EPR spin trapping evaluation of ROS production in human fibroblasts exposed to cerium oxide nanoparticles: Evidence for NADPH oxidase and mitochondrial stimulation

Article abstract of DOI:10.1016/j.cbi.2012.08.007

To better understand the antioxidant (enzyme mimetic, free radical scavenger) versus oxidant and cytotoxic properties of the industrially used cerium oxide nanoparticles (nano-CeO₂), we investigated their effects on reactive oxygen species formation and changes in the antioxidant pool of human dermal and murine 3T3 fibroblasts at doses relevant to chronic inhalation or contact with skin. Electron paramagnetic resonance (EPR) spin trapping with the nitrone DEPMPO showed that pretreatment of the cells with the nanoparticles dose-dependently triggered the release in the culture medium of superoxide dismutase- and catalase-inhibitable DEPMPO/hydroxyl radical adducts (DEPMPO-OH) and ascorbyl radical, a marker of ascorbate depletion. This DEPMPO-OH formation occurred 2 to 24 h following removal of the particles from the medium and paralleled with an increase of cell lipid peroxidation. These effects of internalized nano-CeO₂ on spin adduct formation were then investigated at the cellular level by using specific NADPH oxidase inhibitors, transfection techniques and a mitochondria-targeted antioxidant. When micromolar doses of nano-CeO₂ were used, weak DEPMPO-OH levels but no loss of cell viability were observed, suggesting that cell signaling mechanisms through protein synthesis and membrane NADPH oxidase activation occurred. Incubation of the cells with higher millimolar doses provoked a 25-60-fold higher DEPMPO-OH formation together with a decrease in cell viability, early apoptosis induction and antioxidant depletion. These cytotoxic effects could be due to activation of both the mitochondrial source and Nox2 and Nox4 dependent NADPH oxidase complex. Regarding possible mechanisms of nano-CeO₂-induced free radical formation in cells, in vitro EPR and spectrophotometric studies suggest that, contrary to Fe²⁺ ions, the Ce³⁺ redox state at the surface of the particles is probably not an efficient catalyst of hydroxyl radical formation by a Fenton-like reaction in vivo.

Full text of DOI:10.1016/j.cbi.2012.08.007

Chemico-Biological Interactions 199 (2012) 161–176
Contents lists available at SciVerse ScienceDirect
Chemico-Biological Interactions
journal homepage: www.elsevier.com/locate/chembioint
EPR spin trapping evaluation of ROS production in human fibroblasts exposed
to cerium oxide nanoparticles: Evidence for NADPH oxidase and
mitochondrial stimulation
Marcel Culcasi a, Laila Benameur a,b, Anne Mercier a, Céline Lucchesi a, Hidayat Rahmouni a, Alice Asteian a,
Gilles Casano a, Alain Botta b, Hervé Kovacic c, Sylvia Pietri a,
⇑
a CNRS, Aix Marseille Université, Institut de Chimie Radicalaire UMR 7273, Equipe ‘‘Sondes Moléculaires en Biologie et Stress Oxydant’’, Marseille, France
b Aix Marseille Université, CNRS, EA 1784/FR 3098 ECCOREV, Laboratoire de Biogénotoxicologie et Mutagenèse Environnementale, Marseille, France
c Aix Marseille Université, INSERM UMR 911, CRO2, Marseille, France
a r t i c l e i n f o
a b s t r a c t
Article history:
To better understand the antioxidant (enzyme mimetic, free radical scavenger) versus oxidant and cyto-
toxic properties of the industrially used cerium oxide nanoparticles (nano-CeO2), we investigated their
effects on reactive oxygen species formation and changes in the antioxidant pool of human dermal
and murine 3T3 fibroblasts at doses relevant to chronic inhalation or contact with skin. Electron para-
magnetic resonance (EPR) spin trapping with the nitrone DEPMPO showed that pretreatment of the cells
with the nanoparticles dose-dependently triggered the release in the culture medium of superoxide dis-
mutase- and catalase-inhibitable DEPMPO/hydroxyl radical adducts (DEPMPO–OH) and ascorbyl radical,
a marker of ascorbate depletion. This DEPMPO–OH formation occurred 2 to 24 h following removal of the
particles from the medium and paralleled with an increase of cell lipid peroxidation. These effects of
internalized nano-CeO2 on spin adduct formation were then investigated at the cellular level by using
specific NADPH oxidase inhibitors, transfection techniques and a mitochondria-targeted antioxidant.
When micromolar doses of nano-CeO2 were used, weak DEPMPO–OH levels but no loss of cell viability
were observed, suggesting that cell signaling mechanisms through protein synthesis and membrane
NADPH oxidase activation occurred. Incubation of the cells with higher millimolar doses provoked a
25–60-fold higher DEPMPO–OH formation together with a decrease in cell viability, early apoptosis
induction and antioxidant depletion. These cytotoxic effects could be due to activation of both the mito-
chondrial source and Nox2 and Nox4 dependent NADPH oxidase complex. Regarding possible mecha-
nisms of nano-CeO2-induced free radical formation in cells, in vitro EPR and spectrophotometric
studies suggest that, contrary to Fe2+ ions, the Ce3+ redox state at the surface of the particles is probably
not an efficient catalyst of hydroxyl radical formation by a Fenton-like reaction in vivo.
Received 6 February 2012
Received in revised form 31 July 2012
Accepted 3 August 2012
Available online 22 August 2012
Keywords:
Free radicals
ESR spin trapping
Cerium oxide nanoparticles
Cell signaling
Fibroblasts
Cytotoxicity
Ó 2012 Elsevier Ireland Ltd. All rights reserved.
1. Introduction
Abbreviations: ACV, acetovanillone; AFR, ascorbyl free radical; CAT, catalase; CL,
chemiluminescence; DCIP, 2,6-dichlorophenolindophenol; DEPMPO, 5-(diethoxy-
phosphoryl)-5-methyl-1-pyrroline N-oxide; DMEM, Dulbecco’s modified Eagle’s
medium; DMSO, dimethylsulfoxide; DPI, diphenyleneiodonium; DTPA, diethylene-
triaminepentaacetic acid; EPR, electron paramagnetic resonance; EDTA, ethylene-
diaminetetraacetic acid; GPx, glutathione peroxidase; GSH, reduced glutathione;
GSSG, glutathione disulfide; KH, Krebs–Henseleit medium; LDH, lactate dehydro-
genase; MDA, malondialdehyde; MTT, 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl-
tetrazolium bromide; nano-CeO2, cerium oxide nanoparticles; PAO, phenylarsine
oxide; PBS, phosphate-buffered saline; ROS, reactive oxygen species; SOD, super-
oxide dismutase; TBA, 2-thiobarbituric acid; TEMPO, 2,2,6,6-tetramethyl-4-piperi-
done-N-oxyl radical.
Engineered materials containing cerium oxide (CeO2) nanopar-
ticles, nanoceria (nano-CeO2), are widely used in industry as diesel
fuel additives, catalysts, semiconductors, solar cells or oxygen sen-
sors [1] and also offer promising perspectives in medical diagnostic
or therapy [2]. Perhaps the most remarkable physicochemical
property of CeO2 is that its surface contains oxygen vacancies
releasing electrons which may reduce ceric ions (Ce4+) into cerous
ions (Ce3+), and in nano-CeO2 both the Ce4+/Ce3+ redox cycle and
oxygen exchange are dramatically increased due to its nanoscale
dimension [3].
⇑
Corresponding author. Address: Sondes Moléculaires en Biologie et Stress
Oxydant, CNRS UMR 7273, Institut de Chimie Radicalaire, Aix-Marseille Université,
Centre scientifique de Saint-Jérôme, Case 522, Avenue Escadrille Normandie
Niemen, F-13397 Marseille, France. Tel.: +33 (0) 491 288 579; fax: +33 (0) 491
288 758.
An important consequence of this unique redox chemistry is
that nano-CeO2 is a potential free radical scavenger and antioxi-
dant. For this main reason nano-CeO2 has been tested in a number
E-mail address: sylvia.pietri@univ-amu.fr (S. Pietri).
0009-2797/$ - see front matter Ó 2012 Elsevier Ireland Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.cbi.2012.08.007

162
M. Culcasi et al. / Chemico-Biological Interactions 199 (2012) 161–176
of biological systems where formation of reactive oxygen species
(ROS), including free radicals, is suspected to occur. Thus nano-
CeO2 has been found protective against cellular damage induced
by toxicants [4,5], radiation [6] or in pathological situations such
as cardiac [7] or brain [8] ischemia/reperfusion. As a mechanistic
explanation for these observations physicochemical assays, includ-
ing electron paramagnetic resonance (EPR), have proposed that
nano-CeO2 can alternatively behave like superoxide dismutase
(SOD) [9,10] or catalase [11].
Paradoxically many papers about the toxicity of nano-CeO2 in
relation with ROS production have been published. Hence when
nano-CeO2 was added to the culture medium of a variety of normal
cells [12–14] or administered to healthy animals [15–17] a number
of adverse effects were reported such as cytotoxicity, genotoxicity,
lipid and protein oxidative damage, antioxidant depletion, lifespan
reduction, systemic inflammation or cardiovascular alterations. Be-
cause cytotoxicity linked to oxidative stress was observed upon
in vivo administration of CeCl3 to rats [18,19] it was investigated
whether cerous ions could efficiently trigger ROS formation
in vitro. Based on the EPR spin trapping technique two studies
[19,20] proposed that Ce3+ can convert hydrogen peroxide (H2O2)
into the highly-reactive hydroxyl radical (HOÅ) by a Fenton-like
mechanism (Eq. (1)), thereby setting the conditions for oxidative
stress to occur:
ylsulfoxide (DMSO), acetovanillone (apocynin, ACV), phenylarsine
oxide (PAO), diphenyleneiodonium (DPI), 3-(4,5-dimethylthiazol-
2-yl)-2,5-diphenyltetrazolium bromide (MTT), Rhodamine 123,
Hoechst 33342, 7-amino-actinomycin (7-AAD+), 2,2,6,6-tetra-
methyl-4-piperidone-N-oxyl radical (TEMPO), catalase (CAT) and
xanthine oxidase from bovine liver, and SOD from bovine erythro-
cytes, were from Sigma–Aldrich Chimie (Saint Quentin Fallavier,
France). Annexin V-APC was from BD Biosciences, France. The spin
trap 5-(diethoxyphosphoryl)-5-methyl-1-pyrroline N-oxide (DEP-
MPO) was synthesized and purified according to [26,27]. For cell
culture, Dulbecco’s modified Eagle’s medium (DMEM), HEPES buf-
fer, phosphate-buffered saline (PBS), trypsin–EDTA, Ham’s F12
medium and fetal calf serum (FCS) were obtained from Life Tech-
nologies Inc. (Gaithersburg, MD, USA). Doubly distilled deionized
water was used throughout.
Analytical grade Ce salts and derivatives used in this study
(from Sigma–Aldrich) were Ce(III): CeCl3, Ce2(SO4)3 and Ce2(CO3)3,
and Ce(IV): Ce(SO4)2, (NH4)2Ce(NO3)6 and CeO2. Nano-CeO2, the
formula of which being CeO2(HNO3)0.5(H2O), were provided by
Rhodia Chemicals (France) and are ellipsoidal monocrystallites of
cerianite with a mean diameter of 7 nm and a specific surface area
of 400 m2/g [14]. Stock solutions of CeO2 particles were prepared
by vortexing for 10 min suspensions of the compound (1–3 mg/
ml) in either DMEM (cell incubations) or water (test-tube controls)
prior to use.
Ce3þ þ H2O2 þ Hþ ! Ce4þ þ HOÅ þ H2O
ð1Þ
1H, 13C and 31P NMR spectra were recorded on a Bruker AVL 300
spectrometer. Chemical shifts d are reported as ppm relative to
external 85% H3PO4 (31P) or tetramethylsilane as internal standard
(other nuclei). High resolution mass spectrometry–electron spray
ionization (HRMS–ESI) analysis was performed using a QStar Elite
instrument (AB Sciex, USA).
Besides this potential mechanism of nano-CeO2 toxicity, a
pH-dependent but H2O2-independent oxidase-like activity of the
particles has been recently demonstrated in vitro for a series of or-
ganic substrates or in low pH cell organelles [21] and tumoral
microacidic environment [22] upon nanoceria internalization. This
prooxidant activity of nano-CeO2 has been used to design targeted
anticancer agents [21,22]. Another possible oxidative mechanism
not involving the transient formation of HOÅ may involve the inter-
action of the Ce3+ redox state of CeO2 with H2O2 to convert
adsorbed organics into hydroperoxides [23].
2.2. Modified procedure for mitoPBN synthesis
[4-[4-[[(1,1-Dimethylethyl)-oxidoimino]methyl]phenoxy]butyl]
triphenylphosphonium bromide (mitoPBN; 3) was synthesized in
three steps using a modification (Scheme 1) of the procedure
reported in [28]. The first step involved synthesis of 4-(4-iodobut-
oxy) benzaldehyde (1) from 4-hydroxybenzaldehyde and 1,4-
diiodobutane in the presence of K2CO3 by adapting a protocol
In most of the mechanisms cited above, the Ce4+/Ce3+ redox cy-
cle will trigger oxidative stress only if cellular sources of H2O2 and/
or its biological precursor superoxide (O2Åꢀ) would become acti-
vated. In this regard adding nano-CeO2 to several biological sys-
tems activates oxidative stress genes and apoptosis [13] or
stimulates the release of inflammatory cytokines and chemokines
[17].
At the molecular level knowledge is still lacking concerning the
nature and cellular sources of the species produced in response to
nano-CeO2 interactions and the relationship between their poten-
tial ROS-generating capability, their impact on cell signaling and
the development of pathophysiological outcomes [24]. Using a
wide array of spectral and biochemical techniques, and a cell trans-
fection approach, we explored herein some chemical and biochem-
ical facets of cell/nano-CeO2 concentration-dependent interaction
and the possible cellular H2O2 generating sources [25] involved
(i.e., the mitochondrial respiratory chain and NADPH oxidase) in
two models of fibroblasts.
described for
x-dibromoalkanes [29]. Compound (1), which was
recovered with a high purity, was then converted into the triphen-
ylphosphonium iodide (2), which ultimately led to mitoPBN (3) by
reductive condensation with 2-methyl-2-nitropropane. These two
latter synthetic steps used the conditions described in [30]. The
experimental detail and the characterization of the intermediates,
including NMR and HRMS–ESI analysis, is given in the appendix.
2.3. Cell culture procedures and treatment with nano-CeO2
2.3.1. Experiments on normal human fibroblasts
Primary cultures of human fibroblasts were isolated by the out-
growth method using infant foreskins obtained after circumcision
[14]. The dermis was cut into small pieces of 0.5–1 mm3 under
sterile conditions. The tissue pieces were seeded in culture dishes
and incubated in DMEM supplemented with 10% FCS, 2 mM l-glu-
2. Materials and methods
tamine, 1 mM sodium pyruvate, 100 U/ml penicillin and 100 lg/ml
2.1. Chemicals, reagents and analytical techniques
streptomycin. Fibroblasts were cultured at 37 °C in a humidified
atmosphere of 5% CO2 in air in antibiotic-free DMEM and isolated
cells were generally obtained after 2 weeks where the culture
medium was changed every 2 days.
Solvents, starting materials and reagents, including diethylene-
triaminepentaacetic acid (DTPA), ethylenediaminetetraacetic acid
(EDTA), H2O2, 1,1,3,3-tetramethoxypropane (TMP), 2-thiobarbitu-
ric acid (TBA), 2,6-dichlorophenolindophenol (DCIP), 2-deoxy-d-ri-
bose (deoxyribose), 2(N-morpholino) ethanesulfonic acid (MES),
reduced glutathione (GSH), glutathione peroxidase (GPx), dimeth-
Isolated fibroblasts (105 cells/ml) were plated in 25 cm2 flasks
in a humidified atmosphere of 5% CO2 in air and exposed at 37 °C
for 2 or 24 h to freshly prepared dilutions of nano-CeO2 stock
solution (i.e., 6 ꢁ 10ꢀ5; 6 ꢁ 10ꢀ4 or 6 ꢁ 10ꢀ1 g/l, corresponding to

