Carcinogens and cancers in freshwater fishes.

Epizootics of neoplasms in freshwater fish species are considered in relation to circumstantial and experimental evidence that suggest that some epizootics of neoplasia of hepatocellular, cholangiocellular, epidermal, and oral epithelial origin may be causally related to contaminant exposure. Although there is concern for the safety of consuming fish affected with neoplasms, this concern may be misdirected as direct transmission of cancer by ingesting cancerous tissue would seem unlikely. Of greater concern is the matter of toxic and cancer-causing chemicals present in edible fish that exhibit neoplasia as a symptom of past exposure via residence in a polluted waterway. There is ample evidence to suggest that contaminant chemicals ingested via contaminated Great Lakes fish may already be affecting both human and ecosystem health, but these effects are subtle and may require new approaches to the study of the affected systems.


Introduction
Fish neoplasia has recently received considerable emphasis as an early warning system for the detection of carcinogens in the aquatic environment (1,2). This idea is supported mostly by circumstantial evidence (1)(2)(3)(4)(5). There is, however, a growing body offield-related, laboratory-generated data, which also supports a link between certain kinds of aquatic pollution and cancer in fish (6)(7)(8)(9)(10). The present paper will not deal as much with the linkage of waterborne carcinogens as a causative agent of fish neoplasia, as with identifying the kinds of freshwater species affected with epizootic neoplasia, the anatomical sites and histologic types of neoplasms, prevalence data for specific geographic sites, and a review ofinformation concerning levels ofcancer-causing chemicals in these species. No attention will be given to detailed descriptions of the histologic picture or classifications of tumors in fish. Neoplasms in fish can vary widely in their degree ofcellular atypia and biologic behavior. In the present paper, it is convenient to discuss these growths under the inclusive term "neoplasm,," recognizing that such variation exists.
While we will not limit our discussion to Great Lakes fisheries, some emphasis will be given to fish cancer and contaminant problems ofthe Great Lakes and tributary waters. The title ofthe present conference emphasizes the potential ofaquatic food resources as conveyers of chemical carcinogens to the human population. The conference title carries the implication that only those species used for commercial fisheries would be of interest. Our topic of discussion involves freshwater fish populations exhibiting neoplasia. Although most of the species involved will not be subject to commercial fishing, all are consumed to some extent. The public is concerned about the safety ofconsuming cancerous fishper se. In a sense, this concern is misdirected. After all, it is possible to induce liver cancer in rainbow trout by exposing embryos to as little as 1 ng ofaflatoxin B1, yet in the adult tumor-bearing trout, the presence ofthe chemical that provoked the stimulus will not be detectable. It is conceded that the risk of transmitting cancer by ingesting cancerous fish tissue is an unknown. However, since most freshwater fish is cooked before being consumed, it is expected that risks ofdirect transmission ofan oncogenic DNA fragment or virus would be unlikely. Rather, our concern here is for the ingestion ofchemical contaminants in edible fish tissue. Particularly, this is ofconcern where cancerous fish are taken from heavily polluted water where they are chronically exposed to a wide range ofchemical contaminants.

Fish Species Affected with Neoplasms
Of ffie more than 200 freshwater fish species there are probably no more than 50 freshwater fish species commonly kept by anglers. There are more than 80 freshwater fish species seldom keptby anglers. These include various lampreys, gars, minnows, killifishes, madtoms (small catfishes ofthe genus Schilbeodes), sculpins, darte, and several members ofthe herring family that are found in freshwater, e.g., gizzard shad. The freshwater locales and species where epizootics have been discovered do not usually foster or support major commercial fisheries. On the other hand, all ofthe species involved are edible and are consumed and fished commercially to some extent in other locations. Thus, there is always the possibility of migration of fish from heavily contaminated areas to cleaner areas wheredty may conceivably be taken in either a commercial or sport fishery.
