Environmental and Molecular Mutagenesis 42:155–165 (2003)
Use of the Comet Assay for Studying Environmental Genotoxicity Comparisons Between Visual and Image Analyses Nanthawan Avishai, Claudette Rabinowitz, and Baruch Rinkevich* National Institute of Oceanography, Israel Oceanographic and Limnological Research, Haifa, Israel In order to evaluate the applicability of different measurement parameters employed in the comet assay for analyzing environmental samples, fish hepatoma (RTH-149) cells were exposed to concentrations of the model genotoxic agent hydrogen peroxide (H2O2; 1, 5, and 10 M) and to five water samples from sites along the Kishon River, the most polluted river in Israel. DNA damage was scored in parallel by visual and computer-image (Viscomet) analyses using 12 different parameters. Each parameter exhibited a different profile of responses. The four visual parameters were highly sensitive to the lowest (1 M) H2O2 concentration (1.8 –7.0-fold of the control). At 10 M H2O2 exposure, the visual parameter, percentage severe damage, showed the highest (40.3-fold) response while four other parameters, tail area, tail extent
moment (Viscomet), mean actual tail length and cumulative tail length (visual analysis), also had substantially elevated responses (8 –11-fold). We found that the DNA damage induced by field samples was similar in magnitude to the damage induced by 1 M H2O2, with only some of the parameters being highly sensitive to the damage. Only about one-half of the parameters could distinguish four significant levels of genotoxicity among the five sampling sites, while the remaining parameters detected only three levels. It is concluded that the choice of parameters for analyzing genotoxicity in ecotoxicological studies should be made in accordance with the characteristics of each parameter. Environ. Mol. Mutagen. 42:155–165, 2003. © 2003 Wiley-Liss, Inc.
Key words: comet assay; computer-image analysis; environmental monitoring; genotoxicity; visual scoring
INTRODUCTION The single cell gel electrophoresis (SCGE) assay, also known as the comet assay, is a rapid, reliable and sensitive method for evaluating DNA damage induced in individual cells by physical and chemical agents [Kim et al., 2002]. It ¨ stling and Johanson [1984] as an was first introduced by O assay for detecting DNA double-strand breaks in irradiated mammalian cells under neutral conditions. Singh et al. [1988] and Olive et al. [1990] independently modified the assay by developing alkaline versions (pH ⬎13 and pH ⫽ 12.3, respectively). Since the introduction of the alkaline modification, the breadth of applications and the number of investigations using this assay have increased almost exponentially. The alkaline comet assay is capable of detecting a wide variety of DNA damages such as DNA single-strand breaks, double-strand breaks, DNA-DNA/DNA-protein cross-links, oxidatively induced base damages, alkali-labile sites, and sites undergoing DNA repair [Mitchelmore and Chipman, 1998a; Tice et al., 2000]. It also has been employed to visualize DNA degradation due to necrosis and apoptosis [Kizilian et al., 1999; Singh, 2000]. The comet assay is recognized as one of the most sensitive methodologies available for DNA strand break detection [Collins et al., 1997a] and is distinguished by being simple, fast, and © 2003 Wiley-Liss, Inc.
effective, even for extremely small samples of cells, and applicable to cells from any organ of eukaryotic organisms [Mitchelmore and Chipman, 1998b]. The assay has been widely used for studies in genetic toxicology [Ding et al., 1999; Musatov et al., 1999; Tice et al., 2000; Sasaki et al., 2002], environmental genotoxicity [Mitchelmore and Chipman, 1998b; Kammann et al., 2000, 2001; Avishai et al., 2002], and for clinical [Somorovska´ et al., 1999; Rajaguru et al., 2002], radiation biology [Chaubey et al., 2001; Lyons and O’Brien, 2002] and DNA repair investigations [Olive et al., 1990; Kim et al., 2002]. In contrast to controlled laboratory studies, the assessment of genotoxicity in aquatic organisms in their natural Grant sponsor: Ministry of the Environment, Israel; Grant sponsor: Hal’ha (Israeli environmental fund). *Correspondence to: B. Rinkevich, National Institute of Oceanography, Israel Oceanographic and Limnological Research, Tel-Shikmona, POB 8030, Haifa 31080, Israel. E-mail:
[email protected] Received 7 April 2003; provisionally accepted 5 June 2003; and in final form 26 June 2003 DOI 10.1002/em.10189
156
Avishai et al.
