Aa

Ca

b

a

ARRAA

K�GIDD

1

tcasca

toPp

Cr

oT

1h

Mutation Research 757 (2013) 158– 166

Contents lists available at ScienceDirect

Mutation Research/Genetic Toxicology andEnvironmental Mutagenesis

j o ur nal hom epa g e: www.elsev ier .com/ locate /gentoxCo mm u ni t y add ress : www.elsev ier .com/ locate /mutres

ssessment of the in vitro �H2AX assay by High Content Screening as novel genotoxicity test�

arolina Garcia-Cantona,b,∗, Arturo Anadonb, Clive Mereditha

British American Tobacco, Group Research and Development, Regents Park Road, Southampton, Hampshire SO15 8TL, United KingdomDepartment of Toxicology and Pharmacology, Universidad Complutense de Madrid, Madrid, Spain

r t i c l e i n f o

rticle history:eceived 22 March 2013eceived in revised form 7 July 2013ccepted 7 August 2013vailable online 27 August 2013

eywords:H2AXenotoxicity

a b s t r a c t

The �H2AX assay is widely used as a marker of DNA damage in multiple scientific fields such as can-cer biomarker, clinical studies and radiation biology. In particular, the in vitro �H2AX assay has beensuggested as a novel in vitro genotoxicity test with potential as a pre-screening tool. However, to date,limited assessments have been carried out to evaluate the sensitivity, specificity and accuracy of thein vitro �H2AX assay.

In this study, the microscopy-based system combining automated cellular image acquisition with soft-ware quantification for High Content Screening (HCS) has been used for the first time to evaluate thein vitro �H2AX assay. A panel of well-characterised genotoxic and non-genotoxic compounds was selected

n vitroNA damageouble strand breaks

to assess the performance of the in vitro �H2AX assay in the human bronchial epithelial cell line BEAS-2B.The results obtained during this preliminary assessment indicate that the in vitro �H2AX assay has

a high accuracy (86%) as a result of high sensitivity and specificity (86–92% and 80–88% respectively).Our data highlight the potential for �H2AX detection in HCS as a complement to the current regulatorygenotoxicity battery of in vitro assays. We therefore recommend more comprehensive assessments toconfirm the performance of the in vitro �H2AX assay by HCS with a more extensive set of compounds.

© 2013 The Authors. Published by Elsevier B.V. All rights reserved.

. Introduction

The field of genetic toxicology testing in the 21st century faceswo challenges recently described by Mahadevan et al. [1]; “The firsthallenge is to take full advantage of new technologies to improve ourbility to assess the impacts of chemically induced genetic damage. Theecond is to use these technologies to reliably assess new and existinghemicals for genetic toxicity potential more efficiently, cost-effectivelynd with less reliance on animal models”.

In particular, current advisory bodies such as the Commit-ee of Mutagenicity (COM) and the International Conferencen Harmonisation of Technical Requirements for Registration of

harmaceuticals for Human Use (ICH), recommend two options forre-clinical genotoxicity testing [2,3]:

� This is an open-access article distributed under the terms of the Creativeommons Attribution License, which permits unrestricted use, distribution andeproduction in any medium, provided the original author and source are credited.∗ Corresponding author at: British American Tobacco, Group Research and Devel-

pment, Regents Park Road, Southampton, Hampshire SO15 8TL, United Kingdom.el.: +44 02380 588 969; fax: +44 02380 793 076.

E-mail address: Carolina Garcia-Canton@bat.com (C. Garcia-Canton).

383-5718/$ – see front matter © 2013 The Authors. Published by Elsevier B.V. All rights ttp://dx.doi.org/10.1016/j.mrgentox.2013.08.002

Option 1i. A test for gene mutation in bacteria.

ii. A cytogenetic test for mammalian chromosomal damage(in vitro metaphase chromosome aberration test or in vitromicronucleus) or an in vitro mouse lymphoma Tk gene muta-tion assay.

iii. An in vivo test for genotoxicity, generally a test for chromo-somal damage using rodent hematopoietic cells.

Option 2i. A test for gene mutation in bacteria.

ii. An in vivo test for genotoxicity assessment with two differenttissues.

The recommended in vitro mammalian tests are very laborious,especially due to their manual scoring requirements. Additionally,they have the limitation of reporting a large number of “false posi-tives” [4]. For pharmaceuticals, both options require the additionof an in vivo test. However, this generates an increase in animaluse and might lead to delays in product development. Moreover,in the cosmetic industry, the use of animal testing is banned by

the EU 7th Amendment Directive [5]. In order to avoid unneces-sary in vivo testing, alternative in vitro assays could be used as partof the weight of evidence to support non-relevant in vitro positi-ves that could be negative in vivo, also known as “false positive” [6].

reserved.

tion R

Ahpoioma�iDd�idiciobttbwaicsc

flAmmHtbb(tnfariipam

tltaotccctrDs�sr

C. Garcia-Canton et al. / Muta

lternative in vitro assays could also become part of a pre-screeningigh throughput filter tool to discard potential genotoxic com-ounds early on during development. Some of these alternativer “non-core” in vitro assays have been reviewed and classifiednto mature, maturing or emerging technologies [7]. A numberf emerging technologies are related to the understanding of theechanisms of DNA damage. One novel endpoint of DNA dam-

ge is the phosphorylation of the histone 2AX at serine-139 namedH2AX, in response to double-strand breaks (DSB). This early event

n the DNA damage response (DDR) mechanism is triggered whenNA damage occurs to avoid genome instability and was firstescribed by Rogakou et al. [8]. Since then, the measurement ofH2AX has increasingly been used as a biomarker of DNA damage

n multiple scientific fields such as in vitro mechanistic studies andrug development, in vivo DNA damage and repair studies and clin-

cal trials [9,10]. In recent years, the automation of microscopy inellular-based assays and the potential for higher throughput havencreased its applicability as a complement to the existing batteryf in vitro genotoxicity assays [11]. However, the main efforts haveeen focused on measuring �H2AX by flow cytometry [12,13], evenhough such methodologies have been described to be less sensitivehan microscopic analysis [14]. The sensitivity of the microscopy-ased method is based on the possibility to identify single fociithin the cell’s nucleus [15]. Other technical benefits include the

bility to score attached cells directly from the micro-well platenstead of adding an extra step in the methodology to generate aell suspension. Moreover, the cell sample is not destroyed duringcoring, allowing cell distribution, morphology and other qualityontrol checks to be carried out post-scoring.

