Carcinogen adducts as an indicator for the public health risks of consuming carcinogen-exposed fish and shellfish.

A large variety of environmental carcinogens are metabolically activated to electrophilic metabolites that can bind to nucleic acids and protein, forming covalent adducts. The formation of DNA-carcinogen adducts is thought to be a necessary step in the action of most carcinogens. Recently, a variety of new fluorescence, immunochemical, and radioactive-postlabeling procedures have been developed that allow the sensitive measurement of DNA-carcinogen adducts in organisms exposed to environmental carcinogens. In some cases, similar procedures have been developed for protein-carcinogen adducts. In an organism with active metabolic systems for a given carcinogen, adducts are generally much longer lived than the carcinogens that formed them. Thus, the detection of DNA- or protein-carcinogen adducts in aquatic foodstuffs can act as an indicator of prior carcinogen exposure. The presence of DNA adducts would, in addition, suggest a mutagenic/carcinogenic risk to the aquatic organism itself. Vertebrate fish are characterized by high levels of carcinogen metabolism, low body burdens of carcinogen, the formation of carcinogen-macromolecule adducts, and the occurrence of pollution-related tumors. Shellfish, on the other hand, have low levels of carcinogen metabolism, high body burdens of carcinogen, and have little or no evidence of carcinogen-macromolecule adducts or tumors. The consumption of carcinogen adducts in aquatic foodstuffs is unlikely to represent a human health hazard. There are no metabolic pathways by which protein-carcinogen or DNA-carcinogen adducts could reform carcinogens. Incorporation via salvage pathways of preformed nucleoside-carcinogen adducts from foodstuffs into newly synthesized human DNA is theoretically possible.(ABSTRACT TRUNCATED AT 250 WORDS)


Introduction
The exposure ofaquatic organisms to carcinogenic chemicals has long been recognized. Considerable attention has been paid to the possibility of cancer in aquatic species as a result of exposure to environmental contaminants (1). However, less attention has been paid to the direct public health consequences of eating carcinogen-contaminated aquatic foodstuffs. Recently, the discovery ofhigh tumor frequencies in certain wild fish populations has focused public attention on carcinogen contamination of the aquatic environment and heightened awareness of the potential problems ofconsuning carcinogen-contaminated fish. The potential magnitude ofthe problem is shown by recent risk estimates, which have indicated that people eating sports fish from the Great Lakes may be exposed to a carcinogenic risk from pesticides and PCBs that is orders of magnitude higher than their risk from drinking Great Lakes water (2).
A variety ofchemical carcinogens are actively metabolized by aquatic species. Metabolism of carcinogens typically generates electrophilic metabolites that can bind to nucleic acids and protein, forming covalent adducts. Exposure ofan aquatic organism to a metabolizable carcinogen thus has the potential for contaminating an aquatic foodstuff not only with the carcinogen Bris Columbia Cawa Resrch Cer, VXnove, BC, Canada V5Z L3. itself, but also with its metabolites and with adducts of the metabolites with cellular macromolecules. These metabolites and adducts must be independently considered in assessing the potential health dangers ofeating the organism. On the positive side, adducts and metabolites may be useful in assessing carcinogen exposure ofaquatic organisms. This paper explores how adduct measurements can be used in assessing the exposure of fish populations to carcinogens. Attention is focused on DNA adducts ofPAHs (polycyclic aromatic hydrocarbons). The general remarks, however, are likely to be applicable to other carcinogens and to carcinogen adducts to proteins.

PAHs in Foodstuffs
The original concern over PAHs in foodstuffs was focused on the contamination of food by smoking or barbecuing (3) and whether or not this might be linked to the high rates of stomach cancer in areas such as Iceland where smoked foods compose an important part ofthe traditional diet (4). This concern continues to be expressed; a recent example is the possible relationship between native smoked meat and the high incidence of stomach cancer in Nigeria (S). Kramers has recently evaluated the potential human intake and risks of PAHs in the diet and in air pollution (6). The estimated PAH intake from the diet exceed PAH in take from air; however, the risk from the diet was estimated to be less than the risk from air (6). An analysis ofthe average Swedish diet has indicated that cereals, vegetables, and oils are the major sources ofdietary PAHs (7). Such analyses, however, do not take into account food preferences in individuals, which could lead to deviations from the average. While there are no good epidemiological studies linking PAH-contaminated foodstuffs to human cancer, the fact that PAHs are powerful carcinogens in animals has led public health authorities in some countries to enact regulations limiting their concentrations in foodstuffs. Similarly, widely publicized dietary recommendations for the reduction ofcancer risk have suggested limiting the consumption of smoked food products (8).