environment is a complicated task, mainly because of the relatively low levels of genotoxicants and the existence of multiple potential genotoxic pollutants often encountered as complex mixtures [Mitchelmore and Chipman, 1998b]. The interactions between these genotoxicants and the organism’s DNA lead to a variety of damage [Schnurstein and Braunbeck, 2001]. The level of DNA strand breaks can be used as an early indicator for exposure to a wide variety of genotoxic agents and a sensitive endpoint for detecting DNA damage [Nacci et al., 1996]. The comet assay therefore, may serve as a nonspecific biomarker for the actual genotoxic impact on organisms [Cottle and Fe´rard, 1999]. Although the comet assay methodology is straightforward and does not require sophisticated equipment, the analysis of the comet images is not so simple. Two approaches, visual analysis and computerized image analysis, are used to measure DNA breakage in the comet assay. In the visual analysis process, comets are classified on the basis of their morphology and DNA damage is evaluated as an increase in the percentage of cells with comets [Devaux et al., 1997; Avishai et al., 2002; Heuser et al., 2002], with the comets categorized either by grading their size on a scale [Kobayashi et al., 1995; Wilson et al., 1998; Musatov et al., 1999; Tice et al., 2000; Kamer and Rinkevich, 2002], or by measuring tail lengths [Devi et al., 2001; Avishai et al., 2002]. This approach does not need any special equipment. It is reliable and fast [Kammann et al., 2000] but, to a certain extent, the results rely on subjective decisions made by the investigator. Furthermore, visual comet analysis requires experienced investigators. Differences in skill levels between investigators may influence data comparisons between experiments [Robbino et al., 1999]. Computerized image analysis (commonly used commercial packages include Komet [Kinetic Imaging, UK], Comet analysis system [Loats Associates, USA], Viscomet [Impuls, Germany] and the Comet assay [Perceptive Instruments, UK]) provide additional measurement criteria as compared with the visual approach, including measurements for tail length [Bo¨cker et al., 1999; Boeck et al., 2000; Uhl et al., 2000], tail extent moment [Yusuf et al., 2000], tail moment [Coughlan et al., 2002; Lee and Kim, 2002; Fatur and Filipicˇ, 2002], and percentage tail DNA [Mitchelmore and Chipman, 1998a; Boeck et al., 2000]. The commercial applications are generally sold as combined software/hardware packages and are rather expensive [Kon´ica et al., 2003]. This approach also requires a high-resolution CCD camera for the image analysis system to achieve maximum sensitivity. The aim of this study was to evaluate the applicability of the comet assay measurement parameters that are used in assessing the genotoxicity of environmental samples. We compared the genotoxicity of water samples from several localities along the Kishon River, the most polluted river in Israel. Genotoxicity was determined by performing the comet assay on cultured fish hepatocytes exposed to the
samples in vitro, and the resulting comets were evaluated by the two comet assay analysis approaches, visual analysis and computerized (Viscomet) analysis. The different measurement parameter responses for the field samples were compared with parameter responses resulting from comet assays conducted with a model genotoxic agent (H2O2). The sensitivities of the different damage parameters characteristic to each approach were assessed and their significant qualities determined. MATERIALS AND METHODS Chemicals Normal melting-point agarose was supplied by Seakem (BioWhittaker, Rockland, ME) and low melting point agarose was provided by Hispanagar (Burgos, Spain). All other chemicals used in the comet assay were of an analytical grade, obtained from Frutarom (Haifa, Israel) and Sigma-Aldrich (St. Louis, MO). The tissue culture medium and supplements were provided by the Biological Industries Company (Beith Haemeck, Israel).
Sampling Procedures Five water samples were obtained from four localities along a gradient of water genotoxicity in the Kishon River, Israel (Fig. 1a) as described [Kamer and Rinkevich, 2002]. Two were from the highly polluted Haifa Fishing Harbor (32° 48.5⬘N 35° 01.5⬘E), a surface sample and a sample from a depth of ⬃3.0 m. Additional samples were taken from the Yigael Yadin Bridge (32° 48.0⬘N 35° 03.0⬘E), Histadrut Bridge (32° 47.5⬘N 35° 03.0⬘E) and from the relatively clean area of Kiryat Haroshet (32° 42.0⬘N 35° 05.0⬘E). Each sample was collected in a 50-ml polypropylene tube (Greiner, Germany) and transported to the laboratory at the National Institute of Oceanography, Haifa. The pH of the samples was adjusted to 7 using 1–3% (v/v) of 1 M Hepes buffer. The samples were then filtered through 0.22-m disposable filters. To correct the osmolarity to ⬃220 mOsM, water samples from Kiryat Haroshet were mixed at a ratio of 1:1 with double-strength Dulbecco’s Minimal Essential Medium (DMEM, 2⫻), supplemented with 5% fetal calf serum (FCS), 2 mM L-glutamine, 1 mM sodium pyruvate, 1 mM Hepes buffer and 1% penicillin-streptomycinamphotericin B (PSA; stock solution: 10,000 U/ml penicillin, 10,000 g/ml streptomycin and 25 mg/ml amphotericin B). The samples were coded, stored at 4°C overnight, and tested the following day. Artificial Kishon water (AKW: [1⫻], pH 7.3: 467 mM NaCl, 11.05 mM KCl, 9.81 mM CaCl2 䡠 2H2O, 45.19 mM MgSO4 䡠 7H2O and 30 mM MgCl2 䡠 6H2O; stock solution of AKW (2⫻) [Rinkevich and Rabinowitz, 1993] was diluted with DMEM) and DMEM, which have the same pH and salinity as the Kishon water samples, were used as controls.
Cell Culture Fish cell lines are becoming important in vitro tools in aquatic toxicology [reviewed in Hightower and Renfro, 1988; Fent, 2001]. The fish hepatoma cell line RTH-149, which originated from an alfatoxin-induced hepatoma of an adult rainbow trout (Oncorhynchus mykiss) [Fryer et al., 1981], was obtained as a gift from Dr. Funkelstein (IOLR, Haifa, Israel). This cell line has been used in a variety of studies, such as fish virology [Lannan et al., 1984] and liver enzyme expression [Zafarullah et al., 1989; Olsson et al., 1990; Flouriot et al., 1995], and was recently found to serve as a sensitive indicator of pollution in aquatic monitoring programs [Avishai et al., 2002; Kamer and Rinkevich, 2002]. The cells were cultured in 25-cm2 flasks using DMEM supplemented as above, at 20°C in a 5% CO2 atmosphere. Three days before the experiment, confluent cultures were
Comet Assay Analyses
157
Fig. 1. a: The lower part of the Kishon River depicting sampling sites. b: The different cell damage stages (0-D) used for scoring the comet assay (visual approach).
dislodged using 0.25% trypsin-EDTA in calcium-magnesium-free PBS and seeded into 24-well plates (Nunc, Denmark) at a concentration of 2 ⫻ 104 cells/well in 0.5 ml medium. Cultures normally reached 95% confluence on the day of the experiment (⬃2 ⫻ 105 cells/well).
