Mutagenesis vol.11 no.5 pp.497-504, 1996

The photomutagenicity of fluoroquinolones in tests for genemutation, chromosomal aberration, gene conversion and DNAbreakage (Comet assay)

Andree-Anne Chltelat1, Silvio Albertini andElmar Gocke

Pharma Division, Preclinical Research, Department of Toxicology,F.Hoffmann-La Roche Ltd, CH-4070 Basel, Switzerland

'To whom correspondence should be addressed

The ability of fluoroquinolones to cause light-inducedadverse effects has been established in experimental studiesand clinical observations. The formation of active oxygenspecies appears to be responsible for this activity. Photo-mutagenicity tests with bacterial, lower eukaryotic andmammalian cells were performed with three fluoro-quinolones (Fleroxacin, Ciprofloxacin and Lomefloxacin).After concomitant irradiation with simulated solar light(with a reduced UVB component), weak increases in thenumber of revertants were observed in Salmonella typhimu-rium TA104 and TA100. No photomutagenic activity wasdetected in Saccharomyces cerevisiae D7. In the chromo-somal aberration (CA) test with Chinese hamster V79cells the number of aberrant metaphases was markedlyincreased. In the Comet assay with mouse lymphoma cells,evidence of extensive DNA breakage was obtained. Allthree compounds showed similar potencies in the Cometand Ames assays while there was a clear gradation ofpotencies in the CA assay (Lomefloxacin>Fleroxacin>Ciprofloxacin), which conformed with reports on the relat-ive potencies regarding phototoxicity. The oxygen radicalscavengers catalase, superoxide dismutase and Nfl'-dimethylurea modulated the photoclastogenicity and photo-toxicity but had no influence on the effects in the Cometand Ames tests. It thus appears that different kinds ofmechanism are responsible for toxicity and clastogenicityon the one side and DNA breakage and gene mutation onthe other side. Pre-irradiation of the test articles did notlead to enhanced genotoxicity, indicating the involvementof very short lived genotoxic agents. The results endorsethe advice to avoid excessive light exposure during antibiotictherapy with fluoroquinolones.

IntroductionThe fluoroquinolones are second-generation antibiotics derivedfrom nalidixic acid. They are employed very successfully inthe treatment of a broad spectrum of infectious diseases.

Fluoroquinolones are potent mutagens in bacteria. Theiractivity depends on interaction with the bacterial gyrase(topoisomerase II) enzyme. Effects in mammalian cells aregenerally much reduced or absent. Several reviews on themutagenicity and carcinogenicity of fluoroquinolones andmammalian topoisomerase-interactive agents have been pub-lished recently (Holden et al., 1989; Fort, 1992; Andersonand Berger, 1994; Ferguson and Baguley, 1994; Albertiniet al, 1995).

In the early 1970s, nalidixic acid was reported to cause

phototoxic reactions in clinical practice. More recently thephototoxicity of the fluoroquinolones has been established ina number of experimental studies and in some clinical casereports (Wagai and Tawara, 1992; G.Klecak et al, unpublisheddata). Cutaneous photoreactions are evoked when a photosensi-tizing drug, usually absorbing in the UVA range, is present insufficient amounts in die skin and exposure to light from thesun or from artificial sources induces die formation of reactivemolecules, causing damage to cellular components. Exagger-ated sunburn, scaling and blisters may be experienced.

The photochemistry of the fluoroquinolones indicates that,upon irradiation with UVA, active oxygen species (AOS) areformed. These molecules (-02", H2O2, -OH and singlet oxygen)are generated in a variety of endogenous or exogenous pro-cesses. They are known to be capable of damaging cellularcomponents, including DNA (Levin et al., 1982; Ames et al,1993). In particular, these molecules have been shown toproduce clastogenic effects (Philips et al, 1989; Cantoni et al,1994; Youssefi et al, 1994; Duell et al, 1995). Therefore itwas of interest to investigate the photomutagenic potential ofthe fluoroquinolone antibiotics. Apart from providing evidencefor the clastogenic potential, we report on the ability of thetest articles to induce mitotic conversion in yeast, mutationsin the Ames test and DNA breakage in mouse lymphoma cellsunder concurrent irradiation with UV light An assessment ofrisk regarding photocarcinogenicity is attempted by comparingthe findings to results obtained with the exemplary photo-mutagen 8-methoxypsoralen.

Material and methods

Chemicals and irradiation sourceThe chemical structures of the three fluoroquinolones tested, together withtheir absorption spectra, are shown in Table I. The scavengers superoxidedismutase and AW-dimethylurea were obtained from Fluka (Switzerland)and catalase from Boehringer Mannheim (Germany).

The SUNTEST CPS accelerated exposure machine (Heraeus, Germany)has an emission spectrum between 290 and 800 nm. As the radiation sourcethe machine contains a xenon arc with special filter devices so that the spectralcomposition simulates solar radiation. In order to reduce the intrinsic mutagenicaction of the irradiation we passed the light through a 3 mm glass windowpane (Ames test) or through the covers of the plastic cell containers [yeast,chromosome aberration (CA) and Comet tests], which completely or partiallyremoves the UVB component. For detailed emission spectra and UV dosedeterminations, see Chitelat et al. (1993a). The exposure rates (UVA) wereset to 0.5 mW/cm2 for the Ames and Comet assays, 0.4 mW/cm2 for the CAassay and 0.65 mW/cm2 for the yeast assay, as measured with an RM2 UV-meter (Dr Gr6bel GmbH, Germany).

