NCBI Bookshelf. A service of the National Library of Medicine, National Institutes of Health.

Alberts B, Johnson A, Lewis J, et al. Molecular Biology of the Cell. 4th edition. New York: Garland Science; 2002.

  • By agreement with the publisher, this book is accessible by the search feature, but cannot be browsed.
Cover of Molecular Biology of the Cell

Molecular Biology of the Cell. 4th edition.

Show details

The Preventable Causes of Cancer

The development of a cancer generally requires many steps, each governed by multiple factors—some dependent on the genetic constitution of the individual, others dependent on his or her environment and way of life. A certain irreducible background incidence of cancer is to be expected regardless of circumstances: mutations can never be absolutely avoided, because they are an inescapable consequence of fundamental limitations on the accuracy of DNA replication, as discussed in Chapter 5. If a human could live long enough, it is inevitable that at least one of his or her cells would eventually accumulate a set of mutations sufficient for cancer to develop.

Nevertheless, there is evidence that avoidable environmental factors play some part in the causation of most cases of the disease. This is demonstrated most clearly by a comparison of cancer incidence in different countries: for almost every cancer that is common in one country, there is another country where the incidence is several times lower (Table 23-1). Because migrant populations tend to adopt the pattern of cancer incidence typical of the host country, the differences appear to be due mostly to environmental, not genetic, factors. From such data it is estimated that 80–90% of cancers should be avoidable, or at least postponable. Unfortunately, different cancers have different environmental risk factors, and a population that happens to escape one such danger is usually exposed to another. This is not, however, inevitable. There are some subgroups whose way of life substantially reduces the total cancer death rate among individuals of a given age. Under the current conditions in the United States and Europe, approximately one in five people die of cancer. But the incidence of cancer among strict Mormons in Utah, for example, is only about half that among Americans in general. Cancer incidence is also low in certain relatively affluent populations in Africa.

Table 23-1. Variation Between Countries in the Incidence of Some Common Cancers.

Table 23-1

Variation Between Countries in the Incidence of Some Common Cancers.

Although such observations on human populations indicate that cancer can often be avoided, it has been difficult in most cases to identify the specific environmental risk factors or to establish how they act. We will first look at what has been learned about the external agents that are known to cause cancer. We will then consider some of the triumphs, and difficulties, in finding ways in which human cancer can be prevented. The problem of treatment will be discussed in the last section, after we have examined the molecular biology of the disease.

Many, But Not All, Cancer-Causing Agents Damage DNA

The agents that can cause cancer are many and varied, but the easiest to understand are those that cause damage to DNA, and so generate mutations. These cancer-causing mutagens include chemical carcinogens, viruses, and various forms of radiation—UV light and ionizing radiation such as gamma rays and alpha particles from radioactive decay.

Many quite disparate chemicals have been shown to be carcinogenic when they are fed to experimental animals or painted repeatedly on their skin. Examples include a range of aromatic hydrocarbons and derivatives of them such as aromatic amines; nitrosamines; and alkylating agents such as mustard gas. Although these chemical carcinogens are diverse in structure, they have at least one property in common—they cause mutations. In one popular test for mutagenicity, the carcinogen is mixed with an activating extract prepared from rat liver cells (to mimic the biochemical processing that occurs in an intact animal) and is added to a culture of specially designed test bacteria; the resulting mutation rate of the bacteria is then measured (Figure 23-17). Most of the compounds scored as mutagenic by this rapid and convenient assay in bacteria also cause mutations or chromosome aberrations when tested on mammalian cells. When mutagenicity data from various sources are analyzed, one finds that the majority of identified carcinogens are mutagens.

Figure 23-17. The Ames test for mutagenicity.

Figure 23-17

The Ames test for mutagenicity. The test uses a strain of Salmonella bacteria that require histidine in the medium because of a defect in a gene necessary for histidine synthesis. Mutagens can cause a further change in this gene that reverses the defect, (more...)

