Cytotoxicity of η6-areneruthenium-based molecules to glioblastoma cells and their recognition by multidrug ABC transporters

Article abstract of DOI:10.1016/j.ejmech.2018.02.026

A new series of amphiphilic η⁶-areneruthenium(II) compounds containing phenylazo ligands (group I: compounds 1a, 1b, 2a and 2b) and phenyloxadiazole ligands (group II: compounds 3a, 3b, 4a and 4b) were synthesized and characterized for their anti-glioblastoma activity. The effects of the amphiphilic η⁶-areneruthenium(II) complexes on the viability of three human glioblastoma cell lines, U251, U87MG and T98G, were evaluated. The azo-derivative ruthenium complexes (group I) showed high cytotoxicity to all cell lines, whilst most oxadiazole-derivative complexes (group II) were less cytotoxic, except for compound 4a. The cationic complexes 2a, 2b and 4b were more cytotoxic than the neutral complexes. Compounds 2a and 2b caused a significant reduction in the percentage of cells in the G0/G1 phase, with concomitant increases in the G2/M phase and fragmented DNA in the T98G cell line. The η⁶-areneruthenium(II) compounds were also tested in cell lines that overexpress the multidrug ABC transporters P-gp, MRP1 and ABCG2. Compounds 2b and 4a were substrates for the P-gp protein, with resistance indexes of 8.6 and 1.9, respectively. Compound 2b was also a substrate for ABCG2 and MRP1 proteins, with lower resistance indexes (1.8 and 1.6, respectively). The contribution of multidrug ABC transporters to the cytotoxicity of compound 2b in T98G cells was evidenced, since verapamil (a characteristic inhibitor of MRP1) increased the cytotoxicity of compound 2b at concentrations up to 20 μmol L^?1, whilst GF120918 and Ko143 (specific inhibitors of P-gp and ABCG2, respectively) had no significant effect. In addition, we showed that compound 2b interacts with glutathione (GSH), which could explain its cellular efflux by MRP1. Our results showed that the amphiphilic η⁶-areneruthenium(II) complexes are promising anti-glioblastoma compounds, especially compound 2b, which was cytotoxic for all three cell lines, although it is transported by the three main multidrug ABC transporters.

Full text of DOI:10.1016/j.ejmech.2018.02.026

Accepted Manuscript
6
Cytotoxicity of η -areneruthenium-based molecules to glioblastoma cells and their
recognition by multidrug ABC transporters
Jaqueline Pazinato, Otávio M. Cruz, Karine P. Naidek, Amanda R.A. Pires, Eduard
Westphal, Hugo Gallardo, Hélène Baubichon-Cortay, Maria E.M. Rocha, Glaucia R.
Martinez, Sheila M.B. Winnischofer, Attilio Di Pietro, Herbert Winnischofer
PII:
S0223-5234(18)30154-5
10.1016/j.ejmech.2018.02.026
EJMECH 10203
DOI:
Reference:
To appear in: European Journal of Medicinal Chemistry
Received Date: 8 November 2017
Revised Date: 6 February 2018
Accepted Date: 9 February 2018
Please cite this article as: J. Pazinato, Otá.M. Cruz, K.P. Naidek, A.R.A. Pires, E. Westphal, H.
Gallardo, Héè. Baubichon-Cortay, M.E.M. Rocha, G.R. Martinez, S.M.B. Winnischofer, A. Di Pietro,
6
H. Winnischofer, Cytotoxicity of η -areneruthenium-based molecules to glioblastoma cells and their
recognition by multidrug ABC transporters, European Journal of Medicinal Chemistry (2018), doi:
10.1016/j.ejmech.2018.02.026.
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to
our customers we are providing this early version of the manuscript. The manuscript will undergo
copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please
note that during the production process errors may be discovered which could affect the content, and all
legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT

ACCEPTED MANUSCRIPT
6
Title: Cytotoxicity of
η
-areneruthenium-based molecules to glioblastoma cells and their
recognition by multidrug ABC transporters
1
,5
2
1
#
2,5
Jaqueline Pazinato , Otávio M. Cruz , Karine P. Naidek , Amanda R. A. Pires , Eduard
3
4
5
2
Westphal , Hugo Gallardo , Hélène Baubichon-Cortay , Maria E. M. Rocha , Glaucia R.
2
2
5
1
Martinez , Sheila M. B. Winnischofer *, Attilio Di Pietro and Herbert Winnischofer *
1
Department of Chemistry, Universidade Federal do Paraná, Curitiba, Paraná, 81531-980,
Brazil
2
Department of Biochemistry and Molecular Biology, Universidade Federal do Paraná,
Curitiba, Paraná, Brazil
3
Department of Chemistry and Biology, Universidade Tecnológica Federal do Paraná,
Curitiba, Paraná, 81280-340, Brazil
4
Department of Chemistry, Universidade Federal de Santa Catarina, Florianópolis, Santa
Catarina, 88040-900, Brazil
5
Equipe Labellisée Ligue 2014, Institut de Biologie et Chimie des Protéines, BMSSI UMR
5
086, CNRS/Université Lyon 1, Lyon, France
*
Corresponding
authors.
(H
sheilambw@ufpr.br (SMB Winnischofer)
Tel.
+55-41-3361-3181;
Tel.
E-mail.
hwin@ufpr.br;
E-mail.
hwin.ufpr@gmail.com
Winnischofer);
+55-41-3361-1673;
#
Present address: Universidade do Estado de Santa Catarina, Centro de Ciências
Tecnológicas, Joinville, Santa Catarina, 89219-710, Brazil

ACCEPTED MANUSCRIPT
ABSTRACT
6
A new series of amphiphilic η -areneruthenium(II) compounds containing phenylazo ligands
(group I: compounds 1a, 1b, 2a and 2b) and phenyloxadiazole ligands (group II: compounds
3
a, 3b, 4a and 4b) were synthesized and characterized for their anti-glioblastoma activity.
6
The effects of the amphiphilic η -areneruthenium(II) complexes on the viability of three
human glioblastoma cell lines, U251, U87MG and T98G, were evaluated. The azo-derivative
ruthenium complexes (group I) showed high cytotoxicity to all cell lines, whilst most
oxadiazole-derivative complexes (group II) were less cytotoxic, except for compound 4a. The
cationic complexes 2a, 2b and 4b were more cytotoxic than the neutral complexes.
Compounds 2a and 2b caused a significant reduction in the percentage of cells in the G0/G1
phase, with concomitant increases in the G2/M phase and fragmented DNA in the T98G cell
6
line. The η -areneruthenium(II) compounds were also tested in cell lines that overexpress the
multidrug ABC transporters P-gp, MRP1 and ABCG2. Compounds 2b and 4a were substrates
for the P-gp protein, with resistance indexes of 8.6 and 1.9, respectively. Compound 2b was
also a substrate for ABCG2 and MRP1 proteins, with lower resistance indexes (1.8 and 1.6,
respectively). The contribution of multidrug ABC transporters to the cytotoxicity of
compound 2b in T98G cells was evidenced, since verapamil (a characteristic inhibitor of
-1
MRP1) increased the cytotoxicity of compound 2b at concentrations up to 20 µmol·L , whilst
GF120918 and Ko143 (specific inhibitors of P-gp and ABCG2, respectively) had no
significant effect. In addition, we showed that compound 2b interacts with glutathione (GSH),
6
which could explain its cellular efflux by MRP1. Our results showed that the amphiphilic η -
areneruthenium(II) complexes are promising anti-glioblastoma compounds, especially

ACCEPTED MANUSCRIPT
compound 2b, which was cytotoxic for all three cell lines, although it is transported by the
three main multidrug ABC transporters.
KEYWORDS : areneruthenium, ABC transporters, multidrug resistance, glioblastoma.
1
. Introduction
Glioblastoma (GBM) is one of the most aggressive forms of human astrocytoma; the
median survival of patients diagnosed with GBM is 12–15 months, with approximately 5% of
patients surviving for 5 years after diagnosis [1, 2]. The aggressiveness of glioma cells can be
attributed to their high rates of proliferation, migration and invasion [3]. In spite of advances
in various treatment regimens, such as surgical resection, external beam irradiation and
chemotherapy, glioma treatments remain a significant challenge to oncologists [4].
Chemotherapy is an integral component of GBM treatment, and the major impediments to
delivery of therapeutic agents to the brain include a severe toxicity, the selective permeability
of the blood–brain barrier (BBB), and the presence of efflux pumps that form an
insurmountable obstacle to the accumulation of drugs in the brain [5].
The multidrug resistance (MDR) phenomenon may be related to a number of factors,
the main one being the resistance to drug transport across the plasma membrane, caused by
increased expression of transporter proteins belonging to the ABC superfamily (ATP-binding
cassette)[6, 7]. These multidrug ABC transporters use the energy generated by ATP
hydrolysis to translocate a wide range of endogenous and exogenous substrates across the
membrane against a sharp concentration gradient [8-11].
Currently, 48 members of the ABC transporter superfamily have been identified in
humans [12-14], and three of them are involved in MDR, ABCB1 (MDR1/P-pg), ABCC1
(MRP1) and ABCG2 (BCRP). In glioma cells, overexpression of MDR1/P-glycoprotein (P-
gp) and the MDR-associated protein family (MRP1–MRP9) also contributes to the

