The social life of microbes in chronic infection

Abstract

Chronic infections place a significant burden on healthcare systems, requiring over $25 billion in treatment annually in the United States alone [1, 2]. Notably, the majority of chronic infections, which include cystic fibrosis (CF), chronic wounds, otitis media, periodontitis, urinary tract infections, and osteomyelitis, are considered polymicrobial and are often recalcitrant to antibiotic treatment [1–9]. Although we know that diverse communities of microbes comprise these infections, how microbes interact and the impacts of these interactions on human disease are less understood. Here, we discuss recent advances in our understanding of how bacteria communicate in chronic infection, with a focus on Staphylococcus aureus and Pseudomonas aeruginosa, and we highlight outstanding questions and controversies in the field.

“Hello, is anybody out there?”

In nature, bacteria rarely exist in isolation, but instead live in complex polymicrobial communities. The earliest microscopes revealed the existence of multispecies communities in human infection, and work from microbiologists as early as the 1870s reported phenomena resulting from interactions of bacteria existing in multispecies communities in infection [10]. Ever since, understanding the species composition and how microbes interact in chronic bacterial infection has been a major focus of research. Collectively, a large body of work has revealed that microbes communicate through diverse mechanisms that include the exchange of diffusible molecules and direct cell-cell contact, and these interactions allow microbes to sense and respond to neighboring cells to coordinate group behavior. Importantly, polymicrobial interactions have clinically relevant outcomes in animal infection models, including increased tolerance to antibiotics and increased virulence [4, 6, 11–15]. Additionally, decades of culture-based research, augmented recently with the introduction of culture-independent methods, have uncovered the general community composition in many chronic infections. While understanding community composition is important, many questions remain including: Are the microbes in chronic infection interacting? If so, at what distances are interactions between microbes in chronic infection occurring? And, do interactions between microbes, both inter- and intra-species, impact disease outcomes in chronic infection? In this review, we will discuss our current understanding of intra- and inter- species interactions in chronic infection, focusing on the model organisms Staphylococcus aureus and Pseudomonas aeruginosa in cystic fibrosis (CF) and chronic wounds.

Two key players in chronic infection: P. aeruginosa and S. aureus

Two types of chronic infection that have been at the center of polymicrobial infection research are: CF lung infection and chronic wounds. CF is a recessive genetic disease caused by mutations in the gene encoding the cystic fibrosis transmembrane conductance regulator, an ion channel that conducts chloride and bicarbonate across epithelial cell membranes. A consequence of this mutation is the accumulation of viscous mucus in the lungs and on other mucosal surfaces. Mucus (sputum) provides nutrients and a scaffold in which microbes can reside and cause infection. A hallmark of CF is chronic polymicrobial lung infections that begin in early childhood and persist throughout the lives of CF patients [1, 16]. Of note, bacterial lung infection is the primary cause of morbidity and mortality in CF patients [16]. Chronic wounds, which are defined as wounds that fail to heal in an orderly or timely manner [17], place a significant burden on healthcare systems. Chronic wound infections frequently fail to respond to treatment and in some cases ultimately necessitate amputation [2, 6, 17]. In addition, it has been estimated that up to 2% of the population in developed countries will experience a chronic wound in their lifetime [2]. On average, six microbial species can be found within chronic wounds [6].

S. aureus and P. aeruginosa are two of the most frequently isolated microbes from both CF and chronic wound infections [1, 6]. S. aureus, a Gram-positive coccus harboring more than 50 virulence factors and immune evasion strategies, is both a leading cause of bacterial infection worldwide and a frequent human commensal [18–23]. P. aeruginosa is a Gram-negative opportunistic pathogen, with a large genome that allows for the production of numerous secreted molecules and virulence factors [24]. In addition to their role in pathogenesis, a large body of research has focused on characterizing intra- and inter-species interactions between these organisms and their co-infecting partners (Fig. 1).

Figure 1. Summary of known inter- and intra-species interactions in the model organisms S. aureus and P. aeruginosa in in vitro and in animal infection models.

