The nasal microbiome in asthma

Abstract

Background: Nasal microbiota may influence asthma pathobiology.

Objective: We sought to characterize the nasal microbiome of subjects with exacerbated asthma, nonexacerbated asthma, and healthy controls to identify nasal microbiota associated with asthma activity.

Methods: We performed 16S ribosomal RNA sequencing on nasal swabs obtained from 72 primarily adult subjects with exacerbated asthma (n = 20), nonexacerbated asthma (n = 31), and healthy controls (n = 21). Analyses were performed using Quantitative Insights into Microbial (QIIME); linear discriminant analysis effect size (LEfSe); Phylogenetic Investigation of Communities by Reconstruction of Unobserved States; and Statistical Analysis of Metagenomic Profiles (PICRUSt); and Statistical Analysis of Metagenomic Profiles (STAMP). Species found to be associated with asthma activity were validated using quantitative PCR. Metabolic pathways associated with differentially abundant nasal taxa were inferred through metagenomic functional prediction.

Results: Nasal bacterial composition significantly differed among subjects with exacerbated asthma, nonexacerbated asthma, and healthy controls (permutational multivariate ANOVA, P = 2.2 × 10^-2). Relative to controls, the nasal microbiota of subjects with asthma were enriched with taxa from Bacteroidetes (Wilcoxon-Mann-Whitney, r = 0.33, P = 5.1 × 10^-3) and Proteobacteria (r = 0.29, P = 1.4 × 10^-2). Four species were differentially abundant based on asthma status after correction for multiple comparisons: Prevotella buccalis, P_adj = 1.0 × 10^-2; Dialister invisus, P_adj = 9.1 × 10^-3; Gardnerella vaginalis, P_adj = 2.8 × 10^-3; Alkanindiges hongkongensis, P_adj = 2.6 × 10^-3. These phyla and species were also differentially abundant based on asthma activity (exacerbated asthma vs nonexacerbated asthma vs controls). Quantitative PCR confirmed species overrepresentation in asthma relative to controls for Prevotella buccalis (fold change = 130, P = 2.1 × 10^-4) and Gardnerella vaginalis (fold change = 160, P = 6.8 × 10^-4). Metagenomic inference revealed differential glycerolipid metabolism (Kruskal-Wallis, P = 1.9 × 10^-4) based on asthma activity.

Conclusions: Nasal microbiome composition differs in subjects with exacerbated asthma, nonexacerbated asthma, and healthy controls. The identified nasal taxa could be further investigated for potential mechanistic roles in asthma and as possible biomarkers of asthma activity.

Keywords: 16S rRNA sequencing; Nasal; asthma; microbiome; upper airway; upper respiratory tract.

Figure 1. Trend toward increased nasal bacterial alpha diversity with greater asthma activity

Rarefaction plot showing the total branch length using Faith's phylogenic diversity (PD) is plotted against fraction of retained reads for healthy controls (n=21) and subjects with non-exacerbated asthma (n=31) and exacerbated asthma (n=20). Differences were not statistically significant at P≤0.05.

Figure 2. Distinct bacterial composition in subjects with exacerbated asthma, non-exacerbated asthma, and healthy controls

At the phylum level, Bacteroidetes and Proteobacteria were significantly enriched in subjects with non-exacerbated and exacerbated asthma relative to controls (P=5.1×10^-3 and P=1.4×10^-2, respectively). These phyla were also differentially abundant in three-way comparisons based on asthma activity. Phyla with <0.1% abundance are aggregated as “Others”.

Figure 3. Linear discriminant analysis (LDA) demonstrated distinct bacterial genera enriched in exacerbated and non-exacerbated asthma

Genera (g) with P<0.05 and LDA score >2 were considered significant and are shown here with notation for their corresponding phylum (p).

Figure 4. Heatmap showing abundance of species significantly associated with asthma status and asthma activity

P_adj for two-way comparison based on asthma status (i.e. exacerbated asthma and non-exacerbated asthma vs. healthy control) using the Wilcoxon-Mann-Whitney test with adjustment for multiple testing via the Benjamini Hochberg method if first shown in black font. P_adj for three-way comparison based on asthma activity (i.e. exacerbated asthma vs. non-exacerbated asthma vs. healthy control) using the Kruskal-Wallis test with adjustment for multiple testing using the Benjamini Hochberg method is then shown in blue font. Each column of the heatmap represents an individual.

Figure 5. qPCR validation of Prevotella buccalis (A) and Gardnerella vaginalis (B) overrepresentation in subjects with asthma relative to controls

Pan bacteria, a measure of total bacterial load, was used as a “reference gene” to normalize each sample. t-test was used to assess difference in fold change between asthmatic and healthy samples.

Figure 6. Glycerolipid metabolism associated with asthma activity

(A) Metagenomic inference showed that glycerolipid metabolism was significantly different between subjects with exacerbated asthma, non-exacerbated asthma and healthy controls. (B) Pairwise comparison showed that this pathway was significantly decreased in subjects with exacerbated asthma vs. healthy controls, as well as in subjects with non-exacerbated asthma vs. healthy controls.