Page 555 - The Toxicology of Fishes

Chemical Carcinogenesis in Fishes 535

Washington. Hepatocellular and cholangiocellular carcinomas were found in 27% of 75 fish examined.
Sediments from the contaminated region contained particularly high concentrations of aromatic hydro-
carbons and a variety of nitrogen-containing aromatic compounds. Much lower concentrations of aro-
matic hydrocarbons were found in the reference site and fish collected there were free of neoplasms.
“There are parallel tendencies for changes in disease prevalence and changes in pollution level over
a long period of time” and “A reduction in the level of pollution results in a decline in the prevalence
of the disease” are two closely related requirements. Both were met in analyses by Baumann and
Harshbarger (1995). After a coal coking facility was closed in 1983, polycyclic aromatic hydrocarbons
(PAHs) in sediment and in brown bullhead catfish tissues showed marked decline when measured in
1987 (Baumann and Harshbarger, 1995). Over the same time span, in the 3- to 4-year old resident catfish,
liver cancer prevalence declined to about one quarter of the 1982 frequency.
“The disease occurs regularly in heavily polluted waters but is seldom if ever encountered in [clean]
waters.” Many field investigations have met this criterion (Malins et al., 1985, 1987; Pinkney et al.,
2001; Vogelbein et al., 1990). The study by Pinkney et al. (2001) was particularly relevant as the epizootic
was not only of liver neoplasia but pancreatic neoplasms as well. This finding had never before been
reported. “Field data are confirmed by long-term experiments.” Extracts of contaminated sediment from
sites where resident wild fish showed elevated tumor prevalence produced tumors of the same type in
exposed laboratory surrogate fish (Fabacher et al., 1991). “Disease prevalence is related to body burdens
of pollutants.” The extensive field investigations from Alaska to southern California reported in Myers
et al. (1994) demonstrated this relationship. This was accomplished directly by establishing body burdens
for some contaminants and indirectly by quantifying fluorescence absorbing compounds in bile for others
(e.g., PAHs). Pinkney et al. (2001) provide an excellent example of this in their study of reference and
contaminated sites in the tidal Potomac River. According to Peters et al. (1987), the probability that
pollution is a cause of disease increases when two or more of these criteria are met. From the above,
we can conclude that there is a strong, site-specific association between environmental contamination
and liver cancer (neoplasia) in fishes.

Multistage Carcinogenesis

Carcinogenesis was believed to arise by a multistage process involving initiation, promotion, and
progression (Figure 12.1). This has been a long-standing paradigm that is now shifting, yet much of the
fish-related literature is still based on it. More recent literature in the medical arena shies away from the
traditional paradigm and refers instead to agents as either genotoxic or nongenotoxic and are discussed
later. It was in rodent experimental systems that this stepwise process was first demonstrated and used
to investigate numerous compounds. It is also important to remember that the initiation–promotion model
was first described in skin and subsequently substantiated in liver, colon, lung, prostate, and mammary
gland of rodents. For our considerations in this chapter on carcinogenesis in fishes, only skin, liver, and
perhaps colon apply; in the case of colon, fish may lack a distal portion of the gut that is strictly analogous.
Given the lack of demonstration of initiation and promotion in fish skin neoplastic development, we
need only consider liver. As will be shown below, liver carcinogenesis in trout and medaka has included
initiation and promotion. Progression has received little attention overall, but at least one report (Okihiro
and Hinton, 1999) established the occurrence of this stage in medaka.
As mentioned previously, the initiation–promotion–progression hypothesis is somewhat dated. Other
early general rules, such as “All carcinogens are mutagens” and “All mutagens are initiators,” are also
now known to be false. There remain compounds and conditions that alter gene expression without
causing mutations and may promote (by either increasing tumor yield or reducing the time between
exposure and appearance of tumor). Also, from studies that addressed the proportion of known carcin-
ogens that were detected by Ames assay (about 60%), it is apparent that there are nongenotoxic
carcinogens. Some, therefore, prefer to reject the old paradigm and instead discuss carcinogens more
broadly as either genotoxic or nongenotoxic. Given that the majority of publications relating to fish are
based on this paradigm, however, we shall discuss this topic within that original framework.