Anti-proliferative effects of copper(II) complexes with hydroxyquinoline-thiosemicarbazone ligands

Article abstract of DOI:10.1016/j.ejmech.2017.01.031

The possibility to influence the physiological concentration of copper ions through the careful choice of ligands is emerging as a novel intriguing strategy in the treatment of pathologies such as cancer and Alzheimer. Thiosemicarbazones play an important

Full text of DOI:10.1016/j.ejmech.2017.01.031

Accepted Manuscript
Anti-proliferative effects of copper(II) complexes with Hydroxyquinoline-
Thiosemicarbazone ligands
Dominga Rogolino, Andrea Cavazzoni, Anna Gatti, Matteo Tegoni, Giorgio Pelosi,
Vincenzo Verdolino, Claudia Fumarola, Daniele Cretella, Pier Giorgio Petronini,
Mauro Carcelli
PII:
S0223-5234(17)30041-7
10.1016/j.ejmech.2017.01.031
EJMECH 9182
DOI:
Reference:
To appear in: European Journal of Medicinal Chemistry
Received Date: 29 August 2016
Revised Date: 31 October 2016
Accepted Date: 21 January 2017
Please cite this article as: D. Rogolino, A. Cavazzoni, A. Gatti, M. Tegoni, G. Pelosi, V. Verdolino, C.
Fumarola, D. Cretella, P.G. Petronini, M. Carcelli, Anti-proliferative effects of copper(II) complexes with
Hydroxyquinoline-Thiosemicarbazone ligands, European Journal of Medicinal Chemistry (2017), doi:
10.1016/j.ejmech.2017.01.031.
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
Anti-proliferative Effects of Copper(II) Complexes with
Hydroxyquinoline-Thiosemicarbazone Ligands
±
$§
±
±
±
Dominga Rogolino *, Andrea Cavazzoni , Anna Gatti , Matteo Tegoni , Giorgio Pelosi , Vincenzo
†
‡
§
§
§
±
Verdolino , Claudia Fumarola , Daniele Cretella , Pier Giorgio Petronini , Mauro Carcelli
±
Dipartimento di Chimica and Consorzio Interuniversitario di Ricerca in Chimica dei Metalli nei
Sistemi Biologici, Università di Parma, Parco Area delle Scienze 17/A, 43124 Parma, Italy
§
Department of Clinical and Experimental Medicine, University of Parma
†
Department of Chemistry and Applied Biosciences, Eidgenössusche Technische Hochschule
Zürich, CH-8093 Zurich, Zurich, Switzerland
‡
Facoltà di Informatica, Istituto di Scienze Computazionali, Università della Svizzera Italiana,
6
900 Lugano, Ticino, Switzerland
*
corresponding author
Prof. Dominga Rogolino
Dipartimento di Chimica and Consorzio Interuniversitario di Ricerca in Chimica dei Metalli nei
Sistemi Biologici, Università di Parma
Parco Area delle Scienze 17/A, 43124 Parma, Italy
dominga.rogolino@unipr.it
+39 0521 905419
$
Co-corresponding author: authors equally contributed to the present work
1

ACCEPTED MANUSCRIPT
ABSTRACT
The possibility to influence the physiological concentration of copper ions through the careful
choice of ligands is emerging as a novel intriguing strategy in the treatment of pathologies such as
cancer and Alzheimer. Thiosemicarbazones play an important role in this field, because they offer a
wide variety of potential functionalizations and different kinds of coordination modes. Here we
report the synthesis of some 8-hydroxyquinoline thiosemicarbazone ligands containing an ONN’S
donor set and their Zn(II) and Cu(II) complexes. The metal complexes are characterized in solution
and in the solid state and the X-ray structure of one of the copper(II) complex is reported. The
Cu(II) complexes were characterized also by means of quantum mechanical calculations. The
Cu(II) complexes displayed cytostatic activity in different cancer cell models. In particular, the
most active Cu(II) complex significantly inhibited cell proliferation with an IC value lower than 1
5
0
µM; this effect was associated with a block of the cell cycle in the G /M phase. This Cu(II) complex
2
induced neither the production of reactive oxygen species (ROS) nor the accumulation of p53
protein, suggesting the lack of DNA damage.
KEYWORDS: thiosemicarbazone; copper(II) complex; 8-hydroxyquinoline; proliferation; ROS;
anticancer
ABBREVIATIONS: amino-terminal copper and nickel binding (ATCUN), circular dichroism
(CD), dimethylformamide (DMF), dimethyl sulfoxide (DMSO), diethylenetriamine (Dien),
ethylenediaminetetraacetic acid (EDTA), human serum albumin (HSA), 4-(2-hydroxyethyl)-1-
piperazine ethanesulfonic acid (HEPES), poly (ADP-ribose) polymerase protein (PARP), root mean
square deviation (RMSD), reactive oxygen species (ROS), non-small-cell lung carcinoma (NSCLC)
2

ACCEPTED MANUSCRIPT
1
. Introduction
Thiosemicarbazones are a well-known class of ligands, with a broad spectrum of biological
activities, but newly, recovering the brilliant title of a recent paper, [1] they “turn from the old to
new”. α-N-heterocyclic thiosemicarbazones, in fact, seem to be promising candidates for cancer
therapy, in particular 3-aminopyridine-2-carboxaldehyde thiosemicarbazone (Triapine, Scheme
1A), which has been already investigated in numerous clinical phase trials, and some di-2-
pyridylketone thiosemicarbazones (Scheme 1B; di-2-pyridylketone-4-cyclohexyl-4-methyl-3-
thiosemicarbazone is entering clinical phase I studies) [2]. Notably, some thiosemicarbazones are
reported to overcome P-gp-mediated drug resistance [3,4]. The reason of the anti-neoplastic activity
of thiosemicarbazones is not completely understood, but it is in general connected with their
possibility to chelate intracellular iron and to interact with the activity of the enzyme ribonucleotide
reductase, which is very sensitive to the iron concentration. The chelation properties of
thiosemicarbazones can be exploited also to target copper(II) and copper(II) homeostasis [5,6]. Iron
and copper are involved in the Haber-Weiss reaction and in the production of oxygen reactive
species, so iron and copper complexes of thiosemicarbazones are supposed to act by induction of
oxidative stress [5,7], but other pathways have probably to be considered [8,9]. Triapine and di-2-
pyridylketone thiosemicarbazones are fundamentally NN’S tridentate ligands, but also N2S2
tetradentate chelating bis(thiosemicarbazone) ligands and their Cu(II) complexes (Scheme 1C) have
shown promising anticancer activity [10,11,12]. Moreover, the copper(II) complex of diacetyl
bis(thiosemicarbazone) has shown hypoxia selectivity and it seems interesting as
radiopharmaceutical for the imaging of hypoxic tissues [13]. Redox processes are involved also in
the antineoplastic action of bis-thiosemicarbazones. The activity of the free ligands is attributed to
their chelation properties towards essential ions as copper, so that the active species is the neutral
copper(II) bis(thiosemicarbazone) complex. It enters the cell and, once inside, Cu(II) is reduced to
3

