BIOMARKERS OF EXPOSURE OF BROWN BULLHEADS (AMEIURUS NEBULOSUS) TO CONTAMlNANTS

IN THE LOWER GREAT LAKES

A thesis submitted to the Cornmittee on Graduate Studies

in Partial Fulfilment of the Requirements for the degree of Master of Science

in the Faculty of Arts and Science

TRENT UNIVERSITY Peterborough, Ontario, Canada

BY

Q Lba Dm Arcand, WB6

Watershed Ecosystems MSc. Program

May, 1997

Acquisitions and Bibliographie Senrices

Acquisiins et secvices biMiographques

The author bas gmted a nonexc1usive licence dowing the

L'auteur a accordé une licence non exclusive permettant B la Bibliothèqye nationale du Canada de

Nationai h i of Canada to reprodiiire,pr&ta, distn'buerou reproduce, loan, distniute or seJî copies ofMer thesis by any mems vendredescopiesdesath&se& and in any form or fmslong qgelqye manière a sous q u e I v foane que ce soit pour mettre des exemplaires de cette thèse a la disposition des personnes intéressées.

The author retains ownership ofthe copyright m M e r thesis. Neither the thesis nor substantial erdracts ikom it may be printed or otherwise reproduced with the author's permission.

L'auteia conserve la propiété du droit d'autem qui protège sa thèse. Ni la thèse nides adraits substantiels de celle-ci ne d o m être imprimés ou

Biornarker techniques were applied to brown bullheads from contaminated regions in the lower Great Lakes (i.e. Detroit Rier, Black River and Hamilton Harbour) and mmpared to the same parameters measured in bullheads from the relatively dean sites of Old Woman Creek and the Bay of Quinte. Bullheads h m the contaminated sites showed elevated liver somatic indices (LSI), increased ethoxyresorufin4eethylase (EROD) activty, decreasedhepatic retinoid stores and a greater incidence of hepatic neoplasms. (FA&)

Fluorescent aromatic compounds

in the bile were also elevated in bullheads h m the Black River and

Hamilton Harbour in camparison to fish collected from the reference sites of Old Woman Creek and the Bay of Quinte. Bullheads from the Bla& River also showed

an increased incidenœ of micronudeated hepatocytes. The data showed that the above biomarker techniques were sensitive indicators of exposure to aromatic contaminants in the lower Great Lakes. Biomarker data also suggested that bullheads from the Detroit R i e r may have been exposed to a greater degree of environmental contamination than bullheads from the Black River and Hamilton Harbour.

Principal component analysis (PCA) dernonstrated that biomarker

patterns in Detrol River bullheads were different from Black River and Hamilton Harbour bullheads; bullheads from these latter two sites showed a more similar ordination. In addition, biomarker patterns in bullheads Rom Old Woman Creek and the Bay of Quinte were very similar and much different than those found in bullhead

populations ftom the three mntaminated sites. The overall aim of this project was to evaluatethe useof biomarkertechniques for monitoringin situ exposure of brown

bullheads to aromatic wmpounds in the Great Lakes. The suite of biomarker parameters used in this study could serve as usefi~lmonitoring tools and aid in establishing benchmarks of contamination in aie amas of study. As remedial actions pro-

in these contaminatecl regions. biomarker values measured in

bullheads should approach those of reference site fish.

ACKNOWLEDGEMENTS

There are many people tha deserve acknowledgement fw their help and support in this research project First of all. i would like to thank my thesis supervisor Dr. Chris Metcalfe for his ongoing support and insight throughout this project I am ako greathl to cornmittee members Dr. DarrellTomkins and Dr. Tom Whillans for their support and cornilents on this manuscript. Thanks is also ofkred to Erin Bennett and Peter Hoy for their assistance in the field and laboratory; to Dr.

Kelly Munkittrick and Dr. Paul Baumann for their help with fish collection and use of laboratory faalioes. and to Scott Brown for allowhg me to vise his lab to perform

retinoid analysis. This projed was funded through a Tri-Council Eco-Research Grant administered by the McMaster Ecoresearch Program.

TABLE OF CONTENTS

ACKNOWLEDGEMENTS

TABLE OF CONTENTS LlST OF FIGURES LlST OF TABLES LIST OF PLATES

INTRODUCTION Overview Site Descriptions Histopathology Metabolism of PAHs Genotoxic response Induction of MFOs FluorescentAromatic Cornpounds (FACs) in bile Retinoid metabolism Tocopherol (vitamin E) Projed goalslobjectiies

METHODS Sampling sites and fish capture Sacrifice, tissue removal and storage Spine aging Liver histopathobgy Micronucleus assay Injection of allyl formate Liver perbsion and hepatocyte suspension Staining Micronucleus scon'ng procedure EROD analysis Preparationof pms (S9) and microsomes 2.6.2. ER00 assay 2.6.3 Protein measurement 2.7 Bile FAC analysis 2.8 Retinoid and Tocopherol analysis

viii

2.8.1 2.8.2

2.8.3 2.9 2.9.1 2.9.2 2.1 0

Standard preparation Tissue homogenization and extraction HPLC analysis Laboratory EKposures Fish capture and husbandry Exposun to sediment Statstical aaaalysh

Resultr Age and liver somatic indices Histopathology Extemal lesions Grossly visible hepatic lesions Liver pathology Hepatocytic alterations Hepatic cholangiocyüc lesions Hepatic neoplastic iesions Melanomacrophagecounts Micronucleus Induction EROD induction Bile FACs Laboratory Exposureto Hamilton Harbour Sediment Hepatic Retinoids Tocopherol Analysis PCA of biomarker responses Discussion Exposure of bullheads to aromatic compounds Histopathology Micronucleus Assay EROD lnducction Bile FACs Reünoid Analysis Tocopherol Analysis PCA /Site Cornparisons Contaminated Sites Reference Sites 5.0

Conclusions

7.0

Appendices A. Raw data (age. length. weight. sex. liver weight and gonad weight) for individual bullheads collecteci from the five study sites (spring and faII 1994).

B. Hepatic micronucieus/binuckate data for individual bullheads C. Melanomacrophage counb

D. Histopathological reagents and procedures. E. Hepatic EROD data (standard curves and EROD adivity) in individual bullheads collected from the five study sites F. Bile FAC concentrations in individual bullheads collected from the five study sites. G. Hepatic retinoid data (reagents, standard curves and hepatic retinoid concentrations) for individual fish colleded Rom the five study sites. H Median values of biomarker data.

LIST OF FIGURES Figure 2.1

Brown bullhead collection sites in the lower Great Lakes

Figure 3.1 -

Comparison of hepatic EROD activîty in individual brown bullheads h m al1 sites (faIl 1994). except Hamilton Harbour. derived from assays using pms and microsomes.

Figure 3.2

HPLC chromatogram of bile FAC analysis.

Figure 3.3

Mean EROD adivity and concentration of FACs in bile of brown bullheads sacrificed at 0.4 and 12 days after 72 hours exposure to Hamilton Harbour sediment.

Figure 3.4

HPLC chromatogram of retinoid analysis showing a) retinoid standards and b) retinoid compounds detected in the liver of a brown bullhead ftom Old Woman Creek.

Figure 3.5

Mean concentration of hepatic retinyl palmitate and dehydroretinol in brown bullheads from the lower Great Lakes.

Figure 3.6

Mean concentration of hepatic dehydroretinol esters 1 and 2 in brown bullheads fiom contarninated and relatively clean regions of the lower Great Lakes.

Figure 3-7

Relationship between mean hepatic retinyl palmitate and mean hepatic EROD actiiity in brown bullheads wllected from 5 sites in the lower Great Lakes.

Figure 3.8

Relationship between hepatic retinyl palmitate and hepatic EROD adivity in individual bullheads from the lower Great Lakes.

