Abstract
Purpose:
Investigate associations of maternal social experiences with offspring epigenetic age acceleration (EAA) from birth through mid-childhood among 205 mother-offspring dyads of minoritized racial and ethnic groups.
Methods:
We used linear regression to examine associations of maternal experiences of racial bias or discrimination (0=none, 1–2=intermediate, or 3+=high), social support (tertile 1=low, 2=intermediate, 3=high), and socioeconomic status index (tertile 1=low, 2=intermediate, 3=high) during the prenatal period with offspring EAA according to Horvath’s Pan-Tissue, Horvath’s Skin & Blood, and Intrinsic EAA clocks at birth, 3 years, and 7 years.
Results:
In comparison to children of women who did not experience any racial bias or discrimination, those whose mothers reported highest levels of racial bias or discrimination had lower Pan-Tissue clock EAA in early (−0.50 years; 90% CI: −0.91, −0.09) and mid-childhood (−0.75 years; −1.41, −0.08). We observed similar associations for the Skin & Blood clock and Intrinsic EAA. Maternal experiences of discrimination were not associated with Pan-Tissue EAA at birth. Neither maternal social support nor socioeconomic status predicted offspring EAA.
Conclusions:
Children whose mothers experienced racial bias or discrimination exhibited slower EAA. Future studies are warranted to confirm these findings and establish associations of early-life EAA with long-term health outcomes.
Keywords: Epigenetic age acceleration, experiences of discrimination, social exposures, prenatal period, maternal-offspring dyads, cohort study
INTRODUCTION
The aging epidemic has forged interest in patterns and determinants of aging 1–3. Recently, social determinants of health have risen to the forefront as important drivers of age-related non-communicable conditions 4, with evidence in humans 5,6, non-human primates 7–9 and rodents 8,10 revealing an indisputable link between unfavorable social experiences and chronic disease risk. The effect of social experiences on health is hypothesized to operate, in part, through epigenetic mechanisms 11–13. Of relevance to aging research are genome-wide changes in DNA methylation that occur as individuals get older 14–17. Epigenetic clocks are functional biomarkers of developmental and aging processes 18–20 that capitalize on age-related changes to DNA methylation 21,22. One biomarker of recent interest is epigenetic age acceleration (EAA) – i.e., the estimated difference between epigenetic age and chronological age – as it outperforms other molecular markers of aging, including telomere length 23 and is a strong predictor of morbidity and mortality 24–28.
Studies in adults have found that adverse social experiences predict higher EAA 29–32 with stronger effects observed for experiences occurring earlier vs. later in life 30,33. Our group recently showed that EAA starts before birth 34, pointing toward the importance of considering how prenatal exposures shape the aging process. To date, only a handful of studies have examined the relationship between maternal social or psychosocial experiences and offspring epigenetic aging. The findings are inconsistent, with some studies indicating faster offspring epigenetic aging following unfavorable maternal experiences, and others suggesting the opposite. In a combined analysis of two cohorts in the Netherlands and Singapore, McGill et al. reported positive associations of maternal prenatal anxiety with higher EAA among male offspring through age 4 years 35. Similarly, in the U.K.-based ALSPAC cohort, maternal adverse childhood experiences (ACEs) predicted higher offspring EAA at birth, again primarily in males 36. On the other hand, in the Finnish PREDO Study, Suarez et al. found that maternal history of pre-pregnancy depression and higher antenatal depressive symptoms were each associated with lower epigenetic gestational age at birth in both sexes 37. In the California Salinas Valley CHAMACOS study, Nwanaji-Enweren et al. reported both positive and negative associations of maternal ACEs with offspring epigenetic age acceleration at ages 7, 9, and 14 years, with similar patterns by sex, and also noted that maternal ACEs predicted longer offspring telomere length, indicative of slower biological aging 38. While the inconsistent findings may reflect differences in specific maternal exposures of interest, method of assessing epigenetic age, and study population characteristics, they also highlight a need for additional research in this area.
In this exploratory pilot study, we investigated associations of maternal social experiences with offspring EAA from birth through mid-childhood among 205 mother-offspring pairs of minoritized racial and ethnic groups. Specifically, we sought to examine associations of maternal experiences of racial bias or discrimination, social support, and indicators of socioeconomic disadvantage during the prenatal period with three measures of EAA at birth and across early and mid-childhood: (1) Horvath’s Pan-Tissue clock, (2) Horvath’s Skin & Blood clock, and the (3) Intrinsic EAA (IEAA) clock. We focus only on mother-offspring pairs from minoritized groups considering the increasingly recognized importance of experiences of racial bias or discrimination on a range of maternal/child outcomes 39,40 and to reduce the possibility of exposure misclassification among white women who experience bias or discrimination for non-race-related reasons. We hypothesized that maternal experiences of racial bias or discrimination, lower social support, and socioeconomic disadvantage will be associated with persistent differences in offspring EAA across early-life.
