Comet Assay measurements: a perspective

T. S. Kumaravel & Barbara Vilhar &

Stephen P. Faux & Awadhesh N. Jha

Received: 16 July 2007 /Accepted: 8 October 2007 / Published online: 27 November 2007# Springer Science + Business Media B.V. 2007

Abstract The Comet Assay or single cell gel elec-trophoresis assay is one of the very widely used assaysto microscopically detect DNA damage at the level of asingle cell. The determination of damage is carried outeither through visual scoring of cells (after classifica-tion into different categories on the basis of tail lengthand shape) or by using different commercially avail-able or public domain software (which automaticallyrecognise the extent of damage). In this assay, theshape, size and amount of DNAwithin the ‘comet’ playimportant roles in the determination of the level ofdamage. The use of a software in particular also pro-vides a range of different parameters, many of whichmight not be relevant in determining the extent of DNAdamage. As a large number of factors could influencethe shape, size, identification and determination ofinduced damage, which includes the scoring criteria,

staining techniques, selection of parameters (whilstusing the software packages) and appearance of‘hedgehog’ or ‘clouds’, this article aims (a) to providean overview of evolution of measurements of DNAdamage using the Comet Assay and (b) to summariseand critically analyse the advantages and disadvan-tages of different approaches currently being adoptedwhilst using this assay. It is suggested that judiciousselection of different parameters, staining methodsalong with inter-laboratory validation and harmonisa-tion of methodologies will further help in making thisassay more robust and widely acceptable for scientificas well as regulatory studies.

Keywords Comet Assay . Single cell gelelectrophoresis (SCGE) . CometAssaymeasurements .

Image analysis . Clouds . % Tail DNA .

Olive tail moment (OTM) . Genetic toxicology

Introduction

The Comet Assay or single cell gel electrophoresis(SCGE) assay is a rapid, sensitive and relatively simplemethod for detecting DNA damage at the level ofindividual cells (Singh et al. 1988; Comet Assay in-terest group website: http://cometassay.com/). It com-bines the simplicity of biochemical techniques fordetecting DNA single strand breaks (strand breaks andincomplete excision repair sites), alkali labile sites andcross-linking, with the single cell approach typical of

Cell Biol Toxicol (2009) 25:53–64DOI 10.1007/s10565-007-9043-9

T. S. Kumaravel (*) : S. P. FauxToxicology Group,Advanced Technologies (Cambridge) Ltd.,210 Cambridge Science Park,Cambridge CB4 0WA, UKe-mail: ts.kumaravel@atcbiotech.com

B. VilharDepartment of Biology, Biotechnical Faculty,University of Ljubljana,Ljubljana, Slovenia

A. N. JhaSchool of Biological Sciences, University of Plymouth,Plymouth PL4 8AA, UK

cytogenetic assays. Several reviews have been pub-lished in recent years to highlight the procedures,advantages and limitations of this assay in genotox-icological, ecotoxicological and biomonitoring studies(Collins 2004; Dixon et al. 2002; Fairbairn et al. 1995;Lee and Steinert 2003). The assay has also been suc-cessfully implemented in plant cells under laboratoryconditions (Gichner et al. 2004, 2006). The main ad-vantages of the Comet Assay include: (a) the collectionof data at the level of the individual cell, allowingmore robust statistical analyses, (b) the need for a smallnumber of cells per sample (<10,000), (c) sensitivity fordetecting DNA damage and (d) use of any eukaryotesingle cell population both in vitro and in vivo,including cells obtained from exposed human popula-tions and aquatic organisms for eco-genotoxicologicalstudies and environmental monitoring (Collins et al.1997a; Dixon et al. 2002; Lee and Steinert 2003; Jha2004). The importance of this assay has also beenrealised in regulatory genotoxicological studies (Ticeet al. 2000; Hartmann et al. 2003; Burlinson et al.2007), and there is a move to replace some traditionalassays (e.g. liver unscheduled DNA synthesis assay) inregulatory genotoxicological studies with in vivoComet assay. In combination with certain bacterialenzymes (e.g. formamidopyrimidine glycosylase,endonuclease III, uracil-DNA glycosylases, etc.),which recognise oxidised purines and pyrimidinebases, this assay has been used to determine oxidativeDNA damage that has been implicated in several healthconditions (Collins et al. 1993, 1997a, b, 2001a, b;Kruman et al. 2002). This assay has also been used toshow protective effects of different dietary factors inchemo-preventive studies (Bichler et al. 2007; Collinset al. 2001a, b). In combination with the fluorescencein situ hybridisation (FISH) technique (Comet-FISH),the application of this assay has also been extended todetermine sequence or gene-specific damage andrepair (Santos et al. 1997; McKenna et al. 2003) aswell as of possible diagnostic use (Kumaravel andBristow 2005). In addition, the assay is being used intranslational research to assess whether tumour radio-sensitivity (Fisher et al. 2007) and chemo-sensitivity(Smith et al. 2007) can be determined. This wouldallow clinicians to individualise patient management,allocating cancer therapy to those for whom it will beof most benefit and reducing the likelihood of patientsreceiving toxic (and as such ineffective) therapy. Inview of its wide applications and uses, based on

PubMed/Web of Science, in the last 10 years, morethan 4,000 papers have been published in peer-reviewed scientific journals, which reflect its popularity.

