journa l homepage: www.e lsev ier .com/ locate /gentoxCommuni ty address : www.e lsev ier .com/ locate /mutres
enotoxicity in herring gulls (Larus argentatus) in Sweden and Iceland
alldora Skarphedinsdottir a,∗, Gunnar Thor Hallgrimssonb, Tomas Hanssona, Per-Åke Hägerrotha,irgitta Liewenborga, Ulla Tjärnlunda, Gun Åkermana, Janina Barsiene c, Lennart Balka
Department of Applied Environmental Science, Unit of Environmental Toxicology and Chemistry, Stockholm University, SE-106 91 Stockholm, SwedenInstitute of Biology, University of Iceland, IS-101 Reykjavík, IcelandInstitute of Ecology, Vilnius University, Akademijos 2, LT-08412 Vilnius 21, Lithuania
r t i c l e i n f o
rticle history:eceived 2 February 2010eceived in revised form 16 June 2010ccepted 23 June 2010vailable online 17 July 2010
eywords:NA adductsicronuclei
a b s t r a c t
Adult and young herring gulls (Larus argentatus) in Sweden and Iceland were investigated with respectto DNA adducts, analysed with the nuclease-P1 version of the 32P-postlabelling method, and micronu-cleated erythrocytes. Three important aims were: (1) to estimate the degree of exposure to genotoxicenvironmental pollutants in the Baltic Sea area and Iceland, (2) to evaluate the utility of the investigatedbiomarkers in birds, and (3) to investigate if there was any relationship between genotoxic effects andthiamine deficiency. The results demonstrate that both Swedish and Icelandic herring gulls are exposedto genotoxic pollution. Urban specimens have higher levels of DNA adducts than rural specimens, butbackground exposure to genotoxic environmental pollutants, such as PAHs, is also significant. In the her-
ring gull the general level of DNA adducts in the liver seems to be higher than in fish. DNA adducts weremost abundant in the liver, followed by the kidney, intestinal mucosa, and whole blood, in decreasingorder. The frequency of micronucleated erythrocytes was probably slightly elevated in all the investigatedsites, reflecting a significant background exposure. The level of DNA adducts was unrelated to the fre-quency of micronucleated erythrocytes, and both these variables were unrelated to symptoms of thiaminedeficiency. The investigation confirmed the utility of DNA adducts, and probably also micronucleatederythrocytes, as biomarkers of genotoxicity in birds.
. Introduction
During the last decades, many bird species breeding in Swedennd the Baltic Sea area have suffered from conspicuous populationeclines [1,2] and some of them are even endangered over theirntire range, e.g. the Fennoscandian subspecies of the lesser black-acked gull (Larus fuscus fuscus) [3]. In a recent study, paralysis andass mortality in the herring gull (Larus argentatus) and several
ther bird species in the Baltic Sea area were explained by thiamineeficiency, which was also suggested as a possible cause for thebserved population declines [4]. In the present study, genotoxicityas measured in the same specimens that were examined with
espect to thiamine deficiency [4]. Hence, one important aim was tonvestigate if there was any relationship between genotoxic effectsnd thiamine deficiency.
So far, birds have received very little attention in studies of envi-onmental genotoxicity. Previous work on genotoxicity in herringulls included the effect of industrial pollution on mutation ratesuantified by multi-locus DNA fingerprinting [5,6] and on DNA
damage measured as strand length (median molecular length ofDNA) [7]. In the present study, genotoxicity was measured as DNAadducts and frequency of micronucleated erythrocytes. Two impor-tant aims were: (1) to estimate the degree of exposure to genotoxicenvironmental pollutants in the Baltic Sea area and Iceland, and (2)to evaluate the utility of these biomarkers in birds. To corroborateour results we investigated the relationship between the biomarkerresponses and the actual pollution level of the surroundings. TheBaltic Sea has historically received large inputs of pollutants and iscurrently considered to be one of the most polluted seas in theworld [8] with eutrophication and pollution affecting its entireecosystem [9].
Genotoxic substances produce chemical and/or physical mod-ifications to the DNA, and damaged DNA may lead to a rangeof consequences. Several studies have demonstrated relation-ships between DNA damage and reduced fitness, such as geneand protein dysfunction, tumour initiation, growth impairment,embryonic malformations, reduced fecundity, and negative effects
on longevity [10–15]. Polycyclic aromatic hydrocarbons (PAHs) area major class of genotoxic pollutants commonly present in pol-luted areas. The genotoxic effects of PAHs are well known, andespecially the formation of DNA adducts [e.g. 16–18]. Adducts areformed when reactive electrophilic metabolites of chemicals bind
Fig. 1. The investigated regions. Adult herring gulls (Larus argentatus) were collectedin the County of Södermanland, the County of Blekinge, and the County of Skåne inSpSo
criDt[[o
Mndbcom
2
2
arpw(ojswitp
weden and on the Reykjanes peninsula and at Djúpivogur on Iceland. Herring gullulli were collected in the County of Värmland and the County of Södermanland inweden and on the Reykjanes peninsula on Iceland. The County of Skåne was thenly urban site (Malmö harbour). The other sites were rural.
ovalently to the DNA. Such reactive electrophilic metabolites alsoeadily attack protein, RNA, and other cell constituents, therebyncreasing the risk for a wide range of dysfunctions. Consequently,NA adducts may be used both as a measure of the direct geno-
oxic effect, and as an indicator of a wider range of toxic exposure10,19]. DNA adducts have been widely used as a biomarker in fishe.g. 18,20,21] and mussels [e.g. 22,23], and are considered to bene of the best biomarkers of PAH exposure.
The formation of micronuclei is another genotoxic effect.icronuclei are small DNA-containing bodies outside the cell
ucleus, formed by chromosome breakage, centromere or spindleysfunction, or defective cytokinesis (cell division) [24]. They maye generated by aneugens, which damage the spindle, and/or bylastogens, which induce chromosomal breaks [25]. The frequencyf micronucleated erythrocytes has been used as a biomarker forore than 25 years [26].
