Genetic Toxicology

ELSEVIER

Mutation Research 371 (1996) 165-173

Genotoxicity of the laxative drug components emodin, aloe-emodin and danthron in mammalian cells: Topoisomerase II mediated? Stefan O. Miiller, Inge Eckert, Werner K. Lutz, Helga Stopper * Department ~?fToxicology, Unit'ersity of Wi~r=burg, Versbacher Str. 9, 97078 Wiirzburg, Germany Received 31 January 1996; revised 21 May 1996; accepted 6 August 1996

Abstract 1,8-Dihydroxyanthraquinones are under debate as plant-derived carcinogens that are found in laxatives, food colors, and possibly vegetables. Published genotoxicity data are controversial, and so three of them (emodin, danthron and aloe-emodin) were tested in a number of in vitro assay systems. All three compounds induced tk-mutations in mouse lymphoma L5178Y cells. Induction of micronuclei also occurred in the same cell line, and was dose-dependent, with the potency ranking being danthron > aloe-emodin > emodin. In a DNA decatenation assay with a network of mitochondrial DNA of C. fasciulata, all three test compounds inhibited the topoisomerase II-mediated decatenation. Danthron and aloe-emodin, but not emodin, increased the fraction of DNA moving into comet tails when tested at concentrations around 50 txM in single-cell gel-electrophoresis assays (SCGE; comet assay). Comet assays were also used in modified form to determine whether pretreatment of the cells with the test compounds would reduce the effects of etoposide, a potent topoisomerase II inhibitor. All three test chemicals were effective in this pretreatment protocol, with danthron again being the most potent. Given clearcut evidence of their genotoxic activity, further research on the human cancer risk of these compounds may be warranted.

KeDvords: Micronucleus; Comet assay; Emodin; Danthron; Aloe-emodin; 1,8-Dihydroxyanthraquinone: Mouse L5178Y cell; Topoisomerase II

1. Introduction H y d r o x y a n t h r a q u i n o n e s are the active principle of m a n y plant-derived drugs [1], of which laxatives

Abbreviations: DMSO, dimethyl sulfoxide; EMS, ethylmethanesulfonate; FITC, fluoresceinisothiocyanate; gpt, guaninphosphoribosyltransferase; hprt, hypoxanthinphosphoribosyltransferase; kDNA, kinetoplast DNA; SCGE. single cell gel electrophoresis; TFT~ trifluorothymidine; tk, thymidine kinase * Corresponding author. Tel: (0)931-201-3427, Fax: (0)931201-3446.

such as aloe and senna are the most widely used. Plant extracts are also used for the treatment of kidney and bladder stones ( R u b i a tinct.) and as mild sedatives ( H y p e r i c u m perf.). Naturally occurring hydroxyanthraquinones have also been used as dyes in food industry [2] and have been found in food plants [3]. Most of the h y d r o x y a n t h r a q u i n o n e s are present as pharmacologically inactive glycosides in plant extracts which can be activated by glycosidic cleavage in vivo by organisms in the intestinal flora [4]. The 1,8-dihydroxyanthraquinone danthron (Fig. 1;

0165-1218/96/$15.00 Copyright © 1996 Elsevier Science B.V. All rights reserved. PII S0165- 1218(96)00092-4

166

S.O. Miiller et al./Mutation Research 371 (1996) 165-173 HO

O

OH

H O ~ C H a O Emodin

HO

O

OH

O Danthron

HO

0

OH

~CH2OH O Aloe-Emodin

Fig. 1. Chemicalstructuresof the investigated1,8-dihydroxyanthraquinones. synonym: chrysazin), produced an increased incidence of intestinal tumors in rats, when fed at 1% in the diet [5] and adenomatous hyperplasias with cystic glands of the caecum and liver tumors in mice when fed at 0.2% in the diet [6]. An increased relative risk for colorectal cancer in humans has been reported among users of 1,8-dihydroxyanthraquinone-containing laxatives [7,8]. Several 1,8-dihydroxyanthraquinones have been assayed for their genotoxicity in vitro. Emodin (Fig. 1), a 1,8-dihydroxyanthraquinone found in laxative drugs, induced mutations in the Ames test but usually only in the presence of metabolic activation [9-13]; this compound is also said to be antimutagenic in Salmonella typhimurium [14,15]. Mutagenicity of emodin in mammalian cells is controversial. Bruggemann and van der Hoeven [10] reported that emodin is not mutagenic in the V79 hprt assay with and without metabolic activation, whereas Westendorf et al. [12] reported a weak positive response

