Molecular Biology of the Cell - Volume 3 - Alberts B., Bray D., Lewis J., Raff M., Roberts K., Watson J. 1994

From Cells to Multicellular Organisms
Cancer
Cancer as a Microevolutionary Process

An animal Organism can be viewed as a community or ecosystem where individual units are Cells that reproduce by division and are organized into conglomerates or Tissues united by joint activity. Earlier, when discussing tissue function, we acted as "ecologists": we were interested in Cell birth and death, their habitats, territorial limits, population size maintenance, and so on. The Structure/133.html">Discussion focused on just one important ecological concept—natural Selection—while Mutations and competition among somatic cells were omitted. The reason for this is that a healthy organism is a highly specific cell community in which altruistic tendencies prevail over competition for all cell types except one: any somatic cell is doomed to die without leaving descendants, yet by its very existence it ensures the preservation of Germ Cells—the only ones that have a chance to survive and perpetuate themselves in offspring. There is no paradox here, since the organism is a clone and the genotypes of somatic and germ cells are identical: by sacrificing themselves for the benefit of germ cells, somatic cells contribute to the spread of copies of their own genes.

Class="center">

Fig. 21-1. Malignant tumors typically give rise to metastases (secondary tumors), which is precisely what makes Cancer so difficult to treat. The figure shows the areas of the Bone Marrow where metastatic foci are commonly found in prostate carcinoma. (Union International Centre of Cancer, TNM Atlas: Illustrated Guide to the Classification of Malignant Tumors, 2nd ed. Berlin: Springer, 1986.)

Thus, unlike free-living cells (such as Bacteria), which are in constant competition in the struggle for survival, the cells of a multicellular organism are doomed to cooperation. In this situation, any mutation that causes a departure from altruistic behavior in individual members of such a "cooperative" threatens its very existence. Therefore, mutations, competition, and natural selection that begin to operate within the somatic cell population are signs of pathology. This is precisely the type of pathology that occurs in cancer; the latter represents a disease in which individual cells strive only for their own prosperity to the detriment of their neighbors, ultimately destroying the entire cell community and perishing along with it.

In this section, we will examine The Development of cancer as a microevolutionary process occurring within the organism's cell population over months and years, yet subject to the same factors—mutations and natural selection—that govern the long-term evolution of All living organisms.

21.1.1. Tumors differ According to the cell type from which they originate [2]

Cancer cells are characterized by two "innate" properties that are preserved in their progeny: first, they multiply without the restrictions that apply to normal cells, and second, they invade and occupy spaces in the body normally intended for other cells. It is the combination of these traits that makes cancer particularly dangerous. An individual defective cell that multiplies no faster than its normal neighbors will do little harm, no matter how unpleasant its features, but if its proliferation gets out of control, it gives rise to a tumor, or neoplasm—an inexorably growing mass of abnormal cells. If these remain a compact cluster, the tumor is considered benign, and surgical removal generally leads to a complete cure. A malignant tumor, conversely, is one whose cells are capable of invading surrounding tissues. Invasiveness typically implies The ability to destroy barriers, enter the bloodstream and Lymphatic vessels, and form secondary tumors (metastases) in various PARTS OF THE body (Fig. 21-1). The more extensive the metastases produced by cancer, the harder it is to eradicate. Cancerous tumors are classified according to the tissue and cell type from which they originate. Those of epidermal origin are called carcinomas, while those of Connective Tissue or Muscle origin are called sarcomas. Malignant neoplasms not falling into these two broad categories include Various Forms of leukemias, arising from hematopoietic cells, and tumors originating from Nervous system cells. Table 21-1 lists the incidence and mortality rates for the most common types of cancer in the U.S. Each category is subdivided into numerous groups according to the specific cell type, body localization, and tumor structure; many of these names are traditional and lack a modern rational basis. Alongside the nomenclature for malignant tumors, there are related terms to designate benign tumors: an example is an adenoma—a benign tumor of glandular Organs and mucous membranes lined with cuboidal or columnar epithelium—which corresponds to the malignant tumor known as adenocarcinoma (Fig. 21-2). Chondroma and chondrosarcoma are the benign and malignant tumors of Cartilage tissue, respectively. About 90% of malignant tumors in humans are carcinomas, which is likely due both to the high proliferative activity of Epithelial Tissues and to the fact that the latter are more frequently exposed to harmful Physical and Chemical influences that promote malignant transformation.

Table 21-1. Cancer Incidence and Mortality in the United States, 1986

¹⁾ Excludes non-melanoma skin cancer, as most cases are easily treated and many are not even registered.

The largest contribution to global statistics comes from five groups of cancers: lung, stomach, breast, colon and rectum, and cervix uteri, with a total of over 6 million new cancer cases now registered annually. Note that only about half of those diagnosed die from the disease. (Data for the U.S.: American Cancer Society, Cancer Facts and Figures, 1986.)

Fig. 21-2. An adenoma (benign tumor of glandular tissue) and an adenocarcinoma (the corresponding malignant tumor) differ sharply from one another. There are many varieties of these tumors; those affecting the mammary gland are shown schematically here.

Each cancer tumor has its own characteristic features and peculiarities reflecting its origin. For example, cells of epidermal basal cell carcinoma arising from a keratinocyte stem cell typically continue to synthesize cytokeratin Intermediate filaments, whereas melanoma cells derived from skin pigment cells frequently, though not always, form pigment granules. Generally speaking, cancers of different cellular origins are entirely distinct diseases. In the example considered, basal cell carcinoma is only locally invasive and rarely metastasizes, whereas melanoma is much more malignant and rapidly gives rise to numerous metastases (such behavior is reminiscent of the migratory tendency of normal pigment cell precursors during ontogenesis—see Section 16.6.5). Surgical removal of basal cell carcinoma is usually straightforward and leads to a complete cure, whereas a malignant melanoma that has begun to metastasize can often no longer be eradicated, and a fatal outcome is inevitable.

