SKIN

That benzo[a]pyrene hydroxylase can be induced in cultured human skin was first demonstrated in 1972.¹⁰³ Foreskins from children who were circumcised 2–4 d after birth were shown to contain an enzyme that hydroxylates the carcinogen benzo[a]pyrene (BaP), and induction of the enzyme (by a factor of 2–5) was demonstrated when the foreskins were cultured for 16 h in the presence of 10 µM benz[a]anthracene. Among a group of 13 skin samples studied, control enzymatic activities extended over a threefold range and were not correlated with race, age of mother, or medications given to mother or infant.² The enzyme had an absolute requirement for NADPH and molecular oxygen and was completely inhibited by CO; these findings suggested the involvement of a species of cytochrome P-450 in the hydroxylation reaction. The presence of this heme protein in low concentrations in cutaneous tissue was later demonstrated by Bickers et al.²² Coal-tar products, which are widely used in dermatologic practice in conjunction with exposure to ultraviolet light (e.g., the Goeckerman regimen for psoriasis^62,63), were also shown to induce aryl hydrocarbon hydroxylase (AHH) significantly in patients with dermatologic disease where the coal tar was applied, but not in skin distant from the site of application.²¹ In concurrent studies in neonatal rats,²¹ although (as in humans) distant skin sites were unaffected by the coal-tar application, AHH activity in the livers of treated animals increased to more than 20 times the control values. Among five identifiable constituents of coal tar studied²¹ for their AHH-inducing properties in human skin, BaP was the most potent; pyrene and anthracene also caused significant induction of the enzyme. In isolated cultured human hair follicles, Vermorken et al.,¹⁸⁵ using radioactive BaP as substrate, not only demonstrated the presence of the hydroxylase, but identified the formation of the 3-OH, 7,8-dihydro-7,8-dihydroxy, and 9,10-dihydro-9,10-dihydroxy metabolites of this PAH.

BaP clearly has cytotoxic effects on cultured human skin fibroblasts, although relatively high concentrations are required for cytotoxicity. Milo et al.¹²³ studied the influence of the three carcinogenic PAHs, 7,12-dimethylbenzanthracene (7,12-DMBA), 3-methylcholanthrene (3-MC), and BaP on mixed-function oxidase^* activity, cell proliferation kinetics, and DNA damage in cultured fibroblasts. They found that only BaP, at 10 µg/ml or higher, affected all the cellular metabolic characteristics examined. 7,12-DMBA at 6 µg/ml or higher induced the mixed-function oxidase system and stimulated DNA synthesis; 3-MC at concentrations as high as 15 µg/ml produced no significant cellular alterations. Similarly, 5-fluoro-7,12-DMBA, anthracene, and phenanthrene had no effects on these human cells. The authors concluded that BaP alone could initiate all the biochemical events probably necessary to trigger transformation of human cells in vitro.

PAH-induced cytotoxicity to cultured human fibroblasts has also been demonstrated by Strniste and Brake¹⁷² and Aust et al.⁸ In the former study, normal fibroblasts and xeroderma pigmentosum (XP) cells were used, and BaP was “activated” by light radiation (near ultraviolet), rather than enzymes. Photoactivation (at 300–400 nm) of BaP produced at least three identifiable quinones (1,6-, 3,6-, and 6,12- isomers), as well as more hydrophilic products, depending on the duration of light exposure. Formation of these products was oxygen-dependent. The irradiation products led to several types of DNA damage, with covalently bound hydrocarbons constituting the major lesion under all conditions studied. XP cells were more sensitive to damage than normal cells (by a factor of 1.7–2), and sensitivity increased by a factor of 10 when long-wavelength ultraviolet light was used. 7,12-DMBA, 3-MC, and BaP were also examined; the order of phototoxicity was 7,12-DMBA>BaP>benzo[e]pyrene>3-MC.

In the study of Aust et al.,⁸ a human epithelial cell-mediated cytotoxicity and mutagenicity assay system for BaP was developed with human fibroblasts as the target cells. Lethally x-irradiated human kidney-carcinoma epithelial cells were cocultivated with human XP skin fibroblasts (XP12BE) lacking excision-repair capability for BaP-DNA adducts. Under defined conditions, the frequency of mutation to 6-thioguanine resistance and PAH binding to DNA were shown to be concentration-dependent. Two principal BaP-DNA adduct peaks could be identified—a major peak consistent with an adduct standard synthesized from the anti-isomeric 7,8-dihydrodiol-9,10-epoxide of the hydrocarbon and a minor peak consistent with the syn-isomeric form of this metabolite. The results are consistent with those in other reports on BaP adducts formed in human explant tissue from lung,¹⁶¹ colon,¹⁰ esophagus,⁶⁸ and bronchus,⁷⁵ and they represent an advance in the development of sensitive assay systems for detecting biologic responses to human epithelial-cell activation of BaP.

Direct neoplastic transformation of human fibroblasts by carcinogens has also been demonstrated. In the study of Kakunaga,⁷⁹ normal human adult fibroblasts exposed to the carcinogen 4-nitroquinoline 1-oxide underwent malignant transformation in a process requiring numerous cell divisions. When injected into athymic (nude) mice, the transformed cells produced solid tumors at the site of inoculation. Because it could not be metabolically activated by the target cells used, 3-MC was unable to effect transformation; the use of other PAHs and induced microsomes with high concentrations of cytochrome P-450 to activate 3-MC was not examined. Normal human foreskin cell populations were neoplastically transformed in studies by Milo and DiPaolo^124,125 with a number of non-PAH carcinogens; and treatment with a tumor promoter alone (phorbol ester) has been shown⁹⁰ to induce neoplastic transformation in fibroblasts from humans genetically predisposed to cancer (familial polyposis of the colon). Thus, it can be inferred that cells already in an “initiated state” as a result of a genetic defect represent a novel fibroblast system that may provide a means for exploring separately the roles of initiators and promoters in carcinogenesis. Painter¹³⁵ used HeLa cells to develop a rapid screening test to detect agents that damage human DNA. The test measures thymidine uptake into the cells at various times (principally 1–2 h) after treatment with a presumptive carcinogen or mutagen. In this test system, BaP was inert unless metabolically activated by incubation with rat-liver microsomes. Brookes and Duncan²⁶ compared the effects of PAHs on primary human embryo cells and HeLa cells. Fibroblasts from skin, lung, muscle, and gut were cultured and treated with ³H-labeled BaP and 7,12-DMBA. Both hydrocarbons were metabolized in the cultures, 7,12-DMBA more slowly than BaP; among the cell types studied, lung fibroblasts metabolized the compounds more efficiently than others and retained this capacity well in subculture. The binding of BaP and 7,12-DMBA to DNA, RNA, and protein of these primary lung-derived fibroblasts was also studied (metabolism of each hydrocarbon exceeded 75% during the 48–96 h of treatment) and was found to be significantly greater for BaP than for 7,12-DMBA. The data in this study also established parallelisms, at least for BaP, between hydrocarbon binding to cellular macromolecules in fibroblast cultures derived from mouse embryos and those derived from human lung cells. Such parallelism is of more than casual interest, in view of the susceptibility of the mouse to hydrocarbon carcinogenesis and the known correlation between hydrocarbon-DNA binding and cancer-producing activity in mouse skin.

The effects of pyrene and BaP in the human diploid fibroblast culture WI-38 were studied by Weinstein et al.¹⁸⁸ Neither caused significant damage (compared with controls), as assessed by mitotic index or chromosomal breaks after 1-h pulse exposures. However, metabolic activation of BaP with microsomes resulted in a dramatic decrease in mitotic index and a significant increase in breakage. Microsomal incubation did not alter the inertness of pyrene in this test system. Freeman et al.⁶⁰ have made interesting observations on comparative aspects of hydrocarbon metabolism in skin epithelial and fibroblast cultures. A comparison of the ability of epithelial-cell colonies and of fibroblast colonies from the same 13 subjects to metabolize BaP to a water-soluble form demonstrated clearly the markedly greater metabolizing capacity of epithelial cells. There was a 20-fold difference in this capacity of epithelial cells; within individual subjects, the ability of epithelial cells to metabolize the PAH exceeded that of fibroblasts by as much as a factor of 40. There appeared to be a major effect of culture age (6–55 d) on the ability of epithelial cells to metabolize BaP.

A direct toxicity of BaP to normal human epithelial-cell cultures has been described by Dietz and Flaxman.⁵⁰ This toxicity was reflected in a dramatic reduction in epidermal-cell outgrowth, a decrease in mitotic indexes, a loss of the well-ordered cell relationships, and the early appearance of giant cells ranging in diameter from 100 to 200 µm. In a clinical study that clearly could not be carried out today, Cottini and Mazzone⁴⁵ (in 1939) applied a 1% solution of BaP daily (up to 120 d) to the skin of 26 normal subjects and patients with various dermatologic disorders and examined the gross and histologic consequences. The sequential epidermal changes, of which gross pigmentation and verrucae were the most frequent, and histologic alterations (which regressed within several months when treatment was terminated) led the authors to conclude that “benzopyrene, if applied to human skin for protracted periods, would be carcinogenic as it is in animals.”

The metabolism of benz[a]anthracene, 7,12-DMBA, and BaP by human mammary epithelial-cell aggregates in culture has been investigated by Grover et al.⁶⁶ with nonneoplastic tissue obtained from eight patients undergoing reduction mammoplasty. All three PAHs were metabolized to water-soluble and organic-solvent-soluble products; the latter included K-region and non-K-region dihydrodiols. The major dihydrodiols detected as metabolites of the parent PAHs were the 8,9-dihydrodiols of BaP and 7,12-DMBA and the 9,10-dihydrodiol of BaP. The hydrocarbons bound to the proteins and DNA of the epithelial cells, but there were wide differences between different PAHs in extent of binding between tissue preparations from different patients. Some of the PAH-deoxyribonucleoside adducts formed from 7,12-DMBA and BaP appeared to have been produced through reactions of bay-region diol-epoxides with DNA, but little reaction with DNA was detected in tissue preparations treated with BaP.

Unscheduled DNA synthesis induced by DNA-damaging chemicals has been measured in nonreplicating human fibroblasts by autoradiographic methods that are not readily applicable to organotypic epithelial-cell cultures. To evaluate the range of chemical sensitivity and DNA-repair responses of human skin epithelial cells, Lake et al.⁹⁵ developed a semiquantitative in vitro method for measuring unscheduled DNA synthesis in normal foreskin epithelial cells. On serial subculture of organotypic primary skin cultures, the unscheduled DNA synthesis response elicited by 3-MC decreased in parallel with the ability of cells to metabolize PAHs to water-soluble metabolites. The working hypothesis was that procarcinogens that are efficiently activated by human skin-specific metabolism will be detected with unscheduled DNA synthesis as an end point.

LIVER AND INTESTINE

Obana et al.¹³⁰ analyzed quantitatively and qualitatively the PAH content of samples of human liver and fatty tissue. Six samples of liver and 10 samples of fat were obtained at autopsy from 10 persons who died of unstated causes (although the tissues were reported to be “free from cancer”). Smoking habits, occupations, etc., were not described. The tissue samples analyzed were quite large (40–120 g), and the PAHs were determined without complex pretreatment. Table 6–1 shows the analytic results for liver, and Table 6–2 the comparable data for fat tissue. Note that PAH concentrations are expressed as parts per trillion (ppt), not parts per billion (ppb), and are in general extremely low. Nevertheless, the data indicate that, on the average, the PAH concentration in liver was one-third that in fat. Pyrene had the highest concentration, followed by anthracene. Although the number of samples was small, no sex or age differences were evident. The known carcinogens benz[a]anthracene and dibenz[ah]anthracene were not detected in either tissue. However, BaP was detected in small amounts (20 ppt) in both liver and fat. This finding should be compared with that of Tomingas et al.,¹⁷⁸ who detected BaP at 1–15,000 ppb in human bronchial-carcinoma tissue (24 samples). Obana et al.¹³⁰ called attention to the fact that the PAHs in the human tissues they examined were different, in both concentration and composition, from the PAHs that had been identified in marine samples. For example, pyrene was found in oysters at 7–52 ppb and in Wakame seaweed at 12–41 ppb; the comparable figures for BaP were 0.3–2.6 ppb and 0.6–9 ppb, respectively. Moreover, although pyrene was the most abundant PAH in all cases, the next most abundant in the human tissues was anthracene, whereas in the marine samples the next most abundant were benzo[e]pyrene and benzo[b]fluoranthene. These qualitative and quantitative distinctions, especially the marked concentration differences between nontumorous¹³⁰ and tumorous tissues¹⁷⁸ and between a common food source in the area and the human tissue samples, need to be recalled in evaluating the importance of the food content of PAHs, as well as the role of malignant pathology, when trying to determine the significance of the body or tissue burden of these hydrocarbons.

