Skip to main content

Thank you for visiting You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.

Antenatal prediction of postpartum depression with blood DNA methylation biomarkers

A Corrigendum to this article was published on 22 October 2013


Postpartum depression (PPD) affects 10–18% of women in the general population and results in serious consequences to both the mother and offspring. We hypothesized that predisposition to PPD risk is due to an altered sensitivity to estrogen-mediated epigenetic changes that act in a cell autonomous manner detectable in the blood. We investigated estrogen-mediated epigenetic reprogramming events in the hippocampus and risk to PPD using a cross-species translational design. DNA methylation profiles were generated using methylation microarrays in a prospective sample of the blood from the antenatal period of pregnant mood disorder patients who would and would not develop depression postpartum. These profiles were cross-referenced with syntenic locations exhibiting hippocampal DNA methylation changes in the mouse responsive to long-term treatment with 17β-estradiol (E2). DNA methylation associated with PPD risk correlated significantly with E2-induced DNA methylation change, suggesting an enhanced sensitivity to estrogen-based DNA methylation reprogramming exists in those at risk for PPD. Using the combined mouse and human data, we identified two biomarker loci at the HP1BP3 and TTC9B genes that predicted PPD with an area under the receiver operator characteristic (ROC) curve (area under the curve (AUC)) of 0.87 in antenatally euthymic women and 0.12 in a replication sample of antenatally depressed women. Incorporation of blood count data into the model accounted for the discrepancy and produced an AUC of 0.96 across both prepartum depressed and euthymic women. Pathway analyses demonstrated that DNA methylation patterns related to hippocampal synaptic plasticity may be of etiological importance to PPD.


Postpartum depression (PPD) occurs in approximately 10–18% of women and results in significant morbidity in both mother and child, with offspring risks including low self-esteem, low intellectual skills, child abuse and infanticide.1, 2, 3, 4, 5, 6 Women with mood disorders are at an increased risk of PPD;7 however, the benefits of psychiatric treatment must be carefully weighed against the potential risks of in utero exposure of the offspring to treatment. Antidepressant treatment during pregnancy can result in increased miscarriage rates in early pregnancy and have been associated with low birth weight, preterm birth and birth defects with some classes of antidepressants.8 Limited information is available on the long-term neurocognitive effects of in utero antidepressant exposure.8

PPD occurs up to 4 weeks following parturition according to the Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition (DSM-IV) criteria and follows a dramatic drop in the circulating levels of estradiol (E2) and progesterone (P4). Although depression risk is not predicted by serum levels of gonadal hormones in humans,9 rapid withdrawal from these hormones appears to be an important factor in establishing PPD. In an important experiment, women with a previous history of PPD subjected to supraphysiological doses of E2 and P4 experienced significantly depressed mood symptoms relative to controls upon hormone withdrawal,10, 11 suggesting that the at-risk population exhibits a predisposition to PPD through unknown mechanisms that is triggered by gonadal hormone withdrawal. DNA methylation may represent the link between estrogen and its effects on mood. Indeed, it has previously been demonstrated that E2 administration in vitro can modify DNA methylation at multiple locations downstream of an estrogen-response element.12

Given that fluctuations in estrogen coincide with PPD symptoms and can be antidepressant when administered as a treatment,7, 13, 14, 15, 16 we hypothesized that predisposition to PPD risk is due to an altered sensitivity to estrogen-mediated epigenetic changes that act in a cell autonomous manner detectable in the blood. In this study, we performed a multitiered translational approach to predict PPD status in a prospective cohort using DNA methylation from both the human blood and the hippocampus of mice administered E2. We first define genomic regions of E2-mediated epigenetic change in E2-treated mice and investigated the relationship between E2-induced DNA methylation and PPD risk at syntenic regions in humans. Finally, we use E2-induced methylation models generated in the mice to predict PPD status in the humans.

Materials and methods

Experimental animals

C57BL/6J mice were ovariectomized at 8 weeks of age. At the time of surgery, mice were randomized to receive (subcutaneous implantation) a Silastic capsule (i.d. 1.02 mm; o.d. 2.16 mm) containing 5 mm of dry-packed 17β-E2 (n=5 per group per timepoint). Controls received an empty capsule. Analysis of serum demonstrated consistent levels of E2 in the blood and at 1, 2 and 4 weeks, which was predictive of an increase in uterus weight over those time points (Supplementary Figure 1).

Affymetrix DNA methylation profiling

DNA methylation was assessed in mice using methods described previously17, 18 using HpaII and HinP1I enzymes. Following quality control assessment through Agilent BioAnalyzer-based visualization, the unmethylated fraction of genomic DNA was hybridized to Affymetrix GeneChip Mouse Tiling Promoter 1.0R Arrays at the JHMI Deep Sequencing and Microarray Core facility (Baltimore, MD, USA). Affymetrix CEL files were background corrected and quantile normalized using the AffyTiling package in R, yielding normalized log2-transformed M values representative of the DNA hypomethylation profile per sample. Differentially methylated regions (DMRs) were calculated using the BioTile algorithm ( Identified DMRs were refined by filtering out those not flanked within 1 kb of the DMR boundary by either an HpaII or HinP1I restriction site based on the mouse mm8 genome build sequence. Microarray data is located under GEO accession no.: GSE43460.

Human sample

We recruited 93 pregnant women with a history of either major depression or bipolar disorder (I, II or not otherwise specified) and prospectively followed them during pregnancy and after delivery to identify genetic and clinical characteristics that precede the development of a postpartum depressive episode. Approximately one-third of the sample had bipolar disorder. The average age of the participants was 30.6 years and 70% of the sample was Caucasian. Participants were managed by their treating psychiatrist as clinically indicated and were evaluated during each trimester for pregnancy, and then 1 week, 1 month and 3 months postpartum. Women were classified as being depressed if they met the DSM-IV criteria for a major depressive episode based on a psychiatric interview at each time point (first, second and third trimester and 1 week and 1 month postpartum). We analyzed a subgroup of 32 women euthymic during the third trimester (prepartum euthymic), 34.4% of this subsample (N=11) became depressed within the first 4 weeks postpartum and met the DSM-IV criteria for major depressive episode. A second subgroup of 19 women depressed during pregnancy (prepartum depressed) was assessed in subsequent analyses as an independent replication cohort, of which N=12 remained depressed within the first 4 weeks postpartum and met the DSM-IV criteria for major depressive episode. The trimester of blood draw is depicted in Supplementary Table 1.

Illumina DNA methylation profiling

Samples quality assessment and microarray analysis were conducted at The Sidney Kimmel Cancer Center Microarray Core Facility at Johns Hopkins University using Illumina’s Infinium Human Methylation450 Beadchip Kit (WG-314-1001; Illumina Inc., San Diego, CA, USA) according to the manufacturer’s manual. Images were processed in Illumina’s iScan scanner (Illumina Inc.) and data were extracted using Methylation Module of GenomeStudio v.1.0 Software (Illumina Inc.). Illumina probe type was corrected using the Beta2M function in the wateRmelon package in R. Methylation status of each CpG site was calculated as β (beta)-value based on following definition:

β-Value=(signal intensity of methylation−detection probe)/(signal intensity of methylation−detection probe+signal intensity of non-methylation−detection probe+100).

Microarray data is located under GEO accession no.: GSE44132.

Cell subtype analysis

We quantified cell subfraction percentages for CD8 T cells, CD4 T cells, B cells, monocytes and granulocytes by inputting DNA methylation signatures of 473 loci into an algorithm designed for quantification of the above cell types using DNA methylation proxies from HM450 arrays.20 Before cell-type proportion calculation for the prepartum depressed cohort, DNA methylation values at the 473 loci were transformed by subtracting the residuals from a linear model of the mean DNA methylation values of three cross-batch controls from the prepartum euthymic cohort (batch 1) vs the mean DNA methylation values from two replicates of the same sample run in the prepartum depressed cohort (batch 2).

Sodium bisulfite pyrosequencing

Bisulfite conversion was carried out using EZ DNA Methylation Gold Kit (Zymo Research, Irvine, CA, USA) according to the manufacturer’s instructions. Nested polymerase chain reaction amplifications were performed with a standard polymerase chain reaction protocol in 25-ml volume reactions containing 3–4 μl of sodium-bisulfite-treated DNA, 0.2 μM primers and master mix containing Taq DNA polymerase (Sigma-Aldrich, St Louis, MO, USA). Primer sequences can be found in Supplementary Table 2. Polymerase chain reaction amplicons were processed for pyrosequencing analysis according to the manufacturer’s standard protocol (Qiagen, Gaithersburg, MD, USA) using a PyroMark MD system (Qiagen) with Pyro Q-CpG 1.0.9 software (Qiagen) for CpG methylation quantification.

Statistical analysis

All statistical tests were performed in R ( Using an Anderson–Darling test from the nortest package, all distributions of data that rejected the null hypothesis of normality were subsequently evaluated with non-parametric tests. All statistical tests performed were two-tailed and a P<0.05 is considered significant. Unless otherwise specified, ‘±’ denotes the standard error of the mean.

Weighted genome coexpression network analysis

Weighted genome co-expression network analysis (WGCNA)21 was performed using the WGCNA package in R. In the mouse comparisons, 3606 mean DMR values were used with a power of 20 and minimum module size of 10. For all human analyses, 13 091 nominally significant loci in the combined comparison of PPD (N=11) to non-PPD (N=21) euthymic cohort women were used for correlation with a power of 10 and minimum module size of 10.


Identification of hippocampal targets of E2-mediated DNA methylation change

We sought to identify hippocampal DMRs in the mouse associated with E2 exposure to model the molecular changes occurring during heightened estrogen levels in pregnancy. We chose to utilize hippocampal tissue because effects of E2 on mood are believed, in part, to be localized in the hippocampus, based on numerous studies including knockout experiments,22 17β-E2 administration experiments23 and selective estrogen receptor antagonists and agonists13, 24, 25 demonstrate anxiolytic and antidepressant effects of E2 exposure in rodents. Furthermore in rodent models, E2 administration has been shown to increase synaptic plasticity and dendritic spine density within the hippocampus,14, 15 while withdrawal from pregnancy levels of E2 results in decreased hippocampal brain-derived neurotrophic factor (BDNF) expression16 and suppressed hippocampal neurogenesis.26 We identified 891 significant DMRs before correction for multiple testing. Of these, 380 DMRs exhibited a decrease and 511 exhibited an increase in DNA methylation in response to E2 (Figure 1 and Supplementary Table 3). Gene ontology (GO) analysis using GOstat27 identified a number of significantly enriched GO categories within genes proximal to the identified DMRs (Table 1). Motif enrichment analysis of genomic sequences of the top 100 significant E2 DMRs as well as estrogen receptor-β promoter methylation-correlated regions identified an enrichment for the SP-1 and estrogen receptor transcription factor-binding motifs (Supplementary Figure 2).

Figure 1

Estradiol (E2)-mediated DNA methylation change is associated with postpartum depression (PPD) risk. (a) Volcano plot depicting the difference in DNA methylation between women who suffered PPD vs those who did not (x axis) against the negative natural log of the P-value of association between groups (y axis). (b) A volcano plot depicting DNA methylation differences between the ovariectomy (OVX) and OVX+E2 groups per DMR (x axis) and the—natural log of the P-value for each comparison. Horizontal red lines depict the significance threshold of 5%. (c) Scatter plot of the –log of the P-value of association to discovery sample PPD status and the effect size of E2-mediated DNA methylation change at 103 overlapping loci nominally significant in both humans and mice. (d) Scatter plot of the difference between PPD and non-PPD women in the discovery sample (y axis) as a function of that in the replication sample (x axis).

PowerPoint slide

Table 1 Over-represented GO categories in E2-responsive DMRs

PPD DNA methylation differences are correlated with E2-mediated epigenetic change

We split the human sample into a discovery sample and replication sample consisting of N=6 and 5 women who would and N=12 and 9 who would not develop PPD, each with 35% PPD to 65% non-PPD samples. In the discovery sample, we performed a probe-wise Student’s t-test between PPD and non-PPD cases. We cross-referenced genomic locations of the E2 DMRs from the mouse data with syntenic loci located on the human microarray (Figure 1b). Synteny was calculated based on the relative position of the implicated DMR (Mouse array) or individual CpG locus (Human array) from the closest proximal transcription start site within conserved sequence regions, as established by the UCSC Genome Browser Liftover tool. Owing to the nature of the enzymatic enrichment used in the mouse array experiment, a CpG locus was considered overlapping if it was adjacent within 200 bp of the implicated DMR. In total, 1578 human CpGs were located within the nominally significant mouse E2 DMRs. Pathway analysis of genes associated with overlapping loci using the g.Profiler analysis suite28 identified a single significant GO category (GO: 0010646, frequency observed=0.19, expected=0.024, P=0.046) for ‘regulation of cell communication’ and an enrichment of SP-1 (M00196_4, frequency observed=0.51, expected=0.021, P=0.0084) and AP-2 (M00800_3, frequency observed=0.54, expected=0.021, P=0.0029) transcription factor-binding motifs.

We next attempted to correlate the mean DNA methylation difference between PPD and non-PPD samples and E2-mediated DNA methylation fold change. No correlation was observed across the 1578 overlapping loci (Spearman’s ρ=−0.028, P=0.27). We refined the interrogated data set to 103 loci exhibiting nominally significant association to PPD status and observed significant correlations in both the discovery sample (Spearman’s ρ=0.21, P=0.030) and the replication sample (Spearman’s ρ=0.2, P=0.042). The P-value of association to PPD in the discovery sample was also correlated with E2 DMR effect size (Spearman’s ρ=−0.19, P=0.05) (Figure 1c), suggesting that more robust PPD associations occur at targets of larger E2-mediated DNA methylation change. Furthermore, the mean PPD minus non-PPD value was significantly correlated across the discovery and replication cohorts (Spearman’s ρ=0.32, P=0.0011) (Figure 1d). Permutation testing (20 000 iterations) demonstrated that randomly selected groupings of 103 loci did not correlate better between cohorts (P=5 × 10−5) nor with E2 DMR fold changes in either the discovery or replication samples (P=0.016 and 0.02, respectively). This analysis suggests that the degree to which the discovery and replication cohorts agree is strongly influenced by their localization to syntenic regions of E2-mediated epigenetic reprogramming.

We evaluated the mean PPD minus non-PPD DNA methylation status at the nominally significant PPD associations in the prepartum depressed cohort (N=103 loci) and identified a trend for a positive correlation with the fold change at syntenic E2 DMRs (Spearman’s ρ=0.19, P=0.054). A positive correlation of mean methylation difference between the 1578 loci marked as E2 responsive was also observed between the prepartum depressed and euthymic cohorts (Spearman’s ρ=0.078, P=0.002). Cumulatively, these results support our previous hypothesis that PPD risk may be mediated by an enhanced sensitivity to E2-mediated epigenetic reprogramming.

Identification of DNA methylation biomarkers predictive of PPD

We next reasoned that if estrogen is important for PPD risk, we should be able to predict PPD status based on the degree to which E2 reprograms DNA methylation in the mouse. For each of the 1578 mouse E2 DMRs that overlapped with the human data set, we modeled the mean DNA methylation signature per DMR against the E2 treatment status. In a locus-specific manner, we inputted the human DNA methylation levels per individual in the discovery sample and attempted to predict PPD status using logistic regression. For each locus, the area under the curve (AUC) metric was used to measure prediction accuracy. We then attempted to combine biomarkers to increase predictability using the following algorithm (Supplementary Figure 3a). Linear discriminant analysis was used to combine loci in a forward step-wise manner such that model included loci were those that increased the AUC of the discovery sample until the value was maximized. This set of loci was then used to predict PPD status in the replication sample. The algorithm returned two loci at CpGs cg21326881 and cg00058938, corresponding to the promoter regions of the HP1BP3 and TTC9B genes, respectively, which resulted in an AUC of 0.92 in the discovery sample and 0.9 in the replication sample (Supplementary Figure 3b). A genome-wide significance for this biomarker set of P=0.041 was determined by Monte Carlo permutation analysis.

Pyrosequencing validation of identified biomarkers

We performed sodium bisulfite pyrosequencing to validate the microarray findings in the human sample at CpG dinucleotides located within the region chr 1: 20 986 692–20 986 676 (strand −, human genome build hg18) and chr 19: 45 416 573 (strand +, human genome build hg18), located upstream of HP1BP3 and TTC9B, respectively. PPD status was significantly associated with the HP1BP3 microarray and pyrosequencing data and was significantly correlated between methods (Figures 2a–c and Table 2). DNA methylation for TTC9B was significantly associated with PPD status for both the microarray and pyrosequencing data and was significantly correlated between the two methods (Figures 2e–g and Table 2).

Figure 2

Validation of biomarker loci. Boxplots of the percentage of DNA methylation in the non-postpartum depression (PPD) and PPD groups for HP1BP3 microarray (a) and pyrosequencing (b) and TTC9B microarray (e) and pyrosequencing (f) values. Scatter plots of the % DNA methylation difference between PPD minus non-PPD samples in the prepartum euthymic sample obtained by pyrosequencing (y axis) and microarray (x axis) is depicted for HP1BP3 (c) and TTC9B (g). Boxplots of the percentage of DNA methylation in the non-PPD and PPD groups for HP1BP3 pyrosequencing (d) and pyrosequencing (h) values obtained from the independent replication cohort of prepartum depressed women.

PowerPoint slide

Table 2 Descriptive statistics of biomarker loci

Using HP1BP3 and TTC9B pyrosequencing values in the prediction linear discriminant model, we obtained an AUC of 0.87 for the prepartum euthymic sample, which included three additional women not assessed via microarray (PPD N=13, non-PPD N=22). AUC values did not vary significantly when determined for blood collected in each trimester separately (AUC 1st=0.86, AUC 2nd=0.80, AUC 3rd=1). We next evaluated the performance of the biomarker loci on blood taken from the prepartum depressed sample. While the relative direction of TTC9B association with PPD status was similar to the prepartum euthymic women, it was not significantly different (Figure 2h and Table 2). For HP1BP3 the direction of association was significant but in the opposite direction to that observed in the prepartum euthymic cohort (Figure 2d and Table 2). Linear discriminant model prediction of PPD status in this cohort returned an AUC of 0.12, which represents an 88% prediction accuracy of PPD status in the reverse direction.

Biomarker replication is influenced by blood cellular heterogeneity

We hypothesized that the discrepancy between the prepartum euthymic and depressed cohorts may be related to differences in blood cell-type counts between the two groups. Various experiments have identified elevated granulocytes and decreased CD8 and CD4 T-cell and associated cytokine profiles in individuals exhibiting depressed mood.29, 30 Using DNA methylation proxies in the 19 prepartum depressed and 32 prepartum euthymic women, we determined that cell-type proportions of CD8 T cells, CD4 T cells, B cells and monocytes were significantly reduced in the depressed prepartum group, whereas cross-batch controls exhibited nonsignificant differences in the opposite direction (Supplementary Table 4). Pyrosequencing DNA methylation values for HP1BP3 were evaluated against all cell types in an additive linear model and identified a trend with monocyte proportions (b=−1.11±0.6, P=0.07). We subsequently evaluated the ratio of monocytes to the summed proportions of CD8 T cells, CD4 T cells, B cells and granulocytes, and observed a significant association with prepartum depression status (cell ratio, depressed=0.021±5.2 × 10−4, euthymic=0.032±3.3 × 10−4, P=2.1 × 10−4) but not PPD status (cell ratio, PPD=0.028±5 × 10−4, non-PPD= 0.028±4.2 × 10−4, P=0.86) (Figure 3a), the distribution of which was similar but opposite to that of HP1BP3 DNA methylation (Figure 3b). This monocyte to non-monocyte cell-type ratio was negatively correlated with DNA methylation of HP1BP3 (Spearman’s ρ=−0.37, P=0.0074) (Figure 3c), whereas TTC9B was not associated (Spearman’s ρ=−0.22, P=0.11). Linear regression modeling was performed for PPD diagnosis against an interaction of HP1BP3 DNA methylation with the cell-type ratio, with TTC9B DNA methylation as a covariate. The model was significantly associated with PPD (R2=0.38, P=1.9 × 10−4), as were all model terms including DNA methylation of HP1BP3 (β=−0.22±0.075, P=0.0044), TTC9B (β=−0.033±0.0081, P=1.6 × 10−4), the cell-type ratio (β=−49.66±14.64, P=0.0014), and the interaction between HP1BP3 DNA methylation and cell-type ratio (β=8.03±2.4, P=0.0016). Using a bootstrapping method, we predicted PPD status for each individual using the linear model and obtained an AUC of 0.82 (Figure 3d).

Figure 3

Cell proportion and biomarker DNA methylation predict postpartum depression (PPD). (a) Boxplot of the ratio of monocyte percentage over the sum of T-cell, B-cell and granulocyte percentages as a function of prepartum depression status and PPD diagnosis. (b) Boxplot of the HP1BP3 DNA methylation percentage as a function of prepartum depression status and PPD diagnosis. (c) Scatterplot of the ratio of monocyte percentage over the sum of white blood cell and monocyte percentages as a function of the HP1BP3 DNA methylation percentage. (d) Receiver operator characteristic (ROC) curve of the sensitivity (y axis) vs specificity (x axis) of PPD prediction from the linear model of the HP1BP3 DNA methylation and cell-type ratio interaction, with TTC9B DNA methylation as a covariate. The solid line represents the ROC curve from the proxy-based cell proportion measurement and the dashed line represents that of the complete blood count-derived subsample.

PowerPoint slide

Importantly, the cell proxy analysis only takes into account the relative percentage of various cell types, but not the overall white blood cell (WBC) count. Where available, prepartum WBC counts and proportions of lymphocytes, granulocytes and monocytes were obtained from complete blood count (CBC) data (N=17 women). CBC derived total WBC counts were negatively correlated with the proxy-derived monocyte to non-monocyte ratio (Spearman’s ρ=−0.7, P=0.02), suggesting that the decreased cell-type ratio observed in the prepartum depressed group may be indicative of elevated WBC counts and depression-associated inflammation. This effect appeared to be driven by a positive correlation of WBC count with granulocyte proportion (Spearman’s ρ=0.92, P=2.2 × 10−16), which is consistent with the above-cited elevations in granulocyte levels with depression.29 The ratio of CBC-derived monocyte to non-monocyte (lymphocytes and granulocytes) ratio did not correlate with those derived by DNA methylation proxy (Spearman’s ρ=0.24, P=0.36). We limited the analysis to only those 11 samples where CBC data was derived from within the same trimester as the blood draw used for microarray analysis and observed a significant correlation (Spearman’s ρ=0.66, P=0.044). We attempted to predict PPD status via bootstrap analysis across all 17 individuals using the linear model generated above with CBC data-based monocyte to non-monocyte ratios in place of proxy-based ratios and generated a highly accurate prediction of PPD status (AUC=0.96) (Figure 3d).

Functional classification of HP1BP3 and TTC9B

We attempted to ascertain the function of HP1BP3 and TTC9B loci bioinformatically by using the STRING database31 (Supplementary Figure 4) and by performing weighted gene coexpression network analysis21 against DNA methylation of the HP1BP3 CpG (cg21326881) and TTC9B CpG (cg00058938) as the target variables for correlation (Supplementary Figure 5). The resulting networks of HP1BP3- and TTC9B-coregulated genes were strikingly anticorrelated (Spearman’s ρ=−0.76, P=2.2 × 10−16), suggesting that HP1BP3 and TTC9B are coregulated (Supplementary Figure 5b). Resultantly, we limited networks to those demonstrating significant non-parametric correlation between module membership and correlation significance per group and identified two coregulated networks significantly associated across both genes (Module 1: HP1BP3 ρ=0.56, P=8.8 × 10−4, TTC9B ρ=−0.54, P=0.0015; Module 2: HP1BP ρ=0.45, P=0.0087, TTC9B ρ=−0.46, P=0.0081) (Supplementary Figures 5c and d). No significantly enriched pathways were identified by g.Profiler in Module 2; however, Module 1 contained a number of significantly enriched KEGG (Kyoto Encyclopedia of Genes and Genomes) pathways that can be tied to the antidepressant functions of E2 in the hippocampus (Supplementary Table 5).

We applied weighted genome coexpression network analysis within the PPD and non-PPD women separately, as well as within the mouse E2 DMR data to ascertain the normal coregulation pattern of HP1BP3 and TTC9B genes. The pattern of gene coregulation was positively correlated between HP1BP3 and TTC9B in non-PPD cases and mice, but anticorrelated in PPD cases (Supplementary Figure 6), and is consistent with the proposed heightened sensitivity to E2-mediated epigenetic reprogramming in the PPD group.


We addressed the hypothesis that regions of E2-mediated epigenetic change may predict PPD risk. Numerous correlations linking E2-mediated epigenetic change with DNA methylation changes occurring in the PPD risk population were identified in both the original prepartum euthymic cohort as well as in the independent replication cohort of women depressed during pregnancy. Cumulatively, the results suggest a systematic increase in DNA methylation change occurs in the blood of the PPD group during a period where pregnancy hormones are at high levels. As gonadal hormone levels have been shown not to predict PPD risk, these data provide suggestive evidence that the underlying risk in this group may be related to an increased sensitivity for epigenetic change in response to normal levels of circulating hormones. It is important to consider that the sample sizes interrogated in the mouse experiments were small, and that higher powered experiments may identify additional genomic regions of E2-responsive DNA methylation change in the hippocampus. The findings of enriched SP-1 binding sites and increased evidence for hippocampal long-term potentiation-associated genes in E2-responsive DMRs is consistent with the known downstream transcription factor activation32, 33, 34, 35 as well as antidepressant functions of E2 exposure in the hippocampus,36 and adds confidence to the assertion that we are detecting true E2 DMRs.

CpG methylation levels at two loci within the HP1BP3 and TTC9B genes were identified as biomarkers predictive of PPD. Both genes have ties to estrogen signaling, as HP1BP3 was identified to associate with estrogen receptor β based on tandem affinity purification assays performed on MCF-7 breast cancer cells37 and TTC9B expression has been shown to be responsive to gonadal hormones.38 Owing to the circulating nature of estrogen, the identification of these markers in peripheral blood may be a marker of estrogen-mediated epigenetic changes occurring in the hippocampus and potentially conferring risk to phenotype based on its actions in the brain. The functional relevance of TTC9B may be linked to hippocampal synaptic plasticity as tetratricopeptide repeat containing domains such as that found in TTC9B have been shown to inhibit HSP90-mediated trafficking of AMPA (α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid receptor) receptors critical for hippocampal long-term potentiation/long-term depression.39

Although there have been numerous attempts to generate biomarkers for PPD,40, 41, 42, 43, 44, 45 few studies report a high prediction accuracy. To our knowledge, the identified biomarkers represent the first prospective epigenetic biomarkers capable of predicting PPD status with over 80% accuracy from blood. Segregation of the sample by the trimester of blood collection did not appear to affect prediction accuracy. These results suggest that epigenetic variation at biomarker loci is established early on during pregnancy and may represent a latent epigenetic status in the PPD risk group independent of pregnancy. The clinical implications of this finding are that early screening of those at risk for PPD may be possible, allowing an earlier direction of clinical treatment course.

The high prediction accuracy of the identified biomarkers replicated in an independent cohort of women who were depressed during pregnancy. In this group, the PPD status was segregated with 88% accuracy; however, the prediction was in the opposite direction, driven by differences at the HP1BP3 locus. An analysis of cell subfraction distributions across cohorts identified a difference in the ratio of monocytes to lymphocytes and granulocytes significantly decreased in the depressed cohort that appeared to account for the discrepancy. Our data are consistent with genome-wide expression studies of WBCs taken from women after parturition that demonstrated an association of immune system-related genes with depression scores.45 Incorporation of the DNA methylation biomarkers with cell count data enabled the prediction of PPD status in the entire cohort of 51 women with an AUC of 0.82. A potential confounding factor is that DNA methylation between the prepartum euthymic and depressed cohorts was assessed in two separate batches, as all initial analyses were performed on the euthymic cohort only. To control for this, we normalized DNA methylation levels at all 473 loci used for blood count proxy analysis using a cross-batch control. The predicted cell-type proportions at these controls showed moderate but nonsignificant batch effects between cohorts (Supplementary Table 4); however, the effects were in the opposite direction to the prepartum mood status association observed, suggesting that this association is a true effect of prepartum mood status. In addition, the significant correlation observed with CBC-derived values adds confidence to assertion that the proxy-derived values are representative of actual cell sub-type proportions. Finally, the linear model incorporating CBC-derived cell proportions generated a highly accurate prediction of PPD status (AUC=0.96). Owing to the small size of the subsample used for this prediction, larger prospective cohorts will be required to validate the predictive efficacy of this model. Cumulatively, our data suggest that cell count information in combination with DNA methylation at HP1BP3 and TTC9B successfully and accurately predicts PPD status independent of prepartum mood status.

The results of this study suggest that an increased sensitivity to E2-based epigenetic reprogramming may represent a molecular mechanism of predisposition to PPD risk. Future studies will be needed to rigorously test this hypothesis and track epigenetic changes through the course of pregnancy in women at risk and not at risk for PPD. The investigated population was in women with a previous history of mood disorders; however, studies investigating the efficacy PPD prediction in the general population will need to be determined. Accurate prediction of PPD status will enhance the clinical management of psychiatric treatment during the course of pregnancy.

Accession codes


Gene Expression Omnibus


  1. 1

    O’Hara MW . Postpartum depression: what we know. J Clin Psychol 2009; 65: 1258–1269.

    Article  Google Scholar 

  2. 2

    Soufia M, Aoun J, Gorsane MA, Krebs MO . SSRIs and pregnancy: a review of the literatur]. Encephale 2010; 36: 513–516.

    CAS  Article  Google Scholar 

  3. 3

    Field T . Prenatal depression effects on early development: a review. Infant Behav Dev 2011; 34: 1–14.

    Article  Google Scholar 

  4. 4

    Breese McCoy SJ . Postpartum depression: an essential overview for the practitioner. South Med J 2011; 104: 128–132.

    Article  Google Scholar 

  5. 5

    Cuijpers P, Brannmark JG, van Straten A . Psychological treatment of postpartum depression: a meta-analysis. J Clin Psychol 2008; 64: 103–118.

    Article  Google Scholar 

  6. 6

    Hirst KP, Moutier CY . Postpartum major depression. Am Fam Physician 2010; 82: 926–933.

    PubMed  PubMed Central  Google Scholar 

  7. 7

    Marcus SM . Depression during pregnancy: rates, risks and consequences—Motherisk Update 2008. Can J Clin Pharmacol 2009; 16: e15–e22.

    Google Scholar 

  8. 8

    Pinette MG, Wax JR . The management of depression during pregnancy: a report from the American Psychiatric Association and the American College of Obstetricians and Gynecologists. Obstet Gynecol 2010; 115: 188–189, author reply 189.

    Article  Google Scholar 

  9. 9

    Studd JW . A guide to the treatment of depression in women by estrogens. Climacteric 2011; 14: 637–642.

    CAS  Article  Google Scholar 

  10. 10

    Bloch M, Schmidt PJ, Danaceau M, Murphy J, Nieman L, Rubinow DR . Effects of gonadal steroids in women with a history of postpartum depression. Am J Psychiatry 2000; 157: 924–930.

    CAS  Article  Google Scholar 

  11. 11

    Bloch M, Rubinow DR, Schmidt PJ, Lotsikas A, Chrousos GP, Cizza G . Cortisol response to ovine corticotropin-releasing hormone in a model of pregnancy and parturition in euthymic women with and without a history of postpartum depression. J Clin Endocrinol Metab 2005; 90: 695–699.

    CAS  Article  Google Scholar 

  12. 12

    Kangaspeska S, Stride B, Metivier R, Polycarpou-Schwarz M, Ibberson D, Carmouche RP et al. Transient cyclical methylation of promoter DNA. Nature 2008; 452: 112–115.

    CAS  Article  Google Scholar 

  13. 13

    Walf AA, Frye CA . ERbeta-selective estrogen receptor modulators produce antianxiety behavior when administered systemically to ovariectomized rats. Neuropsychopharmacology 2005; 30: 1598–1609.

    CAS  Article  Google Scholar 

  14. 14

    MacLusky NJ, Luine VN, Hajszan T, Leranth C . The 17alpha and 17beta isomers of estradiol both induce rapid spine synapse formation in the CA1 hippocampal subfield of ovariectomized female rats. Endocrinology 2005; 146: 287–293.

    CAS  Article  Google Scholar 

  15. 15

    ter Horst GJ . Estrogen in the limbic system. Vitam Horm 2010; 82: 319–338.

    CAS  Article  Google Scholar 

  16. 16

    Suda S, Segi-Nishida E, Newton SS, Duman RS . A postpartum model in rat: behavioral and gene expression changes induced by ovarian steroid deprivation. Biol Psychiatry 2008; 64: 311–319.

    CAS  Article  Google Scholar 

  17. 17

    Kaminsky ZA, Tang T, Wang SC, Ptak C, Oh GH, Wong AH et al. DNA methylation profiles in monozygotic and dizygotic twins. Nat Genet 2009; 41: 240–245.

    CAS  Article  Google Scholar 

  18. 18

    Schumacher A, Kapranov P, Kaminsky Z, Flanagan J, Assadzadeh A, Yau P et al. Microarray-based DNA methylation profiling: technology and applications. Nucleic Acids Res 2006; 34: 528–542.

    CAS  Article  Google Scholar 

  19. 19

    Guintivano J, Arad M, Tamashiro KL, Gould TD, Kaminsky ZA . BioTile, A Perl based tool for the identification of differentially enriched regions in tiling microarray data. BMC Bioinform 2013; 14: 76.

    Article  Google Scholar 

  20. 20

    Houseman EA, Accomando WP, Koestler DC, Christensen BC, Marsit CJ, Nelson HH et al. DNA methylation arrays as surrogate measures of cell mixture distribution. BMC Bioinform 2012; 13: 86.

    Article  Google Scholar 

  21. 21

    Langfelder P, Horvath S . WGCNA: an R package for weighted correlation network analysis. BMC Bioinform 2008; 9: 559.

    Article  Google Scholar 

  22. 22

    Walf AA, Koonce CJ, Frye CA . Estradiol or diarylpropionitrile decrease anxiety-like behavior of wildtype, but not estrogen receptor beta knockout, mice. Behav Neurosci 2008; 122: 974–981.

    CAS  Article  Google Scholar 

  23. 23

    Osterlund MK, Witt MR, Gustafsson JA . Estrogen action in mood and neurodegenerative disorders: estrogenic compounds with selective properties-the next generation of therapeutics. Endocrine 2005; 28: 235–242.

    Article  Google Scholar 

  24. 24

    Lund TD, Rovis T, Chung WC, Handa RJ . Novel actions of estrogen receptor-beta on anxiety-related behaviors. Endocrinology 2005; 146: 797–807.

    CAS  Article  Google Scholar 

  25. 25

    Walf AA, Rhodes ME, Frye CA . Antidepressant effects of ERbeta-selective estrogen receptor modulators in the forced swim test. Pharmacol Biochem Behav 2004; 78: 523–529.

    CAS  Article  Google Scholar 

  26. 26

    Green AD, Galea LA . Adult hippocampal cell proliferation is suppressed with estrogen withdrawal after a hormone-simulated pregnancy. Horm Behav 2008; 54: 203–211.

    CAS  Article  Google Scholar 

  27. 27

    Beissbarth T, Speed TP . GOstat: find statistically overrepresented Gene Ontologies within a group of genes. Bioinformatics 2004; 20: 1464–1465.

    CAS  Article  Google Scholar 

  28. 28

    Reimand J, Kull M, Peterson H, Hansen J, Vilo J . g:Profiler—a web-based toolset for functional profiling of gene lists from large-scale experiments. Nucleic Acids Res 2007; 35: W193–W200.

    Article  Google Scholar 

  29. 29

    Shimamiya T, Terada N, Wakabayashi S, Mohri M . Mood change and immune status of human subjects in a 10-day confinement study. Aviat Space Environ Med 2005; 76: 481–485.

    PubMed  Google Scholar 

  30. 30

    Lutgendorf SK, Lamkin DM, DeGeest K, Anderson B, Dao M, McGinn S et al. Depressed and anxious mood and T-cell cytokine expressing populations in ovarian cancer patients. Brain Behav Immun 2008; 22: 890–900.

    CAS  Article  Google Scholar 

  31. 31

    Snel B, Lehmann G, Bork P, Huynen MA . STRING: a web-server to retrieve and display the repeatedly occurring neighbourhood of a gene. Nucleic Acids Res 2000; 28: 3442–3444.

    CAS  Article  Google Scholar 

  32. 32

    Guido C, Panza S, Santoro M, Avena P, Panno ML, Perrotta I et al. Estrogen receptor beta (ERbeta) produces autophagy and necroptosis in human seminoma cell line through the binding of the Sp1 on the phosphatase and tensin homolog deleted from chromosome 10 (PTEN) promoter gene. Cell Cycle 2012; 11: 2911–2921.

    CAS  Article  Google Scholar 

  33. 33

    Bartella V, Rizza P, Barone I, Zito D, Giordano F, Giordano C et al. Estrogen receptor beta binds Sp1 and recruits a corepressor complex to the estrogen receptor alpha gene promoter. Breast Cancer Res Treat 2012; 134: 569–581.

    CAS  Article  Google Scholar 

  34. 34

    Ruegg J, Cai W, Karimi M, Kiss NB, Swedenborg E, Larsson C et al. Epigenetic regulation of glucose transporter 4 by estrogen receptor beta. Mol Endocrinol 2011; 25: 2017–2028.

    CAS  Article  Google Scholar 

  35. 35

    Vivar OI, Zhao X, Saunier EF, Griffin C, Mayba OS, Tagliaferri M et al. Estrogen receptor beta binds to and regulates three distinct classes of target genes. J Biol Chem 2010; 285: 22059–22066.

    CAS  Article  Google Scholar 

  36. 36

    Ramanan VK, Kim S, Holohan K, Shen L, Nho K, Risacher SL et al. Genome-wide pathway analysis of memory impairment in the Alzheimer’s Disease Neuroimaging Initiative (ADNI) cohort implicates gene candidates, canonical pathways, and networks. Brain Imaging Behav 2012; 6: 634–648.

    Article  Google Scholar 

  37. 37

    Nassa G, Tarallo R, Ambrosino C, Bamundo A, Ferraro L, Paris O et al. A large set of estrogen receptor beta-interacting proteins identified by tandem affinity purification in hormone-responsive human breast cancer cell nuclei. Proteomics 2011; 11: 159–165.

    CAS  Article  Google Scholar 

  38. 38

    Cao S, Iyer JK, Lin V . Identification of tetratricopeptide repeat domain 9, a hormonally regulated protein. Biochem Biophys Res Commun 2006; 345: 310–317.

    CAS  Article  Google Scholar 

  39. 39

    Gerges NZ, Tran IC, Backos DS, Harrell JM, Chinkers M, Pratt WB et al. Independent functions of hsp90 in neurotransmitter release and in the continuous synaptic cycling of AMPA receptors. J Neurosci 2004; 24: 4758–4766.

    CAS  Article  Google Scholar 

  40. 40

    Kuijpens JL, Vader HL, Drexhage HA, Wiersinga WM, van Son MJ, Pop VJ . Thyroid peroxidase antibodies during gestation are a marker for subsequent depression postpartum. Eur J Endocrinol 2001; 145: 579–584.

    CAS  Article  Google Scholar 

  41. 41

    Skrundz M, Bolten M, Nast I, Hellhammer DH, Meinlschmidt G . Plasma oxytocin concentration during pregnancy is associated with development of postpartum depression. Neuropsychopharmacology 2011; 36: 1886–1893.

    CAS  Article  Google Scholar 

  42. 42

    Yim IS, Glynn LM, Schetter CD, Hobel CJ, Chicz-Demet A, Sandman CA . Prenatal beta-endorphin as an early predictor of postpartum depressive symptoms in euthymic women. J Affect Disord 2010; 125: 128–133.

    Article  Google Scholar 

  43. 43

    Licinio J . Potential diagnostic markers for postpartum depression point out to altered immune signaling. Mol Psychiatry 2010; 15: 1.

    CAS  Article  Google Scholar 

  44. 44

    Albacar G, Sans T, Martin-Santos R, Garcia-Esteve L, Guillamat R, Sanjuan J et al. Thyroid function 48h after delivery as a marker for subsequent postpartum depression. Psychoneuroendocrinology 2010; 35: 738–742.

    CAS  Article  Google Scholar 

  45. 45

    Segman RH, Goltser-Dubner T, Weiner I, Canetti L, Galili-Weisstub E, Milwidsky A et al. Blood mononuclear cell gene expression signature of postpartum depression. Mol Psychiatry 2010; 15: 93–100, 102.

    CAS  Article  Google Scholar 

Download references


We thank The Solomon R and Rebecca D Baker Foundation for their generous support of this research. This work was funded, in part, by MH093967 to TDG and a NARSAD 2010 Young Investigator Award to ZK. The Sidney Kimmel Cancer Center Microarray Core Facility at Johns Hopkins University was supported by NIH Grant P30 CA006973. All animal treatments were approved by the University of Maryland, Baltimore Animal Care and Use Committee and were conducted in full accordance with the NIH Guide for the Care and Use of Laboratory Animals. Human subjects research was conducted under IRB protocol no. 00008149 and subjects were collected with funding from K23 MH074799-01A2 to JP.

Author information



Corresponding author

Correspondence to Z A Kaminsky.

Ethics declarations

Competing interests

ZK, JP and TDG are coinventors listed on a patent application for DNA methylation at biomarker loci related to PPD. The remaining authors declare no conflict of interest.

Additional information

Supplementary Information accompanies the paper on the Molecular Psychiatry website

Supplementary information

PowerPoint slides

Rights and permissions

Reprints and Permissions

About this article

Cite this article

Guintivano, J., Arad, M., Gould, T. et al. Antenatal prediction of postpartum depression with blood DNA methylation biomarkers. Mol Psychiatry 19, 560–567 (2014).

Download citation


  • biomarker
  • DNA methylation
  • estrogen
  • HP1BP3
  • postpartum depression
  • TTC9B

Further reading


Quick links