In plain terms

The claim being assessed: During normal mammalian aging, progressive loss of cell-type-specific epigenetic regulatory information is a primary rate-limiting cause of organismal functional decline; selectively restoring the prior youthful regulatory state, while retaining cell identity and without genotoxic selection, dedifferentiation, tumorigenesis, unrelated repair effects, and proliferative effects, should reproducibly reverse established aging phenotypes and extend healthspan.

Where the evidence stands: Γ — Plausible mechanism. There is a believable biological story and partial support, but the causal step has not been demonstrated.

What this mechanism does: This makes the process worse or keeps it going, but something else starts it.

Strongest evidence for it: Epigenetic states are causally manipulable and functional reversal is demonstrated in constrained mouse tissues, most convincingly aged retina. The evidence does not isolate progressive information loss from damage responses or establish restoration of a defined youthful state as the mediator, primary rate-limiting cause, or broadly replicated healthspan intervention.

What decisive evidence would look like

In plain terms, researchers would change the proposed cause, confirm it changed as intended, and measure a meaningful health outcome while checking other explanations. The result should be able to support or count against the claim.

Relative explanatory fit: Alternative ranked ahead

What this means for epigenetic 'information loss' and ageing reversal

Cells change their patterns of gene regulation as organisms age, and experiments can alter some of those patterns. In limited mouse tissues, such manipulation has been associated with functional improvement. The much broader claim—that loss of epigenetic information is the primary cause of normal human ageing and that restoring a youthful state reverses ageing—has not been demonstrated.

Implications for health and treatment: The work is a research programme, not an available anti-ageing treatment. It motivates experiments that must separate genuine restoration from stress responses, cell selection, regeneration, or measurement changes, and then test durability and cancer risk. For now, it does not justify clinics or products claiming to reset biological age, reverse ageing, or replace prevention and care for diseases that become more common with age.

Limits for individual care: There is no agreed clinical measurement of the proposed lost information, no completed human study demonstrating organism-wide functional age reversal, and no long-term safety map for interventions that reprogramme cells. A change in an epigenetic clock or biomarker is not proof that a person has become younger or healthier. This page is not medical advice: do not pursue unproven reprogramming or gene-based interventions in place of established care, screening, vaccination, and risk-factor treatment.

Medical information notice: This review provides general health information, not medical advice. It cannot diagnose a condition or replace an individual discussion with a qualified clinician.

This assessment rests on 19 sources and 24 individually checked claims. 24 of those claims were verified against the original paper.

A low grade is not an accusation. It means the causal work has not been done yet, which is the ordinary state of most actively studied mechanisms in biology.

At a glance

Sub-claims assessed separately

ID	Sub-claim	Type	Level	Pivotal card
SC1	Cell-type-specific epigenetic regulatory patterns progressively change or become disordered with age.	existence	Β	—
SC2	Experimentally induced loss or scrambling of epigenetic information is sufficient to cause aging-like functional decline independently of genotoxic injury, cell loss, inflammation, selection, and mutation.	sufficiency	Γ	Card ↗
SC3	Partial reprogramming or another chromatin-state intervention reverses an established functional aging phenotype while retaining tissue identity.	reversibility	Ε	Card ↗
SC4	Functional benefit is mediated by restoration of the prior youthful epigenetic state rather than other transcriptional, regenerative, stress-response, proliferative, or selective effects.	mechanism	Γ	—
SC5	Epigenetic information loss is a primary rate-limiting cause, and selective restoration broadly reverses aging and extends healthspan.	primacy	Γ	—

Visual Evidence Cards · Pivotal Subclaims

Pivotal subclaim illustrations

Dark-theme simulated-micrograph renders and concise causal evidence summaries for the pivotal subclaims that govern this review's grade. Click any card to inspect the master 2400×2400 render.

Master 2400×2400 render ↗

Master 2400×2400 render ↗

Figures

Explanatory figures summarizing this review. They carry no evidential weight of their own; where a figure and the review differ, the review governs.

Full assessment

From here the review is written for a technical reader: caps and thresholds, competing models, the independence basis, and the citation ledger.

Provenance: 19 opened sources (17 Tier A primary + 2 Tier B correspondence); 24 load-bearing claims; 24 verified; 0 unverified; 0 contradicted. Search cutoff: 2026-09-30. The supplied packet was used as methodology, not evidence. This edition is the cross-model red-team synthesis of a single authored review; the grade is unchanged (Γ / High). See the red-team synthesis note near the end and redteam/REDTEAM_REPORT.md.

Adversarial review: Generated by Sol 6.1 (in-package provenance string: "OpenAI Codex with HSEF v5.4; cross-model red-team synthesis by Claude (Anthropic)"); adversarially reviewed by Claude Opus 5.5 (Anthropic) on 2026-10-14. 3 corrections applied (see ADVERSARIAL REVIEW AND APPROVAL). Status: Approved as revised.