A few of these carcinogens act directly on the target DNA, but generally the more potent ones are relatively inert chemically and become damaging only after they have been changed to a more reactive form by metabolic processes—notably by a set of intracellular enzymes known as the cytochrome P-450 oxidases. These enzymes normally help to convert ingested toxins into harmless and easily excreted compounds. Unfortunately, their activity on certain chemicals generates products that are highly mutagenic. Examples of carcinogens activated in this way include the fungal toxin aflatoxin B1 (Figure 23-18) and benzo[a]- pyrene, a cancer-causing chemical present in coal tar and tobacco smoke.

Figure 23-18. Metabolic activation of a carcinogen.

Figure 23-18

Metabolic activation of a carcinogen. Many chemical carcinogens have to be activated by a metabolic transformation before they will cause mutations by reacting with DNA. The compound illustrated here is aflatoxin B1, a toxin from a mold (Aspergillus flavus (more...)

The Development of a Cancer Can Be Promoted by Factors That Do Not Alter the Cell's DNA Sequence

Not all substances that favor the development of cancer are mutagens, however. Some of the clearest evidence comes from studies done long ago on the effects of cancer-causing chemicals on mouse skin, where it is easy to observe the stages of tumor progression. Skin cancers can be elicited in mice by repeatedly painting the skin with a mutagenic chemical carcinogen such as benzo[a]pyrene or the related compound dimethylbenz[a]anthracene (DMBA). A single application of the carcinogen, however, usually does not by itself give rise to a tumor or any other obvious lasting abnormality. Yet it does cause latent genetic damage—mutations that set the stage for a greatly increased incidence of cancer when the cells are exposed either to further treatments with the same substance or to certain other, quite different, insults. A carcinogen that sows the seeds of cancer in this way is said to act as a tumor initiator.

Simply wounding skin that has been exposed once to such an initiator can cause cancers to develop from some of the cells at the edge of the wound. Alternatively, repeated exposure over a period of months to certain substances known as tumor promoters, which are not themselves mutagenic, can cause cancer selectively in skin previously exposed to a tumor initiator. The most widely studied tumor promoters are phorbol esters, such as tetradecanoylphorbol acetate (TPA), which behave as artificial activators of protein kinase C and hence activate part of the phosphatidylinositol intracellular signaling pathway (discussed in Chapter 15). These substances cause cancers at high frequency only if they are applied after a treatment with a mutagenic initiator (Figure 23-19).

Figure 23-19. Some possible schedules of exposure to a tumor initiator (mutagenic) and a tumor promoter (nonmutagenic) and their outcomes.

Figure 23-19

Some possible schedules of exposure to a tumor initiator (mutagenic) and a tumor promoter (nonmutagenic) and their outcomes. Cancer ensues only if the exposure to the promoter follows exposure to the initiator and only if the intensity of exposure to (more...)

As one might expect for genetic damage, the hidden changes caused by a tumor initiator are irreversible: thus, they can be uncovered by treatment with a tumor promoter even after a long delay. The immediate effect of the promoter is apparently to stimulate cell division (or to cause cells that would normally undergo terminal differentiation to continue dividing instead). In the region that had previously been exposed to the initiator, this proliferation results in the growth of many small, benign, wartlike tumors called papillomas. The greater the prior dose of initiator, the larger the number of papillomas induced; it is thought that each papilloma (at least for low doses of the initiator) consists of a single clone of cells descended from a mutant cell that the initiator has produced. How tumor promoters work is not certain, and different tumor promoters are likely to work in different ways. One possibility is that they simply induce expression of growth-controlling genes that had been mutated before the promoter was applied: a mutation that makes a gene product hyperactive will not show its effects until the gene is expressed. Another possibility is that the promoter temporarily releases the cell from an inhibitory influence that normally overrides the proliferation-inducing effect of the mutation; as a result, the cell is enabled to divide and grow into a large cluster of cells (Figure 23-20).

Figure 23-20. The effect of a tumor promoter.

Figure 23-20

The effect of a tumor promoter. The tumor promoter expands the population of mutant cells, thereby increasing the probability of tumor progression by further genetic change.

A typical papilloma might contain about 105 cells. If exposure to the tumor promoter is stopped, almost all the papillomas regress, and the skin regains a largely normal appearance. In a few of the papillomas, however, further changes occur that enable growth to continue in an uncontrolled way, even after the promoter has been withdrawn. These changes seem to originate in an occasional single papilloma cell, at about the frequency expected for spontaneous mutations. In this way, a small proportion of the papillomas progress to become cancers. Thus, the tumor promoter apparently favors the development of cancer by expanding the population of cells that carry an initial mutation: the more such cells there are and the more times they divide, the greater the chance that at least one of them will undergo another mutation carrying it one step further toward malignancy.

