Combining [Arene-Ru] with azocarboxamide to generate a complex with cytotoxic properties

Article abstract of DOI:10.1002/chem.201404448

Azocarboxamide (azcH) has been combined for the first time with [Ru-Cym] to generate metal complexes with N,N- and N,O-coordination mode, [(Cym)Ru(azc)Cl] and [(Cym)Ru(azcH)Cl]⁺ [PF₆]^-. Geometric and electronic structures of the complexes are reported along with their in vitro activities against different tumour cell lines and preliminary results on solution chemistry. Compound [(Cym)Ru(azc)Cl] exhibited remarkable cytotoxic properties. It was cell-type specific and had comparable IC₅₀ values towards both cancer cells and their drug-resistant subline. A tenfold increase in the sensitivity towards [(Cym)Ru(azc)Cl] was noted for the tumour cells with depleted intracellular glutathione (GSH) level, suggesting the essential role of GSH in cell response to this compound.

Full text of DOI:10.1002/chem.201404448

DOI: 10.1002/chem.201404448
Communication
&
Antitumor Agents
Combining [Arene–Ru] with Azocarboxamide to Generate
a Complex with Cytotoxic Properties
Michael G. Sommer,[a] Petra Kureljak,[b] Damijana Urankar,[c] David Schweinfurth,[a]
[b]
[a]
[c]
[b]
[a]
´
Nikolina Stojanovic, Martina Bubrin, Martin Gazvoda, Maja Osmak,* Biprajit Sarkar,*
[c]
ˇ
and Janez Kosmrlj*
properties of the azo group.[7] In addition, the presence of the
amide nitrogen also should make electron removal (after de-
protonation) from these ligands easy. Furthermore, the pres-
ence of three different donor atoms in these ligands makes
possible the formation of two different five-membered chelat-
ing rings on metal coordination (N,N- and N,O-)[8] and introdu-
ces coordination ambivalence in these ligands.
Abstract: Azocarboxamide (azcH) has been combined for
the first time with [Ru–Cym] to generate metal complexes
with N,N- and N,O-coordination mode, [(Cym)Ru(azc)Cl]
and [(Cym)Ru(azcH)Cl]+[PF6]ꢀ. Geometric and electronic
structures of the complexes are reported along with their
in vitro activities against different tumour cell lines and
preliminary results on solution chemistry. Compound
[(Cym)Ru(azc)Cl] exhibited remarkable cytotoxic proper-
ties. It was cell-type specific and had comparable IC50
values towards both cancer cells and their drug-resistant
subline. A tenfold increase in the sensitivity towards
[(Cym)Ru(azc)Cl] was noted for the tumour cells with de-
pleted intracellular glutathione (GSH) level, suggesting the
essential role of GSH in cell response to this compound.
Figure 1. Ru–Cym complex 1 with pap which has been reported to show
catalytic anticancer activity and azcH used in this work.
Azo compounds are a known class of redox-active ligands,
with phenylazopyridine (pap) being a prominent example of
this class. The RuII complex 1 with pap as a ligand has been re-
cently shown to display excellent cytotoxic properties against
cancer cells (Figure 1).[1,2] Ruthenium complexes now belong to
a class of compounds recognized for their anti-tumour behav-
iour, with [indH]trans-[Ru(N-ind)2Cl4] (KP1019) and [imiH]trans-
[Ru(N-imi)(S-dmso)Cl4] (NAMI-A) being in phase I of clinical
trials.[3] In the context of cytotoxicity, it is known that azocar-
boxamides are cytotoxic to cancer cells and increase their sen-
sitivity against some metal-based drugs, such as cisplatin.[4]
Hence, it is surprising that a direct coordination of such ligand-
to-metal complexes and their effect on tumour cells has not
been investigated yet. Azocarboxamides,[5] like related azo
compounds[6] are redox-active, owing to the electron-acceptor
Herein, we present N-2-diphenyldiazenecarboxamide (azcH)
and the Ru–Cym (Cym=p-cymene) complexes, [(Cym)-
Ru(azc)Cl] (2) and [(Cym)Ru(azcH)Cl]+[PF6]ꢀ (3[PF6]), in which
both coordination modes N,N- and N,O- have been observed
(Scheme 1).
