Mutation Research, 280 (1992) 17-27 0 1992 Elsevier Science Publishers B.V. All rights reserved

MUTGEN

17 0165-1218/92/$05.00

01781

Evaluation of genotoxicity of tert.-butylhydroquinone in an hepatocyte-mediated assay with V79 Chinese hamster lung cells and in strain D7 of Saccharomyces cerevisiae C.G. Rogers a, B.G. Boyes a, T.I. Matula b and R. Stapley ’ ’ Toxicology Research Dioision, Food Directorate, h Drug Toxicology Diuision, Drug Directorate and ’ Biostatisrics and Computer Applications Diuision, Food Directorate, Health and Welfare Canada, Ottawa KIA OL2 (Canada) (Received 30 July 1991) (Revision received 8 January 1992) (Accepted 14 January 1992)

Keywords: Sister-chromatid exchange; tert.-Butylhydroquinone

Thioguanine-resistance

mutation;

Gene

conversion;

Reverse

mutation;

Summary tert.-Butylhydroquinone (TBHQ) has been reported to be genotoxic in some short-term assays but non-genotoxic in others. We have examined cytotoxicity and genotoxicity of TBHQ, a principal metabolite of the phenolic antioxidant 2(3)-tert.-butyl-4-hydroxyanisole (BHA), in an hepatocyte-mediated assay with V79 Chinese hamster lung cells including both sister-chromatid exchange (SCE) and thioguanine-resistance (TGR) endpoints. The ability of BHA and of TBHQ to elicit a genotoxic response in Saccharomyces cerevisiae strain D7 was also investigated. In V79 cytotoxicity tests, TBHQ without hepatocytes produced a 50% reduction in colony formation at 4.2 pg/ml and was lethal to 100% of the cells at concentrations above 5 pg/ml. At partially cytotoxic dose levels, (0.17-3.4 pgg/ml of medium), TBHQ sometimes increased significantly the frequency of SCE. TBHQ also produced sporadic statistically significant increases in the mutation frequency at the HGPRTase (TGR) gene locus when tested alone or with activation by rat or hamster hepatocytes. Mitotic gene conversion and reverse. mutation were not induced in strain D7 of Saccharomyces cereuisiae by exposure to BHA or to TBHQ for 4 h at concentrations as high as 200 pg/ml for BHA or 500 pg/ml for TBHQ, either alone or with activation by rat-liver S9. Incubation of the yeast cells with BHA or TBHQ for 24 h in growth medium without activation also did not induce genotoxic activity. The slight and sporadic response to TBHQ in the V79 test system may indicate weak genotoxicity which is sensitive to slight differences in test conditions. The classification and test strategies adopted for compounds such as TBHQ could have important implications for regulatory decisions and for the validation of short-term tests.

Correspondence: Dr. C.G. Rogers, Toxicology Research Division, Banting Research Centre, Health and Welfare Canada, Ross Avenue, Ottawa, Ont KlA 0L2 (Canada).

While many chemicals are unequivocally genotoxic and others are negative in every system in which they are tested, some chemicals are diffi-

OH tert-butyl wi&c

hydroqulnone (C6H,b~.4-(w,

F.W. 166.22

I I 1

2

3

4

5

6

TBHP (pg/ml)

Fig. 1. Percent colony formation (percent survival) of V79 cells treated with ret?.-butylhydroquinone (TBHQ) (insert) in the absence of hepatocytes.

cult to classify as either genotoxic or non-genotoxic. These may be reported in the literature as “equivocal”, “indeterminate”, “having weak evidence for genotoxicity”, or as producing conflicting results in different test systems. In assessing the performance of short-term tests with different genetic endpoints as predictors of carcinogenicity, such chemicals may be arbitrarily classified as genotoxic or non-genotoxic, with consequent influence on regulatory decisions and on the evaluation of short-term tests. One of the chemicals producing inconsistent results in different assay systems is tert.-butylhydroquinone (TBHQ) (Fig. 11, a principal metabolite of 2(3)-tert.-butyl-4-hydroxyanisole (BHA) (El-Rashidy and Niazi, 1983; Armstrong and Wattenberg, 1985). BHA is a phenolic antioxidant and is used extensively as a food additive. Recently, it was reported by Nouaim and Dorange (1988) that BHA was mutagenic to yeast

