 mm with puri-
fied water and incubated at 378C; data were collected at different
times by reversed-phase HPLC.
Figure 11. Gel electrophoretic mobility pattern of pET28a plasmid DNA incu-
bated with platinum(II) complexes (160 mm) at pH 7.4. Lanes 1–6: carboplatin
(0, 12, 24, 48, 72, 96 h) + DNA; Lanes 7–12: DN604 (0, 12, 24, 48, 72, 96 h) +
DNA.
Log Pow determination
The log Pow determination of DN604 was conducted by using the
shake flask method. Excess DN604 was dissolved in KH2PO4/NaOH
(0.1m, pH 7.4) buffer solution presaturated with n-octanol for 48 h
at 378C. The solution was filtered to remove undissolved platinu-
m(II) compound. The initial concentration of platinum content was
determined by RP-HPLC analysis. Subsequently, the platinum(II) so-
lution was added to an equal volume of n-octanol (presaturated
with KH2PO4/NaOH buffer for 48 h). The heterogeneous mixture
was shaken vigorously for 2 h before centrifuging for 15 min to
achieve phase separation. The final concentration of platinum con-
tent in the aqueous phase was again determined by RP-HPLC, and
the water–octanol partition coefficient was calculated. In addition,
all the aqueous solution was diluted 10 fold before HPLC study
due to the high concentration. The experiment was performed in
triplicate.
Figure 12. Gel electrophoretic mobility pattern of pET28a plasmid DNA incu-
bated with platinum(II) complexes (1280 mm) at pH 7.4. Lanes 1–11: 0, 12, 24,
36, 48, 72, 96, 120, 144, 168, 192 h. a) carboplatin +DNA; b) DN604 + DNA.
In vitro cytotoxicity studies
Cytotoxicity of the developed compounds against DU145 (human
prostate cancer cell line), K562 (human leukemia cell lines),
MGC803 (human gastric cancer cell line), NB4 (human leukemia
cell lines), SE1116 (human colorectal cancer cell line), SK-BR-
3(human breast carcinoma cell line), SMMC-7721(human hepatocel-
lular carcinoma cell line), BGC823 (human gastric cancer cell line),
(m, 2H; CH2 of DACH), 1.44–1.47 (m, 2H; CH2 of DACH), 1.80–1.84
(m, 2H; CH2 of DACH), 2.06 (s, 2H; CH2 of DACH), 3.66–3.76 (m,
4H; CH2 of cyclobutyl), 5.25 (m, 2H; NH2), 5.95–5.98 ppm (m, 2H;
NH2); 195Pt NMR ([D6]DMSO): d=À1934.6 ppm; MS (ESI): m/z (%):
Chem. Eur. J. 2014, 20, 15216 – 15225
www.chemeurj.org
15223
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
A549 (human non-small-cell lung cancer cell lines), MCF-7 (human
breast carcinoma cell line), A2780 (human ovarian carcinoma cell
line), and HUVEC (normal human umbilical vein endothelial cell)
was determined by means of the colorimetric assay MTT [3-(4,5-di-
methylthiazol-2-yl)-2,5-diphenyltetrazolium bromide]. The cells
were plated at 5000 cells per well in 96-well culture plates with
culture medium and were incubated for 24 h at 378C in a water-at-
mosphere (5% CO2). The compounds with desired concentrations
were obtained by dissolving in water and diluting with culture
medium. The diluted solutions of complexes were then treated
with the cells for 48 h at 378C in a 5% CO2 incubator. Subsequent-
ly, 10 mL of a freshly diluted MTT solution (5 mgmLÀ1 in PBS) were
added to each well and the plates were incubated for 4.5 h. The
media with MTT solution were removed with 100 mL of DMSO solu-
tion, then cell survival was evaluated by measuring the absorbance
at 540 nm with an automatic microplate ELISA reader. IC50 values
were calculated from the plot of cell viability (%) against com-
pound concentration (mM). All experiments were carried out in
triplicate.
evaluate the toxicity of the compounds in vivo. These mice or rats
were divided into five groups, respectively, administered at the
dose of 10, 30, 60, 90, and 120 mgkgÀ1. The drug was administered
intravenously once. After administration, all external morphological
and behavioral changes, numbers of dead and time to death, as
well as other toxic effects were recorded. According to the mortali-
ty of mice or rats observed, the LD50 values for each drug were cal-
culated by using SPSS software with the Bliss method.
