Document kD32kNydr7K4z3eM5JJZQN3BD

Mutagenesis vol. 22 no. 5 pp. 321327, 2007 Advance Access Publication 18 June 2007 doi:10.1093/mutage/gem020 Leukaemia-specific chromosome damage detected by comet with fluorescence in situ hybridization (comet-FISH) Downloaded from http://mutage.oxfordjournals.org at UCSF Library and Center for Knowledge Management on June 4, 2010 Patricia A. Escobar, Martyn T. Smith, Ananth Vasishta, Alan E. Hubbard and Luoping Zhang* Division of Environmental Health Sciences, School of Public Health, 140 Warren Hall, University of California, Berkeley, California 94720, USA Acute myeloid leukaemia (AML) is associated with exposure to benzene and treatment with chemotherapeutic agents. It is thought to arise from damage to specific regions of DNA, resulting in chromosome rearrangements or loss. For instance, a deletion on the long arm of chromosome 5 [e.g. del(5q31)] is common in AML patients previously treated with alkylating agents, such as melphalan, or exposed to benzene. Translocations of the MLL gene at 11q23 are frequently observed in AML arising from treatment with topoisomerase II inhibitors, such as etoposide. Our goal was to determine whether or not breakage at 5q31 and 11q23 is selectively induced by these chemical agents. To address this question, the comet assay combined with fluorescence in situ hybridization (cometFISH) was used to detect DNA breakage in the specific chromosomal regions in an in vitro model. TK6 lymphoblastoid cells were exposed to melphalan, etoposide or the benzene metabolite, hydroquinone (HQ), at various concentrations. HQ, melphalan and etoposide induced DNA breaks at both 5q31 and 11q23 chromosome regions in a dose-dependant manner. However, HQ produced significantly more DNA damage at 5q31 than at 11q23. Etoposide produce slightly more DNA damage at 11q23 and melphalan had a somewhat greater effect at 5q31, but not significantly so. Thus, HQ and melphalan act similarly, perhaps explaining some similarities between benzene- and alkylating agent-induced AML. CometFISH also appears to be a useful approach for detecting and comparing damage to specific chromosome regions of significance in leukaemogenesis. Introduction Leukaemogenesis is a complex, multi-step process that involves transformation of a haematopoietic progenitor cell through the induction of several types of chromosomal aberrations and gene mutations followed by clonal expansion (1,2). The cause of most types of leukaemia is largely unknown, but several factors have been directly implicated, including inherited mutations, ionizing radiation and chemical exposure (3,4). Chemotherapeutic treatment with alkylating agents and topoisomerase II inhibitors can cause therapy-related acute myeloid leukaemia (AML) (57). However, depending on the agent used, patients display different cytogenetic characteristics (5,6). AML induced by alkylating agents, such as melphalan, usually develops 710 years after exposure and commonly displays a deletion of all or part of chromosomes 5 and 7. Previous in vitro studies have shown that melphalan preferentially induces breaks and structural aberrations in chromosomes 5, 7, 11 and 17 (810). Deletion of 5q31 was also detected in multiple myeloma patients treated with melphalan (11). Topoisomerase II inhibitors, such as etoposide, induce AML with a shorter latency period of 15 years after exposure and are usually associated with balanced translocations involving chromosome band 11q23, the location of the MLL gene (7,12). Topoisomerase II enzymes catalyse breakage of double-stranded DNA during replication, transcription and chromosomal segregation. Topoisomerase II inhibitors stabilize the complex of topoisomerase II, leaving a broken DNA 5# strand end, slowing ligation, triggering double-strand break repair (illegitimate recombination) and, in some instances, apoptosis (12). Occupational exposure to benzene also induces leukaemia including AML (13,14). Benzene is metabolized in the liver to its primary metabolite phenol, which is hydroxylated to hydroquinone (HQ) (15). HQ can be further oxidized in the bone marrow to 1,4-benzoquinone (16,17). This pathway is currently thought to play a major role in benzene-induced leukaemia. Leukaemias resulting from benzene exposure show many types of chromosomal aberrations including the loss of all or part of chromosomes 5 and 7, such as del(5q31) (1820). 