Reports of PAH carcinogens in edible aquatic organisms date back at least three decades to the work ofCahnmann and Kuratsune (9), who measured a number ofcarcinogens in the tissue of oysters collected from a polluted environment. A survey of PAH carcinogens in fresh and tinned commercial seafoods has indicated that while PAH levels in vertebrate fish are low or not detectable, appreciable levels of PAHs are found in shellfish and in crustaceans such as crab and lobsters (10). In addition to such contamination of widely distributed foodstuffs, people may be exposed to carcinogens in noncommercial fish or shellfish harvested by sports fisherman, often from heavily polluted waters. PAH contmination is not limited to the animal kdngdom, as PAHs have also been detected in seaweeds (11). PAH contamination of aquatic food resources thus appears to be widespread, and public consumption of PAH-contaminated seafoods appears to represent an appreciable exposure to these carcinogens (12).

Effects of PAHs on Aquatic Organisms
In the past, most work on carcinogen levels in aquatic organisms has concentrated on the problems of chronic or catastrophic contamination of the marine environment by petroleum products. The main focus of the work was the acute and chronic toxicity of pollution to aquatic organisms themselves, and little ifany attention was paid to the public health aspects of carcinogen-contaminated seafoods. Much of this earlier work on PAHs in the aquatic environment has been reviewed by Neff (13) and by Stegeman (14).
More recently, studies have begun to examine the possible role of pollution-associated PAH as a causative agent for tumors in wild fish populations (I). The best-studied example ofcarcinogen induction of fish tumors is the case of liver tumors in bottomdwelling marine flatfish in Washington State. Tumors appear to be related to sediment levels ofPAHs (1,(15)(16)(17). In the fresh water environment, brown bullheads from a variety of PAHcontaminated areas such as the Black, Buffalo, and Fox rivers have been reported to have an elevated incidence ofliver tumors (1). Bile from such bullheads exhibits fluorescent metabolites resulting from the metabolism ofPAHs (18). Baumann has suggested that PAHs are the most likely carcinogens to be responsible for the high rate ofliver tumors in brown bullheads in contaminated environments (19).
For invertebrates, a large number ofstudies have shown a relationship between high levels of PAHs in contaminated environments and the accumulation of substantial body burdens of PAHs. However, there is a general consensus that there is no good evidence for any relationship between environmental PAHs and invertebrate neoplastic diseases (1).

Adduct Formation and Methods for Measuring Adducts
Ifan environmental carcinogen is not metabolized, it tends to accumulate in organisms until a steady-state body burden is reached at the point that excretion of the compound equals the rate ofuptake from the environment. Ifmetabolism takes place, its effect is typically to form hydrophilic metabolites that are rapidly excreted and tend not to accumulate in the body. This sharply lowers or even eliminates the steady-state body burden of the parent compound. Thus, from the point of view of carcinogen contmination ofedible organisms, metabolism is good. From the point ofview ofan organism encountering environmental carcinogens, carcinogen metabolism is usually bad. Reactive electrophilic intermediates formed during the metabolism ofcarcinogens can react with cellular macromolecules, forming covalent adducts. The formation of DNA-carcinogen adducts is thought to be a necessary step in the action ofmost carcinogens (20).
