Chemico-Biological Interactions 232 (2015) 85–93

Contents lists available at ScienceDirect

Chemico-Biological Interactions

journal homepage: www.elsevier .com/locate /chembioint

Anti-cancer effects of cerium oxide nanoparticles and its intracellularredox activity

http://dx.doi.org/10.1016/j.cbi.2015.03.0130009-2797/� 2015 Elsevier Ireland Ltd. All rights reserved.

⇑ Corresponding authors.E-mail addresses: camala@ibiss.bg.ac.rs (M. Pešic), xenia@imsi.bg.ac.rs

(K. Radotic).

Milica Pešic a,⇑, Ana Podolski-Renic a, Sonja Stojkovic a, Branko Matovic b, Danica Zmejkoski b,Vesna Kojic c, Gordana Bogdanovic c, Aleksandra Pavicevic d, Miloš Mojovic d, Aleksandar Savic e,Ivana Milenkovic e, Aleksandar Kalauzi e, Ksenija Radotic e,⇑a Institute for Biological Research ‘‘Siniša Stankovic’’, University of Belgrade, Despota Stefana 142, 11060 Belgrade, Serbiab Department of Material Science, Vinca Institute of Nuclear Sciences, Vinca, Serbiac Institute of Oncology Sremska Kamenica, Sremska Kamenica, Serbiad Faculty for Physical Chemistry, University of Belgrade, Studentski trg 12-16, Belgrade, Serbiae Institute for Multidisciplinary Research, University of Belgrade, Kneza Višeslava 1, 11000 Belgrade, Serbia

a r t i c l e i n f o

Article history:Received 8 January 2015Received in revised form 13 February 2015Accepted 17 March 2015Available online 24 March 2015

Keywords:Cerium oxideOxygen vacanciesFree radicalsCytotoxicityFlow cytometryElectron spin resonance spectroscopy

a b s t r a c t

Data on medical applications of cerium oxide nanoparticles CeO2 (CONP) are promising, yet informationregarding their action in cells is incomplete and there are conflicting reports about in vitro toxicity.Herein, we have studied cytotoxic effect of CONP in several cancer and normal cell lines and theirpotential to change intracellular redox status. The IC50 was achieved only in two of eight tested cell lines,melanoma 518A2 and colorectal adenocarcinoma HT-29. Self-propagating room temperature methodwas applied to produce CONP with an average crystalline size of 4 nm. The results confirmed presenceof Ce3+ and O2� vacancies. The induction of cell death by CONP and the production of reactive oxygenspecies (ROS) were analyzed by flow-cytometry. Free radicals related antioxidant capacity of the cellswas studied by the reduction of stable free radical TEMPONE using electron spin resonance spectroscopy.CONP showed low or moderate cytotoxicity in cancer cell lines: adenocarcinoma DLD1 and multi-drugresistant DLD1-TxR, non-small cell lung carcinoma NCI-H460 and multi-drug resistant NCI-H460/R, whilenormal cell lines (keratinocytes HaCaT, lung fetal fibroblasts MRC-5) were insensitive. The most sensitivewere 518A2 melanoma and HT-29 colorectal adenocarcinoma cell lines, with the IC50 values beingbetween 100 and 200 lM. Decreased rate of TEMPONE reduction and increased production of certainROS species (peroxynitrite and hydrogen peroxide anion) indicates that free radical metabolism, thusredox status was changed, and antioxidant capacity damaged in the CONP treated 518A2 and HT-29 cells.In conclusion, changes in intracellular redox status induced by CONP are partly attributed to theprooxidant activity of the nanoparticles. Further, ROS induced cell damages might eventually lead tothe cell death. However, low inhibitory potential of CONP in the other human cell lines tested indicatesthat CONP may be safe for human usage in industry and medicine.

� 2015 Elsevier Ireland Ltd. All rights reserved.

1. Introduction

Cerium oxide CeO2 (CONP) crystal lattice, consists of a ceriumcore enveloped by an oxygen lattice. These particles have an inter-esting feature, coexistence of Ce3+ and Ce4+ ions and the ability ofoxygen vacancies formation on their surface, which enables themto interact with and modulate free radicals. In such way thesenanoparticles can pass many cycles of reactions with free radicals

and thus regenerate themselves in every cycle [1,2]. This is aunique feature making CONP advantageous in comparison withthe other nanoparticles. Besides, compared with other rare earthelements, CONPs have a high hydrogen-absorbing capacity, suchthat reactions with H2, O2, or H2O occur more readily, which mayaccount for its regenerative capacity as a catalyst [3]. As such, theydisplay antioxidant/antiradical activity [4,5] that encouraged stud-ies of pharmacological potential of these nanoparticles, as well astheir biomedical applications [6–8].

Some studies have shown protective effect of the CONP towardsoxidant-mediated apoptosis [9,10]. However, in certain studiesCONP induced oxidative stress either in vitro or in vivo [11,12]. It

86 M. Pešic et al. / Chemico-Biological Interactions 232 (2015) 85–93

seems that environmental conditions, such as pH, direct oppositeCONP activity regarding its oxidant or antioxidant properties[13]. In a neutral pH environment such as in normal cells, CONPsare cytoprotective and act as antioxidants. However, in an acidicpH environment characteristic for cancer cells CONPs may becytotoxic and act as pro-oxidants.

Reactive oxygen species (ROS) accumulation in cells is generallyassociated with undesired effects. ROS generation, which alsoincludes the generation of reactive nitrogen species (RNS) con-tributes to diabetes, neurodegenerative diseases, aging, and cancer.In cancer, ROS can trigger its development and progressionthrough unique adaptation to oxidative stress and deregulationof antioxidant enzymes expression [14]. It is now widely acceptedthat ROS can interfere in intracellular processes, thus leading to theinjuries. It has been demonstrated that several types of nanoparti-cles act as potent free radical scavengers and antioxidants. Thusnanoparticle antioxidants may counteract disorders/diseaseswhich are related to oxygen levels and ROS production [15].

