Inhibition of Carcinogenesis by Hormones
Although a number of hormones may act as endogenous carcinogens, presumably through their promoting action (Pitot, 1993), in several examples treatment with exogenous hormones actually inhibited the development of neoplasia. Chedid et al. (1980) demonstrated that the administra- tion of corticosteroids as well as adrenocorticotrophic hormone (ACTH) inhibited the develop- ment of hepatocellular carcinomas after administration of aflatoxin B1. Similarly, administration
of the neural hormone neurotensin inhibited the induction of pancreatic carcinomas by azaserine
in rats (Tatsuta et al., 1991). This latter finding was somewhat surprising, since chronic adminis- tration of this hormone had previously been shown to induce growth of the pancreas as well as to enhance exocrine pancreatic secretion. The mechanism of this effect was not clear; but in an- other observation by Russo et al. (1990), a biological mechanism for the effect of an exogenous hormone was proposed. These investigators had observed that mammary carcinogenesis is in- hibited in rats that had completed a pregnancy prior to exposure to the carcinogenic polycyclic hydrocarbon DMBA. As a mechanism, they proposed that the placental hormone, chorionic go- nadotropin (hCG), was the mediator of these effects. They demonstrated that administration of hCG simulated mammary gland differentiation, depressing the labeling indices of the terminal structures in the glands of hCG-treated animals. This induction of terminal differentiation thus prevented the “fixation” of DMBA initiation (Chapter 7) and inhibited carcinogenesis by this agent. The potential for the use of hCG in the human as a chemopreventative agent for mam- mary cancer has never been adequately studied or exploited.
Endogenous Hormones as Modifiers of Viral Carcinogenesis
Since genetic factors of the host are important in the expression of oncogenic viruses in vivo, one might suspect that hormones and growth factors, as products of specific genes in the host, might affect the expression of viruses within the host. A prime example of such an effect is the induction of transcription of the mouse mammary tumor virus (MMTV) genome by glucocorti-
coids (cf. Beato, 1991). Treatment of cells with one or more integrated copies of the MMTV genome with glucocorticoids leads to an accumulation of MMTV particles (cf. Beato, 1991). The mechanism of this hormonal stimulation involves the interaction of glucocorticoid and progestin receptors with a specific hormone-responsive element (HRE) present in a proximal region of the U3 portion of the long terminal repeat (LTR) (see Chapter 4). This HRE contains four overlapping recognition sites to which the activated receptors bind (cf. Günzburg and Salmons, 1992). In many cells the levels of glucocorticoid and progestin receptors appear to be the limiting factor in MMTV transcription (cf. Beato, 1991). However, other hormones and growth factors, including prolactin and transforming growth factor β, also appear to modulate expression of MMTV (cf. Beato, 1991; Munoz and Bolander, 1989). In these latter instances, the exact mechanism by which these factors enhance or inhibit MMTV transcription is not clear at present. Other oncogenic viruses, including the SV40 DNA papovavirus (Chapter 4), contain hormone-responsive elements in their genomes that may, in turn, alter the expression of the virus in the host (e.g., Zuo et al., 1997). From these few examples, then, one may appreciate the dra- matic effects of hormones as endogenous modifiers in virtually all types of carcinogenesis.
Aging as an Endogenous Modifier of Carcinogenesis
Cancer is a disease that increases with age both in the human and in animal populations that have been studied thus far. Persons 65 years of age and older bear the greatest burden of cancer, with 55% of all neoplasms occurring in this age group (Yancik and Ries, 1991). In the human, however, this increase is not continuous from birth in that, from shortly after birth until the age of puberty, in the United States and many other Western countries, cancer is the second most common cause of death from disease (Chapter 1). The high incidence of cancer at an early stage appears to reflect genetic as well as intrauterine and neonatal environmental influences. How- ever, the number of individuals with cancer at this early age is an extremely small fraction of the total, as can be seen in Figure 8.2. The curve in Figure 8.2 is actually a composite of a number of curves, each for a specific histogenetic type of neoplasm. Examples of the general types of curves are shown in Figure 8.3A and B. In this instance, a number of the major cancers seen in older age groups—i.e., prostate, rectum, pancreas, etc.—increase exponentially after midlife (Figure 8.3A). However, in several neoplasms more commonly seen in the younger age groups, the curves are quite different, as noted in Figure 8.3B. Neoplasms of the bone as well as acute lymphoblastic leukemia show a biphasic curve, with a significant increase in the younger age groups, a low plateau during middle age, and then a continuing increase in the older age groups above age 55. Cancer of the testis may be considered a cancer of young adults, with its peak incidence at about 30 years of age, decreasing then almost exponentially for most of the remain- der of the life span, with a small increase between ages 70 and 80. A somewhat different pattern of breast cancer incidence in the human is noted, with a relatively rapid increase between ages 20 and 40 and a flattening out of the curve after age 40, most dramatically seen in the 40- to 50-year-old range, with increases as age increases after that (Kessler, 1992). Interestingly, there is a gradual decline in breast cancer incidence after age 55 in Japanese women (Kodama et al., 1992).
