🧬 Aging Wiki

in-vivo-partial-reprogramming-therapy

Properties1
aliasesin-vivo OSK therapy, partial reprogramming therapy, Yamanaka factor therapy, transient reprogramming therapy, epigenetic reprogramming therapy

17 min read

In-vivo partial reprogramming therapy

The intervention: delivery of Yamanaka reprogramming factors — OCT4, SOX2, KLF4, and optionally c-MYC (OSK or OSKM) — into living somatic tissues via viral vector, mRNA-LNP, or small-molecule mimicry, using pulsed/transient expression schedules to reverse epigenetic age marks without inducing pluripotency or erasing cell identity. This page covers the intervention as a clinical-development entity: delivery modalities, industrial pipeline, translation barriers, and trial status.

For the underlying biology (mechanism, experimental evidence, critical caveats, extrapolation table), see partial-reprogramming (verified). For the narrow AAV-vector-specific view, see aav-osk (R23b sibling, to be seeded).

The clinical concept

Full Yamanaka reprogramming erases cell identity, creating iPSCs with teratoma potential. Partial reprogramming delivers the same factors transiently or cyclically, restoring youthful epigenetic patterns while the cell retains its lineage identity. The core idea is that epigenetic drift — the progressive disorder of methylation and chromatin states — is a reversible cause of cellular aging, not merely a correlate (see information-theory-of-aging, verified).

The therapeutic implication: if you can pulse the reprogramming signal, you may be able to “rewind” the epigenetic clock in aged tissues without the safety liabilities of full reprogramming. As of 2026-05-08, all direct rejuvenation evidence remains preclinical (mouse models for OSK gene therapy; C. elegans lifespan extension for chemical reprogramming). The first regulated human Phase 1 trial is NCT07290244 (Life Biosciences ER-100, AAV2-OSK intravitreal in optic-nerve conditions; n=18, RECRUITING since March 2026 1) — safety-focused with no epigenetic-age efficacy endpoint. No partial-reprogramming trial in any aging-as-indication context has yet been initiated.

Delivery modalities

The modality choice is the central clinical-development question. Each trades off duration of expression, safety, tissue targeting, and manufacturing complexity differently.

AAV-OSK (current preclinical standard)

Adeno-associated virus (AAV) carrying a polycistronic OSK cassette. The academic gold standard for in-vivo partial reprogramming experiments.

Lu 2020 (Nature) — The landmark proof-of-concept: AAV2-mediated OSK delivery to mouse retinal ganglion cells (RGCs) restored axon regeneration after optic nerve crush and reversed vision loss in a glaucoma model. DNA methylation age was restored toward a youthful signature in RGCs 2. This study established: (1) tissue-restricted AAV delivery is feasible for OSK; (2) TET1/TET2 demethylases are required (TET3 is not); (3) the retina is a tractable initial target for human translation. needs-human-replication

Browder 2022 (Nature Aging) — Systemic AAV-OSKM in normally aged (25-month) C57BL/6J mice: showed reduced expression of aging-associated gene sets and some markers of tissue aging after 7 months of cyclic expression; did not report lifespan extension as a primary endpoint 3. This is the broadest systemic in-vivo test of partial reprogramming in a normally aged (non-progeria) mouse. needs-replication

AAV liabilities in humans: immunogenicity against the capsid; one-time dosing due to immune memory (re-dosing is blocked); manufacturing scale; tissue-specific serotype matching; no AAV-delivered gene-editing is currently approved for systemic aging indications. For AAV-specific technical considerations, see aav-osk.

mRNA-LNP-OSK (emerging; transient by design)

Lipid nanoparticle-encapsulated mRNA encoding OSK. Key advantages over AAV:

  • No genome integration — mRNA does not incorporate into the nucleus; no insertional-mutagenesis risk.
  • Inherently transient — mRNA half-life in cells is hours to days; multi-factor expression terminates automatically after each dose.
  • Re-dosable — no lasting immune memory of the delivery vehicle comparable to AAV; multiple doses feasible in principle.
  • Manufacturing precedent — LNP platform de-risked by COVID-19 mRNA vaccines; manufacturing infrastructure exists.