Rare studies of the CONP activity in cancer showed that CONPpossess anti-invasive properties towards some cancer cells [16],induces radio-sensitisation [17], and simultaneously radio-protection of normal cells by regulating antioxidant enzymes andquantity of ROS [18,19].

Potential adverse health and ecological and environmentalimplications associated with the exposure of CONP are very lim-ited, and the results vary among research groups, from absenceof negative effects [11,20] to less or more deleterious effects onthe organs and accumulation in some organs such are kidneysand lungs [21,22]. The final in vivo effects of CONP may dependon the way of their application and require further studies [6].Because of wide potential for human usage in industry andmedicine, we need more information about CONPs cytotoxicity inhuman cells.

In this study, we have evaluated the anti-cancer effects of CONPin vitro and its selectivity towards human cancer cells. To that end,the cytotoxic effect of CONP was investigated in different humannormal and cancer cell lines. In order to see whether CONP inducechanges in free radicals related antioxidant capacity of the cells, westudied total free radical production by the reduction of stable freeradical TEMPONE in living cells using electron spin resonance (ESR)spectroscopy. Also, production of ROS/RNS and induction of celldeath were analyzed by flow-cytometry. We chose pairs of thenormal and cancer cell lines, to see if the effect of CONPs, regardingtheir interaction with free radical related antioxidant capacity ofthe cells differs in these cells.

2. Materials and methods

2.1. Drugs

The 10 mM CONP suspension was prepared in ethanol and dis-persed for 20 min by using a sonicator (Bandelin Sonorex, Berlin,Germany). Before the experiment the suspension was diluted in50% ethanol and vortexed. Cisplatin (CPt) was obtained fromPfizer (Perth) Pty. Ltd., Bentley, Australia. CPt was kept at roomtemperature. Before treatment, CONP and CPt were freshly dilutedin sterile water.

2.2. Reagents

RPMI 1640 medium, Dulbecco’s Modified Eagle’s Medium(DMEM), penicillin–streptomycin solution, antibiotic–antimycoticsolution, L-glutamine and trypsin/ethylenediaminetetraacetic acid(EDTA) were purchased from PAA, Vienna, Austria. 4-oxo-2,2,6,6-tetramethylpiperidine-N-oxyl (TEMPONE), fetal bovine serum

(FBS), dimethyl sulfoxide (DMSO), sulforhodamine B (SRB) andMTT were obtained from Sigma–Aldrich Chemie GmbH,Germany. Annexin-V-FITC (AV)/Propidium Iodide (PI) kit waspurchased from Abcam, Cambridge, UK. Dihydrorhodamine(DHR) was obtained from Molecular Probes�, Invitrogen, CA, USA.Nitrate of Ce was from Aldrich, USA, and NaOH from Zorka, Serbia.

2.3. Synthesis of the CONP

Nanometric size ceria powder particles with fluorite-type struc-ture were obtained by applying self-propagating room tempera-ture method (SPRT) method using as starting materials nitrate ofCe and NaOH, according to Matovic et al. (2013) [23]. DuringSPRT synthesis hand mixing of nitrates with NaOH was performedin alumina mortar for �10 min until the mixture got light brown.After being exposed to air for 4 h, the mixture was suspended inwater. Rinsing out of NaNO3 was performed in centrifuge –Centurion 1020D, at 3500 rpm for 10 min. This procedure wasrepeated three times with distilled water as well as with ethanol.Obtained powders are dried at 70 �C [23].

The compositions of reacting mixtures were calculatedaccording to nominal composition of the final reaction product.Preparation of CeO2 powder was performed according to reaction:

2½CeðNO3Þ3 � 6H2O� þ 6NaOHþ ð1=2� dÞO2

! 2CeO2�d þ 6NaNO3 þ 15H2O ð1Þ

Obtained powder was characterized by X-ray diffraction and byRaman spectroscopy.

Crystal structure was identified by X-ray diffraction (XRD) usingfiltered Cu Ka radiation (Siemens D5000). XRD was also used toevaluate the crystallite size and lattice parameter. Before measure-ment the angular correction was done by high quality Si standard.Lattice parameters were refined from the data using the leastsquare procedure. Standard deviation was about 1%.

The Raman spectra were obtained using a U-1000 (Jobin-Yvon)double monochromator in back scattering geometry. The Ramanspectra were excited by the 514 nm line of an Ar+ ion laser andtaken at room temperature. In order to avoid sample heating weused a cylindrical focus and the laser power was kept lower than10 mW.

The specific surface area was measured by nitrogen adsorptionaccording to a Brunauer–Emmett–Teller (BET) method. Theadsorption characteristics of synthesized and annealed sampleswere determined. Adsorption and desorption isotherms of N2 weremeasured at �196 �C using the gravimetric McBain method. Thespecific surface area, SBET, pore size distribution, mesoporeincluding external surface area, Smeso, and micropore volume,Vmic, for the samples, were calculated from the isotherms. Pore sizedistribution was estimated by applying BJH method [24] to thedesorption branch of isotherms and mesopore surface and microp-ore volume were estimated using the high resolution as plotmethod [24–26]. Micropore surface, Smic, was calculated by sub-tracting Smeso from SBET. Also, the particle size, DBET, was calculatedfrom the specific surface area, S, using the following expression:DBET = 6/S�q, assuming isometric particles and using the bulkdensity, q, of CeO2 (7.2 g/cm3).

2.4. Cells and cell culture

Non-small cell lung carcinoma NCI-H460, human lung fetalfibroblast MRC-5 and colorectal adenocarcinoma cell lines DLD1and HT-29 were purchased from American Type CultureCollection (Rockville, MD, USA). NCI-H460/R multi-drug resistantcancer cells were selected from NCI-H460 cells by continuoustreatment with stepwise increasing concentrations of doxorubicin

20 30 40 50 60 70 80

Inte

nsity

(a.u

.)

