Functional relevance of NLRP3 inflammasome-mediated interleukin (IL)-1β during acute allergic airway inflammation

Abstract

Overall asthmatic symptoms can be controlled with diverse therapeutic agents. However, certain symptomatic individuals remain at risk for serious morbidity and mortality, which prompts the identification of novel therapeutic targets and treatment strategies. Thus, using an adjuvant-free T helper type 2 (Th2) murine model, we have deciphered the role of interleukin (IL)-1 signalling during allergic airway inflammation (AAI). Because functional IL-1β depends on inflammasome activation we first studied asthmatic manifestations in specific inflammasome-deficient [NACHT, LRR and PYD domains-containing protein 3 (NLRP3^−/−) and apoptosis-associated speck-like protein containing a caspase recruitment domain (ASC^−/−)] and IL-1 receptor type 1^−/− (IL-1R1^−/−) mice on the BALB/c background. To verify the onset of disease we assessed cellular infiltration in the bronchial regions, lung pathology, airway hyperresponsiveness and ovalbumin (OVA)-specific immune responses. In the absence of NLRP3 inflammasome-mediated IL-1β release all symptoms of AAI were reduced, except OVA-specific immunoglobulin levels. To address whether manipulating IL-1 signalling reduced asthmatic development, we administered the IL-1R antagonist anakinra (Kineret®) during critical immunological time-points: sensitization or challenge. Amelioration of asthmatic symptoms was only observed when anakinra was administered during OVA challenge. Our findings indicate that blocking IL-1 signalling could be a potential complementary therapy for allergic airway inflammation.

Introduction

Over the last decade, the importance of inflammasome-mediated release of interleukin (IL)-1β has been shown in a variety of models, and moreover has provided the basis for the treatment of several autoinflammatory diseases such as rheumatoid arthritis, gout and type 2 diabetes [1–3]. In addition, several studies have demonstrated that IL-1β plays an important role during bronchial asthma development. Indeed, asthmatic patients were shown to harbour elevated levels of IL-1β in their bronchoalveloar lavage (BAL) [4]. Other studies have further reported that this proinflammatory cytokine promotes the infiltration of eosinophils and inflammation within the lung and encourages both mast and T cell activation [5,6]. Currently, 300 million people are estimated to be affected by asthma worldwide [Global Initiative for Asthma (GINA) 2010, Bethesda, MD, USA] and the prevalence continues to increase. Although current therapeutic strategies, such as corticosteroids and long- or short-acting beta agonists, can control attacks and disease progression, a quarter of a million people still succumbed to the syndrome in 2009 (GINA 2009) [7,8]. Asthma is characterized by dominant T helper type 2 (Th2) immune responses, including enhanced IL-4, IL-5 and IL-13 responses, allergen-specific immunoglobulin production, eosinophilia, airway inflammation, bronchoconstriction and airway hyperresponsiveness (AHR) [9–11]. Thus, when assessing the onset of asthma in murine models it is imperative to address both cellular and pathological aspects.

The secretion of functional bioactive IL-1β, however, depends upon the inflammasome which mediates, via caspase-1 activation, the cleavage of the inactive cytokine precursor (pro-IL-1β) into the active form (IL-1β). In general, this multi-protein complex contains proteins such as central oligomerization domain (NACHT), a leucine-rich repeat (LRR) domain, apoptosis-associated speck-like protein containing caspase recruitment domain (ASC) and caspase-1 [12–14]. For instance, the well-studied NACHT, LRR and PYD domains-containing protein 3 (NLRP3) inflammasome plays an essential role in a variety of scenarios and is triggered by numerous pathogens such as parasites [15], fungi [16], bacteria [17] or danger signals such as adenosine triphosphate (ATP) [18] or crystalline silica and asbestos [19,20]. However, its role during allergic asthma remains somewhat controversial. For example, IL-1 receptor type 1 (IL-1R1)-deficient mice present reduced allergic responses and lung inflammation [21,22] and ovalbumin (OVA)-treated NLRP3^−/− mice exhibit severely dampened lung inflammation, OVA-specific immunoglobulins, cytokine responses and eosinophil infiltration [23]. In contrast to the latter, Allen et al. demonstrated that allergic airway inflammation within NLRP3-deficient mice is similar to wild-type (WT) control mice [24]. Both these NLRP3 inflammasome studies were performed using C57BL/6 mice, which are renowned for their dominant Th1-based immune responses and reduced allergic airway inflammation when compared to Th2-biased BALB/c mice [25–28]. Therefore, we have assessed the role of the NLRP3 inflammasome and functional IL-1β during OVA-induced allergic airway inflammation (AAI) in a BALB/c murine model and observed altered antigen-specific Th responses, reduced eosinophil infiltration and airway inflammation in mice deficient for NLRP3, ASC and the IL-1R1. To analyse the dynamics of functional IL-1β in more detail, we blocked IL-1 signalling using the IL-1R antagonist (IL-1Ra) anakinra. The application of this therapy ameliorated allergic airway inflammation only during the challenge phase, and implies that blocking IL-1 signalling could complement current asthma treatment regimens.