Various kinds of neoplasms, including hepatic neoplasms, have been documented to occur in freshwater teleosts (Table 1),  (20,21) inland lakes, New York Dermal melanoma Sudbury River, Massachusetts,  (20,29) 'Neoplasms exhibit a range of invasive potential from noninvasive to locally extensive invation to an occasional neoplasm exhibiting metastatic growth. bHistorical data (1941); 166 tumor-bearing fish were studied (160/166 had oral tumors), indicating the tumors must have been common in these populations. Fish (40/166) also had epidermal tumors located at sites other than the mouth (e.g., barbels, head, body).
CCell of origin is uncertain. and the epizootiology ofthese diseases has recently been reviewed by Harshbarger and Clark (I1). These authors have concluded on the basis ofconverging evidence from numerous investigations involving both field and laboratory data that liver neoplasms, and possibly epidermal neoplasms, in certain bottom-feeding species are most specifically correlated to exposure to chemical contminants; e.g., historical (epizootic liver neoplasms found only after 1940); experimental carcinogen exposure (ofthe 30-odd carcinogens tested, all produced some liver neoplasms); physiological/biochemical evidence (metabolizing enzymes in the liver lead to DNA adducts); experimental carcinogenesis studies with contaminated sediment (skin painting, feeding of contaminated benthos to flounder, trout fty injections). The elegant studies ofhepatic neoplasms in English sole (12)(13)(14)(15) and more recently, investigations of hepatic neoplasms in winter flounder (16,17) are paticularly supportive ofthe idea that neoplasms in some fish species are due to environmental contaminant exposure. Currently, there are no studies of environmental carcinogenesis in any freshwater species that are comparable to the extensive interdisciplinary investigtions ofthe relationships between aromatic hydrocarbon pollution and hepatic neoplasia in marine flatfishes. On the other hand, it is likely that discoveries relevant to the relationships between aromatic hydrocarbon pollution and hepaic neoplasia in marine flat-fishes may well be applicable to fresh water environments having the same or similar sediment contminant profiles.
Among the various feshwater species listed in lible 1, brown bulheads and white suckers appear circumntially to be highly sensitive sentinel animals. Both fish species are bottom feeders and, at one or more locations, have exhibited oral, epidermal, and hepatic neoplasms.
White suckers (Catastomus commersoni) from some Great Lakes sites, in particular, western Lake Ontario, have exhibited high frequencies oforal papillomas. On the basis ofa report by Sonstegard (27) of C-type virus particles observed by electron microscopy and tissue fractions contining reverse transcriptase activity, this neoplasm was suspected to have a viral etiology. No viruses were detcted in subsequent ultrastructures studies conducted by Smith et al. (30). The development ofthese neoplasms may also be influenced by pollution, since the frequency of occurrence increases dramatically from a low ofaround 6% in the eastern basin, to a high of 39% in the western part of the basin near the Oakville-Burlington, Ontario, Canada, area (27).
Hepatic neoplasms of both cholangiocellular and hepatocellular origin also occur in white suckers. Dawe et al. (31) indicated that 3 of 12 white suckers from Deep Creek Lake, Maryland, had cholangiocellular neoplasms. It should be noted that only 5 ofthe fish examined by Dawe et al. were large adults, i.e., of a size/age range where frank neoplasms are normally observable. In other words, since cancer in fish, as in mmas, is an age-related disease, tumor prevalence comparisons must be on an age-adjusted basis. Thus, the incidence of this condition in Deep Creek Lake (3 of5 adults) could have been quite high.
Hepatic neoplasms of both bile ductular and hepatocellular origin have been found to occur in the white sucker populations inhabiting the industialized western basin ofLake Ontario, i.e., the liver neoplasms occurred in the same populations exhibiting high frequencies oforal papillomas (29). In these fish, bile duct neoplasms were present in 6% and hepatocellular neoplasms were present in 2 to 3 % of suckers from the polluted Lake Ontario sites (n = 456) versus none in suckers fromthe control sites (Lake Simcoe and Lake Huron, n = 108). Both types ofhepatic neoplasms have also been observed in white suckers from polluted urban-industrial sites on eastern Lake Erie and the upper Niagara River, in the vicinity ofBuffalo, New York (20. In addition, several cases of hepatic, gastrointestinal, and pancreatic neoplasms have been observed in fish ofa related genus, Moxosta sp., e.g., red horse suckers (20).
