Non-invasive epigenetic protein signatures associated with childhood obesity

Childhood obesity is a major predictor of lifelong chronic disease, yet the biological mechanisms linking early-life exposures to adiposity remain incompletely understood. Inflammatory and metabolic proteins have been implicated in obesity-related pathways, but their clinical utility in children is limited by biological variability and the invasiveness of sample collection. DNA methylation (DNAm)-derived protein proxies, or EpiScores, offer a stable, non-invasive alternative tool for capturing systemic physiological processes and may provide novel insights into the early biological embedding of obesity risk.

We examined associations between 105 DNAm-derived protein EpiScores and body mass index (BMI) z-scores in a socioeconomically diverse cohort of 31 school-aged children. EpiScores were generated using the MethylDetectR platform, and linear regression models, adjusted for sex, age, race, saliva cell-type proportions, and socioeconomic status (SES; maternal education and annual household income), were used to evaluate associations. Parallel models excluding SES were compared to quantify confounding. False discovery rate (FDR) correction was applied. Functional enrichment and protein-protein interaction (PPI) analyses were performed using STRINGdB.

Fifteen EpiScores demonstrated nominal significance with BMI z-score (p ≤ 0.05) and were further evaluated using a discovery-level FDR threshold (FDR q ≤ 0.20). VEGFA, MMP-12, MMP-1, and CDL5 showed strong positive associations, while PAPP-A, Resistin, and TGF-α were inversely related to BMI z-score. Adjustment for socioeconomic status (SES) modestly altered several effect estimates, most notably for MMP-1 and CCL17, indicating partial social confounding. Functional enrichment revealed overrepresentation of cytokine-mediated signaling, extracellular matrix organization and angiogenesis pathways, highlighting coordinated immune, metabolic and vascular remodeling processes.

Salivary DNAm-derived protein EpiScores capture biologically coherent and socially patterned molecular signatures of adiposity in children. These DNAm-based proxies provide a noninvasive means of detecting early alterations in metabolic regulation and may inform future longitudinal studies integrating biological and socioeconomic determinants of obesity risk.