Scheme 1. Preparation of 2 and 3[PF6]; [Ru]=[(Cym)RuCl2]2.
Complexes 2 and 3[PF6] were synthesized by the treatment
of [(Cym)RuCl2]2 with azcH. Whereas the presence of a base led
to the formation of the neutral complex 2, the absence of
a base and the presence of the chloride abstracting reagent
NH4PF6 led to the formation of 3[PF6] (Scheme 1).
[a] M. G. Sommer, Dr. D. Schweinfurth, Dr. M. Bubrin, Prof. Dr. B. Sarkar
Institut fꢀr Chemie und Biochemie, Anorganische Chemie
Freie Universitꢁt Berlin, Fabeckstraße 34–36, 14195, Berlin (Germany)
E-mail: Biprajit.Sarkar@fu-berlin.de
´
[b] P. Kureljak, N. Stojanovic, Dr. M. Osmak
ˇ
´
der Boskovic Institute
Division of Molecular Biology, Ru
¯
ˇ
Azocarboxamide, azcH, and the complexes were character-
ized by single crystal X-ray diffraction (Figure 2). The crystal
structure of azcH contains two symmetry-independent mole-
cules in the unit cell (Figure S1 in the Supporting Information).
The bond lengths are within the expected range, with the C=O
bond length of 1.222(2) ꢀ and the N=N bond length of
1.253(2) ꢀ confirming their double bond character (Table S1 in
Bijenicka cesta 54, HR-10000 Zagreb (Croatia)
E-mail: Maja.Osmak@irb.hr
ˇ
[c] Dr. D. Urankar, Dr. M. Gazvoda, Prof. Dr. J. Kosmrlj
Faculty of Chemistry and Chemical Technology, University of Ljubljana
ˇ
ˇ
Askerceva 5, SI-1000, Ljubljana (Slovenia)
E-mail: Janez.Kosmrlj@fkkt.uni-lj.si
Supporting information for this article is available on the WWW under
http://dx.doi.org/10.1002/chem.201404448.
Chem. Eur. J. 2014, 20, 17296 – 17299
17296
ꢁ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Communication
the Supporting Information). The structure of 2 also contains
two symmetry-independent molecules in the unit cell and 0.25
dichloromethane per molecule of 2. In 2, the ruthenium centre
at 1643 and 1644 cmꢀ1 for 2 and 3[PF6], respectively. The C=O
bond is likely weakened to the same extent due to negative
charge in the case of 2 and p-backbonding in the case 3[PF6].
The compound azcH displays an irreversible reduction step
at ꢀ1.33 V in CH2Cl2/0.1m Bu4PF6. Compound 2 displays a rever-
sible oxidation wave at 0.81 V and an irreversible reduction
wave at ꢀ1.22 V. On the other hand, 3+ shows a reversible re-
duction step at ꢀ0.35 V and an irreversible second reduction
step at ꢀ1.16 V (Figure S7 in the Supporting Information). No
oxidation wave was observed for 3+ within the solvent
window. Thus, it is seen that the azcꢀ makes the complex 2
susceptible towards oxidation, as would be expected for an
amide bound to a RuII centre. For 3+ on the other hand, metal
coordination renders the reduction step facile and reversible,
as is known for other neutral azo ligands. Such a metal-centre
induced ease of reduction is prevented in 2 possibly through
the negative charge of azcꢀ.
has
a piano-stool-like half-sandwich coordination, and is
bound to azcꢀ through the amide N3 and the azo N1 atom.
The RuꢀN1 and RuꢀN3 bond lengths are comparable.
