Hallmark Causal Graph — Edge Data

⚠️ Auto-extracted by Claude on 2026-05-05. Each edge’s wiki-page-evidence field points to a verified (or verified-partial) atomic page where the supporting claim lives. Follow those links for n, p-values, and methodology. Evidence-strength ratings: strong = multiple independent lines or >1 replicated interventional study; moderate = single well-powered study or consistent mechanism without interventional proof; weak = mechanistic inference with limited direct evidence; disputed = contradictory or heavily contested evidence.

Edge list

upstream-hallmark	downstream-hallmark	evidence-strength	key-citation	wiki-page-evidence
genomic-instability	cellular-senescence	strong	Tyner 2002 (p53 m-allele); Rodier 2009 (DDR → SASP two-arm model)	genomic-instability verified · cellular-senescence verified
genomic-instability	chronic-inflammation	strong	Rodier 2009 (ATM-dependent SASP); Jaiswal 2014 (CHIP TET2-macrophage IL-1β/IL-6)	genomic-instability verified · chronic-inflammation verified
genomic-instability	clonal-hematopoiesis	strong	Jaiswal 2014 (n=17,182; HR 11.1 hematologic cancer; HR 2.0 CHD); Genovese 2014 (n=12,380)	genomic-instability verified · hematopoietic-stem-cells verified
telomere-attrition	cellular-senescence	strong	Bodnar 1998 / Hayflick; Celli & de Lange 2005 (TRF2 → DDR in MEFs); Bernardes de Jesus 2012 (AAV-TERT reverses)	telomere-attrition verified
telomere-attrition	stem-cell-exhaustion	strong	p21 KO in late-generation Terc−/− mice rescues stem cell function; HSC serial-transplantation telomere requirement	telomere-attrition verified · stem-cell-exhaustion verified
telomere-attrition	genomic-instability	moderate	Dysfunctional telomeres processed as DSBs; end-to-end fusions → breakage-fusion-bridge cycles amplify genomic instability	telomere-attrition verified · genomic-instability verified
epigenetic-alterations	cellular-senescence	moderate	Yang 2023 ICE: site-specific DSBs redistribute chromatin regulators → epigenetic clock acceleration → aging-like phenotypes; OSK reversal	epigenetic-alterations verified · information-theory-of-aging verified
epigenetic-alterations	stem-cell-exhaustion	moderate	HSC myeloid bias: methylation drift at lymphoid loci (Beerman 2010) correlates with and may causally underlie myeloid skewing with age	stem-cell-exhaustion verified · hematopoietic-stem-cells verified
cellular-senescence	chronic-inflammation	strong	SASP → IL-6, IL-8, IL-1β; Hickson 2019 D+Q senolysis → circulating SASP reduction; CANTOS (anti-IL-1β MACE HR 0.85) provides reverse-evidence	cellular-senescence verified · chronic-inflammation verified · sasp verified
cellular-senescence	stem-cell-exhaustion	moderate	Senescent niche stromal cells (SASP) disrupt HSC and satellite cell niches; Sousa-Victor 2014 geriatric satellite cells → p16+ irreversible senescence	cellular-senescence verified · stem-cell-exhaustion verified · satellite-cells verified
cellular-senescence	altered-intercellular-communication	moderate	SASP is the primary paracrine signal disruption in aged tissue; SASP factors (IL-6, MMPs, TGF-β) disrupt niche homeostatic signals at multiple tissue sites	cellular-senescence verified · sasp verified · altered-intercellular-communication verified
deregulated-nutrient-sensing	disabled-macroautophagy	strong	mTORC1 → ULK1 Ser757 inhibition → autophagy suppression; genetic epistasis: autophagy KO abolishes CR/rapamycin longevity benefit in worms, flies	deregulated-nutrient-sensing verified · disabled-macroautophagy verified · mtor verified · ampk verified
deregulated-nutrient-sensing	cellular-senescence	moderate	mTOR drives SASP translation in senescent cells (via 4E-BP1 / MK2 axis); mTOR inhibition reduces SASP output (verified senomorphic mechanism)	deregulated-nutrient-sensing verified · cellular-senescence verified · mtor verified
deregulated-nutrient-sensing	mitochondrial-dysfunction	moderate	mTOR suppresses mitophagy (ULK1 inhibition); AMPK activates mitophagy and biogenesis (PGC-1α); chronic mTOR over-activation → mitophagy impairment → organelle accumulation	deregulated-nutrient-sensing verified · mitochondrial-dysfunction verified
deregulated-nutrient-sensing	stem-cell-exhaustion	moderate	mTOR hyperactivation forces HSC exit from quiescence; rapamycin in aged mice improves HSC function; IIS regulates HSC quiescence via FOXO	deregulated-nutrient-sensing verified · stem-cell-exhaustion verified
disabled-macroautophagy	loss-of-proteostasis	strong	Atg5 KO (Hara 2006) + Atg7 KO (Komatsu 2006) → ubiquitin/p62+ inclusions + neurodegeneration; autophagy epistasis for proteostasis maintenance in post-mitotic cells	disabled-macroautophagy verified · loss-of-proteostasis verified · atg7 verified · atg5 verified