AT A GLANCE

Verdict: Epigenetic dysregulation is causally manipulable and can reinforce aging phenotypes, but current evidence does not show that progressive loss of a defined information store is the primary cause of normal mammalian aging or that retrieval of that store mediates broad organismal reversal. The submitted claim is a conjunction, and each arm sits differently: the causation arm ("information loss causes aging") is Γ — a real in-vivo perturbation (ICE) that does not isolate information loss from genotoxic injury and does not establish primacy; the reversal arm ("restoring the youthful state reverses aging") is Ε for tissue-level reversal in normally aged mouse tissue (clearest in retina) but Γ for broad organismal reversal and for stored-state mediation. Scope-matched aggregation gates the whole-organism claim at Γ.

Plain-language summary

The hypothesis says aging is mainly a loss of the regulatory instructions that keep cells in a youthful state, and that restoring those instructions should reverse decline.

The evidence earns Γ: the mechanism is plausible and partly supported. Epigenetic patterns change with age, deliberate chromatin and reprogramming interventions can move those patterns, and several mouse experiments improved selected functions. The clearest result is restored retinal function in aged mice. Other studies are molecular only, use injury or progeroid models, are small or unreplicated, or report null functional outcomes and toxicity.

What remains unproven is the strong causal chain: that a defined store of epigenetic information progressively disappears during normal aging, that this loss is the primary rate-limiting cause, and that recovery of that information itself mediates broad healthspan reversal. Current interventions change many biological processes at once, and no completed human study demonstrates organismal aging reversal.

Claim boundary and preregistered decomposition

The submitted wording is a conjunction. It joins a primary-cause claim to a reversal and mediation claim. The gradeable form was locked before formal scoring:

During normal mammalian aging, progressive loss of cell-type-specific epigenetic regulatory information is a primary rate-limiting cause of organismal functional decline; selectively restoring the prior youthful regulatory state, while retaining cell identity and without genotoxic selection, dedifferentiation, tumorigenesis, unrelated repair effects, and proliferative effects, should reproducibly reverse established aging phenotypes and extend healthspan.

“Epigenetic information” therefore cannot mean whichever methylation site, histone mark, transcript, or cell population changes after treatment. A qualifying state must be specified before outcomes, measured in the relevant cell type, selectively perturbed, and linked to function by a mediation or rescue design. Molecular-age movement is target engagement unless function and mediation are also shown.

The aggregation rule was scope-matched. The broad statement is gated by SC5; a positive eye, brain, injury, cell-culture, worm, or progeroid result cannot lift the organismal natural-aging claim.

Subclaim scorecard

Scope-matched aggregation gives Γ through SC5. Weakest-link would give Β through the descriptive SC1; a primary-subclaim rule using SC5 also gives Γ. The one-rung alternative is reported rather than hidden.

COMPETING MECHANISTIC MODELS

1. Interacting hallmarks (rank 1; Ε). Epigenetic drift is a reciprocal causal component. Damage, senescence, inflammation, metabolism, proteostasis, stem-cell failure, and chromatin state can cause and reinforce one another.
2. Regeneration and state transition (rank 2; Ε). Partial reprogramming improves plasticity, repair, stress responses, extracellular matrix, or cell composition without retrieving a uniquely stored youthful state.
3. Epigenetic-information primacy (nominal; rank 3; Γ). Regulatory information is the primary lost variable and reprogramming recovers a backup copy.
4. Genotoxic trauma and survivorship (rank 4; Γ). ICE phenotypes arise from double-strand-break injury, inflammation, cell loss, or selection rather than isolated information loss. This is the aging-arm null model named in the hypothesis index, and it is now a matter of published record: a peer-reviewed critique and the authors' response contest exactly this point [18,19].
5. Clock and composition (rank 5; Γ). Molecular-age scores partly track cell mixture and physiological state, so resetting them need not reverse aging causes.

The interacting-hallmarks model best fits the combined positive, null, and toxic results. The regeneration model explains why pleiotropic developmental factors can improve selected functions while broad primacy and mediation remain unproven. The nominal model explains the direction of ICE and retinal findings, but it lacks the selective perturbation and rescue that would distinguish recovery of stored information from a newly induced compensatory state. This tournament informs the role classification; it does not alter the evidence letter.

The clean separator between ranks 1 and 2 would be a lineage-resolved, double-strand-break-free epigenome rewrite in naturally aged animals that restores a prespecified state, reverses an established functional deficit, loses that benefit when the state is blocked, and regains it when the same state is restored again.

HYPOTHESIS STATE EVALUATION FRAMEWORK (HSEF v5.4)

1. What “loss of information” currently demonstrates

Age-associated methylation, chromatin, and transcriptional changes are real and reproducible [1,2,4,5]. They support SC1 at Β. They do not show that a particular pattern stores the causal instructions for youth, that information rather than cell composition or damage response has been lost, or that the pattern is rate-limiting.

