|Home | About | Journals | Submit | Contact Us | Français|
Tuberculosis (TB), along with AIDS and malaria, is one of the three major killers among infectious diseases. New approaches to preventing, diagnosing, and curing TB are needed, which depend on a better understanding of Mycobacterium tuberculosis and the host. The National Heart, Lung, and Blood Institute convened a working group to develop recommendations for future TB research, including genetic aspects of the disease. The following areas were identified: (1) animal model research to improve understanding of persistence, reactivation, and granulomatous reactions; (2) preclinical studies aimed at shortening treatment of TB; (3) new resources for manipulating and characterizing the M. tuberculosis genome, proteome chips for more specific diagnoses, and studies of genes that appear to be essential but whose functions are not known; (4) prospective studies associated with clinical trials in populations with or at risk of TB to advance development of diagnostics and prognostics; (5) new quantitative and bioinformatic approaches to study the interaction between M. tuberculosis and the infected host and how this influences the infection process; (6) molecular characterization of M. tuberculosis genome diversity and phylogenetic analysis; (7) coordinated studies of human genome scans; (8) genetic epidemiology studies; (9) activities to foster knowledge dissemination, education, and training; and (10) coordination between the National Institutes of Health, the Gates Foundation, the Global Alliance for Tuberculosis Drug Development, and other organizations
Tuberculosis (TB), one of the earliest recorded human diseases, is still one of the biggest killers among infectious diseases. The international death toll due to its causative agent Mycobacterium tuberculosis (Mtb) is approximately 1.9 million people each year (1). The same study reported that there are approximately 8.8 million new infections annually, and that one-third of the world's population is infected by this pathogen. In the United States, where the infection rate is low, about 10 to 15 million people are estimated to have latent TB infection (http://www.cdc.gov/nchstp/tb/pubs/TaskForcePlan/TOC.htm). Previous review articles in the AJRCCM have described various aspects of the disease and its social implications (2–4).
Infection can lead to a potentially fatal outcome or an asymptomatic quiescent or latent state, but one in which latent bacteria can be reactivated to cause disease. Immunocompromised individuals and those subject to the ills of poverty are particularly at risk. The frightening public health problem of TB exists despite the widespread use of the Mycobacterium bovis bacillus Calmette-Guérin live attenuated vaccine and several antibiotics. This vaccine is variable in protection and does not prevent reactivation or reinfection (5), and the antibiotics in current use are losing their efficacy because of emerging drug resistance (6, 7). To rationally develop new strategies, we must learn how these bacteria circumvent host defenses to survive in infected human hosts, sometimes for years, and can then be reactivated to cause disease. It is essential to understand why relatively few individuals exposed to Mtb do contract the disease (1), because defining genetic determinants and correlates of host resistance/susceptibility to TB can help in the design of more efficient vaccines. It is equally important to understand why some clinical Mtb strains are more virulent than others, at least in animal models (8), and how Mtb reacts to the host at different stages of infection. Host responses to different Mtb clinical strains are also an important area for study. Successful research in these areas will aid in the design of new therapies.
Most TB infections are initiated by the respiratory route of exposure now that milk products are generally pasteurized, at least in the developed world. One study in 1978, before the AIDS epidemic, showed that 85% of new TB cases were pulmonary (9). Thus, the different forms of the disease that are discussed below usually arise from dissemination of the bacilli from the infected lung. TB in many cases follows a general pattern as described by Wallgren (10), who divided progression and resolution of the disease into four stages. In the first stage, dating from 3 to 8 weeks, after Mtb is inhaled in aerosols and implants in alveoli, the bacteria are disseminated by the lymphatic circulation to regional lymph nodes in the lung, forming the so-called primary or Ghon complex. At this time, conversion to tuberculin reactivity occurs. The second stage, lasting about 3 months, is marked by hematogenous circulation of bacteria to many organs, including other parts of the lung, and at this time sometimes fatal disease can occur in the form of tuberculous meningitis or miliary (disseminated) TB.
Pleurisy or inflammation of the pleural surfaces can occur during the third stage, lasting 3 to 7 months, causing severe chest pain, but this stage can be delayed for up to 2 years. It is believed that this condition is caused by either hematogenous dissemination or the release of bacteria into the pleural space from subpleural concentrations of bacteria. The free bacteria or their components are believed to interact with sensitized T lymphocytes that are attracted and then proliferate and release inflammatory cytokines (11). The last stage, resolution of the primary complex, where the disease does not progress, may take up to 3 years. In this stage, longer developing extrapulmonary lesions—for example, those in bones and joints and frequently presenting as chronic back pain—can appear in some individuals. In addition, other organ systems (e.g., genitourinary, nervous, circulatory) can be affected.
