Attempts to develop a mechanistic understanding of the effects of environmental estrogens on fish are increasingly conducted at the level of gene expression. Appropriate application of real-time PCR in such studies requires the use of a stably expressed 'housekeeping' gene as an internal control to normalize for differences in the amount of starting template between samples.

We sought to identify appropriate genes for use as internal controls in experimental treatments with estrogen by analyzing the expression of eight functionally distinct 'housekeeping' genes (18S ribosomal RNA [18S rRNA], ribosomal protein l8 [rpl8], elongation factor 1 alpha [ef1a], glucose-6-phosphate dehydrogenase [g6pd], beta actin [bactin], glyceraldehyde-3-phosphate dehydrogenase [gapdh], hypoxanthine phosphoribosyltransferase 1 [hprt1], and tata box binding protein [tbp]) following exposure to the environmental estrogen, 17α-ethinylestradiol (EE₂), in the fathead minnow (Pimephales promelas). Exposure to 10 ng/L EE₂ for 21 days down-regulated the expression of ef1a, g6pd, bactin and gapdh in the liver, and bactin and gapdh in the gonad. Some of these effects were gender-specific, with bactin in the liver and gapdh in the gonad down-regulated by EE₂ in males only. Furthermore, when ef1a, g6pd, bactin or gapdh were used for normalization, the hepatic expression of two genes of interest, vitellogenin (vtg) and cytochrome P450 1A (cyp1a) following exposure to EE₂ was overestimated.

Based on the data presented, we recommend 18S rRNA, rpl8, hprt1 and/or tbp, but not ef1a, g6pd, bactin and/or gapdh, as likely appropriate internal controls in real-time PCR studies of estrogens effects in fish. Our studies show that pre-validation of control genes considering the scope and nature of the experiments to be performed, including both gender and tissue type, is critical for accurate assessments of the effects of environmental estrogens on gene expression in fish.