1 3

J Biol Inorg ChemDOI 10.1007/s00775-015-1307-x

ORIGINAL PAPER

SOD mimetic activity and antiproliferative properties of a novel tetra nuclear copper (II) complex

Sagiv Weintraub1 · Yoni Moskovitz1 · Ohad Fleker1 · Ariel R. Levy1 · Aviv Meir1 · Sharon Ruthstein1 · Laurent Benisvy1 · Arie Gruzman1

Received: 24 June 2015 / Accepted: 21 October 2015 © SBIC 2015

Introduction

The DNA crosslinker, cisplatin, is widely used as an anti-cancer drug by more than 50 % of patients suffering from different types of cancer [1, 2]. During the last 30 years, encouraged by cisplatin’s success as a first-choice drug for many human cancers, medicinal and coordination chem-ists have worked in concert to introduce novel metal-based anticancer drugs [1, 3–6]. Such drugs aim to improve tox-icity and the drug-resistance problems associated with cis-platin. For instance, cisplatin is known to have neurotox-icity and nephrotoxicity side effects, especially in children [7–9]. In this context, novel chemotherapeutic metal-based compounds containing metals such as titanium, ruthenium, tin, and rhodium have been recently reported [10, 11]. Additionally, a number of copper complexes have recently been shown to have anticancer properties [12–16].

The metal-based anticancer drugs’ mode of action, for example, that of cisplatin, involves DNA interactions that inhibit cell replication. In addition to this anti-DNA activ-ity, the mechanism underlying the SOD mimetic effect has been considered as a possible means of anticancer activ-ity for metal-containing anticancer drugs. Superoxide dismutases (SODs) are vital metalloenzymes (e.g., cyto-plasmic CuZnSOD-1, mitochondrial manganese SOD-2 (MnSOD), and extracellular CuZnSOD-3) that catalyze the dismutation of superoxide radical anion (O2

·−), thereby protecting the cells against oxidative damage and apopto-sis [17–19]. The majority of cancer cells have increased rates of metabolism as well as an increased production of radical oxidative species (ROS) such as O2

·− [20–22], as compared with normal cells. Moreover, it has been shown that in cancer cells all of the three above-mentioned SOD enzymes are depleted, resulting in an excess of O2

·−, which

Abstract The search for novel anticancer therapeu-tic agents is an urgent and important issue in medicinal chemistry. Here, we report on the biological activity of the copper-based bioinorganic complex Cu4 (2,4-di-tert-butyl-6-(1H-imidazo- [1, 10] phenanthrolin-2-yl)phenol)4]·10 CH3CN (2), which was tested in rat L6 myotubes, mouse NSC-34 motor neurone-like cells, and HepG-2 human liver carcinoma. Upon 96 h incubation, 2 exhibited a significant cytotoxic effect on all three types of cells via activation of two cell death mechanisms (apoptosis and necrosis). Com-plex 2 exhibited better potency and efficacy than the canon-ical cytotoxic drug cisplatin. Moreover, during shorter incubations, complex 2 demonstrated a significant SOD mimetic activity, and it was more effective and more potent than the well-known SOD mimetic TEMPOL. In addition, complex 2 was able to interact with DNA and, cleave DNA in the presence of sodium ascorbate. This study shows the potential of using polynuclear redox active compounds for developing novel anticancer drugs through SOD-mimetic redox pathways.

Keywords Tetra nuclear copper (II) complex · Apoptosis · SOD mimetic · Cytotoxicity · Anticancer drugs

Electronic supplementary material The online version of this article (doi:10.1007/s00775-015-1307-x) contains supplementary material, which is available to authorized users.

* Laurent Benisvy benisvy@gmail.com

* Arie Gruzman gruzmaa@biu.ac.il

1 Department of Chemistry, Bar-Ilan University, 5290002 Ramat Gan, Israel

J Biol Inorg Chem

1 3

after conversion to the perhydroxy radical (HO2·), induces

tumor formation and progression [23]. On the other hand, it has been shown that artificial over-expression of the SOD enzymes in cancer cells leads to a significant anti-tumor effect [24–26]. Though the pharmacological use of SOD enzymes would constitute a therapeutic modality, it is limited by classical protein pharmacokinetic problems such as instability in oral administration, short plasma half-life, insignificant membrane permeability, and immu-nological reactions [27]. Therefore, to overcome these drawbacks, small molecules capable of mimicking SOD activities have been developed [19, 28]. Most of these SOD mimetics possess at least one redox-active metal ion, as in the active site of the natural SODs [e.g., copper (II), ferrous and manganese (II)]. For example, metalloporphy-rin complexes of iron, manganese, and copper have been shown to possess antitumor activity by converting superox-ide to hydrogen peroxide and by a subsequent Fenton-like reaction, consequently, generating a hydroxyl radical that kills cells [3, 29]. Dozens of mono- and dinuclear com-plexes, predominantly with a copper (II) ion, have shown an excellent ability to mimic SOD activity [16, 30, 31]. These include complexes of amino acid residues, peptides, salicylate derivatives, as well as macrocyclic and tetraden-tate Schiff-bases [32–36]. It is important to mention that among all reported SOD mimetics, DNA-targeting copper

(II)-based complexes, in particular, have shown many promising characteristics that made them suitable for anti-cancer drug development. One of the important reasons for that phenomenon is related to the fact that one of the most active anticancer drugs, bleomycin, is a copper-activated, natural chemotherapeutic medicine [37]. For example, based on this molecule, and using Cu(II) and 1,10-phen-anthroline as a ligand, Kellet and co-authors developed a complex (SOD and catalase mimetic) that could initiate free radical DNA scission [38]. Importantly, among many copper (II)-based complexes [38, 39], compound [Cu (3-methoxysalicylic acid) (1,10-phenanthroline)], which was reported by O’Connor et al. [40] (Scheme 1), exhib-ited an antiproliferative effect even in cisplatin-resistant cancer cells [40].

Recently, we have designed a hetero-ditopic N,O-/N,N-pro-ligand (LH2, Scheme 1) that has a redox-active N,O-phenolimidazole chelating unit fused with a N2-phenan-throline chelating unit [41]. The pro-ligand LH2 can bind metal ions, either in their neutral form via the N2-phenan-throline binding mode, as in the octahedral mono-nuclear iron complex [Fe(LH2)3][ClO4]2 (1) (Scheme 1), or in the dianionic N,O-/N,N bridging mode, as in the neutral tetra-nuclear copper(II) complex Cu4L4 (2) [41] (Scheme 1). Herein, we report the first investigation of these compounds regarding their possible anticancer effects.