. Methods
.1. Bird material and sampling
Adult herring gulls (L. argentatus) were collected in three regions in Swedennd two regions in Iceland (Fig. 1) from mid-May to mid-August 2004. The Swedishegions were: the County of Skåne (Malmö harbour), an urban site with high anthro-ogenic pollution load (Table 1), and the Counties of Blekinge and Södermanland,hich are both rural sites with comparatively lower anthropogenic pollution load
Table 1). Herring gull pulli were collected in the County of Värmland and the Countyf Södermanland, which are both rural sites. The Icelandic regions were the Reyk-
anes peninsula (Reykjanes and Miðnesheiði) and Djúpivogur, which are two ruralites without known anthropogenic pollution sources. The two Icelandic regionsere equivalent with respect to DNA adducts and micronucleated erythrocytes
n the adult herring gulls, and were therefore pooled and used as a single con-rol for the Baltic Sea area. Icelandic pulli were collected only on the Reykjaneseninsula (not Djúpivogur). The adult herring gulls were caught with cannon-nets
n Research 702 (2010) 24–31 25
(Iceland) or fyke or hand nets (Sweden), whereas the pulli were collected by handor with a hand-net. The birds were transported to the laboratory, where they wereexamined and sampling was performed. The occurrence of paralytic symptoms wasnoted. Sacrifice was performed by cervical dislocation. The condition was estimatedby visual inspection on a four-grade scale based on the amount of fat on innerorgans: 0 = emaciated (completely devoid of body fat), 1 = poor, 2 = intermediate,and 3 = good. Liver-body index (LBI) was obtained as 100 times the liver weightdivided by the body weight. The LBI values fell into two distinct groups (withoutoverlap). The group with LBI ≥ 2.5% were considered to suffer from hepatomegaly.The exact age of the full-grown birds was not determined. The weight range of thepulli was 100–670 g and, accordingly, all of them had started to eat food provided bythe parents. Collection of birds in the field for invasive sampling was performed incompliance with permits issued by the Swedish Environmental Protection Agencyand the Icelandic Ministry for the Environment.
2.2. Sample preparation
Before sacrifice, blood was sampled directly from the heart with a 10-mL syringe.Blood smears for determination of the frequency of micronucleated erythrocyteswere prepared with the wedge procedure described by Vives Corrons et al. [29]. Theblood smears were air-dried, fixed for 10 min in 99.5% methanol, air-dried again,and kept in the dark until staining and counting of micronucleated erythrocytes.Tissue samples for DNA-adduct analysis were taken from the liver, kidney, intestinalmucosa, and whole blood. The samples were placed in cryotubes, snap-frozen inliquid nitrogen and kept in a −80 ◦C freezer until analysis.
2.3. Chemicals
Giemsa stain (11700), micrococcal endonuclease (N-3755), nuclease P1 (7160),RNase A (R-4642), spermidin (S-2626), spleen phosphodiesterase (P-9041), stan-dard DNA (salmon sperm, D-1626), and Trizma® base (T1503) were obtainedfrom Sigma–Aldrich Sweden AB (Stockholm, Sweden). �-Amylase (102814), phe-nol (1814303), proteinase K (1000144), RNase T1 (109193), and T4-polynucleotidekinase (3′-phosphatase free, 838292) were obtained from Roche Diagnostics Scandi-navia AB (Bromma, Sweden). Radio-labelled ATP ([�-32P]ATP) with specific activity3000 Ci/mmol (110 TBq/mmol) was obtained from GE Healthcare Bio-Sciences(Uppsala, Sweden). The benzo[a]pyrene standard adduct, 7R,8S,9S-trihydroxy,10R-(N2-deoxyguanosyl-3′-phosphate)-7,8,9,10-tetrahydro-benzo[a]-pyrene (B[a]PDE-dG-3′p), was obtained from Midwest Research Institute (Kansas City, MO,USA). Cellulose (MN-301) was obtained from Machery-Nagel (Düren, Germany).Vinyl strips (PVC foil, 0.2-mm thickness), used for the groundwork of thepolyethyleneimine cellulose sheets were obtained from Andrén & Söner (Stockholm,Sweden). Ultima GoldTM scintillation fluid (6013329) was obtained from CiAB (Sol-lentuna, Sweden). All other chemicals were of analytical purity and were obtainedfrom common commercial sources.
2.4. Analysis of DNA adducts
Our experience of other vertebrates is that DNA-adduct levels are higher in theliver than in other organs. Therefore, in order to avoid values below the detectionlimit in the kidney, intestinal mucosa and whole blood, a stepwise procedure wasused. DNA adducts were first measured in the liver of 25 adult herring gulls. Thena sub-sample, consisting of specimens with high levels of liver DNA adducts, wasselected for measurement of DNA adducts in the kidney, intestinal mucosa, andwhole blood. The Swedish part of the sub-sample consisted of the specimens withranks 1–3, 7, and 12 (three highest plus two among the 12 highest with respect toliver DNA adducts), whereas the Icelandic part of the sub-sample consisted of thespecimens with ranks 1–6 (six highest with respect to liver DNA adducts).
The tissue samples were semi-thawed and the DNA was extracted and purifiedaccording to Dunn et al. [16], and Reichert and French [30] with slight modifi-cations according to Ericson and Balk [31]. DNA adducts were analysed with thenuclease-P1 version of the 32P-postlabelling method [32] with some modificationsdescribed below. Purified DNA was hydrolysed to 3′-nucleoside monophosphatesby incubation of 12.5-�L aliquots at 37 ◦C for 4 h with micrococcal endonuclease(24 mU/�g DNA) and spleen phosphodiesterase (3.2 mU/�g DNA) in a solution of0.1 mM CaCl2 and 10 mM succinate buffer, pH 6.0. DNA adducts were enrichedby the nuclease-P1 method: normal nucleotides were hydrolysed to nucleosidesby addition of 0.41 �g nuclease P1 per �g DNA and incubation of the sample at37 ◦C for 30 min. After adjustment of the pH to approximately 7.8 by the additionof 1 �l of 0.5 M Tris–HCl, the mixture was evaporated to dryness in a SpeedVac®
centrifuge (Savant Instrument Inc. Farmingdale, NY, USA). The DNA adducts wereradio-labelled by incubation at 37 ◦C for 30 min with a mixture containing 8.3 Uof T4-polynucleotide kinase, 3.1 MBq [�-32P]-ATP (110 TBq/mmol), 10 mM MgCl2,
10 mM dithiothreitol, 3.5 mM spermidine, and 50 mM Tris–HCl, pH 7.6.
The radio-labelled DNA adducts were separated by means of multidirectionalthin-layer chromatography (TLC) on polyethyleneimine cellulose thin-layer sheetsprepared according to Reichert and French [30] with a Desaga spreader (Desaga,Heidelberg, Germany). Chromatography solvents used for the separation wereessentially the same as those described by Reddy and Randerath [32]:
26 H. Skarphedinsdottir et al. / Mutation Research 702 (2010) 24–31
Table 1�PAH concentrations in surface sediment and blue mussel (Mytilus edulis).