21.1.2. In most cases, a cancer tumor develops from a single abnormal cell [3]

In most cases, THE ORIGIN OF a cancer can be traced back to a single isolated primary tumor; this suggests that they arise through divisions from a single cell with certain hereditary alterations that allow its progeny to "outgrow" their neighbors. However, by the time it is detected, a typical tumor usually consists of at least a billion cells (Fig. 21-3), among which are many normal cells—such as fibroblasts in the supporting connective tissue that typically surrounds a cancer tumor. Proving that all cancer cells in a given tumor form a clone originating from a single abnormal cell is not easy, but data now exist that confirm the monoclonal nature of cancer. For instance, in virtually all patients with chronic myelogenous leukemia, leukocytes differ from normal ones by a specific chromosomal abnormality known as the Philadelphia chromosome, which is formed by a translocation between the long arms of Chromosomes 22 and 9 (see Fig. 21-4). It is highly unlikely that the genetic event responsible for this anomaly occurred simultaneously in multiple cells of a single individual; the hypothesis that all leukemic cells are descendants of a single mutant cell appears much more plausible. Indeed, when the DNA at the translocation site was cloned and its Primary Structure determined, it was found that the breakpoint and reunion points of the translocated fragments are identical in all leukemic cells of a given patient, yet vary between different patients (differing by several hundred or thousand Base Pairs). This is precisely the result one would expect if leukemia always stems from a unique event in a single cell.

Fig. 21-3. Growth of a typical human tumor (tumor diameter shown on a logarithmic scale). Years may pass before the tumor becomes noticeable.

Another way to demonstrate the monoclonal origin of cancer is provided by The phenomenon of X-chromosome inactivation (see Section 10.3.9). A normal female organism is a random mixture or mosaic of two classes of cells: those in which the paternal X chromosome is inactivated, and those in which the corresponding maternal chromosome is inactivated. Such inactivation occurs in each cell during early Embryogenesis, so that the progeny of a dividing somatic cell always have the same X chromosome inactivated as the parent cell. Consequently, X-chromosome inactivation—paternal or maternal—can serve as a heritable marker to trace the Lineage of cells within an organism. In the vast majority of examined tumors, both benign and malignant, the same X chromosome was inactivated in all tumor cells; this provides strong evidence that each tumor is the progeny of a single cell (Fig. 21-5).

21.1.3. Most cancers appear to begin with changes in The nucleotide sequence of cellular DNA [4]

If a defective cell gives rise to a tumor, it must pass on its abnormality to its progeny, meaning the damage must be heritable. Therefore, the first problem we encounter when trying to understand The Nature of cancer is whether this heritable defect results from a genetic change—that is, an alteration in the DNA nucleotide sequence—or an epigenetic change, where the pattern of Gene Expression alters without changing the primary Introduction/20.html">DNA Structure. Heritable epigenetic shifts reflecting "cellular memory" (see Sections 10.3 and 16.2.8) are a feature of normal development, manifested in the Stability of the differentiated state (Section 17.1.1) and in phenomena such as X-chromosome inactivation (Section 10.3.9), and there is no reason to immediately rule out the involvement of similar processes in tumorigenesis. For one rare and unusual type of cancer—teratocarcinoma (Section 16.2.6)—there is indeed evidence supporting an epigenetic origin. Nevertheless, there are strong grounds to believe that most cancers are caused specifically by genetic alterations (although epigenetic changes may also contribute to the subsequent progression of the disease). Specifically, this means that There is a hidden anomaly in the nucleotide sequence of the tumor cells' DNA, which can frequently be detected. We have already mentioned chronic myelogenous leukemia, and similar Examples will be encountered later. However, this does not imply that a genetic alteration is the initial step leading to cancer. A more accurate statement is that most carcinogenic agents cause genetic changes and, conversely, agents that cause genetic changes induce cancer. This correlation between carcinogenesis and mutagenesis is clearly observed for three classes of agents: chemical carcinogens (which typically cause local Changes in the DNA nucleotide sequence); ionizing radiation (such as X-rays), which usually leads to chromosome breaks and translocations; and Viruses, which introduce foreign DNA into The Cell. We will examine The Role of viruses in oncology later; for now, we will focus on chemical carcinogens.

Fig. 21-4. A translocation between chromosomes 9 and 22 is responsible for the development of chronic myelogenous leukemia in humans. The smaller of the two defective chromosomes is called the Philadelphia chromosome (Ph¹) (this anomaly was first described in Philadelphia).

Fig. 21-5. The mosaic pattern of X-chromosome inactivation indicates the monoclonal origin of cancer. Due to the probabilistic nature of X-chromosome inactivation during early embryogenesis, virtually all normal tissues in a female organism are a mixture of cells with different inactivated X chromosomes—paternal or maternal—with inactivation being inherited by all progeny of the cell. If tumor cells are tested for the expression of an X-linked marker gene, virtually all of them are found to have the same X chromosome inactivated. This is a strong argument in favor of the idea that all tumor cells are descendants of a single progenitor cell.

Fig. 21-6. Many chemical carcinogens can induce mutations by interacting with DNA only after they have been activated—that is, after undergoing specific metabolic transformations within the organism. The compound shown in the figure is aflatoxin B₁, a metabolic byproduct of the mold Aspergillus flavus oryzae, which proliferates on grain and peanuts stored in warm, humid conditions. This substance is considered one of the key factors increasing the incidence of liver cancer in tropical regions.