Scheme 1 Structure of [Cu (3-methoxysalicylic acid)(1,10-phenanthroline)], LH2, Complex 1, and Complex 2

J Biol Inorg Chem

1 3

Our data show that complex 2 is highly cytotoxic towards NSC-34 motor neuron-like cells and HepG-2 car-cinoma cells and that it is more potent than cisplatin. Com-plex 2 possesses no appreciable catalase activity, but it has exhibited a remarkable SOD-like activity, higher than that of the well-known SOD mimetic agent TEMPOL. To the best of our knowledge, this is the first time that a tetra-nuclear complex was reported to have a significant SOD mimetic-dependent cytotoxic effect in vitro, combined with the ability to cleave DNA in a therapeutic concentration range.

Materials and methods

A 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bro-mide, pBR322 DNA, amsacrine, cisplatin, and gel elec-trophoresis reagents were purchased from Sigma Aldrich (Rehovot, Israel). Santa Cruz (Dallas, TX, USA) supplied TEMPOL. All tissue culture reagents, PBS tablets, and medium were purchased from Biological Industries (Beit Haemek, Israel). The SOD and catalase assay kits were supplied by Cayman Chemicals (Ann Arbor, MI, USA). The xanthine oxidase kit was purchased from BioVision (Milpitas, CA, USA). The apoptosis detection kit was sup-plied by Medical and Biological Laboratories (Nagoya, Japan). DNA plasmid and HindIII restriction endonuclease were purchased from New England Biolabs (Ipswich, MT, USA). All materials were analytical grade and were used without purification unless otherwise noted. The ligand LH2 and complex 2 were synthesized as recently reported by Mathias et al. [41].

Synthesis and characterization of 1

A solution of Fe(ClO4)2·6H2O (0.0712 g, 0.1965 mmol, 1 eq.) in DMF (5 mL) was added to a solution of LH2 (0.25 g, 0.589 mmol, 3 eq.) and dissolved in a minimum volume of DMF. The reaction mixture was stirred at room temperature for ca. 1 h. The red solution was then evapo-rated under vacuum to yield a red oil. The latter was dis-solved in a minimum amount of methanol and the solu-tion was left in the open air for 2–3 days, after which time red crystals had formed. These were filtered with a Büchner funnel, washed with Et2O, and dried under vacuum to yield 0.260 g (95 %) of fine powder of [FeL3](ClO4)2 (1). Elemental analysis: Calc. for [C81H84N12O3Fe][ClO4]2·3DMF·6H2O: C 58.25, H 6.35, N 11.32, O 17.24; Found: C 58.33, H 6.76, N 11.70, O 17.52 %. NMR (1H, 300 MHz, DMSO): δ 14.34 (s, 1H, OH/NH), 13.21 (s, 1H, OH/NH), 9.23 (m, 2H, ArH), 8.19 (s, 1H, PhOH), 7.85 (m,

2H, ArH), 7.49 (m, 2H, ArH), 7.46 (s, 1H, PhOH), 1.49 (s, 9H, tBu), 1.42 (s, 9H, tBu). MS ES(+): m/z 664 {M2+}. UV/vis (CH2Cl2): λmax/nm (ε/M-1 cm−1): 524 (1930), 486sh (1370), 411sh (1600), 343 (6050), 300sh (10,700), 285 (15,000). IR (ZnSe): ν (cm−1): 2963, 2923, 2881, 1660 (DMF), 1607 (C=C), 1549, 1508, 1486, 1476, 1466, 1441, 1409, 1389, 1372, 1311, 1286, 1251, 1226, 1203, 1144, 1089, 1028, 994, 936, 888, 862, 818, 815, 778, 729(CH), 666, 639, 620, 580.

Cell culture

NSC-34 cells were cultured in Dulbecco’s modified eagle medium (DMEM) supplemented with 20 % heat inacti-vated foetal bovine serum (FBS), 1 % l-glutamine, and 1 % penicillin–streptomycin solution. HepG-2 was also cultured in DMEM supplemented with 10 % FBS, 1 % l-glutamine, and 1 % penicillin–streptomycin. In addition, a solution of 0.2 % amphotericin B was added [42]. The skeletal muscle cell line (L6) was cultured in minimum essential medium (Alpha-MEM) containing 2.2 gm/L sodium bicarbonate, supplemented with 10 % FBS, 1 % l-glutamine, and 1 % penicillin–streptomycin solution [43]. All cells were grown at 37 °C in a humidified atmosphere, in the presence of 5 % CO2.

Cytotoxicity assay

All test compounds were dissolved in DMSO and the maximum percentage of DMSO present in all wells was 0.1 % (v/v). Each compound solution was added to the wells in triplicate in the concentration range indicated in the legends. Following a 96-h cell exposure to each com-pound, the cells underwent an MTT assay [44]. Briefly, cells were incubated with 200 μL of MT (5 mg/mL) in 0.1 M PBS, at 37 °C in a humid atmosphere with 5 % CO2 for 2 h. The media was then gently aspirated from test cul-tures, and 200 μL of DMSO was added. The plates were then shaken for 2 min, and the absorbance was detected at 570 nm using a BioTek plate reader (Biotek, Winooski, VT, USA).

UV/Vis spectrophotometry

The stability of complex 2 and the free ligand were inves-tigated under biologically relevant conditions (37 °C, pH 7.4, PBS). Briefly, 2.5 µL of complex 2 (0.1 mM) solution in DMSO and 10 µL of free ligand (0.2 mM) solution in DMSO were added to 1 mL of PBS, and after the indicated incubation times, each sample was transferred to a 2 mm quartz cuvette and the spectra were recorded between

J Biol Inorg Chem

1 3

wavelengths 200 and 2500 nm using a Varian Cary 5000 UV–Vis-NIR Spectrophotometer (Santa Clara, CA, USA).

