Increasing the bioavailability of RuIII anticancer complexes through hydrophobic albumin interactions

Article abstract of DOI:10.1002/chem.201302671

A series of pyridine-based derivatives of the clinically successful Ru ^III-based complexes indazolium [trans-RuCl₄(1 H-indazole)₂] (KP1019) and sodium [trans-RuCl₄(1 H-indazole)₂] (KP1339) have been synthesized to probe the effect of hydrophobic interactions with human serum albumin (hsA) on anticancer activity. The solution behavior and protein interactions of the new compounds were characterized by using electron paramagnetic resonance (EPR) and UV/Vis spectroscopy. These studies have revealed that incorporation of hydrophobic substituents at the 4′-position of the axial pyridine ligand stabilizes non-coordinate interactions with hsA. As a consequence, direct coordination to the protein is inhibited, which is expected to increase the bioavailability of the complexes, thus potentially leading to improved anticancer activity. By using this approach, the lifetimes of hydrophobic protein interactions were extended from 2 h for the unsubstituted pyridine complex, to more than 24 h for several derivatives. Free complexes were tested for their anticancer activity against the SW480 human colon carcinoma cell line, exhibiting low cytotoxicity. Pre-treatment with hsA improved the solubility of every compound and led to some changes in activity. Particularly notable was the difference in activity between the methyl-and dibenzyl-functionalized complexes. The former shows reduced activity after incubation with hsA, indicating reduced bioavailability due to protein coordination. The latter exhibits little activity on its own but, following treatment with hsA, exhibited significant cytotoxicity, which is consistent with its ability to form non-coordinate interactions with the protein. Overall, our studies demonstrate that non-coordinate interactions with hsA are a viable target for enhancing the activity of Ru^III-based complexes in vivo. Albumin: friend or foe? A series of Ru-based anticancer candidates has been synthesized with pyridine-based ligands (see figure), to probe their non-coordinate interactions with albumin. Significant stabilization of hydrophobic binding was demonstrated and correlated with reduced coordination of the complexes with the protein. Biological testing suggests that this is a promising strategy for enhancing the solubility and anticancer activity of Ru complexes.

Full text of DOI:10.1002/chem.201302671

FULL PAPER
DOI: 10.1002/chem.201302671
Increasing the Bioavailability of RuIII Anticancer Complexes through
Hydrophobic Albumin Interactions
Michael I. Webb,[a] Boris Wu,[a] Thalia Jang,[a] Ryan A. Chard,[a] Edwin W. Y. Wong,[a]
May Q. Wong,[b] Donald T. T. Yapp,[b] and Charles J. Walsby*[a]
Abstract: A series of pyridine-based
tein is inhibited, which is expected to
increase the bioavailability of the com-
plexes, thus potentially leading to im-
proved anticancer activity. By using
this approach, the lifetimes of hydro-
phobic protein interactions were ex-
tended from 2 h for the unsubstituted
pyridine complex, to more than 24 h
for several derivatives. Free complexes
were tested for their anticancer activity
against the SW480 human colon carci-
noma cell line, exhibiting low cytotox-
icity. Pre-treatment with hsA improved
the solubility of every compound and
led to some changes in activity. Particu-
larly notable was the difference in ac-
tivity between the methyl- and diben-
derivatives of the clinically successful
RuIII-based
complexes
indazolium
[trans-RuCl4(1H-indazole)2] (KP1019)
and sodium [trans-RuCl4(1H-inda-
zole)2] (KP1339) have been synthesized
to probe the effect of hydrophobic in-
teractions with human serum albumin
(hsA) on anticancer activity. The solu-
tion behavior and protein interactions
of the new compounds were character-
ized by using electron paramagnetic
resonance (EPR) and UV/Vis spectros-
copy. These studies have revealed that
incorporation of hydrophobic substitu-
ents at the 4’-position of the axial pyri-
dine ligand stabilizes non-coordinate
interactions with hsA. As a conse-
quence, direct coordination to the pro-
zyl-functionalized
complexes.
The
former shows reduced activity after in-
cubation with hsA, indicating reduced
bioavailability due to protein coordina-
tion. The latter exhibits little activity
on its own but, following treatment
with hsA, exhibited significant cytotox-
icity, which is consistent with its ability
to form non-coordinate interactions
with the protein. Overall, our studies
demonstrate that non-coordinate inter-
actions with hsA are a viable target for
enhancing the activity of RuIII-based
complexes in vivo.
Keywords: albumin · cytotoxicity ·
hydrophobic interactions · ligand
design · ruthenium
Introduction
Amongst the most promising ruthenium-based drug can-
didates are negatively charged octahedral RuIII complexes
with axial azole ligands, an equatorial plane of chloride li-
gands, and charge compensation provided either by the pro-
tonated forms of the azole ligands or sodium ions
(Figure 1).[6] These compounds were first described by Kep-
pler and co-workers in the early 1990s,[7] and continue to
serve as the benchmark for activity in the development of
new Ru-based anticancer agents. “Keppler-type” complexes
typically exhibit conventional antineoplastic properties and
have demonstrated significant activity in colorectal cancer
cell lines.[8] One such compound, indazolium [trans-
RuCl4(1H-indazole)2] (KP1019) (Figure 1) has drawn partic-
ular attention after successfully completing phase I clinical
trials.[2a,8b] Although KP1019 ultimately failed in phase II
trials, primarily due to its poor solubility,[9] further clinical
development of its more soluble sodium-compensated ana-
logue, KP1339 (Figure 1), is continuing.[8e,10]
Ruthenium compounds are drawing increasing attention as
the next generation of metal-based anticancer agents.[1] Clin-
ical studies of Ru chemotherapeutics have reported signifi-
cant antineoplastic and antimetastatic activity, with very low
levels of side effects.[2] The on-going development of Ru-
based anticancer complexes is promoted by their fundamen-
tal chemical properties, which include 1) variable ligand-ex-
change rates, which are typically favorable with respect to
rates of cell division,[3] 2) a range of oxidation states, which
are accessible under physiological conditions (i.e., II–IV),[4]
and 3) properties which are tunable by ligand design.[5]
[a] Dr. M. I. Webb, B. Wu, T. Jang, R. A. Chard, Dr. E. W. Y. Wong,
Prof. C. J. Walsby
Department of Chemistry, Simon Fraser University
Burnaby, BC, V5A 1S6 (Canada)
Fax : (+1)778-782-3765
Significant effort has been expended to uncover the bio-
logical targets and anticancer mechanisms of Keppler-type
compounds by using an array of analytical methods.[9,11] Al-
though their specific mode of action remains unclear, it is
widely accepted that these complexes are prodrugs, with
aqueous ligand exchange playing a significant role in their
physiological speciation.[8b] Although their azole ligands are
E-mail: cwalsby@sfu.ca
[b] M. Q. Wong, Dr. D. T. T. Yapp
Department of Experimental Therapeutics
BC Cancer Agency
Vancouver, BC, V5Z 4E6 (Canada)
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/chem.201302671.
Chem. Eur. J. 2013, 19, 17031 – 17042
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
17031

