Associations of Maternal Exposure and Characteristics with Placental Transcription Regulation and Consequences on Fetal Growth

The placenta is a key organ in pregnancy and its dysfunction can lead to pregnancy complications as well as adverse maternal and fetal outcomes. While evidence for adverse consequences of maternal characteristics (such as obesity) or environmental exposures (e.g. particulate matter under 2.5 μm in diameter, PM2.5) on the course and outcomes of pregnancy has been described, specific mechanisms are largely unknown. The transcriptome plays important physiological roles in growth, development, and metabolism, and may help elucidate these mechanistic links. This dissertation project investigated non-coding elements of the placental transcriptome and transcriptomic regulation via epigenetic mechanisms (long non- coding RNAs, lncRNAs and DNA methylation, DNAme) in relation to maternal pre-pregnancy obesity and prenatal PM2.5 exposure. Objectives: This dissertation project had three aims: 1) investigate associations of maternal PM2.5 exposure with placental lncRNA expression and subsequent effects on birthweight, 2) investigate associations of pre-pregnancy BMI on placental lncRNA expression and subsequent effects on birthweight, and 3) investigate associations of pre-pregnancy obesity with differential methylation of genome-wide placental CpG sites and differential transcription of placental mRNA and lncRNA, as well as explore pre-transcriptional regulation of RNAs by DNAme in the placenta. Methods: Aims 1 and 2 were addressed using data and placental samples from the CANDLE (N = 776) and GAPPS (N = 205) cohorts within the ECHO PATHWAYS Consortium. CANDLE, the Conditions Affecting Neurocognitive Development in Early Childhood cohort, is a pregnancy cohort based in Shelby County, TN. GAPPS, the Global Alliance to Prevent Prematurity and Stillbirth, is a pregnancy biorepository of participants recruited in Seattle and Yakima, WA. Aim 3 utilized data and placental samples from the Fetal Growth Studies (FGS) cohort (N = 301), a pregnancy study on women recruited from twelve medical centers across the United States. Maternal prenatal exposure to PM2.5 was estimated via residential address history using spatio-temporal air pollution models developed at the University of Washington, while maternal pre-pregnancy BMI (ppBMI) was calculated from self-reported height and pre- pregnancy weight at initial study visits. Placental transcriptome-wide RNA and epigenome-wide DNAme were assayed using Illumina HiSeq sequencers and Illumina Infinium Human Methylation 450 Beadchip, respectively. To address these study aims, linear regression models were used to examine associations of pre-pregnancy PM2.5 exposure and maternal obesity with placental lncRNA (CANDLE, GAPPS, FGS), DNAme (FGS), and gene expression (FGS). Contextually appropriate covariate and precision variables determined a priori were included in the models, and effect modification by infant-sex was examined using both sex-stratified analyses and additional inclusion of exposure–infant-sex interaction terms. All analyses used a false discovery rate calculation to control for multiple testing. Results: In CANDLE, first-trimester PM2.5 was positively associated expression of the lncRNA LINC00702 (logFC=0.103, FDR p-value=0.027) and second-trimester PM2.5 was negatively associated expression the lncRNA AC105345.1 (logFC=-0.163, FDR p-value=0.028). In CANDLE, interaction of PM2.5 with infant-sex was observed for the lncRNA,LINC00339 (interaction FDR<0.001). Additionally, in CANDLE, 47 transcripts were significantly associated with second-trimester PM2.5 among female infants. Further two transcripts, LINC00467 and AC023157.3, were expressed at higher and lower levels, respectively, among male infants in GAPPS. In CANDLE, ppBMI was associated with increased expression of three lncRNAs (ERVH48-1, AC139099.1, CEBPA-DT) among males, and pre-pregnancy obesity (ppOB) was associated with decreased expression of one lncRNA, ZNF225-AS1, among females. In GAPPS, ppBMI was associated with decreased expression of two lncRNA transcripts (AP000879.1 and AL365203.2) among male infants. In FGS, methylation at five CpG sites, cg11844079[PIGM], cg22591875[MEDAG], cg27278787[DLGAP1], and cg04069951[CD81], was associated with ppOB. Additionally, ppOB was associated with decreased methylation of cg09858237[WDR16; STX8], cg15606914[H1FOO], and cg00311799[CTSF] among male infants. We also observed significant interaction of ppOB and infant-sex on methylation of cg08956510[NME1]. Finally, ppOB was associated with increased expression of the ACE2 gene (logFC=0.922, FDR=0.033) and decreased expression of the lncRNA RP4-55D20.2 (logFC=-1.699, FDR<0.001) among female infants. Conclusions: Overall, in the current study, we found potential evidence supporting associations of maternal PM2.5 exposures during pregnancy and pre-pregnancy BMI with alterations in the placental transcriptome and epigenome. Further, evidence indicated that some of these associations, including lncRNAs and DNAme related to cell proliferation, growth, and transcription control, were infant-sex dependent, suggesting mechanisms that may also influence the known differences in placental morphology and development between males and females. These results need to be replicated in other populations with similar characteristics and exposure distributions to further explore their implications. In addition, functional studies are needed to characterize downstream effects of reported transcriptome and epigenome findings. Our gained knowledge can add mechanistic insight that facilitates efforts to promote optimal health for mother and child over the course of pregnancy.