interference of engineered nanoparticles with in vitro toxicity assays

14
INORGANIC COMPOUNDS Interference of engineered nanoparticles with in vitro toxicity assays Alexandra Kroll Mike Hendrik Pillukat Daniela Hahn Ju ¨ rgen Schnekenburger Received: 14 October 2011 / Accepted: 1 March 2012 / Published online: 11 March 2012 Ó Springer-Verlag 2012 Abstract Accurate in vitro assessment of nanoparticle cytotoxicity requires a careful selection of the test systems. Due to high adsorption capacity and optical activity, engineered nanoparticles are highly potential in influencing classical cytotoxicity assays. Here, four common in vitro assays for oxidative stress, cell viability, cell death and inflammatory cytokine production (DCF, MTT, LDH and IL-8 ELISA) were assessed for validity using 24 well- characterized engineered nanoparticles. For all nanoparti- cles, the possible interference with the optical detection methods, the ability to convert the substrates, the influence on enzymatic activity and the potential to bind proinflam- matory cytokines were analyzed in detail. Results varied considerably depending on the assay system used. All nanoparticles tested were found to interfere with the optical measurement at concentrations of 50 lg cm -2 and above when DCF, MTT and LDH assays were performed. Except for Carbon Black, particle interference could be prevented by altering assay protocols and lowering particle concen- trations to 10 lg cm -2 . Carbon Black was also found to oxidize H 2 DCF-DA in a cell-free system, whereas only ZnO nanoparticles significantly decreased LDH activity. A dramatic loss of immunoreactive IL-8 was observed for only one of the three TiO 2 particle types tested. Our results demonstrate that engineered nanoparticles interfere with classic cytotoxicity assays in a highly concentration-, par- ticle- and assay-specific manner. These findings strongly suggest that each in vitro test system has to be evaluated for each single nanoparticle type to accurately assess the nanoparticle toxicity. Keywords Engineered nanoparticles Cytotoxicity assays Interference Cytokine adsorption Abbreviations H 2 DCF-DA 2 0 ,7 0 -Dichlorodihydrofluorescein diacetate DCF 2 0 ,7 0 -Dichlorofluorescein MTT (3-(4,5-Dimethylthiazol-2-yl)-2,5- diphenyltetrazoliumbromide) INT (2-(4-Iodophenyl)-3-(4-nitrophenyl)-5- phenyl tetrazolium chloride) MTS 3-(4,5-Dimethylthiazole-2-yl)-5- (3-carboxymethoxyphenyl)-2- (4-sulfophenyl)-2H-tetrazoliumin Introduction The increased global production of engineered nanoparti- cles (NPs) and their application in various fields require a detailed understanding of their potential toxicity. In vitro and in vivo studies demonstrated that NPs may display a higher toxic potential than fine-sized particles of identical This article is published as a part of the Special Issue ‘‘Nanotoxicology II’’ on the ECETOC Satellite workshop, Dresden 2010 (Innovation through Nanotechnology and Nanomaterials ? Current Aspects of Safety Assessment and Regulation). Alexandra Kroll and Mike Hendrik Pillukat contributed equally to this article. A. Kroll M. H. Pillukat D. Hahn J. Schnekenburger (&) Biomedical Technology Center, Westfa ¨lische Wilhelms-Universita ¨t, Albert-Schweitzer-Campus 1 A14, 48149 Mu ¨nster, Germany e-mail: [email protected] Present Address: A. Kroll Eawag, Swiss Federal Institute of Aquatic Science and Technology, Du ¨bendorf, Switzerland 123 Arch Toxicol (2012) 86:1123–1136 DOI 10.1007/s00204-012-0837-z

Upload: juergen-schnekenburger

Post on 19-Aug-2016

214 views

Category:

Documents


1 download

TRANSCRIPT

INORGANIC COMPOUNDS

Interference of engineered nanoparticles with in vitrotoxicity assays

Alexandra Kroll • Mike Hendrik Pillukat •

Daniela Hahn • Jurgen Schnekenburger

Received: 14 October 2011 / Accepted: 1 March 2012 / Published online: 11 March 2012

� Springer-Verlag 2012

Abstract Accurate in vitro assessment of nanoparticle

cytotoxicity requires a careful selection of the test systems.

Due to high adsorption capacity and optical activity,

engineered nanoparticles are highly potential in influencing

classical cytotoxicity assays. Here, four common in vitro

assays for oxidative stress, cell viability, cell death and

inflammatory cytokine production (DCF, MTT, LDH and

IL-8 ELISA) were assessed for validity using 24 well-

characterized engineered nanoparticles. For all nanoparti-

cles, the possible interference with the optical detection

methods, the ability to convert the substrates, the influence

on enzymatic activity and the potential to bind proinflam-

matory cytokines were analyzed in detail. Results varied

considerably depending on the assay system used. All

nanoparticles tested were found to interfere with the optical

measurement at concentrations of 50 lg cm-2 and above

when DCF, MTT and LDH assays were performed. Except

for Carbon Black, particle interference could be prevented

by altering assay protocols and lowering particle concen-

trations to 10 lg cm-2. Carbon Black was also found to

oxidize H2DCF-DA in a cell-free system, whereas only

ZnO nanoparticles significantly decreased LDH activity. A

dramatic loss of immunoreactive IL-8 was observed for

only one of the three TiO2 particle types tested. Our results

demonstrate that engineered nanoparticles interfere with

classic cytotoxicity assays in a highly concentration-, par-

ticle- and assay-specific manner. These findings strongly

suggest that each in vitro test system has to be evaluated for

each single nanoparticle type to accurately assess the

nanoparticle toxicity.

Keywords Engineered nanoparticles � Cytotoxicity

assays � Interference � Cytokine adsorption

Abbreviations

H2DCF-DA 20,70-Dichlorodihydrofluorescein diacetate

DCF 20,70-Dichlorofluorescein

MTT (3-(4,5-Dimethylthiazol-2-yl)-2,5-

diphenyltetrazoliumbromide)

INT (2-(4-Iodophenyl)-3-(4-nitrophenyl)-5-

phenyl tetrazolium chloride)

MTS 3-(4,5-Dimethylthiazole-2-yl)-5-

(3-carboxymethoxyphenyl)-2-

(4-sulfophenyl)-2H-tetrazoliumin

Introduction

The increased global production of engineered nanoparti-

cles (NPs) and their application in various fields require a

detailed understanding of their potential toxicity. In vitro

and in vivo studies demonstrated that NPs may display a

higher toxic potential than fine-sized particles of identical

This article is published as a part of the Special Issue

‘‘Nanotoxicology II’’ on the ECETOC Satellite workshop,

Dresden 2010 (Innovation through Nanotechnology and

Nanomaterials ? Current Aspects of Safety Assessment and

Regulation).

Alexandra Kroll and Mike Hendrik Pillukat contributed equally to

this article.

A. Kroll � M. H. Pillukat � D. Hahn � J. Schnekenburger (&)

Biomedical Technology Center, Westfalische

Wilhelms-Universitat, Albert-Schweitzer-Campus 1 A14,

48149 Munster, Germany

e-mail: [email protected]

Present Address:A. Kroll

Eawag, Swiss Federal Institute of Aquatic Science

and Technology, Dubendorf, Switzerland

123

Arch Toxicol (2012) 86:1123–1136

DOI 10.1007/s00204-012-0837-z

composition (Duffin et al. 2007; Landsiedel et al. 2010;

Oberdorster et al. 2005; Singh and Nalwa 2007). Apart

from size and chemical composition, a variety of other

physicochemical properties including crystallinity (Sayes

et al. 2006; Wang et al. 2008), agglomeration state

(Grassian et al. 2007), surface coating (Cho et al. 2007; Pan

et al. 2009), shape (Rothen-Rutishauser et al. 2010) and

microporosity (Rabolli et al. 2010) have recently been

identified as critical determinants of NP toxicity. More-

over, evidence is accumulating that NP toxicity is driven

by interplay of several of these properties and is difficult to

predict a priori (Johnston et al. 2009; Rabolli et al. 2010;

Shaw et al. 2008, Kroll et al. 2011).