M. Culcasi et al. / Chemico-Biological Interactions 199 (2012) 161–176
163
Scheme 1. Modified synthesis of mitoPBN (3).
a range of 0.22
l
M–2.2 mM). The supernatants were timely col-
At the end of incubation, cells were resuspended in binding buffer
and analyzed by flow cytometry on a FACSCalibur flow cytometer
(BD Biosciences) according to [32].
lected, placed in cryotubes and either quickly frozen in liquid
nitrogen for delayed assessment of their ascorbyl free radical
(AFR)–DMSO content [31] or stored at ꢀ80 °C for later extracellular
lactate dehydrogenase (LDH) assay (see below). The remaining
adherent cells were trypsinized and PBS-resuspended (giving
ꢂ50,000 cells/ml) for lipid peroxidation and protein measurement
(see below).
To investigate the mechanisms involved in the short- (2 h) or
long- (24 h) term effects of nano-CeO2 on cells the same parame-
ters as described above were measured in additional fibroblasts
that were (i) first pretreated with the NADPH oxidase inhibitors
2.3.2. Experiments on transfected human fibroblasts
Plasmids for p47phox wild-type, p47-M1 (dominant negative
mutant) and p47-M8 (constitutively active mutant) were kindly
provided by Dr. L.S. Terada (University of Texas, USA). p47-M1
plasmid (W193R mutation) has been shown to lack binding func-
tion of the first p47 SH3 domain, a function critical to assembly
of the oxidase [33]. p47-M8 plasmid (S303, 304, 328D mutation)
mimics phosphorylation of three critical serines and appears to in-
duce unmasking of the tandem SH3 regions needed for oxidase
assembly and activation [33]. Human dermal fibroblasts were har-
vested in single cell suspension by treatment with trypsin–EDTA,
and subsequently transfected by amaxa nucleofector according to
the manufacturer’s protocol. Transfected cells were seeded in six
well plates at a density of 20,000 cells/cm2 and incubated for
24 h with 6 ꢁ 10ꢀ1 g/l nano-CeO2 in DMEM. Following incubation
the supernatants were collected and stored at ꢀ80 °C for later
LDH assay (see below). Transfection controls for p47-M1 and
p47-M8 plasmids were done using nucleofection without plasmid
(control WP cells) and transfection with p47phox wild-type. Con-
PAO (0.1 mM), DPI (5 lM) or ACV (0.3 mM) for 30 min, 1, or 4 h,
respectively, and, after removal of the incubation medium, reincu-
bated with 6 ꢁ 10ꢀ5–6 ꢁ 10ꢀ1 g/l nano-CeO2, or (ii) coincubated
for up to 24 h with nano-CeO2 and either DEPMPO (0.1 mM), CAT
(2 mU/ml) or a mixture of SOD (5 U/ml) and CAT (2 mU/ml), or
(iii) first exposed to 6 ꢁ 10ꢀ1 g/l nano-CeO2 for 23 h before coincu-
bating with 10
lM DPI for 1 h, or (iv) pretreated with mitochon-
dria-targeted mitoPBN (20
l
M) for 1 h followed by removal of
cell medium and incubation for 24 h with 6 ꢁ 10ꢀ1 g/l nano-CeO2
in DMEM, a concentration reported to affect the metabolic mito-
chondrial activity [14].
Other more specific indices of cell viability were determined on
fibroblasts coincubated up to 48 h with graded nano-CeO2 doses
trol WP cells incubated for 24 h with H2O2 (0.5 and 500
used as positive controls.
lM) were
(10ꢀ5ꢀ1 g/l, corresponding to concentrations ranging 0.036
lM–
3.36 mM), i.e., Trypan blue exclusion and MTT assays, the latter
being an indicator of mitochondrial dehydrogenase activity. The
protocol for the MTT assay was adapted from [22]: following incu-
bation of the cells with nano-CeO2, MTT (0.5 mg/ml) was added to
the medium, incubation was extended for 1 h and the conversion
of MTT to a purple formazan precipitate was monitored at 570 nm.
Nano-CeO2-induced mitochondrial toxicity was assessed by the
retention of Rhodamine 123, a membrane-permeable fluorescent
cationic dye. Cells were first incubated for 24 h with the particles
2.3.3. Preparation of samples of nano-CeO2 exposed human fibroblasts
for spin trapping analysis
Following exposure to nano-CeO2, the culture medium of each
flask was removed and stored for LDH assay. Cells were washed
three times with PBS and cells were reincubated for 15 min at
37 °C in 5 ml of DTPA (0.3 mM)- and DEPMPO (20 mM)-containing
PBS in order to reach 105 cells/ml. The supernatant was then trans-
fered into cryotubes and immediately frozen in liquid nitrogen
prior to EPR analysis. This procedure was applied to prevent any
physiologically unrelevant, nano-CeO2-catalyzed formation of
DEPMPO spin adducts (see results and discussion).
as described in the viability tests, then Rhodamine 123 (10
lM)
and Hoechst 333412 (10 g/ml) were successively added to the
l
medium and incubation was extended for 15 min. After being
washed twice with calcium- and magnesium-containing PBS, cells
were observed under an oil immersion 40ꢁ/1.4 plan apochromat
objective on a confocal Leica SP5 microscope (Leica Microsystems,
Nanterre, France).
The effect of nano-CeO2 on cell apoptosis was determined using
quantification of Annexin V and 7-AAD staining, a combination
allowing the differentiation among viable cells (Annexin V nega-
tive, 7-AAD negative), early apoptotic cells (Annexin V positive,
7-AAD negative) and late apoptotic or necrotic cells (Annexin V po-
sitive, 7-AAD positive). Following a 48-h exposure with nano-CeO2
(10ꢀ5ꢀ1 g/l) cells were harvested and stained for 20 min with An-
nexin V and 7-AAD in binding buffer containing 0.1 M HEPES (pH
7.4), 1.4 M NaCl and 25 mM CaCl2 at room temperature in the dark.
2.3.4. Preparation of whole cell lysates and western blotting
The adherent cells from flasks assayed by spin trapping as de-
scribed above were scraped and lysed in bromophenol blue-free
Laemmli buffer supplemented with complete protease inhibitor
cocktail (Roche Applied Science, France) and centrifuged for
10 min at 10,000g. Supernatant protein concentration was deter-
mined using the bicinchoninic acid assay-reducing agent compati-
ble (Pierce, France) following the manufacturer’s instructions.
Equal amounts of protein from lysates were resolved by 12%
SDS–PAGE and transferred to Hybond ECL nitrocellulose mem-
branes (Amersham Pharmacia Biotech, France). Membranes were
blocked in low-fat milk and incubated with primary antibodies
overnight at 4 °C. As primary antibodies, 0.5 lg/ml of mouse

164
M. Culcasi et al. / Chemico-Biological Interactions 199 (2012) 161–176
monoclonal anti-actin (Sigma–Aldrich, clone AC15), 1
l
g/ml of
the supernatant and 100% loss of viability. The percentage of LDH
released was calculated with respect to the total amount which
is the sum of the enzymatic activity in the lysate and in the culture
medium.
rabbit polyclonal anti-Nox4 (SantaCruz Biotechnology, sc-30141)
and a 1/500 dilution of rabbit polyclonal anti-p47-phox (Upstate
Biotechnology, 07–001) were used. Membranes were washed and
subsequently incubated for 1 h with an anti-mouse or anti-rabbit
HRP-conjugate secondary antibody. Proteins were quantified using
an enhanced chemiluminescence detection kit (Amersham Phar-
macia Biotech, France).
The determination of intracellular SOD, CAT and GPx activities
was performed at room temperature in samples prepared from
105 scraped 3T3 cells homogenized in 0.5 ml of appropriate cell ly-
sis buffers according to reported procedures [37]. Color develop-
ment was determined at 440, 540 and 340 nm for SOD, CAT and
GPx, respectively. Data are expressed in UI/106 cells according to
standard calibration curves obtained from commercial enzymes
and are representative of 6–12 independent experiments. Oxidized
carbonylated proteins were determined at 450 nm with the OxySe-
lectTM protein carbonyl ELISA detection kit (Cell Biolabs, Inc., USA)
using a 96-wells microplate reader and bovine serum albumin as
the standard. Data are expressed in nmol/mg cell lysate according
to standard calibration curves. The viability of 3T3 cells incubated
with nano-CeO2 in the range 6 ꢁ 10ꢀ5–6 ꢁ 10ꢀ1 g/l was measured
by the MTT assay as previously described [22].
2.3.5. Experiments on murine 3T3 fibroblasts
3T3 NIH fibroblasts (ATCC–LGC Promochem, Molsheim, France)
were maintained in culture at 37 °C in a 5% CO2-humidified atmo-
sphere in DMEM containing 1% glucose and supplemented with
10% FCS, 100 U/ml penicillin and 100 mg/ml streptomycin. Cells
were plated in 24-well dishes and medium was replenished every
2–3 days until confluency which was checked by microscopic
observation.
For cell exposure 10-ll aliquots of nano-CeO2 stock solution
were transferred into each culture well of confluent cells previ-
ously filled with phenol red-free DMEM containing 1% glucose, in
order to reach a final volume of 0.5 ml/well and final nano-CeO2
concentrations of 6 ꢁ 10ꢀ4 or 6 ꢁ 10ꢀ1 g/l. Cell cultures were then
incubated for 24 h at 37 °C in a 5% CO2-humidified atmosphere,
culture medium was sampled for extracellular LDH assay and
scraped cells were stored at ꢀ80 °C until antioxidant and lipid per-
oxidation assays. Additional inhibitions with PAO, SOD, CAT or
mitoPBN were performed as described above for human fibro-
blasts. In all these experiments the controls represented cells being
incubated in DMEM alone for the same period.
2.5. EPR spectroscopy
2.5.1. Cell cultures
In the culture medium of cells undergoing oxidative stress a
moderate DEPMPO concentration of 10–20 mM was considered
appropriate to allow EPR detection of spin adducts from ROS
[38]. In contrast at a concentration too low to give significant spin
adduct formation (i.e., 0.1 mM) the nitrone was rather expected to
play a protective role against antioxidant depletion [27]. Thus
thawed samples containing 20 mM DEPMPO were directly intro-
duced into a 10-mm quartz flat cell for EPR examination while that
containing 0.1 mM of the nitrone were first added to DMSO (1:1 v/
2.4. Biochemical assays
Nano-CeO2-treated scraped 3T3 fibroblasts were successively
thawed and homogenized (ꢂ105 cells/0.5 ml) on ice in cold MES
buffer (0.04 M; pH 6.5) containing 1 mM EDTA and centrifuged
for 15–20 min (10,000g) at 3 °C. The concentration of total gluta-
thione was measured in the supernatants by the enzymatic recy-
cling method involving 5,50-dithio-bis-2-(nitrobenzoate) which
reacts with the sulfhydryl group of GSH using a MP96 microplate
reader (SAFAS, Monaco) equipped with a 405–414 nm filter. Gluta-
thione disulfide (GSSG) was determined after derivatization of GSH
with 2-vinylpyridine [34]. The concentration of GSH was calculated
from the difference between total glutathione and GSSG. Data are
expressed in nmol/106 cells and are the means of at least 10 exper-
iments made in duplicate.
v) and then introduced into calibrated 50-ll glass capillary tubes
(Hirschmann Laborgeräte, Eberstadt, Germany) for AFR-DMSO
measurement.
The EPR spectra were recorded at room temperature with a Bru-
ker ESP 300 spectrometer (Karlsruhe, Germany) operating at
9.8 GHz with a 100-kHz modulation frequency and a microwave
power of 10 mW. Other parameters for DEPMPO spin trapping
(AFR–DMSO) were: modulation amplitude, 0.497 (0.883) G; recei-
ver gain, 2 ꢁ 105 (5 ꢁ 105); time constant, 40.96 (81.92) ms; field
resolution, 4096 (2048) points; sweep width, 140 (45) G; sweep
rate, 3.34 (1.07) G/s; number of averaged scans, 2 (4). Spectral
acquisition was initiated 1 min after complete thawing of the sam-
ple (DEPMPO spin trapping) or 2 min after addition of DMSO. EPR
signals were quantitated by double integration of the simulated
spectra using the Winsim program [39] and radical concentrations
were estimated by calibration with either TEMPO (for spin ad-
ducts) or freshly prepared AFR–DMSO (from sodium ascorbate)
in PBS. Data represent means of 6–12 independent experiments
for each tested experimental condition.
Ascorbate content in cell homogenates was determined by an
assay involving reduction of DCIP dye by ascorbic acid [35]. Data
are expressed in nmol/106 cells and are the means of at least 10
independent experiments made in triplicate.
Lipid peroxidation was assessed in human or 3T3 fibroblasts by
measuring MDA–TBA levels as described in [36]. One milliliter of
the cell homogenate supernatant, treated with 0.5 ml acetic acid
(20%) and 0.5 ml TBA (10 g/l), was incubated for 20 min at 95 °C,
cooled down at room temperature and assayed at 532 nm. A cali-
bration curve was obtained from standard MDA–TBA samples pre-
pared by using 0.25 ml aliquots of 0.1–10 mM TMP solutions.
Protein determination was performed in two randomly selected
flasks or wells [36]. Data are representative of 6–12 independent
experiments.
Release of cytoplasmic LDH was determined by assaying LDH
activity in the supernatant of human or 3T3 cells using a commer-
cial detection kit (Roche Diagnostic, Mannheim, Germany) and
data (in (UI/l/mg protein) are representative of 6–12 independent
experiments. To estimate the total LDH content a control measure-
ment was performed for each set of experiments, by treating cells
treated with 1% Triton X-100 to induce both total LDH release in
2.5.2. Test-tube controls
Experiments aimed at forming or inhibiting typical DEPMPO
spin adducts were carried out in 20 mM KH2PO4 buffer (pH 7) con-
taining 1 mM DTPA (termed as ‘KH–DTPA’ throughout). The initial
free radical generator was either the H2O2 (1 mM)–FeSO4 (1 mM)
Fenton system or a similar reagent where 1–10 mM Ce compound
was used instead of Fe2+. Both systems were allowed to react with
22–50 mM DEPMPO alone or in combination with 20 mU/ml SOD
or excess competitors such as DMSO or EtOH (15%, v/v). In another
set of experiments Ce salts or nano-CeO2 (1–10 mM) were com-
pared to Fe3+ (2 mM, as FeCl3) as catalysts of the nucleophilic syn-
thesis of the DEPMPO/OH adduct (DEPMPO–OH) from 50 mM
aqueous DEPMPO. EPR signals were recorded with a flat cell using
typical conditions as described above. Except the microwave