In this study, microscopy has been selected as an alternative toow cytometry for the detection of �H2AX in vitro. The CellomicsrrayScan® VTI platform (ThermoScientific, USA) is an automatedicroscopy system that combines cellular image acquisition fromulti-well microplates with software quantification analysis forigh Content Screening (HCS) [16]. The sensitivity and specificity of

he in vitro �H2AX assay to detect genotoxic potential was assessedy testing a panel of 22 compounds that had been recommendedy the European Centre for the Validation of Alternative MethodsECVAM) and used in previous assessments of in vitro genotoxicityests [17,18]. The compounds selected represent different mecha-ism of action that could lead to the formation of DNA damage in the

orm of DSB directly or indirectly as previously described by Bonnernd colleagues [19]. Bleomycin sulphate, for example, was used aseactive oxygen species generator known to directly produce DSBsn the DNA. Etoposide, on the other hand, is a topoisomerase IInhibitor that increases the number of topoisomerase-DNA com-lexes causing an increase in collisions between the complexesnd the replication forks. The collisions result in DSB by an indirectechanism.Selecting an appropriate cell line could also affect the sensi-

ivity of the assay. Recently published recommendations on celline selection suggest the use of human p-53 competent cell lineso reduce the incidence of “false positives” [20]. To develop thessay, we selected the normal phenotype-derived BEAS-2B cellsbtained from non-cancerous lung epithelium cells and immor-alised by Reddel and colleagues [21]. When measuring �H2AX,ell lines derived from normal tissues, such as BEAS-2B cells, areonsidered to be more reliable than those from pre-malignant andancer tissues. Tumorigenic cell lines have an impairment to func-ions such as the DNA damage repair mechanisms and cell cycleegulation [22]. The phosphorylation of H2AX is an early event inNA damage response and is also involve in cell cycle progres-

ion [23,24], therefore, malignant transformation could affect theH2AX response. For instance, there have been reports showingignificant amount of �H2AX in the absence of DSB, possibly as aesponse to “oncogenic stress” [25]. BEAS-2B cells, however, have

esearch 757 (2013) 158– 166 159

been reported to have limited metabolic capabilities reducing itsability to detect compounds that require metabolic activation, i.e.,pro-toxicants [26]. Therefore, this study included the addition ofa standard rat-derived subcellular fraction as external source ofmetabolic activation.

The preliminary results obtained were used to define the sen-sitivity and specificity of the assay using previously recommendedcriteria [17]. The results were also compared with existing in vitro�H2AX relevant data published in the scientific literature. Overall,the in vitro �H2AX assay by HCS has shown potential as an in vitrogenotoxicity assay that could be included as a complementary toolto the existing battery of in vitro regulatory genotoxicity assays.Moreover, the high throughput technology and reproducibility ofthe in vitro �H2AX assay by HCS could be a very useful tool inpre-screening for in vitro genotoxicity.

2. Materials and methods

2.1. Chemicals

All compounds were of the highest purity available from Sigma–Aldrich (UnitedKingdom) except MNNG (Toronto Research Chemicals, Canada). They were dis-solved in an appropriate vehicle and further dilute to a final concentration of 1%(v/v) vehicle in cell culture medium. Selection of doses and treatment times followedcurrent ICH regulatory guidelines [3]. Therefore, the maximum concentration testedwas 1 mM unless precipitation was observed.

2.2. Cell culture

The human bronchial epithelial cell line BEAS-2B, purchased from ATCC(USA), was seeded into culture vessels that had been pre-coated with 0.03 mg/mLPureCol® bovine collagen solution (Nutacon, The Netherlands). Cells were main-tained in Bronchial Epithelial Growth Medium (BEGM®) at 37 ◦C and 5% CO2 in ahumidified incubator. BEGM® was prepared by supplementing Bronchial EpithelialBasal Medium with growth supplements provided in the manufacturer’s BEGM®

SingleQuot® kit (Lonza Group Ltd., Belgium) containing: bovine pituitary extract,hydrocortisone, human epidermal growth factor, epinephrine, insulin, triiodothyro-nine, transferrin, gentamicin/amphotericin-B and retinoic acid. BEAS-2B cells werecultured and expanded in-house. All cultures were negative for mycoplasma. Addi-tionally, the cells were authenticated using the short tandem repeat profiling toconfirm the nature of the cell cultures (LGC Standards, United Kingdom) [27].

2.3. Treatments

The day before the treatment, BEAS-2B cells were seeded into 96-well, black,clear-bottom microplates (PerkinElmer, United Kingdom) at a concentration of8 × 104 cells/mL.

Cells were maintained at 37 ◦C in an atmosphere of 5% CO2 in air and tested withcompounds for either 3 h or 24 h. For compounds that required metabolic activation,the treatment was for 3 h with and without the metabolic activation system. Sincesome of these compounds are known to bind DNA, a 24 h recovery period treatmentwas also included in the experimental design to quantify potentially delayed DNAdamage (3 + 24 h).

The external source of metabolic activation used in this study was rat hepaticS9 (Aroclor 1254-induced animals) (Moltox, USA), used as the standard system formetabolic activation in regulatory in vitro assays.

The preparation of the S9 mix (0.5%, v/v, final concentration in medium) wascarried out in accordance with the suppliers’ instructions. The concentration of sub-cellular fraction used was below the typically used range (1–10%, v/v) due to anintrinsic cytotoxicity observed in BEAS-2B cells. Briefly, Reagent A (Moltox, USA) amixture of phosphate-buffered salt solution and glucose-6-phosphate and ReagentB (Moltox, USA) containing NADP were mixed at a proportion of 1:3.4 creating aclear solution named co-factor mix. The co-factor mix was then used to dilute therat S9 to a 10% (v/v) intermediate solution (S9 mix). The intermediate solution wasfurther diluted in BEGM® media to prepare the treatment solution containing thetest compound.

2.4. Immunostaining

Following treatment, the compound or recovery media was aspirated fromthe cells and the plates were processed for �H2AX immunostaining followingmanufacturer’s recommendations (ThermoScientific, USA). Briefly, cells were fixed

with 4% paraformaldehyde (100 �L/well) and incubated for 15 min at room tem-perature. After fixation, plates were washed twice with 100 �L/well phosphatebuffered saline (PBS) (0.14 M sodium chloride, 0.003 M potassium chloride, 0.002 Mpotassium phosphate and 0.01 M sodium phosphate). The cells were then perme-abilised for 15 min by incubating them at room temperature with 100 �L/well of

1 tion R

pwptTp1abgabwwa

2

AA

2fdwdeactfl

bep

nt

ctndacwR

aaqD

2

�u

d

t[cmo

n

TG

60 C. Garcia-Canton et al. / Muta

ermeabilisation buffer (PBS with 1% Triton X-100). After permeabilisation, the cellsere washed twice with 100 �L/well of PBS. Blocking buffer (ThermoScientific pro-rietary formulation supplemented with 2% foetal bovine serum) (100 �L/well) washen added for 15 min at room temperature to avoid unspecific antibody binding.he blocking buffer was aspirated and 50 �L primary antibody solution (0.1% �H2AXrimary antibody in PBS) was added to each well and plates were incubated for

h at room temperature. After the incubation, the primary antibody solution wasspirated and plates washed twice with PBS II (PBS with 1% Tween-20) followedy one wash with PBS. Then, 50 �L secondary antibody (DyLightTM 549 conjugatedoat anti-mouse IgG) containing Hoechst dye DNA staining solution (0.2% secondaryntibody and 0.01% Hoechst in PBS) was added to each well and plates were incu-ated for 45 min at room temperature in the dark. After the incubation, the wellsere washed twice with PBS II followed by one wash with PBS. Finally, the wellsere filled with 200 �L of PBS and sealed for scanning. Plates were stored at 4 ◦C

fter the scan or while awaiting the scan.

.5. Imaging analysis

Images from the wells at 20× magnification were taken using the Cellomicsrrayscan® VTI platform (Thermo Scientific, USA) and analysed with the Targetctivation Bioapplication software V.6.6.1.4.

The algorithm parameters for object selection was set to count a minimum of50 cells per well, giving a minimum of 1500 cells per concentration tested. Two dif-erent nuclear stains were measured. Nuclear DNA fluorescence intensity (Hoechstye) was assessed in channel 1 to identify viable cells. Micrographies were obtainedith a fixed exposure time of 0.04 s. Object area and shape parameters were defineduring the algorithm set up to fit the specific nuclei morphology of the cell line, gen-rating the criteria to define valid nuclei. The valid nuclei from channel 1 showed

green overlay that served as the target to measure the fluorescence intensity ofhannel 2. Rejected nuclei from channel 1 showed an orange overlay and no fur-her measures are taken from them. No cell rejection in channel 1 was based onuorescence intensity.

