The effectiveness of a chemical or physical carcinogen in inducing neoplasia is not only depend- ent on its structural and energetic properties, but also on the administered dose and the potency of the agent itself. The latter characteristic for chemicals will be considered in a later chapter (Chapter 13) and to some extent has already been spoken to for radiation carcinogenesis in rela- tion to LET and RBE (see above). Both practical and theoretical considerations of quantitative aspects of chemical and radiation carcinogenesis must be taken into account in considering the effect of a specific dose of such agents in producing neoplasia. Theoretical dose-response curves are given in Figure 3.27, in which the dose in arbitrary units produces an effect, in this case cancer. The numbers on each of the curves indicate the number of “hits” that were required to produce that specific effect. Thus, if only a single hit is required for the production of cancer by radiation or by the ultimate form of a chemical carcinogen, an exponential curve beginning at the origin would be produced, and no “threshold” (dose below which there is no effect) would be seen. On the other hand, when a larger number of hits is required to produce an effect, a distinct threshold is noted. Biological data from both humans and experimental animals have demon- strated both types of responses, although the “no threshold” response is generally assumed for radiation and chemical carcinogenesis. Undoubtedly, a number of factors are involved in these differences, including hormonal relations within the organism, DNA repair capabilities of the
Figure 3.27 A family of dose-response curves relating tumor incidence to the dose of radiation in arbi- trary units. For the sake of convenience, the actual dose is related to that dose producing 50% of the maxi- mal effect or maximal tumor response. The numbers n = 1, 2, 10, 50 indicate the number of “hits” that were required to produce that specific response.
radiated cells, position in the cell cycle at the time of carcinogen administration, and the rate of cell proliferation in the target tissue. In particular, Swenberg et al. (1987) have pointed out that increased cell proliferation is a common phenomenon during chemical carcinogenesis associ- ated with exposures to relatively high doses of carcinogenic chemicals.
Some of the more extensive and earlier studies on dose-response relationships of chemical carcinogens were carried out by Druckrey and associates (cf. Port et al., 1976). These studies were carried out by the chronic administration of chemicals at daily or weekly intervals by dif- ferent dosage regimens. By investigating the median tumor induction times (the time at which half of the animals had developed neoplasms), as in Figure 3.27, they found this parameter to be directly related to the daily dose of the carcinogen administered plus a constant. When animals receiving a daily dose of a chemical carcinogen, 4-dimethylaminostilbene, were compared with animals receiving twice the daily dose but only every other week, the cumulative incidences of neoplasms in both groups were still the same. The target organ of a carcinogen may also depend on the dose; e.g., when a relatively few high doses of diethylnitrosamine are given, primarily kidney cancer resulted, whereas at much lower daily doses, liver cancer was the principal neo- plasm induced. Although Druckrey’s studies did seem to suggest that threshold doses of chemi- cal carcinogens may occur, more recent investigations with extremely large numbers of rats (Peto et al., 1991) or mice (Brown and Hoel, 1983) dosed chronically with chemical carcinogens showed no evidence of threshold (no effect) doses.
In the case of radiation carcinogenesis, the relationship of dose to carcinoma incidence has been assumed to lack a threshold (National Academy of Sciences, 1990). However, both from animal studies as well as those in humans, evidence for thresholds does exist. Figure 3.28 is a schematic diagram of the induction of a specific neoplasm in mice exposed to various dose for- mats of ionizing radiation (Tannock and Hill, 1992). Neoplasms induced by intraperitoneal frac- tionated doses, given at a specific dose/time (day, week, etc.), exhibited distinct thresholds or no-effect levels (Tannock and Hill, 1992). Kohn and Fry (1984) have reported that in specific human situations, relatively small doses of radiation produce relatively greater carcinogenic ef- fects than do high doses, such as those used to treat cancer. This phenomenon may reflect a saturation point for radiation-induced carcinogenesis—i.e., a dose of radiation above which no further cancers are induced and cancer production decreases, possibly owing to cytotoxicity. Ap- parent thresholds have also been seen for leukemias resulting from exposure to atomic bomb radiation at Nagasaki, Japan (Kondo, 1990).
One of the principal reasons for the continued emphasis on the presence or absence of a threshold is the need to estimate the risk to humans of carcinogens at very low doses. In general, as seen later (Chapter 11), exposure of humans to carcinogens occurs primarily at low doses. The extrapolation of high-dose data to very low doses may underestimate (Stenbäck et al., 1981) or overestimate (Figures 3.26 and 3.27) (Kondo, 1990) risks of neoplasia. Although safety con- siderations may direct evaluations of human risk strictly from the standpoint of the lack of threshold levels of chemical and physical carcinogens, practical consideration as well as expand- ing scientific knowledge makes such a blanket interpretation of dose-response relationships more and more untenable (Chapter 13).
In the case of neoplasms induced by ionizing radiation, it has become apparent that several factors may be involved. Lieberman and Kaplan (1959) demonstrated that the induction of leu- kemia by ionizing radiation in mice may be completely prevented by the removal of the thymus gland shortly after birth. Later, Kaplan and associates (Declève et al., 1976) demonstrated that radiation-induced leukemia in these animals occurred by the “activation” of a leukemogenic vi- rus that normally occurs in the specific strain of animals used and that was activated by the ion- izing radiation. However, Loutit and Ash (1978) question the direct causative effect of the virus in radiation-induced leukemia, suggesting that these infectious agents were passengers rather
Figure 3.28 Schematic diagram of induction of a specific tumor type in mice exposed to various doses of ionizing radiation given to the whole body based on a review of a number of different in vivo results. Curve A: Tumors induced by single acute doses of low-LET ionizing radiation. Curve B: Tumors induced by single acute doses of high-LET radiation. Curve C: Tumors induced by fractionated doses (e.g., 1 Gy/day), of low-LET radiation. Curve D: Tumors induced by fractionated doses (e.g., 0.5 Gy/day), of high-LET radiation. (From Tannock and Hill, 1992, with permission of the authors and publisher.)
than causative in this condition. Later studies (cf. Janowski et al., 1990) indicated that the vi- ruses involved in radiation-induced lymphomagenesis in mice played a role in the development of neoplasia rather than in its initiation (Chapter 8). Oncogenic viruses have also been isolated from radiation-induced osteosarcomas of rodents, although their relationship to the causation of these neoplasms is unclear (cf. Janowski et al., 1990). In male mice of the CBA/H strain, a single dose of ionizing radiation can induce acute myeloid leukemia. Months before the appearance of the leukemia, bone marrow cells exhibiting a specific deletion/rearrangement of chromosome 2 (Chapter 6) can be observed. However, other factors appear to be involved in the development of overt leukemia in these animals (cf. Janowski et al., 1990).
Major sites of radiogenic neoplasia in humans as well as in rodents include the mammary gland, bone marrow, skin, lung, and GI tract and a variety of other sites in nonhuman primates (Broerse et al., 1989). Earlier studies by Cole and Nowell (1965) demonstrated the interesting phenomenon of the potentiation of carcinogenesis by high-energy radiation when a mitotic stim- ulus was simultaneously administered to the animal. One of the best examples of this is the in- duction of radiation-induced hepatomas in rodents; partial hepatectomy or the administration of sublethal doses of a hepatotoxic chemical, such as carbon tetrachloride, markedly increases the incidence of hepatomas when radiation is given at specific times in relation to the operation or the administration of the chemical (Wiley et al., 1973). In addition, unlike most examples of the chemical induction of neoplasia, acute x-irradiation of mice of certain strains leads to a dramatic
increase in the incidence of acute leukemias not only in exposed animals but also in their off- spring, persisting for several generations (cf. Nomura, 1991). This effect argues strongly for a direct action of ionizing radiation on germ cells, resulting in a genetic propensity for neoplastic development in later generations.