👤 Jakob Linseisen

Changes in DNA methylation can increase or suppress the expression of health-relevant genes. We investigated for the first time the relationship between habitual food consumption and changes in DNA methylation. The German KORA FF4 and KORA Fit studies were used to study the change in methylation over a median follow-up of 4 years. Only subjects participating in both surveys and with available dietary and methylation data were included in the analysis (n = 465). DNA methylation was measured using the Infinium MethylationEPIC BeadChip (Illumina), resulting in 735,527 shared CpGs across both studies. Generalized estimating equation models with an interaction term of exposure and time point were used to analyze the association of 34 food groups, folic acid, and two dietary patterns with changes in DNA methylation over time. The results were corrected for genomic inflation. Significant interaction terms indicate different effects between both time points. We observed only a few significant associations between food intake and change in DNA methylation, except for cream and spirit consumption. The annotated genes include CLN3, PROM1, DLEU7, TLL2, and UGT1A10. We identified weak associations between food consumption and DNA methylation change. The differential results for cream and spirits, both consumed in low quantities, require replication in independent studies. Show less

The present study gives further evidence for the recently found association between variants of the fatty acid desaturase 1 fatty acid desaturase 2 (FADS1 FADS2) gene cluster and PUFA in blood phospholipids and explores this association for cellular fatty acids in erythrocyte membranes. In a subgroup of adults participating in the Bavarian Nutrition Survey II, a cross-sectional population-based study conducted in Bavaria, Germany, allelic variation in three selected loci of the FADS1 FADS2 gene cluster was analysed and used for haplotype construction. Associations with plasma phospholipid PUFA (n 163) and PUFA in erythrocyte membranes (n 535) were investigated by regression analysis. All haplotypes of the original five-loci haplotypes of our previous study could be replicated. In addition, associations with serum phospholipid PUFA were confirmed in the present data set. Although less pronounced, associations between FADS1 FADS2 haplotypes and PUFA in erythrocyte membranes, particularly arachidonic and dihomo-gamma-linolenic acid, could be established. We provide the first replication of the association of the FADS1 FADS2 gene cluster with PUFA in blood phospholipids. For the first time, such associations were also shown for PUFA in cell membranes. Show less