Genomics and Public Health

Breakthroughs in biology are changing our world. Just as chemistry and physics had broad ramifications in the preceding centuries, the New Biology unleashed by the Human Genome Project and associated developments will send ripples through many aspects of 21stcentury life and will be influential in improving the health of the public. The public health sciences will be essential for interpreting the health significance of genetic variation and the gene/environment interactions at the core of most diseases and biological phenomena. The combination of genomics and public health sciences will be critical to achieve the vision of predictive, personalized, preventive health care and community health services (Table 1).

Public and media interest was intense when President Bill Clinton and Prime Minister Tony Blair jointly announced accelerated progress by the public- and private-sector sequencing programs in June 2000. The details of those “blueprints” for the 25,000 to 30,000 genes of the human genome sequence were published in February 2001. These genes, through a variety of pathways, produce an even larger number of proteins, which then undergo numerous structural modifications critical to their functions; thus, the “proteome” is much more numerous and complex than the genome.

Raw genome information reveals very little directly about human health. Genome sequence information must be linked with information about nutrition and metabolism; behaviors; diseases and medications; and microbial, chemical, and physical exposures from the environment in order to understand the environmental/genetic interactions that ultimately affect human health (Table 2). Broadly, genetics and public health share salient attributes. Both focus on populations. Both seek to elucidate the larger patterns to be found among individual variations in genetic predispositions to diseases, sensitivity to environmental exposures, and responsiveness to preventive and therapeutic interventions—within and across population subgroups. Both are aware of the legacy and risks of discrimination on social and racial grounds. Thus, both explicitly recognize the importance of cultural, societal, and ethnic contexts, often explored as part of the Human Genome Project’s trailblazing investments in its Ethical, Legal, and Social Implications (ELSI) component.

Viewing health care as a part of public health, we can see that genetics provides a bridge between medicine and community-based public health, most directly in the setting of clinical genetics. Counseling and treatment of individual patients often must be expanded to nuclear or extended families. Screening for genetic predispositions involves workers and other populations suspected of having higher-than-background risks because of their exposure to potential hazards, their family history, or their ethnic background. Community outreach should be aimed not only at earlier diagnosis and treatment but at preventing problems by reducing nongenetic risk factors. The accelerating pace of discoveries and applications has put a premium on education about genetics for health professionals and education about public health for geneticists. Both disciplines have critical roles to play.

Epidemiology and randomized trials

During the past 15 years, there has been a remarkable transformation of epidemiology, highlighted by a dramatic move beyond simple statistical associations to cause-and-effect research that aims to identify the mechanisms of disease. Gene expression and protein expression are likely to be very useful early indicators of a developing disease. If this research is successful, then it will be possible to use gene or protein expression as a biomarker that will allow clinicians to identify preliminary signs of diseases with long latency periods before clinical symptoms and signs become manifest. Once etiological hypotheses have been generated and tied to credible potential mechanisms, investigators can devise clinical trials of prevention strategies that modify or remove relevant risk factors. Possible interventions include behavior change, such as helping smokers to quit. Other options include the use of antioxidant vitamins, natural products such as folic acid, or pharmaceuticals such as COX-2 inhibitors for chemoprevention.

Finding effective chemical interventions that do not have undesirable side effects is not easy. In the past few years, postmenopausal estrogen medications to reduce heart attack risks, beta-carotene to prevent lung cancers, and COX-2 inhibitor drugs to prevent the recurrence of colon polyps have been found to produce serious adverse effects. On the positive side, several new anticancer drugs (Gleevec, Herceptin, Iressa, and Avastin) successfully target specific molecular mechanisms and help well-defined patient subgroups. When the target and mechanism are precisely identified, the likelihood of unintended consequences is greatly reduced. For infectious agents, vaccines can be especially effective preventive interventions. For diseases related to environmental and occupational exposures, an effective combination is monitoring aimed at reducing emissions and exposures plus periodic checkups for workers to identify early clinical or laboratory indications of disease.

Molecular epidemiology and clinical research studies using “gene chips” to conduct global analyses of gene expression have yielded patterns that reveal striking differences between normal cells and cancer cells of the prostate, lung, breast, and other tissues. Different patterns are found also between tumor specimens from patients with localized cancers as compared with patients with invasive or metastatic cancers of the same type. The ability to distinguish among tumors that appear identical to the pathologist and surgeon should make it easier to prescribe appropriate treatment much earlier in the progression of disease and thus improve prospects for the patient.

