Document DXwgxojxYYYw02QLrZK7RNMo

'.4lo'.jn ^ i.\c on the - I r.iifanmental and Mech- \ p 25. ts<'.sl r>poi<>fc vf S3C* h>r<rpljsias in .^-hydroxy- 65. 1005. . Mill P- VVeiv uirition and its 32. 237. v76). OccupatioVol. 271. y induced dis- & Weisburger. jtral sterols on i skin o mice. of saccharin. In Report .Vo. 1 'arker. p. 153. Mechanisms of amine carcinoarcinogenic and j. Edited by S. 53. Monograph Institute. Deth- LChemical car nee o f Poiions. sen & M. O. :k. Metabolism of reh TreaBec i. 241. er tumour pro- < van de Poll. Value of Shortvaluation Elsensterdam. 31). Systematic aoint approach. Tobacco and Med. 300. 894. P p c h r n l Q j\|0 ;r t\,ryjrmn PrcN*Lid GENETIC AND NONGENETIC EVENTS IN NEOPLASIA W. T. Stott. R. H. Reitz, A. M. Schumann and P. G. W atanabf. Toxicology Research Laboratory. Dow Chemical USA, Midland. M l 4Xb4l), USA (Received 2 February /9HI) Summary-- It has become increasingly evident that all chemical carcinogens do not act via the same mechanism of tumorigenicity. Based upon the extent of a chemical's interaction with DNA, a general classification scheme of various mutational and nonmutational theories of chemical carcinogenesis ts presented. Compounds that directly interact with DNA are classified as genotoxic whereas those that do not interact directly with DNA are classified as epigenetic carcinogens. Under each general heading, several mutational and nonmutational mechanisms of carcinogenesis are believed to be possible. Data are also presented to support the existence of one such mecnanism. an epigenetic-mutational theory of chemical carcinogenesis based upon recurrent cytotoxicity. In this case, increased regenerative DNA synthesis in response to tissue injury is believed to result in an enhancement of the normal spontaneous mutation rate, conceivably leading to a ceilular transformation. The carcinogenic risk posed by such epigenetic carcinogens appears to differ greatly from th3t posed by gcnotoxic carcinogens. Thus, con sideration of data conce-ning the possioie mechanism of carcinogenicity of a chemical, along with pharmacokinetic data, will allow a belter understanding of bioassay results and a more accurate assess ment of carcinogenic risk. Introduction Frequently in the routine rodent carcinogenesis bioassay of a chemical high levels of a test compound are administered for prolonged periods of time and. at the termination of the study, the numbers of tumours formed are counted. If a statistically significant exce-ss of a particular tumour is observed the compound is classified as an animal carcinogen and an extrapola tion of risk-is made to humans. Vet in most cases, without knowledge of the fate of the chemical in several species (i.e. pharmacokinetics) as well as of the resultant macromolecular events to help understand ing of the bioassay results, there is little chance of a meaningful assessment of human carcinogenic risk. It is the integration of these three data bases (bioassay, pharmacokinetics and macromolecular events) to more realistically assess the potential carcinogenic risk of a given chemical that has been a continuing endeavour in our laboratory over the past several years. Recently, we have been involved in trying to understand the different mechanisms of tumour for mation and their implications for risk assessment, and. more specifically, the role of cytotoxicity in the carcinogenic process. This latter area of research will be the focus of this paper. Classification of proposed mechanisms of chemical car cinogenesis Since the first experimental documentation of chemically induced carcinogenesis in animals by Yamagiwa &. Ichikawa (1918). numerous theories have been proposed to explain the mechanisms of chemical carcinogenesis. Apart from the various dif ferentiation. protovirus and viral infection theories of carcinogenesis (Temin, 1974). most theories regarding the mechanising) of chemical carcinogenesis may be Generally classified ns 'genetic' or 'epigenetic' (nongc- netic) in nature. As discussed below, the distinction between these two general classes of mechanisms is of great significance in carcinogenic risk assessment. As used here, chemicals displaying a genetic mechanism of carcinogenesis are defined as those involved in di rect interaction with cellular DNA. In most cases this interaction is considered to result in a mutational event; however there may also be cases of nontr.uiauonal genetic mechanisms of carcinogenesis (e g. Hol liday, 1979). Chemicals displaying an epigenetic mech anism of tumorigenesis are defined as those that do not directly interact with cellular DNA, and yet their action may indirectly result in a mutational event. Thus, the terms genetic and epigenetic in chemical carcinogenesis, as used here, refer to the likelihood of direct chemical-DNA interaction, and in either case may or may not involve a mulational event. The most common interpretation of a genetic mechanism of tumorigenesis is embodied in the soma tic mutation theory of carcinogenesis first proposed by Boveri (929). Basically, this theory dictates that the direct interaction of a chemical with DNA le.g. by alkylation, intercalation) can result in a somaiic cell mutation (iia point or deletion mutaiion. gene dupli cation or gross chromosomal disruptions) which may or may not ultimately lead to a transformed or neo plastic cell. As reviewed by Hanawalt. Friedberg & Fox (1973) mitigating factors in this progression are the DNA-repair enzymes which remove the chemi cally induced `lesion' before (excision repair) or after (post-replication repair) DNA replication (i.c. before the lesion is fixed as a mutation). It also appears that errors in this repair, possibly by an inducible errorprone DNA-repair enzyme system analogous to SOS repair observed in bacteria, may be ro'ponsible fur the mutation rather than the initial chemicallyinduced mispairing of bases (Kondo. 197 & 1977; San .k: Stich. 1975, Sarasin & Benoit. 1980: Wilkin. 1976). 