2θ Degree

111

200 22

0

311

400 33

1Fig. 1. Room temperature X-ray diffractogram of CeO2�y sample. Atomic planes areindicated by Miller index (hkl).

6

M. Pešic et al. / Chemico-Biological Interactions 232 (2015) 85–93 87

(5–100 nM) for a period of 3 months [27]. DLD1-TxR multi-drugresistant cancer cells were selected from DLD1 cells after continu-ous exposure to stepwise increasing concentrations of paclitaxel(60–600 nM) for a period of 10 months [28]. Melanoma 518A2 cellline was kindly provided by Dr. Sanja Mijatovic (Institute forBiological Research ‘‘Sinisa Stankovic’’, Belgrade, Serbia), whileHaCaT cell line (normal human keratinocytes) was obtained fromCLS – Cell Lines Service, Eppelheim, Germany. NCI-H460, NCI-H460/R, DLD1 and DLD1-TxR cell lines were grown in RPMI 1640medium with 10% fetal bovine serum (FBS), L-glutamine, andantibiotic–antimycotic mixture, while 518A2, HT-29, MRC-5 andHaCaT cells were cultured in DMEM supplemented with 10% FBS,4 g/L glucose, L-glutamine (2 mM) and 5000 U/mL penicillin,5 mg/mL streptomycin solution. All cell lines were sub-culturedat 72 h intervals using 0.25% trypsin/EDTA and seeded into a freshmedium at the density of 8000 cells/cm2.

2.5. Sulforhodamine B assay

Cells were grown in 25 cm2 tissue flasks, harvested bytrypsinization, seeded into flat-bottomed 96-well tissue culture

300 400 500 600 700

Inte

nsity

(a.u

.)

Raman shift (cm-1)

603 cm-1

459 cm-1

Fig. 2. Room temperature Raman spectra of CeO2�y sample. The band at around459 cm�1 in the Raman spectrum is active triply degenerate F2g mode, and anadditional peak at 603 cm�1 represents the second order Raman mode centered atO2� vacancy position.

plates and incubated overnight. Investigated cell lines NCI-H460,NCI-H460/R, DLD1, DLD1-TxR, 518A2, HT-29, MRC-5 and HaCaTcells were seeded at 2000 cells/well. CONP treatment (50–500 lM) lasted 72 h. The cellular proteins were stained with sul-forhodamine B (SRB) assay, following slightly modified protocolof Skehan et al. (1990) [29]. Briefly, the cells in 96-well plates werefixed in 50% trichloroacetic acid (50 lL/well) for 1 h at 4 �C, rinsedin tap water and stained with 0.4% (w/v) sulforhodamine B in 1%acetic acid (50 lL/well) for 30 min at room temperature. The cellswere then rinsed three times in 1% acetic acid to remove theunbound stain. The protein-bound stain was extracted with200 lL 10 mM Tris base (pH 10.5) per well. The optical densitywas read at 540 nm, with correction at 670 nm (LKB 5060-006Micro Plate Reader, Vienna, Austria).

IC50 value was defined as concentration of each drug thatinhibited cell growth by 50%. IC50 was calculated by linear regres-sion analysis using Excel software.

2.6. Visualization of CONP in cell culture

518A2 cells were grown in 25 cm2 tissue flasks, harvested bytrypsinization, seeded into chamber slides and incubated over-night. After 72 h, live untreated cells and cells treated with100 lM CONP were observed under the light microscope (LeicaMicrosystems DMIL, Wetzlar, Germany). The experiments wererepeated three times and representative images are presented inFig. 7.

2.7. Cell death detection

The percentages of apoptotic, necrotic and viable cells weredetermined by Annexin-V-FITC (AV) and Propidium Iodide (PI)labeling. HT-29 cells were plated and incubated overnight in 6-well plates at density of 100,000 cells/well. Cells were subjectedto single treatment with 20 lM of CPt or with 100 and 200 lMof CONP. After 72 h, the attached and floating cells were collectedby centrifugation. The cells pellet was re-suspended in 100 lL ofbinding buffer containing 5 lL AV and 5 lL PI according to manu-facturer’s instructions. After the incubation period (5 min at roomtemperature), additional 400 lL of binding buffer was added and

0.0 0.2 0.4 0.6 0.8 1.00

1

2

3

4

5

n m

mol

/ g

P / P0

Fig. 3. Nitrogen adsorption isotherms, as the amount of N2 adsorbed as function ofrelative pressure for samples heat-treated at different temperatures. Solid symbols– adsorption, open symbols – desorption. Obtained isotherms are of type IV andwith a hysteresis loop which is assosiated with mesoporous materials.

Table 1Porous properties of the CeO2 powder samples. Specific surface areas calculated byBET equation, SBET, are shown. Smeso, Smic, Smic – mesopore including external surfacearea, micropore surface, micropore volume, respectively.

SBET (m2/g) Smeso (m2/g) Smic (m2/g) Vmic (cm3/g)

209 89 120 0.071

88 M. Pešic et al. / Chemico-Biological Interactions 232 (2015) 85–93

AV/PI staining was analyzed within 1 h by flow-cytometry. Thefluorescence intensity (green FL1-H and red FL2-H) was measuredon FACSClibur flow-cytometer (Becton Dickinson, Oxford, UnitedKingdom). In each sample, 10,000 cells were recorded (gated toexclude cell debris), and the percentages of viable (AV� PI�), earlyapoptotic (AV+ PI�), apoptotic and necrotic (AV+ PI+), and alreadydead (AV� PI+) cells were analyzed by CellQuest Pro data analysissoftware.