Photomutagenicity testsAmes test. The Salmonella tester strains TA100, TA102 and TA104 wereobtained from B.N.Ames (University of California, Berkeley, CA) and aredescribed elsewhere (Ames el al., 1975; Levin et al., 1982; Maron and Ames,1983). Overnight cultures (20 ml) were inoculated from frozen aliquotscontaining 10% dimethylsulphoxide. Strain TA102 was grown with 0.3 ngtetracycline per ml nutrient broth (NB) medium to ensure an adequate copynumber of the pAQl plasmid (Albertini and Gocke, 1988). The overnightcultures were centrifuged and resuspended in an equal volume of coldphosphate buffer. For irradiation 1 ml of the cell suspension was pipetted intoplastic Petri dishes (diameter 5 cm). Phosphate buffer, test compound and

© UK Environmental Mutagen Society/Oxford University Press 1996 497

by guest on Decem

ber 15, 2015http://m

utage.oxfordjournals.org/D

ownloaded from

A.-A.CMtelat, S.AIbertini and E.Gocke

Table I. Structures and absorption spectra of fluoroquinolones tested

Fleroxacin (Quinodis)

6,8-Difluoro-1-(2-fluoroethyl)-1,4-dihydro-7-(4-methyl-1 -piperazinyl)-4-oxo-3-quinolinecarboxylic acid

Lomefloxacini i r

180-380 nm

rac-1-Ethyl-6,8-difluoro-1,4-dihydro-7-(3-methyl-1 -piperazinyl)-4-oxo-3-quinolinecarboxylic acid

Ciprofloxacin

1-Cyclopropyl-6-fluoro-1,4-dihydro-4-oxo-7-(1 -piperazinyl)-3-quinolinecarboxylic acid

i i i r180-360 n m

~\ i r180-360 nm

scavenger solutions (where appropriate) were added to give a final volume of5 ml. During irradiation the suspensions were stirred continuously. Afterirradiation the suspensions were kept at room temperature (22°C) in the darkto reach a total incubation time of the cells with the fluoroquinolones of ~60min. Thereafter 1.5 ml of the irradiation mix Was transferred to Eppendorfincubation tubes centnfuged in a Sigma 2MK centrifuge at 10 000 r.p.m. for5 min, washed once and resuspended in I ml phosphate buffer. Next. 300 Jilof cell suspension. 200 (il of his/bio solution and 100 j_il of NB medium werepipetted onto each plate The further handling of the plates was as for thestandard plate incorporation method (Maron and Ames. 1983) Positivecontrols were sodium azide (TA100), MMC (TA102) or methylglyoxal (TA104)

in the absence of irradiation, and 8-methoxypsoralen (8-MOP) (TA102) inits presence

Saccharomyccs ctrevisiat D7 test. Cells were obtained from FK.Zimmermann(TU. Darmstadt. Germany) Photomutagenicity tests were performed asdescribed elsewhere (Chetelat et al, 1993a). Briefly, yeast cells were irradiatedin tissue culture clusters (24 wells from COSTAR. Cambridge. MA) Incubationsamples were prepared with 980 uj of cell suspension (in potassium phosphatebuffer) and 20 |il of test substance, reference substance, solvent or placebo.The cells were immediately exposed to irradiation and stirred continuously.After treatment, aliquots of the cell suspensions were pipetted into soft agar

498

by guest on Decem

ber 15, 2015http://m

utage.oxfordjournals.org/D

ownloaded from

The photomutagenldty of fluoroquinolones

tubes and poured onto plates lacking tryptophan (detection of convertants)or after appropriate dilution onto plates of synthetic complete medium(determination of viable cells). The further handling of the plates was as forthe standard assay procedure (Zimmermann el al., 1975). Positive controlswere 4-NQO in the absence and chlorpromazine in the presence of irradiation.

CA lest. The V79 permanent cell line, originally derived from a biopsy ofChinese hamster lung, was obtajned from CCR (Rossdorf, Germany). Stocksof cells free from mycoplasma contamination (McGarrity el al., 1984) werepreserved in liquid nitrogen and the cells were regularly subcultured in tissueculture flasks at 37°C in a 95ft air, 5ft COT atmosphere. The cells from non-confluent flasks were detached using 0.25ft trypsin and a single cell suspensionwas prepared in culture medium consisting of minimal essential medium,10ft- fetal calf serum and amphotericin. Of this suspension 5 ml, containing~1.5X 104 cells, was seeded onto individual sterile glass slides in Quadripermculture dishes (Heraeus, Germany) and incubated for ~2 days. At the time oftreatment the cells reached a confluency of ~30-5Oft. The estimated cellcycle time was ~12 h. The photomutagenesis test methods are described inCh£telat el al. (1993b) Briefly, the culture medium was removed and replacedby phosphate-buffered saline (PBS) containing the solvent or serial dilutionsof the test chemicals and scavenger solutions (where appropriate). A set ofcultures were irradiated with the SUNTEST CPS exposure machine. Othercultures were kept at room temperature in the dark. At the end of the treatmentall cultures were washed twice with PBS and incubated in culture mediumfor another 18 h, until chromosome preparation. Metaphases with chromosomenumbers between 21 and 23 were analysed for aberrations. As positive controlchlorpromazine, in presence of irradiation, was used.

