Challenges for the Development of New Antimicrobials—Rethinking the Approaches: Report of a Workshop
Committee on New Directions in the Study of Antimicrobial Therapeutics: New Classes of Antimicrobials
In 1974, Lewis Thomas described the highest form of medical technology as “the kind that is so effective that it seems to attract the least public notice; it has come to be taken for granted … [and is] exemplified best by … the contemporary use of antibiotics and chemotherapy for bacterial infections … [which] comes as the result of a genuine understanding of disease mechanisms” (Thomas 1974, pp. 34-35). This pronouncement was overoptimistic and premature. Our understanding of the ability of microorganisms to evade modern chemotherapy and to evolve strategies for inactivating our most potent antibiotics was in fact rudimentary, and we are now faced with substantial infectious-disease challenges. In the face of newly emerging infectious organisms, the global crisis in antibiotic resistance, and the threat of bioterrorism, there is a need to invigorate the basic science and technology of anti-infective chemotherapy. To do so, the mechanisms of infectious disease must be better understood, based on a deeper appreciation of microbial physiology, a comprehensive understanding of antibiotic resistance, and a renewed commitment to the discovery of novel antimicrobial molecules and therapies.
There are several indications that new approaches are required to combat emerging infections and the global spread of drug-resistant bacterial pathogens. One is the pattern in rates of death from infectious disease in the 20th century: from 1900 to 1980, the rate dropped from 797 per
100,000 people to 36 per 100,000 people, a reduction by a factor of more than 20 and a testament in part to the efficacy of antibiotics (Armstrong et al. 1999). However, from 1980 to 2000, that rate doubled, largely because of HIV but also due to the spread of drug-resistant bacterial pathogens, such as methicillin-resistant Staphylococcus aureus (MRSA), vancomycin-resistant enterococci, multiple-drug-resistant gram-negative bacteria, and multiple-drug-resistant tuberculosis (Cohen 2000). While the rise in mortality is due partly to infection in more seriously ill or immunocompromised patients, there is no doubting the need for new strategies and new molecules to treat pathogens that are resistant to nearly the full array of contemporary antibiotics. We are at a critical point, not seen since the pre-antibiotic era, at which infections caused by some bacterial pathogens are untreatable.
A second indication of the need for novel antibacterial therapeutics is the almost 40-year innovation gap between introductions of new molecular classes of antibiotics: fluoroquinolones in 1962 and the oxazolidinone linezolid in 2000 (Walsh 2003a,b). A third indication is the recent trend by several large pharmaceutical companies to leave the antibacterial and antifungal therapeutic arenas, suggesting a future decrease in scientific expertise in antibacterial-drug discovery and development skills (Projan 2003; Shlaes 2003). A technology gap is developing and widening, as research on and development of new antimicrobial agents are being de-emphasized or abandoned by many pharmaceutical companies.
Treatment of microbial infections—bacterial, fungal, and viral—selects for the emergence of resistant organisms that may be rare in the initial population but become increasingly prevalent under selective drug pressure. In fact, the presence of an antibiotic can accelerate mutation and recombination in bacterial populations and contribute directly to its own obsolescence (Cirz et al. 2005). This is in addition to resistance that may develop outside of the clinical setting; for example, resistance to penicillin had been documented even before its first widespread clinical use (Abraham and Chain 1988). Resistance is prevalent, heritable, and ancient.
The need for new generations of anti-infective agents, and in particular new antibacterial agents, is constant, as the emergence of resistance is largely a question of when and not if. Medicinal chemists have been highly successful over the last 50 years in reshaping the scaffolds of earlier antibiotics, both natural and synthetic; for example, current antibiotics include the fourth generation of beta lactams and the third generation of
macrolides. However, significantly new approaches and strategies for breakthrough molecules have not been forthcoming.
Antibiotic resistance affects more than one or a few patients: the global accumulation of resistant bacteria threatens everyone’s health. Once a problem associated only with the sickest patients in intensive-care wards, antibiotic-resistant bacteria have become widespread in communities throughout the world. Resistance genes are not distributed randomly in bacterial populations but are commonly clustered in multiple-drug-resistant strains with resistance spread together. The frequency of international travel, combined with the lack of worldwide standards of antibiotic use, exacerbates the problem. The result is an acceleration of the spread of resistance around the globe and in every environment. All stakeholders recognize that the current antibiotic-resistance crisis is associated with a predictable, inexorable loss of efficacy of our current antimicrobial arsenal, but substantial economic, regulatory, and scientific barriers to the development of new antimicrobial agents and therapies persist (Nathan 2004).
This report arises from extensive discussions at a brainstorming workshop organized by the National Research Council of the National Academies under the sponsorship of the National Institute of Allergy and Infectious Diseases. This workshop was planned by the Committee on New Directions in the Study of Antimicrobial Therapeutics: New Classes of Antimicrobials (see Appendix B). Some 40 persons attended the workshop, held on May 23-24, 2005, in Washington, DC, to address strategies for new generations of antimicrobials (see Appendix C for a workshop agenda and participant list). The committee felt that identification of a class of antimicrobials that would be effective in the treatment of a full range of microorganisms—bacteria, viruses, and fungi—without also being detrimental to the host was unlikely. Thus, the workshop was structured around the development of antibiotics that would be effective against bacteria. However, several of the ideas described in this report (e.g., employing natural microbiota to combat pathogens) might also be applicable as treatment strategies against viruses and fungi. The accompanying report on immunomodulation offers additional discussion on treatments that might be effective against the wider range of microorganisms.
This report has four major sections: a discussion of the challenge of antibiotic resistance at the population and molecular levels, the importance of understanding bacterial communities and resident microbiota for the discovery of new antimicrobial therapies, consideration of biological processes that can guide strategic approaches to antibiotic development, and
strategies for discovering new natural and synthetic molecules, including novel screening approaches to bacterial targets. It is hoped that his report will help guide the next decade of antimicrobial research and development.
Antibiotic Resistance Is Inevitable
The goal of the workshop was to identify novel approaches to the development of antimicrobial therapeutics. However, workshop discussions made it clear that even the most innovative antibiotics will be made obsolete, at some point, by the inevitable emergence of resistance. Therefore, the committee concluded that it is worthwhile to identify research that would help to surmount the problem of resistance or at least slow its emergence. The recommendations in this section, although they do not lead directly to the development of novel antibiotics, could be important in increasing the useful lifespan of current and future antibiotics.
Bacteria predate humans by billions of years and have evolved a complex series of coping mechanisms that enable them to survive under harsh conditions and in the presence of numerous toxic metabolites. Most antibiotics discovered during the golden age of antibiotics (about 1945-1960) are natural products, produced for the most part by bacteria themselves (Clardy and Walsh 2004). These bioactive secondary metabolites—so called because they are not essential for cell growth or reproduction—may be produced by bacteria to provide a competitive growth advantage by killing susceptible neighbors in the environment, or they may be signaling molecules that have other functions and whose antibiotic activity at high dosages is a side effect. In either case, organisms that produce antibiotics—and organisms that have evolved to live near them—harbor specific and potent resistance strategies that inactivate or otherwise protect them from the antibiotics’ toxic effects.
Synthetic antibiotics, such as the fluoroquinolones, would appear to be less susceptible to causing resistance, given that bacteria would not have had millennia of exposure to them. Thus, evolved mechanisms of resistance would be less likely to exist. However, ubiquitous and promiscuous efflux systems have evolved to protect microorganisms from diverse toxic small molecules of natural origin, and these systems often provide cross-protection against such non-natural products. As a result, genes that encode resis-
tance elements are embedded in the genomes of virtually all bacteria; these hard-wired resistance genes are inherited in vertical fashion, providing continuous protection against toxic agents in a bacterial species even in the absence of prior exposure.
