HEALTHY AGING: THE PROMISE, AND PERILS, OF GEROSCIENCE

ABSTRACT

Aging is a recent development in human history with nearly half of all median life expectancy gains occurring in the last two to three centuries. Aging is the strongest risk factor for medical conditions that predominately account for morbidity, mortality, and health care costs: cancer, heart disease, stroke, arthritis, and neurodegenerative disease. The geroscience hypothesis postulates that mechanisms of aging simultaneously drive multiple chronic illnesses and functional decline/disability, and intervening in the rate of aging can prevent multiple diseases. Geroscience now relies on 12 identified “hallmarks” or “pillars” of aging that involve alterations in genomic stability/repair, telomere length, epigenetics, proteostasis, macroautophagy, nutrient-sensing, mitochondrial function, cellular senescence, stem cell regeneration, intercellular communication, inflammation, and the microbiome. These pillars are mechanisms/pathways associated with aging, and evidence for causal associations is rapidly becoming more robust. Geroscience-based interventions may reduce illness burden—preserving function and independence—to a greater degree than addressing illnesses one by one. However, the pathway from promise to success is riddled with obstacles and potential pitfalls. Even success can have negative impacts on human populations and our planet that will require major shifts in society. Both the promises and perils of geroscience are likely to shape medical research and ethical debate for years to come.

INTRODUCTION: AGE, RISK OF CHRONIC DISEASE, AND FUNCTIONAL DECLINE

Aging is a relatively recent achievement in human history (Figure 1). In the Bronze Age (∼2000 BC), average life expectancy was only 35–40 years; that did not significantly improve even up through the Middle Ages (∼1000 AD). However, in the last few centuries, food security, housing, sanitation, and modern medical advances have increased life expectancy. By the late twentieth century, the median lifespan in the 10 longest-lived countries approached nearly 80 years. Epidemiologic studies of a growing population of older adults have clearly demonstrated that aging is the major risk factor for many chronic diseases. For example, tremendous increases in rates of cancer, cardiovascular disease, neurodegenerative diseases such as dementia, and osteoarthritis (1) are observed in those in advanced ages.

Fig. 1.

Human survival curves over time. Survival of a substantial population into a second half century is a relatively new event in human history. Geroscience has the potential to “rectangularize” the survival curve so that lifespan approaches maximum projected age for a majority of humans.

Aging, however, is more than disease, or even the accumulation of multiple diseases (i.e., multimorbidity). Aging occurs at the molecular, cellular, organ, and integrative levels of biology, which also increases the risk of functional decline, disability, and loss of independence. Further, measures of integrative physiology, frailty, and resilience are the best predictors of life expectancy in seniors—more so than disease occurrence. For example, an individual’s gait speed at age 65 is a powerful predictor of life expectancy as demonstrated in many different studies/populations (2). A decline of 0.1m/s in walking speed predicts a 3- or 3.5-year reduction in life expectancy for men or women, respectively. However, this relationship is not linear; regardless of walking speed, maximum lifespan tends to converge around age 100, with the very oldest recorded human living to just over 120. Thus, maximum lifespan is relatively fixed—at least thus far—but median life expectancy can be substantially extended; this has been termed “rectangularization” of the survival curve (Figure 1).

Importantly, a long lifespan doesn’t always mean prolonged health. If one accumulates disease, functional limitations, and disability despite longer years of life, this can be referred to as a reduction in health span—years of healthy, functional, and meaningful life. If one is to truly achieve optimal human aging, the goal would be to extend health span right up to the time of death so that health span approximates lifespan. This goal takes a different approach than the one used to address individual diseases. Treating one chronic disease may actually increase the risk of living to the point of accumulating another and another leading to multimorbidity and disability and, thus, separating health span from lifespan.

The Geroscience Hypothesis

If aging is the primary risk factor for many diseases, addressing the underlying mechanisms of aging could prevent multiple diseases and improve both health span and lifespan at the same time. In fact, it has been hypothesized that addressing aging to a degree first achieved by calorie restriction (CR) in rodents in the first half of the twentieth century would have a significantly greater impact on lifespan—and health span—through multi-disease prevention than curing cancer AND heart disease AND stroke AND diabetes (R. Miller, personal communication and figure at https://www.richmillerlab.com/Drugs). This is the promise of a new approach called geroscience.