In brown bullheads, hepatic neoplasia, including hepatocellular and cholangiocellular types, have been observed to occur in bullheads from Deep Creek Iake, Maryland (31), the Fox River, IL (32), industrialized urban aquatic areasofeastern Lake Erie inthevicinity ofBuffalo, New York (20), theBlackand Cuyahoga Rivers near Cleveland, Ohio (33), and the St. Mary's River, Michigan (33). In both the Blackand Cuyahoga, aswellas the urban-industrial sites nearBuffilo, sediments areknowntobe heavily contaminated with aromatic hydrocarbons ( Table 2). Although circumstantial evidence ofhigh sediment loadings of aromatic hydrocarbons to thepresenceofneoplasms is suggestive of a role for these chemicals, hepatic neoplasms have also been found to occur in brown bullhead populations from four of six ostensibly nonpolluted inland lakes and reservoirs sampled in the stateofNewYork(M. Wolfe, unpublisheddata). Althoughthese lakes are reputedto be nonpolluted, theremaybe naurlly occurring carcinogens or as yetunrecognized sources ofcarcinogenic pollutants into these lakes. Also, it should be recognized that neoplasms that present a histologically similar picture can have differentcauses, e.g., viralskinpapillomasoflaboratory rodents look the same as chemically induced skin papillomas. Recent work by Kurelec et al. (34), using 32P-postlabeling methodology, indicates that generally similar quantities of hydrophobic DNA adducts were present in hepatic tissue from cyprinid fishes collected from both polluted (Sava River) and nonpolluted (Korana River) river environments in Yugoslavia. It should be noted that the levels ofDNA adducts found in these nontumorous cyprinid fishes from two Yugoslavian rivers were much lower than levels detected in fish from polluted U.S. environments (8-10). Perhaps even more significantly, in the experiments conducted in the U.S., fish exposed in the laboratory to extracts of sediments containing high concentrations of aromatic hdocarbons developed hepatic DNA adduct patterns that were virtually identical to those produced by 32p_ postlabeling analysis ofhepatic DNA from wild, tumor-bearing fish collected from the same polluted environments used as the source of sediments (10; A. E. Maccubin and J. J. Black, unpublished data). Epizootics ofepidermal and oral neoplasms also occur in brown bullheads. To an extent, these epidermal and oral neoplasms tend to co-occur in sites where liver neoplasia has also been found.
An interesting epizootic ofboth hepatic and dermal neoplasm was found in the case ofsauger (Stizostedion canadense) and to a lesser extent, walleyes (S. utreum), from Torch Lake and the connecting Keweenaw Waterway, Houghton County, Michigan (24). Torch Iake is a large, deep lake (surface area 1077 ha; mean depth 17 m; maximum depth 35 m) located in Michigan's upper peninsula. The area is rich in copper deposits, and mines and finig of coer were a major part ofthe area's economy. Since the turn ofthe century, copper ore, extracted from mines in the area, was crushed at mills located along the shore of this lake.
iilings dumiped in the lake were later dredged from the lake and reprocessed when the price of copper rose enough to assure economic feasibility. The reprocessing continued until 1968, when the mills closed. Although no known chemical carcinogens have been found in high concentrations, the sediments contain high concentrations of some metals, including Cr, Ni, Zn, and especially copper. The copper ore reprocessing used a flotation process (in part) in which various combinations of wood creosote, pine tar, and other organic chemicals such as xnthates (dialkyl-dithiocarbonates) were used as frothing agents for the flotation of the finely milled copper particles.
Sauger from Torch Lake have nearly a 100% incidence of hepatocellular neoplasm [neoplasms present in 23 of23 sauger examined in 1979-80, and in 54 of55 large adult sauger sampled in 1983 (35)]. Walleyes eamin fromthe lbrch Lake in 1979-80 also exhibited hepatic neoplasms (hepatic neoplasms were present in 3 of 11 speimens.) Sampling ofte walleye population in 1983 indicated very few hepatic neoplasms were evident, suggestive of a decreased prevalence between 1979 and 1983. In 1983, only 4 small hepatic neoplasms were found among 106 walleyes sampled from the entire Keweenaw Waterway, including 25 walleyes from Torch Lake.