Cell Exposures The medium was changed 2 hr before the experiment. Experiments began by exposing cells independently, in triplicate to 50%-diluted Kishon water samples for 2 hr at 20°C. The comet assay calibration of RTH-149 cells with the genotoxic agent H2O2 was performed as positive control, using 0, 1, 5, and 10 M H2O2 concentrations. After 2 hr of exposure, the culture plates were placed on a tray over ice. The medium was removed; the cells were washed with 0.5 ml of calciummagnesium-free PBS, and then dislodged with 0.25 ml of 0.25% trypsinEDTA solution. After trypsinization, the cells were washed with cold PBS. Cell viability was determined using the trypan blue dye-exclusion assay. Wells having cell survival ⬎90% were processed for the comet assay.
Comet Assay Kishon water samples vary widely in their chemical (i.e. salinity) characteristics. Previously, we used comet assay testing with RTH-149 cells for determining the genotoxicity of several marine, estuary and freshwater samples, and established a methodology for these analyses, including pH and salinity adjustments [Kamer and Rinkevich, 2002]. Additional methodology used in the present study followed the protocol of Singh et al. [1988], with slight modification. In brief, 10 l of cell suspension (⬃2 ⫻ 105 cells) was mixed with 90l of 0.65% low-melting agarose and spread on a Star-frost microscope slide, pre-coated with 0.65% normal melting agarose. After 20 min of solidification at 4°C, a third layer containing 120l of 0.65% low-melting agarose was placed on top and left at 4°C for
an additional 20 min to solidify. The cells were then lysed by immersing the slides overnight in a freshly prepared lysis solution (2.5 M NaCl, 100 mM EDTA, 10 mM Tris, 1% Triton X-100, 10% DMSO, pH 10.0) at 4°C. After lysis, the slides were washed three times in cold water for 5 min and placed in a horizontal gel electrophoresis apparatus containing freshly prepared electrophoresis buffer (1 mM EDTA, 300 mM NaOH, pH ⫽ 13.0) for 20 min to allow DNA unwinding. Electrophoresis was then carried out at 20 V and at a starting current of 300 mA for 20 min at 4°C. Thereafter, the slides were neutralized with three washes of 0.4 M Tris, pH ⫽ 7.5, for 5 min, dehydrated with ethanol, and dried. The slides were stained with 60 l of 20 g/ml ethidium bromide solution and viewed under a fluorescent microscope using a U-MNG filter (Olympus, Germany). All steps were conducted in the dark to prevent additional DNA damage.
Comet Evaluation Slides were examined in parallel using visual analysis and Viscomet image analysis software. The same images were scored by both types of analyses. The visual analysis was conducted at 400⫻ magnification. A total of 150 randomly chosen cells from triplicate slides were examined for each sample (50 cells per slide). Each cell was scored visually as belonging to one of five specific damage stages based on the relative intensity of the head and tail fluorescence (from undamaged DNA, stage 0, to maximally damaged DNA, stage D; Fig.1b). Undamaged DNA (stage 0) has no tail, damage “stage A” has a tail length equal to or shorter than the length of the nucleus head diameter, damage “stage B” has a tail length 1.1–3.5 times, on average, longer than the head diameter, damage “stage C” has a tail length greater than 3.5 times, on average, the head diameter, and damage “stage D” has no “head” since all DNA migrated to the tail. Stages A to D received numerical values of 1 to 4, respectively. The number of cells in each stage was counted. The parameter “percentage cells with comet” was derived from the distribution of cells in damaged stages A–D, for each
158
Avishai et al.
treatment The “damage score” for the 50 cells examined on a slide ranged, therefore, between 0 (where DNA was intact in all cells) to 200 (where the DNA in all cells migrated into the tails). The head diameter and the total tail length (in m) of each cell were also measured. Actual tail length was calculated from total comet length minus the head diameter, and the mean average tail length was calculated from three independent samples. Cumulative tail length was calculated as the sum of the tail lengths from each independent sample. Image analysis was performed at 200⫻ magnification. The cell images were projected onto a high-resolution Heper-HAD™ (Sony, Japan) CCD camera (8 bits [Applitec, Israel, LIS-700]) and analyzed with Viscomet image analysis software using the MV Delta frame grabber (Matrix Vision, Germany). DNA damage was measured using the parameters: comet extent (the distance from the leading head edge to the trailing edge of the tail), comet total area (the number of pixels in the binary comet image), comet total intensity (sum of all pixel intensities from the grayscale image in calibrated units), tail area (minimum tail area in the binary image), tail extent (sum of all distances of each horizontal scan line from the first signal pixel to the last signal pixel divided by number of scan lines), tail length (the distance in m from the head center to the end of the tail), percentage tail DNA (percentage of DNA in tail), and tail extent moment (tail length ⫻ percentage tail DNA). Slides were coded and a single investigator analyzed all slides by both approaches throughout this study to minimize scoring variation.
Statistical Analysis Mean and standard deviation values were calculated for each treatment from three independent samples. A statistical program, SPSS 11.0 for Windows, was used for data analyses. Analysis of variance (ANOVA) or a nonparametric Kruskal-Wallis test was performed depending on the distribution and variance of the data. The MANOVA test, linear regression analysis and Pearson’s correlations were also performed to test for correlation among parameters.