In the last 15 years, advances in aging science have increasingly relied on a framework referred to as the “hallmarks” of aging (HoA) (3,4). These hallmarks or pillars have been identified by three lines of evidence: (a) they are consistently associated with aging in a time-dependent manner, (b) experimentally accentuating the hallmark/pillar accelerates aging, and (c) therapeutic interventions that mitigate the hallmark/pillar result in slowing or even reversal of age-related phenotypes. In the first description of the HoA (3), nine cellular processes were identified: DNA instability, telomere shortening, epigenetic changes, altered proteostasis, dysregulation of nutrient sensing, mitochondrial dysfunction, cellular senescence, stem cell exhaustion, and modified intercellular communication. In a 2023 update (4), the authors suggested three additional hallmarks: altered macroautophagy, chronic inflammation, and dysbiosis (changes in microbiome). In general, these hallmarks can be thought of as the “fingerprints” of aging, not clearly, yet, as the cause(s) of aging. Changes in the HoA are possibly even compensatory mechanisms to alleviate some effects of aging, but increasingly robust data have begun to link the hallmarks in a causal relationship to chronic diseases, frailty, and functional decline.

The geroscience hypothesis posits that interventions based on the HoA can enhance lifespan AND health span by simultaneously modifying the risk/development of multiple chronic diseases and physical/cognitive decline (5–7) (Figure 2). A wide array of proposed interventions has been investigated in various animal models and led to focused investigations in humans (8) as outlined in the following sections.

Fig. 2.

The geroscience hypothesis, interventions, and chronic disease/disability prevention. The hallmarks of aging (HoA) are cellular mechanisms associated with aging; enhancing these pathways speeds aging, while mitigating these pathways slows the development of age-related disease and functional decline (4). Identifying interventions informed by the HoA may (a) lead to globally slower aging and simultaneous reductions in multiple diseases at the same time or (b) suggest critical interventions for disease-specific therapy/prevention.

Intervening in the Aging Process: The “Math” of Geroscience

Addition by Subtraction

Without question, calorie restriction (CR) is the most robust intervention shown to enhance longevity and health span. First described in the 1930s, restricting calories without loss of required nutrients has now been shown to retard aging in yeast, worms, flies, rodents, and nonhuman primate models. In recent human studies, a 25% CR for two years slowed evidence of aging by 2–3% in some surrogate markers of aging but not in others (9,10). Animal model studies demonstrating lifespan enhancement generally require CR for either the entire lifespan or at least the entire time after allowing for reproductive capacity. While the benefits are undeniable, true CR to the levels required seems to be a very impractical approach in humans. Intermittent fasting has garnered attention recently and may be an alternate method (11), but it is not clear if this will have the same long-term benefits or be more acceptable long term.

Recent data suggest that all calories are not “created equal” and that restriction of specific dietary elements may have a similar impact as CR does (12). Protein restriction (PR), particularly branched-chain amino acids (BCAA) and most notably isoleucine (Ile), shows significant promise (13). Reduction in Ile alone rather than overall CR or PR has profound benefits in mouse models (reduced body mass, improved glucose tolerance, lower cancer and frailty rates) while increasing lifespan (males > > females). Minimal data in humans suggest that higher concentrations of BCAA are associated with obesity, insulin resistance, and multimorbidity, but this research is very preliminary and no causal associations can be suggested at this time (14).