ACCEPTED MANUSCRIPT
development of chemoresistance [15]. In addition, the function of the ABCG2 transporter was
described in both the BBB and stem-like glioma cells [16]. Some reports suggest that in 70%
of GBM cases, the intrinsic drug resistance is mostly regulated by the overexpression of
MRP1 [17, 18]. Therefore, in order to improve the therapeutic effect on malignant glioma, it
is essential to overcome the obstacles of the BBB and MDR.
The standard therapy for GBM includes temozolomide (TMZ) [19], an
imidazotetrazine derivative, which exerts its cytotoxic effects by methylating specific DNA
6
sites. Methylation at the O position of guanine in DNA is generally considered the most
critical [20], resulting in DNA fragmentation, disrupting DNA replication, and causing tumor
suppression and apoptosis. Some preclinical studies have used TMZ together with cisplatin in
6
the treatment for GBM, due to the effect of cisplatin on reducing the activity of O -
methylguanine-DNA methyltransferase (MGMT) in vitro [21], suggesting a possible
synergistic interaction [22, 23].
6
Organometallic
η
-areneruthenium(II) derivatives have been intensively investigated
as potential antitumor agents [24-26], as alternatives to classical platinum complexes in
6
cancer treatments. The
η
-areneruthenium(II) complexes have been considered to show some
6
advantages over platinum: the amphiphilicity of the η -areneruthenium(II) portion due to the
hydrophobic moiety on the aromatic ring, balanced by the hydrophilic metal ion, probably
allows the biological compatibility and should allow easy uptake by the cells. Another
important feature is the hydrolysis of the Ru-Cl bonds, forming aquaruthenium species,
–
–
resulting from replacement of Cl by H O due to the low concentration of Cl in the
2
cytoplasm. The aqua complexes of ruthenium are more successive for substitution than the
chloride complexes [27, 28].
6
6
The prototype anticancer
η -areneruthenium(II) complex η -(p-MeC H Pri)Ru(P-
6 4
pta)Cl RAPTA-C was first investigated by Allardyce et al. [29]. This compound is slightly
2

ACCEPTED MANUSCRIPT
cytotoxic to tumor cells in vitro and generally free of toxicity to healthy cells, even after
prolonged exposure at millimolar concentrations. However, Rapta-C is very active in vivo; it
inhibits pulmonary metastases in CBA mice [28] and is also an antimetastatic agent [30-32].
6
Dougan et al. [33] reported the cytotoxic activity of η -areneruthenium(II) derivatives
containing phenylazo ligands with activity against the A2780 ovarian and A549 lung-cancer
cell lines. The authors attributed this crucial role in biological activity to the phenylazo ligand.
6
The present study investigated a series of eight amphiphilic η -areneruthenium(II)
compounds, the contribution of the further hydrophobicity of the long alkyl chain, as well as
the contribution of the phenylazo (group I) and phenyloxadiazole (group II) residues to the
cytotoxicity in human glioma cell lines. In addition, we investigated the ability of the
ruthenium complexes to be effluxed by three multidrug ABC transporters.
2
. Results and discussion
2
.1. Chemistry
6
The strategy adopted to synthesize the amphiphilic η -areneruthenium(II) compounds enabled
us to generate charged and neutral molecules with modulated hydrophobicity. Scheme 1
illustrates the strategy used to obtain these amphiphilic complexes.
Scheme 1
1
The ESI-MS, Elemental Analysis, H RMN and FTIR analyses gave consistent results for the
molecular structure and purity for the samples. The cationic compounds 2a, 2b, 4a and 4b
+
showed the molecular ion peak (M ) in the ESI-MS(+) spectrum, with a characteristic
isotopomeric distribution (Figures SI 1, 2, 3 and 4). On the other hand, the neutral compounds

ACCEPTED MANUSCRIPT
1
a, 1b, 3a and 3b did not show peaks with m/z coincident with the molecular mass of the
compound, but rather peaks consistent with fragments and the free ligand (Figures SI 5, 6, 7
and 8), revealing the instability of the complexes under the conditions of the experiment. In
1
the H NMR spectra obtained in CDCl the chemical shifts did not show significant changes
3
for the complexes compared to the free ligands, except for the resonance peaks assigned to the
H and H of the pyridine residue of the amphiphilic ligand. Table 1 shows selected values of
α
β
the chemical shifts for H and H of the complexes and free ligands (a complete description
α
β
6
for all δ and ∆δ values is shown in SI). For the η -areneruthenium(II) compounds, H signals
α
were assigned to the doublet at δ = 9.3 ppm, which is 0.4–0.6 ppm up-shifted in comparison
to the free ligands. The neutral compounds 1a,b and 3a,b exhibited the H signal
α
corresponding to two equivalent hydrogens, whereas the cationic compounds 2a,b and 4a,b
exhibited a resonance corresponding to one hydrogen only. The H for group I appeared as a
β
doublet at δ = 7.67 and 7.68 ppm for 1a and 1b, respectively, and at δ = 7.73 and 7.76 ppm
for 2a and 2b, respectively. The H for group II appeared as a doublet at δ = 7.99 ppm for 3a
β
and 3b, and at δ = 7.72 and 7.75 ppm for 4a and 4b, respectively. For the neutral compounds,
the H was 0.01–0.05 ppm up-shifted in comparison to the free ligands, while for the cationic
β
compounds, this up-shift was 0.4 ppm. The benzene hydrogen signal was assigned to the
singlet corresponding to six hydrogens at δ = 5.7 and 6.0–6.1 ppm for neutral and cationic
complexes, respectively. The free benzene showed this resonance at δ =7.38 ppm and the
precursor Ar Ru Cl at δ = 5.95 ppm. The values of the chemical shift were affected by the
2
2
4
aromatic ring current effect, the ruthenium(II) bonded to the N-pyridyl, and by the neighbor
ligand chlorido through ruthenium bonding and through a space effect [34-36]. This last effect
showed a direct relationship to the chemical shifts of the pyridine H and benzene H, since the
α
resonances appeared up-shifted in compounds that have two chlorido ligands in comparison to

ACCEPTED MANUSCRIPT
compounds that have only one chlorido ligand. However, H is more sensitive to the
β
surroundings of the ruthenium(II) center and the azo or oxadiazo residues, in which the
resonance in the compounds ranges from δ = 7.67 to δ = 7.99 ppm depending on the adjacent
aromatic residue. These results provide evidence of the mono- or bidentated coordination of
the amphiphilic ligand to the compounds. FTIR results are shown in SI (Figures SI 17, 18 and
–
1
Tables SI 9 to 16). The spectra of compounds 1a ,2a, 3a and 4a showed bands at 3070 cm
assigned to the aromatic C-H bond, present in the benzene group; intense bands at 2920 and
–
1
2
850 cm assigned to aliphatic C-H stretching vibration, consistent with the long C chain;
12
–
1
~
1600–1410 cm assigned to C-N stretching vibrations present in the phenylazopyridine or
–
1
phenyloxadiazolepyridine residues; and ~1025 cm due to C-O-C stretching vibration.
Compounds 1b, 2b, 3b and 4b showed very similar spectra, except for the absence of the
aliphatic C-H stretching bands. The ionic compounds 2a,b and 4a,b showed an intense band
–
6
–1
associated with the νPF superimposed on the πCH vibration at 835 cm . Usually the band
–
1
6
frequencies up-shifted 4–10 cm from the η -areneruthenium(II) compounds to the free
6
ligands. Nevertheless the FTIR results confirmed the presence of the η -areneruthenium(II)
portion in addition to the amphiphilic ligand. Elemental Analysis indicated the high purity of
the samples, considering the ionic nature of the bidentated complexes.
Table 1
The azo- and oxadiazole- free ligands showed strong π-π* bands under UV, which
were red-shifted in the spectra of the ruthenium complexes. The azo-complexes 1a and 1b
showed the π-π* band at 380 nm (Figure SI 19A), whilst 2a and 2b showed this band at 452
nm (Figure SI 19B). The oxadiazole- complexes 3a and 3b showed the π-π* band at 333 nm
(Figure SI 19C). In these complexes, another π−π* band appeared as a shoulder with maxima

