Alzheimer Disease

Neurodegenerative Disorder MONDO:0004975 Pathograph 44 Show in embeddings browser Dementia Neurodegenerative Disease

Alzheimer disease is a progressive neurodegenerative disorder characterized by cognitive decline, memory loss, and behavioral changes. It is the most common cause of dementia, involving the accumulation of amyloid-beta plaques and neurofibrillary tangles in the brain, leading to neuronal death and brain atrophy.

Ask OpenScientist

Ask a research question about Alzheimer Disease. OpenScientist will conduct autonomous deep research using the Disorder Mechanisms Knowledge Base and PubMed literature (typically 10-30 minutes).

Submitting...

Do not include personal health information in your question. Questions and results are cached in your browser's local storage.

◆

Subtypes

Early-Onset Alzheimer's Disease

Type of Alzheimer's that occurs in individuals younger than 65 and is often associated with genetic factors.

Show evidence (2 references)

PMID:30707186 SUPPORT

"Early-onset Alzheimer disease (AD) is defined as having an age of onset younger than 65 years. ... Early-onset AD comprises about 5% to 6% of cases of AD and includes a substantial percentage of phenotypic variants that differ from the usual amnestic presentation of typical AD. Characteristics..."

This reference clearly supports the definition of Early-Onset Alzheimer's Disease as occurring in individuals younger than 65 and being often associated with genetic factors.

PMID:25998117 SUPPORT

"Young-onset dementia is hereditary, multifactorial, or sporadic. The most common hereditary dementias include Alzheimer disease."

This reference supports the idea that early-onset Alzheimer's disease can be hereditary, aligning with the statement's mention of genetic factors.

Late-Onset Alzheimer's Disease

The most common form of Alzheimer's, occurring in those aged 65 and older.

Show evidence (3 references)

PMID:24429902 SUPPORT

"In both cases, the disease results in severe cognitive dysfunction, among other problems, and the late-onset form of the disease is now considered to be the most common cause of dementia among the elderly."

The literature confirms that late-onset Alzheimer's Disease is the most common form of Alzheimer's and primarily affects those aged 65 and older.

PMID:34120901 SUPPORT

"BACKGROUND: Young onset dementia is associated with a longer time to diagnosis compared to late onset dementia."

The study differentiates between young onset and late onset Alzheimer's Disease, with late onset typically affecting older adults.

PMID:18667359 SUPPORT

"Alzheimer's disease accounts for 60% whereas vascular dementia accounts for approximately 30% of the prevalence. Early-onset familial forms of dementia with single-gene defects occur in Latin America, Asia, and Africa."

Supports Alzheimer's disease as the leading cause of dementia (~60% of cases); early-onset familial single-gene forms are noted as a distinct minority category, consistent with late-onset being the predominant form.

Comorbidities

Alzheimer Disease <-> Cancer

Alzheimer Disease <-> Type 2 Diabetes Mellitus

Influenza <-> Alzheimer Disease

◈

Mechanistic Hypotheses

Amyloid Cascade Model

amyloid_cascade_model CANONICAL Early-Onset Alzheimer's Disease Late-Onset Alzheimer's Disease

Evidence balance 5 support 6 refute

Amyloid-beta production, oligomerization, and plaque deposition are modeled as upstream drivers of tau pathology, synaptic toxicity, glial activation, and progressive cognitive decline, especially in APP/PSEN-associated familial Alzheimer disease.

Retained as CANONICAL, but not sufficient as a complete disease explanation: tau burden, vascular injury, immune state, lysosomal clearance, aging, and genetic background modulate how amyloid pathology maps to symptoms. Best read as a necessary rather than sufficient initiating trigger, applying most strongly to familial disease and to the amyloid-first sporadic subtype. Four independent lines qualify it and are curated above: a tau-first spatiotemporal subtype (PMID:37433038), the Cochrane finding that amyloid removal produces no clinically meaningful benefit (PMID:41985900), oldest-old resilience attributable to co-pathology rather than amyloid tolerance (PMID:41942821), and an APP-dependent C99 endolysosomal arm that does not run through amyloid-beta (PMID:31043483). Assessed from the OpenScientist amyloid_cascade_model report (kb/hypotheses/Alzheimer_Disease/); that report's claim that BACE1 inhibitors raise C99 is backwards and was not curated — BACE1 generates C99, and the cited source reports that BACE1 inhibition rescues the cathepsin deficits.

Pathograph links

Amyloid Plaque Formation -> Neurofibrillary Tangle Formation INDIRECT KNOWN INTERMEDIATES Amyloid-beta accumulation is modeled as an upstream trigger of tau hyperphosphorylation and neurofibrillary tangle formation. Intermediates: Amyloid-beta oligomer and fibril toxicity inducing tau hyperphosphorylation.

Amyloid Plaque Formation -> Synaptic Dysfunction DIRECT Soluble amyloid-beta assemblies are modeled as direct toxic contributors to synaptic dysfunction and cognitive impairment.

Amyloid Plaque Formation -> Neuroinflammation INDIRECT KNOWN INTERMEDIATES Amyloid-beta accumulation and activated microglia can form a reinforcing inflammatory cycle. Intermediates: Microglial activation and inflammatory mediator release.

Neuroinflammation -> Amyloid Plaque Formation DIRECT Inflammasome activation in microglia can reinforce amyloid-beta aggregation through extracellular ASC specks, making amyloid-to-glia ordering partly reciprocal rather than a one-way edge.

Mitochondrial Quality-Control Failure -> Amyloid Plaque Formation INDIRECT KNOWN INTERMEDIATES Impaired mitochondrial quality control is modeled as a contributor to amyloid-beta accumulation through reduced mitophagy-linked microglial plaque clearance and increased neuroinflammatory stress. Intermediates: Reduced microglial phagocytosis of extracellular amyloid-beta plaques.; Suppression or amplification of neuroinflammation depending on mitophagy state.

Vascular Dysfunction -> Amyloid Plaque Formation INDIRECT KNOWN INTERMEDIATES Blood-brain barrier dysfunction and neurovascular-unit impairment are modeled as contributors to brain amyloid-beta accumulation and impaired clearance. Intermediates: Impaired amyloid-beta efflux and neurovascular clearance.

Glymphatic Clearance Failure -> Amyloid Plaque Formation INDIRECT KNOWN INTERMEDIATES Reduced perivascular clearance raises the steady-state interstitial amyloid-beta concentration, favoring aggregation and plaque deposition. Intermediates: Reduced CSF-interstitial fluid exchange raising interstitial amyloid-beta concentration.

Autophagy-Lysosomal Dysfunction -> Amyloid Plaque Formation INDIRECT KNOWN INTERMEDIATES Impaired autophagy and lysosomal degradation are modeled as contributors to amyloid-beta accumulation through reduced aggregate clearance. Intermediates: Reduced amyloid-beta clearance through defective autophagic flux and lysosomal degradation.

Neuronal Endosomal-Lysosomal Acidification Failure -> Amyloid Plaque Formation INDIRECT KNOWN INTERMEDIATES Amyloid-beta accumulated within de-acidified autolysosomes of individual neurons is released when those neurons rupture, so the plaque is modeled as originating from a dying neuron rather than from secreted peptide. Intermediates: Lysosomal membrane permeabilization, cathepsin release and lysosomal cell death.

Oligodendrocyte and Myelin Dysfunction -> Amyloid Plaque Formation INDIRECT KNOWN INTERMEDIATES Myelin dysfunction is modeled as increasing amyloid deposition both directly, by concentrating amyloidogenic processing in axonal swellings, and indirectly, by diverting plaque-clearing microglia to myelin damage. Intermediates: Axonal accumulation of amyloid precursor protein processing machinery.; Competition for the disease-associated microglial response.

Deep research

OpenScientist citations 28 citations 2026-05-18T21:27:36.693751

Show evidence (11 references)

PMID:25941885 SUPPORT

"The 'amyloid cascade hypothesis' posits that an extracellular build-up of amyloid-beta oligomers (Abeta-os) and polymers (fibrils) subsequently inducing toxic hyperphosphorylated (p)-Tau oligomers (p-Tau-os) and neurofibrillary tangles starts the sporadic late-onset Alzheimer's disease (LOAD)..."

Supports amyloid-beta accumulation as an upstream causal model that induces downstream tau pathology.

PMID:22351073 SUPPORT

"Generation of amyloid peptide (Abeta) is at the beginning of a cascade that leads to Alzheimer's disease (AD)... soluble assembly states of Abeta peptides can cause cognitive problems by disrupting synaptic function..."

Supports amyloid-beta generation as an initiating cascade event with synaptic consequences.

PMID:39496485 SUPPORT Model Organism

"APP-A673T significantly downregulated β-cleavage and attenuated the production of Aβ and amyloid pathology in the brains of these animals."

In vivo arm of the bidirectional genetic argument: introducing the protective Icelandic APP variant into Swedish-mutation-free App knock-in mice reduces beta-cleavage and amyloid pathology, complementing the causal familial mutations at the other end of the dose range.

+ 8 more references

Tau Neurodegeneration Model

tau_neurodegeneration_model CANONICAL Early-Onset Alzheimer's Disease Late-Onset Alzheimer's Disease

Evidence balance 4 support 3 refute

Tau hyperphosphorylation and aggregation into neurofibrillary tangles are modeled as proximate drivers of neuronal dysfunction, microtubule disruption, neurodegeneration, and clinical progression.

Pathograph links

Neurofibrillary Tangle Formation -> Synaptic Dysfunction INDIRECT KNOWN INTERMEDIATES Hyperphosphorylated tau and tau aggregates are modeled as toxic contributors to impaired synaptic plasticity and neurotransmitter release. Intermediates: Tau-mediated microtubule disruption and synaptic toxicity.

Neurofibrillary Tangle Formation -> Necroptotic Neuronal Death DIRECT Hyperphosphorylated tau is modeled as an upstream trigger of necrosome assembly, connecting the entry's central proteinopathy to the execution step by which neurons are proposed to die.

Neurofibrillary Tangle Formation -> Intercellular Tau Transmission via Extracellular Vesicles DIRECT Intracellular pathological tau is packaged by Arc into neuronal extracellular vesicles and released, providing a route for cell-to-cell spread of tau seeds.

Neuroinflammation -> Neurofibrillary Tangle Formation INDIRECT UNKNOWN INTERMEDIATES Microglial state is modeled as a context-dependent modifier of amyloid-associated tau seeding and spreading; TREM2-dependent disease-associated microglia can restrain tau propagation, while a simple pathogenic NLRP3-to-tau edge remains contested. Intermediates: TREM2-dependent plaque-associated disease-associated microglial activation.; Context-dependent inflammasome, cytokine, and complement signaling.

Mitochondrial Quality-Control Failure -> Neurofibrillary Tangle Formation INDIRECT KNOWN INTERMEDIATES Impaired mitophagy is modeled as a contributor to tau hyperphosphorylation and tau pathology, but its position relative to amyloid and glial activation remains unresolved. Intermediates: Failure to remove defective mitochondria and downstream mitochondrial stress.

Autophagy-Lysosomal Dysfunction -> Neurofibrillary Tangle Formation INDIRECT KNOWN INTERMEDIATES Impaired autophagy and lysosomal degradation are modeled as contributors to tau aggregate accumulation through reduced proteostatic clearance. Intermediates: Reduced tau aggregate clearance through defective autophagic flux and lysosomal degradation.

Intercellular Tau Transmission via Extracellular Vesicles -> Neurofibrillary Tangle Formation INDIRECT KNOWN INTERMEDIATES EV-delivered seed-competent tau is taken up by recipient neurons and seeds aggregation of endogenous tau, propagating tangle pathology along connected circuits. Intermediates: EV-tau uptake by recipient neurons and templated seeding of endogenous tau.

Selective Vulnerability of RORB+ Entorhinal Excitatory Neurons -> Neurofibrillary Tangle Formation INDIRECT UNKNOWN INTERMEDIATES Cell-intrinsic vulnerability of RORB+ entorhinal excitatory neurons is modeled as determining where neurofibrillary inclusions first form, giving tau pathology its characteristic anatomical starting point.

Senescent Cell Accumulation -> Neurofibrillary Tangle Formation INDIRECT KNOWN INTERMEDIATES Senescent cells are modeled as promoting tau hyperphosphorylation and tangle deposition; the direction is established by the inverse experiment, in which clearing them prevents both. Intermediates: Senescence-associated secretory phenotype and gliosis.

Microglial Lipid Droplet Accumulation -> Neurofibrillary Tangle Formation INDIRECT KNOWN INTERMEDIATES Factors secreted by lipid-droplet-containing microglia drive tau phosphorylation and neurotoxicity in an APOE-dependent manner, linking the glial lipid state to tau pathology. Intermediates: Secreted microglia-derived factors acting on neurons.

Tau-Induced Nucleocytoplasmic Transport Failure -> Neurofibrillary Tangle Formation INDIRECT KNOWN INTERMEDIATES Nup98 mislocalized into the neuronal cytoplasm facilitates tau aggregation, closing a feed-forward loop in which tau-driven transport failure promotes further tau aggregation. Intermediates: Cytoplasmic Nup98 acting as an aggregation-promoting factor.

Deep research

OpenScientist citations 38 citations 2026-05-18T21:45:56.028675

Show evidence (7 references)

PMID:21509508 SUPPORT

"Extracellular accumulations of Abeta, hyperphosphorylation of tau and intracellular neurofibrillary tangle formation have been the hallmarks of Alzheimer's Disease (AD)."

Supports tau hyperphosphorylation and tangle formation as central Alzheimer disease pathology.

PMID:19542604 SUPPORT

"Tau protein is the principal component of the neurofibrillary tangles found in Alzheimer's disease (AD), where it is hyperphosphorylated on serine and threonine residues."

Identifies hyperphosphorylated tau as the principal neurofibrillary tangle component in Alzheimer disease.

PMID:39211286 SUPPORT Model Organism

"Our findings indicate that tau seeding is an early event that occurs well before the appearance of AT8-positive NFT."

Places seed-competent soluble tau upstream of the tangles the model names, and the same study shows that blocking Ser214 phosphorylation reduces propagation in mouse primary neurons and seeding in a Drosophila tauopathy, supplying a perturbation arm. Two mouse models plus Drosophila only.

+ 4 more references

Synaptic Failure Convergence Model

synaptic_failure_convergence_model CANONICAL Early-Onset Alzheimer's Disease Late-Onset Alzheimer's Disease

Evidence balance 7 support 1 refute

Amyloid-beta, tau, inflammatory, vascular, oxidative, and infectious stressors converge on synaptic plasticity, neurotransmitter release, and network function, producing cognitive decline.

Pathograph links

Amyloid Plaque Formation -> Synaptic Dysfunction DIRECT Soluble amyloid-beta assemblies are modeled as direct toxic contributors to synaptic dysfunction and cognitive impairment.

Synaptic Dysfunction -> Memory Loss INDIRECT KNOWN INTERMEDIATES Synaptic plasticity failure and aberrant neurotransmitter release are modeled as proximate drivers of cognitive decline, including memory impairment. Intermediates: Hippocampal and cortical network dysfunction.

Oxidative Stress -> Synaptic Dysfunction INDIRECT UNKNOWN INTERMEDIATES Oxidative neuronal injury is modeled as one convergent stressor that worsens synaptic and network function.

Mitochondrial Quality-Control Failure -> Memory Loss INDIRECT KNOWN INTERMEDIATES Mitochondrial quality-control failure is modeled as one upstream stressor that can worsen cognitive outcomes, but cognitive resilience may decouple pathology burden from symptoms. Intermediates: Amyloid, tau, neuroinflammatory, and synaptic consequences of impaired mitophagy.

Vascular Dysfunction -> Synaptic Dysfunction INDIRECT KNOWN INTERMEDIATES Reduced perfusion and neurovascular-unit injury are modeled as contributors to neuronal and synaptic dysfunction. Intermediates: Inadequate microcirculation in active brain regions.

HSV-1 Reactivation in RORB+ Glutamatergic Neurons -> Synaptic Dysfunction INDIRECT UNKNOWN INTERMEDIATES HSV-1 reactivation in selectively vulnerable cortical neurons is modeled as a possible upstream contributor to loss of synaptic integrity and cognitive-network failure.

Adaptive Immune T Cell Response to Tau Pathology -> Synaptic Dysfunction INDIRECT KNOWN INTERMEDIATES T cell-derived IFN-gamma and cytotoxic activity, engaged downstream of tau pathology, amplify neuronal and synaptic loss; IFN-gamma/PD-1 blockade reduces brain atrophy. Intermediates: IFN-gamma signaling and CD8+ cytotoxic effector function driving neuronal loss.

Complement-Mediated Microglial Synapse Elimination -> Synaptic Dysfunction DIRECT Complement tagging and microglial engulfment remove intact synapses, producing the synapse loss that underlies impaired plasticity and neurotransmission.

Interneuron Dysfunction and Network Hypersynchrony -> Memory Loss INDIRECT KNOWN INTERMEDIATES Loss of inhibitory control and network hypersynchrony are modeled as direct contributors to cognitive impairment; restoring Nav1.1 in a model reverses both the hypersynchrony and the memory deficit. Intermediates: Degraded gamma oscillations and hippocampal-cortical network hypersynchrony.

Deep research

OpenScientist citations 37 citations 2026-05-18T22:06:17.357864

Show evidence (8 references)

PMID:27662312 SUPPORT

"Compelling evidence suggests that soluble amyloid-beta (Abeta) and hyperphosphorylated tau serve as toxins in the dysfunction of synaptic plasticity and aberrant neurotransmitter (NT) release at synapses consequently causing a cognitive decline in Alzheimer's disease (AD)."

Directly supports synaptic dysfunction as a convergence point downstream of amyloid-beta and tau toxicity.

PMID:12973746 SUPPORT

"Histological studies indicate loss of pyramidal neurones and their synapses in Alzheimer's disease (AD), this together with biochemical evidence suggests presynaptic (and postsynaptic) glutamatergic hypoactivity."

Supports loss of synapses and altered neurotransmission in Alzheimer disease.

PMID:40491249 SUPPORT Human Clinical

"CSF SV2A was lower in AD compared to CN participants. Within the AD group, CSF SV2A was highly correlated with SV2A PET."

Concordant fluid and imaging measurement of synaptic density itself rather than of its inferred consequences. Small: 21 participants with early Alzheimer disease and 7 cognitively normal controls, all scanned with carbon-11 UCB-J. Other cerebrospinal-fluid synaptic proteins were higher in Alzheimer disease and unrelated to SV2A PET, which sharpens the specificity of the SV2A measure.

+ 5 more references

Neuroimmune-Glial Amplification Model

neuroimmune_glial_amplification_model ALTERNATIVE Early-Onset Alzheimer's Disease Late-Onset Alzheimer's Disease

Evidence balance 9 support 1 refute

Microglia, astrocytes, complement, and inflammasome pathways are modeled as stage-dependent disease-modifying mechanisms that can respond to amyloid-beta and tau pathology, reinforce amyloid aggregation and inflammatory injury, or restrain amyloid-associated tau spread through protective phagocytic and plaque-barrier functions.

Retained as ALTERNATIVE despite strong GWAS support because microglial and astrocyte programs appear context-dependent: the same innate immune pathways can mediate protective amyloid/tau clearance or pathogenic complement, inflammasome, and synapse-loss amplification depending on disease stage, cell state, and genetic background. ALTERNATIVE is the right value for exactly what the OpenScientist neuroimmune_glial_amplification_model report concluded — an established disease-amplifying mechanism that is not a sufficient standalone causal model — and its proposed status values (PARTIALLY SUPPORTED, DISEASE MODIFIER) are not in MechanisticHypothesisStatusEnum. That report's microglia-to-A1-astrocyte citation (PMID:41345807) was rejected in assessment and deliberately not curated: it is a neuroHIV study of HIV-1 Tat whose Alzheimer content is background restatement of the Liddelow paradigm.

Pathograph links

Neuroinflammation -> Oxidative Stress DIRECT Activated glial cells are modeled as sources of inflammatory mediators and oxygen radicals that amplify oxidative neuronal injury.

Neuroinflammation -> Vascular Dysfunction INDIRECT KNOWN INTERMEDIATES Neuroinflammation is modeled as a contributor to blood-brain barrier breakdown and neurovascular-unit dysfunction. Intermediates: Inflammatory remodeling of cerebral microvessels and blood-brain barrier injury.

Mitochondrial Quality-Control Failure -> Neuroinflammation INDIRECT KNOWN INTERMEDIATES Impaired mitochondrial quality control is modeled as licensing NLRP3 inflammasome activation through cytosolic mitochondrial DNA release, in addition to reduced mitophagy-mediated resolution of glial activation. Intermediates: Cytosolic mitochondrial DNA release licensing NLRP3 inflammasome assembly.; Reduced mitophagy-dependent resolution of microglial activation.

PARP1-Mediated Parthanatos -> Neuroinflammation INDIRECT KNOWN INTERMEDIATES PARP1 overactivation amplifies NF-κB-dependent neuroinflammatory responses and promotes microglial activation. Intermediates: PARP1-driven NAD+ depletion and parthanatos trigger NF-κB signaling.; AIF-MIF complexes activate inflammasome and inflammatory cytokine cascades.

Necroptotic Neuronal Death -> Neuroinflammation INDIRECT UNKNOWN INTERMEDIATES Necroptosis is lytic: unlike apoptosis it releases intracellular contents, so necroptotic neuronal death is modeled as feeding the glial inflammatory response rather than being silently cleared.

Senescent Cell Accumulation -> Neuroinflammation DIRECT The senescence-associated secretory phenotype is modeled as a chronic proinflammatory input to glial activation; senolytic clearance reduces neuroinflammation.

Microglial Lipid Droplet Accumulation -> Neuroinflammation DIRECT Lipid-droplet-accumulating microglia are a proinflammatory, phagocytically defective state, so their accumulation is modeled as amplifying neuroinflammation while reducing amyloid clearance capacity.

Deep research

OpenScientist citations 38 citations 2026-05-18T22:25:37.317379

Show evidence (10 references)

PMID:28019679 SUPPORT

"NOD-like receptor (NLR) family, pyrin domain containing 3 and 1 inflammasomes, present in myeloid cells and neurons, respectively, represent key components of the innate immune reaction observed in Alzheimer patient brains."

Supports inflammasome-mediated innate immune activation in Alzheimer disease brains.

PMID:23930978 SUPPORT

"A vicious cycle of inflammation has been formed between Abeta accumulation, activated microglia, and microglial inflammatory mediators, which enhance Abeta deposition and neuroinflammation."

Supports a self-reinforcing loop between amyloid-beta deposition and microglial inflammation.

PMID:36361780 SUPPORT Other

"Microglia work to reestablish efficiency and stop further degeneration in the early stages of AD but mainly fail in the illness's later phases."

Review-level support for treating microglial activation as stage-dependent rather than uniformly protective or uniformly damaging.

+ 7 more references

Vascular and Blood-Brain Barrier Clearance Model

vascular_bbb_clearance_model ALTERNATIVE Late-Onset Alzheimer's Disease

Evidence balance 5 support 1 refute

Cerebral blood flow changes, neurovascular-unit injury, blood-brain barrier dysfunction, and impaired amyloid-beta clearance are modeled as causal or reinforcing contributors to Alzheimer disease progression.

Retained as ALTERNATIVE. The mechanistic evidence is strong and now molecular (APOE4-graded pericyte degeneration through cyclophilin A and MMP-9, PICALM linking the two endothelial clearance routes, hypoperfusion raising amyloid in a transgenic model), but the human causal proof is missing: no interventional trial has shown that correcting vascular dysfunction prevents or slows amyloid accumulation, and the rrAD trial (NCT02913664) is pending with neurocognitive rather than serial amyloid PET primary outcomes. The model's reach is also bounded at tau — see the vascular-to-tau knowledge gap. Assessed from the OpenScientist vascular_bbb_clearance_model report; that report's ARIA-E incidence figure of 22-36% is not in the Cochrane review it cites (PMID:41985900 reports an absolute risk difference of 107 more per 1000) and was deliberately not curated, nor was its CL:0000669 lead, which names the pericyte cell type rather than pericyte death.

Pathograph links

Deep research

OpenScientist citations 35 citations 2026-05-18T22:42:25.259499

Show evidence (6 references)

PMID:28902142 SUPPORT

"Thus, current evidence suggests that BBB dysfunction may causatively and consequently contribute to AD pathogenesis, forming a vicious cycle between brain Abeta accumulation and neurovascular unit impairments during disease progression."

Supports a bidirectional causal cycle between blood-brain barrier dysfunction, amyloid-beta accumulation, and neurovascular injury.

PMID:26898552 SUPPORT

"There is accumulating evidence suggesting that changes in brain perfusion are present long before the clinical symptoms of Alzheimer's disease (AD), perhaps even before amyloid-beta accumulation or brain atrophy."

Supports early cerebral perfusion changes as part of the vascular model.

PMID:25757756 SUPPORT Human Clinical

"we show accelerated pericyte degeneration in AD APOE4 carriers >AD APOE3 carriers >non-AD controls, which correlates with the magnitude of BBB breakdown to immunoglobulin G and fibrin"

Human post-mortem molecular mechanism for the barrier arm, which this hypothesis otherwise stated only at review level. Cyclophilin A and MMP-9 accumulate in pericytes and endothelial cells in an APOE4-graded fashion; LRP1 was reduced similarly in APOE4 and APOE3 carriers, so the genotype effect runs through the CypA-MMP-9 route rather than through LRP1 abundance.

+ 3 more references

Glymphatic Clearance Model

glymphatic_clearance_model EMERGING Late-Onset Alzheimer's Disease

Evidence balance 3 support 1 refute

Failure of the sleep-dependent perivascular (glymphatic) clearance route is modeled as an upstream contributor to amyloid-beta accumulation, acting through reduced CSF-interstitial fluid exchange rather than through transcytotic blood-brain-barrier efflux (vascular_bbb_clearance_model) or intracellular degradation (autophagy_lysosomal_clearance_model). Held as EMERGING because the human evidence is dominated by imaging surrogates and small intrathecal-tracer cohorts, and because the microscopic transport mode underpinning the mechanism is itself contested; see the glymphatic_dysfunction module for the curated controversy.

Deliberately kept separate from vascular_bbb_clearance_model. The two describe different anatomical routes for amyloid-beta egress (perivascular CSF-ISF exchange versus receptor-mediated transcytosis across the blood-brain barrier) and should not be merged even though both are "clearance" models. PMID:41762118 is curated here rather than on vascular_bbb_clearance_model, where the OpenScientist report placed it, for that reason: it measures perivascular influx and clearance, not barrier transcytosis.

Pathograph links

Show evidence (4 references)

PMID:27893874 SUPPORT Human Clinical

"When controlling for age, loss of perivascular AQP4 localization was associated with increased amyloid-β burden"

Human postmortem dose-response between loss of the glymphatic conduit protein and amyloid burden, the core claim of this hypothesis group.

PMID:33004510 SUPPORT Other

"The glymphatic system, which clears the brain of protein waste products, is mostly active during sleep."

Establishes the sleep-gated clearance premise that distinguishes this hypothesis group from the other clearance models in this entry.

PMID:41762118 SUPPORT Model Organism

"AppNL-F mice showed reductions in glymphatic influx and clearance at 6 months, preceding substantial Aβ plaque deposition."

Temporal ordering in a knock-in model that avoids overexpression artefacts: the clearance deficit appears before substantial plaque, and tracks loss of parenchymal border macrophages, which acute intracisternal amyloid-beta alone reproduces in wild-type mice.

+ 1 more reference

Autophagy-Lysosomal Clearance Model

autophagy_lysosomal_clearance_model EMERGING Early-Onset Alzheimer's Disease Late-Onset Alzheimer's Disease

Evidence balance 8 support

Defective autophagic flux, lysosomal transport, mitophagy, and aggregate clearance are modeled as mechanisms that promote amyloid-beta accumulation, tau accumulation, mitochondrial stress, and downstream cognitive impairment.

Retained as EMERGING because PSEN1-linked lysosomal acidification evidence provides a strong mechanistic anchor and mitophagy perturbation links mitochondrial quality control to amyloid, tau, inflammatory, and cognitive readouts. The broader contribution of autophagy-lysosomal and mitochondrial clearance defects to late-onset Alzheimer disease remains incompletely resolved in human longitudinal data. The descriptive claim and the causal claim should be kept apart when reading this block. That autophagic and endolysosomal failure is an early, Alzheimer-specific feature is well supported by human post-mortem tissue and by pathway-level genetics; that it is a primary driver rather than an early consequence is not, because the human evidence is cross-sectional and the perturbations are rodent. The therapeutic claim is weaker still: no autophagy-enhancing therapy has completed phase III testing in Alzheimer disease, so that arm is entirely preclinical. The OpenScientist autophagy_lysosomal_clearance_model report asked for these three claims to carry separate statuses (ESTABLISHED, EMERGING, SPECULATIVE); the schema has one status per hypothesis and neither ESTABLISHED nor SPECULATIVE is a permitted value, so the distinction is recorded here instead.

Pathograph links

PARP1-Mediated Parthanatos -> Oxidative Stress INDIRECT KNOWN INTERMEDIATES PARP1 hyperactivation and NAD+ depletion impair mitochondrial redox homeostasis, sustaining oxidative stress through disruption of antioxidant SIRT1 and mitophagy. Intermediates: NAD+ depletion impairs SIRT1-dependent mitochondrial quality control.; Dysregulated mitophagy fails to clear damaged mitochondria.

PARP1-Mediated Parthanatos -> Mitochondrial Quality-Control Failure INDIRECT KNOWN INTERMEDIATES PARP1 hyperactivity and NAD+ depletion impair mitochondrial dynamics, mitophagy, and ATP production, contributing to quality-control collapse. Intermediates: NAD+ depletion reduces mitochondrial ATP production.; SIRT1 dysregulation impairs mitochondrial biogenesis and turnover.

Neuronal Endosomal-Lysosomal Acidification Failure -> Autophagy-Lysosomal Dysfunction DIRECT APP-βCTF inhibition of the v-ATPase is modeled as one upstream cause of the broader autophagic-flux failure curated in that node.

Deep research

OpenScientist citations 25 citations 2026-05-18T23:00:45.134591

Show evidence (8 references)

PMID:20541250 SUPPORT Model Organism

"In PS1 null blastocysts, neurons from mice hypomorphic for PS1 or conditionally depleted of PS1, substrate proteolysis and autophagosome clearance during macroautophagy are prevented as a result of a selective impairment of autolysosome acidification and cathepsin activation. These deficits are..."

Demonstrates that PS1 is required for v-ATPase V0a1 targeting to lysosomes; PSEN1 dysfunction therefore directly impairs lysosomal acidification and autophagosome clearance, anchoring the early-onset Alzheimer disease mechanism for this hypothesis group.

PMID:30742114 SUPPORT Model Organism

"Mitophagy diminishes insoluble Aβ1-42 and Aβ1-40 and prevents cognitive impairment in an APP/PS1 mouse model through microglial phagocytosis of extracellular Aβ plaques and suppression of neuroinflammation."

Supports mitophagy as a mitochondrial quality-control mechanism that can alter amyloid burden, neuroinflammation, and cognitive impairment in an Alzheimer disease mouse model.

PMID:30124770 SUPPORT Computational

"We found that both gene sets were enriched for genetic Alzheimer's associations: an enrichment score 3.67 standard deviations from the null model (P = 0.00012) was detected for AphagEndoLyso"

Population-level genetic support for the pathway rather than for a single gene: an 891-gene autophagy and endolysosomal set is enriched for Alzheimer risk variants across three genome-wide association studies with a combined n of 62,415. Pathway-enrichment analysis over published summary statistics, hence COMPUTATIONAL.

+ 5 more references

HSV-1 Reactivation Model

hsv1_reactivation_model EMERGING Late-Onset Alzheimer's Disease

Evidence balance 2 support

Latent HSV-1 reactivation in selectively vulnerable RORB+ glutamatergic neurons is modeled as a possible upstream contributor to neuronal vulnerability and later synaptic-network failure.

Retained as EMERGING. The evidence is human, cell-type-resolved, and hypothesis-generating, but causality between HSV-1 expression and Alzheimer disease progression remains unresolved.

Pathograph links

Deep research

OpenScientist citations 33 citations 2026-05-19T02:39:53.213307Z

Show evidence (2 references)

PMID:42094473 SUPPORT Human Clinical

"Integrative single-nucleus analyses resolve direct evidence of HSV-1 expression in RORB+ glutamatergic neurons, implicating viral reactivation in a neuronal population progressively lost during dementia."

Supports HSV-1 reactivation in a selectively vulnerable Alzheimer-relevant neuronal population.

PMID:42094473 SUPPORT Human Clinical

"identifying recurrent herpes simplex virus 1 (HSV-1) reactivation in healthy but not pathological post-mortem human brain tissue"

Supports the temporal interpretation that HSV-1 reactivation may precede loss of vulnerable neurons rather than simply appear in end-stage pathological tissue.

Arc-Dependent EV-Mediated Tau Propagation Model

ev_mediated_tau_propagation_model EMERGING Early-Onset Alzheimer's Disease Late-Onset Alzheimer's Disease

Evidence balance 7 support

Cell-to-cell (prion-like) spread of tau pathology is modeled as being driven, in part, by packaging of pathological tau into neuronal extracellular vesicles (EVs). The activity-regulated, capsid-forming neuronal protein Arc binds tau directly and is required for efficient release of tau in EVs; Arc-tau co-packaging seeds tau aggregation in recipient neurons, propagating tangle pathology along connected circuits. In this model EV-tau release is partly protective for the donor neuron (eliminating toxic intracellular tau) but drives intercellular transmission of seed-competent tau.

EMERGING. Demonstrated in primary neurons, rTg4510 tau-transgenic / Arc-KO mice, and human postmortem brain EVs (Tyagi et al., Cell 2026). EV-tau is one of several proposed routes of tau spread (free/naked tau uptake via LRP1, tunneling nanotubes, trans-synaptic transfer); the relative in vivo contribution of each, and how Arc levels modulate EV-tau release as disease progresses, remain to be resolved. By 8 months in the transgenic model, tau pathology was similar between Arc-KO and control, so Arc loss does not overtly accelerate late-stage pathology despite blocking transmission. OpenScientist deep research (July 2026; kb/hypotheses/Alzheimer_Disease/ev_mediated_tau_propagation_model) judged the model PARTIALLY SUPPORTED (EMERGING), with the following qualifications: (1) the Arc-tau binding and Arc-KO transmission phenotype rest on a single source (Tyagi et al., 2026) with no independent replication; (2) most extracellular tau is free-floating rather than EV-encapsulated (PMID:29495441), so the quantitative in vivo weight of the EV route versus LRP1/HSPG free-tau uptake and tunneling nanotubes is unresolved; (3) Abeta deposition suppresses Arc expression by 14-58% (PMID:19556514), potentially attenuating the mechanism in plaque-rich regions; and (4) the model was tested with P301L (4R) tau, not authentic AD-type 3R/4R paired helical filaments, and ARC carries no GWAS/GenCC association with AD. Supporting biology (Arc capsid EV transfer, PMID:29328916; EV-tau seeding, PMID:27030011; endolysosomal-escape gating, PMID:33417012) is independently established. Reassessment of that report adds three things. The single-source dependency carries a disclosed commercial interest: the senior author of PMID:42372723 co-founded VNV, LLC and consults for Aera Therapeutics, which licenses patents covering Arc capsids, which is the context in which independent replication matters most. The Arc capsid work (PMID:29328916) demonstrates transfer of Arc's own mRNA, so it establishes plausibility of the vehicle and not of tau as cargo. And the 8-month convergence of tau pathology between Arc-knockout and control animals, which the report uses to argue for restricting the model to early disease, is not in the abstract this repository holds; no early-stage restriction should be curated from it without the full text. Status stays EMERGING. See the Arc replication-and-genetic-support knowledge gap.

Pathograph links

Deep research

OpenScientist citations 26 citations 2026-07-06T11:06:45.270406

Show evidence (7 references)

PMID:42372723 SUPPORT Model Organism

"Strikingly, intercellular tau transmission is almost absent in Arc KO mice."

Loss of Arc nearly abolishes neuron-to-neuron tau transmission in mice, supporting Arc-dependent EV packaging as a mechanism of prion-like tau spread.

PMID:42372723 SUPPORT Human Clinical

"Moreover, Arc levels in brain-derived EVs isolated from human Alzheimer's disease (AD) brains show a strong positive correlation with phosphorylated EV-tau levels."

Human AD brain-derived EVs link Arc levels to phosphorylated EV-tau, extending the EV-mediated propagation model to human disease.

PMID:29328916 SUPPORT In Vitro

"Endogenous Arc protein is released from neurons in extracellular vesicles that mediate the transfer of Arc mRNA into new target cells"

Arc self-assembles into virus-like capsids and is released from neurons in EVs that transfer cargo to recipient cells, establishing the biological plausibility for Arc packaging tau into EVs.

+ 4 more references

APOE Risk-Modulation Model

apoe_risk_modulation_model ALTERNATIVE Late-Onset Alzheimer's Disease

Evidence balance 2 support

Apolipoprotein E (APOE) genotype is modeled as the strongest common-variant modifier of late-onset Alzheimer disease risk, acting upstream of and across multiple pathologies rather than through a single pathway. In proportion to APOE4 gene dose, APOE4 accelerates amyloid-beta seeding/aggregation and impairs its clearance, exacerbates tau-mediated neurodegeneration and neuroinflammation, promotes glial lipid/cholesterol accumulation, and impairs blood-brain-barrier integrity, whereas APOE2 and the rare protective variants APOE3 Christchurch (R136S) and APOE3 Jacksonville (V236E) blunt these effects. The model predicts that lowering or structurally modifying APOE — especially APOE4 — should ameliorate several AD pathologies simultaneously, motivating APOE-directed therapeutics (ASO knockdown, AAV-APOE2, anti-aggregation small molecules).

ALTERNATIVE. APOE modulation overlaps and interacts with the amyloid, tau, neuroimmune, vascular/BBB, and lipid models rather than competing with them; it is curated as a distinct hypothesis because the protective-variant (Christchurch/Jacksonville) and APOE-lowering evidence points to APOE itself as an actionable node upstream of multiple pathologies. Sex- and ancestry-dependent APOE risk and the precise molecular mechanism (aggregation, receptor binding, lipidation) remain incompletely resolved.

Pathograph links

Show evidence (2 references)

PMID:28959956 SUPPORT Model Organism

"P301S/E4 mice develop markedly more brain atrophy and neuroinflammation than P301S/E2 and P301S/E3 mice, whereas P301S/EKO mice are largely protected from these changes."

Isoform-dependent modulation of tau-driven neurodegeneration and neuroinflammation (E4 > E3/E2, and near-protection when APOE is absent) anchors APOE as a cross-pathology risk modifier acting beyond amyloid.

PMID:31686034 SUPPORT Human Clinical

"The individual had two copies of the APOE3 Christchurch (R136S) mutation, unusually high brain amyloid levels and limited tau and neurodegenerative measurements."

A homozygous APOE3 Christchurch carrier resisted autosomal-dominant AD for roughly three decades despite heavy amyloid, dissociating amyloid burden from tau and neurodegeneration and implicating APOE receptor-binding avidity as a protective lever.

Adaptive Immune (T Cell) Neurodegeneration Model

adaptive_immune_tcell_model EMERGING Early-Onset Alzheimer's Disease Late-Onset Alzheimer's Disease

Evidence balance 7 support 1 refute

Beyond innate microglial responses, a clonally expanded, predominantly cytotoxic (CD8+) T cell response is modeled as an active driver of tau-associated neurodegeneration. In tauopathy — but not pure amyloidosis — microglia are required to recruit T cells to sites of tau pathology; infiltrating T cells transition from activated to exhausted states with restricted TCR clonality, and IFN-gamma/PD-1 signaling amplifies neuronal loss. In the tau context the model predicts that depleting T cells or blocking IFN-gamma/PD-1 signaling should be neuroprotective. Critically, this therapeutic valence is tau-scoped, not universal: in a pure-amyloid context PD-1 checkpoint blockade instead drives an IFN-gamma-dependent myeloid response that clears Aβ and improves cognition, so "block IFN-gamma/PD-1" is not a general prescription. The model frames the cytotoxic-CD8 arm as a shared, targetable axis across tau-driven neurodegenerative disease.

EMERGING. The core functional evidence (T cell depletion, IFN-gamma/PD-1 inhibition, cDC1 ablation) is from mouse tauopathy models, with correlative human support (increased cytotoxic T cells in AD brain/CSF tracking tau, not amyloid). The antigen-presenting route is now resolved to cDC1 cross-presentation (PMID:41890046), but the recognized antigen(s) remain unidentified — and two findings temper the "antigen-driven clonal CD8" reading: aging expands antigen-independent virtual-memory CD8 T cells that are clonal and cytotoxic without priming (PMID:42432776), and the clonally expanded CD8 TCRs recovered from human AD CSF were specific to Epstein-Barr virus antigens, not tau (PMID:31915375). IFN-gamma is also not CD8-exclusive (NK cells are a parallel source tracking cognitive decline, PMID:11268360), and whether B cells contribute remains open. See the antigen-identity and IFN-gamma-valence knowledge-gap discussions. Complements — does not replace — the innate neuroimmune_glial_amplification_model. Human cross-tauopathy support is now curated from FTLD-tau (PMID:37703311) and chronic traumatic encephalopathy (PMID:41957675), which also bound the microglia-as-recruiter step: that step is strong in the mouse tauopathy perturbation and unsupported in human FTLD-tau, where microglial activation is minimal and T cell recruitment is chemokine-driven. The OpenScientist adaptive_immune_tcell_model report describes the infiltrating cells as CD103+ tissue-resident memory cells and proposes an ontology binding on that basis; the cited paper (PMID:40993111) reports the opposite marker polarity — CD103-negative — so neither the label nor the binding was curated. Its reading of PMID:15549735 as showing regulatory T cells to be neuroprotective is also inverted (eliminating them attenuated the lesion) and was not curated.

Pathograph links

Deep research

OpenScientist citations 23 citations 2026-07-11T13:56:46.720028

Show evidence (8 references)

PMID:36890231 SUPPORT Model Organism

"We found that mice with tauopathy but not those with amyloid deposition developed a unique innate and adaptive immune response and that depletion of microglia or T cells blocked tau-mediated neurodegeneration."

Depletion of either microglia or T cells blocks tau-mediated neurodegeneration in a tauopathy model, establishing the adaptive T cell arm as functionally required rather than a bystander.

PMID:36890231 SUPPORT Model Organism

"Inhibition of interferon-γ and PDCD1 signalling both significantly ameliorated brain atrophy."

IFN-gamma and PD-1 (PDCD1) blockade reduce brain atrophy, nominating specific, druggable nodes within the adaptive-immune neurodegeneration axis.

PMID:41890046 SUPPORT Model Organism

"We demonstrate that tauopathy mice deficient in cDC1 are markedly protected against tau-mediated neurodegeneration and display a selective decrease in brain CD8+ T cell infiltration and glial reactivity."

Independent Holtzman-group follow-up: ablating conventional type-1 dendritic cells (cDC1) — the cross-presenting antigen-presenting cell — protects tauopathy mice and reduces brain CD8+ T cell infiltration, resolving the antigen-presenting route that primes the pathogenic CD8 response.

+ 5 more references

Anticholinergic Muscarinic Burden Model

anticholinergic_muscarinic_burden_model EMERGING

Evidence balance 2 support 1 no evidence

Sustained pharmacologic blockade of central muscarinic acetylcholine receptors — the property shared by first-generation antihistamines (diphenhydramine), tricyclic antidepressants, bladder antimuscarinics, and antiparkinson anticholinergics — is modeled as removing M1-dependent tonic restraint on amyloidogenic APP processing and on GSK3-beta-mediated tau phosphorylation. Under this model, cumulative anticholinergic exposure is a modifiable disease-risk factor rather than only a cause of the reversible cognitive impairment classically attributed to these drugs.

EMERGING, and deliberately not promoted further. The mechanistic arm is MODEL_ORGANISM only (M1 knockout and dicyclomine in 3xTg-AD/Tg-SwDI mice); the human arm is entirely observational and has two unresolved threats. (1) Reverse causation: prodromal dementia produces insomnia, depression, and urinary urgency, which are the indications for these drugs. (2) Class specificity: the largest UK studies concentrate the signal in antidepressant, urological, and antiparkinson anticholinergics, with antipsychotics and antiepileptics significant in QResearch (PMID:31233095) but antipsychotics null in CPRD (PMID:29695481), and the largest antihistamine-specific cohort finds a dose-response for second-generation agents that barely enter the CNS as well as for first-generation ones (PMID:38935035). Read that last result carefully in both directions: the gradient is steeper for the CNS-penetrant class (first-generation 1.13/1.29/1.51 across rising cumulative dose versus second-generation 1.11/1.19/1.26, the authors concluding that first-generation agents carry the higher risk), which is what a central mechanism predicts, while the residual non-null second-generation signal is left unexplained by that mechanism and is the open puzzle. Do NOT read the antihistamine nulls as a clean negative: the CPRD analysis records only prescriptions, and its authors note that over-the-counter supply is invisible in their data and that antihistamine use is therefore underestimated, which matters most for exactly the drug that prompted this curation. See the two attached discussions before strengthening any claim here.

Pathograph links

Muscarinic M1 Receptor Signaling Loss -> Amyloid Plaque Formation INDIRECT KNOWN INTERMEDIATES Withdrawal of M1 signaling shifts APP processing away from the ADAM17/alpha-secretase route toward amyloidogenic cleavage, increasing amyloid-beta available for deposition. Intermediates: Reduced ADAM17 (alpha-secretase) activation and consequent diversion of APP to beta/gamma-secretase cleavage.

Muscarinic M1 Receptor Signaling Loss -> Neurofibrillary Tangle Formation INDIRECT KNOWN INTERMEDIATES Loss of M1 signaling de-represses GSK3-beta and alters protein kinase C activity, promoting tau hyperphosphorylation and tangle formation. Intermediates: Increased GSK3-beta activity and altered protein kinase C signaling driving tau hyperphosphorylation.

Show evidence (3 references)

PMID:16504943 SUPPORT Model Organism

"We further demonstrate that administration of dicyclomine, an M1 antagonist, exacerbates the Abeta and tau pathologies."

Pharmacologic M1 antagonism — not merely genetic loss — worsens both amyloid and tau pathology, which is the mechanistic premise the model extrapolates from to chronic human anticholinergic exposure.

PMID:25621434 SUPPORT Human Clinical

"A 10-year cumulative dose-response relationship was observed for dementia and Alzheimer disease"

The cumulative dose-response in a prospective cohort is the human observation the model is built to explain.

PMID:38063254 NO_EVIDENCE Human Clinical

"The evidence from RCTs was of very low certainty so cannot support or refute the hypothesis that actively reducing or stopping prescription of medications with anticholinergic properties can improve cognitive outcomes in older people."

The interventional test of the model is uninformative in both directions, which is why this hypothesis group is curated EMERGING rather than ALTERNATIVE or CANONICAL.

Complement-Dependent Microglial Synapse Elimination Model

complement_synaptic_pruning_model ALTERNATIVE Early-Onset Alzheimer's Disease Late-Onset Alzheimer's Disease

Evidence balance 9 support

Synapse loss — the pathological change that correlates best with cognitive decline — is modeled as an active, receptor-mediated phagocytic process rather than as passive degeneration secondary to amyloid or tau toxicity. The classical complement cascade that prunes surplus synapses during development (C1q tagging, C3 opsonization, microglial CR3 engagement) is modeled as being reactivated early in disease, so that microglia engulf structurally intact synapses. Soluble amyloid-beta oligomers and phosphorylated tau both act as upstream triggers, which makes this the mechanism by which two otherwise separate proteinopathies converge on the same cellular endpoint.

ALTERNATIVE rather than CANONICAL. The necessity evidence (C1q, C3 and CR3 blockade) is entirely mouse; the same paper that supplies the best human evidence for glial synapse ingestion states that direct human evidence for glial involvement in synapse removal remained to be established, and implicates MFG-E8 rather than complement. Curated as a distinct hypothesis group rather than folded into neuroimmune_glial_amplification_model because it makes a specific, falsifiable claim about a named cascade acting on a named substrate (the synapse), and because its therapeutic prediction — anti-C1q antibody — is being tested clinically. See the attached HUMAN_MODEL_MISMATCH discussion before strengthening the human claim.

Pathograph links

Deep research

OpenScientist citations 23 citations 2026-08-28T07:54:33.743613

Show evidence (9 references)

PMID:27033548 SUPPORT Model Organism

"C1q, the initiating protein of the classical complement cascade, is increased and associated with synapses before overt plaque deposition."

Places complement tagging of synapses upstream of plaque deposition, which is what distinguishes this model from synapse loss as a late consequence of established amyloid pathology.

PMID:27033548 SUPPORT Model Organism

"Inhibition of C1q, C3, or the microglial complement receptor CR3 reduces the number of phagocytic microglia, as well as the extent of early synapse loss."

Blocking three separate steps of the same cascade each reduces synapse loss, establishing the pathway as required rather than merely present.

PMID:30392797 SUPPORT Model Organism

"At synapses, C1q decorated perisynaptic membranes, accumulated in correlation with phospho-Tau, and was associated with augmented microglial engulfment of synapses and decline of synapse density."

Independent laboratory, and a tau rather than amyloid driver, reaching the same C1q-microglia-synapse axis — the strongest evidence that the model is not specific to amyloid models.

+ 6 more references

Endolysosomal Origin ("Inside-Out" Amyloid) Model

endolysosomal_origin_model EMERGING Early-Onset Alzheimer's Disease Late-Onset Alzheimer's Disease

Evidence balance 8 support

Amyloid pathology is modeled as beginning inside the neuron rather than in the extracellular space. Endocytic-pathway activation with enlargement of Rab5-positive early endosomes is the earliest recognized neuronal change in sporadic Alzheimer disease, preceding plaque deposition; APP-beta-C-terminal fragment accumulation then inhibits the lysosomal v-ATPase, autolysosomes fail to acidify, and amyloid-beta builds up within de-acidified autolysosomes until the neuron ruptures and its contents become the plaque core (the PANTHOS pattern). Under this model the senile plaque is the tombstone of a dead neuron, not a deposit of secreted peptide, which inverts the direction of the classical extracellular cascade.

EMERGING, and deliberately curated as a competitor to — not a restatement of — autophagy_lysosomal_clearance_model. That model says defective clearance lets amyloid accumulate; this one says the plaque is *generated* inside the autolysosomal compartment of a neuron that then dies. The distinction is testable and matters therapeutically: it predicts that lowering APP-βCTF or restoring v-ATPase activity acts upstream of anything an anti-amyloid antibody can reach. Held EMERGING because the human evidence is strong for the endosomal lesion (Cataldo, human postmortem) but thin for PANTHOS itself, where the only verified human statement is that the pattern is present in Alzheimer brains. The v-ATPase mechanism and the PANTHOS phenotype come from the same laboratory, so they are not independent replications of each other.

Pathograph links

Deep research

OpenScientist citations 24 citations 2026-08-28T08:09:47.527572

Show evidence (8 references)

PMID:10880397 SUPPORT Human Clinical

"These abnormalities were evident in pyramidal neurons of the neocortex at preclinical stages of disease when Alzheimer-like neuropathology, such as Abeta deposition, was restricted to the entorhinal region."

Human postmortem evidence that neuronal endosomal abnormality is present at preclinical stages, before neocortical amyloid deposition — the temporal ordering the model requires.

PMID:10880397 SUPPORT Human Clinical

"By contrast, endosomes were normal in size at advanced stages of familial AD caused by mutations of presenilin 1 or 2, indicating that altered endocytosis is not a consequence of Abeta deposition."

The paper's own internal control cuts both ways: it shows the endosomal phenotype is not merely downstream of amyloid, but also that it is absent in presenilin-mutation familial disease, so the mechanism is not universal across Alzheimer genetic subtypes.

PMID:15465622 SUPPORT Human Clinical

"Double-immunofluorescence using numerous Abeta antibodies showed that intracellular Abeta localized principally to rab5-positive endosomes in neurons from AD brains and was prominent in enlarged endosomes."

Locates amyloid-beta inside the enlarged endosomal compartment in human Alzheimer brain, connecting the endosomal lesion to the peptide itself.

+ 5 more references

Selective Neuronal Vulnerability Model

selective_neuronal_vulnerability_model EMERGING Late-Onset Alzheimer's Disease

Evidence balance 3 support

Alzheimer disease is modeled as a disease of specific, molecularly definable neuronal populations rather than of the cortex as a whole: locus coeruleus noradrenergic neurons and entorhinal cortex layer II excitatory neurons degenerate first and disproportionately, and within the entorhinal cortex the vulnerable excitatory population is marked by RORB. Under this model the question "why these cells?" is a mechanistic question in its own right — cell-intrinsic properties determine where tau pathology and neuronal loss begin, so aggregate burden alone cannot explain the anatomy of the disease.

EMERGING as a mechanistic model, though the underlying observations are long-established and essentially uncontested as *descriptions*. What is unresolved is the mechanism: RORB marks a population, it is not known to be the cause of that population's vulnerability, and no experiment has yet shown that manipulating it changes susceptibility. This group is curated separately from the amyloid and tau models because it addresses a question neither answers — the anatomical selectivity of the disease — and because it supplies the missing link to the entry's HSV-1 node, which is built on the same RORB+ population but cited only through the viral-reactivation work.

Pathograph links

Deep research

OpenScientist citations 19 citations 2026-08-28T08:24:46.334923

Show evidence (3 references)

PMID:33432193 SUPPORT Human Clinical

"We identified RORB as a marker of selectively vulnerable excitatory neurons in the entorhinal cortex and subsequently validated their depletion and selective susceptibility to neurofibrillary inclusions during disease progression using quantitative neuropathological methods."

Single-nucleus transcriptomics of human postmortem brain identifies the vulnerable population molecularly and confirms its depletion by independent quantitative neuropathology.

PMID:8699259 SUPPORT Human Clinical

"Decreases in individual lamina were even more dramatic, with the number of neurons in layer II decreasing by 60% and in layer IV by 40% compared with controls."

Stereological neuron counts showing that laminar loss is already severe at the mildest clinically detectable stage, establishing selectivity as an early rather than end-stage feature.

PMID:27513978 SUPPORT Human Clinical

"The long gap between NFT accumulation and neuronal loss suggests that a second trigger may be necessary to induce neuronal death in AD."

Qualifies any simple "tangles kill the vulnerable cell" reading: in the locus coeruleus, tangle accumulation and neuronal loss are separated by a long interval, so vulnerability to tau inclusion and vulnerability to death are not the same property. Stated by the authors as an interpretation, not a measurement.

Necroptosis Model of Neuronal Death

necroptosis_model EMERGING Early-Onset Alzheimer's Disease Late-Onset Alzheimer's Disease

Evidence balance 4 support

The mode of neuronal death in Alzheimer disease is modeled as necroptosis — RIPK1/RIPK3-triggered, MLKL-executed programmed necrosis — rather than apoptosis. Activated necrosome components are found in granulovacuolar degeneration bodies, a long-recognized but mechanistically unexplained Alzheimer lesion, and their regional burden tracks neuronal loss. This model is significant because it names an executioner: it makes neuronal death a druggable step rather than the passive endpoint of upstream pathology.

EMERGING. The human evidence is correlative and rests substantially on phospho-MLKL and phospho-RIPK immunostaining of postmortem tissue, where antibody specificity and postmortem interval are known problems. The causal arm comes from human neurons xenografted into a mouse amyloid brain — a system whose own headline finding is that mouse neurons in the same brain do not show the phenotype, which is an argument for the model's human relevance and a reminder that rodent neurodegeneration models may miss the death mechanism entirely. Distinct from, and not yet reconciled with, the PARP1 parthanatos route curated in this entry; both are caspase-independent regulated necrosis and no work has established which dominates, or whether they act in different cells or stages.

Pathograph links

Deep research

OpenScientist citations 18 citations 2026-08-28T07:43:02.809538

Show evidence (4 references)

PMID:28758999 SUPPORT Human Clinical

"We found that necroptosis was activated in postmortem human AD brains, positively correlated with Braak stage, and inversely correlated with brain weight and cognitive scores."

Human postmortem dose-response between necroptosis activation and both pathological stage and cognitive outcome.

PMID:31802237 SUPPORT Human Clinical

"GVDn + neurons inversely correlated with neuronal density in the early affected CA1 region of the hippocampus and in the late affected frontal cortex layer III."

Anchors the mechanism to a specific, classically recognized neuropathological lesion and shows its burden tracks neuronal density in both an early- and a late-affected region.

PMID:37708272 SUPPORT Model Organism

"Down-regulation of MEG3 and inhibition of necroptosis using pharmacological or genetic manipulation of receptor-interacting protein kinase 1 (RIPK1), RIPK3, or mixed lineage kinase domain-like protein (MLKL) rescued neuronal cell loss in xenografted human neurons."

Converts the human correlation into a causal claim: blocking three separate necroptosis effectors each rescues loss of human neurons in vivo.

+ 1 more reference

Cellular Senescence Model

cellular_senescence_model EMERGING Late-Onset Alzheimer's Disease

Evidence balance 6 support 1 no evidence

Accumulation of senescent cells — permanently cell-cycle-arrested but metabolically active, and secreting a proinflammatory senescence-associated secretory phenotype — is modeled as an active driver of tau pathology, neuroinflammation, and neuronal loss rather than a passive marker of brain aging. The model's distinguishing prediction is therapeutic and unusually direct: removing senescent cells, genetically or with senolytic drugs, should reduce pathology and preserve cognition even when the senescent cells are a small fraction of the tissue.

EMERGING. Causal evidence remains murine, but it is three independent clearance experiments rather than two: genetic p16 ablation in tauopathy (PMID:30232451), senolytic clearance of plaque-associated oligodendrocyte progenitor cells (PMID:30936558), and senolytic treatment of late-stage tau mice reducing tangle density and neuron loss (PMID:30126037). The cell-of-origin question is the real dispute, and framing it as "mouse says glia, human says neurons" is too clean: PMID:30126037 is a mouse and human study placing the senescent state in tangle-bearing neurons, agreeing with the human eigengene survey (PMID:35531351) and disagreeing with both other mouse studies. What conflicts is the cell-type assignment across studies, not the species. It matters because senolytics kill the cells they target: a therapy validated against senescent glia would, on the neuronal reading, be aimed at postmitotic neurons. Human interventional data now exist and are uninformative in both directions — the SToMP-AD phase 1 open-label trial (NCT04063124, n=5) established central nervous system penetrance of dasatinib and acceptable safety, with no change in cognition or imaging and stable CSF amyloid-beta and tau. See the attached CONTROVERSY discussion. The standing methodological objection to the human survey is that it identifies senescence with a derived eigengene rather than a gold-standard marker; note separately that canonical CDKN2A/p16 is badly under-detected in single-nucleus data, which cuts against any negative call in the rarer cell types.

Pathograph links

Deep research

OpenScientist citations 16 citations 2026-08-28T06:45:05.233096

Show evidence (7 references)

PMID:30232451 SUPPORT Model Organism

"Here we show a causal link between the accumulation of senescent cells and cognition-associated neuronal loss."

Genetic clearance of p16-positive cells as they arise prevents tau pathology and neuronal loss, which is the causal claim the model rests on.

PMID:30936558 SUPPORT Model Organism

"Senolytic treatment of AD mice selectively removed senescent cells from the plaque environment, reduced neuroinflammation, lessened Aβ load, and ameliorated cognitive deficits."

Pharmacological rather than genetic clearance, in an amyloid rather than tau model, reaching the same endpoint — the independent replication the therapeutic prediction needs.

PMID:35531351 SUPPORT Human Clinical

"More than 97% of the senescent cells were excitatory neurons and overlapped with tau-containing neurofibrillary tangles (NFTs)."

The largest human survey supports senescence being present and tangle-associated, but assigns it overwhelmingly to excitatory neurons — contradicting the glial cell-type assignment on which both mouse clearance experiments are built.

+ 4 more references

Myelin and Oligodendrocyte Dysfunction Model

myelin_oligodendrocyte_model EMERGING Late-Onset Alzheimer's Disease

Evidence balance 3 support

Age-related breakdown of myelin and of oligodendrocyte support for the axon is modeled as an upstream risk factor for amyloid deposition rather than a downstream consequence of it. Myelin damage concentrates the amyloidogenic processing machinery in axonal swellings and increases cleavage of amyloid precursor protein; separately, it diverts disease-associated microglia toward myelin debris and away from plaques, so the same lesion both raises amyloid production and lowers its clearance. APOE4 is modeled as acting partly through this route, via aberrant cholesterol deposition in oligodendrocytes and reduced myelination.

EMERGING. The causal claim — myelin dysfunction drives amyloid deposition — is mouse-only; the human work establishes that oligodendrocyte cholesterol dysregulation and reduced myelination occur in APOE4 carriers but not that they cause amyloid deposition in people. The "microglial distraction" half of the mechanism should be read as SPECULATIVE rather than merely less independently replicated: a targeted search for independent primary support of that clearance edge found none, which is a checked absence rather than an unexamined gap. This group was curated in part to give the entry's oligodendrocyte precursor cell plasma proteomic age gap biomarker a lineage mechanism to attach to, but that justification does not survive inspection. The biomarker comes from a plasma clock spanning more than forty cell types (PMID:42297981) in which the strongest APOE4-linked Alzheimer signal is astrocyte aging, so it is consistent with this model without privileging it. Both corrections are curated in full on the Oligodendrocyte and Myelin Dysfunction node.

Pathograph links

Deep research

OpenScientist citations 14 citations 2026-08-28T08:48:09.781343

Show evidence (3 references)

PMID:37258678 SUPPORT Model Organism

"Here we identify genetic pathways of myelin dysfunction and demyelinating injuries as potent drivers of amyloid deposition in mouse models of AD."

Multiple independent myelin-mutant crosses each increase amyloid deposition, establishing the direction of causation in the mouse.

PMID:37258678 SUPPORT Model Organism

"Mechanistically, myelin dysfunction causes the accumulation of the Aβ-producing machinery within axonal swellings and increases the cleavage of cortical amyloid precursor protein."

Supplies the subcellular mechanism linking the myelin lesion to increased amyloid production.

PMID:36385529 SUPPORT Human Clinical

"We show that altered cholesterol localization in the APOE4 brain coincides with reduced myelination."

Human postmortem evidence that oligodendrocyte and myelin pathology is real in the APOE4 brain, supplying the human leg the mouse causal work lacks.

Network Hyperexcitability and Interneuron Dysfunction Model

network_hyperexcitability_model EMERGING Early-Onset Alzheimer's Disease Late-Onset Alzheimer's Disease

Evidence balance 7 support 2 refute

Cognitive decline is modeled as arising in part from a failure of inhibition rather than only from loss of excitatory synapses: amyloid-beta impairs parvalbumin-expressing inhibitory interneurons through reduced levels of the interneuron-predominant voltage-gated sodium channel subunit Nav1.1, degrading gamma oscillations and permitting network hypersynchrony and epileptiform activity. The model predicts that subclinical epileptiform activity should be common in Alzheimer disease, should track faster decline, and should be a treatable contributor to symptoms rather than an incidental finding.

EMERGING, and the therapeutic arm should be read as qualified rather than merely "mixed". The mechanistic case in mice is strong: two independent perturbations — restoring Nav1.1 (PMID:22541439) and reducing tau (PMID:21228179) — each restore inhibition and abolish epileptiform activity. The human limb of the Nav1.1 claim remains a postmortem protein-level observation. The prevalence claim is now replicated in an independent cohort with detection across the Alzheimer continuum including preclinical disease (PMID:38263073), which strengthens it considerably beyond the single seed study. The treatability prediction is where the model is weakest. The largest and longest trial, HOPE4MCI (164 participants, 78 weeks, low-dose extended-release levetiracetam), was negative on its primary endpoint (PMID:38356475), and the LEV-AD crossover trial was null overall with benefit confined to its prespecified epileptiform-positive subgroup (PMID:34570177). The honest reading is that any benefit is subgroup-restricted and that this has not yet been tested prospectively in a biomarker-selected population. A competing account also has to be carried: neuronal hyperactivity appears before plaque formation and is inducible by soluble amyloid-beta directly (PMID:22592800), so interneuron failure may be one component of an amyloid-initiated network phenotype rather than its first step. Curated separately from synaptic_failure_convergence_model because it still makes the opposite claim about what fails first: inhibition, not excitation.

Pathograph links

Deep research

OpenScientist citations 23 citations 2026-08-28T07:27:26.242199

Show evidence (9 references)

PMID:22541439 SUPPORT Model Organism

"Restoring Nav1.1 levels in hAPP mice by Nav1.1-BAC expression increased inhibitory synaptic activity and gamma oscillations and reduced hypersynchrony, memory deficits, and premature mortality."

Gain-of-function rescue of a single interneuron-specific channel subunit corrects oscillations, hypersynchrony and memory, establishing interneuron failure as causal rather than correlative in this model.

PMID:27696483 SUPPORT Human Clinical

"Subclinical epileptiform activity was detected in 42.4% of AD patients and 10.5% of controls (p = 0.02)."

Prospective, blinded extended EEG/MEG monitoring showing the predicted hyperexcitability is present in a large minority of patients with no seizure history.

PMID:27696483 SUPPORT Human Clinical

"However, patients with subclinical epileptiform activity showed faster declines in global cognition, determined by the Mini-Mental State Examination (3.9 points/year in patients with epileptiform activity vs 1.6 points/year in patients without; p = 0.006), and in executive function (p = 0.01)."

Links the electrophysiological finding to the clinical outcome the model predicts it should affect.

+ 6 more references

Autophagy-Senescence-Inflammasome (ASI) Axis

asi_axis_reciprocal_cycle EMERGING Early-Onset Alzheimer's Disease Late-Onset Alzheimer's Disease

Evidence balance 3 support

Impaired mitophagy, senescent glia, and inflammasome activation are modeled as reciprocally reinforcing rather than independent: defective mitochondrial quality control licenses inflammasome activation via mitochondrial danger signaling, and inflammasome/IL-1 signaling in turn reinforces the senescence-associated secretory phenotype, closing a self-sustaining cycle of neuroinflammation rather than a one-way cascade.

EMERGING. Connects three mechanisms this entry already models separately — Mitochondrial Quality-Control Failure, Senescent Cell Accumulation, and the NLRP3 inflammasome-activation pathway on the Neuroinflammation node — into the reciprocal loop PMID:42419491 proposes. That review is cited only for the top-level claim (see evidence above); the two arms that close the loop (autophagy/mitochondria to inflammasome, and inflammasome back to senescence) are each supported by a real mechanistic study, but neither is an Alzheimer-specific or CNS-specific result, which is why this is EMERGING rather than ALTERNATIVE. The pre-existing senescence-to-neuroinflammation edge (SASP driving glial activation) is Alzheimer-model evidence already curated on the Senescent Cell Accumulation node.

Pathograph links

Neuroinflammation -> Senescent Cell Accumulation INDIRECT UNKNOWN INTERMEDIATES Inflammasome/IL-1 signaling is modeled as reinforcing rather than merely resulting from glial senescence, closing a reciprocal loop with the existing senescence-to-neuroinflammation edge rather than a one-way cascade.

Show evidence (3 references)

PMID:42419491 SUPPORT REVIEW SYNTHESIS

"we synthesise emerging evidence supporting an integrated 'Autophagy-Senescence-Inflammasome (ASI) axis', in which reciprocal interactions among impaired autophagy, senescent glia, and inflammasome signalling establish a self-sustaining cycle of neuroinflammation"

States the paper's own proposed reciprocal triad structure. Cited only as the source of this hypothesis grouping, not as evidence for any individual causal edge below — this is a narrative review, and the edges each carry their own primary-study evidence.

PMID:21151103 SUPPORT INDIRECT In Vitro

"Depletion of autophagic proteins promoted the accumulation of dysfunctional mitochondria and cytosolic translocation of mitochondrial DNA (mtDNA) in response to lipopolysaccharide (LPS) and ATP in macrophages."

Primary mechanistic support for the autophagy/mitochondria-to-inflammasome arm of the cycle, in macrophages rather than microglia specifically, so curated as indirect for this brain-specific entry.

PMID:23770676 SUPPORT INDIRECT In Vitro

"Treatment with Caspase 1 or IL-1R inhibitors (Fig 6i, S6i) but not TGFBRI inhibitors (Fig S6j), blunted the expression of SASP components during OIS, suggesting that activation of the inflammasome was a cause rather than an effect of the SASP."

Primary mechanistic support for the inflammasome-to-senescence closing arm of the cycle, in an oncogene-induced-senescence fibroblast/epithelial model rather than CNS glia, so curated as indirect here.

N-Glycan Hyperglycosylation Metabolic Driver Model

n_glycan_hyperglycosylation_model EMERGING

Evidence balance 2 support 1 refute

Excess protein N-glycosylation in the Alzheimer brain is modeled as an active contributor to cognitive decline rather than a passive by-product of neurodegeneration. Hexosamine-pathway flux into N-glycan biosynthesis increases, raising the net abundance of N-linked glycans (high-mannose and bisecting species prominent) in cortex and hippocampus; lowering that flux genetically (PGM3 knockdown) or pharmacologically (oligosaccharyltransferase inhibition) improves memory in amyloid and tau mouse models, and feeding the pathway with oral glucosamine worsens it. The model predicts that glucosamine supplementation is harmful specifically in people who already have the disease, not in the healthy brain.

EMERGING, and it should stay there until replicated outside the originating laboratory. The case rests on one paper (PMID:42265388). Its human spatial-glycomics discovery cohort is three AD and three control frontal cortices, extended by a Braak-stratified FFPE series; the mechanistic and interventional work uses female 5xFAD and PS19 mice at 8-9 months with four to seven animals per arm, two-week interventions, and a single behavioural readout (social recognition memory). The PGM3 knockdown and NGI-1 effects on memory did not come with changes in amyloid-beta or GFAP, so the model places the effect off the amyloid and astrocytic axes rather than upstream of them. Two tensions must be carried. First, direction: site-specific glycoproteomics finds loss of N-glycosylation on synaptic proteins in AD brain (PMID:41042700), and the originating authors reconcile this by noting that they measure net released N-glycan abundance rather than peptide-level occupancy — a plausible reconciliation that no study has yet tested on the same tissue. Second, the originating authors themselves raise the possibility that increased glycosylation under cerebral glucose hypometabolism is compensatory, which is the opposite causal reading of the same observation; their perturbation data argue against it in mice but not in humans. The glucosamine human arm is discussed on the environmental entry, where large population cohorts point the other way for incident dementia.

Pathograph links

Brain N-Glycan Hyperglycosylation -> Memory Loss INDIRECT UNKNOWN INTERMEDIATES Reducing N-glycan biosynthesis restores social recognition memory in 5xFAD and PS19 mice, and increasing it with oral glucosamine worsens memory in 5xFAD mice, without changes in amyloid-beta or astrocytic GFAP. The intermediate steps between excess glycosylation and impaired memory are not identified; the glycoproteomic data point to neuronal membrane proteins but no specific glycoprotein has been shown to carry the effect.

Show evidence (3 references)

PMID:42265388 SUPPORT Model Organism

"Genetic knockdown of glycan biosynthetic enzymes improves cognitive outcomes in AD mice whereas oral glucosamine supplementation impairs them."

Bidirectional perturbation in the same model: lowering glycan biosynthesis helps and raising it hurts, which is the causal core of the model.

