Pf bacteriophage is associated with decline in lung function in a longitudinal cohort of patients with cystic fibrosis and Pseudomonas airway infection

Abstract

Background:

The Pseudomonas filamentous bacteriophage (Pf) infects Pseudomonas aeruginosa (Pa) and is abundant in the airways of many people with cystic fibrosis (CF) (pwCF). We previously demonstrated that Pf promotes biofilm growth, as well as generates liquid crystals that confer biofilms with adhesivity, viscosity and resistance to clearance. Consistent with these findings, the presence of Pf in sputum from pwCF has been linked to chronic Pa infection and more severe exacerbations in a cross-sectional cohort study.

Methods:

We examined the relationships between Pf and clinical outcomes in a longitudinal study of pwCF. Sputum Pa and Pf concentrations were measured by qPCR, as well cytokines and active neutrophil elastase by standardized assays. Recorded clinical data, including spirometry and microbiological results, were analyzed for associations with Pf. Finally, lung explants from pwCF in this cohort who underwent lung transplantation were examined for presence of liquid crystals within secretions.

Results:

In explanted lungs from pwCF with known Pf infection we demonstrate areas of birefringence consistent with liquid crystalline structures within the airways. We find that high concentration of Pf in sputum is associated with accelerated loss of lung function, suggesting a potential role for Pf in the pathogenesis of CF lung disease. We also find Pf to associate with increased airway inflammation and an anti-viral cytokine response.

Conclusion:

Pf may serve as a prognostic biomarker and potential therapeutic target for Pa infections in CF.

1. Introduction

Cystic fibrosis (CF) is a genetic disease characterized by progressive lung function decline, often leading to respiratory failure and premature death [1]. Chronic bacterial infection, primarily by Pseudomonas aeruginosa (Pa), exacerbates this decline, contributing to increased mortality [2]. Efforts to eradicate Pa infection are made at first acquisition [3]. However, these efforts are hampered by its ability to form resilient biofilms making clearance challenging, particularly given constant environmental exposure [4]. Consequently, a significant proportion of people with CF (pwCF) develop chronic Pa infection by adulthood, which is further complicated by antibiotic resistance, posing significant treatment challenges [5].

In recent years, CF care has been revolutionized by highly effective cystic fibrosis transmembrane receptor (CFTR) modulators (HEMT), small molecules that target the faulty protein and restore the conductance of chloride and bicarbonate ions. These treatments have notably improved lung function and quality of life [6]. However, airway inflammation and persistent infections, particularly with Pa, remain challenging even after initiating these therapies [7,8].

The Pseudomonas filamentous (Pf) bacteriophage (phage) is a virus that infects Pa and integrates into the bacterial genome as a prophage [9]. When growing in a biofilm state, the bacteria release numerous virions [10]. The protein coats of these long virions are highly negatively charged and capable of organizing polymers in the CF airway (DNA, mucins, actin, and alginate) into liquid crystalline structures [11]. This liquid crystal enhances adhesion and viscosity of the biofilm [12].

We have previously shown that Pf is detected in the sputum of many pwCF with Pa infection, and that it correlates with poor outcomes in a cross-sectional study [13]. Pf also triggers an immune response in both in vitro and in vivo models of chronic infection, shifting the host response towards an anti-viral mode aiding in the persistence of chronic bacterial infections [14–16].

We hypothesized that Pf presence provides a selective advantage to Pa in the CF airway, leading to increased bacterial fitness and worse outcomes over time. In this longitudinal study, we confirm the persistence of Pf throughout chronic Pa infection and its association with heightened airway inflammation. Elevated Pf levels in sputum correspond to a faster decline in lung function. Our findings suggest that Pf may contribute to chronic Pa infection in CF and could be a promising therapeutic target to combat or prevent the pathogenicity of Pa infections.

2. Methods

2.1. Clinical study design & cohort

We have conducted a prospective longitudinal cohort study in 121 pwCF at the Stanford CF Center from January 2016 to December 2021. Patients were enrolled in an unbiased manner, in both pediatric and adult CF clinics at Stanford. This study was approved by the Stanford University IRB (# 37232). Prior to inclusion in the study informed consent was obtained from each subject for sputum collection and biobanking.