S. aureus is depicted in in yellow and P. aeruginosa in green. Positive (cooperative) interactions are shown with an arrow (→) while negative (competitive) interactions are shown with an inhibitory arrow (⊣). Depicted interactions (clockwise from right): 1) Inhibition of S. aureus quorum sensing (QS) by non-cognate auto-inducing peptide (AIP) quorum sensing molecules produced by distantly related S. aureus strains or other staphylococcal species (blue) [32, 51]; 2) Peptidoglycan produced by S. aureus increases P. aeruginosa virulence and production of pyocyanin [60]; 3) Secreted Protein A produced by S. aureus coats P. aeruginosa cells leading to increased persistence and reduced phagocytosis by neutrophils [11]; 4) S. aureus has increased expression of the secreted virulence factors alpha-toxin and Panton-Valentine leucocidin and reduced expression of protein A in a murine model of infection through an unknown mechanism [61]; 5) Expression of the quorum sensing system (Agr-system) in S. aureus increases virulence through higher production of virulence factors including toxins (alpha-toxin, gamma-toxin), proteases (SplA-F, SspA, SspB, ScpA), and leukocidins (LukAB, LukGH) [19, 20, 28, 32, 43]; 6) Quinolones (e.g. 2-n-heptyl-4-hydroxyquinoline N-oxide (HQNO)) secreted by P. aeruginosa induce cell lysis and block the electron transport chain in S. aureus, resulting in a switch from a respiratory to a fermentative metabolism [15, 63]; 7) Expression of the quorum sensing systems in P. aeruginosa (LasI-LasR; RhlI-RhlR) increases virulence through higher production of secreted molecules including elastase, pyocyanin, and quinolones [25, 26, 30, 31, 33].

Intra-species interactions

Quorum sensing (QS), defined as the use of self-produced signals that accumulate in the environment in proportion to cell density to alter gene expression [25], has been at the center of research on intra-species communication. In Gram-negative bacteria, QS is primarily mediated by small molecules called acyl-homoserine lactones (AHLs), and these molecules show species specificity dictated by the length and chemical substitutions of the acyl group. In contrast, Gram-positive bacteria primarily use small peptides, known as autoinducing peptides (AIPs), for QS that vary widely in sequence and structure, which provides species specificity. Collectively, studies over the past 50 years have elucidated a number of QS-regulated bacterial responses in vitro and in animal infection models that control diverse behaviors such as biofilm formation, virulence factor production, and carbon utilization [20, 25–33]. Table 1 contains a number of examples of these responses. However, although QS genes coordinate a number of behaviors important in acute infection and pathogenesis [15, 20, 21, 25, 28, 34], the role of QS in chronic human infection is less clear. For example, there is significant evidence that over time, mutations accumulate in chronic infection isolates that inactivate QS systems in both S. aureus and P. aeruginosa [34–40]. In addition, transcriptomic studies indicate that the expression of QS genes is reduced in chronic human infection in both S. aureus and P. aeruginosa [41–44]. Thus, it is possible that although QS is important for establishing infection, a fully intact QS network is not required for fitness in long-term infections, perhaps due to the relative cost of producing public goods (e.g. secreted enzymes & toxins). However, recent work indicates that while mutations in a central QS system often occur in P. aeruginosa human chronic infection isolates, many QS controlled genes are still expressed in a QS-dependent manner, as their control has been usurped by other QS systems [45, 46]. Therefore, the long-term role and importance of QS in chronic infection remains to be conclusively defined.

Table 1.

Examples of intra-species interactions in S. aureus and P. aeruginosa

Organism	Type of Interaction	Bacterial Response	Citation(s)
S. aureus	Quorum sensing	Dispersal from biofilms: ↑ proteases, ↑ nuclease, ↑ PSMs, ↓ adhesins Virulence: ↑ α-toxin, ↑ γ-toxin, ↑ leukocidins	[9, 19, 20, 27, 28, 32, 50]
	Type VII secretion	↓ growth of strains lacking immunity proteins	[48]
P. aeruginosa	Quorum sensing	Access to new nutrients: ↑ proteases (LasA, LasB, AprA), ↑ siderophores (pyochelin, pyoverdine) Immune evasion: ↑ LasA, ↑ AprA (complement degradation), ↑ RhlAB (necrosis of phagocytic cells), ↑ pyocyanin (neutrophil damage) Enhanced colonization: ↑ LecA (paralysis of cilia), ↑ hydrogen cyanide (arrest of cellular respiration), ↑ pyocyanin (paralysis of cilia, reduced host cellular respiration)	[25, 26, 30, 31, 33]
	Type VI secretion	↓ growth of strains lacking immunity proteins	[47]

Abbreviations: PSM: phenol-soluble modulin

In addition to QS-based communication using diffusible signals, bacteria can also communicate through contact-dependent mechanisms. One example is through the use of the type VI secretion system (T6SS), which is a syringe-like protein nanomachine that is widespread in Gram-negative bacteria and is used to translocate effector proteins into neighboring cells. P. aeruginosa senses penetration of its outer membrane by the T6SS of another microbe and responds with a counter-attack [47]. Additionally, in S. aureus, the nuclease effector of the type VII (T7SS), a recently described secretion system found in Firmicutes and Actinobacteria, inhibits the growth of strains that lack the immunity proteins through the induction of DNA damage [48]. Interestingly, many Gram-negative bacteria encode multiple T6SS in their genome, for example P. aeruginosa has three systems. However, the biological significance of multiple T6SSs in microbe-microbe interactions and the role of competitive secretion systems like T6SS and T7SS during infection have not been fully elucidated.

Inter-species interactions

Chronic infections typically contain complex and diverse communities of both commensal and pathogenic microbes [3, 5–9, 49, 50]. Both cooperative and competitive interactions have been described between microbes in chronic infection (Table 2, and in these comprehensive reviews [3, 7, 8, 49, 50]), and characterizing the mechanisms and types of communication that occur is crucial for understanding the biology of these infections. Interactions between microbes are often mediated by the production of secreted molecules, including enzymes and small molecules, although other mechanisms including contact-dependent mechanisms, such as T6SS and T7SS, to compete for resources are also used. These interactions between microbes are driven by both the composition of the community as well as the host environment. Here, we give examples of bacterial cross-talk, as well as cooperative and competitive interactions that have been described in chronic infections, and any impacts that these interactions have on disease outcomes.

Table 2.

Examples of inter-species interactions with S. aureus and P. aeruginosa

Responding Organism	Interacting partner	Type of Interaction	Mechanism	Response/Outcome	Citation
S. aureus	Other staphylococci & Non-related S. aureus strains	Cross-talk/Competitive	Competitive binding by AIP to AgrC	Inhibition of quorum sensing, reduced virulence	[20, 51]
	C. albicans	Cooperative	Binding to hyphae	↑ biofilm formation, ↑ antibiotic tolerance	[14, 58]
	P. aeruginosa	Cooperative	Unknown	↑ virulence: ↑ α-toxin, ↑ PVL	[61]
	P. aeruginosa	Competitive	Lysing by quinolones	Cell lysis	[2, 74]
	P. aeruginosa	Competitive/Cooperative	Inhibition of respiration	↑ fermentation, ↑ antibiotic tolerance	[15, 63]
P. aeruginosa	S. aureus/Gram-positives	Cooperative	Sensing peptidoglycan	↑ virulence	[60]
	S. aureus	Competitive	Sensing of PSMs	Alter motility: ↑ single cell motility, invade S. aureus colonies	[75]
	C. albicans	Cooperative	Secreted ethanol from C. albicans	↑ biofilm formation	[57]

Abbreviations in Table 2: PVL: Panton-Valentine Leukocidin; PSM: phenol-soluble modulin

Bacterial cross-talk

Although QS is primarily considered an intra-species form of communication, there are increasing examples of bacterial “cross-talk” where sensing or inhibition of QS signals from other species can alter gene expression [51, 52]. In a recent example, the AIP molecule of the coagulase-negative human commensal Staphylococcus caprae was shown to block S. aureus QS, resulting in reduced skin colonization and infection in a murine model [51]. Additionally, a recent paper by Wellington and Greenberg showed that although Gram-negative AHL QS molecules have been thought to be highly species specific, QS receptors displayed a range of sensitivity to different AHL molecules, and most receptors responded strongly to at least one non-self signal [52]. The ability to modify gene expression of competitor species in infection could provide a fitness advantage as well as prevent invasion of new organisms into sites of colonization.

Cooperative interactions

Many described inter-species interactions in chronic infection result in clinically important outcomes, including increased tolerance to antibiotics and host immune defenses [3, 4, 12, 13, 15, 53–56], increased biofilm formation [3, 5, 9], and increased bacterial fitness, which is often termed synergy [7, 8, 13, 49]. For example, ethanol produced by Candida albicans [57] increases biofilm formation in P. aeruginosa, and S. aureus uses the hyphae of C. albicans as a substratum to form more robust biofilms [14, 58]. Also, P. aeruginosa has a higher tolerance to gentamicin in polymicrobial infection when compared to mono-infection in a murine chronic wound infection model [59]. Microbe-microbe interactions have also been shown to increase virulence factor production in S. aureus and in P. aeruginosa. For example, P. aeruginosa can sense peptidoglycan and increase virulence factor production in vitro and in a fly infection model [60]. In addition, staphylococcal protein A produced by S. aureus inhibits neutrophil phagocytosis of P. aeruginosa [11]. Similarly, the presence of P. aeruginosa induced the expression of S. aureus alpha-toxin and Panton-Valentine leukocidin in a porcine wound model when compared to infection with S. aureus alone [61]. Taken together, cooperative interactions, like the examples given, often lead to increased disease severity, increased tolerance to antibiotics, and make polymicrobial chronic infections difficult to treat and resolve.

