Jump to main content or area navigation.

Contact Us

Extramural Research

2006 Progress Report: A High Throughput Zebrafish Embryo Gene Expression System for Screening Endocrine Disrupting Chemicals

EPA Grant Number: R831301
Title: A High Throughput Zebrafish Embryo Gene Expression System for Screening Endocrine Disrupting Chemicals
Investigators: Callard, Gloria V.
Institution: Boston University
EPA Project Officer: Mustra, David
Project Period: October 1, 2003 through September 30, 2007
Project Period Covered by this Report: October 1, 2005 through September 30, 2006
Project Amount: $400,000
RFA: Development of High-Throughput Screening Approaches for Prioritizing Chemicals for the Endocrine Disruptors Screening Program (2003)
Research Category: Economics and Decision Sciences , Children's Health , Endocrine Disruptors



This project investigates the proposition that perturbations in the normal amount or timing of a hormone-regulated gene product can be taken as evidence of chemical exposure and used as end-points in a screening assay to detect endocrine disrupting chemicals (EDCs). Our approach is to demonstrate that real-time quantitative (q)-PCR analysis of expressed genes in living zebrafish embryos can be used to detect and characterize: (1) multiple subclasses of EDCs; (2) multiple gene- and tissue-targets of a given EDC subclass; and (3) EDCs that target genes both upstream and downstream of receptor binding. The objectives of the research project are to: (1) address an urgent need for regulators to better predict which of the estimated 87,000 chemicals in the environment have the potential to disrupt hormone-dependent processes of physiology, reproduction, and development; (2) provide biologically relevant criteria for prioritizing chemicals for further testing; and (3) help interpret reports of reproductive and developmental abnormalities in wildlife and humans. Two major classes of persistent environmental pollutants with potential endocrine disrupting effects are estrogen-like chemicals (xenoestrogens [XE]) that interact with estrogen receptors (ER) and estrogen signaling pathways and dioxin-like chemicals (DLC) that interact with arylhydrocarbon receptors (AhR) to perturb AhR- and ER-signaling pathways. Thus, the current project focuses on specific molecular markers of ER- and AhR-mediated actions and effects. In addition, the project is viewed as a proof-of-principle study to develop approaches and protocols with universal applicability for screening chemicals with effects on any other nuclear receptor signaling pathway of interest, and to establish the feasibility of procedures and endpoints that are amenable to high throughput modifications.

Progress Summary:

Design and Validation of qPCR Primers and Assays to Measure Genes Expressed in ER and AhR Signaling Pathways, and Genetic Markers of Development and/or EDC Effect

We have developed real time qPCR assays for 14 different zebrafish genes and 10 additional killifish genes and are using these genes as mechanism-βased markers of EDC action and effect in embryos in vitro and/or in adult fish in the laboratory or in the natural environment. Genes include P450aromB (marker of estrogen biosynthesis and estrogen effect in brain); P450aromA (marker of estrogen biosynthesis in ovary); ERα, -βα, and -βb (markers of estrogen action and effect); vitellogenin (vtg; marker of estrogen effect in liver); cyp1A and AhR2 (markers of dioxin action and effect in liver); Nurr1, Nor1, and Nur77 (orphan nuclear receptor mediators of neural, neuroendocrine and gonadal development and function); ER-related receptors (ERR) , -βα, -βb and -γb (orphan nuclear receptor mediators of diverse physiological processes); GAP43, a-tubulin, TH1 and -2 (markers of neurodevelopment and neural functions); and b-actin (universal marker for normalizing target gene expression).

P450 aromatase B and -A mRNA Analysis. Assay development and validation in embryonic, larval, and adult male and female zebrafish have been completed for these two genes and serve as a template for assay development of all other targeted mRNAs. Our assay is sensitive, precise, gene-specific, and has a very low intra- and interassay coefficient of variation (<10%). Reproducibility is excellent within and across operators when tested in multiple experiments and beginning at different starting points within a single experiment (e.g., RNA extraction, reverse transcription, PCR). Results confirm and extend data reported with earlier methods and reveal much greater estrogen-inducibility (>150-fold) than with semiquantitative RT-PCR. Each gene of interest is normalized to an internal standard (actin), after correction for gene-specific differences in amplification efficiency, which is more accurate and cost efficient than multiple idealized external standard curves. Appropriate choice of primers assures specificity with SYBR green, which is more cost effective than TaqMan probes for high-throughput applications. The small amount of input rtRNA (80 ng/microwell) enables us to assay greater than 10 genes in triplicate in a single embryo although, in practice, we pool embryos to minimize biological variation. As well, the assay is applicable to qPCR analysis of P450aromB and and other mRNAs in field-collected killifish adults and embryos (see Greytak et al. 2005).

