Introduction
While existing paradigms of clonal infection passed down from Robert Koch (1884) have been valuable in treating acute and epidemiologic infections, these standards have begun to shift with respect to chronic, persistent infections. Chronic infections are typically polyclonal (same species, different strains) or polymicrobial (different species), which can be a mechanism for persistence (
Ehrlich et al., 2005). Genetic and phenotypic diversity allows an infecting community of microbes to persist in the face of the host immune response (
Leid et al., 2002;
Jesaitis et al., 2003;
Leid et al., 2005;
Leid et al., 2009) and antimicrobial therapy (
Xu et al., 1998;
Keren et al., 2004a,
2004b;
Roberts and Stewart, 2004;
Boles et al., 2005;
Ehrlich et al., 2005;
Ehrlich et al., 2008;
Hu and Ehrlich, 2008). Despite significant efforts to understand the composition of multi-microbial communities, it is clear that the complexity of polymicrobial infections has been underestimated (
Hall-Stoodley et al., 2006;
Stephenson et al., 2010). This review highlights the importance of understanding the community interactions among species (and higher taxa) that influence pathogenicity.
Polymicrobial communities of bacteria, fungi, and viruses thrive in nearly every niche in the human body. These communities are essential for the health of the host. However, a perturbation of the human environment through introduction of pathogens, environmental factors, or antibiotic overuse, can result in a shift from a healthy community to one that is pathogenic. Pathogenicity is not only dependent on the specific species within a community, but also their relative abundances (
Jenkinson and Lamont, 2005;
Ehrlich et al., 2008). Increased abundance of pathogens, or the decreased abundance of their protective commensal counterparts, can disrupt homeostasis and provide a platform for disease.
As a result of phenotypic heterogeneity and genome plasticity, polymicrobial infections are inherently persistent and difficult to treat using traditional therapies. Therefore, an understanding of the mechanisms that drive persistence and novel virulence factor production is necessary. While it is understood that many chronic infections are caused by matrix-enclosed biofilms, there are a number of interactions that occur as a result of the biofilm mode of growth that require further study. Biofilms provide an ideal setting for horizontal gene transfer (HGT), which can result in acquisition and expression of novel virulence genes (
Maeda et al., 2006;
Tribble et al., 2007;
Ehrlich et al., 2010). Ehrlich and colleagues recently put forth the distributed genome hypothesis, which postulates that a polyclonal (or polymicrobial) population of bacteria can undergo HGT to generate diversity that allows the community to persist in the face of the host immune system and antimicrobial therapy (
Ehrlich et al., 2010). Additional interactions occur within a biofilm community, including quorum sensing, physical contact, and metabolite perception that can influence the community behavior at the population level. Population level virulence factors are distinct from those expressed by individual bacterial cells since they require cooperation (and sometimes antagonism) from the microbial community (
Hu and Ehrlich, 2008). Examples of population level virulence include biofilm formation, horizontal gene transfer, quorum sensing and two component regulatory systems (TCRS) that otherwise alter expression of virulence factors expressed at the individual bacterial cell level (
Hu and Ehrlich, 2008). An increased understanding of how these factors are generated and/or regulated can lead to development of novel therapeutics that target those microorganisms that are responsible for disease while not harming the healthy microbial community. Here, we review recent efforts to understand polymicrobial infection with a focus on respiratory and diseases of the ear, nose and oral cavity, with a focus on the chronic inflammatory disease, rhinosinusitis. Other recent reviews have been published on another important polymicrobial disease, chronic wounds (
Bader, 2008;
Burmolle et al., 2010;
Wolcott and Dowd, 2011), which are not discussed here.