Antibiotic-resistance genes, however, are not confined to bacterial genomes. They are also frequently found on mobile genetic elements (plasmids, transposons, and integrons) that readily pass horizontally from organism to organism, even across species boundaries, thereby circumventing the standard parent-to-progeny route of genetic flow (Levy and Marshall 2004). The frequency of selection for such events and for the acquisition of genetic elements increases with increased exposure to antibiotics. It is therefore not surprising that locales in which antibiotic use is rampant—such as hospitals, farms, and child-care settings—are prime sources of organisms that harbor these genetic vehicles. Furthermore, genetic elements passed between organisms in this way often collect several resistance genes; this process is, again, selected for by increased antibiotic use that has given rise to multiple-drug-resistant (MDR) organisms, some of which are untreatable—or nearly so—with the current arsenal of approved drugs. MDR organisms have changed from being primarily a health-care sector problem to being a source of community-acquired disease as patients return to their homes earlier than previously, often remaining on courses of antibiotics.
Antibiotic Resistance Is Manageable
1. Predicting Resistance
The inevitability of resistance is well accepted by researchers in the field, but there are barriers to collecting and sharing data on resistance among diverse geographic centers and among individual health-care settings within a single geographic region. Efforts to overcome technical and jurisdictional obstacles will improve the ability to monitor resistance, anticipate its spread, and inform health-care practitioners of its existence in the area.
Antimicrobial resistance grows as strains of bacteria that carry and exchange resistance genes spread throughout a population or region. Knowledge of resistance in bacteria from prior infections thus helps to target both treatment of new infections and efforts to contain resistance locally and globally as information about recent infections can anticipate antimicrobial
resistance in new situations. Tens of thousands of clinical and basic research laboratories throughout the world generate resistance data. But very few labs submit these data to appropriate databases that could allow local analysis or linking with a surveillance network.
The effectiveness of surveillance data can be enhanced by integrating with other types of information. For instance, molecular studies of resistance can help explain observed resistance phenotypes. Comparison with data on antibiotic usage allows estimation of and potential for the management of antibiotic selection. Data on resistance in non-pathogenic organisms, a potential reservoir of new resistance genes, could help anticipate the emergence of new resistance and to develop predictive diagnostics prospectively.
Surveillance of resistance can and should build on existing resources. Clinical laboratories in more than eighty countries have begun to build databases and link them into international networks using free software (WHONET) downloadable from a World Health Organization Web site (O’Brien et al. 2001).1 This type of surveillance data can be complemented and cross-validated by data on isolates collected from clinical laboratories for selected studies in public health reference laboratories or in those supported by pharmaceutical companies. Several companies now collaborate with the Alliance for the Prudent Use of Antibiotics (APUA) to merge their data for these types of meta-analyses (Stelling et al. 2005). Obviously, data in such a database should be as up-to-date as possible and thus able to identify pockets of resistance as they occur.
Surveillance needs to be implemented on a grand scale and over a long period to identify trends and provide data for population studies. The surveillance network should not only be nationwide, but linked to international efforts to integrate worldwide data seamlessly. Surveillance should not be limited to the health-care sector. Mechanisms of resistance to any new antibiotic may already exist in nature, so any resistance encountered in nonpathogenic organisms in the environment or antibiotic producers should also be entered into the database. Clinicians and developers of diagnostics would then be aware of resistance mechanisms that may be encountered in the clinic. Such an integrated database would greatly enhance the ability to develop predictive diagnostics that could be rapidly brought on line as resistance elements move around the globe.
Available at <http://www.who.int/drugresistance/whonetsoftware> at the time of publication.
2. Detecting Resistance
Detection of resistance is rooted largely in the century-old technology of growth susceptibility. As a whole, medical microbiology has not adopted state-of-the-art molecular diagnostic measures, and the barriers to gene- or protein-based diagnostics have been substantial. First, in non-sterile sites, such as skin and the gastrointestinal tract, identifying the specific agent causing disease is difficult; even in sterile tissues, such as blood, present-day detection methods are often not sensitive enough to detect disease-causing organisms.
Second, the cost associated with molecular tests is often prohibitive. Nevertheless, improved diagnostics could have a revolutionary effect. For example, if a physician could know at the bedside which organism is causing a particular infection and whether that organism is resistant to common antibiotics, treatment could be tailored appropriately. Consequently, antibiotics would be used in a specific fashion, selecting only those likely to be effective; this procedure of judicious and specific antibiotic use would thus help extend the useful lifetime of new antibiotics.
Diagnostics able to identify the etiology and antimicrobial susceptibility of all infections could target therapy precisely and eliminate the use of antibacterial agents in patients who do not even have a bacterial infection. If they are done early, such tests could avoid untargeted therapy during the days needed by current diagnostics. In a recent study, polymerase-chain-reaction testing took 6 hours to identify the etiology of 76% of community-acquired pneumonia cases, while older tests took several days to identify 49.5% (Templeton et al. 2005).
Specifically, research into diagnostic tests that can reliably and quickly identify pathogenic organisms and their resistance profiles should be encouraged. Development of such tests may be difficult but could lead to significant advancement in the treatment of infectious disease. To effect the greatest reduction in inappropriate antibiotic use, such tests would need to be so rapid and reliable that clinicians would be comfortable waiting for the results before beginning antibiotic treatment. Determining the necessary degree of reliability for these diagnoses is an open challenge as the tests must not only lead to successful diagnosis of the pathogen, but also have the confidence of clinicians.
A major issue in resistance is that not only disease-causing organisms, but also other resident organisms and the host itself are exposed to an antibiotic. Minimizing exposure through precise choice of antibiotic is critical
in preventing the emergence of resistance by reducing selection in off-target organisms. That is, the use of a narrow-spectrum antimicrobial agent optimized for use against the disease-causing organism would be less likely to select for resistance in non-targeted microorganisms. Advanced diagnostics discussed above will facilitate tight targeting of pathogens and thereby enable the productive exploration of target-specific antibiotics. The advances could include selective interruption of organism-specific processes, such as virulence mechanisms, adhesion of surface antigens, and resistance mechanisms. Enhancing the host response at the site of infection is a potential creative approach to activating toxic molecules where they are needed. New tissue-specific delivery vehicles would greatly help to decrease the exposure of non-target species to antibiotics. However, it should be noted that these strategies will not eliminate resistance as resistant microorganisms also arise from the use of antibiotics in non-clinical settings such as agricultural use.
3. Deterring Resistance
How resistance elements are selected and spread throughout microbial communities is largely unknown. Understanding the fundamental principles underlying how pathogenic organisms and normal microbiota communicate and exchange genetic information is a key to the ability to manage the spread of resistance.
Research on the molecular mechanisms that facilitate resistance is also warranted. For example, do some antimicrobial agents inherently activate mutagenic pathways that can lead to resistance (e.g., Cirz et al. 2005)? If so, are there classes of molecules that are less susceptible to this action? Similarly, are there antimicrobial targets that are less tolerant of mutations selected by the presence of antibiotics?
Furthermore, how such issues as antibiotic dosage and scheduling, antibiotic mixtures, and interactions with other drugs affect the emergence of resistance is not well understood. The example of amoxicillin/clavulanate potassium (Augmentin), a highly successful combination of an antibiotic and an inhibitor of resistance, should be emulated (Matti et al. 1998). Combination therapy to inhibit the emergence of resistance has also been used in the treatment of HIV (HAART therapy) and tuberculosis (isoniazid, rifampin, and pyrazinamide) (Finch et al. 2003). Leveraging knowledge of molecular mechanisms of resistance in the development of selective inhibitors has the potential to rescue the activity of proven antibiotics that
have well-established pharmacological and disease profiles (Wright 2000). Extending the clinical lifetimes of proven antibiotics in this fashion holds great promise. The challenge will be to selectively target the most important resistance mechanisms.
Recommendations on Resistance
To respond to issues of resistance to antimicrobials, the committee recommends the following research directions and action items:
A-1.1 Establishment of a simple and readily searchable antibiotic resistance database into which participating institutions would upload resistance data in real time.
A-1.2 New rapid diagnostics to detect pathogens and their resistance to inform therapy in real time.
A-1.3 Development of strategies that will selectively target pathogenic organisms while avoiding targeting the host and beneficial or benign organisms.
A-1.4 Identifying the sources of resistance mechanisms, their evolution, and the ways in which they are spread in microbial communities, to elucidate the various ways in which resistance can be manifest.