M. Culcasi et al. / Chemico-Biological Interactions 199 (2012) 161–176
165
power, all other settings were adapted depending on the sample
and are noted in each figure legend.
adduct which would have formed if HOÅ radicals had first attacked
the competitor [36]. The signal of Fig. 1A was dose-dependently
inhibited by addition of the chelator DTPA (0.5–5 mM) in 20 mM
KH2PO4 (final pH, 7.1; Fig. 1C). Consistent with a nucleophilic addi-
tion mechanism [45], mixing DEPMPO and nano-CeO2 in pure
methanol afforded DEPMPO–OMe (recorded as a mixture of trans
(cis) diastereoisomers: 76% (24%); aN = 13.03 (12.86) G; aP = 40.16
2.6. Estimation of the interaction of Ce compounds with superoxide
and hydrogen peroxide in vitro
Superoxide radicals were produced from xanthine (50 lM)-xan-
thine oxidase (10 mU/ml) in EDTA (0.1 mM)-supplemented 0.2 M
(46.91) G; aHb = 7.48 (7.29) G; aH = 1.73
G (not resolved))
c
(Fig. 1D). Since these data strongly suggest that DEPMPO–OH seen
in Fig. 1A can form by a HOÅ–unrelated mechanism involving ions
contained in nano-CeO2 that can be inactivated by chelation, we
first incubated aqueous DEPMPO (50 mM) with 10 mM of any of
the Ce(III) salts CeCl3, Ce2(CO3)3 or Ce2(SO4)3 for up to 30 min. Un-
der these conditions no EPR signal was observed as illustrated by
Fig. 1E for cerous carbonate. When the same experiment was per-
formed with 1–10 mM cerium sulfate, a Ce4+ salt, a strong signal
was detected (Fig. 1F), assigned to DEPMPOX (aN = 7.14 G;
aP = 38.69 G; aHb = 3.69 G (2H)), an oxidized form of DEPMPO
[46]. However Fig. 1G shows that using aqueous CeO2 suspensions
yielded weak DEPMPO–OH signals (about 80-fold lower than with
nano-CeO2) that were inhibited in the presence of a chelated phos-
phate buffer (Fig. 1H). This suggests that unlike free Ce3+/Ce4+ ions,
covalently bound Ce(IV) in CeO2 behaves as a weak catalyst in a
nucleophilic-type reaction with DEPMPO and that this property
is strongly enhanced in the nano-structured material.
The mechanism not related to true spin trapping by which DEP-
MPO–OH forms in the presence of CeO2 or nano-CeO2 in aqueous
medium remains to be elucidated. In this regard inverted spin trap-
ping, which primarily requires single electron oxidation of DEP-
MPO, appears unlikely due to the high oxidation potential of this
nitrone (E(DEPMPOÅ+/DEPMPO) = 2.24 V versus NHE in acetonitrile
[47]). On the other hand water is a very poor nucleophile and bind-
ing of a metal ion to the negatively charged oxygen of the nitronyl
function is required for the Forrester–Hepburn reaction to occur
substantially [42]. In line with Fig. 1C and H, the recent finding that
phosphate anions can link to cerium and modify the Ce3+/Ce4+ re-
dox cycle at the surface of nano-CeO2 [48] suggests an indirect
involvement of these DTPA-chelatable ions despite they do not
participate directly in the observed build up of DEPMPO–OH.
sodium phosphate buffer, pH 7.5 and their formation was moni-
tored for 10 min by lucigenin (5
l
M) chemiluminescence (CL) as
Åꢀ
described in [40]. This system was used to (i) evaluate the O2
scavenging effect of a series of Ce derivatives (0.005–1 mM) as
compared to SOD, and (ii) to check whether their pre-incubation
up to 24 h with SOD (5–20 mU/ml) at 37 °C in 0.2 M KH2PO4, pH
7.5 will modify enzymatic activity. In control experiments the po-
tential of tested Ce derivatives to inhibit xanthine oxidase was
evaluated spectrophotometrically by monitoring uric acid forma-
tion from xanthine.
The deoxyribose assay [41] was used as another method to as-
sess if HOÅ radicals are formed in the reaction of Ce derivatives with
H2O2. Briefly, reaction tubes contained 1 mM H2O2, 1.1 mM EDTA
and either 0.2 mM FeSO4 or various Ce compounds (0.2–1 mM)
in 20 mM KH2PO4 (final volume, 1 ml), pH 7.4. Control samples
contained 5–15 mM of the HOÅ radical scavenger mannitol or
KNO3, as a model for nitrate which enters the nano-CeO2 composi-
tion (see above). The mixtures were incubated at 37 °C for 1 h and
the MDA–TBA assay was carried out as described above. To check
for any Ce-chelating effect of the buffer, the experiments were re-
peated in HEPES and all measurements were compared against
blanks where water was used instead of H2O2 or Ce compound.
Data are the means of 6–10 experiments/concentration.
2.7. Statistical analysis
All values are expressed as the means SEM unless otherwise
noted. Statistical analysis was carried out using Student’s t test
or one-way ANOVA followed by a a posteriori Newman–Keuls or
Duncan’s tests. Intergroup differences were considered significant
at P < 0.05.
3. Results and discussion
3.2. Evidence against Ce3+– and nano-CeO2-catalyzed hydroxyl radical
formation from hydrogen peroxide by a Fenton-like reaction
3.1. Formation of DEPMPO hydroxyl radical adducts from aqueous
nano-CeO2 by a mechanism independent from free radical generation
A proposed explanation for the toxic properties of nano-CeO2
reported in some biological systems was that, as Fe2+ or Cu+ salts,
Ce3+ is capable of splitting H2O2 into harmful HOÅ species by a Fen-
ton reaction [19,20]. Upon mixing 8 mM H2O2, 22 mM DEPMPO
and 10 mM nano-CeO2 in KH–DTPA a strong EPR signal was imme-
diately observed, consisting of a mixture of DEPMPO–OH (63%) and
37% DEPMPO–OOH, the DEPMPO/O2Åꢀ spin adduct (Fig. 2A). In gen-
eral the best way to simulate DEPMPO–OOH EPR signals is to con-
sider the major trans diastereoisomer is subjected to a dynamic
chemical exchange between two rotamers [26,36]. However for
the simulation shown in Fig. 2A we obtained a good fit by assum-
ing the trans-DEPMPO–OOH signal is composed of two static com-
ponents with equal contribution, having the hyperfine splitting
constants: aN = 13.20 (13.12) G; aP = 50.76 (49.61) G; aHb = 11.60
In biological spin trapping four major pathways can lead to the
formation of HOÅ spin adducts of five-membered ring pyrroline N-
oxides such as DEPMPO, namely (1) direct scavenging of the radi-
Åꢀ
cal, (2) reduction of the O2 spin adduct by enzymes such as GPx,
(3) nucleophilic addition of water catalyzed by metal ions such as
Fe3+ (Forrester–Hepburn mechanism) [26,42], and (4) hydrolysis of
the cation radical resulting from single electron oxidation of the
nitrone (inverted spin trapping) [43], the two latter mechanisms
not requiring the presence of HOÅ radicals in the system.
To evaluate whether Ce compounds could act as catalysts for
nucleophilic addition DEPMPO (50 mM) was incubated with
10 mM nano-CeO2 in deionized water (final pH ꢂ3). A strong EPR
signal characteristic of DEPMPO–OH was readily detected
(Fig. 1A) which was simulated as a cis (trans) mixture of diastere-
oisomers (relative to the diethoxyphosphoryl group): 46% (54%)
with parameters aN = 14.05 (14.05) G; aP = 47.29 (47.29) G and
aHb = 14.24 (12.73) G, see structure in Fig. 1) [44]. When the aque-
ous DEPMPO and nano-CeO2 were mixed in the presence of excess
DMSO, DEPMPO–OH was still the only detected species (Fig. 1B)
instead of the expected 12-line signal of DEPMPO/methyl radical
(10.38) G; aH 1 = 1.01 (0.62) G and aH (6 H) = 0.39 (0.29) G. In
c
c2
agreement with a previous study [36] the weak additional signal
seen in Fig. 2A (see the low-field, non-composite doublet in the ex-
panded region) was assigned to cis-DEPMPO–OOH (aN = 13.35 G;
aP = 40.59 G; aHb = 9.95 G; aH = 1.35 G; ꢂ25% of the total DEP-
c
MPO–OOH signal) and we postulate it may be the non attributed
species reported in a recent EPR study describing the interaction
of H2O2 with CeCl3 [20].

166
M. Culcasi et al. / Chemico-Biological Interactions 199 (2012) 161–176
DEPMPO+nano-CeO /H O
2
2
A
(EtO)2(O)P
H3C
H
simulated
N
O
OH
DEPMPO-OH
+ 22% DMSO
B
C
+ 1 mM DTPA/Buffer
× 10
× 4
DEPMPO+nano-CeO /MeOH
2
D
(EtO)2(O)P
H3C
H
N
O
OCH3
simulated
DEPMPO-OMe
3+
DEPMPO+Ce /H O
2
E
F
× 10
DEPMPO+Ce(SO ) /H O
4 2
2
(EtO)2(O)P
H3C
O
N
O
simulated
DEPMPOX
DEPMPO+CeO /H O
2
2
G
× 10
× 10
H
+ 1 mM DTPA/Buffer
20 G
Fig. 1. EPR spectra following the addition of nano-CeO2 or typical Ce(III) or Ce(IV) salts (10 mM) to DEPMPO solutions. (A) From nano-CeO2 in water. (B) Same as A in the
presence of 22% DMSO. (C) Same as A in KH2PO4 (20 mM)-DTPA (1 mM). (D) From nano-CeO2 in pure methanol. (E) From Ce2(CO3)3 in water. (F) From Ce(SO4)2 in water. (G)
From CeO2 in water. (H) Same as G in KH2PO4 (20 mM)-DTPA (1 mM). DEPMPO concentration was 50 mM except in (B) and (F): 22 mM. The simulations correspond to the
following adducts and sets of main hyperfine splitting constants: (A): 46% cis- + 54% (trans)-DEPMPO–OH: aN = 14.05 (14.05) G; aP = 47.29 (47.29) G and aHb = 14.24 (12.73) G;
(D): 24% cis- + 76% (trans)-DEPMPO-OMe: aN = 12.86 (13.03) G; aP = 46.91 (40.16) G; aHb = 7.29 (7.48) G; (F): DEPMPOX: aN = 7.14 G; aP = 38.69 G; aHb = 3.69 G (2H). Single-
scan spectra were acquired at 9.79 GHz 1–3 min after mixing using parameters: resolution, 4096 points/scan; microwave power, 10 mW; modulation frequency, 100 kHz;
modulation amplitude, 0.44 G (except for (E): 0.79 G); time constant, 40.96 ms; gain, 2 ꢁ 104 (except for (D): 3.2 ꢁ 104 and (E): 5 ꢁ 104); sweep rate, 3.81 G/s except for (D):
1.91 G/s. Spectral intensities scalings are relative to trace A.
When 1–10 mM of the Ce(IV) compounds Ce(SO4)2 and CeO2
were used instead of nano-CeO2, a slightly weaker EPR signal ap-
peared immediately, with a superoxide versus hydroxyl radical ad-
duct proportion above 70% (Fig. 2C). However, using 10 mM of the
Ce(III) derivatives Ce2(SO4)3 or CeCl3 in the system, mixtures of
DEPMPO–OOH (>80%) with DEPMPO–OH only appeared when
incubation was prolonged for more than 25 min (Fig. 2D and E).
Furthermore, addition of 20 mU/ml SOD in all [H2O2-DEPMPO–Ce
compound] systems caused no change of the detected EPR signal
(Fig. 2F). Because of the strong chelating properties of KH–DTPA
used in these experiments, the results at this point suggest that
nano-CeO2 catalyzes DEPMPO adduct formation from H2O2 more
efficiently than its Ce(III) and Ce(IV) components, by a mechanism
not involving formation of O2 or nucleophilic synthesis of DEP-
MPO–OH. Thus the detection of DEPMPO–OOH in Ce(IV) experi-
ments signals here appears consistent with the earlier
Åꢀ

