Epigenome-wide Association Study Analysis of Calorie Restriction in Humans, CALERIETM Trial Analysis

Abstract

Calorie restriction (CR) increases healthy life span and is accompanied by slowing or reversal of aging-associated DNA methylation (DNAm) changes in animal models. In the Comprehensive Assessment of Long-term Effects of Reducing Intake of Energy (CALERIETM) human trial, we evaluated associations of CR and changes in whole-blood DNAm. CALERIETM randomized 220 healthy, nonobese adults in a 2:1 allocation to 2 years of CR or ad libitum (AL) diet. The average CR in the treatment group through 24 months of follow-up was 12%. Whole blood (baseline, 12, and 24 months) DNAm profiles were measured. Epigenome-wide association study (EWAS) analysis tested CR-induced changes from baseline to 12 and 24 months in the n = 197 participants with available DNAm data. CR treatment was not associated with epigenome-wide significant (false discovery rate [FDR] < 0.05) DNAm changes at the individual-CpG-site level. Secondary analysis of sets of CpG sites identified in published EWAS revealed that CR induced DNAm changes opposite to those associated with higher body mass index and cigarette smoking (p < .003 at 12- and 24-month follow-ups). In contrast, CR altered DNAm at chronological-age-associated CpG sites in the direction of older age (p < .003 at 12- and 24-month follow-ups). Although individual CpG site DNAm changes in response to CR were not identified, analyses of sets CpGs identified in prior EWAS revealed CR-induced changes to blood DNAm. Altered CpG sets were enriched for insulin production, glucose tolerance, inflammation, and DNA-binding and DNA-regulation pathways, several of which are known to be modified by CR. DNAm changes may contribute to CR effects on aging.

Keywords: Caloric restriction; Epigenome; Human aging.

Figure 1.

Study design. (A) CALERIE^TM trial design. Two hundred and twenty participants were randomly assigned to either 25% calorie restriction (CR) or ad libitum (AL) at a 2:1 ratio. Of the 220 participants assigned, 218 started and 188 completed the intervention. (B) Blood samples were collected from participants at baseline and 12- and 24-month follow-ups. DNA was isolated and stored. DNA methylation was assayed with Illumina methyl EPIC bead chip arrays. After quality control and normalization, epigenome-wide association study (EWAS) analysis tested CALERIE^TM intervention effects at 12- and 24-month follow-ups at each of 828,613 CpG sites. Finally, we conducted secondary analysis comparing results from CALERIE^TM EWAS with results from published EWAS of BMI, cigarette smoking, and chronological age.

Figure 2.

Quantile-quantile (QQ) plots of p-value distributions from epigenome-wide association study (EWAS) analysis of CALERIE^TM treatment effects at 12- and 24-month follow-ups. The figure shows QQ plots for EWAS of blood DNA methylation changes in response to CR at 12 months (Genomic Inflation—0.97; Panel A) and 24 months (Genomic Inflation—0.99; Panel B).

Figure 3.

Distributions of test statistics from epigenome-wide association study (EWAS) analysis of CALERIE^TM treatment effects for CpG sites identified in published EWAS of body mass index, cigarette smoking, and chronological age. The figure shows box plots of CALERIE^TM EWAS test statistics for 3 groups of CpGs sites for each target phenotype. The blue-shaded box plot shows CALERIE^TM EWAS test statistics for CpG sites that exhibit lower levels of DNA methylation in association with the target phenotype in published EWAS of independent samples. The red-shaded box plot shows CALERIE^TM EWAS test statistics for CpG sites that exhibit higher levels of DNA methylation in association with the target phenotype in published EWAS. The gray-shaded box plot shows test statistics for CpG sites not associated with the target phenotype in published EWAS. Box plots are drawn for CALERIE^TM EWAS results from 12- and 24-month follow-ups. Stars indicate p-value threshold for comparisons based on a Wilcoxon Rank Sum Test. (* < .05, ** < .005, *** < .0005). Panel A graphs data grouped according to EWAS of body mass index (BMI) by Wahl et al. (18). The figure illustrates reversal of BMI-associated DNAm changes in response to CALERIE^TM intervention. Panel B graphs data grouped according to EWAS of cigarette smoking by Joehanes et al. (21). The figure illustrates reversal of smoking-associated DNAm changes in response to CALERIE^TM intervention. Panel C graphs data grouped according to EWAS of chronological age by McCartney et al. (19). The figure illustrates induction of older-chronological-age-associated DNAm changes in response to CALERIE^TM intervention, although only for sites exhibiting increased DNAm in older as compared to younger people.