MATERIALS AND METHODS
Study population
This analysis used data from Project Viva, a pre-birth cohort of mother-child pairs recruited 1999–2002 from multi-specialty clinics in Massachusetts (Atrius Harvard Vanguard Medical Associates 41. Of the 2,128 enrolled mother-child pairs, 708 had high-quality data on whole-genome DNA methylation in cord blood (n=485) and blood collected from offspring in early childhood (age 3–6 years, n=120) or mid-childhood (age 6–10 years, n=460). Given our interest in maternal experiences of racial bias or discrimination, we restricted the analysis to mother-offspring pairs of minoritized populations based on the mothers’ self-reported race and ethnicity, which we view as a social construct as opposed to a form of biological determinism 40. At enrollment, we asked mothers, “Which of the following best describes your race or ethnicity?” Mothers had a choice of one or more of the following mutually exclusive racial/ethnic groups: Hispanic or Latina, white or Caucasian, Black or African American, Asian or Pacific Islander, American Indian or Alaskan Native, and Other (please specify). For the participants who selected ‘Other,’ we compared the specified responses to the US census definition for the other five race and ethnicities and reclassified them where appropriate. If a participant chose more than one racial or ethnic group, we classified them as “>1 racial or ethnic group.” For this analysis, we excluded white mothers and used four racial/ethnic categories: “Black”, “Hispanic,” “Asian” and “More than 1 race or other.” The analytic sample comprised 205 mother-offspring dyads and with DNA methylation data in cord (n=126), early childhood (n=25; range 3–6 years), and mid-childhood blood (n=149; range 6–11 years). Fifteen children contributed DNA methylation data at all three time-points.
Exposure: maternal social experiences
Experiences of discrimination (EOD)
At the early pregnancy visit (~10 gestational weeks), women reported experiences of racial bias or discrimination via an adapted version 39 of the validated Experiences of Discrimination (EOD) measure 42. The questionnaire comprises eight questions inquiring on lifetime experiences of any racism, discrimination, or bias. Participants responded ‘Yes’ or ‘No’ to the prompt ‘I have experienced unfair or bad treatment because of my race or ethnicity’ for each of eight different situational domains: at school, getting hired or getting a job, at work, getting housing, getting medical care, getting service in a store or restaurant, on the street or in a public setting, and from the police or in the courts. We summed ‘Yes’ responses to produce an EOD score for self-reported experiences of racial discrimination (range 0–8), with the total number of domains conceptualized as an indicator of overall exposure 42,43. We categorized EOD responses as none (EOD=0), moderate (EOD=1–2), and high (EOD=3+) as previously done 39,42.
Social support
We assessed social support during the prenatal period using the Turner Support Scale 44, which comprises 10 questions that prompts the respondent to rank, on a scale from 1 (most support – i.e., strongly agree with statement about positive support) to 4 (least support – i.e., strongly disagree with statement about positive support), the amount of support and affection that they received from their partner (five questions) and family/friends (five questions). Women without a partner at the early pregnancy visit skipped the first five questions, resulting in minimum possible scores of 10 and 5 for partnered and unpartnered women, respectively; and maximum possible scores of 40 and 20 for partnered and unpartnered women, respectively. In the analysis we categorized women into tertiles, using separate cut-points for partnered and unpartnered women. We reversed ranked the social support score such that the lowest tertile corresponded with receipt of the lowest social support (least favorable) and the top tertile corresponded with the highest social support (most favorable).
Prenatal SES index
As an indicator of socioeconomic disadvantage, we used an SES index that was associated with differential epigenome-wide DNA methylation from birth through mid-childhood in this cohort 45. Components of the index include maternal educational attainment, marital status, annual household income, and receipt of public assistance assessed via questionnaires at the early pregnancy visit. For the analysis, we assessed the index as an ordinal variable with the lowest tertile representing the lowest SES and the top tertile representing the highest SES.
Outcomes: epigenetic clocks and estimation of EAA
We extracted DNA using the PureGene kits (Qiagen, Valencia, CA) from cord blood leukocytes and nucleated red blood cells collected at delivery, and from peripheral leukocytes collected during early and mid-childhood. DNA extracted from buffy coat was bisulfite treated (EZ DNA Methylation-Gold Kit, Zymo Research, CA, USA). DNA methylation values were derived using the Infinium HumanMethylation450 BeadChip (Illumina, CA, USA) array 46, following a standard DNA methylation quantification protocol 47 that we published 45. DNA methylation QA/QC procedures are described in the Supplemental Material.
Using DNA methylation data from each life stage (birth, early childhood, mid-childhood), we derived two epigenetic clocks developed for use in multi-ethnic populations, each with utility for estimating biological age across multiple life stages and tissue types: Horvath’s Pan-Tissue clock 48,49 and the Skin & Blood clock 50. Both clocks were developed using training data from cord blood and multiple tissue types including peripheral leukocytes collected from pediatric populations (<18 years), thereby facilitating comparability in estimates of epigenetic age across the multiple life stages of interest in this study. Details on derivation, QA/QC, and performance of these clocks in Project Viva are published 34 and in the Supplemental Material.
For the analysis, the outcome of interest is epigenetic age acceleration (EAA) – a continuous, normally-distributed variable estimated as the residuals obtained from regressing chronological age on epigenetic clock age. EAA can be negative, indicative of slower biological aging/development relative to the sample average; positive, indicative of accelerated biological aging/development, relative to the sample average; or approximately 0, indicating that a child’s biological age is equivalent to the sample average. We calculated EAA based on the Pan-Tissue and Skin & Blood clocks, and additionally considered a third assessment of age-related immunological changes calculated from the Pan-Tissue clock, called Intrinsic Epigenetic Age Acceleration (IEAA) 51. We chose to focus on EAA rather than epigenetic age because the former is independent of chronological age and has been linked to age-related outcomes later in life 24–28.