Another example of adding years by “subtraction” can be found in the clearance of harmful cells, metabolites, or byproducts that accumulate with age and may lead to disease and disability. The recently approved drugs to remove amyloid beta or amyloid precursor protein to treat Alzheimer’s disease work in this fashion. The same theoretical framework underlies the targeting of senescent cells (SC) to address aging more broadly. SC accumulate with age in nearly every organ/compartment (e.g., the immune system, liver, brain, adipose). SC no longer divide nor function well in their terminally differentiated state. Instead, they typically acquire a highly inflammatory state termed the senescence-associated secretory phenotype (SASP) [reviewed in (15)]. The inflammatory cytokines produced by SC impair the function of normal, non-senescent cells in a “bystander effect;” thus, clearance of the SC could improve function not only by clearing nonfunctional inflammatory cells but also by improving the function of surrounding cells. In animal models, SC can be cleared by immunologic manipulations or significantly reduced in number by administration of “senolytic” drugs; clearance of SC very effectively addresses a number of age-related diseases and enhances physical and cognitive function in aged animal models (15). Clearance of SC in humans with senolytics requires more attention to potential toxicities, but several early trials with drugs such as dasatinib plus quercetin or fisetin show significant promise and tolerability in humans (16,17). A combination strategy in which SC are cleared and their re-accumulation inhibited by other geroscience interventions could prove most beneficial.

Addition by Substitution

One of the most novel approaches to geroscience has been heterochronic parabiosis. This technique joins the circulation of a young animal to an old animal to assess the potential for circulating factors that accelerate or mitigate aging. Remarkably, there is evidence of accelerated aging in young animals and retarded aging in older animals, and the effect is long lasting even if the animals are later separated (this effect is time dependent: the longer they share circulation, the more long lasting the effects are) [reviewed in (18)]. This has led to the identification of several “geronic” (e.g., CCL11, miR-29c-3p) and “anti-geronic” (e.g., GDF11, PF4) factors in blood; administration of these specific compounds can partially recapitulate the effects of parabiosis (19).

Another “substitution” mechanism that holds promise for mitigating aging is microbiota transfer. Emerging data demonstrate that dysbiosis is a HoA and that transfer of a “young” gut microbiota into older animals can mitigate many HoA in the brain, vascular system, and other organs (20,21) and can improve body composition (22) and physical fitness (23).

Addition by Addition

As evidence linking the hallmarks of aging to lifespan and health span has become clearer, a number of interventions by “addition” have become more realistic (i.e., adding a compound or activity to modify geroscience pathways). In animal models, this has been most thoroughly investigated in the National Institute of Aging’s Interventions Testing Program (ITP) that was initiated in 2002. The ITP is a platform designed to identify agents that extend lifespan and health span in mice (Figure 3). Investigators at any university, institute, or company/organization can recommend interventions for testing by submitting an application that is peer reviewed. If approved, the intervention is tested in male and female genetically heterogeneous (UM-HET3) mice at three separate facilities, and common endpoints are determined by strict criteria. In some instances, additional testing starting at different ages can identify variation in effects with early versus midlife administration. A full list of compounds tested or in progress and links to results can be found at https://www.nia.nih.gov/research/dab/interventions-testing-program-itp/supported-interventions.

Fig. 3.

Interventions Testing Program (ITP) application, selection, and testing workflow. Both lifespan and health span are measured, and biorepository of tissues is available. Nine compounds have been found to significantly extend lifespan to date (https://www.nia.nih.gov/research/dab/interventions-testing-program-itp/about-itp).

As of 2024, nine compounds have been found to induce significant lifespan enhancement in one or both sexes with early-life administration: aspirin (males only), rapamycin (both sexes), 17α-estradiol (males only), acarbose (both sexes, but much greater effect in males), nordihydroguaiaretic acid (NDGA, males only), Protandim (males only), canagliflozin (males only), 16-hydroxyestriol (males only), and possibly glycine (small effects; borderline significance); several of these have demonstrated midlife benefits on lifespan (e.g., rapamycin, rapamycin plus metformin, 17α-estradiol, canagliflozin). Importantly, sex differences are often apparent and can even be in the opposite direction. For example, canagliflozin, started at six months, leads to a 9% increase in lifespan in male mice (p=0.02) and no change in female mice. However, when started at 16 months, the effect is greater in males [14% increase in lifespan (p=0.004)] but results in a 6% decline (p=0.03) in female lifespan. Canagliflozin sex-specific metabolism differences may underlie this difference (24). Perhaps not surprisingly, combination therapy may hold the greatest promise. Rapamycin is the most robust life-prolonging compound tested to date and is effective in both sexes. Metformin shows no significant change in lifespan when used alone but augments the life-prolonging effect of rapamycin, again in both sexes (25).

