Assessment of airway microbiota and inflammation in cystic fibrosis using multiple sampling methods

Abstract

Rationale: Oropharyngeal (OP) swabs and induced sputum (IS) are used for airway bacteria surveillance in nonexpectorating children with cystic fibrosis (CF). Molecular analyses of these airway samples detect complex microbial communities. However, the optimal noninvasive sampling approach for microbiota analyses and the clinical relevance of microbiota, particularly its relationship to airway inflammation, is not well characterized.

Objectives: The goals of this study were to compare molecular analyses of concurrently collected saliva, OP swabs, IS, and expectorated sputum (ES) from children with CF and to determine the association between microbiota, lung function, and airway inflammation.

Methods: Saliva, OP swabs, IS, and ES were collected from 16 children with CF. Spirometry was performed.

Measurements and main results: Respiratory and saliva samples (n = 61) were sequenced for bacterial microbial communities, and total and CF-specific bacterial quantitative PCR assays were performed. Airway samples underwent conventional culture for CF-specific pathogens. Neutrophil elastase, IL-1β, IL-1ra, IL-6, Il-8, TNF-α, and vascular endothelial growth factor were measured in ES and IS. Sequencing results from individual subjects were similar across samples, with greater between-subject than within-subject variation. However, Pseudomonas and Staphylococcus were detected in higher relative abundance from lower airways (ES and IS) compared with paired upper airway samples (OP and saliva). Pseudomonas, Staphylococcus, and Enterobacteriaceae correlated with increased airway inflammation. Divergence between microbiota in upper airway compared with lower airway samples, indicating greater differences between communities, was associated with increased sputum neutrophil elastase.

Conclusions: Bacteria detected in IS samples resemble ES samples, whereas OP samples may underrepresent bacteria associated with airway inflammation. Divergence of lower airway communities from upper airway was associated with airway inflammation and may portend disease progression.

Keywords: DNA sequencing; bacterial infection; microbiome; oropharyngeal swab; sputum.

Figure 1.

Bacteria detected by sequencing from each participant by sample type. Relative abundances (RAs) of the most prominent taxa are displayed in the bar graph. Participants A through P are ordered by increasing β-diversity (Hill number, absolute Hill number; specified in A indicating greater difference in community between sample types). In specimens where total RA displayed is less than 100%, the proportion not displayed was made up of other taxa present in low RA (<1%). Total bacterial load (log₁₀ bacterial ribosomal RNA copy number/ml) is given below each sample. Positive or negative culture results for cystic fibrosis pathogens (Psuedomonas aeruginosa, Staphylococcus aureus, Haemophilus influenzae, Stenotrophomonas maltophilia, and Streptococcus milleri group) are indicated by + and – below the bar graphs. AF = amplification failed; ES = expectorated sputum; Hill’s #, absolute Hill number; Hn = Hill number normalized for number of samples (shown in each panel); IS = induced sputum; NC = not collected; ND = assay not done; OP = oropharyngeal swab; S = saliva.

Figure 2.

Principal component analysis (PCA). The first two principal components explain 26% of the variability across samples. Pseudomonas is negatively associated with commensal taxa along PCA 1, whereas Staphylococcus is associated with PCA 2. Letters within the PCA plot indicate the subject from whom the sample was obtained and correspond to the letters in Figure 1. Samples from most individual subjects tend to cluster together. Sample types are displayed by color indicated in the key.

Figure 3.

Comparison of relative abundance (RA) between paired specimens. Box plots show the median difference in RA for prominent bacterial taxa and traditional cystic fibrosis–associated taxa detected by sequencing from expectorated sputum (ES) and induced (IS) sputum compared with paired oropharyngeal (OP) and salivary (S) samples. Samples were excluded if RA for the taxon was zero in both paired samples. Pseudomonas and Staphylococcus RAs were higher in lower airway samples compared with upper airways, whereas Veillonella, Rothia, and Streptococcus RAs were higher in upper airway samples. Enterobacteriaceae appears increased in lower airways but was not statistically significant. Boxes indicate median and 25 to 75% interquartile range. Diamonds indicate means; circles indicate values outside the interquartile range. Statistically significant differences are indicated by *P ≤ 0.05 and **P ≤ 0.01.

Figure 4.

Associations between bacterial communities, lung function, and inflammation. The heat map shows Spearman’s correlation coefficients for ecology metrics (Shannon diversity index, Evenness) and relative abundance (RA) of prominent bacterial taxa related to lung function (FEV₁) and airway inflammatory markers: neutrophil elastase (NE), IL-8, IL-6, TNF-α, vascular endothelial growth factor (VEGF), IL-1ra, and IL-1β. Red shades indicate positive correlations; blue shades indicate negative correlations. Positive (up) and negative (down) correlations are also indicated by the direction of the arrow symbol in the heat map. Correlations with a value of ≤|0.5| have a P value of ≤0.05.