Aquatic Toxicology 79 (2006) 65–77

Genotoxicant accumulation and cellular defence activation in bivalveschronically exposed to waterborne contaminants from the Seine River

B. Rocher a, J. Le Goff c, L. Peluhet b, M. Briand c, H. Manduzio a, J. Gallois d, M.H. Devier b,O. Geffard b, L. Gricourt a, S. Augagneur b, H. Budzinski b, D. Pottier c,

V. Andre c, P. Lebailly c, J. Cachot a,∗a Laboratory of Ecotoxicology (LEMA), UPRES-EA 3222, IFRMP 23, University of Le Havre, 25 rue Philippe Lebon,

B.P. 540, 76058 Le Havre Cedex, Franceb LPTC UMR-5472 CNRS, University of Bordeaux I, Bordeaux, France

c GRECAN EA-1772, University of Caen-Basse Normandie and CLCC F. Baclesse, Caen, Franced Laboratoire Departemental Franck Duncombe, Caen, France

Received 10 March 2006; received in revised form 15 May 2006; accepted 16 May 2006

Abstract

aege(SgiAowao©

KG

2dorfppdTf

0d

The aim of the present work was to investigate genotoxicant accumulation and biological responses of zebra mussels and blue mussels collectedlong a pollution gradient in the Seine estuary and in the Seine Bay. The sampling area included three contaminated and one reference sites forach species. The study focused on polyaromatic hydrocarbons (PAH), lindane, polychlorobiphenyls (PCB) and metals known to be potentialenotoxicants and/or reactive oxygen species (ROS) inducers. Enzymatic activities related to cellular defence systems including the phase IInzyme glutathione S-transferase (GST) and three antioxidant enzymes superoxide dismutase (SOD), catalase (CAT) and glutathion peroxydaseGPx) were measured in gills. DNA adducts and DNA strand breaks (Comet assay) were measured in digestive gland and hemocytes, respectively.pecies differences were observed in metal accumulation (As and Pb), GPx activity and DNA adduct formation. A marked upstream–downstreamradient was reported for PAH body burden and to a lesser extent for PCB and metals with the highest values measured just downstream thendustrialized area of Rouen. GST and SOD activities in gills of bivalves were positively related to PAH and metals body burden, respectively.ctivation of those cellular defences may prevent accumulation of electrophilic metabolites and free radicals and thus may protect DNA andthers macromolecules against oxidation and adduction. Although DNA strand breaks and bulky adducts were detected in both species, levelsere relatively low and no significant site differences were observed in June 2003. Our results indicate a clear relationship between genotoxicant

ccumulation and positive activation of detoxification and antioxidant systems but poor consequences in term of DNA damage for wild populationf mussels inhabiting the Seine estuary.

2006 Elsevier B.V. All rights reserved.

eywords: Blue mussel; Zebra mussel; Metals; PAH; PCB; Bulky DNA adducts; Comet assay; Cu/Zn superoxide dismutase; Glutathione peroxydase; Catalase;lutathione S-transferase

Abbreviations: CAT, catalase; CDNB, 1-chloro-2,4-dinitrobenzene; dAMP,′-deoxyadenosine 3′-monophosphate; DRZ, diagonal radioactive zone; dw,ry weight; EDTA, ethylene-diamine tetra-acetic acid; GPx, glutathione per-xidase; GRd, glutathione reductase; GST, glutathione S-transferase; GSH,educed glutathione form; GSSG, oxidized glutathione form; IEF, isoelectricocusing; NADPH, reduced nicotinamide adenine dinucleotide phosphate; PAH,olyaromatic hydrocarbons; PCB, polychlorobiphenyls; POP, persistent organicollutants; RALs, relative adduct levels; RNO, Reseau National d’Observationu milieu marin; ROS, reactive oxygen species; SOD, superoxide dismutase;BT, tri-butyltin; TLC, thin layer chromatography; SEVAG, solution of chloro-

orm and isoamyl alcohol 24/1 (v/v)∗ Corresponding author. Tel.: +33 2 32 74 43 03; fax: +33 2 32 74 43 14.

E-mail address: Jerome.Cachot@univ-lehavre.fr (J. Cachot).

1. Introduction

For decades, numerous field studies including data from theReseau National d’Observation de la qualite du mileu marin,RNO (Claisse et al., 1992) and from the NST Mussel WatchProject (O’Connor, 1998) have demonstrated the utility of usingmussels for pollution monitoring of coastal marine areas. Indeed,mussels are sedentary filter-feeder known to accumulate highlevels of certain metals and organic compounds and to exhibitcellular and physiological responses providing a time-integratedmeasurement of environmental contamination as well as inducedbiological effects (Sole et al., 1995). However most of field stud-ies focused on marine mussels or oysters or a few fresh water

166-445X/$ – see front matter © 2006 Elsevier B.V. All rights reserved.oi:10.1016/j.aquatox.2006.05.005

66 B. Rocher et al. / Aquatic Toxicology 79 (2006) 65–77

bivalve species but little information is available on bivalvesliving in estuaries. In fact, it is of great concern since (i) mostof pollutant originated from human activity are transported tothe sea via rivers, (ii) a large part of contaminants adsorbed toparticulates is trapped in the estuarine area and (iii) highly fluc-tuating environmental conditions might represent an additionalstress factor for estuarine species.

Molluscs, and bivalves in particular, possess a wide rangeof defences to prevent toxic effects of chemicals at the cellu-lar level: (i) metallothioneins and multi xenobiotic resistanceproteins that actively reduce cellular entrance of toxicants,(ii) detoxification and antioxidant systems allowing neutral-ization and elimination of parent compounds, metabolites andby-products and finally (iii) DNA repair and protein and lipidturnover which maintain cell integrity. Most of those processesare driven by inducible proteins whose expression and/or activ-ity levels have been shown to be related to the cellular contentof certain toxicants. For instance, numerous components of thedetoxification and antioxidant systems in mollusc species havebeen shown to be specifically induced by metals or PAH incontrolled laboratory conditions (Livingstone, 1994; Regoli andPrincipato, 1995; Boutet et al., 2004) or by complex mixture ofpollutants in the field (Cossu et al., 1997; Nasci et al., 1998;Gowland et al., 2002) and hence are thought to be particularlysuitable for pollution biomonitoring.

In the present study, biomarkers were selected according totssiaCwCosaTm(gTsptddhba(ocuani

also oxidative DNA lesions (8-OHdG) and bulky DNA adducts(Canova et al., 1998) were reported in digestive gland of Mytilussp. treated with benzo[a]pyrene. In addition, high levels of DNAadducts were measured in indigenous or transplanted musselsfrom industrial and urban areas (Sole et al., 1996; Ericson et al.,2002).

This study was conducted in the frame of an integrated andmultimarker field study investigating the genotoxic risk in theSeine estuary. Data concerning contamination of sediments andtheir genotoxic and embryotoxic activities were reported in thisissue (Cachot et al., this issue). The present paper focused oncontamination and biological responses of two bivalve species,Dreissena polymorpha and Mytilus edulis, living in brackish orsalt water areas of the Seine estuary. The first step aimed todetermine PAH, PCB and metal contents of bivalves throughoutthe studied area in order to highlight the most impacted sites. Ina second step, the levels of oxidative stress and DNA damagewere evaluated in mussels in order to link effects to genotoxicantexposure. Finally, an attempt was made at comparing the levelsof chemical accumulation and biological responses between thetwo bivalve species.

2. Materials and methods

2.1. Sampling of bivalves

bteopiwicaPP(a

TLS

S

YOLCVLAL

he typology of the contamination in the Seine River. Indeed,everal studies have documented the high contamination ofediments of the Seine River by a cocktail of pollutants includ-ng PAH and PCB and metal species, notably mercury and zincnd to a lesser extent copper and cadmium (Minier et al., 2006;achot et al., this issue). PAHs are well known damaging agentshich are able to elicit DNA adduct formation (Venier andanova, 1996; Akcha et al., 2000) but also to generate reactivexygen species (Mitchelmore et al., 1998) in mussels. Metalpecies like copper have also been shown to generate oxyradicalsnd oxidative DNA damage in rat hepatocytes (Li et al., 2002).he glutathione S-transferases (GST) are quantitatively theost important phase II enzymes of the detoxification system

Fitzpatrick et al., 1997). They catalyse conjugation reaction oflutathione with various organic compounds including PAH.hese enzymes also play a role in protection against oxidativetress by catalysing a selenium-independent glutathione-eroxidase activity (Prohaska, 1980). Three components ofhe antioxidant system were also investigated. The superoxideismutases (SOD) are oxido-reductases which catalyse theismutation of the superoxide anion into molecular oxygen andydrogen peroxide. The cytosolic Cu/Zn SOD have previouslyeen characterized in gills and digestive gland of the blue musselnd the acidic isoform was shown to be inducible by copperManduzio et al., 2003). Catalase (CAT) catalyses the reductionf hydrogen peroxide in water. Glutathione peroxidases (GPx)atalyse mainly the reduction of organic peroxides to alcoholssing reduced glutathione. When the first lines of defencesre overtaken, oxidative damage and adducts can occur atucleophilic sites of cellular macromolecules. Dose-dependentncrease of DNA strand breaks (Mitchelmore et al., 1998) but