Just like in 2, the ruthenium centre in 3[PF6] possesses
a three-legged piano-stool-like coordination, and the ligand
azcH is coordinated to it through the azo N1 atom, and the
O atom of the carbonyl group. The RuꢀN and RuꢀO lengths
are in the expected range. In both 2 and 3[PF6], the N=N azo
bond is slightly elongated compared to free azcH (Table S2 in
the Supporting Information). This is likely a consequence of
back-donation from a filled dp orbital of RuII to a p* orbital of
the azo-containing ligand. The distance from the centre of the
arene ring to ruthenium is 1.703(1) and 1.686(1) ꢀ for 2 and
3[PF6], respectively.
Azocarboxamide, azcH, and the complexes were character-
ized by NMR and IR spectroscopy (Table S3, Figures S2–S5 in
the Supporting Information). Heteronuclear multiple bond co-
herence (gs-HMBC) technique enabled us to determine
15N NMR chemical shifts for the azo N1 nitrogen atoms. Coordi-
nation of the azcH (dN1 =144.3 ppm) to the ruthenium resulted
in remarkably large[9] upfield 15N coordination shifts (in CD3OD)
of ꢀ115 and ꢀ94 ppm in 2 (dN1 =29.3 ppm) and 3[PF6] (dN1
50.1 ppm), respectively. NMR studies in CD2Cl2 showed that 3+
can be converted to 2 in the presence of NEt3 (Figure S6 in the
Supporting Information). The carbonyl stretching band appears
The in situ one-electron oxidized species 2+ is EPR silent at
room temperature. At 110 K, a spectrum is observed with g//=
1.921 and g =2.208 (Figure S8a in the Supporting Informa-
?
tion). The average g-value of 2.112 and the g-anisotropy (Dg)
of 0.287 are clear indication of a predominantly metal-centred
spin, and 2+ can be formulated as [(Cym)RuIII(azc)ꢀCl]+. The
one-electron reduced 3 on the other hand shows an EPR
signal in fluid solutions at room temperature with a g-value of
2.001. Hyperfine coupling, possibly to N atoms is also partially
resolved. The g-value, the appearance of the signal at room
temperature, and the total width of the signal clearly point to-
C
wards a ligand-centred spin, and the formulation of 3 as
II
·
Cꢀ
[(Cym)Ru (azcH) Cl]
Information).
(Figure S8b
in
the
Supporting
In vitro activity of compound 2 was examined towards sev-
eral human tumour cell lines of different origin: cervical carci-
noma cells (HeLa), lung carcinoma cells (H460), colorectal carci-
noma cells (HCT-116), breast adenocarcinoma cells (MDA-MB-
231) as well as the laryngeal carcinoma cells (HEp-2) and their
carboplatin, cisplatin and curcumin resistant subline (7T) were
analysed (Table 1). Compound 2 exhibits high cytotoxic activity
against all the examined tumour cell lines with very low IC50
values ranging from 3.4–15.1 mm. The most sensitive were
HEp-2 cells (3.4 mm), followed by MDA-MB-231 and HeLa cells,
whereas H460 and HCT-116 cells were less sensitive. The 4.4-
fold ratio in IC50 for the most and the least sensitive cells lines
indicates the cell-type specific cytotoxicity of 2. The potency of
2 against HeLa cells was twice that of cisplatin, but it was com-
parable against HEp-2 cells. It should be emphasized that the
cytotoxicity of 2 towards HEp-2 cells and their drug-resistant
subline 7T are similar. Because drug-resistance is the major ob-
stacle for successful cancer treatment, this result is of great im-
portance for the potential use of 2 in the treatment of drug-re-
sistant tumours.
Azocarboxamide, azcH, was nontoxic (IC50 >100 mm) towards
HeLa cells and was not considered further for the tests, where-
as the solubility of complex 3[PF6] in the growth medium was
too low to allow such studies. This low solubility is presumably
due to the hydrophobic nature of the PF6ꢀ anion.
Figure 2. ORTEP plots of 2 (top) and 3[PF6] (bottom). Ellipsoids are drawn at
50% probability. Counter anions, hydrogen atoms and solvent molecules
have been omitted for clarity.
Chem. Eur. J. 2014, 20, 17296 – 17299
www.chemeurj.org
17297
ꢁ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Communication
Table 1. Antiproliferative activity of azcH and 2 compared to cisplatin.