Region Matrix WGS 84coordinates
Year �PAH conc.[�g/g dw]
Reference
County of Östergötlanda (20 km offshore) Surface sediment (0–2 cm) 58.17405 N17.47651 E
2005 0.957b Supporting data fromSGU [27]
County of Blekinge (5 km offshore) Surface sediment (0–2 cm) 56.08324 N15.01798 E
2005 0.893b Supporting data fromSGU [27]
County of Skåne (15 km NW of Malmö, offshore) Surface sediment (0–2 cm) 55.79291 N12.83950 E
County of Södermanland (rural site) Blue mussel (Mytilus edulis) 58.82321 N17.63555 E
2000 0.040d [23]
County of Östergötlanda (rural site) Blue mussel (Mytilus edulis) 58.05631 N16.81342 E
2000 0.100d [23]
Western Iceland (Fiord Hvalfjörður, rural site) Blue mussel (Mytilus edulis) 64.36333 N21.49500 W
2000 0.240d [23]
a Southern neighbour to the County of Södermanland.b Sum of 15 PAHs measured in both investigations [27,28]: naphthalene, acenaphthene, fluorene, phenanthrene, anthracene, fluoranthene, pyrene, benzo[a]anthracene,
c ,3-cd]as bel
(
••••
idpi
lDtq5R
pd(wD
hntol
Fmmseta
2.7. Data analysis
The statistical analysis included the parametric methods analysis of vari-ance (ANOVA), analysis of covariance (ANCOVA), regression models, the Waldtest, and the Pearson correlation, as well as the non-parametric methods
hrysene, benzo[b]fluoranthene, benzo[k]fluoranthene, benzo[a]pyrene, indeno[1,2c Arithmetic mean of 11 stations. When the concentration of an individual PAH w
0.0005 �g/g dw).d Sum of 32 PAHs [23].
D1: 1.0 M sodium phosphate, pH 6.0;D3: 7.25 M urea, 3.86 M lithium formate, pH 3.5;D4: 7.25 M urea, 0.42 M Tris–HCl, 1.36 M lithium chloride, pH 8.0;D5: 1.7 M sodium phosphate, pH 6.0.
The TLC sheets were eluted with D1; dipped for a few seconds (∼2 cm elution)n 0.38 M lithium formate, pH 3.5; eluted with D3 in the direction opposite to D1;ipped for a few seconds (∼2 cm elution) in 0.42 M Tris–HCl, 0.68 M lithium chloride,H 8.0; eluted with D4 perpendicular to D3; dipped for a few seconds (∼2 cm elution)
n distilled water; and eluted with D5 in the same direction as D4.Liquid-scintillation counting was performed with a Packard Tri-Carb 2100TR
iquid-scintillation counter (Packard Instrument Company Inc., Meriden, CT, USA).NA was quantified by its absorption at 260 nm with a GeneQuant spectropho-
ometer (Pharmacia Biotech, Uppsala, Sweden). The DNA adducts were located anduantified on the thin-layer sheets with a PhosphorImager SITM and ImageQuant.0 software (Molecular Dynamics, Sunnyvale, CA, USA), essentially as described byeichert et al. [33].
Control material, used to assure the quality of all the analytical steps, were: (a)ure salmon-sperm DNA (negative control); (b) the standard DNA adduct B[a]PDE-G-3′p (positive control); and (c) liver tissue from benzo[a]pyrene-exposed perchPerca fluviatilis) (positive control). The three standards were processed in parallelith the samples, and the results (Fig. 2) strongly suggest a faultless assay for theNA-adduct analysis.
When the DNA-adduct level was below the detection limit, the value was set toalf of the background. The results were expressed as nmol DNA adducts per mol
ormal nucleotides (nmol/mol). Liver DNA adducts differ from other biomarkers inhat every value above a certain limit is indicative of a toxic effect, since the naturalccurrence is below this limit. At least in many fish, the limit is 1 nmol/mol, wheniver DNA adducts are measured in the way described here [34].
ig. 2. Typical autoradiograms for the negative (A) and positive (B and C) controlaterial used to assure the quality of all the analytical steps in the 32P-postlabellingethod for the measurement of DNA adducts: (A) pure salmon-sperm DNA, (B) the
tandard DNA adduct B[a]PDE-dG-3′p, and (C) liver tissue from benzo[a]pyrene-xposed perch (Perca fluviatilis). The three standards were processed in parallel withhe samples, and the results strongly suggest a faultless assay for the DNA-adductnalysis.
pyrene, dibenzo[a,h]anthracene, and benzo[ghi]perylene.ow the detection limit (0.001 �g/g dw), the value was set to half the detection limit
2.5. Analysis of micronucleated erythrocytes
The blood smears were stained with 4% Giemsa until the erythrocyte nuclei weresufficiently coloured (≥10 min). The slides were rinsed in distilled water, air-dried,and coded for blind scoring by a single observer, with a Leitz DMRBE microscope at1000× magnification. The frequency of micronucleated erythrocytes was scored for10,000 erythrocytes in each investigated specimen. Only cells with intact cellularand nuclear membranes were included. The criteria for scoring a micronucleus were:(1) round or ovoid shape; (2) size smaller than or equal to one-third of the size of thecell nucleus; and (3) micronucleus clearly detached from the cell nucleus, but withcolour and structure similar to it. A typical micronucleated herring gull erythrocyte,fulfilling the criteria, is shown in Fig. 3.
2.6. Supporting data
Data on PAH concentrations in surface sediment (0–2 cm) from the Swedishcoast were provided by the Geological Survey of Sweden (SGU). The sampling andPAH analysis are described by Sánchez-García et al. [27].
Fig. 3. A typical micronucleus (arrow) in an erythrocyte of an adult herring gull(Larus argentatus). The slide was stained with Giemsa.