2.8. ROS/RNS detection

Flow-cytometric analysis of dihydrorhodamine (DHR) fluores-cence intensity was used to detect peroxynitrite anion and hydro-gen peroxyde anion level in cells. Peroxynitrite (ONOO�) is an RNS-unstable structural isomer of nitrate, NO3

�. Peroxynitrite is an oxi-dant and nitrating agent. Its formation in living systems is throughto the reaction of the free radical superoxide with the free radicalnitric oxide. 518A2, HaCaT and HT-29 cells were plated and incu-bated overnight in 6-well plates at 300,000 cells/well. Cells weretreated with 200 lM of CONP at 37 �C. After 24 h, adherent cellswere harvested by trypsinization and incubated in adequate med-ium with 1 lM DHR for 30 min at 37 �C in the dark. Cells were sub-sequently washed twice in PBS and DHR fluorescence was analyzedby flow-cytometry (green fluorescence at FL1-H channel). Meanfluorescence intensity (MFI) was calculated after correction forauto-fluorescence.

2.9. ESR spectroscopy

518A2, HaCaT and HT-29 cells were seeded into 6-well plates(300,000 cells/well), treated with CONP for 24 h, trypsinized, har-vested and washed twice in phosphate buffer saline (PBS).Aliquots (29 lL, containing 100,000 cells) of cell cultures(untreated and treated separately with 100, 200 lM CONP) were

0 2 4 6 8 100

10

20

30

40

50

60

ΔVp/ Δ

r p (cm

3 /g n

m)

rp (nm)

Fig. 4. Pore size distribution (PSD) for ceria powder sample depicts mesoporousstructure with most of the pores radius between 2 and 4 nm.

mixed with 1 lL of 0.75 mM TEMPONE (the final concentrationof TEMPONE was 25 lM). The ESR spectra of these samples wererecorded using a Bruker ELEXSYS II X-band spectrometer equippedwith SHQE cavity under the following conditions: microwavefrequency 9.84 GHz, microwave power 6.325 mW, field center3510 G, sweep width 75 G, modulation frequency 100 kHz,modulation amplitude 2.0 G, time constant 29.30 ms. TEMPONEreduction kinetics was analyzed for 20 min after its addition tothe cell cultures The experimental points were fitted using a

Fig. 5. Sensitivity of cancer and normal cells to CeO2 (CONP) treatment. Theinhibitory effect of CONP on cell growth was assessed by SRB assay after 72 htreatment. The absorbance levels obtained in HT-29, DLD1 and DLD1-TxR (A), MRC-5, NCI-H460 and NCI-H460/R (B), 518A2 and HaCaT cells (C) with different CONPconcentrations. Average ± S.D. values were calculated from three independentexperiments carried out in five replicates (n = 5). Statistical significance betweentreatments and untreated control is presented as p < 0.05 (⁄), p < 0.01 (⁄⁄) andp < 0.001 (⁄⁄⁄). The IC50 values for HT-29 and 518A2 were calculated by Excelsoftware.

A B

Fig. 6. Visualization of CONP in cell culture. Untreated melanoma 518A2 cells (A) and the same cells treated with 100 lM CONP (B). Live imaging of cells was generated byLeica Microsystems DMIL, Wetzlar, Germany. The arrows point to average sized CONP aggregates. Scale bar = 50 lm.

M. Pešic et al. / Chemico-Biological Interactions 232 (2015) 85–93 89

nonlinear Nelder–Mead fitting algorithm supplied with MATLAB6.5. The one-exponential model was used for fitting theTEMPONE reduction curves.

2.10. Statistics

All experiments were carried out independently three times(day to day variation). The number of replicates was three (flow-cytometric analysis and ESR spectroscopy) to five (cytotoxicitytesting). After confirmation of normality (Gaussian distribution),unpaired student t-test was applied in order to compare untreatedand treated samples. In the case of flow-cytometric analysis,10,000 cells per each sample were analyzed by CellQuest Pro dataanalysis software. Obtained MFI values were compared byunpaired student t-test. Showed results are representation of oneexperiment, while given percents are average of three independentexperiments.

3. Results and discussion

3.1. Synthesis and characterization of CONP

The X-ray diffraction pattern for the CONP powder is shown inFig. 1. XRD analysis reveals that all peaks for each sample were sig-nificantly broadened indicating small crystallite size and/or strain.It exhibits very diffuse diffraction lines, in such way that someatomic planes are impossible to indicate (hkl: 200, 400, 551,420). It was found that obtained nanomaterial has a crystallite size,calculated with Scherrer method [30], about 4 nm. Calculation ofthe average crystallite size (D) was performed on the basis of thefull width at half maximum intensity (FWHM) of the all reflectionsfor CeO2 by using Scherrer’s formula:

Dh k l ¼Kk

b cos hð2Þ

where k is the wavelengths of the X-rays, h is diffraction angle, b iscorrected half-width for instrumental broadening b = (bm � bs), bm

is observed half-width and bs is half-width of the standard CeO2

sample. Since CeO2 has a cubic structure (isometric), it means thatcrystallite shape is spherical.

Pure ceria has Raman active triply degenerate F2g mode at465 cm�1 due to symmetrical stretching of the Ce–O vibrationalunit in the eightfold coordination [31]. Besides the main band ataround 459 cm�1 in the Raman spectrum of pure ceria powderthere is an additional peak at 603 cm�1 (Fig. 2). This mode can beascribed to the second order Raman mode centered at O2� vacancyposition [32]. The appearance of this mode can be explained by the

lattice expansion in nanopowder ceria due to the increasing con-centration of oxygen vacancies and presence of the Ce3+ sites [33].

Nitrogen adsorption isotherms, as the amount of N2 adsorbed asa function of relative pressure at �196 �C, are shown in Fig. 3.According to the IUPAC classification [25] the isotherms are of typeIV and with a hysteresis loop which is associated with mesoporousmaterials. Specific surface areas calculated by BET equation, SBET,are listed in Table 1. It is obtained a relative high value SBET,(>200 (m2/g). Pore size distribution (PSD) of the ceria sampleshows that samples are mesoporous with most of the pores radiusbetween 2 and 4 nm (Fig. 4). The particle size calculated from thespecific surface area confirms that grain size is about 4 nm.