Evaluating NP toxicity therefore requires a testing

strategy that includes a detailed physicochemical charac-

terization of each individual particle type as well as a set of

in vitro assays that allow for a rapid screening of the vast

number of existing and newly developed NP variants.

Currently, however, standardized in vitro techniques to

accurately assess NP toxicity have not yet been established.

Recent studies suggest that classical cell-based assays

designed for testing chemicals may not provide reliable

data since nanomaterials can interfere with assay compo-

nents and readout systems (Kocbach et al. 2008; Kroll et al.

2009; Monteiro-Riviere et al. 2009).

Due to their large surface area and high surface energy,

nanoparticles can adsorb assay reagents or reporter

dyes, thereby distorting the assay outcome. For carbon

nanomaterials, such as single-walled carbon nanotubes

(SWCNTs) or Carbon Black (CB), adsorption of MTT-

formazan, a reporter dye that indicates metabolic activity in

MTT assays, has been described by different groups of

investigators (Casey et al. 2007; Monteiro-Riviere and

Inman 2006; Monteiro-Riviere et al. 2009; Worle-Knirsch

et al. 2006). Additionally, SWCNTs seem to adsorb

essential nutrients from cell culture media resulting in

indirect cytotoxic effects (Casey et al. 2008; Guo et al.

2008). In both cases, the high adsorption capacity of carbon

nanomaterials generates false-positive data resulting in an

overestimation of their toxicity; notably, CB also seems to

adsorb proinflammatory cytokines leading to false-negative

results in some immunoassays and, consequently, to an

underestimation of its potential toxicity (Brown et al. 2010;

Kocbach et al. 2008). This demonstrates NP interference

with assay systems as cause for contradictory testing results.

Verifying cytotoxicity data with at least two or more

independent test systems as it has been suggested earlier

(Monteiro-Riviere et al. 2009) may thus not be sufficient to

eliminate misleading results and incorrect interpretations.

Recent studies indicate that also metal oxide NPs

interact with assay components or reporter dyes of in vitro

test systems. Ultrafine superparamagnetic iron oxide NPs

seem to interfere with MTS-based cell viability assays

(Doak et al. 2009), a confounding factor that results in

an underestimation of their potential toxicity. Moreover,

superparamagnetic iron oxide NPs can cause significant

changes in the cell culture medium such as depletion of

Cl- ions or denaturation of proteins that, in turn, can result

in toxic effects on cells (Mahmoudi et al. 2010). Finally,

Veranth et al. have found TiO2 and SiO2 to cause artifacts

in immunoassays by adsorbing the proinflammatory cyto-

kine IL-6 (Veranth et al. 2007).

Additional challenges for NP cytotoxicity screening

arise from the intrinsic optical absorbance of NP disper-

sions. Because most of the classic in vitro toxicity assays

are based on optical readouts, the high absorption or

scattering of nanoparticles may interfere with the detection

method. Carbon nanomaterials are known to absorb light of

the visible spectrum (Wells and Smith 1941) and may

therefore generate an absorbance at the same wavelength as

that used to detect the colored assay product. Metal oxide

NPs frequently found in sunscreens, such as TiO2 and ZnO,

absorb and scatter UV radiation as well as visible light

(Wolf et al. 2001) and may possibly influence absorbance

readouts of colorimetric assays. In fact, TiO2 NPs have

been recently shown to change the color and increase the

opalescence of experimental resin matrices (Yu et al.

2009). However, the influence of NPs on optical detection

is difficult to estimate in advance, as optical activity does

not only depend on the chemical composition but also on

other properties such as particle size, shape and crystal-

linity as well as on the dispersion agent. A number of NPs

also seem to interact directly with fluorescent dyes, thereby

altering their emission intensity. Silver NPs, for instance,

have been shown to quench fluorescent dyes presumably by

binding them onto their surface (Sabatini et al. 2007). Gold

nanoparticles have been found to quench DCF fluorescence

upon NP-induced oxidative stress (Pfaller et al. 2010).

Dextran-coated iron oxide NPs displayed interference with

the fluorescence emission of DCFH. The interference was

more pronounced with Fe3O4 than with Fe2O3 NPs, dem-

onstrating that even differences in the oxidation state can

impact the test results (Doak et al. 2009). Testing NPs with

established assays is therefore challenging and requires a

careful evaluation of the test systems applied.

In this study, we utilized a set of 24 well-characterized

engineered nanoparticles and assessed their possible

interference with classical cytotoxicity assays. We selected

frequently used in vitro assays that determine four different

cytotoxic endpoints (oxidative stress (DCF), metabolic

activity (MTT), cell death (LDH) and immune responses

(IL-8 ELISA)) and evaluated their suitability for measuring

toxic effects of NPs on cells. We developed standardized

protocols for the experimental setup including the prepa-

ration of NP dispersions, exposure conditions and test

procedures. For each of the 24 NPs, we tested its possible

1124 Arch Toxicol (2012) 86:1123–1136

123

interference with the optical detection methods, its ability

to convert the substrate, its influence on the lactate dehy-

drogenase activity as well as its adsorption of cytokines.

The NPs represented inorganic compounds manufactured

for a wide range of industrial products including three

different TiO2 particle types and, additionally, CB. Within

the framework of the German project NanoCare, the NPs

had been previously characterized in detail regarding their

physicochemical properties (Landsiedel et al. 2010; Nano-

Care Project 2009).

Materials and methods

Nanoparticles

The engineered NPs were part of the BMBF project

NanoCare materials and were thoroughly characterized

regarding relevant physicochemical properties such as size

distribution, surface area, crystallinity, crystal phase and

solubility (Landsiedel et al. 2010; NanoCare Project 2009).

Crystalline quartz particles (DQ12, mean particle size of

5 lm) were kindly provided by Evonik Degussa GmbH

(Marl, Germany), and fine-sized ZnO particles (mean

particle size of 350 nm) were purchased from Sigma-

Aldrich Chemie GmbH (Munich, Germany).

Chemicals

If not otherwise stated, all chemicals were obtained from

Sigma-Aldrich Chemie GmbH (Munich, Germany). 20,70-Dichlorodihydrofluorescein diacetate (H2DCF-DA) was

obtained from Invitrogen GmbH (Karlsruhe, Germany).

Sodium bicarbonate and sodium pyruvate were from Lonza

AG (Koln, Germany). Dulbecco’s modified Eagle’s med-

ium (DMEM, E15-877), RPMI 1640 (E15-848) and FBS

Gold (A15-151) were supplied by PAA Laboratories

GmbH (Pasching, Germany). Phenol red–free MEM was

obtained from Biochrome AG (Berlin, Germany).

Cell lines and culture conditions

A549 cells were obtained from the American Type Culture

Collection (ATCC, Rockville, USA). Their authenticity

was determined by DSMZ (German Collection of Micro-

organisms and Cell Cultures) via DNA fingerprinting.

Cultured cells were propagated at 37 �C and 5 % CO2.

Cells were grown to confluence within three to 4 days in

DMEM supplemented with 10 % v/v heat-inactivated fetal

calf serum (FBS Gold) and 2 mM L-glutamine and then

transferred to new culture plates at a dilution of 1:6. Cells

were cultured for a maximum of 20 passages with regular

testing for mycoplasma contamination by PCR. To assure

constant cell quality, the cell cycle state and metabolic

activity of the cell cultures were monitored prior to each

experiment with propidium iodide staining followed by

flow cytometry or with the MTT assay. To prepare cell

monolayers for testing NP interference, cells were seeded

in 96-well plates at a density of 3 9 104 cells per well in a

total volume of 100 ll and incubated at 37 �C, 5 % CO2

for 24 h. To avoid any influence of the possible photocat-

alytic activity of the NPs, DCF, MTT and LDH assays

were performed in the absence of UV light.