disabled-macroautophagy	mitochondrial-dysfunction	strong	Impaired mitophagy allows damaged, ROS-generating mitochondria to accumulate; PINK1/Parkin pathway requires functional autophagy machinery downstream	disabled-macroautophagy verified · mitochondrial-dysfunction verified · mitophagy verified
disabled-macroautophagy	chronic-inflammation	moderate	Autophagy limits SASP secretion and clears DAMP-generating debris; impaired autophagy → p62 accumulation → NF-κB activation	disabled-macroautophagy verified · nf-kb verified
mitochondrial-dysfunction	cellular-senescence	disputed	Wiley et al.: mitochondrial dysfunction (without DNA damage) induces MiDAS variant of senescence via ROS → mTORC1 → NAD+/NADH imbalance; counter-evidence: senescence → secondary mito dysfunction (Anderson 2019 cardiomyocytes)	mitochondrial-dysfunction verified · cellular-senescence verified · sasp verified
mitochondrial-dysfunction	chronic-inflammation	moderate	Mitochondrial DAMPs (mtDNA, cardiolipin, TFAM) → cGAS-STING, TLR9, NLRP3 inflammasome → NF-κB; failed mitophagy → NLRP3 priming	mitochondrial-dysfunction verified · chronic-inflammation verified · nf-kb verified
mitochondrial-dysfunction	loss-of-proteostasis	moderate	Reduced mitophagy flux allows damaged organelle accumulation; elevated mtROS damages proteins → aggregate load increases	mitochondrial-dysfunction verified · loss-of-proteostasis verified
loss-of-proteostasis	neurodegeneration	strong	Hara 2006 + Komatsu 2006 ATG KO → neurodegeneration; aggregates in AD (Aβ/tau), PD (α-syn), ALS (TDP-43) are all proteostasis-failure outputs	loss-of-proteostasis verified · neurodegeneration verified-partial · alzheimers-disease verified-partial
chronic-inflammation	stem-cell-exhaustion	moderate	Inflammaging cytokines (TNF-α, IL-6) → myeloid HSC bias (Beerman 2010); SASP impairs satellite cell activation (TGF-β from SASP)	chronic-inflammation verified · stem-cell-exhaustion verified · hematopoietic-stem-cells verified
chronic-inflammation	atherosclerosis	strong	CANTOS: canakinumab anti-IL-1β MACE HR 0.85 (95% CI 0.74–0.98, p=0.021; Ridker 2017, n=10,061); plaque senescent-cell clearance ~60% in mouse (Childs 2016)	chronic-inflammation verified · atherosclerosis verified-partial
chronic-inflammation	clonal-hematopoiesis (cardiovascular outcome)	moderate	CHIP TET2-mutant macrophages → excess IL-1β/IL-6 → cardiovascular risk (Jaiswal 2014 HR 2.0 CHD); bidirectional: CHIP drives chronic inflammation	chronic-inflammation verified · hematopoietic-stem-cells verified
dysbiosis	chronic-inflammation	moderate	LPS translocation → TLR4 → NF-κB → cytokines; Claesson 2012 ELDERMET: long-stay subjects (highest frailty/dysbiosis) had highest IL-6/CRP; diet is the dominant upstream driver per Claesson 2012	dysbiosis verified · chronic-inflammation verified
dysbiosis	altered-intercellular-communication	moderate	Microbiome-derived metabolites (SCFAs, urolithins, TMAO, bile acids) are systemic signaling molecules; dysbiosis disrupts this communication network	dysbiosis verified · altered-intercellular-communication verified
altered-intercellular-communication	stem-cell-exhaustion	weak	Heterochronic parabiosis (Conboy 2005): young systemic factors rescue aged satellite cell function; the causal factor(s) are contested (GDF11 controversy)	altered-intercellular-communication verified · stem-cell-exhaustion verified · satellite-cells verified
stem-cell-exhaustion	sarcopenia	strong	Satellite cell functional decline (Sousa-Victor 2014 p16+ geriatric arrest; Conboy 2005 niche rescue) is a primary contributor to sarcopenia; multi-factorial but established	stem-cell-exhaustion verified · sarcopenia verified
stem-cell-exhaustion	immunosenescence	strong	HSC myeloid bias → reduced naive lymphocyte output → immunosenescence; consistent across species	stem-cell-exhaustion verified · hematopoietic-stem-cells verified · immunosenescence verified-partial
stem-cell-exhaustion	chronic-inflammation (feedback)	moderate	CHIP-bearing TET2-mutant macrophages overproduce IL-1β/IL-6 → amplifies systemic inflammaging; closes a feed-forward loop	stem-cell-exhaustion verified · chronic-inflammation verified · hematopoietic-stem-cells verified
epigenetic-alterations	deregulated-nutrient-sensing	weak	SIRT1 (histone deacetylase) decline with NAD+ loss links epigenetic maintenance to nutrient-sensing pathway activity; indirect mechanistic link via NAD+ availability	epigenetic-alterations verified · sirtuin verified · deregulated-nutrient-sensing verified