Liabilities: current LNP formulations traffic predominantly to the liver after systemic dosing; retinal or CNS targeting requires local delivery or modified LNP chemistry. OSK protein half-life post-translation is short; whether sufficient TF dose reaches the nucleus to drive epigenetic changes at clinically meaningful scale has not been demonstrated in aged mammals in published data as of 2026-05-06. needs-replication

2025–2026 field expansion (R34 recency batch)

The 2025–2026 partial-reprogramming literature consolidates several mechanistic advances and provides the first cautionary lifespan/toxicity data in mammals for chemical approaches:

  • Mesenchymal drift as a unifying transcriptomic target (Lu 2025, Cell; Altos Labs). Across 40+ human tissues and 20 diseases, age- and disease-associated upregulation of mesenchymal-fate gene expression (“mesenchymal drift,” MD) correlates with mortality and disease progression; OSK partial reprogramming reduces MD prior to dedifferentiation 4. First Altos Labs first-author publication on canonical partial-reprogramming biology.
  • Chemical reprogramming extends C. elegans lifespan (Schoenfeldt 2025, EMBO Molecular Medicine; Ocampo lab + EPITERNA). A two-chemical optimized cocktail derived from chemical-induced partial reprogramming improved genomic-instability and epigenetic-alteration markers in aged human cells, ameliorated senescence and oxidative stress, and extended C. elegans lifespan and healthspan in vivo 5. This is the first invertebrate lifespan extension via chemical partial reprogramming. EPITERNA is an Ocampo-affiliated commercial spinout.
  • But chemical reprogramming has dose-dependent mammalian toxicity (Mitchell 2026, Aging Cell; Gladyshev lab, Harvard). In genetically-diverse UM-HET3 mice, low-dose chemical-reprogramming-cocktail osmotic-pump infusion had no transcriptomic-age effect on kidney or liver and only modest OXPHOS expression changes; high-dose caused drastic body-weight loss requiring euthanasia, hepatic lipid-droplet accumulation, and mitochondrial stress 6. Supersedes the optimistic Yang 2023 Aging in-vitro chemical-reprogramming result with an in-vivo mammalian negative finding. The simple in-vivo translatability of in-vitro chemical reprogramming is now in doubt. contradictory-evidence
  • Engram-cell OSK rejuvenates cognition (Berdugo-Vega 2026, Neuron; Gräff lab, EPFL). Cell-type-specific OSK delivery to memory-trace neurons in aged + AD-model mice restored learning and memory to young-mouse levels 7. Shifts the field’s translation question from “tissue dosing” toward “cell-type-specific dosing” — bypassing the mass-distribution problem.
  • Improved comparative reprogrammable mouse-strain tools (PicĂł 2025, Cell Reports; Ocampo lab). Continuous OSKM expression remains lethal across strains; cyclic short-term and organ-avoidance variants partially mitigate the body-weight-loss/premature-death liabilities; new strains released as community resources 8. Tooling progress, not new biological claims.
  • Altos Labs publishes on ex-vivo organ rejuvenation (Haoui 2026, Cell Stem Cell). First Altos Labs perspective piece outlining a near-term clinical strategy: combining partial reprogramming with ex-vivo machine-perfusion organ-preservation platforms to rejuvenate suboptimal donor organs before transplant 9. Decouples partial-reprogramming clinical translation from systemic in-vivo dosing — applies it to organs in isolation. May be the first commercially deliverable partial-reprogramming product class.
  • Local AAV-OSK in osteoarthritis (Liu 2026, Experimental & Molecular Medicine). Intra-articular AAV-OSK delivery preserved chondrocyte vitality, reduced subchondral bone thickening, promoted cartilage hyalinization in OA mice; TET2-dependent 10. Adds OA to the validated AAV-OSK indication list.

Small-molecule / chemical reprogramming

Transcription-factor-free induction of a reprogramming-like epigenetic state using chemical cocktails. Yang et al. 2023 (Aging) identified six chemical cocktails (including inhibitors of GSK-3α/β [CHIR-99021], TGF-β [E-616452/RepSox], and epigenetic modifiers including the HDAC inhibitor valproic acid) that partially reverse transcriptomic aging signatures in human senescent fibroblasts in vitro 11. The primary readout was transcriptomic age (clocks trained on RNA-seq), not DNA methylation-based epigenetic age. No MEK inhibitor is present in these cocktails. needs-human-replication needs-replication

Earlier precedent: Hou et al. 2013 (Science) generated induced pluripotent-like cells from mouse somatic cells using small molecules only — validating that pluripotency induction does not require protein TFs, in principle 12. needs-human-replication

Chemical approaches are early-stage. The “partial” vs “full” reprogramming distinction is harder to control with small molecules than with transient mRNA. No in-vivo aged-animal data for chemical partial reprogramming has been published in a high-impact venue as of 2026-05-06.