567 p U Uf >3721 568 W. T. Stott. R. H. Reitz. A. M. Schumann and P. G, Watasabf. Comings' (1973) `general theory of carcinogenesis' 'integrative theory' of carcinogenesis of Trosko & provides a possible scenario of events leading to a Chang (1978). These authors propose*that neoplasms state of anaplasia after a mutation occurs. In this case may arise from a mutagenic event alone (regulatory the significant events would be a double mutation of locus) if the genes affected are in a transcribable state, one or more regulatory genes (normally present in all from a mutagenic event if there is a coupled nonmuta r cells), in turn derepressing corresponding structural tional alteration of the mutated genes (promotion), or genes capable of coding for cellular transforming from the abnormal derepression of g$nes at critical factors. In the case oi chemical carcinogenesis follow developmental stages that prevent normal gene regu ing an apparent two-step progression towards neo lation. plasia ('two-step theory of carcinogenesis'), this dere Based upon the apparent realization that some pression of genes following a mutational event (in chemicals may cause tumours wrr several different itiation) may be promoted by epigenetic stimuli, mechanisms of action (i.e. there is no single unifying chemical or physical (Boutwell, 1974; Trosko & mechanism), it has been possible to propose a classifi Chang, 1978). cation scheme for chemical carcinogens even though Substantial evidence has been accumulated to link mechanistic details remain to be elucidated. Weis- mutagenic events with the neoplastic transformation burger & Williams (1980) have thus proposed that of cells.' As reviewed by Trosko & Chang (1978), find carcinogenic chemicals be categorized under the gen ings regarding the clonal nature of tumours, the eral division of genotoxic chemicals (direct-acting or mutagenicity of many carcinogens, the correlation of primary carcinogens, procarcinogens or secondary high cellular mutation frequency with increased carcinogens, inorganic carcinogens) and epigenetic cancer rates among humans suffering from deficien carcinogens (solid-state carcinogens, hormones, im- cies in DNA-repair enzymes, the correlation of in munosupressive agents, co-carcinogens, promoters). A vitro DNA damage with cell transformation frequen diagrammatic representation of the various theories cies and neoplasia, the involvement of mutation in the of carcinogenesis is presented in Fig. 1. initiation phase of some hydrocarbon-induced carci nogenesis. the effects of age on mutagenesis and the Concept of the cytotoxic mechanism of tumour incidence of various hereditary tumours, are all sup- formation .portive of a mutagenic origin of cancer. However, several findings cannot be reconciled with a muta An epigenetic mechanism of tumorigenesis that is genic basis of carcinogenesis. These have included ob not frequently considered in terms of tumour produc servations regarding the fact that all carcinogens are tion is that of recurrent cytotoxicity. The simplest not mutagens (Jensen, 1974; Lippman, 1975; Segaloff; example of this mechanism is the production of sar 1975). the discrepancy between gene mutations and comas following subcutaneous injection of inert sol cell transformation rates in short- v. long-lived ani utions such as saline or glucose. It is envisaged that mals (Huberman. Mager & Sachs. 1976; Peto, 1977). repeated subcutaneous injection of these nonreactive the induction of tumours by plastics- or metal-film materials results in inflammation, necrosis and cellu implantation and hormone imbalance (Berenblum, lar division which may increase the error rate in nor 1978). the ability of teratoma cells to produce appar mal DNA replication or DNA repair, resulting in cri ently normal cells upon transplantation into normal tical-site mutation and ultimately in a transformed or mosaic mouse blastocytes (Mintz & Illmensee, 1975; neoplastic cell. In general terms, this theory represents Papaioannou. McBurney, Gardner & Evans. 1975) a modification of Virchow's `irritation theory' in and the ability of Lucke frog adenocarcinoma cells to which hyperplasia was believed to be the driving force reproduce normal cells upon nuclear transplantation behind carcinogenesis (Berenblum, 1944). It is likely into anucleate eggs indicating a totipotency of the that some chemicals may illustrate both genetic and tumour-cell genes (McKinnell. Degeins & Labat, epigenetic effects. In such an instance it will be impor 1969). Thus, there is also substantial evidence for the tant to determine which mechanism is predominantly existence of one or more epigenetic nonmutational responsible for the formation of tumours. mechanism of carcinogenesis. With the concept of testing maximum doses in ani Several theories have attempted to reconcile these mals in order to enhance the probability of detecting different mechanisms of carcinogenesis either by pro a carcinogenic response, one of the primary issues is posing an alternative to mutagenesis or by accommo the relevance of such testing in terms of the real risk dating both genetic mutational and epigenetic non for humans of a chemical shown to be carcinogenic mutational mechanisms. A recent example of the under such test conditions. Frequently in such studies former situation is the nonmutagenic theory proposed tumours are observed at dose levels that also cause by Holliday (1979). This theory proposes that gene recurrent cytotoxic responses, and a logical question regulation, as directed by .specific DNA base methyla- would be whether an analogy to sarcoma production tion and recognition by several DNA methylases. may following subcutaneous injection may be made. In be altered during repair of DNA-chemical alkylation understanding this relationship, it-is enlightening to sites. The resultant loss;of specific methylated recog explore the molecular basis for increasing mutation as nition sites for DNA methylases results in loss of gene a consequence of increased DNA synthesis and cell regulatory control and ultimately in a transformed division. cell. Since the direct interaction of the chemical with DNA does not exist in tiro as a pristine molecule the DNA is required to deregulate genes in this non- even in the absence of a measurahle genotoxic chal muiagenic theory, it is still consistent with a genelic lenge. So-called spontaneous