TOXICOLOGICAL SCIENCES 108(2), 472–481 (2009)

doi:10.1093/toxsci/kfp014

Advance Access publication January 28, 2009

Sintered Indium-Tin-Oxide (ITO) Particles: A New Pneumotoxic Entity

Dominique Lison,*,1 Julie Laloy,* Ingrid Corazzari,† Julie Muller,* Virginie Rabolli,* Nadtha Panin,* Francxois Huaux,*

Ivana Fenoglio,† and Bice Fubini†

*Industrial Toxicology and Occupational Medicine unit, Catholic University of Louvain, Brussels, Belgium; and †Dipartimento di Chimica IFM,

Interdepartmental Center ‘‘G. Scansetti’’ for Studies on Asbestos and other Toxic Particulates and Interdepartmental Center for Nanostructured Interfaces andSurfaces, Universita degli Studi di Torino, 7 10125 Torino, Italy

Received October 22, 2008; accepted January 14, 2009

Indium-Tin-Oxide (ITO) is a sintered mixture of indium-

(In2O3) and tin-oxide (SnO2) in a ratio of 90:10 (wt:wt) that is

used for the manufacture of LCD screens and related high

technology applications. Interstitial pulmonary diseases have

recently been reported in workers from ITO producing plants.

The present study was conducted to identify experimentally the

exact chemical component responsible for this toxicity and to

address possible mechanisms of action. The reactivity of respirable

ITO particles was compared with that of its single components

alone or their unsintered 90:10 mixture (MIX) both in vivo and

in vitro. For all endpoints considered, ITO particles behaved as

a specific toxic entity. In vivo, after a single pharyngeal

administration (2–20 mg per rat), ITO particles induced a strong

inflammatory reaction. At day 3, the inflammatory reaction (cell

accumulation, LDH and protein in bronchoalveolar lavage fluid)

appeared more marked with ITO particles than with each oxide

separately or the MIX. This inflammatory reaction persisted and

even worsened after 15 days. After 60 days, this inflammation was

still present but no significant fibrotic response was observed. The

cytotoxicity of ITO was assessed in vitro in lung epithelial cells

(RLE) and macrophages (NR8383 cell line). While ITO particles

(up to 200 mg/ml) did not affect epithelial cell integrity (LDH

release), a strong cytotoxic response was found in macrophages

exposed to ITO, but not to its components alone or mixed. ITO

particles also induced an increased frequency of micronuclei in

type II pneumocytes in vivo but not in RLE in vitro, suggesting the

preponderance of a secondary genotoxic mechanism. To address

the possible mechanism of ITO toxicity, reactive oxygen species

production was assessed by electron paramagnetic resonance

spectrometry in an acellular system. Carbon centered radicals

(COO-�) and Fenton-like activity were detected in the presence of

ITO particles, not with In2O3, SnO2 alone, or the MIX. Because

the unsintered mixture of SnO2 and In2O3 particles was unable to

reproduce the reactivity/toxicity of ITO particles, the sintering

process through which SnO2 molecules are introduced within the

crystal structure of In2O3 appears critical to explain the unique

toxicological properties of ITO. The inflammatory and genotoxic

activities of ITO dust indicate that a strict control of exposure is

needed in industrial settings.

Key Words: indium oxide; ITO; occupational lung disease;

sintering.

Indium-Tin-Oxide (ITO) is a sintered material composed of

indium- (In2O3) and tin-oxide (SnO2), generally made up in

a 90:10 ratio (wt:wt). This material is mainly used for the

manufacture of LCD screens and related high technology

applications because of its unique properties of high electrical

conductivity, transparency and mechanical resistance. Thin

films of ITO are most commonly deposited on surfaces by

electron beam evaporation, physical vapor deposition, or

a range of sputtering techniques. The specific electronic

properties of ITO result from the introduction of a high density

of free electrons and oxygen vacancies in the In2O3 crystal

structure through the sintering with SnO2.

The medical literature indicates the occurrence of interstitial

pulmonary diseases in workers from ITO plants. Two cases of

interstitial pneumonia were first reported in workers in Japan

(Homma et al., 2003, 2005). Later, a cross-sectional study was

carried out in a Japanese factory where 3 out of 115 workers

had developed interstitial pulmonary changes accompanied by

pulmonary emphysema. Among all these workers, a relation

was reported between exposure assessed through the level of

indium in serum and the biochemical (KL-6 in serum) and

pulmonary imaging (HR-CT scan) changes (Chonan et al.,2007). A second cross-sectional study carried out in Japanese

workers from ITO manufacturing and recycling plants

confirmed the risk of interstitial lung disease, based on the

finding of increased serum biomarkers of lung toxicity (KL-6,

SP-A, and SP-D) in exposed workers (Hamaguchi et al., 2008).

The exact chemical responsible for this lung toxicity is

unknown and existing experimental data are not directly useful

to discriminate between the toxicity of ITO and that of indium

or tin oxides. Experimental studies have been conducted with

indium selenide (Morgan et al., 1997), indium phosphide

(Gottschling et al., 2001; Oda, 1997), or indium arsenide

(Tanaka et al., 1996, 2000) and all reported inflammatory and

fibrotic reactions in the lungs of rats or hamsters. Specific data

1 To whom correspondence should be addressed at Industrial Toxicology

and Occupational Medicine unit, Avenue E. Mounier, 53.02, 1200 Brussels,

Belgium. Fax: þ32-2-764-53-38. E-mail: dominique.lison@uclouvain.be.

� The Author 2009. Published by Oxford University Press on behalf of the Society of Toxicology. All rights reserved.For permissions, please email: journals.permissions@oxfordjournals.org

by guest on February 6, 2016http://toxsci.oxfordjournals.org/

Dow

nloaded from

for indium oxide are not available. Only one study in hamsters

indicates that repeated intratracheal instillation of ITO particles

induced lung inflammation after 16 weeks, but less than with

indium phosphide (Tanaka et al., 2002).