A-1.5 Development of strategies that target and selectively block antibiotic resistance mechanisms to rescue antibiotic activity.
A-1.6 Exploration of the effect of antibiotic usage, alone and in combination, on the development of resistance.
A MICROBIAL COMMUNITY APPROACH TO NEW ANTIBACTERIALS
Characterization of Communities of Microbiota
There is growing evidence of the important role played by resident microbiota in offering protection from infectious disease. Rather than continuing the traditional approach of killing bacteria wherever they occur, there is a need to develop new antimicrobial strategies aimed at subtle manipulation of bacterial behavior. Such therapies would favor natural host defenses and the maintenance of the normal microbiota to keep growth of
pathogenic species in check. At the outset, design of strategies for novel antibiotics should include exploration of strategies for exploiting beneficial and commensal bacteria in fighting infections in sites where normal microbiota reside.
To develop such therapies, a deeper understanding of the diversity and ecology of the normal human microbiota and how these communities are established and stably maintained is needed. At present, understanding of human microbiota communities and their true diversity and ecology is limited (Eckburg et al. 2005; Hooper and Gordon 2001; Wilson 2005; Nataro et al. 2005). Precise definition of these associations in human health and in disease will allow the development of nontraditional therapeutics aimed at manipulating bacteria and their environment to enhance the maintenance and proliferation of the normal microbiota and inhibit the growth of pathogens.
The committee recommends the following to deepen understanding of natural microbiota:2
A-2.1 Characterization and enumeration of the normal resident microbiota in human hosts.
A-2.2 Understanding the relationship between resident microbiota populations and human health.
Manipulating Bacterial Signaling and Communication
The last decade has taught that bacteria do not live independent lives but, rather, communicate within and among species by using a variety of secreted signal molecules (Miller and Bassler 2001). Production, detection and response to these molecules allow bacteria to take a census of the population and synchronize behavior on a population-wide scale. This process, called quorum sensing, is critical for many pathogens because expressing virulence genes as a group ensures that pathogenicity factors are released only when bacterial numbers are sufficient to guarantee success against the host (Donabedian 2003; Williams 2002). More complete understanding of the chemicals that bacteria use for signaling and how bacteria integrate and interpret chemical information in their environment would allow investigation of their use in antibacterial treatments.
At present, only four predominant classes of molecules used for communication are known: acylhomoserine lactones in gram-negative bacteria (Parsek et al. 1999), oligopeptides in gram-positive bacteria (Lazazzera and Grossman 1998), γ-butyrolactones in the streptomycete subset of gram-positive bacteria (Chater and Horinouchi 2003), and a furanone called AI-2 that is used for signaling in diverse bacterial species (Miller et al. 2004; Chen et al. 2002). The chemical lexicon is probably much larger than is currently recognized, and a continued study of cell-cell signaling with an emphasis on further definition of the chemical moieties used should reveal new classes of molecules that convey information about the community.
Manipulation of bacterial cell-cell signaling systems has potential use in novel antimicrobial therapies (Williams 2002; Dong et al. 2001). Enhancing growth-promotion signals of the normal microbiota at the expense of non-indigenous species might restore the normal microbial balanced state. Alternatively, specifically interrupting signaling between pathogens or giving improper signals might cripple the pathogens and make them easier to kill with standard antibiotics or by the immune system. Chemical communication between bacteria is critical for establishing and maintaining complex structured communities, such as biofilms (Davis et al. 1998; Costerton et al. 1994). Disruption of cell-cell signaling systems might provide novel opportunities for antibiotic therapy (Hentzer et al. 2003; Merritt et al. 2003; Ren et al. 2002). Furthermore, it is possible that the host recognizes and responds to bacterial signaling molecules (Chun et al. 2004), and understanding whether and how this occurs could lead to therapies for priming or boosting host defenses.
Beyond chemicals used for quorum-sensing cell-cell communication, bacteria make and release a rich variety of compounds, and enormous amounts of information could be encoded in these molecules. Bacteria probably interpret these compounds for important information about the species composition of the environment, the growth conditions, the vitality of the community, and so on. Streptomycetes are known to produce an extraordinary collection of so-called secondary metabolites (Bibb 2005). These chemicals have been and continue to be mined for those with desired activities, such as anticancer properties. The compounds are viewed as a rich storehouse of novel pharmaceuticals, but why are the bacteria making and releasing them? How do the bacteria recognize and respond to the information encoded in these chemicals? Normal human microbiota might not be as prolific in chemical production and release as the streptomycetes, but they also release complex chemical mixtures, and current understand-
ing of these molecules is inadequate. Identification and characterization of the molecules present and knowledge of how they can affect bacteria may allow researchers to begin to manipulate the chemical environment of bacteria. Such manipulation could trigger the growth of indigenous species for repopulation during or after infection or inhibit the growth or modulate the behavior of pathogens.
The committee recommends several research areas related to bacterial communication:
A-3.1 Identification of the signals and signal-transduction systems in bacterial communication.
A-3.2 Determination of whether and how the resident microbiota and pathogens communicate with host cells and respond to the immune system.
A-3.3 Development of strategies to manipulate chemical signaling in the host microbiota and pathogens.
Anecdotal information suggests that the microbiota plays an active role in defense against pathogens. For example, treatment of Salmonella infections with antibiotics that kill both salmonellae and the normal microbiota increases the longevity of the carrier state. This indicates that the normal microbiota participates in eliminating salmonellae (Neill et al. 1991). Unfortunately, as discussed above, the composition of the human normal microbiota is poorly defined (Hooper and Gordon 2001). In fact, the vast majority of the species present in normal human microbiota have never been cultivated. Enhanced knowledge of normal microbiota could be used to invent ecologically-based therapies that favor proliferation of beneficial organisms and limitation of pathogens (Baker 2005). Probiotic therapies that enhance the metabolic and signaling activities of beneficial bacteria need to be rigorously studied and tested.
Probiotic strategies aimed at ecological control, rather than at killing bacteria, could have the added benefit of lowering the spread of community-acquired drug-resistant bacteria. For example, the spread of MRSA is a major problem in communities of people who come into contact with others who are being aggressively treated for the resistant organism (Saravolatz et al. 1982; CDC 1981; Goetz et al. 1999). It can be envisioned that not
only the patients treated with traditional antibiotic therapies be followed up with probiotic therapies, but also targeted, healthy human populations. The purpose would be to reduce the distribution of antibiotic-resistant bacteria in communities that are at risk for the spread of drug-resistant bacteria (such as sports teams, jail populations, and nurses). Such probiotic therapies could extend the lifespan of traditional broad-spectrum antibiotics because their use in the healthy community essentially converts broad-spectrum drugs into narrow-spectrum ones. It should be noted that immunocompromised patients are likely to have a different response to probiotic therapies than healthy individuals, so the needs of different populations and individuals should be considered.
The committee recommends the following research areas in the use of probiotic therapies:
A-4.1 Research to determine which bacteria should be used in probiotic therapies.
A-4.2 Research to identify delivery mechanisms that would be most effective for probiotic therapies.
UNDERSTANDING BIOLOGICAL PROCESSES TO DEVISE NEW ANTIBIOTICS
Routes to anti-infective research and development can differ dramatically depending on how the goals are articulated. At first glance, there would seem to be only one criterion for a successful anti-infective chemotherapeutic: an agent with excellent pharmacologic properties that kills or inhibits disease-causing microorganisms without harm to the host. Traditionally, this has meant a broad-spectrum drug that kills a wide array of pathogens. This criterion has limited the number of suitable drugs and the strategies used to discover them.
Advances in molecular diagnostics may support greater reliance on narrow-spectrum antimicrobials and allow clinicians to make a specific diagnosis of infection before choosing a therapy. Perhaps even more important, it may be possible to explicitly exclude specific etiologic agents and obviate the use of broad-spectrum treatments in case the cause of the infection was not as originally suspected. Such technology is likely to be available for hospitalized patients thought to have infections of body compartments that
are normally sterile, such as blood and cerebrospinal fluid. The technical challenges are greater in working with samples that are normally polymicrobial, such as sputum.