Overall, these observations have suggested that interactions
with hsA can be tuned by the properties of the axial ligands.
Based on these observations, we have hypothesized that
the bioavailability of Keppler-type complexes could be en-
hanced by promoting the hydrophobic interactions with hsA
and consequently inhibiting the formation of coordinated
species. To explore this concept we have synthesized a series
of pyridine-based KP1019/1339 analogues. These compounds
have a range of substituents at the 4’-position of the axial
pyridine ligands, generating complexes with different ten-
dencies to interact with the hydrophobic regions of hsA.
The ligands chosen for these studies were: pyridine (Pyr, 1),
4-methylpyridine (MePyr, 2), 4-phenylpyridine (PhPyr, 3).
diphenyl-4-pyridyl-methane (DiPhenPyr, 4), and 4-(4-nitro-
benzyl)pyridine (NBenzPyr, 5) (Figure 1). EPR and UV/Vis
spectroscopy has been used to monitor how each of these li-
gands impacts on protein binding. We found that the differ-
ent substituents have a marked effect on the stability of
non-coordinate interactions, which correlates with their an-
ticipated ability to interact with hydrophobic protein re-
gions. Each complex was synthesized with both protonated
pyridine ligands (series a) and sodium (series b) as compen-
sating cations (Figure 1). The latter were required to make
the compounds sufficiently soluble for testing of their anti-
cancer activity.
Figure 1. The Keppler-type RuIII-based anticancer agents undergoing clin-
ical evaluation, KP1019 and KP1339, and their pyridine-based analogues
synthesized in this study.
generally kinetically inert,[11g,12] the chloride ligands of these
complexes readily exchange in aqueous solution, a process
that has a direct impact on the activity of the complexes by
generating aquated species.[11b,e,13] In the case of KP1019,
aqueous exchange is particularly significant because the
mono-aqua derivative of the complex is insoluble and rapid-
ly precipitates out of solution.[14]
Under physiological conditions, Keppler-type complexes
have been shown to bind readily to serum proteins,[15] in par-
ticular to human serum albumin (hsA).[11c,f,16] For KP1019
the presence of serum proteins inhibits its precipitation due
to the formation of soluble protein-bound species.[11b] In the
case of hsA, the complexes can bind either through direct
coordination to amino acid side chains,[6,8b,14b,15,17] or in
a non-coordinate fashion, likely within the hydrophobic
binding domains of the protein[18] as found in Sudlowꢁs
sites I and II.[19]
To see how the improved stability of the hydrophobic in-
teractions correlates with anticancer activity, we have tested
the compounds against the SW480 human colon carcinoma
cell line. In these experiments, we have screened the com-
plexes both after preparation in buffered solution and fol-
lowing initial incubation with hsA. These studies indicate
that the enhancement of non-coordinate interactions can im-
prove the activity of the complexes both through inhibition
of protein coordination and by increasing solubility.
Results and Discussion
We have previously used electron paramagnetic resonance
(EPR) spectroscopy to characterize both coordinate and
non-coordinate interactions of KP1019, and its cytotoxic
imidazole analogue, imidazolium [trans-RuCl4(1H-imida-
zole)2] (KP418), with hsA.[14b] These studies have shown that
the axial heterocyclic ligands of these complexes play a cen-
tral role in their protein binding behavior. KP1019 forms
non-coordinate interactions with hsA very rapidly, with the
resulting species converted readily to protein-coordinated
complexes over time.[14b] By contrast, KP418 slowly forms
non-coordinate interactions with hsA following aqueous
ligand exchange. These interactions are persistent and the
formation of coordinated species does not occur even after
prolonged incubation with the protein.[14b,20] The ability of
KP1019 to rapidly form non-coordinate interactions with
hsA has been attributed to enhanced interactions of its inda-
zole ligands with the hydrophobic domains of hsA,[14b] as
compared to the imidazole ligands of KP418.[20] This has
been implicated in the lower side effects of KP1019.[5c,14b]
Synthesis: Compounds 1a–5a and 1b–5b were synthesized
by using procedures derived from the original syntheses of
KP1019 and KP1339, respectively.[7] An alternative synthesis
and characterization of complex 1a has been described else-
where.[21] However, this is the first report of the sodium ana-
logue of this compound (i.e., compound 1b) and all the
other complexes, with either type of counterion. The identi-
ty and purity of the compounds was confirmed by elemental
analysis, NMR, and EPR spectroscopy as well as X-ray crys-
tallography analysis.
There have been a number of reports of derivatives of
RuIII-based Keppler-type complexes with different heterocy-
clic nitrogen ligands.[22] However, to the best of our knowl-
edge, this is the first study using systematic functionalization
at a single position on the coordinated heterocyclic ligands
to manipulate the properties of the complexes.
Crystal structures: The structures of compounds 1a, 2a, 3b,
4a, and 5a were determined by X-ray crystallography and
17032
www.chemeurj.org
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2013, 19, 17031 – 17042

Increasing the Bioavailability of RuIII Anticancer Complexes
FULL PAPER
ruthenium metal center for all of the complexes are listed in
Table S4 in the Supporting Information. In all cases the two
pyridine ligands are coordinated trans through the heterocy-
ꢀ
clic nitrogen atom with an average Ru N bond length of
2.097 ꢂ, which is slightly longer than the reported value for
both KP418 (2.079 ꢂ)[23] and KP1019 (2.061 ꢂ).[7b] An equa-
torial plane of four chloride ligands completes the coordina-
tion sphere with bond lengths ranging from 2.335–2.384 ꢂ,
which are within experimental error of KP418 (2.342–
2.356 ꢂ) and KP1019 (2.358–2.372 ꢂ). This type of coordi-
nation environment is typical of analogous complexes.[22]
Aqueous solution behavior: Due to the clinical success of
KP1019, the solution behavior of this complex has been ex-
tensively studied by using a variety of techniques.[11bd,13c,24]
Aqueous exchange of a chloride ligand occurs within mi-
nutes under physiological conditions, resulting in the precip-
itation of the insoluble mono-aquated derivative.[14a] This
precipitation event can be inhibited in the presence of other
ligands, particularly proteins or other biomolecules.[11e,14a]
For the compounds 1a–5a and 1b–5b reported here, we
first characterized their ligand-exchange processes in phos-
phate-buffered saline (PBS, pH 7.4) to determine whether
their fundamental solution behavior was similar to KP1019.
Because RuIII is paramagnetic (low-spin d5, S=1= ), EPR
2
spectroscopy can be used to identify changes in the ligand
environment around the metal center by analyzing the ob-
served g values, the line widths, and the signal intensities.
From these studies, species produced by ligand exchange
can be identified, along with their rates of formation. Fur-
thermore, processes that produce diamagnetic species, such
as reduction to give RuII (low-spin d6, S=0), can be moni-
tored by a loss of the overall signal intensity due to the pro-
duction of “EPR-silent” species. To track the solution be-
havior of the various RuIII complexes, incubation under
physiological conditions (pH 7.4, 378C) was performed for
up to 2 h, followed by freezing in liquid nitrogen after se-
lected time points. EPR measurements of frozen solutions
then provided a “snap-shot” of the paramagnetic species
present at each time interval and their relative concentra-
tions.
Figure 2. Crystal structures of the anions of complexes 1a, 2a, 3b, 4a,
and 5a. Structures are drawn at the 50% probability level. Counterions
as well as coordinated solvent molecules are omitted for clarity.
The ability of the RuIII complexes 1a–5a and 1b–5b to
undergo multiple ligand-exchange steps complicates the
analysis of their EPR spectra by producing overlapping sig-
nals that require deconvolution. Spectral analyses were per-
formed by simulating the spectra of each individual species
present in solution, as defined by their g values and line
widths. The spectrum from each spectral component was
subsequently multiplied by a weighting factor, correspond-
ing to the relative concentration of each RuIII species pres-
ent. The weighted spectra of each species identified were
then summed together to reproduce the experimental data.
To ensure that a unique solution was found for the spectral
parameters of each species, only the weighting factors were
varied between simulations of the EPR spectra at different
incubation time points for each complex. We have previous-
ly demonstrated this approach in studies of KP1019, KP418,
are shown in Figure 2. For clarity, the counterions for each
complex are omitted, along with any additional co-crystalliz-
ing solvent molecules. The diverse properties of the axial li-
gands required a variety of solvent conditions to produce
suitably diffracting crystals. The structure of each anion type
was determined, and other spectroscopic and analytical
techniques were used to confirm the synthesis of the com-
pounds with the alternate counterions. The crystal structure
of complex 1a has been described previously,[21] but none of
the other compounds have been reported or characterized
crystallographically.
For all of the structures solved, a similar distorted octahe-
dral geometry is observed. The bond lengths around the
Chem. Eur. J. 2013, 19, 17031 – 17042
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
17033