RESULTS Hydrogen peroxide (H2O2), a powerful oxidizer, was used as a standard to establish dose-response relationships in this study. The visual analysis revealed that the exposure of RTH-149 cells to increasing concentrations of H2O2 for 2 hr induced significantly elevated levels of DNA damage (Fig. 2a,b). A 2-hr exposure to 1M H2O2 increased the percentage of cells with damaged DNA by almost threefold as compared with the controls, but only a minimal percentage (3.3%) of stage D cells were detected. At 5 M, almost all treated cells (94.7%) were damaged, with most comets indicating severe damage to the DNA (stages C, D; Fig. 2a). At 10 M, all treated cells were damaged and 82% of cells exhibited severe DNA damage. Analysis of the five parameters used (percentage cells with comets, percentage cells with severe damage, damage score, mean actual tail length, cumulative tail length; Fig. 2b) showed different profiles of response to elevated H2O2 concentrations. Within the range of H2O2 concentrations that were tested, linear dose-response curves were found for concentrations of 5 M and below. Damage score and the percentage of cells with comets were the most sensitive parameters in this range. The eight computerized parameters of the Viscomet program (Fig. 3a,b) had different response profiles with in-
creasing H2O2 concentrations. While parameters like comet total area, tail area (Fig. 3a) and percentage DNA in the tail (Fig. 3b) had linear dose-response relationships in the 0 –10 M H2O2 range, the other parameters had linear responses only in the 0 –5 M range. With the exception of percentage cells with comets (Fig. 2b), all the visual (Fig. 2b) and Viscomet (Fig. 3a,b) parameter measurements were significantly different at each of the H2O2 concentrations (Duncan’s test, P ⬍ 0.05), indicating that they are all reasonably sensitive as DNA damage indicators. However, comet total intensity was not significantly different at P ⬍ 0.01 for 1 and 5 M H2O2, and comet tail area was not different at P ⬍ 0.01 for the control and 1 M H2O2 treatments, indicating that these two parameters are less sensitive than the others. Comet total intensity, comet total area, and tail area (Viscomet analysis) were not statistically different for the control and 1 M H2O2 at P ⬍ 0.001. All parameters from both approaches, except comet total intensity, had Pearson’s correlation coefficients in the range of 0.7– 0.9 (data not shown), indicating moderate to strong correlations to one another. The four MANOVA tests (Wilkin’s Lambda, Pillai’s Trace, Hotelling-Lawley Trace, and Roy’s greatest root) produced an overall F ⬍ 0.001, indicating that the results from both the visual and image analysis approaches were not significantly different. The linear regression correlation coefficients (R2) were calculated for each parameter as a function of H2O2 concentration. One visual parameter (percentage cells with severe damage) and five Viscomet parameters (comet total area, percentage tail DNA, tail extent, tail extent moment, and tail area) had strong relationships with dose (R2 ⫽ 0.95, 0.99, 0.93, 0.91, 0.95, and 0.99, respectively). The percentage of cells with comets and comet total intensity had weak relationships with H2O2 concentration (0.44 and 0.13), while the rest of the parameters had moderate relationships (0.84 – 0.86) with H2O2 doses. Of this latter group, two visual parameters (mean actual tail length and cumulative tail length) and a Viscomet parameter (tail length) had the same R2 (0.86), indicating that these three parameters have the same strength of relationship to dose. In order to compare the strength of the responses for all the measurement parameters from the two analysis approaches, we calculated the DNA damage responses in terms of relative increase over control values (Fig. 4). Expressed in these terms, the responses to the H2O2 damage differed widely. Percentage of severely damaged cells showed the strongest responses, and a 40.3-fold increase was measured at 10 M H2O2. Tail area, tail extent moment (Viscomet analysis), mean actual tail length and cumulative tail length (visual analysis) gave responses of 8 –11-fold (10 M H2O2). Comet total area (Viscomet) and damage score (visual) were intermediate in sensitivity, producing fourand six-fold increases, respectively (damage score was, however, highly responsive at 0 –5 M concentrations). The other six parameters (a single parameter of the visual anal-
Comet Assay Analyses
159
Fig. 2. Visual analysis of comets produced by exposing fish hepatoma RTH-149 cells to H2O2 for 2 hr. a: Percentage of cells exhibiting damage for each damage stage. Scale bars reveal significant differences using
Duncan’s test at P ⬍ 0.001. b: Different parameters used in the visual approach. Mean actual tail length and cumulative tail length were measured in m (mean ⫾SD).
ysis, percentage cells with comets; and five Viscomet parameters, comet extent, comet total intensity, percentage tail DNA, tail extent and tail length) had lower responses to the H2O2 treatment. The genotoxicities of five water samples from localities along the Kishon River were evaluated in the comet assay
using both the visual analysis and image analysis approaches (Table I, Fig. 5). Both types of analyses distinguished several levels of genotoxicity for each tested parameter (Duncan’s test, P ⬍ 0.05). The visual analysis for percentage of cells with damage (Fig. 5) detected damage from treatment with all five samples. The damage produced
160
Avishai et al.
Fig. 3. Dose-response relationships for hydrogen peroxide (2O2) exposure of fish hepatoma RTH-149 cells using different Viscomet software parameters (mean ⫾SD). Comet total area and tail area were measured in m2, comet extent, tail extent, and tail length were measured in m, comet total intensity and tail extent moment were measured in arbitrary units.
by the samples also differed from one another, revealing four statistically different levels of genotoxic damage among the five sites. The DNA damage produced by the water samples was mainly categorized as stage B or C; only the sample from the most polluted area (Fishing Harbor at 3.0-m depth) produced a significant percentage of stage D cells (3.3%). The visual analysis parameters (Table I) showed either four different levels of genotoxicity above controls (percentage cells with comets, damage score) or three levels (mean actual tail length, cumulative tail length). The Fishing Harbor site (at 3.0-m depth) had the greatest level of genotoxicity by all parameters, while the Kiryat Harochet site had the lowest. Eight parameters were employed in the computerized analysis (Table I). All five sampling sites produced significantly higher levels of DNA damage than the controls (Duncan’s test, P ⬍ 0.05). Overall, the Viscomet parameters ordered the genotoxicity of the samples similarly to the visual analysis parameters. The Fishing Harbor at 3.0-m depth produced the greatest responses and Kiryat Harochet produced the lowest. Four significant grades of genotoxicity above controls were obtained using percentage tail DNA, tail extent, tail length, and tail extent moment. The other four parameters (comet extent, comet total intensity, comet total area, and tail area) showed only three significantly different levels of genotoxicity above the controls.
