To our knowledge this is the first study to examine whether specific aspects of the airway microbiota are related to relevant clinical or physiologic features of asthma. Our findings suggest that bacterial diversity, variations in community composition, and the relative abundance of specific phylotypes, are associated with the degree of bronchial hyperresponsiveness in asthmatics administered inhaled corticosterids. This culture-independent study also expands the repertoire of organisms of potential relevance to asthma pathogenesis within this cohort beyond those previously implicated1,2
. These data, together with recent studies in other asthmatic patients7
, suggest that the microbiome of the airways, as with other discrete host niches, may be an important contributor to asthma pathophysiology and to the heterogeneity of disease39–41
Overall, these findings highlight two key aspects in defining airway health. The existence of a bronchial microbiota suggests that the bronchial tree is unlikely to be completely sterile, and that the composition of its microbiome may directly or indirectly modulate airway function. The role of airway microbiota in airway disease is under-investigated compared with other human niches such as the intestinal microbiota, which has a demonstrated role in chronic inflammatory gastrointestinal diseases10, 42
. That a bronchial epithelia-associated microbiota exists is not entirely unexpected, particularly in the setting of airway disease and/or inhaled corticosteroid (ICS) therapy. Specific bacterial communities persist in the airways despite antimicrobial treatment in patients with COPD and cystic fibrosis8,19
. Recently, Hilty and colleagues also described airway microbiota in patients with mild to moderate asthma (all on ICS therapy), COPD, and healthy subjects7
. Those with airway disease had an increased prevalence of pathogenic Proteobacteria compared to controls, although relationships to clinical measurements were not examined in their study.
Complex microbial communities are now recognized to reside in various host mucosal sites43
, where perturbations of community structure are associated with disease6, 8, 10, 42
. Moreover, the airway epithelium is increasingly recognized as important in immunologic responses to environmental and microbial exposures44
. Thus, the airway microbiome could potentially influence presentations of asthma. As such, our findings of specific community relationships with bronchial hyperresponsiveness, in the context of a well-characterized asthmatic cohort receiving standardized baseline treatment, implicate additional bacterial groups in asthma. These include a Nitrosomonas
spp. possessing a functional nitric oxide reductase45
. Their relative abundance may be a microbial indicator of airway concentrations of nitric oxide which, though not assessed in this study, reflect airway inflammation46
and correlate with measures of airway hyperresponsiveness47
. Also notable is Oxalobacter formigenes
(Oxalobacteraceae), an anaerobic bacterium known to be susceptible to macrolide antibiotics48
. Other bacteria of interest include the Comamonadaceae and Sphingomonadaceae. Members of the Comamonadaceae, previously identified in cystic fibrosis patients49
, possess steroid-responsive degradation pathways50, 51
, and their presence could plausibly be related to the selective pressure of ICS therapy. This raises the intriguing possibility that steroid non-responsiveness, as observed in some asthmatics despite adherence to therapy52
, may be due in part to the presence of airway bacteria with steroid-degrading capacity.
Given the existence of airway microbiota, it is conceivable that coincident presence of multiple potentially pathogenic bacteria contribute to persistent airway inflammation in asthma. This possibility is supported by the finding of neutrophilia in bronchial airway samples of some patients with severe asthma despite regular use of high-dose inhaled corticosteroids53
. Such inflammation could represent an appropriate response by the host to inappropriate airway microbial colonization. For example, Sphingomonadaceae, members of which were significantly correlated with bronchial hyperresponsiveness, are characterized by the presence of cell membrane glycosphingolipids, which are recognized by and can activate invariant natural killer T (iNKT) cells resulting in induction of IL-4 and IL-1354, 55
. Glycolipid activation of iNKT cells can also induce airway hyperreactivity independent of conventional CD4+
. While the presence of iNKT cells in asthmatic airways remains controversial57, 58
, our findings suggest a possible role for Sphingomonadaceae in airway pathophysiology and the possibility that the observed variability of iNKT cell populations in asthmatics may depend in part on the relative abundance of these species.
We recognize that this pilot study is limited by the relatively small number of subjects, especially of healthy controls, and by the absence of asthmatics not taking an ICS. Nonetheless, the data indicate that airway bacterial burden in individuals without airway disease is much lower than in asthma patients requiring ICS therapy. As all asthmatics were required to be on standardized ICS for the parent study23
, we are unable to infer whether the higher bacterial burden among asthmatics is a function of the disease itself or ICS treatment, although the relationships of bacterial burden, and the composition and diversity of the microbiota, to bronchial hyperresponsiveness, are robust. Determining the effects of ICS use on the airway microbiome will require further investigation and is an important question, given the wide use of ICS therapies in airway diseases, which, among patients with chronic obstructive pulmonary disease, has been associated with an increased risk of pneumonia59
The array-based analysis in this study identified a greater diversity of airway microbiota than has previously been described, although to our knowledge there has been only one other study involving asthmatics that applied 16S rRNA phylogenetic analysis7
. In addition to clone library-sequencing validations, our results also agreed with those of Hilty et al
., in that we detected all bacterial phyla and genera identified in their study of the lower airways of eleven adult and thirteen pediatric asthma patients on ICS therapy. The PhyloChip platform permits in-depth analysis of relatively large numbers of samples using a standardized assay that applies several levels of stringent criteria for determining the microbiota profile in a given sample. It is impractical to perform sequence corroboration for every array-based positive hit. Moreover, the extent of microbial diversity detected is dependent on the community structure and the depth of sequencing performed60
. Our study included the largest number of clones sequenced to date for asthmatic airway samples and confirmed the presence of many array-detected phylotypes, indicating that complex microbiota do exist in this niche.
Finally, we observed that asthmatics exhibiting a significant decrease in bronchial hyperresponsiveness post-clarithromycin treatment in the parent study, possessed greater pre-treatment airway bacterial diversity than did non-responders. Despite the small number of subjects in this exploratory analysis, the trend is interesting given that prior studies have also observed decreased bronchial responsiveness following prolonged macrolide therapy4, 61
Furthermore, while the antimicrobial susceptibility of the entire airway microbiota is unknown, conceivably a multitude of community members may be sensitive to macrolides, resulting in a reduction of bacterial diversity and/or burden upon treatment and manifesting clinically in a reduction in bronchial responsiveness. Thus, while the efficacy of macrolides in airway disease is often attributed to their anti-inflammatory properties, their effectiveness could also reflect an extended spectrum of antibacterial activity against members of the airway microbiota, whose composition, as in the setting of ICS, may influence outcomes.
In summary, several features of the airway microbial community are significantly associated with the degree of bronchial hyperresponsiveness among patients with sub-optimally controlled asthma. Despite limitations, these study findings are notable in providing the first evidence for the potential functional, physiologic and clinical relevance of the airway microbiome in asthma. Our results suggest that variations in airway microbiome structure and function may exert distinct effects that contribute to asthma heterogeneity and provide novel targets and hypotheses for future studies on disease mechanisms.
Relationships between airway microbiota and clinical features of asthma were investigated using culture-independent approaches. Among sub-optimally controlled adult asthmatics, bronchial hyperresponsiveness correlated strongly with airway microbiota composition and diversity.