PMID:42265388 SUPPORT Human Clinical

"A retrospective analysis of electronic health records from patients with AD with varying disease severity shows that glucosamine supplementation is associated with accelerated AD progression and worsened survival."

The human arm the model is built to explain. Observational, single health system, exposure ascertained from clinical notes; see notes.

PMID:41042700 REFUTE INDIRECT Human Clinical

"Our findings suggest that the loss of N-glycosylation may contribute to the pathogenesis of AD through impairing synaptic transmission and promoting protein aggregation."

An independent site-specific glycoproteomic study of post-mortem AD brain (nine cases, nine controls) reaches the opposite directional conclusion: reduced glycosylation at synaptic and aggregation-prone sites. INDIRECT because it measures per-site occupancy on enriched glycopeptides, not the net abundance of released N-glycans the hyperglycosylation model rests on.

Reactive Astrocyte Subtype Resilience Model

reactive_astrocyte_resilience_model EMERGING Late-Onset Alzheimer's Disease

Evidence balance 2 support

Cognitive resilience, the preservation of cognition despite Alzheimer neuropathology at a given Braak stage, is modeled as depending in part on which reactive state cortical astrocytes adopt. Reactive astrocyte subtypes expressing neuroprotective programs (for example EMP1 in the PLSCR1-marked subtype) are proposed to buffer neurons against pathology, while subtypes with a more severe, neurotoxic-skewed reactive profile (high FKBP5, MAOB and CHI3L1) are proposed to accompany the transition to dementia. The claim is that astrocyte reactive-state composition, not only amyloid and tau burden, separates resilient from cognitively impaired donors.

EMERGING because the only direct support is one cohort's transcriptomic association, reported with the direction of causation left open. It is separate from neuroimmune_glial_amplification_model, which treats glial reactivity as amplifying injury; this model claims a reactive state can be protective and that the protective and harmful states are distinguishable subtypes. A re-analysis of the source paper's supplementary tables (see the OpenScientist assessment under kb/hypotheses/) found the aggregate neuroprotective-versus-neurotoxic marker contrast null, astrocytes ranked 16th of 27 subclasses for separating resilient from impaired donors, and astrocyte scores rising with every neuronal subclass within donors; the EMP1 and FKBP5 directions come from AD-versus-control tables, not from a resilience contrast.

Deep research

OpenScientist citations 1 citations 2026-09-26T02:38:30.241620

Show evidence (2 references)

PMID:42778763 SUPPORT PRIMARY RESULT Computational

"reactive astrocyte subtypes with altered neuroprotective and neurotoxic gene expression that likely confer cognitive resilience"

The seed observation for this model, from computational phenotype scoring of prefrontal single-nucleus data in the PsychAD cohort. The authors' "likely confer" is an inference from association; no perturbation tests whether the astrocyte state causes resilience.

PMID:37248300 SUPPORT Human Clinical

"Our findings suggest astrocyte reactivity as an important upstream event linking Aβ with initial tau pathology, which may have implications for the biological definition of preclinical AD and for selecting CU individuals for clinical trials."

Independent human biomarker evidence that astrocyte reactivity modulates how amyloid translates into downstream pathology, the plausibility premise for astrocyte state shaping clinical expression.

Glial Inflammatory and ER-Stress Model of Depression in Alzheimer Disease

glial_depression_in_ad_model EMERGING Late-Onset Alzheimer's Disease

Evidence balance 1 support

Depressive symptoms arising in people with Alzheimer disease are modeled as having a glial cellular substrate distinct from the general Alzheimer inflammatory signal: astrocyte states carrying inflammatory and endoplasmic reticulum stress programs are proposed to be enriched in Alzheimer donors with depression compared with Alzheimer donors without it, and to overlap with glial changes seen in major depressive disorder.

EMERGING on a single cohort. Whether depression within Alzheimer disease shares mechanism with major depressive disorder is stated by the source as unresolved. Linked to the Behavioral Changes phenotype and the Neuroinflammation node, but no causal edge is drawn because the direction between glial state and depressive symptoms is not established. A re-analysis of the source paper's supplementary tables (see the OpenScientist assessment under kb/hypotheses/) found the ER-stress limb not robust: unfolded-protein-response genes were enriched in the AD-Depression astrocyte cluster only against an assumed whole-genome background, not against the genes actually tested, none was among the depression-associated astrocyte DEGs, and HSPA5 and CALR appear among the general AD astrocyte DEGs.

Deep research

OpenScientist citations 4 citations 2026-09-26T03:03:53.809776

Show evidence (1 reference)

PMID:42778763 SUPPORT PRIMARY RESULT Computational

"We also identified many PACs for multiple phenotypes, including the astrocytes between AD and depression showing specific gene expression patterns such as inflammation and endoplasmic reticulum stress pathways."

The seed observation. The comparison was within Alzheimer donors (with versus without depression), so the shared Alzheimer signal is removed by design. It is an association from computational scoring and does not show that the glial state produces the mood symptom.

smORF-Encoded Microprotein Model

smorf_microprotein_model EMERGING Late-Onset Alzheimer's Disease

Evidence balance 2 support

Thousands of small open reading frames (smORFs) that are absent from the reviewed reference proteome are nonetheless translated in the human frontal cortex, and the microproteins they encode carry their own Alzheimer-associated expression changes. The model holds that some of these microproteins take part in the disease rather than being incidental translation, and - the part that makes it a distinct model rather than a restatement - that a microprotein's regulation can be decoupled from the annotated main ORF at the same locus. Conventional proteomic and transcriptomic workflows quantify the annotated product, so a microprotein that falls in Alzheimer disease while its host locus does not is invisible to them. The model therefore predicts a class of disease-associated protein changes that existing Alzheimer omics could not have detected, and it is testable in exactly that form: rebuild the search space to include smORFs and the changes should appear.

Curated as EMERGING, not ALTERNATIVE. The model competes with nothing - it does not displace the amyloid, tau, or glial models, and it makes no claim about their ordering. What it asserts is that an entire layer of the protein-coding output has gone unmeasured in Alzheimer proteomics, and that at least one member of that layer has a demonstrable cellular function. The weak point is the step from association to mechanism: the atlas establishes that microproteins change in Alzheimer disease, and a single case (micro-MKKS63) carries a functional experiment behind it. The other roughly thousand do not yet.

Pathograph links

Micro-MKKS63 Microprotein Depletion -> Microglial Oxidative Phosphorylation Failure DIRECT Loss of the microprotein lowers microglial oxidative phosphorylation. Curated as DIRECT because the knockout removes the microprotein itself and the respiratory deficit follows, with no intermediate proposed - but the demonstration is in an immortalized human microglial line, not in cortex, so the edge is a cell-biological result imported into a human pathograph rather than a human observation.

Show evidence (2 references)

PMID:42736449 SUPPORT DIRECT PRIMARY RESULT Human Clinical

"Of 3,001 unreviewed MPs amenable to PROSIT interrogation, 1,067 (35.6%) were strong, 1,433 (47.8%) moderate, 343 (11.4%) weak and 158 (5.3%) insufficient."

The premise the model rests on, at scale and with a stated confidence gradient: over a thousand microproteins absent from the reviewed reference proteome carry strong-grade spectral support in postmortem human frontal cortex, graded against deep-learning-predicted spectra rather than asserted.

PMID:42736449 SUPPORT DIRECT PRIMARY RESULT Human Clinical

"we found that 39% of testable pairs showed disease-associated changes independent of the mORF"

The decoupling claim, which is what separates this model from known locus-level changes: for a large minority of smORF-mORF pairs the Alzheimer-associated change is not recoverable by measuring the annotated product at the same locus.

Discussions and Knowledge Gaps

Which temporal causal ordering among amyloid-beta accumulation, tau spread, microglial/astrocytic activation, mitochondrial quality-control failure, and synaptic/cognitive decline best explains stage-specific Alzheimer disease progression and cognitive resilience?

KNOWLEDGE GAP OPEN gap_ad_amyloid_tau_glia_mitochondria_ordering_resilience

Attached to

pathophysiology#Amyloid Plaque Formation pathophysiology#Neurofibrillary Tangle Formation pathophysiology#Neuroinflammation pathophysiology#Mitochondrial Quality-Control Failure pathophysiology#Synaptic Dysfunction

Amyloid plaques, tau tangles, glial activation, mitochondrial stress, and cognitive decline are all established Alzheimer disease features, but the disease entry should not imply a single fixed ordering across all stages. Current evidence supports amyloid-triggered tau and glial pathways, reciprocal inflammasome-to-amyloid amplification, protective TREM2-dependent microglial containment of amyloid-associated tau spread, mitochondrial quality-control effects on amyloid/tau/inflammation, and cognitive reserve that can decouple pathology burden from symptoms. The open curation question is which edges are upstream drivers, downstream amplifiers, parallel state changes, or resilience modifiers in preclinical, prodromal, and dementia-stage Alzheimer disease.

Proposed experiments

Longitudinal amyloid-tau-glia-mitophagy-resilience ordering cohort

longitudinal multimodal cohort study

exp_ad_longitudinal_pathology_glia_mitophagy_resilience

Follow amyloid-negative, amyloid-positive tau-negative, and amyloid-positive tau-positive older adults across preclinical, prodromal, and dementia stages with amyloid PET, tau PET, plasma and CSF glial biomarkers, candidate mitochondrial/mitophagy biomarkers, structural and functional imaging, and cognitive reserve measures. The study should test whether glial and mitochondrial changes precede tau spread, follow amyloid/tau deposition, amplify clinical decline, or mark resilient compensation.

Assays

amyloid positron emission tomography tau positron emission tomography cerebrospinal fluid biomarker profiling plasma biomarker profiling neuropsychological assessment

Readouts

Glial activation temporal precedence

pathophysiology#Neuroinflammation

Plasma GFAP, CSF sTREM2, inflammasome, cytokine, and complement readouts modeled relative to amyloid PET conversion, tau PET spread, and cognitive change.

plasma GFAP measurement CSF soluble TREM2 measurement complement biomarker profiling

Direction: POSITIVE

Interpretation: Glial-marker elevation before regional tau spread would support an upstream amplifier role; elevation only after amyloid/tau burden would support downstream response.

Mitochondrial quality-control state

pathophysiology#Mitochondrial Quality-Control Failure

Mitophagy and mitochondrial stress markers assessed against amyloid, tau, glial, and cognitive trajectories.

Mitophagy GO:0000422

Direction: NEGATIVE

Interpretation: Declining mitophagy markers before amyloid/tau or glial acceleration would support mitochondrial quality-control failure as an upstream or parallel driver rather than only a late downstream injury marker.

Pathology-symptom decoupling

phenotypes#Memory Loss

Cognitive trajectories stratified by education, occupation, engagement, social-network, and functional-network measures among participants with comparable amyloid, tau, glial, and mitochondrial marker burden.

neuropsychological assessment functional connectivity imaging

Direction: NEGATIVE

Interpretation: Preserved cognition despite high pathology burden would support cognitive resilience as a modifier that should be curated separately from the core amyloid-tau-glia pathophysiology chain.

Controls

Amyloid-negative age-matched controls

Older adult participants without amyloid PET positivity at baseline.

Pathology-burden matched resilience strata

Participants matched for amyloid/tau/glial burden but differing in cognitive reserve and longitudinal cognitive decline.

Decision criterion

The ordering model should be revised according to which biomarker changes temporally precede tau spread and cognitive decline after adjustment for pathology burden and resilience variables; a resilience discussion should remain separate if cognitive trajectories decouple from amyloid, tau, glial, and mitochondrial burden.

Would support

pathophysiology#Neuroinflammation pathophysiology#Mitochondrial Quality-Control Failure mechanistic_hypotheses#neuroimmune_glial_amplification_model mechanistic_hypotheses#autophagy_lysosomal_clearance_model

Show evidence (1 reference)

PMID:37924152 SUPPORT Human Clinical

"Brain amyloid was strongly associated with plasma GFAP and ptau-181 and to a lesser extent with plasma NfL."

Demonstrates feasibility and relevance of measuring astrocytic GFAP alongside amyloid, tau, and neurodegeneration markers in human cohorts.

Posed 2026-06-03T00:00:00Z

Seeded for issue 3661. This discussion deliberately separates established pathologies from the unresolved causal ordering among amyloid, tau, microglial/astrocytic inflammation, mitochondrial quality control, and cognitive resilience.

Show evidence (5 references)

PMID:36911732 SUPPORT Other

"the temporal and spatial changes in microglial phenotype, the interactions among microglia, Aβ, tau, and neurons"

Review-level support that the unresolved issue is temporal and spatial ordering among microglia, amyloid-beta, tau, and neuronal injury rather than the mere presence of those processes.

PMID:37308616 SUPPORT Other

"these proteins do not act in isolation but form part of a pathological network."

Supports representing amyloid-beta and tau as embedded in a multi-cellular pathological network that includes reactive astrocyte states.

PMID:30742114 SUPPORT Other

"Our findings suggest that impaired removal of defective mitochondria is a pivotal event in AD pathogenesis and that mitophagy represents a potential therapeutic intervention."

Supports including mitochondrial quality-control failure in the ordering gap rather than treating mitochondrial stress only as a generic oxidative consequence.

+ 2 more references

Is HSV-1 reactivation in RORB+ glutamatergic neurons CAUSAL for their selective loss in Alzheimer's disease, or a marker of cells already destined to die?

KNOWLEDGE GAP OPEN disc_hsv1_causality

Attached to

pathophysiology#HSV-1 Reactivation in RORB+ Glutamatergic Neurons

Cross-sectional post-mortem evidence establishes the association but cannot resolve causal direction. The answer determines whether antivirals (or pre-emptive HSV-1 suppression) are candidate disease-modifying therapies, or merely a downstream readout of cells already committed to neurodegeneration.

Proposed experiments

HSV-1 reactivation perturbation in human cortical neuron-glia organoids

controlled perturbation experiment

exp_hsv1_organoid_reactivation_causality

Introduce controlled HSV-1 latency/reactivation into human cortical neuron-glia organoids enriched for glutamatergic neurons, then compare RORB+ neuron survival and neuroinflammatory state against matched mock and antiviral-rescue controls.

Model systems

Human cortical neuron-glia organoid

Human pluripotent-stem-cell-derived cortical organoid with glutamatergic neurons and glial support cells, used to test whether viral reactivation precedes neuronal vulnerability in a disease-relevant human cellular context.

ORGANOID namo:Organoid link

human NCBITaxon:9606

RORB+ glutamatergic neuron CL:0000679 microglial cell CL:0000129

Perturbations

Induced HSV-1 reactivation

pathophysiology#HSV-1 Reactivation in RORB+ Glutamatergic Neurons

Establish latent HSV-1 infection and trigger controlled reactivation to test whether reactivation is sufficient to drive selective RORB+ glutamatergic-neuron loss.

viral process GO:0016032

Readouts

RORB+ glutamatergic neuron survival

pathophysiology#HSV-1 Reactivation in RORB+ Glutamatergic Neurons

Single-cell and imaging readout of whether RORB+ glutamatergic neurons are selectively depleted after reactivation.

single-cell transcriptomic profiling high-content imaging

Direction: NEGATIVE

Interpretation: Selective loss after reactivation, reduced by antiviral rescue, would support HSV-1 reactivation as a causal injury mechanism.

Neuroinflammatory activation

pathophysiology#Neuroinflammation

Cytokine, glial activation, and stress-response measurements to test whether viral reactivation creates an inflammatory state upstream of neuron loss.

inflammatory response GO:0006954

multiplex cytokine profiling single-cell transcriptomic profiling

Direction: POSITIVE

Controls

Mock-reactivated isogenic organoids

Matched organoids handled identically without HSV-1 reactivation.

Antiviral rescue arm

Reactivated organoids treated with antiviral suppression before readout.

Decision criterion

HSV-1 reactivation should temporally precede selective RORB+ neuron loss, and antiviral rescue should attenuate both viral signal and neuronal loss.

Would support

pathophysiology#HSV-1 Reactivation in RORB+ Glutamatergic Neurons

Would refute

pathophysiology#HSV-1 Reactivation in RORB+ Glutamatergic Neurons

Show evidence (1 reference)

PMID:42094473 SUPPORT Human Clinical

Human post-mortem evidence motivates a perturbational organoid test that can distinguish causal reactivation from end-stage association.

Posed 2026-05-16T00:00:00Z

Seeded from PR 2789 alongside the new HSV-1 pathophysiology node. See the broader discussion of how `discussions:` (this layer) relates to a proposed structural `knowledge_gaps:` layer in https://github.com/monarch-initiative/dismech/issues/2617#issuecomment-4467637580 The interventional VALAD (PMID:41405855) and observational anti-herpetic prevention evidence (PMID:41467972, PMID:41779765) were surfaced by mining the existing OpenScientist deep-research report for this hypothesis (kb/hypotheses/Alzheimer_Disease/hsv1_reactivation_model/openscientist.md; two fresh OpenScientist re-runs both hit the provider's 2 h job ceiling and produced no output). Each PMID and snippet was independently verified against the fetched abstract before curation, per the deep-research anti-hallucination SOP; in the process one DR-reported figure was found misattributed (the report cited OR=1.46 for the HSV-AD meta-analysis PMID:40934136, whose actual abstract reports OR=1.32 case-control / 20% cohort), so that number was not used.

Show evidence (4 references)

PMID:42094473 SUPPORT Human Clinical

Establishes the association in post-mortem human brain but the cross-sectional design cannot distinguish whether HSV-1 reactivation drives the loss of RORB+ neurons or merely marks neurons destined to die.

PMID:41405855 REFUTE Human Clinical

"greater cognitive worsening with valacyclovir than placebo"

The phase II VALAD randomized controlled trial (JAMA 2026) of valacyclovir 4 g/day vs placebo over 78 weeks in HSV-seropositive early symptomatic AD found no cognitive benefit and, if anything, greater decline in the treatment arm (between-group ADAS-Cog difference 3.93, 95% CI 1.03-6.83, P=.01). Direct interventional evidence that anti-HSV therapy does not help ESTABLISHED symptomatic AD — refuting the treatment corollary while leaving the upstream reactivation hypothesis and the pre-symptomatic-prevention question open (the same treatment-vs-prevention split seen in the zoster-vaccine natural experiments).

PMID:41467972 SUPPORT Human Clinical

"diagnosed and treated versus diagnosed but untreated (aHR=0.77, 95%"

Meta-analysis of 14 cohort studies in non-demented adults >=50 covering HSV and VZV finds anti-herpetic treatment associated with lower incident dementia (aHR 0.77 for diagnosed-and-treated vs diagnosed-but-untreated). Observational prevention-side signal complementing the null VALAD treatment trial and bridging to the zoster-vaccine evidence (VZV included).

+ 1 more reference

Do the clonally expanded, cytotoxic CD8+ T cells that drive tau-associated neurodegeneration recognize a tau-derived (or other brain) antigen, or are they antigen-independent (e.g., age-expanded virtual-memory CD8 T cells)?

KNOWLEDGE GAP OPEN gap_ad_tcell_antigen_identity

Attached to

pathophysiology#Adaptive Immune T Cell Response to Tau Pathology

The adaptive_immune_tcell_model reads "restricted TCR clonality" as evidence of an antigen-driven CD8 response, and cDC1 cross-presentation is now established as the priming route (PMID:41890046) — but the recognized antigen(s) remain unidentified. Two findings keep the antigen-driven interpretation open: aging independently expands antigen-independent virtual-memory CD8 T cells that are clonal and cytotoxic without foreign-antigen priming (PMID:42432776); and the clonally expanded CD8 TCRs recovered from human AD CSF were shown to be specific to Epstein-Barr virus antigens rather than tau (PMID:31915375). Distinguishing a tau-antigen-driven response from an antigen-independent (or virally-primed) bystander clonal expansion is the single most important unresolved edge in this model and gates whether tau-antigen-directed tolerization is even a coherent therapeutic strategy.

Proposed experiments

Brain-infiltrating CD8 clone antigen-discovery screen

antigen-specificity T cell reactivity screen

exp_ad_tcell_antigen_discovery

Isolate brain- and CSF-infiltrating CD8+ T cell clones from tauopathy mice and human AD/primary-tauopathy tissue, pair single-cell TCR sequencing with peptide-MHC tetramer and pMHC-screen reactivity assays against tau-derived peptides, candidate self-antigens, and viral (EBV) antigens, and fate-map virtual-memory versus conventional-memory origin.

Assays

single-cell T cell receptor sequencing peptide-MHC tetramer binding assay

Readouts

Dominant-clone antigen specificity

pathophysiology#Adaptive Immune T Cell Response to Tau Pathology

Fraction of dominant brain-infiltrating CD8 clones with demonstrable reactivity to tau-derived peptides versus viral/self antigens versus no detectable cognate antigen (virtual-memory phenotype).

Direction: POSITIVE

Interpretation: Dominant clones reactive to tau peptides would support a tau-antigen-driven response; predominance of viral-reactive or antigen-inexperienced virtual-memory clones would reframe the arm as cytotoxic-lymphocyte- rather than tau-antigen-driven.

Posed 2026-07-11T00:00:00Z

Seeded from an OpenScientist hypothesis deep-research run on adaptive_immune_tcell_model; all cited PMIDs independently verified against their abstracts before curation.

Is IFN-gamma/PD-1 signaling net-neurotoxic or net-protective in Alzheimer disease, and how much of the IFN-gamma effect attributed to clonal CD8 T cells is actually contributed by NK cells?

KNOWLEDGE GAP OPEN gap_ad_tcell_ifng_pd1_valence_and_source

Attached to

pathophysiology#Adaptive Immune T Cell Response to Tau Pathology pathophysiology#Neuroinflammation

The model predicts IFN-gamma/PD-1 blockade is neuroprotective, based on mouse tauopathy where inhibiting IFN-gamma or PD-1 ameliorates brain atrophy (PMID:36890231). But the therapeutic valence reverses by pathology: in an amyloid context, PD-1 checkpoint blockade drives an IFN-gamma-dependent myeloid response that clears Aβ and improves cognition (PMID:26779813), so a universal "block IFN-gamma/PD-1" prescription is not supported. Separately, IFN-gamma is not CD8-exclusive — NK cells produce IFN-gamma/TNF-alpha that negatively correlates with cognition in AD (PMID:11268360) — so some effects attributed to clonal CD8 may be innate-lymphoid in origin. Resolving the context-dependent valence and the true cellular source of pathogenic IFN-gamma is required before the node's therapeutic predictions can be curated as directional.

Proposed experiments

Cell-type-resolved IFN-gamma modulation across tau vs amyloid models

comparative conditional-knockout intervention study

exp_ad_ifng_source_valence_matched_models

Apply matched IFN-gamma/PD-1 modulation and cell-type-specific conditional Ifng deletion (CD8-restricted vs NK-restricted vs total) across tau-only, amyloid-only, and mixed mouse models, reading neurodegeneration and cognition to separate the direction of the IFN-gamma effect from its cellular source.

Assays

conditional gene knockout neuropsychological assessment

Readouts

IFN-gamma source-and-valence effect on neurodegeneration

pathophysiology#Adaptive Immune T Cell Response to Tau Pathology

Change in atrophy/neuronal loss when IFN-gamma is deleted from CD8 versus NK cells, and when IFN-gamma/PD-1 is blocked, in tau-only versus amyloid-only contexts.

Direction: NEGATIVE

Interpretation: Protection only with CD8-restricted IFN-gamma loss in the tau context (with harm or no benefit in amyloid) would confirm a tau-scoped, CD8-driven pathogenic axis; protection with NK-restricted loss would reassign part of the effect to innate lymphocytes.

Posed 2026-07-11T00:00:00Z

Seeded from an OpenScientist hypothesis deep-research run on adaptive_immune_tcell_model; all cited PMIDs independently verified against their abstracts before curation.

Does the cytotoxic CD8+ T-cell arm that is functionally required for tau-mediated neurodegeneration in mouse tauopathy actually contribute to human Alzheimer disease, given that adaptive-immune/CD8/IFN-gamma/MHC-I genes carry no detectable inherited AD risk while innate-microglial genes dominate the human genetic-association signal?

HUMAN MODEL MISMATCH OPEN mismatch_ad_tcell_arm_not_inherited_risk

Attached to

pathophysiology#Adaptive Immune T Cell Response to Tau Pathology

In mouse tauopathy the adaptive CD8+ T-cell arm is functionally required — T-cell depletion, IFN-gamma/PD-1 inhibition, and cDC1 ablation are each neuroprotective (PMID:36890231, PMID:41890046) — yet human inherited AD risk shows no adaptive-immune signal. A gene-set enrichment of Alzheimer disease genetic-association targets from the Open Targets Platform (disease MONDO:0004975; the top 200 of 13,367 targets ranked by the genetic_association datatype score; retrieved and independently reproduced on 2026-07-11) finds a curated 35-gene innate-microglial set strongly over-represented (18/35 overlap; odds ratio 115.1; Fisher one-sided p = 1.79e-27; the overlap includes TREM2, PLCG2, INPP5D, ABCA7, BIN1, SORL1, CR1, CD33, APOE, CLU, MS4A6A, PICALM), while a 46-gene adaptive-immune/CD8/IFN-gamma/MHC-I set is not enriched at all (0/46 overlap; odds ratio 0.0; p = 1.0). The null adaptive signal is robust across top-N cutoffs (100-1000) and gene-background choices (19,000/20,000), and adaptive-immune genes are absent even from the top 2,000 ranked targets. This coheres with the published human-genetics consensus that microglia and innate immunity are the central inherited determinants of AD susceptibility (PMID:41888907) and that AD risk loci act specifically through microglial eQTLs (PMID:40670704). The mismatch is mechanistically meaningful: it implies the pathogenic adaptive T-cell arm — potent in mouse tauopathy — is most plausibly an acquired/downstream response to tau pathology rather than an upstream heritable driver of AD, reconciling the mouse functional data with the absence of adaptive-immune inherited risk and cohering with the model's own caveats (EBV-specific CSF TCRs, age-expanded antigen-independent virtual-memory CD8). Caveats: this is a coarse gene-set enrichment (not colocalization or fine-mapping); most AD GWAS are amyloid-weighted diagnosis studies whereas the model is tau-scoped; and the LD-complex MHC region may be under-represented at target-level association scores.

Proposed experiments

Cell-type-resolved colocalization of AD GWAS loci with CD8 T-cell vs microglial QTLs

GWAS-QTL colocalization analysis

exp_ad_celltype_coloc_adaptive_vs_microglial

Move beyond coarse enrichment to statistically colocalize Alzheimer disease GWAS loci with CD8+ T-cell versus microglial eQTL/caQTL maps, directly testing whether any adaptive-immune regulatory variants influence inherited AD risk rather than acting only downstream of established tau pathology.

Assays

expression quantitative trait locus colocalization chromatin accessibility quantitative trait locus mapping

Readouts

Adaptive-immune vs microglial colocalization signal

pathophysiology#Adaptive Immune T Cell Response to Tau Pathology

Number and posterior probability of AD loci colocalizing with CD8+ T-cell QTLs versus microglial QTLs.

Direction: POSITIVE

Interpretation: Robust colocalization of AD loci with CD8 T-cell QTLs would revise the downstream-only reading and support an inherited adaptive-immune contribution; continued microglia-exclusive colocalization would confirm the acquired/downstream interpretation of the T-cell arm.

Tau-endophenotype-stratified genetic enrichment

gene-set enrichment analysis

exp_ad_tau_stratified_genetic_enrichment

Repeat the adaptive-immune-vs-microglial enrichment against genetic-association scores derived from tau-PET or CSF-tau endophenotype GWAS rather than amyloid-weighted clinical-diagnosis GWAS, the fairer test of the tau-scoped model.

Assays

genome-wide association study

Readouts

Adaptive-immune enrichment under tau-endophenotype GWAS

pathophysiology#Adaptive Immune T Cell Response to Tau Pathology

Odds ratio and significance of adaptive-immune gene-set enrichment when targets are ranked by tau-endophenotype genetic association.

Direction: POSITIVE

Interpretation: Emergence of adaptive-immune enrichment specifically under tau-endophenotype GWAS (absent under amyloid-weighted diagnosis GWAS) would indicate the coarse diagnosis-GWAS null understated a tau-specific inherited contribution.

Posed 2026-07-11T00:00:00Z

Seeded from a scoped OpenScientist code-execution deep-research run on adaptive_immune_tcell_model that queried the Open Targets Platform GraphQL API and computed the enrichment. The reported numbers (13,367 total targets; top-200 genetic_association score range 0.461-0.954; first-15 targets PSEN1, APP, CR1, PSEN2, APOE, TREM2, PLCG2, MS4A6A, HFE, EPHA1, APH1B, ACE, H2BC4, BIN1, MPO; innate OR 115.1 p 1.79e-27; adaptive OR 0.0 p 1.0) were independently reproduced to the digit by re-running the same GraphQL query and Fisher exact test before curation. Query: disease(efoId ->MONDO_0004975) associatedTargets orderByScore:"genetic_association", top 200 with score > 0; background 20,000 protein-coding genes; one-sided Fisher exact per gene set.

Do preclinical reports that low-dose ionizing radiation below 100 mSv, at selected doses and regimens, reduce Alzheimer-relevant amyloid burden and neuroinflammation or improve cognition in humans without delayed harm, and where does the response change from potentially protective to neurotoxic?

HUMAN MODEL MISMATCH OPEN mismatch_ad_ldir_dose_regimen_translation

Attached to

pathophysiology#Amyloid Plaque Formation pathophysiology#Neuroinflammation pathophysiology#Synaptic Dysfunction phenotypes#Memory Loss

A 2026 review describes potentially protective effects only at specific low-dose regimens and explicitly reports no consensus on the dual mechanisms, safety threshold, or long-term outcomes because of small samples, ethical constraints, and inter-model variability. The abstract synthesizes preclinical evidence and prospective applications rather than establishing a reproducible human Alzheimer disease response. This entry therefore does not assert a protective environmental edge or add low-dose-radiation mechanism nodes. A single descriptor modifier would also be misleading: modifiers encode one biological state, not a dose- and regimen-conditioned reversal of effect. If the response is later established, detrimental and protective arms should be represented separately with explicit exposure context and arm-specific evidence. The question is kept at entry level because the reported reduction in amyloid-beta deposition maps directly to the existing Alzheimer pathograph; a reusable hormesis module would require a conserved causal chain, a key conformance target, and validated recurrence across at least two disorders, none of which this single review establishes.

Proposed experiments

Harmonized low-dose ionizing-radiation dose-response study

harmonized cross-model dose-response study

exp_ad_ldir_dose_regimen_translation

Apply a harmonized absorbed-dose, dose-rate, radiation-quality, and fractionation matrix across human iPSC-derived hippocampal neuron-astrocyte-microglia cultures and an Alzheimer-relevant in vivo model, then compare the prespecified biomarker directions with longitudinal data from consented cohorts receiving clinically indicated imaging or occupational monitoring. Report absorbed dose in mGy for experimental systems and effective dose in mSv for human cohorts rather than treating the units as interchangeable. No participant should receive radiation solely for this study. Follow-up must be long enough to detect delayed neurocognitive or genotoxic harm rather than stopping at an acute adaptive response.

Perturbations

Low-dose ionizing-radiation dose and regimen matrix

pathophysiology#Amyloid Plaque Formation

Compare sham exposure with prespecified single and fractionated doses spanning an absorbed-dose range chosen to correspond to the review's below-100-mSv human exposure definition while holding radiation quality and dosimetry constant; include a higher-dose positive-control arm only in non-participant experimental systems, never in observational cohorts.

exposure to ionizing radiation ECTO:7000047

Readouts

Amyloid-beta deposition response

pathophysiology#Amyloid Plaque Formation

Quantify soluble and deposited amyloid-beta across dose, dose-rate, and fractionation conditions and in longitudinal human biomarkers.

amyloid-beta quantification

Direction: NEGATIVE

Interpretation: A replicated decrease confined to a defined low-dose window would support the proposed protective amyloid arm; a monotonic increase or failure to replicate would not.

Microglial inflammatory-state response

pathophysiology#Neuroinflammation

Measure microglial inflammatory-state markers, cytokines, and phagocytic function across the harmonized exposure matrix.

single-cell transcriptomic profiling cytokine profiling

Direction: NEGATIVE

Interpretation: Reduced inflammatory signaling together with preserved phagocytic function would support an anti-inflammatory arm; a transient marker shift without functional benefit would not.

Synaptic and cognitive outcome

pathophysiology#Synaptic Dysfunction

Measure synaptic plasticity and network activity in experimental systems and longitudinal cognitive change in the observational human cohorts.

electrophysiological recording neuropsychological assessment

Direction: NEGATIVE

Interpretation: Molecular changes count as neuroprotective only if they accompany improved or preserved synaptic function and a concordant human cognitive direction.

Delayed oxidative and genotoxic injury

pathophysiology#Oxidative Stress

Measure persistent oxidative stress, DNA damage, and delayed neuronal or cognitive injury after the acute response window.

DNA damage assay oxidative stress biomarker profiling

Direction: POSITIVE

Interpretation: Delayed injury at a nominally protective dose would refute a net-benefit interpretation even if acute amyloid or inflammatory markers improve.

Controls

Sham-exposed matched controls

Isogenic cultures and matched animals handled identically without ionizing-radiation exposure, plus unexposed comparison participants for the observational cohorts.

Dosimetry and regimen controls

Match radiation quality, absorbed dose, dose rate, fractionation, age, and disease stage as closely as each system permits, and publish the conversion assumptions, so a unit or regimen difference is not misread as a species effect.

Decision criterion

Promote an Alzheimer-specific protective arm only if a preregistered non-monotonic dose-response is independently replicated, improves amyloid, inflammatory, and synaptic outcomes without delayed harm, and shows a directionally concordant human signal. Define separate harmful and protective exposure-mechanism links only after their dose and regimen boundaries are supported; do not infer a clinical safety threshold from model-system data alone.

Posed 2026-08-08T00:00:00Z

Scope resolution for GitHub issue 8135: entry-level structured uncertainty was chosen over asserted mechanism nodes or a cross-disease hormesis module.

Show evidence (2 references)

PMID:42551708 SUPPORT Other

"Nevertheless, emerging evidence indicates that LDIR administered at specific doses and regimens may exert neuroprotective effects via hormesis: it reduces Aβ deposition, promotes anti-inflammatory microglial polarization, and enhances hippocampal neurogenesis, thereby ameliorating cognitive impairment."

The review motivates the protective arm of the question but limits it to selected doses and regimens; it does not establish a human Alzheimer disease mechanism or a safe intervention threshold.

PMID:42551708 SUPPORT Other

"Constrained by ethical considerations, sample size limitations, and inter-model variability, no consensus has been reached regarding its dual mechanisms, safety thresholds, and long-term outcomes."

Directly supports retaining this as an open model-to-human mismatch rather than asserting either a protective pathograph or a general hormesis module.

Does chronic exposure to the non-selective anticholinergic drugs used by older adults reproduce, in the human brain, the acceleration of amyloid and tau pathology that M1-selective antagonism and M1 ablation produce in transgenic mouse models?

HUMAN MODEL MISMATCH OPEN mismatch_m1_antagonism_mouse_vs_human_anticholinergic_exposure

Attached to

pathophysiology#Muscarinic M1 Receptor Signaling Loss

The mechanistic case that anticholinergic burden is disease-modifying rather than merely symptom-masking rests on mouse work that differs from human exposure in three ways. (1) Receptor selectivity: dicyclomine and the M1(-/-) allele isolate M1, whereas diphenhydramine, tricyclics and oxybutynin block M1-M5 and, in the antihistamine case, H1 as well — so the human exposure engages targets whose net effect on APP processing is not established, and the mouse result cannot simply be scaled up. (2) Baseline pathology: the mouse experiments manipulate M1 against a pre-existing transgenic amyloid/tau burden, so they demonstrate acceleration of ongoing pathology, not initiation in a healthy brain, which is what the population-level risk claim requires. (3) CNS exposure: the human epidemiology assigns risk by dispensing records, not by brain receptor occupancy, and the one large antihistamine cohort finds a dose-response for second-generation agents that do not appreciably cross the blood-brain barrier, alongside a steeper one for first-generation agents that do (1.13/1.29/1.51 versus 1.11/1.19/1.26 across rising cumulative dose). The between-class ordering is what a CNS-penetration-dependent mechanism predicts; what it does not predict is that the second-generation arm is non-null at all, and distinguishing residual confounding by indication from a genuine peripheral or non-muscarinic contribution needs measured central occupancy rather than dispensing proxies. Resolving this determines whether reducing anticholinergic burden is a prevention strategy or only good symptomatic geriatric practice.

Proposed experiments

Muscarinic occupancy and AD biomarker trajectory under chronic anticholinergic exposure

prospective biomarker cohort study with receptor occupancy imaging

exp_anticholinergic_occupancy_biomarker_cohort

Follow cognitively normal older adults initiating chronic CNS-penetrant anticholinergics, matched initiators of peripherally restricted or second-generation comparators, and non-users, with serial plasma and CSF amyloid and phospho-tau measures and amyloid/tau PET. Quantify actual central muscarinic receptor occupancy in a subset by PET rather than inferring CNS exposure from dispensing data, and test whether biomarker trajectory tracks occupancy rather than prescription class.

Assays

muscarinic receptor positron emission tomography amyloid positron emission tomography tau positron emission tomography cerebrospinal fluid biomarker profiling plasma biomarker profiling

Readouts

Amyloid trajectory versus central muscarinic occupancy

pathophysiology#Muscarinic M1 Receptor Signaling Loss

Rate of change in amyloid PET signal and plasma amyloid-beta 42/40 modeled against measured central muscarinic occupancy, with prescription class as a competing predictor.

amyloid positron emission tomography plasma amyloid-beta 42/40 measurement

Direction: POSITIVE

Interpretation: Amyloid accumulation scaling with measured central occupancy would support the muscarinic mechanism in humans; accumulation tracking prescription class irrespective of occupancy would instead implicate confounding by indication.

Is the observed association between cumulative anticholinergic exposure and incident dementia causal, or is it produced by reverse causation and confounding by indication — and why is the signal concentrated in drug classes other than the antihistamines?

CONTROVERSY OPEN controversy_anticholinergic_dementia_reverse_causation_class_specificity

Attached to

pathophysiology#Muscarinic M1 Receptor Signaling Loss

The evidence is genuinely conflicting rather than merely thin. Supporting a causal reading: a cumulative dose-response in a prospective cohort with the most recent 12 months of exposure excluded (PMID:25621434); persistence of the association in windows 5-20 years before diagnosis (PMID:31233095); a dose-response that survives an active comparator, with bladder anticholinergics carrying higher dementia risk than mirabegron in patients who all have overactive bladder (PMID:41066055); and interaction with APOE4 and CSF AD pathology rather than an APOE-independent effect (PMID:32878992). Against it: prodromal dementia causes insomnia, depression and urinary urgency, which are these drugs' indications, and the Richardson analysis explicitly names this as an alternative reading of its own result (PMID:29695481); the class distribution is wrong for a shared muscarinic mechanism, since gastrointestinal anticholinergics were null while antidepressant, urological and antiparkinson drugs were not; the largest antihistamine-specific cohort finds a non-null dose-response even for second-generation agents that lack meaningful CNS penetration (PMID:38935035, though its gradient is steeper for the first-generation class, 1.13/1.29/1.51 versus 1.11/1.19/1.26, so the between-class ordering is concordant with a central mechanism and only the non-null second-generation arm is anomalous); a propensity-matched target trial emulation finds no antihistamine signal (PMID:42250644); an inhaled peripherally restricted antimuscarinic shows an effect of questionable clinical importance and a null as-treated estimate (PMID:40388132); and the only randomized test, deprescribing, is uninformative (PMID:38063254). A third position is that the antihistamine-specific nulls are not informative either way. The CPRD full text reports that antihistamines were non-significant on any-prescription exposure but showed a tentative effect above 365 defined daily doses, that only about 0.3% of the sample reached that exposure on prescription, that over-the-counter supply is invisible to the database so antihistamine use is underestimated, and that the antihistamine result needs independent confirmation. Because diphenhydramine is largely bought over the counter, every prescription-database study in this literature is weakest exactly where the public question is strongest, and the two studies that do capture antihistamine exposure directly disagree with each other (PMID:38935035 positive and dose-dependent, PMID:42250644 null). Curation consequence: the exposure is recorded as PREDISPOSES, and the entry should not be strengthened to a causal claim on observational evidence alone — nor weakened to "refuted for antihistamines" on the strength of nulls drawn from data that cannot see the exposure.

Proposed experiments

Long-horizon anticholinergic deprescribing trial with dementia endpoints

randomized controlled deprescribing trial

exp_anticholinergic_deprescribing_dementia_rct

Randomize older adults with high anticholinergic burden and no cognitive impairment to structured deprescribing with non-anticholinergic substitution versus usual care, powered for incident MCI and dementia over at least five years rather than the one-to-three-month cognitive test outcomes of existing trials. Stratify by APOE4 and baseline amyloid status, and include a prespecified negative-control arm substituting a peripherally restricted agent, so that a null result can distinguish "no causal effect" from "insufficient CNS exposure contrast".

Assays

neuropsychological assessment amyloid positron emission tomography plasma biomarker profiling

Readouts

Incident dementia after burden reduction

pathophysiology#Muscarinic M1 Receptor Signaling Loss

Incidence of adjudicated MCI and dementia in the deprescribing arm relative to usual care over five or more years.

neuropsychological assessment

Direction: NEGATIVE

Interpretation: Reduced incidence after burden reduction would support causation; unchanged incidence despite a verified reduction in central anticholinergic exposure would favor reverse causation and confounding by indication.

Do plasma proteomic cellular aging clocks retain their Alzheimer disease risk stratification in younger and non-European-ancestry populations, and do the probabilistic protein-to-cell-type assignments that define them reflect true cell of origin?

KNOWLEDGE GAP OPEN gap_cellular_aging_clock_generalizability_and_cell_assignment

Attached to

pathophysiology#Neuroinflammation pathophysiology#Neurofibrillary Tangle Formation

The astrocyte, oligodendrocyte precursor cell and inhibitory neuron age gaps curated in this entry rest on two assumptions that the source study itself flags as unresolved. First, the derivation cohorts (GNPC, UK Biobank, NSHD) are predominantly older and of European ancestry, so the reported hazard ratios and the APOE-by-astrocyte-aging interaction may not transfer to younger or more diverse populations - a consequential gap for a marker proposed as an instrument of personalized risk stratification. Second, the protein-to-cell-type mapping is probabilistic: it is bounded by the cell types catalogued in the Human Protein Atlas, rests on a twofold transcriptomic enrichment threshold, and cannot fully disentangle plasma proteins released from more than one source. A "cell-type-specific" clock whose inputs are only enriched rather than exclusive is a weaker claim than the name implies, and orthogonal validation against neuropathology, CSF glial markers or neuroinflammation imaging has not yet been done. Until both are addressed, these age gaps should be curated as prognostic associations rather than as measurements of the aging state of a particular brain cell.

Show evidence (3 references)

PMID:42297981 SUPPORT Human Clinical

"Due to the fact cohorts were predominantly older and Caucasian, broader validation in younger and more diverse populations is essential to extend the generalizability of results reported in this study."

The authors' own statement of the generalizability limitation underlying the first half of this gap.

PMID:42297981 SUPPORT Human Clinical

"Analysis was based on and is thus restricted to cell types cataloged in the Human Protein Atlas, leaving certain specialized cell populations underrepresented."

Documents the coverage bound on which cell types can have a clock at all.

PMID:42297981 SUPPORT Human Clinical

"Accordingly, plasma proteins arise from multiple cellular processes not explicitly disentangled here"

The authors concede that a plasma protein assigned to one cell type may have multiple sources, which is the substance of the cell-assignment half of this gap.

Does live-attenuated herpes zoster (VZV) vaccination causally reduce dementia incidence at the population level — i.e., is the neurotropic-herpesvirus contribution to Alzheimer disease causal and preventable by vaccination — given that two regression-discontinuity analyses of the same age-based vaccine rollout reach opposite conclusions?

KNOWLEDGE GAP OPEN gap_ad_herpesvirus_zoster_vaccination_dementia_prevention

Attached to

pathophysiology#HSV-1 Reactivation in RORB+ Glutamatergic Neurons pathophysiology#Neuroinflammation

The neurotropic-herpesvirus hypothesis predicts that reducing herpesvirus reactivation (e.g., by zoster vaccination) should lower dementia risk. Multiple independent quasi-experimental studies now bear on this: three report a protective association and one eligibility-based analysis is null. On the positive side: the Welsh regression-discontinuity (RD) analysis of the date-of-birth eligibility cutoff (PMID:40175543) found that receiving the live vaccine reduced new dementia diagnoses by 3.5 percentage points over 7 years (~20% relative reduction, stronger in women); an independent Canadian analysis of natural experiments in Ontario (PMID:41579903) found that birth just before versus just after the vaccine-eligibility cutoff decreased new dementia diagnoses by 2.0 percentage points over 5.5 years, corroborated by synthetic-control comparisons against the same birth cohorts in provinces with no vaccination programme; and a vaccine-type natural experiment (DOI:10.1038/s41591-024-03201-5) found the recombinant (Shingrix) vaccine associated with even lower dementia risk than the live vaccine, again stronger in women. Against a detectable effect: an English Hospital Episode Statistics RD analysis by an overlapping author group (DOI:10.64898/2026.07.20.26358345) reproduced the expected sharp drop in hospital-coded shingles but found no detectable intention-to-treat effect of vaccine ELIGIBILITY on hospital-coded dementia, robust across specifications, placebo cutoffs, negative controls, and a separately held HES extract. A methodological reconciliation has been proposed — that the null result reflects ITT dilution by imperfect uptake plus insensitive hospital-coded ascertainment, given that dementia is largely coded in primary care and on death certificates — but it does not currently settle the disagreement, for two reasons. First, the null study tested that explanation itself: comparator analyses in Welsh data with linked primary-care and death data did not suggest its results were driven by reliance on hospital data. Second, the receipt-versus-eligibility distinction does not separate the positive studies from the null one: the Canadian headline estimate (PMID:41579903) is itself an eligibility contrast at the birth-date cutoff, like the English one, so ITT dilution alone cannot explain why the Canadian RD detects an effect where the English HES RD does not. Only the Welsh headline estimate is scaled to vaccine receipt. The designs do still differ in dementia ascertainment (primary-care records and death certificates versus hospital-coded diagnoses) and in country, health system, and follow-up length. Whether the disagreement reflects outcome ascertainment, other residual design differences, or a genuinely absent or much smaller population-level effect is unresolved — which determines how strongly the herpesvirus arm should be weighted as a causal, modifiable contributor to Alzheimer disease, and motivates a randomized controlled trial of zoster vaccination for dementia prevention.

Posed 2026-07-23T00:00:00Z

Seeded from a medRxiv paper review of DOI:10.64898/2026.07.20.26358345 (Hamilton, Geldsetzer, Davey Smith et al., 2026), then extended with a literature investigation of the wider zoster-vaccine/dementia natural-experiment landscape (Canadian Lancet Neurology RD, PMID:41579903; recombinant-vs-live vaccine-type natural experiment, Nature Medicine 2024, DOI:10.1038/s41591-024-03201-5). All four cited references were fetched into references_cache/ and their snippets verified as exact substrings of the cached abstracts before curation; the medRxiv item is a non-peer-reviewed preprint (is_preprint: true) and is curated as an epidemiological ITT counterweight, not as mechanistic proof. The residual tension is left open rather than adjudicated here: the ascertainment/dilution reconciliation is recorded as a proposed but unsettled explanation, since the null study's own Welsh comparator analyses with linked primary-care and death data did not support it, and the eligibility-versus-receipt contrast does not cleanly separate the positive studies from the null one. Complements the HSV-1-specific antiviral-suppression knowledge gap attached to the same reactivation node.

Show evidence (5 references)

PMID:40175543 SUPPORT Human Clinical

"reduced the probability of a new dementia diagnosis"

Welsh regression-discontinuity natural experiment supporting a dementia-preventing/-delaying effect of zoster-vaccine receipt (reported as a 3.5-percentage-point / ~20% relative reduction over 7 years), the positive pole of this open question.

PMID:41579903 SUPPORT Human Clinical

"decreased the probability of receiving a new dementia diagnosis by an absolute difference of 2·0 percentage points (95% CI 0·4-3·5, p=0·012) over a 5·5-year follow-up"

The Canadian study's own primary regression-discontinuity FINDING (Ontario Jan 1, 1946 eligibility threshold), reproducing the protective signal in a different country and health system. Note this is an eligibility contrast rather than a receipt-scaled estimate, which is why it also bears on the ITT-dilution explanation for the null English analysis.

DOI:10.1038/s41591-024-03201-5 SUPPORT Human Clinical

"the recombinant vaccine is associated with a significantly lower risk of dementia in the 6 years post-vaccination"

Vaccine-type natural experiment (transition from live to recombinant zoster vaccine) showing the recombinant (Shingrix) vaccine is associated with even lower dementia risk than the live vaccine, adding a vaccine-platform dimension the eligibility-based null design cannot capture and motivating a confirmatory randomized trial.

+ 2 more references

Which cells are senescent in the human Alzheimer brain — astrocytes and microglia, plaque-associated oligodendrocyte progenitor cells, or excitatory neurons — and does the answer invalidate the therapeutic rationale for senolytics in Alzheimer disease?

CONTROVERSY OPEN controversy_senescent_cell_identity_glia_versus_neurons

Attached to

pathophysiology#Senescent Cell Accumulation

The evidence conflicts on cell identity, and the conflict is consequential rather than academic, because senolytics work by killing the cells they target. Note first what the conflict is NOT: it is not a species split. A fourth study (PMID:30126037) reports both that microdissected tangle-bearing neurons from human post-mortem Alzheimer brain carry a senescence expression profile, and that in four transgenic mouse models it is neurofibrillary tangles rather than amyloid plaques that show the senescence-like phenotype. That is mouse and human evidence agreeing on NEURONS. The disagreement is between studies, not between model systems. Four studies give three answers. Genetic clearance of p16INK4A-positive cells in a tau mouse model attributes the pathology to senescent astrocytes and microglia and shows that removing them prevents tangle deposition and neuronal loss (PMID:30232451). A senolytic study in an amyloid model finds the senescent cell to be the plaque-associated oligodendrocyte progenitor cell and states explicitly that astrocytes, microglia and oligodendrocytes do not show the phenotype (PMID:30936558) — a direct contradiction of the first study's assignment, not merely a different emphasis. The largest human dataset, a senescence-eigengene analysis of roughly 140,000 nuclei from 76 postmortem brains, assigns more than 97% of senescent cells to excitatory neurons overlapping with neurofibrillary tangles (PMID:35531351). If the human result is right, a senolytic given for Alzheimer disease would be directed at postmitotic neurons rather than at dispensable glia, which is a different risk-benefit calculation from the one the mouse studies motivate. Both mouse results are causal and both human-relevant readings cannot be correct as stated. The main methodological objection to the human study is that it identifies senescence by a derived eigengene rather than a gold-standard marker; the main objection to the mouse studies is that the markers and models differ between them. Two further facts about the human study bear on how much weight it can carry, both checkable in its own methods. First, it is not an independent cohort: it generated no new sequencing data and re-scored two existing dorsolateral prefrontal cortex datasets, stating that "Datasets were accessed through Accelerating Medicines Partnership - AD (AMP-AD41) with Synapse IDs syn18485175 and syn21126462" — Mathys and Zhou respectively. Its cell-type call therefore inherits those studies' cell-type annotations and the neuron-versus-glia capture characteristics of their single-nucleus preparations. Second, to compare cell types it oversampled to equalize them, reporting that "we repeated the data of each astrocyte 104, endothelial cell 2,919, excitatory neuron 10". Nominal counts are then balanced, but the effective number of independent observations behind a rare glial or vascular cell type is far smaller than behind an excitatory neuron. That is a reason to treat a negative result in the rarer types cautiously, rather than a demonstration that the positive result in neurons is an artifact. A further consideration now bears on the therapeutic stake: the first human senolytic trial in Alzheimer disease has reported (SToMP-AD, NCT04063124; PMID:37679434, PMID:40274471). It was open-label and uncontrolled in five participants, so it cannot test efficacy, but it establishes that the question is no longer purely preclinical and that a trial can be run before the cell-identity question is settled. Its biomarker result — stable CSF amyloid-beta and tau alongside a rise in CSF interleukin-6 — is not what a simple "clear the senescent cells, lower the SASP, reduce pathology" reading predicts, and is a further reason not to treat the mouse clearance results as directly transferable. Resolving this needs marker-agnostic single-cell senescence profiling in human tissue with orthogonal confirmation, and ideally a cell-type-restricted clearance experiment. The most direct available test is to re-score senescence signatures per cell type in a cohort with enough nuclei per glial class to give the glial hypotheses a fair chance — synapse:syn26223298 (SEA-AD, 84 donors, 139 molecular cell types, openly downloadable matrices) — and to check the neuron arm against geo:GSE129308, where tangle-bearing and tangle-free somas from the same donors make the senescence-overlaps-tangles claim testable against measured rather than inferred tangle status. One caveat governs how that second test must be run: PMID:30126037 derived its senescence expression profile FROM microdissected tangle-bearing neuron transcriptomes, so re-scoring that same signature in a design that contrasts tangle-bearing with tangle-free somas is partly circular. Making the test informative requires a senescence signature defined independently of tangle status. Both are curated in this entry's `datasets:` block. Until then the Senescent Cell Accumulation node deliberately lists all four candidate cell types rather than choosing one.

Raised during a cell-biology review of this entry. All three references were fetched into references_cache/ and their snippets verified as exact substrings before curation; the two quoted methods statements were read from the cached full text of PMID:35531351. Note the entry separately carries an oligodendrocyte precursor cell plasma proteomic age gap biomarker, which is consistent with — but does not adjudicate — the OPC arm of this dispute.

Does the senile plaque form from amyloid-beta secreted into the extracellular space, or is it the residue of a single neuron that accumulated amyloid-beta in de-acidified autolysosomes and then ruptured?

CONTROVERSY OPEN controversy_plaque_origin_intraneuronal_versus_secreted

Attached to

pathophysiology#Neuronal Endosomal-Lysosomal Acidification Failure pathophysiology#Amyloid Plaque Formation

The two accounts make incompatible claims about where amyloid pathology begins, and they imply different therapeutic targets. The extracellular account, which the Amyloid Plaque Formation node and the amyloid_cascade_model encode, treats the plaque as a deposit of secreted peptide produced by sequential beta- and gamma-secretase cleavage of amyloid precursor protein; it is the premise of amyloid-directed monoclonal antibody therapy, which acts in the extracellular compartment. The inside-out account holds that autolysosome acidification declines in neurons well before extracellular deposition, that amyloid-beta and APP-βCTF build up inside enlarged de-acidified autolysosomes, and that individual neurons showing the resulting PANTHOS pattern are the principal source of senile plaques (PMID:35654956). Its human support is genuine but narrower than its mouse support: enlarged Rab5-positive endosomes are the earliest known intraneuronal change in sporadic disease and are present at preclinical stages (PMID:10880397), and intraneuronal amyloid-beta localizes principally to those endosomes (PMID:15465622), but the only verified human claim about PANTHOS itself is that the pattern is present in Alzheimer brains — not that it is the source of human plaques, a claim its authors scope to amyloid precursor protein transgenic models where APP is overexpressed. A further complication sits inside the human data: endosomes were of normal size in advanced presenilin-mutation familial disease (PMID:10880397), so the endosomal route appeared not to generalize across Alzheimer genetic subtypes, even though PSEN1 is independently required for lysosomal acidification (PMID:20541250). That tension is now partly resolved rather than merely recorded: in isogenic human iPSC neurons, PSEN1 mutants do accumulate β-CTFs and do enlarge Rab5-positive endosomes (PMID:31416668), so the postmortem negative may reflect end-stage tissue or the limits of an endosome-size readout rather than absence of the mechanism in that subtype. Two things sharpen how this could be settled. First, the discriminating question is not transcriptomic: organelle pH, v-ATPase subunit assembly, peptide sub-cellular location and plaque-to-neuron correspondence are imaging, proteomics and quantitative-neuropathology phenotypes, so none of the single-nucleus datasets curated in this entry can adjudicate them. Second, and more usefully, the relevant human data type already exists: laser-capture proteomics of human amyloid plaques against neighbouring tissue reports that endosomal and lysosomal proteins are particularly highly enriched within plaques (PMID:35418158). That is qualitatively what an intraneuronal origin predicts. It does not exclude glial or dystrophic-neurite lysosomal contributions, so it does not settle the question — but it does mean the crux is re-analyzable against existing human proteomics rather than blocked on data that do not exist. Deciding it still needs quantitative human neuropathology establishing what fraction of plaques have a neuronal origin, in brains not overexpressing APP.

Raised during a cell-biology review of this entry. Curated as a CONTROVERSY rather than a KNOWLEDGE_GAP because the two accounts are both actively defended and make opposing claims, rather than one of them being simply unstudied. The PANTHOS phenotype and the APP-βCTF/v-ATPase mechanism come from the same laboratory and are not independent replications of one another.

Does the C1q/C3/CR3 complement pathway that mediates microglial synapse elimination in Alzheimer mouse models operate in the human Alzheimer brain, where the only direct evidence of glial synapse ingestion implicates MFG-E8 and involves astrocytes as well as microglia?

HUMAN MODEL MISMATCH OPEN mismatch_complement_synapse_pruning_mouse_versus_human_opsonin

Attached to

pathophysiology#Complement-Mediated Microglial Synapse Elimination

The necessity evidence for the complement route is entirely murine, and it is strong: blocking C1q, C3, or the microglial receptor CR3 each reduces early synapse loss (PMID:27033548), and a C1q-blocking antibody rescues synapse density in a tau model (PMID:30392797). The human evidence establishes the phenomenon but not the pathway. The study that directly demonstrates glial synapse ingestion in human Alzheimer tissue states that direct human evidence for glial involvement in synapse removal remained to be established, and finds that blocking MFG-E8 — not complement — reverses the elevated engulfment, with astrocytes participating alongside microglia (PMID:37652017). Human corroboration of complement specifically is better than this discussion originally recorded, and the correction is worth stating precisely. Beyond C1q in postsynaptic densities, C3 protein is elevated in Alzheimer patient brains including at synapses, with cerebrospinal fluid C3 tracking tau (PMID:31433986); complement dysregulation with C1q deposited on tau aggregates extends to human tauopathies where amyloid is absent (PMID:42271460); and common-variant genetics implicates two complement loci, CR1 and CLU (PMID:29504051). The genetic association matters most, because unlike every postmortem correlation it cannot be a downstream consequence of established pathology. What none of this establishes is the executing step. No human study shows glia phagocytosing structurally intact synapses via complement opsonization. The phenomenon is human-confirmed and the pathway is human-implicated; the join between them is not. This matters because anti-C1q antibody therapy is in clinical development on the strength of the mouse pathway. If the dominant human opsonin is MFG-E8, or if astrocytes rather than microglia are the principal effector, the mouse result could be real and the therapy still miss its target. A second, separate caution comes from within the mouse literature itself: C3 deletion protects synapses and cognition while increasing plaque burden (PMID:28566429), so complement inhibition trades amyloid clearance against synapse preservation rather than being unambiguously beneficial. Resolving this needs human tissue or human cellular models in which the complement and MFG-E8 routes are blocked separately and compared, and ideally a synaptic-density readout from the ongoing anti-C1q trials. Note that the single-nucleus datasets curated in this entry cannot close this gap: they can show where and when complement genes are induced and whether the complement and MFG-E8 routes are expressed in the same cells, but engulfment is a phagocytic event and opsonin identity is a protein-level question, neither of which is visible in transcript abundance.

Raised during a cell-biology review of this entry. All four references were fetched into references_cache/ and their snippets verified as exact substrings before curation. The MFGE8 gene is listed on the pathophysiology node alongside the complement genes to keep this alternative visible at the point of use.

Which tau species is the proximate neurotoxic entity in Alzheimer disease — soluble pretangle assemblies or mature neurofibrillary tangles — and is the glial state that separates demented from cognitively resilient people at matched tangle burden a required intermediary or a parallel process?

KNOWLEDGE GAP OPEN gap_ad_tau_toxic_species_and_glial_cofactor

Attached to

mechanistic_hypotheses#tau_neurodegeneration_model pathophysiology#Neurofibrillary Tangle Formation

Two claims that would materially change how this entry reads are currently unsupported by anything citable, and both came out of the OpenScientist tau_neurodegeneration_model report with citations that do not carry them. The first is that soluble pretangle tau rather than the mature tangle is the primary toxic species, which would relocate the toxicity of this hypothesis upstream of the lesion it is named for. The sentence the report quotes for it is background framing in PMID:41695614, not that paper's result; what the paper does establish is that pretangle moieties marked by pS422, TOC1 and TNT2 are present as neuropil threads and neuronal inclusions in default-mode-network hubs from Braak stage III, with stepwise increases through Braak IV to V. The mirror claim, that tangle formation is a protective sequestration step, is worse served: PMID:31456657 is an in vitro study of how EFhd2 alters tau liquid liquid phase separation and measures no toxicity, no neurons and no protection. The oligomer-versus-tangle question is a real and live controversy; neither citation can settle it, and it should not be curated as evidence. The second claim is that glial activation is a required intermediary converting tau pathology into clinical disease. The observation behind it is solid and is curated on the hypothesis block: at matched Braak III-IV burden, demented and resilient brains differ in glial phenotype, including in visual cortex carrying no overt tau (PMID:35534858, 55 brains). But the design is cross-sectional post-mortem comparison. Nothing cited perturbs glia and measures the degenerative or clinical consequence at a fixed tau burden, which is what necessity would require. Resolving both needs the same kind of study: a perturbation with a tau-species-resolved and a glia-resolved readout in the same system, ideally human tissue or human cellular models, with tangle burden held constant.

Raised from the OpenScientist tau_neurodegeneration_model assessment. Recorded as a gap rather than as evidence precisely because the report's two supporting citations are framing sentences rather than results; see the assessment sidecar for the per-claim reasoning.

What is the relative contribution of each upstream stressor — amyloid, tau, complement-mediated pruning, vascular and glymphatic failure, oxidative injury, gut-derived inflammation, infection — to synaptic failure in Alzheimer disease, and what observation would falsify the convergence model as stated?

KNOWLEDGE GAP OPEN gap_ad_synaptic_convergence_relative_stressor_weights

Attached to

mechanistic_hypotheses#synaptic_failure_convergence_model pathophysiology#Synaptic Dysfunction

The convergence model's descriptive breadth is also its methodological weakness, and the OpenScientist synaptic_failure_convergence_model report says so about its own hypothesis: by accommodating nearly every known Alzheimer mechanism as an upstream stressor, the model risks becoming unfalsifiable. The concrete form of that critique is that no study quantifies how much of the synaptic loss each stressor accounts for, so no arm of the model can be weakened by evidence without weakening all of them equally. The arm-by-arm support is also markedly uneven, which the summary verdict hides. The report's own link-strength table scores five of eleven causal links as established, five as emerging, and the infectious arm as speculative against a negative randomised trial of valacyclovir. The amyloid, tau and complement arms rest on multiple independent lines; the gut-vesicle route into complement pruning (PMID:40731189) and the glymphatic-to-synapse association (PMID:41454064) each rest on one study; and the infectious arm belongs in the separate hsv1_reactivation_model block rather than here. What would move this is a formalisation rather than more association studies: a weighted causal graph fitted to multimodal human data in which stressor contributions are estimated jointly and can be driven to zero. Note also that synaptic loss appears to sit at the pathology-to-symptom step rather than at pathology onset — hippocampal synaptic density is similar in controls and in cognitively unimpaired people with biological Alzheimer disease, and falls once impairment appears (PMID:40465636) — so any such model has to explain the threshold, not only the slope.

Raised from the OpenScientist synaptic_failure_convergence_model assessment. The report recommended no status change and the CANONICAL status is unchanged; this records the self-critique, which was the most useful thing in the report and would otherwise have been lost. The report's ontology lead binding complement-mediated synapse elimination to GO:0150076 was rejected in assessment and not curated: that term is the neuroinflammatory response.

Does neurovascular and blood-brain barrier dysfunction cause or accelerate tau pathology in Alzheimer disease, or does its explanatory reach stop at amyloid accumulation while tau follows its own propagation logic?

KNOWLEDGE GAP OPEN gap_ad_vascular_to_tau_causal_link

Attached to

mechanistic_hypotheses#vascular_bbb_clearance_model pathophysiology#Vascular Dysfunction

This is the boundary that keeps the vascular model at ALTERNATIVE, and the entry did not previously record it. The evidence that vascular injury acts on amyloid is now substantial and partly interventional: hypoperfusion raises soluble and then insoluble amyloid in a transgenic model within one to three months (PMID:28667120), APOE4-graded pericyte degeneration tracks barrier breakdown in human post-mortem tissue (PMID:25757756), and a common risk locus routes into barrier efflux through PICALM (PMID:30041013). Nothing comparable connects the vascular arm to tau, which propagates by templated, prion-like spread along connected circuits. Because tau burden rather than amyloid is the stronger correlate of cognitive decline, a model that explains amyloid accumulation but not tau has bounded reach even where it is correct. Two further observations sharpen the gap rather than closing it. Cerebral blood flow associations with tau invert by disease stage — negative in mild cognitive impairment, positive in established disease (PMID:40884970) — so the vascular arm does not even relate to tau monotonically. And no human interventional trial has shown that correcting vascular dysfunction prevents or slows amyloid accumulation. The rrAD trial (NCT02913664) randomises 640 cognitively normal hypertensive adults to intensive vascular risk reduction, exercise, or both, but its primary outcomes are neurocognitive with MRI at baseline and 24 months, so it will not deliver the serial amyloid PET endpoint this question needs.

Raised from the OpenScientist vascular_bbb_clearance_model assessment, which identified this as the hypothesis's most important limitation. The trial identifier was checked against the cached rationale-and-methods publication (PMID:30826452), which carries NCT02913664.

Does the Arc-dependent extracellular-vesicle route for tau propagation replicate outside the founding laboratory, and does the complete absence of human genetic association for ARC argue against it carrying a material share of tau spread in Alzheimer disease?

KNOWLEDGE GAP OPEN gap_ad_arc_ev_tau_replication_and_genetic_support

Attached to

mechanistic_hypotheses#ev_mediated_tau_propagation_model pathophysiology#Intercellular Tau Transmission via Extracellular Vesicles

Two independent reasons to hold this mechanism at EMERGING, neither of which is a criticism of the founding result. The first is single-source dependency with a disclosed commercial interest. Arc binding tau, and the near-abolition of intercellular tau transmission in Arc-knockout animals, both come from PMID:42372723 and from nowhere else; the paper appeared online only shortly before the provider search, so the absence of replication is expected rather than suspicious. What makes it worth recording is that the senior author discloses co-founding VNV, LLC and consulting for Aera Therapeutics, which licenses patents covering Arc capsids. Single-source results with commercial interest in the mechanism are exactly where independent replication carries the most weight. A replication should also use authentic Alzheimer-type 3R/4R paired helical filaments rather than the P301L 4R tau of the original model. The second is the absence of any human genetic footprint. A provider survey of UniProt, Open Targets, the GWAS Catalog and GenCC found no genetic, GWAS, somatic-mutation or drug-target association between ARC and Alzheimer disease. That absence is a legitimate structural argument, and it contrasts with the competing tunnelling-nanotube route, which does connect to Alzheimer risk genes. It is recorded here as a gap rather than as evidence because none of the four surveys was committed with a query, release version, timestamp or saved response, and an Open Targets score quoted to four decimal places is not reproducible without a release version. Re-derive before relying on it. A third, separate question the same report raised should not be mistaken for a finding: whether hyperexcitability induces Arc, which raises vesicle tau release, which raises excitability in recipient neurons, closing a self-amplifying loop. Each link has separate support but no experiment closes the circuit, and the amyloid-Arc suppression already curated on this hypothesis cuts against the loop operating where Alzheimer pathology is heaviest. It belongs here as a question, never as pathophysiology edges.

Raised from the OpenScientist ev_mediated_tau_propagation_model assessment. The report's proposed tau-strain snippet for PMID:33693528 was rejected as a lightly reworded paper title and not curated; its 8-month pathology-convergence claim is not in the cached abstract and no early-stage restriction of the model was made on the strength of it.

Does glucosamine supplementation worsen cognitive decline or survival in people who already have Alzheimer disease, and is any harm confined to brains that are already hyperglycosylated?

KNOWLEDGE GAP OPEN gap_glucosamine_in_established_ad

Attached to

environmental#Oral glucosamine supplementation pathophysiology#Brain N-Glycan Hyperglycosylation mechanistic_hypotheses#n_glycan_hyperglycosylation_model

Mouse data show glucosamine raising brain N-glycans and worsening memory in 5xFAD mice but not in wild-type mice, and one health-system EHR analysis associates documented use with worse survival in dementia. Large population cohorts associate the same supplement with lower or unchanged dementia incidence. The two bodies of evidence ask different questions (progression after diagnosis versus incidence), and all of the human data are observational, so whether a widely used over-the-counter supplement is harmful in established disease is unresolved.

Proposed experiments

Randomized placebo-controlled glucosamine trial in biomarker-confirmed Alzheimer disease

randomized double-blind placebo-controlled trial

exp_glucosamine_rct_established_ad

Randomize people with biomarker-confirmed mild cognitive impairment or mild dementia due to Alzheimer disease who already take glucosamine to continuation or blinded placebo substitution, with cognitive, functional and survival endpoints and, where feasible, CSF or imaging readouts of glycosylation. Stratify by osteoarthritis severity to separate the supplement from its indication.

Assays

neuropsychological assessment

Readouts

Rate of cognitive decline

phenotypes#Memory Loss

Direction: INCREASED

Interpretation: Faster decline on continued glucosamine than on placebo would support the exacerbation claim; no difference would argue that the EHR signal reflected confounding.

Decision criterion

The environmental entry should move from EXACERBATES to a documented null if a trial powered for cognitive decline finds no difference, and the hypothesis group can advance past EMERGING only if the trial shows harm together with evidence that glucosamine alters brain glycosylation in humans.

Would support

mechanistic_hypotheses#n_glycan_hyperglycosylation_model

Would refute

mechanistic_hypotheses#n_glycan_hyperglycosylation_model

Show evidence (1 reference)

PMID:42265388 SUPPORT Human Clinical

"These findings emphasize the urgent need for a well-designed, double-blind clinical trial to systematically evaluate the impact of glucosamine on cognitive decline in patients with dementia."

The originating authors identify a randomized trial as the missing test of their EHR association.

Posed 2026-09-27T00:00:00Z

Is protein N-glycosylation increased or decreased in the Alzheimer brain, and can a net rise in released N-glycans coexist with site-specific loss of glycosylation on synaptic proteins?

KNOWLEDGE GAP OPEN gap_n_glycan_direction_in_ad_brain

Attached to

pathophysiology#Brain N-Glycan Hyperglycosylation pathophysiology#Synaptic Dysfunction

MALDI imaging of released N-glycans finds a net increase in human and mouse AD brain, while site-specific glycoproteomics of post-mortem AD brain finds reduced glycosylation at many sites, enriched on synaptic channels and receptors, and proposes that loss of glycosylation drives synaptic dysfunction. A third glycoproteomic study finds stage-associated changes in glycan structure at individual sites without a single direction. The methods measure different quantities, so the findings need not conflict, but they point therapy in opposite directions: reducing N-glycan biosynthesis is protective under one reading and harmful under the other.