DISCUSSION Several computer image analysis packages (e.g., Komet, Viscomet, and Comet assay) are widely used to score DNA damage in the comet assay. It is believed that these packages, which measure the 10 most commonly cited parameters in the literature, provide more objective and detailed information than visual analysis [Kobayashi et al., 1995]. Several investigators [Olive et al., 1990; Yusuf et al., 2000; Schnurstein and Braunback, 2001] consider tail extent/Olive moment to be the most useful parameter to describe DNA damage, while others [Mitchelmore and Chipman, 1998a; Boeck et al., 2000] indicate that percentage tail DNA is the most satisfactory parameter. The alternative approach, visual (or manual microscopic) [Kobayashi et al., 1995] scoring, has also been used successfully in several biomonitoring studies [Devaux et al., 1997; Kammann et al., 2001; Avishai et al., 2002; Kamer and Rinkevich, 2002]. With all the parameters that are available, it remains unclear which parameter is superior for specific applications. We carried out a parallel scoring protocol with both the visual and the Viscomet analyses to assess their utility for evaluating genotoxicity in a natural setting where mixtures of unknown chemicals occur. In addition, the responses of each parameter to a single model genotoxic agent were compared. Table II compares the major characteristics of the 12
Comet Assay Analyses
161
Fig. 4. Relative increase over control values for visual and computerized comet parameters that evaluate the genotoxicity of hydrogen peroxide in fish hepatoma RTH-149 cells. Continuous lines depict data for the three parameters used in visual approach; dash lines show eight parameters used in Viscomet approach. The insert chart depicts the full range for the “percentage cells with severe damage” parameter.
parameters for DNA damage employed in the visual and Viscomet analyses. We concentrated on the linearity (doseresponse curves) of the responses to H2O2 damage, the significance of the responses produced by different H2O2 concentrations, the fold increase in the parameters produced
by H2O2 damage and the significance of the measurements made on different field samples (Table II). The Kishon water DNA damage responses were comparable in magnitude to the 0 –1 M H2O2 responses. As the Kishon River is one of the most polluted areas in Israel [Avishai et al.,
470.0 ⫾ 178.1b 449.5 ⫾ 184.7b 30.5 ⫾ 7.0b 21.6 ⫾ 6.1b 13.0 ⫾ 2.0b 1438.9 ⫾ 394.7b 2457.1 ⫾ 505.1c,d 41.6 ⫾ 8.0b
*All data are shown as mean average ⫾SD. a– e Significant groups using Duncan’s test at P ⬍ 0.05.
527.5 ⫾ 168.8b,c 31.0 ⫾ 8.7b,c 32.0 ⫾ 6.9b,c
10.6 ⫾ 3.4b,c
537.8 ⫾ 84.7b 639.9 ⫾ 266.5c,d 33.4 ⫾ 3.3b,c 24.8 ⫾ 3.8c 15.6 ⫾ 3.8c,d 1430.4 ⫾ 54.9b 2310.2 ⫾ 171.1b 44.3 ⫾ 3.1b 731.7 ⫾ 355.3c 42.0 ⫾ 17.6c,d 41.3 ⫾ 13.6c,d
14.6 ⫾ 7.1c
541.2 ⫾ 68.0b 672.4 ⫾ 31.5c,d 33.7 ⫾ 1.9c 25.4 ⫾ 1.5c 17.5 ⫾ 1.8d,e 1447.6 ⫾ 225.3b 2278.1 ⫾ 328.3b 44.5 ⫾ 2.7b 745.8 ⫾ 111.9c 44.3 ⫾ 6.8d 41.3 ⫾ 5.8d
14.9 ⫾ 2.2c
791.4 ⫾ 32.1d 916.4 ⫾ 146.8e 41.5 ⫾ 2.8e 32.6 ⫾ 2.8e 18.8 ⫾ 3.3e 1789.4 ⫾ 160.4d 2564.1 ⫾ 369.8d 52.7 ⫾ 2.9d 1200.0 ⫾ 182.0d 73.7 ⫾ 21.1e 60.0 ⫾ 9.2e
24.0 ⫾ 3.6d
662.1 ⫾ 190.6c 777.1 ⫾ 158.1d,e 37.3 ⫾ 3.7d 28.4 ⫾ 2.9d 17.6 ⫾ 2.2d,e 1598.94 ⫾ 316.4c 2343.8 ⫾ 264.1b,c 48.2 ⫾ 4.2c 797.5 ⫾ 97.4c 50.3 ⫾ 8.5d 50.7 ⫾ 7.0d
16.0 ⫾ 1.9c
280.3 ⫾ 52.8a 237.1 ⫾ 67.4a 23.1 ⫾ 1.5a 14.5 ⫾ 1.4a 9.8 ⫾ 2.6a 1125.8 ⫾ 172.7a 2136.4 ⫾ 285.3a 33.7 ⫾ 2.1a 191.7 ⫾ 29.8a 14.3 ⫾ 0.6a 18.0 ⫾ 3.5a
3.8 ⫾ 0.6a
221.5 ⫾ 19.5a 23.6 ⫾ 0.6a 14.7 ⫾ 0.7a 8.8 ⫾ 0.7a 1173.4 ⫾ 44.3a 2221.3 ⫾ 69.9a,b 34.3 ⫾ 0.7a 235.8 ⫾ 39.6a,b 4.7 ⫾ 0.8a,b 16.7 ⫾ 2.1a 22.7 ⫾ 1.2a
Damage score Sampling site
Control (DMEM) Control (AKW) Fishing Harbor (surface) Fishing Harbor (3-m depth) Yigael Yadin Bridge Histadrut Bridge Kiryat Harochet
Tail extent moment Tail length Tail extent Comet total area
Viscomet analysis
Tail DNA (%) Comet total intensity Comet extent Cumulative tail length (M)
Visual analysis
Mean actual tail length (M) % Cells with comet
TABLE I. Responses in Comet Assays Testing the Genotoxicity of Kishon Water Samples in Fish Hepatoma RTH-149 Cells*
291.5 ⫾ 21.7a
Avishai et al. Tail area
162
