This blog was written by Romilly Hodges MS CNS CDN IFMCP
Last month I was honored to be invited to speak at the American Nutrition Association annual summit on the topic Targeting Biological Age with Diet and Lifestyle. Not only was it a much-appreciated opportunity to connect face-to-face with colleagues, but it was also exciting to witness the interest in biological age clocks and their potential clinical application.
In this article, I want to share with you some of the clocks that have so far been most used in scientific research, as well as the findings of some of the recent studies that are applying them. As researchers continue to “stretch the legs” of these assessment tools, it is interesting and notable to understand these applications and results. Research continues to indicate that these aging assessment tools are broadly useful and will continue to be refined over time.
Key Takeaways from this Article
- Scientists continue to use a variety of non-proprietary biological age clocks that have been developed over the past decade (i.e. research is not limited to the most recently developed clocks);
- Research broadly continues to support the utility of DNA methylation-based biological age clocks in predicting health outcomes, although there is more variability in individual studies looking at the impact of specific interventions;
- Interventions that are deliberately designed to target and optimize DNA methylation patterns may be expected to have the biggest impact on measures of biological age.
Epigenetic Clocks – An Update
Since the first epigenetic BioAge clocks captured the attention of scientists around the world for their head-turning accuracy at predicting chronological age from patterns in DNA methylation (the epigenetic mark that these clocks use in their measurement and analysis), the development and use of algorithms based on DNA methylation that can predict your chronological age, morbidity, mortality, and rate of aging has expanded rapidly.
There are now several epigenetic clocks regularly used in research studies. Here are the ones that are available to researchers (i.e., not proprietary), and most commonly used and published on:
- Horvath 2013: A pan-tissue clock based on 353 CpG sites (CpGs are where DNA methylation marks occur). It was trained on chronological age, although remarkably it has been shown to predict morbidity and mortality better than chronological age itself.
- Hannum 2013: A blood-based clock that uses 71 CpGs. It was also trained on chronological age but has, similar to the Horvath 2013 clock, been shown to predict morbidity and mortality better than chronological age.
- Skin and Blood 2018: A blood- and skin-specific clock based on 291 CpG sites. Trained on chronological age.
- PhenoAge 2018: A multi-tissue clock based on 513 CpG sites. It was trained on “phenotypic” age (mortality risk) derived from a machine-developed algorithm of 9 blood-based clinical biomarkers plus chronological age. The biomarkers are readily found on standard blood tests and there is also a free PhenoAge calculator that predicts biological age using the biomarkers without needing to measure DNA methylation.
- GrimAge 2019: Using 1030 CpGs, this clock was trained on biomarkers of plasma protein levels, plus a smoking estimator, and was developed to predict lifespan.
- DunedinPoAm 2020: A blood-based clock based on 46 CpG sites that is set as a Pace of Aging estimator using longitudinal data (which distinguishes it from other clocks that don’t use longitudinal data). It is set to predict morbidity and mortality.
- GrimAge2 2022: An update to the GrimAge 2019 clock that adds CRP and HbA1C as additional plasma proteins that feed into its algorithm. It is reported to be a more accurate predictor than the original GrimAge clock.
- DunedinPACE 2022: An update to the DunedinPoAm that adds 8 more years of longitudinal data, plus 173 CpG sites. It is reported to be superior to the DunedinPoAm.
(list adapted and updated from Piani et al., 2023)
In general, the “second” generation clocks, which switched to being set to predict aging-related events, morbidity, and mortality, have been demonstrated to outperform the “first generation” clocks trained on chronological age. In the list above, this begins at PhenoAge 2018. However, all of the clocks above continue to be used in research studies and demonstrate usefulness in predicting morbidity and mortality moreso than chronological age alone. In addition, the earlier clocks, such as Horvath 2013, while trained on chronological age were intended to assess more than chronological age alone – as Dr. Horvath describes, they were intended to investigate whether manifestations of “age acceleration” could be captured within DNA methylation patterns. So far, research appears to support this hypothesis.
On the Utility of Epigenetic BioAge Clocks
“Epigenetic Age Acceleration (EAA) may be a valid biomarker associated with healthy longevity…and may be used for risk stratification and risk estimation of future functional and cognitive aging.” – Jain et al. (2022) published in JAMA Network Open.