Edge summary statistics

Total directed edges: 34
Strong: 11
Moderate: 17
Weak: 3
Disputed: 1 (mitochondrial-dysfunction ↔ cellular-senescence, both directions)

Nodes with most outbound edges (most causally upstream)

Hallmark	Outbound edges	Role
genomic-instability	3	Proximal damage accumulator; drives DDR-senescence and CHIP-inflammation
deregulated-nutrient-sensing	4	Central metabolic regulator; drives autophagy, senescence, mito, stem cells
cellular-senescence	3	Key intermediate; drives inflammation, niche disruption, signal noise
disabled-macroautophagy	3	Autophagy decline drives proteostasis, mitochondria, inflammation
mitochondrial-dysfunction	3	Cross-tier node; drives inflammation, proteostasis, senescence (disputed direction)

Nodes with most inbound edges (most causally downstream / integrative)

Hallmark	Inbound edges	Role
chronic-inflammation	7	Downstream integrator of nearly every upstream hallmark
cellular-senescence	4	Key intermediate: receives from proximal damage, sends to integrative outcomes
stem-cell-exhaustion	5	Terminal integrator across multiple organ systems
loss-of-proteostasis	3	Receives from autophagy, mitochondria; outputs to neurodegeneration
mitochondrial-dysfunction	3	Cross-tier: receives from nutrient sensing, autophagy

Bidirectional edges (mutual reinforcement loops)

These edges are positive feedback loops — once both hallmarks are active, they amplify each other:

Edge	Loop mechanism
cellular-senescence ↔ chronic-inflammation	SASP drives inflammation; inflammaging promotes paracrine senescence
dysbiosis ↔ chronic-inflammation	LPS drives inflammation; inflammaging disrupts gut barrier
stem-cell-exhaustion ↔ chronic-inflammation	CHIP-macrophages secrete IL-1β; inflammaging impairs HSC function
mitochondrial-dysfunction ↔ cellular-senescence	MiDAS senescence (disputed); senescent cells accumulate dysfunctional mitochondria

Edge-blocker tracker

R45 schema addition (2026-05-20). The edge table above captures causal evidence. This section captures the intervention-tractability and experimental-resolution layer — what would it take to either unlock a clinical intervention at the upstream node, or tighten/refute a disputed edge? Each row links to a type: experiment page in experiments/ where the proposed resolving study lives.

Three columns are tracked per edge: upstream-tractability, resolving-experiment, experiment-scale. Each is defined below; the populated table is in the next subsection.

Column 1 — upstream-tractability

How feasible is it, with 2026 clinical tooling, to therapeutically modulate the upstream node such that the downstream edge effect is reduced? This is a per-edge rating rather than a per-node one, because the edge may depend on a specific sub-mechanism that is more or less tractable than the upstream hallmark overall. Aligned with the intervention-tractability: field on hallmark pages (R14 schema).

