Micronucleus induction in mice exposed to diazoaminobenzene or its metabolites, benzene and aniline: implications for diazoaminobenzene carcinogenicity

Mutation Research 521 (2002) 201–208

Micronucleus induction in mice exposed to diazoaminobenzene or its metabolites, benzene and aniline: implications for diazoaminobenzene carcinogenicity Nancy B. Ress a , Kristine L. Witt b , Jing Xu c , Joseph K. Haseman a , John R. Bucher a,∗ a

National Institute of Environmental Health Sciences, 79 Alexander Drive, Mail Drop EC-34, Research Triangle Park, Triangle Park, NC 27709, USA b ILS, Research Triangle Park, Triangle Park, NC 27709, USA c SITEK Research Laboratories, Rockville, MD 20850, USA

Received 14 August 2002; received in revised form 10 September 2002; accepted 11 September 2002

Abstract Diazoaminobenzene (DAAB), a manufacturing intermediate metabolized primarily to the known carcinogens benzene and aniline, has been identified as an impurity in a number of dyes and coloring agents that are components of cosmetics, food products, and pharmaceuticals. Several structural analogs of DAAB are carcinogenic as well. DAAB was selected for metabolism and toxicity studies by the National Toxicology Program (NTP) based on the potential for human exposure, positive Salmonella data, and lack of adequate toxicological data. In the toxicology studies in mice, DAAB exhibited properties similar to benzene and aniline. Because both these metabolites induce micronuclei (MN) in rodent bone marrow erythrocytes, DAAB was tested for induction of micronuclei in male B6C3F1 mice. DAAB was administered twice by corn oil gavage at 24 h intervals, at doses of 25, 50, and 100 mg/kg per day. In addition, comparative micronucleus tests were conducted with benzene, aniline, and a mixture of benzene plus aniline; doses were based on the respective molar equivalents of each metabolite to DAAB. It was hypothesized that any observed increase in micronuclei seen in DAAB-treated mice would be due primarily to the effects of the benzene metabolite, as benzene is a more potent inducer of chromosomal damage than aniline. Results of this study showed that DAAB and benzene were effective inducers of micronuclei, with stronger responses noted for DAAB at higher doses. Positive results were also obtained with the mixture of benzene and aniline, although the magnitude of the response was lower than for DAAB. Aniline gave a weak positive response at doses exceeding its molar equivalent to 100 mg/kg DAAB. Overall, the data indicated that DAAB is a potent inducer of micronuclei in mice, and its activity appears to be closely related to the activity of benzene, one of its primary metabolites. The results are consistent with a prediction of carcinogenicity for DAAB. Crown Copyright © 2002 Published by Elsevier Science B.V. All rights reserved. Keywords: Bone marrow; Chromosome damage; Carcinogenicity; Erythrocyte; Mixtures

1. Introduction ∗ Corresponding author. Tel.: +1-919-541-4532; fax: +1-919-541-4255. E-mail address: [email protected] (J.R. Bucher).

Diazoaminobenzene (DAAB) is used as an intermediate during organic synthesis and in the

1383-5718/02/$ – see front matter. Crown Copyright © 2002 Published by Elsevier Science B.V. All rights reserved. PII: S 1 3 8 3 - 5 7 1 8 ( 0 2 ) 0 0 2 4 1 - 3

202

N.B. Ress et al. / Mutation Research 521 (2002) 201–208

Fig. 1. Proposed pathway for DAAB metabolism [5].

