Synthesis, characterization and cytotoxic activity of gallium(III) complexes anchored by tridentate pyrazole-based ligands

Article abstract of DOI:10.1016/j.jinorgbio.2010.01.003

Reactions of GaCl₃ with pyrazole-containing ligands of the pyrazole-imine-phenol (HL¹-HL³) or pyrazole-amine-phenol (HL⁴-HL⁶) types led to the synthesis of well-defined [GaL₂]⁺ h

Full text of DOI:10.1016/j.jinorgbio.2010.01.003

Journal of Inorganic Biochemistry 104 (2010) 523–532
Contents lists available at ScienceDirect
Journal of Inorganic Biochemistry
journal homepage: www.elsevier.com/locate/jinorgbio
Synthesis, characterization and cytotoxic activity of gallium(III) complexes
anchored by tridentate pyrazole-based ligands
Francisco Silva a, Fernanda Marques a, I.C. Santos a, António Paulo a, António Sebastião Rodrigues b,
José Rueff b, Isabel Santos a,
*
a Unidade de Ciências Químicas e Radiofarmacêuticas, Instituto Tecnológico e Nuclear, Estrada Nacional 10, 2686-953 Sacavém, Portugal
b CIGMH, Department of Genetics, Faculty of Medical Sciences, New University of Lisbon, Rua da Junqueira 96, 1349-008 Lisboa, Portugal
a r t i c l e i n f o
a b s t r a c t
Reactions of GaCl3 with pyrazole-containing ligands of the pyrazole-imine-phenol (HL1–HL3) or pyrazole-
amine-phenol (HL4–HL6) types led to the synthesis of well-defined [GaL2]+ homoleptic complexes (1–6).
Complexes 1–6 were characterized by elemental analysis, ESI-MS (electrospray ionization-mass spec-
trometry), IR and NMR spectroscopies, and in the case of Complex 1 also by X-ray diffraction analysis.
In complexes 1–3, the pyrazole-imine-phenolate ligands act as monoanionic chelators that coordinate
to the metal in a meridional fashion, while 4–6 contain monoanionic and facially coordinated pyra-
zole-amine-phenolate ligands. Complexes 1–3 have a greater stability in solution compared to 4–6,
which have shown a more pronounced tendency to release the respective ancillary ligands. The cytotox-
icity of 1–6 and of the respective ligands (HL1–HL6) was evaluated against human prostate cancer cells
PC-3 and human breast cancer cells MCF-7. The substituents of the phenolate rings strongly influenced
the cytotoxicity of the compounds. Complexes 3 and 6 that contain chloride substituents at the phenolate
rings have shown the highest cytotoxicity, including in the cisplatin-resistant PC-3 cell line. The cytotoxic
profile of 3 and 6 is very similar to the one displayed by the respective anchor ligands, respectively HL1
and HL6. The cytotoxic activity of 3 and 6 is slightly increased by the presence of transferrin, and both
complexes provoke cell death mainly by induction of apoptotic pathways.
Article history:
Received 5 July 2009
Received in revised form 5 January 2010
Accepted 11 January 2010
Available online 20 January 2010
Keywords:
Gallium(III) complexes
Pyrazole-based ligands
X-ray structures
Cytotoxicity
Apoptosis
Ó 2010 Elsevier Inc. All rights reserved.
1. Introduction
In recent years, the search for Ga(III) complexes potentially use-
ful as anticancer drugs has been based on bidentate and tridentate
The clinical success of cisplatin in the treatment of several hu-
man malignant tumors motivated in recent years strong research
efforts on the finding of alternative metal complexes with potential
usefulness as anticancer drugs [1]. So far, some of the most prom-
ising results have been reported for gallium, which exhibits anti-
neoplastic activity even in the form of simple Ga(III) salts, such
as the nitrate or chloride. However, the use of Ga(III) salts as anti-
cancer drugs presents some drawbacks, which are mainly related
with their toxicity and relatively poor bioavailability. To overcome
these limitations and seeking for compounds with more favorable
pharmacokinetics, Ga(III) complexes anchored by multidentate
chelators started to be evaluated in recent years as alternatives
to gallium salts. In particular, the complexation of Ga(III) by mul-
tidentate chelators is expected to prevent hydrolysis processes,
while improving bioavailability and cell membrane permeation
of the compounds [1,2].
chelators, containing aromatic or N-heterocyclic rings and combin-
ing different donor atoms such as nitrogen, sulfur and oxygen. So
far, the more promising results were reported for tris(-8-quinolino-
lato)gallium(III) (KP46) and tris(3-hydroxy-2-methyl-4H-pyran-4-
onato)gallium(III), two neutral complexes anchored by bidentate
donors that are currently undergoing clinical evaluation [1,2].
Compared to gallium salts, several monocationic and homoleptic
Ga(III) complexes anchored by tridentate chelators have also pre-
sented an enhanced cytotoxic activity against human tumor cells,
namely cisplatin-resistant cell lines [3–8]. Although the mecha-
nisms involved in the anticancer activity of Ga(III) complexes are
not fully understood, it is generally considered that the critical cel-
lular target of gallium is the enzyme ribonucleotide reductase [1].
More recently, it has been shown that cationic Ga(III) complexes
with N,N0O-donors, containing pyridine, amine and phenolate
coordinating groups, can act as potent inhibitors of the proteasome
activity, which emerged therefore as a novel therapeutic target of
gallium-based anticancer agents [9]. All together, these results
point out that Ga(III) complexes with tridentate chelators deserve
to be further investigated as cytotoxic agents. In this context, the
* Corresponding author.
E-mail address: isantos@itn.pt (I. Santos).
0162-0134/$ - see front matter Ó 2010 Elsevier Inc. All rights reserved.
doi:10.1016/j.jinorgbio.2010.01.003

524
F. Silva et al. / Journal of Inorganic Biochemistry 104 (2010) 523–532
finding of compounds with improved pharmacological and phar-
macokinetics properties and the understanding of the mechanisms
that are involved on their antiproliferative action are still very
important issues in contemporary inorganic medicinal chemistry.
In this contribution, we report on the synthesis and character-
ization of cationic and homoleptic Ga(III) complexes (1–6) an-
chored by (N, N0, O) pyrazolyl-containing chelators, as well as on
the evaluation of their interest as cytotoxic metallopharmaceuti-
cals. As shown in Fig. 1, we have explored pyrazolyl-imine-pheno-
late ligands (HL1–HL3) and the respective pyrazolyl-amine-
phenolate congeners (HL4–HL6). By choosing these ligands, we ex-
pected to clarify how the rigidity and ring substituents would
influence the structure, stability and cytotoxic activity of the corre-
sponding Ga(III) complexes (1–6) of the [GaL2]+ type.
Complexes 1–6 were characterized by the common analytical
techniques, including multinuclear (1H, 13C and 71Ga) NMR in the
case of 1–3 and X-ray diffraction analysis for 1. The biological eval-
uation of the ligands HL1–HL6 and complexes 1–6 comprised the
screening of their in vitro cytotoxic activity in the human cancer
cell lines MCF-7 and PC-3. For the more cytotoxic complexes, their
ability to induce apoptosis and the influence of transferrin on their
cytotoxic activity were also studied and will be reported herein.
for 2 h [11]. HL2 and HL3 were isolated as yellow microcrystalline
solids, after removing the solvent under vacuum and washing the
residue with n-hexane and diethyl ether. The corresponding
amines, HL5 and HL6, were obtained by reduction of HL2 and HL3
with a twofold excess of NaBH4 in dry methanol, as we have previ-
ously reported for HL4. The purification of HL5 and HL6 was done
by evaporation of methanol, redissolution of the residue in CH2Cl2
and washing of the organic solution with aqueous K2CO3. HL5 and
HL6 were isolated as yellow microcrystalline solids after removal of
dichoromethane and washing of the obtained residue with n-
hexane.