1,4-Benzoquinone has also been reported to be a topoisomerase II inhibitor and may affect the MLL gene at chromosome band 11q23 (21,22). In the present study, we have attempted to determine if leukaemogenic chemicals induce damage specifically at the 5q31 and 11q23 regions involved in leukaemogenesis, and critically to determine if HQ acts more like melphalan or etoposide. The comet assay (single-cell gel electrophoresis) is a simple method for measuring DNA single- and double-strand breaks in eukaryotic cells. The more DNA breaks produced, the greater the fluorescent intensity of the comet tail relative to the head. These DNA breaks may be induced by oxidative stress (2325). Recently, the comet assay has been combined with the specificity of fluorescence in situ hybridization (FISH) (26,27) to allow simultaneous detection of overall DNA damage and DNA breaks in specific regions of the genome. The use of a fluorescent-labelled DNA probe homologous to the region of interest in conjunction with comet (cometFISH) permits the detection of DNA breaks within that region. CometFISH was introduced in 1997 (26) to study the spatial distribution of specific chromosome sequences and chromatin fibres in comet cells. Subsequently, it was used to examine initial DNA damage and subsequent repair in the TP53 gene region of bladder carcinoma cells after gamma irradiation exposure (28,29). It has recently been applied to study damage in a number of specific genes (2832). Here, we have used cometFISH to compare the DNA damaging effects of melphalan, etoposide and HQ specifically at bands 5q31 and 11q23. Our a priori hypothesis was that the alkylating agent melphalan would produce more damage at *To whom correspondence should be addressed. Tel: 510 643 5189; Fax: 510 642 0427; Email: luoping@berkeley.edu The Author 2007. Published by Oxford University Press on behalf of the UK Environmental Mutagen Society. All rights reserved. For permissions, please e-mail: journals.permissions@oxfordjournals.org. 321 P. A. Escobar et al. band 5q31 than at 11q23 and, conversely, that the topoisomerase II inhibitor etoposide would cause more breakage at band 11q23 than at 5q31. For the benzene metabolite HQ, it was unclear which location, if any, would be preferred as it has been described as a topoisomerase II inhibitor (21), but is also known to cause deletion of 5q31 (20,33). In order to test these hypotheses, we exposed a commonly used model, the human lymphoblastoid cell line, TK6, to different concentrations of melphalan, etoposide and HQ and then applied cometFISH with probes specific for chromosome bands 5q31 and 11q23. Methods Human cells and culture conditions Human lymphoblastoid cell line TK6 was obtained from the American Tissue Culture Collection (Rockville, MD). TK6 was cultured in RPMI-1640 medium with L-glutamine supplemented with 1% penicillinstreptomycin (Invitrogen, Carlsbad, CA) and 10% foetal bovine serum (Omega Scientific Inc., Tarzana, CA). Cells were incubated at 37C, in a 5% CO2 atmosphere. Chemical treatment and cytotoxicity The TK6 cells were exposed to various concentrations of HQ (Aldrich, Milwaukee, WI), melphalan and etoposide (Sigma, St Louis, MO). Dimethyl sulfoxide was used as the vehicle for melphalan and etoposide and was present in cell cultures at a final concentration of 0.25%. HQ was dissolved in RPMI. When TK6 cells were in exponential growth phase, they were treated with these three chemicals for 1 h and then harvested for cell viability test, cell cycle analysis, comet assay and cometFISH. All treatments were performed in duplicate for each dose and each experiment was repeated three times. Cytotoxicity was measured utilizing the Trypan Blue exclusion assay. The cells were mixed with Trypan blue 0.04% (Gibco, Carlsbad, CA) in a 1:1 ratio and the percentage of dead cells was determined for each chemical treatment at various concentrations. The percentage of dead cells was ,10% at non-cytotoxic doses. Cell cycle analysis An aliquot of the cell suspension was mixed with cold Hanks balanced saline solution (HBSS) without Ca and Mg (Invitrogen) and washed twice. Then, cold 70% ethanol was added to the cells and left at 4C overnight. The cells were washed with HBSS, and RNAse A (1 mg/ml) and propidium iodide (PI) (34) were added. The solution was stored at 4C until analysis. The cell cycle analysis was performed on a Beckman-Coulter EPICS XL-MCL. Using the intensity of PI fluorescence as a measure of the total DNA content of individual cells, cell cycle populations were quantified from a standard count of 10 000 cells. Cell cycle analysis was performed at 1 and 24 h after exposure. Comet assay The alkaline comet assay was performed according to Singh et al. (23) with some modifications. Briefly, super-frosted microscope slides were pre-coated by dipping each slide into 1% normal melting agarose and air-dried overnight. Seventyfive microlitres of low-melting agarose (Fisher Scientific, Pittsburgh, PA) mixed with 1 105 cells was placed on the pre-coated slide. Several slides were prepared at each dose to be used for both comet and cometFISH assays. A glass coverslip was placed on top and slides were kept at 4C for 10 322 min to allow the gel to