manufacture of dyes and insecticides [1]. DAAB has been identified as a contaminant in D&C Red No. 33 and FD&C Yellow Nos. 5 and 6, which are permitted for use in ingested and externally applied drugs and cosmetics [2,3]. The maximum level of DAAB allowable in FD&C Yellow Nos. 5 and 6 is 40 ppb and in D&C Red No. 33 is 125 ppb. However, it has been identified in commercial products at concentrations up to 439 ppb with an average level of 99 ppb and in drugs at concentrations up to 110 ppb [3]. Previous work has shown that DAAB is reduced by P450 reductase or gut microflora to aniline and a phenyl diazenyl radical that decomposes to form benzene and nitrogen (Fig. 1; [4]). In support of this pathway, benzene, aniline and their metabolites were detected in the urine of mice following oral administration of DAAB, in the urine of rats treated intravenously, orally, and dermally, and in the blood of rats treated orally with DAAB [5]. The metabolites detected in the urine composed approximately 60% of the administered DAAB dose [5]. Benzene also was detected in the breath of rats receiving DAAB suggesting that benzene and its metabolites were available systemically [5]. Metabolites of benzene and aniline were formed in an in vitro study using human liver slices incubated with DAAB [5]. DAAB exhibits toxicological properties similar to its metabolites benzene and aniline. Methemoglobinemia, increased Heinz body formation, splenomegaly, splenic hematopoiesis, atrophy of the lymphoid tissue, and epidermal hyperplasia and ulceration were observed in rats and mice exposed dermally to DAAB

for 2 weeks [6]. Increased methemoglobin and Heinz body formation, as well as splenotoxicity, are common responses seen in short-term in vivo studies with aniline and as such are strongly associated with aniline formation in DAAB-treated animals [7,8]. Splenic hematopoiesis also has been observed in mice following chronic exposure to benzene [6]. Lymphoid depletion of the spleen was observed in a 2-year oral gavage study with benzene [6]. Similar effects were seen in DAAB-treated animals and were associated with benzene formation [9]. The similarity in the toxic response between benzene and DAAB is not as great as that for aniline and DAAB; however, there is concordance in the site of action. The present studies were performed to further explore the association between the toxicological effects of DAAB and its metabolites. Both benzene and aniline have been shown to induce increased frequencies of micronucleated erythrocytes in bone marrow and peripheral blood of mice [10–14]. However, benzene is more potent than aniline in in vivo mutagenicity assays. Based on the metabolism of DAAB, an increase in micronuclei (MN) in the bone marrow of mice treated with DAAB was expected. It also was anticipated that, based on the relative potencies of benzene and aniline in induction of MN, the response seen in DAAB exposed animals would be more closely related to the formation of benzene than aniline. To test this hypothesis, mouse bone marrow MN tests were conducted with DAAB, benzene, aniline, and a combination of benzene and aniline that was equimolar to DAAB.

N.B. Ress et al. / Mutation Research 521 (2002) 201–208

2. Materials and methods 2.1. Chemicals DAAB (CAS# 136-35-6; Lot No.: A010092501, 97% pure) and Benzene (CAS# 71-43-2; Lot No. 00358 CS, 99.9% pure) were obtained through the National Institute of Environmental Health Sciences from the Analytical Chemistry Laboratory, Midwest Research Institute (Kansas City, MO). Aniline (CAS# 62-53-3; Lot No. 51K0111; 99% pure) was obtained from Sigma. Additional standard laboratory chemicals and reagents used in the MN experiments included cyclophosphamide (CAS# 6055-19-2; Sigma, St. Louis, MO), corn oil (CAS# 8001-30-7; Sigma), acridine orange (CAS# 65-61-2; Sigma), fetal bovine serum (Irvine Scientific, Santa Ana, CA) and phosphate buffered saline (Gibco, Grand Island, NY). 2.2. Test animals Male B6C3F1 mice were obtained at 4 weeks of age from Taconic Laboratory Animals and Services, Germantown, NY. The animals were quarantined for at least 10 days and then maintained until they were 9–16 weeks of age, when they were randomly assigned to dosage groups for the studies. The animals were permitted food (NIH-07, Zeigler Bros. Inc.) and tap water ad libitum, with a 12 h light/dark cycle, at 23 ± 3 ◦ C and 50 ± 15% humidity. All studies were conducted at SITEK Research Laboratories, Rockville, MD, accredited by the American Association for the Accreditation of Laboratory Animal Care (AAALAC, Rockville, MD). Institutional Animal Care and Use Committees approved the experimental protocols. Animal use was in accordance with the United States Public Health Service policy on humane care and use of laboratory animals. 2.3. Micronucleus tests All test compounds were given by gavage in a volume of 10 ml/kg body weight in corn oil. The standard three-exposure protocol used by the National Toxicology Program (NTP) [15], was modified as described below. Male B6C3F1 mice, five per treatment group, were administered the test compounds twice, at 0 and at 24 h. Vehicle control animals received corn

