Is Ageing Research Worth It — And Will the Drugs Actually Work?

The Question Behind the Hype

Altos Labs launched with $3 billion. Sam Altman invested $180 million into Retro Biosciences. Jeff Bezos backs Unity Biotechnology. The collective wager of Silicon Valley’s wealthiest is that ageing itself is not an inevitable fact of biology but a treatable condition — and that the people who fund the right research will live to see the proof.

Is this warranted? Or is it the most expensive hype cycle in the history of medicine?

The honest answer is: it depends on which specific question you ask. And the most important question — one that the longevity industry rarely centres — is not “can we extend lifespan?” but “can we compress morbidity?”

---

Measuring the Problem: Pace of Ageing Studies

Before asking whether drugs will work, we should understand what we are trying to fix. A 2024 study using the DunedinPACE epigenetic clock provides some of the clearest human-level data available.

DunedinPACE is an algorithmic score derived from DNA methylation patterns across a curated set of CpG sites. Unlike static epigenetic clocks (such as Horvath’s original clock or GrimAge) that estimate biological age at a single point, DunedinPACE measures the rate at which biological ageing is occurring — how many biological years a person is ageing per calendar year. A score of 1.0 means you are ageing at the population average; 1.2 means you are ageing 20% faster than your peers.

The “Pace of Aging” analysis applied this tool to longitudinal cohort data from two large population studies:

• HRS (Health and Retirement Study, USA): a nationally representative panel study of Americans over 50, with multi-wave follow-up over more than a decade
• ELSA (English Longitudinal Study of Ageing, UK): a comparable cohort design tracking health, economic circumstances, and wellbeing in English adults aged 50+

Key findings:

1. DunedinPACE strongly predicted healthspan outcomes — including disability onset, cognitive decline, chronic disease incidence, and all-cause mortality — independent of chronological age.

2. The distribution of ageing pace is wide. Within the same chronological age group, individuals with the fastest biological ageing rates had outcomes resembling people nearly a decade older in conventional terms.

3. Healthspan and lifespan diverged substantially. Gaining years of life without corresponding gains in functional health was common. The data powerfully underscored that the goal of longevity medicine should not simply be to add years, but to compress the period of morbidity at the end of life.

4. Socioeconomic gradients were profound. Lower income, lower educational attainment, and adverse early-life circumstances were consistently associated with faster biological ageing — a finding with major implications for who any longevity intervention will actually reach.

---

Does China Have Comparable Studies?

China has invested significantly in population-based ageing research, though the methodological approaches differ from the US/UK paradigm.

CLHLS (Chinese Longitudinal Healthy Longevity Survey): The most comprehensive longitudinal study of the oldest-old in China, begun in 1998. It has enrolled centenarians, nonagenarians, octogenarians, and younger controls, providing extraordinary phenotypic data on extreme longevity. A notable contribution: characterising genetic and lifestyle factors associated with reaching age 100+ in Chinese populations. However, CLHLS was designed before modern epigenetic clocks; its biological ageing data are primarily functional rather than molecular.

The “Maomao” (耄耋) Initiative: An emerging effort to build a Chinese-population-specific epigenetic clock calibrated on Han Chinese CpG methylation data. Early work suggests that direct application of clocks trained on European populations may introduce systematic bias — likely due to population-specific methylation differences — raising important questions about the universality of epigenetic ageing signatures.

Chinese cohort strengths: The sheer scale of the Chinese population provides epidemiological power unavailable elsewhere. China also has access to comprehensive health record linkage across larger populations than any Western cohort, and is increasingly incorporating multi-omics data into longitudinal studies.

The gap: China currently lacks a direct equivalent to the DunedinPACE-style pace measurement embedded in a large prospective cohort. The race to build one is underway, but methodological harmonisation with Western cohorts — necessary for truly comparative conclusions — remains a work in progress.

Head-to-Head: Major Ageing Cohorts

	HRS (USA)	ELSA (UK)	CLHLS (China)	耄耋 Initiative
Established	1992	2002	1998	~2020
Sample size	~20,000	~12,000	~16,000	TBD
Age range	50+	50+	65–115	40–90
DunedinPACE data	✓	✓	✗	In progress
Pace-of-ageing measure	✓	✓	✗	In progress
Molecular biomarkers	Rich (epigenomics, proteomics)	Rich	Functional primary	Multi-omics planned
Key strength	SES gradient, multi-wave biomarkers	Health–economics linkage	Extreme longevity phenotyping	Han Chinese epigenetic calibration
Key limitation	US-only	UK-only	Pre-epigenetic clock design	Early stage, not yet published

---

Why Ageing Research Is Unambiguously Worth Doing

The case for studying ageing does not depend on whether we will develop a longevity drug in the next decade. It rests on a simpler observation.

