Peter Attia MD357 ‒ A new era of longevity science: models of aging, rapamycin trials, biological clocks, & more
CHAPTERS
- 0:00 – 9:12
Brian Kennedy's path from the Buck Institute to Singapore
Peter Attia welcomes Brian Kennedy, who describes the Buck Institute, its funding, and the interest in longevity that took off around 2017-2018. They discuss whether the resveratrol episode slowed the field, and what Kennedy's academic program in Singapore does.
- •The Buck was the first institute devoted solely to aging; Kennedy became its second CEO in 2010
- •Funding was NIH-heavy and philanthropy for aging only grew around 2017-2018
- •Kennedy suggests the resveratrol failure probably slowed investor interest
- •His Singapore program uses a pipeline from yeast, worms, flies, killifish and mice to humans
- •The program runs about thirty-five PIs on varied aging-related projects
- 9:12 – 17:48
What causes aging? Gerophysics and the limits of the hallmarks
Attia asks what causes aging, and Kennedy explains why he wants physicists and equations involved. He argues the hallmarks of aging are outputs of an interconnected network, and that inflammation keeps surfacing as a central driver in his experiments.
- •A gerophysics conference in Singapore aimed to bring modelers and aging biologists together
- •Kennedy says healthy aging is about maintaining a responsive homeostatic network
- •He thinks the idea of fixing each hallmark separately is misleading
- •Every lifespan-extending intervention reduces chronic inflammation, and his biologic-age clocks keep pointing to inflammation
- •He is unsure whether inflammation is a cause or a response, but thinks modifying it changes outcomes
- 17:48 – 23:34
mTOR biology, rapamycin, and immune resilience
Kennedy explains how his yeast screen led to mTOR and why dialing it down extends lifespan. They discuss how rapamycin ended up as an immunosuppressant and why weekly low-dose use may not suppress immunity.
- •Kennedy and Matt Kaeberlein's yeast deletion screen hit mTOR and protein translation genes
- •mTOR is a nutrient-responsive kinase that fits with calorie restriction data
- •Aging raises baseline mTOR, so interventions restore the dynamic range of youth
- •Rapamycin's immunosuppression comes at high doses and in combination with other drugs
- •Kennedy believes weekly dosing with falling trough levels is not immunosuppressive and may protect against respiratory infections
- 23:34 – 36:40
Aging as a linear decline in resilience with oscillating health
Kennedy describes Peter Fedichev's valley-and-hills model of resilience, in which linear damage lowers the hills and failure risk rises exponentially. He and Attia discuss how interventions mostly affect the oscillating component rather than the underlying slope.
- •Resilience is described as hills the ball must cross to reach failure states, which shrink with age
- •Linear accumulation of damage in clocks is overlaid with an oscillating component
- •Most interventions seem to affect the oscillation, so they work around the edges
- •Five to ten extra years of healthspan would still be a revolution
- •Treating individual diseases cannot work if many failure states exist
- 36:40 – 48:17
Can anything bend the slope? Noise, reprogramming, and the epigenome
Attia asks whether rapamycin could change the slope and then poses a sci-fi scenario about a 130-year lifespan. Kennedy speculates about lowering noise in the system and about reprogramming, and doubts the primacy of the epigenome in aging.
- •Kennedy expects rapamycin to have a modest effect size in humans rather than change the slope
- •Lowering noise or the energy in the system is one speculative route
- •Long-lived yeast mutants often carry fitness costs, so timing of interventions matters
- •Reprogramming might replace damaged cells
- •Many types of clocks can predict age, so methylation clocks show association, not proven causality
- 48:17 – 52:13
Immortality, frailty, and the case for funding late-life healthspan
Attia asks whether immortality is impossible, and Kennedy says it is near zero probability but not ruled out. They turn to the frailty of centenarians and Kennedy's plea for far more aging research funding.
- •Kennedy sees the odds of eliminating death from aging as non-zero but close to zero
- •Centenarians show a phase shift of about twenty years, not an escape from decline
- •Attia asks whether too much attention goes to immortality instead of the last decade of healthspan
- •Kennedy wants aging funded closer to cancer levels, with both translational and basic science questions
- •Government and foundation funding is described as under threat
- 52:13 – 57:24
Which chronic diseases are most inevitable, and why Alzheimer's research stalled
Attia and Kennedy compare the big four chronic diseases and agree cardiovascular and metabolic disease are the most preventable. Kennedy thinks cancer is the least modifiable and criticizes narrow mechanistic focus in Alzheimer's research.
- •Both see cardiovascular and metabolic disease as the most modifiable
- •Cancer is seen as the most inevitable, driven by mutations and immune changes
- •Kennedy thinks inflammation and immune roles in dementia were underestimated
- •Field lineages and prizes can lock research onto a few mechanisms
- •Natural aging in mice may model Alzheimer's drivers better than engineered disease models
- 57:24 – 1:08:04
Designing a human rapamycin trial in Singapore
Returning to rapamycin, Kennedy details his six-month human study and discusses what outcomes might show geroprotection. They debate commercial epigenetic clocks, pulse wave velocity, possible controls, and Kennedy's own experience with rapamycin and running.