Channel 2 measures whole nuclei fluorescence intensity of the secondary anti-ody (DyLightTM 549) in valid nuclei. Micrographies were obtained with a fixedxposure time of 0.06 s. All intensities are automatically calculated by the softwareer pixel in the identified target area of every valid object.

With all the micrographies parameters quantified, the software generated aumber of measurements to report, in this study we selected two for their relevanceo the �H2AX endpoint:

“Selected Object Count per Valid Field” is the average number of objects (viableells) selected for analysis per valid field in the well. These data were obtained fromhe identified nuclei in channel 1. Untreated wells would have a higher number ofuclei per field while toxic doses would have lower number of nuclei per field. Theseata were converted into cell viability data. The vehicle treated counts were defineds 100% cell viability. The cell counts in the compound treated wells were thenompared to those in the vehicle-treated wells, and the percentage cell viabilityas calculated. This comparison against the vehicle control data is referred to aselative Cell Counts (RCC) and expressed as percentage.

“Mean average intensity in channel 2” is the average intensity in channel 2 ofll pixels within the overlay mask identified in valid nuclei in channel 1. The meanverage intensity is measured in intensity units and was reported as �H2AX fre-uency (intensity units) in this study. This measurement relates to the number ofSBs present in the cells.

.6. Data analysis and criteria

All experiments were repeated at least 3 independent times. The results for bothH2AX intensity and RCC for all tested compounds were graphically representedsing GraphPad Prism software v.6.

The evaluation criteria for the in vitro �H2AX assay selected in this study isescribed in Table 1 and was previously presented by Smart et al. [12].

The in vitro to in vivo predictivity of the assay was defined by the calculation ofhe “Sensitivity”, “Specificity” and “Accuracy” defined previously by Kirkland et al.28]. The sensitivity is defined as the ability of the assay to detect in vitro genotoxic

ompounds. These compounds should give a positive response in in vitro mam-alian cell genotoxicity tests. They are also in vivo genotoxic compounds and are

ften classified as carcinogenic compounds [17].Specificity, on the other hand, is defined as the ability of the assay to discrimi-

ate non-genotoxic compounds. These compounds should give a negative response

able 1enotoxicity evaluation criteria for the in vitro �H2AX assay, adapted from [12].

�H2AX response Classification

>1.5-fold �H2AX @ RCC > 25% Genotoxic (+)<1.5-fold �H2AX @ RCC 100–0% Non-genotoxic (−)>1.5-fold �H2AX @ RCC < 25% “False” positive;

cytotoxic-drivengenotoxicity (C)

1.5-fold �H2AX @ RCC ≥ 25% Equivocal (±)

esearch 757 (2013) 158– 166

in in vitro mammalian cell tests. When there is data available in in vivo genotoxicitytesting, the results are usually negative. They are often non-carcinogenic or if car-cinogenicity is observed in vivo, the mode of action assumed is non-mutagenic [17].

Some of the compounds tested are classified as artifactual in vitro genotoxic posi-tives (false positive). These compounds should give a negative response in in vitromammalian cell tests but often have been reported to produce positive genotoxicity,especially at high cytotoxic concentrations. They are commonly negative in in vivogentotoxicity tests [17].

Sensitivity = A

B

Specificity = C

D

Accuracy = A + C

B + D

where A is the number of in vivo genotoxic compounds that tested positive in thein vitro �H2AX assay evaluated in this study, B is the total number of in vivo genotoxiccompounds tested. C is the number of in vivo non-genotoxic compounds that testednegative in the in vitro �H2AX assay, and D is the total number of in vivo non-genotoxic compounds tested.

3. Results

The assessment of the in vitro �H2AX assay by HCS wasperformed by treating BEAS-2B cells with different control com-pounds. The increase of histone 2AX phosphorylation (�H2AX)labelled by immunostaining and detected by automated fluores-cence microscopy was used as an indication of DNA double strandbreaks. An illustration �H2AX detection is given in Fig. 1 after 3 htreatment of the cells with etoposide.

For this study, we selected a series of genotoxic compoundswith different mechanisms of action in order to have a compre-hensive selection of DNA-damaging agents. Direct DNA-damagingcompounds interact directly with the DNA causing different degreeof DNA damage ranging from point mutations to crosslinks. Ethylmethanesulfonate, mitomycin C and methylnitronitrosoguanidine(MNNG) were used in this study as they interact directly with theDNA. Similarly, reactive oxygen species (ROS) generators such asbleomycin sulphate, also used here, generate ROS molecules thathave the potential to oxidise DNA. Some direct DNA-damagingcompounds have to be metabolically activated before their toxicform is active, these compounds are known as pro-toxicants.The limitations of some cell systems used in in vitro toxicolog-ical testing to activate these compounds implies that externalsources of metabolic activation would be required to carry out sub-sequent endpoint analysis. Benzo(a)pyrene, cyclophosphamide,2-acetylaminofluorene and aflatoxin B1 were used as pro-toxicantmodels during this study.

Indirect DNA-damaging compounds interfere with cellularfunctions involved in DNA repair, synthesis or mitosis. Damage inthese fundamental functions could ultimately translate into DNAdamage. In this assessment, amsacrine and etoposide were used astopoisomerase inhibitors while aphidicolin and zidovudine wereused as DNA synthesis inhibitors.

Aneugenic compounds affect mitosis mechanisms such asmicrotubules synthesis and cytoeskeleton morphology. Thesechanges could lead to aneuploidy in daughter cells (abnormalchromosomal number), cell cycle deregulation and so on, ratherthan actual DNA damage. These compounds are included here inthe non-genotoxic carcinogenic class. In this study, colchicine andnocodazole were used as model compounds.

Artifactual in vitro compounds typically show a positiveresponse in some in vitro genotoxicity assays but are not genotoxicin vivo. False positive responses are commonly produced when

toxic concentrations are tested [4,17]. Here, 2,4-dichlorophenol andd,l-menthol were used as examples. Negative in vitro genotoxiccompounds selected have been extensively tested, consistently giv-ing negative results for genotoxicity or carcinogenicity both in vitro

C. Garcia-Canton et al. / Mutation Research 757 (2013) 158– 166 161

F ompoa ody. Ha

afc

bga

3

o2cos

drqteltYsief�t

ig. 1. Example of �H2AX detection after 3 h treatment with one of the genotoxic c combination of primary anti-�H2AX and secondary fluorophore-conjugated antibnd orange overlay the rejected nuclei (orange arrow).

nd in vivo [17,18]. Acetonitrile, 3-amino-1,2,4-triazole, dimethylormamide, ethylene glycol, d-mannitol and sodium chloride werehosen as model negative compounds for this work.

The sensitivity and specificity of the in vitro �H2AX assay haseen calculated by the number of in vitro genotoxic and non-enotoxic compounds that the assay could discriminate during thisssessment as previously defined [17,28].

.1. In vitro �H2AX assay assessment

Table 2 summarised the results obtained during the assessmentf the in vitro �H2AX assay by HCS performed in this study on BEAS-B cells using Smart et al. evaluation criteria (Table 1) [12]. Foromparison purposes we have included published data availablen regulatory in vitro genotoxicity tests as well as other �H2AXtudies where these control compounds have been used.