The next critical step for epidemiology and clinical trials will be the identification of circulating biomarkers, particularly proteins, that make it feasible to diagnose cancers (or other diseases) much earlier. These protein biomarkers might be tumor proteins secreted or released during cell turnover. They also might be autoantibodies that the body makes to combat tumor antigens; in this case, the immune response represents a much-needed “biological amplification,” with much higher concentrations of the antibodies than of the original tumor antigen in the circulation. In fact, the most favorable outcome will be achieved when we can diagnose the development of cancers before they grow big enough to be visualized by x-ray or imaging, have a molecularly targeted treatment for the particular mechanism of that tumor, and have sufficient confidence that the diagnosis is correct and that a treatment will be effective to treat the whole patient without waiting for the tumor to grow, develop additional mutations, and metastasize.

Information technology

Database structures, software tools, data mining, and molecular modeling are critical elements of modern molecular research. Generally, they need to be coupled with sophisticated statistical design and analyses. Thus, scientists from a wide swath of mathematical fields have been recruited to studies of the genome, gene expression, protein expression, and metabolic patterns in health and disease. Under the broad banner of “systems biology,” researchers seek to link studies at each of these molecular levels with physical manifestations of disease or health problems that can be observed by a physician. As the biomedical literature has grown, natural language processing has emerged as a valuable approach to automated searching of the literature and databases.

An example of the emergent properties of databases is a report by graduate student Dan Rhodes and his faculty colleagues at the University of Michigan. They merged results of genetic analyses of many different types of cancers and uncovered several types of gene overexpression that were common to more than one type of cancer. This opens the possibility of developing therapeutic interventions that could be successful in some patients across a variety of cancers. Drugs that might work on several different kinds of cancers are of particular interest in overcoming what has been called the “pharmacogenomics nightmare” for drug development companies. We are finding that individuals can differ significantly in their response to therapies. For example, the drug Herceptin is particularly effective for the 15 percent of breast cancer patients who overexpress a specific chemical receptor, and the drug Iressa works only in the 10 percent of lung cancer patients with a mutation in a particular part of a different receptor. Patients without these specific molecular features will not respond. From the company perspective, a drug that works on only 10 percent of lung cancer patients has a far less attractive market than one that works on all or at least is prescribed for most patients just to try. On the other hand, a drug that is effective in treating lung cancer in patients with a specific genetic characteristic might also be found to be effective against other types of cancers with that characteristic. In that case, a company might find that the market for the drug is attractively large, and the overall benefit for patients could also be much greater, with a far more favorable benefit/risk ratio for patients.

Ecogenetics

After decades of polarized views pitting “genetic” versus “environmental” or “nature” versus “nurture” as the cause of various diseases, there is now widespread recognition that the more appropriate concept is “ecogenetics”—the realization that interacting genetic and environmental factors together influence predisposition to, or resistance to, developing specific diseases. The National Institute for Environmental Health Sciences has led the development of the subfields of “toxicogenomics” and now “toxicoproteomics.” The goal is to identify molecular signatures for exposures, early effects, and differential susceptibility to chemical agents that cause cancers, mutations, birth defects, and organ system dysfunction. This work is still in an early phase.

Ecogenetics fits into broad public health constructs for dealing with health risks of environment origin. In fact, the statutory and regulatory rationale for ecogenetics studies is quite explicit. The Occupational Safety and Health Act of 1972 mandated that health standards be set “such that no worker shall suffer adverse effect … [even] if exposed at the maximal permissible level for a working lifetime [45 years].” Like other physicians who see patients with workplace exposure-related clinical conditions, I have heard patients ask the logical question, “Why me, Doc? I’m no less careful than the next guy.” That is a question that we hope to answer by exploring what factors determine individual susceptibility.

The Clean Air Amendments of 1977 required that allowable levels of ozone, nitrogen dioxide, carbon monoxide, sulfur dioxide, lead, and particulate matter be set “so as to protect the most susceptible subgroup in the population.” That proviso can be met only if there are studies to define the most susceptible subgroup and the levels of exposure that are hazardous for that subgroup. In the case of the ozone standard in inhaled air, the Environmental Protection Agency (EPA) based its update in 1979 on susceptibility for the large population subgroup with asthma, bronchitis, or emphysema (3 to 5 percent of the general population); they could have proposed persons with cystic fibrosis, a smaller and probably more susceptible subgroup that had not been studied. Finally, the Food Quality Protection Act of 1996 requires EPA and Food and Drug Administration regulators to address risks for vulnerable or unusually exposed subgroups. In response, the EPA has given special attention to the estimation of risks to children from exposures to pesticides or pesticide residues.