H. Skarphedinsdottir et al. / Mutation Research 702 (2010) 24–31 27
Fig. 4. Liver DNA-adducts in adult herring gulls (Larus argentatus) from Swedenand Iceland (control). Only the County of Skåne (urban site) differed from Iceland(f
Wrn9QWta9
3
3
tIhftDwS(IbobPctaawtPdoi
oiouf
Fig. 5. Liver, kidney, intestinal mucosa, and whole blood DNA adducts in a sub-sample (see Section 2.4) of adult herring gulls (Larus argentatus) from the Countyof Skåne (urban site) and Iceland (control). The box illustrates the quartiles andthe whiskers extend to the most extreme values that are not outliers, which
P = 0.0090, regression model with the regions represented as dummy variables,ollowed by the Wald test). Bars: mean; error bars: 95% confidence intervals.
ilcoxon–Mann–Whitney test (with exact P-values) and the Spearman rank cor-elation. Data were analysed with parametric methods when suitable, and withon-parametric methods otherwise. All significance tests were performed at the5% significance level (˛ = 0.05). In the box plots, the box illustrates the quartiles:1, Q2, and Q3. Two fences are defined as Q1 − 1.5x(Q3 − Q1) and Q3 + 1.5x(Q3 − Q1).hiskers are drawn extending from the ends of the box to the most extreme values
hat are still inside the fences. Observations that fall outside the fences are regardeds possible outliers and are indicated by dots [35]. The software Intercooled Stata.2 (StataCorp LP, College Station, TX, USA) was used for the analyses.
. Results
.1. DNA adducts
In adult herring gulls, liver DNA adducts were detectable in allhe 25 Swedish specimens, and in 15 of the 17 Icelandic specimens.n all specimens with detectable liver DNA adducts the level wasigher than 1 nmol/mol. Individual liver DNA adduct values ranged
rom 1.5 to 73 nmol/mol in the Swedish material and from unde-ectable to 45 nmol/mol in the Icelandic material. The highest liverNA adduct levels were found in the urban site County of Skåne,hich differed from Iceland (Fig. 4; P = 0.0090). The rural County of
ödermanland and County of Blekinge did not differ from IcelandFig. 4; P > 0.05). In the sub-sample from the County of Skåne andceland, DNA adducts were most abundant in the liver, followedy the kidney, intestinal mucosa, and whole blood in decreasingrder (Fig. 5). DNA-adduct levels in the intestinal mucosa and wholelood were higher in the County of Skåne than in Iceland (Fig. 5;= 0.017, P = 0.0079), and there was a strong tendency towards aorresponding difference for the kidney (Fig. 5; P = 0.13). In Sweden,here was a correlation between liver DNA adducts and kidney DNAdducts (Pearson r = 0.73, P = 0.038, n = 19). The occurrence of DNAdducts in the four tissues was unrelated to sex and/or condition, asell as to the occurrence of hepatomegaly and/or paralytic symp-
oms observed in some of the investigated specimens (AN(C)OVA,> 0.05, not shown). The typical DNA-adduct pattern was a diffuseiagonal radioactive zone (DRZ), indicative of a broad spectrumf DNA adducts, but there were also some more distinct spots,ndicative of particular DNA adducts (Fig. 6).
In herring gull pulli, liver DNA adducts were detectable in 5
f the 7 Swedish specimens, and in none of the 6 Icelandic spec-mens. In the rural Swedish sites County of Värmland and Countyf Södermanland, individual liver DNA-adduct levels ranged fromndetectable to 4.9 nmol/mol. The County of Värmland differedrom Iceland (Fig. 7; P = 0.024), and there was a tendency towards
are indicated by dots [35]. Intestinal mucosa, and whole blood DNA adductlevels differed between the County of Skåne and Iceland (P = 0.017, P = 0.0079,Wilcoxon–Mann–Whitney test), and there was a tendency towards a similar dif-ference for kidney DNA adducts (P = 0.13).
a corresponding difference between the County of Södermanlandand Iceland (Fig. 7; P = 0.11).
3.2. Micronucleated erythrocytes
In adult herring gulls, individual values of the frequency ofmicronucleated erythrocytes ranged from 0 to 0.8‰ in both Swedenand in Iceland. None of the Swedish regions differed from Iceland(Fig. 8; P > 0.05). There was no correlation between the occurrenceof micronucleated erythrocytes and DNA adducts in any organ(Spearman rank correlation, P > 0.05). Neither was the frequencyof micronucleated erythrocytes related to sex or condition, nor tothe occurrence of hepatomegaly or paralytic symptoms observedin some of the investigated specimens (AN(C)OVA, P > 0.05, notshown).
4. Discussion
4.1. DNA adducts
Under the assumption that liver DNA-adduct values above1 nmol/mol are indicative of a genotoxic effect [34], the resultsshow that adult herring gulls in both Sweden and Iceland areexposed to genotoxic agents. The typical DRZ pattern, observedin many specimens, has been reported previously in other verte-brate species exposed to complex mixtures of PAHs from knownand unknown sources [e.g. 20,36]. In Sweden there was a positivecovariation between the level of liver DNA adducts and the PAHconcentration in nearby surface sediments (Table 1). The highestliver DNA-adduct levels and the highest surface-sediment PAH con-centrations, were both found in the County of Skåne, the urban site.The rural sites in Sweden had higher PAH concentrations in surfacesediment than a rural site in Iceland, whereas the PAH concentra-tions in the blue mussel (Mytilus edulis) differed in the oppositedirection (Table 1). The liver DNA adduct levels in herring gull pulli
indicated a lower degree of exposure to genotoxic agents in Ice-landic rural sites than in Swedish rural sites. Although partiallyambiguous, these data allow us to conclude that there is a signif-icant background exposure to PAHs, which is difficult to predict,both in Sweden and Iceland.
28 H. Skarphedinsdottir et al. / Mutation Research 702 (2010) 24–31
F ntestif duct l
dtstcamnto
rcp
ig. 6. Typical autoradiograms of 32P-postlabelled DNA adducts in the liver, kidney, irom County of Skåne (urban site), Sweden (A and B) and Iceland (C and D). DNA ad
Significant background exposure to genotoxic agents has beenemonstrated previously in herring gulls [7] and wild brookrout (Salvelinus fontinalis) [37] from remote areas without knownources of pollution, except atmospheric deposition. The impor-ance of atmospheric transfer of genotoxic pollutants has beenonfirmed in a study of laboratory mice exposed in situ to ambientir in a polluted industrial area near steel mills [38] with docu-ented genotoxic effects on herring gulls [5,6]. The occurrence of
on-repaired DNA adducts both in the liver and brain of the brookrout was interpreted as a result of long-time exposure to low levels
f genotoxic pollutants [37].
