Interplay between host-microbe and microbe-microbe interactions in cystic fibrosis

Abstract

The respiratory tract of individuals with cystic fibrosis is host to polymicrobial infections that persist for decades and lead to significant morbidity and mortality. Improving our understanding of CF respiratory infections requires coordinated efforts from researchers in the fields of microbial physiology, genomics, and ecology, as well as epithelial biology and immunology. Here, we have highlighted examples from recent CF microbial pathogenesis literature of how the host nutritional environment, immune response, and microbe-microbe interactions can feedback onto each other, leading to diverse effects on lung disease pathogenesis in CF.

Graphical Abstract

1.1. Background

Individuals with cystic fibrosis (CF) harbor lifelong respiratory infections that are typically polymicrobial. The most frequently cultured bacteria from CF airways are Staphylococcus aureus, Pseudomonas aeruginosa, Haemophilus spp., Stenotrophomonas maltophilia, Achromobacter spp., Burkholderia cepacia complex, and non-Tuberculous Mycobacteria (1). Recently, a growing body of CF respiratory microbiome literature has identified additional bacterial taxa present in CF respiratory secretions, including many genera containing anaerobic species like Prevotella spp. and Veillonella spp. (reviewed in (2)). Furthermore, fungal (reviewed in (3)) and respiratory viral pathogens (reviewed in (4)) frequently co-infect along with the bacterial constituents in the respiratory tract, with important implications for disease pathogenesis.

In this review, based on a symposium at the 16th ECFS Basic Science Conference, Dubrovnik, Croatia in March 2019, we summarize emerging work on the topics of host-microbe and polymicrobial interactions in CF respiratory disease (Figure 1). First, we focus on 3 nutrients (iron, mucins, and glucose) in order to highlight mechanistic examples of how the host nutritional environment drives changes in microbial pathogenesis that feedback into host-microbe and microbe-microbe interactions. Second, we discuss examples of how the host immune response may drive changes in the respiratory microbiota composition, including the temporal shift from S. aureus to P. aeruginosa dominance commonly observed during the late teenage to early adulthood years. Third, we highlight emerging research addressing how host-microbe and microbe-microbe interactions influence antimicrobial susceptibility of bacterial biofilm communities. Finally, we conclude with a measured reflection on contemporary challenges and considerations for researchers in the field of CF respiratory host-pathogen interactions.

Figure 1. Cross section of a large airway in a CF lung, depicting examples of how microbial interactions with each other, the nutritional and physiochemical environment, and the host can feedback into diverse changes in disease pathogenesis.

