JOURNAL OF

Inorganic
Biochemistry
Journal of Inorganic Biochemistry 98 (2004) 402–412
www.elsevier.com/locate/jinorgbio

The hydrolysis of the anti-cancer ruthenium complex NAMI-A

affects its DNA binding and antimetastatic activity:
an NMR evaluation
Marina Bacac a,b,c, Anna C.G. Hotze a, Karlijn van der Schilden a, Jaap G. Haasnoot a,
Sabrina Pacor b, Enzo Alessio d, Gianni Sava b,c, Jan Reedijk a,*
a
Leiden Institute of Chemistry, Gorlaeus Laboratories, Leiden University, P.O. Box 9502, 2300 RA Leiden, The Netherlands
b
Department of Biomedical Sciences, University of Trieste, via L. Giorgieri 7, 34127 Trieste, Italy
c
Callerio Foundation Onlus, via A. Fleming 22-31, 34127 Trieste, Italy
d
Department of Chemical Sciences, University of Trieste, via L. Giorgieri 7, 34127 Trieste, Italy
Received 15 October 2003; received in revised form 19 November 2003; accepted 2 December 2003

Abstract

The coordination of the antimetastatic agent NAMI-A, [H2 im][trans-RuCl4 (dmso-S)(Him)], (Him ¼ imidazole; dmso ¼ di-
methyl sulfoxide), to the DNA model base 9-methyladenine (9-MeAde) was investigated in water. NMR spectroscopy was first
applied for the study of the molecular stability and hydrolysis of NAMI-A in aqueous solution over a range of pH (3.0–7.4) and
chloride ion concentrations (0–1 M) at 37.0 °C. In physiological conditions (phosphate buffer, pH 7.4) NAMI-A disappears from the
solution in 15 min due to chloride and dmso hydrolysis, leading to uncharacterised poly-oxo Ru species. Conversely, at lower pH
(3.0–6.0) and in water (pH
5.5), only a partial dmso hydrolysis occurs, slowly forming the [trans-RuCl4 (H2 O)(Him)] complex.
This latter species coordinates to 9-MeAde (via the N7 of 9-MeAde), forming the [trans-RuCl4 (9-MeAde)(Him)] complex. NAMI-
A and [trans-RuCl4 (H2 O)(Him)] give comparable intracellular ruthenium concentrations and accumulate in KB cells (human
mouth carcinoma) and accumulate these at the G2 /M phase, while poly-oxo Ru species do not, and their cell uptake is reduced to
50%. On the contrary, G2 /M arrest and protein content in the murine metastatic cell line metGM, are not influenced by NAMI-A
hydrolysis. Hydrolysed NAMI-A species apparently are easier taken up by the metGM cells, showing intracellular ruthenium
concentrations one order of magnitude greater than those of intact NAMI-A. Therefore, it is proposed that the selective anti-
metastatic activity of NAMI-A during in vivo experiments can be attributed to its hydrolysed species.
Ó 2003 Elsevier Inc. All rights reserved.

Keywords: Ruthenium; NAMI-A; Hydrolysis; Coordination to 9-methyladenine; Cell cycle arrest at G2 /M phase

1. Introduction antimetastatic activity of ruthenium–dmso complexes

The success of cisplatin and related platinum com-
plexes as anticancer agents stimulated the search for
other active transition-metal complexes in cancer che-
motherapy [1,2]. Among all ruthenium-based anticancer
agents, ruthenium–dmso complexes are believed to have
great potential because of their selectivity for solid tu-
mor metastases and low host toxicity [3]. The selective
*
Corresponding author. Tel.: +31-71-5274459; fax: +31-71-5274671.
E-mail address: reedijk@chem.leidenuniv.nl (J. Reedijk).
was observed in the early 1980s with cis and trans-
[RuCl2 (dmso)4 ], (dmso ¼ dimethyl sulfoxide), but the
most relevant improvement in the ruthenium anticancer
chemistry was reached with NAMI, Na[trans-
RuCl4 (dmso-S)(Him)] [4], and NAMI-A, [H2 Im][trans-
RuCl4 (dmso-S)(Him)] [5] (Fig. 1) the last one being the
only ruthenium complex that successfully finished a
Phase I clinical trial. Differently from cisplatin and
platinum related compounds, biological testing of ru-
thenium–dmso compounds pointed out that DNA is not
the only target responsible for their antimetastatic ac-
tivity [5–7]. NAMI-A selectively reduces the formation

