Targeting iron uptake to control Pseudomonas aeruginosa infections in cystic fibrosis

Introduction

Pseudomonas aeruginosa is an aerobic Gram-negative bacterium which is widespread in the terrestrial environment. It is extremely robust and capable of surviving in challenging and varied environmental niches, as exemplified by its isolation from jet plane fuel and bottles of disinfectant fluid [1]. This adaptability is conferred by its large genome (approximately 6 Mb) and ability to survive as either a planktonic organism or as a member of a codependent bacterial community within the confines of a “biofilm” [2, 3].

The genetic plasticity and biofilm-forming attributes of P. aeruginosa make it a highly successful pathogen in multiple disease settings in eukaryotes. In humans, P. aeruginosa is an opportunistic pathogen, which is responsible for life-threatening acute infections in burn victims and other critically ill patients, as well as chronic infections and acute exacerbations in patients with respiratory diseases [4–6].

Iron is essential to the survival of virtually all prokaryotes and eukaryotes. The importance of iron to P. aeruginosa is exemplified by the fact that 6% of its transcribed genes are iron-responsive [7]. The concentration of bioavailable iron is a powerful regulator of P. aeruginosa behaviour, influencing intercellular communication and biofilm formation [7].

In this review we explore how iron availability within the lung influences the development of chronic P. aeruginosa biofilm infection in people with the autosomal recessive genetic disorder cystic fibrosis (CF), and examine current research into how the iron dependency of P. aeruginosa may be targeted therapeutically.

The susceptibility of the CF airway to infection

The CF airway is inherently prone to infection. In health, the luminal surface of the respiratory epithelium is coated with airway surface liquid (ASL), comprised of mucins, immune cells and antimicrobial peptides. ASL traps and kills inhaled pathogens which are then rapidly cleared by the mucociliary escalator. In CF, impaired function of the CF transmembrane conductance regulator (CFTR) on respiratory epithelial cells results in increased reabsorption of water from the airway lumen and dehydration of the ASL, with consequent slowing of mucociliary clearance [8]. In addition, defective CFTR-mediated bicarbonate export has been shown in animal models to result in a fall in ASL pH and further inhibition of ASL antimicrobial activity [9]. A similar acidic environment exists in human disease [10]. These alterations in the biophysical properties of ASL are compounded by deficits in airway innate immune defences, including defective iron sequestration and degradation of antimicrobial peptides by high concentrations of endogenous and bacterial-derived proteases, which produce an environment conducive to chronic infection [11, 12].

As CF lung disease progresses, plugging of distal airways by dehydrated, inspissated mucus creates microaerobic or frankly anaerobic pockets within the normally aerobic environment [13, 14]. This low oxygen environment drives phenotypic adaptation in incumbent bacteria and promotes the survival of bacteria capable of existing at low oxygen tensions [13]. Bacterial respiration may further lower oxygen tensions and potentially contribute to alteration in the pH of ASL, which will further impair the bactericidal effects of several antibiotics (especially aminoglycosides) commonly used in CF [13, 15].

Respiratory tract infections in CF begin very early in life [16]. Initial intermittent infections are typically caused by the common respiratory pathogens Staphylococcus aureus and nontypeable Haemophilus influenzae [17]. By adulthood, a chronic polymicrobial airway infection develops, with P. aeruginosa becoming the dominant pathogen in 80% of cases [17, 18]. Chronic P. aeruginosa infection leads to an increased rate of lung function decline, morbidity and mortality [19]. Recent culture-independent (metagenomic) microbiological techniques suggest that a wide range of additional bacterial species may also infect the CF airway (including anaerobes), although little is currently known about the pathological significance of these microbes [20, 21]. A key factor in the interplay between host tissues and bacterial pathogens is the management of iron metabolism. The lung is exposed daily to a high oxygen concentration, and unbound iron in atmospheric particulate matter can potentially catalyse the formation of reactive oxygen species, as can ferrous and ferric iron in ASL. This provides the lung with unique challenges with regards to iron homeostasis [22]. Airway cells rapidly sequester iron to prevent the generation of damaging free radicals, and to withhold this key nutrient from inhaled pathogens. This is achieved through uptake of nonprotein-bound iron by divalent metal-ion transporter 1 on the apical surface of bronchial epithelial cells and by the secretion of the iron chelating proteins lactoferrin and transferrin into ASL [23].

