EWAS in a polyphenol dense, DNA methylation-targeted, controlled diet and lifestyle study.
BACKGROUND: Dietary and lifestyle factors can influence DNA methylation patterns. We previously reported epigenetic age attenuation following a controlled study using an 8-week polyphenol-dense, DNA methylation-targeted diet and lifestyle intervention in healthy males (Methylation Diet and Lifestyle Study), with phytonutrient/polyphenol-rich foods (green tea, oolong tea, curcumin, garlic, and berries) being most predictive of this effect. METHODS: Here we conducted an epigenome-wide association study (EWAS) in 38 participants from the Methylation Diet and Lifestyle Study. The intervention included a dietary pattern intentionally rich in substrate and cofactor nutrients for methylation pathways, and components known to alter DNA-methyltransferase (DNMT) enzyme activity. In line with prior EWAS studies with small sample sizes where FDR-significant findings are unlikely, we used pre-specified nominal P-value thresholds (0.001, 0.0001) for the exploratory analyses. RESULTS: At P < 0.001 (unadjusted), 676 differentially methylated loci (DML) were identified in the intervention group versus 286 in controls. At P < 0.0001 (unadjusted), 50 DML were identified in the intervention group compared to 13 in controls. Fifteen DML were in transcription start site-proximal regions of genes including those involved in zinc homeostasis and nutrient sensing, development and pluripotency, proteostasis and genome stability, tumor suppression, and synaptic function. A group-by-time interaction analysis identified 70 intervention-specific DML at P < 0.0001, with nominal enrichment including autophagy, mTOR signaling, and chromatin remodeling pathways. A regional DMR analysis identified 128 within-group and 129 interaction-specific DMRs. DMR functional enrichment analyses revealed convergent nominal associations with lipid metabolism (alpha-linolenic acid, lipoic acid, biosynthesis of unsaturated fatty acids, PPAR signaling, cholesterol homeostasis), central energy metabolism (TCA cycle, glycolysis/gluconeogenesis, pentose phosphate, pyruvate), and nutrient sensing (PI3K-Akt, mTOR, AMPK, autophagy as well as other pathways). As expected for the limited cohort size and short intervention duration, none of the single CpG findings or enrichment analyses survived multiple test correction and are therefore considered exploratory and hypothesis-generating only. CONCLUSION: This EWAS identified a larger number of nominally changing CpGs in the intervention group compared to controls as well as biologically coherent methylation changes. These findings provide mechanistic hypotheses for previously observed epigenetic age attenuation. Replication in larger cohorts, longer intervention durations, and functional validation remain essential.