0162-0134/$ - see front matter Ó 2003 Elsevier Inc. All rights reserved.
doi:10.1016/j.jinorgbio.2003.12.003
 M. Bacac et al. / Journal of Inorganic Biochemistry 98 (2004) 402–412
- H+
N3
SOMe2
4 2 H H
Cl
Cl N
Ru
NH
5 1 6
Cl Cl 5 N7
1N
N3
4 2
2 N 4 N9
3 speciation of NAMI-A in aqueous solution in a wide
NH CH3
5 1
Fig. 1. Chemical structure and numbering scheme of NAMI-A,
[H2 im][trans-RuCl4 (dmso-S)(Him)] (left) and 9-methyladenine (9-Me-
Ade) (right).
and growth of lung metastases of malignant tumors such
as MCa mammary carcinoma, Lewis lung carcinoma
(LLC) and TS/A adenocarcinoma [8]. In vitro studies
excluded cisplatin-like cytotoxicity of NAMI-A for
primary tumor cells and showed an effect on their cell
proliferation, which might be a consequence of the
transient arrest at G2 /M phase of cell cycle [5,8]. Besides
cell cycle arrest, NAMI-A treatment of metastatic cells
induced modifications of cell membrane morphology
towards a less aggressive phenotype [9], down-regulated
CD44 expression and reduced their invasion ability
(manuscript in preparation).
Despite extensive in vitro and in vivo studies, the
exact mechanism of action of NAMI-A has not been
elucidated as yet and further investigations are needed
to identify its real biological target(s). Central interest is
focused on two fronts: (i) the investigation of the spe-
ciation of NAMI-A and other ruthenium-based com-
pounds in biological media such as body fluids, culture
media (to identify the metabolites formed) and (ii) the
interaction of the ruthenium complexes with biological
targets such as proteins and DNA [1,10]. In this view,
NAMI (the corresponding sodium salt of NAMI-A) has
already been investigated for the interaction with DNA.
These studies pointed out some interesting differences in
comparison to cisplatin that might account for different
biological activity between these two compounds [11]. In
particular, NAMI appeared to modify the double helix
of DNA in a different manner than cisplatin and to
coordinate to DNA considerably faster than cisplatin
[11]. The faster binding of NAMI to DNA was attrib-
uted to its higher rate of hydrolysis in comparison to
cisplatin. In fact, NAMI is not inert in physiological
solution but easily undergoes hydrolytic transforma-
tions [4,12], known to be the first necessary steps for the
interaction of metal-based drugs with biological targets
[13].
Although NAMI has been investigated for its inter-
action with DNA, NAMI-A has not been studied as yet.
Moreover, it is still unclear whether species derived from
hydrolysis of NAMI-A still maintain a pharmacological
403
activity comparable to that of the parent compound, or
whether hydrolysis of ligands and subsequently differ-
ences in charge of the appearing species reduce the
ability to cross cell membranes and to target biological
molecules.
Therefore, the aim of the present research is to in-
8
vestigate the interaction of NAMI-A with DNA (using
9-methyladenine as model base Fig. 1(right)) and the
range of pH and chloride ion concentrations by means
of NMR spectroscopy. Second, the biological activity of
some hydrolysis species of NAMI-A was studied by
evaluating parameters correlated with cell viability such
as cell cycle, protein synthesis and ruthenium uptake.
2. Results
2.1. Effects of pH on hydrolysis of NAMI-A: [trans-
RuCl4 (dmso-S)(Him)] ! [mer-RuCl3 (dmso-S)(Him)-
(H2 O)]
Hydrolytic transformation of NAMI-A (5 mM) was
monitored by 1 H NMR spectroscopy at 37.0 °C in
phosphate buffer (0.05 M phosphate, 0.15 M NaCl, in
the range of pH from 3.0 to 7.4) and in water (D2 O,
pH
5.5). Fig. 2(a) shows time-dependent 1 H NMR
spectra of the hydrolysis of NAMI-A in phosphate
buffer (pH 7.4). Immediately after dissolution of NAMI-
A (Fig. 2(a), t0 ), six signals are identified: four broad
signals (due to ligands bound to paramagnetic Ru(III))
and two sharp signals (due to the H2 imþ counter ion),
assigned according to already published data for NAMI
[4]. In fact, the NMR of NAMI and NAMI-A are al-
most identical. The broad signal at ca. )15 ppm was
assigned to the dmso-S ligand of NAMI-A while signals
at )3.5, )5.6 and )7.8 ppm were assigned to H5 and
H2;4 , respectively, of the Him ligand. The sharp reso-
nances at 8.69 and 7.49 ppm (1:2 integration ratio),
correspond, respectively, to H2 and H4 ,5 of the imi-
dazolium counter ion (H2 imþ ). Acetone (1 lL/mL) was
added to the buffer as a reference for the integration, as
it gives a sharp signal at 2.06 ppm that does not overlap
with the resonances of NAMI-A or its hydrolysis
species.
As a result of the hydrolysis of one chloride at pH
7.4, the resonances of NAMI-A are replaced within few
minutes by other signals. In particular, the broad reso-
nance at ca. )10 ppm was attributed to dmso-S of the
first hydrolysis species of NAMI-A, [mer-RuCl3 (dmso-
S)(Him)(H2 O)], (NAMI-AH2 O ), while the resonance at
ca. 0.6 ppm was attributed to the H5 of the Him ligand
of the same species (Fig. 2(a), dotted arrows). The res-
onances of H2 and H4 of Him in NAMI-AH2 O are most
likely too broad to be detected, as previously reported
[12].
404 M. Bacac et al. / Journal of Inorganic Biochemistry 98 (2004) 402–412

(a)

acetone

free
dmso

Him H 5 NAMI-AH2O dmso NAMI-AH2O

t 45 min

t 15- 20 min
t 10 min

t0
ppm 5 0 -5 -10 -1 5
H 2, H 4,5
H 2im+ Him H 5 Him H 2,4 dmso NAMI-A

(b) (c)
pH 7.4
3 3 pH 3.0
NAMI-A NAMI-A
NAMI -AH2O
LN peak area (dmso)

2 dm so 2

1 1 dm so

0 0
25 50 20 40 60 80
-1 -1

-2 Time (min) -2 Time (min)

NAMI-AH2O
[mer-RuCl3(dmso-S)(Him)(H2O)] SOMe2 OH2
SOMe2 Cl Cl Cl Cl
Cl Ru Ru
OH2
Ru Cl Cl Cl Cl
SOMe2 Cl Cl OH2 Him Him
Cl Cl Cl
Him OH2 NAMI-A [trans-RuCl4(Him)(H2O)]-
Ru 45 min. Ru
15 min. OH2
Cl Cl Cl Cl dmso released after 60 min
Cl Cl pH 6.0 40%
Him Him
Ru pH 5.0 11%
NAMI-A Cl Cl pH 3.0 4%
D2O (pH 5.5) 9%
Him
POLYNUCLEAR
[trans-RuCl4(Him)(H2O)]- Species

Fig. 2. (a) 300 MHz 1 H NMR spectra recorded during the hydrolysis of NAMI-A (5 mM) in phosphate buffer (0.05 M phosphate, 0.15 M NaCl), pH
7.4, 37.0 °C. (b, c) Transformation pathways of anionic NAMI-A and hydrolysis species formed in phosphate buffer (b, left) pH 7.4 and (c, right)
pH < 6.0, at 37.0 °C, monitored through integration of the dmso resonances at )15 (NAMI-A), )10 (NAMI-AH2 O ), and 2.7 ppm (dmso released) in
the 1 H NMR spectra. Scheme below the graph (left): hydrolysis species formed at pH 7.4, graph (right) species formed at pH < 6.0 and % dmso
released in function of pH. Note. For graphical reasons only the cis isomer of the [RuCl3 (H2 O)2 (Him)] species is designated in the scheme left; we
have no experimental evidence of the geometry of this species, either cis or trans or a mixture of the two isomers.

Integration of the dmso-S resonances at )15 and )10 although it is not stable in time. In fact, it continues to
ppm was used to monitor the disappearance of NAMI- hydrolyse, with a probable dissociation of at least one
A and the formation and subsequent disappearance of further chloride, until it disappears completely from the
NAMI-AH2 O (Fig. 2(b)). The signals for NAMI-A and solution after ca. 45 min (Fig. 2(b) and the scheme below
released dmso follow smooth curves; the curves for the figure). No new resonances attributable to hydrolytic
NAMI-AH2 O are less accurate and not easy to fit. Nev- species derived from NAMI-AH2 O could be detected,
ertheless, it was possible to determine that NAMI-AH2 O very likely because of the formation of polynuclear
is the predominant species in solution after 15–20 min, species (we have no experimental evidence of the ge-
 M. Bacac et al. / Journal of Inorganic Biochemistry 98 (2004) 402–412 405

ometry of this species, either cis or trans or a mixture of 3

the two isomers). Gradual darkening of the solution in

LN peak area (dmso

this condition, from light orange to dark brown, indi- 2

cated formation of poly-oxo Ru species as the result of

NAMI-A)
advanced hydrolysis of NAMI-A, as previously reported 1

for NAMI [12]. The gradual increase in intensity of the

0
sharp resonance at 2.72 ppm (free dmso) indicated that 3 6 9 12 15 18 21 24 27
slow hydrolysis of dmso from NAMI-A (and from its
-1
hydrolysis products) occurs simultaneously to that of 0M
chloride ligands (Fig. 2(b)). However, integration of the (a) -2 1M
signals of the free ligands indicated that dmso hydrolysis Time (min) 0.5 M
is only partial and that Him hydrolysis is negligible.
3
The situation at pH < 6 is different. No detectable
chloride hydrolysis occurs and only dmso is partially

LN peak area (dmso

2
released, slowly forming the [trans-RuCl4 (H2 O)(Him)]