Red epithelial cells represent host-microbe interactions that lead to an increased host response (e.g. increased inflammatory signaling) and yellow represents a decreased host response (e.g. reduced IL-8 secretion, inactivation of antimicrobial peptides). Alternating red and yellow epithelial cells indicate that both increased and decreased cytotoxicity or host immune-stimulatory phenotypes have been described. Peach epithelial cells are used when an interaction with the host was not described. The light-yellow shading lining the apical surface of the epithelium and surrounding microbial aggregates represents airway surface liquid (ASL) and thickened airway mucus, respectively. Black inhibitor arrows (—|) indicate that the phenotype blocks phagocytosis. A) P. aeruginosa (green rods) senses iron availability as a signal to increase biofilm matrix production (alginate; light green), leading to downregulation of the flagellum and reduced inflammation. Alginate production is associated with reduced killing of S. aureus (yellow cocci). Additionally, respiratory inhibitors produced by P. aeruginosa (including pyocyanin and HQNO; green dots) select for small colony variants (SCV) of S. aureus, a phenotype linked to increased antimicrobial tolerance or resistance and associated with accelerated lung function decline in a pediatric CF cohort. B)ld Host-secreted metabolites (red dots), including succinate and glutamate, stimulate the proton motive force (PMF) in P. aeruginosa, leading to enhanced killing by aminoglycoside antibiotics. Grey rods represent dead P. aeruginosa. C) Injection of the ExoS toxin into airway epithelial cells by P. aeruginosa’s type III secretion system stimulates production of a host phospholipase (sPLA2-IIA) that can lyse S. aureus and other Gram-positive bacteria. D) S. aureus (yellow cocci) modulates interaction of P. aeruginosa with the host. S. aureus secreted products dampen IL-8 production in response to P. aeruginosa. S. aureus secreted Staphylococcal protein A (SpA; peach dots) binds to the Psl polysaccharide in the biofilm matrix and type IV pili on the cell surface of P. aeruginosa, where it can subsequently protect P. aeniginosa cells from phagocytosis by neutrophils by binding and blocking the Fc region of IgG (purple), as well as protect P. aeruginosa from killing by tobramycin. E) The β-glucan rich biofilm matrix (light green) of C. albicans (brown) can bind and sequester antimicrobials, thereby protecting S. aureus. F) Fermentation of mucin by anaerobes releases short-chain fatty acids (SCFAs) and amino acids (teal dots). SCFAs stimulate an epithelial innate immune response, whereas both SCFAs and amino acids serve as a nutrient source for P. aeruginosa and other bacteria. G) Pf bacteriophage (red lines) is upregulated by P. aeruginosa during anaerobic and biofilm growth and has multiple effects on biofilm formation and the host response to infection. Pf phage promotes bacterial attachment to mucin and serves a structural role in the biofilm. Pf phage was also associated with reduced cytotoxicity and pro-inflammatory cytokine production, and alteration of macrophage polarization that lead to reduced P. aeruginosa phagocytosis. Furthermore, in separate studies, both Pf phage and respiratory viral infection have been shown to trigger an antiviral response that limits phagocytosis by murine macrophages and promotes P. aeruginosa biofilm growth in association with polarized human airway epithelial cells, respectively. H) S. aureus glucose fermentation leads to acidification of mucus and ASL (represented by H⁺), leading to changes in virulence factor production and impairment of the host innate epithelial response to infection. Lactate secreted by S. aureus as a result of glucose fermentation can promote P. aeruginosa biofilm growth (not pictured).

2.1. Feedback between microbe-nutrient and host-microbe interactions in the CF respiratory tract

Several research groups have characterized the nutritional and physiochemical environment of the lower respiratory tract (LRT) in CF in order to better model chronic CF respiratory infection. The most abundant carbon and nitrogen sources present are most likely to be host derived, including lipids (lung surfactant), peptides and amino acids (due to degradation of proteins by host proteases (5)), mucins, DNA (largely from neutrophil NETs), as well as glucose (especially in CFRD) and lactate (reviewed in (6,7)). Furthermore, the lower airways are a site of constant competition between microbes and the host for iron, in which the host’s mechanisms of nutritional immunity are at odds with bacteria and viruses alike (reviewed in (8)). While the host likely contributes the most abundant nutrient sources in the LRT, how the microbial populations present sense, respond to, and metabolize these nutrients drives host-microbe and microbe-microbe interactions. In this section, we will highlight examples of three such host-microbe interactions that demonstrate how microbial interactions with the host nutritional environment feed back into diverse changes in disease pathogenesis.

2.2. Dysregulation of host nutritional immunity and competition for iron driving microbial community behavior

In a healthy lung, the airway lumen is depleted for iron through a process called ‘nutritional immunity’. Exogenous iron present in the airways is transported intracellularly by lactoferrin and transferrin where it is stored bound to ferritin. In CF, high iron levels are present due to inflammation and resulting tissue damage, leading to abundant heme and iron-bound ferritin, lactoferrin, hemoglobin, and transferrin in airway secretions. The presence of host and bacterial proteases, as well as the low oxygen tension and acidic pH of airway secretions, results in the majority of iron in the CF lung being in the ferrous rather than ferric form (9). Regardless of the state—bound or unbound, ferric or ferrous—CF pathogens have numerous strategies for iron acquisition in this environment (reviewed in (10)). In addition to supporting microbial growth, iron can act as a signal for P. aeruginosa to form biofilms, dense aggregates of bacteria encased in an extracellular matrix that protects the bacteria from killing by antimicrobials or the host (11). Iron-mediated induction of alginate production by P. aeruginosa likely shapes other aspects of disease pathogenesis from both a host and a pathogen perspective. For example, alginate production is associated with a diminished host inflammatory response in part due to downregulation of the flagellum (12), increased mucin secretion by ferret tracheal epithelial cells (13), and decreased staphylolytic activity in a co-culture model of CF respiratory infection (14). Thus, whereas nutritional immunity drives intense competition for iron between the host and any microbes that find their way into a healthy lung, microbes can also sense iron availability due to dysregulated nutritional immunity in CF, in which case it acts as signal that drives microbial community behavior.

