Problems in consideration of rodent hepatocarcinogenesis for regulatory purposes

Hepatoproliferative lesions of rodents are frequently reported in petitions containing pathology data from chronic toxicity and carcinogenicity studies submitted to the Center for Food Safety and Applied Nutrition of the Food and Drug Administration. The Pathology Branch of the Office of Scientific Analysis and Support evaluates these data, which are submitted in support of the safe use of food additives, color additives, and other regulated products. Data are reviewed for the adequacy of the information provided, the terminology used to describe the reported lesions, and the overall scientific rationale used in interpreting the biological significance of the observed lesions. When questions arise during the review process, additional data, information, or clarification are sought from the petitioner. Microslides may be requested from the petitioner so that an independent evaluation of the lesions may be conducted. Several examples of recent evaluations of hepatoproliferative lesions are presented to illustrate some of the problems encountered during the review process and to demonstrate the procedures and approaches used in the evaluation of hepatocellular lesions within the center.

Analysis of loss of heterozygosity in murine hepatocellular tumors

Because allelotype analysis of tumors has been important in the identification of new tumor suppressor genes, here we studied loss of heterozygosity (LOH) in a well-defined animal experimental system. We analyzed spontaneous liver tumors from C3HHc x C57BL/6J (B6C3F1) mice and urethane-induced hepatocellular tumors from (C3H/He x Mus spretus) x C57BL/6JBy (HSB) interspecific mice. A total of 95 different genetic markers were tested: 13 in 24 B6C3F1 tumors, 76 in 58 HSB tumors, and six in both groups. Minisatellite finger-printing analysis detected one case of LOH and less than 1% genomic rearrangements in polymorphic and nonpolymorphic bands, respectively. There were no changes at hepatocellular susceptibility loci or at markers homologous to loci frequently lost in human hepatocellular carcinomas. Therefore, our results suggest that LOH and genomic rearrangements are uncommon in mouse hepatocellular tumors.

Maximum dosage level in testing low-toxicity chemicals for carcinogenicity in rodents

Owing to the lack of sufficient theoretical and empirical information, the initial guidelines regarding animal carcinogenicity testing of chemicals adopted the most conservative approach possible. One of the recommendations was that non-toxic chemicals be tested at a level as high as 5% of the diet. Since then, a wealth of information has been accumulated, which indicates that such highly exaggerated dosage levels are not only unnecessary but produce scientifically misleading and regulatorily detrimental results that impede the development and evaluation of useful chemicals, including human drugs. This paper presents the rationale supporting the necessity of revision of the outdated maximum level of dietary exposure from 5% to 1% or 1000 mg kg-1 day-1 when the test chemical is administered in drinking water or by gavage.

Classification of carcinogens: polemics, pedantics, or progress?

Rodent carcinogens may, for physiological or other reasons, induce cancer by a variety of mechanisms which vary in their ability to affect humans. While the current approach of some regulatory agencies to carcinogen risk assessment and regulation may possibly be justified with most genotoxic carcinogens, this is not true with all nongenotoxic carcinogens. Mechanisms attributable to high dose toxicity occasioned by misuse of the maximum tolerated dose concept, imbalancing of homeostasis, unphysiological conditions, and induced cellular proliferation are reviewed. The greatest present need for meaningful regulation of carcinogens is to obtain public acceptance of the fact that some carcinogens are species specific and probably will not exert their effects in humans.

Adjudicating cancer causation: scientific, political, and legal conflicts

Lawsuits concerning cancer causation resort to scientific argumentation. Yet, the apparent ambiguities of science confuse the courts, the juries, and the public. This is especially so with regard to official regulatory definitions of cancer causation that carry the weight of law. At the heart of this problem is a prevailing misunderstanding of science and the scientific method, and of the limits of current scientific knowledge about cancer. Moreover, current regulatory policies encourage the public to perceive official cancer risk assessments as if they were scientifically derived and accepted, even though official fine print readily admits they are not. Some recent court decisions have begun to recognize these difficulties with a body of precedent, and this may result in future rulings influenced more by objective appraisals than by reliance upon official but contingent assumptions.

Occurrence and relevance of chemically induced benign neoplasms in long-term carcinogenicity studies

Recent carcinogenicity studies conducted and evaluated by the National Toxicology Program/National Institute of Environmental Health Sciences were examined to determine the frequency of chemically increased incidences of neoplasia. Many of the chemicals originally selected for study were chosen because of an a priori suggestion that they might be carcinogens. Of the 143 chemical studies evaluated, usually involving male and female rats and mice, 42 (29%) did not induce any neoplasms, 20 (14%) gave marginal or equivocal neoplastic responses, and 81 (57%) showed positive neoplastic responses in one or more of the 524 species-gender experiments. Of these 81 positive studies, 60 (74%) were considered positive based on malignant neoplasia, 16 (20%) were positive due primarily to benign neoplasia, but had supporting evidence of malignant neoplasia in the same organ/tissue, and 5 (6%) were positive based only on benign neoplasia. These five chemicals are a) allyl isothiocyanate (transitional cell papillomas of the urinary bladder in male rats), b) 2-amino-4-nitrophenol (tubular cell adenomas of the kidney in male rats), c) asbestos intermediate range chrysotile (adenomatous polyps of the large intestine in male rats), d) decabromodiphenyl oxide (neoplastic nodules of the liver in male and female rats), and e) nitrofurazone (fibroadenomas of the mammary gland in female rats and benign mixed tumors and granulosa cell tumors of the ovary in female mice). For all but one of these lesions (mammary gland), the occurrence in historic controls is low. Thus, only 5 of the 143 chemicals studied (3.5%) induced benign neoplasia alone, and those observed benign neoplasms are known to progress to malignancy. Accordingly, we consider chemically induced benign neoplasia to be an important indicator of a chemical's carcinogenic potential in rodents, and believe it should continue to be made an integral part of the overall weight-of-the-evidence evaluation process for identifying potential human health hazards.

Critical parameters in the quantitation of the stages of initiation, promotion, and progression in one model of hepatocarcinogenesis in the rat

Critical parameters in the quantitation of altered hepatic foci (AHF) developing during multistage hepatocarcinogenesis in the rat include: 1) the enumeration of AHF induced by test agents as well as those AHF occurring spontaneously in livers of untreated animals; 2) the volume percentage or fraction of the liver occupied by all AHF as a reflection of the total number of altered cells within the liver and the degree of tumor promotion which has occurred; and 3) the phenotype of individual AHF as determined by multiple markers with serial sections. These parameters, especially the number of AHF, should be corrected by the presence of spontaneous AHF which increase with the age of the animal, more so in males than females. While accurate estimation of the background level of spontaneous AHF can be important in demonstrating that a carcinogenic agent does not possess the ability to increase the numbers of AHF above the background level, a better method to distinguish the effectiveness and relative potencies of agents as initiators or promoters is reviewed. The relative effectiveness of four different markers--gamma-glutamyltranspeptidase (GGT), a placental form of glutathione S-transferase (GST), canalicular ATPase, and glucose 6-phosphatase (G6Pase)--was described for the chemicals C.I. Solvent Yellow 14 and chlorendic acid as promoting agents in males and females. C.I. Solvent Yellow 14 is a more effective promoting agent in females than males, and AHF exhibit extremely low numbers scored by GGT. On the other hand, the numbers of AHF present in livers of male rats promoted by this agent are more than twice those seen in livers of female animals, possibly owing to the effectiveness of this agent as an initiator in the male but not the female. Very few AHF, especially in the male, are scored by GGT during chlorendic acid promotion. The distribution of phenotypes with these markers also differs in the spontaneous AHF appearing in the livers of animals fed 0.05% phenobarbital on either a crude NIH-07 or AIN-76 purified diet. Such studies emphasize the extreme dependence of the promoting stage of hepatocarcinogenesis on environmental factors of sex, diet, and the molecular nature of the promoting agent itself. The hallmark of the final stage of progression in the development of hepatocellular carcinomas is aneuploidy, which may be reflected by phenotypic heterogeneity within individual AHF, termed foci-in-foci. The implications of such quantitative analyses during hepatocarcinogenesis induced by specific agents in relation to the specific action of the agent at one or more of the stages of hepatocarcinogenesis are discussed.

IARC and HHS lists of carcinogens: regulatory use based on misunderstanding of the scope and purpose of the lists

In the last two decades there has been a tremendous increase in data on carcinogenic activity in experimental animals. While there have been few additions to the list of human carcinogens based on human data, the number of carcinogens based on animal data continues to increase unabated. The International Agency for Research on Cancer (IARC) list of carcinogens grew out of the IARC Monograph Series. The evidence classification system used to prepare the IARC lists in 1980, 1982, and 1988 is based on the sufficiency, i.e., strength, of the evidence of carcinogenic activity in one or more studies, not a full weight-of-the-evidence evaluation of all relevant data. Titles of categories of animal evidence referring to human risk potential were based on a presumption: "for practical purposes . . . as if." No evaluation was made of the predictive relevance of animal data to human risk. The IARC listing did not involve evaluation of potency or mechanism and was intended as a useful input but not as a basis for regulatory or legislative decisions. The Department of Health and Human Services (HHS) lists in the Annual Reports on Carcinogens are selected from the IARC lists and from reports of positive bioassay experiments conducted by the National Toxicology Program (NTP). The reports on the NTP bioassays relate to the strength of the evidence in each experiment and recognize that a "wider analysis" is necessary for determination of human risk. Because of a misunderstanding of the limited scope of the analysis involved, the IARC and HHS lists have recently been used as a basis for legislative and regulatory decisions. Examples of unanticipated use of the lists as triggers for regulatory and legislative decisions will be discussed. Some recommendations to mitigate the consequences of past unanticipated use of the lists and to prevent further misuse are discussed.

Carcinogenicity of chemicals: the weight of evidence

1. The evidence discussed here is derived from epidemiology, long-term bioassays in laboratory animals, and predictive short-term tests. 2. Epidemiological data are obtained directly from human studies and are most compelling when they demonstrate a large relative risk and a clear dose-response in association with a distinctive tumour type. Exposure to a suspected carcinogen and the doses involved are, however, often difficult to determine, and the most sophisticated epidemiological methods are relatively insensitive. There are no epidemiological data for most occupational/environmental chemicals. 3. Long-term bioassays can present major problems in design, interpretation and extrapolation. Particular difficulties are associated with the planning of appropriate dose levels and the occurrence of certain tumours at high incidence in both control and test groups. Results from animal bioassays set priorities for concern and action but they cannot be reliably used for quantitative assessment of human risk. 4. Evidence of potential carcinogenicity derived from short-term predictive tests, involving a wide variety of systems with diverse end-points, is increasingly important. Emphasis is placed on the need for more in vivo procedures with a broadening of the scope of somatic cell targets.

The contribution of the IARC monographs program to the identification of cancer risk factors

The differences between cancers that occur as a consequence of occupational exposure and other cancers are not only their preventability but, more importantly, their social unacceptability. Occupational cancer occurs, by definition, among individuals who have been exposed to carcinogens because of their occupation, and most, if not all, of these individuals are drawn from the less favoured social classes. This is probably one reason why mortality from cancer and from all causes is greater in people in classes IV and V than in those in I and II. We cannot precisely quantify the proportion of cases, among the total number of cancer cases observed, attributable to occupational exposures. It is very likely, however, in some of the most industrialized countries, although not in certain developing countries, that the number of those cancers that are indisputably due to occupational exposure is not increasing and is perhaps decreasing. This is due to the combined effect of two factors: the banning of certain chemicals, as, for instance, aromatic amines--even if this did not take place in all countries and, where it did, not at the same time; and improved working conditions, as, for instance, in the case of vinyl chloride. We do not know, however, to what extent low levels of exposure to which the general population is commonly exposed, as well as workers in occupations where levels of exposure to carcinogens have recently been significantly reduced, play a role in the causation of human cancer. While it is important to stress that a large proportion of the chemicals to which humans are exposed, either because of their occupation or in the general environment, and for which experimental evidence of carcinogenicity is available, have not been the object of epidemiological surveys, it is also important to realize that epidemiological methods are generally insufficient to provide reliable information on risks generated by low levels of exposure. It is certainly important to encourage epidemiological surveillance, but it should at the same time be made clear that the epidemiological approach will never entirely replace the considered use of experimental data in the implementation of primary prevention.

Activation of a cellular proto-oncogene in spontaneous liver tumor tissue of the B6C3F1 mouse

The controversy surrounding the interpretation of observed increases in the high spontaneous liver tumor incidence of the B6C3F1 mouse after administration of certain chemical agents necessitates a mechanistic understanding into the nature of tumor development in this particular strain of mouse. Recently, cancer genes (oncogenes) have been detected in the DNA from a variety of human tumors and tumor cell lines. These genes have been implicated to play a role in the transformation of normal cells into cancerous ones. To investigate the role that cellular oncogenes might play in the development of spontaneous liver tumors in the B6C3F1 mouse, DNA was isolated from spontaneously occurring liver tumors and transfected into NIH 3T3 fibroblasts. DNA from this tumor tissue was capable of transforming NIH 3T3 cells from 82% of the animals examined strongly suggesting the presence of an active cellular oncogene. In contrast, DNA isolated from surrounding non-tumorous liver tissue and liver tissue from non-tumor bearing mice did not cause any transformation in the NIH 3T3 assay. These data demonstrate that the active cellular oncogene is not present in the hepatic tissue via a germ-line transmission but is activated only in those cells of the tumor tissue. Experiments using Southern blot hybridization analysis have identified this active cellular oncogene to be a member of the ras oncogene family. Identification of this cellular oncogene will now allow the evaluation of factors which might modify its expression. These future studies will lead to an increased understanding of potential mechanisms by which hepatic tumors are enhanced and should provide more informed estimates of risk for man based on bioassay data generated in this strain of mouse.

Identification and classification of carcinogens: procedures of the Chemical Substances Threshold Limit Value Committee, ACGIH. American Conference of Governmental Industrial Hygienists

The Chemical Substances Threshold Limit Value Committee of the American Conference of Governmental Industrial Hygienists has refined its procedures for evaluating carcinogens. Types of epidemiologic and toxicologic evidence used are reviewed and a discussion is presented on how the Committee evaluates data on carcinogenicity. Although it has not been conclusively determined whether biological thresholds exist for all types of carcinogens, the Committee will continue to develop guidelines for permissible exposures to carcinogens. The Committee will continue to use the safety factor approach to setting Threshold Limit Values for carcinogens, despite its shortcomings. A compilation has been developed for lists of substances considered to be carcinogenic by several scientific groups. The Committee will use this information to help to identify and classify carcinogens for its evaluation.

Detection of mutations in bacteria and of DNA damage and amplified DNA sequences in mammalian cells as a systematic test strategy for elucidating biological activities of chemical carcinogens

The interdisciplinary evaluation of risks from carcinogens utilizes, inter alia, data on the activities of the compounds in short-term assays. A systematic approach is being used to determine mutagenesis in bacteria (the study of direct activities and specific modes of metabolic activation), DNA damage within primary mammalian cells (DNA single-strand breaks and persistence of damage, by a method extendable to the in vivo situation) and amplified DNA sequences in cultured cells (as an endpoint probably relevant to carcinogenesis). This test combination was expected to reduce some of the shortcomings of other batteries of tests, which suffer from a lack of appropriate metabolic conversion of compounds, irrelevancy of genetic endpoints and pharmacokinetic limitations. Furthermore, as each assay in the test strategy differs from the others only by one of the parameters described above, a reasonable understanding of divergent test results from assay to assay was anticipated. Several substances were investigated to elucidate why their activities in short-term assays and in carcinogenesis experiments do not correlate. The substances were N-nitrodimethylamine, for which formaldehyde is the reactive intermediate in bacterial mutagenesis but not in mammalian cells or in vivo, N-nitrosodiethanolamine, a carcinogen that must be activated by external alcohol dehydrogenase to be mutagenic in bacteria, N-nitrosodialkylamines, with unique organotropism in vivo for which organ-specific activation was studied in vitro, N-nitroso compounds that are inactivated in vivo but not in vitro, and components of the aristolochic acid mixture which may be metabolized oxidatively or reductively, as well as numerous miscellaneous compounds that were expected to be genotoxins on account of their chemical structure. In addition to the assessment of genotoxicity, the results obtained in individual tests of this strategy yield important data on mechanisms of activity, such as organ-specific activation and deactivation, species variations, in vitro/in vivo correlation and persistence or repair of damage.

The effect of dietary administration of Disperse Blue 1 on the urinary system of the Fischer 344 rat

Disperse Blue 1 (containing 50% lignosulphonate dispersants) was fed to Fischer 344 rats at dietary levels of 0.01 and 0.1% for 19 months and at 1.0% for 6 months. Fischer 344 rats were also given the dye by gavage at 1 g/kg for 1-3 days or in the diet at 0.5 or 1% for 4 days, and corresponding dietary levels of the colouring without dispersant were also fed for 4 days. Bladders and kidneys were examined after the 1-4 day treatments, in animals dying or killed from month 6 to termination (19 months) in the chronic study and in those killed at wk 5, 9 and 17. At the latter three times, autoradiography following injection of tritiated thymidine showed increased DNA synthesis in the urothelium of high-dose rats, but no other increased labelling in any group. Bladder lesions were seen only at the 1.0% level, epithelial erosion with adhering dye particles being seen by day 4, calculi and hyperplasia by wk 5 and squamous metaplasia by wk 9. The calculi contained more dye in males than in females and more calcium in females. By month 6, dye particles were embedded in the bladder wall, with some evidence of histiocyte accumulation in their vicinity. Two papillomas and one carcinoma, but no leiomyosarcomas, were diagnosed. The earliest tumours, two papillomas, were detected at wk 17. Tumour incidence following surgical removal of calculus was about double that in rats not subjected to surgery and the incidence of normal bladders at month 19 was higher in the latter group. Compound-related effects in the kidneys--inflammation, pelvic epithelial hyperplasia and tubular degeneration and regeneration with interstitial fibrosis--were seen only in the high-dose group. Dye present in the tubules and renal pelvis persisted in many rats for a year after cessation of treatment.

Nutrition, cancer, and aging: an annotated review. I. Diet, carcinogenesis, and aging

The interrelationships of diet and carcinogenesis are discussed with the focus on aging. To establish whether the elderly are more susceptible to dietary carcinogens and whether dietary prevention of cancer is a reasonable goal for this population, the mechanisms of chemical carcinogenesis, the age-related metabolic and physiologic changes, and the current cancer preventive dietary strategies are reviewed. Vulnerability to dietary carcinogens results from a combination of factors that may increase or decrease the occurrence of cancer in the elderly, and it is, therefore, a very individualized feature, unpredictable when based solely on a subject's age. Dietary prevention of cancer may be effective in advanced age, and the dietary guidelines of the National Academy of Sciences should be implemented in this population.

Role of fat, fiber, nitrate, and food additives in carcinogenesis: a critical evaluation and recommendations

This review presents a critical, select evaluation of the amount and type of fat or fiber in nutritional carcinogenesis, with the emphasis being on cancer development in the mammary gland and large bowel. The role of nitrate and nitrosation is described in relation to risk for cancers of the head and neck (especially the esophagus) and cancers of the stomach and the liver. Systematic tests of increasing complexity to delineate possible carcinogenic risk in food additives and contaminants are described. Specific recommendations stemming from these evaluations are made as to dietary recommendations designed to reduce cancer risk.

Permissible concentrations of chemicals in air and water derived from RTECS entries: a "rash" chemical scoring system

Many chemicals are of concern to human health, but only a few have epidemiologically derived risk estimates. About 45,000 chemicals are listed in RTECS, most of which have had some testing in subhuman models. RTECS entries range from cellular effects through organoleptic damage to lethality, with many pathological endpoints listed, including mutagenic changes, irritation, teratogenesis, cancer, mortality, etc. However, it is difficult to extend any biological test results to human risk assessments. If the results are extended, the degree of validity is highly uncertain. This paper describes a logical basis for using the entire complex spectrum of test results to evaluate the overall toxicological potency of a chemical to be assayed (i.e., an interviewing chemical) and describes how to derive tentative, permissible concentrations in air and water for any particular chemical for which no regulatory guidance exists. This approach has been tested for 16 reference chemicals discussed in NIOSH Criteria Documents, EPA-CAG reports, etc. The evaluations are uncomplicated, but occasionally it is difficult to match RTECS entries for two different chemicals. Difficult comparisons may require some familiarity with experimental design and the toxicological literature. One important product of this novel approach is that a distribution or array of potency values is obtained for any chemical evaluated. This distribution reflects many uncertainties stemming from low statistical power, experimental design, pharmacological processes, interspecies variability, dose rate, biological effect monitored, route of treatment, etc. The array of relative values for a particular chemical reflects many different biological and physical conditions. The distribution of the array helps to index a composite toxicological profile for many different biological effects resulting from numerous treatment protocols. To minimize the effect of extreme sensitivity of certain (perhaps novel) biological test models, possible errors in the RTECS data-base, and possible human pharmacological insensitivity to a particular chemical and/or a particular route of administration, we consider the interquartile range (i.e., the central 50%) of the array of relative potency values between two chemicals being compared as a practical measure of uncertainty. Thus, the range in response derived from variability in relative potency should be useful in addressing the range of response in man as estimated from extrapolations of test data.

Detection of a cellular oncogene in spontaneous liver tumors of B6C3F1 mice

An active cellular oncogene was demonstrated in hepatocellular neoplasms arising spontaneously in 24-month-old B6C3F1 mice. DNA isolated from the tumorous tissue and transfected into NIH 3T3 cells showed an 82 percent (9 of 11 animals) frequency of foci induction. In contrast, DNA isolated from the surrounding nontumorous hepatic tissue from the same animals and DNA from other 24-month-old B6C3F1 mice without tumors did not cause transformation in the NIH 3T3 cell assay. This strain of mouse is used extensively in carcinogen bioassays, and the observed high frequency of transformation (82 percent, compared to 10 to 20 percent in humans) supports the concept that the B6C3F1 mouse is hypersusceptible to liver tumor development. It also emphasizes the need to further understand the mechanisms of oncogene activation in animals used for long-term studies of toxicity and oncogenicity before evaluating potential human risk.