Four eras of aging research
In 1920, a French surgeon named Serge Voronoff — Russian-born, trained under Alexis Carrel (who we will meet again shortly, and not in flattering circumstances) — began grafting thin slices of chimpanzee testicle tissue onto the testicles of elderly human men. He claimed this would restore youth, strength, and vitality. By the mid-1920s, he had performed the operation hundreds of times, set up a private monkey farm in Algeria to ensure supply, and presented the technique to the International Congress of Surgeons in London to an audience of thousands. The British Medical Journal ran a broadly favorable review. A British surgical delegation traveled to Algeria to inspect his colony and came back cautiously impressed, though they noted some discomfiting irregularities in his record-keeping. Irving Berlin wrote a song about monkey glands. The Marx Brothers performed it.
By the 1940s Voronoff’s work was universally understood to be useless — the grafted tissue was simply rejected, producing nothing but scar tissue and, presumably, deep existential horror in retrospect. When he died in 1951 after a fall, very few newspapers reported it, and those that did mocked him. In the 1990s, some scientists proposed that his experiments had accidentally introduced HIV into the human population via contaminated primate tissue. That claim was later debunked, which was probably a relief to his legacy but didn’t otherwise improve it.
I bring up Voronoff not because he’s an entertaining footnote (though he is) but because he’s structurally representative of something that keeps happening in this field. The pattern runs roughly: someone announces a breakthrough intervention for aging, it works great in preliminary data, powerful institutions get involved, and then it quietly fails to pan out in rigorous testing. The interventions change every fifty years — monkey glands, yogurt bacteria, free radical scavengers, resveratrol, senolytics — but the narrative skeleton stays uncomfortably similar.
A 2008 paper by Fishman, Powell, and Wise in the Journal of Aging Studies used the sociologist Thomas Gieryn’s concept of “boundary work” to trace how biogerontology constituted itself as a discipline in the United States. The central problem they identified: from the very beginning, legitimate aging researchers had to fight against two crippling stigmas simultaneously. First, the public conflated their work with the centuries-old tradition of fraudulent rejuvenation quackery (of which Voronoff is only the most flamboyant example). Second, other scientists thought the project of “controlling” aging was either scientifically impossible or ethically forbidden — interfering with the natural order of death, the sort of hubris that gets you turned into a cautionary myth.
What makes this lens powerful is that it explains a lot of otherwise puzzling institutional choices: why mainstream gerontologists spent enormous energy publicly denouncing anti-aging supplements, why the NIH was so reluctant to create a dedicated aging institute, why the field kept insisting that it was “studying aging” rather than “trying to cure aging” even when that’s clearly what they meant. And it creates the central irony the whole history turns on: the scientists doing the most “boundary work” — the loudest denouncers of charlatanism — sometimes turned out to be generating the most spectacular failures themselves.
What follows is a periodization of biogerontology into the four major eras of modern biology, with an honest accounting of what each era actually produced, what methods it was using, and where things went sideways. I’ve tried to verify specific claims against primary or secondary sources rather than just trusting the standard heroic narratives. Some of those narratives hold up; some don’t.
Era I · 1859 – 1900
Darwin’s Shadow and the Question Nobody Knew How to Ask
The obvious thing to say about the Darwinian era is that the Origin of Species didn’t really address aging. Darwin was building a theory of species change over generations; the deterioration of individual organisms was mostly a data point he used, not a phenomenon he was trying to explain. But the Darwinian framework immediately created intellectual pressure to explain aging in terms of natural selection, and the responses to that pressure defined the first era of biogerontological theory.
The first serious response came from August Weismann in 1882, and it was more interesting than it’s usually given credit for. Weismann is often mentioned in one sentence — “he distinguished between mortal somatic cells and immortal germ cells” — but his actual contribution was more complicated and more modern-sounding than that summary implies. He proposed at least two distinct mechanisms that were often conflated: a “wear and tear” model (organisms accumulate damage the way machines do) and a “programmed death” model (organisms are selected to die in order to make room for younger, more adaptable individuals). He waffled between these and eventually substantially revised the programmed death version after criticism, but both ideas persisted independently in the literature. The fundamental question — is aging an adaptive program or a byproduct of other processes? — is one that researchers are still arguing about in 2026. Weismann’s versions of both positions are philosophically recognizable forerunners of what modern researchers mean by those terms.
The methods available in this era were: essentially none. Observation. Anatomy. Comparative notes about how long different animals live. Weismann’s theoretical work had essentially zero empirical content in the modern sense; he was doing biology-flavored natural philosophy. This wasn’t shameful by the standards of the time, but it meant the first era of biogerontology produced frameworks rather than findings.