ERα, -βa, and -βb; vtg; cyp1a1, AhR2; Nurr1, Nor1, Nur77; GAP43, β-tubulin; TH1 and -2 mRNA Analysis. Primer design, validation, and qPCR assays have been established for these mRNAs in zebrafish embryos with or without EDC treatment. Interim results have been described in meeting abstracts (Burnam, et al., 2005; McCurley, et al., 2005; Sawyer, et al., 2005). Additionally, the same assays have been applied to measure changes in gene expression in zebrafish embryos treated with environmental samples (Novillo, et al., 2005) and to measure genetic markers in killifish from EDC-polluted and unpolluted environments (Greytak and Callard, 2006).

mRNA Splice Variants. Approximately 60 percent of human genes are alternatively spliced, which increases functional versatility but confounds PCR-βased methods of expression analysis. Our findings show that certain EDC and environmental mixtures are able to induce splice variants of gene markers in the embryo expression system. Thus, careful selection of qPCR primers and in-assay validation using both authentic plasmids and RNA extracts from exposed embryos must be applied gene-by-gene to assure accurate quantitation.

Intron-Retained P450aromB Splice Variant. Testing of sediment elutriates from polluted freshwater ponds in our zebrafish embryo assay led to identification of a novel P450aromB splice variant containing a retained intron 8, which has an in-frame stop codon but lacks a nearby poly-A addition site (Novillo, 2005; 2006). This variant most likely is subject to nonsense-mediated decay; however, if translated, the resultant protein would be functionally inert because of deletion of the heme-binding and catalytic domains. Results of semiquantitative RT-PCR/Southern transfer analysis show that the ratio of variant:normal mRNAs increases upon exposure of embryos to contaminated sediments; that dioxin mimics sediment effects; and that estrogen increases normal mRNA but decreases variant accumulation. Moreover, the variant form is the earliest P450aromB transcript detected during embryogenesis, indicating it is an immature product. These results provide evidence of a heretofore unrecognized mechanism of EDC action. Currently, we are attempting to design qPCR primers that will allow us to discriminate between the variant and normal P450arom mRNAs in the same sample, and plan to use variant:normal ratios as an additional indictor of EDC effect in screening assays. Initial findings have been described in meeting abstracts (Novillo, et al., 2005), and a full manuscript is in preparation. (Novillo, et al., 2006)

ERα-Deletion Variants. Sequence analysis of end-to-end PCR products, together with Northern analysis, show that ERα has multiple mRNA splice variants in zebrafish (Yershov, Novillo, Hoover, and Callard, unpublished data). We have observed some of the same ERα, -βα, and -βb mRNA variants in killifish; specifically, ER mRNAs that are truncated at the 3’-ends (e.g., lacking normal ligand binding domain and AF2 domain). Truncated or deletion ER variants have also been described in human breast cancer cells, where they have been shown to act as dominant negatives. In theory, therefore, EDC-induced changes in the ratio of different ER forms in a given tissue could alter responsiveness to circulating estrogen. Interestingly, killifish from a polychlorinated biphenyl (PCB)-polluted environment are estrogenized (as indicated by estrogen markers) (Greytak, et al., 2005; 2006) but hyporesponsive to administered estrogen (as shown in dosing experiments) (Greytak and Callard, 2005). We currently are seeking to determine whether there are differences in the type or quantity of ER variants in killifish from estrogenic versus reference environments and whether variant accumulation is regulated by estrogen and/or dioxin exposure.