The large range of liver DNA-adduct levels in the adult her-ing gulls, both in Sweden and Iceland, indicates that there may beonfounding factors. Possible confounding variables are age, foodreferences, and local migration. Herring gulls may become more
nal mucosa, and whole blood of four individual adult herring gulls (Larus argentatus)evels are expressed as nmol DNA adducts per mol normal nucleotides (nmol/mol).
than 20 years old, and liver DNA adducts may accumulate withtime. Some DNA adducts are very persistent, and a positive rela-tionship between age and the occurrence of DNA adducts has beendemonstrated [39]. The herring gull is an omnivore, and analysisof its stomach contents has revealed that it feeds on fish, molluscs,crustaceans, birds, bird eggs, small mammals, insects, and vegeta-tion [40–43]. Some herring gulls follow fishing boats, and may thusbe exposed to pollutants in the smoke plume, whereas other feedon refuse dumps. Both on the European mainland and in Iceland,the herring gull is a non-migrant bird, although some juvenile her-
ring gulls may migrate from Iceland to the coast of Western Europeduring the winter [44,45]. This rules out the possibility of varia-tion in exposure due to seasonal migration, but does not excludevariation in exposure due to local migration, e.g. between rural andurban areas.
H. Skarphedinsdottir et al. / Mutation Research 702 (2010) 24–31 29
Fig. 7. Liver DNA-adducts in herring gull (Larus argentatus) pulli from Swe-den and Iceland (control). The box illustrates the quartiles and the whiskersedWl
hDtegmmaiDm(c
Fig. 8. Frequency of micronucleated erythrocytes in adult herring gulls (Larus argen-
TL
N
TM
xtend to the most extreme values that are not outliers, which are indicated byots [35]. The County of Värmland (rural site) differed from Iceland (P = 0.024,ilcoxon–Mann–Whitney test), and there was an obvious tendency towards a simi-
ar difference between the County of Södermanland (rural site) and Iceland (P = 0.11).
To our knowledge, only two studies on DNA adducts in birdsave been published previously. Schilderman et al. [46] analysedNA adducts in the kidney, liver and lung of city pigeons in cen-
ral Amsterdam, Maastricht, and Assen (The Netherlands). Østbyt al. [47] analysed DNA adducts in the liver of laboratory-hatchedlaucous gulls (Larus hyperboreus) fed with field-collected, environ-entally polluted herring gull and great black-backed gull (Larusarinus) eggs from Vardø in northern Norway. The liver DNA-
dduct levels in the pigeons [46] were similar to the levels found
n Swedish and Icelandic adult herring gulls, whereas the liverNA-adduct levels in the glaucous gulls [47] were 1–2 orders ofagnitude higher, both in the exposed group and the control group
Table 2). The results of the latter study are obviously not directlyomparable with the other results presented in Table 2. Also in
.D. = not detectable.a Range of the means of the investigated groups.b Investigations performed by the same laboratory, i.e. directly comparable values.c Range of the medians of the investigated groups.
able 3icronucleated erythrocytes (MNE) in birds and fish.
Country/sea area Type of animal Species
Sweden/Iceland Adult bird Herring gull (Larus argentatus)Mexico Adult bird 38 speciesLithuania Bird embryo Black-headed gull (Larus ridibundus)Germany Adult fish Eelpout (Zoarces viviparous)Baltic Proper Adult fish Flounder (Platichthys flesus)Baltic Proper Adult fish Perch (Perca fluviatilis)Baltic Proper Adult fish Eelpout (Zoarces viviparous)
a Range of the means of the investigated groups.
tatus) from Sweden and Iceland (control). None of the Swedish regions (one urbanand two rural sites) differed from Iceland (P > 0.05, regression model with the regionsrepresented as dummy variables, followed by the Wald test). Bars: mean; error bars:95% confidence intervals.
general, comparisons of DNA-adduct levels measured by differ-ent laboratories should be made with caution. For example, theremay be differences in the quantification of DNA. In our study, DNAwas quantified spectrophotometrically by its absorption at 260 nm,whereas other laboratories may quantify the DNA by labelling ofthe normal DNA nucleotides. The latter method results in a higherestimate of the DNA-adduct level, because the labelling efficiencyof normal nucleotides is generally only 25–50%.
Table 2 also presents liver DNA-adduct levels in fish, whichwere analysed at our laboratory, and thus are directly compara-ble with the liver DNA-adduct levels in the herring gulls. Liver
DNA adducts seem to be more abundant in the herring gullsthan in the fish, especially since many of the herring gulls camefrom rural areas, whereas most of the fish came from pollutedareas. Differences in metabolising enzyme systems, DNA repair,turnover of cells, feeding, digestion, and age are all factors that
Type of exposure DNA add. [nmol/mol] Reference
Urban and rural sites 9.3–30a This studyb
Rural sites N.D.–3.2c This studyb
Central Amsterdam, Maastricht and Assen 7.74–12.04a [46]Environmentally polluted food (eggs) 1070–1250a [47]Unexposed control specimens 500–810a [47]PAH gradient (aluminium smelter) 2–29a [20]b
PAH gradient (aluminium smelter) 3.3–13.5a [20]b
Creosote contaminated site 1–7a [48]b
Pulp mill effluent polluted site 12–22a [49]b
Unexposed control sites 0.5–2a [49]b
Active fishing harbours 14–33a [50]b
Type of exposure MNEa [‰] Reference
Urban and rural coastal sites 0.18–0.28 This studyZoological garden and other captivity 0–1.58 [51,52]A semi-urban and a rural site 0.057–4.7 [53–55]Urban and semi-urban coastal sites 0.04–0.66 [56]Urban and rural coastal sites 0.12–1.45 [57]Urban and rural coastal sites 0.45–1.15 [57]Urban and rural coastal sites 0.02–0.55 [57]
3 utatio
ms
wssoTtt
ob
4
mvowsor[1waeepttmiipt1
iteoobth
tmttc
C
A
ClT
[[
[
[
[
[
[
[
[
[
[
[
[
[
[
0 H. Skarphedinsdottir et al. / M
ay contribute to the differences in DNA-adduct levels betweenpecies.
In the city pigeons [46] the highest levels of DNA adductsere found in the liver and kidney, whereas lower levels were
een in the lung. Differences in DNA-adduct levels between tis-ues may be caused by differences in the abundance and activityf activating and deactivating xenobiotic metabolising enzymes.he DNA-adduct levels also depend on, e.g. the type of exposure,he uptake of the pollutants, the efficiency of DNA repair, and theurnover of cells.
