Click for next page ( 18


The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 17
2 Cancer: Its Nature arid Relationship to Diet Before discussing the effects of diet and nutrition on the incidence of cancer, it is useful to review what is known about the nature of can- cer, the basis for suspecting a relationship between diet and cancer, and the stages of carcinogenesis at which diet may exert an effect. The following review is meant for a general audience. To avoid being too long, it oversimplifies several issues. For a more complete cover- age, the reader should turn to two books--Origins of Human Cancer (Hiatt et al., 1977) and Cancer: Science and Society (Cairns, 1978--and a _ _ ~ journal article entitled "The causes of cancer: Quantitative estimates of avoidable risks of cancer in the United States today" (Doll and Peto, 1981). THE NATURE OF CANCER Cancers are populations of cells in the body that have acquired the ability to multiply and spread without the normal restraints. To under- stand how such populations arise and the nature of their abnormality, it is necessary to understand how cells normally control their own behavior. This subject falls within a branch of basic biology that is still not well understood. The Control of Cell Division During the Growth and Replacement of Normal . Tissues The adult human body contains about ten trillion (1013) cells. Some of these cells (e.g., the neurons and striated muscle cells) are incapable of undergoing cell division; some (e.g., the cells of the marrow and the epithelial cells of the gut and skin) are actively dividing throughout our adult life; and others (e.g., the cells of the liver) retain the ability to divide, but multiply rapidly only when tissues are undergoing regenera- tion after having been damaged. During our entire adult life, the gross and microscopic anatomy of the body is preserved by precise systems that regulate cell division. Cancer develops from cells that escape such regulation. Although developmental biologists have for many years studied the operation of these regulatory systems, few of the signals that control cell behavior in multicellular animals have been identified. Certain tissues (e.g., the endocrine glands, the liver, and the bone marrow) serve general (systemic) functions rather than local functions; their role is to add or subs tract substances or cells from blood. The extent 2-1 17

OCR for page 17
18 DIET, NUTRITION, AND CANCER of cell multiplication in such tissues must be under general control and therefore must be regulated by various hormones circulating in the blood. Many of these hormones have been identified, and the general features of their operation, if not the precise molecular mechanism, are now well understood. By contrast, local signals that influence each cell's reaction to its immediate environment have not been identified. Yet, these are the signals that are responsible for the microscopic architecture of each tissue. As the result of numerous experiments on tissue development and regeneration, we know that local signals pass between cells, but we still do not know much about their nature. The entire network of signals serves to maintain the stability and integrity of each organ and tissue, and to protect the organism from any form of uncontrolled growth. For the organism as a whole, the forces of natural selection are inexorably at work: animals that multiply fastest win the race for survival. Within each multicellular organism, however, natural selection must be held at bay. Fortunately, the controls work very well. Although some 1016 cell divisions occur within each human being during his or her lifetime, only about one-third of the U.S. popu- lation will develop a clinically detectable cancer. Various Abnormalities of Cell Behavior Until the signalling systems that impose territoriality upon the cells of the body are better understood, it is going to be difficult to determine the way a cell must be altered to free it from these restraints and allow it to form a tumor. During the early days of cancer research, many people hoped that all cancer cells would prove to be defective in one particular, common feature (e.g., in their ability to respond to a specific restraining signal received by all tissues), but this now seems most unlikely. The different forms of uncontrolled growth that lead to the different varieties of benign and malignant (cancerous) tumors appear to have distinct causes. The adult human body contains several hundred different classes of cell that can be distinguished by their morphology, their relationship to other cells, the chemistry of their products, and their response to various hormones and to changes in their environment. Cells in each of these classes are capable of uncontrolled growth and tumor formation, but cells in some classes seem more at risk than others. For example, cancers derived from nerve cells are confined almost entirely to young children, probably because neurons lose their ability to multiply when embryogenesis is complete. In humans of all ages, cancers are common in the epithelial cells of the gut and skin, perhaps because these cells are continually undergoing division and replacement throughout life. The epithelial cells in the breast seem to be particularly susceptible to certain ill-defined carcinogenic factors during the interval between onset of menstruation and the first pregnancy. In all, several hundred forms of uncontrolled growth are now recognized and, as the result of 2 - 2

