Phenazine Content in the Cystic Fibrosis Respiratory ... - ATS Journals

Phenazine Content in the Cystic Fibrosis Respiratory Tract Negatively Correlates with Lung Function and Microbial Complexity Ryan C. Hunter1,3*, Vanja Klepac-Ceraj4*, Magen M. Lorenzi5, Hannah Grotzinger1, Thomas R. Martin5, and Dianne K. Newman1,2,3 1

Division of Biological Sciences, and 2Division of Geological and Planetary Sciences, California Institute of Technology, Pasadena, California; Howard Hughes Medical Institute, Pasadena, California; 4Department of Molecular Genetics, Forsyth Institute, Cambridge, Massachusetts; and 5Department of Pediatrics, Children’s Hospital Boston, Massachusetts 3

Although much is known about how virulence factors affect pathogens and host tissues in vitro, far less is understood about their dynamics in vivo. As a step toward characterizing the chemistry of infected environments, we measured phenazine abundance in the lungs of patients with cystic fibrosis (CF). Phenazines are redoxactive small molecules produced by Pseudomonas aeruginosa that damage host epithelia, curb the growth of competing organisms, and play physiologically important roles in the cells that produce them. Here, we quantify phenazines within expectorated sputum, characterize the P. aeruginosa populations responsible for phenazine production, and assess their relationship to CF lung microflora. Chemical analyses of expectorated sputum showed that the concentrations of two phenazines, namely, pyocyanin (PYO) and phenazine1–carboxylic acid (PCA), were negatively correlated (r ¼ 20.68 and 20.57, respectively) with lung function. Furthermore, the highest phenazine concentrations were found in patients whose pulmonary function showed the greatest rates of decline. The constituent P. aeruginosa populations within each patient showed diverse capacities for phenazine production. Early during infection, individual isolates produced more PYO than later during infection. However, total PYO concentrations in sputum at any given stage correlated well with the average production by the total P. aeruginosa population. Finally, bacterial community complexity was negatively correlated with phenazine concentrations and declines in lung function, suggesting a link to the refinement of the overall microbial population. Together, these data demonstrate that phenazines negatively correlate with CF disease states in ways that were previously unknown, and underscore the importance of defining in vivo environmental parameters to better predict clinical outcomes of infections. Keywords: phenazine; cystic fibrosis; Pseudomonas aeruginosa; biofilm; microbiome

Cystic fibrosis (CF), the most common inherited chronic disease in Caucasians, arises from mutations in the gene encoding the cystic fibrosis transmembrane conductance regulator (CFTR) (1).

(Received in original form March 4, 2012 and in final form July 24, 2012) * These two authors contributed equally to this work. This work was supported by the Howard Hughes Medical Institute (D.K.N.); by the Canadian Cystic Fibrosis Foundation (R.C.H); and by a grant from the Cystic Fibrosis Foundation Center to Children’s Hospital Boston for the sequencing of samples. V.K.-C. is presently at the Department of Biological Sciences, Wellesley College, 106 Central St., Wellesley, MA 02481. Correspondence and requests for reprints should be addressed to Dianne K. Newman, Ph.D., Division of Biology, 147-75, California Institute of Technology, 1200 E. California Blvd., Pasadena, CA 91125. E-mail: [email protected] This article has an online supplement, which is accessible from this issue’s table of contents at www.atsjournals.org Am J Respir Cell Mol Biol Vol 47, Iss. 6, pp 738–745, Dec 2012 Copyright ª 2012 by the American Thoracic Society Originally Published in Press as DOI: 10.1165/rcmb.2012-0088OC on August 3, 2012 Internet address: www.atsjournals.org

Since the discovery of the CFTR gene (2), our understanding of CF lung disease progression has increased substantially. However, unresolved controversies remain regarding the pathogenesis of CF lung infections and the direct link between CFTR mutations and the microbial colonization of airways (3). It is generally accepted that in the respiratory tract, abnormal CFTR function leads to deficiencies in chloride and bicarbonate secretions, and to excessive water adsorption that dehydrates the liquid layer bathing the epithelial surface (1). The subsequent accumulation of viscous airway secretions, impaired mucociliary clearance, abnormal mucosal immune response, decreased pH of the airway surface liquid (ASL), and reduced killing by ASL antimicrobials all favor the establishment of chronic infections (1, 4–8). In turn, bacterial colonization of the airways incites a persistent inflammatory response, leading to irreversible pulmonary decline and eventual respiratory failure (9). Chronic colonization by Pseudomonas aeruginosa is recognized as a major determinant of pulmonary decline, yet among patients who acquire this pathogen (regardless of socioeconomic status, healthcare provider, and therapy compliance) (10), marked variation exists in the rate of disease progression. Although recent studies have begun to investigate the link between antibiotic use, CFTR genotype, microbial community composition, and rates of pulmonary decline, the basis of this variation remains poorly understood (11, 12). In a range of infectious contexts, a surge in studies has suggested the importance of polymicrobial interactions as a determinant of host health (13–15). This is particularly true of the CF airways where, in addition to P. aeruginosa, a diverse microbial community exists within both pediatric and adult populations (12, 16–25). Cross-sectional profiling has revealed that the airway microflora is most complex during early stages of disease, and total species diversity is thought to decrease with age (12, 22). Some variation in community composition has been attributed to antibiotics as well as to the introduction of P. aeruginosa into the airways. However, the arrival of P. aeruginosa appears to result in a stochastic shift in the existing microbial community composition, or else additional environmental factors help shape the respiratory tract microbiome (12). Thus, defining the effects of lung environment on the overall composition of the CF lung bacterial community may hold value in understanding the etiology of CF disease progression. Factors that might shift the pulmonary environment and the resident bacterial community include a diverse suite of metabolites produced by the constituent members of the microflora. Among these is pyocyanin (PYO), the characteristic blue–green pigment of P. aeruginosa comprising one member of a class of redox-active compounds known as phenazines. Both PYO and its precursor, phenazine-1–carboxylic acid (PCA), exert a range of toxic effects upon the host and competing microorganisms, many of which can be attributed to their redox activity and the generation of reactive oxygen species (Figure 1) (26–29). Although phenazine-