2.2.1. N-(2-hydroxy-3-methoxy-benzyl-2-pyrazolethyl)imine (HL2)
Starting from 0.1 g (0.9 mmol) of N-2-(aminoethyl)pyrazole
were obtained 0.163 g (0.66 mmol) of HL2. Yield: 74%. Element.
Anal. (%). Found: C, 62.85; N, 17.09; H, 7.17; Calc. for
C13H15N3O2?0.1 H2O: C, 63.19; N, 17.01 H, 6.20. IR (KBr, cmꢀ1):
1630 (tC@N), 3125 (t
OAH). 1H NMR dH (300 MHz, CDCl3, s = singlet,
d = doublet, m = multiplet, tr = triplet): 13.36 (1H, s, OH), 8.04 (1H,
s, N@CH), 7.51 (1H, d, H(3/5)-pz), 7.33 (1H, d, H(3/5)-pz), 6.89 (1H,
m, Ph), 6.83 (1H, m, Ph), 6.75 (1H, m, Ph), 6.13 (1H, tr, H(4)-pz),
4.34 (2H, tr, NACH2), 4.04 (2H, pz-CH2), 3.89 (3H, s, O–CH3). 13C
NMR dC (75.4 MHz, CDCl3): 167.81 (N@CH), 150.93 (Ph), 148.04
(Ph), 140.86 (C(3/5)-pz), 130.56 (Ph + C(3/5)-pz), 118.32 (Ph),
113.98 (Ph), 105.02 (C(4)-pz), 58.74 (CH2), 56.05 (CH3–O), 52.18
(CH2).
2. Experimental section
2.1. General procedures
The syntheses were carried under a nitrogen atmosphere, using
standard Schlenk techniques and dry solvents, while the work-up
procedures were performed under air. The compounds N-(2-ami-
2.2.2. N-(2-hydroxy-3,5-dichloride-benzyl-2-pyrazolethyl)imine (HL3)
Starting from 0.2 g (1.8 mmol) of N-2-(aminoethyl)pyrazole
were obtained 0.45 g (1.58 mmol) of HL3. Yield: 88%. Element.
Anal. (%). Found: C, 48.07; N, 13.46; H, 3.78. Calc. for
C12H11N3OCl2?H2O: C, 47.70; N, 13.91; H, 4.34. IR (KBr, cmꢀ1):
noethyl)pyrazole,
N-(2-hydroxybenzyl-2-pyrazolethyl)imine
(HL1) and N-(2-hydroxybenzyl-2-pyrazolethyl)amine (HL4) were
prepared according to the procedures described in the literature
[10,11]. 1H and 13C NMR spectra were recorded in a Varian Unity
1620 (tC@N), 3100 (t
OAH). 1H NMR dH (300 MHz, CDCl3): 13.87
(1H, s, OH), 7.85 (1H, s, N@CH), 7.48 (1H, d, H(3/5)-pz), 7.26 (1H,
d, H(3/5), pz), 7.25 (1H, m, Ph), 6.96 (1H, m, Ph), 6.12 (1H, tr,
H(4)-pz), 4.39 (2H, tr, NACH2), 4.03 (2H, pz-CH2). 13C NMR dC
(75.4 MHz, CDCl3): 166.45 (N@CH), 155.98 (Ph), 140.93 (C(3/5)-
pz), 131.04 (Ph + C(3/5)-pz), 133.36 (Ph), 124.67 (Ph), 119.14
(Ph), 105.41 (C(4)-pz), 58.24 (CH2), 51.77 (CH2).
300 MHz spectrometer and 71Ga NMR spectra in
a Bruker
Avance-500 MHz NMR spectrometer. 1H and 13C chemical shifts
(ppm) were referenced with the residual solvent resonances rela-
tive to tetramethylsilane. The 71Ga NMR spectra were recorded in
ppm relative to a Ga(NO3)3 external reference in D2O. IR spectra
were recorded as KBr pellets on a Bruker Tensor 27 spectrometer.
C, H and N analyzes were performed on an EA 110 CE instruments
automatic analyzer. Electrospray ionization-mass spectrometry
(ESI-MS) of complexes 1–6 was carried out with a QITMS instru-
ment, using acetonitrile solutions of the compounds.
2.2.3. N-(2-hydroxybenzyl-3-methoxy-2-pyrazolethyl)amine (HL5)
Starting from 0.1 g (0.41 mmol) of HL2 were obtained 0.081 g
(0.33 mmol) of HL5. Yield: 80%. Element. Anal. (%). Found: C,
61.38; N, 15.84; H, 6.24. Calc. for C13H17N3O2?0.5 H2O: C, 60.92;
N, 16.39; H, 7.08. IR (KBr, cmꢀ1): 3300 ( OAH). 1H
tNAH), 3090 (t
2.2. Synthesis of the ligands
NMR dH (300 MHz, CDCl3): 7.52 (1H, d, H(3/5)-pz), 7.39 (1H, d,
H(3/5)-pz), 6.81 (1H, m, Ph), 6.74 (1H, m, Ph), 6.58 (1H, m, Ph),
6.24 (1H, tr, H(4)-pz), 4.26 (2H, tr, pz-CH2), 3.98 (2H, s, NACH2),
3.86 (3H, s, O–CH3), 3.09 (2H, tr, NACH2). 13C NMR dC (75.4 MHz,
CDCl3): 147.96 (Ph), 146.98 (Ph), 140.07 (C(3/5)-pz), 129.99
(Ph + C(3/5)-pz), 115.42 (Ph), 110.52 (Ph), 105.53 (C(4)-pz), 55.97
(CH3–O), 51.56 (CH2), 51.07 (CH2), 47.78 (CH2).
The new pyrazole-imine-phenolate ligands HL2 and HL3 were
synthesized and purified as previously described for the congener
HL1 by reacting N-2-(aminoethyl)pyrazole with an equivalent
amount of the respective salicilaldeyde, in dry EtOH under reflux
2.2.4. N-(2-hydroxybenzyl-3,5-dichloride-2-pyrazolethyl)amine (HL6)
Starting from 0.227 g (0.8 mmol) of HL3 were obtained 0.192 g
(0.67 mmol) of HL6. Yield: 84%. Element. Anal. (%). Found: C,
48.07; N, 13.46; H, 3.78. Calc. for C12H13N3OCl2?0.9 H2O: C,
47.67; N, 13.90; H, 4.93. IR (KBr, cmꢀ1): 3200 (
tNAH), 3125 (tOAH).
1H NMR dH (300 MHz, CDCl3): 7.52 (1H, d, H(3/5)-pz), 7.39 (1H, d,
H(3/5)-pz), 6.87 (1H, m, Ph), 6.85 (1H, Ph), 6.27 (1H, tr, H(4)-pz),
4.26 (2H, d, pz-CH2), 3.96 (2H, s, NACH2), 3.08 (2H, NACH2). 13C
NMR dC (75.4 MHz, CDCl3): 152.70 (Ph), 139.96 (C(3/5)-pz),
130.75 (Ph + C(3/5)-pz), 126.87 (Ph), 124.64 (Ph), 121.60 (Ph),
105.79 (C(4)-pz), 52.09 (CH2), 50.52 (CH2), 47.77 (CH2).
