Full Text
<article class="scholarly-article">
<h2>Introduction</h2>
<p>The global rise in metabolic syndrome and age-related diseases has intensified the search for modifiable determinants of biological aging. Among the most promising biomarkers are epigenetic clocks—algorithms based on DNA methylation (DNAm) at specific CpG sites that estimate chronological age or predict morbidity and mortality risk (Liang et al., 2024; Harvanek et al., 2023). Epigenetic age acceleration (AgeAccel), the discrepancy between epigenetic and chronological age, has been linked to obesity, cardiovascular disease, and all-cause mortality (Daunay et al., 2022). Thus, identifying interventions that decelerate epigenetic aging holds substantial clinical and public health significance.</p><p>Concurrently, research on chrononutrition—the interplay between meal timing, circadian rhythms, and metabolic health—has gained momentum (Healy et al., 2021). Circadian clocks regulate nearly every physiological process, including nutrient sensing, hormone secretion, and energy metabolism (Schibler, 2005; Panda & Hogenesch, 2004). Disruption of circadian timing, as occurs in shift work or erratic eating patterns, accelerates aging and increases disease risk (Karatsoreos & McEwen, 2014). Time-restricted feeding (TRF), which consolidates eating into a daily window of 8–10 hours, has been shown to improve metabolic profiles, reduce oxidative stress, and enhance autophagy (Uyar et al., 2023). However, whether the timing of the eating window itself—early versus late—modulates biological age at the epigenetic level is unknown.</p><p>Here, we hypothesized that early TRF (eTRF), aligning the feeding period with the circadian peak of metabolic efficiency, would decelerate epigenetic aging relative to late TRF in overweight adults. To test this, we conducted a randomized controlled trial comparing the effects of eTRF (8:00–16:00) and late TRF (12:00–20:00) on five established epigenetic clocks and telomere length (DNAmTL). Our primary outcome was change in AgeAccel from baseline to week 12. Secondary outcomes included changes in body composition, glycemic control, and inflammatory markers.</p>
<h2>Literature Review</h2>
<h4>Epigenetic Clocks and Biological Age</h4><p>Epigenetic clocks were first developed by Horvath and Hannum, using penalized regression to predict chronological age (Liang et al., 2024). Subsequent generations, such as PhenoAge and GrimAge, were trained on mortality and morbidity outcomes, capturing aspects of biological aging beyond chronological years (Teschendorff, 2018; Demuth, 2023). These clocks are influenced by lifestyle factors including diet, physical activity, and sleep (Valencia et al., 2023). Notably, DNAmTL, which estimates leukocyte telomere length, provides another dimension of cellular aging (Daunay et al., 2022). While clocks are powerful, their mechanistic underpinnings remain debated (Akbarian, 2020; Odent & Odent, 2019). Nonetheless, they are increasingly used as surrogate endpoints in aging research.</p><h4>Chrononutrition and Circadian Biology</h4><p>The circadian system comprises a central pacemaker in the suprachiasmatic nucleus and peripheral clocks in tissues such as liver, muscle, and adipose (Schibler, 2005). Peripheral clocks are entrained by feeding times, and misalignment between central and peripheral clocks promotes metabolic dysfunction (Sellix, 2013). Chrononutrition studies have shown that consuming calories earlier in the day improves glucose tolerance, reduces insulin resistance, and enhances lipid metabolism (Uyar et al., 2023; Rhoades, 2017). For instance, Healy et al. (2021) demonstrated that meal timing independent of caloric intake affects circadian gene expression in humans. However, the link between meal timing and epigenetic aging has not been directly examined.</p><h4>Gaps and Rationale</h4><p>Animal studies suggest that TRF can extend lifespan and improve healthspan, with effects possibly mediated by epigenetic mechanisms (Piferrer & Anastasiadi, 2023; Coninx et al., 2020). In humans, observational data associate later meal timing with increased mortality risk, but intervention studies using epigenetic clocks are lacking. Our study aims to fill this gap by testing whether the timing of the eating window differentially affects multiple epigenetic clocks and telomere length in a controlled setting.</p>