ACCEPTED MANUSCRIPT
below 240 nm. Exceptions were complexes 4a and 4b, which in addition to the π−π* band at
3
36 nm, showed another band at 270 nm (Figure SI 19D). The cationic bidentated complexes
which the π-π* red-shifted in comparison to the neutral monodentated complexes, probably
due to the additional N-binding which stabilizes the π orbitals.
The azo-derivative compounds showed the half-wave potential related to phenylazo
reduction at more positive potentials than the free ligands (Figure SI20 and Table 2). This
6
property increases the redox accessibility in biological media of the η -areneruthenium
complexes in comparison to the free organic ligands. The cyclic voltammograms of the group
I compounds (1a, 1b, 2a and 2b) are characterized by a poorly reversible reduction process at
–
0.8 V for compounds 1b and 1b (Figure SI21) and –0.2V for compounds 2a and 2b (Figure
) in acetonitrile (Fig. 1C, D). In dichloromethane, compound 2a shows the reduction
1
potential shifted to –0.15 V (Figure 1A). Table 2 shows peak potential values for group I
compounds. Group II compounds did not undergo any redox process in the potential window
analyzed. The redox process observed in the group I cyclic voltammograms is attributed to
–
-
the 1 e reduction of the phenylazo residue to the phenylazo radical anion ({-N=N-} •)[37].
For compounds 2a and 2b, this reduction occurs at more-positive potentials than for
compounds 1a and 1b, which agrees with the bidentated structure and cationic nature of
compounds 2a and 2b. For compounds 2a and 2b, the product of this reduction is not stable
and is followed by a coupled chemical reaction/transformation[33, 38]. Therefore, the product
of this electrochemical/chemical reaction oxidizes at 0.1V, which is more evidenced in
acetonitrile (Figure 1C, D) than in dichloromethane (Figure 1A, B).
Figure 1
Table 2

ACCEPTED MANUSCRIPT
2
.2. Biological evaluation
6
The effect of the amphiphilic η -areneruthenium complexes on tumor cell viability was
assessed in three human GBM cell lines, U251, U87MG and T98G (Table 3). The glioma
6
cells were treated with the η -areneruthenium complexes for 72 h and the cytotoxicity was
measured by the MTT method. Table 2 shows the IG values (concentrations producing 50%
5
0
growth inhibition) for the different compounds. The group I compounds (azo complexes)
showed cytotoxic activity on all these cell lines. Compounds 2a and 2b were the most
cytotoxic. Most compounds of group II (oxadiazole complexes) were less cytotoxic, except
for compound 4a. From the results summarized in Table 2, some tendencies could be traced:
i) The phenylazo group played a special role in the cytotoxicity, since group I compounds
were all cytotoxic toward the glioblastoma cell lines, while only compound 4a in group II
showed high cytotoxicity. ii) The more-hydrophobic compounds, i.e., those containing the C12
chain, had a tendency to be more cytotoxic than the compounds containing a methoxy group,
except for the effect of compounds 1a and 1b on T98G cells. Increased hydrophobicity of the
6
ligand probably facilitated cellular uptake. iii) The cationic η -areneruthenium(II) complexes
were more cytotoxic than the neutral complexes, as shown by the IG values for 2a, 2b and
5
0
4
a compounds. Probably the increased amphiphilicity provided by the charged species
balanced by the hydrophobic chain played a role in the cell uptake of these compounds.
6
Palmucci et al. [39] described a series of neutral and cationic acylpyrazolonate η -
areneruthenium(II) complexes with aliphatic groups in the acyl moiety, which display
cytotoxicity in vitro against human ovarian carcinoma cells (A2780 and its resistant
counterpart A2780cisR, a variant cell line with acquired resistance to cisplatin). The authors
reported that complexes were highly cytotoxic to both A2780 and A2780cisR cells, and the
cationic complexes bearing a long hydrophobic chain were the most cytotoxic, as seemed also
to occur with our compounds.

ACCEPTED MANUSCRIPT
Table 3
To evaluate whether the effects of the compounds were attributable to changes in the cell-
cycle progression, we conducted DNA assays, using propidium-iodide staining and flow
cytometry analysis, in permeabilized glioblastoma cells (Figure 2A and 2B). T98G cells
–
1
treated with compounds 2a and 2b at 25 µmol·L showed a significant reduction in the
percentage of cells in the G0/G1 phase, with a concomitant increase in the G2/M phase; no
change in the progression cycle was detected for the other compounds (Figure 2A). These
alterations of the cell-cycle progression were intensified by treatment with compounds 2a and
–
1
2
b at 100 µmol·L (Figure 2B). At this higher concentration, compounds 1a, 1b and 4a also
caused reductions of the percentage of cells in the G0/G1 cell-cycle phase. In addition, the
content of fragmented DNA in T98G cells increased after treatment with the compounds
(Figure 2C). Treatment of T98G cells with compound 2a resulted in an increase in the content
–
1
of fragmented DNA at the lower concentration of 25 µmol·L . At the highest concentration
–
1
(
100 µmol·L ), compounds 1a, 1b, 2a, 2b and 4a also caused significant increases in
fragmented DNA (Figure 1C). Taken together, the results showed that the phenylazo
ruthenium compounds (group I) could alter the cell-cycle progression, with a concomitant
increase in the proportion of hypodiploid cells, a hallmark of the late stage of apoptosis.
These alterations were also observed for compound 4a, the most cytotoxic compound in
group II.
Figure 2
Some cell mechanisms can be associated with glioma chemoresistance, including
overexpression of multidrug ABC transporters, which promotes the efflux of antitumor drugs

ACCEPTED MANUSCRIPT
from the cell, thus decreasing their therapeutic effects. To elucidate whether ABC transporters
6
played a role in the differential cytotoxic effect of these η -areneruthenium(II) compounds, we
investigated whether these compounds were specifically toxic to cells overexpressing ABC
transporters such as P-gp, MRP1, and ABCG2. Essentially, the compounds that were
transported showed higher IG50 values for resistant cells (ABC transporter-overexpressing
cells) than for the parental cell lines. The resistance index (RI) was calculated to determine the
degree of acquired resistance of each cell line to its selective compound. RI was calculated
from the following equation: RI = IG resistant cell/IG parental cell. An RI value > 1.0
50
50
indicates the ability of the particular ABC transporter to confer resistance to a compound.
6
For the η -areneruthenium(II) compounds belonging to group I, compounds 1a, 1b and 2a
were not substrates for the multidrug ABC transporters studied here (Table 4). In addition, the
group II compounds 3a, 3b and 4b were neither substrates for the multidrug ABC transporters
(
Table 4) nor highly cytotoxic (IG > 100) for all cell lines analyzed (healthy and tumor cell
50
lines) (Table 3). However, compound 2a showed high cytotoxic activity both for the parental
cell lines (NIH 3T3, HEK 293 and BHK21) and for their resistant counterparts (Table 4). This
compound was the most cytotoxic compound for the glioblastoma cell lines (Table 3). In
contrast, compounds 2b and 4a were substrates for P-gp, with resistance indexes of 8.6 and
1
.9 respectively. Compound 2b was also a substrate for ABCG2, with a resistance index of
1
.8 (Tables 4 and 5). Both compounds are cationic, which agrees with the previous
observation that P-gp and ABCG2 promote a phenotype of resistance to cationic compounds
40]. Despite their efflux, compounds 2b and 4a retained cytotoxicity for all the glioblastoma
[
cell lines analyzed (Table 3). Compound 2b was also a substrate for MRP1, with a resistance
index of 1.6 (Tables 4 and 5). The dynamics of transport of compound 2b is shown in Figure
3
.