3.2. Anti-cancer activity of CONP

The growth inhibition effects of CONP were studied in six differ-ent human cancer cell lines (two of them were multi-drug resis-tant: NCI-H460/R and DLD1-TxR), and in two normal human celllines (lung fetal fibroblasts-MRC-5 and keratinocytes-HaCaT)(Fig. 5). The effects of CONP were assessed by SRB assay after 72 h.

Effect in HT-29 cells was examined in parallel with anothercolorectal adenocarcinoma cell line DLD1, and in its correspondingmulti-drug resistant variant DLD1-TxR (Fig. 5A). The IC50 valuefor CONP was attained at 183 lM in HT-29 cells, while in DLD1and DLD1-TxR cell lines IC50 was not reached in the screenedrange. CONP cytotoxicity in non-small cell lung carcinoma cells(NCI-H460 and NCI-H460/R) was compared next to normal fetalbronchial cells – MRC-5 (Fig. 5B). Although CONP was ineffectivein all three cell lines, it was selective towards cancer cells showinggrowth stimulating effect in normal cells (Fig. 5B). CONP exertedthe best effect in melanoma 518A2 cells with IC50 of 125 lM(Fig. 5C). It was not active in normal keratinocytes-HaCaT indicat-ing the greatest selectivity towards melanoma cells (Fig. 5C).

HT-29 and 518A2 cells showed the strongest sensitivity toCONP. In the tested range of concentrations, the IC50 was achievedonly in HT-29 and 518A2 cells. Other cell lines including normalcells were obviously several-fold less sensitive compared to colonand melanoma cells, respectively (Fig. 5).

The results imply that CONP was more efficient in some types ofcancer cells. Since the HT-29 and 518A2 cells were most sensitiveto the treatment, we have further examined the anti-cancer activ-ity of CONP in this cell line. CONP did not show the difference incytotoxicity between multi-drug resistant cancer cells – NCI-H460/R and DLD1-TxR and their sensitive counterparts – NCI-H460 and DLD1, respectively (Fig. 5A and B). Therefore, its efficacymight not depend on the presence of multi-drug resistantphenotype.

Fig. 7. Reduction of the TEMPONE radical in the HaCaT, 518 A2 and HT-29 cell lines,untreated and treated with 100 or 200 lM CONP. The reduction was followedduring 20 min since addition of TEMPONE to the cells. The data are obtained fromthree independent experiments.

Table 2The reduction rates of the TEMPONE free radical for the three cell lines, treated with100 and 200 lM CONP, and for control cells. Average ± S.D. values were calculatedfrom three independent experiments.

Cell line Treatment Rate constant k (s�1)

HACAT Control (6.3 ± 0.9) � 10�4

100 lM CeO2 (4.8 ± 0.7) � 10�4

200 lM CeO2 (3.3 ± 0.6) � 10�4

518A2 Control (6.9 ± 0.9) � 10�4

100 lM CeO2 (4.0 ± 0.8) � 10�4

200 lM CeO2 (4.5 ± 0.7) � 10�4

HT 29 Control (1.6 ± 0.2) � 10�3

100 lM CeO2 (1.5 ± 0.2) � 10�3

200 lM CeO2 (1.3 ± 0.1) � 10�3

90 M. Pešic et al. / Chemico-Biological Interactions 232 (2015) 85–93

The uptake, localization and effect of CONP in the cells dependon their size, surface charge and agglomeration inside the cells.Particles of diameter less than 20 nm are longer present in the cells

than larger particles [13,34]. We have noticed that the largestagglomerations of particles observed in the cell culture of the mostsensitive 518A2 cells were between 600 and 5000 nm (Fig. 6B). Thelocalization of these particles was mainly in or next to the nucleus.Other studies have shown that CONP can accumulate in variousparts of the cell, like cytoplasm and nucleus, but also in some othercell compartments.

Thus, pH is one of the important factors affecting direction ofCONP activity; neutral pH promoting cytoprotective effect andacidic pH leading to cytotoxic effect [13]. This might be a reasonof different influence of CONP on the cancer and normal cells.Although destroying the cancer cells and simultaneously protect-ing normal cells is a good basis for the application of the CONPin the cancer therapy [16], various types of cancer cells may differin intracellular pH and redox status. In sensitive cancer cells pH isacidic, promoting prooxidant effect of the nanoparticles, whilemulti-drug resistant cells resemble more alkaline pH as it is in nor-mal cells [35]. Thus, antioxidant effect of CONP could be present innormal and in multi-drug resistant cells. This could be partly thereason for different response to CONP that we have observed intested cancer and normal cell lines. The fact that non-small celllung carcinoma cells NCI-H460 and NCI-H460/R as well as colorec-tal carcinoma DLD1 and DLD1-TxR cells poorly react to the CONPtreatment might reflect their different redox status and protectiveactivity compared to melanoma 518A2 and colorectal HT-29 cells.

In order to examine possible effect of the CONP on the antioxi-dant processes in the cells, we measured reduction of the stablefree radical TEMPONE introduced into the cells of three cell linestreated with 100 and 200 lM CONP: HaCaT, 518A2 and HT-29.The rate of TEMPONE reduction in the observed time interval isdirectly proportional to the quantity of the free radicals producedfrom the cells, and/or to the activity of intracellular antioxidantmechanisms [36]. Thus TEMPONE reduction is also a reflection ofthe antioxidant capacity of the cells. The results were comparedwith control, untreated cells, in the same experiment (Fig. 7). Thereduction rates of the TEMPONE free radical for the three cell linesare shown in the Table 2. It is evident that the slope of decay of theTEMPONE radical differs in different cell lines (Fig. 7). Besides, thereduction rate is for one order of magnitude higher in the HT-29cells than in the other two cell lines (Table 2), showing much fasterradical metabolism and/or higher total radical content in this cellline. In all cell lines CONP induce decrease in free radical contentin the cells, evident from the decrease in the TEMPONE reductionrate. In the HT-29 cell line the CONP efficiency was considerablylower comparing to the other two lines, and similar for both nano-particle concentrations (Fig. 7 and Table 2). The damaging effect of100 lM CONP on free radical metabolism or antioxidant defensecapacity was more expressed in 518A2 cells (evidenced by thehigher TEMPONE decay rate) than in HaCaT cells, indicatingCONP-induced change in redox status in the sensitive line. This

Fig. 8. CONP induced increase of ROS levels in 518A2 and HT-29 cells. Flow-cytometric profiles of untreated and 200 lM CONP treated 518A2, HaCaT and HT-29 cellsevaluated after labeling with 1 lM DHR.