This distinction matters because a predictor can report age accurately without causing age. A DNA-methylation clock may move after an intervention because cell state, cell mixture, inflammation, proliferation, or metabolism moved. A 2026 Mendelian-randomization analysis of epigenetic-age-acceleration proxies found no consistent healthspan association, although those proxies are not the same thing as the proposed information store [17].

2. The ICE sufficiency experiment

ICE is the most direct attempt to make the causal claim. I-PpoI recognizes 20 mouse-genome targets, 19 non-coding and including rDNA. Transient induction produced double-strand breaks, delayed aging-like physiological changes, faster methylation clocks, altered chromatin, and impaired function months later [1]. The authors detected no residual cutting at one or ten months and no increase in mutations at assayed sites or genome-wide. These controls make a simple persistent-nuclease or mutational explanation less likely.

They do not isolate epigenetic information. ICE starts with whole-body genotoxic stress. Direct exclusion of apoptosis, cell-cycle change, and early senescence was performed in MEFs; the full article does not report the corresponding in-vivo cell-death, inflammatory, or clonal-selection analysis. Cell autonomy was not established, and the authors could not exclude rDNA cutting as a contributor. OSK rescue in the ICE report mainly moved molecular and histological readouts rather than restoring whole-organism function. The 2024 correction clarifies the tamoxifen regimen used to limit I-PpoI activity and genotoxic stress; it does not supply the missing in-vivo specificity controls [1].

This concern is not only a reviewer inference: it is the substance of a published, peer-reviewed exchange. A 2024 Matters Arising argues that the information theory of aging has not been experimentally isolated and that the ICE phenotype may reflect the response to induced DNA damage and survivorship rather than loss of a defined epigenetic information store [18]; the authors' response defends the interpretation and the sufficiency of the reported controls [19]. The exchange is unresolved. It maps directly onto the genotoxic-trauma-and-survivorship model (competing model 4) and reinforces, rather than changes, the specificity ceiling. SC2 therefore remains Γ.

3. Functional reversal: real, but bounded

The strongest experiment is the retina. Four weeks of AAV-OSK restored visual acuity and retinal-ganglion-cell electrophysiology in mice treated at 11 months and tested at 12 months [2]. Retinal cell or axon number did not increase, supporting improvement in existing cells. Tet1 or Tet2 knockdown blocked the functional benefit and counteracted molecular-age reversal. This is genuine causal functional evidence in a normally aged tissue, and it justifies Ε for the mouse-tissue SC3.

The experiment does not complete the submitted causal chain. OSK broadly changes transcription and chromatin; TET dependence shows that demethylation participates, but not that recovery of a defined stored state mediates the benefit. TET1 catalytic activity alone did not reproduce the full regenerative effect. Benefit was not measured at 18 months because corneal opacity interfered, which leaves advanced-age durability unresolved rather than establishing a clean negative result [2].

Other functional findings add breadth but not decisive convergence:

• Cyclic OSKM improved a laminopathy progeroid phenotype and injury responses, while stronger exposure caused weight loss, mortality, or teratomas [3].
• One small AAV9-OSK experiment in 124-week-old male mice reported median lifespan near 133 versus 142.5 weeks and remaining life of 8.86 versus 18.5 weeks, with better frailty [6]. It had 20 animals per arm, lacked a contemporaneous AAV control, selected healthy survivors for clock assays, and remains unreplicated.
• Targeting OSK to Cdkn2a-positive cells improved mouse aging-model outcomes, but only abstract-level verification was available here [7].
• Neuron-restricted cyclic expression from 6 to 11 months improved memory in eight control and nine treated mice; treatment began in middle age, and continuous expression from birth caused about 60% mortality [9].
• EZH2 overexpression improved aged-liver steatosis, fibrosis, and glucose tolerance after a short intervention, showing that chromatin manipulation can alter function without identifying a stored backup state [14].

These results establish intervention efficacy in constrained mouse contexts. They do not establish broad organismal reversal or primacy.

4. Molecular rejuvenation is not the same endpoint

Long-term cyclic OSKM during physiological mouse aging shifted selected methylation clocks and improved epidermal proliferation and skin fibrosis [4]. Gross wound closure did not notably improve, muscle regeneration was not significantly changed, open-field activity did not improve, and survival did not differ. A separate one-cycle study shifted methylomic, transcriptomic, and serum-metabolite profiles in old mice but did not test healthspan function and observed transient pancreatic histological changes during induction [5].

In human fibroblasts, maturation-phase transient reprogramming shifted molecular-age measures by roughly three decades and restored some cell properties [8]. The experiment used three donors, selected the approximately one-quarter of cells reaching the intended intermediate state, and found migration improvement in only two of four comparisons. This is useful cellular evidence, but it is not reversal of aging in people.