The ITP also examines evidence of benefits in health span through functional assessments of physical and cognitive performance and examination of organs and collected tissues. In general, the compounds that enhance lifespan tend to improve health span as well, but there is some variation. Examples include NDGA does not increase lifespan in females but does demonstrate health span benefits (25) and Ile restriction that, when started at six months, does not extend life in either sex but does extend health span (13).

Exercise

One final “addition” strategy is exercise. While there is no ITP trial of exercise to demonstrate lifespan benefits, there is epidemiologic evidence linking physical activity to longevity (26), broad evidence that exercise beneficially influences many aspects of health span, and exercise has profound effects on the HoA [reviewed in (27,28)]. An exhaustive review of exercise and health is well beyond the scope of this review, but there is robust evidence that a sedentary lifestyle exacerbates the HoA (29) while exercise generally attenuates the HoA (30).

Could We Feasibly Explore the Geroscience Hypothesis in Humans?

Studies in geroscience span yeast, worms, flies, rodents, and even nonhuman primates with quantifiable success. These models have the advantage of controllable circumstances and more importantly shorter lifespans to allow survival and health span to be assessed. A number of key questions underlie the challenges of translating geroscience interventions into humans (31). Human clinical trials would require intervening in a highly heterogeneous population over many years to assess outcomes. Also, unlike animal models where disease is not detected or treated, humans receive evidence-based treatments to prevent and treat disease; to do anything else would be unethical. Thus, geroscience interventions would have to add quality years over/above standard medical care. Further, aging is not a disease; it is a condition that would require new clinical trial paradigms, regulatory oversight, biomarker/surrogate marker development, biobanking, and novel endpoints, which might include diseases but might also include geriatric syndromes or functional outcomes for which FDA does not provide an “indication.” So how might we study longevity and health span interventions in humans? The Translational Geroscience Network (TGN) is collaborating on early human trials (32). The TGN is at the forefront of translating geroscience into humans, currently has >30 compounds in various phases of clinical study, and is attempting to answer the following questions.

Can a High-Risk Group Be Identified?

A primary consideration for human trials is to pick a high-risk group—one that demonstrates risk of age-related disease, functional limitation, or disability—to allow assessments of clinical endpoints prior to lifespan. Specific patients demonstrate evidence of “accelerated aging” that might allow shorter duration endpoints or identification of biomarkers and relevant outcomes. These include patients with a history of chemotherapy, a prolonged ICU stay, a chronic infection such as HIV, and conditions of chronic stress (8). However, whether those patients truly represent aging versus disease/treatment-specific risk factors is hotly debated and somewhat limits this option (33).

A more prudent approach might be to identify groups of heterogeneous older adults that demonstrate an elevated risk for outcomes of interest—development of age-related diseases (i.e., multimorbidity), geriatric syndromes (e.g., falls, frailty), and/or functional decline (physical or cognitive). A number of risk indices have been validated to predict such outcomes, and some are very simple to calculate. For example, gait speed at age 65 years is a highly integrative measure that has been robustly validated in many observational studies to predict life expectancy in men and women (2). Other slightly more complex options include composite measures of physiology and resilience such as the Health ABC Physiological index. This simple index consisting of tertiles of systolic blood pressure, forced vital capacity, digit symbol substitution test, cystatin-C, and specific cut points for serum fasting glucose (<126; 126–142; >142) when calculated at baseline in 70- to 79-year-olds is highly predictive of mortality and incident disability (34) (Figure 4).

Fig. 4.

The Health ABC Modified Physiological index predicts mortality in older adults. This composite index, a simple, five-component score [based on tertiles of systolic blood pressure, forced vital capacity, digit symbol substitution test, cystatin-C, and specific cut points for serum fasting glucose (<126; 126–142; >142)], is normally distributed in adults ages 70–79 years and strongly predicts mortality (and incident disability, not shown) over a mean follow-up of 9.34 (±2.85) years (34). The hazard ratio for mortality after multivariate adjustment was 1.19 for each one-point increase in the index (95% CI 1.14–1.24).

Are There Biomarkers That Could Help Identify Risk and/or Be Used as Surrogate Endpoints?