Élie Metchnikoff is conventionally credited with coining the term “gerontology” around 1903 and publishing the first systematic scientific work on aging in The Prolongation of Life: Optimistic Studies (1907). He proposed that aging was caused primarily by the toxicity of intestinal bacteria and overactive macrophages — a hypothesis he called autointoxication. His recommended intervention was fermented milk, which he believed suppressed harmful gut microbes. He had been observing long-lived Bulgarian peasants who consumed yogurt. He won the Nobel Prize in 1908 (for phagocytosis, not for aging research), which gave his ideas a credibility they perhaps didn’t quite deserve on their own merits.
This is the moment to note something the standard histories gloss over: Metchnikoff’s autointoxication theory was wrong. Not subtly wrong or prematurely wrong — just wrong, in ways that were fairly clear by the 1920s. The gut microbiome hypothesis is now back in the longevity conversation in a more sophisticated form, but the original Metchnikoff version was essentially folk medicine dressed in laboratory language. He deserves credit for founding the field and for the phagocytosis work, but treating the yogurt hypothesis as a serious scientific contribution requires some charitable reinterpretation.
The era also produced Charles Sedgwick Minot’s 1907 The Problem of Age, Growth, and Death, which tried to track the rate of growth against the rate of senescence and drew the conclusion that aging begins at fertilization. This was mostly wrong too, but it was rigorously wrong — it used actual quantitative methods, engaged with real data — and represents a real step toward biogerontology as an empirical discipline rather than theoretical speculation.
What did Era I actually establish? Conceptually: the distinction between germline and soma, the evolutionary framework for thinking about why aging exists, and the question of whether aging is “programmed” or the result of accumulating damage. Empirically: almost nothing. Methods: observation, comparative biology, some histology. The theoretical landscape at the end of the era is strikingly recognizable from the modern vantage point, which is either impressive (they identified the right questions) or depressing (we spent another century on them).
Era II · 1900 – 1953
The Charlatanism Era, and the One Real Result
The rediscovery of Mendel’s work in 1900 — by de Vries, Correns, and von Tschermak, independently and nearly simultaneously — transformed biology, but it took a long time for the transformation to reach gerontology. The immediate effect on aging research was almost nil. Genetics in the early twentieth century was focused on inheritance, speciation, and development. Aging as a topic was considered peripheral, difficult, and professionally risky — which is when you get the charlatans moving in.
The rejuvenation medicine of the 1920s and 1930s is genuinely bizarre by modern standards, but it was operating in a context where endocrinology was brand new, sex hormones hadn’t been isolated yet, and the placebo effect was not well understood. Eugen Steinach, a Viennese physiologist with genuine research credentials, proposed in 1918 that vasectomy could rejuvenate aging men by redirecting hormonal energy previously devoted to sperm production. He performed the operation on humans with Robert Lichtenstern. William Butler Yeats had a “Steinach operation” in 1934 and reported a renewed creative vitality. Sigmund Freud, who had throat cancer and was in considerable pain, reportedly had one too. The endocrine logic was based on a misunderstanding of how testosterone production works, but the idea wasn’t completely insane given what was known at the time — it was just very, very wrong.
Voronoff is the more famous case, but he was downstream of a broader cultural moment in which the glands of young animals seemed to hold the key to vitality and the mechanisms of endocrinology seemed just tantalizingly within reach of therapeutic use. The 1927 British delegation that traveled to Algeria to inspect his work found that his animal records were poorly controlled, that his “prize specimens” had been pre-selected to look impressive, and that there was no reliable evidence of grafted tissue surviving and functioning. The scientific community largely turned on him by the late 1930s. But note what the turn required: not theoretical refutation, but the slow accumulation of failure to replicate and failure to control. This is a pattern.
The most important event of this era for subsequent science actually happened at the Rockefeller Institute, where Alexis Carrel — Nobel laureate, genuine surgical pioneer — claimed in 1912 to have established a culture of chicken heart cells that he kept alive and dividing for the next thirty-four years. Since chickens live for about ten years, this was taken as evidence that cells were intrinsically immortal; aging was therefore something that happened to cells from outside, not something inherent to cells themselves. The culture was maintained until Carrel’s death in 1944 and then ceremoniously terminated.