C. J. Walsby et al.
and a number of other RuIII-based anticancer drug candi-
dates.[14b,25]
Complexes 1a and 1b as well as 2a and 2b had sufficient
ACHTUNGTRENNUNG
aqueous solubility to be dissolved directly in PBS at a 3 mm
concentration and studied by EPR spectroscopy. The other
complexes, 3a–5a and 3b–5b, required the addition of 10%
DMSO to give sufficient concentrations. Immediately after
dissolution, similar uniaxial EPR spectra were observed for
each of the parent RuIII complexes, consistent with the tet-
ragonal symmetry of the RuIII centers prior to ligand ex-
change (1b-C1 and 4a-C1 in Figure 3). Although the per-
pendicular part (g? ) of the EPR spectrum was well resolved
in each case, gj j could not be detected. A similar behavior
has been reported in EPR measurements of KP1019 and
KP418, and has been assigned to line broadening due to g
strain.[14b,26] In the earlier study of KP1019 and KP418,
rapid-passage, dispersion-mode measurements determined
gj j =1.20 for both complexes. Given the structural and
bonding similarity of the ruthenium centers with the com-
pounds reported here, the value of gj j is likely to be very
similar. Thus, a value of gj j =1.20 was used for the simula-
tion of 1a–5a and 1b–5b, with
a large line width
(400 Gauss), but it should be noted that this has essentially
no impact on the shape of the spectra in the perpendicular
spectral region. Simulation of the perpendicular region gave
values in the range g? =2.64–2.66, and line widths of LW? =
105–200 Gauss (see Figures S1, S2, and S6–S15 as well as Ta-
bles S1 and S2 in the Supporting Information). The different
counterions had no effect on the simulations for the spectra
of complexes 1a and 1b as well as 2a and 2b. Similar EPR
data was observed for complexes 3a and 3b, whereas for
complexes 4a and 3b as well as 5a and 5b, the spectra from
the pyridinium-compensated complexes were significantly
broader than those with sodium ion compensation. This
broadening is likely due to the highly hydrophobic nature of
the pyridinium cations, leading to some aggregation of the
complexes in solution.[25a]
Incubation of 3 mm solutions of each complex in PBS at
378C resulted in both precipitation and the formation of
some soluble derivatives. We have previously used EPR
spectroscopy to characterize ligand-exchange rates for both
KP1019 and KP418.[14b] In this work, we found rates of aque-
ous ligand exchange for the new pyridine analogues that are
similar to KP1019. As expected, increasing the hydrophobic-
ity of the axial ligands increased the rate of precipitation
with incubation under physiological conditions. EPR meas-
urements report primarily on the complexes in solution, be-
cause aggregation of precipitated compounds leads to spec-
tral broadening. As a result, in these studies precipitation is
reflected in a decrease in the overall EPR signal intensity
with increasing incubation. For all of the complexes studied
here, prolonged incubation was accompanied by a color
change from clear yellow to dark green. In the case of com-
plexes 3a–5a and 3b–5b significant amounts of a green-blue
precipitate were also observed after longer incubation times.
These observations are consistent with the formation of the
mono-aqua derivatives, as reported for KP1019.[14a,27]
Figure 3. EPR measurements of: a) complex 1b and b) complex 4a, incu-
bated in PBS at 378C, and spectral deconvolution of spectra collected
after 30 and 120 min of incubation. As shown by the spectral simulations,
one new rhombic signal is observed for complex 1b (1b-C2), whereas
two axially symmetric species are observed for complex 4a (4a-C2 and
4a-C3). Experimental parameters: frequency=9.38 GHz, microwave
power=2.0 mW, time constant=40.96 ms, modulation amplitude=6 G,
average of five two-minute scans. For EPR measurements of the solution
behavior of the other complexes, see Figures S1 and S2 in the Supporting
Information. For spectral deconvolution of spectra collected at each incu-
bation time point for each complex, see Figures S6–S15 in the Supporting
Information. For spectral parameters used in each simulation see Ta-
bles S1 and S2 in the Supporting Information.
To characterize the behavior of the complexes remaining
in solution, EPR measurements were performed after incu-
bation times of 0, 30, and 120 min. Spectral components
17034
www.chemeurj.org
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2013, 19, 17031 – 17042

Increasing the Bioavailability of RuIII Anticancer Complexes
FULL PAPER
were analyzed by simulation, as described above and in the
Experimental Section. In the case of complexes 1a and 1b
a single new species was observed, which was identified by
a rhombic EPR spectrum with g=[2.44, 2.26, 1.72] and line
widths of [130, 70, 300 Gauss], labeled 1b-C2 in Figure 3.
This is consistent with the formation of the mono-aqua com-
plex formed by the exchange of an equatorial chloride
ligand. Similarly, for complexes 2a and 2b, only one new
species was detected following dissolution. This was ob-
served after 30 min of incubation: labeled 2a-C2 and 2b-C2
in Figures S8 and S9 in the Supporting Information with g=
[2.47/2.45, 2.27/2.26, 1.78/2.00] and line widths of [125, 100/
50, 400 Gauss]. The similarity in the spectral parameters to
complexes 1a-C1 and 1b-C1 indicates that these complexes
also form a mono-aqua complex through chloride exchange.
Under physiological conditions nitrogen atom donor li-
gands are generally considered to be kinetically inert,[28]
leaving only the equatorial chloride ligands available for ex-
change with water molecules. To confirm this for complexes
1a and 1b as well as 2a and 2b, NMR measurements were
made of the sodium-compensated complexes 1b and 2b in
buffered D2O solutions. These solutions were incubated at
where L is an axial pyridine-based ligand, DMSO takes up
the opposite axial position, and charge compensation is pro-
vided by a pyridinium or sodium cation. Therefore, the first
species to form, that is, C2, is assigned to exchange of an
axial pyridine ligand for DMSO. The second type of species,
that is, C3, which forms with longer incubation, has spectral
parameters that match those of the NAMI-A-type com-
plexes after the DMSO ligand had been exchanged for
a water molecule, indicating that the same process is occur-
ring here for complexes 3a–5a.
As further evidence that this process involves exchange of
a pyridine-based ligand for DMSO, we repeated the experi-
ments with 50 and 100% DMSO. The signal from complexes
3a-C2–5a-C2, was observed with increased intensity (see
Figure S26 in the Supporting Information), which is consis-
tent with the formation of the DMSO complex. 1H NMR
measurements could not be used in this case to detect ex-
changed pyridine-based ligands because their signals were
masked by the pyridinium counterions, whereas the signal
intensity was too low for paramagnetic NMR spectroscopy
to detect coordinated DMSO molecules.
The replacement of pyridine ligands by DMSO is not un-
precedented,[30] but has not been reported for KP1019 under
physiological conditions.[2c] To confirm this, KP1019 was dis-
solved in PBS with 10% DMSO and measured by EPR
spectroscopy (see Figure S58 in the Supporting Informa-
tion). The resulting spectra are identical to those previously
reported for the complex in DMSO-free buffer.[11b,14b] Fur-
thermore, with prolonged incubation at 378C the EPR spec-
trum was unchanged and no precipitation occurred. This
demonstrates that DMSO enhances the solubility of KP1019
by inhibiting the formation of the insoluble mono-aqua spe-
cies, but does lead to an exchange of its indazole ligands.
The ligand-exchange processes of each compound were
also characterized by UV/Vis spectroscopy. Samples were
prepared in PBS at a concentration of 200 mm. Following the
EPR experiments, solutions of complexes 3a–5a and 3b–5b
included 10% DMSO. The samples were incubated at 378C
within the spectrophotometer with measurements taken
every 10 min for two hours.
The UV/Vis spectra of complexes 1a and 1b (Figure 4
and Figures S27 and S28 in the Supporting Information) are
essentially identical with absorption bands in both the ultra-
violet and the visible range. The most intense peak is ob-
served at l=257 nm, and is assigned to the auxochrome
signal from the pyridine ligand.[31] The next most intense ab-
sorbance peak occurs at l=281 nm, which is consistent with
pyridine p–p* transitions for RuIII complexes.[32] A third
peak at l=360 nm is within the typical range of ligand-to-
metal charge-transfer (LMCT) transitions for octahedral
RuIII complexes with bound nitrogen heterocycles.[33] With
incubation, the absorbances at l=257 and 360 nm decrease,
whereas the peak at l=281 nm shifts slightly to longer
wavelength. A well-defined isosbestic point (l=347 nm) is
observed, which is consistent with only one ligand-exchange
pathway. This is in agreement with a single aquation step as
determined by EPR spectroscopy, whereas the decrease in
1
378C for 2 h and subsequently analyzed by using H NMR
spectroscopy to monitor the pyridine or 4-methylpyridine li-
gands, respectively (see Figures S61 and S62 in the Support-
ing Information). In both cases, no free pyridine-based li-
gands were observed, thereby confirming that only chloride
ligand exchange occurs.
No soluble ligand-exchanged complexes were observed
for the sodium-compensated complexes 3b–5b. The more
hydrophobic ligands of these compounds, as compared to
complexes 1a and 1b as well as 2a and 2b, likely led to
highly insoluble aquated derivatives, which precipitate rap-
idly and were thus not observed by EPR spectroscopy.
Although the pyridinium-compensated complexes 3a–5a
also precipitate readily, two new uniaxial EPR signals with
relatively low intensity are observed. The first species to
form (g? =2.44, gj j =1.76; LW? =90 and LWj j =125 Gauss)
is labeled 4a-C2 in Figure 3 and correspondingly for com-
plexes 3a and 5a in Figures S10 and S14 in the Supporting
Information. With further incubation this was replaced by
a second new species (g? =2.30, gj j =1.88; LW? =50 and
LW j j =50 Gauss), identified as 4a-C3 in Figure 3 and simi-
larly for the other complexes in the Supporting Information.
Interestingly, the uniaxial EPR spectra for each species indi-
cate tetragonal symmetry, which requires exchange of the
pyridine-based ligands. To the best of our knowledge, this
has not been reported previously for Keppler-type com-
plexes under physiological conditions.
As described above, to achieve sufficient solubility for
EPR measurements, DMSO was added to the solutions of
3a–5a and 3b–5b, a procedure that is also common in the
preparation of these types of compounds for biological test-
ing. The spectral parameters of the first species formed from
complexes 3a–5a, match those previously reported for
“NAMI-A-type” pyridine complexes.[25a,29] These complexes
have the general formula [trans-RuCl4ACTHNUGRTENNU(G 1H-L)ACHUTTGNREN(NUNG DMSO-S)],
Chem. Eur. J. 2013, 19, 17031 – 17042
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
17035