Fig. 2 illustrates a summary of the graphical results obtaineduring the assessment of the in vitro �H2AX assay. The dataepresented is expressed as mean fluorescence (�H2AX fre-uency) ± standard deviation (SD) of three separate experimentso show the reproducibility between experiments. The RCC isxpressed as a percentage of the vehicle-treated control calcu-ated for the three separate experiments. The X-axis representshe compound concentration (�M) in logarithmic scale, the left-axis represents the �H2AX frequency in absolute units (inten-ity units) and the right Y-axis represents the cell viability as RCCn percentage relative to vehicle-treated control. In those selected

xamples, the direct-DNA damaging compound, Ethyl methanesul-onate (Fig. 2A) produced a statistically significant increase in theH2AX frequency compared to vehicle-treated control at concen-rations above 500 �M after treatment for 24 h without a reduction

unds, etoposide. The nuclei of the cells were stained with Hoechst and �H2AX withoechst micrographs indicate with a green overlay the scored nuclei (green arrow)

in cell viability below 80%. Ethyl methanesulfonate is typically usedin DNA damage and repair studies for its mutagenic activity [39].In contrast, the treatment with the pro-toxicant B(a)P in the pres-ence of aroclor-induced rat S9 mix (Fig. 2B) produced a significantincrease in �H2AX frequency at concentrations above 8 �M at 3 h(RCC > 80%) and concentrations above 16 �M at 3 h followed by 24 hrecovery period (RCC > 60%). In the absence of S9 mix B(a)P did notproduce any increase in �H2AX frequency above vehicle-treatedcontrols at either timepoint.

B(a)P is a polycyclic aromatic hydrocarbon present in cigarettesmoke which is oxidised by cytochrome P450 enzymes, primarilyCYP1A1 but also CYP1A2 and CYP1B1.

Treatment for 3 h with ROS generator, bleomycin sulphate(Fig. 2C), produced a statistically significant increase in �H2AXfrequency at concentrations above 0.3 �M without a reduction incell viability (RCC > 95%). However, treatment for 24 h producedan increase in �H2AX frequency above 1.5-fold at concentrationsabove 0.02 �M with a dose-related decrease in cell viability (RCCfrom 65% to 33%).

Both, topoisomerase and DNA synthesis inhibitors, etoposideand zidovudine, produced a gentoxic response in the in vitro �H2AXassay. Etoposide treatment (Fig. 2D) produced a positive responseafter 3 h at concentrations above 7.8 �M without a reduction incell viability below 80%. The range of concentrations used for the24 h treatments was lower as high cytotoxicity was observed in theinitial range finder (data not shown). Treatment with etoposide for24 h produced a positive response at concentrations above 0.24 �M

with a cell viability of 75%. A dose-related decrease in cell viabilitywas observed after 24 h treatment when etoposide concentrationincreased (Supplementary data). Contrarily, only treatment for 24 hwith zidovudine (Fig. 2E) produced a significant increase in �H2AX

162 C. Garcia-Canton et al. / Mutation Research 757 (2013) 158– 166

1 0 1 00

1 00 0

0

2 0

4 0

6 0

8 0

1 0 0

0

2 0

4 0

6 0

8 0

1 0 0

1 2 0

E th y l m e th a n e s u lfo n a tec o n c e n tra t io n (µ M)

γH2 A

XFrequency

(Intensi tyUnits) R

CC(%

contro

l)

V eh ic le

#

1 0 1 00

1 00 0

0

5 0

1 0 0

1 5 0

2 0 0

0

2 0

4 0

6 0

8 0

1 0 0

1 2 0

Z id o v u d in ec o n c e n tra t io n (µ M)

γH2A

XFrequency

(IntensityUnits) R

CC(%

contr o

l )

V eh ic le

#

1 1 0 1 00

0

2 0

4 0

6 0

8 0

1 0 0

0

2 0

4 0

6 0

8 0

1 0 0

1 2 0

B (a )Pc o n c e n tra t io n ( µ M )

γH2A

XFrequenc y

(Intensi tyUnits) R

CC(%

contro

l)

V eh ic le

#

*

1 0 1 00

1 00 0

0

2 0

4 0

6 0

8 0

1 0 0

0

2 0

4 0

6 0

8 0

1 0 0

1 2 0

C o lc h ic in ec o n c e n tra t io n (µ M )

γH2A

XFrequency

(IntensityUnits) R

CC(%

contro

l)

V eh ic le

#

a

0 .1 1 1 0 1 0

00

1 0 0

2 0 0

3 0 0

4 0 0

5 0 0

6 0 0

0

2 0

4 0

6 0

8 0

1 0 0

1 2 0

B le o m y c in s u lfa tec o n c e n tra t io n (µ M )

γH2A

XFrequency

(IntensityUnits) R

CC(%

cont ro

l)

V eh ic le

#

*

1 0 1 00

1 00 0

0

5 0

1 0 0

1 5 0

2 0 0

0

2 0

4 0

6 0

8 0

1 0 0

1 2 0

2 ,4 -D ic h lo ro p h e n o lc o n c e n tra t io n (µ M )

γH2 A

XFrequency

(IntensityUnits) R

CC(%

contro

l)

V eh ic le

*#

0 .1 1 1 0 1 0

01 00 0

0

1 0 0

2 0 0

3 0 0

4 0 0

5 0 0

6 0 0

0

2 0

4 0

6 0

8 0

1 0 0

1 2 0

E to p o s id ec o n c e n tra t io n (µ M)

γH2A

XFr equenc y

(IntensityUnits) R

CC(%

contro

l)

V eh ic le

*

#

1 0 1 00

1 00 0

0

2 0

4 0

6 0

8 0

1 0 0

0

2 0

4 0

6 0

8 0

1 0 0

1 2 0

D -M a n n ito lc o n c e n tra t io n (µ M)

γH2 A

XFrequency

(Intensi tyUnits) R

CC( %

contro

l)

V eh ic le

A

B

C

D

E

F

G

H

Fig. 2. Example data from the assessment of the in vitro �H2AX assay. Left Y-axis represents �H2AX frequency (continuous line) and right Y-axis represents RCC (% control)(dotted line). Compounds: [A] ethylmethane sulfonate, [B] B(a)P, [C] bleomycin sulphate, [D] etoposide, [E] zidovudine, [F] colchicine, [G] 2,4-dichlorophenol and [H] d-mannitol. Triangle (–�–) represents short treatment and circle (–�–) represents long treatment. Asterisk (*) indicates the minimum concentration that shows genotoxicitycompared to the vehicle treated control after short treatment. Hash (#) indicates the minimum concentration that shows genotoxicity compared to the vehicle treated controlafter long treatment.

C. Garcia-Canton et al. / Mutation Research 757 (2013) 158– 166 163

Table 2Control compounds tested in the in vitro �H2AX assay by High Content Screening in 96-well plates, and data from published data on regulatory in vitro genotoxicity assaysand other �H2AX studies. Positive results (+), negative (−), equivocal (±), cytotoxic (C).

