Page 308 - The Toxicology of Fishes

288 The Toxicology of Fishes

Effects of ROS on mRNA Expression

The transcription factors mentioned above are involved in numerous signaling pathways, including
pathways related to cell division and differentiation, immunological response, cytokine expression and
inflammatory response, xenobiotic metabolism, and many others (Finkel and Holbrook, 2000; Martindale
and Holbrook, 2002; Sauer et al., 2001); thus, it is not surprising that exposure to high levels of ROS
induces the expression of many genes. A small sampling of the many genes reported to be regulated (in
at least some cases) in eukaryotes by ROS include c-fos and c-myc (cellular protooncogenes), gadd45
and gadd153 (growth arrest and DNA damage genes), γ-glutamyl transpeptidase, GCL (both subunits),
some GST isoforms, some UGT isoforms, epoxide hydrolase, heme oxygenase, MnSOD and ECSOD
but not usually CuZnSOD, metallothionein, GPX1, interleukin 8, cytochrome IV, thioredoxin, and an
aldo-keto reductase (Brady et al., 1997; Burczynski et al., 1999; Crawford, 1999; Dhakshinamoorthy et
al., 2000; Prestera et al., 1993; Schull et al., 1991). The expression of some of these (and other) genes
is regulated by ROS via the antioxidant response element, as discussed below. As indicated by this partial
listing, although many of the genes identified as upregulated by ROS could clearly play important
antioxidant roles, many others are not classical antioxidants. Furthermore, the degree of induction
observed in the classical antioxidants is often relatively low; heme oxygenase, not a classical antioxidant,
appears to be one of the most responsive markers of oxidative stress in cells, as it shows a high degree
of induction as well as specificity for oxidative stress as an inducer (Crawford, 1999). Further confusing
the picture are observations that, in some cases, classical antioxidants are upregulated to a greater degree
by inducers other than typical oxidants; for example, MnSOD is upregulated 20- to 100-fold by cytokines
such as interleukin 1 (IL-1) and tumor necrosis factor (TNF), compared to a usually less than 5-fold
induction by prooxidants such as xanthine oxidase, paraquat, iron, copper, and t-butyl hydroperoxide
(Crawford, 1999; Stralin and Marklund, 1994; Valentine and Nick, 1999; Visner et al., 1990).
The expression of many genes is also downregulated by oxidative stress; among these are many
mitochondrial genes (Crawford 1999; Fujii and Taniguchi, 1999; Morel and Barouki, 1999) and a large
number of nuclear-encoded genes involved in the immune response, cell replication, carbohydrate
metabolism, hormonal responses, phase I xenobiotic metabolism, and other cellular activities (Barouki
and Morel, 2001; Morel and Barouki, 1999; Nebert et al., 2000). As in the case of ROS-mediated gene
induction, these results seem likely to reflect not only adaptive changes directed at decreasing oxidative
damage but also alterations in pathways in which ROS normally play a role as signaling molecules or
simple oxidative damage to cellular macromolecules that mediate gene expression in these pathways.

Antioxidant Response Element (ARE)-Mediated Gene Regulation

The only known molecular mechanism for specifically responding to high levels of ROS in eukaryotes
is the activation of a relatively large number of genes via an enhancer element termed the antioxidant
response element (ARE), or electrophile response element (Dalton et al., 1999; Itoh et al., 1999; Nguyen
et al., 2003b; Rushmore et al., 1991; Wasserman and Fahl, 1997). Many genes have been shown to be
regulated by AREs, and many others are suspected to be ARE regulated, based on the presence of ARE
sequences in the promoter regions of those genes, the inducibility of those genes by chemicals that either
generate or scavenge ROS, and the regulation of those genes by putative ARE transcription factors such
as Nrf1 and Nrf2 (discussed below). A list of genes known or suspected to be regulated by the ARE,
grouped according to the type of evidence currently available in the literature, is presented in Table 6.1.
The identity and role of all of the transcription factors that bind to the ARE are not yet clearly
defined but are an area of very active investigation. Early suggestions that AP-1 was involved turned
out to be inaccurate, except in those cases where the ARE is part of a functional 12-O-tetrade-
canoylphorbol-13-acetate (TPA) response element (TRE) (Dalton et al., 1999), such as the case of
the human QR1 gene (Jaiswal 1994b). Thus, although the ARE bears strong sequence similarity to
the TRE, they are not the same. Furthermore, although there are similarities in terms of the transcription
factors that bind the two sites, they are not identical, and the combinations that bind are usually
different. At this point, the best evidence suggests that ARE transcription factors include one or more
CNC-bZIP proteins (e.g., Nrf1 and especially Nrf2), perhaps as part of a heterodimer with a small