M. Pešic et al. / Chemico-Biological Interactions 232 (2015) 85–93 91

indicates that HaCaT cell line had more preserved free radicalmetabolism, and thus redox status, in comparison with the twoother cell lines. However, this effect is similar in these two celllines for 200 lM CONP treatment, since the 512A2 reacted simi-larly to both CONP concentrations and magnitude of HaCaTresponse was more pronounced after treatment with higher nano-particle concentration (Fig. 7 and Table 2).

In order to check if TEMPONE itself affected cell proliferation,we cultured HT-29 cell line in presence of 25 lM TEMPONE, underthe same growth conditions as for screening of CONP. The

TEMPONE had no inhibitory effect on the growth of HT-29 cells(results not shown).

To further study induction of free radicals in the CONP treatedcells, ROS/RNS production by 200 lM CONP was assessed by DHRstaining in melanoma 518A2 cells, normal keratinocytes HaCaTand colorectal HT-29 cells (Fig. 8). Intensity of DHR in 518A2 andHT-29 after CONP treatment significantly increased, while inHaCaT cells it was almost unchanged in comparison with untreatedcontrol (Fig. 8). These results imply that the rise of the ROS/RNSspecies predominantly detected by this method, namely

Fig. 9. Cell death induction by CONP in cancer cells. Analysis performed onuntreated and HT-29 cells treated with 20 lM CPt or with 100 and 200 lM CONP.The samples were analyzed for green fluorescence (Annexin-V-FITC, FL 1) and redfluorescence (Propidium Iodide, FL 2) by flow-cytometry. The assay distinguishesviable cells (AV� PI�), apoptotic cells (AV+ PI�), late apoptotic and necrotic cells(AV+ PI+) and secondary necrotic or dead cells (AV� PI+). (For interpretation of thereferences to color in this figure legend, the reader is referred to the web version ofthis article.)

92 M. Pešic et al. / Chemico-Biological Interactions 232 (2015) 85–93

peroxynitrite anion and hydrogen peroxyde anion, may decreaseantioxidant defense capacity of the cells and increase the level ofoxidative stress, which is also reflected in decreased rate ofTEMPONE reduction in the cells (Fig. 7). The preserved antioxidantcapacity may be the reason why normal HaCaT cells were lessvulnerable to CONP than 518A2 and HT-29 cancer cells.

Culcasi et al. (2012) [37] have shown, by DEPMPO spin trappingof hydroxyl radical in an EPR study, that pretreatment of thehuman dermal and murine 3T3 fibroblasts with the CONP dose-dependently triggered the release in the culture medium ofsuperoxide dismutase- and catalase-inhibitable DEPMPO/hydroxylradical adducts (DEPMPO–OH) and ascorbyl radical. They proposedthat both the mitochondrial stimulation and NADPH oxidase com-plex are involved in this response. The results provided substantialsupport for a dual mechanistic evidence of oxidative stressassociated with exposure to CONP, with an initial signaling atmicromolar doses regulated by thiol-containing compounds. InCONP-induced cytotoxicity at millimolar doses, the mitochondriaare a major source of ROS. The authors observed the strongdecrease of CAT, GPx and GSH levels followed by increase inhydrogen peroxide content after CONP exposure. Here weshowed, by decreased rate of the TEMPONE reduction and byincreased production of certain ROS/RNS species (peroxynitriteand hydrogen peroxyde anion), that whole free radical metabolismwas changed in the CONP treated 518A2 and HT-29 cancer cells.These CONP induced changes in redox intracellular status arepartly attributed to the prooxidant activity. In our study cells weretreated by micromolar concentrations of CONP, but CONP accumu-lation in the cells was evident (Fig. 6) indicating the increase intheir internal concentration. Thus, the higher vulnerability of518A2 and HT-29 cancer cells to CONP treatment may beexplained through cytotoxicity of increased ROS species, namely

peroxynitrite and hydrogen peroxyde anions, and inefficacy ofantioxidant enzymes. Further, ROS induced cell damages mighteventually lead to the cell death. A recent study has shown thatthe genotoxic effect of CeO2 nanoparticles can be modulated byinhibition of cellular hydrogen peroxide release [38].

To investigate the potential of CONP to induce cell death, HT-29cells treated with CONP or CPt were subjected to Annexin-V-FITC/Propidium Iodide staining. The results are summarized in Fig. 9.The percentage of apoptotic and necrotic cells after CONP treat-ment increased in comparison with untreated control. However,necrosis was the predominant way of cell dying. Results obtainedwith 100 and 200 lM of CONP did not considerably differ fromthe cell death induction observed with 20 lM CPt (Fig. 9). Ourexperiments indicated the potential for expanding studies to useCONP as additional treatment of melanoma and colorectal carci-noma, which would selectively eliminate cancer cells and leavehealthy cells intact. Considering CONP low toxicity towards normalcells, combination of CONP with some other cytostatic drug mightbe promising anticancer strategy. Another possibility is to useCONP as an adjuvant in radiotherapy considering that applicationof radiation therapy in some types of cancer (such as colon carci-noma) is complicate due to the specific location of disease and highprobability for damaging surrounding tissues.