M. Culcasi et al. / Chemico-Biological Interactions 199 (2012) 161–176
167
Cerium compounds
Iron (II)
A
G
H
2.5 G
× 10
simulated
B
× 10
C
D
× 10
× 10
25 G
Hydrogen peroxide
E
F
I
× 10
× 10
× 10
Fig. 2. EPR spectra of DEPMPO radical adducts obtained during incubation of hydrogen peroxide alone or under Fenton conditions using Ce compounds (left part) or Iron(II)
(right part) as catalysts. Samples contained 22 mM DEPMPO and 1 mM DTPA in 20 mM phosphate buffer, pH 7.0 and were incubated at room temperature for 1–3 min, except
when indicated. The concentrations of Ce compounds and FeSO4 were 10 and 1 mM, respectively. (A) nano-CeO2 + 8 mM H2O2; (B) same as (A), in the presence of 15% DMSO;
(C) Ce(SO4)2 + 8 mM H2O2; (D) CeCl3 + 8 mM H2O2; (E) same as (D), after 1 h incubation; (F) same as (E) in the presence of 20 mU/ml SOD; (G) FeSO4 + 1 mM H2O2; (H) same as
(G), in the presence of 15% DMSO; (I) H2O2 (8 mM) alone, after 1 h incubation. Simulations were consistent with (G) DEPMPO–OH and (H) DEPMPO–Me in studies with
Iron(II), and with DEPMPO–OOH/DEPMPO–OH mixtures with superoxide adduct contribution of (A) 58%; (B) 63%; (C) 72%; (E) 82%; (F) 62%; (I) 77% in other studies.
Spectrometer settings were as in the legend of Fig. 1, except receiver gain, 4 ꢁ 104 (A–E) and 2 ꢁ 104 (F–H), and number of accumulated scans, 6 (E, F, I). The selected region
shows the low-field doublet of cis-DEPMPO–OOH. Spectral intensities scalings are relative to traces (G) and (H).
mechanism proposed by Czapski et al. [49] for the reduction of Ce4+
to Ce3+ by H2O2 in acidic medium, via intermediate formation of
Ce compounds did not significantly affect the control [DEPMPO–
OOH + DEPMPO–OH] mixture during the first 30 min following
incubation.
the perhydroxyl radical HOOÅ, the conjugate acid form of O2
Åꢀ
(Eq. (2) and (3)):
As these data, in particular those for Ce(III) compounds, are in
disagreement with an initial formation of HOÅ as proposed by
Heckert et al. for CeCl3 [20] (Eq. (1)), we used an alternative meth-
od by evaluating the extent to which Ce compounds can cause oxi-
dative damage to deoxyribose [41]. Table 1 lists the percentages of
deoxyribose (1 mM) degradation in phosphate buffer at pH 7.0 for
test compounds and combinations as compared to the value found
when the Fenton reagent [FeSO4 (0.2 mM) + H2O2 (1 mM)] was ap-
plied for 1 h at 37 °C. In this system, Fe2+-induced formation of
MDA–TBA dose-dependently decreased to the baseline level (i.e.,
in the absence of iron or H2O2) by addition of the HOÅ scavenger
mannitol in the range 0.2–15 mM, all absorbances being similar
when HEPES was used instead of KH2PO4 (not shown). Of the
tested Ce compounds used instead of Iron(II) in this assay only
nano-CeO2 (up to 1 mM) and, in a lesser extent, Ce(IV) derivatives
caused a weak, yet significant concentration-dependent damage to
deoxyribose but since this effect was not modified by 15 mM man-
Ce4þ þ H2O2 ! Ce3þ þ HOOÅ þ Hþ
ð2Þ
ð3Þ
Ce4þ þ HOOÅ ! Ce3þ þ O2 þ Hþ
To clarify if HOÅ is nevertheless formed in [nano-CeO2 + H2O2]
mixtures we checked for the presence of spin adducts from car-
bon-centered radicals by adding excess competitors such as etha-
nol or DMSO. When 15% DMSO was added to
a
[Fe2+
(1 mM) + H2O2 (1 mM)] Fenton reagent in the presence of 22 mM
DEPMPO the expected strong DEPMPO–OH signal of Fig. 2G was
immediately replaced by the signal of the methyl radical adduct
(DEPMPO-Me; aN = 15.14 G; aP = 47.66 G; aHb = 22.09 G; Fig. 2H)
[36]. However repeating the competition reaction with ethanol
or DMSO on 10 mM nano-CeO2 (Fig. 2B) or any of the other tested

168
M. Culcasi et al. / Chemico-Biological Interactions 199 (2012) 161–176
sure situations, i.e., within the range 10ꢀ5–10ꢀ1 g/l, which
corresponds to 0.036 M–0.36 mM [5,12,13,22,24]. Incubating
Table 1
Effect of hydrogen peroxide and various ions and Ce compounds in the degradation of
deoxyribose.
l
the cells at 37 °C for 24 h with the lowest doses of nano-CeO2
(i.e., 10ꢀ5–10ꢀ4 g/l) resulted in no significant cytosolic release of
LDH into the medium as compared to control cells incubated in
DMEM, which had a LDH activity of 10.7 1.9 U/l/mg protein. This
LDH release in cell supernatant did not exceed 2% of total cell LDH
activity (i.e., 689 12.4 U/l/mg protein). Cell monolayer staining
with Trypan blue showed that none of these two lowest doses in-
duced cell death after a 24-h incubation with the particles. Increas-
ing the nano-CeO2 dose resulted in a decrease of cell viability, as
vizualized in Fig. 3A for 10ꢀ1 g/l. The same dose-related loss of via-
bility was observed using the MTT assay (Fig. 3B). In contrast incu-
bating the cells for 24 h with nano-CeO2 induced no significant
mitochondrial toxicity even et the highest tested dose, as assessed
by Rhodamine 123 staining assay (Fig 3C). These results are com-
parable to that of Alili et al. [22] who recently reported no toxic
properties of nano-CeO2 up to 10ꢀ1 g/l in human dermal fibro-
blasts. Interestingly, application of the highest nano-CeO2 concen-
trations for 24 h resulted in a significant increase of total cellular
protein content, reaching 1.71 0.04 mg/ml and 1.53 0.06 mg/
ml for 6 ꢁ 10ꢀ1 g/l and 6 ꢁ 10ꢀ4 g/l, respectively as compared to
the normal level of 0.76 0.03 mg/ml in control cells (P < 0.05).
This protein biosynthesis, which mostly occurred at a non-toxic
concentration of nano-CeO2, paralleled with the increase of the to-
tal cell count.
Earlier studies have shown that low doses of CeCl3 stimulated
cell proliferation in cardiac fibroblasts [51] and the synthesis of
collagen in the rat heart in vivo [18], that could possibly be medi-
ated by an early formation of O2Åꢀ capable of promoting growth re-
sponses prior to any modifications in cell morphology and viability
[51]. On the other hand, pre-incubated nano-CeO2 has been shown
to remain internalized within the cytoplasm and vesicles of fibro-
blasts even after external particles were completely removed from
culture plates by washing [14]. From these considerations, it could
be of interest to explore any potential delayed ROS production pro-
moted by nano-CeO2 after removal of the medium.
In this context the washing protocol following nano-CeO2 expo-
sure, which was considered to comply with our required lack of
prolonged contact between the Ce particles and the spin trap,
could also give crucial information on the behavior of particles as
continuous intracellular ROS generators. Therefore it was adopted
in the next EPR studies where the culture medium of fibroblasts
incubated for 2 or 24 h with graded concentrations of nano-CeO2
was then carefully washed with PBS before a 20 mM DEPMPO solu-
tion in DTPA (0.3 mM)-supplemented PBS was introduced. DEP-
MPO-trappable species still released in the medium at this time
were then allowed to accumulate for 15 min at 37 °C in order to
improve EPR detection. A control sample obtained by mixing
6 ꢁ 10ꢀ4 g/l nano-CeO2 with 20 mM DEPMPO in DMEM for
15 min gave no detectable background EPR signal (Fig 4A).
Preliminary cytotoxicity studies showed no significant effect on
LDH activity when normal fibroblasts were incubated for 1 h with
either 20 or 40 mM DEPMPO (6.4 0.9 and 9.2 2.3 U/l/mg pro-
tein, respectively) compared to DMEM-treated cells (10.7 1.9 U/
l/mg protein, <2% of total LDH activity). Incubation of the cells
for 2 h in plain DMEM gave no EPR signal in the supernatant
(Fig. 4B) whereas strong DEPMPO–OH signals were detected when
6 ꢁ 10ꢀ5 g/l nano-CeO2 was applied during the incubation phase
(Fig. 4C). The mean DEPMPO–OH concentration significantly in-
creased when either the incubation time or nano-CeO2 concentra-
tion increased, e.g., it was almost 4-fold higher when shifting from
6 ꢁ 10ꢀ5 g/l to 6 ꢁ 10ꢀ4 g/l nano-CeO2 and 24 h incubation
(Fig. 4F). For both nano-CeO2 concentrations and incubation times,
adding 2 mU/ml CAT or [2 mU/ml CAT + 5 U/ml SOD] during incu-
bation provoked a significant partial or total inhibition of the DEP-
Added compound
% of Deoxyribose degradation
+H2O2 (1 mM) no H2O2
Fe(II) as FeSO4 (0.2 mM)
100a
0.36 0.01
0.23 0.01
0.19 0.01
2.04 0.10
no Fe2+
0.79 0.01
0.90 0.01
1.66 0.01
1.91 0.01
4.09 0.36*
4.14 0.28*
0.47 0.01
1.13 0.01
1.43 0.16
2.98 0.45
Fe(II) as FeSO4 (0.2 mM) + mannitol
Ce(IV) as Ce(SO4)2 (0.2 mM)
Ce(IV) as Ce(SO4)2 (0.2 mM) + mannitol
Ce(IV) as Ce(SO4)2 (1 mM)
Ce(IV) as Ce(SO4)2 (1 mM) + mannitol
Ce(IV) as CeO2 (0.2 mM)
Ce(IV) as CeO2 (1 mM)
Ce(IV) as (NH4)2Ce(NO3)6 (0.2 mM)
Ce(IV) as (NH4)2Ce(NO3)6 (1 mM)
Ce(IV) as (NH4)2Ce(NO3)6 (1 mM) + mannitol 2.57 0.56
Ce(III) as CeCl3 (0.2 mM)
Ce(III) as CeCl3 (1 mM)
Ce(III) as CeCl3 (1 mM) + mannitol
nano-CeO2 (0.2 mM)
nano-CeO2 (0.2 mM) + mannitol
nano-CeO2 (1 mM)
5.46 0.41*
5.54 0.61*
0.74 0.01
1.54 0.01
1.62 0.01
5.35 0.33*
5.52 0.16*
11.05 0.36*
10.46 0.25*
0.67 0.01
0.55 0.01
2.54 0.11
2.46 0.47
7.83 0.54*
6.72 0.61*
12.86 0.72*
12.75 0.81*
nano-CeO2 (1 mM) + mannitol
ꢀ
NO3ꢀ as KNO3 (0.2 mM)
NO3 as KNO3 (1 mM)
a
Data are calculated relative to the value for [FeSO4 + H2O2] set at 100% and are
expressed as means SD (n = 3). One-way ANOVA followed by Duncan t test.
P < 0.05 versus iron-free buffer and [FeSO4 + H2O2]. The incubation medium
*
contained 1 mM deoxyribose in EDTA (1.1 mM)-supplemented phosphate buffer
(20 mM), pH 7.4. The concentration of the HOÅ scavenger mannitol was 15 mM.
nitol HOÅ radicals are probably not involved. As a confirmation
omitting H2O2 in the samples had no significant effect on MDA–
TBA release suggesting deoxyribose degradation may result from
direct oxidation by ceric ions, which are known to oxidize sugars
such as glucose [50]. We also checked that this oxidation was
ꢀ
not induced by NO3 (a constituent of nano-CeO2) when intro-
duced instead of Fe or Ce in the sample (Table 1).
If the two above in vitro approaches suggested that the direct
interaction between Ce3+ and H2O2 does not rapidly form HOÅ,
the mechanism by which the mixtures of DEPMPO adducts seen
in Fig. 2A–F and reported in [20] are formed with a delay is still un-
known. In an attempt to gain insights we found that omitting CeCl3
in the mixture that yielded the spectrum of Fig. 2E afforded a very
similar background signal after 1 h incubation (Fig. 2I), where still
DEPMPO–OOH was the dominant species (77%) over DEPMPO–OH.
Although H2O2 is not prone to significant daylight homolysis it is
possible that low levels of HOÅ progressively form, which can then
react with excess hydrogen peroxide to form HOOÅ, both radical
species being trapped by DEPMPO.
3.3. In human fibroblasts micromolar concentrations of nano-CeO2
induced ROS formation at non-toxic levels via NADPH oxidase
activation
The cell effects of nano-CeO2 in relation to ROS formation
[12,13,16] or inhibition [4,5] have been studied by many investiga-
tors previously and significant changes in a variety of biochemical
indices of oxidant stress were found to occur over relatively long
periods (>24 h). Since the results of Fig. 1C evidenced weak but
detectable artifactual DEPMPO–OH formation when the nitrone
was in contact with nano-CeO2 in buffers, we considered that the
best spin trapping strategy to evaluate ROS release (especially
H2O2) from cells exposed to the particles would be to avoid pro-
longed coincubation with DEPMPO.
We first investigated the effects of nano-CeO2 on human fibro-
blasts at several concentrations relevant to common chronic expo-