Tier	Definition	Worked example
`none`	No known intervention modulates the upstream node in any model organism for this edge. The mechanism is described but not actionable.	(Currently no edges in this matrix sit here; included for completeness as the bottom of the scale.)
`low`	Preclinical candidates only (mouse, in-vitro, cell-culture proof-of-concept). Either no human safety data exists, or human Phase 1 is ongoing without an efficacy readout. The mechanism is real but not deployable.	`[[genomic-instability]] → [[cellular-senescence]]` — DDR-restoration in aged human tissue has no validated tool; AAV-CIRBP and NMR-style HR-enhancement are preclinical.
`moderate`	At least one clinical-stage candidate exists with human safety data, but efficacy is null, contested, or indication-bounded. Includes drugs FDA-approved for non-aging indications without aging-context efficacy translation (rapamycin off-label after PEARL 2025 + RAPA-EX-01 2026 nulls); biomarker-only readouts without hard endpoints; gene therapies in single-tissue Phase 1; mechanism-validated but trial-pending.	`[[epigenetic-alterations]] → [[cellular-senescence]]` — CR slows DunedinPACE (CALERIE-2); OSK retina-bounded in mouse (Lu 2020); no systemic human tool.
`high`	At least one validated human intervention exists with documented engagement of THIS specific edge. Hard endpoint or strong biomarker; replication across labs; an aging-context indication is established.	`[[cellular-senescence]] → [[chronic-inflammation]]` — D+Q reduces circulating SASP markers (Hickson 2019, n=9, p<0.05); UBX1325 intravitreal +5.6 ETDRS letters in BEHOLD DME (NEJM Evidence 2025).

Per-edge nuance: a node may carry intervention-tractability: high on its hallmark page but receive a lower per-edge rating in this matrix if the available intervention engages a different sub-mechanism than what this edge depends on. Example: cellular-senescence is high at the node level, but the cellular-senescence → stem-cell-exhaustion edge is high only contingent on whether senolytic clearance generalizes from D+Q-tractable adipose/lung tissues to satellite-cell or HSC niches — a sub-question not yet clinically resolved. Use the per-edge rating; when it differs from the node-level field, document the rationale briefly in body.

Distinction from evidence-strength (in the edge table above): evidence-strength rates how well-established the causal edge is in nature; upstream-tractability rates how well we can intervene on it clinically. The two are orthogonal: an edge can be strong evidence + low tractability (well-established cascade we can’t yet act on), or disputed evidence + moderate tractability (the mechanism is contested but a drug is in the clinic anyway).

Column 2 — resolving-experiment

The experiment that would either (a) resolve a disputed/weak edge, (b) unlock an intervention at the upstream node, or (c) close a missing-mechanism gap. Wikilink to the experiments/ page; null if not yet drafted.

Column 3 — experiment-scale

Feasibility tier. Mirrors the scale: field on the experiment page.

The fundamental boundary is between lab and clinic. Lab work has no living human participants — it operates on cells, tissues, animal models, or ex-vivo human samples (cadaver-derived tissue, biobank specimens, residual surgical material). Clinical work has living human participants under IRB + (where a drug or device is involved) IND oversight, with informed consent and ICH-GCP discipline. The regulatory floor, timeline, and cost step-change at this boundary.

Tier	Subjects	Regulatory floor	Typical n	Typical duration	Typical cost	Worked example
`small-lab`	Cells, tissues, animal models, ex-vivo human samples. No living human subjects.	IBC (biosafety); IACUC if vertebrate animals; IRB if human-derived samples but no living-subject interaction (tissue bank, residual material).	Cell-line/donor n=3–20; mouse cohort n=10–60.	6–24 months.	$30 k -$ 500k.	Four proposed ECM-crosslink experiments: MIDAS cell-type stratification, AGE-breaker LC-MS replication, FAOD directed evolution, glucosepane monoclonal antibody.
`small-clinic`	Living human participants. Phase 1 or single-arm Phase 2. Typically first-in-human or indication-bounded.	IRB + IND (drugs) or IDE (devices); ICH-GCP; informed consent; FDA pre-IND meeting.	n=10–100.	1–3 yr (trial + analysis).	$1 M -$ 10M (academic); higher commercial.	AAV-TERT Phase 1 in cardiomyopathy; OSK indication-bounded Phase 1 in optic neuropathy; TRIIM-X-class thymic regeneration Phase 1/2.
`large-clinic`	Randomized controlled trials with hard or surrogate endpoints. Phase 2 / Phase 3. Single institution or small consortium.	Full IRB + IND + DSMB; ICH-GCP; site-monitoring discipline.	n=100–1000.	2–5 yr.	$10 M -$ 100M.	Biomarker-stratified senolytic Phase 2 (Farr 2024 quercetin follow-on); senescent-cell-PET-tracer-enabled trial design (proposed senescent-cell PET tracer, Phase 2+); PEARL-style rapamycin healthy-adult RCT.
`multi-site`	Large-scale Phase 3 / Phase 4 / large prospective epidemiology. Distributed across institutions or countries.	Full clinical-trial discipline at each site; international harmonization (ICH); regulatory submission for approval.	n=1,000–50,000+.	3–8 yr.	$50 M -$ 500M+.	CANTOS-scale anti-IL-1β cardiovascular outcomes (n=10,061); Mannick-style mTORC1 immune-aging Phase 3.
`multi-decade`	Lifespan trials in humans; generational longitudinal cohorts (NHANES, UK Biobank, Framingham); cross-generational genetic studies.	Long-term cohort discipline; data-governance + privacy frameworks; institutional commitment > PI tenure.	n=10,000+ typical for cohort.	10–50+ yr.	$100M–billions over the life of the cohort.	A true human-lifespan RCT for any aging intervention (does not yet exist); longevity-escape-velocity-relevant longitudinal endpoint cohorts.

Why the lab→clinic boundary matters for the matrix: when a resolving-experiment sits at small-lab scale, it can be funded by a single PI grant, run in 1–2 years, and produce a binary signal that updates the matrix directly. When the resolving experiment is small-clinic or higher, the resolution timeline is 3+ years from funding, the cost grows by orders of magnitude, and the matrix update is gated by trial readouts that lie outside any single research program’s control. This is the matrix’s leverage-discovery function: rows where the resolving experiment is small-lab are operationally unblocked even when the underlying biology is not yet fully understood — they are where a small program can directly tighten the matrix without waiting on industry or large-consortium trials.

small-lab includes preclinical animal in-vivo work despite the name; the term refers to the absence of living human subjects + the single-PI scale, not literally to bench-only work. A mouse-cohort efficacy study (e.g., AAV-TERT in aged C57BL/6) is small-lab until it expands to a primate or human Phase 1, at which point it becomes small-clinic. The boundary tracks subjects + regulatory floor, not the literal location of the work.