ACCEPTED MANUSCRIPT
Cu(I); finally, reduction is accompanied by dissociation of the metal ion and general poisoning of
the cell [12].
In medicinal chemistry, 8-hydroxyquinoline is a very interesting pharmacophoric group: it is a
“privileged structure” [14], whose biological properties are generally related to its chelating abilities
[15]. Some 8-hydroxyquinoline derivatives [16] and also some of their metal complexes [17] have
shown in vitro cytotoxicity against human cancer cell lines.
In light of these considerations and of some previous results [18], we decided to study the chelating
properties of 8-hydroxyquinoline thiosemicarbazones 1-3 (Scheme 2), in particular towards
copper(II) and, for comparison, also towards the non-redox-active zinc(II) ion. These ligands are
diprotic, as the bis-thiosemicarbazones, but ONN’S tetradentate. The corresponding complexes 4-9
(Scheme 2) were characterized in solution and in the solid state, and, for the copper(II) complexes,
also by means of quantum mechanical calculations and X-ray diffraction. Finally, in vitro
antineoplastic properties are reported, together with some preliminary investigations on the
mechanism of action of the copper(II) complexes.
2
. Experimental
2.1
Materials and methods. Chemistry. All reagents of commercial quality were purchased from
Sigma-Aldrich and used without further purification. The purity of the compounds was determined
by elemental analysis and verified to be ≥ 95% for all synthesized molecules. NMR spectra were
recorded at 25 °C on a Bruker Avance 400 FT spectrophotometer. The ATR-IR spectra were
recorded by means of a Nicolet-Nexus (Thermo Fisher) spectrophotometer by using a diamond
-1
crystal plate in the range of 4000-400 cm . Elemental analyses were performed by using a FlashEA
112 series CHNS/O analyzer (Thermo Fisher) with gas-chromatographic separation. Electrospray
1
mass spectral analyses (ESI-MS) were performed with an electrospray ionization (ESI) time-of-
flight Micromass 4LCZ spectrometer. MS spectra were acquired in positive EI mode by means of a
4

ACCEPTED MANUSCRIPT
DEP-probe (Direct Exposure Probe) mounting on the tip of a Re-filament with a DSQII Thermo
Fisher apparatus, equipped with a single quadrupole analyzer. UV–Vis spectra were recorded on an
Evolution 260 Bio Thermo spectrophotometer by using cells of 1 cm path length.
2
2
.2
Synthesis
.2.1 Synthesis of 1-3. The thiosemicarbazone ligands were synthesized by condensation reaction
of the proper thiosemicarbazide with 8-hydroxyquinoline-2-carboxaldehyde [18, 19 ].
Thiosemicarbazide (1 mmol) was dissolved in warm ethanol (15 mL) and added of a solution of 8-
hydroxyquinoline-2-carboxaldehyde (1 mmol, 0.173 g) in ethanol (15 mL). The resulting mixture
was heated under reflux for 4 h; the precipitate was subsequently filtered, washed with cold ethanol
and dried under vacuum.
1
8
-Hydroxyquinoline-2-carboxaldehyde thiosemicarbazone (1). Yellow powder. Yield = 89%. H-
NMR (DMSO-d , 25°C), δ: 11.84 (s, 1H, NH), 9.84 (s, 1H, HO), 8.42-8.44 (2H, HC=N+ArH);
6
1
8
.31-8.27 (d+s, 3H, ArH+NH ); 7.37-7.44 (m, 2H, ArH); 7.10 (d, 1H, ArH, J = 7.1 Hz). H-NMR
2
(
MeOD-d , 25°C), δ: 8.24 (3H, HC=N+ArH); 7.47 (t, 1H, ArH, J = 8.0 Hz); 7.39 (d, 1H, ArH, J =
4
+
8
.4 Hz) 7.13 (d, 1H, ArH, J = 8.2 Hz). MS (EI, 70 eV, positive ions) m/z (%) = 245.9 ([M] , 100).
-1
IR (cm ): ν +νNH2= 3383, 3230; νNH = 3145; νC=C +δN-H + νC=N = 1605; νC=S= 1083, 840.
OH
8-Hydroxyquinoline-2-carboxaldehyde–4-ethyl-3-thiosemicarbazone (2). Yellow powder. Yield =
1
6
9%. H-NMR (DMSO-d , 25°C), δ: 11.85 (s, 1H, NH), δ 9.84 (s, 1H, HO), 8.84 (s, 1H,NH), 8.42
6
(d, 1H, ArH, J = 8.4 Hz), 8.31 (d, 1H, ArH, J = 8.3 Hz), 8.26 (s, 1H, HC=N), 7.38-7.46 (m, 2H,
ArH), 7.11 (d, 1H, ArH, J = 7.1 Hz), 3.64 (q, 1H, CH , J = 7.3 Hz), 1.19 (t,1H, CH , J = 7.4 Hz).
2
3
+
-1
MS(EI, 70 eV, positive ions) m/z (%) = 274.1 ([M] , 100). IR (cm ): ν +ν
= 3400, 3298; νNH =
OH
NH2
3
8
8
142; νC=N = 1544; νC=S= 1098, 838.
-Hydroxyquinoline-2-carboxaldehyde–4-phenyl-3-thiosemicarbazone (3). Yellow powder. Yield =
1
1%. H-NMR (DMSO-d , 25°C), δ: 12.24 (s, 1H, OH), 10.37 (s, 1H, HN), 9.88 (s,1H, NH), 8.61
6
(d, 1H, ArH, J = 8.6 Hz), 8.37 (s, 1H, HC=N), 8.30 (d, 1H, ArH, J = 8.3 Hz), 7.58-7.38 (m, 6H,
5

ACCEPTED MANUSCRIPT
1
ArH), 7.27 (t, 1H, ArH, J = 7.2 Hz), 7.13 (d, 1H, ArH, J = 7.1 Hz). H-NMR (MeOD-d , 25°C), δ:
4
8
1
.32-8.26 (m, 3H, ArH+ HC=N), 7.67 (d, 2H, ArH, J = 7.6 Hz), 7.47-7.39 (m, 4H, ArH), 7.26 (t,
H, ArH, J = 7.4 Hz), 7.14 (d, 1H, ArH, J = 7.6 Hz). MS-ESI (positive ions) m/z (%) = 323
+
+
+
-1
(
[M+H] , 100); 345 ([M+Na] , 90) 361 ([M+K] , 50). IR (cm ): ν +νNH2= 3408, 3320; νNH =
OH
3138; νC=N = 1534; νC=S= 1091, 836.
2.2.2 Synthesis of the complexes 4-9. 0.1 g (0.4 mmol) of ligand were dissolved in degassed
methanol under nitrogen and the pH was adjusted to 8-9 by adding NaOH 1M, resulting in an
orange solution. An equimolar amount of Cu(CH COO) 2H O dissolved in methanol was added.
3
2
2
Immediately, a dark precipitate formed, and the suspension was stirred for 2 hours at r.t.. The dark
red powder was filtered off and washed with methanol.
The synthesis of Zn(II) complexes 7-9 proceeded in the same way, but by using Zn(CH COO)2
3
2
H O and ethanol as solvent.
2
+
-
(
4). Dark red powder. Yield: 67%. MS-ESI (positive ions) m/z (%) = 308 ([CuL+H] , 100). IR (cm
1
)
: νNH2 = 3246; νC=C +δN-H +νC=N = 1620; ν = 1155, 838. Anal. Calcd. for C H CuN OS 0.5H O:
CS 11 8 4 · 2
C 41.70; H 2.86; N 17.68. Found: C 41.47; H 2.80; N 17.40.
+
(
5). Dark red powder. Yield: 89%. MS-ESI (positive ions) m/z (%) = 336 ([CuL+H] , 100); 358
+
+
+
-1
(
[CuL+Na] , 50); 673 ([Cu L +H] , 20); 695 ([Cu L +Na] , 40). IR (cm ): ν = 1571; ν = 1148,
2 2 2 2 C=N CS
8
1
34. Anal. Calcd. for C H CuN OS H O: C 44.12; H 3.99; N 15.83. Found: C 44.61; H 4.33; N
13 12 4 · 2
5.60. Crystals suitable for X-ray diffraction analysis were obtained by slow evaporation of a
methanolic solution of the powder.
+
(
6). Dark red powder. Yield: 76%. MS-ESI (positive ions) m/z (%) = 384 ([CuL+H] , 100); 769
+
+
-1
(
[Cu L +H] , 60); 1152 ([Cu L +H] , 50). IR (cm ): ν = 1595; ν = 1123, 827. Anal. Calcd. for
2 2 3 3 C=N CS
C H N OSCu 0.5 H O: C 51.97, H 3.33, N 14.26. Found: C 52.03, H 3.11, N 14.29.
17
12
4
2
+
(
7). Orange powder. Yield: 67%. MS-ESI (positive ions) m/z (%) = 269 ([H L+Na] , 100); 309
2
+
1
(
[ZnL+H] , 80). H-NMR (MeOD-d , 25°C), δ: 8.41 (d, 1H, ArH, J = 8.42 Hz), 8.13 (s, 1H,
4
HC=N), 7.65 (d, 1H, ArH, J = 7.65 Hz), 7.39 (t, 1H, ArH, J = 7.40 Hz), 7.06 (d, 1H, ArH J = 7.06
6

