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

Institute of Medicine (US) Forum on Emerging Infections; Burroughs T, Knobler S, Lederberg J, editors. The Emergence of Zoonotic Diseases: Understanding the Impact on Animal and Human Health: Workshop Summary. Washington (DC): National Academies Press (US); 2002.

Cover of The Emergence of Zoonotic Diseases

The Emergence of Zoonotic Diseases: Understanding the Impact on Animal and Human Health: Workshop Summary.

Show details

3Factors of Emergence

VARIATION AND INTERSPECIES TRANSMISSION OF INFLUENZA A VIRUSES

, Ph.D.

Department of Virology and Molecular Biology, St. Jude Children's Research Hospital

The problem with influenza viruses is that, unlike other respiratory tract viruses, they undergo considerable antigenic variation. Antigens, which are located on the surface of the viruses, are the substances, typically glycoproteins, that stimulate an immune response. Influenza viruses carry two types of antigens: hemagglutinin (HA) and neuraminidase (NA). Both types undergo two forms of variation. The first type, called drift, involves minor changes in the antigens. New drift strains emerge constantly, giving rise to yearly epidemics and forcing the medical community to revise the viral strains used in vaccines. The second type of antigenic variation, called shift, involves major changes in the virus's genetic makeup. These new variants are of even greater concern: new strains to which most humans have no immunity appear suddenly, and the resulting pandemics vary from serious to catastrophic. While we currently know the complete genome sequences of many influenza viruses, we do not understand the molecular basis of pathogenesis and are unable to predict which combinations of viral genes have pandemic potential.

Studies on the ecology of influenza viruses have led to the hypothesis that all mammalian influenza viruses originate from a reservoir of viruses in aquatic birds, particularly ducks. In wild birds, the viruses are spread by fecal–oral transmission through the water supply. Initial transmission of avian influenza viruses to mammals, including pigs, horses, and humans, and to domestic birds, including chickens and turkeys, probably also occurs by fecal contamination of water. Another method of transfer is by feeding pigs untreated garbage or the carcasses of dead birds. After transmission to humans or other mammals, the method of spread of influenza is mainly respiratory.

Influenza represents one of the major success stories for the World Health Organization (WHO). To cope with the variability of influenza, WHO maintains a network of more than 100 laboratories that constantly survey influenza viruses, and this information is then analyzed in four reference centers. Based on these efforts, WHO makes annual recommendations for those virus strains to be included in the current vaccine in order to stay abreast of genetic drift.

The less well resolved problem of influenza is the pandemics, which occur at irregular intervals and are to date unpredictable. In the past century, three viral subtypes have caused pandemics in humans: the Spanish flu of 1918–19, which was caused by the H1N1 subtype; the Asian flu of 1957, which was caused by the H2N2 subtype; and the Hong Kong flu of 1968, which was caused by the H3N2 subtype. The H1N1 and H3N2 subtypes also have caused disease outbreaks in pigs, and the H3N8 and H7N7 subtypes have caused outbreaks in horses.

The Role of Swine in the Emergence of New Influenza Viruses

Generally, influenza viruses are host specific, and viruses from one host rarely establish stable lineages in another host species. Although whole viruses may rarely transmit, gene segments can cross the species barrier through the process of genetic reassortment. Pigs have been postulated to play an important role in the process of genetic reassortment by acting as the “mixing vessel” for such events. Pigs, unlike humans, seem to be readily infected by avian viruses, and most, if not all, avian HA subtypes are capable of replicating in swine. Researchers have proposed a molecular mechanism for the susceptibility of swine to avian virus infection. Viral receptors called sialyloligosaccharides, which are present on the pig tracheal cells, possess the ability to bind to both types of viruses, with human viruses preferentially binding in one location and manner and avian viruses preferentially binding in another location and manner. Thus, pig tracheal cells can be infected not only by human influenza viruses but also by avian viruses. However, the direct chicken-to-human transmission of H5N1 viruses, observed during the 1997 flu outbreak in Hong Kong, argues that factors in addition to receptor specificity must be involved in influenza interspecies transmission.

Influenza in swine is an acute respiratory disease, the severity of which depends on many factors, including pig age, virus strain, and secondary infections. Currently, three main subtypes of influenza virus are circulating in different swine populations throughout the world: H1N1, H3N2, and H1N2.

In North America, Asia, and much of Europe, viruses of the H1N1 subtype are the most commonly isolated. The circulating H1N1 viruses differ, however, in the origins of their genomic components. The H1N1 viruses in North America and Asia belong to the classical swine lineage, which is genetically related to human H1N1 viruses responsible for the 1918 Spanish influenza pandemic. In contrast, all eight genes of the H1N1 virus circulating in Europe are phylogenetically related to the avian lineage that entered pigs in about 1979. The avian-like H1N1 virus also is present in the United Kingdom, although the virus of current concern is a reassortant H1N2 virus with gene segments derived from both human and avian lineages.

Viruses of the H3N2 subtype circulate in Asia and Europe but have been infrequently isolated in North America. The most recent outbreak of this subtype in North America occurred in 1998, when a severe influenzalike illness was observed in pigs on a farm in North Carolina. Additional outbreaks among swine herds soon occurred in Minnesota, Iowa, and Texas. Genetic analysis of the viruses showed that at least two different genotypes were present. The initial North Carolina isolate contained gene segments similar to those of the human (HA, NA, PB1) and classical swine (NS, NP, M, PB2, PA) lineages, whereas the isolates from the other states contained genes from the human (HA, NA, PB1), swine (NS, NP, M), and avian (PB2, PA) lineages. Serological surveillance indicates that the latter triple reassortant virus has spread throughout the pig population of United States.

Examples of Recent Outbreaks in Birds

The avian H5N1 influenza virus that was transmitted to poultry and humans in 1997 in Hong Kong caused high mortality in both species, killing more than 70 percent of chickens and six of the 18 infected humans. (Hong Kong's location at the crossroads of many trade routes makes it particularly susceptible to the outbreak of new diseases.) Surveillance studies revealed that two antigenically and genetically distinguishable variants of H5N1 were circulating among avians and humans. There was no correlation between lethality in humans and one or other of the variants. The slaughter of approximately 1.6 million chickens during a 2-day period stopped the further spread of the virus to humans. The failure of H5N1 to transmit from human to human, and the slaughter of poultry in a matter of days before another strain of influenza (H3N2) began circulating in humans in Hong Kong, probably prevented the generation of reassortants with pandemic potential.

Also in 1997, the avian H9N2 virus struck the live poultry market in Hong Kong. In 1998–99, observers reported that the virus had transmitted to humans and pigs. The initial report of five human cases in southern China was confirmed by the isolation of H9N2 viruses from two children in March 1999 in Hong Kong. The children had typical influenza and recovered. The isolates were genetically similar to H9N2 isolates found in quail. Further characterization of the viruses revealed that the human isolate from Hong Kong and the quail isolates shared similar genetic traits with the H5N1-like viruses from chickens and humans in Hong Kong in 1997. Thus, while avian influenza viruses can transmit directly to humans and cause disease, additional mutations and/or reassortant events are probably required to permit efficient human-to-human spread.

The fact that these viral strains can transmit to and cause respiratory disease in humans confirms that the surface glycoproteins can fulfill their primary functions in mammals. Indeed, several lines of genetic evidence suggest that some of these strains have a special propensity for interspecies transmission. The continued circulation of such viruses in poultry, especially in quail, alerts us to the continuing need for active surveillance for these viruses in humans and pigs in this region.

Recent Advances in Understanding Influenza Viruses

A major advance in the ability to manipulate the genome of influenza viruses occurred in 1990, when scientists established a “reverse genetic” system, which permits the generation of influenza viruses containing genes derived from cloned DNA, or cDNA. A further major advance occurred in 1999 when other researchers demonstrated the generation of influenza A viruses entirely from cloned cDNA with high efficiency. These advances permit complete manipulation of all genes of influenza viruses, which means that it is now possible to tailor-make future live attenuated vaccine strains and to define all of the functional domains in the viral genes and their interaction with the host. Thus, resolution of the molecular basis of pathogenesis will be possible in the near future, and the domains responsible for interspecies transmission and ability to spread in new hosts will eventually be known. With this information in hand, it may be possible to predict which influenza viruses have pandemic potential in humans.

Another potential advantage is that it will be possible to resolve the question of the pathogenicity of the 1918 Spanish influenza. When the total genetic sequence of the 1918 virus is obtained, scientists will be able to recreate the virus. While of great scientific value, this possibility also raises considerable concern; such a study should be done only if the benefits warrant the risk and only if all work is performed in high-level biosafety laboratories. Once researchers have made the virus, tested its ability to interact with cultured cells, and determined which host genes are turned on or turned off, it will then be necessary to study the virus in an animal model, perhaps the mouse or the minipig. Before this happens, however, the research community and society must fully consider the ethics and safety of doing these experiments.

Preparing Our Defenses

Along with advancing our scientific knowledge, we also must improve our ability to detect and respond to new emerging strains of influenza virus, particularly those that appear suddenly and are capable of spreading over large areas. Many countries have prepared plans to cope with the next pandemic, which is considered imminent. Such documents must be updated as new information becomes available.

The cornerstone of pandemic preparedness is surveillance, both of humans and lower animals and birds, for if we develop the ability to predict which combinations of genes have pandemic potential, we must then maintain active surveillance to detect them. Viral surveillance will continue to be the key to providing time for the preparation of vaccines ahead of worldwide spread. In the interim between detection of a pandemic and vaccine availability, it will be essential to have adequate supplies of antiviral drugs, which means that urgent attention should be given to ensuring strategic stockpiles.

Overall, however, the reality is that we are not well prepared to cope with a pandemic, even a moderately severe one. We have identified where our weaknesses are, but we have not brought resources to bear on their solution.

ASSESSING THE THREAT AND THE OPPORTUNITIES ACROSS THE SPECTRUM OF ZOONOTIC DISEASES

, Ph.D.

