Abigail Gaylord

Prenatal exposure to ambient air pollutants has been associated with adverse child health across different populations. In the absence of direct air pollution exposure measures, epigenetic biomarkers can serve as indicators of cumulative exposure. Previously, we developed a methodological pipeline to build a prediction model to distinguish infants highly exposed to air pollutants in utero based on DNA methylation in cord blood samples. Here, we test and refine this pipeline using data from multiple pregnancy cohort sites included in the Environmental influences on Child Health Outcomes (ECHO) Cohort. Levels of prenatal exposure to fine particulate matter (PM2.5), nitrogen dioxide (NO2), and 8-hour maximum ozone (O3) were modeled using geocoded address data and machine learning. DNA methylation was measured using 450K or EPIC arrays. Training and validation sets were created and matched based on exposure level, cohort site, and array type to test the performance of the pipeline prediction model using area under the curve metrics. Among the three exposures, epigenetic-based predictions of PM2.5 and NO2 produced higher area under the curve values (0.64–0.68 and 0.62–0.95, respectively) than those for O3 (0.50–0.75). We identified differences in the performance of the pipeline model across geographical regions, exposure windows, and thresholds used to distinguish “high” and “low” air pollutant exposure. Overall, this pipeline enabled the estimation of air pollution from DNA methylation data, though careful consideration of model inputs is required. Further studies will test the performance of the CpG sites identified by the pipeline to predict child health outcomes.

Humid heat exposures over pregnancy impact birth outcomes but mechanisms are poorly described. We leveraged the Ghana Randomized Air Pollution and Health Study (GRAPHS) to examine associations between maximum shaded wet bulb globe temperature (WBGT) and cord blood extracellular vesicle (EV)-associated microRNA. We also considered heat index (HI) and dry air temperature as secondary exposures. Cord blood was collected at delivery and EVs isolated and RNAs sequenced and small RNA libraries constructed. We assigned participants’ trimester and whole pregnancy average heat exposures using satellite reanalysis heat exposures at the level of participant community of residence. We then performed linear regression to examine associations between prenatal heat exposures and log2-transformed miRNA counts and considered significance with an unadjusted P-value ≤ .01 and |beta| > 0.2. In 44 participants, we identified 91 microRNAs that were detected in at least 70% of samples. We identified that miR-146b-5p (β = 0.94 [95% CI = 0.27, 1.6], per 1-degree increase in WBGT); miR-381-3p (β = 2.23 [1.3, 3.15]), miR-378a-3p (β = −1.4, [−2.42, −0.39]), and miR-744-5p (β = 1.51 [0.46, 2.57]); and miR-126-3p (β = −0.75 [−1.2, −0.29]) were associated with first-, second-, and third-trimester WBGT, respectively. miR-1307-3p (β = 2.27 [0.61, 3.94] per 1-degree increase in WBGT) was associated with WBGT averaged over the whole pregnancy. Similar findings were observed with HI. In the GRAPHS cohort, we identified that humid-heat exposure over pregnancy alters the cord blood EV-associated miRNA profiles.

Household air pollution (HAP) has been associated with adverse pregnancy and birth outcomes, but the underlying mechanisms remain unclear. Pollutants can cross the placenta, potentially causing dysregulation of the crucial organ. Placental microRNAs (miRNAs) may serve as biomarkers of placental health, but studies of prenatal air pollutant exposure and placental miRNAs using non-targeted approaches have been scarce. We leveraged personal air monitoring data from mothers enrolled in the Ghana Randomized Air Pollution and Health Study to estimate prenatal exposure levels to carbon monoxide (CO) and particulate matter <2.5 µm (PM2.5). Placental tissue small RNA was sequenced and aligned to miRbase v22. We used sparse principal components analyses (sPCAs) to identify candidate placental miRNAs associated with prenatal CO (N = 133) and PM2.5 (N = 85) exposure. Associations between candidate miRNAs and prenatal exposures were assessed using linear regressions. We identified four placental miRNAs upregulated with prenatal CO exposure (miR-128–3p, miR-423–3p, miR-671–3p, and miR-744–5p) and five downregulated miRNAs (miR-29b-3p, miR-30e-5p, miR-101–3p, miR-130a-3p, and miR-376b-3p) at P < 0.1. Among female infants only, five miRNAs were downregulated (miR-101–3p, miR-130a-3p, miR-19b-3p, miR-106b-5p, and miR-301a-3p) and one upregulated (miR-22–3) with CO exposure (P < 0.1). We did not identify any associations between placental miRNAs and prenatal PM2.5 exposure. Our results support associations of prenatal CO exposure with differential expressions of critical placental miRNAs that have been implicated in placental disorders by previous studies. This is the first study to examine associations of these exposures with placental miRNAs using a non-candidate approach and lays the groundwork for targeted studies of placental miRNAs associated with prenatal HAP exposure.

Atopic dermatitis (AD) is a prevalent inflammatory skin disease with complex pathogenesis. Both skin and gut microbiota influence AD, with Staphylococcus aureus , in particular, exacerbating the disease. However, the relationship between S. aureus colonization in the gut and skin, and whether it affects AD, remains unclear. Using a combination of culture-based methods, microbiome analysis, and genome sequencing of S. aureus from multiple body sites of children with and without AD, we found that the gut represents a major reservoir of genetically diverse S. aureus that is transmitted to the skin, including mutants associated with worse disease. We validated this association between S. aureus gastrointestinal colonization and AD in an independent human cohort and demonstrated its direct effect on disease in an infantile AD mouse model, wherein S. aureus gastrointestinal colonization worsened skin inflammation. Overall, this study identifies a previously unrecognized S. aureus reservoir, with implications for microbiota-targeting therapies in AD.

To describe tDRs in human milk EVs and their associations with maternal body mass index, age, dietary indices, breastfeeding frequency, season and time of milk collection in a Latina population. We sequenced small RNAs from EVs from 109 mature human milk samples collected at 1 month after delivery in the Southern California Mother's Milk Study. We grouped tDRs using hierarchical clustering and clusters were compared across tDR characteristics. We analyzed associations of tDRs with intrinsic maternal variables (body mass index, age), maternal nutrition (caloric intake, Healthy Eating Index, Dietary Inflammatory Index), and variables related to feeding and milk collection (breastfeeding frequency, season and time of milk collection) using negative binomial models. We identified 338 tDRs expressed in 90% or more of milk EV samples, of which 113 were identified in all samples. tDR-1:26-Gly-CCC-1-M4 accounted for most reads (79%). Pathway analysis revealed a wide array of biological processes and disease mechanisms across the four tDR clusters. tDRs were associated with season of collection, time of collection, breastfeeding frequency, and the dietary inflammatory index. tDRs are abundant in milk EVs and may be sensitive to maternal diet, seasonality, time of day, and breastfeeding frequency.