Polychlorinated Biphenyl Exposure Causes Gonadal Atrophy and Oxidative Stress in Corbicula fluminea Clams

Abstract

Polychlorinated biphenyls (PCBs) are widespread environmental contaminants that have been linked to oxidative and other toxic effects in both humans and wildlife. Due to recent environmental health concerns at a PCB contaminated Superfund site near Raleigh, NC, we used a common clam species (Corbicula fluminea) as surrogates to isolate the effects of PCBs on threatened bivalves native to the region. Under controlled laboratory conditions, clams were exposed to 0, 1, 10, or 100 ppb Aroclor 1260 in the ambient water for 21 days. Measured biomarkers spanned a range of effective levels of biological organization including low molecular weight antioxidants, lipid-soluble antioxidants, and whole tissue radical absorption capacity. These data were augmented by use of histological evaluation of whole samples. Aroclor 1260 significantly increased reduced glutathione (GSH) and total protein concentrations at all treatments levels. Significant decreases were measured in all treatments in γ-tocopherol and total oxidant scavenging capacity (TOSC) and α-tocopherol values in the 100 ppb exposure. Histologically, Aroclor 1260 caused significant gonadal atrophy, effacement of gonad architecture with accumulations of Brown cells, and inflammation and necrosis in digestive glands and foot processes. Our results indicate that oxidative mechanisms play a significant role in the decreased health of these clams due to exposure to Aroclor 1260. The changes in the gonads of exposed clams suggest that a serious threat to bivalve reproduction exists due to PCB exposure.

Introduction

Polychlorinated biphenyls (PCBs) are ubiquitous contaminants in the environment due to their early use as high boiling point, stable compounds in electrical transformers, heat transfer applications, vacuum oils and other industrial applications. Due to widespread lack of containment from operations such as reclamation and storage of electrical components, PCBs were still being introduced into the environment at many sites until the late 1990s. The inherent stability and toxicity of many of the congeners have resulted in PCBs being a persistent environmental problem (Safe, 1994). These highly lipophilic compounds adsorb to soil and sediment readily, with log octanol-water coefficients ranging from 3.76–8.26. They are present in the sediment and water column in aquatic environments, making them available to bioaccumulate and induce negative effects in native fauna (Livingstone, 2001).

The National Priorities List, maintained by the EPA, currently shows more than 500 sites across the country polluted with Aroclor, dibenzofurans, and dioxins. Polychlorinated biphenyls have been detected in a variety of organisms at these sites (Tanabe et al., 1987). Their effects, while congener dependant, include both cancer and non-cancer endpoints. Mechanistically, PCBs affect lipid metabolism, endocrine function, and are implicated in contaminant-stimulated reactive oxygen species production in aquatic organisms at many trophic levels (Kelly et al., 1998; Livingstone, 1990, 2001). The lengthy list of PCB-related maladies, both in humans and wildlife, includes reproductive dysregulation, immune system damage, nervous system disorders, skin disorders such as chloracne, and an array of sensory defects (Kuratsune, 1996; Shimizu et al., 2003).

Suspension feeding invertebrates living at polluted sites serve as useful indicators of PCB contamination (Peterson et al., 1994; Colombo et al., 1995; Labrot et al., 1999; Cataldo et al., 2001). The Asiatic clam, Corbicula flumenia, an invasive species initially introduced into the Northwestern United States in the early 20th century, is now widespread in freshwater ecosystems of the United States (McMahon and Wilbur, 1983). In North Carolina, they are present in most river and lake systems at relatively high densities. The wide distribution of the Asiatic clam and its nuisance designation has facilitated its use as an environmental sentinel (Colombo et al., 1995; Labrot et al., 1999; Barfield et al., 2001; Fournier et al., 2005).