Professor, Department of Biology, Amherst College

Attention to emerging diseases has focused largely on acute infectious diseases that have caused outbreaks after entering humans from other host species. Most of these diseases have very limited potential for spreading in human populations, particularly in wealthy countries. Of far greater potential importance to public health is discovery of infectious causes of the widespread and damaging chronic diseases.

The logic leading to this conclusion emphasizes that the global emergence of highly virulent, acute infectious diseases requires special sets of conditions that are rarely met. Vectorborne diseases, for example, may persist evolutionarily in a highly virulent form, but only a very small proportion of vectorborne pathogens have the characteristics necessary to be transmitted persistently from person to vector to person. None of these pathogens have demonstrated this ability under conditions found in modern wealthy countries, characterized by screened houses, air conditioning, and primarily indoor life. Similarly, the newly emergent pathogens that cause deadly acute infections and are transmitted from person to person by air would need to have characteristics that enable them to be transmitted readily from sick hosts, particularly durability in the external environment, if they are to maintain transmission cycles. The smallpox virus and tuberculosis bacterium have these characteristics. The pathogens that have attracted the most public attention in recent years, such as Ebola and hanta viruses, do not.

Evolutionary principles, current evidence, and the recent track record of recognizing infectious causation indicate that many if not most of the damaging chronic diseases are caused by infection. Some of the candidate pathogens use humans as primary hosts. Others infect humans zoonotically. In contrast to the oft-mentioned examples of newly emerging acute infectious diseases, the most damaging chronic diseases are already globally distributed and prevalent. Some of these diseases are now causing damage in human populations that is comparable to the damage that is merely feared for emerging acute infectious diseases. Atherosclerosis is the most damaging chronic disease in this category, accounting for about half of the deaths in wealthy countries. Other such diseases include schizophrenia, bipolar disorder, Alzheimer's disease, diabetes, and breast cancer. Considering the damage, prevalence, and persistence of these diseases in human populations, investment of intellectual and economic resources in the investigation of infectious causation of these illnesses may be more beneficial than investments in efforts to monitor, study, and control the spread of the acute infectious diseases that make for sensational headlines but pose a relatively small global threat to human populations.

These investments may provide particularly great health benefits because the most damaging manifestations of infection tend to occur when the infection in humans is no longer transmissible. Consequently, strategic use of antibiotics may enable the causative agents to be controlled indefinitely without the evolution of antibiotic resistance. Recent evidence suggests, for example, that schizophrenia may be caused by infection with Toxoplasma gondii, which is transmitted in its natural cycle between cats and rodents. T. gondii damages the mental health of rodents in ways that facilitate capture of the rodents by cats, and hence its transmission to cats. Because T. gondii is not transmissible from humans, evolution of resistance should be negligible if an antitoxoplasmal drug is used only for human infection.

This strategy for controlling antibiotic resistance, referred to as “dead-ending,” requires that at least two and preferably more than two antibiotics be available, so that one antibiotic can be used for the human infection and another can be used for the infection in the reservoir host. This distinction is particularly apparent for T. gondii, because cats are domestic animals that receive extensive medical treatment. When the natural hosts are unlikely to be the target of antibiotic treatment, as with the Lyme disease spirochaete Borrelia burgdorferi, the principles of antibiotic dead-ending are less relevant.

The principles of dead-ending apply to vaccine use as well. Although evolutionary escape from vaccines has been documented only for a few pathogens, we can expect that the continued use of vaccines and generation of new vaccines will lead to more examples, particularly when pathogens are prone to genetic variation through high rates of mutation or genetic recombination. To reduce this danger, vaccines for humans should be antigenically different from vaccines generated for reservoir hosts. If, for example, T. gondii becomes widely recognized as a cause of schizophrenia and other damaging diseases, then the demand for a vaccine against T. gondii will increase for humans and for cats (to prevent infections in humans). If the same vaccine is used for both people and cats, then the use in cats would create a cumulative selective pressure favoring vaccine escape. The use in humans, however, would not generate a cumulative selective pressure because humans are dead-end hosts. To preserve the effectiveness of the best vaccine for humans, an antigenically different vaccine needs to be generated for cats.

Developing different antibiotics or vaccines for humans and reservoir hosts is not as formidable as it may seem, because the constraints are not as severe for nonhuman recipients as they are for humans. A higher frequency of adverse reactions would be more acceptable for nonhuman hosts than for humans, as would certain kinds of adverse reactions. Neuronal damage sufficient to reduce a person's IQ by 5 percent, for example, would not be acceptable for humans but probably would be acceptable for cats.

If the reservoir host is not a valued animal, then discovery of zoonotic causes of already common chronic diseases may provide other less technologically sophisticated options. For example, one of the major candidates for breast cancer is mouse mammary tumor virus (MMTV) or a closely related virus. The normal host for MMTV is the house mouse, Mus domesticus, in which the virus causes mammary tumors. This virus has been found much more frequently in tissue from human breast cancer than in surrounding healthy tissue, and breast cancer is associated geographically with the distribution of M. domesticus. Details of transmission to humans are unknown. If each human infection is directly acquired from M. domesticus, then local extermination of M. domesticus may directly protect humans from breast cancer. If transmission occurs from person to person with occasional reseeding from M. domesticus (analogously to the transmission of yellow fever virus or Yersinia pestis), then extermination may indirectly and diffusely protect the human population.

Lyme disease is an example of a zoonotic agent that is already recognized as a cause of chronic disease, but the spectrum of chronic illnesses caused by B. burgdorferi appears to be broadening. Evidence indicates, for example, that B. burgdorferi is responsible for some cases of chronic diseases that have been diagnosed as multiple sclerosis, motor neuron disease, arthritis, paralysis, or myocarditis. Indeed, B. burgdorferi has been referred to as “The Great New Imitator” because of its potential involvement in chronic diseases that have been previously categorized as other diseases.

This use of the term “imitator,” however, illustrates how an overarching trend that has been occurring in studies of chronic diseases may be inadvertently obscured. During the past half-century, a steadily increasing number of chronic diseases have been accepted as infectious. When a portion of a disease category is so recognized, that portion is typically given a new name to distinguish it from the rest of the category (e.g., reactive arthritis and neuroborelliosis), but this experience has not been used prospectively as a model for allocating research effort. Doing so would involve searching for the agents that will permit a subdivision of each of the umbrella categories, such as multiple sclerosis, schizophrenia, motor neuron disease, chronic fatigue syndrome, obsessive compulsive disorder, atherosclerosis, stroke, and Alzheimer's disease. Just as we now consider such diseases as hepatitis and pneumonia to be collections of different diseases with distinct infectious etiologies, we can expect a variety of infectious etiologies for each of these umbrella diseases. If we instead search for the infectious cause of an umbrella disease, then we risk being misled by studies that do not find an association with a particular agent because that agent is rare in the study area. The ongoing resolution of hepatitis and arthritis illustrates how diverse the infectious causation of a chronic disease can be. The potential applicability to other highly damaging chronic diseases is apparent from the current evidence on infectious causation of diseases for which causation is still controversial. Atherosclerosis, for example, is associated with infections by Chlamydia pneumoniae, Porphyromonas gingivalis, Actinobacillus actinomycetocomitans, Bacillus forsythus, and cytomegalovirus, with each of these pathogens being found in the atherosclerotic plaques, and some being shown to cause atheromas in animal models. Sporadic Alzheimer's disease has been similarly linked to C. pneumoniae and human herpes simplex virus type 1.

Much of the controversy over infectious causation stems from discrepancies between research teams that are unable to replicate associations with their own versions of the assays. This problem may occur because research will tend to achieve consensus most readily for those infectious diseases that are detectable even when detection techniques vary greatly, leaving in their wake those infectious diseases that are detectable only with very specific versions of experimental protocols.

The candidate pathogens for atherosclerosis and Alzheimer's disease are regularly transmitted between humans. The ambiguities due to discrepancies between research teams may be even greater for diseases that are sometimes caused by zoonotic agents, because these agents are less likely to be detected in humans where the zoonotic reservoirs are absent. Thus, if breast cancers are caused in part by MMTVs that are transmitted directly to humans from M. domesticus, then studies might be confirmatory in New York, where M. domesticus is present, but not in Japan, where M. domesticus is absent. In Japan, another pathogen, such as Epstein Barr virus, might be playing a relatively more important role. Similarly, geographic variation in pathogens might help explain why studies have found the zoonotic borna disease virus to be associated with schizophrenia in Japan, but not in other areas, where T. gondii, human herpes simplex virus type 2, and an endogenous retrovirus have been associated with schizophrenia.

The proposed importance of infectious causation of chronic diseases emphasizes an irony in the attention devoted to emerging infectious diseases over the past two decades. This attention was triggered largely by the AIDS experience, in which a lethal disease arose from an exotic source and spread pandemically. This experience gave credence to concerns that other exotic diseases might similarly emerge. Concern focused on the most conspicuous examples—acute infectious diseases, such as Ebola, lassa fever, and hanta disease. But AIDS is a chronic disease syndrome. The irony, therefore, is that the alarm bell was rung in response to an emerging chronic disease syndrome, yet most of the subsequent attention has been devoted to emerging acute infectious diseases. The past two decades have not generated examples of new globally spreading acute infectious diseases that have been highly damaging to human populations; nor was there such an example for the entire 20th century. Resurgences of long-recognized global threats, such as influenza, have occurred, but concern over such resurgences was present before the recent interest in emerging infectious diseases.

This recent history therefore suggests that concern over the future threat of emerging diseases needs redirection. In poor countries, the resurgence of long-recognized acute infectious diseases represents a grave danger. In both poor and rich countries, grave dangers are posed by the long-standing chronic diseases that are or may soon be recognized as caused by infection. Leaders of the effort to awaken concern over infectious diseases have emphasized the danger from resurgence of known acute infectious diseases, and to some extent the growing recognition of infectious causation of chronic diseases, but most of the media attention has focused on the exotic acute diseases. A broader emphasis on studies of infectious causation of chronic diseases and the distribution of current knowledge about these diseases may be needed to direct the attention of researchers, policy makers, and the public to support efforts to identify and reduce the greatest threats to human health.

PRACTICES AND POLICIES TO PROTECT HUMAN HEALTH FROM ANTIBIOTIC-RESISTANT PATHOGENS

, D.V.M., Ph.D.

