Abstract
Bacterial colonisation of the airways is associated with increased risk of childhood asthma. Immunoglobulin (Ig)E against bacterial antigens has been reported in some asthmatics, suggesting a role for bacterial-specific type-2 immunity in disease pathogenesis. We aimed to investigate relationships between bacterial-specific IgE amongst teenagers and asthma susceptibility.
We measured titres of IgE against Haemophilus influenzae, Streptococcus pneumoniae and Staphylococcus aureus in 1,380 teenagers, and related these to asthma symptomatology and immunophenotypes.
IgE titres against S. aureus-derived enterotoxins were highest amongst atopics and were associated with asthma risk. Surprisingly, IgE titres against H. influenzae and S. pneumoniae surface antigens were higher, not stratified by atopy and independently associated with decreased asthma risk.
The positive association between type-2 immunity to S. aureus and asthma phenotypes probably reflects IgE-mediated effector cell activation via enterotoxin super antigens which are secreted in soluble form. The contrasting benign nature of type-2 immunity to H. influenzae and S. pneumoniae antigens may reflect their lower availability in soluble forms that can crosslink IgE receptors. We theorise that instead they may be processed by antigen presenting cells and presented to type-2 memory cells leading to mucosal secretion of interleukin (IL)-4/IL-13, a mechanism widely recognised in other tissues to attenuate T-helper-1 associated bacterial-induced inflammation.
Lower respiratory tract infections have been linked in many studies to the development and expression of asthma during childhood, especially amongst atopics 1. The primary focus has been upon viruses, particularly rhinovirus 2, 3, and bystander airway tissue damage during host-antiviral responses is widely believed to play a key role in driving wheezing disease pathogenesis towards chronicity. Considerably less attention has been paid to bacteria, with the exception of Staphylococcus aureus (SA) strains which produce soluble enterotoxins. These molecules function as superantigens and can directly activate a large proportion of peripheral T-cells 4, triggering polyclonal production of both T-helper (Th)1 and Th2 cytokines and synthesis of enterotoxin-specific immunoglobulin (Ig)E. Colonisation with superantigen-producing SA has been suggested to play a role in a range of atopy-associated diseases, including atopic dermatitis 5, rhinosinusitis and allergic rhinitis 6, 7, as well as adult asthma 8, 9 and childhood wheeze 10. Recently, colonisation of the upper airways during early childhood with bacterial pathogens, including nonenterotoxin-producing strains, has been linked with subsequent development of persistent asthma 11, but the pathogenic mechanisms are unknown and the interpretation of these findings remains controversial 12.
The presence of underlying type-2 immune responses to nonenterotoxin-producing bacteria in asthmatics, in particular bacterial-specific IgE, has been known for some time 13, 14. Moreover these responses have recently been shown to be markedly increased following severe asthma exacerbations in children 15 suggesting that antigens from these bacteria gain entrance to the immune system via the inflamed airway mucosa. The significance of these findings is unclear, and more detailed investigations of relationships between host immune responses, including IgE, to different classes of mucosal dwelling bacteria and susceptibility to asthma are needed. In particular, there is a requirement for baseline data on the spectrum of IgE responsiveness to bacteria in the normal population. Herein, we report on IgE responses to three potential bacterial pathogens, Haemophilus influenzae (HI), Streptococcus pneumoniae (SP) and enterotoxin-producing SA, in a community cohort of 1,380 teenagers, and the positive- and negative-risk associations between antibody titres and underlying asthma-associated clinical phenotypes.
MATERIAL AND METHODS
Study subjects
The West Australian Pregnancy Cohort study (Raine Study) is a longitudinal birth cohort; mothers were enrolled from antenatal clinics from the main tertiary maternity hospital in Perth, Western Australia, and were not selected on any criteria other than having enrolled for antenatal care at the hospital 16. Analyses presented here were based on data collected at the 14-yr follow-up of this cohort, and did not utilise information collected at other ages. By studying asthma in 14-yr-old children we are biasing towards asthma that is more closely associated with atopy and more likely to persist into adult life. All aspects of the current study were approved by our institutional human ethics committee (Princess Margaret Hospital Ethics Committee, Princess Margaret Hospital, Subiaco, Perth, Australia) and parents gave written consent, with assent given by the teenagers.
