What are challenges and perspectives in gene-environment interactions in asthma and allergies? - Kauffmann and Demenais, 2012 - Article

Summary with the article: What are challenges and perspectives in gene-environment interactions in asthma and allergies? - Kauffmann and Demenais, 2012

Asthma and allergy are diseases that develop through genetic and environmental effects and their interactions. In the past most research focused on genes and the environment was seen as less important. However nowadays the focus lies more on research into gene x environment (GxE) interactions and developing new methods to identify those. However, it remains a challenge to understand the GxE interactions of complex diseases.

Gene-environment interaction can be a statistical interaction, which are all effects that are not pure main effects. Qualitative interactions require interaction terms on any measurement scale, but quantitative interactions are scale dependent. Biological interaction is the dependence of the effect of one factor at cellular or molecular level on the presence or absence of the other. Synergy means that the joint effect of two factors is greater than the sum of the excess risks from both factors.

With studying GxE interactions in asthma and allergy, we are able to identify novel genes and environmental risk factors having an effect in only a subgroup of subjects. It can also help to increase our knowledge of complex biological mechanisms and pathways that underlie the disease. The study into relations between asthma and allergies and genes and environmental factors is important, because the prevalence of these diseases has increased dramatically. It might also help us develop new drugs. However research has so far been able to identify only few genes involved in asthma.

What is known about GxE interactions for asthma?

With candidate interaction the gene and environment chosen are assumed to be involved in a physical or chemical reaction. Often studied in asthma is the CD14 gene, the HLA genes and Toll-like receptor genes. There are also candidate interactions that correspond to lower levels of knowledge than mechanical interactions. Without a priori knowledge on the interaction between genes and environment, the level of complexity to detect the GxE interactions depends on whether the gene and environment are both known, or only is known or neither is known. It also depends on if the studied phenotype is the whole disease or a disease-related endo phenotype or sub phenotype.

An example in which both the gene and environment that have an effect on the disease are known is the interaction between the 17q21 locus and exposure to ETS in early life in people with early-onset asthma. Only in early-onset asthma there is an increased risk of asthma by 17q21 variants, which is heightened by ETS exposure in early life.

Limited candidate regions have been found when the environment was known. In the candidate gene studies researchers mainly look at ETS exposure, air pollution, farming and occupational exposures. The difficulties in finding GxE interactions might be because of the small number of genetic polymorphisms investigated and the small sample sizes. A way to simplify the study of GxE interactions at genome-wide level when the environment is known is to conduct a GWAS of environment-related sub phenotypes. In this way the potential association of the catenin α3 gene with diisocyanate-induced asthma was found.

There have been genome-wide environment interaction studies for asthma to explore the interactions between farm-related exposure and genome-wide SNPs. Interactions with rarer SNPs were found, but this still has to be replicated.

When both the gene and environment are unknown there are few possibilities for research. Most well-validated GxE interactions were found by using knowledge of biological processes, effect of environmental exposures and/or pathophysiologic mechanisms. They are now trying to improve the agnostic approaches of GxE interactions.

What challenges are there in detecting GxE interactions and possible solutions?

The authors did a literature search on asthma and GxE interactions. The environmental factors that were often studied were smoking, farming, indoor environment, occupational environment, air pollution and macro environment. None of the 11 articles they looked at showed a very clear interaction and all had different interpretations of this. Therefore we need to develop a systematic manner to identify the issues raised by GxE interactions and propose solutions.

Asthma is a heterogeneous disease and there are several variants. More associations between SNPs with a complex disease were found using GWAs for well-defined diseases than for less well-defined diseases that affected multiple organ systems. Improving the phenotype definition could be done by taking into account the physiologic and biological phenotypes of the disease process. We know that phenotypes involved in allergy, inflammation and airway remodelling are associated with asthma, which have been explored separately, but these studies should be expanded.

There have been attempts to define sub phenotypes of asthma based on combinations of clinical characteristics. Today we try to include biology. By using unbiased and statistically-based approaches we can discover new phenotypes from a number of features that are related to the disease occurrence, evolution, therapy, and biological and physiologic disease-related phenotypes. Time is an important factor in the pathophysiology of asthma and therefore we could look at the gene, environment and time interactions. It might be important to consider the time of asthma onset in the GxE interactions.

We can get more statistical power to identify genes and GxE interactions when doing a joint analysis of multiple phenotypes. It is often necessary to combine data to get an adequate sample size. Methods to do this have to be improved. Another challenge is the lack of uniformity in methods to collect phenotypic and environmental data across studies.

The use of biomarkers in GxE interaction studies is limited because of several problems: the availability of the appropriate markers for the question under study, the availability of the environmental or biological samples to do the measurements and that they often only reflect recent exposure. What we want to study are the chronic exposures. The time of exposure might be more important than the dose, because it could explain the lung growth, immunologic maturation, epigenetic marks and asthma.

For the population, a cohort effect of exposure could be useful. Possible solutions could be long-term longitudinal studies and studies with subjects of different ages, including retrospective information on the childhood environment, the macro environment and international studies.

When doing research, other exposures than just the environmental factor under the study should be accounted for. It could, for instance, be that the effect of an environmental factor is prevented by another one. A way to avoid this problem is a pathway-based approach, which takes several genes and environmental factors related to the biological pathways into account. However, difficulties can still remain. There could be confounding effects when spatially clustering air pollution, temperature, and social determinants, such as the geographic area the people live in. With using multilevel modelling you can overcome this issue.

Specifically interesting are the methods that incorporate previous knowledge. Social determinants of environmental exposures can occur at the collective level and at the individual level. Therefore these social determinants should be taken into account with stratified analyses, multilevel modelling and causal models.

Another issue is that social determinants might result in non-independence between genes and environment in the population, such as for smoking dependence. This is influenced by both genes and social determinants at individual and population levels. By using appropriate study designs and analytic methods you can overcome the complications for the analysis of GxE interactions. Selection biases could also occur. In studies in families you should take into account the shared environment. Consideration is also necessary for assortative mating, which is related to risk factors of asthma.

Until today most GxE interaction studies examine single genetic markers. However, the joint analyses of multiple genetic markers within a locus by a GWAS have been promising. Potential functional genes in GWAS loci can also be discovered by combining the data from GWASs of disease and gene expression levels. There is also more focus nowadays on association signals that are not significant by themselves, but might become significant when combined. The methods to do this depend on the study design, analysis strategy, use of biological knowledge and statistical methodology. Using pathway-based analysis for GxE interactions might be promising.

To add more complexity, genes can also interact with other genes. Not only SNPs can be involved, but also copy number variants (CNVs) can interact with environmental factors. The methods to determine CNV location and copy number from SNP array data will have to be improved, however.

A major issue in studies into GxE interaction is statistical power. To get the same magnitude for an interaction effect as for a main effect, the sample size needs to be four times as large. Increased power is possible by enriching your sample for subjects with a family history of disease or one having extreme phenotypes or extreme environmental exposures. It could also be an option to use a ‘cocktail’ method, which is a method that combines the most appealing aspects from different methods.

How are new knowledge and tools integrated?

Epigenetics might play a big role in gene regulation and might be implicated in the cause of many diseases. It plays a role in T-cell differentiation and regulation and environmental factors associated with asthma induce epigenetic changes in experimental settings. With new technologies we can explore the epigenetic alterations. It makes it possible to not only identify genetic and environmental components of DNA methylation patterns, but also to understand how these methylation patterns mediate the effect of known environmental factors and GxE interactions on asthma-related phenotypes. The gut microbiome influences the development of inflammatory disease, and the composition of this gut microbiome, the airway microbiome or both might be important to determine the risk of asthma development and could contribute to chronic asthma. The gut and airway microbiota could be modified by environmental exposures. Therefore it could be interesting to do more research into the microbiota.

There need to be new approaches to integrate various molecular data and link them to the disease process. Interactions between the genome, the proteome, the environment and multiple disease-related phenotypes could give insight into the complex mechanisms underlying the disease process and to progress toward personalised intervention. With the paradigm of ‘think globally, act locally’ new drug targets and biomarkers could be identified and local interventions to cure diseases could be generated.

What can be concluded from this study?

Many environmental exposures associated with asthma have been discovered in epidemiological studies. Genetic studies have identified genetic loci associated with asthma. Amongst others, the window of exposure and the window of expression of the disease over the lifespan are important. There have only been few GxE interactions found. There have also been few studies on the joint effects of environmental factors, genetic factors or both. Interactions among genes, environmental factors and GxE interactions are poorly documented.