Profile feature - Caroline Relton and the ARIES programme
Combining epigenetics with epidemiology could better our understanding of ageing and disease processes. The ARIES programme looks to that future.
10 September 2012
The Accessible Resource for Integrated Epigenomics Studies (ARIES) programme is a two-year, £1.5M programme to investigate epigenetics – how DNA is modified so genes are turned on and off – and how such changes in gene expression are translated into the fundamental functioning of cells and tissues in the body.
ARIES is a joint venture between Newcastle University and the University of Bristol and draws on data from the ALSPAC (Avon Longitudinal Study of Parents and Children) birth-cohort study, which has followed mother-child pairs for 21 years and amassed a unique wealth of biological samples and clinical data.
What's special about the ALSPAC data that your ARIES project is using?
A major advantage is that it had the foresight to collect samples during mother's pregnancy, such as umbilical cord blood when children were born. After that samples were taken from children at various ages as they grew up. That's a huge asset and to my knowledge there are no other studies with that density of sampling over two decades.
What biological material are you sampling?
We'll be analysing DNA in the ALSPAC samples of peripheral blood (lymphocyte) DNA and umbilical cord blood DNA too. Ideally, you want to sample lots of different tissues, but it's not always possible with invasive biopsies, so we're able in parallel with ALSPAC to complement that data with clinical samples from other sources, such as blood and brain, and blood and liver, to get a better idea of how changes in blood DNA compare with DNA in tissues of interest.
What has ALSPAC taught us about epigenetics so far?
Because ALSPAC has got such a wealth of data collected over 21 years (with more planned) we can look at exposures during pregnancy and see how these impact on the epigenome. Then we can look at how epigenetic changes predict certain characteristics both in terms of normal development and disease.
For example, research done so far has shown us how exposure to air pollution in early years can change epigenetic markers at age seven. Other work shows there are epigenetic signatures in umbilical cord blood DNA that are related to body composition factors in childhood, such as height and BMI (body-mass index).
Initially we are going to try to identify differentially methylated regions (DNA expression is modified by addition of methyl groups, called methylation) – areas of the genome that are different between for example birth and childhood, mothers and their children or between diseased and normal (e.g. asthma or eczema and unaffected individuals) people – then look in more detail and use bioinformatics tools to look at the biological pathways involved.
So you are combining epigenetics with classic epidemiology?
It's a field that's just emerging. ALSPAC is one study that has been fortunate enough to get funding to catalyse the field of what we would call epigenetic epidemiology. And we're in the process of producing a huge amount of data to generate a whole range of hypotheses from early life influences on methylation to later life influences associated with things such as mental health, fatty liver disease and bone density.
How is the ARIES programme structured?
There are three arms. First, we're doing genome-wide Illumina methylation array analysis of 1000 samples of ALSPAC children at three time points and their mums at two time points; then sequencing the whole methylome (methylation across a whole genome) of 10 mother-child pairs.
What DNA sequencing technology are you using?
We are using the latest release of the Illumina array, the HumanMethylation450 BeadChip. It's the most cost-effective way of profiling the numbers of individual samples we have. We will be running a minimum of 5000 arrays – no other technique can give that coverage at that cost – at around $200 per sample.
What's special about this sequencing technique?
The Illumina array measures around 475,000 individual CpG sites (cytosine-phosphate-guanine methylation sites across the genome) which represent 1% of all known potential methylation sites; so it's an overview of whole genome but at single DNA base-pair resolution. It is possible to use this information in association with information on whether these sites are co-located with transcription binding (where protein-making enzymes attach to DNA) and how they control gene expression by making transcription binding sites accessible or inaccessible. We are also interested in areas that are not obvious zones for gene regulation.
You mentioned three arms to ARIES...
Second, we're doing human tissue profiling using the same Illumina platform where we also run an expression array which gives a genome-wide snapshot of gene expression, whether a gene is active or suppressed.
We are profiling tissues because epigenetic patterns are tissue specific, so although we learn a lot from peripheral blood DNA, some people think there may be some limitations to that because peripheral blood may not accurately reflect signatures in all tissues.
And the third strand?
Bioinformatics to glue it all together. We are building a custom data integration system to take in experimental data and integrate it with publically available data on genotypes, gene expression and literature mining, so we'll be able to browse and visualise areas of the genome we're interested in, for example how methylation patterns change over time, and how they relate to other tissues in the body.
It's building on the Ondex project, a previous BBSRC-funded project, which takes bioinformatics know-how developed in collaboration with Rothamsted Research because of their knowledge and expertise developing that system (for managing data from diverse and heterogeneous datasets).
Have you specific collaborations to share the data?
The data we generate will be publically available but we have also developed specific collaborations. We have a BBSRC Brazil Partnering Award, and we're already working with the Federal University of Pelotas, Brazil, as they are keen to collaborate in this area. We will be delivering a course on translating the epigenetic tools and technologies we have developed here in the UK to build capacity in epigenetics research in Brazil.
And it's a similar picture in Canada. We have collaboration with the University of Alberta. They are particularly interested in perinatal health and want to feed off our tools and the techniques we're developing and apply them in a collaborative way in their environment. Scientists from their institution plan to visit us to see how we're setting up our bioinformatics systems.
What else might the information do?
There is so much scope to exploit the data as we are adding genome-wide DNA methylation to a study that already has a huge wealth of information. We can align epigenetic data with a wide range of data from ALSPAC and with the UK10K project (sequencing the DNA of 10,000 people in the UK) to provide more insight into the more general patterns of DNA methylation, and how this relates to genetic variation, development and disease.
What can epigenetics tell us about healthy ageing and disease processes?
There's a huge interest in the epigenetics of certain diseases associated with environmental conditions, as well as those with established early life drivers including psychiatric illness such as depression, schizophrenia and autism. There is also a lot of interest in contributing factors to cardiovascular disease risk, metabolic perturbations, diabetes and metabolic syndrome.
Epigenetics may also play a major part in congenital anomalies such as cleft palate and neural tube defects. These conditions all have a strong link with folate metabolism, and folate provides the methyl groups added to DNA during methylation. You get such profound changes in methylation immediately post-conception it makes complete sense that if you don't have enough folate around then it will alter your ability to reset and reprogramme the genome. I would not be in the least surprised to find that there is a role for epigenetics in these congenital disorders.
What first attracted you to the epigenetics field?
It's new, and it's a new dimension to genetics. I'm really interested in it as at the complex interface of genetics and the environment. My background is in both molecular genetics and epidemiology.
What can epidemiology bring to epigenetics?
Epidemiology provides approaches to uncover associations between exposures and outcomes. Epigenetics is one mechanism that might explain the connection in any given association. Epidemiology, including genetic epidemiology, can provide really useful tools for the field of epigenetics that can help for example to decipher whether epigenetic changes are a cause or a consequence of a certain disease.
What will epigenetics have taught us in 20 years?
I believe that it's going to highlight new pathways to disease. It's going to give us much more insight into how the environment and our genome work together to influence normal development, and disease too.
When did you first become interested in science?
At school I always loved biology, especially as a teenager. Then I taught science in high school and after five years I went back to university do a PhD. I think that route to an academic career really suited me and what I learnt teaching has helped me teach undergraduate and postgraduate students too, as even in a research intensive job, teaching is a very important aspect of what we do.
Actually, I have an identical twin sister who is a scientist too; she works for a biotech company in Cambridge, USA.
So you could get your methylomes sequenced and compare them…
We keep talking about volunteering for a twin study but we're rarely in the same country!