cells. They attributed their findings to the possible formation of one or more quinone metabolites of BHA. Although TBHQ is non-mutagenic in the Salmonella/microsome assay (Abe and Sasaki, 1977; Hageman et al., 1988; Matsuoka et al., 1990), TBHQ can produce chromosomal aberrations in bone-marrow cells of mice (Giri et al., 1984) and in the Chinese hamster fibroblast cell line CHL in the presence of rat liver S9 (Matsuoka et al., 1990). Phillips et al. (1989) have also reported that TBHQ (which can autoxidise in solution to produce hydrogen peroxide) and an end-product of TBHQ oxidation, tert.-butylquinone (TBQ) were clastogenic to Chinese hamster ovary (CHO) cells. Unpublished findings also suggest that TBHQ may be mutagenic to V79 Chinese hamster lung cells and positive in the L5178Y mouse lymphoma test (reviewed by van Esch, 1986). Although TBHQ is permitted as an antioxidant in foods in the U.S.A. and in 9 other countries but is not permitted in Canada, conflicting results have emphasized the need for additional evaluation of its genotoxicity (van Esch, 1986). In the present in vitro study, we have further evaluated genotoxicity [thioguanine resistance (TGR) and sister-chromatid exchange (SCE)] of TBHQ in an hepatocyte-mediated assay with V79 Chinese hamster lung cells. The possibility that BHA, or one of its quinone metabolites, may be genotoxic to yeast cells (Nouaim and Dorange, 1988) lead us also to re-examine the ability of BHA and of TBHQ to induce mitotic gene conversion and reverse mutation in the diploid strain D7 of Saccharomyces cereuisiae. We tested TBHQ in the V79 system, with and without activation, 6 different times. In the studies to be described here, individual experiments for SCE and TGR produced sporadic positive results (p < 0.01) when TBHQ treatment levels were compared with concurrent controls. The presence of significant experiment-treatment interaction precluded an assessment based on data pooled across experiments. Materials

and methods

Medium

Williams’ medium E (WE) (Flow Laboratories, ICN Biomedicals, Mississauga, Ont.), supple-

19

mented with 10% fetal bovine serum (Bocknek Laboratories, Toronto, Ont.), r_-glutamine (2.0 mM) (Flow Laboratories McLean, VA) and gentamicin (10 pgg/ml) (Schering Corp., Kenilworth, NJ) was used for growth and maintenance of cells and for suspension of hepatocytes. Cells

The V79 cells used in this study were kindly donated by Dr. E. Elmore, Northrop Services, Research Triangle Park, NC. At intervals, after storage in liquid nitrogen, cells were pre-grown in antibiotic-free medium and found to be free of mycoplasma as determined by the Hoechst fluorochrome stain technique (Chen, 1977). Cultures of S. cereuisiue, diploid strain D7, were grown in complete medium (YPD) containing 1% yeast extract, 2% peptone and 2% glucose (Zimmermann, 1975; Zimmermann et al., 1975; Matula and Downie, 1984). Chemicals

tert.-Butylhydroquinone (TBHQ) (reagent grade, 99%) was from Aldrich Chemical Co., Milwaukee, WI; ethyl methanesulfonate (EMS) was from Eastman Kodak Co., Rochester, NY; butylated hydroxyanisole (BHA) [containing approximately 95% 3-tert.-butyl-4-hydroxyanisole and 5% 2-tert.-butyl-6hydroxyanisolel; 6-thioguanine (TG); 7,12-dimethylbenz[a]anthracene (DMBA) and dimethyl sulfoxide(DMS0) were from Sigma Chemical Co., St. Louis, MO; colcemid and collagenase were from Gibco Canada Ltd., Burlington Ont.; 5-bromo-2’-deoxyuridine (BrdUrd) was from Boehringer Mannheim Canada, Ltd., Montreal P.Q.; the fluorochrome stain 33258 was from Hoechst Canada, Ltd., Montreal, P.Q. TBHQ, BHA and DMBA were each dissolved in DMSO. TBHQ and DMBA were further diluted in WE medium so that the final DMSO concentration did not exceed 0.2% v/v. BHA was further diluted in YPD medium for genotoxicity testing in the yeast assay with S. cereuisiae. TG was dissolved in 1 N potassium hydroxide, diluted in WE medium (with pH adjusted to 7.0-7.2) and sterilized by membrane filtration (0.2 ~1). EMS was dissolved and diluted directly in WE medium.