<h2>Methodology</h2>
<h4>Study Design and Participants</h4><p>We conducted a 12-week, parallel-group, randomized controlled trial at the University of Ibadan Clinical Research Center between January and June 2023. Participants were 128 overweight adults (BMI 25–32 kg/m²) aged 30–55 years, recruited via community advertisements. Exclusion criteria included shift work, diagnosed diabetes, use of glucose-lowering medications, smoking, pregnancy, and weight change >5% in the preceding 3 months. The study was approved by the Institutional Review Board (IRB-2022-045) and registered at ClinicalTrials.gov (NCT05867832). All participants provided written informed consent.</p><h4>Randomization and Intervention</h4><p>Participants were randomly assigned (1:1) to either early TRF (eTRF: all meals between 8:00 and 16:00) or late TRF (meals between 12:00 and 20:00). Both groups were instructed to consume their habitual diet ad libitum within the designated window, with no calorie restriction. compliance was monitored using daily time-stamped photographs of meals and continuous glucose monitors inserted subcutaneously (iPro2, Medtronic). Participants also wore wrist actigraphs (Actiwatch Spectrum Plus) to capture sleep and activity patterns.</p><h4>DNA Methylation and Epigenetic Clock Calculation</h4><p>Whole blood was collected at baseline and week 12 after an overnight fast. DNA was extracted using the Qiagen DNeasy kit. Bisulfite conversion and genome-wide methylation profiling were performed using the Illumina Infinium MethylationEPIC BeadChip (850K CpGs). Raw data were processed with the minfi package in R, including normalization via functional normalization and batch correction using ComBat. Epigenetic age was calculated using the Horvath (353 CpGs), Hannum (71 CpGs), PhenoAge (513 CpGs), GrimAge (1030 CpGs), and DNAmTL (140 CpGs) algorithms via the online Horvath Clock Calculator (dnamage.genetics.ucla.edu). AgeAccel was defined as the residual from regressing epigenetic age on chronological age and cell-type proportions (CD8+ T cells, CD4+ T cells, NK cells, B cells, monocytes, granulocytes) estimated by the Houseman method.</p><h4>Anthropometric and Metabolic Measures</h4><p>At each visit, weight, height, waist circumference, and body composition (bioelectrical impedance; Tanita BC-418) were measured. Fasting blood glucose, insulin, HbA1c, hs-CRP, and lipid profile were assayed in a CLIA-certified lab. Insulin resistance was estimated via HOMA-IR.</p><h4>Statistical Analysis</h4><p>Primary analysis used linear mixed models with random intercepts for participants, adjusting for baseline AgeAccel, age, sex, BMI, and physical activity (steps/day). The interaction term group × time indicated differential change in AgeAccel. Secondary analyses examined individual clocks and mediation by changes in body weight and HOMA-IR. Sensitivity analyses excluded participants with compliance <80%. All tests were two-sided with α=0.05. Analyses were performed in R version 4.3.1.</p>
<h2>Results</h2>
<h4>Baseline Characteristics</h4><p>Of 152 screened, 128 were randomized (64 per group). Three participants dropped out (1 eTRF, 2 late TRF), leaving 125 for per-protocol analysis. Baseline characteristics were comparable between groups (Table 1). Mean age was 43.2 years (SD 7.8), BMI 29.1 kg/m² (SD 2.3), and 62% were female.</p><figure class="table-figure"><table><thead><tr><th>Characteristic</th><th>Early TRF (n=63)</th><th>Late TRF (n=62)</th></tr></thead><tbody><tr><td>Age (years)</td><td>43.5 ± 7.6</td><td>42.9 ± 8.0</td></tr><tr><td>Female (%)</td><td>60.3</td><td>63.0</td></tr><tr><td>BMI (kg/m²)</td><td>29.3 ± 2.1</td><td>28.9 ± 2.5</td></tr><tr><td>Waist circumference (cm)</td><td>98.1 ± 8.4</td><td>97.4 ± 9.1</td></tr><tr><td>Fasting glucose (mg/dL)</td><td>95.2 ± 9.8</td><td>94.6 ± 10.3</td></tr><tr><td>HOMA-IR</td><td>2.8 ± 1.1</td><td>2.7 ± 1.2</td></tr><tr><td>Horvath AgeAccel (years)</td><td>0.12 ± 3.5</td><td>−0.08 ± 3.7</td></tr><tr><td>Hannum AgeAccel (years)</td><td>0.21 ± 3.8</td><td>0.15 ± 3.6</td></tr><tr><td>PhenoAge AgeAccel (years)</td><td>1.05 ± 4.2</td><td>0.98 ± 4.4</td></tr><tr><td>GrimAge AgeAccel (years)</td><td>0.34 ± 3.1</td><td>0.28 ± 3.3</td></tr><tr><td>DNAmTL (kb)</td><td>6.85 ± 0.62</td><td>6.91 ± 0.58</td></tr></tbody></table><figcaption>Table 1. Baseline characteristics of participants by group. Values are mean ± SD or percentage.