Cigarette Smoke Primes the Pulmonary Environment to IL-1α/CXCR-2–Dependent Nontypeable Haemophilus influenzae–Exacerbated Neutrophilia in Mice

Cigarette smoking has become one of the greatest health concerns of the 21st century (1–3). Despite declining smoking rates in developed nations, global smoking prevalence has reached epidemic levels with ∼20% of the world’s adult population considered a smoker (4). Clinical evidence suggests that pulmonary infections in smokers are significantly more severe than in nonsmokers (1–3). The effects of cigarette smoke during pulmonary infection may be life threatening to vulnerable populations such as individuals suffering from chronic obstructive pulmonary disease (COPD) (5). Ninety percent of COPD cases are attributed to chronic cigarette consumption and the progressive loss of lung function that characterizes COPD is periodically exacerbated by microbial infections (6). Understanding the extent to which cigarette smoke modifies host defense mechanisms and predisposes an individual for exacerbated infectious episodes will be crucial knowledge for future health care strategies.

Cigarette smoke is an unusual stimulus for the lung mucosal environment, with both immune activating and immune suppressing characteristics. Cigarette smoke exerts damaging and proinflammatory effects in the lungs but also can directly suppress innate and adaptive immune processes (7–10). Studies in animal models have demonstrated that exposure to cigarette smoke exacerbates inflammatory responses elicited by several different bacterial agents, including nontypeable Haemophilus influenzae (NTHi), Pseudomonas aeruginosa, and Streptococcus pneumonia (11–14). In all studies, the cellular composition of the bacteria-exacerbated inflammatory response was neutrophilic in nature. This exacerbated neutrophilia seemed to contrast findings that suggest that cigarette smoke suppresses the expression of proinflammatory mediators such as TNF-α in response to bacterial stimuli (15, 16). A more detailed examination of the inflammatory processes in a murine model of NTHi-exacerbated, cigarette smoke–induced inflammation yielded the observation that certain well-characterized inflammatory mediators were suppressed, while other proinflammatory factors, not normally induced by NTHi, were now being expressed (14). This altered phenotype was observed specifically at the level of the alveolar macrophage; a critical orchestrator of immune responses in the lung (17). Although the altered lung phenotype is now being characterized, the specific mediators of neutrophilic inflammation within this altered response have not been identified.

Animal models have identified the importance of IL-1R1 signaling to neutrophilic inflammation elicited by cigarette smoke (18–20). The cytokines IL-1α and IL-1β activate IL-1R1 (21). Precursor IL-1α is biologically active as a 31-kDa protein (22). In contrast, IL-1β is initially synthesized as pro–IL-1β (also 31 kDa), which is biologically inactive and requires proteolytic cleavage. Additional mechanisms involved in the neutrophil recruitment process include the neutrophil associated chemokine receptor CXCR2, which is activated in response to CXCL-1, -2, and -5 (23, 24). Neutrophils are a source of molecular mediators which contribute to the irreversible loss of lung function associated with chronic smoking and COPD, such as proteases, which can break down the extracellular matrix of the lung (25–27). Neutrophil accumulation is associated with the development of emphysematous lung destruction often observed in COPD patients (28). IL-1–mediated neutrophil accumulation is an important aspect of cigarette smoke–induced inflammation that may be exacerbated in the context of a subsequent bacterial infection.

In the current study, we sought to investigate the role of IL-1 signaling in cigarette smoke–induced neutrophil accumulation and investigate the role of IL-1 in the context of a bacterial exacerbation of cigarette smoke–induced inflammation. We found that cigarette smoke–induced neutrophilia was dependent on IL-1α produced by alveolar macrophages. In addition to stimulating the production IL-1α, cigarette smoke predisposed alveolar macrophages to produce exacerbated levels of IL-1α in response to bacterial stimuli. We then investigated the importance of a cigarette smoke–exposed lung environment primed for exacerbated IL-1α production in an in vivo model of NTHi infection. This model elucidated an exacerbated neutrophilic response to NTHi in smoke-exposed mice that was dependent on IL-1α. In addition, the important neutrophil-attracting IL-1–dependent chemokine was epithelial cell–derived CXCL5. These data demonstrate the important role of IL-1 signaling in the altered inflammatory response elicited in a smoke-exposed lung, providing mechanistic insight into this potentially pathogenic phenomenon.

Six- to 8-wk-old female BALB/c and C57BL/6 mice were purchased from Charles River Laboratories (Saint-Constant, PQ, Canada) and The Jackson Laboratory (Bar Harbor, ME). Mice deficient in CXCR2 on a BALB/c background and mice deficient in IL-1R1 on a C57BL/6 background were purchased from The Jackson Laboratory. IL-1α knockout (KO) mice and IL-1β KO mice were on a C57BL/6 background and bred in-house (29). Mice were housed under specific pathogen-free conditions with ad libitum access to food and water and subjected to a light–dark cycle of 12 h. All experiments were approved by the Animal Research Ethics Board at McMaster University.