For immunoassays, cells were seeded at a density of

2 9 104 cells per well in a total volume of 100 ll and

cultured as described above. After centrifugation (15 min,

5009g, RT), supernatants were removed from the assay

plates, diluted 20-fold in DMEM/10 % FBS Gold and

stored at -20 �C until assayed by immunoassays.

Sterilization and dispersion of nanoparticles in cell

culture medium

NPs were sterilized and dispersed according to a stan-

dardized protocol as described previously (Schulze et al.

2008). In particular, NP aliquots of 19.2 mg in powder

form or predispersed as supplied by the manufacturer were

transferred into sterile snap-lid glasses together with a

small magnetic stir bar and subsequently exposed to 30-Gy

gamma irradiation. The dispersions were stirred at 900 rpm

for 1 h at room temperature in 6 ml of DMEM/10 % FBS.

Dilutions of this stock dispersion were prepared immedi-

ately and stirred for 24 h at 900 rpm at room temperature.

All NP dispersions (10 lg cm-2 in DMEM/10 % FBS)

were tested to be free of endotoxins using the Limulus

Amebocyte Lysate (LAL) Kinetic-QCL� kit (50-650U),

Lonza GmbH, Wuppertal, Germany.

Interference with the detection of oxidative stress

The formation of reactive oxygen species (ROS) as indi-

cators of oxidative stress is commonly detected using

the fluorescein derivative H2DCF-DA (H2DCF-DA: 20,70-dichlorodihydrofluorescein diacetate; Jakubowski and

Bartosz 2000). H2DCF-DA penetrates cell membranes, is

hydrolyzed by cellular esterases and is converted via

reactive oxygen species (ROS) to the fluorescent oxidation

product DCF. The interference of NPs with the optical

detection of DCF fluorescence was assessed by replacing

the assay substrate H2DCF-DA by defined amounts of

fluorescent DCF. Cell monolayers prepared as described

above were incubated for 1 h at 37 �C with 100 ll NP

dispersions at concentrations of 0.01, 0.1, 1, 10, 50 or

100 lg cm-2. As control, stirred cell culture medium

(100 ll per well) was used to replace the NP dispersions.

Different concentrations of DCF (100, 50, 10, 1, 0.1 or 0.01

Arch Toxicol (2012) 86:1123–1136 1125

123

nM) diluted in Krebs–Ringer buffer (KRB; pH 7.4,

114.2 mM NaCl, 3 mM KCl, 1.5 mM K2HPO4 9 3 H2O,

10 mM HEPES, 3.996 mM D-glucose, 1.405 mM CaCl2,

2.562 mM MgCl2) were added either directly or after the

cells had been washed with KRB. Fluorescence was

immediately monitored in a spectrophotometer (excitation

485 nm, emission 520 nm; NOVOstar and FLUOstar,

BMG Labtech GmbH, Offenburg, Germany). Each exper-

iment was repeated at least seven times with seven

replications.

To test the catalytic activity of NP dispersions in terms

of H2DCF-DA oxidation, empty 96-well plates were

incubated with NP dispersions (10, 5, 1, 0.1 or 0.01 lg

cm-2) or stirred cell culture medium (control) and incu-

bated for 1 h at 37 �C. After washing with 100 ll KRB, the

H2DCF-DA assay solution was added and incubated for

1 h at 37 �C. Subsequently, the plates were washed twice

with KRB, and fluorescence was recorded twice, immedi-

ately and after an additional incubation for 3 h at 37 �C.

Each experiment was repeated at least three times with

seven replications.

Interference with the detection of metabolic activity

The MTT assay determines cellular metabolic activity by

the reduction of the yellow tetrazolium salt MTT (3-(4,5-

dimethylthiazol-2-yl)-2,5-diphenyltetrazoliumbromide) to

a purple MTT-formazan. The interference of NP disper-

sions with the optical detection of MTT-formazan was

assessed using different mixtures of the reaction product

MTT-formazan (1.2, 0.8, 0.24, 0.024 and 0.0024 mM)

with the substrate (2.4 mM oxidized MTT) to cover the

whole spectrum of light absorption. MTT-formazan was

obtained by incubating A549 cell monolayers with oxi-

dized MTT. In particular, cell monolayers grown as

described above were washed twice with 100 ll PBS and

incubated with 2.4 mM oxidized MTT in 110 ll PBS for

2 h. To dissolve the MTT-formazan crystals, 100 ll of

15 % SDS/50 % DMSO was added; 96-well plates were

shaken at 300 rpm overnight at 37 �C, and the superna-

tants were pooled for further experiments. To verify that

MTT had been fully reduced, cell supernatants were

recovered after incubation with oxidized MTT and added

to fresh cell monolayers. Different NP concentrations of

0.01, 0.1, 1, 10, 50 and 100 lg cm-2 in dispersion were

added to cell monolayers (see above) and incubated for

24 h at 37 �C. Stirred cell culture medium (100 ll per

well) was used as control. Mixtures of reduced and oxi-

dized MTT (100 ll per well) were directly added either

after NP incubation or after the cells were washed with

KRB. Light absorption was measured in a spectropho-

tometer (excitation 485 nm, emission 520 nm; NOVOstar

and FLUOstar, BMG Labtech GmbH; Offenburg,

Germany) at 550 nm and for reference at 670 nm. Each

experiment was repeated three times with seven

replications.

The influence of NP dispersions on the reduction of

MTT was determined in empty wells incubated with NP

dispersions (10, 5, 1, 0.1 and 0.01 lg cm-2) or stirred cell

culture medium (control) and washed with KRB The

standard MTT solution (2.4 mM MTT) was added and

incubated for 3 h at 37 �C. To dissolve the MTT-formazan

crystals, 100 ll of 15 % SDS/50 % DMSO was added and

shaken overnight at 37 �C. Light absorption was measured

at 550 nm. Each experiment was repeated three times with

seven replications.

Interference with the measurement of cell death

Cell death is usually quantified by the measurement of

lactate dehydrogenase (LDH) activity in cell supernatants.

LDH reduces INT (2-(4-iodophenyl)-3-(4-nitrophenyl)-5-

phenyl tetrazolium chloride) in the presence of NADH ?

H? (reduced b-nicotinamide adenine dinucleotide) yielding

a red water-soluble formazan, which is quantified by the

measurement of light absorption (Nachlas et al. 1960). The

interference of NP dispersions with the optical detection of

INT reduction was analyzed with different mixtures of

reduced and oxidized INT. Reduced INT was obtained by

incubating supernatants of A549 cell monolayers prepared

as described above with oxidized INT according to Kor-

zeniewski and Callewaert (Korzeniewski and Callewaert

1983). To verify that INT had been entirely reduced, part of

the reaction product was incubated according to the same

protocol with fresh cell supernatant and reaction mixture

but without oxidized INT. Cell monolayers prepared as

described above were then exposed to NP dispersions at

concentrations of 0.01, 0.1, 1, 10, 50 and 100 lg cm-2 for

24 h at 37 �C. Stirred cell culture medium (100 ll per

well) was used as control. Supernatants (50 ll) were

transferred to empty 96-well plates and mixtures of

reduced INT (1.75, 1.16, 0.35, 0.035 and 0.0035 mM,

respectively) with 3.5 mM oxidized INT were added. The

light absorption at 492 nm was monitored immediately.

Each experiment was repeated three times with seven

replications.

