Department of Population Medicine, Harvard Medical School, Harvard Pilgrim Health Care Institute, Boston, MA, USA.
Centre de Recherche du Centre Hospitalier Universitaire de Sherbrooke (CHUS), Sherbrooke, QC, Canada.
Int J Obes (Lond). 2021 Apr;45(4):860-869. doi: 10.1038/s41366-021-00750-4. Epub 2021 Jan 27.
Fetal exposure to maternal excess adiposity and hyperglycemia is risk factors for childhood adverse metabolic outcomes. Using data from a prospective pre-birth cohort, we aimed to further understand the prenatal determinants of fetal metabolic programming based on analyses of maternal adiposity and glycemic traits across pregnancy with childhood metabolomic profiles.
This study included 330 mother-child pairs from the Gen3G cohort with information on maternal adiposity and glycemic markers at 5-16 (visit 1) and 24-30 (visit 2) weeks of pregnancy. At mid-childhood (4.8-7.2 years old), we collected fasting plasma and measured 1116 metabolites using an untargeted approach. We constructed networks of interconnected metabolites using a weighted-correlation network analysis algorithm. We estimated Spearman's partial correlation coefficients of maternal adiposity and glycemic traits across pregnancy with metabolite networks and individual metabolites, adjusting for maternal age, gravidity, race/ethnicity, history of smoking, and child's sex and age at blood collection for metabolite measurement.
We identified a network of 16 metabolites, primarily glycero-3-phosphoethanolamines (GPE) at mid-childhood that showed consistent negative correlations with maternal body mass index, waist circumference, and body-fat percentage at visits 1 and 2 (ρ = -0.14 to -0.21) and post-challenge glucose levels at visit 2 (ρ = -0.10 to -0.13), while positive correlations with Matsuda index (ρ = 0.13). Within this identified network, 1-palmitoyl-2-decosahexaenoyl-GPE and 1-stearoyl-2-decosahexaenoyl-GPE appeared to be driving the associations. In addition, a network of 89 metabolites, primarily phosphatidylcholines, plasmalogens, sphingomyelins, and ceramides showed consistent negative correlations with insulin at visit 1 and post-challenge glucose at visit 2, while positive correlation with adiponectin at visit 2.
Prenatal exposure to maternal higher adiposity and hyperglycemic traits and lower insulin sensitivity markers were associated with a unique metabolomic pattern characterized by low serum phospho- and sphingolipids in mid-childhood.
胎儿暴露于母体肥胖和高血糖是儿童不良代谢结局的危险因素。本研究利用前瞻性出生队列的数据,旨在通过分析妊娠期间母体肥胖和血糖特征与儿童代谢组学特征,进一步了解胎儿代谢编程的产前决定因素。
本研究纳入了 Gen3G 队列中的 330 对母婴,在妊娠 5-16 周(第 1 次就诊)和 24-30 周(第 2 次就诊)时记录了母亲的肥胖和血糖标志物。在儿童中期(4.8-7.2 岁),我们采集了空腹血浆样本,使用非靶向方法测量了 1116 种代谢物。我们使用加权相关网络分析算法构建了相互关联的代谢物网络。我们估计了妊娠期间母体肥胖和血糖特征与代谢物网络和单个代谢物之间的斯皮尔曼偏相关系数,调整了母亲的年龄、生育次数、种族/民族、吸烟史以及儿童采集血样时的性别和年龄。
我们鉴定出一个由 16 种代谢物组成的网络,主要是甘油-3-磷酸乙醇胺(GPE),该网络在儿童中期与母体体重指数、腰围和体脂百分比在第 1 次和第 2 次就诊时呈负相关(ρ=-0.14 至-0.21),与第 2 次就诊时的餐后血糖呈负相关(ρ=-0.10 至-0.13),与 Matsuda 指数呈正相关(ρ=0.13)。在该鉴定出的网络中,1-棕榈酰基-2-二十二碳六烯酰基-GPE 和 1-硬脂酰基-2-二十二碳六烯酰基-GPE 似乎是导致这些关联的主要原因。此外,一个由 89 种代谢物组成的网络,主要是磷脂、溶血磷脂、神经鞘脂和神经酰胺,与第 1 次就诊时的胰岛素和第 2 次就诊时的餐后血糖呈负相关,而与第 2 次就诊时的脂联素呈正相关。
胎儿在妊娠期间暴露于母体较高的肥胖和高血糖特征以及较低的胰岛素敏感性标志物,与儿童中期血清磷酸和神经鞘脂水平降低的独特代谢模式有关。