ABSTRACT: Epigenetics presents a dynamic approach to assess complex individual variation in obesity susceptibility. However, few studies have examined epigenetic patterns in preschool-age children, despite the relevance of this developmental stage to trajectories of weight gain, because of difficulties obtaining blood tissue samples. This proof of principle study examined DNA methylation in 92 saliva samples, comparing Latino preschool children of normal weight mothers (Body Mass Index [BMI] <27 kg/m2 and WC <90 cm) to children of obese mothers (BMI >30 kg/m2 and WC >100 cm). We hypothesized that salivary DNA methylation patterns in Latino preschool age children born of normal weight vs obese weight mothers would be: 1) associated with maternal BMI phenotype in continuous linear regression analysis; 2) saliva could demonstrate epigenetic variation across individuals; and 3) preschool child saliva would be differentially methylated when comparing those children with obese versus normal weight mothers. One hundred and nineteen CpG sites were significantly (p-value <1.56 X 10-5, p-value adjusted <.05) associated with maternal BMI in linear regression models controlling for childâs age, gender, and BMI. Of these 119 CpG sites, 41 were found within the transcription start site, 5â UTR, 3â UTR, or another regulatory region outside of the gene body. Saliva, a practical human tissue to obtain in naturalistic settings and in pediatric populations, was confirmed to be a viable medium for genome-wide epigenetic testing with maternal weight. Although not identical to results yielded from other human tissue types (i.e., cord blood samples), saliva findings indicate potential epigenetic differences in Latino preschool children at risk for pediatric obesity. This proof of principle study examined DNA methylation in 92 saliva samples, comparing Latino preschool children of normal weight mothers (Body Mass Index [BMI] <27 kg/m2 and WC <90 cm) to children of obese mothers (BMI >30 kg/m2 and WC >100 cm). Antropometry was measured objectively according to a standardized protocol.Saliva from preschool Latino children at risk for obesity (BMI>50% < 95% participating in WIC/SNAP programs) was collected using the Oragene DNA saliva kit following a strict data collection protocol. DNA extraction was performed as per DNA Genotek's recommendations using the PrepIT L2P reagent. Extracted DNA was stored in individually barcoded cryovials at â80 degrees Fahrenheit. For children, saliva was obtained using the âbaby brushâ approach, in which small sponges attached to plastic handles are inserted between cheek and gumline to absorb saliva .Arrays were processed using standard protocol [34], with 3 samples randomly selected to serve as duplicates and 1 sample run with HapMap DNA to test functionality of reagents. Duplicates were measured for high technique consistency with Pearson correlation coefficient (>.99). Methylation data were quality controlled using Illumina GenomeStudio (V2011.1), Methylation module (V1.9.0). Samples with lower than 98% call rate (i.e. <485,000 probes) were excluded. Any non-specific cross-reacting probes, probes carrying common SNPs (MAF >1%), or any probes with p-values greater than 0.05 for more than 20% of the sample were sequentially excluded. Validation via pyrosequencing was conducted.