Industrial pipeline (as of 2026-05-08)

CompanyFounders / backingModalityDisclosed focusStatus
Life BiosciencesDavid Sinclair (co-founder)AAV2-OSK intravitreal (ER-100)Optic neuropathies (OAG, NAION); dry AMD as next indicationPhase 1 RECRUITING — NCT07290244 1, opened March 2026, n=18, single-dose safety/tolerability 4
Altos LabsEric Verdin (CEO), Richard Klausner; Bezos Expeditions, SoftBank; Shinya Yamanaka on scientific advisory boardUndisclosed (AAV + mRNA + ex-vivo organ perfusion under investigation)Multi-tissue partial reprogramming; ex-vivo organ rejuvenation (Haoui 2026 perspective)Preclinical; no IND disclosed as of 2026-05-08; first author-publication output 2025–2026 49
NewLimitBrian Armstrong (Coinbase), Jake HwangChemical / mRNA (undisclosed)T-cell rejuvenation as initial focusPreclinical
Retro BiosciencesJoe Betts-LaCroix; OpenAI-adjacent backing (Altman)Undisclosed (partial reprogramming + plasma protein targets)Longevity platformPreclinical
EPITERNA SAAlejandro Ocampo, Kevin PerezChemical reprogramming (small-molecule cocktails)Lifespan/healthspan extension via chemical partial reprogrammingPreclinical; C. elegans in-vivo lifespan extension published 2025 5 but mammalian in-vivo toxicity issues identified by independent group (Mitchell 2026) 6

unsourced — company pipeline details come from publicly available press releases and investor communications, not peer-reviewed publications. Pipeline contents and focus areas change; this table reflects best-available public information as of 2026-05-06 and should be refreshed each lint pass.

Translation barriers

BarrierDetail
Cancer risk (OSKM)c-MYC is a proto-oncogene; sustained or inadvertently high expression drives tumorigenesis. OSK (without c-Myc) preferred for clinical development. Even OSK: OCT4 and SOX2 have oncogenic co-activity in certain cellular contexts. Long-term tumor surveillance data in aged animals are absent.
Tissue-specific dosing windowThe therapeutic window between “sufficient epigenetic rejuvenation” and “partial/full dedifferentiation” is tissue-specific and not characterized in any human tissue. Post-mitotic cells (neurons, cardiomyocytes) may have tighter windows than cycling cells.
Which biomarker readoutEpigenetic clocks diverge in their response to interventions (see dunedinpace-2022, grimage-2019, horvath-clock-2013). The field has not converged on which clock should serve as a primary endpoint in a partial-reprogramming trial; the most intervention-responsive (DunedinPACE) is also the youngest and least validated for this context.
Regulatory noveltyIn-vivo partial reprogramming has no regulatory precedent. It is not a gene therapy in the traditional sense (no therapeutic gene insert), not a small molecule, and not a cell therapy — it falls into a regulatory gray zone. FDA classification will affect manufacturing and trial design requirements.
Delivery to aged tissuesAged tissues show reduced AAV transduction efficiency; LNP targeting beyond liver is unproven at scale; small-molecule penetration to specific aged-cell subtypes requires tissue distribution validation.
Replication gapThe foundational OSK experiments originate from a single laboratory (Sinclair group). Independent systemic replication at scale has not been published as of 2026-05-06. This is the primary epistemic caveat for the entire field. See partial-reprogramming § Critical caveats.

Relationship to the information theory of aging

In-vivo partial reprogramming is the primary proposed therapeutic intervention of information-theory-of-aging (verified). That hypothesis holds that aging is caused by the progressive loss of the epigenome’s ability to read the genome correctly — analogous to accumulated noise in an analog recording — and that the cell retains a “backup copy” of youthful epigenetic programs accessible via OSK-driven demethylation.