EPR measurement

X-band CW-EPR (continuous wave EPR) spectra of com-plex 2 were recorded using an E500 Elexsys Bruker spec-trometer (Billerica, MA, USA) operating at 9.0–9.5 GHz. The spectra were recorded at room temperature using a microwave power of 20.0 mW, a modulation amplitude of 3.0 G, a time constant of 120 ms, and a receiver gain of 60.0 dB. The samples were measured in 0.8 mm capillary quartz tubes (VitroCom, Mountain Lakes, NJ, USA).

DNA binding study (CD measurement)

The DNA binding experiments were performed in distilled water (pH 7.0) using a pBER322 DNA plasmid (20 μg/mL). DNA linearization was conducted as previously described [45]. Solutions of complex 2 (final concentration, 15 μM), LH2 (final concentration, 60 μM), and amsacrine (25 μM) were added separately to the DNA sample and incubated for 30 min at room temperature. Each spectrum was averaged from two successive accumulations at a scan rate of 50 nm/min, keeping a bandwidth of 1.0 nm. Spectra were recorded between wavelengths 180 and 400 nm in a 1 mm optical path length, using a Chirascan spectrometer (Applied Photophysics, Surrey, UK).

Agarose gel electrophoresis

A DMSO solution containing the different compounds in an Eppendorf tube was treated with the pBR322 plasmid DNA (5 µL of 20 µg/mL) in Tris–HCl buffer (10 mM, pH 8.0). The content was incubated for 5 h at 37 °C and then loaded after being mixed with 1 µL loading buffer onto a 1 % agarose gel. The electrophoresis was performed at a constant voltage (80 V) until the bromophenol blue trav-elled through 75 % of the gel. The plasmid bands were vis-ualized by viewing the gel under a UV transilluminator and then photographed.

DNA cleavage

Plasmid pBER322 DNA was incubated with sodium-l ascorbate (at twice the complex concentration) and with DMSO, complex 2, or LH2 for 5 h in Tris–HCl buffer at 37 °C. The electrophoresis was performed at a constant voltage (80 V) until the bromophenol blue travelled through 75 % of the gel (1 % agarose). The plasmid bands were vis-ualized by viewing the gel under a UV transilluminator and then were photographed.

SOD activity assay

The O2·− dismutase activities of the metal complexes were

assessed using a modified nitro-blue-tetrazolium (NBT) assay with a xanthine–xanthine oxidase system as the source of O2

·−. The quantitative reduction of NBT to blue formazan by O2

·− was followed spectrophotometrically at 450 nm using a BioTek plate reader (Biotek, Winooski, VT, USA). Assays were run in 230 µL of buffer. Results are graphed as the percentage inhibition of NBT reduction for three concentrations. Tabulated results were derived from linear regression analyses and are given as the concentra-tion (μM) equivalent of 1 unit of bovine erythrocyte SOD activity (supplied with the kit). A unit of SOD activity is the concentration of the complex or enzyme that induces 50 % inhibition in the reduction of NBT (referred to as the IC50 value).

Catalase activity assay

The ability of the complexes to disproportionate hydrogen peroxide was assayed using a Cayman Catalase Assay Kit. Measurements were taken in 240 µL of buffer. The activ-ity of the samples was calculated from the average absorb-ance of each sample at 540 nm using a BioTek plate reader (Biotek, Winooski, VT, USA).

Xanthine oxidase activity assay

The possible inhibition of xanthine oxidase by complex 2 was investigated using a xanthine oxidase activity assay kit that was purchased from BioVision (Milpitas, CA, USA). Measurements were taken according to the manufacturer’s instructions.

Apoptosis/necrosis assay

Apoptosis/necrosis analysis was determined using FACS analysis (Becton–Dickinson, Franklin Lakes, USA). Hep-G2 cells were treated with complex 2 (5 µM) or the free ligand (20 µM) for 48 h. Results of the cell sorting were analysed using Cell Quest software (Becton–Dickinson, Franklin Lakes, NJ, USA). A total of 10,000 events were acquired. The cell populations were displayed as a dot plot divided into four quadrants with Annexin V-FITC fluorescence (x-axis) versus propidium iodide fluorescence (y-axis). Both reagents were supplied with the kit.

Statistical analysis

Results are given as mean ± SEM, n = 3. Statistical sig-nificance (p < 0.05) was calculated among experimental

J Biol Inorg Chem

1 3

groups using the two-tailed Student’s t test using the web resource: http://www.graphpad.com/quickcalcs/ttest1.cfm.

Results

In vitro cytotoxic activity

The cytotoxic effect of LH2, as “free” or coordinated in neutral and dianionic forms in complexes 1 and 2, respec-tively, has been investigated in the following cancer cells: NSC-34 (mouse neuroblastoma/mouse motor neurons) that represents a central neural system tumor [46], HepG-2 (human liver hepatocellular carcinoma), which represents a somatic tumor, and rat L6 skeletal muscle myotubes as non-cancerous control cells. These cells spontaneously dif-ferentiate at near confluency in low serum concentrations from myoblasts to myotubes, which are multinuclear cellu-lar clusters with many morphological similarities to native

muscle myofibrils. Such a transformation interrupts the cellular proliferation and therefore L6 myotubes are in use as a control for highly differentiated non-dividing cells in in vitro experiments.

Figure 1a, b shows that both LH2, and complex 2 have significant cytotoxic effects in both types of cancer cells. Within the determined optimal conditions (an incubation time of 96 h and a compound concentration of 5 μM) about 80 % of cells were killed by the cytotoxic effect that was induced in an almost identical manner by 2 and LH2. In contrast, when the ligand was bound to a Fe(II) center, as in complex 1, no significant cytotoxic effect was observed (Fig. 1a, b).

In addition, as shown in Fig. 1c, the elevated cytotoxic effects of LH2 and 2 appear to be non-discriminatory, which was also observed under similar conditions in L6 myotubes. For example, L6 cell viability was dramatically reduced to 8.14 ± 0.15 % for 2 (Fig. 1c). It is important to mention that under these experimental conditions free

Fig. 1 Cytotoxicity activity (MTT) of test compounds after 96 h incubation in: a NSC-34 cells. b Hep-G2 cells. c L6 myotubes. (Pt cisplatin; CN acetonitrile); n = 3

J Biol Inorg Chem

1 3

copper ions (e.g., CuCl2), acetonitrile, and cisplatin did not exhibit a cytotoxic effect towards cancer cells (see ESM, Figure 1). Thus, both the free ligand and the neutral tetra-copper (II) complex 2 appear to be highly cytotoxic to both cancerous as well as non-cancerous cells, showing no selectivity. In contrast, the strong N2-binding mode of LH2 to the Fe(II) ion in complex 1 appears to prevent the cyto-toxicity of the ligand.