Recent Diet & Lifestyle Studies Using These Epigenetic Clocks
Below are the more recent notable papers reporting on the potential effects of diet and lifestyle interventions on BioAge as measured by epigenetic clocks. It is interesting to note that the clocks are often (though not always) able to pick up expected changes after interventions otherwise known to be beneficial. We will continue to learn more about the application and usefulness of these clocks as research progresses.
An 18-month polyphenol-enriched Mediterranean diet (Green-MED) directly moved the needle on DNA methylation sites, moreso than basic healthy dietary guidelines or a plain Mediterranean diet, the other two intervention arms of the trial. The Green-MED intervention included additional green tea and a high-polyphenol green shake containing Mankai. Researchers identified 1573 differentially methylated regions in the green-MED arm of the study compared to 377 and 174 in the healthy dietary guidelines and plain Mediterranean diet arms, respectively. In addition, those with greatest adherence to the Green-MED intervention finished with a lower BioAge as measured by the Li and Hannum clocks. No other clocks registered differences, despite the changes in DNA methylation.
Clocks used: Horvath 2013, Hannum 2013, Li 2018, Skin and Blood 2018, PhenoAge 2018, GrimAge2 2022, DunedinPACE.
Takeaways: Polyphenols appear to be particularly epigenetically active and can reshape DNA methylation patterns. Incidentally, the Diet & Lifestyle program in the DNA methylation-targeted diet & lifestyle intervention reported on below added an estimated 2,900 mg of polyphenol plant compounds for the intervention group. This is more than double the estimated added polyphenols in the Green-MED arm of this study of 1,240 mg.
Calorie restriction is one of the interventions that has delivered the most success in improving longevity – at least in animal studies, where lifespan extensions have been reported of up to 300% compared to ad libitum (free) eating. Observations of “gentle” calorie reductions in long-lived populations such as Okinawans suggest potential benefits in humans, although harms from more aggressive implementations have been reported, too
In this CALERIE trial, participants followed a calorie-restricted diet (targeting 25% calorie reduction, although the average achieved was actually about 12% reduction) for 24 months. Benefits were observed in terms of weight loss and improved metabolic health. Researchers identified a modest reduction in the DunedinPACE clock, but not in either of the PhenoAge and GrimAge clocks.
Clocks used: PhenoAge 2018, GrimAge 2019, DunedinPACE 2022
Takeaways: The lack of BioAge response as measured by the GrimAge and DunedinPACE clocks is surprising. It is possible that the lack of adherence to the target calorie reduction, or potential nutrient deficits induced by the long time period of caloric restriction may have influenced the results. We’ll be keeping an eye out for more research on caloric restriction and BioAge.
As a follow-up to the clinical pilot study published in 2021 (more below), this case series adds to the evidence that an intervention that is deliberately designed to target DNA methylation by providing epinutrients including methyl donor substrates and phytonutrient and nutrient DNA methylation modulators may favorably impact BioAge. Added phytonutrient content totaled an estimated 2,900 mg, which is high even compared to other plant-rich diets. In this case series, participants averaged a BioAge reduction of -4.60 years.
In addition, adherence to stated intervention goals ranged from 71 to 97 percent, suggesting that 100 percent adherence is not necessary to benefit from the intervention.
Clocks used: Horvath 2013.
Takeaways: Although small, these two studies provide early evidence that a carefully designed intervention, intended to impact epigenetic age, may favorably influence BioAge, more than other interventions. Larger studies are needed to confirm these findings but these early indications are very promising.
The DNA Methylation Diet & Lifestyle intervention used in this study can be found in Dr. Fitzgerald’s Younger You program for BioAge reversal.
The Sister Study is a large study of over 50,000 women in the USA and Puerto Rico between the ages of 35-74. Researchers in this study mapped eating patterns of participants to four healthy eating indices: (1) DASH diet; (2) Healthy Eating Index-2015; (3) Alternative Healthy Eating Index-2010; (4) Alternative Mediterranean diet. Biological age was recorded as lower for all four eating indices across all four clocks – Hannum 2013, Horvath 2013, PhenoAge 2018, and GrimAge 2019.
PhenoAge and GrimAge recorded the most significance in the association between diet quality and reduced BioAge. The biggest dietary effects were observed in women who did not meet physical activity guidelines.
Clocks used: Hannum 2013, Horvath 2013, PhenoAge 2018, GrimAge 2019
Takeaways: Higher diet quality is associated with a reduction of epigenetic age estimates and may be particularly helpful for women who are not meeting exercise guidelines.
