A Refined View of Airway Microbiome in Chronic Obstructive Pulmonary Disease at Species and Strain-Levels
Journal Articles
Overview
Research
Identity
Additional Document Info
View All
Overview
abstract
Little is known about the underlying airway microbiome diversity in chronic obstructive pulmonary disease (COPD) at in-depth taxonomic levels. Here we present the first insights on the COPD airway microbiome at species and strain-levels. The full-length 16S rRNA gene was characterized from sputum in 98 COPD patients and 27 age-matched healthy controls, using the Pacific Biosciences sequencing platform. Individual species within the same genus exhibited reciprocal relationships with COPD and disease severity. Species dominant in health can be taken over by another species within the same genus but with potentially increasing pathogenicity in severe COPD patients. Ralstonia mannitolilytica, an opportunistic pathogen, was significantly increased in frequent exacerbators (fold-change = 4.94, FDR P = 0.005). There were distinct patterns of interaction between bacterial species and host inflammatory mediators according to neutrophilic or eosinophilic inflammations, two major airway inflammatory phenotypes in COPD. Haemophilus influenzae, Moraxella catarrhalis, Pseudomonas aeruginosa, and Neisseria meningitidis were associated with enhanced Th1, Th17 and pro-inflammatory mediators, while a group of seven species including Tropheryma whipplei were specifically associated with Th2 mediators related to eosinophilia. We developed an automated pipeline to assign strain-level taxonomy leveraging bacterial intra-genomic 16S allele frequency. Using this pipeline we further resolved three non-typeable H. influenzae strains PittEE, PittGG and 86-028NP with reasonable precision and uncovered strain-level variation related to airway inflammation. In particular, 86-028NP and PittGG strains exhibited inverse associations with Th2 chemokines CCL17 and CCL13, suggesting their abundances may inversely predict eosinophilic inflammation. A systematic comparison of 16S hypervariable regions indicated V1V3 instead of the commonly used V4 region was the best surrogate for airway microbiome. The full-length 16S data augmented the power of functional inference, which slightly better recapitulated the actual metagenomes. This led to the unique identification of butyrate-producing and nitrate reduction pathways as depleted in COPD. Our analysis uncovered finer-scale airway microbial diversity that was previously underappreciated, thus enabled a refined view of the airway microbiome in COPD.