Figure 3.9

Relationship between mean hepadic dehydroredinol and hepatic EROD activi in bullheads collected Rom five sites in the lower Great Lakes.

Figure 3.10 Relationship between hepatic dehydroretinol and hepatic EROD activity in individual bullheads collecteci h m the lower Great Lakes. Figure 3.1 1 Principal component analysis of biomarker data (EROD acüvity, bile FACs. LSI and dehydroretinol) in individual bullheads colleded fiom five study sites in the lower Great takes.

LIST OF TABLES Table 3.1

Age and live?somatic index (mean & std) of selected (3-5 years of age) brown bullheads wllected in the spmg and fall of 1994 from the lower Great Lakes,

Table 3.2.

Epidennal and hepatic pathology in brown bullheads collected fiom the lower Great Lakes.

Table 3.3

The number of micronucleated hepatocytes and binucleated hepatocytes (rnean & ad) in bullheads collected from the Black River and Old Woman Creek (spring 1994) and from Hamilton Harbour and the Bay of Quinte (fall 1994).

Table 3.4

Mean hepatic EROD activity in brom bullheads (measured using both hepatic pms and microsomes) collected from the lower Great Lakes in the spring and fall of 1994.

Table 3-5

Concentration of bile FACs measured in brown bullheads measured befote and after alterations to chromatographie conditions.

Table 3.6

Mean FACs in bile of brown bullheads fiom contaminated and relatively clean sites in the lower Great Lakes.

54

Component loadings and percent of total variance explained by the first and second principal components.

64

Table 3.7

LIST OF PLATES Plate 3.1 Plate 3.2 Plate 3.3

A chobngioma in tk liwr of a brawn bullhead collected fram the Detroit Rier (fall 1994).

43

A cholangiocarcinornafound in the liver of a brown bullhead Rom the Bladr River (spring 1994).

44

A hepatocellularcarcinoma fwnd in the Iiver of a brown bullhead from the Black River (spring 1994).

44

1.O Introduction 1 . ûvewiew

The heatth of aquatic organisms in the Great Lakes may be campromised as

a result of past and present discharges of environmental contaminants. For instance. it is suspected that chemical contaminants are responsible for the high prevalenœ of tumours in Great Lakes benthic fish (Harshbarger and Clark, 1990). Sediments in many industrialüed areas are heavily polluted with organic compounds such as polychlofinated biphenyb (PCBs), polynuclear aromatic hydrocarbons (PAHs) and organochlorine pesticides. as well as heavy metals (Furlong et al.. 1988; MetaIfe et al., 1990; Fabacher et al.. 1991; Baumann et al., 1991).

Conventional research has relied upon analytical data on tissue body

burdens and chemical concentrations in water and d i m e n t to assess the possible toxic impacts of these cornpounds in the Great Lakes. These analflcal methods

are extrernely sensitive and provide quantitative information on specifÏc contaminants. but they have several shortcomings. In paraailar, chemical analysis provides little useful information for compounds that are rapidly metabolized and excreted by vertebrate organisms, such as PAHs- Scientists and regulatory agencies are beginning to develop and utilize indicators of organism health to complement traditional chemical analysis, and to provide more rapid and inexpensive approaches for monitoring toxic impacts of pollutants in the aquatic environment"Biomarkers," sometimes rekrred to as bioindicators, are defined as

physiologid,biochemical andlor histopathological alterations that occur as a result of exposure to environmental pollutants (Huggetî et al., 1992). Biomerker studies

involve sampling organisms from a contaminateci site and comparing them to the same species collecteci at an ecologically similar, but relatively uncontaminated reference site. Typically, the biomarker approach involves analysis of multiple parameters to assess exposure and toxic impads of environmental contaminants. This is necessary because there is currently only a limitecl understanding of the variables that influence biomarkers. A multiple parameter approach allow for a more comprehensive assessrnent of chemical exposures and impacts. Biomarkers are indicators of the cumulative exposure of organisms as they inhabit areas with changing contaminant patterns. Varying degrees of exposure of fish can be attributed to species mobility, changes in industrial oualows and physical disturbances (Varanasi et al, 1989). lndicators can also be measured at various levels of biologicalorganization. For example, biochemical aiterations can be eady indicators of environmental change, whereas histopathological alterations may be indicatorsof long-terni exposure to chemical pollutants. Thus, a biomarker approach that incorporates muitiple parameters over several levels of biological organization is recommended to demonstrate exposure of fish to chemical contaminants. Numerous biomarker techniques have been proposed and utilized in the laboratory and more recentiy in field studies, but further investigation is essential to develop and validate these methods. Despite the potential benefits of biomarkers for evaluating in siru exposure of fish to contaminants, there are limitations to these techniques. Typically, biomarker data for fish are highly variable. Some sources

of variability indude swes

reproductive status, sex, age, diet and mobility

(McCarthy and Shuggart I W O ) . This variabili is most obvious when fish occupy areas where the distribution of contaminants is not uniforni and exposure at time of colledion may deviate from mean exposure intsg-

with the. Thus, in order to

assess cumulative efkts. it is important to use some biomarker parameters that reflect chronic exposure. It should also be noted that some biomarkers are general indicators of exposure to chemical contaminants and these parameters cannot be used to identiry exposure to specific classes of chemicais. Limitations on the use of general indicators is not a reason b reject biomarker data. but indicates a need to understand, validate and interpret many of aiese indicator responses in order to develop the techniques to their full potential.

A suite of genotoxic, biochemical

and histopathological biomarkers were uülized in this study. The biotransfonnation of chemical compounds in mammab, birds and teleosts most often occurs through a cornplex enzyme system of cytochrome P450 associated monooxygenases (MFOs). Exposure to xenobiotic ampounds results in the induction of MFOs. More specifically. cytochrome P45O1A (CYP4501A) is induced by exposure to environmental contaminants such as PAHs, planar PCBs, halogenated dibenzodioxins and furans and pulp miIl effluent (Payne et al., 1987; Munkittrick et al., 1991; Sleiderink et al., 1995; Bucheli and Fent, 1995). Ethoxyreso~n-CMeethylase (EROD) acüvity often is used as a biochemical marker of CYP4501A atiiity in fish.

Aromatic hydrocarbons (AH) are transfomieci by MF0 enzymes into polar metabolites that c m be excreted into the bile (Varanasi et al, 1989). The detecüon of fluorescent metabolites of aromatic hydmcarbons in biliary fluid indicates recent

3

exposure to aromatic compounds (Krahn et al.. 1984). Retinoid (vitamin A) metabolism in mammals and birds is also aitered by exposure to planat aromatic contaminants such as PCBs and dioxins (Spear et al., 1989,1992). Aithough there are IimRed data on the e W of ammatic hydrocarbons

on retinoids in fish (Palace and Brown, 1994), there is potentialfor using aiterations to retinoid metabolism as a bioindicator of contaminant exposure.

Some PAH metabolitesare reactive compounds that bind to DNA and cause genetic damage. Genotoxic effects include DNA damage, gene mutations and chromosomal abenations. There are currentiy few reliable and efficient techniques for assessing genotoxic responses in aquatic o ganisms. Recently,a piscine hepatic micronucleus test was shown to be a sensitive indicator of genotoxicity in rainbow trout, Oncohynchus mykiss (VVïlliarns and Metcalfe, 1992). In this study, the piscine hepatic micronucleus assay was used as a biomarker of in situ exposure to genotoxic wntaminants. In addition, histopathological analysis of hepatocytic and cholangiocytic elements of fish her was used to a

m the rh situ exposure of fish

to environmental carcinogens. The overall aim of this project was to evaluate the use of biomarker

techniques for monitoring in situ exposure of benthic fish to aromatic contaminants in the Great Lakes. Reœnt research has shown that biomarkers can be used to

monitor manne fish species (Krahn et al., 1984; Collier and Varanasi, 1991; Stein et al., 1991; Eggens et al., 1995; Sleiderink et al., 1995). In this study, biomarker techniques were applied to populations of b w n bullheads (Ameiutus nebulosus);

a benthic teieost distributeci throughout the kwer Great Lakes. The biomarker data in bullheads inhabithg several Great Lakes "Areas of Conœm" knom to have high sediment burdens of PAHs (Le.Hamilton Harbour, Wmiï River. Black River) were compared to the data from bullheads captured from two relatively clean referenœ sites (Le. Bay of Quinte, Old Wornan Creek).

