Impact of Caloric Restriction on Metabolomic Aging
Principal Investigator
Name
Waylon Hastings
Degrees
B.S., B.A., M.S., Ph.D.
Institution
Texas A&M AgriLife Research
Position Title
Assistant Professor
About this CDAS Project
Study
IDATA
(Learn more about this study)
Project ID
IDATA-95
Initial CDAS Request Approval
Dec 9, 2025
Title
Impact of Caloric Restriction on Metabolomic Aging
Summary
Aging research suggests that targeting individual disease processes will generate only modest improvements in health outcomes Instead, it has been proposed that the most effective means to improve health is to target aging processes directly (2, 3). One of the most promising geroprotective interventions is caloric restriction (CR). Work in animal models, including non-human primates, has shown extensive evidence for increase in lifespan and healthspan following CR initiated in adulthood or midlife. The foremost data for the effects of CR in humans comes from the NIA-supported Comprehensive Assessment of the Long-term Effects of Reducing Energy (CALERIE) study. Using data from CALERIE, researchers have provided evidence for significant improvements to cardiometabolic health, liver functioning, and immune health via decreased inflammation following moderate CR over a span of two years. Moreover, CR was shown to reduce the rate of physiological aging as measured using a panel of blood chemistry biomarkers. Impacts of CR on molecular measures of biological aging, namely epigenetic clocks and telomere length, were mixed, leaving it unclear whether and how CR impacts processes of biological aging.
The exact mechanism(s) linking CR to observed increases in lifespan are varied and actively contested. Predominate theories include attenuation of oxidative damage, alterations to glucose-insulin signaling, and reduced inflammation, each of which may provide independent and interactive contributions. One appealing avenue to explore these processes simultaneously is metabolomics, the comprehensive profiling of metabolites, their precursors, and their derivatives. Metabolomics has been successfully applied to study the biological processes implicated in several age-related diseases including cancer, cardiovascular disease, and diabetes. Moreover, the technique has also been used to explore changes in the rate of biological aging in the form of a metabolomic age score. Constructed using NMR spectroscopy in urine, this metabolomic age score was significantly associated with risk for mortality and age-related decline. With diverse functions spanning intracellular and extracellular domains, metabolomics represents a prime data source to disentangle conflicting findings observed of CR between physiological and molecular measures of biological aging.
The overarching goal of this proposal is to describe the distribution of metabolomic aging in the general population via metabolomic age scores, and then explore the impact of caloric restriction on changes in metabolomic aging. Metabolomic data from different cohorts across the Consortium of Metabolomics Studies (COMETS) will be linked using common IDs attributable to a shared measurement platform (Metabolon) to form a robust reference population. Three metabolomic indices of biological aging will be constructed using elastic net regression as representation of age-related change in the metabolome (MetaboAge1), physiological dysregulation (MetaboAge2), and metabolomic-mediated risk for mortality (MetaboAge3).
Aims
Aim 1: Generate metabolomic aging scores and test associations with health and functioning
Metabolites considered for inclusion in MetaboAge measures is restricted to analytes quantified in greater than 90% of CALERIE samples at each time point (n=1,117). Metabolomic data from each COMETS study (e.g., IDAT) will be matched to CALERIE using shared identifiers from the Metabolon platform (i.e., Comp_id). The array of possible metabolites included in MetaboAge measures will be restricted with the inclusion of each new COMETS cohort such that any analyte considered for inclusion in the MetaboAge panel must be a) measured in 90% of CALERIE samples at each time point, and b) measured in 75% of the total reference population sample.
After reducing the array of possible metabolites, the reference population is restricted to nonpregnant participants and screened for outliers. Following approaches established by first-generation epigenetic clocks, we will conduct elastic net regression on chronological age to identify a parsimonious of age-related metabolomic changes. MetaboAge2 is constructed as a individuals’ deviation (via Mahalanobis distance) from from a theoretically ideal state represented by a multi-dimensional mean vector µ= (m1, m2,…mn) of metabolites values is calculated in a restricted sample of non-obese individuals aged 20-30 with no reported chronic disease. MetaboAge3 is trained to directly capture all-cause mortality risk following a two-phase statistical procedure. In the first phase, elastic net regression is implemented in a cox proportional hazards framework to identify a parsimonious set of metabolites explaining variation in time to mortality. These metabolites are extracted in the second phase, and Gompertz regression models are used to convert metabolomic mortality hazards into MetaboAge scores with meaningful units of years. After training in the full dataset, we will implement MetaboAge scores within each cohort to test associations with health and functioning across a wide variety of domains spanning sociodemographic strata (age, sex, race, BMI), physical/mental functioning (e.g., cognition, muscle strength, gait speed), lifestyle factors (income, education, stress, smoking, drinking), morbidity, and mortality.
Aim 2: Implement metabolomic aging scores in CALERIE™ to test if CR reduces the rate of metabolomic aging
We will conduct analysis to test the hypothesis that CR slows metabolomic aging by compute change-scores for each MetaboAge measure as the difference in MetaboAge values at the 12-month and 24-month follow-up assessments relative to baseline values (i.e., ΔMetaboAge12 = MetaboAge12 - MetaboAgeBL; ΔMetaboAge24 = MetaboAge24 - MetaboAgeBL). We will conduct analysis of these change scores using two complementary approaches: 1) Intent-to-Treat (ITT) analysis and 2) Effect-of-Treatment-on-the-Treated (TOT) analysis. Effect-sizes for both analyses are scaled according to the distribution of the MetaboAge measures at pre-treatment baseline. Specifically, MetaboAges are differenced from chronological ages and standard deviations for these age-difference values are used for scaling. In this manner, treatment effects are interpreted as Cohen’s d.
Aim 3: Investigate functional significance of metabolites in metabolomic age scores
We will conduct network-based enrichment of the metabolites comprising metabolomic aging scores to investigate possible mechanisms mediating differences in metabolomic aging. Functional enrichment will be performed in a using the FELLA package in R.
Collaborators
Waylon Hastings Texas A&M AgriLife