ACCEPTED MANUSCRIPT
-1
Hz), 6.87 (d, 1H, ArH, J = 6.89 Hz). IR (cm ): ν
= 3435, 3366; νC=C +δN-H +νC=N = 1592; νCS=
NH2
1
096, 837. Anal. Calcd. for C H N OSZn 0.5 C H OH: C 43.32; H 3.33; N 16.84. Found: C 43.40;
11 8 4 · 2 5
H 3.25; N 16.98.
+
(
8). Orange powder. Yield: 92%. MS-ESI (positive ions) m/z (%) = 241 ([H L-S] , 100); 336
2
+
-1
(
[ZnL+H] , 70). IR (cm ): ν
= 3331; νC=N = 1594; ν = 1099, 836. Anal. Calcd. for
NH2 CS
C H N OSZn 0.5 C H OH: C 46.61; H 4.19; N 15.53. Found: C 46.60; H 4.03; N 15.82.
13
12
4
·
2
5
+
(
9). Orange powder. Yield: 72%. MS-ESI (positive ions) m/z (%) = 291 ([H L-S] , 100); 385
2
+
1
(
[ZnL+H] , 50). H-NMR (MeOD-d , 25°C), δ: 8.43 (d, 1H, ArH, J = 8.42 Hz), 8.31 (s, 1H,
4
HC=N), 7.69-7.65 (m, 3H, ArH), 7.43 (t, 1H, ArH, J = 7.40 Hz), 7.33 (m, 2H, ArH), 7.09-7.07 (m,
-1
2
H, ArH), 6.89 (d, 1H, ArH, J = 7.42 Hz). IR (cm ): ν
= 3321; νC=N = 1594; ν = 1097, 837.
NH2 CS
Anal. Calcd. for C H N OSZn 0.5 H O: C 51.72; H 3.32; N 14.19. Found: C 52.10; H 3.49; N
1
7
12
4
·
2
13.85.
2.3
Studies in solution
Diethylenetriamine (Dien) trihydrochloride was prepared by dissolving high purity Dien in a small
amount of ethanol; addition of few drops of concentrated HCl resulted in the appearance of a white
microcrystalline precipitate which was filtered off and dried (purity > 99 % by potentiometric
analysis). A Thermo Orion 720A pH-meter connected with a Hamilton glass electrode was used for
pH measurements and potentiometric data collection. A 0.1 M KCl solution in methanol:water 9:1
(v/v) was used to fill the reference compartment of the electrode. The HEPES 25 mM buffer
methanol:water 9:1 (v/v) solution at pH 7.4 was prepared as follows: solid HEPES (0.59 g) was
suspended in 100 mL of a methanol:water 9:1 (v/v) mixture. Few drops of concentrated (10 N)
aqueous NaOH solution were added until pH 7.4 was reached and complete dissolution of the solid
was observed. Calibration of the glass electrode using buffers in methanol:water 9:1 (v/v) solutions
was performed immediately before its use [20]. Stock solutions of the ligands (C = 1-1.1 mM, L =
L
1
-3) were prepared in DMF and used within few days. Stock solutions of CuCl and ZnCl in water
2 2
(
C ca. 0.016 M, C ca. 0.018 M) were prepared by weight of the salts and their titre determined
Cu Zn
7

ACCEPTED MANUSCRIPT
using standardized EDTA solutions [21] Titrant metal solutions were obtained by dilution of the
stock solutions in methanol:water 9:1 (v/v), and prepared at ca. 4 mM concentration. Titrant
solution of Dien (CDien ca. 11 mM) was prepared by weight from solid trihydrochloride salt. The
salt was dissolved in HEPES 25 mM buffer methanol:water 9:1 (v/v) solution, and the pH corrected
to 7.4 by addition of small aliquots of concentrated (10 N) aqueous NaOH solution.
2.3.1 UV-visible spectrophotometric titrations. The UV-vis spectra were collected using a Thermo
Evolution 260 Bio spectrophotometer provided with a Peltier thermostat, and quartz cuvettes with 1
cm path length. Speciation diagrams were calculated using the HySS 2009 software [22]. The
2
+
spectrophotometric titrations of the ligands 1-3 with Zn were carried out as follows. Solutions of
0
the ligands (C = 4-41 uM) were prepared in the cuvette by diluting with HEPES 25 mM buffer
L
methanol:water 9:1 (v/v) solution (pH 7.4) a proper amount of the ligands stock solutions in DMF.
2
+
The obtained ligand solutions were titrated with Zn titrant solutions up to Zn:L 1.71 – 2.29. The
2
+
Cu /L (L = 1-3) systems could not be studied by direct titrations with the metal since the stability
constants of the copper(II) complexes are too high (vide infra). Competition UV-vis titration
experiments were instead carried out using Dien as competing ligand. Sample solutions containing
2
+
Cu and the ligands in 1:1 ratio were prepared in the cuvette, in HEPES 25 mM buffer
methanol:water 9:1 (v/v) solution (pH 7.4). These solutions were titrated with the Dien titrant
2+
solution, up to Dien:Cu 13:1. Due to the relative slow attainment of the equilibrium conditions,
each spectrum was collected 7 min after titrant addition and mixing. The logarithms of the
conditional stability constants were calculated from the spectral dataset using HyperQuad 2006
software [23]. For each system, data from different titrations were treated together. Speciation
diagrams were calculated using the HySS 2009 software [22]. Continuous variation (Job’s)
experiments were carried out by preparing 11 solutions with constant CCu/Zn + C (ca. 41 µM; L =
L
1
-3), and variable C /(C +CCu/Zn) molar fractions in the 0 – 1 range. All samples were prepared in
L L
HEPES 25 mM buffer methanol:water 9:1 (v/v) solution at pH 7.4.
8

ACCEPTED MANUSCRIPT
2
+
2.3.2 Study of stability under aerobic condition. The stability under aerobic conditions of the Cu
2
+
complexes with 1-3 was studied by preparing Cu :ligand solutions (1:1 ratio) directly in the
cuvette (CCu ca. 40 uM) by adding proper amounts of copper(II) and ligands stock solutions.
HEPES 25 mM buffer methanol:water 9:1 (v/v) solution at pH 7.4 was used as medium. The
cuvettes were sealed with a rubber stopper. A needle was inserted in the rubber stopper to allow
maintain aerobic conditions in the cuvette and limiting the evaporation of the solvent. The solutions
were left under stirring in the presence of atmospheric dioxygen, and the spectra registered every 30
min. for 10 h. There is a negligible variation of the spectra within this time frame.
2
+
2
.3.3 Potentiometry. The potentiometric titrations of Dien and of the Cu /Dien systems were
-1
carried out in methanol:water 9:1 (v/v) mixtures at T = 298.2 ± 0.1 K and I = 0.1 mol L (KCl)
under a N stream, using 50 mL samples. The potentiometric apparatus for the automatic data
2
acquisition was already described [24,25]. The KOH solution used for potentiometric titrations was
prepared by diluting a 10N KOH aqueous solution (Merck) to obtain a ca. 0.1 N KOH solution in
methanol:water 9:1 (v/v). The titrant KOH solution was standardized in methanol:water 9:1 (v/v)
2
+
against potassium hydrogen phthalate. The Cu solution was prepared as reported above (stock
solutions of CuCl ). Both methanol and water used to prepare the samples for potentiometric
2
titrations were boiled, cooled and stored under N atmosphere, and used within few days. The
2
+
electrode was calibrated in terms of [H ] by titrating HCl solutions with a standard KOH solution
[26]. The Dien titrant samples were prepared by dissolving a weighted amount of the solid Dien
-3
-1
trihydrochloride salt in methanol:water 9:1 (v/v). Three samples (6.0–6.6·10 mol L ) were titrated
2
+
with KOH in the pH range 3.6-11.1. The Cu /Dien samples were prepared using a weighted
2
+
amount of solid Dien trihydrochloride salt, to which aqueous Cu was added, followed by proper
amounts of methanol and water to afford the final titrant samples in methanol:water 9:1 (v/v). Three
2
+
.
-3
-1
titrations were carried out with Dien:Cu molar ratios from 2.5 to 3.1 (C = 2.2-2.6 10 mol L )
Cu
2
+
in pH range 3.1-11.2. The protonation and Cu complexation constants were calculated from the
2
3
potentiometric data using the HyperQuad 2009 program. Least-squares treatment was performed
9