The many informative results of the DNA-adduct analysisbtained here add to the ample evidence of the utility of thisiomarker.
.2. Micronucleated erythrocytes
To our knowledge, only a few studies on the frequency oficronucleated erythrocytes in birds have been published pre-
iously. Zúniga-Gonzáles et al. [51,52] measured the frequencyf micronucleated erythrocytes in 38 species of captive birds,hich were all claimed to be healthy. Stoncius et al. [53–55] mea-
ured the frequency of micronucleated erythrocytes in embryosf black-headed gulls (Chroicocephalus ridibundus) in two natu-al populations in Lithuania. In the 38 species of captive birds51,52], the means of the investigated groups ranged from 0 to.58‰ (Table 3). In 13 species, no micronucleated erythrocytesere found at all [51,52]. The median frequency was 0.05‰,
nd only a few species had high frequencies [51,52]. Differ-nces between species may be caused, e.g. by differences in thefficiency of the elimination of micronucleated erythrocytes. Com-arison between species should thus be made with caution. Inhe black-headed gull embryos [53–55] the means of the inves-igated groups ranged from 0.057 to 4.7‰. High frequencies of
icronucleated erythrocytes were found in embryos from eggsncubated less then 13 days, and low frequencies were foundn embryos from eggs incubated longer [53–55]. The authorsropose that this difference is explained by onset of elimina-ion of micronucleated erythrocytes by the spleen around day3.
The similar frequencies of micronucleated erythrocytes in all thenvestigated herring gulls make it difficult to know if they representhe natural background frequency in this species or are an effect ofxposure to genotoxic agents. The results fall within the rangesbtained in other birds and fish (Table 3). The very low frequenciesf micronucleated erythrocytes in most bird species investigatedy Zúniga-Gonzáles et al. [51,52] make it more probable, however,hat the observed frequency of micronucleated erythrocytes in theerring gulls is caused by background exposure to genotoxic agents.
Our results indicate that the frequency of micronucleated ery-hrocytes is a useful biomarker in birds, although this frequency
ay be lower than in other animal classes. Blood samples are rela-ively easy to obtain and allow non-invasive sampling. The responseo genotoxic agents may, however, be lower than for other types ofells, since mature erythrocytes do not divide [58].
onflict of interest
None.
cknowledgements
We are thankful to Kenneth Bengtsson (County of Skåne), Learlsson (County of Blekinge), and Hallgrimur V. Gunnarsson (Ice-
and) for help with the collection of the herring gulls in the field.he study was funded by Formas, The Swedish Research Council,
[
[
n Research 702 (2010) 24–31
and Signhild Engkvists Stiftelse och Stiftelsen Olle Engkvist Byg-gmästare.
References
[1] L. Karlsson, S. Ehnbom, G. Walinder, A comparison between ringing totals atFalsterbo, SW Sweden, ringing totals at Ottenby, SE Sweden, and point countsfrom the Swedish breeding bird census during 20 years (1980–1999), OrnisSvec. 15 (2005) 183–205.
[2] R. Ottvall, L. Edenius, J. Elmberg, H. Engström, M. Green, N. Holmquist, Å.Lindström, M. Tjernberg, T. Pärt, Population trends for bird species breed-ing in Sweden, Report 5813, Swedish Environmental Protection Agency(Naturvårdsverket), Stockholm, Sweden (2008) 123 pp. (In Swedish withEnglish summary).
[3] M. Hario, J.P. Hirvi, T. Hollmen, E. Rudback, Organochlorine concentrations indiseased vs. healthy gull chicks from the northern Baltic, Environ. Pollut. 127(2004) 411–423.
[4] L. Balk, P.-Å. Hägerroth, G. Åkerman, M. Hanson, U. Tjärnlund, T. Hansson, G.T.Hallgrimsson, Y. Zebühr, D. Broman, T. Mörner, H. Sundberg, Wild birds ofdeclining European species are dying from a thiamine deficiency syndrome,Proc. Natl. Acad. Sci. U.S.A. 106 (2009) 12001–12006.
[5] C.L. Yauk, J.S. Quinn, Multilocus DNA fingerprinting reveals high rate of heri-table genetic mutation in herring gulls nesting in an industrialized urban site,Proc. Natl. Acad. Sci. U.S.A. 93 (1996) 12137–12141.
[6] C.L. Yauk, G.A. Fox, B.E. McCarryc, J.S. Quinn, Induced minisatellite germlinemutations in herring gulls (Larus argentatus) living near steel mills, Mutat. Res.452 (2000) 211–218.
[7] G.A. Fox, P.A. White, S. Trudeau, C. Theodorakis, L.J. Shutt, S.W. Kennedy, K.J. Fer-nie, DNA strand length and EROD activity in relation to two screening measuresof genotoxic exposure in great lakes herring gulls, Ecotoxicology 14 (2005)527–544.
[8] M. Fitzmaurice, The new Helsinki convention on the protection of the marineenvironment of the Baltic Sea area, Mar. Pollut. Bull. 26 (1993) 64–67.
[9] B.-O. Jansson, K. Dahlberg, The environmental status of the Baltic Sea in the1940s, today and in the future, Ambio 28 (1999) 312–319.
10] B. Kurelec, The genetic disease syndrome, Mar. Environ. Res. 35 (1993) 341–348.11] M.P. Lesser, J.H. Farrell, C.W. Walker, Oxidative stress, DNA damage and p53
expression in the larvae of Atlantic cod (Gadus morhua) exposed to ultraviolet(290–400 nm) radiation, J. Exp. Biol. 204 (2001) 157–164.
12] M.S. Myers, L.L. Johnson, T. Holm, T.K. Collier, J.E. Stein, U. Varanasi, Toxico-pathic hepatic lesions in subadult English sole (Pleuronectes vetulus) from PugetSound, Washington, USA: relationships with other biomarkers of contaminantexposure, Mar. Environ. Res. 45 (1998) 47–67.
13] W.L. Rose, R.M. Nisbet, P.G. Green, S. Norris, T. Fan, E.H. Smith, G.N. Cherr, S.L.Anderson, Using an integrated approach to link biomarker responses and phys-iological stress to growth impairment of cadmium-exposed larval topsmelt,Aquat. Toxicol. 80 (2006) 298–308.
14] W.J. Sadinski, G. Levay, M.C. Wilson, J.R. Hoffman, W.J. Bodell, S.L. Ander-son, Relationship among DNA-adducts, micronuclei, and fitness parameters inXenopus laevis exposed to benzo(a)pyrene, Aquat. Toxicol. 32 (1995) 333–352.