Edge-blocker table

Edge	Upstream tractability	Resolving experiment	Scale
genomic-instability → cellular-senescence	low	null (open: AAV-CIRBP or NMR-cGAS HR-enhancement Phase 1 — needs separate proposal)	small-clinic
genomic-instability → chronic-inflammation	low	null (open: CHIP-clone-targeted JAK inhibition Phase 2 stratified by clone-VAF)	large-clinic
genomic-instability → clonal-hematopoiesis	low	null (open: allele-selective ASO against driver SNVs — DNMT3A, TET2 hotspots)	small-clinic
telomere-attrition → cellular-senescence	low	null (open: AAV-TERT Phase 1 in cardiac or BMF indication)	small-clinic
telomere-attrition → stem-cell-exhaustion	low	null	small-clinic
telomere-attrition → genomic-instability	low	null	—
epigenetic-alterations → cellular-senescence	moderate	null (open: indication-bounded OSK Phase 1 — optic neuropathy, Lu 2020 precedent)	small-clinic
epigenetic-alterations → stem-cell-exhaustion	moderate	null	—
cellular-senescence → chronic-inflammation	high	null (open: biomarker-stratified senolytic Phase 2 — requires senescent-cell PET tracer, see below)	large-clinic
cellular-senescence → stem-cell-exhaustion	high	null	—
cellular-senescence → altered-intercellular-communication	high	null	—
deregulated-nutrient-sensing → disabled-macroautophagy	high	null (open: 4E-BP1-biased rapalog Phase 1; mTORC2-sparing selective mTORC1 inhibitor)	small-clinic
deregulated-nutrient-sensing → cellular-senescence	high	null	—
deregulated-nutrient-sensing → mitochondrial-dysfunction	high	null	—
deregulated-nutrient-sensing → stem-cell-exhaustion	high	null	—
disabled-macroautophagy → loss-of-proteostasis	high	null (open: TFEB-activator gene therapy Phase 1 in lysosomal storage disease)	small-clinic
disabled-macroautophagy → mitochondrial-dysfunction	high	null	—
disabled-macroautophagy → chronic-inflammation	high	null	—
mitochondrial-dysfunction → cellular-senescence (disputed)	moderate	MIDAS cell-type stratification (proposed)	small-lab
mitochondrial-dysfunction → chronic-inflammation	moderate	null	—
mitochondrial-dysfunction → loss-of-proteostasis	moderate	null	—
loss-of-proteostasis → neurodegeneration	moderate	null (open: engineered-hydrolase Phase 1 for A2E in atrophic AMD — LysoSENS proof-of-concept)	small-clinic
chronic-inflammation → stem-cell-exhaustion	high	null	—
chronic-inflammation → atherosclerosis	high	null (canakinumab CANTOS-precedent; cost-prohibitive; IL-6-axis follow-ons in trial)	—
chronic-inflammation → clonal-hematopoiesis	high	null	—
dysbiosis → chronic-inflammation	moderate	null (open: synthetic-consortium Phase 1 in IBD → aging-context FMT trial with frailty primary)	small-clinic
dysbiosis → altered-intercellular-communication	moderate	null	—
altered-intercellular-communication → stem-cell-exhaustion (weak; GDF11)	moderate	null (open: heterochronic plasma dilution Phase 2 — Conboy lab follow-on)	small-clinic
stem-cell-exhaustion → sarcopenia	low	null	—
stem-cell-exhaustion → immunosenescence	low	null (open: TRIIM-X Phase 2 readout pending — recombinant FOXN1 / IL-7-LA Phase 1 follow-on)	small-clinic
stem-cell-exhaustion → chronic-inflammation	low	null	—
epigenetic-alterations → deregulated-nutrient-sensing (weak)	moderate	null	—

Read across:

11 of 34 edges flow from a low-tractability upstream node — these are the structural blockers. Almost all of Tier 1 (genomic instability, telomere attrition) is in this bucket.
17 edges flow from a moderate or high-tractability node but have null resolving experiments — these are interventions waiting for trial readouts, dose discovery, or biomarker development; they are operationally blocked rather than biologically blocked.
Only 1 edge (mito → senescence, disputed) is currently resolved-by a small-lab experiment. This is the lowest-cost / fastest-feedback row.

Missing-node blockers

R45 schema addition. The López-Otín 12-hallmark frame underweights two damage classes that the SENS frame treats as primary: extracellular matrix crosslinks (GlycoSENS) and intracellular indigestible aggregates (LysoSENS). Both feed into existing hallmark nodes and have characteristic downstream phenotypes, but are not first-class nodes in the matrix above. Surfacing them here lets the leverage ranking capture them. Each missing node is anchored to existing wiki process pages.

ECM crosslinks (GlycoSENS)

Canonical pages: glucosepane (verified), advanced-glycation-end-products, pentosidine, carboxymethyl-lysine, age-crosslink-breakers (verified)

Why missing from the hallmark frame: López-Otín’s altered-intercellular-communication hallmark loosely absorbs ECM-stiffening as a “communication” phenotype, but ECM crosslinking is structurally a damage-accumulation class — molecularly distinct, mechanistically irreversible, and not driven by signaling deregulation. Folding it into intercellular communication obscures its primary causal role.

Inbound edges (drivers):

upstream	→ downstream	evidence-strength	key-citation
hyperglycemic-state / chronic methylglyoxal flux (not a hallmark)	→ ECM-crosslinks	strong	Sell 2005: skin glucosepane ~2000 pmol/mg nondiabetic age-90 vs ~5000 pmol/mg DM age-matched
mitochondrial-dysfunction	→ ECM-crosslinks	moderate	Mitochondrial ROS → lipid peroxidation → α-dicarbonyl flux (MG, glyoxal) → Schiff bases on long-lived ECM lysine/arginine
deregulated-nutrient-sensing	→ ECM-crosslinks	moderate	mTOR/IIS deregulation → impaired mitophagy → mitochondrial ROS (above); also hyperglycemia in insulin-resistant states

Outbound edges (downstream effects):

edge	downstream	evidence-strength	key-citation
ECM-crosslinks	→ altered-intercellular-communication	strong	ECM stiffness disrupts mechanotransduction (YAP/TAZ pathway) in fibroblasts, satellite cells, HSC niche
ECM-crosslinks	→ chronic-inflammation	moderate	RAGE engagement → NF-κB; positive-feedback amplification via RAGE upregulation; rage
ECM-crosslinks	→ arterial-stiffening	strong	PWV correlates with skin autofluorescence; arterial collagen glucosepane drives intrinsic wall stiffness independent of vasomotor tone
ECM-crosslinks	→ stem-cell-exhaustion	weak-moderate	Substrate stiffness alters fibroblast/satellite cell/MSC fate decisions (Engler 2006 mechanotransduction precedent); aged-niche stiffness → niche-dysfunction

Tractability: low. No clinically-deployable AGE breaker exists. Alagebrium failed Phase 3 BENEFICIAL (Hartog 2011, peak VO₂ p=0.06). Yang 2003 critique (model α-dicarbonyl crosslinks ≠ real-tissue Maillard crosslinks) is unresolved. Revel Pharmaceuticals had a glucosepane-enzyme reproducibility setback (industry-watcher reporting, unverified primary). Tier 3 defense (mature-crosslink cleavage) is genuinely empty across all of biology — no mammalian, fungal, bacterial, or archaeal enzyme cleaves glucosepane or pentosidine.

Resolving experiments (proposed; not yet run):

AGE-breaker LC-MS replication — does any current “AGE breaker” cleave glucosepane in aged human cadaver tissue?
FAOD directed evolution — engineer a Tier-3 defense that doesn’t exist in biology
Glucosepane monoclonal antibody — enable population-level IHC quantification + AGE-breaker screening throughput

Intracellular indigestible aggregates (LysoSENS)

Canonical pages: lipofuscin (verified-false; closed-access blocker), 7-ketocholesterol

Why missing from the hallmark frame: López-Otín’s loss-of-proteostasis hallmark covers protein aggregates (Aβ, tau, α-syn) but indigestible lipid-protein-crosslink aggregates (lipofuscin, A2E, oxidized cholesterol byproducts) are mechanistically distinct: they overwhelm lysosomal degradation by being chemically un-cleavable rather than by exceeding capacity. They accumulate predominantly in long-lived post-mitotic cells (cardiomyocytes, neurons, RPE) that cannot be cleared by cell turnover.

Inbound edges (drivers):

upstream	→ downstream	evidence-strength
disabled-macroautophagy	→ LysoSENS-aggregates	strong (failed autophagic clearance permits aggregate growth)
mitochondrial-dysfunction	→ LysoSENS-aggregates	moderate (mtROS → lipid peroxidation → lipofuscin precursors)
chronic-inflammation	→ LysoSENS-aggregates	moderate (oxLDL → 7-ketocholesterol intracellular accumulation in macrophages, RPE)

Outbound edges (downstream effects):

edge	downstream	evidence-strength
LysoSENS-aggregates	→ loss-of-proteostasis	moderate (overwhelmed lysosomes spillover-impair protein degradation)
LysoSENS-aggregates	→ cellular-senescence	weak (lipofuscin marks senescent cells; causal direction contested — Chiavacci 2026 Greenland shark shows lipofuscin compatible with preserved cardiac function)
LysoSENS-aggregates	→ mitochondrial-dysfunction	moderate (dysfunctional lysosomes → impaired mitophagy → bidirectional with autophagy/mito)

Tractability: low. No human-deployable tool for clearing intracellular indigestible aggregates exists. Underdog Pharmaceuticals (A2E in atrophic AMD) is the lead clinical program; preclinical only. Engineered-hydrolase strategy (soil-bacteria-derived cleaving enzymes targeted to lysosomes) is the canonical SENS-LysoSENS approach.

Resolving experiments: null (open: engineered-hydrolase Phase 1 in atrophic AMD would be the first clinical proof-of-concept). No small-lab-tractable resolving experiment currently drafted.

🧬 Aging Wiki

Explorer

causal-graph-data