ESPE Abstracts

Background: Children born SGA are known to develop cardiometabolic conditions in adulthood¹. Nothing is known about the relationship of the transcriptome (gene expression) and epigenome (DNA methylation) to birth size and the future development of cardiometabolic disease.

Aim: To identify, I) differences and functional links between epigenome age-7years, transcriptome age-9years associated and birth size in a normal population; II) links between the transcriptome and epigenome in childhood and adult cardiometabolic risk.

Study Design: Normal children (n=6487) from the Avon Longitudinal Study of Parents and Children were assigned to groups based on birth size using bodyweight and gestation and divided into groups using population level 10^th and 90^th centiles. Adverse cardiometabolic risk at age-17years was defined by the National Heart Lung and Blood Institute criteria of prehypertension using systolic and diastolic blood pressure². Blood epigenome and transcriptome were available from 980 children age-7years and 947 children age-9years respectively. Hypernetworks were used to integrate differentially expressed genes (DEGs) and methylated points (CpGs), identifying functional links. Random Forest, a machine learning approach, was used to determine the predictive value of 'omic data presented as the area under the curve of the receiver operating characteristic (AUC).

Results: Unsupervised analysis identified significant differences between birthweight groups in the whole transcriptome and epigenome (p-value range, 6.6X10^-16<P<1.0x10^-2). Pre-hypertensive participants at age-17years were distinguished from normotensive participants and 'omic differences were defined (232DEGs; 2.0x10^-6<P<1.0x10^-2 & 830CpGs; 1.1x10^-8<P<1.0x10^-2). The pre-hypertensive group was enriched for children born small who caught up by age-7years (155/611 unhealthy/healthy compared to 1979/12746 in the other groups; 1.6-fold, P<1x10^-5). This group had a greater height velocity during their catch-up period than the normotensive participants (1.2-fold, P=0.027). Hypernetwork integration of 'omic data identified a functional relationship between DEGs at age-9years and CpGs at age-7years (55DEGs, 520CpGs). Identified CpGs grouped into 5 chromosomal regions (corrected-P<6.7x10^-5) including the genes SLC44A2 and PON1 with known associations to lipid metabolism. Genetic associations are known between PON1 and SGA². Random forest analysis was able to accurately predict the presence of pre-hypertension aged 17 from the early life 'omic data (epigenome age-7years AUC:0.982 and transcriptome age-9years AUC:0.973).

Conclusions: We have identified an integrated 'omic signature in childhood which defines a subset of individuals, born small who catch up by age-7years and are at increased risk of later cardiometabolic disease.

¹ Barker etal. (1988) BMJ 297(6641):134-135.

² Chobanian etal. (2003) JAMA 289(19):2560-2571.

³ Infante-Rivard (2010) Am.J.Epidemiol 171(9):999-1006.