The catalytic activity of NPs was tested in cell-free

wells pretreated with NP dispersions (10, 5, 1, 0.1 and

0.01 lg cm-2, respectively) or with stirred cell culture

medium (control). Supernatants (50 ll) were mixed with

50 ll of the LDH assay solution (3.5 g l-1 INT,

0.089 g l-1 PMS and 0.9 g l-1 NAD in 0.2 M Tris-Base

pH 8.2 and 5 g l-1 lactic acid). Light absorption was

measured continuously for 1 h at 37 �C in a spectropho-

tometer. Each experiment was repeated three times with

seven replications.

1126 Arch Toxicol (2012) 86:1123–1136

123

Bovine LDH (Sigma-Aldrich Chemie GmbH, Munich,

Germany) was used to determine the influence of NP dis-

persions on LDH activity. NP dispersions or stirred cell

culture medium (control) were mixed with 0.3 U ml-1 LDH

in cell-free wells (100 ll per well) and incubated for 24 h at

37 �C. Supernatants (50 ll) were transferred to new 96-well

plates, and 50 ll of the LDH assay solution was added. Light

absorption at 492 nm was recorded continuously for 1 h at

37 �C in a spectrophotometer. Each experiment was repe-

ated three times with seven replications.

Determination of IL-8 with the enzyme-linked

immunosorbent assay

A549 cells were seeded on 96-well plates and incubated

with NP dispersions (1, 10 and 50 lg cm-2, respectively),

dispersed quartz DQ12 as positive and stirred cell culture

medium as negative control. Cytokine levels in the diluted

samples were determined using the Human IL–8 Opt-

EIATM–ELISA-Kit (BD Biosciences; Heidelberg, Germany)

according to the manufacturer’s protocol with minor

modifications: Dilution of samples (1:20 for the determi-

nation of IL-8 in supernatants from cell cultures, undiluted

for samples from cell-free assays) and of protein standard

was carried out in DMEM/10 % FBS Gold (PAA Labo-

ratories GmbH; Pasching, Germany) and not in dilution

buffer (PBS/10 % FBS Gold). The IL-8 protein standards

were diluted from 300 pg ml-1 down to 4.69 pg ml-1. The

ELISA plates were washed by an automated Opsys Plate

Washer MW (Dynex Technologies Inc., Chantilly, VA,

USA). Absorbance was measured at 450 nm with the plate

reader FLUOstar OPTIMA (BMG Labtech GmbH; Off-

enburg, Germany). Each experiment was repeated twice

with three replications.

Interference of NPs with immunoassays

Cell-free experiments were carried out with different con-

centrations of recombinant IL-8 protein standard (ranging

from 4.69 to 300 pg ml-1, from the Human IL-8 Opt-

EIATM–ELISA-Kit) instead of cells and without medium

exchange before the addition of nanoparticles (50 lg

cm-2) or stirred control medium, respectively. After

incubation and centrifugation (15 min, 5009g, RT), the

supernatants were removed from the assay plates, and IL-8

concentration was determined by ELISA. Two independent

experiments were performed with each of the 24 nano-

particles and controls.

Selected NPs (TiO2 1–3 NPs, CB) were tested for

cytokine adsorption at three concentrations (1, 10 and

50 lg cm-2) and with four different concentrations of

recombinant IL-8 (18.25–150 pg ml-1). These experi-

ments were performed twice with two replications each.

Calculations and statistical analyses

Results are presented as mean values with standard devi-

ations. Statistical significance was compared to the

respective on-plate negative control using Student’s t test

or ANOVA according to Tukey for p = 0.05.

Results

For the analysis of NP interference with common in vitro

toxicity test systems, we assessed 24 engineered NPs using

cytotoxicity assays monitoring oxidative stress (DCF), cell

death (LDH), cellular metabolic activity (MTT) and the

release of the proinflammatory cytokine IL-8 (ELISA). All

experiments were performed with A549 cells that are

commonly used as an in vitro model for lung toxicity.

These cells display features of type II lung alveolar epi-

thelial cells (Stearns et al. 2001). The NPs used in this

study included particles with the same chemical composi-

tion but with different physical properties, for example,

three types of TiO2 or ZrO2 NPs (NanoCare Project 2009).

Fine-sized ZnO (FZO) and crystalline quartz (DQ12) par-

ticles were used as reference materials. Dispersions of

nanoparticles were prepared in a standardized procedure as

previously described (Schulze et al. 2008). Moreover, each

NP dispersion was tested to be free of lipopolysaccharides,

that is, endotoxins that induce inflammatory responses or

oxidative stress in cultured cells (Esch et al. 2010; Schulze

et al. 2008).

To obtain a more comprehensive view of the possible

influences of NPs on the test systems, we used the complete

set of 24 engineered NPs and tested each NP for optical

interference, for interaction with the substrates, for the

inhibition of enzymes (LDH) and for adsorption of cyto-

kines (IL-8) from cell supernatants.

Optical interference

The fluorescence-based DCF assay determines the forma-

tion of intracellular reactive oxygen species (ROS) via

oxidation of the non-fluorescent substrate H2DCF-DA to

the fluorescent DCF product. To investigate the interfer-

ence of NP dispersions with the fluorescence measurement

of DCF, we used dilutions of fluorescent DCF covering the

full detection range of the spectrofluorometer and deter-

mined their fluorescence emission at 520 nm in medium

or in the presence of NP dispersions (0.1, 1, 10 and

50 lg cm-2) by following the assay protocols (for details,

see Materials and methods). All NP types tested interfered

with the detection of DCF fluorescence at concentrations of

10 lg cm-2 (32 lg ml-1) and above with CB displaying

the strongest influence observed. Interference of CB and

Arch Toxicol (2012) 86:1123–1136 1127

123

TiO2 3 NPs with the measurement of DCF fluorescence is

shown exemplarily in Fig. 1a, b. DCF fluorescence was

significantly reduced in the presence of 1–50 lg cm-2 NP

dispersions of CB (Fig. 1a) and Al–Ti–Zr 3 (data not

shown), irrespective of the DCF concentrations applied. NP

dispersions of TiO2 3 and all other NPs tested displayed

interference at particle concentrations of 10–50 lg cm-2

(see Fig. 1b as an example).

To minimize particle interference, the classical DCF

assay protocol was modified and NP dispersions were

removed by washing the cell monolayers prior to the

incubation with DCF (for details, see Materials and

methods). These assay adaptations nearly eliminated DCF

signal alteration when particle concentrations of

1 lg cm-2 (CB, Fig. 1c) or 10 lg cm-2 and below were

used (TiO2 3, Fig. 1d and all other NPs tested, data not

shown).

The MTT assay is a dye-based assay that determines

cellular metabolic activity by colorimetric quantification.

In metabolically active cells, the yellow tetrazolium

substrate MTT is reduced to a purple insoluble MTT-for-

mazan and results are obtained by quantitation of colori-

metric absorption. We assessed the interference of the 24

engineered NPs with the measurement of MTT and MTT-

formazan. To this end, mixtures of oxidized MTT and

MTT-formazan that covered the whole spectrum of light

absorption from the yellow reactant to the purple product

were prepared and measured in the presence or absence of

NP dispersions with different particle concentrations. In

Fig. 2a, b, MTT light absorption in the presence of CB and

TiO2 3 is shown exemplarily. Light absorption of each

MTT/MTT-formazan mixture was strongly distorted in the

presence of all nanoparticle dispersions tested when parti-

cle concentrations of 10 lg cm-2 and above were used

(Fig. 2a, b). The effect was most pronounced in mixtures

containing higher concentrations of MTT-formazan (0.8

and 1.2 mM). We therefore modified the MTT assay pro-

tocols and applied the NP dispersions for 24 h but washed

them from the cell monolayer prior to the incubation with

MTT. CB still interfered with light absorption of MTT at

Fig. 1 Interference of CB (a, c) and TiO2 3 (b, d) NPs with DCF

fluorescence. Fluorescence of five different DCF concentrations was

measured in the presence of A549 cells pre-incubated for 1 h with

dispersions of four different NP concentrations in DMEM/10 % FBS.

