Cyto- and genotoxicity of a vanadyl(IV) complex with oxodiacetate in human colon adenocarcinoma (Caco-2) cells: potential use in cancer therapy

The complex of vanadyl(IV) cation with oxodiacetate, VO(oda) caused an inhibitory effect on the proliferation of the human colon adenocarcinoma cell line Caco-2 in the range of 25–100 μM (P < 0.001). This inhibition was partially reversed by scavengers of free radicals. The difference in cell proliferation in the presence and the absence of scavengers was statistically significant in the range of 50–100 μM (P < 0.05). VO(oda) altered lysosomal and mitochondria metabolisms (neutral red and MTT bioassays) in a dose–response manner from 10 μM (P < 0.001). Morphological studies showed important transformations that correlated with the disassembly of actin filaments and a decrease in the number of cells in a dose response manner. Moreover, VO(oda) caused statistically significant genotoxic effects on Caco-2 cells in the low range of concentration (5–25 μM) (Comet assay). Increment in the oxidative stress and a decrease in the GSH level are the main cytotoxic mechanisms of VO(oda). These effects were partially reversed by scavengers of free radicals in the range of 50–100 μM (P < 0.05). Besides, VO(oda) interacted with plasmidic DNA causing single and double strand cleavage, probably through the action of free radical species. Altogether, these results suggest that VO(oda) is a good candidate to be evaluated for alternative therapeutics in cancer treatment.


Introduction
A variety of vanadium compounds have been synthesized so far in an effort to optimize the antitumoral effects of this metal (Aubrecht et al. 1999;Cortizo et al. 2001;Djordjevic 1995).In particular, vanadium antineoplastic activity has been demonstrated both in vivo and in vitro (Narla et al. 2001;Molinuevo et al. 2004;Molinuelo et al. 2008;Samanta et al. 2008;Kostova 2009).In this aspect, the chemical and biochemical characteristics of ligands play a major role.
In aqueous media, strong chelating ligands are very important because they display the ability of trapping different metal species (Harding and Moksdi 2000).This process becomes particularly significant in living systems, where different metals are acquired, transported and stored mostly by low-molecular-weight compounds involving multidentate oxygen donors from carboxylate, hydroxamate and catecholate ligands (Lippard and Berg 1984;Costas et al. 2004).For vanadyl(IV) cation, complexation with organic ligands is crucial to maintain the solubility of the cation under physiological conditions and to increase its bioavailability (Etcheverry and Cortizo 1998).
Among the family of multidentate oxygen donor species, oda = oxodiacetate, O(CH 2 COO -) 2 holds an OOO donor group and can complex metal ions by forming chelate rings (Baggio et al. 2003).This ligand has been extensively used and several structures containing this anion in different environments have been obtained (Chao et al. 2003).
As it is mentioned above, vanadium compounds are very interesting from a pharmacological point of view due to their different therapeutics applications (Djordjevic 1995;Etcheverry and Cortizo 1998).
Vanadium compounds can exert their antitumoral actions through different mechanisms such as: reduction of cellular growth rate (antiproliferative effect), cytotoxic and genotoxic actions that may be triggered by the increment of oxidative stress or through different interactions at DNA level.
It is known that some inorganic compounds display DNA cleavage activity and for this reason they are considered inorganic nucleases (Butenko et al. 2009).These compounds are relevant as possible therapeutic agents in cancer chemotherapy.The V IV -complex vanadyl(acetylacetonate), V IV O(acac) 2 , has been proposed as a particularly efficient insulin-enhancing compound (Crans 2000).Besides, a recent study (Fu et al. 2008) has demonstrated its anticancer potential and mechanism of action in a human hepatoma cell line: V IV O(acac) 2 blocked the cell cycle permanently at the G1 phase on HepG2 cells.
The aim of the present study is to extend our previous observations on the bioactivity of VO(oda) on tumoral cells, focusing the attention on its cytotoxic and genotoxic effects in the human colon adenocarcinoma cell line Caco-2.

Preparation of VO(oda) solutions
VO(oda) (97% purity determined by thermal analysis) was synthesized and characterized according to literature (Del Rı ´o et al. 2003).
Fresh stock solutions of the complex and the free ligand were prepared in distilled water at 100 mM and diluted according to the concentrations indicated in the legends of the figures.

Cell culture and incubations
Caco-2 human colon adenocarcinoma cell line was grown in DMEM containing 10% FBS, 100 IU/ml penicillin and 100 lg/ml streptomycin at 37°C in 5% CO 2 atmosphere.
Cells in a 75 cm 2 flask were grown until reach 70-80% of confluence.Then, the cells were subcultured using 0.1% trypsin-1 mM EDTA in a Ca 2? -Mg 2? free phosphate buffered saline (PBS).For experiments, cells were grown in multi-well plates.When cells reached the desired confluence, the monolayers were washed with DMEM and were incubated under different conditions according to the experiments.