Fig. 1. Ligands of the pyrazole-imine-phenol (HL1–HL3) and pyrazole-amine-
phenol (HL4–HL6) type used in this work to prepare homoleptic Ga(III) complexes.

F. Silva et al. / Journal of Inorganic Biochemistry 104 (2010) 523–532
525
2.3. Synthesis of the gallium complexes
NACH2), 4.46 (4H, br, pz-CH2). 13C NMR dC (75.4 MHz, CD3CN):
173.43 (N@CH), 157.52 (Ph), 140.04 (C(3/5)-pz), 132.12 (Ph + C(3/
5)-pz), 135.19 (Ph), 123.56 (Ph), 117.78 (Ph), 106.69 (C(4)-pz),
60.09 (pz-CH2), 49.94 (NACH2). 71Ga NMR dGa (152.5 MHz, CD3CN):
25.97.
2.3.1. General procedure for the synthesis of Complexes 1–3 and 5
Inside a nitrogen-filled glove-box, a solution of the desired li-
gand (L) and NEt3 in dry methanol was stirred at room temperature
for 30 min. Then, solid GaCl3 was added and the resulting solution
was stirred for 1 h at room temperature. NH4PF6 was added and
the mixture was stirred for another hour. The Ga complexes pre-
cipitated as microcrystalline solids, which were recovered by cen-
trifugation, washed with cold methanol and dried under vacuum.
2.3.6. Gallium(III) bis[N-(2-hydroxybenzyl-2-pyrazolethyl)amine]
chloride (4)
Starting from 0.032 g (0.18 mmol) of GaCl3 were obtained
0.074 g (0.14 mmol) of complex [Ga(L4)2]Cl (4) in the form of a
white microcrystalline solid. Yield: 76%. Element. Anal (%). Found:
C, 50.85; N, 14.57; H, 5.26. Calc. for C24H28GaN6O2Cl?1.7 H2O: C,
50.72; N, 14.79; H, 5.57. ESI-MS (m/z): Found: 500.9 [M]+. Calc:
The syntheses were performed using
GaCl3:NH4PF6) molar ratio of the reagents.
a 2:2:1:1 (L:NEt3:-
2.3.2. General procedure for the synthesis of Complexes 4 and 6
Inside a nitrogen-filled glove-box, a solution of the desired li-
gand (L) and an equimolar amount of NEt3 in dry methanol was
stirred at room temperature for 30 min. Then, one-half equivalent
of solid GaCl3 was added and the resulting solution was stirred for
1 h at room temperature. After this time, the complexes precipi-
tated as microcrystalline solids that were recovered by centrifuga-
tion, washed with cold methanol and dried under vacuum.
501.1 [M]+. IR (KBr, cmꢀ1): 3125 ( NAH). 1H NMR dH (300 MHz,
t
DMSO-d6): 8.35 (2H, br, H(3/5)-pz), 7.86 (2H, br, H(3/5)-pz), 6.95
(2H, m, Ph), 6.86 (2H, m, Ph), 6.56 (2H, d, Ph), 6.42 (2H, br, Ph),
6.15 (2H, H(4)-pz), 5.05 (2H, NH), 4.48 (4H, br, CH2), 3.75 (4H, br,
CH2).
2.3.7. Gallium(III) bis[N-(2-hydroxy-3-methoxy-benzyl-2-
pyrazolethyl)amine] hexafluorophosphate (5)
Starting from 0.046 g (0.26 mmol) of GaCl3 were obtained
0.101 g (0.17 mmol) of complex [Ga(L5)2]PF6 (5) in the form of a
white microcrystalline solid. Yield: 65%. Element. Anal (%). Found:
C, 42.66; N, 11.40; H, 5.28. Calc. for C26H32GaN6O2PF6?3.2 H2O: C,
42.61; N, 11.47; H, 5.28. ESI-MS: (m/z): Found: 561.0 [M]+. Calc:
2.3.3. Gallium(III) bis[N-(2-hydroxybenzyl-2-pyrazolethyl)imine]
hexafluorophosphate (1)
Starting from 0.041 g (0.23 mmol) of GaCl3 were obtained
0.096 g (0.15 mmol) of complex [Ga(L1)2]PF6 (1) in the form of a
white microcrystalline solid. Yield: 65%. Element. Anal (%). Found:
C, 44.86; N, 13.47; H, 3.10. Calc. for C24H24GaN6O2PF6: C, 44.82; N,
13.07; H, 3.76. ESI-MS: (m/z): Found: 497.0 [M]+. Calc: 497.1 [M]+.
561.1 [M]+. IR (KBr, cmꢀ1): 3135 ( NAH). 1H NMR dH (300 MHz,
t
CD3CN): 8.01 (2H, br, H(3/5)-pz), 7.55 (2H, br, H(3/5), pz), 6.68
(2H, d, Ph), 6.48–6–43 (4H, m. Ph), 6.18 (2H, br, H(4)-pz), 5.30
(2H, tr, CH2), 4.83 (2H, br, NH), 4.28 (2H, m, CH2), 3.89 (2H, m,
CH2), 3.78 (6H, s, O–CH3), 3.48 (2H, m, CH2), 3.33 (2H, CH2), 2.49
(2H, m, CH2).
IR (KBr, cmꢀ1): 1625 ( C@N). 1H NMR dH (300 MHz, CD3CN): 8.48
t
(2H, s, N@CH), 7.66 (2H, d, H(3/5),pz), 7.26 (6H, m, Ph + H(3/5)-
pz), 6.27 (4H, m, Ph), 6.22 (2H, tr, H(4)-pz), 5.01 (2H, br, NACH2),
4.61 (2H, NACH2), 4.33 (4H, pz-CH2). 13C NMR dC (75.4 MHz,
CD3CN): 174.79 (N@CH), 167.36 (Ph), 140.21 (C(3/5)-pz), 137.32
(Ph + C(3/5)-pz), 134.45 (Ph), 123.18 (Ph) 117.33 (Ph) 106.73
(C(4)-pz), 59.99 (CH2), 50.44 (NACH2). 71Ga NMR dGa (152.5 MHz,
CD3CN): 24.07.
2.3.8. Gallium(III) bis[ N-(2-hydroxy-3,5-dichloride-benzyl-2-
pyrazolethyl)amine] chloride (6)
Starting from 0.025 g (0.14 mmol) of GaCl3 were obtained
0.054 g (0.08 mmol) of complex [Ga(L6)2]Cl (6) in the form of a pale
violet microcrystalline solid. Yield: 57%. Element. Anal (%). Found:
C, 40.99; N, 11.85; H, 3.00. Calc. for C24H24GaN6O2Cl5?1.6 H2O: C,
40.93; N, 11.93; H, 3.89. ESI-MS (m/z): Found: 638.9 [M]+. Calc:
2.3.4. Gallium(III) bis[N-(2-hydroxy-3-methoxy-benzyl-2-
pyrazolethyl)imine] hexafluorophosphate (2)
Starting from 0.027 g (0.15 mmol) of GaCl3 were obtained
0.057 g (0.08 mmol) of complex [Ga(L2)2]PF6 (2) in the form of a
greenish microcrystalline solid. Yield: 54%. Element. Anal (%).