ACCEPTED MANUSCRIPT
Table 4
Table 5
Figure 3
To prove that compound 2b indeed acts as a substrate for the three ABC transporters, we
performed additional experiments by using selective inhibitors for P-gp (GF120918/elacridar)
and ABCG2 (Ko143) and a specific cell line expressing an inactive mutant MRP1 (BHK21-
MRP1m). The cytotoxic properties of compound 2b were evaluated both in the parental cell
lines (NIH 3T3, HEK 293 and BHK21) and in their resistant counterparts (NIH-P-gp, HEK-
ABCG2 and BHK-MRP1m). As shown in Figure 3A, B and C, the use of the selective
inhibitors or the inactive mutant cell line abolished the resistance phenotype (RI index almost
=
1).
It is well established that MRP1 is constitutively overexpressed in cells derived from
astrocytomas and glioblastomas [41]. Inhibition of MRP1 had a significant effect on the
response of GBM cells to chemotherapeutics in both primary and recurrent GBM cells [42].
In the glioma cell model, the T98G cell line was reported to express high levels of P-gp and
MRPs [15]. Here, the cytotoxicity analyses showed different IG values among the different
5
0
–
1
cells for compound 2b, with T98G cells displaying the highest IG50 (37.13 µmol·L ) in
–
comparison to the other glioblastoma cell lines analyzed (U87MG cells: IG = 4.90 µmol·L
5
0
1
–1
;
U251 cells: IG = 8.03 µmol·L ). This result can be explained by the high expression of
50
MRP1 mRNA in T98G cells, whereas the U251 and U87MG cells had a low expression [15].
It has been established that the MRP1 protein transports only anionic compounds. Therefore,
it is likely that the cationic compound 2b needs to be converted into a conjugated anion, to
enable its efflux by MRP1. One possible explanation is its interaction with glutathione (GSH),
since MRP1 [43, 44] is often associated with high levels of GSH, which is the most-abundant

ACCEPTED MANUSCRIPT
antioxidant species present in the cell. As it is constituted by a tripeptide (γ-Glu-Cys-Gly),
GSH is negatively charged at physiological pH. GSH is often involved in detoxification
processes through conjugated species [29]. Cole et al. [44] reported the role of MRP1 in
transporting GSH and GSH conjugates with drugs.
In the electrochemical characterization described above, we observed the reduction process
–
attributed to the 1 e reduction of the phenylazo residue to the phenylazo radical anion ({-
-
N=N-} •) at E = –0.2 V for compounds 2a and 2b. Once the reduction-potential value is
cp
close to the oxidation potential of GSH to disulfide-glutathione (GSSH), compounds 2a or 2b
could react directly with GSH, promoting its oxidation. The ratio of the amounts of
GSSG/2GSH mainly determines the redox state of a cell, since it determines the redox
potential of the GSSG/2GSH pair. According to Schafer et al. [45] the ratio of GSSG/2GSH
correlates with the biological state of the cell; cells in the proliferation stage contain a large
amount of GSH and E ~ –240 mV; whereas, differentiating cells have an E ~ –200 mV; and
decreases in GSH with increases in the GSSG/2GSH ratio (E ~ –170 mV) induce apoptosis.
Therefore, compounds 2a and 2b should be able to cause an imbalance in the redox state of
the cell and initiate a set of cell signals, leading to apoptosis. This could explain the
cytotoxicity of these compounds.
6
Dougan et al. [33] reported that azopyridine-based η -areneruthenium(II) complexes have the
ability to catalyze GSH oxidation, through a mechanism that involves an intermediate
resulting from N=N bond reduction, followed by glutathione disulfide formation. To obtain
more insight, we investigated the interaction between compound 2b and GSH by UV-vis
–
1
spectroscopy. Figure 4 shows a 10 mmol·L GSH solution spectrum (black line) that shows a
strong band absorption at 290 nm. After 10 s, the solution of GSH with compound 2b showed
the blue line spectrum illustrated in Figure 4, where, in addition to the GSH band at 290 nm,
the compound 2b π−π* band appears at 440 nm. After 1 h (Figure 4, red line), the 290-nm

ACCEPTED MANUSCRIPT
GSH band vanished and the absorption below 250 nm increased. These changes are similar to
what occurred when GSH was treated with one equivalent of hydrogen peroxide (Figure 4,
dashed black line). The 440 nm π−π* band was up-shifted to 410 nm, revealing that 2b was
modified in the solution, possibly through changes in the coordination sphere of the
ruthenium center. Probably the ruthenium complex undergoes solvolysis and some of the
ligands are substituted. However, the 410-nm maximum is not consistent with the free
azopyridine ligand, indicating that this amphiphilic ligand is still bonded to the ruthenium
complex. Some of the species produced during the catalytic process of GSH oxidation should
be effluxed, which could explain why 2b is a substrate of MRP1. After 48 h, the solution of
GSH and compound 2b produced a spectrum similar to that observed after 1 h, but with a
further increased band below 250 nm (data not shown). These results are consistent with the
catalytic oxidation of GSH by compound 2b.
Figure 4
To assess the functional role of multidrug ABC transporters in the cytotoxicity of compound
2
b in glioma models, T98G cells were treated with increasing concentrations of compound 2b
in the absence or in the presence of modulators of P-gp (GF120918), MRP1 (verapamil) and
ABCG2 (MBLII 141 and Ko143). We found that verapamil further increased the cytotoxicity
–
1
of compound 2b. On the other hand, MBLII 141 showed a small effect at 2.5 µmol·L ,
whereas GF120918 and Ko143 produced no significant effect on the cytotoxicity of
compound 2b (Figure 5).
Despite the cationic nature of compound 2b, we did not observe significant changes in its
cytotoxicity to these cell lines in the presence of the selective inhibitor GF120918. Again, this
result can be explained by conversion of 2b into anionic species by interaction with GSH,

ACCEPTED MANUSCRIPT
since T98G cells are rich in GSH; T98G cells contain 450 nmol of GSH per mg of protein, a
larger amount than in other cell lines such as BHK-21 or HeLa cells [46].
Figure 5
Summarizing, we demonstrated that compounds belonging to group I were all cytotoxic to
glioma cell lines, whereas only compound 4a from group II exhibited high cytotoxicity. The
more-hydrophobic compounds containing long-alkyl-chain ligands tended to be more
6
cytotoxic than compounds having a short methoxy group. The cationic η -areneruthenium(II)
compounds were more cytotoxic than the neutral compounds. This cytotoxicity was
attributable to alterations of cell-cycle progression and increased contents of fragmented
DNA, suggesting an apoptotic mechanism. Two compounds were substrates for the multidrug
ABC transporters: compound 2b (for P-gp, ABCG2 and MRP1) and 4a (for P-gp).
Specifically, for compound 2b, in addition to being a substrate for the three transporters
studied here, this compound was also cytotoxic for the glioblastoma cell lines. We showed
that compound 2b interacts with GSH. This constitutes a possible explanation of its efflux by
MRP1 and its cytotoxic effect, which was further increased in the presence of verapamil in
the T98G cells. In contrast, compound 2a showed high cytotoxicity to the glioblastoma cell
lines, and it was not a substrate for the transporters studied here. In conclusion, we showed
that compounds 2a and 2b are promising for glioblastoma treatments. Further investigations
are required to better understand their mechanism in the tumor cells.
EXPERIMENTAL SECTION
Characterization of the compounds

ACCEPTED MANUSCRIPT
All the reagents and solvents were of analytical grade and were purchased from commercial
sources (Merck, Sigma-Aldrich, Fluka, and Acros) and used as received. A Perkin-Elmer
2
400 Series II CHN Elemental Analyzer was used to characterize the composition of the
1
samples. H NMR spectra were recorded at 298K using a Varian Mercury Plus spectrometer
operating at 400 MHz, or a Bruker DPX 200 MHz spectrometer operating at 200 MHz.
Deuterated chloroform (CDCl ) or deuterated DMSO (DMSO-d ) was used as the solvent and
3
6
TMS as the internal standard. The mass spectra (ESI-MS) were obtained in a Thermo Fisher
Scientific LTQ XL Linear Ion Trap Mass Spectrometer, using sample solutions at a
–2
–1
concentration of 10 mg mL in dichloromethane. High-Resolution Mass Spectra were
recorded in a Bruker micrOTOF-Q II mass spectrometer, using an APPI ionization source. A
Bruker Vertex 70 spectrophotometer was used to acquire FTIR spectra of the compounds
–1
embedded in KBr pellets, in the range of 400 to 4000 cm . An Agilent 8453 or Shimadzu UV
401 PC spectrophotometer was used to acquire the UV-vis spectra of the compounds in
chloroform solutions. For GSH studies, the UV-vis spectrum of compound 2b at 100 µM with
2
–
1
1
0 mmol·L GSH in DMSO added, was monitored over 48 h. The GSH was oxidized to
–
1
GSSH by addition of H O at 10 mmol·L . The electrochemical experiments were performed
2
2
+
on an Ivium Compactstat potentiostat/galvanostat in a three-electrode cell; Ag/Ag (0.01
–
1
mol·L in acetonitrile) reference electrode, platinum disc (diameter 3 mm) working electrode
–
1
and a stainless-steel counter electrode. Experiments were conducted in 0.1 mol·L TBAClO
4
electrolyte solution in dichloromethane or acetonitrile (depending on the solubility of the
compound).
Scheme 2

