Results of the Pacific Coast portion of the NBSP for Cycles I-V (1984-88) indicated that multiple bottomfish species (including flathead sole, English sole, starry flounder, white croaker, and black croaker) inhabiting areas proximate to highly urbanized or industrialized regions were affected by pathological conditions in the kidney and especially the liver at prevalences that are significantly higher than those detected in the same species from relatively uncontaminated, nonurban reference or comparison sites. Moreover, in multivariate analyses of disease risk in individual fish that control for age and gender, fish from these urban sites showed increased relative risks of kidney and especially liver lesions. In addition to this general geographic pattern of lesion distribution, specific lesion categories in both the liver and kidney of these fish species were significantly associated, in logistic regression analyses that simultaneously control for mean age and gender ratio at the sites, with measures of multiple contaminant groups present in sediments, and in stomach contents, liver tissue and bile from fish captured at these sites. In individual fish, the relative risk of occurrence for a number of lesion categories in both liver and kidney increased significantly with increasing fish age, pointing to the necessity of accounting for fish age in intersite comparisons of lesion prevalences, or in analyses relating these prevalences to parameters of potential or actual exposure to chemical contaminants. In contrast, only one lesion type in a single species in the liver (neoplasms in white croaker) or kidney (necrosis in English sole) was associated with fish gender in any species.

Age as a Risk Factor

The risk of occurrence of several hepatic lesion categories in multiple species examined in this study increased significantly with age, consistent with previous results showing a significant influence of age on lesion prevalence in similar studies in bottomfish species (Baumann et al. 1991 (walleye, Stizostedion vitreum and brown bullhead, Ameiurus nebulosus); Murchelano and Wolke 1985, and Moore 1991 (winter flounder); Rhodes et al. 1987, and Barrick et al. 1985 (English sole); Kranz and Peters 1985 (ruffe, Gymnocephalus cernua); and Sloof and Zwart 1982 (bream, Abramis brama)). Specifically, the relative risk of neoplasm occurrence in individual English sole increased directly with age, consistent with previous studies on this species in Puget Sound (Rhodes et al. 1987, Barrick et al. 1985). The risk of neoplasm occurrence also significantly increased with age in individual hornyhead turbot and white croaker. Similarly, preneoplastic foci of cellular alteration and nonneoplastic proliferative lesions were significantly more likely to occur in older starry flounder and white croaker; this relationship paralleled previous results in English sole (Rhodes et al. 1987). However, in contrast to the Rhodes study, the present results indicated that age was not a significant factor influencing relative risk of either foci of cellular alteration or the proliferative lesions in English sole. The most likely reasons that these lesions, especially the foci of cellular alteration, were not linked to advancing age in this study is the lower number of individuals analyzed (413 vs. 1083) and the lower lesion prevalences detected as a result of selection of a higher proportion of sampling sites categorized as relatively uncontaminated. In the study by Rhodes et al. (1987), seven of the eight sampling sites were located in highly urban areas of Puget Sound that also had sediments characterized by high levels of anthropogenic contaminants (Malins et al. 1984), and hepatic lesion prevalences were consequently far higher than those detected in the present study.

For the specific degenerative/necrotic lesions, there was no relationship between relative risk and age in individual fish of any species, consistent with previous findings in English sole (Rhodes et al., 1987). The risk of occurrence of the nonspecific necrotic hepatic lesions was positively associated with age only in individual hornyhead turbot. The reasons for this surprising finding are unclear, especially since previous epizootiological studies in English sole (Rhodes et al. 1987) have not shown the risk of hepatic necrosis to be related to increasing fish age.

The risk of hydropic vacuolation was also positively and consistently associated with increasing fish age in starry flounder and white croaker, the two species in which this lesion was commonly detected. A similarly strong positive association between this lesion and fish age has previously been established for winter flounder from the East Coast of the United States (Moore 1991; Johnson et al.1992), as well as for rock sole (Pleuronectes bilineata ) from Puget Sound, Washington (Myers et al. 1992), suggesting that this contaminant-associated lesion may reflect the cumulative effects of long-term exposure to particular hepatotoxic contaminants, as postulated by Moore (1991).

Overall, the most meaningful positive associations between fish age and increased risk of particular lesion types among the species examined existed for neoplasms, foci of cellular alteration, nonneoplastic proliferative lesions, and hydropic vacuolation.

Gender as a Risk Factor

Gender, as determined in individual fish, was not consistently shown as a risk factor for any hepatic lesion category among the species examined; the single exception was the predominance of hepatic neoplasms detected in male white croaker. Only six individuals in this species for which gender data were available were affected with hepatic neoplasms, of which five were males. Considering the low number of cases upon which this potential relationship was based, gender as a risk factor for neoplasms in white croaker cannot yet be regarded as conclusively established. Similar studies have not generally shown any effect of gender on risk of hepatic lesions in pleuronectids (Moore 1991, Rhodes et al. 1987, Barrick et al. 1985), although recent studies in winter flounder suggest gender-associated differences in neoplasm prevalences (males were more likely to be affected), possibly as a result of differences in migratory behavior of male and female flounder (Johnson et al.1992). The findings of the present study, therefore, generally confirm the lack of importance of fish gender as a risk factor for idiopathic hepatic disease in wild fish.

Chemical Measures in Sediment, Stomach Contents, Liver Tissue, and Bile as Risk Factors

In general, hepatic lesions were most likely to occur in fish captured at sites with chemically contaminated sediments, and all classes of hepatic lesions in at least one species showed significant, positive risk factors for chemical groups measured in sediment, stomach contents, liver tissue, or bile (FACs-L and FACs-H). The discussion below represents an initial attempt to interpret the significance of the numerous and complex associations identified between hepatic lesion types and parameters of contaminant exposure, by using logistic regression analyses to relate lesion prevalences to measures of potential contaminant exposure, dietary uptake, hepatic bioaccumulation, and uptake and metabolism of aromatic compounds determined at the sampling sites, especially in terms of the toxicologic meaning that can be attributed to these associations. This analysis is not intended to determine, nor is it capable of assessing, the chemical etiology of the lesions detected among the multiple species examined; rather, the objective of this analysis was to provide clues regarding which of the many contaminants these species are exposed to may play a role in the development of lesions, and to provide an epizootiological basis for more definitive cause-and-effect laboratory studies. The data on chemical analyses of sediments, stomach contents, and liver tissue were derived from analyses of composite samples collected from sediments and individual fish captured each time a site was sampled, rather than from individual animals. Moreover, the biliary FACs data was in the form of mean site-specific FACs-L and FACs-H levels determined from individual analyses of 10-12 specimens of a species in a particular year. Hence, it was not possible in this analysis to assess the influence of contaminant exposure on risk of disease in individual fish, and consequently the number of samples within a species that were available for this analysis was relatively low. Another factor significantly limiting meaningful interpretation of the toxicologic relevance of these associations is the paucity of information relating hepatic tissue burdens of the chemical contaminants measured in this study to pathological effects in fish or other vertebrates.

Whenever possible, any associations shown between lesion types and measures of actual exposure to contaminants via dietary uptake (levels in stomach contents), bioaccumulation (hepatic levels), or metabolism (biliary FAC levels) have been regarded as more toxicologically relevant than associations between lesion prevalences and measures of potential exposure (i.e., contaminants in sediments). However, because an estimated measure of longer term exposure to AHs, such as persistent xenobiotic-DNA adducts in liver assessed by the 32P-postlabelling technique (Varanasi et al. 1989b), was not yet available for this data set, AHs in sediments and recent AH exposure as reflected in mean biliary FAC levels were considered as measures of AH exposure. Another important factor complicating the interpretation of the chemical exposure data was the universally lower sample size for measures of contaminants in liver tissue and stomach contents as compared to sediments, so that risk factors in sediments were more commonly identified and were perhaps more statistically reliable. In cases where the level of a contaminant class in sediment was identified as a risk factor and was significantly correlated with levels in liver or stomach contents, but the same contaminant class in liver or stomach contents was not identified as a risk factor, the risk factor in sediments was nonetheless considered toxicologically meaningful. Actual contaminant levels shown are derived from the data presented in a companion Technical Memorandum documenting contaminant levels in sediments and fish tissues from the NBSP, Cycles I-V (Landahl et al., in prep.).