The lung is highly adept at iron detoxification and iron is barely detectable in normal airway secretions. The resulting lack of accessible iron inhibits the growth of infectious bacteria. However, respiratory secretions and sputum from patients with CF contain micromolar concentrations of iron, making this micronutrient more readily available to inhaled pathogens (airway iron indices from existing studies are presented in table 1) [24–28]. In vitro data suggest that this increase in lung iron may partly be due to defective iron handling by CF bronchial epithelial (CFBE) cells [12].

Neutrophils represent the first line of cellular defence against bacterial pathogens and also participate in iron-withholding by secreting lactoferrin and lipocalins. Lipocalin 2 binds and inactivates bacterial-derived iron scavenging molecules (siderophores), although it is not thought to bind to the P. aeruginosa-derived siderophores [29, 30]. The role of lipocalin 2 in the setting of polymicrobial infection has not been explored in CF, although serum levels increase when patients develop an increased infective burden [31].

The development of P. aeruginosa biofilms in CF airways

Following initial airway infection, planktonic P. aeruginosa undergoes rapid phenotypic and genotypic adaptation to prevent immune recognition. This is achieved by the formation of a biofilm, which offers physical protection and downregulation of virulence factors [2, 32].

Biofilms comprise an extracellular matrix (ECM) of exopolysaccharides, extracellular DNA (eDNA) and proteins produced by the resident bacteria. By trapping essential nutrients and providing a physical barrier to host immune attack, biofilms offer a survival advantage to embedded bacteria. In the CF lung it is proposed that P. aeruginosa binds abnormal mucins present in ASL to form biofilm “rafts” which float on the respiratory epithelium [32]. Established biofilm infections cannot be eradicated with currently available antibiotics or by the host's neutrophilic inflammatory response [33].

Biofilm development is largely determined by its environment and available nutrients [34]. In vitro, biofilms develop complex three-dimensional structures containing phenotypically distinct subpopulations of bacteria connected by water channels formed within the ECM [35]. Iron is essential as a bacterial nutrient, and lack of iron interferes with biofilm development [36]. Iron also contributes to the structural integrity of the biofilm by cross-linking exopolysaccharide strands [37].

P. aeruginosa biofilm development is dependent on cell–cell communication. Quorum sensing is a population density dependent form of communication employed by bacteria to control the synthesis of key regulatory proteins. Quorum sensing is integral to all activities of the bacterial community, including biofilm formation. P. aeruginosa employs three quorum sensing systems (Las, Rhl and Pseudomonas quinolone signal (PQS)), each of which is iron responsive [38–41].

Under conditions of limited iron availability, both the Las and Rhl systems are activated [38, 39]. The relationship between the PQS system and iron is complex. PQS is able to operate as an iron chelator, thereby controlling the activation of the Las and Rhl systems through iron limitation [42]. Conversely, PQS synthesis is increased under conditions of both iron limitation and excess [40].

An important gene cluster under the control of Rhl is the rhlAB operon that regulates production of the biosurfactant rhamnolipid. Rhamnolipid acts as a “wetting agent”, reducing surface tension and promoting surface-associated movement (twitching motility). Correctly timed production of small quantities of rhamnolipid is critical for the production of water channels within the core of mature P. aeruginosa biofilms, through which motile bacteria are able to travel. In vitro, inhibition of rhamnolipid production leads to the formation of flat, thick, immature biofilms [43]. In contrast, excessive rhamnolipid promotes the dispersal of mature biofilms and causes newly formed biofilms to be thin and flat [44, 45]. When iron availability is limited, increased rhamnolipid production and twitching motility prevents biofilm development, or triggers biofilm dispersal, depending on the stage of biofilm maturation [44, 46].

Paradoxically, supraphysiological iron concentrations appear to be detrimental to biofilm development. Normal human plasma contains 20–25 μM of iron, <1 μM of which is protein-bound. Biofilms grown in medium containing 100 μM of iron contain less eDNA, fail to develop complex macrocolonies and are more susceptible to antimicrobials compared with biofilms grown in an equivalent 1 μM iron medium [47]. Similarly, exposing established biofilms to medium containing 200 μM of ferric ammonium citrate triggers dispersal events and facilitates antibiotic killing [48].