Cell proliferation: crystal violet assay
A mitogenic bioassay was carried out as described by Okajima et al. (1992) with some modifications.Briefly, cells were grown in 48 well plates.When cells reached 70% confluence, the monolayers were washed twice with serum-free DMEM and incubated with medium alone (basal) and different concentrations (10-100 lM) of the complex or the free ligand.Besides, incubations with a mixture of 50 lM of vitamin C and E and increasing concentrations of the complex were also carried out.Cells were pre-incubated with the scavengers of ROS during 2 h previous to the addition of different concentrations of the complex.Following pre-incubations, different concentrations of VO(oda) were added for 24 h at 37°C.After this treatment, the monolayers were washed with PBS and fixed with 5% glutaraldehyde/PBS at room temperature for 10 min.After that, they were stained with 0.5% crystal violet/25% methanol for 10 min.Then, the dye solution was discarded and the plate was washed with water and dried.The dye taken up by the cells was extracted using 0.5 ml/well 0.1 M glycine/HCl buffer, pH 3.0/30% methanol and transferred to test tubes.Absorbance was read at 540 nm after a convenient sample dilution.

NR assay
The NR accumulation assay was performed according to Borenfreund and Puerner (1984).Cells were plated in 96 well culture plates (2.5 9 10 4 cells/well).Cells were treated with different VO(oda) concentrations for 24 h at 37°C in 5% CO 2 in air.After treatment, the medium was replaced by one containing 100 lg/ml NR dye and cells were incubated for other 3 h.Then, NR medium was discarded, the cells were rinsed twice with warm (37°C) PBS (pH 7.4) to remove the non-incorporated dye, and 100 ll of 50% ethanol, 1% acetic acid solution was added to each well to fix the cells releasing the NR into solution.The plates were shaken for 10 min, and the absorbance of the solution in each well was measured in a Microplate Reader (7530, Cambridge technology, Inc, USA) at 540 nm, and compared with wells with untreated cells.Optical density was plotted as percentage of control.

MTT (methyl tetrazolium) assay
The MTT assay was based on a report previously described (Mosmann 1983).Briefly, cells were seeded in a 96-multiwell dish, allowed to attach for 24 h and treated with different VO(oda) concentrations for 24 h.After this treatment, the medium was changed and cells were incubated with 0.5 mg/ml MTT under normal culture conditions for 3 h.Cell viability was marked by the conversion of the tetrazolium salt MTT (3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl-tetrazolium-bromide) to a colored formazan by mitochondrial dehydrogenases.Color development was measured spectrophotometrically in a Microplate Reader (7530, Cambridge technology, Inc, USA) at 570 nm after cell lysis in DMSO (100 ll/well).Cell viability is shown graphically as percent of the control value.

Cell morphology
Cells were grown on glass coverslips and incubated under control conditions (without complex addition) or with different concentrations of VO(oda) in serumfree DMEM for 24 h.Then, the cells were fixed and stained with Giemsa (Cortizo et al. 2000;Etcheverry et al. 2007).Samples were observed under light microscopy.

Cytoskeleton actin determination
A fluorescence technique was used to visualize actin cytoskeleton filaments.Cells were grown on glass coverslips until a 70% of confluence; after that the cells were incubated for 24 h at 37°C with different concentrations of VO(oda).Then, the cells were fixed, permeabilized during 4 min at room temperature using absolute ethanol (cold at -20°C), washed with PBS and blocked with non-fat milk for 2 h.After washing, the cells were incubated with Phalloidin-FITC for 2 h.The samples were washed twice and mounted in slides.Labeled cells with green fluorescence were observed using a fluorescence microscope.

Determination of reactive oxygen species (ROS) production
The possible VO(oda)-induced oxidative stress in the Caco-2 cells was evaluated by measurement of intracellular production of ROS after incubation of the cell monolayers with different concentrations of the complex during 4 h at 37°C.ROS generation was determined by oxidation of DHR-123 to rhodamine by spectrofluorescence as we have previously described (Cortizo et al. 2000).A mixture of 50 lM of vitamin C and E was used to evaluate the role of the oxidative stress in the complex cytotoxicity under similar concentrations and incubation time to those employed in the crystal violet assay.