The Comet Assay is based on the ability of neg-atively charged loops/fragments of DNA to be drawnthrough an agarose gel in response to an electric field.The extent of DNA migration depends directly on theDNA damage present in the cells. It should be notedthat DNA lesions consisting of strand breaks aftertreatment with alkali either alone or in combination withcertain enzymes (e.g. endonucleases) increase DNAmigration, whereas DNA–DNA and DNA–proteincross-links result in retarded DNA migration comparedto those in concurrent controls (Tice et al. 2000).

In this assay, a suspension of cells is mixed withlow melting point agarose and spreads onto a micro-scope glass slide. After lysis of cells with detergent athigh salt concentration, DNA unwinding and electro-phoresis is carried out at a specific pH. Unwinding ofthe DNA and electrophoresis at neutral pH (7–8)predominantly facilitates the detection of doublestrand breaks and cross links; unwinding and electro-phoresis at pH 12.1–12.4 facilitates the detection ofsingle and double strand breaks, incomplete excisionrepair sites and cross-links; whereas unwinding andelectrophoresis at a pH greater than 12.6 expressesalkali labile sites in addition to all types of lesionslisted above (Miyamae et al. 1997). When subjectedto an electric field, the DNA migrates out of the cell,in the direction of the anode, appearing like a ‘comet’.The size and shape of the comet and the distributionof DNA within the comet correlate with the extent ofDNA damage (Fairbairn et al. 1995).

The determination of the shape, size and amount ofDNA within comets is therefore a very importantattribute of the assay if the induced damage is to beevaluated accurately. In parallel with several technicaland procedural evolutions to make the assay morerobust, several approaches have also evolved toquantify the extent of damage more reliably, reproduc-ibly and meaningfully. Such quantification includesboth visual examinations (i.e., photographic, occulo-meter or non-specific image analysis systems) or byuse of commercially available or public domainspecific image analysis software packages. Althoughvisual examinations give a fairly good indication ofDNA damage and can be used in situations that areconsidered appropriate or where specific softwarepackages are not available, the use of specific image

54 Cell Biol Toxicol (2009) 25:53–64

analysis software is considered to be reliable, repro-ducible, which provide simultaneously a range ofparameters and additional information (e.g. the distri-bution of DNA within the comet tail, total cellularDNA content), and may also indicate different phasesof cell cycle distribution, which can be useful in theinterpretation of the data. Such specific softwarepackages also facilitate easy statistical analyses, plot-ting and documentation of the data.

Despite being a very popular choice to determineDNA damage, there are still some concerns over themethodology used, and the type and quality of dataproduced using this assay. Given the importance ofdifferent measurements in determining the extent ofDNA damage (and repair) in this very widely usedassay, this article aims to (a) analyse the develop-ments and our current understanding of differentComet Assay measurements (b) analyse their relativeimportance or use and (c) highlight future develop-ments and perspectives.

Historical perspectives and evolutionof measurement procedures

The Comet Assay was first introduced by Ostling andJohanson in 1984. This was a neutral version of theComet Assay, and interestingly, they used quitesophisticated techniques of image analysis for quanti-fication of the comets, using acridine orange (AO) as

the DNA binding dye (Fig. 1). The fluorescence wasmeasured with a Leitz MPV2 microscope photometerwith a ×40 objective using a Phloemopak filterblockH2 giving excitation at 390–490 nm. The emitted lightfrom individual cells passed an emission filter (longpass 525 nm). Green fluorescence was then measuredusing a circular diaphragm first over the head and thenover different positions on the comet tail. The

Fig. 1 Image analysis used by Ostling and Johanson (1984;modified for clarity). They measured fluorescence (using photo-metre) in the head and in the tail at a distance of 50 μm from thecentre of the head using acridine orange as the DNA binding dye

Table 1 List of various Comet Assay parameters used inpublished literature

Parameters

Cell areaComet coefficient of varianceComet distribution momentComet extentComet inertiaComet meanComet modeComet optical intensityComet skewComet standard deviationHead coefficient of varianceHead distribution momentHead DNAHead extentHead inertiaHead meanHead modeHead optical intensityHead skewHead standard deviationLength/heightOlive tail momentTail coefficient of varianceTail distribution momentTail DNATail extentTail extent momentTail inertiaTail lengthTail meanTail modeTail optical intensityTail skewTail standard deviation

Tail length, Tail DNA and DNA distribution profile in the Tailare primary Comet Assay measurements (obtained by fluores-cent densitometric profiles of the comets). All other measure-ments are derived from the three primary Comet Assaymeasurements (adapted from Kumaravel and Jha 2006).