Inclusion criteria included a diagnosis of CF, ability to perform pulmonary function testing (PFT), and ability to produce a sputum sample by spontaneous expectoration. Patients with prior lung transplantation were excluded, and study involvement was ceased if transplantation took place after enrollment. Sputum samples were obtained at routine clinical visits, acute visits and hospitalizations if the subject was able to expectorate.

11 healthy control subjects underwent sputum induction with nebulized hypertonic saline as per Cystic Fibrosis Foundation Therapeutic Development Network’s standard operating procedure [17] for inclusion in the cytokine profiling comparisons.

2.2. Sputum processing and Pf quantification

Sputum was collected in sterile specimen cups and frozen at −80 °C for later batch analysis. From this specimen, an aliquot of 200–500uL was kept frozen unprocessed. The remainder of the sample was weighed and diluted 1:10 with phosphate buffered solution, spun to remove cell pellet, filtered to remove cells and the supernatant fluid, which we call “sputum fluid,” was aliquoted and saved for cytokine analysis and other measurements. Protease inhibitors were added to half of the aliquots.

We used our previously published methods for processing, DNA extraction and quantitative polymerase chain reaction (qPCR) for identification of Pf [18]. Briefly we utilize mechanical homogenization followed by DNA extraction using QIAmp DNA Mini kit [11,14,18].

To identify Pa, primers targeting the rpIU gene were utilized. In addition to Pf-conserved primers (PA0717), we also utilized a second set of primer that targets a sequence of the coat protein gene, CoaB (PA0723), to improve sensitivity in detecting Pf.

A subject was considered Pa+ or Pf+ if any of their samples during the study period had Pa or Pf detected by qPCR, respectively. A subject was categorized as Pa+Pf+^HIGH if the Pf concentration by qPCR was >2 logs than the Pa concentration by qPCR in any of that subjects’ sputum samples, otherwise subjects were categorized as Pa+Pf+^LO. We reasoned we would see clinical associations with higher Pf concentrations and higher Pf production (greater difference in Pf and Pa concentration) given many effects of Pf phage many are concentration dependent [11], specifically the physical changes in biofilm viscosity.

2.3. Elastase quantification

Active human neutrophil elastase (aHNE) was measured in sputum fluid (without protease inhibitors added) diluted 1:200 by tagged immunoassay (ProteaseTag, ProAxsis, Belfast) in samples where enough sputum was collected for both qPCR and sputum fluid banking.

2.4. Sputum cytokine profiling

Aliquots of sputum fluid with protease inhibitors added had G-CSF, GM-CSF, IL-1β, IL-8, IL-10, Leptin, MIP-1α, TNFα, and TNF β quantified by multiplex immunoassay utilizing either a plate-based Luminex assay (eBiosciences, Thermo Fisher Scientific, Waltham, MA) or glass slide-based antibody array for protein detection (Ray Biotech, Peachtree Corners, GA). These are interchangeable and well validated methods that allow for work with small samples with high sensitivity.

2.5. Clinical data

Clinical data from the day of sample collection and from routine encounters after enrollment was recorded. Fields captured included anthropometrics, age, genotype, forced expiratory volume in one second (FEV₁), pancreatic status, cystic fibrosis related diabetes (CFRD) diagnosis, HEMT use, hospitalizations, antibiotic courses, and microbiology culture results. All PFT data after enrollment was captured (not only on days of sputum sample collection) for determination of rate of change of lung function over time. Leeds criteria was used to define chronic Pa infection (>50 % of sputum cultures with Pa in the preceding 12 months) [19].

Study data were collected and managed using REDCap electronic data capture tools hosted at Stanford University funded by Stanford CTSA award number UL1 TR001085 from NIH/NCRR.

2.6. Human lung tissues

For enrolled subjects who underwent lung transplantation, sections of explanted lungs were obtained from the Pathology Department at Stanford University (IRB #11197) if available. Paraffin embedded sections were stained with routine H&E and evaluated by light microscopy at 400X for presence of plugged bronchial airways. An unstained corresponding section was then co-registered for location of the plugged airway and imaged at 400X with a benchtop light microscope fitted with a custom-built rotating polarizer. The images captured were then analyzed with ROTOPOL software to quantify the degree of birefringence present by measuring the optical anisotropy (|sinδ|) as the circularly polarized light wave passed through the specimen [20]. Measures of |sinδ| were taken in 20 locations within the airways of each explant section.

