Author:

Abstract:

Background DNA methylation in gene-promoter regions is usually related to repressed gene expressions1. DNA methylation can be affected by environmental factors such as diet and lifestyle2, which are also associated with sarcopenia3. This suggests a potential connection between DNA methylation and sarcopenia. However, the association between nutrient intake, physical activity and sarcopenia-related DNA methylation remains unknown. Objectives To explore the relationship between nutrient intake, physical activity and sarcopenia-related methylation in promoter regions. Methods 24 older Caucasian women (72.5 ± 4.2 yr) were identified as “sarcopenic” from a group of 247 (aged 65-80 yrs) by cutoff points of lower quintile hand grip strength (26 kg) and skeletal muscle index (6.75 kg/m2)4. A non-sarcopenic group (n=24, 70.5 ± 3.3 yr) was selected by generally age-matching with the sarcopenic one. Nutrient intake (n=20 in the sarcopenic group) and physical activity score (PASE) were assessed by questionnaires. DNA methylation of whole blood sample was assessed using Infinium MethylationEPIC BeadChip. Differentially methylated CpG sites at promoter regions were identified by comparing methylation values (β values) of sarcopenic women with non-sarcopenic ones (p-value < 0.01). Methylation scores were calculated by summing individual β values in hyper- and hypomethylated CpG sites, respectively. Participants were split into a training set (n=30) and a test set (n=14). Lasso method was applied for variable selecting and model building in the training set with methylation scores as dependent variables. Standardized values of age, BMI, nutrient intake and PASE were included as candidate independent variables. The model was further evaluated in the test set. Models with the highest R2 in the training and the test set were retained. Results In the sarcopenic group, promoter regions of 1621 genes were significantly differentially methylated (420 hypomethylated, 1201 hypermethylated) compared to the non-sarcopenic group. Lasso regression showed that the hypermethylated score was negatively related to BMI, PASE and protein intake (R2 = 0.7% in the test set) and the hypomethylated score was positively related to BMI, protein and selenium intake (R2 = 9.6% in the test set). Conclusion BMI, physical activity, protein and selenium intake are mildly associated with sarcopenia-related DNA methylation in promoter regions, indicating their possible influence on sarcopenia.