1.2 Site Descriptions

There is a large magnitude of difbrenœ in the type and degree of environmental contamination p-nt

in Hamiiton Harbour, the Detroit River and the

Black River in comparison to the chosen referenœ sites of Old Woman Creek and the Bay of Quinte. For example, Hamilton Harbour is a highly industrialized area

that receives discharge of aromatic hydrocarbons predominantly Rom two steel

mills. Concentrations of PAHs, PCBs and metals in Hamilton Harbour greatly exceed sediment quality guidelines (Krantzberg and Boyd, 1992). In addition, benzo[a]pyrene, a csrcinogenic PAH was found at 4.78 uglg in sediment along the south east shore (Metcab et al., 1988). Sediment extracts from this region induœd hepatic neoplasms in rainbow trout (Metcave et al., 1988; Balch et al., 1995b). It

is important to note that contaminants in the harbour region are not unifomly distributed and that bullheads used in this study were collected outside the highly contaminated main harbour region. Similar to Hamilton Harbour, sediments in the Black River are mntarninated with aromatic compounds. Once again, the steel industry is a major source of

contaminant discharge in this region. Liver tumours in bullheads Born this region have been well documenteci over the last decade (Baumann et al., 1987, 1990, 1996). Recently, Baumann and Harshbarger (1995) reported a dedine in sediment

PAH concentration and Iiver neoplasms in bullheads aRer the closing of the USX

coking facility in 1983 and subsequent dedine in steel miIl producüon. For exampie, concentrations of total PAHs in sediment from the Black River decreased from 1096 uglg dry sediment in 1980. to 382.2 uglg dry sediment in 7982. By 1987 total PAH compounds in the sediment had dropped to 4.27 ug/g dry sediment (Baumann and Harshbarger, 1995).

The Detroit River is a highly industrialized connecting channel to the lower Great Lakes, which serves as a source of contaminants to the western basin of Lake Erie. Over 200 chernical compounds have k e n identifid in this river system; more specificaliy in the Trenton Channel, a major depositional zone bordering the

United States (Maccubbin and Ersing, 1991). Concentrations of total PAHs and total PCBs measured in channel sediment were as high as 130 uglg and 13 uglg dry weight respectiveiy. Because of the dificuky in collecting fish in this

highly contaminateci region, bullheads used in this study were collected just west of this channel.

The comparative or reference sites chosen for this study are relatively clean

in cornparison to the contaminateci study sites. For example, Old Woman Creek is a National Estuarine Sanctuary located about 32 km miles west of the Black River, and receives no industrial point source discharges. It has often served as a

referenœ site for bullhead tumour studies in the Black River (Baumann et al., 1987, 6

1Q9O). Hawever, PAH contaminetion atbauted to creosotson railway ties, may be

present in a small kcalbed region of the creek (Baumann et al.. 1996). Bullheads were also collecteci hwn Hay Bay, an inlet of the Bay of Quinte in eastern Lake Ontario (refened to as Bay of Quinte bullheads). Bullheads in this region, as well as perch and walleye are an invaluable resource to area commercial fishennan. Aîthough, Bay of Quinte is by no means a pristine site, impairment in this region is

on a much different scale from that in the Detroit River, Hamilton Harbour and the Black River.

1.3 Histopathology

Epizootiological data over the last decade have documented high prevalences of epithelial and hepatic tumours in brown bullheads from industrialized regions (Baumann et al., 1987, 1990. 1996; Harshbarger and Clark 1990; Maccubbin and Ersing 1991). Baumann et al. (1990) reported that 30% of bullheads collected from the Black Riîer had grossly visible liver tumours. Histopathological analysis of the same fish revealed biliary and hepatic lesions, with cancerous hepatic neoplasms occurMg in 38% of the fish. In addition, skin painting of bullheads with sediment extract from the Black and Buffalo Rivers induced epidermal hyperplasiaand papillomas (Black et al., 1985). Bullheads inhabithg the Detrol River demonstrated an elevated prevalence of liier cancer (Maccubbin and Ersing, 1991a). Hayes et al. (1990) reporteâ high prevalences of both epidemal papillomas and demal carcinomas in bullheads from Hamilton Harbour. In addition,

laboratory exposure of rainbow bout to extracts fiom Hamilton Harbour sedirnent induced hepatoceliular carailornas (MetcaHc et al.. 'i988,IggO).

The hepatic

lesions commoniy reported in the above dudies and also surveyed in this study include: hepatocel!ular alteradions (hepatic ceIl necrosis, basophilie and eosinophilic foci); cholangiolar alterations (anaplasticîhyperplastic bile duct epithelium and cholangiofibrosis);

and

hepatic

neoplastic

lesions

(cholangioma,

cholangiocarcinoma and hepatacellular carcinoma). Grossly visible extemal lesions include skin lesions. oral papillomas and morphologically altered (tnincated) barbels.

1.4 Metabolism of PAHs

PAHs are transformed by MF0 enzymes through oxidation and hydroxylation to more polar metabolites that can be excreted in the bile. However, metabolic transformation is also responsible for the conversion of PAHs to readive intemediates that bind to DDNA and exert mutagenic and carcinogenic effects (Conney 1982; Varanasi et al., 1989; Sikka et al., 1WOb). One of the most potent PAH compounds, benzo[a]pyrene (BaP), has been used widely to demonstrate the

role of biotransformation in the activation of aromabic hydrocarbons. BaP is a well known mammalian carcinogen and its carcinogenic properties have been demonstrated in teleosts (Hendncks et al., 1985, Metcalfe et al., 1988).

BaP is oxidized by cytochrome P450 enzymes to produce arene oxides,

w hich are further transfomied to hydroxyderatives. These intemediates can be

conjugated with glucuronide, gluWhione and sulfate moieties (Gelboin, 1980). Conjugation ads as a detoxifying mechanism mat increases polarity and enhances excretion in bile. However, as stated previousiy, aromaüc cornpounds may also be activated by biotransfomation

enzymes. Of concem is the pathway whereby BaP-

7,û-epoxide is converted to BaP-ï,&dihydrodiol which can be furthsr transfomed into the ultimate carcinogenicfomw of cis- and trans-6aP-7,84ioI-Q. 1Oepoxides. BaP-7,8 dihydradiol (precursor to BaP-7.8diol-9.10epoxide) was the major metabolite found in bullheads exposed to BaP in the laboratory and DNA adduds weie also found in aiese fish (Sikka et al., 1990a, 1990b). Dunn et al. (1987) found sïmilar DNA adducts in bullheads captured from the Detroit and Buffalo Rivers. Unrepaired DNA adducts can result in a number of detrimental effects, including

lesions that can lead to cytotoxicity, mutagenicity and carcinogenicity (Bresnick, 1985).

1.5

Genotoxic Responses Genotoxicity is a general terni used to describe damage to DNA. A nurnber

of tests have been employed to assess exposure of fish to genotoxic contaminants.

including the detection of DNA-adducts, the measurement of chromosomal aberrations using analysis of sister chromatid exchanges (SCE) and formation of micronuclei. Micronucleus formation can oGcur in any proliferating ceIl that has been exposed to a clastogenic agent. The test has been widely used for rodent

genotoxicity assays through examination of lymphocytes (Heddle et al., 1983) and

hepatocytes (Tates et al., 1980; Braithwaite and Ashby, 1988).

