Interventions by hallmark — matrix

This is the wiki’s primary deliverable for the query: “What interventions target [hallmark], and what stage is each at?”

This page is a navigational overlay / synthesis MOC — all quantitative claims and mechanistic details live on atomic compound, intervention, and hallmark pages. This page links, ranks, and contextualises; it does not duplicate primary facts.


Self-monitoring watchdog blocks

These three Dataview blocks surface drift whenever this page is opened. They should be reviewed at every seeding batch and lint pass.

Watchdog 1 — Intervention count per hallmark (flag zeros)

TABLE WITHOUT ID rows.file.link AS "Interventions", length(rows) AS "Count"
FROM "molecules/compounds" OR "interventions"
FLATTEN hallmarks AS h
GROUP BY h
SORT length(rows) ASC

If any hallmark shows zero linked interventions, it must be flagged as a gap. The Dataview result is canonical because hard-coded counts can drift. As of 2026-08-02, telomere-attrition and stem-cell-exhaustion both have linked intervention pages; the tASO ingest added a new direct preclinical class to each.

Watchdog 2 — Stale clinical-stage entries

TABLE WITHOUT ID file.link AS Compound, clinical-stage AS Stage, verified-date AS "Verified", translation-gap AS "Gap"
FROM "molecules/compounds" OR "interventions"
WHERE contains(["phase-2","phase-3","phase-2-trials","phase-4","fda-approved"], string(clinical-stage))
  AND (date(verified-date) < date(today) - dur(180 days) OR !verified-date)
SORT verified-date ASC

Clinical-stage data decays quickly. Any row returning here should trigger a ClinicalTrials.gov re-check per sops/finding-aging-specific.md.

Watchdog 3 — R14 fields missing

LIST FROM "molecules/compounds" OR "interventions"
WHERE !translation-gap OR !next-experiment

Should return empty after R14 close (2026-05-05). Any result here means a compound or intervention page is incomplete and will not render correctly in the matrix below.


How to read the matrix

  • Rows = 12 hallmarks of aging (López-Otín 2023 framework).
  • Per-hallmark Dataview block pulls all compound/intervention pages that link to that hallmark via hallmarks: or target-hallmarks: frontmatter.
  • Columns: Compound | Mechanisms | Stage | Evidence level | Translation gap | Next experiment.
  • Mechanistic-tier and intervention-tractability are hallmark-level properties from the hallmark MOC frontmatter.
  • The class-level synthesis commentary after each block is original synthesis, not duplicating atomic-page claims.

Note on field asymmetry: type: compound pages use hallmarks: for hallmark linkage; type: intervention pages use target-hallmarks:. Until field normalisation, the per-hallmark Dataview blocks below use separate WHERE clauses joined by OR to capture both. This is a known schema inconsistency flagged for a future cleanup pass.


Cellular senescence

Mechanistic tier: intermediate | Intervention tractability: high

TABLE WITHOUT ID file.link AS Compound, mechanisms AS Mechanism, clinical-stage AS Stage, human-evidence-level AS "Evidence", translation-gap AS "Gap", next-experiment AS "Next experiment"
FROM "molecules/compounds" OR "interventions"
WHERE contains(hallmarks, [[cellular-senescence]])
  OR contains(target-hallmarks, [[cellular-senescence]])
SORT clinical-stage DESC

Class-level synthesis: This is the best-resourced hallmark for pharmacological intervention. Senolytics (dasatinib+quercetin, fisetin, navitoclax, A1331852, quercetin) and senomorphics (rapamycin, canakinumab via SASP suppression) converge here. The D+Q combination has the strongest human proof-of-target-engagement (Hickson 2019, n=9: −35% p16 mRNA in human tissue). However, no senolytic has yet reached a hard-endpoint Phase 3 RCT. The BCL-xL-selective PROTAC class (DT2216, PZ15227) is the highest-priority preclinical advance — it would resolve the navitoclax thrombocytopenia blocker. Translation gap: phase-3-rct-needed dominates. See senolytics, cellular-senescence.


Chronic inflammation (inflammaging)

Mechanistic tier: integrative | Intervention tractability: high

TABLE WITHOUT ID file.link AS Compound, mechanisms AS Mechanism, clinical-stage AS Stage, human-evidence-level AS "Evidence", translation-gap AS "Gap", next-experiment AS "Next experiment"
FROM "molecules/compounds" OR "interventions"
WHERE contains(hallmarks, [[chronic-inflammation]])
  OR contains(target-hallmarks, [[chronic-inflammation]])
SORT clinical-stage DESC

Class-level synthesis: This hallmark has the highest-quality human evidence of any inflammaging-targeted class — CANTOS (canakinumab, n=10,061, Phase 3 RCT) demonstrated MACE reduction via IL-1β blockade. The primary translation gap is cost (translation-blocked-cost for canakinumab). The anti-inflammatory convergence of multiple compound classes here (metformin, egcg, quercetin, taurine, rapamycin, urolithin-a, senolytics, senomorphics, caloric restriction) reflects the integrative nature of inflammaging: upstream damage signals from multiple hallmarks funnel into this node. anti-il-11-antibodies add a preclinical ligand/receptor-targeted class with late-life mouse healthspan and survival evidence but no human aging trial. See chronic-inflammation, senomorphics.


Deregulated nutrient sensing

Mechanistic tier: intermediate | Intervention tractability: high

TABLE WITHOUT ID file.link AS Compound, mechanisms AS Mechanism, clinical-stage AS Stage, human-evidence-level AS "Evidence", translation-gap AS "Gap", next-experiment AS "Next experiment"
FROM "molecules/compounds" OR "interventions"
WHERE contains(hallmarks, [[deregulated-nutrient-sensing]])
  OR contains(target-hallmarks, [[deregulated-nutrient-sensing]])
SORT clinical-stage DESC

Class-level synthesis: The densest intervention cluster in the wiki. Four distinct drug classes converge — mTOR inhibitors (rapamycin), AMPK activators (metformin), NAD precursors (NMN, NR), and dietary restriction (caloric-restriction). Rapamycin has ITP-validated mouse lifespan extension; metformin’s TAME trial (~2028) is the landmark human test. NAD+ precursors reliably raise blood NAD+ but have been null on aging endpoints in Phase 2 RCTs to date. Caloric restriction is the only intervention with DunedinPACE clock-slowing evidence in humans (CALERIE, n=197). See deregulated-nutrient-sensing.


Disabled macroautophagy

Mechanistic tier: intermediate | Intervention tractability: high

TABLE WITHOUT ID file.link AS Compound, mechanisms AS Mechanism, clinical-stage AS Stage, human-evidence-level AS "Evidence", translation-gap AS "Gap", next-experiment AS "Next experiment"
FROM "molecules/compounds" OR "interventions"
WHERE contains(hallmarks, [[disabled-macroautophagy]])
  OR contains(target-hallmarks, [[disabled-macroautophagy]])
SORT clinical-stage DESC

Class-level synthesis: Multiple interventions induce autophagy by distinct routes: mTOR inhibition (rapamycin), AMPK activation (metformin, egcg), polyamine EP300 inhibition (spermidine), and direct mitophagy induction (urolithin-a). Epistatic evidence from model organisms shows autophagy is a required downstream effector for CR, rapamycin, and IIS-reduction longevity — the strongest mechanistic convergence in the wiki. Human evidence is weakest here: urolithin-a’s ATLAS RCT missed its primary endpoint; spermidine has only observational human data. See disabled-macroautophagy.


Mitochondrial dysfunction

Mechanistic tier: intermediate | Intervention tractability: moderate

TABLE WITHOUT ID file.link AS Compound, mechanisms AS Mechanism, clinical-stage AS Stage, human-evidence-level AS "Evidence", translation-gap AS "Gap", next-experiment AS "Next experiment"
FROM "molecules/compounds" OR "interventions"
WHERE contains(hallmarks, [[mitochondrial-dysfunction]])
  OR contains(target-hallmarks, [[mitochondrial-dysfunction]])
SORT clinical-stage DESC

Class-level synthesis: The broadest coverage of any hallmark by compound count (creatine, NMN, NR, urolithin-a, taurine, metformin, egcg, caloric-restriction all link here). Exercise is the most reliable mitochondrial biogenesis intervention in humans but is not yet a seeded compound/intervention page. NAD+ precursors have been disappointing on mitochondrial endpoints in human RCTs (Yoshino 2021: muscle NAD+ unchanged, respirometry unchanged). Urolithin-a’s mitophagy mechanism has biomarker support (phospho-Parkin increased at both doses in ATLAS) but hard-endpoint trials pending. Elamipretide (cardiolipin stabiliser) is on the roadmap but unseeded. See mitochondrial-dysfunction.


Epigenetic alterations

Mechanistic tier: proximal | Intervention tractability: moderate

TABLE WITHOUT ID file.link AS Compound, mechanisms AS Mechanism, clinical-stage AS Stage, human-evidence-level AS "Evidence", translation-gap AS "Gap", next-experiment AS "Next experiment"
FROM "molecules/compounds" OR "interventions"
WHERE contains(hallmarks, [[epigenetic-alterations]])
  OR contains(target-hallmarks, [[epigenetic-alterations]])
SORT clinical-stage DESC

Class-level synthesis: Only three compound/intervention pages currently link to this hallmark (NMN, NR, caloric-restriction). This under-represents the field: partial reprogramming (OSK/OSKM), HDAC inhibitors (butyrate), and SIRT1/6 activators all operate here but lack atomic pages. The DunedinPACE clock-slowing signal for caloric restriction (CALERIE) is the strongest human evidence. Partial reprogramming is the highest-potential frontier with zero clinical-stage representation. Translation gap: preclinical-only for the most promising interventions. See epigenetic-alterations, information-theory-of-aging. gap/needs-seeding: partial-reprogrammer, HDAC-inhibitor, SIRT6-activator compound pages.


Loss of proteostasis

Mechanistic tier: intermediate | Intervention tractability: moderate

TABLE WITHOUT ID file.link AS Compound, mechanisms AS Mechanism, clinical-stage AS Stage, human-evidence-level AS "Evidence", translation-gap AS "Gap", next-experiment AS "Next experiment"
FROM "molecules/compounds" OR "interventions"
WHERE contains(hallmarks, [[loss-of-proteostasis]])
  OR contains(target-hallmarks, [[loss-of-proteostasis]])
SORT clinical-stage DESC

Class-level synthesis: Direct intervention coverage remains sparse. EGCG links through autophagy/proteasome effects, while il-11-gene-silencing now contributes a preclinical ovarian-matrix example in which limiting collagen deposition changes a long-lived extracellular proteome. Spermidine and rapamycin operate strongly through autophagy but are tagged primarily to disabled-macroautophagy rather than this hallmark. Chaperone enhancers, proteostasis-specific targets and LAMP-2A stabilisers remain major unseeded gaps. gap/needs-seeding


Stem cell exhaustion

Mechanistic tier: integrative | Intervention tractability: low

TABLE WITHOUT ID file.link AS Compound, mechanisms AS Mechanism, clinical-stage AS Stage, human-evidence-level AS "Evidence", translation-gap AS "Gap", next-experiment AS "Next experiment"
FROM "molecules/compounds" OR "interventions"
WHERE contains(hallmarks, [[stem-cell-exhaustion]])
  OR contains(target-hallmarks, [[stem-cell-exhaustion]])
SORT clinical-stage DESC

Class-level synthesis: The matrix now links multiple intervention families: heterochronic blood-product approaches, senolytics, mTOR inhibition, caloric restriction, stem-cell replacement, partial reprogramming, telomerase strategies, and telomeric-antisense-oligonucleotides. The Oppezzo 2026 tASO study is notable because HSPC function improved without increasing LSK abundance or elongating telomeres, but it remains a single preclinical program with only two human donors ex vivo. low tractability remains appropriate because no intervention is clinically validated to restore stem-cell function in normal human aging.


Altered intercellular communication

Mechanistic tier: integrative | Intervention tractability: moderate

TABLE WITHOUT ID file.link AS Compound, mechanisms AS Mechanism, clinical-stage AS Stage, human-evidence-level AS "Evidence", translation-gap AS "Gap", next-experiment AS "Next experiment"
FROM "molecules/compounds" OR "interventions"
WHERE contains(hallmarks, [[altered-intercellular-communication]])
  OR contains(target-hallmarks, [[altered-intercellular-communication]])
SORT clinical-stage DESC

Class-level synthesis: anti-il-11-antibodies and il-11-gene-silencing now provide explicit cytokine-ligand/receptor examples on this axis, but their aging evidence remains preclinical. Broader coverage is still incomplete: canakinumab, senolytics and senomorphics act on intercellular inflammatory signaling but are not consistently tagged to this hallmark. gap/needs-tagging


Dysbiosis

Mechanistic tier: integrative | Intervention tractability: moderate

TABLE WITHOUT ID file.link AS Compound, mechanisms AS Mechanism, clinical-stage AS Stage, human-evidence-level AS "Evidence", translation-gap AS "Gap", next-experiment AS "Next experiment"
FROM "molecules/compounds" OR "interventions"
WHERE contains(hallmarks, [[dysbiosis]])
  OR contains(target-hallmarks, [[dysbiosis]])
SORT clinical-stage DESC

Class-level synthesis: Metformin (via gut-microbiome-modulation) currently links to this hallmark via frontmatter, and the dietary intervention class pages — prebiotics, probiotics (ad-hoc seed 2026-05-22), postbiotics, fmt — are all tagged with target-hallmarks: [[dysbiosis]]. Urolithin-a has a strong mechanistic link (postbiotic produced by gut bacteria, microbiome-mediated mechanism) but is not tagged here. gap/needs-tagging: urolithin-a should add dysbiosis to hallmarks.


Genomic instability

Mechanistic tier: proximal | Intervention tractability: low

TABLE WITHOUT ID file.link AS Compound, mechanisms AS Mechanism, clinical-stage AS Stage, human-evidence-level AS "Evidence", translation-gap AS "Gap", next-experiment AS "Next experiment"
FROM "molecules/compounds" OR "interventions"
WHERE contains(hallmarks, [[genomic-instability]])
  OR contains(target-hallmarks, [[genomic-instability]])
SORT clinical-stage DESC

Class-level synthesis: Zero compound/intervention pages link here. The hallmark page documents senolytics, NAD+ precursors (PARP1 substrate), and caloric restriction as tangential interventions, but none target genomic instability as a primary mechanism. This is a correct reflection of the therapeutic landscape — no intervention yet directly restores DDR fidelity in aged human tissue. low tractability is accurate. The most tractable near-term approach (NAD+ precursors → PARP1 substrate replenishment) is speculative. gap/needs-tagging: NMN and NR have a DDR-support mechanism but are not tagged here; decision pending on whether a mechanism-speculative tag is appropriate.


Telomere attrition

Mechanistic tier: proximal | Intervention tractability: low

TABLE WITHOUT ID file.link AS Compound, mechanisms AS Mechanism, clinical-stage AS Stage, human-evidence-level AS "Evidence", translation-gap AS "Gap", next-experiment AS "Next experiment"
FROM "molecules/compounds" OR "interventions"
WHERE contains(hallmarks, [[telomere-attrition]])
  OR contains(target-hallmarks, [[telomere-attrition]])
SORT clinical-stage DESC

Class-level synthesis: Three intervention pages link here: telomerase-activators, aav-tert, and telomeric-antisense-oligonucleotides. They cover telomerase activation, TERT delivery/telomere recapping, and suppression of persistent telomeric DDR without elongation, respectively. AAV-modTERT has reached an early human trial in dilated cardiomyopathy; tASOs remain preclinical and their human evidence is two donors ex vivo. low tractability remains appropriate because no approach has demonstrated clinical benefit for normal human aging or resolved long-term cancer/genomic-surveillance risk.


Gaps summary

HallmarkCompound pages linkedTractabilityPrimary gap
cellular-senescence5+ (high coverage)highPhase 3 RCT missing
chronic-inflammation5+ (high coverage)highCost/access blocker for strongest agent
deregulated-nutrient-sensing5+ (high coverage)highHuman hard-endpoint missing
disabled-macroautophagy4+ (good coverage)highHuman endpoint missing
mitochondrial-dysfunction5+ (broad coverage)moderateNAD+ precursor null in humans; exercise unseeded
epigenetic-alterations3 (partial)moderatePartial reprogramming unseeded
loss-of-proteostasis1 (egcg only)moderateMultiple compound pages missing
stem-cell-exhaustion10+ intervention classeslowNo validated restoration of aged human stem-cell function; tASO remains preclinical
altered-intercellular-communication2 intervention classesmoderateIL-11 ligand/receptor targeting is preclinical for aging; broader tagging remains incomplete
dysbiosis1 (metformin only via gut-microbiome-modulation)moderateUrolithin-a re-tagging + FMT seeding needed
genomic-instability1 (uv-protection)lowUV-protection added R42 — first direct-mechanism intervention (UVB → AP-1/NF-κB → MMP / CPD photoaging axis); coverage gap partially resolved
telomere-attrition3 intervention classeslowHuman efficacy and long-term genomic/cancer safety unresolved

Priority propagation candidates (frontmatter changes needed on existing pages, not new pages):

  1. rapamycin — add [[stem-cell-exhaustion]] to hallmarks:
  2. senolytics — add [[altered-intercellular-communication]] and [[stem-cell-exhaustion]] to target-hallmarks:
  3. senomorphics — add [[altered-intercellular-communication]] to target-hallmarks:
  4. canakinumab — add [[altered-intercellular-communication]] to hallmarks:
  5. urolithin-a — add [[dysbiosis]] to hallmarks:
  6. caloric-restriction — add [[stem-cell-exhaustion]] to target-hallmarks:

Class-level summary by tractability tier

Tier 1 — Clinical-stage interventions with human RCT anchor

ClassBest clinical agentStrongest human evidenceHallmarks covered
IL-1β neutralizercanakinumabCANTOS Phase 3 RCT (n=10,061) MACE HR 0.85chronic-inflammation
senolyticdasatinib+quercetinHickson 2019 proof-of-target (n=9); Justice 2019 functional endpoints (n=14)cellular-senescence, chronic-inflammation
AMPK activatormetforminTAME trial in progress; observational longevity signalderegulated-nutrient-sensing, mitochondrial-dysfunction
mTOR inhibitorrapamycinNIA ITP mouse lifespan extension (multi-site validated)deregulated-nutrient-sensing, disabled-macroautophagy

Tier 2 — Phase 2 RCT data, hard endpoint pending

ClassBest agentPhase 2 findingKey gap
mitophagy inducerurolithin-aATLAS: muscle biomarkers improved; primary endpoint (peak power) NSeffect-size-too-small
NAD precursornmn / nrBlood NAD+ elevated; muscle/functional endpoints nulleffect-size-too-small / biomarker-only
senolytic (fisetin)fisetinPhase 2 ongoing; mouse: 50% p16+ cell reductionbiomarker-only

Tier 3 — Preclinical / supplement stage

ClassAgentStrongest evidenceHuman gap
BCL-xL-selective PROTACDT2216, PZ15227 (no page)Preclinical mouse: thrombocytopenia-sparing senolysisNo IND filing yet
partial reprogrammerOSK (no page)Lu 2020 retinal reversal; Yang 2023 systemic ICENo human data; teratoma risk
gene therapyAAV-TERT (no page)Bernardes 2012: 24% mouse lifespan extensionNo Phase 1

Maintenance protocol

Seeder-agent checklist (R16 update to .claude/agents/wiki-seeder.md)

Every new type: compound or type: intervention page must:

  1. Set hallmarks: (compound) or target-hallmarks: (intervention) with at least one entry — verified against the hallmark page frontmatter.
  2. Populate R14 fields: translation-gap: (categorical) and next-experiment: (free-text one-liner).
  3. Verify the compound’s mechanisms: values match at least one class in intervention-classes. If no class fits, add the class to intervention-classes.md before shipping.
  4. After creating the page, open this matrix page and confirm the new row appears in the correct per-hallmark Dataview block. If it does not appear (due to a Dataview rendering issue or field mismatch), flag in the seeder summary.

Lint-pass extension (sops/lint-pass.md § “Intervention matrix coverage”)

The following checks have been added to the R16 lint protocol:

  1. Every type: compound and type: intervention page must have translation-gap: and next-experiment: populated. (Already enforced by R14 lint check; repeated here for emphasis.)
  2. Every mechanisms: value on a compound page must match a class on frameworks/intervention-classes.md. Flagged as #gap/class-normalisation-needed. After normalisation pass, lint should verify no orphan mechanism values remain.
  3. Every hallmark should be checked for linked interventions — use the live Dataview rather than the hard-coded Gaps Summary. Any zero must be classified as a true therapeutic-landscape gap or a tagging/coverage gap. Telomere attrition is no longer zero after seeding telomerase activation, AAV-TERT, and tASO pages.
  4. Watchdog 1 (intervention count per hallmark) must be reviewed at every seeding batch to catch newly zero-linked hallmarks.
  5. Watchdog 2 (stale clinical-stage entries) must be resolved within 180 days — trigger ClinicalTrials.gov re-check per sops/finding-aging-specific.md.

Self-monitoring summary

  • Watchdog 1 (intervention count) runs on page open — reviews live Dataview state.
  • Watchdog 2 (stale clinical-stage) surfaces any compound/intervention entry older than 180 days by verified-date.
  • Watchdog 3 (R14 fields missing) surfaces pages needing frontmatter population.
  • The Gaps Summary table above is manually maintained but should be updated after each propagation pass or seeding batch.

See also