Most humans who are infected with TB do not exhibit progression of the disease. One-third of exposed HIV-negative individuals become infected and, of this number, 3 to 10% develop the disease in their lifetimes, whereas HIV-positive individuals infected with Mtb have a 7 to 10% chance of developing TB each year (12). TB in HIV-positive adults, whether resulting from reactivation or new infection, is frequently pulmonary (13), but extrapulmonary TB is often associated with AIDS and has been accepted as one criterion for an AIDS-defining diagnosis; its frequency among patients with AIDS ranges from 25 to 66% (14). It is believed that the host inflammatory response, especially overproduction of the cytokine tumor necrosis factor α (TNF-α), plays a major role in tissue damage associated with active TB (15). In the absence of antibiotic treatment or if Mtb resistant to front-line antibiotics is involved in the infection, over 50% of individuals with active TB die (14). This high mortality rate results from widespread destruction of the patient's lungs.
In the human host, Mtb usually enters the alveolar passages in the form of aerosols, where its initial contact is with resident macrophages and dendritic cells (DCs). The bacteria are phagocytosed in a process that is initiated by bacterial contact with macrophage mannose and/or complement receptors (16). On entry into an immunologically naive host macrophage, Mtb initially resides in an endocytic vacuole called the phagosome that does not progress into a later endosome. This avoids the normal endocytic maturation cycle observed with phagocytosed particles (i.e., phagosome–lysosome fusion), in which bacteria encounter a hostile environment that includes acid pH, reactive oxygen intermediates, and lysosomal enzymes. Reactive nitrogen intermediates (RNIs) such as nitric oxide (NO) and its derivatives can arise from the activity of inducible NO synthase (iNOS) in numerous cells, including activated macrophages. RNIs are major elements in antimycobacterial activity in the mouse (17). Although iNOS activity or RNIs are not usually observed in human macrophages from cell lines or macrophages derived in vitro from blood monocytes of healthy humans, alveolar and tissue macrophages from lungs of patients with TB show high levels of functional iNOS (18). In addition, infected lung tissue from patients with TB shows high levels of nitrotyrosine, a product of NOS activity, as well as elevated levels of iNOS activity (19). It is not known how Mtb sometimes avoids the killing activity of macrophages that is observed with other pathogens. However, cell wall components like lipoarabinomannan and secreted proteins like the 19-kD lipoprotein interact with macrophage receptors, and these interactions can modulate the signaling systems used by infected macrophages and DCs to activate the innate immune response (20). For example, the 19-kD protein binds to Toll-like receptor 2, and this interaction results in the repression of the interleukin (IL)-12 and the major histocompatibility complex (MHC) class II presentation pathways (21).
On the basis of much of the above evidence, many researchers favor the possibility that the macrophage has the major role in killing Mtb in the human host (22), and one research goal is to identify macrophage effector molecules. In the mouse, these include but are not limited to RNIs; others, some of which may be under the control of the LRG47 protein (23), remain to be identified. Another hypothesis is that the key cell types responsible for killing Mtb during infections are lymphocytes that produce granulysin and perforin, although granulysin seems to be lacking in the mouse and no clear role for perforin in the control of murine TB has been defined. This line of inquiry tends to focus on the strength on human cytotoxic lymphocyte (CTL) responses and how they may be enhanced by vaccination strategies.
The relative ease of working in tissue culture has provided much useful information on Mtb entrance and trafficking in the macrophage and on other responses of the infected cells, but much less is known about how the bacterium survives and grows during later stages of infection in the lung. Infected macrophages, through their production of chemokines, attract unactivated monocytes, lymphocytes, and neutrophils (24), none of which kill the bacteria efficiently (22). In some infected people, the disease does not progress past this stage and bacteria are ultimately cleared completely. Other individuals form granulomatous focal lesions composed of macrophage-derived giant cells and lymphocytes. This process is generally an effective means for containing the spread of the bacteria. A strong cell-mediated immune (Th1) response is necessary for granuloma formation, and TNF-α is especially important in this process (25). In the enclosed granuloma, bacilli-loaded macrophages are killed, and this results in the formation of the caseous center that is surrounded by a cellular zone of fibroblasts, lymphocytes, and blood-derived monocytes (26). Although Mtb bacilli are postulated to be unable to multiply within this caseous tissue, due to its acidic pH, low availability of oxygen, and the presence of toxic fatty acids, some organisms may remain dormant but alive, for decades. The strength of the host cell-mediated immune response determines whether an infection is arrested here or progresses to the next stages. This enclosed infection is referred to as “latent” TB and can persist throughout a person's life, in an asymptomatic and nontransmissible state. With efficient cell-mediated immunity, the infection may be arrested permanently at this point. The granulomas subsequently heal, leaving small fibrous and calcified lesions. However, if an infected person cannot control the initial infection in the lung or if a latently infected person's immune system becomes weakened by immunosuppressive drugs, HIV infection, malnutrition, diabetes mellitus, aging, or other factors, the center of the granuloma can become liquefied in an unknown manner and then serves as a rich medium in which the now-revived bacteria can replicate very rapidly. At this point, viable Mtb can escape from the granuloma and spread within the lungs (active pulmonary TB). It is at this stage that Mtb can enter alveoli and bronchi of the lung, and the infected individual becomes infectious because the bacilli can be disseminated to others by coughing or expectoration. Mtb can also be spread to other tissues via the lymphatic system and the blood (miliary or extrapulmonary TB). When the disease progresses to this point, intensive antibiotic therapy is essential for survival (26).