Single cell gel electophoresis (Comet) assay. Mouse lymphoma L5178Y tk+'~cells were grown up to a density of 4x 105 cells/ml. After centnfugation theywere resuspended in Isoton II buffer (Coulter), preincubated with the testcompound for 20 min (room temperature) and irradiated in 6-well plates (-106

cells/1.9 ml). After irradiation cells were centnfuged and resuspended inRPMI 1640 medium; 50 nl (2X 105 cells) was then added to 950 JJ.1 of lowmelting agarose (0.5%), and 50 u.1 of this suspension was pipetted onto a22X26X4 mm slab of 2% normal agarose, in which was embedded a plasticholder. The samples were incubated for 1 h in lysis buffer (1 % lauroylsarcosine,2.5 M NaCl, 50 mM EDTA, 100 mM NaOH, 10 mM Tris, pH 10, 1% TritonX-100), washed and transferred into the electrophoresis buffer (1 mM EDTA,300 mM NaOH). After equilibration for 20 min electrophoresis was run for20 min at 25 V in a horizontal chamber (length 32.5 cm). The samples werethen washed and placed into staining/rcnaturation solution (400 mM Tris,pH 7.5) for 5 min. The slabs were washed, excessive water was drained byplacing the comer on a paper towel, then 50 u.1 of ethidium bromide solution(20 ng/ml) was pipetted onto the agar and after 5 min a coverslip was added.Cells were classified into four categories with the aid of a Zeiss Axiophotmicroscope under 400-fold magnification: class 1: cells without any or withonly a few (<10) trailing fluorescing particles; class 2; cells with a thintrailing streak of particles; class 3: cells with strong comet; and class 4:comets with decomposed nucleus. Positive controls were H2O2 in the absenceand chlorpromazine in the presence of irradiation.

Assay with pre-irradiated samplesAmes test. The fluoroquinolone solutions were pipetted into phosphate bufferand irradiated in the plastic Petri dishes (volume 4 ml). After irradiation, 1.2ml of the solutions were immediately added to 300 (il of cell suspension inEppendorf incubation tubes.

CA test. The fluoroquinolone solutions (5 ml) were pre-irradiated in theQuadriperm dishes. Immediately afterwards samples were pipetted into newQuadriperm dishes containing the V79 slide cultures.

For both assays, exposure concentrations of the cells to the test chemicals,incubation times and further handling were analogous to the photomutagenesisprocedure described above.

Statistical analysesFor the chromosomal aberration tests statistical significance was calculatedusing the Fisher exact test (Richardson et al. 1989); probability values <5%were accepted as significant For the other tests no statistics were performed;however, results were shown to be reproducible in at least one, generallyseveral, repeat assays.

ResultsBacterial assays (Ames test)The spontaneous level of mutant colonies of strain TA104 is~470 colonies per plate. Control experiments showed that

exposure to solar simulated light, filtered through a glass paneto cut off the UVB component, caused at most a marginalincrease, if any. Exposure to Fleroxacin and Lomefloxacin inthe absence of irradiation had a more definite but still weakeffect (maximal increase to ~700 colonies per plate). Cipro-floxacin was more toxic, which did not allow for the recognitionof this 'dark' effect at the concentrations shown here.

The cells were exposed to the three fluoroquinolones atconcentrations of 10, 31.6 and 100 Jig/ml and concomitantlyirradiated with UVA doses of 90, 300 and 600 mJ/cm2 (3, 10and 20 min of irradiation). Under these conditions all threecompounds caused increasing numbers of mutant coloniescompared with the respective controls without test compoundand/or without irradiation (Figure 1). Concomitant irradiationincreased the mutant values to maximally 927 for Fleroxacin(100 fig/ml, 600 mJ/cm2), 1159 for Lomefloxacin (100 Hg/ml,600 mJ/cm2) and 1063 for Ciprofloxacin (10 Hg/ml, 300 mJ/cm2; toxicity reduced the mutant values at higher Ciprofloxacinand UV doses).

The three fluoroquinolones caused marginal photomutageniceffects in strain TA100 (data not shown). The mutant valueswere increased by a factor of <2. Ciprofloxacin was alsostudied in strain TA102. Because of the intrinsic genotoxicityof the gyrase inhibitors in this strain only concentrations ofup to 0.32 |ig/plate were usable. No additional increase of themutant numbers was induced by the concomitant light exposure(data not shown), which provides evidence that UV exposuredoes not enhance the gyrase-mediated genotoxicty (by enhan-cing the formation or the stability of the cleavable complex).

We evaluated whether the mutagenic action of the fluoro-quinolone photoproducts in strain TA104 could be modulatedby addition of scavengers (catalase, which inactivates H2O2,MTV'-dimethylthiourea, which acts predominantly as OH scav-enger, and superoxide dismutase, which converts O2~ to H2O2).The experiments were performed with Ciprofloxacin andLomefloxacin because they had caused more definite photomu-tagenic effects in the Ames assay. However, as seen in Figure2, there was no difference in the effect when the threescavengers were used.