ModifÎed

micronucleus assays have b e n useâ with fish erythrocytes (Hooftman and Raat, 1982) and fish hepatoqtes (Williams and Metcalfe, 1992).

Compounds capable of inducing structural changes in chromosomes are referred to as clastogens.

In the presence of a clastogen. a portion of a

chromosome rnay fàil to migrate to the metaphase plate dunng anaphase, and fiil to be incorporatecl into the nucleus of the newiy bmed daughter cell. These extranuclear chromatin bodies which lie in the same focal plane as the principal nucleus are termed micronuclei and serve as an indicator of genotoxicity. Micronuclei are usually 115 to 1/20 the size of the main nucleus (Heddle, 1983). Larger chromatin masses may represent whole ~hromosomesthat were not incorporatedwith the nucleus as a result of a spindle apparatus malfundion, and thus do not represent true clastogenicity.

There are advantages to examining hepatocytes for micronuclei, as the Iiver is the primary site of biochemicaltransformationof compounds to reactive genotoxic intemiediates. However. in order to detect micronudeus formation, it is imperative that actively proliferating cells are analyred, as micronuclei are produced after mitosis. Since the liver is a relatively quiescent organ. proliferation must be araficially induced. The hepatic necrogen, aly( formate has been used to induce a regenerative response in fish and in vivo assays with rainbow trout showed that increased micronudeus frequencies were evident after exposure to ethyl methansulfonate (EMS), rnitornycin C (MMC) and diethylnitrosamine (DEN) (Williams and Metcab, 1992).

10

This assay is easily adaptableto h situ monitoring. However, there are many biological fadors that influenœ genotoxic responses that must be understood.

Species, age, sex, reproductivestatus, health and water temperature are believed to play a role in the expression of genotoxic effeds in fish (Al-Sabti and MetcaHe,

1995). In addition, inwnsistencies in identifying and scoring micronudei among researchers can influence data interpretation. An understanding of the factors affecting variability, and standardization of scoring critena are required to develop the hepatic micronucleus assay as a biomarker technique.

1.6 Induction of MF-

Biochemical alrations provide one of the earliest indicaton of exposure to environmental pollutants (Stegeman et al., 1992). The induction of MF0 activity (i.e. increased synthesis of CYP450 proteins) is a biochemical response that has been studied in mammab since the eariy 1950's (Conney, 1982). Today, MF0

activity in fish is used as a biochemical indicator of exposure to waterbome chemical contaminants (Payne et al., 1987; Collier et al., 1992; Haaschet al.. 1993; Eggens et al.. 1995). The specific MF0 enzyme complex CYP4501A is commonly

induced by planar organic compounds that interact with the Ah receptor such as PCBs, PAHs and dioxins (Kleinow et al., 1987; Payne et al., 1987). €muent from

pulp and paper mills is also an inducer of MF0 activity (Munkittrick et al., 1991;

Kloepper-Samsand Bsnton 1994). Induction of CYP4501A b commonly measured in the laboratory as an increase in ethoxyresorufin-Odeethylase (EROD) or aryl

hydrocarbon hydroxylase (AHH). A strong corea l to in

exists between these two

marker enzymes (Collier et al., 1992) and the ER00 assay is ncm more commonly used because of its sensiovii, ease and safhty (Varanasi et al. 1989). Interlaboratory comp~sonsof MF0 activity have been conducteci to help standardize and understand variability associated with tissue handling, enzyme fractions and instrumentation (Munkiidc et al., 1993). However, research is required to further understand and account for the environmental and biological factors that influence MF0 activity in situ. Elevated EROD activity has been reported in many marine and fkeshwater fish exposed to a variety of chernical ~ontaminants(Payne et al., 1987). For

example, winter flounder (heumnectes americanus) inhabithg a mal tar contaminated estuary displayed a seven-fold increase in EROD induction that closely mapped concentrations of PAHs in the sediment. (Mgneir et al., 1994). EROD induction has been demonstrated in mach (Rubtus M7us) and gizzard shad

(Domsoma cepedianum) ewperimentally exposed to /3-naphthoflavone and BaP, respectively ( Levine et al., 1994; O'Hste et al., 1995). Elevated MF0 activii has been observed in minor carp (CypMus carpio) exposed to BaP and chrysene (van der Weiden et al., 1994), bullheads exposed to aromatic hydrocarbons from Hamilton Harbour sediments (Balch et al., 1WSa), and channel catfish (Ictalunrs punctatus)and largemouth bass (Micmptenrs salmoides) caged in a PAH and P C E contaminated river (Haasch et al., 1993).

Conversely, EROD and associated

hepatic biochemical parameters were not found to be reliable indicators of exposure

to chernical contaminant$ in two in situ studies with bullheads (Fabacher and Baumann. 1985; Gallagher and Di Giulio. 1989) and with roach collecteci from polluted river systems (van der Oost et al., 1994). This poor MF0 response may have k e n due to an impaired or inhibitcd metabolic system as a resuit of ewposure to high contaminant levels (Fabacher and Baumann. 1985; Gallagher and Di Giulio, 7989). MF0 responses are dependent upon a number of biological and environmental factors such as species, sex, reproductive status, and temperature (Payne et al., 1987, Varanasi et al, 1989). The greatest influence on MF0 adivity appears to be reproductive status (Hodson et al., 1991). Jimenez et al. (1990) reported that MF0 activ'i in redbreast sunfish (Lepomisauritus) was greatest in the fall and lowest in females during spring spawning. White suckers exposed to bleached kraft miIl effluent (BKME) during spawning had lower MF0 responses than fish colleded from the same sites during the summer (Munkittrick et al., 1991). AHH activity in lake trout was also lowest during spawning (Luxon et al., 1987).

Fish age, size and sex did not influence AHH activity in lake trout (Luxon et al., 1987). but size was a factor in MF0 a c W i in brook trout (Addison and Willis.

1982). Caffish exposed to pnaphthoflavone in water temperatures of 9% had lmer MF0 induction than when temperatures of 2S0Cwere used (Haasch et al.. 1993). However, Sleiderink et al. (1995) demonstrated increased EROD activity in North

Sea dab (Limande limanda) at lower water temperatures, whereas Eggens et al. (1995) found no differenœ in EROD acbjvi in English sole (Pamphrys vetulus) held

at water temperatures of 4 O , 12' and 200 for 14 days. Thus, an understanding of factors that contribute to variabilii W of Ucmost importancewhen interpreüng MF0 responses.

1.7 FluorescentAromatic Compounds (FACs) in Bile

Aromatic compounds are reediiy metabolizeâ in the liver and excreted into the bile (Varanasi et al, 1989).

Therefore, chemical analysis of aromatic

hydrocarbons in fish tissues generally reveals levels that are near or below detection lima (Johnston and Baumann, 1989). Much work has been conducteâ on BaP metabolism in brown bullheads and individual metabolites in the bile (Sikka

.

et al., 1WOa, 1991; Steward et al., 1990). For example BaP-7,8diol has been

found to be a prominent metabolite in the bile of brown bullheads exposed to BaP (Sikka et al.. 1990). To dernonstrate the role of the hepatobiliary system in the

metabolism

of BaP, Steward et al. (1990) injected brown bullheads with

radiolabelleci BaP and anaîyzed bik and extrahepatic tissue after 24 and 72 hours of exposure. It was demonstrated that after 72 houn, the concentration of BaP-

associated radioactivity was 6-42 times higher in biliary fluid than in abdominal organs, demonstrating hepatobiliary secretion as the primary route of elimination.