Found: C, 43.25; N, 11.74; H, 4.93. Calc. for C26H28GaN6O4PF6?H2O:
C, 43.30; N, 11.65; H, 4.19. ESI-MS: (m/z): Found: 557.0 [M]+. Calc:
639.0 [M]+. IR (KBr, cmꢀ1): 3125 ( NAH). 1H NMR dH (300 MHz,
t
CD3CN): 8.07 (2H, br, H(3/5)-pz), 7.66 (2H, br, H(3/5)-pz), 7.22
(2H, br, Ph), 6.90 (2H, br, Ph), 6.43 (2H, br, H(4)-pz), 5.24 (2H, tr,
CH2), 4.93 (2H, br, NH), 4.37 (2H, d, CH2), 3.82 (2H, d, CH2), 3.63
(2H, m, CH2), 3.40 (2H, d, CH2), 2.61 (2H, m, CH2).
557.1 [M]+. IR (KBr, cmꢀ1): 1620 ( C@N). 1H NMR dH (300 MHz,
t
CD3CN): 8.46 (2H, s, N@CH), 7.65 (2H, d, H(3/5)-pz), 7.24 (2H, d,
H(3/5)-pz), 6.92 (2H, m, Ph), 6.64 (2H, m, Ph), 6.57 (2H, m, Ph),
6.20 (2H, tr, H(4)-pz), 5.36 (2H, br, NACH2), 4.55 (2H, br, NACH2),
4.32 (4H, br, pz-CH2), 3.73 (4H, O–CH3). 13C NMR dC (75.4 MHz,
CD3CN): 178.99 (N@CH), 162.34 (Ph), 156.63 (Ph), 144.47 (C(3/
5)-pz), 139.03 (Ph + C(3/5)-pz), 125.56 (Ph), 120.81 (Ph), 110.93
(C(4)-pz), 64.21 (pz-CH2), 60.58 (OCH2), 54.33 (NACH2). 71Ga
NMR dGa (152.5 MHz, CD3CN): 27.99.
2.3.8.1. X-ray diffraction analysis. The X-ray diffraction analysis of 1
has been performed on a Bruker AXS APEX CCD area detector dif-
fractometer, using graphite monochromated Mo K
a radiation
(0.71073 Å). Empirical absorption correction was carried out using
SADABS [12]. Data collection and data reduction were done with
the SMART and SAINT programs [13]. The structure was solved
by direct methods with SIR97 [14] and refined by full-matrix
least-squares analysis with SHELXL-97 [15] using the WINGX
[16] suite of programmes. All non-hydrogen atoms were refined
anisotropically. The remaining hydrogen atoms were placed in cal-
culated positions. Molecular graphics were prepared using ORTEP3
[17]. A summary of the crystal data, structure solution and refine-
ment parameters are given in Table 1.
2.3.5. Gallium(III) bis[N-(2-hydroxy-3,5-dichloride-benzyl-2-
pyrazolethyl)imine] hexafluorophosphate (3)
Starting from 0.019 g (0.11 mmol) of GaCl3 were obtained
0.039 g (0.05 mmol) of complex [Ga(L3)2]PF6 (3) in the form of a
greenish microcrystalline solid. Yield: 45%. Element. Anal (%).
Found: C, 36.45; N, 10.95; H, 2.25. Calc. for C24H20Ga-
N6O2Cl4PF6ꢁ0.2H2O: C, 36.74; N, 10.71; H, 2.58. ESI-MS: (m/z):
2.4. Cell viability assays in human tumor cell lines
Found: 634.9 [M]+. Calc: 634.9 [M]+. IR (KBr, cmꢀ1): 1620 (
tC@N).
2.4.1. Experiments with the complexes and respective ligands
The tumor cell lines MCF-7 and PC-3 (ATCC) were cultured in
DMEM (Dulbecco’s Modified Eagle Medium) containing Gluta-
Max I (MCF-7) and RPMI 1640 (PC-3) culture medium (Gibco)
1H NMR dH (300 MHz, CD3CN): 8.50 (2H, s, N@CH), 7.75 (2H, d,
H(3/5)-pz), 7.49 (2H, d, H(3/5)-pz), 7.36 (2H, m, Ph), 7.21 (2H, m,
Ph), 6.29 (2H, tr, H(4)-pz), 5.17 (2H, br, NACH2), 4.60 (2H, br,

526
F. Silva et al. / Journal of Inorganic Biochemistry 104 (2010) 523–532
Table 1
replaced with fresh one. Compounds 3 and 6, dissolved in DMSO,
Crystallographic data for [Ga(L1)2]PF6 (1).
were added to the wells and after 24 h the medium was removed,
and cells were washed with binding buffer (BB) (HEPES 0,1 M, NaCl
1,4 M, CaCl2 25 mM, pH = 7,4). The cytotoxic drug taxol (Paclitaxel)
Chemical formula
Formula weight
Crystal system
Space group
a/Å
b/Å
c/Å
b/deg
V(Å3)
Z
C24H24F6GaN6O2P
643.18
Monoclinic
Pn
8.3515(5)
12.1952(7)
12.9603(6)
102.345(2)
1289.46(12)
2
150(2)
1.657
1.209
3.00–25.68
3633
was used as a positive control (100 nM). Annexin V (Sigma) (1
lg/
ml) and propidium iodide (Chemicon Int) (2 g/ml) in BB was
l
added to each well for 20 min at 37 °C, in the dark. After this per-
iod, cells were washed with BB, and Hoechst 33258 in BB was
added (1 lg/ml) for 10 min. Cells were finally washed with BB
and mounted with Anti fade Mounting medium (VectaSHIELDÒ
H-1000), and analyzed by fluorescence microscope (LEICA DMLB),
equipped with the appropriate filters and attached to a digital cam-
era (applied imaging) and to a personal computer. Images were
captured from each slide using Cytovision (v3.0) capture software.
Duplicate assays with replicates were performed and at least five
different fields were analyzed per dose per experiment.
T/K
q
(calcd.)/gcmꢀ3
lVoK
a
mmꢀ1
h range for data collection (°)
n°of data
n° of params
361
R indices (all data)
R1 = 0.0295
wR2 = 0.0611
R1 = 0.0271
wR2 = 0.0600
0.898
R indices [I > 2r(I)]
2.6. Statistical analysis
GOF
Statistical analysis was done with GraphPad Prism software to
assess if there was a significant difference between the IC50 values
of complexes 1–3 and ligands HL1–HL3. A p-value < 0.0001 was
considered statistically relevant.
supplemented with 10% FBS and 1% penicillin/streptomycin at
37 °C in a humidified atmosphere of 95% of air and 5% CO2
(Heraeus, Germany). The cells were adherent in monolayers
and, when confluent, were harvested by digestion with 0.05%
trypsin–EDTA (Gibco) and seeded farther apart.