mutations appear to mechanism of tumorigenicity as defined above. An occur at a rate dependent upon exposure to unavoid example of the latter `coexistence* type of theory is the able exogenous (e.g. cosmic radiation) and endogen- .pY-fr-TH jrr; r* GENETIC Direct Interaction with genotic material (DNA) EPIGENETIC No direct Interaction with DNA MUTATIONAL EVENT e.g. Mutation In criticol site on DNA (somatic mutation theory) NONMUTATIONAL EVENT* e.g. Loss of DNA regulation 'markers' due to repair of alkylation sites (Holliday, 1979) MUTATIONAL EVENT e.g. Repeated cytotoxicity leading to enhanced number of errors In normal replication and critical NONMUTAT1ONAL EVENT e.g. Induction of preexisting oncogenic factors by alteration of specific cellular metabolic functions Solid-state carcinogenesis Hormone-Induced carcinogenesis Genetic and nongenetic events in neoplasia TRANSFORMED CELL t' * > l Fig. I. General classification of some of the proposed mechanisms of carcinogenesis, excluding the differentiation, oncogenic, viral infection and protovirus theories, The role of DNA repair in mutagenesis is not included. Believed to be a readily reversible alteration (Holliday. 1979), jL *>t'V it*') r "* -/ W w. i. aiorr, K. H. Keita A. M. Schumann and P. G. Watanabe ous factors, DNA-repuir competency (e.g. relative to age) and the basic thermo-stability of the DNA mol ecule itself. With regard to the latter point it has been estimated that thermal decomposition may account for between 2000 and 10,000 depurinations, several hundred depyrimiclinations and several base deamina tions of DNA molecules in a mammalian cell per gen eration (Lindahl & Karlstrom, 1973; Lindahl & Nyberg, 1972 & 1974). These thermal lesions can result in base transitions (e.g. deamination of 5methylcytosine to form thymine) which conceivably may be `fixed* (irreparable) by DNA replication or repair processes resulting in a functional mutation. Indeed, errors in the process of DNA replication itself may occur as a result of base mispairing. DNA poly merase base selection errors and mismatch repair errors (see review by Hartman, 1980). Based upon es timates of the mutation rales per human gene per sexual generation, it has been suggested that 10% of all human gametes contain a new mutation of their own plus several inherited mutated genes as well (Drake, 1978). Thus the mere fact that DNA synthesis and cellular division is enhanced following cytotoxic responses will result in a decreased time for repair of `naturally' occurring DNA lesions before replication, an increased number of replication errors, and an enhanced somatic mutation rate. Additionally,' since errors in DNA may occur throughout the cell cycle or during DNA synthesis, the DNA-repair process can be thought of as a con tinuous process. However, as noted, some forms of DNA repair are error-free while others may be errorprone. Therefore the fidelity of DNA repair can affect the somatic mutation rate during ceil division. A recent report by Shank & Barrows (1980) on studies of the noncarbonatious chemical hydrazine has shown that cytotoxic doses of hydrazine cause abnor mal methylation of nucleic acid bases in the liver. Furthermore, evidence was presented which suggests that at hepatotoxic doses other chemicals (e.g. carbon tetrachloride) produce a similar alteration in the DNA as a result of cytotoxicity. Whether this effect is due to an alteration in the DNA polymerase or in the fidelity of DNA repair following DNA replication is not known, but it is an example of how a cytotoxic agent at high dose levels can cause abnormal nucleic acid bases to occur in cellular DNA. DNA synthesis and cell division can also enhance the susceptibility to mutation by exogenous and en dogenous agents. For example, Berman, Tong .& Wil liams (1978) have reported an increased mutation fre quency in actively dividing adult rat-liver epithelial cells exposed to the mutagenic chemical methyl methanesulphonate compared with that in quiescent (non dividing) cells. In this study the number of a2aguanine resistant colonies per 106 colony-forming units was used as a measure of mutation frequency. The muta tion frequency in the actively dividing cultures wasincreased five-fold when compared with the quiescent culture, indicating that dividing cells were more sus ceptible to mutation. Finally. DNA synthesis and cell division can also enhance mutation frequency by altering the amount of DNA repair prior to DNA replication. An example .of this is given by the work of Maher, Dorney, Mendrala, Konze-Thomas & McCormick (1979) using normal human fibroblasts and fibroblasts lacking normal DNA excision repair. Cell survival and muta tion rate in cells exposed to ultraviolet (UV) radiation was observed to be related to DNA-repair compe tency. Survival of normal cells was higher and muta tion rate lower per UV dose than for deficient cells. Normal cells were also able to survive a usually lethal and mutagenic UV dose by being held in confluence (nondividing) for a period of time before assessment oi survival and mutation rate, presumably because this allowed a greater time for DNA repair to occur before replication. In addition to a molecular basis for enhancing mutagenic events as a result of increased DNA syn thesis, there is an experimental basis for the role of DNA synthesis in carcinogenesis. Tumours often de velop in chronically inflamed or scarred tissue; colon cancer is frequently observed in patients with chronic colitis; skin cancer occurs in burn scars; liver tumours are associated with chronic cirrhosis of the liver (Ber- enblum, 1944; Chan, 1975; Laroye, 1974). Repeated tissue damage with a physical agent (dry ice) and resultant regeneration has induced tumours in mice (Berenblum, 1929). Physical trauma such as partial hepatectomy has also been shown to enhance the tumorigenic effect of thioacetamide, N-nilrosodi- methylamine (NDMA) and N-nitrosodiethylamine (Craddock, 1978; Date, Gotoskar & Bhide, 1976). Dimethylbenzanthracene-induced skin tumours pro moted by the classic phorbol esters have been ob served to be inhibited by anti-inflammatory steroids which inhibit inflammation. DNA synthesis and cell proliferation (Slaga, Fisher, Viaje. Berry. Bracken, LeClerc & Miller, 1978; Weeks, Slaga, Hennings, Gleason & Bracken. 