solidify. Then a second layer of lowmelting agarose was added and the slides were kept at 4C for another 10 min to allow the gel to solidify. Slides were placed in cold lysing solution (2.5 M NaCl, 100 mM EDTA, 10 mM Trizma base and 1% Triton X-100, pH 10) and kept at 4C until electrophoresis was performed. The Trizma base was from Sigma and other reagents in the lysing solution were purchased from EM Scientific (Darmstadt, Germany). Unwinding was performed in an alkaline buffer (300 mM NaOH and 1 mM EDTA, pH 13) for 25 min and electrophoresis was run in the same alkaline buffer at 8C at 0.73 V/cm, for 35 min. After electrophoresis, the slides were neutralized with 0.4 M Tris, pH 7.5. The gels were dehydrated with 100% ethanol, airdried overnight and stored in a dry place before being processed for comet or cometFISH analysis. All slides were blind-coded before they were stained with SYBR Green (Molecular Probes, Eugene, OR). The cell scoring was performed under a 25 objective on a fluorescence microscope using an FITC filter. Twenty-five randomly chosen cells per slide and two slides per dose were analysed with the CometScan (MetaSytems, Germany) imaging software. CometFISH Comet slides from solvent controls and two concentrations of each of the three chemicals tested were subjected to FISH analysis. The slides were immersed for 20 min in absolute ethanol. All slides were hybridized with two specific chromosome probes, 5q31 and 11q23 (Vysis, Downers Grove, IL), each in separate spots outlined with a diamond pencil. Dual-colour human chromosome band probes 5q31 (200 kb) and 11q23 (350 kb) were used. To each spot, 10 ll of probe mixture was added (2 ll chromosome probe, 1 ll of salmon sperm DNA and 7 ll hybrid solution from Vysis). Coverslips (22 22 mm) were placed over each spot and sealed with rubber cement. An automatic denaturation and hybridization procedure was performed using the HyBrite denaturation/hybridization system (Vysis). Cells and probes were denatured at 68C for 10 min and hybridized at 37C for 72 h. After hybridization, the rubber cement was carefully removed and the slides were placed in water for a few seconds to allow the coverslips to slip off and prevent gel rupture. The slides were washed in 50% formamide/ 2 SSC at 45C for 3 min and 0.1 SSC at room temperature for 2 min, three times. The cells were counter-stained with 1 lg/ml 4#,6-diamidino-2-phenylindole (DAPI). The hybridization signals were observed with a fluorescence microscope under triple band-pass filter for DAPI/FITC/Texas Red (excitation at 405, 490 and 570 nm, respectively; emission at 460, 525 and 635 nm, respectively). In each spot, only the spectrum orange signal was taken into account for the cometFISH analysis because they represented the 5q31 and 11q23 chromosome bands. CometFISH scoring We designed a 4 4 table to classify and evaluate the comet FISH results (Figure 4). The figure describes four categories that quantify site-specific DNA breakage by the number of DNA probe signals (IIV) and four categories that describe the localization of the probe signal in the comet (head, head and tail or only tail) in relation to the presence of a comet tail which correlates with overall DNA damage (AD). A total of 150 300 cells were randomly scored for each chemical (50100 cells per dose). Each cell was categorized according to Figure 4 and the cumulative counts were used for the data analysis. The cometFISH experiments were repeated three times. Downloaded from http://mutage.oxfordjournals.org at UCSF Library and Center for Knowledge Management on June 4, 2010 Chemically induced damage to 5q31 and 11q23 Downloaded from http://mutage.oxfordjournals.org at UCSF Library and Center for Knowledge Management on June 4, 2010 Statistical analysis The comet parameter used for data analysis was the percentage of DNA in the tail (% tail intensity) and linear regression analysis was performed to determine doseresponse relation- ship shown as a Ptrend. The data generated from the cometFISH assay, using Figure 4, were categorical and were analysed using an order logistic regression, which returns odds ratios (ORs) for an increase in the outcome by one category (35). The ordered logistic regressions analysis assessed the distribution of the chromosome probe signals among the different cell damage categories. Because there were multiple measurements per slide, our inference was based on robust standard errors (36), which adjusts for the residual correlation of measurements made on the same slide. The analyses were performed to determine which chromosomes were more affected by the chemicals of interest, and so we included interaction terms (dose chromosome). For all analyses, we treated dose categorically by including the appropriate dummy variables in the model. For instance (using the data in Table III as an example), consider the following ordered logistic regression model [see Models for Ordinal Scales in McCullagh and Nelder (35) for equivalent notation]: ! log cjDose; Chrom 1 cjDose; Chrom 5hj b1IDose525 b2IDose550 b3IChrom511 b4IDose525IChrom511 b5IDose550IChrom511; where cj(Dose, Chrom) 5 P(Y j|Dose, Chrom), Y is the ordered outcome (probe signals: 0/12, 2/34, 3/57, 4/!8) and I(statement) 5 1 if `statement' is true, 0 otherwise. In this model, the test of the null H0: b4 5 0 tests whether the effect (OR for moving up one category) of HQ 25 versus 0 is the same for both chromosomes 11 and 5; likewise, the test of the null H0: b5 5 0 tests the equality of the OR's for 50 versus 0. We used the Wald test of these coefficients (based on the robust standard errors) to derive the P-values reported below in the Results. Percentage (%) A 100 80 60 40 20 0 % Tail Intenisty Viability 0 10 25 50 Hydroquinone [M] 75 Percentage (%) B 100 80 60 40 20 0 0 0.5 2.5 5 Melphalan [M] 10 25 Percentage (%) C 100 80 60 40 20 0 0 0.5 2.5 5 Etoposide [M] 10 25 Fig. 1. Cell viability and DNA damage in TK6 cells exposed to (A) HQ, (B) melphalan, and (C) etoposide for 1 h. DNA damage is represented as the percentage of DNA intensity present in the comet tail (% tail intensity). There was a positive doseresponse relationship by linear regression analysis for all three chemicals tested: Ptrend for (A) 0.04, (B) 0.006 and (C) 0.005. Each bar represents the mean % tail intensity of three independent experiments. Error bars indicate standard error of the mean. Cell viability is represented as each diamond point by the percentage of viable cells detected by Trypan blue staining. Results Overall DNA damage measured using the comet assay Exposure of human TK6 cells to three chemical leukaemogens, HQ, melphalan and etoposide, produced a significant increase in overall DNA damage, represented as an increase of the % tail intensity. HQ (up to 75 lM) caused increasing DNA damage (Ptrend 5 0.04) without significant cytotoxicity (Figure 1A). Melphalan (Ptrend 5 0.006) and etoposide (Ptrend 5 0.005) also caused DNA breakage with increasingly more DNA being present in the comet tails, without cytotoxicity at doses up to 25 lM (Figure 1B and C, respectively). Of the three compounds, etoposide was the most effective at producing DNA damage with % tail intensity content approaching 70% at doses above 5 lM. Cell cycle analysis Cell cycle analysis by flow cytometry revealed that HQ produced S-phase arrest at concentrations !25 lM (Figure 2A). Low doses of melphalan produced G2-phase arrest, whereas higher doses resulted in S-phase arrest (Figure 2B). At the highest (25 lM) dose, however, it produced G1/S phase arrest increasing the percentage of cells in G1 phase drastically (Figure 2B). Etoposide caused arrest at G2/M phase and increasing numbers of cells in G2 phase with increasing doses (Figure 2C). Site-specific DNA breaks detected by cometFISH FISH probes were targeted to the p and q arms of chromosome 5 at the regions 5p15.5 and 5q31 (Figure 3A). The appearance of probe-specific fluorescent signals in the tail, after HQ exposure at 50 lM (Figure 3B), showed breakage at these locations on the chromosome. Using the scoring criteria described in Figure 4, we quantified the amount of site-specific probe breakage in TK6 cells induced by the three compounds. Figure 4 shows the distribution of the probe fluorescence between the head and tail of the comet, comparing categories A and B with C and D. If all or some of the probe fluorescence is located in the comet tail (categories C and D), this shows significant breakage of the DNA at the probe location. Additional probe signals appearing in the head of the comet could result from aneuploidy, gene amplification and duplication rather than breakage. 323 P. A. Escobar et al. A 100 80 60 G1 S G2/M DNA Content 40 20 0 0 B 100 80 10 25 50 Hydro quinone [M] 75 DNA Content 60 40 20 0 DMSO 0.5 2.5 5 Melphalan [M] C 100 80 10 25 DNA Content 60 40 20 0 DMSO 0.5 2.5 5 Etoposide [M] 10 Fig. 2. Cell cycle analysis of TK6 cells exposed to different concentrations of (A) HQ, (B) melphalan and (C) etoposide. Cell cycle profiles were measured by PI staining to determine the percentage of cells in G1, S and G2/M phases. Each point represents the mean of at least two independent experiments. HQ at 25 and 50 lM caused a highly significant shift in location of the probe fluorescence from the head to the tail for FISH probes specific to both bands 5q31 and 11q23 (Table I). To test our a priori hypothesis, we directly compared the HQinduced specific loci breakage on both bands and found that HQ treatment caused significantly more breakage at 5q31 (OR: 5.34) in the tail than at 11q23 (OR: 2.82) (P 5 0.02, Table II). Melphalan, at