203

oil alone. The positive control mice received 20 mg/kg cyclophosphamide by gavage. Twenty-four hours after the second treatment, the mice were killed with CO2 and smears of the bone marrow cells obtained from the femurs were prepared. Air-dried smears were fixed in absolute methanol and stained with acridine orange [16]. Slides were evaluated at 1000× magnification using epi-illuminated fluorescence microscopy (450–490 nm excitation, 520 emission). Two thousand polychromatic erythrocytes (PCE) were scored per animal for frequency of MN cells. In addition, the percentage of polychromatic erythrocytes (%PCE) among 200 total erythrocytes was determined as a measure of chemical-induced bone marrow toxicity. Three in vivo MN tests were conducted: a range finding study, a preliminary investigation, and a definitive MN test. In the range finding study, benzene doses were selected based on information found in the literature and with a goal of obtaining an unequivocal response. Thus, mice were exposed to 1000, 1500, 2000, 2250, or 2500 mg/kg benzene. Because of a lack of information in the literature, a range finding study using doses of 50, 100, or 500 mg/kg DAAB was conducted. Information from the range finding study was used to determine doses for the preliminary investigation and definitive MN test. In the preliminary investigation, mice received DAAB, benzene, aniline, or a mixture of benzene and aniline. The composition of the mixture of benzene and aniline was based on the relative percentage of benzene and aniline in the DAAB molecule by weight (data not shown). Due to the large amount of interanimal variability observed in MN response in the preliminary study, a definitive MN test using the same treatment groups (Table 1) was conducted. In the definitive MN test, the doses of benzene and aniline for the single component studies and for the mixture study were adjusted to compare MN-PCE Table 1 Doses (mg/kg) used for the definitive MN test Group

DAAB

Benzene

Definitive MN test 1 25 10 2 50 20 3 100 40 4 5

Aniline

Mixture (benzene + aniline)

12 23 47 120 470

10 + 12 20 + 23 40 + 47

204

N.B. Ress et al. / Mutation Research 521 (2002) 201–208

induction based on their respective molar equivalents to DAAB (Table 1). Two higher aniline doses were also used (Groups 4 and 5, Table 1) to provide better dose response information than was obtained in the preliminary study which used doses no higher than 100 mg/kg.

Table 2 Range finding study: comparison of DAAB and benzene Chemical DAAB

Dose N (mg/kg) 0 50 100

5 5 4

Benzene

3. Results 3.1. Range finding study Oral doses of 50, 100, and 500 mg/kg DAAB were given to mice. All mice in the 500 mg/kg DAAB group died after a single treatment. In the 100 mg/kg DAAB group, one of five test animals died. Based on these