1979). Finally, as noted above, tumours can be induced a! the site of subcutaneous injection of nonreactive chemicals, such as glucose. ' saline or distilled water (Grasso & Golberg, 1966). Having explored the molecular as well as experi mental basis for the effects of cytotoxic responses on 3 the carcinogenic process, the obvious question arises 1 about the impact of cell division and increased DNA | synthesis on genotoxic chemical carcinogens. For 1 genotoxic carcinogens it is likely that cytotoxicity will | cause an increased incidence and/or a decreased | latency in the production of tumours. A good | example of this is the induction of liver tumours by f NDMA (Terracini, Magee & Barnes, 1967). At low | doses of NDMA that preclude cytotoxic effects h the | liver, tumours can still be induced. At higher levels g where both cytotoxic as well as genotoxic effects act ij in concert, liver tumours are observed at an increased ! frequency as well as at a decreased latency. For chemical carcinogens displaying a cytotoxic mechan ism with little to no genotoxic activity it is likely that a prolonged, recurrent cytotoxic response throughout ! a large portion of the animal's life will be required for i the induction of tumours. I The implications of these carcinogenic mechanisms I are as follows. (1) For genotoxic carcinogens, the j defence mechanisms such as DNA repair, detoxifica- J tion and excretion of the reactive species will be criti- j cal for modifying the carcinogenic response. Risk csti- J mates need to consider the dose-related kinetics of the J lesions produced and the extent and persistence of | those genetic lesions. (2) For epigenetic carcinogens I tv Genetic and nongenetic events in neoplasia 571 causing tumours via a cytotoxic mechanism of tumorigenieiiy, tumours will be induced at doses that pro duce recurrent cytotoxicity and no carcinogenic risk would be predicted at doses that preclude any tissue injury. It is expected that chemical carcinogens will rep resent a spectrum of activity ranging from genotoxic to epigenetic Those chemicals having strictly genetic activity are often those that either are. or can be acti vated to, strongly electrophilic species. Chemicals with intermediate activity may have both a genetic as well as an epigenetic component to their tumorigenic response and some chemicals that show little to no genetic activity may produce tumours strictly through the mechanism of recurrent cytotoxicity or other epi genetic mechanisms. This is not to say that chemical reactivity is always directly relatable to carcinogenic potency but only that such compounds show a pro pensity to react with macromolecules. Obviously, the qualitative character or reaction site specificity of these reactions will have a profound impact upon the outcome of the DNA-chemical interaction. Indeed, a majority or interactions would not be expected to result in any biologically significant event because they involve noncritical-site alkylation or repair. It is also important not to rule out endogenous factors such as genetic predisposition working in concert with epigenetic mechanisms since all species and strains are susceptible to the development of particu lar tumours that are not related to exogenous chemi cal exposure. This concept is of paramount impor tance' in assessing the relevance of extrapolating test results in sensitive animal species to humans. Experimental approaches: some examples The experimental approach taken in our laboratory has been to develop some objective criteria to differ entiate between the genetic effect and the epigenetic, cytotoxic mechanism of tumour production. The two parameters that were selected as representative of a genetic effect were (1) DNA damage, measured by alkylation of DNA in vivo and (2) DNA repair, induced presumably by damage to the DNA mol ecule. While there are inherent limitations to using total DNA alkylation as an index of genetic interac tion, due to the overall lack of information on specific critical sites of reaction on all of the nucleic acid bases, it was one of the only means of assessing the potential for in vivo genetic interaction. NDMA was selected as a genetic-acting carcinogen and the extent of DNA alkylation and of DNA repair following treatment with the test chemicals were compared with those after NDMA treatment. NDMA was also selected because it represented an electrophilic methylating agent and the chemicals we were testing were low-molecular-weight molecules. If high-moleculnr^ght heteroaromalic molecules are being tested, bvn2o[o]pyrene or the naphthylamines may be appropriate. For evidence of epigenetic, cytotoxic c,lCas* DNA synthesis in the target tissue was Measured and classic histopathology was also carried l>i. The chemicals that were tested were perchloro* -'*cnc*chloroform and 1,4-dioxane, hi tesuhs of the carcinogen bioassay on perc oroethylene conducted by the National Cancer Institute and of studies conducted in our laboratory are summarized as follows. Mice exposed by gavage to approximately 500 or lOQOmg/kg/day throughout their lifetime showed an increase in liver tumours (hepatocellular carcinomas), but rats exposed to equally high doses did not show any tumorigenic re sponse (National Cancer Institute, 1977). .In inhala tion studies, exposure to 300 or 600 ppm perchloroethylenc daily for 1yr (subsequent observation for 18 months) did not prove to be tumorigenic in rats (Rampy, Quash Leong & Gehring. 