Proposed experiments

Paired released-glycan imaging and site-specific glycoproteomics on the same AD tissue

cross-sectional post-mortem tissue study

exp_paired_glycan_imaging_and_site_glycoproteomics

On matched sections from the same post-mortem AD and control brains, quantify released N-glycans by MALDI imaging and site occupancy by site-specific glycoproteomics, then test whether net glycan abundance and synaptic site occupancy move in opposite directions within the same samples and brain regions.

Assays

MALDI mass spectrometry imaging of released N-glycans site-specific N-glycoproteomics

Readouts

Net released N-glycan abundance

pathophysiology#Brain N-Glycan Hyperglycosylation

Protein N-linked glycosylation GO:0006487

Direction: INCREASED

Interpretation: An increase alongside reduced synaptic site occupancy in the same samples would reconcile the two literatures; a decrease would refute the hyperglycosylation node in that tissue.

Decision criterion

If released-glycan abundance is not increased when measured side by side with site occupancy, the hyperglycosylation node should be downgraded to HYPOTHETICAL.

Would support

pathophysiology#Brain N-Glycan Hyperglycosylation

Would refute

pathophysiology#Brain N-Glycan Hyperglycosylation

Posed 2026-09-27T00:00:00Z

Show evidence (3 references)

PMID:42265388 SUPPORT Human Clinical

"Our work complements these findings by measuring the net abundance of released N-glycans across brain regions rather than peptide-level occupancy."

The originating authors' own statement that they and the site-specific studies measure different quantities.

PMID:41042700 SUPPORT Human Clinical

"Our findings suggest that the loss of N-glycosylation may contribute to the pathogenesis of AD through impairing synaptic transmission and promoting protein aggregation."

The competing directional claim this discussion asks to be resolved.

PMID:36309312 SUPPORT Human Clinical

"We observed AD-associated differences in the number of antennae, frequency of fucosylation, bisection, and other monosaccharides at individual glycosylation sites among samples from our three groups."

A third, independent dataset showing that N-glycan structure changes across asymptomatic and symptomatic AD at individual sites, without reducing to a single direction.

Does the respiratory deficit produced by micro-MKKS63 knockout in an immortalized human microglial line occur in microglia in the Alzheimer brain, where the microprotein is reduced rather than absent and where the depletion was measured in bulk tissue that cannot say which cell type lost it?

HUMAN MODEL MISMATCH OPEN mismatch_micro_mkks63_hmc3_versus_human_microglia

Attached to

pathophysiology#Micro-MKKS63 Microprotein Depletion pathophysiology#Microglial Oxidative Phosphorylation Failure experimental_models#HMC3 human microglial micro-MKKS63 CRISPR-Cas9 knockout lines

Two claims are being joined here, and each is sound on its own. The human claim is that micro-MKKS63 falls in symptomatic Alzheimer cortex; it comes from tandem-mass-tag quantitation across hundreds of ROSMAP donors. The cellular claim is that losing the microprotein impairs oxidative phosphorylation; it comes from two independent CRISPR knockout lines in HMC3 cells. The join between them is the weak point, and it fails in two distinct ways rather than one. First, cell type. The depletion is a bulk-tissue measurement. Microglia are a small minority of cortical cells, so a bulk decrease is more likely to reflect the majority populations than the microglial one, and cell-type attribution cannot be recovered from the tandem-mass-tag data at all. It is entirely possible that micro-MKKS63 falls in neurons while microglial levels are unchanged, in which case the human observation and the cellular experiment would be about different cells. The microprotein's coexpression network is enriched for mitochondrial biology, which is consistent with a role in any high-demand cell type and does not adjudicate this. Second, dose and context. The knockout abolishes the microprotein; the disease shows a partial decrease. For micro-MKKS63 specifically the log2 fold change is -0.063 - about a 4% reduction - which is typical rather than unusually small: the median absolute log2 fold change across all significant microproteins in the stringent model is 0.0969, which the authors attribute to cellular heterogeneity in bulk postmortem tissue. Whether a reduction of that size produces any respiratory change is untested, and there is no dose-response series. HMC3 is additionally an SV40-immortalized fetal-derived line grown in standard culture, a system whose metabolic setpoint and transcriptional state are known to diverge from microglia in aged tissue. Resolving this needs the cell-type question answered first, because the dose question is only worth asking in whichever cell type actually loses the microprotein. Nothing currently curated in this entry can answer it: single-nucleus datasets report transcripts, and a smORF's transcript is usually shared with the host locus, so transcript-level data cannot separate micro-MKKS63 from canonical MKKS - which is precisely the decoupling the finding rests on.

Raised while curating the microprotein atlas into this entry. Recorded as HUMAN_MODEL_MISMATCH rather than KNOWLEDGE_GAP because the evidence is not absent: it exists in a model system, and its translational validity is the open question. The companion discussion gap_micro_mkks63_causal_role_in_alzheimer_disease covers the separate question of causal direction.

Show evidence (1 reference)

PMID:42736449 SUPPORT DIRECT PRIMARY RESULT Human Clinical

"Effect sizes reflect directional associations rather than large shifts, as median absolute log2(fold change) among significant stringent-model MPs was 0.0969, consistent with cellular heterogeneity in bulk postmortem tissue"

Supports the dose half of the mismatch, and does so in the authors' own words: the disease-associated microprotein changes are small directional shifts in bulk tissue, not the complete loss the knockout imposes. The cellular-heterogeneity attribution is also the cell-type half of the mismatch stated from the other direction.

Is micro-MKKS63 depletion a contributor to Alzheimer disease, a consequence of it, or an incidental correlate - and if it contributes, through what downstream step?

KNOWLEDGE GAP OPEN gap_micro_mkks63_causal_role_in_alzheimer_disease

Attached to

pathophysiology#Micro-MKKS63 Microprotein Depletion pathophysiology#Microglial Oxidative Phosphorylation Failure mechanistic_hypotheses#smorf_microprotein_model

The Microglial Oxidative Phosphorylation Failure node is deliberately terminal, and this discussion records why. Microglial bioenergetic failure has obvious candidate consequences in this entry - impaired phagocytic clearance of amyloid, a shift toward a proinflammatory state feeding Neuroinflammation, convergence on the mitochondrial nodes already curated - and not one of them was tested. Drawing any of those edges would convert a plausible narrative into a curated causal claim, which is the failure mode this entry's other discussions were written to avoid. Direction is equally unsettled. Postmortem tissue in advanced disease cannot distinguish a change that drove pathology from one produced by it, and Alzheimer cortex is a tissue in which mitochondrial function, cell composition, and translation are all already disturbed - so a microprotein that is mitochondria-associated has several routes to falling as a bystander. There is no genetic evidence to break the tie: the MKKS locus carries no known Alzheimer association, and the canonical protein's disease association is with McKusick-Kaufman and Bardet-Biedl syndromes, which the microprotein result does not touch and which are not evidence about this ORF. What would move it: a cell-type-resolved measurement of the microprotein in Alzheimer and control cortex, to establish which cell loses it; a partial knockdown series in human microglia matched to the disease-scale reduction, with phagocytosis and cytokine readouts rather than respirometry alone, to establish whether the deficit propagates; and a variant-level lookup at the smORF in Alzheimer genome-wide association data, which is the only design among these that could establish direction rather than association.

Proposed experiments

Cell-type-resolved micro-MKKS63 quantitation in Alzheimer and control cortex

cell-type-resolved quantitative proteomics of postmortem tissue

exp_micro_mkks63_celltype_resolved_depletion

Quantify micro-MKKS63 in microglia, neurons, astrocytes and oligodendrocytes separately from Alzheimer and non-Alzheimer dorsolateral prefrontal cortex, using cell-type-sorted or laser-capture material with a microprotein-aware search database and a targeted assay for the smORF's unique tryptic peptides. The experiment exists to answer which population loses the microprotein, which is a precondition for interpreting the knockout result at all.

Perturbations

Alzheimer disease status

pathophysiology#Micro-MKKS63 Microprotein Depletion

Not an applied intervention but the contrast under study: symptomatic Alzheimer donors against non-Alzheimer donors, matched for age, sex and postmortem interval.

Assays

fluorescence-activated nuclei or cell sorting targeted parallel reaction monitoring mass spectrometry

Readouts

Micro-MKKS63 abundance per cell type

pathophysiology#Micro-MKKS63 Microprotein Depletion

Abundance of micro-MKKS63 unique peptides in each sorted population, expressed against the same population in control donors, with canonical MKKS peptides quantified alongside as the decoupling control.

Direction: DECREASED

Interpretation: A microglia-selective decrease would join the human observation to the cellular experiment. A decrease confined to neurons would mean the two are about different cells and that the knockout result, while real, models something the disease does not do.

Would support

pathophysiology#Microglial Oxidative Phosphorylation Failure

Supporting outcome

Micro-MKKS63 is reduced in Alzheimer microglia specifically, while canonical MKKS is unchanged in the same cells.

Refuting outcome

Micro-MKKS63 is unchanged in Alzheimer microglia, with the bulk-tissue decrease accounted for by neurons or another non-microglial population.

Listed first because it gates the others: a dose-response knockdown in microglia is not worth running if microglia are not the cells that lose the microprotein.

Disease-scale graded knockdown of micro-MKKS63 in human microglia

graded loss-of-function series in human iPSC-derived microglia

exp_micro_mkks63_graded_knockdown_microglial_function

Titrate micro-MKKS63 loss in human induced-pluripotent-stem-cell-derived microglia across a range spanning the reduction seen in Alzheimer cortex up to complete loss, leaving the canonical MKKS ORF intact, and measure respiration alongside the microglial functions that would connect a bioenergetic deficit to this entry's existing pathograph.

Perturbations

Graded micro-MKKS63 depletion

pathophysiology#Micro-MKKS63 Microprotein Depletion

An allelic or knockdown series producing partial reductions matched to the disease-scale effect size as well as complete loss, with the main MKKS ORF preserved throughout.

MKKS locus (micro-MKKS63 smORF) hgnc:7108

Assays

extracellular flux respirometry amyloid-beta phagocytosis assay multiplexed cytokine profiling

Readouts

Respiratory capacity across the depletion series

pathophysiology#Microglial Oxidative Phosphorylation Failure

Basal, ATP-linked and maximal oxygen consumption at each depletion level, in a human microglial system rather than an immortalized line.

Direction: DECREASED

Interpretation: A deficit appearing only at complete loss would mean the published knockout overshoots the disease; a graded deficit reaching down to the disease-scale reduction would make the bioenergetic claim relevant to Alzheimer cortex.

Amyloid-beta phagocytic capacity

pathophysiology#Microglial Oxidative Phosphorylation Failure

Uptake and degradation of fibrillar amyloid-beta by microglia at each depletion level.

Direction: DECREASED

Interpretation: This is the readout that would justify an edge from the bioenergetic node into the amyloid arm of this entry. Without it the node stays terminal, which is its current state.

Would support

pathophysiology#Microglial Oxidative Phosphorylation Failure

Supporting outcome

Disease-scale depletion reproduces the respiratory deficit and impairs amyloid phagocytosis or shifts cytokine output.

Refuting outcome

Only complete loss affects respiration, and microglial function is unchanged at disease-scale depletion.

Raised while curating the microprotein atlas into this entry, alongside mismatch_micro_mkks63_hmc3_versus_human_microglia, which covers the separate translational-validity question. The proposed experiments are ordered: the cell-type-resolved measurement gates the knockdown series.

⚙

Pathophysiology

Amyloid Plaque Formation

Accumulation of amyloid-beta proteins in the brain, forming extracellular plaques that disrupt cell function and communication between neurons.

Amyloid Cascade Model Neuroimmune-Glial Amplification Model Synaptic Failure Convergence Model Tau Neurodegeneration Model

Neurons CL:0000540 Microglia CL:0000129 Astrocytes CL:0000127

APP hgnc:620 PSEN1 hgnc:9508 PSEN2 hgnc:9509

Protein misfolding GO:0006457 ⚠ ABNORMAL Amyloid-beta generation by gamma-secretase cleavage of APP GO:0034205 Amyloid-beta aggregation into fibrils GO:1990000 ↑ INCREASED Neuroinflammation GO:0150076

Cerebral Cortex Hippocampus Subcortical Regions

Downstream

Neurofibrillary Tangle Formation Amyloid-beta accumulation is modeled as an upstream trigger of tau hyperphosphorylation and neurofibrillary tangle formation.

Synaptic Dysfunction Soluble amyloid-beta assemblies are modeled as direct toxic contributors to synaptic dysfunction and cognitive impairment.

Neuroinflammation Amyloid-beta accumulation and activated microglia can form a reinforcing inflammatory cycle.

Show evidence (8 references)

PMID:28320296 SUPPORT

"These intra- or extracellular insoluble aggregates (fibers or plaques) are hallmarks of many neurodegenerative pathologies including Alzheimer's disease (AD)..."

This reference supports the statement by confirming that amyloid plaques are a hallmark of Alzheimer's disease and discusses the formation and role of amyloid aggregates in AD.

PMID:26322584 SUPPORT

"Alzheimer disease (AD) is characterized by the accumulation of amyloid plaques, which are predominantly composed of amyloid-beta peptide."

This reference supports the statement by indicating that the accumulation of amyloid plaques, composed of amyloid-beta peptide, is a characteristic of Alzheimer's disease.

PMID:22351073 SUPPORT

This reference supports the statement by mentioning that amyloid-beta peptides lead to Alzheimer's disease and cause synaptic dysfunction.

+ 5 more references

Neurofibrillary Tangle Formation

Intracellular accumulation of hyperphosphorylated tau protein, forming twisted fibers that disrupt cellular transport and eventually lead to neuronal death.

Arc-Dependent EV-Mediated Tau Propagation Model Necroptosis Model of Neuronal Death Synaptic Failure Convergence Model Tau Neurodegeneration Model

Neurons CL:0000540

MAPT hgnc:6893 GSK3B hgnc:4617

Protein hyperphosphorylation GO:0006468 ↑ INCREASED Microtubule destabilization GO:0007019 Neurofibrillary tangle assembly GO:1902988 ↑ INCREASED

Tau-protein kinase activity (GSK3-beta and other tau kinases) GO:0050321 ↑ INCREASED

Entorhinal Cortex Hippocampus Neocortex

Downstream

Synaptic Dysfunction Hyperphosphorylated tau and tau aggregates are modeled as toxic contributors to impaired synaptic plasticity and neurotransmitter release.

Necroptotic Neuronal Death Hyperphosphorylated tau is modeled as an upstream trigger of necrosome assembly, connecting the entry's central proteinopathy to the execution step by which neurons are proposed to die.

Intercellular Tau Transmission via Extracellular Vesicles Intracellular pathological tau is packaged by Arc into neuronal extracellular vesicles and released, providing a route for cell-to-cell spread of tau seeds.

Show evidence (10 references)

PMID:21509508 SUPPORT

"Extracellular accumulations of Abeta, hyperphosphorylation of tau and intracellular neurofibrillary tangle formation have been the hallmarks of Alzheimer's Disease (AD)."

The article discusses the role of tau hyperphosphorylation and neurofibrillary tangle formation in Alzheimer's Disease, supporting the statement about these processes.

PMID:36001963 SUPPORT

"In Alzheimer's disease and other neurodegenerative tauopathies, tau is found hyperphosphorylated and aggregated in neurofibrillary tangles."

This article supports the statement by mentioning the hyperphosphorylation and aggregation of tau in neurofibrillary tangles in Alzheimer's Disease.

PMID:37266762 SUPPORT

"Tau is a microtubule-associated binding protein in the nervous system that is known for its role in stabilizing microtubules throughout the nerve cell. It accumulates as beta-sheet-rich aggregates and neurofibrillary tangles, leading to an array of different pathologies."

The article confirms that hyperphosphorylated tau leads to the formation of neurofibrillary tangles and disrupts microtubule stabilization, supporting the statement.

+ 7 more references

Synaptic Dysfunction

Progressive loss of synapses and impaired neurotransmitter signaling, leading to disrupted neuronal communication and cognitive decline.

Synaptic Failure Convergence Model

Neurons CL:0000540

Neurotransmitter release GO:0007269 ↓ DECREASED Synaptic plasticity GO:0048167 ↓ DECREASED

Synapses Neuronal dendrites

Computational models

Digital Alzheimer's Disease Diagnosis (DADD) patient-specific brain model DIGITAL TWIN Python 3.10; The Virtual Brain; NumPy; SciPy; scikit-learn The personalized lp parameter represents synaptic degeneration through modeled cortical hypo-inhibition and hyperexcitation.

Downstream

Cortical Structural Connectivity Degeneration Early synaptic alterations precede structural degradation in vulnerable medial-temporal white-matter pathways along the aging-dementia continuum.

Memory Loss Synaptic plasticity failure and aberrant neurotransmitter release are modeled as proximate drivers of cognitive decline, including memory impairment.

Show evidence (3 references)

PMID:27662312 SUPPORT

The statement is supported by the reference which discusses the role of neurotransmitter release and synaptic plasticity in Alzheimer's Disease, leading to cognitive decline.

PMID:12973746 SUPPORT

The reference supports the statement by discussing the loss of synapses and impaired glutamatergic neurotransmitter signaling in Alzheimer's Disease.

PMID:33232936 SUPPORT Model Organism

"In the present study, we examined alterations in several GABA-specific targets in the APP/PS1 mouse model at different ages... Overall, our study provides evidence of altered GABAergic signaling in an amyloid model of AD at a time point consistent with AD-related deficits."

The reference supports the statement by providing evidence of altered GABAergic signaling and its role in Alzheimer's Disease.

Cortical Structural Connectivity Degeneration

Loss and reorganization of long-range cortical connections disrupt large-scale brain-network communication during the preclinical and clinical Alzheimer disease continuum.

Cerebral cortex UBERON:0000956 White matter UBERON:0002316

Computational models

Digital Alzheimer's Disease Diagnosis (DADD) patient-specific brain model DIGITAL TWIN Python 3.10; The Virtual Brain; NumPy; SciPy; scikit-learn Model inversion estimates patient-specific connectivity degeneration and neuroplastic rewiring from the similarity between simulated and observed EEG signals.

Downstream

Memory Loss Progressive grey- and white-matter network disconnection accompanies the transition from mild memory impairment to broader cognitive decline.

Show evidence (1 reference)

PMID:40450374 SUPPORT Computational

"DADD parameters captured Alzheimer’s-related micro- and macroscopic structural neurodegeneration mechanisms (see Methods)."

The patient-specific model identifies structural connectivity degeneration as an Alzheimer-related mechanism represented in its digital biomarkers.

Neuroinflammation

Chronic activation of immune responses in the brain, contributing to neuronal damage and disease progression.

Amyloid Cascade Model Autophagy-Senescence-Inflammasome (ASI) Axis Neuroimmune-Glial Amplification Model Tau Neurodegeneration Model Vascular and Blood-Brain Barrier Clearance Model

Microglia CL:0000129 Astrocytes CL:0000127

TREM2 hgnc:17761 NLRP3 hgnc:16400 PYCARD (ASC) hgnc:16608

Immune activation GO:0002253 ↑ INCREASED Microglial cell activation GO:0001774 ↑ INCREASED Astrocyte activation GO:0048143 ↑ INCREASED Cytokine production GO:0001816 ↑ INCREASED Phagocytosis GO:0006909

Inflammasome adaptor activity (NLRP3 PYD filament nucleation; ASC/PYCARD bridging to caspase-1) GO:0035591 Caspase-1 activation by the assembled inflammasome (NLRP3, ASC/PYCARD) GO:0140608 TREM2-TYROBP microglial receptor signaling GO:0004888

Downstream

Amyloid Plaque Formation Inflammasome activation in microglia can reinforce amyloid-beta aggregation through extracellular ASC specks, making amyloid-to-glia ordering partly reciprocal rather than a one-way edge.

Neurofibrillary Tangle Formation Microglial state is modeled as a context-dependent modifier of amyloid-associated tau seeding and spreading; TREM2-dependent disease-associated microglia can restrain tau propagation, while a simple pathogenic NLRP3-to-tau edge remains contested.

Oxidative Stress Activated glial cells are modeled as sources of inflammatory mediators and oxygen radicals that amplify oxidative neuronal injury.

Vascular Dysfunction Neuroinflammation is modeled as a contributor to blood-brain barrier breakdown and neurovascular-unit dysfunction.

Senescent Cell Accumulation Inflammasome/IL-1 signaling is modeled as reinforcing rather than merely resulting from glial senescence, closing a reciprocal loop with the existing senescence-to-neuroinflammation edge rather than a one-way cascade.

Show evidence (12 references)

PMID:28019679 SUPPORT

The reference supports the involvement of NLRP3 inflammasome activation in Alzheimer's disease, which is part of the immune activation process contributing to neuroinflammation.

PMID:32061803 SUPPORT

"The response of astrocytes to the presence of Abeta, as well astrocytic and microglial interaction and inflammatory cytokine release is also discussed, highlighting a cyclical behaviour of these cells in contributing to AD pathogenesis."

This reference supports the role of astrocytes and microglia in cytokine production and their interaction contributing to Alzheimer's disease pathology.

PMID:35406803 SUPPORT

"Once activated, microglial cells, which are brain-resident immune cells, exert several key actions, including phagocytosis, chemotaxis, and the release of pro- or anti-inflammatory mediators."

This reference supports the involvement of microglia in immune activation, cytokine production, and phagocytosis in Alzheimer's disease.

+ 9 more references

Oxidative Stress

Imbalance between the production of reactive oxygen species and the brain's ability to detoxify them, leading to cellular damage.

Synaptic Failure Convergence Model

Free radical production GO:1903409 ↑ INCREASED Antioxidant defense GO:0006979 ↓ DECREASED

Mitochondria GO:0005739 Cell membranes GO:0005886

Downstream

Synaptic Dysfunction Oxidative neuronal injury is modeled as one convergent stressor that worsens synaptic and network function.

Show evidence (4 references)

PMID:34416493 SUPPORT

"Emerging evidence suggests that accumulated oxidative stress may be one of the key mechanisms causing cognitive aging and neurodegenerative diseases such as Alzheimer's disease (AD)."

The literature supports the role of oxidative stress, including the imbalance between ROS production and antioxidant defenses, in the development of Alzheimer's disease.

PMID:27888001 SUPPORT

"Depending on its level, HNE exerts harmful or protective effects associated with the induction of antioxidant defense mechanisms. These effects make HNE a key player in maintaining redox homeostasis, as well as producing imbalances in this system that participate in aging and the development of..."

This reference supports the involvement of oxidative stress and the imbalance in ROS/RNS in the pathology of neurodegenerative diseases.

PMID:25772897 SUPPORT

"Key elements driving neurodegeneration include microglia activation, chronic oxidative injury, accumulation of mitochondrial damage in axons, and age-related iron accumulation in the human brain."

This reference supports the involvement of oxidative stress and mitochondrial damage in neurodegeneration but does not specifically focus on Alzheimer's disease.

+ 1 more reference

Mitochondrial Quality-Control Failure

Mechanism confidence: Provisional

Impaired mitophagy and accumulation of damaged mitochondria are modeled as a mitochondrial stress layer that can feed into amyloid-beta accumulation, tau hyperphosphorylation, neuroinflammation, and cognitive decline.

Amyloid Cascade Model Autophagy-Lysosomal Clearance Model Autophagy-Senescence-Inflammasome (ASI) Axis Neuroimmune-Glial Amplification Model Synaptic Failure Convergence Model Tau Neurodegeneration Model

Neurons CL:0000540 Microglia CL:0000129

Mitophagy GO:0000422 ↓ DECREASED

Mitochondria GO:0005739

Hippocampus Cerebral Cortex

Downstream

Amyloid Plaque Formation Impaired mitochondrial quality control is modeled as a contributor to amyloid-beta accumulation through reduced mitophagy-linked microglial plaque clearance and increased neuroinflammatory stress.

Neurofibrillary Tangle Formation Impaired mitophagy is modeled as a contributor to tau hyperphosphorylation and tau pathology, but its position relative to amyloid and glial activation remains unresolved.

Memory Loss Mitochondrial quality-control failure is modeled as one upstream stressor that can worsen cognitive outcomes, but cognitive resilience may decouple pathology burden from symptoms.

Neuroinflammation Impaired mitochondrial quality control is modeled as licensing NLRP3 inflammasome activation through cytosolic mitochondrial DNA release, in addition to reduced mitophagy-mediated resolution of glial activation.

Show evidence (3 references)

PMID:30742114 SUPPORT Human Clinical

"mitophagy is impaired in the hippocampus of AD patients"

Supports impaired mitophagy as a human Alzheimer disease mitochondrial quality-control abnormality.

PMID:30742114 SUPPORT Model Organism

Supports the modeled link from mitochondrial quality-control restoration to reduced amyloid burden, neuroinflammation, and cognitive impairment in an Alzheimer disease mouse model.

PMID:42778763 SUPPORT PRIMARY RESULT Computational

"enhanced excitatory/inhibitory imbalance and mitochondrial dysfunction in cognitively impaired AD donors"

Associates neuronal mitochondrial dysfunction with the cognitive expression of Alzheimer pathology in human cortex. In the full text the signal is upregulation of the mitochondrial transcription repressor MTERF1 in excitatory neurons of cognitively impaired versus resilient donors at matched late Braak stage. This is a transcriptional association from computational phenotype scoring; it does not measure mitophagy.

Vascular Dysfunction

Alterations in cerebral blood flow and blood-brain barrier integrity, contributing to neuronal dysfunction and amyloid accumulation.

Amyloid Cascade Model Synaptic Failure Convergence Model Vascular and Blood-Brain Barrier Clearance Model

Endothelial cells CL:0000115 Pericytes CL:0000669

LRP1 hgnc:6692

Blood-brain barrier regulation GO:1905603 ↕ DYSREGULATED Cerebral blood flow GO:0120275 ↓ DECREASED LRP1-mediated vascular amyloid-beta clearance GO:0097242 ↓ DECREASED

Cerebral blood vessels Blood-brain barrier

Downstream

Amyloid Plaque Formation Blood-brain barrier dysfunction and neurovascular-unit impairment are modeled as contributors to brain amyloid-beta accumulation and impaired clearance.

Synaptic Dysfunction Reduced perfusion and neurovascular-unit injury are modeled as contributors to neuronal and synaptic dysfunction.

Show evidence (5 references)

PMID:28902142 SUPPORT

This reference supports the statement by indicating that blood-brain barrier (BBB) dysfunction contributes to Alzheimer's disease (AD) pathogenesis, including amyloid-beta (Abeta) accumulation and neurovascular impairments.

PMID:26898552 SUPPORT

This reference supports the statement by highlighting the role of cerebral blood flow changes in the early stages of Alzheimer's disease, which is consistent with the described vascular dysfunction.

PMID:36293539 SUPPORT

"Aberrant remodeling of microvesselsis associated with blood-brain barrier breakdown, development of neuroinflammation, inadequate microcirculation in active brain regions, and leads to the dysfunction of the neurovascular unit and progressive neurological deficits."

This reference supports the statement by describing how microvascular remodeling and blood-brain barrier breakdown contribute to neurovascular unit dysfunction and neurological deficits, relevant to Alzheimer's disease.

+ 2 more references

Glymphatic Clearance Failure

Mechanism confidence: Provisional

Failure of the sleep-dependent perivascular (glymphatic) route by which cerebrospinal fluid exchanges with interstitial fluid and carries interstitial amyloid-beta out of the brain. This is a clearance arm distinct from the transcytotic blood-brain-barrier / LRP1 efflux route modeled in Vascular Dysfunction: it operates along perivascular spaces and depends on perivascular polarization of astroglial aquaporin-4, which is lost in aged and Alzheimer brains independently of age. Reduced clearance raises the steady-state interstitial amyloid-beta concentration and so feeds amyloid plaque formation rather than replacing it.

Amyloid Cascade Model Glymphatic Clearance Model

Astrocyte CL:0000127

Amyloid-beta clearance GO:0097242 ↓ DECREASED Cerebrospinal fluid circulation GO:0090660 ↓ DECREASED

Astrocyte end-foot GO:0097450

Perivascular space UBERON:0014930

Downstream

Amyloid Plaque Formation Reduced perivascular clearance raises the steady-state interstitial amyloid-beta concentration, favoring aggregation and plaque deposition.

Show evidence (5 references)

PMID:27893874 SUPPORT Human Clinical

"Perivascular AQP4 localization was significantly associated with AD status independent of age"

Human postmortem evidence that the molecular lesion underlying glymphatic clearance failure is present in Alzheimer disease independently of age.

PMID:29997300 SUPPORT Human Clinical

"Clearance of the tracer substance was delayed in the dementia cohort."

In vivo human intrathecal-tracer evidence of delayed brain clearance in a dementia cohort relative to near-healthy references.

PMID:33004510 SUPPORT Other

"glymphatic failure may constitute a therapeutically targetable final common pathway"

States the convergence claim that motivates modeling glymphatic clearance failure as a shared upstream contributor in the dementias.

+ 2 more references

Autophagy-Lysosomal Dysfunction

Impaired autophagy and lysosomal degradation pathways leading to accumulation of protein aggregates and cellular dysfunction.

Amyloid Cascade Model Autophagy-Lysosomal Clearance Model Tau Neurodegeneration Model

Neurons CL:0000540

Autophagy GO:0006914 ↓ DECREASED Lysosomal degradation GO:0007041 ↓ DECREASED Protein catabolic process GO:0030163 ↓ DECREASED

Hippocampus Cerebral Cortex

Downstream

Amyloid Plaque Formation Impaired autophagy and lysosomal degradation are modeled as contributors to amyloid-beta accumulation through reduced aggregate clearance.

Neurofibrillary Tangle Formation Impaired autophagy and lysosomal degradation are modeled as contributors to tau aggregate accumulation through reduced proteostatic clearance.

Show evidence (2 references)

PMID:20541250 SUPPORT Model Organism

Supports this node's core mechanism by showing that PS1 loss prevents autophagosome clearance through impaired autolysosome acidification and cathepsin activation, with failed v-ATPase V0a1 lysosomal targeting as a causal mechanism.

PMID:42429504 SUPPORT Other

"Dysfunction of the autophagy-lysosomal pathway is believed to be the main factor leading to the accumulation of amyloid-β and tau proteins, which are also pathological features of AD."

Integrative review synthesizing evidence for autophagy-lysosomal dysfunction as a primary driver of amyloid-β and tau accumulation in Alzheimer disease, matching the quoted snippet. (The same review also separately summarizes a potential role for oligodendrocytes, but that is not covered by this snippet.)

HSV-1 Reactivation in RORB+ Glutamatergic Neurons

Mechanism confidence: Provisional

Latent herpes simplex virus 1 (HSV-1) reactivates specifically within RORB+ glutamatergic neurons — a cortical neuronal population selectively vulnerable in Alzheimer's disease. Petabase-scale mining of human sequencing data and single-nucleus analyses detect viral transcripts in this population in HEALTHY (but not pathological) post-mortem brain tissue, consistent with a model in which viral reactivation precedes the progressive loss of this neuronal population during dementia and so is no longer detectable in end-stage disease.

HSV-1 Reactivation Model Synaptic Failure Convergence Model

RORB+ glutamatergic neuron CL:0000679

HSV-1 release from latency GO:0019046 ↑ INCREASED HSV-1 genome replication GO:0019079 ↑ INCREASED

Cerebral Cortex

Downstream

Synaptic Dysfunction HSV-1 reactivation in selectively vulnerable cortical neurons is modeled as a possible upstream contributor to loss of synaptic integrity and cognitive-network failure.

Show evidence (2 references)

PMID:42094473 SUPPORT Human Clinical

The preprint mines >10 petabytes of human sequencing data and uses single-nucleus analyses on post-mortem human brain to localize HSV-1 transcripts specifically to RORB+ glutamatergic neurons, the same cortical population progressively lost in dementia — directly supporting this provisional pathophysiology node.

PMID:42094473 SUPPORT Human Clinical

"identifying recurrent herpes simplex virus 1 (HSV-1) reactivation in healthy but not pathological post-mortem human brain tissue"

Key temporal observation: HSV-1 reactivation is detected in healthy post-mortem brain tissue but is absent from pathological tissue, consistent with a model in which reactivation precedes (rather than follows) the loss of vulnerable neurons.

Intercellular Tau Transmission via Extracellular Vesicles

Mechanism confidence: Provisional

Pathological tau spreads cell to cell in Alzheimer's disease, in part by being packaged into neuronal extracellular vesicles (EVs) and released for uptake by recipient neurons (and microglia), where the delivered tau seeds aggregation of endogenous tau. The activity-regulated, capsid-forming neuronal protein Arc binds tau directly (with higher affinity for phosphorylated tau) and, together with the I-BAR protein IRSp53, drives release of seed-competent tau from dendrites in EVs. Arc and tau are co-packaged in mouse and human brain-derived EVs, and in human AD brain EVs Arc levels correlate with phosphorylated EV-tau. Loss of Arc reduces EV-tau and tau seeding potential and nearly abolishes neuron-to-neuron tau transmission, while causing intracellular tau to accumulate in donor neurons — consistent with EV-tau release being partly protective for the donor but a driver of pathology spread.

Arc-Dependent EV-Mediated Tau Propagation Model Tau Neurodegeneration Model

Neurons CL:0000540 Microglia CL:0000129

ARC hgnc:648 MAPT hgnc:6893 BAIAP2 (IRSp53) hgnc:947

Extracellular vesicle biogenesis GO:0140112 ↑ INCREASED Tau release in extracellular vesicles GO:0009306 ↑ INCREASED Tau seeding of aggregation in recipient neurons GO:1990000 ↑ INCREASED

Neuronal dendrites Synapses Hippocampus Entorhinal Cortex

Downstream

Neurofibrillary Tangle Formation EV-delivered seed-competent tau is taken up by recipient neurons and seeds aggregation of endogenous tau, propagating tangle pathology along connected circuits.

Show evidence (4 references)

PMID:42372723 SUPPORT In Vitro

"We find that the neuronal gene Arc is critical for the release of tau in neuronal extracellular vesicles (EVs) via a direct protein-protein interaction."

Establishes Arc as a required mediator of tau release in neuronal EVs acting through a direct Arc-tau protein-protein interaction, demonstrated in neuronal cultures and purified-protein binding assays.

PMID:42372723 SUPPORT Human Clinical

"Both Arc and tau are co-packaged in mouse and human brain-derived EVs."

Arc-tau co-packaging is observed in human (and mouse) brain-derived EVs, supporting the EV route of tau spread in human tissue.

PMID:42372723 SUPPORT Human Clinical

"Moreover, Arc levels in brain-derived EVs isolated from human Alzheimer's disease (AD) brains show a strong positive correlation with phosphorylated EV-tau levels."

In human AD brain EVs, Arc abundance tracks phosphorylated EV-tau, connecting Arc-dependent EV packaging to pathological tau in human disease.

+ 1 more reference

PARP1-Mediated Parthanatos

Mechanism confidence: Provisional

PARP1 (poly(ADP-ribose) polymerase 1) plays a dual role in Alzheimer's disease: physiological nuclear PARP1 activity is essential for DNA repair and memory consolidation, but overactivity triggered by oxidative stress and amyloid-beta-induced DNA damage leads to excessive NAD+ and ATP consumption. PARP1 overactivation causes NAD+ depletion, triggering AIF (apoptosis-inducing factor) release and formation of AIF-MIF complexes that drive parthanatos, a caspase-independent programmed cell-death mechanism. This PARP1-driven pathway converges with neuroinflammation (NF-κB pathway activation), mitophagy dysregulation, and disruption of SIRT1-mediated neuroprotection, collectively contributing to neuronal death and disease progression.

Autophagy-Lysosomal Clearance Model Neuroimmune-Glial Amplification Model

Neurons CL:0000540 Microglia CL:0000129

PARP1 hgnc:270 SIRT1 hgnc:14929

DNA damage response GO:0006974 ⚠ ABNORMAL Neuroinflammation GO:0150076 ↑ INCREASED

Downstream

Neuroinflammation PARP1 overactivation amplifies NF-κB-dependent neuroinflammatory responses and promotes microglial activation.

Oxidative Stress PARP1 hyperactivation and NAD+ depletion impair mitochondrial redox homeostasis, sustaining oxidative stress through disruption of antioxidant SIRT1 and mitophagy.

Mitochondrial Quality-Control Failure PARP1 hyperactivity and NAD+ depletion impair mitochondrial dynamics, mitophagy, and ATP production, contributing to quality-control collapse.

Show evidence (1 reference)

PMID:42413719 SUPPORT Other

"PARP1 exhibits context-dependent duality: its physiological nuclear expression in hippocampus neurons is essential for memory consolidation and decreases early in cognitive impairment, suggesting a correlative association with synaptic malfunction. In contrast, overactivity of PARP1 resulting..."

Directly supports PARP1-mediated parthanatos as a convergence mechanism in Alzheimer's disease, linking Aβ-induced oxidative stress to NAD+ depletion, AIF-driven programmed necrosis, neuroinflammation amplification, and loss of mitochondrial neuroprotection.

Adaptive Immune T Cell Response to Tau Pathology

Mechanism confidence: Provisional

Downstream of tau pathology and microglial activation, clonally expanded T cells — predominantly CD8+ cytotoxic T cells — infiltrate the brain in a spatiotemporal pattern that tracks tau (not amyloid) pathology and correlates with neuronal loss. Microglia are required to recruit these T cells; once in the parenchyma the cells shift from activated toward exhausted states with restricted TCR clonality, and IFN-gamma/PD-1 signaling amplifies neurodegeneration. Depletion of T cells (or of microglia) is neuroprotective in tauopathy models, and IFN-gamma or PD-1 blockade reduces brain atrophy, identifying the adaptive immune response as an active, druggable contributor to tau-driven neurodegeneration rather than a bystander.

Adaptive Immune (T Cell) Neurodegeneration Model Synaptic Failure Convergence Model

CD8+ cytotoxic T cell CL:0000625 Microglia CL:0000129

T cell chemotaxis into the brain GO:0010818 ↑ INCREASED T cell mediated cytotoxicity GO:0001913 ↑ INCREASED Interferon-gamma (type II interferon) production GO:0032609 ↑ INCREASED

Hippocampus Entorhinal Cortex

Downstream

Synaptic Dysfunction T cell-derived IFN-gamma and cytotoxic activity, engaged downstream of tau pathology, amplify neuronal and synaptic loss; IFN-gamma/PD-1 blockade reduces brain atrophy.

Show evidence (7 references)

PMID:36890231 SUPPORT Model Organism

Establishes a tauopathy-specific adaptive immune response and shows that depleting T cells (or microglia) blocks tau-mediated neurodegeneration.

PMID:36890231 SUPPORT Model Organism

"Numbers of T cells, especially those of cytotoxic T cells, were markedly increased in areas with tau pathology in mice with tauopathy"

Cytotoxic (CD8+) T cells accumulate specifically where tau pathology is present in mouse tauopathy, anchoring the cell type of this node (human AD brain accumulation is captured separately by the human PMID:31915375 item).

PMID:36890231 SUPPORT Model Organism

"T cell numbers correlated with the extent of neuronal loss, and the cells dynamically transformed their cellular characteristics from activated to exhausted states along with unique TCR clonal expansion."

Links T cell burden to neuronal loss and documents the activated-to-exhausted transition with clonal TCR expansion characteristic of an antigen-specific response.

+ 4 more references

Muscarinic M1 Receptor Signaling Loss

Loss of signaling through the M1 muscarinic acetylcholine receptor, arising either from degeneration of basal forebrain cholinergic input or from sustained pharmacologic antagonism by anticholinergic drugs. M1 signaling tonically favors non-amyloidogenic ADAM17-mediated alpha-secretase cleavage of APP and restrains GSK3-beta; withdrawing it shifts APP processing toward the amyloidogenic route and permits tau hyperphosphorylation.

Anticholinergic Muscarinic Burden Model

Neurons CL:0000540 Cholinergic neuron CL:0000108

CHRM1 hgnc:1950 ADAM17 hgnc:195 GSK3B hgnc:4617

M1 muscarinic (Gq/PLC-coupled) receptor signaling GO:0007207 ↓ DECREASED Amyloidogenic APP processing GO:0034205 ↑ INCREASED

M1 muscarinic acetylcholine receptor activity GO:0016907 ↓ DECREASED

Hippocampus Cerebral Cortex Basal forebrain

Downstream

Amyloid Plaque Formation Withdrawal of M1 signaling shifts APP processing away from the ADAM17/alpha-secretase route toward amyloidogenic cleavage, increasing amyloid-beta available for deposition.

Neurofibrillary Tangle Formation Loss of M1 signaling de-represses GSK3-beta and alters protein kinase C activity, promoting tau hyperphosphorylation and tangle formation.

Show evidence (1 reference)

PMID:16504943 SUPPORT Model Organism

"We further demonstrate that administration of dicyclomine, an M1 antagonist, exacerbates the Abeta and tau pathologies."

Establishes that pharmacologic M1 blockade, not only genetic ablation, worsens AD-like pathology.

Complement-Mediated Microglial Synapse Elimination

Mechanism confidence: Provisional

Microglia actively engulf and remove structurally intact synapses after those synapses are tagged by the classical complement cascade. C1q binds perisynaptic membranes, C3 opsonizes the synapse, and microglia engage it through complement receptor 3 (CR3, the ITGAM/ITGB2 heterodimer). This is the same pathway that prunes surplus synapses during normal development, reactivated in the adult brain. Both soluble amyloid-beta oligomers and phosphorylated tau drive it, so it is a point at which the two proteinopathies converge on one cellular execution step. Because synapse loss is the pathological change that correlates best with cognitive decline, this node supplies the cellular mechanism that the Synaptic Dysfunction node otherwise leaves implicit.

Complement-Dependent Microglial Synapse Elimination Model Neuroimmune-Glial Amplification Model Synaptic Failure Convergence Model

Microglia CL:0000129 Astrocyte CL:0000127 Neurons CL:0000540

C1QA hgnc:1241 C3 hgnc:1318 ITGAM (CR3) hgnc:6149 MFGE8 hgnc:7036

Complement-mediated synapse pruning GO:0150062 ↑ INCREASED Complement activation GO:0006956 ↑ INCREASED Microglial engulfment of synaptic material GO:0006909 ↑ INCREASED

Hippocampus Synapses

Downstream

Synaptic Dysfunction Complement tagging and microglial engulfment remove intact synapses, producing the synapse loss that underlies impaired plasticity and neurotransmission.

Show evidence (4 references)

PMID:27033548 SUPPORT Model Organism

"C1q is necessary for the toxic effects of soluble β-amyloid (Aβ) oligomers on synapses and hippocampal long-term potentiation."

Places C1q as a required intermediate between soluble amyloid-beta oligomers and both synaptic and electrophysiological damage.

PMID:27033548 SUPPORT Model Organism

"Finally, microglia in adult brains engulf synaptic material in a CR3-dependent process when exposed to soluble Aβ oligomers."

Demonstrates the engulfment step itself in the adult brain and shows it requires the microglial complement receptor.

PMID:30392797 SUPPORT Model Organism

"A C1q-blocking antibody inhibited microglial synapse removal in cultured neurons and in Tau-P301S mice, rescuing synapse density."

Pharmacological rescue of synapse density by blocking C1q, in a tau model — the basis for treating this as a therapeutic target rather than a description.

+ 1 more reference

Neuronal Endosomal-Lysosomal Acidification Failure

Mechanism confidence: Provisional

Activation and enlargement of Rab5-positive early endosomes is the earliest recognized intraneuronal change in sporadic Alzheimer disease, present in neocortical pyramidal neurons at preclinical stages and, in Down syndrome, decades before classical neuropathology. The APP beta-C-terminal fragment (APP-βCTF) — not amyloid-beta itself — binds the v-ATPase V0a1 subunit and competitively blocks assembly of the V1 subcomplex, so autolysosomes fail to acidify and cathepsins are not activated. Amyloid-beta then accumulates inside enlarged, de-acidified autolysosomes; in the most compromised neurons these vacuoles pack into perikaryal rosettes (the PANTHOS pattern) before lysosomal membrane permeabilization kills the cell. This node therefore models amyloid pathology as beginning inside the neuron.

Amyloid Cascade Model Autophagy-Lysosomal Clearance Model Endolysosomal Origin ("Inside-Out" Amyloid) Model

Neurons CL:0000540 Pyramidal neuron CL:0000598

APP hgnc:620 PSEN1 hgnc:9508

Lysosomal lumen acidification GO:0007042 ↓ DECREASED Endocytic pathway activation GO:0006897 ↑ INCREASED

Enlarged Rab5-positive early endosome GO:0005769

Cerebral Cortex Hippocampus

Downstream

Amyloid Plaque Formation Amyloid-beta accumulated within de-acidified autolysosomes of individual neurons is released when those neurons rupture, so the plaque is modeled as originating from a dying neuron rather than from secreted peptide.

Autophagy-Lysosomal Dysfunction APP-βCTF inhibition of the v-ATPase is modeled as one upstream cause of the broader autophagic-flux failure curated in that node.

Show evidence (6 references)

PMID:10880397 SUPPORT Human Clinical

"These results identify EP activation as the earliest known intraneuronal change to occur in sporadic AD, the most common form of AD."

Human postmortem study positioning endocytic-pathway activation as the earliest intraneuronal lesion in sporadic Alzheimer disease.

PMID:10880397 SUPPORT Human Clinical

"In Down syndrome, early endosomes were significantly enlarged in some pyramidal neurons as early as 28 weeks of gestation, decades before classical AD neuropathology develops."

Establishes how far the endosomal lesion precedes classical neuropathology in the APP-trisomic human condition.

PMID:15465622 SUPPORT Human Clinical

"We found by ELISA and immunocytochemistry that the appearance of enlarged endosomes coincided with an initial rise in soluble Abeta40 and Abeta42 peptides, which preceded amyloid deposition."

Ties the endosomal lesion in human brain to the earliest rise in soluble amyloid-beta, before deposition.

+ 3 more references

Selective Vulnerability of RORB+ Entorhinal Excitatory Neurons

Mechanism confidence: Provisional

Neuronal loss in Alzheimer disease is not uniform. Entorhinal cortex layer II excitatory neurons and locus coeruleus noradrenergic neurons degenerate first and disproportionately, and within the entorhinal cortex the vulnerable excitatory population is marked by RORB. These neurons are selectively susceptible to neurofibrillary inclusion and are depleted as pathology advances, while neighbouring populations in the same tissue are relatively spared. This node models cell-intrinsic vulnerability as a determinant of where the disease starts, which aggregate burden alone does not explain.

Selective Neuronal Vulnerability Model Tau Neurodegeneration Model

RORB+ excitatory neuron CL:0000679 Pyramidal neuron CL:0000598

RORB hgnc:10259

Neurofibrillary tangle assembly in vulnerable neurons GO:1990000 ↑ INCREASED

Entorhinal Cortex Locus coeruleus

Downstream

Neurofibrillary Tangle Formation Cell-intrinsic vulnerability of RORB+ entorhinal excitatory neurons is modeled as determining where neurofibrillary inclusions first form, giving tau pathology its characteristic anatomical starting point.

Show evidence (6 references)

PMID:33432193 SUPPORT Human Clinical

Identifies the vulnerable population molecularly in human postmortem brain and confirms both its depletion and its tangle susceptibility by independent neuropathology.

PMID:8699259 SUPPORT Human Clinical

"These results support the conclusion that a marked decrement of layer II neurons distinguishes even very mild AD from nondemented aging."

Stereological evidence that laminar-selective entorhinal neuron loss is already established in the mildest clinically detectable disease and is not a feature of normal aging.

PMID:27513978 SUPPORT Human Clinical

"As the Braak stage increases by 1 unit, the LC volume decreases by 8.4%."

Quantifies stage-wise loss in a second selectively vulnerable population, the locus coeruleus. Partial for this node because the population is noradrenergic rather than the RORB+ entorhinal one named here.

+ 3 more references

Necroptotic Neuronal Death

Mechanism confidence: Provisional

Neurons die by necroptosis — a caspase-independent, programmed form of necrosis in which RIPK1 and RIPK3 assemble a necrosome in which RIPK3 phosphorylates and activates MLKL, which then permeabilizes the plasma membrane. Activated necrosome components are detectable in granulovacuolar degeneration bodies, a classical Alzheimer lesion that had no mechanistic account, and their regional burden is inversely correlated with neuronal density. The mechanism gives the entry an explicit terminal cell-death step, which the amyloid, tau, and synaptic nodes otherwise leave unstated.

Necroptosis Model of Neuronal Death Neuroimmune-Glial Amplification Model

Neurons CL:0000540

RIPK1 hgnc:10019 RIPK3 hgnc:10021 MLKL hgnc:26617

Necroptotic process GO:0070266 ↑ INCREASED

Hippocampus Cerebral Cortex

Downstream

Neuroinflammation Necroptosis is lytic: unlike apoptosis it releases intracellular contents, so necroptotic neuronal death is modeled as feeding the glial inflammatory response rather than being silently cleared.

Show evidence (6 references)

PMID:28758999 SUPPORT Human Clinical

"We found that necroptosis was activated in postmortem human AD brains, positively correlated with Braak stage, and inversely correlated with brain weight and cognitive scores."

Human postmortem evidence relating necroptosis activation to both pathological stage and clinical severity.

PMID:31802237 SUPPORT Human Clinical

"We demonstrated that all three activated necrosome components can be detected in GVD lesions (GVDn+, i.e. GVD with activated necrosome) in neurons, that they colocalize with classical GVD markers, such as pTDP-43 and CK1δ, and similarly to these markers detect GVD lesions."

Localizes the complete activated necrosome to granulovacuolar degeneration bodies in human neurons, connecting the mechanism to an established lesion.

PMID:31802237 SUPPORT Human Clinical

"GVDn + neurons inversely correlated with neuronal density in the early affected CA1 region of the hippocampus and in the late affected frontal cortex layer III."

Relates necrosome burden to neuronal density in both an early- and a late-affected region, supporting the link to actual neuron loss.

+ 3 more references

Senescent Cell Accumulation

Mechanism confidence: Provisional

Cells that have entered an irreversible cell-cycle arrest but remain metabolically active accumulate in the aging and Alzheimer brain, expressing p16INK4A/CDKN2A and p21/CDKN1A and secreting a proinflammatory senescence-associated secretory phenotype. Clearing them — genetically or with senolytic drugs — prevents gliosis, tau hyperphosphorylation and tangle deposition, reduces amyloid load and neuroinflammation, and preserves cognition in mouse models. Which cell type is senescent is genuinely unresolved: mouse work implicates astrocytes and microglia in tauopathy and oligodendrocyte progenitor cells around plaques, while the largest human survey assigns over 97% of senescent cells to excitatory neurons.

Autophagy-Senescence-Inflammasome (ASI) Axis Cellular Senescence Model Neuroimmune-Glial Amplification Model Tau Neurodegeneration Model

Astrocyte CL:0000127 Microglia CL:0000129 Oligodendrocyte precursor cell CL:0002453 Excitatory neuron CL:0000679

CDKN2A hgnc:1787

Cellular senescence GO:0090398 ↑ INCREASED

Hippocampus Cerebral Cortex

Downstream

Neurofibrillary Tangle Formation Senescent cells are modeled as promoting tau hyperphosphorylation and tangle deposition; the direction is established by the inverse experiment, in which clearing them prevents both.

Neuroinflammation The senescence-associated secretory phenotype is modeled as a chronic proinflammatory input to glial activation; senolytic clearance reduces neuroinflammation.

Show evidence (5 references)

PMID:30232451 SUPPORT Model Organism

"Here we show a causal link between the accumulation of senescent cells and cognition-associated neuronal loss."

Genetic clearance of p16-positive cells as they arise prevents tau pathology and neuronal loss, establishing senescence as causal in this model.

PMID:30936558 SUPPORT Model Organism

"Senolytic treatment of AD mice selectively removed senescent cells from the plaque environment, reduced neuroinflammation, lessened Aβ load, and ameliorated cognitive deficits."

Pharmacological rather than genetic clearance, in an amyloid rather than a tau model, reproducing the same set of benefits.

PMID:30936558 SUPPORT In Vitro

"Direct exposure of cultured OPCs to aggregating Aβ triggered cell senescence."

Identifies aggregating amyloid-beta as a sufficient senescence-inducing stimulus for oligodendrocyte progenitor cells, supplying the upstream trigger.

+ 2 more references

Oligodendrocyte and Myelin Dysfunction

Mechanism confidence: Provisional

Age-related loss of myelin integrity and of oligodendrocyte metabolic support for the axon is modeled as an upstream contributor to amyloid deposition rather than a consequence of it. Myelin damage concentrates the amyloidogenic processing machinery within axonal swellings and increases cleavage of cortical amyloid precursor protein, while simultaneously diverting disease-associated microglia toward myelin debris and away from plaques — so one lesion both raises amyloid production and lowers its clearance. In APOE4 carriers, cholesterol is aberrantly retained in oligodendrocytes and myelination is reduced, giving the strongest genetic risk factor a route through this cell type.

Amyloid Cascade Model Myelin and Oligodendrocyte Dysfunction Model Neuroimmune-Glial Amplification Model

Oligodendrocyte CL:0000128 Oligodendrocyte precursor cell CL:0002453 Microglia CL:0000129

APOE hgnc:613 APP hgnc:620

Myelination GO:0042552 ↓ DECREASED Cholesterol metabolic process in oligodendrocytes GO:0008203 ↕ DYSREGULATED

Cerebral Cortex White matter

Downstream

Amyloid Plaque Formation Myelin dysfunction is modeled as increasing amyloid deposition both directly, by concentrating amyloidogenic processing in axonal swellings, and indirectly, by diverting plaque-clearing microglia to myelin damage.

Show evidence (5 references)

PMID:37258678 SUPPORT Model Organism

"Here we identify genetic pathways of myelin dysfunction and demyelinating injuries as potent drivers of amyloid deposition in mouse models of AD."

Multiple independent myelin-mutant crosses each increase amyloid deposition, establishing direction of causation in the mouse.

PMID:37258678 SUPPORT Model Organism

"Despite successful induction, amyloid disease-associated microglia (DAM) that usually clear amyloid plaques are apparently distracted to nearby myelin damage."

The clearance half of the mechanism: myelin damage competes for the same microglial population that would otherwise remove plaques.

PMID:36385529 SUPPORT Human Clinical

"We show that altered cholesterol localization in the APOE4 brain coincides with reduced myelination."

Human postmortem evidence that oligodendrocyte cholesterol dysregulation and reduced myelination are real features of the APOE4 brain.

+ 2 more references

Glial and Pericyte Mesenchymal Drift

Astrocytes, oligodendrocytes and pericytes in the Alzheimer's disease brain up-regulate the mesenchymal drift gene signature, which persists in age-matched comparisons; neurons, microglia and oligodendrocyte precursor cells did not show it in the same reanalysis. This is evidence at the level of a transcriptional signature only. Whether the drifting glia actually change morphology, deposit matrix, or contribute to disease progression is untested, so the node carries no downstream edges.

astrocyte CL:0000127 oligodendrocyte CL:0000128 brain pericyte CL:2000043

mesenchymal drift (aberrant acquisition of mesenchymal identity) GO:0048762 ⚠ ABNORMAL

Show evidence (2 references)

PMID:42680479 SUPPORT REVIEW SYNTHESIS Human Clinical

"In a reanalysis of published single-nucleus RNA-sequencing data from AD and control brain, Lu et al. found significant upregulation of MD genes in astrocytes, oligodendrocytes, and pericytes, but not in neurons, microglia, or oligodendrocyte progenitor cells (OPCs); this pattern persisted after..."

Review summary of the Alzheimer's single-nucleus result in Lu et al. 2025 (PMID:40816266), naming the three cell types bound here and the three in which the signature was not found. Graded HUMAN_CLINICAL for the human brain data it describes.

PMID:42680479 SUPPORT REVIEW SYNTHESIS Other

"It remains unclear whether transcriptomically drifting astrocytes and microglia undergo true cytoskeletal changes, dismantle their networks, or deposit fibrotic matrices in the brain, as mesenchymal cells do in other tissues."

Records the limit of the evidence: the drift is shown as a signature, not as a physical change in the cells. Evidence source is OTHER because this is the review's own assessment.

Microglial Lipid Droplet Accumulation

Mechanism confidence: Provisional

A distinct microglial state defined by accumulation of cytoplasmic lipid droplets. In the aging brain these lipid-droplet-accumulating microglia are defective in phagocytosis, generate high levels of reactive oxygen species and secrete proinflammatory cytokines. In Alzheimer disease the corresponding human state is marked by expression of the long-chain acyl-CoA synthetase ACSL1 and is most abundant in APOE4 homozygotes; fibrillar amyloid-beta induces it in an APOE-dependent manner, and factors secreted by these microglia phosphorylate tau and are neurotoxic. The state therefore couples the strongest common genetic risk factor to a specific, dysfunctional glial phenotype.

APOE Risk-Modulation Model Neuroimmune-Glial Amplification Model Tau Neurodegeneration Model

Microglia CL:0000129

ACSL1 hgnc:3569 APOE hgnc:613

Lipid droplet organization GO:0034389 ↑ INCREASED Microglial phagocytosis GO:0006909 ↓ DECREASED

Lipid droplet GO:0005811

Hippocampus Cerebral Cortex

Downstream

Neurofibrillary Tangle Formation Factors secreted by lipid-droplet-containing microglia drive tau phosphorylation and neurotoxicity in an APOE-dependent manner, linking the glial lipid state to tau pathology.

Neuroinflammation Lipid-droplet-accumulating microglia are a proinflammatory, phagocytically defective state, so their accumulation is modeled as amplifying neuroinflammation while reducing amyloid clearance capacity.

Show evidence (3 references)

PMID:38480892 SUPPORT Human Clinical

"Through single-nucleus RNA sequencing of brain tissue in Alzheimer's disease, we have identified a microglial state defined by the expression of the lipid droplet-associated enzyme ACSL1 with ACSL1-positive microglia being most abundant in patients with Alzheimer's disease having the APOE4/4 genotype."

Human postmortem single-nucleus evidence defining the state and tying its abundance to APOE4 homozygosity.

PMID:38480892 SUPPORT In Vitro

"In human induced pluripotent stem cell-derived microglia, fibrillar Aβ induces ACSL1 expression, triglyceride synthesis and lipid droplet accumulation in an APOE-dependent manner."

Establishes fibrillar amyloid-beta as the inducing stimulus and APOE as required, in human microglia.

PMID:31959936 SUPPORT Model Organism

"These cells, which we call 'lipid-droplet-accumulating microglia' (LDAM), are defective in phagocytosis, produce high levels of reactive oxygen species and secrete proinflammatory cytokines."

Characterizes the functional deficits of the state. Partial for this node because the work describes the aging brain and makes no Alzheimer-specific claim.

Interneuron Dysfunction and Network Hypersynchrony

Mechanism confidence: Provisional

Amyloid-beta impairs parvalbumin-expressing inhibitory interneurons, in part through reduced levels of Nav1.1 (SCN1A), the voltage-gated sodium channel subunit these cells depend on to fire at high frequency. Loss of parvalbumin-cell output degrades gamma oscillations and releases the network into hypersynchrony, producing epileptiform activity. In patients this appears as subclinical epileptiform discharges detectable in a large minority of people with Alzheimer disease who have never had a clinical seizure, and their presence tracks faster cognitive decline. The node models cognitive impairment as arising partly from failure of inhibition, not only from loss of excitatory synapses.

Network Hyperexcitability and Interneuron Dysfunction Model Synaptic Failure Convergence Model

Parvalbumin-expressing inhibitory interneuron CL:4023018 GABAergic neuron CL:0000617

SCN1A hgnc:10585

Inhibitory synaptic transmission GO:0051932 ↓ DECREASED

Nav1.1 voltage-gated sodium channel activity GO:0005248 ↓ DECREASED

Hippocampus Cerebral Cortex

Downstream

Memory Loss Loss of inhibitory control and network hypersynchrony are modeled as direct contributors to cognitive impairment; restoring Nav1.1 in a model reverses both the hypersynchrony and the memory deficit.

Show evidence (5 references)

PMID:22541439 SUPPORT Model Organism

"Electroencephalographic recordings in hAPP mice revealed spontaneous epileptiform discharges, indicating network hypersynchrony, primarily during reduced gamma oscillatory activity."

Establishes the electrophysiological phenotype and its coupling to reduced gamma activity in an amyloid model.

PMID:22541439 SUPPORT Human Clinical

"Supporting this hypothesis, hAPP mice and AD patients had decreased levels of the interneuron-specific and PV cell-predominant voltage-gated sodium channel subunit Nav1.1."

Contains the human limb of the Nav1.1 claim. Graded PARTIAL because the human observation is a protein-level measurement reported alongside the mouse result rather than a separate human study.

PMID:27696483 SUPPORT Human Clinical

"Subclinical epileptiform activity was detected in 42.4% of AD patients and 10.5% of controls (p = 0.02)."

Prospective blinded extended EEG and MEG monitoring quantifying the predicted hyperexcitability in patients with no seizure history.

+ 2 more references

Tau-Induced Nucleocytoplasmic Transport Failure

Mechanism confidence: Provisional

Pathological tau binds nucleoporins of the nuclear pore complex and disrupts their structural and functional integrity, impairing nuclear import and export in both tau-transgenic mouse brain and human Alzheimer brain tissue. Nup98 mislocalizes from the nuclear envelope into the cytoplasm of tangle-bearing neurons and, once there, accelerates tau aggregation in vitro — a feed-forward arrangement in which the transport lesion promotes the aggregation that caused it. The same Nup98 mislocalization is found across primary tauopathies and tracks regional tau burden.

Tau Neurodegeneration Model

Neurons CL:0000540

MAPT hgnc:6893 NUP98 hgnc:8068

Nucleocytoplasmic transport GO:0006913 ↓ DECREASED

Nuclear pore complex GO:0005643

Cerebral Cortex Hippocampus

Downstream

Neurofibrillary Tangle Formation Nup98 mislocalized into the neuronal cytoplasm facilitates tau aggregation, closing a feed-forward loop in which tau-driven transport failure promotes further tau aggregation.

Show evidence (3 references)

PMID:30189209 SUPPORT In Vitro

"Here, we show that tau can directly interact with nucleoporins of the nuclear pore complex (NPC) and affect their structural and functional integrity."

Establishes the direct physical interaction between tau and nuclear pore complex components.

PMID:30189209 SUPPORT Human Clinical

"Pathological tau impairs nuclear import and export in tau-overexpressing transgenic mice and in human AD brain tissue."

Demonstrates the functional transport deficit in human Alzheimer brain tissue as well as in the mouse model.

PMID:36601621 SUPPORT Human Clinical

"In the occipital cortex, which is relatively spared from pathological tau accumulations in these primary tauopathies, the localization of nucleoporin 98 was not significantly altered."

Regional internal control in human tissue: the nucleoporin lesion is absent where tau pathology is absent, tying it specifically to tau burden rather than to neurodegeneration generally.

Brain N-Glycan Hyperglycosylation

Mechanism confidence: Provisional

Net abundance of protein N-linked glycans rises in the Alzheimer brain, across grey and white matter in human frontal cortex and most markedly in cortex, hippocampus and thalamus in amyloid (5xFAD) and tau (PS19) mouse models. In human grey matter the increase tracks Braak stage, whereas in white matter it appears early and does not progress. Stable-isotope pulse-chase tracing in 5xFAD mice attributes it to increased glycan biosynthesis rather than impaired lysosomal degradation or salvage, with more label entering the nucleotide-sugar donors UDP-hexose and UDP-HexNAc. The extra UDP-GlcNAc is channelled selectively into N-glycosylation: O-GlcNAcylation and hyaluronan show a mild decrease in the same mouse brains, and the free N-glycan precursors N-acetylglucosamine and glucosamine-6-phosphate are depleted in human AD grey matter, as expected if they are being consumed. Lowering the flux improves memory in mice without changing amyloid-beta or GFAP.

N-Glycan Hyperglycosylation Metabolic Driver Model

Neuron CL:0000540

PGM3 hgnc:8907

Protein N-linked glycosylation GO:0006487 ↑ INCREASED UDP-N-acetylglucosamine biosynthesis (hexosamine pathway flux) GO:0006048 ↑ INCREASED Protein O-GlcNAcylation GO:7770074 ↓ DECREASED Hyaluronan biosynthesis GO:0030213 ↓ DECREASED

Cerebral Cortex UBERON:0000956 Hippocampus UBERON:0002421 Thalamus UBERON:0001897

Downstream

Memory Loss Reducing N-glycan biosynthesis restores social recognition memory in 5xFAD and PS19 mice, and increasing it with oral glucosamine worsens memory in 5xFAD mice, without changes in amyloid-beta or astrocytic GFAP. The intermediate steps between excess glycosylation and impaired memory are not identified; the glycoproteomic data point to neuronal membrane proteins but no specific glycoprotein has been shown to carry the effect.

Show evidence (9 references)

PMID:42265388 SUPPORT Human Clinical

"However, spatial glycomics analysis uncovered a particularly striking pattern: glycan abundance was dramatically elevated across both white and grey matter regions in AD samples"

Post-mortem human frontal cortex, three AD cases and three age-, sex- and post-mortem-interval-matched controls, imaged by MALDI spatial glycomics.

PMID:42265388 SUPPORT Human Clinical

"In grey matter, a heat map of glycan abundance demonstrated a progressive increase in glycosylation across Braak stages, with the most pronounced changes observed in later stages"

Stage dependence in a second, Braak-stratified human series: the grey matter increase grows with pathological severity.

PMID:42265388 SUPPORT Human Clinical

"suggesting that hyperglycosylation in white matter may be an early but non-progressive event."

Qualifies the regional picture: the white matter increase is transient, so the progressive signal is a grey matter one.

+ 6 more references

Micro-MKKS63 Microprotein Depletion

Mechanism confidence: Provisional

The MKKS locus in human dorsolateral prefrontal cortex translates two products, and the dominant one is not the annotated one. A 63-amino-acid microprotein encoded by a small ORF at the locus - designated micro-MKKS63, catalogued only in unreviewed TrEMBL as Q9HB66 - is the predominant detectable protein product there, yielding over a thousand peptide-spectrum matches across the ROSMAP tandem-mass-tag cohort, while no peptide from the canonical 570-amino-acid MKKS protein (Q9NPJ1, the McKusick-Kaufman syndrome entry) was recovered at all. The same imbalance holds in a search of a public meta-database of 66,700 mass spectrometry runs, where canonical MKKS accounts for under 0.1% of the locus's peptide-spectrum matches. Micro-MKKS63 abundance is significantly lower in symptomatic Alzheimer disease than in non-Alzheimer donors (111 versus 187 donors, log2 fold change -0.063, Cohen's d -0.596, FDR 2.42 x 10-5), and RNA reads mapping uniquely to its smORF coding sequence anchor a coexpression network enriched for mitochondrial biology across all three Gene Ontology domains - whereas the network anchored to the canonical ORF shows no enrichment in either condition. The depletion is therefore a microprotein-level change at a locus whose annotated product carries none of the signal, which is why no previous Alzheimer proteomic study could have found it.

smORF-Encoded Microprotein Model

MKKS locus (micro-MKKS63 smORF) hgnc:7108

Mitochondrion GO:0005739 Mitochondrial outer membrane GO:0005741

Dorsolateral Prefrontal Cortex

Downstream

Microglial Oxidative Phosphorylation Failure Loss of the microprotein lowers microglial oxidative phosphorylation. Curated as DIRECT because the knockout removes the microprotein itself and the respiratory deficit follows, with no intermediate proposed - but the demonstration is in an immortalized human microglial line, not in cortex, so the edge is a cell-biological result imported into a human pathograph rather than a human observation.

Show evidence (5 references)

PMID:42736449 SUPPORT DIRECT PRIMARY RESULT Human Clinical

"the predominant detectable protein is a 63-aa smORF-encoded MP cataloged in unreviewed TrEMBL (Q9HB66), which we designated micro-MKKS63."

Identifies the microprotein this node is about and establishes that it, rather than the annotated protein, is what the locus predominantly yields as protein in human cortex.

PMID:42736449 SUPPORT DIRECT PRIMARY RESULT Human Clinical

"No peptides from the canonical 570-aa MKKS sequence were detected across the ROSMAP TMT-MS samples, while micro-MKKS63 yielded over 1,000 PSMs"

Quantifies the translational imbalance at the locus. The absence of canonical peptides is what makes the microprotein the only thing at this locus an Alzheimer proteomic study could have been measuring.

PMID:42736449 SUPPORT DIRECT PRIMARY RESULT Human Clinical

"TMT quantitation showed significantly decreased micro-MKKS63 abundance in sym-AD versus non-AD"

The depletion itself, measured in human postmortem cortex across the ROSMAP diagnostic contrast.

+ 2 more references

Microglial Oxidative Phosphorylation Failure

Mechanism confidence: Hypothetical

Human microglia lacking micro-MKKS63 respire less. Two independent polyclonal CRISPR-Cas9 knockout HMC3 lines, each using single-guide RNAs that disrupt the smORF while leaving the main MKKS ORF intact, showed significant reductions in basal oxygen consumption rate, ATP-linked oxygen consumption and maximal respiration against non-targeting controls on the Seahorse XF Mito Stress Test, with both lines reducing oxygen consumption independently. The node records the bioenergetic deficit itself. Whether microglia in the Alzheimer brain carry it, and whether it contributes to the disease rather than accompanying it, are both open - which is why the node is curated as HYPOTHETICAL and left without downstream edges into the established Alzheimer chain.

Microglia CL:0000129

Oxidative phosphorylation GO:0006119 ↓ DECREASED Mitochondrial ATP synthesis GO:0042776 ↓ DECREASED

Mitochondrion GO:0005739

Experimental models

HMC3 human microglial micro-MKKS63 CRISPR-Cas9 knockout lines CELL LINE Removes the microprotein and measures respiration, establishing that micro-MKKS63 is required for normal oxidative phosphorylation in human microglial cells.

Show evidence (1 reference)

PMID:42736449 SUPPORT DIRECT PRIMARY RESULT In Vitro

"with consistent effects across two independent sgRNA lines, supporting a role for micro-MKKS63 in microglial oxidative phosphorylation."

Establishes the node and its reproducibility: the respiratory deficit appears in two separately derived knockout lines, which is what distinguishes it from a single-clone artifact.

⬡

Pathograph

Use the checkboxes to hide or show graph categories. Hover nodes for evidence and cross-linked metadata.

●

Phenotypes

Memory Loss VERY_FREQUENT Cognitive HP:0002354

The earliest and most prominent symptom.

Show evidence (1 reference)

PMID:31724515 SUPPORT

"Memory impairment has been considered as one of the earliest clinical hallmarks of Alzheimer's disease."

This reference supports the statement that memory loss is a very frequent and early diagnostic symptom of Alzheimer's disease.

Executive Dysfunction VERY_FREQUENT Cognitive HP:0033051

Difficulty in planning, decision-making, and judgment.

Show evidence (1 reference)

PMID:24011643 SUPPORT

"Impairment of executive functions is common in neurodegenerative disorders such as Alzheimer's disease."

The literature states that executive dysfunction, which includes difficulties in planning, decision-making, and judgment, is common in Alzheimer's disease.

Behavioral Changes FREQUENT Psychiatric HP:0000708

Includes irritability, depression, and apathy.