The submitted wording requires functional aging phenotypes. Clocks, transcriptomes, histone marks, metabolomes, collagen expression, and selected cell assays cannot be counted again as organismal rejuvenation unless they mediate a prespecified functional outcome.

5. Negative evidence and safety boundaries

The negative results are mechanistically informative rather than side notes. Continuous systemic OSKM can cause rapid weight loss, liver and intestinal failure, and premature death [10]. Other in-vivo reprogramming schedules can produce teratomas or dysplastic tumours when identity control fails [11,12]. In genetically heterogeneous male mice, a tolerated chemical-reprogramming dose did not improve liver or kidney transcriptomic age, while higher doses caused weight loss, lipid accumulation, and kidney-injury signals [13].

These studies do not show that carefully titrated local OSK is always unsafe. They show that “restoring youth” is not a unitary beneficial operation: outcome depends on factor set, dose, timing, cell type, tissue distribution, baseline age, and identity control. That context dependence contradicts an unqualified broad reversal clause.

A local AAV-OSK study in surgically induced osteoarthritis improved cartilage and fibrosis in young-adult mice [15]. A chemical cocktail extended median worm lifespan by 42% [16]. Both are useful leads; neither tests natural mammalian aging or proves stored-information mediation.

6. Mediation and primacy

For SC4 to cross Δ, a study must selectively restore the prespecified regulatory state while holding broad developmental-factor effects constant, then show that blocking that restoration blocks function and restoring it again rescues function. No reviewed study does this. TET knockdown is the best mechanistic narrowing, but TET enzymes regulate many methylation and transcriptional processes. ICE moves aging-like phenotypes forward with DNA damage; OSK moves selected outcomes backward with a pleiotropic developmental programme. Symmetry of direction is suggestive, not a formal mediation test.

For SC5, “primary” additionally requires necessity, sufficiency, and rate limitation across normal aging. Current evidence does not show that preventing the proposed information loss substantially delays organismal aging, that selective loss initiates normal aging without damage, or that selective restoration dominates other interventions. Epigenetic dysregulation therefore fits reinforcing / stabilizing factor under direction-procedure step 5.

7. Overall adjudication

The submitted broad proposition is Γ — Mechanistically plausible, partial system support, with High confidence. The high confidence concerns the present ceiling, not a claim that epigenetic regulation is unimportant. The evidence consistently shows causal manipulability plus a recurring specificity and generalization gap.

The confidence attaches to the ceiling because a large, heterogeneous body of primary work — positive, mixed, null, and toxic — converges on the same missing pieces: no selective, damage-free information-loss experiment, no measured stored-state mediation, no replicated cross-tissue or cross-species healthspan result, and no human efficacy. It is worth naming a countervailing limitation on the positive results, which is a distinct axis: several of the strongest positive findings come from a small, overlapping set of groups (the ICE and retinal-OSK work from one lab lineage; several partial-reprogramming studies from another), so independent replication is thin (pitfall 9, replication confused with reproduction). That concentration is why the subclaim confidences on the positive arms (SC2, SC3, SC4) sit at Moderate even while the ceiling on the broad claim is held with High confidence.

The strongest defensible narrower statement is:

Age-associated epigenetic and transcriptional states are causally manipulable, and partial reprogramming can improve selected functional phenotypes in normally aged mouse tissues; whether recovery of a defined youthful epigenetic information state mediates broad healthspan reversal remains unproven.

8. Steelman

One rung higher (Δ): ICE produced delayed multisystem decline without detected persistent cuts or mutation; retinal OSK restored an established function and required TET enzymes; several interventions improved mouse cognition, frailty, survival, liver function, or repair; and more than one intervention class changed aging-relevant outcomes. This is more than association.

One rung lower (Β): The information construct lacks an agreed independent measurement, ICE starts with DNA damage, most restoration experiments express pleiotropic developmental factors, many endpoints are molecular clocks or omics, the broad lifespan result is small and unreplicated, and human efficacy is absent.

Decision: Δ overall is blocked by the overexpression, single-species, biomarker, and specificity caps on mediation and primacy, plus the missing selective-state and replicated broad-healthspan experiment. Β overall is rejected because controlled functional interventions clearly exceed association within SC3 and the broad causal idea has partial system support.

9. Momentum, stability, and divergence

• Momentum — Mechanistic only. A registered Phase 1 local ocular OSK study, NCT07290244, and active targeted-reprogramming screens address delivery and tissue function. They do not test systemic normal-aging primacy or broad healthspan.
• Stability — Uncertain. Independent, vector-controlled replication of late-life survival and healthspan would raise the score. Failure, toxicity, or lack of functional benefit in that replication would weaken broad reversal while leaving the retinal result intact.
• Divergence — Severe overstatement. Field-facing primary-cause and reversal wording implies Η; assessed Γ; gap +4 rungs.

