Prenatal exposure to environmental air pollution and psychosocial stress jointly contribute to the epigenetic regulation of the serotonin transporter gene in newborns

Antenatal exposures to maternal stress and to particulate matter with an aerodynamic diameter of less than 2.5 μm (PM2.5) have been independently associated with developmental outcomes in early infancy and beyond. Knowledge about their joint impact, biological mechanisms of their effects and timing-effects, is still limited. Both PM2.5 and maternal stress exposure during pregnancy might result in altered patterns of DNA methylation in specific stress-related genes, such as the serotonin transporter gene (SLC6A4 DNAm), that might, in turn, influence infant development across several domains, including bio-behavioral, cognitive and socio-emotional domains. Here, we investigated the independent and interactive influence of variations in antenatal exposures to maternal pandemic-related stress (PRS) and PM2.5 on SLC6A4 DNAm levels in newborns. Mother–infant dyads (N = 307) were enrolled at delivery during the COVID-19 pandemic. Infants’ methylation status was assessed in 13 CpG sites within the SLC6A4 gene’s region (chr17:28562750–28562958) in buccal cells at birth and women retrospectively report on PRS. PM2.5 exposure throughout the entire gestation and at each gestational trimester was estimated using a spatiotemporal model based on residential address. Among several potentially confounding socio-demographic and health-related factors, infant’s sex was significantly associated with infants’ SLC6A4 DNAm levels, thus hierarchical regression models were adjusted for infant’s sex. Higher levels of SLC6A4 DNAm at 6 CpG sites were found in newborns born to mothers reporting higher levels of antenatal PRS and greater PM2.5 exposure across gestation, while adjusting for infant’s sex. These effects were especially evident when exposure to elevated PM2.5 occurred during the second trimester of pregnancy. Several important brain processes (e.g., synaptogenesis and myelination) occur during mid-pregnancy, potentially making the second trimester a sensitive time window for the effects of stress-related exposures. Understanding the interplay between environmental and individual-level stressors has important implications for the improvement of mother-infant health during and after the pandemic.


