2
Keynote Addresses

CANCER, THE ENVIRONMENT, AND THE MEDIA

Samuel Donaldson

I know a little bit about cancer and the environment, because I am a member of the cancer club—the melanoma branch—and since receiving this diagnosis, I have paid more attention to developments in cancer research than I did in the past. We have learned a lot about the effects of the environment in causing cancer, in some cases, far too late.

When I was in the army in the 1950s, I wangled an assignment to French-man Flat in Nevada. In 1958, along with some of my buddies, I crouched in a trench 3,000 yards from a tower where a nuclear device was exploded. The device turned out to be no larger than the ones we dropped on Hiroshima and Nagasaki, but witnessing this event made me a believer in the power of the nuclear bomb. The ground shook and the countryside—the desert—came in on us. We had our hands over our closed eyes in a 6-foot slit trench, yet I saw bright daylight. Then, when the implosion occurred and the countryside came back in on us, we jumped out of the trench. At this point, the fireball was just decreasing in luminance. It was mainly white. In a few moments, as we sat there watching, white rain began to fall on us. The winds had changed. They carted us quickly to water trucks, which were about 2 or 3 miles away, and we stripped and they hosed us down. The photo tags we wore showed we had been exposed only to background radiation.

If we can teach young people early on these good habits, we have a better chance of facing down melanoma in the future.

Sam Donaldson

I do not think that this incident caused my melanoma; rather, overexposure to the sun is the more likely culprit. We now know to wear sunscreens and



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 11
Cancer and the Environment: Gene-Enviroment Interaction 2 Keynote Addresses CANCER, THE ENVIRONMENT, AND THE MEDIA Samuel Donaldson I know a little bit about cancer and the environment, because I am a member of the cancer club—the melanoma branch—and since receiving this diagnosis, I have paid more attention to developments in cancer research than I did in the past. We have learned a lot about the effects of the environment in causing cancer, in some cases, far too late. When I was in the army in the 1950s, I wangled an assignment to French-man Flat in Nevada. In 1958, along with some of my buddies, I crouched in a trench 3,000 yards from a tower where a nuclear device was exploded. The device turned out to be no larger than the ones we dropped on Hiroshima and Nagasaki, but witnessing this event made me a believer in the power of the nuclear bomb. The ground shook and the countryside—the desert—came in on us. We had our hands over our closed eyes in a 6-foot slit trench, yet I saw bright daylight. Then, when the implosion occurred and the countryside came back in on us, we jumped out of the trench. At this point, the fireball was just decreasing in luminance. It was mainly white. In a few moments, as we sat there watching, white rain began to fall on us. The winds had changed. They carted us quickly to water trucks, which were about 2 or 3 miles away, and we stripped and they hosed us down. The photo tags we wore showed we had been exposed only to background radiation. If we can teach young people early on these good habits, we have a better chance of facing down melanoma in the future. Sam Donaldson I do not think that this incident caused my melanoma; rather, overexposure to the sun is the more likely culprit. We now know to wear sunscreens and

OCR for page 11
Cancer and the Environment: Gene-Enviroment Interaction hats, and if we can teach young people early on these good habits, we have a better chance of facing down melanoma in the future. The point of telling the story about the atomic explosion was that exploding the atmosphere was something that seemed necessary during the Cold War, but we now know that exposure to strontium-90 and other noxious radiation from atmospheric testing could eventually kill us all. In 1963, led by President John F. Kennedy, we signed the first test ban treaty. For many years, nearly no country has tested in the atmosphere and we are healthier because of it. What is important about this story is that there was such broad agreement in the scientific community about the health effects of such testing that it was impossible for policymakers to ignore the evidence. Let me just tell you at this point what you already know about those of us in the news media. We look for the people who are the “odd steppers”—who march to different drummers. If you think about it, you really want us to do that. In many areas, it is the odd steppers—Copernicus, Galileo—who turn out to be right. If the news media ignores such people because conventional wisdom contradicts them, then a lot gets lost. For the media to do its job, we must seek out all opinions, even if they are unpopular and unconventional. I recognize that in science this works against you. For example, in 1964 I was at a press conference when Dr. Luther Terry presented his first great Surgeon General’s report on smoking. As a smoker and a reporter, I had to pay attention. So, I actually studied the graphs and looked at the charts and supporting documents and was almost instantly converted into a believer. Of course, it took me seven more years before I could finally kick the habit and not smoke another cigarette. However, during that period and long after, there were many other voices saying that the link between smoking and cancer had not yet been proven. Of course, many of these voices were coming from the tobacco industry’s “smoking and health” divisions, even into the 1990s. Today, to a large extent, those voices have been discredited. Even so, the fact that they existed and came from platforms that seemed to give them standing, helped dilute the message that smoking can kill you. I recently moderated a panel on global warming, during which young people were asking questions. One young lady asked this distinguished panel whether global warming was really occurring. She asked, “Is it true? Does everybody agree on this?” One of the scientists gave a very good answer, telling her that several years ago an international panel of scientists developed a consensus statement that the global temperature will rise anywhere from 3.5 to 9 degrees within 100 years if nothing is done about present carbon dioxide emissions and other causative factors. He also responded that the news media would always be able to find one scientist who will say, “No, it is not true.” He lamented that the media would then report that some scientists say global warming is a fact and others say it is not, which will confuse the issue. He was quite right. In another example, in the mid-1970s the House Select Committee on As-