Gene Discovery and Characterization. To assemble a panel of estrogen- and dioxin-responsive genes that is broadly representative of each tissue-type and developmental process during embryogenesis in zebrafish, we are using multiple complementary gene-finding approaches:

Cloning and Characterization of Zebrafish Nurr1, Nor1, and Nur77. Although mammalian studies show that all three members of the NR4 family of orphan nuclear receptors (Nurr1, Nor1, Nur77) are highly expressed in neural tissues; are essential for development of dopaminergic neurons (Nurr1), hippocampal neurons (Nor1), or neuroendocrine functions (Nur77); and are regulated by estrogen (Nur77), these genes have not been previously cloned and characterized in zebrafish nor have they been investigated as possible mediators of EDC action on the developing central nervous system. We have cloned and characterized all three NR4 genes in zebrafish and shown that they are constitutively expressed and forskolin inducible as early as 48 hpf. In situ hybridization shows that expression begins initially in neural tissues, including retina. Nur77, but not Nor1 or Nurr1, is inducible by estrogen and estrogen-like chemicals. See abstracts and manuscript in preparation by Sawyer, et al. (2005; 2006).

Microarray Analysis of Zebrafish Embryos. An unbiased genome-wide search (microarray analysis, Compugen), identified 91 estrogen-regulated and 316 dioxin-regulated genes in zebrafish embryos (of 17,000 total). Of these, 16 genes were coregulated, and 18 were oppositely regulated by estrogen and dioxin signifying intersection of the two pathways on multiple gene targets (McCurley, 2005; 2006).

Subtractive cDNA Library Construction. Subtractive hybridization (PCR select/suppression amplification) is being used to identify genes turned on or off during neuroregeneration in adult zebrafish. This panel of genes will be inspected for overlap with genes identified as estrogen- and/or dioxin-responsible by microarray analysis of whole embryos. The identified genes are expected to provide valuable reagents for monitoring the progression of neurodevelopment in zebrafish embryos and for testing effects of known and suspected EDCs and other classes of neurotoxicants.

Cloning and Characterization of Killifish ERRα, -βa, -βb and -γb. To extend investigation of environmental EDC effects to other members of the nuclear receptor superfamily, we have collaborated in a study to clone and characterize ERR in killifish (Tarrant, et al., 2006). ERRs are orphan nuclear receptors with diffuse but incompletely defined functions (in mammals). In killifish, ERRα, but not other ERR genes, is downregulated by estrogen in heart.

Characterization of Embryonic Gene Expression Profiles and Optimization of Treatment Protocols

EDC screening by analysis of gene expression in zebrafish embryos has the advantages of an in vitro assay with the added value of data from an intact organism. Critical to development of this approach is a treatment protocol that maximizes specificity, sensitivity, and responsiveness to a broad range of EDCs. Different treatment protocols were evaluated using authentic ER (estradiol, E2) and AhR (dioxin) ligands and qPCR to measure changes in our panel of target genes (Burnam, et al., 2005; 2006; Sawyer, et al., 2005; 2006). Collectively, studies with living zebrafish embryos show that effects are ligand-, gene-, tissue-type-, stage of development-, dose- and time-dependent, confirming our hypothesis that no one cell-free- or in vitro assay can be relied on to identify all chemicals with EDC activity.

Developmental Profiles. First, the developmental pattern of expression was determined for each gene in untreated embryos/larvae (0-120 hpf). As expected, the time of onset and subsequent changes in expression were gene-specific and corresponded to known morphogenetic events (e.g., onset of P450aromB was much earlier than vtg).

Cumulative Treatment Effects. To determine effects of exposure duration, different concentrations of E2 (0.0001, 0.001, 0.01, 0.1, and 1 μM) or dioxin (0.001, 0.01, 0.1, and 1 nM) were added to media between 24-48 hpf, 24-72 hpf, 24-96 hpf, or 24-96 hpf. Results indicate that the embryonic estrogen response system is gene-specific as measured by magnitude (vtg > P450aromB > ERα); sensitivity (P450aromB > vtg > ERα); and specificity (ERβa, -βb, actin, cyp1A1, AhR2 = no change). The AhR response system was also gene specific: cyp1A1 (200X increase), AhR2 (2X increase) but the ~EC50 was similar for both. P450aromA responses were inconsistent: two experiments, 15X downregulation; two experiments, no change; one experiment, 1X upregulation. P450aromB, ERα, ERβa, -βb, AhR1, AhR2, vtg, and actin were unresponsive. Predictably, fold induction of each gene increased with duration of exposure. In general, however, prehatch embryos (< 72 hpf) had relatively weak responses when compared to posthatch larvae (> 72 hpf).