Hepatic neoplasms--Because neoplasms were predominantly detected in English sole, significant contaminant-exposure related risk factors demonstrated by logistic regression were identified primarily in this species. The most common risk factors consistently detected among the compartments were the AHs, including the lower and higher molecular weight AHs and total AHs, especially in sediments. These relationships support the one established between neoplasm prevalence and mean biliary FAC levels in this species in the present study and are consistent with the association between neoplasms and AH exposure shown in previous field and laboratory studies with English sole (Malins et al. 1984; Krahn et al. 1986b; Landahl et al. 1990; Myers et al. 1990, 1991; Schiewe et al. 1991), in studies in other species of wild fish (Black 1983; Baumann et al. 1982, 1991), and by the demonstrated hepatocarcinogenicity of genotoxic AHs in other laboratory studies (Black et al. 1985, 1988; Hendricks et al. 1985; Schultz and Schultz 1984; Metcalfe et al. 1988, 1990; Hawkins et al. 1990). Total AH levels in sediment ranged from undetectable at the Nisqually River site to 5,900 ng/g in Elliott Bay, comparable to levels detected in the previous studies where the relationship between AH exposure and neoplasm occurrence was first established in English sole (Malins et al. 1984), as well as previously reported data for Pacific Coast NBSP sites (Varanasi et al. 1988, 1989a). The fact that AHs in stomach contents were not similarly associated with neoplasms is not regarded as anomalous, considering the much lower sample size for stomach contents (9) than for sediments (22).

In sediments and most importantly in liver tissue, the covarying PCBs also were contributory factors for neoplasm risk in English sole, confirming previous findings in this species (Malins et al. 1984). The PCB levels in sediments (up to 500 ng/g in Elliott Bay ) and liver (up to 11,000 ng/g in Elliott Bay English sole) appear to be toxicologically meaningful, independent of their covariance with the AHs. Polychlorinated biphenyls are not regarded as potent genotoxic initiators of hepatic carcinogenesis (Hayes et al. 1985; Silberhorn et al. 1990), do not readily form covalent adducts with cellular DNA, and show minimal mutagenic activity as either individual congeners or complex PCB mixtures (Safe 1984, 1989). The more highly chlorinated PCB mixtures are hepatocarcinogenic in rodents, but they are thought to act not as complete carcinogens or initiators but primarily as potent epigenetic promoters of hepatic carcinogenesis in mammals (Williams and Weisburger, 1986; Safe 1984, 1989) by virtue of induction of cytochrome P450 enzymes, with consequent hepatotoxicity and stimulation of cell proliferation (O'Brien 1967; Zinkl 1977; Silberhorn et al. 1990). This mechanism is especially applicable to the toxic coplanar congeners (Silberhorn et al. 1990) often present in complex PCB mixtures. Results of laboratory studies on the modulating effects of PCBs on hepatocarcinogenesis in fish have shown variable results, dependent upon the timing of administration and the type of initiator used. In aflatoxin-induced hepatic neoplasia in rainbow trout, PCBs are profoundly inhibitory to hepatocarcinogenesis (Hendricks et al. 1977; Shelton et al, 1983, 1986), whereas they enhance hepatocarcinogenesis when fed simultaneously with diethylnitrosamine (DEN) (Shelton et al. 1984) or have no effect on DEN hepatocarcinogenicity when fed prior to DEN exposure (Fong et al. 1988). Morever, recent studies have shown strong promotional activity for PCBs in hepatocarcinogenesis initiated by dimethylbenz[a]anthracene in rainbow trout (Hendricks et al. 1990). The existence of PCBs as a risk factor for neoplasm occurrence in wild English sole suggests a similar promotional role for these compounds in hepatocarcinogenesis that may be independent of exposure to and covariance with the genotoxic initiators represented among the AHs in sediments. High prevalences of toxicopathic lesions, including neoplasms, have been found in English sole captured from a site in Puget Sound with sediments containing extremely high levels of AHs but low levels of PCBs (Malins et al. 1985b), and similar toxicopathic lesions were rarely detected in winter flounder captured from a site in New Bedford Harbor with high PCB levels in sediments and liver tissue (3,000 ng/g and 39,000 ng/g, respectively) and relatively low AH levels in sediment (Johnson et al. 1992). Consequently, while PCBs may play a promotional role in the development of hepatic neoplasms or function as hepatotoxicants in concert with other contaminants such as AHs, PCB exposure alone is not likely to be sufficient to induce the toxicopathic hepatic diseases in the bottomfish species examined in this study.

The DDTs in liver tissue were also a strong risk factor for hepatic neoplasms, with levels ranging from 51 ng/g in composited English sole livers from the Nisqually River reference site to 1,100 ng/g in Elliott Bay sole. The latter levels are regarded as toxicologically significant, and considering the hepatocarcinogenic or promotional effects of this nongenotoxic compound in rodents (Williams and Weisburger 1986; IARC, 1974, 1982) and fish (Halver, 1967), the DDTs should be regarded as a potential etiologic factor in hepatic neoplasia in English sole, even in light of their covariance with liver PCB levels.

Dieldrin in sediments was also a risk factor for neoplasms in English sole, but considering the extremely low levels detected (undetectable to 2 ng/g in Elliott Bay), the fact that dieldrin in liver was not a risk factor, and the relatively strong correlation with sediment levels of other risk factors for hepatic neoplasms (AHs and Metals 1), this compound is not considered a strong candidate as an etiologic agent of hepatic neoplasms in English sole. Moreover, previous studies in rainbow trout on the carcinogenic/co-carcinogenic or promotional effects of dietary dieldrin (5 ppm) on aflatoxin-induced carcinogenesis showed only a slight, but statistically insignificant co-carcinogenic effect, and dieldrin failed to induce any histopathologic effects in rainbow trout liver when fed alone (Hendricks et al. 1979). The carcinogenic potential of this nongenotoxic cyclodiene pesticide is not unequivocal in mammalian systems, and remains a subject of controversy (Murphy, 1986).

Metals 1 (summed levels of copper, zinc, lead and tin) in sediments are considered toxicologically implausible as etiological factors for hepatic neoplasia. These metals are unlikely etiological factors because relatively low levels are detected (up to 420 ng/g in Elliott Bay), they lack hepatocarcinogenic potential (Goyer 1986), and they strongly covary with sediment risk factors with known initiating or promotional potential in hepatocarcinogenesis such as AHs, PCBs, and DDTs. The absence Metals 1 as a risk factor indicating actual hepatic bioaccumulation is additional evidence that they are not likely to be involved in the development of hepatic neoplasms. However, although a detailed treatment of the processes that regulate levels of bioaccumulation of the above metals in fish is beyond the scope of this discussion, teleosts are known to effectively regulate tissue levels of trace elements such as lead, copper, and zinc via various mechanisms that alter their uptake, transport, sequestration, detoxification, and excretion, such that simple tissue levels of these elements may not be accurate measures of degree of exposure (Phillips and Rainbow 1989). Furthermore, many toxic metals exhibit a rapid turnover in their target tissue in mammalian vertebrates, and no simple relationship exists between administered dose and effective target dose for many metals, making it difficult to relate an observed lesion to concentration of a trace element in tissue or to apply the concept of critical target tissue concentration to the evaluation of metal toxicity (Foulkes 1990).

In summary, the xenobiotic compounds most likely to be etiologic factors in the development of liver neoplasms are the AHs, PCBs, and DDTs.

Because black croaker were only introduced as a target species in 1987, insufficient data existed for a meaningful logistic regression analysis of the contaminant exposure-related parameters at sites where black croaker were captured. Nonetheless, sites (Shelter Island and north San Diego Bay) where significantly higher prevalences of neoplasms and foci of cellular alteration in black croaker were detected, as compared to the reference site at outside Mission Bay, also showed significantly higher levels of AHs, PCBs and DDTs in sediments, and PCBs and DDTs in liver (McCain et al. 1992). These results are in general agreement with the types of chemical contaminants that were risk factors for neoplasms in English sole, as discussed above, as well as for foci of cellular alteration in several species, as discussed below.

Foci of cellular alteration--Putatively preneoplastic foci of cellular alteration were also associated with multiple contaminant exposure-related risk factors in English sole, white croaker, starry flounder, and to a lesser extent in hornyhead turbot. In general, risk factors for this lesion type were similar to those for the neoplasms, which would be expected considering the generally accepted preneoplastic potential attributed to these lesions (Farber and Sarma 1987; Williams 1987; Zerban et al. 1989). Consequently, similar toxicologic significance can be ascribed to the chemical exposure-related risk factors for foci of cellular alteration as those for the neoplasms. As was the case for hepatic neoplasms, the most common and important risk factors were AHs, especially for this lesion category in English sole and white croaker. For example, the highest levels of total AHs in sediment from sites where white croaker were captured were detected at Hunters Point (10,000 ng/g), where a significant prevalence of foci of cellular alteration was also shown. These findings parallel the importance of mean biliary FACs (FACs-H and L for English sole, FACs-L only for white croaker) as risk factors for foci of cellular alteration in these species.

Several chlorinated hydrocarbon classes were also strongly associated with the occurrence of foci of cellular alteration. Polychlorinated biphenyls in sediment and liver was a risk factor, although only in English sole, and PCBs in stomach contents was a risk factor in English sole (up to 1,100 ng/g in Elliott Bay sole) as well as starry flounder (up to 490 ng/g at the Castro Creek site in San Francisco Bay). Consistent with their promotional role in hepatocarcinogenesis, PCBs have also been repeatedly shown to enhance the formation of preneoplastic focal lesions in rodent liver (Safe 1984, 1989). The DDTs in liver tissue (English sole) and stomach contents (English sole and starry flounder (up to 140 ng/g at Castro Creek)) were also associated with this lesion group. Dieldrin in sediment as a risk factor for foci of cellular alteration in hornyhead turbot is not considered meaningful, as levels were extremely low (only as high as 1 ng/g at west Santa Monica Bay). Therefore, there were no toxicologically meaningful risk factors identified for foci of cellular alteration in hornyhead turbot.

As with the neoplasms, Metals 1 in sediment as a risk factor for foci of cellular alteration in English sole is not considered toxicologically significant and exists only because of its covariance with levels of the more meaningful risk factors represented by low and high molecular weight AHs, total AHs, and PCBs.

To summarize, as was true for the neoplasms, the covarying AHs and PCBs, as well as DDTs appear to be the most likely xenobiotic factors measured in this study that may be involved in the etiology of these preneoplastic focal lesions in the affected species. However, these three contaminant groups were not consistently identified as risk factors among the three species. The AHs were risk factors only for English sole and white croaker, whereas measures indicating actual uptake of PCBs and DDTs were risk factors only in English sole and starry flounder.

Nonneoplastic proliferative lesions--Risk factors related to contaminant exposure similar to those identified for neoplasms and foci of cellular alteration were also identified for non-neoplastic proliferative lesions, a lesion category primarily represented by hepatocellular regeneration, biliary proliferation, presumptive oval cell proliferation (Sell 1990), and cholangiofibrosis. These lesion types are inducible by exposure to hepatotoxic and hepatocarcinogenic chemicals in laboratory studies in mammals (e.g., PCBs (Kimbrough et al. 1972)) and fish (Schiewe et al. 1991, Hendricks et al. 1984, Nunez et al. 1990) and are regarded as biomarkers of contaminant exposure in field studies on wild fish (Myers et al. 1987, 1990, 1991, 1992; Moore 1991; Hinton and Lauren 1990b; Hinton et al. 1992; Varanasi et al. 1992a). Moreover, hepatocellular regeneration, oval cell proliferation, and biliary proliferation are thought to represent early steps involved in the histogenesis of hepatocellular and biliary neoplasia (Columbano et al. 1981, Sell 1990), acting to fix the molecular lesions in DNA induced by carcinogen exposure into the genome by cell proliferation and representing compensatory proliferative responses to the early cytotoxicity typically associated with carcinogen exposure (Farber and Sarma 1987).

Significant risk factors for the nonneoplastic proliferative lesions were shown only in English sole and white croaker. The AHs were again the most common risk factors, with all three measures of AHs accounting for significant proportions of the intersite prevalence variation in English sole (sediments,biliary FACs-L and H, and stomach contents) and white croaker (sediments and biliary FACs-L and H only).

The DDTs were a consistent risk factor in sediments, stomach contents, and liver tissue for the nonneoplastic proliferative lesion category in both English sole and white croaker. The DDTs were highly correlated with sediment AHs at sites where English sole were captured, but not at white croaker sites. Levels of DDTs at white croaker sites were as high as 670 ng/g in sediments (San Pedro Outer Harbor), with extremely high levels detected in liver (26,000 ng/g) and stomach contents (7,600 ng/g) of croaker from the same site. Hepatic levels of DDTs were as high as as 1,100 ng/g in English sole (see section above on hepatic neoplasms). Exposure to DDTs has been shown to induce liver necrosis in brown trout (Salmo trutta ) and guppies (Poecilia reticulata) (King 1962), and hepatocellular vacuolar degeneration, hypertrophy, and necrosis in four species of East Indian fish (Malthur 1962). Therefore, DDTs should be seriously regarded as a toxicologically meaningful risk factor for nonneoplastic proliferative lesions in both species, as possible sequelae to hepatotoxicity. This is especially true for white croaker considering the high bioconcentration factor within liver, a target organ for toxicity of DDTs (Murphy 1986).

The PCBs, which strongly covaried with DDTs and AHs in sediments and stomach contents and with DDTs in liver for English sole and were correlated with DDTs in sediments, liver, and stomach contents for white croaker, were also risk factors for this lesion in both species. Specifically, PCBs in both sediment (< 500 ng/g in English sole and < 430 ng/g in white croaker) and stomach contents (< 1,100 ng/g in English sole and < 3,500 ng/g in white croaker) were a risk factor in both species; as a variable indicating actual bioaccumulation (liver), PCBs were a risk factor only in English sole (< 11,000 ng/g). In view of the probable mitogenic effects of PCBs on the liver (Silberhorn et al. 1990), their promotional potency in hepatocarcinogenesis initiated by certain genotoxic compounds, and the documented induction of lesion types included within this category such as cholangiofibrosis (Kimbrough et al. 1972; Kimbrough and Linder 1974), exposure to PCBs must be considered a significant factor in the etiology of these proliferative lesions, especially in English sole.

Chlordanes and dieldrin were also risk factors for the proliferative lesions, but only in English sole. However, for the same reasons discussed above for the foci of cellular alteration, and in view of the relatively low levels detected in English sole stomach contents (up to 50 ng/g chlordanes at Commencement Bay) and sediments (up to only 3 ng/g chlordanes at West Santa Monica Bay and 2 ng/g dieldrin at Elliott Bay), these pesticides are not regarded as toxicologically meaningful etiologic factors. Dietary uptake of hexachlorobenzene is also not considered a meaningful risk factor in white croaker in light of the low levels detected (< 10 ng/g). Similarly, the association of Metals 1 with this lesion category in English sole is probably not relevant to the etiology of lesions within this class, for the same reasons as stated above for the neoplasms and foci of cellular alteration.

In summary, the AHs, PCBs, and DDTs again appear to be the most toxicologically meaningful factors associated with these nonneoplastic proliferative lesions.

Specific degenerative/necrotic lesions--A number of risk factors in several species were also identified for specific degenerative/necrotic lesions (SDN), represented primarily by megalocytic hepatosis and hepatocellular nuclear pleomorphism. These lesion types are inducible in rodents and fish by exposure to a spectrum of hepatotoxic/carcinogenic or promoting compounds, as reviewed in Myers et al.(1987), including AHs, PCBs (Koller and Zinkl 1973; Zinkl 1977; Kimbrough and Linder 1974), DDTs, aflatoxin B1 and other naturally occurring hepatotoxic compounds including microcystin, a toxin present in the blue-green alga Microcystis (Michael Kent, Fisheries and Oceans Canada, Pacific Biological Research Station, Nanaimo, British Columbia, Canada V9R 5K6. Pers. commun., May 1992). These lesions have also been induced in English sole by multiple injections of an organic-solvent fraction from a sediment collected at a creosote-contaminated sediment (Schiewe et al. 1991). The same lesions have also been statistically associated with sediment AH levels and biliary FAC levels (Krahn et al. 1986b, Myers et al. 1990) and liver PCB levels (Myers et al. 1992) in English sole, as well as with biliary FAC levels in rock sole (Myers et al. 1992). Moreover, these lesions are considered as reliable biomarkers of contaminant exposure in wild fish (Hinton and Lauren, 1990b; Hinton et al., 1992), including English sole (Myers et al., 1987, 1990, 1991, 1992) and white croaker (Malins et al. 1987a; Myers et al. 1991).

Chemical exposure-related risk factors for SDN were shown in English sole, white croaker, and starry flounder in the present study. Variables related to AH exposure were again the most frequently identified risk factors, especially in English sole and starry flounder, which is supported at least in English sole by the association shown between this lesion category and both types of biliary FACs measures. Moreover, biliary FACs-L and FACs-H were risk factors for SDN in white croaker. Sediment AHs were risk factors only in English sole and starry flounder; this association was further supported by AHs in stomach contents, but only for English sole. Levels of AHs in sediment at English sole sites have been outlined above and are toxicologically significant; for starry flounder sites, total AH levels in sediment ranged from 5 ng/g at the Chukchi Sea site to 10,000 ng/g at Hunters Point, the latter level interpreted as toxicologically significant.

Consistent with their general covariance with AHs in sediment, the PCBs in sediment were again shown as positive risk factors for the specific degeneration/necrosis lesion category in all three species; this relationship was supported by positive associations between lesion occurrence and bioaccumulation of PCBs in liver (English sole, starry flounder, and white croaker) and exposure via dietary components (English sole and white croaker). Levels of PCBs in all three compartments have been mentioned previously for English sole and white croaker (except liver levels, which were up to 15,000 ng/g at Long Beach) and are interpreted as toxicologically relevant; sediment PCB levels at starry flounder sites ranged from 1 ng/g at Bodega Bay to 140 ng/g at Hunters Point, with PCBs in liver ranging from 230 ng/g in Chukchi Sea flounder to 7,000 ng/g at Hunters Point. This level of exposure to PCBs and extent of bioconcentration in liver should also be considered toxicologically meaningful, especially considering that experimental PCB exposure in mammals can induce hepatic lesions within this category (Frith and Ward 1980; Koller and Zinkl 1973; Zinkl 1977; Kimbrough and Linder 1974). Moreover, PCB exposure in teleosts has also induced hepatic lesions similar to those included within this category (Hinton et al. 1978).

The DDTs were also consistent risk factors in sediments, stomach contents, and liver tissue for SDN in English sole and white croaker. Although levels covaried with PCBs within each compartment measured for white croaker, DDTs typically accounted for a higher proportion of the intersite prevalence variation in specific degeneration/necrosis than did the PCBs, suggesting DDT exposure as a more likely factor in the etiology of this lesion type in this species. Moreover, dietary exposure and hepatic bioaccumulation of DDTs (up to 7,600 ng/g and 26,000 ng/g, respectively, at west Santa Monica Bay) are certainly high enough to have toxicologic significance. Hepatocellular cytomegaly and karyomegaly have been experimentally induced in mice by exposure to DDT (Frith and Ward 1980).

Individual non-DDT pesticides were also risk factors for SDN in English sole, starry flounder, and white croaker. However, these risk factors were inconsistently identified among the compartments when consistency might have been expected on the basis of correlations among the compartments. For example, chlordanes were a risk factor in English sole only in stomach contents and sediment, whereas correlations among all compartments were especially strong (Table 21). Because of this lack of consistency, and the relatively low levels detected, chlordanes are not regarded as a meaningful risk factor for SDN in English sole. Chlordanes (in sediment and stomach contents) were also a risk factor in white croaker; sediment levels ranged from undetectable to 15 ng/g at Long Beach. Levels of chlordanes in stomach contents of white croaker were higher: up to 160 ng/g at Long Beach. Nonetheless, considering these relatively low levels of exposure, the lack of expected risk factor consistency based on strong intercorrelations among the compartments (Table 34), the absence as a risk factor showing actual bioaccumulation in liver tissue, and the strong correlations with PCB levels among the compartments (Tables 32 and 33), chlordanes are not a likely etiologic factor for SDN in white croaker. In starry flounder, both chlordanes and dieldrin in liver were also risk factors for SDN. The highest hepatic levels of chlordanes and dieldrin detected, respectively, were 330 and 300 ng/g at Hunters Point, demonstrating a degree of actual hepatic bioaccumulation of these pesticides. Dieldrin can induce hepatocellular hypertrophy in experimental mammals (Hodge et al. 1967) and hepatic histopathologic changes such as necrosis, hepatocellular vacuolar degeneration and hypertrophy, and hepatocyte pleomorphism in fish (Malthur 1965, 1975); furthermore, it is considered hepatocarcinogenic in mice (IARC 1974). Chlordanes also can induce histopathologic effects in liver of experimental mammals (Ingle 1965), and fish (Eller, unpubl. observ. in Couch 1975), and are considered epigenetic promoters of carcinogenesis (Williams and Weisburger 1986). The levels detected in liver of starry flounder suggest that these pesticides could be contributory factors in the genesis of SDN in this species.

Finally, potential exposure to or uptake of trace metals in the group Metals 1 appeared to influence the prevalence of SDN in English sole, starry flounder, and white croaker. However, because of the strong covariance of Metals 1 levels with more toxicologically relevant risk factors, such as AHs, DDTs and PCBs for English sole (sediments and stomach contents, Tables 18 and 19), AHs and PCBs for starry flounder (sediments only, Table 25), and DDTs and especially PCBs for white croaker (sediments and stomach contents, Tables 32 and 33), as well as the low degree of bioaccumulation of these metals in liver (generally below 300 ng/g), these associations are probably not toxicologically meaningful. However, tissue concentration of metals within this group may not be an accurate index of exposure (Phillips and Rainbow 1989; Foulkes 1990). Metals 2 (summed levels of nickel, chromium, and selenium) in liver (up to only 59 ng/g at Dana Point) as a risk factor for SDN in white croaker is also probably of little toxicologic importance and exists only because of the strong correlation between this risk factor and liver DDTs (Table 35).

In summary, as was the case for the preceding hepatic lesion types, SDN was most meaningfully associated with exposure to AHs, PCBs, and DDTs, with the single exception of starry flounder, where levels of dieldrin and chlordanes bioaccumulation in liver are high enough to have potential toxicologic significance.

Necrosis--Chemical exposure-related risk factors were rarely and inconsistently identified for the nonspecific necrotic lesions represented primarily by hepatocellular coagulative necrosis. Although this specific lesion type is regarded as a relatively reliable biomarker of contaminant exposure that needs further confirmation in wild fish (Hinton and Lauren 1990b, Hinton et al. 1992) and is inducible by exposure to a number of hepatotoxic agents, statistical associations between this lesion and contaminant exposure in field studies conducted with several species by our group have been inconsistently identified (Malins et al. 1980, Rhodes et al. 1987, Myers et al. 1991, Myers unpubl. data). In the present study, the only risk factor that was common to more than one species (English sole and white croaker) was DDTs in liver, an index of actual uptake and bioconcentration. The hepatotoxicity of DDT is well documented in mammals (Murphy 1986) and fish (King 1962, Malthur 1962), and the toxicologic significance of levels of exposure in these two species has been previously discussed. No other risk factors for this lesion in English sole were shown. Other risk factors associated with necrosis in white croaker were AHs and PCBs in sediment (at levels that were toxicologically significant), biliary FACs-H, and hexachlorobenzene in stomach contents. Despite the hepatotoxic potential of hexachlorobenzene (Murphy 1986), the extremely low levels detected in this study (< 10 ng/g) indicate that its association with necrosis in white croaker is probably spurious. Similarly, the association of Metals 1 in liver tissue with necrosis in starry flounder is also of questionable meaning, since the metals included in this group are not regarded as potent hepatotoxicants (Goyer 1986), and levels in starry flounder liver were relatively low (< 280 ng/g at Southampton Shoal ).

In summary, with the possible exception of DDTs, no chemical exposure parameter was consistently and meaningfully associated with necrosis in more than one species. Therefore, we regard this nonspecific lesion type as a less reliable biomarker of contaminant exposure, even in species that are apparently susceptible to the effects of such exposure as evinced by the presence of other hepatic lesions.

Hydropic vacuolation--In both species in which this unique lesion type was commonly detected (starry flounder and white croaker), multiple risk factors related to contaminant exposure were identified in sediments, stomach contents, liver tissue, and bile. We are aware of no exactly analogous lesion in any vertebrate model system that has been experimentally induced by exposure to specific toxicants. However, this is a presumably degenerative lesion with limited proliferative potential as shown in winter flounder (Moore 1991); it affects primarily biliary epithelial cells and also hepatocytes in more severe cases, as documented in winter flounder (Murchelano and Wolke 1985, 1991; Bodammer and Murchelano 1990; Moore et al.1989; Moore 1991; Stehr 1990), white perch (Camus and Wolke 1991); starry flounder and rock sole (Stehr 1990, Stehr et al. 1991), and here reported in white croaker. However, unlike in winter flounder, this lesion in starry flounder and white croaker does not commonly co-occur with neoplasms. The geographic distribution of this lesion type in winter flounder (Murchelano and Wolke 1985, 1991; Moore 1991; Johnson et al.1992), rock sole and starry flounder (Myers et al. 1992), as well as its documented statistical association with indices of contaminant exposure in winter flounder (Johnson et al. 1992) and starry flounder (Myers et al. 1992) strongly indicate this lesion is a reliable and highly specific biomarker of contaminant exposure effects in wild fish, as suggested by others (Moore 1991, Hinton and Lauren 1990b, Hinton et al. 1992). Specifically, risk factors of chemical exposure in winter flounder for this urban-associated lesion type are AHs, DDTs, and chlordanes (Johnson et al. 1992); in starry flounder from Puget Sound the lesion is associated with AH exposure, as reflected in biliary FAC-H levels (Myers et al. 1992). In general, the data in the present report reinforce these relationships. In parallel with the association shown between this lesion and levels of biliary FACs-L and FACs-H in starry flounder in the present report, AH exposure and dietary uptake (from sediments and stomach contents data) were significant risk factors in this species; in contrast, the only risk factor indicating AH exposure in white croaker was levels of biliary FACs-L. However, this latter finding is not surprising in light of the complete lack of correlation between levels of both FAC measures and AH measures in either the sediments or stomach contents (Table 34)

Chlorinated hydrocarbons were more frequently occurring risk factors for hydropic vacuolation in both species. Positive risk factors found at levels with toxicologic significance were DDTs in stomach contents and liver (starry flounder), and PCBs in sediment, stomach contents (starry flounder only), and liver (both species). Several non-DDT pesticides, including hexachlorobenzene, chlordanes and dieldrin, were also risk factors for hydropic vacuolation. In starry flounder, hepatic levels of chlordanes and dieldrin were as high as 330 ng/g and 300 ng/g, respectively, at Hunters Point, while chlordanes in stomach contents (< 17 ng/g) and hexachlorobenzene in sediment (< 1 ng/g) were quite low. Based on the hepatic levels of both chlordanes and dieldrin, these chemicals cannot be disregarded as toxicologically significant risk factors of hydropic vacuolation in starry flounder. However, experimental attempts to induce this lesion type in winter flounder by intraperitoneal injection or dietary exposure to high levels of chlordanes were unsuccessful (Moore 1991). In white croaker, the only non-DDT pesticide risk factor was hexachlorobenzene in sediments (< 1 ng/g ) and liver tissue (up to 9 ng/g at Long Beach); hexachlorobenzene exposure, therefore, appears to have no toxicologic significance. Metals 1 in sediments as a risk factor in starry flounder is similarly interpreted as having no toxicologic significance; it simply was a covarying factor with AHs and PCBs.

In summary, AHs, PCBs, DDTs, chlordanes and dieldrin in starry flounder, and AHs and PCBs in white croaker, appear to be the most toxicologically relevant xenobiotic risk factors for hydropic vacuolation. These risk factors of chemical exposure generally agree with the original suggestion (Bodammer and Murchelano 1990) that this lesion type may be related to exposure to industrial organochlorines and with subsequent hypotheses put forth by Moore (1991) that stress the potential role of epigenetic promoting compounds in the genesis of this lesion in winter flounder. However, the role of AH exposure in the genesis of this lesion type should not be minimized, considering the fact that low prevalences of hydropic vacuolation (<7%) were detected in winter flounder from a site in New Bedford Harbor that had relatively high (3,000 ng/g) sediment levels of PCBs and low to moderate concentrations of AHs in sediment (<2500 ng/g HAHs, <500 ng/g LAHs). The role of AH exposure in the development of this lesion is further complicated by the unsuccessful attempts to experimentally induce this lesion in winter flounder by chronic dietary exposure to benzo(a)pyrene (Moore 1991). In addition, because PCBs and DDTs are readily taken up and extensively bioaccumulated but slowly metabolized, hepatic concentrations of these and other chlorinated compounds may reflect chronic exposure to other contaminants that co-occur with them in sediments, but which are readily metabolized and not bioaccumulated, such as the AHs (Varanasi et al. 1992b). Future incorporation of bioindicators that estimate chronic exposure to AHs (i.e., xenobiotic-DNA adducts) should help to clarify the relative importance of AHs, with respect to the chlorinated hydrocarbons, in the development of this lesion type in affected feral fish. However, laboratory studies with starry flounder, white croaker, and winter flounder exposed to whole sediments, organic-solvent extracts of urban sediments from sites showing high prevalences of this lesion, or model compounds or mixtures are needed to definitively determine how various classes of AHs and chlorinated hydrocarbons, acting either alone or together, contribute to the genesis of hydropic vacuolation in affected teleosts.

Fluorescent Aromatic Compounds (FACs) in Bile as Risk Factors in Individual Fish

In this first attempt to relate AH exposure to hepatic lesion risk in individual fish, levels of FACs in bile, especially metabolites of the lower molecular weight aromatic hydrocarbons fluorescing at naphthalene wavelengths, were associated with an increased relative risk for certain hepatic lesions, but only in individual English sole and white croaker. These results generally agreed with previously demonstrated statistical or inferential associations between idiopathic hepatic lesion prevalences and mean levels of biliary FACs as an estimate of recent exposure to AHs in English sole, rock sole and starry flounder (Krahn et al. 1984, 1986b; Myers et al. 1990, 1991, 1992) and white croaker (Malins et al. 1987a). Although these findings support biliary FACs measurement as a useful estimate of AH exposure, as well as strengthen the linkage between AH exposure and the presence of hepatic lesions in several species, biliary FACs only estimate recent exposure and are not a direct reflection or predictor of hepatic lesions (Myers et al. 1990) that are thought to be inducible by subchronic to chronic exposure. The preceding observation is especially true for migratory bottomfish species such as winter flounder (Johnson et al. 1992). Moreover, no evidence exists to suggest that differences in FAC levels result from alterations in AH metabolism, bile production, or other biliary function caused by the presence of hepatic lesions (Myers et al. 1990).

Hepatic lesions showing an increased relative risk in individual fish due to biliary FAC levels were SDN in English sole and white croaker, and foci of cellular alteration in white croaker; biliary FAC level was not a risk factor in individual fish for neoplasms in either English sole or white croaker, nor was it a risk factor for foci of cellular alteration in the other species in which this lesion type was detected (e.g., starry flounder and English sole) or for any of the other hepatic lesion categories in any species. In all of these cases increased risk was associated only with levels of FACs-L. The inconsistency shown between biliary FAC levels and hepatic lesion risk in analyses done on individual fish in comparison to those performed on a site basis is of concern considering the strong linkages between lesion prevalences and biliary FACs-H levels previously shown in English sole (Krahn et al . 1984, 1986b) from Puget Sound, Washington. However, this previously established linkage between lesion prevalence and levels of biliary FACs-H was established in studies where the majority of sampling sites were characterized by high levels of high and low molecular weight AHs in the sediments as well as higher mean levels of biliary FACs-H, and where the prevalences of hepatic lesion types among these contaminated sites were higher (e.g., >20% neoplasms, >32% foci of cellular alteration, >86% SDN) than those detected in the present study. The strategy of site selection for the NBSP was biased against selecting highly contaminated sites near point sources of pollutants; consequently, lower hepatic lesion prevalences were detected among the sites and species examined. Moreover, the associations previously established between neoplasms, foci of cellular alteration, SDN, and levels of biliary FACs-H were based only on Spearman rank correlations between lesion prevalence and mean levels of FACs-H, and not by logistic regression analyses that account for both age and gender.

The logistic regression analysis of lesion risk versus biliary FACs in individual fish in the present study should be regarded only as an initial attempt to relate lesion risk to AH exposure in individual fish and was done because this measure of chemical exposure was the only one available in individual fish. There are several problems inherent to this approach that may account for the inconsistent and insignificant results obtained in this analysis. First, the logistic regression analyses on individual fish are not fully inclusive of the histopathology data, because bile was only collected and analyzed from a small subset (generally < 30%) of the specimens histologically examined and reflected in the site-specific lesion prevalence data. For example, the data set for English sole from which bile was collected was composed of only 166 individuals, in comparison to the more than 500 fish reflected in the site-specific prevalence data. Second, the frequency of occurrence of lesions such as neoplasms and foci of cellular alteration in this data set was quite low, making it difficult to establish a linkage between lesion risk and FAC levels. Third, because biliary FAC levels only reflect recent AH exposure it would be less likely to establish a linkage between exposure and lesion types that are considered inducible over the chronic term such as neoplasms. On the other hand, lesions that are considered inducible over a shorter time frame than neoplasms (e.g., SDN and foci of cellular alteration) and that occurred more frequently than neoplasms would be more likely to be associated with biliary FAC levels. In fact, the strongest correlation shown in the original study linking biliary FAC levels to various types of idiopathic hepatic lesions (Krahn et al., 1986) was between the prevalence of megalocytic hepatosis (the major component of SDN) and mean biliary FACs-H. Similarly, the only lesions linked to biliary FAC levels in the present study by logistic regression analyses in individual fish were foci of cellular alteration in white croaker and SDN in croaker and English sole; neoplasms were not linked to biliary FACs by this individual analysis in any species.

Future studies reflecting a larger data set for biliary FACs in each primary target species and including more indices of contaminant exposure in individual fish (e.g., xenobiotic-DNA adducts as an estimate of longer term exposure to AHs) will continue to apply this approach to strengthen the linkage between hepatic lesions and AH exposure.

Summary of Relationships Between Hepatic Lesions and Risk Factors Related to Contaminant Exposure

Because of the strong covariance of AHs, PCBs, and DDTs in sediments, dietary components, or liver tissue, it is not possible to separate or quantify the relative influence of particular chemical contaminant groups with respect to the etiology of the hepatic lesions in the species examined. The existence of the PCBs and DDTs as strong risk factors for several classes of hepatic lesions in English sole, white croaker, and starry flounder may be due to their well known bioaccumulation and persistence in fish tissue, thus serving as general indices of long-term exposure to not only chlorinated hydrocarbons, but also other co-occurring compounds such as AHs (Varanasi et al. 1992b). While some of the AH risk factors in sediment and stomach contents are corroborated as positive risk factors for hepatic lesions by the biliary FACs data, levels of metabolites in bile are reliable measures only of relatively recent exposure to AHs (Krahn et al. 1986a; Collier and Varanasi 1991). Since AHs are readily metabolized and do not bioaccumulate in fish tissue (Varanasi et al. 1989c; Stein et al. 1984, 1987), there is no currently reliable measure of chronic AH exposure, with the possible exception of xenobiotic-DNA adducts (Varanasi et al. 1989b). Unfortunately, parallel data on xenobiotic-DNA adducts for individual fish or composites on a by-site basis were not available for inclusion in the present study. Consequently, it was not possible to assess chronic exposure to AHs in the species examined in this study, possibly explaining some of the inconsistencies in associations between AH exposure and certain hepatic lesion types. Recent addition of xenobiotic-DNA adduct measurements into the NBSP and incorporation into multivariate logistic regression analysis in future studies should help to strengthen the relationship between AH exposure and certain hepatic lesions. While carcinogenic compounds such as AHs may need to be present at certain threshold levels to initiate the hepatocarcinogenic process in feral fish, the importance of the chlorinated hydrocarbons in the etiology of these lesions must also be recognized on the basis of their hepatotoxicity, epigenetic promotional activity in hepatocarcinogenesis, and the general chronic proliferative effects due to continuous exposure to these compounds.

Age as a Risk Factor

The relative risk for certain renal lesion categories in several species was significantly influenced by fish age with results that were generally consistent with those of previous studies in bottomfish species from Puget Sound. For example, in English sole, renal lesions within the categories of sclerotic, necrotic, and proliferative lesions all have shown a significantly increased relative risk attributable to fish age, ranging from 1.2 (necrotic lesions) to 1.5 (sclerotic or depositional lesions) for each additional year of age (Rhodes et al. 1987). The present results showed that older fish had an increased risk of necrotic lesions in English sole, starry flounder, white croaker, and black croaker. The sclerotic lesions, primarily represented by mesangiosclerosis, were also more commonly detected in older flathead sole, English sole, and starry flounder. Proliferative renal lesions were not associated with fish age in any species in this study. The significance of age as a factor in renal necrotic lesions is not understood. The relationship of sclerotic lesions to age is probably due to the fact that increases in mesangial matrix quantity and density (i.e., resulting in a morphologic diagnosis of mesangiosclerosis) occur as a cumulative nonspecific response to injury in a variety of chronic renal disease states (Schillings and Stekhoven 1980), as well as occurring in the aging process in vertebrates, including humans (Finn 1983; Wehner 1968). Moreover, the susceptibility of the mammalian kidney to injury has been shown to increase with age (Finn 1983).

Gender as a Risk Factor

In agreement with previous results in English sole (Rhodes et al. 1987), a relationship between fish gender and renal lesion prevalence was rarely shown by logistic regression analyses in individual fish. In fact, the only case where a lesion category preferentially affected either gender was necrotic lesions in English sole, in which males were disproportionately affected. However, this pattern was not shown in any other species affected by necrotic lesions, and a relationship between risk of any renal lesion type and gender has not been previously reported in any study in wild fish.

Chemical Measures in Sediment, Stomach Contents, Liver Tissue, and Bile as Risk Factors

Far fewer contaminant exposure-related risk factors determined on a by-site basis were identified for renal lesions than for hepatic lesions, and these associations were not often consistently shown among the compartments. Moreover, the same risk factors for a particular lesion category were rarely identified in multiple species. Meaningful interpretation of the toxicologic relevance of these associations is severely hampered by the paucity of information on the histopathologic effects of environmental contaminants on fish kidney as assessed in either field biomonitoring surveys or chronic exposure studies in the laboratory, the lack of experimental studies in fish or mammals relating tissue levels of contaminants to renal pathology, and the absence of any data in this study documenting bioaccumulation of chemical contaminants in kidney tissue. In addition, previous studies on English sole in Puget Sound have shown a significant co-occurrence of kidney lesions with idiopathic hepatic lesions (McCain et al. 1982), making it difficult to separate the well-documented association of contaminant exposure with liver lesions from those that may exist with kidney lesions. Although hepatic and renal lesion co-occurrence analyses were not done in the present study, prevalences of kidney lesions tended to be higher at sites also showing high prevalences of liver lesions (e.g., English sole and starry flounder). Nonetheless, because the fish kidney receives the major proportion of postbranchial blood, it should be more seriously considered as a target organ for the effects of environmental contaminants (Hinton and Lauren 1990a), and future studies by our laboratory will continue to assess the suggested relationships between renal lesions and contaminant exposure, as summarized below.

Proliferative lesions--Proliferative lesions in the kidney were associated with several indices of contaminant exposure measured in sediments, liver tissue, and bile. Aromatic hydrocarbon exposure, as reflected by both sediment AH levels and levels of biliary FACs-H, influenced the prevalence of this lesion category in English sole; however, AHs in sediments covaried strongly with PCBs, which were also risk factors in sediment and liver tissue for this lesion in English sole.

Polychlorinated biphenyl exposure and bioaccumulation (liver levels, up to 15,000 ng/g in Elliott Bay) were also risk factors in flathead sole. Although PCBs do not directly induce renal lesions in mammals, they do potentiate the nephrotoxicity of other compounds such as chloroform and carbon tetrachloride by induction of xenobiotic metabolizing enzymes in the kidney, leading to enhanced formation of nephrotoxic metabolites (Kluwe and Hook 1980). The PCBs, therefore, may represent toxicologically significant risk factors of renal disease in fish. Moreover, degenerative lesions have been induced in teleost kidney by exposure to Aroclor 1254 (Nestel and Budd 1975).

Exposure to DDTs also was linked to proliferative lesions in flathead sole (liver levels only, < 970 ng/g in Elliott Bay) and in white croaker, (both sediment and liver levels (< 670 and 26,000 ng/g respectively, at San Pedro Outer Harbor)). The DDTs are also not generally recognized as direct nephrotoxicants, although as an inducer of microsomal enzymes they may also act as potentiators of other environmental nephrotoxicants and have induced histological damage, such as necrosis of the renal tubular epithelium in teleost kidney (King 1962, Malthur 1962, Buhler et al. 1969).

Chlordanes and dieldrin were risk factors, but almost strictly in flathead sole, and were meaningful only in liver, since levels in sediment were exceedingly low (< 1 ng/g for English and flathead sole). Maximum hepatic levels of dieldrin in English sole and flathead sole were only 34 and 48 ng/g, respectively. However, chlordanes levels in flathead sole liver were as high as 160 ng/g at Elliott Bay. Although bioaccumulation of this pesticide was relatively low, exposure to a similar non-DDT cyclodiene pesticide, lindane (ß-hexachlorocyclohexane), has been shown to induce the degenerative lesion, glomerular hyalinosis (Wester and Canton 1986) in medaka; several cyclodiene pesticides are also nephrotoxic to the mammalian kidney tubule (Kluwe 1981).

Metals 1 in sediments and liver tissue were risk factors in English sole, and sediment levels covaried with AHs and PCBs. The presence of Metals 1 in liver tissue (up to 300 ng/g at Nisqually Reach) as a risk factor for proliferative lesions in English sole may have toxicologic significance, considering that one of the components of this group, copper, is nephrotoxic in winter flounder at water-borne doses as low as 1,000 µ g/l Cu⁺², inducing necrosis of the tubular epithelium (Baker 1969). Metals 2 levels in sediment (e.g., 150 ng/g in Elliott Bay) as a risk factor for this lesion in flathead sole may also be toxicologically relevant in light of one of the component metals, chromium, being a known nephrotoxicant in mammals (Hook and Hewitt 1986), as well as inducing degenerative changes in the kidney of rainbow trout (Van der Putte et al. 1981). However, uptake of these metals into the liver appears to be minimal, with a maximum of 19 ng/g detected in flathead sole from Valdez. This association is, therefore, probably of little consequence. Overall, considering their ability to directly or indirectly induce nephrotoxicity in the form of degenerative/necrotic changes in elements of the nephron and the possibility that the proliferative renal lesions represent sequelae to these lesions, the chemical risk factors represented by Metals 1, DDTs, chlordanes, PCBs, and possibly the AHs should be considered toxicologically meaningful.

Sclerotic lesions--Risk factors independent of age that were associated with this lesion category were identified in flathead sole, English sole, starry flounder, and white croaker. Aromatic hydrocarbons (up to 4,800 ng/g total AHs in Elliott Bay sediments) were risk factors only in flathead sole, consistent with the association shown between the prevalence of this lesion and mean levels of biliary FACs-H.

Several chlorinated hydrocarbon classes were risk factors for sclerotic lesions. Polychlorinated biphenyl exposure was associated with this lesion in two species, as reflected by risk factors in sediments (flathead sole, < 500 ng/g PCBs at Elliott Bay, covarying with AHs (Table 38)) and liver tissue (flathead sole, see above; starry flounder, up to 7,000 ng/g at Hunters Point). Bioaccumulation of DDT in liver was a risk factor in flathead sole and starry flounder. As was true for the proliferative lesions, individual non-DDT pesticides were inconsistent risk factors, primarily represented by liver levels in flathead sole and English sole. Specifically, in both species, liver levels of hexachlorobenzene (< 150 ng/g in English sole, < 18 ng/g in flathead sole), chlordanes (< 250 ng/g in English sole, < 160 ng/g in flathead sole) and dieldrin (< 34 ng/g in English sole, < 48 ng/g in flathead sole) were risk factors. For white croaker, only chlordanes in sediment (< 15 ng/g) were a risk factor. Based on these levels of potential exposure and bioaccumulation, chlordanes could be considered a toxicologically relevant risk factor in English and flathead sole; hexachlorobenzene exposure is probably only meaningful in English sole, and it is not likely that dieldrin exposure is a factor in the genesis of sclerotic renal lesions in these two species.

Metals exposure as reflected in liver burden (Metals 1) was a risk factor only in white croaker. The hepatic levels found (< 490 ng/g at San Pedro Outer Harbor) and the presence of nephrotoxic components in this group of metals (e.g., copper) suggest that this may also be a toxicologically relevant risk factor.

In summary, chemical risk factors with etiological potential and toxicological relevance for sclerotic renal lesions were the PCBs, DDTs, some non-DDT pesticides, Metals 1, and possibly the AHs.

Necrotic lesions--Similar risk factors as those shown for the other renal lesion categories existed for the necrotic lesions, primarily in flathead sole, English sole, and starry flounder. However, the consistency of these risk factors among the compartments was generally low, and very few risk factors were common to more than one species.

For example, AH exposure, as reflected by sediment levels of low and high molecular weight AHs and supported by the biliary FACs-H as a risk factor, had a significant influence on prevalence of necrotic lesions only in flathead sole.

Covarying sediment PCBs were also a risk factor in this species, but this association was not corroborated by the PCB levels in the liver as a risk factor. Liver PCBs were a risk factor only in starry flounder, but were not corroborated as a risk factor by sediment levels of PCBs at flounder sampling sites.

Exposure to DDT was a risk factor in flathead sole, starry flounder, and white croaker; however, there was no consistency among the compartments measured for any species, and actual DDT exposure as reflected by hepatic levels was a risk factor only in starry flounder and white croaker. However, the existence of DDT exposure as a risk factor for the necrotic renal lesions is particularly relevant in light of the experimental induction, including via dietary exposure (Buhler et al. 1969), of necrotic and degenerative lesions in the tubular epithelium of the nephron in multiple teleost species (King 1962, Malthur 1962).

Other toxicologically significant risk factors within the chlorinated compound classes were individual non-DDT pesticides, including liver levels of hexachlorobenzene (English sole only, < 150 ng/g), dieldrin (starry flounder only, < 300 ng/g), and chlordanes (English sole, liver (< 250 ng/g) and stomach contents (< 50 ng/g)). Although chlordanes in sediment were also a risk factor for white croaker (< 15 ng/g), and in stomach contents of starry flounder (< 17 ng/g), these maximum detected levels were too low to be toxicologically relevant.

The final risk factor of potential toxicologic importance for renal necrosis, in both flathead sole and starry flounder, was Metals 1 in sediments; however, Metals 1 in liver tissue was not a risk factor for this lesion in either species. As stated earlier, this group of metals contains copper, which is a nephrotoxicant in mammals (Hook and Hewitt 1986), as well as in winter flounder (Baker 1969). Maximum detected sediment levels were 320 ng/g and 240 ng/g at flathead sole and starry flounder sites, respectively.

Overall, the toxicologically relevant risk factors of chemical exposure for the necrotic lesions among the species affected were the AHs (flathead sole), PCBs (starry flounder), DDTs (starry flounder and white croaker), hexachlorobenzene (English sole), dieldrin (starry flounder), chlordanes (English sole), and possibly Metals 1 (flathead sole and starry flounder). The latter risk factor can only be regarded as potentially significant, since it was not a risk factor in a compartment showing actual exposure, and data on bioaccumulation of these metals in kidney tissue were not available.

Fluorescent Aromatic Compounds (FACs) in Bile as Risk Factors in Individual Fish

The relationship between biliary FAC levels and renal disease has not been previously addressed in any field studies on wild fish. In this study, biliary FACs were rarely identified as a risk factor for renal disease, and only in a few species. As described above on a by-site basis, mean biliary FACs-H levels were associated with proliferative lesion prevalence in English sole and with sclerotic lesions in flathead sole; these relationships were not, however, confirmed in logistic regression analyses on individual fish, possibly for the reason that these latter analyses only included a small proportion (~30%) of the renal histopathology data reflected in the by-site analyses, as was the case for the liver lesions. On the other hand, necrotic lesions were linked to levels of biliary FACs-L in individual starry flounder, but not as assessed on a by-site basis. The only relationship shown in both types of analyses between levels of biliary FACs-H and renal disease was for necrotic lesions in flathead sole; this relationship may be simply due to the fact that the site with the highest prevalence of these lesions was Elliott Bay, which also showed the highest levels of biliary FACs-H in this species (up to 410 ng/g).

Considering the relative lack of information on the effects of aromatic hydrocarbon exposure on teleost kidney, the toxicologic meaning of these statistical relationships are unclear. Although AHs are not widely regarded as nephrotoxicants in mammalian systems (Hook and Hewitt 1986) or fish (Hinton and Lauren 1990a), laboratory studies on English sole exposed to BaP or an organic-solvent extract of sediment from a creosote-contaminated site have shown statistically higher incidences, as compared to controls, of necrotic and proliferative lesions in the tubular segments of the nephron (Schiewe et al. 1991).

It is not surprising that temporal changes in prevalences of hepatic lesions were not detected in the species tested, considering the relatively few data points (four to five) for each site tested. Although the Spearman rank correlation method is robust to departures from the typical assumptions of statistical tests, it detects only monotonic trends in the prevalence data, and the likelihood of trend detection was limited by the few data points in the analysis. Therefore, trends in which lesion prevalence may have increased over the first three cycles and decreased since then would not be detected. Logistic regression analyses of hepatic lesion prevalences were also done to account for possible differences in age class or gender composition in a species sampled at a particular site over the sampling periods within this study. This analysis also showed no significant monotonic changes in prevalences, suggesting that hepatic lesion prevalences are not substantially increasing or decreasing over time at the sites and species tested. However, future analyses which include more data points will enable the use of other methods such as meta-analysis (Mullen and Rosenthal 1985), and they may provide more conclusive results upon which to base environmental policy decisions.

In a comprehensive histopathological and epizootiological study in multiple species of bottomfish sampled at sites on the Pacific Coast of the United States between 1984 and 1988, several primary target species were identified as having significant prevalences of toxicopathic lesions in liver and kidney. These lesions were, in many cases, statistically related to exposure to various classes of environmental contaminants and, hence, can serve as biomarkers of contaminant-induced effects in wild fish. The primary target species affected with these lesions were flathead sole, English sole, starry flounder, hornyhead turbot, white croaker, and black croaker. Other species were affected by these lesions at relatively low and homogeneous site-specific prevalences, or the species were sampled less intensely because of a restricted distribution among the sites, and thus were not considered useful target species for assessment of histopathologic biomarkers of contaminant-induced effects. These species included fourhorn sculpin, Arctic flounder, yellowfin sole, Pacific staghorn sculpin, barred sandbass, spotted sandbass, spotted turbot, diamond turbot, California tonguefish, and California halibut.

Of the species chosen as reasonable primary target species for the investigation of the epizootiological relationship between contaminant exposure and histopathologic effects, toxicopathic hepatic lesions were detected primarily in fish from urban sites, especially in English sole, starry flounder, white croaker, and black croaker. Many of these urban sites showed significantly elevated relative risks for at least one hepatic lesion type in these species, while controlling for the variables of fish age and gender. Of greatest importance was the demonstration in English sole, starry flounder, and white croaker of positive statistical associations between exposure to contaminant classes and increased risk of these hepatic lesions. Prevalences of hepatic lesions grouped within the categories of neoplasms, foci of cellular alteration, nonneoplastic proliferative lesions, specific degeneration/necrosis, and hydropic vacuolation were most commonly associated with exposure to and uptake and metabolism of AHs, PCBs, and DDTs at exposure or bioaccumulation levels that were toxicologically significant. However, certain chemical classes were associated with these lesions in a species-dependent pattern, such that a particular lesion type affecting multiple species was not always associated with exposure to the same chemical class(es). Chemical risk factors of hepatic disease were less commonly identified for nonspecific necrotic hepatic lesions. The individual non-DDT pesticides (chlordanes and dieldrin) were toxicologically relevant risk factors only for specific degeneration/necrosis and hydropic vacuolation in a single species (starry flounder).

However, an important caveat to the utility of these hepatic lesions as biomarkers of such exposure is that fish age must be accounted for when performing these statistical analyses, since age is a risk factor in several species that significantly influences the probability of occurrence of hepatic neoplasms, foci of cellular alteration, nonneoplastic proliferative lesions, as well as hydropic vacuolation. Fish gender was generally not an important risk factor for any toxicopathic hepatic lesion.

Not all toxicopathic hepatic lesion types were detected or consistently associated with contaminant exposure in the multiple species examined, suggesting that not all teleosts respond similarly to exposure to the same classes of toxicants (Collier et al. 1992), and that not all of the hepatic lesions identified in this and other studies can be used as histopathologic biomarkers of contaminant exposure in all species. Specifically, in flathead sole and hornyhead turbot, no consistently useful hepatic biomarkers of contaminant-induced effects were identified. In English sole, hepatic neoplasms, foci of cellular alteration, nonneoplastic proliferative lesions, and specific degeneration/necrosis were associated with exposure to AHs and hepatic bioaccumulation of PCBs and DDTs; nonspecific necrotic lesions were associated only with DDT bioaccumulation in liver. Hydropic vacuolation was not detected in English sole. In starry flounder, reliable biomarkers of contaminant exposure included the most prevalent lesion, hydropic vacuolation, which was associated with exposure to AHs, and hepatic bioaccumulation of PCBs, DDTs, chlordanes and dieldrin, as well as the less prevalent lesions of foci of cellular alteration (associated with dietary uptake of PCBs and DDTs) and specific degeneration/necrosis (associated with potential exposure to AHs and PCBs, and hepatic bioaccumulation of PCBs, chlordanes and dieldrin). Hepatic neoplasms were not detected in starry flounder. All lesion types in white croaker with the exception of neoplasms were associated with exposure to one or more classes of contaminants. Specifically, these were foci of cellular alteration (potential and actual AH exposure), nonneoplastic proliferative lesions (potential and actual AH exposure; potential PCB exposure and actual uptake in the diet; potential DDT exposure, actual dietary uptake, and hepatic bioaccumulation), specific degenerative/ necrotic lesions (actual AH exposure; potential PCB exposure, actual uptake in the diet and hepatic bioaccumulation; potential DDT exposure, actual dietary uptake and hepatic bioaccumulation), nonspecific necrotic lesions (potential and actual AH exposure; potential PCB exposure; and hepatic bioaccumulation of DDTs), and hydropic vacuolation (actual AH exposure; hepatic bioaccumulation of PCBs). In black croaker, not enough chemistry or histopathology data existed to perform a logistic regression analysis in this study; however, the geographic distribution of prevalences of hepatic neoplasms and foci of cellular alteration strongly suggest an association with exposure to AHs, and hepatic bioaccumulation of PCBs and DDTs (McCain et al. 1992). Hydropic vacuolation was not detected in this species.

For the first time in a histopathologic biomonitoring study, associations of potential toxicologic significance between contaminant exposure-related risk factors and prevalences of suspected toxicopathic renal lesions were identified, although far less frequently than for liver lesions; species demonstrating these risk factors were flathead sole, English sole, starry flounder, and white croaker. Proliferative, sclerotic, and necrotic lesion category prevalences were irregularly associated among these species with toxicologically significant levels of potential exposure to, dietary uptake of, or hepatic bioaccumulation of AHs, PCBs, DDTs, chlordanes, dieldrin, summed levels of copper, zinc, lead, and tin (Metals 1), and occasionally with summed levels of nickel, chromium, and selenium (Metals 2). However, as with the liver lesions, these relationships were not highly consistent, and the same risk factors were rarely identified for the same lesion type affecting all four species.

As was true for the hepatic lesions, not all renal lesion categories were identified as meaningfully associated with contaminant exposure in all species. Specifically, in flathead sole, proliferative lesions were associated with hepatic bioaccumulation of PCBs, DDTs, and chlordanes; sclerotic lesions were associated with potential and actual AH exposure, and hepatic bioaccumulation of PCBs, DDTs, and chlordanes; necrotic lesions were associated with potential and actual AH exposure, and potential exposure to DDTs and PCBs. In English sole, the proliferative lesions showed meaningful associations with potential and actual AH exposure, potential exposure to and hepatic bioaccumulation of PCBs, and potential exposure to and hepatic bioaccumulation of Metals 1; sclerotic and necrotic lesions were associated only with hepatic bioaccumulation of chlordanes and hexachlorobenzene. In starry flounder, proliferative lesions were not associated with any chemical exposure-related risk factor; sclerotic lesions were linked to hepatic bioaccumulation of DDTs and PCBs; and prevalences of necrotic lesions showed associations with actual exposure to AHs, hepatic bioaccumulation of PCBs, DDTs, and dieldrin, and potential exposure to Metals 1. There were no renal histopathologic biomarkers of contaminant exposure in hornyhead turbot. In white croaker, proliferative lesions were associated with potential exposure to and hepatic bioaccumulation of DDTs; sclerotic lesions were meaningfully linked only with hepatic bioaccumulation of Metals 1; and necrotic lesions were associated only with hepatic bioaccumulation of DDTs.

Because the AHs, PCBs, DDTs and Metals 1 tended to significantly covary in sediments especially, and less commonly in stomach contents or liver/bile (AHs and their metabolites), it is clear that the bottomfish in this study are exposed simultaneously to a complex mixture of contaminants. The assignment of a relative importance for any of these significant chemical exposure-related risk factors with respect to the etiology of pollution-associated hepatic or renal disease is, therefore, not possible on the basis of epizootiological data alone. Laboratory exposure studies are essential in order to conclusively establish the roles of specific toxicants in the etiology of pollution-associated disease in wild fish. However, the existence of statistically significant and consistent relationships between these risk factors and hepatic and renal lesions provides strong epizootiological evidence supporting the role of chemical contaminants in the etiology of these lesions, and clearly indicates their utility as biomarkers of contaminant exposure effects in native fish species sampled in biomonitoring studies.

We thank Drs. Tracy Collier and Ed Casillas for reviewing the manuscript, and Sharon Giese and Barbara Bennett for editorial assistance. These studies were supported by the Office of Ocean Resources Conservation and Assessment (NOAA National Ocean Services) as part of the National Status and Trends Program, and by the NOAA Coastal Ocean Program. Appreciation is also given to the personnel of the NOAA ships McArthur and Miller Freeman, research vessels Harold W. Streeter and Sea Otter, and the U.S. Fish and Wildlife Service vessel Curlew for sample collection.