We hypothesized that, in view of its electronical properties,

and by analogy with other materials such as hard metals (WC-

Co) (Lison et al., 1995), ITO might behave as a very specific

toxic entity and act as a source of reactive oxygen species

(ROS).

We therefore (1) compared experimentally the pulmonary

response to ITO particles (inflammation, fibrosis, genotox-

icity), their individual components alone or mixed without

sintering, (2) examined the cytotoxicity of the same particles in

cell culture models, and (3) evaluated the capacity of ITO

particles and their individual components alone or mixed to

produce ROS or to initiate oxidative damage in an acellular

system,.

MATERIALS AND METHODS

Particles tested. In an attempt to reproduce the type of particles the

diseased Japanese workers had been exposed to, ITO particles were obtained by

milling 90:10 tile fragments (Umicore Indium Products, Providence, RI) with

a tungsten carbide Mix Mill 8000 (Spex Certiprep, Metuchen, NJ). The sub-635

mesh sieve fraction of the resulting powder was used as test material. Tin-

(SnO2) and indium-oxide (In2O3) powders were supplied by the same

company. The main characteristics of these particles are summarized in

Table 1. Elemental composition was determined by inductively coupled plasma

mass spectrometry, size distribution was assessed with a Microtrac X100

(Microtrac, Montgomeryville, PA) and specific surface area (Brunauer,

Emmett, and Teller method) was measured on a Tristar 3000 (Micromeritics

Instrument Corporation, Norcross, GA).

To test the hypothesis that ITO particles would induce an elective reactivity/

toxicity greater than the addition of the effects induced by its individual

components, we also prepared a mixture of In2O3/SnO2 (MIX) by manually

blending nine weight parts of In2O3 with one part of SnO2 in an agate mortar,

without sintering. Two types of crystalline silica particles were used as positive

controls: Min-U-Sil 5 quartz (specific surface area 5.2 m2/g, Berkley Springs,

VA) for electron paramagnetic resonance (EPR) experiments, or DQ12 quartz

(specific surface area 3.5 m2/g, Dorentrup, Germany) for in vitro cytotoxicity

tests. Hard metal (WC-Co) particles used as positive control for genotoxicity

experiments were as described previously (d50, 2 lm; specific surface area,

1.76 m2/g) (De Boeck et al., 2003). All particles used in cellular and animal

studies were heated at 200�C during 2 h in order to remove any possible

endotoxin contamination, and suspended freshly in a 0.9% saline solution. The

particle suspensions were further diluted in buffer for EPR measurements, cell

culture medium or saline for in vivo experiments, and vigorously sonicated

immediately before use.

In vivo experiments. Female Wistar rats (190–220 g) were obtained from

the local breeding facility and maintained in a conventional environment with

free access to water and dried food. The particles were administered in the

lungs by pharyngeal aspiration under anesthesia with 18 mg Ketalar (Warner-

Lambert, Zaventem, Belgium) and 0.5 mg Rompun (Bayer, Leverkussen,

Germany) per rat given ip. The particles were suspended in 300 ll of sterile

saline, except for the highest dose of ITO particles for which a volume of

500 ll was used. The particle suspension was placed posterior on the throat and

the tongue, which was held until the suspension was aspirated into the lungs

(Rao et al., 2003). Control animals were treated with an identical volume of

sterile saline.

At selected time intervals, the animals were sacrificed with an ip injection of

sodium pentobarbital (60 mg per rat), the thoracic cavity was exposed and the

lungs were perfused with sterile saline via the right ventricle.

The inflammatory response was assessed by analyzing bronchoalveolar

lavage (BAL) fluid and histologically. The trachea was dissected and

cannulated to perform a BAL with five iterative washes of the lungs using

6 ml of sterile saline. The samples were centrifuged at 4�C during 10 min at

1500 rpm to separate the supernatant that was used for biochemical analyses

(LDH activity, total protein concentration, tumor necrosis factor [TNF]-a) and

the pellet resuspended in 2 ml of phosphate buffered saline for cellular counts.

LDH activity, total protein concentration, total and differential cell counts were

performed as described previously (Lasfargues et al., 1995). TNF-a was assayed

with a commercial enzyme-linked immunosorbent assay (OptEIA, Pharmingen,

San Diego, CA).

Separate animals were treated with the same particles for histological

analysis of the lungs. The lungs were fixed in 3.7% formaldehyde and

embedded to perform 5 lm thick slices that were stained with hematoxilin-

eosin or Masson green trichrome.

The fibrotic response was also assessed biochemically in separate groups of

animals by measuring the lung hydroxyproline (Biondi et al., 1997) and soluble

collagen content (Sircoll, Biocolor, Belfast, Ireland) as reported previously (van

den Brule et al., 2007).

The in vivo genotoxic potential of ITO particles was evaluated by measuring

the induction of micronuclei (MN) in type II lung epithelial cells of rats as

described previously (De Boeck et al., 2003; Muller et al., 2008) with slight

modifications. Type II epithelial cells were isolated from the lung of rats treated

with ITO particles and digested with elastase and panning on plates coated with

IgG as reported by Dobbs et al. (1986).

The protocol of these studies was approved by the local ethical committee

for the use of animals in research (Universite catholique de Louvain, faculty of

Medicine).

