Back to Section 2

SECTION 3

FLATFISH REPRODUCTIVE INJURY STUDY

Introduction

Reproductive impairment is potentially one of the most damaging effects of aquatic pollution on marine fish, because of its impact on population growth and consequently on fisheries productivity (Donaldson 1990, Kime 1995). Environmental contaminants may have direct toxic effects on germ cell tissue, or may disrupt the endocrine mechanisms that regulate reproduction and early development, causing inhibited or abnormal gonadal development or reduced fertility (Guillette 1994, Muller et al. 1995, Gray et al. 1996). Because of recent research on environmental estrogens and other hormone-mimicking compounds, the ecological and health impacts of endocrine disrupting chemicals has become a subject of increasing concern (Colburn et al. 1993) to both the public and to resource management agencies, including the National Oceanographic and Atmospheric Administration (NOAA) and NMFS.

Sediments from several areas of Puget Sound, including Commencement Bay and the Hylebos Waterway, are polluted with xenobiotic compounds such as chlorinated pesticides, PCBs, and PAHs (Malins et al. 1980, 1984; Appendix D), which are known or suspected disrupters of endocrine function and pose a potential threat to the reproductive health of marine fish which reside in these areas. As part of an effort to assess the biological impacts of AHs and CHs on flatfish, over the past 10 years the ECD of the NWFSC has developed a comprehensive research program to evaluate the effects of these contaminants on fish reproduction. Initial studies focused primarily on the reproductive status of gonadally maturing adult female fish, because of their sensitivity to contaminant effects and the importance of their contribution to the population. Field studies showed that a significant proportion of female flatfish from contaminated areas in Puget Sound, such as the Duwamish Waterway in Seattle and Eagle Harbor near Bainbridge Island, exhibit reduced reproductive success compared to females from minimally contaminated sites (Johnson et al. 1988, Casillas et al. 1991). Among the reproductive effects observed were inhibition of oocyte development, inhibition of spawning, depressed plasma estradiol levels, reduced egg weight, increased larval abnormalities, and reduced viability of offspring. Another study suggested that some female English sole from contaminated areas do not make their spawning migrations (Collier et a. 1992). These field data have been incorporated into population projection models, and the results suggest that the decreased reproductive potential associated with exposure of certain flatfish species (e.g., English sole) to contaminants may be sufficient to reduce the growth rate of sole subpopulations from contaminated areas such as the Duwamish Waterway and Eagle Harbor if it is not offset by mitigating factors such as immigration of fish from minimally contaminated sites (Landahl and Johnson 1993, Landahl et al. 1997). Contaminated sites may be population "sinks" that reduce the potential productivity of the Puget Sound sole population as well as its resiliency in the face of other environmental stressors such as climate change or overfishing. Our data suggest that exposure to PAHs and to a lesser extent, exposure to PCBs, are substantial risk factors for reduced reproductive success. Laboratory investigations have also shown effects on circulating levels of sex hormones due to contaminant exposure (Stein et al. 1991b, Johnson et al. 1995), which are consistent with field results.

These previous studies were used as the basis for the Reproductive Injury portion of the Hylebos Waterway Fish Injury Study. In this assessment the focus was on detecting contaminant-associated changes in gonadal development and reproductive steroid concentrations in female English sole.

Methods

Sample Collection and Analysis

Adult female English sole were collected by otter trawl from the Hylebos Waterway and a reference site in Colvos Passage (Figure 3.1), following protocols outlined in the Sampling and Analysis Plan (see Appendix A). Sampling was conducted approximately every 4-6 weeks from September 1994 through April 1995, during the season when vitellogenesis (i.e., egg development) normally occurs in this species (Johnson et al. 1991). Approximately 30 fish were collected at each site for every sampling period. Otoliths were collected for age determination, blood samples were collected for plasma 17-ß estradiol analyses, and liver, kidney, and ovary were collected for histopathological examination and staging of ovarian development. Liver, bile and stomach contents were collected for assessment of contaminant exposure, and liver was also collected for measurement of DNA adducts. Samples were not collected for analysis of hepatic CYP1A activity, because this measure is known to be substantially down-regulated during the reproductive process in female English sole (Collier et al. 1992).

Chemical and biological analyses (i.e., fish age determination, chemical analyses of tissue and stomach contents for chlorinated and aromatic hydrocarbons, measurement of FAC concentrations in bile, DNA adducts in liver, and estradiol concentrations in plasma, determination of somatic indices, and histological analyses) were carried out as described in Appendix A. Specific individual analytes measured for each chemical class are listed in Appendix B.

The three somatic indices monitored were condition factor, hepatosomatic index (HSI), and gonadosomatic index (GSI). Condition factor is a measure of the fish's body weight relative to its size. HSI is measure of liver weight relative to total body weight, and can be used as an indicator of pollutant-related increases in liver size. GSI is a measure of gonad weight relative to total body weight, and is used as an indicator of gonadal development.

The types of liver lesions monitored in this study are described in detail in Section 2 and Appendix A. Ovarian lesions monitored included atresia or resorption of yolked and non-yolked oocytes (eggs), late stage atretic follicles where the stage of the oocyte undergoing resorption could not be determined, and inflammatory lesions. Because a certain amount of mild atresia is normal in female English sole (Johnson et al. 1991), only moderate to severe atresia was considered as constituting a lesion in this study. Ovarian lesions and lesion severity rankings are described in detail in Appendices A and C.

In this study, biliary FAC concentrations were adjusted for the protein content of the bile sample. This correction adjusts for variations in bile metabolite levels associated with feeding status of the sampled fish (Collier and Varanasi 1991). Procedures for measuring biliary protein and calculating protein-corrected biliary FAC concentrations are described in Appendix A. Also, concentrations of toxic PCBs and other selected CHs in liver tissue (e.g., HCB and DDTs) were measured in samples from individual fish using the HPLC/PDA method (Krahn et al. 1994). In other portions of the study, liver CHs were measured in composite samples using gas chromatography with electron capture detection (GC/ECD) (Sloan et al. 1993). Concentrations of HCBD were not measured in liver tissue of flatfish collected for the Flatfish Reproductive Injury Study, because it could not be readily quantitated using the HPLC/PDA technique. However, we chose to use the HPLC/PDA method because it allowed us to measure CH concentrations in small tissue samples from individual fish, increasing our ability to identify statistical associations between contaminant body burdens and specific types of reproductive dysfunction.

Statistical Methods

The data from these injury assessment investigations were placed in an integrated data management system. Statistical tests were then performed 1) to evaluate relationships between contaminant exposure and biological effects, and 2) to compare values from fish captured in the Hylebos Waterway with values for fish captured in reference areas. ANOVA and analysis of covariance (ANCOVA) were used to identify statistically significant intersite differences in chemical concentrations and biological parameters. Linear regression analysis and Spearman-Rank non-parametric correlation analysis were used to evaluate relationships between biochemical responses to contaminant exposure (e.g., tissue contaminant concentrations, biliary FACs, DNA adducts) and indicators of biological effects (e.g., plasma estradiol concentrations and somatic indices). For parametric statistical tests such as ANOVA and linear regression, data were normalized as necessary through log-transformation prior to analysis. All biological and chemical data required log-transformation, with the exception of fish length and condition factor, which were already normally distributed, and hepatic and ovarian pathology data, to which these types of analyses were not applied. For the ANCOVA examining the effect of site of capture on length-weight relationships in English sole, both fish gutted weight and length were log-transformed, although both variables conformed reasonably well to a normal distribution, to improve the linear fit of the model and remove skewness in the residuals. Prevalences of ovarian and liver lesions at the two sampling sites were compared using the G-statistic or chi-square analysis. Logistic regression analysis (Breslow and Day 1980) was used to evaluate relationships between binomial or proportional outcome variables (e.g., absence or presence of vitellogenic eggs in the ovary, and lesion occurrence). Significance levels (i.e., p values) for individual statistical analyses are reported in the Results section. In those cases where chemical or biochemical measurements were below detection limits, a value of 50% of the detection limit was used so the data could be incorporated into statistical calculations. (See Appendix A for details on treatment of detection limits for specific analyses. Individual samples with values below detection limits are identified in the Case Narratives, Appendix C.)

Results

Chemical Contaminant Exposure

Concentrations of contaminants or their metabolites in stomach contents, tissues, and bile of English sole collected from Hylebos Waterway and Colvos Passage during the 1994 reproductive season are shown in Figures 3.2-3.7.

Aromatic and chlorinated hydrocarbons in stomach contents

English sole from the Hylebos Waterway had significantly higher concentrations of both LACs and HACs, PCBs, DDTs, HCBD, and HCB in stomach contents than fish from Colvos Passage (Figure 3.2). Concentrations of all these contaminants were from 1 to 3 orders of magnitude greater in stomach contents of Hylebos fish than in stomach contents of reference fish. Concentrations of several pesticides (i.e., aldrin, lindane, dieldrin, and chlordanes) were also measured in English sole stomach contents, but their concentrations were low at both sites. Mean concentrations of aldrin, lindane, and dieldrin were < 1 ng/g wet weight at both sites, and concentrations were below detection limits in a high proportion of the samples tested (see Case Narratives, Appendix C). Although concentrations of chlordanes were significantly higher (p < 0.0001) in stomach contents of sole from the Hylebos Waterway than in sole from Colvos Passage, they were still relatively low (1.5 ± 0.3 ng/g wet wt vs. 0.20 ± 0.06 ng/g).

Chlorinated hydrocarbons in liver and metabolites of aromatic hydrocarbons in bile

English sole from Hylebos Waterway also showed significantly higher concentrations of contaminants in bile and liver tissue, compared to fish from Colvos Passage. Liver concentrations of HCB, a chemical which is characteristic of contaminant profiles of sediment from the Hylebos Waterway (Malins et al. 1980, Appendix D), were over 40 times higher in Hylebos Waterway English sole than in Colvos Passage English sole (Figure 3.3). As noted in the Methods section, concentrations of HCBD, another chemical characteristic of Hylebos Waterway sediments (Malins et al. 1980, Appendix D), were not measured in liver tissue of flatfish collected for this portion of the Fish Injury Study. However, in composite samples of liver collected from English sole as part of the Flatfish Toxicopathic Injury Study (Section 2), HCBD concentrations were significantly higher in Hylebos Waterway fish than in Colvos Passage fish. Concentrations of biliary FACs, an indicator of exposure to ACs, were 2 to 4 times higher in sole from the Hylebos Waterway as in sole from Colvos Passage (Figure 3.4).

Hylebos Waterway English sole also had significantly higher concentrations of DDTs (Figure 3.5) as well as a number of toxic PCB congeners in liver tissue than Colvos Passage English sole (Table 3.1; Figures 3.6-3.7). The mean PCB congeners measured by the HPLC/PDA analytical technique was over 12 times as high in liver of Hylebos English sole as in liver of Colvos Passage English sole; in one-way ANOVA, the values were statistically different at p < 0.0001 (Figure 3.6). The mean TCDD TEQ ⁶ concentration of PCBs in English sole livers from Hylebos (37 pg/g wet weight) was also much higher (p < 0.0001) than the mean TCDD TEQ concentration in English sole livers from Colvos Passage (3.7 pg/g wet weight) (Figure 3.7). Moderately chlorinated biphenyls (i.e., PCBs 118, 138, and 153) were the most abundant PCB congeners measured in the English sole livers from both Hylebos and Colvos Passage (Figure 3.6). Several dioxin-like PCBs were also measured in the English sole livers, with the di-ortho (i.e., PCBs 170 and 180) and mono-ortho substituted PCBs (i.e., PCBs 105 and 118) being the most prevalent (Figure 3.6). Of the three non-ortho substituted congeners (PCBs 77, 126 and 169) measured, only one liver sample from the Hylebos site contained a non-ortho substituted PCB (PCB 77); the concentration was 0.6 pg/g wet weight.

Biological Indicators of Contaminant Exposure

DNA adducts in liver

Levels of DNA adducts in liver, which are an indicator of contaminant-related DNA damage, were over 7 times as high in sole from the Hylebos Waterway as in sole from Colvos Passage (Figure 3.8). The difference in mean adduct concentrations was statistically significant at p < 0.0001.

Toxicopathic Liver Lesions

Hylebos Waterway fish had significantly higher prevalences of several types of liver lesions than fish from Colvos Passage (Figure 3.9). These lesions included neoplasms, or liver tumors; FCA, which are preneoplastic lesions; and SDN, which is a type of degenerative lesion whose development is linked to chemical contaminants, particularly ACs. At Colvos Passage, less than 2% of fish sampled had neoplasms or SDN, and about 3% had preneoplastic lesions. In comparison, approximately 6% of fish from the Hylebos Waterway had neoplasms, 13% had preneoplasms, and 10% had SDN. Additional information on liver pathology and specific chemical risk factors associated with the development of liver lesions in English and rock sole is provided in the Flatfish Toxicopathic Injury Study portion of this report (Section 2).

Condition Factor, Growth, and Hepatosomatic Index

Hylebos Waterway sole were significantly larger (i.e., of greater length) for their age between years 1 and 5 than fish from Colvos Passage (Figure 3.10a). After age 5, fish from the two sites did not differ significantly in size at age (Figure 3.10b). Comparison of the length-weight relationships in sole from the two sites (Figure 3.11) showed that the Hylebos fish of younger age classes also tended to weigh more for their size than fish from Colvos Passage (i.e., the intercept of the length-weight regression line was significantly higher for Hylebos fish than for Colvos fish), but their weight did not increase as rapidly with length as the weight of Colvos Passage fish (i.e., the slope of the length-weight regression line was significantly lower for Hylebos fish than for Colvos fish; see ANCOVA table in Figure 3.11). For age 5 or below, mean condition factor was significantly higher in Hylebos Waterway fish than in Colvos Passage fish; however, for fish age 6 or above, condition factor was the same at the two sites (Figure 3.12). Hylebos Waterway English sole also had significantly higher HSI than fish from Colvos Passage (Figure 3.13).

Reproductive Dysfunction

Ovarian Development

The timing of the seasonal reproductive cycle was similar in female English sole from Colvos Passage and Hylebos Waterway (Figure 3.14), but the proportion of adult females that produced yolked eggs at the peak of the reproductive season tended to be lower in the Hylebos Waterway than at Colvos Passage. In September, when the study was initiated, ovaries of approximately 30 to 40% of adult females (> 300 mm in length) from both sites contained yolked eggs. At the Colvos Passage reference site, by October, this percentage had increased to 65%, rising to 80% in December and remaining relatively stable through January. The proportion of maturing females in Hylebos Waterway was approximately 80% in October, but declined to 65% in December, and to only 55% in January. In February, fish apparently left Colvos Passage to migrate to spawning areas, as no adult fish could be found at this sampling time. However, vitellogenic females (i.e., females with yolked eggs in the ovary) were still present in significant numbers at the Hylebos site in February; approximately 30% of the animals collected were vitellogenic. By April, fish had completed spawning, and no females with developing eggs were observed at either Colvos Passage or the Hylebos Waterway. Overall, approximately 80% (n = 44) of adult females (> 300 mm in length) from Colvos Passage collected between October and February were undergoing ovarian development, whereas only 58% (n = 52) of females of comparable size were developing in the Hylebos Waterway. This would suggest that the number of adult females developing was decreased by about 22% in the Hylebos Waterway.When the percentage of females entering vitellogenesis at each age class was examined at the two sites, it was found that Hylebos fish began to mature at a younger age than Colvos fish (Figure 3.15). In the Hylebos Waterway 40 to 50% of fish below 3 years of age were maturing, whereas no fish below 3 years of age were developing at the Colvos Passage site. Approximately 20% of three year old Colvos females were maturing, as compared to 50% of Hylebos females. However, for older age classes (i.e., 5 to 10 years of age) over 70% or more of Colvos females were maturing, while the proportion of maturing fish from the Hylebos rarely exceeded 50%.

Logistic regression analysis (Figure 3.16A) indicated that Colvos Passage English sole were more likely to undergo reproductive development as they grew older, at least up to the age of about 8 or 9, when the probability of maturation approached 1.0. In Hylebos fish, however, no association between age and probability of reproductive development was found; the probability of gonadal development was approximately 0.5, regardless of age class (Figure 3.16B).

Ovarian lesions

Although prevalences of moderate to severe ovarian lesions were relatively low (< 20%) in the English sole sampled as part of this study, lesion prevalences generally tended to be higher in sole from the Hylebos Waterway than in sole from Colvos Passage (Figure 3.17). Significant differences were observed for late stage atresia, which was found in 5.7% of Hylebos Waterway fish sampled during the reproductive season (September through February), but was not detected in Colvos Passage fish, and inflammatory lesions, which were found in 5.7% of Hylebos fish but only 0.9% of Colvos fish. An exception to the general pattern was early stage atresia of yolked oocytes; this lesion was found in 3.5% of Colvos Passage fish, but only 0.8% of Hylebos Waterway fish. Overall, about 16% of Hylebos fish had one or more moderate to severe ovarian lesions (i.e., one or more ovarian lesions in Figure 3.17) in comparison with 9% of Colvos Passage fish, representing an increase of about 7% of the population.

Gonadosomatic Index

Mean GSI (Figure 3.18), which is a measure of gonad size, expressed as per cent body weight, was significantly lower in adult sole (> 5 years of age) sampled from the Hylebos Waterway during the peak period of gonadal development (October-February) than in sole from Colvos Passage. Mean GSI for fish for Colvos Passage was approximately 4.1%, while for Hylebos fish it was approximately 3.5%.

-estradiol Concentrations.

Mean plasma -estradiol concentration (Figure 3.19A) was significantly lower in adult sole (> 5 years of age) sampled from the Hylebos Waterway during the peak period of gonadal development (October-February) than in sole from Colvos Passage sampled during the same time period. Mean estradiol values were about 2700 pg/ml in Hylebos fish, in comparison to around 3800 pg/ml in Colvos Passage fish. Plasma estradiol concentrations normalized for gonad weight are shown in Figure 3.19B. This analysis showed that plasma estradiol per gram of ovary tissue was significantly lower in Hylebos fish than in Colvos Passage fish, suggesting either reduced production or increased clearance of the hormone in the Hylebos fish.

Chemical Risk Factors

Gonadal Development

For English sole under five years of age, the probability of precocious sexual maturation was significantly associated with high tissue concentrations of several types of chemical contaminants which co-occur in the Hylebos Waterway. Several classes of CHs were significant risk factors for precocious maturation (Table 3.2). The strongest correlation was observed with liver HCB (p = 0.0004) (Figure 3.20). The probability of precocious maturation also increased with increasing concentrations of DDTs (p = 0.01), PCBs (p = 0.002), and TCDD TEQs (p = 0.002) in liver. Increasing levels of DNA adducts in liver, an indicator of chronic exposure to genotoxic PAHs, were associated with an increased probability of precocious maturation (p = 0.004). Significant associations (0.002 < p < 0.003) were also found between biliary FAC concentrations and the probability of precocious maturation.

In contrast, for English sole 5 years of age and older, the probability of gonadal development declined as chemical concentrations in tissues and bile increased (Table 3.2). Of the exposure indicators measured, concentrations of FACs in bile were the most strongly associated with a decreased likelihood of ovarian development (p = 0.08, 0.05, and 0.008 for FACs _BaP, FACs_PHN, and FACs_NPH, respectively). The fitted relationships of FACs _BaP and FACs_NPH with the probability of ovarian development are shown graphically in Figure 3.21. The probability of ovarian development also tended to decline as concentrations of PCBs (p = 0.11), DDTs (p = 0.16), TCDD TEQs (p = 0.14), and DNA adduct levels (p = 0.29) in liver increased but the relationships were not as strong as those for FACs. Liver HCB concentrations showed no relationship with the likelihood of ovarian development for sole 5 years of age and older (p = 0.72).

Ovarian Lesions

Elevated biliary FAC levels were a significant risk factor (0.004 < p < 0.01) for late stage ovarian atresia in English sole (Table 3.2). However, only relatively weak associations were observed between indicators of contaminant exposure and other types of ovarian lesions.

GSI and Plasma Estradiol Concentrations

In fish under age 5, most indicators of contaminant exposure were significantly positively correlated with GSI and plasma estradiol concentrations (Table 3.3). In stepwise multiple regression utilizing all of these measurements, the best predictor of GSI and plasma estradiol concentrations was liver HCB concentration, accounting for about 40-50% of the variance in both reproductive indicators (Table 3.4).

For fish of age 5 and older, the indicators of contaminant exposure tended to be negatively correlated with GSI and plasma estradiol concentrations (Table 3.3). For both plasma estradiol concentrations and GSI, the strongest relationships were found with biliary FACs (0.015 < p < 0.06). Significant or near significant negative correlations were also observed between plasma estradiol concentrations and PCBs (p = 0.03), DDT (p = 0.06), and TEQ (p = 0.07) concentrations in liver, and between PCBs and GSI (p = 0.09). Liver HCB concentrations and DNA adducts showed no significant relationship with either plasma estradiol concentrations or GSI in older fish.

Discussion of Major Findings

Gravid female English sole living in the Hylebos Waterway show evidence of exposure to several classes of toxic contaminants, including aromatic hydrocarbons, PCBs, pesticides, and hexachlorobenzene.

Reproductively maturing female English sole in the Hylebos Waterway showed clear signs of exposure to low and high molecular weight aromatic hydrocarbons, PCBs, DDTs, HCB, and HCBD. Concentrations of all these contaminants were from 1 to 3 orders of magnitude greater in stomach contents of Hylebos fish than in stomach contents of reference fish. Moreover, the same contaminants were also present in bile or tissues of Hylebos Waterway English sole at concentrations much greater than those found in reference fish. For example, liver concentrations of HCB, a chemical that is characteristic of sediment contaminant profiles from the Hylebos Waterway (Malins et al. 1980, Appendix D), were over 40 times as high in Hylebos Waterway English sole as in Colvos Passage English sole. Biliary FAC concentrations, an indicator of exposure to aromatic compounds, were several times higher in sole from the Hylebos Waterway than in sole from Colvos Passage. Hylebos Waterway English sole also had significantly higher concentrations of PCBs in liver tissue than Colvos Passage English sole. Concentrations of HCBD could not be measured in liver of sole collected for the Flatfish Reproductive Injury Study, because the HPLC/PDA technique, which was used to allow detection of PCB congeners in liver samples from individual fish, does not detect this compound. However, both liver and stomach contents concentrations of HCBD were significantly elevated in composite samples from Hylebos English sole collected for the Flatfish Toxicopathic Injury Study (Section 2). The HCBD concentration in stomach contents of English sole sampled for the Flatfish Reproductive Injury Study was comparable to that reported for sole collected for the Flatfish Toxicopathic Injury Study, so it is likely that elevated concentrations of HCBD are also present the liver tissue of reproductively active female English sole.

Mean biliary FAC concentrations in English sole from the Hylebos Waterway were generally comparable to biliary FAC concentrations observed in fish from other contaminated sites within Puget Sound and from other urban embayments around the United States. At approximately 770 ng/mg bile protein, biliary FAC_BaP concentrations in English sole from the Hylebos Waterway were very similar to typical biliary FAC_BaP concentrations in English sole from the Duwamish Waterway (e.g., approximately 590 ng/mg bile protein (Stein et al. 1992)), and slightly higher than concentrations measured in sole collected in outer Commencement Bay and Elliott Bay (300 to 400 ng/mg bile protein) (Hom, unpubl.⁴). Typical biliary FAC_BaP concentrations in benthic fish species from other urban sites along the west Coast are 280 ng/mg bile protein in West Santa Monica Bay near Los Angeles, 220 ng/mg bile protein in Southampton Shoal in San Francisco Bay, and 330 ng/mg bile protein in Oakland Estuary (Hom, unpubl⁴).

The mean concentration of PCB congeners⁷ (as measured by HPLC/PDA) in livers of Hylebos Waterway sole was about 1300 ng/g wet weight, while total PCB concentrations in composite liver samples (determined by GC/ECD) ranged from about 2,000 to 4,000 ng/g wet weight (see Section 2, Flatfish Toxicopathic Injury Study), or approximately 10,000 to 20,000 ng/g dry weight, using a wet wt/dry wt ratio of 5. The difference between the PCB concentrations reported for the two methods is due to the larger number of congeners detected with the GC/ECD method. The PCB concentrations reported for the Hylebos Waterway in this study are somewhat higher than previously reported total PCB concentrations of 4,000 ng/g dry weight in livers of English sole from a site outside the Waterway in Commencement Bay (Varanasi et al. 1989), and comparable to levels reported for English sole from Elliott Bay (approximately 10,000 ng/g dry weight (Varanasi et al. 1989)). Typical mean total PCB concentrations in livers of benthic fish from other urban sites along the west Coast are: 4,000 to 5,000 ng/g dry weight in San Francisco Bay, 7,000 ng/g dry weight in San Diego Bay, and 10,000 to 20,000 ng/g dry weight at sites such as Long Beach and West Santa Monica Bay in the Los Angeles area (Varanasi et al. 1989).

Moderately chlorinated PCBs (i.e., PCBs 153, 138, 118), which are the major components of certain technical PCB mixtures (McFarland and Clarke 1989), were the most abundant PCB congeners measured in the English sole livers from the Hylebos Waterway. These PCB congeners have high bioaccumulation factors and have been measured in high concentrations in a wide range of marine organisms (de Boer et al. 1993, Loizeau and Abarnou 1994, Corsolini et al. 1995). Several dioxin-like PCBs, including di-, mono- and non-ortho substituted congeners, were also measured in the English sole livers. Of these three classes of PCBs, the non-ortho substituted congeners are generally considered to be the most toxic, and the di-ortho substituted congeners the least toxic (Safe 1994). The di-ortho (i.e., PCBs 170 and 180) and mono-ortho substituted PCBs (i.e., PCBs 105 and 118) were the most prevalent; of the three non-ortho substituted congeners (PCBs 77, 126 and 169) measured by our HPLC/PDA method, only one liver sample from the Hylebos site contained a non-ortho substituted PCB (PCB 77), at a concentration of 0.6 ng/g wet weight. These results are consistent with those of an earlier survey of tissue PCB concentrations in hepatopancreas of Dungeness crab and muscle tissue of English sole collected from Commencement Bay (Ylitalo et al. 1995). As in the present study, PCB 77 was the only non-ortho substituted PCB congener detected in these tissues.

The mean TCDD TEQ concentration of PCBs in English sole livers from Hylebos was determined to be 37 pg/g, wet weight (n = 56). A similar mean TCDD TEQ concentration (mean = 30 pg/g wet weight; n = 10) was calculated in crab hepatopancreas samples collected from Commencement Bay (Ylitalo et al. 1995). The mean TCDD TEQ concentration in English sole livers from Colvos Passage was calculated to be 3.7 pg/g, wet weight (n = 48). This level is similar to the mean TCDD TEQ concentration determined in Dungeness crab hepatopancreas from Useless Bay, another nonurban Puget Sound site was 7.4 pg/g wet weight (n = 10) (Ylitalo et al. 1995).

Contaminant exposure was associated with several types of injury in female English sole from the Hylebos Waterway, the most notable of which were precocious sexual maturation of juvenile sole, and inhibited gonadal development in adult sole.

Chemical concentrations in stomach contents, bile, and liver of Hylebos Waterway English sole were comparable to contaminant levels associated with liver disease and impaired reproductive success in Puget Sound English sole (Johnson et al. 1988, 1993; Casillas et al. 1991; Stein et al. 1992; Myers et al. 1994, 1997) and in other fish species from contaminated sites in the United States, Canada, and Europe (Kime 1995). Moreover, these levels of contaminant exposure were linked to several types of biological injury in sole from the Hylebos Waterway. Like the animals examined in the Flatfish Toxicopathic Injury Study (see Section 2), English sole sampled as part of the Flatfish Reproductive Injury Study showed evidence of both DNA damage and liver disease. Levels of DNA adducts in liver, which are an indicator of contaminant-related DNA damage, were over 7 times higher in sole from the Hylebos Waterway than in sole from Colvos Passage. Mean concentrations of DNA adducts (~40 nmol/mol DNA bases) were similar to adduct levels found at other contaminated sites in Puget Sound, such as the Duwamish Waterway and Eagle Harbor (Myers et al. 1994, 1997). Hylebos Waterway fish also had significantly higher prevalences of several types of liver lesions than fish from Colvos Passage, including neoplasms, or liver tumors; FCA, which are preneoplastic lesions; and SDN, which is a type of degenerative lesion whose development is linked to chemical contaminants, particularly ACs. Additionally, HSI values in Hylebos sole were generally 1.5 to 2 times as high as in Colvos Passage sole. Similar increases in relative liver weight have been observed in sole from other contaminated sites in Puget Sound, such as the Duwamish Waterway and Eagle Harbor (Johnson et al. 1997). Increased HSI is commonly found in fish and other animals exposed to toxicants, especially those that induce cell proliferation, and is a general indicator of contaminant-related changes in liver structure and function (Heath 1987).

Hylebos fish also showed dramatic changes in their patterns of reproductive development. Particularly striking was the tendency of the Hylebos fish to show precocious maturation. No fish below 3 years of age showed signs of sexual development at the Colvos Passage site, whereas in the Hylebos Waterway 40 to 50% of these fish were maturing. Approximately 20% of three year old Colvos females were maturing, as compared to 50% of Hylebos females. In contrast, in older Hylebos fish (i.e., those above 5 years of age), gonadal development appeared to be inhibited. At Colvos Passage, 70 to 80% of females examined were typically maturing at these age classes, while the proportion of maturing fish from the Hylebos rarely exceeded 50%.

Both aromatic and chlorinated hydrocarbons emerged as potential risk factors for precocious maturation. Of the compounds measured, the strongest risk factor was HCB, a chemical that is found at particularly high concentrations in the Hylebos Waterway (Malins et al. 1980, Appendix D). However, other chlorinated hydrocarbons, such as DDTs and PCBs, were also strong risk factors for precocious maturation and could play a role in the development of this abnormality. Indicators of exposure to ACs (biliary FACs and DNA adducts) also showed strong associations with precocious maturation. Because all of these classes of chemicals co-occur in Hylebos Waterway sediments, and fish are exposed to them simultaneously, it is difficult to establish clear cause-and-effect relationships between specific individual chemicals and the reproductive effects observed. Although the mechanisms through which CHs such as HCB or PCBs might induce precocious maturation in Hylebos Waterway English sole are not known, some of these chemicals, including DDTs and certain PCB congeners such as PCB 153, which are common in Hylebos Waterway sediment, (Appendix D) have estrogenic or other hormone-like activity (Safe 1994, Bitman and Cecil 1970, McKinney and Waller 1994, Kelce et al. 1995, Li et al. 1994). Exposure to HCB has been associated with altered growth, precocious egg production, and autoimmune deficiencies in other species (Baturo et al. 1995, Schielen et al. 1993).

The precocious maturation exhibited by female sole from the Hylebos may also be partially related to their size, nutritional status, or growth rate. In fish age 5 and under, Hylebos fish tended to be about 5 to 15% larger and have a 5 to 10% higher condition factor than Colvos fish of the same age. The difference was most marked in age 1 fish. However, the average length of Hylebos fish at 1 and 2 years of age (240 to 260 mm) was still smaller than the average length at first reproduction in female English sole, which is generally 280-300 mm (Johnson et al. 1991, Lassuy 1989).

In contrast to precocious maturation, inhibited ovarian development in adult sole was most closely associated with exposure to aromatic hydrocarbons, and showed only weak correlations with tissue CH concentrations. Liver HCB concentrations showed no association with inhibited ovarian development in adult sole. Interestingly, DNA adducts, which are an indicator of chronic exposure to genotoxic PAHs, showed no relationship with the probability of gonadal development in older fish. This, in combination with the fact that biliary FAC_NPH levels showed a particularly strong association with inhibited gonadal development in adult sole, raises the possibility that it may be the low molecular weight or non-genotoxic PAHs that are the causative agents for this abnormality. This is consistent with results of laboratory studies showing that exposure of female Atlantic croaker to naphthalene alone, in the absence of higher molecular weight PAHs, led to reduced gonadal growth and increased ovarian atresia (Thomas and Budiantara 1995).

The finding of inhibited ovarian development in adult Hylebos sole parallels our observations in adult female English sole from other sites in Puget Sound (Johnson et al. 1988, 1993). Adult female sole from creosote-contaminated sites in Eagle Harbor, for example, which are exposed to high concentrations of AHs, but relatively low levels of CHs, exhibit a particularly low rate of gonadal development as well as other changes in reproductive endocrine function. The overall pattern of reproductive dysfunction in female English sole from the Hylebos Waterway is also quite similar to the pattern observed in female sole from the Duwamish Waterway. The proportions of non-reproducing adult females are similar at the two sites, and adult sole from the Hylebos Waterway exhibit other abnormalities, such as reduced plasma estradiol concentrations and ovarian lesions, which are also observed in Duwamish Waterway sole (Johnson et al. 1988). Additional study in needed to determine whether juvenile sole at sites with contaminant profiles similar to that of the Hylebos Waterway also exhibit precocious maturation.

It is unlikely that these changes in growth and reproductive development in Hylebos Waterway sole represent a genetic adaptation of this fish stock to its particular habitat, because the breeding strategy of English sole is unlikely to lead to the formation of distinct localized genetic stocks (Waples 1987). Sole residing at nearshore sites such as the Hylebos Waterway and Colvos Passage generally do not constitute discrete breeding populations, but migrate to common breeding areas to spawn (Johnson et al. 1991, Lassuy 1989). Furthermore, English sole have waterborne eggs and larvae that are transported by currents from the spawning grounds to nearshore nursery sites (Lassuy 1989). Consequently, there is probably considerable genetic mixing among sole sub-populations from different areas of the Sound. Typically, marine fish with breeding strategies similar to English sole do not form localized stocks and show relatively little genetic adaptation to local environmental conditions (Waples 1987).

Spawning success and egg and larval viability were not measured in this assessment. However, in a previous study, fish from the Duwamish Waterway, whose PCB and PAH exposure levels are comparable to those reported for Hylebos fish in this study, showed a reduction in egg fertilization success of 8% and a reduction in larval viability of 15% in comparison with fish from an uncontaminated reference site (Casillas et al. 1991). Duwamish fish also exhibited changes in egg size, tending to produce smaller eggs than fish of comparable maturity from reference areas (Johnson et al. 1997). Because of the similar exposure regimens, such effects might also be expected to occur in Hylebos English sole.

In the Hylebos Waterway, 40-50% of juvenile sole showed precocious maturation and 20-30% of adult sole showed inhibited gonadal development. Substantial proportions of fish also exhibited DNA damage and hepatic lesions.

Proportions of female English sole from the Hylebos Waterway showing contaminant associated injury ranged from approximately 20 to over 90%, depending on the type of biological effect under consideration. Ninety-six per cent of fish from Hylebos Waterway had DNA adduct levels in liver above the mean reference level of ~ 6 nmol/mol bases in Colvos Passage fish. Additionally, 13% of Hylebos fish exhibited one or more toxicopathic liver lesions. Between 40 and 50% of Hylebos fish showed precocious sexual maturation, entering vitellogenesis at age 1 or 2, and the proportion maturing at age 3 was about 30% higher than at Colvos Passage. Among older females, the proportion of females maturing was 20 to 30% lower than that observed in Colvos Passage.

In male English sole collected from the Hylebos Waterway as part of the Flatfish Toxicopathic Injury Study (Section 2), indicators of contaminant exposure, liver lesion prevalences and levels of DNA adducts in liver were similar to levels in female fish from the Flatfish Reproductive Injury Study. Additionally, a testicular neoplasm was observed in one male sole from the Hylebos Waterway. This rare lesion has been observed by our laboratory only once before, in a male sole from the Duwamish Waterway, and the role that chemical contaminants may play in its development is unclear. These data suggest the potential for reproductive injury in male fish from the Hylebos Waterway. However, the reproductive function of males from the Hylebos Waterway has not yet been evaluated, so the level of injury to this part of the population cannot be quantified at this time. Although the effects of chemicals on male fish may be a cause for concern, our initial injury assessment focused on female fish because we had evidence from previous studies (Johnson et al. 1988, 1993; Casillas et al. 1991, Collier et al. 1992, Landahl and Johnson 1993) to suggest that there was a high risk of reproductive impairment in female English sole from the Hylebos Waterway. Such data are lacking for male English sole.

The reproductive injury observed in Hylebos fish would reduce the number of eggs and presumably the number of larvae contributed by these animals to the English sole population in Commencement Bay and southern Puget Sound, potentially reducing the overall productivity of the sole population.

Because a complex array of factors can influence fish population growth rates, it is difficult to predict with high certainty the consequences of contaminant-related injury to female English sole for sole populations in the Hylebos Waterway and Commencement Bay. The altered pattern of reproductive development in Hylebos fish, however, would clearly have an impact on the number of eggs produced by fish of different age classes from this site. On the basis of the maturity schedules and length-at-age relationships observed in the present study, and the relationships between fish length and egg production deduced from previous studies (Johnson et al. 1997), projected egg production for each age class of English sole from the Hylebos Waterway and Colvos Passage was calculated. The results are shown in Figure 3.22. These egg production estimates do not take into account the probability that the fish will successfully spawn, but rather estimate the potential number of eggs produced based on fish age and probability of gonadal maturation. The calculations show that although Hylebos Waterway fish begin producing eggs earlier than Colvos Passage fish, after about age 4 they produce only 50 to 75% as many eggs as Colvos Passage fish. The cumulative egg production over the first 10 years of life would be about 1,730,000 eggs for a Colvos Passage fish, but only about 70% of that number (1,195,000 eggs) for a Hylebos Waterway fish.

Previous studies with English sole from other contaminated sites in Puget Sound suggest that the reproductive output of Hylebos Waterway sole may be further decreased by contaminant-related declines in spawning ability and egg and larval viability. Currently, we have no data on the spawning success of Hylebos Waterway English sole. However, for a comparable site, the Duwamish Waterway, the proportion of vitellogenic fish that spawned successfully in laboratory trials was 54%, as compared to 90% for fish from an uncontaminated reference area at Port Susan, Washington. Fertilization and larval viability rates for those fish that spawned were 52% and 74% for reference fish, as compared with 44% and 59% for Duwamish fish (Casillas et al. 1991; Landahl and Johnson 1993). If we assume that conditions in the Hylebos Waterway and the Duwamish Waterway are similar, then we would predict that spawning success and egg and larval viability rates for Hylebos sole would be similar o those observed for Duwamish Waterway sole. Using the Duwamish Waterway data, in combination with the egg production estimates discussed in the preceding paragraph, projected production of viable larvae was calculated for each age class of English sole from the Hylebos Waterway and Colvos Passage (Figure 3.22). These calculations suggest that for most of their lives, Hylebos Waterway fish would produce only about 25% of the viable offspring produced by Colvos Passage English sole. Over ten years, a Colvos Passage fish would produce an estimated 599,000 larvae, while a Hylebos Waterway fish would produce about 167,000 larvae, 28% of the number produced by the Colvos Passage fish. It is possible that the viability of eggs and larvae from precociously maturing Hylebos sole may be even lower than predicted based on the Duwamish results, as studies of English sole collected from Puget Sound spawning sites at University Point and Duwamish Head suggest that egg and larval viability tend to be lower in small and young fish than in their older and larger counterparts (Collier et al. 1992). Additionally, altered testicular function in male fish might also reduce the reproductive contribution of Hylebos Waterway fish to the English sole population. Although we did not examine effects on male fish as part of the present study, other researchers have shown that exposure to estrogenic or other endocrine-disrupting contaminants can reduce testicular growth and sperm production in male fish (Jobling et al. 1996).

The alterations in reproductive development observed in female English sole from the Hylebos Waterway would clearly reduce the number of eggs and larvae contributed by fish from this site to the English sole population in Commencement Bay and southern Puget Sound. These contaminant-related declines in reproductive output could have significant adverse consequences for the population as a whole, if their effects are not offset by density-dependent changes in recruitment, immigration, or other compensatory mechanisms. Moreover, declines in reproductive output may contribute to a reduction in the overall resilience of the English sole population in Puget Sound in the face of multiple environmental stressors such as overfishing, climate change, or destruction or alteration of nursery or other critical habitats.

In conclusion, the results of the Flatfish Reproductive Injury Study indicate that reproductive function is significantly altered in female fish from the Hylebos Waterway, and that this alteration is associated with exposure to chemicals that are characteristic of the contaminant profile in Hylebos Waterway sediments. Because several classes of chemicals co-occur in Hylebos Waterway sediments and fish are exposed to a complex mixture of contaminants, our ability to establish unambiguous cause-and-effect relationships between individual chemicals and the types of reproductive dysfunction observed is restricted. Both chlorinated hydrocarbons and aromatic hydrocarbons emerged as potential risk factors for precocious maturation, while aromatic hydrocarbons were more closely associated with inhibited gonadal development in adult sole. Although the population-level risk of contaminant exposure in Hylebos Waterway could be determined with greater certainty with better field data on actual recruitment and migration patterns in English sole, the reduced reproductive output of Hylebos Waterway sole could potentially contribute to declines in the productivity and resilience of the English sole population in southern Puget Sound.

Table of Contents