Typically, bivalves with higher PCB burdens demonstrate increases in both primary and secondary antioxidant systems. Animals from chronically polluted sites typically have higher resistance to episodic exposure compared to animals from control sites (Rodriguez-Ortega et al., 2002). Uptake of PCBs through contaminated or spiked algal diets has been well documented for bivalves (Thompson et al., 1999; Chu et al., 2000). The recent EPA rediscovery of unacceptable Aroclor 1260 concentrations in aquatic species in a recreational reservoir near Raleigh, NC, and the knowledge that at-risk species of freshwater mussels inhabit downstream waterways, prompted field studies with sentinel C. flumenia to determine the potential biological effects of contaminants in the reservoir on freshwater bivalves. Marked gonadal atrophy and other tissue damage were apparent after the clams were held for 3 weeks in the reservoir (Lehmann, 2006).

However, because the clams were deployed in the environment, the actual effects of Aroclor exposure could not be differentiated from that of other stressors present in the system. In this study, we describe the effects of Aroclor 1260 on Asiatic clams exposed to known concentrations under controlled laboratory conditions. The responses of a suite of primary and ancillary biomarkers of oxidative stress at different levels of biological organization along with histopathology were used to measure the effects of this PCB mixture on the clams. We hypothesized that the effects of PCBs on bivalves are mediated through oxidative stress.

Materials and Methods

Corbicula fluminea clams, 14–20 mm in length, were collected from a local, un-impacted watershed (Lake Wheeler, Wake County, NC) and held in the laboratory for a minimum of 2 weeks in a closed recirculating system (~1100 liters) with sand substrate to assure their health prior to the experiments. While in holding, clams were fed a continuous slow drip of live Scenedesmus green algae. Corbicula (5 or 6 per container) were then randomized into 4 L beakers, 3 beakers per treatment. Two centimeters of sterile sand were placed in the bottom of each beaker as substrate since clams without substrate show alterations in biomarkers of oxidative stress (Vidal et al., 2002a) (Figure 1).

Exposure beakers were placed into a water bath at 20 ± 0.5′C for the duration of the exposure, in a chemical fume hood. Sand was sterilized by baking at 300′C for 12 hours and then thoroughly rinsing and rehydrating with reverse osmosis purified water. Air was provided by slow bubble from a standard aquarium type air pump and sterilized air stone. Exposure water (3 L per container) was reverse osmosis purified (18 + MΩ resistance) and reconstituted to 50 uS/s conductivity with Instant Ocean synthetic sea-salt (Marineland, Moorpark, CA) in order to mimic the conductivity of the local reservoir. Clams were treated at 0 (vehicle control), 1, 10, or 100 ppb Aroclor 1260 in a static renewal configuration. Aroclor 1260 was purchased from Chem Service (West Chester, PA).

Water changes (50%) were performed twice weekly and included feeding 100 ul of commercially available concentrated algal culture (Reed Mariculture, Campbell, CA) as described by Chu and co-workers (2003). Aroclor was dosed into the water changes in 100 ul ethanol carrier and mixed well for 30 minutes to assure dissolution (1.5 L water changed twice per week). Control clams were dosed with 100 ul ethanol containing no Aroclor 1260. We chose to run the exposures for 21 days as current literature and previous pilot studies suggested this is sufficient time for uptake and effects in bivalves (Rodriguez-Ariza et al., 2003).

Preliminary results from our laboratory indicated that oxidative effects and histological changes were also present at 21 days of exposure. Water quality testing was performed weekly and ammonia levels were never above detection limits (0.25 mg/L). Total dissolved organic carbon (DOC) was tested for at the NC State University Department of Soil Chemistry Analytical Services Laboratory in weekly water samples from each treatment vessel (3.3 ± 1.4 mg C/L). DOC levels did not vary significantly between treatments or experiments. The entire clam exposure regime was repeated.

