micronucleus test to study the genotoxicity of four trihalomethanes. F.Le Curieux1-2, L.Gauthier3, F.Erb2 and D.Marzin1*4. 'Laboratory of Genetic Toxicology, ...
Mutagenesis vol.10 no.4 pp.333-341, 1995
Use of the SOS chromotest, the Ames-fluctuation test and the newt micronucleus test to study the genotoxicity of four trihalomethanes
F.Le Curieux1-2, L.Gauthier3, F.Erb2 and D.Marzin1*4 'Laboratory of Genetic Toxicology, Pasteur Institute of Lille, 1 rue du Professeur Calmette, BP 245, 59019 Lille Cedex, 2 Department Toxicology—Hydrology—Hygiene, Faculty of Pharmacy, 3 rue du Professeur Laguesse, BP 83, 59006 Lille Cedex and 3 Centre de Biologic du Diveloppement, UA 675 (CNRS), University Paul Sabatier, 118 Route de Narbonne, 35062 Toulouse Cedex, France •"To whom correspondence should be addressed at: Institut Pasteur de Lille, Laboratoire de Toxicologie g£n£tique, 1 rue du Professeur Calmette, BP 245, 59019 Lille Cedex, France
Three short-term assays (the SOS chromotest, the Amesfluctuation test and the newt micronucleus test) were carried out to evaluate the genotoxicity of four trihalomethanes (chloroform, bromodichloromethane, chlorodibromomethane and bromoform). With the SOS chromotest, all the chemicals studied except chloroform were found to induce primary DNA damage hi Escherichia coli PQ37. In the Ames-fluctuation test, only bromoform showed mutagenic activity on Salmonella typhimurium strain TA100. The newt micronucleus assay detected a clastogenic effect on the peripheral blood erythrocytes of Pleurodeles waltl larvae for bromodichloromethane and bromoform. It appeared that the presence of bromine substituent(s) generally led to significant genotoxic activity. Moreover, the use of the metabolic system significantly increased the genotoxicity of the brominated trihalomethanes in the SOS chromotest Unlike previous investigations in which the SOS chromotest was always the least interesting assay, this study exhibited the good efficiency of this in vitro test on E.coli for the detection of trihalomethanes with bromine substituents.
Introduction The chlorinated and/or brominated trihalomethanes were the first organic halides to be identified as chlorination by-products in drinking water samples (Rook, 1974; Bellar et al, 1974): since the late 1970s and the early 1980s, bromoform (BF), bromodichloromethane (BDCM), chlorodibromomethane (CDBM) and above all chloroform (CF) were the subject of increasing concern. Consequently, many investigations were implemented in the different fields of toxicology to assess the health effects hazards related to the occurrence of these chemicals in drinking water. It was shown that chlorinated drinking water extracts induced genotoxic activity in vitro (Kool et al., 1982; Horth, 1989) and that the genotoxicity detected was related to the amount of AOX (organohalides adsorbable on activated carbon) contained in the drinking water sample (Kool et al., 1984). The genotoxicity observed in drinking water also appeared to be due to non-volatile chlorination by-products (Fielding and Horth, 1986; Meier, 1988). It is now widely accepted that the genotoxicity detected in drinking water mainly originates from the reaction of chlorine with natural organic substances i.e. humic substances © Oxford University Press
which leads to the formation of organohalogenated derivatives. This issue was reviewed by Meier (1988). In the present study, three short-term genotoxicity tests were implemented as follows: • The SOS chromotest is an in vitro assay showing primary DNA damage on Escherichia coli. This assay is a simple, efficient and rapid genotoxicity test that can easily be adapted to the study of environmental water samples (Bourbigot et al, 1986; Xu et al, 1987; Dutka et al, 1986, 1987; Langevin et al, 1992). The main advantage of the SOS chromotest lies in the fact that its results are obtained within 8 h. • The Ames-fluctuation test is an in vitro test detecting point mutation on Salmonella typhimurium. This assay is the modification, in liquid medium, of the widely used Ames/ Salmonella assay. Because of its greater sensitivity, the fluctuation test is very well suited to the search for mutagenicity in water samples (Wilcox and Denny, 1985; Monarca et al, 1985). The