Clinical phenotyping
Current asthma was defined as wheeze or cough without a cold in the last 12 months, plus use of any asthma medication in the last 12 months in children with a doctor diagnosis of asthma ever. Current rhinoconjunctivitis was assessed on the basis of parental response to a standardised questionnaire regarding the child's symptoms (runny, blocked or itchy nose in the presence of runny or itchy eyes) over the preceding 12 months.
All spirometry measurements were performed in the morning (10:00–12:00 h) and conducted according to American Thoracic Society guidelines with the subjects seated 17. Standard variables, forced vital capacity (FVC), forced expiratory volume in 1 s (FEV1), forced expiratory flow at 25–75% of FVC and FEV1/FVC were calculated.
Subjects were eligible to participate in methacholine (MCh) challenge testing if they obtained reproducible spirometry, had an FEV1 ≥80% of predicted, had no respiratory illness in the last 14 days and had withheld from their asthma medications for the standard period 17. MCh challenges were performed with a Koko® digidoser (nSpire Health Inc., Longmont, CO, USA) using a modified dosimeter technique 18. Subjects were given an initial saline dose followed by doubling doses of MCh (0.0625, 0.125, 0.25, 0.5, 1, 2, 4 and 8 mg·mL−1) delivered via DeVilbiss 646 nebulisers (DeVilbiss Healthcare, Somerset, PA, USA). Testing was stopped once a patient's FEV1 had fallen by ≥20% or the highest dose was administered. The dose–response slope was calculated as the two-point slope of the curve regardless of the fall in FEV1 19.
Lung function was represented in the statistical analysis by the FEV1/FVC ratio. In a community population of adolescents where restrictive lung disease would be unexpected, this ratio is most likely to represent airway obstruction in a manner that is independent of body habitus, lung size and sex.
Immunological phenotyping
Atopic family history was obtained from the questionnaire completed by the primary caregiver. Questions were asked about first-degree relatives (biological mother, biological father or sibling) seeking information about asthma, eczema, rhinoconjunctivitis and other allergies. Parents were asked to provide information about whether the allergies had been confirmed and by whom. Information was recorded separately about maternal and paternal allergic history. A child was considered to have a positive family history if one or more first-degree relatives had confirmed asthma, eczema or rhinoconjunctivitis.
Current atopic status, including serum titres of allergen-specific IgE, was determined as detailed previously 20. Subjects were considered to be atopic if they had total IgE of ≥300 kU·L−1 and/or specific IgE of ≥0.35 kU·L−1 for any of the following allergens: house dust mite (HDM), rye grass pollen, cat, couch grass, mould mix, peanut or food mix. Total IgE as well as specific IgE to HDM (Dermatophagoides pteronyssinus), rye grass pollen (Lolium perenne), cat, couch grass (Cynodon dactylon), mould mix (Penicillium notatum, Cladosporium herbarum, Aspergillus fumigatus, Candida albicans, Alternaria alternata and Helminthosporium halodes), peanut and food mix (egg white, milk, fish, wheat, peanut and soybean) and to a mixture of SA enterotoxins (SAE) (Staphylococcal enterotoxin A, Staphylococcal enterotoxin B and toxic shock syndrome toxin 1), were measured by ImmunoCAP (Phadia AB, Uppsala, Sweden) 8.
IgE antibodies to outer membrane protein P6 and pneumococcal surface protein C (PspC) from HI and SP, respectively, were assayed by modifications to the procedure described to use humanised IgE antibody for absolute quantitation 21. A microtitre plate dissociation-enhanced immunofluoresence assay (DELFIA™) was performed where antigen coating was standardised by capturing His-tagged recombinant antigens with anti-His monoclonal antibody coated as described previously 22. The assay was calibrated by interpolating the results from a titration curve constructed with recombinant Derp2 captured by the same procedure and a standardised (IU·mL−1) humanised IgE anti-Derp2 (Indoor Biotechnolgies, Charlottesville, VA, USA). The lower limit of detection was 0.1 ng·mL−1 (0.041 kU·L−1) for P6 and PspC. The P6 antigen was prepared as described previously 15 and the PspC (aa 1 to 445) antigen as previously described by Brooks-Walter et al. 23.
Peripheral blood mononuclear cells (PBMC) were cryopreserved and subsequently cultured as described previously 20. PBMC were cultured for 48 h in AIM-V medium with 4×10−5 M 2-mercaptoethanol alone or with 200 ng·mL−1 Staphylococcal enterotoxin B (SEB; Sigma-Aldrich, Castle Hill, Australia). 0.6×106 cells were cultured in round-bottomed 96-well plates (Nunc, Roskilde, Denmark) at a concentration of 1.2×106 cells·mL−1. Measurements of cytokines and other inflammatory mediators are as detailed previously 20 and in the supplementary material.
Statistical analyses
Univariate analysis of group differences in IgE titres were examined using the Mann–Whitney U-test, and bivariate relationships were assessed using Spearman's correlation. Continuous variables were log10 transformed for use in regression analyses. To allow log10 transformation, cytokine delta values ≤0 were converted to 0.01; antibody values below the limits of detection were ascribed a value equivalent to half the limit of detection. Variables in multivariate analyses were selected based on univariate association with outcome at p<0.10. Bronchial hyperresponsiveness (BHR) dose slope was analysed with linear regression, asthma risk and BHR risk with logistic regression, and asthma severity with ordinal regression. Analyses were performed using SPSS software (SPSS Inc., Chicago, IL, USA).
RESULTS
IgE responses to SAE in atopic and non-atopic teenagers
IgE responses against SAE are common amongst teenagers (19.3% of cohort; table 1 in supplementary material) and more frequent amongst atopics (26.8% versus 8.0% in non-atopics). Moreover as shown in figure 1a the mean SAE-IgE titres were also preferentially increased amongst atopics, suggesting that responsiveness is linked to the atopic phenotype. This gradation is similar if results are expressed as percentage with IgE titres above assay detection limits (52.7% versus 22.3% respectively; table 1 supplementary material).
SAE-IgE and risk for asthma
Figure 1b shows SAE-specific IgE titres within the cohort after stratification by atopy and clinical phenotype, and demonstrates the consistent relationship between SAE-IgE titre and the presence of respiratory symptoms in atopics. Further analysis by univariate logistic regression (table 1) confirmed this association. In particular, when we stratified by atopy status SAE-IgE was a significant risk factor amongst atopics for asthma and BHR, while amongst non-atopics SAE-IgE was only a risk factor for BHR. Figure. 1c further examines relationships between SAE-specific IgE and asthma. For this analysis the population was grouped on the basis of serum SAE-IgE levels (undetectable plus population terciles amongst responders), and asthma risk was computed for each group by univariate logistic regression; this risk appears to increase progressively with antibody titre. This association was explored further via logistic regression modelling, focusing on the atopics identified in table 1 as the “at risk” group. In addition to bacterial-specific IgE titres, these analyses utilised a range of data derived from the study cohort relating to respiratory, inflammatory and atopy-associated functions 20. A two-step process was undertaken as shown in table 2. First, univariate logistic regression was performed to identify individual asthma risk variable, including a range related to intensity of underlying atopy. Of note, SAE-specific cytokine responses (interleukin (IL)-5, IL-10 and IL-13) were also included. The final multivariate regression model combined these significant variables, as shown in table 2. The list of variables identified as independently associated with asthma risk was dominated by measures related directly or indirectly to conventional atopy, notably aeroallergen-specific IgE and a marker of eosinophil activation. The IgE component of the SAE response did not remain as a significant risk variable but was supplanted by the SAE-specific cytokine response in the form of IL-10.
SAE-IgE and risk for BHR
A potential association between SAE-IgE responses and risk for BHR development is also suggested by the results in table 1. Accordingly, multivariate linear regression analyses to predict BHR dose slope were performed employing variables that were significant in initial univariate regressions. In contrast to asthma, univariate analyses showed that SAE-IgE was a risk factor for BHR for both non-atopic and atopic subjects (table 2 in supplementary material), so these were assessed separately in multivariate analyses. Within the atopic population SAE-IgE does not survive as a significant risk factor for BHR in the multivariate model (table 3); however, IL-5 responsiveness to SAE displays a strong positive correlation with degree of BHR. In contrast, SAE-IgE survives as an independent risk factor in the non-atopic subjects.
IgE responses to HI and SP antigens and risk for respiratory symptoms
The spectrum of IgE titres to HI P6 and SP PspC antigens within the study cohort are shown in figure 2a and table 1 in the supplementary material. IgE responses to these antigens are more frequent and stronger than those against SAE and, in contrast to the pattern seen for SAE, the atopic and non-atopic subpopulations display comparable responses to P6 and PspC. Moreover, IgE responses to these antigens are strongly intercorrelated within both atopics (rho 0.658; p = 0.000) and non-atopics (rho 0.623; p = 0.000) but are unrelated to responses to SAE (table 3 in supplementary material).
Univariate analyses on the whole population stratified by P6- and PspC-IgE titre (fig. 2b) and by atopic status (table 4) revealed relationships markedly different from those observed for SAE-IgE. In particular, IgE titres against these non-enterotoxin antigens associated inversely with asthma risk (fig. 2b versus fig. 1c, and table 4 versus table 1). Of note, amongst the atopics both P6-IgE and PspC-IgE remained as significant protective factors in relation to asthma risk in multivariate analyses, which included allergen specific IgE measures as confounders (table 5).
DISCUSSION
The most intensively studied bacteria in relation to pathogenesis of inflammatory airway diseases are enterotoxin-producing SA. Their secreted products can trigger production of T-cell cytokines including IL-5, IL-10, IL-13 and interferon-γ 24, 25, and can also lead to inhibition of T-regulatory cell activity 26.
In this cohort IgE responses to SAE were higher and more frequent amongst atopics, in particular those with airway symptoms (fig. 1a). Moreover, in univariate models SAE-specific IgE titre was a significant risk factor for asthma and BHR, particularly amongst atopics (table 1). SAE-IgE titres in atopics correlate strongly with those against conventional aeroallergens and total IgE (data not shown), and the SAE-IgE effects in these subjects relating to asthma risk were no longer significant after adjusting for overall IgE levels and the potential confounders identified in initial univariate analyses (table 2). This suggests that the direct disease promoting effects of IgE against SAE may be secondary to those resulting from IgE against the common aeroallergens, and/or may be restricted to subjects with severe wheeze as suggested previously 8–10. Alternatively, SAE-IgE may be a surrogate for other covert aspects of the Th2-associated response to SAE, including the effects of SAE on polyclonal stimulation of IgE production and/or on IgE independent mechanisms. It is pertinent to note earlier findings that T-cells from a subset of atopic children are hyperresponsive to stimulation with the archetypal SEB 27, 28, which triggers cytokines including IL-5, IL-10 and IL-13. Accordingly, we included assessment of PBMC responses to SEB as part of the immunophenotyping of this cohort. As noted in table 2, SEB-induced IL-5, IL-10 and IL-13 responses were associated with asthma risk in atopics in univariate analyses, and inclusion of these variables in the final multivariate model identified the IL-10 component of the SEB response as an independent associate of asthma risk. Despite the anti-inflammatory action of this cytokine in relation to allergen-specific Th2 responses 29, its production in excess and/or in the wrong context has been associated with increased susceptibility to wheezing-related symptoms 20, 30. Our findings are also consistent with a recent report linking susceptibility to asthma symptoms following occupational lipopolysaccharide (LPS) exposure to levels of LPS-induced IL-10 production by PBMC 31.
The association between SAE-IgE and risk for BHR was further examined by multiple linear regression, incorporating potential confounders identified by univariate linear regression. As found for asthma, SAE-IgE was not a significant risk factor for BHR in atopic teenagers (table 3). However, this model does contain SEB-induced IL-5 which can drive eosinophilia, in addition to eosinophil protein X, a product of activated eosinophils. Heightened eosinophil-related responses were previously identified as the strongest risk factor for wheeze and BHR in atopic school children 32.