Materials and methods

Mice

BALB/c and C57BL/6 WT mice were purchased from Harlan® (Borchen, Germany). ASC-, NLRP3- and IL-1R1-deficient mice on a BALB/c background (back-crossed at least nine generations) were a kind gift from Professor Jürg Tschopp (University of Lausanne, Switzerland). Mice were bred under specific pathogen-free conditions in the animal facilities at the Institute of Medical Microbiology, Immunology and Hygiene (Munich, Germany) in accordance with national and European Union guidelines 86/809. Experimental mice were sex- and age-matched and the study was approved by the Regierung von Oberbayern, Munich, Germany (Animal Licence number Az. 55·2.1·54–2532-67-12).

OVA-induced allergic airway inflammation model and anakinra treatment

Female BALB/c WT, ASC-, NLRP3- and IL-1R1-deficient mice were thrice sensitized subcutaneously (s.c.) in the neck with 10 μg ovalbumin (OVA) grade VI (Sigma-Aldrich, Taufkirchen, Germany) or phosphate-buffered saline (PBS) (control mice) on days 0, 7 and 14 without adjuvant and challenged consecutively by aerosol inhalation on days 26–28 with 10 μg OVA grade V (Sigma-Aldrich). On day 31 (3 days after the last OVA challenge) mice were killed and assessed for allergic airway development (Supporting information, Fig. S1a). For IL-1-type cytokine blocking experiments, mice were treated with consecutive doses of anakinra (Kineret®; Biovitrum, Stockholm, Sweden) during OVA-sensitization or challenge. For details, see schemes in Supporting information, Fig. S1b and S1c. Experiments were performed using either 15 mg/kg (i.p), 100 mg/kg [intraperitoneally (i.p.)] or 150 mg/kg (s.c.) in the neck per day.

Measurement of airway hyperresponsiveness (AHR)

Airway hyperresponsiveness (AHR) to methacholine (Sigma-Aldrich) was determined using the Flexivent system (Scireq, Montreal, Canada). Following anaesthesia the trachea was intubated with a 1·2 mm tracheal cannula and the lungs ventilated mechanically at a respiratory frequency of 150 breaths per min, a tidal volume of 10 ml/kg and a positive end-expiratory pressure of 3 ml H₂O. After exposing mice to aerosolized PBS to retrieve the baseline value, bronchoconstriction was induced using increasing concentrations (1, 2·5, 5, 10, 25 and 50 mg/ml in PBS) of aerosolized methacholine using an ultrasonic nebulizer. Dynamic resistance was recorded over 1-min intervals (every 5 s) after exposure to defined doses of methacholine via a standardized inhalation manoeuvre (SnapShot-150) [29,30].

Analysis of bronchoalveolar lavage

Mice were euthanized with Narcoren® (Merial, Halbergmoos, Germany) and lungs were flushed twice with 1 ml PBS containing proteinase inhibitor cocktail tablets (Roche Diagnostics, Mannheim, Germany). The obtained BAL was weighed and centrifuged at 230 g for 5 min (4°C). Resulting supernatants were frozen at −20°C and the cytokine content determined using commercially available enzyme-linked immunosorbent assay (ELISA) kits (eBiosciences, Frankfurt, Germany). The cell pellets were resuspended in PBS containing 2% fetal calf serum (FCS) and counted. To analyse the composition of immune cells within the BAL, 150 μl of cell suspension was centrifuged on glass slides at 400 rpm for 5 min using the Shandon Cytospin 3 centrifuge (Thermo Scientific®, Hamburg, Germany). Glass slides were dried overnight at room temperature (RT) and stained using the Diff-Quick staining set (Medion Diagnostics, Langen, Germany) according to the manufacturer's instructions. Cell differentiation was performed as described previously [29,30].

OVA-specific immune responses

Erythrocyte-depleted mediastinal lymph node (LLN) cells (2 × 10⁵) from OVA- or PBS-treated mice were co-cultured in the presence or absence of 20 μg/ml OVA grade VI at 37°C in RPMI-1640 medium (PAA, Linz, Austria) containing 5% FCS, penicillin/streptomycin, sodium-pyruvate, non-essential amino acids and β-mercaptoethanol (all PAA). After 72 h culture supernatants were analysed for cytokine levels by ELISA.