Iron acquisition by P. aeruginosa

P. aeruginosa may take up iron from either haem or nonhaem iron sources (fig. 1). Two haem uptake systems have been described in P. aeruginosa (Phu and Has) [49]. The Phu system relies on direct binding of haem or haem-containing proteins to a membrane-bound receptor, whereas the Has system secretes a haem-binding protein (HasAp) which is reabsorbed through the Has receptor (HasR) when bound to haem [50, 51]. The P. aeruginosa genome contains a third haem receptor-encoding gene (hxuC); however, its functional regulation has yet to be characterised [52]. It is unknown whether haem uptake systems are employed by P. aeruginosa within the CF lung; however, patients frequently have frank blood in their sputum and subclinical bleeding into the airway is probably common.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1–

Pseudomonas aeruginosa iron acquisition pathways.

Haem is an uncommon iron source in the natural environment and P. aeruginosa must also be capable of scavenging nonhaem iron, which under aerobic conditions is most probably present in the poorly soluble ferric (Fe³⁺) form. P. aeruginosa (and other bacteria and fungi) therefore produces high-affinity iron chelating siderophores [53]. Siderophores are secreted by P. aeruginosa into the local environment to chelate free iron and “strip” iron from host iron-binding proteins.

Two distinct siderophores have been characterised in P. aeruginosa: pyoverdine and pyochelin. >50 distinct pyoverdine subtypes have been characterised and are responsible for the distinctive yellow-green fluorescence of certain pseudomonads [54]. Pyoverdines are the primary siderophore produced by P. aeruginosa, with one of three distinct subclasses being produced by individual strains [55].

Pyochelin is considered a secondary siderophore in P. aeruginosa, having a much lower iron binding affinity than pyoverdine [52, 53]. Pyochelin appears to have less influence on the biofilm forming capacity of P. aeruginosa than pyoverdine, and its importance for iron acquisition during clinical airway infections is unclear [36, 53]. In addition to acquiring iron using autologous siderophores, P. aeruginosa has a high capacity to take up iron-laden siderophores produced by other bacteria and fungi [52].

P. aeruginosa, while naturally an aerobic bacterium, is capable of adapting to low oxygen environments such as those encountered within plugged CF airways. Within these regions of low oxygen tension and low pH there is potential for the redox status of iron to change to the more “soluble” ferrous (Fe²⁺) form, but there are currently no data on this scenario in CF lung disease. Ferrous iron may be acquired by P. aeruginosa by passive diffusion or uptake through the FeoB receptor, although the role of these mechanisms in the clinical setting is at present unclear [56].

P. aeruginosa iron acquisition systems are tightly controlled by the ferric uptake regulator (Fur). Fur acts both directly and indirectly, through extracytoplasmic sigma factors (including PvdS), to limit iron absorption [57]. Under iron-replete conditions, Fur binds ferrous iron and attaches to a consensus sequence (Fur-box) in the promoter region of genes instrumental in iron acquisition, thus suppressing their transcription [58]. In the presence of iron, Fur inhibits iron conservation strategies by suppressing the production of two small RNAs (PrrF1 and PrrF2) [59]. In the absence of iron these small RNAs are synthesised and facilitate inhibition of genes that encode “nonessential” iron-containing proteins, thereby maintaining the cytoplasmic iron pool for essential use [60]. Under low iron environments siderophore synthesis increases and nonessential iron-consuming processes are downregulated. Several excellent comprehensive reviews of the iron acquisition systems employed by P. aeruginosa have recently been published [36, 52, 53, 57, 60], but the above overview highlights the central role of iron in P. aeruginosa biofilm development.

Targeting bacterial iron acquisition as a therapeutic strategy

The critical role of iron in P. aeruginosa survival and biofilm formation may represent a potential “Achilles’ heel” in the defensive armamentarium of this fastidious pathogen. Thus considerable research endeavours on a variety of fronts are being undertaken to develop novel therapeutic strategies based on disruption of bacterial iron homeostasis. These therapeutic strategies may be particularly important in CF where host iron homeostatic mechanisms appear to be abnormal.

Strategies to limit iron in the setting of a polymicrobial infection

Any new intervention directed against P. aeruginosa must consider the potential impact on copathogens, as suppression of the dominant pathogen may allow the emergence of other, potentially more harmful, infections.

In common with P. aeruginosa, other commonly isolated CF airway pathogens, including S. aureus, H. influenzae and Burkholderia cepacia complex (BCC), are capable of biofilm development and each have an absolute requirement for iron [92–96].