M. Culcasi et al. / Chemico-Biological Interactions 199 (2012) 161–176
169
Fig. 3. Effect of nano-CeO2 on viability and apoptosis of human fibroblasts. Cells were exposed to the particles (10ꢀ5ꢀ1 g/l, corresponding to the range 0.036
l
Mꢀ3,66 mM)
for 24 h (AꢀD) or 48 h (E) before processing and analysis. (A) Trypan blue staining. (B) MTT assay. (C) RH-123 fluorescence. (D) Effect of NADPH inhibitors using MTT assay.
Cells were pretreated with the NADPH oxidase inhibitors PAO (0.1 mM) or ACV (0.3 mM) for 30 min or 4 h, respectively, before being exposed to 6 ꢁ 10ꢀ1 g/l nano-CeO2 in
fresh buffer. DPI was added 1 h before (5 lM) or at the 23rd hour (10 lM; DPI after nano) of nano-CeO2 exposure (E) Annexin V and 7-AAD staining. Symbols represent: h,
viable cells (Annexin V- and 7-AADꢀ); P, early apoptotic cells (Annexin V + and 7-AADꢀ), and j, late apoptotic cells (Annexin V + and 7-AAD+). Data represent means SEM
(n = 3–6 independent experiments performed in triplicate). One-way ANOVA followed by Newman–Keuls test: ⁄P < 0.01 versus respective control in DMEM and §P < 0.05 for
DPI and PAO pretreated cells versus cells exposed to 6 ꢁ 10ꢀ1 g/l nano CeO2.
MPO–OH signal, respectively (Figs. 4D and E). This observation
strongly suggests that cells loaded with nano-CeO2 are
continuously activated to release H2O2 and O2Åꢀ. The intra- versus
extracellular site of their interaction with the spin trap yet depends
on several factors such as cell penetration of DEPMPO [52] and the
relatively low stability of DEPMPO–OOH towards cell reductants or
transition metals [26,52,53]. The partial inhibition by SOD of the
EPR signal in Fig. 4D could be explained considering that even
when exogenously added, the enzyme can scavenge intracellularly
formed O2 [53]. Of interest, formation of DEPMPO–OH adducts
consecutive to extracellular release of H2O2 has been documented
previously in murine fibroblasts in situation of oxidant stress [38].
Åꢀ

170
M. Culcasi et al. / Chemico-Biological Interactions 199 (2012) 161–176
Fig. 4. DEPMPO–OH formation following exposure of human fibroblasts to nano-CeO2 and effect of inhibitors. After removal of Ce particles from the culture medium cells
were reincubated in DEPMPO (20 mM)- and DTPA (0.3 mM)-containing PBS for 15 min at 37 °C. EPR spectra were obtained in the supernatant by signal averaging 2 scans
using the spectrometer settings given in the legend of Fig. 1, except for modulation amplitude, 0.497 G; receiver gain, 2 ꢁ 105 and sweep rate, 3.34 G/s. (A) EPR signal
following a 15-min contact between 6 ꢁ 10ꢀ4 g/l nano-CeO2 and 20 mM DEPMPO in DMEM. Other typical EPR signals were recorded after incubation for 2 h with; (B) plain
DMEM; (C) 6 ꢁ 10ꢀ5 g/l nano-CeO2; (D) same as; (C) + 2 mU/ml catalase; (E) same as; (D) + 5 U/ml SOD. (F) Variation of DEPMPO–OH levels with exposure time, nano-CeO2
concentration and inhibitors. As described in Fig. 3, concentrations of PAO and ACV were 0.1 mM and 0.3 mM, respectively and data represent means SEM (n = 6–12). One-
way ANOVA followed by Newman–Keuls test: ⁄P < 0.05 versus nano-CeO2 with same concentration and incubation time; §P < 0.05 versus 2 h nano-CeO2.
Of the mechanisms that could lead to the CAT-inhibitable DEP-
MPO–OH signal of Fig. 4E a Fenton reaction involving mobilized
transition metals such as iron, but not Ce3+ (see above) is likely to
occur instead of nucleophilic synthesis because of the presence of
DTPA in the buffer. It has also been shown that human fibroblasts
can produce large amounts of H2O2/O2Åꢀ over many hours by stim-
ulation of NADPH oxidase located in the outer membrane [54].
Pertinent to this hypothesis, the two NADPH oxidase inhibitors,
ACV (0.3 mM) and PAO (0.1 mM), decreased the EPR signal by 60
and 85%, respectively (Fig. 4F). In these experimental conditions
PAO and ACV did not significantly affect the cell viability in the
MTT assay when preincubated before nano-CeO2 exposure at
6 ꢁ 10ꢀ4 g/l. Thusviabilitiesfor these ACV- or PAO-treated cells were
90.3 1.5% and 97.0 1.2%, respectively as compared to 91.3 1.8%
innano-CeO2 treatedcontrolcells. Sincethese inhibitions of ROSpro-
duction were not complete, as expected from the optimal low toxic
concentrations used [54,55], NADPH oxidase may not be the unique
cellular source of ROS being activated by nano-CeO2.
To understand the apparent discrepancy between the lack of
global cytotoxicity of the low doses of nano-CeO2 (up to 10ꢀ1 g/l)
and the fact that they yet trigger ROS generating systems such as
membrane NADPH oxidase, we looked for earlier indices of oxidant
stress. Thus experiments were repeated on human fibroblasts in
order to assess the time course of nano-CeO2-induced ascorbate re-
lease or lipid peroxidation, and the effects of inhibitors of O2Åꢀ and/
or H2O2, or NADPH oxidase. In biological fluids AFR–DMSO EPR sig-
nal (giving a doublet with aH = 1.81 G) is a reliable marker of ascor-
bate concentration [31]. Fig. 5A shows that ascorbate was mostly
released in the cell supernatant at 2 h incubation, at a concentra-
tion dependent on the nano-CeO2 dose. Again, the key role of
H2O2 and NADPH oxidase activation in this early cell damage
was supported by the observed inhibition by 2 mU/ml CAT and
0.1 mM PAO. The EPR signal was also significantly decreased by
coincubation with 0.1 mM DEPMPO, likely by a mechanism involv-
ing direct scavenging of free radicals at the nitronyl site [38]. Par-
allel to this ascorbate leakage fibroblasts exposed to Ce particles
showed increased lipid peroxidation, with a peak of membrane
MDA–TBA formation at 24 h incubation of 6 ꢁ 10ꢀ4 g/l nano-
CeO2, and a similar inhibition profile (Fig. 5B).
Together the results presented in this section tend to demon-
strate that micromolar concentrations of nano-CeO2 can partially
Åꢀ
activate NADPH oxidase to release ROS such as H2O2/O2 (with a
preferential role of Ce(IV) redox state). Rather than sustaining sig-
nificant lipid peroxidation and cell death, this activation has the
characteristics of cell signaling, leading to protein synthesis and
cell proliferation. These particular properties may explain why
low nano-CeO2 has been found either detrimental in cardiac stud-
ies [18,51] or neuroprotective in in vitro [4] and in vivo [8] exper-
iments by increasing cell regeneration and survival. This general
role in ROS-mediated signaling is shared by Nox1–5 enzymes
which specifically regulate physiological processes including cell
growth, apoptosis or differentiation [56]. Thus the next step of
our study was to focus on the specific role of Nox2 and Nox4 iso-
forms because (i) they are localized in endosomes, or in the nu-
cleus and the endoplasmic reticulum, respectively [56], all
subcellular sites that can be reached by nano-CeO2 agglomerates
as they are uptaken in the cytoplasm [5,13,14,21], and (ii) human
fibroblasts mainly express these two proteins [57].
3.4. Simultaneous stimulation of NADPH oxidase and mitochondrial
sources of ROS are involved in the toxicity of millimolar concentrations
of nano-CeO2 in human fibroblasts
To assess the role of Nox2, Nox4 and identify other ROS sources
sustaining the development of toxicity in human fibroblasts we
incubated the cells for 24–48 h with higher (above 10ꢀ1 g/l) nano-

M. Culcasi et al. / Chemico-Biological Interactions 199 (2012) 161–176
171
Fig. 5. Levels of AFR–DMSO and MDA–TBA following short- and long-term exposure of human fibroblasts to non toxic concentrations of nano-CeO2, and effect of inhibitors.
(A) EPR signals of ascorbyl free radical (AFR) were obtained at 9.8 GHz from cell supernatant:DMSO (1:1) mixtures by signal averaging 4 scans with instrument parameters:
microwave power, 10 mW; modulation frequency, 100 kHz; modulation amplitude, 0.883 G; time constant, 81.92 ms; gain, 5 ꢁ 105 and sweep rate, 1.07 G/s. (B) Lipid
peroxidation was measured in homogenates from untreated (DMEM) or nano-CeO2-treated cells. Concentrations of inhibitors were as in the legend of Fig. 3, and DEPMPO was
0.1 mM. Data represent means SEM (n = 6–12). One-way ANOVA followed by Newman–Keuls test: ⁄P < 0.05 versus nano-CeO2 with same concentration and incubation
time;§P < 0.05 versus 2 h nano-CeO2.
CeO2 concentrations. Cell damage occurred when 6 ꢁ 10ꢀ1 g/l or
above nano-CeO2 was used, as significant LDH was released extra-
cellularly (LDH activity was 137.08 5.1 U/l/mg protein; P < 0.05
versus control), corresponding to 22.1 2.6% of control total cell
LDH. This cell necrosis was concomitant to a ꢂ30–50% loss of viabil-
ity, as demonstrated by the MTT assay (Fig. 3B). This finding is in
line with the reported 45% impairment of the metabolic mitochon-
drial dehydrogenase activity of human fibroblasts exposed to
6 ꢁ 10ꢀ1 g/l nano-CeO2 for 24 h using the WST-1 assay [14]. In
our study despite there was significant cell death (Fig. 3B) and
detachment (not shown) at these high nano-CeO2 concentrations,
staining with either Rhodamine 123 or Hoechst 33342 did not show
any evidence of significant compromized mitochondrial function or
sign of chromatin condensation, respectively (data not shown).
Pharmacological NADPH oxidase inhibition by pretreatment with
0.1 mM PAO partially reduced the cellular toxicity of 6 ꢁ 10ꢀ1 g/l
nano-CeO2, while ACV was not efficient to protect cells (Fig. 3D).
This protective effect of PAO could be related to the finding that a
similar dose of PAO provided 90% inhibition of H2O2 production in
human fibroblasts after only 10 min incubation [54]. In our study
the failure of ACV to protect cells might reflect a more toxic and/
or pro-oxidant activity of the inhibitor, a feature already reported
in non phagocytic cells [58]. On another hand, preincubation of
DPI, a general flavoprotein inhibitor that almost completely blocks
NADPH oxidase activity as compared to PAO or ACV [59], again par-
tially preserved the viability of cells exposed to high nano-CeO2
doses (Fig. 3D) showing that NADPH oxidase is likely one but not
the sole of the ROS source involved in cell injury. This effect was
even less marked when DPI was added only at the 23rd hour of
nano-CeO2 treatment.
Taken together, the results suggests that in the presence of
nano-CeO2, mitochondria may also function as a source of signal-

172
M. Culcasi et al. / Chemico-Biological Interactions 199 (2012) 161–176
ing molecules (such as H2O2/O2Åꢀ) that initiate early mechanisms
such as cell proliferation and apoptosis, leading to mitochondrial
dysfunction as described for several other toxic metals [60 and ref-
erence therein]. As a confirmation, we found a significant amount
of early apoptotic cells upon incubation for 48 h with 6 ꢁ 10ꢀ1 or
1 g/l nano-CeO2 (Fig. 3E).
Using the spin trapping protocol of Fig. 4 the mean DEPMPO–
OH levels recovered in the supernatant of normal cells exposed
to cytotoxic nano-CeO2 were 25–60 times more elevated than
when non toxic concentrations were applied for the same period
(Fig. 6A) showing an increased ROS production and suggesting an-
other source than NADPH oxidase has been activated. Indeed
NADPH oxidases and mitochondrial respiratory chain represent
two main sources of H2O2/O2Åꢀ production in cells. Thus we inves-
tigated whether this second source of ROS was implicated by pre-
Next, we compared the mitochondrial versus NADPH oxidase
sources of ROS that become activated by 6 ꢁ 10ꢀ1 g/l nano-CeO2
by measuring the effect on the DEPMPO–OH signal of DPI. Com-
pared to nano-CeO2-treated (24 h) fibroblasts, a 1-h pretreatment
or addition to the particles for the last incubation hour with
DPI decreased DEPMPO–OH levels by 80 and 50%, respectively
(Fig. 6A). This result implies that overstimulation of NADPH oxi-
dase also participates in the observed ROS-induced cytotoxicity
of the high nano-CeO2 concentration.
It was reported that DPI can also inhibit mitochondrial ROS pro-
duction [59], that is NADPH oxidase inhibition may not account for
the effect on DEPMPO–OH seen in Fig. 6A. Therefore to gain insight
into the involvement of NADPH oxidase in the cytotoxicity of nano-
CeO2 we used a molecular approach with transfected cells. Since
Nox2-dependent ROS production relies on p47phox phosphoryla-
tion, a function critical to assembly and activation of the oxidase
[56,57], the effect of the overexpression of wild-type p47phox,
dominant negative mutant of p47phox (p47-M1) and constitutive
active mutant of p47phox (p47-M8) on nano-CeO2 (6 ꢁ 10ꢀ1 g/l)-
induced DEPMPO–OH formation was first evaluated. It has been
shown that P47phox needs to be phosphorylated on different ser-
ine residues to bind p22phox and activate NADPH oxidase activity
[33]. Also, the P47-M1 mutant inhibits the binding of p47phox to
p22phox and the P47-M8 mutant mimics serine phosphorylation
treating the cells for 1 h with 20 lM mitoPBN, a lipophilic,
mitochondria-targeted spin trap reported to efficiently protect
Åꢀ
cells from the consequences of O2 formation [28]. Whereas no
toxicity was found for mitoPBN at 50
fibroblasts (with a mean LDH release of 12.7 2.4 UI/ mg protein,
n = 6) a lower concentration of 20 M significantly decreased DEP-
lM in normal untreated
l
MPO–OH formation by ca. 60% in treated cells (Fig. 6A). These re-
sults unambiguously demonstrate the mitochondrial source of
ROS is activated by cytotoxic doses of nano-CeO2.
Fig. 6. Involvement of NADPH oxidase and mitochondria in ROS production from human fibroblasts ROS exposed to 6 ꢁ 10ꢀ1 g/l nano-CeO2 for 24 h. (A) DEPMPO–OH levels
following nano-CeO2 exposure in human fibroblasts determined by the cell washing, DEPMPO (20 mM) spin trapping and EPR recording protocols described in materials and
methods. Preceding the spin trapping phase, the added inhibitors in control cell medium were the NADPH oxidase inhibitor DPI (1 h before (5
lM) or at the 23rd hour (10 lM)
of nano-CeO2 exposure) or the specific mitochondrial radical inhibitor mitoPBN (20 M; 1 h before nano-CeO2 exposure). (B) DEPMPO–OH levels in nano-CeO2-treated
l
fibroblasts transfected either with no plasmid (control WP), p47phox wild-type, dominant negative or constitutive active mutant of p47phox (p47-M1 and p47-M8
respectively). Inset represents p47phox protein expression and their respective loading control (actin) in untreated transfected cells. (C) Lactate dehydrogenase (LDH) release,
expressed as percentage of LDH released into the cell medium with respect to total LDH, determined after 24-h exposure with or without 6 ꢁ 10ꢀ1 g/l nano-CeO2 in
transfected human fibroblasts as described in (B). Hydrogen peroxide after 24-h exposure at 500 lM in the medium was used as a positive control. (D) Nox4 and actin
expression in untreated and nano-CeO2 exposed human normal fibroblasts. The right panel represents the densitometric analysis of nano-CeO2 on Nox4 expression relative to
actin (from 2 independent experiments). Data on ROS production represent means SEM (n = 6–12). One-way ANOVA followed by Newman–Keuls test: ⁄P < 0.05 versus
nano-CeO2 treatment in corresponding control cells. Data on LDH release represent means SEM (n = 4–6 independent experiments performed in triplicate). One-way
ANOVA followed by Newman–Keuls test: àP < 0.05 versus p47-M8 nano-CeO2 treated cells.