Chronic disease associated with polymicrobial communities
Oral polymicrobial communities
While the oral cavity harbors a diverse array of microorganisms, the past 30 years have focused on molecular studies of the bacterial ecology of the mouth. It is now estimated that over 750 bacterial species are part of the healthy and diseased oral flora (
Dewhirst et al., 2010). Of these taxa, greater than 50% of the oral microbial species have yet to be cultivated. In the oral community, surface receptors and metabolic cooperation mediate the interactions among microbes in a process termed co-aggregation (
Kolenbrander, 2000;
Kolenbrander et al., 2002;
Palmer et al., 2003;
Jakubovics and Kolenbrander, 2010). Early oral communities typically consist of facultative anaerobic species such as
Streptococcus,
Actinomyces,
Veillonella, and
Neisseria (
Nyvad and Kilian, 1987;
Palmer et al., 2003;
Li et al., 2004). Mid to late colonizers primarily include anaerobic species
Fusobacterium nucleatum and
Porphryomonas gingivalis, among others (
Periasamy and Kolenbrander, 2009;
Jakubovics and Kolenbrander, 2010). Interbacterial interactions mediate the colonization and virulence of
P. gingivalis, and can ultimately lead to periodontal disease. For example,
P. gingivalis is more virulent in a mouse model when it is part of a community with
Actinobacillus actinomycetemcomitans,
F. nucleatum,
Tannerella forsythia, or
Treponema denticola (
Chen et al., 1996;
Feuille et al., 1996;
Kesavalu et al., 1998). While the oral mucosa has provided an important view into the structure and function of complex microbial communities, the past decade of research has revealed several other chronic diseases mediated by pathogenic communities of bacteria and fungi.
Cystic fibrosis
The genetic disease cystic fibrosis (CF) is characterized by recurrent and chronic infections in the lung. There is a wealth of knowledge of the CF lung microflora, however co-infection is often underreported due to the focus on
P. aeruginosa mediated infection (
Lee et al., 2005;
Stone and Saima, 2007;
Bjarnsholt et al., 2010;
Høiby et al., 2010;
Park et al., 2011). Colonization of the human lung occurs in phases. Early colonizers include
Hemophilus influenzae and
Staphylococcus aureus (Stone and Saima, 2007).
Stenotrophomonas maltophilia,
Burkholderia cepacia complex, and
Pseudomonas aeruginosa are typically late-colonizers of CF patients (
Klinger and Thomassen, 1985). Several other species colonize the CF lung including
Achromobacter xylosoxidans (Lambiase et al., 2011), non-tuberculoid mycobacteria (
Brown, 2010),
Pandoraea spp. (
Moore et al., 2002),
Ralstonia spp. (
Stelzmueller et al., 2006), and
Inquilinus spp. (
Wellinghausen et al., 2005). The implication of these species in morbidity and mortality is poorly understood.
In CF patients,
P. aeruginosa infection peaks around 18-24 years of age (CF Foundation National Patient Registry, 2004) and is responsible for much of the morbidity and mortality in CF. There is a negative association between
S. aureus infection and
P. aeruginosa, largely attributed to
P. aeruginosa production of the antagonistic molecule 4-hydroxy-2-heptylquinolone-
N-oxide (HQNO) (
Hoffman et al., 2006). Although co-infection in the lung is underreported, it likely occurs in all stages of lung infection in CF patients. A recent report by Anzaudo et al. (
2005) demonstrated that mixed species infections were reported 31% of the time (
Anzaudo et al., 2005). This is supported by past studies by Hoiby (
1974) and Burns et al. (1998) where mixed species communities were present most of the time (
Hoiby, 1974;
Burns et al., 1998). Burns et al. (
1998) reported the presence of 2.9 species of microbes from each patient in an elaborate study that included over 500 CF patients from 69 centers across the United States. Thus, it is likely that interactions among the infecting microbes play an important role in the pathogenesis of lung infections in the CF patient.
Chronic otitis media
Chronic otorhinolaryngologic infections are often polymicrobial in nature. Several species of bacteria living as attached communities have been implicated in chronic otitis media (OM). Otitis media is a common childhood infection and is characterized by inflammation of the middle ear (
Bakaletz, 2010). In a study by Hall-Stoodley et al. (
2006), bacterial biofilms were visualized in the middle ear of 92% of children with otitis media with effusion and recurrent culture negative OM. In addition to visualization of biofilms, all of the OM effusions were PCR-positive for at least one organism, while only 22% were culture positive. This demonstrates the importance of molecular techniques in understanding biofilm-mediated infection, since attached communities typically do not show up as culture positive under standard clinical culturing techniques.
