Complement is implicated in asthma pathogenesis, but its mechanism of action in this disease remains incompletely understood. In this study, we investigated the role of properdin (P), a positive alternative pathway complement regulator, in allergen-induced airway inflammation. Allergen challenge stimulated P release into the airways of asthmatic patients, and P levels positively correlated with proinflammatory cytokines in human bronchoalveolar lavage (BAL). High levels of P were also detected in the BAL of OVA-sensitized and challenged but not naive mice. Compared with wild-type (WT) mice, P-deficient (P−/−) mice had markedly reduced total and eosinophil cell counts in BAL and significantly attenuated airway hyperresponsiveness to methacholine. Ab blocking of P at both sensitization and challenge phases or at challenge phase alone, but not at sensitization phase alone, reduced airway inflammation. Conversely, intranasal reconstitution of P to P−/− mice at the challenge phase restored airway inflammation to wild-type levels. Notably, C3a levels in the BAL of OVA-challenged P−/− mice were significantly lower than in wild-type mice, and intranasal coadministration of an anti-C3a mAb with P to P−/− mice prevented restoration of airway inflammation. These results show that P plays a key role in allergen-induced airway inflammation and represents a potential therapeutic target for human asthma.
Allergic asthma is a chronic inflammatory disease characterized by pulmonary eosinophilia, increased serum IgE levels, airway hyperresponsiveness (AHR), mucous hypersecretion, and structural remodeling of airways (1). Current treatment for asthma such as systemic corticosteroids and inhaled β agonists are far from optimal, and there is a need to develop novel therapeutic approaches, particularly for those patients with severe asthma (2). Asthma is believed to be driven by inappropriate Th2-dominated immune responses to environmental allergens (3–6). In addition, the role of innate immunity in asthma has also attracted interest of investigators (7–10). Complement is a key component of the host innate immunity, and previous clinical and laboratory studies have suggested an important role of the complement system in this disease (11–20). A better understanding of the activation mechanisms and effectors of complement in asthma would aid the development of novel anticomplement therapies.
Properdin (P) is a plasma glycoprotein and the only known positive regulator of the complement cascade (21). It binds to and stabilizes the alternative pathway (AP) C3 convertase C3bBb (21, 22) and, in some cases, may also serve as a platform to form new C3bBb convertases on the cell surface (23). P is best known as a promoter of AP complement activation in the context of host defense, and its deficiency in humans increases the risk for Neisseria meningitidis infection (24–27). There is considerable evidence to suggest that P may also play a critical role in AP complement-mediated tissue injury, for example, in the setting of ischemia reperfusion injury or inflammatory joint destruction (28, 29). In contrast, P deficiency or inhibition in a murine model of fH-related C3 glomerulopathy exacerbated glomerular disease (30), suggesting that the role of P in AP complement-mediated diseases may be complex and potentially context specific. Previous studies have found the AP complement and anaphylatoxin receptors to be involved in murine models of asthma, but whether and how P might play a role in this disease is not known. In this study, we tested the hypothesis that P contributes to the pathogenesis of allergen-induced airway inflammation and that targeting P dampens Th2 and Th17 immune responses. Our study provides proof of concept for therapeutic targeting of P in allergic asthma.
Materials and Methods
Human patient samples
All subjects gave their informed consent, and the study was approved by the Institutional Review Board of the Thomas Jefferson Medical College. Deidentified bronchoalveolar lavage (BAL) samples were obtained from study subjects as described before (31). In brief, healthy subjects without asthma and subjects with allergic asthma and rhinitis were recruited for the study and screened to assess suitability. Screening consisted of medical history and physical examination, followed by skin testing for allergy to common aeroallergens. All subjects were nonsmokers and were not taking any chronic medications. Asthmatics met the National Institutes of Health/National Heart, Lung, and Blood Institute expert panel criteria for the diagnosis of asthma, and the diagnosis was confirmed by spirometry and responsiveness to methacholine (32). In an effort to reduce variability, only a single allergen was used, ragweed Ag E (Amb a I), and patients were studied outside of ragweed season. The concentration of the lung-delivered dose of ragweed Ag was determined by serial intradermal skin testing and was 100-fold higher than that required to cause a minimum positive skin wheal, based on our established protocol (31). In brief, on day 1, the subject presented between the hours of 7:00 and 9:00 am and underwent BAL with 150 ml saline in 50-ml aliquots in a lingular segment. This was immediately followed by Ag instillation into a right middle lobe segmental bronchus. For safety reasons, a 10-fold test dose preceded instillation of the full challenge dose. Both test and challenge volumes were 5 ml. On day 2, the challenged segment was lavaged in the same way as on day 1. For this study, paired BAL samples from individuals before and after allergen challenge were available from asthmatic patients only.
