Chemical structure, reactivity, and carcinogenicity of halohydrocarbons.

This review summarizes studies concerning the covalent binding of [14C]TCE to rat liver microsomal protein and exogenous DNA, in vitro, the enhancement of this binding by inducers of mixed-function oxidases, and inhibition of binding by inhibitors of these enzymes. Furthermore, recent studies on this type of binding in various strains of mice and rats of both sexes and using microsomal preparations from various organs are briefly reviewed. Other work reviewed here concerns the synthesis of TCE epoxide and its reaction with nucleophiles since it is believed that TCE epoxide is the activated carcinogenic intermediate of TCE. The utility of structural prognostication of carcinogenic activity and the importance of considering possible metabolic pathways for other chlorinated olefins is also discussed.


Introduction
Since the time of the discovery of the carcinogenicity of vinyl chloride (VC) in laboratory animals (1,2) and humans (3) there has been an extensive renewed interest in and researches on halogenated hydrocarbons, both unsaturated and saturated (4)(5)(6).
The carcinogenicity findings with VC led to studies on its metabolism (7) and possible modes of action (5,8,9). Simultaneously studies were undertaken in a number of laboratories on other compounds in this class, some of which are also being examined in our laboratory.
Based on its chemical structure, reactivity and possible metabolic pathways, we predicted that trichloroethylene (TCE) will be carcinogenic, particularly to the liver (10,11). Studies were performed by us on the binding of [14C]TCE to rat liver microsomal proteins in vitro (12,13), as well as the synthesis of the epoxide of TCE and its reactions with nucleophilic trapping agents (14). More recently, we have also completed studies on the in vitro binding of ['4C]TCE to microsomal proteins and to DNA in mice and rats of various strains *Laboratory of Organic Chemistry and Carcinogenesis, Institute of Environmental Medicine, New York University Medical Center, New York, New York 10016. and both sexes and using microsomal preparations from various organs (15).
The studies outlined in the previous paragraph, as well as some as yet unpublished work from this laboratory, will be briefly reviewed in this report.

Structural Prognostication of Carcinogenicity
Past studies on structure-activity correlations in this laboratory have dealt with direct-acting carcinogens of the epoxide (16) and halo ether (17,18) classes of compounds. Based on such studies it has become possible, within specific classes of compounds, to suggest the likelihood of carcinogenicity. This subject was recently reviewed (11). Moreover, this approach has important implications in occupational cancer and this aspect was discussed at a recent symposium of the NationaJ Academy of Sciences (19).
The same principles of structure-activity relationships apply also to indirect-acting carcinogens, i.e., those which have to be metabolically activated to carcinogenic intermediates. However, for these compounds known or likely metabolic pathways for activation and for detoxification need also to be considered. This is, in fact, of overriding importance and some of these possible metabolic pathways will be discussed in a later section of this report.