OCR for page 17
Cancer: Its Nature and Relationship to Diet 19 some 100 years of clinical study involving a vast number of patients, the usual behavior of each of these kinds of tumor is now well estab- lished. The range of cell behavior is very great. At one extreme there are such trivial abnormalities as the localized growth of melanocytes, which creates the common freckle. (Indeed, in fair-skinned people who are frequently exposed to sunlight, freckles are regarded as a normal abnormality.) At the other extreme is malignant melanoma, a cancer arising from skin melanocytes. This form of cancer is often rapidly fatal because it has a marked tendency to undergo swift dissemination through the bloodstream to all organs of the body. Between these ex- tremes, all levels of severity can be found. Many tumors vary little in their behavior from one patient to the next. For example, the com- monest tumor of the uterus (the benign fibroma or leiomyoma) arises in the smooth muscle of the uterus and can grow to enormous size, but the cells of the tumor virtually never invade the surrounding tissue or spread to distant sites. Similarly, the commonest tumor of the facial skin (the basal cell carcinoma) is locally very invasive, but it also never spreads to distant sites. Other tumors, such as those in the breast, are much less predictable; some of them spread quickly and others do not. The pathological classification of all these growth abnormalities (or neoplasias) depends on (1) the tissue of origin, (2) the type of cell - involved, and (3) most importantly, whether the abnormal cells are con- fined to their original location (in which case the tumor is classified as benign) or have invaded the surrounding tissues or "metastasized" to distant sites (in which case the tumor counts as a cancer). Cancers that arise in the epithelial cells are called carcinomas, and they account for more than 90% of all human deaths from cancer. Cancers that arise in the progenitors of the circulating cells of the blood are called leukemias (if the abnormal cells are circulating) and lymphomas or myelomas (if the cancer affects lymphocytes that tend to be localized in the lymph nodes or the bone marrow, respectively). Cancers that arise in fibrous con- nective tissue or bone are called sarcomas. Together these nonepithelial cancers account for slightly less than 10Z of deaths from cancer. The pathological classification goes further and subdivides the carcinomas according to the tissue of origin (e.g., hepatocarcinoma) or the behavior of the component cells (e.g., adenocarcinoma and squamous cell carcinoma). Cancer and Cell Heredity It seems likely that most cancers arise from the proliferation of a single altered cell. Every tissue of the body has been shown to be made up of a very large number of small hereditarily similar nests (fam- ilies) of cells. Therefore, any large piece of tissue will contain cells that are members of several of these families (i.e., they come from different branches of the family tree of cells descended from the original fertilized egg). Cancers, however, prove almost without 2 - 3