in this test in the absence of metabolic activation. Morita et al. [16] tested mouse FM3A cells at the hprt locus without metabolic activation and found a clear mutagenic response, although only when toxicity was apparent. Emodin has been found to induce sister chromatid exchanges in CHO cells without metabolic activation [3]. 2-Hydroxyemodin, the presumed mutagenic metabolite of emodin in the Ames test [17], has also been assayed for its mutagenicity in mammalian cells, and appears to be somewhat less mutagenic than emodin [16]. Another laxative constituent, aloe-emodin (Fig. 1), was also mutagenic in the Ames test in the absence of metabolic activation [18], but showed reduced activity when a metabolic activation system was present [12,19]. In mammalian cells [12], aloeemodin produced a weakly positive response in the V79 hprt mutation assay without metabolic activation. Heidemann et al. [18] reported an effect of aloe-emodin in the Chinese hamster ovary cell chromosome aberration test with and without metabolic activation, but did not observe any mutagenic activity in the V79 hprt mutation assay, either with or without metabolic activation. Danthron (Fig. 1) was formerly used as laxative. This compound was positive in the Ames test with and without metabolic activation [12,13], but resembled emodin, in exhibiting antimutagenic activity in this same assay under modified conditions [14]. Danthron was negative in the V79 hprt mutation assay when tested in the absence of metabolic activation [20]. Danthron did, however, induce chromosomal damage in human lymphocytes in vitro [21] or in V79 cells [22], both without metabolic activation. Given the use of anthraquinone drugs as herbal remedies and their occurrence as natural constituents in food plants, increased knowledge about their genotoxic potential is obviously required. Many (but not all) investigations carried out to date in mammalian cells have suggested that they may have genotoxic effects. We therefore decided to investigate the genotoxicities of emodin, aloe-emodin and danthron in (i) mouse TFT-resistance L5178Y tk +/ mutation assays, and (ii) micronucleus assays with mouse lymphoma L5178Y cells. We also wished to study mechanistic aspects of an observed genotoxicity. Hydroxyanthraquinones are planar molecules that are presumed to intercalate into DNA [23]. Other

S.O. M~ller et a l . / Mutation Research 371 (1996) 165-173

anthracene derivatives are known to inhibit topoisomerase II [24]. Topoisomerase II participates in a number of cellular processes including chromosome segregation [25,26] and maintenance of genomic stability [27]. Inhibition or interference with topoisomerase II activity at critical stages of the cell cycle could lead to chromosome breakage [28]. To test this hypothesis we employed a cell-free topoisomerase II-specific decatenation assay and the comet assay (single cell gel electrophoresis) to measure DNA breakage events. 2. Materials and methods Chemicals. Aloe-emodin (CAS no. 481-72-1), danthron (117-10-2) and emodin (518-82-1) were acquired from Roth company (Karlsruhe, Germany). m-Amsacrine (5430 l- 15-4), etoposide (334-15-42-0), EMS (62-50-0), Hoechst 33258, hypoxanthine, methotrexate, thymidine, TFT, Tween 20, vinblastine and FITC-conjugated goat anti human antibody were purchased from Sigma Chemie GmbH (Deisenhofen, Germany). DMSO and formamide were from Aldrich Company Europe (Steinheim, Germany) and CREST-Serum (anti-kinetochore-antibody) was purchased from Biermann GmbH (Bad Nauheim, Germany). Test compounds were dissolved in DMSO. The final concentration of DMSO in the cell culture medium did not exceed 1%. Cell culture. Mouse L5178Y cells, clone 3.7.2c, were cultured in suspension in RPMI-1640 supplemented with antibiotics, 0.25 m g / m l L-glutamine, 107 ~ g / m l sodium pyruvate, and 10% heat inactivated horse serum (all from Sigma Chemie GmbH, Deisenhofen, Germany). Cell cultures were grown in a humidified atmosphere with 5% CO 2 in air at 37°C. Mutation assay. Cultures of mouse L5178Y cells were treated with methotrexate before each experiment to kill pre-existing TFT-resistant cells. To accomplish this, cells were incubated for 12 h in culture medium plus methotrexate (0.3 txg/ml), thymidine (9 p~g/ml), hypoxanthine (15 p~g/ml) and glycine (22.5 ~ g / m l ) . The cells were then incubated for at least 48 h in the same medium without methotrexate. To measure chemically induced mutants using the in situ procedure, cultures containing 1 X l 0 6 cells in 5 ml medium were treated