Serum-containing medium (DMEM/10 % FBS) was used as control.

In a and b, the experimental procedure followed the classic DCF

protocol; in c and d, an additional washing step was included before

the addition of DCF. Mean values with standard deviations were

obtained from seven independent experiments with seven replications

each

1128 Arch Toxicol (2012) 86:1123–1136

123

10 lg cm-2 and above (Fig. 2c), but for all other NPs

tested, these assay adaptations eliminated interference with

the measurement of MTT/MTT-formazan mixtures up to

particle concentrations of 10 lg cm-2 (see Fig. 2d as an

example).

In LDH assays, the yellow tetrazolium salt INT is used to

detect active LDH released into the cell culture medium from

damaged cells. LDH converts lactate to pyruvate, thereby

generating NADH ? H? that, in turn, is used to reduce INT

to a red formazan. Interference of NPs with INT light

absorption was determined with mixtures of reduced and

oxidized INT (for details, see Materials and methods). For all

NPs, interference could be detected with particle concen-

trations of 50 lg cm-2 but not with concentrations of

10 lg cm-2 and below (see representative Fig. 3).

Influence on substrate conversion

We determined the ability of the NP dispersions to convert

the substrates H2DCF-DA, MTTox and INTox. To this

end, the substrates were added to cell-free 96-well plates

pre-incubated with NP dispersions at different particle

concentrations. When H2DCF-DA was incubated for 1 h,

we could not detect DCF fluorescence signals with any of

the particles used (see Fig. 4a, b displaying representative

examples). However, 4 h after the addition of H2DCF-DA,

we detected DCF signals in the presence of CB dispersions

that increased in a concentration-dependent manner

(Fig. 4a). All other NPs had no effect (Fig. 4b).

Influence on enzymatic activity

Because of their large surface area and high surface energy,

NPs may be able to adsorb enzymes on their surface and

possibly inactivate the adsorbed protein.

We therefore analyzed the complete set of NPs regard-

ing their effects on the enzymatic activity of lactate

dehydrogenase. LDH assays were performed in cell-free

wells containing bovine LDH in the presence of NP dis-

persions with different particle concentrations (0.01, 0.1, 1,

5 and 10 lg cm-2). At particle concentrations of up to

5 lg cm-2, we could not detect any significant influence

Fig. 2 Interference of CB (a, c) and TiO2 3 (b, d) NPs with MTT-

formazan light absorption. NP dispersions with four different particle

concentrations in DMEM/10 % FBS were added to A549 monolayers

and incubated for 24 h. Serum-containing medium (DMEM/10 %

FBS) was used as control. Mixtures of reduced and oxidized MTT

were added either directly (a, b) or after washing with KRB (c, d).

Results are displayed as arbitrary units [AU]. Mean values and

standard deviations were obtained from three independent experi-

ments with seven replications each

Arch Toxicol (2012) 86:1123–1136 1129

123

on LDH activity. However, at 10 lg cm-2, a slight but

significant inhibition of LDH activity was observed with

ZnO NPs (Fig. 5, NZO), whereas no other NPs tested

displayed any significant effect on LDH activity. Inhibition

of LDH activity could also be observed in the presence of

fine-sized ZnO particles, suggesting a composition-depen-

dent rather than a size-/surface-dependent effect (Fig. 5,

FZO).

Effects of NPs on cellular IL-8 levels

The levels of the proinflammatory peptide IL-8 were

assessed in A549 cells after a 24-h exposure to 50 lg cm-2

of each of the 24 engineered NPs and reference particles

(FZO, DQ12). For nine out of 24 NPs, a significant

induction of cellular cytokine production was observed

(Fig. 6a–d). Strikingly, after exposure to TiO2 2 NPs, the

cells displayed IL-8 secretion below control levels,

whereas in cells exposed to TiO2 1, IL-8 production

resembled control levels (Fig. 6b). TiO2 1 and TiO2 2 NPs

are identical particles from the same production process

and differ only by predispersion. TiO2 1 was obtained in

powder form, whereas TiO2 2 was predispersed by the

supplier. TiO2 3 was of similar size but of different crys-

talline structure and particle morphology. In cells stimu-

lated with TiO2 3, IL-8 levels were even higher than those

observed for the reference DQ12 and higher than any other

NPs tested (Fig. 6b).

Fig. 3 Interference of CB (a) and TiO2 3 (b) NPs with INT-formazan

light absorption. Light absorption of INTred/INTox mixtures was

measured in the presence of A549 cells pre-incubated with NP

dispersions of four different particle concentrations in DMEM/10 %

FBS. Results are displayed as arbitrary units [AU]. Mean values and

standard deviations were obtained from three independent experi-

ments with seven replications each. *Significantly different from

control (DMEM/10 % FBS) for p = 0.05

Fig. 4 NP-mediated conversion of H2DCF-DA in a cell-free assay.

Cell-free microtiter plates were pre-incubated with CB (a) or TiO2 3

(b) dispersions of five different particle concentrations, washed and

incubated with H2DCF-DA for 1 and 4 h, respectively. Serum-

containing medium (DMEM/10 % FBS) was used as control. Mean

values and standard deviations were obtained from three experiments

with seven replications each

1130 Arch Toxicol (2012) 86:1123–1136

123

Fig. 5 Inhibition of LDH

activity in the presence of NPs.

Particle dispersions (10 lg

cm-2) were mixed with bovine

LDH and incubated for 24 h.

LDH activity was determined

with INT followed by the

measurement of INT-formazan

absorption. Mean values and

standard deviations displayed as

% of control (DMEM/10 %

FBS) were obtained from three

independent experiments with

seven replications each.

*Significantly different from

control for p = 0.05

Fig. 6 IL-8 levels in A549 culture supernatants after 24-h exposition to nanoparticles. All particles were tested in triplicate in two independent

assay series each. Statistical significance was determined by Student’s t test with p \ 0.05 considered as significant (*)

Arch Toxicol (2012) 86:1123–1136 1131

123

Effects of NPs on IL-8 levels in cell-free systems

To test the influence of NPs on measurable IL-8 protein

levels, we determined the absorption of IL-8 ELISA stan-

dards with seven different IL-8 concentrations (from 4.69

to 300 pg ml-1) in the presence of NP dispersions with

particle concentrations of 50 lg cm-2 in a cell-free system.

Of the 24 NPs tested, 23 NPs did not display any sig-

nificant effect on the IL-8 standard curve after incubation

for 24 h (Fig. 7a–d). In the presence of TiO2 2, however, a

dramatic loss of immunoreactive IL-8 was observed with

all IL-8 standard concentrations tested (Fig. 7b). In con-

trast, TiO2 1 from the same production process (but not

predispersed) did not influence measurable IL-8 protein

levels. These results resemble those observed for the

detection of cellular IL-8 levels (Fig. 6b) and suggest an

adsorption of IL-8 by TiO2 2 NPs.

Adsorption of IL-8 by NPs

We further investigated the NP concentration-dependent

potential interleukin binding to a selection of NP types.

Dispersions of TiO2 NPs (1–3) and CB with different

particle concentrations (1, 10 and 50 lg cm-2) were added

to four different IL-8 protein concentrations in medium and

incubated for 24 h prior to the measurement by ELISA. As

shown in Fig. 8, we did not observe the depletion of

immunoreactive IL-8 with any of the CB, TiO2 1 and 3

dispersions tested (Fig. 8a–c).

As expected, measurable IL-8 levels were dramatically

reduced in dispersions containing 50 lg cm-2 TiO2 2 NPs

(Fig. 8d), presumably due to the adsorption of the cytokine

on the particle surface. Only a weak decrease in detectable

IL-8 protein was observed for dispersions with 10 lg cm-2

TiO2 2, and particle concentrations of 1 lg cm-2 did not

affect IL-8 measurements in a significant manner (Fig. 8d).

