Mapping Exposure-Induced Immune Effects: Connecting the Exposome and the Immunome
EXIMIOUS Newsletter 01 | April 2022
ENVIRONAGE and LIFELINES | The effects of our environment on the health of children and the development of immune diseases
Age-related diseases, such as cardiovascular diseases, dementia and diabetes, have their origins early in life. Environmental factors including air pollution, can already affect the health of the baby from conception onwards. ENVIRONAGE (Environmental Influences on Early Ageing) is a birth cohort that to date includes more than 2200 mother-child pairs in Belgium. The cohort was first established in 2010 in cooperation with the hospital Oost-Limburg (Belgium), to investigate the influence of environmental exposures during pregnancy and early life on the health of children. Within the EXIMIOUS consortium, researchers from the University of Hasselt will collect and reuse data from this cohort from the prenatal period, at age 4-6 and at age 9-12, to investigate the effect of genes and environment on the health of children and gain more insight into the underlying mechanisms of diseases. This allows the researchers to look for predictive markers of ageing and health linked to environmental exposures. For this cohort, pregnant women giving birth in the hospital Oost-Limburg are asked to participate in this study. The participating mothers first fill out a questionnaire, and after giving birth their placenta, urine and blood is collected. Finally, the child’s blood pressure is measured at the hospital. About 4-6 years later, these children are invited at the university for a follow-up study, wherefrom 650 children have participated until now. At the age of 9-12 years old these children are invited once more for a second follow-up, of which to this date, almost 100 children have already participated. During each of these follow-up visits the researchers try to gather more insight not only in terms of the health of these children, but also if they have been exposed to for example pollutants, medication or cigarettes’ smoke. The University of Hasselt aims to assess about 200 more children within the next 2 years. The health of these participants is assessed during these follow-up studies by performing multiple preclinical measurements, including blood pressure measurements, cognitive tests, bone density measurements but also by taking a blood and a urine sample.
Glossary:
Biobank: generally defined as a collection of human biological samples and associated information organized in a systematic way for research purposes (ScienceDirect, n.d.)
Black carbon: (BC) is the soot-like by-product of wildfires and fossil fuel consumption, able to be carried long distances via atmospheric transport (Elias, 2021).
Immunomic: All the genes and proteins that constitute the immune system are collectively known as the immunome; the immunome is a vastly complex and highly regulated structure that protects against infection and preserves health. (Biancotto & McCoy,2014)
Taking blood samples in children is challenging, but gives the researchers a very clear idea of the health of these children as well as where they are exposed to, is an ideal matrix to identify biomarkers and provides in depth information about molecular underlying mechanisms. Researchers at the University of Hasselt recently developed a novel method to estimate a child’s personal
exposure to black carbon (an important compound of air pollution) using blood or urine. As EXIMIOUS aims to shed light on the association between these children’s exposures and how the immune system works, the collection of blood and urine samples is essential. Lastly, extensive information about health, lifestyle and exposures is also collected by taking questionnaires from both child and mother each sampling time.
EXIMIOUS collects data within ENVIRONAGE but also uses data and biobanked samples collected during the last 12 years of this study to help find an answer to whether prenatal and postnatal environmental exposures relate to specific immunomic profiles early in life.
The University of Hasselt will also work on samples from adults and eldery within LIFELINES, a large biobank from the Netherlands. LIFELINES collects data and samples from 167.000 participants since 2006. Participants from three generations are followed for at least 30 years, to obtain insights into healthy ageing and the main factors relating to the onset and progression of diseases. EXIMIOUS will use LIFELINES samples (plasma, DNA and urine) together with data collected from these patients who have developed Rheumatoid Arthritis, Systemic Lupus Erythematosus and Type-I-diabetes. By measuring their exposures including black carbon exposure, EXIMIOUS aims to determine whether exposure to black carbon can predict the disease risk of these autoimmune diseases.
References:
Elias, S. (2021). Threats to the Arctic. Chapter 11 – Changes in Terrestrial Environments. 323-365. Elsevier.
https://doi.org/10.1016/B978-0-12-821555-5.00015-2.
ScienceDirect (n.d.) Biobank – an overview. Retrieved March 26, 2022, from https://www.sciencedirect.com/topics/nursing-and-health-professions/biobank
Biancotto, A., & McCoy, J. P. (2014). Studying the human immunome: the complexity of comprehensive leukocyte immunophenotyping. Current topics in microbiology and immunology, 377, 23–60. https://doi.org/10.1007/82_2013_336
Understanding work exposures: where AI and epidemiological analyses meet | Method development in the Danish DOC*X and DOC*X-Generation register-based cohorts
Danish registers provide an exceptional opportunity within health research. They are based on the unique personal identification number assigned to each Dane or person residing in Denmark. This number makes it possible to merge data from different registers on an individual level, as was done in construction of the two Danish cohorts included in the EXIMIOUS project: DOC*X and DOC*X-Generation. The cohorts pose the possibility to investigate a large range of common and rare work exposures over a timespan of more than 35 years. The main cohort is DOC*X, the Danish Occupational Cohort with eXposure data. The register includes data from a single survey in 1970 and register-based data continuously collected from 1976 onwards. The yearly registration
of employment status of each person (industry and job codes) serves as point of departure. Furthermore, each person is linked to health and administrative data, e.g., any contacts had with hospitals, education, income, and family members.
Glossary:
Epidemiology: the field that studies the distribution of disease in human populations and factors determining that distribution. It concerns itself with groups of people rather than individual patients (Mullner RM, 2020).
Job Exposure Matrix (JEM): A cross-tabulation between workplace hazards[/exposures] and occupational title (Choi, 2020). Each occupational title will receive a value of exposure. JEMs are commonly used to assign exposure level to a specific job group in occupational epidemiological studies, where no exposure data is accessible such as register-based studies.
Neural network: A complex computational system made up of “artificial neurons”, inspired by biology.
Register-based cohort: a collection (usually at national level) of information about individuals with a common trait.
The cohort includes more than million persons with a median occupational history of 15 years and between 2.0 to 2.9 million persons indicated as employed each year (Flachs et al, 2019). A range of different Job Exposure Matrices (JEMs) provide information on work exposures by link to the yearly job code assigned to each person. The matrices are constructed on job level, and each job is assigned a specific exposure value. Furthermore, four lifestyle JEMs have been developed. Based on four large population cohorts in Denmark, values of body mass index, smoking, alcohol consumption and fruit/vegetable intake can be assigned to each job code (Petersen et al, 2018).
The second cohort is the DOC*X-Generation cohort, which includes women from DOC*X who become mothers and their children. Data from this cohort analysed within EXIMIOUS will cover approximately the same time period as included in the DOC*X cohort. We expect the cohort to include more than 1.2 million pregnancies divided between more than 600,000 women. The same type of health and administrative data collected for DOC*X will also be retrieved from the registers for the DOC*XGeneration cohort and combined with JEMs.
In the two cohorts, we have the opportunity to investigate several work factors in relation to: i) the development of autoimmune disease after years of exposure within the workers themselves; or ii) exposure to work factors during pregnancy and the children’s risk of developing autoimmune diseases. Data will be analysed using machine learning and neural networks as well as epidemiological analyses.
Machine learning and neural networks
Modern artificial intelligence (AI) is evolving at a rapid pace, presenting researchers with new opportunities to handle large dataset, which traditionally have been hard to exploit. Within EXIMIOUS, experts in occupational health from the National Research Centre for the Working Environment (NFA) and Region Hovedstaden (RegionH), as well as AI developers from Biogenity and the Babraham Institute (BI), collaborate closely to develop new approaches to analyse large datasets. While rigorous statistical models don’t scale favourably to datasets like DOC*X, machine learning can deal with large and heterogenous data quite well. In addition, the strategy is centred on giving the models the tools to explain what information they have learned; this way, previous domain knowledge can be integrated with new insights from machine learning, in turn helping the execution of follow-up epidemiological analyses.
Epidemiological analyses
The results obtained by the machine learning and neural networks analyses, will inform hypotheses relating to the association between specific work exposures, either as a single or combined factor, and development of autoimmune disease. The hypotheses will be explored by NFA and RegionH using traditional epidemiological analyses. Furthermore, air pollution exposure modelled at the home address (by Aarhus University) will also be investigated, and we include external collaborators concerning JEMs (University of Utrecht). In the DOC*X cohort, cumulated exposure over time will be calculated for each work factor and investigated by regression analyses, alone and in combination, in relation to the risk of autoimmune disease later in life. In the DOC*X-Generation cohort, exposure during pregnancy to selected work factors will be investigated by regression analyses, also as a single factor or in combination, and their children’s risk of autoimmune disease during childhood.
EXIMIOUS represents a unique opportunity to bring data scientists, clinical experts, and institutions closer together, and allow us to better understand the interplay between established epidemiological methods and new technologies. This can open up exciting avenues for the future and generate new hypotheses to investigate for a better understanding, and ultimately prevention of, autoimmune diseases in society.
References:
Mullner RM. Epidemiology. In: Encyclopaedia Britannica. Publisher: Encyclopaedia Britannica. Date Published: 2020.04.23. https://www.britannica.com/science/epidemiology. Accessed 2022.03.09.
Choi BK. Developing a Job Exposure Matrix of Work Organization Hazards in the United States: A Review on Methodological Issues and Research Protocol. Safety and Health at Work 2020;11:397-404.
Flachs EM, Petersen SEB, Kolstad HA, Schlünssen V, Svendsen SW, Hansen J, Budtz-Jørgensen E, Andersen JH, Madsen IEH, Bonde JPE. Cohort Profile: DOC*X: a nationwide Danish occupational cohort with eXposure data – an open research resource. International Journal of Epidemiology 2019;48(5):1413-1413k.
Petersen SB, Flachs EM, Prescott EIB, Tjønneland A, Osler M, Andersen I, Juel K, Budtz-Jørgensen E, Kolstad HA, Schlünssen V, Bonde JP. Job-exposure matrices addressing lifestyle to be applied in register-based occupational health studies. Occupational & Environmental Medicine 2018;75:890-897.