Compound (CAS number) IARC Group Regulatory in vitro tests [29–32] Other �H2AX �H2AX by HCS

Ames In vitro mammalian test

Direct DNA-damaging compoundsEthyl methanesulfonate (62–50-0) 2B + + + [12] + (500 �M)Mitomycin C (50-07-7) 2B ± + + [12] + (0.47 �M)MNNG (70-25-7) 2A + + + [33] + (0.49 �M)

Pro-toxicant compoundsBenzo(a)pyrene (50-32-8) 1 + + + [12] + (7.8 �M)Cyclophosphamide (6055-19-2) 1 + + + [12] + (1000 �M)2-Acetylaminofluorene (53-96-3) N/A + + + [12] –Aflatoxin B1 (1162-65-8) 1 + + + [34] + (4 �M)

Reactive oxygen species generatorBleomycin sulfate (9041-93-4) 2B + + + [35] + (0.019 �M)

Topoisomerase inhibitorsAmsacrine (m-AMSA) (54301-15-4) 2B + + N/A + (0.2 �M)Etoposide (33419-42-0) 1 – + + [36] + (0.24 �M)

Nucleotide/DNA synthesis inhibitorsAphidicolin (38966-21-1) N/A – + + [37] + (1.56 �M)Zidovudine (30516-87-1) 2B + + N/A + (125 �M)

Aneugen compoundsColchicine (64-86-8) N/A – ± N/A + (7.8 �M)Nocodazole (31430-18-9) N/A + + [38] C

Artifactual in vitro genotoxic positives (false positive)2,4-Dichlorophenol (120-83-2) N/A ± ± N/A + (1000 �M)d,l-Menthol (15356-70-4) N/A – ± N/A –

In vitro genotoxic negativesAcetonitrile (75-05-8) N/A – ± N/A –3-Amino-1,2,4-triazole (61-82-5) 3 – – N/A –Dimethyl formamide (68-12-2) 3 – – N/A –

fa

nc7w2r

ed

TP

Ethylene glycol (107-21-1) N/A –

d-Mannitol (69-65-8) N/A –

Sodium chloride (7647-14-5) N/A –

requency at concentrations above 125 �M with a cell viabilitybove 65%.

The aneugenic compound, colchicine (Fig. 2F), produced a sig-ificant increase in �H2AX frequency compared to vehicle-treatedontrols after 24 h continuous treatment at concentrations above.8 �M (RCC > 25%). However, as the concentration increased, thereas a reduction in �H2AX frequency and at concentrations above

50 �M the �H2AX frequency was equivalent to background

esponse (<1.5-fold).

The artifactual genotoxic compound 2,4-dichlorophenol gen-rated a significant increase in �H2AX frequency at the highestose tested (1000 �M) after both 3 and 24 h treatments (Fig. 2G).

able 3erformance results based on the 22 compounds tested in this study.

Results Ames test

Aneugens considered genotoxic compoundsNo. of genotoxic compounds clearly positive (A) 9

No. of genotoxic compounds tested (B) 14

Sensitivity (%) 64%

No. of non-genotoxic compounds clearly negative (C) 7

No. of non-genotoxic compounds tested (D) 8

Specificity (%) 88%

Accuracy (%) 73%

Aneugens considered non-genotoxicNo. of genotoxic compounds clearly positive (A) 9

No. of genotoxic compounds tested (B) 12

Sensitivity (%) 75%

No. of non-genotoxic compounds clearly negative (C) 8

No. of non-genotoxic compounds tested (D) 10

Specificity (%) 80%

Accuracy 77%

– N/A –– –[12] –– –[36] –

There was no significant reduction in cell viability after the3 h treatment (RCC > 95%), however, after 24 h treatment, therewas a dose-related reduction in cell viability from 100% to 40%RCC.

Treatments with the non-genotoxic d-mannitol (Fig. 2H) did notproduce any significant increase in �H2AX frequency compared tovehicle-treated controls at any timepoint. The cell viability was alsomaintained at around 100% RCC at all timepoints.

Overall, the majority of tested compounds, both in vitro geno-toxic and non-genotoxic, produced the expected increment in�H2AX frequency (above and below 1.5-fold increase respectively).All the results are graphically represented in Supplementary data

In vitro Mammalian test �H2AX by HCS

13 1214 1493% 86%

5 78 8

63% 88%82% 86%

12 1112 12

100% 92%5 8

10 1050% 80%77% 86%

1 tion R

am

3

wt

cgactnagg

io

rlcrlc

4

pidaanimemaa

idcmccstmwlbedrowh

pt

64 C. Garcia-Canton et al. / Muta

nd discussed in more detail in the in vitro �H2AX assay perfor-ance section.

.2. In vitro �H2AX assay performance

The results obtained in the in vitro �H2AX assay assessmentere used to calculate the sensitivity, specificity and accuracy of

his single assay to predict in vivo genotoxicity [28].The sensitivity (or ability to detect genotoxic compounds) cal-

ulated for this preliminary assessment showed that 91.6% ofenotoxic compounds could be detected if aneugenic compoundsre considered non-genotoxic. When aneugenic compounds wereonsidered as indirect-acting genotoxic compounds, the sensi-ivity fell to 85.7%. For specificity (or ability to discriminateon-genotoxic compounds), this preliminary assessment indicatedn 80% specificity if aneugenic compounds are considered non-enotoxic or 87.5% if aneugenic compounds are considered to haveenotoxic potential.

The accuracy of the in vitro �H2AX assay by HCS in this prelim-nary assessment is therefore, calculated to be 86.4%, independentf the aneugenic compound classification.

To put our performance results in context with existing in vitroegulatory genotoxicity tests, we have evaluated previously pub-ished data for the 22 compounds tested in this study (Table 2) toalculated the sensitivity, specificity and accuracy of the currentegulatory in vitro tests. The results in Table 3 have been calcu-ated considering aneugens as both genotoxic and non-genotoxicompounds.

. Discussion and conclusions

The in vitro �H2AX assay has been previously proposed as aotential complement to the existing battery of in vitro genotoxic-

ty assays [40]. Although, measuring �H2AX as a biomarker of DNAamage is commonly used in current scientific research, there is

limited number of studies focused on assessing the sensitivitynd specificity of the in vitro �H2AX assay [12,13,36]. More data iseeded to support its move from emerging to maturing technology

n genetic toxicology testing as described previously [7]. To date, theain system employed to measure �H2AX has been flow cytom-

try [12,13]. Here, we present the first study where automatedicroscopy-based HCS in multi-well microplates has been used to

ssess the sensitivity, specificity and accuracy of the in vitro �H2AXssay.

Although flow cytometry incorporates high throughput qual-tative scoring, the sensitivity of detection for �H2AX has beenescribed to be lower than in microscopy [14]. Additionally, flowytometry scoring requires cells to be in suspension; however, theajority of cell lines used in in vitro toxicology testing are adherent

ells. The requirement for a cell suspension increases the time andomplexity of the methodology as a trypsin-mediated detachmenttep will be required. Moreover, the use of trypsin has been showno affect the cell viability and can cause DNA damage during detach-

ent [41]. In contrast, HCS applies the sensitivity of microscopyith an automated high-throughput system for acquiring cellu-

ar images and software to aid with the qualitative scoring. Otherenefits include the scoring of samples in situ in multi-well plates,limination of the detachment step and storage of images as rawata for quality control purposes. HCS methods could also incorpo-ate cell cycle information if required by measuring DNA contentr multiplexing nuclei dyes [42,43]. An overview of the technologyith current capabilities, limitations and potential developments

as been published by Bickle [11].