10. Decisive experiments

1. Selective information-loss sufficiency. In adult mice, scramble a preregistered cell-type regulatory circuit with a double-strand-break-free epigenome editor and no developmental transcription factors. Include inactive editors, neutral loci, lineage and clonal tracking, mutation and damage assays, inflammation and cell-death measures, reverse editing, and functional rescue.
2. Assignment-preserving mediation. In old animals with baseline deficits, compare OSK with a selective rewrite that recreates the same prespecified state, plus OSK in which that state change is blocked. Functional benefit must follow the state, disappear on blockade, and reappear on rescue; clocks remain secondary.
3. Cross-species late-life healthspan replication. Replicate a fixed regimen in multiple mouse strains and a second mammalian species with vector controls, both sexes, complete necropsy, all-animal molecular sampling, tumour surveillance, disability-free survival, and organ-function endpoints.

11. Knowledge gaps

• A noncircular operational definition of epigenetic information that distinguishes causal state from age association, cell composition, and downstream damage response.
• Identification of where a youthful backup state is stored and whether reprogramming retrieves it or creates a compensatory state.
• In-vivo exclusion of apoptosis, inflammation, clonal selection, rDNA effects, and non-cell-autonomous injury in ICE.
• Quantitative mediation of function by a measured epigenetic state rather than regeneration, stress response, proliferation, or cell selection.
• Independent replication of late-life survival and cross-tissue healthspan in both sexes, multiple backgrounds, and a second mammalian species.
• Long-term durability, identity stability, organ toxicity, and neoplasia surveillance.
• Completed human functional efficacy evidence; none was identified.

12. Methods, access depth, and corrections

The supplied HSEF v5.4 packet was treated as methodology and not as evidence. The claim, five subclaims, scope-matched aggregation, functional-endpoint rule, and competing models were fixed in the accompanying preregistration before formal scoring. The primary rubric was metabolism/aging; epigenetics/gene-regulation supplied the specificity rules.

All 24 load-bearing claims were checked against 19 opened sources - 17 Tier A primary reports plus the two Tier B correspondence items added in the red-team pass. Of the 17 Tier A records, twelve primary reports were read in full, one publisher article was read partially at the Results and Discussion level, and four were checked at abstract and metadata depth; both Tier B items were read at bibliographic/status depth. Access depth is declared for every ledger record. Full text was used wherever it was lawfully accessible in this environment.

PubMed status metadata was checked on 2026-09-30. The 2024 Cell correction is attached to the ICE parent record and not counted as an independent study. The published 2026 Aging Cell chemical-reprogramming toxicity article replaces its 2025 bioRxiv version for adjudication. Scite was unavailable, so no support, mention, or contrast counts were inferred.

This was not a blinded calibration run. No independently authored sealed target was supplied.

Red-team synthesis note

This edition is a cross-model red-team pass: the review was authored by one model family and adversarially re-examined and synthesized by a different one. The audit confirmed the original adjudication and changed no score. Every reference was re-resolved against PubMed (title, first author, year, journal, volume, pages); all 17 primary sources and the 2024 erratum matched, the ER-100 trial identifier resolved, and two load-bearing numeric sets (retinal restoration and the late-life survival figures) were re-verified against full text and found exact — no confabulated identifiers were present. Two substantive changes were made, both currency and completeness rather than scoring: (1) the peer-reviewed published critique of the keystone ICE study and the authors' response were added [18,19] and woven into the ICE section and competing model 4, where the review had already reached the same conclusion independently; and (2) the confidence discussion now names the lab-concentration limitation on the positive evidence (pitfall 9). The grade, subclaim levels, aggregation, caps, role of mechanism, and M/S/D are unchanged. Full detail is in redteam/REDTEAM_REPORT.md and redteam/SYNTHESIS_CHANGE_LOG.md.

ADVERSARIAL REVIEW AND APPROVAL

Authoring model: Sol 6.1 (the package's own provenance string reads "OpenAI Codex with HSEF v5.4; cross-model red-team synthesis by Claude (Anthropic); validator 2.2"; that string is house provenance and was not altered). Adversarial reviewer: Claude Opus 5.5 (Anthropic). Date: 2026-10-14.

Standing of the earlier approval. The only earlier adversarial record for this slug is the prepublication-enrichment log of 2026-10-06, which approved the dated research addendum for this case (9/9 identifiers, 0 findings) and whose stated scope explicitly excludes the posted review editions. No earlier approval of this review edition exists. The 9 October 2026 reconciliation spliced in a template clause implying otherwise; that clause has been rescoped (finding E2), and the present section is this edition's first adversarial adjudication.

What the 2026-10-09 revision changed. The revision added one block — the "Consensus provenance reconciliation" section inside the framework chapter — and changed nothing in the assessment: the grade (Γ), confidence (High), role of mechanism (reinforcing / stabilizing factor), the five subclaim levels, the scope-matched aggregation, the caps, M/S/D, the reference set and the counts are all as in the red-team synthesis edition. No revision-introduced substantive regression was found. The block's own disclaimers (that provider summaries do not establish examination of original passages, that no comparison count is inferred, and that shared template wording about an amyloid-specific follow-up is not credited to this topic) are accurate and were retained.