In vitro experiments. The rat lung epithelial cell line (RLE), originating

from type II lung epithelial cells of a male F344 rat, was kindly provided by Dr

C. Albrecht (Dortmund, Germany) and cultured in Ham’s F12 medium with

10% fetal bovine serum (FBS), 1% penicillin/streptomycin, 1% L-glutamine (all

obtained from Gibco, Grand Island, NY). The NR8383 alveolar macrophage

cell line, originating from the BAL of a normal Sprague-Dawley rat, was

TABLE 1

Characteristics of Tested Particles

In2O3 SnO2 ITO

In2O3 90%,

SnO2 10 %

Ni (%) < 0.001 < 0.001 < 0.001

Fe (%) < 0.001 0.005 < 0.001

Zn (%) nd 0.001 < 0.001

Cu (%) < 0.001 0.001 < 0.001

Pb (%) < 0.001 0.001 < 0.001

Sb (%) < 0.001 0.001 < 0.001

Water solubility Insoluble Insoluble Insoluble

Median mass

diameter (d50) (lm)

~6 0.7

more than 90% < 5

7

75% < 20

58% < 10

41% < 5

Specific surface area

(m2/g)

0.79 4.62 0.74

Density (g/cm3) 7.179 6.950 7.179

SINTERED INDIUM-TIN-OXIDE (ITO) PARTICLES 473

by guest on February 6, 2016http://toxsci.oxfordjournals.org/

Dow

nloaded from

obtained from Dr C. Albrecht and cultured in Kraighn’s modification of Ham’s

F12 medium (F12-K) containing 2mM glutamine, modified to contain 1.5 g

sodium bicarbonate/l and supplemented with 15% FBS, 1% penicillin/

streptomycin, 1% L-glutamine. The cells were kept at 37�C in a humidified

incubator with 5% CO2.

The different cell types were incubated for 24 h with the test particles

dispersed in their respective cell culture medium without FBS and cytotoxicity

was assessed by measuring the release of LDH in the culture medium as

described previously (Lison and Lauwerys, 1990).

NR8383 cells were also incubated with ITO particles in the presence of

Trolox (Sigma, St Louis, MO) or butylated hydroxytoluene (BHT, Sigma) in

order to examine the role of ROS in the cytotoxic response. Cells were

preincubated with the antioxidants 2 h before and during the exposure period in

the absence of FBS.

The primary genotoxic potential of ITO particles was evaluated by assessing

the in vitro induction of MN in RLE cells as reported previously (Muller et al.,2008).

Free radical studies. The release of radical species was monitored by

EPR spectrometry (Miniscope 100 X-band EPR spectrometer, Magnettech,

Germany) by using the spin-trapping technique with 5-dimethyl-1-pyrroline-

N-oxide (DMPO, Alexis-Biochemicals, Lausen, Switzerland). To assess the

cleavage of C–H bonds, 45 mg of ITO, 45 mg of MIX, 40.5 mg of In2O3,

4.5 mg of SnO2 or 45 mg of Min-U-Sil 5 quartz were suspended in 0.5 ml of

potassium phosphate buffer (0.25M, pH 7.4) containing DMPO (0.085M)

and sodium formate (1M). The production of hydroxyl radicals was assessed

in the presence of H2O2 by incubating 45 mg of ITO, 45 mg of MIX, 40.5 mg

of In2O3, 4.5 mg of SnO2 or 45 mg of Min-U-Sil 5 in 1.25 ml of potassium

phosphate buffer (0.2M, pH 7.4) containing DMPO (0.034M) and H2O2

(0.080M). To verify the possible contribution of artifacts formed by the

addition of water to DMPO, 45 mg of ITO or MIX were suspended in 1.25 ml

of potassium phosphate buffer (0.2M, pH 7.4) containing DMPO (0.034M),

H2O2 (0.08M), and CH3OH (8M). For the detection of superoxide radicals,

7 mg of ITO, 7mg of MIX, 6.3 mg of In2O3, 0.7 mg of SnO2 or 7 mg of Min-

U-Sil 5 were suspended in 0.75 ml of potassium phosphate buffer (0.33M, pH

7.4) containing DMPO (0.057M). All suspensions were incubated during

60 min at 25�C and filtered through a 0.20 lm membrane before recording the

EPR spectra. All the experiments were repeated twice with similar results.

Statistical analyses. Data are presented as means ± SEM. Differences

among experimental groups were assessed with an ANOVA followed, when

appropriate, by a Student-Newman-Keuls test for multiple comparisons and

a linear trend test. Statistical significance was set at p < 0.05.

RESULTS

Pulmonary Effects in Response to ITO Particles

Acute response. The capacity of ITO and related particles

to induce an acute lung reaction was assessed 3 and 15 days

after particle administration. A BAL was performed in rats

instilled with ITO (2 or 20 mg), In2O3 (1.8 or 18 mg), SnO2

(0.2 or 2 mg), MIX (2 or 20 mg), or a saline solution (NaCl

0.9%, controls). LDH activity, a marker of cytotoxicity, was

strongly and significantly increased by the administration of

ITO particles, but not in a dose-dependent manner (Fig. 1A).

The MIX, SnO2 or In2O3 alone did not modify LDH activity.

Similar patterns were observed for other inflammatory

parameters such as the total protein concentration (Fig. 1B)

and the total cell count (Fig. 1C). The total cell count was

increased significantly with 18 mg In2O3. Cytospin prepara-

tions showed that increased BAL cellularity reflected an overall

expansion of macrophages, neutrophils, lymphocytes, and, to

a lesser extent, eosinophils in all experimental groups. At

day 15, inflammatory parameters were further increased

compared with day 3 after the administration of ITO particles

(Figs. 1D–F).

The pathological study revealed that the particles were

dispersed in the alveolar spaces after administration of the

lower dose of ITO particles (Fig. 2B); more aggregates were

detected with the higher dose (Fig. 2F). Alveolitis, character-

ized by a thickening of the alveolar wall was moderate in the

lung of rats given the lower dose of 2 mg (Fig. 2B), whereas it

was more pronounced and diffuse with the higher dose of

20 mg (Fig. 2C). The formation of nodules around particles

was also noted in the lung of animals dosed with 20 mg of ITO

particles. These nodules included mononuclear and poly-

morphonuclear cells (Fig. 2D).