%PCE

0.40 ± 0.10 6.30 ± 0.20∗∗ 13.00 ± 1.14∗∗

68.0 55.4 56.8

P < 0.001a

2.4. Micronucleus data analysis The MN results were tabulated as the mean frequency of MN-PCE/1000 PCE per animal within a treatment group, plus or minus the standard error of the mean among animals. The frequencies of MN-PCE were analyzed by the Micronucleus Assay Data Management and Statistical Software Package [17] that tested for increasing trend over exposure groups using a one-tailed Cochran–Armitage trend test. The trend test was followed by pairwise comparisons between each exposure group and the control group using an unadjusted one-tailed Pearson chi-square test. In the presence of excess binomial variation, as detected by a binomial dispersion test, the binomial variance of the Cochran–Armitage test was adjusted upward in proportion to the excess variation. The %PCE data were analyzed by an analysis of variance (ANOVA) test based on individual animal data; pairwise comparisons between the exposure and control groups were made using a two-tailed Student’s t-test. For the MN test, an individual trial was considered positive if the trend test was ≤0.025, and the P-value for any single exposure group was ≤0.025/N where N is the number of exposure groups. For between group comparisons, ANOVA procedures were used to assess the significance of differences. The variance-stabilizing Freeman–Tukey Poisson transformation was used in these analyses [18]. For arriving at an overall conclusion, statistical as well as biological factors are considered.

MN-PCE/1000

0 1000 1500 2000 2250 2500

5 5 5 5 5 5

0.40 18.80 23.10 22.10 24.20 30.10

± ± ± ± ± ±

0.10 1.85∗∗ 1.76∗∗ 1.45∗∗ 2.28∗∗ 2.86∗∗

68.0 47.5 47.6 37.9 40.9 46.5

P < 0.001a Cyclophosphamideb

20

5

13.30 ± 0.72

62.7

Data are presented as the mean MN-PCE/1000 PCE ± S.E.M. a One-tailed Cochran–Armitage trend test. b Positive control. ∗∗ Pairwise comparison of treated group to corresponding vehicle (corn oil) control. A single control group was used for both test chemicals (P < 0.001).

results, the maximum tolerated dose (MTD) of DAAB was determined to be 100 mg/kg. No toxicity was observed at the lowest dose of 50 mg/kg DAAB. In contrast, benzene was tested up to a level of 2500 mg/kg with no lethality. Both DAAB and benzene induced highly significant increases in MN-PCE at all dose levels tested (Table 2). The response seen with DAAB was remarkable in that the dose levels that were tested were 10- and 20-fold lower than those of benzene, yet the MN-PCE response was similar in magnitude to that observed in benzene-treated animals and it equaled the response seen in the positive control group. 3.2. Definitive MN study Significant increases in MN-PCE were observed in the bone marrow of mice in all groups exposed to DAAB or benzene (Table 3). Significant increases in MN-PCE were also observed in two of the five aniline treatment groups, but these were not well correlated with dose. Based on the lack of dose response, and the small magnitude of the increases in MN frequency, the aniline response was judged to be a weak positive response. In mice exposed to the mixture, significant dose-related increases in MN-PCE were observed in

N.B. Ress et al. / Mutation Research 521 (2002) 201–208

205

Table 3 Definitive MN study Chemical

Group

Dose (mg/kg)

N

MN-PCE/1000a

1 2 3

0 25 50 100

5 5 5 5

0.70 2.10 5.00 9.00

DAAB

± ± ± ±

0.25 0.75∗ 1.75∗∗ 2.45∗∗

%PCE 57.0 61.2 60.4 63.4

P < 0.001b Benzene 1 2 3

0 10 20 40

5 5 5 5

0.70 2.40 2.50 6.30

± ± ± ±

0.25 0.40∗ 1.65∗∗ 2.10∗∗

57.0 60.6 55.2 63.9

P < 0.001b Aniline 1 2 3 4 5

0 12 23 47 120 470

5 5 5 5 5 5

0.70 1.20 2.60 1.40 2.10 3.30

± ± ± ± ± ±

0.25 0.90 0.95∗ 1.05 0.75 2.95∗∗

57.0 64.0 68.9 59.7 64.6 63.6

P = 0.001b Mixture (benzene/aniline) 1 2 3

0 10/12 20/23 40/47

5 5 5 5

0.70 2.00 3.00 3.80

± ± ± ±

0.25 1.60∗ 1.25∗∗ 1.90∗∗

57.0 59.2 62.4 65.5

P < 0.001b Cyclophosphamidec

20

5

17.9 ± 2.58

57.3

Comparative induction of MN by DAAB, benzene, aniline, and a mixture of benzene plus aniline using equimolar doses. Data are presented as the mean MN-PCE/1000 PCE ± S.E.M. a Significance of pairwise comparison of treated group to vehicle (corn oil) control. A single control group was used for all test chemicals. b One-tailed Cochran–Armitage trend test. c Positive control. ∗ P < 0.005. ∗∗ P ≤ 0.001.