1978). It is impor tant to emphasize that the B6C3Ft mice in which liver tumours were enhanced by perchloroethylene treatment have an average control incidence of hepatocellular carcinoma of 10-15% (some nearly 50%). Therefore, this species and strain has a marked genetic predisposition for liver neoplasms. - Studies designed to characterize differences in the pharmacokinetics of perchloroethylene in rats and mice showed the following (Table 1; Schumann, Quast & Watanabe, 1980; Watanabe, Reitz, Schu mann, McKenna, Quast & Gehring, 1980). When exposed to a 10 ppm perchloroethylene atmosphere for 6 hr mice metabolized and activated nine times more perchloroethylene than did rats. This resulted in a seven-fold greater macromolecular binding in the liver of mice than rats. Importantly, binding to hepa tic DNA in mice at a tumorigenic dose level of 500 mg/kg was not observed. The detection limit for the perchloroethylene-DNA binding studies in mice was 10 alkylations per 10s nucleotides. When parameters of cytotoxicity were assessed (Table 1) the liver weights were increased in treated mice but not in rats. DNA synthesis caused by tissue injury was increased by 82% in the mice but was not significantly increased in rats, and finally histopathological alterations in the liver were evident in mice but not in rats. Another possible indication of direct genetic inter action is that of gross chromosomal damage and of in vitro mutagenesis, if a possible nonmutational gene deregulation mechanism of carcinogenesis (e.g. Holli day's theory) is ruled out. When mutagenicity data for perchloroethylene were examined (Table 2), it was_ noted that negative results in Salmonella have been reported from three laboratories and there has been one report of a positive finding in the Ames-type strain of Salmonella typhimurium. Further, in a cyto genetics study no chromosomal abnormalities attribu table to perchloroethylene exposure were observed. The overall conclusion is that perchloroethylene has little to no genetic activity. In summary of the perchloroethylene data, DNA alkylation was not detectable at a detection limit of ten alkylations per 106 nucleotides; DNA-rcpair assays were not conducted because they are generally less sensitive than the alkylation studies. Mutagenicity data is predominantly negative in bacterial systems as well as negative in an animal cytogenetics study. Cytotoxic responses indicated by increased DNA syn thesis, two-fold over that of control, and histopathological examination were evident in the mouse at tumorigenic dose levels. In contrast, a classic genotoxie carcinogenic agent (e.g. NDMA, ethylnitrosourea) would alkylate DNA at a level of hundreds to thousands per 10* nucleotides; DNA repair would J W. T. Stott. R. H. Reitz, A. M. Schumann and P. G. Watanabe Table l. Perchloroethylene metabolism. genoto.xicity and cytotoxicity data (Sc/iumif/jrt, Quast <Sc Watanabe, !9$i); Watanabe, Reitz, Schumann, McKenna Jt Gehring, JV80) Response in Perchloroethylene 'A treatment Parameter f Rats Mice 6 Inhalation exposure to 10ppm for 6 hr Single oral dose of 500 mg/kg body weight Twelve oral doses of 500 mg/kg body weight/day Total metabolized Otmol-equiv/kg body weight) Total macromolecular binding (pmol-equiv/g hepatic protein) Hepatic DNA binding Liver weight:body weight (% increase) Hepatic DNA synthesis (% increase)* Hepatic histopathological changesf 105 002 -- ` 5 -- 89-5 0147 ND 25 + ND = Not detected Present as a percentage increase in DNA synthesis in treated compared with control animals as measured by [6-JH]ihYmidine uptake. tA treatment-related response was observed (-j-Xor no treatment-related response was observed (--). larly was DNA relate 10 an Cy thesis treats lation weigh 15 me illustrate a dose-response relationship; in vitro muta genicity would be clearly evident. In contrast to these genotoxic effects, there would be little histopathological alteration or increased DNA synthesis in the target organs at low, though tumorigenic, dosages, thus providing evidence of a lack of a significant role of an epigenetic, cytotoxic mechanism. Our overall conclusion is that the liver tumours observed in mice following lifetime administration of high doses of perchloroethylene were a result of the recurrent hepatotoxic effect of perchloroethylene which caused an enhancement of the normal background liver tumour incidence in a strain of mice with genetic predisposition towards the development of such tumours. Most importantly, hepatotoxic effects in ani mals and humans would be observed only at dose levels that are well above levels encountered in the occupational environment. Thus the relevance of the data that has been obtained for mice which suggests a carcinogenic risk to humans exposed at occupational or environmental levels of perchloroethylene is highly questionable. The next example-is that of chloroform (Reitz, Quasi, Stott, Watanabe & Gehring. 1930). The cardnogenidty data for male mice given chloroform are summarized in Table 3. This represents a compilation of more than one study, but these data clearly show that following administration of chloroform at high doses male mice are susceptible to the development of liver and kidney tumours. It is noteworthy that at I7mg/kg/day no significant incidence of tumours at any site was observed. At high doses chloroform also induces tumours in the rat, but for the sake oi space we shall limit discussion here to mouse data. At an oral dose level of 240 mg chloroform/kg body weight, mouse liver and kidney DNA was alkylated at a rate of three and one alkylations per 10* nucleotides respectively (Table 4). These values represent the upper bounds of chloroform alkylation since no actual DNA adducts have been identified. In contrast, * I * | f 5 I | j| % | S f | ^ K K serv sity loci In tion chlor 10s revie' form Table 2. In vitro mutagenicity data and cytogenetics data on perchloroethylene Bioassay system Results* Reference Escherichia coti Salmonella strains Cytogenetics + Greim, Bonse, Radwan. Reichert Sc Henschler (1975) National Institute for Occupational Safety and Health (1977). Cern.i & Kypcnova (19771 Barisch, Malaveille, Bnrbin & Planche'(l979) National Toxicology Program (I9it0) Ccrna & Kypcnova (1977) Positive (+ ) and negative (--) responses are indicated. t I-* Genetic and nongcnetic events in neoplasia ^ 573 Table 3. Data from tumorgenicity assays uf chloroform gu rrt orally to male mice Chloroform dose Strain Percentage of excess tumours (site) Reference 277 mg-kg/day, 5 days/wk for 78 wk 60 mgkg/day. 6 days/wk for 80 wk 60 mg/kg/day. 6 days/wk for 80 wk 17 mg-kg/day, 6 days/wk for 80 wk B6C3F, C57BL ' ICI ICI 92 (liver) ND 23 (kidney) ND ND = Not detected t National Cancer Institute (1976) ' Roe, Palmer. Worden & Van Abb (1979) Roe et al. (1979) Roe el al. (197?) h a oie is highly ? 