Scope of checks. All 19 references were re-verified against PubMed metadata by PMID where printed and by DOI where the review cites a PMC record: 19/19 resolve to the stated title, first author, journal, year, volume and pages, and 19/19 DOIs match. No reference is retracted or carries an expression of concern. The one linked notice is real and correctly paired: reference 1 (Yang 2023, Cell 186(2):305-326.e27, PMID 36638792) carries the published erratum PMID 38428398 at Cell 187(5):1312-1313, exactly as the review states. The two Tier B items are correctly tiered and labelled: PMID 38428390 (Timmons and Brenner, Cell 187(5):1101-1102) and PMID 38428391 (Yang and colleagues' response, 1103-1105) are both indexed as Letters. The 2026 chemical-reprogramming toxicity article (PMID 41589348, Aging Cell 25(2):e70390) does update the 2025 bioRxiv preprint PMID 40667171, as recorded, and the preprint is not double-counted. Nine load-bearing claims were checked against the original abstracts: the ICE keystone result; OSK restoring youthful methylation and vision with TET1/TET2 dependence (reference 2); progeroid lifespan extension and wild-type injury recovery under cyclic OSKM (reference 3); multi-omic reversal in naturally aged tissues after a single transient cycle (reference 5); the 124-week-old AAV-OSK cohort with a 109% increase in median remaining lifespan and improved frailty (reference 6); premature death with hepatic and intestinal failure after systemic reprogramming (reference 10); teratoma and cancer outcomes from uncontrolled or prematurely terminated reprogramming (references 11 and 12); the EZH2 single-factor liver rejuvenation screen (reference 14); and C. elegans-only lifespan extension from chemical reprogramming (reference 16). All nine were supported as stated, including the review's insistence that the lifespan evidence is progeroid, worm or single-cohort rather than replicated normal-aging mammalian healthspan.

Rejected findings (checked, no change).

1. Aggregation. Scope-matched is preregistered and applied correctly. SC5 is the subclaim whose scope matches the central organismal claim and it returns Γ; SC3's Ε is explicitly scope-restricted to normally aged mouse tissue with functional endpoints and is reported separately rather than lifting the headline. The review states the rule in the claim-boundary section and reports weakest-link (Β) and primary-subclaim (Γ) as alternatives. This is the correct reading of the rule, and the validator's logic.aggregation check agrees.
2. Confidence High with a Γ letter. The rationale rests on evidence density and reproducibility - many primary studies across positive, mixed, null and toxicity results converging on the same missing specificity, mediation, replication, cross-species and human bridges - not on access depth or on reviewer belief. That is the framework's intended use of the band, and the framework explicitly allows a Γ with high confidence.
3. Role of mechanism "reinforcing / stabilizing factor" at derivation step 5. Steps 1 to 4 fail in order: the mechanism is not merely produced by the phenotype, a dependency is demonstrated in at least one direction, it is not shown to be necessary, sufficient and rate-limiting, and it is not shown to be required as a permissive condition. Step 5 is the first match.
4. Divergence "Severe overstatement" with a four-rung gap (field-facing language implying Η against an assessed Γ). The arithmetic is shown and M/S/D are stated not to change the letter.
5. Momentum "Mechanistic only". The artifact names the ER-100 Phase 1 local ocular OSK safety study and active preclinical targeted-reprogramming and transcription-factor screens, and states that none tests systemic normal-aging primacy or broad healthspan. A Phase 1 ocular safety study cannot cross the organismal boundary, so "Decisive work underway" would be wrong here; "Mechanistic only" is correct.
6. Counting the two Tier B correspondence items as sources. They are tiered B, labelled as Matters Arising and response, used only to establish that the genotoxic-trauma alternative is a matter of published record, and excluded from the Tier A primary count. Correct handling.
7. "Scite was unavailable, so no support, mention, or contrast counts were inferred." Accurate as a stated limitation, and not contradicted by the spliced consensus block, which concerns Consensus browser artifacts rather than Scite.
8. Renderer defect that prints the nominal model as its own "Best-supported rival". Not applicable: this edition presents the tournament as ranked prose with the nominal explicitly labelled rank 3, and models_evaluated[0] is the rank-1 non-nominal model.
9. Builder whitespace defect. Checked and absent: the only candidate matches in this document are correct tokens (Tet1, Tet2, Cdkn2a, article e-locators such as 305-326.e27, e13578 and eadg1777), so no whitespace repair was applied.

