Analysis of Unique and Overlapping Patterns of Gene Expression After Treatment of Zebrafish Embryos with Estradiol and/or Dioxin

To increase the utility of the zebrafish gene expression bioassay for assessing EDC effects on multiple gene targets and tissue types, and to expand our understanding of genetic overlap between estrogen receptor (ER) and arylhydrocarbon receptor (AhR) mediated signaling pathways, we conducted microarray analysis of zebrafish embryos exposed to estradiol and dioxin alone or in combination. Of >16,000 probe sets on the array, 34 were regulated by estradiol (E2), 86 by dioxin (TCDD) and 109 by E2+TCDD (as chosen by >2-fold change and p<0.1). Of 62 genes selected for verification by QPCR, 14 genes, or 22% were reproducibly up- or down-regulated, offering potential additional target genes for screening of estrogen- and dioxin-like EDC. The majority of these successful hits, 11, were TCDD-responsive. In addition, all of the target genes routinely evaluated in this laboratory (AroB, Vtg1, and Esr1: E2-responsive; Cyp1a: TCDD- responsive; Vtg1: E2+TCDD-responsive) were verified with these arrays, testifying to the power of microarray analysis in finding reliable responsive genes. However, over two-thirds of the novel up- or down- regulated probes were not annotated as zebrafish genes, and many of the identified genes were changed only 2- to 3-fold, an effect often not reproduced by QPCR. Additional responsive genes were identified for each treatment condition, and while low levels of expression and low magnitude fold changes make those for E2 responsiveness or interaction between E2 and TCDD response unlikely to serve as robust biomarkers, their response to EDCs may assist in understanding the regulation of existing biomarkers and zebrafish endocrine systems as a whole. In addition to a source for potential EDC screening biomarkers, these studies provide an entry point to further study physiological effects of ER and AHR ligands and cross-talk between these signaling pathways, in a physiologically relevant in vivo model.

12