NAMI-AH2O )
complex (Fig. 2(c) and the scheme below). Integration of 1

the resonance signal of free dmso as a function of pH

values pointed out that the rate of dmso hydrolysis de-
creases by lowering the pH (scheme below the Fig. 2(c)).
This decrease contributes to the better stability of
NAMI-A in solution and elongates its half-life to several
hours (in comparison to t1=2 of 20 min at pH 7.4). The
behaviour of NAMI-A in water is similar to that ob-
tained at lower pH since the dissolution of the compound
(5 mM) results in pH 5.5 that decreases in time (to
4)
due to the acidity of the coordinated water molecule. As
for solutions with pH < 6, the two predominant species of
NAMI-A in water (besides the cation) are the original
molecule [trans-RuCl4 (dmso-S)(Him)] and [trans-
RuCl4 (H2 O)(Him)] during several hours.
2.2. Effects of Cl concentration on NAMI-A hydrolysis:
[trans-RuCl4 (dmso-S)(Him)] ! [mer-RuCl3 (dmso-S)
(Him)(H2 O)]
The effects of increasing chloride ion concentration (0–
1 M) on the hydrolysis of NAMI-A were investigated in
phosphate buffer (pH 7.4, 37.0 °C). Increase of [Cl ] in
solution slows down the first chloride hydrolysis from
NAMI-A (Fig. 3(a)) and also dmso hydrolysis (data not
shown), influencing the formation of NAMI-AH2 O . It
also interferes with the stability of NAMI-AH2 O in solu-
tion since it slows down chloride hydrolysis from this
species: when no NaCl is added to the buffer, NAMI-
AH2 O reaches maximum values already after 15 min and
disappears rather fast (after 40 min, Fig. 3(b)). At [NaCl]
1 M, NAMI-AH2 O reaches its maximum values only after
30 min, remains present in solution for longer time, and
finally disappears gradually after 60 min (Fig. 3(b)).
2.3. Coordination to 9-methyladenine: [trans-RuCl4 (dm-
so-S)(Him)] ! [trans-RuCl4 (H2 O)(Him)] + 9-Me-
Ade ! [trans-RuCl4 (9-MeAde)(Him)]
The coordination of NAMI-A to the DNA model
base 9-methyladenine (9-MeAde) was studied by NMR
0
10 20 30 40 50 60 1M
-1
0.5 M
-2
0M
(b) -3 Time (min)
Fig. 3. (a, b) Influence of chloride ion concentration on (top) the sta-
bility of 5 mM NAMI-A obtained by the integration of dmso-S res-
onance at )15 ppm, and on (lower graph) the formation and
disappearance of the [mer-RuCl3 (dmso-S)(Him)(H2 O)] species ob-
tained by integration of dmso-S resonance at )10 ppm. Reaction was
followed in phosphate buffer (0.05 M phosphate), pH 7.4, 37.0 °C.
Range of [Cl] : 0–1 M.
spectroscopy in situ, at 37.0 °C, in water. As reported in
the hydrolytic study (see above) only the dmso ligand
hydrolyses slowly from NAMI-A in water and forms the
[trans-RuCl4 (H2 O)(Him)] complex. Therefore, the po-
sition occupied by the labile water molecule is the only
available coordination site for a DNA model base under
these conditions. Fig. 4 shows some 1 H NMR spectra of
the reaction (during 7 h) between NAMI-A and 9-Me-
Ade (1:4). At t0 only the resonances of the uncoordi-
nated 9-MeAde at 8.18, 8.04 and 3.78 ppm (assigned by
NOE difference spectroscopy to H2 , H8 and CH3 , re-
spectively), and of unreacted NAMI-A were identified.
The paramagnetic properties of Ru(III) cause severe
shifting and broadening of the resonances of ligands
coordinated to the metal ion (as seen for dmso and Him
ligands of NAMI-A). Therefore, it is expected that,
upon coordination, the resonances of 9-MeAde are
shifted and broadened with respect to the free ligand.
Thus, the two new and relatively broad signals of equal
intensity that appear at 13.16 and 1.05 ppm after 1 h of
reaction were attributed to coordinated 9-MeAde in
[trans-RuCl4 (9-MeAde)(Him)] .
The signals for the H2 and H8 protons of coordinated
9-methyladenine have been identified by T1 measure-
ments using inversion recovery. Namely, nuclear relax-
ation is strongly affected by interaction with unpaired
406 M. Bacac et al. / Journal of Inorganic Biochemistry 98 (2004) 402–412
1 0
Fig. 4. 300 MHz H NMR spectra recorded during the coordination of
9-methyladenine (20 mM) to NAMI-A (5 mM) in D2 O, 37.0 °C. (t0 )
NAMI-A and 9-methyladenine immediately after addition; resonances
at 8.18, 8.04 and 3.78 ppm correspond to H2 , H8 and CH3 , respec-
tively, of free 9-methyladenine. (7 h) Broad and shifted signals at 13.16
and 1.05 ppm (arrows) are relative to H2 and H8 , respectively, of co-
ordinated 9-MeAde in [trans-RuCl4 (9-MeAde)(Him)] .
electrons. Therefore, protons close to the Ru(III) centre
have relatively short relaxation times. Hence, they can be
distinguished from the protons of the free ligand which
display large T1 values, i.e., in the order of hundreds of
milliseconds. The new signals appearing at 13.16 and 1.05
ppm exhibit T1 values of 27 and 13 ms, respectively, in-
dicating coordination of the ligand to Ru(III). A third
signal at 5.15 ppm, which probably belongs to the CH3
group of 9-methyladenine, has been identified with the T1
measurement. The signal is shifted almost under the
water signal but has been recognized by its short T1 value,
i.e., 19 ms. However, it was impossible to identify the
resonances of the Him ligand of the [trans-RuCl4 (9-Me-
Ade)Him] complex, which is reasonable considering
that only ca. 10% of NAMI-A coordinates to 9-MeAde.
This is in agreement with the fact that also the Him res-
onances in the [trans-RuCl4 (H2 O)Him] and [mer-
RuCl3 (dmso-S)(Him)(H2 O)] complexes could not be
identified (with the exception of H5 of the Him in the
[mer-RuCl3 (dmso-S)(Him)(H2 O)] complex) and might
overlap with those of NAMI-A.
The assignment of the H2 and H8 resonances of the
coordinated 9-MeAde in [trans-RuCl4 (9-Me-
Ade)(Him)] was performed by monitoring their pH-
dependence. Free 9-MeAde exhibits N7 protonation at
pH < 2, and N1 protonation at pH
4 [14]. Protonation
of N1 occurs at lower pH when a metal is bound at N7
[15–17]. The absence of N7 protonation for pH < 2
(no shifting of the resonances of [trans-RuCl4 (9-Me-
Ade)Him] ), and the protonation of N1 of bound 9-
MeAde at pH
3 (Fig. 5), clearly indicate that 9-MeAde
coordinates to ruthenium through its N7 atom and
consequently exclude N6 coordination. The signal at
13.16 ppm shows a larger pH-dependent shifting (Fig. 5)
and was thus assigned to H2 of coordinated 9-MeAde,
16
14 H2
12
10
chemical shift (ppm)
8
2 H8