Cytotoxicity

The cytotoxicity of TBHQ was determined in the absence of hepatocytes by measurement of the effect on plating efficiency as described by Rogers and Heroux-Metcalf (1983). TBHQ was tested at dose levels ranging from 0 to 8.3 pg/ml of medium. Liter perfusion Rat hepatocytes. Primary hepatocytes were prepared immediately before use by in vivo collagenase perfusion of the liver of a 250-300 g male Fischer 344 rat as described previously (Williams et al., 1977; Rogers and H&roux-Metcalf, 1983; Rogers et al., 1985). Hamster hepatocytes. The liver of a 100-200 g male golden Syrian hamster was perfused in situ with collagenase via the portal vein by the method of Maslansky and Williams (1982) with minor modifications as described in Rogers et al. (1985). Mutagenicity

Mutagenicity was evaluated in a hepatocytemediated mutation assay according to protocols of Langenbach et al. (1978, 19811 with some modifications (Rogers et al., 1985). The period of exposure to chemicals, with or without activation by rat or hamster hepatocytes, was 48 h, after which the chemicals were removed and the cultures reseeded into flasks (75 cm*) at a density of 1.9 x 10” viable cells per 15 ml of Williams’ E medium. Following a 6-day period of expression, the cells were reseeded into 20 culture dishes (Corning, 60 mm) at 20 x lo3 viable cells per dish. For the selection of mutants, 6-thioguanine was added immediately after seeding to give a final concentration per dish of 10 pg/ml. Mutant colonies were counted 8 days later. Sister-chromatid exchange V79 cells were incubated

24 h in the presence or absence of the test chemical with and without rat or hamster hepatocytes, and then analyzed for SCE (Perry and Wolff, 1974; Latt et al., 1981) as described in detail elsewhere (Rogers et al., 1988; Boyes et al., 1991).

20

Yeast assay

Mitotic gene conversion in S. cerevisiae strain D7 was estimated at the trp5 locus and reverse mutation at the ilvZ-92 locus as described elsewhere (Zimmermann, 1975; Zimmermann et al., 1975; Matula and Downie, 1984) following treatment with BHA or TBHQ. Three protocols were used: (i) 4 h of exposure in citrate buffer, pH 3.6 as in Nouaim and Dorange (1988); (ii) 4 h of exposure with metabolic activation by rat-liver S9 at pH 7.4 as in Vleminckx et al. (1987); and (iii> 24 h of exposure in the growth medium without S9 as in Matula and Downie (1984).

Statistical methodology

A preliminary test for experiment-treatment interaction was conducted for the SCE and TGR data. Treatments without hepatocyte activation, treatments with rat hepatocyte activation and treatments with hamster hepatocyte activation were analyzed separately. The results indicated significant experiment-treatment interaction (p < 0.01) for each group of treatments indicated above. The presence of significant experimenttreatment interaction precluded an assessment based on pooling of the data across experiments, and as such, separate analyses were conducted for each experiment. Within each experiment with V79 cells, TGR and SCE data were analyzed separately. The method of analysis is described in Boyes et al. (1991). For thioguanine resistance data, the average number of mutant colonies/dish, based on 20 replicated dishes per treatment group, was analyzed. The assumption was made that the average number of mutant colonies/dish for treatment or control groups had a Poisson distribution. Plots of average number of mutant colonies/dish versus the variance of the number of mutant colonies/dish revealed this assumption to be reasonable. For TGR data, the average number of mutant colonies per dish that exceeded the 95% or 99% upper confidence limits (u.c.1.‘~) for the concurrent controls were identified. Analyses of SCE data were performed on the natural log scale as previous experience has shown this to be the most appropriate scale (Boyes et al., 1990); and

values which exceeded the 95% or 99% u.c.l.‘s for the concurrent controls were identified. Results In V79 cells, the relationship between percent colony formation (an indication of plating efficiency) and TBHQ concentration is shown in Fig. 1. TBHQ was tested in the absence of hepatocytes at concentrations ranging from 0 to 8.3 pg/ml. The chemical showed slight but increasing cytotoxicity to V79 cells at concentrations ranging from 2.0 to 3.4 pg/ml. Colony formation was reduced to 50% of the control at about 4.2 pg/ml and the antioxidant was lethal to 100% of the cells at dose levels above 5 pg/ml. Geometric means of SCE per metaphase and results of treatment-control comparisons are shown in Fig. 2. The frequency of SCE was increased by TBHQ without activation at 1.70 pg/ml (only in Expts. 1 and 3) at 0.17 pg/ml with rat hepatocyte activation (only in Expt. 3) and at 0.17 kg/ml with hamster hepatocyte activation (only in Expt. 6). The average number of thioguanine-resistant mutant colonies per dish and results of treatment-control comparisons are provided in Fig. 3. Without hepatocytes, TBHQ was non-mutagenic at the HGPRT (TGR) gene locus at all concentrations in each of Expts. 1 through 5, but was positive at 1.70 and 3.40 pg/ml in Expt. 6. When activated by rat hepatocytes, 0.17 pg/ml TBHQ was positive in Expt. 1 and negative in Expts. 2 and 3; at 1.7 pg/ml, TBHQ was positive in Expts. 1 and 2, but was negative in Expt. 3; whereas at 3.4 pg/ml, TBHQ was negative in all 3 experiments. With activation by hamster hepatocytes (Fig. 31, TBHQ was negative except at 0.17 pg/ml in Expt. 5. In the latter Expt., the control value with hamster hepatocytes was unusually low and mutagenic activity with 0.17 ug/ml TBHQ was higher than in Expt. 4 or Expt. 6. These results could be interpreted as equivocal. The cloning efficiencies for the various treatment groups when tested alone and with activation by rat or hamster hepatocytes are shown for each experiment in Table 1. Using this information, the number of mutants per lo6 surviving cells could be calculated for each treatment group