Zebra mussels, Dreissena polymorpha Pallas are fresh waterivalves abundant and widely distributed all along the course ofhe Seine River (Akopian et al., 2001). Blue mussels, Mytilusdulis, are typically marine bivalves forming natural coloniesn rocky shore of the Seine Bay. A preliminary sampling cam-aign was conduced in June 2002 focusing on two highlympacted sites, La Bouille and Oissel and one reference freshater site, Yville-sur-Seine. A second campaign encompass-

ng eight sites in the Seine estuary and in the Seine Bay wasarried out in June 2003 (Table 1). Le Moulard was chosens a marine reference site because of the very low levels ofAH, CB153, organochlorinated pesticides and metals such asb, Cd, Hg and Ni reported in blue mussels from this areaRNO database). Samplings were performed from the shoret low tide. Zebra mussels were collected on rocks and pon-

able 1ocalization of sampling sites and characteristics of bivalves collected in theeine estuary and in the Seine Bay in June 2003

ite Species Length(mm)b

Weight (g)c

ville-sur-Seinea Dreissena polymorpha 25 ± 1 1.7 ± 0.2issel (left bank) Dreissena polymorpha 24 ± 2 1.9 ± 0.6a Bouille (left bank) Dreissena polymorpha 24 ± 3 2.4 ± 0.5audebec (left bank) Dreissena polymorpha 23 ± 2 1.7 ± 0.4illerville Mytilus edulis 37 ± 3 5.0 ± 1.1e Havre Mytilus edulis 45 ± 3 9.3 ± 2.9ntifer Mytilus edulis 42 ± 3 7.0 ± 1.3e Moularda Mytilus edulis 39 ± 3 6.9 ± 1.3

a Reference sites.b Mean length ± standard deviation for 30 individuals.c Mean total wet weight ± standard deviation for 30 individuals.

B. Rocher et al. / Aquatic Toxicology 79 (2006) 65–77 67

toons under the water surface while Mytilus were sampled onemerged rocks. Thirty individuals were randomly collected ateach site for length and weight measurements (Table 1). In addi-tion, about two hundreds 22–27 mm length zebra mussels andthe same number of 35–45 mm blue mussels were collected ateach station for chemical and biological analyses. Dissectionswere performed on freshly collected living animals. Gills anddigestive gland were removed and immediately deep frozen inliquid nitrogen or carbonic ice and then conserved at −80 ◦Cuntil further biological analyses. The whole soft body from 20blue mussels or 50 zebra mussels were pooled, deep frozen inliquid nitrogen and conserved at −20 ◦C for further chemicalanalyses.

2.2. Chemical analyses

Analyses were mainly focused on known environmen-tal mutagens and carcinogens including 21 PAH, 7 PCB, 1chlorinated pesticide (lindane) and 9 metal species includingchromium, cobalt, nickel, arsenic, cadmium, mercury, lead, cop-per and zinc (IARC monographs).

2.2.1. Organic analysesFreeze-dried mussels were extracted by microwave-assisted

extraction with dichloromethane (30 W; 10 min). The internalsot

eoaceTb(a

emG

2

7scA1dtmdcv

2.3. Enzymatic measurements

2.3.1. Preparation of protein extractsGills from three to four zebra mussels but from only one indi-

vidual blue mussel were homogenized in 10 mM Tris, 10 mMEDTA, 5 mM 2-� mercaptoethanol and 1% Triton X100, pH7.5, for 30 s using an electric potter. The homogenates weretwice sonicated briefly using an ultrasonic processor (Fisher-Bioblock, France) and then centrifuged at 9000 × g, at 4 ◦C for15 min. After centrifugation, supernatants were collected andused immediately for determination of enzyme activities. Pro-tein content of each resulting supernatant was measured on analiquot according to the method of Bradford using bovine serumalbumin as standard (Bradford, 1976). For each sampling site,six different protein extracts were prepared. All enzymatic activ-ities were measured at least twice from each protein extract.

2.3.2. Glutathione S-transferase activity (GST, EC 2.5.1.18)GST activity was measured at 340 nm according to the

method of Habig et al. (1974), using the 1-chloro-2,4-dinitrobenzene (CDNB) and reduced glutathione (GSH) as sub-strates (Sigma–Aldrich, France). Activities were expressed asmilliunits per milligram of proteins.

2.3.3. Catalase activity (CAT, EC 1.15.1.1)

fcp

2E

Bircpw

2

prBrmSocwgoaca

tandards were gravimetrically added prior to the extraction andne extraction blank was performed with each series of extrac-ion.

For PCB analysis (Thompson and Budzinski, 1999), thextract was then purified by shaking with sulphuric acid. Therganic phase was neutralised with deionised water, dried onnhydrous sodium sulphate and reconcentrated. A second purifi-ation step on silica micro-columns was performed. PCB wereluted with a mixture of pentane/dichloromethane (90/10, v/v).he final extract was reconcentrated in isooctane and analyzedy gas chromatography coupled to electron capture detectorGC–ECD). PCB congeners 30, 103, 155, and 198 were useds internal standards.

PAH were analyzed by gas chromatography/mass spectrom-try after purification of organic extracts on alumina and silicaicrolumns. Detailed protocol was described previously (Leoff et al., 2006).

.2.2. Inorganic analysesFreeze-dried mussels (0.5 g dw) have been digested with

0% nitric acid for 2 h at 100 ◦C. After digestion volume ofamples was completed to 25 mL with MilliQ water. Metalontent was measured by ICP-MS HP 4500 (HP 4500 Series,gilent Technologies, USA) using an external calibration. Ag L−1 multi-elemental solution was used to prepare the stan-ard solutions. The rhodium was used as an internal standardo evaluate the instrumental drift. The accuracy of the deter-

ination procedure was assessed by the analysis of a stan-ard reference material NBS SRM 1566A (oyster tissues). Theoncentrations measured were in agreement with the certifiedalues.

CAT activity was determined according to Aebi (1984) byollowing the decrease in absorbance at 240 nm due to H2O2onsumption. Results were expressed as units per milligram ofroteins.

.3.4. Se-dependent glutathione-peroxidase activity (GPX,C 1.11.1.9)

Se-dependent GPX activity was measured using theIOXYTECK GPx-340

TMkit (OxysInternational, Inc., USA),

n a coupled enzyme system where the GSSG formed in the GPXeaction was converted to the reduced GSH form by GRD. Theonsumption of NADPH was monitored at 340 nm and directlyroportional to the GPX activity in the sample. GPX activityas expressed as milliunits per milligram of proteins.

.3.5. Superoxide dismutase activity (SOD, EC 1.15.1.1)Determination of SOD activity was performed on gel after

roteins separation by isoelectric focusing (IEF). IEF was car-ied out in a Multiphor II equipment (Amersham Pharmaciaiotech, France) on gels with pH range 4.0–5.0 for a better

esolution. The pI values were determined according to IEFarkers (Amersham Pharmacia Biotech, France). Revelation ofOD activity was performed immediately after the achievementf electrophoresis using the nitrobluetetrazolium (NBT) photo-hemical method of Beauchamp and Fridovitch (1971). The gelsere first soaked in 2.5 mM NBT for 30 min, in darkness, underentle agitation. They were then incubated in a solution madef 1 M phosphate buffer (pH 7.8) containing 0.04 mM riboflavinnd 0.5% TEMED for 30 min, once again in the dark and underonstant agitation. Detection of SOD activity and isoforms waschieved by soaking gels in 10 mM phosphate buffer (pH 7.8)

68 B. Rocher et al. / Aquatic Toxicology 79 (2006) 65–77

under light condition. Achromatic bands in the gels indicatedSOD activity. The gels were scanned and subsequently analyzedusing the ImageMaster TotalLab software developed by Amer-sham Pharmacia Biotech (France). The total SOD activity is thesum of activities of each isoform and was expressed as a per-centage of the value measured at the reference site.

2.4. DNA damage

DNA strand breaks and bulky DNA adducts were measuredon different tissues coming from the same group of individuals.The alkaline comet assay (Shing et al., 1988) was performed onhemocytes in order to measure levels of DNA strand breaks andalkali labile sites. Twenty-six to 30 freshly collected musselswere individually analyzed for each sampling site. Hemolym-phe was collected from the posterior adductor muscle of livingmussels with a hypodermic syringe. Hemocytes were then con-centrated by centrifugation 10 min at 1000 rpm. Negative con-trols consisting of mononuclear leukocytes from one healthyman without any known exposure to genotoxic agents, wereprocessed together with mussel samples. The comet assay wasperformed under minimal illumination essentially as describedby Shing et al. (1988) but at room temperature with some minormodifications (Lebailly et al., 1997). Briefly, cells were sus-pended in 200–300 �L prewarmed low melting point agarose(0.5% in PBS) and put on a microscope slide prewashed with1AtEa4aEttbentwwa1rstlDadnn

b(D

previously (Le Goff et al., 2006). DNA was extracted from poolsof 6–8 digestive glands. Three different pools were prepared foreach sampling sites. Final results were expressed as the rela-tive adduct levels (RALs) according to the following relation:RAL = 1.22 × 10−6 cpmadducts/cpmdAMP.

2.5. Statistical analysis

The non-parametric Wilcoxon’s test was applied to com-pare chemical and biological data between contaminated andcorresponding reference sites. Differences were considered assignificant for P ≤ 0.05. Variations of enzymatic activities andDNA damage levels between sites were analyzed using one wayanalysis of variance (ANOVA) complemented by the Dunnett’stest. The Levene’s test was applied to check the equality ofvariances. For adducts, Log-transformation was performed tonormalized data distribution. The non-parametric Spearman’srank correlation test was applied to find out potential correla-tion between the different studied parameters. Correlation wasconsidered as significant if P ≤ 0.05.