doses that did not produce cytotoxicity or cell cycle changes, induced only modest breakage of probes to bands 5q31 and 11q23 and movement of the probes into the comet tail (Table I). This effect on bands, 5q31 (OR: 1.88) and 11q23 (OR: 1.32), was almost the same, but the OR at 5q31 was slightly more pronounced (Table II). Etoposide produced a large shift of probe fluorescence from the comet head to the tail especially at the higher dose of 2.5 lM. The effect of etoposide was slightly greater on band 11q23 (OR: 9.09) than 324 Fig. 3. CometFISH staining in human TK6 cells. Human TK6 cells were hybridized with a dual colour probe for chromosome bands 5q31 (42) and 5p15.5 (green). (A) Non-treated control cells. (B) Cells exposed to 50 lM HQ. The arrows show the probe signals located in the comet head and tail. on 5q31 (OR: 6.79), but this difference was not statistically significant (Table II). Another method of scoring the amount of site-specific damage in each band is to count the number of probe signals detected by the fluorescent probes in each cell. We therefore counted fluorescent probe spots and categorized them into four categories, IIV: I (12 spots), II (34 spots), III (57 spots) and IV (.8 spots) (Figure 4). Categorization in this fashion showed that HQ and melphalan caused site-specific damage and a shift from categories I and II to categories III and IV at both 25 and 50 lM (Table III). These two compounds induced a slightly higher shift at 5q31 than at 11q23. However, etoposide tended to induce more damage at 11q23 than 5q31, but these differential effects were not statistically significant (Table IV). Thus, in contrast to spot counting, the categorization of probe fluorescence into head and tail displayed a better assessment of the effects of the three leukaemogens to specific chromosomes. The data were further classified into two more categories (Figure 4); number of probe signals in comets with no tail (A) and comets with tails (B, C and D), and into what we considered low (AI, AII and BI), medium (AIII, BII, BIII and CII) and high (CIII, CIV, DIII and DIV) effects. The results from these classifications did not provide any additional information over that presented in Tables I and III. Discussion Measuring classical chromosomal aberrations in metaphase cells, either by Giemsa staining or FISH, is the most frequently used method to detect chromosomal damage. This method is time consuming and requires skilled personnel. We have explored the potential of cometFISH as a simpler alternative method to assess damage to specific chromosomal regions. CometFISH was introduced in 1997 and has been recently applied to study damage to the telomeres (37,38) and to specific genes such as p53, dihydrofolate reductase (DHFR) and the O6-methylguanine DNA methyltransferase (MGMT) Downloaded from http://mutage.oxfordjournals.org at UCSF Library and Center for Knowledge Management on June 4, 2010 Downloaded from http://mutage.oxfordjournals.org at UCSF Library and Center for Knowledge Management on June 4, 2010 Table I. Distribution of 5q31 and 11q23 probe signals in the comet head, head and tail or tail, in relation to chemical dose Treatment Dose (lM) Localization (%)a 5q31 11q23 Head Head/tail Tail Head Head/tail Tail HQ 0 25 50 Melphalan 0 0.5 2.5 Etoposide 0 0.5 2.5 96.67b 3.33 30.13 55.13 20.39 64.47 73 27 46.67 36 45.75 33.99 80.92 16.45 43.48 51.55 6 69.33 0 86.09 13.25 14.74 35.71 53.90 15.14 39.08 48.34 0 77 22 17.33 47.41 47.4 20.26 66 34 2.63 78.15 21.19 4.97 42.11 53.95 24.67 1.97 58.55 0.66 10.39 12.58 1 5.19 0 0.66 3.94 39.48 aLocalizations are classified from Figure 4 as head 5 A B, head/tail 5 C and tail 5 D. bThe data show the cumulative counts of three independent experiments; in each experiment, approximately 50100 cells were analysed. The percentage is calculated from 150 to 300 cells representing 100% total. Table II. ORs and statistical significance after ordered logistic regression analysis comparing the distributions of 5q31 and 11q23 probe signals in the comet head, head and tail or tail Chemical treatment ORa P-valueb 5q31 11q23 HQ Melphalan Etoposide 5.34 2.82 1.88 1.32 6.79 9.09 0.02 0.19 0.41 aOR for trends, i.e. the OR for increasing one category of damage from jumping one dose category of the chemical. For example, the OR of 5.34 represents the ratio of odds of comparing one category of damage to the next one higher for changing dose of HQ from 0 to 25 (or 25 to 50) among chromosome 5. bComparison between chromosome bands 5q31 and 11q23. Chemically induced damage to 5q31 and 11q23 (2832,34,39). To our knowledge, there are only two studies that used chromosome probes to study specific chromosomal damage (27,40) and these studies used whole-chromosome painting. In the present study, we quantified the amount of DNA