Discussion

In this study, we tested four classical in vitro assays for

their validity to assess NP toxicity and optimized the pro-

tocols to avoid particle interference. We used 24 well-

characterized engineered NPs (Kroll et al. 2011) and

investigated each NP for its interference with the optical

detection methods as well as its interaction with the dyes,

Fig. 7 Screening of nanoparticles for ELISA interference. Different

particle dispersions (50 lg-2) were incubated for 24 h together with

recombinant human IL-8 protein standards. Graphs shown represent

typical resulting standard curves of one out of two independent assay

series with seven different concentrations ranging from 300 to

4.69 pg ml-1 IL-8 with or without nanoparticles in DMEM ? 10 %

fetal bovine serum plus appropriate controls

1132 Arch Toxicol (2012) 86:1123–1136

123

enzymes or peptides used in DCF, MTT, LDH and IL-8

ELISA assays. Although a number of studies have shown

interaction of carbon-based nanomaterials with reporter

dyes used in cell viability assays (Belyanskaya et al. 2007;

Casey et al. 2007; Monteiro-Riviere et al. 2009; Worle-

Knirsch et al. 2006), only very limited information is

available for inorganic compounds and for their possible

interference with optical detection. Recently, Pfaller et al.

reported enhanced fluorescence intensities when cell-free

DCF assays were performed in the presence of 4.5-nm gold

NPs. This spectral interference was thought to result from

an interaction of the excited fluorophores with free elec-

trons in the metals (Pfaller et al. 2010). Here, we observed

an apparent reduction of DCF fluorescence transmission

with all of the 24 nanoparticles tested, already at particle

concentrations from 10 lg cm-2. The effect was most

pronounced in the presence of CB and can easily be

explained by its ability to absorb light in the visible spec-

trum (Wells and Smith 1941). Since both the excitation

(485-nm) and the emission (520-nm) wavelengths of the

DCF marker lie within the spectrum of visible light, CB

may absorb not only the emitted DCF fluorescence but also

the excitation energy, thereby preventing excitation of the

fluorophore. The other nanoparticles represented metal

oxides, metal hydroxides and metal carbonates that give

whitish or yellowish opaque dispersions in serum-con-

taining cell culture medium. The decreased DCF fluores-

cence signal measured in the presence of these NP

dispersions is therefore likely to result to a larger extent

from light scattering rather than from absorption. Remov-

ing NPs from the solution by including washing or cen-

trifugation steps before measuring DCF fluorescence is

therefore essential to avoid methodological artifacts. Given

the fact that the interference of NPs is concentration

dependent and that NPs cannot be removed entirely from

the solution, NP dispersions applied should also be limited

to concentrations that do not interfere with the measure-

ment of DCF signals. For the set of NPs tested here, a

maximum concentration of 10 lg cm-2or 32 lg ml-1 (for

CB 1 lg cm-2) can be used without optical interference

when performing an optimized DCF assay protocol. Since

interference with the fluorimetric measurement depends on

Fig. 8 Testing of different concentrations of selected nanoparticles

for ELISA interference. Different nanoparticle dispersions in three

concentrations (1, 10 and 50 lg cm-2) were incubated for 24 h with

recombinant human IL-8 protein standard concentrations. Data shown

represent four assays (a–d) with four different concentrations ranging

from 150 to 18.25 pg ml-1 IL-8 with or without nanoparticles

dispersions (performed four times). Statistical significance was

determined by Student’s t test with p \ 0.05 considered as significant

(*) and with p \ 0.001 considered as highly significant (**)

Arch Toxicol (2012) 86:1123–1136 1133

123

the physico-chemical properties of the NPs used, maximum

concentrations for the toxicity assessment have to be

determined for each individual particle type.

Without prior adaptation, the DCF assay would lead to

false-negative results and an underestimation of NP tox-

icity. Strikingly, in the case of CB, we also observed DCF

production in a cell-free assay possibly due to its oxidative

properties. This, in turn, leads to false-positive results and

an overestimation of the toxic effect. Therefore, our results

suggest that DCF assays used in classical toxicology

studies need to be further optimized for the use of engi-

neered NPs. Apart from lowering particle concentrations to

non-interfering levels and removing particles by washing,

we suggest testing the oxidative potential of each NP and

subsequently limiting the incubation time accordingly.

We also demonstrate interference of the engineered NPs

with the MTT assay. In particular, we found a concentra-

tion-dependent increase in MTT-formazan light absorption

for each individual NP. The contribution of CB to the

absorption of MTT-formazan is likely to result mainly

from its light absorptive capacity since we could not detect

a cell-free conversion of oxidized MTT in the presence of

CB as it has been suggested earlier (Monteiro-Riviere et al.

2009). The enhanced MTT-formazan light absorption

observed in the presence of TiO2 and of all other NPs may

be basically due to light scattering effects. Such effects

have been reported for TiO2 NPs of 40 nm size. When

applied to resin composites, these particles significantly

reduced the translucency and increased opalescence (Yu

et al. 2009). However, the contribution of NPs to the MTT

light absorption signal does not only increase with the NP

concentration used but also with the concentration of

reduced MTT-formazan present in the MTTred/MTTox

mixtures, indicating different mechanisms of interference

that cannot be predicted a priori. Therefore, determining

background signals for NPs with MTT in cell-free controls

may not be sufficient to avoid this bias in the results. Our

findings suggest that also in the MTT protocols, washing

steps should be included and NP concentrations should be

limited to non-interfering levels (10 lg cm-2 for the NP

used here) to reduce the amount of NPs present during the

measurement. However, in the case of CB, these assay

adaptations were not appropriate. Results obtained with the

optimized assay protocol indicate the interference of CB

even at particle concentrations of 10 lg cm-2. Due to its

high adsorption capacity, CB may adsorb to the surface of

the microtiter plate or the cell monolayer and may not be

efficiently removed by washing. The interference of CB

with the measurement of MTT will lead to false-negative

results and consequently to an underestimation of its

cytotoxicity. MTT assays are therefore not useful for

testing the cytotoxicity of CB. The suitability of MTT

assays to assess the in vitro toxicity of nanomaterials has

been discussed previously. Carbon nanotubes have been

demonstrated to adsorb MTT-formazan crystals, thereby

preventing solubilization (Belyanskaya et al. 2007; Casey

et al. 2007; Worle-Knirsch et al. 2006). Recently, meso-

porous silica NPs have been shown to interfere with the

MTT test by accelerating the exocytosis of formazan

crystals likely due to the perturbation of intracellular ves-

icle trafficking by silica NP uptake (Fisichella et al. 2009).

Silver nanoparticles were found to interact directly with

MTT, suggesting greater than 100 % viability of nano-

particle-exposed cells (Sadik et al. 2009). These studies

and our data demonstrate the pleiotropic ways in which

nanomaterials can interfere with the MTT assay and sug-

gest careful validation when employing this assay for NP

toxicity evaluation.

Our results also demonstrate the interference of all 24

NPs with the optical detection of reduced INT. Because

INT-formazan also absorbs in the spectrum of visible light,

the mechanisms underlying this interference might be

similar to those discussed for the NP-increased MTT

light absorption. Contrary to the interference with MTT

light absorption, however, an enhancement of INT light

absorption by NPs gives false-positive results and leads to

an overestimation of NP toxicity. Since reduced INT is

present in the supernatant of cells, NPs cannot be removed

by washing prior to the measurement. Removal of NPs by

centrifugation depends highly on the NP type and might

only be applicable when individual particles are tested.