Cell Biol Toxicol (2009) 25:53–64 55

background fluorescence adjacent to the cells wassubtracted. The time of illumination and betweenmeasurements were standardised to minimise fadingbias. They presented their results in terms of the ratioof fluorescence (Fx) at distance × micrometre on thetail versus fluorescence at the centre of the head (Fo).Based on their observation, they concluded that 50 μmgives the best resolution of the method used. Based onour current understanding however, as mentioned later,this measurement is not considered to be robust.

Singh et al. (1988) developed the alkaline versionof the Comet Assay in which they used the length ofDNA migration (tail length) to quantify the extent ofdamage. Subsequently, several research groups pub-lished papers in which various Comet Assay param-eters were used (Table 1). However, with time, mostof them were not of frequent or wide use. Of notableimportance is the publication by Olive et al. (1990),who used the concept of the tail moment to describeDNA migration. The tail moment calculated by Oliveet al. (1990) came to be known as the Olive tailmoment (OTM). This parameter is considered to beparticularly useful in describing heterogeneity withina cell population, as OTM can pick up variations inDNA distribution within the tail.

Although image analysis on comets has beenpreferable for continuity in assessing DNA damageby this method, some groups have been working onsimple, less time consuming visual scoring methodsthat do not require special image analysis software.Collins et al. (1995) published a visual scoring methodthat classifies comets from grades 0–4 (Fig. 2). In thisapproach, for example, if 100 comets are scored and

each comet assigned a value of 0 to 4 according to itsclass, the total score for the sample gel will be between0 and 400 “arbitrary units.” Visual scoring is rapid aswell as simple and should appeal to those exploring theusefulness of the technique without the need to investin expensive analytical equipment or software pack-ages (Collins 2004).

Another important publication on visual scoringwas by Kobayashi et al. (1995). The scoring systemused by them grouped comets into five stages (Fig. 3),but they did not calculate the total score for each gel.Furthermore, Kobayashi et al. (1995) and Collinset al. (1997a) showed that the results of visual scoringcorrelated very well with image analysis measure-ments. It is interesting to note that visual scoringbased on Collins et al. (1995) is becoming popularespecially in biomonitoring and DNA repair studies.About 70 biomonitoring studies have reported DNAdamage using visual scoring criteria (Moller 2006).Moreover, visual scoring has the potential to be usedfor inter-laboratory comparisons.

Significance of staining procedures in cometmeasurements

Whether the comets are scored by image analysis orvisual scoring, good staining of comets is of para-mount importance. Various fluorochromes, which weretraditionally used to stain DNA, chromosomes ornuclei, are being used to stain the comets (i.e. ‘head’and ‘tail’). Ethidium bromide (EB) is most commonlyused to stain the DNA on Comet Assay slides (Singh

Fig. 2 Visual classificationsuggested by Collins et al.(1995). Images of comets(from lymphocytes), stainedwith DAPI. They representclasses 0–4 as used forvisual scoring

56 Cell Biol Toxicol (2009) 25:53–64

et al. 1988), followed by 4, 6-diamidino-2-phenyl-indole (DAPI, Gedik et al. 1992). EB is an intercalat-ing dye that binds more efficiently to double-strandedDNA than to single-stranded DNA. DAPI bindspredominantly to the major groove of the DNA. Theamount of dye binding to the DNA is proportional tothe amount of DNA present and, hence, the amount oflight emitted after excitation with ultraviolet light ofappropriate wavelength. It is important that we usevery low concentrations of the dye, as higher concen-trations will saturate the system. If the light emitted bythe comets is very intense, the image analysis softwarecannot accurately define certain Comet Assay meas-urements like OTM, because the centre of gravity (CG)of DNA distribution is not defined correctly (Fig. 4).Other dyes used are SYBR® green/gold (Tice et al.1998), AO, YOYO dye (Singh et al. 1994) andpropidium iodide. In addition, non-fluorescent staining,such as silver stain, has also been used by someworkers (Kizilian et al. 1999; Garcia et al. 2007).

For general genotoxicity testing purpose, EB is anexcellent choice of Comet Assay stain. EB produces abright fluorescence, which does not fade easily. Unlikethe SYBR® green stain, EB does not fade during theprocess of image capturing. This gives the option towork in dim light rather than in completely dark rooms.Furthermore, EB at concentrations in the range of

2–20 μg ml−1 gives good quality staining of DNAwith a low background signal. We would thereforerecommend EB as the first choice of stain for highthroughput genetic toxicological studies. YOYO dye

Fig. 4 Schematic diagram explaining how concentration ofDNA binding fluorescent dye may affect Comet Assay measure-ments using image analysis. A A high concentration of DNAbinding dye is used, and the tail fluorescence is fully saturated.B Optimal concentration of DNA binding dye is used and thecenter of gravity of DNA distribution is properly defined