Fluorometric determination of GSH and GSSG cellular levels
GSH and GSSG levels were determined in culture as follows.Confluent Caco-2 monolayers from 24 well dishes were incubated with different doses of VO(oda) for 24 h at 37°C.Then, the monolayers were washed with PBS and harvested by incubating with 250 ml Triton 0.1% for 30 min.For GSH determination, 100 ml aliquots were mixed with 1.8 ml of ice cold phosphate buffer (Na 2 HPO 4 0.1 M-EDTA 0.005 M pH 8) and 100 ml o-phthaldialdehyde (0.1% in methanol) as it was described by Hissin and Hilf (1996).For the determination of GSSG, 100 ml aliquots were mixed with 1.8 ml NaOH 0.1 M and ophthaldialdehyde as before.Previously, to avoid GSH oxidation the cellular extracts for GSSG determination were incubated with 0.04 M of N-ethylmaleimide (NEM).Fluorescence at an emission wavelength of 420 nm was determined after excitation at 350 nm.Standard curves with different concentrations of GSH or GSSG were processed in parallel.Protein content in each cellular extract was quantified using Bradford assay (Bradford 1976).The ratio GSH/GSSG was calculated for all the experimental conditions.

Single cell gel electrophoresis (SCGE) assay
For detection of DNA strand breaks the SCGE ('Comet') assay was used in the alkaline version, based on the method of Singh et al. (1988) with minor modifications.Under alkaline conditions, DNA loops containing breaks loose supercoiling, unwind and are released from the nuclei and form a 'Comet-tail' by gel electrophoresis.For this experiment, 2 9 10 4 cells were seeded in a 12-well plate; 24 h later the cells were incubated with various concentrations of VO(oda).After treatment, cells were suspended in 80 ll of 0.5% low melting point agarose (pH 7.4) and immediately pipetted onto a frosted glass microscope slide precoated with a layer of 0.5% normal melting point agarose to promote a firm attachment of the second layer.After these layers had solidified at 4°C for 5-10 min, slides were immersed in ice-cold freshly prepared lysis solution (2.5 M NaCl, 10 mM Tris, 100 mM Na 2 EDTA, 1% Triton X-100, 10% DMSO, pH 10) at darkness for 1 h (4°C) in order to lyse the cells, remove cellular proteins and to permit DNA unfolding.Immediately after, slides were put in a horizontal electrophoresis tank containing 1 mM Na 2 EDTA, 0.3 M NaOH (pH 12.7) and then electrophoresis was performed for 30 min at 25 V (4°C).Afterwards, slides were neutralised with a solution comprising 0.4 M Tris-HCl (pH 7.5) and stained with 4 0 ,6-diamidino-2-phenylindole (DAPI; Vectashield mounting medium H1200).Analysis of the slides was performed in an Olympus BX50 fluorescence microscope equipped with appropriated filter combination with an X63 fluorescence objective.Cellular images were acquired with the Leica IM50 Image Manager (Imagic Bildverarbeitung AG).A total of 50 randomly captured cells per experimental point of each experiment was used to determine the tail moment (product of tail length by tail DNA percentage) using a free Comet scoring software (Comet Score version 1.5).Two parallel slides were performed for each experimental point.Independent experiments were repeated twice.