The saugers from these waters were heavily parasitized with srl species oflarval matodes (liver) and nematodes (heart, pericardial cavity, and mesentaries) (36). It is not known to what extent the parasitic injury contributes to the presence ofeither the dermal growths or the hepatic neoplasms, but any strict correlation to either was inapparent. In addition to a total of 66 sauger and 106 walleyes examined from the Torch Lake-Keweenaw system, 86 white suckers, 23 redhorse suckers, 21 brown bullhads, 90yellow perch, 23 smallmuhbass, and 62 northern pike were also necropsied and examined for neoplasms (35). Neoplasms were not evident in any other species ammined. The white suckers, redhorse suckers, and yellow perch are infected by may of the same parasites found in the sauger.  It may be relevant to note that the saugers constitute the only Torch Lake-Keweenaw Waterway fish population with members old enough to have been alive at or near the time that the ore reprocessing and active disbursement oforganic chemical to the lake was ongoing. Regardless of these circumstances, no conclusions have been reached relative to the exact cause of the tumors in these fish.
Dermal neoplasms were also present among 23 of the 55 sauger examined from Torch Lake and 6 of 106 walleyes collected from the entire Keweenaw Waterway. The histologic appearance of these growths was consistent with a diagnosis of dermal (ossifying) fibrosarcoma. Similar growths are relatively common in walleye populations from various other locations (Lake Erie, Detroit River, Tittabwassi River, Lakes in western Canada, etc.). Studies have shown that these neoplasms in walleyes from Lake Oneida, New York, appear to have a retrovirus etiology on the basis of evident virus particles (25), seasonal regression patterns (26), which may be related to temperature effects upon high reverse transcriptase activity in dermal sarcoma tissue (P.R. Bowser, personal communication), and demonstrated transmission of the disease in juvenile walleyes by cell-free extracts (P.R. Bowser, personal communication).

Fish Consumption
Viable freshwater commercial fisheries do exist, especially in the Great Lakes. For example, in 1982, the commercial fishing harvest for the Great Lakes exceeded 116 million pounds. Included in this statistic was approximately 64 million pounds from Lake Erie (37). In the upper portion of the Mississippi River, through the states ofMinnesota, Wisconsin, and Iowa, the commercial catch alone totals nearly 3 million pounds per year (38). Approximately 44% of the catch from these waters is composed ofbottom-feeding carp and buffalo. Some species harvested from freshwater (not included in the above catch statistics) are not consumed by humans but are used in the production of animal and pet feeds.
Although commercial fishing remains an important industry in some areas ofthe Great Lakes, more and more states are emphasizing the considerable economic value of sport and recreational fisheries over commercial fishing. In the Ohio and New York waters of Lake Erie, the commercial fishing has recently been eliminated through a buy out of the commercial fleet. As indicated above, the Great Lakes support major recreational and sport fisheries. At least two dozen species of fish are commonly caught by anglers from these waters. Walleye are a prized game and food fish, and stocks ofcatchable-size walleye in the western basin of Lake Erie are estimated to exceed 25 million fish (37). The sport/recreational catch oftrout and Pacific salmon from the Great Lakes (in total, five species) easily exceeds 12 million pounds. The catch oftrout and salmon in Lake Ontario taken by the open water (trolling) fishery alone is in excess of 2 million pounds (39). This is approximately equal to the combined commercial harvest of all species taken from that lake. This figure does not include sizable numbers oftrout and salmon caught by anglers fishing from the lakeshore or the additional large harvest that takes place when these fish ascend tributary streams for spawning. The total of all sport and recreational fishing trips made in the U.S. in 1985 was estimated at 735,400,000; 1 lb of fish per trip yields the same number in pounds (40).