2.3. Mucus degradation by anaerobes leading to bacterial cross-feeding and feed-forward effects on inflammation

Dysregulated mucus secretion is a hallmark of CF, occurring in response to diverse signals ranging from tissue damage to bacterial and respiratory viral infections. Accumulation of thickened airway mucus generates heterogeneous microenvironments with steep pH and oxygen gradients, the latter of which is due to a combination of O₂ depletion by neutrophils as they generate an oxidative burst (15) and respiration by microbial communities present on the surface of the mucus plug (16). Mucin glycoproteins represent an abundant nutrient source for microbes in the CF airways. P. aeruginosa and other traditional CF pathogens inefficiently metabolize mucin; however, mucin fermenting bacteria such as Prevotella spp. and Veillonella spp. can improve the utility of mucins for the rest of the microbial community by generating short chain fatty acids (SCFAs; propionate and acetate) and free amino acids (17,18). SCFAs also induce concentration-dependent effects on inflammatory cytokine production and inducible nitric oxide synthase (iNOS) expression in CF airway epithelial cell cultures, potentially leading to delayed resolution of inflammation and neutrophil persistence in the airways (19).

Beyond its role as a nutrient source, mucus can mediate P. aeruginosa-bacteriophage dynamics with the potential to modulate the host immune response. Filamentous Pf phage genes were among the most highly upregulated genes during anaerobic respiration (20). Furthermore, Pf phage promotes P. aeruginosa adherence to mucin and leads to reduced invasiveness, inflammation, and susceptibility to phagocytosis by macrophages (21). Two recent studies have further elucidated a role for Pf phage in chronic CF P. aeruginosa infections. Pf phage was shown to be present in high densities in sputum from one quarter to one third of CF individuals in Danish and American cohorts, respectively (22). The authors went on to demonstrate relationships between Pf phage concentration in sputum and a number of clinical disease correlates (including an inverse relationship between Pf phage and FEV₁), as well as demonstrated how Pf phage can sequester a variety of antibiotics used in CF, leading to increased tolerance. Pf phage was also shown to elicit a maladaptive innate anti-viral response in vitro and in a murine burn model, promoting increased P. aeruginosa survival (23). Human respiratory virus infections and the resulting innate antiviral response is known to dysregulate nutritional immunity and promote P. aeruginosa biofilm growth, raising the question of whether Pf phage may confer a similar benefit to P. aeruginosa in the respiratory environment (24). Together, these studies suggest that anaerobic niches produced when mucus accumulates can be a site for microbe-microbe and microbe-bacteriophage interactions that alter disease pathogenesis.

2.4. S. aureus growth on glucose has implications for bacterial interspecies interactions and host-microbe interactions

The airway surface liquid of a healthy lung contains glucose concentrations approximately 12 times lower than in the bloodstream. However, airway glucose homeostasis is dysregulated during periods of airway inflammation, as a result of increased paracellular glucose flux, and during hyperglycemia, due to the increased transepithelial glucose gradient (reviewed in (25)). In individuals with CF-related diabetes (CFRD), glucose is present in the airway surface liquid (ASL) at even higher concentrations (approximately 4.0 ±2.1 mM) compared to individuals with CF or diabetes alone (26). Treatment with insulin is associated with fewer exacerbations, as well as preserved or increased lung function in recent clinical trials (reviewed in (27)), and metformin reduced hyperglycemia-induced P. aeruginosa and S. aureus growth in an in vitro airway cell culture model (28). Furthermore, CFRD and S. aureus colonization were associated with worse sinus disease in a study of adults with CF chronic rhinosinusitis (29). Together, these studies suggest glucose may be an important nutrient source in the airways and that targeting this nutrient may improve outcomes for individuals with CF.