21

5: 26

EXPERIMENT

$

IFi

3

f 22

EXPERIMENT

5

EXPERIMENT

6

CONTROL EMS DMBA TEHO

O.l7fig/ml 1.70pgiml

WITH RAT HEPATOCYTES

HEPATOCYTES

Fig. 2. Geometric

TABLE

means

of SCE per metaphase

WITHOUT HEPATOCYTES

*

WITH HAMSTER HEPATOCYTES

in V79 cells treated with TBHQ with and without hepatocytes. * p < 0.05; * * p < 0.01.

activation

by rat or hamster

1

CLONING EFFICIENCIES FOR VARIOUS TREATMENT HAMSTER HEPATOCYTES Treatment

Cloning

efficiency

GROUPS ALONE AND WITH ACTIVATION

BY RAT OR

a (%o)

Rat

Hamster

Expt. 1

Expt. 2

Expt. 3

Expt. 4

Expt. 5

Expt. 6

(1.70 @g/ml) (3.40 *g/ml)

92 79 86 81 100 92

112 112 93 119 52 75

107 90 101 99 75 86

111 114 128 102 69 77

102 74 102 106 85 88

67 68 77 101 128 119

With hepatocytes Control DMBA (2.6 pg/ml) TBHQ (0.17 @g/ml) (1.70 *g/ml) (3.40 pg/ml)

81 64 94 95 103

101 105 92 73 60

98 105 93 75 72

108 98 112 62 55

123 93 103 83 67

80 81 141 99 122

Withouf hepatocytes Control EMS (155 wg/ml) DMBA (2.6 pg/ml) TBHQ (0.17 pg/ml)

a Cloning

efficiency

= (number

of colonies/number

of cells seeded) X 100.

22

external activation (Table 4). Frequencies of gene conversion or reverse mutation were not significantly increased at dose levels as high as 200 pg/ml of medium (BHA) or 500 pg/ml of medium (TBHQ), with any of the treatment protocols (Tables 2, 3 and 4). Although in one experiment (Table 4), BHA at 100 pg/ml increased the apparent number of reverse mutations 3-fold over that of the untreated control, this result was not considered significant since the number of surviving cells was reduced to 3%.

in each experiment by dividing the average number of mutant colonies per dish by the corresponding cloning efficiency value in Table 1. However, it did not appear feasible to attach parametric or non-parametric statistical significance to mutants per lo6 surviving cells since this was the ratio of a Poisson distributed variable (average number of mutant colonies per dish) and an approximately normally distributed random variable (cloning efficiency). For this reason, the data for mutation frequencies were not reported as mutants per lo6 surviving cells. In order to duplicate as closely as possible the assay conditions of Nouaim and Dorange (1988), gene conversion and reverse mutation were measured in the diploid strain D7 of S. cerevisiae after 4 h of exposure to BHA or TBHQ without activation at pH 3.6 (Table 2). Genotoxicity was also determined after 4 h of exposure to BHA or TBHQ with metabolic activation by rat-liver S9 at pH 7.4 (Table 3); and after 24 h of exposure to BHA or to TBHQ in the growth medium without TABLE

Discussion

Evaluation of the performance of short-term genotoxicity tests as predictors of human carcinogenicity has occupied genetic toxicologists for several years. Species and dosing differences as well as the suitablility of various short-term tests for particular classes of chemicals and for indicating different steps in the process of carcinogenesis make this task particularly difficult, especially

2

GENE CONVERSION AND REVERSE MUTATION IN DIPLOID EXPOSURE TO BHA OR TBHQ WITHOUT ACTIVATION (pH 3.6) Treatment

Survivors

(pg/mll

(%)

Untreated control 0

Gene conversion (trp+/106 survivors) a

STRAIN

Ratio

D7

OF

Reverse mutation (ilv+/lO’ survivors)

(T/c)’

S. cereuisiae:

EFFECT

Ratio (T/c)’ b

7.8(272) d

_

0.3(22)