The aims of new criteria for successful anti-infective chemotherapeutics are to preserve the efficacy of each agent as long as possible by delaying the emergence of drug resistance and to spare the normal microbiota as much as possible. The normal microbiota is viewed as containing invaluable allies in combating microbial pathogenesis by protecting niches against new microbial competitors and sustaining the species diversity that impedes virulence (Foster 2005) and by helping to preserve the integrity and function of epithelia and the immune system (Rakoff-Nahoum et al. 2004; Hooper et al. 2001).
With these considerations in mind, the following additional criteria for successful anti-infective chemotherapy are offered. Some agents should have a broad spectrum and others a moderate or narrow spectrum, but in each case, their use should be restricted to specifically defined human populations so that they target as few microbial species as the clinical situation warrants. Combination therapy should be encouraged, but each drug in a combination need not be required to kill microorganisms on its own.
Biological Understanding Required to Intervene in Microbial Pathogenesis
Antibiotic development has focused on the identification of “essential” targets whose inhibition is lethal under conditions of maximal microbial proliferation. A fresh approach would be to revise the operational definition of essentiality so that it more accurately reflects the biological reality: Which genes are essential to the pathogen in vitro under conditions that are relevant in the host? Which genes are essential to the pathogen in specific host environments, including polymicrobial communities on epithelial surfaces, where the microorganism of interest may represent a relatively minor planktonic population; in monomicrobial populations deep in tissues, where the pathogen may attain high population densities; and in biofilms in either kind of site? (If genes expressed in a pathogen only when it is the sole microbial species yield products that are needed for pathogenesis, targeting their products could spare the normal microbiota.) Which microbial genes are essential in combination, in such a way that the joint inhibition of their products produces synthetic lethality?
An even more drastic departure from convention is to set aside the
issue of essentiality and ask if and how bacterial entrance to or withdrawal from the cell cycle, production of virulence factors and toxins, exchange or mutation of DNA, and activation of programmed cell death can be affected. These questions should be answered in the context of understanding how bacteria communicate with others that occupy the same environment and are of the same and other species. Of particular interest are microbial decision pathways that have evolved over long periods under high selective pressures. The biology of stress responses and the DNA exchange and mutagenic DNA repair processes that contribute to antibiotic resistance may be instructive (Cirz et al. 2005). A better understanding of microbial regulatory pathways, including identification of global regulators and natural chemical signals that activate them, would also be helpful. When antibiotic producers and resisters compete in natural environments, what controls the dynamic of the relationship so that neither prevails? Finally, attention needs to be paid not just to microorganisms themselves, but also their “mobilomes”—the genomes and transcriptomes of their phages, plasmids, and integrons.
Tools needed to study these questions include in vitro culture systems with pathophysiologically-relevant concentrations of oxygen, iron, and carbon sources; biologically-relevant growth surfaces; and, in some cases, mixed microbial populations with systems-biological approaches to the analysis of prokaryotes in microbial communities.
Alternatives to Direct Killing of Microorganisms
The standard goal of eliminating disease-causing microorganisms without harming the host can be complemented by an alternative goal of simply suppressing their pathogenic behavior. In immunocompetent hosts, the latter goal may be enough to allow the immune system to control less serious infections, which account for most antibiotic prescriptions. In immunocompromised hosts or those with life-threatening infections, it may be optimal to combine a bactericidal antibiotic and an inhibitor of bacterial regulatory pathways that controls stress responses.
In any case, killing all pathogens with an extrinsic antimicrobial agent may not be necessary or even desirable. For example, the eponymous gold pigment of Staphylococcus aureus is a carotenoid that protects the bacterium from oxidative injury by host immune cells. Interruption of bacterial pigment synthesis makes the pathogen much easier for host cells to kill but does not impair bacterial growth in vitro (Liu et al. 2005). Simply limiting
the spread of the pathogen may be enough for the host microbiota to outcompete the invader.
An anti-infective agent that does not kill is less likely to select for drug resistance as the selection pressure is less intense. Identification of such agents is difficult, however, because such a target would be missed by conventional strategies that seek inhibitors of enzymes that are essential for survival of diverse species of pathogens in vitro. For example, identifying a potential new target for antibiotic development such as the synthesis of the staphylococcal pigment mentioned above that is involved solely in (a) the virulent behavior of (b) a single bacterial species (c) in vivo would probably be missed by conventional strategies that seek inhibitors of enzymes that are (a) essential for survival of (b) diverse species of pathogens (c) in vitro.
Rather than seeking inhibitors, one could seek to develop chemical agents that mimic a natural signal activating a regulatory pathway counterproductive to the pathogen. For example, an agent that selectively triggers a program of replication in a pathogen might be used in conjunction with an antibiotic that inhibits microbial protein synthesis. Protein-synthesis inhibitors are generally bacteriostatic but might be selectively bactericidal to a species whose proliferative program is induced. Combating infectious disease by interfering with microbial signaling mimics a major strategy used by the immune system.
Targeting bacterial invasion or colonization rather than bacterial survival may be a valid strategy for prophylactic intervention in specific host populations, such as those in intensive-care units and burn wards rife with drug-resistant bacteria and those in dormitories or barracks during outbreaks of meningococcal meningitis.
It is futile to target some aspects of microbial pathogenesis that are manifest only in early stages of disease, such as colonization and transepithelial invasion, if the patient is not diagnosed until the time of advanced disease.
Most traditional anti-infectives work by blocking enzymes, but we may also be able to inhibit bacterial decision pathways controlled by transcription factors that mediate protein-protein and protein-nucleic acid interactions. For example, it may be possible to exploit the dynamic turnover critical to such functions by identifying compounds that stabilize such interactions.
Strategic Approaches to Discovery of Anti-infectives for Use in Combinations
Broad new classes of targets for antibiotics have been described elsewhere (Nathan 2004). Some are enzymes involved in the synthesis of products not targeted by antibiotics in current use, such as isoprenoids (Walsh 2003a) or those involved in ATP generation (Andries et al. 2005). Others dispense with inhibition of biosynthesis altogether and focus on other aspects of the life cycle of macromolecules, such as their repair and degradation (Nathan 2004) or secretion (Brown and Wright 2005). Here, we focus on an even broader question: how to exploit synthetic lethality through combination chemotherapy. A striking example was recently reported in Mycobacterium tuberculosis through the joint disruption of two isocitrate lyases (Muñoz-Elías and McKinney 2005).
One may create a library of mutants of the pathogen in which each gene has been conditionally disrupted. Each clone in this library can be subjected to saturation signature-tagged transposon mutagenesis (Hensel et al. 1995) to generate a sublibrary. Microarray hybridization techniques can be used to identify the genes in the sublibrary that are essential for the microorganism to survive. This determination can be made in vitro under diverse conditions and ex vivo after recovery of surviving elements of the sublibrary from experimentally infected animals (Sassetti and Rubin 2003).
Another approach is to seek synthetic lethality in the disruption of a given gene in the presence of an antibiotic. For example, a workshop participant noted that there are transposon-induced null mutations in bacterial genes that impart a counterintuitive fitness increase in the presence of antibiotics. These mutations may impair a programmed bacterial cell-death response to the antibiotics. Such mutations can be introduced into the parental strain, and the new mutant can be used for saturation transposon mutagenesis to discover the genes that, when disabled, convert the antibiotic-resistant strain into an antibiotic-sensitive one.
Still another approach to synthetic lethality is to use one molecule that is already known to be an antibiotic and to screen for another that inhibits the efflux pumps that transport the antibiotic out of the cell. Such pumps limit the effectiveness of the antibiotic, thereby anticipating a possible mechanism of resistance. It may even be practical to target the small number of ATP synthases that power the pumps rather than attempt to block the pumps directly. At the same time, chemoinformatic tools are needed to
identify features of compounds that improve their likelihood of accumulating in bacteria through improved entry and diminished efflux.