ACCEPTED MANUSCRIPT
4
-(4-dodecyloxyphenylazo)pyridine (L1a). The synthesis of this ligand was described in a
previous paper from our group [47].
2
-Hydrazinepyridine (6). A mixture of 2-chloropyridine 2.00 g (17.6 mmol) and 12 mL
hydrazine monohydrate (65%) was added to a 25-mL round-bottom flask and stirred under
reflux for 7 h. After this period, the solution was cooled and the product extracted with diethyl
ether (2 × 20 mL). The organic phase was dried over anhydrous Na SO , filtered, and stored
2
4
in a freezer. The resulting precipitate was collected by filtration and used in the next step
1
without further purification. m.p.: 41.8–44.3 °C (lit. 45 °C) [48]. H NMR (200 MHz, CDCl )
3
δ ppm: 3.81 (broad, 2H, -NH ), 6.12 (broad, 1H, -NH-), 6.67 (m, 2H, Py-H), 7.46 (ddd, J =
2
8
.5 Hz, 7.1 Hz and 1.8 Hz, 1H, Py-H), 8.10 (m, 1H, Py-H).
2
-(4-hydroxyphenylazo)pyridine (7). 1.27 g (11.8 mmol) hydroquinone, 8 mL sulfuric acid
and 100 mL distilled water were added to an Erlenmeyer flask. To this solution and under
strong stirring, 2-hydrazinepyridine (6) dissolved in 20 mL water was added dropwise, and
the reaction mixture was stirred at room temperature for a further 90 min. The pH was
adjusted to ≈ 6 using NaOH aqueous solution (20%), and the resulting precipitate was filtered
and washed with distilled water. The crude product was purified by recrystallization from
isopropanol/water, affording 1.95 g (83%) of the desired azo compound after the two-step
1
procedure. m.p.: 232 °C (decomposition). H NMR (400 MHz, DMSO-d ) δ ppm: 6.87 (d, J =
6
8
.6 Hz, 2H, Ar-H), 7.50 (m, 1H, Py-H), 7.65 (d, J = 8.2 Hz, 1H, Py-H), 7.87 (d, J = 8.6 Hz,
1
3
2
H, Ar-H), 7.99 (m, 1H, Py-H), 8.67 (d, J = 4.4 Hz, 1H, Py-H), 10.50 (broad, 1H, -OH). C
NMR (100.6 MHz, DMSO-d ) δ ppm: 112.86, 116.10, 124.93, 125.53, 138.67, 145.19,
6
1
49.18, 161.90, 163.13.

ACCEPTED MANUSCRIPT
5
-(4-dodecyloxyphenyl)tetrazole (9). This compounds was prepared by a two-step synthesis
according to a previously described procedure [49].
2
-(4-dodecyloxyphenylazo)pyridine (L2a). In a 100-mL round-bottom flask were added
.40 g (7.03 mmol) 2-(4-hydroxyphenylazo)pyridine (7), 1.84 g (7.38 mmol) 1-
1
bromododecane, 1.94 g (14.1 mmol) K CO and 40 mL butanone. The reaction mixture was
2
3
refluxed under stirring for 24 h, cooled to room temperature, filtered, and washed with
butanone. The crude product obtained after removal of the solvent was purified by column
chromatography using silica and CH Cl as eluent, affording 2.43 g (94%) of an orange solid.
2
2
1
m.p.: 79.0 – 79.9 °C. H NMR (400 MHz, CDCl ) δ ppm: 0.88 (t, J = 6.6 Hz, 3H, CH ), 1.21
3
3
–
2
1.41 (m, 16H, -CH -), 1.47 (m, 2H, -CH -), 1.82 (m, 2H, -CH CH O-), 4.05 (t, J = 6.6 Hz,
2 2 2 2
H, -CH O-), 7.01 (d, J = 9.0 Hz, 2H, Ar-H), 7.36 (m, 1H, Py-H), 7.79 (d, J = 8.2 Hz, 1H, Py-
2
1
3
H), 7.87 (m, 1H, Py-H), 8.05 (d, J = 9.0 Hz, 2H, Ar-H), 8.71 (d, J = 4.7 Hz, 1H, Py-H). C
NMR (100.6 MHz, CDCl ) δ ppm: 14.09, 22.65, 25.96, 29.11, 29.31, 29.33, 29.52, 29.55,
3
2
1
9.60, 29.62, 31.88, 68.40, 114.71, 115.17, 124.56, 125.69, 138.16, 146.66, 149.40, 162.68,
+
+
63.11. Q-TOF/MS for C H N O [M H] : Calculated: 368.2696; Found: 368.2690.
23
34
3
2
-(4-dodecyloxyphenyl)-5-(4-pyridyl)-1,3,4-oxadiazole (L3a). In a round-bottom flask
equipped with a condenser and a drying tube (CaCl ), 0.60 g (4.88 mmol) isonicotinic acid
2
(
8), 1 drop DMF and 5 mL SOCl were added and refluxed for 3 h. After this period, the
2
remaining unreacted SOCl was removed by vacuum distillation, and 1.77 g (5.36 mmol) 5-
2
(4-dodecyloxyphenyl)tetrazole (9) and 15 mL anhydrous pyridine were added to the round-
bottom flask and refluxed for a further 20 h. The solution was then cooled to room
temperature, poured into a 300-mL ice/water bath and basified to pH ≈ 12 with a NaOH
aqueous solution (10%). The precipitate was separated by filtration, washed with water, and

ACCEPTED MANUSCRIPT
recrystallized from isopropanol/H O solution, affording 1.75 g (88%) of a white solid. m.p:
2
1
9
4.8–95.4 °C. H NMR (400 MHz, CDCl ) δ ppm: 0.87 (t, J = 6.8 Hz, 3H, -CH ), 1.22-1.41
3
3
(
m, 16H, -CH -), 1.47 (m, 2H, -CH -), 1.82 (m, 2H, -CH CH O-), 4.03 (t, J = 6.5 Hz, 2H, -
2 2 2 2
OCH -), 7.02 (d, J = 8.8 Hz, 2H, Ar-H), 7.97 (d, J = 6.0 Hz, 2H, Py-H), 8.06 (d, J = 8.8Hz,
2
1
3
2
H, Ar-H), 8.82 (d, J = 6.0 Hz, 2H, Py-H). C NMR (100.6 MHz, CDCl ) δ ppm: 14.10,
3
2
2.66, 25.96, 29.07, 29.32, 29.34, 29.53, 29.56, 29.61, 29.63, 31.89, 68.33, 115.08, 115.42,
+
+
1
20.20, 128.91, 131.19, 150.77, 162.19, 162.39, 165.50. QTOF/MS for C H N O [M H] :
25 34 3 2
Calculated: 408.2646; Found: 408.2640.
2
-(4-dodecyloxyphenyl)-5-(2-pyridyl)-1,3,4-oxadiazole (L4a). Under argon atmosphere,
0
.80 g (6.50 mmol) picolinic acid (10), 2.15 g (6.50 mmol) 5-(4-dodecyloxyphenyl)tetrazole
(11), 0.08 g (0.65 mmol) DMAP and 40 mL anhydrous toluene were added to a 100-mL
Schlenk flask. After 10 min, 1.60 g (7.80 mmol) DCC was added, with an additional 10 mL
toluene. The mixture was stirred for 10 min at room temperature and for 6 h at 100 °C. The
reaction mixture was cooled to room temperature and 40 mL ethyl acetate was added. The
precipitate was filtered, washed with ethyl acetate, and the solvent of the combined organic
phases removed under reduced pressure. The crude product was purified by column
chromatography using silica and a mixture of CH Cl : ethyl acetate (gradient from 100:0 to
2
2
9
5:5) as eluent. A final recrystallization from heptane afforded 2.07 g (78%) of a white solid.
1
m.p.: 100.7–101.5 °C. H NMR (400 MHz, CDCl ) δ ppm: 0.87 (t, J = 6.8 Hz, 3H, -CH ),
3
3
1
.17–1.38 (m, 16H, -CH -), 1.46 (m, 2H, -CH -), 1.80 (m, 2H, -CH CH O-), 4.02 (t, J = 6.5
2 2 2 2
Hz, 2H, -OCH -), 7.00 (d, J = 9.0 Hz, 2H, Ar-H), 7.45 (ddd, J = 7.8 Hz, 4.9 Hz and 1.2 Hz,
2
1
H, Py-H), 7.88 (ddd, J = 7.8 Hz, 7.8 Hz and 1.8 Hz, 1H, Py-H), 8.13 (d, J = 9.0 Hz, 2H, Ar-
H), 8.29 (ddd, J = 7.8 Hz, 1.2 Hz and 1.0 Hz, 1H, Py-H), 8.80 (ddd, J = 4.9 Hz, 1.8 Hz and
1
3
1
.0 Hz, 1H, Py-H). C NMR (100.6 MHz, CDCl ) δ ppm: 14.10, 22.65, 25.96, 29.08, 29.32,
3

