USGS - Pesticide National Synthesis Project Pesticides Report

Reconnaissance of 17ß -Estradiol, 11-Ketotestosterone, Vitellogenin, and Gonad Histopathology in Common Carp of United States Streams: Potential for Contaminant-Induced Endocrine Disruption

By Steven L. Goodbred, Robert J. Gilliom, Timothy S. Gross, Nancy P. Denslow, Wade L. Bryant, and Trenton R. Schoeb

U.S. Geological Survey Open-File Report 96-627

DISCUSSION

Use of Biomarkers as Indicators of Potential Endocrine Disruption

An important issue with regard to use of sex steroid hormones as biomarkers of potential endocrine disruption is the natural temporal variability of hormones in fish. This is particularly true for this reconnaissance in which one-time samples were collected from a wide geographic area. The sampling period of this study (late August to early December) is after carp spawn (March to mid August [Panek, 1987]), during a time of gonadal recrudescence, and before water tem-peratures go below 15°C (Down and others, 1990). Samples were collected during this part of the reproductive cycle to minimize hormone variance, which is much greater in the spring through summer spawning period (Barry and others, 1990; Chang and Chen, 1990; and Down and others, 1990). Even within this part of the reproductive cycle, sex steroid hormones among individual fish at a site can vary up to 30-fold (Chang and Chen, 1990; Down and others, 1990; and Folmar and others, 1996). In this study, however, the 10th and 90th percentiles of 17ß-estradiol and 11-ketotestosterone in individual males and females were within a factor of 5 of each other for most sites. Nevertheless, detecting possible endocrine disruption through differences in sex steroid hormones, even within the same period of the reproductive cycle, is difficult because of natural variability. Generally, subtle effects of potential endocrine disrupters are unlikely to be detected, and only the strongest influences of contaminants are likely to be evident.

In addition to the individual sex steroid hormones---17ß-estradiol and 11-ketotestosterone---their ratio was used in this study as an indicator of possible endocrine disruption. Folmar and others (1996) concluded that the ratio of 17ß-estradiol to testosterone appears to be a sensitive marker of abnormal sex steroid concentrations in carp, but has little functional significance because normal ranges have not been established. However, Hileman (1994) concluded that a specific ratio of estrogen to testosterone is necessary for sexual differentiation in developing animals and that alteration of the ratio can result in incomplete or improper gonadal development. Additionally, the balance between these two hormones determines a fish's phenotype, which includes sex characteristics, differentiation of the brain and behavior, and development of other reproductive organs (Lehninger, 1982; and Hunter and Donaldson, 1983). There may be an acceptable range of proportions of female to male sex steroid hormones at various stages in a fishes life cycle, and the range may be most critical in immature and developing fish. In this study, values of the E₂/11KT ratio were mostly below 1.0 for males, and mostly above 1.0 for females. Although ranges have not been established for classifying fish as normal or not, extreme values of the ratio compared to other fish, or correlations between the ratio and contaminant levels, are useful indicators of potential endocrine disruption.

Vitellogenin induction in male fish has also been presented as evidence of endocrine disruption (Purdom and others, 1994 and Folmar and others, 1996), and vitellogenin was detected in one or more male fish at over half the study sites. However, the level of vitellogenin was never over 1 mg/mL in any individual, far below the normal range found in females. Purdom and others (1994) and Folmar and others (1996) found vitellogenin induction in male carp below sewage treatment plants (STP) and their results suggested that exposure to estrogenic compounds in the effluent, such as alkylphenol-ethoyxlates could be the cause. Folmar and others (1996) also showed that an increase in 17ß-estradiol was not responsible for this vitellogenin induction, but a reduction in testosterone lowered the ratios of 17ß-estradiol/testosterone in males at the STP impacted site. Some of our sites where vitellogenin was found in male carp may have been influenced by sewage effluent, and substances in the effluent may be partly responsible for the induction. However, vitellogenin also was detected in males at several minimally contaminated sites with no sewage effluent, such as Don Pedro Reservoir in California (DPR), which indicates that some male fish have low background vitellogenin during some portion of the reproductive cycle. Male fish have a vitellogenin gene that is not usually expressed, but male fish of several species not exposed to endocrine disrupting compounds have been documented to have low concentrations of vitellogenin present (Copeland and Thomas, 1988 and Goodwin and others, 1992).