OCR for page 17
20 DIET, NUTRITION, AND CANCER exception to contain cells from only one family (i.e., they must have arisen within a single family)--a finding that is most easily explained by assuming that each cancer is descended from a single abnormal cell. However, it is also clear that even though cancers are in this sense clonal, a considerable amount of modification by variation and natural selection can occur during the growth of each cancer. For example, although a particular cancer may be marked, from the start, with a particular chromosomal abnormality, further abnormalities may be added during its subsequent growth. It is as if an early step had permitted uncontrolled growth and the operation of natural selection, which in turn allowed the progressive evolution of increasingly abnormal and more rapidly multiplying types of cells. Because the average cancerous growth will amount to many millions of cells before it becomes detectable, it may already have undergone considerable selection for the fittest vari- ants arising spontaneously within the population. So, even if two sep- arate cancers were to start off with the same underlying abnormality, they could have very different characteristics by the time the diagnosis of cancer becomes possible. This makes it very difficult to be certain whether the great diversity of phenotypic characteristics observed in most forms of cancer means that there has to be great diversity in the ways of producing cancer. The Varying Incidence of Cancer It is abundantly clear that the incidence of all the common cancers in humans is being determined by various potentially controllable exter- nal factors, because people in different parts of the world suffer from different kinds of cancer, depending on their habits, diet, and customs rather than on their ethnic origins. Thus, when people migrate from one country to another they tend to acquire the pattern of cancer that is characteristic of their new home. This is surely the most comforting fact to come out of all cancer research, for it means that cancer is, in large part, a preventable disease. Next, it is also clear that some carcinogens tend to be associated with specific cancers. For example, cigarette smoke is the major cause of the common bronchogenic carcinoma of the lung, but it does not cause the less common mesothelioma of the lung. Asbestos causes both meso- theliomas and bronchogenic carcinomas. Certain aniline dyes (especially 2-naphthylamine) cause bladder cancer, but little of any other kind of cancer. A similar specificity probably also applies to cancers whose cause or causes have not yet been identified, because the incidences of the different kinds of cancer tend to vary independently. Although the causes of most cancers that are common in affluent in- dustrialized nations have not yet been identified, epidemiological data suggest certain general conclusions about the nature of these causes. Apart from lung cancer (which has become much more prevalent during this century as more and more people have taken up smoking), the only common 2 - 4

OCR for page 17
Cancer: Its Nature and, Relationship to Diet 21 cancers to have changed much in incidence during the 20th century are stomach and uterine cancers, both of which have become much less common. So it seems probable that most cases of the cancers that are common today are not being caused by the products of modern industry. In fact, the same conclusion can be reached in quite a different way, because the incidence and age-specific death rate from many of the common nonrespi- ratory cancers has been found to be higher in certain nonindustrialized nations like New Zealand than in the United States. It seems likely, therefore, that the common cancers not attributable to smoking are re- lated, for the most part, not to industrialization but to various other long-standing features of our lifestyle, especially diet. Experimentally Induced Carcinogenesis Before considering how ingredients of the diet (or any other factor that varies from one population to another) could determine the inci- dence of cancer, it would be helpful if we had more information about the kinds of event that can turn a normal cell into a cancer cell. In principle, this information can come from experimental studies of car- cinogenesis. Much of this report is concerned with certain experimental systems for producing (or preventing) cancer in animals. These experiments show that a variety of treatments and agents affect the incidence of cancer. If only one general class of agent or treatment had proved to be carcino- genic, it would then have been clear what kinds of substance we should, if possible, be trying to eradicate from our lifestyle or environment. In fact, a bewildering array of agents and treatments have been shown to influence the incidence of cancer in animals. The most widely studied method for producing cancer is to expose an animal to repeated doses of any physical or chemical agent that damages DNA and causes mutations. In certain instances it has been possible to show that the cancers produced by these agents are indeed arising as the consequence of damage to DNA. This has led people to postulate that the substances that are the important determinants of cancer in humans will, for the most part, prove to be agents that produce mutations. The idea is attractive because the cancer cell plainly has a defect that can be inherited by all its descendants. Moreover, several quick and inexpen- sive methods have been developed for detecting mutagens in our environ- ment. However, it is clear that many, perhaps most, examples of carcin- ogenesis in laboratory animals actually proceed by way of a sequence of steps, some of which are brought about by the mutagens, whereas other, later steps are driven by agents (promoters) that are not mutagenic. One of the most fully studied examples of "multistep" carcinogenesis is the induction of skin cancer in mice. The first step, referred to as initiation, can be produced by any of a wide range of mutagens. The step appears to occur rapidly and to be essentially irreversible and is assumed 2 - 5