- •The study uses 5 mg of rapamycin weekly in healthy 40-to-60-year-olds, roughly 150 to 200 participants
- •Measures include clocks, cytokines, pulse wave velocity, DEXA, strength and cognition
- •Kennedy notes he runs poorly within 24 hours of taking rapamycin but trains well days later
- •Attia cites Kaeberlein's duplicate commercial tests that disagreed with each other
- •Attia suggests a calorie-restriction arm as a control, and Kennedy says multiple studies are planned
- 1:08:04 – 1:13:41
Running a human ITP: funding and the alpha-ketoglutarate trial
Kennedy describes his lab's pipeline of human studies and their cost, and Attia highlights this as an efficient place to direct funding. They then discuss alpha-ketoglutarate, from the worm screen to the mouse results and the completed human trial.
- •Each human study costs roughly one to one and a half million US dollars
- •Attia argues translational research like this is high-leverage use of money
- •AKG came out of a worm screen and increased mouse lifespan by five to ten percent while reducing frailty
- •A time-release form is used and Kennedy discloses involvement with PDL Health
- •The placebo-controlled AKG trial with a three-month follow-up is finished and being analyzed
- 1:13:41 – 1:26:51
Urolithin A and sublingual NAD
Kennedy discusses urolithin A mouse data and unpublished targets, then his skepticism about NR and NMN. He describes how a sublingual NAD product combined with AKG appeared to improve his own exercise performance.
- •Urolithin A reduced frailty in male mice but not females, so Kennedy is repeating it
- •Attia suggests human trials use fat oxidation, zone two efficiency and muscle biopsies
- •Kennedy is skeptical that NR and NMN raise NAD much and saw little in his mice
- •The sublingual product carries 100 mg NAD plus apigenin, a CD38 inhibitor
- •Kennedy describes lower respiratory rate at a given heart rate, calling it an N-of-1 observation
- 1:26:51 – 1:34:29
Spermidine, 17-alpha estradiol, and HRT
The conversation covers spermidine's mouse lifespan effect and Kennedy's work on ovarian aging. They then discuss HRT, testosterone, and why hormone therapy uptake appears low in Asia.
- •Spermidine suppressed high-fat diet metabolic dysfunction and extended lifespan in Kennedy's mice
- •Geroprotectors tend to extend fertility in mice
- •Kennedy says the question should be whether any reason exists for a woman not to take HRT
- •Attia says testosterone fears were debunked while women still labor under the Women's Health Initiative
- •Kennedy sees very low HRT use in countries he visits
- 1:34:29 – 1:45:02
Biological clocks and a clinical chemistry clock
Attia sets the bar for clocks as outperforming chronologic age in predicting mortality. Kennedy describes his group's NHANES analysis, in which a blood-marker clock beats methylation clocks and shows undertreated preconditions raise biologic age.
- •First-generation clocks predict age, second-generation clocks predict outcomes like mortality
- •Some first-generation methylation clocks predicted mortality worse than chronologic age
- •GrimAge and PhenoAge beat chronologic age, but the clinical chemistry clock did better
- •The clock uses about fifty standard lab parameters, many from a complete blood count
- •Treated 65-year-olds had lower biologic age than untreated ones, suggesting earlier aggressive optimization
- 1:45:02 – 1:53:31
Rapamycin, metformin, GLP-1s, SGLT2 and muscle as predictors of healthspan
Attia asks which drugs are most likely to be geroprotective in healthy people. Kennedy is skeptical about metformin, rates rapamycin, GLP-1s and SGLT2 inhibitors comparably, and worries about muscle loss. They discuss why strength and fitness matter for healthspan.
- •Kennedy is skeptical of metformin in healthy people and more optimistic about rapamycin
- •He treats GLP-1 and SGLT2 effects in healthy people as unproven and worries about lean mass
- •Short-lived control mice mean a result may be only longevity normalizing
- •Kennedy thinks muscle and fitness matter but may mainly square the curve
- •Attia says exercise would be worth it even if it shortened life by six months because of healthspan
- 1:53:31 – 1:59:21
Why combining too many interventions may backfire
Kennedy supports people making their own health decisions but warns that stacking many compounds is likely to cancel effects. He describes adding one intervention at a time and measuring, and what he would do with far greater funding.
- •In mice, combinations are more likely to cancel than add
- •Kennedy compares taking twenty pills to mixing twenty paint colors
- •He tries one or two things at a time and measures before and after
- •With a hundred million dollars he would run multifactorial studies and identify compound targets
- •Singapore's small, compliant population is seen as a good place to test healthspan at scale
- 1:59:21 – 2:02:07
AI's role in aging research
Attia asks how AI could accelerate aging research. Kennedy says his lab already uses it for clocks, signal detection and drug picking, and expects AI eventually to suggest the questions to ask.
- •AI is used for clocks, signal detection and drug selection
- •One project aims to improve how large language models answer longevity questions
- •AI is not yet good at telling researchers what to ask next
- •Attia wonders how much training data experimental biology would need
- •Half of Kennedy's lab is now dry lab
- 2:02:07 – 2:12:31
Longevity clinics, safety, and closing thoughts
Kennedy names the trends that worry him, including unproven gene therapy, stem cell practitioners and growth hormone, and explains why he works with clinics that value transparency and safety. Attia reflects on how he prescribes rapamycin, and the pair close with the NHANES clock paper and combining ITP winners.
- •Kennedy is excited about synergistic combinations and new classes of interventions that alter the linear damage accumulation
- •He jokes that billionaires have become the model organisms of aging research
- •He asks clinics for transparency and safety and thinks engagement beats isolation
- •Both are cautious on growth hormone because long-term cancer safety data is lacking
- •Attia says fewer than ten percent of his patients take rapamycin and he frames it probabilistically
