Multi-omic analysis of biological aging biomarkers in long-term calorie restriction and endurance exercise practitioners: A cross-sectional study

Abstract

Calorie restriction (CR) and physical exercise (EX) are well-established interventions known to extend health span and lifespan in animal models. However, their impact on human biological aging remains unclear. With recent advances in omics technologies and biological age (BioAge) metrics, it is now possible to assess the impact of these lifestyle interventions without the need for long-term follow-up. This study compared BioAge biomarkers in 41 middle-aged and older adult long-term CR practitioners, 41 age- and sex-matched endurance athletes (EX), and 35 sedentary controls consuming Western diets (WD), through PhenoAge: a composite score derived from nine blood-biomarkers. Additionally, a subset of participants (12 CR, 11 EX, and 12 WD) underwent multi-omic profiling, including DNA methylation and RNAseq of colon mucosa, blood metabolomics, and stool metagenomics. A group of six young WD subjects (yWD) served as a reference for BioAge calculation using Mahalanobis distance across six omic layers. The results demonstrated consistently lower BioAge biomarkers in both CR and EX groups compared to WD controls across all layers. CR participants exhibited lower BioAge in gut microbiome and blood-derived omics, while EX participants had lower BioAge in colon mucosa-derived epigenetic and transcriptomic markers, suggesting potential tissue-specific effects. Multi-omic pathway enrichment analyses suggested both shared and intervention-specific mechanisms, including oxidative stress and basal transcription as common pathways, with ether lipid metabolism uniquely enriched in CR. Despite limitations due to sample size, these findings contribute to the broader understanding of the potential anti-aging effects of CR and EX, offering promising directions for further research.

Keywords: biological aging; calorie restriction; multi‐omic; oxidative stress; physical exercise.

FIGURE 1

Comparative analysis of bioage measures and correlations with body composition metrics in calorie restriction, endurance exercise, and western diet groups. (a) Boxplots illustrating the six BioAge measures across the CR, EX, and WD groups. All p‐values have been adjusted for multiple comparisons using the Tukey HSD method. (b) Correlation plot displaying the relationships between BMI, body fat, and the six BioAge measures. The upper triangle shows pairwise Pearson correlation coefficients, the diagonal presents standardized distributions of each biomarker, and the lower triangle features pairwise scatterplots for each combination of measures.

FIGURE 2

Multi‐omic signatures of calorie restriction and endurance exercise. (a) Heatmap illustrating the multi‐omic signature identified in the comparison of CR versus WD across various omic layers. (b) Circos plot depicting the correlation structure among the biomarkers constituting the multi‐omic CR signature. (c) Heatmap illustrating the multi‐omic signature identified in the comparison of EX versus WD across various omic layers. (d) Circos plot highlighting the correlation structure among the biomarkers in the multi‐omic EX signature.

FIGURE 3

Enrichment of KEGG Pathways in CR and EX Signatures. Bar plots displaying the significantly enriched KEGG pathways in the comparison of CR versus WD (top) and in the comparison of EX versus WD (bottom). Enrichment analyses are based on four omic layers: Gut microbiome metagenomics, metabolomics, epigenomics (DNAm), and transcriptomics (mRNA). Pathway names are presented on the y‐axis, while FDR‐adjusted p‐values (−log10 transformed) are shown on the x‐axis.

FIGURE 4

Enrichment of Inflammatory in CR and EX Signatures. Bar plots displaying the significantly enriched inflammatory pathways in the comparison of CR versus WD (left) and in the comparison of EX versus WD (right). Enrichment analyses are based on epigenomics (DNAm), and transcriptomics (mRNA) layers. Pathway names are presented on the y‐axis, while FDR‐adjusted p‐values (−log10 transformed) are shown on the x‐axis.