Fig. 3 Schematic of visual classification of comets byKobayashi et al. (1995). They represent comets of Types 1 to 5

Fig. 5 The model image of a comet. The light emitted from acomet on the slide is detected as an image. The image of a realcomet on the microscope slide is shown in (a). A simplified modelimage of comet (b) can be used to demonstrate the measurementof Comet Assay parameters. An image is composed of separatepixels (c). The size of this model image of a comet is 35×25pixels. The images that cameras record, have many pixels (e.g.750×550, 1,300×1,030), so individual pixels cannot be dis-cerned in (a) unless the image is considerably enlarged

Cell Biol Toxicol (2009) 25:53–64 57

gives a strong fluorescence signal and is particularlyuseful when cells with low DNA content are used inthe Comet Assay. However, YOYO dye is expensiveand cannot be stored for longer periods of time.SYBR® green/gold also gives a bright fluorescence,but fading during the process of scoring is definitely aproblem. SYBR® gold stains both double-strandedand single-stranded DNA and is considered betterthan SYBR® green. Although silver staining isconsidered to be cheaper in the sense that it doesnot require use of a fluorescence microscope, itappears to be more time consuming and have a lowerresolution compared to fluorescence staining. It also

requires proper optimisation, as it may incur a lot ofbackground staining.

Principles of image analysis in Comet Assay

It is difficult to ascertain who used commerciallyavailable software for the first time. Presently,different software packages are used for measurementof comet parameters on the basis of image analysis.Although these software packages may differ slightlythe way they calculate DNA damage, the underlyingprinciples of image analysis are the same (Vilhar2004; http://www.botanika.biologija.org/exp/comet/Comet-principles).

The first step is to visualise the comets under thefluorescence microscope fitted with appropriate filters

Fig. 6 Conversion of light intensity to grey values on animage. Information on an image is coded as grey values. Animage of a comet is composed of separate pixels (a), whereeach pixel has a grey value, as shown in (b) for a small imagearea outlined in red in (a). The model image is an 8-bit imagewith available grey values 0–255. The relationship between thegrey value of a pixel on the image and light intensity(fluorescence) that a camera element detects is linear (c)

Fig. 7 Segmentation of the comet image. During the segmen-tation step, the regions of interest (ROI) for measurement ofcomet head and tail are defined. In some software packages, theuser can interactively draw rectangles that define the head andthe tail ROI (a; box segmentation). In others, the head and thetail are detected automatically and outlined (b; close fittingsegmentation)

58 Cell Biol Toxicol (2009) 25:53–64

(depending on the fluorochrome used) and capturethe image using a camera. The camera records theintensity of light emitted from each point on thecomet and converts it into electrical signals. Theseelectrical signals are then sent to the computer alongwith their coordinates, and the computer decodes thesesignals and displays the image on the screen (Fig. 5).An image is composed of small dots called pixels.Each pixel represents one light sensitive element ofthe camera, and the numbers corresponding to light

intensity detected for each pixel (grey values) arestored in a computer file (Fig. 6). Once the image isconverted into numbers as shown in Fig. 6, the actualimage analysis is initiated. The next step is to definethe head and the tail of the comet. Different softwarepackages use different approaches to detect the headand the tail (Fig. 7). Once the ‘head’ and ‘tail’ regionare defined, the tail length is measured in terms ofpixels, which are then converted into microns. Thelight intensities originating from the head and tail

Fig. 8 Grey values forthe head, the tail and thebackground regions of themodel comet image. Boxsegmentation is shown, withhead indicated in blue, thetail in orange and the back-ground in yellow. Thepixel column eight is in bluefont in the table. The pixelcolumn number is indicatedin italics

Cell Biol Toxicol (2009) 25:53–64 59

parts of the comet are used to calculate differentcomet parameters (Fig. 8). The background fluores-cence is subtracted from head and tail intensities toget the true fluorescence as described in Table 2.

Standardisation of Comet Assay measurements

The Comet Assay has mainly remained an assay ofacademic and scientific interest until quite recently.Currently, the Comet Assay has however the potential tobe used as a tool in genotoxicity testing and regulatorysubmissions for new chemicals and mixtures (Tice et al.2000; Hartmann et al. 2003; Kumaravel and Jha 2006).In an attempt to make the assay more sensitive andreliable, several research groups have come out with

unique procedures and specialised measures of DNAmigration. Attempts have been made to make this assaywidely acceptable, by correlating the results with otherwell-established assays (e.g. micronucleus assay) whilstdetermining the genotoxicity (Raisuddin and Jha 2004).Before it can be accepted as a regulatory tool, thisassay has to be harmonised in terms of its methodol-ogy and interpretations and should be demonstrated tobe reliable, accurate and transferable between labora-tories. An expert panel at an International Workshopon Genotoxicity Test Procedures (IWGTP) held inWashington, DC, in 1999, identified minimal experi-mental and methodological standards necessary toensure that the results of the Comet Assay studieswould be acceptable as being informative by knowl-edgeable scientists and regulatory authorities in this