Lower eukaryotic system (Saccharomyces cerevisiae D7)For the evaluation of the photomutagenic potential of the threefluoroquinolones, gene conversion was selected as the geneticendpoint. Earlier investigations with chlorpromazine and 8-MOP had shown that conversion is an appropriate endpointfor measuring photomutagenic effects (Che'telat et al., 1993a).Cells were treated with different concentrations of the threefluoroquinolones (50, 200 and 400 Hg/ml) and UV doses of3800/75 and 7600/150 mJ/cm2 (UVA/UVB component). Thedata indicate induction of gene conversion at the trp5 locusby UV light alone (a factor of 3—4 above non-irradiatedcontrol). Addition of the fluoroquinolones did not increase thefrequencies of convertants above the corresponding irradiatedcontrols (Table II).

CA testExposure to the fluoroquinolones and concomitant irradiationwith UVA doses of up to 500 mJ/cm2 produced a clearconcentration-related increase in the frequency of aberrantcells as well as of the number of aberrations per cell in thedose range evaluated (6-200 (Jg/ml at a UVA dose of 500mJ/cm2; Figure 3). Statistically significant increases in thefrequency of aberrant cells were seen at concentrations of 6.5jig/ml for Lomefloxacin and 13 |ig/ml for the other two

499

by guest on Decem

ber 15, 2015http://m

utage.oxfordjournals.org/D

ownloaded from

A.-A.CWtelat, S.AIbertini and E.Gocke

Photo-Ames (TA104) with three fluoroquinolones

1200

1000

UVAV

mJ/cm

Fig. 1. Photomutagenicity results of Fleroxacin (Hero), Lomefloxacin (lome) and Ciprofloxacin (cipro)' induction of gene mutation in Salmonella typhimunumTA104. Cells were exposed to the three fluoroquinolones at concentrations of 10, 31.6 and 100 Hg/ml and concurrently irradiated with UVA doses of 0, 90,300 and 600 mJ/cm2

compounds. The concentrations at which more than oneaberration per cell was observed were 25 (ig/ml for Lomeflox-acin, 100 |ig/ml for Fleroxacin and 200 H-g/rnl for Ciprofloxacin.For Fleroxacin and Lomefloxacin the frequency of aberrantmetaphases decreased at high concentrations probably due tocell cycle delay.

We evaluated whether the mutagenic action of the fluoro-quinolone photoproducts could be modulated by addition ofscavenger compounds such as catalase, DMTU or SOD. Theexperiments were performed at concentrations of 13 and 25M-g/ml of each of the fluoroquinolones. All three scavengers,at all concentrations, significantly reduced the frequency ofaberrant metaphases as well as the number of aberrations percell. The effect was more apparent for Ciprofloxacin andFleroxacin at relatively lower aberration frequencies than forLomefloxacin at a high level of damage where multipleaberrations are more common (Figure 4).

Comet assayIn the absence of UV irradiation, the fluoroquinolones did notinduce DNA damage in mouse lymphoma cells as visible inthe Comet assay (Figure 5A). When the cells were concurrentlyirradiated, a dose-dependent shift to cells with more andmore extensive comets was apparent (Figure 5B shows arepresentative experiment with Fleroxacin). Comet length wasnot a good parameter for quantification of the damage, so weclassified the cells into the four categories described in Mat-erials and methods. UV irradiation alone caused only a marginalshift from class 1 to class 2; at Fleroxacin concentrations of3 and 10 \ig/m\ most cells were in class 2, at 30 |ig/ml mostcells were in class 3 and at 100 u:g/ml most cells were in class4. Data on the relative potencies of the three test compoundsin this test did not conform with the results obtained in theCA assay, but rather with the results in the Ames test. InFigure 6, the fraction of heavily damaged cells (classes 3 and

CATSOD *"

concentration (ug/ml)

irrad.unirrad.

DMTU

Fig. 2. Effect of the three scavengers CAT, SOD and DMTU on thephotomutagenic activity of Ciprofloxacin (10 Ug/ml; UVA 600 mJ/cm2) instrain TA104

4) is plotted as a function of dose. Fleroxacin seems toproduce slightly less photo-damage than either Ciprofloxacinor Lomefloxacin.

The scavengers CAT, SOD and DMTU did not modulatethe effects in the Comet assay, as shown in Figure 7. Noobvious change in the fraction of heavily damaged cellscompared with the level in absence of scavengers was seen.

P're-irradiation of the fluoroquinolonesSolutions of the three fluoroquinolones were irradiated with aUVA dose of 600 mJ/cm2 (for the Ames test) or 200 mJ/cm2

(for the CA test) and immediately afterwards added to thebacteria (strain TA104) or the V79 cells. No increases in thenumber of revertant colonies or chromosomal aberrations wereapparent (data not shown) when compared with the non-irradiated samples. This observation indicates that the geno-toxic agents are very short lived, as expected for AOS.

500

by guest on Decem

ber 15, 2015http://m

utage.oxfordjournals.org/D

ownloaded from

The photomutageniclty of fluoroqninolones

Table II. EffectsCiprofloxacin

UVA/UVB(mJ/cm2)

Cone, (ug/ml)

Fleroxacin0 (Plac.)0 (H2O)50200400

Chlorpro.3050

Lomeflox.0 (Plac.)0 (H,O)50200400

Chlorpro.3050

Ciproflox.0 (H2O)50200400

Chlorpro.3050

of SUNTEST CPS

0/0

*£ surv.