TABLE 6–1

PAHs in Human Liver.

TABLE 6–2

PAHs in Human Fat.

The liver contains the highest concentration of cytochrome P-450 in the body. The activity of the pathway by which heme, the prosthetic group of this heme protein, is synthesized can be greatly induced by a host of foreign chemicals and can approach the rates of heme synthesis in erythroid cells; and the enzymatic capacity of hepatic cells to carry out the biotransformations that characterize the great variety of PAH metabolites formed in vitro, and probably in vivo, has been well defined. Only selected PAH transformations catalyzed by liver cells are reviewed here, with some emphasis on the relationships of PAH- and drug-metabolizing capabilities and on recent data indicating that carcinogen metabolism may be increased by direct actions on relevant membrane-bound enzymes, as well as by the conventionally assumed process of increased de novo synthesis of enzyme protein, i.e., induction.

Dybing et al.⁵⁴ have examined the in vitro metabolism and metabolic activation of several carcinogenic PAHs in subcellular fractions from seven human livers. The patients all suffered from total cerebral infarction and were serving as potential kidney donors (maintained temporarily by life-support systems) at the Huddinge University Hospital in Sweden. At the appropriate time, liver extirpation was performed; within 20 min after the procedure, perfusion had been completed and the tissue frozen in liquid nitrogen. This study may mimic the enzymatic properties of human liver cells in the living subject as closely as experimentally possible, other than by direct biopsy or surgery in a living patient. Because of the unique source of the tissue studied, some of the data merit recording here. Microsomal cytochrome P-450 content (seven livers) was 0.16–0.60 nmol/mg of protein, with a mean±S.D. of 0.36±0.15, and AHH activity averaged 175±138 pmol/mg of protein per minute; one sample had a value of 483 pmol/mg. These activities are approximately the same as those in liver microsomes of untreated mice and rats. AHH activities expressed per nanomole of cytochrome P-450 varied by a factor of 2.8 among the seven liver samples.

Conney and co-workers^{27,28,40,41,73,81,82,94,156,157,174} conducted a series of studies of direct liver-cell metabolism of carcinogens and compared such metabolism with that of drugs. They established that carcinogen metabolism may be increased not only through enzyme induction, but through enzyme activation as well. They compared the oxidative metabolism of BaP with that of antipyrine, hexobarbital, coumarin, zoxazolamine, and 7-ethoxycoumarin in 32 adult liver samples obtained at autopsy some 8–20 h after death. When enzyme activity for one substrate was plotted against enzyme activity for a second substrate for each of the 32 liver samples, significant correlations were found. For example, for BaP paired against antipyrine, hexobarbital, zoxazolamine, and coumarin, the correlation coefficients were 0.85, 0.72, 0.69, and 0.57, respectively. Some drug-drug metabolizing activities also showed a high correlation (e.g., antipyrine and hexabarbital, r=0.79; antipyrine and coumarin, r= 0.72), whereas in other instances, metabolizing capacities did not have a high correlation, e.g., carcinogen vs. drug (BaP and 7-ethoxycoumarin, r=0.35) and drug vs. drug (e.g., hexobarbital and 7-ethoxycoumarin, r=0.37). The findings raise the possibility that an in vivo drug-metabolism assay (e.g., a plasma disappearance-rate study of a suitable test drug) might predict some carcinogenmetabolizing capabilities of a person and suggest the presence in humans of multiple monooxygenase systems for the substrates studied, as well as their heterogeneous distribution in the population. Individual differences in the rates of metabolism of BaP (7-fold) in these and other liver samples studied⁴¹ were considerable, although they did not approach the known species differences⁴⁰ in rates of metabolism of drugs.

The effects of PAH administration on the metabolic disposition of a specific carcinogen, such as BaP, have not been studied in humans, but Schlede et al.^156,157 recently examined the metabolic disposition of radiolabeled BaP in rats, and the results probably can be extrapolated to humans. Pretreatment of rats with unlabeled BaP greatly increased the plasma disappearance rate of a tritiated dose of the same compound given intravenously; the effect was especially marked during the first 5 min after the intravenous dose of the radiolabeled material, and the increased rate lasted for at least 6 h. This effect of pretreatment with the compound was paralleled by a lower concentration of [³H]BaP in brain, liver, and fatty tissues; similar but more varied results were observed in lung tissue. These influences of BaP pretreatment on a later intravenous dose of the ³H-labeled chemical were also observed when the radiolabeled PAH was administered orally. 3-MC and 7,12-DMBA pretreatment of animals produced comparable effects on the metabolic disposition and tissue contents of radiolabeled BaP. Pyrene, anthracene, and phenobarbital had little or no such effect on the in vivo disposition of this compound. In other studies, the biliary excretion of [¹⁴C]BaP was shown to be increased by pretreatment with the unlabeled compound; however, no increase in excretion into bile of the ¹⁴C-labeled metabolites of BaP was observed after prior administration of this PAH. These findings suggest that conversion of BaP to its metabolites may be the rate-limiting step in the biliary excretion of the compound. Phenobarbital had no effects on the pharmacokinetics (plasma disappearance) of [¹⁴C]BaP and its metabolites; this drug might stimulate the conjugation of hydroxylated BaP derivatives before their excretion into bile. The relevance of these findings to humans is related not only to the (probably) qualitatively similar pharmacokinetics of such a chemical as BaP—particularly its extensive excretion via bile—but also to the potentially important use of selected drugs, singly or multiply, to increase disposal of PAHs and their metabolites from the body by stimulating conjugation and biliary excretion and by increasing otherwise innocuous metabolic biotransformations.

Prough et al.¹⁴³ have also studied BaP metabolism in human liver, kidney, and lung, and they characterized the metabolites formed by HPLC techniques. Tissue samples were obtained within 1–5 h after death, and assays were completed within the succeeding 2–4 h. In the analysis of metabolites formed from BaP, quinones, three classes of dihydrodiols, and two classes of phenols were categorized. Table 6–3 summarizes the rates of formation of these metabolites by tumor, liver, kidney, and lung microsomes and presents comparable data in rodents. There was a very large variation in the human metabolism of BaP, compared with that demonstrated in studies carried out concurrently in rodents. That might reflect, as the authors noted, either the controlled environment of the animals studied or a genetic variation in humans. In the human liver, activities for metabolite formation were substantially lower than those in rat microsomal fractions, and there were significant differences in the BaP-metabolite profiles. A greater proportion of benzene-ring metabolites was formed by human lung microsomes than by human liver and kidney microsomes (or rodent lung microsomes). The relative increase in the 9, 10-dihydrodiol product, as well as some increase in the 7,8-dihydrodiol metabolite, accounted for the larger portion of this difference among lung, liver, and kidney microsomes. There is an apparent biologic inconsistency in these findings: although the human lung is the principal site of PAH tumorigenesis, as the authors observed, the 9,10-dihydrodiol product is suggested to have little biologic activity on further metabolism, whereas other tissues, such as the liver, formed large concentrations of the 7,8-dihydrodiol, a “proximate carcinogen.” Nevertheless, the findings are important, providing not only data on comparative rates of formation of metabolites constituting the “HPLC profiles” in man and rodents, but also intertissue metabolic profiles of BaP biotransformation for three major organ systems in man. Thus, they extended the earlier findings of Selkirk et al.¹⁶⁰ on liver cells and lymphocytes.

TABLE 6–3

Rates of Formation of Benzo[a]pyrene Metabolites by Hamster, Rat, and Human Microsomes.

A major advance in defining the role of the liver in PAH metabolism and the factors that regulate liver monooxygenase activity has been the demonstration that hepatic microsomal oxidation of specific substrates can be directly increased in vitro, in addition to their property of being induced in vivo, by various chemical treatments.^27,28,73,174 Conney and colleagues have shown that 7,8-benzoflavone added to homogenates of human liver samples can increase the rate of BaP hydroxylation by a factor of up to 11. Benzoflavone also increased some drug hydroxylations substantially, although benzoflavone at concentrations lower by a factor of about 100 paradoxically inhibited these reactions. Marked individuality for activating and inhibiting effects of benzoflavone were noted, and no significant effects on the oxidation of some drug substrates (e.g., coumarin and hexobarbital) were observed.

The enhancing effect of 7,8-benzoflavone on BaP oxidation was shown to extend to other flavonoids, such as flavone, nobiletin, and tangeritin. Related compounds—such as apigenin, chrysin, fisetin, flavonone, galargin, hesperitin, kaempferal, morin, myricitin, naringerin, and quercetin—inhibited BaP oxidation. The stimulatory effect of 7,8-benzoflavone on BaP 9,10-dihydrodiol oxidation to bay-region epoxides was also studied and shown to have significant species-specific characteristics. With untreated hamster microsomes, more than 60% of the total metabolites of the hydrocarbon were bay-region diol-epoxides, whereas human liver formed less than 5% of such metabolites. Addition of 7,8-benzoflavone to the microsomal incubations dramatically stimulated the formation of these metabolites in human (and rabbit) liver microsomes. The stimulatory effects of flavonoids on hydrocarbon oxidation have recently been shown to occur in vivo as well,¹⁰⁰ so the biologic importance of this phenomenon for the intact host is likely to be considerable. The mechanism of flavonoid activation of BaP hydroxylation has recently been explored in detail by Huang et al.⁷³ The flavone stimulates the NADPH reduction of cytochrome P-450, but not that of cytochrome c, by NADPH-cytochrome c reductase; this finding supports the idea that the catalytic sites for these substrates of the reductase are different. Other evidence that these catalytic sites are different has also recently been provided by the studies of Yoshinaga et al.^197–200

Metabolic transformation of PAHs and their binding to cellular macromolecules in cultured human gut tissues have been described. Harris et al.⁶⁸ examined the metabolic fate of BaP and several other compounds in cultured esophogeal explants from eight patients, six of whom had esophageal carcinomas. Metabolism of the ³H-labeled PAH to water-soluble metabolites varied among the eight patients over the range of 1–68% of total metabolism; the variation found within a single case, however, in relation to different anatomic segments of the esophagus (proximal, middle, and distal) was quite narrow—2%. In spite of the 68-fold variation in metabolism among subjects, the patterns of conjugates formed from metabolites in general were qualitatively similar: sulfate esters, 21–55%; glucuronide conjugates, 7–37%; and glutathione conjugates, 24–66%. Most of the radioactivity of the organic-solvent-soluble metabolites of BaP cochromatographed with authentic metabolites of this compound, including its proximate carcinogenic, (−)-trans-7,8-dihydro-7,8-dihydroxy, derivative. Despite the predominant occurrence of esophageal cancer in the distal segment, the patterns of metabolites formed in all segments of the organ were similar. Binding of ³H-labeled PAHs to DNA and protein was detected in all eight cases, with binding to protein greater than to DNA in each instance. Binding among the eight cases varied 99-fold, and at least three hydrocarbon-DNA adducts, including the specific guanine adduct, were recognized. The adducts appeared identical with those previously found in human colon and bronchus; and the patterns of both metabolism and adduct formation with BaP were analogous to those found in experimental animals susceptible to the carcinogenic action of PAHs.