Table 2 Subtraction of the background signal for each pixel column of the comet on the model comet image

The data for the boxed segmentation of the model image are shown in Fig. 8. The data for pixel column eight in Fig. 8 are shown in blue.In this example, a narrow rectangle of 4-pixel height is used as a background (see Fig. 8). The percentage of DNA in the tail is 10,920/29,793=0.37=37%. The background corrected comet, head or tail intensities are used in all further calculations of the comet parameters

60 Cell Biol Toxicol (2009) 25:53–64

field (Tice et al. 2000; Hartmann et al. 2003). Theexpert panel recognised that different methods havebeen used to analyse ‘comets’ in the assay. At theIWGTP, however, the expert panel did not recommendany particular measurement of comet migration to bemore useful than any other measure. However, theexpert panel did recommend that, when using derivedmeasurements (e.g. tail moment), data on primarymeasurements (e.g. tail length and % Tail DNA)should also be presented in the analyses (Tice et al.2000). Similar recommendations were put forwards byHartmann et al. (2003). However, for peer-reviewpublications, normally, only one set of measurements(e. g. either tail moment or % Tail DNA) is presented.

Kumaravel and Jha (2006) defined the most reliablecomet measurements that would truly reflect the extentof DNA damage induced by low linear energy transferionising radiation. The authors approached this ques-tion by performing alkaline Comet Assay on humanperipheral blood lymphocytes irradiated with gradeddoses of 137Cs gamma radiation and correlating thevarious comet measurements with the radiation dose.As DNA damage produced is directly proportional tothe radiation dose, any change in dose should bereflected in proportional change in the comet measure-ments. They concluded that only a few comet measure-ments provided by the image analysis softwarecorrelated well with gamma radiation dose. Furtherretrospective analysis from in vitro and in vivo experi-ments using chemicals also suggested that OTM and %Tail DNA gave good correlations with the dose ofgenotoxic agents used and were the most reliablecomet measurements. Statistically, the authors did notfind much difference between OTM and % Tail DNAin analysing extent of DNA damage.

Although OTM appeared to be the most statisti-cally significant measurement (Kumaravel and Jha2006), the inter-laboratory comparison of resultsseems to be difficult for this parameter. OTM iscalculated as a product of two factors: the percentageof DNA in the tail (%Tail DNA) and the distancebetween the intensity centroids (centres of gravity) ofthe head and the tail along the x-axis of the comet.Hence, OTM is an absolute parameter with ameasurement unit μM (Vilhar 2004; http://www.botanika.biologija.org/exp/comet/Comet-principles).This requires that the image analysis system isgeometrically calibrated before comet measurement(i.e. the number of pixels per micrometre is known for

different microscope objectives). If the system is notcalibrated, inter-laboratory comparisons are difficult.On the other hand, many published reports quoteOTM values without a unit (micrometre), whichmakes inter-laboratory comparisons impossible. Inaddition, OTM calculation includes the distancebetween the intensity centroids of the head and thetail, which depends upon conditions of electrophore-sis (e.g. electrophoresis time), and algorithms used todefine the CG of DNA distribution vary amongdifferent software packages. Under these circum-stances, it is advisable to use % Tail DNA forregulatory purposes and for inter-laboratory compar-isons. In addition, for studies involving multipleelectrophoresis runs, the % Tail DNA, rather thanOTM, would be a better descriptor of DNA damagefor all the reasons given above. Kumaravel and Jha(2006) recommended that, for scientific purposes,both OTM and % Tail DNA could be used. Based onthese findings, the Fourth IWGT also recommendedthe use of % Tail DNA for regulatory studies(Burlinson et al. 2007). The Japanese Centre for theValidation of Alternative Methods (JaCVAM) are alsorecommending % Tail DNA for their inter-laboratoryComet Assay trials. As JaCVAM initiative aims tofacilitate Organization for Economic Cooperation andDevelopment acceptance of Comet Assay as a regula-tory tool, this parameter (i.e. % Tail DNA) would playan important role in harmonisation of the laboratoryprotocols and inter-laboratory comparison of the data(Burlinson B, communication at the 7th InternationalComet Assay Workshop, Coleraine, UK, 2007).

Identification and measuring the ‘Clouds’

‘Clouds’ or ‘hedgehogs’ are important observationsin most Comet Assay experiments. These are cellswith extensive DNA migration that are outside themeasurement capabilities of the image analysissystem or may give inappropriate measurementswhen image analysis is used. Clouds are thereforescored only by visual analysis. They are character-ised by a small or absent head with a highly diffusedtail that is physically separate from the head.Accurate identification of clouds comes with prac-tice and experience; hence, it is important that cloudsare identified correctly whilst scoring slides usingimage analysis systems. Ideally, they should be

Cell Biol Toxicol (2009) 25:53–64 61

scored visually and recorded alongside the results ofimage analysis.