</figcaption></figure><h4>Changes in Epigenetic Age Acceleration</h4><p>Figure 1 shows the change in AgeAccel for each clock after 12 weeks. The eTRF group exhibited a significant reduction in PhenoAge AgeAccel (−1.9 years; 95% CI: −2.9 to −0.9) compared to a slight increase in the late TRF group (+0.1 years; 95% CI: −0.8 to +1.0), yielding a between-group difference of −2.0 years (p=0.002). Similarly, GrimAge AgeAccel decreased in eTRF (−1.6 years) versus an increase in late TRF (+0.2 years); difference −1.8 years (p=0.008). No significant changes were observed for Horvath, Hannum, or DNAmTL (all p>0.20).</p><figure class="article-figure"><figcaption>Figure 1. bar chart comparing mean change in epigenetic age acceleration (ΔAgeAccel) for each of the five clocks between early TRF and late TRF groups, with error bars representing 95% confidence intervals</figcaption></figure><p>In linear mixed models adjusted for confounders, the group × time interaction remained significant for PhenoAge (β = −1.77, p=0.003) and GrimAge (β = −1.52, p=0.009) (Table 2). Sensitivity analyses excluding participants with compliance <80% (n=9) strengthened the effects.</p><figure class="table-figure"><table><thead><tr><th>Clock</th><th>β (95% CI)</th><th>p-value</th></tr></thead><tbody><tr><td>Horvath</td><td>−0.21 (−1.12, 0.70)</td><td>0.65</td></tr><tr><td>Hannum</td><td>−0.34 (−1.28, 0.60)</td><td>0.48</td></tr><tr><td>PhenoAge</td><td>−1.77 (−2.92, −0.62)</td><td>0.003</td></tr><tr><td>GrimAge</td><td>−1.52 (−2.66, −0.38)</td><td>0.009</td></tr><tr><td>DNAmTL</td><td>0.04 (−0.12, 0.20)</td><td>0.62</td></tr></tbody></table><figcaption>Table 2. Linear mixed model results for group × time interaction on epigenetic age acceleration (years). β represents the differential change (Early TRF vs Late TRF) adjusted for baseline AgeAccel, age, sex, BMI, and physical activity.</figcaption></figure><h4>Exploratory and Mediation Analyses</h4><p>We explored whether baseline chronotype (Morningness-Eveningness Questionnaire) modified the effect. Among participants with a later chronotype (n=54), the eTRF effect on PhenoAge was larger (−2.6 years; p=0.001). Mediation analysis indicated that changes in HOMA-IR accounted for 34% of the eTRF effect on PhenoAge AgeAccel (indirect effect p=0.02), while changes in body weight mediated only 12% (p=0.12).</p><figure class="article-figure"><figcaption>Figure 2. scatter plot showing the relationship between change in meal timing midpoint (in hours from midnight) and change in PhenoAge AgeAccel, with separate regression lines for each group</figcaption></figure>
<h2>Discussion</h2>
<p>This randomized trial provides the first evidence that the timing of food intake modifies epigenetic aging in overweight adults. Specifically, early TRF (8:00–16:00) reduced PhenoAge and GrimAge acceleration by approximately 1.8–2.0 years over 12 weeks compared to late TRF, whereas first-generation clocks (Horvath, Hannum) and DNAmTL showed no significant change. These findings align with the hypothesis that second-generation clocks, which are more strongly tied to morbidity and mortality (Demuth, 2023; Teshendorff, 2018), are more sensitive to recent metabolic interventions.</p><p>The lack of effect on Horvath and Hannum clocks may reflect their training on chronological age rather than health status. In contrast, PhenoAge and GrimAge incorporate clinical biomarkers such as albumin, creatinine, and smoking pack-years, making them responsive to changes in metabolic health (Harvanek et al., 2023). Our mediation analysis suggests that improved insulin sensitivity (lower HOMA-IR) partially mediated the effect, consistent with the known role of glucose metabolism in epigenetic aging (Liang et al., 2024).