INTRODUCTION
The last decades have seen a growing appreciation for the roles of environmental exposures in shaping brain development and influencing later risk for psychopathology, with increasing attention directed toward prenatal exposures [1].Pregnancy represents an important window of susceptibility, characterized by rapid growth and brain development, which makes the fetus particularly vulnerable to any insults and changes in the intrauterine environment with possible long-term consequences [2].Antenatal exposures to maternal stress and to particulate air pollution are increasingly recognized as risk factors, being independently associated with altered developmental outcomes in early infancy and beyond [3,4].Initial evidence also suggests that timing of exposure during gestation might influence the nature and degree of these effects, with environmental stressors interfering with specific fetal brain processes occurring at different gestational ages.For example, key neurogenetic events occurs during fetal cortical histogenesis with a predominance of neuronal proliferation and migration processes early in pregnancy, whereas growth of axons and dendrites and synaptogenesis during the second and third trimesters [5].
Amassing evidence now indicates that higher than typical levels of maternal stress during pregnancy are associated with a wide range of biological, behavioral, cognitive and socio-emotional outcomes in children (reviewed in 4).The COVID-19 outbreak, and related lockdown, represented an unprecedented stressful experience for families worldwide, raising serious concerns for its potential broad impact on vulnerable populations such as pregnant women and their offspring.Evidence from pregnant cohorts around the world suggests that women experienced elevated symptoms of stress, anxiety and depression during the COVID-19 pandemic [6][7][8][9].Furthermore, early reports indicate that antenatal exposure to maternal pandemic-related stress (PRS) is associated with early infants' developmental outcomes, including temperament [10][11][12][13], socio-emotional and socio-cognitive development [14,15], motor outcomes [16] and even brain connectivity patterns [17].
Exposure to air pollution during pregnancy is another wellacknowledged risk factor for later physical and mental development [3].The most investigated air pollutant is fine particulate matter with an aerodynamic diameter of less than 2.5 microns (PM 2.5 ), for which strong molecular evidence exists concerning its neurotoxic effects [18][19][20].Mounting data prospectively links antenatal PM 2.5 exposure and a range of neurodevelopmental outcomes in children including poorer cognitive development [21], higher behavioral problems [22] and structural brain alterations [23].
The mechanisms that underline the associations between antenatal environmental exposures and altered neurodevelopment in the offspring are largely unknown and likely encompass neuroendocrine, inflammatory and epigenetic pathways [24][25][26].Concerning the latter, DNA methylation, which involves adding a methyl group to a DNA base cytosine in a cytosine nucleotide-phosphate-guanine nucleotide (CpG) sequence [27], is the most studied epigenetic marker.In particular, methylation of genes crucially involved in stress regulation and mental health, such as methylation of the serotonin transporter (SLC6A4 DNAm) gene, is thought to play a crucial role in mediating the effects of antenatal exposures on fetal development [28].Few studies investigated the association between some measures of maternal antenatal stress and infants' SLC6A4 DNAm and these have yielded mixed findings [29][30][31][32].Specifically, DeLano and colleagues [30] reported higher exposure to maternal community disadvantage during pregnancy to be associated with higher infants' SLC6A4 DNAm 4 weeks after birth in a relatively small (N = 53) high-risk sample.Likewise, Wankerl and collaborators [32] reported a positive association between prenatal stress and adults' levels of SLC6A4 DNAm at specific CpG sites.In contrast, Devlin and colleagues [31] found a negative association between maternal depressive symptoms in the second (but not third) trimester of pregnancy and newborns' SLC6A4 DNAm levels, whereas no association between antenatal maternal stress and levels of methylation of the SLC6A4 gene has also been reported [29].
Initial recent evidence indicates a positive association between maternal antenatal PRS and infants' SLC6A4 DNAm at birth [10].Furthermore, higher SLC6A4 DNAm was reported at birth in mothers and infants exposed to the COVID-19 lockdown in the second and third trimester of pregnancy [33].Likewise, evidence exists for an association between antenatal exposure to PM2.5 and altered patterns of DNA methylation in newborns [34] particularly in several genes that are involved in oxidative stress, inflammation, mitochondrial function, fetal growth and cardio-respiratory health outcomes.Some of these associations were also observed later in childhood [35].Noteworthy, air pollution particles have been found to be able to pass into and across the placental barrier [36,37] thus, potentially, directly affecting fetal development.Even though the direction of the association remains unclear, there is converging evidence for a window of higher susceptibility to air pollutants exposure at the beginning of pregnancy [38,39].Indeed, some studies suggest a positive correlation between PM exposure during the first trimester and placental global DNA methylation levels [40], while others bring evidence for a significant inverse correlation [41] in the same gestational period.
Importantly, environmental exposures are not likely to represent isolated events.Rather, multiple environmental stressors are likely to occur simultaneously, with the final developmental outcomes being the result of their complex interplay [42,43].In this light, it is noteworthy that neurodevelopmental disorders are characterized by complex and heterogeneous phenotypes and multifactorial etiologies [44].On one hand, multiple stressors can contribute to a cumulative allostatic burden, reflecting the overall wear-and tear of the body from repeated or chronic exposure to environmental stressors, leading to epigenetic and physiological changes.This is in line with a "multiple hit" [45] or "cumulative stress" hypothesis [46].For example, antenatal exposure to air pollution and maternal stress can both contribute to a series of biological events, involving for example oxidative stress or inflammation, which are thought to play a central role in fetal programming [25,26].On the other hand, effect modifications are also possible, for example, with the effects of an individual-level exposure (e.g., maternal antenatal stress) being modified by the simultaneous environmental exposure (e.g., air pollution).Literature on the joint impact of prenatal exposures to environmental and psychosocial stressors is still sparse [47].A larger negative effect of antenatal exposure to PM 2.5 on birth weight has been found among women living in lower socioeconomic status neighborhoods, used as a proxy for psychosocial stress [48].Likewise, stronger associations between antenatal PM 2.5 exposure and child's asthma or wheeze have been reported among children antenatally exposed to higher levels of maternal stress [49,50].A better understanding of the joint effects of antenatal environmental and psychosocial stress-related exposures on infant development and on the possible underlying biological mechanisms is needed.The current study begins to fill this gap by investigating, for the first time, the independent and interactive influence of variations in antenatal exposures to maternal PRS and PM 2.5 on SLC6A4 DNAm levels in newborns.As associations may differ as a function of timing of gestation, in a supplementary set of analyses we explore the effects of trimester-specific PM 2.5 levels, jointly with maternal PRS, on DNA methylation.We predicted higher levels of SLC6A4 DNAm in infants antenatally exposed to both higher levels of maternal PRS and PM 2.5.We made no a priori predictions for possible timing effects of PM 2.5 exposure on levels of DNA methylation due to limited available literature.