100.00102.06100.88101.4796.17

88.7980.24

100.00102.06101.1897.6495.28

88.7980.24

100.0097.2599.3196.11

91.0885.58

irradiation

0/0

Conv.

7.086.656.438.727.06

7.646.99

7.086.657.006.958.36

7.646.99

6.186.597.838.10

7.297.75

on Saccharomyces

250/5

tt surv.

99.41103.54

81.7140.71

99.41103.54

81.7140.71

104.35

91.5346.68

cerevisiae D7

250/5

Conv.

8.617.41

25.9961.59

8.617.41

25.9961.59

6.80

25.7580.39

in the absence and presence

3800/75

Sc surv.

89.0999.7189.3891.7497.05

89.0999.7194.99

102.9593.81

88.5697.25

100.2383.07

of Fleroxacin,

3800/75

Conv.

19.5414.5015.5112.2210.03

19.5414.5014.2914.0415.09

13.7012.4713.2417.63

Quinodis, Lomefloxacin and

7600/150

'He surv.

87.6188.4992.3386.1471.09

87.6188.4992.92

100.5988.79

86.0475.9776.8977.80

7600/150

Conv.

27.9518.0022.0416.4417.84

27.9518.0020.0014.9617.94

18.6219.5823.5129.41

Percentage of survival on synthetic complete medium and frequency of colonies grown on trp-less medium per 10° cells (convertants) calculated from thevalues shown in the Appendix Tables. Abbreviations: surv.: survival; Conv.: convertant; Cone.: concentration; Chlorpro.: chlorpromazine; Plac.: placebo.

Discussion

The phototoxic action of fluoroquinolones has been attributedto the UVA-induced production of AOS (Wagai and Tawara,1992). These molecules (O2-, H2O2, OH and singlet oxygen)are known to be damaging to cellular components, includingDNA (e.g. Ames et al, 1993). Of the standard Ames tester-strain set, TA100 is known to be weakly responsive to themutagenic activity of AOS (Epe et al., 1989). TA102 combinessensitivity towards oxygen radicals with a responsivenesstowards crosslinking agents (Levin et al, 1982). The non-standard tester strain TA104, which was developed as a'precursor' in the construction of TA102, is particularly sensit-ive towards AOS (Levin et al, 1982).

Strain TA102 is normally our first choice in photomutagen-esis experiments (Chdtelat etai, 1993a). In the present context,however, we expected that this strain would not be veryinformative because interaction of the fluoroquinolones withthe gyrase leads to a marked increase in mutant colony numbersin the absence of light already at very low concentrations(submicrogram range, Albertini et al., 1995). This effect wouldbe likely to mask any 'additional' photomutagenic activitycaused by a different mode of action. In order to assess thepossibility that irradiation would actually enhance the gyrase-mediated effect (i.e. by enhancing the formation or stabilityof the so-called 'cleavable complex') we performed a controlphotomutagenicity test with TA102, which, however, showed

no radiation enhancement of the 'dark' mutagenicity of theantibiotics. The mutagenic effect of the gyrase inhibitors isdependent on an intact excision repair pathway (Levin et al.,1982; Ysern et al, 1990; Gocke, 1991; Albertini et al, 1995)which is present in strain TA102 but not in the other Amesstandard tester strains. Oxidative mutagens do not require theexcision repair process. The repair-deficient strain TA104 hasbeen shown to be at least as sensitive to AOS (Levin et al,1982) as strain TA102, and we expected it not to be responsiveto the 'dark' mutagenicity of the fluoroquinolones.

In fact, the gyrase inhibitors weakly increased the numberof mutant colonies in strain TA104 in absence of UV light.However, the effect was much reduced compared with theactivity in TA102. With concomitant irradiation, the mutantfrequencies were enhanced in strains TA104 and TA100.Compared with the photoclastogenicity, the photogenotoxiceffect in these Salmonella strains was, however, small.

No photogenotoxic activity of the quinolones was observedin the S.cerevisiae D7 gene conversion assay. Treatment withhydrogen peroxide has been shown to induce gene conversionin strain D7M1 of S.cerevisiae (Kaneko et al, 1988). Morerecently Brennan and co-workers reported that AOS signific-antly increase the frequency of interchromosomal recombina-tion in S.cerevisiae (Brennan et al, 1994). The sensitivity ofthe assay towards the oxidative agent is much lower whencompared, for example, with the clastogenicity tests: Kanekoet al reported a sizable increase in gene conversion at 50 mM

501

by guest on Decem

ber 15, 2015http://m

utage.oxfordjournals.org/D

ownloaded from

A.-A.Ch6telat, S.AIbertinl and E.Gocke

20 40 60 Rri 100 1 ?r i40 160 'BO 20'i

concentration uq/ml

Fig. 3. Photoclastogenicity results of Fleroxacin, Lomefloxacin andCiprofloxacin: induction of structural chromosome aberrations in V79 cells.Cells were exposed to the three fluoroquinolones at concentrations from 6 to200 Hg/ml and concomitantly irradiated with a UVA doses of 500 mJ/cm2.

lome 13ug/ml

(lero 25ug/ml

cipro 25ug/ml

Fig. 4. Modulation of the photoclastogenic activity of the fluoroquinolonesby scavengers (CAT 15 ng/ml, 1000 U/ml; SOD. I mg/ml, 500 U/ml;DMTU: 104 |ig/ml. 1 mM). Frequency of aberrant cells after treatment withFleroxacin, Ciprofloxacin and Lomefloxacin and irradiation with an UVAdose of 500 mJ/cm2.

of H2O2 while we recently observed clastogenicity at 0.025 mMof H2O2 (B.Miller, unpublished results).