The therapeutic implication follows directly: if the backup is intact and accessible, periodic epigenetic “resets” could, in principle, prevent or reverse age-associated functional decline. Yang 2023 (Cell) is the strongest current evidence that the causal direction runs from epigenetic disruption → accelerated aging and that OSK can reverse it, at least in the ICE artificial-induction mouse model 13.

The full iPSC reprogramming background: Lapasset et al. 2011 (Genes & Development) showed that full iPSC reprogramming of senescent and centenarian human cells erased all detectable markers of cellular aging — elongating telomeres, restoring youthful gene expression, correcting mitochondrial function 14. Importantly, this required a six-factor cocktail (OCT4, SOX2, KLF4, c-MYC, NANOG, LIN28 — abbreviated OSKMNL), not the standard four-factor OSKM; the standard four factors failed to efficiently reprogram senescent cells. This is the historical foundation: full reprogramming “resets” aging, but destroys cell identity. Partial reprogramming attempts to harvest only the epigenetic reset component. needs-replication — erasing aging markers in vitro does not imply clinically useful in-vivo rejuvenation.

Transcriptomic-clock evidence (2026). Tyshkovskiy et al. 2026 applied multi-species transcriptomic clocks to reprogramming data and found that iPSC reprogramming reduced both chronological and mortality tAge (i.e., molecular rejuvenation), with the strongest rejuvenation signal in the epithelial-to-mesenchymal-transition (EMT/MET) module clock, consistent with the dedifferentiation and mesenchymal-to-epithelial transition that drives early reprogramming 15. This gives reprogramming a quantitative, pathway-resolved rejuvenation readout, and — importantly — the same study showed reprogramming’s rejuvenation signature overlaps those of caloric restriction and embryonic “ground zero” rejuvenation (shared down-regulation of Cdkn1a, Ccl5, S100a9, Lgals3), tying partial reprogramming to a common cross-intervention rejuvenation axis. Notably, transcriptomic tAge moved under reprogramming whereas DNA-methylation age did not respond to some damage perturbations (γ-irradiation, hTERT) — so transcriptomic and epigenetic clocks capture complementary aspects of the reset.

Extrapolation table (preclinical → human)

DimensionStatusNotes
Pathway conserved in humans?partialOCT4/SOX2/KLF4 are human TFs; TET demethylase mechanism conserved; but specific aging-relevant CpG loci differ between mouse and human epigenetic clocks
Phenotype conserved in humans?unknownNo human somatic-cell partial reprogramming data published as of 2026-05-06
Replicated in humans?noAll direct partial reprogramming experiments in mouse models only

Limitations and gaps

needs-human-replication — No human partial reprogramming study has been published or initiated.

needs-replication — Core OSK/OSKM in-vivo rejuvenation claims are from a single laboratory (Sinclair group at Harvard); Ocampo 2016 (Belmonte group) is independent support but uses a different system (OSKM, progeria model). Systemic independent replication is lacking.

long-term-unknown — No completed study has demonstrated that cyclic OSK/OSKM extends maximum lifespan in normally aged mice from a standard laboratory strain (as distinct from progeria models).

dose-response-unclear — The therapeutic window between epigenetic rejuvenation and partial dedifferentiation has not been systematically characterized in aged animals, let alone humans. Tissue-specific dosing parameters are unknown.

unsourced — Industrial pipeline entries sourced from public communications, not peer review. Refresh at each lint pass.

Cross-references

Footnotes

Footnotes

  1. NCT07290244 · “Evaluating ER-100 for Safety in People With Glaucoma or Non-Arteritic Anterior Ischemic Optic Neuropathy (Optic Nerve Conditions)” · Sponsor: Life Biosciences Inc. · Phase 1 single-dose · n=18 (planned) · status (as of 2026-05-08): RECRUITING · started 2026-03-02; primary completion estimated 2027-05; full completion 2032-03 · indications: open-angle glaucoma + NAION · primary endpoint: safety/tolerability through dox-activation period · ClinicalTrials.gov v2 API verified 2026-05-08 ↩ ↩2