In view of the steadily rising U.S. per capita consumption of seafood (15.4 lbs yearly in 1987), the role of commercial aquaculture can be expected to increase. Freshwater aquaculture is already a rapidly growing and important industry. Recent production figures are appoximately 51 x 106 lbs for trout, 280 x 106 catfish, and 65 x 106 for crayfish (mostly wild captured in Louisisana), and an additional 20 x 106 of miscellaneous species are harvested (41). Disclosure of chemical contamination of fishery products resulting from aquaculture would undoubtably undermine the faith ofthe consumer and would have significant impacts upon this developing industry.

Aquatic Contamination
It is a fact of life that most freshwater fish, especially those taken from the Great Lakes, carry variable quantities of trace contaminants in their tissues. It is well known that contminants, especially chlorinated organics, accumulate in part as a function ofthe lipid content ofthe organism. The highest concentrations are reached in fatty species such as lake trout, which occupy the terminal predatory niches of their respective food chains. Consumable fish are unique in that they constitute the only large group of predatory species consumed by North Americans.
Although a compendium ofall compounds identified in Great Lakes fish tissues is beyond the scope ofthe present discussion, some feeling for the magnitude of this problem may be gained from the following considerations. Studies by Hesselberg and Seelye (42) identified 476 compounds present in Great Lakes fish tissues (Table 3). Ofthese 476 compounds, 29 were isomers of polychlorinated biphenyls (PCBs), 9 were chlorinated pesticides, 15 were chlorinated industrial chemicals, 5 were polycyclic aromatics, and over 150 were oxygen-containing compounds (metabolites?) such as phenols, esters, and/or carboxylic acids. Lake trout from Lake Michigan and walleyes from Lake St. Clair contained the largest number of chemicals (167 and 215 compounds, respectively). Only 8 compounds were identified from hatchery-reared lake trout, used as controls (42). Currently, fish from Lake Ontario (Table 4) are considered to carry the highest amounts of many of these anthropogenic compounds (43). The dramatic concentrations of chlorinated hydrocarbons in Great Lakes fish result from only trace detectable amounts found in water. For some conaminants such as PCBs in lake trout and other salmonids, the gradient between concentrations found in water and the fish may exceed 1 million. Thus, to receive an equivalent dose of PCBs contained in a single meal of Great Lakes trout or salmon (assuming fish contains 1.5 mg PCB, water contains 3 ng/L, and assuming water is consumed at a rate of 5 L/day), a human being would have to drink Great Lakes water for more than 200 years. Thus, it is not surprising that several reports have indicated that Great Lakes fish are a major dietary source ofexposure to these stable organochlorines (44,45). A joint U.S. National Research Council-Royal Society ofCanada committee concluded that the human population living in the Great Lakes basin is exposed to and accumulates appreciably more toxic substances than other parts of North America (46). Compounds of concern (mostly polluted harbor areas near major cities) Aerial transport Worst-case scenarios: Lake Ontario, Lake Michigan aOnly eight compounds were identified in clean water controls.
Although the presence of contaminants in the lakes has been well documented, the possible health effects ofthe contaminants upon biota, including man, is less well understood. Studies of fish-eating birds have shown effects ranging from complete reproductive failures in the late 1960s to congenital malformations in the late 1970s to gradual improvements in reproductive success in the 1980s (47). Work by Casterline et al. (48) has indicated that cleaned up organochlorine residues containing fractions derived from edible fish tissues were potent inducers of arylhydrocarbon hydroxylase (measured in TCDD equivalents) when tested in an in vitro bioassy using the H4IIE rat hepatoma cell line. Furthermore, recent work by Tillitt et al. (49), using the same bioassay in a blind study, has shown that not only are these residues accumulated in fish-eating birds from the Great Lakes, but the assay results were correlated to both the rates ofdeformities and colony (egg) mortality.