When INT light absorption has to be measured in the

presence of NPs, the concentration of NPs should be lim-

ited to a maximum concentration that does not interfere

with the colorimetric detection. For the NPs tested in this

study, a maximum of 10 lg cm-2 was used without par-

ticle interference.

We have further determined the influence of NPs on the

enzymatic activity of LDH and observed a partial decrease

in LDH activity in the presence of ZnO NPs and fine-sized

ZnO. Since we found similar inhibition effects with ZnO

NPs and fine-sized ZnO and a similar solubility of both

ZnO particle types in cell culture medium (data not

shown), inhibition of LDH activity may be triggered by

released zinc ions as it has been described earlier for ZnCl

(Carlsson et al. 1993). In contrast to interference of NPs

with INT absorption, this enzymatic inactivation generates

false-negative results and makes ZnO NPs appear less toxic

as they actually are. This again demonstrates that even

within one single assay NP interference can influence the

toxicity outcome in different directions leading to highly

biased results and invalid conclusions.

Another confounding factor within NP toxicity screen-

ing might be the ability of NPs to specifically adsorb

proteins on their surface (Cedervall et al. 2007; Lundqvist

et al. 2008). This is particularly important when proteins

1134 Arch Toxicol (2012) 86:1123–1136

123

are used as reporters to indicate particle toxicity as it is the

case when assessing the proinflammatory potential of NPs.

A variety of NPs has been reported to adsorb proin-

flammatory cytokines. Metal oxide NPs such as TiO2 and

SiO2 seem to adsorb IL-6 (Veranth et al. 2007), and CB

was found to bind several different cytokines, for example,

GM-CSF, TGF-ß, TNFa, IL-6 and IL-8 (Brown et al. 2010;

Kim et al. 2003; Kocbach et al. 2008; Monteiro-Riviere

and Inman 2006; Seagrave et al. 2004; Val et al. 2009). In

contrast to the studies of Kocbach et al. and Brown et al.,

we could not detect any IL-8 adsorption by CB particles of

14 nm size (Brown et al. 2010; Kocbach et al. 2008).

Although Brown et al. used the same NP type as we did in

our study, their particles were suspended in PBS, whereas

we used serum-containing cell culture medium for the

preparation of NP dispersions (Schulze et al. 2008).

Kocbach et al. 2008 reported that cytokine binding could

be nearly completely reduced by adding serum proteins to

particle suspensions (Kocbach et al. 2008). Moreover,

cytokine binding has been found to increase with increas-

ing particle concentrations (Brown et al. 2010; Kocbach

et al. 2008). Thus, limiting NP concentrations to a maxi-

mum that does not display cytokine binding is of particular

importance to evade underestimating the proinflammatory

potential of engineered NPs. Our results demonstrate that

CB dispersions in serum-containing medium with particle

concentrations up to 50 lg cm-2 do not display significant

cytokine binding in cell-free controls. However, this does

not hold true for all 24 engineered NPs tested. One of the

three TiO2 NPs (TiO2 2) adsorbed considerable amounts of

cytokine when particle concentrations of 50 lg cm-2 and

below were used. As TiO2 2 differs from TiO2 1 only by

predispersion and TiO2 1 did not significantly decrease

measurable IL-8 protein levels, it can be assumed that apart

from chemical composition (Val et al. 2009) and surface

area (Brown et al. 2010), even minor modifications such as

predispersion may influence the cytokine binding proper-

ties of engineered NPs.

In conclusion, this study highlights the non-biological

artifacts obtained with classic cytotoxicity assays when

testing NPs. In particular, we observed a concentration-

dependent interference of all 24 engineered NPs with the

optical measurements used to determine the oxidative

stress (DCF), cellular metabolic activity (MTT) and cell

viability (LDH). Influence of NPs on substrates or enzymes

depended on the particle composition and, in the case of

cytokine binding, also on particle predispersion. Our results

suggest that classic cytotoxicity assays have to be further

developed, validated for each particle tested and restricted

to NP concentrations below interfering levels. Assay

adaptations have to be ascertained by a series of control

experiments for each single NP variant to obtain reliable

NP toxicity data.

Acknowledgments This work was supported by grants of the

German Federal Ministry of Education and Research (BMBF projects

NanoCare and Cell@Nano) and the state NRW (NanoPaCT). We

thank Birgit Phillip for excellent technical assistance.

Conflict of interest The authors declare that they have no conflict

of interest.

References

Belyanskaya L, Manser P, Spohn P, Bruinink A, Wick P (2007) The

reliability and limits of the MTT reduction assay for carbon

nanotubes-cell interaction. Carbon 45:2643–2648

Brown DM, Dickson C, Duncan P, Al-Attili F, Stone V (2010)

Interaction between nanoparticles and cytokine proteins: impact

on protein and particle functionality. Nanotechnology 21:215104

Carlsson H, Prachayasittikul V, Bulow L (1993) Zinc ions bound to

chimeric His4/lactate dehydrogenase facilitate decarboxylation

of oxaloacetate. Protein Eng 6:907–911

Casey A, Herzog E, Davoren M, Lyng FM, Byrne HJ, Chambers G

(2007) Spectroscopic analysis confirms the interactions between

single walled carbon nanotubes and various dyes commonly used

to assess cytotoxicity. Carbon 45:1425–1432

Casey A, Herzog E, Lyng FM, Byrne HJ, Chambers G, Davoren M

(2008) Single walled carbon nanotubes induce indirect cytotoxicity

by medium depletion in A549 lung cells. Toxicol Lett 179:78–84

Cedervall T, Lynch I, Lindman S, Berggard T, Thulin E, Nilsson H,

Dawson KA, Linse S (2007) Understanding the nanoparticle-

protein corona using methods to quantify exchange rates and

affinities of proteins for nanoparticles. Proc Nat Acad Sci USA

104:2050–2055

Cho SJ, Maysinger D, Jain M, Roder B, Hackbarth S, Winnik FM

(2007) Long-term exposure to CdTe quantum dots causes

functional impairments in live cells. Langmuir 23:1974–1980

Doak SH, Griffiths SM, Manshian B, Singh N, Williams PM, Brown

AP, Jenkins GJ (2009) Confounding experimental considerations

in nanogenotoxicology. Mutagenesis 24:285–293

Duffin R, Tran L, Brown D, Stone V, Donaldson K (2007)

Proinflammogenic effects of low-toxicity and metal nanoparti-

cles in vivo and in vitro: highlighting the role of particle surface

area and surface reactivity. Inhalation Toxicol 19:849–856

Esch RK, Han L, Foarde KK, Ensor DS (2010) Endotoxin contam-

ination of engineered nanomaterials. Nanotoxicology 4:73–83

Fisichella M, Dabboue H, Bhattacharyya S, Saboungi ML, Salvetat

JP, Hevor T, Guerin M (2009) Mesoporous silica nanoparticles

enhance MTT formazan exocytosis in HeLa cells and astrocytes.

Toxicol In Vitro 23:697–703

Grassian VH, O’Shaughnessy PT, Adamcakova-Dodd A, Pettibone

JM, Thorne PS (2007) Inhalation exposure study of titanium

dioxide nanoparticles with a primary particle size of 2 to 5 nm.