OCR for page 11
Cancer and the Environment: Gene-Enviroment Interaction sassinations assembled 20 top forensic scientists to look at the autopsy photographs from the Kennedy assassination. Nineteen of these scientists said, despite what the Zapruder film appears to show, that President Kennedy was killed by a shot from the rear. They postulated that the large flash seen on his forehead and the appearance of his neck jerking back on the film is the result of an exit wound, that is, representing the natural reflexes of the muscles when there is a trauma to the back of the head. One forensic scientist, said, “No, I think that could be an entry wound,” forever leading the popular press and the film industry to beliefs about conspiracy and cover-up. What I am saying to the scientific community is this: as it pursues, discusses, and reports its work, to the extent possible it must make clear the degree to which there is unanimity in the scientific community. That is, give the media a sense of how widespread the agreement is on a given issue. It is not our mission as reporters to be in on the science of science, but to report it. I am not going to work against the scientific community, but if I hear somebody with credibility that disagrees, I am going to call them and see what they think. I am not suggesting that scientific findings must be compromised, but rather that the magnitude of scientific differences should be explained. For example, in the case of my global warming story, one of the panelists later told me that he actually thinks the rise in temperature due to global warming over the next 100 years is likely to be closer to 9 degrees than 3.5 degrees. That helped me put the widely touted range of 3.5 to 9 into better perspective. It would also be useful if the general scientific community believes the outlier, or odd stepper, is really wrong, they should say it loud and clear, and avoid the tendency toward professional courtesy. I know with every passing year that my odds improve, but the same cannot be said for all cancer patients. Each one is an individual. Sam Donaldson It has been six years since I had a melanoma tumor removed from my right groin. I know with every passing year that my odds improve, but the same cannot be said for all cancer patients. Each one is an individual. The graphs and charts showing cancer rates and survival give us a sense of the field, but any one point on the graph is an individual, whose risks and odds of survival might differ. All of us who have had cancer know that once it strikes, it is always stalking. It is always in the back of your mind, and we want to avoid its occurence in others as much as its recurrence in our own bodies. Thus, when you are looking for the relationship between, for example, the hole in the ozone layer and cancer, I can’t help but think how important this information is to the future incidence of melanoma. The scientific community seems unanimous in saying that chlorofluorocarbons help destroy the ozone, particularly in the southern hemisphere. If we don’t do something about this, my

OCR for page 11
Cancer and the Environment: Gene-Enviroment Interaction division of the melanoma cancer club will grow, rather than decrease. If there is broad agreement about the need to act, then the scientific community must provide a solid front. I appreciate that this is a challenge—that scientists are by their very nature skeptical. To the extent that the scientific community can reach broad agreement, we will all be in your debt. If there is broad agreement about the need to act, then the scientific community must provide a solid front. Sam Donaldson GENES AND THE ENVIRONMENT IN CANCER ETIOLOGY Joseph F. Fraumeni, Jr. For some time, the epidemiologic evidence has suggested that the bulk of cancer in the population is related to environmental exposures, which are broadly defined here to include lifestyle factors, such as smoking, nutrition, and reproductive variables. Although genetic mechanisms are fundamental to the development and progression of all forms of cancer, the actual role of inherited susceptibility as an etiologic factor has been very difficult to assess. The causes of cancer in the population can be assigned to one of four broad categories: (1) inherited susceptibility alone; (2) environment alone; (3) interactions between genes and the environment; or (4) a “spontaneous” category of tumors that may arise randomly from the play of chance (see Figure 2-1). It is in the category of gene–environment interactions that tremendous interest is build- FIGURE 2-1 Categories of cancer causation in the general population. This figure illustrates the potential interaction of genes and the environment in the development of cancer. For example, the absence of genetic (-) and environment (-) influences reults in the development of a spontaneous tumor. SOURCE: Adapted from Knudson (1996). Reprinted with permission.

