Hepatic transcriptional responses of zebrafish to chronic, low dose pharmaceutical exposures

Our study makes use of zebrafish and has determined hepatic transcriptional changes after exposure to four single pharmaceuticals, a pharmaceutical mixture (MIX), and wastewater effluent (WWE) exposures. The pharmaceuticals chosen include acetaminophen, carbamazepine, gemfibrozil, and venlafaxine. We have performed chronic (6 week), low concentration (0.5 and 10 μgL-1 or 5 and 25% effluent) exposure of male and female fish and previously determined a range of effects (reproductive, histological, hormonal) after exposure to single compounds, a mixture of these compounds or wastewater effluent. Herein we determine by microarray the hepatic transcriptional responses of male and female zebrafish from the same exposures. Hepatic transcriptional responses in zebrafish to four common, environmentally relevant pharmaceuticals (acetaminophen, carbamazepine, gemfibrozil, venlafaxine), a mixture of equal concentrations of the four pharmaceuticals (MIX), and dilutions of wastewater effluent (WWE) were determined after a 6 week chronic exposure to a low (0.5 μg L-1 or 5% effluent) or high (10 μg L-1 or 10% effluent) concentration, including water or DMSO controls. Adult wild-type zebrafish were housed in aquaria at a density of 4 fish per liter in 1:1 sex ratio, with triplicate tanks for each treatment group housing 50 adult zebrafish each. Tanks were dosed every three days for six weeks after a 90% system water change-out. After 6 weeks of exposure, zebrafish were sacrificed, weighed and tissues (GI tract with livers) were fixed in RNA Later and stored at -80°C. For RNA extraction, livers were stripped from the GI tract, livers from 10-20 fish per gender were pooled to a maximum 100 mg of tissue, and total RNA extracted. RNA was bioanalyzed and then labeled with Cy3 and 1.65 ug Cy3 labeled cDNA hybridized to a customized Agilent zebrafish single-channel 4 x 44K V2 microarray by the University Health Network Microarray Centre (Toronto, Canada). Raw array data obtained from the University Health Network Microarray Centre were extracted using Agilent's feature extraction software using background detrending (spatial and multiplicative). Prior to normalization, Cy3 values below 5 were set to 5. The data were then normalized using the non-linear scaling method based on rank invariant probes. After normalization but before statistical analyses, probes not significantly above background in all microarrays were removed. None of the probes were saturated for Cy3 signal on any microarray, so no further filtering was applied. Statistical tests were performed using MeV. All data was log transformed and median centered for each probe. For each experiment, a two-factor ANOVA was run for sex, treatment, and their interaction with p-value based on 1000 permutations of the data and alpha of 0.01. For exposures with two treatments (i.e. VEN-ACE or GEM-MIX), there were five treatment groups (Control, Low 1, High 1, Low 2, High 2); for exposures with single compounds or effluent (i.e. CBZ or WWE), there were three treatment groups (Control, Low, High). Probes found significant for treatment and/or interaction from the two-factor ANOVA, were secondarily examined using Rank Product (RP) analysis to provide lists of probes up- and down-regulated between control, low dose, and high dose for each treatment separately for males and females (VEN, ACE, CBZ, GEM, MIX, WWE). For this, a two-class unpaired RP analysis was performed using 1000 permutation of the data and alpha of 0.01 with a false discovery rate not exceeding 10%.