damage induced by HQ, melphalan and etoposide in chromosome regions 5q31 and 11q23 using the cometFISH technology. These chemicals and chromosome locations were chosen because of their relationship to chemically induced leukaemia (5,7). Our a priori hypothesis was that the alkylating agent melphalan would produce more damage at band 5q31 than at 11q23 and, conversely, that the topoisomerase II inhibitor etoposide would cause more breakage at band 11q23 than at 5q31. For the benzene metabolite HQ, it was unclear which location, if any, would be preferred as it has been described as a topoisomerase II inhibitor (21), but is also known to cause deletion of 5q31 (20,33). Surprisingly, HQ exposure produced significantly more DNA breakage at 5q31 than at 11q23. Melphalan induced a modest overall effect, but tended to produce more breakage at 5q31 than 11q23 (Table I). Melphalan is a bifunctional alkylating agent that has the capability to induce DNA cross-linking. Cross-linking causes a retardation in DNA migration during the electrophoresis, which could account for the apparently low overall DNA breakage we observed with melphalan in the comet assay (Figure 1). Etoposide induced considerable DNA damage in a dose-dependant fashion, showing more damaging effects on chromosome 11q23 than at 5q31 at the higher dose. Etoposide is a topoisomerase II inhibitor that induces a wide variety of DNA damage, but our findings are in agreement with those of Ng et al. (41), which showed an apparent preference for the MLL gene at chromosome 11q23. In the studies presented here, etoposide induced more DNA breaks than HQ, and HQ induced more than melphalan (Table III). Melphalan and etoposide showed preferential effects that were in line with our a priori hypotheses, but to our surprise there was evidence of Fig. 4. CometFISH scoring criteria. 325 P. A. Escobar et al. Downloaded from http://mutage.oxfordjournals.org at UCSF Library and Center for Knowledge Management on June 4, 2010 Table III. Distribution of 5q31 and 11q23 probe signals in comets grouped by number of probe signalsl Chemical treatment Dose (lM) Percentage of comets with given number of probe signals (%)a 5q31 11q23 HQ Melphalan Etoposide 12 34 57 !8 12 34 57 !8 0 35.33b 54 9.33 1.34 44.37 45.03 9.97 0.66 25 5.77 39.74 41.03 13.46 12.99 33.12 31.82 22.08 50 12.50 30.92 42.11 14.47 17.88 38.41 34.44 9.27 0 37 43 19 1 27 46 23 4 0.5 18 30 36.67 15.33 20.78 38.31 35.06 5.85 2.5 29.41 30.07 17.65 22.87 32.67 40 22 5.33 0 32.89 51.97 13.16 1.98 45.7 35.1 14.57 4.63 0.5 22.98 37.27 31.68 8.07 17.76 31.58 42.77 7.89 2.5 6.67 20 38.67 34.66 2.63 23.68 41.45 32.24 aDNA probe signals are classified from Figure 4 as I 5 12 spots, II 5 34 spots, III 5 57 spots and IV ! 8 spots. bThe data show the cumulative counts of three independent experiments; in each experiment, approximately 50100 cells were analysed. The percentage is calculated from 150 to 300 cells representing 100% total. Table IV. ORs and statistical significance after ordered logistic regression analysis comparing the distributions of 5q31 and 11q23 probe signals in comets grouped by number of probe signals Chemical treatment ORa P-valueb 5q31 11q23 HQ Melphalan Etoposide 2.40 2.11 1.48 0.91 3.38 3.84 0.71 0.39 0.52 aOR for trends, i.e. the OR for increasing one category of damage from jumping one dose category of the chemical. bComparison between chromosome bands 5q31 and 11q23. a greater effect of HQ on 5q31 than at 11q23. This tends to contradict the hypothesis that HQ acts as an etoposide-like topo II inhibitor in one proposed mechanism for benzene-induced leukaemia (21,42). It does, however, agree with the notion that benzene-induced leukaemia mediated through the quinone pathway has similar characteristics to therapy-induced leukaemias produced by alkylating agents in which loss of band 5q31 is a common clonal event (43). Each of the chemicals tested produced a different pattern of DNA damage at bands 5q31 and 11q23. Typically, comet FISH data are classified as either head or tail effects. We devised a system to classify the damage using a 4 4 table (Figure 4). This helped us determine the presence of the chromosome probe in the head and/or tail of the comet and to quantify the level of site-specific breakage by measuring the number of specific chromosome signals present. As described in Methods, DNA damage was classified into four categories A, B, C and D that described the localization of probe signal in the comet (head, head and tail or only tail). Categorizing in this manner revealed differences between the chemicals that may be functionally important. We also tried scoring the amount of breakage in each chromosome band by counting the number of signals detected by the fluorescent probes in each cell, dividing them into four categories, IIV: I (12 spots), II (34 spots), III (57 spots) and IV (.8 spots) (Figure 4). Categorizing by this second procedure revealed fewer differences in the comparison between the three chemicals presented here, but may be useful in future studies. 