-2
2 4 6 8 10
pH
Fig. 5. pH titration of [trans-RuCl4 (9-MeAde)(Him)] .
which is closest to the N1 protonation site. The resonance
at 1.05 ppm was therefore assigned to H8 of coordinated
9-MeAde. This assignment is in agreement with the T1
values of the H2 and H8 , since the proton closest to the
Ru(III) centre (which is the H8 ), displays the shortest T1
value in comparison to the more distant H2 (13 vs. 27 ms).
Therefore, the coordination mode of 9-MeAde to
NAMI-A is different from the reported N6 coordination
observed in a structurally related complex, i.e. [trans-
RuCl3 (9-MeAde)(dmpt)2 ]  H2 O, (dmpt ¼ 5,7-dimeth-
yl[1,2,4]triazolo[1,5-a]pyrimidine) [18]. The bidentate
N6–N7 coordination, found in [cis-Ru(bpy)2 (9-Me-
Ade)](PF6 )2 and [a-Ru(azpy)2 (9-MeAde)](PF6 )2 (bpy ¼
2,20 -bipyridine; azpy ¼ 2-phenylazopyridine) [19] was
also excluded, since the [trans-RuCl4 (H2 O)(Him)]
complex has only one available coordination site.
No variations of chemical shifts were seen for H8 and
H2 by increasing the pH to 7, while for pH > 7 the
complex started to decompose with gradual loss of
‘‘paramagnetic’’ signals and darkening of the solution.
ESI-MS (electrospray mass spectrometry) experiments,
performed on the reaction solution with the aim of
gaining further evidence on the nature of the product,
were unsuccessful and this technique proved to be in-
adequate for investigating these systems. The reaction
between NAMI-A and 9-MeAde was also run adding
the DNA model base to an aged solution of the complex
(3 days in 10 mM NaClO4 in D2 O at room tempera-
ture), and in the presence of 10% dmso. 9-MeAde co-
ordinated already after 30 min when the aged solution of
NAMI-A was used, in which ca. 40% of the dmso ligand
was hydrolysed. On the contrary, 9-MeAde did not co-
ordinate to NAMI-A when 10% of dmso was added to
the solution, as dmso hydrolysis from NAMI-A was
inhibited in agreement with previous studies [20]. These
observations unambiguously proved that the dmso site
in NAMI-A becomes the coordination site of 9-MeAde.
 M. Bacac et al. / Journal of Inorganic Biochemistry 98 (2004) 402–412
9-MeAde did not coordinate to NAMI-A in phos-
phate buffer at pH 7.4, where only its hydrolysis species
and free 9-MeAde were found after overnight reaction
together with black colour of solution due to poly-oxo
Ru species (data not shown). Besides the hydrolytic in-
stability of NAMI-A at pH 7.4, the absence of detect-
able coordination of the model base to ruthenium in
phosphate buffer is attributed to competition with OH
(high affinity ligand for Ru(III)) [21].
2.4. Effects of NAMI-A hydrolysis on cell proliferation
The effects of NAMI-A and of its hydrolysis species
on cell proliferation were evaluated on human ovarian
carcinoma cell line A2780 (normal) and A2780cisres
(resistant to cisplatin). Two different hydrolysis species
of NAMI-A were tested and compared to the original
molecule of NAMI-A; these were the poly-oxo Ru
species, obtained by complete hydrolysis of NAMI-A in
PBS, and a mixture of NAMI-A and [trans-
RuCl4 (H2 O)(Him)] (ca. 1.5:1), derived from hydrolysis
of NAMI-A in water (Section 5). Regarding this last
solution, it was impossible to further increase the
amount of [trans-RuCl4 (H2 O)(Him)] species, since
longer hydrolysis of NAMI-A in water led to the for-
mation of poly-oxo Ru species. For easier description,
NAMI-A fresh is indicated as (1), poly-oxo Ru species
as (2) and the mixture of NAMI-A and [trans-
RuCl4 (H2 O)(Him)] as (3). Cell growth was determined
by MTT proliferation assay 24 h after treatment with 1
mM ruthenium (originally as NAMI-A).
NAMI-A (1) and (3) show comparable reduction of
cell proliferation of A2780 cells in comparison to con-
trols (***p < 0:001, Table 1), and both species are more
active in comparison to NAMI-A (2) ($$$ p < 0:001,
Table 1). (Data are considered to be statistically signif-
icant if the probability factor ‘‘p’’ is *p < 0:05, **p <
0:01 and ***p < 0:001.)
Effects on A2780cisres were comparable to those on
the A2780 cell line (data not shown) suggesting that the
mechanisms responsible for cisplatin resistance do not
interfere with activity of NAMI-A nor with that of its
Table 1
Reduction of cell proliferation (MTT test, A2780 cells), G2 /M arrest and intracellular ruthenium content (KB and metGM cells), 24 h after treatment
with NAMI-A (0.1–1 mM)
Sample 1 mM 0.1 mM
MTT OD (590 nm) % G2 /M phase
A2780 KB
Control 2.57  0.06 13.4  1.00
(1) 1.79  0.07***;$$$ 20.8  1.20***;$$$
(2) 2.29  0.16* 14.8  1.0
(3) 1.72  0.15***;$$$ 27.0  1.0***;$$$
(1) NAMI-A immediately dissolved, (2) poly-oxo Ru species, (3) mixture of NAMI-A and [trans-RuCl4 (H2 O)(Him)] (1.5:1) (solutions were
prepared according to Section 5). Statistical analysis, Student–Newman–Keuls test, *p < 0:05; **p < 0:01; ***p < 0:001 vs. controls; $$$ p < 0:001
(1), (3) vs. (2); ^^ p < 0:01 (2) vs. (1); °°p < 0:01; °°°p < 0:001 (2) vs. (1) and (3).
407
hydrolysis species. Data regarding the effects of 0.1 mM
NAMI-A on cell proliferation of KB cell line have al-
ready been reported elsewhere and indicated a statisti-
cally significant reduction of cell proliferation in
comparison to controls [8,22]. The same study pointed
out a transient cell cycle arrest of KB cells at G2 /M phase.
Therefore, we used this cell line for further investigations
concerning the effects of NAMI-A and its hydrolysis
species on the cell cycle distribution of tumor cells.
2.5. Effects of NAMI-A hydrolysis on cell cycle distribu-
tion of KB cells
The same species mentioned above were evaluated at
the total Ru concentration of 0.1 mM. NAMI-A (1)
increased KB cells at G2 /M phase by 64% in comparison
to controls (***p < 0:001, Table 1), similarly to NAMI-
A (3) (***p < 0:001, Table 1). Only a weak G2 /M arrest
was observed with NAMI-A (2), (*p < 0:05, Table 1).
Statistically significant differences were observed not
only in comparison to controls but also within the tested
species, and consisted in differences between NAMI-A
(1), (3) and NAMI-A (2) and between NAMI-A (1) and
NAMI-A (3) ($$$ p < 0:001, Table 1). The first one in-
dicates the inability of poly-oxo Ru species to affect cell
cycle distribution of KB cells, the second one indicates