A whole body cigarette smoke exposure system (SIU-48; Promech Lab AB [Vintrie, Sweden]) was used. Mice were exposed to 12 3R4F reference cigarettes (Tobacco and Health Research Institute, University of Kentucky, Lexington, KY) with filters removed, for 50 min, twice daily, for 4 or 5 d/wk for 8 wk. Details of the exposure protocol were reported previously (30). Control mice were exposed to room air only.

The nontypeable Haemophilus influenzae (NTHi) strain 11P6 (provided by Dr. S. Sethi, VA Medical Research, Buffalo, NY) was used. This clinical strain of NTHi was isolated from the sputum of a COPD patient during acute exacerbation. NTHi was initially grown on chocolate agar plates containing 1% isovitalex (BD Biosciences, Franklin Lakes, NJ). Colonies of NTHi were then grown to log phase in 10 ml brain–heart infusion (BHI) broth (Difco, Fisher Scientific, Ottawa, ON, Canada) supplemented with Hemin and NAD (Sigma-Aldrich, Oakville, ON, Canada). The inoculated brain–heart infusion + Hemin + NAD broth was maintained on a rotary shaker at 37°C until an OD value of 0.7–0.8 was obtained at a 600-nm wavelength. CFU was predicted from the OD value based on a previously generated standard curve. Before alveolar macrophage stimulation and infection of mice, NTHi was washed three times with PBS, resuspended, and diluted to a ratio of 10 CFU/cell (for alveolar macrophage culture) and 10⁶ CFU/35 μl (for mouse infection).

Prior to bronchoalveolar lavage (BAL), the right lung was tied off and homogenized in 2 ml PBS using a Polytron PT 2100 homogenizer (Kinematica, Switzerland) at 21,000–25,000 rpm for 3–6 s. BAL fluid was collected after instilling the lungs with 250 μl ice-cold PBS and then with 200 μl PBS. Total cell number in the BAL was determined using a hemocytometer. Cytospins were prepared and stained with Hema 3 (Biochemical Sciences, Swedesboro, NJ). Five hundred cells were counted per cytospin for determination of percent mononuclear cells (MNC) and percent neutrophils (NEU). Differential cell counts were calculated using this percentage and the total cell number (TCN).

BAL fluid was collected after instilling the whole lung with 1 ml of ice cold PBS. This process was repeated 5 times for maximal recovery of macrophages. BAL cells were resuspended in 500 μl fresh PBS. Alveolar macrophages were identified and counted using a hemocytometer. BAL cells were resuspended in RPMI 1640 medium supplemented with 10% FBS (Sigma-Aldrich, Oakville, ON, Canada), 1% l-glutamine, 1% penicillin/streptomycin, and 0.1% 2-ME (Invitrogen, Grand Island, NY) and cultured in polystyrene flat-bottom 96 well plates at 50,000 alveolar macrophages/well. Cells were incubated at 37°C and 5% CO₂ for 1 h to facilitate adherence. Non-adherent cells were removed by washing three times with warm PBS. Cells were cultured for 24 h with RPMI 1640 medium alone, in the presence of 1 μg/ml Pam₃CSK₄, 1 μg/ml LPS (InvivoGen, San Diego, CA), or 10 CFU/cell NTHi. Cell supernatants were collected and stored at −20°C for the measurement of TNF-α, IL-1α, and IL-1β protein levels by ELISA. In addition, RNA was isolated from alveolar macrophages after 4 h of LPS stimulation using the RNeasy mini Kit with optional DNase step (Qiagen, Mississauga, ON, Canada). cDNA was synthesized with SuperScript II Reverse Transcriptase (Invitrogen), and TaqMan real-time RT-quantitative PCR (qPCR) was carried out with the ABI PRISM 7900HT Sequence Detection System (Applied Biosystems, Foster City, CA). Target gene expression was normalized to that of the housekeeping gene GAPDH in the same sample and expressed as fold increase over the control unstimulated group.

IL-1α, IL-1β, CXCL1, CXCL2, CXCL5, and CXCL7 protein levels in lung homogenates were determined by ELISA (R&D Systems, Minneapolis, MN) as per the manufacturer’s instructions.