all three dose groups. Based on the dose response, positive trend, and magnitude of response, the mixture was considered to be positive. No differences in the induction of MN-PCE were observed in mice exposed to 25 mg/kg DAAB and in mice exposed to doses of benzene, aniline or the 1:1 molar mixture equivalent to 25 mg DAAB (Group 1; Table 3; Fig. 2). In mice exposed to 50 mg/kg DAAB or its molar equivalents of benzene, aniline or the mixture (Group 2), the induction of MN-PCE was significantly greater (two-fold) in mice treated with DAAB than those treated with benzene or aniline. However, in the same group, DAAB was not different than the mixture (P = 0.08). In mice exposed to 100 mg/kg

DAAB or its molar equivalents of benzene, aniline or the mixture (Group 3), induction of MN-PCE was significantly greater in DAAB-treated animals than those exposed to the mixture, benzene, or aniline. In addition, induction of MN-PCE in benzene treated animals in Group 3 was significantly greater than mice exposed to aniline or the mixture.

4. Discussion Previous NTP sponsored studies have shown that DAAB is metabolized almost completely to benzene, aniline, and their metabolites in (Fig. 1; [5]).

206

N.B. Ress et al. / Mutation Research 521 (2002) 201–208

Fig. 2. Definitive MN study. Comparative induction of MN in the bone marrow of B6C3F1 mice exposed to DAAB, benzene, aniline or a mixture of benzene and aniline. Data are presented as the mean MN-PCE/1000 PCE ± S.D.—“a”: significantly higher than aniline exposed mice (P < 0.05; variance-stabilizing Freeman–Tukey Poisson transformation); “b”: significantly higher than benzene-exposed mice (P < 0.05; variance-stabilizing Freeman–Tukey Poisson transformation); “m”: significantly higher than mixture exposed mice (P < 0.05; variance-stabilizing Freeman–Tukey Poisson transformation). Group 1: DAAB (25 mg/kg), benzene (10 mg/kg), aniline (12 mg/kg), mixture (benzene + aniline (10 + 12)); Group 2: DAAB (50 mg/kg), benzene (20 mg/kg), aniline (23 mg/kg), mixture (benzene + aniline (20 + 23)); Group 3: DAAB (100 mg/kg), benzene (40 mg/kg), aniline (47 mg/kg), mixture (benzene + aniline (40 + 47)).

In a 16-day dermal toxicity study of DAAB in rats and mice, responses characteristic of aniline toxicity (splenotoxicity, methemoglobinemia, and increased Heinz body formation) were observed [9]. While similarities were also seen between DAAB effects and benzene toxicity, the toxicological responses observed in rats and mice only showed concordance in site of action [9]. The NTP studies were used as the basis for a prediction that DAAB would be carcinogenic if tested in a 2-year bioassay based on its metabolism to the known carcinogens benzene and aniline, and their metabolites. The current study was performed to provide additional supporting evidence of similarities in response between DAAB and benzene for a toxicological endpoint that is predictive for carcinogenesis, i.e. MN formation. It was anticipated that this information would further support the prediction of its carcinogenicity based on an additional short-term endpoint. DAAB, benzene, and the mixture gave positive responses in the mouse bone marrow MN test (Table 3). This is the first reported evidence of DAAB-induced chromosomal damage. However, benzene has been extensively studied for genotoxicity in humans and animals [19]. In humans, strong evidence for the