the important role of dmso hydrolysis in favouring G2 /
M arrest of same cells.
In agreement with the found G2 /M arrest, NAMI-A
(1) and (3) caused a statistically significant increase of the
protein content in comparison to controls; the average
fluorescence intensity (MFI) relative to the protein con-
tent of treated samples was found 206  2.1 for (1),
228  3.9 for (3) vs. 185  7.1 for control, p < 0:001,
whereas NAMI-A (2) did not influence this parameter
(MFI, 177  11.7 for (2) vs. 1857.1 for control, p > 0:05).
2.6. Effects of hydrolysis on cell uptake of ruthenium in
KB cells
Total ruthenium content inside KB cells, determined
by atomic absorption spectroscopy, indicated that
lg Ru/106 cells
metGM KB metGM
12.5  1.20 – –
15.7  0.30* 0.02  0.002 0.03  0.006
17.6  1.00**;^^ 0.01  0.005°° 0.34  0.027°°°
14.6  3.56 0.02  0.006 0.07  0.029
408 M. Bacac et al. / Journal of Inorganic Biochemistry 98 (2004) 402–412
NAMI-A (1) and (3) reached comparable values of in-
tracellular ruthenium concentration (ca. 0.02 lg Ru/
1  106 cells, Table 1). The intracellular ruthenium
concentration of NAMI-A (2) was reduced by about
50% (Table 1) indicating decreased uptake of poly-oxo
Ru species, derived from advanced hydrolysis of
NAMI-A.
2.7. Effects of hydrolysis on cell cycle distribution of
metGM cells
Similarly to KB cells, metGM cells were treated with
solutions (1), (2) and (3), at the total Ru concentration
of 0.1 mM. For a better evaluation, these cells were
treated for three consecutive times with the above
mentioned solutions of NAMI-A and analysed for cell
cycle distribution, protein content and intracellular ru-
thenium concentration after each treatment.
Differently from KB cells, NAMI-A (1), (2) and (3)
gave comparable effects on cell cycle arrest of metGM
cells, 24 h after the first, second and third treatments.
The first treatment with NAMI-A (1), (2) and (3) re-
sulted in comparable S phase arrest (data not shown)
that shifted to G2 /M arrest after the second treatment
(Table 1), and was still present after the third treatment.
After the second treatment, a statistically significant
difference in G2 /M arrest was observed between NAMI-
A (1) and (2), with (2) being more active (^^ p < 0:01,
Table 1). It is interesting to point out that both NAMI-
A (1) and NAMI-A (2) significantly reduce protein
content in metGM cells after the first and the second
treatment (MFI, about 750 vs. 866  30 for the control
after first, and about 925 vs. 1070  30 after the second
treatment, p < 0:05).
2.8. Effects of hydrolysis on cell uptake of ruthenium in
metGM cells
Significant differences were also observed in the in-
tracellular ruthenium content between NAMI-A and its
hydrolysis species. It appears that cellular uptake of
ruthenium by metGM cells is favoured by hydrolysis of
NAMI-A, since (2) showed increased uptake already
after the first treatment (Table 1). As shown, the intra-
cellular ruthenium concentration of (2) is about 10-fold
higher than that of (1) and about fivefold higher than
that of (3) (lg Ru/1  106 cells: 0.03 for (1), 0.34 for (2),
0.07 for (3)). The same pattern of intracellular ruthe-
nium concentrations was obtained after the second and
the third treatment (data not shown).
3. Discussion
To be able to understand the coordination of NAMI-
A to biological targets, it was first considered necessary
to elucidate the stability and hydrolytic transformations
of NAMI-A in aqueous solution, since a good under-
standing of these processes is the crucial step to identify
the active species and to unravel their mode of interac-
tion with DNA and other biological targets [13].
Hydrolysis of NAMI-A was studied by NMR spec-
troscopy in a wide range of pH (3.0–7.4) and of chloride-
ion concentrations (0–1 M) to reproduce the conditions
that NAMI-A might encounter in vivo. Blood plasma
presents a pH of 7.4, while the hypoxic environment of
solid tumors offers pH values between 6.0 and 7.0. Some
cell compartments, like endosomes that handle trans-
ferrin, possess even lower pH values, i.e.
5.5, and it is
also possible to reach extreme conditions like pH 2.0 in
the stomach, or pH 8.0 in urine. Studies were performed
at 37.0 °C, i.e., a temperature at which hydrolysis and
binding to target biomolecules occur in host organisms.
In physiological conditions (phosphate buffer, pH
7.4), NAMI-A undergoes both chloride and dmso hy-
drolysis while forming a mixture of hydrolysis species
responsible for the disappearance of the parent com-
pound from the solution in about 15 min. Increased
chloride-ion concentration delays the first hydrolysis
step of NAMI-A and improves the stability of [mer-
RuCl3 (dmso-S)(Him)(H2 O)] (NAMI-AH2 O ) in solution.
This behaviour, indeed, suggests that hydrolytic de-
composition of NAMI-A in body fluids might be altered
by chloride ion concentration, similarly to what happens
for cisplatin.
At pH < 6.0 the chemical metabolism of NAMI-A
changes, and the complex becomes much more inert.
Chloride hydrolysis becomes negligible and slow dmso
hydrolysis generates the reactive aquated species, [trans-
RuCl4 (H2 O)(Him)] . Treatment of NAMI-A with ex-
cess 9-methyladenine in water (pH 5.5) resulted in the
partial formation of the [trans-RuCl4 (9-MeAde)(Him)]
complex, whose resonances appeared clearly in the 1 H
NMR spectrum after 2 h of reaction. NMR pH titration
experiments indicated that 9-MeAde coordinates to ru-
thenium through its N7 atom.
NAMI-A was not found to coordinate significantly to
9-MeAde in phosphate buffer (pH 7.4), where it com-
petes with OH , a high affinity ligand for Ru(III), and
undergoes both chloride and dmso hydrolysis forming a
mixture of unidentified poly-oxo Ru species. However, it
cannot be excluded that these species might also coor-
dinate to 9-MeAde, but in this case their signals are too
broad and too weak to be detected by 1 H NMR spec-
troscopy. In support of this hypothesis is the fact that
NAMI, the corresponding Naþ salt of NAMI-A, forms
intrastrand bifunctional adducts on polymeric DNA
[11]. Although with different techniques, this study
pointed out that NAMI binds both guanine and adenine
(with guanine being its preferential binding site), and in
this view it behaves similarly to NAMI-A, which is