Hunter, Klepac-Ceraj, Lorenzi, et al.: Phenazine Content and CF Lung Decline

induced oxidative stress may confer a competitive advantage to P. aeruginosa, these compounds have also been demonstrated to contribute to bacterial virulence in the absence of competitors, implying additional physiological roles (30–32). Indeed, phenazines have been shown to function as extracellular electron shuttles that facilitate iron acquisition (33), modulate redox homeostasis (34), serve as intercellular signals (35), affect biofilm architecture (36, 37), and support survival during hypoxia (38). In addition to benefiting P. aeruginosa, however, phenazines may provide an advantage to other organisms capable of using them. Alternately, even if they cannot directly use phenazines per se, some bacteria may potentially adapt to the environmental conditions phenazines help generate (e.g., the consumption of oxygen (39)), and rise in prominence. Although phenazines have been recovered at high concentrations from the respiratory tracts of patients with CF (40), little is known of how their presence relates to disease progression. Given the range of toxic effects that phenazines exert on host tissue and the potential for these compounds to affect bacterial physiology, we hypothesized that phenazine concentrations would rise as pulmonary function declined, and would correlate with a less complex CF microbial community composition. Here, we present data from a cross-sectional study of adult patients with CF at Children’s Hospital Boston that tested this hypothesis.

MATERIALS AND METHODS Sample Collection Forty-seven participants with CF were recruited during scheduled visits to Children’s Hospital Boston (CHB). Inclusion criteria comprised a diagnosis of CF, based on genotyping and sweat chloride testing (. 60 mEq/L according to pilocarpine electrophoresis), and chronic P. aeruginosa infection (positive culture for . 1 yr). Sputum was processed within 4 hours of expectoration by shearing through a 1-ml syringe, and was homogenized in equal volumes of 1 mM Sputolysin (Calbiochem, San Diego, CA) for 30 minutes. An aliquot was used for bacterial isolation, and leftover samples were stored at 2808 C for subsequent analysis. Coded clinical data were also obtained, and included participant age, CFTR genotype, antibiotic treatments, and FEV1% (for further details, see Table E1 in the online supplement) (41). Patients were grouped by disease states according to published guidelines (42). Rates of disease progression (DFEV1%/Dt) for each patient were calculated as outlined in Figure E1.

739

Bacterial Isolation Homogenized sputum was serially diluted 10-fold in PBS and plated on Pseudomonas Isolation Agar (BD Difco, Sparks, MD). Colonyforming units (CFUs) were counted after incubation at 378 C for 5 days, and total P. aeruginosa loads were recorded as CFU/g sputum (Table E1). Up to 20 representative isolates from each patient were selected and confirmed as P. aeruginosa via PCR (43).

Sputum Phenazine Quantification Homogenized sputum was centrifuged (16,000 3 g for 30 min), and supernatants were filtered through SpinX columns (0.22-mm pore size; Corning, Corning, NY; 16,000 3 g for 10 min). Phenazine content was quantified by high-performance liquid chromatography (33). Concentrations were determined based on Pseudomonas-negative samples spiked with known concentrations of PYO and PCA standards.

Quantification of Phenazine Production by P. aeruginosa Isolates Triplicate cultures of each isolate were prepared in deep 96-well plates containing 500 ml of lysogeny broth. Strain PA14 (30) was used as a control. Plates were incubated at 378 C, with shaking at 300 rpm. After 24 hours, fresh plates were inoculated and grown for another 24 hours. Each sample was measured at l ¼ 500 nm to determine cell density and centrifuged (4,000 3 g for 20 min) to remove cells, and supernatants were measured at l ¼ 691 nm to quantify PYO. Concentrations were calculated using the extinction coefficient for PYO (44), and normalized to cell density.

DNA Extraction, Amplification, Sequencing, and Analysis DNA was extracted from sputum and sequenced at Research and Testing Laboratories (Lubbock, TX) (45, 46). Tag-encoded FLX 454 pyrosequencing using Roche/454 Titanium chemistry (454 Life Sciences, Branford, CT) was performed, using 16S primers Gray28F 59-GAGTTTGATCNTGGCTCAG-39 and Gray519R 59-GTNTTACNGCGGCKGCTG-39 (Integrated DNA Technologies, Coralville, IA) (45–48). FASTA-formatted sequences and corresponding quality scores (QCs) were extracted from standard flowgram format (SFF) data files, using GS Amplicon software (454 Life Sciences). Sequences were analyzed with the Quantitative Insights into Microbial Ecology (QIIME; version 1.4.0) pipeline, using default parameters (49). Only sequences > 200 nt in length with QCs greater than 25 that contained no ambiguous bases were included. UCLUST, version 5.1 (http://www.drive5.com/ usearch), was used on the basis of greater than 94% identity to process reads. Sequences were checked for putative chimeras using UCHIME and the greengenes core set (50). BLAST to greengenes was used to assign taxonomy (51). QIIME pipeline a_rarefaction.py was used to generate a diversity metrics and rarefaction curves.

Statistical Analyses Statistical methods may be found in the online supplement.

RESULTS Study Cohort

Figure 1. Phenazine cycling by Pseudomonas aeruginosa. Schematic of bacterial reduction and auto-oxidation of pyocyanin. Bacterial cells mediate a two-electron transfer to the oxidized form of pyocyanin (PYO), generating its reduced form. Reduced pyocyanin is readily oxidized by molecular oxygen (or other terminal oxidants), generating reactive oxygen species. The redox activity and oxygen radical generation by pyocyanin and other phenazines have broad implications for microbial communities in the airways of patients with cystic fibrosis (CF).