The measurernent of xenobiotic compounds in the biliary fiuid of fish has

been proposeâ as a rnethod of monitoring exposure of fish to aromatic wmpounds (Krahn et al., 1984; Collier and Varanasi, 1991;Lin et al., 1994). Krahn et al. (1984) developed a relatively simple HPLC fluorescence technique to estimate the relative

concentrations of fluorescent aromaüc compounds (FACs) in bile. This prodoes not require sampîe deanup. nor does it require resoIuüon of individual

metabolites in order to obtain an estimate of the relative concenûations of FACs, HPLC chmmatograrns of fish bile collected from sites contaminated mth aromatic

hydrocarbons showed high concentrations of cornpiex matures of FACs (Krahn et al., 1984; Johnston and Baumann. 1989; Lin et al.. 1994). For example Krahn et ai. (1984) demonstrated that English sole captured from Puget Sound showed total fluorescence at wavelengais characteristic of naphthalene. phenanthrene and BaP that were 9-20 fold higher than fluorescence in bile from reference fish. Bullheads captureci and analyzed Rom the polluted Bladc and Cuyahoga Rivers demonstrated elevated bile fluorescence when wrnpared to reference fish (Johnston and Baumann, 1989; Lin et al., 1994). as did winter flounder caged in a polluted river system for 3 months (Beyer et al., 1994). FACs have been related to other biomarken used to assess the presence and impact of chernical contaminants. For example, concentrations of PAHs in the

sedirnent have been directly correlated with FACs in fish bile (Krahn et al., 1986; Johnson and Baumann, 1989). Liver lesions in fish have been associated with higher levels of biliary FACs (Krahn et al, 1984). Relationships between MF0 activity and bile FACs in winter flounder and English sole has also been reported (Collier and Varanasi, 1991). Most recently, research has demonstrated that the feeding status of the fish plays an important role in the relative conœntrations of FACs in the bile. Haasch et al. (1993) reported that adively feeding fish have lower FACs in the bile, as 15

dgiesoitn

of fwd causes movemnt of biliary ffuid. Lower FACs in adively feeding

fish has also been obsenred in captive English sole (Collier and Varanasi, 1991). Currently, bile FACs are considemi to be an indicator of relativeiy recent exposure to aromatic hydrocaibons.

1.8 Retinoid Metabolism Many physiological and developmental processes are dependent on

retinoids, induding differentiation and maintenance of epithelial cells, vision, growth and reproduction (Pawson, 1981). Vitamin A compounds occur in nature as retinol, retinal and retinoic acid. In anmals, ingested m r o t e n e is conveited to retinal and further reduced in a reversible reacüon in the intestine to retinol. Retinol is then

transported through the hepatic portal system and stored in the Iiver as retinol esters. The predominant storage ester is retinyl palmitate which can be rapidly convertecl back to retinol and released from the liver when required by target tissues. Exposure to environmental contaminants can result in reduced or altered metabolism of vitamin A and subsequent detrimental health effects (Zile, 1992) . Halogenateci aromatic compounds such as PCBs and dioxins are known to alter retinoid metabolism in mammals (Mercier et al., 1990) and birds (Spear et al., 1992).

Alteration of vitamin A metabolism in rodents exposed to 2,3,7,8-

tetrachloradibenzo-p-dioxin (TCDD) and PCBs have been extensively studied (Zile,

1992). More recently, studies have been directed at retinoid levels in wildlife exposed in situ to environmental contaminants. For example, herring gull colonies

in Lake Ontario with high tissue conœnûaüons of TCDD were %und to have correspondingly lower liver mün~f concentrations (Spear and Moon, 1986). More reœntly, the molar ratios of ren itoI

and tetinol palmitate in Great Lakes herring gull

eggs were correlatecl wiai concentrations and toxic equivalent factors of polychlorinated dibenzog-dioxin (PCDD) and polychlorinated dibenzo-pfuran (PCDF) in the eggs (Spear et al., 1990). Spear et al. (1992) also found a negative correlation between AHH induction and Iiver retinol in hennig gulls.

Much of the knowiedge of retinoids and their metabolkm has been derived from experimental studies with mammals and birds. However, there are a number of reports that pertain to dietary vitamin deficiencies in various fish species

(Tavekijakam et al., 1994). For example, cherry salmon (Oncorfiynchus masou) deprived of vitamin A in their diet showed reduced growth, truncated snouts, reddened fins and an increased mortality when compared to control fish (Tavekijakam et al., 1994). However, there has been Iimited research on retinoid levels in fish exposed to chernical contaminants (Palace and Brown, 1994). Recently, Spear et al. (1992) examined a population of white suckers from a contaminated river system and compared vitamin A levels to those fmm a relatively clean site. Resub showed that fish h m the contaminated site had lower levels of retinol and retinol palmitate. In addition lake char (Sabelinus namaycush) exposed to coplanar 3,3',4,4',5pentachlorobip~had reduced retinol, dehydroretinol and

retinyl palmitate concentrations in the liver (Palace and Brown, 1994). It has been determined that dehydroretinol is more predorninant than retinol in fieshwater fish than in marine fish and rnammals (Dyke, 1965, Guillou et al.. 1989). In addition, 17

seals f@dPCBladen fish dispfayed a large redudion in plasma retinol and thymxin (Brouwer et al.. 1989). These studies support the hypothesis derived h m marnmalian and avian studies that exposure of fish to halogenatecl aromatic contaminants andlor relateci compaunds aiters hepatic retinoid stores.

1.9 Tocopheiof (Viimin E)

Tocopherol (Wimin E) is an antioxidant vitamin. known for its role in protecting cellular and subcellular membranes. Studies îndicate that fish with low levels of vitamin E are more susceptible to disease.

For example, goldfish

(Carassius aumtus) raised in a bboratory and fed a diet with inadequate levels of tocopherol displayed deterioration of the tail and fins that led ta fungal infections and death (Huerkamp et al. 1988). In addition, rainbow trout fed a low vitamin E

diet demonstrated vitamin E depletion in liver microsornes and increased binding of BaP.7.8diol to D M ; perhaps suggesting an increased susceptibility to carcinogenesis (Williams et al.. 1992). An increase in vitamin E intake appears to enhanœ immunological responses. For example, Wise et al. (1993) reported that channel catfïsh (Ictalurus punctatus) fed a vitamin E-enriched diet displayed an enhanced immunological response to a bacterial pathogen. To date. few studies have been conducted to evaluate the potential of tocopherol as a bioindicator of

contaminant exposure in fish.

2.0 Goab and ObjBCfjVes The ovemll aim of this study mis to evaluate the use of multiple biomarkers to assess in situ exposure of benthic fish to aromatic contaminants in the Great Lakes. The brown bullhead is a good indicator species for this study because of its close contact with sediments and its ubiquitousdistribution. It is hypothesueâ that biomaiker responses in bullheads h m the contaminated regions should difler h m biomarkers in referenœ site bullheads. The abilii of these biomarker techniques to differentiatebetween the degree of contaminantexposure in bullhead populations from the contaminated study sites will also be assessed. The findings of this study

could be helpful in bench marking current exposure of brown bullheads to aromatic contaminants. This is important as Remedial Action Plans are being implemented

in Areas of Conœm in the Great Lakes. Biomarker responses should decline with rernedial action, and thus the suite of parameters may provide a useful means of monitoring the benefits of remediation with tirne.

2.0 METHODS 2.1 Sampling Sibs and Fbh Capture

Sampling sites in Lake Ontario and Lake Erie are illustrateci in Figure 2.1. Brown bullheads were collected from three contaminated sites in the lower Great Lakes (Black River, Detroit River and Hamibn Harbour) and compared to fish captured from two relatively ckan reference areas (Bay of Quinte and Old Woman

Creek).