reasonable since the cation (Naþ or H2 imþ ) does not
 M. Bacac et al. / Journal of Inorganic Biochemistry 98 (2004) 402–412
influence the behaviour of the anion (at least not in
aqueous solution). Although not presented herewith, we
also observed that NAMI-A coordinates to 9-ethyl-
guanine with the same efficacy as to 9-methyladenine.
The biological activity of the two main hydrolysis
species (i.e., the poly-oxo Ru species and [trans-
RuCl4 (H2 O)(Him)] ) was investigated, and compared to
that of the parent compound NAMI-A. Hydrolysis affects
the biological activity of NAMI-A in different ways in
human carcinoma cell lines and murine metastatic cells.
Stronger reduction of cell proliferation on the human
carcinoma cell line A2780 is achieved with solutions of
fresh NAMI-A and of [trans-RuCl4 (H2 O)(Him)] than
with a solution of poly-oxo Ru species. Surprisingly, al-
though the uptake in KB cells of the two former species
was found to be almost similar, [trans-RuCl4 (H2 O)
(Him)] was found more active than fresh NAMI-A in
arresting the cell cycle of KB cells in the G2 /M phase. In
contrast to the species 1 and 3, the poly-oxo Ru species
were found to be only slightly able to enter KB cells, to
arrest them at G2 /M phase and to affect protein synthesis.
Unlike KB cells, the cell cycle and protein synthesis in
metGM cells is not influenced by hydrolysis of NAMI-A,
since poly-oxo Ru species and NAMI-A gave comparable
effects. Contrary to KB cells, hydrolysis seems to favour
ruthenium uptake in metGM cells; in fact, the intracel-
lular ruthenium concentration increased up to ten times in
the case of poly-oxo Ru species in comparison to NAMI-
A and up to thirty times compared to the uptake of poly-
oxo Ru species in KB cells.
It is to be noted that in the solid tumor cell lines (KB
and A2780) reduction of proliferation occurs in the
presence of NAMI-A and [trans-RuCl4 (H2 O)(Him)] ,
and that these species do have the ability to bind to 9-
MeAde (and probably also to DNA). In metGM cell, on
the other hand, the poly-oxo Ru species are taken up
easily, but these species do not bind to 9-MeAde.
4. Conclusions
Even though DNA is not expected to be the major
biological target for the antimetastatic activity of
NAMI-A, the present study indicates that it is likely that
this compound interacts with DNA, once it enters tumor
cells and reaches the nuclei. The molecular stability re-
quested for the binding of NAMI-A to DNA can be
achieved at relatively low pH, such as in a tumor cell. In
these conditions chloride hydrolysis may be slowed
down and the dmso ligand can be hydrolysed from
NAMI-A resulting in the reactive aqua-complex [trans-
RuCl4 (H2 O)(Him)] , that subsequently coordinates to
DNA. These findings differ from reports on the inter-
action of NAMI-A with albumin, achieved after chlo-
ride hydrolysis [23], suggesting that interactions of
NAMI-A with biologically relevant molecules might
409
involve different hydrolysis species formed through dif-
ferent hydrolysis pathways. While DNA binding re-
quires a rather intact species, binding to albumin might
well occur after chloride hydrolysis.
The different biological activity found for NAMI-A
and its hydrolysis species between KB and metGM cells
might account for the selective antimetastatic activity of
NAMI-A. In the solid tumor KB cell line, the poly-oxo
species have lost activity, whereas these species do arrest
metastatic cells at G2 /M phase and reduce their protein
synthesis. Interestingly, the poly-oxo species show a
larger potential to enter metastatic metGM cells and
reach a significantly higher intracellular ruthenium
concentration than fresh NAMI-A. Interestingly, the Ru
uptake for the poly-oxo species is about thirty times
more for the metastatic cells compared to the solid tu-
mor cells, which suggests the importance of this species
for the antimetastatic activity. Therefore, it is proposed
that the selective antimetastatic activity of NAMI-A
during in vivo experiments might be, indeed, attribut-
able to these poly-oxo Ru species, i.e., species probably
more abundant after days of treatment.
5. Experimental
5.1. Materials
NAMI-A, [H2 im][trans-RuCl4 (dmso-S)(Him)] was
prepared according to earlier published experimental
section [5]. 9-Methyladenine was synthesised by methyl-
ation of adenine with methyl iodide, as previously de-
scribed [24].
5.2. Hydrolysis experiments
Hydrolysis was performed in phosphate buffer (0.05
M phosphate, 0.15 M NaCl) at pH 3.0, 5.0, 6.0, 6.5, 7.0,
7.4 and in water (D2 O). Reactions were followed di-
rectly in 5 mm NMR tubes at 37.0 °C, on freshly pre-
pared 5 mM solutions of NAMI-A (2.3 mg/mL
phosphate buffer). Spectra were recorded automatically
every 2 min (pH 7.4, 7.0), 5 min (pH 6.5, 6.0), 10 min
(pH 5.0, 3.0) and 30 min (D2 O) on a Bruker 300 DPX
spectrometer at 300.13 MHz. pH was measured by
laboratory pH Meter (PHM220, MeterLabÒ, Radiom-
eter Analytical S.A), no correction for the pD values was
performed. All experiments were performed at least
twice. Acetone (1 lL/mL) was added to the buffer and
used as the reference signal for the integration, as it gives
a sharp signal (2.06 ppm) that does not overlap with the
resonances of NAMI-A or its hydrolysis species. 1 H
NMR spectra were calibrated on the acetone peak set at
2.06 ppm. Normal 1 H NMR spectra were obtained us-
ing a spectral window of 100 ppm and an acquisition
time of 1.0 s.
410 M. Bacac et al. / Journal of Inorganic Biochemistry 98 (2004) 402–412
5.3. Effect of Cl concentration
Influence of chloride ion concentration on the hy-
drolysis of NAMI-A was studied in phosphate buffer
(0.05 M phosphate, pH 7.4) additioned with NaCl to the
desired concentration of Cl (0–1 M). Hydrolysis was
followed directly in 5 mm NMR tubes on freshly pre-
pared 5 mM solutions of NAMI-A, at 37.0 °C.
5.4. Coordination to DNA model base
Coordination of 9-methyladenine to NAMI-A was
performed in three different solutions; (a) in water
(D2 O), (b) in D2 O containing NaClO4 (10 mM); and (c)
in phosphate buffer (0.05 M phosphate, 0.15 M NaCl,
pH 7.4). In (a) and (c) NAMI-A was freshly dissolved in
solution (5 mM) and a fourfold molar ratio of 9-
methyladenine was added directly to the 5 mm NMR