The exact origin of clouds is not clear, but it isassumed that apoptotic cells lead to clouds. This hasbeen observed in Rat-1 cells exposed to irradiationwith 10 Gy gamma rays, where the cells underwentapoptosis 24 h after irradiation and produced cloudsin Comet Assay (Kumaravel TS; unpublished obser-vations). Interestingly, cells treated with hydrogenperoxide for approximately 5 min and processedimmediately (assuming that apoptotic process cannotbe initiated and completed in 5 min) also gave rise toclouds. Moreover, cells treated with high doses ofgamma radiations and processed immediately gave asimilar response. Clouds are routinely seen in in vitroand in vivo Comet Assay experiments (particularlywhere cells are collected by scraping, e.g. stomach).These observations suggest that, in addition toapoptosis, clouds are also induced by high levels ofDNA damage as well as in necrotic cells. It is alsoobserved that, sometimes, identical experimentalconditions can result either in measurable comets orin clouds. This phenomenon is particularly seen intreatment with hydrogen peroxide and methyl meth-anesulphonate (Collins 2004; Speit et al. 2004) whenused in conjunction with enzymes such as endonu-cleases. To confirm whether clouds really representapoptotic cells, more experiments under differentexposure conditions are required bearing in mind thatapoptosis is an irreversible process. There is insuffi-cient information in literature on this issue, and morestudies are required to elucidate this phenomenon.

There are several questions on how to integrateclouds with other Comet Assay measurements. Theusual practice is to determine the percentage of cloudson each slide. This data is usually presented alongwith other Comet Assay measurements and cytotox-icity data where an increase in clouds parallels anincrease in DNA migration, but this is not always thecase. Good scientific judgement should be used ininterpreting these data, and further work is necessaryto assess how to integrate clouds with other CometAssay measurements.

Comet assay combined with FISH

The Comet Assay has also been combined with FISHtechnique (Comet-FISH) to investigate the localisa-

tion of specific gene domains within an individual cell(e.g. p53, her-2). The position of the fluorescenthybridisation spots in the comet head or tail indicateswhether the sequence of interest lies within or in thevicinity of a damaged region of DNA. Although notmany studies have been performed using this tech-nique, it has a number of potential uses in DNA repairand genomic instability studies. The measurementsthat can be collected from Comet-FISH experimentsare the position (either in head or tail) and number(number of fluorescence spots) of FISH signals afterDNA damage. In the assay, depending on the proberegion and probe length, the signals can be split orjust migrate to the tail. The location of FISH signals(either split or intact) in either head or tail appears tobe the best indicator for DNA damage using Comet-FISH. The splitting of signals appears to be randomevents depending on whether the DNA damagingagent targets the vicinity of the gene/locus specificindicator of interest. The p53 gene is a well-studiedexample where the signals split and migrate to the tailimmediately after irradiation (McKenna et al. 2003;Kumaravel and Bristow 2005). When allowed torepair for a period of time, the signals return back tothe head. More work is necessary to standardise themeasurements for the Comet -FISH technique beforeit is adopted for routine use.

Use of control cells in Comet Assay measurements

In a typical Comet Assay, electrophoresis methodsand differences in cell preparations create a significantsource of variation for the measurements. Suchvariation sometimes makes it difficult to compareresults between laboratories, and even within the samelaboratory. To overcome this problem, use of slidesprepared from cells containing a known amount ofDNA damage (also called control cells) are included inevery electrophoresis run. Control cells consistentlyproduce comets with predetermined DNA migration inthe tails. Some researchers define acceptance criteriafor experiments based on DNA migration observed incontrol cells. If the DNA migration in control cellsdoes not fall within the laboratory’s historical controlvalues, the data generated from that electrophoresis runare rejected. Some researchers use the data for DNAmigration in control cells to normalise with those fromother samples. Control cells are produced from stable

62 Cell Biol Toxicol (2009) 25:53–64

cells that have been irradiated with known amount ofgamma radiation, aliquoting them in small cryovialsfollowed by immediate flash freezing. There are somecommercial sources of control cells such as Trevigen®,who prepare their control cells by treating them withknown concentrations of etoposide. There was norecommendation for use of control cells in any of theIWGT workshops. We will recommend the use ofcontrol cells in each and every electrophoresis runas a best practice to generate high quality CometAssay data.