To better understand the cytotoxic activity of LH2 and 2, additional in vitro assays were conducted. The inactive complex 1 was not tested in additional experiments.

Stability of complex 2 in aqueous medium

To rule out the possibility that the cytotoxic effect of com-plex 2 is attributed to the free ligand activity, resulting from the dissociation of the complex under physiological conditions, the complex stability was monitored by UV/Vis and EPR spectroscopies. As shown in Fig. 2, the UV/vis spectra of complex 2 (0.1 mM, in PBS buffer at pH 7.4, 37 °C) remains identical for at least 96 h, and no decrease in the intensity of the complex characteristic absorption bands (at 286 and 345 nm [41]) was observed. Further-more, X-band CW-EPR measurements were carried out on aqueous solutions of complex 2 at room temperature at 0, 24, and 96 h incubation times. As shown in Fig. 3, the characteristic isotropic single signal of complex 2 [41] at ca. g = 2.1 remained identical during the entire incubation time, up to 96 h. Moreover, no additional signals of free Cu(II) ions were observed, indicating that the structure of the tetracopper(II) complex 2 remained unchanged (Fig. 3). These findings support the hypothesis that the cytotoxic effect of complex 2 is not related to the cytotoxic activity of the free ligand and that the cytotoxicity pathways of both compounds have distinct mechanisms.

DNA binding and cleavage studies

The potential ability of LH2 (60 μM) and complex 2 (15 μM) to interact with DNA was investigated with super-coiled pBR322 plasmid DNA and was monitored by circu-lar dichroism (CD). The significant differences in the bio-logical activity between complex 2 and cisplatin indicate that both compounds have different cytotoxic mechanisms. Therefore, instead of cisplatin, the well-known intercala-tion agent amsacrine was used as a positive control for the experiment. The CD DNA absorption rate of amsacrine has been well studied in the 180–400 nm region [47–50]. Thus, the DNA and test compounds were run at this spe-cific wavelength range, as shown in Fig. 4. The CD spec-tra of pure DNA (20 μg/mL) were recorded as a standard value (Fig. 4b). All bands obtained by this measurement corresponded to DNA bands that were reported previously

in the literature. As shown in Fig. 4a, all three compounds interact with DNA. Amsacrine, as predicted, binds to DNA and markedly changes the native DNA structure; as indicated by a blue shift in the DNA peak from 225 nm (Fig. 4b) to 209 nm (Fig. 4a). This specific band repre-sents the helicity of DNA under native conditions [50]. In addition, the intensity of the amsacrine signal in the range of 200–220 nm increased dramatically compared with pure DNA. Moreover, both compounds (complex 2 and LH2) also interacted with DNA. However, the mode of interaction was slightly different. Complex 2 led to a

Fig. 2 UV/Vis spectra of complex 2 (10 µM) in phosphate-buffered saline (pH 7.4) at 37 °C recorded at 24 and 96 h incubation time

Fig. 3 X-band CW-EPR spectra of complex 2 in fluid solution [5 µM in 80 % PBS, pH 7.4/20 % DMSO] recorded at RT with incubation times of 0 h (A), 24 h (B) and 96 h (C). The measurement was con-ducted as described in “Materials and methods”

J Biol Inorg Chem

1 3

stronger blue shift than did amsacrine (202 nm compared with 209 nm) and to a lower signal intensity (Fig. 4a). LH2 induced the most significant blue shift in the corre-sponding band (195 nm), but the signal intensity values were much lower than the corresponding amsacrine and complex 2 values (Fig. 4a). Both molecules changed the intensity of the CD signal. An additional band (275 nm) in the DNA CD measurement occurred because of the stack-ing interactions of the DNA bases [50]. This band is very sensitive to any interactions between DNA and the differ-ent ligands. For example, the intensity and wavelength of the absorption of this band correlate with the aggregation effect of amsacrine on the DNA duplex [50]. As shown in Fig. 4c, amsacrine indeed increased the intensity and red-shifted this band as compared with native DNA. Interest-ingly, complex 2 did not affect the position of the band but

decreased the intensity of the signal as compared to DNA. Conversely, LH2 did not affect the intensity of the band but similarly to amsacrine, it induced a red shift of the band (to 283 nm).

In addition to the binding activity, the possible effect of compound 2 on DNA relaxation was tested by aga-rose gel electrophoresis. As shown in Fig. 4d, complex 2 (15 µM) and LH2 (60 µM) did not appreciably affect the DNA structure and more than 70 % of the DNA remained supercoiled. However, in the presence of sodium ascorbate reducing agent, the supercoiled DNA was totally cleaved by complex 2, as shown in Fig. 4e. This observation fur-ther suggests that complex 2 can interact with supercoiled DNA and that the presence of redox-active Cu(II) ions [that can be reduced to Cu(I)] is essential for supercoiled DNA cleavage.

A B

C

wavelength

CD(m

deg)

CD(m

deg)

wavelength

CD(m

deg)

wavelength

E

D

NC

LCSC

NC

LCSC

1 2 3 4 5 6

1 2 3 4 5 6

Sodium Ascorbate [120 µM]

180 200 220 240 260 280 300 320 340 360

230 240 250 260 270 280 290 300 310 320 350340330

200 220 240 260 280 300 320 340 360 380

Fig. 4 DNA binding and cleavage analysis determined by CD and gel electrophoresis. a CD signal spectra of the pBR322 plasmid DNA (20 μg/mL) in the presence of complex 2 (15 μM), LH2 (60 μM), and amsacrine (25 μM). b CD signal spectra zoom in of the pBR322 plasmid DNA (20 μg/mL). c Zoom in on the CD signal spectra of pBR322 plasmid DNA (20 μg/mL) in the presence of complex 2 (15 μM), LH2 (60 μM), and amsacrine (25 μM) at the 230–350 nM range. d DNA binding determined by gel electrophoresis. DNA bands designation: supercoiled DNA (SC), linear circular (LC) nicked cir-cular (NC). Plasmid DNA was linearized with Hind III restriction

endonuclease and recovered by phenol extraction and ethanol precipi-tation. Lane 1 marker; lane 2 DNA; lane 3 treated by Hind III DNA; lane 4 DNA + 1 % DMSO; lane 5, DNA + complex 2 (25 µM); lane 6 DNA + LH2 (60 µM). e Plasmid pBER322 DNA (20 μg/mL) was incubated without and with complex 2, DMSO, LH2 and sodium ascorbate for 1 h in Tris–HCl buffer (pH = 7.4) at 37 °C. The plas-mid bands were visualized by viewing the gel under a UV transillu-minator. Lane 1 marker; lane 2 DNA; lane 3 DNA + 1 % DMSO; lane 4 DNA; lane 5 DNA + complex 2 (25 µM); lane 6 DNA + LH2 (60 µM)