Covariates
Research staff collected information on maternal prenatal smoking and parity via questionnaires at the early pregnancy visit. We calculated gestational age at delivery using the date of the mother’s last menstrual cycle prior to the index pregnancy or second trimester ultrasound where available, or if the estimates differed by >10 days. We calculated maternal pre-pregnancy body mass index (BMI; kg/m2) using self-reported pre-pregnancy weight collected at enrollment and clinically-measured height. Prenatal medical records provided information on mode of delivery and offspring sex.
Data analysis
First, we assessed the relationship between the three maternal social exposures with offspring EEA, separately, during each life stage using linear regression. We included maternal prenatal smoking as confounders per expert recommendations for studies examining perinatal determinants of epigenetic endpoints 52 and offspring sex given known sex differences in DNA methylation.
Next, we considered EAA across early and mid-childlood as repeated outcomes using mixed-effects models with random intercepts and an unstructured corrrelation matrix. This approach is complementary to the above analysis, but more efficiently leverages longitudinal data by including all participants to contribute to the standard error of the model, so long as they have at least outcome measurement. In multivariable anlayses, we adjusted for maternal prenatal smoking and child sex (Model 1), followed by additional adjustment for EAA at birth to account for variation in biological age at birth (Model 2). Nota bene, we adjusted for EAA at birth as a covariate rather than include it as part of the repeated measures outcomes as EAA assessed in cord leukocytes may not be comparable to EAA estimated from peripheral leukocytes during childhood.
In sensitivity analysis, we adjusted for cell type composition at the time of sample collection to ensure that associations are not completely driven by variation in cell type composition, acknowledging that it could be a mediator 53,54.
For all models, we used α=0.10 as the threshold for statistical significance. We selected a relaxed α due to the relatively small sample and correlated nature of the three maternal social exposures and three offspring EAA outcomes at three life stages. However, when interpreting results, we focus on consistency in the direction and magnitude of associations for the relationships of interest within and across life stages, rather than statistical significance. Analyses were conducted in R, version 4.3.0 55 and SAS 9.4 (SAS Institute Inc., NC, USA).
RESULTS
Descriptive statistics
The average age of mothers at enrollment was 29.8±6.3 years. Forty-five percent of the women identified as Black, 24.4% as Hispanic, 15.1% as Asian, and 15.6% identified as multi-racial or ‘other.’ Most women (76.1%) were married/cohabiting. Forty-four percent graduated college and 35.4% had an annual household income >$70,000. Forty-six percent of offspring were female. Table 1 shows additional characteristics of the sample and the distribution of the maternal social exposures and offspring EAA.
Table 1.
Background characteristics of 205 Project Viva mother-offspring pairs of minoritized racial/ethnic groups overall, and among subgroups with DNA methylation (DNAm) data at birth (n=126), and during early (n=25) and mid-childhood (n=149).
| Overall | Cord blood DNAm data | Early childhood DNAm data | Mid-childhood DNAm data | |
|---|---|---|---|---|
| n = 205 | n = 126 | n = 25 | n = 149 | |
|
| ||||
| Sociodemographic and background characteristics | ||||
|
| ||||
| Mother’s age at enrollment (years) | 29.8 ± 6.3 | 29.9 ± 6.1 | 30.9 ± 6.2 | 29.7 ± 6.4 |
| Nulliparous at index pregnancy | 43.4% (89) | 45.2% (57) | 36.0% (9) | 43.6% (65) |
| Pre-pregnancy BMI | 25.7 ± 5.9 | 25.4 ± 5.8 | 27.4 ± 6.8 | 25.7 ± 7.0 |
| Prenatal smoking habits | ||||
| Never smoked | 73.7% (151) | 73.8% (93) | 68.0% (17) | 73.8% (110) |
| Former smoker | 9.8% (20) | 11.1% (14) | 8.0% (2) | 10.1% (15) |
| Smoked during pregnancy | 16.6% (34) | 15.1% (19) | 24.0% (6) | 16.1% (24) |
| Mother’s race and ethnicity | ||||
| Black | 44.9% (92) | 38.9% (49) | 32.0% (8) | 49.0% (73) |
| Hispanic | 24.4% (50) | 26.2% (33) | 24.0% (6) | 22.8% (34) |
| Asian | 15.1% (31) | 16.7% (21) | 28.0% (7) | 13.4% (20) |
| >1 race/ethnicity or other | 15.6% (32) | 18.3% (23) | 16.0% (4) | 14.8% (22) |
| Married or cohabitating | 76.1% (156) | 77.8% (98) | 76.0% (19) | 73.2% (109) |
| Education level | ||||
| No college | 20.5% (42) | 17.5% (22) | 24.0% (6) | 21.5% (32) |