Director, Center for Veterinary Medicine, Food and Drug Administration, U.S. Department of Health and Human Services

Antimicrobial resistance is one of the highest-priority issues facing the Food and Drug Administration's (FDA) Center for Veterinary Medicine (CVM). The evidence of harm to the public from certain uses of antimicrobials in food-producing animals continues to grow, so CVM is taking steps to address the problem. We hope to deal with this issue through regulatory changes in the way we manage and approve drugs; through improved monitoring systems; through better risk assessment, which we think is a critical area of need; and through education.

The issue of antimicrobial resistance has been around for some 30 years. The National Academy of Sciences, through the National Research Council (NRC) and the Institute of Medicine (IOM), has taken up this issue and provided input to FDA and the public. The NRC's first report, which basically was the first risk assessment, was released in 1980. The NRC was asked to address the issue of whether subtherapeutic use of antimicrobials in feed for food animals was a potential hazard to human health. The report concluded that existing data neither proved nor disproved the potential hazards.

That study was followed in 1998 by an IOM report titled Human Health Risks with Subtherapeutic Use of Penicillin and Tetracycline in Animal Feeds. The report concluded that the study committee was unable to define a substantial body of direct evidence that established a definite human health hazard from the subtherapeutic use of these drugs in food animal feed. However, the committee did find considerable indirect evidence of that human health hazard. Another IOM report in 1998, on the benefits and risks of using drugs in food animals, reached a stronger conclusion: “There is a link between the use of antibiotics in food animals, the development of resistant microbes, and the zoonotic spread of pathogens to humans.”

Because of the mounting evidence of risk to humans, FDA/CVM believes that there are issues we must address. From a regulatory standpoint, however, an issue as complex as antimicrobial resistance presents a tremendous challenge.

To begin the process, we published in November 1998 a document titled Guidance for Industry #78. This document affirmed FDA's position that it has the authority to regulate not just the toxic effects of drug residues, which has been FDA's traditional role, but also the microbial effects from the antibiotics and antimicrobials that FDA regulates. In the document, we asked that two types of questions be answered for approval of a drug. First, we asked for information regarding the quantity of resistance that would be created from the use of an antimicrobial in food animals. We want to know which organisms are affected, how much resistance is likely to be created, and at what rate would resistance be likely to develop. Second, we asked for information about the change in animal enteric bacteria that are human pathogens that would come from the use of the drug. From a scientific standpoint, the problem we face is that we have a very limited ability to predict the rate and extent of antimicrobial resistance that could result from the use of an animal drug.

A month later, in December 1998, CVM issued a second document, Microbial Effects of Antimicrobial New Animal Drugs Intended for Use in Food-Producing Animals, also known simply as the “Framework Document.” (The document is available on CVM's web site at www.fda.gov/cvm.) This document lays out a conceptual, risk-based approach for regulating antimicrobial drugs so that resistance is minimized. The primary public health goal is to ensure that significant human antimicrobial therapies are not lost due to antimicrobial use in food-producing animals. The framework document was not intended to be regulation. Instead, it was an attempt to lay out what we considered to be a rational approach to dealing with this issue and to do so in a way that would be informative for a variety of stakeholders, including the animal drug industry, animal producers, scientists, and the general public.

What we came up with was a fairly straightforward, risk-based approach to dealing with the regulatory issues of antimicrobial resistance. First, we thought in terms of assessing the risk. We want to know how important a particular antimicrobial is for human medicine. The greater the importance, the less risk we would be willing to accept. In addition, we want to know what would be the likelihood that humans would be exposed to resistant pathogens as a result of the drug's use for food animals. This is a typical exposure-times-hazard type of risk analysis.

The framework document considers five main issues:

Drug categorization

Drug categorization is determining how important these drugs are in human medicine. As a basis for our risk analysis, we proposed to categorize antimicrobials according to their importance to human medical therapy. This determination had to come before we could determine what kind of regulatory approach we should take, so that the regulatory burden would be commensurate with the risk.

We proposed three categories. In Category 1 would be drugs of greatest importance to human medicine. These are drugs that would be essential for treating serious or life-threatening diseases in humans where there are no, or very few, satisfactory therapeutic alternatives. Simply put, these are the drugs of last resort with a life-threatening disease. We would put drugs in Category 1 particularly if they were important for the treatment of foodborne disease, because many zoonotic diseases are foodborne in nature. We also would put drugs in this category when resistance to alternative antimicrobial therapy may limit therapeutic success. Category 2 drugs would include those that are considered drugs of choice or that are important for the treatment of potentially serious disease, whether foodborne or otherwise, but satisfactory alternative therapy exists. Category 3 drugs would include those with very little or no use in human medicine.

Preapproval studies

We are trying to determine what kind of information drug sponsors should provide that will have some kind of predictive value regarding the drug's ultimate effects. We want the preapproval studies to indicate how rapidly resistance will emerge in pathogens of concern and to what degree. Preapproval studies will need to consider the proposed use of the drug in question. Is it going to be used in an entire herd or flock of animals or just for individual animals? How is the drug going to be used? How is it going to be administered? Once we have answers, we can design studies that will give us some predictive value. We also are considering asking for studies on pathogen load. After a drug is administered to an animal, is there a rebound effect in the number of pathogenic organisms that are found in the animal's intestinal tract, and, if so, what can be done to minimize the food safety risk when those animals go to slaughter?

Thresholds

Before a new drug is used under field conditions, we cannot know in advance exactly which microbes are going to develop resistance, how fast the resistance will develop, and what the overall impact is going to be. But it is possible to describe the events that would make us concerned. CVM has outlined a possible approach for establishing thresholds to stimulate discussion. This approach would attempt to link an unacceptable level of human health impact to a level of resistance in animals. CVM acknowledges the complexity of establishing such thresholds and continues to seek further scientific input.

Monitoring

Perhaps the main reason the issue of resistance has been studied for so long without any resolution is that we have never had a good system for collecting data; that is, we lacked good surveillance and monitoring systems in the field. For years all such reports were based on anecdotal information. We felt that the only way to really get on top of this problem was to put a system in place to start tracking resistance. So, in 1996, we started the National Antimicrobial Resistance Monitoring System (NARMS). This system uses information that already was being collected through two other federal programs. The first is the Centers for Disease Control and Prevention (CDC)'s FoodNet system, which gives us a random sampling of food-borne diseases that occur in humans around the country. The second is the Department of Agriculture's (USDA) Food Safety Inspection Service (FSIS) program for collecting slaughterhouse samples, obtained under USDA's Hazard Analysis and Critical Control Point program, which gives us information about pathogens in food from animals.

With NARMS, we know both the incidence of resistant foodborne infections in humans and in animals. Thus, we can look to see if there is a correlation between those two sets of information to spot increases in resistance that may be related to the use of antimicrobials in food animals. We also can determine whether attempts to mitigate the rise in resistance have a positive or negative effect on the level of resistance.

Drug use information

We want information about how drugs are being used—in which species of animals they are being used, under what conditions they are being used, and how much of the drugs are being used—in order to be able to relate drug use to the development of resistance. From a regulatory point of view, this is an important step, because we have to justify, on a scientific basis, why we might be taking an adverse regulatory action against a product. Having that chain of evidence in place is absolutely critical for us to be able to sustain our decision. Therefore, we now are developing guidance and regulations on new requirements for companies to report to FDA, on a yearly basis, information on drug sales and use. The information we will require includes the amounts of antimicrobial agents used in each food animal species, the routes of administration, and the claim made for their use.

After a year of review, FDA publicly presented the framework document in December 1999, and we asked for additional comments. Among the approximately 40 comments received, several common themes emerged. First, whatever approach FDA takes, it should be well founded in science. There was some concern that antimicrobial resistance is a politically hot issue, and that FDA therefore may feel pressure to do things that are not well based in science. We were cautioned against that. Second, any regulatory actions should follow an extensive quantitative risk assessment. There was concern that FDA had not done an adequate job of determining how much risk is involved with the use of these products. We were advised to undertake such risk assessments.

In FDA's first major attempt to conduct a quantitative risk assessment for microbial resistance, we focused on the use of fluoroquinolones in poultry and its effects on Campylobacter. Fluoroquinolones are administered to poultry in their drinking water; the drug is used to treat entire flocks, because it is impractical to treat individual birds. The model we developed is designed to directly assess the impact on human health that occurs from the use of this drug in chickens. The model determines the illness that results in humans from drug-resistant Campylobacter infections attributable specifically to fluoroquinolone use in chickens, and it relates the prevalence of resistant Campylobacter infections due to use of the drug in chickens to the prevalence of Campylobacter in chickens. In this way, we can predict what to expect in humans when we see a certain amount of resistance in Campylobacter in chickens.

We began with this model, in part, because we expected the relationship between fluoroquinolone use and development of resistance in Campylobacter to be fairly straightforward. In other words, we wanted to tackle an easier problem first, before moving on to those expected to be scientifically more complex. (Indeed, before this project, we tested several more complex assessment models but found that the uncertainty surrounding the assumptions was so great that the result was a finding of risk somewhere between zero and one—an outcome not at all helpful.) Based on this experience, we believe that our risk assessment model should work fairly well for most enteric pathogens. We are now conducting another risk assessment of the new drug Synercid—a more daunting challenge, since it involves the indirect transfer of resistance from animal enterococcal organisms to human enterococcal organisms.

ECOLOGICAL SOURCES OF ZOONOTIC DISEASES

, M.D.

Professor, Department of Pathology, University of Texas Medical Branch

Ecology is the branch of biology that deals with the interrelationships between organisms and their environment. Ecological factors play an extremely important role in the epidemiology of zoonotic diseases, since these infections are usually environmentally acquired, and most emerging zoonotic disease outbreaks result from ecological changes. Because of their complex ecology, zoonotic agents are difficult to study in their natural habitat; consequently, most zoonotic disease research now focuses on the pathogenesis of the microbes or on their biochemical and genetic characteristics. This focus on the microbe has led to a myopic view of zoonotic disease. Despite spectacular achievements in microbial genetics and genomics, we still do not really understand how most zoonotic agents are maintained in nature or how they respond to environmental (usually anthropogenic) changes, nor do we understand the precise ecological factors that lead to human infection and emergence. Consequently, textbook descriptions of the epidemiology of most zoonotic diseases are at best simplistic. If we want to prevent or control zoonotic diseases, we first must better understand the ecology of their respective etiologic agents.

The Institute of Medicine's 1992 report on emerging infectious diseases identified six factors that shape their emergence: human demographics and behavior, technology and industry, economic development and land use, international travel and commerce, microbial adaptation and change, and breakdown of public health measures. Within this framework, consider the following three emerging zoonotic diseases, which illustrate the complex interaction of various ecologic factors on disease emergence and how little we really understand about the basic ecology of most zoonotic disease agents.

Venezuelan Hemorrhagic Fever

In 1989, physicians in central Venezuela began to report cases of a severe hemorrhagic illness that was initially thought to be dengue hemorrhagic fever. Subsequent clinical and epidemiologic studies demonstrated that this was a new disease, which was given the name Venezuelan hemorrhagic fever (VHF), and that the etiologic agent, designated Guanarito virus, was a novel member of the Tacaribe complex of New World arenaviruses. VHF is sporadic and cyclic in occurrence and quite localized in distribution. To date, approximately 250 human cases have been reported; the mortality rate has been about 30 percent. VHF affects mainly male agricultural workers; it appears to be restricted to rural areas in two states in the central plains of Venezuela. The majority of cases have occurred during the dry season, when there is considerable agricultural activity in the endemic region. VHF has many clinical and epidemiological similarities with the other arenaviral hemorrhagic fevers (Lassa fever, Argentine hemorrhagic fever, and Bolivian hemorrhagic fever), so it is assumed that humans generally acquire the disease by inhalation of virus in aerosols of infected rodent excreta.

Epidemiologic studies in the VHF-endemic region have implicated the cane mouse, Zygodontomys brevicauda, as the principal reservoir host of Guanarito virus and the probable source of the virus to humans. Cane mice experimentally infected with Guanarito virus develop a persistent non-immunizing infection and chronically shed infectious virus in their urine and saliva. Z. brevicauda is a grassland species and reaches high densities in fallow fields and abandoned pastures, as well as in tall grass along fence lines.

The broad geographic distribution of the various members of the Arenaviridae, their marked genetic diversity, and their intimate association with specific rodent species suggest that arenaviruses are ancient viruses that have coevolved or cospeciated with their rodent hosts. Thus, it is unlikely that Guanarito is a new virus or even that VHF is a new disease in humans. If so, then why was VHF not recognized until 1989?

The endemic region of Guanarito virus and VHF was once largely covered with forest, but during the past 50 years a significant part of the original forest has been cut down to create agricultural land. Much of the deforested land is now used for pasture or cultivation. The transformation from forest to agricultural land has created a highly favorable environment for grassland rodents, such as Z. brevicauda, and their population densities are probably much higher now than before. The economic prosperity brought by agriculture in turn has attracted many human migrants into the region. The resulting ecological changes have brought more people into contact with more infected cane mice, resulting in the emergence of VHF.

Observers noted a similar scenario in the 1950s with the emergence of Argentine hemorrhagic fever (AHF). It has been suggested that conversion of the Argentine prairie to cropland for corn and wheat production favored vesper mice (Calomys laucha and C. musculinus), the natural reservoirs of Junin virus, the etiologic agent of AHF. These two rodent species are common in terrain disturbed by humans. Thus, in both VHF and AHF, changes in land use and human demographics appear to have created conditions favorable for the emergence of new zoonoses.

Yellow Fever

Yellow fever (YF) is a severe viral hemorrhagic fever that is transmitted by the bite of infected mosquitoes. The current geographic distribution of the disease includes tropical regions of sub-Saharan Africa and South America. YF exists in two forms: an endemic or sylvan cycle, thought to involve monkeys and certain forest mosquitoes that breed in tree holes, and an epidemic or urban cycle that involves humans and the domestic mosquito Aedes aegypti. During the 17th, 18th, and 19th centuries, YF caused major urban epidemics in Africa, European port cities, and the Americas. YF, and fear of it, played an important role in the settlement and commerce of the New World during that period. For example, in 1878 a massive YF epidemic devastated the lower Mississippi Valley; it is estimated that more than 100,000 people were infected and approximately 20,000 people died. The resulting panic and interruption of commerce caused major financial losses.

Because of the havoc and high mortality resulting from urban YF epidemics and the earlier success of American efforts at controlling the disease in Cuba and Panama by sanitation and vector control, Brazil embarked on a nationwide campaign to eradicate Ae. aegypti during the 1930s. The early success of that program, coupled with the introduction of the pesticide DDT, led the Pan American Sanitary Bureau to initiate a hemisphere-wide Ae. aegypti eradication campaign in 1947. By 1972, this mosquito had been eliminated from Central America and most of South America, and YF disappeared in urban areas.

The development of YF vaccine also played an important role in elimination of the urban disease from the Americas. The 17D YF vaccine was first used in Brazil in 1937, and during the next 30 years many millions of persons throughout the Americas received the vaccine. However, following the eradication of Ae. aegypti and the disappearance of urban YF, many countries in South America stopped mass vaccination campaigns. The current policy in most countries in the region is to wait for the appearance of sylvan cases and then to vaccinate people living in areas around the outbreaks in order to prevent spread of the disease to urban areas.

Ironically, during the past 20 years, as many South American countries discontinued or deemphasized YF vaccination campaigns, Ae. aegypti, the urban YF vector, was reinfesting the same countries. The mosquito now occupies almost its entire preeradication geographic distribution, including most cities and towns in the Amazon Basin. The reemergence of dengue and dengue hemorrhagic fever in South America during the same period attests to the current widespread distribution and abundance of Ae. aegypti in the hemisphere. Recent dengue epidemics in the cities of Belem and Manaus in northern Brazil, in Iquitos and Pulcalpa in eastern Peru, and in Santa Cruz in eastern Bolivia demonstrate that Ae. aegypti has reinfested many urban communities in the Amazon Basin and now exists in close proximity to areas where sylvan YF occurs. Consequently, public health officials are deeply concerned that urban YF will reemerge in the Americas.

Sylvan YF is still endemic in tropical forested regions of South America. From 1969 to 1995, the Pan American Health Organization reported a total of 3,924 cases of sylvan YF in the hemisphere. This undoubtedly is an underestimate, since milder cases of the disease often are not recognized and the more severe forms can be confused with other endemic diseases, such as hepatitis B and D, leptospirosis, and dengue hemorrhagic fever.

Current knowledge about the ecology of sylvan YF in South America is still incomplete. The most widely accepted view is that YF virus moves in “epizootic waves” through the Amazon Basin in a sylvan cycle involving principally monkeys and arboreal mosquitoes of the genera Haemogogus and Sabethes. Human exposure to YF virus is strongly linked to occupational activities, such as forest clearing, lumbering, road construction, and jungle military maneuvers, that bring people into contact with the sylvan vectors. Consequently, most cases of sylvan YF occur in adult males.

Two recent outbreaks of YF, in Peru and Bolivia, illustrate some of the factors contributing to the emergence of sylvan YF and the increasing risk of urban outbreaks. During the late 1980s and early 1990s, terrorist activities of the Shining Path guerillas in eastern Peru drove many civilians out of the region. Following the defeat of the Shining Path, people moved back into the region to reclaim their land. As part of this migration, the government brought in tens of thousands of contract laborers from the impoverished highlands to help clear land and plant crops in rainforests in the eastern Andean foothills. Thus, a large number of people who lacked immunity to the disease moved into an area where YF virus was endemic. In 1995, about 500 cases of YF were reported from Peru, many of them among these contract workers. Since the region is quite remote and isolated, few of the patients were hospitalized. However, if the outbreak had occurred among a more mobile group, such as soldiers, mineral prospectors, or coca producers, some of the cases undoubtedly would have been evacuated by air and transported to urban areas for treatment. In that scenario, the risk of YF introduction into an urban area with Ae. aegypti would have been much greater.

A second and potentially more dangerous YF outbreak occurred in 1997–98 in Santa Cruz, a city in the tropical lowlands of southeastern Bolivia. In recent years, Santa Cruz, which has a population of about 1 million people, has become quite prosperous because of large deposits of natural gas, rich agricultural land, and its proximity to major urban centers in Brazil and Argentina. Dengue is endemic in Santa Cruz; it appeared after the reintroduction of Ae. aegypti in 1980. Sylvan yellow fever also is endemic in rural areas around the city.

During the past decade, the Bolivian government privatized its mines in the country's highlands; this led to many unprofitable mines being closed and to massive unemployment and economic hardship in a region where most of the country's population lives. Many unemployed miners and their families migrated into the area around Santa Cruz because of the economic prosperity and the availability of agricultural land and jobs. Most of the migrants, as well as a majority of the residents of Santa Cruz, lacked immunity to yellow fever virus. Many of the migrants established shanty towns around the city's periphery. Initially, such communities have poor sanitation and lack running water; thus, people by necessity store water in their houses and discard refuse in their yards, thereby creating multiple breeding sites for Ae. aegypti. Some of these urban migrants work as day laborers on farms and plantations near the city, where they have potential exposure to YF virus.

Between December 1997 and June 1998, six cases of yellow fever were confirmed in Santa Cruz; five of the cases were fatal. Five of the patients lived in the southern sector of the city in areas of substandard housing. Follow-up epidemiologic investigations indicated that several of the patients had had recent exposure in rural areas where YF is known to be endemic. But two of the patients reportedly had not left the city during the incubation period, suggesting urban transmission of YF. Emergency vaccination of persons living in close proximity to the YF cases and intensified vector control were instituted, and no further cases were reported.

In these two examples of recent YF outbreaks, human demographics and migration, economic change and land use, and the breakdown of public health measures all contributed to reemergence of the disease.