Following the BAL, lungs were fixed at 30 cm H₂O pressure in 10% formalin for histological assessment. After a minimum of 24 h formalin fixation, lungs were paraffin embedded, and 4-μm slices were cut. Tissues were treated for 45 min in 0.01 M citrate buffer to retrieve Ag and incubated for 1 h with either anti–IL-1α polyclonal goat Ab or anti-CXCL5 polyclonal rabbit Ab (R&D Systems) diluted in UltrAb diluent (Thermo Scientific, Rockford, IL) at 7 μg/ml. Immunohistochemistry was developed with goat on rodent HRP probe (Biocare, Concord, CA) or anti-rabbit Dakocytomation HRP (Dako Canada, Burlington, ON, Canada) and counterstained in Meyer’s solution.

Lung slices were generated as described previously (18). Briefly, lungs were inflated with 37°C agarose (Sigma-Aldrich) prepared to 2% in HBSS supplemented with HEPES (Sigma-Aldrich). The agarose was allowed to cool. Lung lobes were dissected and maintained in an ice-cold HBSS solution prior to and during slicing. Slices (120 μm thick)were generated using a vibratome (Leica Microsystems, Concord, ON, Canada). Approximately 40 slices were isolated from each mouse lung. RNA was isolated, and RT-qPCR was performed as described for alveolar macrophage cultures.

Forced oscillation measurements were performed using the FlexiVent ventilator system (SCIREQ, Montreal, QC, Canada). Mice were sedated and anesthetized with 10 mg/kg xylazine (Bayer Healthcare, Berlin, Germany) and 30 mg/kg sodium pentobarbital (Ceva, Lenexa, KS), both given by i.p. injection. Mice were then immobilized by i.p. injection of 10 mg/kg rocuronium bromide (Ω Laboratories, Montreal, QC, Canada), whereupon each mouse underwent tracheostomy using a blunted 18-gauge needle. Oxygen saturation and heart rate were continuously monitored using an infrared pulse oxymeter (Biox 3700; Ohmeda, Boulder, CO). The animal was next connected to a computer-controlled Flexivent ventilator. All mice were ventilated with 150 breaths/min, with an applied pressure limit of 30 cm H₂O. A snapshot perturbation maneuver was performed with forced oscillation perturbation consequently applied. Maximal pressure-regulated pressure-volume loops were finally generated to obtain maximal vital (total) lung capacity and static compliance. Following data collection, animals were removed from the ventilator and immediately euthanized via terminal exsanguination.

H&E-stained slides were generated from formalin-fixed paraffin-embedded lungs. Three lung slices were generated for each sample and used in the analysis. Each lung section was imaged in its entirety, and the airspace size was quantified by Pneumometrics Software (version 1) as described previously (31).

Data are expressed as mean ± SEM. Statistical analysis was carried out using SPSS Software (IBM, Armonk, NY). Levene’s test for equality of variances was used to adjust for differences in data variability between groups. The Univariate General Linear Model was used, and independent t tests were applied subsequently for two-group comparisons. Differences with p < 0.05 were considered statistically significant. Inflammatory cell count data for key experiments were pooled for from two independent experiments.

To model cigarette smoke-induced inflammation, C57BL/6 mice were cigarette smoke-exposed for 4 d. We have previously shown that 4 d of cigarette smoke exposure is sufficient to elicit lung neutrophilic inflammation (30). Cigarette smoke exposure led to a significant increase in IL-1α and IL-1β in the lung homogenates (Fig. 1A, 1B). IL-1α KO and IL-1β KO mice were exposed to cigarette smoke for 4 d, and despite both IL-1α and IL-1β being upregulated in wild-type animals, only IL-1α deficiency resulted in significantly reduced neutrophilia where the mean neutrophil counts was reduced from 1.080 × 10⁵ (± 2.028 × 10⁴ SEM) cells/ml in the WT mice to 1182 (± 988 SEM) cells/ml in the IL-1α KO mice (Fig. 1A, 1B). This reduction in neutrophil infiltration was associated with a decrease in the neutrophil recruiting chemokines CXCL1 and CXCL5 (Fig. 1A). Interestingly IL-1β deficiency had no effect of neutrophilia and resulted in less CXCL1 but not significantly less CXCL5, indicating that CXCL5 is a better biomarker of neutrophilic inflammation in this experimental system (Fig. 1B).

Immunohistochemistry has previously shown that alveolar macrophages stain positive for IL-1α in a 4 d smoke exposure protocol (14). To assess whether the critical source of IL-1α is a hematopoietic cell, we generated bone marrow chimeric mice with IL-1α KO and wild-type mice, as depicted in Fig. 1C. After 4 d of smoke exposure, mice deficient in IL-1α expression in their hematopoietic cells had an almost complete abrogation in pulmonary neutrophil infiltration (Fig. 1C). In contrast, when IL-1α was deficient in radioresistant structural cells, there was no significant reduction in neutrophilia. These data in conjunction with our previous findings provide strong evidence that the alveolar macrophage is crucial to the neutrophilic response to cigarette smoke through the production of IL-1α.

Because of our finding that IL-1α derived from alveolar macrophages drives smoke-induced neutrophilia, we sought to further investigate the phenotype of this cell within the context of cigarette smoke. In addition to the accumulation of neutrophils, cigarette smoke exposure led to the expansion of the monocyte/macrophage population in the lung lumen, as evident from the accumulation of these cells in BAL fluid (Fig. 2A). In addition to more monocytes/macrophages in the BAL, alveolar macrophages were observed to have an altered phenotype. The morphology of these cells changed in the smoke-exposed pulmonary environment, because smoke-exposed alveolar macrophages were larger, had an irregular shape, and contained large vesicle structures (Fig. 2B) (14).

In our current study, we demonstrate that alveolar macrophages isolated from cigarette smoke-exposed mice spontaneously produced IL-1α when placed in culture (Fig. 2C). Smoke-exposed macrophages also produced significantly more IL-1α in response to stimulation with the TLR2 and TLR4 ligands Pam₃CSK₄ and LPS as well as the live bacteria NTHi (Fig. 2C). The increased IL-1α was regulated at the transcriptional level as IL-1α mRNA was significantly increased in smoke-exposed macrophages placed in medium alone or following LPS stimulation (Fig. 2D). Although alveolar macrophages from cigarette smoke-exposed mice were primed to produce more IL-1α in response to bacterial stimuli, these same macrophages produced significantly less TNF-α in response to Pam₃CSK₄, LPS, and NTHi (Fig. 2E). IL-1β was measured in the supernatants of macrophages from smoke-exposed mice as well, and although IL-1β signal could be detected, the levels were below the limit of the assay to produce quantifiable results (Fig. 2F). These data suggest that cigarette smoke increases the number of monocytes/macrophages in the lung environment and primes alveolar macrophages to produce exaggerated levels of IL-1α in response to bacterial ligands.

We have established the importance of the IL-1α/IL-1R1 signaling axis in cigarette smoke–induced neutrophilic inflammation and that cigarette smoke primes alveolar macrophages to produce excessive amounts of IL-1α in response to bacteria. These data strongly imply a role for IL-1 signaling in the context of cigarette smoke and bacterial infection. To investigate this, we established a model where BALB/c mice were infected with NTHi after 8 wk of smoke exposure. Cigarette smoke–induced cellular inflammation was significantly increased by bacterial infection (Fig. 3A); the hallmark of this exacerbated inflammation was a significant increase in neutrophils.

CXCR2 ligands have previously been shown to be important in neutrophil recruitment in an animal model of cigarette smoke-induced inflammation, as a CXCR2 inhibitor attenuated neutrophilia (24). To assess the role of CXCR2 in bacterial exacerbation of cigarette smoke–induced inflammation, CXCR2 KO mice were exposed to 8 wk of cigarette smoke, followed by NTHi infection. Although the NTHi-exacerbated neutrophilia was not as dramatic in this experiment, it was still statistically significant and CXCR2 deficiency resulted in a significant reduction in the neutrophils observed in the BAL (Fig. 3B). To determine the CXCR2 ligands involved, we assessed CXCL1, CXCL2, and CXCL5 in BALB/c mice exposed to cigarette smoke, followed by NTHi infection. CXCL1 and CXCL2 were increased by NTHi infection in room air-exposed mice, and this increase was, interestingly, close to significantly attenuated in smoke-exposed mice (Fig. 3C for CXCL1; data not shown for CXCL2). Although we do not completely understand the reduced CXCL1 levels as neutrophilia increased, CXCL5 seems to be the crucial CXCR2 ligand as CXCL5 was the only CXCR2 ligand assessed that was both induced by NTHi and further increased in the smoke-exposed and NTHi-infected group of mice (Fig. 3C).

Our previous data suggest the importance of structural cells in neutrophil recruitment (18), such as epithelial cells. It would be an intuitive supposition that these epithelial cells could be a source of neutrophil recruiting chemokines such as CXCL5. Immunohistochemistry using an anti-CXCL5 Ab identified epithelial cells of the airways and alveoli as being a potential source of CXCL5 (Fig. 3D). Although some staining of alveolar macrophages was observed with CXCL5 immunohistochemistry, we were unable to measure CXCL5 in ex vivo cultures of macrophages from mice that were exposed to cigarette smoke for 8 wk and stimulated for 24 h with NTHi (data not shown). To demonstrate that, although the macrophage is not producing CXCL5, the lung tissue is primed for its production by cigarette smoke, CXCL5 was measured in precision cut lung slices, an ex vivo culture system where all lung cell types are present. Increased CXCL5 mRNA expression was detected in lung slices generated from 8 wk smoke-exposed mice compared with room air-exposed mice following stimulation with NTHi (Fig. 3E). These data suggest that the source of CXCL5 was the lung epithelium and that it is transcriptionally upregulated by smoke and NTHi. Our findings establish that NTHi exacerbates cigarette smoke–induced neutrophilia in a CXCR-2–dependent manner and that the most relevant CXCR2 ligand in this system is likely CXCL5.

After establishing the importance of IL-1 signaling in driving cigarette smoke–induced neutrophilia and characterizing a model of bacterial exacerbation of cigarette smoke–induced neutrophilia, we sought to investigate the role of IL-1 signaling in the smoke and NTHi experimental system with IL-1R1 KO mice. The exacerbated cellular inflammation elicited by cigarette smoke and NTHi was significantly attenuated in IL-1R1 KO mice (Fig. 4A). The attenuated inflammation was evident on the level of total cell number, monocytes/macrophages, and neutrophils. Of note, IL-1R1 deficiency did not impact the cellular profile in room air exposed mice following NTHi infection. These data strongly imply that neutrophil recruitment to the lungs observed in our smoke exposure model is dependent on IL-1 signaling and that this signaling axis is further activated upon NTHi challenge.

To study mechanisms of attenuated neutrophilia, we assessed CXCR2 chemokine expression in the BAL of IL-1R1 mice exposed to cigarette smoke, followed by NTHi infection (Fig. 4B). Although CXCL1 expression was increased, CXCL5 was significantly reduced in IL-1R1 KO mice. These observations demonstrate the importance of IL-1 signaling for the induction of CXCL5, which represents a biomarker of neutrophilia as both CXCL5 and neutrophilia were attenuated in IL-1R1 KO mice.

We have demonstrated that cigarette smoke exposure primes alveolar macrophages to produce exacerbated levels of IL-1α in response to bacteria and that IL-1 signaling is important in an in vivo model of cigarette smoke and bacterial inflammation. We subsequently sought to investigate the specific role of IL-1α in this in vivo system. In our bacterial exacerbation model, IL-1α levels were significantly increased by the combination of cigarette smoke and NTHi when compared with either stimulus alone (Fig. 5A). This differed from the observed levels of IL-1β in this model in which NTHi induced IL-1β, but this induction was not significantly greater in the smoke and NTHi samples (Fig. 5A). We next used immunohistochemistry to examine the source of IL-1α (Fig. 5B). The cell type that demonstrated the highest signal was the alveolar macrophage, consistent with our findings in Fig. 2. These findings suggest that IL-1α, and not IL-1β, drives the exacerbated response. To test this hypothesis, IL-1α and IL-1β KO mice were smoke-exposed and subsequently infected with NTHi, and cellular inflammation was assessed in BAL fluid. IL-1α deficiency led to a significant decrease in the total cells observed in the BAL, and this reduction was the result of decreases in both monocyte/macrophage and neutrophil recruitment to the lungs (Fig. 5C). This differed from the BAL cell counts in IL-1β KO mice, which did not elaborate a reduced inflammatory response (Fig. 5D). These data demonstrate a role for IL-1α in driving NTHi-exacerbated responses in vivo through the priming of alveolar macrophages.

To investigate the implications of disrupted IL-1 signaling on the clearance of bacteria from the lungs, bacterial burden was assessed in lung homogenates generated from 8 wk smoke-exposed, NTHi-infected mice. Bacterial burden was significantly increased in smoke-exposed mice that were deficient in IL-1R1 and CXCR2 (Fig. 6). Of note, bacterial burden was only increased in IL-1R1 KO mice that were cigarette smoke exposed. Bacterial burden was not significantly increased in IL-1α KO mice. To investigate the consequences of decreased bacterial clearance on chemokine expression CXCL1, CXCL2, and CXCL5 were measured in the BAL from the CXCR2 KO experiment (Supplemental Fig. 1). The increased NTHi burden was accompanied by a marked increase in these chemokines in CXCR2 KO mice. These data suggest that the increased cellular inflammation in cigarette smoke–exposed mice aids in the control of bacterial burden.

Chronic cigarette smoke–induced inflammation is believed to drive lung pathology associated with COPD, and specifically, IL-1–driven inflammatory mechanisms have been implicated in this process (6, 20). The current study describes a model of exacerbated IL-1–dependent inflammation in a model of smoke exposure. To test whether this exacerbated inflammatory response can induce measurable changes in lung physiology and pathology over time, a model of long-term cigarette smoke exposure and repeated NTHi infection was used (Fig. 7A). Using Flexivent technology, pressure-driven pressure volume curves were generated, and cigarette smoke–exposed lungs had a greater volume at lower pressures when compared with room air–exposed mice, and this increase was even greater in lungs from smoke-exposed mice that received repeated NTHi infections (Fig. 7B). For a statistical quantification, these data were used to calculate the compliance of the lungs (Fig. 7C). Histology slides also were generated from this experiment, and changes in the size of airspaces were quantified (Fig. 7D). The combined stimulus of smoke and NTHi led to a significantly more compliant lung than cigarette smoke and control PBS. Lung destruction was also enhanced by repeated NTHi administration as indicated by airspace enlargement. These data suggests that exacerbated inflammation in this model is accompanied by changes to the lung physiology and increased pathology, which supports previous observations that NTHi leads to greater levels of the collagen breakdown product hydroxyproline in smoke-exposed mice (14), indicative of extracellular matrix destruction.

Cumulatively, the data presented in this study demonstrate that cigarette smoke induces IL-1α production in alveolar macrophages and alters the phenotype of these macrophages such that they produce excessive levels of IL-1α in response to a bacterial stimulus resulting in exacerbated neutrophilia (Fig. 8). This excessive response may underlie a phenomenon where cigarette smoke–induced lung pathology is accelerated by bacterial infection.

Cigarette smoke’s impact on immune inflammatory processes elicited by pulmonary infection affects a large population of individuals worldwide. It is of particular concern to COPD patients as microbial-driven exacerbations are a significant health concern, yet the mechanisms of these disease exacerbations are not well understood. Most experimental work studies either cigarette smoke–induced inflammation or bacterial infection in isolation without extrapolating these results to experimental models that combine infectious agents and cigarette smoke. The objective of the current study was to investigate the cellular and molecular mechanisms that contribute to bacterial exacerbation of cigarette smoke–induced inflammation. This approach elucidated a novel mechanism by which cigarette smoke can elicit an exacerbated inflammatory response to bacteria.

To address the objectives of this study, we used a murine model of cigarette smoke exposure and intranasal infection with NTHi. Our smoke exposure system elicits neutrophilic inflammation. This was a central readout of our study because of the importance of chronic neutrophilia in the pathogenesis of COPD (32, 33). Our experimental system also reproduces other clinical hallmarks associated with cigarette smoking, such as airspace enlargement and changes in ventilation and perfusion following prolonged cigarette smoke exposure (31). The observation that carboxyhemoglobin and cotinine levels measured in our system are comparable to human smokers further validates this experimental system (30). To model an infection that is clinically relevant in the context of a COPD patient, the strain of NTHi chosen to be used in this study was isolated from a COPD patient experiencing an exacerbation (34). It is for these reasons that we feel this experimental approach can provide meaningful insight into the inflammatory pathways engaged by cigarette smoke and bacteria that will be relevant to a COPD patient experiencing an exacerbation because of a bacterial infection.

The murine model possesses some unique characteristics in terms of the differences between the mouse strains used. We have previously compared the inflammatory response between BALB/c and C57BL/6 mice (14, 30), and there are differences in the response of the different strains to cigarette smoke. BALB/c mice have a more neutrophilic response, whereas C57BL/6 mice have more monocyte recruitment. This trend was observed again in the current study, because the exacerbation of cigarette smoke induced cellular inflammation was predominantly neutrophilic in nature and less so in C57BL/6 mice. This highlights the disadvantage of using multiple strains; however, even though it was not as dramatically increased in C57BL/6 mice, NTHi-induced neutrophilia was still significantly exacerbated by cigarette smoke. Investigating the differences between the two strains may be of interest in future studies as it could help to identify what factors in an outbred population may predispose to exacerbated responses. Another strain difference observed in this study was the bacterial burden in smoke exposed mice. We have previously published the phenomenon of increased clearance of NTHi in a smoke-exposed lung because of an increase in NTHi-specific Abs induced by cigarette smoke (14, 35). In the current study. we observed this increased clearance for the BALB/c strain but not C57BL/6 mice. A possibly explanation for this is that we have changed the supplier of C57BL/6 mice since the initial study; however, we have not measured NTHi-specific Ab expression in these mice because this was not the focus of the current study. When interpreting results from experiments using a murine smoke exposure system, it is important to consider the strain of mice used.