Show evidence (2 references)

PMID:38157881 SUPPORT

"Neuropsychiatric or behavioral symptoms of dementia encompass a series of disorders, such as anxiety, depression, apathy, psychosis, and agitation, all commonly present in individuals living with dementia."

The abstract confirms that behavioral changes, including irritability, depression, and apathy, are common in Alzheimer's disease.

PMID:28413709 SUPPORT

"Using the median split approach, greater apathy and lower depression were associated with poorer awareness on the Self-Consciousness Scale (respectively: odds ratio ... = 4.8, p = .03; OR = 4.84, p = .04), and the PCRS (only apathy: OR = 9.3, p = .003)."

This study indicates that apathy and depression are significant behavioral symptoms in Alzheimer's disease.

Aphasia FREQUENT Neurologic HP:0002381

Difficulty with speech and understanding language.

Show evidence (1 reference)

PMID:24035593 SUPPORT

"Language disorders of degenerative origin are frequently tied to Alzheimer disease (AD) the different variants of which can result in primary and secondary aphasia syndromes."

The literature clearly supports that aphasia, which involves difficulty with speech and understanding language, is a frequent neurologic symptom in Alzheimer's Disease.

Apraxia OCCASIONAL Neurologic HP:0002186

Difficulty with motor tasks despite intact motor function.

Show evidence (1 reference)

PMID:36375032 SUPPORT

"Limb apraxia is a common early sign of Alzheimer's disease (AD) and is thought to occur specifically in early-onset (before the age of 65) AD."

The literature indicates that limb apraxia is a common early sign of Alzheimer's disease, supporting the statement that apraxia can occur in Alzheimer's disease.

Agnosia OCCASIONAL Neurologic HP:0010524

Inability to recognize objects or people.

Show evidence (1 reference)

PMID:31449049 SUPPORT Human Clinical

"both low-level and high-level visual perception disorders seem quite common in Alzheimer's disease, including, on a low-level, loss of visual field, decreased acuity and contrast sensitivity, and impaired color vision, and on a high-level, impaired color vision, motion perception, visuospatial..."

Review of visual-perceptual deficits in Alzheimer's disease reporting that high-level disorders including object agnosia and prosopagnosia are common, directly supporting agnosia as a recognized (occasional) neurologic feature of the disease.

Seizures OCCASIONAL Neurologic HP:0001250

Usually non-motor and therefore easily missed. Clinically overt epilepsy is the visible fraction of a much larger network-hyperexcitability problem: extended EEG and MEG monitoring detects subclinical epileptiform activity in roughly 42% of patients with no seizure history, against about 10% of controls (PMID:27696483). Mechanistically linked to the Interneuron Dysfunction and Network Hypersynchrony node.

Show evidence (3 references)

PMID:23835471 SUPPORT Human Clinical

"Epilepsies were most often complex partial seizures (47%) and more than half were nonconvulsive (55%)."

Characterizes the seizure semiology in Alzheimer disease as predominantly complex partial and nonconvulsive, which is why the phenotype is under-recognized clinically.

PMID:23835471 SUPPORT Human Clinical

"Patients with AD who had epilepsy presented with cognitive decline 5.5 years earlier than patients with AD who did not have epilepsy (64.8 vs 70.3 years; P = .001)."

Associates the presence of epilepsy with earlier onset of cognitive decline. The cohort is retrospective and referral-based, so the interval should not be read as a population estimate.

PMID:27696483 SUPPORT Human Clinical

"Extended monitoring detects subclinical epileptiform activity in a substantial proportion of patients with AD."

Establishes that the epileptiform phenotype is far more common than overt seizures once monitoring is sensitive enough to detect it.

🧬

Genetic Associations

APP (Genetic Mutation)

Gene: APP hgnc:620 gene page

Show evidence (6 references)

PMID:1365885 SUPPORT

"Four mutations involving amino acid substitutions in exons 16 and 17 of the amyloid precursor protein (APP) gene, have been identified which co-segregate with the disease in some families multiply affected by early onset Alzheimer's disease."

This reference supports the association of APP genetic mutations with early-onset Alzheimer's disease.

PMID:33789815 SUPPORT

"This is the first case demonstrating that a low-degree APP gene-dose increase suffices to cause EOAD with prominent amyloid-beta/tau pathology."

This reference supports the role of APP gene mutations in causing early-onset Alzheimer's disease.

PMID:36306459 SUPPORT

"Numerous mutations in amyloid precursor protein (APP) and presenilin 1 and 2 (PSEN1 and PSEN2) have been identified for EOAD, but they can only account for a small proportion of EOAD cases."

This reference supports the association of APP genetic mutations with early-onset Alzheimer's disease.

+ 3 more references

PSEN1 (Genetic Mutation)

Gene: PSEN1 hgnc:9508 gene page

Show evidence (6 references)

PMID:31296348 SUPPORT

"Presenilin 1 (PSEN1) gene mutations are the major known genetic cause of early-onset Alzheimer's disease."

The study reports a novel PSEN1 mutation associated with early-onset Alzheimer's disease, supporting the genetic association.

PMID:35487021 SUPPORT

"At present, three early-onset AD genes (APP, PSEN1, PSEN2) and one late-onset AD susceptibility gene apolipoprotein E (APOE) have been determined."

The reference confirms that PSEN1 is one of the genes associated with early-onset Alzheimer's disease.

PMID:36951251 SUPPORT

"A genetic association study was conducted to examine ADAD AAO in 340 individuals with the PSEN1 E280A mutation."

The study discusses a large cohort with a specific PSEN1 mutation associated with early-onset Alzheimer's disease, reinforcing the genetic link.

+ 3 more references

PSEN2 (Genetic Mutation)

Gene: PSEN2 hgnc:9509 gene page

Show evidence (3 references)

PMID:35491795 SUPPORT

"The established causative mutations in the APP, PSEN1, and PSEN2 can explain less than 1%, Alzheimer's disease (AD) patients. Of the identified variants, the PSEN2 mutations are even less common."

The reference confirms that PSEN2 mutations are among the causative factors for early-onset Alzheimer's disease.

PMID:36701017 SUPPORT

"Early-onset AD (EOAD) was defined as AD occurring before age 65. Although it has a high genetic risk, EOAD due to PSEN2 variation is very rare."

The reference supports the association of PSEN2 with early-onset Alzheimer's disease, although it notes that such cases are rare.

PMID:32741831 SUPPORT

"Early-onset familial Alzheimer disease (EOFAD) is caused by heterozygous variants in the presenilin 1 (PSEN1), presenilin 2 (PSEN2), and APP genes."

The reference explicitly states that early-onset familial Alzheimer's disease can be caused by PSEN2 mutations.

APOE (Risk Factor)

Gene: APOE hgnc:613 gene page

Show evidence (4 references)

PMID:28959956 SUPPORT Model Organism

"APOE4 is the strongest genetic risk factor for late-onset Alzheimer disease."

Establishes APOE4 as the leading genetic risk factor for late-onset AD.

PMID:31686034 SUPPORT Human Clinical

"The individual had two copies of the APOE3 Christchurch (R136S) mutation, unusually high brain amyloid levels and limited tau and neurodegenerative measurements."

The homozygous APOE3 Christchurch (R136S) variant is associated with resistance to autosomal-dominant AD despite high amyloid, evidence for a protective APOE variant acting on tau/neurodegeneration.

PMID:34586832 SUPPORT Model Organism

"Mice expressing APOE3-Jac have reduced amyloid pathology, plaque-associated immune responses, and neuritic dystrophy."

The APOE3-Jacksonville (V236E) variant reduces APOE aggregation and amyloid pathology, a second protective APOE variant supporting APOE aggregation as a therapeutic target.

+ 1 more reference

BIN1 (Risk Factor)

Gene: BIN1 hgnc:1052 gene page

Show evidence (5 references)

PPR:PPR1263744 Preprint · not peer-reviewed SUPPORT Computational

"eQTLGen tested 16,875 genes for AD risk and identiﬁed 204 FDR-signiﬁcant genes, led by BIN1."

eQTLGen blood eQTL analysis identified BIN1 as the lead AD-risk gene among 204 FDR-significant hits, establishing it as a primary genetic risk factor distinct from amyloid-burden genetics.

PPR:PPR1263744 Preprint · not peer-reviewed SUPPORT Computational

"CEACAM16-AS1 led cortex, frontal cortex BA9, anterior cingulate BA24, and hippocampus, whereas BIN1 led cerebellum"

Multi-region eQTL analysis (cortex, frontal cortex BA9, anterior cingulate BA24, hippocampus, cerebellum) confirms BIN1 as the lead gene in cerebellum, establishing brain-region-specific eQTL support.

PPR:PPR1263744 Preprint · not peer-reviewed SUPPORT Computational

"BIN1 is the strongest broad AD-risk anchor and has external disease-state expression support."

Integrated analysis across eQTL, colocalization, MetaBrain, and disease-state expression layers converges on BIN1 as the primary non-APOE AD-risk architecture anchor.

+ 2 more references

TREM2 (Risk Factor)

Gene: TREM2 hgnc:17761 gene page

NLRP3 (Genetic Variant)

Gene: NLRP3 hgnc:16400 gene page

PYCARD (Genetic Variant)

Gene: PYCARD hgnc:16608 gene page

LRP1 (Genetic Variant)

Gene: LRP1 hgnc:6692 gene page

Show evidence (1 reference)

PMID:19098903 SUPPORT Human Clinical

"Western blot analysis of small cerebral leptomeningeal arterial blood vessel homogenates from AD individuals with CAA showed elevated levels of SRF and MYOCD, Aβ accumulation and significantly lower levels of LRP, compared with age-matched, non-demented controls without CAA"

LRP1 protein is reduced in post-mortem cerebral arterial vessels from AD patients with CAA, alongside vessel-wall amyloid-beta accumulation, supporting the reduced-expression part of this record.

💊

Medical Actions

Cholinesterase Inhibitors

Action: cholinesterase inhibitor therapy Ontology label: Pharmacotherapy NCIT:C15986

Medications that slow the breakdown of acetylcholine to help with memory and cognitive function (e.g., donepezil, rivastigmine).

Show evidence (5 references)

PMID:24807367 SUPPORT

"Donepezil, galantamine and rivastigmine are commonly used AChEIs in pharmacotherapy for AD, slowing the progression and controlling the symptoms of AD."

The literature supports that cholinesterase inhibitors, including donepezil and rivastigmine, are used to help with symptoms of Alzheimer's disease by slowing the progression and controlling cognitive symptoms.

PMID:28671413 SUPPORT

"Cholinesterase inhibitors, memantine, and a combination of a cholinesterase inhibitor and memantine have produced statistically significant but clinically small delays in various domains of cognitive and functional decline in select patients with Alzheimer disease."

The literature supports the use of cholinesterase inhibitors in delaying cognitive decline in Alzheimer's disease.

PMID:35608903 SUPPORT

"Two classes of drug - cholinesterase inhibitors (donepezil, galantamine and rivastigmine) and memantine - are widely licensed for dementia due to Alzheimer's disease."

The literature confirms that cholinesterase inhibitors are licensed for use in Alzheimer's disease to alleviate symptoms and delay disease progression.

+ 2 more references

Anti-Amyloid Monoclonal Antibody Therapy

Action: Pharmacotherapy NCIT:C15986

Agent: lecanemab NCIT:C175105 donanemab NCIT:C166484 monoclonal antibody NCIT:C20401

Platform: Monoclonal antibody

Amyloid-beta-directed monoclonal antibodies (lecanemab, donanemab) are disease-modifying immunotherapies for early Alzheimer disease (mild cognitive impairment or mild dementia with confirmed amyloid pathology). They clear brain amyloid-beta and produce a modest slowing of cognitive and functional decline; use is limited by amyloid-related imaging abnormalities (ARIA) and infusion reactions, and requires amyloid confirmation plus MRI monitoring.

Show evidence (4 references)

PMID:36449413 SUPPORT Human Clinical

"Lecanemab, a humanized IgG1 monoclonal antibody that binds with high affinity to Aβ soluble protofibrils, is being tested in persons with early Alzheimer's disease."

Identifies lecanemab as an amyloid-beta protofibril-directed monoclonal antibody developed for early Alzheimer disease.

PMID:36449413 SUPPORT Human Clinical

"Lecanemab reduced markers of amyloid in early Alzheimer's disease and resulted in moderately less decline on measures of cognition and function than placebo at 18 months but was associated with adverse events."

Phase 3 CLARITY-AD trial (NCT03887455) shows lecanemab lowers amyloid burden and modestly slows cognitive/functional decline versus placebo, supporting the disease-modifying anti-amyloid immunotherapy class.

PMID:37459141 SUPPORT Human Clinical

"Among participants with early symptomatic Alzheimer disease and amyloid and tau pathology, donanemab significantly slowed clinical progression at 76 weeks in those with low/medium tau and in the combined low/medium and high tau pathology population."

Phase 3 TRAILBLAZER-ALZ 2 trial (NCT04437511) demonstrates that donanemab significantly slows clinical progression in early symptomatic Alzheimer disease, anchoring the second FDA-reviewed anti-amyloid immunotherapy in this class.

+ 1 more reference

Tau-Directed Antisense Oligonucleotide Therapy

Action: Pharmacotherapy NCIT:C15986

Agent: diranersen NCIT:C1291

Platform: Antisense oligonucleotide RNase H knockdown Delivery: Unformulated (free uptake) Targeting: Unconjugated Chemistry: 2′-MOE

RNA target: MAPT hgnc:6893 MAPT mRNA

Diranersen (BIIB080 / IONIS-MAPTRx) is an investigational, intrathecally administered antisense oligonucleotide that binds MAPT mRNA and recruits RNase H1 to degrade it, lowering production of both intracellular and extracellular tau. It is the first tau-lowering strategy to demonstrate target engagement in Alzheimer disease and represents a mechanistically distinct disease-modifying approach from the anti-amyloid antibodies: rather than clearing extracellular amyloid-beta, it reduces the tau substrate that hyperphosphorylates and aggregates into neurofibrillary tangles, the pathology most tightly correlated with cognitive decline. The phase 1b MAPTRx trial (NCT03186989) showed dose-dependent >50% mean reductions in CSF total-tau, and the phase 2 CELIA study (NCT05399888) in mild cognitive impairment / mild AD dementia used the Clinical Dementia Rating-Sum of Boxes (CDR-SB) as its primary endpoint. Per Biogen/Ionis topline results reported at AAIC 2026, CELIA reduced CSF total-tau by ~50-65% across doses and showed a slowing of clinical decline versus placebo, although it did not meet its primary endpoint; the FDA granted diranersen Fast Track designation in 2025.

Mechanism Target:

INHIBITS Neurofibrillary Tangle Formation — Diranersen indirectly inhibits this mechanism upstream rather than blocking aggregate assembly directly: RNase H1-mediated MAPT mRNA knockdown lowers the total tau available to become hyperphosphorylated and aggregate into neurofibrillary tangles.

Show evidence (3 references)

PMID:37095250 SUPPORT Human Clinical

"We sought to inhibit MAPT expression with a tau-targeting antisense oligonucleotide (MAPTRx) and reduce tau levels in patients with mild AD."

Phase 1b MAPTRx trial establishes diranersen as an antisense oligonucleotide that lowers tau by inhibiting MAPT expression in mild Alzheimer disease.

PMID:37095250 SUPPORT Human Clinical

"Dose-dependent reduction in the CSF total-tau concentration was observed with greater than 50% mean reduction from baseline at 24 weeks post-last dose in the 60 mg (four doses) and 115 mg (two doses) MAPTRx groups."

Demonstrates dose-dependent CSF total-tau lowering (target engagement) with MAPTRx, the pharmacodynamic basis for the tau-directed ASO approach.

PMID:37095250 SUPPORT Human Clinical

"randomized 3:1 to intrathecal bolus administrations of MAPTRx or placebo every 4 or 12 weeks during the 13-week treatment period"

Documents the intrathecal (CSF) route of administration required for CNS delivery of the unconjugated ASO.

NMDA Receptor Antagonist

Action: Pharmacotherapy NCIT:C15986

Agent: memantine CHEBI:64312

Medication that regulates glutamate activity to improve symptoms (e.g., memantine).

Show evidence (3 references)

PMID:12768511 SUPPORT

"Memantine, an antagonist of the glutamatergic NMDA receptor, has been recently approved for the treatment of advanced AD. Due to its action mechanism, memantine is considered a neuroprotective drug, whose utility has been demonstrated in preclinical studies, and a useful symptomatic treatment..."

The abstract confirms that memantine, an NMDA receptor antagonist, is used to treat Alzheimer's Disease by regulating glutamate activity.

PMID:27662322 SUPPORT

"Studies indicate that the distinct outcomes of NMDAR-mediated responses are induced by regionalized receptor activities, followed by different downstream signaling pathways. The activation of synaptic NMDARs initiates plasticity and stimulates cell survival. In contrast, the activation of..."

This abstract highlights that memantine, an NMDA receptor antagonist, helps in blocking the negative effects of excessive NMDAR activity in Alzheimer's Disease.

PMID:20943326 SUPPORT INDIRECT

"The hypothesis proposed is restoration of medium spiny neurons in Huntington's disease using neural progenitor cell implantation and attenuation of glutamate mediated excitotoxicity using a partial glutamate antagonist - Memantine. Memantine can block the NMDA receptors and will prevent excess..."

Although primarily discussing Huntington's Disease, this abstract supports the idea that memantine, an NMDA receptor antagonist, regulates glutamate activity, which is relevant to Alzheimer's Disease treatment.

Cognitive Therapy

Action: behavioral counseling Ontology label: Behavioral Counseling NCIT:C181743

Platform: Behavioral / lifestyle

Non-pharmacological interventions to maintain cognitive function.

Show evidence (5 references)

PMID:35621327 SUPPORT

"Cognitive stimulation was found to be an effective intervention for people with moderate Alzheimer's disease because it helped to maintain memory function, executive functions, and attention."

This study specifically highlights the effectiveness of cognitive stimulation in maintaining cognitive functions in patients with moderate Alzheimer's disease.

PMID:28671413 SUPPORT

"Cognitive stimulation programs show benefit in maintenance of cognitive function and improved self-reported quality of life in patients with mild to moderate Alzheimer disease."

This reference supports the statement by indicating that cognitive stimulation programs help maintain cognitive function in patients with mild to moderate Alzheimer's disease.

PMID:37428401 SUPPORT

"CT can improve discrete cognitive functions, but durability is limited and real-world utility is unclear. CR treatments are holistic and flexible and, therefore, most promising but are difficult to simulate and study under rigorous experimental conditions."

A review of cognitive interventions in Alzheimer and Parkinson disease. Cognitive training improves specific functions with limited durability, and cognitive rehabilitation is the most promising but least rigorously tested approach, so the support for maintaining cognitive function is real but modest.

+ 2 more references

Supportive Care

Action: Supportive Care NCIT:C15747

Includes occupational therapy, speech therapy, and caregiver support.

Show evidence (3 references)

PMID:28809650 SUPPORT

"Occupational therapy practitioners play a significant role in supporting adults with Alzheimer's disease and related major neurocognitive disorders, as well as their caregivers, through all phases of the disease process."

The reference highlights the role of occupational therapy in supporting individuals with Alzheimer's disease, which aligns with the statement that includes occupational therapy as part of supportive care.

PMID:38883339 SUPPORT

"The telephone hotline is a useful component of dementia care in Germany and an important contribution to the National Dementia Strategy."

The reference discusses the importance of caregiver support through telephone counseling, aligning with the statement that includes caregiver support as part of supportive care.

PMID:29361068 SUPPORT

"This article first describes the educational, information, and support needs of individuals living dementia and their family caregivers across all stages of Alzheimer's."

The reference discusses the support needs of individuals with Alzheimer's disease and their caregivers, which aligns with the statement that includes caregiver support as part of supportive care.

Lifestyle Modifications

Action: Lifestyle Therapy NCIT:C15900

Physical exercise, mental stimulation, and healthy diet to potentially slow disease progression.

Show evidence (4 references)

PMID:35503939 SUPPORT

"The aim of this review is to emphasize the importance of mental activity and aerobic physical exercise as one of the most important health-related activities which may delay the onset or slow down the progression of Alzheimer's dementia."

The review highlights the importance of mental and physical exercise in potentially slowing the progression of Alzheimer's disease.

PMID:32579499 SUPPORT

"Combined with the prevention of AD risk factors such as heart disease, diabetes, and with more recent evidence, microbiome dysfunction, there is a substantial foundation for diet as a modifiable risk factor and preventative measure for AD."

The review suggests that a healthy diet can be a preventative measure for Alzheimer's disease, supporting the role of lifestyle modifications.

PMID:37321363 SUPPORT

"Results showed that eating a healthy diet with plenty of fruits and vegetables, and participation in leisure and physical activities may protect against cognitive decline and cognitive impairment among oldest-old regardless of the APOE genotype."

The systematic review indicates that lifestyle factors such as diet and physical activities may protect against cognitive decline, supporting the statement.

+ 1 more reference

🌍

Environmental Factors

Cumulative exposure to strong central anticholinergic medication

Curated deliberately as PREDISPOSES, not as an established cause. Three findings keep it there. First, no randomized evidence: the Cochrane review of anticholinergic deprescribing found very low certainty evidence that can neither support nor refute a cognitive benefit of stopping these drugs (PMID:38063254). Second, class specificity: the large UK studies concentrate the signal in antidepressant, urological and antiparkinson drugs (PMID:29695481, PMID:31233095), and a propensity-matched target trial emulation finds no antihistamine signal (PMID:42250644). Third, the one large cohort reporting an antihistamine dose-response reports it for second-generation agents too (PMID:38935035), which do not meaningfully enter the CNS. That third point cuts both ways and should not be quoted as a failed negative control: the gradient is steeper for the first-generation, CNS-penetrant class (1.13/1.29/1.51 across rising cumulative dose versus 1.11/1.19/1.26 for second-generation), which is concordant with a central mechanism; it is the residual non-null second-generation signal, not the comparison between the classes, that the muscarinic mechanism fails to account for. IMPORTANT COUNTERWEIGHT, from the CPRD full text rather than its abstract: the antihistamine null there is weaker than it looks. Antihistamines were tested and were not significant on any-prescription exposure, but the authors report a tentative effect above 365 defined daily doses, describe the overall antihistamine association as small and short of their significance threshold, state that over-the-counter supply is not captured in the database so antihistamine use is underestimated, and call for independent confirmation specifically for antihistamines. Since diphenhydramine is predominantly an over-the-counter drug, exposure measures built from prescriptions systematically undercount precisely the exposure of interest, so the antihistamine nulls are substantially nulls of exposure misclassification rather than clean refutations. The claim that survives all of this unambiguously is the acute one (PMID:11570937), which is why diphenhydramine is listed as potentially inappropriate in older adults regardless of how the dementia question resolves.

Years of cumulative exposure to CNS-penetrant muscarinic antagonists — first-generation antihistamines such as diphenhydramine, tricyclic antidepressants, bladder antimuscarinics, and antiparkinson anticholinergics — is associated with incident dementia and Alzheimer disease in a dose-dependent manner across several large observational cohorts. Curated as a candidate modifiable risk factor, NOT as an established cause; see notes and the attached discussions.

Show evidence (4 references)

PMID:25621434 SUPPORT Human Clinical

"The most common anticholinergic classes used were tricyclic antidepressants, first-generation antihistamines, and bladder antimuscarinics."

Identifies the drug classes constituting the exposure, including the first-generation antihistamines of which diphenhydramine is the exemplar.

PMID:40707785 SUPPORT Human Clinical

"moderate certainty of increased risk with anticholinergics"

A GRADE-rated umbrella review across 68 meta-analyses places the anticholinergic-dementia association at moderate certainty, the current ceiling for this exposure.

PMID:11570937 SUPPORT Human Clinical

"Diphenhydramine administration in older hospitalized patients is associated with an increased risk of cognitive decline and other adverse effects with a dose-response relationship."

Establishes drug-specific, dose-responsive acute cognitive harm from diphenhydramine — the well-supported near-term effect, distinct from the contested long-term dementia claim.

+ 1 more reference

Mechanism Target:

PREDISPOSES Muscarinic M1 Receptor Signaling Loss — Chronic anticholinergic exposure is the pharmacologic route by which M1 muscarinic signaling is persistently withdrawn, the state the mouse work links to accelerated amyloid and tau pathology. PREDISPOSES rather than TRIGGERS: the human evidence is observational, dose-dependent, and does not establish that the exposure initiates disease.

Show evidence (2 references)

PMID:25621434 SUPPORT Human Clinical

"A 10-year cumulative dose-response relationship was observed for dementia and Alzheimer disease"

Cumulative-dose dependence is the signature expected if sustained receptor blockade, rather than transient pharmacologic effect, is what matters.

PMID:32878992 SUPPORT Human Clinical

"aCH+ participants had increased risk of progression to MCI"

Anticholinergic exposure predicts incident MCI in cognitively normal adults, with the effect amplified by APOE4 and CSF AD pathology — interaction with AD biology rather than an independent cognitive effect.

Oral glucosamine supplementation

The human evidence for harm and the human evidence against it answer different questions and should not be netted against each other. The originating EHR analysis (PMID:42265388) is about outcomes after diagnosis, in one health system, with exposure from keyword matching of clinical notes in roughly 8% of patients, and with ADRD pooling several dementias. It is exposed to confounding by indication and frailty: glucosamine is taken for osteoarthritis and joint pain, which track age, immobility and multimorbidity, and the reported adjustment covers age, sex and demographics. The UK Biobank cohorts (PMID:36978077, PMID:36514123, PMID:37689747) are about incidence in people without dementia at baseline and find inverse or null associations; a fourth UK Biobank analysis (PMID:37158699) also finds no association with incident dementia (HR 1.06, 95% CI 0.99-1.14). None of these is randomized. The originating authors call for a double-blind trial, which is the only design that separates a glucosamine effect from the reasons people take it.

Glucosamine, taken widely as a joint supplement, enters the hexosamine pathway directly and so feeds N-glycan biosynthesis. In 5xFAD mice it raises brain N-glycans and worsens memory, while in wild-type mice it does neither. In a single-health-system electronic health record analysis, documented glucosamine use was associated with higher mortality in AD and related dementias and with more frequent conversion from mild cognitive impairment to dementia, but not with mortality in mild cognitive impairment. Curated as a candidate disease-exacerbating exposure in established disease, NOT as a risk factor for developing it: several large UK Biobank cohorts find glucosamine use associated with lower or unchanged incidence of dementia and Alzheimer disease.

Show evidence (8 references)

PMID:42265388 SUPPORT Human Clinical

"Our analysis revealed that glucosamine usage was associated with a 25% increase in mortality risk among ADRD patients"

The survival signal in the dementia cohort. ADRD here pools Alzheimer disease with vascular, Lewy body and frontotemporal dementia, so it is not an Alzheimer-specific estimate.

PMID:42265388 SUPPORT Human Clinical

"In this analysis, we observed a 25% increase in the proportion of patients with MCI transitioning to ADRD in the glucosamine user group"

The progression signal: more conversion from mild cognitive impairment to dementia among documented users.

PMID:42265388 SUPPORT Human Clinical

"By contrast, glucosamine usage did not significantly impact mortality risk in the MCI cohort"

The within-study contrast the authors use to argue that harm depends on established disease; it is what restricts this entry to EXACERBATES.

+ 5 more references

Mechanism Target:

EXACERBATES Brain N-Glycan Hyperglycosylation — Glucosamine bypasses the rate-limiting step of the hexosamine pathway and supplies precursor for N-glycan biosynthesis. EXACERBATES rather than TRIGGERS: in mice it raises brain N-glycans only on an AD background and leaves wild-type brains unchanged.

Show evidence (2 references)

PMID:42265388 SUPPORT Model Organism

"Conversely, glucosamine supplementation significantly increased glycosylation across the brain, as shown by global heat map analysis"

Oral glucosamine for two weeks raises brain N-glycans in 5xFAD mice, at a dose scaled from a human intake of about 2.5 g per day.

PMID:42265388 SUPPORT Model Organism

"Strikingly, glucosamine supplementation did not result in hyperglycosylation in WT mouse brains"

The same regimen has no effect on wild-type brains, which is why the effect is curated as exacerbation of an existing state.

🔬

Biochemical Markers

Amyloid Beta (Aβ42) (Elevated)

Context: Found in cerebrospinal fluid and brain tissue.

Show evidence (4 references)

PMID:19661632 SUPPORT

"The core candidate CSF biomarkers Abeta42, total tau (T-tau), and phosphorylated tau (P-tau) have been shown to have a high diagnostic performance to identify AD also in the early phase of the disease."

The literature confirms that Abeta42 is a core biomarker found in cerebrospinal fluid (CSF) and is elevated in Alzheimer's disease.

PMID:31958088 SUPPORT

"Genetic predispositions associated with metabolism of the amyloid-beta protein precursor underlie familial Alzheimer's disease; a form of dementia characterized by early disease onset and elevated levels of cortical amyloid-beta."

The literature confirms that elevated levels of amyloid-beta, including Abeta42, are found in the brain tissue of individuals with Alzheimer's disease.

PMID:38987603 SUPPORT

"beta-amyloid plaques contained a mixture of fibrils, some of which were branched, and protofilaments, arranged in parallel arrays and lattice-like structures."

This reference supports the presence of beta-amyloid in brain tissue, which includes Abeta42, in Alzheimer's disease.

+ 1 more reference

Astrocyte Plasma Proteomic Age Gap (Elevated)

Context: Blood-based cell-type-specific aging clock. Plasma proteins are mapped to their putative cell of origin using Human Protein Atlas single-cell transcriptomic data, and a machine-learning model estimates astrocyte biological age; the age gap is the difference between that estimate and chronological age. "Extreme" agers are those in the upper tail of the age-gap distribution and "youthful" agers the lower tail.

Pathograph Readouts

Predicts Neuroinflammation Positive Prognostic

An elevated astrocyte age gap reports an aged astrocyte proteomic state and predicts incident Alzheimer disease years before diagnosis. It is linked to the astrocytic arm of the neuroinflammation node, but the study measures plasma protein signatures rather than brain astrocyte reactivity directly.

Show evidence (1 reference)

PMID:42297981 SUPPORT Human Clinical

"Individuals with extreme astrocyte aging demonstrated a 12.59-fold increased risk of incident AD compared to those with youthful aging."

Supports treating the astrocyte age gap as a prognostic readout that anticipates clinical Alzheimer disease.

Correlates With Neurofibrillary Tangle Formation Positive Diagnostic

In the NSHD Insight-46 substudy the astrocyte age gap tracked plasma pTau-217 burden, an orthogonal blood measure of amyloid and tau pathology, linking the clock to tau pathology rather than to cognition alone.

Show evidence (1 reference)

PMID:42297981 SUPPORT Human Clinical

"we found significant associations between plasma phosphorylated tau-217 (pTau-217) burden and several neuronal and glial cell types, including astrocytes"

Provides the orthogonal pathology correlate for the astrocyte age gap, supporting a correlational readout against the tau pathology node.

Show evidence (4 references)

PMID:42297981 SUPPORT Human Clinical

"Individuals with extreme astrocyte aging demonstrated a 12.59-fold increased risk of incident AD compared to those with youthful aging."

Quantifies the prognostic association between the astrocyte plasma proteomic age gap and incident Alzheimer disease over 15 years of UK Biobank follow-up.

PMID:42297981 SUPPORT Human Clinical

"Individuals who were homozygous for APOE4 and had extreme astrocyte aging showed the highest cumulative incidence of 38.3% over 15 years of follow-up, compared to 12.6% for homozygotes with normal astrocyte aging."

Shows that astrocyte aging status stratifies absolute AD incidence within the highest-risk APOE genotype, supporting use of the age gap as a risk modifier on top of genotype.

PMID:42297981 SUPPORT Human Clinical

"extreme astrocyte aging consistently identified individuals at elevated AD risk compared to those with normal and youthful astrocyte aging, demonstrating that astrocyte aging provides independent risk stratification beyond APOE genotype"

States the authors' conclusion that the astrocyte age gap adds risk information independent of APOE genotype.

+ 1 more reference

Oligodendrocyte Precursor Cell Plasma Proteomic Age Gap (Elevated)

Context: Blood-based cell-type-specific aging clock (see Astrocyte Plasma Proteomic Age Gap for the derivation). Oligodendrocyte precursor cells were the most prominent cross-sectional association with clinical Alzheimer disease in the Global Neurodegeneration Proteomics Consortium cohort, though pancreatic endocrine cells and inhibitory neurons showed a comparable effect size.

Pathograph Readouts

Correlates With Neurofibrillary Tangle Formation Positive Diagnostic

The oligodendrocyte precursor cell age gap tracked plasma pTau-217 burden in the Insight-46 substudy, tying this glial clock to measured tau pathology as well as to cognitive stage.

Show evidence (1 reference)

PMID:42297981 SUPPORT Human Clinical

"we found significant associations between plasma phosphorylated tau-217 (pTau-217) burden and several neuronal and glial cell types, including astrocytes (coefficient beta (ꞵ) = 1.08, adjusted P = 3.92 × 10−7), oligodendrocyte precursor cells (ꞵ = 1.02, adjusted P = 1.19 × 10−6)"

Names oligodendrocyte precursor cells explicitly, with a pTau-217 regression coefficient comparable to that of astrocytes, in the NSHD Insight-46 substudy.

Show evidence (3 references)

PMID:42297981 SUPPORT Human Clinical

"AD was associated with accelerated aging across a wide range of cell types, most prominently oligodendrocyte precursor cells"

Establishes the cross-sectional case/control association in the GNPC cohort and the relative prominence of the oligodendrocyte precursor cell signal.

PMID:42297981 SUPPORT Human Clinical

"oligodendrocyte precursor cell aging showed a consistent stepwise increase with worsening cognitive impairment"

Establishes the graded relationship between the oligodendrocyte precursor cell age gap and Clinical Dementia Rating stage.

PMID:42297981 SUPPORT Human Clinical

"we observed an association between worsened cognition and elevated age gaps across a variety of cell types, with oligodendrocyte precursor cells demonstrating one of the strongest associations"

Independent cohort (NSHD Insight-46) association with the Preclinical Alzheimer Cognitive Composite, supporting the marker as a correlate of early cognitive decline.

Inhibitory Neuron Plasma Proteomic Age Gap (Elevated)

Context: Blood-based cell-type-specific aging clock (see Astrocyte Plasma Proteomic Age Gap for the derivation). The finding is notable for its selectivity: inhibitory neuron aging associated with Alzheimer disease while excitatory neuron aging did not.

Pathograph Readouts

Correlates With Synaptic Dysfunction Positive Diagnostic

An elevated inhibitory neuron age gap, in the absence of a matching excitatory neuron signal, is read as a plasma correlate of the selective interneuron vulnerability that contributes to excitation/inhibition imbalance and synaptic dysfunction in Alzheimer disease.

Show evidence (1 reference)

PMID:42297981 SUPPORT Human Clinical

"while aging of inhibitory neurons was linked to AD, aging of excitatory neurons was not"

The selectivity of the association is what licenses reading this marker against the interneuron/synaptic arm rather than neurodegeneration generally.

Show evidence (2 references)

PMID:42297981 SUPPORT Human Clinical

"while aging of inhibitory neurons was linked to AD, aging of excitatory neurons was not"

Establishes the cell-type selectivity that distinguishes this marker from a generic neuronal aging signal.

PMID:42297981 SUPPORT Human Clinical

"inhibitory neuron age gaps increased with higher CDR scores across cohorts"

Shows the inhibitory neuron age gap tracks dementia severity stage, not only case/control status.

🔬

Diagnosis

Neuropsychological Tests

Assess cognitive function including memory, language, and problem-solving.

Show evidence (4 references)

PMID:23072720 SUPPORT

"A comprehensive neuropsychological examination encompassing several cognitive domains can provide a pattern of altered and preserved functions that is helpful to early detection, differential diagnosis and even prognosis of progression in predementia stages."

The article mentions that neuropsychological tests encompass several cognitive domains, which include memory, language, and problem-solving.

PMID:29851873 SUPPORT

"The focused history and mental status examination remain essential tools for the evaluation and diagnosis of neurologic disorders affecting cognition, language, and behavior."

The article highlights the importance of mental status examinations, which are part of neuropsychological tests, in assessing cognitive functions including memory, language, and behavior.

PMID:37244373 SUPPORT

"The results demonstrate significant correlations between UEF cognitive score and mini-mental state examination (MMSE), Mini-Cog, Category fluency, Benson complex figure copy, Trail making test, and Montreal cognitive assessment (MOCA)."

The study shows that neuropsychological tests like MMSE, Mini-Cog, and others are used to assess cognitive functions, which include memory, language, and problem-solving.

+ 1 more reference

Brain Imaging

MRI and PET scans to detect brain changes such as shrinkage and amyloid plaques.

Show evidence (4 references)

PMID:22173295 SUPPORT

"Brain imaging researchers have contributed to the scientific understanding, early detection and tracking of AD. ... They have developed ground-breaking methods, including positron emission tomography (PET) ligands to measure fibrillar amyloid-β (Aβ) deposition, new magnetic resonance imaging..."

The literature supports the use of MRI and PET scans for detecting brain changes such as shrinkage and amyloid plaques in Alzheimer's Disease.

PMID:33640881 SUPPORT

"Glucose metabolism reduction and brain volume losses are widely reported in Alzheimer's disease (AD). ... The AD group had significantly reduced volume in the hippocampus and DMN regions (P < 0.001) relative to that of normal controls determined by using ROI analysis."

The literature supports the use of brain imaging techniques like MRI and PET to detect brain changes such as shrinkage and amyloid plaques in Alzheimer's Disease.

PMID:34127752 SUPPORT Model Organism

"Amyloid plaques are a hallmark of Alzheimer's disease (AD) that develop in its earliest stages. Thus, non-invasive detection of these plaques would be invaluable for diagnosis and the development and monitoring of treatments."

The literature supports the use of MRI and PET scans for detecting amyloid plaques in Alzheimer's Disease.

+ 1 more reference

Cerebrospinal Fluid Analysis

Measurement of amyloid-beta, total tau, and phosphorylated tau levels.

Show evidence (5 references)

PMID:19661632 SUPPORT

This reference explicitly mentions the use of amyloid-beta, total tau, and phosphorylated tau levels in cerebrospinal fluid for diagnosing Alzheimer's disease.

PMID:12975285 SUPPORT

"Total tau, Ptau-181, and amyloid-beta1-42 levels in CSF, obtained by lumbar puncture, were determined by sandwich enzyme-linked immunosorbent assay."

This study evaluates levels of total tau, phosphorylated tau 181, and amyloid-beta in the cerebrospinal fluid of patients with Alzheimer's disease, supporting the statement.

PMID:36510321 SUPPORT

"Alzheimer's disease is characterized by an abnormal increase of phosphorylated tau (pTau) species in the CSF."

This reference supports the statement by discussing the increase of phosphorylated tau in cerebrospinal fluid in Alzheimer's disease.

+ 2 more references

Blood-Based Biomarkers (plasma p-tau217)

Minimally invasive plasma assays now approach the accuracy of amyloid-PET and CSF for detecting AD pathology, greatly expanding access to biomarker-based diagnosis. Plasma phosphorylated-tau 217 (p-tau217) and the p-tau217/Aβ42 ratio are the strongest performers for classifying amyloid status and predicting progression; the plasma Aβ42/40 ratio shows the earliest detectable abnormalities. In May 2025 the first blood test for brain amyloid (a p-tau217/β-amyloid 1-42 ratio) received FDA approval. Blood biomarkers complement — and are increasingly used to triage before — CSF and PET testing, which remains relevant given that anti-amyloid immunotherapies require confirmed amyloid positivity.

Show evidence (2 references)

PMID:32722745 SUPPORT Human Clinical

"Among 1402 participants from 3 selected cohorts, plasma P-tau217 discriminated AD from other neurodegenerative diseases, with significantly higher accuracy than established plasma- and MRI-based biomarkers, and its performance was not significantly different from key CSF- or PET-based measures."

Plasma p-tau217 discriminates AD from other neurodegenerative diseases with accuracy exceeding other blood/MRI biomarkers and comparable to CSF/PET, supporting its use as a blood-based diagnostic biomarker.

PMID:32722745 SUPPORT Human Clinical

"To examine plasma tau phosphorylated at threonine 217 (P-tau217) as a diagnostic biomarker for AD."

Establishes plasma p-tau217 (tau phosphorylated at threonine 217) as the diagnostic biomarker evaluated in this study.

Montreal Cognitive Assessment

Brief cognitive screening instrument developed to help first-line physicians detect mild cognitive impairment. Its validation study administered it, alongside the MMSE, to patients with mild Alzheimer's disease, patients with mild cognitive impairment and healthy elderly controls, and it detected every patient in the mild Alzheimer's disease group at the published cut-off.

MoCA LOINC:72133-2

Reference Ranges

Total score [MoCA] (MoCA validation study, 2005 (mild Alzheimer's disease with MMSE score of 17 or more, mild cognitive impairment, and healthy elderly controls; a community clinic and an academic center))

Below the cut-off (cognitive impairment) (< 26 {score}) At or above the cut-off (≥ 26 {score})

The abstract states the cut-off as "Using a cutoff score 26" and does not say in its own text on which side a score of exactly 26 falls; the bands put 26 in the unimpaired tier and 25 and below in the impaired tier, a placement the cited abstract does not itself state. A later meta-analysis of MCI validation studies (PMID:28731508) reports that "the originally suggested cutoff score of 26/30 leads to an inflated rate of false positives, particularly for those of older age and/or lower education" and recommends 23; it concerns MCI rather than Alzheimer's disease, so it is not recorded as a separate range here.

Show evidence (2 references)

PMID:15817019 SUPPORT PRIMARY RESULT Human Clinical

"Using a cutoff score 26, the MMSE had a sensitivity of 18% to detect MCI, whereas the MoCA detected 90% of MCI subjects. In the mild AD group, the MMSE had a sensitivity of 78%, whereas the MoCA detected 100%."

States the cut-off of 26 and the MoCA's sensitivity for mild Alzheimer's disease at that cut-off.

PMID:28731508 SUPPORT BACKGROUND Human Clinical

"Some studies have indicated that the originally suggested cutoff score of 26/30 leads to an inflated rate of false positives, particularly for those of older age and/or lower education."

Confirms that 26 is the originally suggested cut-off and records the known false-positive caveat in older and less-educated people.

Total score [MoCA] (Taiwanese cohort of 67 patients with mild-to-moderate Alzheimer's disease, 36 with dementia with Lewy bodies and 62 healthy participants without dementia, aged 60 to 90 (Alzheimer's disease versus controls))

At or below 21 (screen positive for Alzheimer's disease) (< 22 {score}) 22 or above (≥ 22 {score})

The abstract states "The cut-off point of the MoCA for AD was 21/22"; the bands read the x/y notation as 21 or below positive and 22 or above negative, which the abstract does not spell out.

Show evidence (1 reference)

PMID:23919950 SUPPORT PRIMARY RESULT Human Clinical

"The cut-off point of the MoCA for AD was 21/22 with a sensitivity of 95.5% and a specificity of 82.3% (area under the curve (AUC): 0.945)"

States the MoCA cut-off separating Alzheimer's disease from controls in this cohort.

Copyright © Tina Brosseau for Center for Diagnosis & Research on Alzheimer's disease (CEDRA). Used with permission. The instrument is bound at its LOINC panel code and the cut-offs at its total-score code. The instrument's items are not reproduced in this entry; only published thresholds are recorded.

Show evidence (1 reference)

PMID:15817019 SUPPORT PRIMARY RESULT Human Clinical

"The MoCA and MMSE were administered to all participants, and sensitivity and specificity of both measures were assessed for detection of MCI and mild AD."

The instrument was validated for detection of mild Alzheimer's disease as well as MCI.

Mini-Mental State Examination

The most common screening measure of cognitive function. MMSE score ranges have been mapped onto Clinical Dementia Rating stages so that the MMSE can stand in for the CDR when staging dementia in Alzheimer's disease.

MMSE LOINC:72107-6

Reference Ranges

Total score [MMSE] (863 probands (524 with probable Alzheimer's disease, 92 with questionable dementia, 247 with memory complaints but no objective cognitive impairment), MMSE ranges mapped onto CDR stages)

Severe dementia (0–11 {score}) Moderate dementia (11–21 {score}) Mild dementia (21–26 {score}) Questionable dementia (26–30 {score}) No dementia (≥ 30 {score})

The abstract states "The MMSE ranges were 30 for no, 26-29 for questionable, 21-25 for mild, 11-20 for moderate, and 0-10 for severe dementia." The integer ranges are written as half-open bands. The same abstract reports that the MMSE "performed poorly in the separation between CDR stages zero and 0.5", with only moderate agreement for no dementia (kappa=0.44) and fair agreement for questionable dementia (kappa=0.28), so the two upper bands are the least reliable.

Show evidence (2 references)

PMID:16473978 SUPPORT PRIMARY RESULT Human Clinical

"The MMSE ranges were 30 for no, 26-29 for questionable, 21-25 for mild, 11-20 for moderate, and 0-10 for severe dementia."

States the MMSE severity ranges recorded as bands.

PMID:16473978 SUPPORT PRIMARY RESULT Human Clinical

"The MMSE can be used as a surrogate measure for the CDR for the staging of dementia in AD."

The ranges are a staging scheme for dementia in Alzheimer's disease.

Total score [MMSE] (Taiwanese cohort of 67 patients with mild-to-moderate Alzheimer's disease, 36 with dementia with Lewy bodies and 62 healthy participants without dementia, aged 60 to 90 (Alzheimer's disease versus controls))

At or below 24 (screen positive for Alzheimer's disease) (< 25 {score}) 25 or above (≥ 25 {score})

The abstract states "For the MMSE, the cut-off points for AD and for DLB from NC were all 24/25"; the bands read the x/y notation as 24 or below positive and 25 or above negative, which the abstract does not spell out.

Show evidence (1 reference)

PMID:23919950 SUPPORT PRIMARY RESULT Human Clinical

"For the MMSE, the cut-off points for AD and for DLB from NC were all 24/25, with a sensitivity of 88.1% and a specificity of 85.5% for AD (AUC: 0.92)"

States the MMSE cut-off separating Alzheimer's disease from controls in this cohort.

Copyright © Vicki King for Psychological Assessment Resources. Used with permission. The instrument is bound at its LOINC panel code and the cut-offs at its total-score code. The instrument's items are not reproduced in this entry; only published thresholds are recorded.

Show evidence (1 reference)

PMID:35308911 SUPPORT

"Assessing cognitive impairment is important for diagnostic, clinical management, and research purposes. The Folstein Mini-Mental State Examination (MMSE) is the most common screening measure of cognitive function."

The MMSE is the most commonly used cognitive screening measure, which is why it is bound as a measurement here.

Clinical Dementia Rating

Global dementia staging instrument rated from information collected from both the patient and an informant. Its stages (0, 0.5, 1, 2 and 3) grade Alzheimer's disease from no dementia through questionable, mild, moderate and severe dementia. Its sum of boxes is the primary endpoint of several trials recorded in this entry's treatments.

CDR LOINC:72088-8

Copyright © Washington University Charles F. and Joanne Knight Alzheimer's Disease Research Center (Knight ADRC). Used with permission. LOINC carries the CDR only as a panel. A search of the LOINC 2.83 table (the Tuva Project extract pinned in data/loinc/MANIFEST.yaml) for "CDR", "dementia rating" and "sum of boxes", and of the Monarch KG for "Clinical Dementia Rating", "CDR global", "Sum of boxes score [CDR]" and "Total score [CDR]", found no global-score or sum-of-boxes code, so the CDR stages are not recorded as interpretation bands: a numeric band on a panel code would put a score threshold on the instrument's identity. The instrument's items are not reproduced in this entry.

Show evidence (2 references)

PMID:16473978 SUPPORT BACKGROUND Human Clinical

"The Clinical Dementia Rating (CDR) has demonstrated high validity and reliability for this purpose, but it requires a considerable amount of data to be collected both from the patient and from an informant."

The CDR is the validated instrument for distinguishing severity levels of Alzheimer's disease, rated from patient and informant data.

PMID:16473978 SUPPORT PRIMARY RESULT Human Clinical

"The MMSE discriminated well between CDR stages 0.5, 1, 2, and 3 but performed poorly in the separation between CDR stages zero and 0.5."

Names the CDR stages used to stage dementia in Alzheimer's disease.

📊

Prevalence

Global

Point Prevalence 480 per 100,000 (447–521) >1 in 1,000

An estimated 57.4 million people worldwide were living with dementia in 2019 (GBD 2019; PMID:34998485), projected to reach 152.8 million by 2050. Alzheimer disease accounts for 60–70% of dementia cases. Applying 60–70% to the 2019 global dementia prevalence of ~745 per 100,000 (57.4 M / 7.7 B population) yields an AD-specific point prevalence of approximately 447–521 per 100,000 (midpoint ~480); this is a derived estimate, not directly quoted from a single source.

Show evidence (1 reference)

PMID:34998485 SUPPORT

"We estimated that the number of people with dementia would increase from 57.4 (95% uncertainty interval 50.4-65.1) million cases globally in 2019 to 152.8 (130.8-175.9) million cases in 2050."

Global Burden of Disease Study 2019 analysis providing country-level estimates of dementia prevalence and forecasts to 2050.

📊

Related Datasets

Molecular characterization of selectively vulnerable neurons in Alzheimer's Disease geo:GSE147528

Single-nucleus RNA-seq of caudal entorhinal cortex and superior frontal gyrus from post-mortem brains spanning the progression of tau neurofibrillary pathology. The dataset from which RORB was identified as a marker of selectively vulnerable excitatory neurons.

Homo sapiens SINGLE CELL RNA SEQ n=20

entorhinal cortex UBERON:0002728 frontal cortex UBERON:0001870

PMID:33432193

Underlies the Selective Vulnerability of RORB+ Entorhinal Excitatory Neurons node. Accession resolved against the GEO API with `just verify-datasets`. Re-deriving the RORB result from this dataset is not an independent test of it.

Molecular signatures underlying neurofibrillary tangle susceptibility in Alzheimer's disease geo:GSE129308

Transcriptomes of single somas bearing neurofibrillary tangles compared with tangle-free somas isolated from the same human Alzheimer brains. The within-donor design controls for donor-level confounding when asking what distinguishes an aggregation-prone from an aggregation-resistant neuron.

Homo sapiens SINGLE CELL RNA SEQ n=27

cerebral cortex UBERON:0000956

PMID:35882228

The sharpest available human contrast for three separate questions curated in this entry: which neurons are selectively vulnerable, whether senescent cells are the tangle-bearing excitatory neurons reported in PMID:35531351, and whether necroptosis effectors are enriched in pathology-bearing neurons. Accession resolved against the GEO API with `just verify-datasets`. Note that the GEO record's own linked citation is PMID:41620473, a later reuse of the series (Dharshini et al. 2026 on layer 4 resilience); the originating study is Otero-Garcia et al. 2022 (PMID:35882228), whose title matches the series title. Do not take the GEO-linked PMID as the source publication here.

APOE4/4 is linked to damaging lipid droplets in Alzheimer's microglia geo:GSE254205

Single-nucleus RNA-seq of human Alzheimer brain stratified by APOE genotype, defining the ACSL1-positive lipid-droplet microglial state that is most abundant in APOE4/4 carriers.

Homo sapiens SINGLE CELL RNA SEQ n=102

cerebral cortex UBERON:0000956

PMID:38480892

Source data for the Microglial Lipid Droplet Accumulation node. Because the donors are APOE-genotyped, the same nuclei also bear on whether the glial lipid phenotype is microglia-specific or shared with the oligodendrocyte lineage. Accession resolved against the GEO API with `just verify-datasets`.

Single-nucleus chromatin accessibility and transcriptomic characterization of Alzheimer's Disease geo:GSE174367

Paired single-nucleus RNA-seq and ATAC-seq of human Alzheimer cortex, so expression changes can be checked against chromatin accessibility in the same cell types.

Homo sapiens MULTI OMICS n=230

cerebral cortex UBERON:0000956

PMID:34239132

The paired assay is what makes this dataset useful beyond replication: it allows a senescence or necroptosis expression claim to be tested for regulatory support rather than transcript abundance alone. Accession resolved against the GEO API with `just verify-datasets`.

Single-nucleus transcriptome analysis reveals dysregulation of angiogenic endothelial cells and neuroprotective glia in Alzheimer's disease geo:GSE157827

Independent human cortical single-nucleus RNA-seq cohort with endothelial and glial coverage.

Homo sapiens SINGLE CELL RNA SEQ n=21

cerebral cortex UBERON:0000956

PMID:32989152

Curated as a replication cohort rather than for a specific node. Accession resolved against the GEO API with `just verify-datasets`.

A single-cell atlas of the human cortex reveals drivers of transcriptional changes in Alzheimer's disease in specific cell subpopulations geo:GSE138852

Human entorhinal cortex single-nucleus atlas resolving cell-subpopulation transcriptional change in Alzheimer disease.

Homo sapiens SINGLE CELL RNA SEQ n=8

entorhinal cortex UBERON:0002728

PMID:31768052

Small donor count (8); curated as a replication cohort, not as a primary discovery set. Accession resolved against the GEO API with `just verify-datasets`.

Distinct amyloid-b and tau associated microglia profiles in Alzheimer's disease geo:GSE148822

Human single-nucleus microglial profiling separating amyloid-associated from tau-associated microglial states.

Homo sapiens SINGLE CELL RNA SEQ n=95

cerebral cortex UBERON:0000956

PMID:33609158

Directly relevant to the entry's repeated claim that glial responses are pathology-context-dependent rather than uniform, including the amyloid-versus-tau split curated on the Neuroinflammation node. Accession resolved against the GEO API with `just verify-datasets`.

Diverse human astrocyte and microglial transcriptional responses to Alzheimer's pathology geo:GSE160936

Human astrocyte and microglial transcriptional responses to Alzheimer pathology, profiled across donors.

Homo sapiens SINGLE CELL RNA SEQ n=24

cerebral cortex UBERON:0000956

PMID:34767070

Glia-focused, and so a natural place to characterize astrocyte and microglial responses to Alzheimer pathology. It is not a place to settle the cell-identity dispute curated on the Senescent Cell Accumulation node: it profiles glia only, so it carries no neuronal comparator and cannot adjudicate a glia-versus-neuron contrast, which needs both cell classes measured in the same assay. It can test whether a glial senescence signature is present; it cannot test whether glia rather than excitatory neurons are the senescent population. Accession resolved against the GEO API with `just verify-datasets`.

SEA-AD (Seattle Alzheimer's Disease Brain Cell Atlas) synapse:syn26223298

Multimodal atlas of 84 aged donors spanning the full range of Alzheimer neuropathology, with single-nucleus RNA-seq, ATAC-seq and multiome, MERFISH spatial transcriptomics, and quantitative neuropathology, resolved to 139 molecular cell types across middle temporal gyrus and Brodmann area 9.

Homo sapiens MULTI OMICS n=84

temporal cortex UBERON:0016538 frontal cortex UBERON:0001870

PMID:39402379

Access is split, and the distinction matters for anyone planning an analysis: the raw FASTQs are controlled (Sage/Synapse data use agreement), but the nuclei-by-gene matrices are openly downloadable and are also served through CELLxGENE and the Allen Brain Cell Atlas. `synapse:` accessions have no open per-record metadata API, so `just verify-datasets` reports this one as UNSUPPORTED rather than OK — that is the expected result, not a failure. This is the only human dataset here with enough nuclei per glial class to give the astrocyte, microglia and oligodendrocyte-progenitor senescence hypotheses a fair test against the excitatory-neuron claim.

PsychAD_NPS synapse:syn60084804

PsychAD single-nucleus RNA-seq of dorsolateral prefrontal cortex from 1,494 donors (about 6.3 million nuclei, 27 cell subclasses and 65 subtypes), with donor-level annotation of Alzheimer diagnosis, Braak stage, clinical dementia rating, cognitive resilience and neuropsychiatric symptoms including depression and agitation. PMID:42778763 analysed the 584 donors carrying Alzheimer-related phenotypes to score phenotype-associated cells.

Homo sapiens SINGLE CELL RNA SEQ n=1494

dorsolateral prefrontal cortex UBERON:0009834

PMID:42778763

Distributed through the AD Knowledge Portal under its data use requirements. The title is the Synapse project name as registered (DataCite 10.7303/syn60084804). As with SEA-AD, `synapse:` accessions have no open per-record metadata API, so `just verify-datasets` reports this one as UNSUPPORTED. Unlike SEA-AD it carries neuropsychiatric-symptom annotation, which makes it the dataset here that can separate depression-associated cell states from the Alzheimer signal within the same donors.

MEG3 activates necroptosis in human neuron xenografts modeling Alzheimer's disease geo:GSE195455

RNA-seq of human H9-derived neurons seven days after transduction with either a control virus or a MEG3 lentiviral vector, isolating the transcriptional consequence of MEG3 gain of function in human neurons.

Homo sapiens BULK RNA SEQ n=12

PMID:37708272

The causal arm of the necroptosis model, and open access. Its value is as a reference signature: the programme MEG3 overexpression induces in human neurons can be scored against human Alzheimer neurons in the atlases above, which is the actual discriminating test. Note the parent SuperSeries geo:GSE195458 also contains xenograft bulk RNA-seq (geo:GSE195457); the study deposited no human post-mortem cohort of its own, so its claim that MEG3 is raised in patients is not backed by a deposited human dataset.

Ageing-associated myelin dysfunction drives amyloid deposition in mouse models of Alzheimer's disease geo:GSE178304

SuperSeries of mouse single-nucleus and bulk transcriptomics testing whether myelin dysfunction drives amyloid deposition, including grey versus white matter under myelin abnormality and microglial profiling on a 5xFAD background.

Mus musculus MULTI OMICS n=30

PMID:37258678

Mouse only, and deliberately curated as such: this is the data establishing the causal direction (myelin dysfunction to amyloid) that no human observational dataset can establish, and it is the pairing partner for the human oligodendrocyte arm. The human counterpart from the APOE4-myelination work (PMID:36385529) is Synapse syn38120890 and is fully controlled-access with no GEO deposit, so SEA-AD's open oligodendrocyte matrices are the practical human arm instead. Accession resolved against the GEO API with `just verify-datasets`.

Identification of Microproteins in AD Using DIA Mass Spectrometry pride:PXD071500

Raw mass spectrometry underlying the brain microprotein atlas: the HMC3 human microglial experiments and a small data-independent-acquisition dorsolateral prefrontal cortex pilot cohort. The HMC3 runs are the ones this entry depends on - they detect endogenous micro-MKKS63 in a human microglial line while recovering no peptide from the canonical 570-amino-acid MKKS protein, which is the in-cell counterpart of the translational imbalance seen in postmortem cortex.

Homo sapiens PROTEOMICS

dorsolateral prefrontal cortex UBERON:0009834

PMID:42736449

Underlies the Micro-MKKS63 Microprotein Depletion and Microglial Oxidative Phosphorylation Failure nodes. Accession resolved against the PRIDE API with `just verify-datasets`; PRIDE is not a reference-cached prefix, so there is no committed cache file and the check makes a live call on every run. The title is copied verbatim from the PRIDE record and names only the DIA arm, although the deposition also covers the HMC3 runs - the repository's own title is what belongs in this slot, so it is not corrected here. The large ROSMAP and MSBB cohorts that the atlas reanalyzed are not curated as separate dataset records on this entry: they are AMP-AD controlled-access Synapse resources (syn3219045, syn3159438) rather than open accessions, and ROSMAP single-nucleus data are already represented here by synapse:syn26223298.

🔬

Clinical Trials

NCT03486938 PHASE_III COMPLETED

HOPE4MCI: a randomized, double-blind, 78-week trial of AGB101, an extended-release low-dose (220 mg) levetiracetam formulation selected for its ability to normalize heightened hippocampal activity, in 164 participants with mild cognitive impairment due to Alzheimer disease. The largest test to date of the network-hyperexcitability model's therapeutic prediction. ClinicalTrials.gov registers the study under the combined designation PHASE2/PHASE3; the schema enum has no combined value, so `phase` is recorded here as PHASE_III, the pivotal-efficacy designation under which the trial is reported.

Target Phenotypes: Memory impairment HP:0002354

Show evidence (3 references)

PMID:38356475 REFUTE Human Clinical

"The estimated difference between arms is -0.10 (95% CI: -0.85, 0.58), which was not statistically significant."

Negative on the primary endpoint (Clinical Dementia Rating Sum of Boxes) in an unselected population, which is the strongest single constraint on the claim that normalizing network activity is therapeutically actionable.

PMID:38356475 SUPPORT Human Clinical

"One hundred and sixty-four participants were randomized to placebo (n = 83) or AGB101 (n = 81), an extended-release formulation of low dose (220 mg) levetiracetam."

Establishes the trial's size, randomization and intervention.

clinicaltrials:NCT03486938 SUPPORT Other

"Participants will be randomized to receive placebo or AGB101 (220 mg), once daily for 78 weeks."

Trial registration record establishing the intervention, dose and duration.

NCT04063124 PHASE_I COMPLETED

SToMP-AD: an open-label, proof-of-concept phase 1 trial of oral senolytic therapy (dasatinib plus quercetin) in five participants with early symptomatic Alzheimer disease, testing central nervous system penetrance, safety and feasibility rather than efficacy. The first clinical test of the cellular senescence model in Alzheimer disease.

Target Phenotypes: Memory impairment HP:0002354 Impaired executive functioning HP:0033051

Show evidence (2 references)

PMID:37679434 SUPPORT Human Clinical

"In cerebrospinal fluid (CSF), D levels were detected in four participants (80%) ranging from 0.281 to 0.536 ml-1 with a CSF to plasma ratio of 0.422-0.919%; Q was not detected."

Establishes central nervous system penetrance for dasatinib but not quercetin, which is the trial's primary question and a precondition for any brain-directed senolytic strategy.

clinicaltrials:NCT04063124 SUPPORT Other

"The purpose of this pilot study is to evaluate whether a combination of two drugs, dasatinib (D) and quercetin (Q)"

Trial registration record establishing the intervention and pilot design.

NCT05399888 PHASE_II ACTIVE_NOT_RECRUITING

CELIA — a randomized, double-blind, placebo-controlled, parallel-group study of diranersen (BIIB080), a tau-targeting antisense oligonucleotide, in subjects with mild cognitive impairment due to Alzheimer disease or mild Alzheimer disease dementia. Diranersen or placebo is given by intrathecal injection every 12 or 24 weeks over a 76-week placebo-controlled period followed by a long-term extension, with the Clinical Dementia Rating-Sum of Boxes (CDR-SB) as the primary efficacy measure.

Target Phenotypes: Dementia HP:0000726 Memory impairment HP:0002354

Show evidence (2 references)

clinicaltrials:NCT05399888 SUPPORT Human Clinical

"The main question researchers are trying to answer is if diranersen can slow the worsening of AD more than placebo."

Defines CELIA's primary objective — whether the tau-targeting ASO diranersen slows Alzheimer disease progression versus placebo.

clinicaltrials:NCT05399888 SUPPORT Human Clinical

"an injection into the fluid around the spinal cord (cerebrospinal fluid)"

Confirms the intrathecal route of administration used to deliver diranersen to the CNS.

🧫

Experimental Models

HMC3 human microglial micro-MKKS63 CRISPR-Cas9 knockout lines CELL_LINE

The design is what makes this model informative rather than merely convenient. The guides disrupt the smORF while preserving the main MKKS ORF, so a respiratory deficit in these lines is attributable to loss of the 63-amino-acid microprotein and not to loss of the 570-amino-acid MKKS protein - a separation that no whole-locus knockout could make, and the reason this is the experiment that turns a coexpression correlation into a functional claim. Endogenous micro-MKKS63 is detectable by mass spectrometry in unedited HMC3 extracts with no peptide recovered from the canonical MKKS protein, mirroring the pattern in postmortem cortex, so the line reproduces the locus's translational imbalance before anything is edited.

Organism

human NCBITaxon:9606

Cell source

HMC3, an SV40-immortalized human microglial line (ATCC CRL-3304), transduced with lentiCas9-Blast and then with either of two single-guide RNAs disrupting the micro-MKKS63 smORF, or a non-targeting control guide. Marker-positive populations were recovered by fluorescence-activated cell sorting and editing was confirmed by PCR across the cut site followed by Sanger sequencing and TIDE analysis, so the lines are polyclonal edited populations rather than clonal knockouts.

Publication

PMID:42736449

Pathograph links

Microglial Oxidative Phosphorylation Failure perturbs fidelity: low observes: cellular Removes the microprotein and measures respiration, establishing that micro-MKKS63 is required for normal oxidative phosphorylation in human microglial cells.

Limitations: Fidelity is LOW because the model establishes the requirement without modeling the disease state. HMC3 is an immortalized line under standard culture oxygen and glucose, not aged cortical microglia in situ; the knockout abolishes the microprotein where Alzheimer cortex shows a partial decrease of roughly 4% (log2 fold change -0.063); and nothing downstream of respiration - phagocytosis, cytokine output, amyloid handling - was measured, so the model says nothing about whether the deficit matters for the disease.

Divergences from the human mechanism

population mismatch qualifying

HMC3 is an immortalized fetal-derived microglial line maintained in culture. The node describes microglia in the aged human cortex, whose transcriptional state, metabolic substrate availability and disease context differ; cultured microglial lines are known to drift from the in situ microglial phenotype.

cause unrepresented qualifying

The knockout deletes the smORF outright. In Alzheimer cortex micro-MKKS63 is reduced rather than abolished, and whatever process drives that reduction - transcriptional, splicing-level, or translational - is absent from the model. The model therefore represents a complete loss standing in for a partial one, and represents none of its cause.

Readouts

Basal oxygen consumption rate

Seahorse XF Mito Stress Test basal oxygen consumption rate, normalized to total protein per well by bicinchoninic acid assay, in each knockout line against the non-targeting control.

Oxidative phosphorylation GO:0006119

Direction: DECREASED

Interpretation: A reduction in resting respiratory flux is the most direct available measure of the node; both knockout lines reduced it independently, which is what distinguishes the result from a single-clone artifact.

Show evidence (1 reference)

PMID:42736449 SUPPORT DIRECT PRIMARY RESULT In Vitro

"Seahorse Mito Stress Test analysis showed significant reductions in basal oxygen consumption rate (OCR), ATP-linked OCR and maximal respiration across KO lines relative to nontargeting controls"

Reports this measurement as the first of the three named reductions.

ATP-linked oxygen consumption rate

Oligomycin-sensitive fraction of oxygen consumption in the same Mito Stress Test.

Mitochondrial ATP synthesis GO:0042776

Direction: DECREASED

Interpretation: Separates respiration coupled to ATP synthesis from proton leak, so a decrease here locates the deficit in useful energy output rather than in membrane integrity alone.

Show evidence (1 reference)

PMID:42736449 SUPPORT DIRECT PRIMARY RESULT In Vitro

"Seahorse Mito Stress Test analysis showed significant reductions in basal oxygen consumption rate (OCR), ATP-linked OCR and maximal respiration across KO lines relative to nontargeting controls"

Reports this measurement as the second of the three named reductions.

Maximal respiration

Oxygen consumption after uncoupling, reporting the ceiling of the electron transport chain rather than its resting rate.

Oxidative phosphorylation GO:0006119

Direction: DECREASED

Interpretation: A lower ceiling indicates reduced respiratory capacity rather than reduced demand, which is the distinction that makes the basal deficit interpretable as impairment.

Show evidence (1 reference)

PMID:42736449 SUPPORT DIRECT PRIMARY RESULT In Vitro

"Seahorse Mito Stress Test analysis showed significant reductions in basal oxygen consumption rate (OCR), ATP-linked OCR and maximal respiration across KO lines relative to nontargeting controls"

Reports this measurement as the third of the three named reductions.

Reserve respiratory capacity

Spare capacity between basal and maximal respiration, measured in the same assay but reported only in the figure, not in the results text.

Oxidative phosphorylation GO:0006119

Direction: DECREASED

Interpretation: The one readout in this panel where the two knockout lines disagree: reserve capacity falls significantly in the sgRNA1 line but not in the sgRNA2 line, and the results text names only the other three parameters. Recorded because a partially discordant readout is part of the result, and omitting it would make the panel look more uniform than it is.

Show evidence (1 reference)

PMID:42736449 SUPPORT DIRECT PRIMARY RESULT In Vitro

"we generated two independent polyclonal CRISPR–Cas9 KO HMC3 lines using single-guide RNAs (sgRNAs) disrupting the micro-MKKS63 smORF while preserving the mORF"

Attests that this model is informative for the linked node: it is a microprotein-specific loss-of-function in the cell type the node names, with the main ORF left intact so the manipulation is attributable.

Raw mass spectrometry for the HMC3 work is deposited under pride:PXD071500, curated in this entry's datasets block.

Show evidence (1 reference)

PMID:42736449 SUPPORT DIRECT PRIMARY RESULT In Vitro

"MS on HMC3 protein extracts identified robust micro-MKKS63 detection (PXD071500), with no peptides recovered from the canonical 570-aa MKKS protein, mirroring the pattern seen across ROSMAP brain tissue."

Establishes that the unedited line reproduces the translational imbalance seen in human cortex, which is what makes it an appropriate background for a microprotein-specific knockout rather than an arbitrary cell line.

🧮

Computational Models

Digital Alzheimer's Disease Diagnosis (DADD) patient-specific brain model Jupyter notebooks and NumPy arrays Python 3.10; The Virtual Brain; NumPy; SciPy; scikit-learn DIGITAL_TWIN

A 76-region cortical neural-mass model in The Virtual Brain whose synaptic degeneration, structural-connectivity degeneration, and compensatory rewiring parameters are personalized by inversion against task EEG. The inferred parameters serve as interpretable digital biomarkers for early Alzheimer disease risk and progression.

Repository ↗ PMID:40450374

Variable	Model ID	Unit
Connectivity degeneration Severity of large-scale structural-connectivity loss, ranging from 0 to 1.	`cp`	dimensionless
Synaptic degeneration Severity of local synaptic degeneration and cortical excitation-inhibition alteration, ranging from 0 to 1.	`lp`	dimensionless
Neuroplastic connectivity rewiring Compensatory short-range connectivity rewiring parameter.	`np`	dimensionless

Pathograph links

Cortical Structural Connectivity Degeneration measures fidelity: moderate Model inversion estimates patient-specific connectivity degeneration and neuroplastic rewiring from the similarity between simulated and observed EEG signals.

Limitations: Personalization was evaluated in a single-center cohort enriched for subjective cognitive decline; the inferred parameters are model-dependent latent biomarkers rather than direct measurements of white-matter loss, and external multicenter validation was not reported.

Readouts

Connectivity degeneration parameter

Direction: INCREASED

Interpretation: The inferred cp parameter was higher in the subjective-cognitive-decline group than in controls, but it remains an indirect, model-derived measure.

Show evidence (1 reference)

PMID:40450374 SUPPORT Computational

"Statistical analysis (Fig. 2D) confirmed significant differences across groups in connectivity degeneration (U = 1078.5, p = 0.029, d = −0.6) while differences in synaptic degeneration were non-statistically significant (U = 880.0, p = 0.60)."

Reports the group difference for the model-inferred connectivity-degeneration parameter while preserving the negative synaptic-degeneration result.