clastogenicity of benzene comes from observations of significant increases in numerical and structural chromosomal damage in lymphocytes of benzene-exposed workers [20–23,19]. Animal studies also provide convincing evidence of benzene genotoxicity and support human case reports and epidemiological studies in which chromosomal damage was linked with benzene exposure. Benzene consistently has been shown to increase the frequency of MN and chromosome aberrations in mouse bone marrow and peripheral blood [24–28]. Choy et al. [10] also showed that male mice were more sensitive to the genotoxicity of benzene than other sex/species. In the present study, aniline was negative in the preliminary investigation (data not shown). However, studies in the literature have shown that aniline is a weak inducer of MN at doses approaching the MTD [11,29]. Therefore, in the definitive trial the doses for aniline were extended upward to what was thought to approach an MTD in order to determine if, in our system, aniline-induced MN at higher doses than those used for comparison to DAAB. Even at higher doses, aniline gave only a weak positive response (Table 3). One of the major toxic events resulting from aniline exposure is methemoglobinemia [30,8,7]. In a 16-day

N.B. Ress et al. / Mutation Research 521 (2002) 201–208

dermal toxicity study of DAAB, a treatment-related decrease in the erythroid mass evidenced by a decrease in hematocrit, hemoglobin, and erythrocyte counts suggested a developing anemia. The erythron decrease was accompanied by an increased bone marrow response as indicated by increased reticulocytes in rats and mice and nucleated erythrocytes in rats. This information coupled with the very strong MN response to DAAB in the range finding study suggested that an increased production of reticulocytes resulting from an aniline-induced anemia might increase the probability for MN induction from the benzene also present. However, consistent changes in %PCE were not noted in the bone marrow of treated animals. This may be because the 48-h exposure duration prior to analysis did not allow sufficient time for development of a bone marrow response. Thus, it appeared unlikely that increased erythropoiesis contributed significantly to MN induction from DAAB. The molecular structure of DAAB contains a benzene ring, an aniline moiety, and nitrogen, with their relative percentages being 40, 47, and 13%, respectively. The doses selected for the definitive MN study were chosen based on the molar contribution of each of these components to DAAB. The doses selected for the definitive MN study were chosen based on the molar contribution of each component to DAAB. In addition, a mixture of benzene and aniline was tested to determine if both components, when given simultaneously, acted similarly to DAAB. DAAB was equal in potency to all compounds tested in Group 1. However, when the dose was increased (Groups 2 and 3) DAAB was significantly greater in potency compared to benzene or aniline. In Group 3, DAAB also was significantly greater than the mixture, and while statistical significance (P = 0.08) was not achieved in Group 2, when comparing the mixture and DAAB, DAAB appeared to be more potent. These observations were duplicated in the preliminary investigation (data not presented). In Group 3, MN induction in benzene-exposed animals was significantly greater than that observed in animals exposed to the mixture or to aniline. These results demonstrate that benzene and its metabolites play a major role in the induction of MN in DAAB-treated animals. Whether aniline acts to potentiate the induction of MN by benzene is unknown. However, our results suggest that, at doses representative of the contribution of aniline to DAAB, aniline or its

207

metabolites alone do not play an important role in MN induction. It also has been shown that metabolism of DAAB to benzene, aniline, and their metabolites results in the production of a phenyl radical (Fig. 1; [4]). The increased frequency of MN observed in mice treated with DAAB compared to benzene or the mixture could be a result of phenyl radical production, thus, increasing the genotoxicity of DAAB. However, possible production of the phenyl radical from DAAB in bone marrow has not been investigated. In summary, these studies show that, like benzene, DAAB is a potent inducer of MN in mouse bone marrow. At higher doses DAAB was more potent than equimolar doses of benzene, aniline, or a mixture of these compounds. These results provide evidence of the genotoxicity of DAAB and an association with a characteristic benzene bone marrow response. The findings support the predicted carcinogenicity of DAAB [9].