Participants and procedures
Women were recruited from May 2020 to February 2021 in ten neonatal units in Northern Italy as part of a longitudinal multi-centric research project, the MOM-Cope Study (for full details see [51]).Inclusion criteria were: maternal age over 18 years, absence of prenatal and perinatal diseases or injuries, and term delivery (i.e., from 37 + 0 to 41 + 6 weeks of gestation).As the data collection occurred during the COVID-19 health emergency, a negative PCR test for SARS-CoV-2 at delivery was an additional inclusion criterion.Women were first contacted at the childbirth classes or immediately following birth.Three hundred seven mother-infant dyads agreed to participate and provided complete data concerning maternal PRS and infant methylation at the neonatal assessment (6-48 h after birth).Samples characteristics are reported in Table 1.Mother-infant dyads were all enrolled in major urban centers of Northern Italy, such as

Measures
Maternal PRS.At delivery mothers retrospectively reported on their PRS) during pregnancy through an ad-hoc questionnaire (Supplementary File S1) [51].The questionnaire included six 5-point Likert scale items (1, not at all; 5, very much) on the emotional stress response to the COVID-19 emergency during pregnancy.An average maternal antenatal PRS score was obtained, ranging from 1 (low) to 5 (high).
Antenatal exposure to air pollution.Based on prenatal address locations of the mother-infant dyads, fine particulate matter (PM 2.5 ) ground-level concentrations were derived from an ensemble of air quality model results, as provided by the 'Copernicus Atmospheric Monitoring Service', one of the six services that form Copernicus, the European Union's Earth observation program which looks at our planet's environment [see https://atmosphere.copernicus.eu/].The air quality model results (we are using here the 'reanalysis' product) were spatially and temporally averaged to obtain measures of the PM 2.5 concentrations at the municipality level as monthly PM 2.5 average values.Total exposure throughout pregnancy, as well as exposure during each trimester, were computed based on birth date.The geographical distribution of air pollution exposure in our sample is displayed in Fig. 1 and further detailed in Supplementary Fig. S2.
Infant SLC6A4 DNAm.Infant's saliva was collected using the OraCollect for Pediatrics kit OC-175 (DNA Genotek, Ottawa, Canada) between 6 and 24 h from delivery.Methylation assessment was conducted according to previous validated procedures [33,52].The genomic DNA was extracted following manufacturer's protocols and its quality was assessed using a Qubit fluorometer Invitrogen, Thermo Fisher Scientific, Waltham, Massachusetts, USA.The methylation status of 13 CpG sites within the Exon 1 region of the SLC6A4 gene (chr17:28562750-28562958) was assessed by PCR amplification of bisulfite-treated DNA followed by Next Generation Sequencing on a NEXTSeq-500 (Illumina, San Diego, California, USA).Samples with less than 1000 reads (N = 13) were excluded.Target regions were selected based on available literature on the association between early stress exposures and epigenetic regulation of the serotonin transporter gene [28].The position of each CpG site is reported in Supplementary File S1.

Data reduction
A principal component analysis (PCA) was used to reduce the number of CpG sites into a smaller set of factors which accounted for shared variance, as done in prior works [10,53].The PCA was carried out setting a varimax rotation, suppressing coefficients lower than 0.30, and extracting principal components (PCs) based on eigenvalues greater than one [e.g., [54]].A four-factor solution, where the targeted CpG sites were aggregated in four main factors, showed the best fit to the data (

Plan of analysis
Variables were first examined for outliers and skewness.Samples greater than 3 SD from the mean (N = 7) were removed.Pearson correlations and independent samples t tests were employed to explore the potential effect of sociodemographic and health-related variables on infant methylation levels.All variables found to be significantly associated with the outcomes examined were included as covariates in subsequent analyses.Separate hierarchical regression analyses were performed to evaluate the independent and interactive effects of maternal PRS and exposure to air pollution on infant methylation levels.Covariates were entered in the first step.Maternal PRS and exposure to air pollution were entered in a second step to assess their independent effect on infant's methylation levels, whereas their interactive effect was included in a third step.Maternal PRS and exposure to air pollution were centered at the sample means and entered  in the models as continuous variables, infant's sex was centered at males.Significant interactions were examined with simple slope analyses probing the effects of one independent variable (i.e., maternal PRS) on the outcome (i.e., infant methylation levels) at higher (+1 SD) and lower (−1 SD) levels of the other independent variable (i.e., antenatal PM 2.5 exposure).Furthermore, significant interactions were followed up by supplementary exploratory analyses to investigate possible timing effects related to the trimester of exposure to air pollution on infant's SLC6A4 DNAm levels.Specifically, as pollution levels at each trimester were moderately correlated, three separate regression models were performed including antenatal PM 2.5 exposure at each trimester, rather than the total estimate across gestation, as a predictor variable along with maternal PRS on infants' PC1 SLC6A4 DNAm levels.All the analyses were performed using Jamovi 2.2.5.0 for Windows (The Jamovi Project, 2021).