Although naturally occurring cancers do not necessarily arise through the specific sequence of distinct initiation and promotion steps just described, their evolution must be governed by similar principles. They too will evolve at a rate that depends both on the frequency of mutations and on the influences affecting the survival, proliferation, and spread of selected mutant cells.

Viruses and Other Infections Contribute to a Significant Proportion of Human Cancers

As far as we know, viruses and other infections play no part in the majority of human cancers. However, a small but significant proportion of human cancers, perhaps 15% in the world as a whole, are thought to arise by mechanisms that do involve viruses, bacteria or parasites. The main culprits, as shown in Table 23-2, are the DNA viruses. Evidence for their involvement comes partly from the detection of viruses in cancer patients and partly from epidemiology. Liver cancer, for example, is common in parts of the world (Africa and Southeast Asia) where hepatitis-B viral infections are common, and in those regions the cancer occurs almost exclusively in people who show signs of chronic hepatitis-B infection.

Table 23-2. Viruses Associated with Human Cancers.

Table 23-2

Viruses Associated with Human Cancers.

The precise role of a cancer-associated virus is often hard to decipher because there is a delay of many years from the initial viral infection to the development of the cancer. Moreover, the virus is responsible for only one of a series of steps in the progression to cancer, and other environmental factors and genetic accidents are also involved. Like the cancer-causing chemicals we discussed earlier, viruses can either alter a cell's DNA directly or act as tumor promoters. As we shall explain later, DNA viruses frequently carry genes that can subvert the control of cell division in the host cell, causing uncontrolled proliferation. DNA viruses that operate in this manner include the human papillomaviruses (Figure 23-21); some of these viruses cause warts, while others infect the uterine cervix and are implicated in the development of carcinomas of the cervix.

Figure 23-21. A human papillomavirus.

Figure 23-21

A human papillomavirus. (Courtesy of Norman Olson.)

In some other cancers, viruses seem to have additional, indirect tumor-promoting actions; the hepatitis-B virus may, for example, favor the development of liver cancer by doing damage that provokes cell division in the liver, as well as by altering cell growth control directly. In AIDS, the human immunodeficiency virus (HIV) promotes development of an otherwise rare cancer called Kaposi's sarcoma by destroying the immune system, thereby permitting a secondary infection with a human herpes virus (HHV-8) that has a direct carcinogenic action. Chronic infection with parasites and bacteria may also promote the development of some cancers. For example, infection of the stomach with the bacterium Helicobacter pylori, which causes ulcers, appears to be a major cause of stomach cancer. And bladder cancer in some parts of the world is associated with infection by the blood fluke, Schistosoma haematobium, a parasitic flatworm.

Identification of Carcinogens Reveals Ways to Avoid Cancer

Tobacco smoke is by far the most important environmental cause of cancer in the world today. Other comparably important chemical causes of cancer in humans remain to be identified. It is sometimes thought that the main environmental causes of cancer are the products of a highly industrialized way of life—the rise in pollution, the enhanced use of food additives, and so on—but there is little evidence to support this view. The idea may have come in part from the identification of some highly carcinogenic materials used in industry, such as 2-naphthylamine and asbestos. In fact, except for the increase in cancers caused by smoking, and a remarkable decrease in stomach cancer, the incidence of the most common cancers for individuals of a given age has not changed very much during the course of the twentieth century (Figure 23-22).

Figure 23-22. Age-adjusted cancer death rates, United States, 1930-1996.

Figure 23-22

Age-adjusted cancer death rates, United States, 1930-1996. Selected death rates adjusted to the age distribution of the US population in 1970, are plotted for (A) females and (B) males. Note the dramatic rise in lung cancer for both sexes, following the (more...)

