Chronic Pseudomonas aeruginosa Lung Infection Is IL-1R Independent, but Relies on MyD88 Signaling

Pseudomonas aeruginosa is an opportunistic pathogen known to cause acute pneumonia as well as persistent chronic infections. The prevalence of P. aeruginosa infections has increased over the past decade and represent a major clinical concern. P. aeruginosa is a frequent infection in intensive care unit and a leading cause of mechanical ventilation, trauma, or viral infections, with a significantly higher mortality compared with other pathogens (1). P. aeruginosa is also intrinsically associated with chronic obstructive pulmonary disease and cystic fibrosis (CF) and is a common cause of recurrent infections (1, 2). The persistent colonization of the airways by P. aeruginosa is associated with chronic inflammation and remodeling accompanied by a progressive decline of respiratory function, leading ultimately to premature mortality (3). Despite considerable advances in antimicrobial chemotherapy, treatment of chronic infections is problematic because of antibiotic resistance (4). Thus, there is a need of novel classes of anti-infectious agents as alternative or complementary strategies to antibiotic. However, the pathophysiological mechanisms that govern host immune responses to chronic P. aeruginosa infections remain to be fully deciphered.

Pattern-recognition receptors such as TLR and IL-1R family members are involved in the detection of pathogens and the initiation of inflammatory responses after infections (5). The adapter protein MyD88 (6) is shared by receptors of the TLR and IL-1R families and is critical to control bacterial lung infections, including mycobacterial infection (7–9), but also early acute P. aeruginosa infection (10, 11). Type 1 IL-1R (IL-1R1) heterodimerize with IL-1Ra upon IL-1β binding (5). IL-1β is prototypal proinflammatory cytokine that is synthesized as a precursor protein, pro–IL-1b, under the control of the MyD88–NF-κB axis and required a proteolytic cleavage mainly mediated by the class I cysteine protease caspase-1 (ICE1) (12). ICE1 requires activation in response to pattern-recognition receptor signaling, triggered by a multiprotein intracellular platform called inflammasomes (10, 11, 13).

Proinflammatory cytokines, including the ones from the IL-1 family, have been suggested to be critical for the protection against P. aeruginosa, although a complete understanding of their effects according to the infectious mode of the bacteria has not been achieved yet. Previous reports suggested that IL-1β expression was mandatory for host protection against acute infection (14). During P. aeruginosa chronic colonization, it has been shown that NLRP3 inflammasome is activated leading to the production of IL-1β and the NLRP3 inhibitor MCC950 inhibited inflammation in CF mice (15). In addition, we have reported a critical role of IL-1β in the induction of deleterious inflammatory response in the context of CF-associated mutations (16).

In this study, we revisited the role of IL-1β/IL-1R1 and MyD88 signaling in host resistance against a chronic P. aeruginosa lung infection using the clinical strain P. aeruginosa RP73. We report in this study that IL-1β is not essential, whereas MyD88 is required to control chronic infection as completed MyD88-deficient mice develop a severe and lethal lung pathology. Interestingly, genetic deletion of MyD88 in alveolar type I (ATI) and alveolar type II (ATII) epithelial cells or in myeloid cells does not impair the control of P. aeruginosa chronic infection. Thus, host resistance is independent of MyD88 signaling by IL-1R1 to control the inflammatory response and host resistance to chronic P. aeruginosa infection.

MyD88 (17) and IL-1R1–deficient mice (18), as well as cell-specific, MyD88-deficient mice generated in-house (19), all on C57BL/6 background, and littermate controls were bred at TAAM (UPS44, CNRS, Orléans, France). All mice were housed under specific pathogen–free conditions and had access to food and water ad libitum. All animals were followed twice daily. All animal experiments complied with the current European legislative, regulatory, and ethical requirements and were approved by Ethics Committee for Animal Experimentation of CNRS Campus Orléans (CLE CCO 2012-042).

The P. aeruginosa strain RP73 strain (gift from Pr. A. Bragonzi, San Raffalele Hospital, Milano, Italy) was cultured on agarose beads as previously described (20), and mice were infected with freshly prepared inoculum of 1 × 10⁶ CFU by endopharyngeal administration under isoflurane inhalation anesthesia, whereas control mice received uncoated agar beads. As an acute infection control, we used the P. aeruginosa 01 strain at 1 × 10⁶ CFU, as previously described (21). Tobramycin (Sigma-Aldrich) was administered by gavage at 100 mg/kg on days 2 and 3 after RP73 P. aeruginosa infection, and mice were killed 6 h after the last administration.

Mice were injected i.p. with anti–IL-1β Ab (gift from Dr. H. Gram, Novartis Pharma, Basel, Switzerland) at the dose of 10 µg/g body weight in 200 µl. Control mice received the same dose and volume of the isotype control Ab.

Lung total weights were recorded at sacrifice and expressed as a percentage of body weight. Lung homogenates were prepared in 1.5 ml of isotonic saline solution using a Dispomix tissue homogenizer (Medic Tools, Bern, Switzerland). The 10-fold serial dilutions of lung homogenates were plated on brain heart infusion agar plates (BioValley France). Plates were incubated at 37°C and 5% CO₂, and the numbers of CFU were enumerated after 24 h. In vitro antibiotic resistance was assessed in bacterial suspension cultures with increasing tobramycin concentrations, known as minimal inhibitory concentration assay.

After sacrifice, bronchoalveolar lavage fluid (BALF) was collected by cannulating the trachea and washing the lung with 1 ml saline at room temperature. The lavage fluid was centrifuged at 400 × g for 10 min at 4°C, and the supernatant was stored at –80°C for analysis. The cell pellet was resuspended in 1× PBS and counted in a hemocytometer chamber, and cytospins were prepared to count the neutrophils, macrophages, lymphocytes, and eosinophils as described before (16).

TNF-α, IL-1β, IL-6, and keratinocyte-derived chemokine (KC; CXCL-1) as well as myeloperoxidase (MPO) levels in lung homogenates were measured by ELISA (Duoset Kit; R&D Systems, U.K.) according to the manufacturer’s instructions. Total protein concentration to assess respiratory barrier integrity was analyzed in BALF using Bradford assay with BIORAD DC Protein assay kit (Bio-Rad Laboratories, France), according to the manufacturer’s instructions.

Lungs were fixed in 4% buffered formalin, dehydrated in ethanol, and embedded in paraffin. Sections (3 μm) were stained with H&E, and all lung sections were evaluated by light microscopy by three independent observers (J.P., T.M., and B.R.) for recruitment of inflammatory cells in bronchi, bronchiolitis, and respiratory epithelial injury using semiquantitative severity score (0–5) as described before (16).

Statistical evaluation of differences between the experimental groups was done by Mann–Whitney nonparametric test or log-rank test for survival. All tests were performed with GraphPad Prism (GraphPad Software, San Diego, CA; www.graphpad.com). A p value <0.05 was considered significant and was symbolized with an asterisk (*), and **p < 0.01 and ***p < 0.001. A p value >0.05 was considered NS.

To establish a chronic model, we prepared agar beads coated with the clinical mucoid RP73 strain (provided by Prof. A. Bragonzi). First, the course of infection with the nonmucoid PAO1 (not on beads) and the mucoid RP73 strain was compared. Agar beads coated with the P. aeruginosa RP73 strain were administered at 1 × 10⁶ CFU by endopharyngeal instillation to BL6 mice. Infection with both strains caused clinical signs disease, but all mice inoculated with the P. aeruginosa RP73 strain survived, whereas mice infected with the 1 × 10⁶ CFU PAO1 strain died within 2 d upon infection (Fig. 1A), as reported earlier (21). Infected BL6 mice lost ∼15% of body weight at 48 h but recovered over the next 10 d and survived the infection (Fig. 1A, 1B). Lung weights as surrogate marker of inflammation were increased in P. aeruginosa RP73 strain–infected mice on days 1 and 3 and were elevated up to day 21, suggesting chronic lung infection (Fig. 1C). To ascertain the chronicity of the infection, the viable bacteria recovered from BALF and lung homogenates were assessed. Viable P. aeruginosa RP73 were detected in the lung at day 1 at 3 × 10⁶ CFU counts and decreased to 1 × 10⁶ CFU at day 7 and were stable till day 21. The BALF CFU counts at day 1 were 1 × 10⁵ CFU and increased significantly to 3 × 10⁶ CFU at day 14 and declined to 1 × 10⁵ CFU at day 21 (Fig. 1D). Inflammatory cell recruitment in the BALF was initially characterized by abundant neutrophils peaking at 5 × 10⁶ on day 3 and reduced on day 7. By contrast, macrophages were increased from day 3 to 14, whereas lymphocyte counts were elevated from day 1 to 14. Total protein in BALF as a measure of respiratory barrier damage was detected from day 1 to 7 (Fig. 1E). Further, P. aeruginosa RP73 strain infection was associated with significantly increased MPO as a marker of neutrophil infiltration, CXCL1/KC, IL-1β, and IL-6 in lung (Fig. 1F).