results of this assay are obtained in 3-4 days. • The newt micronucleus test is an in vivo assay demonstrating chromosomal aberrations on peripheral blood erythrocytes of the amphibian Pleurodeles waltl larvae. This assay is also able to detect loss or gain of chromosomes or chromatids and thus aneuploidy. This test has been, since its beginning, intended for use in the monitoring of fresh, polluted or drinking water genotoxicity (Jaylet et al, 1986, 1987; Fernandez et al, 1989). A previous study had shown the greater sensitivity of the newt micronucleus test compared to the mouse micronucleus test for the detection of three carcinogens (Le Curieux et al, 1992). The results of a whole micronucleus test are obtained in 3—4 weeks. Another limitation of this assay on the newt P.waltl is that the natural laying period of this animal is from September to May. An extension of the laying period can be obtained by injecting LHRH (luteinizing hormone releasing hormone) to the two partners (Fernandez et al, 1993). These three tests present a diversity and a complementarity in terms of biological materials used, genetic endpoints investigated (primary DNA damage, gene mutations and chromosomal aberrations) and duration of the exposure to the genotoxicant (Le Curieux et al, 1993). The aim of the present work was to carry out the SOS chromotest, the Ames-fluctuation test and the newt micronucleus test to assess the genotoxicity of four trihalomethanes found in drinking water samples (CF, BCDM, CDBM and BF). Materials and methods Chemicals The main characteristics of the four trihalomethanes studied are shown in Table I. Genotoxicity tests SOS chromotest. The tester strain E.coli PQ37, was kindly given by M.Hofnung (Institut Pasteur, Paris, France). Its genetic features were described by Quillardet and Hofnung (1985). The SOS chromotest was performed as
333
F.Le Curieux et al
Table I. Characteristics of the four trihalomethanes studied Chemicals Name
Abbreviation Molecular formula
Chloroform Bromodichloro me thane Chlorodibromome thane Bromoform
CF BDCM CDBM BF
CHCI3 BrCHCl2 CICHBr2 CHBr3
Mol wt (g/mol)
Density
Solubility (mg/1 H2O at 20°C)
Solvent used
Supplier
Purity (%)
CAS No.
119.38 163.83 208.3 252.75
1.49 1.98 2.45 2.82
8000 — 3200 —
DMSO DMSO DMSO DMSO
Fluka Fluka Fluka Fluka
IR1 >98 >97 97
67-66-3 75-27-4 124-48-1 75-25-2
"Consistent infra-red spectrum.
Table II. Results of the SOS chromotest on the four trihalomethanes studied Chemicals
Chloroform (CF) Bromodichloromethane (BDCM) Chlorodibromomethane (CDBM) Bromoform (BF)
S9*
Range of concentrations studied (|ig/ml in the assay)
Threshold toxic concentrations6 (Hg/ml in the assay)
10-10 000 10-10 000 3-1000 1-3000 3-3000 1-3000 3-1000 3-3000
3000 3000 1000 1000 1000 1000 1000 1000
Range of genotoxic concentrations (Hg/ml in the assay) 1;
c d
(6O-2O0) 3-300 30-300 10-300 60-300 10-300
Maximum induction factor
1.07 1.10 1.58 3.29 1.84 2.58 1.86 2.60
Test performed in the presence (+) or the absence ( —) of S9-mix metabolic system. ""Lowest concentration inducing a decrease of >5fJ% in the alkaline phosphatase compared to solvent control. c —, No genotoxic effect detected on E.coli PQ37. ''Weak genotoxic activity (induction factor close to 1.5). recommended by these researchers and following the adaptation made by Marzin et al. (1986). After incubation, the mixtures were divided into two series, one for p-galactosidase activity measurement (an induction assay) and the other for alkaline phosphatase (a control of protein synthesis). In our study, a chemical is considered as toxic when it induces a decrease of >50% in the alkaline phosphatase activity compared to the solvent control. The measurement of enzymatic activities was performed using microplates and an automatic microplate reader as described by Xu et al. (1989). The genotoxic activity for concentration c may be expressed in the ratio Rc = p/p where P represents P-galactosidase activity (in mlU) and p, phosphatase alkaline activity (in mlU). The induction factor for a compound at concentration c is defined as lc = ff^/Zfo- ' n which Ro is the spontaneous ratio measured in the blank test (solvent control). Experiments were performed with and without the metabolic system for every trihalomethane