| Some college | 34.1% (70 ) | 34.1% (43) | 28.0% (7) | 35.6% (53) |
| College degree | 24.4% (52) | 25.4% (32) | 20.0% (5) | 25.5% (38) |
| Graduate degree | 20.0% (41) | 23.0% (29) | 28.0% (7) | 17.5% (26) |
| Annual household income | ||||
| ≤$20,000 | 12.7% (23) | 7.9% (8) | 4.8% (1) | 15.3% (20) |
| >$20,000 - $40,000 | 30.4% (55) | 26.6% (30) | 23.8% (5) | 32.8% (43) |
| >$40,000 - $70,000 | 21.5% (39) | 27.4% (31) | 28.6% (6) | 19.1% (25) |
| >$70,000 | 35.4% (64) | 38.9% (44) | 42.9% (9) | 32.8% (43) |
| Receipt of public assistance | 19.8% (36) | 13.3% (15) | 17.4% (4) | 22.3% (29) |
| Gestational age at delivery (weeks) | 39.4 ± 1.7 | 39.4 ± 1.8 | 38.8 ± 1.6 | 39.4 ± 1.7 |
| Child is female sex | 45.9% (94) | 45.2% (57) | 52.0% (13) | 47.7% (71) |
|
| ||||
| Social exposures | ||||
|
| ||||
| Experiences of discrimination (EOD) | ||||
| Mean ± SD | 2.1 ± 2.0 | 2.1 ± 2.0 | 1.7 ± 2.0 | 2.2 ± 2.1 |
| Median | 2.00 | 2.00 | 1.00 | 2.0 |
| Min, Max | 0, 8 | 0, 8 | 0, 7 | 0, 8 |
| Social support score | ||||
| Women with a partner | ||||
| Mean ± SD | 14.1 ± 4.0 | 13.9 ± 3.9 | 14.5 ± 3.6 | 14.4 ± 4.0 |
| Median | 13.0 | 13.0 | 13.5 | 13.0 |
| Min, Max | 9, 28 | 10, 28 | 10, 23 | 9, 27 |
| Women without a partner | ||||
| Mean ± SD | 6.8 ± 2.1 | 6.4 ± 2.2 | 5.0 a | 7.0 ± 2.3 |
| Median | 6.0 | 5.0 | 5.0 a | 6.0 |
| Min, Max | 5, 10 | 5, 10 | 5.0 a | 5, 10 |
| Prenatal SES index | ||||
| Mean ± SD | −0.56 ± 0.94 | −0.58 ± 0.91 | −0.79 ± 0.85 | −0.60 ± 0.93 |
| Median | −0.89 | −0.87 | −1.09 | −0.99 |
| Min, Max | −1.65, 1.53 | −1.65, 1.49 | −1.65, 1.13 | −1.65, 1.53 |
Abbreviations: BMI - body mass index; DNAm - DNA methylation; SES - socioeconomic status
n = 1, therefore no estimate of SD or range.
Bivariate associations
Experiences of discrimination (EOD) and social support scores were inversely correlated (Pearson’s R2=−0.24; P=0.002), such that more EOD correlated with lower social support. Prenatal SES was mildly positively correlated with social support (Pearson’s R2=0.13; P=0.10). EOD and the prenatal SES index were uncorrelated (Pearson’s R2=−0.03; P=0.72).
Table 2 shows correlations among chronological and epigenetic age based on the Pan-Tissue and Skin & Blood clocks. All correlation coefficients were positive and increased in magnitude from low at birth (Pearson’s R2 0.08 to 0.19) to moderate at early (0.28 to 0.61) and mid-childhood (0.49 to 0.59).
Table 2.
Pearson’s correlation coefficients (R2) among chronological age, Horvath's Pan Tissue Clock, and Skin & Blood Clock at birth and during early (median age 3.1 y) and mid-childhood (median age 7.8 y) among 205 Project Viva mother-offspring pairs of minoritized racial/ethnic groups.
| Chronological age | Pan-Tissue | Skin & Blood | |
|---|---|---|---|
|
| |||
| Birth (n = 126) | |||
|
| |||
| Chronological age | 1.00 | 0.08 | 0.14 |
| Pan Tissue | 1.00 | 0.19 | |
| Skin & Blood | 1.00 | ||
|
| |||
| Early childhood (n = 25) | |||
|
| |||
| Chronological age | 1.00 | 0.49 | 0.28 |
| Pan Tissue | 1.00 | 0.61 | |
| Skin & Blood | 1.00 | ||
|
| |||
| Mid-childhood (n=149) | |||
|
| |||
| Chronological age | 1.00 | 0.49 | 0.59 |
| Pan Tissue | 1.00 | 0.58 | |
| Skin & Blood | 1.00 | ||
Multivariable analysis
Supplemental Table 1 shows associations of maternal social experiences with offspring EAA, separately by life stage. We found an inverse association of maternal EOD with offspring EAA during early and mid-childhood, but no associations at birth. In comparison to children whose mothers reported no EOD, those born to women with EOD scores of 1–2 (−0.71 years, 90% CI: −1.18, −0.24) or 3+ (−0.50 years, 90% CI: −0.91, −0.09) had lower EAA in early childhood according to the Pan-Tissue clock. We noted similar inverse associations of EOD with EAA according to the Skin & Blood clock and IEAA during early childhood, and according to all three clocks during mid-childhood. On the other hand, we did not observe consistent associations of social support or the prenatal SES index with EAA according to any of the clocks. We observed similar associations when assessing EAA longitudinally across early and mid-childhood. In comparison to children whose mothers reported no EOD, those whose mothers had highest EOD exhibited lower EAA across early and mid-childhood according to the Pan-Tissue clock (−0.69, 90% CI: −1.23, −0.13), Skin & Blood clock (−0.49, 90% CI: −0.91, −0.07), and IEAA (−0.58, 90% CI: −1.15, −0.01) (Table 3, Model 1). Adjustment for cord blood EAA did not change associations for the Pan-Tissue clock or IEAA (Table 3, Model 2).
Table 3.