167

with DMSO (final concentration 1%; vehicle control) or 1,8-dihydroxyanthraquinones with increasing doses until insolubility. EMS was used as a positive control. Incubation was performed for 4 h, then the cells were washed twice with fresh medium. After that, from each tested culture, 0.5 X 10 6 cells were added to 50 ml of semi-solid culture medium (containing 0.25% granulated agar, Baltimore Biological Laboratories, USA) and plated into two plastic 100mm culture dishes and allowed to solidify at room temperature. TFT-resistant cells were selected by adding an overlay of TFT (final concentration 8 l~g/ml) in 10 ml semisolid medium after an expression time of 42 h. Cloning efficiency was determined by adding 600 cells to 100 ml of semi-solid medium in three 100 mm culture dishes. All plates were incubated for a total of 9 to 12 days at 37°C in 5% CO 2 for colony growth. An automatic colony counter was used to count the number of TFT-resistant colonies. In uitro micronucleus assay. Exponentially growing mouse L5178Y cells were treated for 4 h. The vehicle control was 1% DMSO. After removing the chemicals by centrifugation and medium replacement, the cells were incubated for 15 h (expression time). The cells were then brought onto glass slides by cytospin-centrifugation and were fixed with methanol ( - 2 0 ° C , 1 h). To stain nuclei and micronuclei, the slides were incubated with Hoechst 33258 (5 ~ g / m l , 3 rain), washed twice with Ca- and Mg-free PBS buffer for 2 rain and mounted for microscopy. Numbers of nuclei and micronuclei were scored at a magnification of 1250 X. Each data point represents the mean of three slides, with 2000 nuclei being evaluated on each slide. All experiments were repeated with consistent results. Kinetochore analysis. Kinetochore staining was achieved by incubating the fixed cell preparations (after washing 5 min in PBS/0.1% Tween 20) with CREST serum for 75 min in a humidified chamber at 37°C. After washing twice for 5 min in PBS/0.5% Tween 20 again, the cells were incubated as before for 30 min with FITC-conjugated goat-anti-human antibody (diluted 1:100 in PBS/0.5% Tween 20, pH 7.4), washed again twice in PBS/0.1% Tween 20 and counterstained with Hoechst 33258 (5 txg/ml, 3 min). Micronuclei were analyzed for the presence of a kinetochore signal at a magnification of 1250 X.

168

S.O. MMler et a l . / Mutation Research 371 (1996) 165-173

Topoisomerase H assay. Topoisomerase II assays were performed according to the protocol provided by TopoGen, Inc. (Columbus, USA). The reactions contained 150 ng kDNA, 4 U of topoisomerase II (170 kDa form) and the test chemicals (dissolved in DMSO, final concentration 10%) in 20 ~zl reaction buffer (30 mM Tris-HC1 pH 7.6, 3 mM ATP, 15 mM 2-mercaptoethanol, 8 mM MgC12, 60 mM NaC1). The reactions were incubated for 15 min at 37°C and terminated with 1% of SDS, followed by proteinase K (20 ~ g / m l ) treatment for 15 min at 37°C. After addition of 0.1 vol. gel loading buffer (0.25% bromphenol blue, 50% glycerol), the samples were extracted once with an equal volume of chloroform/isoamylalcohol (24:1). The blue upper layer was loaded onto a 1% agarose gel. The products were resolved by gel electrophoresis in 1 × TAE buffer for 30 rain at 100 V, which separated the catenated kDNA from the decatenated DNA monomers. Gels were stained with SYBR TM Green (Molecular Probes Europe, The Netherlands) and photographed. Comet assay. Comet assays were performed according to Singh et al. [29] with slight modifications. Mouse L5178Y cells were incubated as indicated in the legend to Fig. 4. Cells were then washed and embedded in low melting agarose (0.5%) which was layered onto fully frosted microscope slides that had been coated with a layer of 0.75% normal agarose (diluted in Ca- and Mg-free PBS buffer). A final layer of 0.5% low-melting agarose was added on top. Slides were immersed in a jar containing cold lysing solution (1% Triton X-100, 10% DMSO and 89% of 10 mM Tris/1% sodium laurylsarcosine/2.5 M NaC1/100 mM Na2EDTA (pH 10) for lysis at 4°C (1 h). Then, slides were pretreated for 15 min in electrophoresis buffer (300 mM NaOH/1 mM N%EDTA (pH 13) and after that exposed to 25V/300 mA for 20 min. Preincubation and electrophoresis were performed in an ice bath. Slides were neutralized for 3 X 5 min in 0.4 M Tris, pH 7.5 and DNA was stained by adding 50 txl of ethidium bromide (20 txg/ml) onto each slide. Cells were analyzed with a 1250X magnification and using computer-aided image analysis. Images of at least 50 cells (25 from each of two slides) were evaluated by the use of the software program NIH Image 1.54 (NIH, USA). Two areas were selected in each pic-

ture: the whole cellular DNA including the tail region of the comet and a region containing only the tail region of the comet. The integrated densities (sum of the gray values of each pixel in the selection) were measured in each selection and the percentage in the tail region of the comet was calculated. This number represents the amount of DNA in the tail and is referred to as 'tail (%)' in Fig. 4. Medians and standard errors for each treatment are given.