Consumption rates ofcontaminated fish (mostly salmon) by humans and the transfer of PCBs, a potential causative agent of these effects, has been well documented in serum, cord blood, and breast milk (50,51). Recent studies by Jacobson et al. of human health effects in relation to consumption ofcontaminated Great Lakes trout and salmon, have clearly shown that subtle neurological and growth deficits occurred in a dose-related fashion in exposed infants (52,53). The validity of the putative PCB-contaminated Great Lakes fish consumption-human health effects linkage in these studies was strengthened by independent observations of similar neurological deficits among a group of North Carolina infants that were correlated to the concentrations of PCBs in breast milk, irrespective of fish consumption, as a possible route ofmaternal exposure to these compounds (54). A recent study by Daly et al. (55), in which laboratory rats were fed contaminated Lake Ontario salmon, was strongly supportive of the idea that subtle neurological effects resulting in measurable behavioral changes occur as a direct result of contaminant exposure through consumption of contaminated fish.
Aerial transport is a continuing source of many toxic contaminants (e.g., DDT, PCBs, TCDD), but in addition to this more or less global transport, in the Great Lakes there are 42 identified heavily polluted areas ofconcern. These are mostly industrialized areas in and around various cities/harbors located throughout the Great Lakes (46). Fish from these contaminated harbor areas often carry a broad range of industrial and anthropogenic compounds that are not present in fish from nonindustrialized parts of the Great Lakes ecosystem (56).
In many of these areas, a major problem involves in-place pollutants (contaminated sediments). Sediments from urban areas are often heavily contaminated with PAH compounds. Because PAHs are actively metabolized and excreted by fish, members ofthis class ofcompounds do not accumulate in fish to the same extent that chloroorganics do. Although PAH compounds can be detected in fish taken from areas heavily contaminated with PAHs, the most prevalent PAHs in edible fish tissue are low molecular weight, noncarcinogenic PAHs, with only trace amounts of carcinogens (21,57) and their metabolites (58). Limited data for PAH compounds in fish indicate that oily, bottom feeding fish such as carp do accumulate higher residue levels than less oily species (59). Conversely, PAH do accumulate to significant concentrations in benthic food chain organisms (57), which apparently lack or are deficient in the appropriate mixed-function oxidase enzymes. These benthic food chain organisms constitute a known, quantifiable route of PAH exposure to fish that use these organisms for food (60). Furthermore, there is ample evidence to suggest that, once ingested, PAHs can either be metabolized to biologically active metabolites that are covalently bound to cellular DNA and/or excreted as conjugates in the bile (57,61,62). Some fishing occurs even in aquatic areas known to be badly polluted, despite fish consumption advisories or health warnings that have been issued by fisheries and/or health agencies. In some of these heavily contaminated urban areas, local anglers fish for and consume significant quantities of chemically contaminated fish. The consumption of these fish by anglers and potential Dieldrin Coho salmon Lake Ontario 0.03 4g/g (43) cis-Chlordane Coho salmon Lake Ontario 0.05 g/g (43) trans-Chlordane Coho salmon Lake Ontario 0.02 Ag/g (43) Mirex Coho salmon Lake Ontario 0.14 jzg/g health risks associated with consumption ofthese contaminated fish has not been well studied. Ironically, in the most heavily polluted urban areas, heavily exposed bottom-feeding species such as carp, catfish, suckers, and freshwater drum often make up the bulk ofthe catch. Projected cancer risks associated with the consumption ofcontminated sport and recreationally caught fish appear to be significant even at levels that fall below FDA action levels (63).

Conclusions
In conclusion, there is a chain ofevidence to suggest that contaminants in aquatic systems are affecting the health status of aquatic organisms as well as the humans consuming contaminated fish from these systems. For the most part, the effects appear to be subtle, and the significance of these effects is not well understood. New approaches to the problems of recognizing and measuring these subtle effects will be required. Cancers in some species of fish from some locations appear to be wholly or partly due to exposure to xenobiotic chemicals. Although some neoplasms in fish may be large and relatively easy to detect, the initiating event involving the formation ofDNA adducts and possible promoting effects ofchronic exposure involve very subtle events. Against this background, the rationale for further studies of neoplasia in native fish populations is obvious.