Evidence for the Placenta-Brain Axis: Multi-Omic Kernel Aggregation Predicts Intellectual and Social Impairment in Children Born Extremely Preterm

Abstract

BackgroundChildren born extremely preterm are at heightened risk for intellectual and social impairment. There is increasing evidence for a key role of the placenta in prenatal neurodevelopmental programming, suggesting that the placenta serves as a role in the origins of neurodevelopmental outcomes.MethodsWe examined associations between genomic and epigenomic profiles in the placenta and assessed their ability to predict intellectual and social impairment at age 10 years in 379 children from the Extremely Low Gestational Age Newborn (ELGAN) cohort. Assessment of intellectual ability (IQ) and social function was completed with the Differential Ability Scales-II (DAS-II) and Social Responsiveness Scale (SRS), respectively. Genome-wide mRNA, CpG methylation and miRNA were assessed with the Illumina Hiseq 2500, HTG EdgeSeq miRNA Whole Transcriptome Assay, and Illumina EPIC/850K array, respectively. We conducted genome-wide differential mRNA/miRNA and epigenome-wide placenta analyses. These molecular features were then integrated for a predictive analysis of IQ and SRS outcomes using kernel aggregation regression.ResultsWe found that genes with important roles in placenta angiogenesis and neural function were associated with intellectual and social impairment. Multi-omic predictions of intellectual and social function were strong, explaining approximately 10% and 12% of the variance in SRS and IQ scores via cross-validation, respectively.ConclusionsOur findings demonstrate that aggregating information from biomarkers within and between molecular data types improves prediction of complex traits like social and intellectual ability in children born extremely preterm, suggesting that traits influenced by the placenta-brain axis may be omnigenic.

Genetic control of fetal placental genomics contributes to development of health and disease

“…Dysfunction of transcription factor regulation in the placenta has also shown profound effects on childhood traits [19][20][21][22] . Although combining genetics, transcriptomics, and epigenomics lends insight into the influence of placental genomics on complex traits 23 , genome-wide screens for GTAs that integrate different molecular profiles and generate functional hypotheses require more sophisticated computational methods.…”

“…We imputed expression into individual-level ELGAN genotypes ( = 729) (Online Methods). Controlling for race, sex, gestational duration, inflammation of the chorion, and maternal age, as described before and in Online Methods 23 , we tested for associations for 6 representative traits measured at birth or at 24 months: neonatal chronic lung disease, head circumference Z-score, fetal growth restriction, birth weight Z-score, necrotizing enterocolitis, and Bayley II Mental Development Index (MDI) at 24 months . Shown in Figure 4B and Supplemental Table S9, at FDR-adjusted < 0.05, we detected negative associations between SEC11A GReX and birthweight Z-score (effect size: -0.248, 95% adjusted 57 and ERP29 (-3.332, [-4.987,-1.677]).…”

“…Genotypes and multi-omic (mRNA, miRNA, and CpG methylation) data were collected from umbilical cord blood and the fetal side of the placenta of subjects enrolled in the ELGAN 33 Study, as described in previous work 23,40,78 and in detail in Supplemental Methods. Genotype data was assayed on Illumina Million Quad and HumanOmniExpress-12 v1.0 arrays 79 .…”

“…For binary traits, we used a logistic regression with the same predictors and covariates. These covariates have been previously used in placental genomic studies of neonatal traits 23,78,96 because of their strong correlations with the outcomes and with placental transcriptomics and methylomics.…”