'i Tor eitz, >i ]. V... -Jirci- ; ;; oroform are ? compilation -'[early show arm at-high v j elopment of i 'thy that at tumours at - ' roform also -ike of space Jta. Tin. kg body alk)lated at ' nucleotides present the i since no In contrast. r tt a dose of 10 mg NDMA/kg results in approximately 540 alkylations per 106 nucleotides in the liver. Simi larly. a tumorigenic dose of 240 mg chloroform/kg was not observed to induce hydroxyurea-resistant DNA repair (Table 4) while NDMA showed a doserelated increase in DNA repair following doses of 3, 10 and 20 mg.-kg body weight. Cytotoxicity, as evidenced by increased DNA syn thesis in both the liver and kidney of chloroformtreated mice, showed an increasing dose-response re lationship at 60 and 240 mg chloroform/kg body weight, but it was normal at a nontumorigenic dose of 15 mg/kg (Table 4). Concomitant with the increase in DNA synthesis, histopathological alterations were ob served at the higher dose levels. On the other hand, NDMA, at 3 mg kg body weight, showed no propen sity towards increasing DNA synthesis or histopatho logical alterations as evidence of cytotoxicity. In summary of the chloroform data. DNA alkyla tion in the liver following a tumorigenic dose level of chloroform was a maximum of three alkylations per I06 nucleotides. DNA repair was not detectable. A review of the literature has also shown that chloro form is not mutagenic in a mammalian cell and several bacterial mutagenesis assay systems (Kirkland. Smith & Van Abb, 1981: Reitz et aLt 19S0). Evidence of chloroform's cytotoxicity has come from studies using mice dosed with a tumorigenic level of 240 mg/kg body weight. DNA synthesis was observed to be increased 14-fold in the liver and 25-fold in the kidneys of these animals, and this was associated with histopathological observations of toxicity. It appears that the genetic potential of chloroform is nil to very little, while the component of recurrent cytotoxicity is very significant. Thus, as with perchloroethylene, it appears that chloroform causes tumours in mice via a cytotoxic mechanism rather than by a purely genetic mechanism. The final example is that of 1,4-dioxane. When chronically administered to rats at a concentration of 0-75 to I-8/ in their drinking-water, 1,4-dioxane has been observed to cause an excess incidence of nasal and hepatocellular carcinomas (Argus, Arcos & Hoch-Lieeti, 1965; Hoch-Lgeti, Argus & Arcos. 1970; Kociba, McCollister, Park, Torkelson & Gehring, 1974). Ingestion of lower doses (01%) in drinkingwater (Kociba et at. 1974) and inhalation of 1,4-diox ane (Torkelson, Leong, Kociba, Richter & Gehring, *' 1 '*1-D Table 4. Data on chloroform and bi-nitrosodimethylamine (NDMA] genotoxietty and cyto toxicity in male CD-I mice (Reitz, Quasi, Stott, IVatanube & Gehring. JVX0) Treatment Hepatic DNA Binding (alkylations/ 104 nucleotides) Rlepair* DNA synthesist Histopathological in changest in Liver Kidney Liver Kidney Chloroform 240 mg/kg 60 mg/kg 15 mg/kg NDMA: 20 mg/kg 10 mg/kg 3 mg'kg 3 -- -- -- 540 -- ND x 14 x 25 + -- x2-2 x 8-2 -- -- ND ND -- 7-37 -- -- -- 304 -- -- -- 1-60 ND -- -- + + - -- -- -- N D = Not detected ` Ratio of treated control values. Values > 10 indicate an increase in DNA repair, tPresented as multiples of control vahies as measured hy [6--lU]ihymidine incorporation. A treatment-related response was observed ( + ) or no treatment-related effect was ob served { - ) . Chloroform was administered as a single oral dose; NDMA was given as a single ip injection. 574 W. T. Stott. R. H. Reitz. A. M. Schumann and P. G. Watanabe 1974) have failed to produce excess tumours. At a tumorigenic dose level (1015 mg/kg/day or 1% in water). 1,4-dioxane was also observed to cause pro* nounced hepatic degenerative and regenerative changes (Kociba et at. 1974). No genotoxicily has been observed in SpragueDawley rats dosed with a single oral carcinogenic dose level or l,4rdioxane (1000mg/kg); no increased DNA repair or hepatic DNA alkylation was delected. In vitro mutagenicity assays carried out following the methods of Ames, Durston, Yamasaki & Lee (1975) and using S. typhimurium strains TA98. 100, 1535, 153? and 1538 also gave negative results. No DNA repair was detected in an in vitro primary rat hpato cyte RNA repair assay following the method of Wil liams (1977). Yet, as shown in Tabic 5, repealed administration of a carcinogenic dose level of 1,4-dioxane (1000mg/kg body weight/day) was observed to be cytotoxic to rat hepatic tissue as evidenced by a l*5-fo]d increase in hepatic DNA synthesis and abnor mal histopathology relative to the controls. No cyto toxic changes were observed in rats dosed with a nontumorigenic level of 1,4-dioxane (10mg/kg body weight/day). The lack of genotoxicity and the observed cytotoxi city of 1.4-dioxane at tumorigenic dose levels indi cates that as with perchloroethylene and chloroform in the mouse, 1,4-dioxane may cause tumours in the rat via an epigenetic-cytotoxic mechanism of action. However, the pronounced hepatocellular hypertro phic response observed in rats exposed to a high dose level of 1.4-dioxane (Table 5) and the observation that 1.4-dioxane induces hepatic drug-metabolizing enzyme systems in the rat (this laboratory) also sug gest another epigenetic mechanism of tumorigenic action characteristic of metabolic inducing agents such as phnobarbital. It has been suggested that this mechanism of action may be a result of degranulation of rough endoplasmic reticulum resulting in alter ations in protein synthesis and gene expression or by a hyperplasia-induced enhancement of the expression of pre-existing oncogenic factors (Tennekes, 1979; Wright, Akntowa & Wooder. 