Overall, the experiment indicated that ITO particles

induced a strong inflammatory lung reaction. At day 3, the

inflammatory reaction appeared more marked with ITO

particles than with each oxide separately or their mixture.

After 15 days, the inflammatory reaction to ITO persisted and

had even worsened, with a statistically significant dose-effect

relationship.

Delayed response. We then assessed the capacity of ITO

and related particles to induce a lung fibrotic response.

Collagen deposition was evaluated by measuring hydroxypro-

line and soluble collagen lung content 60 days after a single

particle administration. For both markers, no significant change

was found among the different groups suggesting that the

particles did not induce a fibrotic reaction in this experimental

model (Fig. 3). These results were confirmed by the

pathological study which did not reveal fibrosis 60 days after

the administration of ITO particles. The damage induced by

ITO was characterized by an alveolitis accompanied with

a thickening of the alveolar wall and the presence of

macrophages, lymphocytes and polymorphonucleated neutro-

phils (Fig. 3D). Perivascular inflammatory infiltrates, particles

(Fig. 3E) as well as a proteinacious material (Fig. 3F) were also

found in the alveolar lumen.

Overall, we concluded that ITO particles induced a protracted

inflammation but no fibrotic reaction in the lung of exper-

imental animals.

Tumor necrosis factor-a. TNF-a is a key mediator in

inflammation and fibrosis (Huaux, F. 2007). We, therefore,

examined the production of this cytokine after administration

of the different particles in rats. At the inflammatory stage (day

3), no significant increase of TNF-a was observed in BAL fluid

of treated rats (not shown). At the fibrotic stage (day 60),

a slight but significant increase of TNF-a (76.5 ± 18 versus

34.1 ± 10.3 pg/ml in controls) was found in animals treated

with the higher concentration of ITO particles (20 mg per rat).

No increase was observed after administration of the other

particles.

474 LISON ET AL.

by guest on February 6, 2016http://toxsci.oxfordjournals.org/

Dow

nloaded from

Genotoxic response in epithelial cells. In view of the

inflammatory reaction induced by ITO particles, which might

potentially lead to genotoxic events in epithelial cells (Schins

and Knaapen, 2007), we first examined the in vivo genotoxic

potential of ITO particles. The induction of MN was assessed

in type II pneumocytes 3 days after the administration of

2 doses of ITO (2 mg/rat was selected as a dose causing a clear

inflammatory reaction and 0.5 mg/rat was chosen as a subin-

flammatory dose), WC-Co was used as a positive control (De

Boeck et al., 2003). Figure 4 clearly shows that WC-Co caused

a strong increase in MN frequency and that ITO particles also

induced, although to a lesser extent, an increase in MN

frequency only at the inflammatory dose. Because this

genotoxic effect could be mediated by inflammatory cells and

their ROS products rather than by ITO particles themselves

(Schins and Knaapen, 2007), we assessed the capacity of ITO

particles to cause genotoxic effects in vitro in RLE cells. We

did not find increased MN frequencies in RLE cells treated

with ITO particles (50, 100, or 200 lg/ml), suggesting the

preponderance of secondary genotoxicity phenomena.

In Vitro Cytotoxicity Studies

To further document the toxicity of ITO particles, we then

studied cellular models (macrophages or epithelial cells)

potentially relevant for the lung response to ITO particles.

Epithelial cells. Because existing clinical studies con-

ducted in workers exposed to ITO (Chonan et al., 2007;

Hamaguchi et al., 2008) reported biochemical changes

reflecting epithelial toxicity (increased KL-6, SP-A and/or

SP-D serum levels) we first examined the response of epithelial

cells to ITO and related particles. Rat lung epithelial cells

FIG. 1. ITO particles induce an elective acute lung inflammatory response. Particles of SnO2 (0.2 or 2 mg per rat), In2O3 (1.8 or 18 mg per rat), MIX (2 or 20

mg per rat), ITO (2 or 20 mg per rat) were administered by pharyngeal aspiration to female Wistar rats (controls were treated with an equivalent volume of 0.9%

NaCl vehicle). LDH activity (A, D), total proteins (B, E), and cell number (C, F) were measured in BAL 3 days (A–C) or 15 days (D, E, F) after treatment. After 15

days, only ITO particles were tested. Data are presented as means ± SEM. (n ¼ 5). ***p < 0.001, **p < 0.01 relative to NaCl controls (Student-Newman-Keuls

test). At day 15, the dose-effect relationship was statistically significant (linear trend test, p < 0.001 for the three parameters).

SINTERED INDIUM-TIN-OXIDE (ITO) PARTICLES 475

by guest on February 6, 2016http://toxsci.oxfordjournals.org/

Dow

nloaded from

(RLE) were exposed during 24 h to SnO2 alone (200 lg/ml),

In2O3 alone (200 lg/ml), ITO (50, 100, or 200 lg/ml), the

mixture of SnO2/In2O3 (MIX; 200 lg/ml), crystalline SiO2 as

a positive control (200 lg/ml), or to culture medium (negative

control). The cytotoxic response was assessed by measuring

the LDH release after 24 h exposure. No effect of the particles

was observed except for cells exposed to DQ12 silica

particles (Fig. 5A). Microscopical observation showed that

ITO particles were not well taken up by RLE cells (Fig. 5B)

Macrophages. Macrophages play a major role in the

initiation of the cascade of events leading to inflammatory

reactions induced by particles in general. A significantly

increased LDH release was observed in the supernatant of

NR8383 cells exposed during 24 h to In2O3, ITO particles, the

SnO2/In2O3 mixture (MIX) as well as to silica but not SnO2

(Fig. 5C). This response was dose-dependent for ITO particles

and of a similar intensity than that induced by silica. Thus,

when compared on a surface area basis, ITO (0.74 m2/g) was

about five times more cytotoxic to NR8383 cells than silica

(3.5 m2/g). ITO particles were readily taken up by NR8383

cells (Fig. 5D).