tube. In (b) NAMI-A was aged in the solution for at
least 3 days at room temperature prior to reaction with
9-methyladenine according to [11]. The influence of ex-
cess dmso on the coordination of 9-MeAde to NAMI-A
was investigated in a 5 mM solution of NAMI-A (D2 O)
added with 20 mM 9-MeAde and 10% dmso-d6 . In all
cases, reaction was followed in situ in 5 mm NMR tubes,
at 37.0 °C, recording spectra every 30 min. During re-
action in D2 O, the pH spontaneously decreased from 5.5
to ca. 3.5.
5.5. Inversion recovery
1
H NMR were recorded using the inversion recovery
pulse sequence with the following typical values: spectral
width of 100 ppm, acquisition time of 200 ms, delay time
(tau) in the range between 10 and 300 ms. The delay
between consecutive scans was generally 100 ms. Anal-
yses were performed on a Bruker 300 DPX spectrometer
at 300.13 MHz.
5.6. pH titration
pH titration was performed at 25.0 °C on a 5 mM
D2 O solution of NAMI-A added with 20 mM 9-MeAde
in a 5 mm NMR tube, adjusting the pH by DCl (0.1 N)
and NaOD (0.1 N). Direct pH-meter readings were used
to estimate the pKa values (no correction for the pD
values was performed).
5.7. ESI-MS
Mass spectra analyses were performed at the Gorla-
eus Laboratories of Leiden University on a Finningan
MAT 900 instrument, equipped with an electrospray
interface (ESI) and at the API I ion-spray mass spec-
trometer (Applied Biosystems SCIEX), Department of
Biochemistry, Biophysics and Macromolecular Chem-
istry, University of Trieste, Italy.
5.8. Cell cultures
A2780 (human ovarian carcinoma) and A2780 cis-
platin-resistant cell lines were maintained in DulbeccoÕs
modified EagleÕs Medium (DMEM) (Gibco BRLe, In-
vitrogen Corporation, NL) supplemented with 10% fetal
calf serum, 1% penicillinG sodium (100 units/mL), 1%
streptomycin (100 lg/mL), and 1% Glutamax 100 (all
from Duchefa Biochemie BV, Netherlands, NL). KB
cell line (ECACC No. 86103004) was cultured in EagleÕs
medium added with 10% FCS, 1% non-essential amino
acids 100, 1% sodium pyruvate 100 mM, 1% L -gluta-
mine, 1% penicillin/streptomycin 100 buffered with 5%
NaHCO3 (20 lM) (all from EuroCloneÒ, Paington,
UK). Cells were harvested from confluent monolayers
by 0.05% trypsin solution. metGM cell line (for refer-
ence, see [10]) was maintained in RPMI 1640 medium
(Sigma Chemical) added with 10% FBS, 1% L -glutamine
100, 1% penicillin/streptomycin 100, 1% non-essen-
tial amino acids 100, 0.5% gentamycin 100, 0.37%
sodium pyruvate 100 mM (all from EuroCloneÒ).
5.9. In vitro treatment with NAMI-A
KB and metGM cells were seeded in 6-wells plates
(0.5  106 cells/4 mL/well, Corning CostarÒ, New York,
USA) and left to adhere for 48 h at 37.0 °C, 5% CO2 . In
the case of metGM cells, plastic wells were pre-coated
with MatrigelÒ (30 lg/100 lL) to allow adhesion of cells
that usually grow in suspension. Two main hydrolysis
species of NAMI-A were tested and compared to fresh
NAMI-A (1): poly-oxo Ru species, formed by advanced
hydrolysis of NAMI-A in PBS (2) and [trans-
RuCl4 (H2 O)(Him)] , formed by hydrolysis of NAMI-A
in water (3). Solution (1) was prepared by dissolving
NAMI-A immediately before treatment, while solutions
(2) and (3) were prepared by hydrolysing NAMI-A in
PBS and in water, respectively, for 4 h at 37.0 °C. In
these conditions, at t0 the cells received the solution (1)
containing 100% of NAMI-A, the solution (2) contain-
ing 100% of poly-oxo Ru species (also confirmed by the
black colour of the solution), and the solution (3)
containing a mixture of NAMI-A and [trans-
RuCl4 (H2 O)(Him)] species (ca. 1.5:1). Regarding the
solution (3), it was impossible to further increase the
amount of [trans-RuCl4 (H2 O)(Him)] species in the so-
lution, since longer hydrolysis in water led to the for-
mation of poly-oxo Ru species. All solutions were
sterilised before treatment by filtration (filters 0.22 lm,
Sartorius Minisart AG, G€ ottingen, Germany) and tested
at the final concentration of 0.1 mM. After incubation
 M. Bacac et al. / Journal of Inorganic Biochemistry 98 (2004) 402–412
with NAMI-A (1 h at 37.0 °C in PBS, pH 7.4), the cells
were washed twice with PBS and incubated for further
24 h in complete medium, before analyses. In the case of
repeated treatments, the same procedure was performed
for three consecutive times. (PBS, phosphate-buffered
saline; 8 g NaCl, 0.2 g KCl, Na2 HPO4 1.15 g, 0.2 g
KH2 PO4 for 1 L, pH 7.4).
5.10. MTT colorimetric assay
Cell proliferation was evaluated by MTT colorimetric
assay [25] based on the reduction of the tetrazolium salt,
MTT [3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazo-
lium bromide] by actively growing cells to produce a
blue formazan product. Briefly, cells were seeded into 96
well plastic plates (2000 cells/well) and treated with 100
lL NAMI-A (1 mM) for 1 h, controls in PBS. After
treatment, cells were allowed to grow for further 24 h in
complete medium. Each well was then added with 50 lL
MTT solution (5 mg/mL in PBS, Sigma Chemical) and
incubated in thermostat for 4 h. At the end of incuba-
tion time, formazan crystals were solubilised with 100
lL dimethyl sulfoxide and optical density was measured
at 570 nm (Automated Microplate Reader EL311s,
BIO-TEKÒ Instruments, VT, USA).
5.11. Cell cycle analysis
(a) Double staining FITC/PI: 0.5  106 cells were
fixed in 70% EtOH for at least 4 h, washed twice with
PBS and allowed to balance in PBS for 2 h. Pellets were
stained with 0.5 mL of a PBS solution containing 10 lg
PI, FITC 0.25 ng fluorescein isothiocyanate (FITC) and
4 lg RNase (all from Sigma Chemicals) and left over-
night before flow cytometric analysis.
5.12. Flow cytometry analyses
Samples were analysed by the CoulterÒ EpicsÒ XLe
Flow Cytometer (Coulter Corporation, MI, FL, USA)
at the LINFA laboratories (Callerio Foundation ON-
LUS). For each sample, at least 10000 events were ac-
quired. Data were elaborated with the WinMDI vs. 2.7
(Scripps Research Institute, La Jolla, CA, USA) and
with MulticycleÒ (for cell cycle distribution).
5.13. Atomic absorption spectroscopy
Intracellular ruthenium content in treated cells was
determined by atomic absorption spectroscopy. Samples
were processed according to a modification of the pro-
cedures described by Tamura et al. [26]. All specimens