Generally speaking, no single specific case of cancer can be attributed entirely to a single cause. As we will see, cancer is typically the result of a chance convergence within a single cell of several independent events whose consequences have a cumulative effect. The frequency of these events is influenced in various ways by the cell's external environment, allowing us to view cancer as the outcome of a probabilistic process positively driven by combinations of external factors (see Section 21.1.6). Nevertheless, certain carcinogenic exposures can elevate the probability of critical events to the point where they become practically inevitable—ensuring that at least one cell in the body turns cancerous. A classic textbook example of this is 2-naphthylamine, which was used in the chemical industry early this century: at a British chemical plant, all workers involved in its distillation (and thus subjected to prolonged exposure to high doses) eventually developed Bladder cancer.

A wide variety of chemical substances prove to be carcinogenic when fed to experimental animals or repeatedly applied to their skin. Some act on target cells in their original form, but many must first be converted into a more active state—a process usually mediated by an intracellular enzyme system known as cytochrome P-450 oxidases. Normally, these Enzymes convert incoming poisons and lipid-soluble xenobiotics into harmless, easily excreted compounds. However, The oxidation of certain substances by this system generates products that act as direct carcinogens (Fig. 21-6). Although known chemical carcinogens are structurally diverse, most share at least one common property: the ability to induce mutations. Mutagenicity can be demonstrated using various Methods; one of the most widely accepted is the Ames test, in which a carcinogen is mixed with a rat liver cell extract (serving an activating role) and added to a culture of specially selected ("tester") bacteria. The mutation rate in such a bacterial culture serves as a measure of the test substance's mutagenicity (Fig. 21-7). Most compounds that show mutagenicity in this test are also capable of inducing mutations and/or Chromosomal aberrations in mammalian cells, and their chemical structure suggests they chemically react with DNA. A comparison of mutagenicity data obtained via various methods with carcinogenicity data from in vivo tumor induction studies reveals that most known carcinogens are mutagens, and conversely, most mutagens exhibit carcinogenic activity.

Fig. 21-7. The Ames test for mutagenicity. This assay uses a specific Histidine-requiring mutant strain of the bacterium Salmonella typhimurium (his^-), which cannot synthesize histidine and therefore requires it for growth. If the test substance is mutagenic, the his^- gene can revert to the wild type under the Influence of the compound. The resulting revertant bacteria are able to grow on a histidine-free medium. To increase the sensitivity of the test, a DNA Repair system mutation is introduced into The Genome of the tester strain, making these bacteria particularly susceptible to DNA-damaging agents. Most compounds that demonstrate mutagenicity in this test are carcinogenic, and conversely, carcinogens generally exhibit high mutagenicity.

There is, however, an important—though relatively small—group of carcinogens that are not mutagens. Below we will discuss (Section 21.1.7) how non-mutagenic substances can promote cancer development by influencing The behavior of already existing mutant cells. First, however, we must examine the frequency with which such mutant cells arise during normal organismal development.

21.1.4. A Single Mutation Is Not Sufficient to Cause Cancer [1, 5]

The total number of cell divisions in a human lifetime can be estimated at around 10¹⁶, whereas in a mouse—which consists of fewer cells and has a shorter lifespan—the corresponding figure is about 10¹². Even in the absence of Environmental Mutagens, mutations occur spontaneously at a rate of roughly 10^-6 mutations per gene per Cell Cycle (a value constrained by the fundamental limits of DNA Replication and Repair processes; see Section 5.3.2). Thus, over the course of a human lifetime, each individual gene may undergo approximately 10 billion different mutations; in a mouse, this number can reach one million. One would expect that among these arising mutant cells, many will have undergone alterations in genes that regulate Cell Division, and consequently, they may ignore the constraints normally imposed on cell proliferation. From this perspective, The problem of cancer is not why it arises at all, but why it arises so rarely.

Clearly, mammals must possess fail-safe mechanisms—double or even more robust—protecting us from being overrun by mutant cell clones that hold a selective advantage over our normal, healthy cells. Indeed, if a single mutation in a given gene were sufficient to turn a healthy cell cancerous, we would not be viable organisms. There is abundant evidence that the onset of cancer requires the coincidence of several independent, highly improbable events within a single cell. One line of evidence comes from epidemiological studies of cancer incidence as a function of age. For most types of cancer, incidence increases steeply with age—typically varying as the third, fourth, or even fifth power of age (Fig. 21-8). By contrast, if a single mutation (with a constant probability) were enough to cause cancer, there would be no such age-dependent acceleration. Epidemiological statistics suggest that, on average, transforming a normal cell into a tumor cell requires between three and seven independent random events, each of low probability. The number of such events is generally lower in leukemias and higher in solid tumors.

Fig. 21-8. Mortality from Colorectal Cancer in the United States over a one-year period as a function of age: (A) plotted on standard linear coordinates; (B) the same data plotted on double logarithmic coordinates. Cancer incidence increases dramatically with age (the slope of the curve in panel B is approximately 5). Apparently, turning a normal cell into a cancer cell requires multiple independent random events to occur within it. (U. S. Department of Health, Education and Welfare; Vital Statistics of the United States, Vol. II: Mortality. Washington, D. C.: U. S. Government Printing Office, 1968.)

Now that many specific mutations responsible for cancer development have been identified, it has become possible to study the effects of such mutant genes in Transgenic Mice (see Section 5.6.10). As we will see below (Section 21.2.6), the results obtained through this approach provide additional (and more direct) evidence supporting the hypothesis that a single mutation is insufficient to cause cancer. This hypothesis is also reinforced by numerous earlier studies on tumor progression—the process by which an initially subtle disruption in cellular behavior gradually evolves into "full-blown" cancer. These studies have provided insight into the multiple alterations required to transform a normal cell into a tumor cell, as well as the factors that trigger these changes.

21.1.5. Tumors Develop Slowly from Mildly Altered Cells [1, 5, 6]

If we examine The Link Between an external cause of cancer (where such a cause is obvious) and the onset of intensive disease progression, we find that they are almost always separated by a long latency period: the incidence of Lung Cancer rises sharply only after 10–20 years of smoking; the incidence of leukemia in Hiroshima and Nagasaki remained low for the first 5 years after the atomic bombing and peaked only after 8 years; industrial workers exposed to chemical carcinogens rarely show signs of related cancers until 10–20 years or more have passed since exposure, and so on (Fig. 21-9). During this prolonged incubation period, future tumor cells undergo a series of sequential changes. A striking example is chronic myelogenous leukemia. This disease begins with an elevated WHITE Blood CELL count that is not immediately fatal and persists for several years before transforming into a much more rapidly progressing illness that typically leads to death within a few months. In the initial chronic phase, leukemic cells differ from normal cells only by the presence of a chromosomal translocation (see Section 21.1.2). In the subsequent acute phase, the hematopoietic system becomes overrun with cells that, In addition to the aforementioned chromosomal anomaly, exhibit several new alterations. Evidently, Cells of the original clone that have undergone secondary mutations—forcing them to divide faster (or undergo more cell divisions before terminal differentiation)—begin to outpace both normal Blood Cells and their sister cells bearing only the primary defect. Carcinomas and other solid tumors are believed to develop in a similar manner. Although most human solid tumors are not diagnosed until relatively late stages, early developmental stages can occasionally be observed. A classic example is Cervical cancer. Cervical tumors originate from a Stratified Epithelium structurally similar to the epidermal layer of the skin (see Section 17.4.2). Normally, only cells in the basal layer proliferate, with new cells gradually migrating toward the epithelial surface while differentiating into flattened, keratin-rich, non-dividing cells that are eventually shed from the outermost layer (Fig. 21-10A). However, analyzing a sufficient number of such epithelial samples from different women reveals frequent patches of Dysplasia, where dividing cells are found far beyond the basal layer, indicating a disruption in the normal differentiation process (Fig. 21-10B). Cells are shed from the epithelial surface at abnormally Cytology/cytology/16.html">Early stages of differentiation, which is why dysplasia can be detected by taking an epithelial scrape and examining it under a Microscope (Fig. 21-11). Typically, such dysplastic patches are harmless and may even spontaneously regress; occasionally, however, they progress over several years into foci of so-called carcinoma in situ (Fig. 21-10C). In this more dangerous condition, the normal pattern of cell division and differentiation is severely disrupted—all layers of the epithelium now consist of proliferating, undifferentiated cells that often vary widely in size and karyotype. Nevertheless, these defective cells remain confined to one side of the basement membrane within the epithelium. At this stage, a complete cure is still possible by surgically removing or destroying the altered tissue. Without Treatment, the pathological lesion may remain benign or even regress, but in 20–30% of cases it progresses further (usually over a span of several years) to form true invasive cervical cancer (Fig. 21-10D). At this point, tumor cells breach the integrity of the epithelium, penetrate the basement membrane, and invade the underlying connective tissue. From this moment on, the efficacy of Surgical treatment declines progressively as invasive growth spreads.

Fig. 21-9. Latency period distribution for bladder cancer in a group of 78 men occupationally exposed to the carcinogen 2-naphthylamine; subgroups are categorized by duration of exposure. (After F. Cairns, Cancer: Science and Society. San Francisco: Freeman, 1978, with modifications; based on M. H. C. Williams, in Cancer, Vol. III, R. W. Raven, ed., London: Butterworths, 1958.)

Fig. 21-10. Stages in the development of cervical carcinoma. In dysplasia, the superficial cell layer retains some features of differentiation, but it is incomplete, and dividing cells are found abnormally far from the basal layer. In carcinoma in situ, cells across all layers are actively dividing and phenotypically undifferentiated. True malignancy manifests when cells breach the basement membrane and begin invading the underlying connective tissue. Several years may elapse between the first signs of dysplasia and The formation of a fully malignant tumor.

Fig. 21-11. Photomicrographs of cells from cervical scrapes (Papanicolaou or "Pap smear" technique). (A) Normal: cells are large and well differentiated with densely condensed nuclei. (B) Dysplasia: cells are at various stages of differentiation, interspersed with completely immature forms. (C) Invasive carcinoma: all visible cells are undifferentiated, with scanty Cytoplasm and relatively large nuclei; blood cells destroyed within carcinoma-ulcerated areas are visible in the Background debris. (Courtesy of E. Miller.)

21.1.6. Tumor Progression Involves Successive Cycles of Mutation and Natural Selection [6, 7]

In the previous section, we examined two very different examples illustrating this principle: broadly speaking, cancer development can be described as a process whereby an initial population of mildly altered cells—descendants of a single mutant progenitor cell—moves "from bad to worse" through successive cycles of mutation and natural selection. Because the element of chance in this evolution is very high, it typically spans many years, and most people die of other causes before cancer has time to fully develop. Understanding the causes of cancer requires identifying the factors that can accelerate this process.