Cell line
IC50 [mm] ꢁSD after 72 h incubation
azcH
2
cisplatin
HeLa
HEp-2
7T
MDA-MB-231
H460
HCT-116
>100
N.D.[b]
N.D.[b]
N.D.[b]
N.D.[b]
N.D.[b]
7.7ꢁ1.0
3.4ꢁ1.4
4.4ꢁ2.0
5.4ꢁ1.5
11.5ꢁ6.5
15.1ꢁ5.5
16.3ꢁ3.6[a]
3.33[c]
11.0[c]
N.D.[b]
N.D.[b]
N.D.[b]
[a] Reference [10a]. [b] Not determined. [c] Reference [10b].
Figure 3. The influence of compound 2 on HeLa cells with (c) and with-
out (control: g) 18 h pre-treatment with nontoxic dose of buthionine sul-
foximine (BSO, 0.01 mm), as determined by MTT assay. Pooled data from
three experiments (mean ꢁS.D.).
Hydrolysis of the metalꢀhalide bond into a more reactive
aqua complex is suggested to be a typical activation pathway
of the metal-based anticancer drugs including the arene–
ruthenium complexes.[2] An aqueous solution of 2 was studied
in D2O at 238C over 48 h with time-dependent NMR spectros-
copy, which indicated a combination of hydrolysis and arene
loss. Within this time complex 2 underwent hydrolysis into
[(Cym)Ru(azc)(D2O/OD)]+/0 to the extent of 47% as determined
by the ratio of integrals (Figure S9 in the Supporting Informa-
tion). Arene loss accounted for 7% of the initial complex 2 as
indicated by the presence of free Cym resonances in the spec-
tra. Accordingly, HPLC and electrospray ionization mass spec-
trometry (ESI-MS) analyses gave two chromatographic peaks
(Figure S10 in the Supporting Information). The less polar com-
pound gave ESI-HRMS peak at m/z 496.0722 for [2+H]+
whereas the more polar gave peaks at m/z 460.0968 and
501.1235, assignable to [2ꢀCl]+ and [2ꢀCl+CH3CN]+, respec-
tively, interpreted as a result of in-source collision-induced dis-
or ligand precursor in coordination to a transition metal. As
a result of the coordination ambivalence of azcH, two arene–
Ru complexes were easily prepared. Complex 2 was highly cy-
totoxic with IC50 values in the low micromolar range. The activ-
ity of 2 was cell-type specific and comparable in cancer cells
and their drug-resistant subline. A nearly tenfold increase in
the sensitivity of the tumour cells towards 2 was noted when
these were pre-treated with redox modulator BSO, indicating
that a combined treatment with these two compounds may
have potential as a selective therapeutic strategy.[11]
Acknowledgements
sociation of [(Cym)Ru(azc)(D2O/OD)]+/0
.
Fonds der Chemischen Industrie (FCI) and the Deutscher Aka-
demischer Austauschdienst (DAAD), Ministry of Science, Educa-
tion and Sport of the Republic of Croatia (Projects 098-
0982913-2748; 098-0982913-2850 and Joint Project BI-HR/12-
13-028), and The Ministry of Education, Science and Sport, Re-
public of Slovenia, the Slovenian Research Agency (P1-0230)
and postdoctoral grant to M.G. (430-168/2013/114) are grate-
fully acknowledged for the financial support. Dr. S. M. Mobin is
kindly acknowledged for solving the structure of azcH. Techni-
Tumour cells display a redox metabolism that is different
from that of the normal cells. Depletion of cellular glutathione
(GSH), the major scavenging system of reactive oxygen species
(ROS), has a profound effect on tumour cell survival with high
therapeutic selectivity.[11,12] To assess whether 2 might be in-
volved in the reaction with GSH, we first incubated these two
compounds in H2O at 238C. A clean formation of new species
was evident by the appearance of the third chromatographic
peak in HPLC ESI-MS and was assigned to the thiolato complex
[(Cym)Ru(azc)(GS)] (compare Figures S10 and S11 in the Sup-
porting Information). Although seen in Ru–arene chemistry,[13]
the formation of [(Cym)Ru(azc)(GS)] from 2 and GSH is in
a sharp contrast to the RuII–pap complex 1, where the sulf-
hydryl group of GSH undergoes addition to the diazene N=N
bond into a ruthenium-hydrazo intermediate.[1] Next, we inves-
tigated the effect of 2 on GSH in HeLa cells that were pre-incu-
bated with a nontoxic dose of buthionine sulfoximine (BSO),
a specific inhibitor of GSH synthesis, which is often used to de-
plete the cellular level of GSH.[11,14] A remarkable, nearly tenfold
increase in the sensitivity of the cells towards 2 was noted as
compared to the BSO-untreated cells. The IC50 decreased from
6.42 to 0.66 mm (Figure 3), suggesting an essential role of GSH
in cell response to compound 2.[15]
´
cal assistance of Mrs. Ana Rosic is also acknowledged.