OCR for page 17
22 DIET, NUTRITION, AND CANCER to be a direct consequence of damage to the DNA of those cells in the skin that normally proliferate and differentiate to produce the scaly cells that protect the body surface. To complete the process, it is then necessary to expose the skin for a prolonged period to various promoting agents that modify or accelerate the normal process of cell turnover. Because initiation is essentially irreversible, the process of promotion will produce cancer, even if a long time is allowed to elapse between initiation and promotion. The nature of the events taking place during these later stages of carcinogenesis is still very obscure. For example, it is not simply a matter of forcing the initiated cell to divide quickly, because certain agents that provoke cell proliferation do not act as promoters. Most promoting agents (e.g., the irritant phorbol esters in croton oil) have a wide range of effects on the physiological functions of cells, espe- cially on reactions taking place in the cell membrane. But there is no reason to believe that they have any direct action on DNA. Unlike in- itiation, the steps driven by promoters must be to some extent rever- sible because the effects of prolonged exposure to a promoter such as cro ton oil are lost if too long an interval is allowed between each application. To complicate matters still further, there is good evi- dence that promotion can itself be divided into two classes of events, each of which can be driven and inhibited by specific agents. So far, however, the exact molecular biology of these later events in carcino- genesis remains obscure. The predominant effects of the mutagens commonly used in experiments in animals are localized changes in DNA, usually affecting only one or two base pairs. There are, however, other ways to induce cancer in ani- mals that do not involve local changes in base pair sequence. For exam- ple, approximately one-quarter of all viruses known to infect vertebrates are capable of causing cancer in some animal or other. Many of these viruses probably are carcinogenic because they lead to novel juxtaposi- tions of genes from the virus and the host. Perhaps for this reason they tend to cause particular kinds of cancer (e.g., lymphomas, leukemias, or sarcomas) rather than act as general nonspecific carcinogens. Certainly, there is little evidence that they are acting as mutagens. Certain cancers can be produced simply by transplanting cells to novel sites in the body where they can multiply without the usual re- straint or by placing them next to inert solid surfaces such as plas- tic or metal. It seems unlikely that mutation plays any part in these processes, especially since certain of the cancers produced in this way will recover their normal restrained behavior when they are returned to their normal location. One conspicuous (and neglected) group of cancers results from over- feeding--a treatment that is obviously not mutagenic. The incidence of these cancers can be reduced by reductions in food intake. These effects 2 - 6

OCR for page 17
Cancer: Its Nature and Relationship to Diet 23 of nutrition on the incidence of several kinds of spontaneous cancer in animals are very much like some of the correlations between diet and cancer that have been observed by epidemiologists--correlations that involve most of the major forms of cancer. Laboratory studies of carcinogenesis are therefore offering us a choice of possible mechanisms for the formation of the major cancers in humans. There appears to be no way of guessing which are likely to be the important mechanisms of carcinogenesis in humans until further epi- demiological studies have been conducted to isolate and identify the variables that determine cancer rates in humans. THE CAUSES OF CANCER The commonest cancer in Western nations is lung cancer. Its inci- dence is related to each nation's consumption of cigarettes--not to its level of industrialization. Thus, it is safe to be a nonsmoker in the United States, but it is not safe to be a smoker anywhere, even in the clean air of a country like Finland. A somewhat similar observation has been made for the next two most common cancers: cancer of the large intestine and breast. Both of these cancers are associated in some way with long-standing affluence, but they are apparently not linked to in- dustrialization. Thus, the incidence of both cancers in an industrialized country like Czechoslovakia is not nearly as high as it is in New Zealand, which has one of the highest rates for both cancers despite its lack of the oil and coal required for chemical and manufacturing industries and its dependence on dairy and agricultural products for income. The incidence of the other major cancers also varies greatly from one nation to the next, but not in the way most of us must have been led to believe from reading the many news reports about newly discovered car- cinogens in our environment. Apart from cigarettes, the causes of most of today's cancers do not bear any simple and direct relationship to the intended and unintended products of industry and to what might be called the more unnatural features of modern life. This is not to say that in- dustrial exposure is harmless, but simply that relatively few of the middle-aged and older members of our current population have been exposed to great occupational hazards. For those who were exposed, the hazards are real and inexcusable. In addition, we have to remember that the time course of carcinogenesis commonly extends over 20 years or more. This means that we have to be very concerned about the possible long-term effects resulting from exposure to novel hazards. Investigation of the causes of cancer has been an important branch of cancer research. Early in the course of such studies it became apparent that genetic factors were not responsible for international differences in cancer incidence. When groups of people migrate from one country to another, thereby changing their environment and way of life,