3. Results and discussion

3.1. Mutagenicio, and micronucleus induction Mutagenic activity as measured by the induction of TFT-resistance in mouse L5178Y cells was assayed for the 1,8-dihydroxyanthraquinones emodin, danthron and aloe-emodin (Table 1). All three compounds induced a moderate increase in mutant fraction. Toxicity as shown by a reduction in the cloning efficiency was low in the applied dose range. A dose-dependent induction of micronuclei by all three compounds was observed in the same cell line (Fig. 2). Emodin yielded fewer micronuclei than danthron and aloe-emodin. Micronuclei were analyzed for the presence of stainable kinetochore protein (Table 2). While the positive control vinblastine mainly induced kinetochore-containing micronuclei, the vast majority of anthraquinone-induced micronuclei provided no kinetochore signals. This implies that chromosomal fragments rather than whole chromosomes were generated, thereby indicating that the compounds were probably clastogenic [30,31]. Thus, all three 1,8-dihydroxyanthraquinones appeared to be genotoxic in mammalian cells. Previously published data from mammalian mutation assays by other investigators were not consistently positive (see for example [10,12,16,18,20]). However, the earlier data relates to the hemizygous hprt locus, a locus that is known to be less sensitive (especially for the detection of compounds that induce large deletions or multi-locus lesions, see Refs. [32,33]). Our study made use of the highly sensitive heterozygous thymidine kinase (tk) locus in the L5187Y cell line. To support the mutagenicity data we tested the

S.O. Miiller et al. / Mutation Research 371 (1996) 165-173

169

Table 1 Mutagenicity of emodin, danthron and aloe-emodin in the mouse lymphoma L5178Y tk ÷ / - assay Substance

Control (DMSO) EMS Emodin

Control (DMSO) EMS Danthron

Control (DMSO) EMS Aloe-Emodin

Concentration

Experiment I

Experiment II

(p~M) efficiency

Cloning fraction

Mutant mutant

Relative efficiency fraction a

Cloning fraction

Mutant mutant

(1%) 2000 37 55.5 74 111 (1%) 2000 21 42 63 84 126 (1%) 2000 37 55.5 74 l 11

0.65 0.73 0.67 0.62 0.71 0.44 0.52 0.27 0.69 0.40 0.52 0.46 0.52 0.27 0.36 0.55 0.65 0.51

142 1161 269 300 214 386 277 969 609 1195 1042 891 277 969 406 651 348 686

l 8.2 1.9 2.2 1.6 2.8 I 3.5 2.2 4.3 3.8 3.2 1 3.5 1.5 2.4 1.3 2.5

0.93 0.70 0.76 0.69 0.71 0.57 0.52 0.38 0.30

120 206 208 162 211 284 435 1447 520

l 1.7 1.7 1.4 1.8 2.4 1 3.3 1.2

0.35 0.41 0.43 0.38 0.26 0.23 0.37 0.26

629 873 326 1695 639 661 1005 731

1.5 2.0 1 5.2 2.0 2.0 3.1 2.2

Relative fraction ~

Mutant fraction relative to the control mutant fraction.

1,8-dihydroxyanthraquinones in another mammalian cell line for mutagenicity with a heterozygous locus (gpt in AS52 Chinese hamster cells). The highest relative mutation frequencies obtained were 4.9 (28 IxM emodin; titer: 0.42), 6.1 (32 IxM danthron; titer: 0.27) and 4.7 (28 txM aloe-emodin; titer: 0.48). In this experiment the positive control EMS (4 mM) showed a relative mutation frequency of 9.2 (titer: 0.49) and the solvent control titer was 0.49. Thus the in vitro assays used here appeared to be highly suitable for the detection of the types of genotoxic damage induced by these compounds. The

cell systems we used are probably not capable of metabolically activating these compounds [34] in which case the effects we observed may well reflect activation-independent genotoxicity.