H. influenzae,
Streptococcus pneumoniae, and
Moraxella catarrhalis are the bacteria predominantly responsible for chronic OM, and have been documented in or on human tissue by confocal scanning laser microscopy (CSLM), FISH, and PCR-based identification methods (
Hall-Stoodley et al., 2006;
Krishnamurthy et al., 2009;
Armbruster et al., 2010;
Weimer et al., 2010). Specific microbial interactions between OM bacteria can mediate the incidence and duration of infection. For example, a study by Krishnamurthy et al. (
2009) demonstrated that co-infection with
S. pneumoniae and
M. catarrhalis led to increased OM incidence, increased bacterial load, and longer duration in a mouse model than co-infection with
S. pneumoniae and
H. influenzae (
P<0.001) (
Krishnamurthy et al., 2009). This study highlights the importance of co-bacterial interactions in chronic disease.
Chronic rhinosinusitis
In addition to chronic OM, mixed species communities of bacteria and fungi are implicated in the common inflammatory disease, chronic rhinosinusitis (CRS). CRS affects ~20% of the population in the United States with associated medical costs of ~$6 billion annually (
Lanza and Kennedy, 1997;
Cherry and Woodwell, 2002;
Ly and McCaig, 2002). A retrospective study by Chan and Hadley (
2001) demonstrated polymicrobial communities in a culture-based review of charts of 83 patients. Approximately 12% of patients had mixed species communities. The predominant pathogens were coagulase-negative staphylococci (31% of isolates),
H. influenzae (25%),
S. pneumoniae (12%),
M. catarrhalis (10%),
P. aeruginosa (7%), alpha-hemolytic streptococci (5%) and
S. aureus (3%) (
Chan and Hadley, 2001). In 2004, the first published report of biofilms in CRS demonstrated the presence of attached communities in 6 of 6 frontal sinus stents through the use of scanning electron microscopy (SEM) (
Perloff and Palmer, 2004). Following this initial observation, several studies demonstrated the polymicrobial nature of biofilms implicated in CRS. A recent culture-based study by Prince and colleagues (2008) showed polymicrobial communities in 71% of 157 CRS patients. This study examined biofilm forming capacity of
S. aureus and
P. aeruginosa isolated from mucopurulence, and found a positive correlation between biofilm forming bacteria and prior endoscopic sinus surgery (
Prince et al., 2008), suggesting that the ability to form biofilms was one factor in the persistence of the disease.