studied. To ensure the validity of the assay, a positive control was included in each experiment The positive controls used were 4-nitroquinoline 1-oxide (4NQO; 1 Hg/ml) without S9-mix and benzo[a]pyrene (B[a]P; 30 Hg/ml) with S9-mix. Compounds were tested at least twice (two independent assays) using six experimental points for each dose. A compound is considered as an SOS repair system inducer in E.coli if the four following conditions were fulfilled (Olivier and Marzin, 1987): (i) the induction factor was >1.5; (ii) the p-galactosidase activity was significantly increased compared to the solvent control; (iii) the induction factor versus concentration graph shows a dose—effect relationship; and (iv) the result was reproducible. Ames-fluctuation test. The tester strain, S.typhimurium strain TAI00, was kindly provided by B N.Ames (University of California, Berkeley, USA). The genetic features of the strain were described by Maron and Ames (1983). We used strain TA100 in all the experiments because it is considered to be the strain most sensitive to water mutagens and particularly to chlorinated byproducts (Loper, 1980; Forster et al., 1983; Harrington et al., 1983; Vartiainen and Liimatainen, 1986; Fielding and Horth, 1987; Meier, 1988). The Amesfluctuation test was performed as described by Hubbard et al. (1984). This assay is a modification of the Ames test: briefly, the compound under study is exposed to bacteria in a liquid medium in many replicate cultures (96-well microplate) instead of the agar plate used in the Ames assay. After the 3-day incubation, bromothymol blue (600 |ig/ml) was added. Positive wells (containing prototrophic mutants) turned yellow, whereas negative wells remained green. For each experiment, spontaneous reversion in response to the solvent used (DMSO) was included. The positive response
334
to a standard mutagen was also performed to ensure the validity of the test, i.e. sensitivity of the bacteria and of the metabolic system when used. The positive control used was 1 ng/ml 4-NQO without the metabolic system and 60 Hg/ml cyclophosphamide (CPA) with the metabolic system All the trihalomethanes studied were tested with and without the metabolic system. Compounds were tested at least twice (two independent assays) using three experimental points for each concentration. A chemical was considered as toxic when it produced a significant decrease in the number of positive wells compared to the previous concentration. The statistical significance of the results was assessed with the x 2 test: P < 0.05 for x2 > 3 84 and P < 0.01 for %2 > 6.63 (Green et al, 1976). A compound was considered as mutagenic: (i) if it induced a statistically significant increase in the number of positive wells compared to the solvent control; (ii) if a dose—effect relationship was noticeable; and (iii) if the result was reproducible. Newt micronucleus test. The test-organisms, P.waltl larvae, were provided by the Centre de Biologic du DeVeloppement, University P.Sabatier (Toulouse, France) and acclimatized for 1 week before the experiment. The assay was performed as described by Jaylet et al. (1986) and according to the recommendations of the French standard AFNOR T-90-325 (1987) revised in 1992 (AFNOR, 1992). Briefly, larvae at stage 53 of the developmental table established by Gallien and Durocher (1957) were used for the experiment, since at this stage the mitotic index is at a maximum (Deparis, 1973). The highest concentration to be used in the 12-day micronucleus assay was defined as half the minimum concentration that led to detectable physiological disturbances (weight loss, swelling, reduction in food intake, swimming in circles, etc.) in a 6-day preliminary toxicity test. In every glass container, the water (containing the dissolved chemical at the required concentration) was renewed daily. After 12 days of treatment, blood samples were taken from every animal (15-20 larvae per concentration) by cardiac puncture into an hepannized rrucropipette. The slides (one for each animal) were stained with Masson acid Hemalun and the number of micronucleated erythrocytes was counted in a sample of 1000 erythrocytes. As there is not a normal distribution of micronucleus frequencies, median values and quartiles were calculated instead of mean values. The statistical method of MacGill et al. (1978), a quick and reliable test adapted to small sample sizes and values that do not have normal distribution, was used to analyse the results. A compound was considered as genotoxic: (1) if it induced a statistically significant increase in the number of micronucleated erythrocytes compared to the control (P < 0.05); and (ii) if the median for the treated group was at