3.2. Inhibition of topoisomerase H The genotoxicity of our three test-compounds may be related to their ability to intercalate into DNA. Since many known topoisomerase II inhibitors, including anthracycline derivatives are DNA-intercalating compounds [24], we tested the hypothesis

Table 2 Presence of kinetochore signals in micronuclei induced in mouse L5178Y cells Substance

Concentration

Evaluated micronuclei (n)

Kinetochore-positive micronuclei (n)

Kinetochore-positive micronuclei (%)

Control (DMSO) Vinblastine Emodin Danthron Aloe-Emodin

1% 11 nM 74 IxM 21 ~M 37 tzM

211 300 300 201 300

34 239 19 11 11

16 80 6 6 4

170

S.O. Miiller et al. /Mutation Research 371 (1996) 165-173 Micronucleated cells / 1000 cells

140 2O

15 ; 120

/

. o4,

10

o° i

5

100

. . . . . . .

. - *

o~

•

~,.~o~,~

G

oo

w

~

~o .

•

0 0

,#

20 40 60 80

Emodin[UM]

NW

•l 8O MC

60

), " " / /

40

/ •

/

/

20

" •

/

1

/

0

20

40

60

Emodin

"--m--Aloe-Emodin

3

4

5

Fig. 3. Assay for topoisomerase II activity and its inhibition. All lanes loaded with 150 ng kDNA. Lanes 2-5 after incubation with 4 U topoisomerase II. Lanes I and 2 without inhibitor; lane 3:1 mM emodin; lane 4:1 mM danthron; lane 5:1 mM aloe-emodin. NW, network DNA; MC, monocircle DNA.

80

c [pM]

i

2

- i - Danthron

Fig. 2. Micronucleus induction by 1,8-dihydroxyanthraquinones in mouse L5178Y cells treated with emodin, aloe-emodin and danthron.

a more potent inhibitor than danthron or aloe-emodin (Table 3).

that the anthraquinones investigated here inhibit the activity of topoisomerase II. The activity of this enzyme was assayed by the decatenation of kDNA, a catenated network of mitochondrial DNA rings isolated from Crithidia fasciculata [35]. The decatenation reaction was monitored by the appearance of 2.5 kilobase DNA monomers following gel electrophoresis (Fig. 3). Emodin, danthron and aloe-emodin reduced the amount of monomer DNA generated by topoisomerase II, indicating that all three compounds were capable of inhibiting this enzyme. Emodin was Table 3 Inhibition of topoisomerase II by 1,8-dihydroxyanthraquinones

Percentage of catenated DNA a Percentage of monocircle DNA b

3.3. Comet assay While topoisomerase II-inhibiting activity has clearly been shown in a cell-free assay it is not clear whether the compounds can also exert that effect in intact cells. We therefore applied the comet assay (SCGE). Agents such as m-amsacrine and etoposide form the so-called 'cleavable complex' with topoisomerase II and DNA [24]. For the duration of these 'cleavable complexes', the DNA contains double

in a DNA decatenation assay

Control 1 (kDNA)

Control 2 (kDNA + topo II)

Emodin l mM

Danthron 1 mM

Aloe-Emodin 1 mM

100 0

31 39

94 0

85 15

87 10

a Percent catenated kDNA (band NW in Fig. 2) of the total DNA (sum of all detectable bands). The respective amounts of DNA were calculated as the integrated densities of the respective DNA bands (program NIH Image 1.54). b Percent monocircle kDNA (band MC in Fig. 2) of the total DNA (sum of all detectable bands).

S.O. Miiller et al. /Mutation Research 371 (1996) 165-173

strand breaks that are only held together by the enzyme. Thus, m-amsacrine and etoposide-mediated DNA break frequencies that can be measured in certain assays like alkaline elution (e.g. [36-38]) or the comet assay (e.g. [39-41]) were used to signal interference with topoisomerase II activity, mAmsacrine and etoposide induced the formation of comets (Fig. 4, upper panel). Danthron and also aloe-emodin, but not emodin induced DNA damage in the comet assay (Fig. 4, upper panel), albeit at concentrations 1000-fold higher than were required for m-amsacrine. Tail (%) 50

40

i l)i

30

iiiiii 20

iiiill

10

¢ 80--

171

If inhibition of topoisomerase II by the 1,8-dihydroxyanthraquinones is not mediated by the formation of the 'cleavable complex', but by preventing the attachment of the enzyme to the DNA, the regular comet assay may not be very sensitive, since no 'cleavable complex' mediated DNA strand breaks are induced. If in that situation a 'cleavable complex' forming topoisomerase II inhibitor like etoposide is applied subsequently to the 1,8-dihydroxyanthraquinones, the etoposide induced damage should be reduced, since formation of the 'cleavable complex' through etoposide activity is not possible without prior attachment of the enzyme to the DNA, which would be prevented by the 1,8-dihydroxyanthraquinones. Therefore, we used a combination treatment, in which we applied the 1,8-dihydroxyanthraquinones first and added etoposide later and investigated whether the DNA cleavage induced by etoposide is less efficient than without 1,8-dihydroxyanthraquinone pretreatment. We included a combination of the DNA intercalator m-amsacrine and etoposide to determine whether the subsequent treatment with two known topoisomerase II inhibitors with different mechanisms of action in the comet assay actually yields a reduction of DNA damage (i.e., less than additive; positive control). Indeed, the etoposide-induced comet-formation was reduced by the pretreatment with m-amsacrine and all three 1,8-dihydroxyanthraquinones, with danthron being the most and emodin being the least effective of the 1,8-dihydroxyanthraquinones (Fig. 4, lower panel).