3. Results

3.1. Contamination of biota by mutagenic and carcinogenicpollutants

nbaicCwbfgFss

wbg(PtmcrIoMi(bT

00% ethanol and precoated with normal agarose (0.8% in PBS).third layer of 0.5% LMP agarose was added. The slide was

reated with a freshly made lysis solution (2.5 M NaCl, 0.1 MDTA, 10 mM Tris adjusted to pH 10, 10% dimethyl sulfoxidend 1% Triton X-100 both freshly added) for at least one night at◦C. To allow DNA unwinding, the slide was then incubated infreshly made electrophoresis buffer (0.3 M NaOH and 1 mMDTA, pH > 13) for 40 min in order to remove all lysis solu-

ion. Finally, the slide was placed in a horizontal electrophoresisank (Hoeffer HE99 X©), immersed in fresh electrophoresisuffer, and exposed to 0.7 V cm−1 for 15 min (300 mA). Afterlectrophoresis, the slide was washed twice in freshly madeeutralization buffer (0.4 M Tris adjusted to pH 7.5 with concen-rated HCl). After drying overnight at 4 ◦C, the slide was stainedith 50 �L ethidium bromide solution (2 �g mL−1), coveredith a cover slide and observed at 200× magnification usingBX41 fluorescence microscope (Olympus) equipped with a

00 W lamp and a filter wg2 (excitation filter: 510–550 nm; bar-ier filter: 590 nm). Each mussel was tested in triplicate in theame assay and at least 200 randomly selected cells for each ofwo different slides (at least 400 cells/sample) were visually ana-yzed. Cells were eye-graded into four categories depending onNA damage level: non-damaged, slightly damaged, damaged

nd atypic nuclei. Slightly damaged and damaged nuclei wereifferentiated according to the length of the comet tail. Atypicuclei contained highly degraded DNA spread all around theuclei.

Bulky DNA adducts were measured from digestive gland ofivalves using the nuclease P1-enhanced 32P-postlabelling assayReddy and Randerath, 1986). The detailed procedure used forNA extraction and DNA adduct measurement was described

Several potentially carcinogenic metals including cobalt,ickel, arsenic, cadmium, mercury and lead were detected inivalves (Table 2). As a general trend, body burden of total met-ls appeared significantly higher (P < 0.05) in zebra mussels thann blue mussels. In fact, differences were shown according to theonsidered metal species. Mean concentrations of Co, Ni, andd were about two- to three-fold higher in the former specieshereas As and Pb were two to four times more abundant inlue mussels. For a same species, no significant site to site dif-erence was evidenced for total metal body burdens. Besides thiseneral trend, localized peaks of contamination were observed.or instance, higher mercury levels were measured in musselsampled at Le Havre in June 2003 but also during a previousampling campaign in June 2001 (data not shown).

Numerous potentially mutagenic and carcinogenic PAHere detected in bivalves (Table 3). As a general trend, PAHody burdens showed a clear upstream–downstream decreasingradient with zebra mussels containing higher levels of PAHP = 0.075) in comparison to blue mussels (Table 3). The highestAH concentrations were reported in zebra mussels collected inhe upper Seine estuary (up to 990 ng g−1 dw total PAH) but blueussels from Villerville and Le Havre, which are known to be

hronically exposed to the Seine River’s plume, still containedelatively high amount of PAH (up to 447 ng g−1 dw total PAH).n contrast, those compounds were barely detectable in bivalvesbtained from the reference sites of Yville-sur-Seine and Leoulard (Table 3). Although mutagenic PAH were measured

n the same proportions in bivalves (62–73%) and sediments76–82%), carcinogenic ones were clearly under represented inivalve tissues (9–28%) in comparison to sediments (39–47%).his feature could stem from the predominance of four-ring

B. Rocher et al. / Aquatic Toxicology 79 (2006) 65–77 69

Table 2Metals body burdens of zebra mussels and blue mussels

Compound Yville/SeineDreissena

OisselDreissena

La BouilleDreissena

CaudebecDreissena

VillervilleMytilus

Le HavreMytilus

AntiferMytilus

Le MoulardMytilus

Cr <0.5 <0.5 <0.5 <0.5 <0.5 <0.5 <0.5 <0.5Co 1.0 0.9 1.0 0.8 0.4 0.5 0.5 0.2Ni 3.5 3.3 3.9 4.1 1.2 1.5 1.4 0.9As 4.5 2.8 2.7 2.9 8.0 8.4 10.4 27.0Cd 1.0 1.6 2.6 2.5 0.7 0.7 0.6 0.6Hg 1.7 1.7 2.4 1.7 1.5 5.0 1.8 1.9Pb 0.9 2.4 1.5 1.7 3.0 4.0 2.9 1.8Cu 9.3 12.9 17.3 10.7 8.4 8.4 7.7 5.5Zn 77.5 94.9 93.0 90.2 72.3 64.2 60.8 53.8

Sum of metals 99.4 120.5 124.4 114.5 95.6 92.8 86.1 91.8

Bivalves were collected in June 2003 at eight locations in the Seine estuary and in the Seine Bay. Concentrations are expressed in �g g−1 dry weight.

PAH in bivalves and lower accumulation of six-ring PAH(Fig. 1A) in comparison to sediments. Benzo[a]anthracene andbenzofluroranthene were shown to be the two predominantcarcinogenic PAH in both bivalve species while dibenzoan-thracenes were barely detectable (Table 3). The ratio betweenbenzo[e]pyrene and benzo[a]pyrene in both bivalves wasalways exceeding 1 indicating possible metabolization of B[a]P(Fig. 1B). Actually, species differences in PAH intake and

biotransformation might exist since six-ring PAH were shownto be relatively more abundant in blue mussels (Fig. 1A) andone noticed a three-fold higher mean B[e]P/B[a]P ratio for bluemussels than for zebra mussels.

Seven different congeners of PCB were detected in bivalvetissues (Table 4). A slight decrease of PCB body burden wasobserved from the upper Seine estuary to the Seine Bay. Meanconcentrations of total PCB were very close between zebra mus-

Table 3PAH body burdens of zebra mussels and blue mussels

Compound Yville/SeineDreissena

OisselDreissena

La BouilleDreissena

CaudebecDreissena

VillervilleMytilus

Le HavreMytilus

AntiferMytilus

Le MoulardMytilus

Naphtalene 4.5 5 3 5 1.4 2 0.3 <0.3Acenaphthylene 0.4 4 3 2 5 3 2.2 <0.3Acenaphtene 1.2 4 3 3 5 3 1.5 <0.3Fluorene 2.6 5 5 3 7 6 4 1.7Phenanthrene (PHEN) 10.0 14 24 4 28 24 12 3Anthracene <0.1 2 7 0.8 2 3 1.4 <0.1Fluoranthene (FLUT) 12.3 95 131 18 69 90 33 3Pyrene 4.7 131 198 60 77 81 21 1.4Benzo[a]anthracene (BaA) 1.4 60 104 21 20 16 5 0.7Triphenylene (TRYF) + chrysene

(CHRY)6.5 171 253 91 57 52 14 1.2

Benzo[b]fluoranthene(BbF) + benzo[j]fluoranthene(BjF) + benzo[k]fluoranthene(BkF)

3.0 71 115 53 57 39 19 1

BBD

PIB∑

.7∑∑

FP

Bb

enzo[e]pyrene 1.3 56 84enzo[a]pyrene (BaP) <0.1 16 29ibenzo[a,h]anthracene(DahA) + dibenzo[a,c]anthracene(DacA)

<0.1 2 4

erylene 1.0 2 3ndeno[1.2.3-cd]pyrene (IcdP) <0.1 6 10enzo[g,h,i]perylene (BghiP) <0.1 8 15

PAHa 48.9 652.1 989

PAHmutb 33.3 (68) 443.1 (68) 684.6 (69

PAHcarc 4 (9) 154 (24) 263 (26)

luoranthene/Pyrene 2.6 0.7 0.7henanthrene/Anthracene – 7.0 3.4

ivalves were collected at eight locations in the Seine estuary and in the Seine Bayrackets are percentages of considered compounds vs. total PAH.a Sum of all analyzed PAH.b Sum of mutagenic PAH: PHEN, FLUT, BaA, CHRY, TRYF, BbF, BjF, BkF, BaP,c Sum of carcinogenic PAH: BaA, BbF, BjF, BkF, BaP, IcdP, DahA and DacA.

48 65 50 17 113 11 8 3 0.1

2 4 2 0.8 <0.1

3 7 5 2 <0.14 13 8 4 1.68 19 14 7 0.5

339.8 447.0 404.1 147.4 15.2

) 215.7 (63) 277.7 (62) 251.5 (62) 97.2 (66) 11.1 (73)

94 (28) 105 (23) 72 (18) 31 (21) 3 (22)0.3 0.9 1.1 1.6 2.15.0 16.0 8.0 8.6 –

in June 2003. Concentrations are expressed in ng g−1 dry weight. Numbers in

DahA, DacA, IcdP and BghiP.

