Page 880 - The Toxicology of Fishes

860 The Toxicology of Fishes

Tillitt, 1999). The extracts were able to induce dioxin-like symptoms of toxicity that were predicted by
the concentrations of PCBs, PCDDs, and PCDFs measured in the extract. The dose–response relationships
developed in these experiments produced LD values based on the TEF/TEQ approach that matched
50
previously measured values in lake trout and rainbow trout. The complex mixture of HAHs taken from
lake trout from Lake Michigan was able to induce the same endpoints of toxicity observed in the feral
fry, and this relationship was predicted by the concentrations of HAHs and an additive model of toxicity.
This was strongly affirmative evidence which meets the criterion of predictive performance.

Coherence
The coherence criterion refers to the coherence of information on the association with known facts about
life history, biology, and toxicology. Is the information about the association consistent with this knowl-
edge? Is the relationship biologically plausible? Does the information conflict in some way with theo-
retical models? Can the suspected agent cause the symptoms and effect? This criterion is generally
evaluated from the basis of coherence with theoretical models, biological knowledge, and toxicological
information about the association. We have seen through the detailed studies in zebrafish the mechanisms
whereby dioxin and dioxin-like chemicals (e.g., HAHs) produce their characteristic signs of blue sac
syndrome in developing embryos and larvae. Although the mechanism of dioxin toxicity in lake trout
has not been elucidated to this same extent as it has in zebrafish, the endpoints of HAH toxicity are the
same in both these species and a similar pathway through the AhR occurs in salmonines, including lake
trout. The sensitivity of lake trout toward dioxin has been established (Spitsbergen et al., 1991; Walker
et al., 1991) and indeed is the most sensitive species of fish tested to this point (Elonen et al., 1998).
Complementary to the sensitivity of lake trout toward dioxin and other HAHs, the exposure of lake trout
to HAHs has also been established in the lower Great Lakes (Baumann and Whittle, 1988; DeVault et
al., 1989; Hickey et al., 2006). Certainly, confounding exposures to other organochlorines, most notably
DDT and its metabolites, also occurred during this time (Baumann and Whittle, 1988; Hickey et al.,
2006). DDT can cause a swim-up syndrome in salmonines (Macek, 1968) and may well have contributed
to the lack of recruitment observed in lake trout from the lower Great Lakes; however, concentrations
of DDT in lake trout were not found to be above toxicity thresholds on a consistent basis. Based on the
exposures of lake trout to HAHs during the last half of the 20th century and the sensitivity of this species
toward HAHs, it is certainly plausible that an association between HAHs in the Great Lakes and the
lack of recruitment in lake trout occurred over that time period.
Biological field data is also coherent with the premise of an association of HAHs and the lack of
recruitment of lake trout in the lower Great Lakes during this same period. Pathologies of dioxin-like
toxicity were observed in adult lake trout during this time, as well as in the developing fry. Fry from
the lower Great Lakes developed blue sac syndrome and suffered from swim-up mortalities that were
consistent with HAH-induced effects. During this same period, lake trout from Lake Superior had
significantly smaller concentrations of HAHs, were not observed to have symptoms of HAH toxicity,
and were not suffering from a lack of recruitment into the population. Thus, the biological information
on lake trout from the Great Lakes supported the plausibility of this association.
Finally, the toxicological information, particularly our knowledge regarding dose–response relation-
ships for HAHs in lake trout, supports a causal relationship between HAHs and population-level effects
in lake trout from the Great Lakes over the last half of the 20th century. Correlative studies comparing
concentrations of chemicals in salmonine eggs taken from the Great Lakes with mortality have reported
conflicting results. A number of studies found no relationship between HAH contaminants in eggs and
early-life-stage mortality of the resulting fry (Fitzsimons, 1995; Williams and Giesy, 1992; Zint et al.,
1995), while others have observed a correlation between these factors (Ankley et al., 1991; Mac and
Schwartz, 1992; Mac et al., 1993). This type of experimental design, however, is fraught with problems
and might not allow clear relationships to be resolved. Many factors that may well contribute to mortality
in the developing fry may not be measured; for example, the quality of the accuracy of chemical
measurement of HAHs has changed and improved over the past two decades, so early studies may not
have quantified the HAHs or other chemicals accurately. Additionally, most of the correlative studies
did not measure all of the HAHs known to be present, and many of these studies were conducted at a
time when the concentrations of HAHs in salmonines from the Great Lakes were reduced and near