NR8383 cells exposed in vitro to ITO particles (up to 200

lg/ml) did not show an increased production of TNF-a (not

shown).

We concluded that, in the present cytotoxicity assays, ITO

particles appear to have a greater cytotoxic activity in

macrophages than in lung epithelial cells.

FIG. 2. The acute reaction to ITO particles (day 3) is characterized by an alveolitis and the formation of inflammatory nodules. Lung sections of control rats

(A, E), rats treated with 2 mg (B) or 20 mg of ITO particles (C, D, E, F) stained with hematoxylin and eosin. Magnification: 320 (A–D) or 350 (E, F).

476 LISON ET AL.

by guest on February 6, 2016http://toxsci.oxfordjournals.org/

Dow

nloaded from

Assessment of the Capacity to Release ROS

As a next step to explore the potential toxicity of

ITO particles, we assessed their capacity to generate free

radicals in an aqueous suspension by means of the EPR/spin-

trapping technique with DMPO. Sodium formate was first

chosen as a target to assess the capacity of samples to induce

a cleavage of a C-H bond, a reaction representing the first

step to the oxidative damage of biomolecules. A carbon

centred radical (COO-�) was formed in the presence of

ITO particles, not with In2O3, SnO2 alone or the simple

SnO2/In2O3 mixture in proportions identical to ITO

(Fig. 6A).

We also tested the capacity of the samples to generate

hydroxyl radicals (HO�) via a Fenton-like reaction when

suspended in the presence of H2O2 (Fig. 6C). HO� radicals

were detected with ITO, but also with In2O3 alone or mixed

with SnO2, although to a lesser extent. Almost no HO� radicals

were detected with SnO2 only.

When the results were expressed not as a function of the

amount of dust, but of the particle surface area involved in the

reactions (insets in Figs. 6B and 6D), ITO was the most

reactive dust even in comparison with the Min-U-Sil quartz

used as positive control.

To exclude the possibility that the detected DMPO/HO�adducts were an artefact formed by the nucleophilic addition of

water to DMPO (Hanna et al., 1992), we performed the same

experiments in the presence of an excess of methanol. Under

these conditions, methanol being more nucleophilic than water,

the formation of DMPO/CH3O� would be observed (Burkitt

et al., 1995). In our experimental conditions the DMPO/CH3O�was not detected, thus indicating that the observed spectra were

not the result of DMPO degradation but represented a real

hydroxyl radical yield.

To examine the possible formation of superoxide radicals in

the presence of ITO particles, we performed the experiments in

the absence of formate. We did not detect the DMPO/HOO�

FIG. 3. Absence of fibrotic response in the lung of rats treated with a single administration of ITO particles. Hydroxyproline (A) and soluble collagen (B) lung

content after 60 days. Each column represents the average ± SEM of five observations. The results were analyzed statistically with the ANOVA and no significant

difference was noted. Lung sections (magnification 320) from saline (C) or ITO-treated rat (2 mg per rat: D–F).

SINTERED INDIUM-TIN-OXIDE (ITO) PARTICLES 477

by guest on February 6, 2016http://toxsci.oxfordjournals.org/

Dow

nloaded from

adduct, which would have been formed by the direct reaction

of HOO� (hydroperoxyl radicals, the conjugated acid of

superoxide) with DMPO or the DMPO/HO� adduct, which

may derive from the decomposition of DMPO/HOO� in

solution (Finkelstein et al., 1982) (data not shown).

We concluded that only ITO is reactive in the cleavage of the

C-H bond in the formate ion and that it is more active in

Fenton-like activity than its individual components alone or

combined in the In2O3/SnO2 mixture. Thus, ITO behaves as

a new reactive entity able to produce COO�� and HO� radicals.

To examine the relationship between ROS producing

capacity and toxicity, an attempt was made to block the

cytotoxicity of ITO particles with antioxidants. In NR8383

cells, we were not able to reduce the cytotoxicity of ITO

particles by coincubating the cells in the presence of up to

1mM of Trolox or BHT (not shown).

DISCUSSION

The elective capacity of ITO particles to produce free radicals,

to cause cytotoxicity in macrophages and to trigger an

inflammatory lung reaction in experimental animals, demon-

strated in the present study, contributes to explain the pulmonary

manifestations reported in exposed workers (Chonan et al.,2007; Hamaguchi et al., 2008; Homma et al., 2003, 2005). The

in vivo genotoxic potential of ITO particles for lung epithelial

cells indicates a capacity to induce lung cancers.

All the results, from free radical production to in vivopulmonary responses, consistently indicate that ITO is

a specifically reactive/toxic entity, different from its single

components considered separately or blended. Because the

simple mixture of SnO2 and In2O3 particles was unable to

FIG. 4. In vivo genotoxic effects of ITO particles in type II pneumocytes of

rats. The frequency of micronuleated cells was assessed in type II pneumocytes

isolated from rats treated 3 days before with a single pharyngeal aspiration of

ITO (0.5 or 2 mg per rat), WC-Co (2 mg per rat, positive control) or saline

(controls). Each column represents the mean ± SEM of three to four animals.