J Biol Inorg Chem

1 3

Cell death pathways

The classical cell death pathways of apoptosis and necro-sis [51] were investigated for both the ligand and com-plex 2 using fluorescence-activated cell sorting (FACS) analysis. To discriminate between apoptotic and necrotic cells, dual cell staining was performed with fluorescent Annexin V and propidium iodide (PI), respectively. Such a method can show the levels of apoptosis and necrosis detected by flow cytometry [52, 53]. Experiments were conducted using HepG-2 cells after 48 h of incubation because after incubating for 72 h (and obviously similarly for 96 h, the time we used for MTT) the amount of live cells would be insufficient to conduct the FACS analysis. For complex 2 (Fig. 5a), out of the total amount of dead cells, 38.5 % were detected as late apoptosis, 19.5 % were detected as necrotic, and 42.0 % exhibited both late apop-totic and necrotic signs. Regarding the ligand (in a concen-tration chosen to be four times higher than 2, i.e. 20 µM, as described for MTT viability assays), the apoptotic cell death pathway is clearly dominant, with 30.67 % for early and 59.6 % for late apoptosis, respectively. Only 9.72 % of the dead cells were necrotic (Fig. 5b).

It is important to mention that the free ligand alone, LH2, induced apoptosis more effectively than did com-plex 2. The percentage of only apoptotic cells that was detected after the cells were exposed to the free ligand was around 79.0 %, compared with ~59 % for complex 2. Such a difference in the amount of dead cells may indi-cate that the two compounds have different mechanisms of action.

SOD activity

The potential SOD mimetic activities of both compound 2 and LH2 were examined using the xanthine/xanthine-oxidase-based method, which is used for generating the superoxide radical [54]. The cell membrane-permeable SOD mimetic TEMPOL was used as a positive control [55, 56]. The obtained results are reported in SOD equiva-lent activity units and are shown in Fig. 6a. The calculated EC50 values are presented in Fig. 6b. The experimental data indicated that ligand LH2 displayed no apparent SOD-like activity, whereas compound 2 indeed exhibited SOD-like activity with an EC50 of 1.9 µM. This value is almost iden-tical to the value reported by O’Connor et al. [40] where EC50 = 1.72 µM for the [Cu (3-methoxysalicylic acid) (1,10-phenanthroline)]2+ complex. This evidence also revealed that complex 2 is approximately 2.5-fold more potent than TEMPOL (EC50 = 5.48 µM).

In order to rule out the possibility that the SOD-like activity of complex 2 was not related to the direct inhibi-tion of xanthine oxidase by the complex, the xanthine oxi-dase activity of complex 2 was investigated using a com-mercially available xanthine activity kit, which measures the generation of hydrogen peroxide as a result of the conversion of xanthine to uric acid by the enzyme. Such a test was a crucial since Cu (II)-containing complexes such as [Cu(II)(β-citryl-l-glutamate)] and even Cu(II) by itself have shown significant xanthine oxidase inhibition [57, 58]. The ESM Figure 2 shows that complex 2 did not have any inhibitory effect on xanthine oxidase. Moreover, as expected, the amount of hydrogen peroxide that was

Fig. 5 Determination of apoptosis/necrosis using flow cytometry analysis. The apoptotic effect of complex 2 (a 5 µM) and LH2 (b 20 µM) after a 48 h incubation in Hep-G2 cells was determined by the Annexin V/PI staining method. Each panel shows negative (via-

ble) cells (lower left quadrant), annexin V-positive (early apoptotic) cells (lower right quadrant), PI-positive (necrotic) cells (upper left quadrant), or annexin V and PI double-positive (late apoptotic) cells (upper right quadrant)

J Biol Inorg Chem

1 3

detected in treatment using the complex 2 samples was dose-dependently higher compared with the control meas-urements, suggesting that hydrogen peroxide was produced both by the xanthine oxidase reaction and by dismutation of the superoxide radical (which was synthesized in the process of oxidizing xanthine) by complex 2. Hydrogen peroxide, which came from both sources, was determined by the kit.

Catalase activity

Although the catalytic site of the catalase enzyme is occu-pied by an iron metal ion, several copper complexes have exhibited catalase activity [59, 60]. The ability of complex 2 to disproportionate hydrogen peroxide was investigated using an H2O2 colorimetric detection method (Cayman Catalase Assay Kit). This method is based on the fact that catalase exhibits peroxidase activity in which low molec-ular weight alcohols (in our case, methanol) can serve as electron donors for its own oxidation towards formalde-hyde. The obtained formaldehyde reacts with chromogen

(4-amino-3-hydrazino-5-mercapto-1,2,4-trizole) to pro-duce a purple bicyclic heterocycle(5,6,7,8-tetrahydro- [1, 2, 4] triazolo[4,3b] [1, 2, 4, 5] tetrazine-3-thiol), which is detected by a colorimeter [61, 62]. As shown in ESM Figure 3, complex 2 displays significant catalase-like activ-ity (5 µM of compound 2 converted 0.2 ± 0.02 µM of hydrogen peroxide in 1 min), as compared with the ligand, which is negligible. However, the catalase-like activity of complex 2, calculated to be 0.0244 U/mg (by extrapolat-ing a standard curve of bovine liver catalase), is extremely low, compared with a well-known SOD/catalase mimetic manganese-based complex (EUK-134) [56], with an activ-ity equal to 26 U/mg.