The importance of IL-1 is increasingly being shown to be a major signaling factor in initiating immune inflammatory responses to cigarette smoke, but the biologically relevant source has not been definitively identified. In addition, there is evidence of the clinical relevance of IL-1 signaling as increased levels of both IL-1α and IL-1β have been observed in human smokers (18). In this study, we demonstrate that cigarette smoke–induced neutrophilia is IL-1α–dependent and redundant of IL-1β. This observation is supported by findings by us and others, showing an IL-1α dependency for neutrophil recruitment in response to cigarette smoke exposure (18, 36). We have previously reported that alveolar macrophages produce IL-1α in response to cigarette smoke based on observations from immunohistochemistry of lung samples (18); however, these findings did not provide any indication as to how important alveolar macrophage-derived IL-1α was in the process of neutrophilic inflammation. Using bone marrow chimeric mice, we demonstrate that hematopoietic cells are the critical source of IL-1α. The data presented in the current study are novel and conclusively demonstrates that the relevant source of IL-1α is a hematopoietic cell-type. This finding can be considered in conjunction with our previous histology data to conclude that neutrophilic inflammation elicited by cigarette smoke is dependent on IL-1α produced by a population of alveolar macrophages. It has previously been hypothesized that alveolar macrophages are orchestrators of inflammation within the context of COPD (17). A recent study has shown that depleting macrophages prevents lung damage and the loss of lung function elicited by cigarette smoke (37). The findings of our current study reinforce the crucial role of the alveolar macrophage in cigarette smoke–induced inflammation and identify IL-1α as the key signaling molecule for this cell type’s action. Specifically, our data suggest that cigarette smoke primes the lung, and specifically the alveolar macrophage, to exacerbated responses to bacterial stimuli (outlined in Fig. 8).

We have observed that cigarette smoke-induced IL-1α leads to the production of the neutrophil attracting chemokine CXCL5; however, the IL-1α/CXCL5 signaling axis is not unique to cigarette smoke-induced inflammation because this pathway is also involved in sterile inflammation (38). The importance of IL-1α in the inflammatory response to sterile tissue damage and cell death has been validated in multiple model systems (38–40). Similar to cigarette smoke, macrophages are the sensors of necrotic debris in models of sterile inflammation and are the crucial producers of IL-1α (41). Experimental models of cigarette smoke exposure have been known to induce apoptosis and necrosis (42, 43). It is currently unclear how the insult of cigarette smoke is sensed and the inflammatory response of cigarette smoke may be driven in part by the sensing of cigarette smoke-induced apoptosis/necrosis and the release of damage-associated molecular patterns. The findings of the current study suggest that cigarette smoke–induced inflammation may involve overlap with components of the sterile inflammatory response and that prolonged exposure to a sterile inflammatory stimulus may have implications for subsequent bacterial infection.

A key observation of this study is that cigarette conditions the lung environment to a fundamentally altered response. These findings are of critical importance because it suggests that molecular pathways engaged in a healthy, noninflamed tissue may be distinct to inflammatory pathways engaged in an inflamed tissue, such as the environment generated by cigarette smoke. Our findings suggest that smoke-exposed macrophages take a pragmatic approach and favor further activation of the IL-1α pathway upon challenge with a secondary stimulus that would normally engage different signaling and effector pathways. Molecularly, IL-1R1 signaling pathways were redundant in NTHi-induced inflammatory processes in room air–exposed control mice. Contrasting this observation, cigarette smoke exposure primed the lungs to an exaggerated inflammatory response that involved distinct molecular signaling pathways; IL-1R1 dependency was only observed in cigarette smoke–exposed mice. This represents a novel finding and suggests that mucosal immune responses are shaped by the immune inflammatory history of the tissue. This general phenomenon could explain the exacerbated inflammatory response observed in our experimental model. Taken together, our observations establish the critical importance of the lung mucosal environment in determining inflammatory responses engaged by environmental agents.

The findings of the current study provide some of the first mechanistic insight, to our knowledge, into the mediators that drive bacterial exacerbations of cigarette smoke–induced inflammation. These findings form the basis for future studies to dissect the relative importance of bacterial exacerbation in the pathology elicited by cigarette smoke because exacerbations are associated with a greater loss of lung function in COPD (44–46). Work has begun in this area, as a previous study demonstrated that repeated challenge with the viral stimulus polyinosinic:polycytidylic acid led to increased airspace enlargement (47). A recent study that uses cigarette smoke exposure and two administrations of heat-killed NTHi demonstrated that both stimuli were necessary to induce measurable increases in pathology as measured by airspace enlargement and an increase in lung compliance (48). This study characterized exacerbated inflammation in their model but did not examine potential mechanisms. Our study also uses the flexivent system to measure changes in lung physiology, which has been established as capable of detecting differences in a model of cigarette smoke exposure (37), in addition to scoring histology sections with pneumometric software. Here we add to these studies and show that bacterial infection can accelerate cigarette smoke–induced changes to lung physiology and pathology. Our focus on IL-1 signaling and neutrophilia synergizes with previous work that suggests these to be important in driving lung pathology in smoke-exposure models. Churg et al. (20) have shown that mice deficient in IL-1R1 have attenuated airspace enlargement after 6 mo of cigarette smoke exposure and also observed the IL-1 dependency of neutrophilia in the context of cigarette smoke. Our bacterial exacerbation model demonstrates that NTHi infection increases the levels of IL-1α and neutrophil numbers in the lung. We have provided the proof-of-principle data that shows that repeated exacerbation of these inflammatory mechanisms led to measurable differences in lung function and future studies will examine this in greater detail.