OCR for page 17
24 DIET, NUTRITION, AND CANCER they tend to leave behind the cancers typical of their homeland and ac- quire those of their new country. It is almost as if they are offering us the results of a calculated experiment; indeed, to make the experiment more perfect, several migrant groups have tended to intermarry for sev- eral generations rather than to outbreed with the other inhabitants of their new country. Although studies of migrants have some drawbacks, they have made an invaluable contribution to research on cancer in humans. For example, the fact that blacks and Japanese within the United States develop the spectrum of cancers that is generally typical for the U.S. population but different from that in Africa and Japan tells us that most cancers must have external causes and, in principle, should therefore be preventable. Certain causes of cancer have been fairly obvious for a long time. For example, it was not difficult to connect sunlight with skin cancer. Similarly, it seemed likely that the ever-increasing number of lung cancers would prove to be related to something that people were breath- ing. For most cancers, however, likely candidates for the causes were not immediately apparent. The list of forces that play upon us, and are likely to change if we move from one country to another, is not impossibly long. Such forces obviously include the air we breathe, the water we drink, and the food we eat, as well as the way we prepare the food and apportion it into meals. The list should include the infectious diseases that we contract and perhaps, in a more general sense, some measure of our contact with other living creatures. In addition to the radiation in sunlight, we are exposed to other forms of radiation that vary somewhat in intensity from one region to another. Lifestyles in various countries differ enormously. Cultures vary in family size, in the age at which reproduction starts, in the stresses and strains placed on their populations, and in many other ways. Some of these variables can be precisely measured. For example, it is possi- ble to estimate accurately the extent to which time spent indoors in cold climates increases our exposure to the background radiation that emanates from many building materials, and to speculate that this increased expo- sure might account for some of the extra incidence of cancer in the north- eastern United States. Similarly, it is easy to measure exposures to in- fectious agents. Thus, it was a straightforward exercise to demonstrate that Burkitt's lymphoma in Africa bears some relationship to a conjunction of endemic malaria and infection with Epstein-Barr virus. At the opposite end of the scale, certain other variables are virtually impossible to quantitate. For example, some people have suggested that stress can cause cancer, but this hypothesis cannot easily be tested, and there is, as yet, no reason to think that it is true. For most of the variables, however, measurements can be made, but they are not straightforward. Thus, it will require some care to trace 2 - 8

OCR for page 17
Cancer: Its Nature and, Relationship to Diet 25 all the threads of causality to their source. The cancers are fairly easy to record, in terms of either incidence or mortality, but the environmental variables are not. This is especially true for diet. During the past 200 years, there have been major changes in the nutri- tional content of our diet and in our exposure and response to infec- tious diseases. The effects can be seen most readily in migrants. For example, the children of migrants to the United States are, on average, taller and live longer than their parents, just as their parents were taller than their ancestors. Because nutrition has strong effects on growth, physiology, and longevity, it was natural to suspect that the effects could extend to cover susceptibility to cancer. This hypothe- sis has been examined in both epidemiological studies and laboratory experiments, but the investigation of the association between diet and cancer is far from complete. Epidemiologists found it relatively easy to demonstrate a correla- tion between the diets consumed by modern affluent societies and the incidence of cancers in such organs as the breast, colon, and uterus. But it is much more difficult to determine exactly which, if any, of the dietary components are responsible. For example, certain interested parties formerly argued that the association between lung cancer and smoking was not causal; instead, they suggested that the kind of people who smoke are the kind of people who, for some quite independent reason, develop lung cancer. This argument had to be resolved by prospective studies of groups of people who had stopped smoking. Exactly the same questions now arise about components of our diet: are the associations causal or coincidental? Unfortunately, it is much harder to find out what someone is eating than whether or not they smoke. It is important therefore that we prepare ourselves for a period of uncertainty, be- tween our present realization that diet affects cancer and our eventual ability to offer the public a precise formula for minimizing the inci- dence of cancer. Although the formula is still not known, we do have some estimate of the benefits it would bestow. Judging from the observed differ- ences in cancer rates among populations with different diets, it is highly likely that the United States will eventually have the option of adopting a diet that reduces its incidence of cancer by approximately one-third, and it is absolutely certain that another third could be prevented by abolishing smoking. Those reductions would be roughly equivalent to the reduction in mortality from the infectious diseases brought about by improved hygiene and better health care delivery during the 19th century. This prediction can be made with confidence because some major cancers have already been controlled. For example, the mortality due to stomach cancer in the United States has dropped sharply during the past 50 years--from first to sixth place on the list of most common cancers--a change brought about presumably by some alterations in our diet that took place during that period. 2 - 9