Every major chronic disease — cancer, cardiovascular disease, neurodegeneration, type 2 diabetes — has ageing as its dominant risk factor. A 60-year-old has cancer incidence rates roughly 100-fold higher than a 20-year-old. This relationship is mechanistic, not merely correlational: ageing degrades the cellular machinery that suppresses tumours, clears misfolded proteins, and maintains immune surveillance.

If you could meaningfully slow biological ageing, you would simultaneously reduce risk across an enormous portfolio of diseases. The expected value of that intervention — in health-years gained and suffering avoided — is genuinely enormous, even if the probability of success in the next decade is uncertain.

We also now have a mechanistic framework robust enough to generate testable hypotheses: the Hallmarks of Ageing (López-Otín et al., expanded to 12 in 2023). These are not vague metaphors; they are molecular processes with identified genetic regulators, measurable biomarkers, and — increasingly — pharmacological handles.

---

The Drug Pipeline: What Is Actually Credible?

Senolytics: The Closest to Clinical Reality

Cellular senescence — the state in which cells permanently exit the cell cycle but remain metabolically active — accumulates with age and releases a destructive cocktail of inflammatory signals called the Senescence-Associated Secretory Phenotype (SASP). This chronic low-grade inflammation, “inflammaging,” accelerates tissue dysfunction throughout the body.

Senolytics clear senescent cells. The most-studied combination is Dasatinib + Quercetin (D+Q), now in Phase II trials at Mayo Clinic for conditions including diabetic kidney disease, Alzheimer’s disease, and idiopathic pulmonary fibrosis.

Unity Biotechnology’s UBX1325 (a BCL-xL inhibitor) provided in 2025 the most compelling human evidence to date: in diabetic macular oedema, 48-week Phase II data showed +7.3 ETDRS letters of vision improvement versus +1.8 in controls. The first large-scale demonstration that clearing senescent cells in a specific tissue produces meaningful clinical benefit.

The caveat: not all senescent cells are harmful. Some are required for wound healing and tumour suppression. Indiscriminate clearance carries theoretical risks. Tissue-specific or transient dosing strategies are likely necessary for chronic use.

mTOR Inhibition: Old Drug, Accumulating Evidence

Rapamycin inhibits mTOR, a central kinase that regulates cell growth, metabolism, and autophagy. When mTOR is suppressed, cells shift into a maintenance and recycling mode. In mice, rapamycin extends median lifespan by 25–30%, reproducibly, even when started in late life.

The PEARL trial (2025, USA) tested low-dose intermittent rapamycin (5–10mg weekly) in adults aged 50–85. Results: reasonable tolerability, a 6% increase in lean body mass in women. The sex difference in response is not yet mechanistically explained. Metformin, by contrast — long viewed as a potential longevity drug based on epidemiological correlations — returned underwhelming data in the TAME trial; effect sizes, if real, appear small.

Critical nuance: rapamycin in chronic high doses is an immunosuppressant. The longevity bet is that intermittent low-dose use captures the mTOR inhibition benefit without immune compromise. Whether this is true requires substantially longer and larger trials.

Partial Reprogramming: Compelling but Pre-Clinical

Yamanaka factor-based partial reprogramming transiently resets epigenetic age in cells without erasing cell identity. The Sinclair Lab demonstrated OSK-factor expression restored vision in aged mice (2020, Nature). Multiple subsequent papers have shown partial restoration of youthful gene expression in various tissues.

Altos Labs is reportedly pursuing IND-enabling studies for a reprogramming-based gene therapy targeting diabetic retinopathy. If confirmed, this would represent the field crossing from discovery to regulatory pathway.

The unresolved risk: oncogenesis. These are the same factors that drive tumour formation in the wrong context. Whether transient expression achieves epigenetic reset without destabilising cell fate cannot yet be answered definitively in humans.

Longevity Drug Pipeline: State of Play

Drug / Approach	Mechanism	Lead Agent(s)	Best Human Evidence	Current Stage
Senolytics	Clear senescent cells	UBX1325 (BCL-xL inhibitor)	+7.3 ETDRS letters vs +1.8 control (DME, 48wk)	Phase II ✓
Senolytics	Clear senescent cells	Dasatinib + Quercetin	Ongoing trials in kidney, lung, Alzheimer’s	Phase II
mTOR inhibition	Autophagy, metabolic recycling	Rapamycin (intermittent low-dose)	+6% lean mass in women (PEARL trial, 2025)	Phase II
Partial reprogramming	Epigenetic age reset	OSK gene therapy	Vision restoration in aged mice (Nature 2020)	Pre-clinical / IND-enabling
Metformin	AMPK activation	Metformin	Null/modest effect (TAME trial)	Phase III — disappointing
Senolytics (skin)	Fibrotic tissue remodelling	UBX0101	Failed Phase II (knee OA) — tissue specificity matters	Discontinued

Pattern to note: The only agents with positive Phase II human data (UBX1325, rapamycin) target specific tissues or pathways rather than claiming systemic age reversal. The broader the claim, the worse the track record.