Zoonotic Visceral Leishmaniasis

From a worldwide public health perspective, zoonotic visceral leishmaniasis (ZVL) is one of the most important emerging parasitic diseases. ZVL is endemic in rural areas of tropical America, southern Europe, north Africa, sub-Saharan regions of east and west Africa, southwest Asia, and China. The disease typically affects young children and immunocompromised adults. In the New World, ZVL is caused by Leishmania chagasi, which is transmitted by the sand fly Lutzomyia longipalpis. In the Old World, the disease is caused by L. infantum, which is transmitted by sand flies of the genus Phlebotomus, subgenus Larroussius. (Although the parasites causing ZVL in the New and Old Worlds still carry different names, many parasitologists now believe that they are in fact the same organism.) Dogs serve as the principal domestic reservoir of both parasites, while wild canids, including foxes, jackals, and wolves, serve as the major sylvan reservoirs.

In most locations where ZVL occurs, the essential maintenance cycle of the parasite in nature is presumed to involve a transmission cycle between wild canids (including feral and stray dogs) and sand flies living in caves and rock crevices. The parasite is introduced into the domestic cycle when infected wild animals visit houses to scavenge for food. During such visits, so-called “peridomestic” sand flies feed on the infected wild animals, pick up the parasite, and subsequently transmit it to dogs, which then act as domestic reservoirs. Once introduced into a community, the parasite can be maintained in a dog–insect–dog transmission cycle. Occasionally, some humans are infected directly by being bitten by sand flies, but humans are not thought to play a significant role in maintenance of the parasite. Introduction of the parasite into new regions (including occasionally the United States) occurs when infected dogs are transported from endemic areas to nonendemic areas.

Several factors are contributing to the rise of ZVL as a major public health problem. First, and probably the most important, are land use and demographic changes. This is most apparent in tropical Latin America, where massive destruction of primary forests, together with rapid human population growth and the concomitant development of new farmland and rural settlements, have led to conditions that now support large populations of the vector Lu. longipalpis, as well as large populations of the major reservoir hosts, dogs and foxes. As a consequence, ZVL now occurs in many regions of Latin America where it was not found previously.

In addition, ZVL recently has begun to appear in suburban areas of several major Brazilian cities, including Rio de Janeiro, where dogs alone now seem to be the major reservoir of the parasite. In most Latin American countries during the past 30 years, there has been a major migration of people from rural to urban areas. As noted in the case of Santa Cruz, new migrants typically settle in hastily constructed shanty towns on the periphery of large cities. These settlements are usually overcrowded, with inadequate housing and poor sanitation. The migrants often bring with them dogs, chickens, and pigs that they keep in or around their houses. These conditions create an excellent habitat for Lu. longipalpis, and the density of this insect in both houses and animal shelters may reach very high levels in such communities.

Another factor in the recent emergence of ZVL has been the elimination in many regions of public campaigns to control malaria by spraying houses with insecticides to kill mosquitoes. One of the side benefits of such spraying was a reduction in the number of sand flies present in a community. In most areas of the Mediterranean and Latin America, where active spray campaigns were in place, the incidence of ZVL decreased. However, as malaria control programs were discontinued, the number of ZVL cases increased, presumably due to increased numbers of sand flies and increased parasite transmission. A similar pattern also has been observed with other forms of leishmaniasis and with sand fly fever following the cessation of house-spraying programs.

Still other important factors involve human behavior and microbial adaptation. Until recently, ZVL was largely a disease of malnourished children. Tests in ZVL-endemic areas indicate that many people are infected with the parasite but that most well-nourished persons with normal cellular immune responses usually develop only mild, self-limited infections. However, the immunosuppression caused by HIV infection has changed this pattern. In the Mediterranean region, ZVL is now a common coinfection among HIV-positive adults. Some of these cases appear to be caused by reactivation of old inactive ZVL infections, while others represent new infections in immunosuppressed persons. In addition to natural transmission route by the bite of infected sand flies, there is accumulating evidence of direct person-to-person transmission of the parasite among intravenous drug users (many with HIV infection) sharing contaminated syringes.

The association between HIV and ZVL is especially ominous for another reason. Spanish investigators have reported infection of sand flies (P. perniciosus) by feeding the insects on blood from ZVL patients coinfected with HIV. The ease with which the insects were infected suggests that such patients could potentially serve as urban reservoirs of the parasite, establishing a focus of human–sand fly–human transmission and thereby eliminating the need for infected dogs to maintain a domestic cycle. Some scientists suggest that such a pattern has occurred with another disease: kala-azar. According to this argument, L. donovani, the etiologic agent of kala-azar, began as a true animal parasite and then evolved into a zoonosis. Finally, the animal reservoir host was completely eliminated, so that L. donovani is now maintained solely in a human–sand fly–human cycle. Since L. chagasi/L. infantum and its vectors already have adapted to the “peridomestic” environment, one wonders if, with the assistance of HIV, this parasite might evolve (emerge) in a manner similar to that of L. donovani.

These three examples are meant to illustrate the importance of ecologic factors in the maintenance and emergence of zoonotic diseases. The current paradigm of biomedical research, which focuses on mechanistic studies at the cellular, molecular, and genetic levels, frequently overlooks the importance of ecological factors in the development of human disease. Zoonotic diseases provide some of the best examples that the factors responsible for human illness involve much more than cellular immune response and gene expression. Furthermore, the ultimate control of zoonotic diseases probably depends more on our understanding of their epidemiology than of their molecular biology.

VECTORBORNE ZOONOTIC DISEASES

, Ph.D.

Chief, Arbovirus Diseases Branch, Division of Vector-Borne Infectious Diseases, National Center for Infectious Diseases, Centers for Disease Control and Prevention

A number of vectorborne zoonotic diseases, both bacterial and viral, have emerged in recent years or are now increasing in many regions. The agents that cause some of these diseases, such as plague and tularemia, also are of concern because of their potential use by bioterrorists. Thus, vector-borne zoonotic diseases, which are commonly transmitted from animals to humans by ticks or mosquitoes, are of immediate and growing concern in the United States and around the world.

From a public health viewpoint, viral diseases present particularly complex problems. There are more than 500 known arboviruses, or viruses transmitted by arthropods, and at least 110 of them have been associated with human disease, either in the form of natural occurrences or as laboratory-acquired infections. Thus, diagnosing new disease outbreaks can be a formidable task. In the United States, the good news is that we face only a few types of vectorborne viral diseases. The bad news is that these diseases are caused by members of a number of different virus families, including flaviviruses, which cause West Nile fever, and togaviruses, which cause several types of encephalitis. Many of the diseases caused by these viruses show very similar clinical symptoms.

A number of factors may be contributing to what appears to be an increase in arboviral diseases worldwide. Viruses may enter areas where they previously had not existed, as happened with the West Nile virus that emerged in the New York City vicinity in 1999. In addition, more people moving into areas where they have not normally lived, such as people moving into newly cleared forest areas in South America, may become exposed to the viruses. It also may be that members of the medical community in many regions have gained increased awareness of the possibility of encountering new viral diseases and are more likely to test for their presence. For example, in one study of LaCrosse encephalitis in the mid-Atlantic region of the United States, we found that as awareness of the disease increased among health care personnel, more cases were detected. Taken together, these factors suggest that arboviral diseases are more common than has been suspected and that some new viruses are appearing or old viruses are appearing in new places.

The case of West Nile virus may illustrate how the federal government responds to an emerging infectious disease, as well as some of the policy issues and problems that can arise. The virus was discovered in Uganda in 1937, and major outbreaks of the disease periodically occurred worldwide. Before its introduction to the United States, the most recent human or veterinary outbreaks had occurred in Romania in 1996 and in Italy in 1998. There also was a large human outbreak in Russia concurrent with the outbreak in the United States. The CDC sent a team to investigate the outbreak in Romania, and the information gained proved valuable when veterinarians and physicians in New York began to detect encephalitis in horses, birds, and humans but could not immediately determine the causative virus. Collaborations between CDC and other international health agencies also helped diagnose this outbreak. In the months following the outbreak, scientists in CDC's Arbovirus Diseases Branch assisted local and state personnel in identifying the disease agent and in characterizing its geographic spread.

Among the questions that still remain is how the West Nile virus reached the United States. Several routes have been suggested. Someone infected with the virus may have traveled by airplane to New York, where they were bitten by a mosquito that picked up the virus. The mosquito, in turn, may have transmitted the virus to a bird, perhaps a crow or a sparrow, since these species have now been found to carry the virus. Once established in the bird population, the natural disease transmission cycle could begin. Humans may have played a part in other ways as well. Through either legal or illegal activities, someone may have brought in animals, perhaps birds, that carried the virus, which was then transmitted to mosquitoes. Alternatively, infected mosquitoes may have hitched a ride by ship, breeding in some remnant of water during the journey. (While bioterrorism remains a threat with some arboviruses, evidence suggests that this was not the case in this instance.) In addition, it is possible that the virus entered the country “naturally,” carried by infected birds that had been blown across the ocean by storms, but this now appears to be a less likely route. CDC, in conjunction with other public and private agencies and organizations, continues to investigate the possible origins of the virus, though a number of technical problems may keep researchers from ever solving this mystery.

The initial investigations to identify the West Nile virus involved collaboration among many parties, both public and private, and at the federal, state, and local levels. By many accounts, this collaboration did not always work as smoothly as might have been hoped. CDC has begun to establish a comprehensive national response plan to deal with the sort of problems that can arise during such a multijurisdictional investigation of a disease outbreak. Among the lessons that have emerged, for example, is that there should be centralized coordination of investigations that involve multistate disease outbreaks. Responsibility for coordinating data typically rests with CDC, and in doing so the agency must address such issues as maintaining the security of shared information; protecting states' rights to privacy; ensuring that surveillance methods are consistent from state to state; and maintaining a common, easy-to-use disease database.

Dealing with the West Nile outbreak also has reinforced the importance of having a strong relationship between the medical, public health, and veterinary communities, especially at the local level. In addition, it is important for all members of the human and animal health communities to maintain awareness of unusual disease possibilities. Finally, the experience has pointed out a dilemma in devising efforts to control emerging arboviral diseases: that is, while vaccines potentially can prevent arboviral diseases, pharmaceutical companies typically are not interested in developing such vaccines. Many of these diseases either threaten too few people or threaten people, often those in developing countries, who cannot afford to pay for vaccines, thus reducing the economic incentives of vaccine development.