There has been extensive research directed toward measuring DNA-carcinogen adducts. Levels of adducts in tissues of carcinogen-exposed animals are low, with typical DNA modification levels being less than one adducted nucleotide per million normal nucleotides. Until recently, adducts could only be measured in a laboratory situation, using radioactively labeled carcinogens. A variety of new fluorescence, immunochemical, and radioactive-postlabeling procedures have, however, now been developed that allow the sensitive measurement of DNA-carcinogen adducts in organisms exposed to environmental carcinogens (21). In some cases, similar procedures have been developed for protein-carcinogen adducts. These new techniques are beginning to be used with human subjects in investigations commonly termed "biochemical epidemiology" or "molecular epidemiology" (21)(22)(23)(24)(25). Similar approaches can be applied to wild animal populations exposed to environmental carcinogens. Table 1 lists some of the characteristics of three different classes ofprocedures for measuring DNA-carcinogen adducts in isolated DNA. Ifa carcinogen forms a fluorescent adduct, then sensitive procedures for measuring the adduct by fluorescence of native or degraded DNA may be developed. This technique has been applied to PAH adducts, but it is not widely applicable to adducts ofother carcinogens. Immunoassay procedures have a similar sensitivity, but are more generally applicable. However, such assays depend on the availability of well-characterized monoclonal or polyclonal antibodies, which can be difficult and time-consuming to develop. Furthermore, immunoassays require chemical synthesis ofthe adduct to act as an antigen in the Field studies of carcinogen adducts in organisms from chemically contaminated environments are in their infancy. Using 32P-postlabeling, we have recently examined DNA from the livers ofbrown bullheads from the Buffalo and Detroit rivers, and compared these samples with DNA from bullheads raised in clean aquariums (35). Samples were analyzed by both the nuclease P1 and HPLC enrichment versions of the 32P-postlabeling assay (36), with selective chromatography to isolate aromatic DNA adducts. Adducts show up as black spots or zones on autoradiograms of thin-layer chromatograms (Fig. 1). Aquarium-raised bullheads exhibited background spots, which may represent either endogenous adducts or adduction from common environmental contaminants. Fish from polluted areas showed, in addition, a strong diagonal radioactive zone, which we interpret as resulting from multiple overlapping adduct spots. The chromatographic and enzymatic behavior ofthe adducts is consistent with their identification as multiple DNA adduct species arising from mixtures ofPAHs, although this identification has as yet not received any direct confirmation (35). Also using 32P-postlabeling, Reichert has reported the presence of a similar zone ofputative aromatic DNA adducts in DNA from the livers of marine flatfish from polluted areas (37).
Consistent with the high metabolic capability ofvertebrate fish for PAHs, body levels ofPAHs in fish appear to be either very low or not detectable (I). Even in fish from the heavily polluted Black River, levels oflower molecular weight aromatic hydrocarbons were orders of magnitude below the levels in sediments, and higher molecular weight, carcinogenic PAHs were undetectable (38). Early work on PAH metabolism in marine organisms has been reviewed by Stegeman (14). Recent research has begun to delineate subtle details ofcarcinogen metabolism in fish cells in culture (27) and to investigate the metabolism ofcarcinogens and other foreign compounds in wild fish (28). In vertebrate fish, PAH metabolism is generally very rapid. In at least some species of fish, ingested PAHs are extensively metabolized in the intestine, and only a fraction of the applied dose reaches the circulation (29). When top minnows were exposed to waterborne radiolabeled benzo(a)pyrene, 64 to 70% of the labeled material taken up in a 24-hr exposure was already in the gallbladder or gut, indicating rapid metabolism and excretion ofthe carcinogen (30).
Consistent with the effective metabolism of PAHs by fish, DNA and protein adducts of benzo(a)pyrene were readily detected in English sole injected with benzo(a)pyrene (31). For an equivalent dose ofbenzo(a)pyrene, levels of DNA adducts in English sole and starry flounder were higher than in rats (32). When bluegill sunfish were exposed to benzo(a)pyrene, fluorescent techniques indicated the presence ofboth DNA and protein adducts (33). Similarly, in laboratory experiments using radioactive carcinogens, English sole exposed to sediments contaminated with aromatic hydrocarbons and radiolabeled benzo(a)pyrene readily formed carcinogen-macromolecule adducts (34). In this study, steady-state levels ofbenzo(a)pyrene adducts in tissue were reached after approximately 4 days.
The consensus from a number of investigations is that invertebrates have only a marginal capability for metabolizing and eliminating PAHs (1,14). For example, the half-life of benzo(a)pyrene in environmentally contaminated mussels has been reported to be greater than 2 weeks (39). A large proportion of the metabolism that does takes place in shellfish does not appear to proceed via the action of cytochrome P.450 and produces quinones rather than diolepoxides as a major product (40). Such metabolites do not have the biochemical characteristics necessary for DNA adduct fornation. Consistent with this, it has been reported that mussels cannot activate benzo(a)pyrene to mutagenic metabolites (41). A similar situation may hold true for aromatic amine carcinogens. It has been shown that carcinogen metabolism in carp liver microsomes is similar to that of mammalian species, but metabolism in mussels is different and does not involve cytochrome P450 monooxygenases (42). Consistent with the low rate of PAH metabolism in shellfish, a variety of bivalve species have been repeatedly found to contain high levels of pollution-associated PAHs (1).