Butylated hydroxyanisole (BHA) 50 84 100 79 150 60 200 46

7.8(206) 6.9(177) 8.4(1621 8.2(121)

1.0 0.9 1.1 1.1

OS(291 OS(24) 0.7(25) 0.7(211

1.7 1.7 2.3 2.3

tert.-Butylhydroquinone (TBHQ) 100 88 200 72 400 62 500 52

8.6(250) 7.8(1841 8.4072) 11.9(2031

1.1 1.0 1.1 1.5

0.5(30) 0.407) 0.3(121 0.7(251

1.7 1.3 1.0 2.3

Ethyl methanesulfonate (EMS) e 1000 66

64.1(1385)

8.2

29.4(1270)

a b ’ d e

100

trp+= gene conversion at the trp.5 (ttyptophanl gene locus. ilv+ = reverse mutation at the i/u-92 (isoleucine) gene locus. Ratio (treatment/untreated control). Numbers in parentheses denote total numbers of colonies on triplicate Direct-acting control.

plates.

98.0

OF

4 h

23

when some compounds behave in an inconsistent or contradictory fashion in various test systems that display different genetic endpoints. We have assessed intensively one such compound, TBHQ, in an attempt to elucidate its behaviour in V79 cells and in yeast. As shown here, the cytotoxicity of TBHQ to V79 cells, in the absence of hepatocytes, was about 15fold higher than that reported previously for BHA (Rogers et al., 1985). Unlike studies with nitrosamines (Poiley et al., 1980; Phillipson and Ioannides, 19841, the present results do not suggest that hamster hepatocytes were more effective than rat hepatocytes in the activation of TBHQ to products(s) genotoxic to V79 cells. However, as in previous work (Rogers et al., 19851, the induction of SCE by DMBA (Fig. 2) and the increase in TGR mutation frequency with DMBA (Fig. 3) were notably greater with activation by hepatocytes from the rat than TABLE

from the hamster. It is generally recognized (Bradley et al., 1981; Madle et al., 1986) that enzyme patterns in tissue homogenates do not necessarily represent those in vivo, and that the use of hepatocytes may more closely represent conditions of metabolism in intact cells. For these reasons we did not use S9 with V79 cells. The inability of BHA or of TBHQ to induce mitotic gene conversion or reverse mutation in strain D7 of S. cerevisiue under any of the test conditions employed in this study is in contrast to the positive results reported for BHA by Nouaim and Dorange (1988), and does not support their suggestion that a quinone metabolite may be responsible for genotoxic activity of BHA in the yeast cell assay. In our studies with V79 cells, at somewhat cytotoxic dose levels (as high as 3.4 pg/ml), TBHQ did not increase consistently the frequency of SCE. Mutation to thioguanine resis-

3

GENE CONVERSION AND REVERSE MUTATION IN DIPLOID STRAIN D7 OF S. cerevisioe: SURE TO BHA OR TBHQ WITH ACTIVATION BY RAT LIVER S9 (pH 7.4) Treatment

Survivors

(pg/ml)

(%o)

Untreated control 0

100

Butylated hydroxyanisole (BHA) 50 95 100 150

80 77

tert. -Butylhydroquinone (TBHQ) 111 50 100 150 200

105 105 105

Ethyl methanesulfonate (EMS) e 88 1000

Gene conversion (trp+/106 survivors) a

Reverse mutation (ilv+/lOh survivors)

Ratio

(T/c)’

EFFECT

OF 4-h EXPO-

Ratio (T/c)’ b

_

1.0(44)

_

6.8(111)

0.9

1.1(18)

7.0(96) 7.3(94)

0.9 1.0

1.1(14) 1.6(21)

1.1 1.1 1.6

6.8(169) 7.2(169) 6.3(150) 6.2(144)

0.9 0.9 0.8 0.8

0.8(42) 0.8(38) 0.9(44) 0.8(35)

0.8 0.8 0.9 0.8

39.6(594)

5.2

53.7(804)

53.7

18.8(442)

2.5

10.2(477)

10.2

7.6071)

d

Dimethylnitrosamine (DMN) f 22 000 a b ’ d ’ f

104

trp+= gene conversion at the trp5 (tryptophan) gene locus. ilv+ = reverse mutation at the ilu-92 (isoleucine) gene locus. Ratio (treatment/untreated control). Numbers in parentheses denote total numbers of colonies on triplicate Direct-acting control. Indirect-acting control.

plates.