Associations (β [90% CI]) of maternal experiences of discrimination, social support, and prenatal socioeconomic (SES) status index with offspring epigenetic age acceleration as repeated measures across early (median age 3.1 y) and mid-childhood (median age 7.8 y) among Project Viva mother-offspring pairs of minoritized racial/ethnic groups.
| Pan-Tissue | Skin & Blood | IEAA | |
|---|---|---|---|
|
| |||
| Model 1 | |||
|
| |||
| Experience of discrimination | n=141 | n=141 | n=141 |
| 0 (lowest) | 0.00 (Reference) | 0.00 (Reference) | 0.00 (Reference) |
| 1-2 | −0.40 (−0.94, 0.16) | −0.21 (−0.63, 0.21) | −0.31 (−0.88, 0.25) |
| 3+ (highest) | −0.69 (−1.23, −0.13) | −0.49 (−0.91, −0.07) | −0.58 (−1.15, −0.01) |
| P-trend | 0.05 | 0.06 | 0.10 |
| Social support | n=138 | n=138 | n=138 |
| Tertile 1 (lowest) | −0.48 (1.00, 0.04) | −0.18 (−0.57, 0.22) | −0.32 (−0.83, 0.19) |
| Tertile 2 | −0.02 (−0.57, 0.53) | −0.20 (−0.62, 0.22) | 0.12 (−0.42, 0.66) |
| Tertile 3 (highest) | 0.00 (Reference) | 0.00 (Reference) | 0.00 (Reference) |
| P-trend | 0.15 | 0.46 | 0.32 |
| Prenatal SES index | n=118 | n=118 | n=118 |
| Tertile 1 (lowest) | −0.35 (−0.93, 0.25) | 0.08 (−0.35, 0.50) | −0.38 (−0.95, 0.20) |
| Tertile 2 | 0.00 (Reference) | 0.00 (Reference) | 0.00 (Reference) |
| Tertile 3 (highest) | −0.39 (−1.01, 0.25) | 0.05 (−0.39, 0.50) | −0.41 (−1.01, 0.19) |
| P-trend | 0.96 | 0.93 | 0.98 |
|
| |||
| Model 2 | |||
|
| |||
| Experiences of discrimination | n=72 | n=72 | n=72 |
| 0 (lowest) | 0.00 (Reference) | 0.00 (Reference) | 0.00 (Reference) |
| 1-2 | −0.62 (−1.32, 0.08) | −0.47 (−1.02, 0.09) | −0.61 (−1.38, 0.16) |
| 3+ (highest) | −1.06 (−1.75, −0.39) | −0.47 (−1.01, 0.08) | −0.90 (−1.66, −0.16) |
| P-trend | 0.02 | 0.20 | 0.07 |
| Social support | n=69 | n=69 | n=69 |
| Tertile 1 (lowest) | −0.38 (−0.99, 0.24) | 0.20 (−0.32, 0.72) | −0.18 (−0.82, 0.46) |
| Tertile 2 | 0.39 (−0.31, 1.08) | −0.03 (−0.63, 0.56) | 0.58 (−0.16, 1.30) |
| Tertile 3 (highest) | 0.00 (Reference) | 0.00 (Reference) | 0.00 (Reference) |
| P-trend | 0.36 | 0.55 | 0.67 |
| Prenatal SES index | n=61 | n=61 | n=61 |
| Tertile 1 (lowest) | 0.17 (−0.56, 0.90) | −0.03 (−0.59, 0.53) | 0.17 (−0.58, 0.92) |
| Tertile 2 | 0.00 (Reference) | 0.00 (Reference) | 0.00 (Reference) |
| Tertile 3 (highest) | 0.20 (−0.57, 1.00) | 0.68 (0.09, 1.26) | 0.28 (−0.52, 1.11) |
| P-trend | 0.97 | 0.10 | 0.86 |
Abbreviations: IEAA: Intrinsic epigenetic age acceleration. Bolded values indicate statistical significance at alpha = 0.10.
Model 1: Estimates are adjusted for prenatal smoking habits (never/former/current) and offspring sex.
Model 2: Model 1 + cord blood measure of epigenetic age acceleration.
Sensitivity analysis
Adjustment for cell type composition at the time of epigenetic age assessment did not change results (Supplemental Tables 2 and 3). Therefore, we focus on estimates from non-cell-type adjusted estimates from the main analysis to avoid overadjustment 56.
DISCUSSION
In this pilot study of 205 mother-offspring pairs of minoritized racial and ethnic groups, maternal experiences of discrimination corresponded with lower EAA in offspring across early (age 3 years) and mid-childhood (age 7 years). While these findings contrast with an adult literature indicating that unfavorable social experiences correlate with higher EAA 33,57–61, there could be differential implications of EAA before vs. after reproductive maturity. Indeed, our findings align with a recent study of Latino families in which Clausing et al. found that maternal experience of discrimination was related to lower epigenetic age measured in saliva of children 62. Additionally, the broader construct of maternal psychosocial stress has been linked to lower offspring EAA at birth in the PREDO cohort 37. The inverse association observed in this, and other studies may reflect life history trade-offs in which costs incurred during early-life accumulate and negatively affect future growth and somatic maintenance 63–66. Still, several studies report findings in the opposite direction. For example, unfavorable prenatal experiences like pregnancy complications 67, prenatal anxiety 35, and maternal ACEs 38 are associated with higher offspring EAA during childhood and adolescence. Such inconsistencies emphasize a need for additional research in this area.
Of note, we did not find associations between any maternal exposures and EAA at birth. Beyond the possibility of no true association, this finding may reflect the fact that we used epigenetic clocks developed to assess epigenetic age across multiple life stages. Such clocks may be less sensitive to differences in gestational age than those developed to specifically assess epigenetic age in newborns 68–70. Additionally, neither maternal social support nor the prenatal SES index predicted offspring EAA. The null results could be because this study took place in a mid- to upper-income population in eastern Massachusetts, resulting in a relatively narrow range of SES that may have limited our ability to detect associations. Moreover, we excluded white participants given a central focus on the experiences of racial discrimination which likely reduced variation in social experiences.