To explore the relationship between phenazine content, bacterial community composition, and the progression of lung disease, we analyzed expectorated sputum from a cohort of 47 adult patients (age range, 18–60 yr; mean age, 34.2 yr; median age, 33 yr) with a confirmed diagnosis of CF. Patients were selected so that their FEV1% values spanned a wide range of pulmonary disease severity (range, 16–128; mean, 64.7; median, 57) (Table 1). At the time of sputum collection, most patients were being treated with azithromycin and/or tobramycin. Of patients not receiving azithromycin/tobramycin, all but two were being treated with one or more classes of other antibiotics (Table E2). Antibiotic

740

AMERICAN JOURNAL OF RESPIRATORY CELL AND MOLECULAR BIOLOGY VOL 47 2012 TABLE 1. DISTRIBUTION OF PHENAZINE CONCENTRATIONS BETWEEN STAGES OF DISEASE SEVERITY (FEV1%) Obstruction

FEV1%

n

Normal Mild Moderate Severe

. 90 70–89 40–69 , 40

13 5 19 10

Age in Years (Average) 31.4 34.8 36.5 33.0

6 6 6 6

Average Total PHZ (mM)

9.3 12.2 10.4 10.9

12.9 30.3 42.7 86.8

6 6 6 6

3.8 12.6 5.9 18.0

Average PYO (mM) 7.7 10.5 25.2 46.8

6 6 6 6

2.7 3.0 3.7 8.5

Average PCA (mM) 5.2 19.7 17.5 39.9

6 6 6 6

1.3 11.6 4.1 11.5

Definition of abbreviations: PCA, phenazine-1–carboxylic acid; PHZ, phenazine; PYO, pyocyanin.

treatments were found to exert no significant effect on phenazine concentrations or differences in microbiota composition between patients (data not shown). Elevated Sputum Phenazine Concentrations Are Negatively Correlated with Declines in Pulmonary Function

Although phenazines were previously detected in expectorated sputum (40), there is little understanding of how their presence relates to disease progression. Accordingly, our primary goal was to determine the relationships between sputum PYO and PCA concentrations and clinical states (as determined by FEV1%). Using HPLC to quantify phenazines, we found a significant negative correlation between concentrations of both PYO (Figure 2A; r ¼ 20.68) and PCA (Figure 2B; r ¼ 20.57) and the severity of lung function impairment. Pairwise comparisons between patient groups revealed no significant differences in total phenazine concentrations between patients with normal FEV1% values (. 90) and those with mild lung obstruction (FEV1%, 70–89) (Table 1). However, patients with moderate (FEV1%, 40–69) and severe (FEV1%, , 40) lung obstruction showed a significant progressive increase in total phenazine concentrations relative to those with normal lung function. We observed similar differences in phenazine concentrations between stages of disease severity for both PYO and PCA individually, although relative amounts were found to vary between patients. In fact, as FEV1% values declined, a trend toward high concentrations of either PCA or PYO, but rarely both (Figure 2C), was evident. Next, we sought to investigate associations between rates of disease progression and phenazine concentrations. To assess this relationship, archived FEV1% results for each patient were used to calculate the rate of change in pulmonary function during a 2-year period, spanning 1 year before expectoration to 1 year afterward. Significant negative correlations were found between

total phenazine concentrations (r ¼ 20.41), PYO (r ¼ 20.41), and PCA (r ¼ 20.36) and rates of lung function decline (Figure 3). In general, patients with stable FEV1% baselines or minor improvements in pulmonary function (DFEV1%/Dt . 0) showed little total phenazine in their sputum. In contrast, those exhibiting rapid rates of disease progression often showed elevated phenazine concentrations, which could be attributed to increases in both PYO (Figure 3B) and PCA (Figure 3C). These results collectively demonstrate a correlation between sputum phenazine concentrations and rates of pulmonary decline in adult patients with CF chronically infected with P. aeruginosa. Pseudomonas Isolates, Depending on Stage of Infection, Produce Variable Amounts of Pyocyanin

Despite extensive laboratory evidence for both the beneficial and detrimental roles of phenazines in bacterial physiology, little understanding exists regarding how phenazine concentrations relate to microbial populations in vivo. In fact, P. aeruginosa isolates taken years apart from the same patient show an attenuation of phenazine production over time (52, 53), leading to speculation that phenazines are not necessary for chronic P. aeruginosa colonization of the airways (52). In this study, the detection of PYO and PCA within sputum at all stages of infection (and greater abundance in patients with severe pulmonary obstruction) was inconsistent with this hypothesis. Therefore, our next goal involved determining whether the recorded PYO and PCA concentrations within sputum reflected phenazine production phenotypes of P. aeruginosa isolates from the same sample. To achieve this, we characterized numerous isolates (i.e., 14–20) from each patient with respect to phenazine production capacity in laboratory culture. We focused on one phenazine, PYO, which is easily quantifiable using a high-throughput assay, and which has been the subject of recent phenotype studies of clinical P. aeruginosa isolates during periods of acute exacerbation (54).

Figure 2. Phenazine concentrations are correlated with clinical stages of disease progression. (A) Pyocyanin (r ¼ 20.68; P , 0.001) and (B) phenazine-1–carboxylic acid (r ¼ 20.57; P , 0.001) concentrations in expectorated sputum significantly increase as pulmonary function (FEV1%) declines. (C) Relative concentrations of sputum PYO and phenazine-1–carboxylic acid (PCA), as detected within the same sputum sample. Data points represent a single sputum sample from an individual patient.


741

Figure 3. Phenazine concentrations are negatively correlated with the rate of pulmonary function decline over a 2-year period. Each data point represents a single patient. In general, patients showing improvement in lung function (ΔFEV1%/Δt . 0) had little phenazine content in their sputum. This correlation was statistically significant for (A) total phenazine (r ¼ 20.41; P ¼ 0.004), (B) PYO (r ¼ 20.41; P ¼ 0.004), and (C) PCA (r ¼ 20.36; P ¼ 0.014).