Preliminary analyses
Preliminary unadjusted bivariate correlations among study variables showed weak associations between infant's PC1 SLC6A4 DNAm levels and respectively, PM 2.5 levels across gestation (r(292) = 0.19, p = 0.001) and maternal antenatal PRS levels (r(292) = 0.12, p = 0.034).In contrast, maternal PRS levels and exposure to PM 2.5 were not significantly associated (r(318) = 0.04, p = 0.48).In a series of univariate correlation analyses we evaluated the associations between sociodemographic factors (i.e., parental age and education) and health-related (i.e., birth weight, gestational age, Apgar scores, delivery mode) factors and infant SLC6A4 DNAm levels.These variables were not significantly associated with infant's methylation levels (all ps > 0.05) and were trimmed from subsequent models.Independent samples t tests showed significant sex-differences in infant SLC6A4 DNAm levels at CpG site 06 t(305) = 2.28, p = 0.02 with higher levels of methylation in females as compared to males.Furthermore, sexdependent differences in levels of methylation soon after birth have been previously reported [55].Thus, infant's sex was included as a covariate in subsequent models.
Results of the hierarchical linear regression analyses predicting infants' PC2 SLC6A4 DNAm showed no significant independent or interactive effects of maternal PRS and air pollution exposure on infants' PC2 SL6CA4 DNAm levels (all ps > 0.05).

C
Fig. 3 Association between maternal PRS and infants' SLC6A4 DNAm PC1 as a function of antenatal exposure to air pollution at each trimester of pregnancy.Association between maternal PRS and infants' SLC6A4 DNAm PC1 as a function of antenatal exposure to air pollution at each trimester of pregnancy: panel A refers to 1st trimester (β = −0.08,p = 0.18), B refers to 2nd trimester (β = 0.12, p = 0.02) and C refers to 3rd trimester (β = 0.05, p = 0.48).Note.For illustrative purposes, the association is plotted for infants antenatally exposed to higher (+1 SD), average and lower (−1SD) levels of PM 2.5 at each trimester.Bands represent 95% CI.

