The Peri/postnatal Epigenetic Twins Study (PETS) is a unique cohort of 250 mothers and their twins. Women were recruited from three Melbourne hospitals in mid pregnancy, which enabled measurement of maternal and fetal factors at multiple time points and minimization of recall bias. We collected multiple biological specimens at birth (cord blood, cords, placenta and buccal tissue) and collected samples of blood and buccals when infants were 18 months of age. The study is ongoing.
The main aim of PETS is to study the plasticity of epigenetic marks and the genes they control during the intrauterine period and in early childhood. We also aim to tease out the genetic, common and unique environmental factors that influence the epigenome. A further aim is to focus on the early life determinants of cardiovascular health.
Using methylation analysis on a locus-specific and genome-wide scale, we have shown that monozygotic (identical) and dizygotic (fraternal) twins can exhibit a wide range of epigenetic discordance at birth. We found that the contribution of genetic variation to epigenetic variation was 5-12%, varying by tissue, which agrees with other studies and is an important consideration in epigenome-wide association studies.
We found strong relationships between shared maternal environments such as gestational diabetes with DNA methylation in all tissues at birth. We also found tissue-specific and region-specific associations of methylation with shared environment, implying that in many cases, findings from one tissue will not be able to be extrapolated others. On a genome scale, imprinted genes and genes involved in metabolic pathways, principally protein metabolism, were associated with multiple maternal environments. However, the largest component of epigenetic variation at birth was nonshared environment – the environment that, due to twins having their own umbilical cord and placenta (or share of a placenta), is specific to each twin. Genes associated with response to environment were most differentially methylated within pairs. Size and morphology of the cord and placenta had significant associations with DNA methylation at birth. In addition, differences in DNA methylation and expression of genes associated with metabolism and cardiovascular development were associated with birth weight discordance, providing a possible mechanistic link between birth weight and complex disease in later life.
Moving to the first 18 months of postnatal life, we found that one third of the genome is epigenetically dynamic during this period and that genes involved in development, differentiation and aging were more likely to change during this time, suggesting a continuum of epigenetic change from early development to aging. Twin pairs also showed a wide range of within-pair dynamics during this time, exhibiting not also epigenetic drift but also convergence, possible because of a discrepancy between discordant prenatal and concordant postnatal environment.
In summary, twins have added an extra dimension with which to understand the plasticity and dynamics of early human development. As twins do not differ significantly with singletons in health outcomes, we expect our finding will be more widely applicable.