C. J. Walsby et al.
(KP418), non-coordinate protein interactions are also ob-
served, however, they form at a much slower rate.[14b,20] The
reduced rate of hsA binding by KP418 has been suggested
to leave more of the complex free in vivo, and is possibly re-
sponsible for its observed toxicity.[39] In light of these re-
ports, we have studied the ability of complexes 1a–5a and
1b–5b to bind to hsA, both through hydrophobic interac-
tions and through direct coordination.
The complexes 1a–5a and 1b–5b were incubated for 0
and 30 min as well as 1, 2, 6, and 24 h at 378C in solutions
of hsA in PBS. Protein-bound fractions were then isolated
by using centrifugal ultrafiltration and studied by EPR spec-
troscopy. This approach allows us to observe exclusively
hsA-bound RuIII species while removing contributions from
free complexes in solution.[14b,25] The resulting spectra for
select incubation times are shown in Figure 5 for complexes
1a, 3a, and 5a, and in the Supporting Information (Fig-
Figure 4. UV/Vis spectra for complex 1a incubated in PBS for 2 h at
378C.
the peak at l=257 nm is consistent with the observed pre-
cipitation in the sample.
ACHTUNGTRENNUNG
Similar absorbances and spectral changes with incubation
were observed for complexes 2a and 2b (Figures S29 and
S30 in the Supporting Information), which is consistent with
a single aquation step, as also observed by EPR spectrosco-
py. Complexes 3a–5a and 3b–5b had spectra dominated by
broad, intense features that did not change significantly with
incubation (Figures S31–S36 in the Supporting Information).
These absorbances are from p–p* transitions of the phenyl-
based substituents. However, subtle changes in the signal in-
tensities indicate ligand-exchange processes.
Overall, these studies demonstrate that aqueous solubility
is a significant issue for all of these compounds, with precipi-
tation occurring readily in buffered solutions. However, con-
tributions from soluble mono-nuclear RuIII species are also
notable and there is some diversity in the mechanisms of
their formation due to the different axial ligands.
parent compounds of each complex were observed even at
the earliest time point (labeled “0 min” in Figure 5). This
was confirmed by spectral simulation by using the parame-
ters determined from the complexes dissolved directly in
buffer, and identified previously as species C1. Because
these measurements are of isolated protein fractions, they
demonstrate that each complex readily binds to hsA without
exchanging ligands, indicating non-coordinate interactions
with the protein. Furthermore, the EPR spectra of the pyri-
dinium-compensated compounds 4a and 5a exhibited nar-
rower line widths than in buffer alone (labeled “4a-C1-hsA”
and “5a-C1-hsA” in Figures S22 and S24 in the Supporting
Information), indicating that protein interactions improve
the solubility of the complexes.
Aquated derivatives of the complexes 1a and 1b as well
as 2a and 2b were also observed to form non-coordinate in-
teractions with hsA (see Figures S16–S19 in the Supporting
Information). Simulation of the EPR spectra from these
species gave parameters that agreed within experimental un-
certainty with the corresponding compounds in buffer alone.
For complexes 3a as well as 4a and 4b the NAMI-A-type
complexes formed by axial ligand exchange were again ob-
served as minor contributors in the experimental spectrum
(see Figures S20, S22, and S23 in the Supporting Informa-
tion), whereas only the parent complex for complexes 3b,
5a, and 5b was visible (see Figures S24 and S25 in the Sup-
porting Information), demonstrating the influence of protein
interactions on ligand-exchange processes. In previous stud-
ies, we have also observed aquated derivatives of KP418
and NAMI-A interacting with hsA.[14b,25b]
Interactions of complexes with hsA: Human serum albumin
is the most abundant protein in the circulatory system, and
has been identified as a primary target in the transportation
of many drugs to their sites of activity.[34] Several studies
have shown that hsA interactions are prevalent for RuIII
complexes under physiological conditions.[11f,16,35] The pro-
pensity of albumin to form interactions with a variety of
pharmaceuticals has been reported[36] and is considered to
be an important factor in their delivery to their sites of
action. Indeed, the use of albumin in drug delivery has de-
veloped into a promising area of medicinal chemistry.[37]
However, the majority of the uses of albumin rely on coor-
dinate interactions between the pharmaceutical and the pro-
tein.[38] Here, we focus on optimizing non-coordinate hsA in-
teractions, in part to reduce coordination to the protein and
thereby increase the bioavailability of the complex.
Further incubation of each complex with hsA resulted in
a gradual decrease in the signals from the parent compounds
and their exchanged derivatives, coupled with the appear-
ance of new broad features in their EPR spectra. Spectral
simulations revealed two new signals in all cases, with rela-
tive intensities and formation rates that varied for each com-
plex. The first new species, labeled “hsA-1” in Figure 5 (see
Figures S16–S25 in the Supporting Information for each
We have previously used EPR methods to show that
KP1019 forms rapid non-coordinate interactions with
hsA,[14b] which are gradually converted to coordinate protein
interactions after further incubation under physiological
conditions. In the case of its toxic imidazole analogue
17036
www.chemeurj.org
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2013, 19, 17031 – 17042

Increasing the Bioavailability of RuIII Anticancer Complexes
FULL PAPER
Figure 5. EPR measurements of complexes a) 1a, b) 3a, and c) 5a incubated in PBS (1 mm, and 4% DMSO for complexes 3a and 5a) with hsA
(0.5 mm) at 378C, and spectral deconvolution of all spectra collected (shown in red) from the individual components (top of each column). Experimental
parameters: frequency=9.38 GHz, microwave power=2.0 mW, time constant=40.96 ms, modulation amplitude=6 G, average of five two-minute scans.
For EPR measurements of the hsA binding behavior of the other complexes, see the Supporting Information, Figures S3–S5. For spectral deconvolution
of spectra collected at each incubation time point for each complex, see the Supporting Information, Figures S16–S25. For spectral parameters used in
each simulation see the Supporting Information Tables S1 and S2.
complex), has a rhombic spectrum with g1 =2.38–2.61, g2 =
2.24–2.29, and g3 =1.20, and line widths of: [LW1 =175–400,
LW2 =100–150, and LW3 =400 Gauss]. The second species
also has a rhombic EPR signal, labeled “hsA-2” in Figure 5,
with g1 =2.38–2.45, g2 =2.16–2.22, and g3 =1.71–1.78, and
line widths of: [LW1 =80–150, LW2 =80–200, and LW3 =250–
300 Gauss] (see Tables S1 and S2 in the Supporting Informa-
tion for parameters for each complex). In each case, the dis-
tinctive g values and line widths of these species are consis-
tent with direct coordination to hsA through ligand ex-
change. Histidine coordination has previously been implicat-
ed in the protein binding of KP1019[17,40] and other RuIII
complexes,[2d,25b,41] and is expected to be the coordination
mode seen here. The rhombic EPR spectra observed are
consistent with coordination at equatorial positions, whereas
the presence of two distinct signals indicates concurrent
aquation processes. Given that hsA-1 has a greater signal in-
tensity, particularly at early time points, this is likely the spe-
cies formed by a single exchange with a histidine imidazole.
Subsequent aqueous exchange likely then produces the
second protein-bound species hsA-2.
Ru per protein molecule.[17,18] In this work we used a 2:1
ratio of each complex to hsA, to allow for the possibility of
different binding modes. Strong signals from hsA-1 and
hsA-2 were observed for every complex, demonstrating that
protein coordination is highly favored and that the rutheni-
um center remains in the 3+ oxidation state. Furthermore,
the signal intensity from the coordinated species was main-
tained even after 24 h of incubation, demonstrating how ef-
fectively these interactions stabilize the compounds in solu-
tion. A similar behavior has been reported previously for
KP1019.[11b,14]
To determine the effect of the different axial ligands of
complexes 1a–5a and 1b–5b on interactions with hsA, spec-
tra of each complex were collected at different incubation
time points and analyzed by simulation (see Figures S16–
S25 in the Supporting Information). From this analysis the
relative total signal intensities from non-coordinate and co-
ordinate interactions could be determined from the weight-
ing factors used to scale individual spectral components in
each simulation. Every complex showed significant non-co-
ordinate interactions with hsA after minimal incubation
times. However, the rate that these species were replaced by
Previous studies of RuIII-based anticancer candidates have
reported the coordination of upwards of five equivalents of
protein-coordinated
complexes
varied
dramatically
Chem. Eur. J. 2013, 19, 17031 – 17042
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
17037