In early 2000, CDC published preliminary guidelines for surveillance, prevention, and control of West Nile virus in the United States. The agency developed these guidelines with extensive input from a variety of stakeholders, including arbovirologists, epidemiologists, laboratory personnel, vector control specialists, wildlife biologists, and state and local health and agriculture officials. The goals include monitoring geographic and temporal spread of the virus; rebuilding vectorborne disease surveillance infrastructure at the state level; developing more effective strategies for surveillance, prevention, and control; defining regional distribution and incidence of other arbovirus diseases; and providing up-to-date national and regional information on West Nile and other vectorborne diseases.

Federal, state, and local public health officials are now using the guidelines in an effort to minimize the public and veterinary health impact of West Nile virus. To aid in this control effort, Congress has given CDC $2.7 million to enhance surveillance efforts in states along the Eastern seaboard, because of concern that migrating birds will carry the virus southward from New York. More recently, the Department of Health and Human Services (DHHS) has provided CDC with an additional $5 million to expand surveillance efforts into other states, although on a more limited basis. All of these surveillance programs will be conducted through cooperative agreements between CDC and state agencies, often the state health department.

Note: Since this IOM Forum, the CDC has held additional meetings, in collaboration with other federal and state agencies, to review the effectiveness of the various activities related to West Nile virus, and it has issued updated guidelines to enhance surveillance, prevention, and control efforts. In 2001, Congress allocated another $21 million to CDC for West Nile surveillance, prevention, and control.

MATHEMATICAL MODELS AND PREDICTORS OF DISEASE OUTBREAKS

, Ph.D.

Senior Scientist, Center for Medical, Agricultural, and Veterinary Entomology, Agricultural Research Service, U.S. Department of Agriculture

Since World War II, the U.S. Department of Agriculture's Agricultural Research Service has worked with the Department of Defense (DoD) on developing mathematical models of human diseases. During the past decade, our laboratory in the Center for Medical, Agricultural, and Veterinary Entomology has developed risk assessment models for a number of diseases, including malaria, Lyme disease, and dengue hemorrhagic fever. The laboratory also has developed models for predicting the population dynamics of some of the mosquito species that serve as disease vectors, such as the mosquitoes that transmit Venezuelan equine encephalitis.

Among their strengths, models can help researchers and public health officials integrate large quantities of disparate information, thus permitting more sophisticated analysis of the factors involved in a particular disease or at a particular location. Models also can highlight critical unknown factors, and they can be used as a quick “first test” of various hypotheses before launching more costly and time-consuming evaluations in the field.

Perhaps the laboratory's most comprehensive modeling effort has focused on dengue fever. We have developed two types of models. One type is an entomological model, which can be used to predict with considerable accuracy how many Aedes aegypti mosquitoes (the vectors of dengue) there are in a given area. The second type is a transmission model, which combines input from the first model with detailed information about human population dynamics to predict how many people in the area are likely to become infected and develop dengue fever. Both models are built on a detailed knowledge base. Through earlier work with yellow fever, scientists have developed a thorough understanding of Aedes mosquitoes, including understanding of the mathematical relationships involved in how they propagate under particular conditions and how they spread disease. This level of understanding does not exist for any other arthropod vector. Validation studies comparing simulation results and predictions with actual field measurements conducted in a wide variety of locations indicate that the accuracy of the models is adequate.

These models are finding a variety of applications. For example, the models can help public health officials in a community threatened by dengue evaluate possible control measures. In most cases, it is impossible to completely eradicate all mosquitoes, since available insecticides and cleanup campaigns that target the containers in which mosquitoes breed are not completely effective. The models are meant to optimize control strategies using several methods in concert. An important use of the models in control projects is in teaching the dynamics of dengue as a function of a host of variables, such as weather, previous epidemics, and control measures.

To help make the insight of models more accessible for community officials, we have recently used the models to develop estimates of transmission thresholds. Put simply, this model can predict, for a specific set of environmental and population conditions, the relationship between the number of mosquito pupae present in an area (as a measure of the ultimate population of adult mosquitoes) and the number of cases of human disease that this vector population would be expected to cause. Thus, rather than conducting a full-scale test with the models, officials simply can conduct a survey to count the pupae present in their community and then compare the results to established standards. If the count reaches a certain level, then the officials know they need to undertake a control program of cleaning up or covering the most productive types of breeding containers. This type of information linking the extent of potential control measures to expected disease outcomes has been lacking in most control programs. The new method of targeted source reduction is being considered by the World Health Organization and the Pan American Health Organization and is in operation in countries in Southeast Asia and South America.

One encouraging recent finding from extensive work with this model is that it may not be too difficult for communities to mount effective mosquito control programs that will prevent or greatly reduce the spread of dengue fever. We have learned that rather than having to reduce overall mosquito populations in an area, it may be possible to focus control measures on a limited number of locations, such as specific types of containers in domestic areas, where most of the mosquitoes breed. Thus, local public health officials can conduct a rapid survey to identify the 1 percent or so of containers that, from our tests, may give rise to perhaps 99 percent of the mosquitoes in the community. Targeting these populations represents a far more manageable and less expensive undertaking. However, surveys in the Americas and in Southeast Asia indicate that this approach is not suitable for all situations.

In addition to these “mechanistic” types of models, which typically involve relatively rote processing of huge quantities of disease-specific and location-specific information, another type of modeling is showing promise. This type involves remote sensing, in which models incorporate data on weather and geography, often compiled by satellites operated by the National Atmospheric and Space Administration and the National Oceanic and Atmospheric Administration. Our laboratory, for example, currently is working with these agencies, as well as with academic and industry researchers, to convert daily satellite imagery into constantly updated grids depicting weather on a 1.6-kilometer basis. Using this detailed information, we are working to derive models for such things as plant growth and pest management. With further advancement, it also may be possible to derive models of vectorborne diseases, including West Nile virus and St. Louis encephalitis. Such models not only may prove useful in predicting the spread of naturally occurring diseases but may help in predicting how a biological agent released by terrorists might spread geographically and from one animal community to another before finally reaching humans.

Of course, modeling has its limits—and, indeed, its potential has often been oversold. One major problem, for example, is that for many diseases we simply lack sufficient background for developing reliable models; we do not adequately understand the disease system and its interactions with the environment, nor do we have reliable series of disease incidence to validate the models. In many cases, such a knowledge base is likely to remain elusive for years to come. But capitalizing on the recognized strengths of models—in both practical application and research—has provided at least some help in the effort to predict, manage, and even prevent some emerging zoonotic diseases.

THE ROLE OF NATIVE BIRDS AND OTHER WILDLIFE ON THE EMERGENCE OF ZOONOTIC DISEASES

, Ph.D.1 and , Ph.D.2.

1 Executive Director, Salton Sea Science Subcommittee, U.S. Geological Survey
2 U.S. Geological Survey

Wildlife can be an important source of transmission of infectious disease to humans. One potential transmission route involves hunting and fishing, both common activities in the United States and worldwide. For example, during 1996, approximately 11 million Americans, about 40 percent of the total population 16 years of age and older, took part in some recreational activity relating to wildlife and fish. Another potential route of infection focuses on urban and suburban environments. These locations are of special concern because of their increasing role as wildlife habitat, the greater interface between humans and wildlife that takes place within those environments, the paucity of knowledge about disease in those wildlife populations, and the general lack of orderly management for wildlife within those environments.

In the wild, several trends are contributing to the growing importance of zoonotic diseases. First, the spectrum of infectious diseases affecting wildlife today is greater than at any time during the previous century. Second, the occurrence of infectious diseases has changed, from sporadic, self-limiting outbreaks that generally resulted in minor losses to frequently occurring events that generally result in major losses of wildlife. Third, disease emergence has occurred on a worldwide scale in a broad spectrum of wildlife species and habitats.

Given the scope of the problem, current disease surveillance efforts are inadequate. Few state wildlife agencies allocate personnel and resources to address wildlife disease, despite their statutory responsibility for managing nonmigratory wildlife. Some state agencies provide minimal support for regional programs based at universities. At the federal level, the primary surveillance effort is conducted by the National Wildlife Health Center, operated by the U.S. Geological Survey. Outside of government, some veterinary schools, agriculture diagnostic laboratories, and other programs provide additional information on animal diseases, primarily by examining carcasses of dead wildlife submitted for analysis, and individual university-based researchers carry out a variety of studies.

Typically, information about the occurrence of disease in free-ranging wildlife is derived from surveys and mortality events in areas where wildlife observations by agencies and the public are frequent enough to detect their occurrence before carcasses are removed by scavengers and predatory animals. The result is that disease occurrence is grossly underreported, heavily biased toward mortality events, and biased toward species of special concern and interest, such as game and endangered species. Therefore, the available information should be viewed as the “proverbial tip of the iceberg” relative to disease activity within wildlife populations.

In general, mammals are the most important source of zoonoses transmitted by wildlife. However, birds are involved in the transmission of a number of serious zoonoses, especially vectorborne diseases. This is of special concern because of the greater geographic movement of many bird species. The 1999 outbreak in New York City of West Nile fever, which afflicted 62 people and killed 7 of them, serves as an example. After extensive study, scientists determined that the virus apparently was carried by crows and transmitted to humans by mosquitoes. (However, it is not known how the virus was initially introduced to the region or, indeed, to the United States.) Waterfowl, such as Canada geese and mallard ducks, represent a particular threat for disease transmission to humans. These species are becoming increasingly common in urban and suburban areas, both because of habitat losses elsewhere and because of the current trend in landscape planning toward creating planned communities that feature “miniestates,” natural areas, and golf courses—environments that are attractive to waterfowl. The urban/suburban environment also has become an increasingly important habitat for songbirds. Salmonella typhimerium has emerged as an important pathogen, causing large-scale epizootics, usually in association with bird feeding. The magnitude of the potential human–songbird interaction is reflected by the 1996 expenditure for bird food in the United States of approximately $2.7 billion, with nearly 39 million people participating in this activity during that year.