Stage-Related Changes in Sensitivity and Responsiveness. To further evaluate the ontogeny of estrogen and dioxin response systems, embryos were treated with a range of E2 (0.0001, 0.001, 0.01, 0.1, and 1 μM) or dioxin concentrations (0.001, 0.01, 0.1, and 1 nM) for 24 hours at each of four developmental windows: 24-48, 48-72, 72-96, and 96-120 hpf. Results confirmed that prehatch embryos, when compared to later stages, are relatively unresponsive to E2 and dioxin, and also revealed that responses between 72-96 hpf were at or near the magnitude of responses at 96-120 hpf. For routine screening, therefore, a 24-hour treatment period between 72-96 hpf was selected as optimal. This paradigm minimizes chemical effects on morphogenesis per se (as opposed to gene regulation) and has the advantage that less than 96 hpf larvae are less motile than later larvae and hence easier to treat and collect.

Screening of Representative Chemical Classes

Using standardized qPCR assays and embryo treatment protocols, we have begun to test diverse known and suspected EDC alone and in combination and to evaluate effects of environmental samples. Results show that the zebrafish assay can detect ER and AhR ligands, alone or in combination, chemicals that can be metabolized to ER ligands, and estrogen-like and dioxin-like chemicals that are present as mixtures in environmental samples. Responses are gene-specific and dose-related. Depending on the presence and relative concentration of authentic E2, agonist versus antagonist effects of relatively weak estrogens can be discerned.

ER Ligands. Comparing E2 with bisphenol A (BPA) and diethylstilbestrol (DES) at different concentrations, estrogen responsive genes were found to be ligand- and gene-specific: P450aromB, DES > E2 > BPA; vtg, E2 > DES > BPA. The relatively weak agonist, 4-HydroxyPCB51, upregulated estrogen responsive genes at high doses (agonist) but in combination with E2 displayed antagonist effects.

AhR Ligands. Experiments comparing TCDD, PCB 126, DMBA, beta-naphthoflavone (BNF), and α-naphthoflavone (an AhR antagonist) alone and in combination are in progress.

ER-AhR Ligand Interactions. To evaluate how the two signaling pathways intersect in the whole organism, embryos were treated with E2 or TCDD alone or in combination at different doses and ratios. Results suggest that each pathway antagonizes the other, and that the combination of two chemical classes is additive on toxicity.

Aromatizable and Non-Aromatizable Steroids. The aromatizable androgen testosterone (T) at different concentrations (0.1, 1, 10, and 100 µM) upregulated estrogen responsive genes, and the response was blocked by the ER antagonist ICI (1 and 10 µM). These results indicate that zebrafish embryos are able to convert T to E2, which in turn activates ER and estrogen responsive genes. Paradoxically, nonaromatizable androgens (DHT, and its metabolite 5a, 3b-androstandiol) upregulated estrogen responsive genes at relatively high doses, and the response was blocked by ICI. These results are consistent with studies showing that 5α, 3β-androstandiol is a relatively weak but effective ER ligand.

Miscellaneous Suspected EDC. To evaluate effects of the pesticides atrazine, vinclozolin, and M2 (a vinclozolin metabolite), embryos were exposed to different doses in a cumulative treatment paradigm. With the exception of cyp1A1, which was induced by 10 µM vinclozolin, no changes in gene expression were observed. Thus, using zebrafish embryos, we cannot confirm reported effects of these compounds on aromatase expression or estrogen signaling.

Environmental Samples. Water and sediment were collected from polluted and reference ponds near the Massachusetts Military Reservation Superfund site on Cape Cod, Massachusetts. Impacted ponds had turtles with reproductive abnormalities. Water from the impacted pond induced dioxin responsive genes (AhR2, cyp1A1), while the sediment elutriates from the same pond induced estrogen responsive genes (vtg, P450aromB, ERα) and increased the intron-retained P450aromB variant. These results suggest that the zebrafish embryo bioassay has potential for isolating and identifying EDCs in the natural environment.