The genotoxicity assessment of the in vitro �H2AX assayerformed in this study involved the treatment of the pheno-ypically normal cell line BEAS-2B to a panel of genotoxic and

esearch 757 (2013) 158– 166

non-genotoxic compounds following ICH guidelines [3]. One ofthe major limitations of current in vitro assays is the rate of“false” or misleading positives, non-carcinogenic compounds thatproduce a positive response in vitro [4]. In this study, all the well-established in vitro genotoxic negative compounds tested produceda negative response at both treatment timepoints (3 and 24 h).Of the two previously reported artifactual in vitro genotoxic posi-tive compounds, d,l-menthol produced a clear negative result. Bycontrast, 2,4-dichlorophenol produced a positive response at thehighest concentration tested after both 3 and 24 h treatments, how-ever, this statistically significant increase in �H2AX frequency wasonly achieved in one experiment. The same inconsistent responsehas been reported for 2,4-dichlorophenol in other in vitro geno-toxicity studies and summarised recently [20]. Although thereis no evidence for carcinogenicity of 2,4-dichlorophenol in vivo,there have been some studies indicating endocrine disruption inthe zebra fish model [44]. The hormonal imbalance caused by2,4-dichlorophenol could potentially elevate the levels of estro-gens, an endogenous genotoxic agent. These events suggest that2,4-dichlorophenol have an indirect DNA-damage effect caused byhormonal imbalance [45,46].

The acceptance criteria described for the �H2AX assay by flowcytometry [12] were applied to our data to evaluate the relevanceof using the same approach for the �H2AX assay by HCS. Ourdata showed that all increments in �H2AX frequencies over thevehicle-treated control observed were also identified by the 1.5-fold increase “cut off” criterion defined by Smart and colleagues.The RCC levels selected as a measure of cytotoxicity was the secondcriterion employed to monitor that increases in �H2AX frequencieswere not caused solely by a toxic effect.

The performance of the in vitro �H2AX assay by HCS to detectdirect-acting genotoxic compounds showed a high sensitivity. Onlythe pro-toxicant 2-acetylaminofluorene (2AAF) failed to show apositive response. 2AAF is an aromatic amine not classified byIARC regarding its carcinogenicity. However, the IARC monographdescribing some aromatic amines refers to its mechanism of actionas formation of DNA-adducts by different metabolites [47]. 2AAFis mainly activated by CYP1A2 present in the standard rat S9 mix.Although, in this study 2AAF show a dose-related increase in �H2AXfrequency, the response was never greater than 1.5-fold. In a pre-vious study, Zhou and colleagues reported that the ability of 2AAFto induce DSBs depends on the cell type tested [48]. They obtaineda positive response at 90 �M in Chinese hamster CHL cells aftershort treatments only, while no positive response was observed inhuman ambion FL cells at any timepoint. Similarly, Tsamou et al.report a positive response at 50 �M after 24 h treatment in HepG2cells, known to have a degree of metabolic competency [13]. Smartet al. report a positive response at a concentration above 125 �Mafter 3 h treatment in the presence of 1% (v/v) of rat S9 mix [12].In our experiments, we tested 2AAF up to a concentration of only125 �M as precipitation was observed at higher doses. However,in our experimental design for the testing of pro-toxicants, thelimiting factor was the inherent toxicity produced by the rat S9to the cells. At the minimum concentration of 1% (v/v) S9, typi-cally used in vitro, the cell viability fell below 55%, which is notcompatible with robust testing for DNA damage. The reduction inconcentration to 0.5% (v/v) S9 maintained the cell viability above80% but could have compromised the metabolic activation of 2AAF.

In the case of indirect-acting genotoxic compounds, the in vitro�H2AX assay performance was also highly sensitive. The cellulartargets affected by these compounds are essential for the DNA syn-thesis and repair, both functions are intrinsic of the cell cycle. All the

indirect-acting genotoxic compounds tested in this study produceda positive response in the in vitro �H2AX assay.

During the in vitro �H2AX assay assessment, we also testedtwo aneugenic compounds. We define these as non-genotoxic

tion R

cpmiaceC2recc[atcmbBlfaed

ft(mpaw�sgIfimeHps

tetatuiaatf

F

mioBCr

[

[

[

[

[

[

[

[

C. Garcia-Canton et al. / Muta

arcinogens as they can cause aneuploidy [49]. Aneugenic com-ounds do not damage DNA directly, instead, they affect cellularechanisms involve in the cell cycle process to the point of forc-

ng the cell to programmed death, i.e., apoptosis [50]. In thisssessment, nocodazole produced a genotoxic response at highlyytotoxic concentrations (RCC < 25%) after 24 h treatment. Thisffect is considered an artifactual genotoxic positive result [51].olchicine, on the other hand, produced a genotoxic response after4 h treatment with cell viability around 40%. These contrastingesults could be influence by the different levels of cytotoxicityxerted by the compounds. In general, treatments with aneugenicompounds could lead to DNA damage by products liberated to theulture media during cell death and other mechanisms of toxicity51]. Therefore, a small degree of clastogenicity could be observedfter treatment with aneugenic compounds. This effect could leado an increase in �H2AX frequency above the vehicle-treatedontrol when aneugens are tested, and potentially an additionalechanism of indirect DNA damage. Increases in �H2AX have

een associated to other mechanisms such as apoptosis [52,53].riefly, during a late-stage event of apoptosis, fragmentation of cel-

ular DNA could produce an increase in the frequency of �H2AXollowed by a reduction of both �H2AX levels and cell viabilitys a result of the programmed cell death. Thus, it is consid-red important to report both �H2AX frequency and cell viabilityata.

The performances of other in vitro regulatory assays calculatedrom the panel of compounds tested in this study (Table 3), showedhat the Ames test is better at discriminating negative compoundsspecificity). Other in vitro mammalian tests (i.e., Micronucleus test,

ouse lymphoma test, etc.) are better at detecting positive com-ounds (sensitivity). Our observations concur with more extensivessessments such as the evaluation carried out by Kirkland et al.ith over 800 compounds [28]. The performance of the in vitroH2AX assay by HCS obtained in this preliminary assessmenthows a consistently high level of both sensitivity and specificityiving as a result a greater accuracy than the other in vitro tests.t is as yet not possible to compare the sensitivity and speci-city of the in vitro �H2AX assay by HCS with previously reportedethods such as flow cytometry. Only a limited subset of the ref-

rence compounds have been used across all studies (Table 2).owever, it would be interesting for future studies to perform aarallel comparison between HCS and flow cytometry using theame experimental conditions, e.g., cell lines.

Overall, the in vitro �H2AX assay by HCS results obtained duringhis preliminary assessment are in line with previously publishedvaluations [12,13,54]. However, we believe that the use of HCS forhe scoring of micro-well plates and detecting �H2AX may offerdvantages in terms of higher throughput and sensitivity. Initially,he reproducibility and accuracy of the data gathered supports these of the in vitro �H2AX assay as a complementary tool to the exist-

ng battery of genotoxicity assays, principally as a pre-screeningssay. Nevertheless, a more comprehensive assessment including

more extensive set of compounds would be required to confirmhe performance of the in vitro �H2AX assay by HCS and to provideurther support for future validation trials.

unding

This work was supported by Group Research and Develop-ent of British American Tobacco (Investments) Ltd. as part of

ts research programme focusing on reducing the health impact

f tobacco use. C. Garcia-Canton and C. Meredith are employees ofritish American Tobacco. A. Anadón is employee of the Universityomplutense of Madrid and has not received any funding for thisesearch.

[

esearch 757 (2013) 158– 166 165

Conflict of interest statement

None.