Compared with the two microbial test systems, the chromo-somal aberration test with V79 cells and the Comet assay withmouse lymphoma cells proved to be exquisitely sensitivetowards the photogenotoxic activity of the fluoroquinolones.Since oxidative mutagens are known to act predominantly asclastogens by the introduction of DNA single strand breaks,this is in line with expectations.

The modulating activity of several scavengers further sup-ports the hypothesis that the photoclastogenic effects can beattributed to AOS generated by irradiation of the fluoro-quinolones and does not involve the topoisomerase-mediatedpathway of genotoxicity of these compounds.

In analogy to the reduction of phototoxic effects (Wagaiand Tawara, 1992), we found that the enzyme catalase, whichinactivates H2O2, and the OH scavenger, W,//'-dimethylurea,reduced the photoclastogenic effect of irradiated fluoro-

quinolones. In our experiments, we also observed a protectiveeffect by superoxide dismutase (SOD), which converts O2 toH2O2. This is in contrast to the study on phototoxicity of thefluoroquinolones, where SOD was reported to increase toxicity(Wagai and Tawara, 1992), and the findings of Duell et al.(1995), who report an enhancement of the clastogenicity ofthe xanthine/xanthine oxidase system by SOD. It is, however,in line with the findings of Emerit and others (see Emerit,1994), who observed protection by SOD against the clastogenicaction of a variety of oxidative agents, including, for example,the xanthine/xanthine oxidase system. Apparently, some as yetundefined culture conditions are responsible for the divergenteffects of SOD, and additional factors such as adaptive responseof the cell lines and pre-existing resistance to, for example,H2O2 oxidative stress in V79 variants may also play a role(Cantoni et al., 1994). The observation that the exogenouslyadded enzymes CAT and SOD, which do not enter the cells,are protective against the photoclastogenic effects of thefluoroquinolones suggests that the reactive AOS responsiblefor the effects are either long lived enough to travel from themedium to the nucleus or that indirect effects, e.g. damage tothe membranes, are mediating the clastogenicity. Formationof secondary clastogenic material of cellular origin, calledclastogenic factors (CF), has been implicated in the clastogenic-ity of a wide variety of processes causing oxidative stress(Emerit, 1994). It appears that CFs are chemically veryheterogeneous. Degradation products of arachidonic acid maybe candidates, but at present the whole process is still rathermysterious.

Our observation that the scavengers were not able to protectagainst the photomutagenicity of the fluoroquinolones in theAmes and Comet assays is most easily explained by attributingthe effects observed in these two test systems to reactiveproducts generated in the immediate vicinity of the DNA sothat there is no chance for the scavengers to intercept themolecules. More puzzling is the observation that in the Amesand Comet assays the three fluoroquinolones appear to beabout equipotent (with Fleroxacin possibly slightly less active),while in the CA test and in the phototoxicity tests there is aclear gradation, with Lomefloxacin being the most potent,followed by Fleroxacin and then Ciprofloxacin. The photo-chemical reactivity of Lomefloxacin is clearly higher than thatof Ciprofloxacin, with Fleroxacin somewhere in the middle(K.H.Pfoertner, unpublished data). Thus, one would expect aclear gradation of activity independent of whether the respons-ible AOS were generated close to the DNA or at a greaterdistance. The discrepancy might be explained by the assump-tion that gene mutations and DNA breakage are caused by asubtype of reaction products which are generated in equalamounts from the three compounds or, alternatively, it mightbe that the three quinolones are present within the cells atdifferent concentrations, counterbalancing the different photo-reactivities.

While the underlying mechanisms are not yet clear, theobservations might suggest that the type of damage measuredin the Comet assay is not responsible for the toxic andclastogenic effects, but could be responsible for the gene-mutation effect.

Several studies on the photocarcinogenesis of fluoro-quinolones have been reported. UV doses which, in combina-tion with the drugs, produced severe, dose-dependent toxicityled to development of skin tumours in Swiss and Skh-1mice (Johnson et al.. 1989). A comprehensive study with

502

by guest on Decem

ber 15, 2015http://m

utage.oxfordjournals.org/D

ownloaded from

The photomutagenldty of fluoroqulnolones

100concentration (ug/ml)

concentration (ug/ml)

Fig. 5. Results of the Comet assay with Fleroxacin in mouse lymphomacells without (A) and with concurrent irradiation with 500 rrJ/cm2 (B).Classification: class 1: cells without any or with only a few (<10) trailingfluorescing particles; class 2: cells with a thin trailing streak of particles;class 3: cells with strong comet; and class 4: comets with decomposednucleus.

concentration (ug/ml) 100

Fig. 6. Results of the Comet assay with Fleroxacin, Lomefloxacin andCiprofloxacin at a UVA dose of 500 mJ/cm2. The fraction of cells in classes3 plus 4 is given.

subphototoxic dose levels was performed (Urbach et al, 1995;G.Klecak, unpublished data) with 8-MOP and a number offluoroquinolones, including those studied in the experiments

0>

u&aa•a

none

cipro. 50 ug/ml

lome, 10 ug/ml

flero, 10 ug/mlCAT

SOD DMTU

Fig. 7. Modulation of the photogenotoxicity of the three fluoroquinolones inthe Comet assay. The fraction of cells in classes 3 plus 4 is given (CAT: 15ug/ml, 1000 U/ml; SOD: 0.6 mg/ml, 300 U/ml; DMTU: 104 ug/ml, 1 mM).

presented here. All fluoroquinolones studied were capableof enhancing UVA-induced phototumorigenesis. Almost alltumours were benign except in the 8-MOP-positive controls(which received markedly lower UV doses) and the Lomeflox-acin-treated mice. It is notable that median latent periodsincreased in the order Lomefloxacin—»Fleroxacin—»Cipro-floxacin. Thus, relative potencies parallel the relativephototoxic/photoclastogenic potencies rather than the relativeDNA-damaging and gene mutation-inducing potencies, asobserved in the Comet and the Ames test. It appears that thefluoroquinolones are compounds which might be useful toinvestigate the relative contributions of phototoxicity—withsubsequent enhanced cell proliferation—and DNA damage andmutation to the processes of skin tumour formation.

Conclusion

With the prior knowledge of the phototoxic mechanisms thepositive responses in the photomutagenicity tests were to beexpected. AOS are produced in many exogenous cellularprocesses as well as by exogenous agents. Cells have developeda multitude of defence mechanisms against oxidative damage,including accumulation of glutathione, ascorbate, tocopherolsand carotinoids (Ames et al, 1993). These defences are partlymissing in established cell lines cultured in vitro. But even inthe whole animal, antioxidant defences are not perfect. It isestimated that the genetic material of a human cell experiencesup to 10 000 oxidative hits per day. DNA repair enzymesremove most but not all of the lesions formed. Apoptosis ofskin cells (i.e. sunburn)—as would potentially be induced bythe phototoxic action of the fluoroquinolones—represents afurther level of defence.

Nevertheless, whether human antibiotic therapy with thefluoroquinolones could represent a significant additional factorin the formation of skin cancer has to be discussed. Clearly,it is advisable that humans treated with these antibiotics shouldavoid extensive exposure to sunlight or artifical UVA light

In the context of risk assessment, it is of interest to considerthe health risk associated with PUVA treatment, which consistsof exposure to 8-MOP in combination with extensive UVAirradiation. 8-MOP is the photomutagenic compound parexcellence: the genetic lesions involve cross-links as well asoxidative damage. Photoclastogenicity of 8-MOP plus UVAappears to be largely due to AOS (Youssefi et al, 1994). In

503

by guest on Decem

ber 15, 2015http://m

utage.oxfordjournals.org/D

ownloaded from

A.-A.CWtelat, S-Albertini and E.Gocke

the toxicological comparison with the fluoroquinolones thePUVA case would, therefore, represent a very extreme 'worstcase' scenario.

Despite the strong experimental evidence for the photomuta-genicity of 8-MOP, no unequivocal indication for an increasedrisk of human skin cancer was apparent in a multitude ofepidemilogical studies. Those patients who developed skincancer were exposed to large doses of UVA, administered overmany years, and other risk factors also existed (Stern, 1989;Studniberg and Weller, 1993). In those cases where tumoursdeveloped, the patients had received at least 5 years ofPUVA treatment. On this basis, treatment with fluoroquinoloneantibiotics is not expected to pose an appreciable risk, evenwhen simple precautions against excessive light exposure arenot taken, in which case the patient will experience repeatedlythe uncomfortable immediate reaction of sunburn. But light-induced adverse effects will not occur in the first place ifsimple precautions are taken during the relatively short periodsof antibiotic therapy with the fluoroquinolones (at most afew weeks).

AcknowledgementsWe are indebted to Ms B Bartsch. Ms M -E Brun. Ms J Muller-Freudenreich.Mr G.MUller and Ms A.Wespieser for their excellent technical assistance

ReferencesAlbertini.S. and Gocke.E. (1988) Plasmidcopy number and mutant frequencies

in S.lyphimurium TAI02. Environ Mol. Mutagen., 12, 353-363.Albertini.S., Chetelat,A.-A.. Miller.B , Muster.W.. Pudjadas.E.. Strobel.R

and Gocke.E (1995) Genotoxicity of 17 gyrase- and four mammaliantopoisomerase II-poisons in prokaryotic and eukaryotic test systems.Mutagenesis, 10, 343-351.

Ames.B.N , Shigenaga,M K and Hagen.T.M (1993) Oxidants. antioxidants.and the degenarative diseases of aging Proc. Natl Acad. Sci. USA, 90.7915-7922.

Ames.B.N., McCannJ and Yamasaki.E (1975) Methods for detectingcarcinogens and mutagens with the Sa/monW/a/mammalian-microsomemutagenicity test. Mutat Res, 31, 347-364

Anderson,R.D. and Berger.N.A (1994) Mutagenicity and carcinogenicity oftopoisomerase-interactive agents. Mulat. Res. 309, 109-142

Brennan.RJ., Swoboda,B.E.P. and Schiestl.R.H. (1994) Oxidative mutagensinduce intrachromosomal recombination in yeast. Mutai. Res.. 308. 159-167.