Fig 2-1 Brown bullhead collection sites in the Lower Great Lakes. (referenœ sites are indicated by filied squares)

Black River and Old Woman Cr&

bullheads were captured in May and September

of 1994 with ovemight sets with a fyke net. Bullheads from the Black River were

collected along a 2 km stretch of the river bank occupied by steel miIl operations. Bullheads from Hamilton Harbour (September of 1994) and the Bay of Quinte (October 1994) were captured by trap net

Hamilton Harbour bullheads were

collected outside of the main harbour area, towards Cootes Paradise. Bay of Quinte bullheads were adually collected from Hay Bay, an infet of the Bay of Quinte. Detroit River bullheads were eledroshodced in October of 1994, just down river of the Trenton Channel. Approximately 20 fish were cdlected fmm each site

and a number of biomarker techniques were performed on each fish. At selected

sites, an additional 15-20 bullheads were captured and allyl formate (AF) 20

administered as part of the micronucleus assay. 2.2 Sacrifice, thmue mmoval and storage Fish were transported to a nearby labontory in well aerated holding tanks and sacrificed wiai an overdose of MS-222 (Sigma. St. Louis, MO.), fdlowed by cervical dsio lcao itn.

The fish were weighed and gross extemal abnormalities were

recorded. All bullheads were sexed, gonadal weight was recorded and the left pectoral spine was removed for aging. Bik was removed from the gall bladder with a 23-gauge disposable syringe and transfened to a Nalgene cryovial and

immediately placed in liquid nitmgen. The Iiverwas removed and quickly weighed. If the micronucleus test was k i n g conduded, the Iiver was perfused with a 1% trisodium citrate solution as an anti-coagulant (Sigma, S t Louis. MO.) and 1.5-2.0 g of liver was used immediately for this test. Perfusion was perfomed by

inserting the tip of the wash bottle containing the citrate solution into the hepatic artery opening on the dorsal side of the liver and flushing the liver until the tissue

tumed to a crearn cokur. Similar poftïons of the liver were excised and rinsed with

cold potassium chloride (KCI) and quick frozen in liquid nitrogen for EROD and vitamin analysis (retinoid and towpherol). In addition, the posterior kidney was removed and stored at -80°C for retinoid and tocopherol analysis. L i e r tissue was also collecteci for histopathology and placed in Bouins fixative. Bullheads that received AF injectionswere aged, sexed and proœssed for the micronucleus assay.

No other biomarker techniques were performed on the fish used in MN-AF assays.

2.3 Spine Aging Bullheadspines were dislocated fiom the pectoraljoint and placed in maked envelopes for drying. Spines were decalcified in 10%hydrochloric aciâ (HCI) for 20 houn and transfened to individual glas 5 mL viais containing 50% isopropyl alcohol. To detenine the age of the fish, a 1 mm thick cross-section of the spine was -oned

just below the outbranching node of the spine using a singleedged

razor blade. The thin sections were soaked in water until translucent and then exarnineâ under a dissecting microscope. The age of the fish was determineci by counting the number of winter growth periods as indicated by dark circular bands.

2.4 LIVER HISTOPATHOLOGY A small portion of the liver tissue (appmximateiy 1 gram) was sedioned from the whole Iiver, placed in a tissue cassette and placed in Bouins fixative for 24-48 hours. The tissues were then washed with ninning water to remove the fixative and

were transferred to 70% ethanol. Livers were embedded in paraffin wax (melting p0int=54.6~C)using a model T/P 8000 tissue processor (Amencan Optical). The schedule for embedding of tissues in paraffin is listed in Appendix D. Briefiy, tissue

cassettes were transferred through a seriss of increasing conœntmüons of ethanol

(70%, 90%, 100%) and 100% xylene prior to placement in liquid paraffin. Tissues were then blocked in parafin and sectioned at a thickness of 6 pm using a rotary microtome. Sections were laid in a 25OC water bath and adhered to a glass microscope slide with albumin fixative (BDH, Toronto) and air dried.

Paraffinwax-

removecimai 100% xyieneandtransfiened airough a series

of decreasing concentrations of ethanol(100%, 90%. 70%) and distilled water to rehydmte the tissues (Appendix D). SMes were plaœd in Elrichs haematoxylin (BDH. Toronto) for four minutes, washed mth ninning tap water for 10 minutes and stained with eosin (Fisher, NJ) for 1 minute.

Staining was completed with

placement in 95% and 100% ethanol, and 100% xylene for 1-2 minutes each i ns (Appendix D). Secto

were air dried, mounted with DPX resin and examined by

light micmscopy for the presenœ andlor absence of histopathological lesions.

Histopathological lesions were classified according to the criteria used by Hayes et al. (1990). Hinton and Lauren (1992) and Baumann et al. (1990, 1996). Lesions were grouped into three categoties: hepatWc

alterations (cell necrosis,

basophilie and eosinophilic foci); hepatic cholangiocytic alterations (anaplastic or

hyperpbstic bile du& epithelium, cholangiofibrosis); and hepatic neoplastic lesions (cholangiornas, cholangiocarcinomas, and hepatocellularcarcinomas). In addition, melanomacrophage œnters (MMC) were counted per square œntimetre of tissue section. This was done by placing a 1 cm2grid over the section of liver prepared for histopathological analysis and enumerating MMC present in the hepatic parenchyma and pancreatic tissue. All Iivers were processed for histopathologicalanalysis with the exception of liver tissue from bullheads collecteci in September of 1994 from the Black River and

Old Woman Creek. At this time, other remarchers invohred in a long-term tumour survey of bullheads from these regions, utilized a portion of the liven collecteci to

conduct their own histological examination.

23

2.5 MICRONUCLEUS ASSAY 2.5.1 Injection of Allyl FormaOe

Brown bullheads captured from the Black River (n=l3) and Old Woman

Creek (n=14) and fiom Hamiiton Harbour (n=20) and the Bay of Quinte (n=20) were injected intrapentoneally(ip.) with allyl forniate (ICN, Cleveland) at a dose of I O uL per 100 g of fish and then placed in well aerated holding tanks for 72 hours prior to sacrifice. As rnentioned previously, collecüon of fish occuned in May of 1994 for Black River and Old Woman Creek bullheads and in September of 1994 for bullheads h m Hamilton Harbour and the Bay of Quinte. Injection occurred on the posterior ventral surface of the fish, adjacent to the pectoral sphes using a 50 uL glass syringe (Hamilton) with a 23 gauge disposable needle. Allyl formate, a hepatic necrogen, was used to induce hepatocyte regeneration in these fish (Williams and Metcalfe, 1992). Bullheads (n=20 per site) were also wlleded from each site and processed for the micronucleus assay without injection of allyl formate. These bullheads were sacrificed on the day of collecüon.

2.5.2 Liver PerfUsion and Hepatocyb Suspension Liver perfusion and preparation of hepatocyfes suspensions were perfomed according to the method of Williams and Metcab (1992). Bullhead liver was removed and perfused with a 1%trisodium citrate solution as an anti-coagulant (Sigma, St. Louis, MO). Perfusion was done by inserting the tip of a wash bottle containing the citrate solution into the hepatic artery opening on the dorsal side of

the liver. The Iiver tumed a aeam ookur when it was successfully perfused. Approm'mately, 1.5 to 2.0 g of tissue was exased from the whole liver, gentiy diced and washed with additional citrate. The section of Iher was placed in a Nalgene centrifuge tube with 57 mL of collagenase solution (Collagenase. Sigma Chernical, St Louis, MO). Collagenase was used to stnp hepatocytes h m connecüve cellular components. The mixture was vortexed every 4-5 minutes for a total of 30 minutes or until the suspension became cloudy. The supernatant was then removed and

centrifuged at high speed (3540 x g) using a table top centrifuge (ICE) for 3 minutes. At the end of centrifugation, the supernatant was discarded and the rernaining pellet containing hepatocytes and other cell types was resuspended in 1-2 drops of the 1% citrate solution. The cell mixture was drawn up into a pasteur pipette and a drop of liquid (size of a dime) was lowered ont0 the slide and then drawn back up into the pipette. This procedure left only a thin layer of the hepatocyte suspension on the slide. The slides were dried with the aid of a slidewaner and transfked to Camoy's fixative for one hour. SIides were air-dried ovemight priorto staining.