1979). Thus from the available data it appears that repeated exposure to high doses of 1,4-dioxane may cause tumours in rats via one or more epigenetic (mutational or nonmutational) mechanisms of action. Conclusion Obviously, much work remains to be done to fill in the gaps in our understanding of the mechanisms of chemical carcinogenesis. Future work to define criti cal sites on macromolecules (DNA, histones, nonhis tone proteins) that may be important in the process of gene expression and cellular transformation will be of particular importance. However our understanding of the mechanisms of chemical carcinogenesis has reached a point whereby this knowledge may be ufced to generally classify carcinogens and in human risk assessment, even though the more definitive mechan istic details remain to be elucidated. The studies cited above, and others, have shown that all carcinogens do not act by the same mechanism. Some, which have genetic mechanisms of carcinogenicity, pose greater risks.than others, which act by epigenetic mechan isms. Finally, it should be stressed that chemicals hav ing an apparent cytotoxic mechanism of action are tumorigenic only at toxic dose levels, have reversible cytotoxicity, require multiple dosages, and appear to have realistic thresholds. REFERENCES Ames. B. N,, Durston, W. E , Yamasaki E & Lee; F. D. (1975). Carcinogens are mutagens: A simple test system combining liver homogenates for activation and bacteria for detection. Proc. natn. Acad. Sci. U.S.A. 70, 223!. Argus. M. F Arcos. J. C. & Hoch-Ligcli. C. (1965). Studies on the carcinogenic activity of protein denaturing agents: hepatocarcinogenicity or dioxane. J. Natn. Cancer Inst. 35. 949. Bartsch. Malaveille, C , Barbin. A. & Planche, G. (1979). Mutagenic, and alkylating metabolites of halo- ethylene. chlorobuladienes and dicnlorobutenes pro duced by rodent or human liver tissues. Archs Toxicol. 41, 249. Berenblum, 1. (1929). Tumour-formation following freezing with carbon dioxide snow. Br. J. exp. Path. 10, 179. Berenblum. 1. (1944). Irritation and carcinogenesis. Archs Path. 38, 233. Berenblum. I. (1978). Established principles and unresolved problems in carcinogenesis. J. natn. Cancer Inst. 60, 723. Berman. J. J,, Tong. C. & Williams, G. M. (1978). Enhance ment of mutagenesis during cell replication of cultured liver epithelial cells. Cancer Lett. 4, 277. Boutweil, R. K. (1974). The function and mechanism of promotors of carcinogenesis. CRC Crir. Rec. Toxicol. 2, 419. Boveri, T. (1929). The Original o f Malignant Tumors. Wil liam and Wilkins, Baltimore, MD. Ccrna. M. & Kypenova. H. (1977). Mutagenic activity of chloroelhylenes analysed by screening system tests. Mutation Res. 46, 214. Chan, C. H. (1975). Primary carcinoma of the liver, Afed. Clins N. Am. 59, 9S9. Table 5. Data on cytotoxicity o f 1,4-dioxane in Sprague-Dawley rats Treatment Dose Liver weight: (me/kg/day) body weight* Hepatic DNA synthesis* Hepatic histopathology t Exposure rid drinkingwater for 11 wk 10 1000 x 100 x M2 x 1-23 x 1-50 + Expressed as multiples of the control value. tTrealmenl-relaicd ejects were observed ( + ) or no treatment-related effects were observed (-k Genetic and nongenetic eventi in neoplasia 575 :'t pfOCOS . ..... Nr " ;d;n4 ; 3 h^ -.j* be used risii !>c mcchantodies cited ;cmotens do . which have pose greater -no mechan-eniicals hav- of action are jve reversible ml appear to ;; : i Y\ -i. u d If & Lee. F. D. pie test system n and bacteria 1.70, 2231. (1965). Studies rtn denaturing jn e. J. Natn. sii j. U j: '-j ,.;ti ii; -i' 3c Planche, G. $ I n of halo- ' ! robutencs pro- i^ . Archs Toxicol. i JTie Freezing .:iiJ iu, 179. oogenesis. Archs i vr , and unresolved tr jnsf. 60, 723. v. iI978). Enhance- 1 uion oT cultured 'll A4 mechanism of Rit. Tox/coL 2, mi Tumors. W1- i] igemc activity or r| '$ >stem tests, - i t -f the liver. MeJ. j1 js 1 tj b fi ' j *r.j .t . j] k V\ `1 N li Comings. D. E. (1973). A general theory of carcinogenesis. Proc. nain. Acad. Sa. US.A. 70, 3324. Craddock, V. M. (1973). Cell proliferation and induction of liver cancer. In Primary Lifer Tumors. Edited by H. Remmer, H. M. Bolt, H. Popper & P. Bannaschi. p. 377. M.I.T. Press Limited. Boston, MA. D3te, P. A.. Gotoskar, S. V. Sc Bhide, S. V. ( 1976X Effect of partial hepa'teclomy on tumor incidence and metab olism of mice fed ihioacetamide. J. nain Cancer nut. 56, 493. Drake, J. W. (197S). Some guidelines for determining maxi mum permissible levels of chemical mutagens. In Advances in Modern Toxicology, Vol. 5. Edited by W. G. Flamm Sc. M. A. Mehlman. p. 9. Hemisphere Publishing Corp., New York. Grasso, P. & Golberg. L. (1966). Subcutaneous sarcoma as an index of carcinogenic potency. Fd Cosmet. Toxicol. 4, 297. Greim. H-, Bonse, G , Radwan, Z , Reichert, D. & Henschler, D. (1975). Mutagenicity in vitro and potential carcinogenicity or chlorinated ethylenes as a function of metabolic oxirane formation. Biochem. Pharmac. 24, 2013. Hanawalt. P. C , Friedberg, E. C. Sc Fox, C. F. (197S). DNA Repair Mechanisms. Academic Press Inc. New York. Hartman. P. E. (1980). Bacterial mutagenesis: Review of new insights. Envir. Mutagen, 2, 3. Hoch-Ligeii. C , Argus, M. F. & Arcos. J. C (1970). Induc tion of carcinomas of the nasal cavity of rats by dioxane. Br. J. Cancer 24, 164. Holliday. R. (1979). A new theory of carcinogensis. Br. J. Cancer 40. 5 13. Huberman. E.. Mager, R. & Sachs, L. (1976). Mutagenesis and transformation of normal cells by chemical carcino gens. Sature, Lend. 264, 360. Jensen, E. V. (1974). Some newer endocrine aspects of breast car.cer. S ew Engl. J. Med. 291, 1252. Kirkland. D. J.. Smith, K. L. & Van Abb, N. J. (1931). Failure of chloroform to induce chromosome damage or sister-chromatid exchanges in cultured human lympho cytes and failure to induce reversion in Escherichia coli, Fd Cornier. Toxicol. 19, 651. Kociba, R. J, McCollister, S. B,, Park. C , Torkelson, T. R. Sl Gehring. P. J. (1974). 1,4-Dioxar.e, I. Results of a 2-year insestion study in rats. Toxic, appi Pharmac. 30, 275. Kondo, S. (1976). Misrepair model for mutagenesis and carcinogenesis. In Fundamentals in Prevention o f Cancer. Edited by P. N. Magee, S. Takayama, T. Sigimura Sc T. Matsushima, p. 417. Tokyo University Press, Tokyo/ University Park Press, Baltimore. MD. Kondo. S. (1977). A test for mutation theory of cancer: Carcinogenesis by misrepair or DNA damaged by 4-miroquinoline 1-oxide. Br. J. Cancer 35, 595. Laroye. G. J. (1974). How efficient is immunological sur veillance against cancer and why does it fail? Lancer I. 1097. Lindahl. T. Sc Karlstrom. O. (1973). Heat-induced depyrimidalion oi deoxyribonucleic acid in neutral solution, Biochemistry. N.Y. 12, 5151. Lindahl. T. Sc Nyberg. B. (1972). Rate of depurination of native Blochemisiry, /ST.Y. 11, 3610. Lindahl. T. & Nyberg, B. (1974). Heat-induced deaminaHon of cytosine residues in deoxyribonucleic acid. Biochemistry, N.Y. 13, 3405. Lippman. M. (1975). Steroid hormone receptors in human roalignancy. Life Sci. 18. 