Cytokine and OVA-specific immunoglobulin determination by ELISA

To decipher in-situ cytokine levels, weighed lung samples were placed in 500 μl of RPMI-1640 medium (without supplements) and homogenized using the T10 basic Ultra-Turrax® disperser (IKA, Staufen, Germany). Samples were then centrifuged at 16 000 g for 10 min (4°C) and the resulting supernatant frozen at −20°C. The cytokine content was then determined by ELISA. OVA-specific immunoglobulin (Ig)E and IgG1 levels were measured in the sera of individual mice [30]. In brief, 96-well ELISA plates (Nunc, Langenselbold, Germany) were coated overnight (4°C) with either 1 μg (for IgE) or 0·1 μg (for IgG1) OVA grade V diluted in 50 mM Tris (Roth®, Karlsruhe, Germany) solution containing 3% bovine serum albumin (BSA) (PAA; blocking buffer). After washing and blocking, sera was diluted in blocking buffer (1:200–1:100 000 dilutions) and standards of mouse α-OVA IgE or IgG1 antibodies (Biozol, Eching, Germany) were applied in twofold serial dilutions and incubated overnight (4°C). Subsequently, plates were washed and α-mouse IgE or IgG1 biotinylated detection antibodies (Biozol) were applied and incubated for 2 h (RT). After further washing, streptavidin–horseradish peroxidase (HRP) conjugate (R&D Systems GmbH, Wiesbaden, Germany) was added and plates were incubated for 30 min (RT) in the dark. After a final washing step, BD OptEIA™ TMB substrate (BD, Heidelberg, Germany) was applied and reactions were stopped with 2 M H₂SO₄. Finally, optical densities (ODs) were determined at 450 nm using the Sunrise™ ELISA microplate reader (Tecan, Crailsheim, Germany). The concentration of the samples was then calculated according to the standard curve.

Histochemistry and evaluation of lung inflammation

Paraffin-embedded sections (3 μm) from the left lungs of individual mice were stained with PAS (periodic acid-Schiff), which allows the detection of goblet cells in lung basilar membranes. Sections were analysed microscopically for tissue inflammation and goblet cell hyperplasia according to previously described methods [29,30]. In short, tissue inflammation was determined by the degree of visual thickness of the basal membrane which was graded on a scale from 0 to 3 (inflammation score). To determine goblet cell hyperplasia, goblet cells within the basal membrane were counted and results are presented as the percentage of goblet cells per length of basal membrane.

Dendritic cell generation and in-vitro inflammasome assays

Bone marrow-derived dendritic cells (BMDC) from WT C57BL/6, WT BALB/c and ASC^−/− BALB/c mice were generated using 10 μg/ml granulocyte–macrophage colony-stimulating factor (GM-CSF) (Peprotech, Rocky Hill, NJ, USA), as described previously [15]. To assess IL-1β activation, 1 × 10⁵ BMDC were stimulated in RPMI medium containing 10% FCS and supplements (PAA) for 6 h with lipopolysaccharide (LPS) (5ng/ml; Invivogen, San Diego, CA, USA) and thereafter with ATP (5 mM; Roth) for 1 h. IL-1β in the culture supernatant was measured by ELISA according to the manufacturer's guidelines (eBiosciences).

Statistical analysis

Statistical differences were analysed using GraphPad Prism 5 software (San Diego, CA, USA). Parametrically distributed data were analysed using unpaired t-tests or one-way analysis of variance (anova). Parametric data are represented as mean ± standard deviation.

Results

Lack of NLRP3 inflammasome reduces the development of allergic airway inflammation