References [1] R.J. Lewis Sr., Hawley’s Condensed Chemical Dictionary, 12th ed., Van Nostrand Reinhold Co., NY, 1993, p. 366. [2] J.E. Bailey Jr., Determination of 1,3-diphenyltriazene and azobenzene in D&C Red No. 33 by solvent extraction and reversed-phase high-performance liquid chromatography, J. Chromatogr. 321 (1985) 185–197. [3] S. Palmer, R.A. Mathews, The role of non-nutritive dietary constituents in carcinogenesis, Surg. Clin. North Am. 66 (1986) 891–915. [4] M.B. Kadiiska, K.S. De Costa, R.P. Mason, J.M. Mathews, Reduction of 1,3-diphenyl-1-triazene by rat hepatic microsomes, by cecal microflora, and in rats generates the phenyl radical metabolite: an ESR spin-trapping investigation, Chem. Res. Toxicol. 13 (2000) 1082–1086. [5] J.M. Mathews, K.S. De Costa, Absorption, metabolism, and disposition of 1,3-diphenyl-1-triazene in rats and mice after oral, i.v., and dermal administration, Drug Metab. Dispos. 27 (1999) 1499–1504. [6] National Toxicology Program (NTP), Toxicology and Carcinogenesis Studies of Benzene (CAS# 71-43-2) in F344/N Rats and B6C3F1 Mice (Gavage Studies), Technical Report Series No. 289, NIH Publication No. 86-2545, US Department of Health and Human Services, Public Health Service, National Institutes of Health, Research Triangle Park, NC, 1986. [7] M.F. Kahn, X. Wu, B.S. Kaphalia, P.J. Boor, G.A.S. Ansari, Acute hematopoietic toxicity of aniline in rats, Toxicol. Lett. 92 (1997) 31–37.

208

N.B. Ress et al. / Mutation Research 521 (2002) 201–208

[8] C.R. Short, C. King, P.W. Sistrunk, K.M. Kerr, Subacute toxicity of several ring-substituted dialkylanilines in the rat, Fundam. Appl. Toxicol. 3 (1983) 285–292. [9] National Toxicology Program (NTP), Report on the Metabolism, Toxicity, and predicted Carcinogenicity of Diazoaminobenzene (CAS# 136-35-6), Toxicity Report Series Number 73, NIH Publication Number 01-4412, US Department of Health and Human Services, Public Health Service, National Institutes of Health, Research Triangle Park, NC 27709, 2002. [10] W.N. Choy, J.T. MacGregor, M.D. Shelby, R.R. Maronpot, Induction of micronuclei by benzene in B6C3F1 mice: retrospective analysis of peripheral blood smears from the NTP carcinogenesis bioassay, Mutat. Res. 143 (1985) 55–59. [11] J.T. MacGregor, C.M. Wehr, P.R. Henika, M.D. Shelby, The in vivo erythrocyte micronucleus test: measurement at steady state increases assay efficiency and permits integration with toxicity studies, Fundam. Appl. Toxicol. 14 (1990) 513–522. [12] J. Ashby, D.A. Vlachos, H. Tinwell, Activity of aniline in the mouse bone marrow micronucleus assay, Mutat. Res. 263 (1991) 115–117. [13] C. Westmoreland, D.G. Gatehouse, Effects of aniline hydrochloride in the mouse bone marrow micronucleus test after oral administration, Carcinogenesis 12 (1991) 1057–1059. [14] K.L. Witt, A. Knapton, C.M. Wehr, G.J. Hook, J. Mirsalis, M.D. Shelby, J.T. MacGregor, Micronucleated erythrocyte frequency in peripheral blood of B6C3F1 mice from short-term, prechronic, and chronic studies of the NTP carcinogenesis bioassay program, Environ. Mol. Mutagen. 36 (2000) 163–194. [15] M.D. Shelby, G.L. Erexson, G.J. Hook, R.R. Tice, Evaluation of a three-exposure mouse bone marrow micronucleus protocol: results with 49 chemicals, Environ. Mol. Mutagen. 21 (1993) 160–179. [16] M. Hayashi, T. Sofuni, M. Ishidate Jr., An application of acridine orange fluorescent staining to the micronucleus test, Mutat. Res. 120 (1983) 241–247. [17] ILS, Micronucleus Data Management and Statistical Analysis Software, Version 1.4a, Integrated Laboratory Systems, P.O. Box 13501, Research Triangle Park, NC 27707, 1990. [18] N.L. Johnson, S. Kotz, Discrete Distributions, Houghton Mufflin Company, Boston, 1969, p. 99. [19] Agency for Toxic Substances and Disease Registry (ATSDR), Toxicological Profile for Benzene, US Department of Health and Human Services, Public Health Service, National Institutes of Health, Research Triangle Park, NC, 1997.