Acknowledgements

The authors thank Mrs. A. Banerjee for her technical supportand Dr. E. Minet for his feedback during the preparation of thismanuscript.

Appendix A. Supplementary data

Supplementary data associated with this article can befound, in the online version, at http://dx.doi.org/10.1016/j.mrgentox.2013.08.002.

References

[1] B. Mahadevan, R.D. Snyder, M.D. Waters, R.D. Benz, R.A. Kemper, R.R. Tice,A.M. Richard, Genetic toxicology in the 21st century: reflections and futuredirections, Environ. Mol. Mutagen. 52 (2011) 339–354.

[2] COM, A Strategy for Testing of Chemicals for Genotoxicity, 2011 http://www.iacom.org.uk/guidstate/documents/COMGuidanceFINAL2.pdf

[3] ICH, Topic S2(R1), Guidance on Genotoxicity Testing and Data Interpreta-tion for Pharmaceuticals Intended for Human Use, 2011 http://www.ich.org/fileadmin/Public Web Site/ICH Products/Guidelines/Safety/S2 R1/Step4/S2R1 Step4.pdf

[4] D. Kirkland, S. Pfuhler, D. Tweats, M. Aardema, R. Corvi, F. Darroudi, A. Elhajouji,H. Glatt, P. Hastwell, M. Hayashi, P. Kasper, S. Kirchner, A. Lynch, D. Marzin, D.Maurici, J.R. Meunier, L. Muller, G. Nohynek, J. Parry, E. Parry, V. Thybaud, R.Tice, J. van Benthem, P. Vanparys, P. White, How to reduce false positive resultswhen undertaking in vitro genotoxicity testing and thus avoid unnecessaryfollow-up animal tests: report of an ECVAM Workshop, Mutat. Res. 628 (2007)31–55.

[5] European Commission, Directive 2003/15/EC of the European Parliament andof the Council of 27 February 2003 amending Council Directive 76/768/EECon the approximation of the laws of the Member States relating to cosmeticproducts, Off. J. Eur. Union L66 (2003) 26–35.

[6] D. Kirkland, M. Aardema, N. Banduhn, P. Carmichael, R. Fautz, J.R. Meunier, S.Pfuhler, In vitro approaches to develop weight of evidence (WoE) and mode ofaction (MoA) discussions with positive in vitro genotoxicity results, Mutagen-esis 22 (2007) 161–175.

[7] A.M. Lynch, J.C. Sasaki, R. Elespuru, D. Jacobson-Kram, V. Thybaud, M. De Boeck,M.J. Aardema, J. Aubrecht, R. Benz, S.D. Dertinger, G.R. Douglas, P.A. White, P.A.Escobar, A. Fornace Jr., M. Honma, R.T. Naven, J.F. Rusling, R.H. Schiestl, R.M.Walmsley, E. Yamamura, J. van Benthem, J.H. Kim, New and emerging technolo-gies for genetic toxicity testing, Environ. Mol. Mutagen. 52 (2011) 205–223.

[8] E.P. Rogakou, D.R. Pilch, A.H. Orr, V.S. Ivanova, W.M. Bonner, DNAdouble-stranded breaks induce histone H2AX phosphorylation on serine 139,J. Biol. Chem. 273 (1998) 5858–5868.

[9] J.S. Dickey, C.E. Redon, A.J. Nakamura, B.J. Baird, O.A. Sedelnikova, W.M. Bonner,H2AX: functional roles and potential applications, Chromosoma 118 (2009)683–692.

10] C.E. Redon, A.J. Nakamura, O.A. Martin, P.R. Parekh, U.S. Weyemi, W.M. Bon-ner, Recent developments in the use of gamma-H2AX as a quantitative DNAdouble-strand break biomarker, Aging 3 (2011) 168–174.

11] M. Bickle, The beautiful cell: high-content screening in drug discovery, Anal.Bioanal. Chem. 398 (2010) 219–226.

12] D.J. Smart, K.P. Ahmedi, J.S. Harvey, A.M. Lynch, Genotoxicity screening via thegammaH2AX by flow assay, Mutat. Res. 715 (2011) 25–31.

13] M. Tsamou, D.G. Jennen, S.M. Claessen, C. Magkoufopoulou, J.C. Kleinjans, J. H.van Delft. Performance of in vitro gammaH2AX assay in HepG2 cells to predictin vivo genotoxicity, Mutagenesis 27 (2012) 645–652.

14] J.P. Banath, P.L. Olive, Expression of phosphorylated histone H2AX as a surro-gate of cell killing by drugs that create DNA double-strand breaks, Cancer Res.63 (2003) 4347–4350.

15] O.A. Sedelnikova, E.P. Rogakou, I.G. Panyutin, W.M. Bonner, Quantitativedetection of (125)IdU-induced DNA double-strand breaks with gamma-H2AXantibody, Radiat. Res. 158 (2002) 486–492.

16] D.L. Taylor, A personal perspective on high-content screening (HCS): from thebeginning, J. Biomol. Screen. 15 (2010) 720–725.

17] D. Kirkland, P. Kasper, L. Muller, R. Corvi, G. Speit, Recommended lists of geno-toxic and non-genotoxic chemicals for assessment of the performance of newor improved genotoxicity tests: a follow-up to an ECVAM workshop, Mutat.

Res. 653 (2008) 99–108.

18] P.W. Hastwell, L.L. Chai, K.J. Roberts, T.W. Webster, J.S. Harvey, R.W. Rees, R.M.Walmsley, High-specificity and high-sensitivity genotoxicity assessment in ahuman cell line: validation of the GreenScreen HC GADD45a-GFP genotoxicityassay, Mutat. Res. 607 (2006) 160–175.

1 tion R

[

[

[

[[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

66 C. Garcia-Canton et al. / Muta

19] W.M. Bonner, C.E. Redon, J.S. Dickey, A.J. Nakamura, O.A. Sedelnikova, S. Solier,Y. Pommier, GammaH2AX and cancer, Nat. Rev. Cancer 8 (2008) 957–967.

20] P. Fowler, K. Smith, J. Young, L. Jeffrey, D. Kirkland, S. Pfuhler, P. Carmichael,Reduction of misleading (“false”) positive results in mammalian cell genotox-icity assays. I. Choice of cell type, Mutat. Res. 742 (2012) 11–25.

21] R.R. Reddel, Y. Ke, B.I. Gerwin, M.G. McMenamin, J.F. Lechner, R.T. Su, D.E. Brash,J.B. Park, J.S. Rhim, C.C. Harris, Transformation of human bronchial epithelialcells by infection with SV40 or adenovirus-12 SV40 hybrid virus, or transfectionvia strontium phosphate coprecipitation with a plasmid containing SV40 earlyregion genes, Cancer Res. 48 (1988) 1904–1909.

22] D. Hanahan, R.A. Weinberg, The hallmarks of cancer, Cell 100 (2000) 57–70.23] O. Fernandez-Capetillo, A. Lee, M. Nussenzweig, A. Nussenzweig, H2AX: the

histone guardian of the genome, DNA Repair (Amst.) 3 (2004) 959–967.24] M. Downey, D. Durocher, gamma H2AX as a checkpoint maintenance signal,

Cell Cycle 5 (2006) 1376–1381.25] M.P. Svetlova, L.V. Solovjeva, N.V. Tomilin, Mechanism of elimination of phos-

phorylated histone H2AX from chromatin after repair of DNA double-strandbreaks, Mutat. Res. 685 (2010) 54–60.