OCR for page 17
26 DIET, NUTRITION, AND CANCER What cannot be predicted is the exact way in which we will discover the precise changes that ought to be made in the nation's diet. As this report points out several times (especially in Chapter 3), it is not easy to determine precisely what people are eating now, and it is even more difficult to learn what they were eating many years ago when the seeds were presumably being sown for the cancers they now have. As shown in later chapters, it has been possible to develop in lab- oratory animals reasonable facsimiles of the common cancers in humans. By studying these cancers, it may be possible for the experimentalist to uncover certain important variables that the epidemiologist would only discover with difficulty. THE INFLUENCE OF DIET ON EXPERIMENTALLY INDUCED CANCERS Most laboratory and domesticated animals have a significant incidence of cancer during old age. These cancers tend to be affected by changes in diet in the same way as cancer in their human counterparts. Thus, a reduction in total intake of food or specific food items tends to lower the incidence of both spontaneous and chemically induced cancer in most strains of rats and mice. The main exceptions to this rule occur when some dietary restriction leads to a deficiency disease involving some particular tissue, thereby raising the incidence of cancer in that tissue. For example, deficiency of methyl donors such as choline leads to liver damage and raises the incidence of spontaneous liver cancer in rats. These examples of the influence of diet on experimentally induced cancers are not easily investigated because the underlying mechanisms and molecular biology of the cancers are not understood. Indeed, the effects of diet were often treated as if they were simply a nuisance, being yet another variable standing between the investigators and their assays for carcinogenicity. The Early Stages of Carcinogenesis As mentioned earlier, the most generally accepted concept of carcin- ogenesis is that it is a prolonged process that starts when an animal or human being is exposed to some mutagen (initiator) that can interact with the DNA. Because chemical initiators have to be reactive to interact in this way, they are usually unstable and cannot persist very long in the environment. Thus, a more usual carcinogenic sequence is exposure to a stable but toxic chemical (e.g., aflatoxin B1) that has to be detoxified in an organ such as the liver and, in the process, is turned into a highly reactive derivative that interacts with the DNA of the liver cell. In short, most carcinogenic initiators are created within the body as the result of metabolic activation. This opens the way for a number of very complicated effects during carcinogenesis. 2-10