ACCEPTED MANUSCRIPT
o
c 2
o
c
by minimization of the sample standard deviation σ = [Σ w (E – E ) /(n – m)]½, where E and E
i
i
i
i
i
i
are the observed and calculated e.m.f. values, respectively, n is the number of observations and m is
2
the number of refined parameters. The statistical weights w were put equal to 1/σ , where σ is the
i
i
i
expected error on each observed e.m.f. value (0.2 mV). A pK value of 14.41, determined from
w
glass electrode calibration, was employed. For each system, data from different titrations were
2
+
2+
treated together. The conditional formation constants of the [Cu(Dien)] and [Cu(Dien)2]
2
2
complexes at pH = 7.4 were calculated by the HySS 2009 program, taking into account the
2
+
formation of hydrolytic species of Cu [27,28].
.3.4 Interaction of the complex 4 with albumin. The study of the interaction of the Cu(II)
2
complex 4 with albumin has been carried out in aqueous buffer HEPES solution 25 mM pH 7.4. A
stock solution of commercial WT human serum albumin (HSA, Sigma) was prepared by weight at
300 µM concentration in the buffer solution. Due to the limited solubility in water, the stock
solution of 4 (25 mM) was prepared by weight in DMSO. A 25 mM solution of CuCl ·2 H O was
2
2
also prepared by weight in DMSO. The samples for CD data collection in the visible range were
prepared as follows: 2 ml of the HSA stock solution were added with either 10 µl of 4 or 10 µl of
CuCl solutions (control). The reference HSA spectrum was collected on a sample obtained by
2
adding 10 µl DMSO to 2 ml of the HSA stock solution. The HSA:Cu ratio was 2.4. The samples for
CD data collection in the UV range were prepared as follows: 250 µl of the HSA stock solution
were added with either 2.5 µl of 4 or 2.5 µl of CuCl solutions (control). The reference HSA
2
spectrum was collected on a sample obtained by adding 2.5 µl DMSO to 250 µl of the HSA stock
solution. The HSA:Cu ratio was 1.2. The blank spectrum was collected on a 250 µl HEPES buffer
solution, added with 2.5 µl of DMSO. The spectra were collected on a Jasco J-715 UV-visible
spectropolarimeter provided with a Peltier thermostat. Matched quartz cells of 10 or 1 mm path
length were used for collecting spectra in the visible or UV range, respectively. The temperature
was set to 25.0 °C.
1
0

ACCEPTED MANUSCRIPT
2.4 Crystallography.
A single crystal of 5 was mounted on a glass fibre and the intensity data were collected with a
SMART APEX2 diffractometer with Bruker AXS CCD detector using Mo-Ka radiation and a
graphite crystal monochromator [λ(Mo-Kα) = 0.71073 Å]. The SAINT [29] software was used for
integrating reflection intensities and scaling, and SADABS [30] for absorption correction. The
structure was solved by direct methods using SIR97 [31] and refined by full-matrix least-squares on
all F2 using SHELXL97 [32] implemented in the WinGX package [33]. All the non-hydrogen
atoms in the molecules were refined anisotropically. The hydrogen atoms were partly found and
partly placed in the ideal positions using riding models. The structure was solved by direct methods
and difference Fourier synthesis using the SHELX suite of programs as implemented within the
WINGX software. Thermal ellipsoid plots were generated using the program ORTEP-333
integrated within the WinGX suite of programs.
The crystal of 5 is monoclinic, space group P2 /c; cell parameters: a = 6.9669(6), b = 13.063(1), c =
1
3
1
5.831(1) Å, β = 102.595(1)°, V = 1406.1(4) Å . The asymmetric unit is formed by a single
-
3
-1
molecule of formula C H CuN OS, Mr = 335.87, Z = 4, Dc = 1.59 g cm , µ = 1.756 mm , F(000)
1
3
12
4
=684. A semi-empirical absorption correction, based on multiple scanned equivalent reflections, has
been carried out and gave 0.5934 < T < 0.7454. A total of 17393 reflections were collected up to a θ
range of 26.5° (±8 h, ±16 k, ±19 l), 2914 unique reflections (Rint = 0.049). All non-hydrogen atoms
were refined anisotropically. The hydrogen atoms were placed in ideal positions and refined using
riding models.
CCDC1483582
http://www.ccdc.cam.ac.uk/data_request/cif; see also ESI).
.5 Calculations.
contains
the
supplementary
crystallographic
data
of
5
(
2
Molecular orbital calculations were carried out using the D01 development version of Gaussian09
series of programs [34]. Two sets of geometries have been employed for electronic structure
calculations: 1) pure square planar CuL complexes (SET-1); 2) Jahn-Teller distorted octahedral
1
1

ACCEPTED MANUSCRIPT
CuL complexes with two coordinating water molecules (SET-2). Several different DFT levels and
basis sets have been employed: the hybrid functional B3LYP density functional [35,36,37] using
the 6-31++G** (H. C. N. O. S) plus ECP-Lanl2DZ (Cu) (BS-I) and the hybrid functional of Truhlar
and Zhao M06 density functional method [ 38 ] using the 6-31++G** (H. N. C. O. S)
[39,40,41,42,43 44,45,46,47,48] and Los Alamos ECP plus TZ - Lanl2TZ(f) (Cu) [49,50,51] basis
set (BS-II). Geometry SET-1 has been investigated both at BS-I and BS-II level, whereas SET-2 has
been treated at the level BS-II only. Optimized geometries and energies in the gas phase were
computed with tight convergence criteria and the “nosymm” options were used for all
optimizations. Cartesian coordinates in mol2 format for the optimized geometries and electronic
energies for all compounds are provided in the Supporting Information. Vibrational frequencies
were computed in the gas phase at the same level of theory and were used without scaling since the
MO6 frequencies agree quite well with experimental values for a wide range of second and third
period compounds [ 52 ]. Thermal corrections and free energies were calculated by standard
statistical thermodynamic methods53 using the unscaled DFT frequencies and the ideal gas/rigid
rotor/harmonic oscillator approximations. Population analysis has been calculated within the
Natural Bond Order approximation [54,55,56,57].
2
2
.6
Biology.
.6.1 Cell culture. Human non-small cell lung cancer (NSCLC) cell line A549, human breast
cancer cell line MCF-7 and human MSTO mesothelioma cell line were obtained from American
Type Culture Collection (ATCC) (Manassas, VA, USA) and cultured as recommended at 37°C in a
humidified atmosphere of 5% CO and 95% air. Stock solutions of 10 mM compounds were
2
prepared in DMSO and diluted in fresh medium for use. The final concentration of DMSO never
exceeded 0.1% v/v.
2.6.2 Analysis of cell proliferation and cell cycle. Cells were seeded into 96-well plate and
exposed to various treatments. After 72 hours, cell proliferation was evaluated by crystal violet
staining as previously described [58]. Briefly, crystal violet was solubilized in PBS with 0.2%
1
2

