Page 272 - The Toxicology of Fishes

252 The Toxicology of Fishes

A detailed description of these can be found elsewhere (Ankley et al., 2006; Cossins and Crawford,
2005; Denslow et al., 2005; Ju et al., 2007; Larkin et al., 2003b); here, we briefly describe some of these
approaches and discuss their application to fish toxicology.
Several polymerase chain reaction (PCR)-based methods are available to assess differential gene
expression—for example, between control and chemically treated fish. Methods such as differential
display PCR (ddPCR), suppressive subtractive hybridization (SSH), and representational difference
analysis (RDA) are considered unbiased in that they involve no a priori selection of target genes and
therefore can be used for gene discovery. These methods can identify genes that are either induced
(upregulated) or repressed (downregulated), but all three have high rates of false positives; thus, genes
identified as differentially expressed by these methods must be confirmed by more robust assays such
as real-time reverse transcription PCR (RT-PCR). In some cases, genes identified by ddPCR, SSH, or
RDA are used to construct a macroarray or microarray for subsequent use in evaluating gene expression
in a larger number of samples (see below). Both ddPCR and SSH have been used to reveal differential
gene expression in fish exposed to toxicants (Table 5.3).
Other methods for unbiased discovery of differentially expressed genes involve high-throughput
analysis of transcript abundances in two different samples. Two powerful techniques are serial analysis
of gene expression (SAGE) (Velculescu et al., 1995) and massively parallel signature sequencing (MPSS)
(Brenner et al., 2000), both of which provide short sequence tags of 20 to 21 bp that are usually unique
and can be mapped to genome sequences to determine the genes from which they came. Both SAGE
and MPSS are quantitative in that tag abundances (the number of times each tag appears) are directly
related to transcript abundances in the original samples. Although there is one report of SAGE applied
to fish (Knoll-Gellida et al., 2006), neither SAGE nor MPSS has yet been used in the context of fish
toxicology. In addition to SAGE and MPSS, the recently developed 454 parallel sequencing technology
(Emrich et al., 2007; Margulies et al., 2005; Sogin et al., 2006) is likely to be even more powerful for
transcriptional profiling in a variety of applications, including fish toxicology.
The first use of microarrays (DNA chips) in fish was by Gracey et al. (2001), who created custom
cDNA microarrays to measure the transcriptional response of the goby (Gillichthys mirabilis) to hypoxia.
Microarrays (cDNA and oligonucleotide) and macroarrays are now widely used in fish biology, including
toxicology. Most of the available microarray resources are targeted to zebrafish (Handley-Goldstone et
al., 2005; Linney et al., 2004a; Mathavan et al., 2005; Ton et al., 2002), salmonids (Rise et al., 2004;
von Schalburg et al., 2005; Vuori et al., 2006), flounder (Williams et al., 2003), carp (Cossins et al.,
2006; Gracey et al., 2004), or Fundulus (Oleksiak et al., 2001, 2002).
Several recent reports illustrate the power of microarray-based transcriptional profiling in fish to
provide insight into mechanisms of toxicity or to identify candidate biomarkers of exposure or effect
(Table 5.3). In one study, the mechanism of valproic acid (VPA) teratogenesis was investigated in
zebrafish embryos. Gene expression profiles after VPA exposure were similar to those observed after
exposure to inhibitors of histone deacetylase (HDAC), suggesting that HDAC inhibition plays a role in
VPA teratogenesis (Gurvich et al., 2005). Several groups have used microarrays to investigate the effects
of TCDD. Handley-Goldstone et al. (2005) found that CYP1A was the gene most strongly induced in
whole zebrafish embryos exposed to TCDD early in development, confirming the dominance of this
widely studied response to TCDD and other AhR agonists. More interestingly, these authors also
measured altered expression of genes encoding components of cardiac muscle sarcomeres, including
myosin and troponin T2; these changes suggest an explanation for cardiomyopathy seen in fish and other
vertebrates (Handley-Goldstone et al., 2005). Altered gene expression has also been measured directly
in hearts of larval zebrafish exposed to TCDD, demonstrating distinct responses in heart as compared
to the rest of the larvae (Carney et al., 2006). Among the changes in cardiac gene expression occurring
prior to signs of cardiovascular toxicity were increases in genes encoding xenobiotic-metabolizing
enzymes (CYP1A, CYP1B1, CYP1C1, sulfotransferase) and those involved in cell signaling. Organ-
specific changes in gene expression also were seen in medaka exposed to TCDD (Volz et al., 2005,
2006). Dramatic differences in the direction of change were seen between liver (changes dominated by
induction) and testis (many genes repressed). This study also demonstrated how gene expression profiling
can be fruitfully combined with histopathological analysis, an example of “phenotypic anchoring” of
microarray data (see also Luo et al., 2005; Moggs, 2005; Moggs et al., 2004; Paules, 2003). Gene