Changing environmental conditions provide selective pressures that drive adaptive changes in animal populations [1, 2]. Among the many environmental stressors that drive adaptation, the presence of toxic chemicals--naturally derived or anthropogenic--can exert strong effects, in part through their ability to affect the survival of sensitive early developmental stages. Although the acute effects of chemicals are widely studied and adaptation to acute effects of pesticides in invertebrates such as insects is well known, the impact of long-term, multi-generational exposure to chemicals on naturally occurring populations of vertebrate animals is not well understood.
One species that has emerged as a valuable model for investigating evolutionary adaptations to chemical exposure is the Atlantic killifish, Fundulus heteroclitus. This estuarine teleost has a long history as a subject for research in environmental biology [3–5], and studies over the past two decades have identified several populations of this species that have evolved tolerance or resistance to toxic chemicals [6, 7]. Prominent among these are killifish populations that have developed resistance to toxic polynuclear aromatic hydrocarbons (PAHs) and halogenated aromatic hydrocarbons (HAHs) such as 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD) and planar polychlorinated biphenyls (PCBs) .
Evolved resistance of F. heteroclitus to PAHs or HAHs, first noted in Newark, NJ [9, 10], has also been described in killifish from the Elizabeth River, VA [11–13], New Bedford Harbor, MA [14, 15], and several more moderately contaminated sites in New England [8, 16]. At all of these sites, killifish embryos, larvae, and adults are much less sensitive to acute toxicity of HAHs and PAHs as compared to fish from less contaminated reference sites. They also exhibit reduced sensitivity to the induction of cytochrome P450 1A (CYP1A), a widely used marker of altered gene expression in response to these compounds. In fish, mammals, and other vertebrate animals, both the induction of CYP1A and the toxic effects of PAHs and HAHs are controlled by the aryl hydrocarbon receptor (AHR), a ligand-activated, bHLH-PAS protein [17–19]. Thus, the results of these studies on PAH/HAH-resistant killifish suggest that certain AHR-regulated genes--or possibly the AHR pathway generally--have become desensitized in the affected populations.
The New Bedford Harbor (NBH) killifish population, the focus of this study, is resistant to the effects of a variety of AHR ligands, including PAHs, β-naphthoflavone, non-ortho-substituted PCBs, 2,3,7,8-tetrachlorodibenzofuran, and TCDD. The resistance to induction of CYP1A is present at all life stages and in all tissues, is heritable, and occurs at the level of mRNA, suggesting a transcriptional effect [8, 14, 15, 20, 21]. Recent findings suggest that the resistant phenotype is the result of genetic rather than epigenetic mechanisms [22–25].
Previous studies of reduced sensitivity to altered gene expression in PAH/HAH-resistant killifish, including those in NBH, have focused almost exclusively on induction of CYP1A, as measured by changes in CYP1A mRNA, protein, or activity. The role of CYP1A in the mechanism of PAH and HAH embryotoxicity in fish is not yet clear and is likely to be complex. Some studies have demonstrated that CYP1A is not involved in the mechanism of TCDD toxicity in fish , whereas this enzyme appears to play a protective role in PAH toxicity [27–30]. An important unanswered question is whether the reduced sensitivity to gene induction in affected populations is specific to CYP1A or a small subset of AHR target genes or, alternatively, reflects a global (i.e. genome-wide) reduction in responsiveness to all AHR-mediated changes in gene expression. The objective of the present study was to address this question using microarray-based gene expression profiling.
The development of microarray resources for F. heteroclitus [31, 32] has facilitated the comparison of gene expression profiles in individuals from HAH-sensitive and resistant populations. Studies using a 384-gene metabolic array to survey basal gene expression in brain  and liver  of untreated adult fish found a small number of changes in fish from polluted sites that could represent either induced or adaptive (evolved) changes. More recently, a second generation array containing 7,000 spots (genes) was generated using cDNA libraries from all 40 developmental stages of F. heteroclitus . At the only developmental stage investigated (stage 31, corresponding to approximately ~6 days post-fertilization (dpf)), there were few differences in basal gene expression profiles between polluted and reference sites , suggesting that the most important differences in gene expression may only occur under conditions of chemical exposure.
In the present study, we used custom microarrays and deep transcriptome sequencing (RNA-Seq using 454 Life Sciences technology) to examine gene expression profiles in control and PCB-treated killifish embryos and early larvae spawned by fish from NBH (PCB-contaminated site) and Scorton Creek, MA (SC; reference site ). We compared expression profiles in embryos and early larvae that had been exposed to vehicle (DMSO) or a potent non-ortho-PCB (PCB-126; 3,3',4,4',5-pentachlorobiphenyl) early in development and then sampled at 5, 10, or 15 dpf. The results suggest that resistance to altered gene expression in NBH embryos represents a genome-wide loss of sensitivity.