[20] T.M. Smith, L. Zhang, L.-Q. Xi, The mechanism of benzeneinduced leukemia: a hypothesis and speculations on the causes of leukemia, Environ. Health Perspect. 104 (1996) 1219– 1225. [21] F. Marcon, A. Zijno, R. Crebelli, A. Carere, T. Veidebaum, K. Peltonen, R. Parks, M. Schuler, D. Eastmond, Chromosome damage and aneuploidy detected by interphase multicolour FISH in benzene-exposed shale oil workers, Mutat. Res. 445 (1999) 155–166. [22] L. Zhang, N. Rothman, Y. Wang, R.B. Hayes, S. Yin, N. Titenko-Holland, M. Dosemeci, Y.-Z. Wang, P. Kolachana, W. Lu, L. Xi, G.-L. Li, M.T. Smith, Benzene increases aneuploidy in the lymphocytes of exposed workers: a comparison of data obtained by fluorescence in situ hybridization in interphase and metaphase cells, Environ. Mol. Mutagen. 34 (1999) 260– 268. [23] V. Kasuba, R. Rozgaj, K. Sentija, Cytogenetic changes in subjects occupationally exposed to benzene, Chemosphere 40 (2000) 307–310. [24] M.D. Shelby, K.L. Witt, Comparison of results from mouse bone marrow chromosome aberration and micronucleus test, Environ. Mol. Mutagen. 25 (1995) 203–313. [25] R.R. Tice, D.L. Costa, R.T. Drew, Cytogenetic effects of inhaled benzene in murine bone marrow: induction of sister chromatid exchanges, chromosomal aberrations, and cellular proliferation inhibition in DBA/2 mice, Proc. Natl. Acad. Sci. U.S.A. 77 (1980) 2148–2152. [26] R. Ciranni, R. Barale, A. Marrazzini, N. Loprieno, Benzene and the genotoxicity of its metabolites. I. Transplacental activity in mouse fetuses and in their dams, Mutat. Res. 208 (1988) 61–67. [27] C.A. Luke, R.R. Tice, R.T. Drew, The effect of exposure regiment and duration on benzene-induced bone-marrow damage in mice. I. Sex comparison in DBA/2 mice, Mutat. Res. 203 (1988) 251–271. [28] S. Suzuki, H. Atai, Y. Hatakeyama, M. Hara, S. Nakagawa, Administration-route-related differences in the micronucleus test with benzene, Mutat. Res. 223 (1989) 407–410. [29] E. George, M. Andres, C. Westmoreland, Effects of azobenzene and aniline in the rodent bone marrow micronucleus test, Carcinogenesis 11 (1990) 1551–1556. [30] Chemical Industry Institute of Toxicology (CIIT), 104-Week Chronic Toxicity Study in Rats: Aniline Hydrochloride, Final Report, Research Triangle Park, NC, 1982.