While culture-based techniques provide insight into the mixed species communities of CRS, sensitive molecular studies are being employed to discern the microbial composition of attached microbial communities. In collaboration with the United States Naval Medical Center in San Diego, our laboratory demonstrated the presence of biofilms in 9/11 CRS samples and 2/3 controls (
Sanderson et al., 2006;
Healy et al., 2008). The predominant biofilm bacteria in these samples identified by species-specific probes were
H. influenzae, although
P. aeruginosa,
S. epidermidis,
S. aureus, and
S. pneumoniae were also detected. Fungi were successfully identified in communities with bacteria by phyla-specific genetic probes (
Sanderson et al., 2006;
Healy et al., 2008). In a study of 50 CRS patients and 10 non-CRS controls, Foreman et al. (
2009) used FISH and identified biofilms in 72% CRS and 0% non-CRS. In these studies, the
S. aureus dominated the biofilms. The difference in the predominant observed species between these studies is likely a result of sampling technique, fixing, or staining of the tissue samples. Importantly though, both of these studies demonstrated co-infection in CRS. To further investigate the microbial flora in CRS, our laboratory employed 454 pyrosequencing in an attempt to speciate the microbes in 200 CRS and 80 non-CRS controls. From these efforts, it is clear that the sinuses, in the healthy and diseased state, contain multiple species of microorganisms (bacteria and fungi). Although some taxa are increased in CRS (Clostridia, Actinobacteria), others are increased in non-CRS (Flavobacteria, Bacteroidetes) (Fig. 1). Patient exposure to antibiotics, steroids, cigarette smoke, and individual immune status played a role in modulating the composition of the sinus community. Recently, Stephenson and colleagues (2010) used 454 pyrosequencing to speciate the microbial communities in 18 CRS and 9 non-CRS patients. They demonstrated an average of 10 organisms per patient, where standard culture techniques detected only 2. Anaerobes dominated the CRS microbial communities that were reported in this study (
Stephenson et al., 2010). Microbial community composition can differ among patients and can influence disease severity, as demonstrated in an important retrospective study by Foreman and Wormald. In these studies, polymicrobial biofilms adversely affected preoperative disease severity. Patients with single species
H. influenzae biofilms had less severe disease and regained normalcy soon after FESS and patients with predominant
S. aureus biofilms suffered from more severe disease, and had complicated post-operative recovery (
Foreman and Wormald, 2010). These studies highlight the importance of defining the species in diseased communities for diagnostic and effective treatments. However, very few studies to date have examined the functional interactions that occur between bacteria that could lead to increases in disease pathology.
To define diseased communities, an understanding of healthy sinonasal communities is also important. Healthy microbial communities were elucidated in the upper respiratory tract of 62 volunteers by Charlson and colleagues (2010) using 454 pyrosequencing.This study demonstrated the stability of microbial communities in non-smoking healthy patients, including bacteria of the phyla bacteriodetes and actinobacteria. The microbial communities were disrupted and microbial diversity was increased when patients were exposed to cigarette smoke, indicating an unhealthy state (
Charlson et al., 2010). An endoscopically guided culture-based study of the healthy sinus demonstrated the prevalence of
S. aureus and coagulase-negative staphylococci, diptheroids, and
Propionibacteria acnes (
Nadel et al., 1999). All of these bacteria were found in CRS diseased tissue using 454 sequencing in our studies, above. This suggests that the relative abundance of potential pathogens in a microbial community may play a role in disease development, and is in agreement with the ideas discussed by Ehrilch et al. (2008), and Jenkinson and Lamont (
2005). Overall, it is clear that CRS is caused or exacerbated by attached microbial communities, and these communities are often comprised of multiple species. Further investigation of community composition is necessary to develop more reliable therapeutic approaches for CRS.
Mechanisms of microbial interactions: physical and chemical mediators of virulence
Interspecies interactions are important mediators of microbial attachment and virulence. These interactions span the spectrum from antagonistic to mutualistic, depending on the species and environment (
Palmer et al., 2001;
Hoffman et al., 2006). The past few years has seen an increase in research devoted to understanding cross-species and cross-kingdom interactions (
Hogan and Kolter, 2002;
Waters and Bassler, 2005;
Ryan and Dow, 2008;
Peters et al., 2010). It is imperative that we continue to examine interspecies interactions in disease, so new vaccines and therapeutics can be developed to combat more complex diseases.