Neuroplastic connectivity rewiring parameter

Direction: DECREASED

Interpretation: The inferred np parameter was lower in the subjective-cognitive-decline group than in controls, consistent with reduced compensatory rewiring; it remains an indirect model-derived quantity.

Show evidence (1 reference)

PMID:40450374 SUPPORT Computational

"Additionally, neuroplastic rewiring levels were significantly higher in CTR subjects (U = 474.0, p = 0.004, d = 0.8)."

Reports the largest standardized group difference among the personalized DADD parameters and establishes lower rewiring in the SCD group.

Show evidence (1 reference)

PMID:40450374 SUPPORT Computational

"Personalized digital biomarkers reflecting individual neurodegeneration levels were reconstructed through model inversion by maximizing the similarity between experimental and simulated signals."

Establishes how the patient-specific connectivity and degeneration parameters were inferred.

Synaptic Dysfunction measures fidelity: low The personalized lp parameter represents synaptic degeneration through modeled cortical hypo-inhibition and hyperexcitation.

Limitations: The lp parameter was not significantly different between the subjective-cognitive-decline and control groups, and a neural-mass parameter cannot distinguish synapse loss from other causes of altered excitation-inhibition balance.

Readouts

Synaptic degeneration parameter

Direction: UNCHANGED

Interpretation: Patient-specific lp inferred by EEG-constrained model inversion; no consistent group-level direction was established.

Show evidence (1 reference)

PMID:40450374 SUPPORT Computational

"Simulated neurodegeneration is regulated by parameters cp and lp, governing connectivity and synaptic degeneration, respectively."

Defines lp as the model's synaptic-degeneration readout; the null group comparison is encoded by direction and interpretation.

Show evidence (1 reference)

PMID:40450374 SUPPORT Computational

"Biophysically, lp reflects cortical hypo-inhibition and hyperexcitation"

Defines the physiological interpretation of the modeled synaptic-degeneration parameter.

Findings

DADD digital biomarkers distinguished subjective cognitive decline from healthy controls with seven percentage points greater accuracy than standard EEG biomarkers.

Show evidence (1 reference)

PMID:40450374 SUPPORT Computational

"Robustly distinguished SCD from healthy participants, improving classification accuracy by 7% compared to standard EEG biomarkers"

Reports the publication's headline comparison between the DADD-derived and conventional EEG biomarker classifiers.

DADD-derived digital biomarkers identified participants positive for Alzheimer CSF biomarkers with 88% accuracy and predicted clinical cognitive-decline conversion with 87% accuracy.

Show evidence (1 reference)

PMID:40450374 SUPPORT Computational

"Identified patients positive for CSF biomarkers of AD with 88% accuracy (significantly outperforming standard EEG biomarkers, which achieved 58% accuracy); iii) Predicted follow-up conversions to clinical cognitive decline with 87% accuracy (compared to 54% accuracy for standard EEG biomarkers)."

Reports the resampling performance stated in the publication abstract; it is not external prospective validation.

Notes

Repository inspected at commit bf480f5cbba429639fcd6f1cb8dc92081a4f5975 (2025-11-28). It contains three notebooks and a structural-connectivity NumPy array under GPL-2.0, but no pinned dependency manifest, release, or automated tests. The notebooks import The Virtual Brain and scientific-Python packages; execution was not attempted because the repository does not specify compatible package versions. The public notebooks support simulations and model inversion, while patient EEG data are not deposited in the repository.

Show evidence (1 reference)

PMID:40450374 SUPPORT Computational

"We introduce an early diagnostic pipeline based on the Digital Alzheimer's Disease Diagnosis (DADD) digital twin model, which derives personalized AD biomarkers from non-invasive electroencephalographic (EEG) recordings."

Establishes the model as an EEG-personalized Alzheimer disease digital twin.