70 B. Rocher et al. / Aquatic Toxicology 79 (2006) 65–77

Fig. 1. Relative distribution of 21 two- to six-ring PAH (A) andbenzo[e]pyrene/benzo[a]pyrene ratio (B) in soft tissues of zebra mussels andblue mussels. Bivalves were collected at eight different locations in the Seineestuary and in the Seine Bay in June 2003. One asterisk indicates significantdifferences at P ≤ 0.05.

sels (390 ng g−1 dw) and blue mussels (404 ng g−1 dw) and inboth case, PCB profile was characterized by the predominanceof congeners CB138 and CB153 and to a lesser extent CB101.In both species, a significant enhancement of total PCB con-tent (P = 0.048) was observed in individuals from contaminatedareas in comparison to those collected at the reference sites.Lindane was measured at relatively lower levels in blue mus-sels (0.93 ng g−1 dw) than in zebra mussels (1.75 ng g−1 dw)but no significant differences were observed between speciesor between contaminated areas and reference sites (Table 4).

3.2. Activities of defence enzymes

3.2.1. Superoxide dismutase activitySOD activity was measured in gills using a semi-quantitative

method directly on gels after native isoelectrofocusing. Migra-tion patterns were similar for both bivalve species showing threemain acidic isoforms of SOD. The two major ones were observedat pI 4.7 and 4.6, respectively, while the third one at pI 4.55 rep-resented less than 20% of the total SOD activity. The relativeactivity of each isoform was generally constant while total SODactivity, defined as the sum of the three isoform activities, washighly variable according to the sampling location (Table 5).Overall, SOD activities were significantly higher in zebra mus-sels from the Seine estuary in comparison to blue mussels fromt

Table 4Organochlorine content of soft tissues of zebra mussels and blue mussels

Compound Yville/SeineDreissena

OisselDreissena

La BouilleDreissena

CaudebecDreissena

CB 28 2 30 31 16CB 52 3 46 50 35CB 101 8 90 120 90CB 118 13 66 75 53CB 138 10 104 119 94CB 153 10 123 142 119CB 180 1 31 41 37�PCBs 47 490 579 444

Lindane 1.6 2.7 1.5 1.2

Bivalves were collected at eight locations in the Seine estuary and in the Seine Bay in

Table 5Glutathione S-transferase (GST), superoxide dismutase (SOD), catalase (CAT) andMytilus edulis sampled in June 2003 at eight locations in the Seine estuary and in the

Site Species GSTa

Yville/Seine Dreissena 794 ± 95Oissel Dreissena 1639 ± 91*La Bouille Dreissena 1218 ± 176*Caudebec Dreissena 1192 ± 136*Villerville Mytilus 1051 ± 80*Le Havre Mytilus 794 ± 101Antifer Mytilus 836 ± 113Le Moulard Mytilus 828 ± 143

Values are expressed as mean ± S.D. (N = 6). Asterisks indicate significant differenceMoulard for blue mussels.

a mU mg−1 protein except for catalase in U mg−1 protein.b Percentage of the reference site i.e. Yville-sur-Seine for Dreissena and Le Moula

he Seine Bay (P < 0.0001) whereas no significant difference

VillervilleMytilus

Le HavreMytilus

AntiferMytilus

Le MoulardMytilus

16 18 9 138 42 25 3

136 124 46 8763 78 43 5

131 133 81 10176 174 109 14

24 22 11 0.4583 591 324 120

1.8 0.9 0.9 0.1

June 2003. Concentrations are expressed in ng g−1 dry weight.

glutathione-peroxidase (GPX) activities in gills of Dreissena polymorpha andSeine Bay

SODb CATa GPXa

100 ± 33 287 ± 28 5.1 ± 0.8279 ± 60* 434 ± 49* 4.3 ± 0.8320 ± 112* 380 ± 43* 5.0 ± 1.5191 ± 80 335 ± 48 5.1 ± 1.3142 ± 19 * 468 ± 72 13.8 ± 2.5*

84 ± 15 269 ± 54* 9.3 ± 0.2

89 ± 7 336 ± 46 8.3 ± 1.1

100 ± 17 415 ± 56 8.5 ± 1.1

s (P ≤ 0.05) with reference site i.e. Yville-sur-Seine for zebra mussels and Le

rd for Mytilus.

B. Rocher et al. / Aquatic Toxicology 79 (2006) 65–77 71

was evidenced between animals of both species collected at thetwo reference sites (P = 0.979). Levels reported in gills of zebramussels collected in the Seine estuary at La Bouille (+220%)and Oissel (+179%) were significantly higher (P < 0.05)than values measured in individuals collected at Yville-sur-Seine. In the Seine Bay, only mussels collected at Villervilleexhibited significantly higher SOD activity (+40%, P < 0.05)when compared to individuals from the reference site of LeMoulard.

3.2.2. Glutathione S-transferase activityThe GST activity in gills of bivalves varied according to the

location of the sampling sites (Table 5). The lowest activities,about 800 mU mg−1 protein, were reported in animals collectedat the two reference sites of Yville-sur-Seine and Le Moulard.Zebra mussels sampled in the Seine estuary exhibited a signif-icant (P < 0.05) increase of their GST activity, i.e., by +106%at Oissel and about +50% at La Bouille and Caudebec in com-parison to Yville-sur-Seine. In the Seine Bay, mean GST activ-ity was significantly higher in mussels collected at Villerville(+27%, P < 0.05) in comparison to specimens sampled at LeHavre, Antifer and Le Moulard. Overall, the mean GST activi-ties were significantly higher (P < 0.0001) in zebra mussels fromthe Seine estuary in comparison to blue mussels from the SeineBay.

3

((fnfp

3

(mB(ps(easH

3

smtGb

Fig. 2. Bulky DNA adducts in digestive gland of zebra and blue mussels. (A)Relative adduct levels (RALs) per 108 nucleotides (mean ± S.D.) and (B) meannumber ± S.D. of individual spots at the different sampled sites. Bivalves werecollected in the Seine estuary and in the Seine Bay in June 2003. Analyses wereperformed by the nuclease P1-enhanced 32P-postlabelling assay on three poolsof 6–8 individuals per sampled site. One asterisk indicates significant differencesat P ≤ 0.05 and three asterisks at P ≤ 0.001.

diagonal radioactive zone (DRZ) for the Seine estuary musselswhile very few and more scattered spots were detected at thereference site.

In June 2003 a marked decline of DNA adduct levels wasobserved at La Bouille (−86%), Oissel (−92%) and to lesserextent at Yville-sur-Seine (−32.5%) and no more significantdifference could be evidenced between sites (Fig. 2A). Bluemussels from the Seine Bay showed significantly (P = 0.03)higher mean levels of adducts and also significantly highermean number of individual spots as compared to zebra mus-sels (Fig. 2A and B). In contrast, no significant differences wereobserved between mussels originated form Villerville, Le Havreand Antifer. Unexpectedly, repeated measurements on three dif-ferent pools of digestive glands revealed a 40-fold higher levelof DNA adducts in blue mussels from the reference site of LeMoulard (Fig. 2A). Overall, 14 different spots were identifiedand most of them were observed in both species. The four mainadducts, found in more than one third of all samples, were clearlydistributed close to the DRZ (data not shown).

The comet assay, performed on zebra mussels collected inJune 2002, revealed significantly lower (P < 0.05) levels of dam-aged hemocytes in individuals collected at the reference site incomparison to those sampled in the Seine estuary (Fig. 3A).A completely different pattern was observed in June 2003(Fig. 3B). Firstly, the ratios of atypic nuclei were shown to be

.2.3. Glutathione-peroxidase activityOverall, the mean GPx activity was about two times higher

P < 0.0001) in gills of blue mussels than in zebra musselsTable 5). No significant difference between sites was evidencedor zebra mussels collected in the Seine estuary. In contrast, sig-ificantly higher GPx activity was reported for blue musselsrom Villerville (+62%, P < 0.05) in comparison to those sam-led at Le Moulard.

.2.4. Catalase activityOverall, zebra and blue mussels exhibited similar levels

P = 0.368) of catalase activity in their gills (Table 5). Zebraussels collected in the upper part of the Seine estuary (Laouille, Oissel) exhibited significantly higher catalase activity

P < 0.05) than individuals from Yville-sur-Seine. The generalattern of activity for catalase in gills of zebra mussels was veryimilar to that observed for GST i.e. higher values at Oissel+51%) and a progressive but clear decrease along the Seinestuary (Table 5). Surprisingly, blue mussels from Le Moulardnd Villerville exhibited similar levels of catalase activity whileignificantly lower levels were measured in animals from Leavre (P < 0.05).

.3. Levels of DNA damage

The preliminary campaign carried out in June 2002 revealedignificantly higher levels (P < 0.0001) of bulky adducts in zebraussels collected in the upper Seine estuary in comparison

o those sampled at the reference site of Yville-sur-Seine (Leoff et al., 2006). The adduct profiles were also shown to varyetween the two areas with spots mainly distributed close to the

72 B. Rocher et al. / Aquatic Toxicology 79 (2006) 65–77

Fig. 3. Comet assay analysis of DNA strand breaks in hemocytes of zebra andblue mussels collected in June 2002 (A) and June 2003 (B) at different locationsin the Seine estuary and in the Seine Bay. About 400 cells from each of 26–30individuals per sampling site were visually examined. Cells were graded in fourcategories of damage according to the shape of the nucleus and length of thecomet. One asterisk indicates significant differences at P ≤ 0.05.

extremely high in all samples representing between 52 and 72%of hemocytes against only 7% in the negative control (humanleukocytes). This particular feature was associated with highlevels of necrotic and apoptotic cells (data not shown). Sec-ondly, although frequencies of damaged nuclei (11–19%) werevery close to those measured in 2002, no more significant dif-ference were observed between sites. In fact, hemocytes of bluemussels appeared slightly less damaged than those from zebramussels and for both species large inter-individual variability ofresponses was observed (Fig. 3B).

