The Frontier

The reason this stands above almost everything else in comparative longevity is the completeness of the causal chain: CIRBP overexpression in human cells increased repair success; knockdown in whale cells reduced it; whale cells make fewer deletion errors at repair junctions and resolve γH2AX/53BP1 foci faster; and the chain closes with a cross-species in vivo rescue — CIRBP-overexpressing Drosophila showed extended lifespan and markedly improved radiation survival.

Most 'why does this animal live so long' work stops at a correlation between a gene and a lifespan. This one has knockdown, overexpression, biochemical reconstitution and a cross-species rescue. Remember this one.

Source tier — Primary — Nature, October 2025 (Firsanov / Zacher / Gorbunova / Seluanov, University of Rochester). Tier-1, causally validated.

There is no single 'virtual cell'. CZI's effort has fragmented — sensibly — into a family of narrow AI models, with one genuinely mechanistic simulation project running in parallel and largely separate from it.

Running orthogonally, a true mechanistic whole-cell simulation exists — but only for JCVI-syn3A, a minimal synthetic bacterium with 493 genes: biochemical reaction networks, gene expression, spatial structure and molecular dynamics across 50 replicate runs. That is the closest thing to an actual digital-twin cell, and it is emphatically not what the AI models are doing.

Source tier — Primary preprints (Arc/State, GREmLN) plus one independent critical source (Nature Biotechnology editorial). Performance claims for rBio/GREmLN are self-reported in company blog posts and not independently replicated.

The best-benchmarked result this cycle came from an unexpected corner. Ginkgo Bioworks ran a closed design-build-test-learn loop across six rounds over six months on cell-free protein synthesis, with a model given internet access, a data-analysis toolchain and prior experimental metadata, and a schema-validation layer checking plate layout, controls and reagent constraints before each round reached the robotic carts. Result: superfolder GFP at $422/gram in reaction-component cost against a prior published state of the art of $698/gram. A 40% reduction, benchmarked against a specific named publication rather than a vague claim — that specificity is rare enough to be the reason it is worth taking seriously.

• Recursion (merged with Exscientia) has been cutting pipeline rather than scaling triumphantly — 6 active programmes post-merger. One real clinical number: REC-4881 showed 43% polyp reduction at week 13 deepening to 53% at week 25 in a Phase 2 FAP trial. Its large 'map' claims are data-asset claims, not verified throughput.
• Lila Sciences (Flagship; George Church as chief scientist) raised a $350M Series A in June 2026 on top of a $200M seed. Funding figures are primary; the technical claims about outperforming commercial therapeutics are unaccompanied by independent numbers.
• Two counter-signals worth noting because they cut against the everything-is-accelerating narrative: Insilico Medicine's most detailed public description of its automated lab is, on inspection, 2022 news resurfacing in 2026 search results with no updated throughput figures — and Emerald Cloud Lab and A-Alpha Bio have essentially no substantive 2025–26 public news at all. Absence is itself information.

Naked mole rat — hyaluronan, transferred into mice

Naked mole rats produce cytoprotective high-molecular-weight hyaluronan via a modified HAS2 gene plus suppressed hyaluronidase activity. The causally important result is that the Rochester group transferred naked-mole-rat Has2 into laboratory mice and saw extended median lifespan (~4.4%), improved resistance to spontaneous and chemically-induced tumours, and reduced age-related inflammation — this is the same body of work as the hyaluronan finding itself, not a separate later follow-up, and a previous framing of it here wrongly split one result into two. Note how modest 4.4% is against the 'naked mole rats are immortal' framing.

Turritopsis dohrnii — reprogramming factors in the wild

RNA-seq across all four life-cycle stages shows SIRT3 peaking during the cyst/transdifferentiation stage, SIRT6 peaking during the reversed-polyp stage alongside DNA-repair efficiency, telomerase-associated genes (TERT, EST1A, RTEL1) co-expressed specifically at the cyst stage, and — the striking part — Yamanaka-factor homologs present, with a c-Myc homolog rising steadily through reverse development and peaking at the cyst stage. That directly parallels reprogramming-factor dynamics in induced pluripotency. Preprint, not peer-reviewed; hold loosely.