Wild-type (WT) C57BL/6 mice were obtained from The Jackson Laboratory; P−/− mice with B6/129J mixed background were generated by gene targeting as previously described (33). Homozygous P−/− mice were screened from pups from heterozygotes breeding pairs; WT littermates from the same breeding pairs were used as controls. Mice were used at 6–8 wk of age and housed in a specific pathogen-free facility. All animal experiments were approved by the Institutional Animal Care and Use Committee.
Regents and Abs
Chicken OVA (grade V) was obtained from Sigma-Aldrich; the adjuvant aluminum hydroxide (Alum imject) was purchased from Pierce; Abs and protein standards for mouse IL-4 (capture: purified rat anti-mouse IL-4, clone 11B11, 554434; detection: biotinylated rat anti-mouse IL-4, clone BVD6-24G2, 554390; standard: recombinant mouse IL-4, 550067), IL-5 (capture: purified rat anti-mouse IL-5, clone TRFK5, 554393; detection: biotinylated rat anti mouse IL-5, clone TRFK4, 554397; standard: recombinant mouse IL-5, 554581), IL-17A (capture: purified rat anti-mouse IL-17A, clone TC11-18H10, 555068; detection: biotinylated rat anti-mouse IL-17A, clone TC11-8H4, 555067), INF-γ (capture: purified rat anti-mouse INF-γ, clone R4-6A2, 551216; detection: biotinylated rat anti-mouse INF-γ, clone XMG1.2, 554410; standard: recombinant mouse INF-γ, 554587), IgE (capture: purified rat anti-mouse IgE, clone R35-92, 553416; detection: biotinylated rat anti-mouse IgE, clone R35-72, 553414; standard: purified mouse IgE κ Isotype control, 557079), C5a (capture: purified rat anti-mouse C5a, clone I52-1486, 558027; detection: biotinylated-rat anti-mouse C5a, clone I52-278, 558028; standard: purified mouse C5a, 622597), and C3a (capture: purified rat anti-mouse C3a, clone I87-1162, 558250; detection: biotin-rat anti-mouse C3a, clone I87-419, 558251; standard: purified mouse C3a, 558618) ELISA detection and streptavidin HRP (554066) were purchased from BD Biosciences; recombinant mouse IL-17A protein purchased from eBioscience (14-8171-62) served as mouse IL-17A ELISA standard, mouse anti-human P and anti-mouse P mAbs and control mAbs; MOPC-31C were produced in-house (28). A neutralizing rat anti-mouse C3a mAb, clone 3/11, was purchased from Hycult Biotech. Human C3a ELISA Kit, 550499 was from BD Biosciences; human IL-4 (D4050) and IL-5 (D5000B) Quantikine ELISA kits were purchased from R&D Systems; human IL-13 ELISA Kit (EHIL13) was from Thermo Scientific.
Mouse allergic asthma
A murine allergic asthma model was developed as described before (34, 35); mice were sensitized with 10 μg OVA mixed with 2 mg alum by i.p. injection on days 0 and 14. From day 21, they were challenged daily for 5 consecutive days with 30-min exposure to aerosolized OVA (1% w/v) in a closed chamber. Twenty-four hours after the last challenge, mice were sacrificed for analysis (Fig. 1C). Mice that were sensitized and challenged with vehicle (PBS) served as negative controls (“naive” mice). To block endogenous P at both sensitization and challenge phases, we administrated a mouse anti-mouse P mAb (28) to WT mice (50 μg/g body weight/mouse, i.p.) on days −1, 6, 13, and 20 of the experimental protocol (Fig. 3A). To block P at sensitization phase alone, we administrated the anti-P mAb to WT mice on days −1, 6, and 13 (Fig. 4A). To block P at challenge phase alone, we administrated a single dose of the anti-P mAb to WT mice 1 d before OVA challenge on day 20 (Fig. 5A). To block P locally in the airways, we intranasally instilled 30 μg/mouse anti-P mAb in 20 μl PBS to WT mice immediately before each OVA challenge (Fig. 6A). For reconstitution of P in P−/− mice at the challenge phase, 10 μg/mouse recombinant P in 20 μl PBS was delivered intranasally to P−/− mice immediately before each OVA challenge (Fig. 7A). P−/− mice receiving recombinant P premixed with the anti-P mAb were used as controls. To evaluate the role of C3a, we included a rat anti-mouse C3a mAb (5 μg/mouse) with recombinant P in some of the reconstitution experiments (Fig. 8C).