ACCEPTED MANUSCRIPT
2
1
4
9.52, 29.55, 29.60, 29.62, 31.88, 68.24, 114.89, 115.71, 123.06, 125.57, 129.05, 137.14,
+
+
43.75, 150.19, 162.18, 163.33, 165.58. Q-TOF/MS for C H N O [M H] : Calculated:
2
5
34
3
2
08.2646; Found: 408.2644.
Scheme 3
4
-(4-hydroxyphenylazo)pyridine (12). The synthesis of this compound was described in a
previous report, and can be found in reference [47].
4
-(4-methoxyphenylazo)pyridine (L1b). A mixture of 4-(4-hydroxyphenylazo)pyridine (12)
1
.50 g (7.53 mmol), iodomethane 0.56 mL (9.04 mmol), K CO 2.08 g (15.06 mmol) and 50
2
3
mL acetone was added to a round-bottom flask and stirred at 50 °C for 20 h. After this period,
the suspension was filtered and the solid was washed with hot methanol. The solvent was
removed under reduced pressure and the product was purified by column chromatography
using silica and a mixture of CH Cl :ethyl acetate (95:5) as eluent, affording 0.81 g (50%) of
2
2
1
an orange solid. m.p.: 96.2–97.8 °C. H NMR (200 MHz, CDCl ) δ ppm: 3.91 (s, 3H, -OCH ),
3
3
7
8
1
2
.03 (d, J = 9.3 Hz, 2H, Ar-H), 7.67 (d, J = 6.1 Hz, 2H, Py-H), 7.97 (d, J = 9.3 Hz, 2H, Ar-H),
1
3
.78 (d, J = 6.1 Hz, 2H, Py-H). C NMR (50.3 MHz, CDCl ) δppm: 55.64, 114.37, 116.12,
3
+
25.55, 146.80, 151.18, 157.37, 163.19. Q-TOF/MS for C H N O [M+H] : Calculated:
1
2
12
3
14.0975; Found: 214.0978.
2
-(4-hydroxyphenylazo)pyridine (7). The synthesis of this compound was described above
in scheme 2.
2
-(4-methoxyphenylazo)pyridine (L2b). In a round-bottom flask, 1.540 g (7.53 mmol) 2-(4-
hydroxyphenylazo)pyridine (7), 0.56 mL (9.04 mmol) iodomethane, 2.08 g (15.06 mmol)
K CO and 100 mL acetone were added. The mixture was heated to 50 °C and stirred for 20
2
3

ACCEPTED MANUSCRIPT
h. After this, the mixture was cooled to room temperature, filtered, and the solid washed with
CH Cl . The crude product obtained after removal of the solvent was purified by column
2
2
chromatography using silica and a mixture of CH Cl :ethyl acetate (95:5) as eluent, affording
2
2
1
0
.85 g (53%) of an orange solid. m.p.: 49.5–51.1 °C (lit. 51–52 °C) [50]. H NMR (200 MHz,
CDCl ) δ ppm: 3.91 (s, 3H, -OCH ), 7.03 (d, J = 9.3 Hz, 2H, Ar-H), 7.33–7.41 (m, 1H, Py-H),
3
3
1
3
7
.76–7.94 (m, 2H, Py-H), 8.07 (d, J = 9.3 Hz, 2H, Ar-H), 8.70–8.74 (m, 1H, Py-H). C NMR
(
50.3 MHz, CDCl ) δ ppm: 55.61, 114.29, 115.23, 124.64, 125.70, 138.21, 146.88, 149.44,
3
+
1
5
2
63.02, 163.11. Q-TOF/MS for C H N O [M+H] : Calculated: 214.0975; Found: 214.0976.
1
2
12
3
-(4-pyridyl)tetrazole (14). The synthesis of this compound is described in reference [51].
-(4-methoxyphenyl)-5-(4-pyridyl)-1,3,4-oxadiazole (L3b). In a round-bottom flask
equipped with a condenser and a drying tube (CaCl ), 1.48 g (9.75 mmol) p-anisic acid (15), 1
2
drop DMF and 5 mL SOCl were added, heated to 60 °C and stirred for 16 h. After this
2
period, the remaining SOCl was removed by vacuum distillation, resulting in the crude acid
2
chloride (16). To the round-bottom flask, 1.43 g (9.75 mmol) 5-(4-pyridyl)tetrazole (14) and
1
5 mL anhydrous pyridine were added and the mixture refluxed for a further 24 h. The
solution was then cooled to room temperature, poured into 150 mL ice/water, and the
precipitate separated by filtration. The crude product was purified by column chromatography
using silica and a gradient of CH Cl :ethyl acetate (from 100:0 to 90:10) as eluent. The
2
2
product was further recrystallized over ethanol, affording 1.95 g (79%) of a white solid. Mp:
1
1
92.3–192.8 °C (lit. 188 °C) [52]. H NMR (200 MHz, CDCl ) δ ppm: 3.90 (s, 3H, -OCH ),
3
3
7
.04 (d, J = 9.0 Hz, 2H, Ar-H), 7.97 (d, J = 6.1 Hz, 2H, Py-H), 8.09 (d, J = 9.0 Hz, 2H, Ar-H),
1
3
8
.83 (d, J = 6.1 Hz, 2H, Py-H). C NMR (50.3 MHz, CDCl ) δ ppm: 55.45, 114.57, 115.68,
3
+
1
20.13, 128.88, 131.06, 150.75, 162.20, 162.69, 165.35. Q-TOF/MS for C H N O [M+H] :
14
12
3
2
Calculated: 254.09240; Found: 254.09239.

ACCEPTED MANUSCRIPT
5
-(2-pyridyl)tetrazole (18). In a round-bottom flask equipped with a condenser, 5.00 g (48.1
mmol) 2-cyanopyridine (17), 9.37 g (144.1 mmol) NaN , 7.71 g (144.1 mmol) NH Cl and 20
3
4
mL DMF were added, heated to 130 °C, and stirred for 24 h. The suspension was cooled to
room temperature, poured into 200 mL water/ice, and the pH adjusted to 6 with HCl (10%).
The precipitate formed was separated by filtration and recrystallized over water, affording
1
5
.16 g (73%) of colorless crystals. m.p.: 214.4–216.0 °C (lit. 213 °C) [53]. H NMR (200
MHz, DMSO-d ) δ ppm: 7.55–7.64 (m, 1H, Pyr-H), 8.00–8.09 (m, 1H, Pyr-H), 8.17–8.23 (m,
6
1
3
1
1
H, Pyr-H), 8.72–8.81 (m, 1H, Pyr-H). C NMR (50.3 MHz, DMSO-d ) δ ppm: 122.63,
6
26.10, 138.23, 143.69, 150.06, 154.78
2
-(4-methoxyphenyl)-5-(2-pyridyl)-1,3,4-oxadiazole (L4b). In a round-bottom flask
equipped with a condenser and a drying tube (CaCl ), 1.50 g (9.86 mmol) p-anisic acid (15),
2
1
drop DMF and 5 mL SOCl were added, heated to 60 °C, and stirred for 16 h. After this
2
period, the remaining SOCl was removed by vacuum distillation, resulting in the crude acid
2
chloride (16). To the round-bottom flask, 1.45 g (9.86 mmol) 5-(2-pyridyl)tetrazole (18) and
1
5 mL anhydrous pyridine were added and the mixture refluxed for a further 24 h. The
solution was then cooled to room temperature, poured into 150 mL ice/water, and the product
extracted with ethyl acetate (3 × 50 mL). The solvent of the combined organic phases was
removed under reduced pressure and the product purified by column chromatography using
silica and CH Cl as eluent. The product was further recrystallized over ethanol, affording 1.54
3
1
g (62%) of a white solid. m.p.: 155.0–155.8 °C (lit. 154–156 °C) [54]. H NMR (200 MHz,
CDCl ) δ ppm: 3.89 (s, 3H, -OCH ), 7.03 (d, J = 9.0 Hz, 2H, Ar-H), 7.47 (ddd, J = 7.7 Hz, 4.9
3
3
Hz and 1.2 Hz, 1H, Py-H), 7.90 (ddd, J = 7.9 Hz, 7.7 Hz and 1.7 Hz, 1H, Py-H), 8.16 (d, J =
.0 Hz, 2H, Ar-H), 8.30 (ddd, J = 7.9 Hz, 1.2 Hz and 1.0 Hz, 1H, Py-H), 8.81 (ddd, J = 4.9
9
1
3
Hz, 1.7 Hz and 1.0 Hz, 1H, Py-H). C NMR (50.3 MHz, CDCl ) δ ppm: 55.42, 114.43,
3