Greenland shark — chromatin and Peto's paradox

A 5.9 Gb genome with 37,125 predicted genes and unusually high (83.1%) repetitive content. Expanded gene families cluster around chromatin organisation and chromosome condensation — proposed as a buffer against age-related DNA damage — with specific immune families expanded (TNF ×14, TLR ×31, LRRFIP ×12) and seven positively selected cancer-related genes, offered as a partial species-specific resolution of Peto's paradox. Preprint.

The headline event this cycle was a failure. Unity Biotechnology dissolved in 2025. Its lead compound foselutoclax — a BCL-XL inhibitor targeting senescent cells in diabetic retinal vasculature — reached Phase 2b and *missed* its primary endpoint — it did not establish non-inferiority to aflibercept on best-corrected visual acuity, though it showed a durable signal on some secondary measures — and the company then ran out of cash, laid off all staff including the CEO, and wound up. Clinical promise does not reliably convert into commercial survival in this sector.

The encouraging counterweight: Rubedo Life Sciences' RLS-1496, described as the first GPX4 modulator studied in humans against pathologic senescent cells — mechanistically distinct from BCL-2 apoptosis priming, since GPX4 normally suppresses ferroptotic lipid peroxidation and the approach appears to modulate it to push senescent cells toward death. Company-announced Phase 1 topical results (April 2026) — no peer-reviewed publication, so this sits at company tier despite the specificity of the numbers: reduced senescent-cell burden, ~20% epidermal thickness reduction in psoriasis, meaningful itch improvement in a quarter of atopic dermatitis patients versus none on vehicle, and dose-dependent collagen induction in photoaged fibroblasts. Going after dermatology first, before systemic aging, is a sound runway strategy given Unity's fate.

The catch is that full reprogramming reliably causes loss of somatic identity and, in vivo, teratoma formation — which is why essentially every company is chasing partial reprogramming: cyclic short-pulse expression of OSK (usually dropping Myc, the principal oncogenic driver), aiming to move the epigenome younger without fully dedifferentiating the cell. Where the sweet spot lies between rejuvenation and loss-of-identity is unresolved and is the real bottleneck.

• Life Biosciences' ER-100 is by wide consensus the first human clinical trial of partial reprogramming — OSK delivered by intravitreal injection to retinal ganglion cells for open-angle glaucoma and NAION. FDA cleared the IND on 28 January 2026; Phase 1 began Q1 2026 (NCT07290244).
• NewLimit raised $435M Series C (June 2026, Founders Fund-led), reporting a prototype achieving epigenetic age reversal in old human liver cells, with first-in-human planned for 2027. Factor-combination and mechanism claims are company-sourced, not yet published.
• Altos Labs — four years and ~$3B in — remains largely opaque; the most notable 2026 development is increased media visibility rather than new peer-reviewed results. Worth flagging as under-scrutiny relative to the capital involved.
• Retro Biosciences raised at a $1.8B pre-money valuation, but its lead programme RTR242 targets autophagy restart (mTOR/AMPK-regulated lysosomal recycling), not OSKM chromatin remodelling — a mechanistically different category that often gets loosely folded into 'reprogramming progress'.

The obstacle is unchanged: the FDA has no disease classification for aging itself, which forces every aging-adjacent drug into a narrow disease-specific framing. That is precisely why senolytics keep being tested against single named diseases, and why the most likely first regulatory landmark is veterinary.