Therefore, the data demonstrate that the mucoid P. aeruginosa RP73 strain causes chronic lung injury, chronic infection, bronchitis, and pneumonia in BL6 mice.

To characterize the inflammation in chronic P. aeruginosa infection, we investigated the infected lung tissue in BL6 mice. Unlike the empty agar beads, which do not cause any significant tissue reaction, the P. aeruginosa RP73–loaded beads elicit respiratory epithelial cell damage with cell desquamation visible at day 1 and 3, followed by repair on day 7 (black arrows).

Epithelial injury is associated with acute bronchitis and bronchiolitis, with infiltration by neutrophils, macrophages/monocytes, and lymphocytes in the first few days, which then develops into a chronic biofilm infection with pneumonia at ∼7 d (red arrows), which declines at day 21 (Fig. 2A, 2B).

Biofilm infections are usually associated with antibiotic resistance. We tested the resistance to tobramycin in suspension culture of the P. aeruginosa RP73 strain and confirm that this strain is sensitive to tobramycin (Fig. 2C). Furthermore, administering a high-tobramycin dose (100 mg/kg on days 2 and 3 by gavage) to mice with established chronic RP73 P. aeruginosa lung infection had no effect on the bacterial burden, relative lung weight, and lung inflammation at day 7 postinfection (Fig. 2D). These data reproduce the important feature of an antibiotic resistance of clinical biofilm infection and are in line with previous reports that P. aeruginosa RP73 biofilm infection is resistant to antibiotic treatment.

Having established the chronic biofilm infection model, we asked which factors might reduce host resistance. The common TLR adapter protein MyD88 is critically involved in acute P. aeruginosa infection (22). We report in this study that P. aeruginosa RP73 strain (1 × 10⁶ CFU)–infected, MyD88-deficient mice are highly susceptible to infection and succumb to infection within 4 d, whereas littermate control mice survive (Fig. 3A). MyD88-deficient mice present increased lung weight as surrogate marker of lung inflammation. Total protein levels in BALF were increased indicating respiratory barrier injury/disruption and increased numbers of neutrophils, macrophages, and lymphocytes (Fig. 3B) associated with apoptosis and cell death of neutrophils and epithelial cells (Supplemental Figs. 1, 2). MPO, CXCL1/KC, and IL-6 levels in lung were increased in RP73 P. aeruginosa–infected, MyD88-deficient mice, whereas IL-1β and IL-17A levels were not different (Fig. 3C). The infection was uncontrolled with a massive increase of the bacterial load in lung and BALF (3 log CFU), as compared with littermate control mice (Fig. 3D). Microscopic analysis of the lung of MyD88-deficient mice at day 3 showed a trend for increased acute inflammatory cell infiltration, epithelial cell injury, acute alveolitis, and bronchiolitis as compared with infected littermate controls (Fig. 3E). Therefore, the data demonstrate that MyD88 signaling is critical to control chronic RP73 P. aeruginosa–induced lung epithelial damage, inflammation, and lethal infection.

Having demonstrated that MyD88 is essential to control infection, we asked whether lung epithelial cells are involved in the inflammatory response and control of chronic P. aeruginosa RP73 infection. To address this question, we used cell-specific, MyD88-deficient mice developed in the laboratory and described before (19).

Importantly, we find that P. aeruginosa RP73 strain infection is controlled in mice deficient for MyD88 in AQP5⁺ ATI lung cells (MyD88 floxxAQP5-cre mice, in short MyD88-AQP5 knockout [KO]). Clinical signs and body weight loss, lung weight, and inflammation with macrophage and neutrophil recruitment in the bronchoalveolar lavage and protein extravasation in the airways after P. aeruginosa RP73 infection were reduced in MyD88-AQP5 KO mice as compared with littermate controls (Fig. 4A, 4B). Although MPO in the lung tissue was increased in infected mice, the increase of CXCL1/KC and IL-6 level in the lung seen in complete MyD88-deficient mice was attenuated in MyD88-AQP5 KO mice upon infection, comparable to littermate control mice, whereas IL-1β and IL-17A levels were not different (Fig. 5C). In addition, MyD88-AQP5 KO mice controlled bacterial load in the lung and BALF similar to littermate infected control mice (Fig. 4D). Histologically, the extent of epithelial injury, inflammatory cell infiltration, and bronchiolitis and pneumonia was attenuated in MyD88-AQP5 KO mice comparable to littermate controls, in contrast to complete MyD88 KO mice (Fig. 4E). Therefore, the data obtained from MyD88-AQP5 KO mice show that MyD88 signaling by lung epithelial ATI cells do not contribute to the control of lung inflammation and infection by the P. aeruginosa RP73 strain in mice.

Furthermore, myeloid cell–specific, MyD88-deficient mice (MyD88 floxxlysM-Cre, in short MyD88-lysM KO) controlled infection and inflammation similar to MyD88-AQP5 KO and wild-type littermate controls, unlike complete MyD88-deficient mice (Fig. 4A–E). Therefore, although complete absence of MyD88 pathway impairs the control of chronic P. aeruginosa infection, MyD88 in myeloid cells are dispensable to control chronic P. aeruginosa RP73 infection. However, the LysM promotor is not restricted to myeloid cells but is also expressed in ATII epithelial cells (Supplemental Fig. 3) (23, 24), suggesting that ATII epithelial cells are not required.

Therefore, our data suggest that MyD88 signaling in ATI or ATII epithelial cells, as well as in myeloid cells, is not required because in their absence, P. aeruginosa RP73 infection allows control of inflammation and infection, and the mice survive at least for 3 wk (data not shown), unlike complete MyD88-deficient mice.

We published before that IL-1β is required to control in acute P. aeruginosa infection using PAO1 strain using IL-1R1 KO mice (16) and report in this study increased IL-1β in bronchoalveolar lavage upon P. aeruginosa RP73 strain infection (Fig. 1F). Therefore, we asked whether IL-1 plays a role in chronic P. aeruginosa RP73 biofilm infection and whether the high susceptibility to infection of MyD88-deficient mice is related to IL-1R signaling. Therefore, we infected IL-1R1–deficient mice and found that they survived chronic P. aeruginosa RP73 infection as BL6 mice (data not shown). Lung weight, neutrophil, and lymphocyte recruitment in BALF and lung bacterial load were similar in P. aeruginosa RP73–infected, IL-1R1–deficient and littermate wild-type control mice (Fig. 5A). The increase of MPO, CXCL1/KC, and protein leak in BALF as well as the lung injury and inflammation were similar in infected IL-1R1–deficient and in littermate control mice (Fig. 5B). Thus, IL-1R1–deficient mice control chronic P. aeruginosa infection, as infected BL6 control mice, unlike complete MyD88-deficient mice, which succumb to infection within 4 d.

Finally, to corroborate these data, we asked whether neutralizing IL-1β Ab alters host inflammation and resistance to P. aeruginosa RP73 lung infection in BL6 mice. Blockade of IL-1β by neutralizing Abs (200 µg, injected daily i.p.) did not increase the sensitivity of BL6 mice to P. aeruginosa RP73 infection, as all the inflammatory and infection parameters measured were comparable to untreated infected BL6 mice (Fig. 5C, 5D).

Therefore, our data demonstrate that IL-1R1 signaling is not required to control chronic P. aeruginosa RP73 infection, and IL-1β Ab blockade does not reduce host resistance to chronic P. aeruginosa infection in control B6 WT mice.

Chronic biofilm lung infection with P. aeruginosa is a common complication in patients with CF or other predisposing factors and represents a therapeutic challenge. We established a chronic lung inflammation in BL6 mice using the clinical P. aeruginosa RP73 strain to investigate the role of innate immune pathway MyD88/IL-1R1 in epithelial barrier injury, inflammation, and host resistance. A critical role of MyD88 is established in response to endotoxin (17) and several bacterial infections in mice, like with Burkholderia (9), Listeria (25, 26), Legionella (27), or Mycobacterium tuberculosis (8), to mention a few major infections. A role of MyD88 was shown in the early response to acute P. aeruginosa aerosol infection with defective NF-κB, IL-1β, or TNF response in the absence of the MyD88 pathway (11) while dispensable at 48 h postinfection (10).

Because chemotherapy resistance is a major therapeutic issue of chronic P. aeruginosa biofilm infection in patients, we first investigated whether the clinical P. aeruginosa RP73 strain infection is sensitive to chemotherapy. We found that although the P. aeruginosa strain in suspension culture is sensitive to tobramycin, in vivo chronic P. aeruginosa RP73 lung infection is resistant to tobramycin treatment. We demonstrated before that the acute PAO1 infection is sensitive to meropenem and tobramycin (28). Increased antibiotic resistance of bacteria with biofilm growth is a major therapeutic challenge (29).

