Maternal obesity during pregnancy is associated with neurodevelopmental disorder (NDD) risk. We utilized integrative multi-omics to examine maternal obesity effects on offspring neurodevelopment in rhesus macaques by comparison to lean controls and two interventions. Differentially methylated regions (DMRs) from longitudinal maternal blood-derived cell-free fetal DNA (cffDNA) significantly overlapped with DMRs from infant brain. The DMRs were enriched for neurodevelopmental functions, methylation-sensitive developmental transcription factor motifs, and human NDD DMRs identified from brain and placenta. Brain and cffDNA methylation levels from a large region overlapping mir-663 correlated with maternal obesity, metabolic and immune markers, and infant behavior. A DUX4 hippocampal co-methylation network correlated with maternal obesity, infant behavior, infant hippocampal lipidomic and metabolomic profiles, and maternal blood measurements of DUX4 cffDNA methylation, cytokines, and metabolites. Ultimately, maternal obesity altered infant brain and behavior, and these differences were detectable in pregnancy through integrative analyses of cffDNA methylation with immune and metabolic factors. 

All animals were housed at the California National Primate Research Center (CNPRC). Adult pregnant female rhesus macaques (Macaca mulatta) with male fetuses were selected for this study. All dams ranged in age from 7 to 12 years and were selected for lean and obese groups based on their Body Condition Score (BCS). Obese females had a BCS of at least 3.5 (range 1-5) which correlates with 32% body fat, and lean animals had a BCS of 2-2.5. Animals had maintained a consistent BCS for at least one year prior to selection for the study and pre-study physicals confirmed that none of the selected females were diabetic. All animals were maintained with standard indoor housing conditions at CNPRC and fed nine “biscuits” of commercial chow (High Protein Primate Diet Jumbo; LabDiet; 5047) twice daily while pregnant, received biweekly fresh produce, daily forage mixture, and ad libitum water. The caloric restriction group had the amount of chow restricted to prevent weight gain during pregnancy and the Pravastatin group was given 1 mg/kg of body weight. However, all dams, regardless of group, were provided twelve biscuits twice daily during nursing of 4 months or older infants. Dams were relocated to a single housing room around gestational day 70.  Approximately two weeks later they were paired with a compatible cage mate during daytime hours and were separated prior to feeding times. Dams were allowed to deliver naturally (~165 day gestation length) and mother-infant pairs were raised indoors until offspring were 6 months old. However, five pregnant dams required Cesarian deliveries for post-date pregnancies (~175 gestation days) or as recommended by veterinarians for health reasons, two in the obese group, one in the pravastatin group and two in the lean control group. In those scenarios, infants were successfully reared to 6 months of age by foster dams using established CNPRC protocols. Mother-infant dyads were housed indoors with another compatible mother-infant dyad during daytime hours if possible; however, this was not possible for 1 lean control dyad, 3 obese dyads, and 1 pravastatin dyad. Infant brain samples (hippocampus, hypothalamus, and prefrontal cortex) were collected on postnatal day 180 after infants were anesthetized with ketamine and euthanized with 120 mg/kg pentobarbital. Upon collection, samples were immediately frozen and stored at -80 °C.

Fasting blood was collected in lavender top (EDTA) tubes from mothers once during the 1st and 2nd trimesters and twice during the 3rd trimester after anesthetization with 5-30 mg/kg ketamine or 5-8 mg/kg telazol. Plasma samples were filtered using Amicon Ultra Centrifugal Filters (3k molecular weight cutoff, Millipore) to remove proteins and lipids. Metabolites were extracted from each of the brain tissues as previously described (*1). To 207 µL of either plasma filtrate or brain tissue extract, 23 µL of internal standard containing DSS-d6 was added and samples were placed in 3 mm Bruker nuclear magnetic resonance (NMR) tubes. Proton NMR spectra were acquired on each sample at 25 °C using the noesypr1d pulse sequence on a Bruker Avance 600 MHz NMR spectrometer (Bruker, Billerica, MA, USA). The concentrations of individual metabolites are determined based on the addition of an internal standard, and analyzed using Chenomx NMRSuite (version 8.1, Chenomx Inc) to annotated the metabolites based on Reference Library 10.

(*1) Hasegawa, Y. et al. Optimization of a Method for the Simultaneous Extraction of Polar and Non-Polar  Oxylipin Metabolites, DNA, RNA, Small RNA, and Protein from a Single Small Tissue Sample. Methods Protoc. 3, (2020).