C. J. Walsby et al.
the nitro groups. Increased sta-
bilization of hydrophobic bind-
ing due to electrostatic interac-
tions imparted by this function-
ality is in agreement with the
observation that hsA has a high
affinity for anionic species.[42]
Overall, non-coordinate binding
to hsA was more prolonged for
the pyridinium compensated
complexes, as compared to the
sodium
compensated
com-
plexes, likely reflecting their
aqueous insolubility, thereby
promoting interactions with the
hydrophobic regions of hsA,
and inhibiting ligand exchange.
UV/Vis measurements were
also used to characterize the in-
teractions of the complexes
Figure 6. Relative fractions of coordinate (yellow) and non-coordinate (red) hsA-bound RuIII complexes fol-
lowing 0, 2, and 24 h of incubation at 378C, as determined by simulation of the EPR spectra.
(Figure 6). In the case of the unsubstituted pyridine com-
plexes 1a and 1b, this process was complete after just 2 h of
incubation, with only the protein-bound species hsA-1 and
hsA-2 observed with subsequent incubation. A similar be-
havior was observed for complexes 2a and 2b, except that it
took up to 6 h of incubation until non-coordinate species
were no longer observed. This demonstrates that the methyl
group at the 4’-position of the pyridine rings significantly en-
hances these interactions with the protein. This effect is
even more pronounced in the complexes where the pyridine
ligands have been functionalized with phenyl-based groups.
In the case of complexes 3a and 3b significant contributions
from non-coordinated species were observed after 6 h of in-
cubation, and even after 24 h for the pyridinium-compensat-
ed compound. The protein binding of complexes 4a and 4b
was very similar to the mono-phenyl derivatives, demon-
strating that addition of the second phenyl ring does not sig-
nificantly enhance non-coordinate interactions. Specific in-
teractions with the hydrophobic binding regions, that is, the
Sudlowꢁs sites I and II of hsA,[19] have previously been sug-
gested as targets for interactions with Keppler-type com-
plexes.[14a,18c] Although increasingly hydrophobic groups
would certainly be expected to enhance interactions with
these protein sites, such as observed in the trend in the bind-
ing of complexes 1a–3a and 1b–3b, steric interactions, such
as from the bulky dibenzyl groups of complexes 4a and 4b
are likely also to be an important factor.
with hsA. Each compound (200 mm) was incubated with the
protein (100 mm) at 378C in PBS and monitored for up to
2 h; 4% DMSO were added to solutions of complexes 3a–
5a and 3b–5b to aid in solubility. The time dependence of
the UV/Vis measurements is shown in Figures S37–S46 in
the Supporting Information for all of the compounds.
During the first 2 h of incubation, the spectra do not differ
significantly from those observed in the absence of the pro-
tein, indicating similar species in solution. This is consistent
with the EPR observations of the parent compounds of 1a–
5a and 1b–5b and their aquated derivatives interacting with
hsA through non-coordinate interactions, and may also re-
flect a fraction of complexes that remain free in solution.
Additional samples were prepared of each complex
(600 mm) with hsA (100 mm) to monitor for a weak d–d tran-
sition at high wavelengths associated with coordinate hsA
interactions.[43] Because the EPR experiments demonstrated
that such interactions only dominate after prolonged incuba-
tion, the samples were measured for up to 24 h at 378C. For
each complex, broadening of the spectra and the appearance
of a low intensity band at l=625–657 nm was observed
(Figure 7 for complex 1a, 24-hour spectra for all complexes
are shown in Figures S47–S56 in the Supporting Informa-
tion). A similar feature at l=575 nm has been reported fol-
lowing incubation of KP1019 with hsA[18a] and is also ob-
served for KP1019 under the conditions used in our experi-
ments (see Figure S60 in the Supporting Information). This
absorbance has been proposed to arise from a d–d transition
that is specific for hsA coordination to a surface histidine.[17]
The assignment of this transition for complexes 1a–5a and
1b–5b and the rate that its intensity increases with incuba-
tion are in agreement with the EPR results indicating for-
mation of coordinate interactions of the complexes with
hsA.
Interestingly, the most stabilized non-coordinate interac-
tions are provided by the nitrobenzyl-functionalized com-
pounds 5a and 5b. Even after 24 h of incubation, significant
signals from non-coordinated species were observed, and
the rate of coordination to hsA was lower than any of the
other compounds. Spectral simulation determined that the
non-coordinate signal for complex 5a was the largest, ac-
counting for roughly 33% of the observed EPR intensity
(Figure 6). Comparison with complexes 3a and 3b is partic-
ularly relevant because it demonstrates the importance of
The observation of ligand-dependent stabilization of non-
coordinate interactions with hsA has important implications
for drug design. The coordination of KP1019 to hsA has
17038
www.chemeurj.org
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2013, 19, 17031 – 17042

Increasing the Bioavailability of RuIII Anticancer Complexes
FULL PAPER
Figure 7. UV/Vis spectra for complex 1a incubated with hsA for 24 h at
378C.
been suggested to potentially impact on its observed anti-
cancer activity.[11c] As we have demonstrated here, enhance-
ment of non-coordinate interactions with hydrophobic li-
gands, possibly with the addition of polarized groups, can in-
hibit the formation of coordinated species. This design strat-
egy has the potential to increase the concentration of the
active species in vivo.
As we have shown, non-coordinate interactions with hsA
also enhance the solubility of these complexes. This suggests
that preparation of the compounds with hsA prior to intra-
venous administration could help to stabilize them in solu-
tion while also increasing the concentrations of active spe-
cies. In essence, such an approach can be considered as the
application of non-coordinate protein conjugates. To test the
viability of this approach, the anticancer activity of the new
compounds described here was evaluated with and without
prior incubation with hsA.
&
Figure 8. MTS assay results for complexes 2b (top) and 4b (bottom).
=
^
free complex in solution and =complex pre-incubated with hsA.
tion) had IC50 values of approximately 400 mm, whereas the
other pyridine-based complexes 1a, 1b and 3b–5b were in-
active at their maximum concentrations (Figures 8 and S63
in the Supporting Information). Pre-incubation of the com-
pounds with hsA had notable effects on the activity of the
methyl- (i.e., complexes 2a and 2b) and dibenzyl-substituted
(i.e., complex 4b) complexes. In the case of complexes 2a
and 2b a decrease in activity was observed following incuba-
tion with hsA (Figures 8 and S63 in the Supporting Informa-
tion). This observation is consistent with coordination of the
complexes to the protein, which potentially reduces their
bioavailability. By contrast, pre-incubation of complex 4b
with hsA resulted in an increased activity. In fact, this prep-
aration provided the most activity of any of the tests with an
IC50 value of approximately 150 mm (Figure 8). This is partic-
ularly striking given the inactivity of the free complex. The
observed increase in activity is possibly related to the stabi-
lized hydrophobic interactions with hsA observed for com-
plex 4b, which may enhance the bioavailability of the com-
plex through inhibition of protein coordination.
The low activity of the pyridine-based compounds in this
report is consistent with previous in vitro cell studies of
Keppler-type complexes.[8a,22a,44] However, several studies
have shown that the low intrinsic cytotoxicity predicted in
such studies frequently underestimates the in vivo activity of
these types of compounds, which have shown excellent anti-
tumor properties in several studies.[7a,8c,45] A key example of
this is the sodium-compensated bisindazole complex KP1339
(Figure 1), which is currently undergoing clinical trials,[8e,10]
Biological activity: Anticancer activity screening was per-
formed against the SW480 human colorectal cancer cell line.
This cell line was chosen to allow for comparison with previ-
ous studies of KP1019 and KP1339.[8a,d,11a] In these studies,
the maximum concentrations used in testing of the free
complexes varied due to differences in their aqueous solubil-
ity. This issue was significant enough to prevent meaningful
testing of the pyridinium-compensated complexes 3a–5a.
However, all the remaining compounds, that is, complexes
1a and 2a as well as 1b–5b were subjected to in vitro stud-
ies. To probe the effect of hsA interactions on the activity of
the compounds, two sets of experiments were performed:
1) the compounds were dissolved in cell culture medium
with 1% DMSO and then incubated with the SW480 cells
for 24 h, and 2) the compounds were dissolved in PBS with
1% DMSO and incubated with hsA for one hour at room
temperature, after which the protein-bound fractions were
isolated and tested, as described in the Experimental Sec-
tion.
Screening of the free complexes revealed that they were
relatively inactive against the SW480 cell line. Complex 2b
(Figure 8) and 2a (Figure S63 in the Supporting Informa-
Chem. Eur. J. 2013, 19, 17031 – 17042
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
17039