Physical interactions
An important example of physical interactions between species is co-aggregation. Extensively described in the oral setting, the ability to co-aggregate and co-adhere forms the foundation of early oral biofilms, which can dictate health or disease (
Jenkinson and Lamont, 2005). In short, co-aggregation is a process by which bacteria physically adhere to each other via cell surface ligands (
Kolenbrander et al., 2006). For example, co-aggregation of the pathogen
P. gingivalis with
T. denticola is dependent on surface proteins that contain a 44-kDa Hgp adhesion domain. These domains are found in the arginine-specific cysteine proteinases (rgpA, pgpB), hemagglutinin A (HagA), and the lysine-specific cysteine proteinase (kgp). All three groups represent common surface receptors on
P. gingivalis (
Ito et al., 2010). Co-aggregation can lead to increased virulence and host inflammation when pathogens are allowed to integrate in an oral community. Co-infection of the mid to late colonizer
F. nucleatum and the periodontal pathogen
P. gingivalis led to increased formation of oral lesions in a mouse model than either bacteria alone (
Feuille et al., 1996;
Ebersole et al., 1997). Primary infection with both
P. gingivalis and
F. nucleatum resulted in greater lesion formation. Interestingly, simultaneous infection (given at different sites) and subsequent infection (
F. nucleatum followed by
P. gingivalis) greatly increased the ability of
P. gingivalis to form invasive lesions. When given a virulence-attenuated strain of
P. gingivalis, no lesions could be detected in these mice (
Feuille et al., 1996). These studies suggest that interspecies interactions, including co-aggregation, influence pathogenicity in the host.
The host response is an important factor in virulence of co-aggregated communities, as it can result in tissue damage from chronic inflammation. Guggenheim and colleagues (2009) demonstrated an elevated inflammatory response of human gingival epithelial cells (HGEG) when a 9-species biofilm representing a sub-gingival pathogenic community was introduced during
in vitro culture. It is important to understand co-aggregation in early colonizers of the oral cavity to fully understand healthy vs. diseased states. Palmer et al. (
2001) demonstrated a mutualistic relationship between the early oral colonizers
Actinomyces naeslundii and
Streptococcus oralis. Biofilms were grown
in vitro in unaltered saliva and they demonstrated that biofilm growth only occurred in coculture. After 18 h, both strains flourished together, and could be visualized in contact with each other. The synergy demonstrated by these studies is of major significance to the oral bacterial ecology. Since both
A. naeslundii and
S. oralis are early colonizers of the oral cavity, understanding the interactions required to build a healthy community may lead to therapies that target early oral colonizers to promote health (
Palmer et al., 2001).
Cross-kingdom interactions have a large impact on human health.
Candida albicans, a common fungal member of the skin and mucosal microflora, interacts with bacteria to increase virulence and invasion (
Hogan and Kolter, 2002;
Shirtliff et al., 2009;
Peters et al., 2010).
C. albicans biofilms are found in burn wounds, urinary catheters, and medical implants in intimate association with bacteria (
Hemady, 1995;
Romano Hermann et al., 1999;
Romano and Kolter, 2005;
Falleiros de Padua et al., 2008;
Holcombe et al., 2010). These interactions span from mutualism to parasitism, and often require physical signals.
S. aureus and
P. aeruginosa selectively adhere to
C. albicans filaments, allowing nutrient acquisition and invasion into tissue (
Romano and Kolter, 2005;
Peters et al., 2010). The relationship between
C. albicans and
P. aeruginosa is parasitic.
P. aeruginosa inhibits
C. albicans biofilm formation and kills the filaments on which it forms biofilms (
Hogan and Kolter, 2002;
Holcombe et al., 2010). Hogan and Kolter (
2002) found that both pili and secreted molecules acted to kill
C. albicans hyphae. In a comprehensive microarray study, Holcombe and colleagues (2010) were able to discern potential genes of interest in
P. aeruginosa, and concluded that quorum sensing was involved in the inhibition of
C. albicans. This antagonism has clinical significance and can be exploited to inhibit
C. albicans biofilm infections of catheters and medical implants. An elegant study by Peters et al. (2010) characterized the physical interaction and the proteomic consequences of a dual species
S. aureus-
C. albicans biofilm (
Peters et al., 2010). They demonstrated that
S. aureus selectively bound to the hyphal form of
C. albicans and that physical contact increased
S. aureus production of virulence and growth factors including L-lactate dehydrogenase 1, which confers resistance to host oxidative stress. Additionally, this interaction repressed a global virulence factor repressor in
S. aureus, CodY. Importantly, these interactions only occurred during hyphal growth of
C. albicans, which has considerable clinical significance. It is likely that
S. aureus uses
C. albicans to invade host tissue and cause disease (
Peters et al., 2010). These studies have provided new information about cross-kingdom interactions and will likely have an important impact on the treatment of chronic infections involving complicated bacterial-fungal communities.
Quorum sensing
Quorum sensing has been extensively studied as a bacterial cell-cell signaling system. Briefly, there are four major classes of QS systems. Autoinducer-1 and autoinducer-3 exist in Gram-negative bacteria. Gram-positive bacteria use a system of autoinducing oligopeptides (AIP) regulated by the accessory gene regulator (agr) system. The fourth is autoinducing peptide-2 (AI-2), which we focus on here. AI-2 is a well-characterized interspecies signaling molecule (
Rickard et al., 2006). The
luxS gene controls AI-2 synthesis, and homologs have been demonstrated in a variety of Gram-negative and Gram-positive organisms (
Surette et al., 1999;
Bassler, 1999). Almost half of sequenced bacteria have the
luxS gene, and many more have homologs of
luxS genes, including
lsr (
luxS regulated) that responds to AI-2 (
Taga et al., 2001;
Xavier and Bassler, 2005). The LuxS enzyme catalyzes the formation of 4,5-dihydroxy-2,3-pentanedione (DPD) (
Schauder et al., 2001). DPD spontaneously rearranges to form the active product termed AI-2, which can vary among species. Two genes characterized in
V. harveyi,
luxP and
luxQ, encode components of the AI-2 receptor and homologs have also been observed in sequenced genomes of several bacteria (
Bassler et al., 1994).
A large effort has begun to define the role of AI-2 in gene expression and virulence. Several studies have examined the role of this signaling molecule in virulence factor production in numerous human pathogens including
S. aureus,
S. epidermidis,
H. influenzae,
Escherichia coli,
Vibrio cholerae, and
Bacillus anthracis (
Soni et al., 2007;
Higgins et al., 2007;
Li et al., 2008;
Armbruster et al., 2010;
Jones et al., 2010;
Zhao et al., 2010). For example, in
V. cholerae, AI-2 plays a role in virulence factor production and biofilm formation (
Higgins et al., 2007). Some
S. epidermidis genes, including those that encode for lipase and the pro-inflammatory phenol soluble modulins, are under AI-2 control (
Li et al., 2008). AI-2 signaling in
S. aureus has not been extensively studied, but a role in the inverse regulation of the polysaccharide capsule was recently described ((
Li et al., 2008;
Armbruster et al., 2010;
Zhao et al., 2010). Expression of the virulence genes
yadK and
hhA in
E. coli O157∶H7 are regulated by AI-2 signaling. Introduction of an AI-2 inhibitor decreased their respective gene expression, which validated the role of AI-2 in virulence of this food-borne pathogen (
Soni et al., 2008).
Interestingly, in a mixed species setting, certain bacteria can manipulate AI-2 signaling and cause neighboring species to underestimate species density and misinterpret the environmental signals. This can confer a competitive advantage in multi-species communities. A study by Xavier and Bassler (
2005) showed that
E. coli consumes AI-2 in coculture conditions with
Vibrio. By using
V. harveyi bioluminescence as a measure of QS signaling, they demonstrated the inverse effect of AI-2 interference by
E. coli on
V. harveyi. This effect, while more modest, was also seen in
E. coli-
V. cholerae cocultures (
Xavier and Bassler, 2005). These studies show an interesting antagonism between two important intestinal pathogens. On the other end of the spectrum, Rickard and colleagues (2006) demonstrated that AI-2 was required for mutualistic biofilm growth of the oral bacteria
S. oralis and
A. naeslundii in a saliva fed
in vitro system. This study demonstrated a concentration-dependent signaling system for coculture biofilm growth of
S. oralis and
A. naeslundii. In the absence of AI-2, mixed species biofilms were sparse and contained 10-fold lower biomass of each species. AI-2 was also shown to promote polymicrobial biofilm formation of
H. influenzae and
M. catarrhalis in otitis media (
Armbruster et al., 2010). Importantly, AI-2 promoted increased resistance of biofilms to antibiotics and host immune molecules via indirect pathogenicity. In this study, they identified AI-2 as the key molecule produced by
H. influenzae that promoted
M. catarrhalis biofilm growth and persistence
in vivo (
Armbruster et al., 2010). Collectively, these studies highlight the importance of AI-2 in infection and may represent an ideal target for disruption of polymicrobial communities that are the cause of disease. Further studies of the role of AI-2 in multi species communities are necessary to elucidate the role of this important signaling system.