The problem: no one could replicate it. Carrel’s methods were closely held and the culture was maintained with considerable ceremony and mystique. The most plausible explanation — eventually settled on by Jan Witkowski in a 1980 Medical History paper — is that fresh living cells were inadvertently being introduced whenever nutrient extract prepared from chick embryos was added to the culture, and the confusion was either innocent or not innocent depending on how generously you read Carrel. (Hayflick, who we will get to, was diplomatically charitable in public about this; in private accounts he was less so.) What’s historically significant is that Carrel’s claim dominated the field for nearly fifty years, and it was completely wrong. Biogerontology’s foundational assumption about cell biology was a mistake maintained by the authority of a Nobel Prize winner.
The one genuine empirical result from this era is Clive McCay’s caloric restriction work, published in 1935. McCay, working at Cornell, showed that rats fed a severely restricted diet — roughly 30% fewer calories than controls, while maintaining adequate nutrition — lived substantially longer than ad libitum-fed controls. Mean and maximum lifespan both increased. This is reproducible. It’s been reproduced in yeast, worms, flies, mice, and rats across dozens of labs. Whether it translates to meaningful effects in primates with normal lifespans is genuinely contested (more on this in Era IV), but as a biological phenomenon in rodents it is the most robust result in the entire history of the field. And yet it had almost no theoretical framework to explain it; McCay himself didn’t know why it worked. It would take another fifty years for the relevant mechanistic explanation to arrive.
The Gerontological Society of America was founded in 1945. The NIA’s predecessor programs begin organizing. The era ends with Medawar’s An Unsolved Problem of Biology, his 1951 University College London inaugural lecture published in 1952, which provides the first rigorous evolutionary argument for why aging exists: late-acting deleterious mutations are invisible to natural selection because they only express after reproduction has ceased. This is mutation accumulation theory, and it’s the first theoretical contribution of Era II that doesn’t require substantial qualification to present as correct.
The overall shape of Era II: one real result (caloric restriction), one foundational error (Carrel’s immortal cells), a cultural spectacle of rejuvenation medicine (Voronoff, Steinach), and one late genuine theoretical breakthrough (Medawar). The methods available advanced somewhat — tissue culture arrived, animal experiments became more systematic, endocrinology emerged as a real discipline — but the dominant mode of inquiry was still physiological and observational. The molecular era hadn’t begun.
Era III · 1953 – 2003
The Molecular Revolution and the Institutionalization of Biogerontology
Watson and Crick in 1953 didn’t immediately transform aging research, which continued on a fairly separate track from molecular biology for about a decade. But they created the methodological conditions — a mechanistic understanding of heredity and gene expression — that would eventually make the field’s breakthrough experiments possible. And in the meantime, the theoretical side of biogerontology had an extraordinarily productive decade.
Medawar (1952) gave us mutation accumulation: late-acting bad genes escape selection. George Williams in 1957 published “Pleiotropy, Natural Selection, and the Evolution of Senescence” in Evolution — a paper that built on Medawar and introduced antagonistic pleiotropy, which is probably the most important theoretical idea in the field’s history. Williams noticed something implicit in Medawar that Medawar himself had almost missed: if natural selection is weak against late-acting effects, then it might actually favor genes that help you early in life even if they harm you later. A gene that massively accelerates bone growth in juveniles might be wonderful for reproduction and terrible for your arteries at sixty. Selection doesn’t care; it only sees the early part. This explains something crucial: why aging seems to involve biological programs rather than just random decay. The programs aren’t for aging — they’re for early life success — but they have aging as a side effect. William Hamilton formalized both Medawar’s and Williams’s ideas mathematically in 1966, showing that the “force of natural selection” declines with age in a calculable way. Kirkwood added the “disposable soma” framing in 1977: organisms have limited energy to allocate between reproduction and somatic maintenance, and the evolutionary optimum generally favors undermaintaining the soma.
This body of theory — Medawar, Williams, Hamilton, Kirkwood — is the genuine intellectual foundation of modern biogerontology, and it is actually pretty solid. The specific mechanisms Williams imagined for antagonistic pleiotropy haven’t all been confirmed, but the basic framework has held up reasonably well, and the prediction that longevity genes should show evidence of early-life costs has received support. It’s worth noting, though, that this theoretical framework explains why aging is evolutionarily stable, not how to intervene in it. The practical payoffs would have to come from somewhere else.
They came first from Leonard Hayflick, and they came attached to one of the better institutional scandals in the history of American science.
In 1961, Hayflick — working at the Wistar Institute in Philadelphia, where his job was developing cell lines for vaccine research — noticed that human fetal diploid cells stopped dividing after forty to sixty passages. They didn’t die immediately; they entered a quiescent state he called senescence. He confirmed this in the clearest way possible: he mixed early-passage female cells with late-passage male cells and showed that it was the male cells — distinguished by their Y chromosome — that stopped dividing first, regardless of which culture they were in. The replicative limit was intrinsic to the cell, not an artifact of the culture conditions. Carrel was wrong. Normal cells were mortal.