2002], these results may represent the maximum values for genotoxicity that may be obtained in field studies. Therefore, parameters that best responded to the lower range of H2O2 concentrations may best describe the genotoxicity of field samples. This logic suggests that two parameters from the visual analysis approach (percentage of cells with comets and damage score) and four parameters in the Viscomet analysis (percentage tail DNA, tail extent, tail length, and tail extent moment) may best detect the DNA damage from field samples. In contrast, assays that test the genotoxicity of agents producing high levels of DNA damage (such as laboratory studies of ionizing radiation), may best be performed using other parameters such as the tail extent moment [see also Olive et al., 1990]. This is because responses for this parameter were linear at high concentrations of H2O2 and because significant differences were detected for all the concentrations in the H2O2 dose-response. Tail extent moment (Viscomet) was one of the best parameters for detecting DNA damage within the concentration range of H2O2 used in the present study. This parameter includes a measure of tail length and percentage of DNA in the tail, and therefore is most similar to the visual damage score. Tail extent moment was highly responsive to increasing concentrations of H2O2 (10.5-fold induction compared with control) and was able to distinguish four different genotoxic levels among the five field samples. Moreover, its response was linear at 0 –5 M H2O2 and detected significant differences between the damage produced by all the H2O2 concentrations used. Two visual parameters, percentage cells with comets and damage score, and the Viscomet parameter of percentage tail DNA follow tail extent moment as the best overall parameters for detecting DNA damage. Collins et al. [1997b] indicated that visual scoring and undefined computer image analysis were equally useful for human studies employing the comet assay. Kamman et al. [2000] used both analyses (visual scoring and Image Pro 4.0, Media Cybernetics, USA) to measure the genotoxicity of North Sea marine sediment and found a significant correlation between visual scoring results and percentage tail DNA. Heaton et al. [2002] also found high correlations between both approaches (visual scoring and Komet 4.0 analysis) when analyzing DNA damage in canine and feline leukocytes. Similarly, our statistical evaluation showed medium to high correlations between the measurements made using both approaches (data not shown). In contrast, Kobayashi et al. [1995] compared the sensitivities of parameters from both approaches for detecting the genotoxicity of N-methy-N⬘-nitro-N-nitrosoguanidine, H2O2 and methyl methanesulfonate in human lymphoblast TK6 cells and found higher sensitivity for the visual (manual) scoring over the computer image analysis (New Vision System, Keio Electronic, Japan). Our results also indicate that the visual parameters were sensitive to low (1 M) H2O2 exposure. Visual scoring reached its limitations at higher H2O2 con-
Comet Assay Analyses
163
Fig. 5. Genotoxicity (percentage cells with damage) in fish hepatoma RTH-149 cells treated 2 hr with Kishon water samples. DMEM, control medium; AKW, control, Artificial Kishon water; FH-S, Haifa Fishing Harbor, surface; FH-D, Haifa Fishing Harbor, 3.0-m depth; YB, Yigael Yadin Bridge; HB, Histadrut Bridge; KH, Kiryat Haroshet (refer to Fig. 1.). Mean ⫾SD Scale bar grouping showed significant difference using Duncan’s test at P ⬍ 0.05.
TABLE II. Evaluation of Characteristics of Parameters Used for Measuring Genotoxicity in the Visual and Viscomet Analyses Evaluation of parameter characteristics Type of analysis
Parameter
Linearity (H2O2)
Visual
% cells with comet % severe damage Damage score Mean actual tail length Cumulative tail length
None 0–5 M None 0–5 M 0–5 M
Viscomet
Comet extent Comet total intensity Comet total area % tail DNA Tail extent Tail length Tail extent moment Tail area
0–5 M 0–1 M 0–10 M 0–5 M 0–5 M 0–5 M 0–5 M 0–10 M
Fold increaseb
Significance levels (field)c
Not significant Whole range Whole range Whole range Whole range
Low High Medium to high High High
High High High Low Low
Whole range Not significant at P ⫽ 0.001 Not significant at P ⫽ 0.001 Whole range Whole range Whole range Whole range Not significant at P ⫽ 0.001
Low Low Medium Low Low Low High High
Low Low Low High High High High Low
Significant differences (H2O2 concn)a
Outcomes at 4 concentration points within the range of 0 –10 M H2O2 (P ⬍ .001). As compared with controls within 0 –10 M H2O2 range: high, ⬎8-fold induction; medium, 3– 8-fold induction; low, ⬍3-fold induction. c Significant levels at field samples from five sites: high, 4 significant genotoxic levels above controls; low, 3 significant genotoxic levels above control. a
b
centrations (1–10 M) where the computer image analysis produced superior results. While H2O2 produces mainly single-strand breaks, field samples possess multiple genotoxic pollutants that cause various DNA damages (e.g., single- and double-strand breaks, cross-links). This point
should be taken into consideration when comparing H2O2 results with environmental outcomes in the comet assay. In many instances, very little attention is given to the applicability of the measurement parameters used in the comet assay. Investigators often use different parameters
164
Avishai et al.
based on how they are commonly employed in the literature [Olive et al., 1990; Collins et al., 1996, 1997a; Boeck et al., 2000; Schnurstein and Braunbeck, 2001; Heaton et al., 2002]. The results of this study indicate that the parameters used in analyzing genotoxicity in the comet assay, for example when analyzing the genotoxicity of field samples containing bulky genotoxins, should be chosen carefully. The choice of the parameter(s) used should be made with reference to the parameters. ACKNOWLEDGMENTS The authors are grateful to J. Douek and G. Paz for their assistance and helpful advice. REFERENCES Avishai N, Rabinowitz C, Moiseeva E, Rinkevich B. 2002. Genotoxicity of the Kishon River, Israel: the application of an in vitro cellular assay. Mutat Res 518:21–37. Bo¨cker W, Rolf W, Bauch T, Mu¨ller WU, Streffer C. 1999. Automated comet analysis. Cytometry 35:134 –144. Boeck MD, Touil N, Visscher GD, Vande PA, Kirsch-Volders M. 2000. Validation and implementation of an internal standard in comet assay analysis. Mutat Res 469:181–197. Chaubey RC, Bhilwade HN, Rajagopalan R, Bannur SV. 2001. Gamma ray induced DNA damage in human and mouse leucocytes measured by SCGE-Pro: a software developed for automated image analysis and data processing for comet assay. Mutat Res 490:187–197. Collins AR, Dusˇinska´ M, Gedik CM, Sˇteˇtina R. 1996. Oxidation damage to DNA: do we have a reliable marker? Environ Health Perspect 104:465– 469. Collins AR, Dobson VL, Dusˇinska´ M, Kennedy G, Sˇteˇtina R. 1997a. The comet assay: what can it really tell us? Mutat Res 375:183–193. Collins A, Dusˇinska´ M, Franklin M, Somorovska´ M, Petrovska´ H, Duthie S, Fillion L, Panayiotidis M, Pasˇlova´ K, Vaughan N. 1997b. Comet assay in human biomonitoring studies: reliability, validation and applications. Environ Mol Mutagen 30:139 –146. Cottle S, Fe´rard JF. 1999. Comet assay in genetic ecotoxicology: a review. Environ Mol Mutagen 34:246 –255. Coughlan BM, Hartl MGJ, O’Reilly SJ, Sheehan D, Morthersill C, Van Pelt FNAM, O’Halloran J, O’Brien NM. 2002. Detecting genotoxicity using the comet assay following chronic exposure of Manila calm Tapes semisdecussatus to polluted estuarine sediments. Mar Pollut Bull 44:1359 –1365. Devaux A, Personen M, Monod G. 1997. Alkaline comet assay in rainbow trout hepatocytes. Toxicol In Vitro 11:71–79. Devi KD, Banu BS, Mahboob M, Jamil K, Grover P. 2001. In vitro genotoxic effect of cadmium chloride in mice leukocytes using comet assay. Teratog Carcinog Mutagen 21:325–333. Ding W, Shen H, Zhu H, Lee B, Ong C. 1999. Genotoxicity of microcystic cyanobacteria extract of a water source in China. Mutat Res 442: 69 –77. Fatur T, Filipicˇ M. 2002. Cadmium induced DNA damage in human hepatoma (Hep G2) and Chinese hamster ovary (CHO) cells. Radiol Oncol 36:156 –158. Fent K. 2001. Fish cell lines as versatile tools in ecotoxicology: assessment of cytotoxicity, cytochrome P4501A induction potential and estrogenic activity of chemicals and environmental samples. Toxicol In Vitro 15:477– 488. Flouriot G, Pakdel F, Ducouret B, Valotaire Y. 1995. Influence of xeno-
biotics on rainbow trout liver estrogen receptor and vitellogenin gene expression. J Mol Endrocrinol 15:143–151. Fryer JL, Mccaine BB, Leong JC. 1981. A cell line derived from Rainbow trout (Salmo gairdneri) hepatoma. Fish Pathol 15:193–200. Heaton PR, Ransley R, Charlton CJ, Mann SJ, Stevenson J, Smith BHE, Rawlings JM, Harper EJ. 2002. Application of single-cell gel electrophoresis (comet) assay for assessing levels of DNA damage in canine and feline leukocytes. J Nutr 132:1598S–1603S. Heuser VD, Silva JD, Moriske HJ, Dias JF, Yoneama ML, Freitas TRO. 2002. Genotoxicity biomonitoring in regions exposed to vehicle emission using the comet assay and the micronucleus test in native rodent Ctenimys minutus. Environ Mol Mutagen 40:227–235. Hightower LE, Renfro JL. 1988. Recent applications of fish cell culture to biomedical research. J Exp Zool 248:290 –302. Kamer I, Rinkevich B. 2002. In vitro application of the comet assay for aquatic genotoxicity: considering a primary culture versus a cell line. Toxicol In Vitro 16:177–184. Kammann U, Riggers JC, Theobald N, Steinhart H. 2000. Genotoxic potential of marine sediments from the North Sea. Mutat Res 467:161–168. Kammann U, Bunke M, Steinhart H, Theobald N. 2001. A permanent fish cell line (EPC) for genotoxicity testing of marine sediments with the comet assay. Mutat Res 498:61–77. Kim B, Park JJ, Edlaer L, Fournier D, Haase W, Sautter-Bihl M, Gotzes F, Thielmann HW. 2002. New measurement of DNA repair in the single-cell gel electrophoresis (comet) assay. Environ Mol Mutagen 40:50 –56. Kizilian N, Wilkins RC, Reinhardt P, Ferrarotto C, McLean JRN, McNamee JP. 1999. Silver-stained comet assay for detection of apoptosis. Biotechniques 27:926 –930. Kobayashi H, Sugiyama