3.4. Relationship between contamination and biologicalresponses of mussels

A strong positive relationship was observed between all stud-ied contaminants (Table 6). In addition, body burden of PAH(R’s = 0.85, P < 0.0001, N = 14) and PCB (R’s = 0.62, P = 0.018,N = 14) but not metals or lindane were in clear concordance withlevels reported for sediments (Cachot et al., this issue). GST andSOD activities in gills showed a strong relationship (R’s = 0.87,P = 0.005). In fact, SOD was shown to be strongly related tometal tissue content (R’s = 0.87, P = 0.004) while GST seemedto be more closely related to PAH tissue content (R’s = 0.71,P = 0.05). GPX activity appeared negatively related to metalbody burden (R’s = −0.73, P = 0.04) and to a lesser extent tomost of other contaminants and enzyme activities (Table 6). Thisrtct

d Tabl

e6

Mat

rix

ofSp

earm

an’s

rank

corr

elat

ion

for

biva

lves

data

Para

met

ers

Len

gth

∑PA

Hs

∑PC

Bs

Lin

dane

∑m

etal

sSO

DG

STC

AT

GPX

Add

ucts

Com

et

Len

gth

1−0

.40

(11)

0.22

7−0

.12

(11)

0.72

6−0

.56

(11)

0.07

5−0

.81

(10)

0.00

5−0

.81

(8)0

.014

−0.6

6(8

)0.

077

−0.2

0(8

)0.

629

0.72

(8)0

.045

0.73

(8)0

.040

0.11

(11)

0.75

6∑

PAH

s−0

.40

(11)

0.22

71

0.85

(14)

<0.

001

0.79

(14)

<0.

001

0.70

(13)

0.00

70.

63(8

)0.

091

0.71

(8)0

.050

0.33

(8)

0.42

0−0

.33

(8)

0.41

7−0

.63

(8)

0.09

7−0

.13

(11)

0.70

9∑

PCB

s−0

.12

(11)

0.72

60.

85(1

4)<

0.00

11

0.75

(14)

0.00

20.

34(1

3)0.

255

0.11

(8)

0.80

00.

25(8

)0.

548

0.12

(8)

0.77

90.

27(8

)0.

509

−0.1

6(8

)0.

706

−0.2

3(1

1)0.

501

Lin

dane

−0.5

6(1

1)0.

075

0.79

(14)

<0.

001

0.75

(14)

0.00

21

0.60

(13)

0.02

90.

60(8

)0.

114

0.52

(8)

0.18

0.42

(8)

0.30

1−0

.41

(8)

0.31

3−0

.62

(8)

0.09

80.

11(1

1)0.

757

∑M

etal

s−0

.81

(10)

0.00

50.

70(1

3)0.

007

0.34

(13)

0.25

50.

60(1

3)0.

029

10.

87(8

)0.0

040.

67(8

)0.

069

0.14

(8)

0.73

6−0

.73

(8)0

.04

−0.9

5(8

)<0.

001

0.05

(10)

0.88

1SO

D−0

.81

(8)0

.014

0.63

(8)

0.09

10.

11(8

)0.

800

0.60

(8)

0.11

40.

87(8

)0.0

041

0.87

(8)0

.005

0.53

(8)

0.18

0−0

.67

(8)

0.07

0−0

.73

(8)0

.040

0.51

(8)

0.19

1G

ST−0

.66

(8)

0.07

70.

71(8

)0.0

500.

25(8

)0.

548

0.52

(8)

0.18

00.

67(8

)0.

069

0.87

(8)0

.005

10.

61(8

)0.

108

−0.6

1(8

)0.

109

−0.5

1(8

)0.

194

0.36

(8)

0.38

2C

AT

−0.2

0(8

)0.

629

0.33

(8)

0.42

00.

12(8

)0.

779

0.42

(8)

0.30

10.

14(8

)0.

736

0.53

(8)

0.18

00.

61(8

)0.

108

1−0

.012

(8)

0.97

70.

11(8

)0.

794

−0.0

9(8

)0.

822

GPX

0.72

(8)0

.045

−0.3

3(8

)0.

417

0.27

(8)

0.50

9−0

.41

(8)

0.31

3−0

.73

(8)0

.04

−0.6

7(8

)0.

070

−0.6

1(8

)0.

109

−0.0

12(8

)0.

977

10.

73(8

)0.0

40−0

.80

(8)0

.016

Add

ucts

0.73

(8)0

.040

−0.6

3(8

)0.

097

−0.1

6(8

)0.

706

−0.6

2(8

)0.

098

−0.9

5(8

)<0.

001

−0.7

3(8

)0.0

40−0

.51

(8)

0.19

40.

11(8

)0.

794

0.73

(8)0

.040

1−0

.80

(8)0

.018

Com

et0.

11(1

1)0.

756

−0.1

3(1

1)0.

709

−0.2

3(1

1)0.

501

0.11

(11)

0.75

70.

05(1

0)0.

881

0.51

(8)

0.19

10.

36(8

)0.

382

−0.0

9(8

)0.

822

−0.8

0(8

)0.0

16−0

.80

(8)0

.018

1

The

Spea

rman

’sco

rrel

atio

nco

effic

ient

,the

num

ber

ofsa

mpl

e(i

nbr

acke

ts)

and

the

Pva

lue

(in

italic

s)ar

ein

dica

ted

for

each

para

met

er.S

igni

fican

tdat

aar

ebo

ld-f

aced

.

esult could stem from the highest GPX activity and lower con-aminant content and enzymes activities in blue mussels. Nolear relationship was seen for catalase activity with any con-aminant (Table 6).

Length appeared as a suitable parameter to highlight speciesifferences. Indeed, strong negative relationships were evi-

B. Rocher et al. / Aquatic Toxicology 79 (2006) 65–77 73

denced between length of mussels and metal content or SODactivity and positive relationship between length and GPX activ-ity or DNA adduct levels (Table 6). These features could rea-sonably be attributed to the higher body burden of metals andSOD activity in zebra mussels and conversely higher GPX activ-ity and DNA adduct levels in blue mussels (Tables 2 and 5 andFig. 2).

4. Discussion

Biological responses of environmentally exposed animals areoften barely interpretable because of the high complexity of thepollutant cocktail which can trigger different and sometimescontradictory responses and also because chronic exposure tolow levels of contaminants can lead to physiological mecha-nisms of adaptation and hence may reduce animal sensitivity tocontaminants.

The present study investigated genotoxicant accumulationand biological responses of two different bivalve species chron-ically exposed to waterborne pollutants from the Seine River.The main purpose of this study was to determine the relation-ship between genotoxicants exposure and biological responsesin term of activation of cellular defences and induction of DNAdamage in animals collected along a pollution gradient. Thestudy focused on bivalves because they are filter-feeders andexhibit low biotransformation abilities in comparison to fish (foralrtffCf

caarc

tPttdtlboft2lr

a

reported for other bivalve species worldwide in coastal and estu-arine areas (Hendriks et al., 1998; Chase et al., 2001; Szefer et al.,2002; Bervoets et al., 2004; Liu and Kueh, 2005), except for Hgwhose levels were among the highest reported in the literature.Hg levels above 1 �g g−1 dw have only been reported in “hotspot” contaminated areas (Chase et al., 2001; Amodio-Cocchieriet al., 2003; Garron et al., 2005). Although quite elevated, Pb,Cu and Cd levels in this study were not comparable with theextremely high levels recorded near some polluted industrialareas of the Baltic Sea where, for instance, Cd and Pb levels ashigh as 20–40 �g g−1 dw and 210–264 �g g−1 dw, respectively,were reported (Szefer et al., 2002).

PAH concentrations in mussels were comparablewith those reported for moderately contaminated areas(200–1000 ng g−1 dw) of different bays worldwide (Baumardet al., 1999a,b; Chase et al., 2001; Liu and Kueh, 2005), andone order of magnitude lower at least than in highly anthro-pogenically impacted sites (Telli-Karakoc et al., 2002; Liu andKueh, 2005). The PAH sources can be characterized by meansof molecular indices based on the thermodynamic stability ofvarious isomeric compounds (Baumard et al., 1999a). Indeed,low phenanthrene/anthracene (P/A) ratios (<10) together withhigh fluoranthene/pyrene (F/P) ratios (>1) are indicative of apyrolytic source, whereas much higher P/A ratios (>25) togetherwith low F/P ratios (<1) are characteristic for a petrogenicsource. Mussels from the Seine estuary exhibited a particularpfp

tsRrlooi(bwad

iWttafEic2s

i

review see Livingstone, 1998) and thus are able to accumu-ate relatively high levels of certain toxicants. Bivalves can thusepresent an important source of contamination and hazards forheir predators. On the other hand, filter-feeders are key speciesor aquatic ecosystems because of their central position in theood net and their role in organic matter deposition and recycling.onsequently, population decline could deeply alter ecosystem

unctioning.Further in this issue we documented the high and complex

hemical contamination of sediments of the upper Seine estuarynd their potential genotoxicity and embryotoxicity (Cachot etl., this issue). Authors concluded that those sediments couldepresent a source of hazards for species living or feeding in thisompartment.