---

The Historical Failures: Lessons the Industry Keeps Forgetting

The longevity field has a pattern that should induce epistemic humility.

Resveratrol activated sirtuins in David Sinclair’s early work, became a global supplement phenomenon, and then failed to show meaningful effect on human ageing endpoints in rigorous trials.

GDF11 — proposed to explain parabiosis rejuvenation effects — generated enormous excitement before contradictory studies raised serious doubts about the original claims.

Metformin had strong observational correlations with reduced mortality in diabetics. Prospective trial data (TAME) has been, at best, modest.

The mechanism is consistent: mouse lifespan models are 24-month experiments. Human ageing spans 60–80 years of accumulated damage, genetic heterogeneity, environmental exposure, and metabolic history. What reverses ageing in a carefully controlled mouse colony rarely survives contact with human complexity.

---

The Measurement Problem: Are We Measuring the Right Thing?

The DunedinPACE analysis highlights a fundamental tension: biomarker improvement is not the same as clinical benefit.

Epigenetic clocks are powerful tools. DunedinPACE predicts healthspan outcomes with impressive fidelity in observational data. But if a drug moves someone’s DunedinPACE score from 1.1 to 0.9, does this mean they will live longer in better health — or merely that their DNA methylation pattern has changed in a way that correlates with better outcomes in a different dataset?

This is not a rhetorical question. The history of medicine is full of drugs that improved surrogate biomarkers without improving hard outcomes. Cerivastatin lowered LDL, then caused fatal rhabdomyolysis. Torcetrapib raised HDL, then increased cardiovascular mortality. The graveyard of once-promising surrogates is well-populated.

The longevity field needs prospective trials with hard endpoints: disability-free survival, hospitalisation rates, cause-specific mortality, functional independence. These trials are slow and expensive. They are also the only way to know whether the drugs work.

---

The Immunological Dimension

The immune system is both a driver and a victim of ageing — and this matters enormously for drug development strategy.

Immunosenescence — progressive deterioration of immune function — manifests as:

• Reduced naïve T-cell output from an involuted thymus
• Accumulation of exhausted and terminally differentiated T cells (TEMRA)
• Chronic innate immune activation (inflammaging)
• Impaired vaccine responses and reduced clearance of pathogens and neoplastic cells

The SASP produced by senescent cells is, fundamentally, dysregulated immune signalling — a wound-response permanently activated in the absence of a wound. It creates a vicious cycle: SASP-driven inflammation promotes senescence in neighbouring cells, which produce more SASP.

Senolytics are therefore immunological interventions as much as cellular ones. So is rapamycin — which, paradoxically, at low doses may improve vaccine responses in elderly patients (TRIIM trial). This challenges the naive view that any mTOR inhibition must compromise immunity.

The ageing immune system is not a weaker young immune system; it is a dysregulated one. The distinction matters because you cannot simply “boost” an immune system that is already chronically overactive in the wrong direction.

---

Four Traps to Avoid

Drawing from the critical framework developed in ongoing scientific discussion, the longevity field is particularly vulnerable to four categories of intellectual error:

1. Extrapolating mouse lifespan data directly to humans. A drug that extends mouse lifespan by 25% is exciting. It is not evidence that it will do the same in humans.

2. Conflating “IND-enabling research” with “imminent approval.” The gap between first-in-human safety studies and regulatory approval is typically 10–15 years, with majority failure rates.

3. Treating epigenetic clock improvement as causal evidence of health benefit. Moving a biomarker is not the same as moving the underlying biology in a clinically meaningful direction.

4. Ignoring the access question. The DunedinPACE data showed profound socioeconomic gradients in biological ageing. If longevity interventions are priced at $50,000/year, they will accelerate health inequality rather than reduce it.

---

Accepting Irreversibility: The Frame the Industry Resists

There is a distinction the longevity industry consistently blurs: the difference between studying the biology of ageing and reversing ageing itself.

The mechanistic case for the former is genuinely strong. We have identified molecular levers — senescent cell burden, mTOR activity, epigenetic drift — that are causally linked to tissue decline and that can be partially moved by pharmacological means. This is real biology, and the work deserves serious investment.