Conclusions

In conclusion, ‘measurements’ form an important partof Comet Assay analysis. Robust Comet Assay dataand interpretation depend on good and optimum slidestaining, adoption of robust image analysis practicesand use of reliable and meaningful Comet Assaymeasurement (e.g. % Tail DNA or OTM). As OTMvalues can differ widely between laboratories and/orwith different software packages, % Tail DNA is con-sidered appropriate for regulatory or inter-laboratorycomparison studies. Moreover, for all studies thatinvolve multiple electrophoresis runs, it is recommen-ded that % Tail DNA be used to reduce variability inthe results. It is generally accepted that visual scoringis as comparable as image analysis; however, imageanalysis can provide additional information (e.g.determination of the cell cycle status of cells bymeasuring their DNA content) that may be importantin the characterisation of the genotoxicity of somecompounds. It should be noted that clouds form animportant form of Comet Assay measurement, espe-cially when the DNA damage is extensive. It istherefore important that clouds are actively andaccurately looked for, recorded and appropriatelyinterpreted in Comet Assay experiments. These prac-tices will help to establish Comet Assay as a reliableand robust tool for fundamental biological research, inaddition to hazard and risk assessment, the main aimsof the field of genetic toxicology.

Acknowledgement/Disclaimer BV’s contribution to this ar-ticle was supported by a grant from the Ministry of HigherEducation, Science and Technology of Slovenia (BV: grant no.P1-0212). ANJ has not received any financial support fromAdvanced Technologies (Cambridge) Ltd.

References

Bichler J, Cavin C, Simic T, Chakraborty A, Ferk F, Hoelzl C,et al. Coffee consumption protects human lymphocytesagainst oxidative and 3-amino-1-methyl-5H-pyrido[4,3-b]indole acetate (Trp-P-2) induced DNA-damage: results ofan experimental study with human volunteers. Food ChemToxicol 2007. Epub ahead of print.

Burlinson B, Tice RR, Speit G, Agurell E, Brendler-SchwaabSY, Collins AR, et al. In vivo Comet Assay workgroup,part of the Fourth International Workgroup on GenotoxicityTesting: results of the in vivo Comet Assay workgroup.Mutat Res 2007;627:31–5.

Collins AR. Comet Assay for DNA damage and repair:principles, applications and limitations. Mol Biotechnol2004;26:249–61.

Collins AR, Duthie SJ, Dobson VL. Direct enzymic detectionof endogenous oxidative base damage in human lympho-cyte DNA. Carcinogenesis 1993;14:1733–35.

Collins AR, Ma AG, Duthie SJ. The kinetics of repair ofoxidative DNA damage (strand breaks and oxidisedpyrimidine) in human cells. Mutat Res 1995;336:69–77.

Collins A, Dusinska M, Franklin M, Somorovska M, PetrovskaH, Duthie S, et al. Comet Assay in human biomonitoringstudies: reliability, validation, and applications. EnvironMol Mutagen 1997a;30:139–46.

Collins AR, Mitchell DL, Zunino A, de Wit J, Busch D. UV-sensitive rodent mutant celllines of complementationgroups 6 and 8 differ phenotypically from their humancounterparts. Environ Mol Mutagen 1997b;29:152–60.

Collins AR, Dusinská M, Horská A. Detection of alkylationdamage in human lymphocyte DNA with the comet assay.Acta Biochim Pol 2001a;48:611–4.

Collins BH, Horská A, Hotten PM, Riddoch C, Collins AR.Kiwifruit protects against oxidative DNA damage inhuman cells and in vitro. Nutr Cancer 2001b;39:148–53.

Collins AR. Comet assay for DNA damage and repair:principles, applications and limitations. Mol Biotechnol2004;26:249–261.

Comet Assay interest group website. http://cometassay.com/(accessed 06, 2007).

Dixon DR, Pruski AM, Dixon LRJ, Jha AN. Marine inverte-brate eco-genotoxicology: a methodological overview.Mutagenesis 2002;17:495–507.

Fairbairn DW, Olive PL, O’Neill KL. The Comet Assay: Acomprehensive review. Mutat Res 1995;339:37–59.

Fisher AE, Burke D, Routledge MN. Can irradiation of rectaltumour cells from patient biopsy predict outcome ofradiotherapy? Proceedings of the Genome Stability net-work/United Kingdom Environmental Mutagen SocietyJoint Congress, University of Cardiff, 1–4 July 2007.

Garcia O, Romero I, González JE, Mandina T. Measurementsof DNA damage on silver stained comets using freeInternet software. Mutat Res 2007;627:186–90.

Gichner T, Patkova Z, Szakova J, Demnerova K. Cadmiuminduces DNA damage in tobacco roots, but no DNAdamage, somatic mutations or homologous recombinationin tobacco leaves. Mutat Res 2004;559:49–57.

Gichner T, Mukherjee A, Veleminsky J. DNA staining with thefluorochromes EtBr, DAPI and YOYO-1 in the comet

Cell Biol Toxicol (2009) 25:53–64 63

assay with tobacco plants after treatment with ethylmethanesulphonate, hyperthermia and DNase-I. MutatRes 2006;605:17–21.

Gedik CM, Ewen SWB, Collins AR. Single-cell gel electro-phoresis applied to the analysis of UV-C damage and itsrepair in human cells. Int J Radiat Biol 1992;62:313–20.