In the present study we reported the accumulation of rela-ively high levels of mutagens and carcinogens including PAH,CB and metals in soft tissues of zebra mussels originating from

he Seine estuary and to a lesser extent in blue mussels living athe mouth of the Seine River. The contamination of biota at theifferent locations was shown to be related to the contamina-ion of the sedimentary compartment but with noticeably lowerevels in the former. A main exception to this trend was theioaccumulation of PCB and lindane to levels about one orderf magnitude above concentrations reported for sediments. Thiseature has already been reported for native zebra mussels fromhe Seine River system (Chevreuil et al., 1996; Minier et al.,006) and can result from high uptake efficiency together withow capabilities of metabolization and excretion of organochlo-ines in this species (Bruner et al., 1994).

The Seine estuary and the Seine Bay appeared moder-tely contaminated by metals when comparing data with those

attern i.e. P/A < 10 and F/P < 1, indicating that PAH originatedrom a combination of different pollution sources, pyrolytic,etrogenic and/or diagenic origins.

Total PCB contents in mussels from the Seine estuary andhe Seine Bay ranged between moderate and high in compari-on to the literature (Connors et al., 2001; Chase et al., 2001;enaud et al., 2004). However, the values measured here never

eached the excessively high levels reported for some “hot spot”ocations as in the Rhine–Meuse basin (Hendriks et al., 1998)r in canals and lakes in Flanders (Bervoets et al., 2004). More-ver blue mussels collected in 1992 at stations under the directnfluence of the Seine’s plume exhibited higher contents of PCBAbarnou et al., 2002). Indeed, PCB 153 concentrations variedetween 200 and 500 ng g−1 dw in mussels from the Seine Bay,hereas concentrations recorded in the present study were rel-

tively lower (about 170 ng g−1 dw), confirming the trend of aecreasing PCB contamination in the Seine estuary.

Very little information is available concerning the ecotox-cological significance of the micropollutant levels in mussels.

hen compared to the OSPAR ecotoxicological assessment cri-eria (EAC) (OSPAR, 2000), total PCB contents in mussels ofhe Seine exceeded the highest value (58 ng g−1 dw) of the EACt all sites except at Yville-sur-Seine. Contrary to that observedor sediments, PAH levels in mussels were largely below theAC available for some PAH. Finally, the Hg level in mussels

ssued from Le Havre was largely above the maximum con-entration permitted (2.5 �g g−1 dw) in fishery products (EC,002), and was relatively close to this limit value at all the othersites.

This study also investigated activities of four enzymesnvolved in cellular defence processes. Two main results were

74 B. Rocher et al. / Aquatic Toxicology 79 (2006) 65–77

obtained. Firstly, significantly higher GST, SOD and to a lesserextent catalase activities were measured in gills of mussels col-lected in the Seine estuary and at the mouth of the Seine Riverin comparison to mussels from the reference sites and fromthe Seine Bay. Secondly, a strong relationship was evidencedbetween SOD and GST activities and levels of bioaccumulatedmetals and PAH. Correlation of GST activity with PAH tis-sue content has already been shown for several bivalves speciesincluding blue mussel (Gowland et al., 2002) and green-lippedmussel, Perna viridis (Cheung et al., 2001). Furthermore, up-regulation of GST omega class and Cu/Zn SOD was evidencedin digestive gland of the oyster C. gigas after 3–7 days ofexposure to a mixture of hydrocarbons (Boutet et al., 2004).GST activity in gills of mussels was also shown to be signifi-cantly depressed by exposure to high levels of B[a]P (Akcha etal., 2000). Numerous studies have also documented significantrelationships between antioxidant activities, mainly SOD andGPx, and metal or PAH body burdens in mussels (Porte et al.,1991; Manduzio et al., 2003; Giguere et al., 2003). For instance,Regoli and Principato observed a clear induction of SOD andGPx activities in gills of Mytilus galloprovincialis exposed tocopper in controlled conditions or to complex mixture of met-als in field conditions (Regoli and Principato, 1995). Cheungand collaborators demonstrated that B[a]P is a potent oxidativestress inducer in the marine mussel, Perna viridis since antioxi-dant parameters including GR, GPx and lipid peroxidation havebaabiiaof

asacYoevmoDahsmsbo

Me

mussels collected at that site in June 2003. Two major hypothe-ses could be put forward: (i) exposure to unknown genotoxicants(ii) and/or defect of repair of DNA damage. Only a limited num-ber of genotoxicants were measured in the course of this studyand one can assume that numerous other potential genotoxicantsincluding xenobiotics, phycotoxins or even radionuclei are actu-ally present in this area. Okadaic acid, which is known to induceDNA adduct formation in cell lines (Fessard et al., 1996), wasnot detected in digestive gland of those animals. Strikingly, athree-fold higher level of arsenic was measured in soft tissuesof blue mussels from Le Moulard in comparison to other loca-tions. Interestingly, this compound was shown to enhance B[a]Pgenotoxicity by inhibiting repair of DNA damage (Tran et al.,2002; Hartwig et al., 2003).

Concerning the Comet assay, significant differences wereobserved between sites in June 2002 with higher levels of DNAdamage reported for zebra mussels collected at both contami-nated sites of the Seine estuary. On the contrary, no more signifi-cant difference was observed between sites in June 2003. In addi-tion, a marked elevation of atypic nuclei associated to necroticor apoptotic cells was observed. This last feature unlikely stemfor genotoxicant exposure since a similar trend was observedat all locations. Conditions of transport (isotherm containers)and manipulation of bivalves as well as Comet assay proce-dure cannot be in question since they have been successfullyapplied in June 2002 and negative controls displayed low levelsotbDbiarhetb

ooctr2l1esahs

5

b

een correlated to tissue content of this compound (Cheung etl., 2004). The results reported here clearly suggest the positivectivation of cellular defence mechanisms in individuals fromoth bivalve species chronically exposed to waterborne contam-nants from the Seine River. Activation of those cellular defencess thought to prevent accumulation of electrophilic metabolitesnd free radicals and hence might partially protect DNA andthers cellular macromolecules against oxidation and adductormation.

Besides induction of cellular defences, DNA damage waslso detected in digestive gland and hemocytes of both bivalvepecies. During the preliminary study in June 2002, higher levelsnd larger number of adducts were measured in zebra musselsollected at both sites in the Seine estuary in comparison toville-sur-Seine (Le Goff et al., 2006). A different trend wasbserved in June 2003 with about one magnitude order lower lev-ls of adducts and no more difference between sites. Mean RALalues in blue mussels were significantly higher than in zebraussels although the latter species was exposed to higher levels

f PAH. The lower sensitivity of zebra mussel to PAH-inducedNA adducts might result from its lower ability to biologically

ctivate PAH as compared to blue mussels. Indeed, we reportere a three-fold higher mean B[e]P/B[a]P ratio in blue mus-els than in zebra mussels indicating a lower efficiency of B[a]Petabolization in the latter species. In this respect, zebra mus-

els exposed to B[a]P in controlled laboratory conditions haveeen shown to exhibit lower mean RAL values per microgramf B[a]P than blue mussels (Le Goff et al., 2006).

Unexpectedly, blue mussels from the reference site of Leoulard were shown to contain extremely high DNA adduct lev-

ls. This result was consistently obtained for the three pools of

f atypic nuclei (7%). More probably, necrotic and/or apop-otic processes in relation to poor condition of bivalves mighte involved. Although potential value of the comet assay forNA damage assessment in hemocytes of bivalves has largelyeen documented (Pavlica et al., 2001; Rank and Jensen, 2003),t has also been shown that non-genotoxic stress leading topoptosis can increase DNA fragmentation and comet assayesponses (Steinert, 1996). In addition, temperature of wateras been shown to significantly alter DNA damage baseline lev-ls in hemocytes of zebra mussels and it has been suggestedhat mussel sensitivity towards environmental pollutants coulde temperature-dependent (Buschini et al., 2003).

It is noteworthy that massive mortalities among populationsf zebra mussels and blue mussels inhabiting the Seine estuaryccurred during summer 2003. This population decline coin-ided with unusually high temperatures of surface water, upo +6 ◦C above normal temperatures, together with decrease ofiver flow up to −50% during this period (Ficht and Olivier,003). It has been shown that high temperatures can affect pol-utant intake and physiology of the organism (Fisher et al., 1993,999) thus modifying certain biomarker responses to pollutantxposure (Vidal et al., 2002; Khessiba et al., 2005). In the presenttudy, pollutant accumulation as well as enzymatic responsesnd DNA damage in mussels were likely affected by unusuallyigh seasonal temperatures but at the same level throughout thetudied area.

. Conclusion

In the present study we highlighted differences betweenivalve species in metal accumulation (As and Pb) and GPx

B. Rocher et al. / Aquatic Toxicology 79 (2006) 65–77 75

activity as well as variations in PAH metabolization and sensi-tivity to DNA adduct formation. This work also clearly demon-strated that zebra mussels from the upper part of the Seineestuary were much more exposed to potential oxidizers andgenotoxicants than bivalves from others sites. However, thechemical stress induced in those animals was not sufficient, inusual temperature conditions, to exceed cellular defences andhence to significantly increase DNA damage in adult mussels.Actually, other cellular damage such as oxidations of lipids andproteins and effects at other developmental stages must be inves-tigated to ascertain absence of impact of genotoxicant exposure.Moreover, chemical stress can also increase animal sensitivityto fluctuation of other environmental factors. Indeed, a massivemortality of zebra mussels, likely related to exceptionally hightemperatures of surface water, was evidenced in the Seine estu-ary in August 2003. Finally, It must also be pointed out that byaccumulating high levels of mutagenic and carcinogenic PAH,zebra mussels could represent a major source of contaminationand hazards for crustaceans, fishes and birds feeding in the Seineestuary.