C. J. Walsby et al.
and yet only shows an IC50 of 120 mm against the SW480 cell
line.[8a] This is similar to the best result for our compounds,
which was an IC50 of approximately 150 mm for compound
4b after incubation with hsA. We note also that the activity
of complex 4b with hsA is higher than that of KP1019 with
or without incubation with hsA under the testing conditions
used here (Figure S63 in the Supporting Information). With
such low intrinsic cytotoxicity, the in vivo activity[8a,d,18b] and
clinical success of KP1019 and KP1339 are due in part to
the very low levels of side effects they exhibit.[2b,8b] This
allows for high concentrations of the compounds to be used
during treatments.[2a,b]
It should be emphasized that the design of compounds
1a–5a and 1b–5b was primarily to target protein interac-
tions for comparative studies, rather than activity specifical-
ly. The use of non-coordinate hsA conjugates has been ex-
plored with conventional organic anticancer compounds, in
some cases resulting in improved activity.[46] However, this is
the first such study for RuIII-based bisazole complexes.
Indeed, despite successful clinical trials for both KP1019
and KP1339, relatively few derivatives of these compounds
have been reported.[7a,21,22,47]
Another useful consequence of incubating complexes 1a–
5a and 1b–5b with hsA was the improved solubility. This is
a particularly important observation given the low intrinsic
cytotoxicity of Keppler-type complexes described above. In
the studies reported here, the compounds were dissolved in
culture media and then added to an equal volume of cells
and media, which reduced the testing concentration by
a factor of two. Thus, for the highest concentration solutions
used in testing, that is, 500 mm, an initial solution with a con-
centration of 1 mm was required. At this concentration, pre-
cipitation was significant for compounds 3b–5b, which could
only be prepared in concentrations of half this value or less.
However, if the complexes were first incubated with hsA
they all could be prepared at the maximum concentration
without precipitation. As shown in Figure 8, incubation of
compound 4b with hsA increases the maximum testing con-
centration from 250 mm to at least 500 mm. This suggests an
advantage in preparing these types of compounds with hsA
for clinical studies by preventing precipitation and allowing
for higher concentrations for treatment solutions.
active species prior to treatment. Although DMSO has tra-
ditionally been used for this purpose, as we have shown
here, this can potentially influence ligand-exchange process-
es by coordination of DMSO through replacement of an
azole ligand. Such processes are expected to reduce the cy-
totoxicity of the compounds by producing species analogous
to NAMI-A-type complexes.
Because the origin of the anticancer activity of Keppler-
type complexes remains unknown, the effect of modifica-
tions to their axial azole ligands continues to be hard to ra-
tionalize. In this work we have been able to show how these
ligands can influence interactions with hsA, and how this
correlates with the hydrophobic properties of their substitu-
ents. Although the predicted stabilization of non-coordinate
protein interactions generally correlates with the hydropho-
bic properties of the ligands, closer analysis indicates that
other factors are also important. The observation that the
dibenzyl complex does not provide the anticipated increase
in non-coordinate interactions over the benzyl-substituted
compound suggests that steric factors should also be consid-
ered. Furthermore, the remarkable stability of non-coordi-
nate interactions with the nitrobenzyl compounds suggests
that polarizable substituents could increase the interactions
with the binding sites, possibly through hydrogen bonding.
Generally, the pyridine-based analogues of KP1019/
KP1339 were relative inactive against the SW480 cell line.
However, results from complexes 2a, 2b, and 4a do provide
some insight into the effects of hsA because the former
show a decrease in activity in the presence of the protein,
whereas the later shows an increase. This correlates with our
determination of a greater proportion of coordinate interac-
tions of complexes 2a and 2b with hsA and more non-coor-
dinate interactions for complex 4b. In the case of compound
4b this suggests greater bioavailability, and is particularly
notable because it is inactive as the free complex, but after
pre-treatment with hsA exhibits activity similar to the clini-
cally successful compound KP1339.[8e,10] Overall, these re-
sults suggest that targeting of non-coordinate interactions
with hsA could be a promising direction in the on-going de-
velopment of Keppler-type RuIII-based anticancer com-
plexes.
Experimental Section
Conclusion
Materials: The starting compounds RuCl3·H2O (Aldrich), pyridine (Ana-
chemia), 4-methylpyridine (TCI America), 4-phenylpyridine (Aldrich),
diphenyl-4-pyridyl-methane (Aldrich), and 4-(4-nitrobenzyl)pyridine
(Alfa Aesar) as well as human serum albumin (Aldrich) were used as
purchased.
For Keppler-type RuIII-based anticancer complexes, two key
aspects of their prodrug behavior are aqueous ligand ex-
change, and serum–protein interactions. This can lead to
a variety of species in solution, which may inhibit or pro-
mote their activity. As shown here, hydrophobic interactions
with albumin can potentially enhance the concentration of
active species in vitro. At the most fundamental level, this is
from increased solubilization of the complexes in aqueous
solution through inhibition of precipitation. This suggests
that the preparation of these complexes for intravenous
therapy with hsA could be advantageous for stabilizing the
Crystallographic structure determination: Single-crystal X-ray crystallo-
graphic analysis was performed on a Bruker SMART diffractometer
equipped with an APEX II CCD area detector fixed at a distance of
6.0 cm from the crystal and
a MoKa fine focus sealed tube (l=
0.71073 nm) operating at 1.5 kW (50 kV, 30 mA) and filtered with
a graphite monochromator. Structures were solved by using direct meth-
ods (SIR92) and refined by least-squares procedures in CRYSTALS.[48]
Diagrams of complexes 1a, 2a, 3b, 4a, and 5a were generated by
ORTEP-3 for Windows (v. 2.00).[49] Crystal data, data collection parame-
17040
www.chemeurj.org
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2013, 19, 17031 – 17042