OCR for page 11
Cancer and the Environment: Gene-Enviroment Interaction ing, particularly as advances in molecular biology and genome technology are incorporated into epidemiologic strategies. The earliest recorded observation of an environmental risk factor occurred in Italy just over three centuries ago, when Bernardino Ramazzini reported an unusually high frequency of breast cancer in Catholic nuns, which can now be largely explained by reproductive factors and their effect on endogenous hormones. The next milestone was in 1775, when the British surgeon Percivall Pott (see Figure 2-2) discovered a cluster of scrotal cancer among young chimney FIGURE 2-2 In 1775, Dr. Percivall Pott, a British surgeon, reported one of the earliest observations on environmental cancer. SOURCE: U.S. Government.

OCR for page 11
Cancer and the Environment: Gene-Enviroment Interaction sweeps exposed to soot, which was subsequently found to contain mixtures of carcinogenic polycyclic hydrocarbons. Over the years, alert clinicians continued to provide an early clue for a series of epidemiologic studies of environmental cancer, particularly by identifying case clusters of tumors that were uncommon in the general population and appeared related to a particular exposure (see Table 2-1). The overall importance of the environment to cancer etiology can be roughly gauged by the international variation in statistics, gathered from the periodic volumes entitled Cancer Incidence in Five Continents, published by the International Agency for Research on Cancer. The differences between geographic areas with the highest and the lowest rates have ranged between 50- and 150-fold for melanoma and for cancers of the nasopharynx, prostate, and liver, to about 5-fold for leukemia (Table 2-2). Some of the geographic and ethnic variation can be related to diagnostic and reporting practices, as well as genetic factors for certain tumors such as melanoma, which tends to occur in fair-skinned populations. Nonetheless, the available evidence for most tumors suggests that environmental factors may be driving the patterns: when cancer rates from the lowest-risk countries were substracted from U.S. rates, it was estimated that perhaps 80 percent of all cancers in the United States are related in some way to environmental factors and are, thus, potentially avoidable. More persuasive evidence for environmental factors comes from studies of migrant populations in which the risk of various cancers tends to shift away from the country of origin toward that of the new country. In a case-control study of breast cancer among Asian American women, the risk varied about sixfold according to migration history. The risk was lowest in migrants from rural parts of Asia who had lived less than a decade in the United States, and highest in those born in the United States along with at least three grandparents. Thus, for breast TABLE 2-1 Some Examples of Cancer and Environmental Exposure Cancer Exposure Scrotal cancer Chimney sweeps Liver angiosarcoma Vinyl chloride Acute leukemia Benzene Nasal adenocarcinoma Hardwood dust Bone sarcoma Radium Multifocal skin cancer Arsenic Mesothelioma Asbestos Vaginal cancer DES Kaposi sarcoma AIDS/HIV NOTE: DES = diethylstilbestrol. SOURCE: U.S. Government.