Strengths of our study include prospectively collected data in multi-ethnic mother-offspring dyads; assessment of the women’s lifetime experiences of racial bias or discrimination; and assessment of EAA using multiple clocks and across multiple sensitive periods of development. Our study also has limitations. First, while we used epigenetic clocks developed from populations of diverse ancestries 48,50, the extent to which these clocks are appropriate for assessing epigenetic age at birth 70,71 and across early life is unknown 72. Second, we were underpowered to assess sex-specific associations. Third, our small sample sizes likely limited statistical power to detect smaller effect sizes. Finally, we caution against extrapolating the results beyond the scope of the racial/ethnic groups in this analysis, which comprised mother-offspring pairs of minoritized populations.
Conclusions
In this study of mother-offspring pairs from minoritized racial and ethnic groups and of mid-to-high socioeconomic status, maternal experiences of racial bias or discrimination are associated with slower biological aging in offspring across early and mid-childhood. This finding is consistent with the theory that social stressors contribute to an intergenerational “embedding” of health disparities by slowing development during early-life 73. Future studies are warranted to confirm our findings and assess the relationship between early-life EAA across multiple developmental stages with clinically-relevant outcomes or well-recognized biomarkers of aging. Additionally, considering evolutionary processes, including the adaptive significance of tempo of development, will elucidate explanations for why adverse social experiences compromise health over the lifecourse 74.
Supplementary Material
Acknowledgements and funding:
This work is funded by R01HD034568 and R24ES030894. ZML is supported by the National Science Foundation (NSF) DBI 2010607. AB and AC were supported by the United States National Institutes of Health Grants R01ES031259, R01MD016595 and R25GM143298. WP is supported by U01 DK134981 and ADA-7–22-ICTSPM-08. Funding sources did not play any role in study design; in the collection, analysis, and interpretation of data; in the writing of the report; or in the decision to submit the article for publication.
Abbreviations:
- EAA
Epigenetic age acceleration
- EOD
experiences of discrimination
- IEAA
intrinsic epigenetic age acceleration
- SES
socioeconomic status
Footnotes
Conflicts of interest: The authors have no conflicts to disclose.
Declaration of interests
The authors declare the following financial interests/personal relationships which may be considered as potential competing interests:
Wei Perng, Marie-France Hivert, Andres Cardenas, and Anne Bozack report financial support was provided by National Institutes of Health. Zachary Laubach reports financial support was provided by National Science Foundation. Wei Perng reports financial support was provided by American Diabetes Association.
Publisher's Disclaimer: This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition of a cover page and metadata, and formatting for readability, but it is not yet the definitive version of record. This version will undergo additional copyediting, typesetting and review before it is published in its final form, but we are providing this version to give early visibility of the article. Please note that, during the production process, errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
References
- 1.Ferrucci L, Giallauria F & Guralnik JM Epidemiology of Aging. Radiologic Clinics of North America vol. 46 643–652 Preprint at 10.1016/j.rcl.2008.07.005 (2008). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.United Nation. World Population Prospects 2019 - Volume II: Demographic Profiles. (UN, 2020). doi: 10.18356/7707d011-en. [DOI] [Google Scholar]
- 3.World Health Organization. Decade of healthy ageing baseline report. (2020). [Google Scholar]
- 4.Noren Hooten N, Pacheco NL, Smith JT & Evans MK The accelerated aging phenotype: The role of race and social determinants of health on aging. Ageing Res Rev 73, (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Shonkoff JP, Boyce WT & McEwen BS Neuroscience, molecular biology, and the childhood roots of health disparities: building a new framework for health promotion and disease prevention. Journal of American Medical Association 301, 2252–2259 (2009). [DOI] [PubMed] [Google Scholar]
- 6.Clark AM, DesMeules M, Luo W, Duncan AS & Wielgosz A Socioeconomic status and cardiovascular disease: Risks and implications for care. Nat Rev Cardiol 6, 712–722 (2009). [DOI] [PubMed] [Google Scholar]
- 7.Sanchez MM The impact of early adverse care on HPA axis development: Nonhuman primate models. Horm Behav 50, 623–631 (2006). [DOI] [PubMed] [Google Scholar]
- 8.Sánchez MM, Ladd CO & Plotsky PM Early adverse experience as a developmental risk factor for later psychopathology: Evidence from rodent and primate models. Dev Psychopathol 13, 419–449 (2001). [DOI] [PubMed] [Google Scholar]
- 9.French JA & Carp SB Early-life social adversity and developmental processes in nonhuman primates. Curr Opin Behav Sci 7, 40–46 (2016). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Nishi M Effects of Early-Life Stress on the Brain and Behaviors: Implications of Early Maternal Separation in Rodents. Int J Mol Sci 21, (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Wadhwa PD, Buss C, Entringer S & Swanson JM Developmental Origins of Health and Disease: Brief History of the Approach and Current Focus on Epigenetic Mechanisms. Seminar in Reproductive Medicine 27, 358–368 (2009). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Szyf M The early-life social environment and DNA methylation. Clin Genet 81, 341–349 (2012). [DOI] [PubMed] [Google Scholar]