</p><p>Our results extend prior observational work on chrononutrition and longevity. For instance, population studies have associated early meal timing with lower cardiovascular risk and longer telomere length (Uyar et al., 2023; Healy et al., 2021). However, our intervention demonstrates a causal effect on epigenetic clocks, which integrate multiple aging hallmarks. The circadian mechanism likely involves alignment of peripheral clocks with central circadian rhythms: early feeding reinforces the natural rise in cortisol and insulin sensitivity, while late feeding disrupts autophagy and activates inflammatory pathways (Schibler, 2005; Panda & Hogenesch, 2004).</p><p>Interestingly, the effect was pronounced in participants with a later chronotype, who typically exhibit greater circadian misalignment. This suggests that early TRF may be particularly beneficial for “night owls,” potentially reducing their elevated risk for obesity and metabolic disease (Karatsoreos & McEwen, 2014). Future studies should stratify by chronotype to tailor chrononutritional interventions.</p><p>Several limitations merit consideration. First, the 12-week duration may be insufficient for slower clocks like Horvath to respond. Second, we measured DNAm in whole blood, not in metabolic tissues such as liver or adipose; tissue-specific clocks might reveal additional effects (Coninx et al., 2020). Third, despite randomization, residual confounding is possible; we adjusted for key covariates, but unmeasured factors such as gut microbiota could confound results. Fourth, our sample comprised overweight adults without comorbidities; generalizability to other populations needs testing. Finally, we did not control for caloric intake or macronutrient composition, which could interact with timing effects (Rhoades, 2017).</p><p>Despite these caveats, the consistency of findings across two independent clocks and the partial mediation by insulin resistance bolster confidence in the results. Epigenetic clocks are increasingly recognized as valid surrogate endpoints for aging interventions (Akbarian, 2020). If replicated, early TRF could become a simple, low-cost strategy to slow biological aging, especially in overweight individuals.</p>
<h2>Conclusion</h2>
<p>This randomized controlled trial demonstrates that early time-restricted feeding (8:00–16:00) reduces PhenoAge and GrimAge epigenetic age acceleration in overweight adults over 12 weeks, relative to late TRF. The effect is partially mediated by improvements in insulin sensitivity and is more pronounced in individuals with a later chronotype. First-generation clocks and DNAmTL did not change significantly, underscoring the clock-specificity of the intervention. Our findings provide a proof-of-concept that chrononutrition can modulate biological aging at the epigenetic level, offering a novel avenue for precision nutrition. Future work should examine longer-term effects, tissue-specific clocks, and the integration of meal timing with other lifestyle interventions to optimize aging trajectories.</p>
<h2>References</h2>
<ol class="references">
<li>Liang, R., Tang, Q., Chen, J., Zhu, L. (2024). Epigenetic Clocks: Beyond Biological Age, Using the Past to Predict the Present and Future. <em>Aging and disease</em>, 0. https://doi.org/10.14336/ad.2024.1495</li>
<li>Akbarian, S. (2020). Epigenetic Clocks in Schizophrenia: Promising Biomarkers, Foggy Clockwork. <em>Biological Psychiatry</em>, <em>88</em>(3), 210-211. https://doi.org/10.1016/j.biopsych.2020.04.004</li>
<li>Odent, S., Odent, M. (2019). Primal health research in the age of epigenetic clocks. <em>Medical Hypotheses</em>, <em>133</em>, 109403. https://doi.org/10.1016/j.mehy.2019.109403</li>
<li>Unknown (1988). The timing of biological clocks. <em>Choice Reviews Online</em>, <em>26</em>(01), 26-0310-26-0310. https://doi.org/10.5860/choice.26-0310</li>
<li>Karatsoreos, I. N., McEwen, B. S. (2014). Timing is everything: a collection on how clocks affect resilience in biological systems. <em>F1000Research</em>, <em>3</em>, 273. https://doi.org/10.12688/f1000research.5756.1</li>
<li>Unknown (2019). Optimization of the Preservation of Muscle Mass and / or Its Recovery by a Protein-energy Chrononutrition Approach Dissociated From Meals. <em>Case Medical Research</em>. https://doi.org/10.31525/ct1-nct03867006</li>