Little is known about the mechanism(s) of TB reactivation. Some experiments on “persistence” or “latency” have been performed using chronic mouse infection models in which bacterial numbers can be maintained in a steady state, in the absence of disease. Reactivation of bacterial growth and histopathology in these models can be induced by the administration of the iNOS inhibitors N-iminoethyl-l-lysine (27) and aminoguanidine (28). These results and other data strongly suggest that cell-mediated immunity (i.e., IFN-γ induction of iNOS and Th1 cytokines) plays a major role in limiting bacterial growth in infected animals. It is still not certain whether the bacteria in the chronic disease models are actually viable but nongrowing, which would reflect a true latent state, or whether they are growing and dying at the same rate. An early observation showed that Mtb in chronically infected mice is partially susceptible to isoniazid (29), a drug that only is effective against growing Mtb (30), providing evidence for both explanations—that is, noncycling persistence and balanced growth and death. Recent experiments now strongly suggest that the bacteria in the chronic disease models are actually viable but nongrowing (31). Moreover, mouse models of acute and chronic TB are different from the human disease in many regards. Other animal models of TB, such as guinea pig, rabbit, and, more recently, monkey infections, are closer in some respects to their human counterpart (32, 33). This is especially true for the formation and development of pulmonary granulomas in human TB infections that show caseation and cavitation (34). Another issue with the mouse “ latency and reactivation” experiments is that iNOS plays a major role in these processes. Although iNOS is expressed in tuberculous lesions in humans, as discussed above, there is no direct evidence for a protective role for RNIs in human TB and the role of these compounds in human TB is unclear.
Although we know that strong cell-mediated immune responses are necessary for protection and that CD4+ T cells are needed to end the lymphohematogenous stage of early infection in animal models and humans, we still lack critical knowledge regarding the mechanism(s) that govern the induction and maintenance of latent TB infection in humans. Investigations involving persons who have had active versus latent TB infection suggest that latency might be associated with stronger CD8+ CTL responses (35). In addition, Mtb-specific CD8+ T cells preferentially recognize and lyse heavily infected antigen-presenting cells, suggesting a role for these lymphocytes in immunosurveillance (36). This hypothesis is consistent with experiments in mice in which suppression of CD8+ T cells with anti-CD8 antisera hastens reactivation of TB in a latency model of infection (37). The antigen-presentation pathways involved in inducing CD8+ CTL responses in humans are being actively studied. A currently discussed possibility is that either Mtb or its antigens escape from the phagosome into the cytosol and are presented via proteasome–MHC class I pathways (38). A related hypothesis proposes apoptosis-associated cross-priming in which mycobacterial antigens in the cytosol of the infected macrophages are packaged into apoptotic blebs. These structures subsequently fuse with DCs, the primary antigen-presenting cells to naive T cells, so that the antigen contents of the blebs become cytosolic in the DCs and can be presented efficiently (39, 40). Thus, differences in the apoptosis of monocytes and macrophages during initial infection with Mtb might determine how well microbial antigens enter into apoptosis-associated cross-priming pathways and might thereby influence whether active or latent disease develops. Also, a role for TNF-α in controlling later stages of human TB infection and latency has been postulated, as elevated levels of this cytokine are found in chronically infected human lung tissue (36), and the use of anti–TNF-α agents like infliximab to treat rheumatoid arthritis and Crohn's disease is associated with the reactivation of latent TB (41).
Currently, there is little information on how Mtb responds to the environment of the lung. There is biochemical evidence that the intermediary metabolism of Mtb is different during the course of mouse infections than that observed in bacteria growing in broth culture because fatty acids seem to be a preferred carbon source for Mtb in vivo (42–44), but the significance of these apparent changes in metabolism for acute or chronic infection is not known. Recent experiments have shown that RNIs induce a unique pattern of gene expression in Mtb (45), and specific Mtb genes show different patterns of expression in wild-type as compared with IFN-γ knockout mice (46). This is consistent with Mtb in lungs responding to the onset of cell-mediated immunity.
