Autoimmunity to bactericidal/permeability-increasing protein in bronchiectasis exhibits a requirement for Pseudomonas aeruginosa IgG response

To the Editor:

In bronchiectasis, due to diseases other than cystic fibrosis, idiopathic, genetic and environmental factors alter the airway landscape and immune responses, rendering the patients susceptible to infection, with the Gram-negative bacterium (GNB) Pseudomonas aeruginosa as a major contributor to mortality [1, 2]. The high prevalence of P. aeruginosa in bronchiectasis patients cannot be explained by a single genetic or environmental influence, and immunological permissiveness of bronchiectasis airways to this colonisation remains unexplained [2]. A similar predisposition to this pathogen is characteristic of cystic fibrosis patients, in whom a single genetic mutation (CFTR) shapes the abnormal lung environment.

In cystic fibrosis, where P. aeruginosa colonises the airways of up to 80% of patients, we and others have proposed that the inability of the innate immune system to combat P. aeruginosa infection is related, in part, to an autoimmune antibody response to bactericidal/permeability-increasing protein (BPI), a neutrophil antimicrobial protein [3]. Through high-affinity binding of lipopolysaccharides (LPS) on the bacterial outer envelope, BPI mediates extracellular bactericidal and LPS neutralising functions [4]. Autoantibodies to BPI were first reported in European cystic fibrosis cohorts and confirmed by us in a US cohort of adult cystic fibrosis patients [5, 6]. Notably, autoreactivity to BPI was associated with diminished lung function while in vitro functional studies demonstrated that anti-BPI IgG inhibits its biological activities [7–9].

In this research letter, we ask two critical questions regarding the immunological interactions that shape the bronchiectatic airway permissiveness to P. aeruginosa infection. 1) Do bronchiectasis patients develop autoimmunity to BPI? 2) What is the relationship between autoreactivity to BPI and chronic infection by P. aeruginosa? To address these questions, autoantibodies to BPI in patient sera were measured by ELISA in two bronchiectasis cohorts from the USA: one from Dartmouth Hitchcock Medical Center (DHMC) in Lebanon, NH (n=16), and the other from Oregon Health and Science University (OHSU) in Portland, OR (n=42). Immunoblotting of sera negative by ELISA yielded a low frequency of additional seropositivity (∼10%). BPI autoreactivity was identified at nearly identical frequencies in the DHMC (56%) and OHSU (52%) cohorts (figure 1a and b).

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 1

High anti-bactericidal/permeability-increasing protein (BPI) IgG titers in bronchiectasis (BE) patients associate with chronic Pseudomonas aeruginosa infection in a temporal and pathogen-specific manner. a) Anti-BPI IgG titers detected by ELISA in two BE cohorts: Dartmouth Hitchcock Medical Center (DHMC) (n=16, 37.5% positive) and Oregon Health and Science University (OHSU) (n=42, 45.2% positive). Anti-BPI IgG positive samples >6 U·mL⁻¹; positive cut-off determined as mean of healthy controls (HC)+2sd (n=16), represented by dashed line. b) Representative immunoblots of bronchiectasis serum reactivity to BPI (5 µg) and P. aeruginosa PA14 lysate (10 µg). c and d) Anti-BPI IgG positivity, determined by ELISA and immunoblot, associates with antibody reactivity to P. aeruginosa (PA14 lysate) in both BE cohorts: c) BE DHMC and d) BE OHSU; reactivity to P. aeruginosa determined by ELISA (positive cut-off of >22 U·mL⁻¹ represented by dashed line); filled symbols represent positive reactivity to P. aeruginosa by immunoblot. e) Fold change in anti-BPI and anti-P. aeruginosa IgG titers over two sequential visits in a retrospective longitudinal cohort of BE patients from OHSU (n=34; r=0.623; p=8.412×10⁻⁵ as determined by Spearman correlation analysis). f) Anti-BPI IgG positivity in BE patients (OHSU) is associated with positive P. aeruginosa sputum culture. Positive sputum culture for nontuberculous mycobacterium (NTM) does not associate with anti-BPI IgG positivity in the absence of antibody response to P. aeruginosa. No infection: no current NTM infection. Filled symbols represent positive antibody reactivity to P. aeruginosa.

We and others have reported an association between autoreactivity to BPI and the presence of P. aeruginosa in sputum culture of cystic fibrosis patients [5, 6]. Given these findings, we evaluated the relationship between anti-BPI autoimmunity and chronic P. aeruginosa infection in bronchiectasis by measuring anti-P. aeruginosa IgG titres in patient sera as a serological marker of chronic infection. Healthy control sera exhibited little or no reactivity against PA14 extract (figure 1b). Autoreactivity to BPI in the DHMC bronchiectasis cohort was strongly associated with the existence of an antibody response to P. aeruginosa (n=16, p=0.003) (figure 1c). This association was confirmed in an independent cohort of bronchiectasis patients (OHSU; n=42, p=0.002) (figure 1d). Each bronchiectasis cohort exhibited a dichotomised relationship between anti-BPI autoreactivity and the presence of anti-P. aeruginosa antibodies (figure 1c and d). Together, these findings represent the first report of anti-BPI autoimmunity in bronchiectasis patients in the USA and indicate that autoreactivity to BPI develops specifically in the context of chronic P. aeruginosa infection, independently of a single genetic factor.