Hartmann A, Agurell E, Beevers C, Brendler-Schwaab S,Burlinson B, Clay P, et al. 4th International Comet AssayWorkshop. Recommendations for conducting the in vivoalkaline Comet Assay. Mutagenesis 2003;18:45–51.

Jha AN. Genotoxicological studies in aquatic organisms: Anoverview. Mutat Res 2004;552:1–17.

Kizilian N, Wilkins RC, Reinhardt P, Ferrarotto C, McLeanJRN, Mc-Namee JP. Silver-stained comet assay fordetection of apoptosis. Biotechniques 1999;27:926–30.

Kobayashi H, Sugiyama C, Morikawa Y, Hayashi M, Sufuni T.Comparison between manual microscopic analysis andcomputerised image analysis in single cell gel electropho-resis assay. MMS Commun 1995;3:103–15.

Kruman II, Kumaravel TS, Lohani A, Pedersen WA,Cutler RG, Kruman Y, et al. Folic acid deficiency andhomocysteine impair DNA repair in hippocampal neuronsand sensitize them to amyloid toxicity in experimentalmodels of Alzheimer’s disease. J Neurosci 2002;22:1752–62.

Kumaravel TS, Bristow RG. Detection of genetic instability atHER-2/neu and p53 loci in breast cancer cells usingComet-FISH. Breast Cancer Res Treat 2005;91:89–93.

Kumaravel TS, Jha AN. Reliable Comet Assay measurementsfor detecting DNA damage induced by ionising radiationand chemicals. Mutat Res 2006;605:7–16.

Lee RF, Steinert S. Use of the single cell gel electrophoresis/Comet Assay for detecting DNA damage in aquatic (marineand freshwater) animals. Mutat Res 2003;544: 43–64.

McKenna DJ, Rajab NF, McKeown SR, McKerr G, McKelvey-Martin VJ. Use of the comet-FISH assay to demonstraterepair of the TP53 gene region in two human bladdercarcinoma cell lines. Radiat Res 2003;159:49–56.

Miyamae Y, Iwasaki K, Kinae N, Tsuda S, Murakami M,Tanaka M, et al. Detection of DNA lesions inducedby chemical mutagens using the single-cell gel electro-phoresis (Comet) assay. 2. Relationship between DNAmigration and alkaline condition. Mutat Res 1997;393:107–13.

Moller P. Assessment of reference values for DNA damagedetected by the comet assay in human blood cell DNA.Mutat Res 2006;612:84–104.

Olive PL, Banath JP, Durand RE. Heterogeneity in radiation-induced DNA damage and repair in tumor and normalcells measured using the “Comet” assay. Radiat Res1990;122:86–94.

Ostling O, Johanson KJ. Microelectrophoretic study of radiation-induced DNA damages in individual mammalian cells.Biochem Biophys Res Commun 1984;123:291–8.

Raisuddin S, Jha AN. Relative sensitivity of fish and mammaliancells to sodium arsenate and arsenite as determined byalkaline single cell gel electrophoresis and cytokinesis-block micronucleus assay. Environ Mol Mutagen 2004;44:83–9.

Santos SJ, Singh NP, Natarajan AT. Fluorescence in situhybridization with comets. Exp Cell Res 1997;232:407–11.

Singh NP, McCoy MT, Tice RR, Schneider EL. A simpletechnique for the quantitation of low levels of DNA damagein individual cells. Exp Cell Res 1988;175:184–91.

Singh NP, Stephens RE, Schneider EL. Modifications ofalkaline microgel electrophoresis for sensitive detectionof DNA damage. Int J Radiat Biol 1994;66:23–28.

Smith AJO, Almeida GM, Thomas AL, Jones GD. Comet assaymeasures of irinotecan-induced DNA damage in vitro andin vivo. Proceedings of the Genome Stability network/United Kingdom Environmental Mutagen Society JointCongress University of Cardiff, 1–4 July 2007.

Speit G, Schütz P, Bonzheim I, Trenz K, Hoffmann H.Sensitivity of the FPG protein towards alkylation damagein the comet assay. Toxicol Lett 2004;146:151–8.

Tice RR, Furedi-Machacek M, Satterfield D, Udumudi A,Vasquez M, Dunnick JK. Measurement of micronucleatederythrocytesand DNA damage during chronic ingestion ofphenolphthalein intransgenic female mice heterozygousfor the p53 gene. Environ Mol Mutagen 1998;31:113–24.

Tice RR, Agurell E, Anderson D, Burlinson B, Hartmann A,Kobayashi H, et al. Single cell gel/Comet Assay: guide-lines for in vitro and in vivo genetic toxicology testing.Environ Mol Mutagen 2000;35:206–21.

Vilhar B. Help! There is a comet in my computer! A dummy’sguide to image analysis used in the comet assay. Universityof Ljubljana, http://www.botanika.biologija.org/exp/comet/Comet-principles.pdf (accessed 07, 2007); 2004.

64 Cell Biol Toxicol (2009) 25:53–64