Cantoni.O , Guidarelh^A., Sestili,P, Manello.F, Gazzanelli.G. and Cantabeni.F(1994) Hydrogen peroxide cytoxicity under conditions of normal or reducedcatalase activity in h^O^-sensitive and -resistent Chinese hamster ovary(CHO) cell variants. Toxicol. Lett, 73, 193-199.

Ch£telat,A., Albertini.S., DrespJ.H . Strobel.R. and Gocke.E (1993a) Photo-mutagenesis test development I. 8-Methoxypsora]en. chlorpromazine andsunscreen compounds in bacteria and yeast assays Mutat. Res., 292,241-250.

Chitelat.A., DrespJ.H. and Gocke.E (1993b) Photomutagenesis testdevelopment: II. 8-Methoxypsoralen, chlorpromazine and sunscreencompounds in chromosomal aberration assays using CHO cells Mutat.Res.. 292. 251-258

Duell.T, Lengsfelder.E , Fink.R.. Giesen.R and Bauchinger.M (1995) Effectof activated oxygen species in human lymphocytes. Mutat Res. 336. 29-38

Ement.I. (1994) Reactive oxygen species, chromosome mutations, and cancer.possible role of clastogenic factors in carcinogenesis. Free Radical Biol.Med.. 16. 99-109.

Epe.B., HeglerJ. and Wild.D. (1989) Singlet oxygen as an ultimately reactivespecies in Salmonella typhimurium DNA damage induced by methyleneblue/visible light Carcinogenesis. 10. 2019-2024

Ferguson.L.R. and Baguley.B.C. (1994) Topoisomerase II enzymes andmutagenicity. Environ. Mol. Mutagen., 24. 245—261.

Fort.F.L (1992) Mutagenicity of quinolone antibactenals. Drug Safety. 7,214-222.

Gocke.E. (1991) Mechanism of quinolone mutagenicity in bacteria Mutat.Res., 248. 135-143.

Holden.H.E. BarettJ.F. Huntington.C M.. Muehlebauer.PA andWahrenburg.M.G. (1989) Genetic profile of a nalidixic acid analogue: amodel for the mechanism of sister chromatid exchange induction Environ.Mol. Mutagen.. 13. 238-252.

504

Johnson.B.E. Walker.E M. and Ferguson J. (1989) Quinolone-mducedphotosensitivity. Rev Infect Dis, 11, S1396-1397

Kaneko.M , Leadon.S.A. and Ito.T. (1988) Relationship between the inductionof mitotic conversion and the formation of thymine glycols in yeastSaccharomyces treated with hydrogen peroxide. Mutat. Res., 207. 17-22

Levin,D.E., Hollstein.M.C . Christman.M.F., Schwiers.E.A. and Ames.B N.(1982) A new Salmonella tester strain (TAI02) with A:T base pairs at thesite of mutation detects oxidative mutagens. Proc Natl Acad Sci USA. 79.7445-7449

Maron.D.M. and Ames,B N. (1983) Revised methods for the Salmonellamutagenicity lest. Mutat. Res.. 113. 173—215

McGamty.G J.. Vanarran.V and SaramaJ (1984) Cytogenetic effects ofmycoplasmal infection of cell cultures' a review. In Vitro. 20. 1-18

Phihps.BJ., Caroll.PA., Tee.A C. and Anderson.D. (1989) Microsome-mediated clastogenicity of butylated hydroxyanisole (BHA) in culturedChinese hamster ovary cells: the possible role of reactive oxygen speciesMutat Res., 214, 105-114

Richardson.C, Williams.D A , Amphlett.G., Philhps.B , Allen.J.A. andChanter.D 0 (1989) Analysis of data from in vitro cytogenetic assays InKirkland,D J (ed.), Statistical Evaluation of Mutagenicity Data. CambridgeUniversity Press, Cambridge, pp 141-154.

Stem.R. (1989) PUVA. its status in the United States In Psoralens.T.B.F(ed.) John Libbey Eurotext, Pans.

Studniberg.H.M. and Weller.P. (1993) Photosensitized lipid peroxidation bycinoxacin and its photoproducts. lnvolvment of a derived peroxide in itsphototoxicity J. Am. Acad. Dermutol., 29, 1013-1022

Urbach.F (1995) Drug photosensitized skin tumorigenesis Abstract. 6thCongress of the European Society for Photobiology Cambridge. UK.3-8.9 1995

Wagai.N., and Tawara.K (1992) Possible direct role of reactive oxygens inthe cause of cutaneous phototoxicity induced by five quinolones in miceArch. Toxicol., 66. 392-397

Youssefi.A A , Arutynyan.R and Ement.I. (1994) Chromosome damage inPUVA-treated human lymphocytes is related to active oxygen species andclastogenic factors Mutat. Res.. 309. 185-191.

Ysem.P, Clerch.B , Castano.M., Gilbert.I., Barbe.l. and Llagostem.M (1990)Induction of SOS genes in Escherichia coli and mutagenesis in Salmonellatyphimurium by fluoroquinolones. Mutagenesis 5, 63-66.

Zimmermann.FK , Kem.R. and Rasenberger.H (1975) A yeast strain forsimultaneous detection of induced mitotic crossing over, mitotic geneconversion and reverse mutation. Mutat. Res , 28, 381-388.

Received on February 20, 1996. accepted on April I, 1996

by guest on Decem

ber 15, 2015http://m

utage.oxfordjournals.org/D

ownloaded from