He and his co-author Paul Moorhead submitted this paper to the Journal of Experimental Medicine in part because that journal had published most of Carrel’s work, and they wanted to confront the dogma directly. The journal rejected it. Reviewing what amounts to the most important single paper in the history of biogerontology, the editors said — and I’m paraphrasing only slightly — that the observation of cellular mortality must reflect an artifact of cell culture technique, because it was established that cells were immortal. This is what foundational errors do: they immunize themselves against refutation by redefining anomalous data as error. The paper was eventually published in Experimental Cell Research in 1961, Macfarlane Burnet named the phenomenon “the Hayflick limit” in 1974, and the rest is history. George Gey, the person who established the HeLa cell line, had warned Hayflick before publication: “Lenny, you’re going to get yourself in a lot of trouble if you publish this.”
Then the real trouble started. The WI-38 cell line Hayflick had developed at Wistar — named after the institute and one of his daughters — turned out to be extraordinarily valuable for vaccine manufacturing, because it was clean, well-characterized, and virus-free. By the late 1960s it was being used to make vaccines for rubella, polio, measles, mumps, rabies, and hepatitis A. Billions of doses. When Hayflick moved to Stanford in 1968, he brought ampoules of WI-38 with him, stored in a liquid nitrogen container strapped to the back seat of his Pontiac, two of his children sitting next to it. He distributed the cells to pharmaceutical companies and, when NIH funding for distribution dried up, began charging fees for shipping and handling via a company he set up.
In 1976, the NIH released two investigative reports accusing Hayflick of theft of government property. The cells, they argued, belonged to the federal government because the research had been federally funded. Hayflick sued to prevent release of the reports, arguing Privacy Act violations. He lost the lawsuit. The NIH then physically confiscated his WI-38 cultures while he was at a conference. Six years of litigation followed. The settlement, in 1982 — signed by 85 supporting scientists in Science — essentially vindicated Hayflick on the theft charges and established that the funds in his escrow account were legitimate operating costs. His attorneys took the money for the litigation fees. By this point the Bayh-Dole Act had passed (1980), which allowed researchers to hold title to inventions made with federal funds, making the whole dispute somewhat moot in principle.
What the WI-38 story illustrates is something the standard “boundary work” narrative tends to obscure: the NIA and NIH weren’t simply good-faith arbiters of scientific legitimacy. They were also bureaucratic entities with their own institutional interests, capable of misusing their authority in ways that were scientifically damaging. The person who definitively overturned the field’s foundational error got accused of theft by the federal government for trying to make a cell line available for vaccine production.
The NIA itself was founded in 1974, with Robert Butler — a psychiatrist, gerontologist, and genuine reformer, best known for coining the term “ageism” — as its first director. Butler was passionate, eloquent, and effective at getting congressional support. The achievement of creating a dedicated aging institute at all was real; the Gerontological Society had been pushing for one since the 1960s, and the opposition from other NIH institutes who didn’t want to cede territory was significant. But almost immediately, Congress designated the NIA as the federal lead agency on Alzheimer’s disease, which permanently fractured its mandate. In practice, a very large proportion of NIA funding went and continues to go to Alzheimer’s research — which is valuable but which is quite different from research on the biology of aging per se. When you read accounts of what the NIA “does for biogerontology,” it’s worth asking what the Alzheimer’s funding breakdown actually looks like in any given year, and being somewhat skeptical of the official narrative that these missions are naturally complementary.
The molecular era also produced several other major theoretical contributions. Denham Harman published his free radical theory of aging in 1956 — his original 1955 report was done at the Donner Laboratory under Atomic Energy Commission funding, which gives it an appropriately Cold War flavor — proposing that the accumulation of oxidative damage from reactive oxygen species is the primary driver of aging. He extended this in 1972 to specifically implicate mitochondria as both the source and primary target of free radical damage. This was enormously influential. It spawned an antioxidant supplement industry that is probably the most commercially successful application of any biogerontological theory ever — and which, unfortunately, has not reliably extended lifespan in controlled human trials, and in some cases (high-dose beta-carotene, vitamin E in some contexts) appears to have increased mortality. The free radical theory as originally formulated is probably not wrong exactly, but it was simpler than reality turned out to be. Antioxidants and ROS have complex signaling roles; oxidative stress is not just damage, it’s also information.