***p < 0.001, *p < 0.05 relative to controls (Student-Newman-Keuls test).

FIG. 5. ITO particles are cytotoxic towards macrophages but not lung epithelial cells. The cells were cultured in 24-well plates (2 cm2 per well) and exposed to

SnO2, In2O3, a mixture of SnO2/In2O3 (MIX), ITO, DQ12 quartz (SiO2) dispersed in 1 ml of culture medium. Controls were incubated in culture medium only.

Twenty-four hours after exposure LDH release was assessed in the supernatant of rat lung epithelial cells (RLE, panel A) or NR8383 rat macrophage cells (C).

Each column represents the mean ± SEM. of four observations. ***Highlights a significant difference (p < 0.001) between values measured in treated cells

compared with controls (Student-Newman-Keuls test). ITO particles were readily taken up by NR8383 (D) but not by RLE (B) cells. Microphotographs were taken

after 5 h of incubation with 200 lg ITO/ml (magnification 3200).

478 LISON ET AL.

by guest on February 6, 2016http://toxsci.oxfordjournals.org/

Dow

nloaded from

reproduce the reactivity/toxicity of ITO particles, the sintering

process through which SnO2 molecules are introduced within

the crystal structure of In2O3 appears critical to explain the

unique properties of ITO. This toxicological interaction

between two metallic compounds is similar to what was

previously reported for another lung toxicant, namely hard

metals (Lison et al., 1995), although with some differences.

Hard metals are pseudocomposite materials produced by

a powder sintering process through which tungsten carbide

particles (WC) are cemented into a cobalt metal matrix. In hard

metals, toxic ROS are formed through the reduction of ambient

oxygen by cobalt metal, a process catalyzed at the surface of

WC particles as a consequence of the close contact between

particles, and this interaction could be reproduced with a simple

mixture of Co and WC particles (Lison et al., 1995). In ITO

dusts, the basis of the toxic interaction appears not to be at the

interface between two particles in contact, but at the structural/

molecular level. In ITO, free radicals are most probably

generated at reactive sites caused by the introduction of

substitutional Sn in the overall crystal structure, where electron

density is high (Fan and Goodenough, 2008).

Only ITO, and not its components alone or mixed, showed

the capacity to cause a rupture of C-H bonds, with the

consequent release of carbon centered radicals.

Contrary to hard metals, no oxygenated free radicals were

detected in the absence of formate, suggesting that the reaction

FIG. 6. ITO particles selectively generate reactive oxygen species in aqueous suspension. DMPO spin-trapping spectra were recorded after incubating the

particles suspended in the appropriate buffer during 60 min at 25�C and filtering the liquid through a 0.22-lm membrane. Cleavage of H-C bonds in formate (A):

45 mg of ITO, 45 mg of MIX, 40.5 mg of In2O3, 4.5 mg of SnO2, or 45 mg of Min-U-Sil 5 quartz (positive control) were suspended in 0.5 ml of potassium

phosphate buffer (0.25M, pH 7.4) containing DMPO (0.085M) and sodium formate (1M). In (B), the intensity of the signals is normalized to the surface area of the

analyzed particles. The experiments were performed twice with similar results. The production of HO� radicals (C) was assessed in the presence of H2O2 by

incubating 45 mg of ITO, 45 mg MIX, 40.5 mg of In2O3, 4.5 mg of SnO2, or 45 mg of quartz in 1.25 ml of potassium phosphate buffer (0.2M, pH 7.4) containing

DMPO (0.034M) and H2O2 (0.080M). In (D), the intensity of the signals is normalized to the surface area of the analyzed particles. The experiments were

performed twice with similar results.

SINTERED INDIUM-TIN-OXIDE (ITO) PARTICLES 479

by guest on February 6, 2016http://toxsci.oxfordjournals.org/

Dow

nloaded from

occurs directly at the surface of the particles probably via an

electron transfer assisted reductive cleavage (reaction 1) rather

than via homolytic cleavage following reaction with oxygen-

ated radicals (reaction 2):

HCOO� / COO�� þ Hþ þ e� (1)

HCOO� þ HO� / COO�� þ H2O (2)

The defective structure of ITO may account for such

a reactivity, the cleavage of C-H bonds applies to a large

number of biomolecules and may lead to substantial cell

damage such as observed in the inflammatory response induced

in the lung of the treated rats. ITO exhibited also a Fenton

activity higher than its single components or their mixture,

which may also contribute to the overall inflammatory reaction.

Contrary to silica which was cytotoxic to both cell types

examined, ITO was cytotoxic only to macrophages and not

epithelial cells. This might reflect the fact that larger particles

(ITO, 7 lm) are taken up less efficiently than smaller ones

(silica, 2.2 lm) by epithelial cells compared with macro-

phages (Fig. 5). Another possible explanation for this

difference in cell sensitivity might be that macrophages

produce H2O2, notably during phagocytosis (Forman and

Torres, 2002), which might in the presence of ITO particles

participate in a Fenton-like reaction to produce toxic HO�. It is

noteworthy that the cytotoxicity in NR8383 macrophages

(LDH release) paralleled closely the capacity to produce HO�radicals in an acellular system for the different particles

tested, suggesting that HO� are critical for the cytotoxic

activity in macrophages. We did, however, not obtain

evidence for a role of ROS in the cytotoxicity of ITO

particles. Neither Trolox, a water-soluble analogue of

vitamin-E, or BHT, a lipohilic antioxidant mainly active on

peroxyradicals, had a protective effect in NR8383 cells

challenged with ITO particles. Although this observation may

indicate that ROS produced by ITO particles are not directly

involved in its cytotoxicity, this explanation appears very

unlikely in view of the remarkable association between ROS

producing capacity, cytotoxicity in macrophages and in vivoreactivity. An alternative might be that while in hard metals

the superoxide anion is formed at the WC surface, in ITO the

hydroxyl radical is generated only in the presence of H2O2,

thus in a specific compartment, e.g phagolysosomes, that

would not be sufficiently accessible to antioxidants or where

antioxidants loose their activity. It should also be considered

that the cytotoxicity of particles such as silica is not only due

to a radical mechanism, but largely to a membranolytic

potential, which is related to H-bonding properties (Fubini,

1998). Further studies will be necessary to clarify this issue.