ACCEPTED MANUSCRIPT
1
15.99, 123.05, 125.58, 129.06, 137.13, 143.69, 150.19, 162.53, 163.36, 165.48. Q-TOF/MS
+
+
for C H N O [M H] : Calculated: 254.0924; Found: 254.0928.
1
4
12
3
2
6
The precursor [{(η -C H )ClRu} Cl ] was synthesized following the procedure described in
6
6
2
2
the literature [55, 56].
Synthesis of compounds 1a, 1b, 3a and 3b (neutral complexes)
In a 100-mL round-bottom flask, 0.1 g (0.2 mmol) of the precursor [{(η -C H )ClRu} Cl ]
6
6
6
2
2
was mixed with 0.15 g (0.1 mmol) of the ligand in 20 mL THF. Next, the reaction mixture
was heated under reflux for 6 h. Then, the mixture was cooled to room temperature. After
that, 10 mL ethanol was added to precipitate the product, which was kept under refrigeration
for 12 h. Finally, the product was filtered and washed with ethanol and THF, and dried under
vacuum in a desiccator with anhydrous calcium chloride. For further purification, the
compounds were dissolved in dichloromethane with the minimum volume needed for total
solubilization, and then filtered through Celite. In this step, the excess of the ruthenium
precursor present in the mixture was separated in the Celite. The filtered solution was left to
allow partial evaporation of the dichloromethane and concentration of the mixture. The
mixture was filtered through Celite again to remove additional precipitated precursor. The
procedure was repeated once more. After that, the dichloromethane was totally evaporated
and the mixture suspended in THF and filtered in a sintered glass filter. In this step, the
product was separated from the solution containing the excess of the free ligand. The product
was washed with THF and dried under vacuum.
Compound 1a: Orange solid. Yield 17%. CHN: RuON Cl C H .H O experimental
3
2
29 39
2
1
(
(
(
calculated): %C: 55.6 (54.8), %H: 6.3 (6.5), and %N: 7.0 (6.6). Purity 99.94%. H NMR:
200 MHz, 30 °C, CDCl , TMS) δ ppm: 0.88 (3H, H ), 1.27 (16H, H ), 1.48 (2H, H ), 1.83
3
1
2
3
2H, H ), 4.08 (2H, H ), 5.70 (6H, H ), 7.03 (2H, H ), 7.67 (2H, H ), 7.97 (2H, H7, 7’),
4
5
10
6,6’
8,8’

ACCEPTED MANUSCRIPT
–1
9
.21 (2H, H ). IR (KBr) (ν cm ): 3066, 2924 and 2853 (ν ), 1599 (ν ), 1497 and 1412
9,9’ max CH CC
(
νCN), 1254 (νasCOC), 1146 (δ ), 1015 (ν
), 832 (π ) and 564 (π ).
CH
sCOC
CH
CC
Compound 1b: Yellow solid. Yield 21%. CHN: RuON Cl C H .H O experimental
3
2
18 17
2
1
(
(
(
calculated): %C: 45.6 (44.9), %H: 3.7 (3.9), and %N: 8.7 (8.7). Purity 99.94%. H NMR:
200 MHz, 30 °C, CDCl , TMS) δ ppm: 3.93 (3H, H ), 5.69 (6H, H ), 7.04 (2H, H ), 7.68
3
1
10
6,6’
–
1
2H, H ), 7.99 (2H, H ), 9.22 (2H, H ). IR (KBr) (ν cm ): 3066 (ν ), 1599 (ν ),
8
,8’
7,7’
9,9’
max
CH
CC
1
499 and 1410 (ν ), 1258 (ν
), 1144 (δ ), 1023 (ν
), 832 (π ) and 562 (π ).
sCOC CH CC
CN
asCOC
CH
Compound 3a: Yellow solid. Yield 35%. CHN: RuO N Cl C H .H O experimental
2
3
2
31 39
2
1
(
calculated): %C: 56.1 (55.1), %H: 6.0 (6.1), and %N: 6.4 (6.2). Purity 99.92%. H NMR:
(
200 MHz, 30 °C, CDCl , TMS) δ ppm: 0.88 (3H, H ), 1.27 (16H, H ), 1.48 (2H, H ), 1.83
3
1
2
3
(
2H, H ), 4.05 (2H, H ), 5.72 (6H, H ), 7.05 (2H, H ), 7.99 (2H, H ), 8.08 (2H, H7, 7’),
4 5 10 6,6’ 8,8’
–1
9
.32 (2H, H ). IR (KBr) (ν cm ): 3065, 2928 and 2853 (ν ), 1613 (ν ), 1495 and 1420
9,9’ max CH CC
(
νCN), 1258 (νasCOC), 1173 (δ ), 1030 (νsCOC), 843 (πCH).
CH
Compound 3b: Brown solid. Yield 4%. CHN: RuO N Cl C H .H O experimental
2
3
2
20 17
2
1
(
(
(
calculated): %C: 37.3 (46.0), %H: 2.7 (3.6), and %N: 7.2 (8.0). Purity 99.12%. H NMR:
200 MHz, 30 °C, CDCl , TMS) δ ppm: 3.92 (3H, H ), 5.72 (6H, H ), 7.07 (2H, H ), 7.81
3
1
10
6,6’
–
1
2H, H ), 8.09 (2H, H ), 9.32 (2H, H ). IR (KBr) (ν cm ): 3069 (ν ), 1613 (ν ),
8
,8’
7,7’
9,9’
max
CH
CC
1
495 and 1422 (ν ), 1262 (ν
), 1179 (δ ), 1025 (νsCOC), 840 (πCH).
asCOC CH
CN
Synthesis of compounds 2a, 2b, 4a and 4b (positive charged complexes)
6
In a 100-mL Schlenk flask, 0.1 g (0.2 mmol) of the precursor [{(η -C H )ClRu} Cl ] was
6
6
2
2
mixed in 20 mL methanol. Then, the ligand of interest (added in a stoichiometric amount, 2:1)

ACCEPTED MANUSCRIPT
was dissolved in 10 mL methanol and slowly added to the reaction mixture. The solution was
stirred for 4 h at room temperature and protected from light (with the exception of compound
4
a, which was stirred for 18 h at 120 °C. After that, the solvent was removed by
rotoevaporation to approximately 1 mL and NH PF in a methanol solution was added in
4
6
excess. The mixture was kept under refrigeration for 24 h. Finally, the product was filtered
and washed with methanol and ether, and dried under vacuum in a desiccator with anhydrous
calcium chloride. It was only necessary to purify compound 2b.
Compound 2a: Brown solid. Yield 79%. CHN RuON ClC H PF .H O experimental
3
29 39
6
2
1
(
(
(
(
calculated): %C: 45.5 (46.7), %H: 5.6 (5.5), and %N: 5.9 (5.6). Purity 99.95%. H NMR:
200 MHz, 30 °C, CDCl , TMS) δ ppm:0.89 (3H, H ), 1.28 (16H, H ), 1.48 (2H, H ) 1.87
3
1
2
3
2H, H ), 4.14 (2H, H ), 6.05 (6H, H ), 7.09 (2H, H ), 7.73 (1H, H ), 8.21 (2H, H7, 7’), 8.50
4
5
10
6,6’
9
–1
d, 1H, H and t, 1H, H ), 9.30 (1H, H ). IR (KBr) (ν cm ): 3109, 2924 and 2851 (ν ),
8
8’
9’
max
CH
1
597 (ν ), 1499 and 1409 (ν ), 1260 (ν
), 1150 (δ ), 1020 (ν
), 838 (ν ) and 560
CC
CN
asCOC
CH
sCOC
PF
+
+
(
π ). ESI/MS for M [RuON ClC H ] : Calculated: 582.18; found: m/z 582.34.
CC 3 29 39
Compound 2b: Brown solid. Yield 25%. CHN RuON ClC H PF .H O experimental
3
18 17
6
2
1
(
(
(
calculated): %C: 37.0 (36.6), %H: 3.2 (3.2), and %N: 6.9 (7.1). Purity 99.96%. H NMR:
200 MHz, 30 °C, CDCl , TMS) δ ppm: 3.81 (3H, H ), 6.07 (6H, H ), 7.56 (2H, H ), 7.76
3
1
10
6,6’
–
1
1H, H ), 8.19 (2H, H ), 8.54 (2H, H ), 8.92 (1H, H ). IR (KBr) (ν cm ): 3094 (ν ),
9
7,7’
8,8’
9’
max
CH
1
596 (ν ), 1466 and 1370 (ν ), 1262 (ν
), 1154 (δ ), 1027 (ν
), 832 (ν ) and 558
sCOC PF
CC
CN
asCOC
CH
+
+
(
π ). ESI/MS for M [ RuON ClC H ] : Calculated: 427.88; found: m/z 428.09.
CC 3 18 17
Compound 4a: Green solid. Yield 79%. CHN RuO N ClC H PF .H O experimental
2
3
31 39
6
2
1
(
calculated): %C: 45.3 (47.4), %H: 5.0 (5.3), and %N: 5.8 (5.3). Purity 99.87%. H NMR:

ACCEPTED MANUSCRIPT
(
200 MHz, 30 °C, CDCl3, TMS) δ ppm:0.89 (3H, H ), 1.28 (16H, H ), 1.48 (2H, H ), 1.87
1
2
3
(
2H, H ), 4.07 (2H, H ), 6.12 (6H, H ), 7.06 (2H, H ), 7.73 (1H, H ), 8.09 (4H, H7,7’,8,8’),
4 5 10 6,6’ 9
–1
9
.38 (1H, H ). IR (KBr) (ν cm ): 3094, 2923 and 2853 (ν ), 1609 (ν ), 1499 and 1438
9’ max CH CC
+
+
(
νCN), 1262 (νasCOC), 1179 (δ ), 1025 (ν
), 838 (ν ). ESI/MS for M [RuO N ClC H ] :
sCOC PF 2 3 31 39
CH
Calculated: 622.20; found: m/z 622.31.
Compound 4b: Green solid. Yield 77%. CHN RuO N ClC H PF .H O experimental
2
3
20 17
6
2
1
(
(
(
calculated): %C: 37.2 (38.0), %H: 3.1 (3.0), and %N: 6.8 (6.6). Purity 99.93%. H NMR:
200 MHz, 30 °C, CDCl , TMS) δ ppm: 3.94 (3H, H ), 6.15 (6H, H ), 7.09 (2H, H ), 7.75
3
1
10
6,6’
–
1
1H, H ), 8.11 (4H, H
), 9.35 (1H, H ). IR (KBr) (ν
cm ): 3094 (ν ), 1609 (ν ),
max CH CC
9
7,7,8,8’
9’
1
500 and 1420 (ν ), 1265 (ν
), 1181 (δ ), 1021 (ν
), 836 (ν ). ESI/MS for
sCOC PF
CN
asCOC
CH
+
+
M [RuO N ClC H ] : Calculated: 467.90; found: m/z 468.16.
2
3
20 17
Biology
Cell Cultures. The human embryonic kidney HEK293 cell lines transfected with ABCG2
HEK293 ABCG2), mouse embryonic fibroblast wild-type (NIH3T3) and P-gp-
(
overexpressing (NIH3T3 ABCB1), as well as the human glioblastoma T98G, U87MG and
U251 cell lines, were maintained in Dulbecco’s modified Eagle’s medium (DMEM high
glucose) (Sigma-Aldrich), supplemented with 10% fetal bovine serum (FBS), 1%
–
1
penicillin/streptomycin, and supplemented with 0.75 mg⋅mL G418 (HEK293 ABCG2), or 60
–
1
ng⋅mL colchicin (NIH3T3 P-gp). The mammalian fibroblast baby hamster kidney 21 cells
BHK21) and MRP1-overexpressing cells (BHK21 MRP1) were maintained in Dulbecco’s
modified Eagle’s medium (DMEM/F12), supplemented with 5% fetal bovine serum (FBS),
(
–
1
1
% penicillin/streptomycin, and supplemented for BHK21 MRP1 with 0.1 mg⋅mL
methotrexate.

ACCEPTED MANUSCRIPT
Cytotoxicity Assays. Different cell lines were seeded into 96-well culture plates at an
appropriate concentration to allow a log-linear growth for 3 d. Twenty-four hours after
–
1
plating, the cells were treated with increasing concentrations of 0.1 to 100 µmol·L
compounds for 72 h, at 37 °C under 5% CO . Cell viability was then evaluated with an MTT
2
colorimetric assay [57]. Control experiments were performed with DMEM high glucose
containing 0.1% DMSO (v/v). The results were expressed as percentage of viable cells versus
initial control cells, taken as 100%.
4
Flow cytometric analysis. Human tumor cells (6 × 10 cells/well) were incubated in 12-well
–
1
6
plates (24 h, 37 °C). The cells were then treated with 25 and 100 µmol·L η -
areneruthenium(II) compounds for 72 h. After centrifugation, the pellet was washed twice
.
–1
with PBSA and incubated with a cold staining solution of 50 ꢀg mL propidium iodide, 0.1%
–
1
Triton X-100, 0.1% sodium citrate, and 0.2 mg⋅mL RNAse A (30 min, in the dark) [58]. The
samples were run in a FACSCalibur System, and the DNA histograms obtained were
analyzed to measure their relative DNA content (characterizing G1, S, G2 phase distributions
and the proportion of sub-G1 hypodiploid cells), using Win MDI 2.9 software (developed by
Joseph Trotter).
Statistical analysis. Mean ± SD results were submitted to analysis of variance (one-way
ANOVA) using the GraphPad Prism 5 program. Values of p ≤ 0.05 were considered
statistically significant.
Acknowledgments

ACCEPTED MANUSCRIPT
The authors thank the Ligue Nationale Contre le Cancer (Equipe Labellisée 2014) for
financial support. Amanda Pires thanks CNPq and the Science Without Borders Program
2
014-3/ 2014-5. Jaqueline Pazinato thanks CAPES and the Overseas Sandwich Doctorate
Program 6738/14-9. The authors also thank Elizangela Cavazzini Cesca for the NMR
analyses and discussions, and CEBIME-UFSC for the HR-MS analysis.
REFERENCES
1.
M.E. Salacz, K.R. Watson, D.A. Schomas, Glioblastoma: Part I. Current state of
affairs, J. Missouri State Med. Assoc 108 (2011) 187–194.
.
2.
M. Jhanwar-Uniyal, et al., Glioblastoma: Molecular Pathways, Stem Cells and
Therapeutic Targets, Cancers 7 (2015) 538.
3.
F.B. Furnari, et al., Malignant astrocytic glioma: genetics, biology, and paths to
treatment, Genes Dev. 21 (2007) 2683–2710.
4.
E.G. Van Meir, et al., Exciting new advances in neuro-oncology: the avenue to a cure
for malignant glioma, CA Cancer J. Clin. 60 (2010) 166–193.
5.
A. Nowacek, H.E. Gendelman, NanoART, neuroAIDS and CNS drug delivery.
Nanomedicine 4 (2009) 557–574.
6.
G. Szakacs, et al., Targeting multidrug resistance in cancer, Nat. Rev. Drug Discovery
(2006) 219–234.
Z. Chen, et al., Mammalian drug efflux transporters of the ATP binding cassette
ABC) family in multidrug resistance: A review of the past decade, Cancer Lett. 370 (2016)
53–164.
5
7.
(
1