Future Activities:

Future work will: (1) accelerate routine screening of diverse classes of EDC chemicals and agonist-antagonist combinations using our optimized treatment protocol and qPCR analysis of all zebrafish mRNAs in our panel of genes; (2) complete characterization and quantitation of P450aromB and ER splice variants in control and EDC-treated embryos and adults and test effects of environmental samples from polluted and unpolluted environments to clarify the role of EDCs in impacting normal splice mechanisms; (3) complete collection, processing, and assay of environmental samples from polluted (e.g., New Bedford Harbor and Cape Cod Superfund sites) and reference environments to link laboratory testing to actual problems in the natural environment; (4) expand the panel of genes used as end-points for screening EDCs based on results of microarrays and subtraction libraries; and (5) complete and submit for publication all manuscripts/preliminary reports listed among the abstract/meeting presentations below.

Journal Articles on this Report : 4 Displayed | Download in RIS Format

Other project views: All 54 publications 7 publications in selected types All 7 journal articles

Type Citation Project Document Sources
Journal Article Greytak SR, Champlin D, Callard GV. Isolation and characterization of two cytochrome P450 aromatase forms in killifish (Fundulus heteroclitus): differential expression in fish from polluted and unpolluted environments. Aquatic Toxicology 2005;71(4):371-389. R831301 (2006)
R831301 (Final)
  • Abstract from PubMed
  • Full-text: Science Direct HTML
  • Abstract: Science Direct Abstract
  • Other: Science Direct PDF
  • Journal Article Greytak SR, Callard GV. Cloning of three estrogen receptors (ER) from killifish (Fundulus heteroclitus): differences in populations from polluted and reference environments. General and Comparative Endocrinology 2007;150(1):174-188. R831301 (2006)
    R831301 (Final)
  • Abstract from PubMed
  • Full-text: Science Direct HTML
  • Abstract: Science Direct
  • Other: Science Direct PDF
  • Journal Article Sawyer SJ, Gerstner KA, Callard GV. Real-time PCR analysis of cytochrome P450 aromatase expression in zebrafish: gene specific tissue distribution, sex differences, developmental programming, and estrogen regulation. General and Comparative Endocrinology 2006;147(2):108-117. R831301 (2004)
    R831301 (2006)
    R831301 (Final)
  • Abstract from PubMed
  • Full-text: Science Direct HTML
  • Abstract: Science Direct Abstract
  • Other: Science Direct PDF
  • Journal Article Tarrant AM, Greytak SR, Callard GV, Hahn ME. Estrogen receptor-related receptors in the killifish Fundulus heteroclitus: diversity, expression, and estrogen responsiveness. Journal of Molecular and Cellular Endocrinology 2006;37(1):105-120. R831301 (2006)
    R831301 (Final)
  • Abstract from PubMed
  • Full-text: JME Full Text HTML
  • Abstract: JME
  • Other: JME PDF
  • Supplemental Keywords:

    water, groundwater, sediments, exposure, risk, risk assessment, effects, health effects, ecological effects, human health, bioavailability, metabolism, dose-response, animal, organism, cellular, enzymes, infants, children, metabolism, sex, cumulative effects, chemicals, toxics, PAHs, PNAs, PCBs, Dioxin, metals, heavy metals, solvents, oxidants, organics, pollution prevention; treatment, environmental chemistry, biology, histology, genetics, analytical, measurement methods,, RFA, Health, Scientific Discipline, PHYSICAL ASPECTS, INTERNATIONAL COOPERATION, ENVIRONMENTAL MANAGEMENT, POLLUTANTS/TOXICS, Environmental Chemistry, Health Risk Assessment, Chemicals, Endocrine Disruptors - Environmental Exposure & Risk, endocrine disruptors, Risk Assessments, Biochemistry, Physical Processes, Children's Health, Environmental Policy, Endocrine Disruptors - Human Health, Risk Assessment, health effects, endocrine disruptor screening program, childhood development, zebrafish embryo gene expression system, age-related differences, endocrine disrupting chemicals, exposure, exposure studies, animal model, screening assay, Human Health Risk Assessment, animal models, gene expression, susceptibility, children's vulnerablity, assessment of exposure, endocrine disrupting chemcials, biogeochemistry, zebrafish embryo gene expression, human health risk

    Progress and Final Reports:
    Original Abstract
    2004 Progress Report
    2005 Progress Report
    Final Report

    Top of Page

    The perspectives, information and conclusions conveyed in research project abstracts, progress reports, final reports, journal abstracts and journal publications convey the viewpoints of the principal investigator and may not represent the views and policies of ORD and EPA. Conclusions drawn by the principal investigators have not been reviewed by the Agency.

    Jump to main content.