24

tance by TBHQ was also inconsistent, with sporadic positive responses appearing in all three activation groups (alone, with rat hepatocytes or with hamster hepatocytes) despite test conditions that were controlled as closely as possible. While these positive results were statistically significant, the magnitude of the response in each case was small. The small fluctuations in TGR response that occurred with the controls appeared to be unrelated to those obtained with test chemicals. Other factors that may have contributed to fluctuations in the TGR mutation frequencies are the capability of the V79 cells to produce error-free repair; and the inherent variability of the hepatocyte metabolic activation system in combination with the V79 target cells (Jensen, 1984). The fluctuations seen here in TGR response cannot

TABLE

be explained on the basis of dilution errors, since sporadic positive results for SCE appeared independently of those for TGR. Thus, although TBHQ is a confirmed clastogen in vivo as well as in vitro (Giri et al., 1984; Matsuoka et al., 1990; Phillips et al., 19891, it showed inconsistent activity in our V79 test system for TGR and SCE and was negative in strain D7 of S. cereuisiae. Our results are, however, consistent with findings of non-mutagenicity of TBHQ in the Salmonella assay (Abe and Sasaki, 1977; Hageman et al., 1988; Matsuoka et al., 1990). Ashby et al. (1985) have emphasized the need to determine what constitutes a positive or a negative response in in vitro assays. In our V79 system, the appropriateness of the use of concurrent controls as well as empirically calculated

4

GENE CONVERSION AND REVERSE MUTATION IN DIPLOID EXPOSURE TO BHA OR TBHQ IN GROWTH MEDIUM Treatment

Survivors

(kg/ml)

(%)

Untreated control 0

100

Butylated hydroxyanisole (BHA) 1 86 5 84 10 89 25 86 50 59 100 3 tert-Butylhydroquinone 1 5 10 25 50 100 150 Ethyl methanesulfonate 1000

(TBHQ) 81 72 69 88 110 80 74 (EMS) e 26

Gene conversion (trp+/lO’ survivors) ’

STRAIN

Reverse mutation (ilv+/106 survivors)

Ratio (T/c)’

EFFECT

Ratio (T/c)’ ’

3.5(170) d

_

0.3(28)

_

3.7(157) 2.80 15) 3.2(134) 3.0(124) 2.1(59) 2.8(4)

1.1 0.8 0.9 0.9 0.6 0.8

0.4(33) 0.3(25) 0.4(40) 0.2(21) 0.2(9) 0.9(3)

1.3 1.0 1.3 0.7 0.7 3.0

2.1(85) 4.1045) 2.5(81)

0.6 1.2 0.7 _

0.3(20) 0.5(12) 0.4(26)

1.0 1.7 1.3 _

2.7(144) 2.7(106) 3.5(124)

0.8 0.8 1.0

0.4(38) 0.4(27) OS(37)

1.3 1.3 1.7

139.7(1760)

166.8(4203)

39.9

gene conversion at the trp5 (tryptophan) gene locus. ilv+ = reverse mutation at the ilc-92 (isoleucine) gene locus. Ratio (treatment/untreated control). Numbers in parentheses denote total numbers of colonies on triplicate Direct-acting control.

a trp+= b ’ d ’

D7 OF S. cereuisiae:

plates.

556.0

OF 24 h OF

25

: EXPERMENT

6:

‘!

1.0

h

0.0

0.6

0.4.

0.2.

cially when screening chemicals without replicating experiments. We have thus documented a gray area in genotoxicity testing, which could be interpreted as evidence that TGR and SCE tests performed poorly with the chemical, or that TBHQ itself has genotoxic potential which is very sensitive to slight differences in test conditions. As well as unequivocal compounds, there may very well be compounds of weak or intermediate genotoxicity. The interpretation chosen could have a major impact on regulatory decisions, as well as on future genotoxicity testing strategies. On the basis of the findings reported here, we are left with the question of what would be an appropriate strategy for screening unknown chemicals. SCE, TGR and yeast mutation assays have been used extensively for this purpose, as have other short-term tests. Different endpoints do not correlate perfectly (e.g. Galloway et al., 1987; Ashby, 1990; Boyes et al., 1990; FuEZ et al., 1990; Loveday et al., 1990; Wasserman et al., 1990), and no one of the short-term tests nor the rodent bioassay will always predict human carcinogenesis. To develop an optimal testing strategy will require more information about the performance of specific assays with particular chemical classes. Only then will we be able to discern patterns of response that will clearly indicate appropriate ways to assess new chemicals.