ACCEPTED MANUSCRIPT
TritonX-100 and the absorbance was measured at 570 nm. Data are expressed as percent inhibition
of cell proliferation versus untreated cells. Distribution of cell population in the cell cycle was
determined by propidium iodide staining and flow cytometry analysis [59].
2
.6.3 Western blot analysis. Procedures for protein extraction, solubilisation and analysis by 1-D
PAGE are described elsewhere [60]. Briefly 30-50 µg of proteins from cell lysates were resolved by
.5% SDS-PAGE and transferred to PVDF membranes. Membranes were incubated with anti-
7
phospho-mTOR (Ser2448); anti-mTOR; anti-phospho-P70S6K (Thr421/Ser424); anti-P70S6K;
anti-phospho-AKT (Ser473); anti-AKT; anti Cyclin B1 (Cell Signaling Technology, Beverly, MA,
USA); anti-HSP72 (Stressgene Bioreagents, BE); anti-p53 (Santa Cruz Biotechnology, USA); anti-
Actin (Sigma–Aldrich, St Louis, MO, USA). Blots were then washed and incubated with secondary
HRP-coniugated antibodies (Pierce, Rockford, IL, USA). Immunoreactive bands were visualized
TM
using an enhanced chemiluminescence system (Immobilion
Western Cemiluminescent HRP
Substrate, Millipore, USA). The chemiluminescent signal was acquired by C-DiGit® Blot Scanner
and the spots were quantified by Image Studio™ Software, LI-COR Biotechnology (Lincoln, NE).
2
.6.4 Reactive oxygen species measurements. Intracellular hydrogen peroxide H O and
2 2
-
superoxide anion O were assessed by oxidation of the cell permeable fluorescent probes 5-(and-6)-
2
chloromethyl-20,70-dichlorodihydrofluorescein diacetate, acetyl ester (CM-H2DCFDA, Molecular
Probes®), and dihydroethidium (DHE, Merck Millipore), respectively as previously described [61].
5
Briefly, 3 x 10 cells were incubated in absence or presence of the compound. During the last 30
min of treatment, the cells, loaded with CM-H2DCFDA 5 µM or DHE 5 µM, were left in the dark
at 37 °C, and then trypsinized and analysed on a Beckman-Coulter EPICS XL flow cytometer.
2.6.5 Statistical analysis. Statistical analysis was carried out using GraphPad Prism 5.00 software.
Statistical significance of differences among data was estimated by two-tailed Student’s t test and P
values are indicated where appropriate in the figures.
3
. Results and Discussion
1
3

ACCEPTED MANUSCRIPT
3.1
Synthesis. The thiosemicarbazone ligands 1-3 (Scheme 2) were synthesized in high yields
by condensation of thiosemicarbazide or a 4-N-substituted-3-thiosemicarbazide with 8-
1
hydroxyquinoline-2-carboxaldehyde. They were satisfactorily characterized by H-NMR, IR, mass
spectrometry and microanalysis. The ligands are in the E form in solution, as evidenced by the
1
chemical shift values of the HC=N and NH proton in the H-NMR spectrum in DMSO-d [19,62].
6
1-3, in principle, can exhibit thione–thiol tautomerism: the absence of the S–H stretching band near
–1
–1
2
600 cm and the presence of a N–H stretching band at approximately 3150 cm indicate that the
thione form is the stable one; this is also confirmed by the ν(C=S), coupled with N=C=N stretching
-1
vibrations, observed at about 1100 cm as a strong absorption [63].
The copper(II) and zinc(II) metal complexes 4-9 (Scheme 2) were obtained by reacting the acetate
salts with the ligands 1-3 in ethanol or methanol, adjusting the pH at 8-9 with NaOH. In the IR
-1
-1
spectra of 4-9 the absorptions of the NH and OH groups (around 3100 cm and 3400 cm in the
ligand, respectively) disappeared, evidently because the ligand deprotonated upon complexation. In
–
1
the spectra of the ligands the bands around 1530-1600 cm were assigned to the iminic bond
–
1
stretching vibrations; they underwent shifts of 15-60 cm upon coordination, indicating the
1
involvement of this group in coordination. In the H NMR of the zinc(II) complexes 7 and 9 in
DMSO-d , a partial decoordination of the ligand and the presence of several chemical species were
6
observed. On the other hand, in methanol-d , there is a unique set of signals with a shift of the
4
protonic resonances of about 0.2-0.3 ppm vs the free ligand. Elemental analysis and ESI-mass
analysis indicated the presence of a complex with a 1:1 metal to ligand ratio. Overall, the data
suggested the formation of a 1:1 metal chelate for all the complexes 4-9, with the bis-deprotonated
ligand behaving as an ONN’S donor system. In the case of 5, the proposed structure was confirmed
by X-ray diffraction analysis.
3.2
X-ray discussion. The molecular structure of copper(II) complex 5, as determined by single-
crystal X-ray crystallography, is reported in Figure 1.
1
4

ACCEPTED MANUSCRIPT
The geometry of the Cu(II) coordination is square planar with the four donor atoms deriving from
the dianionic tetradentate ligand, and namely the deprotonated phenolic oxygen, the nitrogen of the
substituted quinoline ring, the iminic nitrogen and the thiolate sulfur. The coordination plane can be
described as a 5-5-5 (O,N,N,S) chelate ring system. The copper atom lies exactly on the average
plane formed by the four donor atoms, but its geometry is slightly distorted from the ideal square
planar being the bite angles N2–Cu–S1 85.5(2)°, N2-Cu1-N1 78.27(2) and N1–Cu–O1 82.3(2)°.
The Cu–S bond distance, 2.220(2) Å, and Cu–N(iminic) distance, 2.011(6) Å, are similar to those
observed in analogous systems [19, 64 ]. Upon coordination, the thiosemicarbazone moiety
undergoes deprotonation and gives an overall dianionic ligand. Thiosemicarbazones normally
present a thione-thiol tautomerism and the deprotonation, concomitant to the coordination of the
metal centre, brings to the stabilization of the thiolate form. This is apparent, in our case, from the
length of the C11–S1 distance of 1.760(8) Å that suggests a thiol character, since it is much longer
than the distances observed in free thiosemicarbazones (the average value of 431 non deprotonated
structures found in the CSD [65,66] is C– S = 1.63(1) Å). The N3–C11 bond distance of 1.320(9)
Å, accordingly, is markedly shorter than the average 1.35(1) Å of the free thiosemicarbazones of the
same set, suggesting a double bond-like character.
The packing in the crystal is characterized by two main features: the hydrogen bond that connects
the terminal amino nitrogen and the phenolic oxygen of an adjacent complex (Figure SI1 and SI2),
and the formation of stackings with π-π interactions of the aromatic systems. The first interaction
leads to the formation of ribbons of molecules developing in the b direction of the unit cell, while
the stackings extend in the a direction.
3.3
Quantomechanical calculation. We first wanted to investigate the electronic effects of the R
groups of the ligands 2 and 3 (Scheme 2) on the structural properties of their complexes (Cu(II)EtL
and Cu(II)PhL, respectively). We fully optimized the square planar structures (SET-1) generated by
2
+
complexation between Cu and the ligands, substituting the functional group R = NHC H in 2
2
5
with R = NHCH (Cu(II)MetL) for simplicity; we left methanol co-crystallization molecules as
3
1
5

ACCEPTED MANUSCRIPT
reported in one structure in the literature [19]. Despite the different electron withdrawal properties
of the R groups, a substantially unmodified chelation ring was obtained (Figure SI3). Then, we
considered the square planar structures optimized starting from a distorted octahedral geometry
obtained placing two water molecules in the axial positions (SET-2). Figure 2 shows the optimized
structures of the aquo complexes Cu(II)MetL (a), Cu(II)EtL (b), and Cu(II)PhL (c).
9
2+
As expected for a d transition metal (i.e. Cu ) in all the three complexes the initial octahedral
geometry relaxes to a square planar, with the originally coordinated water molecules now loosely
interacting either with the metal or the alkoxy group. The square planar configuration is also
supported by the molecular orbital analysis that, in agreement with a doublet spin multiplicity,
shows a remarkable spin polarization effect stabilizing the alpha electrons more than the beta ones.
9
(
Figure SI4). In agreement with a d square planar coordination the hole in the Cu d orbitals is
located in the high-energy dx2-y2(Cu)+σ(ligand) (Figure SI4) and the HOMO (Figure SI5) is
localized on the atoms involved in the metallic coordination with an overall dxz symmetry.
Moreover, the HOMO does not involve the water molecules nearby the metallic center, that would
bring the square planar coordination to a square pyramidal. Lastly, the LUMO is characterized by a
marked dx2-y2 orbital associated to the spin beta vacancy and a substantially ligand-dominant
character for the spin alpha component.
The structural analysis of the complexes reported in Figure 2 interestingly shows that both with the
mild electron donors (i.e. –CH and –CH CH ) and the withdrawal (i.e. Ph) substituents, the
3
2
3
2
thioamide nitrogen results substantially planar (i.e. sp hybridized). This is supported by the co-
planarity between C12-N4(H)-C11 resulting in 179.95, 175.76 and 178.8° dihedral angles,
respectively for –CH , –CH CH and Ph. Furthermore, the average N4-C11 distance (1.35 Å) is
3
2
3
6
7
compatible with primary amides (1.325 Å, bond order 1.2). This partial double bond can generate
E-Z isomers. Despite experimental observation that characterized the E isomer (complex 5), DFT
calculations showed that the E isomer is more stable than the Z one by only 0.5 kcal/mol, that is
below the accuracy of the method. Still, an approximate estimation of the isomerization process
1
6