The relationship between the IgG antibody responses to P. aeruginosa and BPI was further examined through a retrospective longitudinal study of sera from 34 bronchiectasis patients. Serological analyses demonstrated that levels of antibodies targeting BPI and P. aeruginosa changed in the same direction over two consecutive visits; i.e. an increase in anti-P. aeruginosa IgG titres was accompanied by a positive fold change in anti-BPI autoantibody titres, while a reduction in anti-P. aeruginosa antibody levels tracked together with a negative fold change in anti-BPI IgG titres (figure 1e). Therefore, the autoimmune response to BPI follows the same temporal pattern as the humoral response to P. aeruginosa.

The specificity of this interaction with P. aeruginosa exposure and autoantibodies to BPI was also examined in relation to nontuberculous mycobacterium (NTM). Segregation of the OHSU bronchiectasis cohort by sputum culture yielded three patient populations: history of positive P. aeruginosa (with some overlapping NTM positivity within 6 months), NTM positive within 6 months, or no current infection. Serological analyses of each population demonstrated that anti-BPI IgG were absent in patients colonised with NTM or in those without an infection by sputum culture, unless accompanied by a positive anti-P. aeruginosa IgG titre (figure 1f). A positive sputum culture for P. aeruginosa was reported in only 55% of patients positive for anti-P. aeruginosa IgG, indicating that the serological analysis captures prior, as well as current, P. aeruginosa infection [10].

In this study, we report that anti-BPI autoreactivity in bronchiectasis is strongly associated with chronic P. aeruginosa infection, characterised by the presence of anti-P. aeruginosa antibodies. We observed this identical association in two bronchiectasis cohorts from New England and the Pacific Northwest. The synchronised changes in antibody reactivity to P. aeruginosa and BPI in a longitudinal cohort of bronchiectasis patients (figure 1e) further support the model that the breaking of tolerance to BPI is mediated through an association with chronic P. aeruginosa infection. The remarkable conservation of the linkage between autoreactivity to BPI and chronic P. aeruginosa infection in bronchiectasis is heightened by: 1) the heterogeneous genetic nature of bronchiectasis [11] and 2) identical associations in cystic fibrosis [5, 6]. Together, these data argue against a human leukocyte antigen-dependent mechanism by which tolerance is broken, which is further bolstered by a similar relationship between BPI autoreactivity and immune response to P. aeruginosa in a bronchiectasis cohort from Japan [12]. Several potential mechanisms leading to the breach of tolerance to BPI in the context of P. aeruginosa infection warrant investigation: 1) molecular mimicry, 2) cross-activation of immune response due to LPS:BPI complexing and 3) cryptic epitope reveal due to differential BPI processing during inflammation. The latter model has been supported by our previous findings that P. aeruginosa-stimulated neutrophil extracellular trap formation leads to BPI cleavage and a possible reveal of neoepitope(s) [6].

The strength of the association between autoreactivity to BPI and chronic P. aeruginosa infection stands out in marked contrast to other autoantibodies against neutrophil azurophilic granule proteins (anti-neutrophil cytoplasmic antibodies (ANCA)), such as serine proteinase 3 and myeloperoxidase where no one infectious trigger has been defined [13]. In ANCA-associated vasculitis, while infectious stimuli have been implicated in the disease onset, their influence is only relevant in the context of genetic factors [14], unlike the P. aeruginosa–BPI interaction that argues against a genetic component of peptide restriction. Beyond the issue of immunopathogenesis of anti-BPI reactivity lie the functional implications of this autoreactivity. Three main functional roles of BPI have been proposed: 1) bactericidal, via LPS binding and permeabilisation of the GNB membrane; 2) inflammatory, via transport of GNB to dendritic cells; and 3) anti-inflammatory, via LPS neutralisation and down-modulation of monocyte response [15]. The temporal relationship between anti-P. aeruginosa and anti-BPI IgG titres suggests that autoimmune responses to BPI hinder its bactericidal and anti-inflammatory functions to facilitate colonisation/infection by P. aeruginosa. Thus, rather than viewing anti-BPI as a highly linked epiphenomenon of P. aeruginosa infection, these autoantibodies may play a causal role in the perpetuation of infection. In this model, strategies that eliminate anti-BPI reactivity may enhance clearance of P. aeruginosa, leading to improved airway function and clinical outcomes. Creating a model system in which the functional role of BPI and anti-BPI responses can be tested in vivo would seem to be a necessary first step in addressing this question.