Agar beads were prepared following a previously published protocol with minor modifications (Growcott et al., 2011). Briefly, colonies from an overnight culture of P. aeruginosa were diluted with 30 mL sterile PBS to achieve a suspension of 0.3 O.D. at 600 nm. Cultures were washed and the pellets re-suspended in 4 mL sterile PBS. One milliliter of bacterial suspension was added to 10 mL of sterile 2% agar at 52°C and the mixture added to 15 mL heavy mineral oil (52°C) supplemented with 0.02% sorbitan-monooleate (Sigma-Aldrich) and mixed thoroughly. Beads were centrifuged at 10,000 × g at 4°C, extensively washed with sterile PBS and filtered using a 200 μm nylon mesh. Beads were resuspend in 2 × w/v PBS and 500 μL bead suspension (≈ 2E7 colony forming units) was used per animal. Sterile beads were prepared using sterile saline. Each bead solution was plated on blood-agar plates and incubated overnight at 37°C for inoculum validation.

All animal experiments were conducted according to the guidelines of the Federation of European Laboratory Animal Science Associations and approved by the University of Antwerp Ethics Committee. Rats were anesthetized using 100 mg/kg ketamine and 1 mg/kg medetomidine and intubated utilizing a tilting intubation platform (Hallowell EMC). Briefly, anesthetized rats were placed on the tilting platform in a horizontal position. An elastic band was placed behind the front incisors and attached to the platform. Next a 14 G catheter was placed into the trachea with the aid of a guide wire. With 14 G catheter, animals were instilled with agarose beads, extubated, and anesthesia antagonized with 300 μg/kg atipamezole. Animals were placed back in their cages and followed up.

A total of 36 adult male Wistar rats (mean weight 340 g, SD = 23 g, Charles River) were used in this study and randomly assigned into one of the five experimental groups: (i) Non-manipulated animals served as the control group (n = 12; no bead group); (ii) Two sterile agarose (St) bead groups euthanized at day 1 and day 3 post-instillation, respectively (n = 6 animals per group); (iii) Two chronic pneumonia (Pa-bead) groups where animals were instilled with 2 × 10⁷ CFU of P. aeruginosa enmeshed in agarose beads and euthanized at day 1 and day 3 post-infection, respectively (n = 6 animals per group).

Animals were closely monitored for clinical signs of pneumonia (SI Table 1). Euthanasia was performed by isoflurane overdose and blood collected by cardiac puncture utilizing EDTA coated tubes. The trachea was exposed and lungs lavaged using a 14-gauge angiocatheter with 20 mL/kg body-weight of ice-cold sterile PBS. Right lung was removed, washed in sterile PBS and snap frozen in liquid nitrogen. Left lung was washed and fixed overnight in 2% paraformaldehyde and prepared for paraffin embedding.

Lung pathology scoring was performed on H&E stained 5 μm thick paraffin sections, as done previously (Matute-Bello et al., 2011). All lungs used for histology were treated with standardized protocols to allow comparisons between the groups. Lung pathology was assessed blinded by a trained pathologist. Amount of neutrophils for each group were manually counted on 8–10 consecutive images grabbed using 200x magnification per slide for each animal, eosinophils were quantified on 400x magnification. For slides that showed heterogenic staining pattern, most affected areas were imaged for quantification.

Immunohistochemistry was done as described by us previously (Wils et al., 2012). Briefly, antigen retrieval was performed in citrate buffer (0.018 M citric acid.H₂O and 0.082 M sodium citrate.H₂O) using microwave heating. Endogenous peroxidase was blocked using 0.3% H₂O₂ for 20 min. Blocking of non-specific antigens was done using 1:5 diluted normal horse serum in 1% BSA solution in PBS (PBS-BSA) for 30 min. Primary antibodies were diluted in PBS-BSA solution and incubated overnight at 4°C: anti-CD68, 1:200 dilution (Abd Serotec MCA341R); anti-arginase-1, 1:400 dilution (Santa Cruz Sc-18354); and anti-iNOS, 1:100 (Abcam ab15323). Secondary biotinylated antibodies (Jackson Immunoresearch) were used in 1:200 dilution, incubated for 30 min at room temperature, washed and incubated with extravidin-HRP for 30 min followed by DAB (5', 5' diaminobenzedine, Dako) development. Double-labeled immunohistochemistry was performed using anti-CD68 (1:200 dilution), anti-arginase1 (1:400 dilution), and anti-iNOS (1:100 dilution) with DAG-cy3, DAM-cy5 and DAR-cy5 in 1:200 dilutions. Sections were co-stained with 5 μg/mL DAPI (Sigma-Aldrich) and coverslipped using antifading PBS-glycerol mounting medium (Citifluor).

Light microscopy images were grabbed on Olympus UC30 color camera. Quantification was performed by calculating the percentage of stained area using pixel-by pixel analysis after spectral de-convolution of the image using the IHC profiler plug-in in ImageJ v1.47 (Varghese et al., 2014). The mean percentage of positive stained area per animal was used. Slides were randomized before analyses by a blinded investigator. High-resolution images from double-labeled immunofluorescence stained sections were grabbed using a dual spinning disk confocal microscope (UltraView VoX, PerkinElmer) and images analyzed using Volocity (PerkinElmer). Proportions of Arg1+CD68+ and iNOS+CD68+ cells were manually counted by the blinded investigator.

Total RNA from lung tissue was extracted using RNeasy-mini spin columns (Qiagen) after grinding in liquid nitrogen. RNA integrity and concentrations were estimated using RNA-nanochips on Bio-analyzer (Agilent). Extracted RNA was converted to cDNA by reverse transcriptase (RT² First Strand Kit, Qiagen). Quantitative PCR was performed using CFX connect (Bio-Rad) using custom PCR-arrays (Qiagen) for following genes of interest: IL-1β, IL-4, IL-5, IL-6, IL-10, IL-12a, IL-13, IL-17a, IL-17f, IL-22, IL-23a, Ifnγ, Tnfα. Housekeeping genes included were: Ywhag, Actb, Polr2b, Gapdh, Sdha, and Tbp. Additionally, a random genomic DNA contamination control, three reverse transcriptase controls, and three additional PCR controls were used. Data was validated on independently extracted RNA and analyzed using in-house primers (primer sequences available on request) using SYBRGreen assay. Data was analyzed using comparative C_T method as described earlier (Kumar-Singh et al., 2006; Schmittgen and Livak, 2008). Briefly, the combined average Ct value of the control group was used to calculate the fold differences in the study groups.

Data analyses were performed using SPSS version 23 (IBM) and presented as either averages or average fold-differences with the standard errors of mean. Kolmogorov-Smirnov test for normality was used for each data set before testing statistical significance of differences on log-transformed values by one-way ANOVA with post-hoc Bonferroni corrections. Survival analyses were performed using Kaplan Meier estimator with Mantel-Cox log rank test for testing significant differences. Values of all significant correlations (P < 0.05) are given with degree of significance indicated (^*P < 0.05, ^**P < 0.01, ^***P < 0.001).

To study whether a Th2 response could be provoked by the presence of biofilm matrix, we intratracheally instilled animals with sterile agar beads and analyzed lung transcript levels of key cytokines at day 1 (d1) and day 3 (d3) post-bead instillation time points. Compared to non-manipulated control animals (No bead group), sterile beads did not cause an increase in proinflammatory Th1 cytokine expression of Ifnγ, Tnfα, IL-1β, IL-12a at d1 post-bead instillation (Figure 1A, white bars). Analyzing the lung transcript levels for the Th17 cytokine family, we showed increased expression of IL-17a (157-fold, P < 0.001) and IL-22 (112-fold, P < 0.01) at d1 post-bead instillation that were sustained at the d3 time-point (IL-17a, 158-fold, P < 0.001; IL-22, 65-fold, P < 0.01; Figure 1B).

Interestingly, lung transcript levels of Th2 cytokines in sterile agar beads-instilled animals were markedly elevated with increased expression of IL-5 and IL-13 at d1 post-bead instillation (IL-5, 39-fold, P < 0.01; IL-13, 102-fold, P < 0.001) that remained elevated at d3 (IL-5, 30-fold, P < 0.05; IL-13, 26-fold, P < 0.001). Transcript levels of IL-4 were not altered by sterile agar-bead instillation (Figure 1C). Additionally, elevated transcript levels compared to control group were also noticed for immunomodulatory cytokines IL-6 (d1, 38-fold, P < 0.001; d3, 21-fold, P < 0.05) and IL-10 (d1, 15-fold, P < 0.05; d3, 20-fold, P < 0.05; Figure 1D).

IL-5 and IL-13 are potent inducers of eosinophil chemotaxis in the lung (Pope et al., 2001) and IL-13 has been shown to be involved with M2 macrophage polarization causing depressed antibacterial immunity (Knippenberg et al., 2015). Therefore, we next studied these two innate type 2 effector cells. Compared to control animals, sterile beads caused a moderate increase in neutrophil infiltration in lung at both time-points (P < 0.05 for both), confirming previous observations (Cash et al., 1979; Growcott et al., 2011) (SI Figure 1). Additionally, sterile beads also caused an intense eosinophilic infiltration at both studied time-points (P < 0.001 for both; Figures 2A,B,D), especially in regions surrounding bronchioles harboring agarose beads.

Similarly, an intense macrophage infiltration was also associated with sterile beads. However, due to overlapping macrophages especially around the agarose beads, an automated quantification estimating percentage area occupation of immunohistochemical staining was employed as described in Methods. With this protocol, we showed that CD68+ areas were significantly increased at d1 and d3 (two and five times, respectively), compared to control animals (P at least < 0.05 for both; Figures 2A,B,D). This coincided with significantly increased Arg1+ areas at both time-points (P at least < 0.01; Figures 2B,D). These data were confirmed on confocal double-labeled immunofluorescence microscopy showing that 31% of activated CD68+ macrophages were Arg1+ already at d1 and this proportion of M2 macrophages increased to 82% at d3 compared to controls (P < 0.01 for d1; and P < 0.001 for d3 group; Figures 3A,C). Additionally, a proportional drop in numbers of M1 macrophages (iNOS+CD68+) was observed compared to controls (P < 0.001; Figures 3B,D). These data indicate that agarose biofilm structures induce an acute Th2 innate cellular pathology marked by elevated eosinophils and M2 macrophages in rat lung.

To study how relevant agarose/biofilm-induced Th2/Th17 cytokines are compared to the infectious situation, and whether these Th2 cytokines and Th2 innate cells observed in the sterile bead model persist after P. aeruginosa co-challenge, we established a Pa-bead model. In this model, rats were intratracheally instilled with agarose beads enmeshed with P. aeruginosa that led to development of progressive pneumonia evident up to d3 post-infection (SI Figure 2A). Pa-bead animals showed ≈50% mortality at d3 post-infection whereas sterile bead animals showed no mortality (Kaplan-Meier, ^***P < 0.001, Mantel-Cox log rank test; SI Figure 2B).

Intratracheal instillation of P. aeruginosa enmeshed beads caused a drastic reduction in the observed Th2 response with both IL-5 and IL-13 attenuated by up to 20 and 30 times, respectively, compared to sterile bead instilled animals (Figure 1C). This coincided with an expected surge at d1 of all studied proinflammatory cytokines (Ifnγ, 99-fold; Tnfα, 117-fold; IL-1β, 82-fold; IL-12a, 16-fold; P for all < 0.001; Figure 1A). In accordance with earlier studies (Lavoie et al., 2011; Lovewell et al., 2014), a high expression of immunomodulatory cytokines IL-6 (6,600-fold) and IL-10 (211-fold; P for both < 0.001) was also observed (Figure 1D).