ACCEPTED MANUSCRIPT
revealed, as expected, significantly high energy (~10 kcal/mol) barrier justifying a possible kinetic
control. In Table 1 we compare the most relevant structural parameters obtained by the X ray
diffraction analysis of 5, with the relative calculated at the DFT-BS(I) and BS(II) level of accuracy.
In the X-ray structure of 5 there are three five-membered chelation rings. The data calculated for
this structure, with an E configuration around the amide bond, are reported in the DFT-BS(I) 5(E)
and DFT-BS(II) 5(E) columns. Structures with 5,6,4-membered chelation rings are also known
[
19,68]; the structural details calculated assuming this isomer, with the thio- moiety involved in a 4
membered chelation to the metal, are reported in the DFT-BS(I) 4(E) column. Both computational
levels (i.e. BS-(I) and BS-(II) ) predict with good accuracy the X-ray structure of 5; in fact,
extremely low RMSDs (root mean square deviation) on the atomic distances (D) and on the angular
amplitudes (A) with respect the experimental data are obtained.
The DFT-BS(I) 4(E) structure obviously shows significant deviation from the experimental one,
mostly ascribed to the Cu-S length which increases from ~2.35 Å to ~2.5 Å due to the more
constrained four-membered ring. As a consequence, the iminic nitrogen compensates the partial
positive charge created with a shorted Cu-N bond length (i.e. from ~2.08 Å to ~1.97 Å).
Interestingly the two structures (with 5,5,5 membered or with 5,6,4 membered chelation rings) do
not significantly differ in energy, leaving open the discussion for possible isomerization in solution
or in the crystallization process [19].
2
+
3.4
Studies in solution. The conditional stability constants (hereafter indicated as β) of the Zn
complexes with the thiosemicarbazones 1-3 were determined in methanol:water 9:1 (v/v) (25 mM
HEPES buffer, pH 7.4) by direct spectrophotometric UV-visible titrations of solutions of the
2
+
2+
ligands with a Zn solution. A representative spectral set for the titration of 2 with Zn is reported
2
+
in Figure 3 (for 1 and 3 with Zn see Figures SI6 and SI7). The spectrum of the ligands presents
2
+
two bands at 308 and 343 nm, which decrease upon addition of Zn . Two bands at 327 and 395 nm
2
+
associated to the formation of complex species appeared upon addition of Zn , and an isosbestic
point is present at 371 nm.
1
7

ACCEPTED MANUSCRIPT
2
+
The absorbance values at 395 nm for the titration of 2 as a function of the equivalents of Zn added
2
+
shows a titration end-point that is consistent with the addition of 1 eq. of Zn per ligand,
suggesting the presence of a Zn + L = [ZnL] complexation equilibrium (charges omitted, L = 1 – 3;
SI8-10). Plots of the absorbance values at 395 nm from spectra of samples containing different
2
+
Zn /ligand molar fractions, and constant C + C concentrations, clearly evidenced maxima at
Zn
L
molar fraction of ligand = 0.5, confirming the formation of a Zn:L 1:1 species (Job’s plots for 1-3,
Figures SI11-SI13). The treatment of the data of the spectroscopic titrations allowed to determine
the conditional log β values for the [ZnL] complexes, which are reported in Table 2.
2
+
We attempted the study of the complex formation equilibria for 1-3 with Cu through direct
2
+
spectrophotometric titrations of the ligands with the metal ion, in analogy with Zn . The spectral
2
+
dataset for the titration of 2 with Cu and the plot of the absorbance values at 418 nm as a function
of the equivalent of the metal are reported in Figure 4; the spectral changes are fully consistent with
those reported for similar systems [18,19].
A sharp break of the formation curve is observed at 1 eq. of metal added, which indicates that the
logβ of the complex species are too high for being determined by these direct titration experiments.
2
+
It is noteworthy, however, that the titration end-point, corresponding to 1 eq. of Cu added, is
consistent with the results of the Job’s plots carried out for all 1-3 ligands (Figures SI14-SI16).
2
+
Overall, these data suggest that the species formed are Cu :L 1:1, that is [CuL].
We decided to determine the log β values of the copper(II) complexes by competition experiments
based on the [CuL] + L’ = [CuL’] + L reaction, where L’ is the competing polydentate ligand
diethylenetriamine (Dien). These experiments were carried out by titration with Dien of solutions
2
+
containing the [CuL] complexes obtained by addition in situ of 1 eq. of Cu to the ligands. A
2
+
representative titration with Dien of a solution containing Cu and 2 in 1:1 ratio is reported in
Figure 5 (the titrations with the other ligands are reported in the Supporting Information paragraph
as Figures SI17 and SI18). Upon addition of Dien the absorbance of the bands at ca. 308 and 343
nm increases, while the two bands associated to the complex at ca. 418 and 490 nm decrease
1
8

ACCEPTED MANUSCRIPT
concomitantly. It is important to note that each spectrum was collected 7 minutes after the addition
of Dien, as we observed that the spectrum changes within the first 3-4 min after titrant addition and
mixing, while there are no spectral changes after 5 min. This behavior is attributed to a relative slow
kinetics for the displacement of the thiosemicarbazone ligands by Dien.
2+
The conditional log β values of the Cu /Dien complexes at pH 7.4 in methanol:water 9:1 (v/v)
were in turn required to calculate the conditional log β value for the [CuL] complex. We have
2
+
therefore carried out a potentiometric study of the Cu /Dien system in methanol:water 9:1 (v/v)
with the aim to determine the speciation of the system and to calculate the conditional log β values
of the complexes from their global formation constants. The protonation and complex formation
2
+
constants of the Cu complexes with Dien in this medium are reported in the Supporting
Information (Figures SI19 and SI20). The complex species found by potentiometry are
2+
3+
2+
[
Cu(Dien)] , [Cu(HDien)(Dien)] and [Cu(Dien) ] . Known the speciation model, it was possible
2
to calculate the concentration of all copper(II) and Dien species in solution at pH 7.4 using the Hyss
2
2
2
009 software. Noteworthy, for total CCu ca. 40 uM at pH 7.4 the Cu:Dien 1:2 species
3
+
2+
(
[Cu(HDien)(Dien)] and [Cu(Dien) ] ) account to ca. 4 % total copper, therefore negligible. The
2
2
+
conditional binding constant of Dien for Cu was then calculated as:
ꢀꢁꢂꢃꢄꢅ[ꢁꢆꢇꢈꢉꢃ]ꢊ[
[
]
ꢁ
ꢆꢇꢈꢉꢃ
ꢁꢆꢅꢃꢂꢋꢅꢌꢂꢆꢃꢄꢅꢋꢂꢅꢇꢈꢉꢃ][ꢇꢈꢉꢃꢅꢃꢂꢋꢅꢌꢂꢆꢃꢄꢅꢋꢂꢅꢁꢆ]
and it resulted to be 13.50 at pH 7.4.
By treating the spectral dataset as reported in Figure 5 using the conditional binding constant of
2
+
2+
Dien for Cu , the calculated log β values for the Cu + L = [CuL] equilibria are 14.56, 14.67 and
2
+
1
5.65 for L = 1, 2 and 3, respectively (Table 2). It is worth noting that the [Cu(Dien)] complex
-1
-1
has a significant absorption in the visible range only at ca. 600 nm (ε ca. 80 cm M ), its
contribution to the absorbance is less than 0.01 at 600 nm, and thus negligible in the entire visible
1
9