Competitive interactions

In addition to cooperative interactions, many of the interactions between co-infecting microbes in chronic infection are competitive, and are often related to nutrient limitation. During infection, a critical innate immune defense is the sequestration of key nutrients, particularly iron, in a process termed nutritional immunity. This causes infecting bacteria to compete for iron, and many resulting competitive interactions have been described in infection. For example, as described in detail in a review by Bisht, Baishya, and Wakeman in this issue, during iron depletion P. aeruginosa increases production of quinolones that lyse S. aureus [62], releasing intracellular iron that P. aeruginosa can use. In addition to direct competition for nutrients, interactions that alter the metabolism of competing organisms have also been described. For example, the P. aeruginosa-produced quinolone 2-n-heptyl-4-hydroxyquinoline N-oxide (HQNO) inhibits respiration and growth in S. aureus, leading to the adoption of a fermentative metabolism [63], and often the formation of small colony variants (SCVs) [54, 64]. Of note, SCVs have increased tolerance to antibiotics and molecules like pyocyanin [12, 13, 54], indicating that competitive interactions can also be beneficial for microbes and result in increased persistence in infection.

Importance of spatial organization

An emerging area of research in microbiology is to understand the spatial organization of microbes within infection, and how it dictates the types of interactions that occur. In ecology, it is well described that the spatial structure of a system promotes the coexistence of competing organisms in a given environment [65]. However, what defines the spatial parameters of a microbial environment? At a macro-scale, chronic infections provide vast space for a bacterial community to emerge. Although high densities (~10⁹ colony forming units) of bacteria can exist within a chronic infection, a bacterial population of this size would fill less than 0.1% of the volume of a 1 ml sample based on the known volume of Escherichia coli (10⁻¹⁵ liters). Further, there is significant evidence that in infection, bacteria often exist as small dense aggregates of ~10¹–10⁴ cells [55, 56, 66, 67], and multiple studies have indicated that aggregates need to be within microns of each other to interact [68–70]. For example, a recent study [69] found that aggregates of P. aeruginosa must be within ~180 microns of each other to respond to QS molecules. The importance of spatial organization on bacterial physiology and interactions between microbes is increasingly appreciated, with the development a number of advances in imaging technologies, including innovations in fluorescence in situ hybridization labelling [71], and techniques that leverage multiple imaging platforms [72]. These techniques have revealed highly structured environments in a number of chronic human infections [71–73]. Therefore, when taken together, it is necessary to consider the scale and spatial organization of the organisms when assessing microbial interactions in infection, as interactions are likely to only occur locally.

What are the next steps?

Although both inter- and intra-species interactions between co-infecting microbes have been described in numerous animal and in vitro models, some of the most important remaining questions are: 1) do these interactions occur in human infection? and, 2) what, if any, are the impacts on human disease outcomes? In vitro and animal models differ from human infection in many ways, including the size of the inoculum, nutritional status of the infecting bacteria, spatial organization, and the ratio and density of co-infecting microbes. However, the impact of these different factors on bacterial physiology, and the interactions between microbes in infection are not always clear. In addition, although two microbes may exist within a human infection at the macroscale, they may not be spatially organized in a way that allows for interactions. In support of this idea, although it is generally accepted that S. aureus is pushed to fermentative metabolism when grown with P. aeruginosa [15, 63], in our recent work, S. aureus transcriptomes in human CF sputum samples indicated that S. aureus is respiring despite the clinical lab indicating 8/10 of the sputum samples also contained P. aeruginosa [44]. This corroborates the concept that although these bacteria were both contained in the sample, they were likely not localized in a manner that was permissive for them to interact.

We propose that one must always consider that the in vitro and in vivo models used to study bacterial interactions may be forcing interactions that are not present in human infections. We view infection models as tools to discover the types of interactions that could be taking place in a human infection and for exploring molecular bases of these interactions. Ultimately, understanding the types of interactions that occur in human chronic infection will be technically challenging and will likely require significant numbers of longitudinal, observational, and manipulative studies of human-derived samples. These future studies will be critical to truly understanding the role of interactions between microbes during human chronic infection.

Highlights.

• Quorum sensing is frequently lost or downregulated in chronic infection

• Interactions often lead to increased antibiotic tolerance and disease severity

• Spatial organization of microbes in an infection influences their interactions

• Polymicrobial interactions have not been demonstrated in most human infections