OCR for page 11
Cancer and the Environment: Gene-Enviroment Interaction TABLE 2-2 International Variation in Cancer Incidence Type of Cancer H/L Highest Rates Lowest Rates Melanoma 155 Australia Japan Nasopharynx 100 Hong Kong U.K. Prostate 70 U.S. (blacks) China Liver 50 China Canada Cervix uteri 28 Brazil Israel Stomach 22 Japan Kuwait Lung 19 U.S. (blacks) India Colon 19 U.S. (whites) India Bladder 16 Switzerland India Pancreas 11 U.S. (blacks) India Ovary 8 N.Z. (Maori) Kuwait Breast 7 Hawaii (Hawaiian) Israel (non-Jews) Leukemia 5 Canada India   SOURCE: Parkin et al. (1992). Reprinted with permission. cancer, it takes a few generations of acculturation for the rates in migrants to approach the rates in the general U.S. population. Studies so far suggest that this gradient in incidence is related partly to changing reproductive factors, a more westernized diet, and an increase in height and body mass, as well as weight gain during adult life, but there are probably other factors that have not yet been detected. Further evidence for environmental factors appears in the temporal variation in incidence or mortality for certain cancers, although again the trends may be influenced by improvements in detection and reporting. Based on recent statistics from the National Cancer Institute’s (NCI’s) Surveillance, Epidemiology, and End Results (SEER) program,1 a number of cancers are continuing to show substantial average annual increases in incidence, including lung cancer among women due to cigarette smoking; melanoma due largely to sunlight exposure; pleural mesothelioma due to asbestos exposure; non-Hodgkin’s lymphoma in part due to the AIDS epidemic, but largely unexplained; hepatocellular carcinoma due to hepatitis C virus infection; renal adenocarcinoma, particularly in African Americans, possibly related to increases in obesity and the prevalence of hypertension; and esophageal cancer. When we divided esophageal cancer by cell type, again using data from the SEER program, there was a remarkable increase in one cell type, esophageal 1   The SEER program collects and publishes cancer incidence and survival data from 11 populationbased cancer registries and 3 supplemental registries covering approximately 14 percent of the U.S. population (see http://seer.cancer.govformoreinformation).

OCR for page 11
Cancer and the Environment: Gene-Enviroment Interaction adenocarcinoma. This tumor was once uncommon but is now rising more rapidly than any other cancer, by about 8–9 percent per year. In a case-control study of esophageal adenocarcinoma, we have found substantial excess risks associated with gastroesophageal reflux and obesity. Gastroesophageal reflux is also increasing in incidence and may lead to an intestinal-type metaplasia called Barrett’s esophagus, a precursor to adenocarcinoma. It is noteworthy that esophageal adenocarcinoma is much more common in Caucasian men than in African American men, whereas squamous cell carcinoma is six times more common in African American men than in Caucasian men. The vast majority of squamous tumors of the esophagus are strongly related to drinking habits and, to a lesser extent, smoking habits and a low intake of fruits and vegetables. In our case-control study of this tumor, we found that these risk factors interact with one another and contribute to the higher rates in the African American population. We also found a strong effect of low socioeconomic status, which persists after adjusting for other risk factors, contributes to the ethnic differential in risk, and probably reflects a correlated exposure (perhaps a virus or dietary component) that has so far eluded detection. Although the geographic variation of cancer within the United States is far less pronounced than the international patterns, we thought that mapping cancer mortality on a small-area scale (i.e., county level) might uncover spatial clusters of the more common tumors. A strategy was developed, with a stepwise progression of studies ranging from detailed mapping and correlation studies that generated etiologic hypotheses, to field studies in high-rate areas that tested hypotheses. We have recently published an updated atlas of cancer mortality among the Caucasian and African American populations, with the maps and other displays available on an interactive NCI Web site. The computer-generated maps in the new atlas are shown at the level of both the county and the state economic areas (combinations of counties with similar cultural and socioeconomic characteristics) for two time periods, 1950–1969 and 1970–1994. For lung cancer in men, the earlier maps showed elevated rates in urban counties in the Northeast, but the highest rates by far were displayed in a string of counties along the Southeast Atlantic Coast and along the Gulf Coast. In the more recent maps, through 1994, there was a remarkable shift, with the elevated rates moving from the urban Northeast and coastal areas to broad stretches across the Southeast, a pattern that tracks the regional changes that have occurred in smoking habits. The earlier maps for lung cancer prompted a series of case-controlled studies in the high-rate coastal counties, which drew attention to the unexpected scale and impact of asbestos exposures in shipyards, particularly those operating during World War II, and to synergistic interactions between asbestos and cigarette smoking. Another unexpected finding from the earlier maps was the clustering of elevated rates for oral cancer among women in the rural South. This pattern prompted a case-control study in North Carolina that implicated the long-stand-