ACCEPTED MANUSCRIPT
range [69]. For this reason, we calculated the conditional log β of formation of [CuL] by treating the
absorbances in the 400-540 nm range.
3.5
Interaction of 4 with albumin. Albumin (HSA) is a plasma constituent present in
concentrations close to mM [70]. It is known to bind copper(II) at the N-terminal ATCUN fragment
with affinities of the order of pM (pH 7.4) [71]. Therefore, it is a potential competing ligand for
2
+
Cu in physiological media. In order to gain insight on the stability of the Cu(II) complex 4 in the
presence of albumin, we collected the CD spectra in the visible range of solutions containing HSA
2
+
at 300 µM concentration and in 1.2- to 2.4-fold excess with respect of 4. As a fingerprint of Cu
complexation we used the Cotton effect which appears at ca. 500-600 nm in the visible CD spectra,
2
+
that is known to be diagnostic of the binding of Cu to the ATCUN fragment [71]. The visible CD
2
+
spectra of the solutions of HSA in presence of 4 or Cu as chloride salt are reported in Figure 6.
The solution of free HSA does not present, as expected, any CD signal in the 400-800 nm range.
2
+
Conversely, the solution of HSA and Cu as chloride salt presents two CD bands at 490 nm
(positive) and 570 nm (negative), which indicate the binding of copper(II) at the ATCUN site of the
protein. Interestingly, the spectrum of the HSA/4 solution does not present any Cotton effect
associated to the binding of copper(II) to the ATCUN fragment, therefore indicating that the
complex is stable in the presence of the protein. Also the CD spectra in the 300-500 nm range
appear quite interesting (Figure 7).
In this range, the spectra do not allow to differentiate free HSA to that binding the copper(II) ion.
However, it is clear that the HSA/4 solution provides with spectral feature that are not accounted by
2
+
simple Cu ion-protein interaction. The Cotton effect of the HSA/4 solution corresponds very well
in wavelength and, remarkably, also in intensity, with the UV-visible absorption spectrum of 4
(Figure 7). It seems that not only 4 remains undissociated in the presence of HSA, but also it
interacts with the protein to provide a signal in the CD spectrum. It is known that HSA has multiple
binding sites and pockets into which a variety of drugs can be accommodated [72]. The CD data in
our hands do not allow to put forward the structural nature of the HSA/complex adduct, which
2
0

ACCEPTED MANUSCRIPT
might involve the interaction of 4 into a pocket of HSA and/or the formation of hydrogen bonds on
the protein surface. Nevertheless, 4 becomes CD active likely as the result of second sphere
interactions: therefore, not only 4 is stable in the presence of the strong Cu(II) binding agent HSA,
but the protein may also act as a carrier for the complex in plasma medium.
3
.6
compounds to inhibit cell proliferation, we treated A549 and MCF7 cells with 10 µM of the ligands
-3, the copper(II) complexes 4-6 and the zinc(II) complex 9. These tumor cell lines were chosen as
Effect on cell proliferation of A549 and MCF-7 cells. In order to evaluate the ability of the
1
in vitro model for non small cell lung cancer (A549) and breast cancer (MCF7). NSLC, in
particular, is not very sensitive to chemotherapy and/or radiation, and novel therapeutics are
urgently needed. The zinc complex was chosen because zinc(II) ion is not redox active, on the
contrary of copper(II) ion. As shown in Figure 8 (A and B), 1 and 2 reduced cell proliferation up to
50% in A549 cells, whereas in MCF7 cells they exhibited maximal inhibition of cell growth; also 3
and 9 showed a strong growth-inhibitory effect on MCF7 cells, but they were significantly less
active with A549. Only the copper(II) complexes 4-6 completely inhibited cell proliferation in both
cell lines. Compared to the analogous copper complex 6, the zinc coordination compound 9 shows
reduced activity: especially towards A549 cells, inhibition of cell proliferation is around 20%,
compared to a complete inhibition induced by the analogous copper compounds. This indicates a
crucial role of the metal ion features in determining activity. An investigation on the possible
involvement of ROS production in the apoptosis process is presented in the following paragraph.
Based on these results, 4-6 were chosen for further studies. Firstly, A549 cells were treated with
increasing doses of 4-6 for 3 days. As demonstrated in Figure 9, these compounds showed
comparable dose-response curves, with IC50 of 0.5, 0.3 and 0.7 µM, respectively; comparable
effects were obtained in MSTO mesothelioma cancer cells (data not shown). Note that the IC of
5
0
the potent antineoplastic 2-benzoylpyridine 4-ethyl-3-thiosemicarbazone is practically equal, i.e.
IC = 0.593 ± 0.148 µM [73] (for comparison, IC against A549 cells: cisplatin 11.92 ± 2.19 [74];
5
0
50
doxorubicin 1.8 ± 0.85 [75]). The slight differences in IC values could be related to the different
5
0
2
1

ACCEPTED MANUSCRIPT
substituent at the hydrazonic moiety (R = H, ethyl, phenyl) resulting in different lipophilicity and
thus in different ability to carry the metal ion throughout the cell membrane [77].
Then, we investigated the effects of 5 on A549 cell cycle distribution. As shown in Figure 10, 5 at
1
µM increased the percentage of cells in G /M phase compared to untreated cells (19.5 vs 9.6),
2
suggesting that the observed inhibition of cell proliferation was mediated by a block in this phase of
the cell cycle.
3.7
The copper(II) complex 5 did not induce cell stress. Treatment of A549 cells with 5 at 1 µM
concentration induced an increase of Cyclin B expression, indicative of a block in G2/M phase, but
no accumulation of p53 protein was observed (Figure 11).
In order to highlight if compound 5 could cause DNA damage, we performed a western blot
analysis of poly (ADP-ribose) polymerase (PARP) protein activation in A549 cells treated with
compound 5, at 0.1 and 1 µM for 24 and 48 h and with doxorubicin 1 µM, a well-known DNA-
damage inducing agent. The enzymatic activity of PARP is induced upon DNA damage and the
PARP protein is cleaved during apoptosis, which suggested a role of PARP in DNA repair and
DNA damage-induced cell death. As can be seen in Figure 12, compound 5 did not induce PARP
cleavage, whereas after doxorubicin treatment we detected the cleaved (active) form of PARP
enzyme. Therefore, it can be inferred the lack of activity of compound 5 as DNA damaging agent.
Previous studies reported that the anti-proliferative activity of some copper(II) based-compounds is
7
associated with the production of reactive oxygen species and, in general, with the induction of
intracellular stress. To evaluate whether our compounds produce the same effects, firstly we
quantified the levels of HSP72, a well-known stress-induced protein [76] in A549 cells treated with
5
at 1 µM concentration. As reported in Figure 11, the inducible level of HSP72 did not change. In
addition, A549 cells were treated with 0.5 and 1 µM of 5 and ROS formation was evaluated by flow
cytometry using CM-H2DCFDA and DHE oxidant-sensitive fluorescent detection probes. We do
not detect ROS production after treatment with 5 at concentrations comparable or higher than IC50
(data not shown) for 24 and 48 hours. To further confirm a ROS-independent mechanism of action,
2
2