Agarose gel electrophoresis (AGE) of plasmid DNA
The plasmid DNA used for gel electrophoresis experiments was pA1, which consists of a full-length cDNA from Cytochrome P450 CYP3A1 inserted in the PBS plasmid vector (pBluescribe, Strata-gene, UK) and described elsewhere (Fisher et al. 1998).Plasmid DNA was amplified in Escherichia coli DH5a and purified using PureYieldTM Plasmid Midiprep System from Promega.Linear DNA was obtained by digestion of pA1 with HindIII and used as a reference in AGE.This reference was replaced, in some experiments, with plasmid DNA digested with VO(acac) 2 in conditions known to produce the linearized DNA form (Butenko et al. 2009).
DNA concentration per nucleotide base pair (bp) was determined by UV absorption at 260 nm using the extinction coefficient of 13,200 M -1 cm -1 bp -1 .
A 200 lM mother solution of vanadium complex in deionized MilliQ water was freshly prepared for each experiment.Both complex and ligand are soluble in water.
DNA cleavage activity was evaluated by monitoring the conversion of supercoiled plasmid DNA (Sc) to nicked circular DNA (Nck) and linear DNA (Lin).Each reaction mixture was prepared by adding (in this order) 6 ll of water, 2 ll of 100 mM stock pH 7 buffer solution, 2 ll (0.2 lg) of supercoiled pA1 DNA and 10 ll of the aqueous solution of the complex.The buffers used in these experiments were MOPS/NaOH and phosphate/HNO 3 .The final reaction volume was 20 ll and the final metal concentrations tested were 6, 12, 25, 50 and 100 lM.The final buffer concentration was 10 mM.When the reaction involved additional activating agents or radical scavengers, the initial volume of water was reduced to 4 ll, and 2 ll of a solution of the agent or scavenger were added before the metal complex.MPA and oxone were chosen as reducing and oxidizing activating agents, respectively.Several known OH radicals scavengers (Halliwell and Gutteridge 2007) sodium azide, DMSO and NaBz were used in the experiments.The final concentration was 200 lM for activating agents and 40 mM for scavengers.The control samples were prepared in the absence of metal complex.The sample containing hydrogen peroxide was prepared in the same way as those with activating agents or scavengers.The stock solution of *1 mM hydrogen peroxide was freshly prepared from 30% H 2 O 2 , its concentration was measured spectrophotometrically (e 230nm = 74 M -1 cm -1 ) (Mueller et al. 1997).
Agarose powder (1.5 g) from Sigma-Aldrich was weighed following by adding 150 ml of 0.59 TBE (89 mM Tris-Borate, 1 mM EDTA, pH 8.3).The mixture was heated in a microwave oven (at middle power) until the agarose dissolved completely.10 ll (10 lg/ml) of ethidium bromide were added upon cooling the mixture to approximately 65-70°C and poured on a clean casting plate with an appropriate comb and left for approximately 1 h at room temperature to solidify before using.
Samples were incubated for 1 h at 37°C, wrapped up in aluminium foil.After incubation, 5 ll of DNA loading buffer (0.25% bromophenol blue, 0.25% xylene cyanol, 30% glycerol in water) were added to each tube and the solution was loaded onto a 1% agarose gel in TBE buffer.Controls of non-incubated and of linearized plasmid were included in both extremes of a 16-well gel plate.The electrophoresis was carried out for 3 h at 100 V. Bands were visualized under UV light and photographed using an AlphaImager (Alpha Innotech).
Peak areas were measured by densitometry using AlphaEaseFC TM Software from Alpha Innotech.Peak areas for the Sc form were corrected using the factor 1.47 to account for its lower staining capacity by ethidium bromide (Bernadou et al. 1989) and used to calculate the percentage (%) of each form (Sc, Nck and Lin).The variability of results was estimated from the analysis replicate digestions using VO(acac) 2 as a reference standard metallonuclease (Butenko et al. 2009).Standard deviation for repeatability (s r which measures the random variability within each experiment) for the peak area within the same gel was estimated as 0.8% for the Sc form and 6% for the Nck form.The s r of the whole experimental procedure, including the preparation of solutions and incubation was 10, 6 and 3%, for the Sc, Nck and Lin forms, respectively.These values were estimated from analysis of variance (single factor ANOVA) of duplicates from 11 different runs.s r for the peak area was estimated from duplicates of plasmid DNA (not incubated).s r for the whole procedure was estimated from duplicates of samples incubated with metal complex (Butenko et al. 2009).
All experiments were done in duplicate, and the gel with best quality was presented as example.

Statistical analysis
For statistical analysis the following methods were used: One-way ANOVA followed by Newman-Keuls test for comparison of the differences and the Student t test for the comparison of two means.Results represent the mean ± SEM.All the results with P \ 0.05 were considered statistically significant.

Effects of VO(oda) on Caco-2 cell proliferation
As can be observed in Fig. 1, VO(oda) exerted an inhibitory effect on Caco-2 cell proliferation in a dose response manner.This effect was statistically significant in relation to basal condition (P \ 0.001) from 25 lM concentration.IC 50 for the complex in Caco-2 cell line was )100 lM.Besides, the free ligand did not cause any effect on cell proliferation (data not shown).
The inhibition of Caco-2 proliferation was partially ameliorated by scavengers of free radicals as a mixture of vitamin C and E (50 lM).In the presence of vitamins, the deleterious effect of the complex can be seen at a concentration higher than in the absence of scavengers.The differences between these two conditions began to be statistically significant in the range of 50-100 lM (P \ 0.05).
In this figure, the beneficial effect of 50 lM vitamin C plus vitamin E mixture can be seen.
To get a better understanding of the cytotoxic effect of the complex, we next studied its effects on the lysosomal and mitochondrial activity by means of NR uptake and MTT assay, respectively.