Keywords: antitumor agents · azocarboxamide · N,N ligands ·
N,O ligands · ruthenium
[1] a) S. J. Dougan, M. Melchart, A. Habtemariam, S. Parsons, P. J. Sadler,
Inorg. Chem. 2006, 45, 10882–10894; b) S. J. Dougan, A. Habtemariam,
S. E. McHale, S. Parsons, P. J. Sadler, Proc. Natl. Acad. Sci. USA 2008, 105,
11628–11633; c) S. Samanta, P. Ghosh, S. Goswami, Dalton Trans. 2012,
41, 2213–2226; d) I. Romero-Canelꢂn, L. Salassa, P. J. Sadler, J. Med.
Chem. 2013, 56, 1291–1300.
[2] Selected reviews: a) W. Han Ang, P. J. Dyson, Eur. J. Inorg. Chem. 2006,
4003–4018; b) P. C. A. Bruijnincx, P. J. Sadler, Adv. Inorg. Chem. 2009, 61,
1–62; c) A. M. Pizarro, A. Habtemariam, P. J. Sadler, in Topics in Organo-
metallic Chemistry, Vol. 32 (Eds.: G. Jaouen, N. Metzler-Nolte), Springer,
Heidelberg, 2010, pp. 21–56; d) G. Gasser, I. Ott, N. Metzler-Nolte, J.
Med. Chem. 2011, 54, 3–25.
We report on the first synthesis, X-ray, spectroscopic and
electrochemical study of an azocarboxamide (azcH) as a ligand
[3] a) P. J. Dyson, G. Sava, Dalton Trans. 2006, 1929–1933; b) E. S. Antonara-
kis, A. Emadi, Cancer Chemother. Pharmacol. 2010, 66, 1–9; c) G. Sꢃss-
Chem. Eur. J. 2014, 20, 17296 – 17299
www.chemeurj.org
17298
ꢁ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Communication
´ ´ ´
[10] Reported in: a) N. Stojanovic, D. Urankar, A. Brozovic, A. Ambriovic-
Fink, Dalton Trans. 2010, 39, 1673–1688; d) R. Trondl, P. Heffeter, C. R.
Kowol, M. A. Jakupec, W. Berger, B. K. Keppler, Chem. Sci. 2014, 5, 2925–
2932.
ˇ
Ristov, M. Osmak, J. Kosmrlj, Acta Chim. Slov. 2013, 60, 368–374; b) M.
´
Osmak, L. Bizjak, B. Jernej, S. Kapitanovic, Mutat. Res. 1995, 347, 141–
ˇ
ˇ
ˇ
[4] J. Kosmrlj, M. Kocevar, S. Polanc, Synlett 2009, 2217–2235.
150, and determined under the same conditions as used in this work.
[11] D. Trachootham, J. Alexandre, P. Huang, Nat. Rev. Drug Discovery 2009,
8, 579–591.
ˇ
[5] J. Kosmrlj, M. Kocevar, S. Polanc, J. Chem. Soc. Perkin Trans. 1 1998,
3917–3919.