Genotoxfclty of four trihalomethanes
B
2,8 2,6 2,4 22 2 1.8
GC
o & 1,6
I Q Z
1,4 '
1 0,8 0,6 0,4 02 0
10
100
10
1000
CONCENTRATIONS (u.gftnl)
C
100
1000
10000
CONCENTRATIONS Gig/ml)
2.8 2,6 2,4 22 2 1.8
100 (ig/1 (Bellar et al, 1974). CF was classified in group 2B by the International Agency for Research on Cancer (IARC, 1991): it is possibly carcinogenic in humans based on sufficient evidence for carcinogenicity to animals and inadequate data in humans (IARC, 1987). The GV set by WHO for the concentrations of CF in drinking water, was initially 30 (ig/1 (WHO, 1985), but was modified and increased up to 200 (j.g/1 (WHO, 1994). This GV is the concentration in drinking water associated with an excess lifetime cancer risk of 10~5 (one additional cancer per 100 000 of the population ingesting drinking water containing the substance at the GV for 70 years). Most in vitro tests for detecting point mutations or DNA damage are negative for CF (IARC, 1979; De Serres and Ashby, 1981). In human lymphocytes and CHO cells treated in vitro, CF did not induce chromosomal aberrations or sisterchromatid exchanges (White et al., 1979; Kirkland et al., 1981; Perry and Thomson, 1981). It did not cause unscheduled DNA synthesis in vitro in human lymphocytes (Perocco and Prodi, 1981) or in rat hepatocytes (Pienta and Kawalek, 1981; Althaus et ai, 1982). In vivo, CF was inactive in the micronucleus' test on mouse bone marrow cells (Salamone etal, 1981; Tsuchimoto and Matter, 1981; Gocke et al, 1981) and in the unscheduled DNA synthesis test on rat hepatocytes (Mirsalis et al, 1982). The negative results given by the three short-term assays performed in this study agree with the trend reported in the literature. BDCM BDCM is currently used only as a standard in the analysis of drinking water (Strobel and Grummt, 1987). This compound was identified in drinking water as by-product of the reaction of chlorine (added during drinking .water treatment) on natural organic matter in the presence of bromide ions. BDCM was detected in drinking water samples worldwide at concentrations ranging from 1 to 50 ng/1 (IARC, 1991). It was also found in a few untreated (raw) water samples but at much lower concentrations. BDCM is a major organic halide produced by
10
15 20 25 30 35 CONCENTRATIONS (ug/ml)
40
45
Fig. 3. Results for the four trihalomethanes in the newt micronucleus test
marine algae. It was classified in group 2B because it is carcinogenic in animals and possibly carcinogenic in humans (IARC, 1991). WHO fixed at 60 ng/1 the maximum concentration of BDCM permissible in drinking water (WHO, 1994). As for CF, this GV is the concentration corresponding to an additional lifetime cancer risk of 10~5. Our investigation indicated that BDCM was genotoxic on strain PQ37 of E.coli and on newt larvae erythrocytes but not on strain TA100 of S.typhimurium. In previous studies, the results regarding the mutagenicity of BDCM on S.typhimurium strain TA100 were not consistent: Simmon et al. (1977) showed the mutagenic activity of this chemical in the absence of S9-mix, whereas Strobel and Grummt (1987), Mersch-Sunderman (1989) and Khudoley et al (1989) demonstrated a positive effect in the presence of S9-mix only; moreover, several other studies (Mortelmans et al, 1986; Khudoley et al, 1987; Varma et al, 1988) did not find any genotoxicity of BDCM on strain TA100. The 337
F.Le Curieux et al
Table V. Summary of the results obtained with the three tests on the four trihalomethanes studied Chemicals
CF
S9 S9 BDCM S9 S9 CDBM S9 S9 BF S9 S9
Sensitivity1'
Tests'
+ + + +
SOS Chromotest
Ames-fluctuation
Newt micronucleus
_ ± + + + + +
+ 300 -
-c
60 3 30 10 60 10
+ 50 -
All three tests are negative SOS > MN Ames-fluctu. negative SOS positive Ames-fluctu. and MN negative MN > SOS > Ames-fluctu.