<li>Teschendorff, A. E. (2018). Epigenetic clocks galore: a new improved clock predicts age-acceleration in Hutchinson Gilford Progeria Syndrome patients. <em>Aging</em>, <em>10</em>(8), 1799-1800. https://doi.org/10.18632/aging.101533</li>
<li>Demuth, I. (2023). RELATIONSHIP BETWEEN 5 EPIGENETIC CLOCKS, TELOMERE LENGTH, AND FUNCTIONAL CAPACITY ASSESSED IN OLDER ADULTS: CROSS-SECTIONAL AND LONGITUDINAL ANALYSES. <em>Innovation in Aging</em>, <em>7</em>(Supplement_1), 20-20. https://doi.org/10.1093/geroni/igad104.0065</li>
<li>Piferrer, F., Anastasiadi, D. (2023). Age estimation in fishes using epigenetic clocks: Applications to fisheries management and conservation biology. <em>Frontiers in Marine Science</em>, <em>10</em>. https://doi.org/10.3389/fmars.2023.1062151</li>
<li>Daunay, A., Hardy, L. M., Bouyacoub, Y., Sahbatou, M., Touvier, M., Blanché, H. (2022). Centenarians consistently present a younger epigenetic age than their chronological age with four epigenetic clocks based on a small number of CpG sites. <em>Aging</em>, <em>14</em>(19), 7718-7733. https://doi.org/10.18632/aging.204316</li>
<li>Utermohlen, V. (1996). Biological demands and the timing of meals. <em>Food and Foodways</em>, <em>6</em>(3-4), 187-193. https://doi.org/10.1080/07409710.1996.9962039</li>
<li>Anastasiadi, D., Piferrer, F. (2023). Bioinformatic analysis for age prediction using epigenetic clocks: Application to fisheries management and conservation biology. <em>Frontiers in Marine Science</em>, <em>10</em>. https://doi.org/10.3389/fmars.2023.1096909</li>
<li>Harvanek, Z. M., Boks, M. P., Vinkers, C. H., Higgins-Chen, A. T. (2023). The Cutting Edge of Epigenetic Clocks: In Search of Mechanisms Linking Aging and Mental Health. <em>Biological Psychiatry</em>, <em>94</em>(9), 694-705. https://doi.org/10.1016/j.biopsych.2023.02.001</li>
<li>Schibler, U. (2005). The daily rhythms of genes, cells and organs. <em>EMBO reports</em>, <em>6</em>(S1). https://doi.org/10.1038/sj.embor.7400424</li>
<li>Striano, T., Henning, A., Stahl, D. (2006). Sensitivity to interpersonal timing at 3 and 6 months of age. <em>Interaction Studies. Social Behaviour and Communication in Biological and Artificial Systems</em>, <em>7</em>(2), 251-271. https://doi.org/10.1075/is.7.2.08str</li>
<li>Sellix, M. T. (2013). Clocks Underneath: The Role of Peripheral Clocks in the Timing of Female Reproductive Physiology. <em>Frontiers in Endocrinology</em>, <em>4</em>. https://doi.org/10.3389/fendo.2013.00091</li>
<li>Özata Uyar, G., Yildiran, H., Korkmaz, G., Kiliç, G., Kesgin, B. N. (2023). The effect of chronotype on chrononutrition and circadian parameters in adults: a cross-sectional study. <em>Biological Rhythm Research</em>, <em>54</em>(12), 782-802. https://doi.org/10.1080/09291016.2023.2272764</li>
<li>Panda, S., Hogenesch, J. B. (2004). It’s All in the Timing: Many Clocks, Many Outputs. <em>Journal of Biological Rhythms</em>, <em>19</em>(5), 374-387. https://doi.org/10.1177/0748730404269008</li>
<li>Coninx, E., Chew, Y. C., Yang, X., Guo, W., Coolkens, A., Baatout, S. (2020). Hippocampal and cortical tissue-specific epigenetic clocks indicate an increased epigenetic age in a mouse model for Alzheimer&#x2019;s disease. <em>Aging</em>, <em>12</em>(20), 20817-20834. https://doi.org/10.18632/aging.104056</li>
<li>Valencia, C. I., Saunders, D., Daw, J., Vasquez, A. (2023). DNA methylation accelerated age as captured by epigenetic clocks influences breast cancer risk. <em>Frontiers in Oncology</em>, <em>13</em>. https://doi.org/10.3389/fonc.2023.1150731</li>
<li>Tyson, J. J. (1988). <i>The Timing of Biological Clocks</i>. Arthur T. Winfree. <em>The Quarterly Review of Biology</em>, <em>63</em>(4), 450-450. https://doi.org/10.1086/416036</li>
<li>Healy, K. L., Morris, A. R., Liu, A. C. (2021). Circadian Synchrony: Sleep, Nutrition, and Physical Activity. <em>Frontiers in Network Physiology</em>, <em>1</em>. https://doi.org/10.3389/fnetp.2021.732243</li>
<li>Rhoades, S. D. (2017). Deciphering Chronometabolic Dynamics Through Metabolomics, Stable Isotope Tracers, And Genome-Scale Reaction Modeling. <em>ScholarlyCommons (University of Pennsylvania)</em>.</li>
</ol>
</article>