DISCUSSION
This is the first study to investigate synergistic effects of antenatal maternal PRS and PM 2.5 exposure on newborns' SLC6A4 DNAm levels at birth.Higher levels of methylation of the SLC6A4 gene at 6 CpG sites were found in newborns born to mothers reporting higher levels of PRS in pregnancy and greater PM 2.5 exposure across gestation.Further, exploratory analyses provided preliminary evidence suggesting that these effects might become especially evident when exposure to elevated PM 2.5 levels occurred during the second trimester of pregnancy.Of note, the observed associations were found soon after birth and were independent of infant's sex.Evidence from both animal and human studies suggests that pregnancy is an especially vulnerable period for exposure to environmental stressors on the developing fetus [56].Both individual-level stressors, such as maternal symptoms of anxiety, depression and stress during pregnancy, and environmentallevel stressors, such as exposure to air pollution or toxicants, occurring during gestation have been studied and found to be independently associated with behavioral and physiological outcomes in the offspring [3,4].However, literature concerning the effect of co-occurring exposures is still sparse.
We previously showed that variations in levels of maternal PRS reported by women who were pregnant during the COVID-19 outbreak and subsequent lockdown in Northern Italy were associated with the levels of newborns' SLC6A4 methylation and further predicted infant's temperament [9].Our current findings extend those results by showing higher levels of SLC6A4 DNAm at 6 CpG sites in newborns co-exposed to both higher mean PM 2.5 levels across gestation and higher levels of maternal antenatal PRS.In contrast, the association between maternal PRS during pregnancy and infant's methylation was not significant at lower levels of PM 2.5 , raising important conceptual, methodological and clinical implications.
First, the findings of an interactive effect between antenatal exposure to maternal PRS and PM 2.5 levels on newborns' methylation are suggestive of a cumulative or joint effect of individual-level and environmental stressors and are in line with emerging literature on the effects of multiple antenatal exposures [43].In particular, current findings might indicate that concurrent (and possibly prior) exposure to air pollution could exacerbate the effects of exposure to maternal PRS on infant methylation levels.This might occur through sensitization mechanisms whereby exposure to high levels of PM 2.5 sets fetal biological systems to be more susceptible to the effect of concurrent or later adversities.It is important to underline that exposure to air pollution has itself been associated with increased levels of stress in non-pregnant individuals [57,58].Initial evidence also showed a positive association between exposure to PM 2.5 and maternal distress in pregnant women (e.g., [59]).In contrast, we did not report any significant association between PM 2.5 exposure and maternal PRS during pregnancy in the current sample.While it is tempting to hypothesize that these two factors act independently to affect infant methylation levels, further evidence is needed.Noteworthy, several biological pathways are known to be affected by both air pollution exposure and psychological distress and may thus be activated by both exposures to determine epigenetic changes and impact offspring development.These include, but are not limited to, inflammation, oxidative stress and endocrine disruption [26,60].For example, increased maternal systemic oxidative stress and levels of proinflammatory cytokines related to air pollution and stress exposure may lead to placental dysfunction, increase fetal oxidative stress and affect fetal development.
Interestingly, the current findings align with mounting evidence showing that exposure to a broad range of early life adversities is related to a pattern of hyper-methylation at specific CpG sites located within the SLC6A4 exon 1 (for a review see 28).The functional relevance of the current findings as well as possible implications for later risk of health and disease is yet to be determined.It is noteworthy that patterns of methylation within the promoter region of SLC6A4, widely overlapping with the exon 1 CpG island, either alone or in interaction with individuals' genotype, have been implicated in adults' brain function (e.g., [61,62]), stress reactivity (e.g., [63]) and several psychiatric disorders in adulthood [64].
Second, from a methodological account, the current findings emphasize the need for research into the effects of antenatal maternal stress to account for the influence of environmental stressors, as overlooking this aspect might lead to misguiding findings and possibly explain inconsistencies in the literature.
Third, as air pollution is a global health problem, with around 86% of urban population living in places exceeding air quality recommended limits [65] and exposure to PM 2.5 being listed among the top five mortality risk factors worldwide [66], the current findings have important practical implications.In particular, while replication is needed, they suggest that being exposed to lower levels of air pollution might buffer the effects of higher levels of antenatal maternal PRS on infant methylation levels.In this sense, they might indicate that, in light of limited resources, greater priority to efforts to improve environmental air quality should be directed to disadvantaged communities that may be at greater risk of experiencing multiple psychosocial stressors as well.This might also be protective in the event of future major stressful experiences, such as a new pandemic.Ultimately, a multi-level approach that targets environmental stressors as well as individual stressors is likely to make the greatest impact on perinatal and infant outcomes.
Noteworthy, exploratory analyses indicate that the second trimester of pregnancy is the most sensitive window for a significant independent effect of PM 2.5 exposure on infant SLC6A4 DNAm levels.Likewise, a significant interaction between PM 2.5 exposure in the second trimester and maternal PRS during pregnancy on levels of infant SLC6A4 methylation was reported.While averaging PM 2.5 exposure across pregnancy is thought to provide a robust marker of total PM 2.5 exposure across gestation, previous studies indicate timing-related variability in the effects of air pollution exposure on neonatal methylation [67,68].Fetal brain development is characterized by a progression of sensitive time windows, during which specific anatomical and physiological structures undergo rapid development and are particularly vulnerable to environmental stressors [69].Our findings align with emerging evidence suggesting timing-specific association between antenatal adverse exposures and pattern of neonatal DNA methylation.For example, maternal community deprivation (an index derived by combining multiple community-levels indicators such as income and highest level of education) during the second or third trimester of pregnancy was found to be significantly associated with greater infant DNA methylation in eight CpG sites of the SLC6A4 gene [30].Likewise, exposure to the COVID-19 lockdown in the second and third trimester of pregnancy was associated with higher mother and infants' SLC6A4 DNAm [33].Mechanisms for a timing-specific effects are still poorly understood.Noteworthy, important processes in the fetal brain begin in mid-pregnancy, including myelination, starting from the subcortical regions, or synaptogenesis [70].Furthermore, the fetal stress response system is not fully developed until the second trimester.Thus, stress-related exposures at this time might interfere with these important brain development processes [71].
Some study limitations warrant attention.First, while we employed outdoor PM 2.5 exposure data based on mothers' residential addresses during pregnancy, data concerning indoor air pollution exposure are missing.Furthermore, as PM 2.5 concentrations were assessed at the municipality levels, clustering effects might potentially affect current estimates.Secondly, newborns' SLC6A4 DNAm was peripherally assessed in buccal cells.It is unclear how epigenetic variation in the peripheral tissue relates to epigenetic changes within the brain.Partial evidence exists on the cross-tissue consistency of DNA methylation measures in humans.It is plausible that the methylation status of genes that have widespread effects and actions across central and peripheral tissues, such as the SLC6A4 gene, may be tissueand site-specific [72,73].Third, maternal antenatal PRS was assessed through an ad hoc questionnaire developed ad hoc for this study, prioritizing sensitivity to the specific and unprecedented nature of COVID-19 emergency over measure standardization.Noteworthy, women were asked to reply to the items thinking about their pregnancy in general, thus no information about maternal PRS at each trimester of gestation was available.However, it is also important to consider that maternal reports of distress (e.g., anxiety or depression) are characterized by their chronicity and by high levels of within-individual stability throughout pregnancy [74].Thus, prenatal exposure to maternal stress is likely to be relatively continuous over gestation rather than discrete, leaving limited leverage for testing possible timing effects.For this reason, our exploratory analyses on possible timing-effects focused on exposure to air pollution where intensity and timing of exposure can be easily determined.Fourth, it is important to underline that the current study is not able to address the functional and adaptive consequences of the observed patterns of methylation in infants exposed to high levels of antenatal PM 2.5 and maternal PRS.Future studies are needed to investigate how these early mechanisms might predict later infant neurodevelopmental trajectories.Lastly, while newborns' methylation was assessed soon after birth, thus limiting the effects of postnatal influences [75][76][77], we cannot rule out unmeasured confounding factors affecting neonatal levels of methylation and no causal conclusions should be drawn.