Discussion

The potential antiproliferative effect of two (Fe- and Cu-based) complexes bearing the phenol-phenanthroimidazole ligand LH2 was tested in two cancer cell lines: NSC-34 and in HepG-2. As a control for normal cells, non-cancer rat L6 skeletal muscle cells were used.

Cell viability assays showed that the octahedral mono-nuclear iron complex 1 was inactive. In contrast, both the neutral tetra-nuclear copper(II) complex Cu4L4 (2) and the free ligand LH2 induce cytotoxic effect. In addition, cell death pathway, DNA binding, and SOD activity stud-ies indicate that the cytotoxic activity of the free ligand and complex 2 operate through different mechanisms. In particular, the complex 2 exhibits a remarkable SOD-like activity, in combination with DNA cleavage ability; whereas the ligand LH2 lacks both of these properties. Both SOD-like activity and DNA cleavage ability of complex 2 may therefore play a key role in explaining the high cyto-toxicity of the latter towards cancer cells.

It is known that SOD enzymes play a critical role in cancer pathogenesis. The first reported work by L. Ober-ley et al. in 1978 showed that H6 hepatoma cells do not express Mn SOD [63] and that the loss of Mn SOD activ-ity seems to be an important characteristic of tumor cells in general [64]. Additionally, the transfection of MnSOD gene and overexpression of the protein, in hamster cheek pouch carcinoma cells, has shown to lead to a significant cell growth rate decrease of 50 %. [25]. Such antiprolifera-tive and anticancer effects of SOD overexpression in vitro and in vivo were also obtained in many different types of cancer cells and tumors [65]. For example, the volume of human oral squamous carcinoma cell xenografts in mice was reduced by 70 % after MnSOD overexpression [24]. Similar positive results were obtained with human breast cancer cells in vitro [24]. In addition, the overexpression of the CuZnSOD in cancer cell has been shown to induce a similar antiproliferative effect than that of MnSOD [66].

Fig. 6 Measurement of SOD activity. a The SOD mimetic dose response effect of test compounds. b EC50 values. The concentra-tions were calculated as an equivalent for the effect of 1unit of SOD supplied by the kit. The assay was conducted using a commercially available SOD activity kit according to the manufacturer’s provided protocol

J Biol Inorg Chem

1 3

Interestingly, in contrast to cancer cells, in normal cells, the overexpression of SOD is believed to have a cytoprotective effect owing to normal endogenous antioxidant protein lev-els [66].

The molecular basis for the cytotoxic effect of overex-pressed SODs in cancer cells has been explained by the increased dismutation rate of superoxide anion to hydro-gen peroxide (approximately by four orders of magnitude over spontaneous dismutation) [3]. By increasing the rate of H2O2 production over that of spontaneous superoxide dismutation, the effect of SOD on cell growth could be due perturbation of normal signaling pathways; or cell dam-age caused through hydroxyl radical production formed by Fenton-like reaction with metal cations such as Fe2+ or Cu1+ [66–68].

Based on the knowledge obtained from SOD overex-pression studies in cancer cells, it is believed that molecular systems that have SOD-like activity may have potential use as anticancer agents [24]. In that respect the herein reported highly cytotoxic and SOD-mimetic complex 2 may have a similar effect to that of overexpressed SOD in cancer cells. Considering the tetranuclear nature of the complex 2, and its ability to bind exogeneous ligand at labile positions of each copper(II) ions [41], it is most likely that compound 2 acts according to the known SOD enzymatic mechanism (Scheme 2) [38, 69]. Thus, presumably, the Cu(II) centers in complex 2 bind and oxidises superoxide anions to oxy-gen. The resulting reduced Cu(I) form of the complex, in turn, may reduce superoxide anions, which, in the presence

of protons, produces hydrogen peroxide, although this remains to be determined (Scheme 2).

Recently, the hydroxyl radical, which is an extremely toxic substance for cells (inducing DNA cleavage or non-specific reactions with other macromolecules), has been generated from hydrogen peroxide in a Cu(I)/Cu(II) cata-lytic system, as was shown by several researchers [38, 69].

According to the DNA binding assay results, we also hypothesized that the cytotoxic effect of complex 2 can be related to interruptions in the cell division cycle (DNA-related cytotoxicity). Complex 2 binds to DNA as moni-tored by the drastic changes induced in the CD spectrum of DNA. The oxidative DNA cleavage ability of 1,10-phen-anthroline (one of the structural domains in the ligand) in the presence of Cu(II), was demonstrated in the 1970s by Sigman et al. [70]; and contributed to the cytotoxic effect of the copper complexes [71]. Recently, diverse types of different copper-containing complexes with oxidative DNA cleavage ability have been synthesized [72–77]. Most of them exhibited a biological effect only in the presence of some reductant. However, several complexes among them were also able to cleave DNA without the presence of reductant agents [39, 70]. As presented in this work, com-plex 2 exhibited significant DNA cleavage in the presence of sodium ascorbate as a reducing agent. Combining these data with the fact that complex 2 has very efficient super-oxide dismutase mimetic activity, we can conclude that complex 2 can cleave DNA through free radical generation and by these two effects it can induce the observed cyto-toxic results.

Regarding catalase activity, it is commonly believed that complexes exhibiting such activity may have a cell protective effect (with cancer cells such an effect is not wanted) due to their ability to catalytically disproportion-ate H2O2 (generated by the SOD activity) to harmless water and molecular oxygen; preventing thus the formation of cytotoxic OH• by Fenton chemistry [40]. The complex 2 has a low catalase activity compared with the potent SOD mimetic [Cu(ph)(2,2′-bipy)]2

.2H2O, reported by Kel-let et al. [38]. This complex, which also has strong cata-lase activity, was still extremely toxic for cancer cells [38]. Such results have validated the concept that even if catalase activity is exhibited by a complex, the ability of hydrogen peroxide directly or via the formation of the hydroxyl radi-cal to consequently lead to cell damage is high. Moreover, the generation of hydrogen peroxide is sufficient for this effect, as was shown by Kasugai et al. for water-soluble Fe-porphyrin-based SOD mimetics [29].