To supplement the hormone and vitellogenin biomarkers, gonads also were evaluated for abnormalities. Evidence of endocrine disruption in mammals and reptiles has been documented through gonadal histopathology, such as multinuclear eggs, too many eggs in the ovary, and dark bar-shaped structures in tubules of testes (McLachlan and Arnold, 1996). However, this study found only one possible gonadal abnormality (a few basophyllic cells in a testes, which are probably primary ova) out of 438 gonads examined. Other possible endocrine-disruption effects that have been observed in fish, such as reduced ovary size, lower egg viability, and delayed sexual maturity (McMaster and others, 1991; Munkittrick and others, 1992; and Hontela and others, 1995), were not assessed in this study. Thus, the evidence for potential endocrine disruption that is indicated by differences in hormones and vitellogenin among sites, and signif-icant correlations between hormone levels and contaminants, is not confirmed by gonad abnormalities that were evaluated in this study.

Contaminants and Biomarkers

An important objective of this reconnaissance study was to evaluate whether differences in biomarkers among sites may be related to environmental contaminants. There is strong evidence that xenobiotics can induce toxicity and changes in endocrine systems (Atterwill and Flack, 1992), and field studies have found correlations between the levels of endocrine disruption in fish and the types and degree of contaminant exposure (Folmar, 1993). Most of the evidence in teleost fish shows contaminants reducing levels of circulating sex steroid hormones and vitellogenin (Folmar, 1993), although some contaminants, such as cadmium (Sangalang and Freeman, 1974), DDT (Denison and others, 1981), ßi-hexachlorocyclohexane (ß-HCH [Webster and others, 1985]), nonylphenol (Waldock and others, 1994) and PAHs (Janssen and others, 1995), may increase steroid hormones or vitellogenin.

Table 12 summarizes the results of the pair-wise correlation analyses of contaminant-biomarker relations for both males and females. There are five statistically significant (a=0.05) correlations between biomarkers and contaminant groups with four negative and one positive. Of 15 correlations with p values less than 0.25, 14 were negative. The strongest patterns common to both males and females are: (1) negative correlations between the E₂/11-KT ratio and dissolved pesticides, and (2) negative correlations between phenols and both 17ß-estradiol and 11-ketotestosterone. Notable differences between males and females are: (1) the significant negative correlation between 17ß-estradiol and organochlorine pesticides shown in males and not females, and (2) the significant positive correlation between 11-ketotestosterone and dissolved pesticides shown in females and not males.

The significant correlations between dissolved pesticides in water and E₂/11-KT for both males and females are shown in figure 6. Best-fit regression lines for males and females follow the same general pattern, with the greatest reduction in E₂/11-KT occurring from 0.2 to 1 ug/L dissolved pesticides. The lowest E₂/11-KT ratios in both male and female carp were found in the Platte River at Louisville (PR-L) in Nebraska, which had the highest total concentration of dissolved pesticides in water. Studies have shown that some water soluble pesticides, such as atrazine (Babic-Gojmerac and others, 1989 and Simic and others, 1991), alachlor (U.S. Environmental Protection Agency, 1984 and Amdur and others, 1991) and carbaryl (Amdur and others, 1991), all of which were detected at one or more sites in this study, affect endocrine systems. However, the available contaminant data are not sufficient to determine which specific pesticides or groups of pesticides could be responsible for the reduction in E₂/11-KT ratios.

The contaminant group having the most consistent direction of correlation with biomarkers is the phenols, which have negative correlation coefficients with 17ß-estradiol and 11-ketotestosterone in male and female fish (although most are not significant at the 0.05 level). Only the correlation with 11-ketotestosterone in males was statistically significant. Within the phenol chemical group, the alkylated phenols have been shown to bind to the estrogen receptor, displaying estrogenic effects, such as vitellogenin induction in male fish (Waldock and others, 1994 and White and others, 1994).

Organochlorine pesticides showed a significant negative correlation with male 17ß-estradiol. A similar response was seen in largemouth bass from a site in Florida that is contaminated with organochlorine pesticides (Gross and others, 1995). Results from another study (Singh and Singh, 1987) also showed that organochlorine pesticides (lindane and y-benzene hexachloride [y-BHC]) reduced estradiol in fish.

Significant pair-wise correlations with biomarkers were not found for PCBs, phthalates, or PAHs. Others, however, have reported reduced estradiol in fish exposed to PCBs. PCB concentration was found in this study to explain a significant portion of the variance in E₂/11KT in female carp in a multiple regression with dissolved pesticides. Phthalates, widely used chemicals in plastics, are routinely found in sampling of aquatic environments and have been shown by others to affect endocrine systems (Wams, 1987; Treinen and others, 1990; and Laskey and Berman, 1993). Phthalate concentration was found in this study to explain a significant portion of the variance in E₂/11KT in male carp in a multiple regression with dissolved pesticides. Available literature on fish mostly show that PAHs reduce sex steroid hormones and vitellogenin (Johnson and others, 1988; Thomas 1988; Singh, 1989; and Sol and others, 1995), although one study showed that PAHs in contaminated sediment increased steroid hormones and vitellogenin (Janssen and others, 1995). The absence of strong findings that these contaminant groups are correlated with biomarkers at the reconnaissance study sites may be due to one or a combination of several reasons: there are no effects at the levels of exposure studied; effects are subtle in relation to uncontrolled sources of variability; or that various contaminants have effects in opposite directions that confound simple correlation analysis.

Direction of Future Studies

An important objective of this reconnaissance study was to determine the need and priorities for further endocrine disruption studies. Improved information is needed in several areas to evaluate whether endocrine disruption in fish is actually occurring in some streams and, if so, to determine its causes and its effects on fish populations.

Improved knowledge of the occurrence of endo-crine disruption in fish will require more detailed and integrated assessment of the exposure of fish to contaminants over time in relation to reproductive cycles and variability in endocrine systems. Major elements of study that should be included are:

Assessment of biological effects on both individual fish and populations, and how these effects correlate with the degrees of endocrine disruption indicted by biomarkers and other measures.
Inclusion of fish species that are more sensitive to endocrine disruption than carp.
Contemporaneous temporal characterization of contaminant exposure and endocrine status over the complete annual reproductive cycle.
Broader characterization of contaminant exposure (including background levels), such as trace elements, additional pesticides, and nitrates.
Improved pairing of reference sites with contami-nated sites. Such site pairs need to be as similar as possible in all respects, except contaminant levels.

Investigation of biological effects associated with altered endocrine systems is probably the most difficult task. Detailed discussion of methods to assess individual and population-level effects is beyond the scope of this paper. Initially, however, reproductive impairment of individuals should be assessed using a range of techniques on selected fish species, including gonadal somatic index, fecundity, egg size and hatchability, and sperm quantities and qualities. These indicators should be combined with basic characterization of fish populations, such as sex and age distribution, condition, health, and growth. Multiple lines of evidence in assessing reproductive impairment are critical because of the high degree of inherent variability.

Achieving the types of improvements in information on biological effects and contaminant influ-ences outlined above will require a distinct shift in study design from the "one sample from many sites" approach of the reconnaissance, to the "intensive study of a few sites" approach for more detailed study. A paired-site approach is suggested, with one "contaminated" site and one "reference" site located in each of the three regions with the clearest evidence of endocrine disruption: the Northeast, the Mississippi River Basin, and the West. Criteria for selection of contaminated sites should include representation of the regions' streams, historical availability of contaminant and fish population data, availability of comparable species with sufficient numbers to sample, and availability of a comparable reference site. Potential "contaminated" sites for further study include the Shenandoah River at Millville, West Virginia (SR-M) or the Mohawk River at Frankfort, New York (MR-F) in the Northeast; the Platte River at Louisville, Nebraska (PR-L) or the White River at Hazleton, Indiana (WR-H) in the Mississippi River Basin; and the San Joaquin River at Fremont Ford, California (SJR-FF) in the West.

Next - Summary