Abstract P355: Mendelian Randomization Supports Causality Between Maternal Hyperglycemia and Fetal Metabolic Programming by Leptin Epigenetic Modulation
Introduction: In utero exposure to maternal hyperglycemia is associated with increased risk of obesity later in life. Animal studies have suggested that dysregulation of leptin pathways contributes to fetal metabolic programming of obesity, but human studies are limited. Mendelian Randomization can be used to reduce confounding factors and eliminate risk of reverse causation when assessing associations between biomarkers and outcomes using genetic variants as instrumental variables.
Hypothesis: Using Mendelian Randomization, we tested the hypothesis that maternal glycemia was part of causal pathways modulating newborns’ leptin epigenetic regulation.
Methods: We conducted a prospective population-based cohort study of pregnant women and newborns. This study included up to 485 mother-child dyads. We measured maternal anthropometry and glycemia during pregnancy; we collected clinical data, neonatal anthropometric measurements, and cord blood samples at birth. We excluded women treated for gestational diabetes to assess the impact of maternal glycemia across the normal spectrum and avoid confounding by treatment. We built a genetic risk score (GRS10) based on 10 genetic variants known for association with fasting glucose in previous genome-wide association studies in pregnant and non-pregnant populations. We used GRS10 calculated with maternal genetic variants as an instrumental variable representing maternal glycemia. We assessed DNA methylation at 16 CpG sites near LEP (encoding for leptin) in cord blood cells. We measured leptin in cord blood by ELISA (Luminex).
Results: Women were 28±4 years old and had a mean body mass index of 25.4±5.6 kg/m2 in first trimester of pregnancy. Mean maternal fasting glucose was 4.2±0.3 mmol/L at second trimester of pregnancy. Newborns were 3.409±0.464 kg at birth and mean leptin levels were 14.7±13.2 ug/L in cord blood. The GRS10 was associated with maternal fasting glucose (β= 0.0457 per risk allele; SE= 0.007; P=7.76x10-11; N= 467) and was considered as an adequate instrumental variable (r2= 0.087 with maternal fasting glucose). DNA methylation at CpG site cg12083122 demonstrated the strongest association with leptin levels in cord blood (β= -0.17; SE= 0.07; P=0.01; N=170). The association between the instrumental variable GRS10 and DNA methylation at cg12083122 (β= -0.072; SE= 0.037; P=0.05; N=166) was in the expected direction and plausible in effect size when compared to expected observational association with measured maternal fasting glucose and DNA methylation at cg12083122 (βobs= -0.43; βiv = [-0.072 / 0.0457] = -1.58; 2-stage least-square β= -1.82; P=0.17).
Conclusions: Our data support that maternal glycemia is causing epigenetic adaptations at LEP gene locus in newborns, suggesting that the programming effect of maternal glycemia might have life long impact on weight regulation in offspring through leptin epigenomics.
Author Disclosures: M. Hivert: None. C. Allard: None. V. Desgagne: None. J. Patenaude: None. M. Lacroix: None. L. Guillemette: None. M. Battista: None. M. Doyon: None. J. Menard: None. J. Ardilouze: None. P. Perron: None. L. Bouchard: None.
- © 2015 by American Heart Association, Inc.