Work

Cohort Profile: The Peri/post-natal Epigenetic Twins Study

Year: 2011
Type: article
Abstract: Although cells of an individual generally all have the same DNA, each cell type has a unique pattern of active (expressed) and inactive genes. These patterns are established or ‘programmed’ during early development by epigenetic modifications involving small molecules that bind to specific sites in DNA or DNA-packaging proteins (reviewed in Foley et al.1). The most stable and easily measured epigenetic mark is DNA methylation, which occurs at cytosine–guanine (CG) sites. Combinations of epigenetic marks can effectively act as gene-specific ‘volume controls’, turning expression up or down without changing the DNA sequence. Important properties of epigenetic marks are that they are stably passed on through cell division but can be influenced by environmental factors. Levels of epigenetic marks within the human genome change over an individual's life course. However, both animal and human studies suggest that the period of greatest subsequent epigenetic flux is during intrauterine life (reviewed in Foley et al.1). However, few studies have addressed the extent of epigenetic change that occurs thereafter. Age-associated epigenetic change has been observed in adults,2,3 with changes occurring in a subset of genes4,5 possibly under genetic control.3 Other studies have found evidence for ‘epigenetic drift’ within pairs of genetically identical monozygotic (MZ) twins,6,7 leading to increasing within-twin pair epigenetic discordance over time in a subset of genes. There is strong evidence that environmental stimuli during critical periods of pre- and immediate post-natal mammalian development can influence the developmental pathways, inducing permanent changes in metabolism and modulating chronic disease susceptibility. Much of the research in this field (Developmental Origins of Health and Disease)8,9 has focused on evidence associating low birthweight with an increased risk of cardiovascular and other complex diseases in adulthood.10,11 There is also evidence that specific environmental exposures in utero can influence epigenetically regulated developmental pathways. Most of these data come from experimental animal studies of intrauterine undernutrition;8,9,12,13 or human studies exploring intrauterine undernutrition with epigenetic change detected later in life.14 Recent evidence suggests that pre-natal factors other than macronutrient availability can also influence the epigenetic programme. These data come from animal studies of maternal alcohol consumption15 and from human studies of maternal smoking,16 folate intake,17,in vitro fertilization,18 stress19 and socio-economic status20 (reviewed in Foley et al.1). Genetic and apparently stochastic factors also affect epigenetic marks in mammals.21–25 Twin studies have traditionally enabled estimation of genetic, shared and common environmental contributions to variation in phenotype.26,27 They also offer a unique opportunity to unravel the influence of DNA sequence and recorded environmental factors on gene expression and epigenetic marks such as DNA methylation.28–30 MZ twins share maternal, obstetric and genetic factors, but differ in terms of their individual intrauterine environment as influenced by the adequacy of their supply line,31 which can be thought of as the sum of physical (e.g. cord dimensions, placental placement) and biochemical (e.g. nutrients, oxygen) components. Dizygotic (DZ) twins also share maternal and obstetric factors, but their genotypes differ as well as their individual intrauterine environments. As with physical and psychological phenotypes, epigenetic marks can also be treated as quantitative traits and their variance can also be split into genetic and environmental components.30 This study came about to address the lack of knowledge of the rate of epigenetic drift from gestation to early childhood, and the scarcity of data addressing the extent of environmental and genetic influence on epigenetics during these crucial life stages at which the epigenome is at its most sensitive. It was established by a multi-disciplinary team encompassing expertise in epigenetics, biostatistics, epidemiology, obstetrics, paediatrics and nutrition. The majority of funding came from the Australian National Health and Medical Research Council. The Peri/post-natal Epigenetic Twins Study (PETS) aims to study the plasticity of epigenetic marks and the genes they control during the intrauterine period and in early childhood. It also aims to apply the classical twin model to determine the genetic and environmental factors they are influenced by. We focused on obtaining extensive data on maternal factors during pregnancy, collection of multiple biological specimens from different cell lineages in a longitudinal manner, using state-of-the art technologies at both gene-specific and genome-wide levels. We are focusing, in particular, on the potential for epigenetic marks at birth and in early life to provide clues to the causal links between the intrauterine exposures that influence peri-natal phenotype such as birthweight and the risk of chronic diseases in later life. Following ethics approval, women were recruited from multiple pregnancy clinics at three Melbourne hospitals midway through their second trimester (18–22 weeks gestation). Recruiting at this time enabled measurement of maternal and fetal factors at multiple time points; minimized recall bias in dietary questionnaires and enabled us to maintain contact with the women through to birth. Women were not recruited if they planned to leave the area before delivery or had poor English language skills. Our aim was to collect data and samples from 250 mothers and their newborn twins within a 24-month period, a target sample size determined by feasibility constraints. From previous experience, we estimated that approximately 80 pairs would be MZ. In total, 287 pregnant women were recruited over a 32-month period from January 2007 to September 2009. The most common reasons women cited for not participating were: ‘not interested’, being overwhelmed by having a twin pregnancy and anxiety about pregnancy complications such as twin-to-twin transfusion syndrome. After some attrition between recruitment and delivery (see below), we achieved our target of 250. When twins reached 18 months of age, we took repeat biological samples, along with anthropometric measurements and a detailed history of the twins’ health and nutrition since birth (see below). This phase of the study ran from December 2008 to May 2011 and a further follow-up of twins at 5 years of age is currently being planned. Full details of data and sample collection and processing are in Table 1. We aimed to have trained staff attend all deliveries to collect neonatal tissues (cord blood, cords) and buccal (inner cheek) samples within 2 days of delivery. During the course of the study, we also implemented collection of placental tissue in order to examine the level of epigenetic variation present in extra-embryonic tissue. Although it was labour-intensive, this enabled the study of epigenetically diverse tissues from different cell lineages.32 In total, 24 deliveries were missed mainly due to emergency deliveries and/or participants not informing us of impending delivery, but in all of these cases, neonatal buccal cells were collected (by mail if the twins had already left hospital) and access to neonatal (Guthrie) blood spots (which are routinely collected and stored in Victoria) requested. The latter represents a source of genomic DNA from whole blood. Collection of data and samples aSerum, plasma and cells were stored in 1.5-ml cryotubes. We have subsequently measured a limited number of nutritional factors: macronutrients (energy, protein, carbohydrate) from food frequency questionnaires and micronutrients (folate, vitamin B12, homocysteine and vitamin D) from serum/plasma. Collection of data and samples aSerum, plasma and cells were stored in 1.5-ml cryotubes. We have subsequently measured a limited number of nutritional factors: macronutrients (energy, protein, carbohydrate) from food frequency questionnaires and micronutrients (folate, vitamin B12, homocysteine and vitamin D) from serum/plasma. MZ twins may have a shared placenta (monochorionic) or separate placentae (dichorionic). DZ twins may be the same or different sex and are always dichorionic. Thus same-sex dichorionic pairs can be MZ or DZ and require a test of genetic similarity.33 Chorionicity was recorded from a first trimester ultrasound scan and from placental examination by trained staff at delivery. Zygosity was determined by a 12 marker microsatellite test34 using DNA from cord and/or buccal samples. By biobanking tissue samples, we aimed to enable analysis of epigenetic marks and gene expression regulated by these marks. Storing serum and plasma enables analysis of a wide range of nutrients and metabolites. Cryopreservation of viable cells will enable study of epigenetic marks such as DNA methylation and histone modification in addition to gene expression and proteomic studies. Cord blood can be easily separated by density gradient centrifugation into fractions of mononuclear cells (CBMCs), a mixture of mainly T cells and B cells and polymorphonuclear cells (CBPMNCs), consisting mainly of granulocytes.35 We isolated human umbilical vein endothelial cells (HUVECs) from the umbilical cord, as previously described,36 and purified them as a homogeneous (>97% pure) cell population; a small section of whole cord was also snap-frozen. Buccal swabs contain a homogeneous (>97% pure) population of epithelial cells, which have an ectodermal germ cell origin (epidermis, nervous system).37 In total, from 250 twin pairs at birth, we were able to store CBMCs from 52.2% of twin pairs, CBPMNCs from 51% of pairs, placenta from 59% of pairs, HUVECs from 86.9% of pairs, cord from 87.3% of pairs and buccal cells from 96.5% of pairs. Values for blood fractions are low due to suboptimal blood volumes collected at birth. Retrospective collection of Guthrie punches means that whole blood (leucocyte) samples are available from 98% of the cohort. When twins reached 18 months of age, we took repeat biological samples along with anthropometric measurements and a detailed history of the twins’ health and nutrition since birth. Mothers were given the option of allowing a repeat blood sample to be taken from their twins either at the Royal Children's Hospital, Melbourne, or as part of a home visit by our research staff. Buccal cells were obtained from 208 (92.8%) twin pairs remaining in the study at 18 months (see below) and blood from 92 (44.2%) twin pairs. A total of 37 participants were lost from the study between recruitment and delivery for a variety of reasons. A total of 14 women lost one or both twins, 1 moved away from hospital catchment areas, 9 withdrew from the study, 9 were lost due to a temporary staff shortage at one hospital and there were no samples available from four sets of twins (two of which delivered very prematurely). Thus, 250 women and their twins remained in the study immediately after delivery. Between birth and 18 months, 6 mothers withdrew from the study, 7 were uncontactable and 13 failed to return the questionnaires and/or buccal samples. We obtained 18-month data from a total of 224 pairs, 90% of our birth cohort. To maximize participant retention, we attempted to maintain telephone contact with all mothers throughout pregnancy. In addition to relying on mothers informing us of change of address, we sent out birthday cards to twins at 1 and 2 years of age and telephoned mothers if cards were returned due to a change of address. We have also sent out two newsletters to participants during the 18-month follow-up, summarizing our research findings and incorporating a change of contact information form. Table 2 provides data on twin pair-specific factors including mode of conception and birth, zygosity and chorionicity; twin-specific factors including sex and anthropometric measurements including birthweight. One in five pairs were conceived by in vitro fertilization and two-thirds of the pairs was born by caesarean section. The median gestational age at birth was 37 weeks and just over two-fifths of the pairs were monozygotic, slightly higher than average.40 Birthweights and birthweight discordance were representative of twins, in general,41 with a maximum birthweight discordance of 43.3% [(weight of heaviest twin minus weight of lightest twin)/weight of heaviest twin] or an inter-pair z-score difference of 3.08. Summary of twins’ phenotypic data IVF, in vitro fertilization; ICSI, intracytoplasmic sperm injection (a subset of IVF). aFor two additional twin pairs, the first twin was delivered vaginally, the second via caesarean. bZygosity was unable to be determined for three pairs due to insufficient DNA. cCord insertion in relation to the geometric middle of the placenta (central, nearer the centre than the edge; peripheral, nearer the edge than the centre; velamentous, cord arising in membranes surrounding the placenta). dBirthweight standard deviation scores (SDS) were calculated for each newborn using the Microsoft Excel add-in LMS growth (version 2.64) taking into account sex and gestational age.42 Summary of twins’ phenotypic data IVF, in vitro fertilization; ICSI, intracytoplasmic sperm injection (a subset of IVF). aFor two additional twin pairs, the first twin was delivered vaginally, the second via caesarean. bZygosity was unable to be determined for three pairs due to insufficient DNA. cCord insertion in relation to the geometric middle of the placenta (central, nearer the centre than the edge; peripheral, nearer the edge than the centre; velamentous, cord arising in membranes surrounding the placenta). dBirthweight standard deviation scores (SDS) were calculated for each newborn using the Microsoft Excel add-in LMS growth (version 2.64) taking into account sex and gestational age.42 Our research on the PETS cohort has to date focused on using the multiple birth specimens we collected to measure within-pair epigenetic discordance at birth to determine the extent of intrauterine epigenetic plasticity and how this is influenced by environmental and genetic factors. Ongoing research is focusing on changes in epigenetic marks between birth and 18 months of age and the effects of specific intrauterine factors on DNA methylation at birth (at a global and locus-specific level), focusing on maternal nutrition, alcohol intake and assisted reproduction. We first focused on discovering whether individual intrauterine epigenetic plasticity resulted in discordance in DNA methylation in MZ twin pairs at birth, and whether we could find evidence for a genetic effect on discordance in DZ pairs.32 We measured DNA methylation in 56 MZ and 35 DZ twin pairs, within four regions controlling expression within the imprinted IGF2/H19 locus. We studied five cell types: CBMCs, granulocytes, HUVECs, placenta and buccal cells and saw extensive variation in the level of DNA methylation between all individuals, implying that epigenetic differences are widespread within the population. We also saw epigenetic variation between different tissues within each individual, which sheds light on the mechanism by which cells with the same DNA can develop into different cell types. Most interestingly, we saw differences in DNA methylation above the level of technical variation and up to 50% absolute