  2. doi:10.1038/s41586-020-2975-4 · in-vivo · n= not a single-n study (AAV2-OSK delivery to mouse RGCs across multiple cohorts: aged wild-type, optic nerve crush, glaucoma model) · model: aged C57BL/6J mice; retinal ganglion cells · Lu Y et al., Nature 2020 · PDF downloaded locally · restored axon regeneration, vision, and youthful DNA methylation in RGCs; TET1/TET2 demethylase required, TET3 not required needs-human-replication ↩

  3. doi:10.1038/s43587-022-00183-2 · in-vivo · model: aged C57BL/6J mice (25-month baseline; AAV-OSKM, 7 months cyclic expression) · Browder KC et al., Nature Aging 2022 · not locally downloaded (not_oa) · reduced aging-associated gene-expression signatures systemically in normally aged mice; lifespan extension not a primary endpoint needs-replication ↩

  4. doi:10.1016/j.cell.2025.07.031 · Lu JY, Tu WB, Li R et al. (Izpisua Belmonte group, Altos Labs) · in-vivo + in-silico · Cell 2025 · 188(21):5895–5911.e17 · model: gene expression analysis across 40+ human tissues and 20 diseases; OSK partial reprogramming validation in human and mouse cells · identified pervasive “mesenchymal drift” (MD) — upregulation of mesenchymal-fate TFs across cell types in aging/disease — that correlates with reduced patient survival; OSK reduces MD before dedifferentiation; suppression of MD TFs sufficient for epigenetic rejuvenation · Altos Labs first-author publication · disclosure: all authors Altos Labs employees · PMID 40816266 · R34 recency add ↩ ↩2 ↩3

  5. doi:10.1038/s44321-025-00265-9 · Schoenfeldt L, Paine PT, Picó S et al. (Ocampo lab + EPITERNA, U Lausanne) · in-vivo + in-vitro · EMBO Molecular Medicine 2025 · 17(8):2071–2094 · model: aged human cells (chemical-induced partial reprogramming); C. elegans in vivo · two-chemical optimized cocktail improved genomic instability + epigenetic alterations in aged human cells, ameliorated cellular senescence + oxidative stress, extended C. elegans lifespan and healthspan in vivo · PMID 40588563; PMC12340157 (OA) · disclosure: Perez & Ocampo are EPITERNA co-founders/shareholders · R34 recency add ↩ ↩2

  6. doi:10.1111/acel.70390 · Mitchell W, de Magalhães CG, Tyshkovskiy A et al. (Gladyshev lab, Brigham & Women’s / Harvard) · in-vivo + in-vitro · Aging Cell 2026 · 25(2):e70390 · model: aged mouse fibroblasts in vitro; UM-HET3 genetically-diverse middle-aged male mice in vivo · in vitro: chemical reprogramming alters mitochondrial cristae morphology, increases mitochondrial size and fusion. In vivo via implantable osmotic pump: low dose well-tolerated but no transcriptomic-age effect on kidney/liver, only modest OXPHOS expression changes; high dose caused drastic body-weight loss requiring euthanasia, hepatic lipid-droplet accumulation, mitochondrial stress · supersedes overoptimistic in-vitro chemical-reprogramming framing; first negative in-vivo mammalian result · PMID 41589348; PMC12835892 (OA) · contradictory-evidence — directly contradicts EPITERNA C. elegans result 5 in mammalian context · R34 recency add ↩ ↩2

  7. doi:10.1016/j.neuron.2025.11.028 · Berdugo-Vega G, Sierra C, Astori S et al. (Gräff lab, EPFL) · in-vivo · Neuron 2026 · 114(6):1102–1116.e7 · model: aged WT + Alzheimer’s-model mice; engram-cell-targeted OSK gene therapy · cell-type-specific OSK delivery to memory-trace neurons reversed senescence + AD-related cellular hallmarks, restored synaptic-plasticity epigenetic-transcriptional patterns, counteracted AD neuronal hyperexcitability, recovered learning/memory to young-mouse levels · first cognitive-rejuvenation via cell-type-specific OSK · PMID 41672073 · R34 recency add ↩

  8. doi:10.1016/j.celrep.2025.115879 · Picó S, Vílchez-Acosta A, Agostinho de Sousa J et al. (Ocampo lab, U Lausanne / EPITERNA) · in-vivo · Cell Reports 2025 · 44(7):115879 · model: comparative reprogrammable mouse strains · continuous in-vivo OSKM expression remains lethal; cyclic short-term and organ-avoidance variants mitigate weight-loss/premature-death liabilities; new strains released as community tools · PMID 40560729 · R34 recency add ↩