3. Results and discussion
Cell viability was evaluated by using a colorimetric method
based on the tetrazolium salt MTT ([3-(4,5-dimethylthiazol-2-yl)-
2,5-diphenyltetrazolium bromide]), which is reduced by living
cells to yield purple formazan crystals. Cells were seeded in 96-
well plates at a density of 2 ꢂ 104 cells (PC-3) and 5 ꢂ 104 cells
3.1. Synthesis
As depicted in Scheme 1, the novel cationic Ga(III) complexes,
1–6, were obtained by treating gallium(III) chloride with the corre-
sponding ligand in a 1:2 molar ratio, in dry methanol and under
basic conditions. Complexes 1–3 and 5 were isolated as microcrys-
talline solids in moderate yield (45–65%) after counter-ion metath-
esis with sodium hexafluorophosphate. Being less soluble in
methanol, 4 and 6 precipitated directly from the respective reac-
tion mixture, and were isolated as chloride salts in moderate to
high yield (57–76%). All the isolated compounds (1–6) are soluble
in polar organic solvents like dimethyl sulfoxide, but are only spar-
ingly soluble in water. Their characterization has been done by FT-
IR, 1H NMR, ESI-MS and elemental analysis. For some of the com-
pounds, the obtained elemental analysis are somewhat unsatisfac-
tory but the proposed formulations are supported by other
analytical data, such as 1H NMR and ESI-MS data. Complexes 1–3
were also characterized by 13C and 71Ga NMR, as well as by X-
ray diffraction analysis in the case of 1. The analytical data ob-
tained for 1–3, including good elemental analyzes, were consistent
with the presence of pseudo-octahedral [GaL2]+ complexes. ESI-MS
and 1H NMR analysis of 4–6 have shown the presence of the same
type of homoleptic complexes, but have also indicated the pres-
ence of the respective free ligands. As discussed below in more de-
tail, this may reflect the occurrence of dynamic processes in
solution, namely of the intermolecular type, and not necessarily
the contamination with excess of the ligand.
(MCF-7) per well in 200 lL of culture medium and left to incubate
overnight for optimal adherence. After careful removal of the med-
ium, 200 L of a dilution series of the compounds in fresh medium
l
were added and incubation was performed at 37 °C/5% CO2 for 24 h
or 72 h. HL1–HL6 and 1–6 were first solubilized in DMSO, diluted in
medium and added to the cells in final concentrations between
16 nM and 160 lM. The percentage of DMSO in cell culture med-
ium did not exceed 1%. Cisplatin was first solubilized in saline
and then added at the same concentrations used for the other com-
pounds. At the end of the incubation period, the compounds were
removed and the cells were incubated with 200
(500 g/ml). After 3–4 h at 37 °C/5% CO2, the medium was re-
moved and the purple formazan crystals were dissolved in
200 L of DMSO by shaking. The cell viability was evaluated by
lL of MTT solution
l
l
measurement of the absorbance at 570 nm by using a plate spec-
trophotometer (Power Wave Xs, Bio-Tek). The cell viability was
calculated dividing the absorbance of each well by that of the con-
trol wells (cells treated with medium containing 1% DMSO). Each
experiment was repeated at least three times and each point was
determined in at least six replicates.
2.4.2. Experiments with the complexes and transferrin
The effect of transferrin on cell viability either as a single agent
or in combination with 3 and 6 was evaluated using 100 lg/well of
transferrin and increasing concentrations of 3 and 6 (1.6, 16 and
3.2. Mass spectrometry and spectroscopic characterization
80
l
M). MCF-7 or PC-3 cells were plated at a density of 8 ꢂ 103
cells (PC-3) and 1.2 ꢂ 104 cells (MCF-7) per well in 200
lL of cul-
The most significant feature of the IR spectra of complexes 1–3
is the presence of
m ,
(C@N) bands in the range 1625–1620 cmꢀ1
ture medium and were incubated in the presence of the com-
pounds. After 72 h cell viability was evaluated using the MTT
assay as described above.
undergoing a slightly red shift upon coordination to the metal.
The IR spectra of 4–6 showed medium intense bands between
3135 and 3125 cmꢀ1 that were assigned to
m(NAH) of the coordi-
2.5. Apoptosis (annexin assay)
nated ligands. These frequencies compare well with those that
have been reported for other homoleptic Ga(III) complexes an-
chored by related tridentate ligands containing pyridine and phen-
oxide coordinating groups [18].
PC-3 cells were seeded in eight well chamber slides (LABTEK,
Nalge Nunc Int.) and 24 h later the medium was removed and

F. Silva et al. / Journal of Inorganic Biochemistry 104 (2010) 523–532
527
1) HL, NEt3, MeOH, r.t, 1h
1) HL, NEt3, MeOH, r.t., 1h
2) NH4PF6, r.t., 1h
GaCl3
2) NH4PF6, r.t., 1h
HL, NEt3, MeOH, r.t., 1h
PF6
R'
Cl
PF6
R
O
R'
R
O
MeO
N
N
N
N
N
N
O
O
N
Ga
N
H
N
Ga
N
H
H
N
Ga
N
H
O
R
N
N
O
R
N
N
N
N
R'
R'
OMe
1: R=R’=H
4: R=R’=H
6: R=R’=Cl
5
2: R=OMe, R’=H
3: R=R’=Cl
Scheme 1. Synthesis of the Ga(III) complexes.
For 1–3 in acetonitrile solution, the ESI mass spectra in the po-
sitive mode have shown the presence of prominent peaks corre-
sponding to m/z = [GaL2]+ with isotopic patterns in agreement
with the elemental composition of the complexes. Any other peaks
present in the mass spectra of 1–3 displayed very low intensity.
Under the same conditions, the ESI mass spectrometry analysis
of complexes 4–6 gave markedly different results, being observed
quite intense peaks due the protonated ligands (m/z = [L + H]+).
Nevertheless, the molecular-ions corresponding to the homoleptic
Ga(III) complexes (m/z = [GaL2]+) were detected in all the spectra.
Complexes 1–3 were studied by multinuclear (1H, 13C and
71Ga) NMR. In CD3CN solution, the 1H NMR and 13C NMR spectra
of 1–3 are very similar and are consistent with the magnetic
equivalence of the two ligands coordinated to the Ga3+ ion in a
meridional fashion, a coordination mode that was confirmed for
1 in the solid state by X-ray crystallography (see below). As
exemplified for Complex 1 in Fig. 2, such magnetic equivalence
was easily accounted for by the presence of a unique set of reso-
nances assigned to the aromatic protons, to the imine protons
and to the pyrazolyl protons in the 1H NMR spectra of 1–3. These
protons are slightly lowfield shifted (Dd = 0.1–0.3 ppm) compared
to the corresponding protons in the respective free ligands, cor-
roborating the coordination of the phenoxide, imine and pyrazolyl
groups to the metal. By contrast, the CH2 protons from the ethyl-
enic bridge linking the pyrazolyl and imine groups suffer a more
pronounced deshielding (Dd = 0.3–0.6 ppm) upon coordination to
the metal. The CH2 protons adjacent to the imine function origi-
nate two broad signals, appearing between 4.55 and 4.61 ppm
Fig. 2. 1H NMR spectrum of 1 in CD3CN.

528
F. Silva et al. / Journal of Inorganic Biochemistry 104 (2010) 523–532
and integrating each for two protons. The CH2 protons adjacent to
the pyrazolyl ring give rise to a unique broad signal that appears
in the range 3.93 and 4.46 ppm and integrates for four protons.
The assignment of these CH2 resonances has been based on the
study of 1 by 1H/13C HSQC and on the similarity of the 1H and
13C NMR spectra of complexes 1–3. These compounds were also
studied by 71Ga NMR. The obtained spectra have shown the pres-
ence of one relatively broad resonance (see Fig. 3) in the range
+22.1/+28.0 ppm. These values are at the low-end of chemical
shifts that have been reported for octahedral Ga3+ complexes,
Fig. 3. 71Ga NMR spectrum of 1 in CD3CN (ꢃGa(H2O)6).