Autrup et al.,⁹ in an earlier and less detailed but essentially similar study, had reported comparable data based on human colon explants of tumor-free tissue. The binding of labeled BaP to cellular protein was several times higher than that to DNA; however, hydrocarbon binding to DNA correlated with tissue AHH activity (r=0.87), whereas no such correlation existed for protein binding. DNA binding (BaP, picomoles bound per 10 mg) among seven tissue samples varied over a 25-fold range; variation within subjects was small. In an extension of this work, Autrup et al.¹¹ studied the comparative metabolism of BaP (and aflatoxin B1) and hydrocarbon-DNA adduct formation in cultured human and rat colon explants. Adduct formation (in 103 cases) varied over a 125-fold range in the tumor samples and in the same subject over a 3- to 10-fold range in different segments of the organ. A number of hydrocarbon-DNA adducts were identified, and both qualitative and quantitative rat-human differences were demonstrated. Although overall BaP metabolism was similar in rat and human colon tissue, the ratio of organic-solvent-soluble to water-soluble metabolites was higher in the human; sulfate esters predominated in rat colon, whereas equivalent quantities of sulfate esters and glutathione conjugates were formed in the human tissue; and hydrocarbon-DNA binding was distinctly greater in human colon, although, as noted, there was marked variability in adduct formation within a given subject.

The comparative hydration of styrene 7,8-oxide, octene 1,2-oxide, naphthalene 1,2-oxide, phenanthrene 9,10-oxide, benz[a]anthracene 5,6-oxide, 3-MC 11,12-oxide, dibenz[ah]anthracene 5,6-oxide, and BaP 4,5-, 7,8-, 9,10-, and 11,12-oxides to their dihydrodiols was investigated in microsomes from nine human liver samples obtained at autopsy.⁸⁰ The substrate specificity of the epoxide hydratase in human liver microsomes was very similar to that of the epoxide hydratase in rat liver microsomes. Phenanthrene 9,10-oxide was the best substrate for the human and rat epoxide hydratases, and dibenz[ah]anthracene 5,6-oxide and BaP 11,12-oxide were the poorest substrates. Plotting epoxide hydratase activity obtained with one substrate against epoxide hydratase activity for another substrate for each of the nine human livers revealed excellent correlations for all combinations of the 11 substrates studied (r=0.87–0.99). The data suggest the presence in human liver of a single epoxide hydratase with broad substrate specificity, although the results do not exclude the possible presence in human liver of several epoxide hydratases that are under similar regulatory control.

CIRCULATORY SYSTEM

Juchau et al.^76,77 summarized a body of literature bearing on the hypothesis that PAHs may play an important role in the pathogenesis of arteriosclerotic lesions. The validity of this hypothesis apart, these investigators clearly demonstrated that aortic tissues from a number of species, including man, have detectable, albeit low, monooxygenase activities using BaP and 7,12-DMBA as substrates. Enzyme activities were comparable with those characterizing mouse skin. Cytochrome P-450 could be detected in primate aortas, and epoxide hydratase activity for BaP 4,5-oxide was identified in homogenates of the arterial walls of chickens and rabbits. The characteristics of the aortic monooxygenase for BaP resembled those of the enzyme system found in other tissues. It could be markedly induced, for example, by 3-MC, polychlorinated biphenyls (PCBs), and 5,6-benzoflavone; and, surprisingly, aortic homogenates produced higher than expected quantities (by as much as a factor of 28) of alkali-extractable metabolites when hematin was added to the reaction mixtures. Interestingly, hematin has been shown in other studies to degrade, in vitro, components of the monooxygenase system.¹¹¹ The primary BaP metabolites formed in rabbit aortic homogenates were the 3-OH and 9-OH derivatives, phenolic compounds known to be cytotoxic. The authors cited unpublished data to show that the aortic metabolites of BaP form covalent bonds with such macromolecules as calf-thymus DNA. Treatment of chickens with the inducer 3-MC markedly increased the amount of the PAH-DNA adducts, whereas addition of 7,8-benzoflavone in vitro inhibited binding. Aortic enzymes also have been shown to catalyze the formation of mutagenic metabolites from 7,12-DMBA. Thus, both cytotoxic and mutagenic metabolites of PAHs can be generated in vascular tissues. The possible relation of the formation of these compounds to the initial vascular injury that may presage the local development of an atherosclerotic plaque is of considerable interest.

The interaction of benz[a]anthracene and BaP with crystalline human serum albumin in solution has been studied fluorimetrically by Ma et al.¹¹⁰ Equilibrium studies indicated that both PAHs bind to the protein to the same extent. Evidence of energy transfer from the tryptophan residue of the protein (increase in the weak B region—395–420 nm—fluorescence of the PAHs) permitted an assessment of the mean distance between the tryptophan and the bound ligand, thus identifying two different binding sites in the same general area. The authors suggested that structural differences among hydrocarbons, which may greatly affect their orientations on the protein molecule, influence mainly the selection of the binding site, rather than the binding equilibrium.

In vivo BaP associates very little with serum albumin in the presence of lipoproteins. The kinetics of BaP transfer between human plasma lipoproteins have been examined by Smith and Doody¹⁶³ with high-density lipoproteins (HDL), low-density lipoproteins (LDL), and very-low-density lipoproteins (VLDL) prepared from fresh unfrozen human plasma by ultracentrifugal flotation. BaP-lipoprotein interactions were analyzed fluorimetrically, and kinetic measurements were determined by stopped-flow techniques. The half-times of BaP transfer between HDLs, between LDLs, and between VLDLs were 40, 180, and 390 ms, respectively. The transfer of these PAHs among lipoproteins of the same density class was about one-twentieth that of pyrene under the same conditions. The rate of BaP transfer between lipoproteins also decreased with increasing size of lipoprotein; at equilibrium in Vitro, VLDLs contain about 10 times more of the BaP than LDLs, and LDLs contain 20–50 times more than HDLs. The distribution between plasma and erythrocytes is different for 7,12-DMBA, BaP, benzanthracene, and anthracene, the mass of the PAH being associated with red cells (50, 70, 93, and 100%, respectively). Plasma lipid concentrations and the dynamics of lipid and lipoprotein metabolism clearly may have an impact on PAH distribution in blood and into specific tissues. For example, transfer of BaP is quite rapid, compared with the half-time for either hydrolysis of chylomicron triglyceride (about 2–5 min in humans) or clearance of the most abundant lipoproteins from the circulation (3–5 d in humans). The data of Smith and Doody¹⁶³ concerning the role of plasma lipoprotins in the transport of PAHs corroborated and extended the findings of other investigators who examined the interaction of PAHs and plasma proteins.^{13,31,32,57,162}

The specific process of BaP uptake from human LDLs into cultured human cells was examined by Remsen and Shireman.¹⁴⁹ The cell lines used were WI-38, a human embryonic lung-fibroblast line, and GM 1915, a skin-fibroblast line derived from a patient with homozygous familial hypercholesterolemia; the former cells are LDL-receptor-positive, and the latter LDL-receptor-negative. Thus, in these studies, it was possible to explore the role of LDL receptors in the cellular uptake of PAHs that enter the bloodstream transported by chylomicrons and plasma lipoproteins. The results indicated that cellular uptake of the tritiated PAH by both cell lines from delipidated or serum-free medium varied linearly with concentration, whereas incorporation of PAH bound to LDLs was much less and, at higher lipoprotein concentrations, varied nonlinearly. The presence of the PAH in the LDL preparation did not affect the binding of ¹²⁵I-labeled lipoprotein to receptor-positive cells. The study provided several findings of special importance relative to the biologic impact of PAHs—or at least BaP as a model compound—on tissues in vivo. Clearly, although LDLs carry substantial amounts of PAH, the presence of LDL receptors on cells is not necessary for tissue uptake. The fact that PAH bound to LDL was incorporated into cells more slowly than PAH in a delipidated serum or serum-free medium raises questions about the biologic significance of experimental models in which increased incorporation of BaP from particles into lipid vesicles has been demonstrated. The data from these experiments also indicate that cells that may be directly exposed to a PAH (i.e., tracheobronchial, intestinal, and cutaneous cells) before the compound reaches the bloodstream may accumulate PAH in much higher concentrations than cells exposed to the PAH bound to lipoproteins, inasmuch as the latter significantly slowed as well as limited the cellular uptake of BaP. Finally, the report indicated that BaP previously incorporated into WI-38 cells could be substantially removed (by 55–79%) in a 120-min posttreatment study period by 10% delipidated serum or LDL-containing medium. This finding implies a potential for considerable PAH redistribution and a requirement for a not insignificant period for progression of the hydrocarbon from the plasma membrane to the endoplasmic reticulum, where metabolism takes place.

The ability of human monocytes to oxidize BaP and the induction of this enzyme activity by benzanthracene have been demonstrated by several investigators.^16,17,96,145 Lake and colleagues⁹⁶ re-examined this problem with the goal of developing a practical assay for measuring whole-cell metabolism of BaP under highly standardized conditions, eliminating—among other problems—the need for a large volume of blood (50 ml) in the fluorometric assay developed earlier for AHH activitity in this cell type. By measuring whole-cell generation of water-soluble BaP metabolites over a 3-d culture period, using ³H-labeled substrate and closely controlling other characteristics, they provided a useful alternative cell system to that using mitogen-stimulated lymphocytes for characterizing BaP oxidation activity in humans.

Because of the advantage gained by the much greater inducibility of AHH activity (up to 40-fold) in cultured monocytes, compared with mitogen-stimulated lymphocytes (about 5-fold), the monocyte system was used by Okuda et al.¹³¹ to study the contribution of genetic factors to the control of individual variation in AHH inducibility. Ten sets of monozygotic tissues were assayed two to four times and 17 sets of dizygotic tissues one to three times for basal and induced monocyte AHH activity. The results indicated that 55–70% of the individual variation in AHH inducibility of monocytes was genetically determined. Variation in AHH inducibility within subjects in repeat assays was wide and approached the magnitude of the variation between subjects. Thus, a single AHH assay is an imprecise biochemical characterization of a subject. Alternatively expressed, the method then available (late 1977) made it impractical to characterize a population with genetically distinct differences in AHH inducibility. The large intrasubject variation in AHH inducibility of monocytes also indicated that, in addition to the clear genetic influences on this process, unknown environmental or technical factors expressed themselves in the test procedure.

An abundant literature exists related to the monooxygenase activity of lymphocytes; the inducibility of this activity by mitogens, which have the property of stimulating lymphocyte transformation, during which a number of metabolic activities are concurrently greatly increased; and the use of mitogen-stimulated lymphocytes to study the genetic control of AHH in man and its relation to the occurrence of some human cancers—notably those of the lung. Kouri and colleagues have reviewed key aspects of this subject;^91,92 McLemore et al.^119–122 have also provided a detailed analysis of the genetics of AHH and its purported relation to human cancer. Only a brief summary of these findings can be included here.

The identification of AHH activity in lymphocytes in 1972^29,192 and its increase during lymphocyte blastogenesis led quickly to clinical studies, the earliest being that of Kellermann et al.,⁸⁷ in which this induced enzyme activity was measured in cultured lymphocytes of normal controls, non-lung-tumor controls, and lung-cancer patients. In a preceding study in the same year, this group⁸⁶ had examined the genetic variation in AHH activity in lymphocytes of 353 normal subjects and had categorized the population into three groups—low, intermediate, and high responders with respect to AHH inducibility; the population frequencies were about 50%, 40%, and 10%, respectively. The conclusion was reached that the enzyme activity was controlled by two alleles at a single gene locus and that the high and low responders were homozygous and the intermediate group heterozygous for those alleles. In the initial lung-cancer study,⁸⁷ there was a virtual absence of cases in the low-inducibility population, and all but two cases were in the intermediate- and high-inducibility categories. All the lung-cancer cases were in heavy smokers; of the 50 subjects, 48 had an average consumption of two packs of cigarettes per day. When the two control groups (normal subjects and a non-lung-cancer tumor group) and the lung-cancer group were compared for risk of lung cancer, those with intermediate and high inducibility (48 of the 50 lung-cancer cases) had risks for lung cancer 16 and 36 times, respectively, the risk in the low-inducibility group. This study prompted considerable controversy over the next few years, during which the findings of Kellermann and associates were cast in doubt.

A strong correlation (r=0.923) was also found by Kellermann et al.⁸⁵ between the plasma elimination rate of antipyrine and the rate of BaP metabolism in human lymphocytes from a “carefully selected homogeneous” population, compared with the much lower correlation (r= 0.425) found in a “heterogeneous” population. The authors interpreted their findings as supporting the existence of common oxidative systems or common genetic control of the systems for antipyrine and BaP oxidation. Atlas et al.⁷ confirmed that plasma antipyrine half-life is correlated to some extent with AHH inducibility (r=0.84), although no intrasubject correlations were found between AHH inducibility and the oxidation of other drug substrates, such as phenylbutazone and bishydroxycoumarin. Most importantly, this group,⁷ while affirming a significant heritable determinant of AHH inducibility in human lymphocytes, failed to confirm the monogenic model and trimodal distribution of AHH inducibility in the general population, proposed by Kellermann et al.;⁸⁶ rather, the population distributions for AHH inducibility (and for plasma antipyrine half-life) were consistent with polygenic control of both traits in man. In other studies in which the relation of AHH inducibility to the occurrence of lung cancer was re-examined by Paigen et al.,^133,134 low AHH activity was found in half the tumor patients studied, in contrast with the earlier findings of Kellermann et al.,⁸⁷ and no characteristic alterations in this enzyme activity were found in the progeny of these patients. A considerable number of technical problems related to the lymphocyte-AHH assay may confound the results obtained in studies of this enzyme activity and its relation to human cancer, as noted by Kouri et al.⁹¹ However, recent methodologic advances made by this group, particularly the use of cryopreserved lymphocytes and close control of a number of assay variables, have added an important degree of precision to the assay.

Chrysene, one of several PAH derivatives (benzanthracene is another), has been shown by Snodgrass et al.¹⁶⁴ to induce AHH activity in cultured human lymphocytes (from normal subjects) with BaP as substrate. The individual variation in the monooxygenase activity observed with other inducers was also seen with chrysene.

The comparative metabolism of BaP in human lymphocytes and human liver microsomes has been studied by Selkirk et al.,¹⁶⁰ who examined the nature of the metabolites formed by each cellular system. The patterns of metabolites formed in both cell systems had characteristics quite similar to each other, with some exceptions—for example, among the derivatives formed in a 30-min incubation, all three dihydrodiols produced by liver were absent in the lymphocyte incubation mixture. In a 24-h incubation of lymphocytes, however, all three dihydrodiols formed by liver microsomes were also formed by the blood cells, and new metabolite peaks were observed, presumably reflecting more extensive biotransformation of already formed metabolites in the reaction mixture. The authors concluded that, although the ratios of some metabolites may differ and although lymphocytes form several more derivatives than does liver, many identical metabolites are produced in these two human cell types.

Schönwald et al.¹⁵⁸ studied the effect of BaP on sister chromatid exchange in mitogen-stimulated lymphocytes of 11 normal subjects and 18 patients with lung cancer. Patients and controls differed neither with respect to the spontaneous rate of sister chromatid exchange nor in their responses to the hydrocarbon, although it did double the number of exchanges in both population groups.

Barfknecht et al.¹⁵ studied the ability of dichloromethane extracts of automobile diesel soot at high concentrations (100 mg/m³) to induce trifluorothymidine-resistant mutants in human lymphocytes incubated in the presence of rat-liver postmitochondrial supernatant. A significant induction of such mutants was observed. Anthracene, phenanthrene, and their alkylated derivatives accounted for one-fourth of the observed biologic activity. Among eight related compounds, there was general agreement between responses in lymphoblasts and in bacterial test systems. Phenanthrene was an exception, in that it was positive in the human-lymphoblast test system, but negative in bacteria at a concentration 60 times higher. The data in this report indicate that methyl substitution at some sites of anthracene and phenanthrene greatly increases their mutagenicity in both S.typhimurium and human lymphoblasts. A similar effect for chrysene has been observed. Methylations at the 1 and 3 positions of phenanthrene and the 2 and 9 positions of anthracene result in PAHs that are particularly mutagenic in the human and bacterial test systems used. Methylations at other positions had the capability of eliminating the mutagenic activity of the PAH derivative. No correlation between the results of the mutagenesis studies with the soot-derived PAHs and the reported capacity of the compounds studied to elicit neoplastic or carcinogenic responses in test animals could be made.

REPRODUCTION

The title of this section refers collectively to studies related to the ability of some genital tissues (including the placenta) to metabolize or otherwise respond biochemically to PAHs- There is an abundant and detailed literature on transplacental¹⁵¹ and perinatal¹⁸² carcinogenesis. These and related topics in reproduction were reviewed in a 1981 special issue of the Journal of Environmental Pathology and Toxicology ¹⁵⁴ and are not summarized here. It is perhaps appropriate, however, to refer to the report by Sir Percival Pott in 1775,¹⁴² in which there was first described an increased incidence of scrotal cancer in chimney sweeps exposed to soot, and to note that almost 150 yr elapsed before Yamagiwa and Ichikawa¹⁹⁴ demonstrated that the repetitive application of crude coal tar to the rabbit ear produced skin cancer and that the identification of specific carcinogenic coal-tar constituents, such as BaP, required the passage of additional decades.^20,45,88 Over this period, the question of why only scrotal cancers, and not other genital cancers or even other cancers in general, were found in excess in chimney sweeps appears to have remained unanswered.

Grover et al.⁶⁶ investigated the metabolism—including the specific identification of biotransformation products—of three ³H-labeled PAHs by nonneoplastic human mammary epithelial-cell aggregates maintained in culture. The lobuloalveolar units from which these aggregates are derived are thought to be the site of origin of many human mammary carcinomas; two of the PAHs studied, 7,12-DMBA and BaP, are known to be relatively potent mammary carcinogens in rats, whereas benz[a]anthracene is not a mammary carcinogen in rats. Tissues from eight patients were studied. The extent of metabolism of the PAHs is summarized in Table 6–4. There was considerable individual variation in PAH metabolism among the subjects studied, but the formation of water-soluble metabolites by the tissue samples accounted, in each instance, for a major portion of the total of each PAH metabolized. The extent of binding of each PAH to cellular DNA and proteins also varied considerably. Interestingly, the extent to which ³H-labeled metabolites of benz[a]anthracene—a noncarcinogen for mammary tissue in the rat—were bound seemed, from—the limited data obtained, to be consistently lower than the binding displayed by the other two PAHs. The results of chromatographic characterization of PAH-DNA adducts formed suggested that, with BaP, the hydrocarbon was activated by the cultured cells through the formation of anti-BaP 7,8-diol-9,10-oxide, a bay-region diol-epoxide that appears to be responsible for most of the nucleic acid adducts formed in several other biologic systems. The situation was less clear with 7,12-DMBA, although a portion of the adducts formed with this PAH cochromatographed with adducts present in a DNA hydrolysate that had been treated with anti-7,12-DMBA 3,4-diol-1,2-oxide—a derivative that is also classified as a bay-region diol-epoxide. The authors interpreted their data with caution, considering all the factors known to bear on the development of mammary cancer; but the possibility of partial causal relationships among the PAHs, their metabolic transformations, and tumor stimulation is implicit in this work.

TABLE 6–4

Metabolism of Benz[a]anthracene, 7,12-Dimethylbenz[a]anthracene, and Benzo[a]pyrene by Human Mammary Tissue.

Stampfer and colleagues¹⁶⁸ did similar studies with BaP and cultured mammary epithelial cells and fibroblasts. They showed that the breast epithelial cells were 50–100 times more sensitive (growth inhibition) to BaP than the fibroblasts; that the epithelial cells formed adducts as early as 6 h after addition of the PAH to the cultures; and that the adducts between the 7R anti stereoisomer of BaP diol-epoxide and deoxyguanosine predominated at all times and, with two minor adducts that were consistently present, persisted in the epithelial cells for at least 72 h in a BaP-free medium. No adducts were detected in fibroblasts until 96 h after exposure to the PAH, at which time the type and extent of adduct formation were similar to those observed with epithelial cells. As with the report of Grover et al.,⁶⁶ caution concerning the direct relation of these findings to the role of PAHs in mammary carcinogenesis is necessary. On this matter, Stampfer and co-workers¹⁶⁸ stated, however, that “chemical carcinogens, particularly BaP, should not be minimized as possible factors in the initiation of breast cancer.”

Mass et al.¹¹³ studied 26 specimens of normal human endometrium to determine the patterns of metabolism of [³H]BaP in short-term explant cultures. Three of the tissue samples were from postmenopausal women; of the remaining 23, it was possible to approximate the stage of the menstrual cycle at which the tissue was removed during surgery. Eight of the latter subjects were smokers. In summary, it was clear that normal human endometrium could enzymatically convert BaP to a wide variety of oxygenated derivatives that cochromatographed with dihydrodiols, quinones, and monohydroxy products of the PAH; sulfation was also identified. HPLC analysis of metabolites revealed marked individual variation in metabolite formation among the subjects studied; smoking did not account for this difference, but some evidence of hormonal influences on the patterns of PAH metabolism was adduced.

In a study by Dorman et al.,⁵² BaP binding to DNA in human endometrial tissue was studied in samples obtained from 41 subjects and, again, a striking (70-fold) range in the observed specific activities of carcinogen binding to DNA was identified (see Figure 6–1). Tissues obtained late in the proliferative phase or early in the secretory phase of the menstrual cycle had the highest mean specific activity of PAH-DNA binding (Table 6–5). Binding was significantly reduced when tissue specimens from low-estrogen periods of the menstrual cycle were studied. The reason for this apparent association between estrogen content (actually, the estimated phase of the cycle) and PAH-DNA binding is obscure, but clearly merits further study. Such study would have to deal with the important confounding factor of the broad range of individual variation in binding, which may mask systematic but small changes that can occur during a menstrual cycle, but which cannot now be detected.

FIGURE 6–1.

Spectrum of specific activities of binding of [³H] to DNA in human endometrial tissue in vitro. Human endometrial tissue was incubated for 18 hr in organ culture in medium containing 1 M[³H)BP. For each of the 41 specimens of endometrial tissue examined, (more...)

TABLE 6–5

Binding of [³H]Benzo[a] pyrene to DNA in Human Endometrial Tissue Taken Throughout Menstrual Cycle and Before and After Menopause.

Namkung and Juchau¹²⁸ studied the oxidative biotransformation of BaP in preparations of human placental microsomes with HPLC. The investigations revealed that the use of substrate concentrations high enough to ensure zero-order reaction kinetics markedly inhibited the formation of dihydrodiols in the reaction mixtures. The relative quantities of dihydrodiols generated increased with decreasing substrate concentrations between 200 and 2.7 µM. Addition of manganese or ferric ions to reaction mixtures altered the ratios of generated phenols to dihydrodiols. Identical results were obtained with ¹⁴C-and ³H-labeled BaP as substrate. The data suggested that considerable amounts of 7,8-dihydroxy-7,8-dihydro-BaP, a proximate mutagen-carcinogen, may be generated in vivo by placental tissues of women who smoke.

The formation of PAH metabolite-nucleoside adducts when human tumor placental microsomes were incubated with [³H]BaP and salmon sperm DNA has been studied by Pelkonen and Saarni.¹³⁹ There were significant differences between the PAH metabolite patterns and the nucleosidemetabolite complexes formed, compared with rat liver, for example. Specifically, in the human placenta microsomes, the absence of the nucleoside complex of 9-hydroxy-4,5-oxide implied the inability of this tissue to form 4,5-oxides of BaP. Indirect evidence of epoxide hydratase activity in placental tissue was obtained. The extent of PAH-DNA binding in this tissue correlated significantly with both 7,8-diol metabolite formation and fluorometrically determined AHH activity. The question of whether the 7,8-diol-9,10-epoxide of BaP is formed by the human placenta in vivo could not be answered unequivocally, but the authors' inferential conclusion is that it is probably formed in the human host. The interplay of possible genetic influences and clearly established regulatory influences of environmental factors on human placental AHH has been incisively discussed by the same group.¹³⁸

Cigarette-smoking has been shown by Conney and associates^42,189,190 to be one of the most potent and consistent inducers of human placental AHH activity yet identified. In the initial report of the group,¹⁸⁹ the enzymatic hydroxylation of BaP could not be detected in nonsmokers in homogenates of placentas frozen immediately after birth and studied within 48 h. In contrast, the enzyme activity was present in all 11 placentas from women who smoked during gestation, although enzyme activity in this small group did not correlate with the number of cigarettes smoked. BaP administration to pregnant rats also was shown to induce AHH activity in the placenta. The effect was related to PAH dose. This study constituted the first demonstration that compounds in cigarette, smoke could induce a carcinogen-metabolizing enzyme in human tissues. These studies were extended¹⁹⁰ to related enzymatic reactions in human placentas and to other types of pyrenes as probes for AHH-inducing activity in rat placenta (see Table 6–6). Extremely active inducers included chrysene, 1,2-benzanthracene, pyrene, 3 ,4-benzofluorene, and a number of related compounds.¹⁹⁰ The wide variability in the induction of AHH activity in human placentas is exemplified by the data in Table 6–7—a range in activity of the enzyme in smokers approaching 1,000-fold (a nearly 2,000-fold range if smokers are compared with nonsmokers). The basis for this extreme range of responses to a chemical exposure (15–20 cigarettes/d for each subject) is not known. However, data presented by Harris et al. ^70a suggests that pulmonary alveolar macrophages can metabolize BaP to proximate and ultimate mutagens released into extracellular space.

TABLE 6–6

Effect of PAHs in Cigarette Smoke on Benzopyrene Hydroxylase Activity in Rat Placenta.

TABLE 6–7

Variability in Induction of Benzopyrene Hydroxylase Activity in Human Placenta.

LUNG

The respiratory tract comprises an extremely disparate and complex set of tissues containing some 40 different cell types.¹⁶⁶ As Devereux et al.⁴⁷ have noted, whereas pulmonary cytochrome P-450 and the metabolism of xenobiotics have been studied with various preparations of lung tissue (microsomes, isolated perfused lung, cells obtained by pulmonary gavage, direct instillation of xenobiotics in various portions of the respiratory tract, etc.), little is known about the localization of the cytochrome P-450 monooxygenase components in the pulmonary system. This section deals exclusively with the metabolic properties of human respiratory tissues with respect to PAH metabolism, but the lack of information just cited needs to be kept in mind. There are facets of the investigation of Devereux et al.⁴⁷ in rabbits that probably bear significantly on problems of human pulmonary tissue biotransformations that depend on cytochrome P-450; these aspects include the observation that the alveolar macrophage that accumulates PAH has little or no measurable cytochrome P-450 or monooxygenase activity^58,71,148 and that there is selective cellular distribution of cytochrome P-450 species.

The ability of human bronchial epithelial cells to bind and presumably to activate such PAHs as 7,12-DMBA, 3-MC, BaP, and dibenz[ah]anthracene was described by Harris and colleagues in 1974.⁷⁰ Four tissue samples were studied (one control and three lung cancer) in explant cultures, and radiolabeled PAHs were used; radioactivity from all four compounds tested was found in both cytoplasm and nuclei and in all tissue samples studied (see Table 6–8). The number of tissues examined precluded comparisons between normal and tumorous lung PAH metabolism, and no studies of PAH-DNA adduct identification were carried out, although, as noted, radioactivity from the labeled PAHs was found tightly bound to DNA isolated by CsC1 gradient. A more detailed study by this group¹⁹⁵ used tissues obtained from an additional four patients, three of whom had pulmonary malignancy.

TABLE 6–8

Specific Activities of Binding of Tritium-Labeled PAHs to Human Bronchial DNA.

Explants of human bronchi also metabolized BaP and released derivatives that are mitogenic in the Chinese hamster V-79 cell line.⁷² The 7,8-diol of BaP was approximately 5 times more potent as a promutagen than the parent PAH; binding of the diol to DNA was 5–20 times greater than that found with BaP. When 13 samples of bronchial cells were studied with cloned Chinese hamster V-79–4A cells, a positive correlation between DNA-PAH binding (in the cultured bronchial cells) and induction of 0^r (ouabain-resistant) mutants was found, but no correlation between this mutation frequency and AHH activity was identified. This may be attributable, as the authors noted, to the difficulty in correlating AHH activity with the consequences of the multistep pathway of metabolic activation for BaP. The individual variation in mutation frequency was 9-fold, and the variation in binding of PAH to DNA 5-fold. This important investigation pointed the way toward study of the metabolic activation of chemical carcinogens into promutagens and mutagens directly in differentiated epithelial cells derived from human tissues; and the human tissue-mediated mutagen assay opened the possibility of testing the hypothesis that people differ in mutagenic and oncogenic susceptibility to environmental chemicals, depending on individual capacity to activate and deactivate chemical procarcinogens. Autrup et al.¹² compared the metabolism of BaP by cultured tracheobronchial tissues from humans and four other species (mice, hamsters, rats, and cows). They provided evidence that the metabolism of BaP is qualitatively similar in tracheobronchial tissues from humans and from animal species in which PAHs have been shown experimentally to be carcinogenic.

A similar study limited to a comparison of human lung microsomal fractions and rat microsomes was carried out by Prough et al.¹⁴⁴ The results indicated that human microsomes form a higher percentage of dihydrodiol products from BaP than do rat microsomes. The wide variation of PAH metabolite profiles formed by the 15 samples of human lung studied may be due in part to differences in clinical diagnosis when the samples were obtained. Bronchial tissues cultured in a chemically defined medium were exposed to radiolabeled BaP or its metabolites, and their binding to DNA was measured. Radiolabeled metabolites were prepared by incubating the parent PAH with rat liver microsomes and then purifying and identifying with silica gel and HPLC. The binding data showed that (−)-trans-7,8-diol bound to bronchial mucosal DNA to a considerably greater degree (5- to 23-fold) than did BaP; binding was also much greater (25- to 80-fold) than with the (−)-trans-9,10-diol. The trans-7,8-diol constituted 3–6% of the total identified metabolites when human bronchi were exposed to BaP. Diol-epoxides were formed from (−)-trans-7,8-diol in two of the bronchial explants, and strong evidence was provided that the major tumor bronchial mucosal DNA-binding BaP metabolite is in fact derived from (−)-trans-7,8-diol.¹⁹⁵ The specific adducts formed between DNA and the metabolic intermediates of BaP were not isolated, but the author concluded that the predominant bound metabolite is a single enantiomer of diol-epoxide I derived as indicated above.

In an extension of their earlier work, Harris and colleagues⁶⁹ examined the metabolism of BaP and 7,12-DMBA in explants of human bronchus and made a metabolic comparison with human pancreatic duct explants. As in the prior study, both normal and malignant human bronchi (37 subjects) metabolized BaP actively and in generally similar fashion, except for a higher percentage of organic-solvent-extractable metabolites formed by bronchi from noncancer patients. In addition, prior exposure of the bronchial explants to benz[a]anthracene altered the qualitative features of the metabolite profile of BaP, as analyzed by HPLC. Benz[a]anthracene specifically increased the binding of BaP to cellular DNA and the activity of AHH. Among a group of 28 of the patients' tissues studied, 7,12-DMBA was bound to DNA more often (26 of 28) than BaP. In the comparison with pancreatic duct explants, 7,12-DMBA-DNA binding was consistently lower in the latter tissue than in the bronchial explants.

Cohen et al.³⁴ showed, with cultured human bronchial epithelium, that BaP was converted promptly to metabolites that cochromatographed with 9,10-dihydro-9,10-dihydroxy-BaP and 7,8-dihydro-7,8-dihydroxy-BaP. Similar results were obtained with human lung cultures, except that a major metabolite, benzo[a]pyrene-3-yl hydrogen sulfate, was identified. The biologic activity of this sulfate ester of 3-hydroxy-BaP is of interest, because, owing to its physicochemical properties, it could be extremely persistent in man.

Covalent adducts between DNA and BaP in treated cultured explants of peripheral human lung tissue and in the continuous human alveolar tumor cell line were identified by Shinohara and Cerutti.¹⁶¹ From the chromatographic analysis of digests of DNA extracted from these tissues, it was concluded that both the lung specimens and the human alveolar tumor (A549) cells metabolized BaP to diastereomeric 7,8-dihydroxy-9,10-epoxytetrahydro-BaP intermediates that mostly reacted with the exocyclic amino groups of deoxyguanosine to form N²-(10-[7β,8α,9α- and 9β-trihydroxy-7,8,9,10-tetrahydro-benzo[a]pyrene]yl)deoxyguanosine (dGua-BaP I and II). Although comparable amounts of dGua-BaP I and II were formed in A549 cells, dGua-BaP I was the predominant adduct in the DNA of lung specimens from six different donors.

The wide range of metabolic capacities for PAHs exhibited by other human tissues studied also extends to lung tissue, as shown by Cohen et al.³⁵ They observed a 44-fold variation in the ability of short-term organ cultures of peripheral lung tissues from human cancer patients to metabolize BaP to organic-solvent-soluble derivatives. The total amounts metabolized ranged from 1% to 96.2% in a 24-h culture period. The authors concluded that, although caution must be exercised in measuring metabolic activities of human tissues derived from diseased patients, the use of short-term organ explant cultures mimics the in vivo metabolic disposition of PAH better than the use of lymphocyte AHH activity would. A solution to the practical problem of obtaining lung tissue from large populations to study the validity of this conclusion is not apparent.

Kahng et al.⁷⁸ concluded from a study of 11 immediately autopsied subjects that bronchial tissue exposed to benz[a]anthracene produced induction responses of AHH that correlated with induced AHH activity in monocytes from the same subjects. A reconfirmation of the wide range of individual differences in AHH activity of surgically obtained specimens of normal lung tissue (86 subjects) came from a detailed study by Sabadie et al.¹⁵³ Briefly, AHH activity was lower than normal in tumorous lung sections in 73 of the 86 patients; and in 22 tumor tissue samples, no AHH activity was detected at all. Individual variation (excluding the 22 subjects) in lung-tumor AHH activity was 20-fold, which approximated the variation observed in other studies, including those in which PAH-DNA binding and pulmonary tract tissues were studied. BaP metabolite formation was analyzed, and the results generally conformed with the data of other investigators.

Interestingly, BaP (but not pyrene) induces AHH and prolyl hydroxylase activity in neonatal rat lungs in organ culture.⁷⁴ Because prolyl hydroxylase is an indicator of collagen synthesis and increased activity of this enzyme in lung reflects increased collagen formation, the authors, Hussain et al., hypothesized that the earliest events in BaP-induced lung injury may include alterations in collagen metabolism. In a study of the effect of tobacco-smoke compounds on the plasma membrane of cultured human lung fibroblasts, Thelestam et al.¹⁷⁵ examined 464 compounds, of which nearly one-fourth gave rise to severe membrane damage. PAHs proved inactive in this test system; the PAHs tested included anthracene, benz [a]anthracene, chrysene, pyrene, BaP, perylene, fluoranthene, and coronene. The significance of these findings is not entirely clear, but, inasmuch as very large concentrations of the compounds were used (25 mM), the failure of all PAHs tested to cause substantial release of the radiolabeled nucleotide material from the cells suggests that PAH entry into cells of organs in which their carcinogenic potential is expressed does not require as an initial event plasma membrane damage by the active chemical species.

Lung damage by ozone⁵⁹ and nitrates¹⁸³ showed contradictory effects: in the former case, adaptation may become apparent, and, in the latter, susceptibility to infection may increase. In the case of asbestos-produced damage, as well as damage produced by other particles—such as iron oxide, silica, and carbon black—cellular uptake and availability of BaP increase.^97–99 Asbestos, of the several particles tested, was particularly effective in increasing microsomal uptake of the PAH, although clearly adsorption of the PAH on the particles—rather than simple mixture of the two—is required for the increase in cellular uptake to become evident. The relevance of these findings to the phenomenon of particle-PAH cocarcinogenesis is clear.⁹⁹ BaP elution from typical soot from pollution sources, as well as from soot in lungs (11 cases), has been carefully studied by Falk et al.⁵⁶ Strikingly, this PAH could not be recovered from soot in human lungs without malignancy (Table 6–9), whereas the noncarcinogen pyrene could be identified (in much lower concentrations than expected). Adequate controls appeared to ensure that the disappearance of the carcinogenic PAH was a biologic phenomenon taking place in vivo; the authors concluded that elution must have occurred in the host through an undefined mechanism. In another study,¹⁷⁸ involving 21 bronchial carcinomas, a search was made for 12 PAHs in the tissues with chromatographic and fluorescence techniques. Only four of the 12 PAHs sought were found: BaP, fluoranthene, perylene, and benzo[b]fluoranthene (Table 6–10). BaP was found in all tumors; fluoranthene and benzo[b]fluoranthene were sometimes present, as was perylene. Coronene, dibenz[ah]anthracene, pyrene, benz[a]anthracene, chrysene, benzo[ghi]perylene, benzo[k]fluoranthene, and benzo[e]pyrene were, if present, below the limits of detection.

TABLE 6–9

Disappearance of PAHs from Soot in Human Lungs.

TABLE 6–10

PAH a in Human Lung Samples from Surgical Operations.

WORKPLACE EXPOSURE

Schenker in 1980¹⁵⁵ reviewed the question of whether diesel exhaust is an occupational carcinogen and summarized a number of the principal studies (Table 6–11) on the question of cancer incidence in populations of workers exposed to diesel exhaust. Data on environmental and occupational BaP and total suspended particles in various urban and rural sites and specific occupations were also provided (Table 6–12). These epidemiologic data emphasize the conclusion that “the carcinogenicity of workplace exposure to diesel engine exhaust is suggested…but the existing data are sparse and contradictory.” Table 6–11 shows only concentrations of BaP, and the values are in units of micrograms per 1,000 cubic meters. Because the air breathed by a normal adult approximates 15–20 m³/d, the highest PAH concentration shown indicates a potential exposure dose of about 700 µg/d in a work setting (coal and pitch-coking plant) known to have one of the most intense PAH exposures. This figure exceeds by orders of magnitude the exposure produced by the heaviest smoking, and such an occupational locale would thus be expected to elicit detrimental and clearly detectable health effects in man. The same consideration applies to the data on workers in gasworks retort houses and roof tarrers. But beyond these specific occupational sites, the respiratory intake of BaP—even if, for occupational purposes, a person had to remain for 24 h/d in Blackwall Tunnel, London (Table 6–11)—would approximate that from about a pack of old-style cigarettes per day. The improbability of such occupational exposures emphasizes the difficulty of measuring the health hazards of atmospheric PAH sources in the general sense (i.e., in the 28 rural and 24 urban sites depicted in Table 6–11).

TABLE 6–11

Environmental and Occupational Benzo[a]pyrene and Total Suspended Particulate Concentrations.

TABLE 6-12

Epidemiologic Studies of Cancer in Occupations Exposed to Diesel Exhaust.

A number of occupational-epidemiologic studies have emphasized the difficulties of reaching firm conclusions with respect to the direct (or measurable) health risks of PAHs in work environments, whether the suspected hydrocarbon comes from diesel or other automotive exhausts or from chemicals, such as petroleum sources, that are intrinsic in the occupation itself. Battigelli et al.¹⁸ studied 210 locomotive repairmen (average age, 50 yr; average work period, 10 yr) considered to be regularly exposed to diesel exhaust and 154 “control” railroad workers. The studies were carried out in two railroad shops in Pittsburgh, Pa. The clinical data were scanty, and it was not possible to differentiate the exposed from the nonexposed worker population on the basis of pulmonary-function tests. However, smoking clearly impaired the pulmonary functional performance of workers. A somewhat comparable environmental study carried out by El Batawi and Noweir⁵⁵ in two diesel-bus garages in Egypt raised the possibility of clinically detrimental, synergistic effects of smoke and acrolein gas, which is known to be present in exhaust of diesel engines. Ventilatory-function changes over a workshift in coal miners exposed to diesel emission were studied by Reger et al.¹⁴⁷ the only positive finding in this study of 800 men was that smokers suffered consistently greater pulmonary-function decrements over a workshift than nonsmokers. In a retrospective study of mortality statistics,⁸³ Kaplan could identify no higher than normal rates of death from bronchopulmonary carcinoma in workers exposed to fumes from diesel engines among the medical records of 6,500 deceased railroad workers, including 818 deaths from malignant diseases.

Lloyd et al.¹⁰⁹ reported that the mortality from respiratory cancer for men employed in a coke plant was twice the rate generally observed among steelworkers; the whole difference was accounted for by a threefold excess for nonwhite workers. A more detailed analysis¹⁰⁸ showed the following: The excess of respiratory cancer previously reported for coke-plant workers was limited to men employed at the coke ovens, the relative mortality for this disease being 2.5 times that predicted. The greatest part of the excess was accounted for by an almost fivefold risk of lung cancer in men working on the tops of the coke ovens. A 10-fold risk of lung cancer was observed for men employed 5 yr or more at full-time topside jobs; 15 lung-cancer deaths were observed among the 132 men in the topside group, compared with 1.5 expected. The apparent differential in respiratory-cancer rates for white and nonwhite coke-plant workers reported in an earlier paper was accounted for by differing distributions by work area and the unusually high lung-cancer risk for topside workers; lung-cancer mortalities for white and nonwhite coke-plant workers employed at work stations other than topside were comparable. A deficit of deaths from heart disease, previously reported for similar occupational groups, was also seen for coke-oven workers. Coke-plant workers employed only in nonoven areas may be at excess risk of digestive cancer.

A review of the literature on cancer mortality of men employed in the coal-tar industries showed that all these occupations evidence excess cancer at one or more sites. The lung-cancer excess in coke-oven workers also was observed in other groups engaged in coal carbonization, and it appeared that the lung-cancer response was positively correlated with the temperature of carbonization.

Among coke-oven workers studied by Mazumdar et al.,¹¹⁷ excessive deaths from respiratory malignancy were reported. As in the study of Lloyd et al.,¹⁰⁹ there was a tendency for the death rates of nonwhite workers to be higher than those of white workers. Measured concentrations of coal-tar pitch volatiles in the environment of men who worked at the top of coke ovens were 2–3 times higher than in that of men employed at the side of the ovens. High BaP emission, among others, has been measured in the gaseous discharge—including the coal-tar pitch volatiles—of coke ovens in the steel industry, a rough estimate being that 1.8 g of this chemical is emitted per ton of coke produced.¹¹⁷ As in the Lloyd et al. study,¹⁰⁹ the overall cancer-death risk for coke-oven workers was distinctly higher than that for normal persons in the age group over 55 yr, and the age-adjusted death rates for lung cancer showed a strong relationship between extent of exposure to coal-tar pitch volatiles and lung-cancer mortality. The lowest-exposure group¹¹⁷ had death rates similar to those of nonoven workers, but all higher-exposure groups had age-adjusted rates that ranged from 3 to 10 times those of the comparison group with increasing exposure. The data in this study also confirmed the long latency in cancer formation, even under the conditions of high exposure to carcinogens characterizing coke-oven workers; the time between first exposure to coal-tar pitch volatiles and death from lung cancer varied from 10 to 40 yr, with an average of 25 yr.

Toxicologic experience with workers in the developing shale-oil industry is incomplete, although historical evidence indicates that potential health hazards related to malignancy may exist in the processes involved in oil extraction.¹⁸⁷ Some data on the content of BaP and pyrene analogues from shale materials, as reported by Weaver and Gibson,¹⁸⁷ are useful to record here (Tables 6–13 and 6–14). Because the industry is still in its developmental stage in this country, the overall health impact that may be attributed to exposure to these PAHs—as well as to other contaminants, such as arsenic, beryllium, cadmium, lead, mercury, and nickel¹⁸⁷ —is difficult to estimate.

TABLE 6–13

Benzo[a]pyrene Content of Shale Materials, µg/kg (ppb).

TABLE 6–14

PAHs in Shale Retort Oils, ppb.

EXPOSURE TO PAHs VIA THE GASTROINTESTINAL TRACT

The exposure of humans to PAHs may be considered to be almost exclusively via the respiratory and gastrointestinal tracts. Some occupational groups (e.g., the chimney sweeps studied by Pott) may have an intense local cutaneous exposure to PAHs, but the significance of percutaneous absorption of these compounds for the general population is not known. Such substances as the polychlorinated biphenyls²² and constituents of coal tar²¹ can pass through the skin and induce liver oxidative enzymes in animals, so it is possible for some (undoubtedly small) degree of PAH accumulation to occur in humans systemically via skin exposure.

Several major reviews of the importance of water and food as vehicles of human exposure to PAHs have been published in the last 5 yr. These include a special issue of the Journal of Environmental Pathology and Toxicology ¹⁵⁴ devoted to the health aspects of PAHs and several monographs focusing on PAHs in drinking-water sources and on PAHs in the marine environment.^19,129,181

PAHs in Water

It can be stated at the outset that human exposure to PAHs through the ingestion of water is quantitatively insignificant, compared with exposure through food—the contribution of drinking water is estimated to be only about 0.1% of the total PAH derived via the gastrointestinal tract in humans.³ This estimate, carrying with it an implicit assumption of relative biologic safety (at least compared with foods as a source of PAHs), is probably valid except perhaps for some surface-water sources, which, because of location (e.g., downstream from shale-oil effluent or coke-byproduct discharge sites—see Table 5–12 of Santodonato et al.¹⁵⁴ ), may be heavily contaminated by such PAHs as BaP. Groundwater concentrations of this prototype PAH determined in multiple German and American sources are extremely low (see Table 5–11 of Santodonato et al.¹⁵⁴ ), ranging from a fraction of a nanogram per liter to several nanograms per liter. The average “total” PAH content is, of course, greater, but still in the same range. In contrast, low- to medium-concentration contaminated surface waters may contain PAHs 5–20 times higher, and this pollution may be increased by several orders of magnitude in sewage water or in surface waters adjacent to industrial sites. Treatment of surface water to obtain drinking water can nevertheless remove the bulk (95% or more) of the PAHs, particularly with activated-carbon filtration. This reflects the fact that much of the PAH in water subject to pollution is quickly adsorbed on suspended solids or is found in sedimented particulate matter. The majority of PAH entering surface water is concentrated locally; although PAH can probably be considered ubiquitous in water, the amounts involved are substantially lower than those found in air or on land. Neff¹²⁹ has pointed out that, if all PAHs found in the aqueous environment were distributed evenly throughout the oceans and fresh-water bodies, they would be undetectable and inconsequential.

As noted, the PAH content of drinking water is, with an occasional exception, low, as expressed as BaP and total PAH (Table 6–15).¹⁵⁴ Among the general class of PAHs, the compounds that have been detected by high-resolution gas chromatography after extraction from tapwater¹³² are listed in Table 6–16 with their concentrations. Such contamination at a typical, most proximate (tapwater) drinking-water source represents only trace contamination, compared with the PAH content of original fresh-water sources, marine and estuarine waters, fresh-water and marine sediments, and some alcoholic beverages.¹⁷⁹

TABLE 6–15

PAHs in Drinking Water.

TABLE 6–16

PAHs in Tapwater.

The occurrence of PAHs in saltwater sources has for several reasons more potential biologic importance than the occurrence of these compounds in drinking water. The oceans provide a very large surface area for deposition of airborne PAHs via rain and dry fallout. Runoff of PAHs from the land surface also contributes substantially to marine-water content, as do direct effluents from sewage and industry. Carcinogenic PAHs occur in crude and, particularly, refined oils,³ and oil spills may contribute in a major way to marine pollution with these compounds, especially on a local scale. The oceans constitute an ecosystem in which varied animal and plant life can participate in the metabolic processes involved in the uptake, storage, concentration, biotransformation, and discharge of PAHs. Thus, the consumption of fish and shellfish of predominantly saltwater, compared with fresh-water, origin (88% vs. 12% of the seafood in the diet) gives special importance to the PAH contamination of the aquatic environment that these food species inhabit.

PAHs are universally, although unevenly, distributed throughout the marine (saltwater) environment. They are derived principally from atmospheric fallout, terrestrial runoff, and spills of petroleum products. The contribution, if any, of marine organisms to PAH pollution by de novo biosynthesis is unknown. Total PAH entry into the marine environment from petroleum spills is estimated at 17×10⁴ tons/yr, of which BaP would constitute 20–30 tons/yr.¹²⁹ Conservative figures for the total world contribution of industrial and domestic wastewaters to marine pollution with PAH have been estimated to be BaP at about 29 tons/yr and total PAH at 4.4×10³ tons/yr. For terrestrial runoff, the figures are about 118 tons/yr and about 2.9×10³ tons/yr, respectively, and for atmospheric fallout, 500 tons/yr and 50,000 tons/yr. Because the composition of total PAH in these sources varies considerably, it has been suggested¹²⁹ that the figures estimated for BaP input provide a better index of the potential carcinogen input from these sources than do the figures for total PAH.

The majority of PAH in the aquatic environment remains near the point source of contamination and thus is concentrated in coastal waters; here, the bulk of the PAH is in bottom sediments and to a lesser extent in suspended solids or solution. The water solubility of carcinogenic PAHs is very low, but solubilization may be increased by the concurrent presence of detergents and other organic substances. Photodegradation of PAHs in the marine environment can occur variably, depending on the depth and turbidity of water and other factors; but persistence of PAHs is much greater in water than in air, because the particulate matter on which these compounds are mostly adsorbed provides a storage pool from which they may be slowly returned to the water by leaching or through biologic processes involving marine organisms.

The characteristics of marine pollution by PAHs are such as to suggest the occurrence of multiple varieties of discrete ecosystems with relatively high concentrations of these compounds in sediments and local plant and animal species—all existing in a vastly larger aquatic environment characterized by a smaller degree of PAH contamination. In the local marine areas of high PAH pollution—principally river basins and estuarine and coastal waters—the degree of PAH contamination and the PAH composition in water, sediments, and nonmigratory marine life are determined by the nature of the point sources of contamination. In the organisms found in these areas, the PAH composition depends on metabolic processes related to the selective bioconcentration, biotransformation, and accumulation of the PAHs or metabolites or their discharge into the aquatic environment.

The fate of PAHs in marine ecosystems has been studied by Lee et al.,¹⁰¹ who used as a model Prudhoe crude oil enriched with a number of PAHs dispersed into a controlled ecosystem (polyethylene enclosure 2 m wide and 15 m deep) suspended in Sadnich Inlet, Canada. The oil was estimated to contribute PAHs at concentrations ranging from BaP at 100µg to naphthalene at 300×10³ µg per 100 g. Multiple water and sediment sampling, microbial-degradation studies, analysis of bioaccumulation by oysters, and analysis of adsorption to sediments with [¹⁴C]PAH were carried out. The results demonstrated a rapid, marked decrease in PAHs from water (half-life, 3–4 d) and a variable recovery, depending on the PAH, in the sediment. For the low-molecular-weight PAH naphthalene, this recovery was only 7% after 1 wk; for BaP, it was 39%. Oysters rapidly took up all PAHs, but released naphthalene to such an extent that it was not detectable in the organisms after 23 d. In contrast, benz[a]anthracene and BaP were released much more slowly, with estimated half-lives (assuming exponential discharge) of 9 and 18 d, respectively. Thus, the higher-weight PAHs persisted much longer in the organisms than the lower-weight PAHs. Other degradation studies involving mussels collected from oil-contaminated waters also have shown the persistence of the higher-molecular-weight PAHs.^51,53 Evaporative loss of lower-weight PAHs, such as naphthalene, in the upper waters was expected, whereas this would be limited for higher-weight PAHs. Microbial degradation of naphthalene and anthracene was measurably increased in oil-contaminated water, compared with control water (4 h vs. 48 h, respectively, for appreciable degradation)—a finding consistent with those of other studies showing higher numbers of oil-degrading microorganisms in polluted than in control or unpolluted waters.^6,186 Photochemical degradation of PAHs was inferred; for BaP, this was considered to account for an amount that could approximate about 50% of the compound, inasmuch as no microbial degradation of the compound was demonstrated and 40% was recovered in the bottom sediment. The study permitted several conclusions that probably have general relevaace. The half-lives of PAHs in marine waters are short (a few days); for lower-weight PAHs, microbial degradation and evaporative loss may be primary removal processes; for higher-weight PAHs, such as BaP, sedimentation and photodegradation are the most important removal means; and, by inference, for higher-weight PAHs after sedimentation, biologic degradation and interaction between plant and animal life in the sediment are important factors in removal.

These processes (biologic degradation and interactions) have been extensively studied with a wide variety of aquatic species. It is clear that, as with terrestrial fauna, the capacity of marine animal species to effect the metabolic transformation of PAHs can be considered to be universally distributed. Reviews of the results and other aspects of such studies have been published elsewhere,^{3,4,37,38,129,181} and only representative reports are summarized here. PAHs in the marine environment can be metabolized by aquatic bacteria and fungi;¹²⁹ for some species of bacteria, a monocyclic aromatic hydrocarbon, such as benzene, can serve as a sole carbon source. PAHs, such as BaP and benz[a]anthracene, can also be oxidatively metabolized to hydroxylated derivatives comparable with those produced in the livers of vertebrates. PAHs can be degraded to CO₂ to a considerable degree (13–68%¹²⁹ ) by aquatic microorganisms. PAH metabolism by fungi also occurs; these organisms contain cytochrome P-450 and can carry out the initial oxidative metabolism of PAHs in a manner resembling that catalyzed in vertebrate liver. Marine fungi isolated from oil-polluted water or oil slicks have a substantial ability to assimilate petroleum hydrocarbons, and this hydrocarbon-degrading capacity can permit use of a PAH as a growth substrate.¹

Fish and crustaceans (and some worms) can oxidize PAHs—as measured by AHH activity—and cytochrome P-450 has been identified in a number of these species. Most oxidative metabolism in these aquatic animals is in the liver, as it is in mammals. Induction of cytochrome P-450 (not always correlated with an associated P-450-dependent increase in chemical oxidation) in fish has been produced by benz[a]anthracene, chrysene, BaP, and other organic substances^{33,61,129,140,169} to which fish may be exposed in their natural environments or under experimental conditions. The products of the oxidative metabolism of PAHs in fish resemble those produced in mammalian liver and include diols, epoxides, phenols, quinones, and all principal types of conjugates formed from PAH metabolites in mammalian liver.

Seasonal changes in P-450-dependent oxidation have been reported in fish,⁴⁹ and alterations in this enzymatic activity have been related to ambient temperature, food status, and exposure to inducing chemicals in their natural habitat.^48,49 Apart from carrying out biotransformation, the capacity of marine species to accumulate and discharge PAHs from the surrounding waters is important in relation to the pattern of distribution of these compounds in the marine environment and to the use of marine species as food, in view of their contribution to the exposure of humans to PAHs via the gastrointestinal tract.

Marine animals readily accumulate PAHs from the surrounding waters and can discharge both the untransformed PAHs and their metabolic products into the aqueous environment. The rates of release of accumulated PAHs may vary substantially from species to species (and compound to compound), and half-lives can range from hours to many days. The substantial concentration gradients of PAHs that may occur between an organism and its aqueous environment can have importance for man in relation to marine species that are eaten by man or by edible species.^44,102,193 Whether these concentration gradients involve an active uptake mechanism is not known; but they do not depend solely on solubility, inasmuch as polar metabolites of a PAH can be retained longer than the more lipophilic parent compound.¹³ This may be due to the electrophilic nature of these metabolites and their consequent binding to tissue macromolecules.¹²⁹

Oysters have been shown to concentrate hydrocarbons from diesel-oil-contaminated waters to concentrations over 300 µg of total hydrocarbons per gram of wet weight over a 7-wk period.^170,171 These hydrocarbons were rich in aromatics, compared with the contaminating oil. In clean seawater, the hydrocarbon concentrations decreased dramatically (by 90% in 4 wk). Other marine species show the same biologic characteristics, although uptake and release of accumulated hydrocarbons vary. The concentration factor (i.e., tissue vs. water concentration) may reach 1,000-fold²³ in marine animals that cannot escape a contaminated environment. The potential importance for humans of this capacity for bioaccumulation in edible marine species is evident. PAHs can, as expected, accumulate rapidly in fish from contaminated sediments, as McCain et al.¹¹⁸ have shown, although this process is less efficient than uptake from water.

The biologic impact of contaminating PAHs on marine species has been thoroughly reviewed recently^{37,38,46,93,129} and is not summarized here. Toxic effects of these and related pollutants have been described across the spectrum of marine life, from bacteria and fungi to plants and animals; and they range from the “tainting” of commercial species^37,38 to the development of cancer and cancer-like growths in aquatic animals.^46,93

PAHs in Food

The exposure of humans to PAHs from dietary constituents greatly exceeds that from any other sources except specific hazardous occupational settings. PAHs are ubiquitous contaminants of foods and— depending on the extent of atmospheric and soil pollution in crop areas and on methods of processing, preservation, and preparation—can become highly concentrated in selected foodstuffs. At least 100 types of PAHs have been identified in foods.²⁰¹ Some of these have been shown to have well-defined carcinogenic properties in experimental animals. Epidemiologic studies have suggested an association between the consumption of high-PAH foods and gastrointestinal malignancies in selected populations,^106,165,177 but it is difficult to extend this association to the general population or to define the biologic risk of PAHs in foods in more direct terms. Nevertheless, the quantitative dimensions of PAH exposure via the diet and the established carcinogenic potential of some of the compounds frequently identified in foods suggest that the health risks from this source of exposure, although still incompletely defined, may be important for various groups.

Edible marine species may contain variable amounts of PAHs derived principally from polluted terrestrial runoff waters, from marine sediments, and from petroleum-contaminated aquatic environments. As noted above, such environments are largely in-shore (e.g., estuaries and river basins), with pollution diminishing rapidly in the open seas. Bioaccumulation of PAHs in the marine food chain may be substantial in some fauna, and, of course, national predilections for such modes of seafood preparation as smoking^{14,64,67,114,176} can increase to high values the content of PAHs in such foods. The potential for biomagnification of PAHs in aquatic food chains is clear, but the extent to which this process results in contamination of seafood ingested by humans is not known (the subject has been reviewed by Neff¹²⁹ ). For some crustaceans and fish, PAH uptake through the food chain can be more efficient than uptake from the surrounding waters,^44,102,193 and storage of such compounds in these species can be substantial. PAHs thus stored may or may not undergo extensive biotransformation. The processes of storage, uptake, metabolism, accumulation, and excretion have generally large interspecies variation; but crustaceans appear relatively efficient in their uptake of PAHs from food and other sources.¹²⁹ Table 6–17 shows an analysis of PAHs in oysters collected in a moderately polluted harbor area by Cahnmann and Kuratsune.³⁰ The comparative BaP and benzanthracene contents of a variety of foodstuffs are shown in Table 6–18. (Also, see Table 6–21 for similar information on benzo[e]pyrene, chrysene, and dibenz[ah]anthracene.) The extent of and striking variation in PAH contamination of marine species are evident in the data (Table 6–19) of Mix and Schaffer,¹²⁶ who examined BaP concentrations in mussels (Mytilus edulis) in Yaquina Bay, Ore., at multiple sites over a 2-yr period. The variations have a time component, geographic determinants, seasonal and environmental elements, and unknown biologic influences that make generalizations from such data extremely difficult and perhaps impossible. The BaP concentrations in mussels reported by this study exemplify, however, the extent to which marine species have the potential for representing a considerable exposure source of PAHs in the human diet.

TABLE 6–17

PAHs in Extracts from Shucked Oysters.

TABLE 6–18

PAHs in Foodstuffs.

TABLE 6–21

PAHs in Foodstuffs.

TABLE 6–19

Benzo[a]pyrene in M.edulis from Yaquina Bay, Oregon.

A variety of foodstuffs of terrestrial origin have been analyzed for PAH contamination, and many PAHs have been identified. They include the polycyclic compounds listed in Table 6–20, some of which have known carcinogenicity.¹²⁶ The known carcinogens 7,12-DMBA, cholanthrene, and dibenzo[ai]pyrene have also been identified in curing smoke. The relative concentrations of five carcinogenic PAHs in a sampling of foodstuffs are shown in Table 6–21.²⁰¹ It is clear from these data that amounts of some of these foodstuffs that are well within the amounts ingestible within a 1-d period constitute a PAH exposure via the gastrointestinal tract that can greatly exceed the pulmonary exposure of a very heavy smoker to PAHs.

TABLE 6–20

PAHs in Foods.

Large amounts of PAHs can be found in soils and can enter food crops from this source. Table 6–22 ²⁵ shows results of a sampling of soils in the Northeast for BaP. The concentration of PAHs in soil can vary over an enormous range; for the prototype compound BaP, Baum¹⁹ has summarized World Health Organization data showing a range (in micrograms per kilogram of soil) extending from around 100 (nonindustrial sites) through 1,000 (towns and vicinity) and 2,000 (soil near traffic) to 200,000 (soil near an oil refinery) or even over 600,000 (soil directly contaminated by coal-tar pitch). The higher figures reflect particle deposition, local atmospheric fallout, and direct waste discharge; the origin of the PAHs in forest samples (whose soil concentrations range up to 1,300 µg/kg¹⁹ ) is not certain, but must include a large contribution from natural combustion.

TABLE 6–22

Benzo[a]pyrene in Soils.

PAHs in food crops are probably derived in part from polluted soils, although the relative contribution of this source, compared with irrigation water or atmospheric pollution, is not established. PAHs in soils can be translocated to plants, probably through root adsorption, but the extent to which this occurs does not seem to correlate with the PAH content of soil.^89,159,184 Uptake of PAHS may also vary with plant species. The aboveground parts of edible plants can, of course, also concentrate PAHs through surface absorption from deposited dusts containing these compounds. Through this process, the aboveground parts of food crops can accumulate a gradient of PAH contamination exceeding that in root parts by a factor as high as 10,⁸⁹ and the bulk of this contamination in such edible crops as leafy vegetables (e.g., lettuce, spinach, and kale) and tomatoes cannot readily be removed by washing.¹⁸⁴ PAH contamination of irrigation wastes also contributes to an unknown extent to the contamination of edible plants. In the processing of foods, packing materials and additives are other sources of potential PAH contamination.

By far the largest sources of PAH contamination of foods are curing and preserving processes and cooking, especially of meats. Apart from shellfish, the “intrinsic” content of PAHs in most foods is low; for example, uncooked pork and beef may contain only 0.1 µg/kg. This concentration can increase substantially as a result of any cooking process (see Table 6–21) and especially as a result of smoking, curing, or broiling under a direct flame in which food drippings can be pyrolyzed.

PAH contamination of foods associated with smoke-curing results in part from the resinous condensates of liquid smoke flavors and from food combustion products.^{64,106,150,177,191} The type of smoke generation and other characteristics of the smoking process can influence the amounts and types of PAHs produced—e.g., the temperature of combustion, the air supply, the length of smoke ducts, and the density and temperature of the smoke-cure.¹⁷⁷ Domestic smoking clearly produces more PAH than the commercial process,¹⁴ probably because the procedure is less controlled and, as a result, entails heavier and more prolonged smoke exposure.¹⁷⁶ A general survey by the Food and Drug Administration and the U.S. Department of Agriculture of PAH content of smoked foods prepared commercially was reported by Malonoski et al.¹¹²

The broiling of meats over an open flame in which fat drippings can be pyrolyzed probably contributes more to diet-derived PAH exposure than any other method of food processing or preparation. Potent mutagens can be produced from amino acids and proteins in foods by high-temperature cooking.^{36,116,127,167,173,196} This mode of cooking also increases the carcinogenic PAHs in meats to very high concentrations.¹⁰⁶

The concentrations of 15 PAHs found in the outer surfaces of charcoal-broiled steaks by Lijinsky and Shubik¹⁰⁶ are recorded in Table 6–23. These concentrations are not unusually high for broiled or smoked meat (as the data in Table 6–21 indicate), nor for dietary constituents that are known to have a high PAH content, such as yeast. oils some leafy vegetables and fruits, roasted coffee, and teas.^24,64,65,201 PAHs formed by pyrolysis can be derived (at least with pure substrates) from carbohydrates, fatty acids, and amino acids, and the extent of their production depends on temperature.206 The data of Masuda et al.¹¹⁵ (Table 6–24) show the amounts of 19 PAHs formed from combustion of six potential substrates at 500 or 700C. Combustion took place in a nitrogen atmosphere; at 300C, no PAHs were formed from any of the starting materials, but at the highest temperature studied, large amounts were produced from each. Clearly, substantial quantities of PAHs can be formed from these substrates under the pyrolytic conditions used, and, although ordinary pyrolysis takes place in air, the substrates tested are common constituents of foods and common broiling temperatures are within the range of those used in this study.

TABLE 6–23

PAHs in Charcoal-Broiled Steaks.

TABLE 6–24

PAHs Produced by Pyrolysis of Carbohydrates, Amino Acids, and Fatty Acids at Two Temperatures.

The conditions of broiling heavily influence the amounts of PAHs produced. Fatty meat produces more PAH after broiling than lean meat, and it has been suggested¹⁰⁶ that pyrolysis of fats dripping onto red-hot coals is the most likely source of PAHs. PAH production in broiled meat clearly depends, in addition, on the closeness of the meat to the heat source, on whether meat drippings reach the heat source (i.e., heating from the top, rather than the bottom), and on whether cooking is quick at high temperatures or slow at low temperatures.^104–107 Toxins other than PAHs are also produced by high-temperature cooking; these include the mutagenic-carcinogenic amino acid pyrolysis products described by Japanese and American workers and the N-nitroso compounds formed in cured-meat products, especially bacon and ham.⁶⁴ It should be noted that these non-PAH substances can be produced at temperatures distinctly lower than those used in conventional broiling and that a large fraction of them may be volatile; thus, redeposition of these airborne substances and their inhalation during cooking are additional toxin exposures that can be related to the diet.¹⁴⁶

An approximate “balance sheet” of the estimated PAH exposure of humans from air, water, and food is shown in Table 6–25.¹⁵⁴ Despite a degree of inexactness in these figures—especially for foods—they provide a reasonable perspective of the sources of PAHs that might have an impact on man. It should be evident from these estimates that food constitutes the predominant source of PAHs for humans; even if the contribution from smoking were included, the diet would still be the dominant source.

TABLE 6–25

Estimated Daily Human Exposure to PAH from Air, Water, and Food.

The health impact of the PAHs in the human diet is not known, although, as noted above, an association between the intake of these substances in smoked foods and the occurrence of gastrointestinal malignancies in select populations has been inferred. The remarkably large amounts of PAHs that are ingested, compared with those to which the pulmonary system is exposed (even in heavy smokers), makes it clear that there must be tissue-specific factors related to the disposition of or metabolic responses to PAHs that protect the gut from the deleterious impact that might be anticipated from such exposure. The possibility of detrimental effects of diet-derived PAHs on the gastrointestinal system will not be so amenable to quantitation as has been the case with respect to smoking and the development of lung pathology.

An approach to defining the human metabolic impact of diet constituents in general and of charcoal-broiled meats in particular has been taken in the clinical-nutrition studies recently summarized by Anderson et al.⁵ Several dietary factors were shown to influence potently the oxidative metabolism of various drugs used as model substrates for cytochrome P-450- and cytochrome P-448-mediated chemical transformations. It has been shown that isocaloric substitution of dietary protein for carbohydrate substantially shortens the plasma half-times of such drugs as antipyrine and theophylline; i.e., a protein-enriched diet increases the oxidative metabolism of these compounds. Opposite changes were observed during periods of high-carbohydrate feeding. Substitution of protein for fat in the diet (a nonisocaloric change) also stimulates the oxidative metabolism of these drug substrates; however, neither high-unsaturated-fat nor high-saturated-fat diets produce alterations in drug oxidation distinct from those produced by high-carbohydrate diets alone. Thus, with respect to influences on microsomal mixed-function oxidases, carbohydrate and fat in the diet appear to be interchangeable.

Feeding rats charcoal-broiled beef is known to increase intestinal metabolism of phenacetin.¹³⁷ Increased oxidative metabolism of this drug, as well as of antipyrine and theophylline, was also observed in the test subjects after short-term feeding (2 portions/d for 4 d) of normal portions of charcoal-broiled beef at mealtimes.^39,84,136 The effect of broiling (in control diets, the beef was protected from the cooking fire with aluminum foil) was striking; during the test-diet period, there was a pronounced decrease in the mean plasma concentration of phenacetin and a comparable decrease in the area under the curve for plasma phenacetin concentration plotted against time. The ratios of the mean concentrations of metabolite and unchanged phenacetin at each point studied increased markedly during the charcoal-broiled-beef test period, compared with control periods. The findings suggested that charcoal-broiled beef greatly stimulated the metabolism of this model drug substrate in the gastrointestinal tract or during its first pass through the liver. Smaller, but still substantial, increases in antipyrine and theophylline metabolism during the ingestion of the charcoal-broiled-beef test diet were also observed.

These systematic and pronounced effects of specific dietary manipulations on the metabolism of model drug substrates by the cytochrome P-450-dependent mixed-function oxidase system provide a valuable means for defining the metabolic responses of both normal and ill subjects to the ingestion of various foodstuffs or foods prepared in various ways. The physiologic import of such clinical studies can be greatly extended by the judicious selection of suitable chemical substrates for the metabolic systems under investigation. The extent to which individuality in man characterizes specific chemical biotransformations can also be explored by these metabolic techniques. Finally, it may be possible through such clinical studies—in which each subject serves as his own control—to identify patterns of biologic responses to specific foods or food components that might otherwise be obscured by the genetic and environmental diversity of large population groups.