[6] W. Kaim, Coord. Chem. Rev. 2001, 219–221, 463–488.
[7] W. Kaim, Inorg. Chem. 2011, 50, 9752–9765.
[12] a) J. M. Estrela, A. Ortega, E. Obrador, Crit. Rev. Clin. Lab. Sci. 2006, 43,
143–181; b) M. L. Circu, T. Y. Aw, Free Radical Biol. Med. 2010, 48, 749–
´
[8] For selected recent publications on arene ruthenium complexes con-
taining N,N- and N,O-chelating ligands, see: a) A. Habtemariam, M. Mel-
chart, R. Fernꢄndez, S. Parsons, I. D. H. Oswald, A. Parkin, F. P. A. Fabbia-
ni, J. E. Davidson, A. Dawson, R. E. Aird, D. I. Jodrell, P. J. Sadler, J. Med.
Chem. 2006, 49, 6858–6868; b) N. P. E. Barry, J. Furrer, J. Freudenreich,
G. Sꢃss-Fink, B. Therrien, Eur. J. Inorg. Chem. 2010, 725–728; c) N. P. E.
Barry, O. Zava, J. Furrer, P. J. Dyson, B. Therrien, Dalton Trans. 2010, 39,
5272–5277; d) N. P. E. Barry, F. Edafe, B. Therrien, Dalton Trans. 2011, 40,
7172–7180; e) J. Freudenreich, J. Furrer, G. Sꢃss-Fink, B. Therrien, Orga-
nometallics 2011, 30, 942–951; f) G. S. Smith, B. Therrien, Dalton Trans.
2011, 40, 10793–10800; g) F. A. Egbewande, L. E. H. Paul, B. Therrien, J.
Furrer, Eur. J. Inorg. Chem. 2014, 1174–1184; h) Z. Adhireksan, G. E.
Davey, P. Campomanes, M. Groessl, C. M. Clavel, H. Yu, A. A. Nazarov,
C. H. F. Yeo, W. H. Ang, P. Drçge, U. Rothlisberger, P. J. Dyson, C. A.
Davey, Nat. Commun. 2014, 5, 3462; i) A. A. Nazarov, C. Hartinger, P. J.
Dyson, J. Organomet. Chem. 2014, 751, 251–260.
756; c) A. Brozovic, A. Ambriovic-Ristov, M. Osmak, Crit. Rev. Toxicol.
2010, 40, 347–359.
[13] a) A. L. Noffke, A. Habtemariam, A. M. Pizarro, P. J. Sadler, Chem.
Commun. 2012, 48, 5219–5246; b) F. Wang, J. Xu, K. Wu, S. K. Weidt,
C. L. Mackay, P. R. R. Langridge-Smith, P. J. Sadler, Dalton Trans. 2013, 42,
3188–3195.
[14] O. W. Griffith, A. Meister, J. Biol. Chem. 1979, 254, 7558–7560.
[15] For the comparison, in a similar experiment with BSO, a two- to sixfold
increase in the sensitivity of human ovarian cancer cells was observed
for organoiridium(III) complexes that exhibited oxidant catalytic anti-
cancer activities; see: Z. Liu, I. Romero-Canelꢂn, B. Qamar, J. M. Hearn,
A. Habtemariam, N. P. E. Barry, A. M. Pizarro, G. J. Clarkson, P. J. Sadler,
Angew. Chem. Int. Ed. 2014, 53, 3941–3946; Angew. Chem. 2014, 126,
4022–4027.
[9] a) L. Pazderski, Annual Reports on NMR Spectroscopy 2013, 80, 33–179;
b) A. Bolje, S. Hohloch, D. Urankar, A. Pevec, M. Gazvoda, B. Sarkar, J.
Received: July 16, 2014
ˇ
Kosmrlj, Organometallics 2014, 33, 2588–2598.
Published online on November 5, 2014
Chem. Eur. J. 2014, 20, 17296 – 17299
www.chemeurj.org
17299
ꢁ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim