Synthetic glucocorticoids are pharmaceutical compounds prescribed in human and veterinary medicine as anti-inflammatory agents and have the potential to contaminate natural watersheds via inputs from wastewater treatment facilities and confined animal-feeding operations. Despite this, few studies have examined the effects of this class of chemicals on aquatic vertebrates. To generate data to assess potential risk to the aquatic environment, we used fathead minnow 21-d reproduction and 29-d embryo-larvae assays to determine reproductive toxicity and early-life-stage effects of dexamethasone. Exposure to 500 µg dexamethasone/L in the 21-d test caused reductions in fathead minnow fecundity and female plasma estradiol concentrations and increased the occurrence of abnormally hatched fry. Female fish exposed to 500 µg dexamethasone/L also displayed a significant increase in plasma vitellogenin protein levels, possibly because of decreased spawning. A decrease in vitellogenin messenger ribonucleic acid (mRNA) expression in liver tissue from females exposed to the high dexamethasone concentration lends support to this hypothesis. Histological results indicate that a 29-d embryo-larval exposure to 500 µg dexamethasone/L caused a significant increase in deformed gill opercula. Fry exposed to 500 µg dexamethasone/L for 29 d also exhibited a significant reduction in weight and length compared with control fry. Taken together, these results indicate that nonlethal concentrations of a model glucocorticoid receptor agonist can impair fish reproduction, growth, and development.

Effects of bisphenol A (BPA) on ovarian transcript profiles as well as targeted end points with endocrine/reproductive relevance were examined in two fish species, fathead minnow (Pimephales promelas) and zebrafish (Danio rerio), exposed in parallel using matched experimental designs. Four days of waterborne exposure to 10 μg BPA/L caused significant vitellogenin induction in both species. However, zebrafish were less sensitive to effects on hepatic gene expression and steroid production than fathead minnow and the magnitude of vitellogenin induction was more modest (i.e., 3-fold compared to 13,000-fold in fathead minnow). The concentration-response at the ovarian transcriptome level was nonmonotonic and violated assumptions that underlie proposed methods for estimating hazard thresholds from transcriptomic results. However, the nonmonotonic profile was consistent among species and there were nominal similarities in the functions associated with the differentially expressed genes, suggesting potential activation of common pathway perturbation motifs in both species. Overall, the results provide an effective case study for considering the potential application of ecotoxicogenomics to ecological risk assessments and provide novel comparative data regarding effects of BPA in fish.

Animal waste from livestock farming operations can contain varying levels of natural and synthetic androgens and/or estrogens, which can contaminate surrounding waterways. In the present study, surface stream water was collected from 6 basins containing livestock farming operations. Aqueous concentrations of 12 hormones were determined via chemical analyses. Relative androgenic and estrogenic activity was measured using in vitro cell assays (MDA-kb2 and T47D-Kbluc assays, respectively). In parallel, 48-h static-renewal in vivo exposures were conducted to examine potential endocrine-disrupting effects in fathead minnows. Mature fish were exposed to surface water dilutions (0%, 25%, 50%, and 100%) and 10-ng/L of 17α-ethynylestradiol or 50-ng/L of 17β-trenbolone as positive controls. Hepatic expression of vitellogenin and estrogen receptor α mRNA, gonadal ex vivo testosterone and 17β-estradiol production, and plasma vitellogenin concentrations were examined. Potentially estrogenic and androgenic steroids were detected at low nanogram per liter concentrations. In vitro estrogenic activity was detected in all samples, whereas androgenic activity was detected in only 1 sample. In vivo exposures to the surface water had no significant dose-dependent effect on any of the biological endpoints, with the exception of increased male testosterone production in 1 exposure. The present study, which combines analytical chemistry measurements, in vitro bioassays, and in vivo fish exposures, highlights the integrated value and future use of a combination of techniques to obtain a comprehensive characterization of an environmental chemical mixture.

Environmental assessment of complex mixtures typically requires integration of chemical and biological measurements. This study demonstrates the use of a combination of instrumental chemical analyses, effects-based monitoring, and bio-effects prediction approaches to help identify potential hazards and priority contaminants in two Great Lakes Areas of Concern (AOCs), the Lower Green Bay/Fox River located near Green Bay, WI, USA and the Milwaukee Estuary, located near Milwaukee, WI, USA. Fathead minnows were caged at four sites within each AOC (eight sites total). Following 4d of in situ exposure, tissues and biofluids were sampled and used for targeted biological effects analyses. Additionally, 4d composite water samples were collected concurrently at each caged fish site and analyzed for 132 analytes as well as evaluated for total estrogenic and androgenic activity using cell-based bioassays. Of the analytes examined, 75 were detected in composite samples from at least one site. Based on multiple analyses, one site in the East River and another site near a paper mill discharge in the Lower Green Bay/Fox River AOC, were prioritized due to their estrogenic and androgenic activity, respectively. The water samples from other sites generally did not exhibit significant estrogenic or androgenic activity, nor was there evidence for endocrine disruption in the fish exposed at these sites as indicated by the lack of alterations in ex vivo steroid production, circulating steroid concentrations, or vitellogenin mRNA expression in males. Induction of hepatic cyp1a mRNA expression was detected at several sites, suggesting the presence of chemicals that activate the aryl hydrocarbon receptor. To expand the scope beyond targeted investigation of endpoints selected a priori, several bio-effects prediction approaches were employed to identify other potentially disturbed biological pathways and related chemical constituents that may warrant future monitoring at these sites. For example, several chemicals such as diethylphthalate and naphthalene, and genes and related pathways, such as cholinergic receptor muscarinic 3 (CHRM3), estrogen receptor alpha1 (esr1), chemokine ligand 10 protein (CXCL10), tumor protein p53 (p53), and monoamine oxidase B (Maob), were identified as candidates for future assessments at these AOCs. Overall, this study demonstrates that a better prioritization of contaminants and associated hazards can be achieved through integrated evaluation of multiple lines of evidence. Such prioritization can guide more comprehensive follow-up risk assessment efforts.

Adaptive or compensatory responses to chemical exposure can significantly influence in vivo concentration-duration-response relationships. This study provided data to support development of a computational dynamic model of the hypothalamic-pituitary-gonadal axis of a model vertebrate and its response to aromatase inhibitors as a class of endocrine active chemicals. Fathead minnows (Pimephales promelas) were either exposed to the aromatase inhibitor fadrozole (0.5 or 30 μg/l) continuously for 1, 8, 12, 16, 20, 24, or 28 days or exposed for 8 days and then held in control water (no fadrozole) for an additional 4, 8, 12, 16, or 20 days. The time course of effects on ovarian steroid production, circulating 17β-estradiol (E2) and vitellogenin (VTG) concentrations, and expression of steroidogenesis-related genes in the ovary was measured. Exposure to 30 μg fadrozole/l significantly reduced plasma E2 and VTG concentrations after just 1 day and those effects persisted throughout 28 days of exposure. In contrast, ex vivo E2 production was similar to that of controls on day 8-28 of exposure, whereas transcripts coding for aromatase and follicle-stimulating hormone receptor were elevated, suggesting a compensatory response. Following cessation of fadrozole exposure, ex vivo E2 and plasma E2 concentrations exceeded and then recovered to control levels, but plasma VTG concentrations did not, even after 20 days of depuration. Collectively these data provide several new insights into the nature and time course of adaptive responses to an aromatase inhibitor that support development of a computational model (see companion article).