The short answer to this is, not definitively, as yet. A number of “biological clocks” have been proposed [reviewed in (35)]. These “clocks” calculate the difference between chronologic and “biologic” age; however, this can vary by organ system and biochemical/cellular pathway being examined. Epigenomic, transcriptomic, proteomic, and metabolomic changes all have been proposed and have specific advantages and disadvantages. Use of these clocks to act as biomarkers for inclusion/exclusion and/or surrogate endpoints for longevity clinical outcomes for future studies requires validation.

A different approach would be to use biomarkers validated to predict disease, disability, and risk of death in cohort studies that are relatively simple and have been validated to predict important age-related disease and/or functional loss. A critical review of published biomarkers for use in geroscience trials by Justice and colleagues is available (36). This analysis of 250+ biomarkers suggested that blood-based measures of inflammation and intercellular signaling (IL-6, CRP, and TNF receptor-II), stress and mitochondrial health (GDF15), nutrient signaling (IGF-1, insulin), kidney function (cystatin C), cardiovascular health (NT-proBNP), and glucose metabolism (hemoglobin A₁C) are imperfect but best suited to act as potential biomarkers.

What Clinical Endpoints Might Be Useful in Geroscience Clinical Trials?

As noted previously, proving lifespan extension would require exceedingly long and large clinical trials, but clinical endpoints that are age related and provide some evidence of mitigating aging and extending health span would be of great value. Accumulation of multiple chronic diseases, additional vulnerability (i.e., development or worsening in frailty measures), and occurrence of geriatric syndromes particularly functional or cognitive decline are reasonable outcomes of interest (31,37). Furthermore, some provocative tests could serve as outcomes such as vaccine responses that are known to wane with age. In a small human trial, the mTOR inhibitor rapamycin—known to be an immunosuppressant and used in transplant patients to reduce the risk of rejection, but the single most potent compound shown to extend lifespan in the ITP—increased vaccine response in older humans (38). Similar vaccine response measures could be an outcome of interest in larger geroscience trials (39).

What Are the Most Promising Candidate Compounds for Human Geroscience Studies?

As noted earlier, the TGN has multiple compounds currently in various phases of human study. Nearly all of these compounds have consistently and robustly extended lifespan and health span in animal models (e.g., the ITP) or are currently FDA-approved drug classes used to treat diseases but also fit the geroscience hypothesis. Recent analysis suggests that the top candidates of the latter group are SGLT-2 inhibitors (SGLT2i), metformin, bisphosphonates, GLP-1 receptor agonists, acarbose, rapamycin, methylene blue, ACE inhibitors/ARB receptor antagonists, and dasatinib/quercetin (40). Interestingly, several of these (SGLT2i, metformin, acarbose, GLP-1 agonists, and rapamycin) interfere with nutrient sensing pathways and could be considered CR mimetics linking all the way back to the first geroscience intervention in the first half of the twentieth century.

One could take a more traditional approach to disease prevention utilizing the HoA as guidance to likely pathways that lead to age-related disease to identify clinical trial candidate compounds (Figure 2). The strongest associations of HoA with cardiovascular disease are present in microvascular dysfunction, atrial fibrillation, coronary artery disease, and heart failure with preserved ejection fraction (HFpEF) (41). The success of SGLT2i in treating the quintessential age-related CV condition, HFpEF (42,43), was predicted before clinical trials to demonstrate efficacy were developed. It was predicted that inflammation and metabolic pathways impacted by SGLT2i—consistent with the geroscience hypothesis—would be beneficial in HFpEF (44). Subsequent investigations indicate broad effects on multiple HoA by SGLT2i (45,46), perhaps acting as a CR mimetic (47). A similar geroscience-informed approach for disease-specific therapy is being employed for Alzheimer’s disease (48).

Potential Perils of Geroscience

Biological Concerns

Evolutionary biology is a powerful force. Nearly all species have evolved to live until reproduction and rearing is complete and then die off to conserve resources for the generations that follow. Humans have evolved to include intergenerational and group rearing of offspring which alters this relationship, but the tenet that aging leads to vulnerability and eventually an individual’s death makes evolutionary sense. The HoA have primarily focused on late life pathways. Interrupting those mechanisms earlier in life could lead to dire reproductive consequences. Conversely, starting too late may abrogate the benefits of geroscience. As noted earlier, vast sex differences in response to geroscience interventions indicate that it may take different approaches based on sex to achieve desired outcomes. The previously mentioned sex and timing differences and interactions for the SGLT2 inhibitor canagliflozin illustrate both issues (24).