Increasing the Bioavailability of RuIII Anticancer Complexes
FULL PAPER
ters, and details of structure refinement for compounds 1a, 2a, 3b, 4a,
and 5a are listed in Table S3 in the Supporting Information.
were prepared in cell culture media with 1% DMSO, at the maximum
concentration possible for each complex: 1 mm for complexes 1a, 1b, 2a,
and 2b, 500 mm for complexes 3b and 4b, and 750 mm for complex 5b.
Due to their low aqueous solubility, complexes 3a, 4a, and 5a were not
tested. Cells were incubated with each complex for 24 h, after which the
absorbance reading at l=490 nm was measured as an indicator of activi-
ty.
CCDC-968766 (1a), CCDC-968767 (2a), CCDC-968768 (3b), CCDC-
968769 (4a) and CCDC-968770 (5a) contain the supplementary crystallo-
graphic 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.
Additional samples were prepared where the complexes were loaded
onto hsA. Solutions of each complex in PBS (1 mm, 1% DMSO) were
mixed with solutions of hsA in PBS (0.5 mm) and incubated for 10 min at
378C, after which the protein-bound fractions were isolated, as described
above for the preparation of the EPR samples. The fractions were then
incubated at room temperature for one hour prior to being diluted in cell
culture medium and subsequently tested for activity by using identical
conditions as for the free complexes.
Optical measurements: UV/Vis spectra were measured by using a Cary
1E UV-Visible spectrophotometer, connected to a Haake F3 water bath,
which maintained the temperature of each sample at 378C. Spectra were
collected from samples dissolved in a PBS solution containing: NaCl
(137 mm), KCl (2.7 mm), Na2HPO4 (10 mm), and KH2PO4 (2 mm),
pH 7.4, with complexes 3a–5a and 3b–5b having 10% DMSO added to
aid in solubility. Measurements were performed by using 200 mm solutions
of each complex in 1 mL volumes. Protein binding measurements were
performed by using 200 mm solutions of each complex and 100 mm hsA in
1 mL volumes, with 4% DMSO added to the solutions of complexes 3a–
5a and 3b–5b. All samples were measured at 378C for a total of 2 h,
with scans taken at 10 min intervals, at a scan rate of 10 nmsꢀ1. Addition-
al protein samples were prepared to monitor for a weak d–d transition
corresponding to hsA coordination by using a 600 mm concentration of
each complex and a 100 mm concentration of hsA. These samples were in-
cubated at 378C for 24 h with one hour scan intervals.
Acknowledgements
The authors are grateful for the financial support of The Natural Sciences
and Engineering Research Council of Canada (NSERC): C.J.W. for
NSERC Discovery Grant funding, and M.I.W. for an NSERC postgradu-
ate doctoral scholarship (PGS-D). The Canada Foundation for Innova-
tion (CFI), and Simon Fraser University are also acknowledged.
Preparation of EPR samples
Complexes in buffer: Compounds were dissolved in PBS with complexes
3a–5a and 3b–5b having 10% DMSO added to aid in solubility, to give
a concentration of 3 mm, and incubated at 378C for 0, 30, and 120 min.
Each sample was promptly mixed with 30% by volume of glycerol,
which acted as a glassing agent, and frozen in liquid nitrogen until use.
[1] a) A. Bergamo, G. Sava, Dalton Trans. 2011, 40, 7817–7823; b) A.
Bergamo, C. Gaiddon, J. H. M. Schellens, J. H. Beijnen, G. Sava, J.
Inorg. Biochem. 2012, 106, 90–99.
Complexes with hsA: A solution of hsA (600 mL, 0.75 mm) in PBS was
mixed with a 600 mL solution of each complex (1.5 mm), also in PBS with
complexes 3a–5a and 3b–5b having 4% DMSO added to aid in solubili-
ty. The combined solution was then diluted to 4 mL with PBS and incu-
bated at 378C for one of the following time periods: 0, 30 min, 1, 2, 6,
and 24 h. Each 4 mL solution was concentrated down to a volume of less
than 200 mL by using an Amicon centrifugal filter unit (molecular weight
cut-off 30 kDa) by centrifuging at 88C and 4500 rpm for 30 min, or until
a volume of less than 200 mL was attained. The resulting filtered product
was then mixed with 90 mL of glycerol, diluted to a final volume of
300 mL with PBS, transferred to an EPR tube, and immediately frozen in
liquid nitrogen.
[2] a) C. G. Hartinger, M. A. Jakupec, S. Zorbas-Seifried, M. Groessl,
A. Egger, W. Berger, H. Zorbas, P. J. Dyson, B. K. Keppler, Chem.
Biodiversity 2008, 5, 2140–2155; b) F. Lentz, A. Drescher, A. Linda-
uer, M. Henke, R. A. Hilger, C. G. Hartinger, M. E. Scheulen, C.
Dittrich, B. K. Keppler, U. Jaehde, Anti-Cancer Drugs 2009, 20, 97–
103; c) R. E. Aird, J. Cummings, A. A. Ritchie, M. Muir, R. E.
Morris, H. Chen, P. J. Sadler, D. I. Jodrell, Br. J. Cancer 2002, 86,
1652–1657; d) A. Bergamo, A. Masi, P. J. Dyson, G. Sava, Int. J.
Oncol. 2008, 33, 1281–1289; e) J. M. Rademaker-Lakhai, D. van den
Bongard, D. Pluim, J. H. Beijnen, J. H. M. Schellens, Clin. Cancer.
Res. 2004, 10, 3717–3727.
EPR measurements and simulations: EPR measurements were per-
formed at X-band (9.3–9.4 GHz) by using a Bruker EMXplus spectrome-
ter with a PremiumX microwave bridge and HS resonator. Low-tempera-
ture (20 K) experiments used a Bruker ER 4112HV helium temperature-
control system and continuous-flow cryostat. Solution conditions and
spectroscopic parameters were unchanged for each experiment so that
the intensities of the EPR signals from RuIII-based species in different
samples could be compared. The Bruker cryostat system contains
a quartz-insert tube holder that enables reproducible sample placement
within the EPR resonator. Consequently, variation in instrument sensitiv-
ity between measurements was minimal, and automatic tuning of the
spectrometer gave a Q-factor of (7000ꢁ10)%. In experiments with hsA,
the distinctive EPR signal from a minor FeIII human serum transferrin
impurity at g=4.3 also provided a reference for normalizing the RuIII
EPR signal sensitivity.
[3] a) J. Reedijk, Platinum Met. Rev. 2008, 52, 2–11; b) S. Page Educ.
Chem. 2012, 49, 26–29.
[4] C. S. Allardyce, P. J. Dyson, Platinum Met. Rev. 2001, 45, 62–69.
[5] a) M. A. Jakupec, M. Galanski, V. B. Arion, C. G. Hartinger, B. K.
Keppler, Dalton Trans. 2008, 183–194; b) E. Alessio, G. Mestroni,
A. Bergamo, G. Sava, Curr. Top. Med. Chem. 2004, 4, 1525–1535;
c) G. Gasser, I. Ott, N. Metzler-Nolte, J. Med. Chem. 2011, 54, 3–25.
[6] A. Levina, A. Mitra, P. A. Lay, Metallomics 2009, 1, 458–470.
[7] a) K.-G. Lipponer, E. Vogel, B. K. Keppler, Met.-Based Drugs 1996,
3, 243–260; b) W. Peti, T. Pieper, M. Sommer, B. K. Keppler, G.
Giester, Eur. J. Inorg. Chem. 1999, 1551–1555.
[8] a) S. Kapitza, M. Pongratz, M. A. Jakupec, P. Heffeter, W. Berger, L.
Lackinger, B. K. Keppler, B. Marian, J. Cancer Res. Clin. Oncol.
2005, 131, 101–110; b) C. G. Hartinger, S. Zorbas-Seifried, M. A. Ja-
kupec, B. Kynast, H. Zorbas, B. K. Keppler, J. Inorg. Biochem. 2006,
100, 891–904; c) M. H. Seelig, M. R. Berger, B. K. Keppler, D.
Schmahl in Metal Ions in Biology and Medicine, French & European
Publications, New York, 1990, pp. 476–478; d) S. Kapitza, M. A. Ja-
kupec, M. Uhl, B. K. Keppler, B. Marian, Cancer Lett. 2005, 226,
115–121; e) M. M. Henke, H. Richly, A. Drescher, M. Grubert, D.
Alex, D. Thyssen, U. Jaehde, M. E. Scheulen, R. A. Hilger, Int. J.
Clin. Pharmacol. Ther. 2009, 47, 58–60.
All spectra were simulated by using the program WinEPR Simfonia,[50]
which efficiently produces accurate results for S=1= systems. A manual,
2
iterative fitting procedure was employed to analyze the overlapping spec-
tra from multiple RuIII species observed in many samples.
Anticancer activity testing: Complexes were tested for anticancer activity
by using a standard MTS (3-(4,5-Dimethylthiazol-2-yl)-5-(3-carboxyme-
thoxyphenyl)-2-(4-sulfophenyl)-2H-tetrazolium) assay (Promega). SW480
human colon carcinoma cells were obtained from the American Type
Culture Collection (ATCC). The cells were cultured by using Dulbeccoꢁs
modified eagle medium (DMEM), supplemented with 10% fetal bovine
serum (FBS), and l-glutamine. Fresh stock solutions of each free complex
[9] J. B. Aitken, S. Antony, C. M. Weekley, B. Lai, L. Spiccia, H. H.
Harris, Metallomics 2012, 4, 1051–1056.
[10] R. Trondl, P. Heffeter, M. A. Jakupec, W. Berger, B. K. Keppler,
BMC Pharmacol. Toxicol. 2012, 13, A82.
Chem. Eur. J. 2013, 19, 17031 – 17042
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
17041