Acknowledgements

This study was supported by the French scientific pro-gram Seine-Aval. Jeremy Le Goff was a recipient of a fellow-sdwft

R

A

AA

A

A

B

B

B

Bervoets, L., Voets, J., Chu, S., Covaci, A., Schpens, A., Blust, R., 2004.Comparison of accumulation of micropollutants between indigenous andtransplanted zebra mussels (Dreissena Polymorpha). Environ. Toxicol.Chem. 23, 1973–1983.

Boutet, I., Tanguy, A., Moraga, D., 2004. Response of the pacific oysterCrassostrea gigas to hydrocarbon contamination under experimental con-ditions. Gene 329, 147–157.

Bradford, M.M., 1976. A rapid and sensitive method for the quantitationof microgram quantities of protein utilizing the principle of protein–dyebinding. Anal. Biochem. 72, 248–254.

Bruner, K.A., Fisher, S.W., Landrum, P.F., 1994. The role of the zebra mussel,Dreissena polymorpha, in contaminant cycling. I. The effect of body sizeand lipid content on the bioconcentration of PCB and PAH. J. Great LakeRes. 20, 725–734.

Buschini, A., Carboni, P., Martino, A., Poli, P., Rossi, C., 2003. Effectsof temperature on baseline and genotoxicant-induced DNA dam-age in haemocytes of Dreissena polymorpha. Mutat. Res. 537, 81–92.

Cachot, J., Geffard, O., Augagneur, S., Lacroix, S., Le Menach, K., Peluhet,L., Couteau, J., Denier, X., Devier, M.H., Pottier, D., Budzinski, H., thisissue. Evidence of genotoxicity related to high PAH content of sedimentsin the upper part of the Seine estuary (Normandy, France). Aquat. Toxi-col., doi:10.1016/j.aquatox.2006.06.014.

Canova, S., Degan, P., Peters, L.D., Livingstone, D.R., Voltan, R., Venier,P., 1998. Tissue dose, DNA adducts, oxidative DNA damage andCYP1A-immunopositive proteins in mussels exposed to waterbornebenzo[a]pyrene. Mutat. Res. 399, 17–30.

Chase, M.E., Jones, S.H., Hennigar, P., Sowles, J., Harding, G.C.H., Free-man, K., Wells, P.G., Krahforst, C., Coombs, K., Crawford, R., 2001.Gulfwatch: monitoring spatial and temporal patterns of trace metal andorganic contaminants in the Gulf of Maine (1991–1997) with the blue

C

C

C

C

C

C

E

E

F

hip from the Ligue national contre le cancer (Comite Calva-os) and from the Conseil Regional de Basse-Normandie. Weould like to thank Florence Mondeguer (IFREMER, Nantes)

or measurements of okadaic acid and derivates in musselissues.

eferences

barnou, A., Le Guellec, A.M., Loizeau, V., 2002. Les polychlorobiphenyles(PCB) dans l’environnement littoral. In: RNO 2002, Surveillance duMilieu Marin. Ifremer et Ministere de l’Ecologie et du DeveloppementDurable, pp. 13–24.

ebi, H., 1984. Catalase in vitro. Meth. Enzymol. 105, 121–126.kcha, F., Izuel, C., Venier, P., Budzinski, H., Burgeot, T., Narbonne, J.-

F., 2000. Enzymatic biomarker measurement and study of DNA adductformation in benzo[a]pyrene-contaminated mussels, Mytilus galloprovin-cialis. Aquat. Toxicol. 49, 269–287.

kopian, M., Garnier, J., Testard, P., Ficht, A., 2001. Estimating the benthicpopulation of Dreissena polymorpha and its impact in the lower SeineRiver, France. Estuaries 24, 1003–1014.

modio-Cocchieri, R., Amoroso, S., Arnese, A., Cirillo, T., Montuori, P.,Triassi, M., 2003. Pollution by mercury, arsenic, lead, chromium, cad-mium, and polycyclic aromatic hydrocarbons of fish and mussels fromthe gulf of Naples Italy. Bull. Environ. Contam. Toxicol. 71, 551–560.

aumard, P., Budzinski, H., Garrigues, P., Dizer, H., Hansen, P.D., 1999a.Polycyclic aromatic hydrocarbons in recent sediments and mussels(Mytilus edulis) from Western Baltic Sea: occurrence, bioavailability andseasonal variations. Mar. Environ. Res. 47, 17–47.

aumard, P., Budzinski, H., Garrigues, P., Narbonne, J.F., Burgeot, T., Michel,X., Bellocq, J., 1999b. Polycyclic aromatic hydrocarbon (PAH) burden ofmussels (Mytilus sp.) in different marine environments in relation withsediment PAH contamination, and bioavailability. Mar. Environ. Res. 47,415–439.

eauchamp, C., Fridovitch, I., 1971. Superoxide dismutase: improved assaysand an assay applicable to acrylamide gels. Anal. Biochem. 44, 276–287.

mussel, Mytilus edulis L. Mar. Pollut. Bull. 42, 490–504.heung, C.C.C., Zheng, G.J., Li, A.M.Y., Richardson, B.J., Lam, P.K.S.,

2001. Relationships between tissue concentrations of polycyclic aro-matic hydrocarbons and antioxidative responses of marine mussels, Pernaviridis. Aquat. Toxicol. 52, 189–203.

heung, C.C.C., Siu, W.H.L., Richardson, B.J., De Luca-Abbott, S.B., Lam,P.K.S., 2004. Antioxidant responses to benzo[a]pyrene and Aroclor 1254exposure in the green-lipped mussel, Perna viridis. Environ. Pollut. 128,393–403.

hevreuil, M., Blanchard, M., Teil, M.-J., Carru, A.-M., Testard, P., Ches-terikoff, A., 1996. Evaluation of the pollution by organochlorinatedcompounds (polychlorobiphenyls and pesticides) and metals (Cd, Cr,Cu and Pb) in the water and in the zebra mussel (Dreissena poly-morpha Pallas) of the Seine River. Water Air Soil Pollut. 88, 371–381.

laisse, D., Joanny, M., Quintin, J.Y., 1992. Le Reseau Nationald’Observation de la qualite du milieu marin (RNO). Analysis 20,719–722.

onnors, L., Johnson, M.S., Copplestone, D., Leah, R.T., 2001. Recent trendsin organochlorine residues in mussels (Mytilus edulis) from the Merseyestuary. Mar. Environ. Res. 52, 397–411.

ossu, C., Doyotte, A., Jacquin, M.C., Babut, M., Exinger, A., Vasseur, P.,1997. Glutathione reductase, selenium-dependent glutathione peroxidase,glutathione levels, and lipid peroxidation in freshwater bivalves, Uniotumidus, as biomarkers of aquatic contamination in field studies. Ecotox-icol. Environ. Saf. 38, 122–131.

C, 2002. Commission regulation No. 221/2002 of 6 February 2002 amend-ing Regulation (EC) No. 466/2001 (OJ L77 16.03.2001, p. 7) setting max-imum levels for certain contaminants in foodstuffs (OJ L37 07.02.2002,p. 4).

ricson, G., Skarpheðinsdottir, H., Zuanna, L.D., Svavarsson, J., 2002. DNAadducts as indicators of genotoxic exposure in indigenous and transplantedmussels, Mytilus edulis L. from Icelandic coastal sites. Mutat. Res. 516,91–99.

essard, V., Grosse, Y., Pfohl-Leszkowitz, A., Puiseux-Dao, S., 1996. Okadaicacid treatment induces DNA adduct formation in BHK21 C13 fibroblastsand HESV keratinocytes. Mutat. Res. 361, 133–141.

76 B. Rocher et al. / Aquatic Toxicology 79 (2006) 65–77

Ficht, A., Olivier, M., 2003. Suivi de la qualite des eaux de la Seine a l’avalde Poses. In: Bilan de l’annee 2003. Service de la Navigation de la Seine,p. 157.

Fisher, S.W., Gossiaux, D.C., Bruner, K.A., Landrum, P.F., 1993. Investi-gations of the toxicokinetics of hydrophobic contaminants in the zebramussel (D. polymorpha). In: Nalepa, T.F., Schloesser, D.W. (Eds.), ZebraMussels: Biology, Impacts and Control. CRC Press, Boca Raton, FL, pp.465–490.

Fisher, S.W., Hwang, H., Atanasoff, M., Landrum, P.F., 1999. Lethal bodyresidues for pentachlorophenol in zebra mussels (D. polymorpha) undervarying conditions of temperature and pH. Ecotoxicol. Environ. Saf. 43,274–283.

Fitzpatrick, P., O’Halloran, J., Sheehan, D., Walsh, A.R., 1997. Assessmentof a glutathione S-tranferase and related proteins in the gill and digestivegland of Mytilus edulis (L.), as potential organic pollution biomarkers.Biomarkers 2, 51–56.

Garron, C., Gagne, F., Ernst, W., Julien, G., Bernier, M., Caldwell, C., 2005.Mercury contamination of marine sediments and blue mussels (Mytilusedulis) in the vicinity of a mercury cell chlor-alkali plant in Dalhousie,New Brunswick, Canada. Water Qual. Res. J. Can. 40, 1–15.

