Individual longitudinal changes in DNA-methylome identify signatures of early-life adversity and correlate with later outcome

Abstract

Adverse early-life experiences (ELA) affect a majority of the world's children. Whereas the enduring impact of ELA on cognitive and emotional health is established, there are no tools to predict vulnerability to ELA consequences in an individual child. Epigenetic markers including peripheral-cell DNA-methylation profiles may encode ELA and provide predictive outcome markers, yet the interindividual variance of the human genome and rapid changes in DNA methylation in childhood pose significant challenges. Hoping to mitigate these challenges we examined the relation of several ELA dimensions to DNA methylation changes and outcome using a within-subject longitudinal design and a high methylation-change threshold. DNA methylation was analyzed in buccal swab/saliva samples collected twice (neonatally and at 12 months) in 110 infants. We identified CpGs differentially methylated across time for each child and determined whether they associated with ELA indicators and executive function at age 5. We assessed sex differences and derived a sex-dependent 'impact score' based on sites that most contributed to methylation changes. Changes in methylation between two samples of an individual child reflected age-related trends and correlated with executive function years later. Among tested ELA dimensions and life factors including income to needs ratios, maternal sensitivity, body mass index and infant sex, unpredictability of parental and household signals was the strongest predictor of executive function. In girls, high early-life unpredictability interacted with methylation changes to presage executive function. Thus, longitudinal, within-subject changes in methylation profiles may provide a signature of ELA and a potential predictive marker of individual outcome.

Keywords: Adverse childhood experiences; Biomarkers; DNA methylation; Early-life stress; Epigenetics; Executive control; Methylomics; Precision medicine; Stress; Within-subject design.

Fig. 1

Methylation is influenced by age. A) Timeline of sample collection and assessments in 110 infants. B) Heatmap depicting distinct patterns of methylation distinguishing DNA methylation profiles from newborn and 1-year old children. C) Average percentage methylation at selected sites increases with age. D) Using PCA, the first principal component, explaining 25% of the variance, accounts for the age of sample collection. **p < 0.001, bars represent mean, lines represent individual sites.

Fig. 2

localization and gene ontology of the sites differentially methylated (DMS) between neonatal and one year old samples. (A). Chromosomal distribution of the DMS demonstrates that the reside on all autosomes. Numbers on the left denote the percentage of the overall DMS that localize to each chromosome. (B) Alignment of the DMS with genes and their structures: 13983 of the 14037 DMS localized to within 1000 kb of transcription a start site (TSS), and these 13983 DMS associated with 2764 unique genes. (C) Gene ontology identified developmental processes as the key theme of genes associated with DMS between 10-day old and one year old samples of the same child. n = 110 infants.

Fig. 3

Methylation-changes of individual infants between the ages of 10 days and one year predict effortful control at 5 years. A) The first component of the principal component analysis (PCA) of methylation changes in the ∼14,000 differentially methylated sites reflect the average change in methylation (n = 110). B) Average methylation changes from newborn to one year of age of an individual child predicts effortful control performance at five years of age (n = 90). C) Average percent methylation in newborns does not predict outcome. D) Similarly, average percent methylation at one year of age does not predict outcome. Note that analogous results were observed when using all 1.74 million methylated sites, as shown in the Supplemental Fig.S-3. Points represent individual samples, circles = females, triangles = males. Line represents linear regression. R represents Pearson correlation coefficient.

Fig. 4

Unpredictability, assessed using the QUIC, portends functional outcomes at 5 years. A) Income/needs ratio (INR) has a weak association with effortful control in our sample of individuals. B) Correlation of maternal depressive symptoms with effort control suggests a weak negative association in these individuals. C) Measures of maternal sensitivity do not correlate with effortful control at 54 months of age in this sample. D) Unpredictability measured using the Questionnaire of Unpredictability in Childhood (QUIC) is inversely correlated with effortful control score at 54 months. Points represent individual samples, circles = females, triangles = males. Line represents linear regression (n = 90). R represents Pearson correlation coefficient.

Fig. 5

Unpredictability portends child development and may interact with methylation changes over time. A) Unpredictability assessed using the QUIC predicts effortful control at five years of age to a similar degree in both males and females. B) The change in methylation over the first year of life also predicts effortful control at the same age to the same degree in both sexes. n: females = 41, males = 49. C) There is an interaction between unpredictability and change in methylation in females only: for females who experience high unpredictability, the change in methylation over the first year of life predicts effortful control (please see Table 2 for descriptions of significance). In contrast, there is no such interaction observed in males. n: females low = 13, females high = 28, males low = 16, males high = 33. Points represent individual samples, purple = females, green = males. Line represents linear regression. F = females, M = Male. R and p represent Pearson correlation for subgroup. (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.)

Fig. 6

Impact scores identify individuals vulnerable to poor outcomes following the unpredictability dimension of early-life adversity. A) Flow chart of the computation method used to select highly contributing sites. The clumping and thresholding method identifies sites that appear to have a significant interaction but do not correlate highly with other sites that have already been selected. B) Manhattan plot representing the differentially methylated sites (blue) and the distribution across the chromosomes and the corresponding significance score (-log10p) of each site interacting with QUIC to predict effortful control in females. Sites in red are those selected via the clumping and thresholding algorithm. Dotted line is at p = 0.05. C) The significant interaction of QUIC and risk score calculated from top sites according to our model predicts effortful control at 5 years in females who have experienced more adversity n: low = 28, high = 13. Line represents linear regression. R and p represent Pearson correlation for subgroup. (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.)