Acknowledgements
We are grateful to the New Drug Creation Project of the Na-
tional Science and Technology Major Foundation of China
(Project 2015ZX09101032) and the National Natural Science
Foundation of China (Project 21272041) for financial aid to this
work. Zhao would like to thank the Scientific Research Innova-
tion Project for College Graduates of Jiangsu Province (Project
CXLX 0148) and the Scientific Research Foundation of Gradu-
ate School of Southeast University (Project YBJJ1240).
Kinetics study
The ligand-substitution reactions were studied by observing the
change in absorbance at suitable wavelengths under pseudo-first-
order conditions (at least 10-fold excess of l-Met over the platin-
um(II) complexes). A suitable wavelength for kinetic trace was de-
termined by recording spectra of the reaction mixture over the
wavelength range 200–600 nm, and 265 nm was selected to ob-
serve for all the kinetics measurement. All the reactions were stud-
ied at 37.5Æ0.18C and pH 7.4 (Hepes, 0.25m) in the presence of
0.1m NaClO4. The reactions were initiated by mixing an aqueous
solution (1 mL) of the platinum(II) complex with l-Met in water
(1 mL).
Keywords: antitumor agents · drug design · drug discovery ·
kinetics · platinum
[1] A. V. Klein, T. W. Hambley, Chem. Rev. 2009, 109, 4911–4920.
[2] N. J. Wheate, S. Walker, G. E. Craig, R. Oun, Dalton Trans. 2010, 39,
8113–8127.
[3] L. Kelland, Nat. Rev. Cancer 2007, 7, 573–584.
[4] M. Galanski, B. K. Keppler, Anticaner. Agents. Med. Chem. 2007, 7, 55–73.
[5] A. Casini, J. Reedijk, Chem. Sci. 2012, 3, 3135–3144.
[6] M. J. Cleare, J. D. Hoeschele, Platinum Metals Rev. 1973, 17, 2–13.
[7] M. J. Cleare, J. D. Hoeschele, Bioinorg. Chem. 1973, 2, 187–210.
[8] J. Zhao, S. Gou, F. Liu, Y. Sun, C. Gao, Inorg. Chem. 2013, 52, 8163–8170.
[9] Y. Sun, S. Gou, R. Yin, P. Jiang, Eur. J. Med. Chem. 2011, 46, 5146–5153.
[10] X. Liu, H. Shen, H. Zhu, K. Cui, S. Gou, Bioorg. Med. Chem. Lett. 2007, 17,
3831–3834.
[11] J. Zhao, S. Gou, Y. Sun, R. Yin, Z. Wang, Chem. Eur. J. 2012, 18, 14276–
14281.
[12] H. S. Oberoi, N. V. Nukolova, A. V. Kabanov, T. K. Bronich, Adv. Drug Deliv-
ery Rev. 2013, 65, 1667–1685.
Reaction with sodium chloride
DN604 and carboplatin were diluted with distilled water to a final
concentration of 5 mm, and sodium chloride (1m) solution was
also prepared with distilled water. DN604 or carboplatin with equal
volume of sodium chloride solution was mixed and incubated at
378C. The reaction was studied by HPLC-MS analysis at different
time points. The HPLC method was applied as above (HPLC study).
[13] J. Zhao, S. Gou, L. Fang, R. Yin, Z. Wang, Inorg. Chem. 2012, 51, 10317–
10324.
[14] R. Yin, S. Gou, X. Liu, L. Lou, J. Inorg. Biochem. 2011, 105, 1095–1101.
[15] L. Fang, S. Gou, J. Zhao, Y. Sun, L. Cheng, Eur. J. Med. Chem. 2013, 69,
842–847.
Gel electrophoresis study
DNA binding properties of DN604 and carboplatin were investigat-
ed by agarose gel electrophoresis and pET28a plasmid DNA
(50 ng/ml) was used as the target. Desired concentrations of the
platinum complexes were prepared by dilution of the compound
with Tris-H3PO4 (100 mm) buffer at different pH values. pET28a
DNA (5 mL) was added to each tube, and the mixtures of platinum
complexes and pET28a plasmid DNA were then incubated at 378C
for the required time. Subsequently, the agarose gel (made up to
1% w/v) was prepared with TA buffer (50 mm Tris-acetate, pH 7.4).
The mixtures with loading buffer (1 mL) were submitted to electro-
phoresis in agarose gel in TA buffer at 100 V for 90 min. Agarose
gels were then dyed with ethidium bromide (0.5 mgLÀ1) for
20 min. Bands were imaged by using a Molecular Imager (Bio-Rad,
USA) under UV light.