ACCEPTED MANUSCRIPT
we used A549 cells treated with H O (5 mM for 15 minutes) as positive control for ROS
2
2
production, in comparison with A549 cells treated with compound 5 (1 µM for 72 h). Compound 5
Figure 13, blue line) did not induce ROS accumulation compared to control cells (in grey in
Figure 13), whereas H O treatment caused a marked accumulation of ROS species (red line).
(
2
2
Moreover, to further confirm the lack of apoptosis/necrosis after compound 5 exposure, we
performed a morphological analysis of nuclear chromatin with propidium iodide (PI) and Hoescht
3
3342 staining in A549 cells, after treatment with compound 5 (1 µM), and doxorubicin (1 µM) for
2 h and H O (5 mM) for 1 h. As showed in Figure 14, nuclei from A549 control cells (A) and
7
2
2
after compound 5 treatment (B) did not show any sign of apoptosis/necrosis. In contrast,
doxorubicin (C) induced apoptotic cell death, documented by nuclear fragmentation and increased
number of PI-positive shrunken apoptotic nuclei; H O treatment (D) induced necrosis.
2
2
These observations are in accord with mechanicistic studies on copper complexes of Triapine and
other thiosemicarbazones, underlying that thiol-induced intracellular ROS generation might
contribute to the anticancer activity of copper thiosemicarbazone complexes, but it is not the only
determining factor [8]. In the literature the antiproliferativae activity of the Cu(II) complexes is
correlated to diverse mechanisms of action: for example, Cu(II) thioxotriazole complexes induce
cell death via paraptosis [77]. Other copper(II) compounds activated a mechanism, that is based on
the arrest of the cell in the S phase and induction of apoptosis by intrinsic pathway, with increased
ROS accumulation [78]. An interesting evaluation of the Topo IIα inhibition activity of a series α-
heterocyclic-N4-substitued thiosemicarbazones and their corresponding copper complexes was also
reported [79]. Therefore, further studies will be directed to explore the molecular mechanism of
action of the synthetized copper compounds.
4
. Conclusions
2
3

ACCEPTED MANUSCRIPT
The hydroxyquinoline thiosemicarbazone ligands 1-3 form exclusively metal:ligand 1:1 species in
solution with copper(II) and zinc(II) ions, in agreement with the behavior of previously reported,
similar thiosemicarbazones. The solid state structure of the copper(II) complexes is square planar,
as results from quantum mechanical calculations and by the single crystal X-ray diffraction analysis
2
+
2+
of 5. The Cu complexes are more stable than the Zn ones, as expected by the Irving-Williams
series and defined by solution studies. In this connection, note that the inhibition of cell
proliferation produced by the Zn(II) complex 9 (Figure 8) can be due to partial transmetallation and
formation in the cell of the more stable Cu(II) complex [80].
The Cu(II) complexes 4-6 are more active than the corresponding ligands towards NSCLC and
breast cancer cell proliferation, with IC values ranging from 0.3 to 0.7 µM in A549 cell line. On
5
0
the contrary, the zinc(II) complex 9 showed reduced activity. The most active copper(II) complex 5
reduces cell proliferation with IC lower than 1 µM, with an increase of cells in G2/M phase.
5
0
Peculiarity of 5 is the absence of ROS-accumulation, in contrast with other copper(II) complexes
which were shown to induce ROS-dependent cell stress and death [7]. It is known that p53
accumulation is a sign of genotoxic stress accompanied by DNA damage [81]: the Cu(II) complex 5
did not cause increase of p53 levels in A549 cells wt for p53, indicating lack of DNA damage.
Further studies are ongoing to explore the molecular mechanism of action of these compounds and
to improve their selectivity.
Author Contributions
The manuscript was written through contributions of all authors. All authors have given approval to
the final version of the manuscript.
Notes
The authors declare no competing financial interest.
2
4

ACCEPTED MANUSCRIPT
ACKNOWLEDGMENTS
The authors thank “Centro Interfacoltà Misure Giuseppe Casnati” of the University of Parma for
facilities. Thanks are also due to dr. Andrea Griesi for the collection of the UV data and to Dr.
Monica Maffini for technical assistance. This investigation was supported by the association “Noi
per Loro Onlus”, project “Andrea Spadola”, Associazione Volontaria Promozione Ricerca Tumori
(A.VO.PRO.RI.T.) (Parma), Cooperativa sociale “Di Mano in Mano Onlus” (Parma), and
Associazione Augusto per la Vita (Novellara, Reggio Emilia).
2
5

ACCEPTED MANUSCRIPT
Table 1
N
N
O
N
S
O
N
N
Cu
N
Cu
NHC H5
S
NHC H5
2
2
Param.
Exp E
DFT-BS(I) 5(E) DFT-BS(II) 5(E)
DFT-BS(I) 4(E)
Cu-S
2.200
85.5
78.3
2.348
82.3
78.0
2.304
83.4
78.0
2.514
68.3
88.7
N2-Cu-S1
N2-Cu1-N1
N1-Cu-O1
Cu-N(iminic)
C11-S1
82.3
79.8
81.6
83.5
2.011
1.760
1.320
2.081
1.765
1.339
2.080
1.768
1.336
1.966
1.742
1.343
C11-N3
Relative Energy
0
.0
0.0
(
kcal/mol)
RMSD-D
Å)
RMSD-A (°)
0.149
0.106
2.2
0.315
(
4.1
20.1
2
6

ACCEPTED MANUSCRIPT
2
+
2+
Table 2. Logarithms of the conditional formation constants of the Cu and Zn complexes of the
ligands L=1-3 (methanol:water 9:1 v/v, 25 mM HEPES buffer, pH 7.4). The corresponding
affinities of the ligands for the metal ions (K ) are also reported. Standard deviations are in
a
parentheses.
Ligand (L)
log β
Ka
1
-
3
[
[
CuL]
ZnL]
14.56(1)
6.68(1)
2.75(6) 10 pM
0.21(2) µM
2
-
3
[
[
CuL]
ZnL]
14.67(1)
6.13(4)
2.14(5) 10 pM
0.74(7) µM
3
-
4
[
[
CuL]
ZnL]
15.65(1)
7.30(1)
2.24(5) 10 pM
0.050(1) µM
2
7

ACCEPTED MANUSCRIPT
References
[1] Yu, Y.; Kalinowski, D.S.; Kovacevic, Z.; Siafakas, R.A.; Jansson, P.J.; Stefani, Ch.; Lovejoy,
D.B.; Sharpe, P.S.; Bernhardt, P.V.; Richardson, D.R. Thiosemicarbazones from the Old to New:
Iron Chelators That Are More Than Just Ribonucleotide Reductase Inhibitors. J. Med. Chem. 2009,
52, 5271–5294.
[2] Britta, F.; Kryeziu, K; Kallus, S.; Heffeterb, P.; Bergerb, W.; Kowol, Ch. R.; Keppler, B. K.
Nanoformulations of anticancer thiosemicarbazones to reduce methemoglobin formation and
improve anticancer activity. RSC Adv. 2016, 6, 55848–55859.
[
3 ] Pati, M.L.; Niso, M.; Ferorelli, S.; Abate, C.; Berardi, F. Novel metal chelators
thiosemicarbazones with activity at the σ2 receptors and P-glycoprotein: an innovative strategy for
resistant tumor treatment. RSC Adv. 2015, 5, 103131 –103146.
[
4] Fung, K. L.; Tepede, A. K.; Pluchino, K. M.; Pouliot, L. M.; Pixley, J. N.; Hall, M. D.;
Gottesman, M. M. Uptake of compounds that selectively kill multidrug-resistant cells: the copper
transporter SLC31A1 (CTR1) increases cellular accumulation of the thiosemicarbazone NSC73306.
Mol. Pharmaceutics 2014, 11, 2692−2702.
[
5] Santini, C.; Pellei, M.; Gandin, V.; Porchia, M.; Tisato F.; Marzano, C. Advances in copper
complexes as anticancer agents. Chem.Rev. 2013, 114, 815-862.
6] Denoyer, D.; Masaldan, S.; La Fontaine, S.; Cater, M.A. Targeting copper in cancer therapy:
Copper That Cancer'. Metallomics 2015, 7, 1459-1476.
7] Akladios, F.N.; Andrew, S. D.; Parkinson, Ch. J. Increased generation of intracellular reactive
[
'
[
oxygen species initiates selective cytotoxicity against the MCF-7 cell line resultant from redox
active combination therapy using copper–thiosemicarbazone complexes. J Biol Inorg Chem 2016,
21, 407–419.
2
8