In some organisms, such as M. tuberculosis, antibiotic resistance is thought to arise exclusively from heritable mutation of chromosomal genes, not from lateral gene transfer. Even in pathogens that are adept at lateral gene transfer, such as E. coli, diverse antibiotics may induce resistance to others by activating a stress response that includes error-prone DNA repair and leads to chromosomal mutagenesis (Cirz et al. 2005). Combination chemotherapy in such cases could include an inhibitor of the critical errorprone translesional DNA polymerase with dual goals: to kill the pathogen more effectively and to block emergence of drug resistance in surviving cells (Boshoff et al. 2003).
Most radically, it may sometimes be possible to include the product of a host gene as one target in an effective combination therapy. Host-gene products serve as receptors for microorganisms and their toxins, for the trafficking of intracellular pathogens to replicative niches, and other processes that contribute to pathogenesis. What makes the inclusion of a temporarily nonessential host-gene product attractive as a target for combination therapy is that, unlike the microorganism, the host will not mutate to a resistant state. Examples of targets of host rather than microbial origin are the receptors for toxins of Bacillus anthracis; POSH, a ubiquitin protein ligase (Alroy et al. 2005) and ATM, a protein kinase (Lau et al. 2005), both required for replication of HIV; CCR5/CXCR4 antagonists to limit the replication of HIV (Princen et al. 2004); ErbB, a protein kinase required for replication of vaccinia virus (Yang et al. 2005); and Abl, a protein kinase required for dissemination of vaccinia virus (Reeves et al. 2005).
In the area of identifying targets, the committee recommends the following:
A-5.1 Screening for antibiotics in vitro under conditions that are relevant in the host.
A-5.2 Searching for agents that exhibit synthetic lethality.
A-5.3 Learning enough about bacterial behavior to influence actions crucial to disease and drug resistance, for example, by mimicking or blocking bacterial signals.
A-5.4 Exploring inhibition of host targets to thwart infectious disease.
NEED FOR NEW MOLECULES
New Molecules: Natural and Synthetic
Given the long gap in the introduction of new structural classes of antibiotics—38 years between streptogramins in 1962 and linezolid in 2000 (Walsh 2003a,b)—and the inexorable development of resistance to a given antibiotic once it is in widespread clinical use, there is a pressing and recurrent need for new molecules with antibiotic properties. Historically, two lines of discovery have been fruitful: natural products with antibiotic activity and synthetic antibacterial agents. Penicillins, cephalosporins, vancomycin, tetracycline, and aminoglycosides are in the first category, and fluoroquinolones, sulfonamides, and oxazolidinones are in the second. Natural and synthetic molecules are both likely to remain important sources for new antibiotics but offer distinct challenges.
Mining the Natural World: Discover, Diversify, and Deliver
Microorganisms themselves have been the richest source of antibiotics; with 99% of the known microbial species as yet uncultured, these clearly are untapped sources of novel molecules. Realization of their potential requires attention to new techniques for microbial cultivation—including consortia—and exploration of new biological niches of microorganisms. Each time a new set of biological microenvironments has been accessed by chemists, new classes of bioactive natural products have been isolated and structures determined (Koehn and Carter 2005; Clardy and Walsh 2004). A recent example is the isolation of abyssomicin C, which inhibits folate biosynthesis in MRSA; the molecule is produced by the rare actinomycete Verrucosispora collected in a sediment sample from the Japanese sea at a depth of 867 ft (Bister et al. 2004). The emerging field of metagenomics, in which the pooled genetic material of a bacterial community is sequenced without cultivating each individual member, offers the possibility of identifying novel product and biosynthetic pathways (Handelsman 2005).
Although currently unknown or unculturable bacteria are likely to be the source of novel bioactive molecules, even bacteria that are routinely grown in the laboratory may have additional capabilities that are not expressed under standard laboratory conditions. There should be an emphasis on deciphering previously unrecognized secondary metabolite pathways in bacterial genomes and in eliciting the production of end products from
metabolic pathways that are cryptic when producer organisms are grown in standard culture conditions. For example, only three of the predicted 25 polyketide synthase biosynthetic gene clusters in Streptomyces avermitilis have been shown to be active (Omura et al. 2001). Similarly, several sets of antibiotic gene clusters in Bacillus subtilis are expressed only in starvation-induced conditions not generally present in laboratory cultures (Stein 2005).
Continued, even increased, attention should be paid to natural strategies used by prokaryotes and eukaryotes against microorganisms. Hosts respond to microbial infection by secretion of peptidic molecules, such as defensins, to act locally. Narrow-spectrum protein toxins (bacteriocins) are a predominant strategy in natural microbial communities for killing neighboring strains; for example, bacteriocidic microcins secreted by one strain of E. coli kill neighboring strains selectively but have minimal effect on the microbial community. This effect—selectively removing pathogens while leaving commensal bacteria unharmed—would be a desired feature of antibiotics. Efforts should be focused on delivery strategies for peptides, bacteriocins, and phage-based lytic proteins (Yoong et al. 2004).
Bacteria produce bioactive molecules through a series of biosynthetic steps. The idea behind combinatorial biosynthesis is to break the biosynthetic pathways down into modules and combine the modules in a well-characterized host to generate novel end products. The ability to express biosynthetic-pathway genes in heterologous hosts will be required for efficient combinatorial biosynthesis. Ultimately, full realization of the potential of combinatorial biosynthesis will require engineering bacteria that can make the monomeric building blocks that are required for running the assembly lines and expressing a full range of post-assembly-line tailoring enzymes. Ideally, the ability to shuffle protein domains and modules in the polyketide synthase and nonribosomal peptide synthetase pathways and to engineer intersections with terpenoid and other pathways to merge other chemical frameworks onto polyketide and peptide scaffolds will be needed for maximizing natural diversity (Walsh 2004).
Finally, further investigation of the molecular biology underlying bacterial cell death is needed. In most cases, there is little understanding of why bacteria that are susceptible to particular antibiotics die or which traditional targets hold more promise for the development of bactericidal antibiotics. Bacteria, like eukaryotes, may have mechanisms of programmed cell death. If so, the genetic regulatory programs and the biochemical processes associated with the triggering of cell death could provide new targets
for antibiotic development (Engelberg-Kulka et al. 2004). Current understanding of bacterial cell death lags behind understanding of this process in higher organisms.
To maximize the discovery of novel natural molecules and to increase the ability to generate variations on natural molecules, the committee recommends the following:
A-6.1 Increased sampling in diverse environments and increased application of the techniques of metagenomics to identify bioactive compounds produced by currently unknown and uncultured microorganisms.
A-6.2 The development of novel and varied culture conditions to identify cryptic metabolic pathways in currently cultivated strains.
A-6.3 Increased research on the role of host-derived antimicrobial peptides, phage lytic proteins, and bacteriocins in the ecology of host-bacteria interactions to improve delivery strategies for these natural products.
A-6.4 Increased research on combinatorial biosynthesis to allow the most varied possible uses of the novel biosynthetic pathways found in known and unknown organisms.
A-6.5 Increased research on bacterial cell death, including investigation of programmed cell death and how antibacterials kill to exploit new strategies for the elimination of pathogens.
Developing Synthetic Molecules: Diversity, Bioactivity, and Specificity
In addition to discovering and elaborating on the bioactive compounds made by bacteria themselves, the design of synthetic antibiotic molecules should also be pursued. The committee identified three kinds of research that would contribute to greater success in the design of synthetic antibiotics: developing techniques that make it easier to generate diversity in synthetic molecules, increasing understanding of the characteristics that allow molecules to enter and remain in cells, and developing the ability to move beyond using growth inhibition as the measure of a compound’s activity and achieving a more sophisticated understanding of how compounds affect metabolism.
One of the principles of synthetic-molecule construction is modularity, with variable shape and architecture of modular cores; linkers or con-
nectors that are also variable in shape, length, and polarity; and surface functional groups in each module that allow rapid elaboration. A goal of such modularity is to build on initial leads and elaborate them in any direction for optimization. A further goal would be to minimize blunt ends—scaffold elements that prevent expansion—in any modular array of synthetic molecules. Two promising techniques for generating greater diversity in synthetic-molecular construction are “click chemistry” and programming small molecules genetically.