2.5.3 Staining

Slides were stained with Schiff's reagent (BDH, Toronto). This staining technique is specific to DNA and has been used successfully in previous micronucleus applications (Tates et al., 1980, Braithwaite and Ashby, 1988, Williams and MetcaL, 1992). Slides were placed for one hour in a 5 N HCI solution

and transfened to a 1 N HCI solution for 15 minutes. Placement in HCI soiutions was done to degrade nuckic acid covalent bonds. SIides were then placed in SchifPs reagent for one hour and then neutralized in a series of three potassium metabisurne baths (10%) for a total of nine minutes. Sfidea were then washed with

distilled water and counterstained with 2% light green (in 70% ethanol) for two minutes. Finally, the slides were rinsed in distilled water, dried and mounted with DPX.

2.5.4 Micmnucleus Scoring Procedure Slides were examined by light microscopy using an Olympus BHS microscope with oil immersion at 1500x magnification. All scoring was done blind. Prior to cell enumerations, it was necessary to differentiate hepatocytes from other blood cells and debris present on the slide. Erythrocytes were numerous in al1 cell preparations, but were easily distinguished from hepatocytes due to their smaller cell and nuclear size. Some monocytes were present. but were fewer in number than hepatocytes and erythracytes and were easiiy identified by their large cell size and granular cytoplasm. The cytoplasm of hepatocytes was stained unifomly.

Criteria for swing micronuclei were established prior to enurneration. Similar to the technique used by Wlliams and Metcalfe (1992), only isolated hepatocyteswith intact cellular and nuclear membranes were counted. In addition, only micronucleithat were 115 to 1RO the size of the principal nucleus, unrefractive to light. located in the same focal plane and unattached Rom the principal nucleus

were counted as true mimnudei. The nmber of micronuchted hepatocy&eswas counted in 1000 hepatocytes scored per fish. Cornparisons were made between MN ftequencies in Black River and Old Woman Creek bullheads and between

Hamilton Harbour and Bay of Quinte bullheads exposed and not exposed to allyl formate.

2.6 EROD ANALYSIS The EROD assay was performed with both the post-mitochondrial

supernatant (pms) and rnicrosomes. The assay using the microsomal preparations

was perfomed at Trent University whereas the pms EROD assay was performed at the Department of Fisheries and Oœans (DFO), Burlington, Ontario. The

protocols used at both locations were very similar, with the exœption of instrumentation and protein analysis.

Methods outlined below apply to the

microsomal EROD assay performed at Trent University. Protocols for protein analysis and instrumentation used at the Department of Fisheries and Oceans are listed in Appendïx E.

2.6.1 Preparaüon of pms (Sa) and Micmomes

Cryovials containing approximately 1-2 g of h z e n liver were thawed on ice. The tissue was then washed with KCI to remove exœss blood. Approximately 1.5 g of liver tissue was weighed, and HEPES KCI buffer, (HEPES, Sigma, St.

Louis, MO,

KCI), was added in a 4:1 liver weight ratio.

The tissue was

homogenized with five sfrokes of a Potter-Ehrehjelm teflon homogenizer. The homogenatewas cenbifiiged at 9 000 x g for 20 minutes to precipitate protein (pms fracüon). The precipitateor pms was transkned to Nalgene cyoviah and stored at -80°C for 1-2 weeks. To isolate the microsomes, the precipitate was thawed on ice

and then ultracentriruged for one hour at 105 000 x g. After centrifugation, the

supernatant was discardeci and the remaining pellet was scraped h m the side of the Nalgenecentrifugetube and mhed with 1 mL of &orage b d k r consisting of Tris (hydroxymethyl) aminomethane, EDTA, dithiothreitol and 20% w/v glycerol. This

was once again homogenized using a srnall teflon Potter-Elvehjelm homogenizer and the remaining microsomes w r e divided into two portions; one was frozen at 80°C for replication purposes and the other was placed on ice and used

immediately in the EROD assay (microsomal preparation). All tissue homogenization and centrifugation was perfomed at 4OC. and a11 vials and tubes were kept on iœ throughout the procedures.

Two replicates were prepared for each sample, along with a blank. To each glass centrifuge tube (13 mm x 100 mm). the following chernicab were added: 1300 uL Hepes buffer, 20 uL MgS04, 40 uL BSA, 30 uL NADPH and 50 uL

microsomal suspension. The solutions were voftexed immediately after the microsomes were added and tubes were placed in a water bath for 5 minutes at

To prepare madion blanks. 3.2 mL of methanol was pipetted into one sample tube per fish and vortexed. mis was followed by the addition of 30 uL of 7-ethoxyreso~n (?-ER) (Sigma, S t Louis. MO.) into the same tube, followed by vortexing. The 7-ER is substrate for the reaction and rnethanol was used to end the readion.

Subsequently, reacüon tubes were prepared by adding these

substances in reverse order. For example, using a stop watch, 30 uL of 7-ER was pipetted into each reacüon tube at 15 second intervals and vortexed. At 10 minutes

after the first addition of 7-ER, the same procedure was followed by the addition of 3.2 rnL of methanol at 15 second intervak. Thus, total reaction time was 10

minutes. The readion and blank reagent tubes were centrifugeci for 15 minutes at

100 000 x g to pellet the protein. The clear supernatant was rernoved with a pipette and transferred into glass reagent tubes specific for the fluororneter. A Turner

mode1111 fluorometer equipped wiVi a green Turner Wli0-854 bulb was used and the fluorescence was measured at 3x. All values were read and the sample

fluorescencewas determined by subtracting the blank value from the average of the reaction tubes. Standard curves of protein and resolufin adivity were determined for both the fluorometer and the spectrophotorneter (Appendix E).

ER00 activity was deterrnined by:

EROD ACTlVlTY = pmol resorufinlmg protein x readion time

2.6.3. Protein meaiumment Protein was measured using the Bio-Rad assay (Bradford, 1976). using bovine senirn albumin as a standard. The microsornes were diluted 193 with lio itn. distilled water. From that du

50 uL of r n i ~ ~ ~ s o rpreparation nal was added to

5 mL of a Bio-Rad reagent diluted 4:1 with distilleci water. Absorbame was

measured using a spectrophotometer (Spedronic 21, Baush and Lomb) set at 585 nm.

2.7 BILE FAC ANALYSIS

BaP equivalents in the bile of brown bullhead were measured by high performance liquid chromatography (HPLC) with a Waters 600E controller and a Waters 470 fluorescence detector. Bile was thawed on iœ and 5 uL was injected through a 2 rnL injection loop onto a Supelco LC-PAH column (i-d. 4.6 mm, 30 cm length). The solvents used were HPLC grade water wîth 5% glacial aœtic acid, and 100% methanol. The program gradient used for most samples was; 1.O0 mumin of 100% water to 1.O0 mUmin of 100% methanol in 15 minutes, followed by seven

minutes of 100% methanol at 1.00 mUrnin, three minutes of 100% water at 1.5 mumin, followed by 100% water for the duration of the 35 minute nin. Total fluorescence (area under the peaks) was measured at excitation/emission wavelengths of 3801430 nm and compared to fluorescent peak areas of BaP standards (0.36 pglml).

Bile FAC concentration were reported as ng BaP

equivalents per mL of bile. Chromatogramswere anaiyzed with 810 Waters HPLC

software. A new LGPAH d u m n (i-d. 4.6 mm, 30 an length) was installed late in the

study and due to dinierences in badc pressure between the two columns, a slightly new soivent gradient had to be developed. The program gradient commenced with 1.00 mUmin of 100%water to 1-00mL of 100% methanol in 15 minutes, This was

followed by 1-5 mL of methanol for 20 minutes, followed by five minutes of 100% water at 1.00 mL. A similar procedore was used by Lin et al. (1994). Consistency between methods was tested by re-anaiyzing selected samples using the modifiecl gradient.

