Methylome meets maternity ward

Abstract

Pre-eclampsia is a dangerous placental condition that can lead to premature labour, seizures and death of mother and infant. It is the most common pregnancy complication in the developed world, occurring in 5--10% of pregnancies (15--18% in developing countries) [1--3]. Unfortunately, diagnosis of pre-eclampsia is difficult as the clinical symptoms (proteinuria and hypertension) do not present until after mid-gestation [1]. The only definitive treatment option is early-induced delivery of the placenta and baby, despite fetal immaturity. Recent evidence suggests that epigenetics plays a critical role in the onset and severity of pre-eclampsia [1, 4-6], and our extensive genome-wide DNA methylation sequencing of placental tissue from pre-eclamptic and normal pregnancies shows significant differences in methylation. We aim to translate the results from our large-scale epigenetic analysis of pre-eclampsia into a useful tool that can be used clinically to predict women who are at risk of developing this threatening condition of pregnancy. We performed a genome-wide DNA methylation analysis of pre-eclampsia using a technique called reduced-representation bisulfite sequencing (RRBS) [7-9]. RRBS combines the bisulfite conversion of DNA with next-generation sequencing to quantify the methylome, or the set of DNA methylation marks across the genome. RRBS interrogates approximately 4 million CpG sites in the genome [10]. The Illumina HiSeq2000 platform was used to sequence a matched cohort consisting of 15 pre-eclamptic and 15 control placentas to determine which genes are differently methylated in pre-eclampsia. The placental tissue in this study is from the Otago Placental Study; a collaboration between scientists, pathologists, and obstetricians in Dunedin, New Zealand that aims to improve maternal and fetal health outcomes. Our genome-wide methylation analysis of pre-eclamptic and control placentas interrogated approximately 650,000 DNA fragments (per sample) and produced roughly 40 million sequence reads, which corresponds to 1.2 billion analyzable sequences (30 samples × 40 million reads). Our large-scale methylation analysis was performed using in-house bioinformatic and biostatistical pipelines for RRBS data [8, 9], which generated a ranked list of epigenetic differences between pre-eclamptic and control placentas. Stringent quality control and data filtering criteria were used (≥15% difference, p≤0.05) to identify 3061 individual CpG-sites and 246 CpG-rich regions that were differentially methylated in pre-eclampsia. The majority of these loci also display hypomethylation in pre-eclamptic placentas compared to control placentas. Our bioinformatic analysis of the methylome in pre-eclampsia was focused on determining candidate loci for altered methylation in pre-eclampsia. We have generated an extensive database to catalogue this differential methylation. We have used Sequenom MassARRAY [11] to validate a selection of candidate loci, and RNA sequencing has been performed to determine which epigenetic changes are biologically meaningful in pre-eclampsia. Once our candidate differentially methylated loci are confirmed, we will investigate whether they can be useful in a clinical setting. The early detection of methylation changes in pre-eclampsia would provide a feasible route to early diagnosis and potential intervention. The methylation profile of placental DNA is a prime target for clinical assessment of the placenta in early pregnancy [12, 13]. A pregnant woman's blood contains circulating fetal DNA that originates from the placenta [1, 14] and comprises 10% (median) of the total DNA in a pregnant women's plasma during the first and second trimesters of pregnancy (the remaining 90% is maternal) [15, 16]. We aim to isolate the placental DNA from maternal plasma and identify whether the differential methylation of our candidate genes can be detected. We are optimizing techniques to differentiate between maternal DNA and placental DNA in maternal plasma. We will then perform deep sequencing on PCR products to determine the proportion of placental DNA. The Illumina MiSeq platform will be used, which produces approximately 12 million sequence reads and can analyse up to 96 samples at a time. This equates to a minimum output of 125,000 reads per sample if a single genetic locus is analysed. We will quantify the methylation of our candidate pre-eclampsia loci in the maternal plasma of a cohort of New Zealand women that are part of the SCOPE (Screening for Pregnancy Endpoints) study; a multinational collaboration involving New Zealand, Australia, the UK and the USA designed to predict and prevent the major diseases of late pregnancy. We will examine the blood plasma samples from pregnant women taken at both 15 and 20 weeks gestation; a selection of which are from women who were later diagnosed with pre-eclampsia. We will quantify the methylation of our candidate differentially methylated loci and correlate the results with the patient's subsequent clinical outcome to determine whether these loci can be used as predictive biomarkers of pre-eclampsia.