326 The cometFISH approach described here can measure whole-genome DNA damage plus specific chromosomal region damage in a manner that is faster and more streamlined than the classical chromosomal aberration assay. It may also be a useful method for biomarker research in human populations where one wishes to study effects on specific chromosome regions because it does not require metaphase cells for analysis. Acknowledgements This study was supported by grants from P42 ES04705 and P30 ES01896 from the National Institute of Environmental Health Sciences (to M.T.S.). We are grateful to Weihong Guo for her laboratory assistance and generous help with the manuscript. Conflict of interest statement M.T.S. has received consulting and expert testimony fees from law firms representing both plaintiffs and defendants in cases involving exposure to benzene. References 1. Krivtsov, A. V., Twomey, D., Feng, Z. et al. (2006) Transformation from committed progenitor to leukaemia stem cell initiated by MLL-AF9. Nature, 442, 818822. 2. Taki, T. and Taniwaki, M. (2006) Chromosomal translocations in cancer and their relevance for therapy. Curr. Opin. Oncol., 18, 6268. 3. Smith, M. T., Skibola, C. F., Allan, J. M. and Morgan, G. J. (2004) Causal models of leukaemia and lymphoma. IARC Sci. Publ., 157, 373392. 4. Deschler, B. and Lubbert, M. (2006) Acute myeloid leukemia: epidemiology and etiology. Cancer, 107 (9), 2099107. 5. Larson, R. A. and Le Beau, M. M. (2005) Therapy-related myeloid leukaemia: a model for leukemogenesis in humans. Chem. Biol. Interact., 153154, 187195. 6. Pedersen-Bjergaard, J. (2005) Insights into leukemogenesis from therapyrelated leukemia. N. Engl. J. Med., 352, 15911594. 7. Felix, C. A. (2001) Leukemias related to treatment with DNA topoisomerase II inhibitors. Med. Pediatr. Oncol., 36, 525535. 8. Mamuris, Z., Dumont, J., Dutrillaux, B. and Aurias, A. (1990) Specific chromosomal mutagenesis observed in stimulated lymphocytes from patients with S-ANLL. Int. J. Cancer, 46, 563568. 9. Mamuris, Z., Gerbault-Seureau, M., Prieur, M., Pouillart, P., Dutrillaux, B. and Aurias, A. (1989) Chromosomal aberrations in lymphocytes of patients treated with melphalan. Int. J. Cancer, 43, 8086. 10. Mamuris, Z., Prieur, M., Dutrillaux, B. and Aurias, A. (1989) The chemotherapeutic drug melphalan induces breakage of chromosomes regions rearranged in secondary leukemia. Cancer Genet. Cytogenet., 37, 6577. 11. Amiel, A., Yukla, M., Yogev, S., Manor, Y., Fejgin, M. D. and Lishner, M. (2004) Deletion of 5q31 and 7q31 in patients with stable melphalan treated multiple myeloma. Cancer Genet. Cytogenet., 152, 8487. Chemically induced damage to 5q31 and 11q23 Downloaded from http://mutage.oxfordjournals.org at UCSF Library and Center for Knowledge Management on June 4, 2010 12. Felix, C. A., Kolaris, C. P. and Osheroff, N. (2006) Topoisomerase II and the etiology of chromosomal translocations. DNA Repair (Amst.), 5, 10931108. 13. Infante, P. F., Rinsky, R. A., Wagoner, J. K. and Young, R. J. (1977) Leukaemia in benzene workers. Lancet, 2, 7678. 14. Rinsky, R. A., Hornung, R. W., Silver, S. R. and Tseng, C. Y. (2002) Benzene exposure and hematopoietic mortality: a long-term epidemiologic risk assessment. Am. J. Ind. Med., 42, 474480. 15. Snyder, R. (2002) Benzene and leukemia. Crit. Rev. Toxicol., 32, 155210. 16. Smith, M. T., Yager, J. W., Steinmetz, K. L. and Eastmond, D. A. (1989) Peroxidase-dependent metabolism of benzene's phenolic metabolites and its potential role in benzene toxicity and carcinogenicity. Environ. Health Perspect., 82, 2329. 17. Smith, M. T. (1996) The mechanism of benzene-induced leukemia: a hypothesis and speculations on the causes of leukemia. Environ. Health Perspect., 104 (Suppl. 6), 12191225. 18. Zhang, L., Eastmond, D. A. and Smith, M. T. (2002) The nature of chromosomal aberrations detected in humans exposed to benzene. Crit. Rev. Toxicol., 32, 142. 19. Zhang, L., Rothman, N., Wang, Y. et al. (1998) Increased aneusomy and long arm deletion of chromosomes 5 and 7 in the lymphocytes of Chinese workers exposed to benzene. Carcinogenesis, 19, 19551961. 20. Zhang, L., Wang, Y., Shang, N. and Smith, M. T. (1998) Benzene metabolites induce the loss and long arm deletion of chromosomes 5 and 7 in human lymphocytes. Leuk. Res., 22, 105113. 21. Eastmond, D. A., Mondrala, S. T. and Hasegawa, L. (2005) Topoisomerase II inhibition by myeloperoxidase-activated hydroquinone: a potential mechanism underlying the genotoxic and carcinogenic effects of benzene. Chem. Biol. Interact., 153154, 207216. 