Clams were euthanized by an overdose of MS-222 (Argent Labs, Redmond, WA), 300–500 mg/L for 10 minutes (Levine, 2006) and were nonresponsive to physical stimulus and fixation chemicals. Three clams were randomly removed for histopathology from each treatment level. Remaining clams were dissected, blotted dry, weighed and then used to create a composite sample from each beaker. The weight of each composite sample was recorded and then homogenized in 4 volumes (w:v) of PBS buffer containing 1mM EDTA using a BioHomogenizer (ESGE, Switzerland). Aliquots were then taken and processed per the requirements for each assay. Final aliquots were immediately diluted or processed and stored at −80°C until analysis. All chemicals without a specified source were purchased from Sigma-Aldrich (St. Louis, MO).

Specimens were prepared for histopathology by gently prying open the shells and inserting a toothpick between the valves to hold them open. Corbicula were euthanized and immediately fixed in 10% neutral buffered formalin for 24 hours. The visceral mass of each specimen was then carefully removed by transecting the adductor muscle close to the shell with a scalpel and peeling up the mantle. Clams were then bisected and placed into 10% formic acid for 24 hours to remove any residual shell or mineral debris.

After that time, they were placed in 70% ethanol until histological processing. The tissues were routinely processed, embedded in paraffin, sectioned at 5 microns, stained with hematoxylin and eosin (HE), and examined by light microscopy. Lesions were scored by a single pathologist according to the scale outlined in Hurty et al. (2002). In brief, lesions were scored from 0 (no remarkable microscopic abnormalities) to 5 (severe lesions) for each lesion type.

Morphometrics were performed on HE stained slides using Image-Pro Plus (Media Cybernetics, Silver Springs, MD) software. Three random and nonoverlapping fields from each slide at 4 × objective magnification (total of 21.24 mm²) were photographed and each image was analyzed for total cross-sectional area of gonadal acini, including both ovarian and testicular tissues. Regions of lumen or those containing no tissue were subtracted from the total tissue area of each field. Fields containing 25% or more of intestinal or no tissue were discarded and not photographed for purposes of gonadal measurements. Samples were all aligned in the same plane for slide mounting. Results are expressed as a percentage of gonadal cross-sectional area per total tissue area examined.

Reduced glutathione (GSH) levels were measured with commercially available kits according to manufacturer’s instructions (Cayman Chemical, Ann Arbor, MI). The GSH kit uses glutathione reductase and DTNB as a basis for colorimetric detection of glutathione. Protein was measured by use of a bicinchoninic acid (BCA) kit from Pierce (Rockford, IL) using manufacturer’s instructions at a dilution of 1:20. This kit uses BCA as a universal reagent measuring the reduction of copper in solution.

Antioxidants were extracted in amber glass vials using sequential extractions with dichloromethane:hexane (9:1). In brief, samples were combined with the DCM:hexane at 1 ml volume per 0.1 ml sample homogenate. δ-tocopherol was added as a recovery standard. After an overnight shake, the supernatant was removed, and the process repeated for 30 minutes.

A third and final extraction was performed after briefly vortexing. The supernatant was then exchanged to ethanol by reduction under nitrogen and brought up to a final volume of 250 ul. Fifty ul of each sample was injected onto the column and run at 0.8 ml/min. Antioxidants were measured via high-pressure liquid chromatography (HPLC) with electrochemical detection (ECD) on an ESA, Inc. Coularray (Boston, MA). Retinol acetate was used as an internal standard. α-, δ-, and γ-tocopherol, β-carotene, retinol, and coenzymes Q9 and Q10 were measured with the following voltages: +200, +400, +500, +700, −800, −900, +200, +500 mV using an isocratic method and a C18 nucleosil column (Supelco, Bellefonte, PA). Running buffer was 78:20:2 methanol:2-propanol:ammonium acetate. Each sample run was followed by a 1-minute flushing with running buffer that included 10% hexane.

Total oxidant scavenging capacity was assessed according to Winston, Regoli, and coworkers (Winston et al., 1998). Whole homogenate samples were cleaned up using a 10,000 × g centrifugation at 4°C for 10 minutes to remove large cellular debris. Due to experimentally increased protein concentrations in our samples, we did not normalize TOSC values to total protein concentration. Instead, samples are normalized to wet tissue weight to prevent data misinterpretation derived from covariance of the assay with experimentally derived differences in total protein concentrations (Winston et al., 1998). We used whole tissue homogenates in place of S9, cytosolic, or lipid soluble fractions alone, in order to better assess antioxidant capacity in the entire animal.

Samples were diluted 1:5 in 100 mM potassium phosphate buffer with 1 mM EDTA and a total of 10 ul was used in each 1 ml reaction vessel fitted with Mininert valves (Supelco). Samples were reacted with 20mM 2,2′-azobis(2-methylpropionamidine)dihydrochloride (ABAP) and 2mM α-keto-γ-methiolbutyric acid (KMBA) over a period of 120 minutes for the measurement of ethylene gas production as an indicator of susceptibility to peroxy radicals. Samples were run on a Hewlett-Packard Series 5890 gas chromatograph (GC) with flame ionization detection (FID). A 30 m Rt-QPlot capillary column (ResTek, Bellefonte, PA) was used for separation of ethylene from other gases. Running conditions were injector 165°C, oven 40°C, detector 250°C. 0.5 ml of headspace was injected per sample every 12 minutes over the course of the reaction.

Statistical analyses were performed with JMP or SAS (SAS, Inc., Cary, NC). After evaluation of normality and homogeneity of variance, endpoint values were analyzed using one-way analysis of variance (ANOVA) at each treatment concentration. If the f values from ANOVA were significant, data was further analyzed by use of Dunnett’s t-test to compare the results of treatments to control results. A value of p ≤ 0.05 was considered significant. Values shown are means ± standard error. Data from both exposures was combined.

Results

Gross Findings

Low mortality was observed in the initial experiment (2/18 individuals at 10 ppb, 1/19 at 100 ppb). Clams were immediately removed and discarded when it was determined that they were dead. Mortality was determined by gaping and a lack of response to physical stimulus. No mortalities occurred in the repeat experiment. During dissection, in preparation for homogenization, we noted that the visceral mass of some of the 10 and 100 ppb treated clams was bulging and more turgid than normal, had rounding of the normally sharp edges, and exuded clear fluid on cut surface. This was interpreted as generalized edema.

Analytical Findings

Lipid-soluble antioxidant analysis indicated that γ-tocopherol was significantly decreased in all treatments relative to controls (Figure 2). Alpha-tocopherol levels were highly variable among samples, with the 100 ppb dose being reduced in concentration, but not significantly. Ubiquinol (CoQ9 and CoQ10), β-carotene, and retinol levels were typically below detection limits in the dilutions used for analysis. Total reduced glutathione (GSH) levels in whole body were elevated in all treatments compared to control, with exposure to 10 ppb showing the highest increase in concentration. No apparent dose response was discerned in increasing treatment concentrations between 1 and 100 ppb over the 21 days exposure time (Figure 3).

Protein concentrations were significantly increased in all treatments compared to controls (Figure 4). The 1 and 10 ppb Aroclor treatments had TOSC values that were not significantly different from controls, the highest concentration of Aroclor 1260, however, caused a halving in overall antioxidant capabilities (Figure 5). The TOSC assay was only performed on one repetition of the experiment.

Histopathology

Based on grading of lesion severity, there was an apparent concentration response in treatment-related morphologic changes, with an increase in severity with increasing treatment levels in both exposures (Table 1). Corbicula from the control group showed few to no remarkable microscopic lesions. A few controls had scattered hemocytes around the digestive glands, but this was considered to be within normal limits. Corbicula in the PCB-exposed groups had three major lesion types: gonadal atrophy with increased Brown cell infiltration, generalized edema, and focal necrosis and inflammation. Total gonad cross-sectional area decreased significantly in all exposures compared to control clams (Figure 6).

The gonadal acini in the exposed specimens were often shrunken, necrotic/fragmented, and infiltrated with one to several enlarged, pigmented phagocytes (melano-macrophage-like hemocytes or Brown cells—see Figure 7c). While control clams had the normal, clustered or acinar arrangement of ovarian follicles and ovotestes, exposed clams had marked necrosis and loss of oocytes and ovotestes with both a decrease in the overall gonadal cross-sectional area and a decrease in the size of individual acini. Morphometric analysis of whole sections verified that changes in gonadal abundance and density occurred in the exposed groups (Table 1).

Compared to controls, 10 and 100 ppb exposed clams (2/6 and 4/6, respectively) also had diffuse, moderate to-severe edema characterized by increased clear space within connective tissues and paler staining of the viscera (Figure 7d). Digestive glands in exposed clams were also moderately expanded by infiltrates of enlarged (presumably activated) hemocytes. Epithelial cells in some tubules were swollen, pale staining, and had indistinct to frayed cell borders. In some specimens, aggregates of hemocytes were also noted in the loose connective tissue away from the digestive glands.

Discussion

Corbicula exposed to Aroclor 1260 showed increasing evidence of cell injury with increasing concentrations of PCBs. The gross and microscopic findings were similar to those seen in our recent field study in which sentinel clams were deployed in an aquatic system near Raleigh, NC that is heavily contaminated with PCBs (Lehmann, 2006). The marked changes seen in both the field and in the current controlled laboratory study are consistent with our hypothesis that the adverse effects of PCBs are mediated via oxidative stress.

Two mechanisms appear to exist in which PCBs exert their oxidative effects in organisms. The first is direct oxidation of macromolecules due to phase I metabolism by the mixed function oxidase system and quinone cycling of the metabolites (Twaroski et al., 2001), and the second is indirect release of reactive oxygen species by metabolizing enzymes induced by exposure (McLean et al., 2000). Polychlorinated biphenyls interact at the hormonal level by interaction with aryl hydrocarbon (AhR) and other cellular receptors and by interfering with lipid metabolism (Hahn, 1998). Dioxin induced reproductive alterations have been shown in some cases to be AhR-independent (Butler et al., 2004).

Although the AhR is not phylogenetically conserved in mollusks, there may be comparable receptors in bivalves and other invertebrates as evidenced by 7-ethoxyresorufin-O-deethylase (EROD) induction due to halogenated hydrocarbon exposure (Brown et al., 1997; Gardinali and Wade, 1998). Bivalves fed a contaminated diet or exposed via field studies show transfer of PCBs to reproductive tissues, with the highest residual concentration of these lipophilic compounds being found in the gonads (Chu et al., 2003).

Reductions in gonadal size and function are also associated with PCB exposure in rodent studies, even at levels approaching those found as background in human populations (Johnson et al., 1994; Smolowitz and Leavitt, 1996). In the present study, we observed similar reductions in gonadal cross-sectional area in Corbicula exposed to Aroclor 1260. The gonadal atrophy was evident using routine histopathology and was confirmed by morphometric analysis.

Freshwater bivalves serve as highly useful biological sentinels in polluted environments. Corbicula sp. clams have been used for this purpose for many years (Colombo et al., 1995; Labrot et al., 1999). As an invasive species, these clams are now relatively ubiquitous in warm water environments worldwide. Their broad distribution and non-native status makes them ideal sentinels, which provide a consistent set of biomarkers for environmental field studies (Colombo et al., 1995; Barfield et al. 2001; Pampanin et al., 2002; Vidal et al., 2002b). Corbicula are well adapted to shifting environments and oxidative challenge. Their benthic habitat places them into contact with sediment-associated contaminants. These laboratory studies provide a baseline for comparison with field studies using sentinel clams, where multiple factors can confound data interpretation, and for evaluation of the risk of PCB-related health effects in native, endangered wildlife.

No single biomarker reflects the complex interactions of variables that define the health of ecosystems. Common biomarkers of oxidative challenge include antioxidative enzymes (such as superoxide dismutase and catalase), ancillary antioxidative enzyme systems and those involved in metabolism (such as glucose-6-P-dehydrogenase and glutathione S transferase), lipid oxidation state, non-enzymatic antioxidants (such as total reduced glutathione levels, vitamins, urea), and total ability of cellular components to reduce oxidation in vitro (Total Oxidant Scavenging Capacity) (Koremura et al., 1990; Winston et al., 1998; Sheehan and Power, 1999; Lee and Opanashuk, 2004).

Generally, in single compound exposures, there is a rise in enzymatic activity or concentration as well as increases in nonenzymatic defenses. High levels of oxidative compounds cause a sharp decrease in many of these biomarkers indicating that the antioxidative systems employed have been overburdened or compromised. Highly specific biomarkers, such as reduced glutathione concentrations are less relevant at the population level, while more general population level biomarkers give no indication of the mechanism(s) of action of a toxicant.

Therefore, it is of benefit to use an array of biomarkers for determination of toxic effects, preferably at different levels of organization, from molecular markers to genetic alternations to whole animal health indices (Ham et al., 1997; Regoli, 1998; Burton et al., 2005). Included in the analysis of biomarker changes, however, is the need for demonstrable biological or physiological effects of a toxicant that can be compared across phyletic levels. This need is well met by the use of histopathology. This idea is in line with the need to relate the responses of individuals to the responses of populations and communities (Hinton et al., 2005).

Measuring the relative changes of biomarkers in response to environmental and contaminant-related variables will help define the breadth of physiologic changes that can be anticipated in response to specific contaminants. In the current Corbicula study, the gonadal alterations due to Aroclor 1260 exposure points to the role of PCBs in chronic environmental degradation. Further definition of these biomarkers in resident aquatic fauna will help refine our understanding of their value in assessing ecosystem health.

The duration of exposure, in this case 21 days, allowed for a whole animal response to the oxidative stress which the PCBs initiate, but not necessarily the time necessary to achieve steady state concentrations due to the quantity of highly (penta-, hepta-, hexa-) chlorinated PCB congeners in the Aroclor 1260 mixture. According to Rodriguez-Ariza et al., 2003, approximately 50% maximal uptake is achieved by 21 days in sediment associated uptake monitoring, including proportionately more of the lower (mono-, di-, tri-) chlorinated congeners which are more readily metabolized by monooxygenases (Rodriguez-Ariza et al., 2003). Nevertheless, the rate of uptake and resulting tissue concentrations are sufficient to generate a strong response in oxidant scavenging function. In our study, increases in total glutathione, as well as the decreases in α- and γ-tocopherol and the decrease in TOSC levels indicated that the course of exposure was sufficient to allow for accelerated cellular metabolism and to indicate a reduction or possible collapse of overall antioxidant capacity at tested concentrations.

Lipid-soluble antioxidants appear to be a fairly stable component of the complete antioxidant system in organisms (Kohar et al.,1995). Many of the lipid soluble antioxidants have critical function in homeostasis of cells and are regulated strongly to maintain those functions (Saldeen et al., 1999; Jiang et al., 2000). Any measurable change in these lipid soluble antioxidant levels is an indication of a potent oxidative challenge. Gamma-tocopherol is the stronger lipid soluble antioxidant, being more responsive and excreted when consumed in comparison to α-tocopherol.

Alpha-tocopherol is typically recycled in conjunction with ascorbic acid, unlike γ-tocopherol, so that levels should be comparatively more stable (Kohar et al., 1995). In the short term, shifts in relative proportions of tocopherols indicate that the cumulative total of the oxidant side of the oxidant:anti-oxidant balance is increasing and allows for more oxidation of lipids. Lipofuscin, the “wear and tear” pigment, builds up in cells due to oxidative degradation of lipids and is visible as yellow or brown pigment in cells during histological evaluation (Zaroogian and Jackim, 2000).

It follows that reduction in lipid protective molecules may increase lipofuscin accumulation, as seen in this study within Brown cells. In this experiment, the 100 ppb Aroclor 1260 concentration may take defense mechanisms beyond the compensation point of the antioxidant system. The more responsive results demonstrated by γ-tocopherol are likely due to perturbance of lipid bilayers by the highly lipophilic PCBs and resulting quinone cycling oxidative metabolites. Tocopherols have been shown experimentally to reduce thiobarbituric acid reactive substances (TBARS) values during oxidative challenge and are known to directly affect reproduction (Kawai-Kobayashi and Yoshida, 1986).

Alterations in lipid metabolism due to PCB exposure have been demonstrated by Ferreira and Vale in oysters (Ferreira and Vale, 1998). Coincident with the lipid metabolism alterations, PCBs are known to interact with signal transduction as evidenced by the release of insulin in cell culture. While invertebrates are thought to have no aryl hydrocarbon receptor, functional homologs have been described in the literature (Fischer et al., 1998; Brown et al., 2002; Hahn, 2002; Butler et al., 2004). Taken together with the fact that reproductive tissues accumulate high quantities of PCBs, these possible mechanisms may have contributed to the severe changes identified in the gonads of the Corbicula studied.

Reduced glutathione concentrations are used frequently as markers of effects in both field situations and in the laboratory. Experimental supplementation with GSH has been shown to reduce oxidative damage directly induced by some PCB congeners (Slim et al., 2000). However, the variability seen in the literature somewhat offsets the usefulness of GSH as a sole marker of oxidative stress. In combination with other markers, GSH can remain a useful indicator of early responses to oxidative challenge to cells.

In this study, GSH rose in response to increasing concentrations of PCBs, corroborating the results seen in previous studies with bivalves (Sheehan et al., 1995). An increase in the glutathione system components tested indicates that oxidative stress was a major consequence of exposure. Rodriguez-Ariza et al. (2003) noticed a decline in GST levels over a longer term PCB exposure, where the activity continued to fall until 110 days’ exposure, suggesting a slow overburdening of the system due to PCB accumulation while GSH levels rose and fell with time (Rodriguez-Ariza et al., 2003).

The TOSC assay has been used as a generalized indicator of overall antioxidant system status by measuring whole sample in vitro capacity to absorb peroxyl, hydroxyl, or other radicals depending on the radical generating system employed (Regoli et al., 1998; Winston et al., 1998; Regoli, 2000). Since protein production increased by treatment in this experiment, we did not normalize our values to protein concentration to avoid covariance. As noted in the study by Regoli et al. (1998) of Mytilus bivalves, various fractions (cytoplasmic, lipid, s9) have differing degrees of relevance for each contaminant stressor.

Due to the nature of the oxidative stress derived from the Aroclor mixture used in this study, we opted to examine whole homogenized samples in lieu of various fractions in order to more closely evaluate the antioxidant capacity intrinsic to the entire animal. This assay has not previously been extended to Corbicula sp. clams. We found that high concentrations of Aroclor 1260 in the water column caused a significant decrease in TOSC consistent with our hypothesis that PCBs cause significant oxidative stress in clams.

Although the concentrations in this study are three-fold higher than EPA water quality criteria levels for protection of aquatic life, the results are very similar to those evidenced in our prior field studies in the Aroclor polluted Superfund site (Lehmann, 2006). The resistance of clams to lower concentrations of Aroclor is likely related to a reduced ability to metabolically activate the contaminants. TOSC assay values fell in a concentration-dependent manner with increasing concentration of PCB mixture, again consistent with depletion due to oxidative stress.

Microscopically, gonadal atrophy seems to be the most significant population relevant effect of PCB exposure in these clams. This gonadal alteration may be sufficient to reduce fecundity or cause population level alterations in the local environment as noted by a lack of mature eggs in C. virginica oysters after PCB exposure by Chu et al. (2003). Evidence of inflammation and discrete regions of necrosis were found in our study, indicating that overt damage was derived from exposure to Aroclor 1260, possibly due to oxidative cellular damage as noted by Koponen in fish from polluted lake systems (Koponen et al., 2001).

Holding the clams under constant exposure conditions beyond the three week period used in this experiment would likely have produced even more adverse effects at the cellular and organ levels and would probably have resulted in higher mortalities, as noted in the extended study of oxidative effects in C. gallina by Rodriguez-Ariza et al. (2003). Many of our higher dose specimens also had severe generalized edema (anasarca), which may be caused by changes in osmotic balance, hydrostatic pressure, increased tissue permeability due to inflammation, or direct tissue injury. In our study clams, the edema was most likely caused by oxidation of lipid membranes, creating increased permeability as well as profound osmotic disturbances.

The scattered inflammatory infiltrates seen in various tissues were most likely reactive or secondary to direct tissue injury. This injury is an explicit result of oxidative damage and the resulting peroxidation of lipid chains. Patchy necrosis was also visible in histological sections from some of the treated clams. Necrosis is the end result of too much overt cell injury that cannot be repaired. This also fits into the model of oxidative damage being primarily responsible for dysregulation of ionic balance and breakdown in cell structure.

In addition to the extensive necrosis and loss of gonadal tissue, the accumulation of relatively large, pigmented macrophage-like hemocytes or Brown cells amongst the residual necrotic debris of the gonads was interesting. While the origin and specific functions of these cells are still being investigated, recent studies have associated the accumulation of Brown cells in various tissues of bivalves with exposure to organic pollutants (Smolowitz and Leavitt, 1996; Morado and Mooney, 1997). Zaroogian et al. have perhaps most thoroughly characterized the brown cells of the red gland or pericardial gland of clams as fixed tissue cells involved in detoxication of metals (Zaroogian and Jackim, 2000; Zaroogian and Norwood, 2002). This would suggest a liver-like function for the red gland.

However, the Brown cells observed in our study are more likely a subtype of the mobile, phagocytic hemocytes involved in innate immunity. The accumulation of Brown cells in areas of oocyte damage, apparently due to oxidation of lipid membranes, would suggest that these cells serve a somewhat similar function to the melanomacrophages of teleost fish. That is, they probably serve as the “clean-up crew” and accumulate ceroid and other materials that cannot be further broken down. Further research correlating morphology and function of Brown cells is needed.

The specific biomarker responses, as well as the results of whole organism pathology, in this study are consistent with our hypothesis that oxidative damage is a result of exposure to Aroclor 1260, with accumulated damage affecting various levels of biological organization. The combination of oxidative stress biomarkers and the gonadal damage visible at the microscopic level suggests that oxidative stress is a direct consequence of exposure to the PCB mixture. While we are unable to define oxidation as a direct cause of the gonadal atrophy, evidence derived from this study and others indicates that oxidative damage is a direct acting factor in such declines. Relevant changes, whether directly or indirectly due to oxidative stress, occur at the organ and organism levels and will likely result in population wide effects, including reduced fecundity and chronic maladies, if similar uptake of PCBs occur in the field.

Freshwater mussels native to North America are among the most endangered groups of animals on the planet (Williams, 1993; Neves, 1997). Polychlorinated biphenyl contamination is extensive and may be contributing to declines in freshwater mussel fecundity and populations. Endangered bivalves and other aquatic macroinvertebrates will remain at risk due to the extensive number of PCB, dioxin, and dibenzofuran polluted sites in this country. Biologists considering the augmentation or reintroduction of extirpated or extinct macro-invertebrate populations should measure and carefully consider PCB levels at potential restoration sites.