Another concern is the need to ensure that health span is extended, not merely lifespan. Extending disabled or dependent life is not the goal, but quality life that is spread equitably to all corners and all populations has to be the aim (49).

Societal Implications

What if geroscience succeeds? The world population reached 9 billion in 2022, but the growth rate is slowing with numbers expected to peak in ∼2050–2060 and then start to fall. A less populated world inhabited by an older cohort of people is the likely future state. This has already happened in many countries where birth rates have declined either by choice or national policy (as occurred in China in the late twentieth century) and has led to inverted population pyramids (greater number of individuals at older ages than young adult ages). Many social constructs depend on a larger population of young people to support a smaller number of seniors. In 1935, when Social Security was enacted, the average age at retirement was 65 years, and the mean duration of retirement was 12 years. In 2024, the average age at retirement is now 62 years, and the mean duration of retirement is 19 years (age 81 years). This has already stretched programs like Social Security and Medicare in the United States and other social supports around the world.

The issue of extended longevity on late-life government-funded benefits was the subject of an important modeling study published in 2013 (50). The authors modeled delayed aging with reduced mortality and the probability of onset of both disability and chronic conditions (heart disease, cancer, stroke or transient ischemic attack, diabetes, chronic bronchitis and emphysema, and hypertension) by 1.25% for each year of life lived above age 50 (the inflection point when most of these diseases emerge). This reduction was phased in over 20 years. The resulting additional 2.2 years of life expectancy, most of which would be spent in good health, still resulted in a very large impact on overall spending due to the extended life expectancy and increased number of Social Security beneficiaries (Figure 5). The authors applied a proposed “fix” consisting of gradually increasing the eligibility age for Medicare from 65 to 68, and for Social Security from 67 to 68 (extending the Social Security Amendments of 1983). Obviously, greater improvements (5, 10, or even 20 years in life expectancy versus the 2.2 years modeled in this study) would have massive impacts and require much more aggressive “fixes.”

Fig. 5.

Modeling of total government (Medicare, Medicaid, Social Security) spending increases brought on by delaying aging. A 20% decline in the rate of aging over 30 years (2020 to 2050; average lifespan expectancy increases 2.2 years) modeled versus 2010 data would significantly increase government outlays due to an increased number of recipients despite projected reductions in health care costs. However, this is dramatically ameliorated by a relatively modest “eligibility fix”—a gradual increase in the Medicare eligibility age from 65 to 68 and the Social Security eligibility age from 67 to 68. Models and data from (50) (all totals are in 2010 dollars).

Though many dispute the likelihood of greatly extended lifespan being achieved in the foreseeable future, in a recent article in National Geographic (https://www.nationalgeographic.com/magazine/article/half-of-todays-5-year-olds-will-live-to-be-100), a number of longevity researchers suggested most 5-year-olds today will live to be 100 and that reality should cause a reexamination of the entire life course. In modern society, we currently live life in three stages: ∼20 years of education, ∼40–50 years of work, and then retirement for the remainder of our lives. Rather than “fix” the retirement age to make longevity affordable, Professor Andrew Scott from the London School of Business suggests that the three-stage life (education, work, retirement) “is made for a world that no longer exists and will be replaced with a multi-stage life … that is much more flexible.” Dr. Laura Carstensen from the Stanford Longevity Center suggests that “We have an incredible opportunity to redesign our lives … by spreading those additional years throughout life. Think of it more as an extended middle age, than a longer old age.” Gen Z and millennial generations often take a career “break” or employ other strategies to pursue work-life balance (https://www.nytimes.com/2019/09/17/style/generation-z-millennials-work-life-balance.html). Workplace adaptations to keep these workers engaged are required (https://imagine.jhu.edu/blog/2023/04/18/gen-z-in-the-workplace-how-should-companies-adapt/), and those changes have already begun to foster the “extended middle age” suggested by Carstensen.

A final societal consideration is that climate change is inevitably marching on: as the world becomes older, it is also becoming markedly hotter. At first glance, geroscience and climate change concerns may appear at odds, but thoughtful approaches to both can be complementary. As noted by Farrelly, both share the goals of promoting health across the lifespan, rectifying health disparities, and improving economic prospects across generations (51). The issues of climate change are particularly relevant when we examine health disparities. Strong epidemiologic evidence suggests that vulnerable populations have increased exposure to environmental and social factors that exacerbate the HoA which is likely a contributor to increased disease and disability (52). If we are able to bridge geroscience to achieve clinical benefit, societal initiatives in public health promotion, equitable access, and targeting of social determinants will be essential to ensure disparities are not exacerbated (49).

DISCUSSION

Liedtke, Durham: Very nice talk. I wanted to ask about your perspective on how best to identify research subjects or future patients who are prone to pathologic aging. There are now ways to identify neurodegenerative diseases using composite measures that combine biomarkers, clinical measures, and other metrics to produce a “clinical phenotype.” In this day and age, there are also digital biomarkers where we assess how people move, how fast they tap on a cellphone, how fast they speak on a cellphone, etc. If we apply to that the wonderful source of artificial intelligence (AI), it gives us an index that will be very powerful and we might have to do that with different populations. Is that the direction this field is taking? That is my first question, and I have a second.

High, Winston-Salem: Yes, that’s a great question. We are definitely trying to determine the appropriate surrogate markers, endpoints, or biochemical pathways that are clearly linked to mechanism as opposed to epiphenomena and that’s a big problem with aging. A lot of things happen as we age—maybe as compensatory mechanisms and not as a causative pathway or problem. I think AI has a pretty good chance of helping us with this. In every aging disease I’ve ever seen that had a successful intervention, it usually takes more than one thing, it’s not a single change. If you want to prevent falls, you have to do about eight or nine different things to really prevent injurious falls, and that’s true of almost everything in aging. I suspect there’s not going to be one pathway or one hallmark of aging. We’re going to have to figure out if we do one thing, then what compensates and takes over aging. If you hit autophagy, maybe DNA methylation is a problem. If you change the dysbiosis, maybe it’s mitochondrial powerhouse issues. I think it’s going to take something like AI to be able to look at many different possibilities at the same time. In other words, if we tweak one thing, what’s going to happen to that down the road.

Liedtke, Durham: Do you have a favorite cell lineage where there is a saying “aging is aging of the so and so cell”? In neuroscience, people say brain aging is glial aging which has some arguments in favor of it. For centenarians into the super agers, there is a saying, oh immunologically, they are like 40 years old, so then aging is t-cell aging. Do you have a favorite cell lineage where you say this cell’s aging, that’s aging, or are we looking at a combined matrix?

High, Winston-Salem: That’s a great question. The short answer is no. I don’t think a single cell is going to be the issue here and I’ll give you examples of that. Cells at the end of their life often don’t die but become senescent, become very pro-inflammatory, and pump out a lot of cytokines; this is called senescent-associated secretory phenotype (SASP). Senescent cells actually impair the function of cells surrounding them that are not senescent and are normal. If you clear those senescent cells, the rest of the function improves. You can do that in mice immunologically and in a variety of other ways so I don’t think there’s going to be one thing that does it or one cell type that does it. If there could be one organelle that does, it might be mitochondria. Mitochondria might be a key organelle.

Karp, Dallas: Thank you for the excellent talk and for showing Dr. Una Makris’s picture as one of the Williams Scholars. They are one of my star faculty members. Has anyone looked into whether people want to live longer? I know that sounds like a strange question, but in trying to prevent autoimmune diseases, suppose you ask somebody, “Would you take Rituximab if it would prevent you from getting lupus?” I actually ran a session on lupus where one of the speakers talked about rapamycin in mouse models of lupus and the effects on mitochondria and how the mice were living longer. I asked the audience, “How many of you would take rapamycin to live longer?” Nobody raised their hand, so I just wonder whether taking a medicine to gain a few more years of life is what people would do.