OCR for page 11
Cancer and the Environment: Gene-Enviroment Interaction ing practice of snuff dipping. The risks reached as high as fiftyfold for cancers of the gum and buccal mucosa, the tissues in direct contact with the smokeless tobacco product. The study promptly led to congressional hearings and regulatory actions aimed at controlling the advertising and labeling of smokeless tobacco, educational campaigns aimed at young people, and laboratory studies that implicated tobacco-specific nitrosamines as the likely carcinogen. Since our first atlas was published, cancer mapping on a small-area scale has become standard practice in virtually every nation with vital statistics. Most remarkable have been the mortality patterns in China, where there is enormous geographic variation that has provided special opportunities to study environmental cancer. For example, the map for lung cancer among women revealed very high rates in the northeast, where case-control studies have implicated several indoor air pollutants as pulmonary carcinogens (such as cooking oil fumes from wok cooking, polycyclic hydrocarbons from coal-heating stoves, and indoor radon in poorly ventilated houses). In some high-risk areas in China, intervention studies are under way to clarify etiologic factors and hopefully develop preventive measures for certain cancers. In an endemic area for squamous esophageal cancer associated with poor nutritional status, the effects of vitamin and mineral supplements are being evaluated, both in very high risk individuals who have esophageal dysplasia and in the general population. Similar studies are under way in a high-incidence area for gastric cancer, where it is possible to monitor through endoscopy the progression of precancerous lesions (ranging from atrophic gastritis to intestinal metaplasia to dysplasia), which are present in virtually the entire adult population. Since Helicobacter pylori infection is a major risk factor for this cancer, its eradication by antibiotic therapy is one of the treatment arms in our study. The epidemiologic study of risk factors in the U.S. population is far from complete, particularly when exposure assessment is difficult—such as with diet and environmental pollution—and when relative risks are small. As expected, tobacco smoking is the main culprit, accounting for at least 30 percent of all cancer deaths. Alcohol interacts with tobacco to cause tumors of the upper aerodigestive system and liver, and also appears to be involved in breast cancer and possibly in colon and pancreas cancer. It is clear that diet- and nutrition-related variables, including obesity and physical inactivity, are very important risk factors, but there are still many questions about their overall impact on cancer, as well as the specific causative and protective elements in the diet, not to mention the role played by hormonal, metabolic, and other alterations affected by nutritional status. Infectious agents appear to contribute less to the cancer burden in the United States than in developing countries, where cancers of the liver (hepatitis B), cervix (human papilloma virus), and stomach (H. pylori) are very common, but further studies are needed on the possible role of viruses and bacteria in other tumors. In addition, the impact of chemicals from occupational exposure, environ-

OCR for page 11
Cancer and the Environment: Gene-Enviroment Interaction mental pollution, and pharmaceutical agents still has to be clarified, as does ionizing radiation to some extent, including indoor radon. Ultraviolet radiation, of course, is the major cause of skin cancers, but these tumors are usually not reflected in mortality data except for melanoma. The greatest uncertainty at present is genetic or inherited susceptibility, but there is growing evidence that its overall impact on cancer risk is likely to be greater than previously estimated from the demographic patterns of cancer, the low levels of familial risk for common tumors, the low frequency of hereditary cancer syndromes, or the calculations based on twin studies that have sharply partitioned genetic from environmental factors. In addition, as we learn more about the role of genetic susceptibility and its pathways, we are likely to gain further insights into the cancer risks associated with common exposures, including dietary components, endogenous hormones, and environmental hazards, that are acted upon by functional variants of candidate genes. A great deal of progress has been made in linking hereditary cancer syndromes to highly penetrant genes, mainly through family-based studies. As shown in Table 2-3, most of the cloned genes have turned out to be tumor suppressors, while others are proto-oncogenes (such as the RET gene in multiple endocrine neoplasia type II) or mismatch repair genes that are associated with familial nonpolyposis colon cancer. In general, once a gene is mapped to a particular locus, cloning of the gene and characterizing its function soon follow. As we learn more about the hereditary syndromes, it appears that many consist of constellations of tumors that appear to share genetic pathways. This can be seen with germline mutations of BRCA-1 and 2, which are associated with breast cancer, ovarian cancer, and other tumors as well, and with the mis- TABLE 2-3 Familial Syndromes and Cloned Tumor Suppressor Genes   Gene Locus Date Retinoblastoma RB1 13q14 1986 Wilms’ tumor WT1 11p13 1990 Li–Fraumeni syndrome P53 17p13 1990 Familial adenomatous polyposis APC 5q21 1991 Neurofibromatosis 1 NF1 17q11 1990 Neurofibromatosis 2 NF2 22q11 1993 Von Hippel–Lindau syndrome VHL 3p25 1993 Familial melanoma 1 P16 9p21 1994 Tuberous sclerosis 2 TSC2 16p13 1993 Familial breast cancer 1 BRCA1 17q21 1994 Familial breast cancer 2 BRCA2 13q12 1995 Basal cell nevus syndrome PTC 9q22 1996   SOURCE: Adapted from Knudson (1996). Reprinted with permission.

