Background
The chronic lung diseases of bronchiectasis (BR) and chronic obstructive pulmonary disease (COPD) are both associated with recurrent airway infections. COPD is a major cause of death globally, with numbers of deaths rising [
1], and BR is underestimated with incidence rising in the UK by around 6% annually [
2]. Whilst they differ in disease causation, established disease in both is mainly characterised by repeated or persistent heavy bacterial colonization of the damaged lower respiratory tract. Such infection is associated with inflammation, mucus production, and reduced ciliary action, which promotes further infection, inflammation and tissue damage, in a vicious cycle [
3]. Studies have suggested that infection causes disease exacerbation and diminished lung function, which are often proportional to the bacterial load and to reduced diversity [
4,
5]. More recent findings propose more species-rich lung ecologies where alterations in specific bacterial populations, dysbiosis, is at the heart of clinical disease [
6,
7]. Pathogenic bacteria, as determined clinically by microbiological culture of expectorated sputum, are dominated by organisms specific to these diseases including
Pseudomonas aeruginosa, Haemophilus influenzae, Streptococcus pneumoniae and
Moraxella catarrhalis [
8]. Recent studies using DNA-sequencing technology reveal more detailed bacterial ecosystems in the lungs of diseased patients, but with culture approaches mainly corroborated [
9,
10].
P.aeruginosa is considered the major cause of morbidity (increased exacerbations and reduced lung function) and mortality in BR [
11], particularly during chronic infection and mucoid characteristics of the bacterium [
12], which may allow evasion of host immunity. Non-typeable strains of
Haemophilus influenzae (NTHi) are frequently found in BR [
13] and are not targeted by current vaccines. Both pathogens are also common in COPD albeit with a reduced frequency of Pseudomonas infections as compared to BR [
14]. Furthermore, less frequent suppurative infection and sputum production in COPD results in lower detection of pathogenic microbes, implying fewer infections than BR. Failure to produce sputum for microbiology, particularly in younger BR patients and in many COPD patients, as well as intermittent negative cultures, means that immune biomarkers of disease may provide a useful adjunct for directing clinical management.
Knowledge of immunity in BR is limited, but studies suggest immune system genes that are involved in presentation of antigens to CD4
+ T cells, such as HLA-DR1 and DQ5, play a role [
15,
16]. Notably, a role for adaptive immune responses (specific antibodies and T cells) in protection against
P.aeruginosa and H.influenzae, has been demonstrated in human vaccine trials in cystic fibrosis-related bronchiectasis [
17,
18] and in mouse vaccination models [
19,
20]. Furthermore, the above-mentioned lung pathogens appear in individuals with defined immunodeficiencies [
21], underlining the role of antibodies and phagocytes in protection. Whilst healthy individuals are exposed to the same pathogenic organisms as diseased individuals, healthy lungs typically have low levels of bacterial species, reflecting the naso-pharynx [
22]. Immune responses against pathogenic microbes do not cause overt immunopathology in healthy individuals, but may contribute to disease in colonized patients due to continuous immune stimulation by the localised high antigen doses, particularly through excessive Th17 responses that promote neutrophil infiltration [
23]. Together with inflammatory cytokines, neutrophils are abundant in the sputum of BR patients, and decline after antibiotic treatment [
24]. It is possible that dysfunction of both innate and adaptive immunity contribute directly or indirectly to disease in both BR and COPD. The aim of this study was to characterise antibody and T cell responses against key lung microbes in disease-stable patients with BR and COPD, characterised by the Bronchiectasis Severity Index (BSI) and GOLD guidelines, respectively, in comparison to controls (healthy volunteers), and to relate the immune responses to culture-based bacterial colonization, lung function and frequency of exacerbation.
Discussion
This study began by comparing immune responses against common lung pathogens in BR, COPD and HV. The clinical categorisation of the patients followed standard processes and was in keeping with other studies in the field, as were the microbiology results obtained. One expectation was that the degree of exposure to the microbes will be proportional to the magnitude of immune response measured. This was broadly the case for antibody responses, which were higher in BR than COPD and HV, particularly against
P.aeruginosa,
H.influenzae and
S.maltophilia, reflecting rates of positive sputum cultures in BR and COPD. Measurement of isotype components of the antibody responses against
P.aeruginosa showed a high IgG1 component in BR and HV, compared to COPD which had a higher IgM. This may suggest that COPD has reduced isotype switching, which is usually controlled by cognate T cell responses, through CD40:CD40L interaction and through cytokines. Reduced or altered antibody responses as we have seen here could be due to increased regulatory T cells, as have been demonstrated in COPD, which may depress protective immunity [
29].
Having found specific antibody responses to be increased in BR, the question was whether these responses showed a direct dynamic relationship with colonization levels. Sufficient numbers for this analysis were only available in the BR group. Whilst there was a trend for increasing antibodies with colonisation for each individual pathogen, only P.aeurginosa and S.maltophilia showed significance. We categorised patients based on their exacerbation frequency (< 3, ≥ 3) which were validated by showing reducing lung function. Although significant, only modest increases in antibody against P.aeruginosa, M.catarrhalis and H.influenzae were found in BR with ≥ 3 exacerbations compared to < 3. Antibody response only against H.influenzae showed a negative correlation with FEV1% predicted, suggesting it to be a marker of disease and exposure.