60--

40 -

\ \ \ \

\ \ \ \ \

-.J ,,J

)

\

\

\ \

20

0

~× Y.,

]

Calculated sum

.x- v.,

[]

_x ,9'

Sequential treatment

Fig. 4. Comet assay (SCGE). DNA damage is expressed as percent in tail. The medians with standard errors are given for each treatment. Cells were treated for 5 h with DMSO (1%; control), m-amsacrine (51 nM), emodin (55.5 IxM), danthron (62.4 IxM) and aloe-emodin (55.5 I~M), respectively; for 3.5 h with etoposide (255 nM). In the combined exposure experiment (lower panel) cells were incubated with the above indicated concentrations of m-amsacrine ( A M S A / E t o ) , emodin (Emodin/Eto), danthron (Danthron/Eto) and aloe-emodin (Aloe/Eto). At 1.5 h etoposide (Eto) was added and the cells were further incubated for 3.5 h. The overall incubation time was 5 h. The calculated sum of the net effect of the respective single exposures is shown in the respective left columns (striped; substance + Eto) to compare the differences of the combined exposures (right columns, substance/Eto) with the respective single exposures (upper panel)•

172

s.o. Miiller et al. / Mutation Research 371 (1996) 165 173

There is still discussion about the relationship b e t w e e n the types of D N A d a m a g e and the a m o u n t of effect measurable in the c o m e t assay [42] and m e c h a n i s m s different from inhibition o f topoisomerase II may contribute to the o b s e r v e d effects. H o w e v e r , we feel that the most likely interpretation o f these experimental data is that the 1,8-dihydroxyanthraquinones inhibit the interaction b e t w e e n topoisomerase II and D N A . O n e o f the c o n s e q u e n c e s most likely manifests at mitosis, w h e n t o p o i s o m e r a s e II activity is required to separate the chromatids [26]. W h e n the chromatids are not c o m p l e t e l y separated but pulled apart, chromatin bridges are f o r m e d [28]. S o m e of these chromatin bridges may rupture and D N A fragments may be included in micronuclei or lost. D e p e n d i n g on the genes that are lost mutants may also be formed. E m o d i n was more active in the cell-free assay and less active in the cell culture assays than danthron and aloe-emodin. It is possible that the c o m p o u n d s interacted differently with c o m ponents of the culture m e d i u m or the cell before reaching the D N A . A n o t h e r possible explanation may be a different substrate specificity of the two cellular t o p o i s o m e r a s e II subtypes (c~ and [3) only one of which ((x) is used in the decatenation assay. A l t h o u g h the intestinal absorption of a l o e - e m o d i n [18] and the e x p e c t e d concentrations of 1,8-dihydroxyanthraquinones achievable in human tissue by medication or food intake are rather low, local accumulation is possible. W h e t h e r the postulated increased relative risk o f colorectal cancer after the abuse o f anthranoid laxatives [7] is due to the genotoxicity o f their 1,8-dihydroxyanthraquinone contents or to other aspects of laxative abuse and whether anthraquinones present in foods also represent a cancer risk can only be j u d g e d along with information on content and appropriate dose-response considerations.

Acknowledgements This study was supported by the Swiss Federal Office o f Public Health ( B A G grant No. F E 316.95.0500). W e thank Dr. J. Schlatter, monitoring scientist o f the B A G , for valuable advice in all parts of this work, and Mrs. A. Golka and Mrs. N. Herrmann for expert technical assistance.