• Loyal's LOY-002 cleared FDA Target Animal Safety and Reasonable Expectation of Effectiveness reviews as of January 2026, with an Expanded Conditional Approval application planned. If it goes through it becomes the first FDA-approved lifespan-extension drug in any species — built on the STAY study, the largest veterinary clinical trial ever run (1,300 dogs, 70 clinics).
• TAME (Targeting Aging with Metformin) remains unfunded at ~$75M. Design is finalised — 3,000+ participants aged 65–79, 14 sites, 6 years, Wake Forest coordinating — but there is no pharma incentive to sponsor a trial of a cheap generic. Results, if funded, not before 2027–28.
• ARPA-H's PROSPR programme committed up to $144M across 7 teams to build validated aging biomarkers and run Phase 3 trials of approved drugs in healthy older adults — explicitly framed as laying the regulatory track that would have to exist before the FDA could ever accept aging as an indication.

The concrete outcome: Tim Andrews became the longest-surviving genetically-engineered-organ recipient at 271 days before rejection, at which point the pig kidney was explanted and replaced with a matched human kidney — the first person to go pig kidney → human kidney, and notably requiring only about a third of the immunosuppression burden on the human organ. FDA cleared a full Phase 1/2/3 IND in September 2025. United Therapeutics/Revivicor is pursuing a mechanistically more ambitious parallel track: a 'xenothymokidney' co-transplanting thymic tissue to re-educate the recipient's T-cell repertoire toward tolerance rather than relying on pharmacologic suppression.

The field has genuinely moved from 'can this work at all' to 'how long can it last and what exactly causes eventual rejection'. Current ceiling ~9 months in the best case. Real, non-hyped progress — and the moonshot category that advanced most this year.

Organ bioprinting — mostly vaporware, relative to xeno

No vascularised, transplantable, human-scale bioprinted organ has been implanted in a human. The unsolved fundamental is diffusion: cells more than roughly 200 microns from a blood supply die, so any thick construct needs a functioning vascular tree built in, not printed around. Most 'breakthrough' coverage is review-article or content-farm material. Worth flagging explicitly so xenotransplant momentum does not imply bioprinting is catching up. It is not.

Whole-brain emulation — the bottleneck is data, not compute

No organism has ever been fully recorded at single-neuron resolution. Connectomics progress is real but small: complete C. elegans (~300 neurons), fully proofread Drosophila connectomes (~140,000 neurons) for both sexes, and the MICrONS consortium's cubic millimetre of mouse visual cortex (Allen Institute / Baylor / Princeton, not Princeton alone) — ~200,000 cells, 523 million synapses. Mouse brains have ~500× more neurons than a fly; humans ~1,000,000× more. The field is also tiny: an estimated fewer than 500 people worldwide work directly on emulation objectives. Small-organism emulation within a decade is plausible; human whole-brain emulation is on no credible near-term timeline.

Cryonics — genuinely reinvigorated, with falsifiable metrics

Real biophysics progress rather than marketing: first in-house whole-body CT scans validating cryoprotectant perfusion, 'excellent vitrification with minimal ice formation' in roughly 40% of tested porcine kidneys — a concrete, falsifiable number — brain slice cultures surviving 2–3 weeks post-cryopreservation, and exploratory antifreeze-protein gene therapy work (antifreeze proteins bind ice-crystal faces and inhibit further growth by adsorption-inhibition: real physics). The distinction worth internalising is that Until Labs is pursuing *reversible organ* vitrification — a far shorter-horizon, more falsifiable goal (organ banking) than whole-body suspension, and probably the more useful thing to track.

Therapeutic plasma exchange — the defensible version of 'young blood'

A Buck Institute trial (Fuentealba et al., Aging Cell, 2025) found biweekly TPE + IVIG produced an average 2.61-year reduction in epigenetic age, and TPE alone 1.32 years, front-loaded with diminishing returns. Disclose what the file's own rubric requires: the trial was industry-sponsored, which does not invalidate it but does move it off pure primary tier. Mechanistically this supports dilution of accumulated pro-aging plasma factors rather than addition of any specific young factor — a meaningfully more defensible claim than the discredited parabiosis hype cycle, because it requires no magic molecule, only removal.