Generally speaking, The rate of evolution—whether dealing with a population of cells "attempting to act" like cancer within an organism, or a population of organisms adapting to a new environment—must depend on four main parameters: (1) mutation rate, or the frequency of mutations (the probability that any given member of the population undergoes a genetic change); (2) population size; (3) reproduction rate (the average number of progeny generations per unit time); and (4) the selective advantage of the mutant individual, measured as The ratio of surviving fertile offspring produced per unit time by the mutant versus the non-mutant. Selective advantage depends on both the Nature of the mutation and environmental conditions. The situation can be further complicated by heritable epigenetic changes.

The validity of these evolutionary principles is well illustrated by studies of experimental cancer in animals. Reviewing data from these studies conveys a sense of the immense variety of factors influencing human cancer incidence—from smoking (in lung cancer) to the age at which a woman has her first child (for breast cancer). It is abundantly clear, at the very least, that the mutation rate per cell is not the sole variable dictating cancer development.

21.1.7. Cancer development can be promoted by factors that do not alter the nucleotide sequence of cellular DNA [6, 8]

The stages a damaged cell goes through before turning into a cancerous one are most easily observed in the skin. In mice, skin cancer can be induced experimentally by painting the skin with the chemical carcinogen benzpyrene (a component of coal tar and tobacco smoke) or a related compound, dimethylbenzanthracene (DMBA). A single contact with a carcinogen does not normally lead to tumor formation on its own (nor to any other obvious, long-lasting disruption). Yet, it causes a latent genetic lesion that manifests as a sharp increase in tumor incidence upon repeated contact with the same carcinogen or exposure to other agents of a completely different nature. In such situations, the carcinogen is said to act as a tumor initiator. Even simple trauma to skin that has once been exposed to such an initiator can trigger cancer (almost certainly developing from cells located at the wound's edge). At the same time, repeated contact over several months with substances known as tumor promoters (which are not mutagenic on their own) can induce tumor growth selectively in the areas of skin that came into contact with the tumor initiator. Among tumor promoters, the most thoroughly studied are phorbol esters, such as tetradecanoyl phorbol acetate (TPA), which we discussed earlier in a different context as artificial activators of protein kinase C (and thus as agents that activate part of the phosphatidylinositol Intracellular Signaling pathway [see Sec. 12.3.10]). These substances induce cancer at high frequency only after prior exposure to a mutagenic initiator (Fig. 21-12).

Fig. 21-12. Tumor formation resulting from the combined action of a tumor initiator (mutagen) and a tumor promoter (non-mutagen). The disease develops only when contact with the initiator precedes exposure to the promoter, and when the intensity of the promoter's action exceeds a certain threshold. Tumor development also occurs upon repeated exposure to a single initiator.

The latent Damage caused by a tumor initiator, like genetic damage, is irreversible and can therefore manifest upon exposure to a tumor promoter even after a long period of time. The immediate primary effect of a promoter appears to be The stimulation of cell division (or the retention of cells that should have terminally differentiated in a continuing state of division). Numerous small, benign, wart-like growths called Papillomas begin to proliferate at the sites of initiator exposure. The larger the initial dose of the initiator, the greater the number of papillomas formed. It is believed that each papilloma—at least at low doses of the tumor initiator—originates from a single clone of cells, the progeny of a mutant cell generated by the action of the initiator. A tumor promoter and mechanical injury can induce the expression of certain "social control" genes that directly or indirectly affect cell proliferation (see Sec. 13.4). In quiescent epithelium, such genes may remain "silent," and thus any mutations arising in them As a result of initiator action stay hidden; however, the induction of their expression by a tumor promoter or injury-derived factors leads to the manifestation of these mutations and, consequently, altered cell proliferation (Fig. 21-13).

A typical papilloma may contain approximately 10⁵ cells, which is more than a thousand times the number of cells in a normal "epidermal proliferative unit" (see Sec. 17.4.2). If the tumor promoter is removed, virtually all papillomas regress, and the skin returns to an almost normal appearance—consistent with the hypothesis illustrated in Fig. 21-13. However, further changes occur in some papillomas, and they acquire the capacity for uncontrolled growth even after the promoter is removed. These changes apparently arise in single cells within the papillomas at a frequency similar to that of spontaneous mutations. This is the pathway by which a small fraction of all papillomas turns into malignant tumors. Thus, a tumor promoter promotes cancer development (at least in this system) by expanding the population of cells carrying the primary mutation: the more such cells there are, the higher the probability that at least one of them will mutate further in the direction of malignancy. Even if "natural" tumors do not necessarily arise in exact accordance with this described scheme—that is, passing through distinct and Sequential Stages of initiation and promotion—their development follows similar patterns. They will also grow at a rate that depends both on mutation frequency and on Factors influencing the survival, reproduction, and spread of certain types of mutant cells should they arise.

Fig. 21-13. A possible MECHANISM OF ACTION for tumor promoters. According to an alternative hypothesis, a mutant gene may be expressed continuously but produce no phenotypic effect until a promoter activates additional genes required for cell proliferation.

21.1.8. Most cancers are caused by avoidable combinations of environmental factors [9]

Cancer development is a multistage process. The progression of each stage is influenced by numerous factors; some depend on the individual's genetic makeup, while others depend on environmental conditions and lifestyle. Therefore, by changing our environment and/or habits, we can, in principle, significantly reduce the risk of developing virtually any form of cancer. The most striking evidence for this comes from the comparative incidence of cancer across different countries. For almost any type of cancer that is common and widespread in a given country, there is another nation where the incidence of that same cancer is several times lower (Table 21-2). Furthermore, cancer rates in immigrant populations tend to converge toward those of the native population, indicating the leading role of environmental rather than genetic factors. Based on these data, it has been estimated that 80–90% of all cancer cases can be prevented. Unfortunately, different cancers have distinct environmental risk factors, and a country that has successfully minimized one of them holds no advantage over others regarding the rest. For this reason, overall cancer statistics (combining all types of malignancies) reveal a remarkably similar age-specific incidence pattern across all nations. At the same time, certain population groups whose restrained lifestyles lead to markedly reduced cancer mortality do exist; for instance, among orthodox Mormons in Utah, cancer mortality is half the U.S. national average.

Table 21-2. International variations in the incidence of the most common cancers

Data for cancers of the cervix, breast, uterus, and ovaries are for women only; the rest are for men. Cumulative incidence is defined as the percentage of people who develop cancer by age 75 (in the absence of other causes of death); incidence ratios are calculated for the 35–64 age group. (Slightly modified from: R. Doll and R. Peto, The Causes of Cancer. New York: Oxford University Press, 1981.)

Fig. 21-14. Probability of breast cancer in women as a function of age at first childbirth. The older a woman is when she gives birth for the first time, the higher her risk of breast cancer. Apparently, a specific combination of Sex Hormones can promote tumor development. Laboratory studies indicate that a first full-term Pregnancy can cause a stable epigenetic change in breast epithelial cells, dictating their subsequent hormonal responses. Breast cancer risk also correlates with other factors, such as dietary fat intake. (F. Cairns, Cancer: Science and Society. San Francisco: Freeman, 1978. After B. MacMahon, P. Cole, F. Brown, J. Natl. Cancer Inst., 50, 21–42, 1973.)

Although these epidemiological observations strongly suggest that cancer is largely preventable, identifying specific risk factors and their Mechanisms of action remains a formidable challenge. Some of these factors are clearly mutagenic tumor initiators that directly induce genetic alterations; others act predominantly as tumor promoters, expanding the population of cells susceptible to malignant progression via further mutations. The carcinogens found in tobacco smoke, as well as aflatoxin (see Fig. 21-6), belong primarily to the first category; Female Sex Hormones (Fig. 21-14) can be assigned to the second.

Some factors may operate through mechanisms entirely distinct from these two categories—for instance, by inducing heritable epigenetic changes. Of course, identifying and avoiding a carcinogenic factor does not strictly require a complete understanding of how it works; in this regard, cancer Epidemiology has already made great strides and remains a highly promising field of research. Simply clarifying the role of smoking in lung cancer has paved the way for reducing overall cancer mortality in North America and Europe by 30%. Cancer Prevention is not only more appealing than treatment, but given our Current state of knowledge, it is vastly more effective (and significantly cheaper to boot).

21.1.9. Finding cures for cancer is difficult, but not hopeless [10]

Treating cancer is as difficult as weeding a field. Tumor cells can be removed surgically or destroyed using Chemotherapy and radiotherapy, but eradicating every single tumor cell—that is, eliminating them without exception—is exceptionally challenging. Surgery rarely detects all existing metastases, and chemical agents that kill cancer cells are typically toxic to normal cells as well. Even if only a few tumor cells remain in the body, they can proliferate and cause a relapse. Furthermore, unlike normal cells, cancer cells can develop resistance to the drugs used against them. Yet the situation is not hopeless. Despite the difficulties, effective treatments have been developed for certain types of cancer (using anticancer drugs, alone or in combination) that previously carried high mortality rates (among them Hodgkin's lymphoma, testicular cancer, Choriocarcinoma, and certain types of leukemia and childhood malignancies). Modern antineoplastic drugs, surgery, and localized Radiation therapy can restore the majority of patients to a normal life if the disease is detected at a sufficiently early stage. Even when a cure is out of reach, there are ways to prolong life or at least alleviate suffering.

The central objective in clinical cancer research is the selective destruction of tumor cells. Current therapies rely chiefly on relatively subtle differences in proliferation rates, METABOLISM, and radiosensitivity between normal and tumor cells, which inevitably leads to undesirable toxic side effects. Several types of cancer cells are particularly vulnerable to targeted interventions because they are hormone-dependent or display atypical chemical structures on their surface that can be recognized by Antibodies. Overall, however, progress on the much-discussed problem of selective antitumor therapy remains slow, advancing through a combination of hypothesis, trial and error, and rational design.

To discover optimal ways of "taming" the proliferation and spread of tumor cells, it is essential to study and understand the underlying principles (or core mechanisms) of these processes in greater detail.

21.1.10. Tumor growth is frequently associated with impaired Cell Differentiation [11]

Up to this point, we have emphasized that unrestrained cell division is the defining feature of tumor cells. However, many tissues are organized in such a way that even an uncontrolled increase in cell division frequency does not automatically lead to a progressively growing tumor. This principle is illustrated by the example of cervical cancer discussed earlier (Sec. 21.1.5). Like the skin epidermis and many Other types of epithelium, the cervical epithelium normally undergoes continuous self-renewal: differentiated cells are shed from its outer surface and constantly replaced through the division of stem cells in the basal layer (see Sec. 17.4.4). On average, each normal stem cell division yields one daughter stem cell and one cell destined to stop dividing and undergo terminal differentiation. If a stem cell simply divided faster, it would merely accelerate tissue turnover—differentiated cells would form and shed more rapidly, thereby preserving tissue Homeostasis. In contrast, when a transformed cell produces a continuously growing progeny, this balance must be disrupted: either more than half of the daughter cells remain stem cells, or the differentiation process is altered such that differentiating daughter cells retain the capacity for unlimited division, escaping their normal fate at the end of their "lifespan"—shedding (Fig. 21-15).

Fig. 21-15. Stem cell strategies and their role in cell differentiation. Two Types of defects that can lead to the relentless proliferation characteristic of cancer. Note that an elevated rate of stem cell division on its own does not lead to this outcome.

There is good reason to believe that the development of such properties underlies the transition from moderate cervical dysplasia to carcinoma in situ and a fully formed malignant tumor (see Fig. 21-10). Similar assumptions apply to the development of cancer in other tissues that undergo renewal via stem cells, such as the skin, intestinal epithelium, and hematopoietic system. For example, certain forms of leukemia arise from a disruption in the normal differentiation program, where an intermediate precursor of a blood cell lineage continues to divide indefinitely instead of completing differentiation after a strictly determined number of cell cycles (see Section 17.5.5).

In general, mutations or epigenetic changes that block normal cell maturation and transformation into non-dividing differentiated forms must play a critical role in the Pathogenesis of many malignancies. Consequently, there is hope that agents stimulating cell differentiation could prove promising in cancer therapy, either as a primary treatment or as an adjunct to cytotoxics—drugs that kill dividing cells.

21.1.11. To form metastases, tumor Cells must be able to invade through the basement membrane [12]

The principal property of tumors that complicates their treatment by surgery or local radiotherapy is the ability to metastasize. To spread throughout the body, the cells of a typical solid tumor must break free from mechanical contact with neighboring cells, "escape" from the tissue of origin, "crawl" through other tissues to reach a blood or lymphatic vessel, penetrate the basement membrane and the endothelial layer to enter the vessel lumen, and then once again "forge through" the vessel wall in the reverse direction at another site in the body, finally surviving and proliferating in their new environment (Fig. 21-16). The final stages are evidently the most difficult; many tumors release vast numbers of cells into the bloodstream, but only a tiny fraction of these cells prove capable of forming metastases (metastatic colonies).

Fig. 21-16. The Mechanism of metastasis. This example illustrates the spread of a tumor from the lung to the liver. Tumor cells can enter the bloodstream either by crossing the wall of a blood vessel, as shown in the figure, or, as is likely more frequent, via The Lymphatic system. Lymphatic vessels empty their contents (Lymph) into the bloodstream, but tumor cells are often trapped in Lymph Nodes along their path, giving rise to secondary tumors there. Animal studies have shown that out of all the tumor cells entering the blood, only a few possess the ability to establish a tumor at a new site. The "success" of metastasis depends both on The properties of the tissue the tumor cell attempts to invade and on the CHARACTERISTICS OF THE cancer cell itself.

Fig. 21-17. An experiment demonstrating heritable differences among cells of the same tumor regarding their metastatic potential. Cells from a single line are subcloned, and standard aliquots of each subclone are injected into the bloodstream of mice (an assay for the ability to form secondary tumors). The subclones vary significantly in the number of metastases produced per mouse.

Certain types of normal cells, namely leukocytes, inherently possess many or even all of the properties required for dissemination throughout the organism. For most tumors, however, The Emergence of metastatic capacity is likely linked to additional mutations or epigenetic changes. Most likely, such alterations (like others involved in carcinogenesis) arise randomly within the initial tumor population, and only those few cells that acquire the properties necessary for metastasis—and happen to find themselves in a suitable microenvironment—will be able to give rise to secondary tumors. Thus, the cells within a single tumor are heterogeneous in their metastatic potential (Fig. 21-17).

An understanding of the MOLECULAR MECHANISMS OF metastasis could pave the way for developing preventive approaches. Some progress has been achieved in this direction. For instance, it has been shown that tumor cells must, first, carry Laminin receptors on their surface (see Section 14.2.18) to attach to the basement membrane, and second, secrete type IV collagenase to degrade the membrane in order to penetrate it (Fig. 21-18). Antibodies or other agents that block laminin attachment or type IV collagenase activity have been shown to suppress metastasis in experimental animals. It remains to be determined whether such anti-metastatic treatment will prove effective in cancer patients.

21.1.12 Defects in DNA repair, replication, and recombination processes contribute to the development of cancer [1, 13]

We have already noted that tumor onset and the rate of tumor progression from benign to malignant depend on the mutation rate. Mutation rates are increased by both environmental mutagens and intracellular defects in DNA replication, recombination, and repair mechanisms. For example, individuals with xeroderma pigmentosum, a rare inherited disorder, exhibit a defect in the enzyme system required to repair ultraviolet radiation-induced DNA damage (see Section 5.2.8). Consequently, even brief exposure to sunlight can trigger skin cancer. A more generalized predisposition to cancer is observed in Bloom's syndrome, which involves a defect in DNA ligase—an enzyme essential for DNA Replication and repair—as well as in Fanconi anemia and ataxia-telangiectasia, which feature impairments in the same Functions, though less well characterized. In these rare Genetic Disorders, the anomaly is inherited via germ cells and is therefore present in all cells of the body. However, similar Genetic Defects in DNA metabolism can also arise via somatic mutations, and there is reason to believe that such aberrations are a common and major factor in the development of many malignancies.

Fig. 21-18. Penetration of a basement membrane by a tumor cell.

Fig. 21-19. Typical nuclear Morphology anomalies in a cancer cell (in this case, erythroleukemia). The Nucleus of such a cell is unusually large, possesses a nuclear envelope with randomly distributed folds, and contains a nucleolus that is also abnormally enlarged and structurally complex. (Courtesy of D. Friend.)