In this analysis, researchers found that vitamin D supplementation (dosing variable, not reported) was associated with a -1.3 year reduction in the Horvath 2013 clock and that having a higher vitamin D status (25-OH-D) was associated with a lower GrimAge, particularly in women. No changes were associated with the Hannum and PhenoAge clocks.
Clocks used: Horvath 2013, Hannum 2013, PhenoAge 2018, GrimAge 2019.
Takeaways: Vitamin D supplementation was one of the first interventions to show potential to reduce BioAge – Chen et al reported a -1.85 BioAge reduction using 4,000 IU/d in the Horvath 2013 clock, and -1.90 BioAge reduction for 2,000 IU/d in the Hannum 2013 clock back in 2018. We also know that we lose on average 2.9 nmol/L 25-OH-D for every decade of age – in other words, our vitamin D status declines on average as we get older. However, not all clocks appear to pick up vitamin D as an important age-related factor, as also seen in Schöttker et al.’s 2019 paper which found that vitamin D was associated with age but independently from epigenetic age measures.
Over 200 women followed a 24-month program that largely mirrored a Mediterranean diet – increased intake of whole grains, vegetables, legumes, olive oil, and fish as well as reduced intake of sweetened foods, dairy, and red meat. The researchers evaluated BioAge with the GrimAge clock and found a decrease of -0.41 years compared to the control group. The study also had an intervention arm that only consisted of physical activity guidance, which did not result in any reduced BioAge.
Clocks used: GrimAge 2019
Takeaways: This is a modest but still potentially impactful reduction in biological age following a longer intervention time period with a cohort that made changes towards a Mediterranean-style eating pattern.
In this smaller study, improved grip strength and gait speed were associated with lower BioAge in obese older adults. In addition, the 16 obese study participants completed a 12-week weight loss intervention including biweekly exercise sessions and dietary counseling, resulting in a mean weight loss of 4.6kg, an increased grip strength and gait speed, and a reduction in biological age of -0.8, -1.1, and -0.5 years in the Hannum 2013, Horvath 2013, and PhenoAge 2018 clocks, respectively.
Clocks used: Horvath 2013, Hannum 2013, PhenoAge 2018.
Takeaways: These findings link physical function and obesity status with BioAge in older adults, and suggests that weight-reduction interventions may positively impact both. This study adds to other research that has found tissue-specific associations between obesity and advanced biological age here and here. Adolescents in the highest BMI group have also been found to have the most accelerated aging as measured by PhenoAge, GrimAge, DunedinPoAm, and DunedinPACE.
The publication of this paper on the potential of a deliberately targeted, multi-modal dietary and lifestyle intervention aimed at optimizing DNA methylation marked the first of its kind. The dietary component of the intervention was rich in methyl donor nutrients as phytonutrient DNA methylation modulators (estimated added phytonutrient content totaled 2,900 mg, based on unpublished data, higher than estimations of other plant-rich diets). Participants in the intervention arm, males between the ages of 50-72 followed the 8-week program resulting in an average reduction of BioAge of 1.96 compared to the start of the program and a statistically significant difference compared to controls. After the intervention, participants were 3.24 years younger than the control group.
Blood biomarkers for serum folate increased and triglycerides, total cholesterol, and LDL cholesterol decreased following the intervention.
Clocks used: Horvath 2013.
Takeaways: The first study to successfully use a deliberately designed dietary and lifestyle intervention sets a new trajectory for BioAge research.
In a study of 120 elderly healthy individuals based in Italy and Poland, who followed a Mediterranean-like diet for 1 year, researchers identified a statistically significant reduction in epigenetic age, as measured by the Horvath 2013 clock, in Polish females and in those who were epigenetically older at baseline.
What was the Mediterranean-like diet? In this case, it consisted of following the nutrient recommendations for older adults from the European Commission, plus 9 sessions with tailored dietary advice. Participants received Mediterranean Diet-compatible foods including whole-grain pasta, olive oil, and cheese to aid with compliance. Participants also consumed supplemental vitamin D at 400 IU/d.
Clocks used: Horvath 2013.
Takeaways: This study suggests the importance of who is receiving any given intervention since there can be sex-specific differences in outcomes as well as differences based on each individual’s starting health and BioAge.
Research using DNA methylation-based biological age clocks is only expected to continue to increase. To date, studies generally support their utility, finding correlations that are supported by our existing understanding of health-promoting interventions, even as we continue to learn more about their use and interpretation. I expect that interventions that are deliberately designed to target and optimize DNA methylation patterns will continue to have the largest effects on epigenetic biological age.