- 13.Champagne FA Early adversity and developmental Outcomes: Interaction between Genetics, Epigenetics, and social experiences across the Life Span. Perspectives on Psychological Science 5, 564–574 (2010). [DOI] [PubMed] [Google Scholar]
- 14.Fraga MF et al. Epigenetic differences arise during the lifetime of monozygotic twins. Proc Natl Acad Sci U S A 102, 10604–10609 (2005). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Teschendorff AE, West J & Beck S Age-associated epigenetic drift: Implications, and a case of epigenetic thrift? Hum Mol Genet 22, 7–15 (2013). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Boks MP et al. The relationship of DNA methylation with age, gender and genotype in twins and healthy controls. PLoS One 4, 1–8 (2009). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Bjornsson TH et al. Intra-individual Change Over Time in DNA Methylation With Familial Clustering. Journal of American Medical Association 299, 2877–2883 (2008). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Slieker RC et al. Age-related accrual of methylomic variability is linked to fundamental ageing mechanisms. Genome Biol 17, 1–13 (2016). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Fraga MF & Esteller M Epigenetics and aging: the targets and the marks. Trends in Genetics 23, 413–418 (2007). [DOI] [PubMed] [Google Scholar]
- 20.West J, Widschwendter M & Teschendorff AE Distinctive topology of age-associated epigenetic drift in the human interactome. Proceedings of the National Academy of Sciences 110, 14138–14143 (2013). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Bell CG et al. DNA methylation aging clocks: Challenges and recommendations. Genome Biol 20, (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Horvath S & Raj K DNA methylation-based biomarkers and the epigenetic clock theory of ageing. Nat Rev Genet 19, 371–384 (2018). [DOI] [PubMed] [Google Scholar]
- 23.Jylhävä J, Pedersen NL & Hägg S Biological Age Predictors. EBioMedicine 21, 29–36 (2017). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Dhingra R, Nwanaji-Enwerem JC, Samet M & Ward-Caviness CK DNA Methylation Age—Environmental Influences, Health Impacts, and Its Role in Environmental Epidemiology. Current environmental health reports vol. 5 317–327 Preprint at 10.1007/s40572-018-0203-2 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Zheng Y et al. Blood Epigenetic Age may Predict Cancer Incidence and Mortality. EBioMedicine 5, 68–73 (2016). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Marioni RE et al. DNA methylation age of blood predicts all-cause mortality in later life. Genome Biol 16, (2015). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Chen BH et al. DNA methylation-based measures of biological age: meta-analysis predicting time to death. Aging 8, 1844–1859 (2016). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Fransquet PD, Wrigglesworth J, Woods RL, Ernst ME & Ryan J The epigenetic clock as a predictor of disease and mortality risk: A systematic review and meta-analysis. Clin Epigenetics 11, (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Gettler LT et al. Epigenetic aging in children from a small-scale farming society in The Congo Basin: Associations with child growth and family conflict. Dev Psychobiol 62, 138–153 (2020). [DOI] [PubMed] [Google Scholar]
- 30.Austin MK et al. Early-life socioeconomic disadvantage, not current, predicts accelerated epigenetic aging of monocytes. Psychoneuroendocrinology 97, 131–134 (2018). [DOI] [PubMed] [Google Scholar]
- 31.Joshi D, Gonzalez A, Lin D & Raina P The association between adverse childhood experiences and epigenetic age acceleration in the Canadian longitudinal study on aging (CLSA). Aging Cell 22, (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Jovanovic T et al. Exposure to Violence Accelerates Epigenetic Aging in Children. Sci Rep 7, (2017). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Fiorito G et al. Social adversity and epigenetic aging: A multi-cohort study on socioeconomic differences in peripheral blood DNA methylation. Sci Rep 7, (2017). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Bozack AK et al. DNA methylation age at birth and childhood: performance of epigenetic clocks and characteristics associated with epigenetic age acceleration in the Project Viva cohort. Clin Epigenetics 15, 62 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.McGill MG et al. Maternal Prenatal Anxiety and the Fetal Origins of Epigenetic Aging. Biol Psychiatry 91, 303–312 (2022). [DOI] [PubMed] [Google Scholar]
- 36.Dye CK et al. Mother’s childhood adversity is associated with accelerated epigenetic aging in pregnancy and in male newborns. bioRxiv (2023) doi: 10.1101/2023.03.02.530806. [DOI] [Google Scholar]
- 37.Suarez A et al. The Epigenetic Clock at Birth: Associations With Maternal Antenatal Depression and Child Psychiatric Problems. J Am Acad Child Adolesc Psychiatry 57, 321–328.e2 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38.Nwanaji-Enwerem JC et al. Maternal adverse childhood experiences before pregnancy are associated with epigenetic aging changes in their children. Aging 13, (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39.Powell CA et al. Maternal experiences of racial discrimination and offspring sleep in the first 2 years of life: Project Viva cohort, Massachusetts, USA (1999–2002). Sleep Health 6, 463–468 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 40.Flanagin A, Frey T & Christiansen SL Updated guidance on the reporting of race and ethnicity in medical and science journals. The Journal of the American Medical Association 326, 621–627 (2021). [DOI] [PubMed] [Google Scholar]
- 41.Oken E et al. Cohort profile: Project viva. Int J Epidemiol 44, 37–48 (2015). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 42.Krieger N, Smith K, Naishadham D, Hartman C & Barbeau EM Experiences of discrimination: Validity and reliability of a self-report measure for population health research on racism and health. Soc Sci Med 61, 1576–1596 (2005). [DOI] [PubMed] [Google Scholar]