Environmental factors are the driving force for many emerging diseases of wildlife. In general, wildlife disease prevention and control will be most effective when environmental conditions are understood and the anthropogenic actions causing those conditions are addressed. The most important considerations can be classified under the headings of landscape changes, wildlife translocations, and human values.

Landscape changes include both changes to the physical environment and the introduction of exotic species. Changes in the wetlands of the Central Valley of California provide a case in point. These wetlands lie within the Pacific Flyway, one of four primary migration corridors for birds that typically breed in remote northern areas and winter in southern areas of the North American continent and beyond. Because of their location, the wetlands have always been an important stopover area for birds. In recent years, however, approximately 90 percent of the wetlands have been converted to agricultural lands and other uses. As a result, more than 60 percent of the entire Pacific Flyway waterfowl population is now channeled into about 10 percent of the former wetland habitat. Such mass concentrations of birds for prolonged periods facilitate exposure of large numbers of birds to disease agents that may be present. The frequency of outbreaks, variety of bird species involved, and numbers of birds exposed to various pathogens provide a continuum of opportunity for the development of novel host–parasite relations.

Human-created environments and exotic species are other important aspects of landscape change. For example, the Salton Sea, located in the desert of Southern California, was created in the early 1900s. Artificially sustained by agricultural drainwater, this highly saline body of water has the most productive fishery in the world and is one of the crown jewels of avian biodiversity. However, since the 1990s, the ecosystem's species richness has been tarnished by an unprecedented array of disease outbreaks that have killed large numbers of birds. The Salton Sea is an often-repeated contemporary situation that involves the creation of new environments and the mixing within these environments of multiple species that do not have established ecological relations with those that do. The opportunity for disease emergence is a component of the resulting species interactions and environmental changes taking place.

Wildlife translocation, in which humans move free-ranging wildlife from one geographic area to another, is a common conservation tool that has clearly facilitated disease emergence, including zoonoses. An example was the government-directed translocation of raccoons trapped within a known enzootic area for raccoon rabies in the southeastern United States. These rabies-infected animals were the source of a raccoon rabies epizootic in West Virginia that spread to numerous other mid-Atlantic states, some coastal Atlantic states, and New England. The result is that enzootic foci for raccoon rabies are now established in geographic areas where rabies in raccoons previously had either been incidental cases due to epizootics in other species or small, self-limiting events in raccoons. Rabies also has been translocated with foxes and coyotes moved for sporting purposes. Other types of wildlife movements by humans that are contributing to disease emergence are captive rearing of wildlife for release into nature, wildlife rehabilitation and releases, and translocations for commercial purposes. Bovine tuberculosis has recently spilled over from the agriculture industry into white-tailed deer. Establishment of bovine tuberculosis within free-ranging white-tailed deer populations will pose a significant human health threat because of the pursuit of this species by millions within the hunting community.

Disease emergence is as much a social issue as it is a biological issue. Of particular note, the prevailing philosophical attitude among many people within the wildlife conservation community is that disease is a natural event that need not be addressed. (Exceptions are made for transient responses to high-impact mortality events and for limited diagnostic activities in response to public inquiry about cases of visible mortality.) Proponents of this view maintain that impacts on wildlife population rather than on the individual animal should primarily determine whether or not there is a need for actions to be taken. This is a fundamental difference relative to human health and companion animal considerations, where clinical disease in individuals is of prime concern.

A reasonable question is: Why is more not being done to address disease within wildlife populations? One contributing factor is that while most people believe it is possible to deal with disease threats involving humans and domestic animals, similar confidence is lacking regarding our ability to control disease in free-ranging populations of animals. Response to disease in humans, livestock, and companion animals is governed by clear agency mandates supported by statutory authorities, legal mandates, laws, regulations, and other directives. The response process is facilitated by reporting systems, including designated reportable diseases; formal interagency infrastructures for disease diagnosis and control; infrastructures for epidemiological investigations; and systems for clinical treatment and fiscal considerations, among other factors. In general, all of those conditions are either absent or at best rudimentary for agencies with stewardship responsibilities for free-ranging wildlife. As a result, disease outbreaks in wildlife do not have a mandated responsibility to be investigated or dealt with. Because of the differences in agency responsibilities, agriculture agencies do not become involved unless the outbreak is known to be, or has a high probability of being, a disease of major concern for domestic animals. However, even in those instances, the wildlife are under the jurisdiction of wildlife agencies. Similar considerations are involved for zoonoses transmitted by wildlife. Complicating factors include animal rights advocates that give special attention to protecting wildlife, even when serious diseases are involved, and conservation legislation such as the Endangered Species Act and other laws that can constrain actions normally implemented when domestic animals are affected by a disease of major importance.

Free-ranging wildlife populations are under the legal stewardship of state and federal government agencies. Primary responsibilities are vested in different agencies depending on the types of species involved. Shared responsibilities between state and federal agencies generally exist despite one or the other having primary responsibility for a specific situation. This stewardship form of ownership leads to two results. First, there is no private ownership of free-ranging wildlife. Instead, wildlife are held in the public trust for human society. Second, the wildlife stewardship agencies are nonprofit organizations that have little economic incentive to address the costs of disease. Thus, there are no compelling reasons for government agencies to expend resources on disease prevention and little incentive for disease control.

To help in minimizing the emergence of diseases in wildlife, as well as the transmission of such diseases to humans, numerous observers have suggested a variety of actions. These actions include:

  • Interdisciplinary collaboration and cooperation. Federal agencies should develop a tripartite cooperative program to address infectious diseases in humans, in domestic animals, and in wildlife. This program should serve as a focus for regular communications through working groups to address information transfer; to improve response to disease emergencies; to establish priorities for collaborative, focused investigations; and to pursue other areas of mutual interest. The program also should serve as a model and catalyst to stimulate the development of similar cooperative programs between state agencies that would network with the federal program.
  • Database development. Federal agencies should take the lead in developing a common database for disease surveillance and monitoring that can be used to track infectious diseases and the emergence of new diseases. As part of this effort, a work group should be established to develop a listing of “reportable” diseases that are to be entered into the system, with standards for data entry, reporting, and utilization by collaborating agencies and institutions. The CDC has proposed developing a national electronic disease surveillance network for state and federal public health information on emerging infectious diseases, and this network should be expanded to include wildlife surveillance information on emerging zoonotic diseases.
  • Focused collaborative investigations. In ongoing research programs, joint development and planning can help ensure that high-quality specimens, reagents, information, and assays will be provided among collaborators at minimal costs. In other instances, joint budget initiatives will be required to provide the resources needed to carry out the monitoring programs and other focused investigations to address specific diseases. Agencies should develop agreements with one another to facilitate collaborative investigations on issues of mutual interest; such agreements should cover such issues as fund transfers, personnel assignments, and sharing of facilities and technical capabilities.
  • Biological repositories. There is a need to develop and maintain systems for archiving materials from wildlife disease investigations for retrospective and comparative studies. Isolates of infectious disease agents, serum banks, histological specimens, and other biological reference materials need to be organized in a coherent manner that provides ready access to them by qualified investigators willing to work as true collaborators within the area of emerging infectious diseases.
  • Disease ecology. Disease prevention and control activities will be enhanced by greater understanding of the epizootiology of wildlife diseases. This is a fruitful area for interagency and interdisciplinary collaboration. Efforts should extend beyond field and laboratory investigations to include the areas of mathematical modeling and geographic information system technology. These efforts should be supported by expanded databases of information from the physical, biological, and social sciences.
  • Urban wildlife disease studies. Given the increasing importance of urban/suburban environments as habitats for some species of wildlife, this is an important area for collaborative investigation. Such studies have urgency for protecting the well-being of migratory birds and other wildlife, as well as for protecting human health.
  • Public education. A coordinated, ongoing process is needed to provide the general public with timely, accurate information about emerging diseases of wildlife and the importance of such diseases to public health, domestic animals, and the wild animals themselves. One possible route would be for federal agencies to work collaboratively with an independent organization dedicated to public outreach.
  • Emergency response. Collaborative arrangements should be developed to integrate the emergency response capabilities within the public health, domestic animal, and wildlife conservation communities. Response to emerging infectious diseases of wildlife should be augmented as needed by the combined capabilities of the different programs to minimize the potential for establishment and spread of wildlife diseases capable of infecting other species, including humans.
  • Technical forum. Communications need to be improved between officials with responsibility for managing wildlife on public lands and researchers who study diseases of wildlife or diseases transmitted from wildlife to humans and domestic animals. It is not sufficient to rely on the diverse scientific meetings that currently incorporate wildlife diseases as agenda topics. The North American Wildlife Conference can provide an appropriate forum for bringing wildlife disease issues before those individuals who manage public lands, and organizers of the conference should be encouraged to develop a regular forum devoted to emerging diseases. The human health community should develop a reciprocal opportunity for participation by members of the wildlife community.
  • Guidance on landscape change. The expanding human population assures continued landscape changes. Thus, the government and other organizations should become proactive in terms of developing and disseminating information that can help guide land development in a manner that gives greater consideration to disease emergence. Initial actions that should be considered include distributing authoritative publications, sponsoring public forums, and providing consultations on particular problems.

To quote the comic strip character Pogo, “We have met the enemy and he is us.” This observation continues to be demonstrated for emerging diseases. Human arrogance cannot overcome biological processes. However, by replacing arrogance with some humility and by addressing these issues from a truly collaborative perspective, we will be able to improve environmental conditions substantially and impede disease in a manner that will greatly benefit humankind and our planet's biological resources for decades to come.

ANIMAL HUSBANDRY PRACTICES AND RISK FACTORS, WITH PARTICULAR REFERENCE TO BOVINE SPONGIFORM ENCEPHALOPATHY

, Ph.D.