Collection of BAL and cell counts
After serum collection, mice were sacrificed and lungs were lavaged as previously described (34). In brief, the left lung was ligated, trachea was cannulated, and the right lung was lavaged with 500 μl PBS three times. Cell counting was performed on cytospin preparations stained with May–Grünwald–Giemsa (Merck), and at least 200 cells were classified in blinded samples by an independent investigator using standard morphologic criteria. BAL was stored at −80°C until analysis. Cytokines and C3a/C5a concentration in BAL were determined by ELISA using commercial Abs and standards (BD Biosciences and eBioscience) according to the product’s instruction.
Measurement of P concentration in BAL and serum
Mouse P concentrations in BAL and serum were measured using sandwich ELISA. Plates were coated with the anti-P mAb clone 14E1 (28). BAL was diluted with PBS containing 1% BSA, and serum was diluted with PBS containing 1% BSA and 20 mM EDTA. Recombinant mouse P was used as a standard. Plate-bound P was detected by biotinylated anti-mouse P mAb clone 5A6 generated in-house. To measure P in human BAL samples, we used two in-house–generated mouse anti-human P mAbs, clone 8.1 and 24.2, as capture and detection Abs, respectively. Human P purified from plasma was used as a standard to generate standard curves.
Preparation and culture of BAL and lung cells
Naive, OVA-sensitized and -challenged, and OVA-challenged but not presensitized mice were sacrificed; lungs were lavaged with 1 ml PBS three times; and infiltrated cells in BAL were collected by centrifuge and counted. After BAL collection, whole lungs were then dissected, minced by razor blade, and enzymatically digested for 1 h at 37°C in PBS containing 2 mg/ml collagenase A (Roche) and 40 U/ml DNase I (Roche). Lung digests were then filtered (70-μm cell strainer; BD Falcon) to collect total lung cells, and the latter were washed twice in PBS + 2% FBS. BAL and total lung cells were cultured with RPMI-1640 medium containing 10% FBS for 2 d in the presence or absence of OVA (50 ng/ml) at a density of 106 cells/ml, and the supernatant was collected for mouse P detection using ELISA.
Immediately after collection of lavage fluid, lungs were incised and treated with ice-cold 10% formalin before processing for histology. Paraffin sections were cut at 5-μm thickness, mounted on positively charged slides, and stained with H&E or periodic acid–Schiff reagent (PAS). PAS+ airways of lung were quantitated by counting both PAS+ and PAS− airways using light microscopy for a total of three lung sections per animal, and the percentage of PAS+ airways was calculated based on airways counted on all sections in a given experimental group.
Serum IgE detection
Serum was separated by centrifugation and stored at −80°C until analysis. Total IgE levels were determined using ELISA. Purified rat anti-mouse IgE capture mAb (R35-72; BD Biosciences) was used to coat the plates. Purified mouse IgE (C38-2; BD Biosciences) served as a standard, and biotinylated rat anti-mouse IgE (R35-92; BD Biosciences) was used for IgE detection.
Mediastinal lymph node cell culture and stimulation
After mice were sacrificed, mediastinal lymph node (MLN) cells were isolated, prepared as single-cell suspension, and seeded on 96-well plates. Cells were restimulated with OVA (50 ng/ml) or vehicle (PBS) for 72 h, and the supernatant was collected for cytokines detection using ELISA.
Determination of AHR
Twenty-four hours after the last OVA challenge, mice were anesthetized and intubated for measurement of lung resistance (RL) as described before (36). In brief, mice were challenged with increasing concentrations of nebulized methacholine (0, 12.5, 25, 50, and 100 mg/ml) for 10 s (60 breaths/min, 0.5 ml tidal volume) and were mechanically ventilated by 160 breaths/min, tidal volume of 0.15 ml, positive end-expiratory pressure of 2–4 cm H2O. The peak value for RL was measured before and after inhalation of PBS and after each concentration of methacholine, and cumulative concentration-response curves were constructed. Responses are presented as increases in RL above baseline.
Assay of serum AP complement activity in vitro
Serum AP complement activity of anti-P mAb–treated mice was determined using an LPS-dependent ELISA assay as previously described (33). In brief, 96-well plates were precoated with LPS. Serum samples were serially diluted with gelatin–veronal buffer containing EGTA and Mg2+, and incubated on the plates for 1 h at 37°C. After the reaction was stopped by EDTA, a HRP-conjugated goat anti-mouse C3 Ab (55557; MP Biomedicals) was used to detect surface-bound activated C3b.
For experiments containing two groups, difference between the groups was calculated using unpaired t test. For data with nonparametric distributions, Mann–Whitney U test was used. For experiments containing three or more groups, ANOVA was used to determine the levels of difference and Bonferroni was used to perform the posttest of all pairs of data. The p values <0.05 were considered to be significant. Data are expressed as mean ± SEM.
P is released into the BAL of allergen-challenged asthmatic patients and experimental mice
In both humans and mice, P is primarily produced and released by leukocytes (37–40). To evaluate whether P might be secreted by inflammatory cells infiltrating the asthmatic airways where it promotes AP complement activation, we measured P levels in BAL samples of seven individuals with allergic asthma and rhinitis collected before and 24 h after an allergen challenge. As shown in Fig. 1A, although the absolute level of P varied considerably between patients, it was consistently higher in BAL collected after allergen challenge in these individuals. Furthermore, P levels correlated significantly with that of IL-5, IL-4, and IL-13 in postallergen-challenged BAL (Fig. 1B), suggesting that P release into BAL is a significant marker of the allergic response in patients. To further study this phenomenon and experimentally test the role of P in asthma pathogenesis and therapy, we induced airway inflammation in mice by chicken OVA sensitization and airway challenge. Mice were immunized (days 0 and 14) with OVA and alum, and challenged on 5 consecutive days (days 21–25) by exposure to aerosolized OVA followed by examination of airway inflammation and immune responses on day 26 (Fig. 1C). Consistent with data from asthmatic patients, we found a dramatic 6-fold elevation of P level in the BAL of OVA-sensitized and -challenged mice compared with naive mice (Fig. 1C). As expected, no P was detected in BAL samples of either naive or experimental P knockout (P−/−) mice.
To investigate the mechanism of P secretion into BAL, we prepared BAL cells and total lung cells from naive, OVA-sensitized and -challenged, and OVA-challenged but not presensitized mice and cultured them for 2 d with or without OVA stimulation. Examination of the cell culture medium revealed that high levels of P were secreted by BAL and total lung cells from mice sensitized and challenged with OVA, irrespective of whether the cells in culture were stimulated with OVA (Supplemental Fig. 1A). Furthermore, the level of P and number of inflammatory cells in BAL of mice sensitized and challenged with OVA were closely correlated, and each was increased with successive OVA challenges (Supplemental Fig. 1B, 1C). These data suggested that P secretion into BAL was related to airway inflammation arising from allergen sensitization and challenge rather than from nonspecific allergen challenge alone, and that inflammatory cells in the lung were a major source of P detected in BAL.
P deficiency or systemic inhibition in mice alleviates allergen-induced airway inflammation and Th2 immune responses
To determine whether P contributes to allergen-induced airway inflammation and immune responses, we subjected WT and P−/− mice to OVA sensitization and challenge and compared their phenotypes. We assessed airway inflammation and tissue damage by counting the number of inflammatory cells in BAL and by examining lung histology. As shown in Fig. 2A and 2B, we detected significantly fewer numbers of total inflammatory cells, eosinophils, and lymphocytes in P−/− mouse BAL, as well as less pulmonary inflammation and mucous secretion in the mutant mouse lungs by H&E and PAS staining, respectively. Notably, P deficiency did not significantly affect BAL macrophage and neutrophil numbers (Fig. 2A) or serum IgE levels (Fig. 2C). We also evaluated AHR to methacholine challenge of the experimental mice and found P−/− mice to demonstrate significantly reduced AHR compared with WT mice (Fig. 2D).
Because allergic asthma is a typical Th2-driven disease and recent studies have also implicated Th17 cells in the disease (41–43), we next investigated whether P deficiency might affect OVA-specific T cell immune responses. We collected the MLNs from OVA-sensitized and -challenged mice, prepared single-cell suspensions, and then restimulated the cells with OVA in culture for 3 d. By ELISA measurement of cytokines in the cell culture medium, we found that restimulated P−/− mouse lymphocytes produced dramatically reduced amounts of IL-4, IL-5, and IL-17A, but much higher IFN-γ (Fig. 2E). Separately, we measured IL-4, IL-5, IL-17A, and IFN-γ levels in the BAL but did not observe any differences between the two groups of mice (Fig. 2F). Thus, attenuated airway inflammation and AHR in P−/− mice were accompanied by a suppressed Th2 and Th17 immune responses by lymph node cells but were not associated with cytokine changes in the BAL.
To confirm the earlier data from P−/− mice, we blocked P function in WT mice with a mouse anti-mouse P mAb. Based on previously determined pharmacodynamics of the mAb (28), we treated WT mice weekly with either the anti-P or a control mAb (50 μg/g body weight), starting at 1 d (day −1) before OVA immunization (Fig. 3A). The efficacy of the anti-P mAb was confirmed by an LPS-dependent ELISA assay which showed a single injection to effectively suppress serum AP complement activity in WT mice for up to a week (Fig. 3B). Thus, weekly anti-P mAb treatment was expected to continuously block plasma P function in WT mice during the 26-d experimental period (Fig. 3A). Furthermore, the level of immunoreactive P in the BAL of anti-P mAb–treated mice was also diminished (Fig. 3C). Thus, both systemic (in plasma) and local (in lung) P function were blocked under this experimental protocol. We found that compared with control mAb-treated animals, WT mice treated with anti-P mAb displayed less airway eosinophilia (Fig. 3D) and less pulmonary inflammation and mucous secretion in the lung (Fig. 3E); their MLN cells produced lower amounts of the Th2 cytokines IL-4 and IL-5, but higher amount of IFN-γ after OVA restimulation in vitro (Fig. 3F). As in P−/− mice, we observed no difference in serum IgE (Fig. 3G) or BAL cytokines levels (Fig. 3H) in anti-P mAb–treated mice compared with control mAb-treated mice. Thus, blocking P function during allergen sensitization and challenge phases recapitulated the phenotype observed in P−/− mice.
Systemic inhibition of P during the challenge but not the sensitization phase reduces allergen-induced airway inflammation and Th2 and Th17 immune responses
Previous studies of the C5a/C5aR pathway have shown that complement mediators may play different roles in the allergen sensitization and challenge phases of asthma pathogenesis (15, 44, 45). To determine the stage at which P contributed to OVA-induced airway inflammation, we blocked P function in WT mice with anti-P mAb at different time points of the experimental protocol. To inhibit P function during OVA sensitization only, we treated WT mice three times at weekly intervals with anti-P mAb (Fig. 4A). Functional assay confirmed that serum P activity was blocked in the treated mice, but it recovered by the time of first OVA challenge on day 21, and the level of immunoreactive P in the BAL of OVA-challenged mice was high (data not shown). Interestingly, we found no difference in BAL inflammatory cell counts, serum IgE levels, pulmonary inflammation and mucous secretion, BAL cytokine levels, or IL-4, IL-5, and IL-17A production by restimulated MLN cells between anti-P mAb–treated and control mAb-treated mice (Fig. 4B–F). In contrast, IFN-γ production by restimulated MLNs was significantly increased in anti-P mAb–treated mice (Fig. 4E).
To address the role of P during the challenge phase of OVA-induced airway inflammation, we treated OVA-sensitized WT mice with anti-P mAb 1 d before OVA challenge (Fig. 5A). This single i.p. anti-P mAb injection reduced the level of immunoreactive P in BAL to naive mouse level (Fig. 5B). Intriguingly, blocking P function before OVA challenge alone was sufficient to alleviate airway inflammation and injury, and reduce IL-4, IL-5, and IL-17A production by restimulated MLN cells (Fig. 5C–E). It also significantly decreased serum IgE levels (Fig. 5F), an unexpected outcome considering that no change in serum IgE levels was observed in either P−/− mice (Fig. 2) or WT mice continuously treated with anti-P mAb during disease induction (Fig. 3). Blocking P function during OVA challenge did not significantly alter BAL cytokine levels (Fig. 5G), nor did it affect IFN-γ production by restimulated MLN cells (Fig. 5E). The latter result contrasted with the finding of increased IFN-γ production by restimulated MLN cells when P was inhibited during the sensitization phase (Fig. 4E).
Local inhibition of P during allergen challenge alleviates airway inflammation
Because P was found to be released abundantly into BAL in OVA-challenged mice and blocking P systemically at the challenge phase reduced airway inflammation, we investigated whether local inhibition of P in the airways of WT mice before OVA challenge might be effective in reducing lung inflammation. We sensitized WT mice by OVA immunization as usual, and before each daily OVA challenge administered anti-P or a control mAb by intranasal instillation (Fig. 6A). As shown in Fig. 6B and 6C, intranasal delivery of the anti-P mAb markedly diminished immunoreactive P level in BAL but had no impact on serum P in the treated mice. Furthermore, it significantly reduced airway inflammation, AHR, and pulmonary inflammation and mucous production (Fig. 6D–F), as well as decreased IL-17A and increased IFN-γ production by restimulated MLN cells (Fig. 6G). In contrast, local inhibition of P by intranasal route had no effect on serum IgE (Fig. 6H) and BAL cytokine levels except IL-5 (Fig. 6I), nor did it significantly alter IL-4 and IL-5 production by restimulated MLN cells (Fig. 6G). Thus, local inhibition of P in the airways of presensitized mice was sufficient and potentially therapeutic in alleviating lung inflammation and AHR upon allergen re-exposure.
Airway reconstitution of P in P−/− mice restores sensitivity to lung inflammation and injury
To confirm the pathogenic effect of P in the airways, we performed the reverse experiment, that is, by reconstituting OVA-sensitized P−/− mice with recombinant P locally in the airways through intranasal instillation before each OVA challenge (Fig. 7A). To ensure that any effect of the treatment was attributable to P reconstitution per se rather than introduction of potential impurities in the recombinant P preparation, we intranasally reconstituted a second group of P−/− mice with the same recombinant P that had been premixed with anti-P mAb. As shown in Fig. 7B and 7C, we found that intranasal administration of P, but not P pretreated with the anti-P mAb, to P−/− mice before each OVA challenge restored airway inflammation and mucous secretion to a similar or more severe degree than that observed in WT mice. Likewise, IL-4, IL-5, and IL-17A production by restimulated MLN cells was significantly higher in P−/− mice reconstituted with P alone than in mice reconstituted with P mixed with anti-P mAb (Fig. 7D). Notably, airway reconstitution of P at the challenge phase did not alter IFN-γ production by restimulated MLN cells (Fig. 7D) but reduced IFN-γ level in BAL (Fig. 7E). Other BAL cytokines including IL-4, IL-5, and IL-17A showed no significant difference between the groups (Fig. 7E). It is also interesting that although P deficiency did not affect serum IgE levels, intranasal instillation of P at the OVA challenge phase markedly elevated serum IgE (Fig. 7F). By ELISA assay, we confirmed that intranasal P reconstitution led to detection of abundant P in BAL but not in sera of P−/− mice (Fig. 7G). Thus, intranasally reconstituted P must have acted locally in the airways of the mutant mice to cause the observed phenotype changes.
P-dependent C3a production contributes to airway inflammation at the allergen challenge phase
Consistent with previous studies showing a pathogenic role of the complement C3a/C3aR pathway in murine models of asthma (11, 13, 14, 46, 47), we observed a positive correlation between C3a levels and total inflammatory cell counts in WT mouse BAL (Fig. 8A) and between BAL C3a and P levels in allergen-challenged human asthmatic patients (Supplemental Fig. 2A), as well as between human C3a and Th2 cytokines in asthma patient’s BAL (Supplemental Fig. 2B). Furthermore, of several complement gene mutant mouse strains (C3aR, C3, C4, fD) tested in the present model, we found C3aR gene deletion to have the largest protective effect on airway inflammation (Supplemental Fig. 3 and data not shown). Unlike C3a levels, C5a levels in BAL of WT asthmatic mice are not strongly correlated with infiltrated inflammatory cell counts (Supplemental Fig. 4A). Although, compared with that of asthmatic WT mice, there was a trend of reduced C5a level in the BAL of asthmatic P-deficient mice and this reduction in BAL C5a was reversed by intranasal reconstitution of P, the change in BAL C5a levels, in each case, was not statistically significant (Supplemental Fig. 4B). Based on these observations, we hypothesized that as a positive regulator of the AP complement activation, P may have contributed to airway inflammation and injury by promoting local C3a production in the lung. Indeed, ELISA assay revealed a significant reduction in C3a levels in the BAL of experimental P−/− mice compared with WT mice (Fig. 8B). To further evaluate the intermediacy of C3a in the mechanism of action of P in OVA-induced airway inflammation, we coadministered a control or anti-C3a mAb in the P reconstitution experiment of P−/− mice (Fig. 8C). As shown in Fig. 8, although intranasal reconstitution of P together with a control mAb restored sensitivity to airway inflammation and mucous secretion in P−/− mice (Fig. 8D, 8E), and elevated serum IgE levels (Fig. 8F), concurrent intranasal administration of an anti-C3a mAb significantly blunted the effect of P reconstitution (Fig. 8D–F). Anti-C3a treatment completely inhibited the enhancing effect of P reconstitution on IL-17A production, and partially but significantly on IL-5 production, by OVA restimulated MLN cells but did not alter IL-4 or IFN-γ production by restimulated MLN cells (Fig. 8G), nor did it have an effect on BAL cytokine levels (Fig. 8H).Thus, P contributed to OVA-induced airway inflammation in the effector phase of lung disease by promoting complement activation, and its effect, in several respects, was mediated by the C3a/C3aR pathway.