High, Winston-Salem: Yes, I think that’s a great question and there’s a couple of answers to that. First, if you ask older people, “Do you want to keep living?”, almost all of them say yes. Even when it’s not necessarily a quality of life that you and I want, the answer they give is yes as long as I don’t get worse in my function. It’s function that people really want as long as they are alive and then they hope to die quickly. Right? So people do want to live longer if they can live well, but they absolutely do not want to live longer if they can’t function well. They don’t want lifespan at the expense of health span.

Hochberg, Baltimore: Over 40 years ago, my late friend Jim Fries proposed compression of morbidity—not extending lifespan but keeping people healthy and functioning until they abruptly died. Tell us about the contributions of geroscience to compressing morbidity but not expanding lifespan beyond what it is now, or beyond the median lifespan in the United States.

High, Winston-Salem: Late 70s for women.

Hochberg, Baltimore: Okay in the 80s?

High, Winston-Salem: Well, at birth, it’s that low because it’s actually been flat for the last three decades. Life expectancy has not increased in the 10 longest-lived countries in the last three decades, maybe due to obesity or maybe due to other things. This was before COVID so it didn’t have an impact. There is no question that compressing health span is really what we want. In fact, in the Interventions Testing Program, most of the time when we push that lifespan to the rectangle it’s because morbidity is suppressed as well to the end and then they get cancer or something and they die. So, it does both, but there are things that improve health span to the end and don’t expand lifespan. NDGA is an example of that in the Interventions Testing Program. One of the best things you can do to actually compress health span all the way to the end of life is exercise.

Hochberg, Baltimore: Right, so if I could follow up on lifestyle factors.

High, Winston-Salem: Yes.

Hochberg, Baltimore: As opposed to pharmacologic agents.

High, Winston-Salem: Yes, no question. The things we know extend lifespan and health span now are certain diets, especially Mediterranean-like diets, and exercise. Exercise is probably the most consistent. It’s not exactly clear how much exercise you need, but a little bit more is better than whatever you’re doing now unless you’re an ultra-athlete. Just getting up and moving and not being sedentary actually confers benefit—not as much as standard exercise three times a week for 30 to 60 minutes or five days a week for 30 to 60 minutes, but there is a question of whether you should add exercise to anything you do. Should you have a baseline of exercise and then add to that? I would have said yes, but some data suggest that if we’re giving something like rapamycin or metformin at the time of exercise, we may blunt the effects of exercise. We might have to sequence it right or time it right. A lot of work has to be done yet to know how to do this.

Brock, Baltimore: Great talk. Your very first slide showed cardiovascular disease, cancer, and so forth, and you said that they were in aging a risk factor.

High, Winston-Salem: It’s the highest risk factor.

Brock, Baltimore: Highest risk factor, by far dominant; however, you stopped and didn’t really say that aging itself is a disease. In general, there has been a movement toward trying to label it as such. The World Health Organization has restricted that and is still sort of mulling over it. I submit that if we, as academics, thought that aging was a disease, things would take off, and you wouldn’t have Bryan Johnson doing all this project Blueprint stuff. We would have these randomized trials, and we would look at it just like any other disease, which it probably is.

High, Winston-Salem: Thanks. I’m not going to get into a philosophical argument about whether aging is a disease or not. I will say that we know aging is a major risk factor and these fingerprints that we can interrupt with some of these interventions do affect the development of multiple diseases. The Food and Drug Administration (FDA) doesn’t recognize aging as a disease either, so if you want to get an indication for an intervention, you have to actually show that it interferes with a disease. Ending life at the end-of-life lifespan is not a disease, right? There have been some discussions with the FDA about whether preventing mobility decline or frailty is an endpoint that we consider a condition to be avoided or a disease. So far, we haven’t successfully made that argument yet.

Palmer, New York: I want to compliment you on the presentation and remind you of a verse by Algernon Swinburne. One of the stanzas is: “From too much love of living, from hope and fear set free, we thank with brief thanksgiving whatever gods may be that no life lives forever; that dead men rise up never; that even the weariest river winds somewhere safe to sea.”

High, Winston-Salem: Hear, hear!