• Cardiorespiratory fitness. The largest single modifiable mortality signal in medicine, with no observed upper limit. Raise it with intensity, not only with Zone 2.
• Resistance training and the muscle mass it builds — mechanically, metabolically, and endocrinologically (see the mechanism ledger).
• Sleep regularity and depth. Now causally linked in humans to glymphatic clearance of amyloid-beta and tau.
• Protein sufficiency with leucine distributed across feedings, not concentrated in one.
• Zero alcohol, near-zero ultra-processed food, sun protection, and dental hygiene. Unglamorous, cheap, and better evidenced than anything below.

The strongest evidence in this entire dossier points at interventions that are free, widely known, and boring. That is not a disappointing finding — it is the finding.

Urolithin A — the best-evidenced newer addition

A gut-microbiome metabolite of ellagitannins from pomegranate, walnuts and berries; most people are natural low-producers, which is the actual rationale for supplementing rather than eating more pomegranate. It induces mitophagy through PINK1/Parkin: in a damaged mitochondrion with collapsed membrane potential, PINK1 accumulates on the outer membrane instead of being imported and degraded, recruits Parkin (an E3 ubiquitin ligase), which ubiquitinates outer-membrane proteins to tag the organelle for autophagosomal engulfment — with PGC-1α-mediated compensatory biogenesis downstream. A 2025 Nature Aging RCT (N=50, ages 45–70, 1,000mg/day, 4 weeks) confirmed ex vivo mitophagy activation within 48 hours plus expanded naive-like CD8+ T cells with reduced terminal exhaustion, increased CD8+ fatty-acid-oxidation capacity, expanded NK subsets and improved monocyte bacterial uptake. Real mechanism, real recent RCT, quantified endpoints — the strongest combination in this tier.

GlyNAC (glycine + N-acetylcysteine)

Clean mechanism: glutathione synthesis requires glycine, cysteine and glutamate; aging reduces glycine and cysteine availability; NAC and glycine supply the two rate-limiting substrates. Consistent pilot-trial signal — reversed glutathione deficiency, oxidative stress and mitochondrial dysfunction, with improvements in strength, gait speed and cognition that partially reversed after washout, which supports causality. But n≈8–24 per arm and essentially one research group. Promising, not replicated.

Topical rapamycin

An exploratory trial (Chung et al., GeroScience, 2019) found 0.001% topical rapamycin on dorsal hand skin over 6–8 months significantly reduced p16INK4A and increased collagen VII at the basement membrane, with no detectable systemic absorption. Flag the size honestly: only 17 of 36 enrolled completed, 8 biopsies usable. The rationale — mTORC1 hyperactivation drives senescence and SASP, which plausibly contributes to photoaging independent of UV — is coherent, and compounding pharmacies cite this single small study to sell product as if it were settled.

• Metformin now has real, not theoretical, evidence of blunting exercise adaptation — via the same mitohormetic ROS suppression that makes it metabolically protective. A genuine trade-off for anyone training seriously.
• Taurine's founding premise is substantially refuted (levels rise or stay flat with age, not fall). If it entered a stack on the 2023 Science paper, that reason is gone.
• NAD+ precursors: bioavailability confirmed by a 2026 RCT, benefit not measured. The most persistently oversold category in the space.
• Spermidine: exquisite mechanism (eIF5A hypusination), null flagship cognition RCT, and a real bioavailability problem at commercial doses.
• GLP-1 agonists as general geroprotectors: mechanistically plausible, genuinely under-evidenced in healthy populations, with a 2026 preprint suggesting tirzepatide carries greater relative lean-mass loss than semaglutide (1.1–2.0 percentage points excess across 3–12 months). Preprint, 85% bioimpedance rather than DEXA — a signal, not a settled result.