2.8 RETINOID and TOCOPHEROL ANALYSIS

All analysis of retinoids and tocopherol took place at the Freshwater Institute of the Department of Fisheries and Oœans,Winnipeg, Manitoba. LNer and kidney were analyzed for the presenœ of vitamin A and E compounds including retinol,

dehydroretinol, retinyl palmitate and tocopherol, according to the methodsdescribed by Palace and Brown (1994).

2.8.1

Standard Prepantion A complete set of standards was prepared with each extraction series. Each

standard series contained retinol, dehydroretinol, retinyl palmitate and tocopherol. Protocols for standard preparation and the concentrations used are listed in Appendix G. Stock solutions of the intemal standards retinyl acetate (RA) and

tocopherol aœtate (TA) were prepared in ethanol and dluted to obtain concentrations of 50 ng RAI20 uL and 400 ng TA120 uL.

2.8.2 'Tissue Homogenbtion and Extraction

Liver and kidney tissue (100 mg) were chipped away from m e n samples and placed in a cdd test tube and weighed. A 2.0 mL volume of coid distilled water was added to the tissue and polytron homogenizeâ (on iœ) with a small probe at a speed setting of 10 for 10-15 seconds. The homagenate was vortexed and 200 uL was transferred to a 1.5 mL Eppendorf tube. A 200 uL volume of HPLC

grade ethanol containing the intemal standards was added and the via1 was capped and vortexed and set aside for 15 minutes ta precipitatethe proteins. In addition, 200 uL of the intemal standard was placed diredly into an empty amber

autosampler injector (ASI) via1 and drîed. This was done for every eight samples processed. After the proteins had precipitated, the tubes were centrifUged for ten

seconds to clean the sides. The samples were extracted with the addition of 500 uL of ethyl aœtate:hexane (3:2)and the mixture was capped, vottexed and allowed to stand for 60 minutes in the dark at room temperature. Tubes were then centrifuged for an additional 3 W 5 seconds and 250 uL of the top layer was transferred into a 0.7 mL amber AS1 via1with a conical bottom. Vials were placed

in a Speed Vac and evaporated to complete dryness (approximately one hour). Vials were capped and stored at 80°C until HPLC analysis.

2.8.3 HPLC analysis. Retinoid and tocopherol anatysis was perfomed by HPLC using a General Electric (GE) mode1 704 system conbolkr and a GE model 620 data module. Detedors used included a GE mode1 116dual channel UV absorbanœ detector, a GE 420 UV detector and a Waters 470 flwmmeter. W wavelengths were set at 292 nm for tocopherol and tocopherol acetate and et 325 nm fw dehydroretinol and

dehydroretinol esters. The fluororneter was set at excitationlemission wavelengths of 330 nm and 480 nm for retinol, retinol acetate and retinyl palmitate detedon.

AS1 vials of samples and standards were removed ftom the freezer and 20 uL of the mobile phase (aœtonitrile:methanol:water, 7O:2O:l O,vhr/v) was added and vortexed. Samples were injected through a 20 uL sample loop ont0 a bpm-bead

sire Adsorbosphere HS Cla wlumn (i.d. 4.6 mm, 150 mm length) with a guard column. Colurnn temperature was maintained at 26% and samples and standards were eluted isocratically with the mobile phase at a fiow rate of 1.0 mumin for 15

minutes. Chromatograms were analyzed using Gilson HPLC software, version 2.02. Standard curves and regression coefficients for retinoid compounds are listed in Appendix G. Intemal standards were used to correct for extraction efkiencies and a recovery eniciency was determined by spiking liver samples with standard C.

2.9 LABORATORY EXPOSURES

2.9.1

Fish captura and hwbandry

Fish were wliected by trap nets set in the Bay of Quinte in July of 1995 and transferred to Trent University. Bullheads were given an intramuscular (Lm.) injection of Borgal and held in a flow-through system of sand Siterad Otonabae River water a 1 6 O C . Bullheads are prone to disease in the laboratory and Borgal was administered as a broad-range antibiotic. An attempt was made to feed

bullheads during a one week acdimation penod. but regardless of the tirne or type of food ofkred, the fish did not W. In addition. moctalibies began after airee days of holding. Thus, al1 physicaliy healthy bullheads were exposed to experimental

conditions immediately afler the acclirnation perid.

2.9.2 Exposure to Sediment Hamilton Harbour sediment was collecteci in July of 1995 from Randal Reef with a Ponar dredge. Approximately 2 kg of the sediment was placed in a plastic-

lined holding tank (2500 L), creating a 1-2 cm layer on the bottom. Bullheads (n=21) were placed in the holding tank with sediment for 72 hours and then transferred to a clean holding tank. After 72 hours of exposure. there were mortalities among 12 of the 21 treatd bullheads. Fish were sacrificd pnor to exposure (n=3), immediately after exposure (n=3) and at 96 hours(n=3) and twelve days (n-3) post exposure. FACs in the bile were anaiyred according to the

methods listed above. The EROD assay as descrbed in section 2.6.2 was carried

out using the pms at Trent University.

2.10 Stitistical A n a m

Al data were log transformecl in an attempt to reduœ heteroscedacity and to meet n o m a l i i requiernents. Micronucleus data were converted to percent of micronucleated hepatocytes per 1000 hepatocytes scored and then transformed using the arcsine square mot transformation.

Data that did not meet statistical

requirernents for homogeneity of variance and normality were analyzed by nonpararnetric techniques. Thus, analysis of variance (ANOVA) or ANOVA on ranks (Kruskal-Wallis) were used to test for differences between sites. Differenceswere considered statistically significant at pc0.05.

Multiple site cornparisons were

pefiorrned with Dunn's method which is useful for paimise cornpansons with unequal sample skes. Spearrnan rank correlation was used to test for relationships between biomarkers. In addition, principal components analysis (PCA) was perfomed to detemine if the five study sites could be difterentiated based upon the

biomarker response of individual bullheads. Parameters utilized in PCA analysis included hepatic ER00 acüvity, hepatic dehydroretinol, and bile FACs. A more

detailed description of PCA analysis can be found in section 3.9. Statistical analysis

was perfomed with Sigma Stat version 1.O software. with the exception of PCA, which was perforrned with SYSTAT 5.1 software. It is important to note that the majority of statistically significant difierences

observed were based upon non-parametric techniques.

However, data are

presented in the text as mean i standard deviation. The median values utilizedwith non-parametric statistical procedures are listed in Appendix H.

Lastly, non-

detedable concentrations of hepatic retinoids were found in a f&w fish (n=4). Random numbers between zero and the detection Iimit were generated using SYSTAT 5.1 and used in place of non-detectabie data points (Appendix G).

3.0 RESULTS 3.1 Age and Liver Somatic Indices

Brown bullheads sampled h m various sites (spnng and faIl 1994) ranged from 2 to 6 years of age, with some fish from Hamilton Harbour reaching 8 and 9 years.

To ensure a relatively hornogenous age among sampling sites. only the data from fish aged 3 to 5 yeam were used. H i n this age grwp, the mean age of bullheads from Hamilton Harbourwas significantiy greater (4.5I0.73years) than the mean age of fish (collecteci in fall 1994) frwn al1other sites (pe0.05, Dunn's) (Table 3.1). The

mean age of bullheads collected in the fall of 1994 from the remaining study sites ranged fiom 3.1 to 3.7 years. In addition, bullheads colleded from Old Wornan Creek in the spring and fall were significantly older than bullheads collected from the Black River in the spring and fall of 1994 respectively (pc0.05, Dunn's). Relative liver size or the Iiver somatic index (LSI) was detenineci by dividing the lNer weight by body weight minus liver weight. LSls of bullheads from al1 sites

are listed in Table 3.1 . There was a signifiant differenœ in LSI between sites (fall collection) (p