HDHL-Biomarkers: Early life programming of childhood health: a nutritional and epigenetic investigation (ALPHABET)

Lead Research Organisation: University of Bristol


The rising prevalence of obesity, cardiometabolic disease, asthma, osteoporosis and neurodevelopmental disorders over recent decades cannot be fully explained by genetic or adult lifestyle factors. Increasing evidence suggests links to early life environmental factors. If such links exist, then there must be non-genetic mechanisms that retain memory of early life exposures such as suboptimal nutrition during pregnancy or early childhood and that influence health decades later. DNA methylation, the addition of methyl groups to DNA, has been proposed as a possible mechanism because it is responsive to the environment and influences gene expression.

Maternal diet during pregnancy is a modifiable behaviour that could impact on maternal, neonatal and child health. However the specific dietary requirements during pregnancy for optimal fetal growth and development are unknown in large part because effects on health only begin to appear decades after exposure. DNA methylation patterns however should be affected almost immediately and, if they are linked to later health outcomes, should remain throughout life. This project will utilise biological samples and data from existing European longitudinal birth cohorts to investigate the complex relationships between maternal diet summarised by indices of diet quality and inflammatory potential, health outcomes and DNA methylation patterns from birth throughout childhood and into adulthood. Hence, we will be able to observe how prenatal diet modifies infant DNA methylation patterns and if and how these modifications are maintained throughout life and relate to health outcomes in adulthood. These observations may aid development of more refined dietary recommendations for promoting optimal health throughout life.

Technical Summary

DNA methylation is known to be associated with a wide variety of exposures, phenotypes and health outcomes. Animal models have provided evidence that in many cases DNA methylation mediates the effects of exposures such as diet on health outcomes. Direct evidence for similar mediating relationships in humans is sparse partly due to ethical limitations that restrict experimental approaches and analysis of the most relevant tissues. To address these limitations, alternatives to experimentation have been proposed. Structural equation modelling has been used to determine how well causal network models fit observational data, and instrumental variable analysis has been used to interrogate causal relationships. Mendelian randomisation (MR), an instrumental variable approach based on genetic variants, has been used to successfully discriminate between causal and non-causal relationships in health research that were later confirmed by randomised control trials (RCTs). We propose here to apply structural equation modelling to evaluate evidence that the effects of prenatal diet on later health and developmental outcomes are mediated by DNA methylation and to apply Mendelian randomisation to assess the causality of such relationships. In addition to adiposity, cardiometabolic, bone, respiratory and neurodevelopmental health.

We will 1) test associations between prenatal diet as summarised by DII and DASH scores and blood DNA methylation pattern, 2) determine the extent to which these associations change throughout childhood, 3) test associations in adulthood between diet as summarised by DII and DASH scores and blood DNA methylation patterns, 4) test associations between health outcomes (adiposity and cardiometabolic, bone, respiratory and neurodevelopmental health) and blood DNA methylation patterns, and 5) examine evidence that DNA methylation may be on the causal pathway between diet and health outcomes using structural equation modelling and Mendelian randomisation.

Planned Impact

The results of this research programme may have the potential to inform health relevant policy, clinical practice, patient care and health behaviour. This may be achieved through generating evidence for policy and production of, or at least citation of, our findings in evidence based clinical or public health guidelines or policy documents. It is anticipated that our findings will highlight areas of action for policymakers with responsibility for addressing maternal and childhood obesity (key areas already identified both at European and International level). Thus our findings may aid development of new public health strategies for prevention, with an emphasis on early development. For example data relating to nutritional intake during pregnancy may inform revision of existing National guidelines or development of new guidelines or policies for optimal nutrition in pregnancy, which may reduce adverse health outcomes in future generations.

The potential health and socio-economic impacts may be significant considering the health service and societal costs associated with obesity alone, currently and for future generations. Our results will improve understanding of the relationships between maternal nutrition, pregnancy outcome and subsequent child health. For example improved dietary quality and/or reduced pro-inflammatory dietary status achieved through revision of existing guidelines or development of new guidelines for optimal nutrition in pregnancy or development of functional foods may lead to beneficial health effects in women of reproductive years and their offspring which may have long term effects on their health into adulthood.


10 25 50

Description We have strengthened the evidence that a pro-inflammatory diet during pregnancy, for example diets high in red meat and refined sugar, may reduce fetal growth and increase the risk that the child will be overweight and experience difficult emotions and exhibit negative behaviours. This evidence may contribute to refining dietary recommendations for pregnant mothers aimed at improving child health. Obtaining these results required surmounting the major challenge of harmonizing diet and health information across multiple cohort studies which had collected this information in different ways. This harmonization and the resulting dietary scores can now be used for future analyses to improve our understanding of the links between diet and health.
Exploitation Route Papers have been published using the open access model and are publicly available to other researchers. Our results may have public health implications in terms of refining dietary exposure measures and informing the development of more effective evidence-based public health strategies specifically related to pre-pregnancy, pregnancy, and early postnatal life. For example, we provide evidence that a pro-inflammatory diet during pregnancy may negatively impact fetal growth and increase offspring childhood risk of obesity and negative emotional and behavioural symptoms. Data across multiple cohort studies have been collated and harmonised and dietary scores capturing the extent to which diet influences inflammation and hypertension have been generated which will be of use for future analyses. New genome-wide DNA methylation datasets have been generated and will also be of use for future analyses. A new method was developed for identifying associations with improved statistical power.
Sectors Healthcare

Description ALPHABET 
Organisation University College Dublin
Department School of Public Health, Physiotherapy and Population Science
Country Ireland 
Sector Academic/University 
PI Contribution I am a co-investigator on a ERA-HDHL (BBSRC) collaborative research project on the influence of early life nutrition on offspring health led by UCD. Our contribution includes collating multi-centre epigenetic data, running epigenome-wide association studies, meta-analyses and causal analysis methods.
Collaborator Contribution The partners in this collaboration contribute data from various large birth cohort studies and expertise in nutrition and other clinical disciplines
Impact Aubert A, Forhan A, de Lauzon-Guillain B, Chen LW, Polanska K, Hanke W, Mensink-Bout SM, Duijts L, Suderman M, Relton CL, Crozier SR, Harvey NC, Cooper C, McAuliffe FM, Kelleher CC, Philips CM, Heude B, Bernard JY. Deriving the Dietary Approaches to stop Hypertension (DASH) score in women from seven pregnancy cohorts from the European ALPHABET consortium. Nutrients. 2019 Nov 8;11(11). pii: E2706. doi: 10.3390/nu11112706. PMID:31717283. Chen LW, Aubert AM, Shivappa N, Bernard JY, Mensink-Bout SM, Geraghty AA, Mehegan J, Suderman M, Polanska K, Hanke W, Trafalska E, Relton CL, Crozier SR, Harvey NC, Cooper C, Duijts L, Heude B, Hébert JR, McAuliffe FM, Kelleher CC, Phillips CM. Associations of maternal dietary inflammatory potential and quality with offspring birth outcomes: An individual participant data pooled analysis of 7 European cohorts in the ALPHABET consortium. PLoS Med. 2021 Jan 21;18(1):e1003491. doi: 10.1371/journal.pmed.1003491. eCollection 2021 Jan. PMID: 33476335. Chen LW, Aubert AM, Shivappa N, Bernard JY, Mensink-Bout SM, Geraghty AA, Mehegan J, Suderman M, Polanska K, Hanke W, Jankowska A, Relton CL, Crozier SR, Harvey NC, Cooper C, Hanson M, Godfrey KM, Gaillard R, Duijts L, Heude B, Hébert JR, McAuliffe FM, Kelleher CC, Phillips CM. Maternal dietary quality, inflammatory potential and childhood adiposity: an individual participant data pooled analysis of seven European cohorts in the alphabet consortium. BMC Med. 2021 Feb 22;19(1):33. doi: 10.1186/s12916-021-01908-7. PMID: 33612114.
Start Year 2017
Description MRC The Gambia 
Organisation Medical Research Council (MRC)
Department MRC Unit, The Gambia
Country Gambia 
Sector Public 
PI Contribution Collaboration in epigenetics research. Co-investigator on recently funded award to establish a W African Bioresource in nutritional genetics and epigenetics.
Collaborator Contribution Myself and members of my team have committed to helping our colleagues at the MRC Gambia Unit establish a bioresource to support molecular epidemiology research in the future. This will involve researcher exchanges,workshops and training in Bristol and the provision of ongoing support and capacity building in The Gambia.
Impact Capacity building achieved through sharing of good practice.
Start Year 2018
Title Dmrff 
Description Identifies associations in genome-wide DNA methylation datasets with improved power while controlling error rates 
Type Of Technology Software 
Year Produced 2019 
Open Source License? Yes  
Impact Used to analyse data in 14 published studies 
URL https://github.com/perishky/dmrff
Description CAPICE workshop session: "Introduction to EWAS" 
Form Of Engagement Activity A talk or presentation
Part Of Official Scheme? No
Geographic Reach National
Primary Audience Postgraduate students
Results and Impact I presented a workshop session on preparing and analysing genome-wide DNA methylation datasets for associations with phenotypes and exposures. The audience included postgraduate students and postdoctoral researchers attending the 2-day "CAPICE workshop on Mendelian Randomization and Epigenetics".
Year(s) Of Engagement Activity 2019
URL https://www.capice-project.eu/
Description Society for Epidemiological Research Annual Meeting workshop: An introduction to predicting exposures and outcomes using omic biomarkers 
Form Of Engagement Activity Participation in an activity, workshop or similar
Part Of Official Scheme? No
Geographic Reach International
Primary Audience Professional Practitioners
Results and Impact Our workshop provided a primer on the essential considerations and evaluation procedures required when developing predictors of health-related exposures and outcomes using omic datasets. We offered context on how such predictors have been developed in the past, as well as common pitfalls. We introduced how several machine learning techniques can be harnessed to design generalizable and interpretable predictors.
Year(s) Of Engagement Activity 2019
URL https://epiresearch.org/annual-meeting/2019-meeting/