.
O P E N A C C E S S S O U R C E (.PDF) : MedRXiv
Abstract
Manipulations to set back biological age and extend lifespan in animal models are well established, and translation to humans has begun. The length of human life makes it impractical to evaluate results by plotting mortality curves, so surrogate markers of age have been suggested and, at present, the best established surrogates are DNA methylation clocks. Herein we report on a randomized, controlled clinical trial designed to be a first step in evaluating the effect of a diet and lifestyle intervention on biological age. Compared to participants in the control group (n=20), participants in the treatment group tested an average 3.23 years younger at the end of the eight-week program according to the Horvath DNAmAge clock (p=0.018).
Those in the treatment group (n=18) tested an average 1.96 years younger at the end of the program compared to the same individuals at the beginning with a strong trend towards significance (p=0.066 for within group change). This is the first such trial to demonstrate a potential reversal of biological age.
In this study, the intervention was confined to diet and lifestyle changes previously identified as safe to use. The prescribed program included multiple components with documented mechanistic activity on epigenetic pathways, including moderate exercise, breathing exercises for stress, and a diet rich in methyl donor nutrients and polyphenols.
1 INTRODUCTION
Advanced age is the largest risk factor for impaired mental and physical function and many noncommunicable diseases including cancer, neurodegeneration, type 2 diabetes, and cardiovascular disease (Jin et al., 2015; Sen et al., 2016). The growing health-related economic and social challenges of our rapidly aging population are well recognized and affect individuals, their families, health systems and economies. Considering economics alone, using the Future Elderly Model, Goldman showed that delaying aging by 2.2 years (with associated extension of healthspan) could save $7 trillion over fifty years. This broad approach was identified to be a much better investment than disease-specific spending (Goldman, 2017). Thus, if interventions can be identified that extend healthspan even modestly, benefits for public health and healthcare economics will be substantial.
DNA methylation is the addition of a methyl group to cytosine residues at selective areas on a chromosome (e.g. CpG islands, shelf/shore, exons, open sea). Methylation constitutes the best-studied, and likely most resilient of many mechanisms controlling gene expression (Li & Zhang, 2014). Unique among epigenetic markers, DNA methylation can readily and cheaply be mapped from tissue samples. Of 20+ million methylation sites on the human genome, there are a few thousand at which methylation levels are tightly correlated with age. Currently, the best biochemical markers of an individual’s age are all based on patterns of methylation (Horvath & Raj, 2018). This has led some researchers to propose that aging itself has its basis in epigenetic changes (including methylation changes) over time (Field et al., 2018; Johnson et al., 2012; Mitteldorf, 2013; Rando & Chang, 2012).
As of this writing, the best-studied methylation-based clock is the multi-tissue DNAmAge clock (Horvath, 2013). At the time this study design was approved, there were few viable alternatives. Horvath’s DNAmAge clock predicts all-cause mortality and multiple morbidities better than chronological age. Methylation clocks (including DNAmAge) are based on systematic methylation changes with age, with about 60% of CpG sites losing methylation with age and 40% gaining methylation. This is distinct from stochastic changes, “methylation drift”, unpredictable changes which vary among individuals and cell-by-cell within individuals. Systematic methylation changes include hypermethylation in promotor regions of tumor suppressor genes (inhibiting expression) and hypomethylation promoting inflammatory cytokines (promoting expression). Saliva can be considered a good source of high-quality DNA, containing both white blood cells and buccal cells, and is a suitable tissue type to be assessed for the DNAmAge clock (Horvath, 2013; Langie et al., 2017).
The dietary recommendations employed as part of the treatment protocol for this study were based largely on biochemistry and generalized measures of health, because few dietary associations with the DNAmAge clock have yet been established. A modest, but significant, reduction in DNAmAge in individuals consuming a non-specific lean meat, fish and plant-based diet (as measured by blood carotenoids) has been observed (Quach et al., 2017). It is possible that changes of a greater magnitude require a more targeted approach. The dietary intervention used here was also plant-centered, but including a high intake of nutrients that are substrates or cofactors in methylation biosynthetic pathways (e.g. containing folate, betaine), ten-eleven translocation demethylase cofactors and modulators (e.g. alpha ketoglutarate, vitamin C and vitamin A) (Hore, 2017) and polyphenolic modulators of DNA methyl transferases (DNMT) (e.g. curcumin, epigallocatechin gallate (EGCG), rosmarinic acid, quercetin, luteolin). It also included limited nutrient-dense animal proteins (e.g. liver, egg). The diet restricted carbohydrates and included mild intermittent fasting, both designed to lower glycemic cycling. The diet was supplemented daily with a fruit and vegetable powder, also rich in polyphenolic modulators of DNMT activity , and a probiotic providing 40 million CFU of Lactobacillus plantarum 299v. L. plantarum has been shown to be a folate producer in the presence of para aminobenzoic acid (PABA) (Sybesma et al., 2003); it also has been demonstrated to alter gene expression (Hariri et al., 2015).
Lifestyle guidance in this study included a minimum of 30 minutes of exercise per day, at least 5 days per week at an intensity of 60-80 percent of maximum perceived exertion. Exercise is well-known to be broadly beneficial for almost every aspect of health and has been shown to extend mean lifespan in animal models. Exploration of the effect of exercise on the methylome has recently begun. For example, regular tai chi practice was associated with slowing of age-related DNA methylation losses in 500 women (Ren et al., 2012). In another study of 647 women, a lifelong history of exercise was associated with a similar endpoint (White et al., 2013). These results were not reported in terms of the Horvath clock, because it had not yet been developed. One systematic review of human studies found that regular, daily physical activity was associated with lower blood levels of homocysteine, which when elevated, suggests an insufficiency of methylation capacity (e Silva & da Mota, 2014). Excessive exercise may accelerate methylation aging, but this danger has only been observed in elite, competitive athletes (Spólnicka et al., 2018).
Twice-daily breathing exercises that elicit the Relaxation Response were prescribed for stress reduction. It was recently demonstrated that 60 days of relaxation practice designed to elicit the Relaxation Response, 20 minutes twice per day, could significantly reduce DNAmAge as measured by the Zbieć-Piekarska clock in their group of healthy participants (though not in their ‘patient’ group) (Pavanello et al., 2019). Almost a quarter of the DNAmAge CpG sites (85/353) are located in glucocorticoid response elements, pointing to a likely relationship between stress and accelerated aging. Cumulative lifetime stress has been shown to be associated with accelerated aging of the methylome (Zannas et al., 2015). Zannas et al. also reported that dexamethasone, a glucocorticoid agonist, can advance the DNAmAge clock and induce associated transcriptional changes. Dexamethasone-regulated genes showed enriched association of aging-related diseases, including coronary artery disease, arteriosclerosis and leukemias. Other findings include that PTSD contributes to accelerated methylation age (Wolf et al., 2016); and that greater infant distress (lack of caregiver contact) is associated with an underdeveloped, younger epigenetic age (Moore et al., 2017).
This study aimed to optimize sleep, with a recommendation for at least seven hours nightly. Seven hours is generally considered to be healthy (Panel et al., 2015), but the limited data on accelerated aging only relates to extremes of sleep deprivation. A (presumably transient) effect of sleep deprivation on genomewide methylation patterns in blood has been demonstrated (Nilsson et al., 2016). Acceleration of the DNAmAge clock has been associated with insomnia in a sample of 2078 women (Carroll et al., 2017). Carskadon et al (2019) found an association between poor quality / fewer hours of sleep with age acceleration in a small sample of 12 female college students.
This multimodal (“systems”) intervention is reflective of a clinically-used approach that combines individual interventions, each of which carry evidence of favorable influence on the DNA methylome and of which several authors of this study have clinical experience of health benefits. Such interventions likely produce synergistic effects and reduce the possibility of negative effects from one diseasepromoting input canceling out the benefits of another health-promoting input. Dietary and lifestyle interventions, as used here, target upstream influences that are generally considered safe, even over the long term.
By design, an important endpoint of this study was to be Horvath’s DNAmAge clock, to see if it could be potentially slowed or reversed. This is to say we have tentatively accepted the hypothesis that the methylation pattern from which the DNAmAge clock is computed is a driver of aging (and the chronic diseases of aging), thus we expect that attempting to directly influence the DNA methylome using diet and lifestyle to set back DNAmAge will lead to a healthier, more “youthful” metabolism. To date, one small pilot study (Fahy et al., 2019) has been reported to have set back the DNAmAge clock over the course of 12 months by 1.5 (plus the one-year duration of the study) years in humans, using a combination of growth hormone, metformin, DHEA and two dietary supplements. Herein we report comparable initial results based on diet and lifestyle interventions employed for eight weeks (preceded by a one-week washout period).
2 RESULTS
2.1 Methylation clock setback
Compared to participants in the control group (n=20), participants in the treatment group scored an average 3.23 years younger at the end of the eight-week program according to the Horvath DNAmAge clock (p=0.018). Those in the treatment group (n=18) scored an average 1.96 years younger, at the end of the program compared to the same individuals at the beginning with a strong trend towards significance (p=0.066 for within group change). Control participants scored an average of 1.27 years older at the end of the study period, though this within-group increase was not statistically significant (p= 0.153). Comparison of DNAmAge change between treatment and control groups is shown in Figure 1 whereas within group changes for the treatment group are shown in Figure 2.