4. Conclusions

Our results imply that CONP showed low cytotoxicity againstseveral human cancer (colorectal adenocarcinoma DLD1 andmulti-drug resistant DLD1-TxR, non-small cell lung carcinomaNCI-H460 and multi-drug resistant NCI-H460/R) and two normalcell lines (keratinocytes HaCaT, lung fetal fibroblasts MRC-5). TheIC50 was achieved in tested range of concentrations only in twoof eight employed cell lines, melanoma 518A2 and colorectal ade-nocarcinoma HT-29. We showed, by increased production of ROSand by decreased rate of TEMPONE reduction, that free radicalmetabolism/antioxidant capacity and thus redox status was chan-ged in these cells, indicating that ROS induced cell damages thatmight further lead to the cell death. The results also showed thatregarding total free radical metabolism/antioxidant capacity,512A2 and HT-29 cancer cell lines reacted similarly to bothCONP concentrations while magnitude of normal HaCaT cellsresponse was more pronounced when higher nanoparticle concen-tration was applied. Considerably low inhibitory potential of CONPin several cancer cell lines and in two normal cell lines, suggeststhat the usage of CONP in industry or medicine may be safe forhumans. The idea for selective killing of cancer cells by CONP inrespect to the difference between intracellular pH of cancer andnormal cells needs to be further clarified.

Conflict of Interest

The authors declare that there are no conflicts of interest.

Transparency Document

The Transparency document associated with this article can befound in the online version.

Acknowledgment

This work was supported by the grants III45012, III41031,III41005, III45005-4 from the Ministry of Education, Science andTechnology of the Republic of Serbia.

M. Pešic et al. / Chemico-Biological Interactions 232 (2015) 85–93 93

References

[1] F. Zhang, P. Wang, J. Koberstein, S. Khalid, S.W. Chan, Cerium oxidation state inceria nanoparticles studied with X-ray photoelectron spectroscopy andabsorption near edge spectroscopy, Surf. Sci. 563 (2004) 74–82.

[2] M. Fronzi, A. Soon, B. Delley, E. Traversa, C. Stampfl, Stability and morphologyof cerium oxide surfaces in an oxidizing environment: a first-principlesinvestigation, J. Chem. Phys. 131 (2009) 104701–104717.

[3] B.A. Rzigalinksi, K. Meehan, R.M. Davis, Y. Xu, W.C. Miles, C.A. Cohen, Radicalnanomedicine, Nanomedicine (Lond.) 1 (2008) 399–412.

[4] D. Schubert, R. Dargusch, J. Raitano, S. Chan, Cerium and yttrium oxidenanoparticles are neuroprotective, Biochem. Biophys. Res. Commun. 32 (2006)86–91.

[5] J.M. Dowding, S. Seal, W.T. Self, Cerium oxide nanoparticles accelerate thedecay of peroxynitrite (ONOO�), Drug Deliv. Transl. Res. 3 (2013) 375–379.

[6] I. Celardo, J.Z. Pedersen, E. Traversab, L. Ghibelli, Pharmacological potential ofcerium oxide nanoparticles, Nanoscale 3 (2011) 1411–1420.

[7] M.S. Wason, J. Zhao, Cerium oxide nanoparticles: potential applications forcancer and other diseases, Am. J. Transl. Res. 5 (2013) 126–131.

[8] C. Walkey, S. Das, S. Seal, J. Erlichman, K. Heckman, L. Ghibelli, E. Traversa, J.F.McGinnis, W.T. Self, Catalytic properties and biomedical applications ofcerium oxide nanoparticles, Environ. Sci.: Nano (2015), http://dx.doi.org/10.1039/c4en00138a.

[9] I. Ceraldo, M. De Nicola, C. Mandoli, J.Z. Pedersen, E. Traversa, L. Ghibelli, Ce3+

ions determine redox-dependent anti-apoptotic effect of cerium oxidenanoparticles, ACS Nano 5 (2011) 4537–4549.

[10] A. Clark, A. Zhu, K. Sun, H.R. Petty, Cerium oxide and platinum nanoparticlesprotect cells from oxidant-mediated apoptosis, J. Nanopart. Res. 13 (2011)5547–5555.

[11] E.J. Park, J. Choi, Y.K. Park, K. Park, Oxidative stress induced by cerium oxidenanoparticles in cultured BEAS-2B cells, Toxicology 245 (2008) 90–100.

[12] M. Hamrahi-Michak, S.A. Sadeghi, H. Haghighi, Y. Ghanbari-Kakavandi, S.A.Razavi-sheshdeh, M. Torkamani Noughabi, M. Negahdary, The toxicity effect ofcerium oxide nanoparticles on blood cells of male rat, Ann. Biol. Res. 3 (2012)2859–2866.

[13] A. Asati, S. Santra, C. Kaittanis, J.M. Perez, Surface-charge-dependent celllocalization and cytotoxicity of cerium oxide nanoparticles, ACS Nano 4 (2010)5321–5331.

[14] G. Waris, H. Ahsan, Reactive oxygen species: role in the development of cancerand various chronic conditions, J. Carcinog. 5 (2006) 14.

[15] S.F. Elswaifi, J.R. Palmieri, K.S. Hockey, B.A. Rzigalinski, Antioxidantnanoparticles for control of infectious disease, Infect. Disord. Drug Targets 9(2009) 445–452.

[16] L. Alili, M. Sack, A.S. Karakoti, S. Teuber, K. Puschmann, S.M. Hirst, C.M. Reilly,K. Zanger, W. Stahl, S. Das, S. Seal, P. Brenneisen, Combined cytotoxic andanti-invasive properties of redox active nanoparticles in tumor–stromainteractions, Biomaterials 32 (2011) 2918–2929.

[17] M.S. Wason, J. Colon, S. Das, S. Seal, J. Turkson, J. Zhao, C.H. Baker, Sensitizationof pancreatic cancer cells to radiation by cerium oxide nanoparticle-inducedROS production, Nanomedicine 9 (2013) 558–569.

