By Jonathan Thorsen
Chariman: Torben Hansen
Opponent: Debby Bogaert
Opponent: Tine Rask Lich
Asthma is one of the most common chronic disorders in childhood and affects hundreds of millions of people worldwide. While several risk genes and environmental risk factors have been identified, the etiological mechanisms leading to asthma are still not understood. Treatment options focus on counteracting airway inflammation and bronchoconstriction, but no effective means of prevention exist.
The human microbiome has been implicated in the development of many immune-mediated diseases through production of immunoactive metabolic products and direct cellular microbe- host interactions. The airway microbiome, though much less studied than the gut microbiome, is becoming increasingly recognized as a relevant niche with potential effects on host inflammation, both before disease evolves and as trigger of infections and acute asthmatic symptoms.
This PhD thesis is based on data from the Copenhagen Prospective Studies on Asthma in Childhood 2010 (COPSAC2010) cohort of 700 mother-child pairs, recruited during pregnancy week 24 and monitored through childhood with frequent detailed clinical visits to the COPSAC study clinic. The primary end-points for the cohort are development of asthma, allergy and eczema, which are all prospectively diagnosed using pre-defined quantitative criteria. The airway microbiome of the children was examined using 16S rRNA gene amplicon sequencing of hypopharyngeal aspirates collected longitudinally from the age of 1 week. The thesis contains three scientific papers.
In Paper I, we examined the association between the airway microbiota in the first three months of life and the development of asthma by age six years. We found that healthy infants who at 1 month of age had high microbial diversity in the airways and high relative abundances of the taxa
Veillonella and Prevotella, among others, had a substantially increased risk of developing asthma later in childhood. We also found that these children had an altered local airway immune profile at 1 month of age, which also independently predisposed to asthma. We speculate that these findings may signify an early-life interaction between the resident bacteria and the immune system that leads to an inflammation-prone immune trajectory later in childhood.
In Paper II, we focused on the children who developed asthmatic symptoms between 12 and 36 months of age. The children participated in a randomized controlled trial of azithromycin vs. placebo for the treatment of acute asthmatic episodes and were monitored using structured symptom diaries. We previously reported that azithromycin reduced episode duration by more than 40%, but the effect was unrelated to identification of cultured pathogenic bacteria. In this study, we found that children with high microbial diversity and high relative abundances of several taxa in the airways, including Neisseria and Veillonella, had longer asthmatic episodes. This was counteracted by azithromycin treatment, indicating that antimicrobial effects on some of these bacteria may contribute to the observed treatment effect. Intriguingly, this phenomenon seemed to be carried by a latent community structure rather than single pathogens.
In Paper III, we critically examined the features of 16S rRNA gene sequencing datasets and the statistical challenges they raise. We benchmarked the most common statistical methods applied in studies testing Operational Taxonomic Units from 16S amplicons and found that many of these methods had high false positive rates, likely due to statistical assumptions not met by the data. This phenomenon was exacerbated with increasing dataset size for some of the statistical methods, since sparsity increases with number of samples. We furthermore conducted in silicospike-ins of signals and measured the relative performance between methods in distinguishing spiked vs non-spiked OTUs, and showed that some of the statistical methods needed much stronger signals to correctly identify the spike-ins than others. Finally, we showed the relative effects of normalization, count transformation and choice of distance metric in distinguishing between groups in beta-diversity analysis using datasets from the COPSAC2010 cohort in addition to publicly available data from other studies.
In conclusion: Collectively, these finding shed light on the relevance of the airway microbiome for asthma in childhood, both in early life, before disease onset, and during acute episodes in children with recurrent asthmatic symptoms. In the future, strategies for manipulating the airway microbiota may help prevent and alleviate asthma in childhood.