We describe in this article that P, the only known positive regulator of complement activation, plays a significant pathogenic role in allergen-induced airway inflammation and immune responses. We showed that P is released into the BAL of human asthmatic patients in response to airway exposure of allergens, and its level correlated with key Th2 cytokines. Similarly, we detected abundant P in the BAL of mice subjected to OVA-induced experimental asthma. These findings suggested that in presensitized human individuals and mice, P might have been released by inflammatory cells infiltrating the airway during allergen challenge, a hypothesis in line with current understanding of P biosynthesis by leukocytes in humans and mice (37–40). Indeed, in a separate experiment, we showed that BAL cells or total lung cells from OVA-immunized and -challenged mice, but not from naive mice or mice challenged with OVA without prior sensitization, autonomously secreted P in culture independent of further OVA stimulation. Furthermore, kinetic analysis showed P level in BAL to be closely correlated with the number of inflammatory cells in BAL during consecutive days of OVA challenge. It is likely that inflammatory cells infiltrated into lung tissues were present in the total lung cell preparation and contributed to P secretion in culture, but we cannot rule out the possibility that lung epithelial cells from asthmatic mice may also have produced and contributed to P secretion into BAL. By using P−/− mice and a function-blocking mouse anti-P mAb in WT mice, we further demonstrated that P deficiency protected mice from airway inflammation and attenuated Th2 and Th17 immune responses. These data showed that P plays an important role in promoting AP complement activation in asthma, and blocking P may be an effective way to ameliorate the known pathogenic effect of complement in this disease.
Complement is a danger-sensing component of the innate immune system that is critical for host defense (48, 49). However, inappropriately regulated complement, particularly the AP, can cause severe inflammatory and allergic reactions and damage host tissues (50–53). Previous animal studies have shown that complement can regulate Th2 and Th17 immune responses, AHR, and pulmonary inflammation and remodeling (11–18, 47, 54, 55). Some of the earlier studies focused on receptors of the complement anaphylatoxins C3a and C5a, and used either gene-targeted mice or pharmacological reagents to block C3aR and/or C5aR signaling (11–15, 46, 47, 54–56). These studies demonstrated that the C3a/C3aR pathway promoted allergic asthma both at allergen sensitization (47) and challenge phases (54, 55) and regulated IL-17 production (47, 54, 55). In contrast, the role of the C5a/C5aR pathway was found to be more complex, and blocking C5aR activity at the sensitization or challenge phase produced opposing results (15, 44, 57, 58). Other studies examined the roles of C3, fB, C5, and fH in asthma (16, 57, 59–62). Although these works have collectively supported the importance of complement in allergic asthma, questions remain on the regulatory mechanisms of complement activation and on the most optimal and promising target for anti-complement therapy in asthma.
A major finding of this study is that P played a pathogenic role in airway inflammation primarily at the effector phase of disease induction. Using a function-blocking anti-P mAb, we were able to selectively inhibit P activity at either the allergen sensitization or challenge phase or at both phases. Blocking P activity in WT mice at both phases recapitulated data obtained from P−/− mice, that is, mice developed attenuated asthmatic lung injury. Interestingly, blocking P activity at the allergen sensitization phase alone had no effect on airway inflammation or Th2 and Th17 cytokine production by restimulated MLN cells, whereas blocking P activity at the allergen challenge phase alone was sufficient to alleviate lung inflammation and Th2, Th17 immune responses. The only effect that we were able to detect from blocking P activity at the allergen sensitization phase was augmented IFN-γ production by restimulated MLN cells. The fact that intranasally delivering anti-P mAb before OVA challenge had similar therapeutic efficacy to systemically administered anti-P mAb suggests that P plays a pathogenic role locally in the airways. This conclusion was further supported by local P reconstitution experiments and by diminished level of immunoreactive P in the BAL of anti-P mAb–treated mice. We showed that intranasal delivery of recombinant P to P−/− mice before OVA challenge restored sensitivity of these mice to airway inflammation and increased Th2 and Th17 immune responses by restimulated MLN cells.