Trichloroethylene
This widely used industrial chemical and anesthetic was chosen for the first phase of our work on halogenated hydrocarbons because of its structural analogy to VC and to some other known indirectacting carcinogens. TCE was also chosen after consideration of its known and suggested metabolic pathways.
Powell suggested many years ago (20) that TCE (I) is probably metabolized to an epoxide, epoxy-1,1,2-trichloroethane (II). This suggestion was based on his extensive metabolic studies on TCE. Subsequently, we suggested that VC (III) exerts its carcinogenicity by metabolism to the activated carcinogenic intermediate (IV) i.e., chloroethylene epoxide (10). It was pointed out (10) that IV is structurally analogous to the direct-acting carcinogens bis(chloromethyl) ether (V) and glycidaldehyde (VI). The chemistry and carcinogenicity of V and VI were described by us in earlier research reports (16,(21)(22)(23) In the same report (14) we described the reaction of II with a variety of nucleophiles shown in structures VII-XX. The purpose of these organicchemical studies was to obtain suitable agents for intercepting the epoxide, if it were formed from [14C] TCE, using in vitro biochemical experiments with rat liver microsomal preparations. Because of the high reactivity of the epoxide, II, it was important that such intercepting nucleophilic reagents In a recent report the carcinogenicity assays on TCE were described (24). The compound is carcinogenic to the liver by feeding to B6C3F 1 hybrid mice, but not to Osborne-Mendel rats. It should be pointed out that TCE must be considered a weak carcinogen because of the high dosages used in these feeding experiments.
Moreover, there is a concern about the matter of metabolic overload which may inactivate detoxifying enzymes.
Epoxy-1,1,2-trichloroethane (II) was synthesized by the autoxidation of TCE (I) with oxygen and ultraviolet irradiation in the presence of benzoyl peroxide. The product was fully characterized (14) and it was found that the compound has a half-life of react rapidly with II, have sufficient water solubility; that the products couig be readily extracted by organic solvents from an aqueous system, and that the nucleophiles selected are nontoxic to the microsomal enzymes. All these criteria were met by the nucleophiles and some of their products shown in structures VII-XX. These interception experiments are currently underway using [14C]TCE with rat liver microsomal preparations in vitro. For the determination of the covalent interaction of TCE with rat liver microsomal proteins, in vitro, microsomal preparations were obtained using established procedures (25). The [14C]TCE used was of > 99% radiochemical purity as determined by gas chromatograph-mass spectrometry and was used as such. These in vitro binding studies have been described in detail by us (13). In brief, the studies showed that TCE binds covalently to microsomal proteins. The binding was decreased by 7,8-benzoflavone. This compound inhibits the binding of 7,12-dimethylbenz(a)anthracene to DNA, RNA, and protein in mouse skin (26). The binding of TCE to rat liver microsomal protein was blocked by ,B-diethylaminoethyl diphenylpropyl acetate (SKF-525A), which is known to inhibit the metabolism of various substrates of cytochrome P-450 (27). As expected, phenobarbital enhanced the binding of TCE to microsomal proteins. Finally, the effect of an inhibitor of epoxide hydrase (28), 1,2-epoxy-3,3,3-trichloropropane (TCPO) on the same in lvitro system was examined and found to cause an enhancement of TCE binding to microsomal protein. The addition of competing nucleophiles such as reduced glutathione, 1-methyl-2-mercaptoimidazole, and mercaptoethanol all decreased the binding of TCE to microsomal proteins.
All of these findings are consistent with the suggestion that an epoxide intermediate is formed and that such an epuxide may be the activated carcinogenic intermediate of TCE in liver carcinogenesis.
The likely metabolit pathway of TCE in animals and man is summarized in the scheme involving structures I, 11, and XXI-XXIV which is consistent with the metabolic studies and the suggestion of Powell (20) concerning an epoxide intermediate.
In recent experiments (15,30), the binding of TCE to liver microsomal proteins of male B6C3F1 hybrid mice which are susceptible to TCE-induced liver tumorigenesis (24) was found to be significantly higher than the binding of ['4C]TCE to microsomal protein from male Osborne-Mendel rats, which were resistant to TCE-induced hepatocellular carcinoma in the same earlier study (24). The in v itro binding of [14C]TCE to liver microsomal proteins was higher for male compared with female B6C3F1 mice; females have been reported (24) to show a lower incidence of TCE-induced hepatocellular carcinoma compared with males. Microsomal proteins from the lung, stomach, and kidney of Poly(vinyl bromide) (PVBr) and copolymers containing the vinyl bromide (VBr) moiety have been under investigation for possible use as flame retardants.
Because of our continuing interest in flame retardants (31) and because of the structural closeness of vinyl chloride and VBr, the latter compound and its polymer were tested for carcinogenicity in mice. (Both materials were kindly provided by Ethyl Corporation, Detroit, Mich.) The results of these experiments are given in detail in Tables 2 and 3, since they have not been published elsewhere.
Skin application of VBr in acetone solution and PVBr in commercial aqueous latex suspension using female ICR/Ha Swiss mice, 30 per group, did not result in any skin tumors in either group. When tested as initiators with phorbol myristate acetate (PMA) as promoter (two-stage carcinogenesis), the test group (VBr followed by PMA) did not show any significant increase in tumors induced compared with the control group (PMA alone). The  (1) No treatment (160 mice) 0 aThirty female ICR/Ha Swiss mice per group; duration 420 days; PMA 2.5 jig, three times/week. bBoth compounds were completely inactive when tested as whole carcinogens on mouse skin by repeated application, 3 x wkly at the same doses and group sizes given above, also for 420 days. cNumber of mice with squamous carcinoma in parentheses. dOne additional animal bore a subcutaneous fibrosarcoma. positive control group showed the expected high tumor incidence (32). Although VBr has a low boiling point (15.80C/760 mm), it is expected that the high dose used in the skin application experiments, 15 mg/application in 0.1 ml acetone, would result in absorption into the skin of a substantial amount of the material before evaporation.
The two materials were also tested by subcutaneous injection in mice (Table 3); VBr was inactive. PVBr, on the other hand, resulted in a high incidence of liposarcomas. However, the material was injected as an aqueous suspension and hence this interesting finding probably is due to physical carcinogenesis and thus does not bear on the matter of VBr and/or PVBr as chemical carcinogens. To our knowledge inhalation studies on VBr have not, as yet, been reported.