In a single published study on the effect of gallium on planktonic and biofilm grown BCC, strains were exposed to gallium nitrate at concentrations of up to 64 mg·L⁻¹ (∼250 μM Ga³⁺) [97]. Disappointingly, there was little effect seen on either planktonic or biofilm growth. These results have been challenged on the basis that the concentration of gallium used was lower than could be safely administered therapeutically [98]. However, in a similar study examining the effects of gallium maltolate on the growth of S. aureus and S. epidermidis biofilms, equally disappointing results were reported, and minimal inhibitory concentrations far in excess of those that could be safely administered systemically (>3000 mg·L⁻¹) were needed to achieve biofilm inhibition [99].

There are few studies of iron chelator effects on CF bacterial pathogens other than P. aeruginosa (table 3) [100–103]. The effect of the synthetic chelators deferiprone and deferoxamine against a number of staphylococcal species grown in broth cultures has been examined [103]. Deferiprone inhibited growth of all species studied, but desferrioxamine promoted growth in a number of staphylococcal species [103]. Similarly, it has been demonstrated that S. aureus can take up iron hydroxamates such as desferrioxamine and utilise them as an iron source to promote biofilm growth [101, 104].

Translational research and the challenges of targeting P. aeruginosa iron homeostasis in the human lung

Despite the early promise of a number of the agents discussed above in vitro, important questions remain to be answered about their safety and efficacy before advancing to human trials.

The majority of the work presented above has been performed using common laboratory-adapted strains of P. aeruginosa, which vary both genetically and phenotypically from clinical strains isolated from the CF lung. Additionally, studies have considered only a limited number of environmental variables and often use conditions that are distinct from those within the CF lung, where there is reduced oxygen tension, significant amounts of extracellular iron, low pH and a hostile milieu replete with proteases and free radicals [2, 10, 13]. In the very limited work performed with clinical isolates, different responses to iron-targeted therapies have been reported, both between clinical and laboratory strains, and between clinical isolates from different patients [73].

Although there have been no studies of treatments targeting bacterial iron homeostasis under “CF lung conditions”, factors including pH, glucose source and oxygen availability have been shown to affect the biofilm-forming capacity of airway pathogens [32, 73, 101]. Consequently, if new agents are to be successful they must remain active over a wide pH range, and compete with both ferrous and ferric iron acquisition systems.

Iron limitation in vitro triggers the dispersal of motile planktonic bacteria with increased virulence compared to their biofilm-dwelling counterparts, and thus the potential for biofilm disruption to trigger an acute host inflammatory response [105]. To better understand the inflammatory potential of these agents testing in an animal model is desirable; however, representative models of CF airway infection are limited. Mice containing the major CFTR gene mutations (e.g. DeltaF508, G551D) do not develop spontaneous airway infections and P. aeruginosa has to be introduced directly into the mouse lung where it is either spontaneously cleared or results in overwhelming infection [106, 107]. Successful chronic mouse airway infection has been achieved by introducing P. aeruginosa bound to agar beads into the trachea and by contaminating drinking water with P. aeruginosa [108], but how closely this reflects human disease is debated. More recently, pig and ferret models of CF have been developed, which may more closely mimic human respiratory disease [109, 110].

Finally, the route of administration must be considered. The concentrations of gallium required for activity against S. aureus and BCC biofilms are well above those considered safe for systemic delivery in humans, suggesting that inhalation may be the only viable option to safely administer the required dose. Similarly, in vitro studies suggest iron chelators delivered directly to biofilms grown on the apical membrane of CFBE cells inhibit growth more effectively than when they are applied to the basal membrane, suggesting that direct delivery to the airway may also be the preferred mode of delivery for these compounds [76]. The possibility of localised delivery of chelators is supported by in vitro modelling, which has suggested that chelated iron may be effectively aerosolised to a particle size suitable for lung delivery [61].

Conclusions

As our understanding of the biology of bacterial biofilms expands, new therapeutic possibilities present themselves. Given the absolute requirement for iron of P. aeruginosa and other CF airway pathogens, disrupting iron utilisation is an exciting avenue for further research. The results of the safety trial of intravenous gallium are eagerly awaited. Future studies of iron chelation therapy will need to test the efficacy of these agents against clinically relevant P. aeruginosa strains and establish their safety within animal models, before proceeding to human trials.