[16] T. Soldatovic, M. Shoukry, R. Puchta, Z. D. Bugarcic, R. van Eldik, Eur. J.
Inorg. Chem. 2009, 15, 2261–2270.
[17] U. Frey, J. D. Ranford, P. J. Sadler, Inorg. Chem. 1993, 32, 1333–1340.
[18] S. Gou, Patent of China, 2013, CN 102924528A.
[19] S. P. Oldfield, M. D. Hall, J. A. Platts, J. Med. Chem. 2007, 50, 5227–5237.
[20] F. M. Freimoser, C. A. Jakob, M. Aebi, U. Tuor, Appl. Environ. Microbiol.
1999, 65, 3727–3729.
[21] L. E. Gerweck, K. Seetharaman, Cancer Res. 1996, 56, 1194–1198.
[22] M. Stubbs, P. M. McSheehy, J. R. Griffiths, C. L. Bashford, Mol. Med. Today
2000, 6, 15–19.
[23] P. M. Takahara, C. A. Frederic, S. J. Lippard, J. Am. Chem. Soc. 1996, 118,
12309–12321.
[24] S. H. van Rijt, P. J. Sadler, Drug Discov. Today 2009, 14, 1089–1097.
[25] R. N. Bose, S. Moghaddas, E. L. Weaver, E. H. Cox, Inorg. Chem. 1995, 34,
5878–5881.
[26] A. J. Jansen, J. Brouwer, J. Reedijk, J. Inorg. Biochem. 2002, 89, 197–201.
[27] K. J. Barnham, M. I. Djuran, P. del Socorro-Murdoch, J. D. Ranford, P. J.
Sadler, Inorg. Chem. 1996, 35, 1065–1072.
Determination of acute toxicity (LD50)
[28] C. Li, Z. Li, E. Sletten, F. Arnesano, M. Losacco, G. Natile, Y. Liu, Angew.
Chem. Int. Ed. 2009, 48, 8497–8500; Angew. Chem. 2009, 121, 8649–
8652.
LD50 values (median lethal dose) were determined with ICR mice
and rats, for both sexes (six female and six male for each group) to
Chem. Eur. J. 2014, 20, 15216 – 15225
www.chemeurj.org
15224
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
[29] C. Li, R. Huang, Y. Ding, E. Sletten, F. Arnesano, M. Losacco, G. Natile, Y.
Liu, Inorg. Chem. 2011, 50, 8168–8176.
[37] G. Bernhardt, H. Brunner, N. Gruber, C. Lottner, S. K. Pushpan, T. Tsuno,
M. Zabel, Inorg. Chim. Acta 2004, 357, 4452–4466.
[30] N. Summa, T. Soldatovic, L. Dahleburg, Z. D. Bugarcic, R. van Eldik, J.
Biol. Inorg. Chem. 2007, 12, 461–475.
[38] B. A. Howell, R. Rashidianfar, J. R. Glass, B. J. Hutchinson, D. A. Johnson,
Inorg. Chim. Acta 1988, 142, 181–183.
[31] N. Summa, W. Schiess, R. Puchta, N. van E. Hommes, R. van Eldik, Inorg.
Chem. 2006, 45, 2948–2959.
[32] A. Ciancetta, C. Coletti, A. Marrone, N. Re, Theor. Chem. Acc. 2011, 129,
757–769.
[33] X. Wang, H. Li, X. Du, J. Harris, Z. Guo, H. Sun, Chem. Sci. 2012, 3, 3206–
3215.
[39] S. C. Dhara, Indian J. Chem. 1970, 8, 193–194.
[40] G. M. Sheldrick, software for empirical absorption correction, SADABS,
Version 2.05, University of Gçttingen, Germany, 2002.
[41] SHELXTL 6.10, Bruker Analytical Instrumentation, Madison, Wisconsin,
2000.
[34] M. Xie, W. Liu, L. Lou, X. Chen, Q. Ye, Y. Yao, Q. Chang, S. Hou, Inorg.
Chem. 2010, 49, 5792–5794.
Received: June 23, 2014
[35] J. Zou, P. Y. Dou, K. Wang, J. Inorg. Biochem. 1997, 65, 145–149.
[36] M. Avram, C. D. Nenitzescu, M. Maxim, Chem. Ber. 1957, 90, 1424–1432.
Published online on September 26, 2014
Chem. Eur. J. 2014, 20, 15216 – 15225
www.chemeurj.org
15225
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim