ABSTRACT
Previous studies suggest epigenetic alterations may contribute to the association between maternal prenatal depression and adverse offspring outcomes. Developmental researchers have recently begun to examine these associations in relation to epigenetic age acceleration/deceleration, a biomarker of developmental risk that reflects the deviation between epigenetic age and chronological age. In the perinatal period, preliminary studies indicate that maternal prenatal depression may lead to epigenetic age deceleration in newborns, which may predict adverse developmental outcomes. The present study examined the relationship between maternal prenatal exposures (i.e., depression, stress, and SSRI use) and offspring epigenetic age deceleration in 303 mother-offspring dyads. Women were recruited in the first trimester of pregnancy and followed longitudinally until delivery. Maternal depression, perceived stress, and SSRI use were assessed at each prenatal visit. Newborn epigenetic age was determined via cord blood samples. Results indicated maternal prenatal stress was not associated with newborn epigenetic age deceleration (ΔR2 = 0.002; p = 0.37). Maternal prenatal depression was associated with decelerated epigenetic age (ΔR2 = 0.01, p = 0.04), but this relationship did not hold when accounting for maternal use of SSRIs (ΔR2 = 0.002, p = 0.43). Conversely, maternal SSRI use significantly predicted newborn epigenetic age deceleration over and above the influence of maternal depression (ΔR2 = 0.03, p = 0.001). These findings suggest maternal prenatal SSRI use may significantly contribute to the previously documented association between maternal prenatal depression and epigenetic age deceleration. Further studies are needed to examine how these epigenetic differences at birth may contribute to adverse outcomes in later development.
KEYWORDS: Epigenetic Age, prenatal exposures, SSRI, depression
Pregnancy is marked by an increased risk for psychopathology, with nearly a quarter of pregnant women in the United States meeting criteria for major (6.1%) or minor (16.6%) depression [1]. Evidence suggests that maternal depression during pregnancy may negatively impact the foetus, with offspring exposed to maternal prenatal depression demonstrating a variety of adverse outcomes – from developmental delays to internalizing disorders – across the lifespan [2]. These findings align closely with the Developmental Origins of Health and Disease (DOHaD) framework, which posits that environmental factors during the perinatal period can enact biological changes in offspring that lead to health risks in later life [3]. Among these potential biological changes are epigenetic alterations, which may lead to differences in gene expression that may in turn shape developmental trajectories of health and disease. Recent evidence supports this mechanism of action, with preliminary studies demonstrating associations between maternal prenatal depression, DNA methylation (a form of epigenetic regulation that influences gene expression), and offspring developmental outcomes including cognitive functioning and depressive symptomatology [4,5]. However, these studies require replication as the influence of maternal prenatal depression on the offspring epigenome remains unclear.
Unlike the genome, which remains stable across the lifespan, the epigenome is a dynamic system that responds to internal and external influences across development. DNA methylation, the most widely studied epigenetic marker, shows particular malleability during gestation as the foetus undergoes cell differentiation, organ development, and the establishment of immune and stress-response systems [6]. Consequently, exposures during gestation have the potential to unduly impact offspring developmental trajectories or disease susceptibility throughout life [7]. Previous studies have drawn links between prenatal exposures and differences in DNA methylation using a variety of methodologies, the most common being biologically informed candidate-gene approaches and data-driven whole epigenome (i.e., EWAS) approaches [8–10]. Increasingly, though, attention has been dedicated to the impact of prenatal exposures on epigenetic age, a recently identified biomarker of developmental maturity.
Epigenetic age
Epigenetic age (also referred to as DNA methylation age, or DNAm age) refers to the association between specific DNA methylation patterns and chronological age. For the past decade, epigenetic age has set the foundation for a burgeoning field of research examining how epigenetic alterations may function as a biomarker and mechanism for ageing and health [11]. Naturally, evidence indicates that individuals with the same chronological age demonstrate variability in epigenetic age, with some individuals exhibiting methylation patterns characteristic of younger individuals (i.e., decelerated age, or age deceleration) and some of older individuals (i.e., accelerated age, or age acceleration). The difference between chronological age and epigenetic age demonstrates significant predictive utility for a variety of developmental outcomes, over and above the effect of chronological age. Primarily studied in the context of later-life outcomes, epigenetic age acceleration has been found to predict lower cognitive functioning and physical fitness, higher rates of all-cause mortality, and a wide range of chronic conditions in adults [11–13]. Further, studies investigating the factors that may influence epigenetic age have identified chronic stress – as well as proxies for chronic stress, such as socioeconomic status and level of education – as a significant predictor of accelerated epigenetic age in adults [14,15]. Together, this evidence suggests that differences in DNA methylation, reflected by epigenetic age, may serve as a biological mechanism underlying the associations between environmental risk factors and physical and cognitive decline.
Epigenetic gestational age
Given the evidence that accelerated epigenetic age may play a mechanistic role in adult disease and development, preliminary studies have begun to examine the role of epigenetic age in early-life outcomes. These studies, though limited, have identified epigenetic gestational age (i.e., DNA methylation patterns associated with length of gestation) as a potentially informative biomarker for prenatal exposures and developmental outcomes. Contrary to adult studies, lower gestational age acceleration (i.e., deceleration, or an epigenetic age that is lower than the clinically estimated gestational age) has demonstrated predictive utility for adverse health-related outcomes, such as low birth weight [16,17]. This finding is important given the association of low birth weight with cognitive delays, attention deficits, and internalizing problems such as depression and anxiety [18,19]. Interestingly, preliminary findings suggest that gestational age deceleration is directly associated with psychological and neurodevelopmental outcomes. For example, a recent study found that decelerated gestational age significantly predicted adverse outcomes in early childhood including internalizing and total problems on the Child Behaviour Checklist [5]. Of note, this study also found a significant association between maternal prenatal depression and newborn epigenetic age, such that higher levels of maternal depression during pregnancy predicted decelerated epigenetic gestational age. Together, these findings provide preliminary evidence that epigenetic age may mediate the relationship between prenatal exposures such as maternal depression and adverse developmental outcomes.
Prenatal influences on epigenetic gestational age
Given the dearth of studies examining the impact of maternal prenatal depression on infant epigenetic gestational age, further studies are needed to replicate and extend the preliminary findings in this field. As with other research on the influence of maternal prenatal depression on offspring biology and psychopathology, it is important to separate the effects of depression from the effects of common correlates of depression such as psychotropic medications (e.g., selective serotonin reuptake inhibitors (SSRIs)) and elevated levels of perceived stress [20,21]. Although depression and stress are closely related, previous studies have identified overlapping but distinct effects of prenatal depression and prenatal stress on offspring outcomes [8,22]. Given the suggested impact of stress exposure on epigenetic age acceleration in older populations – along with evidence that the brain and epigenome are particularly susceptible to stress during the perinatal period [23,24] – it is important to assess to potential influence of prenatal stress on gestational epigenetic age.
To date, no studies have examined the potential impact of maternal prenatal medication use on epigenetic gestational age. Preliminary evidence from EWAS studies investigating prenatal exposure to maternal SSRI use has suggested that offspring epigenetic alterations may actually be more highly associated with SSRI use than with maternal depression [25], although results are mixed [26,27]. With up to 41% of depressed women choosing to maintain antidepressant use during pregnancy Cohen et al. [28], it is important to determine whether prenatal SSRI exposure may actually be driving the adverse influences of maternal depression on offspring gestational age deceleration.
The present study
Given the prevalence of maternal depression during pregnancy and the cooccurrence of elevated stress and SSRI use during this sensitive period of offspring development, the present study aims to explore the potential associations between maternal prenatal depressive symptoms, maternal prenatal stress, maternal prenatal SSRI use, and offspring epigenetic gestational age deceleration at birth. In line with the associations identified by Suarez and colleagues (2018), we hypothesize (1) that greater levels of maternal prenatal depression will be associated with decelerated epigenetic gestational age. We expect (2) that this association will remain after accounting for maternal prenatal stress and SSRI use, and (3) that these additional prenatal exposures will be associated with further deceleration of offspring epigenetic gestational age.
Method
Study population
Mother-infant dyads were recruited from the Emory Women’s Mental Health Program (WMHP), an outpatient treatment facility for women experiencing psychiatric disorders during the perinatal period. Mothers were followed longitudinally from early pregnancy through delivery, with data collected on maternal medication use, self-reported depressive symptoms, and self-reported perceived stress at each visit. Foetal umbilical cord blood samples were collected at delivery. Our total sample comprised 303 mother-infant dyads (all singleton pregnancies) with available cord blood methylation data. Of these 303 dyads, data on maternal prenatal SSRI use was available for 264 dyads, maternal prenatal depressive symptoms for 256 dyads, and maternal prenatal perceived stress for 249 dyads. There were no differences between those who did and did not have available data on prenatal depressive symptoms, perceived stress, and/or SSRI use. In order to maximize use of the existing data, missing data were multiply imputed using the Multivariate Imputation by Chained Equations (MICE) package in R Buuren and Groothuis-Oudshoorn [29].
Infant DNA methylation and epigenetic gestational age acceleration at birth
Foetal cord blood samples were collected according to standard procedures (described in detail by Schroder et al. [30]). DNA was extracted, processed, and sequenced using the HumanMethylation27 BeadChip (Illumina) at the Emory Biomarker Service Centre. Signal data was normalized to adjust for technical variability between samples using information from Illumina BeadChip negative control probes that are designed to detect a true methylated and unmethylated signal at zero. Signals from methylated and unmethylated bead types were then used to calculate the proportion of DNA methylated at each CpG site.
Epigenetic gestational age was calculated based on a weighted average of 148 CpG sites, according to procedures developed by Knight et al. [16]. In line with expectations, epigenetic gestational age was positively correlated with clinically estimated gestational age (Pearson correlation r = 0.498, p < 0.001], which was determined through ultrasound dating. The difference between epigenetic gestational age and clinically estimated gestational age (i.e., gestational age acceleration/deceleration) was defined by taking the residual of a linear regression of epigenetic gestational age on clinically estimated gestational age.
Infant cord blood cell counts at birth
To account for the potential effects of cell type heterogeneity in foetal cord blood, cell counts were estimated for six cell types [i.e., CD4+, CD8+, CD14+, CD19+, CD56 + T cells, and eosinophils) using the method proposed by Houseman et al. [31].
Maternal prenatal depressive symptoms
Mothers were followed longitudinally over the course of pregnancy, with approximately four to six weeks between visits. At each visit, the Beck Depression Inventory-Second Edition (BDI-II; Beck, Steer & Brown, 1996] was used to measure maternal self-reports of continuous depressive symptom data. The BDI-II is a 21-item widely used measure of depressive symptoms that has been well-validated in both clinical and community samples (Beck, Steer & Carbin, 1988). BDI-II scores were used to calculate an area under the curve (AUC) measure of symptom levels across pregnancy, normalized to 40 weeks to account for differences in length of gestation.
Maternal prenatal stress symptoms
The Perceived Stress Scale [PSS-10; 32] is a 10-item self-report measure used to assess maternal psychological distress. The PSS-10 has been shown to demonstrate high internal consistency (Cronbach’s α = 0.89) and high convergent validity with other well-established measures of perceived stress and anxiety [33]. PSS-10 scores across all prenatal visits were used to calculate an AUC measure of maternal self-reported distress across pregnancy, normalized to 40 weeks’ gestation.
Prenatal SSRI exposure
Exposure to prenatal SSRIs – as well as to other anti-depressants, psychotropics, and substances – was measured using a tracking protocol described by Johnson et al. [34]. Briefly, participants completed weekly tracking sheets that asked about medication and substance use over the past seven days, which were then verified by study physicians. SSRI exposure was calculated by summing the number of weeks that usage was endorsed and standardizing sums to a 40-week pregnancy. Due to a largely bimodal distribution of standardized sums (i.e., 50.8% of mothers with zero weeks of SSRI exposure and 27.7% of mothers with 40 weeks) – a dichotomous variable (SSRI exposure vs. no SSRI exposure) was used in all analyses.
Data analyses
Hypothesis testing was conducted in SPSS through a series of multiple regression analyses. All tests adjusted for cord blood cell type composition. For our replication analysis, we tested the association of maternal prenatal depression with offspring epigenetic age deceleration, adjusting for maternal age at delivery (in line with the study we aimed to replicate [5]). We then extended this replication by running a second analysis, additionally accounting for covariates identified using bivariate correlation analyses: Offspring ethnicity, maternal prenatal SSRI use, maternal prenatal tobacco use, and maternal education (a proxy for maternal socioeconomic status [35]). A third regression analysis, which also accounted for these covariates, was used to identify the role of maternal prenatal stress on offspring epigenetic age deceleration. Of note, we did not include maternal prenatal depression in this model, given its significant overlap with maternal prenatal stress (Pearson’s r = 0.71, p < 0.001). Finally, we conducted a fourth multiple regression analysis to identify the competing role of maternal prenatal SSRI use, accounting for maternal prenatal depression as well as the additional covariates. The first two tests were based on a two-sided p-value, with an alpha of 0.05. The last two tests – which both addressed the third hypothesis – used an alpha of 0.025 in order to adjust for multiple testing, per the Bonferroni correction (α = 0.05/2).
Results
Table 1 displays descriptive statistics for study variables for the present sample. Table 2 provides results from preliminary analyses of relevant covariates (maternal substance use, maternal age at delivery, maternal education, and offspring sex and ethnicity).
Table 1.
Descriptive statistics
| M (SD) | % | |
|---|---|---|
| Predictors | ||
| Maternal Prenatal Depression a | 450.4 (307.5) | |
| Maternal Prenatal Perceived Stress a | 1109.0 (391.5) | |
| Maternal SSRI Use | 49.5% | |
| Outcome | ||
| Gestational Age Deceleration | 0 (1.63) | |
| Covariates | ||
| Maternal Tobacco Use | 18.8% | |
| Maternal Age at Delivery | 32.99 (5.28) | |
| Maternal Education | ||
| High School or Below | 8.9% | |
| Some College | 20.5% | |
| College Degree | 34.0% | |
| Secondary Degree | 36.6% | |
| Offspring Sex | ||
| Female | 50.2% | |
| Male | 49.8% | |
| Offspring Ethnicity | ||
| White | 82.8% | |
| Non-White | 17.2% | |
| Sample Characteristics | ||
| Number of Previous Pregnancies | ||
| 0 | 41.9% | |
| 1 | 38.9% | |
| 2+ | 19.2% | |
| Offspring Gestational Age at Delivery (weeks) | 38.80 (1.24) | |
| Preterm b | 7.3% | |
| Delivery Method | ||
| Vaginal | 65.9% | |
| Caesarean | 34.1% | |
| Offspring Birth Weight (lbs) | 7.36 (0.99) | |
|
a Area Under Curve (AUC) values; BDI-II scores ranged from 0 to 50 (M = 16.27, SD = 11.0) b All preterm infants were considered ‘late preterm’ (34–37 weeks) | ||
Table 2.
Bivariate correlations between clinically estimated gestational age, epigenetic gestational age deceleration, and relevant predictors
| Clinically Estimated Gestational Age |
Epigenetic Gestational Age Deceleration |
|||
|---|---|---|---|---|
| Pearson’s r | p-value | Pearson’s r | p-value | |
| Maternal Prenatal Depression | −0.186 | 0.001 | 0.123 | 0.032 |
| Maternal Prenatal Perceived Stress | −0.151 | 0.01 | 0.064 | 0.27 |
| Maternal Prenatal SSRI Use a | −0.126 | 0.03 | 0.164 | 0.004 |
| Maternal Prenatal Tobacco Use b | −0.042 | 0.46 | 0.201 | <0.001 |
| Maternal Age at Delivery | −0.029 | 0.61 | −0.152 | 0.01 |
| Maternal Education | 0.062 | 0.29 | −0.220 | <0.001 |
| Offspring Sex c | −0.117 | 0.04 | 0.038 | 0.51 |
| Offspring Ethnicity d | −0.036 | 0.53 | 0.124 | 0.03 |
Italics indicate significance at the 0.05 level.
aSSRI Use was coded as 0 (no SSRI use) or 1 (SSRI use)
bTobacco Use was coded as 0 (no tobacco use) or 1 (tobacco use)
cSex was coded as 0 (female) or 1 (male)
dEthnicity was coded as 0 (White) or 1 (non-White)
Replication analysis: role of maternal prenatal depressive symptoms
Results from the replication analysis are displayed in Table 3. A greater level of maternal depressive symptoms during pregnancy was associated with decelerated epigenetic gestational age in newborns (ΔR2 = 0.014, p = 0.04). These findings aligned with results from the study we aimed to replicate [5].
Table 3.
Replication analysis. Multiple regression of maternal prenatal depression predicting epigenetic gestational age deceleration
| Variable | ∆F | ∆R2 | t | p-value |
|---|---|---|---|---|
| Block 1 (Cord Blood Cell Type Composition) | 6.41 | 0.12 | <0.001 | |
| CD4 + T cells | 2.45 | 0.02 | ||
| CD8 + T cells | 1.22 | 0.22 | ||
| CD14 + T cells | 1.06 | 0.29 | ||
| CD19 + T cells | 1.56 | 0.12 | ||
| CD56 + T cells | 0.34 | 0.74 | ||
| Eosinophil Granulocytes | 4.13 | <0.001 | ||
| Block 2 (Maternal Age at Delivery) | 8.81 | 0.03 | −3.00 | 0.003 |
| Block 3 (Maternal Prenatal Depression) | 4.22 | 0.01 | 2.06 | 0.04 |
Italics indicate significance at the 0.05 level.
Extension of replication analysis: role of maternal prenatal depressive symptoms, accounting for relevant covariates
Results from the extended replication analysis, which accounted for covariates significantly correlated with gestational age deceleration, are displayed in Table 4. After accounting for additional covariates (i.e., maternal prenatal SSRI use and tobacco use, maternal education, and offspring ethnicity), level of maternal depressive symptoms during pregnancy was no longer associated with decelerated gestational age in newborns (ΔR2 = 0.002, p = 0.43).
Table 4.
Multiple regression of maternal prenatal depression predicting epigenetic gestational age deceleration, accounting for additional covariates
| Variable | ∆F | ∆R2 | t | p-value |
|---|---|---|---|---|
| Block 1 (Cord Blood Cell Type Composition a) | 6.41 | 0.12 | <0.001 | |
| Block 2 (Maternal Age at Delivery) | 8.81 | 0.03 | 0.003 | |
| Block 3 (Additional Covariates) | 4.66 | 0.06 | <0.001 | |
| Maternal Prenatal Tobacco Use | 2.21 | 0.03 | ||
| Maternal Prenatal SSRI Use | 3.50 | 0.001 | ||
| Maternal Education | −1.75 | 0.08 | ||
| Offspring Ethnicity | 0.65 | 0.52 | ||
| Block 4 (Maternal Prenatal Depression) | 0.61 | 0.002 | 0.78 | 0.43 |
Italics indicate significance at the 0.05 level.
aTable 3 displays all cell types included in Block 1.
Extension analysis: role of maternal prenatal perceived stress
Maternal prenatal perceived stress was not significantly associated with gestational age deceleration. This was the case for both the correlation analysis (Table 2) and the multiple regression analysis, which accounted for relevant covariates (Table 5).
Table 5.
Multiple regression of maternal prenatal perceived stress predicting epigenetic gestational age deceleration, accounting for relevant covariates
| Variable | ∆F | ∆R2 | t | p-value |
|---|---|---|---|---|
| Block 1 (Cord Blood Cell Type Composition a) | 6.41 | 0.12 | <0.001 | |
| Block 2 (Maternal Age at Delivery) | 8.81 | 0.03 | 0.003 | |
| Block 3 (Additional Covariates) | 5.55 | 0.06 | <0.001 | |
| Maternal Prenatal Tobacco Use | 2.21 | 0.03 | ||
| Maternal Prenatal SSRI Use | 3.50 | 0.001 | ||
| Maternal Education | −1.75 | 0.08 | ||
| Offspring Ethnicity | 0.65 | 0.52 | ||
| Block 4 (Maternal Prenatal Stress) | 0.81 | 0.002 | 0.90 | 0.37 |
Italics indicate significance at the 0.05 level.
aTable 3 displays all cell types included in Block 1.
Extension analysis: role of maternal prenatal SSRI use
Maternal SSRI use during pregnancy was significantly associated with gestational age deceleration at birth (Table 6; Figure 1; ΔR2 = 0.03, p = 0.001). This was the case even after accounting for maternal depression during pregnancy (in addition to all other covariates) and after adjusting for multiple testing (α = 0.025).
Table 6.
Multiple regression of maternal prenatal SSRI use predicting epigenetic gestational age deceleration, accounting for relevant covariates
| Variable | ∆F | ∆R2 | t | p-value |
|---|---|---|---|---|
| Block 1 (Cord Blood Cell Type Composition a) | 6.41 | 0.12 | <0.001 | |
| Block 2 (Maternal Age at Delivery) | 8.81 | 0.03 | 0.003 | |
| Block 3 (Additional Covariates) | 3.02 | 0.03 | 0.02 | |
| Maternal Prenatal Depression | 1.35 | 0.18 | ||
| Maternal Prenatal Tobacco Use | 1.70 | 0.09 | ||
| Maternal Education | −1.68 | 0.10 | ||
| Offspring Ethnicity | 0.54 | 0.59 | ||
| Block 4 (Maternal Prenatal SSRI Use) | 10.34 | 0.03 | 3.22 | 0.001 |
Italics indicate significance at the 0.05 level.
aTable 3 displays all cell types included in Block 1.
Figure 1.
Offspring of mothers who used SSRIs during pregnancy exhibit decelerated epigenetic gestational age (GA), as reflected by negative residual scores when epigenetic GA is regressed on clinically estimated GA. The origin represents equivalent epigenetic and clinically estimated GA. Bars indicate 95% confidence intervals
Discussion
Our findings replicate and extend prior work on the factors associated with epigenetic ageing at birth. Specifically, we replicated the previous finding that maternal prenatal depressive symptoms prospectively predict gestational age deceleration at birth [5], but we found that this association was no longer significant when controlling for confounding variables such as maternal SSRI use during pregnancy. Conversely, we found that maternal prenatal SSRI use remained a significant predictor of child epigenetic gestational age deceleration even after controlling for maternal prenatal depressive symptoms and relevant confounders.
Prenatal perceived stress was not associated with alterations in child epigenetic gestational age. This finding was surprising given that previous research supports associations between prenatal stress and adverse birth outcomes such as low birth weight and preterm birth [36]. Prior work has also linked stress exposure to epigenetic age acceleration in adulthood [14,15], leading to the hypothesis that epigenetic ageing may be one mechanism that links chronic stress to negative health outcomes. However, it remains unclear how epigenetic gestational ageing relates to measures of epigenetic ageing at older developmental stages. Health decline is often linked to accelerated epigenetic ageing in adults, but previous research examining prenatal development has linked decelerated epigenetic ageing to adverse outcomes, including low birthweight and internalizing problems [5,37]. Additional research is therefore needed to better understand the relationship between epigenetic gestational ageing and health outcomes in early life, but it seems as though decelerated gestational ageing, instead of or in addition to accelerated ageing, may confer risk for negative outcomes.
Few studies have examined the relationship between maternal prenatal psychopathology on child epigenetic gestational age deceleration. To our knowledge, the only published study that has examined maternal depression and child epigenetic gestational ageing did not assess for maternal psychotropic medication use during pregnancy [5]. The present study suggests the replicated association between maternal prenatal depression and child epigenetic gestational age deceleration may be driven by maternal SSRI use during pregnancy, which is an important consideration for future work.
Previous work in humans and rodents has linked prenatal SSRI exposure to various epigenetic alterations in offspring. For instance, a recent preclinical study demonstrated increased global DNA methylation in the hippocampus and decreased methylation in the cortex of 22-day-old rats following prenatal SSRI exposure [38]. These differences are notable given the role of these brain regions in cognition, development, and psychopathology [39]. In humans, prenatal exposure to antidepressant medication (including, but not specific to, SSRIs) has been shown to predict differential methylation of CpG sites in genes that influence neural development, connectivity, and neurocognitive functioning [25]. However, a systematic review of studies utilizing candidate gene and array approaches to examine the association between prenatal exposure to antidepressant medication and epigenetic alterations in newborn cord blood has yielded mixed results [27]. Half of the studies included in this review identified a significant link between prenatal antidepressant exposure and differential methylation at certain CpG sites in newborn cord blood, but half of the studies found no association. These mixed findings indicate there may be moderators of this association, such as subsequent postnatal exposures or child sex.
In addition to potential epigenetic alterations, prenatal exposure to SSRIs has been linked to a number of phenotypic alterations across development. Experimental studies find that mice exposed to SSRIs in utero exhibit sleep anomalies, increased anxiety and depression-like behaviour including anhedonia and helplessness in the context of stress, and reduced novelty investigation across the lifespan [see review: 40]. Although no clinical studies in humans have examined SSRI-related outcomes beyond childhood, prenatal SSRI exposure is linked to a number of adverse outcomes in the first few years of life. Of particular relevance to the present study, prenatal SSRI use is linked to shortened gestation and higher rates of preterm birth [41,42], which is consistent with our finding that SSRI use predicts epigenetic gestational age deceleration in offspring. Prenatal SSRI exposure is additionally associated with altered HPA axis functioning [43–46] and poor psychomotor development during the neonatal period [47]. Generally, however, studies examining the impact of maternal SSRI use during pregnancy on infant outcomes have largely demonstrated null findings, which may contribute to increases in prenatal SSRI use in recent years [48]. Studies examining childhood outcomes show more consistent negative impacts on neurodevelopment, potentially suggesting the presence of a ‘sleeper effect’ or a latency period. Children exposed to SSRIs prenatally show increased internalizing and externalizing behaviours, reduced task persistence, poorer language and social skills, and lower IQ compared to unexposed controls [49–51]. Crucially, these differences are not only observed with comparison to ‘control’ children of non-depressed mothers but also with children exposed to equivalent levels of maternal prenatal depression without accompanying SSRI exposure.
The present findings should be considered within the context of study limitations. Although our sample was enriched for maternal depression, our participants were predominantly white, highly educated families. Additional research is therefore needed to determine the degree to which our findings generalize to families of diverse ethnicities and socioeconomic status. We were also unable to conduct replication analyses given the limited datasets that contain measures of prenatal maternal stress and depression, prenatal SSRI use, and infant DNA methylation. As an alternative, we attempted to replicate our findings using the present cohort but an alternate predictor of epigenetic gestational age [52]. However, the predictor required methylation data from the HumanMethylation450 BeadChip and the majority of our samples were run on the HumanMethylation27 BeadChip. As such, we were not able to utilize this predictor for replication. Our analyses were also restricted to a single tissue type. Given that DNA methylation patterns are tissue-specific [53,54], methylation patterns observed in cord blood samples are not fully representative of methylation patterns in the newborn brain. However, nonhuman primate studies have identified a strong correlation between methylation patterns in brain tissue and in cord blood samples, which increases our confidence that our findings may generalize to DNA methylation in neonatal brain tissue. Further, prior work supports the notion that SSRIs can cross the placental barrier to impact the developing foetal brain [55]. Nevertheless, the mechanisms by which maternal SSRI use may lead to methylation alterations in the foetal brain remain unknown and warrant further investigation.
We were also limited in our ability to examine how dosage, timing, or duration of SSRI use may impact epigenetic gestational age deceleration. Previous studies have suggested that SSRI exposure during late pregnancy is more strongly associated with adverse offspring outcomes compared to exposure during early pregnancy [42]. There is also evidence that longer duration of SSRI exposure predicts poorer offspring outcomes [43]. As such, future studies would benefit from examining how timing, dosage, and length of prenatal SSRI exposure may differentially impact epigenetic gestational age deceleration. Finally, information about the specific disorder for which SSRIs were prescribed to our participants was not available. It is possible that some women were prescribed an SSRI for anxiety or another psychiatric disorder, which would introduce noise when comparing prenatal SSRI exposure to maternal prenatal depression.
The present study has a number of methodological strengths. Its prospective, longitudinal design eliminated the reporting bias associated with retrospective recall, enabling us to collect more accurate data on maternal depressive symptoms and stress throughout pregnancy. Maternal medication use was also verified by study clinicians. Concurrent measurement of a variety of prenatal exposures enabled us to parse apart the relative contribution of these factors on epigenetic gestational ageing, which has been missing from previous studies that examine stress, depression, or SSRI use in isolation. In addition, many previous studies examining the effects of prenatal SSRI exposure have been unable to parse out confounding influences of other prenatal exposures, such as tobacco. Given that previous studies have demonstrated higher rates of smoking among depressed mothers [56] and that prenatal tobacco exposure has been associated with epigenetic gestational age in previous samples [57] and in the present study, it is particularly important to account for this confound when examining the impact of maternal depression on child epigenetic ageing.
In line with the developmental origins of health and disease (DOHaD) model, the present study corroborates and extends previous research that shows an association between prenatal exposures and epigenetic gestational ageing. In addition to being one of the first studies to examine the impact of prenatal exposures on epigenetic gestational age, it is also the very first study to specifically examine the impact of maternal SSRI use. Our findings specifically suggest that psychotropic medication use (in this case SSRIs) during pregnancy may delay epigenetic gestational ageing. This is particularly meaningful given the suggested role of epigenetic age deceleration in risk for cognitive and emotional problems among children of depressed mothers [5]. Studies such as this are necessary in order to enhance our knowledge of how psychotropic medications interact with the intrauterine environment to shape child development, which will ultimately help women and their providers weigh the benefits and risks of medication use during pregnancy.
Funding Statement
This work was supported by the National Institutes of Health under Grants RC1MH088609 and 5K12HD085850-04. Brooke G. McKenna and Cassandra L. Hendrix are supported by the National Science Foundation Graduate Research Fellowship Program under Grant DGE-1444932.
Disclosure statement
No potential conflict of interest was reported by the authors.
References
- [1].Ashley JM, Harper BD, Arms-Chavez CJ, et al. Estimated prevalence of antenatal depression in the US population. Arch Women’s Mental Health. 2016;19(2):395–400. [DOI] [PubMed] [Google Scholar]
- [2].Field T. Prenatal depression effects on early development: A review. Infant Behav Dev. 2011;34(1):1–14. [DOI] [PubMed] [Google Scholar]
- [3].Silveira PP, Portella AK, Goldani MZ, et al. Origens desenvolvimentistas da saúde e da doença (DOHaD). J Pediatr (Rio J). 2007;83(6):494–504. [DOI] [PubMed] [Google Scholar]
- [4].Cardenas A, Faleschini S, Cortes Hidalgo A, et al. Prenatal maternal antidepressants, anxiety, and depression and offspring DNA methylation: epigenome-wide associations at birth and persistence into early childhood. Clin Epigenetics. 2019;11(1):1–11. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [5].Suarez A, Lahti J, Czamara D, et al. The Epigenetic Clock at Birth: associations With Maternal Antenatal Depression and Child Psychiatric Problems. J Am Acad Child Adolesc Psychiatry. 2018;57(5):321–328.e2. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [6].Spiers H, Hannon E, Schalkwyk LC, et al. Methylomic trajectories across human fetal brain development. Genome Res. 2015;25(3):338–352. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [7].Hannon E, Schendel D, Ladd-Acosta C, et al. Variable DNA methylation in neonates mediates the association between prenatal smoking and birth weight. Philos Trans R Soc B. 2019;374:1770. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [8].Cao-Lei L, Massart R, Suderman MJ, et al. DNA methylation signatures triggered by prenatal maternal stress exposure to a natural disaster: project ice storm. PLoS ONE. 2014;9(9). DOI: 10.1371/journal.pone.0107653 [DOI] [PMC free article] [PubMed] [Google Scholar]
- [9].Mulligan CJ, D’Errico NC, Stees J, et al. Methylation changes at NR3C1 in newborns associate with maternal prenatal stress exposure and newborn birth weight. Epigenetics. 2012;7(8):853–857. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [10].Radtke KM, Ruf M, Gunter HM, et al. Transgenerational impact of intimate partner violence on methylation in the promoter of the glucocorticoid receptor. Transl Psychiatry. 2011;1(June):1–6. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [11].Horvath S, Raj K.. DNA methylation-based biomarkers and the epigenetic clock theory of ageing. Nat Rev Genet. 2018;19(6):371–384. [DOI] [PubMed] [Google Scholar]
- [12].Chen BH, Marioni RE, Colicino E, et al. DNA methylation-based measures of biological age: meta-analysis predicting time to death. Aging (Albany NY). 2016;8(9):1844–1865. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [13].Marioni RE, Shah S, McRae AF, et al. The epigenetic clock is correlated with physical and cognitive fitness in the Lothian Birth Cohort 1936. Int J Epidemiol. 2015;44(4):1388–1396. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [14].Quach A, Levine ME, Tanaka T, et al. Epigenetic clock analysis of diet, exercise, education, and lifestyle factors. Aging (Albany NY). 2017;9(2):419–446. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [15].Zannas AS, Arloth J, Carrillo-Roa T, et al. Lifetime stress accelerates epigenetic aging in an urban, African American cohort: relevance of glucocorticoid signaling. Genome Biol. 2015;16(1):1–12. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [16].Knight AK, Craig JM, Theda C, et al. An epigenetic clock for gestational age at birth based on blood methylation data. Genome Biol. 2016;17(1):1–11. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [17].Simpkin AJ, Howe LD, Tilling K, et al. The epigenetic clock and physical development during childhood and adolescence: longitudinal analysis from a UK birth cohort. Int J Epidemiol. 2017;46(2):549–558. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [18].Huang J, Zhu T, Qu Y, et al. Prenatal, perinatal and neonatal risk factors for intellectual disability: A systemic review and meta- Analysis. PLoS ONE. 2016;11(4):1–12. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [19].Mathewson KJ, Chow CHT, Dobson KG, et al. Mental health of extremely low birth weight survivors: A systematic review and meta-analysis. Psychol Bull. 2017;143(4):347–383. [DOI] [PubMed] [Google Scholar]
- [20].Dipietro JA. Maternal stress in pregnancy: considerations for fetal development. J Adolesc Health. 2012;51(2SUPPL):S3–S8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [21].Olivier JDA, Åkerud H, Sundström Poromaa I. Antenatal depression and antidepressants during pregnancy: unraveling the complex interactions for the offspring. Eur J Pharmacol. 2015;753:257–262. [DOI] [PubMed] [Google Scholar]
- [22].Halligan SL, Murray L, Martins C, et al. Maternal depression and psychiatric outcomes in adolescent offspring: A 13-year longitudinal study. J Affect Disord. 2007;97(1–3):145–154. [DOI] [PubMed] [Google Scholar]
- [23].Lucassen PJ, Naninck EFG, van Goudoever JB, et al. Perinatal programming of adult hippocampal structure and function; Emerging roles of stress, nutrition and epigenetics. Trends Neurosci. 2013;36(11):621–631. [DOI] [PubMed] [Google Scholar]
- [24].Meaney MJ, Szyf M, Seckl JR. Epigenetic mechanisms of perinatal programming of hypothalamic-pituitary-adrenal function and health. Trends Mol Med. 2007;13(7):269–277. [DOI] [PubMed] [Google Scholar]
- [25].Schroeder JW, Smith AK, Brennan PA, et al. DNA methylation in neonates born to women receiving psychiatric care. Epigenetics. 2012;7(4):409–414. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [26].Non AL, Binder AM, Kubzansky LD, et al. Genome-wide DNA methylation in neonates exposed to maternal depression, anxiety, or SSRI medication during pregnancy. Epigenetics. 2014;9(7):964–972. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [27].Viuff A-CF, Pedersen LH, Kyng K, et al. Antidepressant medication during pregnancy and epigenetic changes in umbilical cord blood: a systematic review. Clin Epigenetics. 2016;8(1):0–11. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [28].Cohen LS, Altshuler LL, Harlow BL, et al. Relapse of major depression during pregnancy in women who maintain or discontinue antidepressant treatment. JAMA. 2006;295(5):499–507. DOI: 10.1001/jama.295.5.499 [DOI] [PubMed] [Google Scholar]
- [29].Buuren SV, Groothuis-Oudshoorn K. Mice: multivariate imputation by chained equations in R. J Stat Software. 2010:1–68. [Google Scholar]
- [30].Schroeder JW, Conneely KN, Cubells JF, et al. Neonatal DNA methylation patterns associate with gestational age. Epigenetics. 2011;6(12):1498–1504. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [31].Houseman EA, Accomando WP, Koestler DC, et al. DNA methylation arrays as surrogate measures of cell mixture distribution. BMC Bioinformatics. 2012;13(1). DOI: 10.1186/1471-2105-13-86 [DOI] [PMC free article] [PubMed] [Google Scholar]
- [32].Chan SF, La Greca AM. Perceived Stress Scale (PSS). Encyc Behav Med. 2013;1454–1455. DOI: 10.1007/978-1-4419-1005-9_773 [DOI] [Google Scholar]
- [33].Roberti JW, Harrington LN, Storch EA. Further Psychometric Support for the 10-Item Version of the Perceived Stress Scale. J College Couns. 2006;9(2):135–147. [Google Scholar]
- [34].Johnson KC, Smith AK, Stowe ZN, et al. Preschool outcomes following prenatal serotonin reuptake inhibitor exposure: differences in language and behavior, but not cognitive function. J Clin Psychiatry. 2016;77(2):e176–e182. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [35].Arntzen A, Andersen AMN. Social determinants for infant mortality in the Nordic countries, 1980-2001. Scand J Public Health. 2004;32(5):381–389. [DOI] [PubMed] [Google Scholar]