The vast majority of sputum isolates produced at least small amounts of PYO relative to the wild-type strain PA14 (data not shown). Surprisingly, isolates obtained from healthy patients (FEV1 . 90%) showed significantly higher PYO production compared with those obtained from patients with mild to severe conditions (Figure 4A). In fact, patients with severe lung obstruction (whose sputum phenazine concentrations were highest; see Figure 2) harbored isolates found to produce the least amount of PYO when grown in culture. In addition to differences between patients, within-patient diversity in PYO production was also apparent. Most evident was the heterogeneity shown by P. aeruginosa isolates within sputum from healthy individuals (e.g., FEV1% . 90), compared with isolates obtained from patients at later stages of infection (note the standard deviations in Figure 4A). These results are consistent with previous reports of sequential isolates losing the capacity to produce phenazines (and other virulence factors) over time as infections progress (52, 53, 55). However, a paradoxical relationship is evident between these data and the elevated phenazine concentrations found in sputum from patients with low FEV1% values (compare Figures 2 and 4). If sputum phenazine concentrations are highest in patients with severe lung obstruction, why do these

patients harbor isolates that produce very little phenazine when grown in culture? We hypothesized that low PYO production per cell may be compensated by high cell densities, such that the collective P. aeruginosa population would be sufficient to generate the PYO concentrations detected within sputum. In support of this hypothesis, total sputum PYO determined by HPLC showed a strong statistical dependence (r ¼ 0.43) on the total P. aeruginosa cell load (measured as CFUs) within each sputum sample (Figure 4B). In addition, if the average PYO production of each group of isolates (Figure 4A) is normalized to the total P. aeruginosa load, the collective production by a dense population, despite low levels of PYO production per cell during late stages of infection, could explain elevated phenazine concentrations at later stages of disease (Figure 4C). This has not been explicitly shown here. However, our data clearly demonstrate that laboratory phenotypes of clinical isolates are not necessarily reflective of the environment from which they derive. Rather, the chemistry of their milieu (e.g., phenazine concentrations) and in vivo bacterial population densities must also be taken into account when characterizing microbial infections of the airways.

Figure 4. Phenazine production by clinical isolates of P. aeruginosa. (A) Average in vitro PYO production by isolates from patients at various stages of disease (FEV1%). Each point represents the mean PYO production for 14–20 isolates from a given patient. Values were determined based on the absorbance relative to PA14. Each isolate was measured in triplicate, and error bars represent one standard deviation for each group of isolates. (B) Sputum PYO concentrations show a strong statistical dependence on total sputum P. aeruginosa bacterial load (as determined by colony-forming units [CFU]). (C) Average in vitro PYO production by each group of isolates is normalized to the density of the sputum bacterial population (normalized PYO ¼ mM PYO 3 CFU/g sputum).

742

AMERICAN JOURNAL OF RESPIRATORY CELL AND MOLECULAR BIOLOGY VOL 47 2012

Elevated Phenazine Concentrations Are Correlated with the Overall CF Microbial Population

Based on the range of physiological effects of phenazines on their producers and microbial competitors (26–38), we speculated that their presence in sputum and their increase over time may correlate with the overall microbial community complexity of the CF lung. To assess this, we analyzed the same sputum samples by 454 multiplex pyrosequencing of the 16S rRNA gene. The final dataset of 97,545 sequences included an average of 2,710 6 171 reads/sample. At this depth, a number of organisms present at low abundance would not be detected, and thus the exact number of detected phylotypes is best viewed as an estimate of microbial community composition. Our analysis of relative abundances revealed diverse phylotypes mapping to 261 genera (94% similarity; Figure E2) and 509 species (97% similarity). As expected, P. aeruginosa was the dominant bacterium (in terms of percentage of total bacterial population) in the majority of patients in our cohort. Multiple patients (86% and 36%, respectively) also harbored Streptococcus and Staphylococcus at high abundance, although these phylotypes were significantly more prominent (as determined by the rank abundance method) in patients with low phenazine concentrations (Figure E2). Other dominant taxa in patients with low sputum phenazines included Fusobacteria, Carnobacterium spp., Veillonella, Micrococcus, Gemella, and Spiroplasma. These taxa were all found at lower abundance in the presence of elevated PYO and PCA. The presence of these taxa is consistent with previous reports (17, 20, 24, 25). However, this is the first dataset that links their presence to an environmental variable. Conversely, a striking degree of interindividual phylotype variation was evident in patients with high phenazine content. Despite this population divergence, a distinct relationship in the entire patient cohort was evident between number of phylotypes detected and phenazine concentrations (r ¼ 20.41, 20.51, and 20.49, for PYO, PCA, and total phenazine, respectively) (Figures 5A–5C). Specifically, after phenazines accumulated in the airways, a concomitant significant decrease occurred in the number of unique phylotypes at the genus level. These data are supported by rarefaction analyses that showed increased richness in samples with the lowest phenazine concentrations (Figure E3). Although no distinct community signature was evident that corresponded to elevated phenazine concentrations, these data suggest that elevated phenazine concentrations are correlated with a nonspecific reduction of microbial community complexity. As expected, a significant correlation (r ¼ 0.36) was also evident between FEV1% and phylotype richness (Figure 5D), although this relationship was not as strong as the phenazine–phylotype relationship. Similarly, the total P. aeruginosa load was slightly correlated with population complexity (r ¼ 20.15), although not significantly (Figure 5E), suggesting that community restructuring is not based on the stage of disease or the presence of P. aeruginosa alone.

DISCUSSION Among patients with CF chronically colonized with P. aeruginosa, striking variations exist in the rates of disease progression. These variations have drawn attention to the multifactorial nature of the disease, leading to a growing awareness that the etiology of CF airway infections is polymicrobial. However, little attention has been given to environmental parameters that help define “who’s there” in the lower airways. Here, we demonstrate that two virulence factors produced by P. aeruginosa, PYO and PCA, are examples of environmental variables that can be linked to CF lung function and the complexity of lung microbiota. These results underscore the importance of identifying the environmental

parameters that may affect the ecology of microbial communities in the context of infection. Based on our observations of sputum coloration shifting over time for a given patient (oxidized PYO is blue, whereas reduced PCA is yellow) and the temporal loss of phenazine biosynthetic genes in CF P. aeruginosa isolates (52), we expected phenazine production to peak during the middle stage of the disease (e.g., FEV1% between 50 and 90). However, as pulmonary disease advanced, we were surprised to discover a progressive increase in total sputum concentrations of phenazine, and a progressive increase in both PYO and PCA individually. The relative abundances of these two compounds, however, varied across samples. The accumulation of either compound may be controlled by the differential production or degradation of PCA and PYO as a function of the sputum microenvironment. For example, PYO biosynthesis requires the presence of oxygen (56), and therefore an increased PCA/PYO ratio may reflect lower oxygen tensions within discrete microenvironments of the airways. However, the production of phenazines is likely to be regulated on multiple environmental and molecular levels, many of which have not yet been identified. Variation in phenazine production was also evident in the phenotypes of individual isolates. However, contrary to expectations based on sputum phenazine content, we observed that isolates from patients at late stages of infection produced significantly less PYO in laboratory cultures compared with isolates from earlier stages. Although this overall trend can be explained by observing that at later stages of infection, phenazineproducers are in high abundance, how any individual cell regulates the amount of phenazine it accumulates remains less clear. Factors that repress or trigger phenazine accumulation in the lung may not be captured in laboratory experiments. Alternately, the dynamic evolution of genome content in the lung may explain the attenuation in phenazine production among subpopulations (52, 54, 57–60). For example, phenotypic changes may reflect a decoupling of phenazine production from quorum sensing (QS) at later stages of infection by “hypermutator” populations (54, 57, 58) or mutations of lasR, a common site of mutation in P. aeruginosa clinical isolates (59). Such mutations are known to lead to the rise of “social cheaters” that take advantage of QS-controlled extracellular compounds produced by others in their environment (60). In this context, phenazines could represent a “public good” that supports nonproducing subpopulations. Phenazines have great potential to shape microbial communities, and this is reflected in the variations of microbial community composition we observed. Traditionally, phenazines have been regarded as antibiotics, owing to their redox properties and their concomitant generation of oxygen radicals (61). For example, PYO has been shown to serve as a respiratory inhibitor in S. aureus, conferring a growth advantage to P. aeruginosa when grown in coculture (62). Indeed, we observed a general enrichment of S. aureus in patients with low phenazine content. In addition, phenazines are toxic to Candida albicans and Aspergillus fumigatus, two CF lung fungi (63, 64). Within some microenvironments of the CF lung, however, oxygen may be limiting (65). Under these conditions, phenazines would be expected to be less toxic and manifest their effects in different ways. For example, phenazines, by reducing Fe(III) to Fe(II), can liberate bioavailable iron from host proteins (66). In the laboratory, this was shown to stimulate biofilm formation by P. aeruginosa (33), and other organisms in the same environment may also benefit from this process. Similarly, bacteria that coevolve with phenazine producers may be able to capitalize on other phenazine-mediated processes such as increased ferrous iron availability and oxygen consumption. Because different


743

Figure 5. Bacterial phylotype richness is related to concentrations of phenazine in CF sputum. Less complex bacterial communities correspond to lower concentrations of (A) PYO, (B) PCA, and (C) total phenazines found within sputum samples, in addition to (D) FEV1%. (E) Community complexity was also negatively correlated with total P. aeruginosa load within a given sputum sample, although not significantly (Spearman coefficient ranks; PYO, r ¼ 20.41, P ¼ 0.01; PCA, r ¼ 20.51, P ¼ 0.001; total phenazine, r ¼ 20.49, P ¼ 0.002; FEV1%, r ¼ 20.39, P ¼ 0.02; P. aeruginosa, r ¼ 20.15, P ¼ 0.38).

phenazines exhibit distinct chemical reactivities (for example, PCA is most reactive with Fe(III), whereas PYO is most reactive with oxygen relative to other phenazines [67]), different phenazines could help shape the overall airway microbiome in complementary ways. However, the data presented in this study only demonstrate a correlation between phenazines and reduced microbial diversity. Whether phenazines directly cause these shifts by the mechanisms we have described requires further testing. As phenazines accumulated in the airways, we identified a concurrent acceleration of CF disease progression. Thus, the abundance of PYO and PCA can serve as an indicator for the trajectory of a patient’s clinical status. Because respiratory failure is the primary cause of mortality in patients with CF, a substantial need exists for markers that enable forecasts of disease progression. At present, FEV1% is considered the gold standard, although every test has its limitations. For example, a single FEV1% measurement does not predict future disease progression, nor can pulmonary function tests be performed on patients experiencing hemoptysis or pneumothorax. Sputum nitrates and other inflammatory indices have shown a correlation with lung function (68–70), but few objective measures gauge therapeutic efficacy or serve as predictive indicators (71). The strong relationship between PYO, PCA, and lung function decline demonstrated here warrants the further consideration of phenazines serving as biomarkers in adult patients chronically infected with P. aeruginosa (we are currently testing whether this correlation also holds true in a pediatric cohort).

In conclusion, this case study of phenazine abundance in the CF lung reveals that phenotypes of clinical isolates are not reliable biomarkers of disease progression. Rather, we demonstrate that monitoring changes in the chemistry of infected environments has potential as an important diagnostic aid. Phenazines likely represent just one environmental parameter conditioned by P. aeruginosa that is negatively correlated with the complexity of the microbial community of the CF lung. Going forward, it will be important to consider how multiple factors (including oxygen availability, host genetics, diet, antibiotics, and additional bacteria) change over time to reach a predictive understanding of environmental microbial community dynamics and how they relate to CF disease states. Author disclosures are available with the text of this article at www.atsjournals.org. Acknowledgments: The authors thank the individuals who participated in this study, and Erin Leone and Chao-Yu Guo for data collection and experimental design. The authors also thank Maureen Coleman and other members of the Newman Laboratory, California Institute of Technology, for critically reviewing the manuscript. The Committee on Clinical Investigation, CHLA, approved this study (protocol 09-04-0183).

References 1. Gibson RL, Burns JL, Ramsey BW. Pathophysiology and management of pulmonary infections in cystic fibrosis. Am J Respir Crit Care Med 2003;168:918–951. 2. Riordan JR, Rommens JM, Kerem BS, Alon N, Rozmahel R, Grzelczak Z, Zielinski J, Lok S, Plavsic N, Chou JL, et al. Identification of the

744

3. 4. 5. 6.

7. 8.

9. 10.

11.

12.

13. 14. 15.

16.

17.

18. 19.

20.

21.

22.

23.

24.

25.

AMERICAN JOURNAL OF RESPIRATORY CELL AND MOLECULAR BIOLOGY VOL 47 2012

cystic fibrosis gene: cloning and characterization of complementary DNA. Science 1989;245:1066–1073. Ratjen F. Cystic fibrosis: pathogenesis and future treatment strategies. Respir Care 2009;54:595–605. Rajan S, Saiman L. Pulmonary infections in patients with cystic fibrosis. Semin Respir Infect 2002;17:47–56. Machen TE. Innate immune response in CF airway epithelia: hyperinflammatory? Am J Physiol Cell Physiol 2006;291:C218–C230. Pezzulo AA, Tang XX, Hoegger MJ, Abou Alaiwa MH, Ramachandran S, Moninger TO, Karp PH, Wohlford-Lenane CL, Haagsman HP, van Eijk M, et al. Reduced airway surface pH impairs bacterial killing in the porcine cystic fibrosis lung. Nature 2012;487:109–113. Cohen TS, Prince A. Cystic fibrosis: a mucosal immunodeficiency syndrome. Nat Med 2012;18:509–519. Schroeder TH, Lee MM, Yacono PW, Cannon CL, Gerceker AA, Golan DE, Pier GB. CFTR is a pattern recognition molecule that fibrosis associated pathogen B. cenocepecia, reduced wound healing in vitro. Proc Natl Acad Sci USA 2002;99:6907–6912. Downey DG, Bell SC, Elborn JS. Neutrophils in cystic fibrosis. Thorax 2009;64:81–88. McCormick J, Mehta G, Olesen HV, Viviani L, Macek M Jr, Mehta A. Comparative demographics of the European cystic fibrosis population: a cross-sectional database analysis. Lancet 2010;375:1007–1013. Schaedel C, de Monestrol I, Hjelte L, Johannesson M, Kornfalt R, Lindblad A, Strandvik B, Wahlgren L, Holmberg L. Predictors of deterioration of lung function in cystic fibrosis. Pediatr Pulmonol 2002;33:483–491. Klepac-Ceraj V, Lemon KP, Martin TR, Allgaier M, Kembel SW, Knapp AA, Lory S, Brodie EL, Lynch SV, Bohannan BJM, et al. Relationship between cystic fibrosis respiratory tract bacterial communities and age, genotype, antibiotics, and Pseudomonas aeruginosa. Environ Microbiol 2010;12:1293–1303. Brogden KA, Guthmiller JM, Taylor CE. Human polymicrobial infections. Lancet 2005;365:253–255. Ley RE, Turnbaugh PJ, Klein S, Gordon JI. Microbial ecology: human gut microbes associated with obesity. Nature 2006;444:1022–1023. Dethlefsen L, Mcfall-Ngai M, Relman DA. An ecological and evolutionary perspective on human–microbe mutualism and disease. Nature 2007; 449:811–818. Bittar F, Richet H, Dubus JC, Reynaud-Gaubert M, Stremler N, Sarles J, Raoult D, Rolain JM, Fox D. Molecular detection of multiple emerging pathogens in sputa from cystic fibrosis patients. PLoS ONE 2008;3:e2908. Guss AM, Roeselers G, Newton ILG, Young CR, Klepac-Ceraj V, Lory S, Cavanaugh CM. Phylogenetic and metabolic diversity of bacteria associated with cystic fibrosis. ISME J 2010;5:20–29. LiPuma JJ. The changing microbial epidemiology of cystic fibrosis. Clin Microbiol Rev 2010;23:299–323. Rogers GB, Hoffman LR, Whiteley M, Daniels TWV, Carroll MP, Bruce KD. Revealing the dynamics of polymicrobial infections: implications for antibiotic therapy. Trends Microbiol 2010;18:357–364. Sibley CD, Grinwis ME, Field TR, Eshaghurshan MM, Faria MM, Dowd SE, Parkins MD, Rabin HR, Surette MG. Culture enriched molecular profiling of the cystic fibrosis airway microbiome. PLoS ONE 2011;6:e22702. Willner D, Haynes MR, Furlan M, Schmieder R, Lim YW, Rainey PB, Rohwer F, Conrad D. Spatial distribution of microbial communities in the cystic fibrosis lung. ISME J 2011;6:471–474. Cox MJ, Allgaier M, Taylor B, Baek MS, Huang YJ, Daly RA, Karaoz U, Anderson GL, Brown R, Fujimura KE, et al. Airway microbiota and pathogen abundance in age-stratified cystic fibrosis patients. PLoS One 2010;B23:5:e11044 van der Gast CJ, Walker AW, Stressmann FA, Rogers GB, Scott P, Daniels TW, Carroll MP, Parkhill J, Bruce KD. Partitioning core and satellite taxa from within cystic fibrosis lung bacterial communities. ISME J 2010;5:780–791. Rogers GB, Daniels TWV, Tuck A, Carroll MP, Connett GJ. Studying bacteria in respiratory specimens by using conventional and molecular microbiological approaches. BMC Pulm 2009;9:14–24. Tunney MM, Klem ER, Fodor AA, Gilpin DF, Moriarty TF, Mcgrath SJ, Muhlebach MS, Boucher RC, Cardwell C, Doering G, et al. Use of culture and molecular analysis to determine the effect of antibiotic treatment on microbial community diversity and abundance during exacerbation in patients with cystic fibrosis. Thorax 2012;66:579–584.

26. Baron SS, Terranova G, Rowe JJ. Molecular mechanism of the antimicrobial action of pyocyanin. Curr Microbiol 1989;18:223–230. 27. Denning G, Railsback M, Rasmussen G, Cox C, Britigan B. Pseudomonas pyocyanin alters calcium signaling in human airway epithelial cells. Am J Physiol Lung Cell Mol Physiol 1998;274:L893–L900. 28. Kerr JR, Taylor GW, Rutman A, Hoiby N, Cole PJ, Hoiby N. Pseudomonas aeruginosa pyocyanin and 1-hydroxyphenazine inhibit fungal growth. J Clin Pathol 1999;52:385–387. 29. Allen L, Dockrell DH, Pattery T, Lee DG, Cornelis P, Hellewell PG, Whyte MKB. Pyocyanin production by Pseudomonas aeruginosa induces neutrophil apoptosis and impairs neutrophil-mediated host defenses in vivo. J Immunol 2005;174:3643–3649. 30. Rahme LG, Stevens EJ, Wolfort SF, Shao J, Tompkins RG, Ausubel FM. Common virulence factors for bacterial pathogenicity in plants and animals. Science 1995;268:1899–1902. 31. Mahajan-Miklos S, Tan MW, Rahme LG, Ausubel FM. Molecular mechanisms of bacterial virulence elucidated using a Pseudomonas aeruginosa– Caenorhabditis elegans pathogenesis model. Cell 1999;96:47–56. 32. Lau GW, Ran H, Kong F, Hassett DJ, Mavrodi D. Pseudomonas aeruginosa pyocyanin is critical for lung infection in mice. Infect Immun 2004;72:4275–4278. 33. Wang Y, Wilks JC, Danhorn T, Ramos I, Croal L, Newman DK. Phenazine1–carboxylic acid promotes bacterial biofilm development via ferrous iron acquisition. J Bacteriol 2011;93:3606–3617. 34. Price-Whelan AM, Dietrich LEP, Newman DK. Pyocyanin alters redox homeostasis and carbon flux through central metabolic pathways in Pseudomonas aeruginosa PA14. J Bacteriol 2007;189:6372–6381. 35. Dietrich LEP, Price-Whelan A, Peterson A, Whiteley M, Newman DK. The phenazine pyocyanin is a terminal signaling factor in the quorum-sensing network of Pseudomonas aeruginosa. Mol Microbiol 2006;61:1308–1321. 36. Dietrich LEP, Price-Whelan A, Teal T, Newman DK. Redox-active antibiotics control gene expression and community behavior in divergent bacteria. Science 2008;321:1203–1206. 37. Ramos I, Dietrich LEP, Price-Whelan A, Newman DK. Phenazines affect biofilm formation by Pseudomonas aeruginosa in similar ways at various scales. Res Microbiol 2010;161:187–191. 38. Wang Y, Kern SE, Newman DK. Endogenous phenazine antibiotics promote anaerobic survival of Pseudomonas aeruginosa via extracellular electron transfer. J Bacteriol 2010;192:365–369. 39. Friedheim EAH. Pyocyanin, an accessory respiratory pigment. J Exp Med 1931;54:207–221. 40. Wilson R, Sykes DA, Watson D, Rutman A, Taylor GW, Cole PJ. Measurement of Pseudomonas aeruginosa phenazine pigments in sputum and assessment of their contribution to sputum sol toxicity for respiratory epithelium. Infect Immun 1988;56:2515–2517. 41. Knudson RJ, Slatin RC, Lebowitz MD, Burrows B. The maximal expiratory flow–volume curve: normal standards, variability, and effects of age. Am Rev Respir Dis 1976;113:587–600. 42. Flume PA, O’Sullivan BP, Robinson KA, Goss CH, Mogayzel PJ, Willey-Courand DB, Bujan J, Finder J, Lester M, Quittell L, et al. Cystic fibrosis pulmonary guidelines: chronic medications for maintenance of lung health. Am J Respir Crit Care Med 2007;180:802–808. 43. Spilker T, Coenye T, Vandamme P, Lipuma JJ. PCR-based assay for differentiation of Pseudomonas aeruginosa from other Pseudomonas species recovered from cystic fibrosis patients. J Clin Microbiol 2004; 42:2074–2079. 44. O’Malley YQ, Reszka KJ, Spitz DR, Denning GM, Britigan BE. Pseudomonas aeruginosa pyocyanin directly oxidizes glutathione and decreases its levels in airway epithelial cells. Am J Physiol Lung Cell Mol Physiol 2004;287:L94–L103. 45. Olafson PU, Lohmeyer KH, Dowd SE. Analysis of expressed sequence tags from a significant livestock pest, the stable fly (Stomoxys calcitrans), identifies transcripts with a putative role in chemosensation and sex determination. Arch Insect Biochem Physiol 2010;74:179–204. 46. Smith DM, Snow DE, Rees E, Zischkau AM, Hanson JD, Wolcott RD, Sun Y, White J, Kumar S, Dowd SE. Evaluation of the bacterial diversity of pressure ulcers using bTEFAP pyrosequencing. BMC Med Genomics 2010;3:41. 47. Dowd SE, Callaway TR, Wolcott RD, Sun Y, Mckeehan T, Hagevoort RG, Edrington TS. Evaluation of the bacterial diversity in the feces of cattle using 16S rDNA bacterial tag-encoded FLX amplicon pyrosequencing (bTEFAP). BMC Microbiol 2008;8:125.

Hunter, Klepac-Ceraj, Lorenzi, et al.: Phenazine Content and CF Lung Decline 48. Andreotti R, Perez De Leon AA, Dowd SE, Guerrero FD, Bendele KG, Scoles GA. Assessment of bacterial diversity in the cattle tick Rhipicephalus (Boophilus) microplus through tag-encoded pyrosequencing. BMC Microbiol 2011;11:6. 49. Caporaso JG, Kuczynski J, Stombaugh J, Bittinger K, Bushman FD, Costello EK, Fierer N, Gonzalez Pena A, Goodrich JK, Gordon JI, et al. QIIME allows analysis of high-throughput community sequencing data. Nat Methods 2010;7:335–336. 50. Edgar RC, Haas BJ, Clemente JC, Quince C, Knight R. UCHIME improves sensitivity and speed of chimera detection. Bioinformatics 2011;27:2194–2200. 51. Desantis TZ, Hugenholtz P, Larsen N, Rojas M, Brodie EL, Keller K, Huber T, Dalevi D, Hu P, Andersen GL. Greengenes: a chimerachecked 16S rRNA gene database and workbench compatible with ARB. Appl Environ Microbiol 2006;72:5069–5072. 52. Smith EE, Buckley DG, Wu Z, Saenphimmachak C, Hoffman LR. Genetic adaptation by Pseudomonas aeruginosa to the airways of cystic fibrosis patients. Proc Natl Acad Sci USA 2006;103:8487–8492. 53. Rau MH, Hansen SK, Johansen HK, Thomsen LE, Workman CT, Nielsen KF, Jelsbak L, Hoiby N, Yang L, Molin S. Early adaptive developments of Pseudomonas aeruginosa after the transition from life in the environment to persistent colonization in the airways of human cystic fibrosis hosts. Environ Microbiol 2010;12:1643–1658. 54. Mowat E, Paterson S, Fothergill JL, Wright EA, Ledson MJ, Walshaw MJ, Brockhurst MA, Winstanley C. Pseudomonas aeruginosa population diversity and turnover in cystic fibrosis chronic infections. Am J Respir Crit Care Med 2011;183:1674–1679. 55. Döring G, Parameswaran IG, Murphy TF. Differential adaptation of microbial pathogens to airways of patients with cystic fibrosis and chronic obstructive pulmonary disease. FEMS Microbiol Rev 2010; 35:124–146. 56. Parsons JF, Greenhagen BT, Shi K, Calabrese K, Robinson H, Ladner JE. Structural and functional analysis of the pyocyanin biosynthetic protein PhzM from Pseudomonas aeruginosa. Biochem 2007;46:1821–1828. 57. Driffield K, Miller K, Bostock JM, O’Neill AJ, Chopra I. Increased mutability of Pseudomonas aeruginosa in biofilms. J Antimicrob Chemother 2008;61:1053–1056. 58. Fothergill JL, Panagea S, Hart CA, Walshaw MJ, Pitt TL, Winstanley C. Widespread pyocyanin overproduction among isolates of a cystic fibrosis epidemic strain. BMC Microbiol 2007;7:45.

745

59. Hoffman LR, Kulasekera HD, Emerson J, Houston LS, Burns JL, Ramsey BW, Miller SI. Pseudomonas aeruginosa lasR mutants are associated with cystic fibrosis lung disease progression. J Cyst Fibros 2009;8:66–70. 60. Sandoz KM, Mitzimberg SM, Schuster M. Social cheating in Pseudomonas aeruginosa quorum sensing. Proc Natl Acad Sci USA 2007;104:15876–15881. 61. Mavrodi DV, Blankenfeldt W, Thomashow LS. Phenazine compounds in fluorescent Pseudomonas spp.: biosynthesis and regulation. Annu Rev Phytopathol 2006;44:417–445. 62. Voggu L, Schlag S, Biswas R, Rosenstein R, Rausch C, Gotz F. Microevolution of cytochrome BD oxidase in Staphylococci and its implication in resistance to respiratory toxins released by Pseudomonas. J Bacteriol 2006;188:8079–8086. 63. Gibson J, Sood A, Hogan DA. Pseudomonas aeruginosa–Candida albicans interactions: localization and fungal toxicity of a phenazine derivative. Appl Environ Microbiol 2009;75:504–513. 64. Morales DK, Jacobs NJ, Rajamani S, Krishnamurthy M, Cubillos-Ruiz JR, Hogan DA. Antifungal mechanisms by which a novel Pseudomonas aeruginosa phenazine toxin kills Candida albicans in biofilms. Mol Microbiol 2010;78:1379–1392. 65. Worlitzsch D, Tarran R, Ulrich M, Schwab U, Cekici A. Effects of reduced mucus oxygen concentration in airway Pseudomonas infections of cystic fibrosis. J Clin Invest 2002;109:317–325. 66. Cox CD. Role of pyocyanin in the acquisition of iron from transferrin. Infect Immun 1986;52:263–270. 67. Wang Y, Newman DK. Redox reaction of phenazine antibiotics with ferric (hydr)oxides and molecular oxygen. Environ Sci Technol 2009; 42:2380–2386. 68. Sagel D, Chmiel JF, Konstan MW. Sputum biomarkers of inflammation in cystic fibrosis lung disease. Proc Am Thorac Soc 2007;4:406–417. 69. Mayer-Hamblett N, Aitken ML, Accurso FJ, Kronmal RA, Konstan MW, Burns JL, Sagel SD, Ramsey BW. Association between pulmonary function and sputum biomarkers in cystic fibrosis. Am J Respir Crit Care Med 2007;175:822–828. 70. Anil N, Singh M, Rajwanshi A, Vohra H. Induced sputum nitrites correlates with FEV1 in children with cystic fibrosis. Acta Paediatr 2010; 99:711–714. 71. Ramsey BW, Boat TF. Outcome measures for clinical trials in cystic fibrosis: summary of a cystic fibrosis consensus conference. J Pediatr 1994;124:177–192.