- 43.Krieger N Discrimination and Health. in Social Epidemiology (eds. Berkman LF., Kawachi I. & Glymour MM.) 63–125 (Oxford University Press, 2014). doi: 10.1093/med/9780195377903.001.0001. [DOI] [Google Scholar]
- 44.Turner RJ, Grindstaff CF & Phillips N Social Support and Outcome in Teenage Pregnancy. J Health Soc Behav 31, 43–57 (1990). [PubMed] [Google Scholar]
- 45.Laubach ZM et al. Socioeconomic status and DNA methylation from birth through mid-childhood: A prospective study in Project Viva. Epigenomics 11, 1413–1427 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 46.Sandoval J et al. Validation of a DNA methylation microarray for 450,000 CpG sites in the human genome. Epigenetics 6, 692–702 (2011). [DOI] [PubMed] [Google Scholar]
- 47.Wu S et al. Exposure to low levels of lead in utero and umbilical cord blood DNA methylation in project viva: An epigenome-wide association study. Environ Health Perspect 125, (2017). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 48.Horvath S DNA methylation age of human tissues and cell types. Genome Biol 14, 115 (2013). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 49.Fang F et al. Evaluation of pediatric epigenetic clocks across multiple tissues. Clin Epigenetics 15, 142 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 50.Horvath S et al. Epigenetic clock for skin and blood cells applied to Hutchinson Gilford Progeria Syndrome and ex vivo studies. Aging 10, (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 51.Horvath S et al. An epigenetic clock analysis of race/ethnicity, sex, and coronary heart disease. Genome Biol 17, (2016). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 52.Breton CV et al. Small-magnitude effect sizes in epigenetic end points are important in children’s environmental health studies: The children’s environmental health and disease prevention research center’s epigenetics working group. Environ Health Perspect 125, 511–526 (2017). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 53.Schisterman EF, Cole SR & Platf RW Overadjustment bias and unnecessary adjustment in epidemiologic studies. Epidemiology 20, 488–495 (2009). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 54.Laubach ZM, Murray EJ, Hoke KL, Safran RJ & Perng W A biologist’s guide to model selection and causal inference. Proc Biol Sci 288, (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 55.R Core Team. R: A language and environment for statistical computing. Preprint at https://www.r-project.org/. (2023).
- 56.Aronoff JE et al. Associations between perceived discrimination and immune cell composition in the Jackson Heart Study. Brain Behav Immun 103, 28–36 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 57.Liu Z et al. The role of epigenetic aging in education and racial/ethnic mortality disparities among older U.S. Women. Psychoneuroendocrinology 104, 18–24 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 58.Lawrence KG et al. Association of Neighborhood Deprivation with Epigenetic Aging Using 4 Clock Metrics. JAMA Netw Open 3, (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 59.Schmitz LL et al. The Socioeconomic Gradient in Epigenetic Ageing Clocks: Evidence from the Multi-Ethnic Study of Atherosclerosis and the Health and Retirement Study. Epigenetics 17, 589–611 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 60.Zhao W et al. Education and lifestyle factors are associated with dna methylation clocks in older African Americans. Int J Environ Res Public Health 16, (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 61.Oblak L, van der Zaag J, Higgins-Chen AT, Levine ME & Boks MP A systematic review of biological, social and environmental factors associated with epigenetic clock acceleration. Ageing Res Rev 69, (2021). [DOI] [PubMed] [Google Scholar]
- 62.Clausing ES, Binder AM & Non AL Epigenetic age associates with psychosocial stress and resilience in children of Latinx immigrants. Epigenomics 13, 1677–1699 (2021). [DOI] [PubMed] [Google Scholar]
- 63.Stearns SC The evolution of life histories. (Oxford University Press, 1992). [Google Scholar]
- 64.Kirkwood TBL Understanding the odd science of aging. Cell 120, 437–447 (2005). [DOI] [PubMed] [Google Scholar]
- 65.Kirkwood TBL. Evolution of ageing. Nature 270, (1977). [DOI] [PubMed] [Google Scholar]
- 66.Stearns SC Trade-Offs in Life-History Evolution. Ecology vol. 3 (1989). [Google Scholar]
- 67.Girchenko P et al. Associations between maternal risk factors of adverse pregnancy and birth outcomes and the offspring epigenetic clock of gestational age at birth. Clin Epigenetics 9, (2017). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 68.Wang J & Zhou WH Epigenetic clocks in the pediatric population: when and why they tick? Chin Med J (Engl) 134, 2901–2910 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 69.Knight AK et al. An epigenetic clock for gestational age at birth based on blood methylation data. Genome Biol 17, (2016). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 70.Bohlin J et al. Prediction of gestational age based on genome-wide differentially methylated regions. Genome Biol 17, (2016). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 71.Knight AK et al. An epigenetic clock for gestational age at birth based on blood methylation data. Genome Biol 17, (2016). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 72.Wu X et al. DNA methylation profile is a quantitative measure of biological aging in children. Aging 11, (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 73.Palma-Gudiel H, Fañanás L, Horvath S & Zannas AS Psychosocial stress and epigenetic aging. in International Review of Neurobiology vol. 150 107–128 (Academic Press Inc., 2020). [DOI] [PubMed] [Google Scholar]
- 74.Nesse RM & Stearns SC The great opportunity: Evolutionary applications to medicine and public health. Evol Appl 1, 28–48 (2008). [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.