In both treatment and control groups, global average methylation stayed the same over the course of the study, with no net increase or decrease in the 353 sites that compose the Horvath clock.
3 DISCUSSION
3.1 Significance of Results
The significance of these findings is multi-factorial, but primarily as the first demonstration of potential reversal of epigenetic age in a randomized, controlled clinical trial, accounting for any normal variability in epigenetic methylation. Secondarily, this is the first report of a diet and lifestyle intervention reducing biologic aging. Notably, the scale of potential reduction, while modest in magnitude, may correlate with meaningful socioeconomic benefits, and appears to have the potential to be broadly achievable.
Published in the fall of 2019, the TRIIM study (Fahy et al., 2019) was the first demonstration of a set of interventions setting back the flagship Horvath clock DNAmAge (Horvath, 2013). In TRIIM, a one-year regimen of daily injection of growth hormone plus one prescription drug and three nutritional supplements was shown to set back the DNAmAge clock by 1.5 years in 9 middle-aged men (plus the 1- year study duration = 2.5 years). In the present study, age set-back was achieved in eight weeks, using less expensive, less invasive, and otherwise generally beneficial interventions known to have mechanistic plausibility for affecting methylation pathways.
3.2 Targeting Epigenetics with Diet
The seminal work of Waterland and Jirtle (Waterland & Jirtle, 2003) in the Agouti mouse model marked a defining point in our understanding that nutrition elements could so affect DNA methylation marks as to silence gene expression and dramatically alter phenotype. The power of nutrition to bring about transformative phenotypic changes has held up over the intervening years, most strongly in animal studies, but also in some limited human trials (Waterland & Jirtle, 2003). Both TRIIM and the present study were able to effect changes on the DNA methylome without extra-dietary supplementation of known methyl donor nutrients (e.g., folate, vitamin B12, choline, SAMe or betaine) Illustrating a farreaching regulatory network on DNA methylation and representing a departure from previous studies that manipulated DNA methylation more directly with extra-dietary supplemental folate, B12 and other methyl donor nutrients (Pauwels et al., 2017; Sae-Lee et al., 2018; Waterland & Jirtle, 2003; Zhong et al., 2017).
3.3 Rationale for Not Using Supplemental Methyl Donor Nutrients
In designing the present study, extra-dietary supplementation of methyl donor nutrients was specifically avoided because a growing body of epidemiological evidence indicates potential long-term risks, to which the short-term studies were not sensitive. Although overall data are mixed, and certain conditions (e.g. pregnancy, macrocytic anemia, hyperhomocysteinemia, dietary limitations) often require extra-dietary supplementation, several trials have found a positive association between methyl donor supplementation and increased cancer risk: Published long-term follow up on 2,524 participants in the B-PROOF trial which assessed the effect of 2-3 years of daily supplementation with 400 mcg folic acid and 500 mcg vitamin B12 found an increased risk of overall cancer (HR 1.25, 95% CI 1.00-1.53), p=0.05) and colorectal cancer in particular (HR 1.77, 95% CI 1.08-2.90, p=0.02) (Oliai Araghi et al., 2019). A metaanalysis of 2 trials in Norway similarly reported that 800 mcg folic acid plus 400 mg vitamin B12 daily was associated with increased cancer outcomes and all-cause mortality (Ebbing et al., 2009). In contrast, dietary folate intake from food was found to be inversely associated with non-muscle-invasive bladder cancer progression in a study that also found higher recurrence for folic acid intake (Tu et al., 2018), and baseline dietary folate intake was inversely associated with prostate cancer risk in a trial that subsequently identified an increased risk of prostate cancer in the treatment arm that received 1 mg folic acid per day for 10 years (Figueiredo et al., 2009). Also relevant is the demonstration, albeit in a small study, adding dietary supplements of folic acid, vitamin B6 and vitamin B12 to a vitamin D plus calcium intervention increased biological aging (sex-adjusted odds ratio 5.26 vs vitamin D plus calcium alone) during a 1-year intervention (Obeid et al., 2018).
3.5 Polyphenols as Selective DNA Modulators
The DNAmAge clock is computed from some sites that increase and others that decrease methylation with age, so a net methylation increase would not necessarily be beneficial. Since this study targeted a healthy methylation pattern, not limited to increased methylation, the prescribed diet was rich in TET demethylase-associated nutrients (Hore, 2017) and specific plant polyphenols known to selectively regulate DNMT activity in addition to food-sourced methyl donors. It may be that these compounds assist in elevating methylation substrate and cofactor support from a risky ‘blunt instrument’ to ‘precision surgery’ on the DNA methylome by regulating where methyl groups are applied and removed.