+ 2.5
*+, significant genotoxic activity; the lowest concentration inducing a genotoxic effect is reponed (ng/ml in the assay); ±, significant but weak genotoxic activity; —, no significant genotoxic activity. b As already described in previous studies (Le Curieux el al, 1993, 1994a,b), the sensitivity of a test was defined as follows: for a given compound, test A is more sensitive than test B if the lowest genotoxic concentration (LGC) in test A is lower than the LGC in test B; MN: newt micronucleus test. c Fernandez el al (1993).
same inconsistency was observed in the results on strains TA98 and TA1537, whereas this chemical was negative on strain TA1535 and positive, with S9-mix, on strain TA97. The study of Simon et al. (1977) stressed the difficulty encountered in detecting the mutagenic effect of BDCM: this trihalomethane revealed its activity only when the agar overlay plates were placed uncovered in a desiccator containing BDCM; the classic Ames test procedure did not detect any mutagenicity of this chemical on strain TA100 with or without S9-mix. The difficulty of demonstrating the mutagenic potency of BDCM in the fluctuation test is certainly to be related to the strong volatility of this chemical. In our study, within the range of concentrations tested (which include concentrations with strong bacteriostatic activities) no concentration induced significant mutagenicity. In their review, Ashby and Tennant (1991) considered BDCM as a mutagenic compound in the gaseous phase. Moreover, the work of Hayashi et al. (1988) indicated that BDCM did not lead to the formation of micronuclei in mice bone marrow cells after per os administration. Our positive result in the newt micronucleus test demonstrated the greater sensitivity of this assay compared to the mouse micronucleus test for the detection of the clastogenicity of BDCM. This last point (i.e. greater sensitivity of the newt micronucleus) had already been observed for other chemicals such as p"-naphthylamine (Le Curieux et al., 1992), sodium hypochlorite (Le Curieux et al., 1993) and several chlorinated and brominated acetonitriles (Le Curieux et al., 1994b). Other genotoxic effects of BDCM were reviewed by the National Toxicology Program (1993): this chemical was not mutagenic in the mouse lymphoma L5178Y/TK+/~ assay in the presence of S9-mix. Cytogenetic tests with Chinese hamster ovary (CHO) cells demonstrated no induction of chromosomal aberrations or sister-chromatid exchanges following treatment with BDCM either with or without the metabolic system. Concerning the carcinogenicity of BDCM, 2-year gavage studies indicated clear evidence of carcinogenic activity for male and female F344/N rats and B6C3F! mice (National Toxicology Program, 1993): increased incidences of tubular cell adenomas and adenocarcinomas in the kidney, and adenocarcinomas and adenomatous polyps in the large intestine in male and female rats, increased incidences of tubular cell adenomas and adenocarcinomas in the kidney of male mice, and increased incidences of hepatocellular adenomas and carcinomas in female mice. 338
Regarding the comparison of the sensitivity of the three tests implemented in the present study, the SOS chromotest (3 |ig/ml) was clearly more sensitive than the newt micronucleus test (50 |!g/ml), while the Ames-fluctuation test was negative. This particular point will be addressed later. In comparing the sensitivity of the SOS chromotest and the newt micronucleus test, the significant difference between the biological materials used and, above all, between the genetic endpoints detected in each test, must be emphasized: on the one hand, the SOS chromotest is an in vitro test on cultured bacteria which only detects primary DNA alterations; on the other hand, the newt micronucleus test is an in vivo assay detecting spindle poison activity and/or clastogenic effects in amphibian peripheral erythrocytes. Thus, the SOS chromotest can be positive (through the SOS repair system)'following lesions such as DNA adducts, apuric and apyrimidic sites, strand breaks, cross-links and gaps, whereas the newt' micronucleus test will be positive only if loss or gain-of chromosomes and/or chromosome breakages are induced. .