{ }

Source YAML

click to show

name: Alzheimer Disease
creation_date: '2025-12-04T16:57:31Z'
description: >
  Alzheimer disease is a progressive neurodegenerative disorder characterized by
  cognitive decline, memory loss, and behavioral changes. It is the most common cause
  of dementia, involving the accumulation of amyloid-beta plaques and neurofibrillary
  tangles in the brain, leading to neuronal death and brain atrophy.
disease_term:
  preferred_term: Alzheimer disease
  term:
    id: MONDO:0004975
    label: Alzheimer disease
gene_sets:
- gene_set: MYGENESET:KEGG_ALZHEIMERS_DISEASE
  relationship: CANONICAL_PATHWAY
  note: >-
    KEGG Alzheimer disease pathway.
- gene_set: MYGENESET:WP_ALZHEIMERS_DISEASE
  relationship: CANONICAL_PATHWAY
  note: >-
    WikiPathways Alzheimer disease pathway.
category: Neurodegenerative Disorder
parents:
- Dementia
- Neurodegenerative Disease
has_subtypes:
- name: Early-Onset Alzheimer's Disease
  display_name: Early-Onset Alzheimer's Disease
  description: Type of Alzheimer's that occurs in individuals younger than 65 and is often associated with genetic factors.
  evidence:
  - reference: PMID:30707186
    reference_title: "Early-onset Alzheimer Disease and Its Variants."
    supports: SUPPORT
    snippet: Early-onset Alzheimer disease (AD) is defined as having an age of onset younger than 65 years. ... Early-onset AD comprises about 5% to 6% of cases of AD and includes a substantial percentage of phenotypic variants that differ from the usual amnestic presentation of typical AD. Characteristics of early-onset AD in comparison to late-onset AD include a larger genetic predisposition (familial mutations and summed polygenic risk)
    explanation: This reference clearly supports the definition of Early-Onset Alzheimer's Disease as occurring in individuals younger than 65 and being often associated with genetic factors.
  - reference: PMID:25998117
    reference_title: "Genetic testing and counseling in the diagnosis and management of young-onset dementias."
    supports: SUPPORT
    snippet: Young-onset dementia is hereditary, multifactorial, or sporadic. The most common hereditary dementias include Alzheimer disease.
    explanation: This reference supports the idea that early-onset Alzheimer's disease can be hereditary, aligning with the statement's mention of genetic factors.
- name: Late-Onset Alzheimer's Disease
  display_name: Late-Onset Alzheimer's Disease
  description: The most common form of Alzheimer's, occurring in those aged 65 and older.
  evidence:
  - reference: PMID:24429902
    reference_title: "Etiology and pathogenesis of late-onset Alzheimer's disease."
    supports: SUPPORT
    snippet: In both cases, the disease results in severe cognitive dysfunction, among other problems, and the late-onset form of the disease is now considered to be the most common cause of dementia among the elderly.
    explanation: The literature confirms that late-onset Alzheimer's Disease is the most common form of Alzheimer's and primarily affects those aged 65 and older.
  - reference: PMID:34120901
    reference_title: "Time to Diagnosis in Young Onset Alzheimer's Disease: A Population-Based Study from Central Norway."
    supports: SUPPORT
    snippet: 'BACKGROUND: Young onset dementia is associated with a longer time to diagnosis compared to late onset dementia.'
    explanation: The study differentiates between young onset and late onset Alzheimer's Disease, with late onset typically affecting older adults.
  - reference: PMID:18667359
    reference_title: "Alzheimer's disease and vascular dementia in developing countries: prevalence, management, and risk factors."
    supports: SUPPORT
    snippet: Alzheimer's disease accounts for 60% whereas vascular dementia accounts for approximately 30% of the prevalence. Early-onset familial forms of dementia with single-gene defects occur in Latin America, Asia, and Africa.
    explanation: Supports Alzheimer's disease as the leading cause of dementia (~60% of cases); early-onset familial single-gene forms are noted as a distinct minority category, consistent with late-onset being the predominant form.
prevalence:
- population: Global
  measure_type: POINT_PREVALENCE
  prevalence_class: ABOVE_1_IN_1000
  rate_per_100000: 480
  rate_low: 447
  rate_high: 521
  notes: >-
    An estimated 57.4 million people worldwide were living with dementia in 2019
    (GBD 2019; PMID:34998485), projected to reach 152.8 million by 2050. Alzheimer
    disease accounts for 60–70% of dementia cases. Applying 60–70% to the 2019
    global dementia prevalence of ~745 per 100,000 (57.4 M / 7.7 B population)
    yields an AD-specific point prevalence of approximately 447–521 per 100,000
    (midpoint ~480); this is a derived estimate, not directly quoted from a single source.
  evidence:
  - reference: PMID:34998485
    reference_title: "Estimation of the global prevalence of dementia in 2019 and forecasted prevalence in 2050: an analysis for the Global Burden of Disease Study 2019."
    supports: SUPPORT
    snippet: We estimated that the number of people with dementia would increase from 57.4 (95% uncertainty interval 50.4-65.1) million cases globally in 2019 to 152.8 (130.8-175.9) million cases in 2050.
    explanation: Global Burden of Disease Study 2019 analysis providing country-level estimates of dementia prevalence and forecasts to 2050.
mechanistic_hypotheses:
- hypothesis_group_id: amyloid_cascade_model
  hypothesis_label: Amyloid Cascade Model
  status: CANONICAL
  description: >-
    Amyloid-beta production, oligomerization, and plaque deposition are modeled
    as upstream drivers of tau pathology, synaptic toxicity, glial activation,
    and progressive cognitive decline, especially in APP/PSEN-associated
    familial Alzheimer disease.
  applies_to_subtypes:
  - Early-Onset Alzheimer's Disease
  - Late-Onset Alzheimer's Disease
  evidence:
  - reference: PMID:25941885
    reference_title: "Preventing the spread of Alzheimer's disease neuropathology: a role for calcilytics?"
    supports: SUPPORT
    snippet: "The 'amyloid cascade hypothesis' posits that an extracellular build-up of amyloid-beta oligomers (Abeta-os) and polymers (fibrils) subsequently inducing toxic hyperphosphorylated (p)-Tau oligomers (p-Tau-os) and neurofibrillary tangles starts the sporadic late-onset Alzheimer's disease (LOAD)..."
    explanation: Supports amyloid-beta accumulation as an upstream causal model that induces downstream tau pathology.
  - reference: PMID:22351073
    reference_title: "Synaptic dysfunction in Alzheimer's disease."
    supports: SUPPORT
    snippet: "Generation of amyloid peptide (Abeta) is at the beginning of a cascade that leads to Alzheimer's disease (AD)... soluble assembly states of Abeta peptides can cause cognitive problems by disrupting synaptic function..."
    explanation: Supports amyloid-beta generation as an initiating cascade event with synaptic consequences.
  - reference: PMID:39496485
    reference_title: "The Icelandic Mutation (APP-A673T) Is Protective against Amyloid Pathology In Vivo."
    supports: SUPPORT
    evidence_source: MODEL_ORGANISM
    snippet: "APP-A673T significantly downregulated β-cleavage and attenuated the production of Aβ and amyloid pathology in the brains of these animals."
    explanation: >-
      In vivo arm of the bidirectional genetic argument: introducing the protective
      Icelandic APP variant into Swedish-mutation-free App knock-in mice reduces
      beta-cleavage and amyloid pathology, complementing the causal familial
      mutations at the other end of the dose range.
  - reference: PMID:42128444
    reference_title: "Clinical Meaningfulness of Donanemab in Early Symptomatic Alzheimer Disease: Data From the Randomized Phase 3 TRAILBLAZER-ALZ 2 Trial."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "Donanemab reduced the risk of progression from MCI to mild AD by 33% (hazard ratio [HR] = 0.67; 95% CI 0.52-0.87; p = 0.003) and from mild to moderate AD by 50% (HR = 0.50; 95% CI 0.33-0.78; p = 0.002)."
    explanation: >-
      Partial pharmacological validation of the amyloid target: removing plaque
      amyloid slows clinical stage transition. Note this identifier is a
      clinical-meaningfulness analysis of TRAILBLAZER-ALZ 2 combining prespecified
      and post hoc analyses, not the primary trial report.
  - reference: PMID:41985900
    reference_title: "Amyloid-beta-targeting monoclonal antibodies for people with mild cognitive impairment or mild dementia due to Alzheimer's disease."
    supports: REFUTE
    evidence_source: OTHER
    snippet: "Successful removal of amyloid from the brain does not seem to be associated with clinically meaningful effects in people with mild cognitive impairment or mild dementia due to Alzheimer's disease."
    explanation: >-
      Cochrane review of 17 randomized trials in 20,342 participants; the effect on
      cognition and dementia severity at 18 months is graded trivial. Removing the
      proposed causal species does not produce a clinically meaningful benefit,
      which is the strongest available qualification of the cascade's sufficiency.
  - reference: PMID:39887500
    reference_title: "Biomarker treatment effects in two phase 3 trials of gantenerumab."
    supports: REFUTE
    evidence_source: HUMAN_CLINICAL
    snippet: "Gantenerumab significantly reduced brain amyloid load. Tau positron emission tomography showed no treatment effect in a small subset of participants."
    explanation: >-
      Human interventional evidence bearing on the amyloid-to-tau step: 116 weeks of
      amyloid removal moved amyloid PET and multiple fluid biomarkers but did not move
      tau PET, constraining a simple unidirectional amyloid-drives-tau reading. The
      tau PET substudy was small, so this bounds rather than excludes the link.
  - reference: PMID:37433038
    reference_title: "A data-driven study of Alzheimer's disease related amyloid and tau pathology progression."
    supports: REFUTE
    evidence_source: HUMAN_CLINICAL
    snippet: "In the amyloid-first subtype, extensive neocortical amyloid-β precedes the spread of tau beyond the medial temporal lobe, while in the tau-first subtype, mild tau accumulates in medial temporal and neocortical areas prior to interacting with amyloid-β."
    explanation: >-
      Data-driven subtyping of ADNI tau/amyloid PET and ROSMAP neuropathology
      recovers a tau-first subtype in both, so amyloid does not initiate the sequence
      in every case. Refutes universality, not the cascade itself — the source paper
      is explicit that the hypothesis does not require amyloid to occur first in all
      cases, and tau-first APOE e4 non-carriers largely recapitulate primary
      age-related tauopathy rather than Alzheimer disease.
  - reference: PMID:38685682
    reference_title: "Lessons learned from the failure of solanezumab as a prospective treatment strategy for Alzheimer's disease."
    supports: REFUTE
    evidence_source: OTHER
    snippet: "Unfortunately, this compound showed to accelerate cognitive decline in both asymptomatic and symptomatic trial participants, and this failure of solanezumab further questioned the Aβ cascade hypothesis of AD."
    explanation: >-
      Targeting soluble monomeric amyloid-beta did not help and was judged to worsen
      decline. Graded OTHER deliberately: this sentence is the authors' expert-opinion
      summary of the solanezumab programme in a narrative case history, not a trial
      result, and the underlying EXPEDITION, A4 or DIAN-TU reports should be cited for
      any quantitative claim.
  - reference: PMID:41942821
    reference_title: "Contribution of health history and neuropathologic changes to the likelihood of dementia in those with intermediate/high Alzheimer's pathology: findings from The 90 + Study."
    supports: REFUTE
    evidence_source: HUMAN_CLINICAL
    snippet: "Among 235 participants (mean age at death = 98 years, 70% women), 33% maintained normal cognition."
    explanation: >-
      Amyloid pathology is not deterministic for clinical disease: a third of oldest-old
      people with intermediate or high Alzheimer neuropathologic change died cognitively
      normal. The fraction falls to 22% in the high-change subgroup, so the headline
      number depends on including intermediate change.
  - reference: PMID:41942821
    reference_title: "Contribution of health history and neuropathologic changes to the likelihood of dementia in those with intermediate/high Alzheimer's pathology: findings from The 90 + Study."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "In this study, the presence of neurodegenerative neuropathologic changes other than ADNC and the absence of hypertension distinguish oldest old individuals with dementia from those with normal cognition."
    explanation: >-
      Carries the study's own explanation for the resilience above: co-pathology
      (LATE-NC, Lewy body disease, hippocampal sclerosis, a high count of non-Alzheimer
      changes), not tolerance of amyloid, separates the demented from the cognitively
      normal. Curated alongside the 33% figure so the number is not read as evidence
      for a synaptic or metabolic resilience mechanism.
  - reference: PMID:31043483
    reference_title: "Lysosomal Dysfunction in Down Syndrome Is APP-Dependent and Mediated by APP-βCTF (C99)."
    supports: REFUTE
    evidence_source: IN_VITRO
    snippet: "The modest elevation of endogenous APP-βCTF needed to disrupt lysosomal function in DS is relevant to sporadic AD where APP-βCTF, but not APP, is also elevated."
    explanation: >-
      An APP-dependent but amyloid-beta-independent pathogenic arm: the beta-cleaved
      C-terminal fragment C99, not amyloid-beta, mediates the lysosomal acidification
      failure, and C99 rather than APP is what is elevated in sporadic disease. Scoped
      to lysosomal acidification in a Down syndrome model (fibroblasts, cultured Ts2
      neurons, and Bace1 allele deletion in vivo), not to Alzheimer pathogenesis
      generally.
  notes: >-
    Retained as CANONICAL, but not sufficient as a complete disease explanation:
    tau burden, vascular injury, immune state, lysosomal clearance, aging, and
    genetic background modulate how amyloid pathology maps to symptoms.
    Best read as a necessary rather than sufficient initiating trigger, applying
    most strongly to familial disease and to the amyloid-first sporadic subtype.
    Four independent lines qualify it and are curated above: a tau-first
    spatiotemporal subtype (PMID:37433038), the Cochrane finding that amyloid
    removal produces no clinically meaningful benefit (PMID:41985900), oldest-old
    resilience attributable to co-pathology rather than amyloid tolerance
    (PMID:41942821), and an APP-dependent C99 endolysosomal arm that does not run
    through amyloid-beta (PMID:31043483). Assessed from the OpenScientist
    amyloid_cascade_model report (kb/hypotheses/Alzheimer_Disease/); that report's
    claim that BACE1 inhibitors raise C99 is backwards and was not curated —
    BACE1 generates C99, and the cited source reports that BACE1 inhibition
    rescues the cathepsin deficits.
- hypothesis_group_id: tau_neurodegeneration_model
  hypothesis_label: Tau Neurodegeneration Model
  status: CANONICAL
  description: >-
    Tau hyperphosphorylation and aggregation into neurofibrillary tangles are
    modeled as proximate drivers of neuronal dysfunction, microtubule disruption,
    neurodegeneration, and clinical progression.
  applies_to_subtypes:
  - Early-Onset Alzheimer's Disease
  - Late-Onset Alzheimer's Disease
  evidence:
  - reference: PMID:21509508
    reference_title: "Tau mediated neurodegeneration: an insight into Alzheimer's disease pathology."
    supports: SUPPORT
    snippet: "Extracellular accumulations of Abeta, hyperphosphorylation of tau and intracellular neurofibrillary tangle formation have been the hallmarks of Alzheimer's Disease (AD)."
    explanation: Supports tau hyperphosphorylation and tangle formation as central Alzheimer disease pathology.
  - reference: PMID:19542604
    reference_title: "The microtubule-associated protein tau is also phosphorylated on tyrosine."
    supports: SUPPORT
    snippet: "Tau protein is the principal component of the neurofibrillary tangles found in Alzheimer's disease (AD), where it is hyperphosphorylated on serine and threonine residues."
    explanation: Identifies hyperphosphorylated tau as the principal neurofibrillary tangle component in Alzheimer disease.
  - reference: PMID:39211286
    reference_title: "Phosphorylation at serine 214 correlates with tau seeding activity in an age-dependent manner in two mouse models for tauopathies and is required for tau transsynaptic propagation."
    supports: SUPPORT
    evidence_source: MODEL_ORGANISM
    snippet: "Our findings indicate that tau seeding is an early event that occurs well before the appearance of AT8-positive NFT."
    explanation: >-
      Places seed-competent soluble tau upstream of the tangles the model names,
      and the same study shows that blocking Ser214 phosphorylation reduces
      propagation in mouse primary neurons and seeding in a Drosophila tauopathy,
      supplying a perturbation arm. Two mouse models plus Drosophila only.
  - reference: PMID:40734174
    reference_title: "Axonal tau reduction ameliorates tau and amyloid pathology in a mouse model of Alzheimer's disease."
    supports: SUPPORT
    evidence_source: MODEL_ORGANISM
    snippet: "Brains of 5×FAD mice deficient for hnRNP R contained less phospho-tau aggregates and amyloid-β plaques in the cortex and hippocampus."
    explanation: >-
      Lowering axonal tau — genetically, or with a MAPT antisense oligonucleotide
      that blocks hnRNP R binding to Mapt — reduces both phospho-tau and amyloid
      plaque load and prevents neurodegeneration, so the tau-amyloid relation runs
      in both directions rather than only downstream of amyloid.
  - reference: PMID:35534858
    reference_title: "Changes in glial cell phenotypes precede overt neurofibrillary tangle formation, correlate with markers of cortical cell damage, and predict cognitive status of individuals at Braak III-IV stages."
    supports: REFUTE
    evidence_source: HUMAN_CLINICAL
    snippet: "We found distinct glial responses with increased proinflammatory and decreased homeostatic markers, both in regions with tau tangles (temporal pole) and without overt tau deposits (visual cortex) in demented but not in resilient."
    explanation: >-
      In 55 brains matched at Braak III-IV, tangle burden alone does not determine
      clinical expression: the demented and the cognitively resilient are separated
      by glial phenotype, including in visual cortex that carries no overt tau at
      those stages. Refutes tangle burden as a sufficient proximate driver. The
      design is cross-sectional post-mortem, so the co-factor is an association and
      its necessity is not established — see the toxic-species knowledge gap.
  - reference: PMID:38012285
    reference_title: "TANGO: a placebo-controlled randomized phase 2 study of efficacy and safety of the anti-tau monoclonal antibody gosuranemab in early Alzheimer's disease."
    supports: REFUTE
    evidence_source: HUMAN_CLINICAL
    snippet: "No significant effects were identified in cognitive and functional tests as no dose resulted in a favorable change from the baseline CDR-SB score at week 78 compared to placebo control"
    explanation: >-
      Lowering extracellular N-terminal tau in 654 participants with early Alzheimer
      disease did not slow cognitive or functional decline, despite significant
      target engagement in cerebrospinal fluid. A translational gap for a hypothesis
      curated as CANONICAL.
  - reference: PMID:37643887
    reference_title: "Randomized Phase II Study of the Safety and Efficacy of Semorinemab in Participants With Mild-to-Moderate Alzheimer Disease: Lauriet."
    supports: REFUTE
    evidence_source: HUMAN_CLINICAL
    snippet: "Based on the results of the prespecified coprimary endpoints, this study was negative. While semorinemab had a significant effect on cognition measured by the ADAS-Cog11, this effect did not extend to improved functional or global outcomes."
    explanation: >-
      Second anti-tau antibody failure at a later disease stage; the ADAS-Cog11
      signal (42.2% reduction in decline, p = 0.0008) did not carry to the ADCS-ADL
      coprimary endpoint, MMSE, or CDR-SB, so the trial is negative on its own
      prespecified terms.
- hypothesis_group_id: synaptic_failure_convergence_model
  hypothesis_label: Synaptic Failure Convergence Model
  status: CANONICAL
  description: >-
    Amyloid-beta, tau, inflammatory, vascular, oxidative, and infectious
    stressors converge on synaptic plasticity, neurotransmitter release, and
    network function, producing cognitive decline.
  applies_to_subtypes:
  - Early-Onset Alzheimer's Disease
  - Late-Onset Alzheimer's Disease
  evidence:
  - reference: PMID:27662312
    reference_title: "Stress-Induced Synaptic Dysfunction and Neurotransmitter Release in Alzheimer's Disease: Can Neurotransmitters and Neuromodulators be Potential Therapeutic Targets?"
    supports: SUPPORT
    snippet: "Compelling evidence suggests that soluble amyloid-beta (Abeta) and hyperphosphorylated tau serve as toxins in the dysfunction of synaptic plasticity and aberrant neurotransmitter (NT) release at synapses consequently causing a cognitive decline in Alzheimer's disease (AD)."
    explanation: Directly supports synaptic dysfunction as a convergence point downstream of amyloid-beta and tau toxicity.
  - reference: PMID:12973746
    reference_title: "Glutamatergic systems in Alzheimer's disease."
    supports: SUPPORT
    snippet: "Histological studies indicate loss of pyramidal neurones and their synapses in Alzheimer's disease (AD), this together with biochemical evidence suggests presynaptic (and postsynaptic) glutamatergic hypoactivity."
    explanation: Supports loss of synapses and altered neurotransmission in Alzheimer disease.
  - reference: PMID:40491249
    reference_title: "Cerebrospinal fluid and brain positron emission tomography measures of synaptic vesicle glycoprotein 2A: Biomarkers of synaptic density in Alzheimer's disease."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "CSF SV2A was lower in AD compared to CN participants. Within the AD group, CSF SV2A was highly correlated with SV2A PET."
    explanation: >-
      Concordant fluid and imaging measurement of synaptic density itself rather
      than of its inferred consequences. Small: 21 participants with early Alzheimer
      disease and 7 cognitively normal controls, all scanned with carbon-11 UCB-J.
      Other cerebrospinal-fluid synaptic proteins were higher in Alzheimer disease
      and unrelated to SV2A PET, which sharpens the specificity of the SV2A measure.
  - reference: PMID:41406948
    reference_title: "Plasma growth-associated protein 43 correlates with synaptic loss in Alzheimer's disease."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "Elevated plasma GAP-43 is associated with greater tau aggregation, faster brain atrophy, and accelerated cognitive decline, particularly among cognitively unimpaired individuals."
    explanation: >-
      A blood-based marker of the convergence endpoint, in two Chinese cohorts
      totalling 785 participants with SV2A PET in 85, tau PET in 280 and MRI in 595.
      Ties synaptic degeneration to tau burden and to subsequent decline, and does so
      earliest in people who are still cognitively unimpaired.
  - reference: PMID:40465636
    reference_title: "Examining the association between synaptic density and neurofibrillary tau among cognitively impaired and unimpaired older adults with and without Alzheimer's disease pathology."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "Hp NFT burden strongly correlated with synaptic density, while the ERC showed weaker effects."
    explanation: >-
      Quantifies the tau-to-synapse arm in living people (94 older adults imaged with
      carbon-11 UCB-J, carbon-11 PiB and fluorine-18 MK-6240). The companion result —
      hippocampal synaptic density is similar in controls and in cognitively
      unimpaired people with biological Alzheimer disease — places synapse loss at
      the pathology-to-symptom step rather than at pathology onset.
  - reference: PMID:41454064
    reference_title: Impaired glymphatic function is associated with synaptic loss in cognitive impairment.
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "A lower ALPS index, indicating greater glymphatic impairment, was significantly associated with reduced synaptic density."
    explanation: >-
      First quantitative link in this entry between the separately curated
      glymphatic_clearance_model and synaptic convergence, in 182 participants.
      Curate as an association only: the design is cross-sectional with mediation
      modelling, and DTI-ALPS is an imaging surrogate for glymphatic function rather
      than a direct measure of it.
  - reference: PMID:40731189
    reference_title: Gut-derived bacterial vesicles carrying lipopolysaccharide promote microglia-mediated synaptic pruning.
    supports: SUPPORT
    evidence_source: MODEL_ORGANISM
    snippet: "These bEVs activated microglial Piezo1, consequently precipitating an excessive synaptic pruning process mediated by the C1q-C3 complement pathway."
    explanation: >-
      A mechanistically specific peripheral-to-central route into the complement arm
      of the convergence model: lipopolysaccharide-carrying bacterial extracellular
      vesicles, elevated in patient plasma, cross the blood-brain barrier and drive
      C1q-C3 pruning through microglial Piezo1. Single study, so an emerging edge
      rather than an established one; see the separate
      complement_synaptic_pruning_model block for the downstream pathway.
  - reference: PMID:38829680
    reference_title: "Cognitive integrity in Non-Demented Individuals with Alzheimer's Neuropathology is associated with preservation and remodeling of dendritic spines."
    supports: REFUTE
    evidence_source: HUMAN_CLINICAL
    snippet: "NDAN subjects displayed a higher spine density in regions distant from Aβ plaques versus AD patients."
    explanation: >-
      Convergence on synaptic failure is not obligatory: people with Alzheimer
      neuropathology who remained non-demented preserve dendritic spine density away
      from plaques, with a predominance of immature spines and higher Pin1 than in
      Alzheimer dementia. Human post-mortem DiI labelling and descriptive, so it
      supports a resilience account rather than a causal edge.
- hypothesis_group_id: neuroimmune_glial_amplification_model
  hypothesis_label: Neuroimmune-Glial Amplification Model
  status: ALTERNATIVE
  description: >-
    Microglia, astrocytes, complement, and inflammasome pathways are modeled as
    stage-dependent disease-modifying mechanisms that can respond to amyloid-beta
    and tau pathology, reinforce amyloid aggregation and inflammatory injury, or
    restrain amyloid-associated tau spread through protective phagocytic and
    plaque-barrier functions.
  applies_to_subtypes:
  - Early-Onset Alzheimer's Disease
  - Late-Onset Alzheimer's Disease
  evidence:
  - reference: PMID:28019679
    reference_title: "Inflammasome activation and innate immunity in Alzheimer's disease."
    supports: SUPPORT
    snippet: "NOD-like receptor (NLR) family, pyrin domain containing 3 and 1 inflammasomes, present in myeloid cells and neurons, respectively, represent key components of the innate immune reaction observed in Alzheimer patient brains."
    explanation: Supports inflammasome-mediated innate immune activation in Alzheimer disease brains.
  - reference: PMID:23930978
    reference_title: "Microglia, neuroinflammation, and beta-amyloid protein in Alzheimer's disease."
    supports: SUPPORT
    snippet: "A vicious cycle of inflammation has been formed between Abeta accumulation, activated microglia, and microglial inflammatory mediators, which enhance Abeta deposition and neuroinflammation."
    explanation: Supports a self-reinforcing loop between amyloid-beta deposition and microglial inflammation.
  - reference: PMID:36361780
    reference_title: "Microglia and Alzheimer's Disease."
    supports: SUPPORT
    evidence_source: OTHER
    snippet: "Microglia work to reestablish efficiency and stop further degeneration in the early stages of AD but mainly fail in the illness's later phases."
    explanation: >-
      Review-level support for treating microglial activation as stage-dependent
      rather than uniformly protective or uniformly damaging.
  - reference: PMID:34100905
    reference_title: "Activated microglia mitigate Aβ-associated tau seeding and spreading."
    supports: SUPPORT
    evidence_source: MODEL_ORGANISM
    snippet: "We show that both TREM2KO and microglial ablation dramatically enhance tau seeding and spreading around plaques."
    explanation: >-
      Qualifies the amplification model by showing that TREM2-dependent activated
      microglia can restrain amyloid-associated tau propagation in mouse models.
  - reference: PMID:41633359
    reference_title: "Repression of RIPK1 kinase by INPP5D inhibits expression of diverse proinflammatory mediators and late-onset Alzheimer's disease risk factors."
    supports: SUPPORT
    evidence_source: MODEL_ORGANISM
    snippet: "Microglial INPP5D deficiency cell-autonomously promoted RIPK1-mediated transcriptional induction of diverse LOAD risk genes, proinflammatory cytokines, complements, and ROS mediators"
    explanation: >-
      Connects a late-onset Alzheimer risk gene to the inflammasome: the INPP5D SH2
      domain binds RIPK1 at p-Y383 to suppress its kinase activation, and losing that
      restraint induces Nlrp3, gasdermin D, complement and Toll-like receptor
      signalling in microglia. The most mechanistically specific genetic entry point
      into this hypothesis in the entry.
  - reference: PMID:41789727
    reference_title: "HDAC6 regulates BACE1 stability and NLRP3 inflammasome activation in Alzheimer's disease."
    supports: SUPPORT
    evidence_source: MODEL_ORGANISM
    snippet: "HDAC6 deficiency in 5xFAD mice reduced BACE1 accumulation, Aβ deposition, ASC speck formation, and IL-1β levels, accompanied by improved cognitive performance."
    explanation: >-
      A convergence node between amyloid production and inflammasome activation:
      HDAC6 both stabilises BACE1 by deacetylating K501 and facilitates NLRP3
      activation, so removing it lowers amyloid and inflammasome readouts together.
  - reference: PMID:39194073
    reference_title: Amyloid-β-activated microglia can induce compound proteinopathies.
    supports: SUPPORT
    evidence_source: MODEL_ORGANISM
    snippet: "transplantation of amyloid-β oligomer-activated microglia into the striatum of naïve mice was sufficient to generate all the features of Alzheimer's disease, including widespread tauopathy and synucleinopathy, gliosis, neuroinflammation, synapse loss, neuronal death, and cognitive and motor deficits"
    explanation: >-
      The strongest transfer-of-sufficiency result for microglia as an amplifier:
      the activated cells alone, moved into a naive brain, generate the downstream
      proteinopathies, and the pathology is abolished by microglial depletion or
      anti-inflammatory treatment. Mouse only.
  - reference: PMID:32544086
    reference_title: Impact of TREM2R47H variant on tau pathology-induced gliosis and neurodegeneration.
    supports: SUPPORT
    evidence_source: MODEL_ORGANISM
    snippet: "PS19-TREM2R47H mice had significantly attenuated brain atrophy and synapse loss versus PS19-TREM2CV mice."
    explanation: >-
      The mirror image of the Gratuze result above: the same TREM2 risk variant that
      worsens amyloid-associated tau spread is protective against tau-driven atrophy
      and synapse loss, with reduced microglial phagocytosis of postsynaptic elements
      seen in human Alzheimer brain as well. Direct evidence that the microglial
      contribution changes sign with the pathology present.
  - reference: PMID:40747577
    reference_title: "Early microglial and astrocyte reactivity in preclinical Alzheimer's disease."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "Plasma glial fibrillary acidic protein (GFAP) and cerebrospinal fluid (CSF) soluble triggering receptor expressed on myeloid cells 2 (sTREM2) were increased in preclinical AD."
    explanation: >-
      Places glial reactivity in the preclinical phase in humans: 211 cognitively
      unimpaired participants, with CSF sTREM2 also influencing p-tau181 and
      neurogranin. Cross-sectional structural equation modelling, so it establishes
      early timing but not causal direction.
  - reference: PMID:19589092
    reference_title: "An update on the efficacy of non-steroidal anti-inflammatory drugs in Alzheimer's disease."
    supports: REFUTE
    evidence_source: OTHER
    snippet: "Once the Abeta deposition process has started, NSAIDs are no longer effective and may even be detrimental because of their inhibiting activity on activated microglia of the AD brain, which mediates Abeta clearance and activates compensatory hippocampal neurogenesis."
    explanation: >-
      Records the failure of long-term NSAID and COX-2 trials, including the negative
      rofecoxib secondary-prevention and naproxen/celecoxib primary-prevention
      studies, which refutes a simple anti-inflammatory therapeutic reading of this
      hypothesis. Graded OTHER: this is a 2009 review, and the stage-dependence
      account in the quoted sentence is the author's hypothesis rather than a
      demonstrated result.
  notes: >-
    Retained as ALTERNATIVE despite strong GWAS support because microglial and
    astrocyte programs appear context-dependent: the same innate immune pathways
    can mediate protective amyloid/tau clearance or pathogenic complement,
    inflammasome, and synapse-loss amplification depending on disease stage,
    cell state, and genetic background. ALTERNATIVE is the right value for exactly
    what the OpenScientist neuroimmune_glial_amplification_model report concluded —
    an established disease-amplifying mechanism that is not a sufficient standalone
    causal model — and its proposed status values (PARTIALLY SUPPORTED, DISEASE
    MODIFIER) are not in MechanisticHypothesisStatusEnum. That report's
    microglia-to-A1-astrocyte citation (PMID:41345807) was rejected in assessment
    and deliberately not curated: it is a neuroHIV study of HIV-1 Tat whose
    Alzheimer content is background restatement of the Liddelow paradigm.
- hypothesis_group_id: vascular_bbb_clearance_model
  hypothesis_label: Vascular and Blood-Brain Barrier Clearance Model
  status: ALTERNATIVE
  description: >-
    Cerebral blood flow changes, neurovascular-unit injury, blood-brain barrier
    dysfunction, and impaired amyloid-beta clearance are modeled as causal or
    reinforcing contributors to Alzheimer disease progression.
  applies_to_subtypes:
  - Late-Onset Alzheimer's Disease
  evidence:
  - reference: PMID:28902142
    reference_title: "Blood-Brain Barrier Dysfunction and the Pathogenesis of Alzheimer's Disease."
    supports: SUPPORT
    snippet: "Thus, current evidence suggests that BBB dysfunction may causatively and consequently contribute to AD pathogenesis, forming a vicious cycle between brain Abeta accumulation and neurovascular unit impairments during disease progression."
    explanation: Supports a bidirectional causal cycle between blood-brain barrier dysfunction, amyloid-beta accumulation, and neurovascular injury.
  - reference: PMID:26898552
    reference_title: "The Utility of Cerebral Blood Flow as a Biomarker of Preclinical Alzheimer's Disease."
    supports: SUPPORT
    snippet: "There is accumulating evidence suggesting that changes in brain perfusion are present long before the clinical symptoms of Alzheimer's disease (AD), perhaps even before amyloid-beta accumulation or brain atrophy."
    explanation: Supports early cerebral perfusion changes as part of the vascular model.
  - reference: PMID:25757756
    reference_title: "Accelerated pericyte degeneration and blood-brain barrier breakdown in apolipoprotein E4 carriers with Alzheimer's disease."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "we show accelerated pericyte degeneration in AD APOE4 carriers >AD APOE3 carriers >non-AD controls, which correlates with the magnitude of BBB breakdown to immunoglobulin G and fibrin"
    explanation: >-
      Human post-mortem molecular mechanism for the barrier arm, which this
      hypothesis otherwise stated only at review level. Cyclophilin A and MMP-9
      accumulate in pericytes and endothelial cells in an APOE4-graded fashion; LRP1
      was reduced similarly in APOE4 and APOE3 carriers, so the genotype effect runs
      through the CypA-MMP-9 route rather than through LRP1 abundance.
  - reference: PMID:30041013
    reference_title: The concerted amyloid-beta clearance of LRP1 and ABCB1/P-gp across the blood-brain barrier is linked by PICALM.
    supports: SUPPORT
    evidence_source: IN_VITRO
    snippet: "Late-onset AD risk factor Phosphatidylinositol binding clathrin assembly protein (PICALM) is associated with both ABCB1/P-gp and LRP1 representing a functional link and guiding both proteins through the brain endothelium."
    explanation: >-
      Routes a common-variant risk locus directly into barrier amyloid-beta efflux by
      linking the two clearance proteins, which is what makes this model more than a
      physiological description. Molecular and in vitro, so curated as a mechanism
      rather than as human causal evidence.
  - reference: PMID:28667120
    reference_title: "Chronic cerebral hypoperfusion alters amyloid-β peptide pools leading to cerebral amyloid angiopathy, microinfarcts and haemorrhages in Tg-SwDI mice."
    supports: SUPPORT
    evidence_source: MODEL_ORGANISM
    snippet: "A significant increase in soluble Aβ peptides (Aβ40/42) was detected after 1 month of hypoperfusion in the parenchyma in parallel with elevated APP and APP proteolytic products."
    explanation: >-
      Interventional evidence that hypoperfusion is upstream of amyloid rather than
      downstream: bilateral carotid stenosis raises soluble amyloid within a month,
      with insoluble amyloid, microinfarcts and haemorrhage by three months. Tg-SwDI
      is a cerebral amyloid angiopathy model, which limits generalisation to
      parenchymal-predominant human disease.
  - reference: PMID:40884970
    reference_title: "A comprehensive Investigation of the associations between cerebral blood flow and cerebrospinal fluid biomarkers across the Alzheimer's disease continuum."
    supports: REFUTE
    evidence_source: HUMAN_CLINICAL
    snippet: "In early stages, tau pathology is linked to hypoperfusion, whereas in established AD, elevated tau levels are paradoxically associated with increased CBF-potentially reflecting compensatory or neuroinflammatory mechanisms."
    explanation: >-
      Constrains this hypothesis's assumption of monotonic perfusion decline: across
      416 ADNI participants the tau-flow association inverts by stage, negative in
      mild cognitive impairment and positive in established disease. Cross-sectional,
      so the stage pattern is between-group rather than within-subject.
  notes: >-
    Retained as ALTERNATIVE. The mechanistic evidence is strong and now molecular
    (APOE4-graded pericyte degeneration through cyclophilin A and MMP-9, PICALM
    linking the two endothelial clearance routes, hypoperfusion raising amyloid in
    a transgenic model), but the human causal proof is missing: no interventional
    trial has shown that correcting vascular dysfunction prevents or slows amyloid
    accumulation, and the rrAD trial (NCT02913664) is pending with neurocognitive
    rather than serial amyloid PET primary outcomes. The model's reach is also
    bounded at tau — see the vascular-to-tau knowledge gap. Assessed from the
    OpenScientist vascular_bbb_clearance_model report; that report's ARIA-E
    incidence figure of 22-36% is not in the Cochrane review it cites
    (PMID:41985900 reports an absolute risk difference of 107 more per 1000) and
    was deliberately not curated, nor was its CL:0000669 lead, which names the
    pericyte cell type rather than pericyte death.
- hypothesis_group_id: glymphatic_clearance_model
  hypothesis_label: Glymphatic Clearance Model
  status: EMERGING
  description: >-
    Failure of the sleep-dependent perivascular (glymphatic) clearance route is
    modeled as an upstream contributor to amyloid-beta accumulation, acting
    through reduced CSF-interstitial fluid exchange rather than through
    transcytotic blood-brain-barrier efflux (vascular_bbb_clearance_model) or
    intracellular degradation (autophagy_lysosomal_clearance_model). Held as
    EMERGING because the human evidence is dominated by imaging surrogates and
    small intrathecal-tracer cohorts, and because the microscopic transport mode
    underpinning the mechanism is itself contested; see the
    glymphatic_dysfunction module for the curated controversy.
  applies_to_subtypes:
  - Late-Onset Alzheimer's Disease
  evidence:
  - reference: PMID:27893874
    reference_title: "Association of Perivascular Localization of Aquaporin-4 With Cognition and Alzheimer Disease in Aging Brains."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: >-
      When controlling for age, loss of perivascular AQP4 localization was
      associated with increased amyloid-β burden
    explanation: >-
      Human postmortem dose-response between loss of the glymphatic conduit
      protein and amyloid burden, the core claim of this hypothesis group.
  - reference: PMID:33004510
    reference_title: "Glymphatic failure as a final common pathway to dementia."
    supports: SUPPORT
    evidence_source: OTHER
    snippet: >-
      The glymphatic system, which clears the brain of protein waste products,
      is mostly active during sleep.
    explanation: >-
      Establishes the sleep-gated clearance premise that distinguishes this
      hypothesis group from the other clearance models in this entry.
  - reference: PMID:41762118
    reference_title: Early glymphatic failure in AppNL-F knock-in mice is linked to parenchymal border macrophages loss.
    supports: SUPPORT
    evidence_source: MODEL_ORGANISM
    snippet: "AppNL-F mice showed reductions in glymphatic influx and clearance at 6 months, preceding substantial Aβ plaque deposition."
    explanation: >-
      Temporal ordering in a knock-in model that avoids overexpression artefacts:
      the clearance deficit appears before substantial plaque, and tracks loss of
      parenchymal border macrophages, which acute intracisternal amyloid-beta alone
      reproduces in wild-type mice.
  - reference: PMID:41762118
    reference_title: Early glymphatic failure in AppNL-F knock-in mice is linked to parenchymal border macrophages loss.
    supports: REFUTE
    evidence_source: MODEL_ORGANISM
    snippet: "Four weeks of systemic anti-Aβ antibody treatment efficiently reduced Aβ plaque load and rescued PBMs in some brain regions; however, the treatment did not restore glymphatic function in the AppNL-F model."
    explanation: >-
      Curated alongside the item above because it breaks the return arm of the
      vicious cycle this hypothesis usually assumes: clearing plaque did not restore
      glymphatic function, and the authors conclude the deficit is not driven by
      parenchymal plaque load. The two directions of the loop are therefore not
      symmetric, and the same paper supplies both.
  notes: >-
    Deliberately kept separate from vascular_bbb_clearance_model. The two
    describe different anatomical routes for amyloid-beta egress (perivascular
    CSF-ISF exchange versus receptor-mediated transcytosis across the
    blood-brain barrier) and should not be merged even though both are
    "clearance" models. PMID:41762118 is curated here rather than on
    vascular_bbb_clearance_model, where the OpenScientist report placed it, for
    that reason: it measures perivascular influx and clearance, not barrier
    transcytosis.
- hypothesis_group_id: autophagy_lysosomal_clearance_model
  hypothesis_label: Autophagy-Lysosomal Clearance Model
  status: EMERGING
  description: >-
    Defective autophagic flux, lysosomal transport, mitophagy, and aggregate
    clearance are modeled as mechanisms that promote amyloid-beta accumulation,
    tau accumulation, mitochondrial stress, and downstream cognitive impairment.
  applies_to_subtypes:
  - Early-Onset Alzheimer's Disease
  - Late-Onset Alzheimer's Disease
  evidence:
  - reference: PMID:20541250
    reference_title: "Lysosomal proteolysis and autophagy require presenilin 1 and are disrupted by Alzheimer-related PS1 mutations."
    supports: SUPPORT
    evidence_source: MODEL_ORGANISM
    snippet: "In PS1 null blastocysts, neurons from mice hypomorphic for PS1 or conditionally depleted of PS1, substrate proteolysis and autophagosome clearance during macroautophagy are prevented as a result of a selective impairment of autolysosome acidification and cathepsin activation. These deficits are caused by failed PS1-dependent targeting of the v-ATPase V0a1 subunit to lysosomes."
    explanation: >-
      Demonstrates that PS1 is required for v-ATPase V0a1 targeting to
      lysosomes; PSEN1 dysfunction therefore directly impairs lysosomal
      acidification and autophagosome clearance, anchoring the early-onset
      Alzheimer disease mechanism for this hypothesis group.
  - reference: PMID:30742114
    reference_title: "Mitophagy inhibits amyloid-β and tau pathology and reverses cognitive deficits in models of Alzheimer's disease."
    supports: SUPPORT
    evidence_source: MODEL_ORGANISM
    snippet: "Mitophagy diminishes insoluble Aβ1-42 and Aβ1-40 and prevents cognitive impairment in an APP/PS1 mouse model through microglial phagocytosis of extracellular Aβ plaques and suppression of neuroinflammation."
    explanation: >-
      Supports mitophagy as a mitochondrial quality-control mechanism that can
      alter amyloid burden, neuroinflammation, and cognitive impairment in an
      Alzheimer disease mouse model.
  - reference: PMID:30124770
    reference_title: "Genetic variation within endolysosomal system is associated with late-onset Alzheimer's disease."
    supports: SUPPORT
    evidence_source: COMPUTATIONAL
    snippet: "We found that both gene sets were enriched for genetic Alzheimer's associations: an enrichment score 3.67 standard deviations from the null model (P = 0.00012) was detected for AphagEndoLyso"
    explanation: >-
      Population-level genetic support for the pathway rather than for a single gene:
      an 891-gene autophagy and endolysosomal set is enriched for Alzheimer risk
      variants across three genome-wide association studies with a combined n of
      62,415. Pathway-enrichment analysis over published summary statistics, hence
      COMPUTATIONAL.
  - reference: PMID:30124770
    reference_title: "Genetic variation within endolysosomal system is associated with late-onset Alzheimer's disease."
    supports: SUPPORT
    evidence_source: COMPUTATIONAL
    snippet: "The APOE locus explained most of the MenDisLyso signal (1.16 standard deviations after APOE removal, P = 0.12), but the AphagEndoLyso signal was less affected (3.35 standard deviations after APOE removal, P = 0.00040)."
    explanation: >-
      The contrast that makes the enrichment above evidence for this pathway rather
      than a restatement of APOE: removing the APOE locus collapses the
      Mendelian-lysosomal-disease subset's signal but leaves the broader autophagy
      and endolysosomal signal essentially intact.
  - reference: PMID:9236226
    reference_title: "Increased neuronal endocytosis and protease delivery to early endosomes in sporadic Alzheimer's disease: neuropathologic evidence for a mechanism of increased beta-amyloidogenesis."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "Endosomal alterations were evident in most pyramidal neurons in Alzheimer brain, detectable at early stages of the disease but absent in several other neurodegenerative disorders examined."
    explanation: >-
      Establishes that endosomal enlargement (up to 32-fold by individual endosome
      volume, 2.5-fold per neuron) is both early and disease-specific in human
      tissue. "Detectable at early stages" in cross-sectional post-mortem material
      means the change is present early, not that it precedes amyloid or tau.
  - reference: PMID:8004466
    reference_title: Lysosomal abnormalities in degenerating neurons link neuronal compromise to senile plaque development in Alzheimer disease.
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "alterations of the endosomal-lysosomal system in neurons of the Alzheimer disease brain, which preceded evident degenerative changes"
    explanation: >-
      Hydrolase-positive lysosomes accumulate in affected pyramidal neurons before
      visible degenerative change and far beyond normal aging. This is 1994
      morphology in end-stage tissue, so "before overt degeneration" means before
      degeneration was visible in that neuron, not before disease onset.
  - reference: PMID:28085665
    reference_title: "Impaired retrograde transport of axonal autophagosomes contributes to autophagic stress in Alzheimer's disease neurons."
    supports: SUPPORT
    evidence_source: MODEL_ORGANISM
    snippet: "This interaction impairs dynein recruitment to amphisomes through competitive interruption of dynein-Snapin motor-adaptor coupling, thus immobilizing them in distal axons."
    explanation: >-
      A feedback route by which the aggregate impairs its own clearance: amyloid-beta
      oligomers bind dynein motors and displace the Snapin adaptor, stranding
      amphisomes in distal axons away from somatic lysosomes. Snapin deletion
      reproduces the axonal autophagic stress and Snapin overexpression relieves it,
      supplying the perturbation in both directions.
  - reference: PMID:23103503
    reference_title: "Presenilin-null cells have altered two-pore calcium channel expression and lysosomal calcium: implications for lysosomal function."
    supports: SUPPORT
    evidence_source: IN_VITRO
    snippet: "We report that presenilin loss leads to a lower total lysosomal calcium store despite the buildup of lysosomes found in these cells."
    explanation: >-
      A second candidate mechanism for the presenilin lysosomal defect alongside the
      v-ATPase acidification route above, with altered two-pore calcium channel
      expression. Both come from presenilin-null or presenilin-depleted cells and
      neither has been weighed against the other in patient-derived cells carrying
      heterozygous PSEN1 mutations, so they are curated as complementary readings of
      one genetic lesion rather than as a settled dual mechanism.
  notes: >-
    Retained as EMERGING because PSEN1-linked lysosomal acidification evidence
    provides a strong mechanistic anchor and mitophagy perturbation links
    mitochondrial quality control to amyloid, tau, inflammatory, and cognitive
    readouts. The broader contribution of autophagy-lysosomal and mitochondrial
    clearance defects to late-onset Alzheimer disease remains incompletely
    resolved in human longitudinal data.
    The descriptive claim and the causal claim should be kept apart when reading
    this block. That autophagic and endolysosomal failure is an early,
    Alzheimer-specific feature is well supported by human post-mortem tissue and
    by pathway-level genetics; that it is a primary driver rather than an early
    consequence is not, because the human evidence is cross-sectional and the
    perturbations are rodent. The therapeutic claim is weaker still: no
    autophagy-enhancing therapy has completed phase III testing in Alzheimer
    disease, so that arm is entirely preclinical. The OpenScientist
    autophagy_lysosomal_clearance_model report asked for these three claims to
    carry separate statuses (ESTABLISHED, EMERGING, SPECULATIVE); the schema has
    one status per hypothesis and neither ESTABLISHED nor SPECULATIVE is a
    permitted value, so the distinction is recorded here instead.
- hypothesis_group_id: hsv1_reactivation_model
  hypothesis_label: HSV-1 Reactivation Model
  status: EMERGING
  description: >-
    Latent HSV-1 reactivation in selectively vulnerable RORB+ glutamatergic
    neurons is modeled as a possible upstream contributor to neuronal
    vulnerability and later synaptic-network failure.
  applies_to_subtypes:
  - Late-Onset Alzheimer's Disease
  evidence:
  - reference: PMID:42094473
    reference_title: "Resolving human neuronal herpesvirus reactivation via petabase-scale association studies."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "Integrative single-nucleus analyses resolve direct evidence of HSV-1 expression in RORB+ glutamatergic neurons, implicating viral reactivation in a neuronal population progressively lost during dementia."
    explanation: Supports HSV-1 reactivation in a selectively vulnerable Alzheimer-relevant neuronal population.
  - reference: PMID:42094473
    reference_title: "Resolving human neuronal herpesvirus reactivation via petabase-scale association studies."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "identifying recurrent herpes simplex virus 1 (HSV-1) reactivation in healthy but not pathological post-mortem human brain tissue"
    explanation: Supports the temporal interpretation that HSV-1 reactivation may precede loss of vulnerable neurons rather than simply appear in end-stage pathological tissue.
  notes: >-
    Retained as EMERGING. The evidence is human, cell-type-resolved, and
    hypothesis-generating, but causality between HSV-1 expression and Alzheimer
    disease progression remains unresolved.
- hypothesis_group_id: ev_mediated_tau_propagation_model
  hypothesis_label: Arc-Dependent EV-Mediated Tau Propagation Model
  status: EMERGING
  description: >-
    Cell-to-cell (prion-like) spread of tau pathology is modeled as being driven,
    in part, by packaging of pathological tau into neuronal extracellular vesicles
    (EVs). The activity-regulated, capsid-forming neuronal protein Arc binds tau
    directly and is required for efficient release of tau in EVs; Arc-tau
    co-packaging seeds tau aggregation in recipient neurons, propagating tangle
    pathology along connected circuits. In this model EV-tau release is partly
    protective for the donor neuron (eliminating toxic intracellular tau) but
    drives intercellular transmission of seed-competent tau.
  applies_to_subtypes:
  - Early-Onset Alzheimer's Disease
  - Late-Onset Alzheimer's Disease
  evidence:
  - reference: PMID:42372723
    reference_title: "Arc mediates intercellular tau transmission via extracellular vesicles."
    supports: SUPPORT
    evidence_source: MODEL_ORGANISM
    snippet: "Strikingly, intercellular tau transmission is almost absent in Arc KO mice."
    explanation: >-
      Loss of Arc nearly abolishes neuron-to-neuron tau transmission in mice,
      supporting Arc-dependent EV packaging as a mechanism of prion-like tau
      spread.
  - reference: PMID:42372723
    reference_title: "Arc mediates intercellular tau transmission via extracellular vesicles."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "Moreover, Arc levels in brain-derived EVs isolated from human Alzheimer's disease (AD) brains show a strong positive correlation with phosphorylated EV-tau levels."
    explanation: >-
      Human AD brain-derived EVs link Arc levels to phosphorylated EV-tau,
      extending the EV-mediated propagation model to human disease.
  - reference: PMID:29328916
    reference_title: "The Neuronal Gene Arc Encodes a Repurposed Retrotransposon Gag Protein that Mediates Intercellular RNA Transfer."
    supports: SUPPORT
    evidence_source: IN_VITRO
    snippet: "Endogenous Arc protein is released from neurons in extracellular vesicles that mediate the transfer of Arc mRNA into new target cells"
    explanation: >-
      Arc self-assembles into virus-like capsids and is released from neurons in
      EVs that transfer cargo to recipient cells, establishing the biological
      plausibility for Arc packaging tau into EVs.
  - reference: PMID:27030011
    reference_title: "Extracellular Vesicles Isolated from the Brains of rTg4510 Mice Seed Tau Protein Aggregation in a Threshold-dependent Manner."
    supports: SUPPORT
    evidence_source: MODEL_ORGANISM
    snippet: "extracellular vesicles derived from transgenic mice were capable of seeding tau aggregation in a threshold-dependent manner"
    explanation: >-
      Brain-derived EVs from tau-transgenic mice seed aggregation of monomeric
      tau in recipient cells, supporting EVs as vehicles for seed-competent tau.
  - reference: PMID:33417012
    reference_title: "Exosomes induce endolysosomal permeabilization as a gateway by which exosomal tau seeds escape into the cytosol."
    supports: SUPPORT
    evidence_source: IN_VITRO
    snippet: "tau aggregation was only induced in cells that exhibited permeabilization"
    explanation: >-
      Endolysosomal permeabilization is the gating step by which internalized
      exosomal tau seeds escape into the cytosol to template aggregation,
      supplying the downstream seeding mechanism the model requires.
  - reference: PMID:29495441
    reference_title: "Detection of Aggregation-Competent Tau in Neuron-Derived Extracellular Vesicles."
    supports: SUPPORT
    evidence_source: IN_VITRO
    snippet: "In each case, most tau was free-floating with a small component inside EVs."
    explanation: >-
      Across iPSC-neuron media, CSF, and plasma most extracellular tau is
      free-floating rather than EV-encapsulated, qualifying the quantitative in
      vivo contribution of the EV-tau route.
  - reference: PMID:19556514
    reference_title: "Independent effects of intra- and extracellular Abeta on learning-related gene expression."
    supports: SUPPORT
    evidence_source: MODEL_ORGANISM
    snippet: "reductions in both the number of Arc-activated neurons and the levels of Arc mRNA were seen in the neocortices of depositing mice from all transgenic lines (deficits ranging from 14 to 26%)"
    explanation: >-
      Amyloid-beta deposition suppresses activity-induced Arc expression, so the
      Arc-dependent EV-tau mechanism may be attenuated precisely in plaque-rich
      regions (the "Abeta-Arc paradox").
  notes: >-
    EMERGING. Demonstrated in primary neurons, rTg4510 tau-transgenic / Arc-KO
    mice, and human postmortem brain EVs (Tyagi et al., Cell 2026). EV-tau is one
    of several proposed routes of tau spread (free/naked tau uptake via LRP1,
    tunneling nanotubes, trans-synaptic transfer); the relative in vivo
    contribution of each, and how Arc levels modulate EV-tau release as disease
    progresses, remain to be resolved. By 8 months in the transgenic model, tau
    pathology was similar between Arc-KO and control, so Arc loss does not overtly
    accelerate late-stage pathology despite blocking transmission.
    OpenScientist deep research (July 2026;
    kb/hypotheses/Alzheimer_Disease/ev_mediated_tau_propagation_model) judged the
    model PARTIALLY SUPPORTED (EMERGING), with the following qualifications:
    (1) the Arc-tau binding and Arc-KO transmission phenotype rest on a single
    source (Tyagi et al., 2026) with no independent replication; (2) most
    extracellular tau is free-floating rather than EV-encapsulated (PMID:29495441),
    so the quantitative in vivo weight of the EV route versus LRP1/HSPG free-tau
    uptake and tunneling nanotubes is unresolved; (3) Abeta deposition suppresses
    Arc expression by 14-58% (PMID:19556514), potentially attenuating the mechanism
    in plaque-rich regions; and (4) the model was tested with P301L (4R) tau, not
    authentic AD-type 3R/4R paired helical filaments, and ARC carries no GWAS/GenCC
    association with AD. Supporting biology (Arc capsid EV transfer, PMID:29328916;
    EV-tau seeding, PMID:27030011; endolysosomal-escape gating, PMID:33417012) is
    independently established.
    Reassessment of that report adds three things. The single-source dependency
    carries a disclosed commercial interest: the senior author of PMID:42372723
    co-founded VNV, LLC and consults for Aera Therapeutics, which licenses patents
    covering Arc capsids, which is the context in which independent replication
    matters most. The Arc capsid work (PMID:29328916) demonstrates transfer of Arc's
    own mRNA, so it establishes plausibility of the vehicle and not of tau as cargo.
    And the 8-month convergence of tau pathology between Arc-knockout and control
    animals, which the report uses to argue for restricting the model to early
    disease, is not in the abstract this repository holds; no early-stage
    restriction should be curated from it without the full text. Status stays
    EMERGING. See the Arc replication-and-genetic-support knowledge gap.
- hypothesis_group_id: apoe_risk_modulation_model
  hypothesis_label: APOE Risk-Modulation Model
  status: ALTERNATIVE
  description: >-
    Apolipoprotein E (APOE) genotype is modeled as the strongest common-variant
    modifier of late-onset Alzheimer disease risk, acting upstream of and across
    multiple pathologies rather than through a single pathway. In proportion to
    APOE4 gene dose, APOE4 accelerates amyloid-beta seeding/aggregation and
    impairs its clearance, exacerbates tau-mediated neurodegeneration and
    neuroinflammation, promotes glial lipid/cholesterol accumulation, and impairs
    blood-brain-barrier integrity, whereas APOE2 and the rare protective variants
    APOE3 Christchurch (R136S) and APOE3 Jacksonville (V236E) blunt these effects.
    The model predicts that lowering or structurally modifying APOE — especially
    APOE4 — should ameliorate several AD pathologies simultaneously, motivating
    APOE-directed therapeutics (ASO knockdown, AAV-APOE2, anti-aggregation small
    molecules).
  applies_to_subtypes:
  - Late-Onset Alzheimer's Disease
  evidence:
  - reference: PMID:28959956
    reference_title: "ApoE4 markedly exacerbates tau-mediated neurodegeneration in a mouse model of tauopathy."
    supports: SUPPORT
    evidence_source: MODEL_ORGANISM
    snippet: "P301S/E4 mice develop markedly more brain atrophy and neuroinflammation than P301S/E2 and P301S/E3 mice, whereas P301S/EKO mice are largely protected from these changes."
    explanation: >-
      Isoform-dependent modulation of tau-driven neurodegeneration and
      neuroinflammation (E4 > E3/E2, and near-protection when APOE is absent)
      anchors APOE as a cross-pathology risk modifier acting beyond amyloid.
  - reference: PMID:31686034
    reference_title: "Resistance to autosomal dominant Alzheimer's disease in an APOE3 Christchurch homozygote: a case report."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "The individual had two copies of the APOE3 Christchurch (R136S) mutation, unusually high brain amyloid levels and limited tau and neurodegenerative measurements."
    explanation: >-
      A homozygous APOE3 Christchurch carrier resisted autosomal-dominant AD for
      roughly three decades despite heavy amyloid, dissociating amyloid burden
      from tau and neurodegeneration and implicating APOE receptor-binding avidity
      as a protective lever.
  notes: >-
    ALTERNATIVE. APOE modulation overlaps and interacts with the amyloid, tau,
    neuroimmune, vascular/BBB, and lipid models rather than competing with them;
    it is curated as a distinct hypothesis because the protective-variant
    (Christchurch/Jacksonville) and APOE-lowering evidence points to APOE itself
    as an actionable node upstream of multiple pathologies. Sex- and
    ancestry-dependent APOE risk and the precise molecular mechanism (aggregation,
    receptor binding, lipidation) remain incompletely resolved.
- hypothesis_group_id: adaptive_immune_tcell_model
  hypothesis_label: Adaptive Immune (T Cell) Neurodegeneration Model
  status: EMERGING
  description: >-
    Beyond innate microglial responses, a clonally expanded, predominantly
    cytotoxic (CD8+) T cell response is modeled as an active driver of
    tau-associated neurodegeneration. In tauopathy — but not pure amyloidosis —
    microglia are required to recruit T cells to sites of tau pathology;
    infiltrating T cells transition from activated to exhausted states with
    restricted TCR clonality, and IFN-gamma/PD-1 signaling amplifies neuronal
    loss. In the tau context the model predicts that depleting T cells or blocking
    IFN-gamma/PD-1 signaling should be neuroprotective. Critically, this
    therapeutic valence is tau-scoped, not universal: in a pure-amyloid context
    PD-1 checkpoint blockade instead drives an IFN-gamma-dependent myeloid
    response that clears Aβ and improves cognition, so "block IFN-gamma/PD-1" is
    not a general prescription. The model frames the cytotoxic-CD8 arm as a
    shared, targetable axis across tau-driven neurodegenerative disease.
  applies_to_subtypes:
  - Early-Onset Alzheimer's Disease
  - Late-Onset Alzheimer's Disease
  evidence:
  - reference: PMID:36890231
    reference_title: "Microglia-mediated T cell infiltration drives neurodegeneration in tauopathy."
    supports: SUPPORT
    evidence_source: MODEL_ORGANISM
    snippet: "We found that mice with tauopathy but not those with amyloid deposition developed a unique innate and adaptive immune response and that depletion of microglia or T cells blocked tau-mediated neurodegeneration."
    explanation: >-
      Depletion of either microglia or T cells blocks tau-mediated
      neurodegeneration in a tauopathy model, establishing the adaptive T cell
      arm as functionally required rather than a bystander.
  - reference: PMID:36890231
    reference_title: "Microglia-mediated T cell infiltration drives neurodegeneration in tauopathy."
    supports: SUPPORT
    evidence_source: MODEL_ORGANISM
    snippet: "Inhibition of interferon-γ and PDCD1 signalling both significantly ameliorated brain atrophy."
    explanation: >-
      IFN-gamma and PD-1 (PDCD1) blockade reduce brain atrophy, nominating
      specific, druggable nodes within the adaptive-immune neurodegeneration axis.
  - reference: PMID:41890046
    reference_title: "CD8+ T cells are primed by cDC1 and exacerbate tau-mediated neurodegeneration."
    supports: SUPPORT
    evidence_source: MODEL_ORGANISM
    snippet: "We demonstrate that tauopathy mice deficient in cDC1 are markedly protected against tau-mediated neurodegeneration and display a selective decrease in brain CD8+ T cell infiltration and glial reactivity."
    explanation: >-
      Independent Holtzman-group follow-up: ablating conventional type-1 dendritic
      cells (cDC1) — the cross-presenting antigen-presenting cell — protects
      tauopathy mice and reduces brain CD8+ T cell infiltration, resolving the
      antigen-presenting route that primes the pathogenic CD8 response.
  - reference: PMID:26779813
    reference_title: "PD-1 immune checkpoint blockade reduces pathology and improves memory in mouse models of Alzheimer's disease."
    supports: SUPPORT
    evidence_source: MODEL_ORGANISM
    snippet: "When induced in mice with established pathology, this immunological response leads to clearance of cerebral amyloid-β (Aβ) plaques and improved cognitive performance."
    explanation: >-
      Qualifies the model's therapeutic prediction: in an amyloid context PD-1
      blockade is beneficial (IFN-gamma-dependent Aβ clearance), the opposite of
      the tau-context prediction — so the IFN-gamma/PD-1 valence is tau-scoped,
      not universal.
  - reference: PMID:37703311
    reference_title: Glial reactivity and T cell infiltration in frontotemporal lobar degeneration with tau pathology.
    supports: SUPPORT
    directness: INDIRECT
    evidence_source: HUMAN_CLINICAL
    snippet: "Of note, the CD8+ T cell infiltration was associated with pTau epitopes and microglial and astrocytic markers."
    explanation: >-
      Human, amyloid-independent tissue evidence that CD8+ T cell infiltration tracks
      tau: a post-mortem series of 45 progressive supranuclear palsy, 33 Pick's
      disease and 12 FTLD-MAPT cases against 52 controls, quantified against seven
      phospho-tau epitopes. Extends the tau-scoping of this model beyond Alzheimer
      disease to primary tauopathy, where amyloid is not a confounder.
  - reference: PMID:37703311
    reference_title: Glial reactivity and T cell infiltration in frontotemporal lobar degeneration with tau pathology.
    supports: REFUTE
    directness: INDIRECT
    evidence_source: HUMAN_CLINICAL
    snippet: "Our results highlight that FTLD-tau is associated with astrocyte reactivity, remarkably little activation of microglia, but involvement of adaptive immunity in the form of chemokine-driven recruitment of T lymphocytes."
    explanation: >-
      Qualifies the microglia-as-required-recruiter step, which is strong in the
      mouse tauopathy perturbation but not corroborated here: in human FTLD-tau the
      T cell recruitment is chemokine-driven and microglial activation is minimal.
      Curated so the recruiter step reads as model-organism-strong and
      human-uncertain rather than established in both.
  - reference: PMID:41957675
    reference_title: Meningeal and infiltrating T-cells are associated with repetitive head trauma and tau-mediated neurodegeneration in chronic traumatic encephalopathy.
    supports: SUPPORT
    directness: INDIRECT
    evidence_source: HUMAN_CLINICAL
    snippet: "Meningeal and infiltrating T-cells were elevated in sulci with p-tau depositions and spatially related to MHC2 expressing cells."
    explanation: >-
      Second human tauopathy with the same spatial association, in 58 individuals
      exposed to repetitive head impacts against 18 unexposed controls. The MHC class
      II relationship is the antigen-presentation element of the model observed in
      human tissue.
  - reference: PMID:41957675
    reference_title: Meningeal and infiltrating T-cells are associated with repetitive head trauma and tau-mediated neurodegeneration in chronic traumatic encephalopathy.
    supports: SUPPORT
    directness: INDIRECT
    evidence_source: HUMAN_CLINICAL
    snippet: "In addition, T-cell infiltration correlated with the duration of RHI, as measured by years of sports play, and synaptic loss."
    explanation: >-
      Links T cell burden to exposure dose and to synaptic loss in human tissue,
      which is the outcome this model predicts. Observational and post-mortem, so it
      establishes association rather than the causal role the mouse depletion
      experiments carry.
  notes: >-
    EMERGING. The core functional evidence (T cell depletion, IFN-gamma/PD-1
    inhibition, cDC1 ablation) is from mouse tauopathy models, with correlative
    human support (increased cytotoxic T cells in AD brain/CSF tracking tau, not
    amyloid). The antigen-presenting route is now resolved to cDC1
    cross-presentation (PMID:41890046), but the recognized antigen(s) remain
    unidentified — and two findings temper the "antigen-driven clonal CD8" reading:
    aging expands antigen-independent virtual-memory CD8 T cells that are clonal
    and cytotoxic without priming (PMID:42432776), and the clonally expanded CD8
    TCRs recovered from human AD CSF were specific to Epstein-Barr virus antigens,
    not tau (PMID:31915375). IFN-gamma is also not CD8-exclusive (NK cells are a
    parallel source tracking cognitive decline, PMID:11268360), and whether B cells
    contribute remains open. See the antigen-identity and IFN-gamma-valence
    knowledge-gap discussions. Complements — does not replace — the innate
    neuroimmune_glial_amplification_model.
    Human cross-tauopathy support is now curated from FTLD-tau (PMID:37703311) and
    chronic traumatic encephalopathy (PMID:41957675), which also bound the
    microglia-as-recruiter step: that step is strong in the mouse tauopathy
    perturbation and unsupported in human FTLD-tau, where microglial activation is
    minimal and T cell recruitment is chemokine-driven. The OpenScientist
    adaptive_immune_tcell_model report describes the infiltrating cells as CD103+
    tissue-resident memory cells and proposes an ontology binding on that basis;
    the cited paper (PMID:40993111) reports the opposite marker polarity —
    CD103-negative — so neither the label nor the binding was curated. Its reading
    of PMID:15549735 as showing regulatory T cells to be neuroprotective is also
    inverted (eliminating them attenuated the lesion) and was not curated.
- hypothesis_group_id: anticholinergic_muscarinic_burden_model
  hypothesis_label: Anticholinergic Muscarinic Burden Model
  status: EMERGING
  description: >-
    Sustained pharmacologic blockade of central muscarinic acetylcholine
    receptors — the property shared by first-generation antihistamines
    (diphenhydramine), tricyclic antidepressants, bladder antimuscarinics, and
    antiparkinson anticholinergics — is modeled as removing M1-dependent tonic
    restraint on amyloidogenic APP processing and on GSK3-beta-mediated tau
    phosphorylation. Under this model, cumulative anticholinergic exposure is a
    modifiable disease-risk factor rather than only a cause of the reversible
    cognitive impairment classically attributed to these drugs.
  evidence:
  - reference: PMID:16504943
    reference_title: M1 receptors play a central role in modulating AD-like pathology in transgenic mice.
    supports: SUPPORT
    evidence_source: MODEL_ORGANISM
    snippet: "We further demonstrate that administration of dicyclomine, an M1 antagonist, exacerbates the Abeta and tau pathologies."
    explanation: >-
      Pharmacologic M1 antagonism — not merely genetic loss — worsens both
      amyloid and tau pathology, which is the mechanistic premise the model
      extrapolates from to chronic human anticholinergic exposure.
  - reference: PMID:25621434
    reference_title: "Cumulative use of strong anticholinergics and incident dementia: a prospective cohort study."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "A 10-year cumulative dose-response relationship was observed for dementia and Alzheimer disease"
    explanation: >-
      The cumulative dose-response in a prospective cohort is the human
      observation the model is built to explain.
  - reference: PMID:38063254
    reference_title: Anticholinergic deprescribing interventions for reducing risk of cognitive decline or dementia in older adults with and without prior cognitive impairment.
    supports: NO_EVIDENCE
    evidence_source: HUMAN_CLINICAL
    snippet: "The evidence from RCTs was of very low certainty so cannot support or refute the hypothesis that actively reducing or stopping prescription of medications with anticholinergic properties can improve cognitive outcomes in older people."
    explanation: >-
      The interventional test of the model is uninformative in both directions,
      which is why this hypothesis group is curated EMERGING rather than
      ALTERNATIVE or CANONICAL.
  notes: >-
    EMERGING, and deliberately not promoted further. The mechanistic arm is
    MODEL_ORGANISM only (M1 knockout and dicyclomine in 3xTg-AD/Tg-SwDI mice);
    the human arm is entirely observational and has two unresolved threats.
    (1) Reverse causation: prodromal dementia produces insomnia, depression,
    and urinary urgency, which are the indications for these drugs. (2) Class
    specificity: the largest UK studies concentrate the signal in
    antidepressant, urological, and antiparkinson anticholinergics, with
    antipsychotics and antiepileptics significant in QResearch
    (PMID:31233095) but antipsychotics null in CPRD (PMID:29695481), and the
    largest antihistamine-specific cohort finds a dose-response for
    second-generation agents that barely enter the CNS as well as for
    first-generation ones (PMID:38935035). Read that last result carefully in
    both directions: the gradient is steeper for the CNS-penetrant class
    (first-generation 1.13/1.29/1.51 across rising cumulative dose versus
    second-generation 1.11/1.19/1.26, the authors concluding that
    first-generation agents carry the higher risk), which is what a central
    mechanism predicts, while the residual non-null second-generation signal
    is left unexplained by that mechanism and is the open puzzle. Do NOT read the
    antihistamine nulls as a clean negative: the CPRD analysis records only
    prescriptions, and its authors note that over-the-counter supply is
    invisible in their data and that antihistamine use is therefore
    underestimated, which matters most for exactly the drug that prompted this
    curation. See the two attached discussions before strengthening any claim
    here.
- hypothesis_group_id: complement_synaptic_pruning_model
  hypothesis_label: Complement-Dependent Microglial Synapse Elimination Model
  status: ALTERNATIVE
  description: >-
    Synapse loss — the pathological change that correlates best with cognitive
    decline — is modeled as an active, receptor-mediated phagocytic process
    rather than as passive degeneration secondary to amyloid or tau toxicity.
    The classical complement cascade that prunes surplus synapses during
    development (C1q tagging, C3 opsonization, microglial CR3 engagement) is
    modeled as being reactivated early in disease, so that microglia engulf
    structurally intact synapses. Soluble amyloid-beta oligomers and
    phosphorylated tau both act as upstream triggers, which makes this the
    mechanism by which two otherwise separate proteinopathies converge on the
    same cellular endpoint.
  applies_to_subtypes:
  - Early-Onset Alzheimer's Disease
  - Late-Onset Alzheimer's Disease
  evidence:
  - reference: PMID:27033548
    reference_title: "Complement and microglia mediate early synapse loss in Alzheimer mouse models."
    supports: SUPPORT
    evidence_source: MODEL_ORGANISM
    snippet: "C1q, the initiating protein of the classical complement cascade, is increased and associated with synapses before overt plaque deposition."
    explanation: >-
      Places complement tagging of synapses upstream of plaque deposition, which
      is what distinguishes this model from synapse loss as a late consequence of
      established amyloid pathology.
  - reference: PMID:27033548
    reference_title: "Complement and microglia mediate early synapse loss in Alzheimer mouse models."
    supports: SUPPORT
    evidence_source: MODEL_ORGANISM
    snippet: "Inhibition of C1q, C3, or the microglial complement receptor CR3 reduces the number of phagocytic microglia, as well as the extent of early synapse loss."
    explanation: >-
      Blocking three separate steps of the same cascade each reduces synapse
      loss, establishing the pathway as required rather than merely present.
  - reference: PMID:30392797
    reference_title: "Changes in the Synaptic Proteome in Tauopathy and Rescue of Tau-Induced Synapse Loss by C1q Antibodies."
    supports: SUPPORT
    evidence_source: MODEL_ORGANISM
    snippet: "At synapses, C1q decorated perisynaptic membranes, accumulated in correlation with phospho-Tau, and was associated with augmented microglial engulfment of synapses and decline of synapse density."
    explanation: >-
      Independent laboratory, and a tau rather than amyloid driver, reaching the
      same C1q-microglia-synapse axis — the strongest evidence that the model is
      not specific to amyloid models.
  - reference: PMID:28566429
    reference_title: "Complement C3 deficiency protects against neurodegeneration in aged plaque-rich APP/PS1 mice."
    supports: SUPPORT
    evidence_source: MODEL_ORGANISM
    snippet: "We found that 16-month-old APP/PS1;C3 KO mice performed better on a learning and memory task than did APP/PS1 mice, despite having more cerebral Aβ plaques."
    explanation: >-
      Qualifies the model in an important direction: removing complement protects
      synapses and cognition while *increasing* plaque burden, so complement is
      simultaneously protective for amyloid clearance and harmful for synapses.
      The model must not be read as "complement is uniformly pathogenic".
  - reference: PMID:37652017
    reference_title: "Human astrocytes and microglia show augmented ingestion of synapses in Alzheimer's disease via MFG-E8."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "Here we observe astrocytes and microglia from human brains contain greater amounts of synaptic protein in AD compared with non-disease controls, and that proximity to amyloid-β plaques and the APOE4 risk gene exacerbate this effect."
    explanation: >-
      Confirms in human tissue that glia ingest synapses in Alzheimer disease, but
      nominates MFG-E8 rather than complement as the opsonin and puts astrocytes
      alongside microglia — so it corroborates the phenomenon while leaving the
      molecular pathway open in humans.
  - reference: PMID:31433986
    reference_title: "Complement C3 Is Activated in Human AD Brain and Is Required for Neurodegeneration in Mouse Models of Amyloidosis and Tauopathy."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "In addition, C3 protein is elevated in AD patient brains, including at synapses, and levels and processing of C3 are increased in AD patient CSF and correlate with tau."
    explanation: >-
      Human evidence that the opsonizing component of the cascade is elevated at
      synapses in Alzheimer brain, with a cerebrospinal fluid correlate tracking
      tau. This is stronger human support for complement specifically than the
      postsynaptic-density immunohistochemistry previously curated here.
  - reference: PMID:31433986
    reference_title: "Complement C3 Is Activated in Human AD Brain and Is Required for Neurodegeneration in Mouse Models of Amyloidosis and Tauopathy."
    supports: SUPPORT
    evidence_source: MODEL_ORGANISM
    snippet: "Blocking complement function by deleting C3 rescues plaque-associated synapse loss in PS2APP mice and ameliorates neuron loss and brain atrophy in TauP301S mice, improving neurophysiological and behavioral measurements."
    explanation: >-
      An independent laboratory reproducing the necessity result in both an
      amyloid and a tauopathy model, which is the convergence claim this
      hypothesis group is built on.
  - reference: PMID:42271460
    reference_title: "Dysregulation of complement at the synapse in P301S mice and human tauopathies."
    supports: SUPPORT
    evidence_source: MODEL_ORGANISM
    snippet: "C1q protein levels were markedly increased in brain homogenates from P301S mice compared to WT, accompanied by C1q deposition on tau aggregates."
    explanation: >-
      Extends complement dysregulation to pure tauopathy, where amyloid is absent,
      and localizes C1q to tau aggregates themselves.
  - reference: PMID:29504051
    reference_title: "Updated Meta-Analysis of BIN1, CR1, MS4A6A, CLU, and ABCA7 Variants in Alzheimer's Disease."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "We validated the risk for LOAD with BIN1 (rs744373), CR1 (rs6656401), and ABCA7 (rs376465), as well as the protective association for MS4A6A (rs610932) and CLU (rs11136000) variants."
    explanation: >-
      Inherited common-variant association at two complement loci — CR1, the
      complement receptor 1 gene, and CLU, encoding the complement inhibitor
      clusterin. Genetic association is the one class of human evidence here that
      cannot be a downstream consequence of established pathology, which is why it
      carries weight the postmortem correlations do not.
  notes: >-
    ALTERNATIVE rather than CANONICAL. The necessity evidence (C1q, C3 and CR3
    blockade) is entirely mouse; the same paper that supplies the best human
    evidence for glial synapse ingestion states that direct human evidence for
    glial involvement in synapse removal remained to be established, and
    implicates MFG-E8 rather than complement. Curated as a distinct hypothesis
    group rather than folded into neuroimmune_glial_amplification_model because
    it makes a specific, falsifiable claim about a named cascade acting on a
    named substrate (the synapse), and because its therapeutic prediction —
    anti-C1q antibody — is being tested clinically. See the attached
    HUMAN_MODEL_MISMATCH discussion before strengthening the human claim.
- hypothesis_group_id: endolysosomal_origin_model
  hypothesis_label: Endolysosomal Origin ("Inside-Out" Amyloid) Model
  status: EMERGING
  description: >-
    Amyloid pathology is modeled as beginning inside the neuron rather than in
    the extracellular space. Endocytic-pathway activation with enlargement of
    Rab5-positive early endosomes is the earliest recognized neuronal change in
    sporadic Alzheimer disease, preceding plaque deposition; APP-beta-C-terminal
    fragment accumulation then inhibits the lysosomal v-ATPase, autolysosomes
    fail to acidify, and amyloid-beta builds up within de-acidified autolysosomes
    until the neuron ruptures and its contents become the plaque core (the
    PANTHOS pattern). Under this model the senile plaque is the tombstone of a
    dead neuron, not a deposit of secreted peptide, which inverts the direction
    of the classical extracellular cascade.
  applies_to_subtypes:
  - Early-Onset Alzheimer's Disease
  - Late-Onset Alzheimer's Disease
  evidence:
  - reference: PMID:10880397
    reference_title: "Endocytic pathway abnormalities precede amyloid beta deposition in sporadic Alzheimer's disease and Down syndrome: differential effects of APOE genotype and presenilin mutations."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "These abnormalities were evident in pyramidal neurons of the neocortex at preclinical stages of disease when Alzheimer-like neuropathology, such as Abeta deposition, was restricted to the entorhinal region."
    explanation: >-
      Human postmortem evidence that neuronal endosomal abnormality is present at
      preclinical stages, before neocortical amyloid deposition — the temporal
      ordering the model requires.
  - reference: PMID:10880397
    reference_title: "Endocytic pathway abnormalities precede amyloid beta deposition in sporadic Alzheimer's disease and Down syndrome: differential effects of APOE genotype and presenilin mutations."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "By contrast, endosomes were normal in size at advanced stages of familial AD caused by mutations of presenilin 1 or 2, indicating that altered endocytosis is not a consequence of Abeta deposition."
    explanation: >-
      The paper's own internal control cuts both ways: it shows the endosomal
      phenotype is not merely downstream of amyloid, but also that it is absent in
      presenilin-mutation familial disease, so the mechanism is not universal
      across Alzheimer genetic subtypes.
  - reference: PMID:15465622
    reference_title: "Abeta localization in abnormal endosomes: association with earliest Abeta elevations in AD and Down syndrome."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "Double-immunofluorescence using numerous Abeta antibodies showed that intracellular Abeta localized principally to rab5-positive endosomes in neurons from AD brains and was prominent in enlarged endosomes."
    explanation: >-
      Locates amyloid-beta inside the enlarged endosomal compartment in human
      Alzheimer brain, connecting the endosomal lesion to the peptide itself.
  - reference: PMID:35654956
    reference_title: "Faulty autolysosome acidification in Alzheimer's disease mouse models induces autophagic build-up of Aβ in neurons, yielding senile plaques."
    supports: SUPPORT
    evidence_source: MODEL_ORGANISM
    snippet: "Quantitative analyses confirm that individual neurons exhibiting PANTHOS are the principal source of senile plaques in amyloid precursor protein AD models."
    explanation: >-
      The load-bearing claim of the model — that the plaque originates from a
      single dying neuron — stated for amyloid precursor protein transgenic mouse
      models. The corresponding human claim in the same work is presence of the
      pattern, not that it is the source of human plaques.
  - reference: PMID:37494443
    reference_title: "Lysosomal dysfunction in Down syndrome and Alzheimer mouse models is caused by v-ATPase inhibition by Tyr(682)-phosphorylated APP βCTF."
    supports: SUPPORT
    evidence_source: MODEL_ORGANISM
    snippet: "Lowering APP-βCTF Tyr682 phosphorylation restores v-ATPase and lysosome function in DS fibroblasts and in vivo in brains of DS model mice."
    explanation: >-
      Supplies the molecular mechanism and its genetic rescue: APP-βCTF
      competitively inhibits the v-ATPase V0a1 subunit, so the acidification
      failure is caused by an APP fragment upstream of amyloid-beta itself.
  - reference: PMID:31416668
    reference_title: "A Large Panel of Isogenic APP and PSEN1 Mutant Human iPSC Neurons Reveals Shared Endosomal Abnormalities Mediated by APP β-CTFs, Not Aβ."
    supports: SUPPORT
    evidence_source: IN_VITRO
    snippet: "Importantly, endosomal dysfunction correlated with accumulation of β-CTFs, not Aβ, and could be rescued by pharmacological modulation of β-secretase (BACE)."
    explanation: >-
      Independent laboratory, human isogenic iPSC neurons rather than
      overexpressing mice, dissociating the endosomal lesion from amyloid-beta and
      attributing it to the APP fragment — with a pharmacological rescue. This is
      the strongest non-Nixon-laboratory support for the upstream half of the model.
  - reference: PMID:31416668
    reference_title: "A Large Panel of Isogenic APP and PSEN1 Mutant Human iPSC Neurons Reveals Shared Endosomal Abnormalities Mediated by APP β-CTFs, Not Aβ."
    supports: SUPPORT
    evidence_source: IN_VITRO
    snippet: "APP and PSEN1 mutations had discordant effects on Aβ production but similar effects on APP β C-terminal fragments (β-CTFs), which accumulate in all mutant neurons."
    explanation: >-
      Partly reconciles the scope limitation this hypothesis group carries. Human
      postmortem work found endosomes of normal size in advanced
      presenilin-mutation familial disease, but in isogenic human neurons PSEN1
      mutants do accumulate β-CTFs and do enlarge endosomes — so the discrepancy
      may reflect disease stage or an endosome-size readout rather than the
      mechanism being absent in that subtype.
  - reference: PMID:35418158
    reference_title: "The amyloid plaque proteome in early onset Alzheimer's disease and Down syndrome."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "Endosomal/lysosomal proteins were particularly highly enriched in amyloid plaques."
    explanation: >-
      Laser-capture proteomics of human amyloid plaques against neighbouring
      non-plaque tissue. Endolysosomal enrichment inside the plaque is
      qualitatively what an intraneuronal origin predicts — the plaque should
      contain the remnants of the compartment it came from. It does not exclude
      contributions from glia or dystrophic neurites, so it is consistent with the
      model rather than decisive for it.
  notes: >-
    EMERGING, and deliberately curated as a competitor to — not a restatement of
    — autophagy_lysosomal_clearance_model. That model says defective clearance
    lets amyloid accumulate; this one says the plaque is *generated* inside the
    autolysosomal compartment of a neuron that then dies. The distinction is
    testable and matters therapeutically: it predicts that lowering APP-βCTF or
    restoring v-ATPase activity acts upstream of anything an anti-amyloid antibody
    can reach. Held EMERGING because the human evidence is strong for the
    endosomal lesion (Cataldo, human postmortem) but thin for PANTHOS itself,
    where the only verified human statement is that the pattern is present in
    Alzheimer brains. The v-ATPase mechanism and the PANTHOS phenotype come from
    the same laboratory, so they are not independent replications of each other.
- hypothesis_group_id: selective_neuronal_vulnerability_model
  hypothesis_label: Selective Neuronal Vulnerability Model
  status: EMERGING
  description: >-
    Alzheimer disease is modeled as a disease of specific, molecularly definable
    neuronal populations rather than of the cortex as a whole: locus coeruleus
    noradrenergic neurons and entorhinal cortex layer II excitatory neurons
    degenerate first and disproportionately, and within the entorhinal cortex the
    vulnerable excitatory population is marked by RORB. Under this model the
    question "why these cells?" is a mechanistic question in its own right —
    cell-intrinsic properties determine where tau pathology and neuronal loss
    begin, so aggregate burden alone cannot explain the anatomy of the disease.
  applies_to_subtypes:
  - Late-Onset Alzheimer's Disease
  evidence:
  - reference: PMID:33432193
    reference_title: "Molecular characterization of selectively vulnerable neurons in Alzheimer's disease."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "We identified RORB as a marker of selectively vulnerable excitatory neurons in the entorhinal cortex and subsequently validated their depletion and selective susceptibility to neurofibrillary inclusions during disease progression using quantitative neuropathological methods."
    explanation: >-
      Single-nucleus transcriptomics of human postmortem brain identifies the
      vulnerable population molecularly and confirms its depletion by independent
      quantitative neuropathology.
  - reference: PMID:8699259
    reference_title: "Profound loss of layer II entorhinal cortex neurons occurs in very mild Alzheimer's disease."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "Decreases in individual lamina were even more dramatic, with the number of neurons in layer II decreasing by 60% and in layer IV by 40% compared with controls."
    explanation: >-
      Stereological neuron counts showing that laminar loss is already severe at
      the mildest clinically detectable stage, establishing selectivity as an
      early rather than end-stage feature.
  - reference: PMID:27513978
    reference_title: "Locus coeruleus volume and cell population changes during Alzheimer's disease progression: A stereological study in human postmortem brains with potential implication for early-stage biomarker discovery."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "The long gap between NFT accumulation and neuronal loss suggests that a second trigger may be necessary to induce neuronal death in AD."
    explanation: >-
      Qualifies any simple "tangles kill the vulnerable cell" reading: in the
      locus coeruleus, tangle accumulation and neuronal loss are separated by a
      long interval, so vulnerability to tau inclusion and vulnerability to death
      are not the same property. Stated by the authors as an interpretation, not a
      measurement.
  notes: >-
    EMERGING as a mechanistic model, though the underlying observations are
    long-established and essentially uncontested as *descriptions*. What is
    unresolved is the mechanism: RORB marks a population, it is not known to be
    the cause of that population's vulnerability, and no experiment has yet shown
    that manipulating it changes susceptibility. This group is curated separately
    from the amyloid and tau models because it addresses a question neither
    answers — the anatomical selectivity of the disease — and because it supplies
    the missing link to the entry's HSV-1 node, which is built on the same RORB+
    population but cited only through the viral-reactivation work.
- hypothesis_group_id: necroptosis_model
  hypothesis_label: Necroptosis Model of Neuronal Death
  status: EMERGING
  description: >-
    The mode of neuronal death in Alzheimer disease is modeled as necroptosis —
    RIPK1/RIPK3-triggered, MLKL-executed programmed necrosis — rather than
    apoptosis. Activated necrosome components are found in granulovacuolar
    degeneration bodies, a long-recognized but mechanistically unexplained
    Alzheimer lesion, and their regional burden tracks neuronal loss. This model
    is significant because it names an executioner: it makes neuronal death a
    druggable step rather than the passive endpoint of upstream pathology.
  applies_to_subtypes:
  - Early-Onset Alzheimer's Disease
  - Late-Onset Alzheimer's Disease
  evidence:
  - reference: PMID:28758999
    reference_title: "Necroptosis activation in Alzheimer's disease."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "We found that necroptosis was activated in postmortem human AD brains, positively correlated with Braak stage, and inversely correlated with brain weight and cognitive scores."
    explanation: >-
      Human postmortem dose-response between necroptosis activation and both
      pathological stage and cognitive outcome.
  - reference: PMID:31802237
    reference_title: "Necrosome complex detected in granulovacuolar degeneration is associated with neuronal loss in Alzheimer's disease."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "GVDn + neurons inversely correlated with neuronal density in the early affected CA1 region of the hippocampus and in the late affected frontal cortex layer III."
    explanation: >-
      Anchors the mechanism to a specific, classically recognized neuropathological
      lesion and shows its burden tracks neuronal density in both an early- and a
      late-affected region.
  - reference: PMID:37708272
    reference_title: "MEG3 activates necroptosis in human neuron xenografts modeling Alzheimer's disease."
    supports: SUPPORT
    evidence_source: MODEL_ORGANISM
    snippet: "Down-regulation of MEG3 and inhibition of necroptosis using pharmacological or genetic manipulation of receptor-interacting protein kinase 1 (RIPK1), RIPK3, or mixed lineage kinase domain-like protein (MLKL) rescued neuronal cell loss in xenografted human neurons."
    explanation: >-
      Converts the human correlation into a causal claim: blocking three separate
      necroptosis effectors each rescues loss of human neurons in vivo.
  - reference: PMID:32949047
    reference_title: "Necrosome-positive granulovacuolar degeneration is associated with TDP-43 pathological lesions in the hippocampus of ALS/FTLD cases."
    supports: SUPPORT
    directness: INDIRECT
    evidence_source: HUMAN_CLINICAL
    snippet: "Necrosome-positive GVD was primarily observed in hippocampal regions of ALS/FTLD cases and was associated with hippocampal TDP-43 inclusions as the main predictor of the pMLKL-GVD stage, as well as with the Braak stage of neurofibrillary tangle pathology."
    explanation: >-
      Boundary condition: necrosome-positive granulovacuolar degeneration is not
      specific to Alzheimer disease but tracks proteinopathy more generally, so
      necroptosis is better modeled as a shared execution mechanism than as an
      Alzheimer-defining one.
  notes: >-
    EMERGING. The human evidence is correlative and rests substantially on
    phospho-MLKL and phospho-RIPK immunostaining of postmortem tissue, where
    antibody specificity and postmortem interval are known problems. The causal
    arm comes from human neurons xenografted into a mouse amyloid brain — a system
    whose own headline finding is that mouse neurons in the same brain do not show
    the phenotype, which is an argument for the model's human relevance and a
    reminder that rodent neurodegeneration models may miss the death mechanism
    entirely. Distinct from, and not yet reconciled with, the PARP1 parthanatos
    route curated in this entry; both are caspase-independent regulated necrosis
    and no work has established which dominates, or whether they act in different
    cells or stages.
- hypothesis_group_id: cellular_senescence_model
  hypothesis_label: Cellular Senescence Model
  status: EMERGING
  description: >-
    Accumulation of senescent cells — permanently cell-cycle-arrested but
    metabolically active, and secreting a proinflammatory senescence-associated
    secretory phenotype — is modeled as an active driver of tau pathology,
    neuroinflammation, and neuronal loss rather than a passive marker of brain
    aging. The model's distinguishing prediction is therapeutic and unusually
    direct: removing senescent cells, genetically or with senolytic drugs, should
    reduce pathology and preserve cognition even when the senescent cells are a
    small fraction of the tissue.
  applies_to_subtypes:
  - Late-Onset Alzheimer's Disease
  evidence:
  - reference: PMID:30232451
    reference_title: "Clearance of senescent glial cells prevents tau-dependent pathology and cognitive decline."
    supports: SUPPORT
    evidence_source: MODEL_ORGANISM
    snippet: "Here we show a causal link between the accumulation of senescent cells and cognition-associated neuronal loss."
    explanation: >-
      Genetic clearance of p16-positive cells as they arise prevents tau
      pathology and neuronal loss, which is the causal claim the model rests on.
  - reference: PMID:30936558
    reference_title: "Senolytic therapy alleviates Aβ-associated oligodendrocyte progenitor cell senescence and cognitive deficits in an Alzheimer's disease model."
    supports: SUPPORT
    evidence_source: MODEL_ORGANISM
    snippet: "Senolytic treatment of AD mice selectively removed senescent cells from the plaque environment, reduced neuroinflammation, lessened Aβ load, and ameliorated cognitive deficits."
    explanation: >-
      Pharmacological rather than genetic clearance, in an amyloid rather than tau
      model, reaching the same endpoint — the independent replication the
      therapeutic prediction needs.