Fig. 4. 1H NMR spectra of Complex 3 in DMSO-d6: immediately after preparation of the sample (in the top) and after 24 h in solution at room temperature (in the bottom).

F. Silva et al. / Journal of Inorganic Biochemistry 104 (2010) 523–532
529
and comparable to values found for other gallium complexes con-
taining phenolate donor groups [19,20].
of such configuration [GahNpz1Npz2ihNam1Nam2ihOphen1Ophen2i] in
spite of the poor quality of the crystallographic data. The best
crystal measured did not provide a good quality data set to deter-
mine a satisfactory structure for 6. Nevertheless, the connectivity
of the atoms was determined unambiguously (compound 6 crys-
tallized from a saturated methanol solution as white crystals in
the triclinic space group P1; with cell parameters
The 1H NMR spectra of 4–6 have always shown the presence of
free ligand, even using freshly prepared solutions of the complexes
in DMSO-d6 or CD3CN. The tendency to release the ligand is more
pronounced in dimethyl sulfoxide than in acetonitrile. Due to its
limited solubility in most common solvents, the 1H NMR spectrum
of 4 could only be recorded in DMSO-d6. In this solvent, the signals
due to the protons of the coordinated ligands are considerably
broad (see Fig. 4), which seems to indicate that the complex is
undergoing a dynamic process in solution. However, it was not
possible to perform variable temperature NMR experiments due
to the high freezing point of DMSO-d6. The 1H NMR spectra of 4–
6 in DMSO-d6 have shown the presence of signals assignable to
the respective free ligands, immediately after preparation of the
samples. In the case of Complex 4, the proportion of free ligand
was 1:2.5 immediately after preparation of the sample, rising to
1:1.5 after 2 h at room temperature and remaining fairly constant
(1:1 molar ratio) after 48 and 72 h at room temperature. These
data suggest that the complexes may undergo an intermolecular
process involving, eventually, the replacement of one of the ancil-
lary ligands by dimethyl sulfoxide molecules. We could not iden-
tify signals assignable to any species containing coordinated
DMSO, due most probably to the broadness of the spectra. All to-
gether, these data indicated that 4–6 have a poorer stability in
DMSO solution compared with 1–3.
a = 7.8745(4) Å, b = 10.5419(6) Å, c = 18.0674 Å,
a = 93.784(3)°,
b = 95.814(4)°,
c
= 95.143(3)°, V = 5394(3) Å3). Most probably, the
configuration found in the solid state for 6 is retained in solution
in agreement with the pattern and relative intensity of the peaks
observed in the 1H NMR spectra of 5 and 6 in CD3CN. Compounds
4–6 were not characterized by 13C NMR due to their tendency to
release slowly the respective ancillary ligands when kept in solu-
tion. Using the same conditions as for complexes 1–3, no 71Ga
NMR signal could be obtained for 4–6. 71Ga is a quadrupolar nu-
clei (I = 3/2) and the broadening of the line width of 71Ga NMR
signals can be due to an exchange process, a change in the corre-
lation time sc and the reduction of symmetry about the metal ion
[19]. Taking into consideration the behavior observed in solution
for complexes 4–6 the occurrence of exchange processes is the
most plausible possibility to justify the broadening of the line
and the absence of 71Ga signals.
3.3. X-ray diffraction analysis of 1
In the case of 5 and 6, it has been possible to record the
respective 1H NMR spectra in CD3CN. In this solvent, the reso-
nances appear much narrower and with well-resolved multiplic-
ity, being observed three resonances for the pyrazolyl protons,
one resonance for the NH protons and six multiplets assignable
to the CH2 protons, integrating each for two protons. These data
indicate that both ligands are magnetically equivalent, and also
pointed out that the respective methylenic protons are diastereo-
topic. The magnetic equivalence of the ligands could be accounted
by a facial coordination with a symmetrical all-trans configura-
tion of the NpzNamOphen donor atom sets. For Complex 6, the
determination of its solid state structure confirmed the presence
Suitable crystals of 1 were grown by slow diffusion of diethyl-
ether in an acetonitrile solution of theꢀcompound. The structure
of 1 consists of [Ga(L1)2]+ cations and PF6 anions that establish sev-
eral short intermolecular contacts involving hydrogen atoms from
the pyrazole and phenoxide rings and fluoride atoms from the
counter-ion. An ORTEP diagram of the cation of 1 is shown in
Fig. 5 and a selection of bond lengths and angles is presented in Ta-
ble 2. The coordination geometry around the Ga3+ ion is distorted-
octahedral. Both ligands are coordinated in a meridional fashion
through the oxygen atom from the phenolate group and through
the nitrogen atoms from the imine and pyrazole groups. The two
pyrazole rings, as well as the two phenoxide groups, exhibit a cis
Fig. 5. ORTEP view of Complex 1; thermal ellipsoids are drawn at the 40% probability level.

530
F. Silva et al. / Journal of Inorganic Biochemistry 104 (2010) 523–532
Table 2
Selected bond lengths (Å) and angles (°) for [Ga(L1)2]PF6 (1).
Ga(1)–O(1)
Ga(1)–N(1)
Ga(1)–N(3)
C(1)–O(1)
N(1)–C(7)
N(4)–C(19)
1.920(2)
Ga(1)–O(2)
Ga(1)–N(4)
Ga(1)–N(6)
C(13)–O(2)
N(1)–C(8)
1.912(2)
2.047(3)
2.181(3)
1.308(4)
1.479(4)
1.484(4)
2.034(3)
2.131(3)
1.320(4)
1.277(4)
1.285(4)
N(4)–C(20)
O(1)–Ga(1)–O(2)
O(1)–Ga(1)–N(3)
O(1)–Ga(1)–N(6)
O(2)–Ga(1)–N(3)
O(2)–Ga(1)–N(6)
96.41(10)
174.90(11)
86.18(10)
88.67(10)
176.75(10)
O(1)–Ga(1)–N(1)
O(1)–Ga(1)–N(4)
O(2)–Ga(1)–N(1)
O(2)–Ga(1)–N(4)
91.52(10)
89.28.(10)
89.11(10)
92.83(10)
orientation, while the imine nitrogen atoms are trans to each other.
The cis angles range from 86.2° to 96.4° and the trans angles from
174.9° to 177.8°, which indicates a small distortion of the octahe-
dral geometry as often observed for bis-tridentate Ga(III) com-
plexes. The GaAO (av. 1.916(2) Å), GaANim (av. 2.041(3) Å) and
GaANpz (av. 2.156(3) Å) bond distances can be considered normal,
taking into account the values reported for other Ga(III) complexes
containing phenoxide, imine and pyrazole coordinating groups,
respectively [5,18,21,22].
and 108
between 54 and 223
l
M, while those found for the respective ligands ranged
M (Table 3). These values confirmed that
l
we were in the presence of weak to moderately cytotoxic com-
pounds, being Complex 3 the one that showed the strongest cyto-
toxicity in both cell lines.
With the exception of 2 and HL2 in MCF-7 cells, the complexa-
tion with Ga3+ led to a statistically relevant (p < 0.0001) enhance-
ment of the cytotoxic potencies of the compounds, according to
the following ratios of the respective IC50 values: (i) MCF-7 cell
line: IC50(HL1)/IC50(1) = 6.0, IC50(HL3)/IC50(3) = 4.1; PC-3 cell line:
IC50(HL1)/IC50(1) = 2.1,
IC50(HL2)/IC50(2) = 1.6,
IC50(HL3)/
3.4. Cytotoxicity in human tumor cell lines
IC50(3) = 3.7. Albeit the observed tendency to obtain lower IC50 val-
ues for the complexes, the values that have been found for the
respective ligands are of the same order of magnitude. Therefore,
we can not exclude that the augmented cytotoxic potencies of
the complexes could be the result of a stoichiometric effect, reflect-
ing merely the presence of two molecules of ligand per each com-
plex ion [3].