Based on the data presented here, we can conclude that complex 2 is a potent SOD mimetic, which may explain its cytotoxic activity in a similar manner to the antiprolifera-tive and anticancer effects observed by SOD-overexpressed

SOD

Cu(II)=

Cu(II) + O2- Cu(I) + O2

O2-2H+

H2O2

Cell prolifera�on

X

Scheme 2 The proposed mechanism of Complex 2 action

J Biol Inorg Chem

1 3

cancer cells. However, we cannot rule out the possibility that complex 2 cytotoxic activity is also mediated by DNA binding and oxidative cleavage. Interestingly, our experi-ments have indicated that for the specific cell lines used, cisplatin displayed only a minor cytotoxic effect compared with the significant activity of complex 2. Alternatively, the free ligand (LH2), which was ineffective as a SOD mimetic but binds DNA, induces its biological activity mainly by a DNA-related mechanism of action, which is now under investigation. In addition, the distribution of necrotic/apop-totic cells, which was observed in the free ligand and in the complex 2 treatments, was clearly different. This fact also points out that both compounds have different mechanisms of action.

In summary, we described here a potent and effective cytotoxic copper-based complex 2 with SOD mimetic properties. Complex 2 exhibited a highly potent cytotoxic effect in two types of cancer cells (NSC-34 and HepG-2). Interestingly, it was more potent than the well-known chemotherapeutic drug cisplatin. Moreover, complex 2 was also more potent in the SOD activity assay than was the well-known SOD mimetic molecule TEMPOL. Addi-tionally, complex 2 exhibited very low catalase activity. Taken together, such a combination of biological activi-ties could be used to develop novel metal-based anti-cancer drugs, which might provide an additional option for inorganic chemistry-related therapeutics besides cisplatin.

Acknowledgments This study was partly supported by Bar-Ilan University‘s new faculty Grants for AG and LB. We thank Dr. M. Kanovsky and S. Manch for editing the manuscript.

References

1. Kelland LR (1993) Crit Rev Oncol Hemat 15:191–219 2. Ho JW (2006) Recent Pat Anticancer Drug Discov 1:129–134 3. Huang R, Wallqvist A, Covell DG (2005) Biochem Pharmacol

69:1009–1039 4. Mascini M, Bagni G, Di Pietro ML, Ravera M, Baracco S, Osella

D (2006) Biometals 19:409–418 5. Jin VX, Ranford JD (2000) Inorg Chim Acta 304:38–44 6. Margiotta N, Bergamo A, Sava G, Padovano G, de Clercq E,

Natile G (2004) J Inorg Biochem 98:1385–1390 7. Hanigan MH, Devarajan P (2003) Cancer Therapy 1:47–61 8. Sadler PJ, Guo ZJ (1998) Pure Appl Chem 70:863–871 9. Sastry J, Kellie SJ (2005) Pediatr Hemat Oncol 22:441–445 10. Clarke MJ, Zhu FC, Frasca DR (1999) Chem Rev 99:2511–2533 11. Dyson PJ, Sava G (2006) Dalton Trans 16:1929–1933 12. Zhang SC, Zhu YG, Tu C, Wei HY, Yang Z, Lin LP, Ding J,

Zhang JF, Guo ZJ (2004) J Inorg Biochem 98:2099–2106 13. Balakrishna MS, Suresh D, Rai A, Mague JT, Panda D (2010)

Inorg Chem 49:8790–8801 14. Sanghamitra NJ, Phatak P, Das S, Samuelson AG, Somasunda-

ram K (2005) J Med Chem 48:977–985 15. Zhou H, Zheng CY, Zou GL, Tao DD, Gong JP (2002) Int J Bio-

chem Cell B 34:678–684

16. Devereux M, Shea DO, Kellett A, McCann M, Walsh M, Egan D, Deegan C, Kgdziora K, Rosair G, Mulller-Bunz H (2007) J Inorg Biochem 101:881–892

17. Miller AF (2004) Curr Opin Chem Biol 8:162–168 18. Afonso V, Champy R, Mitrovic D, Collin P, Lomri A (2007)

Joint Bone Spine 74:324–329 19. Salvemini D, Riley DP, Cuzzocrea S (2002) Nat Rev Drug Dis-

cov 1:367–374 20. Saczewski F, Dziemidowicz-Borys E, Bednarski PJ, Grunert R,

Gdaniec M, Tabin P (2006) J Inorg Biochem 100:1389–1398 21. Huang P, Feng L, Oldham EA, Keating MJ, Plunkett W (2000)

Nature 407:390–395 22. Muscoli C, Cuzzocrea S, Riley DP, Zweier JL, Thiemermann C,

Wang ZQ, Salvemini D (2003) Br J Pharmacol 140:445–460 23. Fridovich I (1999) Ann NY Acad Sci 893:13–18 24. Weydert CJD, Smith BB, Xu LJ, Kregel KC, Ritchie JM, Davis

CS, Oberley LW (2003) Free Radical Bio Med 34:316–329 25. Lam EWN, Zwacka R, Engelhardt JF, Davidson BL, Domann

FE, Yan T, Oberley LW (1997) Cancer Res 57:5550–5556 26. Weydert CJ, Waugh TA, Ritchie JM, Iyer KS, Smith JL, Li L,

Spitz DR, Oberley LW (2006) Free Radical Bio Med 41:226–237 27. Riley DP (1999) Chem Rev 99:2573–2587 28. Safavi M, Foroumadi A, Nakhjiri M, Abdollahi M, Shafiee

A, Ilkhani H, Ganjali MR, Hosseinimehr SJ, Emami S (2010) Bioorg Med Chem Lett 20:3070–3073

29. Asayama S, Kasugai N, Kubota S, Nagaoka S, Kawakami H (2007) J Inorg Biochem 101:261–266

30. Devereux M, O’Shea D, O’Connor M, Grehan H, Connor G, McCann M, Rosair G, Lyng F, Kellett A, Walsh M, Egan D, Thati B (2007) Polyhedron 26:4073–4084

31. Devereux M, McCann M, O’Shea D, O’Connor M, Kiely E, McKee V, Naughton D, Fisher A, Kellett A, Walsh M, Egan D, Deegan C (2006) Bioinorg Chem Appl 2006:1–11. doi:10.1155/BCA/2006/80283

32. Brigeliu R, Spottl R, Bors W, Lengfeld E, Saran M, Weser U (1974) Febs Lett 47:72–75

33. Younes M, Weser U (1976) FEBS Lett 61:209–212 34. Younes M, Lengfelder E, Zienau S, Weser U (1978) Biochem

Bioph Res Co 81:576–580 35. Durackova Z, Felix K, Fenikova L, Kepstova I, Labuda J, Weser