  - reference: PMID:35531351
    reference_title: "Profiling senescent cells in human brains reveals neurons with CDKN2D/p19 and tau neuropathology."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "More than 97% of the senescent cells were excitatory neurons and overlapped with tau-containing neurofibrillary tangles (NFTs)."
    explanation: >-
      The largest human survey supports senescence being present and
      tangle-associated, but assigns it overwhelmingly to excitatory neurons —
      contradicting the glial cell-type assignment on which both mouse clearance
      experiments are built.
  - reference: PMID:30126037
    reference_title: "Tau protein aggregation is associated with cellular senescence in the brain."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "Transcriptomic analyses of NFT-containing neurons microdissected from postmortem AD brain revealed an expression profile consistent with cellular senescence."
    explanation: >-
      An independent human postmortem observation, predating the eigengene survey
      and using microdissected tangle-bearing neurons rather than a derived
      signature, that also places the senescent state in neurons.
  - reference: PMID:30126037
    reference_title: "Tau protein aggregation is associated with cellular senescence in the brain."
    supports: SUPPORT
    evidence_source: MODEL_ORGANISM
    snippet: "Using four AD transgenic mouse models, we found that NFTs, but not Aβ plaques, display a senescence-like phenotype."
    explanation: >-
      A third mouse clearance study, and one whose cell-type answer is neuronal
      and tau-specific rather than glial, so the dispute is not simply mouse glia
      versus human neurons.
  - reference: PMID:37679434
    reference_title: "Senolytic therapy in mild Alzheimer's disease: a phase 1 feasibility trial."
    supports: NO_EVIDENCE
    evidence_source: HUMAN_CLINICAL
    snippet: "Secondary cognitive and neuroimaging endpoints did not significantly differ from baseline to post-treatment further supporting a favorable safety profile."
    explanation: >-
      The first human test of the model's distinguishing therapeutic prediction.
      It is an open-label, uncontrolled feasibility study in five participants and
      was not powered for efficacy, so the absent cognitive and imaging signal
      neither confirms nor refutes the model — but the model can no longer be
      described as untested in people.
  - reference: PMID:40274471
    reference_title: "Evaluation of exploratory fluid biomarkers from a phase 1 senolytic trial in mild Alzheimer's disease."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "Mass spectrometry indicated stable levels of amyloid β and tau proteins in CSF, unchanged urinary metabolites, and modest treatment-associated lipid profile changes."
    explanation: >-
      Exploratory biomarkers from the same trial: senolytic treatment left CSF
      amyloid-beta and tau unchanged, which sits awkwardly with a simple "clear
      senescent cells, lower the SASP, reduce pathology" reading of the model.
  notes: >-
    EMERGING. Causal evidence remains murine, but it is three independent
    clearance experiments rather than two: genetic p16 ablation in tauopathy
    (PMID:30232451), senolytic clearance of plaque-associated oligodendrocyte
    progenitor cells (PMID:30936558), and senolytic treatment of late-stage tau
    mice reducing tangle density and neuron loss (PMID:30126037).
    The cell-of-origin question is the real dispute, and framing it as "mouse says
    glia, human says neurons" is too clean: PMID:30126037 is a mouse and human
    study placing the senescent state in tangle-bearing neurons, agreeing with the
    human eigengene survey (PMID:35531351) and disagreeing with both other mouse
    studies. What conflicts is the cell-type assignment across studies, not the
    species. It matters because senolytics kill the cells they target: a therapy
    validated against senescent glia would, on the neuronal reading, be aimed at
    postmitotic neurons.
    Human interventional data now exist and are uninformative in both directions —
    the SToMP-AD phase 1 open-label trial (NCT04063124, n=5) established central
    nervous system penetrance of dasatinib and acceptable safety, with no change in
    cognition or imaging and stable CSF amyloid-beta and tau. See the attached
    CONTROVERSY discussion. The standing methodological objection to the human
    survey is that it identifies senescence with a derived eigengene rather than a
    gold-standard marker; note separately that canonical CDKN2A/p16 is badly
    under-detected in single-nucleus data, which cuts against any negative call in
    the rarer cell types.
- hypothesis_group_id: myelin_oligodendrocyte_model
  hypothesis_label: Myelin and Oligodendrocyte Dysfunction Model
  status: EMERGING
  description: >-
    Age-related breakdown of myelin and of oligodendrocyte support for the axon
    is modeled as an upstream risk factor for amyloid deposition rather than a
    downstream consequence of it. Myelin damage concentrates the amyloidogenic
    processing machinery in axonal swellings and increases cleavage of amyloid
    precursor protein; separately, it diverts disease-associated microglia toward
    myelin debris and away from plaques, so the same lesion both raises amyloid
    production and lowers its clearance. APOE4 is modeled as acting partly through
    this route, via aberrant cholesterol deposition in oligodendrocytes and
    reduced myelination.
  applies_to_subtypes:
  - Late-Onset Alzheimer's Disease
  evidence:
  - reference: PMID:37258678
    reference_title: "Myelin dysfunction drives amyloid-β deposition in models of Alzheimer's disease."
    supports: SUPPORT
    evidence_source: MODEL_ORGANISM
    snippet: "Here we identify genetic pathways of myelin dysfunction and demyelinating injuries as potent drivers of amyloid deposition in mouse models of AD."
    explanation: >-
      Multiple independent myelin-mutant crosses each increase amyloid deposition,
      establishing the direction of causation in the mouse.
  - reference: PMID:37258678
    reference_title: "Myelin dysfunction drives amyloid-β deposition in models of Alzheimer's disease."
    supports: SUPPORT
    evidence_source: MODEL_ORGANISM
    snippet: "Mechanistically, myelin dysfunction causes the accumulation of the Aβ-producing machinery within axonal swellings and increases the cleavage of cortical amyloid precursor protein."
    explanation: >-
      Supplies the subcellular mechanism linking the myelin lesion to increased
      amyloid production.
  - reference: PMID:36385529
    reference_title: "APOE4 impairs myelination via cholesterol dysregulation in oligodendrocytes."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "We show that altered cholesterol localization in the APOE4 brain coincides with reduced myelination."
    explanation: >-
      Human postmortem evidence that oligodendrocyte and myelin pathology is real
      in the APOE4 brain, supplying the human leg the mouse causal work lacks.
  notes: >-
    EMERGING. The causal claim — myelin dysfunction drives amyloid deposition — is
    mouse-only; the human work establishes that oligodendrocyte cholesterol
    dysregulation and reduced myelination occur in APOE4 carriers but not that
    they cause amyloid deposition in people. The "microglial distraction" half of
    the mechanism should be read as SPECULATIVE rather than merely less
    independently replicated: a targeted search for independent primary support
    of that clearance edge found none, which is a checked absence rather than an
    unexamined gap. This group was curated in part to give the entry's
    oligodendrocyte precursor cell plasma proteomic age gap biomarker a
    lineage mechanism to attach to, but that justification does not survive
    inspection. The biomarker comes from a plasma clock spanning more than forty
    cell types (PMID:42297981) in which the strongest APOE4-linked Alzheimer
    signal is astrocyte aging, so it is consistent with this model without
    privileging it. Both corrections are curated in full on the Oligodendrocyte
    and Myelin Dysfunction node.
- hypothesis_group_id: network_hyperexcitability_model
  hypothesis_label: Network Hyperexcitability and Interneuron Dysfunction Model
  status: EMERGING
  description: >-
    Cognitive decline is modeled as arising in part from a failure of inhibition
    rather than only from loss of excitatory synapses: amyloid-beta impairs
    parvalbumin-expressing inhibitory interneurons through reduced levels of the
    interneuron-predominant voltage-gated sodium channel subunit Nav1.1,
    degrading gamma oscillations and permitting network hypersynchrony and
    epileptiform activity. The model predicts that subclinical epileptiform
    activity should be common in Alzheimer disease, should track faster decline,
    and should be a treatable contributor to symptoms rather than an incidental
    finding.
  applies_to_subtypes:
  - Early-Onset Alzheimer's Disease
  - Late-Onset Alzheimer's Disease
  evidence:
  - reference: PMID:22541439
    reference_title: "Inhibitory interneuron deficit links altered network activity and cognitive dysfunction in Alzheimer model."
    supports: SUPPORT
    evidence_source: MODEL_ORGANISM
    snippet: "Restoring Nav1.1 levels in hAPP mice by Nav1.1-BAC expression increased inhibitory synaptic activity and gamma oscillations and reduced hypersynchrony, memory deficits, and premature mortality."
    explanation: >-
      Gain-of-function rescue of a single interneuron-specific channel subunit
      corrects oscillations, hypersynchrony and memory, establishing interneuron
      failure as causal rather than correlative in this model.
  - reference: PMID:27696483
    reference_title: "Incidence and impact of subclinical epileptiform activity in Alzheimer's disease."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "Subclinical epileptiform activity was detected in 42.4% of AD patients and 10.5% of controls (p = 0.02)."
    explanation: >-
      Prospective, blinded extended EEG/MEG monitoring showing the predicted
      hyperexcitability is present in a large minority of patients with no seizure
      history.
  - reference: PMID:27696483
    reference_title: "Incidence and impact of subclinical epileptiform activity in Alzheimer's disease."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "However, patients with subclinical epileptiform activity showed faster declines in global cognition, determined by the Mini-Mental State Examination (3.9 points/year in patients with epileptiform activity vs 1.6 points/year in patients without; p = 0.006), and in executive function (p = 0.01)."
    explanation: >-
      Links the electrophysiological finding to the clinical outcome the model
      predicts it should affect.
  - reference: PMID:38263073
    reference_title: "Subclinical epileptiform activity in the Alzheimer continuum: association with disease, cognition and detection method."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "We found an increased prevalence of SEA in AD subjects (31%) as compared to controls (8%) (p = 0.041; Fisher's exact test), with increasing prevalence over the disease course (50% in dementia, 27% in MCI and 25% in preclinical AD)."
    explanation: >-
      Independent replication of the prevalence excess in a separate cohort, and
      an extension of it: subclinical epileptiform activity is already present in
      preclinical Alzheimer disease and rises across the continuum, which is what
      an early network mechanism predicts rather than a late consequence.
  - reference: PMID:21228179
    reference_title: "Amyloid-β/Fyn-induced synaptic, network, and cognitive impairments depend on tau levels in multiple mouse models of Alzheimer's disease."
    supports: SUPPORT
    evidence_source: MODEL_ORGANISM
    snippet: "Using electroencephalography to examine network effects, we found that tau reduction prevented spontaneous epileptiform activity in multiple lines of hAPP mice."
    explanation: >-
      A second, mechanistically independent perturbation reaching the same
      endpoint: removing tau, rather than restoring an interneuron channel,
      also abolishes epileptiform activity. Two unrelated interventions
      converging on restored inhibition is stronger than either alone.
  - reference: PMID:38356475
    reference_title: "The HOPE4MCI study: A randomized double-blind assessment of AGB101 for the treatment of MCI due to AD."
    supports: REFUTE
    evidence_source: HUMAN_CLINICAL
    snippet: "The estimated difference between arms is -0.10 (95% CI: -0.85, 0.58), which was not statistically significant."
    explanation: >-
      The largest and longest test of the model's therapeutic prediction — 164
      participants over 78 weeks, using a drug selected for its ability to
      normalize hippocampal hyperactivity — was negative on its primary endpoint
      in an unselected population. This is the single strongest reason the
      treatability claim cannot be called established.
  - reference: PMID:34570177
    reference_title: "Effect of Levetiracetam on Cognition in Patients With Alzheimer Disease With and Without Epileptiform Activity: A Randomized Clinical Trial."
    supports: REFUTE
    evidence_source: HUMAN_CLINICAL
    snippet: "Overall, treatment with levetiracetam did not change NIH-EXAMINER composite scores (mean difference vs placebo, 0.07 points; 95% CI, -0.18 to 0.32 points; P = .55) or secondary measures."
    explanation: >-
      Null on its primary endpoint in unselected patients. Read together with
      the negative HOPE4MCI trial, this argues against a treatability claim for
      Alzheimer disease generally.
  - reference: PMID:34570177
    reference_title: "Effect of Levetiracetam on Cognition in Patients With Alzheimer Disease With and Without Epileptiform Activity: A Randomized Clinical Trial."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "although it did not improve the primary outcome, in prespecified analysis, levetiracetam improved performance on spatial memory and executive function tasks in patients with AD and epileptiform activity."
    explanation: >-
      The benefit is confined to the prespecified epileptiform-positive
      subgroup, so any treatability is restricted to a hyperexcitable subgroup
      rather than to Alzheimer disease generally. This is a testable refinement
      of the model, not a rescue of it.
  - reference: PMID:22592800
    reference_title: "Critical role of soluble amyloid-β for early hippocampal hyperactivity in a mouse model of Alzheimer's disease."
    supports: SUPPORT
    evidence_source: MODEL_ORGANISM
    snippet: "Remarkably, in the hippocampus of young mice, we observed a selective increase in hyperactive neurons already before the formation of plaques, suggesting that soluble species of Aβ may underlie this impairment."
    explanation: >-
      Qualifies where the initiating lesion sits. Neuronal hyperactivity appears
      before plaques and is inducible by soluble amyloid-beta directly, so the
      network phenotype can arise upstream of, or in parallel with, interneuron
      failure rather than requiring it as the first step.
  notes: >-
    EMERGING, and the therapeutic arm should be read as qualified rather than
    merely "mixed". The mechanistic case in mice is strong: two independent
    perturbations — restoring Nav1.1 (PMID:22541439) and reducing tau
    (PMID:21228179) — each restore inhibition and abolish epileptiform activity.
    The human limb of the Nav1.1 claim remains a postmortem protein-level
    observation. The prevalence claim is now replicated in an independent cohort
    with detection across the Alzheimer continuum including preclinical disease
    (PMID:38263073), which strengthens it considerably beyond the single seed
    study.
    The treatability prediction is where the model is weakest. The largest and
    longest trial, HOPE4MCI (164 participants, 78 weeks, low-dose extended-release
    levetiracetam), was negative on its primary endpoint (PMID:38356475), and the
    LEV-AD crossover trial was null overall with benefit confined to its
    prespecified epileptiform-positive subgroup (PMID:34570177). The honest
    reading is that any benefit is subgroup-restricted and that this has not yet
    been tested prospectively in a biomarker-selected population.
    A competing account also has to be carried: neuronal hyperactivity appears
    before plaque formation and is inducible by soluble amyloid-beta directly
    (PMID:22592800), so interneuron failure may be one component of an
    amyloid-initiated network phenotype rather than its first step. Curated
    separately from synaptic_failure_convergence_model because it still makes the
    opposite claim about what fails first: inhibition, not excitation.
- hypothesis_group_id: asi_axis_reciprocal_cycle
  hypothesis_label: Autophagy-Senescence-Inflammasome (ASI) Axis
  status: EMERGING
  description: >-
    Impaired mitophagy, senescent glia, and inflammasome activation are modeled
    as reciprocally reinforcing rather than independent: defective mitochondrial
    quality control licenses inflammasome activation via mitochondrial danger
    signaling, and inflammasome/IL-1 signaling in turn reinforces the
    senescence-associated secretory phenotype, closing a self-sustaining cycle
    of neuroinflammation rather than a one-way cascade.
  applies_to_subtypes:
  - Early-Onset Alzheimer's Disease
  - Late-Onset Alzheimer's Disease
  evidence:
  - reference: PMID:42419491
    reference_title: "The autophagy-senescence-inflammasome axis: A novel triad in neurodegenerative diseases?"
    supports: SUPPORT
    quote_role: REVIEW_SYNTHESIS
    snippet: "we synthesise emerging evidence supporting an integrated 'Autophagy-Senescence-Inflammasome (ASI) axis', in which reciprocal interactions among impaired autophagy, senescent glia, and inflammasome signalling establish a self-sustaining cycle of neuroinflammation"
    explanation: >-
      States the paper's own proposed reciprocal triad structure. Cited only as
      the source of this hypothesis grouping, not as evidence for any individual
      causal edge below — this is a narrative review, and the edges each carry
      their own primary-study evidence.
  - reference: PMID:21151103
    reference_title: "Autophagy proteins regulate innate immune responses by inhibiting the release of mitochondrial DNA mediated by the NALP3 inflammasome."
    supports: SUPPORT
    evidence_source: IN_VITRO
    directness: INDIRECT
    snippet: "Depletion of autophagic proteins promoted the accumulation of dysfunctional mitochondria and cytosolic translocation of mitochondrial DNA (mtDNA) in response to lipopolysaccharide (LPS) and ATP in macrophages."
    explanation: >-
      Primary mechanistic support for the autophagy/mitochondria-to-inflammasome
      arm of the cycle, in macrophages rather than microglia specifically, so
      curated as indirect for this brain-specific entry.
  - reference: PMID:23770676
    reference_title: "A complex secretory program orchestrated by the inflammasome controls paracrine senescence."
    supports: SUPPORT
    evidence_source: IN_VITRO
    directness: INDIRECT
    snippet: "Treatment with Caspase 1 or IL-1R inhibitors (Fig 6i, S6i) but not TGFBRI inhibitors (Fig S6j), blunted the expression of SASP components during OIS, suggesting that activation of the inflammasome was a cause rather than an effect of the SASP."
    explanation: >-
      Primary mechanistic support for the inflammasome-to-senescence closing arm
      of the cycle, in an oncogene-induced-senescence fibroblast/epithelial
      model rather than CNS glia, so curated as indirect here.
  notes: >-
    EMERGING. Connects three mechanisms this entry already models separately —
    Mitochondrial Quality-Control Failure, Senescent Cell Accumulation, and the
    NLRP3 inflammasome-activation pathway on the Neuroinflammation node — into
    the reciprocal loop PMID:42419491 proposes. That review is cited only for
    the top-level claim (see evidence above); the two arms that close the loop
    (autophagy/mitochondria to inflammasome, and inflammasome back to
    senescence) are each supported by a real mechanistic study, but neither is
    an Alzheimer-specific or CNS-specific result, which is why this is EMERGING
    rather than ALTERNATIVE. The pre-existing senescence-to-neuroinflammation
    edge (SASP driving glial activation) is Alzheimer-model evidence already
    curated on the Senescent Cell Accumulation node.
- hypothesis_group_id: n_glycan_hyperglycosylation_model
  hypothesis_label: N-Glycan Hyperglycosylation Metabolic Driver Model
  status: EMERGING
  description: >-
    Excess protein N-glycosylation in the Alzheimer brain is modeled as an
    active contributor to cognitive decline rather than a passive by-product of
    neurodegeneration. Hexosamine-pathway flux into N-glycan biosynthesis
    increases, raising the net abundance of N-linked glycans (high-mannose and
    bisecting species prominent) in cortex and hippocampus; lowering that flux
    genetically (PGM3 knockdown) or pharmacologically (oligosaccharyltransferase
    inhibition) improves memory in amyloid and tau mouse models, and feeding the
    pathway with oral glucosamine worsens it. The model predicts that
    glucosamine supplementation is harmful specifically in people who already
    have the disease, not in the healthy brain.
  evidence:
  - reference: PMID:42265388
    reference_title: "Hyperglycosylation is a metabolic driver of Alzheimer's disease."
    supports: SUPPORT
    evidence_source: MODEL_ORGANISM
    snippet: "Genetic knockdown of glycan biosynthetic enzymes improves cognitive outcomes in AD mice whereas oral glucosamine supplementation impairs them."
    explanation: >-
      Bidirectional perturbation in the same model: lowering glycan biosynthesis
      helps and raising it hurts, which is the causal core of the model.
  - reference: PMID:42265388
    reference_title: "Hyperglycosylation is a metabolic driver of Alzheimer's disease."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "A retrospective analysis of electronic health records from patients with AD with varying disease severity shows that glucosamine supplementation is associated with accelerated AD progression and worsened survival."
    explanation: >-
      The human arm the model is built to explain. Observational, single health
      system, exposure ascertained from clinical notes; see notes.
  - reference: PMID:41042700
    reference_title: "Comprehensive and Site-Specific Characterization of Protein N-Glycosylation in AD Samples Reveals Its Potential Roles in Protein Aggregation and Synaptic Dysfunction."
    supports: REFUTE
    directness: INDIRECT
    evidence_source: HUMAN_CLINICAL
    snippet: "Our findings suggest that the loss of N-glycosylation may contribute to the pathogenesis of AD through impairing synaptic transmission and promoting protein aggregation."
    explanation: >-
      An independent site-specific glycoproteomic study of post-mortem AD brain
      (nine cases, nine controls) reaches the opposite directional conclusion:
      reduced glycosylation at synaptic and aggregation-prone sites. INDIRECT
      because it measures per-site occupancy on enriched glycopeptides, not the
      net abundance of released N-glycans the hyperglycosylation model rests on.
  notes: >-
    EMERGING, and it should stay there until replicated outside the originating
    laboratory. The case rests on one paper (PMID:42265388). Its human
    spatial-glycomics discovery cohort is three AD and three control frontal
    cortices, extended by a Braak-stratified FFPE series; the mechanistic and
    interventional work uses female 5xFAD and PS19 mice at 8-9 months with four
    to seven animals per arm, two-week interventions, and a single behavioural
    readout (social recognition memory). The PGM3 knockdown and NGI-1 effects on
    memory did not come with changes in amyloid-beta or GFAP, so the model
    places the effect off the amyloid and astrocytic axes rather than upstream
    of them. Two tensions must be carried. First, direction: site-specific
    glycoproteomics finds loss of N-glycosylation on synaptic proteins in AD
    brain (PMID:41042700), and the originating authors reconcile this by noting
    that they measure net released N-glycan abundance rather than peptide-level
    occupancy — a plausible reconciliation that no study has yet tested on the
    same tissue. Second, the originating authors themselves raise the
    possibility that increased glycosylation under cerebral glucose
    hypometabolism is compensatory, which is the opposite causal reading of the
    same observation; their perturbation data argue against it in mice but not
    in humans. The glucosamine human arm is discussed on the environmental
    entry, where large population cohorts point the other way for incident
    dementia.
- hypothesis_group_id: reactive_astrocyte_resilience_model
  hypothesis_label: Reactive Astrocyte Subtype Resilience Model
  status: EMERGING
  description: >-
    Cognitive resilience, the preservation of cognition despite Alzheimer
    neuropathology at a given Braak stage, is modeled as depending in part on
    which reactive state cortical astrocytes adopt. Reactive astrocyte subtypes
    expressing neuroprotective programs (for example EMP1 in the PLSCR1-marked
    subtype) are proposed to buffer neurons against pathology, while subtypes
    with a more severe, neurotoxic-skewed reactive profile (high FKBP5, MAOB and
    CHI3L1) are proposed to accompany the transition to dementia. The claim is
    that astrocyte reactive-state composition, not only amyloid and tau burden,
    separates resilient from cognitively impaired donors.
  applies_to_subtypes:
  - Late-Onset Alzheimer's Disease
  evidence:
  - reference: PMID:42778763
    reference_title: "AI-based characterization of Alzheimer's disease phenotypes from population-scale single-cell data."
    supports: SUPPORT
    evidence_source: COMPUTATIONAL
    quote_role: PRIMARY_RESULT
    snippet: "reactive astrocyte subtypes with altered neuroprotective and neurotoxic gene expression that likely confer cognitive resilience"
    explanation: >-
      The seed observation for this model, from computational phenotype scoring
      of prefrontal single-nucleus data in the PsychAD cohort. The authors'
      "likely confer" is an inference from association; no perturbation tests
      whether the astrocyte state causes resilience.
  - reference: PMID:37248300
    reference_title: "Astrocyte reactivity influences amyloid-β effects on tau pathology in preclinical Alzheimer's disease."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "Our findings suggest astrocyte reactivity as an important upstream event linking Aβ with initial tau pathology, which may have implications for the biological definition of preclinical AD and for selecting CU individuals for clinical trials."
    explanation: >-
      Independent human biomarker evidence that astrocyte reactivity modulates
      how amyloid translates into downstream pathology, the plausibility
      premise for astrocyte state shaping clinical expression.
  notes: >-
    EMERGING because the only direct support is one cohort's transcriptomic
    association, reported with the direction of causation left open. It is
    separate from neuroimmune_glial_amplification_model, which treats glial
    reactivity as amplifying injury; this model claims a reactive state can be
    protective and that the protective and harmful states are distinguishable
    subtypes. A re-analysis of the source paper's supplementary tables (see the
    OpenScientist assessment under kb/hypotheses/) found the aggregate
    neuroprotective-versus-neurotoxic marker contrast null, astrocytes ranked
    16th of 27 subclasses for separating resilient from impaired donors, and
    astrocyte scores rising with every neuronal subclass within donors; the
    EMP1 and FKBP5 directions come from AD-versus-control tables, not from a
    resilience contrast.
- hypothesis_group_id: glial_depression_in_ad_model
  hypothesis_label: Glial Inflammatory and ER-Stress Model of Depression in Alzheimer Disease
  status: EMERGING
  description: >-
    Depressive symptoms arising in people with Alzheimer disease are modeled as
    having a glial cellular substrate distinct from the general Alzheimer
    inflammatory signal: astrocyte states carrying
    inflammatory and endoplasmic reticulum stress programs are proposed to be
    enriched in Alzheimer donors with depression compared with Alzheimer donors
    without it, and to overlap with glial changes seen in major depressive
    disorder.
  applies_to_subtypes:
  - Late-Onset Alzheimer's Disease
  evidence:
  - reference: PMID:42778763
    reference_title: "AI-based characterization of Alzheimer's disease phenotypes from population-scale single-cell data."
    supports: SUPPORT
    evidence_source: COMPUTATIONAL
    quote_role: PRIMARY_RESULT
    snippet: "We also identified many PACs for multiple phenotypes, including the astrocytes between AD and depression showing specific gene expression patterns such as inflammation and endoplasmic reticulum stress pathways."
    explanation: >-
      The seed observation. The comparison was within Alzheimer donors (with
      versus without depression), so the shared Alzheimer signal is removed by
      design. It is an association from computational scoring and does not show
      that the glial state produces the mood symptom.
  notes: >-
    EMERGING on a single cohort. Whether depression within Alzheimer disease
    shares mechanism with major depressive disorder is stated by the source as
    unresolved. Linked to the Behavioral Changes phenotype and the
    Neuroinflammation node, but no causal edge is drawn because the direction
    between glial state and depressive symptoms is not established. A
    re-analysis of the source paper's supplementary tables (see the
    OpenScientist assessment under kb/hypotheses/) found the ER-stress limb not
    robust: unfolded-protein-response genes were enriched in the AD-Depression
    astrocyte cluster only against an assumed whole-genome background, not
    against the genes actually tested, none was among the depression-associated
    astrocyte DEGs, and HSPA5 and CALR appear among the general AD astrocyte
    DEGs.
- hypothesis_group_id: smorf_microprotein_model
  hypothesis_label: smORF-Encoded Microprotein Model
  status: EMERGING
  description: >-
    Thousands of small open reading frames (smORFs) that are absent from the
    reviewed reference proteome are nonetheless translated in the human frontal
    cortex, and the microproteins they encode carry their own Alzheimer-associated
    expression changes. The model holds that some of these microproteins take part
    in the disease rather than being incidental translation, and - the part that
    makes it a distinct model rather than a restatement - that a microprotein's
    regulation can be decoupled from the annotated main ORF at the same locus.
    Conventional proteomic and transcriptomic workflows quantify the annotated
    product, so a microprotein that falls in Alzheimer disease while its host
    locus does not is invisible to them. The model therefore predicts a class of
    disease-associated protein changes that existing Alzheimer omics could not
    have detected, and it is testable in exactly that form: rebuild the search
    space to include smORFs and the changes should appear.
  applies_to_subtypes:
  - Late-Onset Alzheimer's Disease
  evidence:
  - reference: PMID:42736449
    reference_title: A microprotein atlas of the human frontal cortex in Alzheimer's disease.
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    directness: DIRECT
    quote_role: PRIMARY_RESULT
    snippet: "Of 3,001 unreviewed MPs amenable to PROSIT interrogation, 1,067 (35.6%) were strong, 1,433 (47.8%) moderate, 343 (11.4%) weak and 158 (5.3%) insufficient."
    explanation: >-
      The premise the model rests on, at scale and with a stated confidence
      gradient: over a thousand microproteins absent from the reviewed reference
      proteome carry strong-grade spectral support in postmortem human frontal
      cortex, graded against deep-learning-predicted spectra rather than asserted.
  - reference: PMID:42736449
    reference_title: A microprotein atlas of the human frontal cortex in Alzheimer's disease.
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    directness: DIRECT
    quote_role: PRIMARY_RESULT
    snippet: "we found that 39% of testable pairs showed disease-associated changes independent of the mORF"
    explanation: >-
      The decoupling claim, which is what separates this model from known
      locus-level changes: for a large minority of smORF-mORF pairs the
      Alzheimer-associated change is not recoverable by measuring the annotated
      product at the same locus.
  notes: >-
    Curated as EMERGING, not ALTERNATIVE. The model competes with nothing - it
    does not displace the amyloid, tau, or glial models, and it makes no claim
    about their ordering. What it asserts is that an entire layer of the
    protein-coding output has gone unmeasured in Alzheimer proteomics, and that
    at least one member of that layer has a demonstrable cellular function. The
    weak point is the step from association to mechanism: the atlas establishes
    that microproteins change in Alzheimer disease, and a single case
    (micro-MKKS63) carries a functional experiment behind it. The other roughly
    thousand do not yet.
pathophysiology:
- name: Amyloid Plaque Formation
  description: Accumulation of amyloid-beta proteins in the brain, forming extracellular plaques that disrupt cell function and communication between neurons.
  cell_types:
  - preferred_term: Neurons
    term:
      id: CL:0000540
      label: neuron
  - preferred_term: Microglia
    term:
      id: CL:0000129
      label: microglial cell
  - preferred_term: Astrocytes
    term:
      id: CL:0000127
      label: astrocyte
  biological_processes:
  - preferred_term: Protein misfolding
    modifier: ABNORMAL
    term:
      id: GO:0006457
      label: protein folding
  - preferred_term: Amyloid-beta generation by gamma-secretase cleavage of APP
    term:
      id: GO:0034205
      label: amyloid-beta formation
  - preferred_term: Amyloid-beta aggregation into fibrils
    modifier: INCREASED
    term:
      id: GO:1990000
      label: amyloid fibril formation
  - preferred_term: Neuroinflammation
    term:
      id: GO:0150076
      label: neuroinflammatory response
  locations:
  - preferred_term: Cerebral Cortex
  - preferred_term: Hippocampus
  - preferred_term: Subcortical Regions
  chemical_entities:
  - preferred_term: Amyloid-beta peptides (Aβ40, Aβ42)
  genes:
  - preferred_term: APP
    term:
      id: hgnc:620
      label: APP
  - preferred_term: PSEN1
    term:
      id: hgnc:9508
      label: PSEN1
  - preferred_term: PSEN2
    term:
      id: hgnc:9509
      label: PSEN2
  pathways:
  - preferred_term: Amyloid precursor protein processing
    term:
      id: GO:0042982
      label: amyloid precursor protein metabolic process
  mechanisms:
  - Beta-secretase (BACE1) cleavage of APP
  - Gamma-secretase cleavage of APP C-terminal fragment
  consequences:
  - Synaptic Dysfunction
  - Neuroinflammation
  evidence:
  - reference: PMID:28320296
    reference_title: "Biological Basis for Amyloidogenesis in Alzheimer's Disease."
    supports: SUPPORT
    snippet: These intra- or extracellular insoluble aggregates (fibers or plaques) are hallmarks of many neurodegenerative pathologies including Alzheimer's disease (AD)...
    explanation: This reference supports the statement by confirming that amyloid plaques are a hallmark of Alzheimer's disease and discusses the formation and role of amyloid aggregates in AD.
  - reference: PMID:26322584
    reference_title: "η-Secretase processing of APP inhibits neuronal activity in the hippocampus."
    supports: SUPPORT
    snippet: Alzheimer disease (AD) is characterized by the accumulation of amyloid plaques, which are predominantly composed of amyloid-beta peptide.
    explanation: This reference supports the statement by indicating that the accumulation of amyloid plaques, composed of amyloid-beta peptide, is a characteristic of Alzheimer's disease.
  - reference: PMID:22351073
    reference_title: "Synaptic dysfunction in Alzheimer's disease."
    supports: SUPPORT
    snippet: Generation of amyloid peptide (Abeta) is at the beginning of a cascade that leads to Alzheimer's disease (AD)... soluble assembly states of Abeta peptides can cause cognitive problems by disrupting synaptic function...
    explanation: This reference supports the statement by mentioning that amyloid-beta peptides lead to Alzheimer's disease and cause synaptic dysfunction.
  - reference: PMID:10911965
    reference_title: "Transgenic mouse models of Alzheimer's disease."
    supports: SUPPORT
    evidence_source: MODEL_ORGANISM
    snippet: Alzheimer's disease (AD) pathology is characterized by A beta peptide-containing plaques, neurofibrillary tangles consisting of hyperphosphorylated tau, extensive neuritic degeneration, and distinct neuron loss.
    explanation: This reference supports the statement by describing Alzheimer's disease pathology, including amyloid-beta peptide-containing plaques.
  - reference: PMID:10702374
    reference_title: "Cellular and molecular basis of beta-amyloid precursor protein metabolism."
    supports: SUPPORT
    snippet: beta-amyloid (Abeta) is the main constituent of the amyloidogenic plaques which are a primary pathological hallmark of Alzheimer's disease...
    explanation: This reference supports the statement by confirming that beta-amyloid is the main constituent of amyloid plaques, a hallmark of Alzheimer's disease.
  - reference: PMID:36555791
    reference_title: "PS1 Affects the Pathology of Alzheimer's Disease by Regulating BACE1 Distribution in the ER and BACE1 Maturation in the Golgi Apparatus."
    supports: SUPPORT
    snippet: Neuritic plaques are one of the major pathological hallmarks of Alzheimer's disease. They are formed by the aggregation of extracellular amyloid-beta protein (Abeta)...
    explanation: This reference supports the statement by indicating that neuritic plaques, formed by the aggregation of amyloid-beta protein, are a hallmark of Alzheimer's disease.
  - reference: PMID:25941885
    reference_title: "Preventing the spread of Alzheimer's disease neuropathology: a role for calcilytics?"
    supports: SUPPORT
    snippet: The 'amyloid cascade hypothesis' posits that an extracellular build-up of amyloid-beta oligomers (Abeta-os) and polymers (fibrils) subsequently inducing toxic hyperphosphorylated (p)-Tau oligomers (p-Tau-os) and neurofibrillary tangles starts the sporadic late-onset Alzheimer's disease (LOAD)...
    explanation: This reference supports the statement by discussing the amyloid cascade hypothesis, which involves the build-up of amyloid-beta leading to Alzheimer's disease.
  - reference: PMID:16611586
    reference_title: "The involvement of lipid rafts in Alzheimer's disease."
    supports: SUPPORT
    snippet: The amyloidogenesis occurring in Alzheimer's disease represents a fundamental membrane-related pathology... the amyloid-beta peptide (Abeta), which accumulates extracellularly as plaques in the brains of Alzheimer's disease patients...
    explanation: This reference supports the statement by describing the accumulation of amyloid-beta peptide as plaques in Alzheimer's disease.
  downstream:
  - target: Neurofibrillary Tangle Formation
    description: Amyloid-beta accumulation is modeled as an upstream trigger of tau hyperphosphorylation and neurofibrillary tangle formation.
    hypothesis_groups:
    - amyloid_cascade_model
    - tau_neurodegeneration_model
    causal_link_type: INDIRECT_KNOWN_INTERMEDIATES
    intermediate_mechanisms:
    - Amyloid-beta oligomer and fibril toxicity inducing tau hyperphosphorylation.
    evidence:
    - reference: PMID:25941885
      reference_title: "Preventing the spread of Alzheimer's disease neuropathology: a role for calcilytics?"
      supports: SUPPORT
      snippet: "The 'amyloid cascade hypothesis' posits that an extracellular build-up of amyloid-beta oligomers (Abeta-os) and polymers (fibrils) subsequently inducing toxic hyperphosphorylated (p)-Tau oligomers (p-Tau-os) and neurofibrillary tangles starts the sporadic late-onset Alzheimer's disease (LOAD)..."
      explanation: Supports the modeled edge from amyloid-beta buildup to downstream tau pathology.
  - target: Synaptic Dysfunction
    description: Soluble amyloid-beta assemblies are modeled as direct toxic contributors to synaptic dysfunction and cognitive impairment.
    hypothesis_groups:
    - amyloid_cascade_model
    - synaptic_failure_convergence_model
    causal_link_type: DIRECT
    evidence:
    - reference: PMID:22351073
      reference_title: "Synaptic dysfunction in Alzheimer's disease."
      supports: SUPPORT
      snippet: "Generation of amyloid peptide (Abeta) is at the beginning of a cascade that leads to Alzheimer's disease (AD)... soluble assembly states of Abeta peptides can cause cognitive problems by disrupting synaptic function..."
      explanation: Directly links soluble amyloid-beta assemblies to synaptic dysfunction and cognitive problems.
  - target: Neuroinflammation
    description: Amyloid-beta accumulation and activated microglia can form a reinforcing inflammatory cycle.
    hypothesis_groups:
    - amyloid_cascade_model
    - neuroimmune_glial_amplification_model
    causal_link_type: INDIRECT_KNOWN_INTERMEDIATES
    intermediate_mechanisms:
    - Microglial activation and inflammatory mediator release.
    evidence:
    - reference: PMID:23930978
      reference_title: "Microglia, neuroinflammation, and beta-amyloid protein in Alzheimer's disease."
      supports: SUPPORT
      snippet: "A vicious cycle of inflammation has been formed between Abeta accumulation, activated microglia, and microglial inflammatory mediators, which enhance Abeta deposition and neuroinflammation."
      explanation: Supports a reinforcing edge between amyloid-beta accumulation and neuroinflammatory activation.
- name: Neurofibrillary Tangle Formation
  description: Intracellular accumulation of hyperphosphorylated tau protein, forming twisted fibers that disrupt cellular transport and eventually lead to neuronal death.
  cell_types:
  - preferred_term: Neurons
    term:
      id: CL:0000540
      label: neuron
  biological_processes:
  - preferred_term: Protein hyperphosphorylation
    modifier: INCREASED
    term:
      id: GO:0006468
      label: protein phosphorylation
  - preferred_term: Microtubule destabilization
    term:
      id: GO:0007019
      label: microtubule depolymerization
  - preferred_term: Neurofibrillary tangle assembly
    modifier: INCREASED
    term:
      id: GO:1902988
      label: neurofibrillary tangle assembly
  molecular_functions:
  - preferred_term: Tau-protein kinase activity (GSK3-beta and other tau kinases)
    modifier: INCREASED
    term:
      id: GO:0050321
      label: tau-protein kinase activity
  locations:
  - preferred_term: Entorhinal Cortex
  - preferred_term: Hippocampus
  - preferred_term: Neocortex
  chemical_entities:
  - preferred_term: Tau protein
  genes:
  - preferred_term: MAPT
    term:
      id: hgnc:6893
      label: MAPT
  - preferred_term: GSK3B
    term:
      id: hgnc:4617
      label: GSK3B
  pathways:
  - preferred_term: Tau phosphorylation cascade
    term:
      id: GO:0006468
      label: protein phosphorylation
  mechanisms:
  - Hyperphosphorylation of tau by various kinases
  - Aggregation of hyperphosphorylated tau
  consequence: Synaptic Dysfunction
  evidence:
  - reference: PMID:21509508
    reference_title: "Tau mediated neurodegeneration: an insight into Alzheimer's disease pathology."
    supports: SUPPORT
    snippet: Extracellular accumulations of Abeta, hyperphosphorylation of tau and intracellular neurofibrillary tangle formation have been the hallmarks of Alzheimer's Disease (AD).
    explanation: The article discusses the role of tau hyperphosphorylation and neurofibrillary tangle formation in Alzheimer's Disease, supporting the statement about these processes.
  - reference: PMID:36001963
    reference_title: "Disruption of nuclear envelope integrity as a possible initiating event in tauopathies."
    supports: SUPPORT
    snippet: In Alzheimer's disease and other neurodegenerative tauopathies, tau is found hyperphosphorylated and aggregated in neurofibrillary tangles.
    explanation: This article supports the statement by mentioning the hyperphosphorylation and aggregation of tau in neurofibrillary tangles in Alzheimer's Disease.
  - reference: PMID:37266762
    reference_title: "The Role of Tau Proteoforms in Health and Disease."
    supports: SUPPORT
    snippet: Tau is a microtubule-associated binding protein in the nervous system that is known for its role in stabilizing microtubules throughout the nerve cell. It accumulates as beta-sheet-rich aggregates and neurofibrillary tangles, leading to an array of different pathologies.
    explanation: The article confirms that hyperphosphorylated tau leads to the formation of neurofibrillary tangles and disrupts microtubule stabilization, supporting the statement.
  - reference: PMID:19542604
    reference_title: "The microtubule-associated protein tau is also phosphorylated on tyrosine."
    supports: SUPPORT
    snippet: Tau protein is the principal component of the neurofibrillary tangles found in Alzheimer's disease (AD), where it is hyperphosphorylated on serine and threonine residues.
    explanation: This article supports the statement by identifying hyperphosphorylated tau as the main component of neurofibrillary tangles in Alzheimer's Disease.
  - reference: PMID:28100644
    reference_title: "Multicellular hypothesis for the pathogenesis of Alzheimer's disease."
    supports: SUPPORT
    snippet: These multicellular interactions are initiated by insoluble tangles of phosphorylated tau protein and plaques of amyloid peptides.
    explanation: This article supports the statement by discussing how phosphorylated tau protein tangles initiate multicellular interactions in Alzheimer's Disease.
  - reference: PMID:2135393
    reference_title: "Tau protein and neurodegeneration."
    supports: SUPPORT
    snippet: The most common of these conditions is Alzheimer's disease, in which microtubules are lost from neurites that fill up with filamentous structures. One component of the filamentous structures is the microtubule-associated protein (MAP), tau.
    explanation: The article supports the statement by explaining how tau protein is involved in the formation of filamentous structures in Alzheimer's Disease, leading to microtubule destabilization.
  - reference: PMID:12859672
    reference_title: "Hyperphosphorylation and aggregation of tau in mice expressing normal human tau isoforms."
    supports: SUPPORT
    evidence_source: MODEL_ORGANISM
    snippet: Neurofibrillary tangles are composed of insoluble aggregates of the microtubule-associated protein tau. In Alzheimer's disease the accumulation of neurofibrillary tangles occurs in the absence of tau mutations.
    explanation: This article supports the statement by describing the composition and formation of neurofibrillary tangles from tau protein in Alzheimer's Disease.
  - reference: PMID:31903881
    reference_title: "Targeting Tau Hyperphosphorylation via Kinase Inhibition: Strategy to Address Alzheimer's Disease."
    supports: SUPPORT
    snippet: Hyperphosphorylation of tau protein leads to aggregation of tau into paired helical filaments like structures which are major constituents of neurofibrillary tangles, a hallmark of Alzheimer's disease.
    explanation: The article supports the statement by linking tau hyperphosphorylation and aggregation to the formation of neurofibrillary tangles in Alzheimer's Disease.
  - reference: PMID:31903881
    reference_title: "Targeting Tau Hyperphosphorylation via Kinase Inhibition: Strategy to Address Alzheimer's Disease."
    supports: SUPPORT
    quote_role: REVIEW_SYNTHESIS
    evidence_source: HUMAN_CLINICAL
    snippet: In pathological conditions, abnormal activity of tau kinases such as glycogen synthase kinase-3 β (GSK3β), cyclin-dependent kinase 5 (CDK5), c-Jun N-terminal kinases (JNKs), extracellular signal-regulated kinase 1 and 2 (ERK1/2) and microtubule affinity regulating kinase (MARK) lead to tau hyperphosphorylation.
    explanation: Names GSK3-beta among the tau kinases whose abnormal activity drives tau hyperphosphorylation, supporting GSK3B and tau-protein kinase activity on this node.
  - reference: PMID:23948895
    reference_title: "Tauopathies and tau oligomers."
    supports: SUPPORT
    snippet: Tauopathies are neurodegenerative diseases characterized behaviorally by dementia and neuropathologically by neurofibrillary tangles and neuronal loss.
    explanation: The article supports the statement by describing tauopathies, including Alzheimer's Disease, as being characterized by neurofibrillary tangles and neuronal loss.
  downstream:
  - target: Synaptic Dysfunction
    description: Hyperphosphorylated tau and tau aggregates are modeled as toxic contributors to impaired synaptic plasticity and neurotransmitter release.
    hypothesis_groups:
    - tau_neurodegeneration_model
    - synaptic_failure_convergence_model
    causal_link_type: INDIRECT_KNOWN_INTERMEDIATES
    intermediate_mechanisms:
    - Tau-mediated microtubule disruption and synaptic toxicity.
    evidence:
    - reference: PMID:27662312
      reference_title: "Stress-Induced Synaptic Dysfunction and Neurotransmitter Release in Alzheimer's Disease: Can Neurotransmitters and Neuromodulators be Potential Therapeutic Targets?"
      supports: SUPPORT
      snippet: "Compelling evidence suggests that soluble amyloid-beta (Abeta) and hyperphosphorylated tau serve as toxins in the dysfunction of synaptic plasticity and aberrant neurotransmitter (NT) release at synapses consequently causing a cognitive decline in Alzheimer's disease (AD)."
      explanation: Supports hyperphosphorylated tau as a synaptic toxin in Alzheimer disease.
  - target: Necroptotic Neuronal Death
    description: >-
      Hyperphosphorylated tau is modeled as an upstream trigger of necrosome
      assembly, connecting the entry's central proteinopathy to the execution
      step by which neurons are proposed to die.
    hypothesis_groups:
    - necroptosis_model
    - tau_neurodegeneration_model
    causal_link_type: DIRECT
    evidence:
    - reference: PMID:35971179
      reference_title: "Hyperphosphorylated tau mediates neuronal death by inducing necroptosis and inflammation in Alzheimer's disease."
      supports: SUPPORT
      evidence_source: IN_VITRO
      snippet: "Hyperphosphorylated tau could induce necroptosis in neuronal cells by promoting the formation of the RIPK1/RIPK3/MLKL necrosome."
      explanation: >-
        Establishes phosphorylated tau as sufficient to assemble the necrosome in
        neuronal cells, which is the mechanistic content of this edge.
    - reference: PMID:35971179
      reference_title: "Hyperphosphorylated tau mediates neuronal death by inducing necroptosis and inflammation in Alzheimer's disease."
      supports: SUPPORT
      evidence_source: MODEL_ORGANISM
      snippet: "Furthermore, necroptosis stimulation, NF-κB activation, and cytokine induction have been detected in TauP301S mice and blocking necroptosis markedly ameliorated behavioral defects and excessive neuroinflammation in AD mice."
      explanation: >-
        In vivo confirmation with a pharmacological rescue: blocking necroptosis
        in a tauopathy model improves behaviour and reduces neuroinflammation.
  - target: Intercellular Tau Transmission via Extracellular Vesicles
    description: Intracellular pathological tau is packaged by Arc into neuronal extracellular vesicles and released, providing a route for cell-to-cell spread of tau seeds.
    hypothesis_groups:
    - ev_mediated_tau_propagation_model
    - tau_neurodegeneration_model
    causal_link_type: DIRECT
    evidence:
    - reference: PMID:42372723
      reference_title: "Arc mediates intercellular tau transmission via extracellular vesicles."
      supports: SUPPORT
      evidence_source: MODEL_ORGANISM
      snippet: "Brain EVs purified from transgenic rTg4510 mutant tau mice (rTgWT) crossed with Arc knockout mice (rTgArc KO) contain less tau and reduced tau seeding potential."
      explanation: >-
        Pathological tau is loaded into brain-derived EVs in an Arc-dependent
        manner, linking intraneuronal tangle pathology to EV-mediated tau
        release.
- name: Synaptic Dysfunction
  description: Progressive loss of synapses and impaired neurotransmitter signaling, leading to disrupted neuronal communication and cognitive decline.
  cell_types:
  - preferred_term: Neurons
    term:
      id: CL:0000540
      label: neuron
  biological_processes:
  - preferred_term: Neurotransmitter release
    modifier: DECREASED
    term:
      id: GO:0007269
      label: neurotransmitter secretion
  - preferred_term: Synaptic plasticity
    modifier: DECREASED
    term:
      id: GO:0048167
      label: regulation of synaptic plasticity
  locations:
  - preferred_term: Synapses
  - preferred_term: Neuronal dendrites
  chemical_entities:
  - preferred_term: Acetylcholine
    term:
      id: CHEBI:15355
      label: acetylcholine
  - preferred_term: Glutamate
    term:
      id: CHEBI:14321
      label: glutamate(1-)
  - preferred_term: GABA
    term:
      id: CHEBI:16865
      label: gamma-aminobutyric acid
  pathways:
  - preferred_term: Synaptic vesicle recycling
    term:
      id: GO:0099504
      label: synaptic vesicle cycle
  - preferred_term: Neurotransmitter receptor signaling
    term:
      id: GO:0050804
      label: modulation of chemical synaptic transmission
  consequence: Oxidative Stress
  evidence:
  - reference: PMID:27662312
    reference_title: "Stress-Induced Synaptic Dysfunction and Neurotransmitter Release in Alzheimer's Disease: Can Neurotransmitters and Neuromodulators be Potential Therapeutic Targets?"
    supports: SUPPORT
    snippet: Compelling evidence suggests that soluble amyloid-beta (Abeta) and hyperphosphorylated tau serve as toxins in the dysfunction of synaptic plasticity and aberrant neurotransmitter (NT) release at synapses consequently causing a cognitive decline in Alzheimer's disease (AD).
    explanation: The statement is supported by the reference which discusses the role of neurotransmitter release and synaptic plasticity in Alzheimer's Disease, leading to cognitive decline.
  - reference: PMID:12973746
    reference_title: "Glutamatergic systems in Alzheimer's disease."
    supports: SUPPORT
    snippet: Histological studies indicate loss of pyramidal neurones and their synapses in Alzheimer's disease (AD), this together with biochemical evidence suggests presynaptic (and postsynaptic) glutamatergic hypoactivity.
    explanation: The reference supports the statement by discussing the loss of synapses and impaired glutamatergic neurotransmitter signaling in Alzheimer's Disease.
  - reference: PMID:33232936
    reference_title: "Alterations of GABA B receptors in the APP/PS1 mouse model of Alzheimer's disease."
    supports: SUPPORT
    evidence_source: MODEL_ORGANISM
    snippet: In the present study, we examined alterations in several GABA-specific targets in the APP/PS1 mouse model at different ages... Overall, our study provides evidence of altered GABAergic signaling in an amyloid model of AD at a time point consistent with AD-related deficits.
    explanation: The reference supports the statement by providing evidence of altered GABAergic signaling and its role in Alzheimer's Disease.
  downstream:
  - target: Cortical Structural Connectivity Degeneration
    description: >-
      Early synaptic alterations precede structural degradation in vulnerable
      medial-temporal white-matter pathways along the aging-dementia continuum.
    causal_link_type: INDIRECT_KNOWN_INTERMEDIATES
    intermediate_mechanisms:
    - Deafferentation and axonal degeneration in vulnerable white-matter pathways.
    evidence:
    - reference: PMID:23380151
      reference_title: "Perturbations of neural circuitry in aging, mild cognitive impairment, and Alzheimer's disease."
      supports: SUPPORT
      evidence_source: HUMAN_CLINICAL
      snippet: "The earliest changes are synaptic in nature and are detectable in particularly vulnerable white matter pathways such as the perforant path. These are followed by structural degradation in the transentorhinal region and subsequently neurodegeneration of the hippocampus as a result of accumulating pathology as well as deafferentation from entorhinal input."
      explanation: A human-neuroimaging review proposes a temporal sequence from early synaptic changes to structural degradation and deafferentation in the medial temporal network.
  - target: Memory Loss
    description: Synaptic plasticity failure and aberrant neurotransmitter release are modeled as proximate drivers of cognitive decline, including memory impairment.
    hypothesis_groups:
    - synaptic_failure_convergence_model
    causal_link_type: INDIRECT_KNOWN_INTERMEDIATES
    intermediate_mechanisms:
    - Hippocampal and cortical network dysfunction.
    evidence:
    - reference: PMID:27662312
      reference_title: "Stress-Induced Synaptic Dysfunction and Neurotransmitter Release in Alzheimer's Disease: Can Neurotransmitters and Neuromodulators be Potential Therapeutic Targets?"
      supports: SUPPORT
      snippet: "Compelling evidence suggests that soluble amyloid-beta (Abeta) and hyperphosphorylated tau serve as toxins in the dysfunction of synaptic plasticity and aberrant neurotransmitter (NT) release at synapses consequently causing a cognitive decline in Alzheimer's disease (AD)."
      explanation: Supports synaptic dysfunction as a proximate mechanism for cognitive decline.
- name: Cortical Structural Connectivity Degeneration
  description: >-
    Loss and reorganization of long-range cortical connections disrupt
    large-scale brain-network communication during the preclinical and clinical
    Alzheimer disease continuum.
  locations:
  - preferred_term: Cerebral cortex
    term:
      id: UBERON:0000956
      label: cerebral cortex
  - preferred_term: White matter
    term:
      id: UBERON:0002316
      label: white matter
  biological_scale: TISSUE
  downstream:
  - target: Memory Loss
    description: >-
      Progressive grey- and white-matter network disconnection accompanies the
      transition from mild memory impairment to broader cognitive decline.
    causal_link_type: INDIRECT_KNOWN_INTERMEDIATES
    intermediate_mechanisms:
    - Loss of efficient communication across memory-related cortical networks.
    evidence:
    - reference: PMID:23294972
      reference_title: "Structural networks in Alzheimer's disease."
      supports: SUPPORT
      evidence_source: HUMAN_CLINICAL
      snippet: "On the other hand, much progress has been made in characterizing the longitudinal progression of AD pathology, which includes the formation of \"plaques and tangles\", a distinct topological pattern of atrophy of grey and white matter, and the concurrent decline of specific cognitive functions, beginning with mild memory impairments and ending with general debilitating dementia."
      explanation: A human structural-network review links progressive grey- and white-matter network atrophy with the course from mild memory impairment to dementia.
  evidence:
  - reference: PMID:40450374
    reference_title: "Digital twins and non-invasive recordings enable early diagnosis of Alzheimer's disease."
    supports: SUPPORT
    evidence_source: COMPUTATIONAL
    snippet: "DADD parameters captured Alzheimer’s-related micro- and macroscopic structural neurodegeneration mechanisms (see Methods)."
    explanation: The patient-specific model identifies structural connectivity degeneration as an Alzheimer-related mechanism represented in its digital biomarkers.
- name: Neuroinflammation
  description: Chronic activation of immune responses in the brain, contributing to neuronal damage and disease progression.
  conforms_to: "neuroinflammation_glial_activation#Reactive Microglial and Astrocytic Activation"
  cell_types:
  - preferred_term: Microglia
    term:
      id: CL:0000129
      label: microglial cell
  - preferred_term: Astrocytes
    term:
      id: CL:0000127
      label: astrocyte
  biological_processes:
  - preferred_term: Immune activation
    modifier: INCREASED
    term:
      id: GO:0002253
      label: activation of immune response
  - preferred_term: Microglial cell activation
    modifier: INCREASED
    term:
      id: GO:0001774
      label: microglial cell activation
  - preferred_term: Astrocyte activation
    modifier: INCREASED
    term:
      id: GO:0048143
      label: astrocyte activation
  - preferred_term: Cytokine production
    modifier: INCREASED
    term:
      id: GO:0001816
      label: cytokine production
  - preferred_term: Phagocytosis
    term:
      id: GO:0006909
      label: phagocytosis
  molecular_functions:
  - preferred_term: Inflammasome adaptor activity (NLRP3 PYD filament nucleation; ASC/PYCARD bridging to caspase-1)
    term:
      id: GO:0035591
      label: signaling adaptor activity
  - preferred_term: Caspase-1 activation by the assembled inflammasome (NLRP3, ASC/PYCARD)
    term:
      id: GO:0140608
      label: cysteine-type endopeptidase activator activity
  - preferred_term: TREM2-TYROBP microglial receptor signaling
    term:
      id: GO:0004888
      label: transmembrane signaling receptor activity
  chemical_entities:
  - preferred_term: Cytokines
  - preferred_term: Chemokines
  - preferred_term: Complement proteins
  pathways:
  - preferred_term: NF-κB signaling
    term:
      id: GO:0007249
      label: canonical NF-kappaB signal transduction
  - preferred_term: NLRP3 inflammasome activation
    term:
      id: GO:0044546
      label: NLRP3 inflammasome complex assembly
  genes:
  - preferred_term: TREM2
    term:
      id: hgnc:17761
      label: TREM2
  - preferred_term: NLRP3
    term:
      id: hgnc:16400
      label: NLRP3
  - preferred_term: PYCARD (ASC)
    term:
      id: hgnc:16608
      label: PYCARD
  consequences:
  - Oxidative Stress
  - Vascular Dysfunction
  evidence:
  - reference: PMID:28019679
    reference_title: "Inflammasome activation and innate immunity in Alzheimer's disease."
    supports: SUPPORT
    snippet: NOD-like receptor (NLR) family, pyrin domain containing 3 and 1 inflammasomes, present in myeloid cells and neurons, respectively, represent key components of the innate immune reaction observed in Alzheimer patient brains.
    explanation: The reference supports the involvement of NLRP3 inflammasome activation in Alzheimer's disease, which is part of the immune activation process contributing to neuroinflammation.
  - reference: PMID:32061803
    reference_title: "Cyclical amyloid beta-astrocyte activity induces oxidative stress in Alzheimer's disease."
    supports: SUPPORT
    snippet: The response of astrocytes to the presence of Abeta, as well astrocytic and microglial interaction and inflammatory cytokine release is also discussed, highlighting a cyclical behaviour of these cells in contributing to AD pathogenesis.
    explanation: This reference supports the role of astrocytes and microglia in cytokine production and their interaction contributing to Alzheimer's disease pathology.
  - reference: PMID:35406803
    reference_title: "Microglial Endocannabinoid Signalling in AD."
    supports: SUPPORT
    snippet: Once activated, microglial cells, which are brain-resident immune cells, exert several key actions, including phagocytosis, chemotaxis, and the release of pro- or anti-inflammatory mediators.
    explanation: This reference supports the involvement of microglia in immune activation, cytokine production, and phagocytosis in Alzheimer's disease.
  - reference: PMID:21546088
    reference_title: "Complement in the brain."
    supports: SUPPORT
    snippet: In age related diseases, such as Alzheimer's disease (AD), accumulating amyloid proteins elicit complement activation and a local, chronic inflammatory response that leads to attraction and activation of glial cells that, under such activation conditions, can produce neurotoxic substances, including pro-inflammatory cytokines and oxygen radicals.
    explanation: This reference supports the involvement of complement proteins, cytokines, and oxidative stress in Alzheimer's disease.
  - reference: PMID:23930978
    reference_title: "Microglia, neuroinflammation, and beta-amyloid protein in Alzheimer's disease."
    supports: SUPPORT
    snippet: A vicious cycle of inflammation has been formed between Abeta accumulation, activated microglia, and microglial inflammatory mediators, which enhance Abeta deposition and neuroinflammation.
    explanation: Review of microglia dual roles in AD pathogenesis, describing the vicious cycle between Aβ accumulation, microglial activation, and inflammatory mediator release.
  - reference: PMID:36361780
    reference_title: "Microglia and Alzheimer's Disease."
    supports: SUPPORT
    evidence_source: OTHER
    snippet: "Extracellular amyloid and/or intraneuronal phosphorylated tau in AD can both activate microglia."
    explanation: >-
      Review-level support that both amyloid-beta and phosphorylated tau can sit
      upstream of microglial activation in Alzheimer disease.
  - reference: PMID:37308616
    reference_title: "Functional roles of reactive astrocytes in neuroinflammation and neurodegeneration."
    supports: SUPPORT
    evidence_source: OTHER
    snippet: "This network could involve phenotypic alteration of multiple cell types in the CNS, including astrocytes, which have a major neurosupportive, homeostatic role in the healthy CNS but adopt reactive states under acute or chronic adverse conditions."
    explanation: >-
      Supports modeling astrocytic reactivity as part of a broader pathological
      neuroinflammatory network rather than as an isolated downstream marker.
  - reference: PMID:36357946
    reference_title: "Microglia as Central Protagonists in the Chronic Stress Response."
    supports: SUPPORT
    snippet: Furthermore, exposure to chronic stress alters the phenotype of microglia, a population of innate immune cells that reside in the CNS parenchyma.
    explanation: This reference supports the role of microglia as key players in immune activation and neuroinflammation.
  - reference: PMID:37248300
    reference_title: "Astrocyte reactivity influences amyloid-β effects on tau pathology in preclinical Alzheimer's disease."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "Here, in a biomarker study across three cohorts (n = 1,016), we tested whether astrocyte reactivity modulates the association of Aβ with tau phosphorylation in CU individuals."
    explanation: >-
      Large multi-cohort human biomarker study establishing that astrocyte
      reactivity is not merely a downstream marker but gates whether amyloid
      burden translates into tau phosphorylation in cognitively unimpaired people.
  - reference: PMID:37248300
    reference_title: "Astrocyte reactivity influences amyloid-β effects on tau pathology in preclinical Alzheimer's disease."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "Our findings suggest astrocyte reactivity as an important upstream event linking Aβ with initial tau pathology, which may have implications for the biological definition of preclinical AD and for selecting CU individuals for clinical trials."
    explanation: >-
      Positions glial reactivity upstream of the amyloid-to-tau transition rather
      than downstream of established pathology. Astrocyte reactivity is
      operationalized as plasma GFAP positivity and the design is observational,
      so "upstream" is the authors' inference from a longitudinal association.
  - reference: PMID:42778763
    reference_title: "AI-based characterization of Alzheimer's disease phenotypes from population-scale single-cell data."
    supports: SUPPORT
    evidence_source: COMPUTATIONAL
    quote_role: PRIMARY_RESULT
    snippet: "we identified microglia subpopulations implicated in AD pathology"
    explanation: >-
      Phenotype scoring of about 6 million prefrontal nuclei from over 1,000
      donors ranked microglia, then astrocytes, as the cell subclasses most
      informative for Alzheimer versus control (full text), with
      Alzheimer-associated microglia enriched for known Alzheimer microglial
      pathways.
      Supports glial reactivity as a disease-associated cell state in human
      cortex at population scale; it does not establish direction of causation.
  - reference: PMID:42778763
    reference_title: "AI-based characterization of Alzheimer's disease phenotypes from population-scale single-cell data."
    supports: SUPPORT
    evidence_source: COMPUTATIONAL
    quote_role: PRIMARY_RESULT
    snippet: "We also identified many PACs for multiple phenotypes, including the astrocytes between AD and depression showing specific gene expression patterns such as inflammation and endoplasmic reticulum stress pathways."
    explanation: >-
      Comparing Alzheimer donors with and without depression, which removes the
      shared Alzheimer signal, identified astrocyte states carrying inflammatory
      and endoplasmic reticulum stress programs. This extends the glial
      inflammatory node toward the neuropsychiatric phenotype (depression within
      Alzheimer disease) as an association; the authors note that whether it
      shares mechanism with major depressive disorder is unresolved.
  downstream:
  - target: Amyloid Plaque Formation
    description: Inflammasome activation in microglia can reinforce amyloid-beta aggregation through extracellular ASC specks, making amyloid-to-glia ordering partly reciprocal rather than a one-way edge.
    hypothesis_groups:
    - neuroimmune_glial_amplification_model
    - amyloid_cascade_model
    causal_link_type: DIRECT
    evidence:
    - reference: PMID:29293211
      reference_title: "Microglia-derived ASC specks cross-seed amyloid-β in Alzheimer's disease."
      supports: SUPPORT
      evidence_source: MODEL_ORGANISM
      snippet: "ASC specks released by microglia bind rapidly to amyloid-β and increase the formation of amyloid-β oligomers and aggregates, acting as an inflammation-driven cross-seed for amyloid-β pathology."
      explanation: >-
        Supports a reciprocal causal edge in which microglial inflammasome
        products amplify amyloid-beta aggregation rather than merely responding
        to pre-existing amyloid plaques.
  - target: Neurofibrillary Tangle Formation
    description: Microglial state is modeled as a context-dependent modifier of amyloid-associated tau seeding and spreading; TREM2-dependent disease-associated microglia can restrain tau propagation, while a simple pathogenic NLRP3-to-tau edge remains contested.
    hypothesis_groups:
    - neuroimmune_glial_amplification_model
    - tau_neurodegeneration_model
    causal_link_type: INDIRECT_UNKNOWN_INTERMEDIATES
    intermediate_mechanisms:
    - TREM2-dependent plaque-associated disease-associated microglial activation.
    - Context-dependent inflammasome, cytokine, and complement signaling.
    evidence:
    - reference: PMID:34100905
      reference_title: "Activated microglia mitigate Aβ-associated tau seeding and spreading."
      supports: SUPPORT
      evidence_source: MODEL_ORGANISM
      snippet: "Together, these data suggest that TREM2-dependent activation of the DAM phenotype is essential in delaying Aβ-induced pathological tau propagation."
      explanation: >-
        Supports glial state as a modifier of amyloid-associated tau propagation,
        but in a protective direction for TREM2-dependent disease-associated
        microglia.
    - reference: PMID:39381137
      reference_title: "NLRP3 inflammasome activation and pyroptosis are dispensable for tau pathology."
      supports: REFUTE
      evidence_source: MODEL_ORGANISM
      snippet: "The absence of key components of the NLRP3 inflammasome pathway did not yield a beneficial effect on tau pathology or neurodegeneration in the preclinical Tau-P301S mouse model of AD."
      explanation: >-
        Refutes a simple model in which NLRP3 inflammasome activation is
        necessary for tau pathology in a tau-transgenic mouse context, supporting
        curation of the glia-to-tau edge as unresolved and context-dependent.
  - target: Oxidative Stress
    description: Activated glial cells are modeled as sources of inflammatory mediators and oxygen radicals that amplify oxidative neuronal injury.
    hypothesis_groups:
    - neuroimmune_glial_amplification_model
    causal_link_type: DIRECT
    evidence:
    - reference: PMID:21546088
      reference_title: "Complement in the brain."
      supports: SUPPORT
      snippet: "In age related diseases, such as Alzheimer's disease (AD), accumulating amyloid proteins elicit complement activation and a local, chronic inflammatory response that leads to attraction and activation of glial cells that, under such activation conditions, can produce neurotoxic substances, including pro-inflammatory cytokines and oxygen radicals."
      explanation: Supports activated glia as a source of cytokines and oxygen radicals in Alzheimer disease.
  - target: Vascular Dysfunction
    description: Neuroinflammation is modeled as a contributor to blood-brain barrier breakdown and neurovascular-unit dysfunction.
    hypothesis_groups:
    - neuroimmune_glial_amplification_model
    - vascular_bbb_clearance_model
    causal_link_type: INDIRECT_KNOWN_INTERMEDIATES
    intermediate_mechanisms:
    - Inflammatory remodeling of cerebral microvessels and blood-brain barrier injury.
    evidence:
    - reference: PMID:36293539
      reference_title: "Physiological and Pathological Remodeling of Cerebral Microvessels."
      supports: SUPPORT
      snippet: "Aberrant remodeling of microvesselsis associated with blood-brain barrier breakdown, development of neuroinflammation, inadequate microcirculation in active brain regions, and leads to the dysfunction of the neurovascular unit and progressive neurological deficits."
      explanation: Partially supports coupling among neuroinflammation, blood-brain barrier breakdown, and neurovascular-unit dysfunction.
  - target: Senescent Cell Accumulation
    description: >-
      Inflammasome/IL-1 signaling is modeled as reinforcing rather than merely
      resulting from glial senescence, closing a reciprocal loop with the
      existing senescence-to-neuroinflammation edge rather than a one-way
      cascade.
    hypothesis_groups:
    - asi_axis_reciprocal_cycle
    causal_link_type: INDIRECT_UNKNOWN_INTERMEDIATES
    evidence:
    - reference: PMID:23770676
      reference_title: "A complex secretory program orchestrated by the inflammasome controls paracrine senescence."
      supports: SUPPORT
      evidence_source: IN_VITRO
      directness: INDIRECT
      snippet: "Treatment with Caspase 1 or IL-1R inhibitors (Fig 6i, S6i) but not TGFBRI inhibitors (Fig S6j), blunted the expression of SASP components during OIS, suggesting that activation of the inflammasome was a cause rather than an effect of the SASP."
      explanation: >-
        General mechanistic support for inflammasome/IL-1 signaling reinforcing
        the senescent secretory phenotype, in an oncogene-induced-senescence
        fibroblast/epithelial model rather than CNS glia, hence indirect for
        this entry.
- name: Oxidative Stress
  description: Imbalance between the production of reactive oxygen species and the brain's ability to detoxify them, leading to cellular damage.
  biological_processes:
  - preferred_term: Free radical production
    modifier: INCREASED
    term:
      id: GO:1903409
      label: reactive oxygen species biosynthetic process
  - preferred_term: Antioxidant defense
    modifier: DECREASED
    term:
      id: GO:0006979
      label: response to oxidative stress
  cellular_components:
  - preferred_term: Mitochondria
    term:
      id: GO:0005739
      label: mitochondrion
  - preferred_term: Cell membranes
    term:
      id: GO:0005886
      label: plasma membrane
  chemical_entities:
  - preferred_term: Reactive oxygen species (ROS)
    term:
      id: CHEBI:26523
      label: reactive oxygen species
  - preferred_term: Reactive nitrogen species (RNS)
    term:
      id: CHEBI:62764
      label: reactive nitrogen species
  pathways:
  - preferred_term: Mitochondrial electron transport chain
    term:
      id: GO:0022904
      label: respiratory electron transport chain
  - preferred_term: NADPH oxidase activation
  consequences:
  - Amyloid Plaque Formation
  - Neurofibrillary Tangle Formation
  evidence:
  - reference: PMID:34416493
    reference_title: "Emerging roles of oxidative stress in brain aging and Alzheimer's disease."
    supports: SUPPORT
    snippet: Emerging evidence suggests that accumulated oxidative stress may be one of the key mechanisms causing cognitive aging and neurodegenerative diseases such as Alzheimer's disease (AD).
    explanation: The literature supports the role of oxidative stress, including the imbalance between ROS production and antioxidant defenses, in the development of Alzheimer's disease.
  - reference: PMID:27888001
    reference_title: "Antioxidants and HNE in redox homeostasis."
    supports: SUPPORT
    snippet: Depending on its level, HNE exerts harmful or protective effects associated with the induction of antioxidant defense mechanisms. These effects make HNE a key player in maintaining redox homeostasis, as well as producing imbalances in this system that participate in aging and the development of pathological conditions.
    explanation: This reference supports the involvement of oxidative stress and the imbalance in ROS/RNS in the pathology of neurodegenerative diseases.
  - reference: PMID:25772897
    reference_title: "Pathological mechanisms in progressive multiple sclerosis."
    supports: SUPPORT
    snippet: Key elements driving neurodegeneration include microglia activation, chronic oxidative injury, accumulation of mitochondrial damage in axons, and age-related iron accumulation in the human brain.
    explanation: This reference supports the involvement of oxidative stress and mitochondrial damage in neurodegeneration but does not specifically focus on Alzheimer's disease.
  - reference: PMID:33164705
    reference_title: "Neuroprotective effects of natural compounds on neurotoxin-induced oxidative stress and cell apoptosis."
    supports: SUPPORT
    snippet: Oxidative stress-induced neuronal apoptosis plays a pivotal role in pathogenesis of neurodegeneration.
    explanation: This reference supports the role of oxidative stress in neurodegeneration, which is relevant to Alzheimer's disease.
  downstream:
  - target: Synaptic Dysfunction
    description: Oxidative neuronal injury is modeled as one convergent stressor that worsens synaptic and network function.
    hypothesis_groups:
    - synaptic_failure_convergence_model
    causal_link_type: INDIRECT_UNKNOWN_INTERMEDIATES
    evidence:
    - reference: PMID:34416493
      reference_title: "Emerging roles of oxidative stress in brain aging and Alzheimer's disease."
      supports: SUPPORT
      snippet: "Emerging evidence suggests that accumulated oxidative stress may be one of the key mechanisms causing cognitive aging and neurodegenerative diseases such as Alzheimer's disease (AD)."
      explanation: Supports oxidative stress as a disease-relevant neurodegenerative mechanism, though the specific synaptic edge remains partly inferred.
- name: Mitochondrial Quality-Control Failure
  conforms_to: "mitochondrial_dysfunction#Impaired Mitophagy and Quality Control"
  description: >-
    Impaired mitophagy and accumulation of damaged mitochondria are modeled as a
    mitochondrial stress layer that can feed into amyloid-beta accumulation, tau
    hyperphosphorylation, neuroinflammation, and cognitive decline.
  mechanism_confidence: PROVISIONAL
  cell_types:
  - preferred_term: Neurons
    term:
      id: CL:0000540
      label: neuron
  - preferred_term: Microglia
    term:
      id: CL:0000129
      label: microglial cell
  biological_processes:
  - preferred_term: Mitophagy
    modifier: DECREASED
    term:
      id: GO:0000422
      label: autophagy of mitochondrion
  cellular_components:
  - preferred_term: Mitochondria
    term:
      id: GO:0005739
      label: mitochondrion
  locations:
  - preferred_term: Hippocampus
  - preferred_term: Cerebral Cortex
  mechanisms:
  - Failed clearance of damaged mitochondria through mitophagy.
  - Mitochondria-associated oxidative and inflammatory stress.
  consequences:
  - Amyloid Plaque Formation
  - Neurofibrillary Tangle Formation
  - Neuroinflammation
  - Memory Loss
  evidence:
  - reference: PMID:30742114
    reference_title: "Mitophagy inhibits amyloid-β and tau pathology and reverses cognitive deficits in models of Alzheimer's disease."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "mitophagy is impaired in the hippocampus of AD patients"
    explanation: >-
      Supports impaired mitophagy as a human Alzheimer disease mitochondrial
      quality-control abnormality.
  - reference: PMID:30742114
    reference_title: "Mitophagy inhibits amyloid-β and tau pathology and reverses cognitive deficits in models of Alzheimer's disease."
    supports: SUPPORT
    evidence_source: MODEL_ORGANISM
    snippet: "Mitophagy diminishes insoluble Aβ1-42 and Aβ1-40 and prevents cognitive impairment in an APP/PS1 mouse model through microglial phagocytosis of extracellular Aβ plaques and suppression of neuroinflammation."
    explanation: >-
      Supports the modeled link from mitochondrial quality-control restoration
      to reduced amyloid burden, neuroinflammation, and cognitive impairment in
      an Alzheimer disease mouse model.
  - reference: PMID:42778763
    reference_title: "AI-based characterization of Alzheimer's disease phenotypes from population-scale single-cell data."
    supports: SUPPORT
    evidence_source: COMPUTATIONAL
    quote_role: PRIMARY_RESULT
    snippet: "enhanced excitatory/inhibitory imbalance and mitochondrial dysfunction in cognitively impaired AD donors"
    explanation: >-
      Associates neuronal mitochondrial dysfunction with the cognitive
      expression of Alzheimer pathology in human cortex. In the full text the
      signal is upregulation of the mitochondrial transcription repressor MTERF1
      in excitatory neurons of cognitively impaired versus resilient donors at
      matched late Braak stage. This is a transcriptional association from
      computational phenotype scoring; it does not measure mitophagy.
  downstream:
  - target: Amyloid Plaque Formation
    description: Impaired mitochondrial quality control is modeled as a contributor to amyloid-beta accumulation through reduced mitophagy-linked microglial plaque clearance and increased neuroinflammatory stress.
    hypothesis_groups:
    - autophagy_lysosomal_clearance_model
    - amyloid_cascade_model
    - neuroimmune_glial_amplification_model
    causal_link_type: INDIRECT_KNOWN_INTERMEDIATES
    intermediate_mechanisms:
    - Reduced microglial phagocytosis of extracellular amyloid-beta plaques.
    - Suppression or amplification of neuroinflammation depending on mitophagy state.
    evidence:
    - reference: PMID:30742114
      reference_title: "Mitophagy inhibits amyloid-β and tau pathology and reverses cognitive deficits in models of Alzheimer's disease."
      supports: SUPPORT
      evidence_source: MODEL_ORGANISM
      snippet: "Mitophagy diminishes insoluble Aβ1-42 and Aβ1-40 and prevents cognitive impairment in an APP/PS1 mouse model through microglial phagocytosis of extracellular Aβ plaques and suppression of neuroinflammation."
      explanation: >-
        Supports the inverse edge: restoring mitophagy reduces amyloid-beta
        species and neuroinflammation, so impaired mitophagy is curated as a
        plausible contributor to amyloid accumulation.
  - target: Neurofibrillary Tangle Formation
    description: Impaired mitophagy is modeled as a contributor to tau hyperphosphorylation and tau pathology, but its position relative to amyloid and glial activation remains unresolved.
    hypothesis_groups:
    - autophagy_lysosomal_clearance_model
    - tau_neurodegeneration_model
    causal_link_type: INDIRECT_KNOWN_INTERMEDIATES
    intermediate_mechanisms:
    - Failure to remove defective mitochondria and downstream mitochondrial stress.
    evidence:
    - reference: PMID:30742114
      reference_title: "Mitophagy inhibits amyloid-β and tau pathology and reverses cognitive deficits in models of Alzheimer's disease."
      supports: SUPPORT
      evidence_source: IN_VITRO
      snippet: "Mitophagy enhancement abolishes AD-related tau hyperphosphorylation in human neuronal cells"
      explanation: >-
        Supports the inverse edge from enhanced mitophagy to reduced
        Alzheimer-related tau hyperphosphorylation in human neuronal cells.
    - reference: PMID:30742114
      reference_title: "Mitophagy inhibits amyloid-β and tau pathology and reverses cognitive deficits in models of Alzheimer's disease."
      supports: SUPPORT
      evidence_source: MODEL_ORGANISM
      snippet: "reverses memory impairment in transgenic tau nematodes and mice."
      explanation: >-
        Supports the model-organism functional arm of the same mitophagy-enhanced
        tau-model finding; this item is partial for the tau pathology edge
        because the snippet reports memory impairment rather than tangle burden.
  - target: Memory Loss
    description: Mitochondrial quality-control failure is modeled as one upstream stressor that can worsen cognitive outcomes, but cognitive resilience may decouple pathology burden from symptoms.
    hypothesis_groups:
    - autophagy_lysosomal_clearance_model
    - synaptic_failure_convergence_model
    causal_link_type: INDIRECT_KNOWN_INTERMEDIATES
    intermediate_mechanisms:
    - Amyloid, tau, neuroinflammatory, and synaptic consequences of impaired mitophagy.
    evidence:
    - reference: PMID:30742114
      reference_title: "Mitophagy inhibits amyloid-β and tau pathology and reverses cognitive deficits in models of Alzheimer's disease."
      supports: SUPPORT
      evidence_source: MODEL_ORGANISM
      snippet: "In both amyloid-β (Aβ) and tau Caenorhabditis elegans models of AD, mitophagy stimulation (through NAD+ supplementation, urolithin A, and actinonin) reverses memory impairment through PINK-1 (PTEN-induced kinase-1)-, PDR-1 (Parkinson's disease-related-1; parkin)-, or DCT-1 (DAF-16/FOXO-controlled germline-tumor affecting-1)-dependent pathways."
      explanation: >-
        Supports a causal link from mitophagy stimulation to improved memory in
        amyloid-beta and tau model systems, motivating impaired mitophagy as a
        potential upstream contributor to cognitive decline.
  - target: Neuroinflammation
    description: >-
      Impaired mitochondrial quality control is modeled as licensing NLRP3
      inflammasome activation through cytosolic mitochondrial DNA release, in
      addition to reduced mitophagy-mediated resolution of glial activation.
    hypothesis_groups:
    - autophagy_lysosomal_clearance_model
    - neuroimmune_glial_amplification_model
    - asi_axis_reciprocal_cycle
    causal_link_type: INDIRECT_KNOWN_INTERMEDIATES
    intermediate_mechanisms:
    - Cytosolic mitochondrial DNA release licensing NLRP3 inflammasome assembly.
    - Reduced mitophagy-dependent resolution of microglial activation.
    evidence:
    - reference: PMID:21151103
      reference_title: "Autophagy proteins regulate innate immune responses by inhibiting the release of mitochondrial DNA mediated by the NALP3 inflammasome."
      supports: SUPPORT
      evidence_source: IN_VITRO
      directness: INDIRECT
      snippet: "Depletion of autophagic proteins promoted the accumulation of dysfunctional mitochondria and cytosolic translocation of mitochondrial DNA (mtDNA) in response to lipopolysaccharide (LPS) and ATP in macrophages."
      explanation: >-
        General macrophage mechanism for autophagy/mitochondrial-quality-control
        failure licensing inflammasome activation; not microglia-specific or
        Alzheimer-specific, hence indirect.
    - reference: PMID:30742114
      reference_title: "Mitophagy inhibits amyloid-β and tau pathology and reverses cognitive deficits in models of Alzheimer's disease."
      supports: SUPPORT
      evidence_source: MODEL_ORGANISM
      snippet: "Mitophagy diminishes insoluble Aβ1-42 and Aβ1-40 and prevents cognitive impairment in an APP/PS1 mouse model through microglial phagocytosis of extracellular Aβ plaques and suppression of neuroinflammation."
      explanation: >-
        Supports the inverse edge in an Alzheimer mouse model: restoring
        mitophagy suppresses neuroinflammation, so impaired mitophagy is
        curated as a plausible contributor to it.
- name: Vascular Dysfunction
  description: Alterations in cerebral blood flow and blood-brain barrier integrity, contributing to neuronal dysfunction and amyloid accumulation.
  cell_types:
  - preferred_term: Endothelial cells
    term:
      id: CL:0000115
      label: endothelial cell
  - preferred_term: Pericytes
    term:
      id: CL:0000669
      label: pericyte
  biological_processes:
  - preferred_term: Blood-brain barrier regulation
    modifier: DYSREGULATED
    term:
      id: GO:1905603
      label: regulation of blood-brain barrier permeability
  - preferred_term: Cerebral blood flow
    modifier: DECREASED
    term:
      id: GO:0120275
      label: cerebral blood circulation
  - preferred_term: LRP1-mediated vascular amyloid-beta clearance
    modifier: DECREASED
    term:
      id: GO:0097242
      label: amyloid-beta clearance
  locations:
  - preferred_term: Cerebral blood vessels
  - preferred_term: Blood-brain barrier
  pathways:
  - preferred_term: Neurovascular coupling
  genes:
  - preferred_term: LRP1
    term:
      id: hgnc:6692
      label: LRP1
  consequences:
  - Amyloid Plaque Formation
  - Oxidative Stress
  evidence:
  - reference: PMID:28902142
    reference_title: "Blood-Brain Barrier Dysfunction and the Pathogenesis of Alzheimer's Disease."
    supports: SUPPORT
    snippet: Thus, current evidence suggests that BBB dysfunction may causatively and consequently contribute to AD pathogenesis, forming a vicious cycle between brain Abeta accumulation and neurovascular unit impairments during disease progression.
    explanation: This reference supports the statement by indicating that blood-brain barrier (BBB) dysfunction contributes to Alzheimer's disease (AD) pathogenesis, including amyloid-beta (Abeta) accumulation and neurovascular impairments.
  - reference: PMID:26898552
    reference_title: "The Utility of Cerebral Blood Flow as a Biomarker of Preclinical Alzheimer's Disease."
    supports: SUPPORT
    snippet: There is accumulating evidence suggesting that changes in brain perfusion are present long before the clinical symptoms of Alzheimer's disease (AD), perhaps even before amyloid-beta accumulation or brain atrophy.
    explanation: This reference supports the statement by highlighting the role of cerebral blood flow changes in the early stages of Alzheimer's disease, which is consistent with the described vascular dysfunction.
  - reference: PMID:36293539
    reference_title: "Physiological and Pathological Remodeling of Cerebral Microvessels."
    supports: SUPPORT
    snippet: Aberrant remodeling of microvesselsis associated with blood-brain barrier breakdown, development of neuroinflammation, inadequate microcirculation in active brain regions, and leads to the dysfunction of the neurovascular unit and progressive neurological deficits.
    explanation: This reference supports the statement by describing how microvascular remodeling and blood-brain barrier breakdown contribute to neurovascular unit dysfunction and neurological deficits, relevant to Alzheimer's disease.
  - reference: PMID:35289012
    reference_title: "Pathological changes within the cerebral vasculature in Alzheimer's disease: New perspectives."
    supports: SUPPORT
    snippet: We consider how abnormalities of the constituent cells of the neurovascular unit - particularly of endothelial cells and pericytes - and impairment of the blood-brain barrier (BBB) impact on the pathogenesis of AD.
    explanation: This reference supports the statement by discussing the role of endothelial cells and pericytes in blood-brain barrier impairment and Alzheimer's disease pathogenesis.
  - reference: PMID:19098903
    reference_title: SRF and myocardin regulate LRP-mediated amyloid-beta clearance in brain vascular cells.
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: Western blot analysis of small cerebral leptomeningeal arterial blood vessel homogenates from AD individuals with CAA showed elevated levels of SRF and MYOCD, Aβ accumulation and significantly lower levels of LRP, compared with age-matched, non-demented controls without CAA
    explanation: Post-mortem cerebral arterial vessels from AD patients with CAA have lower LRP1 and more vessel-wall amyloid-beta than controls, supporting decreased LRP1-mediated vascular amyloid-beta clearance on this node.
  downstream:
  - target: Amyloid Plaque Formation
    description: Blood-brain barrier dysfunction and neurovascular-unit impairment are modeled as contributors to brain amyloid-beta accumulation and impaired clearance.
    hypothesis_groups:
    - vascular_bbb_clearance_model
    - amyloid_cascade_model
    causal_link_type: INDIRECT_KNOWN_INTERMEDIATES
    intermediate_mechanisms:
    - Impaired amyloid-beta efflux and neurovascular clearance.
    evidence:
    - reference: PMID:28902142
      reference_title: "Blood-Brain Barrier Dysfunction and the Pathogenesis of Alzheimer's Disease."
      supports: SUPPORT
      snippet: "Thus, current evidence suggests that BBB dysfunction may causatively and consequently contribute to AD pathogenesis, forming a vicious cycle between brain Abeta accumulation and neurovascular unit impairments during disease progression."
      explanation: Supports a causal and consequential cycle between BBB dysfunction, amyloid-beta accumulation, and neurovascular-unit impairment.
  - target: Synaptic Dysfunction
    description: Reduced perfusion and neurovascular-unit injury are modeled as contributors to neuronal and synaptic dysfunction.
    hypothesis_groups:
    - vascular_bbb_clearance_model
    - synaptic_failure_convergence_model
    causal_link_type: INDIRECT_KNOWN_INTERMEDIATES
    intermediate_mechanisms:
    - Inadequate microcirculation in active brain regions.
    evidence:
    - reference: PMID:36293539
      reference_title: "Physiological and Pathological Remodeling of Cerebral Microvessels."
      supports: SUPPORT
      snippet: "Aberrant remodeling of microvesselsis associated with blood-brain barrier breakdown, development of neuroinflammation, inadequate microcirculation in active brain regions, and leads to the dysfunction of the neurovascular unit and progressive neurological deficits."
      explanation: Supports neurovascular-unit dysfunction and inadequate microcirculation as contributors to neurological deficits.
- name: Glymphatic Clearance Failure
  description: >-
    Failure of the sleep-dependent perivascular (glymphatic) route by which
    cerebrospinal fluid exchanges with interstitial fluid and carries
    interstitial amyloid-beta out of the brain. This is a clearance arm distinct
    from the transcytotic blood-brain-barrier / LRP1 efflux route modeled in
    Vascular Dysfunction: it operates along perivascular spaces and depends on
    perivascular polarization of astroglial aquaporin-4, which is lost in aged
    and Alzheimer brains independently of age. Reduced clearance raises the
    steady-state interstitial amyloid-beta concentration and so feeds amyloid
    plaque formation rather than replacing it.
  role: amplifier
  biological_scale: TISSUE
  conforms_to: "glymphatic_dysfunction#Impaired Perivascular CSF-ISF Exchange and Solute Clearance"
  mechanism_confidence: PROVISIONAL
  cell_types:
  - preferred_term: Astrocyte
    term:
      id: CL:0000127
      label: astrocyte
  cellular_components:
  - preferred_term: Astrocyte end-foot
    term:
      id: GO:0097450
      label: astrocyte end-foot
  biological_processes:
  - preferred_term: Amyloid-beta clearance
    modifier: DECREASED
    term:
      id: GO:0097242
      label: amyloid-beta clearance
  - preferred_term: Cerebrospinal fluid circulation
    modifier: DECREASED
    term:
      id: GO:0090660
      label: cerebrospinal fluid circulation
  locations:
  - preferred_term: Perivascular space
    term:
      id: UBERON:0014930
      label: perivascular space
  evidence:
  - reference: PMID:27893874
    reference_title: "Association of Perivascular Localization of Aquaporin-4 With Cognition and Alzheimer Disease in Aging Brains."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: >-
      Perivascular AQP4 localization was significantly associated with AD status
      independent of age
    explanation: >-
      Human postmortem evidence that the molecular lesion underlying glymphatic
      clearance failure is present in Alzheimer disease independently of age.
  - reference: PMID:29997300
    reference_title: "Brain-wide glymphatic enhancement and clearance in humans assessed with MRI."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: >-
      Clearance of the tracer substance was delayed in the dementia cohort.
    explanation: >-
      In vivo human intrathecal-tracer evidence of delayed brain clearance in a
      dementia cohort relative to near-healthy references.
  - reference: PMID:33004510
    reference_title: "Glymphatic failure as a final common pathway to dementia."
    supports: SUPPORT
    evidence_source: OTHER
    snippet: >-
      glymphatic failure may constitute a therapeutically targetable final
      common pathway
    explanation: >-
      States the convergence claim that motivates modeling glymphatic clearance
      failure as a shared upstream contributor in the dementias.
  - reference: PMID:30046111
    reference_title: "Functional aspects of meningeal lymphatics in ageing and Alzheimer's disease."
    supports: SUPPORT
    evidence_source: MODEL_ORGANISM
    snippet: "Impairment of meningeal lymphatic function slows paravascular influx of macromolecules into the brain and efflux of macromolecules from the interstitial fluid, and induces cognitive impairment in mice."
    explanation: >-
      Identifies the downstream drainage leg of this route: perivascular exchange
      empties into meningeal lymphatics, and impairing them slows both influx and
      efflux and impairs cognition.
  - reference: PMID:30046111
    reference_title: "Functional aspects of meningeal lymphatics in ageing and Alzheimer's disease."
    supports: SUPPORT
    evidence_source: MODEL_ORGANISM
    snippet: "Disruption of meningeal lymphatic vessels in transgenic mouse models of Alzheimer's disease promotes amyloid-β deposition in the meninges, which resembles human meningeal pathology, and aggravates parenchymal amyloid-β accumulation."
    explanation: >-
      Connects failure of the drainage leg to increased parenchymal amyloid-beta,
      the same endpoint this node's perivascular arm feeds. Entirely in mice — the
      quantitative contribution of meningeal lymphatic drainage to human
      amyloid-beta clearance is not established.
  downstream:
  - target: Amyloid Plaque Formation
    description: >-
      Reduced perivascular clearance raises the steady-state interstitial
      amyloid-beta concentration, favoring aggregation and plaque deposition.
    hypothesis_groups:
    - glymphatic_clearance_model
    - amyloid_cascade_model
    causal_link_type: INDIRECT_KNOWN_INTERMEDIATES
    intermediate_mechanisms:
    - Reduced CSF-interstitial fluid exchange raising interstitial amyloid-beta concentration.
    evidence:
    - reference: PMID:28899014
      reference_title: "Slow wave sleep disruption increases cerebrospinal fluid amyloid-beta levels."
      supports: SUPPORT
      evidence_source: HUMAN_CLINICAL
      snippet: >-
        Specific disruption of slow wave activity correlated with an increase in
        amyloid-β40
      explanation: >-
        Interventional human evidence that removing the sleep drive for
        glymphatic clearance raises amyloid-beta levels.
    - reference: PMID:29632177
      reference_title: "beta-Amyloid accumulation in the human brain after one night of sleep deprivation."
      supports: SUPPORT
      evidence_source: HUMAN_CLINICAL
      snippet: >-
        one night of sleep deprivation, relative to baseline, resulted in a
        significant increase in Aβ burden in the right hippocampus and thalamus
      explanation: >-
        Human PET evidence that a single night of sleep deprivation increases
        measured brain amyloid burden.
- name: Autophagy-Lysosomal Dysfunction
  conforms_to: "disabled_macroautophagy#Failure of Cytoplasmic Quality Control"
  description: Impaired autophagy and lysosomal degradation pathways leading to accumulation of protein aggregates and cellular dysfunction.
  cell_types:
  - preferred_term: Neurons
    term:
      id: CL:0000540
      label: neuron
  biological_processes:
  - preferred_term: Autophagy
    modifier: DECREASED
    term:
      id: GO:0006914
      label: autophagy
  - preferred_term: Lysosomal degradation
    modifier: DECREASED