[18] J. Colon, N. Hsieh, A. Ferguson, P. Kupelian, S. Seal, D.W. Jenkins, C.H. Baker,Cerium oxide nanoparticles protect gastrointestinal epithelium fromradiation-induced damage by reduction of reactive oxygen species andupregulation of superoxide dismutase 2, Nanomedicine 6 (2010) 698–705.

[19] R.A. Madero-Visbal, B.E. Alvarado, J.F. Colon, C.H. Baker, M.S. Wason, B. Isley, S.Seal, C.M. Lee, S. Das, R. Manon, Harnessing nanoparticles to improve toxicityafter head and neck radiation, Nanomedicine 8 (2012) 1223–1231.

[20] M. Fall, M. Guerbet, B. Park, F. Gouriou, F. Dionnet, J.P. Morin, Evaluation ofcerium oxide and cerium oxide based fuel additive safety on organotypiccultures of lung slices, Nanotoxicology 1 (2007) 227–234.

[21] F.R. Cassee, A. Campbell, J.F. Boere, S.G. McLean, R. Duffin, P. Krystek, I. Gosens,M.R. Miller, The biological effects of sub acute inhalation of diesel exhaustfollowing addition of cerium oxide nanoparticles in atherosclerosis-pronemice, Environ. Res. 115 (2012) 1–10.

[22] S. Aalapati, S. Ganapathy, S. Manapuram, G. Anumolu, B.M. Prakya, Toxicityand bio-accumulation of inhaled cerium oxide nanoparticles in CD1 mice,Nanotoxicology 8 (2014) 786–798.

[23] B. Matovic, J. Dukic, D. Bucevac, Z. Dohcevic-Mitrovic, M. Radovic, S. Boskovic,Synthesis, calcination and characterization of nanosized ceria powdersby self-propagating room temperature method, Ceram. Int. 39 (2013)5007–5012.

[24] E.P. Barret, L.G. Joyner, P.P. Halenda, The determination of pore volume andarea distributions in porous substances. I. Computations from nitrogenisotherms, J. Am. Chem. Soc. 73 (1951) 373–380.

[25] K. Kaneko, C. Ishii, M. Ruike, H. Kuwabara, Origin of superhigh surface area andmicrocrystalline graphitic structures of activated carbons, Carbon 30 (1992)1075–1088.

[26] K. Kruk, M. Jaroniec, K.P. Gadakaree, Nitrogen adsorption studies of novelsynthetic active carbons, J. Colloid Interface Sci. 192 (1997) 250–256.

[27] M. Pešic, J. Markovic, D. Jankovic, I. Markovic, Lj. Rakic, S. Kanazir, S. Ruzdijic,Induced resistance in the human non small cell lung carcinoma (NCI-H460)cell line in vitro by anticancer drugs, J. Chemother. 8 (2006) 66–73.

[28] A. Podolski-Renic, T. Andelkovic, J. Bankovic, N. Tanic, S. Ruzdijic, M. Pešic, Therole of paclitaxel in the development and treatment of multidrug resistantcancer cell lines, Biomed. Pharmacother. 65 (2011) 345–353.

[29] P. Skehan, R. Storeng, D. Scudiero, A. Monks, J. McMahon, D. Vistica, J.T.Warren, H. Bokesch, S. Kenney, M.R. Boyd, New colorimetric cytotoxicity assayfor anticancer-drug screening, Natl. Cancer Inst. 82 (1990) 1107–1112.

[30] P. Scherrer, Bestimmung der grösse und der inneren struktur vonkolloidteilchen mittels öntgenstrahlen, Gött. Nachr. 2 (1918) 98–100.

[31] W.H. Weber, K.C. Hass, J.R. McBride, Second-order scattering, lattice dynamics,and particle-size effects, Phys. Rev. B 48 (1993) 178–185.

[32] J.R. Mebride, K.C. Hass, B.D. Poindexter, W.H. Weber, Raman and X ray studiesof Ce1�xRExO2�y, where RE = La, Pr, Nd, Eu, Gd, and Tb, J. Appl. Phys. 76 (1994)2435–2439.

[33] A.E.C. Palmqvist, M. Wirde, U. Gelius, M. Muhammed, Surfaces of dopednanophase cerium oxide catalysts, Nanostruct. Mater 11 (1999) 995–1007.

[34] L.K. Limbach, Y. Li, R.N. Grass, T.J. Brunner, M.A. Hintermann, M. Muller, D.Gunther, W.J. Stark, Oxide nanoparticle uptake in human lung fibroblasts:effects of particle size, agglomeration, and diffusion at low concentrations,Environ. Sci. Technol. 39 (2005) 9370–9376.

[35] C. Daniel, C. Bell, C. Burton, S. Harguindey, S.J. Reshkin, C. Rauch, The role ofproton dynamics in the development and maintenance of multidrug resistancein cancer, Biochim. Biophys. Acta 832 (2013) 606–617.

[36] N. Kochengisky, H.M. Swartz, Nitroxide Spin Labels, Reactions in Biology andChemistry, CRC Press Inc, 1995.

[37] M. Culcasi, L. Benameur, A. Mercier, C. Lucchesi, H. Rahmouni, A. Asteian, G.Casano, A. Botta, H. Kovacic, S. Pietri, EPR spin trapping evaluation of ROSproduction in human fibroblasts exposed to cerium oxide nanoparticles:evidence for NADPH oxidase and mitochondrial stimulation, Chem. Biol.Interact. 199 (2012) 161–176.

[38] L. Benameur, M. Auffan, M. Cassien, W. Liu, M. Culcasi, H. Rahmouni, P. Stocker,V. Tassistro, J.-Y. Bottero, J. Rose, A. Botta, S. Pietri, DNA damage and oxidativestress induced by CeO2 nanoparticles in human dermal fibroblasts: evidence ofa clastogenic effect as a mechanism of genotoxicity, Nanotoxicology (2015),http://dx.doi.org/10.3109/17435390.2014.968889.