M. Culcasi et al. / Chemico-Biological Interactions 199 (2012) 161–176
173
Table 2
Cell viabilitya and oxidative stress-related parameters in murine 3T3 cells following exposure to nano-CeO2 for 24 h and effect of specific inhibitors.
Inhibitor
Viability
SOD
CAT
GPx
Protein
MDA–TBA
GSH
Ascorbate
activity
activity
activity
carbonyls
(nmol/mg)
(%)
(UI/106 cells)
(mUI/106 cells)
(nmol/10 mg prot)
8.3 1.8
(nmol/106 cells)
20.8 0.01
none (DMEM)
100
8.7 0.9
36.5 2.6
29.5 2.0
2.98 0.75
5.35 0.9
[nano-CeO2] = 6 ꢁ 10ꢀ4 g/l
none
99.1 0.1
10.1 2.8
10.8 1.1
11.8 2.3
12.7 1.8
48.6 3.1*
41.9 2.2
37.7 1.2§
38.8 1.9
37.9 2.1*
28.6 3.1
28.1 1.1§
29.2 1.0
3.09 0.11
3.02 0.09
3.45 0.11
2.78 0.07
14.1 2.7*
10.1 1.9
6.9 1.4§
9.6 2.2
33.7 1.4*
16.4 1.6§
19.9 1.1§
20.6 2.7§
3.93 0.1*
6.24 0.7
6.90 0.2§
6.78 0.4§
CAT
100.1 0.2
99.9 0.3
96.1 0.4
[CAT + SOD]
PAO
[nano-CeO2] = 6 ꢁ 10ꢀ1 g/l
none
64.1 2.4*
4.5 0.5*
5.9 0.8
7.7 1.1§
6.8 0.9
8.9 1.1§
18.5 2.4*
22.9 1.9*
27.8 0.9§
23.9 1.2*
30.9 1.9§
16.8 1.7*
18.9 1.1*
25.2 1.1
18.0 2.1*
25.9 1.7
4.99 0.67*
3.90 0.27
2.81 0.09§
3.99 0.24
2.70 0.91§
27.7 1.5*
14.9 1.2*
10.3 1.2§
19.8 1.7*
9.2 2.9§
9.8 1.1*
2.4 0.9*
2.9 0.8*
4.8 1.1§
3.9 0.9*
5.1 0.8§
CAT
86.1 0.9§
91.1 1.9§
76.1 1.1
89.1 2.1§
13.9 2.0*
17.6 0.9*§
13.9 1.0*
20.3 0.9§
[CAT + SOD]
PAO
mitoPBN
a
Measured by the MTT assay. Data are expressed as means SEM (n = 8–14 independent experiments in duplicate/group). The concentrations of added inhibitors were:
CAT, 2 mU/ml; SOD, 5 U/ml; PAO, 0.1 mM and mitoPBN, 20
l
M. One-way ANOVA followed by Duncan t test.
*
P < 0.05 versus cells incubated in DMEM.
§
P < 0.05 versus nano-CeO2-treated cells at the corresponding dose.
necessary to activate NADPH oxidase [33]. In line with these data
Fig. 6B shows that control transfected cells (control cells without
plasmid, WP) and wild-type p47phox transfected cell in absence
of nano-CeO2 do not display any significant EPR signal. In nano-
CeO2 treated cells a same extent of ROS production was observed
in control WP transfected cells and in cells over-expressing
wild-type p47phox. ROS production in p47-M1 transfected cells
was decreased by 50%, compared to control transfected cells and
wild-type p47phox cells, while it significantly increased with
p47-M8, the constitutive active mutant of p47phox (+40% versus
wild-type p47phox, P < 0.05). Together these data suggest that
nano-CeO2 exposure induces ROS formation in part through Nox2
activation by a mechanism involving p47phox phosphorylation
and translocation.
end (10
lM) of nano-CeO2 exposure, respectively, while mitoPBN
(20 M) had no effect (1.70 versus 1.69 mg/ml in nano-CeO2-over-
l
expressing WT-p47phox cells).
Together the results presented in this section suggest that (i)
Nox2 activation might participate to nano-CeO2 stimulation of pro-
tein synthesis and that this activation is ROS-mediated and (ii) a
large amount of ROS are produced by the mitochondria, a phenom-
enon which is likely in relation with the increase in cell mortality,
decrease in mitochondrial dehydrogenase activity and early apop-
tosis development.
3.5. Impact of nano-CeO2 exposure on the antioxidant and oxidative
stress status in murine 3T3 fibroblasts (Table 2)
In parallel, LDH activity was monitored in supernatants from
transfected cells showing that wild-type p47phox expression does
not increase cell necrosis in absence of nano-CeO2 (Fig 6C) as com-
pared to normal fibroblasts. As a positive control, a significant LDH
In a further series of experiments we investigated if the nano-
CeO2-induced ROS formation evidenced in the above EPR studies
(Figs. 4 and 6) would result in changes in cell viability and an array
of biomarkers of oxidative stress. Being a more abundant cell
source than primary cultures of human cells, we used murine
3T3 fibroblasts which were obtained at 105 cells/ml and incubated
with 6 ꢁ 10ꢀ4–6 ꢁ 10ꢀ1 g/l nano-CeO2 for 24 h. As compared to
untreated fibroblasts cell viability (evaluated by the mitochondrial
dehydrogenase activity assay) at the end of incubation was not al-
release required WP control cells to be treated by 500
lM (Fig 6C),
but not 0.5 M H2O2 for 24 h, in line with the reported properties
l
of the oxidant as a necrotic inductor when used in the millimolar
range [61]. In the presence of nano-CeO2 the expression of the con-
stitutive active form of p47phox (p47-M8) lead to a significant con-
comitant increase in ROS production (Fig 6B) and LDH release (Fig
6C) compared to wild-type p47phox expression. In contrast, the
expression of the dominant negative form of p47phox (p47-M1)
led to a significant decrease of nano-CeO2 induced ROS production
but no significant decrease of LDH activity. These results suggest
that the p47-phox dependent NADPH oxidase is not the only
source mediating nano-CeO2 toxicity in human fibroblasts. A pos-
sible involvement of Nox4 might also be partly suggested as
nano-CeO2 treatment induced an increase in Nox4 expression
(Fig. 6D). In contrast to Nox2, Nox4 does not possess any identified
regulatory subunit and an increase in Nox4-dependent ROS pro-
duction relies exclusively on Nox4 protein induction [56]. To date,
there is no particular clue on the pathway leading to the induction
of Nox4 in presence of nano-CeO2.
Finally, we evaluated the impact of Nox2 activation on nano-
CeO2-induced protein synthesis. After 24 h incubation with
6 ꢁ 10ꢀ1 g/l nano-CeO2 the total protein content in WT-p47phox
overexpressing cells increased by 2.2 0.4 times compared to un-
treated cells (i.e., 0.59 mg/ml) while it decreased by 1.7 0.3 and
raised by 3.8 0.4 times (n = 3) for the p47-M1 and p47-M8 mu-
tants, respectively. Treatment with DPI decreased protein synthe-
tered when the applied dose of nano-CeO2 ranged 6 ꢁ 10ꢀ4
–
6 ꢁ 10ꢀ2 g/l, while it was significantly impaired by 30% with
6 ꢁ 10ꢀ1 g/l of the particles, a pattern paralleled by measurements
of extracellular LDH release (data not shown). In those cells sus-
taining a normal viability in the presence of 6 ꢁ 10ꢀ4 g/l nano-
CeO2, ROS formation was nevertheless demonstrated by (i) the sig-
nificant variations of the levels of ascorbate (decreasing by 25%)
and GSH (increasing by 150%), two antioxidants that function to-
gether in the cell, and (ii) a slight but significant increase of lipid
peroxidation. However these biochemical events are not detrimen-
tal enough as to induce a significant production of protein carbon-
yls in cell lysates, an end-point of protein oxidation which
irreversibly accumulate within cells [62]. From the increase of
specific enzyme activities such CAT and GPx, but not SOD, or the
effects of added antioxidant enzymes or inhibitors, the participa-
tion of the NADPH oxidase/H2O2 system in cell damage can be pos-
tulated. Since GSSG levels fell under the detection limit when cells
were treated by 6 ꢁ 10ꢀ4 g/l nano-CeO2, the elevation of GSH levels
reflects mobilizations from GSSG (given that one mole of GSSG will
yield two moles of GSH) or the synthesis and/or transport of GSH
into the cells. Together with cell viability, all the above oxidative
stress-related biomarkers were markedly depressed when the
sis by 15 and 30% when administered before (5 lM) or at the

174
M. Culcasi et al. / Chemico-Biological Interactions 199 (2012) 161–176
nano-CeO2 dose was increased up to 6 ꢁ 10ꢀ1 g/l and, if PAO only
provided a partial protection, only the mixture of SOD and CAT,
or mitoPBN could efficiently restore all these indices. In parallel
lipid peroxydation and protein carbonyls raised by 3.3 and 1.7
times, respectively.
Table 3
Inhibition of superoxide-induced lucigenin chemiluminescence by Ce compounds.
Added compound
% of luminescence
None
100
Nano-CeO2 (5 M)
Nano-CeO2 (10 M)
Nano-CeO2 (30 M)
Nano-CeO2 (50 M)
CeO2 (5 M)
Ce(SO4)2 (5 M)
CeCl3 (5 M)
SOD
SOD + nano-CeO2 (10 M)
SOD + nano-CeO2 (30 M)
Preincubated [SOD + nano-CeO2 (30 M)]
Preincubated SOD
68.8 2.5
56.4 1.3
27.3 1.1
10.6 1.0
91.3 1.5
54.9 3.5
25.6 1.9
15.7 2.1
10.4 1.4*
22.4 1.9*
32.6 2.9*,§
16.9 1.9
The present data in 3T3 cells provide substantial support for a
dual mechanistic evidence of oxidative stress associated with
exposure to nano-CeO2, with an initial signaling at micromolar
doses regulated by thiol-containing compounds. Similar observa-
tions have been reported in in vivo studies with Ce compounds
[15,63]. In our study the observed inhibitory properties of mitoPBN
and the lack of a clear protective effect of PAO confirm that mito-
chondria is a major source of ROS in nano-CeO2-induced cytotoxic-
ity at millimolar doses. Although PBN and its derivatives are able to
quench O2Åꢀ, this reaction is not fast enough to compete with dis-
mutation (either catalyzed or spontaneous) and therefore the pro-
tective effect of mitoPBN in the present study may arise from
inhibition of secondary oxidation reactions. Nevertheless, this im-
plies that Ce particles could reach the mitochondrial membrane, a
property that was not clearly established in previous transmission
electron microscopy studies of different cell types [5,14,21]. We
also showed that, from the signaling phase, all systems involved
in H2O2 control (i.e., CAT, GPx and GSH) play a primordial role in
cell response to nano-CeO2 exposure. In this regard the cytotoxicity
of high levels of H2O2 in cells, which provokes an irreversible oxi-
dation of thiol-containing compounds [25,56,60], can explain the
strong decrease of CAT, GPx and GSH levels observed here. The cru-
cial cooperation between GSH and the mitochondrial enzyme glu-
tathione peroxidase in the control of H2O2 concentration has been
described in details [60].
Data are calculated relative to the value for (lucigenin + O2Åꢀ) measured 5 min after
addition of xanthine oxidase and are expressed as means SD (n = 6). The incuba-
tion medium contained 5 lM lucigenin, 50 lM xanthine in EDTA (0.1 mM)-sup-
plemented 0.2 M sodium phosphate buffer, pH 7.5. The reaction was initiated by
addition of xanthine oxidase (10 mU/ml). SOD concentration was 20 mU/ml
throughout and preincubation time before addition of xanthine oxidase was 6 h at
37 °C. One-way ANOVA followed by Duncan t test.
*
P < 0.05 versus SOD.
P < 0.05 versus preincubated SOD.
§
ble 2) may partially reflect a nonradical mechanism between the
particles and the enzyme. While a mixture of SOD (20 mU/ml)
and nano-CeO2 (10
compared to each component alone, this inhibitory effect dropped
markedly when nano-CeO2 was present above 30 M. This decline
lM) acted synergistically on lucigenin CL as
l
in lucigenin CL inhibition was even more pronounced when the
particles and the enzyme were coincubated for 6 h at 37 °C before
O2Åꢀ generation was initiated, a result not due to an intrinsic loss of
SOD activity during incubation.
It is generally believed that HOÅ plays an important role in H2O2-
induced cell injury leading to the oxidation of membrane lipids and
proteins and that the molecular mechanism requires the presence
of catalytic metal ions. From our discussion above and literature
data [23] HOOÅ is the most likely intermediate formed when
nano-CeO2 interferes with H2O2 within the cell (Eq. (1)). Then
HOÅ-induced toxicity, if any, will not proceed via Ce3+ catalysis
but probably by ‘conventional’ Fenton chemistry, e.g., with much
more catalytically active iron or copper ions released upon mem-
brane disruption. Furthermore a nonradical, H2O2-independent
oxidase-like activity of Ce compounds, including nanoparticles
has been described for many biologically-occurring organic sub-
stances [19,21,23,50].
By showing the prominent role of Ce3+ versus Ce4+ constituents
Åꢀ
in the O2 scavenging effect of nano-CeO2 in vitro, our data are
consistent with other studies on nanoceria having variable Ce3+
/
Ce4+ surface ratios [9,10]. Previous investigators [9] have proposed
that this active redox [Ce4+/Ce3+] couple in nano-CeO2 possess a po-
tent SOD-like activity as described by Eqs. (4) and (5):
Ce3þ þ OÅ2ꢀ þ 2Hþ ! H2O2 þ Ce4þ
ð4Þ
Ce4þ þ O2Åꢀ ! O2 þ Ce3þ
Besides its direct reaction with O2 (Eq. (5)) it has also been
postulated that Ce4+ can be reduced into Ce3+ in the presence of
H2O2 according to a catalase-mimetic mechanism [11].
ð5Þ
Åꢀ
3.6. Antagonist properties of cerium compounds towards superoxide
and superoxide dismutase (Table 3)
The results of Tables 2 and 3 support the reported significance of
O2Åꢀ scavenging in the protection by nano-CeO2 of cells under differ-
ent oxidative stress [4,6,8] and pH [22] conditions. Moreover there
have been several cell studies describing an upregulation of SOD
activity consecutive to nano-CeO2 treatment at low doses [6,22].
On the other hand the results of SOD preincubations with higher,
but still micromolar concentrations of nano-CeO2 (Table 3) strongly
suggest, however, that their accumulation within the cell may result
In a final series of experiments we first investigated whether
the protection of SOD levels of 3T3 cells seen in Table 2 could be
relevant to O2Åꢀ scavenging properties of nano-CeO2 in micromolar
concentrations. Therefore CL determinations were performed using
Åꢀ
lucigenin (5
l
M) as a chemiluminogenic probe for O2 generated
by the xanthine-xanthine oxidase reaction in the absence or pres-
ence of Ce compounds (5–50 M). Incubation of lucigenin alone
l
Åꢀ
in an overall loss of antioxidant defenses because the intrinsic O2
with Ce compounds (up to 50 lM) for up to 15 min yielded no light
scavenging potential of the particles will not compensate the de-
crease of SOD activity. This assumption of a detrimental effect of
nano-CeO2 overload in cells is in line with studies on bovine eryth-
rocytes showing that CeCl3 can bind the active site of SOD, and while
emission whereas O2Åꢀ-induced lucigenin CL at 5 min was concen-
tration-dependently inhibited by nano-CeO2 suggesting the parti-
Åꢀ
cles are good O2 scavengers. Lucigenin CL inhibition data of
individual Ce compounds at 5 l
M indicated the Ce3+ component
[Ce3+] <15
lM stimulated SOD activity, it was markedly decreased
of nano-CeO2 (tested as CeCl3) to be a far better radical scavenger
compared to Ce4+ (as Ce(SO4)2) or CeO2. In this assay addition of
SOD (20 mU/ml) inhibited 85% of lucigenin CL and uric acid absor-
bance was not modified by Ce compounds showing none of them
inhibit xanthine oxidase (data not shown).
when Ce3+ was added up to 25
lM in the system [64].
4. Conclusions
Second we tested the possibility that depressed SOD levels seen
in cells treated with a millimolar concentration of nano-CeO2 (Ta-
The present study imparts new information on the apparent
contradictory results of the literature where nanoceria is consid-