The “click chemistry” paradigm elaborated by Sharpless and colleagues (Kolb et al. 2001)—such as the coupling of azides and alkynes with copper catalysis in aqueous solution under mild conditions—is a leading example of rapid modular combinatorial chemistry. In favorable cases, the shape of the binding pocket in a target protein can guide covalent couplings into complementarily shaped molecules (Manetsch et al. 2004). The binding of small molecule fragments capable of self-assembly in cavities of target proteins may become a generalizable strategy to produce small-molecule architecture complementary to and with high affinity for target bacterial proteins.
Another promising direction for antimicrobial chemical libraries is genetically programmable small molecules (Li and Liu 2004; Halpin et al. 2004; Halpin and Harbury 2004a,b). DNA tethering can allow enhanced adjacency to promote new, high-yield chemistry and the creation of large libraries from which molecules can be selected for function. In a screen for nanomolar binders to a target bacterial protein, 105 promising molecules could be obtained from very large (for example, 1012 entries) programmable libraries; these 100,000 winners could be put through a further series of functional screens (such as for whole-cell activity) with the prospect that there would be many with high activity for further structural optimization. A key advantage of the DNA-directed programmable approach is the opportunity to evolve molecules to optimize a selected function.
However, the ability to generate diverse molecules will not be enough. A major bottleneck in drug design and evaluation is the optimization process for turning hits into molecules that will work in the host. What is needed is the ability to move smoothly back and forth between changing chemical structure and activity in the host; computational prediction and rapid preparation of related families of molecules must be integrated with pharmacokinetic measurements. Computational biology and systems biology must become central to the evaluation of new molecular scaffolds in infected animals and humans to predict the safety and efficacy of new classes of molecules.
To increase the ability to generate synthetic molecules that are not only diverse but also bioactive in a predictable way, the committee recommends the following:
A-7.1 Development of small-molecule libraries customized for bacterial targets.
A-7.2 Increased research on DNA-directed synthesis of diverse collections of small molecules for screening and selection against bacterial targets.
A-7.3 Greater emphasis on projects that systematically relate chemical structure to biological activity.
Mining Historical Knowledge
Many pharmaceutical and biotechnology companies that have a history of pursuing the identification of new antibiotics probably have files on the development and testing of molecules that could be mined for promising leads. It might also be valuable to interview those who did the work. The workshop and committee members’ experiences have brought forth anecdotes about valuable drugs nearly abandoned because of shifts in corporate policy but then saved by the conviction or special insight of one investigator. Interviews and data mining of now-ignored records might reveal promising molecules on which much work has already been conducted, but that were dropped for reasons unrelated to their efficacy. Other molecules may have encountered obstacles that can now be circumvented by new technology or understanding. Such a historical and biographical approach is unusual in drug development and in biomedical research, but may reinvigorate now-ignored research that showed promise.
Screening Issues: How to Find Functional Properties in Candidate Molecules
If it becomes possible to generate a multitude of diverse molecules, the next challenge is to improve the ability to screen them for antibiotic potential. Cell-based screening, in which the ability of compounds to kill bacteria growing in ideal, monoculture conditions is tested, has the advantage of identifying compounds with the right physical properties to penetrate and persist in cells (and affect their growth). However, because the targets are unknown, it is difficult to predict and test how the activity of the compounds could be enhanced. In contrast, target-based screening identifies
compounds that bind to or inhibit a bacterial target that is believed to be essential for bacterial survival. It has proved challenging to endow compounds that have the desired activity with the ability to penetrate cells and reach their targets. Both target- and cell-based approaches have value, but fresh approaches are needed because of the weaknesses inherent in each (Brown and Wright 2005).
For example, in cell-based screens, it is important to collect more information beyond the single criterion of growth inhibition so that it will be possible to characterize a molecule’s activity and narrow its possible targets (Schreiber 2005). Screens that provide detailed information about intermediate states of bacterial cell perturbation, including gene chip arrays and metabolome profiling, would be valuable. Validation of these screens on the dozens of existing classes of antibiotics would provide the beginning of a comprehensive database. Inclusion of both pathogenic and commensal bacteria in high-density screening arrays would lend a systems-biology perspective and build the detailed resolution to identify sites of compound action. A few hundred antimicrobial small molecules would become standard probes to provide response patterns with which new candidate molecules could be compared.
In target-based screens, greater emphasis needs to be placed on increasing the likelihood of success in cell penetration and persistence of active molecules. The physical-chemical characteristics of cell permeability are poorly understood. Further research is needed to characterize the functional properties of molecules that minimize interaction with bacterial-membrane efflux pumps and allow penetration and persistent accumulation in pathogenic bacteria. The high-density screening techniques developed to improve cell-based screening could also be used to screen for these characteristics. Such efforts may help augment traditional and current chemical libraries to have a greater representation of molecules likely to be successful antibiotics and provide the data to optimize antibiotic-like molecules in future libraries.
As an example of the value of combining cell- and target-based approaches in the development of novel antimicrobials, a narrow-spectrum diarylquinolone, acting against mycobacteria by inhibition of the F0 subunit of ATP synthase (Andries et al. 2005), has recently been discovered by medicinal-chemistry optimization in whole-cell killing assays that used the fast-growing M. smegmatis as an initial M. tuberculosis surrogate. The mode of action was determined by whole-genome sequencing of resistant organ-
isms and then pharmacokinetics in mice optimized before an initial phase I trial for safety and tolerability in humans. For microbial cell-based killing assays as the starting point, this study may become a paradigm of strategic execution.
Both screening approaches would benefit from a publicly available collection of molecules that have been shown to have antimicrobial activity. Many such molecules have been reported over the last 6 decades of antibiotic research but may be languishing in the private coffers of companies not actively developing antimicrobials. Molecules could be collected from those companies or resynthesized. Another avenue worth pursuing would be donation of such molecules, especially naturally occurring ones, by pharmaceutical or biotechnology companies. The establishment of a collection of active antimicrobial compounds, numbering in the thousands, would represent a precious archive available to the research community for information-rich screens. Such a collection might also serve as small molecule-based microarrays for target-based screening.
Finally, measurements of the efficacy of novel therapeutics must use assays that closely mimic in vivo conditions. Traditional tests for antibiotic effectiveness rely on in vitro assays typically of single-species bacterial cultures grown under standardized conditions in defined laboratory media. These research assays do not effectively mimic the environments that bacteria experience in a host and although positive outcomes reliably occur in the laboratory, they are not always good predictors of outcomes in vivo. Treatment failure due to phenotypic tolerance needs to be examined so that therapies that avoid problems associated with noninherited resistance can be developed.
To improve the identification and characterization of bioactive compounds, the committee recommends the following:
A-8.1 Development of cell-based screening techniques that collect detailed information on cell metabolism through gene arrays, metabolome profiling, and other measurements.
A-8.2 Increased research on the chemical properties necessary for cell permeability and retention.
A-8.3 Establishment of a publicly available collection of molecules that have antibiotic activity.
A-8.4 Development of new assays that mimic in vivo conditions.
HOW CAN THIS WORK BE CARRIED OUT?
Not only new scientific approaches, but also legislative and institutional actions may be required to improve the prospects for development of new antimicrobial agents and prolongation of their efficacy. The introduction of three bills in the U.S. Congress bearing on antibiotic use and development suggests that major changes in society’s approach to antibiotics may be on the horizon (Nathan and Goldberg 2005). While it was beyond the scope of the committee to consider non-scientific matters, the participants in the workshop noted that the scientific, regulatory and economic aspects of the development of antimicrobials are extensively interconnected (Nathan 2004).
Abraham, E.P., and Chain, E. (1988) An enzyme from bacteria able to destroy penicillin. 1940. Rev. Infect. Dis. 10(4), 677-8.
Alroy, I., Tuvia, S., Greener, T., Gordon, D., Barr, H.M., Taglicht, D., Mandil-Levin, R., Ben-Avraham, D., Konforty, D., Nir, A., Levius, O., Bicoviski, V., Dori, M., Cohen, S., Yaar, L., Erez, O., Propheta-Meiran, O., Koskas, M., Caspi-Bachar, E., Alchanati, I., Sela-Brown, A., Moskowitz, H., Tessmer, U., Schubert, U., and Reiss, Y. (2005) The trans-Golgi network-associated human ubiquitin-protein ligase POSH is essential for HIV type 1 production. Proc. Natl. Acad. Sci. USA 102, 1478-83.