To examine the possibility ofadduct formation in wild shellfish from PAH-polluted areas, we have performed 32P-postlabeling analysis on DNA from mussels (Mywilus edulis) from the outer harbor of Vancouver, Canada. Clean mussels were from foreshore rocks at a swimming beach, while polluted mussels were growing on creosoted pilings approximately 500 m from the clean samples. PAH levels in mussels were not analyzed, but we have previously shown consistently elevated levels of PAH carcinogens in mussels growing in creosote-contaminated areas (11,43,44). Figure 2 shows autoradiograms of chromatograms derived from DNA of the foot, mantle, and gill of mussels from both areas. Chromatograms exhibited a complex pattern of faint spots that appeared to be tissue specific. However, there were no major differences in the pattern or intensity ofthe spots between mussels from the clean and from the PAH-contaminated area. This indicates that, at least in this preliminary examination, adduct formation in mussels from environmental PAH seems unlikely. Further investigations in which other tissues such as the metabolically active digestive gland (14,40) are examined are clearly warranted before a firm conclusion can be reached about the possibility of PAH adducts in shellfish.

Adducts As Indicators of Environmental Contamination
The presence of DNA adducts of a given carcinogen in the tissue of an aquatic organism provides an unequivocal indication of prior carcinogen exposure of the organism. The exact level of adducts in an organism in a given contaminated environment, however, is likely to be a complex function of such factors as the organisms' age and sex, the time of year, and environmental factors such as food availability and water temperature, which can modify carcinogen metabolism (45). It is therefore likely to prove difficult to demonstrate a consistent quantitative relationship between adduct levels in organisms and environmental contamination.
The interpretation ofadduct levels in aquatic organisms is further complicated by the scarcity ofknowledge about the lifetime of adducts in marine species. In mammals, PAH adduct repair and removal typically takes place with a half-life of days (46): however, adducts of some carcinogens have been detected months after exposure (47). In fish, there are indications that the DNA repair processes that remove adducts are less active than in mammalian species, which would lead to the prediction of long life times for environmentally caused adducts (48). Consistent with this, Varanasi has found that DNA adducts of benzo(a)pyrene were essentially unrepaired in English sole over a period of2 weeks (31P. Recent research has suggested that not all adducts in mammalian DNA are repaired at the same rate (47).
Ifthis holds true for fish, an aquatic orgnism chronically exposed to low levels of a carcinogen might be expected to remove repairable adducts as quickly as they form, but to gradually accumulate unrepairable or slowly repairable adducts. It is these accumulated adducts that are likely to be detected in monitoring programs exanining wild fish. In view ofthis, it seems likely that adducts will have a long half-life in fish and that pollutioninduced adduct levels will represent not so much acute exposure to carcinogens, but the typical contamination levels in the aquatic environment over a period of weeks or months.

Adducts As Indicators of Public Health Dangers
Adduct analysis is generally more costly and difficult than direct analysis for carcinogens, and for this reason if no other it UNPOLLUTED POLLUTED FOOT MANTLE GILL FPcuRi. 2. 32P_postlabeling detection of DNA adducts in mussels. DNA, 2. 5 ig samples.. from the foot. mantle, or gills of mussels from a c;ean and a creosote-contaminated area in the outer Vancouver harbor were analyzed by tme nuclease P1 version of the 32P-pcstlabeling assay, as described elsewhere (35, .36). Adducts separated on thin layer chromatograms were detected by at-oradiogaphy for 65 hr at room temperature. Arrows indicatd backgrund spots not related to environmental contaniination. seems unsuited as a testing method to determine if an edible aquatic organism is at risk for carcinogen contamination. Hewever, the presence in an organism of adducts of a given carcinogen clearly indicates past exposure to the carcinogen and raises the possibility of ongoing contamination with the carcinogen. In addition, the presence of adducts will generally suggest the presence ofcarcinogen metabolites, for which there are generaly no good analytical techniques available and which may pose an appreciable human health hazard. For public health purposes therefore, it seems prudent that where analysis has shown the presence ofcarcinogen adducts. an aquatic organism should be regarded as being carcinogen contaminated.