As above discussed, Complex 6 has shown some tendency to re-
lease its ancillary ligand of the pyrazole-amine-phenolate type
when kept in solution. Hence, one could hypothesize that 6 could
undergo dissociation in the extracellular media, releasing the
Ga3+ cation and the HL6 ligands. Being observed such behavior,
the biological effects induced by the complex should result from
the independent action of their components, i.e. Ga3+ and the
two HL6 ligands. In order to have a better insight into this aspect,
we have evaluated the cytotoxicity of Complex 6 against cancer
PC-3 cells and human breast cancer MCF-7 cells in the presence
of transferrin (500 g/mL of culture medium). This study was run
in parallel for Ga(NO3)3, as well as for Complex 3 that has shown
a greater kinetic inertness compared to 6. The rationale to evaluate
the effect of transferrin on the cytotoxic profile of the compounds
is justified by the well known ability of this iron transport protein
to increase the cellular uptake of Ga3+ [23]. The effect of transferrin
on the biological action of the compounds was assessed using 1.6,
Contrary to complexes 1–3, 4–6 tend to undergo exchange pro-
cesses in DMSO solution with release of the respective ancillary li-
gands, as discussed above. In addition, the analytical data collected
for 4–6 could not assess properly the chemical purity of these com-
pounds. Nevertheless, we have pursued with the in vitro biological
evaluation of 4–6 searching to compare their eventual cytotoxic ef-
fects with those of 1–3. In these studies, we have not quantified the
cytotoxicity of 4–6 by IC50 measurements which have been per-
formed only for 1–3.
The cytotoxic activity of the Ga(III) complexes (1–6) and respec-
tive ligands (HL1–HL6) was evaluated in human prostate cancer PC-
3 cells and human breast cancer MCF-7 cells, within the concentra-
tion range 16 nM–160
comparative purposes, such study was also performed for
Ga(NO3)3. Up to a concentration of 160 M, this gallium salt did
lM and using a colorimetric MTT assay. For
l
not exhibit any cytotoxic effects in both cell lines, being recovered
a percentage of viable cells higher than 80% after 24 h of incuba-
tion. The Ga complexes (1–6) and respective ligands (HL1–HL6)
have shown a low to moderate cytotoxicity against PC-3 and
MCF-7 cells, being 3 and 6 the compounds that have presented
the highest ability to induce cell death in both cell lines. Fig. 6
shows the effect of different concentrations (1.6, 32 and 160 lM)
of complexes 3 and 6 on the viability of the cells, as well as of
the respective free ligands (HL3 and HL6) and cisplatin. At the
16 and 80 lM concentrations for complexes 3 and 6 and for
Ga(NO3)3, with molar ratios of the test compounds to transferrin
micromolar level (1.6
hibit the cell growth and the cell viability was not affected. By
increasing the concentration of complexes 3 and 6 to 32 M some
lM), none of the compounds was able to in-
l
cytotoxic effects started already to be observed, which became
more pronounced for the highest evaluated concentration of
160 lM. At this concentration, 3 and 6 induced a marked reduction
of the cell viability that dropped to values as low as 10% in the case
of Complex 6 and for the MCF-7 cell line. Remarkably, 3 and 6 have
shown a moderate cytotoxicity against the PC-3 cell line, which has
shown a high resistance to cisplatin even at the highest tested con-
Table 3
Cytotoxicity of the Ga(III) complexes (1–3) and their ligands (HL1–HL3) in two human
cancer cell lines.
Compound
IC50 (l
M)a
MCF-7
PC-3
Ga(III) complexes
[Ga(L1)2]PF6 (1)
[Ga(L2)2]PF6 (2)
[Ga(L3)2]PF6 (3)
Ligands
37
6
108 21
centration (160 lM), as can be seen in Fig. 6.
52 13
13.3 3.6
48
20
5
4
For complexes 1–3 and respective ligands (HL1–HL3), we have
measured their IC50 values in the PC-3 and MCF-7 cell lines. Such
measurements were done after 72 h of incubation and using con-
HL1
223 29
67 13
54 16
226 34
75 14
73 24
HL2
HL3
centrations of the several compounds in the range 0.32–160
lM.
The values determined for the complexes spanned between 13.3

F. Silva et al. / Journal of Inorganic Biochemistry 104 (2010) 523–532
531
B
A
HL3
3
100
50
0
100
50
0
HL6
6
Cisplatin
1.6
32
160
1.6
32
160
Concentration (µM)
Fig. 6. Effect of the concentration of complexes 3 and 6, respective ligands (HL3 and HL6) and cisplatin on the viability of MCF-7 cells (A) and PC-3 cells (B).
A
B
3
100
50
0
100
3 + Tf
6
6 + Tf
Ga(NO3)3
Ga(NO3)3 + Tf
50
0
16
1.6
16
80
1.6
80
Concentration (µM)
Fig. 7. Effect of the concentration of complexes 3 and 6, respective ligands (HL3 and HL6) and Ga(NO3)3 on the viability of MCF-7 cells (A) and PC-3 cells (B) in the presence of
increasing concentrations of transferrin.
Fig. 8. Images of PC-3 cells after exposure to compounds 3 and 6 for 24 h, with the indicated concentrations, and processed for apoptosis. Staining was performed with FITC-
annexin V for apoptotic cells (green) and DAPI for nuclei (blue). Propidium iodide staining (red) was also performed to identify necrotic cells but no signal was present. Taxol
was used as a positive control. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
ranging between 1:0.075 and 1:3.75. As can be seen in Fig. 7, in the
presence of transferrin the cytotoxic potencies of the complexes
were always more pronounced than those of the gallium salt, ex-
cept when a 1.6 lM concentration of the compounds was tested
in the MCF-7 cell line. In general, the presence of transferrin mod-
erately increased the cytotoxicity both for the complexes and gal-

532
F. Silva et al. / Journal of Inorganic Biochemistry 104 (2010) 523–532
lium nitrate. These finding could be explained by the extracellular
dissociation of the complexes and a facilitated transport of the re-
leased Ga3+ that contributes to increase, although in a relatively
low extent, their cytotoxicity potencies. Consistently, the cytotox-
icity of the compounds seems to be governed mainly by the pyra-
zole-containing ligands. However, one cannot exclude that the
complexes by themselves have their intrinsic mode of action.
consider that 3 and related complexes with different halogen sub-
stituents (Cl vs. Br vs. I) deserve to be further studied as potential
antitumor drugs, in order to elucidate their mechanism of action
and identify their molecular target(s). In particular, this should in-
clude the evaluation of their capability to act as proteasome inhib-
itors, since the anticancer properties of other Ga(III) complexes
with related asymmetric (N,N0,O)-donor ligands has been recently
associated with their ability to inhibit the proteasome activity [9].
