By Eskil Kreiner-Møller, MD, PhD
Chairman: Torben Hansen
Opponent: Charlotte Brasch-Andersen
Opponent: John Henderson
Asthma is a common heritable disease of the airways with recurrent episodes of symptoms and reversible airflow obstruction that has increased dramatically in prevalence. The disease is highly heterogeneous with varying age at onset and clinical presentation and most likely represents several different subtypes of disease associated with distinct clinical features, divergent underlying molecular mechanisms, and individual treatment responses. Information obtained from genetic studies may be an important way of understanding underlying disease subtypes. Genome wide association studies (GWAS) have enabled the discovery of loci associated with disease, often by meta-analysis of several different studies. This is done by applying general and broad disease definitions. Subsequent follow-up studies are needed that investigate the relationship between the novel findings and specific disease phenotypes, related traits and objective measures in order to disentangle the underlying pathophysiological disease mechanisms for the subtypes of disease.
Several genes and loci have been found to be associated with adult lung function in GWAS, but it is currently unknown at what time in life these genes exert their effect. In paper I, we investigated the known effect of adult lung function loci on the development of lung function and bronchial responsiveness in children from birth until 7 years of age in the COPSAC2000 birth cohort of 411 children. We measured lung function and bronchial responsiveness at one month of age using the raised volume rapid thoracoabdominal compression technique in sedated neonates and repeated the measurements at age 7 utilizing traditional spirometry assessments. Lung function genetic variants identified in adults were not associated with neonatal lung function or bronchial responsiveness, but were associated with lung function growth. This suggests that these loci do not exert their effects prenatally and indicate a potential window of opportunity in early childhood for preventing lung function decline and maintaining respiratory health.
The 17q21 locus harbors common variants that are most robustly associated with persistent and severe childhood onset asthma. However, it is unknown how the 17q21 common variants are associated with the persistence of symptoms into adulthood. In paper II,we investigated the effect of the 17q21 locus on current adult asthma and related traits in the Health2006 study and replicated findings in the B58C birth cohort in a pooled analysis including more than 10,000 individuals. The 17q21 locus was not associated with current adult asthma. This contrasts with findings in children and suggests that this locus is associated with a specific transient asthma subtype with severe symptoms in childhood that do not persist into adulthood.
The GWAS have given valuable insight in many disease processes, but the common genetic variants discovered so far in general confer a very low risk of disease on an individual level. Low frequency variants are speculated to be able to have larger effects on disease development. These variants can be imputed in existing data instead of expensive re-genotyping by using publicly available reference data (a group of individuals genotyped densely). This method is based on an attempt to estimate the genotype at the individual level, using the haplotype structure in the reference data. However, imputation of low frequency markers is problematic and often results in poor quality. In paper III, we developed a new method for improving the imputation of low frequency markers using data from the Rotterdam studies. First we imputed the genome-wide data using a local reference population and took these intermediate data forward and imputed them using a publicly available reference sample. This two-step approach increased imputation quality significantly, especially in the low frequency spectrum.
In conclusion, we observed an age-dependent association between specific genetic loci and asthma and lung function development. These findings suggest that the investigated variants can distinguish subtypes of asthma and lung function trajectories to some degree and that the results can help to identify periods in early life as a possible window of opportunity for targeted intervention. We finally developed a two-step strategy for improving the imputation quality of low frequency variants in large existing genome-wide datasets.