2.7. Statistical analyses

Categorical data in demographics and clinical characteristics were summarized in contingency tables and compared between different Pf status groups by Chi-squared test or Fisher’s exact test as appropriate. Continuous data were summarized by their means and standard deviations, compared between different Pf status groups by ANOVA or Kruskal-Wallis test as appropriate.

Sputum cytokine concentrations were analyzed by generalized estimating equations (GEE) to handle the correlated longitudinal data, account for repeated measures and missing data.

The association between Pf status and the rate of decline in lung function during the study period was assessed by mixed effects model for repeated measurements to take into account repeated measures over time for each individual participant and adjusted for age, sex and HEMT use as potential confounders.

Graphing and statistical analyses were performed using GraphPad Prism 6 (GraphPad Software), R (R computing) and SAS (v.9.4, SAS Institute Cary, NC). Comparisons of mean |sinδ| as a measure of birefringence in lung explant sections were by Kruskal-Wallis test and Dunn’s multiple comparisons.

3. Results

3.1. Study population characteristics

The 121 enrolled subjects were categorized by Pa and Pf status as defined by qPCR results, with 39 subjects negative for Pa (Pa−), 39 subjects positive for Pa (Pa+) but negative for Pf (Pf−), and 43 subjects Pa+ and positive for Pf (Pf+) (Table 1). A subject was categorized as Pa or Pf+ if there was any sample with Pa or Pf detected by qPCR, respectively, at any point during the study. Every Pf+ sample was also Pa+ by qPCR. Of the samples Pa− by qPCR, 6 % (12/191) were Pa+ by sputum culture on the same day. Additionally, 14.7 % (14/95) of Pa− cultures were Pa+ by qPCR, of which all subjects had a history of prior Pa+ cultures.

Table 1.

Demographics and Clinical Characteristics of Cohort Separated by Pseudomonas aeruginosa (Pa) and Pf phage (Pf) Status.

	Pa −		Pa + Pf −		Pa + Pf +		p-value
subjects†	39		39		43
male	26	(67 %)	22	(56 %)	22	(51 %)	0.3561
Age at entry in years (SD)	21.5	(12.9)	27.8	(10.0)	35.1	(12.0)	<0.0001
F508del homozygous	12	(31 %)	13	(33 %)	25	(58 %)	0.0719
F508del heterozygous	18	(46 %)	15	(38 %)	11	(26 %)
Pancreatic insufficiency	36	(92 %)	34	(87 %)	43	(100 %)	0.0623
CFRD	16	(41 %)	15	(38 %)	22	(51 %)	0.1048
impaired glucose tolerance	5	(13 %)	5	(13 %)	11	(26 %)
Leeds Criteria at Entry
Chronic Pa infection	3†	(0.8 %)	28	(72 %)	37	(86 %)	0.0483 ††††
Intermittent Pa infection	12	(31 %)	10	(26 %)	3	(7 %)
No Pa infection	24	(62 %)	1†††	(3 %)	3†††	(7 %)
Leeds Criteria During Study
Becomes chronic	0	(0 %)	2	(5 %)	2	(5 %)	0.0909
Clears Pa infection	5	(13 %)	2	(5 %)	1	(2 %)
Mucoid Pa, % of isolates per subject (SD)			72 %	(39 %)	84 %	(34 %)	0.1788
# of sputum samples contributed per subject (SD)	1.5	(0.9)	2.5	(2.1)	4.1	(2.6)	<0.0001
# samples collected in exacerbation (total)	26	43 %	46	47 %	65	37 %	0.195
Length of follow up in years (SD)	3.1	(1.6)	3.5	(1.8)	3.8	(1.6)	0.1848
Pulmonary Function
FEV1 %-pred at entry (SD)	60.5	(25.7)	58.3	(19.5)	50.9	(22.8)	0.0987
Azithromycin use at entry	19	(49 %)	27	(69 %)	25	(60 %)	0.1827
CFTR Modulator
on HEMT††††† at enrollment	2	(5 %)	5	(13 %)	1	(2 %)	0.2274
started HEMT during study period	18	(46 %)	21	(54 %)	26	(60 %)
no HEMT	19	(49 %)	13	(33 %)	15	(37 %)
Death or Transplant	7	(18 %)	6	(15 %)	9	(21 %)	0.8086

Definition of abbreviations: Pa = P. aeruginosa, Pf = Pf phage, CFRD = cystic fibrosis-related diabetes, FEV₁ = forced expiratory volume in one second, HEMT = highly effective CFTR modulator therapy, CFTR = cystic fibrosis transmembrane receptor.

^†Subject is considered Pf phage or P. aeruginosa positive if either was detected in any of their sputum samples by qPCR. Leeds criteria determined by sputum culture results.

^††These subjects had history of positive sputum cultures in past 12 months but subsequently cleared P. aeruginosa infection,.

^†††These subjects had P. aeruginosa detected by qPCR but not by culture, 2 in Pa+Pf+ group on review had history of positive sputum cultures for Pa >12 months prior, the remaining 2 subjects developed chronic Pa infection during study.

^††††Comparison between Pa+Pf− vs Pa+Pf+ only.

^{†††††}HEMT considered ivacaftor alone for ivacaftor responsive mutation or elexacaftor / tezacaftor / ivacaftor.

3.2. Associations between Pf and clinical characteristics

We evaluated demographics, clinical characteristics, and incidence of complications of CF disease to identify differences in groups of subjects associated with Pa and Pf status. We observed that Pa+Pf+ subjects were older than the Pa+Pf− subjects (p = 0.0197) and were more likely to meet Leeds criteria (Table 1, Fig. 1A-B). Pa+Pf+ subjects had a similar length of follow up to other groups, but contributed more samples than both Pa− (p < 0.0001) and Pa+Pf− (p = 0.0087) during the study window (Table 1, Fig. 1C-D). These data indicate that Pf is associated with chronic stage of infection and may be associated with higher sputum production.

Fig. 1.

Distribution of select clinical characteristics, Pf phage and Pseudomonas Concentrations. (A) Age at entry between groups, analysis by Kruskal Wallis Test with Dunn’s multiple comparisons. (B) Leeds Criteria at entry into study. Analyzed by Fisher’s test, p = 0.0483. (C) Length of follow up (years) for each subject in each patient category. Calculated from first sample provided to last FEV₁ measurement recorded. (D) Lung function as FEV₁ %-predicted at entry. Comparisons in (C) and (D) are not significant by Kruskal Wallis Test with Dunn’s multiple comparisons. (E) Display of Pa and Pf concentrations separated by category of sputum sample. Samples with no Pa and no Pf phage detection are not depicted. Comparison by ANOVA. (F) Comparison of Pa and Pf phage qPCR values in samples with both detected. Dotted line depicts a 1:1 relationship. Mean Pf concentration was 3.88 × 109/mL (1.16 × 1010) and mean Pa concentration was 9.30 × 108/mL (3.14 × 109), corresponding to a 0.62 log higher concentration of Pf. This is evidence of an abundance of Pf phage over the expected values from the presence of Pa alone. Pa = P. aeruginosa. * p > 0.05, ** p < 0.0001.

Conversely, there was no difference in FEV₁ percent predicted at study entry (Table 1, Fig. 1D). In addition, there was no difference between groups in CFRD or glucose intolerance occurrence. There was also no difference in use of azithromycin or HEMT, either at enrollment or starting during the study for any group. No significant difference in either mortality or the rate of lung transplantation between the different groups was noted during the study period.(Table 1)

3.3. Pf outnumber Pa in CF sputum

We then quantified the sputum Pa concentrations by qPCR, which ranged from 10⁴ copies/mL to 10¹⁰ copies/mL. There was no significant difference in Pa concentrations between the Pf+ and Pf− samples (Fig. 1E). Though Pf and Pa concentrations are correlated (r²=0.3276), Pf concentrations were on average 0.62 logs higher than Pa concentrations, indicating active Pf production by Pa in the airway (Fig. 1F). Of the Pa+ subjects, 52 % (43/82) were also Pf+. Of the 43 subjects in the Pf+ group, 62 %(27) had Pf detected by qPCR in all of their sputum samples. This demonstrates Pf is prevalent and persistent in a subset of pwCF.

3.4. CF airways infected with Pa harboring Pf exhibit mucus plugs with crystalline structures

Pf was previously reported under in vitro conditions to promote formation of liquid crystalline structures in biofilms [11,12,21], but whether Pf from natural Pa infection would effectively produce liquid crystal in the CF airway was unknown. To address this, we examined sections of explanted lungs from pwCF in our cohort who underwent lung transplantation. Bronchi with mucus plugs were examined by polarizing microscopy to evaluate for birefringence as an indication of the presence of liquid crystals. We had access to explanted lung tissue from pwCF with the following infection statuses: Pa−, Pa+Pf− and Pa+Pf+. There were two subjects in the Pa+Pf+ category: one pwCF was Pa+Pf+^LOW and the other was Pa+Pf+^HIGH (Fig. 2A–H). Only the mucus plugs in the bronchi of the Pa+Pf+ pwCF contained visible birefringent material (Fig. 2F,H). The quantified degree of birefringence present by optical anisotropy (sin∂) was significantly higher in the Pa+Pf+^HIGH than the Pa+Pf+^LOW explant (mean 0.53 vs 0.16, p < 0.001) (Fig. 2I). This suggests a concentration-dependent effect of Pf in the airway. These data establish that Pf are associated with liquid crystal formation within the CF airways.

Fig. 2.

Polarized microscopy of mucus plugs within airways of explanted CF lungs. Representative section of explanted lungs shown at 400x from a pwCF and no Pa infection (A & B), Pa infected without Pf phage infection (Pa+Pf−) infection (C & D), Pa+ but low concentrations of Pf phage (E & F) and with Pa+ and high concentrations of Pf phage in the sputum (G & H). H&E stain demonstrating mucus plug within the airway (A, C, E, G), and corresponding images of same airway utilizing ROTOPOL software to evaluate birefringence (|sinδ|) (B, D, F, H). (I) Quantitative evaluation of the mucus plugs filling airway sections by polarized microscopy |sinδ| in all 4 subjects. Images were taken in 20 areas of each airway. Statistical comparison by Kruskal-Wallis test and Dunn’s multiple comparisons. p value = * < 0.05, ** < 0.01, *** 2 logs Pa concentration. Pa = P. aeruginosa.

3.5. High production of Pf is associated with significant decline in lung function

Overall, during the study period, a moderate and non-significant change in FEV₁ was noted for the entire cohort (0.23 %-predicted/year, p = 0.35), accounting for repeated measures and adjusting for sex, age and HEMT use. However, the Pa− group had a positive trajectory in their FEV₁ of 0.7 %-predicted/year(p = 0.03). Although the Pa+ group did experience a decline, this was of low magnitude and non-significant at −0.27 %-predicted/year (p = 0.3).

To evaluate if Pf presence was association with lung function and to take into account the effect of Pa, we estimated the rate of decline for the Pa−, Pa+Pf− and Pa+Pf+ groups. Again, only small changes in FEV₁ were noted for the Pa+ positive pwCF whether Pf− or Pf+ (−0.17 and −0.46 %-predicted/year, p = 0.8 for both). Given the varying levels of Pf production (in comparison to Pa concentrations) noted on sputum and findings on lung explants, we evaluated this further by assessing for differences between Pa+Pf+^LOW and Pa+Pf+^HIGH pwCF. We find that those patients with high Pf production had a significant FEV₁ decline of 1.53 %-predicted per year in, which is significantly different that the other groups (p < 0.0029 for all comparisons) (Fig. 3, Table S1). This effect was independent of HEMT use, the only other potential confounder variable to be found significant (Table S1). These data indicate that within a Pa infected airway, production of high concentrations of Pf is associated with faster decline in lung function over time.

Fig. 3.

Yearly rate of change in FEV₁ %-predicted grouped by Pf status. Display of slope and standard error for each group as calculated by mixed effects model for repeated measurements for each individual participant and adjusted for HEMT (final model estimates in Table S1). The Pa+Pf+^HIGH group experienced a significant decline at −1.53 %-predicted/year (p = 0.0029), which was also significantly higher than the other groups (p < 0.05 for all comparisons). p value = * < 0.05, ** < 0.01, *** 2 logs Pa concentration. Pa = P. aeruginosa.

