Epidemiology of Maternal-Fetal Biomarkers, Resilience, Adversity, and Child Epigenetics (EMBRACE)

Using the PrOMIS cohort, the team is investigating how the timing and severity of maternal stress and trauma across the lifespan affect child health outcomes using innovative structured life-course modeling approaches. Their initial findings indicate that cumulative maternal physical abuse (i.e., occurring during childhood, pre-pregnancy, and pregnancy) is associated with higher internalizing and externalizing behavior problems in children, with the strongest effects observed in girls. In addition, they identified sex-specific sensitive periods: the pre-pregnancy for boys and pregnancy for girls. The associations were stronger with increased severity of maternal trauma, underscoring how both the timing and severity of maternal trauma contribute to intergenerational transmission of risk.

The team is also investigating the effects of trauma from one generation to the next through the deoxyribonucleic acid (DNA) methylation function. Using saliva samples from 7-year-old children, they assessed whether maternal abuse timing and severity predict children’s DNA methylation patterns. Exposure to maternal abuse was associated with differences in childhood DNA methylation at 11 CpG sites across the genome. Maternal sexual abuse experienced in adulthood, specifically one year before pregnancy or during pregnancy, was the most relevant exposure associated with alterations in childhood DNA methylation. These findings reinforce evidence for intergenerational transmission of maternal history of trauma through epigenetic mechanisms and highlight the importance of addressing maternal trauma during the perinatal period to help prevent developmental problems in children.

One pregnancy outcome that the team extensively studied is preterm birth (PTB). Increasing evidence shows that PTB is influenced by a complex interplay of psychosocial, neighborhood, environmental, medical, biological, and genetic factors. Cortisol is among the most widely studied biomarkers of stress. Previously, cortisol was primarily measured in saliva, blood, and urine. However, cortisol measured in these tissues primarily captures acute cortisol synthesis and release, and the collection of samples, particularly of blood, is invasive. To address these limitations, Gelaye integrated hair cortisol and cortisone concentration assessments in pregnancy cohorts. In terms of duration and stability over time, these biomarkers provide an integrated measure of glucocorticoids that surpasses measurements in other tissues such as urine, saliva, or plasma. The measures more accurately reflect sustained hypothalamic–pituitary–adrenal (HPA) axis activity. The HPA axis is a primary signaling pathway activated in response to stress and traumatic life events.

Although prior studies have examined the relationship between maternal cortisol concentrations and perinatal outcomes such as PTB, results have been inconsistent. The team investigated the association between corticosteroid levels in pre-pregnancy and first trimester and PTB risk using marginal structural models and a causal inference framework. Their findings indicate that HPA axis dysregulation, measured using hair cortisol and cortisone concentrations, is significantly associated with preterm birth risk. They are now extending this work by analyzing glucocorticoids from the second and third trimesters in another cohort. Elucidating the biological and molecular pathways underlying PTB may ultimately lead to novel prevention and treatment strategies.

In September 2025, Gelaye initiated a new project, the Metabolomics of Perinatal Sleep and Adverse Birth Outcomes Study (Principal Investigator Gelaye). Using existing biospecimens from NICHD cohorts, the study aims to characterize metabolites and metabolic clusters associated with sleep disturbances in women during the antepartum and postpartum periods, using targeted and untargeted ultra-performance LC-MS/GC-MS methods.