Environ Health Perspect 115:397–402

Guo L, Von Dem Bussche A, Buechner M, Yan A, Kane AB, Hurt

RH (2008) Adsorption of essential micronutrients by carbon

nanotubes and the implications for nanotoxicity testing. Small

(Weinheim an der Bergstrasse, Germany) 4:721–727

Jakubowski W, Bartosz G (2000) 2,7-dichlorofluorescin oxidation and

reactive oxygen species: what does it measure? Cell Biol Int

24:757–760

Johnston HJ, Hutchison GR, Christensen FM, Aschberger K, Stone V

(2009) The biological mechanisms and physicochemical char-

acteristics responsible for driving fullerene toxicity. Toxicol Sci

114:162–182

Kim H, Liu X, Kobayashi T, Kohyama T, Wen FQ, Romberger DJ,

Conner H, Gilmour PS, Donaldson K, MacNee W, Rennard SI

Arch Toxicol (2012) 86:1123–1136 1135

123

(2003) Ultrafine carbon black particles inhibit human lung

fibroblast-mediated collagen gel contraction. Am J Respir Cell

Mol Biol 28:111–121

Kocbach A, Totlandsdal AI, Lag M, Refsnes M, Schwarze PE (2008)

Differential binding of cytokines to environmentally relevant

particles: a possible source for misinterpretation of in vitro

results? Toxicol Lett 176:131–137

Korzeniewski C, Callewaert DM (1983) An enzyme-release assay for

natural cytotoxicity. J Immunol Methods 64:313–320

Kroll A, Pillukat MH, Hahn D, Schnekenburger J (2009) Current in

vitro methods in nanoparticle risk assessment: limitations and

challenges. Eur J Pharm Biopharm 72:370–377

Kroll A, Dierker C, Rommel C, Hahn D, Wohlleben W, Schulze-

Isfort C, Gobbert C, Voetz M, Hardinghaus F, Schnekenburger J

(2011) Cytotoxicity screening of 23 engineered nanomaterials

using a test matrix of ten cell lines and three different assays.

Part Fibre Toxicol 8:9

Landsiedel R, Ma-Hock L, Kroll A, Hahn D, Schnekenburger J,

Wiench K, Wohlleben W (2010) Testing metal-oxide nanoma-

terials for human safety. Adv Mater 22:2601–2627

Lundqvist M, Stigler J, Elia G, Lynch I, Cedervall T, Dawson KA

(2008) Nanoparticle size and surface properties determine the

protein corona with possible implications for biological impacts.

Proc Nat Acad Sci USA 105:14265–14270

Mahmoudi M, Simchi A, Imani M, Shokrgozar MA, Milani AS,

Hafeli UO, Stroeve P (2010) A new approach for the in vitro

identification of the cytotoxicity of superparamagnetic iron oxide

nanoparticles. Colloids Surf 75:300–309

Monteiro-Riviere NA, Inman AO (2006) Challenges for assessing

carbon nanomaterial toxicity to the skin. Carbon 44:1070–1078

Monteiro-Riviere NA, Inman AO, Zhang LW (2009) Limitations and

relative utility of screening assays to assess engineered nano-

particle toxicity in a human cell line. Toxicol Appl Pharmacol

234:222–235

Nachlas MM, Margulies SI, Goldberg JD, Seligman AM (1960) The

determination of lactic dehydrogenase with a tetrazolium salt.

Anal Biochem 1:317–326

NanoCare Project (2009) NanoCare: health related aspects of nanom-

aterials. Final Scientific Report, Dechema e.V., Frankfurt a.M

Oberdorster G, Oberdorster E, Oberdorster J (2005) Nanotoxicology:

an emerging discipline evolving from studies of ultrafine

particles. Environ Health Perspect 113:823–839

Pan Z, Lee W, Slutsky L, Clark RA, Pernodet N, Rafailovich MH

(2009) Adverse effects of titanium dioxide nanoparticles on

human dermal fibroblasts and how to protect cells. Small 5:511–

520

Pfaller T, Colognato R, Nelissen I, Favilli F, Casals E, Ooms D,

Leppens H, Ponti J, Stritzinger R, Puntes V, Boraschi D, Duschl

A, Oostingh GJ (2010) The suitability of different cellular in

vitro immunotoxicity and genotoxicity methods for the analysis

of nanoparticle-induced events. Nanotoxicology 4:52–72

Rabolli V, Thomassen LC, Princen C, Napierska D, Gonzalez L,

Kirsch-Volders M, Hoet PH, Huaux F, Kirschhock CE, Martens

JA, Lison D (2010) Influence of size, surface area and

microporosity on the in vitro cytotoxic activity of amorphous

silica nanoparticles in different cell types. Nanotoxicology

4:307–318

Rothen-Rutishauser B, Brown DM, Piallier-Boyles M, Kinloch IA,

Windle AH, Gehr P, Stone V (2010) Relating the physicochemical

characteristics and dispersion of multiwalled carbon nanotubes

in different suspension media to their oxidative reactivity in vitro

and inflammation in vivo. Nanotoxicology 4:331–342

Sabatini CA, Pereira RV, Gehlen MH (2007) Fluorescence modula-

tion of acridine and coumarin dyes by silver nanoparticles.

Journal of fluorescence 17:377–382

Sadik OA, Zhou AL, Kikandi S, Du N, Wang Q, Varner K (2009)

Sensors as tools for quantitation, nanotoxicity and nanomoni-

toring assessment of engineered nanomaterials. J Environ Monit

11:1782–1800

Sayes CM, Wahi R, Kurian PA, Liu Y, West JL, Ausman KD,

Warheit DB, Colvin VL (2006) Correlating nanoscale titania

structure with toxicity: a cytotoxicity and inflammatory response

study with human dermal fibroblasts and human lung epithelial

cells. Toxicol Sci 92:174–185

Schulze C, Kroll A, Lehr CM, Schafer UF, Becker K, Schnekenburger J,

Schulze Isfort C, Landsiedel R, Wohleben W (2008) Not ready to

use—overcoming pitfalls when dispersing nanoparticles in phys-

iological media. Nanotoxicology 2:51–61

Seagrave J, Knall C, McDonald JD, Mauderly JL (2004) Diesel

particulate material binds and concentrates a proinflammatory

cytokine that causes neutrophil migration. Inhalation Toxicol

16(Suppl 1):93–98

Shaw SY, Westly EC, Pittet MJ, Subramanian A, Schreiber SL,

Weissleder R (2008) Perturbational profiling of nanomaterial

biologic activity. Proc Nat Acad Sci USA 105:7387–7392

Singh S, Nalwa HS (2007) Nanotechnology and health safety–toxicity

and risk assessments of nanostructured materials on human

health. J Nanosci Nanotechnol 7:3048–3070

Stearns RC, Paulauskis JD, Godleski JJ (2001) Endocytosis of

ultrafine particles by A549 cells. Am J Respir Cell Mol Biol

24:108–115

Val S, Hussain S, Boland S, Hamel R, Baeza-Squiban A, Marano F

(2009) Carbon black and titanium dioxide nanoparticles induce

pro-inflammatory responses in bronchial epithelial cells: need for

multiparametric evaluation due to adsorption artifacts. Inhalation

Toxicol 21(Suppl 1):115–122

Veranth JM, Kaser EG, Veranth MM, Koch M, Yost GS (2007)

Cytokine responses of human lung cells (BEAS-2B) treated with

micron-sized and nanoparticles of metal oxides compared to soil

dusts. Particle and fibre toxicology 4:2

Wang J, Chen C, Liu Y, Jiao F, Li W, Lao F, Li Y, Li B, Ge C, Zhou

G, Gao Y, Zhao Y, Chai Z (2008) Potential neurological lesion

after nasal instillation of TiO(2) nanoparticles in the anatase and

rutile crystal phases. Toxicol Lett 183:72–80

Wells AJ, Smith WR (1941) The absorption spectrum of suspensions

of carbon black. J Phys Chem 45:1055–1060

Wolf R, Wolf D, Morganti P, Ruocco V (2001) Sunscreens. Clin

Dermatol 19:452–459

Worle-Knirsch JM, Pulskamp K, Krug HF (2006) Oops they did it

again! Carbon nanotubes hoax scientists in viability assays.

Nano Lett 6:1261–1268

Yu B, Ahn JS, Lim JI, Lee YK (2009) Influence of TiO2 nanoparticles

on the optical properties of resin composites. Dent Mater

25:1142–1147

1136 Arch Toxicol (2012) 86:1123–1136

123