CONCLUSIONS
While the literature on the independent effect of antenatal air pollution and maternal stress exposure on infant development is burgeoning, little is known about their interactive effects and possible gestational windows of sensitivity.The current study is among the first to suggest that greater exposure to PM 2.5 , particularly during the second trimester of pregnancy, can exacerbate the effect of maternal antenatal PRS on levels of SLC6A4 DNAm at 6 CpG sites in newborns.Whether and how these early patterns of methylation might alter SLC6A4 gene expression and influence future stress reactivity and risk of disease remains an open question.Understanding the interplay between environmental and individual-level stressors can have important implications for future policy and intervention strategies to ultimately improve infants' development, particularly among disadvantaged communities where multiple stressors are likely to accumulate.

Fig. 1
Fig. 1 Distribution map of pollution and methylation data.Graphical localization and distribution of pollution and methylation data for the present sample.

Table 1 .
Sample characteristics.Milano, Brescia, Pavia.Socio-demographic (i.e., age, educational level, occupational status) and neonatal (i.e., gestational age, birth weight, head circumference, length, Apgar scores, breastfeeding at birth, and mode of delivery) data were obtained from medical records.Maternal PRS was retrospectively self-reported between 12 and 48 h from delivery.Measures of prenatal exposure to PM 2.5 were derived based on prenatal address locations of the mother-infant dyads.Infants' SLC6A4 DNAm levels were assessed between 6 and 24 h after birth from buccal cells, as detailed below.The study was approved by the Ethics Committees of the IRCCS Mondino Foundation and the participating hospitals.All the procedures were performed in accordance with the 2018 Declaration of Helsinki for studies conducted with human participants.All parents provided written informed consent to participate to the study.

Table 2 )
. Principal Component 1 (PC1; composed of 6 CpG sites) and Principal Component 2 (PC2; composed of 3 CpG sites) accounted, respectively, for 31.6% and 12.8% of the total variance in SLC6A4 DNAm and were used in further analyses.Results for the regression models for Principal Component 3 (PC3; composed of 2 CpG sites, 8,8% of the variance) and Principal Component 4 (PC4; composed of 2 CpG sites, 8,3% of the variance) are reported in supplementary results (Supplementary File S3).
Loadings on the respective principal components (PC) are reported in bold.Loadings <0.300 are not reported.

Table 3 .
Hierarchical linear regression analyses predicting infants' SLC6A4 DNA methylation PC1 from antenatal exposure to maternal PRS and full gestation PM 2.5 (Model 1) as well as trimester-specific PM 2.5 levels (Model 2-4).