References [1] Thomson, R. H. (1986) III. Recent Advances, Naturally Occurring Quinones, Chapman and Hall, London. [2] Banthorpe, D. V. and J. J. White (1995) Novel anthraquinones from undifferentiated cell cultures of Galium verum, Phytochemistry, 38, 107-111. [3] Von Wright, A., O. Raatikainen, H. Taipale, S. Kfirenlampi and J. Maki-Paakkanen (1992) Directly acting geno- and cytotoxic agents from a wild mushroom Dermocybe sanguinea, Mutat. Res., 269, 27-33. [4] Hattori, M., T. Akao, K. Kobashi and T. Namba (1993) Cleavages of the O-and C-Glucosyl Bonds of Anthrone and 10,10'-Bianthrone Derivatives by Human Intestinal Bacteria, Pharmacology, 47 (suppl. 1), 125-133. [5] Mori, H., S. Sugie, K. Niwa, M. Takahashi and K. Kawai (1985) Induction of intestinal tumors in rats by chrysazin, Br. J. Cancer, 52, 781-783. [6] Mori, H., S. Sugie, K. Niwa, N. Yoshimi, T. Tanaka and 1. Hirono (1986) Carcinogenicity of Chrysazin in Large Intestine and Liver of Mice, Japanese Journal of Cancer Research, 77, 871-876. [7] Siegers, C. P., E. yon Hetzberg-Lottin, M. Otte and B. Schneider (1993) Anthranoid laxative abuse - a risk fbr colorectal cancer?, Gut, 34(8), 1099-1101. [8] Kune, G. A. (1993) Laxative use not a risk lbr colorectal cancer: data from the Melbourne colorectal cancer study, Z. Gastroenterol., l, 140-143. [9] Tikkanen, L., T. Matsushima and S. Natori (1983) Mutagenicity of anthraquinones in the Salmonella preincuabation test, Mutat. Res., 116, 297-304. [10] Bruggemann, I. M. and J. C. M. van der Hoeven (1984) Lack of activity of the bacterial mutagen emodin in HGPRT and SCE assay with V79 Chinese hamster cells, Mutat. Res., 138, 219-224. [11] B6sch, R., U. Friederich, W. K. Lutz, E. Brocker, M. Bachmann and C. Schlatter (1987) Investigations on DNA binding in rat liver and in Salmonella and on mutagenicity in the Ames test by emodin, a natural anthraquinone, Mutat. Res., 188, 161-168. [12] Westendorf, J., H. Marquardt, B. Poginsky, M. Dominiak, J. Schmidt and H. Marquardt (1990) Genotoxicity of naturally occurring hydroxyanthraquinones, Mutat. Res., 240, 1-12. [13] Krivobok, S., F. Seigle-Murandi, R. Steiman, D. R. Marzin and V. Betina (1992) Mutagenicity of substituted anthraquinones in the Ames/Salmonella microsome system, Mutat. Res., 279(1), 1-8. [14] Hao, N. J., M. P. Huang and H. Lee (1995) Structure-activity relationships of anthraquinones as inhibitors of 7ethoxycoumarin O-deethylase and mutagenicity of 2-amino3-methylimidazo(4,5-f)quinoline, Murat. Res., 328, 183-19 l. [15] Su, H. Y., S. H. Cheng, C. C. Chen and H. Lee (1995) Emodin inhibits the mutagenicity and DNA adducts induced by l-nitropyrene, Mutat. Res., 329, 205-212. [16] Morita, H., H. Umeda, T. Masuda and Y. Ueno (1988) Cytotoxic and mutagenic effects of emodin on cultured mouse carcinoma FM3A cells, Mutat. Res., 204, 329-332.