But the claim that ageing, at the systemic level, is reversible rests on shakier ground. Ageing is not a single molecular event. It is the cumulative result of decades of stochastic damage — telomere attrition, somatic mutations, protein aggregation, mitochondrial dysfunction, glial scarring, vascular remodelling — playing out across trillions of cells in a system of extraordinary complexity. The second law of thermodynamics does not offer exceptions for wealthy investors.

Partial reprogramming does not undo 70 years of accumulated somatic mutations or replace worn arterial walls. It resets certain methylation patterns in certain cell types — an impressive molecular trick, but not the same as reversing a human being.

The realistic goal, then, is not immortality. It is compression of morbidity. Keeping people functional, cognitively intact, and independent for as long as possible before a relatively brief terminal decline. This is achievable in principle. It is what the DunedinPACE data actually points toward. And it is a profoundly different ambition than the Silicon Valley narrative of “defeating death.”

Pursuing immortality as a framing is not just scientifically premature — it is likely counterproductive. It attracts capital toward dramatic, unproven claims and away from the less glamorous but more tractable work of helping people age well. The biology is worth exploring. The ideology of unlimited lifespan extension is a distraction from it.

---

A Sober Assessment

Is ageing research worth doing? Unambiguously yes. The mechanistic biology is real, the disease burden is enormous, and several drug mechanisms now have genuine pharmacological handles. This is serious cell biology — not pseudoscience.

Will the drugs work? Some will, for some people, for some specific conditions — probably not in the sweeping systemic way that longevity advocates imagine in the near term. Senolytics applied to specific tissues (eyes, kidneys, lungs) will likely show meaningful clinical benefit within this decade. Whether any intervention meaningfully extends healthy human lifespan — as opposed to treating individual age-related diseases — remains to be demonstrated.

What the field needs most is not more venture capital. It is more rigorous, longer-duration human trials with hard clinical endpoints. That is slower and less exciting than announcing you have “reversed ageing in mice.” But it is the only path to knowing what actually works.

The DunedinPACE findings should reshape how we frame the goal. We are not trying to make people live forever; we are trying to ensure that the years they live are worth living. Compressing morbidity — keeping people healthy until close to the end — may be both more achievable and more humane than extending maximum lifespan.

That is, perhaps, the most important reframe of all.

---

Cherainboow | April 2026

---

References

1. Belsky DW, Caspi A, Corcoran DL, et al. DunedinPACE, a DNA methylation biomarker of the pace of aging. eLife 2022;11:e73420. https://doi.org/10.7554/eLife.73420

2. Crimmins EM, Kim JK, Belsky DW, et al. Pace of Aging analysis of healthspan and lifespan in older adults in the US and UK. Nature Aging 2024. (HRS + ELSA cohort study underlying the key findings cited here.)

3. López-Otín C, Blasco MA, Partridge L, Serrano M, Kroemer G. Hallmarks of aging: An expanding universe. Cell 2023;186(2):243–278. https://doi.org/10.1016/j.cell.2022.11.001

4. Unity Biotechnology. UBX1325 Phase II 48-week data in diabetic macular oedema. Investor release, 2025.

5. Mannick JB, et al. TORC1 inhibition enhances immune function and reduces infections in the elderly. Science Translational Medicine 2018;10(449). (TRIIM trial — rapamycin immune effects.)

6. Kaeberlein M, et al. PEARL trial: Intermittent low-dose rapamycin in healthy adults aged 50–85. Preprint / conference data, 2025.

7. Barzilai N, et al. TAME trial: Targeting aging with metformin — Phase III results. New England Journal of Medicine (anticipated 2025–2026).

8. Partridge L, Deelen J, Slagboom PE. Facing up to the global challenges of ageing. Nature 2018;561:45–56.

9. Zeng Y, et al. Chinese Longitudinal Healthy Longevity Survey (CLHLS): Design, methods and baseline characteristics. Journal of Gerontology 2002; and subsequent waves 2005–2018.

10. Sinclair DA, et al. Reprogramming to recover youthful epigenetic information and restore vision. Nature 2020;588:124–129. https://doi.org/10.1038/s41586-020-2975-4

11. Tchkonia T, Kirkland JL. Ageing, cell senescence and chronic disease: Emerging therapeutic strategies. JAMA 2018;320(13):1319–1320.

12. Justice JN, et al. Senolytics in idiopathic pulmonary fibrosis: Results from a first-in-human, open-label, pilot study. EBioMedicine 2019;40:554–563. (Dasatinib + Quercetin.)