M. Culcasi et al. / Chemico-Biological Interactions 199 (2012) 161–176
175
ered either a protective antioxidant and free radical scavenger, or
pro-oxidant and cytotoxic. In this regard our work on human and
murine fibroblasts identifies two key factors: (i) the possibility that
particles can accumulate within a biological system following
chronic exposure (a case environmentally relevant to the wide-
A.1.2. [4-(4-Formylphenoxy)butyl]triphenylphosphonium iodide (2)
A reaction mixture containing (1) (3 g, 9.86 mmol) and tri-
phenyl phosphine (4.4 g, 16.79 mmol) was refluxed in toluene
(40 ml) for 12 h in the dark. After cooling down at room tempera-
ture, the mixture evaporated in vacuo, the residue was dissolved in
a minimal amount of dichloromethane and this solution was added
dropwise to excess ether (100 ml) under vigorous stirring. The
resultant yellow residue was allowed to settle for 1 h and the sol-
vent was decanted. The resulting precipitate was again submitted
three times to the above precipitation process and the final residue
was dried in vacuo for 2 h to afford (2) as a pale yellow solid
(5.13 g, 92%). 1H NMR (300 MHz, CDCl3): 1.87 (m, 2 H, P+–CH2-
CH2), 2.25 (2H, quin, J = 6.4 Hz, O–CH2-CH2), 3.85 (m, 2 H, P+–
CH2), 4.22 (t, 2 H, J = 5.7 Hz, OCH2), 6.95 (d, 2 H, J = 8.7 Hz, o-H,
Ar–O-C), 7.6–7.9 (m, 17 H, ArH), 9.86 (s, 1 H, CHO); 31P NMR
(121.5 MHz, CDCl3): 25.24 ppm; 13C NMR (75 MHz, CDCl3): 19.34
(d, J = 4.4 Hz, P+–CH2-CH2), 22.70 (d, J = 50 Hz, P+–CH2), 29.22 (d,
J = 16.00 Hz, P+–CH2-CH2-CH2), 67.12 (O–CH2), 115.01 (o-C, Ar–O–
C), 118.60 (d, J = 95.7 Hz, ꢀP+Ph3 ipso), 130.05 (p-C, Ar–O–C),
130.62 (d, J = 12.6 Hz, ꢀP+Ph3 meta), 132.13 (m-C, Ar–O–C),
133.77 (d, J = 9.9 Hz, ꢀP+Ph3 ortho), 135.23 (d, J = 2.7 Hz, ꢀP+Ph3
para), 163.88 (Ar–O–C), 190.94 (CHO).
spread use of nanoceria), and (ii) the dual signaling/damaging role
Åꢀ
of the H2O2/O2 couple. At micromolar doses (<300
lM), nano-
CeO2 activates Nox2 (and possibly Nox4)-dependent ROS produc-
tion at the membrane, interacts with signaling pathways leading
to protein synthesis activation and increases H2O2- and GSH-re-
lated enzyme activity. However these pro-oxidant properties of
low nano-CeO2 do not lead to any significant cytotoxicity despite
cell antioxidant status is altered. At millimolar doses of nano-
CeO2 both the NADPH oxidase and the mitochondrial sources of
ROS becomes highly activated leading to many known aspects
of cellular oxidative stress such as antioxidant depletion and loss
of viability. In that case mitochondria is supposed to command
alternative signaling pathways leading to early apoptosis. In this
cascade of damaging events where formation of HOÅ is prone to oc-
cur, Ce3+ ions are not likely to play a significant catalytic role, e.g.,
compared to iron. This can be considered as a great advantage for
biomedical applications when non-toxic, micromolar doses of nan-
oceria are used.
A.1.3. [4-[4-[[(1,1-Dimethylethyl)oxidoimino]methyl]phenoxy]butyl]-
triphenylphosphonium bromide (mitoPBN; 3)
Conflict of Interest Statement
MitoPBN (3) was obtained (0.95 g; 46%) by reducing (2) (2.0 g,
3.5 mmol) with Zn powder in dry ethanol, following the experi-
mental procedure reported in [28].
The authors declare that there are no conflicts of interest.
Acknowledgments
References
This work was funded by grants from the CNRS and the Agence
Nationale de la Recherche (INTOX, ANR-08-CES-002 and ROS SIG-
NAL, ANR-10-BLAN-003) at SMBSO-ICR UMR CNRS 7273. LB was
supported by a doctoral grant from Post-Grenelle funding via the
Antiopes network (Impecnano project) and thanks J. Rose and M.
Auffan at CEREGE (UMR 6635 CNRS, Aix-en-Provence, France) for
their kind help. We also thank P. Bernasconi, G. Gosset, A.S. Saba-
tier at DIPTA (Aix-en-Provence, France) and C. Questaigne at IN-
SERM UMR 911 (Marseille, France) for their kind technical
assistance.
[1] F.R. Cassee, E.C. van Balen, C. Singh, D. Green, H. Muijser, J. Weinstein, K.
Dreher, Exposure, health and ecological effects review of engineered nanoscale
cerium and cerium oxide associated with its use as a fuel additive, Crit. Rev.
Toxicol. 41 (2011) 213–229.
[2] A. Karakoti, S. Singh, J.M. Dowding, S. Seal, W.T. Self, Redox-active radical
scavenging nanomaterials, Chem. Soc. Rev. 39 (2010) 4422–4432.
[3] C.T. Campbell, C.H.F. Peden, Chemistry: oxygen vacancies and catalysis on ceria
surfaces, Science 309 (2005) 713–714.
[4] M. Das, S. Patil, N. Bhargava, J.F. Kang, L.M. Riedel, S. Seal, J.J. Hickman, Auto-
catalytic ceria nanoparticles offer neuroprotection to adult rat spinal cord
neurons, Biomaterials 28 (2007) 1918–1925.
[5] T. Xia, M. Kovochich, M. Liong, L. Madler, B. Gilbert, H. Shi, J.I. Yeh, J.I. Zink, A.E.
Nel, Comparison of the mechanism of toxicity of zinc oxide and cerium oxide
nanoparticles based on dissolution and oxidative stress properties, ACS Nano 2
(2008) 2121–2134.
Appendix A. Details of syntheses and characterization
[6] J. Colon, N. Hsieh, A. Ferguson, P. Kupelian, S. Seal, D.W. Jenkins, C.H. Baker,
Cerium oxide nanoparticles protect gastrointestinal epithelium from
radiation-induced damage by reduction of reactive oxygen species and
upregulation of superoxide dismutase 2, Nanomed. Nanotechnol. Biol. Med.
6 (2010) 698–705.
A.1. Synthesis of mitoPBN (3) (Scheme 1)
A.1.1. 4-(4-Iodobutoxy)benzaldehyde (1)
[7] J. Niu, A. Azfer, L.M. Rogers, X. Wang, P.E. Kolattukudy, Cardioprotective effects
To a cold solution of 4-hydroxybenzaldehyde (2 g, 16.39 mmol)
in acetonitrile (25 ml) was added solid K2CO3 (3.85 g, 27.86 mmol).
After stirring for 10 min in an ice-water bath, 1,4-diiodobutane
(25 g, 80.91 mmol) was added dropwise over 30 min, the reaction
mixture was allowed to increase to room temperature over 1 h and
was then stirred at room temperature for 12 h. After removal of the
precipitate, the organic layer was concentrated under vacuum and
the crude product was purified by flash column chromatography
on silica gel (CH2Cl2) to give 4-(4-iodobutoxy)benzaldehyde (1)
as a yellow oil (4.0 g, 13.10 mmol, 80%). 1H NMR (300 MHz, CDCl3):
1.82–2.03 (m, 4 H, O–CH2-CH2; I–CH2-CH2), 3.21 (t, 2 H, J = 6.8 Hz,
I–CH2), 4.01 (t, 2 H, J = 6.0 Hz, O–CH2), 6.93 (dd, 2 H, J = 2.00, 8.9 Hz,
o-H, Ar–O–C), 7.76 (dd, 2 H, J = 2.00, 8.9 Hz, m-H, Ar–O–C), 9.82 (s,
1 H, CHO); 13C NMR (75 MHz, CDCl3): 6.13 (I–CH2), 29.77 (I–CH2-
CH2; O–CH2-CH2), 66.85 (O–CH2), 114.52 (o-C, Ar–O–C), 129.68
(p-C, Ar–O–C), 131.76 (m-C, Ar–O–C), 163.65 (Ar–O–C), 190.53
(CHO). Anal. HRMS–ESI: Calcd for C11H13IO2: 303.9960. Found
303.9963.
of cerium oxide nanoparticles in
cardiomyopathy, Cardiovasc. Res. 73 (2007) 549–559.
[8] A.Y. Estevez, S. Pritchard, K. Harper, J.W. Aston, A. Lynch, J.J. Lucky, J.S.
Ludington, P. Chatani, W.P. Mosenthal, J.C. Leither, S. Andreescu, J.S. Erlichman,
a
transgenic murine model of
Neuroprotective mechanism of cerium oxide nanoparticles in
hippocampal brain slice model of ischemia, Free Radical Biol. Med. 51 (2011)
155–163.
a mouse
[9] C. Korsvik, S. Patil, S. Seal, W.T. Self, Superoxide dismutase mimetic properties
exhibited by vacancy engineered ceria nanoparticles, Chem. Commun. 14
(2007) 1056–1058.
[10] E.G. Heckert, A.S. Karakoti, S. Seal, W.T. Self, The role of cerium redox state in
the SOD mimetic activity of nanoceria, Biomaterials 29 (2008) 2705–2709.
[11] T. Pirmohamed, J.M. Dowding, S. Singh, B. Wasserman, E. Heckert, A.S.
Karakoti, J.E. King, S. Seal, W.T. Self, Nanoceria exhibit redox state-dependent
catalase mimetic activity, Chem. Commun. 46 (2010) 2736–2738.
[12] W. Lin, Y.W. Huang, X.D. Zhou, Y. Ma, Toxicity of cerium oxide nanoparticles in
human lung cancer cells, Int. J. Toxicol. 25 (2006) 451–457.
[13] E.J. Park, J. Choi, Y.K. Park, K. Park, Oxidative stress induced by cerium oxide
nanoparticles in cultured BEAS-2B cells, Toxicology 245 (2008) 90–100.
[14] M. Auffan, J. Rose, T. Orsiere, M. De Meo, A. Thill, O. Zeyons, O. Proux, A.
Masion, P. Chaurand, O. Spalla, A. Botta, M.R. Wiesner, J.Y. Bottero, CeO2
nanoparticles induce DNA damage towards human dermal fibroblasts in vitro,
Nanotoxicology 3 (2009) 161–171.