Recent methodologic advances in Mtb genetics, such as more efficient allelic replacement and transposon mutagenesis (47, 48), as well as the DNA sequencing and annotation of the Mtb H37Rv genome (49) and those of related mycobacteria that have been or are currently being completed by the Sanger Center–Pasteur Institute consortium and by the Institute for Genomic Research (50) have led to a burst of new information on the genetics of Mtb virulence. The demonstration that large-sequence polymorphisms are unique events that generate robust phylogeographic bacterial population structures provides a rational framework for understanding the consequences of genetic variation among clinical isolates of Mtb (51). Recent scientific advances have demonstrated that over 200 genes and the proteins they encode are necessary for Mtb virulence (52, 53). These genetic and molecular approaches have allowed researchers to analyze the genes responsible for differences in virulence that are observed in clinical Mtb strains (8) and to characterize the proteins they encode as being responsible for these differences in ability to cause disease (54). New National Institutes of Health (NIH) initiatives to fund a comprehensive mutant analysis of the Mtb genome are now underway and approximately 900 transposon mutants have been obtained (NO1 30036, awarded to W. Bishai, Jr.). At the same time, some progress has been made in discovering genetic loci in mice that determine susceptibility/resistance to Mtb infection (55–58). Likewise, specific human population cohorts are being analyzed to elucidate the reasons for high susceptibility to TB (59–62). In some cases, correlations have been made with specific human genetic loci (63, 64). A new project to use whole genome scans with multiple single nucleotide poymorphisms (SNPs) to study 1,000 patients with TB and 1,000 control subjects is now underway and should provide valuable information on human genetic loci (A.V. Hill, personal communication, 2005). Thus, genetic tools are now in place for integrating rigorous population genetic analysis of both the pathogen and the host.
The National Heart, Lung, and Blood Institute Working Group identified the following areas as important for future research. It is important to stress that the goals of the workshop did not include vaccine development, clinical management, diagnostics, or drug development. This did not reflect a lack of interest. On the contrary, these issues have such high priority that they are being addressed by other working groups and funding mechanisms.
The authors thank the participants of the workshop who provided the important ideas embodied in this report. These participants included: co-chairs Carl Nathan (New York, NY) and Issar Smith (Newark, NJ); Alexander S.Apt (Moscow, Russia); Samuel M. Behar (Boston, MA); John Chan (Bronx, NY); Zheng Chen (Chicago, IL); Paul Converse (Baltimore, MD); Jeffrey Cox (San Francisco, CA); Ken Duncan (Seattle, WA); Sabine Ehrt (New York, NY); Philip L. Felgner (Irvine, CA); JoAnne Flynn (Pittsburgh, PA); Li M. Fu (Anaheim, CA); Dorothy B. Gail (Bethesda, MD); Anne E. Goldfeld (Boston, MA); James E. Graham (Louisville, KY); Liana Harvath (Bethesda, MD); Sandra Colombini Hatch (Bethesda, MD); Robert L. Hunter (Houston, TX); Gail G. Jacobs (Bethesda, MD); William R. Jacobs, Jr. (Bronx, NY); Chinnaswamy Jagannath (Houston, TX); Gilla Kaplan (Newark, NJ); Douglas S. Kernodle (Nashville, TN); Denise Kirschner (Ann Arbor, MI); Hardy Kornfeld (Worcester, MA); Igor Kramnik (Boston, MA); Mike Kurilla (Bethesda, MD); Yukari C. Manabe (Baltimore, MD); Robin Mason (Bethesda, MD); John Mittler (Seattle, WA); Hannah H. Peavy (Bethesda, MD); Richard Pine (Newark, NJ); Mary Reichler (Atlanta, GA); Jesse Roman (Atlanta, GA); Eric J. Rubin (Boston, MA); James C. Sacchettini (College Station, TX); Erwin Schurr (Montreal, Canada); William K. Scott (Durham, NC); David Sherman (Seattle, WA); Christine Sizemore (Bethesda, MD); Peter Small (Seattle, WA); Andrew A. Vernon (Atlanta, GA); Zhenhua Yang (Ann Arbor, MI); Thomas Zahrt (Milwaukee, WI).
Supported by the Division of Lung Diseases, National Heart, Lung, and Blood Institute, National Institutes of Health, Bethesda, Maryland.
Originally Published in Press as DOI: 10.1164/rccm.200506-997WS on September 28, 2005
Conflict of Interest Statement: None of the authors have a financial relationship with a commercial entity that has an interest in the subject of this manuscript.