143. t fr* Forney, D. J., Mendrala. A. L., Konzc- Thomas, b. & McCormick. J. J. 11979). DNA cxdsinn- repair processes in human cells can eliminate the cyio- *0v,c anJ mutagenic consequences of ultraviolet ir Res. 62, 311. cK,nneIL R. G., Deggins, B. A. & Labai, D. D. (1969). Transplantation of pluripotential nuclei from tripluid frog tumors. Science, N.Y. 165, 394. Mtnlz, B. Si Illmensee, K. (1975). Normal genetically mosaic mice produced from malignant leralocarcinomu celts. Proc. nain. Acad. Sci. U.S.A. 72, 3535. National Cancer Institute. (1976). Carcinttycnvsis BUmssay o f Chloroform. Nat. Tech. Inf. Service No. PB264018/AS. NCI, Bethesda. MD. National Cancer Institute (1977). Bioassay o f Tetrachloroethytene fo r Possible Carcinogenicity. DHEW Publica tion No. 77-813. NCI, Bethesda, MD. National Institute Tor Occupational Safety and Health (1977) . Memorandum to Oificc/Division Directors, NIOSH? Mutagenicity Task Force from G. Taylor, NIOSH. Morgantown, WV, December 9. National Toxicology Program (1980). NTP Technical Bull. Vol. 1(2), April. Papaioannou, V. E., McBuriiey, M. W., Gardner, R. L & Evans, M. J. (1975). Fate of tcratocarcinoma cells injected into early mouse embryos. Nature, Land. 258, 70. Peto, R. (1977). Epidemiology, multistage models and short term mutagenicity tests. In Origins o f Human Cancer. Edited by H. H. Hiatt. J. D. Watson Sc. J. A. Winsten. p. 1403, Cold Spring Harbor Laboratory. Rampy, L. W,, Quast, J. F.. Leong, B. K. J. & Gehring. P. J. (1978) . Results of long-term inhalation toxicity studies on rats of 1,1.1-trichloroeihane and perchloroethylene formulations. In Proceedings o f the First International Congress on Toxicology Edited by G. L. Plaa & W. A. M. Duncan, p. 562. Academic Press, New York. Reitz, R. H , Quast, J. F,, Stoll. W. T,, Watanabe, P. G. & Gehring, P. J. (1980). Pharmacokinetics and macromolecular effects of chloroform in rats and mice: Impli cations for carcinogenic risk estimation. In Water Chforination: Environmental Impact and Health Effects. Vol. 3. Edited by R. L. Jolley, W .A. Brungs. R. B. Cumining Sc V. A. Jacobs, p. 983. Ann Arbor Science Publishers inc., Ann Arbor, ML Roe. F. J. C , Palmer, A. K , Worden, A. N. & Van Abb (1979) . Safety evaluation of toothpaste containing chloroform I. Long-term studies in mice. J. Envir. Path. Toxicol. 2, 799. San, R. H. C. Sc Stich, H. F. (1975). DNA repair synthesis of cultured human cells as a rapid bioassay for chemical carcinogens. Int. J. Cancer 16, 284. Sarasin, A. & Benoit. A. (19S0). Induction o f an errorprone model of DNA repair in UV-irradiaied monkey kidney cells. Mutation Res. 70, 71. Schumann, A. M,, Quast, J. F. Sc Watanabe, P. G. (1980). The pharmacokinetics and macromolecular interactions of perchloroethylene in mice and rats as related to onco genicity. Toxic, appl. Pharmac. 55, 207. Segaloff, A. (1975). Steroids and carcinogenesis. J. Steroid Biochem. 6, 171. Shank. R. C. Sc Barrows, L. R. (1980). Toxicity dependent DNA melhylation-sigmficance to risk analysis. In Health Analysis. Edited by C R. Richmond & P. J. Walsh. Franklin Press, Philadelphia (in press). Slaga. T. J., Fisher, S. M,, Viaje, A , Berry. D. L.. Bracken, W. M., LeClerc. S. Sc Miller, D. R. (1978). Inhibition of tumor promotion by anti-inflammatory agents: An approach to the biochemical mechanism of promotion. In Carcinogenesis. Vol. 2. Mechanisms of Tumor Promo tion and Carcinogenesis. Edited by T. J. Slaga. A. Sivak & R. K. Bout well. p. 173. Raven Press, New York. Tcmin, H. M. (1974). On the origin of the genes for neo plasia: G. II. A. Clowes Memorial Lecture. Ctirtcer Res. 34. 235. ^ Tennekes. H. A. (1979). The Relationship Between Microsunuil F.ncyme induction and Liver Tumor Formalion. Centre for Agricultural Publishing and Documentation.' Wagenigen, Netherlands. \ W. T. Stutt. R. H. Rht2. A. M. Schumann and P. G. Watanabk .erraeini. B.. Magee. P. N. & Barnes, J. M. 11067). Hepatic - pathology in rats on low dietary levels of dimethylniiro- samine. Br. J. Cancer 21, 559. Torkelson. T. R., Leong. B. K. J., Kociba, R. J., Richter, W. A. & Gehring, P. J. (1974). 1.4-Diox ane. II. Results of a 2-year inhalation study in rats. Toxic, uppl. Pharmuc. 30, 2S7. Trosfco. J. E. & Chang. C. C. (1978). Relationship between mutagenesis and carcinogenesis. Phntochem. Photobini. 28, 157. Watanabc. P. G,, Reitz. R. H., Schumann, A. M , McKenna. M. JTMQuast, J. F. & Gehring. P. J. (1930) Implications of the mechanisms of tumorigeniciiy for risk assessment. In The Scientific Basis of Toxicity Assessment. Vol. 6. Edited by H. Witschi. p. 69. Elsevier/ North-Holland Biomedical Press. New York. Weeks. C. E,, Slaga. T. J.. Hennings, H,, Gleason. G. L. & Bracken. W. M. (1979). Inhibition of phorbol esterinduced tumor promotion in mice by vitamin A analog and anti-inflammatory steroid. J. iiam. Cunt.er Inst. 63, 401. Weisburger, J. H. & Williams, G. M. (1980). Chemical car cinogens. In Casureit amt Douirs Toxicology. The Basic Science of Poisons. Edited by J. Deuil. C. D. K1aasen & M. O. Amdur. p. 84. MacMillan Publishing Co., Inc, New York. Williams, G. M. (1977). Detection of chemical carcinogens by unscheduled DNA synthesis in rat liver primary cell cultures. Cancer Res. 37, IS45. Wilkin, E. M. (1976). Ultraviolet mutagenesis and induc ible DNA repair in Escherichia coti. Bact. Rec. 40, 869. Wright. A. S., Akintonwa, D. A. A. & Wood er. M. F. (1977). Studies on the interactions or dieldrin with mam malian liver cells at the subcellular level. Ecotox. cneir. Saf. 1 ,7 . Yamagiwa, K. & Ichikawa. K. (1918). Experimental study of the pathogenesis of carcinoma. J. Cancer Res. 3. I. FJ* ftir Int ma mil IT: cot ied giv SOI act mi pci ein Wi tw epi I ca| re: th: get bu pa, chi ini sh< hej cat 511! gei i& th; [ \\ cn: ter Hi' Cti: ad I pt; F I \