Leslie Orgel contributed the “error catastrophe” theory in 1963 — errors in protein synthesis lead to faulty enzymes that make more errors, in a positive feedback loop — which turned out to be largely wrong empirically but was intellectually serious and prompted real experimental work. The decade from the late 1970s to the late 1980s produced telomere biology: Elizabeth Blackburn, Carol Greider, and Jack Szostak worked out the structure of telomeres and discovered telomerase, earning the 2009 Nobel Prize. This connected the Hayflick limit to a specific molecular mechanism: cells stop dividing because their telomeres shorten with each replication. This is beautiful science. The translational implications — can you prevent aging by keeping telomeres long? — have turned out to be more complicated, because tumors use telomerase to achieve immortality and simply “more telomerase” as a strategy has obvious failure modes.
The moment that actually transformed the field’s ambition came in 1993, and it came from a worm. Cynthia Kenyon at UCSF discovered that a single point mutation in a gene called daf-2 doubled the lifespan of Caenorhabditis elegans. Not average lifespan — maximum lifespan. The animals were active, fertile, apparently healthy; they just lived twice as long. “Back then, people thought that if you studied aging, it was because you probably weren’t a very good scientist,” Kenyon said later. She found a second gene — daf-16 — that was required for the longevity effect: daf-2 works by suppressing daf-16, which encodes a transcription factor that activates protective and longevity-promoting genes when daf-2 is out of the way. Turn off the insulin/IGF-1 signaling pathway, and the worm’s default program shifts from reproduction-oriented short life to stress-resistant long life. The genes were conserved in mammals, including humans.
It’s worth noting that Michael Klass had published evidence for lifespan-extending mutations in C. elegans in 1983, and Tom Johnson had characterized the age-1 gene — which turned out to encode the PI3K upstream of daf-2 — in 1988. Kenyon’s 1993 paper was the clarification and the characterization, not quite the origin. But it was the moment the field internalized that aging was genetically controlled in a deep way, not just subject to gene influences at the margin. A single switch could double your life. That’s a different kind of claim than “antioxidants might help a little.”
The era ends, roughly, with an interesting exercise in “boundary work.” In 2002, Olshansky and fifty other prominent scientists published “No Truth to the Fountain of Youth” in Scientific American, explicitly stating that “no intervention has yet been proven to slow, stop, or reverse human aging, and those who say otherwise are lying.” This was directed at the anti-aging supplement industry and the various human growth hormone clinics that had proliferated in the late 1990s. It was appropriate and accurate. But read it now and it has an ironic quality, because within a decade the same scientific mainstream would be generating its own hype cycle of roughly comparable proportions.
Era IV · 2003 – present
The Genomic Era, or: What Happens When There’s Money
The Human Genome Project completion in 2003 didn’t directly transform biogerontology, but it created the infrastructure — cheap sequencing, genome-wide tools, the ability to ask “which genes?” across the full genome rather than one gene at a time — that made the subsequent decade’s discoveries possible. The field’s transformation in this era is as much financial as it is methodological, which is one of the things that makes it hard to evaluate clearly.
The sirtuin story is the defining scientific narrative of the early genomic era, and it’s an object lesson in how scientific hype compounds. Leonard Guarente’s lab at MIT had been working on sirtuins in yeast since the 1990s; sirtuins are a family of NAD-dependent deacetylases that regulate gene silencing and DNA repair, and overexpression of Sir2 in yeast appeared to extend lifespan. David Sinclair was a postdoc in Guarente’s lab who left for Harvard in 1999, and the two subsequently had a public scientific dispute in 2002 at Cold Spring Harbor about how Sir2 actually works — a dispute that set off a years-long rivalry. In 2003, Sinclair published a Nature paper showing that resveratrol — a compound found in red wine — activated sirtuins in yeast and extended yeast lifespan by 70%.
Resveratrol is found in red wine. This information was widely reported. Red wine sales increased. Sinclair appeared on 60 Minutes and said resveratrol was “as close to a miraculous molecule as you can find.” In 2006, his lab published a Nature paper showing that resveratrol extended the lifespan of obese mice — mice fed a 60% fat diet that was killing them — by activating SIRT1. GlaxoSmithKline bought Sirtris Pharmaceuticals, a company Sinclair co-founded, in 2008 for $720 million. In 2009, Pfizer researchers published a paper noting that the resveratrol-sirtuin activation experiments used a fluorescent substrate for the assay, and that resveratrol was binding the fluorescent tag rather than actually activating SIRT1 on its own. Multiple drug candidates in the Sirtris pipeline failed in human trials, including SRT501, which was discontinued in 2010 due to safety concerns and non-reproducibility. GSK shut Sirtris down entirely in 2013.