The place of individual differences in susceptibility to environmental agents was highlighted 25 years ago in a report from the White House Office of Science and Technology Policy (Table 3) and expanded in the much-cited 1983 National Research Council report Risk Assessment in Federal Government: Managing the Process, commonly called “The Red Book.” Later, a Presidential/Congressional Commission on Risk Assessment and Risk Management created a six-stage Framework for Risk Management that introduced three enhancements: (1) putting each [new] problem into broader public health or ecological context; (2) engaging stakeholders from the start, in order to better inform the characterization of risks, development of options, decisions for risk reduction, and support for recommended actions; and (3) insisting on a postimplementation evaluation of benefits, costs, and unintended hazards.

One of the most active areas of genomic research has been the elucidation of genome sequences of many dozens of bacteria and fungi. Genomics has provided insights about gene content, repetitive sequences, sequence similarities, mobile genetic elements, and large numbers of genes of previously unknown function. Genes that are unique to a given species or virulent strains of a species are potential targets for selective therapy and vaccine development. In general, organisms that need to survive diverse environments have larger genomes with comprehensive biosynthetic pathways, whereas obligate parasites tend to have smaller genomes with adaptations that facilitate an existence entirely dependent on their hosts. Some of these adaptations are clearly important to public health. For example, Mycobacterium tuberculosis has genomic expansions of enzymes involved in lipid metabolism and cell wall biogenesis, which facilitate resistance to anti-tuberculosis (TB) drugs; this organism uses enzymes that seem to enhance its survival in the lung tissue of humans. Various disease-causing organisms contain “pathogenicity islands”: regions of 10 to 200 kilobase pairs with distinctive features that are determinants of bacterial virulence. Conversely, the human host has polymorphisms in genes that alter susceptibility to infections and response to antimicrobial drugs. There are quite prominent examples of ecogenetic relationships between variation in susceptibility and the infectious agents of malaria, TB, HIV-AIDS, cholera, and meningitis-otitis. With the combined Gates Foundation/ National Institutes of Health Grand Challenges Initiative on Global Infectious Diseases, the biology, host variability, and targets of opportunity for new drugs and new vaccines are topics of greatly increased salience. The anticipation of bioterrorism threats puts a high premium on related studies of infectious agents and counterterrrorism strategies.

Nutrition and genetics

The diet is a key source of environmental variables: both nutritive factors and contaminants. Genomics and proteomics can help bring modern biology, chemistry, toxicology, and epidemiology to nutritional sciences. We already know in a general sense that genetic factors are important in common diseases with substantial dietary influences, beginning with obesity, diabetes mellitus, and heart disease; and we know that components of foods can induce carcinogen-activating and detoxifying enzymes that exist in variant forms. We are learning more about specific enzymes. For example, research has revealed that a variant of one enzyme is associated with low folate levels, with associated abnormalities that lead to increased risk of cardiovascular disease and death. Low folate levels are also associated with higher risk of colon cancer in women with a family history of common colon cancers. Treatment with folate is expected to reduce these risks. Administration of folic acid supplements to women before they are pregnant has been demonstrated in clinical trials in multiple countries to markedly reduce the incidence of very serious neural tube closure birth defects such as spina bifida. In fact, that benefit is so dramatic that in 1997 the United States and Canada mandated the fortification of flour products in the food supply with folic acid to ensure that the entire population gains the benefit. Genetic testing can also identify individuals with a predisposition to develop hemochromatosis, a tendency to absorb excessive amounts of dietary iron, with subsequent iron overload and deposition in various organs. Treatment is simple and inexpensive: Just donate a unit of blood every few months. Ironically, a large-scale trial at Kaiser Permanente in California showed that most people with genetic predisposition to this iron-overload disorder are minimally affected, putting a halt to plans to have widespread testing of the population.

Evolutionary aspects of nutritional genetics are quite significant. Our current related epidemics of obesity and diabetes mellitus surely reflect the consequences of the common availability of excess calories for a species that evolved with a need to sustain glucose levels during prolonged and highly variable intervals between eating and to survive frequent famines. Humans evolved to survive famine, not feast. Population differences in digestion of the sugar lactose in milk reflect different times of domestication of milk-producing animals and the introduction of milk after weaning as a major component of the diet. To this day, large proportions of non-Caucasian populations continue to turn off the lactose-digestion enzyme after weaning and so tend to be lactose-intolerant, with digestive discomfort when drinking substantial quantities of milk.

We are also gaining insights into the role of genetic factors in behaviors. There are numerous publications about genetic polymorphisms in dopamine receptors and many other functions related to cigarette smoking and the complications of smoking. Genetic variation in alcohol metabolism and in immediate and long-term organ damage from excessive alcohol intake is known, as well. There can be no doubt that there is important genetic variation in other unhealthful behaviors, probably including a lack of interest in physical activity.