Because the majority of inflammasome activation studies have been performed in C57BL/6 mice, we first investigated whether BMDCs generated from BALB/c mice were able to produce IL-1β in vitro using a standard assay [15]. As shown in the Supporting information, Fig. S2a, BMDC derived from C57BL/6 mice produced high amounts of IL-1β following LPS priming and ATP activation. BALB/c BMDC were also able to secrete sufficient amounts of IL-1β which were significantly reduced using ASC-deficient BMDC, confirming that BALB/c mice possess functional inflammasome activation [31,32]. To determine the requirement of the NLRP3 inflammasome on the development of AAI, groups of WT, ASC- and NLRP3-deficient BALB/c mice were sensitized with PBS or OVA (see Methods and Supporting information, Fig. S1a). The three rounds of sensitization were administered s.c. in the absence of an adjuvant such as aluminium hydroxide (Alum), as it is known that Alum can induce the NLRP3 inflammasome and thus IL-1β secretion [33]. In addition, Conrad et al. demonstrated that this adjuvant-free OVA model can induce AAI in a similar manner to models using adjuvants [29]. In the model applied here, mice were challenged with aerosolic OVA over 3 consecutive days and analysed 72 h thereafter (Supporting information, Fig. S1a). A primary parameter of AAI is the influx of immune cells into the BAL, and the induction of asthmatic symptoms can be observed clearly when the cellular content in OVA-WT mice is compared to that found in PBS-WT groups (Fig. 1a–c). At first glance, levels of leucocyte infiltration were comparable between OVA-treated WT and inflammasome-deficient strains (Fig. 1a). However, upon further analysis, both knock-out strains showed a significant reduction in the percentage of eosinophils (Fig. 1b) and total numbers of eosinophils were reduced significantly in ASC-deficient mice (Supporting information, Fig. S3a). Inflammasome-deficient mice had significantly higher numbers of macrophages (Fig. 1c and Supporting information, Fig. S3b). In contrast, no neutrophil infiltration could be observed in either WT or inflammasome-deficient mice (Supporting information, Fig. S3c). In the control groups of mice, no differences were observed in the aforementioned parameters (Fig. 1a–c and Supporting information, Fig. S3a–c). In terms of AHR, sensitized ASC^−/− and NLRP3^−/− mice showed a significant reduction in airway resistance (Fig. 1d).

Fig. 1.

NLRP3 inflammasome influences eosinophil infiltration and airway hyperresponsiveness. Groups of wild-type (WT) [ovalbumin (OVA), n = 27; phosphate-buffered saline (PBS), n = 11], ASC^−/− (OVA, n = 15; PBS, n = 9) or NLRP3^−/− (OVA, n = 20; PBS, n = 6) mice were sensitized with either PBS or OVA (days 0, 7, 14) and subsequently challenged with OVA (days 26–28) to induce allergic airway inflammation. On day 31, (a) total leucocytes and the percentage of (b) eosinophils and (c) macrophages were determined within the bronchoalveolar lavage (BAL). (a–c) Symbols show mean ± standard deviation (s.d.) of individual mice from three independent experiments and asterisks show significant differences between the indicated brackets (***P < 0·001). (d) Airway resistance in response to methacholine was evaluated in OVA-treated WT (n = 7), ASC^−/− (n = 5) and NLRP3^−/− (n = 7) or PBS-treated mice (n = 14). Symbols represent the mean ± s.d. of pooled data and significant differences (P < 0·05) were detected between WT groups and either (*) NLRP3^−/− or (#) ASC^−/− mice.

Lung inflammation and goblet cell hyperplasia are reduced in NLRP3^−/− and ASC^−/− mice

To verify the finding that NLRP3 inflammasome-deficient mice had reduced experimental asthma, lung sections were analysed for the degree of inflammation and goblet cell hyperplasia (bright red colouring) [29,30]. When compared to lung sections of PBS-sensitized mice (Fig. 2a–c), strong inflammation could be observed clearly in sections from OVA-sensitized and challenged WT mice (Fig. 2d), and this was visibly reduced in ASC^−/− (Fig. 2e) and NLPR3^−/− mice (Fig. 2f). To quantify these impressions, the extent of inflammation (Fig. 2g) and the amount of goblet cell hyperplasia (Fig. 2h) were measured in each individual mouse. When compared to the pathology found in groups of OVA-WT mice, both parameters were reduced significantly in ASC^−/− strains (Fig. 2g,h). NLRP3^−/− mice showed a significant reduction in the inflammation score (Fig. 2g), but only a mild reduction in the percentage of goblet cells (Fig. 2h). These data confirm that the absence of inflammasome components dampens OVA-induced AAI.

Fig. 2.

Severity of lung inflammation depends on NLRP3 inflammasome activation. Lung sections from individual (a,d) wild-type (WT), (b,e) ASC^−/− or (c,f) NLRP3^−/− mice were cut (3 μm) and stained with periodic acid-Schiff (PAS). Sections were assessed microscopically (×20 magnification) to calculate (g) the inflammation score using a severity score from 0 to 3 and (h) goblet cell count per basal membrane length. (a–f) PAS-stained lung sections are representative examples from the different groups of mice. (g–h) Results show combined data from three independent experiments per mouse strain (see Fig. 1 legend). Bars show mean ± standard deviation and asterisks show significant differences between the indicated brackets (***P < 0·001 and *P < 0·05).

NLRP3 inflammasome activation alters Th cytokine responses

OVA-induced allergic airway inflammation is associated with strong OVA-specific Th responses. To assess whether the diminished airway responses in inflammasome-deficient mice were reflected in their Th profiles, various cytokines were measured in the BAL and within cell-culture supernatants from mediastinal lymph node (LLN) cells that were restimulated ex vivo with OVA. Fig. 3a shows that levels of IL-5 and IL-13 but not IFN-γ in the BAL were down-regulated significantly in both ASC^−/− and NLRP3^−/− mice. However, OVA-stimulated cells from the local draining lymph node of inflammasome-deficient mice secreted significantly less IFN-γ but not IL-5, when compared to WT animals. In contrast, IL-10 responses were moderately elevated (non-significantly), indicating possible elevations in regulatory processes in ASC- and NLRP3-deficient mice (Fig. 3b). IL-10 levels within the BAL and IL-13 levels of OVA-stimulated lymph node cells were not altered in inflammasome-deficient mice (data not shown). Interestingly, no IL-17 responses could be detected within the BAL- or OVA-stimulated lymph node cells (data not shown). Differential regulation was not observed in OVA-specific B cell responses, as levels of IgE and IgG1 in the sera were comparable in WT and inflammasome-deficient mice (Fig. 3c).

Fig. 3.

NLRP3 inflammasome activation alters cytokine but not ovalbumin (OVA)-specific immunoglobulin responses. (a) Levels of interleukin (IL)-5, IL-13 and interferon (IFN)-γ were measured within the bronchoalveolar lavage (BAL). (b) Lower limit of the normal range (LLN) cells (2 × 10⁵) from individual mice were stimulated with or without 20 μg/ml OVA for 72 h. Thereafter, levels of IFN-γ, IL-5 and IL-10 were detected in the cell-culture supernatants. (c) OVA-specific immunoglobulin (Ig)E and IgG1 levels were measured in sera from individual mice. (a–c) Symbols and bars show mean ± standard deviation of individually assessed mice from three independent experiments (see Fig. 1 legend). Asterisks show significant differences between the indicated brackets (***P < 0·001; **P < 0·01; *P < 0·05).

Reduced eosinophilia but not lung inflammation and function in the absence of the IL-1R1

Levels of in-situ IL-1β, normalized to individual lung weight, were determined in the lungs of OVA-exposed WT and inflammasome-deficient mice and, surprisingly, no differences were observed on the final day of analysis (Fig. 4a). However, directly after the last OVA exposure (day 28), significantly higher levels of IL-1β were determined in the WT group (Fig. 4b). This suggested that IL-1β plays a critical role during the primary inflammation stages of asthma but is rapidly surpassed by other immune mechanisms. In contrast, IL-18 levels were slightly higher in the WT group on day 28 but significantly higher on day 31 (Supporting information, Fig. S2b and S2c, respectively). To further decipher the role of functional IL-1β, we studied AAI development in mice deficient for the IL-1 receptor type 1 (Fig. 4c–i). Within the BAL, no differences were observed between the two groups of mice in terms of total leucocyte numbers (Fig. 4c). However, OVA-sensitized IL-1R1^−/− mice demonstrated significantly lower levels of eosinophils (Fig. 4d and Supporting information, Fig. S3d), whereas macrophage numbers were comparable to WT mice (Supporting information, Fig. S3e). Again, very few neutrophils could be detected in the BAL (Supporting information, Fig. S3f). With regard to pathology, IL-1R1^−/− mice showed no significant reduction in either inflammation score (Fig. 4e) or goblet cell influx (Fig. 4f). Interestingly, PBS-treated IL-1R1^−/− mice already presented enhanced lung inflammation and goblet cell hyperplasia when compared to naive WT mice (Fig. 4e,f). Moreover, when compared to PBS groups of WT mice, the inflammation in OVA-treated WT groups was enhanced significantly, but this was not the case in IL-1R1^−/− mice (Fig. 4e). Airway resistance levels in PBS-treated IL-1R1^−/− mice were almost equal to those resulting from OVA-treated IL-1R1^−/− mice and both groups were higher than in OVA-treated WT mice (Fig. 4g). Interestingly, levels of IL-18 were equal within BAL of asthmatic WT and IL-1R1-deficient mice (Supporting information, Fig. S2d). Levels of IL-13 and IFN-γ within the BAL were increased in IL-1R1-deficient mice, whereas levels of IL-5 were significantly lower (Fig. 4h). As with all other inflammasome-deficient animals, OVA-specific immunoglobulin levels remained unchanged (Fig. 4i). These data show that functional IL-1β plays a role during AAI and appears to be important for the initiation of immune responses and maintaining the correct balance of cytokines and the cellular composition within the lung.

Fig. 4.

Functional interleukin (IL)-1β influences allergic airway inflammation. In-situ IL-1β levels within lungs of wild-type (WT), ASC^−/− and NLRP3^−/− were measured on (a) day 31 or (b) day 28 [directly after last ovalbumin (OVA)-exposure]. OVA-treated WT (n = 11) and IL-1R1^−/− (n = 15) mice were assessed for (c) total leucocytes, (d) percentage of eosinophils, (e) inflammation score, (f) goblet cell hyperplasia, (h) cytokine responses within the bronchoalveolar lavage (BAL)and (i) OVA-specific immunoglobulin (Ig)E and IgG1 levels in sera. (g) Airway resistance in response to methacholine was evaluated in OVA-treated WT (n = 10) and IL-1R1^−/− (n = 10) or phosphate-buffered saline (PBS)-treated WT (n = 5) and IL-1R1^−/− (n = 5) mice. Results show data from three (a,g) and two (b–f,h–i) independent experiments. (a–f,h–i) Symbols and bars show mean ± standard deviation of individual mice and asterisks show significant differences between the brackets (***P < 0·001; **P < 0·01; *P < 0·05). (g) Significant differences (*P < 0·05) were detected between OVA- and PBS-treated WT groups of mice.

Administration of anakinra during OVA-challenge reduces the development of AAI

Finally, we aimed to decipher whether bioactive IL-1β was critical during the sensitization phase or challenge phase of AAI. IL-1 signalling can be blocked through the administration of drugs, such as the endogenous antagonist IL-1Ra anakinra (Kineret®). Indeed, the use of such drugs has been shown to ameliorate IL-1β-mediated diseases such as rheumatoid arthritis or type 2 diabetes [1,3]. Thus, BALB/c mice were treated with different doses of anakinra during either the sensitization (anakinra_Sens) or challenge (anakinra_Chall) phases of AAI (see Supporting information, Fig. S1b and S1c, respectively). Treatment doses of anakinra were either 15 or 100 mg/kg i.p. or 150 mg/kg s.c., as the administration of anakinra in patients is usually applied in this manner. These data are depicted, respectively, in the Supporting information, Figs. S4 and S5 and Fig. 5. Interestingly, mice injected with any of the employed doses of anakinra during challenge had significantly reduced eosinophil numbers in the BAL, but this was not observed in mice treated during sensitization (Fig. 5a and Supporting information, Figs. S3g, S4a and S5a). When compared to doses of 15 or 100 mg/kg i.p. (Supporting information Figs. S4b and S4c, S5b and S5c), the increased dose of anakinra via the s.c. route led to significantly reduced disease parameters such as lung inflammation and goblet cell hyperplasia (Fig. 5b,c). With regard to immune profiles, levels of IL-13 and IFN-γ, but not IL-5, were strongly down-regulated in the anakinra_Chall group but not the anakinra_Sens group (Fig. 5d, cf. filled squares and grey triangles). These modulations to OVA-specific responses were only partially observed using lower doses of anakinra (Supporting information, Figs. S4d and S5d). Again, levels of OVA-specific IgG1 and IgE were not altered in any of the applied experimental scenarios (Fig. 5e and Supporting information, Figs S4e and S5e). Moreover, no differences in macrophage levels and, again, no neutrophil infiltration could be detected (Supporting information, Figs. S3h and S3i). However, because anakinra treatment during the challenge phase significantly reduced several AAI parameters, we also measured the effects of anakinra on airway hyperresponsiveness. Interestingly, when compared to OVA groups, airway resistance was reduced significantly in anakinra-treated OVA groups (c.f. closed squares and closed triangles in Fig. 5f). Indeed, levels in the anakinra-treated group were comparable to control groups of mice [cf. PBS (open squares) and OVA + anakinra (closed triangles)]. Thus, in contrast to the data sets using IL-1R1-deficient mice, which showed no significant differences in lung pathology and function, the data presented here reveal that specific blocking of IL-1 signalling during the acute challenge phase leads to reduced inflammatory responses against OVA. Moreover, as the PBS IL-1R1 knock-out group showed elevated AHR and moderately increased inflammation we demonstrate that a permanent loss of IL-1 signalling has a fundamental effect in the steady state. In conclusion, the data reveal that specific blocking of IL-1 signalling during the acute challenge phase leads to reduced inflammatory responses against OVA and opens up an avenue for exploration into possible therapeutic strategies.

Fig. 5.