15] C.W. Theodorakis, Integration of genotoxic and population genetic endpointsin biomonitoring and risk assessment, Ecotoxicology 10 (2001) 245–256.
16] B.P. Dunn, J.J. Black, A. Maccubbin, 32P-postlabeling analysis of aromatic DNAadducts in fish from polluted areas, Cancer Res. 47 (1987) 6543–6548.
17] G. Ericson, E. Noaksson, L. Balk, DNA adduct formation and persistence in liverand extrahepatic tissues of northern pike (Esox lucius) following oral expo-sure to benzo[a]pyrene, benzo[k]fluoranthene and 7H-dibenzo[c,g]carbazole,Mutat. Res. 427 (1999) 135–145.
18] J.E. Stein, T.K. Collier, W.L. Reichert, E. Casillas, T. Hom, U. Varanasi, Bioindicatorsof contaminant exposure and sublethal effects in benthic fish from Puget Sound,WA, USA, Mar. Environ. Res. 35 (1993) 95–100.
19] L.R. Shugart, Environmental genotoxicology, in: G.M. Rand (Ed.), Aquatic Tox-icology. Effects, Environmental Fate, and Risk Assessment, Taylor & Francis,North Palm Beach, FL, USA, 1995, pp. 405–419.
20] G. Ericson, E. Lindesjöö, L. Balk, DNA adducts and histopathological lesions inperch (Perca fluviatilis) and northern pike (Esox lucius) along a polycyclic aro-matic hydrocarbon gradient on the coastline of the Swedish Baltic Sea, Can. J.Fish. Aquat. Sci. 55 (1998) 815–824.
21] U. Varanasi, W.L. Reichert, J.E. Stein, 32P-postlabelling analysis of DNA adductsin liver of wild English Sole (Parophrys vetulus) and Winter Flounder (Pseudo-pleuronectes americanus), Cancer Res. 49 (1989) 1171–1177.
22] G. Ericson, H. Skarphedinsdottir, L. Dalla Zuanna, J. Svavarsson, DNA adductsas indicators of genotoxic exposure in indigenous and transplanted mussels,Mytilus edulis, L. from Icelandic coastal sites, Mutat. Res. 516 (2002) 91–99.
23] H. Skarphedinsdottir, G. Ericson, J. Svavarsson, K. Naes, DNA adducts and poly-cyclic aromatic hydrocarbon (PAH) tissue levels in blue mussels (Mytilus spp.)from Nordic coastal sites, Mar. Environ. Res. 64 (2007) 479–491.
24] J.A. Heddle, M.C. Cimino, M. Hayashi, F. Romagna, M.D. Shelby, J.D. Tucker, P.Vanparys, J.T. Macgregor, Micronuclei as an index of cytogenetic damage—past,
present and future, Environ. Mol. Mutagen. 18 (1991) 277–291.
25] G. Wirzinger, L. Weltje, J. Gercken, H. Sordyl, Genotoxic damage in field-collected three-spined sticklebacks (Gasterosteus aculeatus L.): a suitablebiomonitoring tool? Mutat. Res. 628 (2007) 19–30.
26] F. Ayllon, E. Garcia-Vazquez, Induction of micronuclei and other nuclearabnormalities in European minnow Phoxinus phoxinus and mollie Poecilia latip-
utatio
[
[
[
[
[
[
[
[
[
[
[
[
[
[
[
[
[
[
[
[
[
[
[
[
[
[
[
[
[
[
H. Skarphedinsdottir et al. / M
inna: an assessment of the fish micronucleus test, Mutat. Res. 467 (2000)177–186.
27] L. Sánchez-García, I. Cato, Ö. Gustafsson, Evaluation of the influence of blackcarbon on the distribution of PAHs in sediments from along the entire Swedishcontinental shelf, Mar. Chem. 119 (2010) 44–51.
28] H. Guðfinnsson, H. Valdimarsson, J. Briem, S. Jónsson, J. Ólafsson, S. Ólafsdóttir,Á. Gíslason, S.A. Steingrímsson, Current Measurements, Environmental Factorsand Biology of Reyðarfjörður in the Period Late July to the Beginning of October2000, Marine Research Institute (Hafrannsóknastofnunin), Reykjavík, Iceland,2001, 133 pp. (In Icelandic with English summary).
29] J.-L. Vives Corrons, S. Albarède, G. Flandrin, S. Heller, K. Horvath, B. Houwen,G. Nordin, E. Sarkani, M. Skitek, M. Van Blerk, J.-C. Libeer, Guidelines for bloodsmear preparation and staining procedure for setting up an external qualityassessment scheme for blood smear interpretation. Part I: control material,Clin. Chem. Lab. Med. 42 (2004) 922–926.
30] W.L. Reichert, B.L. French, The 32P-postlabeling protocols for assaying levels ofhydrophobic DNA adducts in fish, Report, NOAA Tech. Memo. NMFS-NWFSC-14, National Technical Information Service, Springfield, VA, USA, 1994, 89 pp.
31] G. Ericson, L. Balk, DNA adduct formation in northern pike (Esox lucius)exposed to a mixture of benzo[a]pyrene, benzo[k]fluoranthene and 7H-dibenzo[c,g]carbazole: time-course and dose-response studies, Mutat. Res. 454(2000) 11–20.
32] M.V. Reddy, K. Randerath, Nuclease P1-mediated enhancement of sensitivityof 32P-postlabeling test for structurally diverse DNA adducts, Carcinogenesis 7(1986) 1543–1551.
33] W.L. Reichert, J.E. Stein, B. French, P. Goodwin, U. Varanasi, Storage phosphorimaging technique for detection and quantitation of DNA adducts measured bythe 32P-postlabelling assay, Carcinogenesis 13 (1992) 1475–1479.
34] E. Aas, B. Liewenborg, B.E. Grosvik, L. Camus, G. Jonsson, J.F. Borseth, L. Balk, DNAadduct levels in fish from pristine areas are not detectable or low when analysedusing the nuclease P1 version of the 32P-postlabelling technique, Biomarkers 8(2003) 445–460.
35] A.C. Tamhane, D.D. Dunlop, Statistics and Data Analysis—From Elementary toIntermediate, Prentice Hall, Upper Saddle River, NJ, USA, 2000, 722 pp.