While diverse and potentially more abundant carbon sources are present in the airways of individuals with CF and CFRD, S. aureus is likely to preferentially use glucose over these other carbon sources due to carbon catabolite repression, with ASL glucose concentrations in CFRD being within the range of levels that activate catabolite repression during S. aureus in vitro growth (4mM; (30)). S. aureus can grow aerobically and anaerobically on glucose, and the organic acid lactate is abundantly secreted following glucose fermentation, leading to decreased pH (31) and increased P. aeruginosa biofilm growth (32). Acidification of the ASL as a result of S. aureus glucose fermentation could lead to diverse impairments of mucosal immunity, including increased mucus viscosity and inactivation of antimicrobial peptides, as well as impaired phagocyte function (reviewed in (33)). Glucose catabolism and a resulting decrease in pH is known to inhibit the Agr quorum sensing system in S. aureus, leading to changes in virulence factor expression (reviewed in (34)) with the potential to feedback into alterations to host-microbe and microbe-microbe interactions. Similarly, P. aeruginosa was recently shown to preferentially grow on glucose-derived metabolites from Rothia mucilaginosa (an oral bacterium also present in CF sputum) in vitro, and the authors discuss how cross-feeding from R. mucilaginosa has implications for P. aeruginosa pathogenesis through glutamate production(35).

3.1. Bacterial interspecies interactions mediated by the host innate and adaptive immune response to infection

The microbiota plays a key role in the modulation of host immune system development, function, and functional tuning by various mechanisms (reviewed in (36)). Conversely, the host innate immune response to infection can also modulate microbial community composition. S. aureus and P. aeruginosa frequently co-infect in CF, with a progression toward P. aeruginosa dominance over time that tends to coincide with the age range when lung function declines most rapidly. However, the mechanisms by which S. aureus is progressively replaced by P. aeruginosa in CF airways remains unclear (reviewed in (37)). In this section, we describe examples of how interspecies interactions between bacteria in the CF respiratory environment can be mediated by the host. These studies highlight the importance of considering a particular microbe within the context of the microbiota and the CF airways; the complex interactions that emerge when polymicrobial communities interact with the host would be difficult to predict by examining individual microbes separately.

3.2. Impact of a host phospholipase on microbial interactions in CF

Mammals produce a variety of secreted phospholipase A2 enzymes that are selectively expressed in different tissues and display unique substrate specificities, suggesting that they perform unique functions (reviewed in (38)). The secreted phospholipase A2 (sPLA2) family, including the group IIA sPLA2 (sPLA2-IIA), belongs to a superfamily of mammalian enzymes called PLA2s which hydrolyze membrane phospholipids of both eukaryotic and prokaryotic cells (39) and participate in a variety of processes such as inflammation and host defense mechanisms (39,40). This secreted phospholipase is among the most potent bactericidal agents produced by mammals and is, for example, the principal bactericidal molecule in human tears (41) and broncho-alveolar lavage of patients with acute respiratory distress syndrome (42). In CF, sputum levels of sPLA2-IIA have been shown to increase between the ages of 5 and 20 years old (43). Immuno-histochemical analyses of bronchial explants from individuals with CF revealed a dramatic expression of sPLA2-IIA and confirmed that bronchial epithelial cells (BECs) are the major cell source of this enzyme, although sPLA2-IIA staining was also detected in infiltrating neutrophils (43). Among mammalian PLA2s, sPLA2-IIA has been shown to exert potent bactericidal functions, especially on Gram-positive bacteria such as S. aureus, with no effects on host cells at physiological concentrations (39,40). The sPLA2-IIA levels in CF expectorations were sufficient to kill S. aureus without significant effects on P. aeruginosa. Furthermore, sPLA2-IIA levels are induced in BECs in response to delivery of ExoS toxin by P. aeruginosa’s type III secretion system into BECs (43). This secretion system is found in several Gram-negative bacteria and known to modulate the virulence of these bacteria and their action on host immunity (44). Thus, induction of host-produced sPLA2-IIA by P. aeruginosa may shape the CF airway microbiome by selectively eliminating S. aureus and other Gram-positive species, thereby creating a niche favorable to other bacterial species that are resistant to killing by sPLA2-IIA (including P. aeruginosa). Host-produced sPLA2-IIA is likely to be one part of a complex and multi-factorial process involved in the progressive and selective elimination of S. aureus from CF airways. For example, the presence of S. aureus or its secreted products have been shown to dampen IL-8 production by AECs in response to subsequent P. aeruginosa exposure (45) and protect P. aeruginosa from phagocytosis by neutrophils (46) and killing by tobramycin (47).

Additional mechanisms by which the host innate or adaptive immune response to one bacterial species can be manipulated by the other species have been described beyond P. aeruginosa and S. aureus. In the case of Streptococcus pneumoniae, production of the protease LasB by P. aeruginosa was shown to cleave specific complement molecules and suppress reactive oxygen species production by alveolar macrophages (AM), leading to impairment of S. pneumoniae killing by AM in murine and in vitro cell culture studies (48). Furthermore, a recent longitudinal study of the airway microbiota of infants with CF described an anti-correlation between Streptococcus spp. and Haemophilus spp. abundance (49). The authors hypothesize the low pH of CF ASL could be responsible, citing the observation that S. pneumoniae can outcompete H. influenzae under low pH conditions in an in vitro co-culture system (50). Finally, measures of airway inflammation in CF sputum from adult and pediatric cohorts was observed to be inversely correlated with microbial diversity and relative abundance Streptococcus spp. (and other anaerobes), independent the relative abundance of P. aeruginosa (51). Together, these studies underscore the importance of the host in modulating the airway microbiota composition during the progression of CF respiratory disease.

4.1. Role of interspecies interactions and host-pathogen interactions in susceptibility of biofilms to antimicrobials

The biofilm mode of growth, in which aggregates of microbes are encased in an extracellular matrix that can be either host-derived or self-produced, is thought to play a key role in survival and persistence of microbes in the CF respiratory tract. In this section, we summarize three aspects of biofilm physiology and sociality underpinning their increased tolerance or resistance to antimicrobials used in CF.

4.2. Enzymatic degradation of antibiotics — ‘indirect pathogenicity ‘

‘Indirect pathogenicity’ refers to the protection of a susceptible isolate against the action of an antibiotic due to the enzymatic degradation of the antibiotic by a resistant isolate (52). This passive resistance is not limited to biofilms, but as antibiotic-degrading enzymes can accumulate in the biofilm matrix, it is likely much more efficient under these conditions. Examples include the protection of one species from killing by β-lactam antibiotics through β-lactamases produced by another species (e.g. Moraxella catarrhalis protecting H. influenza (53)) and the protection against aminoglycosides through production of aminoglycoside-modifying enzymes (e.g. P. aeruginosa protecting S. aureus (54)).

4.3. Modification of penetration of antibiotics through the biofilm and into cells

The extracellular matrix produced by cells in a microbial biofilm consists of polysaccharides, proteins and DNA. This slime layer can significantly slow down the penetration of reactive oxidants and cationic molecules (55), as well as that of larger complexes like liposomes and nanoparticles (56). In multispecies biofilms, matrix components produced by one species may provide protection for the other. For example, in mixed Candida albicans — S. aureus biofilms, the extracellular polysaccharide β-l,3-glucan produced by C. albicans protects S. aureus against vancomycin (57).

Penetration of antibiotics through the biofilm is of course only the first step: to exert their antimicrobial activity these antibiotics must make it into the cell. However, interactions between organisms in multispecies communities can lead to altered antimicrobial susceptibility. One such example is through modification of the cell wall thickness and lipopolysaccharides (LPS). The increased tolerance of Streptococcus anginosus to cell-wall acting antibiotics (including vancomycin) when grown in a multispecies biofilm with P. aeruginosa and S. aureus (58) is related to an increased cell wall thickness, and that factors secreted by S. aureus mediate this increased tolerance (59). A second example is the alteration of the LPS structure of P. aeruginosa when grown together with S. aureus, which gave rise to increased β-lactam resistance (60). Taken together, these studies underscore how polymicrobial interactions pose a challenge to predicting antimicrobial susceptibility profiles based on standard clinical laboratory techniques.

4.4. Influence of biofilm metabolism on antimicrobial susceptibility

Biofilms contain regions with low metabolic activity, and this contributes to reduced susceptibility towards antibiotics (reviewed in (61)). While it is clear that metabolic adaptations to the biofilm lifestyle also occur in single-species biofilms, more complex interactions occur in multi-species biofilms (reviewed in (62)), and between microorganisms and the host in host-associated biofilms. These complex interactions increase biofilm heterogeneity and as the microenvironment of the biofilm plays a crucial role in reduced susceptibility, obtaining a better understanding of metabolic adaptations to the biofilm microenvironments will be crucial to obtain a more fundamental understanding of reduced biofilm susceptibility. For example, it was recently observed that metabolites secreted by three-dimensional respiratory epithelial cell cultures (including glutamate and succinate) enhance the activity of aminoglycoside antibiotics against P. aeruginosa biofilms (by stimulating the proton motive force and increasing antibiotic uptake) (63). In contrast, colistin was found to be less effective against P. aeruginosa biofilms in the presence of respiratory epithelial cells (64).

In terms of bacterial interspecies interactions, a number of P. aeruginosa secreted products, including proteases and respiratory inhibitors, have been shown to promote S. aureus biofilm dispersion or to select for the small-colony variant (SCV) phenotype of S. aureus, leading to alterations in antimicrobial tolerance or resistance (65-69). These SCVs have been associated with accelerated lung function decline in a pediatric CF cohort (70), raising the question of whether they are causal or a marker of disease progression. Finally, the link between metabolic adaptations in biofilms and reduced susceptibility to antimicrobial agents creates opportunities for innovative treatment approaches. One such example is the use of hyperbaric oxygen therapy to increase metabolic rates in aerobic organisms and by doing so revert tolerance (71). Similarly, there are several studies in which activation of particular central metabolic pathways (such as the TCA cycle) by providing specific metabolites led to reduced tolerance (reviewed in (61)).

5.1. Summary

The CF respiratory tract contains heterogeneous nutritional and physiochemical microenvironments that drive microbial interactions and pathogenesis.

The lower respiratory tract in cystic fibrosis is a nutritionally complex, heterogeneous, and dynamic environment that supports microbial life during decades-long infections (Figure 1). This complexity is evidenced by the variety of carbon and nitrogen sources available for microbial consumption, ranging from amino acids to simple carbohydrates to complex mucins and DNA. Furthermore, across multiple individuals with CF, there is likely considerable heterogeneity in nutrients present in the airways depending on an individual’s disease state or class of CFTR mutation, the presence of co-morbidities such as CFRD, as well as an individual’s treatment regimen, current exacerbation status, and the functional capacity of their respiratory microbiota. Even within one person, the physiochemical and nutritional environment differs throughout the lower respiratory tract, from the conducting airways where mucus buildup can produce regions of hypoxia and into the bronchioles and alveolar sacs that are drenched in surfactant. These microenvironments are further modified based on localized damage and obstruction, inflammation, and immune infiltrate. Adding to this complexity is variation in the presence, abundance, and relative proximity of different microbial species throughout the respiratory tract, which can further differ at the strain level in terms of accessory genomic contents (72) and within-patient evolved traits (reviewed in (73)). Finally and importantly, the respiratory epithelium and resident or recruited immune cells each represent dynamic and renewing sources of nutrients and physiochemical alterations as the host itself senses and responds to nutritional and infectious cues in its environment. Future studies of CF microbial pathogenesis should consider both the potential for strain-level variability in microbial behaviors and the role of the host as an active participant in shaping host-microbe and microbe-microbe interactions.

5.2. Future directions

Together with the lack of accessible animal models that accurately recapitulate the chronic respiratory infections present in CF, the complex, heterogeneous, and dynamic nature of the CF respiratory environment continues to provide a challenge to the field. Looking ahead, we anticipate continued growth in research at the interface of host-pathogen interactions and microbial ecology in CF. We are optimistic that clinically-relevant hypotheses generated from multi-omic studies (e.g. studies integrating 16S amplicon microbiome or metagenomic sequencing with metatranscriptomics, metaproteomics, or metabolomics (74,75)) of CF clinical specimens can be tested using computational, in vitro, or animal models (including porcine explant models (76)) as appropriate. Furthermore, advances in microscopy methods such as the MiPACT tissue-clearing technique (77) are likely to play an important role in examining the biogeography of putative microbial interactions with each other and the host in situ. Finally, we expect that the fields of CF microbial pathogenesis and ecology will continue to evolve as the use of CFTR modulator therapies (reviewed in (78)) become more widespread and change CF disease.

Highlights.

• Interactions between microbes and nutrients in the CF lung feedback into diverse effects on respiratory disease pathogenesis.

• The respiratory innate or adaptive immune response can influence microbial community structure and function.

• Antimicrobial tolerance or susceptibility of biofilm bacteria can be altered by polymicrobial interactions.