Most of the carcinogenic factors that are known to be significant are by no means peculiar to the modern world. The most potent known carcinogen (by certain assays), and an important cause of liver cancer in Africa and Asia, is aflatoxin B1 (see Figure 23-18), a compound produced by fungi that naturally contaminate foods such as tropical peanuts. And for women, the risk of cancer is powerfully influenced by the reproductive hormones that circulate in the body at different stages of life. Thus, a striking correlation exists between reproductive history and the occurrence of breast cancer (Figure 23-23). The reproductive hormones presumably affect breast-cancer incidence through their influence on cell proliferation in the breast. Clearly, when attempting to identify the environmental causes of cancer, we need to keep an open mind.

Figure 23-23. Effects of childbearing on the risk of breast cancer.

Figure 23-23

Effects of childbearing on the risk of breast cancer. The relative probability of breast cancer developing at some time in a woman's life is plotted as a function of the age at which she gives birth to her first child. The graph shows the value of the (more...)

Epidemiology—the analysis of disease frequency in populations—remains the principal tool for finding environmental causes of human cancer. The approach has enjoyed some notable successes and it promises more to come. Simply by revealing the role of smoking, epidemiology has shown a way to reduce the total cancer death rate in North America and Europe by as much as 30%. The approach works best when applied to a fairly uniform population in which it is easy to distinguish between individuals who were exposed to the agent and those who were not, and when the agent under investigation is responsible for most of the cases of a certain kind of cancer. For example, in the early part of this century, in one British factory, all of the men who had been employed in distilling 2-napthylamine (and were thereby subjected to prolonged exposure) eventually developed bladder cancer (see Figure 23-8); the connection was relatively easy to establish because both the chemical and the form of cancer were uncommon in the general population.

In contrast, it is very hard to identify, by epidemiology alone, everyday environmental factors that favor development of common cancers: most of these factors are probably agents to which we are all exposed to some extent, and many of them probably contribute together to a given cancer's incidence. If, say, eating oranges doubled the risk of colorectal cancer, it is unlikely that we would find it out, unless we had some prior reason to suspect a connection; and, the same goes for the countless other substances that we eat, drink, breathe, and put on our bodies. Even when evidence is obtained that a substance may be carcinogenic, either from epidemiology or from laboratory tests, it may be difficult to decide what level of human exposure is acceptable. Estimating how many cases of human cancer a certain amount of a substance is likely to cause is difficult; balancing this risk against the utility of the substance is more difficult still. For example, certain agricultural fungicides appear to be mildly carcinogenic at high doses in animal tests, but it has been calculated that if they were not used in agriculture the contamination of food by fungal metabolites such as aflatoxin B1 would cause far more cases of cancer than the fungicide residues in food ever could.

Nevertheless, efforts to identify potential carcinogens still have a central place in our struggle with cancer. Prevention of the disease is not only better than a cure; for many types of cancer, it seems also, in our present state of knowledge, to be much more readily attainable.


The rate of tumor evolution and progression is accelerated both by mutagenic agents (tumor initiators) and by nonmutagenic agents (tumor promoters) that affect gene expression, stimulate cell proliferation, and alter the ecological balance of mutant and nonmutant cells. The majority of known cancer-causing agents are mutagens, including chemical carcinogens, certain viruses, and various forms of radiation—such as UV light and ionizing radiation. Because many factors contribute to the development of a given cancer—some of which are under our control—a large proportion of cancers are in principle preventable.

We remain largely ignorant of the principal environmental factors affecting cancer incidence. Of the environmental risk factors that have been identified, however, many can be avoided. These include smoking tobacco and falling prey to infection with cancer-causing viruses such as papillomaviruses or hepatitis B. Epidemiology can be a powerful tool for identifying such causes of human cancer and revealing ways to prevent the disease. The approach does not require knowing how cancer-causing agents work, and it can uncover factors that are not simply mutagens, such as viruses and certain patterns of child-bearing.

Image ch23f8

By agreement with the publisher, this book is accessible by the search feature, but cannot be browsed.

Copyright © 2002, Bruce Alberts, Alexander Johnson, Julian Lewis, Martin Raff, Keith Roberts, and Peter Walter; Copyright © 1983, 1989, 1994, Bruce Alberts, Dennis Bray, Julian Lewis, Martin Raff, Keith Roberts, and James D. Watson .
Bookshelf ID: NBK26897


  • Cite this Page
  • Disable Glossary Links

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...