3.5. Evaluation of the induction of apoptosis in human tumor cell lines
Acknowledgement
Taking into consideration the moderate cytotoxicity exhibited
by complexes 3 and 6 and respective ancillary ligands, particularly
against the cisplatin-resistant PC-3 cell line, we decide to proceed
with the study of their biological activity and we have evaluated if
these compounds induced cell death by apoptotic or necrotic pro-
cesses. Our results show that both 3 and 6 induced cell death
through apoptotic processes (Fig 8). The externalization of phos-
phatidylserine is considered an early event in apoptosis, and can
be visualized using the Annexin V protein which displays a high
affinity for phosphatidylserine in the presence of Ca2+. The usage
of Annexin V–fluorescein conjugates (green signal) allows detec-
tion of cells that have initiated apoptotic processes. Our results
indicate that both 3 and 6 induced apoptosis at the concentrations
evaluated, in agreement with the cytotoxicity results. Necrosis was
not observed, since staining with propidium iodide did not elicit
any signal. The images presented are merged images from blue
(nuclei), green (FITC-annexin V, apoptotic cells) and red (necrotic
cells) filters.
Francisco Silva wishes to thank the Fundação para a Ciência e
Tecnologia (National Foundation for Science and Technology) for
a Doctoral research Grants (SFRH/BD/6227/2001). The authors
would like to acknowledge FEDER and FCT for financial support
(PPCDT/QUI/57632/2004). L. Maria and J. Ascenso are acknowl-
edged for the 71Ga NMR spectroscopic studies. J. Marçalo is
acknowledged for the ESI-MS analyzes which were run on a QITMS
instrument acquired with the support of the Programa Nacional de
Reequipamento Científico (Contract REDE/1503/REM/2005 – ITN)
of FCT and is part of RNEM – Rede Nacional de Espectrometria de
Massa.
Appendix A. Supplementary material
Supplementary data associated with this article can be found, in
the online version, at doi:10.1016/j.jinorgbio.2010.01.003.
References
4. Conclusions
[1] M.A. Jakupec, M. Galanski, V.B. Arion, C.G. Hartinger, B.K. Keppler, Dalton
Trans. (2008) 183–194.
[2] M. Frezza, C.N. Verani, D. Chen, Q.P. Dou, Lett. Drug Des. Discov. 4 (2007) 311–
317.
[3] V.B. Arion, M.A. Jakupec, M. Galanski, P. Unfried, B.K. Keppler, J. Inorg. Biochem.
91 (2002) 298–305.
[4] A.V. Rudnev, L.S. Foteeva, C. Kowol, R. Berger, M.A. Jakupec, V.B. Arion, A.R.
Timerbaev, B.K. Keppler, J. Inorg. Biochem. 100 (2006) 1819–1826.
[5] R. Shakya, F. Peng, J. Liu, M.J. Heeg, C.N. Verani, Inorg. Chem. 45 (2006) 6263–
6268.
[6] A. Dobrov, V.B. Arion, N. Kandler, W. Ginzinger, M.A. Jakupec, A. Rufin´ ska, N.G.
Keyserlingk, M. Galanski, C. Kowol, B.K. Keppler, Inorg. Chem. 45 (2006) 1945–
1950.
[7] C.R. Kowol, R. Berger, R. Eichinger, A. Roller, M.A. Jakupec, P.P. Schmidt, V.B.
Arion, B.K. Keppler, J. Med. Chem. 50 (2007) 1254–1265.
[8] C.R. Kowol, E. Reisner, I. Chiorescu, V.B. Arion, M. Galanski, D.V. Deubel, B.K.
Keppler, Inorg. Chem. 47 (2008) 11032–11047.
[9] D. Chen, M. Frezza, R. Shakya, Q.C. Cui, V. Milacic, C.N. Verani, Q.P. Dou, Cancer
Res. 67 (2007) 9258–9265.
[10] C. Moura, R.F. Vítor, L. Maria, A. Paulo, I.C. Santos, I. Santos, Dalton Trans.
(2006) 5630–5640.
[11] M. Videira, F. Silva, A. Paulo, I.C. Santos, I. Santos, Inorg. Chim. Acta 362 (2009)
2807–2813.
[12] SADABS, Area-Detector Absorption Correction, Brukee AXS Inc., Madison, WI,
2004.
[13] SAINT, Area-Detector Integration Software (Version7.23), Brukee AXS Inc.,
Madison, WI, 2004.
[14] A. Altomare, M.C. Burla, M. Camalli, G. Cascarano, G. Giacovazzo, A. Guagliardi,
A.G.G. Moliterini, G. Polidoro, R.J. Spagna, Appl. Crystallogr. 32 (1999) 115–
119.
Chelators of the pyrazole-imine-phenol (HL1–HL3) and pyra-
zole-amine-phenol (HL4–HL6) type allowed the synthesis of well-
defined [GaL2]+ homoleptic complexes (1–6). In complexes 1–6,
the ligands act as monoanionic and (N,N0,O)-tridentate but display
different coordination modes. The more rigid L1–L3 coordinate to
the Ga3+ ion in a meridional fashion (complexes 1–3) while L4–L6
are facially coordinated (complexes 4–6). Complexes 1–3 have a
greater stability in solution compared to 4–6, which have shown
a more pronounced tendency to release the respective ancillary li-
gands. Most probably, such difference reflects the highest rigidity
of L1–L3.
The cytotoxicity of 1–6 and respective ligands (HL1–HL6)
against human prostate cancer cells PC-3 and human breast cancer
cells MCF-7 is strongly dependent on the substituents of the phe-
nolate rings. Complexes 3 and 6 that contain chloride substituents
at the phenolate rings have shown the highest cytotoxic potency,
including in the cisplatin-resistant PC-3 cell line. In both cell lines,
3 and 6 are more cytotoxic than the respective anchor ligands,
respectively HL1 and HL6, although presenting cytotoxic potencies
of the same order of magnitude. It has been also found that the
cytotoxic activity of 3 and 6 is slightly increased by the presence
of transferrin. All together, these findings could indicate that the
enhanced cytotoxicity of the complexes merely reflects a stoichi-
ometric effect and an eventual contribution of the dissociation of
the complexes. However, at this point, it can not be excluded that
the complexes have their own mode of action.
[15] G.M. Sheldrick, SHELXL-97: Program for the Refinement of Crystal Structure,
University of Gottingen, Germany, 1997.
[16] L.J. Farrugia, Appl. Crystallogr. 32 (1999) 837–838.
[17] L.J. Farrugia, ORTEP-3, J. Appl. Crystallogr. 30 (1997) 565.
[18] C. Imbert, H.P. Hratchian, M. Lanznaster, M.J. Heeg, L.M. Hryhorczuk, B.R.
McGarvey, H.B. Schlegel, C.N. Verani, Inorg. Chem. 44 (2005) 7414–7422.
[19] J.P. André, H.R. Mäcke, J. Inorg. Biochem. 97 (2003) 315–323.
[20] P. Caravan, C. Orvig, Inorg. Chem. 36 (1997) 236–248.
[21] G. Bandoli, A. Dolmella, F. Tisato, M. Porchia, F. Refosco, Coord. Chem. Rev. 253
(2009) 56–77.
The induction of cell death by complexes 3 and 6 involves
mainly apoptotic processes. Apparently, 3 is a more effective apop-
totic inducer. Complex 3, stabilized by a chloro-substituted pyra-
zole-imine-phenol ligand, has also shown the highest cytotoxic
against a cisplatin-resistant PC-3 cell line. For these reasons, we
[22] D.L. Reger, Y. Ding, Organometallics 12 (1993) 4485–4492.
[23] C.R. Chitambar, P. Boon, J.P. Wereley, Clin. Cancer Res. 2 (1996) 1009–1015.