Polychlorinated biphenyls and methylmercury act synergistically to reduce rat brain dopamine content in vitro.

Consumption of contaminated Great Lakes fish by pregnant women is associated with decreased birth weight and deficits in cognitive function in their infants and children. These fish contain many known and suspected anthropogenic neurotoxicants, making it difficult to determine which contaminant(s) are responsible for the observed deficits. We have undertaken a series of experiments to determine the relevant toxicants by comparing the neurotoxic effects of two of these contaminants--polychlorinated biphenyls (PCBs) and methylmercury (MeHg)--both of which are recognized neurotoxicants. Striatal punches obtained from adult rat brain were exposed to PCBs only, MeHg only, or the two in combination, and tissue and media concentrations of dopamine (DA) and its metabolites were determined by high performance liquid chromatography. Exposure to PCBs only reduced tissue DA and elevated media DA in a dose-dependent fashion. Exposure to MeHg only did not significantly affect either measure. However, when striatal punches were simultaneously exposed to PCBs and MeHg, there were significantly greater decreases in tissue DA concentrations and elevations in media DA than those caused by PCBs only, in the absence of changes in media lactate dehydrogenase concentrations. Elevations in both tissue and media 3, 4-dihydroxyphenylacetic acid concentrations were also observed. We suggest that the significant interactions between these two toxicants may be due to a common site of action (i.e., toxicant-induced increases in intracellular calcium and changes in second messenger systems) that influences DA function. The synergism between these contaminants suggests that future revisions of fish-consumption guidelines should consider contaminant interactions.

punches were simultaneously exposed to PCBs and MeHg, there were significantly greater decreases in tissue DA concentrations and elevations in media DA than those caused by PCBs only, in the absence of changs in media lacte dehydrogenase concentrations. Elevations in both tissue and media 3,4-dihydroxyphenylacetic acid concentrations were also observed. We suggest that the significant interactions between these two toxicants may be due to a common site of action (i.e., toxicant-induced increases in intracellular calcium and changes in second messenger systems) that influences DA function. The synergism between these contamiants suggets that future revisions of fish-consumption guidelines should consider contaminat interactions. Key wordk. dopamine, methyimercury, polychlorinated biphenyls, rat brain, striatal slice, synergism. Environ Health Pepect 107: 879-885 (1999). [Online 7 October 1999] btp://ehpnetl.niehs.nih.gov/docs/)999/107p879-885bemis/abstract html An important but daunting task for toxicologists is elucidation of the interactive effects and mechanisms of action of individual toxicants that make up complex mixtures of environmental contaminants. One of the most relevant mixtures of toxicants in the environment is found in contaminated fish, particularly those from the Great Lakes. Epidemiologic studies have demonstrated an association between consumption of contaminated Great Lakes fish by mothers and deficits in growth, development, and cognition in their infants and children (1)(2)(3). Because polychlorinated biphenyls (PCBs), at concentrations measured either in breast milk or in fetal cord blood (4), have been associated with these deficits, it has been assumed that PCBs are the contaminant responsible for these deficits. However, Great Lakes fish, induding salmon, contain many other known and suspected developmental neurotoxicants, including methylmercury (MeHg), chlorinated pesticides, and heavy metals (5,6). Thus, it remains difficult, using only the tools of epidemiology, to determine which contaminant or contaminants are responsible for the observed associations between maternal fish consumption and developmental and cognitive deficits in their infants and children.
To begin to determine which fish-borne contaminants alter central nervous system function, and whether they interact functionally or synergistically, we examined changes in tissue and media concentrations of dopamine (DA) following in vitro exposure of striatal punches from adult rats to PCBs or MeHg alone or in combination. These compounds represent two of the major neurotoxicants found in contaminated Great Lakes salmon. We chose to initially look for possible neurotoxic interactions between these two contaminants because of the extensive literature documenting the neurotoxicity of each toxicant (7)(8)(9). For this initial study we used striatal tissue punches, rather than cells in culture, because the explanted tissue conserves much of the neuronal complexity and many of the neurotransmitter interactions of the intact brain. We examined changes in DA function because developmental exposure of laboratory rodents to lyophilized Great Lakes fish containing the above contaminants results in significant decreases in regional brain concentrations of DA (10).