- [36].Wadhwa PD, Sandman CA, Porto M, et al. The association between prenatal stress and infant birth weight and gestational age at birth: A prospective investigation. Am J Obstet Gynecol. 1993;169(4):858–865. [DOI] [PubMed] [Google Scholar]
- [37].Khouja JN, Simpkin AJ, O’Keeffe LM, et al. Epigenetic gestational age acceleration: A prospective cohort study investigating associations with familial, sociodemographic and birth characteristics. Clin Epigenetics. 2018;10(1):1–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [38].Toffoli LV, Rodrigues GM, Oliveira JF, et al. Maternal exposure to fluoxetine during gestation and lactation affects the DNA methylation programming of rat’s offspring: modulation by folic acid supplementation. Behav Brain Res. 2014;265:142–147. [DOI] [PubMed] [Google Scholar]
- [39].Barch DM, Harms MP, Tillman R, et al. Early childhood depression, emotion regulation, episodic memory, and hippocampal development. J Abnorm Psychol. 2019;128(1):81–95. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [40].Glover ME, Clinton SM. Of rodents and humans: A comparative review of the neurobehavioral effects of early life SSRI exposure in preclinical and clinical research. Int J Dev Neurosci. 2016;51(May):50–72. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [41].Eke AC, Saccone G, Berghella V. Selective serotonin reuptake inhibitor (SSRI) use during pregnancy and risk of preterm birth: a systematic review and meta-analysis. BJOG. 2016;123(12):1900–1907. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [42].Oberlander TF, Warburton W, Misri S, et al. Neonatal outcomes after prenatal exposure to selective serotonin reuptake inhibitor antidepressants and maternal depression using population-based linked health data. Arch Gen Psychiatry. 2006;63(8):898–906. [DOI] [PubMed] [Google Scholar]
- [43].Casper RC, Gilles AA, Fleisher BE, et al. Length of prenatal exposure to selective serotonin reuptake inhibitor (SSRI) antidepressants: effects on neonatal adaptation and psychomotor development. Psychopharmacology (Berl). 2011;217(2):211–219. [DOI] [PubMed] [Google Scholar]
- [44].Hanley GE, Brain U, Oberlander TF. Infant developmental outcomes following prenatal exposure to antidepressants, and maternal depressed mood and positive affect. Early Hum Dev. 2013;89(8):519–524. [DOI] [PubMed] [Google Scholar]
- [45].Pawluski JL, Brain UM, Underhill CM, et al. Prenatal SSRI exposure alters neonatal corticosteroid binding globulin, infant cortisol levels, and emerging HPA function. Psychoneuroendocrinology. 2012;37(7):1019–1028. [DOI] [PubMed] [Google Scholar]
- [46].Pedersen LH, Henriksen TB, Olsen J. Fetal exposure to antidepressants and normal milestone development at 6 and 19 months of age. Pediatrics. 2010;125(3). DOI: 10.1542/peds.2008-3655 [DOI] [PubMed] [Google Scholar]
- [47].Salisbury AL, Wisner KL, Pearlstein T, et al. Newborn neurobehavioral patterns are differentially related to prenatal maternal major depressive disorder and serotonin reuptake inhibitor treatment. Depress Anxiety. 2011;28(11):1008–1019. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [48].Bérard A, Boukhris T, Sheehy O. Selective serotonin reuptake inhibitors and autism: additional data on the Quebec Pregnancy/Birth Cohort. Am J Obstet Gynecol. 2016;215(6):803–805. [DOI] [PubMed] [Google Scholar]
- [49].Nulman I, Koren G, Rovet J, et al. Neurodevelopment of children following prenatal exposure to venlafaxine, selective serotonin reuptake inhibitors,or untreated maternal depression. Am J Psychiatry. 2012;169(11):1165–1174. [DOI] [PubMed] [Google Scholar]
- [50].Oberlander TF, Papsdorf M, Brain UM, et al. Prenatal effects of selective serotonin reuptake inhibitor antidepressants, serotonin transporter promoter genotype (SLC6A4), and maternal mood on child behavior at 3 years of age. Arch Pediatr Adolesc Med. 2010;164(5):444–451. [DOI] [PubMed] [Google Scholar]
- [51].Smearman EL, Hendrix CL, Winiarski DA, et al. School-age social behavior and pragmatic language ability in children with prenatal serotonin reuptake inhibitor exposure. Dev Psychopathol. 2019;1–10. DOI: 10.1017/S0954579418001372 [DOI] [PMC free article] [PubMed] [Google Scholar]
- [52].Bohlin J, Håberg SE, Magnus P, et al. Prediction of gestational age based on genome-wide differentially methylated regions. Genome Biol. 2016;17(1):1–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [53].Bakulski KM, Feinberg JI, Andrews SV, et al. DNA methylation of cord blood cell types: applications for mixed cell birth studies. Epigenetics. 2016;11(5):354–362. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [54].Christensen BC, Houseman EA, Marsit CJ, et al. Aging and environmental exposures alter tissue-specific DNA methylation dependent upon CPG island context. PLoS Genet. 2009;5(8):e1000602. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [55].Rampono J, Simmer K, Ilett KF, et al. Placental transfer of SSRI and SNRI antidepressants and effects on the neonate. Pharmacopsychiatry. 2009;42(3):95–100. [DOI] [PubMed] [Google Scholar]
- [56].Smedberg J, Lupattelli A, Mårdby AC, et al. The relationship between maternal depression and smoking cessation during pregnancy—a cross-sectional study of pregnant women from 15 European countries. Arch Women’s Mental Health. 2015;18(1):73–84. [DOI] [PubMed] [Google Scholar]
- [57].Simpkin AJ, Hemani G, Suderman M, et al. Prenatal and early life influences on epigenetic age in children: A study of mother-offspring pairs from two cohort studies. Hum Mol Genet. 2016;25(1):191–201. [DOI] [PMC free article] [PubMed] [Google Scholar]