C, Morikawa Y, Hayashi M, Sofuni T. 1995. A comparison between manual microscopic analysis and computerized image analysis in the single cell gel electrophoresis assay. MMS Commun 3:103–115. Kon´ica K, Lankoff A, Banasik A, Lisowska H, Kuszewski T, Go´ z´dz´ S, Koza Z, Wojcik A. 2003. A cross-platform public domain PC image-analysis program for the comet assay. Mutat Res 534:15–20. Lannan CN, Winton JR, Fryer JL. 1984. Fish cell lines: establishment and characterization of nine cell lines from salmonids. In Vitro 20:671– 676. Lee R, Kim GB. 2002. Comet assay to assess DNA damage and repair to grass shrimp embryos exposed to phototoxicants. Mar Environ Res 54:465– 469. Lyons NM, O’Brien NM. 2002. Modulatory effect of an algal extract containing astaxanthin on UVA-irradiated cells in culture. J Dermatol Soc 30:73– 84. Mitchelmore CL, Chipman JK. 1998a. Detection of DNA strand breaks in Brown trout (Salmo trutta) hepatocytes and blood cells using the single cell gel electrophoresis (comet) assay. Aquat Toxicol 41: 161–182. Mitchelmore CL, Chipman JK. 1998b. DNA strand breakage in aquatic organisms and the potential value of the comet assay in environmental monitoring. Mutat Res 399:135–147. Musatov SA, Anisimov VN, Andre´ V, Vigreux C, Godard T, Sichel F. 1999. Effect of melatonin on N-nitroso-N-methyurea induced carcinogenesis in rats and mutagenesis in vitro (Ames test and comet assay). Cancer Lett 138:37– 44. Nacci DE, Cayula S, Jackin E. 1996. Detection of DNA Damage in individual cells from marine organisms using the single cell gel assay. Aquat Toxicol 35:135–147. Olive PL, Banath JP, Durand RE. 1990. Heterogeneity in radiation-induced DNA damage and repair in tumor and normal cells measured using the “Comet” assay. Radiat Res 122:86 –94. Olsson PE, Hyllner SJ, Zafarullah M, Andersson T, Gedamu L. 1990. Differences in metallothionein gene expression in primary cultures
Comet Assay Analyses of rainbow trout hepatocytes and RTH-149 cell line. Biochim Biophys Acta 1049:78 – 82. ¨ stling O, Johanson KJ. 1984. Microelectrophoretic study of radiationO induced DNA damages in individual mammalian cells. Biochem Biophys Res Commun 123:291–298. Rajaguru P, Vidya L, Baskarasethupathi B, Kumar PA, Palanivel M, Kalaiselvi K. 2002. Genotoxicity evaluation of polluted ground water in human peripheral blood lymphocytes using the comet assay. Mutat Res 517:29 –37. Rinkevich B, Rabinowitz C. 1993. In vitro culture of blood cells from the colonial protochordate Botryllus schlosseri. In Vitro Cell Dev Biol 29A:79 – 85. Robbino L, Carrozzino R, Puglia CP, Corbu C, Brambilla G. 1999. Correlation between induction of DNA fragmentation and micronuclei formation in kidney cells from rats and humans and tissue-specific carcinogenic activity. Toxicol Appl Pharmacol 161:153–159. Sasaki YF, Kawaguchi S, Kamaya A, Ohshita M, Kabasawa K, Iwana K, Taniguchi K, Tsuda S. 2002. The comet assay with 8 mouse organs: results with 39 currently used food additives. Mutat Res 519:103– 119. Schnurstein A, Braunbeck T. 2001. Tail moment versus tail length- application of an in vitro version of the comet assay in biomonitoring for genotoxicity in native surface waters using primary hepatocytes and gill cells from zebrafish (Danio reio). Ecotoxicol Environ Saf 49:187–196. Singh NP. 2000. Microgel for estimation of DNA strand breaks, DNA protein crosslinks and apoptosis. Mutat Res 455:111–127. Singh NP, McCoy MT, Tice RR, Scheider EL. 1988. A simple technique
165
for quantitation of low levels of DNA damage in individual cells. Exp Cell Res 175:184 –191. Somorovska´ M, Szabova´ E, Vodicˇka P, Tulinska´ J, Barancokova´ M, Fa´bry R, Lı´sˇkova´ A, Riegerova Z, Petrovska H, Kubova J, Rausova K, Dusˇinska M, Collins A. 1999. Biomonitoring of genotoxic risk in workers in a rubber factory: comparison of the comet assay with cytogenetic methods. Mutat Res 445:181–192. Tice RR, Agurell E, Anderson D, Burlinson B, Hartmann A, Kobayashi K, Miyamae E, Rojas E, Ryu JC, Sasaki YF. 2000. Single cell gel/ comet assay: guidelines for in vitro and in vivo genetic toxicology testing. Environ Mol Mutagen 35:206 –221. Uhl M, Helma C, Knasmu¨ller S. 2000. Evaluation of the single cell gel electrophoresis assay with human hepatoma (Hep G2) cells. Mutat Res 468:213–225. Wilson JT, Pascoe PL, Parry JM, Dixon DR. 1998. Evaluation of the comet assay as a method for the detection of DNA damage in the cells of a marine invertebrate, Mytilus edulis L. (Mollusca: Pelecypoda). Mutat Res 399:87–95. Yusuf AT, Vian L, Sabatier R, Cano JP. 2000. In vitro detection of indirect-acting genotoxins in the comet assay using Hep G2 cells. Mutat Res 468:227–234. Zafarullah M, Olsson PE, Gedamu L. 1989. Endogenous and heavy metal ion induced metallothionein gene expression in salmonid tissue and cell line. Gene 83:85–93.
Accepted by— P.L. Olive
