DNA methylation patterns in T lymphocytes are generally stable in human pregnancies but CD3 methylation is associated with perinatal psychiatric symptoms

Abstract

Objectives: To determine whether DNA methylation patterns in genes coding for selected T-lymphocyte proteins are associated with perinatal psychiatric distress or with complications of pregnancy.

Methods: T lymphocyte DNA was obtained from pregnant women across three time points in pregnancy and the postpartum period and epigenetic patterns were assessed using Illumina 450 ​K Methylation Beadchips. Seven selected genes critical for T cell function were analyzed for methylation changes during pregnancy and for associations of methylation patterns with psychiatric distress or with pregnancy complications, with particular attention paid to spatial aggregations of methyl groups, termed 'hotspots,' within the selected genes.

Results: In the candidate gene approach, DNA methylation density within a single cluster of 9 contiguous CpG loci within the CD3 gene was found to be strongly associated with anxiety and depression in mid- and late pregnancy, and weakly associated with the presence of complications of pregnancy. Average DNA methylation density across each of the seven genes examined, and assay-wide, was found to be relatively stable across pregnancy and postpartum, but methylation within the CD3 hotspot was more malleable and changes over time were coordinated across the nine cytosines in the hotspot. CD3 CpGs did not pass array-wide tests for significance, but CpG clusters in two other genes, DTNBP1 and OXSR1, showed array-wide significant associations with anxiety.

Conclusions: Despite the need for tolerating the fetal hemi-allograft, overall DNA methylation patterns in T lymphocytes are generally stable over the mid to late course of human pregnancies and postpartum. However, site-specific changes in DNA methylation density in CD3 appear linked to both symptoms of depression and anxiety in pregnancy and, less strongly, to adverse pregnancy outcomes.

Keywords: Anxiety; DNA methylation; Depression; Epigenetics; Immunity; Pregnancy; Stress; T cell.

Fig. 1

Heat maps showing DNA methylation density at 97 measured loci at the 24–27 week time point, against Hamilton-Anxiey, Hamilton-Depression, Perceived Stress, and complications of pregnancy, delivery, and the postpartum period, for 38 individuals. Scores on clinical scales are sorted from lowest (least distress) to highest (most distress). Pregnancy complications are binary: no complication (sorted to the left) or some complication (sorted to the right). Green indicates greater methylation density, red indicates lower methylation density. Loci where methylation density is correlated with the given clinical scale are marked with asterisks (∗ ​= ​p ​< ​0.1, ∗∗ ​= ​p ​< ​0.05, ∗∗∗ ​= ​p ​< ​0.01). A band of loci within CD3 shows greater methylation density associated with higher scores on the Hamilton-Depression scale. (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.)

Fig. 2

Gene map showing the location of the methylation hotspot in the 3’region of CD3-D and 5′UTR of CD3-G. (hg19 assembly, chr 11, figure courtesy of UCSC genome browser.) Each arrow indicates an individual CpG locus where methylation density is positively associated with anxiety, depression, and antenatal complications of pregnancy (not to scale).

Fig. 3

Schematic illustration of strength of association for methylation density at individual CpG loci within CD3 with selected clinical variables. Top row: methylation data obtained at 24–27 weeks. Second row: methylation data obtained at 34–37 weeks. Columns: left to right, distress data obtained at 24–27 weeks, at 34–37 weeks, and at 4 months postpartum (Spearman correlation). Last column shows complications of pregnancy, delivery, and postpartum (T test). A ‘hotspot’ of 9 contiguous CpG loci within CD3 shows positive associations of DNA methylation density with psychiatric distress. Second trimester distress is more strongly associated than third trimester distress with hotspot methylation across pregnancy. Antenatal distress does not correlate with hotspot methylation density at 4 months postpartum. Hotspot methylation density in the third trimester is also associated with complications of pregnancy.

Fig. 4

DNA methylation density at 24–27 weeks, 34–37 weeks, and 4 months postpartum for 12 individual participants, for cytosines in CD3 located within the hotspot (Fig. 4A) or outside it (Fig. 4B). Methylation changes at hotspot CpGs appear coordinated, while methylation outside the hotspot is more stable and, when altered, changes independently of neighboring CpGs. (A). 9 individual CpGs within the CD3 hotspot. (B). Six individual CpGs outside the CD3 hotspot, each with mean methylation density in similar range to hotspot CpGs. Five CpGs with mean methylation densities above 0.25 were omitted from figure, but showed similar temporal stability.

Fig. 4

DNA methylation density at 24–27 weeks, 34–37 weeks, and 4 months postpartum for 12 individual participants, for cytosines in CD3 located within the hotspot (Fig. 4A) or outside it (Fig. 4B). Methylation changes at hotspot CpGs appear coordinated, while methylation outside the hotspot is more stable and, when altered, changes independently of neighboring CpGs. (A). 9 individual CpGs within the CD3 hotspot. (B). Six individual CpGs outside the CD3 hotspot, each with mean methylation density in similar range to hotspot CpGs. Five CpGs with mean methylation densities above 0.25 were omitted from figure, but showed similar temporal stability.