S.O. Miiller et al. / Mutation Research 371 (1996) 165-173

[17] Masuda, T., K. Hairakawa, N. Morooka, S. Nakano and Y. Ueno (1985) 2-Hydroxy-emodin, an active metabolite of emodin in the hepatic microsomes of rats, Mutat. Res., 149, 327-332. [18] Heidemann, A., H. G. Miltenburger and U. Mengs (1993) The genotoxicity status of senna, Pharmacol., 47, 178-186. [19] Brown, J. P. (1980) A review of the genetic effects of the naturally occuring flavonoids, anthraquinones and related compounds, Mutat. Res., 75, 243-277. [20] Wi~lfle, D., J. Westendorf, C. Schmutte, M. Dominiak, B. Poginski and H. Marquardt (1989) Genotoxicity, cell transformation and growth stimulation (promotion) in vitro induced by hydroxyanthraquinones, Proc. Amer. Canc. Res., 30, 142. [21] Carballo, M. A., M+ D'Aquino and E. I. Aranda (1981) Accion clastogenica de un compuesto antraquinonico en linfocitos humanos, Medicina (Buenos Aires), 41, 531-534. [22] Simi, S., S. Morelli, P. G. Gervasi and G. Rainaldi (1995) Clastogenicity of anthraquinones in V79 and in three derived cell lines expressing P450 enzymes, Mutat. Res., 347, 151156. [23] Swanbeck, G. (1966) Interaction between deoxyribonucleic acid and some anthracene and anthraquinone derivatives, Biochem. Biophys. Acta, 123, 630-633. [24] Anderson, R. D. and N. A. Berger (1994) Mutagenicity and carcinogenicity of topoisomerase-interactive agents, Mutation Research, 309, 109-142. [25] Holm, C., T. Stearns and D. Bostein (1989) DNA topoisomerase II must act at mitosis to prevent nondisjunction and chromosome breakage, Mol. Cell. Biol., 9, 159-168. [26] Downes, C. S., A. M. Mullinger and R. T. Johnson (1991) Inhibitors of topoisomerase II prevent chromatid separation in mammalian cells but do not prevent exit from mitosis, Proc. Natl. Acad. Sci. USA, 88, 8895-8899. [27] Wang, J. R., P. R. Caron and R. A. Kim (1990) The role of DNA topoisomerases in recombination and genomic stability: a double-edged sword'?, Cell, 62, 403-406. [28] Gaulden, M. E. (1987) Hypothesis: some mutagens directly alter specific chromosomal proteins (DNA topoisomerase II and peripheral proteins) to produce chromosome stickiness, which causes chromosome aberrations, Mutagenesis, 2, 357365+ [29] Singh, N. P., M. T. McCoy, R. R. Tice and E. L. Schneider (1988) A simple technique for quantitation of low levels of DNA dammage in individual cells, Exp. Cell Res., 175, 184-191. [30] Stopper, H., I. Eckert, D. Schiffmann, D. L. Spencer and W. J. Caspary (1994) Is micronucleus induction by aneugens an early event leading to mutagenesis?, Mutagenesis, 9(5), 411416.

173

[31] Combes, R. D., H. Stopper and W. J. Caspary (1995) The use of the L5178Y mouse lymphoma cells to assess mutagenic, clastogenic and aneugenic activity of chemicals, Mutagenesis, 10(5), 403-408. [32] Shibuya, M. L., A. M. Ueno, D. B. Vannais, P. A. Craven and C. A. Waldren (1994) Megabase pair deletions in mutant mammalian cells following exposure to amsacrine, an inhibitor of DNA topoisomerase II, Cancer Res., 54, 10921097+ [33] McGregor, D. B., C. Riach, P. Cattanach, I. Edwards, W. Shepherd and W. J. Caspary (1995) A comparison of mutagenic response of L5178Y mouse cells at the tk and hprt loci, Environm. Molec. Mutagen., in press. [34] McGregor, D. B., I. Edwards, C. R. Wolf, L. M. Forrester and W. J. Caspary (1991) Endogenous xenobiotic enzyme levels in mammalian cells, Mutat. Res., 261, 29-39. [35] Marini, J. C., K. G. Miller and P. T. Englund (1980) Decatenation of kinetoplast DNA by topoisomerase, J. Biol. Chem., 255, 4976-4979. [36] Zwelling, L. A., J. Minford, M+ Nichols, R. I. Glazier and S. Shackney (1984) Enhancement of intercalator-induced deoxyribonucleic acid scission and cytotoxicity in murine leukemia cells treated with 5-azacytidine, Biochem. Pharmacol., 33(23), 3903-3906. [37] Kaufmann, W. K., J. C. Boyer, L. L. Estabrooks and S. J. Wilson (1991) Inhibition of replicon initiation in human cells following stabilization of topoisomerase-DNA cleavable complexes, Mol. Cell. Biol., 11(7), 3711-3718. [38] Dubrez, L., F. Goldwasser, P. Genne, Y. Pommier and E. Solary (1995) The role of cell cycle regulation and apoptosis triggering in determining the sensitivity of leukemic cells to topoisomerase I and lI inhibitors, Leukemia 9(6), 1013-1024, Topo I1. [39] Olive, P. L., J. P. Banath and R. E. Durand (1990) Detection of etoposide resistance by measuring DNA damage in individual Chinese hamster cells, J. Natl. Cancer Inst., 82, 779-783. [40] Olive, P. L. and J. P. Banath (1993) Detection of DNA double-strand breaks through the cell cycle after exposure to X-rays, bleomycin, etoposide and 125IdUrd, Int. J. Radiat. Biol., 64, 349-358. [41] Stopper, H., I. Eckert, P. Wagener and W. Schulz (1995) Formation of micronuclei and inhibition of topoisomerase II in the comet-assay in mammalian cells with altered DNAmethylation, Recent Results Cancer Res., in press. [42] McKelvey-Martin, V. J+, M. H. L+ Green, P. Scbmezer, B. L. Pool-Zobel, M. P. De Meo and A. Collins (1993) The single cell gel electrophoresis assay (comet assay): a European review, Mutat. Res., 288, 47-63.