Andries, K., Verhasselt, P., Guillemont, J., Göhlmann, H.W.H., Neefs, J.-M., Winkler, H., Van Gestel, J., Timmerman, P., Zhu, M., Lee, E., Williams, P., de Chaffoy, D., Huitric, E., Hoffner, S., Cambau, E., Truffot-Pernot, C., Lounis, N., and Jarlier, V. (2005) A diarylquinoline drug active on the ATP synthase of Mycobacterium tuberculosis. Science 307, 223-7.
Armstrong, G.L., Conn, L.A., and Pinner, R.W. (1999) Trends in infectious disease mortality in the United States during the 20th century. J. Am. Med. Assoc. 281(1), 61-6.
Baker, M. (2005) Better living through microbes. Nat. Biotechnol. 23, 645-7.
Bibb, M.J. (2005) Regulation of secondary metabolism in streptomycetes. Curr. Opin. Microbiol. 8, 208-15.
Bister, B., Bischoff, D., Ströbele, M., Riedlinger, J., Reicke, A., Wolter, F., Bull, A.T., Zähner, H., Fiedler, H.-P., and Süssmuth, R.D. (2004) Abyssomicin C—a polycyclic antibiotic from a marine Verrucosispora strain as an inhibitor of the p-aminobenzoic acid/ tetrahydrofolate biosynthesis pathway. Agnew. Chem. Int. Ed. Engl. 43(19), 2574-6.
Boshoff, H.I., Reed, M.B., Barry, C.E., 3rd, and Mizrahi, V. (2003) DnaE2 polymerase contributes to in vivo survival and the emergence of drug resistance in Mycobacterium tuberculosis. Cell 113, 183-93.
Brown, E.D., and Wright, G.D. (2005) New targets and screening approaches in antimicrobial drug discovery. Chem. Rev. 105, 759-74.
Centers for Disease Control and Prevention (CDC). (1981) Community-acquired methicillin-resistant Staphylococcus aureus infections—Michigan. Morb. Mortal. Wkly. Rep. (MMWR) 30, 185-7.
Chater, K.F., and Horinouchi, S. (2003) Signalling early developmental events in two highly diverged Streptomyces species. Mol. Microbiol. 48(1), 9-15.
Chen, X., Schauder, S., Potier, N., Van Dorsselaer, A., Pelczer, I., Bassler, B.L., and Hughson, F.M. (2002) Structural identification of a bacterial quorum-sensing signal containing boron. Nature 415, 545-9.
Chun, C.K., Ozer, E.A., Welsh, M.J., Zabner, J., and Greenberg, E.P. (2004) Inactivation of a Pseudomonas aeruginosa quorum-sensing signal by human airway epithelia. Proc. Natl. Acad. Sci. USA 101, 3587-90.
Cirz, R.T., Chin, J.K., Andes, D.R., Crecy-Lagard, V.D., Craig, W.A., and Romesberg, F.E. (2005) Inhibition of mutation and combating the evolution of antibiotic resistance. PLoS Biol. 3, e176.
Clardy, J., and Walsh, C.T. (2004) Lessons from natural molecules. Nature 432, 829-37.
Cohen, M.L. (2000) Changing patterns of infectious disease. Nature 406(6797), 762-7.
Costerton, J.W., Lewandownski, Z., DeBeer, D., Caldwell, D., Korber, D., and James, G. (1994) Biofilms, the customized microniche. J. Bacteriol. 176(8), 2137-42.
Davies, D.G., Parsek, M.R., Pearson, J.P. Iglewski, B.H., Costerton, J.W., and Greenberg, E.P. (1998) The involvement of cell-to-cell signals in the development of a bacterial biofilm. Science 280(5361), 295-8.
Donabedian, H. (2003) Quorum sensing and its relevance to infectious diseases. J. Infect. 46, 207-14.
Dong, Y.H., Wang, L.H., Xu, J.L., Zhang, H.B., Zhang, X.F., and Zhang, L.H. (2001) Quenching quorum-sensing-dependent bacterial infection by an N-acyl homoserine lactonase. Nature 411(6839), 813-7.
Eckburg, P.B., Bik, E.M., Bernstein, C.N., Purdom, E., Dethlefsen, L., Sargent, M., Gill, S.R., Nelson, K.E., and Relman, D.A. (2005) Diversity of the human intestinal microbial flora. Science 308(5728), 1635-8.
Engelberg-Kulka, H., Sat, B., Reches, M., Amitai, S., and Hazan, R. (2004) Bacterial programmed cell death systems as targets for antibiotics. Trends Microbiol. 12, 66-71.
Finch, R.G., Greenwood, D., Norrby, S.R., and Whitley, R.J., eds. (2003) Antibiotic and Chemotherapy, 8th ed. Edinburgh: Churchill Livingstone.
Foster, K.R. (2005) Biomedicine. Hamiltonian medicine: why the social lives of pathogens matter. Science 308, 1269-70.
Goetz, A., Posey, K., Fleming, J., Jacobs, S., Boody, L., Wagener, M.M., and Muder, R.R. (1999) Methicillin-resistant Staphylococcus aureus in the community: a hospital-based study. Infect. Control Hosp. Epidemiol. 20(10), 689-91.
Halpin, D.R., and Harbury, P.B. (2004a) DNA display I. Sequence-encoded routing of DNA populations. PLoS Biol. 2(7), e173.
Halpin, D.R., and Harbury, P.B. (2004b) DNA display II. Genetic manipulation of combinatorial chemistry library for small-molecule evolution. PLoS Biol. 2(7), e174.
Halpin, D.R., Lee, J.A., Wreen, S.J., and Harbury, P.B. (2004) DNA display III. Solid-phase organic synthesis on unprotected DNA. PLoS Biol. 2(7), e175.
Handelsman, J. (2005) How to find new antibiotics. The Scientist 19(19, 10 October), 20.
Hensel, M., Shea, J.E., Gleeson, C., Jones, M.D., Dalton, E., and Holden, D.W. (1995) Simultaneous identification of bacterial virulence genes by negative selection. Science 269, 400-3.
Hentzer, M., Wu, H., Andersen, J.B., Riedel, K., Rasmussen, T.B., Bagge, N., Kuman, N., Schembri, M.A., Song, Z., Kristoffersen, P., Manefield, M., Costerton, J.W., Molin, S., Eberl, L., Steinberg, P., Kjelleberg, S., Hoiby, N., and Givskov, M. (2003) Attenuation of Pseudomonas aeruginosa virulence by quorum sensing inhibitors. EMBO J. 22(15), 3803-15.
Hooper, L.V., and Gordon, J.I. (2001) Commensal host-bacterial relationships in the gut. Science 292, 1115-8.
Hooper, L.V., Wong, M.H., Thelin, A., Hansson, L., Falk, P.G., and Gordon, J.I. (2001) Molecular analysis of commensal host-microbial relationships in the intestine. Science 291, 881-4.
Koehn, F.E., and Carter, G.T. (2005) The evolving role of natural products in drug discovery. Nat. Rev. Drug Discov. 4(3), 206-20.
Kolb, H.C., Finn, M.G., and Sharpless, K.B. (2001) Click chemistry: diverse chemical function from a few good reactions. Agnew. Chem. Int. Ed. 40(11), 2004-21.
Lau, A., Swinbank, K.M., Ahmed, P.S., Taylor, D.L., Jackson, S.P., Smith, G.C., and O’Connor, M.J. (2005) Suppression of HIV-1 infection by a small molecule inhibitor of the ATM kinase. Nat. Cell. Biol. 7, 493-500.
Lazazzera, B.A., and Grossman, A.D. (1998) The ins and outs of peptide signaling. Trends Microbiol. 6, 288-94.
Levy, S.B., and Marshall, B. (2004) Antibacterial resistance worldwide: causes, challenges and responses. Nat. Med. 10(12s), S122-9.