Mechanistically, locally released P in the airways appeared to have contributed to eosinophilia and lung injury primarily through increased C3a production. We detected a significant reduction in C3a levels in the BAL of P−/− mice compared with WT mice subjected to the same OVA-dependent airway disease induction, and there was a positive correlation between P and C3a levels in the BAL of allergen-challenged human asthmatic patients. Furthermore, blocking C3a function in the airways with a neutralizing mAb largely abrogated the pathogenic effect of intranasally reconstituted P protein in P−/− mice. Although we cannot entirely rule out the relevance of the C5a/C5aR pathway in the mechanism of action of P in this model, P deficiency seemed to have less of an effect on local C5a production in the airways. We found C5a levels in the BAL of experimental P−/− mice to be moderately reduced compared with that of WT mice, and conversely, intranasal administration of P to P−/− mice increased BAL C5a levels (Supplemental Fig. 4). However, the differences between these groups did not reach statistical significance. It is possible that nonspecific protease activities contributed more significantly to local C5a production in the airways, whereas C3a generation was more dependent on P-facilitated AP activation.
Whether P acted as a trigger or facilitator of AP complement activation in the airways has not been determined in this study. Recent studies have shown that surface-bound P can act as a platform to assemble new C3 convertases and initiate AP complement activation (23, 33, 63–66). Although direct binding of plasma P to host cells in the absence of deposited C3b has not been demonstrated, newly released P from activated neutrophils was found to be capable of binding to apoptotic cells and trigger AP complement activation (40, 67). Thus, it is possible that locally released P from initially migrated leukocytes may bind to airway epithelial cells and trigger AP complement activation, amplifying inflammatory lung injury. Alternatively, P may have primarily acted as a stabilizer of the AP C3 convertase formed via a tick-over mechanism on asthmatic airway surfaces where further AP amplification may subsequently occur because of loss or reduction of negative complement regulatory activities. Regardless, there is little doubt that AP complement is activated in asthmatic airways as has been demonstrated also by studies of fB−/− mice (59). The balance between positive and negative regulators of AP may critically determine how much complement-mediated injury ensues (30). For example, a previous report has demonstrated the protective role of fH as a negative regulator of AP complement in a murine model of asthma and provided proof of concept favoring therapeutically increasing fH activity in the treatment of asthma (62). Our studies in this article in turn have revealed the pathogenic role of P as a positive regulator of AP complement in asthma and demonstrated the therapeutic efficacy of blocking P in the effector phase of the disease.
The effect of blocking P activity before allergen challenge on Th2 and Th17 immune responses and on serum IgE levels is quite notable. Although we did not observe any effect on Th2 or IL-17A cytokine change in the BAL of experimental mice in association with P inhibition, IL-4, IL-5, and IL-17A production by OVA restimulated MLN cells was invariably affected when P activity was manipulated at the effector phase. It is possible that P-dependent local C3a production in the airways regulated the migration, proliferation, differentiation, and survival of memory Th2 and Th17 cells in MLN but not in the airways. The dissociation between airway inflammation and Th2 and IL-17A levels in the BAL of P−/− mice suggested that these cytokines may not be the sole mediators responsible for pulmonary inflammation. Conversely, inflammatory infiltrates in the BAL of P−/− mice may not be the only source of Th2 and IL-17A cytokine production. Another prominent finding was that blocking P at the allergen challenge phase strikingly reduced total serum IgE levels. In contrast, we observed no serum IgE changes when P was blocked at the allergen sensitization phase and, intriguingly, when P was inhibited at both allergen sensitization and challenge phases, nor did we see a change in serum IgE in P−/− mice. Thus, therapeutically blocking P activity at the effector phase may be more desirable than blocking at both phases in alleviating the adverse allergic immune reactions. The link between serum IgE level and P activity in the airways during allergen exposure remains an observational finding and more studies are required to elucidate the underlying mechanisms. It did appear, however, that the regulating effect of P on serum IgE was mediated also by C3a because reconstitution of P−/− mice at the effector phase increased serum IgE as expected and coadministration of an anti-C3a mAb reversed this effect.
In summary, we have shown in this study that P plays a pathogenic role in allergen-induced airway inflammation and abnormal Th2 and Th17 immune responses. By selectively inhibiting P activity with a function-blocking mAb at different phases of disease pathogenesis and by testing different routes of introduction of the mAb, we were able to determine that P contributed to lung injury locally during allergen challenge only. Furthermore, we demonstrated that P exerted its effect in allergen-induced airway disease primarily through the promotion of AP complement activation and C3a production. Collectively, our data suggest that therapeutic inhibition of P locally in the airways may protect susceptible asthmatic patients from allergen-induced lung inflammation and injury.
This work was supported by National Institutes of Health Grants R21AI103965, RO1AI044970, and RO1AI085596 (to W.-C.S.) and Grants RO1AI072197, RC1ES018505, and P30ES013508 (to A.H.).
The online version of this article contains supplemental material.
The authors have no financial conflicts of interest.