Tumor cells frequently exhibit abnormal Variability in nuclear shape and size (Fig. 21-19), as well as in chromosome number and structure; in clinical practice, changes in nuclear morphology serve as one of the key diagnostic hallmarks of cancer. When tumor cells are cultured, their karyotype often proves to be extremely unstable: gene Amplification or deletion, and the loss, duplication, or translocation of chromosomes (or chromosomal segments) may be observed—all occurring at a much higher frequency than in cultured normal cells. On the one hand, such variability in chromosome number and structure may simply be a consequence of cell cycle acceleration, arising in a differentiated cell poorly adapted to rapid proliferation. On the other hand, it may reflect a heritable defect in the mechanism or regulation of DNA repair, replication, or recombination, stemming from a somatic mutation in any of the numerous genes involved in these complex processes. Such a mutation will increase the probability of all subsequent mutations in other gene groups. Therefore, one would expect this mechanism to be common to cells that have undergone the multiple mutations required for their malignant transformation. Suppose, for example, that the Transformation of a normal cell into a tumor cell requires three mutations in genes controlling cell behavior, and that the probability of each such mutation during a human lifespan is 10^-4 per cell. Then the probability that a single normal cell will manage to "accumulate" these three mutations (even over the specified timeframe) will be 10^-4 × 10^-4 × 10^-4 = 10^-12. But let us now assume that the mutation rate has increased due to a prior mutation in one of the Enzymes of the DNA replication or repair system, reaching 10^-2/cell over a human lifespan. Assuming the probability of this repair/replication system mutation to be the standard 10^-4, we find that this pathway—initiating with a mutation that increases mutability—will lead to a much higher frequency of cancer cell emergence: the cumulative probability of a cell transforming into a cancerous one during a lifetime will be 10^-4 × 10^-2 × 10^-2 × 10^-2 = 10^-10. This is 100 times more probable than in the first case, even though it requires four mutations instead of three.

21.1.13. High mutability of cancer cells promotes the acquisition of resistance to anti-tumor drugs [10, 14]

Regardless of the cause of the high mutability of cancer cells, in most tumors they are highly heterogeneous in many respects and capable of evolving at an alarming rate under The Influence of new selective pressures; naturally, this only exacerbates the challenges of cancer therapy. Prolonged treatment with drugs selectively toxic to dividing cells can eliminate the majority of tumor cells in a patient, but eradicating them completely is rarely achieved—typically, a small fraction of the cells proves resistant to the given drug (or class of drugs). Moreover, exposure to a particular drug can sometimes induce resistance not only to that drug but also to other agents with which the patient's cells have never come into contact.

This phenomenon of multidrug resistance often correlates with a striking karyotypic alteration: cells exhibit additional pairs of tiny chromosomes, known as double minutes, or a homogeneously staining region integrated into one of the normal chromosomes and disrupting its normal banding pattern. Both of these aberrations result from the massive amplification of a small genomic segment (see below, Figs. 21-26 and 21-31). Cloning of such amplified DNA revealed that it frequently harbors a specific gene known as the multidrug resistance gene (mdr 1). It encodes a Plasma Membrane transport ATPase that is thought to prevent the intracellular accumulation of certain classes of lipid-soluble drugs by "pumping" them out of the cell. Amplification of other gene types can also confer a selective advantage to tumor cells—treatment with the Folic acid antagonist methotrexate often leads to the amplification of the Dihydrofolate Reductase (DHFR) gene; in some tumors, as we will see below, certain Proto-oncogenes involved in cell division regulation are similarly amplified (see Section 21.2.8).

Defects in DNA replication, recombination, or repair processes, while making tumor cells evolutionarily more adaptable through increased mutability, simultaneously render them more vulnerable to specific types of interventions. This explains the well-established therapeutic fact that many tumor cells can be killed much more readily than normal cells by irradiation or treatment with specific agents interfering with DNA metabolism. A more comprehensive Study of the molecular mechanisms of DNA replication, repair, and recombination will enable the development of assays to detect impairments in these processes in individual cancer cases. Armed with such information, we could target and destroy rogue cells much more effectively by tailoring drugs that strike at their "weak spots."

Conclusion

Cancer cells, by definition, disobey normal growth controls (and are therefore termed neoplastic) and possess the ability to invade surrounding tissues and establish colonies—metastases (i.e., they are malignant). The capacity for metastasis (secondary tumor formation) complicates surgical treatment. Cancer cells typically retain many Features of the cells from which they originated ("parent" cells). Most tumors develop from a single cell that has undergone a somatic mutation; however, before giving rise to a cancerous tumor, the progeny of this cell must undergo certain changes (possibly several additional mutations). This phenomenon—tumor progression, which typically spans years—represents the evolution of somatic cells driven by mutation and natural selection. The process can be accelerated by mutagenic agents (tumor initiators) and certain non-Mutagenic Factors (tumor promoters) that influence gene expression, stimulate cell proliferation, and alter the "ecological balance" between normal and mutant cells. Because numerous factors contribute to the development of each specific case of cancer, and because The impact of some environmental factors can be eliminated, it is theoretically possible to prevent a significant proportion of oncological diseases.

A vast amount of research in oncology has been devoted to finding treatments that selectively destroy tumor cells without harming their normal neighbors. Rationally solving this problem requires understanding the specific properties of cancer cells that drive their growth, proliferation, and dissemination. For instance, tumor cell proliferation is often apparently coupled with impaired differentiation, where the progeny of a stem cell continue to divide instead of transitioning to a terminal (non-dividing) stage; in principle, proliferation can be suppressed by pushing cells toward differentiation. To become malignant, a tumor cell must acquire the ability to penetrate the basement membrane. This capacity of a cancer cell can be blocked using appropriate antibodies, thereby suppressing metastasis. The mutability of cancer cells is frequently exceptionally high; this accelerates the acquisition of the complex of properties required for neoplastic and malignant phenotypes and facilitates the development of resistance to anti-tumor drugs. Conversely, the DNA metabolism defects underlying such high mutability can render cancer cells highly sensitive to targeted therapy.