OCR for page 17
Cancer: Its Nature and Relationship to Diet 27 For example, a chemical that is not itself a carcinogen can act as a cocarcinogen or an anticarcinogen by stimulating or inhibiting one of the enzymes involved in the metabolism of some carcinogen. An item in the diet could therefore determine the incidence of cancer not only because of the carcinogens that it contains but also because of its various cocarcinogens and anticarcinogens, which can modify the process of carcinogenesis. Metabolic activation has one other important consequence. In some cases, a carcinogen can be partly metabolized in one tissue and then enter the blood and undergo its final activation in some other, distant tissue. Therefore, it is not uncommon to observe that a carcinogen fed to an animal can produce cancer in organs such as breast, brain, lung, or uterus, which are far from the gastrointestinal tract. One other variable is also important in determining the course of initiation. Most cells possess effective methods for repairing DNA. They are therefore able to undo most of the damage caused by initiating carcinogens, if there is sufficient time before they have to duplicate their DNA. It follows that initiators are sometimes made more effective if administered at the same time as some agent that forces rapid cell multiplication. For example, the production of liver cancers in choline- deficient rats, mentioned earlier, proved on further investigation to result from the action of two separate stimuli--an unexpected (and unin- tended) initiating carcinogen in the diet and the intended deficiency of choline, which was acting as a cocarcinogen by destroying the liver and therefore forcing the remaining liver cells to continue regenerating. Thus, these cancers could have been prevented either by adding choline to the diet or by removing the carcinogen from the diet. To summarize, the effect of diet on the initiation of cancer can be quite complex. The early stages of carcinogenesis can involve the simultaneous interaction of several independent variables operating in a variety of ways. But at least this means that the early steps in the formation of many kinds of cancer may be interceptable in any one of several ways. The Late Stages of Carcinogenesis The late stages of carcinogenesis tend to be even more obscure, because they involve reactions that are even less well understood than the biochemistry of metabolic activation, DNA damage, and DNA repair. For most chemically induced (as opposed to spontaneous) cancers in laboratory animals, the process of carcinogenesis seems to go through a succession of stages. The early initiatory steps require exposure to substances that usually are known to be mutagenic. The later stages are brought about by agents (promoters) that affect cell differentiation and provoke cell proliferation. These agents appear to act primarily on pro- cesses occurring in cell membranes, including the responses to certain 2-11

OCR for page 17
28 DIET, NUTRITION, AND CANCER signalling substances and free radicals. But the molecular biology of their action remains obscure despite extensive studies on the subject. This must surely be the outstanding lacuna in experimental cancer re- search. Many investigators believe that these later stages concern the gradual expression of all the mutations produced during the early stages, but several observations do not fit in well with this hypothesis. Whenever dietary experiments discriminate between the early and late stages of carcinogenesis, they usually show that the late stages are most affected by changes in the diet. The mechanisms underlying such effects are not known, but it is clear that normal dietary components can either raise or lower the incidence of cancers that have been initiated by expos- ing animals to carcinogens in the diet or by other routes. The details of many such experiments are described in the body of this report. One interesting feature of these experiments is that their results so closely mimic the human condition. Most laboratory animals fed ad libitum are grossly obese compared to their wild counterparts. If they are placed on diets to maintain their weight within the range that would be found in the wild, their cancer rate tends to drop to very low levels-- unless, of course, they are simultaneously exposed to high levels of some carcinogen. Similarly, obesity has also been associated with higher rates of cancer at some sites in humans. In principle, we should be at least as interested in the late stages of carcinogenesis as in the early stages. Although cancer could in prin- ciple be prevented by blocking events at any stage, only the young would receive much benefit if we removed the initiators from our environment, whereas everyone--old and young alike--could be benefited by blocking the late stages. Unfortunately, because the late events of carcinogenesis are so poorly understood, much more effort has been expended on the study of initiators. Furthermore, searches for environmental hazards and most cost-benefit analyses have centered on eradicating the initiators to which we are exposed, rather than seeking out the promoters. 2-12

OCR for page 17
Cancer: Its Nature and, Relationship to Diet 29 REFERENCES Cairns, J. 1978. Cancer: Company, San Francisco, S cience and Society. W . H. Freeman and Calif . 19 9 pp . Doll, R., and R. Peto. 1981. The causes of cancer: Quantitative esti- mates of avoidable risks of cancer in the United States today. J. Natl. Cancer Inst. 66 :1191-1308. Hiatt, H. H., J. D. Watson, and J. A. Winsten, eds. 1977. Origins of Human Cancer. Cold Spring Harbor laboratory, Cold Spring Harbor, N.Y. 3 volumes, 1,889 pp. 2-13