Subcutaneous anakinra administration suppresses acute allergic airway inflammation. Groups of phosphate-buffered saline (PBS)- or ovalbumin (OVA)-treated wild-type (WT) mice were treated with 150 mg/kg (s.c.) anakinra (Kineret®) during either sensitization (+ anakinra_Sens) or challenge (+ anakinra_Chall) phases. (a) Eosinophil infiltration, (b) inflammation score, (c) goblet cell hyperplasia, (d) cytokine responses [bronchoalveolar lavage (BAL)] and (e) OVA-specific immunoglobulin (Ig)E and IgG1 levels in sera were analysed on day 31. PBS control groups, n = 4 (− anakinra), n = 2 (+ anakinra_Sens) and n = 2 (+ anakinra_Chall) and OVA groups were n = 5 (− anakinra), n = 7 (+ anakinra_Sens) and n = 10 (+ anakinra_Chall). (f) Airway resistance in response to methacholine was evaluated in OVA-treated mice (n = 6), PBS control mice (n = 5), PBS control with anakinra (n = 3) and anakinra-treated OVA groups (n = 8). Anakinra was administered during the challenge phase. Symbols and bars show the mean ± standard deviation of individual mice from two independent experiments. Asterisks show significant differences (analysis of variance or Student's t-test) between the brackets (***P < 0·001; **P < 0·01; *P < 0·05).

Discussion

In vitro, several agents have been shown to trigger the assembly of the NLRP3 inflammasome and the consequent release of functional IL-1β [13,15,19,20]. However, the dysregulation of IL-1β can elicit severe consequences, such as the rare inherited inflammasome-mediated pathophysiology of autoinflammatory diseases, including Muckle–Wells syndrome and familial cold autoinflammatory syndrome [34,35]. In both these cases, the genetic defect lies in multiple point mutations in the NACHT domain of the NLRP3 protein, which results in permanent activation of the NLRP3 inflammasome [36,37]. These conditions are classified as cryopyrin-associated periodic syndromes (CAPS), which are a subset of autoinflammatory disorders [1,38], and are distinguished from autoimmune diseases because they fail to induce antigen-specific T cell responses and high antibody titres [35]. Research correlating NLRP3 inflammasome activation and the pathogenesis of these disorders has exploded during the last decade, and further evidence shows a fundamental role in the development of other diseases, including cancer [1–3,38]. With regard to the latter, murine models of chronic pulmonary fibrotic disorders using silica and asbestos have correlated IL-1β and the onset of inflammation-induced lung cancer [19,39].

The list of diseases that have been treated successfully with the IL-1Ra anakinra (Kineret®), a non-glycosylated recombinant form of the naturally occurring IL-1Ra, encompasses not only autoinflammatory diseases such as Sweet syndrome [40] and relapsing polychondritis [41], but illnesses such as gout and type 2 diabetes [1–3]. Within those studies, anakinra demonstrated high efficacy and no adverse side effects. The data presented here display that the administration of anakinra to OVA-allergic mice ameliorates disease parameters such as eosinophil infiltration and lung inflammation. We blocked IL-1 signalling during either the sensitization or challenge phases, and determined that blocking IL-1 signalling during the challenge phase produced the most beneficial results. In accordance with this, a recent study demonstrated that administration of IL1-alpha only during the challenge phase (in contrast to the sensitization phase) had disease aggravating effects [42]. Although studies have demonstrated that the IL-1 family members IL-18 and IL-33 contribute to allergic airway inflammation [43,44], anakinra specifically blocks IL-1β and IL-1α, and therefore indicates that the observed effects here are independent of these cytokines. However, further studies should be performed to identify whether simultaneous blocking of several IL-1 family members during the challenge phase, such as IL-18 and IL-33, is more effective. Although anakinra has a good safety record, it requires daily injections due to its short half-life (4–6 h). Alternative agents such as rilonacept (Regeneron®; IL-1 trap) and canakinumab (Ilaris®; anti-IL-1β antibody), which have much longer half-lives and specifically block IL-1β, are currently being assessed in drug trials [45,46]. Thus, it will be interesting to decipher whether these alternative treatments or new upcoming drugs dampen allergic responses in man.