In contrast to these findings relating to atopics, the positive association observed amongst non-atopics between BHR and SAE-IgE by univariate analysis (table 1) remained significant after adjustment for potential confounders by logistic regression (table 3). This suggests that SAE-specific IgE may function as an independent risk factor for BHR in non-atopics, implying that despite its low titre this IgE may contribute to airways inflammation in subjects in whom significant IgE of other specificities is extremely rare.
IgE antibodies to nasopharyngeal bacteria have been reported in a number of studies and a recent investigation detected a steep rise of anti-P6 IgE titres in children in the wake of severe asthma exacerbations 15, which implies that sufficient bacterial antigen penetrated the mucosa during this period to stimulate underlying bacterial-specific immunity. These responses are unlikely to be restricted to antibody, given that bacterial products exemplified by HI P6 antigen are also potent inducers of pro-inflammatory cytokines from human macrophages 33 including those at mucosae. Our findings do not identify the precise contribution of anti-bacterial Th2 immunity to the local cytokine milieu in the airways, but this contribution is clearly more complex than a straightforward addition to ongoing inflammation. In particular, HI- and SP-specific IgE is inversely associated with asthma risk in healthy atopic teenagers (table 4, fig. 2b) despite the fact that respective (mean) production levels are higher than corresponding responses to SAE which are positively associated with risk (fig. 1e versus fig. 2b). This discrepancy may be due, in part, to the active secretion of soluble SAE 34. In contrast, PspC antigen of SP would be expected to be mainly encountered when covalently attached to the cell wall 35 and although HI do secrete vesicles that contain the outer membrane antigen P6 36, these are 200-nm structures which may not be readily available for cross-linking IgE-FcR-armed receptors.
Additionally, the finding of circulating specific IgE antibody against HI and SP also automatically denotes the presence of additional forms of underlying Th2 immunity against these antigens. In particular, the presence of IL-4 and/or IL-13-secreting Th cells 37. Indeed, IL-13 production by HI-stimulated PBMC from both atopics and non-atopics has been reported previously for P6 38 and IL-4 production against other undefined antigens in HI extracts has also been reported 39. It is noteworthy in this regard that IL-4 and IL-13 are highly pleiotrophic cytokines with multiple regulatory functions beyond IgE regulation, including an important role in modulation of macrophage activity. However, pre-treatment of macrophages with IL-4 can induce “alternative activation” leading to potentiation of cytokine responsiveness to subsequent microbial exposure 40. Conversely, as demonstrated with human cells from a variety of sources including synovium, blood, peritoneum and the central nervous system 41–44, IL-4/IL-13 can also play an anti-inflammatory role via attenuation of bacterial-induced production of pro-inflammatory cytokines, such as tumour necrosis factor-α, IL-1 and IL-6 by macrophages. Thus, P6-specific and PspC-specific Th2-memory cells of the type required for stimulation of IgE production could serve as an inducible source of such "homeostatic" IL-4/IL-13 in response to transmucosal leakage of these antigens and their uptake by local antigen presenting cells 45, thus helping to attenuate the pro-inflammatory effects of ligands released by the organisms. Follow-up studies are required to elucidate the potential dualistic nature of these poorly understood Th2 responses to mucosal dwelling bacteria.
Acknowledgments
We acknowledge with thanks the skilled technical assistance of J. Tizard, the study families and the members of the Raine Study team who participated. We thank Phadia Diagnostics (Uppsala, Sweden) for the gift of the reagents for determination allergen-specific IgE antibody titres in the cohort.
Footnotes
For editorial comments see page xx.
This article has supplementary material available from www.erj.esjournals.com
Support Statement
The authors are supported by the National Health and Medical Research Council of Australia (Canberra, Australia). The Raine Medical Research Foundation at the University of Western Australia (Perth, Australia) provided funding for the ongoing core Raine Study. Phadia Diagnostics (Uppsala, Sweden) provided the reagents for determination allergen-specific IgE antibody titres in the cohort.
Statement of Interest
Statements of interest for S. Ahlstedt and P.G. Holt and the study itself can be found at www.erj.ersjournals.com/misc/statements.dtl
- Received November 19, 2009.
- Accepted January 14, 2010.
- ©2010 ERS