Environmental Health Perspectives
Both materials were also tested by feeding in rats. The results of these tests are incomplete and will be published elsewhere when the histopathology is complete.

Other Chlorinated Hydrocarbons
Because of the extensive use of chlorinated hydrocarbons in chemical industry, some of them are becoming ubiquitous environmental contaminants and possibly environmental hazards. They may occur in films used as food wrappings, latex paint, rubber, etc. (4)(5)(6) and some of them occur in drinking water. A few of the compounds of concern to us that have been shown to occur in drinking water (33) are listed in Table 4. The presence of these compounds in drinking water must be ascribed to contamination of water supplies by effluents from chemical industry. The carcinogen chloroform (34,35) also occurs in drinking water. However, this compound is also produced from naturally occurring humic materials in the disinfection of water supplies by chlorine (36,37). The latter source must account, at least in part, for its presence in drinking water. The manufacture, chemistry and toxicology of these compounds have been reviewed recently (38). The purpose of this section of the present review is to discuss some compounds which are of current interest and under investigation in many laboratories, including our own.
The compounds of major interest are shown in structures XXVII, XXIX, XXXI, and XXXIII-XXXVIII. These compounds merit special attention for several reasons: (a) the structural analogy of some of these compounds (the chlorinated olefins) to vinyl chloride and TCE and hence possible carcinogenicity; (b) their large-scale production and use in chemical industry leading to environmental pollution by these agents; (c) some of these compounds are known carcinogens in rodents, e.g. ethylene dibromide (39), hexachlorobutadiene (40), and vinylidene chloride (41). Hexachlorobutadiene was recently shown to cause kidney tumors in rats by ingestion, but details of these experiments are not, as yet, available. Ethylene dichloride is discussed elsewhere in this volume (42).
Studies in this laboratory on structures XXVII, XXIX, XXXI and XXIII-XXXVIII are concerned with their chemistry, metabolism, carcinogenicity and determination of the nature of the activated carcinogenic intermediates for those compounds which are known carcinogens. These studies are carried out by using the same approaches described earlier in this review for trichloroethylene, and also including bioassays.
In conclusion, only one additional point will be made, and that concerns the matter of structural prognostication of carcinogenicity already alluded to in an earlier section of this review. For compounds suspected to be carcinogenic, it is of importance to examine the known, as well as postulated metabolic pathways if it appears likely that the parent compound is an indirect-acting carcinogen. As CH =CHCH2OH -CH =CHCHO -) CH2 =CHCO2H an example, the possible intermediates that can be visualized for the metabolism of allyl chloride, XXXIII are shown in structures VI and XXXIX-XLIV. This example is particularly interesting, since two of the possible intermediates, epichlorhydrin (XLII) and glycidaldehyde (VI) are known direct-acting carcinogens (17,43).
The author is indebted to his collaborators whose work is summarized as part of this review and referred to in the refer-