26] C. Garcia-Canton, E. Minet, A. Anadon, C. Meredith, Metabolic characterizationof cell systems used in in vitro toxicology testing: lung cell system BEAS-2B asa working example, Toxicol. In Vitro 27 (2013) 1719–1727.

27] R.W. Nims, G. Sykes, K. Cottrill, P. Ikonomi, E. Elmore, Short tandem repeatprofiling: part of an overall strategy for reducing the frequency of cell misiden-tification, In Vitro Cell. Dev. Biol. Anim. 46 (2010) 811–819.

28] D. Kirkland, M. Aardema, L. Henderson, L. Muller, Evaluation of the ability of abattery of three in vitro genotoxicity tests to discriminate rodent carcinogensand non-carcinogens. I. Sensitivity, specificity and relative predictivity, Mutat.Res. 584 (2005) 1–256.

29] TOXNET Toxicology Data Network, http://toxnet.nlm.nih.gov (accessedFebruary 2013).

30] NTP National Toxicology Program, http://ntp-server.niehs.nih.gov (accessedFebruary 2013).

31] IARC, IARC monographs on the evaluation of carcinogenic risk to humans. Someantiviral and antineoplastic drugs, and other pharmaceutical agents, Interna-tional Agency for Research on Cancer, 76, 2000.

32] IARC, IARC monographs on the evaluation of carcinogenic risk to humans. Over-all evaluation of carcinogenicity: an updating of IARC monographs, vols. 1–42,suppl. 7, International Agency for Research on Cancer, 1987.

33] O. Staszewski, T. Nikolova, B. Kaina, Kinetics of gamma-H2AX focus forma-tion upon treatment of cells with UV light and alkylating agents, Environ. Mol.Mutagen. 49 (2008) 734–740.

34] O. Gursoy-Yuzugullu, H. Yuzugullu, M. Yilmaz, M. Ozturk, Aflatoxin geno-toxicity is associated with a defective DNA damage response bypassing p53activation, Liver Int. 31 (2011) 561–571.

35] S. Kim, D.H. Jun, H.J. Kim, K.C. Jeong, C.H. Lee, Development of a high-contentscreening method for chemicals modulating DNA damage response, J. Biomol.Screen. 16 (2011) 259–265.

36] K. Matsuzaki, A. Harada, A. Takeiri, K. Tanaka, M. Mishima, Whole cell-ELISA tomeasure the gammaH2AX response of six aneugens and eight DNA-damagingchemicals, Mutat. Res. 700 (2010) 71–79.

37] T. Tanaka, X. Huang, H.D. Halicka, H. Zhao, F. Traganos, A.P. Albino, W.Dai, Z. Darzynkiewicz, Cytometry of ATM activation and histone H2AX

[

esearch 757 (2013) 158– 166

phosphorylation to estimate extent of DNA damage induced by exogenousagents, Cytometry A 71 (2007) 648–661.

38] W.B. Dalton, B. Yu, V.W. Yang, p53 suppresses structural chromosome instabil-ity after mitotic arrest in human cells, Oncogene 29 (2010) 1929–1940.

39] R. Munroe, J. Schimenti, Mutagenesis of mouse embryonic stem cells withethylmethanesulfonate, Methods Mol. Biol. 530 (2009) 131–138.

40] C. Garcia-Canton, A. Anadon, C. Meredith, gammaH2AX as a novel endpoint todetect DNA damage: applications for the assessment of the in vitro genotoxicityof cigarette smoke, Toxicol. In Vitro 26 (2012) 1075–1086.

41] K. Wischermann, P. Boukamp, P. Schmezer, Improved alkaline comet assay pro-tocol for adherent HaCaT keratinocytes to study UVA-induced DNA damage,Mutat. Res. 630 (2007) 122–128.

42] S.K. Lyman, S.C. Crawley, R. Gong, J.I. Adamkewicz, G. McGrath, J.Y. Chew, J.Choi, C.R. Holst, L.H. Goon, S.A. Detmer, J. Vaclavikova, M.E. Gerritsen, R.A. Blake,High-content, high-throughput analysis of cell cycle perturbations induced bythe HSP90 inhibitor XL888, PLoS ONE 6 (2011) e17692.

43] G.K. Chan, T.L. Kleinheinz, D. Peterson, J.G. Moffat, A simple high-content cellcycle assay reveals frequent discrepancies between cell number and ATP andMTS proliferation assays, PLoS ONE 8 (2013) e63583.

44] Y. Ma, J. Han, Y. Guo, P.K. Lam, R.S. Wu, J.P. Giesy, X. Zhang, B. Zhou, Disruption ofendocrine function in in vitro H295R cell-based and in in vivo assay in zebrafishby 2,4-dichlorophenol, Aquat. Toxicol. 106–107 (2012) 173–181.

45] E. Cavalieri, K. Frenkel, J.G. Liehr, E. Rogan, D. Roy, Estrogens as endogenousgenotoxic agents--DNA adducts and mutations, J. Natl. Cancer Inst. Monogr. 27(2000) 75–93.

46] S.M. Amer, F.A. Aly, Genotoxic effect of 2,4-dichlorophenoxy acetic acid and itsmetabolite 2,4-dichlorophenol in mouse, Mutat. Res. 494 (2001) 1–12.

47] IARC, IARC Monographs on the evaluation of carcinogenic risks to humans.Some aromatic amines, organic dyes, and related exposures, 99, 2010.

48] C. Zhou, Z. Li, H. Diao, Y. Yu, W. Zhu, Y. Dai, F.F. Chen, J. Yang, DNA damage eval-uated by gammaH2AX foci formation by a selective group of chemical/physicalstressors, Mutat. Res. 604 (2006) 8–18.

49] M.J. Aardema, S. Albertini, P. Arni, L.M. Henderson, M. Kirsch-Volders, J.M.Mackay, A.M. Sarrif, D.A. Stringer, R.D. Taalman, Aneuploidy: a report of anECETOC task force, Mutat. Res. 410 (1998) 3–79.

50] H.Y. Yamada, G.J. Gorbsky, Spindle checkpoint function and cellular sensitivityto antimitotic drugs, Mol. Cancer Ther. 5 (2006) 2963–2969.

51] D. Scott, S.M. Galloway, R.R. Marshall, M. Ishidate Jr., D. Brusick, J. Ashby, B.C.Myhr, International commission for protection against environmental muta-gens and carcinogens. Genotoxicity under extreme culture conditions. A reportfrom ICPEMC Task Group 9, Mutat. Res. 257 (1991) 147–205.

52] B. Mukherjee, C. Kessinger, J. Kobayashi, B.P. Chen, D.J. Chen, A. Chatterjee, S.Burma, DNA-PK phosphorylates histone H2AX during apoptotic DNA fragmen-tation in mammalian cells, DNA Repair (Amst.) 5 (2006) 575–590.

53] Z. Darzynkiewicz, H. Zhao, H.D. Halicka, P. Rybak, J. Dobrucki, D. Wlodkowic,DNA damage signaling assessed in individual cells in relation to the cellcycle phase and induction of apoptosis, Crit. Rev. Clin. Lab. Sci. 49 (2012)

199–217.

54] M. Audebert, A. Riu, C. Jacques, A. Hillenweck, E.L. Jamin, D. Zalko, J.P. Cravedi,Use of the gammaH2AX assay for assessing the genotoxicity of polycyclicaromatic hydrocarbons in human cell lines, Toxicology Letters 199 (2010)182–192.