Chemical messengers
Although QS systems have been well characterized in bacteria and have great importance in virulence factor production and infection, other non-QS systems likely play a role in disease pathology. Bacterial secondary metabolite perception can influence gene expression in mixed-species biofilms and aid in host immuno-avoidance (
Ramsey and Whiteley, 2009). A recent study by Ramsey and Whitely (
2009) demonstrated the importance of metabolite perception in oral biofilms of the pathogen
Aggregatibacter actinomycetemcomitans and the commensal
Streptococcus gordonii.
S. gordonii production of hydrogen peroxide induced the upregulation of
A. actinomycetemcomitans genes
apiA and
katA. Importantly,
apiA, regulated by
oxyR, was critical for resistance to host innate immunity. This study demonstrated a change in gene expression resulting from secondary metabolite signaling that produced a functional response in the bacteria. The cooperation of these organisms created a mechanism for increased persistence. Further
in vivo studies are needed to define these consequences during infection. A well-characterized interaction between
P. aeruginosa and
S. aureus is the result of a non-QS signal, 4-hydroxy-2-heptylquinolone-
N-oxide (HQNO). Hoffman and colleagues (2006) demonstrated both
P. aeruginosa repression of
S. aureus respiration, and a resulting increase of antibiotic resistance by promotion of small colony variants (SCVs) (Hoffman et al., 2006).
S. aureus SCVs are a slow growing phenotype that result from defects in electon-transport that promote persistence in infection (
Proctor et al., 2006). By decreasing respiration, common antibiotics, including aminoglycosides, which target bacterial growth and metabolism, have fewer targets. Further, SCVs are associated with persistent infections including cystic fibrosis (
Kahl et al., 1998), osteomyelitis (
Lattar et al., 2009), and implant device infections (Rolauffs et al., 2002; Baddour and Christensen, 1987). The interspecies interactions highlighted here demonstrate the importance of chemical interactions that can result in increased virulence and persistence.
Future prospects for clinical management of polymicrobial disease
As culture-independent techniques elucidate the complexity of polymicrobial diseases, new treatment strategies must be developed to increase the effectiveness of chronic disease therapeutics. These studies discussed above illustrate the importance of multi species communities in chronic infection. The next step in clinical management of chronic, polymicrobial infection, will be to target the species that drive pathogenesis within a community, while leaving the healthy components intact. In addition, it would be useful to block initial biofilm formation by the pathogenic members of infecting communities. For example, inhibition of co-aggregation between T. denticola and the periodontal pathogen P. gingivalis could be an effective prophylaxis for periodontitis.