Residual limitations. Twelve Tier A records were read in full, one partially and four at abstract/metadata depth, and both Tier B items at bibliographic depth; the ER-100 trial identifier and the two numeric sets re-verified in the red-team pass were confirmed there rather than re-extracted here. Scite was unavailable to the original review, so no citation-context counts exist. The scorecard's review_metadata.reviewer and provenance_header.generated_by strings are house provenance and were not edited. A separate byte-identical copy of this edition exists outside the audit folder at reviews/information-theory-aging/final/; it is listed in the final report for the lead to keep in step.

Validator. The frozen HSEF v5.3 validator (tool version 2.3) was re-run on the scorecard and the revised prose: 34 pass, 0 warn, 0 fail. This is the only package in this group that clears the validator outright, because it is a v5.4 artifact and carries the lay_summary that the tool requires from v5.4 onward. Full output is in validator.txt.

Approval: Following adversarial review and the corrections listed above, the revised document is approved. - Claude Opus 5.5 (Anthropic), 2026-10-14.

REFERENCES

1. Yang JH, Hayano M, Griffin PT, et al. Loss of epigenetic information as a cause of mammalian aging. Cell. 2023;186(2):305-326.e27. PMID 36638792. Record. DOI: 10.1016/j.cell.2022.12.027. Access: full_text_complete — Complete PMC article, figures, methods, supplement captions, limitations, and correction metadata read. Status: erratum: Cell. 2024;187(5):1312-1313; PMID 38428398; tamoxifen-delivery sentence corrected.
2. Lu Y, Brommer B, Tian X, et al. Reprogramming to recover youthful epigenetic information and restore vision. Nature. 2020;588(7836):124-129. Record. DOI: 10.1038/s41586-020-2975-4. Access: full_text_complete — Complete PMC article, figures, extended data, sample sizes, methylation analyses, and TET-dependence experiments read. Status: none_found.
3. Ocampo A, Reddy P, Martinez-Redondo P, et al. In Vivo Amelioration of Age-Associated Hallmarks by Partial Reprogramming. Cell. 2016;167(7):1719-1733.e12. Record. DOI: 10.1016/j.cell.2016.11.052. Access: full_text_complete — Complete PMC article, progeroid survival, wild-type injury studies, dosing, pathology, and safety results read. Status: none_found.
4. Browder KC, Reddy P, Yamamoto M, et al. In vivo partial reprogramming alters age-associated molecular changes during physiological aging in mice. Nat Aging. 2022;2(3):243-253. Record. DOI: 10.1038/s43587-022-00183-2. Access: full_text_complete — Complete article PDF, main and extended figures, wound, muscle, activity, clock, and pathology results read. Status: none_found.
5. Chondronasiou D, Gill D, Mosteiro L, et al. Multi-omic rejuvenation of naturally aged tissues by a single cycle of transient reprogramming. Aging Cell. 2022;21(3):e13578. Record. DOI: 10.1111/acel.13578. Access: full_text_complete — Complete PMC article, cohorts, methylome, transcriptome, metabolome, pathology, and limitations read. Status: none_found.
6. Macip CC, Hasan R, Hoznek V, et al. Gene Therapy-Mediated Partial Reprogramming Extends Lifespan and Reverses Age-Related Changes in Aged Mice. Cell Reprogram. 2024;26(1):24-32. Record. DOI: 10.1089/cell.2023.0072. Access: full_text_complete — Complete PMC article and high-resolution survival figure read, including cohort size, controls, survival, frailty, clock selection, sex restriction, and limitations. Status: none_found.
7. Sahu SK, Reddy P, Lu J, et al. Targeted partial reprogramming of age-associated cell states improves markers of health in mouse models of aging. Sci Transl Med. 2024;16(764):eadg1777. Record. DOI: 10.1126/scitranslmed.adg1777. Access: abstract_metadata — PubMed abstract and complete bibliographic/status record read; publisher full text was inaccessible. Status: none_found.
8. Gill D, Parry A, Santos F, et al. Multi-omic rejuvenation of human cells by maturation phase transient reprogramming. eLife. 2022;11:e71624. Record. DOI: 10.7554/eLife.71624. Access: full_text_complete — Complete PMC article, donor count, cell sorting, clocks, collagen, migration, and limitations read. Status: none_found.
9. Anton-Fernandez A, Roldan-Lazaro M, Valles-Saiz L, et al. In vivo cyclic overexpression of Yamanaka factors restricted to neurons reverses age-associated phenotypes and enhances memory performance. Commun Biol. 2024;7(1):631. Record. DOI: 10.1038/s42003-024-06328-w. Access: full_text_complete — Complete PMC article, induction timing, behavioral groups, molecular results, mortality, and pathology read. Status: none_found.