As mentioned above, the beneficial effects of anakinra were observed upon application during the challenge phase. In studies comparing inflammasome-deficient and WT animals, we were initially puzzled that the levels of IL-1β within the lungs were comparable on the final day of analysis (day 31). However, further investigation revealed that in WT mice, levels were significantly higher directly after the last exposure to OVA (day 28). These data indicated that the first inflammatory response depended on NLRP3 inflammasome activation that led subsequently to the release of functional IL-1β. This hypothesis was verified to some extent using IL-1R1-deficient mice, as these mice also displayed reduced eosinophil infiltration and levels of IL-5 in the BAL. However, they also demonstrated elevated levels of IL-13 and IFN-γ, implying that additional factors play a role in the induction of local Th2 and Th1 responses within the IL-1R1-deficient mice. Another intriguing aspect of these studies was observed in naive IL-1R1 knock-out mice, as it appeared that a continuous loss of functional IL-1 signalling in the steady state had a more fundamental effect because deficient animals showed elevated basal levels of inflammation, goblet cell numbers and, moreover, showed higher airway resistance than OVA-treated WT mice. Such effects were not observed in PBS-sensitized mice that were treated with anakinra, demonstrating that selective blocking of IL-1 signalling for a certain period of time does not influence cellular composition and immune responses per se. In correlation, Th1 and Th2 levels in the BAL were reduced in anakinra-treated (challenge) and inflammasome-deficient asthmatic mice, confirming that NLRP3 inflammasome-mediated IL-1β responses are crucial for the induction of local inflammation within the lung. Besides the functional relevance of IL-1β, we show here that IL-18 is important during AAI. Indeed, levels of IL-18 were reduced significantly in ASC- and NLPR3-deficient strains, whereas levels were comparable between WT and IL-1R1^−/− mice. Because IL-18 is an important Th1-inducing proinflammatory cytokine [43], this may explain the reduced IFN-γ levels observed in asthmatic ASC^−/− and NLPR3^−/− mice but not IL-1R1^−/− mice. With regard to OVA-specific responses of restimulated lung lymph node cells, only IFN-γ responses were dampened in inflammasome-deficient mice and, interestingly, IL-10 levels were actually elevated, albeit not significantly, demonstrating that inflammasome activation influences Th immune responses [15]. IL-17 responses were also measured in BAL and antigen-specific culture assays but no differences between the groups could be observed and, moreover, levels were extremely low in any of the assessed experiments (data not shown). In addition, almost no IL-17A and IL-17F mRNA expression within asthmatic lungs of WT and ASC-deficient mice could be determined using quantitative reverse transcription–polymerase chain reaction (RT–PCR) (data not shown). These results are in contrast to previous findings, as studies have reported that alongside other cytokines, IL-1β induces Th17 immune responses [47–49] that can influence allergic asthma development [50,51]. Moreover, several studies also reported that IL-17 favours neutrophil recruitment leading to the induction of neutrophilia and lung inflammation [52,53]. Interestingly, we could hardly detect neutrophils within the BAL, and the frequency of those cells was extremely low compared to other studies [22,23,29]. This might be a possible explanation as to why no IL-17 responses could be detected in any of our currently described experiments. However, the lack of neutrophils within the lungs of asthmatic BALB/c mice and the total absence of IL-17 responses remains unclear. Further studies should investigate the role of neutrophils in correlation with IL-17 responses within allergic airway diseases of BALB/c mice.

The data presented here demonstrate that NLRP3 inflammasome activation is essential for the initiation of inflammatory responses within the lung and the modulation of Th2 immune responses, but is dispensable for immunoglobulin production, as no differences were observed in levels of OVA-specific IgG1 and IgE. In contrast, Besnard et al. demonstrated that OVA-specific IgE levels were strictly dependent upon NLRP3 inflammasome activation, as well as all the other asthmatic parameters [23]. However, this study was performed with C57BL/6 mice which, in contrast to BALB/c mice, develop less pronounced AAI and have dominant Th1-based immune responses [25–28]. Moreover, Allen et al. showed that allergic airway inflammation within NLRP3^−/− mice with a C57BL/6 background is similar to WT control mice [24]. These contrasting findings demonstrate the complexity of allergic asthma development and highlight the necessity to employ further distinct methods to analyse the role of cytokines in addition to complete knock-out mouse models. Overall, the mechanisms of inflammasome activation during AAI still remain unclear, although the data shown here indicate that the NLRP3 inflammasome may play a larger role in controlling AHR and airway remodelling rather than inflammation in AAI. Nevertheless, it has been reported that ATP levels are increased in asthmatic patients [54] and in the BAL of OVA-treated mice [50]. Interestingly, mice that lack the functional purinergic P2-receptor P2X7, which recognizes extracellular ATP, displayed reduced features of acute and chronic asthma [55], and epithelial damage has been shown to release uric acid and ATP which provokes NLRP3 inflammasome activation [49,56,57]. Thus, the possible release of damage-associated molecular patterns (DAMPs) (i.e. ATP) during allergen exposure could be responsible for NLRP3 inflammasome activation during AAI. Collectively, the data presented within this work demonstrate that NLRP3 inflammasome activation and IL-1β release play a crucial role in the pathogenesis of experimental allergic asthma. Moreover, we show that administration of anakinra during the acute challenge phase ameliorates allergic airway inflammation and suggests that blocking IL-1 signalling could provide the basis for novel treatment strategies.

Disclosure

The authors declare no commercial or financial conflicts of interest.

Author contributions

M. R., K. S., S. S., S. H., H. G. and L. E. L. performed the experiments. M. R., L. E. L., D. B., H. G. and C. P. da C. designed the experiments and provided essential expertise. M. R., C. P. da C. and L. E. L. wrote the manuscript.

Supporting information

Additional Supporting information may be found in the online version of this article at the publisher's web-site: