Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
Chapter 4 CARCINOGENIC POTENCY AND RISK ESTIMATION In this chapter the committee uses information concerning the carcinogenicity of aromatic amines to suggest a means for expressing concepts of carcinogenic potency and for estimating risk. Aromatic amines are excellent candidates for risk assessment. They have been assayed for carcinogenicity in several animal species and strains, and some members of this class (2-naphthylamine, benzidine, 4-aminobiphenyl, and probably phenacetin) are indeed bladder tumor inducers in humans. 1-Naphthylamine, as manufactured at one time, was associated with bladder cancer in workers; this was probably due to a high level (4-10%) of contamination with 2-naphthylamine. Of course, there are no reliable data concerning the level of adventitious human exposure in the workplace or elsewhere, although the amount of chemical administered to laboratory animals is readily determined. Limited data may possibly be obtained from studying iatrogenic carcinogens, e.g., 2-naphthylamine mustard (Thiede and Christensen, 1975), but in such cases, animal data are usually deficient. CARC INOGENI C POTENCY Ire potency of a carcinogen depends on three factors: the dose of carcinogen required to induce tumors, the time to tumor induction, and the percentage of tumor response. For purposes of the following discussion the committee has defined potency as 86
potency = 7 -log dE50 where dE50 is the dose expressed in ~mol/kg body weight/week required to induce tumors in 501 of the animals in a lifetime experiment. This dose is expressed as a logarithm in order to compress the range of values to a readily understandable format and to avoid undue emphasis on small differences in potency that lack biologic significance. The negative of the logarithm is used because potency is inversely related to the dose required to induce tumors, and the number 7 is used to bring all values to a readily comprehensible positive form. As calculated according to the above criteria, typical potencies of a number of carcinogens that induce liver or bladder tumors in rat or mouse are shown in Table 4-1. The data illustrate the range of potency from the most potent (aflatoxin B1) to the least potent (trichloroethylene and saccharin) carcinogens. The calculations used to determine these values are admittedly approximate . For example, an approximation has been made that in attaining a 50% dose response, the dose experimentally shown to give a 25% tumor incidence has been doubled; likewise, doses that lead to tumors in half the lifespan of the animal have been halved to calculate dESo. me se assumptions may well be adequate because the logarithmic scale compresses small differences. The values for the lifespan of several test species and the ir food and water consumption requ irements are compared in Table 4-2. mese are the values used for dE50 calculations. Better approximations would lead to more accurate potency values. 87
Table 4-1 The Potency of a Range of Carcinogens to Rat or Mouse Liver Following Continuous Feeding Chemica 1 Spec ies logic dESO Potency (mg/kg/week ) Aflatoxin B1 Rat 0.67 9.18 Michler's ketone Rat 4.88 4.62 Dimethylnitrosamine Rat 4 .90 4 .00 Carbon tetrachlor ide Ra t 5.27 3 .87 2-Aminoanthraqu ine Rat 6 .72 4 .44 Tr ichloroethylene Mouse 7.03 2. 12 88
Table 4-2 Factors Used for Calculating dE50 in Animals Factors (s) Mouse Rat Dog ifes n (years) 2 5 3 0 0 Food (g/day) 4 6 12 15 300-500 Drinking water (ml/day) 2.1 20.0 Not Gestation (days ~ 21.0 21.~0 63 Weight (q) 25-40 100-SOO 10,000 - a NK ~ not known 89
Certain features, of this potency derivation need emphasis. First, the dose rate Ml/kg body weight/week) is used in preference to the total dose to avoid giving the impression that long-lived species are less sensitive to carcinogens than are shor ter-1 ived species. For example, it has been accepted that dog e are sensitive to the carcinogenicity of 2-naphthylamine, and that mice are less sensitive. This finding is shown in Table 4-3. Total exposures (450-500 weeks in dogs compared to 100 weeks in mice) indicates that dogs and mice are more similarly sensitive to this chemical. The potency as calculated here defines one point on the tumor incidence-dose curve. If one accepts the linear one-hit model (National Academy of Sciences, 1976) of carcinogenesis, which, is probably not universally accurate, as shown later in this chapter, the potency value ef fectively def ines the slope of the dose-response curve. Whatever tumor incidence-dose model is used, the potency values discussed here have a considerable advantage in that only they may be der ived without excessive data extrapolation. In many cases, where two or more doses of a carcinogen have been used, interpolation rather than extrapolation may be required. Tables 4-4 and 4-5 show the potency values for 4-aminobiphenyl and methylene-bisto-chloraniline) in different species. Although the potency of a carcinogenic aromatic amine in different species appears to lie within two orders of magnitude, this impression, to a certain extent, is false; species that fail to respond to an aromatic 90
Table 4-3 Potency of 2-Naphthylamine in Different Spec ies Following Ore1 Administration Species loglo dE50 Potency tm9/kg/week ) Dog 4.92 4.34 Muse 5.23 3.93 Fee t 6 . 72 2 . 74 Hamster 6.72 2.14 91
Table 4-4 to Different Species Route of Species Administration Tissue Potency Dog Oral Bladder 6.22 Mouse Gavage Liver 4.52 Rat Subcutaneous Intestine 4.37 92
Table 4-5 Potency of Methylene-bis (o~chloraniline) in Var ious Species Following Oral Administra tion logic dE50 a Species Tissues (ag/kg/week) Potency Dog Bladder 4.57 ~ 4. 86 Ratb (adequate diet) Lung 4.81 4. 57 Breast 5.15 4. 25 Zymbal ' s gland 4 .85 4 .58 Liver 4.72 4.70 Ratb (low protein diet) Lung 5.08 4 . 35 Breast 5.48 3.95 Zymbal ' s gland 5.18 4.25 a Each value of dEs0 is calculated as life correction x tumor yield correction x dose x 105 ~g/kg/week. It is assumed that ra ts 1 ive 95-104 weeks and consume an average 105 g/food/week and that dogs live an average of 9 years, or 468 weeks. b Only male animals were used. 93
amine are omitted from the tables. It is not possible to decide whether these nonresponsive species are completely insensitive to the par ticular agent or are sensitive at a level too low to be detected in the an imal bioassay. The cutof f point for the sensitivity of a carcinogenesis bioassay is determined first by the toxicity of the chemical being tested and, second, by the maximum amount of test substance it is reasonable to give to an animal during any one period. The use of the maximum~tolerated dose, as recommended in the bioassay protocols of the National Cancer Institute/National Toxicology Program (Sontag et al., 1976), ensures that the highest feasible dose level is used, but that this level may lead to abnormal results due to the intervention of the agent's toxic properties in the carcinogenic process. However, substances with ~ very low level of toxicity may, if the maximum tolerated dose is used, be administered at levels that interfere physiologically with the host, for example, by inducing nutr itional imbalance, and similarly lead to difficulty in interpreting results. At present, it is not possible to predict with confidence that the demonstration of potency of a carcinogen in one animal species means the carcinogen will be as potent in another species. Thus, it may be prudent to assume (since it cannot be tested in bumans} that humans are at least as sensitive to a carcinogen as are the most sensitive species. Consequently, 2-naphthylamine in humans could be assumed to be as potent as it is in dogs (Table 4-3} and methylene-b~s- {o-chloraniline) to be as potent in humans as it is in 94
rats (Table 4-5~. Table 4-6 shows the carcinogenic potency of N,2-fluorenylacetamide in different species and strains of test animals. Ways of predicting carcinogenic potency based on the present knowledge of the meabanisms of carcinogenesis urgently need to be expanded. Crouch and Wilson (1979} discuss interapecies differences in carcinogenic potency, including humans; however, they do not describe how they calculate potency to humans in the absence of re 1 table exposure da ta . HIGH- TO LOW-DOSE EXTRAPOLATION It is now generally regarded as prudent to assume, when extrapolating data from high carcinogenic dose rates in animals to low~ose exposure in humans, that the mathematically simple linear one-hit model is appropriate even if the limited experimental evidence does not necessarily support this conclusion. The relevance of this model, which implies that a single carcinogenic itical receptor interaction is involved, has not been seriously questioned, although it may prove unsuitable in many cases. Chemical carcinogenes is is a multifaceted process . It is generally accepted that the procarcinogen has to be metabolically activated to the ultimate or reactive form, and the ultimate carcinogen then has to interact with its critical target. When the critical target is DNA, this interaction is followed by. DNA replication to flock inn genetic damage, or by DNA repair to restore 95
Table 4-6 The Carcinogen ic Potency of N. 2-Fluorenylacetamide in Different Species and Strains of Test Animala. Species Route Site of Tumor Potency Dog Diet Liver, bladder ~ . 5 Rabbit Gavage Bladder, ureter 4.46 Hamster Diet Gall bladder 4.29 Rat Slonaker (M+F) b Diet Bladder 4.40 Wistar (M) Diet Liver 5.14 Wistar (~) Diet Breast 5.03 Piebald (M+F) Diet Intestine 4.93 Mouse BALB/c (F) Diet Liver 4.17 BALB/c (F) Diet Bladder 4.17 a Some evaluations based on early studies b M = male; F = female. g6 .
the cel Is to quas i-norma 1 ity, and the damaged cells or tumor progen itor cel Is then undergo var ious biolog ic interactions before frank clinical neoplasia is observed. It is cliff icult to conceive all these changes involving only a single interaction between the carcinogen and its critical receptor (s) The s implest biolog ic model of carcinogenesis is the two-stage hypothesis, proposed by Berenblum and Shubik (1947a,b; 1949) in mouse sk in, and now being shown applicable for many other tissues such as liver, bladder, and pancreas. mis model may not effectively describe what may well be a more complex multi-stage process. Initiation and promotion are separate processes induced by different agents. Therefore, they are independent of each other. A complete carcinogen, however, is capable of both initiation and promotion. Since both processes are presumably dose dependent, an exponential rather than linear relationship should exist between dose of carcinogen and tumor response. A linear relationsh ip ply be applicable for pure initiators or pure promoters. The validity of these suggestions is suggested by the large scale [EDoi ~ exper iment conducted by the National Center for Toxicological Research (Staf fa and Mehlman, 1979), in which low levels of N,2-fluorenylacetamide were fed to BALB/c mice. The liver tumor yield was apparently related linearly at low doses to the carcinogen dose; the bladder tumor yield was not (Figure 4-1~. 97
1.0OI o.9o 0.8C ~ 0.70 o 0.60 o.sa o - CE 0.40 g 0 1 a 0.30 0.20 0.10 O _ o _ Liver 33 month O ~ All / Bladder ~ / Liver 24 my I I I ~ I ~ I ~ I I I i I I 20 40 60 80 100 1 20 140 DOSE (ppm) Figure 4-1 Comparison between liver and bladder tumor yield in response to duff ferent doses of N,2-fluorenylacetamide g iven over 33 months. These data are for BALB/c mice, (Staf fa and Hehlman, 1979 ~ . 98
Between 24 and 33 months, BALB/c mice developed a signif icant incidence of naturally occurring liver tumors, which the carcinogen enhanced. The naturally occurs ing incidence of bladder tumors was very low. Emus, one explanation of the dose-response relationships exhibited is that the carcinogen must act as both initiator and promoter in bladder carcinogenicity to give a tumor incidence-dose curve that is very dif ferently shaped than that for the liver, where only promotion occurs. These observations could mean that, although a linear dose-response curve may be appropr late in specif ic cases (possibly with pure initiators or with pure promoters acting on an appreciable spontaneous tumor yield), it may be inappropr late in other cases. The use of linear dose-response relationships may indicate a level of risk higher than actually occurs. This result is possibly benef icial in that it provides estimates at a time when methods for risk estimations are little understood. As ~ isk assessment techniques become more precise it may become possible to make estimates with greater accuracy. The use of these statistical models to estimate possible risk to humans at very low exposures is filled with uncertainty. Studies with animals usually involve exposure to a high level of a single carcinogen and, sometimes, just one modifying agent. In the real, nonexperimental world, humans are exposed to a wide range of carcinogens and carcinogenesis-modifying agents, which may enhance or inhibit cancer development due to low levels of a particular agent. The suggestion that a given agent will induce, for example, one tumor in a population of a million is, under these conditions, tr ite speculation. 99
REFERENCES Carcinogen ic Potency and Risk Estimation Berenblum, I., and P. Shubik. 1947. me role of Proton oil appl ications, associated with a single painting of a carcinogen, in tumour induction of the mouse ' s skin. or . J. Cancer 1: 379-382. Berenblum, I., and P. Shubik. 1947. A new, quantitative approach to the study of the stages of chemical carcinogenes is in the mouse ' s Sk in . Br . J. Cancer 1: 383-391. Berenblum, I., and P. Shubik. 1949. me persistence of latent tumour cells induced in the mouse's skin by a single application of 9:10-dimethyl-1:2-benzanthracene. Br. J. Cancer 3: 384-386. Crouch, E., and R. Wilson. 1979. Interspecies comparison of care inogen ic potency. J . Toxicol . Env iron . Health 5: 1095-1118 . National Academy of Sciences. 1976. Dr inking Water and Health. Safe Drinking Water Committee, Assembly of Life Sciences, National Research Council, Washington, D.C. 938 pp. Sontag, J.M. , N.P. Page, and U. Saffiotti. 1976. Guidelines for Carcinogen ic Bioassay in Small Rodents. NCI-CG-Ir-l. U.S. Department of Health, Education, and Welfare, Public Health Service, National Institutes of Health, National Cancer Institute, Washington, D.C. 65 pp. 100
Staffa, J.A., and M.A. Mehlman, eds. 1980. Risk Assessment (EDo~ Study ~ Inno~ra t ions in Cancer : Proceedings of a Symposium Sponsored by the National Center for Toxicological Research, U. S. Food and Drug Administration, and the American College of Toxicology. J. Environ. Patbol. Toxicol. 3 {3) :1-246. Thiede, T., and B.C. Christensen. 1975 . Tumour s of the bladder induced by chlornaphazine treatment. Ugeskr . Beg. 137: 661-666 ~ in Danish; English abstract] . 101