U (1995) Biometals 8:183–187 36. Arslantas A (2002) Met Based Drugs 9:9–18 37. Dorr RT (1992) Semin Oncol 19:3–8 38. Kellett A, Howe O, O’Connor M, McCann M, Creaven BS,

McClean S, Foltyn-Arfa Kia A, Casey A, Devereux M (2012) Free Radical Bio Med 53:564–576

39. Loganathan R, Ramakrishnan S, Ganeshpandian M, Bhuvanesh NSP, Palaniandavar M, Riyasdeen A, Akbarsha MA (2015) Dal-ton Trans 44:10210–10227

40. O’Connor M, Kellett A, McCann M, Rosair G, McNamara M, Howe O, Creaven BS, McClean S, Kia AFA, O’Shea D, Devereux M (2012) J Med Chem 55:1957–1968

41. Mathias J-L, Arora H, Lavi R, Vezin H, Yufit D, Orio M, Aliaga-Alcade N, Benisvy L (2013) Dalton Trans 42:2358–2361

42. Getter T, Zaks I, Barhum Y, Ben-Zur T, Böselt S, Gregoire S, Viskind O, Shani T, Gottlieb H, Green O, Shubely M, Senderow-itz H, Israelson A, Kwon I, Petri S, Offen D, Gruzman A (2015) ChemMedChem 10:850–861

43. Pasternak L, Meltzer-Mats E, Babai-Shani G, Cohen G, Viskind O, Eckel J, Cerasi E, Sasson S, Gruzman A (2014) Chem Com-mun 50:11222–11225

44. Meltzer-Mats E, Babai-Shani G, Pasternak L, Uritsky N, Getter T, Viskind O, Eckel J, Cerasi E, Senderowitz H, Sasson S, Gruz-man A (2013) J Med Chem 56:5335–5350

45. Munder A, Moskovitz Y, Redko B, Levy A, Ruthstein S, Geller-man G, Gruzman A (2015) Med Chem 11:373–382

J Biol Inorg Chem

1 3

46. Daniel B, Green O, Viskind O, Gruzman A (2013) Amyotroph La Scl Fr 14:434–443

47. Miyahara T, Nakatsuji H, Sugiyama H (2013) J Phys Chem A 117:42–55

48. Zhao C, Ren J, Gregolinski J, Lisowski J, Qu X (2012) Nucleic Acids Res 40:8186–8196

49. Saha B, Islam MM, Paul S, Samanta S, Ray S, Santra CR, Choudhury SR, Dey B, Das A, Ghosh S, Mukhopadhyay S, Kumar GS, Karmakar P (2010) J Phys Chem B 114:5851–5861

50. Jangir DK, Dey SK, Kundu S, Mehrotra R (2012) J Photochem Photobiol 114:38–43

51. Sawai H, Domae N (2011) Biochem Bioph Res Co 411:569–573 52. Vermes I, Haanen C, Steffensnakken H, Reutelingsperger C

(1995) J Immunol Methods 184:39–51 53. van Engeland M, Ramaekers FCS, Schutte B, Reutelingsperger

CPM (1996) Cytometry 24:131–139 54. Goldstein S (1996) Oxford University Press, Oxford 55. Augusto O, Trindade DF, Linares E, Vaz SM (2008) An Acad

Bras Cienc 80:179–189 56. Samai M, Sharpe MA, Gard PR, Chatterjee PK (2007) Free Rad-

ical Bio Med 43:528–534 57. Narahara M, Hamada-Kanazawa M, Kouda M, Odani A, Miyake

M (2010) Biol Pharm Bull 33:1938–1943 58. Hadizadeh M, Keyhani E, Keyhani J, Khodadadi C (2009) Acta

Biochim Biophys Sinica 41:603–617 59. Kaizer J, Csonka R, Speier G, Giorgi M, Reglier M (2005) J Mol

Catal A Chem 236:12–17 60. Kaizer J, Csay T, Speier G, Reglier M, Giorgi M (2006) Inorg

Chem Commun 9:1037–1039 61. Johansson LH, Håkan Borg LA (1988) Anal Biochem

174:331–336

62. Wheeler CR, Salzman JA, Elsayed NM, Omaye ST, Korte DW Jr (1990) Anal Biochem 184:193–199

63. Oberley LWBIB, Sahu SK, Leuthauser SW, Gruber HE (1978) J Natl Cancer Inst 61:375–379

64. Oberley L, Oberley T (1988) Mol Cell Biochem 84:147–153 65. Holley A, Dhar S, Xu Y, St. Clair D (2012) Amino Acids

42:139–158 66. Cerutti P (1985) Science 227:375–381 67. Juarez JC, Manuia M, Burnett ME, Betancourt O, Boivin

B, Shaw DE, Tonks NK, Mazar AP, Donate F (2008) PNAS 105:7147–7152

68. Reddi AR, Culotta VC (2013) Cell 152:224–235 69. Gunther MR, Hanna PM, Mason RP, Cohen MS (1995) Arch

Biochem Biophys 316:515–522 70. Sigman DS, Graham DR, D’Aurora V, Stern AM (1979) J Biol

Chem 254:12269–12272 71. Baranovskii AG, Buneva VN, Nevinsky GA (2004) Biokhim

Biochem (Mosc.) 69:587–601 72. Manikandamathavan VM, Rajapandian V, Freddy AJ, Weyhermüller

T, Subramanian V, Nair BU (2012) Eur J Med Chem 57:449–458 73. Joyner JC, Reichfield J, Cowan JA (2011) J Am Chem Soc

133:15613–15626 74. Kumar P, Gorai S, Kumar Santra M, Mondal B, Manna D (2012)

Dalton Trans 41:7573–7581 75. Dixit N, Koiri RK, Maurya BK, Trigun SK, Höbartner C, Mishra

L (2011) J Inorg Biochem 105:256–267 76. Reddy PR, Shilpa A, Raju N, Raghavaiah P (2011) J Inorg Bio-

chem 105:1603–1612 77. Kao C-L, Tang Y-H, Lin YC, Chiu L-T, Chen H-T, Hsu SCN,

Hsieh K-C, Lu C-Y, Chen Y-L (2011) Nanomed Nanotech Biol Med 7:273–276