  9. doi:10.1016/j.stem.2025.12.011 · Haoui M, Reddy P, Izpisua Belmonte JC (Altos Labs) · perspective · Cell Stem Cell 2026 · 33(1):13–28 · proposes near-term clinical strategy: combining partial reprogramming with ex-vivo machine-perfusion donor-organ-rejuvenation platforms; decouples partial-reprogramming clinical translation from systemic in-vivo dosing · disclosure: all authors Altos Labs · PMID 41512832 · R34 recency add — first Altos Labs perspective on commercial deliverable ↩ ↩2

  10. doi:10.1038/s12276-026-01662-x · Liu YW, Zou JT, Gong JS et al. (Wang ZX group, Xiangya Hospital + Central South University) · in-vivo + in-vitro · Experimental & Molecular Medicine 2026 · 58(3):782–797 · intra-articular AAV-OSK in OA mouse models preserved chondrocyte vitality (no stemness-marker upregulation), reduced subchondral bone thickening, promoted cartilage hyalinization, reduced senescence + DNMT expression; TET2 identified as pivotal mediator · PMID 41786976; PMC13049178 (OA) · R34 recency add ↩

  11. doi:10.18632/aging.204896 · in-vitro · model: human replicatively senescent fibroblasts (n=3 per condition, p-adj < 0.05) · Yang JH, Petty CA et al., Aging (Albany NY) 2023 · PDF downloaded locally · six chemical cocktails (C1–C6, including CHIR-99021 [GSK-3α/β inhibitor], E-616452/RepSox [TGF-β inhibitor], valproic acid [HDAC inhibitor], tranylcypromine, forskolin, and others; no MEK inhibitor) reversed transcriptomic age in human senescent fibroblasts in vitro; primary readout: transcriptomic clock (RNA-seq based); cell identity preserved (no pluripotency markers induced); cocktails C1–C3 most effective, reducing transcriptomic chronological age by >3 years in four days needs-human-replication needs-replication ↩

  12. doi:10.1126/science.1239278 · in-vitro · model: mouse somatic cells · Hou P et al., Science 2013 · not locally downloaded (not_oa) · pluripotent stem cells induced from mouse fibroblasts using small-molecule cocktail only (no TF overexpression); proof-of-concept that TFs are dispensable for full reprogramming in mouse cells needs-human-replication ↩

  13. doi:10.1016/j.cell.2022.12.027 · in-vivo + in-vitro · model: C57BL/6J ICE mice (inducible DSB system; OSK delivered via AAV); human cells in vitro · Yang JH, Hayano M et al., Cell 2023 · PDF downloaded locally · ICE mice showed ~50% faster epigenetic age advance; OSK reversed up to ~57% of clock advance; provides causal evidence for epigenetic-disruption → accelerated aging direction needs-replication (ICE is an artificial induction system) ↩

  14. doi:10.1101/gad.173922.111 · in-vitro · model: senescent (74-year-old donor) and centenarian (92–101-year-old donors) human fibroblasts · Lapasset L et al., Genes & Development 2011 · PDF downloaded locally · full iPSC reprogramming with a six-factor cocktail (OCT4, SOX2, KLF4, c-MYC, NANOG, LIN28; OSKMNL) erased all detectable aging markers (telomere length, gene expression, mitochondrial function) in senescent/centenarian cells; standard four-factor OSKM did not efficiently reprogram senescent cells; foundational evidence that the epigenetic aging program is reversible via full pluripotency induction; note: full reprogramming, not partial reprogramming needs-replication ↩

  15. tyshkovskiy-2026-universal-transcriptomic-hallmarks · doi:10.1038/s41586-026-10542-3 · Nature 2026 · n=11,165 transcriptomes (4 species) incl. OSKM-reprogramming iPSC datasets · multi-species elastic-net/module-specific transcriptomic clocks; mixed-effects/ANOVA, P_adj<0.05 · model: human primary fibroblast → iPSC + multi-tissue · reprogramming reduced chronological + mortality tAge, strongest in EMT/MET module; rejuvenation signature overlaps CR + embryogenesis ↩

Graph View

Backlinks