ITO particles are virtually insoluble at physiological pH and

we show here that they are not rapidly cleared from the lung.

The presence of ITO particles in the lung was detected up to 2

months after administration. ITO particles induced in our

experimental model a strong, progressive and persistent

inflammatory lung reaction, but apparently no fibrosis.

The apparent absence of TNF-a production in the present

experimental model is intriguing because this cytokine is

generally regarded as a key mediator of both inflammatory and

fibrotic lung reactions induced by other inhaled particles such

as crystalline silica (Huaux, 2007).

The fact that ITO particles (2 mg per rat) were able to induce

a genotoxic response in alveolar epithelial cells is not entirely

unexpected because several other inhaled particles which

induce an inflammatory reaction in the lung, including

crystalline silica (10–100 mg/kg) (Driscoll et al., 1997), hard

metals (16.6 mg/kg) (De Boeck et al., 2003), carbon nanotubes

(2 mg per rat) (Muller et al., 2008) are able to cause genotoxic

damage and/or mutations in alveolar epithelial cells. The

mechanism of this genotoxic activity can be either the result of

a direct genotoxic activity of the particles (primary genotox-

icity) or mediated by inflammatory cells recruited in the lung

by the particles which have the capacity to release ROS and

can cause genotoxic damage in epithelial cells (secondary

genotoxicity) (Schins and Knaapen, 2007). In the present

study, we found evidence of genotoxicity in type II epithelial

cells in vivo but not in vitro, which would suggest that a dose of

ITO particles sufficient to cause inflammation is needed to

induce genotoxic effects.

The coherence of our experimental results with existing

human data should also be critically considered. First, the

pathological features of the lung responses observed in the

present study may appear as not entirely compatible with the

pattern of lesions reported in diseased workers. Indeed, clinico-

epidemiological studies reported epithelial damage as reflected

by increased KL-6, SP-A and SP-D serum levels (Chonan

et al., 2007; Hamaguchi et al., 2008) whereas our in vitro data

indicate that macrophages and not epithelial cells are main

targets of ITO particles. It is, however, not excluded that in

human studies the biochemical changes reflecting epithelial

damage is secondary to an inflammatory process driven by

other pulmonary cells, including macrophages. Also, although

Homma et al. (2003) reported interstitial pneumonia without

apparent fibrosis in a 27-y grinder exposed for 3 years, Homma

et al. (2003) reported a granulomatous lung fibrosis in a 30-y

grinder exposed for 6 years. It is possible that the fibrotic

reaction reported in the latter case reflects a more severe form

or a longer course of the disease. In rats, we did not find

evidence that ITO particles could induce a fibrotic reaction but

this may be the consequence of the single administration

protocol that was used. Further experimental studies with

multiple administrations or, preferably, by inhalation exposure

will be necessary to further explore the capacity of ITO

particles to induce a fibrotic lung reaction.

To conclude, we found that ITO particles represent a new

toxicological entity which has the potential to generate high

amounts of ROS and induce lung toxicity in experimental

animals, including inflammatory and genotoxic effects which

are possibly predictive of a carcinogenic potential. These

results contribute to clarify the cause of the lung diseases

480 LISON ET AL.

by guest on February 6, 2016http://toxsci.oxfordjournals.org/

Dow

nloaded from

reported in ITO workers. Adequate industrial hygiene measures

should be implemented to minimise human exposure to ITO

particles in relevant occupational settings. Our results indicate

that ITO particles are almost as or even more toxic than quartz

when considering the surface area doses. This might suggest

that occupational exposure levels for ITO particles might need

to be at least as low as for crystalline silica. Further in vivostudies, possibly by inhalation, and using particles with strictly

similar dimensions, will be needed to further substantiate this

comparison.

ACKNOWLEDGMENTS

The authors acknowledge a grant from Umicore (Brussels,

Belgium) to perform this research; the funding organization did

not have control over the resulting publication. We declare that

there are no conflict of interest.

REFERENCES

Biondi, P. A., Chiesa, L. M., Storelli, M. R., and Renon, P. (1997). A new

procedure for the specific high-performance liquid chromatographic de-

termination of hydroxyproline. J. Chromatogr. Sci. 35, 509–512.

Burkitt, M. J., Tsang, S. Y., Tam, S. C., and Bremner, I. (1995). Generation of

5,5-dimethyl-1-pyrroline N-oxide hydroxyl and scavenger radical adducts

from copper/H2O2 mixtures: Effects of metal ion chelation and the search

for high-valent metal-oxygen intermediates. Arch. Biochem. Biophys. 323,

63–70.

Chonan, T., Taguchi, O., and Omae, K. (2007). Interstitial pulmonary disorders

in indium-processing workers. Eur. Respir. J. 29, 317–324.

De Boeck, M., Hoet, P., Lombaert, N., Nemery, B., Kirsch-Volders, M., and

Lison, D. (2003). In vivo genotoxicity of hard metal dust: Induction of

micronuclei in rat type II epithelial lung cells. Carcinogenesis 24,

1793–1800.

Dobbs, L. G., Gonzalez, R., and Williams, M. C. (1986). An improved method

for isolating type II cells in high yield and purity. Am. Rev. Respir. Dis. 134,

141–145.

Driscoll, K. E., Deyo, L. C., Carter, J. M., Howard, B. W., Hassenbein, D. G.,

and Bertram, T. A. (1997). Effects of particle exposure and particle-elicited

inflammatory cells on mutation in rat alveolar epithelial cells. Carcinogen-

esis 18, 423–430.