Without further information, the a7bsence ofadducts cannot mdicate the absence of carcinogens. The absence ofadducts mav simply be due to the failure of the species in question to meta bolize a carcinogen to reactive species. Furthermore, even in organisms capable of carcinogen metabolism, metabolism and adduct formation are frequently tissue specific. The failure to detect adducts in a given tissue may simply indicate that the wrong tissue was chosen. Only when adducts of a specific carcinogen are known to occur in a specific tissue of a specific species can their absence be used as evidence for the absence of the carcinogen.

Public Health Dangers of Adducts
Carcinogen adducts are the products ofthe covalent binding of highly reactive species with cellular macromolecules. General considerations of chemical kinetics suggest that preformed carcinogen-macromolecule adducts in foodstuffs are thermodynamically unlikely to spontaneously regenerate carcinogens or carcinogen metabolites. The question then rises as to whether carcinogen-adducted protein or DNA as such in edible aquatic organisms could be dangerous.
The majority of carcinogen adducts in the edible tissue of an aquatic organism are likely to be protein adducts. When consumed and digested, these are likely to yield adducted amino acids or peptides that will be either eliminated as such or further broken down by catabolic enzymes. Even if adducted amino acids were activated and incorporated into newly synthesized protein, this would not represent a genetic or carcinogenic threat.
The intestinal digestion of adducted nucleic acids consumed in foodstuffs is likely to yield adducted nucleotides, nucleosides, and bases. Recently, Tierney et al. have demonstrated the rapid excretion of a putative benzo(a)pyrene-deoxyguanosine adduct in rats fed DNA adducted with benzo(a)pyrene (49). Metabolic pathways exist by which preformed nucleotides and bases can be salvaged and converted to nucleoside triphosphates, which can in turn be incorporated into newly synthesized nucleic acids. If adducted nucleotides were to be incorporated into DNA by these pathways, the end result could be genetic damage indistinguishable from that produced by the direct attack of the original carcinogen on DNA. There has been considerable debate over whether mutagenesis (and by extension, carcinogenesis) can occur by such mechanisms (50). DNA synthetic processes inherently have high fidelity and discriminate strongly against the incorporation ofincorrect nucleotides (SI). Incorporation is most likely therefore, for nucleotides differing only slightly from the parent nucleotide, either where the base modification does not affect base pairing or where alternate base-pairing modes exist. Even for nucleotides adducted with as little as a methyl group, evidence suggests that incorporation rates into newly synthesized DNA are extremely low relative to normal nucleotides (50). For larger adducts, the opinion has been expressed that incorporation is extremely unlikely to take place at all (50). On an even more basic level, it can be argued that if a carcinogen-adducted or otherwise abnormal nucleotide is so similar to a normal nucleotide that it can be readily incorporated into DNA, it effectively is a normal nucleotide and is unlikely to cause mutations at the next round of DNA synthesis. An example of this phenomenon is given by the thymidine analog bromodeoxyuridine, which can be incorporated at high rates into DNA with little or no effect on cells. Only a very small fraction of a carcinogen entering an organism is converted into DNA adducts, and absolute levels ofadducts in tissues ofedible organisms are hkble 2. Characteristics of vertebrate rish and shellfish with respect to polycyclic aromatic hydrocarbons. High low. Overall, therefore, the presence ofadducts in seafood does not seem likely to directly pose a danger to human health. Table 2 summarizes the relationship between PAH metabolism, carcinogen body burdens, and genetic risk for vertebrate fish and shellfish. Paradoxically, the same metabolic processes that activate environmental carcinogens to DNA-reactive forms are responsible for the efficient elimination of the parent carcinogen from an organism. An aquatic organism with a high level ofcarcinogen metabolism may be at risk from the carcinogen but may be safe to eat, as steady-state carcinogen levels in the tissue are low. Conversely, an organism that does not metabolize a carcinogen may not itself be at risk, but may accumulate substantial tissue levels of carcinogens and may be a danger to public health. In aquatic environments contaminated by PAH carcinogens, vertebrate fish have active carcinogen metabolism, high DNA-carcinogen adduct levels, but low tissue levels ofthe parent carcinogens. Shellfish, on the other hand, are characterized by low carcinogen metabolism, low DNA-carcinogen adduct levels, and high tissue levels of the parent carcinogens. In keeping with this pattern, there is compelling evidence for the carcinogenicity of PAHs to vertebrate fish, but little or no evidence of their carcinogenicity to shellfish.