176
M. Culcasi et al. / Chemico-Biological Interactions 199 (2012) 161–176
[15] S.S. Hardas, D.A. Butterfield, R. Sultana, M.T. Tseng, M. Dan, R.L. Florence, J.M.
Unrine, U.M. Graham, P. Wu, E.A. Grulke, R.A. Yokel, Brain distribution and
protein 53-induced nuclear protein 1 is a major mediator of p53 antioxidant
function, Cancer Res. 69 (2009) 219–226.
toxicological evaluation of
ceria, Toxicol. Sci. 116 (2010) 562–576.
a
systemically delivered engineered nanoscale
[39] D.R. Duling, Simulation of multiple isotropic spin-trap EPR spectra, J. Magn.
Reson. 104 (1994) 105–110.
[16] H. Zhang, X. He, Z. Zhang, P. Zhang, Y. Li, Y. Ma, Y. Kuang, Y. Zhao, Z. Chai, Nano-
CeO2 exhibits adverse effects at environmental relevant concentrations,
Environ. Sci. Technol. 45 (2011) 3725–3730.
[17] J.Y. Ma, H. Zhao, R.R. Mercer, M. Barger, M. Rao, T. Meighan, D. Schwegler-
Berry, V. Castranova, J.K. Ma, Cerium oxide nanoparticle-induced pulmonary
inflammation and alveolar macrophage functional change in rats,
Nanotoxicology 5 (2011) 312–325.
[40] Y. Li, H. Zhu, P. Kuppusamy, V. Roubaud, J.L. Zweier, M.A. Trush, Validation of
lucigenin (bis-N-methylacridinium) as a chemilumigenic probe for detecting
superoxide anion radical production by enzymatic and cellular systems, J. Biol.
Chem. 273 (1998) 2015–2023.
[41] B. Halliwell, J.M.C. Gutteridge, O.I. Aruoma, The deoxyribose method: a simple
‘‘test-tube’’ assay for determination of rate constants for reactions of hydroxyl
radicals, Anal. Biochem. 165 (1987) 215–219.
[18] B.P. Kumar, S.L. D’Souza, K. Shivakumar, K. Rathinam, Cerium stimulates
protein biosynthesis in rat heart in vivo, Biol. Trace Elem. Res. 50 (1995) 237–
242.
[42] K. Makino, T. Hagiwara, A. Hagi, M. Nishi, A. Murakami, Cautionary note for
DMPO spin trapping in the presence of iron ion, Biochem. Biophys. Res.
Commun. 172 (1990) 1073–1080.
[19] Y. Cheng, Y. Li, R. Li, J. Lu, K. Wang, Orally administrated cerium chloride
induces the conformational changes of rat hemoglobin, the hydrolysis of 2,3-
DPG and the oxidation of heme-Fe(II), leading to changes of oxygen affinity,
Chem. Biol. Interact. 125 (2000) 191–208.
[20] E.G. Heckert, S. Seal, W.T. Self, Fenton-like reaction catalyzed by the rare earth
inner transition metal cerium, Environ. Sci. Technol. 42 (2008) 5014–5019.
[21] A. Asati, S. Santra, C. Kaittanis, J.M. Perez, Surface-charge-dependent cell
localization and cytotoxicity of cerium oxide nanoparticles, ACS Nano 4 (2010)
5321–5331.
[43] K. Ranguelova, R.P. Mason, The fidelity of spin trapping with DMPO in
biological systems, Magn. Reson. Chem. 49 (2011) 152–158.
[44] M. Culcasi, A. Rockenbauer, A. Mercier, J.L. Clément, S. Pietri, The line
asymmetry of electron spin resonance spectra as a tool to determine the
cis:trans ratio for spin-trapping adducts of chiral pyrrolines N-oxides: the
mechanism of formation of hydroxyl radical adducts of EMPO, DEPMPO, and
DIPPMPO in the ischemic-reperfused rat liver, Free Radical Biol. Med. 40
(2006) 1524–1538.
[45] S. Dikalov, P. Tordo, A. Motten, R.P. Mason, Characterization of the high
resolution ESR spectra of the methoxyl radical adduct of 5-
(diethoxyphosphoryl)-5-methyl-1-pyrroline N-oxide (DEPMPO), Free Radical
Res. 37 (2003) 705–712.
[46] J.L. Clément, B.C. Gilbert, A. Rockenbauer, P. Tordo, Radical damage to proteins
studied by EPR spin-trapping techniques, J. Chem. Soc. Perkin Trans. 2 (2001)
1463–1470.
[47] R.J. Singh, H. Karoui, M.R. Gunther, J.S. Beckman, R.P. Mason, B. Kalyanaraman,
Reexamination of the mechanisms of hydroxyl radical adducts formed from
the reaction between familial amyotrophic lateral sclerosis-associated Cu, Zn
superoxide dismutase mutants, Proc. Nat. Acad. Sci. USA 95 (1998) 6675–6680.
[48] S. Singh, T. Dosani, A.S. Karakoti, A. Kumar, S. Seal, W.T. Self, A phosphate-
dependent shift in the redox state of cerium oxide nanoparticles and its effects
on catalytic properties, Biomaterials 32 (2011) 6745–6753.
[22] 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 and anti-
invasive properties of redox-active nanoparticles in tumor-stroma
interactions, Biomaterials 32 (2011) 2918–2929.
[23] P. Ji, L. Wang, F. Chen, J. Zhang, Ce3+-centric organic pollutant elimination by
CeO2 in the presence of H2O2, Chem. Cat. Chem. 2 (2010) 1552–1554.
[24] A. Nel, T. Xia, L. Madler, N. Li, Toxic potential of materials at the nanolevel,
Science 311 (2006) 622–627.
[25] D.P. Jones, Radical-free biology of oxidative stress, Am. J. Physiol. Cell Physiol.
295 (2008) C849–C868.
[26] C. Fréjaville, H. Karoui, B. Tuccio, F. Le Moigne, M. Culcasi, S. Pietri, R. Lauricella,
P. Tordo, 5-(Diethoxyphosphoryl)-5-methyl-1-pyrroline N-oxide:
a new
efficient phosphorylated nitrone for the in vitro and in vivo spin trapping of
oxygen-centered radicals, J. Med. Chem. 38 (1995) 258–265.
[49] G. Czapski, B.H.J. Bielski, N. Sutin, The kinetics of the oxidation of hydrogen
peroxide by cerium(IV), J. Phys. Chem. 67 (1963) 201–203.
[50] C.R. Pottenger, D.C. Johnson, Mechanism of cerium(IV) oxidation of glucose
and cellulose, J. Polym. Sci. A 8 (1970) 301–318.
[27] S. Pietri, T. Liebgott, C. Fréjaville, P. Tordo, M. Culcasi, Nitrone spin traps and
their pyrrolidine analogs in myocardial reperfusion injury: hemodynamic and
ESR implications. Evidence for a cardioprotective phosphonate effect for 5-
(diethoxyphosphoryl)-5-methyl-1-pyrroline N-oxide in rat hearts, Eur. J.
Biochem. 254 (1998) 256–265.
[28] M.P. Murphy, K.S. Echtay, F.H. Blaikie, J. Asin-Cavuela, H.M. Cochemé, K. Green,
J.A. Buckingham, E.R. Taylor, F. Hurrell, G. Hughes, S. Miwa, C.E. Cooper, D.A.
Svistunenko, R.A. Smith, M.D. Brand, Superoxide activates uncoupling proteins
by generating carbon-centered radicals and initiating lipid peroxidation:
studies using a mitochondria-targeted spin trap derived from alpha-phenyl-N-
tert-butylnitrone, J. Biol. Chem. 278 (2003) 48534–48545.
[51] R. Preeta, R.R. Nair, Stimulation of cardiac fibroblast proliferation by cerium: a
superoxide anion-mediated response, J. Mol. Cell. Cardiol. 31 (1999) 1573–
1580.
[52] K. Anzai, T. Aikawa, Y. Furukawa, Y. Matsushima, S. Urano, T. Ozawa, ESR
measurement of rapid penetration of DMPO and DEPMPO spin traps through
lipid bilayer membranes, Arch. Biochem. Biophys. 415 (2003) 251–256.
[53] H. Shi, G. Timmins, M. Monske, A. Burdick, B. Kalyanaraman, Y. Liu, J.L.
Clément, S. Burchiel, K.J. Liu, Evaluation of spin trapping agents and trapping
conditions for detection of cell-generated reacive oxygen species, Arch.
Biochem. Biophys. 437 (2005) 59–68.
[29] L. Yu, R. Cao, W. Yi, Q. Yan, Z. Chen, L. Ma, W. Peng, H. Song, Synthesis of 4-
[(diethylamino)methyl]-phenol derivatives as novel cholinesterase inhibitors
with selectivity towards butyrylcholinesterase, Bioorg. Med. Chem. Lett. 20
(2010) 3254–3258.
[54] S. Arbault, P. Pantano, N. Sojic, C. Amatore, M. Best-Belpomme, A. Sarasin, M.
Vuillaume, Activation of the NADPH oxidase in human fibroblasts by
[30] Y. Xu, B. Kalyanaraman, Synthesis and ESR studies of a novel cyclic nitrone spin
trap attached to a phosphonium group-a suitable trap for mitochondria-
generated ROS?, Free Radical Res 41 (2007) 1–7.
[31] S. Pietri, M. Culcasi, L. Stella, P.J. Cozzone, Ascorbyl free radical as a reliable
indicator of free-radical-mediated myocardial ischemic and post-ischemic
injury. A real-time continuous-flow ESR study, Eur. J. Biochem. 193 (1990)
845–854.
mechanical intrusion of
Carcinogenesis 18 (1997) 569–574.
a
single cell with an ultramicroelectrode,
[55] B.A. t’Hart, J.M. Simmons, S. Knaan-Shanzer, N.P. Bakker, R.P. Labadie,
Antiarthritic activity of the newly developed neutrophil oxidative burst
antagonist apocynin, Free Radical Biol. Med. 9 (1990) 127–131.
[56] D.I. Brown, K.K. Griendling, Nox proteins in signal transduction, Free Radical
Biol. Med. 47 (2009) 1239–1253.
[32] P. N’Guessan, L. Pouyet, G. Gosset, S. Hamlaoui, M. Seillier, C.E. Cano, M. Seux,
M. Culcasi, J. Iovanna, N. Dusetti, S. Pietri, A. Carrier, Absence of tumor
suppressor Tumor Protein 53-Induced Nuclear Protein 1 sensitizes mouse
thymocytes and embryonic fibroblasts to redox-driven p53-independent
apoptosis, Antioxid. Redox Signal. 15 (2011) 1639–1653.
[33] T. Ago, F. Kuribayashi, H. Hiroaki, R. Tayeka, T. Ito, D. Kohda, H. Sumimoto,
Phosphorylation of p47phox directs phox homology domain from SH3 domain
toward phosphoinositides leading to phagocyte NADPH oxidase activation,
Proc. Nat. Acad. Sci. USA 100 (2003) 4474–4479.
[34] O.W. Griffith, Determination of glutathione and glutathione disulfide using
glutathione reductase and 2-vinylpyridine, Anal. Biochem. 106 (1980) 207–
212.
[35] S.T. Omaye, J.D. Turnbull, H.E. Sauberlich, Selected methods for the
determination of ascorbic acid in animal cells, tissues, and fluids, Methods
Enzymol. 62 (1979) 3–11.
[36] M. Culcasi, A. Muller, A. Mercier, J.L. Clément, O. Payet, A. Rockenbauer, V.
Marchand, S. Pietri, Early specific free radical-related cytotoxicity of gas phase
cigarette smoke and its paradoxical temporary inhibition by tar, Chem. Biol.
Interact. 164 (2006) 215–231.
[37] C.R. Wheeler, J.A. Salzman, N.M. Elsayed, S.T. Omaye, D.W. Korte Jr., Automated
assays for superoxide dismutase, catalase, glutathione peroxidase, and
glutathione reductase activity, Anal. Biochem. 184 (1990) 193–199.
[38] C.E. Cano, J. Gommeaux, S. Pietri, M. Culcasi, S. Garcia, M. Seux, S. Barelier, S.
Vasseur, R.P. Spoto, M.J. Pébusque, N.J. Dusetti, J.L. Iovanna, A. Carrier, Tumor
[57] N. Anilkumar, R. Weber, M. Zhang, A. Brewer, A.M. Shah, Nox4 and Nox2
NADPH oxidases mediate distinct cellular redox signaling responses to agonist
stimulation, Arterioscler. Thromb. Vasc. Biol. 28 (2008) 1347–1354.
[58] M. Vejrazka, R. Micek, S. Stipek, Apocynin inhibits NADPH oxidase in
phagocytes but stimulates ROS production in non-phagocytic cells, Biochim.
Biophys. Acta 1722 (2005) 143–147.
[59] A.J. Lambert, A.J. Buckingham, H.M. Boysen, M.D. Brand, Diphenyleneiodonium
acutely inhibits reactive oxygen species production by mitochondrial complex
I during reverse, but not forward electron transport, Biochim. Biophys. Acta
1777 (2008) 397–403.
[60] E. Cadenas, A. Boveris, Mitochondrial free radical production, antioxidant
defenses and cell signaling, The handbook of environmental chemistry
(2005) 219–234, Springer-Verlag, Berlin, Heidelberg.
2
[61] A. Russo, S. Caggia, M. Piovano, J. Garbarino, V. Cardile, Effect of vicanicin and
protolichesterinic acid on human protate cancer cells: role of Hsp70 protein,
Chem. Biol. Interact. 195 (2012) 1–10.
[62] M.J. Davies, R.T. Dean, Radical-mediated protein oxidation: from chemistry to
medicine, Oxford University Press, Oxford, 1997.
[63] M. Kawagoe, F. Hirasawa, S.C. Wang, Y. Liu, Y. Ueno, T. Sugiyama, Orally
administrated rare earth element cerium induces metallothionein synthesis
and increases glutathione in the mouse liver, Life Sci. 77 (2005) 922–937.
[64] P. Liu, L. Ma, S. Yin, F. Hong, Effect of Ce3+ on conformation and activity of
superoxide dismutase, Biol. Trace Elem. Res. 125 (2008) 170–178.