The biochemist Charles Brenner — who had also worked on NAD metabolism and had his own professional disagreements with Sinclair — spent several years on social media in the late 2010s documenting these failures and criticizing what he saw as systematic misrepresentation in Sinclair’s communications. “The global interest in sirtuins and sirtuin activators was such that companies — most notably GSK — spent many billions of dollars trying to get a positive result and could not,” Brenner wrote, “because the so-called sirtuin activators do not activate sirtuins.” Sinclair has continued to defend the broader sirtuin/NAD framework while acknowledging that specific compounds and claims needed revision. His 2019 book Lifespan hit the New York Times bestseller list and contained claims that multiple researchers found overstated. This debate is ongoing and somewhat tribalized, which makes it hard to evaluate from outside.
What’s fair to say: sirtuins are genuinely interesting proteins with real roles in metabolism and stress response. NAD+ levels do decline with age and this does appear to matter. But the specific therapeutic claim — take resveratrol, activate your sirtuins, live longer — failed the regulatory test, the clinical test, and apparently the basic biochemical test. The resveratrol episode is the most expensive example in history of the pattern I described at the opening: spectacular preliminary results, institutional hype, and then quiet failure in rigorous testing. The fact that it happened in a lab at Harvard rather than in a French surgeon’s Algerian monkey colony doesn’t change the structure.
Meanwhile, genuinely important science was happening. In 2009, the NIA’s Interventions Testing Program published results showing that rapamycin — an mTOR inhibitor developed as an immunosuppressant for organ transplant patients — extended the lifespan of genetically heterogeneous mice by 9% in males and 14% in females. The remarkable thing was that the treatment started when the mice were 20 months old, the equivalent of roughly sixty years in humans. Later ITP work showed up to 25% lifespan extension with earlier administration. Rapamycin works by inhibiting mTOR (mechanistic target of rapamycin), which is a nutrient-sensing kinase — essentially, rapamycin tells the cell it’s in a nutrient-poor environment and triggers the same kinds of stress responses and cleanup programs that caloric restriction activates. This is the most reproducible pharmacological lifespan-extending result ever obtained in a mammal, achieved by an independent multi-site trial, and it doesn’t get nearly the attention the resveratrol story did. The resveratrol failure was on 60 Minutes; the rapamycin result was in Nature but didn’t make the evening news. This might tell us something about what we’re optimizing for when we cover aging research.
Rapamycin’s translational outlook is genuinely uncertain, and this uncertainty is legitimate rather than manufactured by skeptics. mTOR inhibition suppresses the immune system — it’s an immunosuppressant, remember — which is why it works for organ transplantation, and which raises obvious concerns about chronic use in healthy people. The doses that extend mouse lifespan are lower than the doses used clinically, and some researchers think the immune effects at longevity doses might be acceptable or even beneficial (there’s evidence that some mTOR inhibition improves vaccine responses in the elderly). But we genuinely don’t know yet, and anyone who tells you confidently that rapamycin will extend healthy human lifespan is outrunning the evidence.
The two most intellectually important contributions of this era are probably the Lopez-Otin “Hallmarks of Aging” framework and Horvath’s epigenetic clock, both published in 2013 — a remarkable year for the field’s conceptual infrastructure.
Carlos Lopez-Otin, Maria Blasco, Linda Partridge, Manuel Serrano, and Guido Kroemer published “The Hallmarks of Aging” in Cell in 2013, identifying nine cellular and molecular hallmarks — genomic instability, telomere attrition, epigenetic alterations, loss of proteostasis, deregulated nutrient sensing, mitochondrial dysfunction, cellular senescence, stem cell exhaustion, and altered intercellular communication — and organizing them into primary (causally upstream), antagonistic (damage responses that themselves cause harm at high levels), and integrative (downstream consequences) categories. The paper is now one of the most-cited in biology. A 2023 update expanded the list to twelve hallmarks.
David Gems published a pointed critique of the Hallmarks framework in 2022, arguing that the nine hallmarks are largely descriptive features of aged organisms rather than true causal mechanisms, and that organizing them into three tiers gives an appearance of causal structure that the evidence doesn’t really support. He’s probably partly right. “Hallmarks” frameworks are organizational tools rather than mechanistic theories; they’re useful for structuring research programs and grant applications, but they don’t themselves explain aging, they catalog it. The fact that the framework keeps expanding — twelve hallmarks now, possibly more later — might suggest it’s accumulating descriptive accuracy at the cost of explanatory depth. That said, “useful lie” is too strong; the framework has genuinely organized research effort productively.