36] B.L. French, W.L. Reichert, T. Hom, M. Nishimoto, H.R. Sanborn, J.E. Stein,Accumulation and dose-response of hepatic DNA adducts in English sole (Pleu-ronectes vetulus) exposed to a gradient of contaminated sediments, Aquat.Toxicol. 36 (1996) 1–16.
37] S. Ray, T. Bieger, D.A. Scruton, 32P-postlabeling analysis of aromatic DNAadducts in liver and brain of wild brook trout (Salvelinus fontinalis), Chemo-sphere 30 (1995) 773–778.
38] C.M. Somers, C.L. Yauk, P.A. White, C.L.J. Parfett, J.S. Quinn, Air pollution inducesheritable DNA mutations, Proc. Natl. Acad. Sci. U.S.A. 99 (2002) 15904–15907.
39] W.L. Reichert, B.L. French, J.E. Stein, Exposure of marine mammals to geno-toxic environmental contaminants: application of the 32P-postlabeling assayfor measuring DNA-xenobiotic adducts, Environ. Monit. Assess. 56 (1999)225–239.
40] L. Hillstrom, M. Kilpi, K. Lindstrom, Diet of herring gulls Larus argentatus duringchick rearing in the Gulf of Finland, Ornis Fenn. 71 (1994) 95–101.
41] M. Kilpi, L. Hillstrom, K. Lindstrom, Egg-size variation and reproductive successin the herring gull Larus argentatus: adaptive or constrained size of the last egg?Ibis 138 (1996) 212–217.
42] U. Kubetzki, S. Garthe, Distribution, diet and habitat selection by four sympatri-cally breeding gull species in the south-eastern North Sea, Mar. Biol. 143 (2003)199–207.
[
[
n Research 702 (2010) 24–31 31
43] R. Noordhuis, A.L. Spaans, Interspecific competition for food between herringLarus argentatus and lesser black-backed gulls L. Fuscus in the Dutch WaddenSea area, Ardea 80 (1992) 114–132.
44] M. Kilpi, Breeding biology of the herring gull Larus argentatus in the northernBaltic, Ornis Fenn. 67 (1990) 130–140.
45] M. Kilpi, Responses of herring gulls Larus argentatus and common gulls L. canusto warm years—early migration and early breeding, Ornis Fenn. 69 (1992)82–87.
46] P.A.E.L. Schilderman, J.A. Hoogewerff, F.J. van Schooten, L.M. Maas, E.J.C. Moo-nen, B.J.H. van Os, J.H. vanWijnen, J.C.S. Kleinjans, Possible relevance of pigeonsas an indicator species for monitoring air pollution, Environ. Health Perspect.105 (1997) 322–330.
47] L. Østby, G.W. Gabrielsen, Å. Krøkje, Cytochrome P4501A induction and DNAadduct formation in glaucous gulls (Larus hyperboreus), fed with environmen-tally contaminated gull eggs, Ecotoxicol. Environ. Saf. 62 (2005) 363–375.
48] G. Ericson, B. Liewenborg, E. Lindesjöö, C. Näf, L. Balk, DNA adducts in perch(Perca fluviatilis) from a creosote contaminated site in the River Ångermanäl-ven, Sweden, Aquat. Toxicol. 45 (1999) 181–193.
49] G. Ericson, Å. Larsson, DNA adducts in perch (Perca fluviatilis) living in a coastalwater polluted with bleached pulp mill effluents, Ecotoxicol. Environ. Saf. 46(2000) 167–173.
50] E. Stephensen, J. Svavarsson, J. Sturve, G. Ericson, M. Adolfsson-Erici, L. Forlin,Biochemical indicators of pollution exposure in shorthorn sculpin (Myoxo-cephalus scorpius), caught in four harbours on the southwest coast of Iceland,Aquat. Toxicol. 48 (2000) 431–442.
51] G. Zúniga-González, O. Torres-Bugarín, J. Luna-Aguirre, A. González-Rodríguez,A. Zamora-Perez, B.C. Gómez-Meda, A.J. Ventura-Aguilar, M.L. Ramos-Ibarra,A. Ramos-Mora, G.G. Ortíz, M.P. Gallegos-Arreola, Spontaneous micronuclei inperipheral blood erythrocytes from 54 animal species (mammals, reptiles andbirds): Part two, Mutat. Res. 467 (2000) 99–103.
52] G. Zúniga-González, O. Torres-Bugarín, A. Zamora-Perez, B.C. Gómez-Meda,M.L. Ramos-Ibarra, S. Martínez-González, A. González-Rodríguez, J. Luna-Aguirre, A. Ramos-Mora, D. Ontiveros-Lira, M.P. Gallegos-Arreola, Differencesin the number of micronucleated erythrocytes among young and adult ani-mals including humans—spontaneous micronuclei in 43 species, Mutat. Res.494 (2001) 161–167.
53] D. Stoncius, J.R. Lazutka, Spontaneous and benzo[a]pyrene-induced micronu-clei in the embryos of the black-headed gull (Larus ridibundus L.), Mutat. Res.538 (2003) 31–39.
54] D. Stoncius, S. Sinkevicius, Spontaneous micronuclei in the black-headed gull(Larus ridibundus L.) embryos in relation to parameters determining nestlingsurvival, Acta Zool. Lit. 13 (2003) 294–298.
55] D. Stoncius, Spontaneous micronuclei in embryos of the black-headed gull(Larus ridibundus L.) populations, Ekologija 1 (2003) 63–66.
56] D. Schiedek, K. Broeg, J. Barsiene, K.K. Lehtonen, J. Gercken, S. Pfeifer, H. Vuon-tisjärvi, P.J. Vuorinen, V. Dedonyte, A. Koehler, L. Balk, R. Schneider, Biomarkerresponses as indication of contaminant effects in blue mussel (Mytilus edulis)and female eelpout (Zoarces viviparus) from the southwestern Baltic Sea, Mar.Pollut. Bull. 53 (2006) 387–405.
57] J. Barsiene, V. Dedonyte, A. Rybakovas, K. Broeg, L. Forlin, J. Gercken, J. Kopecka,L. Balk, Environmental mutagenesis in different zones of the Baltic Sea, ActaZool. Lit. 15 (2005) 90–95.
58] K.R. Carrasco, K.L. Tilbury, M.S. Myers, Assessment of the piscine micronucleustest as an in situ biological indicator of chemical contaminant effects, Can. J.Fish. Aquat. Sci. 47 (1990) 2123–2136.
