Airway microbiota and bronchial hyperresponsiveness in patients with suboptimally controlled asthma

Abstract

Background: Improvement in lung function after macrolide antibiotic therapy has been attributed to reduction in bronchial infection by specific bacteria. However, the airway might be populated by a more diverse microbiota, and clinical features of asthma might be associated with characteristics of the airway microbiota present.

Objective: We sought to determine whether relationships exist between the composition of the airway bacterial microbiota and clinical features of asthma using culture-independent tools capable of detecting the presence and relative abundance of most known bacteria.

Methods: In this pilot study bronchial epithelial brushings were collected from 65 adults with suboptimally controlled asthma participating in a multicenter study of the effects of clarithromycin on asthma control and 10 healthy control subjects. A combination of high-density 16S ribosomal RNA microarray and parallel clone library-sequencing analysis was used to profile the microbiota and examine relationships with clinical measurements.

Results: Compared with control subjects, 16S ribosomal RNA amplicon concentrations (a proxy for bacterial burden) and bacterial diversity were significantly higher among asthmatic patients. In multivariate analyses airway microbiota composition and diversity were significantly correlated with bronchial hyperresponsiveness. Specifically, the relative abundance of particular phylotypes, including members of the Comamonadaceae, Sphingomonadaceae, Oxalobacteraceae, and other bacterial families were highly correlated with the degree of bronchial hyperresponsiveness.

Conclusion: The composition of bronchial airway microbiota is associated with the degree of bronchial hyperresponsiveness among patients with suboptimally controlled asthma. These findings support the need for further functional studies to examine the potential contribution of members of the airway microbiota in asthma pathogenesis.

Figure 1. 16S rRNA bacterial burden in bronchial brushings from asthmatic and healthy subjects

Of subjects analyzed by PhyloChip (total n = 47), asthmatic subjects exhibited greater bacterial burden (measured concentration of 16S rRNA PCR product) than healthy subjects (p = 0.008, Welch’s t-test with log-transformed data). Median and interquartile ranges are noted.

Figure 2. Relationships between variability in airway microbiota composition and community diversity with bronchial hyperresponsiveness

Each circle represents the microbiota present in a single sample (n=44). 2A, 2B. NMDS ordination, based on Sorensen dissimilarity matrices, demonstrates that variability in community composition is most strongly related to bronchial hyperresponsiveness and bacterial burden. Axes have no inherent units and indicate spatial relationships of samples based on their phylogenetic (dis)similarity. Three asthmatic samples were excluded due to extreme mean relative Sorensen distances from the pooled mean. 2C. Increased diversity (higher Shannon indices) is correlated with greater bronchial hyperresponsiveness. (Pearson correlations with log-transformed data.)

Figure 3. Phylogenetic tree based on 16S rRNA representative gene sequences (GenBank accession numbers) of the ~100 bacterial taxa highly correlated with greater bronchial hyperresponsiveness

(p < 0.01, q ≤ 0.015). Colors represent different bacterial families. Asterisks denote taxa with member species previously associated with clinical disease or possessing notable functional features (See also Table E2 in the Online Repository).

Figure 4. Baseline airway microbiota diversity and change in bronchial hyperresponsiveness with clarithromycin treatment

Sixteen asthmatic subjects with pre-treatment samples analyzed by microarray received clarithromycin therapy. Subjects with significant improvement in bronchial hyperresponsiveness post-clarithromycin (n=9) possessed higher pre-treatment airway bacterial diversity than those with no significant response (n=3). Four subjects with incomplete data due to study discontinuation were not included.