Spontaneous carcinogenesis in rodents is also age-related, with relatively few neoplasms appearing within the first 12 months in most but not all strains (Tamano et al., 1988; Yamate et al., 1990; Chandra and Frith, 1992a,b). Interestingly, in a comparison of cancer mortality in the Beagle dog and humans, Albert et al. (1994) demonstrated that the age dependence and total cancer mortality was the same in both sexes in both species, although the total cancer mortality was somewhat greater in female Beagles owing primarily to an increase in the incidence of
Figure 8.2 Cancer incidence rates by 5-year age groups including all races and both sexes. The data for this figure are taken from the National Cancer Institute SEER program from 1983 to 1987. (Adapted from Yancik and Ries, 1991, with permission of authors and publisher.)
breast cancer. Thus, it is likely that, in general, there is little if any difference in the effect of aging on spontaneous carcinogenesis in humans and domesticated and experimental animals.
The reason for the dramatic increase in both the incidence and death rate from cancer as a function of increasing age in these mammalian species is not entirely clear. The two major theo- ries to explain this phenomenon are the following: (1) the aging process itself is responsible for this high incidence of neoplasia; (2) the increasing cancer incidence is due to a continuous effect of environmental influences as organisms age; or (3) both (1) and (2). It is unlikely that specific genetic mechanisms, such as those that may play a role in prepubertal cancer incidence, are in- volved. Peto et al. (1975) attempted to distinguish experimentally between environmental and in- trinsic aging as mechanisms for the increased incidence of neoplasms seen with advancing age, using skin tumor induction in mice by multiple applications of benzo[a]pyrene as a model system.
These authors predicted results from such an experiment on the basis of the two theories stated above. These predictions are seen in Figure 8.4. In hypothesis I, the incidence rate of neo- plasia depends wholly on age and not at all on the duration of the treatment (exposure). In hy- pothesis II, the incidence rate depends wholly on the duration of the treatment (exposure) but not at all on age. These are the two extreme predictions, but the data obtained experimentally (Peto et al., 1975) agree very closely with hypothesis II. These authors concluded that skin carcino- genesis by this agent in this model system was independent of any intrinsic effects of aging, such as failing immunosurveillance (Chapter 19) or age-related hormonal alterations.
When a protocol of epidermal carcinogenesis, which distinguishes the stages of initiation and promotion, is employed in the mouse, an intrinsic effect of the aging animal, especially on the latter process, may be discerned. Van Duuren et al. (1975) were the first to show a general decrease in tumor production when promotion by TPA was carried out at later ages. Later stud- ies by Stenbäck et al. (1981) confirmed this finding and indicated a significant decrease in the
Figure 8.4 Life tables depicting percentages of mice without tumors against age and duration of expo- sure to benzo[a]pyrene under hypotheses I and II. Details of the hypotheses are given in the text. (From Peto et al., 1975, with permission of the authors and publisher.)
promotional efficacy of TPA in aging mice, as evidenced by epidermal carcinogenesis initiated with DMBA on the skin. On the other hand, Loehrke et al. (1983) saw no such effect when initiation was carried out systemically by intragastric administration of DMBA. Other data indi- cate that phenobarbital may be somewhat more effective in promoting the growth and appear- ance of hepatocellular foci (Chapter 7) in aged as compared with young rats (Ward, 1983).