Visiting Scientist, Plum Island Animal Disease Center, U.S. Department of Agriculture

Bovine spongiform encephalopathy (BSE) is caused by a member of a group of agents that collectively are known as transmissible subacute spongiform encephalopathies. These agents cause fatal human and animal neurological diseases, which are characterized by a long symptom-free incubation period followed by a short acute phase. Many such diseases have been described over the years, but attention has been focused on them most recently by the massive outbreak of BSE in cows that started in the United Kingdom in 1986. This outbreak presented a new problem for the agricultural industry because it had no precedent. The disease occurred in many herds but was of low incidence in any individual herd, with one to three animals being affected. The animals became uncoordinated and irritable, fell frequently, and eventually died. The only diseases remotely connected to BSE seemed to be scrapie in sheep and goats and Creutzfeldt-Jakob disease (CJD) and kuru in humans. Histological examination of the brains of infected animals showed that all three diseases caused similar lesions.

Veterinary scientists quickly traced the source of BSE infection during the United Kingdom outbreak to the food concentrates—essentially, tissues of slaughtered cows—that were being fed to cattle to enhance their productivity. The clear question to be answered was why the disease had emerged then, when the feeding of concentrates had been part of animal husbandry for several decades. Why had the disease not emerged earlier? Another critical question was how an infectious agent could survive the severe heat treatment used in the preparation of these concentrates. The only clue to these questions seemed to be the change made in the rendering procedure by which the concentrates were produced—a change in the solvent step that was made to decrease the cost of the process. This change had been made in the 1980s, leading to the suggestion that the incubation period of the disease was about 5 years.

Despite the imposition in 1988 of a ban on the feeding of concentrates, the number of infected animals continued to rise dramatically. The peak in the outbreak occurred in 1992 and 1993, confirming the initial suggestion of a 5-year incubation period. Subsequent pathogenesis studies have shown that cattle fed large amounts of the BSE agent do not develop disease earlier than 36 months, nor do their tissues contain any infectious agent before that time. This led to a ban on any use of meat or other animal tissues for human consumption from animals older than 30 months. (Of course, this raises a question: if an animal develops the disease when it has received the infected concentrate 36 months previously, how safe is the precursor of the infectious agent?)

Early on, many observers considered it unlikely that the agent causing BSE would lead to any problems in humans, since there was no evidence that the infectious agent that causes scrapie in animals causes disease in humans. Indeed, the Southwood Committee established by the British government reached this conclusion in 1989, as did the Spongiform Encephalopathy Advisory Committee (of which I was a member) in 1990. Still, the burning question in many quarters remained: Is beef safe? Such worries had a formidable impact on the British cattle industry.

Then, in March 1996, the U.K. Ministries of Agriculture and Health announced that human cases of a new form of Creutzfeldt-Jakob disease had been detected in a small number of young people. This was unexpected because CJD does not usually occur in young people, yet histological examination of the brains of these victims showed clear similarities to BSE. Subsequent evidence has confirmed that the new disease (now called new variant Creutzfeldt-Jakob disease) and BSE are caused by the same agent.

These new findings had perhaps an even greater impact, on both the cattle industry and the general public, than did the initial observations. Who was going to eat beef from a potentially infected animal? In a world demanding no risk, sales of beef plummeted, and some groups demanded that all 12 million cattle in the British herd be killed. The clear problem was to ensure public health but at the same time attempt to preserve an important and lucrative industry. Balancing these demands would require a delicate balancing act on the government's part.

The outcome, however, did not go well. Many individuals and groups argue that the BSE crisis was badly handled, both by the government and by its scientific advisers. But what were the options? There are several issues that need to be taken into account before judgment is rendered: (1) BSE was a new disease of cattle, (2) the disease was caused by an “old” agent, (3) the disease had a very long incubation period, (4) there was no in vitro diagnostic test, and (5) the assay system for detecting the agent relied on the use of mice and required months to complete.

One lesson seems clear, however: the problem in Britain was so large that the government should have appointed an “overlord” who would have devoted full-time attention to issues as they emerged. Advisory committees are no substitute for day-to-day involvement. (I suggested this approach to the British government in 1991, but it was rejected.) Would such an approach have made any difference? Because of the disease's long incubation period, it may not have done so. But having such a central authority would have ensured that all the scientists who knew something about these agents and the diseases they cause would have been brought into the equation at the earliest possible stage. Neither worrying about “territory” or “turf,” nor the seeking of glory, can solve such problems—and these issues should not cloud the work under way today on other emerging diseases.

NATURAL HISTORY OF SIMIAN IMMUNODEFICIENCY VIRUSES: CLUES TO THE EMERGENCE AND VIRULENCE OF AIDS VIRUSES

, Ph.D.

Visiting Scientist, Aaron Diamond AIDS Research Center

Simian immunodeficiency viruses (SIVs) infect more than 20 species of Old World monkeys and apes, all of them of African origin. Those species naturally infected include chimpanzees, mangabeys, mandrills, baboons, colobus, African green monkeys, and guenons. African green monkeys (genus Chlorocebus) were the first nonhuman primates found to harbor SIV in the wild.

In most cases, SIVs are species specific, meaning that viruses obtained from animals of a given species will phylogenetically cluster together. The species-specific clustering of viral sequences suggests that virus and hosts evolved in parallel and that SIV infections are ancient. However, several exceptions to the species-specific clustering rule indicate that SIVs do occasionally cross species barriers. For instance, viruses from African green monkeys (SIVagm) appear to have been acquired by a talapoin, a patas, a white-crowned mangabey, and two baboons. The frequency of SIVagm cross-species transmission may reflect the fact that African green monkeys are particularly widespread and numerous and thus represent a major SIV reservoir.

SIV infection has not been shown to cause disease in its natural hosts. This issue is obviously difficult to address in the wild but has been addressed by the epidemiological studies of sooty mangabey and African green monkey populations bred in primate centers. SIV seroprevalence typically is low in young animals but rises sharply in juveniles and young adult mangabeys, suggesting that SIV is mainly transmitted through the sexual route.

However, SIVs have the potential to become pathogenic when transferred to new host species, including humans. There is compelling evidence that SIV from sooty mangabeys (SIVsm) is the recent ancestor of the human AIDS viruses HIV-2 and of SIVmac, the virus that causes simian AIDS in rhesus macaques. There also is evidence that SIV from chimpanzees (SIVcpz) is the ancestor of at least some types of HIV-1. The clustering of several human and simian lentivirus pairs on phylogenetic trees indicates that cross-species transmission of SIVs to humans has been a repeated occurrence. Transmission events did not always result in the emergence of pathogenic and highly transmissible AIDS viruses, as indicated by the fact that only two of the six HIV-2 subtypes identified so far were associated with AIDS. HIV-1 types N and O are pathogenic but have a limited epidemic spread as compared to HIV-1 type M, suggesting differences in adaptation to the human host.

The factors responsible for the acquisition of virulence in the human host remain to be elucidated. This issue may be more easily addressed in simian models, by comparing SIV infection in species, such as the sooty mangabey and the rhesus macaque, that are resistant and susceptible to disease.

Sooty mangabeys range from Sierra Leone and Liberia to the western half of Ivory Coast. Converging evidence supports the idea that cross-species transmission of SIVsm to humans is at the origin of HIV-2. Among the evidence are the following observations: (1) SIVsm and HIV-2 are genetically close and share a common genome structure; (2) all known HIV-2 subtypes occur together only within the range of the sooty mangabey; and (3) ample opportunities exist for transmission, since people in the region hunt sooty mangabeys and keep them as pets, and the prevalence of SIVsm in sooty mangabeys is relatively high. The most convincing evidence is based on the phylogenetic clustering of SIVsm and HIV-2 isolated from the same geographic locale, either in Sierra Leone or Liberia.

Evidence for the origin of HIV-1 from chimpanzees can be found in the close genetic proximity between two SIVcpz from Cameroon and HIV-1 type N, which is found exclusively in the same country, indicating both phylogenetic and geographic coincidence between the two viruses. However, only seven SIVcpz have been characterized so far, and none of them cluster closely with the HIV-1 type M and HIV-1 type O. Thus, it may be premature to draw conclusions on the most recent ancestors for these two group of viruses. They are likely to belong to the SIVcpz group, but the detection of an ancestral HIV-1 type M or O in another species cannot be entirely ruled out at this stage. Chimpanzees are known to prey on other monkey species and may thus be exposed to transmission of heterologous SIVs.

One crucial question is whether cross-species transmission of SIVs to humans can readily generate an AIDS virus, or whether further adaptation to the new host is required for the emergence of a virulent HIV strain. Some insight can be obtained from studies of rhesus macaques. When infected with certain SIVsm isolates, such as B670, these monkeys readily develop AIDS, while inoculation with other SIVsm isolates does not appear to cause disease. It is also relevant to note that accidental SIVsm transmission to laboratory and animal workers has been documented in at least two instances, with no cases of AIDS reported thus far. Taken together, these observations suggest that simian immunodeficiency viruses have the potential to cause disease upon transmission to a new host but that this may actually be a rare occurrence.

The most telling piece of evidence is epidemiological. Only two of the six HIV-2 subtypes described so far have spread epidemically. The four individuals infected with subtypes C to F were all healthy (or, in one case, afflicted with a disease that is not associated with AIDS). Thus, the divergent HIV-2 subtypes C to F may represent viruses poorly adapted to the human hosts. It appears likely, then, that either the epidemic HIV-2 subtypes originated from the transmission of specific variants that happened to be pathogenic for humans or that the emergence of pathogenic HIV requires further adaptation to the human host through unknown mechanisms.

What mechanisms might drive the acquisition of SIV virulence in the human host? Experiments in the macaque models have repeatedly shown that serial intravenous passages increase SIV and HIV virulence in this host. Thus, it is possible to draw a parallel and speculate that serial intravenous passages could have contributed to the propagation and the adaptation of SIVsm and SIVcpz in humans. Epidemiologists and historians have documented multiple instances of reuse of nonsterile needles or even of direct arm–arm vaccination in Africa since the beginning of the 20th century. The main reason why serial intravenous passages can promote SIV adaptation is that they provide the setting for successive viral jumps from primary infection to primary infection. A poorly adapted virus would induce a very low viral load and therefore would be very unlikely to be transmitted during the chronic phase of the infection. The only window of time during which transmission could occur would be the few weeks that precede the establishment of the antiviral immune response—that is, the primary infection.

Copyright © 2002, National Academy of Sciences.
Bookshelf ID: NBK98097