The bacterial and fungal communities in the airways of adults with asthma and eosinophilic lung diseases

Abstract

Asthma is the most common long-term respiratory condition, but its aetiology is not fully understood. The incidence has increased alongside urbanisation and altered microbial exposure patterns. The respiratory tract’s bacterial communities (microbiota) undergo dysbiosis in asthma. Eosinophilic inflammation, common in asthma, also occurs in fungal infections and eosinophilic lung diseases (ELDs) including Chronic Eosinophilic Pneumonia (CEP). The CEP microbiome has not yet been described and there are few studies of respiratory fungal communities (mycobiota) in health or disease. The main objective of this thesis was to describe the respiratory microbiota and mycobiota in asthma, ELDs and healthy controls, through two cross-sectional bronchoscopy studies: 1) A multi-centre study of asthmatics and controls (Celtic Fire); 2) A single-centre study of subjects with ELDs and controls (Bronchoscopic evaluation of the Eosinophilic Airway Microbiome, BEAM). Oropharyngeal (throat) swab (OTS) samples (N = 130) and left lower lobe (LLL) brushes (N = 131) were obtained from 131 subjects (Celtic Fire: 66 asthmatics, 44 controls; BEAM: 11 subjects with ELDs, 10 controls). In Celtic Fire additional brushings were obtained when feasible. Samples underwent DNA extraction. Bacterial DNA was quantified and sequenced using 16S rRNA gene qPCR and 16S sequencing (Illumina MiSeq) respectively. Selected samples underwent comparable processes regarding fungi: 18S rRNA gene qPCR and ITS2 sequencing. Sequences underwent pre-processing using Quantitative Insights Into Microbial Ecology and analysis in R. Within individuals, left upper lobe (LUL) and LLL brushes were highly similar. In both studies OTS samples contained c. 100 times more bacterial and fungal DNA than LLL brushes. Upper respiratory tract (URT) bacterial communities were similar to the lower respiratory tract (LRT), but fungal sequencing results varied greatly between URT and LRT. Numerous operational taxonomic units (OTUs) were differentially abundant according to case-control status and other clinical variables e.g. sex, corticosteroid use and smoking history. Analyses have provided insights into relationships between microbes, host and disease.