In this study, using cell-specific deletion of MyD88 signaling in ATI or ATII epithelial cells, as well as of myeloid cells (19), we show that these cells are not involved in host resistance because in their absence, P. aeruginosa RP73 infection and inflammation was controlled, and the mice survived at least for 3 wk, unlike complete MyD88-deficient mice.

Previous studies focused on the role of innate pattern-recognition receptors such as TLRs in P. aeruginosa infection. The consensus of these studies was that the sensing of P. aeruginosa is more complex and not restricted to a single TLR, but it appears the MyD88 adapter protein is essential to initiate the cytokine inflammatory response and control acute P. aeruginosa infection (10, 11, 22). In this study, we extend this finding and report for the first time, to our knowledge, that MyD88 is essential to control chronic P. aeruginosa RP73 biofilm lung infection.

Importantly, we demonstrate that MyD88 signaling in lung epithelial cells is not required to control lung inflammation and infection because ATI-specific, MyD88-deficient mice control inflammation and infection as littermate control and BL6 mice. We confirm that myeloid cells play a redundant role because MlyscrexMyD88-deficient mice control inflammation and infection. Because the LysM promotor is also expressed in ATII cells, we also include that both alveolar cell types are not required (23, 24).

Therefore, we conclude that MyD88 signaling in alveolar epithelial cells, the first contact site for the bacteria and in myeloid cells, is not required to control P. aeruginosa RP73 chronic infection. Previous studies report that radioresistant stromal cells in MyD88 bone marrow chimeric mice control acute P. aeruginosa infection (30). By contrast, Mijares et al. (22), using a transgenic expression of MyD88 by a CC10/Scgb1a1 promoter in the secretory club/basal cells and goblet cells of the bronchiolar region, but not in alveolar AT1 and -2 cells, restored MyD88-deficient phenotype in the acute P. aeruginosa infection model. In this study, using biofilm P. aeruginosa RP73 infection, we show directly that MyD88 signaling in the alveolar epithelium, in either ATI or ATII cells and myeloid cells, plays no role for the high susceptibility to chronic RP73 P. aeruginosa infection, as in their absence, P. aeruginosa infection is controlled.

Single-cell transcriptomic analysis of mouse lungs revealed that LysM is not only expressed in myeloid cells but also in ATII epithelial cells (31, 32). Therefore, we postulate that both myeloid cells or ATI and ATII cells are not required to control inflammation and infection because in their absence, the uncontrolled infection found in complete MyD88-deficient mice is drastically attenuated and comparable to littermate controls.

We did not investigate the role of other cytokines in chronic P. aeruginosa infection but demonstrated a protective role of IL-1, IL-17, and TNF in unrelated bacterial infections. We reported that IL-1 is required to control mycobacterial infection because IL-1R1–deficient mice are highly sensitive to M. tuberculosis (7, 33), whereas other IL-1 family members, including IL-36R, exert no significant effect on mycobacterial infection (34). Further, we demonstrated that both T cell and myeloid cells expressing TNF and TNR1 are essential to control mycobacterial infection (35, 36). We reported before that IL-1 is required for IL-17A differentiation (37). Although Th17 cells play no significant role in M. tuberculosis infection (36), IL-17A is required for host resistance to chronic P. aeruginosa lung infection (38).

Because the MyD88 adapter protein is involved in signaling of IL-1R family members, we investigated the role of IL-1 and report in this study that IL-1R1 signaling is not required for the control of chronic P. aeruginosa–induced lung inflammation and infection. Furthermore, IL-1β Ab blockade did not diminish host resistance in BL6 mice. Therefore, the role of other cell types or other members of the IL-1R family or TLR family depending on MyD88 signaling need to be explored in future studies.

In conclusion, the high susceptibility to chronic P. aeruginosa infection observed in complete MyD88 KO mice is not found in cell-specific deletion of MyD88 in ATI or ATII epithelial and myeloid cells, suggesting that these cells are not required for infection control. Furthermore, host resistance is IL-1R1 independent. Thus, IL-1–targeted therapy for inflammatory diseases in patients with autoimmune inflammatory disease may not increase the susceptibility to chronic P. aeruginosa biofilm infection.

The authors have no financial conflicts of interest.