Li, X., and Liu, D.R. (2004) DNA-templated organic synthesis: nature’s strategy for controlling chemical reactivity applied to synthetic molecules. Agnew. Chem. Int. Ed. 43(37), 4848-70.
Liu, G.Y., Essex, A., Buchanan, J.T., Datta, V., Hoffman, H.M., Bastian, J.F., Fierer, J., and Nizet, V. (2005) Staphylococcus aureus golden pigment impairs neutrophil killing and promotes virulence through its antioxidant activity. J. Exp. Med. 202(2), 209-15.
Manetsch, R., Krasinksi, A., Radic, Z., Raushel, J., Taylor, P., Sharpless, K.B., and Kolb, H.C. (2004) In situ click chemistry: enzyme inhibitors made to their own specifications. J. Am. Chem. Soc. 126(40), 12809-18.
Matti, S.N., Phillips, O.A., Micetich, R.G., and Livemore, D.M. (1998) β-lactamase inhibitors: Agents to overcome bacterial resistance. Curr. Med. Chem. 5, 441-56.
Merritt, J., Qi, F., Goodman, S.D., Anderson, M.H., and Shi, W. (2003) Mutation of luxS affects biofilm formation in Streptococcus mutans. Infect. Immun. 71, 1972-9.
Miller, M.B., and Bassler, B.L. (2001) Quorum sensing in bacteria. Annu. Rev. Microbiol. 55, 165-99.
Miller, S.T., Xavier, K.B., Campagna, S.R., Taga, M.E., Semmelhack, M.F., Bassler, B.L., and Hughson, F.M. (2004) Salmonella typhimurium recognizes a chemically distinct form of the bacterial quorum-sensing signal AI-2. Mol. Cell. 15(5), 677-87.
Muñoz-Elías, E.J., and McKinney, J.D. (2005) Mycobacterium tuberculosis isocitrate lyases 1 and 2 are jointly required for in vivo growth and virulence. Nat Med. 11(6), 638-44.
Nataro, J.P., Cohen, P.S., Mobley, H.L.T., and Weiser, J.N., eds. (2005) Colonization of Mucosal Surfaces. Washington, DC: ASM Press.
Nathan, C. (2004) Antibiotics at the crossroads. Nature 431, 899-902.
Nathan, C., and Goldberg, F.M. (2005) The profit problem in antibiotic R&D. Nat. Rev. Drug Discov. 4(11), 887-91.
Neill, M.A., Opal, S.M., Heelan, J, Giusti, R., Cassidy, J.E., White, R., and Mayer, K.H. (1991) Failure of cioprofloxacin to eradicate convalescent fecal excretion after acute salmonellosis: experience during an outbreak in health care workers. Ann. Intern. Med. 114(3), 195-9.
O’Brien, T.F., Eskildsen, M.A., and Stelling J.M. (2001) Using Internet discussion of antimicrobial susceptibility databases for continuous quality improvement of the testing and management of antimicrobial resistance. Clin. Infect. Dis. 33(Suppl. 3), S118-23.
Omura, S., Ikeda, H., Ishikawa, J., Hanamoto, A., Takahashi, C., Shinose, M., Takahashi, Y., Horikawa, H., Nakazawa, H., Osonoe, T., Kikuchi, H., Shiba, T., Sakaki, Y., and Hattori, M. (2001) Genome sequence of an industrial microorganism Streptomyces avermitilis: Deducing the ability of producing secondary metabolites. Proc. Natl. Acad. Sci. USA 98(21), 12215-20.
Parsek, M.R., Val, D.L., Hanzelka, B.L., Cronan, J.E., Jr., and Greenberg, E.P. (1999) Acyl homoserine-lactone quorum-sensing signal generation. Proc. Natl. Acad. Sci. USA 96, 4360-5.
Princen, K., Hatse, S., Vermeire, K., Aquaro, S., De Clercq, E., Gerlach, L.-O., Rosenkilde, M., Schwartz, T.W., Skerlj, R., Bridger, G., and Schols, D. (2004) Inhibition of human immunodeficiency virus replication by a dual CCR5/CXCR4 antagonist. J. Virol. 78(23), 12996-13006.
Projan, S.J. (2003) Why is big Pharma getting out of antibacterial drug discovery? Curr. Opin. Microbiol. 6, 427-30.
Rakoff-Nahoum, S., Paglino, J., Eslami-Varzaneh, F., Edberg, S., and Medzhitov, R. (2004) Recognition of commensal microflora by toll-like receptors is required for intestinal homeostasis. Cell 118, 229-41.
Reeves, P.M., Bommarius, B., Lebeis, S., McNulty, S., Christensen, J., Swimm, A., Chahroudi, A., Chavan, R., Feinberg, M.B., Veach, D., Bornmann, W., Sherman, M., and Kalman, D. (2005) Disabling poxvirus pathogenesis by inhibition of Abl-family tyrosine kinases. Nat. Med. 11(7), 731-9.
Ren, D., Sims, J.J., and Wood, T.K. (2002) Inhibition of biofilm formation and swarming of Bacillus subtilis by (5Z)-4-bromo-5-(bromomethylene)-3-butyl-2(5H)-furanone. Lett. Appl. Microbiol. 34, 293-9.
Saravolatz, L.D., Markowitz, N., Arking, L., Pohlod, D., and Fisher, E. (1982) Methicillin-resistant Staphylococcus aureus. Epidemiological observations during a community-acquired outbreak. Ann. Intern. Med. 96(1), 11-16.
Sassetti, C.M., and Rubin, E.J. (2003) Genetic requirements for mycobacterial survival during infection. Proc. Natl. Acad. Sci. USA 100, 12989-94.
Schreiber, S.L. (2005) Small molecules: the missing link in the central dogma. Nat. Chem. Biol. 1, 64-6.
Shlaes, D.M. (2003) The abandonment of antibacterials: why and wherefore? Curr. Opin. Pharmacol. 3, 470-3.
Stein, T. (2005) Bacillus subtilis antibiotics: structure, syntheses and specific functions. Mol. Microbiol. 56(4), 845-57.
Stelling, J.M., Travers, K., Jones, R.N., Turner, P.J., O’Brien, T.F., and Levy, S.B. (2005) Integrating Escherichia coli antimicrobial susceptibility data from multiple surveillance programs. Emerg. Infect. Dis. 11, 873-82.
Templeton, K.E., Scheltinga, S.A., van den Eeden, W.C.J.F.M., Graffelman, A.W., van den Broek, P.J., and Claas, E.C.J. (2005) Improved diagnosis of the etiology of community-acquired pneumonia with real-time polymerase chain reaction. Clin. Infect. Dis. 41(3), 345-51.
Thomas, L. (1974) The Lives of a Cell. New York: Viking Press.
Walsh, C. (2003a) Where will new antibiotics come from? Nat. Rev. Microbiol. 1, 65-70.
Walsh, C.T. (2003b) Antibiotics: Actions, Origins, Resistance. Washington, DC: ASM Press.
Walsh, C.T. (2004) Polyketide and nonribosomal peptide antibiotics: modularity and versatility. Science 303, 1805-10.
Williams, P. (2002) Quorum sensing: an emerging target for antibacterial chemotherapy? Expert Opin. Ther. Targets 6, 257-74.
Wilson, M. (2005) Microbial Inhabitants of Humans: Their Ecology and Role in Health and Disease. Cambridge, UK: Cambridge University Press.
Wright, G.D. (2000) Resisting resistance: new chemical strategies for battling superbugs. Chem. Biol. 7, R127-32.
Yang, H., Kim, S.-K., Kim, M., Reche, P.A., Morehead, T.J., Damon, I.K., Welsh, R.M., and Reinherz, E.L. (2005) Antiviral chemotherapy facilitates control of poxvirus infections through inhibition of cellular signal transduction. J. Clin. Invest. 115, 379-87.
Yoong, P., Schuch, R., Nelson, D., and Fischetti, V.A. (2004) Identification of a broadly active phage lytic enzyme with lethal activity against antibiotic-resistant Enterococcus faecalis and Enterococcus faecium. J. Bacteriol. 186(14), 4808-12.