C. J. Walsby et al.
[11] a) P. Heffeter, K. Bock, B. Atil, M. A. R. Hoda, W. Korner, C.
Bartel, U. Jungwirth, B. K. Keppler, M. Micksche, W. Berger, G.
Koellensperger, J. Biol. Inorg. Chem. 2010, 15, 737–748; b) A.
Kung, T. Pieper, R. Wissiack, E. Rosenberg, B. K. Keppler, J. Biol.
[27] B. Cebriꢃn-Losantos, E. Reisner, C. R. Kowol, A. Roller, S. Shova,
V. B. Arion, B. K. Keppler, Inorg. Chem. 2008, 47, 6513–6523.
[28] M. Bacac, A. C. G. Hotze, K. van der Schilden, J. G. Haasnoot, S.
Pacor, E. Alessio, G. Sava, J. Reedijk, J. Inorg. Biochem. 2004, 98,
402–412.
´
Inorg. Chem. 2001, 6, 292–299; c) K. Polec-Pawlak, J. K. Abramski,
J. Ferenc, L. S. Foteeva, A. R. Timerbaev, B. K. Keppler, M. Jarosz,
J. Chromatogr. A 2008, 1192, 323–326; d) J. C. Chen, L. M. Chen,
S. Y. Liao, K. Zheng, L. N. Ji, Dalton Trans. 2007, 3507–3515; e) M.
Groessl, C. G. Hartinger, P. J. Dyson, B. K. Keppler, J. Inorg. Bio-
chem. 2008, 102, 1060–1065; f) M. Groessl, C. G. Hartinger, A.
Egger, B. K. Keppler, Met. Ions Biol. Med. 2006, 9, 111–116; g) A.
Levina, J. B. Aitken, Y. Y. Gwee, Z. J. Lim, M. Liu, A. M. Singharay,
P. F. Wong, P. A. Lay, Chem. Eur. J. 2013, 19, 3609–3619; h) A. A.
Hummer, P. Heffeter, W. Berger, M. Filipits, D. Batchelor, G. E.
Buchel, M. A. Jakupec, B. K. Keppler, A. Rompel, J. Med. Chem.
2013, 56, 1182–1196.
[29] M. I. Webb, Simon Fraser University (Burnaby), 2013.
[30] a) J. M. Santos, C. Cipriano, R. B. Faria, J. D. FigueroaVillar, Can. J.
Chem. 1997, 75, 890–898; b) J. M. Santos, A. L. B. Formiga, J. D.
Figueroa-Villar, J. Mol. Struct. 2002, 608, 143–149; c) J. M. Santos,
R. R. da Silva, A. L. B. Formiga, L. W. Tinoco, J. D. Figueroa-Villar,
Chem. Phys. 2004, 306, 143–151.
[31] T. J. Siek, R. J. Osiewicz, J. Forensic Sci. 1975, 20, 18–37.
[32] P. T. Manoharan, P. K. Mehrotra, M. M. Taquikha, R. K. Andal,
Inorg. Chem. 1973, 12, 2753–2757.
[33] K. Krogh-Jespersen, H. J. Schugar, Inorg. Chem. 1984, 23, 4390–
4393.
[12] C. Anderson, A. L. Beauchamp, Can. J. Chem. 1995, 73, 471–482.
[13] a) G. Sava, A. Bergamo, S. Zorzet, B. Gava, C. Casarsa, M. Coc-
chietto, A. Furlani, V. Scarcia, B. Serli, E. Iengo, E. Alessio, G. Mes-
troni, Eur. J. Cancer 2002, 38, 427–435; b) M. Groessl, E. Reisner,
C. G. Hartinger, R. Eichinger, O. Semenova, A. R. Timerbaev,
M. A. Jakupec, V. B. Arion, B. K. Keppler, J. Med. Chem. 2007, 50,
2185–2193; c) T. Pieper, W. Peti, B. K. Keppler, Met.-Based Drugs
2000, 7, 225–232.
[14] a) F. Kratz, M. Hartmann, B. Keppler, L. Messori, J. Biol. Chem.
1994, 269, 2581–2588; b) N. Cetinbas, M. I. Webb, J. A. Dubland,
C. J. Walsby, J. Biol. Inorg. Chem. 2010, 15, 131–145.
[15] A. R. Timerbaev, C. G. Hartinger, S. S. Aleksenko, B. K. Keppler,
Chem. Rev. 2006, 106, 2224–2248.
[16] C. G. Hartinger, S. Hann, G. Koellensperger, M. Sulyok, M. Groessl,
A. R. Timerbaev, A. V. Rudnev, G. Stingeder, B. K. Keppler, Int. J.
Clin. Pharmacol. Ther. 2005, 43, 583–585.
[17] F. Piccioli, S. Sabatini, L. Messori, P. Orioli, C. G. Hartinger, B. K.
Keppler, J. Inorg. Biochem. 2004, 98, 1135–1142.
[18] a) L. Trynda-Lemiesz, A. Karaczyn, B. K. Keppler, H. Kozlowski, J.
Inorg. Biochem. 2000, 78, 341–346; b) A. K. Bytzek, K. Boeck, G.
Hermann, S. Hann, B. K. Keppler, C. G. Hartinger, G. Koellensperg-
er, Metallomics 2011, 3, 1049–1055; c) O. Dçmçtçr, C. G. Hartinger,
A. K. Bytzek, T. Kiss, B. K. Keppler, E. A. Enyedy, J. Biol. Inorg.
Chem. 2013, 18, 9–17.
[19] a) G. Sudlow, D. J. Birkett, D. N. Wade, Mol. Pharmacol. 1976, 12,
1052–1061; b) G. Sudlow, D. J. Birkett, D. N. Wade, Mol. Pharma-
col. 1975, 11, 824–832.
[34] a) S. Wanwimolruk, D. J. Birkett, P. M. Brooks, Mol. Pharmacol.
1983, 24, 458–463; b) G. J. Quinlan, G. S. Martin, T. W. Evans, Hep-
atology 2005, 41, 1211–1219; c) G. Colmenarejo, Med. Res. Rev.
2003, 23, 275–301.
[35] a) J. Szpunar, A. Makarov, T. Pieper, B. K. Keppler, R. Lobinski,
Anal. Chim. Acta 1999, 387, 135–144; b) K. Polec-Pawlak, J. K.
Abramski, O. Semenova, C. G. Hartinger, A. R. Timerbaev, B. K.
Keppler, M. Jarosz, Electrophoresis 2006, 27, 1128–1135.
[36] a) J. Ghuman, P. A. Zunszain, I. Petitpas, A. A. Bhattacharya, M.
Otagiri, S. Curry, J. Mol. Biol. 2005, 353, 38–52; b) I. Sjoholm, B.
Ekman, A. Kober, I. Ljungstedtpahlman, B. Seiving, T. Sjodin, Mol.
Pharmacol. 1979, 16, 767–777.
[37] a) F. Kratz, B. Elsadek, J. Controlled Release 2012, 161, 429–445;
b) B. Elsadek, F. Kratz, J. Controlled Release 2012, 157, 4–28.
[38] a) F. Kratz, J. Controlled Release 2008, 132, 171–183; b) A. Wunder,
U. Muller-Ladner, E. H. K. Stelzer, J. Funk, E. Neumann, G. Stehle,
T. Pap, H. Sinn, S. Gay, C. Fiehn, J. Immunol. 2003, 170, 4793–4801;
c) E. A. Murphy, B. K. Majeti, R. Mukthavaram, L. M. Acevedo,
L. A. Barnes, D. A. Cheresh, Mol. Cancer Ther. 2011, 10, 972–982.
[39] L. Kersten, H. Braunlich, B. K. Keppler, C. Gliesing, M. Wendelin,
J. Westphal, J. Appl. Toxicol. 1998, 18, 93–101.
[40] C. A. Smith, A. J. SutherlandSmith, B. K. Keppler, F. Kratz, E. N.
Baker, J. Biol. Inorg. Chem. 1996, 1, 424–431.
[41] J. R. Winkler, H. B. Gray, Chem. Rev. 1992, 92, 369–379.
[42] Y. Song, M. R. Gunner, J. Mol. Biol. 2009, 387, 840–856.
[43] a) M. Brindell, I. Stawoska, J. Supel, A. Skoczowski, G. Stochel, R.
van Eldik, J. Biol. Inorg. Chem. 2008, 13, 909–918; b) L. Messori, P.
Orioli, D. Vullo, E. Alessio, E. Iengo, Eur. J. Biochem. 2000, 267,
1206–1213.
[20] L. Trynda-Lemiesz, B. K. Keppler, H. Kozlowski, J. Inorg. Biochem.
1999, 73, 123–128.
[21] A. A. Batista, S. A. Onofre, S. L. Queiroz, G. Oliva, M. R. M.
Fontes, O. R. Nascimento, J. Braz. Chem. Soc. 1997, 8, 641–647.
[22] a) P. Mura, M. Camalli, L. Messori, F. Piccioli, P. Zanello, M. Corsi-
ni, Inorg. Chem. 2004, 43, 3863–3870; b) M. A. Jakupec, E. Reisner,
A. Eichinger, M. Pongratz, V. B. Arion, M. Galanski, C. G. Harting-
er, B. K. Keppler, J. Med. Chem. 2005, 48, 2831–2837; c) E. Reisner,
V. B. Arion, A. Eichinger, N. Kandler, G. Giester, A. J. L. Pombeiro,
B. K. Keppler, Inorg. Chem. 2005, 44, 6704–6716; d) E. Reisner,
V. B. Arion, M. Fatima, C. G. da Silva, R. Lichtenecker, A. Eiching-
er, B. K. Keppler, V. Y. Kukushkin, A. J. L. Pombeiro, Inorg. Chem.
2004, 43, 7083–7093.
[44] P. Heffeter, M. Pongratz, E. Steiner, P. Chiba, M. A. Jakupec, L. El-
bling, B. Marian, W. Korner, F. Sevelda, M. Micksche, B. K. Keppler,
W. Berger, J. Pharmacol. Exp. Ther. 2004, 312, 281–289.
[45] a) B. K. Keppler, W. Rupp, J. Cancer Res. Clin. Oncol. 1986, 111,
166–168; b) M. R. Berger, F. T. Garzon, B. K. Keppler, D.
Schmaehl, Anticancer Res. 1989, 9, 761–765.
[46] G. C. V. Tuan Giam Chuang, U. Kragh-Hansen, M. Otagiri, Pharm.
Res. 2002, 19, 569–577.
[47] a) A. Galeano, M. R. Berger, B. K. Keppler, Arzneim.-Forsch. 1992,
42–1, 821–824; b) C. Anderson, Can. J. Chem. 2001, 79, 1477–1482;
c) P. Mura, F. Piccioli, C. Gabbiani, M. Camalli, L. Messori, Inorg.
Chem. 2005, 44, 4897–4899; d) A. Nikolova, D. Ivanov, R. Buyu-
kliev, S. Konstantinov, M. Karaivanova, Arzneim.-Forsch. 2001, 51,
758–762.
[48] P. W. Betteridge, J. R. Carruthers, R. I. Cooper, K. Prout, D. J.
Watkin, J. Appl. Crystallogr. 2003, 36, 1487.
[49] L. J. Farrugia, J. Appl. Crystallogr. 1997, 30, 565.
[23] B. K. Keppler, W. Rupp, U. M. Juhl, H. Endres, R. Niebl, W. Balzer,
Inorg. Chem. 1987, 26, 4366–4370.
[24] O. M. Ni Dhubhghaill, W. R. Hagen, B. K. Keppler, K. G. Lipponer,
P. J. Sadler, J. Chem. Soc. Dalton Trans. 1994, 3305–3310.
[25] a) M. I. Webb, R. A. Chard, Y. M. Al-Jobory, M. R. Jones, E. W. Y.
Wong, C. J. Walsby, Inorg. Chem. 2012, 51, 954–966; b) M. I. Webb,
C. J. Walsby, Dalton Trans. 2011, 40, 1322–1331.
[26] J. A. Weil, J. R. Bolton, Electron Paramagnetic Resonance: Elemen-
tary Theory and Practical Applications, 2nd ed., Wiley-Interscience,
New York, 2007.
[50] WINEPR SimFonia, V. 1.26, Bruker Analytik GmbH, Copyright
1994–1997.
Received: July 9, 2013
Published online: November 7, 2013
17042
www.chemeurj.org
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2013, 19, 17031 – 17042