10. Parras A, Vilchez-Acosta A, Desdin-Mico G, et al. In vivo reprogramming leads to premature death linked to hepatic and intestinal failure. Nat Aging. 2023;3(12):1509-1520. Record. DOI: 10.1038/s43587-023-00528-5. Access: full_text_partial — PubMed record plus publisher-visible Results and Discussion sections on mortality, weight loss, liver and intestinal dysfunction, and tissue-restricted rescue read. Status: none_found.
11. Ohnishi K, Semi K, Yamamoto T, et al. Premature termination of reprogramming in vivo leads to cancer development through altered epigenetic regulation. Cell. 2014;156(4):663-677. Record. DOI: 10.1016/j.cell.2014.01.005. Access: abstract_metadata — PubMed abstract and complete bibliographic/status record read. Status: none_found.
12. Abad M, Mosteiro L, Pantoja C, et al. Reprogramming in vivo produces teratomas and iPS cells with totipotency features. Nature. 2013;502(7471):340-345. Record. DOI: 10.1038/nature12586. Access: abstract_metadata — PubMed abstract and complete bibliographic/status record read. Status: none_found.
13. Mitchell W, de Magalhaes CG, Tyshkovskiy A, et al. In Vivo Chemical Reprogramming Is Associated With a Toxic Accumulation of Lipid Droplets Hindering Rejuvenation. Aging Cell. 2026;25(2):e70390. Record. DOI: 10.1111/acel.70390. Access: full_text_complete — Complete PMC final article, dose groups, transcriptomics, mitochondrial assays, histology, kidney injury, and toxicity read. Status: updates bioRxiv PMID 40667171; no correction or retraction found.
14. Sengstack J, Zheng J, Aghayev T, et al. Systematic identification of single transcription factor perturbations that drive cellular and tissue rejuvenation. Proc Natl Acad Sci U S A. 2026;123(2):e2515183123. Record. DOI: 10.1073/pnas.2515183123. Access: full_text_complete — Complete PMC article, screen, aged-liver intervention, functional and molecular endpoints, sample sizes, and limitations read. Status: none_found.
15. Liu YW, Zou JT, Gong JS, et al. Local delivery of OSK factors enables partial cellular reprogramming to mitigate osteoarthritis and cartilage fibrosis. Exp Mol Med. 2026;58(3):782-797. Record. DOI: 10.1038/s12276-026-01662-x. Access: full_text_complete — Complete PMC article, young-adult surgical models, vector controls, group sizes, methylation, cartilage, and fibrosis results read. Status: none_found.
16. Schoenfeldt L, Paine PT, Pico S, et al. Chemical reprogramming ameliorates cellular hallmarks of aging and extends lifespan. EMBO Mol Med. 2025;17(8):2071-2094. Record. DOI: 10.1038/s44321-025-00265-9. Access: full_text_complete — Complete PMC article, human-cell assays, worm lifespan and healthspan experiments, dosing, and limitations read. Status: none_found.
17. Feng B, Yang R, Wang GR, et al. Genetically Proxied Leukocyte Telomere Length and Epigenetic Age Acceleration in Relation to Healthspan: A Mendelian Randomization Study. Ann Hum Genet. 2026 Sep 10. Online ahead of print. Record. DOI: 10.1111/ahg.70057. Access: abstract_metadata — PubMed abstract and complete bibliographic/status record read; no full text was accessible. Status: none_found.
18. Timmons JA, Brenner C. The information theory of aging has not been tested. Cell. 2024;187(5):1101-1102. PMID 38428390. Record. DOI: 10.1016/j.cell.2024.01.013. Tier B (Matters Arising / correspondence). Access: abstract_metadata — PubMed bibliographic/status record read; the letter contests whether the ICE experiment isolates loss of epigenetic information from genotoxic injury and survivorship. Status: none_found.
19. Yang JH, Hayano M, Rajman LA, Sinclair DA. Response to: The information theory of aging has not been tested. Cell. 2024;187(5):1103-1105. PMID 38428391. Record. DOI: 10.1016/j.cell.2024.01.014. Tier B (correspondence). Access: abstract_metadata — PubMed bibliographic/status record read; the authors' response defends the interpretation and the reported controls. Status: none_found.

Provenance check: 19 sources (17 Tier A primary + 2 Tier B correspondence); 24 load-bearing claims; 24 verified; 0 unverified; 0 contradicted. Generated with HSEF v5.4; cross-model red-team synthesis edition (2026-09-30); validated with validator 2.2.

Documentary record

Every grade on this site is reproducible from its record. The files below are the posted edition exactly as assessed.

• ADVERSARIAL_REVIEW.md
• SHA256.txt
• adversarial-findings.json
• codex-correction-provenance.json
• consensus-addendum.md
• consensus-validation.json
• edition-approval-audit.json
• information-theory-aging.review.md
• information-theory-aging.scorecard.json
• research-update-2026-10-05.bib
• research-update-2026-10-05.html
• research-update-2026-10-05.json
• research-update-2026-10-05.md
• research-update-2026-10-05.ris
• validation-report.json
• validation-round2-2026-10-14.json
• validation.txt
• video-insert-2026-10-05.md

Sources: 19 · load-bearing claims: 24 · verified: 24 · unverified: 0 · contradicted: 0 · schema 5.4