O-

Acknowledgements Fig. 3. Average number of thioguanine-resistant colonies of V79 cells per dish after treatment with TBHQ with and without activation by rat or hamster hepatocytes.

criteria for indication of a positive or negative result have been demonstrated (Boyes et al., 1990). The results of these studies could be interpreted to support arguments both for and against genotoxicity of this compound, depending on how many replicate experiments are considered. These results illustrate a difficulty in routine screening strategies for unknowns; while some individual experiments may indicate a positive response, others may lead to a negative assessment. Thus it would be easy to misclassify a compound, espe-

The authors thank Claudette H&roux-Metcalf, Irene Langlois, Nicolle Barrett and Robert Downie for excellent technical assistance. We also especially thank Dr. David Blakey for helpful comments and discussion of the manuscript. References Abe, S., and M. Sasaki (1977) Chromosome aberrations and sister chromatid exchanges in Chinese hamster cells exposed to various chemicals, _I. Nat]. Cancer Inst., 58, 1635-1641. Armstrong, K.E., and L.W. Wattenberg (1985) Metabolism of 3-tert.-butyl-4-hydroxyanisole to 3-tert.-butyl-4,Sdihydroxyanisole by rat liver microsomes, Cancer Res., 45, 15071510.

26 Ashby, J. (1990) Are short-term genetic tests useful for predicting carcinogenicity? A panel discussion revisited. Environ. Mol. Mutagen., 16, 324-327. Ashby, J., F.J. de Serres, M. Draper, M. Ishidate Jr., B.H. Margolin, B. Matter and M.D. Shelby (1985) Overview and conclusions of the IPCS collaborative study on in vitro assay systems. In: J. Ashby, F.J. de Serres et al. (Eds.). Progress in Mutation Research, vol. 5, ch. 10, pp. 117-174. Boyes, B.G., C.G. Rogers, K. Karpinski and R. Stapley (1990) A statistical evaluation of the reproducibility of micronucleus, sister chromatid exchange, thioguanine resistance and ouabain resistance assays in V79 cells exposed to ethyl methanesulfonate and 7,12-dimethylbenz[a]anthracene, Mutation Res., 234, 81-89. Boyes, B.G., C.G. Rogers and R. Stapley (1991) Genotoxicity of 1-nitronaphthalene in Chinese hamster V79 cells, Mutation Res., 259, 111-121. Bradley, M.O., B. Bhuyan, M.C. Francis, R. Langenbach, A. Peterson and E. Huberman (1981) Mutagenesis by chemical agents in V79 Chinese hamster cells: a review and analysis of the literature, Mutation Res., 87, 81-142. Chen, T.R. (1977) In situ demonstration of mycoplasma contamination in cell culture by fluorescent Hoechst 33258 stain, Exp. Cell Res., 104, 255-262. El-Rashidy, R., and S. Niazi (1983) A new metabolite of butylated hydroxyanisole in man, Biopharm. Drug Dispos., 4, 389-396. FuEiE, A., D. Horvat and B. Dimitrovid (1990) Mutagenicity of vinyl chloride in man: comparison of chromosome aberrations with micronucleus and sister-chromatid exchange frequencies, Mutation Res. 242, 265-270. Galloway, SM., M.J. Armstrong, C. Ruben, S. Colman, B. Brown, C. Cannon, A.D. Bloom, F. Nakamura, M. Ahmed, S. Duk, J. Rimpo, B.H. Margolin, M.A. Resnick, B. Anderson and E. Zeiger (1987) Chromosome aberrations and sister chromatid exchanges in Chinese hamster ovary cells: Evaluation of 108 chemicals, Environ. Mol. Mutagen., 10, Suppl. 1, 1-175. Giri, A.K., S. Sen, G. Talukder, A. Sharma and T.S. Banerjee (1984) Mutachromosomal effects of tert.-butylhydroquinone in bone-marrow cells of mice, Food Chem. Toxicol., 22, 459-460. Hageman, G.J., H. Verhagen and J.C.S. Kleinjans (1988) Butylated hydroxyanisole, butylated hydroxytoluene and tert.-butylhydroquinone are not mutagenic in the Salmonella/microsome assay using new tester strains, Mutation Res., 208, 207-211. Jenssen, D. (1984) A quantitative test for mutagenicity in V79 Chinese hamster cells, in: B.J. Kilby, M. Legator, W. Nichols and C. Ramel (Eds.), Handbook of Mutagenicity Test Procedures, Elsevier, Amsterdam, pp. 269-290. Langenbach, R., H.J. Freed and E. Huberman (1978) Liver cell-mediated mutagenesis of mammalian cells by liver carcinogens, Proc. Natl. Acad. Sci. (U.S.A.), 75,2864-2867. Langenbach, R., L. Malick and S. Nesnow (1981) Rat bladder cell-mediated mutagenesis of Chinese hamster V79 cells and metabolism of benzo[a]pyrene, J. Natl. Cancer Inst., 66, 913-917.

Latt, S.A., J. Allen, SE. Bloom, A. Carrano, E. Falke, D. Kram, E. Schneider, R. Schreck, R. Tice, B. Whitfield and S. Wolff (1981) Sister chromatid exchanges: A report of the Gene-Tox Program, Mutation Res., 87, 17-62. Loveday, K.S., B.E. Anderson, M.A. Resnick and E. Zeiger (1990) Chromosome aberrations and sister chromatid exchange tests in Chinese hamster ovary cells in vitro, V. Results with 46 chemicals, Environ. Mol. Mutagen., 16, 272-303. Madle, E., G. Tiedmann, S. Madle, A. Ott and G. Kaufmann (1986) Comparison of S9 mix and hepatocytes as external metabolizing systems in mammalian cell cultures: Cytogenetic effects of 7,12_dimethylbenzanthracene and aflatoxin Bl, Environ. Mutagen., 8, 423-437. Maslansky, C.J., and G.M. Williams (1982) Primary cultures and levels of cytochrome P450 in hepatocytes from mouse, rat, hamster and rabbit liver, In Vitro, 18, 683-693. Matsuoka, A., M. Matsui, N. Miyata, T. Sofuni and M. Ishidate Jr. (1990) Mutagenicity of 3-cert.-butyl-4-hydroxy anisole (BHA) and its metabolites in short-term tests in vitro, Mutation Res., 241, 125-132. Matula, T.I., and R.H. Downie (1984) Genetic toxicity of erythrosine in yeast, Mutation Res., 138, 153-156. Nouaim, R., and J.L. Dorange (1988) Effets genotoxiques du 2(31-tert.-butyl-4-hydroxyanisole mis en evidence par des tests avec Saccharomyces cerecisiae, Sci. Aliments, 8, 431445. Perry, P., and S. Wolff (1974) New Giemsa method for the differential staining of sister chromatids, Nature (London), 251, 156-158. Phillips, B.J., P.A. Carroll, A.C. Tee and D. Anderson (1989) Microsome-mediated clastogenicity of butylated hydroxyanisole (BHA) in cultured Chinese hamster ovary cells: The possible role of reactive oxygen species, Mutation Res., 214, 105-114. Phillipson, C.E., and C. Ioannides (1984). A comparative study of the bioactivation of nitrosamines to mutagens by various animal species including man, Carcinogenesis, 5, 1091-1094. Poiley, J.A., R. Raineri, A.W. Andrews, D.M. Cavanaugh and R.J. Pienta (1980) Metabolic activation by hamster and rat hepatocytes in the Salmonella mutagenicity assay, J. Natl. Cancer Inst., 65, 1293-1298. Rogers, C.G., and C. Heroux-Metcalf (1983) Cytotoxicity and absence of mutagenic activity of vomitoxin (4-deoxynivalenol) in an hepatocyte-mediated mutation assay with V79 Chinese hamster lung cells, Cancer Lett., 20, 29-35. Rogers, C.G., B.N. Nayak and C. Heroux-Metcalf (1985) Lack of induction of sister chromatid exchanges and of mutation to 6-thioguanine resistance in V79 cells by butylated hydroxyanisole with and without activation by rat or hamster hepatocytes, Cancer Lett., 27, 61-69. Rogers, C.G., B.G. Boyes, T.I. Matula, C. H&roux-Metcalf and D.B. Clayson (1988) A case report: A multiple endpoint approach to evaluation of cytotoxicity and genotoxicity of erythrosine (FD and C Red No. 3) in a V79 hepatocyte-mediated mutation assay, Mutation Res., 205, 415423.

27 van Esch, G.J. (1986) Toxicology of tert.-butylhydroquinone (TBHQ), Food Chem. Toxicol., 24, 1063-1065. Vleminckx, C., J. Arany, B. Hendrickx and W. Moens (1987) Genotoxic effects of glycidyltrimethylammonium chloride, Mutation Res., 189, 387-394. Wasserman, S.S., M.M. Cohen and S. Schwartz (19901 Factors underlying variation in spontaneous and clastogen-induced sister chromatid exchanges and chromosome breakage frequencies, Environ. Mol. Mutagen., 16, 255-259. Williams, G.M., E. Bermudez and D. Scarramuzzino (1977)

Rat hepatocyte primary culture, III. Improved dissociation and attachment techniques and the enhancement of survival by culture medium, In Vitro, 13, 809-817. Zimmermann, F.K. (197.5) Procedures used in the induction of mitotic recombination and mutation in the yeast Saccharomyces cerecisiae, Mutation Res., 31, 71-86. Zimmermann, F.K., R. Kern and H. Rasenberger (1975) A yeast strain for simultaneous detection of induced crossing over, mitotic gene conversion and reverse mutation, Mutation Res., 28, 381-388.