Methods
We initially exposed striatal punches to MeHg at media concentrations ranging from 1 to 40 pM to determine the concentrations of the toxicant that would alter DA function but not induce tissue damage, defined as significant elevations in media lactate dehydrogenase (LDH) concentrations. Similarly, the PCB doses chosen (10-200 ppm in exposure media), which resulted in altered DA concentrations in the absence of elevated media LDH concentrations, were based on previous striatal slice experiments conducted in our laboratory (11).
After these initial experiments, we characterized the neurochemical effects of combined exposure of striatal slices to PCBs and MeHg by exposing striatal punches from naive adult male rats to a 1:  14 pM-that yielded no effect to minimal effects on our end points without inducing elevations in media LDH concentrations. Stock solutions of each toxicant were dissolved in DMSO and diluted in the exposure medium, resulting in a final DMSO concentration of 0.2%. HPLC analysis of dopamine and its metabolites. After the 4-hr incubation, a 200-jiL aliquot of media from each well was diluted 1:1 with 0.4 N perchloric acid, and striatal punches were homogenized in 300 pL 0.2 N perchloric acid. Supernatants from both the tissue homogenates and the perchloric acid-diluted media were analyzed for DA and its metabolites [3,4-dihydroxyphenylacetic acid (DOPAC) and homovanillic acid (HVA) (media samples only)] by HPLC with electrochemical detection as described in Seegal et al. (13). Measurement ofmedia LDH. In parallel experiments, striatal punches, harvested as described above, were incubated in media containing 0.2% DMSO (control), 4, 10, or 14 pM MeHg only, 100 or 200 ppm PCBs only, or various combinations of PCBs and MeHg. After incubation, the media were removed and analyzed for LDH content using the Sigma Diagnostics LDH Assay (Sigma, St. Louis, MO) based on the technique ofAmador et al. (14).
'Statistical analyses. Data are presented from 14 individual experiments that used a range of PCB and MeHg concentrations tested either alone or in combination. The design of each experiment, using 48 tissue punches assigned to treatments in a counterbalanced manner, allowed for either 8 or 12 conditions to be tested with either 6 or 4 punches at each condition, respectively. It was not possible to test all treatment conditions in a single experiment. Because of the variable number of observations and the variance at each test condition, we analyzed the data using the Brown-Forsythe analysis of variance (ANOVA) test, in which the variances in each experimental treatment are not assumed to be equal (15). This conservative test controls for unequal cell size by weighting the variance according to the number of observations. Dose-response effects of either PCBs or MeHg alone were analyzed using one-way ANOVA with Bonferroni-corrected post hoc t-tests to compare each dose to control. Interactions between varying concentrations of PCBs, in the presence or absence of a constant concentration of MeHg, were analyzed using two-way ANOVA with Bonferroni-corrected post hoc t-tests to compare the effects of each combined treatment with the effect induced by the corresponding concentration of PCBs alone. Finally, because of small variations in the control punch DA concentrations between individual experiments, we expressed the data from each experiment as a percentage of its average control value, thus reducing interexperiment variance. The range of absolute values for control tissue and media samples are presented in "Results."

Results
Experiments that examined various exposure times to PCBs, MeHg, and combinations of both revealed no significant effects on the chosen end points before 4 hr (data not shown). Thus, all data represent 4-hr exposure to the previously described treatment conditions. The data presented, expressed as percent of DMSO control values, should be interpreted with regard to the absolute concentrations of tissue and media DA: 372.9 ± 8.9 and 6.2 ± 1.4 ng/mg tissue protein, respectively. The absolute concentrations of DOPAC in tissue and media exposed to control treatments are 8.9 ± 0.35 and 140.7 ± 7.4 ng/mg tissue protein, respectively. Thus, the apparently large elevations in media DA expressed as a percent of control values actually represent a relatively small change in absolute DA concentrations as compared to the large changes in absolute tissue DA content induced by these toxicants.
Effects ofMeHg on dopamine function.
Exposure of striatal punches to MeHg at concentrations ranging from 1 (Table 1).
Tissue concentrations of DOPAC were significantly reduced after exposure to 4, 10, 20, and 40 pM MeHg (p < 0.05 to p < 0.001), although media DOPAC concentrations were elevated only at the 20 pM dose of MeHg (Table 1). The concentrations of HVA in media from punches exposed to all concentrations of MeHg (except 1 pM) were significantly reduced as compared to control values (p < 0.05 to p < 0.001) (data not shown). On the basis of these results, we chose three concentrations of MeHg (4, 10, Table 1. DA and DOPAC levelsa in tissue and media from striatal punches exposed to MeHg. 'Mean + standard error, expressed as a percent of the dimethyl sulfoxide control. bbn = 6-84 punches (wells) per dose. cAbsolute concentrations of tissue and media DA were 372.9 ± 8.9 and 6.2 ± 1.4 nglmg tissue protein, respectively, and tissue and media DOPAC were 8.9 ± 0.35 and 140.7 ± 7.4 ng/mg tissue protein, respectively. *p < 0.1. **p < 0.05. #p < 0.01. ap ' 0.001. Table 2. DA and DOPAClevels" in tissue and media from striatal punches exposed to Aroclor 1254/1260. aMean ± standard error, expressed as a percent of the dimethyl. sulfoxide control. bn = 23-83 punches (wells) per dose. cAbsolute concentrations tissue and media DA were 372.9 ± 8.9 and 6.2 ± 1.4 ng/mg tissue protein, respectively, and tissue and media DOPAC were 8.9 ± 0.35 and 140.7 ± 7.4 ng/mg tissue protein, respectively. *p < 0.05. **p < 0.01. 'p <0.001. Volume 107, Number 1 1, November 1999 * Environmental Health Perspectives and 14 pM) to examine for potential interactions with PCBs. These three doses of MeHg spanned the range of doses from no effect at 4 pM MeHg to small nonsignificant effects on DA function, in the absence of gross alterations in DA concentrations or changes in media LDH, at 14 pM.
Effects ofPCBs on dopamine function.
Exposure to the Aroclor 1254/1260 mixture at media concentrations ranging from 10 to 200 ppm resulted in a significant dosedependent decrease in tissue DA concentrations (F = 63.64; df= 5, 212; p < 0.001), with all doses significantly different from control at the p < 0.01 or p < 0.001 level (Table 2, Figure 1).
Exposure to PCBs also resulted in a significant dose-dependent increase in media DA concentrations (F= 38.31; df= 5, 178; p < 0.001) with all doses significantly different from control values at the p < 0.05 to p < 0.001 level (Table 2, Figure 2). However, the net effect of PCB exposure was a reduction in total DA concentrations (i.e., punch DA + media DA).
Tissue DOPAC concentrations were significantly elevated after exposure to 100 and 200 ppm PCBs (p < 0.001) ( Table 1). In addition, a significant dose-dependent increase in media DOPAC concentrations (F = 19.81; df= 5, 202; p < 0.001) was observed after exposure to PCBs, with all doses significantly different from control at the p < 0.05 to p < 0.00 1 level. Media HVA concentrations were unchanged after exposure to any doses of PCBs (data not shown). These results are comparable to those reported by Chishti et al. (11)  indicating that the slopes of the lines differ and a significant interaction exists when striatal tissue was exposed to both contaminants.   also significantly decreased tissue DA concentrations as compared to exposure to PCBs alone (p < 0.00 1), resulting in a significant MeHg x PCB statistical interaction (F =5.09; df= 5, 57;p<0.001).
Combined exposure to 14 pM MeHg and varying concentrations of the 1:1 Aroclor 1254/1260 mixture ( Figure IC) also significantly decreased tissue DA concentrations, with combined exposure to all doses of PCB (except 10 ppm PCB) significantly different from exposure to PCBs only at either the p < 0.05 or p < 0.001 level, resulting in a significant MeHg x PCB statistical interaction (F = 3.45; df= 5, 53; p < 0.01). This treatment also increased media DA concentrations ( Figure 2C) with exposure to 14 pM MeHg only; 14 p1M MeHg combined with 40, 100, and 200 ppm PCB was significantly different from exposure to PCBs only at the p < 0.1 to p < 0.01 level, which resulted in a MeHg x PCB statistical interaction (F = 9.39; df= 5, 18; p < 0.001). As with PCBonly exposure, the net effect on total DA (punch + media) was a significant reduction.
DOPAC concentrations were elevated in tissue exposed to 14 pM + 40 ppm PCBs only (p < 0.1) ( Figure 3C) and media from tissue exposed to 14 pM MeHg combined with 200 ppm PCB (p < 0.05) ( Figure 4C)  The addition of 10 or 14 pM MeHg to any of the PCB doses resulted in a simple additive reduction in media HVA concentrations as compared to PCB-only exposed tissue (data not shown). There were no interactive effects observed for HVA after exposure to any of the combinations of PCBs and MeHg.
Effects of toxicant exposure on media concentrations ofLDH. To determine if the toxicants alone or in combination induced nonspecific cytotoxic effects, we measured the concentration of LDH released into media, expressed in international units. We chose to use this measure of tissue viability because the measurement of LDH release into media is a commonly used measure of cellular and tissue toxicity (16,17), particularly when applied to organotypic cultures.
Media LDH levels were unaffected after exposure to MeHg at concentrations < 14 pM (

Discussion
This study is part of a larger project to determine the neurologic (i.e., behavioral and neurochemical) effects of developmental exposure of laboratory rodents to the toxicants present in Great Lakes salmon. Although the concentrations of PCBs and MeHg used in these experiments are higher then those seen in Great Lakes fish (PCB content ranged from 0.84 to 1.9 ppm, whereas mercury concentrations were 0.34 ppm) (5,6), the concentrations of these contaminants are similar to, or lower than, those used in other in vivo and in vitro studies (18)(19)(20)(21). Furthermore, some human populations consume contaminated pilot whale meat that contains PCB concentrations as high as 30 ppm and mercury levels in the range of 3.3 ppm (22), well within our range of doses. As presented, these results provide the first evidence of a significant functional interaction between two of the major neurotoxicants present in contaminated Great Lakes salmon and other freshwater and marine species-PCBs and MeHg. To understand the potential mechanisms for the interactions between these contaminants, it is appropriate to first discuss the mechanisms by which the individual contaminants alter DA function.
PCB effects on dopamine function. The data we gathered in intact animals, PC12 cells, and striatal slices exposed to PCBs suggest two distinct but interdependent mechanisms that may be responsible for the decreases in DA concentrations observed after exposure to PCBs. First, PCBs reduce tissue DA concentrations in the above test systems (10,11,23), most likely due to the inhibition of tyrosine hydroxylase (TH), the rate-limiting enzyme in DA synthesis. Inhibition of TH by PCBs has been demonstrated in striatal slices exposed to NSD-1015, which inhibits aromatic amino acid decarboxylase (AADC). AADC is the enzyme that converts 3,4-dihydroxyphenylalanine (L-DOPA), an intermediate product formed by the TH catalyzed conversion of tyrosine, to DA (24). Thus, PCBs reduce L-DOPA content in striatal slices after exposure to NSD-1015 + PCBs, demonstrating an inhibition of TH activity (12). In addition, Choksi et al. (25) reported reductions in TH activity in rat brain striatal minces exposed to specific ortho-substituted PCB congeners.
Second, in both striatal slices (11) and punches (data presented here) we noted significant elevations in media concentrations of DA, which we believe is a reasonable surrogate for changes that occur in the extracellular space of the tissue punch or the brain. These elevations may be due to PCBenhanced neurotransmitter release and/or inhibition of DA uptake at either the dopamine transporter (DAT) or the vesicular monoamine transporter (VMAT). Thus, Messeri et al. (26) reported elevations in spontaneous DA release from bovine chromaffin cells exposed to PCBs-results similar to those obtained in our experiments. Furthermore, Mariussen et al. (27) showed that ortho-substituted PCBs inhibit DA uptake by VMAT in vesicles isolated from rat striatal synaptosomes, resulting in elevated media levels of DA. The reported elevations of both tissue and media DOPAC concentrations further support the proposed increases in intracellular DA levels because DOPAC is formed from free intracellular DA via the action of intracellular monoamine oxidase (MAO) (28). Thus, increased tissue and extracellular DOPAC concentrations may reflect increased free intracellular DA concentrations elicited by exposure to PCBs. In addition, elevations in free intracellular DA concentrations result in end-product inhibition of TH activity (29,30) and elevations in extracellular DA by a carrier-mediated process that moves DA out of the terminal via DAT (31). Thus, elevations in extracellular DA or exposure to DA agonists result in activation of presynaptic DA autoreceptors and inhibition of DA synthesis at the level of TH (32,33).
HVA is formed from extracellular DA by the catabolic activity of catechol-0-methyltransferase (COMT) and MAO (28). The reported elevations in media DA elicited by exposure to PCBs or MeHg would be expected to result in increased media HVA concentrations. However, the lack of effect on media HVA reported here has previously been observed in our in vitro preparation after exposure to PCBs (11) and most likely represents the effects of relatively short incubation times and the relatively small elevations in absolute media DA concentrations as compared to the large overall reduction in tissue DA content.
In summary, these data provide evidence that PCBs inhibit the synthetic activity ofTH in several model systems. We suggest that the elevations in media DA and, by inference, extracellular DA concentrations represent an initial event induced by PCB exposure which, by the mechanisms described above, contributes to the PCB-induced decrease in DA synthesis.
MeHg efects on dopamine. Both in vivo and in vitro exposure to MeHg alters DA function. Using a mouse striatal slice preparation similar to the one used here, MeHg exposure, at concentrations considerably higher than we used (i.e., 50 pM), resulted in significantly elevated media DA concentrations (20). Similarly, when adult rats were exposed intraperitoneally to MeHg at concentrations ranging from 0.1 to 2 mg/day (for a total dose of 4 or 6 mg), DA concentrations in striatal dialysates obtained by in vivo microdialysis were significantly elevated (21).
Both groups of investigators concluded that the MeHg-induced increases in extracellular DA concentrations may be due to enhanced DA release. These results are similar to those obtained by Komulainen (34), who exposed striatal synaptosomes to MeHg and reported dose-dependent inhibition of uptake of DA (35) demonstrated elevations in intracellular calcium in PC12 cells after exposure to orthosubstituted PCB congeners. These elevations are due to PCB-induced activation of intracellular ryanodine channels that act as release sites for calcium sequestration stores in endoplasmic reticulum and mitochondria. MeHg exposure also results in elevations in intracellular calcium concentrations in both continuous cell lines and primary cells.
Using the calcium-sensitive dye fura 2, Marty and Atchison (36) demonstrated a two-component elevation in intracellular calcium in cerebellar granule cells exposed to MeHg. They concluded that the smaller, more rapid onset increase in intracellular calcium involved MeHg-induced liberation of calcium from inositol 1,4,5-triphosphate (IP3)-sensitive intracellular stores. The second and larger elevation in intracellular calcium involved the influx of extracellular calcium due to MeHg acting on multiple sites of entry of extracellular calcium into granule cells. Furthermore, Minnema et al. (37) showed that exposure of mitochondria isolated from whole rat brain to MeHg results in significant efflux of labeled calcium from these calcium storage organelles. These effects again complement the action of PCBs on the disruption of intracellular calcium homeostasis via sequestration of calcium in mitochondria. Finally, in cultured cerebellar granule cells exposed to MeHg, Sarafian (38) observed increased intracellular calcium as well as 1P3 levels. He demonstrated that MeHg did not directly activate protein kinase C (PKC), but most likely functions through one or both of the reported elevations in the second messengers, calcium and 1P3. Verity (351) published a more complete review of MeHg effects on calcium.
The similar actions of these two environmental toxicants on regulation of intracellular calcium concentrations suggest a potential mechanism that might account for their significant interactive effects on the release of DA into media and on DA synthesis.
Hanneman et al. (40) reported that elevations in intracellular calcium in rat hippocampal slices, induced by exposure to 2,3,7,8-tetrachlorodibenzo--dioxin, increased the activity of PKC. In turn, activation of PKC in both mouse and rat striatal synaptosomes led to phosphorylation of DAT (41,4), resulting in increased extracellular DA concentrations.
Furthermore, an increase in intracellular calcium can also initiate the release of DA from synaptic vesicles docked at the plasma membrane (43). Thus, toxicant-induced elevations in intracellular calcium may lead to both inhibition of DA uptake, mediated by DAT, and enhanced release of intracellular DA-events that ultimately result in decreased DA synthesis (32,33).
It is also conceivable that the observed interactive effects on punch DA concentrations are toxicokinetic in nature (i.e., either MeHg increases the intracellular concentrations of PCBs or PCBs increase intracellular concentrations of MeHg). However, we believe that elevated toxicant levels were not responsible for the observed interactions, particularly in punches exposed to either 4 or 10 pM MeHg plus PCBs, for the following reasons. First, in preliminary studies (data not shown) there were no differences in PCB content between PCB-and PCB + MeHg-exposed punches. Second, there were no detectable differences in media LDH concentrations, except for the 14 ,uM + PCB-exposed punches. These data not only suggest that the preparations were viable, but also suggest that MeHg exposure, at least at the 4 and 10 pM MeHg concentrations used in these experiments, did not significantly alter membrane permeability and, by inference, increase intracellular concentrations of the toxicants.
In summary, we have shown that exposure of rat striatal punches to PCBs and MeHg decreases punch DA content and elevates media DA levels to a greater extent than exposure to the single toxicants alone in the absence of significant changes in media LDH concentrations, which provides a recognized measure of both the integrity of the plasma membrane and the viability of the striatal punch. The magnitude of the decrease in punch DA content after combined exposure to MeHg and PCBs (a 20-50% decrease depending on the level of the contaminants studied) is not as great as that seen in rat brain after developmental exposure to the complex mixture of contaminants found in Great Lakes salmon (10). Although there are methodologic differences between these studies (e.g., in vivo vs. in vitro), we suggest that a major difference is related to the greater number of putative neurotoxicants present in the contaminated fish as compared to the two neurotoxicants studied here. Examination of binary pairs of contaminants (e.g., MeHg + PCBs) is but a preliminary step in the investigation of the neurotoxic potential of complex contaminant mixtures. Thus, we have begun to examine functional interactions between PCBs and other fish-borne contaminants, including the chlorinated pesticide chlordane, and we demonstrated similar changes in striatal punch DA function. Examination of the neurotoxic actions of tertiary and higher order combinations of toxicants provides a logical next step toward understanding the actions of complex mixtures of environmentally relevant toxicants on nervous system function. Thus, our results warrant examination of the impact of other environmental mixtures on human health, suggest that future fish-consumption advisories consider the likelihood that environmental contaminants may act synergistically, and provide a strong rationale for examining for possible interaction effects of this and other contaminant mixtures in the intact developing animal. 9