OCR for page 11
Cancer and the Environment: Gene-Enviroment Interaction match repair genes that predispose to cancer of the colon, endometrium, and other sites. The most striking example of this phenomenon is a familial constellation of cancers among children and young adults that features extremely high risks for soft tissue and bone sarcomas, and breast cancer, along with excesses of brain tumors, acute leukemia, and adrenocortical neoplasms (Li–Fraumeni syndrome). Our recent experience with these families suggests that there are excesses of other cancers arising at an early age, but at lower levels of relative risk. As in other hereditary syndromes, the affected individuals often develop multiple primary cancers, including a predisposition to sarcomas associated with radiotherapy, indicating gene–environment interaction. The variety of tumors in this syndrome has helped to dispel an earlier notion that inherited susceptibility to cancer is site-specific or tissue-specific, and is thus in line with molecular studies suggesting that fundamental biological mechanisms controlling susceptibility and cell proliferation are shared by several, if not all, forms of cancer. In the case of Li–Fraumeni syndrome, the search for an underlying mechanism was unrewarding until the molecular technology was available to uncover germline mutations of p53 in most families. This finding was of special interest, since alterations of this tumor suppressor gene are somatically acquired in more than 50 percent of all cancers in the population. It is important to note that even for highly penetrant cancer genes such as p53 or the retinoblastoma gene RB-1, tumor expression may be affected by other modifying genes or by environmental exposures, such as radiotherapy and smoking. In our cohort study of hereditary retinoblastoma, we found that half the cases developed second cancers by age 50, particularly sarcomas, melanomas, and brain tumors. The cumulative risks were 58 percent in the irradiated group and 26 percent in those who received no radiotherapy, suggesting an interaction between the RB-1 gene and radiation. An extended follow-up of this group has revealed a significant excess of lung cancer that is limited to smokers, again consistent with gene–environment interaction. In contrast, there has been no excess risk of second cancers among the cases with nonhereditary retinoblastoma. Although the single-gene mutations associated with hereditary cancer are highly penetrant and tend to carry high relative and absolute risks, they are relatively rare and appear to account for a small percentage of cancers overall. On the other hand, the polymorphic susceptibility genes or genetic variants are generally associated with low penetrance and low relative and absolute risks, but they are very common in the population (more than 1 percent) and may be involved in a high proportion of cancers through biologic interactions with environmental or endogenous exposures. It is clear that family studies and linkage analyses have been highly successful in identifying major genes and helping to revolutionize our understanding of carcinogenic mechanisms, and they can still provide insights into interactions

