The human body's aging clock keeps its own time
Timekeeping is variable and elusive. The human body's estimated 37 trillion cells, with their complex molecular circuitry, have their own way of keeping time. It turns out cells and the proteins they shed into the bloodstream have a lot to say about how fast we are aging compared to how old we are. What do they have to say?
Studies of so-called aging clocks, based on molecular biology rather than calendar years, were launched about 15 years ago. Today, they have become central to the surging interest in extending the human healthspan. Indeed, the convergence of aging clocks, AI-equipped wearable devices, wellness initiatives, and drugs that tamp down inflammation throughout the body raise a real prospect of our living longer, healthier lives. Biological aging clocks, it appears, can be rewound.
LMP associate professor Anna Prizment, a molecular epidemiologist working with colleagues in the department’s Advanced Research and Diagnostics Laboratory (ARDL) and collaborators at the University of Minnesota, has been leading NIH-funded studies on aging clocks since 2019. In the past two years she’s led or coauthored studies that show aging clocks based on proteomics – the study of the structure and function of proteins -- can measure
-
risk of developing chronic disease such as kidney disease, heart disease, and cancer
- benefit of social networks and support to biological aging.
Proteins and age-related pathology
Prizment and her colleagues at the University and throughout the U.S. have also been exploring cellular senescence, a core aging mechanism associated with non-dividing but still metabolically active cells. These cells secrete pro-inflammatory factors into the bloodstream. They are linked to age-related diseases, dementia in particular. Studies from the University’s Masonic Institute on the Biology of Aging and Metabolism show senescent proteins from blood plasma reflect molecular signatures from the organ or tissue, Prizment said. Because senescence--related proteins associated with aging tend to cause inflammation, they are targets of senomorphic drugs, which serve to mitigate their harmful effects, and senolytic drugs, which remove the cells that produce them.
Prizment’s entry into the aging clocks field six years ago was timely. The technical capabilities for proteomic aging clock analysis – the detection and identification of protein biomarkers in blood – were advancing rapidly. The SomaScan multiplex high-throughput proteomics platform used by ARDL enabled 5,000 specific protein measurements in a single blood plasma sample when it was introduced in 2019, a five-fold increase over five years earlier. Today, that number is 11,000. Prizment expects the number to continue to grow.
Although each of the estimated 20,000 genes in the body's cells makes a protein, a given protein can have many variants based on what’s called post-translational protein modification, amounting to many tens to hundreds of thousands of proteins. But how do we know which ones are important in aging?
That's where machine learning comes in -- to help identify the most important proteins underlying age-related health and disease. Proteomic aging clocks may be superior to epigenetic and transcriptomic clocks because proteins “are much closer to the disease,” Prizment told the journal Science in 2024 in an article entitled “A scientific showdown seeks the biological ‘clock’ that best tracks aging.” Epigenetic clocks that launched the aging clock field track chemical changes to the genome – changes that reflect an individual’s environmental and occupational exposure. But age-related pathologies arise and evolve in the body’s protein-rich tissues and organs. Declining immune system surveillance with age can accelerate the process.
Aging clocks and large longitudinal cohorts
To achieve statistical significance and validation, aging clock investigators need to analyze thousands of blood plasma samples from large, longitudinal patient cohorts. Most of Prizment’s research involves the Atherosclerosis Risk in Communities (ARIC) and the Multi-Ethnic Study of Atherosclerosis (MESA) longitudinal cohorts. The ARIC Study is a cohort of 16,000 African American and white adults from four U.S. communities monitored for more than 35 years. Launched in 2000, MESA is an ethnically diverse, community-based sample of an initial 6,814 men and women aged 45-84 years without known cardiovascular disease at baseline, including white, African American, Hispanic, and participants of Chinese descent recruited from six field centers across the United States.
Prizment and fellow investigators use machine learning models to identify the most important proteins to create aging clocks and predict health outcomes. They found the proteomic aging clocks to be statistically robust, demonstrating comparable performance in predicting health outcomes regardless of the methodological approach or study population. “It is likely that many proteins participate in the same biochemical pathways,” Prizment said. “Also, a given protein can contribute to many different pathways.” Unexpectedly, clocks based on a limited number of proteins perform equally as well as those utilizing hundreds of proteins, suggesting that these proteins are clinically informative. Prizment said that all proteomic clocks examined in her studies can predict age-related diseases many years before they are diagnosed.
As the power of aging clock technologies grows, so does the scale of the field’s ambition. Last year, a Swedish-led team of more than a hundred researchers published “A pan-disease blood atlas of the human proteome” in Science and the UK Biobank announced the launch of what it called the “world’s largest study of blood proteins.” Some 5,400 proteins samples from half a million UK Biobank participants will be analyzed. The plan is to integrate study participants’ proteomics, genomics, and imaging data and track changes in 100,000 volunteer participants over 15 years. Such deep data initiatives may call for deep learning, “an AI approach that few published clocks have employed but that can infer more complex relationships between variables than machine learning can,” according to Science.
Is it safe to assume that what investigators are measuring in blood plasma proteins accurately reflect what is actually occurring in the tissues and organs where proteins reside and function?
“That is a good question,” Prizment said. “Studies including those of our group have been using organ-specific clocks to provide some information about that. We are looking at more specific pathways that are easier to investigate because there are fewer key proteins that are overexpressed in the organ and thus could be more relevant to the disease.
Age acceleration and the exposome
Prizment and collaborators are also leading an NIH-awarded study on socio-demographic factors in aging based on the ARIC and MESA studies. They have shown that greater proteomic aging may be related to lower level of social support and smaller social networks. They are also exploring whether age acceleration is related to what has been called the “exposome.” The exposome is defined as the measure of all the exposures of an individual in a lifetime – exposures from where one lives, where one works, the air one breathes, one’s diet and lifestyle, etc. -- and how those exposures relate to health. Prizment is part of a research team organized by LMP professor and ARDL director Bharat Thyagarajan that is investigating whether different exposures affect diseases through biological aging.
As noted above, biological aging clock work is converging with precision, personalized medicine, widespread wellness and fitness programs, wearable medical devices that monitor vital bodily functions, and anti-aging agents such as senolytics, a class of drugs and natural compounds that clear senescent cells from the body. [See the U of M Medical Discovery Team’s feature article on the biology of aging with a graphic display: The Science of Senolytics.”]
In his 2025 book Super Agers: An Evidence-Based Approach to Longevity, Scripps Institute cardiologist and bestselling author Eric Topol notes that the GLP-1 receptor agonist family of drugs (Ozempic, Wegovy, Zepbound, others), now being taken by an estimated 12 percent of Americans for type-2 diabetes, obesity, a high body mass index, and other conditions, “have been shown to improve various aging clocks in old mice.” The anti-inflammatory effects of these drugs include suppression of brain inflammation “through the gut-brain axis.” That may well slow or even prevent neurodegenerative diseases like Alzheimer’s disease and other dementias. Topol notes that the terms "immunosenescence" and "inflammaging" have been coined to reflect the fact that our immune system and inflammation are highly interconnected and both are influenced by aging.
As Prizment told Science about the potential clinical utility of aging clocks, “Perhaps we could wake people up for lifestyle interventions.” Has the time arrived for a wake-up call?