Health services research

An important part of public health research is focused on the organization, effectiveness, ethics, and costs of community-based and clinically based health care services. As many diagnostic and prognostic genetic and protein tests are introduced and clinical genetic services are extended to more people, society will need well-framed research on what works and what does not, what is safe and what is not, and how best to make useful tests and services cost-effective against the backdrop of burdensome total health care costs and inequitable access to health insurance and health care. In general, information will be needed about the heterogeneity of genetic predispositions and their interactions with nongenetic exposures and other factors that influence disease risks and responses to treatment and preventive interventions. In pursuing these objectives, we must anticipate and recognize cultural, social, ethnic, and racial context to avoid discrimination based on genetic and related information. Community-based research studies should be designed to embrace the following principles: involve community partners from the earliest stages; ensure that community partners have real influence on the project; invite community members to be part of the review, analysis, and interpretation of findings; make sure that relevant research processes and outcomes benefit the studied community; make productive partnership last beyond a single project; and empower community members to initiate projects.

In the United Kingdom, discussion and planning for a BioBank with specimens and information from 500,000 individuals and their families began in 2001. In Iceland and more recently in Estonia, prospective genotype/phenotype studies are already under way, with quite a lot of speculation about the for-profit interests of deCode Genetics in Iceland and access by others to the proprietary findings. The European Union has instituted quality assurance and harmonization of genetic tests, though laboratory participation so far is disappointing. In the United States, the National Institute for Child Health has launched a long-term children’s study, the National Human Genome Research Institute is in the early stages of planning for a gene/environment cohort study, the National Cancer Institute has held a workshop with a similar thrust, and the National Institute for Environmental Health Sciences is assessing personal and community exposure measures for such studies. Planning is likely to take at least several years, given the complexities and costs. Complex choices are required about numbers of individuals to enroll, of what ages, with or without other family members; and how to collect, store, analyze, and share the specimens and data while earning the confidence of participants that their personal data will be kept confidential. At the same time, some participants and their advocates demand that the researchers report in real time to participants any findings that might have serious clinical implications. The principles mentioned above for community-based research should be applied to planning for these prospective studies.

Policy choices

The broad public/private Partnership for Prevention issued a report in 2003 called Harnessing Genetics to Prevent Disease and Improve Health: A State Policy Guide (www.prevent.org). Its stated aims are to help state policymakers to protect consumers; monitor the implications of genetics and genomics for health, social, and environmental goals; and ensure that genetic advances will be tapped not only to treat medical conditions but also to prevent disease and improve health before people become ill. The report is optimistic that the genomic era can lead to personalized health care and pharmacogenetics-enhanced drug development to prevent or better manage chronic diseases, with products and services that include diagnostic tests, drug therapies, and drug-monitoring protocols.

Its key policy finding is that genetics and genomics should be integrated into existing health, social, and environmental policies, rather than establishing stand-alone genetics programs. Policymakers at the state and federal levels should follow the example of Michigan, where a Governor’s Commission on Genetic Policy and Progress adopted an integration perspective and urged that genetic issues be dealt with in the context of overall medical care values and principles.

The case for integration is strong. All health conditions have a genetic basis. Most common diseases result from gene/environment interactions, so genetic advances are likely to extend and expand, not supplant, current practices in medicine, public health, and environmental protection. Because there is wide variation in the extent to which genetic factors affect health risks, a one-size-fits-all policy is inappropriate. Decisions about genetic policies involve complex issues about ethics, costs, benefits, and individual and societal interests. Medical care decisions should be linked with research, insurance, and broader public health policies. Finally, the intersection between genetics and public policy is both immediate and long-term, warranting close monitoring and timely actions in a broad context. Nevertheless, special legislative action seems warranted to prevent discrimination based on genetic tests or traits by insurance companies or employers.

As an Institute of Medicine committee recommended in Who Will Keep the Public Health Healthy? Educating Public Health Professionals for the 21st Century (2002), the states should invest in reciprocal training in genetics for health professionals and in public health for geneticists. The avalanche of genomic information that links specific genetic factors with disease risks will only grow larger. Thus, it is essential, even urgent, to identify, develop, and enhance environmental and behavioral data sets for eco-genetic analyses and make sure that they can be linked to data on genetic variation. If the genetic data are made anonymous, such linkages and the capacity to link genes with risks (genotypes and phenotypes) will be undermined. Simultaneously, a foundation of credible privacy and confidentiality protections, not just for genetic information but for all personal and family medical information, must be established.

We envision breakthrough tests, vaccines, drugs, behavior modification strategies, environmental exposure reductions, and epidemiological surveillance to reduce health risks and cost-effectively manage illnesses and disease risks in the United States and globally (Table 4). But we should not let our enthusiasm for these potential benefits distract us from the slow but necessary work of ensuring that we use these tools wisely for the benefit of the greatest number of people.