Quorum sensing antagonists
Quorum sensing antagonists have been widely studied as potential inhibitors of biofilm formation in disease, but the clinical applications remain difficult because the timing of therapeutic delivery often dictates the success of the therapy. For example, since QS molecules are often produced early on during biofilm formation, because the genes that they regulate are involved in early biofilm formation and maturation, it may be too late to administer these compounds as early as 5 days post exposure. Nevertheless, several studies have demonstrated that a wide array of compounds may be useful for blocking QS, and the resultant downstream virulence, of biofilms
in vitro and
in vivo. Macrolide antibiotics, including azithromycin, are effective at blocking QS and attenuating virulence in
P. aeruginosa in an experimental UTI model of infection (
Bala et al., 2011). In addition, several studies have demonstrated that synthetic and natural furanones in combination with antibiotics can be effective in treating bacterial infections in experimental systems (
Hentzer et al., 2003;
Hentzer et al., 2003;
Bjarnsholt and Givskov, 2007). A 2009 study by Swem et al. elucidated a molecule that acted as a broad-spectrum acyl-homoserine lactone (AHL, Gram negative autoinducer-1) antagonist, and demonstrated reduced virulence of
Chromobacterium violaceum in a
Caenorhabditis elegans model. These studies show potential uses for QS blockers to reduce bacterial virulence and to reduce bacterial load when used in combination with known antibiotics. The challenge of interrupting the right signals at the right time in infection is underscored in a study by Geske (2007). Their studies demonstrated that slight alterations of the molecular structure or concentration of a known QS antagonist, non-native
N-acetylated-L-homoserine lactone, changed the molecule from a strong antagonist to an agonist (
Geske et al., 2007;
Geske et al., 2007). It is likely that design and use of QS antagonists as effective therapeutics will not be as straightforward as the scientific community has hoped and more research in this area, especially in animal models of biofilm disease, is necessary.
Two component response regulator antagonists
Targeting of two component response systems in bacteria may prove to be a means to control virulence in a mixed species population. More than 4000 TCRS have been identified in approximately 400 sequenced genomes (
Galperin, 2006;
Ulrich and Zhulin, 2007). These often control aggressive virulence factors; so antagonizing these systems may prove to be a useful approach to control mixed species communities without eradicating the healthy components. In addition to controlling frank virulence of bacteria, TCRS can also control tolerance of bacteria to antibiotics. For example the VanR-VanS TCRS detect vancomycin and activate the expression of the enzymes VanA, VanH, and VanX, which are required for antibiotic resistance (
Arthur et al., 1992;
Evers and Courvalin, 1996;
Hong et al., 2008). Biofilm formation and resistance to oxidative stress of the dental caries pathogen
Streptococcus mutans, is controlled by the VicRK TCRS (
Deng et al., 2007;
Duque et al., 2011). While the environmental signals that elicit a response are still unclear, recent studies have suggested oxidative stressors and environmental sources of glucan as potential signals for the VicRK TCRS in
S. mutans. A common TCRS in low-GC content Gram-positive bacteria has been explored as a potential antimicrobial target. The WalKR system is highly conserved and controls cell wall metabolism and cell division (
Gotoh et al., 2010). Walrycin A and B were effective as bacteriocidal antimicrobials against
Bacillus subtilis and
S. aureus (
Gotoh et al., 2010). Targeting these response systems can prove useful in attenuating virulence of microbial communities in chronic disease. Since treatment will be specific to a component of a pathogen that is driving disease, therapeutics targeted toward TCRS will be less likely to disrupt healthy communities in non-diseased parts of the patients and will likely have low host-toxicity, leading to less side effects.
Conclusions
The paradigm of one organism, with a corresponding virulence, causing one disease, has been an important concept that has led to a vast understanding of microbes and the diseases they cause. While many acute infections caused by a single organism are easily treated, the past decade of molecular research has been crucial in identifying the complexity associated with chronic diseases. These chronic diseases are often recalcitrant to antibiotic treatment, leaving debridement or removal of affected tissue as the sole approach for clinical therapy. As a result, chronic infections are often more expensive to treat, and contribute to a large socioeconomic burden. Microbiome sequencing of healthy human tissue has started to redefine a new normal in human medicine; no parts of the human body are ‘sterile’ from microbial burden. Since many of these normal flora bacteria are often found in diseased tissues, developing a better understanding of how microbes interact with each other, to produce a functional consequence (health or disease), should be an important goal over the next decade. As community-specific factors are identified, demonstrating the effects of population-level virulence, a new group of therapeutics may be designed that target the community, and not a single pathogen. These new therapies may dramatically advance the treatment of chronic, population-level diseases in humans.
Higher Education Press and Springer-Verlag Berlin Heidelberg
PDF
(250KB)