were dried in crio-vials. A first step of drying was per-
formed overnight at 80 °C, a second step at 105 °C, until
the samples reached the constant dried weight. The de-
composition of the dried samples was carried out with
411
the addiction of 100 lL of tetramethylammonium hy-
droxide (TMAH) at 25% in water (Aldrich Chimica, It-
aly) and 100 lL of milliQ directly in the vials. Volumes
were adjusted to 0.5 mL with milliQ water. Ruthenium
quantitation: the concentration of ruthenium in biolog-
ical samples was measured in triplicate by means of
Graphite Furnace Atomic Absorption Spectrometry
(GFAAS), model SpectrAA-300, supplied with a specific
ruthenium emission lamp (Hollow cathode lamp P/N 56-
101447-00) Varian, Mulgrave, Vic., Australia. Before
each daily analysis session a five point calibration curve
was traced using Ruthenium Custom-Grade Standard
998 mg mL1 , Inorganic Ventures (St. Louis, USA). In
order to correct for possible deterioration of the graphite
furnace during a daily working session, after every twelve
samples the calibration curve was re-traced and standard
was measured every six samples. The lower and higher
limits of quantitation (LLQ, HLQ) were set at the con-
centration levels which corresponds respectively to the
lowest and higher standard concentration employed. The
limit of detection (LOD) was estimated according to:
EURACHEM guide ‘‘The fitness for purpose of ana-
lytical experimental section’’. LLQ, HLQ and LOD were
respectively: 20, 100 and about 10 ng Ru mL1 of sample.
The quantitation of ruthenium was carried out in 10 lL
samples at 349.9 nm with an atomising temperature of
2500 °C, using argon as purge gas at a flow rate of
3.0 L min1 . For further details concerning the furnace
parameter settings, see [27].
5.14. Statistical analysis
Data were submitted to computer-assisted statistical
analysis (Instat II, Graphpad Software, vs. 2.05) using
the Student–Newman–Keuls for analysis of variance
(ANOVA) and t test for grouped data. Data were con-
sidered to be statistically significant if the probability
factor ‘‘p’’ was <0.05, <0.01 and <0.001. Symbols used
to identify the significance are *p < 0:05, **p < 0:01 and
***p < 0:001.
6. Abbreviations
NAMI-A [H2 im][trans-RuCl4 (dmso-S)(Him)]
NAMI-AH2 O [mer-RuCl3 (dmso-S)(Him)(H2 O)]
Him imidazole
dmso dimethyl sulfoxide
dmso-S sulfur-bonded dimethyl sulfoxide
9-MeAde 9-methyladenine
KB human ovarian carcinoma cell line
metGM murine metastatic cell line
PI propidium iodide
FITC fluorescein isothiocyanate
PBS phosphate-buffered saline
412 M. Bacac et al. / Journal of Inorganic Biochemistry 98 (2004) 402–412
Acknowledgements
Marie Curie Training Fellowship (Medicinor re-
search program 1.4.1; Contract No. HPMT-CT-2000-
00192; type PHD20) and Fondazione C&D Callerio
Onlus are acknowledged for financial support of the
Ph.D. fellowship of M. Bacac. Work contributed by
MIUR ‘‘Pharmacological Mechanisms of the Antimet-
astatic Activity of Metal Based Drugs’’, and done within
the framework of the EU COST D20/0005/01 and D20/
0001/00 Working Groups. Special thanks to Fons
Lefeber and Kees Erkelens for their generous help and
technical assistance during NMR measurements and to
Moreno Cocchietto for AAS assistance.
References
[1] M.J. Clarke, F. Zhu, D.R. Frasca, Chem. Rev. 99 (1999) 2511–2533.
[2] J. Reedijk, Proc. Natl. Acad. Sci. USA 100 (2003) 3611–3616.
[3] G. Sava, A. Bergamo, Int. J. Oncol. 17 (2000) 353–365.
[4] E. Alessio, G. Balducci, A. Lutman, G. Mestroni, H. Calligaris,
W.M. Attia, Inorg. Chim. Acta 203 (1993) 205–217.
[5] G. Sava, E. Alessio, A. Bergamo, G. Mestroni, in: M.J. Clarke,
P.J. Sadler (Eds.), Sulfoxide Ruthenium Complexes: Non Toxic
Tools for the Selective Treatment of Solid Tumor Metastases,
1999, pp. 143–170.
[6] G. Sava, S. Zorzet, L. Peressin, G. Mestroni, G. Zassinovich, A.
Bontempi, Inorg. Chim. Acta 137 (1987) 69–71.
[7] E. Alessio, W. Attia, M. Calligaris, D. Cauci, L. Dolzani, G.
Mestroni, C. Monti-Bragadin, G. Nardin, F. Quadrifoglio, G.
Sava, M. Tamaro, S. Zorzet, in: M. Nicolini (ed.), Boston,
Platinum and Other Metal Coordination Compounds in Cancer
Chemotheraphy, N€yhoff 1988, pp. 617–633.
[8] A. Bergamo, R. Gagliardi, V. Scarcia, A. Furlani, E. Alessio, G.
Mestroni, G. Sava, J. Pharmacol. Exp. Ther. 289 (1999) 559–564.
[9] S. Pacor, M. Vadori, F. Vita, M. Bacac, M.R. Soranzo, G.
Zabucchi, G. Sava, Anticancer Res. 21 (2001) 2523–2530.
[10] S.J. Berners Price, P.J. Sadler, Coord. Chem. Rev. 151 (1996) 1–
40.
[11] J. Malina, O. Novakova, B.K. Keppler, E. Alessio, V. Brabec, J.
Biol. Inorg. Chem. 6 (2001) 435–445.
[12] G. Mestroni, E. Alessio, G. Sava, S. Pacor, M. Coluccia, A.
Boccarelli, Metal Based Drugs 1 (1993) 41–63.
[13] A. K€ ung, T. Pieper, R. Wissiack, E. Rosenberg, B.K. Keppler, J.
Biol. Inorg. Chem. 6 (2001) 292–299.
[14] R.M. Izatt, J.J. Christensen, J.H. Rytting, Chem. Rev. 71 (1971)
439–481.
[15] K. Inagaki, M. Kuwayama, Y. Kidani, J. Inorg. Biochem. 16
(1982) 59–65.
[16] T.G. Spiro, Nucleic Acid-Metal Ion Interactions, Wiley, New
York, 1980.
[17] S.J. Lippard, Progress in Inorganic Chemistry, Wiley, New York,
1977.
[18] A.H. Velders, B. van der Geest, H. Kooijman, A.L. Spek, J.G.
Haasnoot, J. Reedijk, Eur. J. Inorg. Chem. (2001) 369–372.
[19] A.C.G. Hotze, M.E.T. Broekhuisen, A.H. Velders, K. van der
Schilden, J.G. Haasnoot, J. Reedijk, Eur. J. Inorg. Chem. (2002)
369–376.
[20] G. Sava, A. Bergamo, S. Zorzet, B. Gava, C. Casarsa, M.
Cocchietto, A. Furlani, V. Scarcia, B. Serli, E. Iengo, E. Alessio,
G. Mestroni, Eur. J. Cancer 38 (2002) 427–435.
[21] M.M. Taqui Khan, G. Ramachandraiah, R.S. Shukla, Inorg.
Chem. 27 (1988) 3274–3278.
[22] A. Bergamo, S. Zorzet, B. Gava, A. Sorc, E. Alessio, E. Iengo, G.
Sava, Anti-cancer Drugs 11 (2000) 665–672.
[23] L. Messori, P. Orioli, D. Vullo, E. Alessio, E. Iengo, Eur. J.
Biochem. 267 (2000) 1206–1213.
[24] E.G. Talman, W. Br€ uning, J. Reedijk, A.L. Spek, N. Veldman,
Inorg. Chem. 36 (1997) 854–861.
[25] H. Tada, O. Shiho, Kuroshima, K. Koyama, M. Tsukamoto, J.
Immunol. Meth. 93 (1986) 157–165.
[26] H. Tamura, T. Arai, M. Nagase, N. Ichinose, Bunseki Kagaku 41
(1992) T13–T18.
[27] M. Cocchietto, G. Sava, Pharmacol. Toxicol. 87 (2000) 193–197.