difference in methylation within both MZ and DZ pairs. These findings suggest that individual intrauterine environment has an effect on epigenetic marks and sheds light on one of the possible mechanisms behind developmental programming of disease risk. We also found that within-pair similarity in methylation, as measured by intraclass correlation coefficient, was consistently smaller across all tissues for DZ twins compared with MZ twins—evidence for a genetic component to methylation level within the IGF2/H19 locus. We next used a genome-wide approach to study intrauterine epigenetic plasticity using microarrays, first using gene expression as proxy for overall epigenetic profile. We used the Illumina Expression BeadChip platform, which contains over 48 000 probes, allows interrogation of six samples simultaneously and produces highly accurate and reproducible data.43 Using RNA extracted from CBMCs and HUVECs from a dozen MZ twin pairs, we found, using a compound measure, that all twin pairs exhibited a range of within-pair discordance above the level of technical variation.44 We also found some evidence that sharing a placenta resulted in lower within-pair expression discordance. The most discordantly expressed genes in both cell types were significantly enriched for those previously shown to be involved in response to external environment, again indicating that differences in intrauterine environment can influence epigenetic profile. Focusing on the relationship between birthweight discordance and expression discordance, we found association between birthweight and expression of genes involved in metabolism, catabolism and cardiovascular function. These data provide a plausible biological mechanism for the Developmental Origins of Health and Disease phenomenon, as exemplified by the link between low birthweight and elevated risk of future complex disease, and imply that such epigenetic changes present at birth can be ‘remembered’ for many years to come. We are currently using a genome-wide approach to measure DNA methylation within four cell types at birth and within one of these cell types at 18 months of age, in MZ and DZ twin pairs. We are using Illumina Infinium HumanMethylation BeadChip arrays, which can accurately measure DNA methylation at more than 27 000 specific CpGs at gene promoters.45 This array has subsequently been updated to incorporate more than 485 000 CpGs throughout the genome including at intergenic regions.46 Data are being analysed in a similar way to the expression array data with additional analysis of age-related changes in DNA methylation and, using the traditional twin model, to estimate genetic and environmental influence on inter-individual variability of DNA methylation at both time points (J-H E Joo et al., manuscript in preparation). This longitudinal study follows twins from the second trimester of their gestation. Data on maternal environment are collected retrospectively about peri-conceptional and first trimester exposures and subsequently to birth in real time, thus minimizing recall bias. However, some recall is also required for the 18-month follow-up. Another strength of our study is that multiple cell types were collected and stored from twins at birth and at 18 months, and can also be collected at subsequent time points in the future. In addition, we have accurate documentation on chorionicity and zygosity for all twin pairs, a combination not common in twin studies. Finally, 250 is a moderately large number for a birth cohort in which biospecimens have been taken; it is not large compared with epidemiological twins studies in which heritability (the proportion of phenotypic variance due to variance in additive genetic factors) is calculated. Mechanisms of data sharing, including from microarrays, are currently being developed. Access to de-identified data and offers of potential collaborations are being managed through the PETS Steering Committee via the corresponding author and all such interactions will be dependent on appropriate ethical approvals and Data/Material Transfer Agreements. The Steering Committee is currently considering establishing a PETS web page with public access. The Australian National Health and Medical Research Council (grant numbers 437015 and 607358 to J.C. and R.S); the Bonnie Babes Foundation (grant number BBF20704 to E.J.); the Sigrid Juselius Foundation (to M.O.); the Academy of Finland (to M.O.); the Finnish Cultural Foundation (to M.O); the Financial Markets Foundation for (grant number the The PETS cohort 250 mothers and their The women were recruited during their from three data were collected and multiple tissue specimens were collected from their twins at birth and again at 18 months of Our aims are to of within-pair epigenetic discordance and drift after birth, to examine the environmental and genetic factors that influence epigenetic marks and to measure epigenetic of birthweight and post-natal We all of the in the our Research and and for technical and and staff from Royal and for their and for the of
Language: English
Open Access status: bronze
OpenAlex ID: W2158320244
FWCI: 2.106
Cites: 46
Cited by: 56
Related to: 8