OCR for page 11
Cancer and the Environment: Gene-Enviroment Interaction with other genes and with epigenetic and environmental factors that modulate risk. However, it now appears that population studies of a case-control or cohort design are needed to clarify and quantify the risks associated with common susceptibility genes and their interactions with exposures. The first associations linking cancer to these genetic variants involved the so-called metabolic genes, including the cytochrome P-450 activating genes such as CYPIA1 and CYP2E1, and the conjugating or detoxifying genes such as GSTM1 and NAT2. By knowing the substrate and the metabolic pathway of the gene, one may learn about mechanisms and the exogenous or endogenous carcinogens that are otherwise difficult to identify. For example, a metabolite of alcohol—acetaldehyde—has been implicated by the heightened risk of alcohol-related oral cancer associated with a rapid-acting genotype of alcohol dehydrogenase-3, which accelerates the metabolism of alcohol to acetaldehyde. In Japan, a gene variant of aldehyde dehydrogenase, which blocks the metabolism of acetaldehyde to acetate, has been linked to oral and esophageal cancer risk. These two observations point to acetaldehyde as the likely carcinogen in alcohol-related cancer. In addition, a protective effect of dietary folate on colon cancer risk is supported by the relation to a methylenetetrahydrofolate reductase gene, while various estrogen- and androgen-metabolizing genes are under active study for breast and prostate cancer. More recently, research on genetic variants has broadened to include other functional classes of cancer susceptibility genes that have more distal or down-stream effects. These include polymorphisms affecting DNA repair and processing, cell cycle control, immune function, inflammation, growth factors, apoptosis, and angiogenesis. Further study of interactions between relatively common alleles and exposures will rely on case-control studies, which may be population or hospital based, but are often embedded within prospective studies. These studies will require sufficent sample sizes for statistical power to evaluate gene–gene and gene–environment interactions, especially when the effects are multiplicative. It also is important to ensure careful epidemiologic design with appropriate control groups, as well as highly accurate genotyping and exposure assessment, which may include acquired genetic biomarkers such as macromolecular adducts and mutational “fingerprints” of exposure. Any misclassification of the gene or the exposure will greatly decrease the power of the study and increase the sample size needed to show an effect. The challenges to population studies are formidable but they can be met. Large sample sizes are possible through collaborative multicenter studies. The development of simple, noninvasive approaches to collecting genomic DNA, such as the mouthwash rinse to collect buccal cells, should decrease the costs of sample collection and improve participation rates. The studies will require not only sound epidemiology, but also specimen processing and repository facilities,

OCR for page 11
Cancer and the Environment: Gene-Enviroment Interaction and close collaboration between epidemiologists, molecular biologists, genomicists, and bioinformaticians. Despite efforts to ensure high-quality and cost-effective experience in genotype data, the sheer volume of information and the complexity of interactions will generate a prodigious number of multiple comparisons and false-positive findings—more so when it becomes possible to conduct whole-genome scans and subclassify tumors on a molecular basis, not to mention new technologies and developments in proteomics and other areas. To avoid publication bias and assist in interpretation, it would seem important to report all the tests that are conducted in a particular study and to make available all negative as well as positive results through comprehensive databases. A major problem at present is the blizzard of positive associations that are being reported on almost a daily basis and are followed in due course by a flurry of contradictory observations by other investigators. Whenever possible, it would seem important to assemble a coalition of research groups to coordinate approaches and conduct parallel case-control studies that can provide quick replication of positive or negative findings, using independent data sets prior to publication. Other potential difficulties with population-based designs include linkage disequilibrium, in which the candidate gene may be a marker allele rather than a disease allele itself, but this problem can be minimized by further progress in sequencing the genome and understanding the biologic relevance of the allele. The potential for population stratification, also known as ethnic confounding, can be handled in the main by appropriate epidemiologic techniques. In conclusion, the available epidemiologic evidence indicates that while environmental exposures drive the demographic patterns for most cancers, there are growing indications that gene variants may have a sizable impact on cancer development by modifying the effects of exogenous or endogenous risk factors and by helping to uncover low levels of relative risk from common exposures in genetically susceptible subgroups. The identification of susceptibility or modifier genes should also help identify potential carcinogens and protective factors acted upon by the gene products and should provide insights into mechanisms and interactions that will multiply the opportunities for preventive intervention. The big challenge for epidemiology now is to develop strategies to ensure that the advances in human genomics are incorporated appropriately into population studies, as well as family-based, and hybrid studies. Joseph Fraumeni The National Institutes of Health has moved in various ways to seize the opportunities in this important area, such as the initiatives of the environmental genome project coordinated by the National Institute of Environmental Health Sciences. Another blueprint is sketched out in the NCI Bypass Budget Proposal

OCR for page 11
Cancer and the Environment: Gene-Enviroment Interaction for Fiscal Year 2002, which summarizes the extraordinary opportunities for investment in cancer research. In a chapter entitled “Genes and the Environment,” NCI outlines a series of objectives and plans that combine epidemiologic and molecular approaches in ways that may enlarge our understanding of cancer etiology and inform new clinical and public health approaches aimed at preventing and controlling cancer. The big challenge for epidemiology now is to develop strategies to ensure that the advances in human genomics are incorporated appropriately into population studies as well as family-based and hybrid studies (consisting of population and family-based components) and that these studies have the power and sensitivity to dissect the environmental and genetic influences on cancer risk.