22. Frantz, C. E., Chen, H. and Eastmond, D. A. (1996) Inhibition of human topoisomerase II in vitro by bioactive benzene metabolites. Environ. Health Perspect., 104 (Suppl. 6), 13191323. 23. Singh, N. P., McCoy, M. T., Tice, R. R. and Schneider, E. L. (1988) A simple technique for quantitation of low levels of DNA damage in individual cells. Exp. Cell Res., 175, 184191. 24. Tice, R. R., Agurell, E., Anderson, D. et al. (2000) Single cell gel/comet assay: guidelines for in vitro and in vivo genetic toxicology testing. Environ. Mol. Mutagen., 35, 206221. 25. Collins, A. R. (2004) The comet assay for DNA damage and repair: principles, applications, and limitations. Mol. Biotechnol., 26, 249261. 26. Santos, S. J., Singh, N. P. and Natarajan, A. T. (1997) Fluorescence in situ hybridization with comets. Exp. Cell Res., 232, 407411. 27. Rapp, A., Bock, C., Dittmar, H. and Greulich, K. O. (2000) UV-A breakage sensitivity of human chromosomes as measured by COMET-FISH depends on gene density and not on the chromosome size. J. Photochem. Photobiol. B, 56, 109117. 28. McKelvey-Martin, V. J., Ho, E. T., McKeown, S. R., Johnston, S. R., McCarthy, P. J., Rajab, N. F. and Downes, C. S. (1998) Emerging applications of the single cell gel electrophoresis (Comet) assay. I. Management of invasive transitional cell human bladder carcinoma. II. Fluorescent in situ hybridization comets for the identification of damaged and repaired DNA sequences in individual cells. Mutagenesis, 13, 18. 29. McKenna, D. J., Rajab, N. F., McKeown, S. R., McKerr, G. and McKelvey-Martin, V. J. (2003) Use of the comet-FISH assay to demonstrate repair of the TP53 gene region in two human bladder carcinoma cell lines. Radiat. Res., 159, 4956. 30. Schaeferhenrich, A., Sendt, W., Scheele, J., Kuechler, A., Liehr, T., Claussen, U., Rapp, A., Greulich, K. O. and Pool-Zobel, B. L. (2003) Putative colon cancer risk factors damage global DNA and TP53 in primary human colon cells isolated from surgical samples. Food Chem. Toxicol., 41, 655664. 31. Horvathova, E., Dusinska, M., Shaposhnikov, S. and Collins, A. R. (2004) DNA damage and repair measured in different genomic regions using the comet assay with fluorescent in situ hybridization. Mutagenesis, 19, 269276. 32. Kumaravel, T. S. and Bristow, R. G. (2005) Detection of genetic instability at HER-2/neu and p53 loci in breast cancer cells sing Comet-FISH. Breast Cancer Res. Treat., 91, 8993. 33. Stillman, W. S., Varella-Garcia, M. and Irons, R. D. (2000) The benzene metabolite, hydroquinone, selectively induces 5q31- and -7 in human CD34CD19- bone marrow cells. Exp. Hematol., 28, 169176. 34. Amendola, R., Basso, E., Pacifici, P. G., Piras, E., Giovanetti, A., Volpato, C. and Romeo, G. (2006) Ret, Abl1 (cAbl) and Trp53 gene fragmentations in comet-FISH assay act as in vivo biomarkers of radiation exposure in C57BL/6 and CBA/J mice. Radiat. Res., 165, 553561. 35. McCullagh, P. and Nelder, J. (1989) Generalized Linear Models. CRC Press, London. 36. Zeger, S. L. and Liang, K. Y. (1986) Longitudinal data analysis for discrete and continuous outcomes. Biometrics, 42, 121130. 37. Hovhannisyan, G., Rapp, A., Arutyunyan, R., Greulich, K. O. and Gebhart, E. (2005) Comet-assay in combination with PNA-FISH detects mutagen-induced DNA damage and specific repeat sequences in the damaged DNA of transformed cells. Int. J. Mol. Med., 15, 437442. 38. Arutyunyan, R., Rapp, A., Greulich, K. O., Hovhannisyan, G., Haroutiunian, S. and Gebhart, E. (2005) Fragility of telomeres after bleomycin and cisplatin combined treatment measured in human leukocytes with the Comet-FISH technique. Exp. Oncol., 27, 3842. 39. Knobel, Y., Glei, M., Weise, A., Osswald, K., Schaferhenrich, A., Richter, K. K., Claussen, U. and Pool-Zobel, B. L. (2006) Uranyl nitrilotriacetate, a stabilized salt of uranium, is genotoxic in nontransformed human colon cells and in the human colon adenoma cell line LT97. Toxicol. Sci., 93, 286297. 40. Harreus, U. A., Kleinsasser, N. H., Zieger, S., Wallner, B., Reiter, M., Schuller, P. and Berghaus, A. (2004) Sensitivity to DNA-damage induction and chromosomal alterations in mucosa cells from patients with and without cancer of the oropharynx detected by a combination of comet assay and fluorescence in situ hybridization. Mutat. Res., 563, 131138. 41. Ng, A., Taylor, G. M. and Eden, O. B. (2006) Genotoxicity of etoposide: greater susceptibility of MLL than other target genes. Cancer Genet. Cytogenet., 164, 164167. 42. Whysner, J., Reddy, M. V., Ross, P. M., Mohan, M. and Lax, E. A. (2004) Genotoxicity of benzene and its metabolites. Mutat. Res., 566, 99130. 43. Irons, R. D. and Stillman, W. S. (1996) The process of leukemogenesis. Environ. Health Perspect., 104 (Suppl. 6), 12391246. Received on November 3, 2006; revised on April 10, 2007; accepted on April 13, 2007 327