The measurement of T cell responses against lung pathogens may be useful for the diagnosis of latent infection, as is the case of the Quantiferon test for
Mycobacterium tuberculosis (Mtb). In this study T cell responses showed an overall tendency for reduction in BR and COPD compared to HV, associated with colonisation status, with responses to
P.aeruginosa and
S.pneumoniae being significantly reduced. This suggests that increased infection and exposure may exhaust the T cell response. Within the BR group T cell responses showed a trend for being highest in the group that had occasional infections, for all pathogens tested. The highest T cell responses were found for
H.influenzae and
M.catarrhalis which coincides with them having intracellular phases that require T cells for efficient immune protection or eradication. T cell responses did not show any associations with exacerbation level. However, increased IFNγ ELIspot T cell responses against
H.influenzae showed significant positive association, albeit weak, with lung function (FEV
1%) and negative association with BSI, which may suggest that T cells are protective against disease, in contrast to antibody responses which showed a negative correlation with FEV
1%, and may simply be associated with more infection. The next aim was to investigate further the nature of the T cells reactive against the two major pathogens,
P.aeruginosa and
H.influenzae, in a sub-group of BR patients who were good responders to the antigens and in comparison to good-responding HV. There was a tendency for IFNγ, IL-2 and IL-17 to be reduced in BR patients compared to HV, suggesting greater antigen exposure, where memory T cells producing IL-2 convert to T cells secreting effector cytokines. Conversely, there was a tendency for IL-10 to be increased in BR for both antigens suggesting their conversion to a regulatory (Tr1) phenotype due to high and sustained antigen exposure at the mucosal surface. IL-4 responses showed a significant increase in BR against
H.influenzae, similar to published work on COPD [
30], but a tendency for the opposite for
P.aeruginosa. This suggests a discrepancy in immune responses between BR and HV, and against the two pathogens, reflecting the fact that T cell response against
H.influenzae was protective against disease. When pathogen-reactive T cells, based on CD69 and CD4 staining, were examined for further key phenotypic markers no differences were found between BR and HV. All reactive cells had high levels of CD49d, a lung homing receptor, but low levels of inflammatory homing receptors and the marker of senescence PD-1.
The measurement of antibodies and T cells specific for
P.aeruginosa and
H.influenzae in patients with BR [
31,
32] and COPD [
30,
33] has previously revealed increased antibody responses associated with repeated infection, but decreased T cell responses, despite CD4
+ T cell presence and oligoclonal TcR T cell expansion in the lungs [
34,
35], suggesting immune dysregulation such as T cell exhaustion. Thus, while immune responses may be protective, or a marker of infection by microbes, their dysregulation may be detrimental to the patient due to reduced protection from infection or through immunopathology as suggested in cystic fibrosis [
36]. Studying responses in disease states is important as this may reveal mechanisms of disease that are direct (via immunopathology) or indirect (via anti-microbial effects) that may provide therapeutic targets. Furthermore, studies of such blood-based immunodiagnostics may be useful for diagnosis and stratification of patients, and their responses to treatment [
31], when microbiology or genomic analysis is not possible or reliable (young BR patients, no sputum, difficult to culture microbes, false negative). Baseline immunity related to contemporaneous microbiota may particularly be a useful way to identify a frequent exacerbator phenotype. With regard to an antibody marker of current colonization with
P.aeruginosa, this data showed 92% specificity (ability to show true negatives) and 73% sensitivity (ability to show true positives) based on the HV mean + 2 sd. This is similar to previous findings [
31].
The strengths of the study were the extensive nature of the immunological investigations carried out on patients, particularly those with BR, who were well characterised clinically and microbiologically. One weakness is that numbers of COPD patients producing sputum, and thus with positive cultures, was too low to allow a sufficiently powered analysis to be undertaken for COPD and so the study focussed on BR after the initial observations (Fig.
1). Furthermore, it would have been useful to have longitudinal data of immune responses and microbiology, and this is the subject of a future study. Another weakness is that microbiological culture is not able to determine the complete microbial makeup in a sample if it contains fastidious unculturable bacteria. We are currently addressing this by carrying out genomic analysis of patient sputum samples as well as microbiological culture. Finally, the cytokine secretion data would have benefited from larger numbers, particularly for HV, and again this is the subject of ongoing work.
Conclusion
In conclusion, exposure to these lung pathogens generates antibody responses of magnitudes that are broadly proportional to the level of exposure and thus disease (exacerbation, reduced lung function), and may be useful markers of disease. T cell responses appear to be reduced in patients with increased infection rates, and are proportional to lung function and BSI for H.influenzae, suggesting that they may be protective against such a pathogen that is partially intracellular. The T cell responses in patients differ little in phenotype from HV, apart from possible subtle cytokine differences that are currently being examined further. The interaction between T cells and antibody-producing B cells, and how the two arms of the adaptive immune response interact and influence innate immunity, and ultimately impact on bacterial infection and disease, is likely to be complex and multifactorial. The data in this study suggests the use of antibodies for Pseudomonas-inducing disease diagnosis, whilst T cells may indicate protective immunity against Haemophilus, suggesting a possible benefit of T cell-inducing vaccines.