Effects of VO(oda) on lysosomal activity: NR assay
The metabolically active lysosomes display the capacity of uptake the NR dye.A loss of lysosomal activity, indicated by a decrease in the uptake of NR was observed when Caco-2 cells were exposed to VO(oda).The cytotoxic effects of VO(oda) in this cell line affects the functions of these organelles.This effect can be observed in Fig. 2.
As can be seen, the lower tested concentrations (2.5 and 5 lM) did not show any difference with the control condition (without complex addition).Nevertheless, at 10 lM a statistically significant decrease in the uptake of the dye by the lysosomes could be detected (P \ 0.001).As the concentrations of VO(oda) increase, a marked cytotoxic effect could be observed in the 25-100 lM range (P \ 0.001).
These results show that the alterations in the lysosomal activity begin at a concentration lower (10 lM) than the dose at which proliferation is statistically diminished (25 lM).At the latter concentration the lysosomal activity was reduced to 50% of the basal.Effects of VO(oda) on mitochondrial activity: MTT assay Alteration in the energetic cell metabolism induced by VO(oda) measures the ability of the mitochondrial succinic dehydrogenases to reduce the methyl tetrazolium salt (MTT assay).The results presented in Fig. 3 show a decrease in the MTT dye reduction to an insoluble violet formazan product when Caco-2 cells were exposed to different complex concentrations (2.5-100 lM).In the presence of VO(oda), a significant decrease (P \ 0.001) was observed in the absorbance for concentration values equal or higher than 10 lM.Cell metabolism inhibition in the 10-25 lM concentration range was lower (but still above 50%) than in the 50-100 lM range.At the higher tested concentrations the decreased in cell metabolism reached a plateau (ca.45%).These results are in parallel with those of the NR uptake, suggesting that the cytotoxic actions of the complex affected the normal activity of mitochondria, contributing also in this way to the deleterious effect of VO(oda).

Morphological changes
Next we determined the effect of this complex on the morphology of Caco-2 cells.Morphological studies were carried out using Giemsa staining and light microscopy.In basal conditions, the cells displayed polygonal shape and big nuclei with numerous nucleoles.Cells present also numerous connections between each other (Fig. 4a).Upon incubation with the 25 and 100 lM complex concentrations, the nuclei showed different alterations in their form such as chromatine condensation and apoptotic bodies.Besides some membrane blebs could be seen especially at the highest dose (Fig. 4c).In addition, the cytoplasm of the cells also showed important transformations such as numerous irregular vacuoles and lost of neighboring connections.These changes increased with the doses and correlated with a decrease in number of cells per field (Fig. 4b, c).An important number of cells died and detached from the monolayers (Fig. 4c).Altogether, these are typical features of apoptosis, a process of cell death induced by the treatment with the complex.

Interactions of VO(oda) with the actin filaments of cytoskeleton
The cytoskeleton network can be observed by the staining of the actin fibers with Phalloidin-FITC, a fluorescent probe with high affinity for the cytoskeleton.The disassembly of the network of actin microfilaments as a function of complex concentration can be observed in Fig. 5.Under normal conditions (Fig. 5a), the actin microfilaments were placed in the direction of the main axis of the cells.In the presence of the vanadium complex (25 and 100 lM), a rearrangement of the actin network could be observed.At 25 lM (Fig. 5b), VO(oda) caused a strong alteration in the actin arrangement but a higher effect was observed at 100 lM (Fig. 5c), a dose at which practically a total lost of the actin network could be seen.At this concentration, a complete disorganization was visualized with the fibers surrounding the nuclei of the cells and the shape of cells completely altered.All these actin changes were in parallel with serious disruption of cell shape and size.This result is in accordance with the morphological and the proliferation studies.

Genotoxicity: induction of DNA damage assessed by the Comet assay
The SCGE (Comet) assay was carried out in order to elucidate the genotoxic effect of VO(oda).In this assay three parameters are currently used: the Comet Length, the Tail Length and Tail Moment which are determined using a scoring software.The two first parameters are based on the migration of DNA.Most commonly, the distance of DNA migration is used to measure the extent of DNA damage.However, if DNA damage is relatively high, the tail increases in fluorescent staining intensity but not in length (Liao et al. 2009).Thus, it is also convenient to use the Tail Moment.This parameter is calculated as: Tail Length 9 DNA amount in the tail.DNA amount is determined through the intensity of fluorescence.
As can be seen from Fig. 6, the treatment with the complex produced an increment in the parameters of the standard alkaline Comet assay (Comet Length, Tail Length and Tail Moment).The genotoxic action of VO(oda) was already detectable at 5 lM (P \ 0.001 for the Comet Length and Tail Length parameters).Besides, VO(oda) caused an increase of Tail Moment value in a concentration dependent fashion (P \ 0.01 at 5 lM and P \ 0.001 from 10 to 25 lM).The Comet assay detects DNA damage due to direct genotoxic impact of different agents or it can also correlate with DNA degradation due to necrosis or apoptosis.The two latter processes could introduce a false positive response due to cytotoxic events.Henderson et al. (1998) indicated that the maximum concentration of tested agent should produce viabilities[75% in order to avoid cytotoxicity-induced false positive responses in the study of genotoxic effects.In our study, the percentage of viable cells observed in the proliferation assay indicates the lack of overt cytotoxic effects in the range of concentrations within 2.5-10 lM, i.e. the viability remains above 75%.However, the DNA damage detected at higher VO(oda) concentrations (25 lM) may be due to secondary events; i.e. related to the cytotoxicity of VO(oda).

Oxidative stress
In order to get a deeper insight into the mechanisms underlying the cytotoxicity of VO(oda), we determined its effect on the oxidative stress using two parameters: the oxidation of dihydrorhodamine 123 to rhodamine (measured by spectrofluorometry) and the changes in the ratio of reduced glutathione (GSH) to oxidized glutathione (GSSG).
The complex produced a great increment in the levels of rhodamine in the range of 50-100 lM (P \ 0.001) (see Fig. 7).Besides, the mixture of vitamins C and E, scavenger of free radicals, caused a decrease in the oxidative stress produced by VO(oda) in the range of 50-100 lM (P \ 0.05).
Glutathione is the major antioxidant found in most mammalian tissues.Its oxidation to GSSG generates a redox buffer that protects cells from oxidative stress.Under oxidant conditions, the GSH/GSSG ratio diminishes.After treatment of Caco-2 tumoral cells with the complex, we found that GSH levels decreased in a dose dependant manner in the range of 50-500 lM (P \ 0.05) (Fig. 8a).Moreover a decrease in the GSH/GSSG ratio was also observed for 100 and 500 lM (P \ 0.05) (Fig. 8b).Effect of VO(oda) on plasmidic DNA: AGE From Fig. 9 it is clear that VO(oda) cleaves DNA under relatively mild conditions (1 h digestion at 37°C).The intensities of the nicked (Nck) bands are much stronger than the supercoiled (Sc) bands in DNA samples digested with VO(oda).This indicates an extensive single-strand cleavage of DNA at relatively low ratios of metal:bp (1.7).Double-strand cleavage of DNA is also detected since a light band corresponding to the linear form is observed in all samples digested with VO(oda).The addition of mercaptopropionic acid (MPA) did not affect the activity of VO(oda) significantly, but increased it slightly.Addition of the strong oxidant oxone (KHSO 5 ) induced a further increase in the activity, observed by the stronger intensity of both the Nck and Lin bands.
It is known that the nature of the pH buffer may affect the observed nuclease activity, particularly when changing from organic to inorganic buffers (Butenko et al. 2009).In order to evaluate any effect of the buffer, parallel experiments with variable complex concentration were carried out at pH 7.0 under two different pH buffers, phosphate and MOPS (Fig. 10).In both cases, the extent of DNA cleavage is closely dependent with complex concentration.The activity of VO(oda) can be already observed starting from the most diluted sample-6 lM-in both buffers.Nevertheless, VO(oda) shows a higher activity under phosphate buffer than under MOPS.
The effect of three radical scavengers-sodium azide (NaN 3 ), dimethyl sulfoxide (DMSO) and sodium benzoate (NaBz)-is shown in Fig. 11.Under MOPS buffer it is clear that NaN 3 quenches the activity of VO(oda) much more strongly than DMSO and NaBz.Unlike NaBz and DMSO, NaN 3 may scavenge singlet oxygen species as well as hydroxyl radicals and other ROS.The fact that the reaction is Results are expressed as the mean ± SEM of three independent experiments, *significant differences versus basal, P \ 0.05 Fig. 9 AGE of plasmid DNA digested with VO(oda) with and without activating agents.The graphic shows percentages of Sc, Nck and linear forms, measured from the corresponding AGE bands.Controls for native DNA and for the linear form are represented as ''DNA'' and ''Lin''. 1 VO(oda), 2 VO(oda) ?MPA, 3 VO(oda) ?Oxone.VO(oda) concentration is 50lM (r.i.= 1.7), solutions are buffered at pH 7.0 with MOPS.Error bars were estimated as ±2 s for 95% confidence intervals reduced more efficiently in the presence of NaN 3 than the other scavengers suggests that singlet oxygen may play a role in the activity of VO(oda) under MOPS buffer.It is interesting to note that some linearization of DNA was still observed in all samples, as if the quenching was only effective for the single-strand cleavage and not the double-strand one.This residual activity points to the possibility of another, nonradicalar mechanism, taking place.
Under phosphate buffer both NaN 3 and NaBz decrease the nuclease activity to the same extent, and no linearization is observed in samples containing radical scavengers.Some residual activity is still observed after the addition of the radical scavengers, though weaker than under MOPS buffer.
These results point to different mechanisms of DNA cleavage by VO(oda) in these buffers.It is possible that under phosphate buffer different species Figure 12 compares the nuclease activity of VO(oda) under phosphate buffer with other active vanadyl complexes such as VO(acac) 2 , VO(hd) 2 and VO(Et-acac) 2 previously tested in similar conditions (Butenko et al. 2009;Costa Pessoa et al. 2007;Cavaco et al. 2009) showing that VO(oda) is quite an efficient nuclease.It has a similar activity to VO(hd) 2 , lower than the nuclease activity of VO(acac) 2 and higher than VO(Et-acac) 2 .
Vanadium complexes that belong to the group of non-platinum metal antitumor agents have shown to possess a wide spectrum of antineoplastic properties.These compounds appear to offer a different alternative for cancer chemotherapy which does not follow the mechanism of action of the platinum complexes (Farrell 1999).
As part of a research project devoted to the investigation of vanadium complexes with potential pharmacological applications, we have tested the antitumoral effects of a complex of vanadyl(IV) cation with a multidentate oxygen donor, oxodiacetate, in a model system of human colon adenocarcinoma Caco-2 cells in culture.In this paper, we have investigated the cyto-and genotoxic effects of VO(oda) and also we have tried to elucidate the putative mechanism of action underlying its effects.We have studied the effects of VO(oda) on cell proliferation, cell morphology and on the cytoskeleton actin fibers.We observed a deleterious effect of the complex on the proliferation of Caco-2 cells in culture (Fig. 1).This finding is in agreement with a previous paper reporting the inhibition of cell proliferation by the same vanadium(IV) complex in a rat osteosarcoma derived cell line (UMR106) (Rivadeneira et al. 2007).In both tumoral cell lines the inhibitory action of VO(oda) begun at 25 lM.The sensitivity of the rat osteosarcoma cells towards the complex seems to be slightly higher than the Caco-2 cells since at 100 lM there was ca 65% of surviving cells versus ca.55% for the UMR106 cells.
The cytotoxic effects of VO(oda) on Caco-2 cells may be due to different causes such as the functional disruption of several organelles like the lysosomes and the mitochondria.
The metabolic activity of the lysosomes may be determined by the uptake of NR supravital dye.Viable cells will take up the dye, so any decrease in the uptake will serve as an indicator of cytotoxicity.On the other hand, the mitochondrial dehydrogenase enzyme activity can be measured through the determination of the reduction of MTT which is colored in its reduced form.After 24 h incubation with different concentrations of the complex, a reduction in these two important metabolic functions could be determined in Caco-2 cells in the 10-100 lM range (Figs. 2, 3).According to these results, the cytotoxic effects began at a lower concentration than the determined decrease in cell proliferation studies with the crystal violet assay.These results are in agreement with previous reports on the action of different vanadium species in several cell types.Decrease cell viability was induced by vanadyl sulphate in  (Butenko et al. 2009;Costa Pessoa et al. 2007;Cavaco et al. 2009).Cleavage efficiency is measured by the slope of the straight line of the logarithm of the percentage of Sc (log SC) with complex concentration tumorigenic and non carcinogenic cells (Holko et al. 2008), as well as the toxicity of vanadate solution in cardiomyocytes (Soares et al. 2008).Moreover, it has been also previously reported that vanadocenes blocked the cell division in human cancer cells by disrupting bipolar spindle formation.Vanadocene treatment caused an arrest at the G2/M phase of the cell cycle.This unique mechanism of anti-mitotic function warrants further development of vanadocene complexes as anti-cancer drugs (Navara et al. 2001).It has also been shown that vanadocene compounds induced apoptosis in human testicular cancer cell lines (Ghosh et al. 2000).It was also reported that the toxicity of vanadyl complexes on Caco-2 monolayer involved F-actin-related change of tight junction and impairment of microvilli.The toxicity was also related to elevated intracellular ROS and their cellular accumulation (Yang et al. 2004).VO(oda) has been shown to display a strong effect on the architecture of the actin cytoskeleton fibers which arrangement was affected in a dose response manner for the vanadyl complex (Fig. 5).The observed disruption of the actin fibres induced by VO(oda) may be related to toxic effects on the mitotic spindle in a way similar to that previously reported for vanadocenes.These changes also may be the cause of the observed apoptotic morphological alterations (Fig. 4) which were similar to those observed in the model of osteosarcoma cells and confirmed by flow cytometry studies (Rivadeneira et al. 2010).It has been reported that changes in actin fibers and cell-tocell contact are involved in the antimetastatic potential of vanadium compounds (Dharwan et al. 1997;Luber et al. 2000).
In an attempt to elucidate the mechanism involved in the effects of VO(oda), we studied the relationship between the oxidative stress, GSH levels and the antiproliferative effect of VO(oda) in the tumoral cells.The results showed that GSH levels as well as the ratio GSH/GSSG (Fig. 8), which is a better marker of the GSH metabolism, decreased in parallel with the increment in the oxidative stress determined by the DHR123 oxidation (Fig. 7).Oxidative stress exerted on cells by other vanadium(IV) compounds has been previously reported (El-Naggara et al. 1998;Younes et al. 1991).
Indeed, this effect has been regarded as one of the major effects implicated in their anti tumorogenic activity, while a rise in GSH levels upon treatments has been observed in therapy resisting cells (Majumder et al. 2006).
On the other hand, it was observed a decrease in Caco-2 cell proliferation in the range of 50-100 lM (Fig. 1).This toxic effect could be ameliorated by the pre-incubation with a mixture of ROS scavengers (mixture of vitamin C and E) (Fig. 1), suggesting that the main mechanism involved in the VO(oda) cytotoxicity was the oxidative stress.
Vanadium compounds may also exert antiproliferative effects via interactions with DNA.It was previously reported that vanadocene complexes interact with DNA's nucleotide phosphate groups forming a labile outer sphere complex via a water group (Harding and Moksdi 2000;Harding et al. 1993;Kopf-Maier et al. 1981;Murray and Harding 1994;Toney et al. 1986).DNA cleavage in vivo and in vitro is also induced by a variety of vanadium compounds (Altamirano-Lozano et al. 1999;Lloyd et al. 1997;Rojas et al. 1996;Sakurai 1994;Sakurai et al. 1995;Sit et al. 1996).Several lines of evidence indicate that V(IV) and V(V) induce DNA cleavage through the production of ROS, especially hydroxyl radicals (OH • ) generated in cells (Sakurai 1994).
Taking into consideration that genotoxicity is a process that can begin before the alterations at cytoplasm level could be detected; we investigated the effect of the complex at DNA level.Comet assay showed that VO(oda) induced alterations in the nuclei of the cells at 5 and 10 lM (Fig. 6), which indicated that the changes in DNA begun before any alteration in cell morphology, actin fibers, oxidative stress and cell number diminution could be observed.
The genotoxicity of VO(oda) was evaluated using the single cell gel electrophoresis (Comet assay).Vanadium has been considered an aneugen (any agent that affects cell division and the mitotic spindle apparatus resulting in the loss or gain of whole chromosomes).However, there is controversial information about the clastogenic effects of vanadium compounds (Rodrı ´guez-Mercado et al. 2003).Our results showed a significant increase in three parameters evaluated with the Comet assay; this indicates that VO(oda) is able to induce DNA single and double-strand breaks and/or alkali-labile damage at very low doses (5 lM) (Fig. 6).Our results agree quite well with previous studies on the effects of vanadium compounds in different cell lines.Vanadium pentoxide produced a clear dose-response in

Fig. 1
Fig. 1 Effect of VO(oda) on Caco-2 cell proliferation.Cells were incubated in serum-free DMEM without (basal) or with different concentrations of VO(oda) (2.5, 5, 10, 25, 50, 75 and 100 lM) at 37°C for 24 h.Results are expressed as % basal and represent the mean ± SEM (n = 9) *, # significant differences versus control, P \ 0.05, o significant differences in relation to the same concentration of the complex in absence of the mixture of vitamins, P \ 0.05

Fig. 3
Fig. 3 MTT assay.Caco-2 cells were incubated with different doses of VO(oda) for 24 h at 37°C.After incubation, cell viability was determined by the conversion of the tetrazolium salt MTT to a colored formazan by mitochondrial dehydrogenases.Color development was measured at 570 nm after cell lysis in DMSO.Results are expressed as % basal and represent the mean ± SEM (n = 16), ***P \ 0.001

Fig. 5
Fig. 5 Effect of VO(oda) on the actin cytoskeleton filaments of Caco-2 cells.Cells were incubated with different concentrations of VO(oda) for 24 h at 37°C.Then, the cells were

Fig. 6 Fig. 7
Fig. 6 SCGE (Comet assay).Induction of DNA damage by VO(oda) in Caco-2 cells was evaluated by a Tail Length, b Comet Length and c Tail Moment.After incubation with VO(oda) for 24 h, cells were lysed.After electrophoresis, nuclei were stained and analyzed.Results are expressed as mean ± SEM (n = 150), ***P \ 0.001

Fig. 8 a
Fig. 8 a GSH level in Caco-2 cells in presence of different VO(oda) concentrations.Results are expressed as the mean ± SEM of three independent experiments, *significant differences versus basal, P \ 0.05.b GSH/GSSG ratio in Caco-2

Fig. 10
Fig. 10 Evolution of DNA degradation with increasing concentration of VO(oda).a Solutions buffered at pH 7.0 with MOPS, b solutions buffered at pH 7.0 with phosphate buffer

Fig. 12
Fig. 12 Comparison of cleavage activity of VO(oda) with other vanadium nucleases: VO(acac) 2 , VO(hd) 2 and VO(Et-acac) 2 (Butenko et al. 2009; Costa Pessoa et al. 2007; Cavaco et al. 2009).Cleavage efficiency is measured by the slope of the straight line of the logarithm of the percentage of Sc (log SC) with complex concentration