A new “clock” may provide the answer to the question: “What makes us age?”
Researchers at Harvard’s Brigham and Women’s Hospital have developed a new epigenetic clock, a machine learning model designed to predict biological age from DNA structure. The clock can distinguish between genetic differences that slow or accelerate aging, predict biological age, and evaluate anti-aging interventions with greater accuracy. The results are published in Nature Aging.
“Previous clocks considered the relationship between methylation patterns and traits known to correlate with aging, but they did not tell us which factors cause the body to age faster or slower,” said corresponding author Vadim Gladyshev, senior research scientist in BWH’s Department of Genetics. “Our clock distinguishes between changes that accelerate and suppress aging, allowing us to predict biological age and evaluate the effectiveness of aging interventions.”
Aging researchers have long recognized the impact that DNA methylation (changes in gene structure that shape gene function) has on the aging process. In particular, certain regions of DNA known as CpG sites are strongly associated with aging. Lifestyle choices such as smoking and diet influence DNA methylation, but so does genetic inheritance, so individuals with similar lifestyles may age at different rates.
“Our clock will distinguish between changes that accelerate aging and those that block aging, predict biological age and assess the effectiveness of aging interventions.”
Lead Researcher Vadim Gladyshev
Existing epigenetic clocks use DNA methylation patterns to predict biological age (actual cellular age, not chronological age), but until now none have distinguished between methylation differences that cause biological aging and those that simply correlate with the process.
First author Kejun (Albert) Yin, a graduate student in the Gradysheff lab, used the large genetic dataset to perform epigenome-wide Mendelian randomization (EWMR), a technique used to randomize data and establish causal relationships between DNA structure and observable traits, on 20,509 CpG sites responsible for eight aging-related traits.
The eight age-related traits include lifespan, ultralongevity (defined as survival beyond the 90th percentile), healthspan (age at first onset of major age-related diseases), frailty index (a measure of frailty based on the accumulation of health impairments over a lifetime), self-rated health, and three broad age-related measures that incorporate family history, socioeconomic status, and other health factors.
With these traits and their associated DNA sites in mind, Ying created three models: CausAge is a general clock that predicts biological age based on causal DNA elements, while DamAge and AdaptAge include only damaging or protective changes. The researchers then analyzed blood samples from 7,036 people aged 18 to 93 from the Generation Scotland cohort, and finally trained their model on data from 2,664 people in the cohort.
Using these data, the researchers created a map pinpointing the CpG sites in humans that drive biological aging. This map will allow researchers to identify causative biomarkers of aging and evaluate how different interventions might extend lifespan or accelerate aging.
Scientists tested the clock’s validity on data collected from 4,651 people in the Framingham Heart Study and the Standard Aging Study. They found that DamAge correlated with adverse outcomes such as mortality, while AdaptAge correlated with lifespan. This suggests that age-related damage may contribute to mortality risk, and protective changes in DNA methylation may contribute to extended lifespan.
Next, the researchers tested the clock’s ability to determine biological age by reprogramming stem cells – the process of returning specialized cells, such as skin cells, to a younger, more defined state that can develop into many different types of cells in the body. When they applied the clock to the newly altered cells, DamAge decreased, indicating less age-related damage during reprogramming, but no particular pattern was seen with AdaptAge.
Finally, the team tested the clock’s performance in biological samples from patients with various chronic diseases, such as cancer and hypertension, as well as samples damaged by lifestyle factors such as smoking. In conditions associated with age-related damage, DamAge consistently increased, while AdaptAge decreased, effectively capturing protective adaptations.
“Aging is a complex process, and we still don’t know what interventions against aging actually work,” Gradyshev said. “Our findings are a step forward for aging research, making it possible to more precisely quantify biological age and evaluate the ability of novel aging interventions to extend lifespan.”
Disclosure: Kejun Ying and Vadim Gladyshev are inventors on patent applications related to the reported research.
This research is supported by the National Institute on Aging, an Impetus grant, and the Michael Antonov Foundation.
Source link