3.6. Pf presence in sputum is associated with inflammatory activity in the airway of pwCF

Given the persistence of Pf in the CF airway, we wanted to investigate if Pf altered immune responses by profiling sputum cytokines among pwCF in our cohort (Table S2). In 241 sputum samples, from 51 subjects with CF and 11 healthy controls, we found by GEE a significant increase in multiple pro-inflammatory cytokines in the Pf+ samples, specifically of G-CSF, GM-CSF, IL-1b, IL-6, Leptin, and MIP-1a, as well as the immunomodulatory cytokine, IL-10 (Fig. 4A-J). In contrast, IL-8 was significantly lower in the presence of Pf (Fig. 4E). Interferon levels were below the limit of detection for most samples which precluded comparisons between groups. These data demonstrate that infection with Pa strains lysogenized by Pf is associated with a distinct immune response.

Fig. 4.

Box Plots of Cytokine Profiles of Sputum by Pf status. Cytokines were measured in sputum fluid and reported as pg/mL. Concentrations were log transformed and analyzed by GEE to account for multiple comparisons and repeated measures. Cytokines of interest and with significant differences between conditions are shown and with p values for the comparison of Pa+Pf− vs Pa+Pf+ shown on each graph. A) G-CSF, (B) GM-CSF, (C) IL-1β, (D) IL-6, (E) IL-8, (F) IL-10, (G) Leptin, (H) MIP-1a, (I) TNF-a, (J) TNF-β. Matrix for p-values for all comparisons presented in Supplemental Table 2. HC = healthy control, Pa− = no P. aeruginosa infection, Pa+Pf− = P. aeruginosa infection but no Pf phage infection, Pa+Pf+ = P. aeruginosa infection with Pf phage infection. * p > 0.05, ** p < 0.0001.

In contrast, there was no significant difference in neutrophil elastase between the Pa−, Pa+Pf− and Pa+Pf+ samples.(Fig S1A). The concentrations of Pf and elastase are correlated but not independently from Pa concentrations or age.(Fig S1B-D) These data indicate that the more rapid decline in lung function seen in the subjects with high Pf production is not likely secondary to proteolytic tissue destruction by neutrophil elastase.

3.7. Co-infections are not associated with Pf presence

To evaluate the associations of Pf with other CF pathogens we evaluated sputum culture results, noting any other bacteria or fungus also grown. Overall no significant differences in the presence of other pathogens was associated with the presence of Pf.(Fig S2A-G) There was also no difference in incidence of allergic bronchopulmonary aspergillosis between subjects with and without Pf.(Fig S2H) These data suggest that Pf may not significantly affect interactions with other pathogens in the CF airway.

4. Discussion

We demonstrate that a high Pf phage production is associated with lung function decline in a longitudinal CF cohort. Building on our previous report of Pf presence in a cross-sectional cohort study [13], we now show Pf associates with lung function trajectory. Prior studies have shown Pf’s ability to elicit an immune response in vitro [14], as well as in murine models of chronic wounds and acute pneumonia [15,16,22]. Similarly, we see differences in sputum cytokine concentrations associated with Pf presence, consistent with exaggerated inflammation with elements of an antiviral response. Additionally, in a subset of subjects who underwent lung transplantation we demonstrate birefringence in secretions within the airway of the Pf+ subjects consistent with liquid crystalline structures which Pf induces when in contact with polymers known to be abundant in the CF airway.

We detect Pf in the sputum of over half of Pa-infected pwCF in our cohort. Although we observe no difference in baseline lung function, we note a higher number of samples per patient in the Pf+ group, potentially indicating increased sputum production due to Pf-induced airway inflammation. Mucus hypersecretion, a known feature of CF-related inflammation, may contribute to this phenomenon. This is also supported by elevated sputum IL-1b levels in this group, which is associated with increased muc5ac and muc5b secretion [23]. Notably, this difference in sputum production cannot be attributed to HEMT use, as there was no significant difference between groups.

While we did not see an increased mortality or the incidence of lung transplant during the study period between groups, we did see an accelerated loss of lung function in those with highest sputum concentrations of Pf in relation to concentration of Pa. This suggests that an abundance of Pf may be necessary to see the deleterious effects – robust liquid crystal biofilms, sequestration of antibiotics and immunologic response. Given that chronically infected CF airways can host multiple strains of the same bacterial organism [24], it is possible that individuals with lower Pf concentrations may have “mixed” infections of Pa, with some strains harboring and producing Pf while co-infecting Pa strains lack the Pf prophage. Though we had suspected that Pf may also provide a selective advantage to other organisms co-existing in the airway, our review of CF sputum culture results did not reveal any trends to support a preferential advantage.

The CF airway is a complex environment with evidence of inflammation noted even before infection onset, followed by a dysregulated immune response once infection occurs [25,26]. Comparing cytokine concentrations in patient sputum, where co-infections and disease severity vary, with controlled in vitro experiments and animal models poses challenges. However, we did observe significant differences in cytokine concentrations in Pf+ sputum samples in comparison to the other categories. Specifically, elevations are seen in G-CSF, GM-CSF, IL-1b, IL-6, MIP-1a, TNFa, and IL-10. Both IL-1b and IL-6 are pro-inflammatory with IL-1b known to be hyper-secreted in CF sputum and associated with mucus impaction and lung disease [23,27,28]. These elevations in IL-1b and IL-6 have not been observed in vitro or animal models of Pf exposure [16,22,29], and may be attributable to the hyperinflammatory response and immune dysregulation associated with CFTR dysfunction. IL-10 is an anti-inflammatory cytokine associated with an anti-viral immune response [30], which has been previously reported as reduced in the CF airway [31], but here was increased in Pf+ sputum. This contrasts the observed decrease in IL-10 in a murine model of acute Pa pneumonia supplemented with Pf [22]. This difference may be explained by the difference in pathophysiology of acute and chronic infection. Interestingly, IL8–the neutrophil-activating cytokine which is classically very elevated in CF sputum—is decreased in the Pf+ samples though still significantly elevated in comparison to the healthy control sputum. The overall association of Pf with cytokines appears to be pro-inflammatory and perhaps anti-viral.

The observed decline in lung function associated with Pf may have multiple mechanistic explanations. It is possible that liquid crystalline biofilms induced by Pf presence promote chronic infection by Pa in the airway by both providing physical protection from immune cells and antibiotics and by shifting the immune response from anti-bacterial to anti-viral. Chronic Pa infection is well established to be associated with accelerated decline in lung function [2]. Additionally, the Pf associated increase in inflammatory cytokines may incite increased mucus secretion potentially exacerbating poor mucociliary clearance which would lead to increased tissue damage. Further study of individuals with Pf+Pa+ infection is warranted to better understand this relationship.

This study is limited by the aspects of an observational longitudinal study including dropout and attrition, as well as the heterogeneous nature of disease in CF, introducing numerous potential confounding variables. This study period includes both the COVID pandemic and the FDA approval of the most widely used HEMT, elexacaftor/tezacaftor/ivacaftor, which has had a dramatic effect on outcomes in CF. Pandemic-related restrictions led to reduced data collection for approximately 12 months, impacting sputum sample collection and performance of PFTs. This may result in misclassification of Pf status, as patients were classified as Pf+ if Pf was detected at any point during the study. Some subjects may have transiently carried Pf, potentially making associations attributable to Pf more difficult to observe. However, the pandemic affected all subjects in the cohort, and HEMT use and initiation were not significantly different between groups. Additionally, HEMT use did not have a confounding effect in our analysis of FEV₁ decline as a function of Pf.

In summary, we have demonstrated Pf production in sputum to be associated with decline in lung function as well as an inflammatory and anti-viral response in a longitudinal cohort study of pwCF. Pf may play a role in the pathogenesis of CF lung disease when chronic Pa infection is present. Pf may be a useful prognostic biomarker which could be used to identify those at risk for precipitous decline and thus signal the need for early and aggressive intervention. A biomarker of this type could prove very useful in the post-HEMT era of CF care where the general trend has been to decrease therapies and interventions. Additionally, Pf may have potential as a therapeutic target in disruption of biofilm and treatment of chronic Pa infection.