Steve Horvath’s 2013 paper in Genome Biology — a single-author paper, which is now almost unheard of in major biology journals — reported the development of a DNA methylation-based age estimator, the “Horvath clock,” that applies to virtually all nucleated human cells and tissues and can estimate chronological age with a median error of about three years. The paper used 8,000 samples from 82 datasets spanning 51 tissue types. The clock works because DNA methylation patterns change predictably with age across tissues in a way that is consistent enough to be machine-learned. The really striking observation Horvath made was that induced pluripotent stem cells — ordinary adult cells reprogrammed to an embryonic-like state by Yamanaka factors — have a DNAm age close to zero. The epigenetic marks of aging are apparently erased by reprogramming.
This observation has driven an enormous amount of subsequent work on “partial reprogramming” — using Yamanaka factors transiently to rejuvenate cells without converting them fully to a pluripotent state (which would cause cancer). Juan Carlos Izpisua Belmonte’s lab at the Salk Institute published foundational results in mice in 2016. Sinclair’s lab reported in 2023 that partial reprogramming could reverse vision loss in aged mice. Altos Labs was founded in 2021 with $3 billion in startup funding — the largest startup round in biotech history — explicitly to pursue reprogramming biology, with Horvath as a principal investigator. Whether partial reprogramming will translate to humans, whether we can do it safely, and whether the epigenetic age of a cell is causally important to aging or just correlated with it — these are open questions. But the phenomenon itself seems real.
The senescence story is the other major narrative of this era, and it’s more nuanced than the press coverage usually implies. Cellular senescence — the state Hayflick identified in 1961 — turns out not to be just an end state; senescent cells actively secrete inflammatory signals (the SASP, senescence-associated secretory phenotype) that damage surrounding tissue. Jan van Deursen’s 2011 paper in Nature showed that genetic elimination of p16-positive senescent cells in progeroid mice dramatically reduced multiple features of accelerated aging. This was a genetic proof of concept. Translating it to pharmacology required identifying drugs that selectively kill senescent cells (“senolytics”), and a 2015 paper from James Kirkland’s group at the Mayo Clinic identified the combination of dasatinib (a leukemia drug) and quercetin (a plant compound) as having this property.
Dasatinib and quercetin are now in numerous human trials. Early results in conditions like idiopathic pulmonary fibrosis and diabetic kidney disease have shown some promise. A phase 1 trial in early Alzheimer’s disease published in 2023 showed that the drug combination could cross the blood-brain barrier and reduce neuroinflammatory markers. Whether this translates to actual cognitive benefit, and whether the combination is safe for long-term use in otherwise healthy people, remains genuinely uncertain. Unity Biotechnology, which licensed senolytic technology and went public in 2018 on the strength of it, has had a difficult few years in the clinic.
And then there’s the money, which is historically unprecedented and which makes the current era both the most exciting and the hardest to evaluate. Calico (Google/Alphabet, 2013), Unity Biotechnology, Alkahest, Rejuvenate Bio, Retro Biosciences, BioAge Labs, Altos Labs ($3B, 2022), and a dozen others have raised billions of dollars for aging biology. This has funded genuinely important science. It has also created incentive structures that reward hype, premature translation, and narrative coherence over mechanistic understanding. The field now contains people who became wealthy on the resveratrol narrative doing the same structural thing with the next narrative. Distinguishing the signal from the noise requires more cynicism than most science reporters apply, because the scientists themselves are often financially entangled with the claims they’re making.
What the current field actually has: one robust pharmacological intervention in mice (rapamycin), a plausible conceptual framework for cellular senescence with early but promising human data, the epigenetic clock as a biomarker that might eventually let us measure whether interventions actually work, and several dozen companies spending venture capital on things that mostly won’t work. What the field claims to have: the imminent capability to meaningfully extend healthy human lifespan. Whether those two things will converge is the central empirical question of the next twenty years.
The thread that runs through all four eras is the “boundary problem” — both in the sociological sense Fishman and colleagues identified (boundary work against charlatanism) and in the more literal scientific sense: the difficulty of establishing what actually causes aging versus what merely correlates with it, what interventions actually work versus what produces impressive preliminary data, and what “working” even means in a species with an eighty-year lifespan and almost unlimited variation in environment and genetics.