Fan, J. C. C., and Goodenough, J. B. (2008). X-ray photoemission spectroscopy

studies of Sn-doped indium-oxide films. J. Appl. Phys. 48, 3524–3531.

Finkelstein, E., Rosen, G. M., and Rauckman, E. J. (1982). Production of

hydroxyl radical by decomposition of superoxide spin-trapped adducts. Mol.

Pharmacol. 21, 262–265.

Forman, H. J., and Torres, M. (2002). Reactive oxygen species and cell

signaling: Respiratory burst in macrophage signaling. Am. J. Respir. Crit.

Care Med. 166, S4–S8.

Fubini, B. (1998). Surface chemistry and quartz hazard. Ann. Occup. Hyg. 42,

521–530.

Gottschling, B. C., Maronpot, R. R., Hailey, J. R., Peddada, S.,

Moomaw, C. R., Klaunig, J. E., and Nyska, A. (2001). The role of oxidative

stress in indium phosphide-induced lung carcinogenesis in rats. Toxicol. Sci.

64, 28–40.

Hamaguchi, T., Omae, K., Takebayashi, T., Kikuchi, Y., Yoshioka, N.,

Nishiwaki, Y., Tanaka, A., Hirata, M., Taguchi, O., and Chonan, T. (2008).

Exposure to hardly soluble indium compounds in ITO production and

recycling plants is a new risk for interstitial lung damage. Occup. Environ.

Med. 65, 51–55.

Hanna, P. M., Chamulitrat, W., and Mason, R. P. (1992). When are metal ion-

dependent hydroxyl and alkoxyl radical adducts of 5,5-dimethyl-1-pyrroline

N-oxide artifacts? Arch. Biochem. Biophys. 296, 640–644.

Homma, S., Miyamoto, A., Sakamoto, S., Kishi, K., Motoi, N., and

Yoshimura, K. (2005). Pulmonary fibrosis in an individual occupationally

exposed to inhaled indium-tin oxide. Eur. Respir. J. 25, 200–204.

Homma, T., Ueno, T., Sekizawa, K., Tanaka, A., and Hirata, M. (2003).

Interstitial pneumonia developed in a worker dealing with particles

containing indium-tin oxide. J. Occup. Health 45, 137–139.

Huaux, F. (2007). New developments in the understanding of immunology in

silicosis. Curr. Opin. Allergy Clin. Immunol. 7, 168–173.

Lasfargues, G., Lardot, C., Delos, M., Lauwerys, R., and Lison, D. (1995). The

delayed lung responses to single and repeated intratracheal administration of

pure cobalt and hard metal powder in the rat. Environ. Res. 69, 108–121.

Lison, D., Carbonnelle, P., Mollo, L., Lauwerys, R., and Fubini, B. (1995).

Physicochemical mechanism of the interaction between cobalt metal and

carbide particles to generate toxic activated oxygen species. Chem. Res.

Toxicol. 8, 600–606.

Lison, D., and Lauwerys, R. (1990). In vitro cytotoxic effects of cobalt-

containing dusts on mouse peritoneal and rat alveolar macrophages. Environ.

Res. 52, 187–198.

Morgan, D. L., Shines, C. J., Jeter, S. P., Blazka, M. E., Elwell, M. R.,

Wilson, R. E., Ward, S. M., Price, H. C., and Moskowitz, P. D. (1997).

Comparative pulmonary absorption, distribution, and toxicity of copper

gallium diselenide, copper indium diselenide, and cadmium telluride in

Sprague-Dawley rats. Toxicol. Appl. Pharmacol. 147, 399–410.

Muller, J., Decordier, I., Hoet, P. H., Lombaert, N., Thomassen, L., Huaux, F.,

Lison, D., and Kirsch-Volders, M. (2008). Clastogenic and aneugenic effects of

multi-wall carbon nanotubes in epithelial cells. Carcinogenesis 29, 427–433.

Oda, K. (1997). Toxicity of a low level of indium phosphide (InP) in rats after

intratracheal instillation. Ind. Health 35, 61–68.

Rao, G. V., Tinkle, S., Weissman, D. N., Antonini, J. M., Kashon, M. L.,

Salmen, R., Battelli, L. A., Willard, P. A., Hoover, M. D., and Hubbs, A. F.

(2003). Efficacy of a technique for exposing the mouse lung to particles

aspirated from the pharynx. J. Toxicol. Environ. Health A 66, 1441–1452.

Schins, R. P., and Knaapen, A. M. (2007). Genotoxicity of poorly soluble

particles. Inhal. Toxicol. 19(Suppl. 1), 189–198.

Tanaka, A., Hirata, M., Omura, M., Inoue, N., Oueno, T., Homma, T., and

Sekizawa, K. (2002). Pulmonary toxicity of indium-tin oxide and indium

phosphide after intratracheal instillation into the lung of hamsters. J. Occup.Health 44, 99–102.

Tanaka, A., Hisanaga, A., Hirata, M., Omura, M., Makita, Y., Inoue, N., and

Ishinishi, N. (1996). Chronic toxicity of indium arsenide and indium

phosphide to the lungs of hamsters. Fukuoka Igaku Zasshi 87, 108–115.

van den Brule, S., Heymans, J., Havaux, X., Renauld, J. C., Lison, D.,

Huaux, F., and Denis, O. (2007). Profibrotic effect of IL-9 overexpression

in a model of airway remodeling. Am. J. Respir. Cell Mol. Biol. 37, 202–209.

SINTERED INDIUM-TIN-OXIDE (ITO) PARTICLES 481

by guest on February 6, 2016http://toxsci.oxfordjournals.org/

Dow

nloaded from