The increased inflammation observed in cigarette smoke-exposed mice following infection with NTHi may not be entirely detrimental to the host as there may be a role for the increased macrophage and neutrophil numbers in the control of bacterial burden. The data presented in Fig. 6 suggest that in the absence of IL-1R1 and CXCR2 bacterial clearance was compromised, but this increased bacterial burden was only observed in the smoke-exposed groups. It is possible that enhanced cellular inflammation is necessary to compensate for other aspects of host defense that are impaired by cigarette smoke. It has been established that cigarette smoke attenuates the phagocytic ability of macrophages, and this impairment has been specifically demonstrated in the context of NTHi (49). In addition to phagocytic activity, cigarette smoke also compromises the production of antimicrobial peptides and disrupts the integrity of the epithelial barrier. Exaggerated cellular recruitment may be necessary for bacterial clearance to compensate for these smoke-induced host defense deficiencies. In addition, the increased bacteria burden likely results in increased sensing of the bacteria, which could explain the increased levels of CXCL1 observed in smoke and NTHi=infected IL-1R1 KO mice and the increased CXCL1, CXCL2, and CXCL5 observed in CXCR2 KO mice. Disrupting the clearance mechanisms could be expected to lead to increased cytokine and chemokine production as the lung attempts to mount an unsuccessful inflammatory response against a stimulus that is not being eliminated and thus continues to propagate the inflammatory response.

Enhanced bacterial burden may seem like a barrier to the efficacy of targeting of IL-1 signaling for therapeutic purposes. This same phenomenon of enhanced bacterial burden was observed when corticosteroids were used as a treatment in our smoke exposure and NTHi infection model (14). In COPD patients, the treatment of an acute exacerbation of COPD with a steroid is recommended to be supplemented with antibiotic treatment whenever an infection is suspected (6, 50). The combined antibiotic and corticosteroid treatment is associated with better patient outcomes and a lower mortality (51). Our data would suggest that future therapeutic strategies that target IL-1 signaling may need to follow the recommended practices as corticosteroid treatment and be accompanied by antibiotic treatment to compensate for host defense mechanisms that are attenuated by cigarette smoke.

Ultimately, the exacerbated response to NTHi in a cigarette smoke–exposed lung appears to be beneficial for the control of bacteria while being detrimental to lung function. The data presented in Figs. 6 and 7 elucidated a biological response that is representative of a double-edged sword. When the increased leukocyte infiltration is attenuated bacterial burden is increased, whereas the increased inflammation associated with repeated administration of NTHi can enhance the damage elicited by cigarette smoke. Although additional experimentation is required to determine the exact role of the IL-1α/CXCL5 signaling axis in this loss of lung function, the necessity of this pathway in the control of bacteria must always be considered when targeting it therapeutically. This biological process might represent an appropriate response of a chronically inflamed lung to bacteria; however, there is still a desire to modify this process because increased inflammation is a key component in COPD exacerbations, which have been shown to accelerate the decline in lung function (52). These data indicate that a balanced approach of controlling the infectious agent while inhibiting the inflammatory response is necessary to mitigate the negative effects of bacterial exacerbations of cigarette smoke–induced inflammation.

The data presented in the current study expand the understanding of the role of IL-1 signaling in cigarette smoke–induced inflammation and elucidates the importance of this signaling axis in the context of a bacterial infection. The crucial source of IL-1α is a hematopoietic cell, reinforcing the perceived importance of the alveolar macrophage in orchestrating the inflammatory response elicited by cigarette smoke. In addition to driving inflammation, cigarette smoke changes the phenotype of alveolar macrophages, priming these cells to produce exacerbated IL-1α in response to bacterial stimulus. In vivo infection with NTHi results in greater IL-1α levels in smoke-exposed mice, and this was accompanied by IL-1R1–dependent CXCL5 production and exacerbated neutrophil recruitment. The current study highlights an important phenotypic change to a mucosal environment with important ramifications to host defense mechanisms in a model relevant to exacerbations of COPD.

We thank Joanna Kasinska for expert technical support, Marie Bailey for secretarial assistance, and Dr. Abraham Roos for discussion and preparation of the manuscript.

R.K. and A.A.H. are employees of MedImmune LLC.