The charlatanism problem has never gone away; it’s evolved. In Era I it was Metchnikoff’s yogurt. In Era II it was monkey glands. In Era III it was the antioxidant supplement industry (which still sells roughly $3 billion annually in the US on the strength of Harman’s 1956 paper). In Era IV it was resveratrol, and it may yet be NAD+ precursors or metformin or any of several other compounds currently being marketed to healthy people for longevity on the strength of animal data and mechanism plausibility. The boundary work — the mainstream researchers publicly distinguishing themselves from the commercial hype — keeps happening, but the mainstream researchers keep producing their own hype cycles. This isn’t because they’re dishonest people; it’s because the incentive structures are very strong, the questions are very hard, and the preliminary results in model organisms are genuinely exciting in ways that make the cautionary qualifications feel like unnecessary pessimism until, some years later, they turn out to be necessary pessimism.
The field has also, genuinely, made progress. The evolutionary theory of why aging exists is worked out and broadly correct. The molecular pathways involved — insulin/IGF-1 signaling, mTOR, sirtuins and NAD metabolism, telomere biology, senescence — are real and are druggable. The epigenetic clock gives us a measurement tool that could, in principle, allow us to test interventions much faster than we could by waiting for humans to die. Rapamycin extends lifespan in mice in a robust and reproducible way that no intervention before it achieved. Partial reprogramming reverses some aging phenotypes in rodents. These aren’t nothing.
The fair assessment at the end of four eras is probably something like: biogerontology has finally, in the last twenty years, become the kind of science it was pretending to be for the first hundred. It has real molecular targets, real biomarkers, real interventions with real effect sizes in organisms more complex than worms. Whether it becomes the kind of science that materially extends healthy human life in the near term — as opposed to producing important basic science while the longevity claims perennially fall twenty years behind schedule — is a question the next era will answer. I’m genuinely uncertain. I’m also somewhat aware that every generation of biogerontologists has thought it was the one finally turning the corner, and most of them have been wrong about the timelines even when they were right about the mechanisms.
The monkey glands are still out there. They’re just wearing better clothes.
[1] The Fishman et al. paper is: Fishman JR, Powell J, Wise D. “Anti-aging science: The emergence, maintenance, and enhancement of a discipline.” Journal of Aging Studies 22 (2008). The Gieryn “boundary work” concept is from: Gieryn TF. “Boundary-work and the demarcation of science from non-science.” American Sociological Review 48 (1983). These are the actual historiographical frameworks being applied here, not my invention.
[2] Weismann’s position on programmed aging is genuinely more complicated than the textbook version. He proposed several distinct and somewhat inconsistent ideas between 1882 and 1892. The “disposable soma” framing commonly attributed to him is cleaner than what he actually wrote, and Kirkwood’s 1977 version is the rigorous form of the idea.
[3] Whether Carrel knew his immortal cells were being contaminated is disputed. Jan Witkowski’s 1980 and 1985 papers in Medical History and Trends in Biochemical Sciences argue he didn’t, and that lab assistants may have been inadvertently introducing embryonic cells when adding nutrient extract. Hayflick was more skeptical in private. Either way, the culture of secrecy around the experiment was Carrel’s choice, and it prevented refutation for decades.
[4] The 2002 “no fountain of youth” paper is: Olshansky SJ, Hayflick L, Carnes BA, et al. “No truth to the fountain of youth.” Scientific American 286(6):92-95 (2002). Hayflick himself signed this, which creates an interesting dynamic given his complex relationship with NIH, which co-signed a related statement. The supplement industry was the explicit target; rapamycin hadn’t arrived yet.
[5] The caloric restriction story in primates is messier than mouse data suggests. The Wisconsin National Primate Research Center rhesus monkey study showed significant survival and health benefits; the NIA-funded National Primate Research Center study showed more equivocal results. The 2012 reconciliation paper suggested the NIA control animals may have been inadvertently calorically restricted at baseline, which would explain the weaker effect. This remains an active methodological debate.
[6] The ITP rapamycin result is: Harrison DE et al. “Rapamycin fed late in life extends lifespan in genetically heterogeneous mice.” Nature 460 (2009). The multi-site testing program design — the same experiment at three independent labs simultaneously, with genetically diverse mice — is methodologically much more rigorous than most lifespan studies in rodents. This is part of why the result deserves more attention than it typically gets.
[7] David Gems’s critique of the Hallmarks framework is: Gems D, de Magalhães JP. “The hoverfly and the wasp: A critique of the hallmarks of aging as a paradigm.” Ageing Research Reviews (2022). This is not a fringe paper; Gems is one of the senior figures in the field. That it hasn’t gotten more mainstream attention reflects the same dynamics that have always operated here.