Interestingly, the surge in IL-17 family of cytokines was much more pronounced. At d1 post-infection, IL-17a and IL-17f were upregulated 25,000 and 15,000-fold, respectively, compared to controls, and ≈20,000 and ≈1,500-fold compared to sterile bead-instilled animals (Figure 1B). These cytokines are not only important in mediating neutrophil chemotaxis (Xu et al., 2014), but are also amongst the important drivers of proinflammatory cytokine expression including IL-1β and Tnfα that we also observed to be elevated in our Pa-bead model. IL-23, another member of the IL-17 family and one of the strongest drivers of Th17 cell differentiation and IL-17 secretion (Bedoya et al., 2013) was also significantly elevated (Figure 1B). Lastly, IL-22 that along with IL-17 has important functions in mucosal immunity against extracellular pathogens, particularly Gram-negative organisms (Aujla et al., 2008; Xu et al., 2014) was also elevated by several 100-folds in lung transcripts (Figure 1B). However, at d3 post-infection, transcripts of the Th17 family subsided to levels comparable to sterile bead-instilled animals. These data suggest that the IL-17 response observed in CF patients is primarily pathogen-driven.

While patient studies have shown an increased arginase activity in BAL fluid of CF patients colonized with P. aeruginosa (Grasemann et al., 2005; Murphy et al., 2010), the individual contribution of the pathogen and the biofilm matrix remain unknown. Additionally, as bacteria are one of the most important drivers of M1 macrophage polarization, we studied whether the increased type 2 innate cellular infiltrates induced by sterile agarose beads persist in the Pa-bead model.

At both d1 and d3 time-points, Pa-bead-instilled animals were characterized by high loads of neutrophils compared to controls (P < 0.01; SI Figure 1) and to the sterile-bead group (P < 0.01 for d1; P < 0.05 for d3; SI Figure 1), as shown previously (van Heeckeren and Schluchter, 2002; Growcott et al., 2011). Interestingly, while eosinophilic counts in the Pa-bead group were reduced at d1 and d3 time points (≈3 and ≈5 times, respectively), these were still highly elevated in the Pa-bead group compared to the controls (P < 0.001; Figures 2C,D). These data are also consistent with sporadic reports of increased eosinophil activation in lung of CF patients with P. aeruginosa pneumonia (Koller et al., 1994).

Importantly, we showed that the Pa-bead group had ≈3-fold increased lung infiltration of CD68+ cells at both d1 and d3 time-points compared to the sterile bead group (P < 0.001; Figures 2C,D). Studying the M1/M2 macrophage markers, we further showed that iNOS+ cells were drastically elevated in the Pa-bead group especially at d1 (d1, 19 times and d3, 5 times, P < 0.01 for both), while Arg1+ cells showed a non-significant increase in infiltration in the Pa-bead model at both time-points compared to sterile bead-inoculated animals (≈3 times for both; Figures 2C,D). Double-labeled immunofluorescence microscopy confirmed these data showing that at d1, 65% of CD68+ activated macrophages were iNOS+ M1 macrophages (P < 0.001; Figures 3B,D), while the population of Arg1+ M2 macrophages was 18% and non-significantly elevated compared to controls (Figures 3B,C). Moreover, at d3, the Arg1+ M2 macrophage proportion increased to 63% of the total activated macrophage population (P < 0.001; Figures 3B,C) while iNOS+ M1 macrophages declined to 18% (P < 0.001; Figures 3B,D). These data indicate that chronic P. aeruginosa infection can sustain the anti-inflammatory M2 cellular innate immune response induced by biofilm matrix-mimicking agarose bead-instillation.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.