Giguere, A., Couillard, Y., Campbell, P.G.C., Perceval, O., Hare, L., Pinel-Alloul, B., Pellerin, J., 2003. Steady-state distribution of metals amongmetallothionein and other cytosolic ligands and links to cytotoxicityin bivalves living along a polymetallic gradient. Aquat. Toxicol. 64,185–200.

Gowland, B.T.G., McIntosh, A.D., Davies, I.M., Moffat, C.F., Webster, L.,2002. Implications of a field study regarding the relationship betweenpolycyclic aromatic hydrocarbons and glutathione S-transferase activityin mussels. Mar. Environ. Res. 54, 231–235.

Habig, W.H., Pabst, M.J., Jakoby, W.B., 1974. Glutathions S-transferases: thefirst enzymatic step in mercapturic acid formation. J. Biol. Chem. 249,

H

H

I

K

L

L

L

L

L

L

Manduzio, H., Monsinjon, T., Rocher, B., Leboulenger, F., Galap, C., 2003.Characterization of an inducible isoform of the Cu/Zn superoxide dis-mutase in the blue mussel Mytilus edulis. Aquat. Toxicol. 64, 73–83.

Minier, C., Abarnou, A., Jaouen-Madoulet, A., Le Guellec, A.-M., Tutundjian,R., Bocquene, G., Leboulenger, F., 2006. A pollution-monitoring pilotstudy involving contaminant and biomarker measurements in the Seineestuary, France, using zebra mussels (Dreissena polymorpha). Environ.Toxicol. Chem. 25, 112–119.

Mitchelmore, C.L., Birmelin, C., Chipman, J.K., Livingstone, D.R., 1998.Evidence for cytochrome P450 catalysis and free radical involvement inthe production of DNA strand breaks by benzo[a]pyrene and nitroaromat-ics in mussel (Mytilus edulis L.) digestive gland cells. Aquat. Toxicol.41, 193–212.

Nasci, C., Da Ros, L., Campesan, G., Fossato, V.U., 1998. Assessment ofthe impact of chemical pollutants on mussel, Mytilus galloprovincialis,from the Venice Lagoon Italy. Mar. Environ. Res. 46, 279–282.

O’Connor, T.P., 1998. Mussel Watch results from 1986 to 1996. Mar. Pollut.Bull. 37, 14–19.

OSPAR, 2000. Quality Status Report 2000 for the North-East Atlantic.OSPAR Commission, London, pp. 41–63, 108 (Chapter 4).

Pavlica, M., Klobucar, G.I.V., Mojas, N., Erben, R., Papes, D., 2001. Detec-tion of DNA damage in haemocytes of zebra mussel using comet assay.Mutat. Res. 490, 209–214.

Porte, C., Sole, M., Albaiges, J., Livingstone, D.R., 1991. Responses ofmixed-function oxygenase and antioxidase enzyme system of Mytilus sp.to organic pollution. Comp. Biochem. Physiol. 100C, 183–186.

Prohaska, J.R., 1980. The glutathione peroxidase activity of glutathione S-transferase. Biochim. Biophys. Acta 611, 87–98.

Rank, J., Jensen, K., 2003. Comet assay on gill cells and hemocytes fromthe blue mussel Mytilus edulis. Ecotoxicol. Environ. Saf. 54, 323–

R

R

R

R

S

S

S

S

S

T

7130–7139.artwig, A., Blessing, H., Schwerdtle, T., Walter, I., 2003. Modulation of

DNA repair processes by arsenic and selenium compounds. Toxicology193, 161–169.

endriks, A.J., Pieters, H., De Boer, J., 1998. Accumulation of metals, poly-cyclic (halogenated) aromatic hydrocarbons, and biocides in zebra musseland eel from the Rhine and Meuse rivers. Environ. Toxicol. Chem. 17,1885–1898.

ARC monographs on the evaluation of carcinogenic risks to humans.http://monographs.iarc.fr/.

hessiba, A., Romeo, M., Aissa, P., 2005. Effects of some environmen-tal parameters on catalase activity measured in the mussel (Mytilusgalloprovincialis) exposed to lindane. Environ. Pollut. 133, 275–281.

ebailly, P., Vigreux, C., Godard, T., Sichel, F., Bar, E., Le Talaer, J.Y., Henry-Amar, M., Gauduchon, P., 1997. Assessment of DNA damage induced invitro by etoposide and two fungicides (carbendazim and chlorothalonil)in human lymphocytes with the comet assay. Mutat. Res. 375, 205–217.

e Goff, J., Gallois, J., Pelhuet, L., Devier, M.H., Budzinski, H., Pottier, D.,Andre, V., Cachot, J., 2006. DNA adduct measurements in zebra mussels,Dreissena polymorpha Pallas. Potential use for genotoxicant biomonitor-ing of fresh water ecosystems. Aquat. Toxicol. 70, 55–64.

i, Y., Seacat, A., Kuppusamy, P., Zweier, J.L., Yager, J.D., Trush,M.A., 2002. Copper redox-dependent activation of 2-tert-butyl(1,4)hydroquinone: formation of reactive oxygen species andinduction of oxidative DNA damage in isolated DNA and cultured rathepatocytes. Mutat. Res. 518, 123–133.

iu, J.H., Kueh, C.S.W., 2005. Biomonitoring of heavy metals and traceorganics using the intertidal mussel Perna viridis in Hong Kong coastalwaters. Mar. Pollut. Bull. 51, 857–875.

ivingstone, D.R., 1994. Recent developments in marine invertebrateorganic xenobiotic metabolism. Toxicol. Ecotoxicol. News 1, 88–94.

ivingstone, D.R., 1998. The fate of organic xenobiotics in aquatic ecosys-tems: quantitative and qualitative differences in biotransformation byinvertebrates and fish. Comp. Biochem. Physiol. A 120, 43–49.

329.eddy, M.V., Randerath, K., 1986. Nuclease P1 mediated enhancement of

sensitivity of 32P-postlabelling test for structurally diverse DNA adducts.Carcinogenesis 7, 1543–1551.

egoli, F., Principato, G., 1995. Glutathione, glutathione-dependent andantioxidant enzymes in mussel, Mytilus galloprovincialis, exposed to met-als under field and laboratory conditions: implications for the use ofbiochemical biomarkers. Aquat. Toxicol. 31, 143–164.

enaud, C.B., Martel, A.L., Kaiser, K.L.E., Comba, M.E., 2004. A compar-ison of organochlorine contaminant levels in the zebra mussel, Dreis-sena polymorpha, versus is unionid attachment, Elliptio complanata,in the Rideau River, Ontario. Water Qual. Res. J. Can. 39, 83–92.

NO database. Reseau National d’Observation de la qualite du milieu marin.MEDD/IFREMER, Quadrige database. http://www.ifremer.fr/envlit/surveillance/index.htm.

hing, N.P., McCoy, M.T., Tice, R.R., Schneider, E.L., 1988. A simple tech-nique for quantitation of low levels of DNA damage in individual cells.Exp. Cell Res. 175, 184–191.

ole, M., Porte, C., Albaiges, J., 1995. The use of biomarkers for assess-ing the effects of organic pollution in mussels. Sci. Total Environ. 159,147–153.

ole, M., Porte, C., Biosca, X., Mitchelmore, C.L., Chipman, J.K., Living-stone, D.R., Albaiges, J., 1996. Effects of the “Aegean Sea” oil spill onbiotransformation enzymes, oxidative stress and DNA adducts in diges-tive gland of the mussel (Mytilus edulis L.). Comp. Biochem. Physiol.113C, 257–265.

teinert, S.A., 1996. Contribution of apoptosis to observed DNA damage inmussel cells. Mar. Environ. Res. 42, 253–259.

zefer, P., Frelek, K., Szefer, K., Lee, C.-B., Kim, B.-S., Warzocha, J., Zdro-jewska, I., Cesielski, T., 2002. Distribution and relationships of tracemetals in soft tissue, byssus and shells of Mytilus edulis trossulus fromthe southern Baltic. Environ. Pollut. 120, 423–444.

elli-Karakoc, F., Tolun, L., Henkelmann, B., Klimm, C., Okay, O., Schramm,K.-W., 2002. Polycyclic aromatic hydrocarbons (PAH) and polychlori-nated biphenyls (PCB) distributions in the Bay of Marmara sea: IzmitBay. Environ. Pollut. 119, 383–397.

B. Rocher et al. / Aquatic Toxicology 79 (2006) 65–77 77

Thompson, S., Budzinski, H., 1999. Determination of polychlorinatedbiphenyls and chlorinated pesticides in environmental biological samplesusing focused microwave-assisted extraction. Int. J. Environ. Anal. Chem.76, 49–60.

Tran, H.P., Prakash, A.S., Barnard, R., Chiswell, B., Ng, J.C., 2002. Arsenicinhibits the repair of DNA damage induced by benzo(a)pyrene. Toxicol.Lett. 133, 59–67.

Venier, P., Canova, S., 1996. Formation of DNA adducts in the gill tissueof Mytilus galloprovincialis treated with benzo[a]pyrene. Aquat. Toxicol.34, 119–133.

Vidal, M.-L., Basseres, A., Narbonne, J.-F., 2002. Influence of temperature,pH, oxygenation, water-type and substrate on biomarker responses in thefreshwater clam Corbicula fluminea (Muller). Comp. Biochem. Physiol.C 132, 93–104.