    term:
      id: GO:0007041
      label: lysosomal transport
  - preferred_term: Protein catabolic process
    modifier: DECREASED
    term:
      id: GO:0030163
      label: protein catabolic process
  locations:
  - preferred_term: Hippocampus
  - preferred_term: Cerebral Cortex
  mechanisms:
  - Impaired autophagic flux and autophagosome-lysosome fusion
  - Defective lysosomal acidification
  - Accumulation of autophagic vacuoles containing Aβ
  consequences:
  - Amyloid Plaque Formation
  - Neurofibrillary Tangle Formation
  evidence:
  - reference: PMID:20541250
    reference_title: "Lysosomal proteolysis and autophagy require presenilin 1 and are disrupted by Alzheimer-related PS1 mutations."
    supports: SUPPORT
    evidence_source: MODEL_ORGANISM
    snippet: "In PS1 null blastocysts, neurons from mice hypomorphic for PS1 or conditionally depleted of PS1, substrate proteolysis and autophagosome clearance during macroautophagy are prevented as a result of a selective impairment of autolysosome acidification and cathepsin activation. These deficits are caused by failed PS1-dependent targeting of the v-ATPase V0a1 subunit to lysosomes."
    explanation: >-
      Supports this node's core mechanism by showing that PS1 loss prevents
      autophagosome clearance through impaired autolysosome acidification and
      cathepsin activation, with failed v-ATPase V0a1 lysosomal targeting as a
      causal mechanism.
  - reference: PMID:42429504
    reference_title: "Mechanisms of autophagy-lysosome pathway impairment in Alzheimer's disease."
    supports: SUPPORT
    evidence_source: OTHER
    snippet: "Dysfunction of the autophagy-lysosomal pathway is believed to be the main factor leading to the accumulation of amyloid-β and tau proteins, which are also pathological features of AD."
    explanation: >-
      Integrative review synthesizing evidence for autophagy-lysosomal
      dysfunction as a primary driver of amyloid-β and tau accumulation in
      Alzheimer disease, matching the quoted snippet. (The same review also
      separately summarizes a potential role for oligodendrocytes, but that is
      not covered by this snippet.)
  notes: Beclin-1/VPS34 complex deficits, p62 accumulation, and faulty autolysosome acidification contribute to impaired clearance of Aβ and tau aggregates. A recent integrative review (PMID:42429504) also flags a potential contributing role for oligodendrocytes alongside neuron-intrinsic mechanisms.
  downstream:
  - target: Amyloid Plaque Formation
    description: Impaired autophagy and lysosomal degradation are modeled as contributors to amyloid-beta accumulation through reduced aggregate clearance.
    hypothesis_groups:
    - autophagy_lysosomal_clearance_model
    - amyloid_cascade_model
    causal_link_type: INDIRECT_KNOWN_INTERMEDIATES
    intermediate_mechanisms:
    - Reduced amyloid-beta clearance through defective autophagic flux and lysosomal degradation.
  - target: Neurofibrillary Tangle Formation
    description: Impaired autophagy and lysosomal degradation are modeled as contributors to tau aggregate accumulation through reduced proteostatic clearance.
    hypothesis_groups:
    - autophagy_lysosomal_clearance_model
    - tau_neurodegeneration_model
    causal_link_type: INDIRECT_KNOWN_INTERMEDIATES
    intermediate_mechanisms:
    - Reduced tau aggregate clearance through defective autophagic flux and lysosomal degradation.
- name: HSV-1 Reactivation in RORB+ Glutamatergic Neurons
  description: >-
    Latent herpes simplex virus 1 (HSV-1) reactivates specifically within RORB+
    glutamatergic neurons — a cortical neuronal population selectively
    vulnerable in Alzheimer's disease. Petabase-scale mining of human
    sequencing data and single-nucleus analyses detect viral transcripts in
    this population in HEALTHY (but not pathological) post-mortem brain
    tissue, consistent with a model in which viral reactivation precedes the
    progressive loss of this neuronal population during dementia and so is no
    longer detectable in end-stage disease.
  mechanism_confidence: PROVISIONAL
  cell_types:
  - preferred_term: RORB+ glutamatergic neuron
    term:
      id: CL:0000679
      label: glutamatergic neuron
  biological_processes:
  - preferred_term: HSV-1 release from latency
    modifier: INCREASED
    term:
      id: GO:0019046
      label: release from viral latency
  - preferred_term: HSV-1 genome replication
    modifier: INCREASED
    term:
      id: GO:0019079
      label: viral genome replication
  locations:
  - preferred_term: Cerebral Cortex
  mechanisms:
  - Reactivation of latent HSV-1 in long-lived cortical neurons
  - Selective viral expression in RORB+ glutamatergic neuron subpopulations
  consequences:
  - Synaptic Dysfunction
  evidence:
  - reference: PMID:42094473
    reference_title: "Resolving human neuronal herpesvirus reactivation via petabase-scale association studies."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "Integrative single-nucleus analyses resolve direct evidence of HSV-1 expression in RORB+ glutamatergic neurons, implicating viral reactivation in a neuronal population progressively lost during dementia."
    explanation: >-
      The preprint mines >10 petabytes of human sequencing data and uses
      single-nucleus analyses on post-mortem human brain to localize HSV-1
      transcripts specifically to RORB+ glutamatergic neurons, the same
      cortical population progressively lost in dementia — directly
      supporting this provisional pathophysiology node.
  - reference: PMID:42094473
    reference_title: "Resolving human neuronal herpesvirus reactivation via petabase-scale association studies."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "identifying recurrent herpes simplex virus 1 (HSV-1) reactivation in healthy but not pathological post-mortem human brain tissue"
    explanation: >-
      Key temporal observation: HSV-1 reactivation is detected in healthy
      post-mortem brain tissue but is absent from pathological tissue,
      consistent with a model in which reactivation precedes (rather than
      follows) the loss of vulnerable neurons.
  downstream:
  - target: Synaptic Dysfunction
    description: HSV-1 reactivation in selectively vulnerable cortical neurons is modeled as a possible upstream contributor to loss of synaptic integrity and cognitive-network failure.
    hypothesis_groups:
    - hsv1_reactivation_model
    - synaptic_failure_convergence_model
    causal_link_type: INDIRECT_UNKNOWN_INTERMEDIATES
    evidence:
    - reference: PMID:42094473
      reference_title: "Resolving human neuronal herpesvirus reactivation via petabase-scale association studies."
      supports: SUPPORT
      evidence_source: HUMAN_CLINICAL
      snippet: "Integrative single-nucleus analyses resolve direct evidence of HSV-1 expression in RORB+ glutamatergic neurons, implicating viral reactivation in a neuronal population progressively lost during dementia."
      explanation: Supports HSV-1 reactivation in a vulnerable neuronal population, while the synaptic dysfunction edge remains an inferred downstream consequence.
  notes: >-
    Provisional / emerging hypothesis. The herpesvirus-reactivation model of
    Alzheimer's pathogenesis is not yet established as canonical, but the
    petabase-scale single-nucleus evidence ties viral reactivation to a
    specific, selectively vulnerable cortical neuronal population (RORB+
    glutamatergic neurons). The signal is present in healthy but not
    pathological brain — interpreted by the authors as reactivation preceding
    the progressive loss of these neurons. Causality between HSV-1
    expression and neuronal loss remains correlative in this work. The
    downstream "Synaptic Dysfunction" consequence is inferred from
    neuronal-population loss rather than directly demonstrated in the cited
    paper. Upstream drivers (e.g., aging-related immune decline, latent HSV-1
    burden) are not yet represented as separate nodes in this entry.
- name: Intercellular Tau Transmission via Extracellular Vesicles
  description: >-
    Pathological tau spreads cell to cell in Alzheimer's disease, in part by being
    packaged into neuronal extracellular vesicles (EVs) and released for uptake by
    recipient neurons (and microglia), where the delivered tau seeds aggregation of
    endogenous tau. The activity-regulated, capsid-forming neuronal protein Arc
    binds tau directly (with higher affinity for phosphorylated tau) and, together
    with the I-BAR protein IRSp53, drives release of seed-competent tau from
    dendrites in EVs. Arc and tau are co-packaged in mouse and human brain-derived
    EVs, and in human AD brain EVs Arc levels correlate with phosphorylated EV-tau.
    Loss of Arc reduces EV-tau and tau seeding potential and nearly abolishes
    neuron-to-neuron tau transmission, while causing intracellular tau to
    accumulate in donor neurons — consistent with EV-tau release being partly
    protective for the donor but a driver of pathology spread.
  mechanism_confidence: PROVISIONAL
  cell_types:
  - preferred_term: Neurons
    term:
      id: CL:0000540
      label: neuron
  - preferred_term: Microglia
    term:
      id: CL:0000129
      label: microglial cell
  biological_processes:
  - preferred_term: Extracellular vesicle biogenesis
    modifier: INCREASED
    term:
      id: GO:0140112
      label: extracellular vesicle biogenesis
  - preferred_term: Tau release in extracellular vesicles
    modifier: INCREASED
    term:
      id: GO:0009306
      label: protein secretion
  - preferred_term: Tau seeding of aggregation in recipient neurons
    modifier: INCREASED
    term:
      id: GO:1990000
      label: amyloid fibril formation
  locations:
  - preferred_term: Neuronal dendrites
  - preferred_term: Synapses
  - preferred_term: Hippocampus
  - preferred_term: Entorhinal Cortex
  chemical_entities:
  - preferred_term: Phosphorylated tau
  genes:
  - preferred_term: ARC
    term:
      id: hgnc:648
      label: ARC
  - preferred_term: MAPT
    term:
      id: hgnc:6893
      label: MAPT
  - preferred_term: BAIAP2 (IRSp53)
    term:
      id: hgnc:947
      label: BAIAP2
  mechanisms:
  - Direct Arc-tau protein-protein interaction packaging tau into EVs
  - IRSp53/I-BAR-dependent EV biogenesis at dendritic membranes
  - Release of seed-competent EV-tau and uptake by recipient neurons
  - Seeding of endogenous tau aggregation in recipient neurons
  consequences:
  - Neurofibrillary Tangle Formation
  evidence:
  - reference: PMID:42372723
    reference_title: "Arc mediates intercellular tau transmission via extracellular vesicles."
    supports: SUPPORT
    evidence_source: IN_VITRO
    snippet: "We find that the neuronal gene Arc is critical for the release of tau in neuronal extracellular vesicles (EVs) via a direct protein-protein interaction."
    explanation: >-
      Establishes Arc as a required mediator of tau release in neuronal EVs
      acting through a direct Arc-tau protein-protein interaction, demonstrated
      in neuronal cultures and purified-protein binding assays.
  - reference: PMID:42372723
    reference_title: "Arc mediates intercellular tau transmission via extracellular vesicles."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "Both Arc and tau are co-packaged in mouse and human brain-derived EVs."
    explanation: >-
      Arc-tau co-packaging is observed in human (and mouse) brain-derived EVs,
      supporting the EV route of tau spread in human tissue.
  - reference: PMID:42372723
    reference_title: "Arc mediates intercellular tau transmission via extracellular vesicles."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "Moreover, Arc levels in brain-derived EVs isolated from human Alzheimer's disease (AD) brains show a strong positive correlation with phosphorylated EV-tau levels."
    explanation: >-
      In human AD brain EVs, Arc abundance tracks phosphorylated EV-tau,
      connecting Arc-dependent EV packaging to pathological tau in human disease.
  - reference: PMID:42372723
    reference_title: "Arc mediates intercellular tau transmission via extracellular vesicles."
    supports: SUPPORT
    evidence_source: MODEL_ORGANISM
    snippet: "rTgArc KO mice have increased accumulation of intracellular tau and a modest increase in cell toxicity early in disease progression."
    explanation: >-
      Without Arc, tau is retained intracellularly and donor-neuron toxicity
      rises early, consistent with EV-tau release being partly protective for the
      donor neuron.
  downstream:
  - target: Neurofibrillary Tangle Formation
    description: EV-delivered seed-competent tau is taken up by recipient neurons and seeds aggregation of endogenous tau, propagating tangle pathology along connected circuits.
    hypothesis_groups:
    - ev_mediated_tau_propagation_model
    - tau_neurodegeneration_model
    causal_link_type: INDIRECT_KNOWN_INTERMEDIATES
    intermediate_mechanisms:
    - EV-tau uptake by recipient neurons and templated seeding of endogenous tau.
    evidence:
    - reference: PMID:42372723
      reference_title: "Arc mediates intercellular tau transmission via extracellular vesicles."
      supports: SUPPORT
      evidence_source: MODEL_ORGANISM
      snippet: "These results show that Arc is critical for the packaging of tau in EVs, which plays a significant role in intercellular tau transmission."
      explanation: >-
        Arc-dependent EV packaging of tau is required for intercellular tau
        transmission, the basis for the seeding edge back onto tangle pathology
        in recipient neurons.
  notes: >-
    PROVISIONAL / EMERGING (ev_mediated_tau_propagation_model). Evidence is from
    primary neurons, rTg4510 tau-transgenic mice crossed to Arc-KO, and human
    postmortem AD brain EVs (Tyagi et al., Cell 2026; PMID:42372723). EV-tau is
    one of several proposed routes of tau spread (free/naked tau uptake via LRP1,
    tunneling nanotubes, trans-synaptic transfer), so this node models a partial
    contributor rather than the sole transmission mechanism. Arc derives from a
    domesticated Ty3/gypsy retrotransposon Gag and forms virus-like capsids,
    placing tau spread within an emerging "endogenous-retroviral/capsid" framing
    of neurodegeneration. Notably, Arc loss blocks transmission without overtly
    accelerating late-stage (8-month) tau pathology in this transgenic model,
    where extreme tau overexpression may mask EV-dependent spread.
- name: PARP1-Mediated Parthanatos
  description: >-
    PARP1 (poly(ADP-ribose) polymerase 1) plays a dual role in Alzheimer's disease:
    physiological nuclear PARP1 activity is essential for DNA repair and memory
    consolidation, but overactivity triggered by oxidative stress and amyloid-beta-induced
    DNA damage leads to excessive NAD+ and ATP consumption. PARP1 overactivation causes
    NAD+ depletion, triggering AIF (apoptosis-inducing factor) release and formation of
    AIF-MIF complexes that drive parthanatos, a caspase-independent programmed cell-death
    mechanism. This PARP1-driven pathway converges with neuroinflammation (NF-κB pathway
    activation), mitophagy dysregulation, and disruption of SIRT1-mediated neuroprotection,
    collectively contributing to neuronal death and disease progression.
  mechanism_confidence: PROVISIONAL
  cell_types:
  - preferred_term: Neurons
    term:
      id: CL:0000540
      label: neuron
  - preferred_term: Microglia
    term:
      id: CL:0000129
      label: microglial cell
  biological_processes:
  - preferred_term: DNA damage response
    modifier: ABNORMAL
    term:
      id: GO:0006974
      label: DNA damage response
  - preferred_term: Neuroinflammation
    modifier: INCREASED
    term:
      id: GO:0150076
      label: neuroinflammatory response
  chemical_entities:
  - preferred_term: NAD+ (nicotinamide adenine dinucleotide)
    term:
      id: CHEBI:15846
      label: NAD(+)
  - preferred_term: ATP (adenosine triphosphate)
    term:
      id: CHEBI:15422
      label: ATP
  - preferred_term: AIF (apoptosis-inducing factor)
  genes:
  - preferred_term: PARP1
    term:
      id: hgnc:270
      label: PARP1
  - preferred_term: SIRT1
    term:
      id: hgnc:14929
      label: SIRT1
  mechanisms:
  - Aβ-induced oxidative stress and DNA damage trigger PARP1 hyperactivation
  - PARP1 overactivity depletes cellular NAD+ and ATP pools
  - NAD+ depletion triggers AIF-MIF complex formation and parthanatos
  - Parthanatos activation amplifies neuroinflammatory cascades via NF-κB
  - PARP1 hyperactivity dysregulates mitophagy and impairs SIRT1-dependent neuroprotection
  consequences:
  - Neuroinflammation
  - Oxidative Stress
  - Mitochondrial Quality-Control Failure
  evidence:
  - reference: PMID:42413719
    reference_title: "Targeting PARP1-dependent parthanatos in Alzheimer's disease: Mechanisms and therapeutic opportunities."
    supports: SUPPORT
    evidence_source: OTHER
    snippet: "PARP1 exhibits context-dependent duality: its physiological nuclear expression in hippocampus neurons is essential for memory consolidation and decreases early in cognitive impairment, suggesting a correlative association with synaptic malfunction. In contrast, overactivity of PARP1 resulting from Aβ-induced oxidative stress and DNA damage induces neurodegeneration via multiple pathways, including NAD+/ATP exhaustion leading to metabolism collapse, creation of the AIF-MIF complex promoting parthanatos, NF-κB-induced neuroinflammation, dysregulation of mitophagy, and disruption of the neuroprotective SIRT1 signaling pathway."
    explanation: >-
      Directly supports PARP1-mediated parthanatos as a convergence mechanism in
      Alzheimer's disease, linking Aβ-induced oxidative stress to NAD+ depletion,
      AIF-driven programmed necrosis, neuroinflammation amplification, and loss of
      mitochondrial neuroprotection.
  downstream:
  - target: Neuroinflammation
    description: PARP1 overactivation amplifies NF-κB-dependent neuroinflammatory responses and promotes microglial activation.
    hypothesis_groups:
    - neuroimmune_glial_amplification_model
    causal_link_type: INDIRECT_KNOWN_INTERMEDIATES
    intermediate_mechanisms:
    - PARP1-driven NAD+ depletion and parthanatos trigger NF-κB signaling.
    - AIF-MIF complexes activate inflammasome and inflammatory cytokine cascades.
    evidence:
    - reference: PMID:42413719
      reference_title: "Targeting PARP1-dependent parthanatos in Alzheimer's disease: Mechanisms and therapeutic opportunities."
      supports: SUPPORT
      evidence_source: OTHER
      snippet: "NF-κB-induced neuroinflammation, dysregulation of mitophagy, and disruption of the neuroprotective SIRT1 signaling pathway"
      explanation: >-
        PARP1 overactivity is modeled as driving NF-κB-mediated neuroinflammation
        through NAD+ depletion and parthanatos-driven inflammatory amplification.
  - target: Oxidative Stress
    description: PARP1 hyperactivation and NAD+ depletion impair mitochondrial redox homeostasis, sustaining oxidative stress through disruption of antioxidant SIRT1 and mitophagy.
    hypothesis_groups:
    - autophagy_lysosomal_clearance_model
    causal_link_type: INDIRECT_KNOWN_INTERMEDIATES
    intermediate_mechanisms:
    - NAD+ depletion impairs SIRT1-dependent mitochondrial quality control.
    - Dysregulated mitophagy fails to clear damaged mitochondria.
    evidence:
    - reference: PMID:42413719
      reference_title: "Targeting PARP1-dependent parthanatos in Alzheimer's disease: Mechanisms and therapeutic opportunities."
      supports: SUPPORT
      evidence_source: OTHER
      snippet: "dysregulation of mitophagy, and disruption of the neuroprotective SIRT1 signaling pathway"
      explanation: >-
        Links PARP1-driven NAD+ depletion to SIRT1 loss and mitophagy dysregulation,
        sustaining oxidative stress and mitochondrial dysfunction.
  - target: Mitochondrial Quality-Control Failure
    description: PARP1 hyperactivity and NAD+ depletion impair mitochondrial dynamics, mitophagy, and ATP production, contributing to quality-control collapse.
    hypothesis_groups:
    - autophagy_lysosomal_clearance_model
    causal_link_type: INDIRECT_KNOWN_INTERMEDIATES
    intermediate_mechanisms:
    - NAD+ depletion reduces mitochondrial ATP production.
    - SIRT1 dysregulation impairs mitochondrial biogenesis and turnover.
    evidence:
    - reference: PMID:42413719
      reference_title: "Targeting PARP1-dependent parthanatos in Alzheimer's disease: Mechanisms and therapeutic opportunities."
      supports: SUPPORT
      evidence_source: OTHER
      snippet: "precursors of NAD+ such as nicotinamide and NMN attenuate Aβ deposition, normalize metabolism, and ameliorate cognitive decline"
      explanation: >-
        Supports PARP1-driven NAD+ depletion and mitophagy dysregulation as
        contributors to mitochondrial failure and therapeutic intervention targets.
  notes: >-
    PROVISIONAL / EMERGING. The paper (PMID:42413719) is a 2026 review clarifying
    PARP1's context-dependent duality and positions PARP1-mediated parthanatos as a
    central convergence point for oxidative stress, DNA damage, NAD+ metabolism, and
    neuroinflammation. Preclinical evidence supports PARP1 inhibitors and NAD+
    supplementation (nicotinamide riboside, NMN) as therapeutic avenues. This node
    integrates with the existing autophagy_lysosomal_clearance_model and
    neuroimmune_glial_amplification_model by defining upstream PARP1-NAD+ metabolic
    triggers and downstream parthanatos-driven inflammatory amplification.
- name: Adaptive Immune T Cell Response to Tau Pathology
  description: >-
    Downstream of tau pathology and microglial activation, clonally expanded
    T cells — predominantly CD8+ cytotoxic T cells — infiltrate the brain in a
    spatiotemporal pattern that tracks tau (not amyloid) pathology and correlates
    with neuronal loss. Microglia are required to recruit these T cells; once in
    the parenchyma the cells shift from activated toward exhausted states with
    restricted TCR clonality, and IFN-gamma/PD-1 signaling amplifies
    neurodegeneration. Depletion of T cells (or of microglia) is neuroprotective
    in tauopathy models, and IFN-gamma or PD-1 blockade reduces brain atrophy,
    identifying the adaptive immune response as an active, druggable contributor
    to tau-driven neurodegeneration rather than a bystander.
  mechanism_confidence: PROVISIONAL
  cell_types:
  - preferred_term: CD8+ cytotoxic T cell
    term:
      id: CL:0000625
      label: CD8-positive, alpha-beta T cell
  - preferred_term: Microglia
    term:
      id: CL:0000129
      label: microglial cell
  biological_processes:
  - preferred_term: T cell chemotaxis into the brain
    modifier: INCREASED
    term:
      id: GO:0010818
      label: T cell chemotaxis
  - preferred_term: T cell mediated cytotoxicity
    modifier: INCREASED
    term:
      id: GO:0001913
      label: T cell mediated cytotoxicity
  - preferred_term: Interferon-gamma (type II interferon) production
    modifier: INCREASED
    term:
      id: GO:0032609
      label: type II interferon production
  chemical_entities:
  - preferred_term: Interferon-gamma
  locations:
  - preferred_term: Hippocampus
  - preferred_term: Entorhinal Cortex
  mechanisms:
  - cDC1 cross-presentation of brain-derived antigen priming CD8+ T cells in secondary lymphoid tissue
  - Microglia-dependent recruitment of primed T cells to sites of tau pathology
  - Clonal expansion of predominantly CD8+ cytotoxic T cells
  - Granzyme K release acting on neuronal protease-activated receptor-1 (PAR-1) to drive tau hyperphosphorylation
  - IFN-gamma and PD-1 (PDCD1) signaling amplifying neuronal loss
  consequences:
  - Synaptic Dysfunction
  evidence:
  - reference: PMID:36890231
    reference_title: "Microglia-mediated T cell infiltration drives neurodegeneration in tauopathy."
    supports: SUPPORT
    evidence_source: MODEL_ORGANISM
    snippet: "We found that mice with tauopathy but not those with amyloid deposition developed a unique innate and adaptive immune response and that depletion of microglia or T cells blocked tau-mediated neurodegeneration."
    explanation: >-
      Establishes a tauopathy-specific adaptive immune response and shows that
      depleting T cells (or microglia) blocks tau-mediated neurodegeneration.
  - reference: PMID:36890231
    reference_title: "Microglia-mediated T cell infiltration drives neurodegeneration in tauopathy."
    supports: SUPPORT
    evidence_source: MODEL_ORGANISM
    snippet: "Numbers of T cells, especially those of cytotoxic T cells, were markedly increased in areas with tau pathology in mice with tauopathy"
    explanation: >-
      Cytotoxic (CD8+) T cells accumulate specifically where tau pathology is
      present in mouse tauopathy, anchoring the cell type of this node (human AD
      brain accumulation is captured separately by the human PMID:31915375 item).
  - reference: PMID:36890231
    reference_title: "Microglia-mediated T cell infiltration drives neurodegeneration in tauopathy."
    supports: SUPPORT
    evidence_source: MODEL_ORGANISM
    snippet: "T cell numbers correlated with the extent of neuronal loss, and the cells dynamically transformed their cellular characteristics from activated to exhausted states along with unique TCR clonal expansion."
    explanation: >-
      Links T cell burden to neuronal loss and documents the activated-to-exhausted
      transition with clonal TCR expansion characteristic of an antigen-specific
      response.
  - reference: PMID:41890046
    reference_title: "CD8+ T cells are primed by cDC1 and exacerbate tau-mediated neurodegeneration."
    supports: SUPPORT
    evidence_source: MODEL_ORGANISM
    snippet: "This priming is required for a large increase of activated CD8+ T cells in the brain which promotes tau-mediated neurodegeneration."
    explanation: >-
      Resolves the antigen-presenting route: cDC1 cross-presentation primes the
      CD8+ T cells whose brain infiltration promotes tau-mediated neurodegeneration.
  - reference: PMID:40993111
    reference_title: "CD103(-)CD8(+) T cells promote neurotoxic inflammation in Alzheimer's disease via granzyme K-PAR-1 signaling."
    supports: SUPPORT
    evidence_source: MODEL_ORGANISM
    snippet: "Ablation of brain CD103-CD8+ T cells in 3xTg-AD mice ameliorates cognitive decline and reduces neuropathology."
    explanation: >-
      Independent lab: ablating the CD103-negative, granzyme-K-producing CD8+ T
      cell subset improves cognition and reduces neuropathology, corroborating the
      cytotoxic-CD8 effector role beyond the founding study.
  - reference: PMID:40993111
    reference_title: "CD103(-)CD8(+) T cells promote neurotoxic inflammation in Alzheimer's disease via granzyme K-PAR-1 signaling."
    supports: SUPPORT
    evidence_source: IN_VITRO
    snippet: "GrK induces neuronal dysfunction and tau hyperphosphorylation in human and mouse cells via protease-activated receptor-1 (PAR-1)"
    explanation: >-
      Provides a molecular effector mechanism: CD8-derived granzyme K acts on
      neuronal PAR-1 to induce tau hyperphosphorylation, linking the T cell arm to
      tau pathology in both human and mouse cells.
  - reference: PMID:31915375
    reference_title: "Clonally expanded CD8 T cells patrol the cerebrospinal fluid in Alzheimer's disease."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "we discovered clonally expanded CD8+ TEMRA cells in the cerebrospinal fluid of patients with Alzheimer's disease."
    explanation: >-
      Human correlative anchor for clonally expanded CD8+ (TEMRA) T cells in AD
      CSF, supporting the clonal-CD8 element of this node in human disease.
  downstream:
  - target: Synaptic Dysfunction
    description: T cell-derived IFN-gamma and cytotoxic activity, engaged downstream of tau pathology, amplify neuronal and synaptic loss; IFN-gamma/PD-1 blockade reduces brain atrophy.
    hypothesis_groups:
    - adaptive_immune_tcell_model
    - synaptic_failure_convergence_model
    causal_link_type: INDIRECT_KNOWN_INTERMEDIATES
    intermediate_mechanisms:
    - IFN-gamma signaling and CD8+ cytotoxic effector function driving neuronal loss.
    evidence:
    - reference: PMID:36890231
      reference_title: "Microglia-mediated T cell infiltration drives neurodegeneration in tauopathy."
      supports: SUPPORT
      evidence_source: MODEL_ORGANISM
      snippet: "Inhibition of interferon-γ and PDCD1 signalling both significantly ameliorated brain atrophy."
      explanation: >-
        Blocking IFN-gamma or PD-1 signaling reduces brain atrophy, supporting a
        causal edge from the T cell response to neurodegeneration/synaptic failure.
  notes: >-
    PROVISIONAL / EMERGING (adaptive_immune_tcell_model). Core functional
    evidence is from mouse tauopathy models (Chen et al., Nature 2023;
    PMID:36890231) with correlative human AD-brain data. The adaptive response is
    much smaller and less consistent in pure amyloidosis, so this node is curated
    as tau-associated. The recognized antigen(s), the cDC1 cross-presentation
    route, and any B cell contribution remain unresolved.
- name: Muscarinic M1 Receptor Signaling Loss
  biological_scale: MOLECULAR
  description: >-
    Loss of signaling through the M1 muscarinic acetylcholine receptor, arising
    either from degeneration of basal forebrain cholinergic input or from
    sustained pharmacologic antagonism by anticholinergic drugs. M1 signaling
    tonically favors non-amyloidogenic ADAM17-mediated alpha-secretase cleavage
    of APP
    and restrains GSK3-beta; withdrawing it shifts APP processing toward the
    amyloidogenic route and permits tau hyperphosphorylation.
  cell_types:
  - preferred_term: Neurons
    term:
      id: CL:0000540
      label: neuron
  - preferred_term: Cholinergic neuron
    term:
      id: CL:0000108
      label: cholinergic neuron
  molecular_functions:
  - preferred_term: M1 muscarinic acetylcholine receptor activity
    modifier: DECREASED
    term:
      id: GO:0016907
      label: G protein-coupled acetylcholine receptor activity
  biological_processes:
  - preferred_term: M1 muscarinic (Gq/PLC-coupled) receptor signaling
    modifier: DECREASED
    term:
      id: GO:0007207
      label: phospholipase C-activating G protein-coupled acetylcholine receptor signaling pathway
  - preferred_term: Amyloidogenic APP processing
    modifier: INCREASED
    term:
      id: GO:0034205
      label: amyloid-beta formation
  locations:
  - preferred_term: Hippocampus
  - preferred_term: Cerebral Cortex
  - preferred_term: Basal forebrain
  genes:
  - preferred_term: CHRM1
    term:
      id: hgnc:1950
      label: CHRM1
  - preferred_term: ADAM17
    term:
      id: hgnc:195
      label: ADAM17
  - preferred_term: GSK3B
    term:
      id: hgnc:4617
      label: GSK3B
  mechanisms:
  - Loss of M1-dependent activation of ADAM17 (alpha-secretase), diverting APP from the non-amyloidogenic to the amyloidogenic cleavage route.
  - Loss of M1-dependent restraint on GSK3-beta and protein kinase C activity, permitting tau hyperphosphorylation.
  consequences:
  - Amyloid Plaque Formation
  - Neurofibrillary Tangle Formation
  downstream:
  - target: Amyloid Plaque Formation
    description: >-
      Withdrawal of M1 signaling shifts APP processing away from the
      ADAM17/alpha-secretase route toward amyloidogenic cleavage, increasing
      amyloid-beta available for deposition.
    hypothesis_groups:
    - anticholinergic_muscarinic_burden_model
    causal_link_type: INDIRECT_KNOWN_INTERMEDIATES
    intermediate_mechanisms:
    - Reduced ADAM17 (alpha-secretase) activation and consequent diversion of APP to beta/gamma-secretase cleavage.
    evidence:
    - reference: PMID:16504943
      reference_title: M1 receptors play a central role in modulating AD-like pathology in transgenic mice.
      supports: SUPPORT
      evidence_source: MODEL_ORGANISM
      snippet: "The mechanism underlying the effect on the Abeta pathology was caused by the selective activation of ADAM17, thereby shifting APP processing toward the nonamyloidogenic pathway, whereas the reduction in tau pathology is mediated by decreased GSK3beta activity."
      explanation: >-
        Identifies ADAM17 as the effector coupling M1 signaling to
        non-amyloidogenic APP processing, so loss of M1 signaling is the
        directional inverse of this rescue.
    - reference: PMID:21704011
      reference_title: "Loss of muscarinic M1 receptor exacerbates Alzheimer's disease-like pathology and cognitive decline."
      supports: SUPPORT
      evidence_source: MODEL_ORGANISM
      snippet: "Ablating the M(1)R increased plaque and tangle levels in the brains of 3xTgAD mice"
      explanation: >-
        Genetic removal of M1 increases plaque burden in an AD model, the
        loss-of-signaling direction this edge asserts.
  - target: Neurofibrillary Tangle Formation
    description: >-
      Loss of M1 signaling de-represses GSK3-beta and alters protein kinase C
      activity, promoting tau hyperphosphorylation and tangle formation.
    hypothesis_groups:
    - anticholinergic_muscarinic_burden_model
    causal_link_type: INDIRECT_KNOWN_INTERMEDIATES
    intermediate_mechanisms:
    - Increased GSK3-beta activity and altered protein kinase C signaling driving tau hyperphosphorylation.
    evidence:
    - reference: PMID:21704011
      reference_title: "Loss of muscarinic M1 receptor exacerbates Alzheimer's disease-like pathology and cognitive decline."
      supports: SUPPORT
      evidence_source: MODEL_ORGANISM
      snippet: "tau hyperphosphorylation and potentiation of amyloidogenic processing in the mice with AD lacking M(1)R were attributed to changes in the glycogen synthase kinase 3β and protein kinase C activities"
      explanation: >-
        Names GSK3-beta and PKC as the kinases mediating tau
        hyperphosphorylation downstream of M1 loss.
  evidence:
  - reference: PMID:16504943
    reference_title: M1 receptors play a central role in modulating AD-like pathology in transgenic mice.
    supports: SUPPORT
    evidence_source: MODEL_ORGANISM
    snippet: "We further demonstrate that administration of dicyclomine, an M1 antagonist, exacerbates the Abeta and tau pathologies."
    explanation: >-
      Establishes that pharmacologic M1 blockade, not only genetic ablation,
      worsens AD-like pathology.
  notes: >-
    Curated as EMERGING (anticholinergic_muscarinic_burden_model), not as part
    of the canonical cholinergic-deficit account of AD symptoms. Every edge here
    rests on MODEL_ORGANISM evidence in 3xTg-AD/Tg-SwDI mice. Two limits matter
    for anyone reusing this node to reason about drug exposure: dicyclomine and
    the M1 knockout are M1-selective, whereas the drugs driving the human
    epidemiology (diphenhydramine, tricyclics, oxybutynin) are non-selective
    across M1-M5 and additionally block H1; and the mouse work manipulates M1
    against an existing transgenic amyloid/tau burden, which is not the same
    experiment as exposing a pathology-free brain. See the attached
    HUMAN_MODEL_MISMATCH discussion. ADAM17 here is the M1-stimulated arm of
    APP alpha-cleavage (PMID:16504943). Constitutive alpha-secretase cleavage of
    APP in cell lines and primary murine neurons is carried by ADAM10: its
    knockdown, but not that of ADAM9 or ADAM17, suppressed APP alpha-cleavage
    (PMID:20676056).
- name: Complement-Mediated Microglial Synapse Elimination
  biological_scale: CELLULAR
  role: amplifier
  mechanism_confidence: PROVISIONAL
  description: >-
    Microglia actively engulf and remove structurally intact synapses after those
    synapses are tagged by the classical complement cascade. C1q binds
    perisynaptic membranes, C3 opsonizes the synapse, and microglia engage it
    through complement receptor 3 (CR3, the ITGAM/ITGB2 heterodimer). This is the
    same pathway that prunes surplus synapses during normal development,
    reactivated in the adult brain. Both soluble amyloid-beta oligomers and
    phosphorylated tau drive it, so it is a point at which the two proteinopathies
    converge on one cellular execution step. Because synapse loss is the
    pathological change that correlates best with cognitive decline, this node
    supplies the cellular mechanism that the Synaptic Dysfunction node otherwise
    leaves implicit.
  cell_types:
  - preferred_term: Microglia
    term:
      id: CL:0000129
      label: microglial cell
  - preferred_term: Astrocyte
    term:
      id: CL:0000127
      label: astrocyte
  - preferred_term: Neurons
    term:
      id: CL:0000540
      label: neuron
  biological_processes:
  - preferred_term: Complement-mediated synapse pruning
    modifier: INCREASED
    term:
      id: GO:0150062
      label: complement-mediated synapse pruning
  - preferred_term: Complement activation
    modifier: INCREASED
    term:
      id: GO:0006956
      label: complement activation
  - preferred_term: Microglial engulfment of synaptic material
    modifier: INCREASED
    term:
      id: GO:0006909
      label: phagocytosis
  locations:
  - preferred_term: Hippocampus
  - preferred_term: Synapses
  genes:
  - preferred_term: C1QA
    term:
      id: hgnc:1241
      label: C1QA
  - preferred_term: C3
    term:
      id: hgnc:1318
      label: C3
  - preferred_term: ITGAM (CR3)
    term:
      id: hgnc:6149
      label: ITGAM
  - preferred_term: MFGE8
    term:
      id: hgnc:7036
      label: MFGE8
  mechanisms:
  - C1q deposition on perisynaptic membranes ahead of plaque deposition
  - C3 opsonization and microglial CR3 engagement of tagged synapses
  - Microglial phagocytic engulfment of intact synaptic elements
  consequences:
  - Synaptic Dysfunction
  evidence:
  - reference: PMID:27033548
    reference_title: "Complement and microglia mediate early synapse loss in Alzheimer mouse models."
    supports: SUPPORT
    evidence_source: MODEL_ORGANISM
    snippet: "C1q is necessary for the toxic effects of soluble β-amyloid (Aβ) oligomers on synapses and hippocampal long-term potentiation."
    explanation: >-
      Places C1q as a required intermediate between soluble amyloid-beta oligomers
      and both synaptic and electrophysiological damage.
  - reference: PMID:27033548
    reference_title: "Complement and microglia mediate early synapse loss in Alzheimer mouse models."
    supports: SUPPORT
    evidence_source: MODEL_ORGANISM
    snippet: "Finally, microglia in adult brains engulf synaptic material in a CR3-dependent process when exposed to soluble Aβ oligomers."
    explanation: >-
      Demonstrates the engulfment step itself in the adult brain and shows it
      requires the microglial complement receptor.
  - reference: PMID:30392797
    reference_title: "Changes in the Synaptic Proteome in Tauopathy and Rescue of Tau-Induced Synapse Loss by C1q Antibodies."
    supports: SUPPORT
    evidence_source: MODEL_ORGANISM
    snippet: "A C1q-blocking antibody inhibited microglial synapse removal in cultured neurons and in Tau-P301S mice, rescuing synapse density."
    explanation: >-
      Pharmacological rescue of synapse density by blocking C1q, in a tau model —
      the basis for treating this as a therapeutic target rather than a
      description.
  - reference: PMID:37652017
    reference_title: "Human astrocytes and microglia show augmented ingestion of synapses in Alzheimer's disease via MFG-E8."
    supports: SUPPORT
    evidence_source: IN_VITRO
    snippet: "Inhibiting interactions of MFG-E8 rescues the elevated engulfment of AD synapses by astrocytes and microglia without affecting control synapse uptake."
    explanation: >-
      The counterpart of the rescue experiment, but through MFG-E8 rather than
      complement, and with astrocytes participating alongside microglia. Graded
      IN_VITRO because this sentence reports the cultured-cell arm — a phagocytosis
      assay in primary human astrocyte and microglia cultures with a blocking
      antibody. The same study's postmortem-tissue observation is curated
      separately as HUMAN_CLINICAL on the hypothesis group.
  downstream:
  - target: Synaptic Dysfunction
    description: >-
      Complement tagging and microglial engulfment remove intact synapses,
      producing the synapse loss that underlies impaired plasticity and
      neurotransmission.
    hypothesis_groups:
    - complement_synaptic_pruning_model
    - neuroimmune_glial_amplification_model
    - synaptic_failure_convergence_model
    causal_link_type: DIRECT
    evidence:
    - reference: PMID:27033548
      reference_title: "Complement and microglia mediate early synapse loss in Alzheimer mouse models."
      supports: SUPPORT
      evidence_source: MODEL_ORGANISM
      snippet: "Here we show in mouse models that complement and microglia mediate synaptic loss early in AD."
      explanation: States the causal edge from the complement-microglia axis to synapse loss.
  notes: >-
    PROVISIONAL. The necessity evidence (C1q, C3 and CR3 blockade, and the C1q
    antibody rescue) is entirely from mouse models; the human evidence establishes
    that astrocytes and microglia ingest more synaptic material in Alzheimer brain
    and that blocking MFG-E8 reverses it, which corroborates the phenomenon
    without confirming the complement pathway in humans. Note also that complement
    is not uniformly harmful here: C3 deletion protects synapses and cognition
    while increasing plaque burden, so this node models a specific, synapse-directed
    arm of complement activity and should not be collapsed into a general
    "complement is pathogenic" claim. MFGE8 is listed among the genes because it
    is the opsonin implicated in the human data, not because the mouse complement
    work identifies it. MFG-E8 is not a complement component: it bridges exposed
    phosphatidylserine to alphav-beta3/beta5 integrins on the engulfing cell, so
    it is a parallel, complement-independent tag for the same engulfment step.
- name: Neuronal Endosomal-Lysosomal Acidification Failure
  biological_scale: CELLULAR
  role: initiator
  mechanism_confidence: PROVISIONAL
  description: >-
    Activation and enlargement of Rab5-positive early endosomes is the earliest
    recognized intraneuronal change in sporadic Alzheimer disease, present in
    neocortical pyramidal neurons at preclinical stages and, in Down syndrome,
    decades before classical neuropathology. The APP beta-C-terminal fragment
    (APP-βCTF) — not amyloid-beta itself — binds the v-ATPase V0a1 subunit and
    competitively blocks assembly of the V1 subcomplex, so autolysosomes fail to
    acidify and cathepsins are not activated. Amyloid-beta then accumulates inside
    enlarged, de-acidified autolysosomes; in the most compromised neurons these
    vacuoles pack into perikaryal rosettes (the PANTHOS pattern) before lysosomal
    membrane permeabilization kills the cell. This node therefore models amyloid
    pathology as beginning inside the neuron.
  conforms_to: "disabled_macroautophagy#Failure of Cytoplasmic Quality Control"
  cell_types:
  - preferred_term: Neurons
    term:
      id: CL:0000540
      label: neuron
  - preferred_term: Pyramidal neuron
    term:
      id: CL:0000598
      label: pyramidal neuron
  biological_processes:
  - preferred_term: Lysosomal lumen acidification
    modifier: DECREASED
    term:
      id: GO:0007042
      label: lysosomal lumen acidification
  - preferred_term: Endocytic pathway activation
    modifier: INCREASED
    term:
      id: GO:0006897
      label: endocytosis
  cellular_components:
  - preferred_term: Enlarged Rab5-positive early endosome
    term:
      id: GO:0005769
      label: early endosome
  locations:
  - preferred_term: Cerebral Cortex
  - preferred_term: Hippocampus
  chemical_entities:
  - preferred_term: APP beta-C-terminal fragment (APP-βCTF)
  genes:
  - preferred_term: APP
    term:
      id: hgnc:620
      label: APP
  - preferred_term: PSEN1
    term:
      id: hgnc:9508
      label: PSEN1
  mechanisms:
  - Endocytic pathway activation with enlargement of Rab5-positive early endosomes
  - APP-βCTF binding to the v-ATPase V0a1 subunit and competitive inhibition of V1 assembly
  - Failure of autolysosome acidification and cathepsin activation
  - Intraneuronal amyloid-beta accumulation within de-acidified autolysosomes
  consequences:
  - Amyloid Plaque Formation
  - Autophagy-Lysosomal Dysfunction
  evidence:
  - reference: PMID:10880397
    reference_title: "Endocytic pathway abnormalities precede amyloid beta deposition in sporadic Alzheimer's disease and Down syndrome: differential effects of APOE genotype and presenilin mutations."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "These results identify EP activation as the earliest known intraneuronal change to occur in sporadic AD, the most common form of AD."
    explanation: >-
      Human postmortem study positioning endocytic-pathway activation as the
      earliest intraneuronal lesion in sporadic Alzheimer disease.
  - reference: PMID:10880397
    reference_title: "Endocytic pathway abnormalities precede amyloid beta deposition in sporadic Alzheimer's disease and Down syndrome: differential effects of APOE genotype and presenilin mutations."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "In Down syndrome, early endosomes were significantly enlarged in some pyramidal neurons as early as 28 weeks of gestation, decades before classical AD neuropathology develops."
    explanation: >-
      Establishes how far the endosomal lesion precedes classical neuropathology
      in the APP-trisomic human condition.
  - reference: PMID:15465622
    reference_title: "Abeta localization in abnormal endosomes: association with earliest Abeta elevations in AD and Down syndrome."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "We found by ELISA and immunocytochemistry that the appearance of enlarged endosomes coincided with an initial rise in soluble Abeta40 and Abeta42 peptides, which preceded amyloid deposition."
    explanation: >-
      Ties the endosomal lesion in human brain to the earliest rise in soluble
      amyloid-beta, before deposition.
  - reference: PMID:37494443
    reference_title: "Lysosomal dysfunction in Down syndrome and Alzheimer mouse models is caused by v-ATPase inhibition by Tyr(682)-phosphorylated APP βCTF."
    supports: SUPPORT
    evidence_source: IN_VITRO
    snippet: "In human DS fibroblasts, the phosphorylated 682YENPTY internalization motif of APP-βCTF binds selectively within a pocket of the v-ATPase V0a1 subunit cytoplasmic domain and competitively inhibits association of the V1 subcomplex of v-ATPase, thereby reducing its activity."
    explanation: >-
      The molecular mechanism of the acidification failure, worked out in human
      Down syndrome fibroblasts.
  - reference: PMID:35654956
    reference_title: "Faulty autolysosome acidification in Alzheimer's disease mouse models induces autophagic build-up of Aβ in neurons, yielding senile plaques."
    supports: SUPPORT
    evidence_source: MODEL_ORGANISM
    snippet: "Autolysosome acidification declines in neurons well before extracellular amyloid deposition, associated with markedly lowered vATPase activity and build-up of Aβ/APP-βCTF selectively within enlarged de-acidified autolysosomes."
    explanation: >-
      In vivo imaging across five mouse models establishing the temporal ordering
      of acidification failure ahead of extracellular deposition.
  - reference: PMID:35654956
    reference_title: "Faulty autolysosome acidification in Alzheimer's disease mouse models induces autophagic build-up of Aβ in neurons, yielding senile plaques."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "This unique pattern, termed PANTHOS (poisonous anthos (flower)), is also present in AD brains."
    explanation: >-
      The full extent of the verified human claim for PANTHOS — presence of the
      pattern in Alzheimer brains, not that PANTHOS neurons are the source of
      human plaques.
  downstream:
  - target: Amyloid Plaque Formation
    description: >-
      Amyloid-beta accumulated within de-acidified autolysosomes of individual
      neurons is released when those neurons rupture, so the plaque is modeled as
      originating from a dying neuron rather than from secreted peptide.
    hypothesis_groups:
    - endolysosomal_origin_model
    - amyloid_cascade_model
    causal_link_type: INDIRECT_KNOWN_INTERMEDIATES
    intermediate_mechanisms:
    - Lysosomal membrane permeabilization, cathepsin release and lysosomal cell death.
    evidence:
    - reference: PMID:35654956
      reference_title: "Faulty autolysosome acidification in Alzheimer's disease mouse models induces autophagic build-up of Aβ in neurons, yielding senile plaques."
      supports: SUPPORT
      evidence_source: MODEL_ORGANISM
      snippet: "Quantitative analyses confirm that individual neurons exhibiting PANTHOS are the principal source of senile plaques in amyloid precursor protein AD models."
      explanation: >-
        The causal edge from the intraneuronal lesion to the plaque, established in
        amyloid precursor protein transgenic mouse models.
  - target: Autophagy-Lysosomal Dysfunction
    description: >-
      APP-βCTF inhibition of the v-ATPase is modeled as one upstream cause of the
      broader autophagic-flux failure curated in that node.
    hypothesis_groups:
    - endolysosomal_origin_model
    - autophagy_lysosomal_clearance_model
    causal_link_type: DIRECT
    evidence:
    - reference: PMID:37494443
      reference_title: "Lysosomal dysfunction in Down syndrome and Alzheimer mouse models is caused by v-ATPase inhibition by Tyr(682)-phosphorylated APP βCTF."
      supports: SUPPORT
      evidence_source: MODEL_ORGANISM
      snippet: "Elevated APP-βCTF Tyr682 phosphorylation in two mouse AD models similarly disrupts v-ATPase function."
      explanation: >-
        Extends the fibroblast mechanism to Alzheimer mouse brain, supporting the
        edge onto the general autophagy-lysosomal node.
  notes: >-
    PROVISIONAL. The human evidence is strong for the endosomal lesion and its
    timing (Cataldo and colleagues, postmortem cohorts spanning preclinical
    disease) and for the βCTF/v-ATPase mechanism in human Down syndrome
    fibroblasts. It is much weaker for PANTHOS: the only verified human statement
    is that the pattern is present in Alzheimer brains, and the claim that such
    neurons are the principal source of plaques is scoped by its authors to
    amyloid precursor protein transgenic models, where APP is overexpressed. The
    PANTHOS phenotype and the v-ATPase mechanism come from the same laboratory and
    are not independent of each other. Note the negative result inside the human
    data: endosomes were of normal size in advanced presenilin-mutation familial
    disease, so this mechanism does not generalize across all Alzheimer genetic
    subtypes — which sits awkwardly beside the separate finding that PSEN1 is
    required for lysosomal acidification (PMID:20541250), curated on the
    Autophagy-Lysosomal Dysfunction node. See the attached CONTROVERSY discussion
    on plaque origin.
- name: Selective Vulnerability of RORB+ Entorhinal Excitatory Neurons
  biological_scale: CELLULAR
  mechanism_confidence: PROVISIONAL
  description: >-
    Neuronal loss in Alzheimer disease is not uniform. Entorhinal cortex layer II
    excitatory neurons and locus coeruleus noradrenergic neurons degenerate first
    and disproportionately, and within the entorhinal cortex the vulnerable
    excitatory population is marked by RORB. These neurons are selectively
    susceptible to neurofibrillary inclusion and are depleted as pathology
    advances, while neighbouring populations in the same tissue are relatively
    spared. This node models cell-intrinsic vulnerability as a determinant of
    where the disease starts, which aggregate burden alone does not explain.
  cell_types:
  - preferred_term: RORB+ excitatory neuron
    term:
      id: CL:0000679
      label: glutamatergic neuron
  - preferred_term: Pyramidal neuron
    term:
      id: CL:0000598
      label: pyramidal neuron
  biological_processes:
  - preferred_term: Neurofibrillary tangle assembly in vulnerable neurons
    modifier: INCREASED
    term:
      id: GO:1990000
      label: amyloid fibril formation
  locations:
  - preferred_term: Entorhinal Cortex
  - preferred_term: Locus coeruleus
  genes:
  - preferred_term: RORB
    term:
      id: hgnc:10259
      label: RORB
  mechanisms:
  - Cell-type-restricted susceptibility of RORB+ excitatory neurons to neurofibrillary inclusion
  - Early and disproportionate depletion of entorhinal layer II neurons
  consequences:
  - Neurofibrillary Tangle Formation
  - Synaptic Dysfunction
  evidence:
  - reference: PMID:33432193
    reference_title: "Molecular characterization of selectively vulnerable neurons in Alzheimer's disease."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "We identified RORB as a marker of selectively vulnerable excitatory neurons in the entorhinal cortex and subsequently validated their depletion and selective susceptibility to neurofibrillary inclusions during disease progression using quantitative neuropathological methods."
    explanation: >-
      Identifies the vulnerable population molecularly in human postmortem brain
      and confirms both its depletion and its tangle susceptibility by independent
      neuropathology.
  - reference: PMID:8699259
    reference_title: "Profound loss of layer II entorhinal cortex neurons occurs in very mild Alzheimer's disease."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "These results support the conclusion that a marked decrement of layer II neurons distinguishes even very mild AD from nondemented aging."
    explanation: >-
      Stereological evidence that laminar-selective entorhinal neuron loss is
      already established in the mildest clinically detectable disease and is not a
      feature of normal aging.
  - reference: PMID:27513978
    reference_title: "Locus coeruleus volume and cell population changes during Alzheimer's disease progression: A stereological study in human postmortem brains with potential implication for early-stage biomarker discovery."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "As the Braak stage increases by 1 unit, the LC volume decreases by 8.4%."
    explanation: >-
      Quantifies stage-wise loss in a second selectively vulnerable population, the
      locus coeruleus. Partial for this node because the population is
      noradrenergic rather than the RORB+ entorhinal one named here.
  - reference: PMID:37292694
    reference_title: "Integrated multimodal cell atlas of Alzheimer's disease."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "Temporal analysis of cell-type proportions indicated an early reduction of Somatostatin-expressing neuronal subtypes and a late decrease of supragranular intratelencephalic-projecting excitatory and Parvalbumin-expressing neurons"
    explanation: >-
      An important scope limit on the excitatory-neuron framing of this node. In
      middle temporal gyrus the earliest compositional loss is a
      somatostatin-expressing inhibitory population, with excitatory and parvalbumin
      populations declining later — so which cell type is "selectively vulnerable"
      depends on the region examined, and the entorhinal excitatory answer does not
      transfer to neocortex.
  - reference: PMID:30559469
    reference_title: "A tau homeostasis signature is linked with the cellular and regional vulnerability of excitatory neurons to tau pathology."
    supports: SUPPORT
    evidence_source: IN_VITRO
    snippet: "We verified that reducing BAG3 levels in primary neurons exacerbated pathological tau accumulation, whereas BAG3 overexpression attenuated it."
    explanation: >-
      A competing account of the same anatomy that has what this node's own marker
      lacks: perturbation evidence. Proteostasis capacity, indexed by BAG3, changes
      tau accumulation bidirectionally when manipulated, whereas no experiment has
      manipulated RORB and altered tau susceptibility or neuronal death.
  - reference: PMID:41928804
    reference_title: "Brain network dynamics determine tau presence while regional vulnerability governs tau load in Alzheimer's disease."
    supports: SUPPORT
    evidence_source: COMPUTATIONAL
    snippet: "We find that whether tau reaches a brain region (presence) and how much tau accumulates there (load) are governed by different mechanisms."
    explanation: >-
      Supports a hybrid rather than a purely cell-intrinsic account: connectivity
      determines where tau arrives, while regional and cell-intrinsic properties
      determine how much accumulates. Cell-intrinsic vulnerability is then one term
      in the explanation of the disease's anatomy rather than the whole of it.
  downstream:
  - target: Neurofibrillary Tangle Formation
    description: >-
      Cell-intrinsic vulnerability of RORB+ entorhinal excitatory neurons is
      modeled as determining where neurofibrillary inclusions first form, giving
      tau pathology its characteristic anatomical starting point.
    hypothesis_groups:
    - selective_neuronal_vulnerability_model
    - tau_neurodegeneration_model
    causal_link_type: INDIRECT_UNKNOWN_INTERMEDIATES
    evidence:
    - reference: PMID:33432193
      reference_title: "Molecular characterization of selectively vulnerable neurons in Alzheimer's disease."
      supports: SUPPORT
      evidence_source: HUMAN_CLINICAL
      snippet: "Alzheimer's disease (AD) is characterized by the selective vulnerability of specific neuronal populations, the molecular signatures of which are largely unknown."
      explanation: >-
        States the selectivity premise while making explicit that the molecular
        basis is unknown, which is why this edge is curated with unknown
        intermediates.
  notes: >-
    PROVISIONAL as a mechanism, though the underlying observations are
    long-established and essentially uncontested as descriptions. The hypothesis
    bundles three claims that should be graded apart: anatomical selectivity is
    established; the vulnerable population being molecularly definable is
    supported; and cell-intrinsic properties causing the selectivity is
    unresolved. RORB marks the vulnerable population, it has not been shown to
    cause the vulnerability, and no perturbation experiment has tested it.
    Three caveats now travel with this node. First, the marker is region-specific:
    RORB identifies the vulnerable excitatory population in entorhinal cortex, and
    in middle temporal gyrus the earliest compositional loss is a
    somatostatin-expressing inhibitory population instead (PMID:37292694), so the
    principle generalizes while this particular marker does not. Second, competing
    accounts explain much of the same anatomy and one of them carries perturbation
    evidence that RORB does not: proteostasis capacity indexed by BAG3 changes tau
    accumulation bidirectionally when manipulated (PMID:30559469). Third, the
    locus coeruleus data show a long gap between tangle accumulation and neuronal
    loss, so susceptibility to inclusion and susceptibility to death are separable
    properties. The best-supported synthesis is hybrid rather than purely
    cell-intrinsic: connectivity determines where tau arrives, regional and
    cell-intrinsic properties determine how much accumulates (PMID:41928804).
    The entorhinal cohort is small, as postmortem single-nucleus cohorts are. This
    node also supplies the population-level anchor that the HSV-1 Reactivation node
    depends on — that node's RORB+ glutamatergic neurons are selectively vulnerable
    independently of any viral claim.
- name: Necroptotic Neuronal Death
  biological_scale: CELLULAR
  role: effector
  mechanism_confidence: PROVISIONAL
  description: >-
    Neurons die by necroptosis — a caspase-independent, programmed form of
    necrosis in which RIPK1 and RIPK3 assemble a necrosome in which RIPK3
    phosphorylates and activates MLKL, which then permeabilizes the plasma membrane. Activated
    necrosome components are detectable in granulovacuolar degeneration bodies, a
    classical Alzheimer lesion that had no mechanistic account, and their regional
    burden is inversely correlated with neuronal density. The mechanism gives the
    entry an explicit terminal cell-death step, which the amyloid, tau, and
    synaptic nodes otherwise leave unstated.
  cell_types:
  - preferred_term: Neurons
    term:
      id: CL:0000540
      label: neuron
  biological_processes:
  - preferred_term: Necroptotic process
    modifier: INCREASED
    term:
      id: GO:0070266
      label: necroptotic process
  locations:
  - preferred_term: Hippocampus
  - preferred_term: Cerebral Cortex
  genes:
  - preferred_term: RIPK1
    term:
      id: hgnc:10019
      label: RIPK1
  - preferred_term: RIPK3
    term:
      id: hgnc:10021
      label: RIPK3
  - preferred_term: MLKL
    term:
      id: hgnc:26617
      label: MLKL
  mechanisms:
  - RIPK1/RIPK3 necrosome assembly and MLKL phosphorylation in neurons
  - Deposition of activated necrosome components in granulovacuolar degeneration bodies
  - MEG3-driven induction of necroptosis in human neurons
  consequences:
  - Neuroinflammation
  evidence:
  - reference: PMID:28758999
    reference_title: "Necroptosis activation in Alzheimer's disease."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "We found that necroptosis was activated in postmortem human AD brains, positively correlated with Braak stage, and inversely correlated with brain weight and cognitive scores."
    explanation: >-
      Human postmortem evidence relating necroptosis activation to both
      pathological stage and clinical severity.
  - reference: PMID:31802237
    reference_title: "Necrosome complex detected in granulovacuolar degeneration is associated with neuronal loss in Alzheimer's disease."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "We demonstrated that all three activated necrosome components can be detected in GVD lesions (GVDn+, i.e. GVD with activated necrosome) in neurons, that they colocalize with classical GVD markers, such as pTDP-43 and CK1δ, and similarly to these markers detect GVD lesions."
    explanation: >-
      Localizes the complete activated necrosome to granulovacuolar degeneration
      bodies in human neurons, connecting the mechanism to an established lesion.
  - reference: PMID:31802237
    reference_title: "Necrosome complex detected in granulovacuolar degeneration is associated with neuronal loss in Alzheimer's disease."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "GVDn + neurons inversely correlated with neuronal density in the early affected CA1 region of the hippocampus and in the late affected frontal cortex layer III."
    explanation: >-
      Relates necrosome burden to neuronal density in both an early- and a
      late-affected region, supporting the link to actual neuron loss.
  - reference: PMID:37708272
    reference_title: "MEG3 activates necroptosis in human neuron xenografts modeling Alzheimer's disease."
    supports: SUPPORT
    evidence_source: MODEL_ORGANISM
    snippet: "Down-regulation of MEG3 and inhibition of necroptosis using pharmacological or genetic manipulation of receptor-interacting protein kinase 1 (RIPK1), RIPK3, or mixed lineage kinase domain-like protein (MLKL) rescued neuronal cell loss in xenografted human neurons."
    explanation: >-
      Causal evidence: blocking three separate necroptosis effectors each rescues
      loss of human neurons in an amyloid-rich brain.
  - reference: PMID:32949047
    reference_title: "Necrosome-positive granulovacuolar degeneration is associated with TDP-43 pathological lesions in the hippocampus of ALS/FTLD cases."
    supports: SUPPORT
    directness: INDIRECT
    evidence_source: HUMAN_CLINICAL
    snippet: "Necrosome-positive GVD was primarily observed in hippocampal regions of ALS/FTLD cases and was associated with hippocampal TDP-43 inclusions as the main predictor of the pMLKL-GVD stage, as well as with the Braak stage of neurofibrillary tangle pathology."
    explanation: >-
      Establishes that necrosome-positive granulovacuolar degeneration is not
      specific to Alzheimer disease, so this node models a shared execution
      mechanism rather than a disease-defining lesion.
  - reference: PMID:29238035
    reference_title: "Long non-coding RNA MEG3 functions as a competing endogenous RNA to regulate ischemic neuronal death by targeting miR-21/PDCD4 signaling pathway."
    supports: SUPPORT
    directness: INDIRECT
    evidence_source: MODEL_ORGANISM
    snippet: "Besides, overexpression of miR-21 protects oxygen-glucose deprivation and reoxygenation (OGD/R)-induced apoptotic cell death. Furthermore, MEG3 functions as a competing endogenous RNAs (ceRNAs) and competes with programmed cell death 4 (PDCD4) mRNA for directly binding to miR-21, which mediates ischemic neuronal death."
    explanation: >-
      Qualifies the proposed upstream trigger rather than the execution step. In
      ischemic injury the same long non-coding RNA drives neuronal death through a
      caspase-dependent, apoptotic route, so MEG3 looks like a general pro-death
      effector whose downstream modality is context-dependent — and the necroptotic
      reading may belong to the amyloid xenograft context rather than to MEG3
      itself.
  downstream:
  - target: Neuroinflammation
    description: >-
      Necroptosis is lytic: unlike apoptosis it releases intracellular contents,
      so necroptotic neuronal death is modeled as feeding the glial inflammatory
      response rather than being silently cleared.
    hypothesis_groups:
    - necroptosis_model
    - neuroimmune_glial_amplification_model
    causal_link_type: INDIRECT_UNKNOWN_INTERMEDIATES
    evidence:
    - reference: PMID:28758999
      reference_title: "Necroptosis activation in Alzheimer's disease."
      supports: SUPPORT
      evidence_source: HUMAN_CLINICAL
      snippet: "In addition, we found that the set of genes regulated by RIPK1 overlapped significantly with multiple independent AD transcriptomic signatures, indicating that RIPK1 activity could explain a substantial portion of transcriptomic changes in AD."
      explanation: >-
        Supports RIPK1 activity shaping the broader Alzheimer transcriptional
        state, of which the inflammatory response is a major component. The
        specific edge to glial activation remains inferred, hence unknown
        intermediates.
  notes: >-
    PROVISIONAL. The human evidence is correlative and depends heavily on
    phospho-MLKL and phospho-RIPK immunodetection in postmortem tissue, where
    antibody specificity and postmortem interval are recognized problems. The
    causal arm comes from human neurons xenografted into a mouse amyloid brain,
    a system whose headline result is that mouse neurons in the same brain do not
    show the phenotype — evidence for human relevance, and a caution that rodent
    models may miss the death mechanism altogether. The node now carries an
    upstream edge from tau pathology (PMID:35971179), which supplies the trigger
    the mechanism previously lacked.
    Two open attributions matter. First, cell type: the evidence placing the
    necrosome in granulovacuolar degeneration is unambiguously neuronal, but
    necroptosis in Alzheimer
    disease is separately framed as an innate-immune programme engaged by microglia
    and astrocytes alongside pyroptosis and apoptosis (PMID:35741014), so any
    transcriptomic necroptosis signal could originate in glia rather than neurons.
    Second, relationship to the entry's PARP1-Mediated Parthanatos node is
    unresolved: both are caspase-independent regulated necrosis, and nothing
    establishes which dominates, or whether they operate in different cells or at
    different stages.
    A methodological limit applies to all transcriptomic tests of this node:
    necroptosis is executed by phosphorylation of MLKL, not by transcription, so
    transcript abundance of RIPK1, RIPK3, MLKL or MEG3 is an indirect proxy at
    best. Cell-resolved expression data can test where the programme is engaged;
    only phospho-protein or imaging data can show that it is executing.
- name: Senescent Cell Accumulation
  biological_scale: CELLULAR
  role: amplifier
  mechanism_confidence: PROVISIONAL
  conforms_to: "cellular_senescence#Senescent Cell Accumulation"
  description: >-
    Cells that have entered an irreversible cell-cycle arrest but remain
    metabolically active accumulate in the aging and Alzheimer brain, expressing
    p16INK4A/CDKN2A and p21/CDKN1A and secreting a proinflammatory
    senescence-associated secretory phenotype. Clearing them — genetically or with
    senolytic drugs — prevents gliosis, tau hyperphosphorylation and tangle
    deposition, reduces amyloid load and neuroinflammation, and preserves
    cognition in mouse models. Which cell type is senescent is genuinely
    unresolved: mouse work implicates astrocytes and microglia in tauopathy and
    oligodendrocyte progenitor cells around plaques, while the largest human
    survey assigns over 97% of senescent cells to excitatory neurons.
  cell_types:
  - preferred_term: Astrocyte
    term:
      id: CL:0000127
      label: astrocyte
  - preferred_term: Microglia
    term:
      id: CL:0000129
      label: microglial cell
  - preferred_term: Oligodendrocyte precursor cell
    term:
      id: CL:0002453
      label: oligodendrocyte precursor cell
  - preferred_term: Excitatory neuron
    term:
      id: CL:0000679
      label: glutamatergic neuron
  biological_processes:
  - preferred_term: Cellular senescence
    modifier: INCREASED
    term:
      id: GO:0090398
      label: cellular senescence
  locations:
  - preferred_term: Hippocampus
  - preferred_term: Cerebral Cortex
  genes:
  - preferred_term: CDKN2A
    term:
      id: hgnc:1787
      label: CDKN2A
  mechanisms:
  - Amyloid-beta-induced senescence of plaque-associated oligodendrocyte progenitor cells
  - Accumulation of p16INK4A-positive senescent glia in tauopathy
  - Senescence-associated secretory phenotype driving neuroinflammation
  consequences:
  - Neurofibrillary Tangle Formation
  - Neuroinflammation
  evidence:
  - reference: PMID:30232451
    reference_title: "Clearance of senescent glial cells prevents tau-dependent pathology and cognitive decline."
    supports: SUPPORT
    evidence_source: MODEL_ORGANISM
    snippet: "Here we show a causal link between the accumulation of senescent cells and cognition-associated neuronal loss."
    explanation: >-
      Genetic clearance of p16-positive cells as they arise prevents tau pathology
      and neuronal loss, establishing senescence as causal in this model.
  - reference: PMID:30936558
    reference_title: "Senolytic therapy alleviates Aβ-associated oligodendrocyte progenitor cell senescence and cognitive deficits in an Alzheimer's disease model."
    supports: SUPPORT
    evidence_source: MODEL_ORGANISM
    snippet: "Senolytic treatment of AD mice selectively removed senescent cells from the plaque environment, reduced neuroinflammation, lessened Aβ load, and ameliorated cognitive deficits."
    explanation: >-
      Pharmacological rather than genetic clearance, in an amyloid rather than a
      tau model, reproducing the same set of benefits.
  - reference: PMID:30936558
    reference_title: "Senolytic therapy alleviates Aβ-associated oligodendrocyte progenitor cell senescence and cognitive deficits in an Alzheimer's disease model."
    supports: SUPPORT
    evidence_source: IN_VITRO
    snippet: "Direct exposure of cultured OPCs to aggregating Aβ triggered cell senescence."
    explanation: >-
      Identifies aggregating amyloid-beta as a sufficient senescence-inducing
      stimulus for oligodendrocyte progenitor cells, supplying the upstream trigger.
  - reference: PMID:35531351
    reference_title: "Profiling senescent cells in human brains reveals neurons with CDKN2D/p19 and tau neuropathology."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "More than 97% of the senescent cells were excitatory neurons and overlapped with tau-containing neurofibrillary tangles (NFTs)."
    explanation: >-
      Supports senescence being present in human Alzheimer brain and associated
      with tangles, while contradicting the glial cell-type assignment on which
      both mouse clearance experiments rest.
  - reference: PMID:30126037
    reference_title: "Tau protein aggregation is associated with cellular senescence in the brain."
    supports: SUPPORT
    evidence_source: MODEL_ORGANISM
    snippet: "Cdkn2a transcript level, a hallmark measure of senescence, directly correlated with brain atrophy and NFT burden in mice."
    explanation: >-
      Relates a canonical senescence marker quantitatively to both tangle burden
      and brain atrophy, tying this node to the tangle and neurodegeneration arms
      of the entry.
  downstream:
  - target: Neurofibrillary Tangle Formation
    description: >-
      Senescent cells are modeled as promoting tau hyperphosphorylation and tangle
      deposition; the direction is established by the inverse experiment, in which
      clearing them prevents both.
    hypothesis_groups:
    - cellular_senescence_model
    - tau_neurodegeneration_model
    causal_link_type: INDIRECT_KNOWN_INTERMEDIATES
    intermediate_mechanisms:
    - Senescence-associated secretory phenotype and gliosis.
    evidence:
    - reference: PMID:30232451
      reference_title: "Clearance of senescent glial cells prevents tau-dependent pathology and cognitive decline."
      supports: SUPPORT
      evidence_source: MODEL_ORGANISM
      snippet: "Collectively, these results show that senescent cells have a role in the initiation and progression of tau-mediated disease, and suggest that targeting senescent cells may provide a therapeutic avenue for the treatment of these pathologies."
      explanation: >-
        States the causal role of senescent cells in initiating and progressing
        tau-mediated disease.
  - target: Neuroinflammation
    description: >-
      The senescence-associated secretory phenotype is modeled as a chronic
      proinflammatory input to glial activation; senolytic clearance reduces
      neuroinflammation.
    hypothesis_groups:
    - cellular_senescence_model
    - neuroimmune_glial_amplification_model
    - asi_axis_reciprocal_cycle
    causal_link_type: DIRECT
    evidence:
    - reference: PMID:30936558
      reference_title: "Senolytic therapy alleviates Aβ-associated oligodendrocyte progenitor cell senescence and cognitive deficits in an Alzheimer's disease model."
      supports: SUPPORT
      evidence_source: MODEL_ORGANISM
      snippet: "Our findings suggest a role for Aβ-induced OPC cell senescence in neuroinflammation and cognitive deficits in AD, and a potential therapeutic benefit of senolytic treatments."
      explanation: >-
        States the modeled edge from amyloid-induced progenitor senescence to
        neuroinflammation and cognitive deficit.
  notes: >-
    PROVISIONAL. All causal evidence is murine; the human evidence is a single
    postmortem observation of senescent oligodendrocyte progenitor cells plus a
    much larger transcriptomic survey that reaches a different cell-type answer.
    Four cell types are listed here deliberately, because the three studies
    disagree — astrocytes and microglia in tauopathy, oligodendrocyte progenitor
    cells around plaques, and excitatory neurons in the human dataset. That
    disagreement has a direct therapeutic consequence, since senolytics kill the
    cells they target: a drug designed to clear senescent glia would, on the human
    data, be aimed at postmitotic neurons. See the attached CONTROVERSY discussion.
- name: Oligodendrocyte and Myelin Dysfunction
  biological_scale: TISSUE
  role: initiator
  mechanism_confidence: PROVISIONAL
  conforms_to: "cns_myelin_failure#Deficient or Unstable CNS Myelin Sheath"
  description: >-
    Age-related loss of myelin integrity and of oligodendrocyte metabolic support
    for the axon is modeled as an upstream contributor to amyloid deposition
    rather than a consequence of it. Myelin damage concentrates the amyloidogenic
    processing machinery within axonal swellings and increases cleavage of
    cortical amyloid precursor protein, while simultaneously diverting
    disease-associated microglia toward myelin debris and away from plaques — so
    one lesion both raises amyloid production and lowers its clearance. In APOE4
    carriers, cholesterol is aberrantly retained in oligodendrocytes and
    myelination is reduced, giving the strongest genetic risk factor a route
    through this cell type.
  cell_types:
  - preferred_term: Oligodendrocyte
    term:
      id: CL:0000128
      label: oligodendrocyte
  - preferred_term: Oligodendrocyte precursor cell
    term:
      id: CL:0002453
      label: oligodendrocyte precursor cell
  - preferred_term: Microglia
    term:
      id: CL:0000129
      label: microglial cell
  biological_processes:
  - preferred_term: Myelination
    modifier: DECREASED
    term:
      id: GO:0042552
      label: myelination
  - preferred_term: Cholesterol metabolic process in oligodendrocytes
    modifier: DYSREGULATED
    term:
      id: GO:0008203
      label: cholesterol metabolic process
  locations:
  - preferred_term: Cerebral Cortex
  - preferred_term: White matter
  genes:
  - preferred_term: APOE
    term:
      id: hgnc:613
      label: APOE
  - preferred_term: APP
    term:
      id: hgnc:620
      label: APP
  mechanisms:
  - Accumulation of amyloidogenic processing machinery within axonal swellings at sites of myelin damage
  - Diversion of disease-associated microglia from plaques to myelin debris
  - APOE4-associated cholesterol retention in oligodendrocytes with reduced myelination
  consequences:
  - Amyloid Plaque Formation
  evidence:
  - reference: PMID:37258678
    reference_title: "Myelin dysfunction drives amyloid-β deposition in models of Alzheimer's disease."
    supports: SUPPORT
    evidence_source: MODEL_ORGANISM
    snippet: "Here we identify genetic pathways of myelin dysfunction and demyelinating injuries as potent drivers of amyloid deposition in mouse models of AD."
    explanation: >-
      Multiple independent myelin-mutant crosses each increase amyloid deposition,
      establishing direction of causation in the mouse.
  - reference: PMID:37258678
    reference_title: "Myelin dysfunction drives amyloid-β deposition in models of Alzheimer's disease."
    supports: SUPPORT
    evidence_source: MODEL_ORGANISM
    snippet: "Despite successful induction, amyloid disease-associated microglia (DAM) that usually clear amyloid plaques are apparently distracted to nearby myelin damage."
    explanation: >-
      The clearance half of the mechanism: myelin damage competes for the same
      microglial population that would otherwise remove plaques.
  - reference: PMID:36385529
    reference_title: "APOE4 impairs myelination via cholesterol dysregulation in oligodendrocytes."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "We show that altered cholesterol localization in the APOE4 brain coincides with reduced myelination."
    explanation: >-
      Human postmortem evidence that oligodendrocyte cholesterol dysregulation and
      reduced myelination are real features of the APOE4 brain.
  - reference: PMID:36385529
    reference_title: "APOE4 impairs myelination via cholesterol dysregulation in oligodendrocytes."
    supports: SUPPORT
    evidence_source: MODEL_ORGANISM
    snippet: "Pharmacologically facilitating cholesterol transport increases axonal myelination and improves learning and memory in APOE4 mice."
    explanation: >-
      Rescue arm showing the oligodendrocyte cholesterol defect is correctable and
      that correcting it improves cognition in an APOE4 model.
  - reference: PMID:32619874
    reference_title: "Disruption of oligodendrocyte progenitor cells is an early sign of pathology in the triple transgenic mouse model of Alzheimer's disease."
    supports: SUPPORT
    evidence_source: MODEL_ORGANISM
    snippet: "Notably, the loss of myelin and OPC sister cells occurred earlier at 6 months in 3xTg-AD, suggesting accelerated aging, although there was not a concomitant decline in OPC numbers at this age, suggesting the observed changes in myelin were not a consequence of replicative exhaustion, but possibly of OPC disruption or senescence."
    explanation: >-
      Evidence for the reverse direction, and a reason not to read this hypothesis
      as settled. In a model carrying amyloid and tau transgenes, myelin and
      oligodendrocyte-progenitor changes appear as an early consequence of the
      pathology — so oligodendrocyte-lineage disruption can be downstream of
      proteinopathy rather than upstream of it.
  downstream:
  - target: Amyloid Plaque Formation
    description: >-
      Myelin dysfunction is modeled as increasing amyloid deposition both directly,
      by concentrating amyloidogenic processing in axonal swellings, and
      indirectly, by diverting plaque-clearing microglia to myelin damage.
    hypothesis_groups:
    - myelin_oligodendrocyte_model
    - amyloid_cascade_model
    - neuroimmune_glial_amplification_model
    causal_link_type: INDIRECT_KNOWN_INTERMEDIATES
    intermediate_mechanisms:
    - Axonal accumulation of amyloid precursor protein processing machinery.
    - Competition for the disease-associated microglial response.
    evidence:
    - reference: PMID:37258678
      reference_title: "Myelin dysfunction drives amyloid-β deposition in models of Alzheimer's disease."
      supports: SUPPORT
      evidence_source: MODEL_ORGANISM
      snippet: "Mechanistically, myelin dysfunction causes the accumulation of the Aβ-producing machinery within axonal swellings and increases the cleavage of cortical amyloid precursor protein."
      explanation: >-
        Supplies the subcellular mechanism for the edge from myelin damage to
        increased amyloid production.
  notes: >-
    PROVISIONAL, and the two halves of the hypothesis sit at different evidence
    levels. That myelin and oligodendrocyte pathology co-occurs with, and is
    entangled with, amyloid and APOE4 biology is well supported across mouse,
    human postmortem and imaging data. That myelin dysfunction is UPSTREAM of
    amyloid deposition is established only in mouse, and the authors of that work
    frame their conclusion explicitly as a working model. No human observational
    dataset can establish the direction — cross-sectional postmortem data can show
    co-occurrence only — and the decisive human material behind the
    APOE4-myelination work (PMID:36385529) is controlled-access through Synapse.
    Reverse-direction evidence is not trivial and is now curated alongside:
    oligodendrocyte-progenitor and myelin changes appear early as a consequence of
    amyloid and tau pathology in 3xTg-AD mice (PMID:32619874).
    The microglial-diversion sub-mechanism should be read as SPECULATIVE rather
    than merely less replicated: a targeted search for independent primary support
    for that clearance edge found none.
    Two qualifications on the APOE4 route. The best-powered human study localizes
    the dominant lipid-droplet phenotype to microglia via ACSL1 (PMID:38480892),
    not to oligodendrocytes, so "APOE4 acts through the oligodendrocyte route" is
    only partly supported. And this node should not be read as validating the
    entry's Oligodendrocyte Precursor Cell Plasma Proteomic Age Gap biomarker:
    that biomarker comes from a plasma clock spanning more than forty cell types
    (PMID:42297981) in which the strongest APOE4-linked Alzheimer signal is
    astrocyte aging, not oligodendrocyte or progenitor aging. The biomarker is
    consistent with this model but does not privilege it.
- name: Glial and Pericyte Mesenchymal Drift
  conforms_to: "mesenchymal_drift#Mesenchymal Drift of Differentiated Cells"
  description: >-
    Astrocytes, oligodendrocytes and pericytes in the Alzheimer's disease brain up-regulate the
    mesenchymal drift gene signature, which persists in age-matched comparisons; neurons, microglia and
    oligodendrocyte precursor cells did not show it in the same reanalysis. This is evidence at the level
    of a transcriptional signature only. Whether the drifting glia actually change morphology, deposit
    matrix, or contribute to disease progression is untested, so the node carries no downstream edges.
  role: central_effector
  biological_scale: CELLULAR
  cell_types:
  - preferred_term: astrocyte
    term:
      id: CL:0000127
      label: astrocyte
  - preferred_term: oligodendrocyte
    term:
      id: CL:0000128
      label: oligodendrocyte
  - preferred_term: brain pericyte
    term:
      id: CL:2000043
      label: brain pericyte
  biological_processes:
  - preferred_term: mesenchymal drift (aberrant acquisition of mesenchymal identity)
    term:
      id: GO:0048762
      label: mesenchymal cell differentiation
    modifier: ABNORMAL
  evidence:
  - reference: PMID:42680479
    reference_title: "Alzheimer's disease through the lens of mesenchymal drift: the collapse of cellular identity."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    quote_role: REVIEW_SYNTHESIS
    snippet: >-
      In a reanalysis of published single-nucleus RNA-sequencing data from AD and control brain, Lu et al.
      found significant upregulation of MD genes in astrocytes, oligodendrocytes, and pericytes, but not in
      neurons, microglia, or oligodendrocyte progenitor cells (OPCs); this pattern persisted after
      restricting the comparison to age-matched donors
    explanation: >-
      Review summary of the Alzheimer's single-nucleus result in Lu et al. 2025 (PMID:40816266), naming
      the three cell types bound here and the three in which the signature was not found. Graded
      HUMAN_CLINICAL for the human brain data it describes.
  - reference: PMID:42680479
    reference_title: "Alzheimer's disease through the lens of mesenchymal drift: the collapse of cellular identity."
    supports: SUPPORT
    evidence_source: OTHER
    quote_role: REVIEW_SYNTHESIS
    snippet: >-
      It remains unclear whether transcriptomically drifting astrocytes and microglia undergo true
      cytoskeletal changes, dismantle their networks, or deposit fibrotic matrices in the brain, as
      mesenchymal cells do in other tissues.
    explanation: >-
      Records the limit of the evidence: the drift is shown as a signature, not as a physical change in
      the cells. Evidence source is OTHER because this is the review's own assessment.
- name: Microglial Lipid Droplet Accumulation
  biological_scale: CELLULAR
  role: amplifier
  mechanism_confidence: PROVISIONAL
  description: >-
    A distinct microglial state defined by accumulation of cytoplasmic lipid
    droplets. In the aging brain these lipid-droplet-accumulating microglia are
    defective in phagocytosis, generate high levels of reactive oxygen species and
    secrete proinflammatory cytokines. In Alzheimer disease the corresponding
    human state is marked by expression of the long-chain acyl-CoA synthetase
    ACSL1 and is most
    abundant in APOE4 homozygotes; fibrillar amyloid-beta induces it in an
    APOE-dependent manner, and factors secreted by these microglia phosphorylate
    tau and are neurotoxic. The state therefore couples the strongest common
    genetic risk factor to a specific, dysfunctional glial phenotype.
  cell_types:
  - preferred_term: Microglia
    term:
      id: CL:0000129
      label: microglial cell
  biological_processes:
  - preferred_term: Lipid droplet organization
    modifier: INCREASED
    term:
      id: GO:0034389
      label: lipid droplet organization
  - preferred_term: Microglial phagocytosis
    modifier: DECREASED
    term:
      id: GO:0006909
      label: phagocytosis
  cellular_components:
  - preferred_term: Lipid droplet
    term:
      id: GO:0005811
      label: lipid droplet
  locations:
  - preferred_term: Hippocampus
  - preferred_term: Cerebral Cortex
  genes:
  - preferred_term: ACSL1
    term:
      id: hgnc:3569
      label: ACSL1
  - preferred_term: APOE
    term:
      id: hgnc:613
      label: APOE
  mechanisms:
  - Fibrillar amyloid-beta induction of ACSL1, triglyceride synthesis and lipid droplet accumulation
  - APOE genotype dependence of the lipid droplet microglial state
  - Secretion of tau-phosphorylating and neurotoxic factors by lipid-droplet-laden microglia
  consequences:
  - Neuroinflammation
  - Neurofibrillary Tangle Formation
  evidence:
  - reference: PMID:38480892
    reference_title: "APOE4/4 is linked to damaging lipid droplets in Alzheimer's disease microglia."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "Through single-nucleus RNA sequencing of brain tissue in Alzheimer's disease, we have identified a microglial state defined by the expression of the lipid droplet-associated enzyme ACSL1 with ACSL1-positive microglia being most abundant in patients with Alzheimer's disease having the APOE4/4 genotype."
    explanation: >-
      Human postmortem single-nucleus evidence defining the state and tying its
      abundance to APOE4 homozygosity.
  - reference: PMID:38480892
    reference_title: "APOE4/4 is linked to damaging lipid droplets in Alzheimer's disease microglia."
    supports: SUPPORT
    evidence_source: IN_VITRO
    snippet: "In human induced pluripotent stem cell-derived microglia, fibrillar Aβ induces ACSL1 expression, triglyceride synthesis and lipid droplet accumulation in an APOE-dependent manner."
    explanation: >-
      Establishes fibrillar amyloid-beta as the inducing stimulus and APOE as
      required, in human microglia.
  - reference: PMID:31959936
    reference_title: "Lipid-droplet-accumulating microglia represent a dysfunctional and proinflammatory state in the aging brain."
    supports: SUPPORT
    evidence_source: MODEL_ORGANISM
    snippet: "These cells, which we call 'lipid-droplet-accumulating microglia' (LDAM), are defective in phagocytosis, produce high levels of reactive oxygen species and secrete proinflammatory cytokines."
    explanation: >-
      Characterizes the functional deficits of the state. Partial for this node
      because the work describes the aging brain and makes no Alzheimer-specific
      claim.
  downstream:
  - target: Neurofibrillary Tangle Formation
    description: >-
      Factors secreted by lipid-droplet-containing microglia drive tau
      phosphorylation and neurotoxicity in an APOE-dependent manner, linking the
      glial lipid state to tau pathology.
    hypothesis_groups:
    - apoe_risk_modulation_model
    - neuroimmune_glial_amplification_model
    - tau_neurodegeneration_model
    causal_link_type: INDIRECT_KNOWN_INTERMEDIATES
    intermediate_mechanisms:
    - Secreted microglia-derived factors acting on neurons.
    evidence:
    - reference: PMID:38480892
      reference_title: "APOE4/4 is linked to damaging lipid droplets in Alzheimer's disease microglia."
      supports: SUPPORT
      evidence_source: IN_VITRO
      snippet: "Additionally, conditioned media from lipid droplet-containing microglia lead to Tau phosphorylation and neurotoxicity in an APOE-dependent manner."
      explanation: >-
        Conditioned-media transfer establishes that the effect on tau is mediated
        by secreted factors rather than contact.
  - target: Neuroinflammation
    description: >-
      Lipid-droplet-accumulating microglia are a proinflammatory, phagocytically
      defective state, so their accumulation is modeled as amplifying
      neuroinflammation while reducing amyloid clearance capacity.
    hypothesis_groups:
    - neuroimmune_glial_amplification_model
    - apoe_risk_modulation_model
    causal_link_type: DIRECT
    evidence:
    - reference: PMID:31959936
      reference_title: "Lipid-droplet-accumulating microglia represent a dysfunctional and proinflammatory state in the aging brain."
      supports: SUPPORT
      evidence_source: MODEL_ORGANISM
      snippet: "RNA-sequencing analysis of LDAM revealed a transcriptional profile driven by innate inflammation that is distinct from previously reported microglial states."
      explanation: >-
        Supports the state being transcriptionally inflammatory and distinct from
        other described microglial states, in the aging rather than Alzheimer
        brain.
  notes: >-
    PROVISIONAL. The Alzheimer- and APOE4-specific evidence is human-first
    (postmortem single-nucleus sequencing plus human iPSC-derived microglia), which
    is unusual for a glial-state claim, but it is recent and independent
    replication of ACSL1 as the defining marker is still accruing. The
    neurotoxicity arm is conditioned-media in vitro, so "lipid droplet microglia
    kill neurons in the Alzheimer brain" is not established. The founding LDAM
    description is an aging study and makes no Alzheimer claim, which is why its
    evidence items are graded PARTIAL here.
- name: Interneuron Dysfunction and Network Hypersynchrony
  biological_scale: TISSUE
  mechanism_confidence: PROVISIONAL
  description: >-
    Amyloid-beta impairs parvalbumin-expressing inhibitory interneurons, in part
    through reduced levels of Nav1.1 (SCN1A), the voltage-gated sodium channel
    subunit these cells depend on to fire at high frequency. Loss of
    parvalbumin-cell output degrades gamma oscillations and releases the network
    into hypersynchrony, producing epileptiform activity. In patients this appears
    as subclinical epileptiform discharges detectable in a large minority of people
    with Alzheimer disease who have never had a clinical seizure, and their
    presence tracks faster cognitive decline. The node models cognitive impairment
    as arising partly from failure of inhibition, not only from loss of excitatory
    synapses.
  cell_types:
  - preferred_term: Parvalbumin-expressing inhibitory interneuron
    term:
      id: CL:4023018
      label: pvalb GABAergic interneuron
  - preferred_term: GABAergic neuron
    term:
      id: CL:0000617
      label: GABAergic neuron
  molecular_functions:
  - preferred_term: Nav1.1 voltage-gated sodium channel activity
    modifier: DECREASED
    term:
      id: GO:0005248
      label: voltage-gated sodium channel activity
  biological_processes:
  - preferred_term: Inhibitory synaptic transmission
    modifier: DECREASED
    term:
      id: GO:0051932
      label: synaptic transmission, GABAergic
  locations:
  - preferred_term: Hippocampus
  - preferred_term: Cerebral Cortex
  genes:
  - preferred_term: SCN1A
    term:
      id: hgnc:10585
      label: SCN1A
  mechanisms:
  - Reduced Nav1.1 (SCN1A) levels in parvalbumin-expressing interneurons
  - Degraded gamma oscillatory activity and loss of inhibitory control
  - Network hypersynchrony and spontaneous epileptiform discharges
  consequences:
  - Memory Loss
  - Synaptic Dysfunction
  evidence:
  - reference: PMID:22541439
    reference_title: "Inhibitory interneuron deficit links altered network activity and cognitive dysfunction in Alzheimer model."
    supports: SUPPORT
    evidence_source: MODEL_ORGANISM
    snippet: "Electroencephalographic recordings in hAPP mice revealed spontaneous epileptiform discharges, indicating network hypersynchrony, primarily during reduced gamma oscillatory activity."
    explanation: >-
      Establishes the electrophysiological phenotype and its coupling to reduced
      gamma activity in an amyloid model.
  - reference: PMID:22541439
    reference_title: "Inhibitory interneuron deficit links altered network activity and cognitive dysfunction in Alzheimer model."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "Supporting this hypothesis, hAPP mice and AD patients had decreased levels of the interneuron-specific and PV cell-predominant voltage-gated sodium channel subunit Nav1.1."
    explanation: >-
      Contains the human limb of the Nav1.1 claim. Graded PARTIAL because the
      human observation is a protein-level measurement reported alongside the
      mouse result rather than a separate human study.
  - reference: PMID:27696483
    reference_title: "Incidence and impact of subclinical epileptiform activity in Alzheimer's disease."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "Subclinical epileptiform activity was detected in 42.4% of AD patients and 10.5% of controls (p = 0.02)."
    explanation: >-
      Prospective blinded extended EEG and MEG monitoring quantifying the predicted
      hyperexcitability in patients with no seizure history.
  - reference: PMID:23835471
    reference_title: "Seizures and epileptiform activity in the early stages of Alzheimer disease."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "The timing of seizure onset in patients with aMCI and AD was nonuniform (P < .001), clustering near the onset of cognitive decline."
    explanation: >-
      Places the emergence of overt seizure activity close to the onset of
      cognitive decline rather than at end stage, consistent with hyperexcitability
      being an early network event.
  - reference: PMID:42778763
    reference_title: "AI-based characterization of Alzheimer's disease phenotypes from population-scale single-cell data."
    supports: SUPPORT
    evidence_source: COMPUTATIONAL
    quote_role: PRIMARY_RESULT
    snippet: "enhanced excitatory/inhibitory imbalance and mitochondrial dysfunction in cognitively impaired AD donors"
    explanation: >-
      Independent human single-nucleus support for excitatory/inhibitory
      imbalance tracking cognitive impairment. The full text contrasts
      cognitively impaired with cognitively resilient donors at matched late Braak
      stage, finding excitatory-neuron downregulation of the potassium channels
      KCNB2 and KCND2 and opposite regulation of ADCY2 and CBLN2 in excitatory and
      inhibitory neurons, and reports SST interneurons strongly depleted of
      Alzheimer-associated cells. It does not examine Nav1.1 or parvalbumin-cell
      firing, so it supports the node's network-imbalance claim, not its SCN1A
      mechanism.
  downstream:
  - target: Memory Loss
    description: >-
      Loss of inhibitory control and network hypersynchrony are modeled as direct
      contributors to cognitive impairment; restoring Nav1.1 in a model reverses
      both the hypersynchrony and the memory deficit.
    hypothesis_groups:
    - network_hyperexcitability_model
    - synaptic_failure_convergence_model
    causal_link_type: INDIRECT_KNOWN_INTERMEDIATES
    intermediate_mechanisms:
    - Degraded gamma oscillations and hippocampal-cortical network hypersynchrony.
    evidence:
    - reference: PMID:22541439
      reference_title: "Inhibitory interneuron deficit links altered network activity and cognitive dysfunction in Alzheimer model."
      supports: SUPPORT
      evidence_source: MODEL_ORGANISM
      snippet: "Restoring Nav1.1 levels in hAPP mice by Nav1.1-BAC expression increased inhibitory synaptic activity and gamma oscillations and reduced hypersynchrony, memory deficits, and premature mortality."
      explanation: >-
        Gain-of-function rescue of a single interneuron channel subunit corrects
        oscillations, hypersynchrony and memory, establishing the edge as causal in
        this model.
    - reference: PMID:27696483
      reference_title: "Incidence and impact of subclinical epileptiform activity in Alzheimer's disease."
      supports: SUPPORT
      evidence_source: HUMAN_CLINICAL
      snippet: "However, patients with subclinical epileptiform activity showed faster declines in global cognition, determined by the Mini-Mental State Examination (3.9 points/year in patients with epileptiform activity vs 1.6 points/year in patients without; p = 0.006), and in executive function (p = 0.01)."
      explanation: >-
        Human correlate of the same edge: hyperexcitability is associated with more
        than twice the annual rate of cognitive decline.
  notes: >-
    PROVISIONAL. The mechanistic arm is entirely in amyloid precursor protein
    transgenic mice, and the human Nav1.1 limb is a postmortem protein-level
    observation reported within the same paper rather than an independent human
    study. The clinical association is prospective and blinded but small — 33
    patients at a single centre, mean age 62 and so skewed toward young-onset
    disease — and observational, so whether epileptiform activity accelerates
    decline or marks a more aggressive phenotype is unresolved; levetiracetam
    trials have been mixed. Curated separately from the synaptic-failure
    convergence account because it makes the opposite claim about what fails
    first: inhibition rather than excitation.
- name: Tau-Induced Nucleocytoplasmic Transport Failure
  biological_scale: MOLECULAR
  mechanism_confidence: PROVISIONAL
  description: >-
    Pathological tau binds nucleoporins of the nuclear pore complex and disrupts
    their structural and functional integrity, impairing nuclear import and export
    in both tau-transgenic mouse brain and human Alzheimer brain tissue. Nup98
    mislocalizes from the nuclear envelope into the cytoplasm of tangle-bearing
    neurons and, once there, accelerates tau aggregation in vitro — a feed-forward
    arrangement in which the transport lesion promotes the aggregation that caused
    it. The same Nup98 mislocalization is found across primary tauopathies and
    tracks regional tau burden.
  cell_types:
  - preferred_term: Neurons
    term:
      id: CL:0000540
      label: neuron
  biological_processes:
  - preferred_term: Nucleocytoplasmic transport
    modifier: DECREASED
    term:
      id: GO:0006913
      label: nucleocytoplasmic transport
  cellular_components:
  - preferred_term: Nuclear pore complex
    term:
      id: GO:0005643
      label: nuclear pore
  locations:
  - preferred_term: Cerebral Cortex
  - preferred_term: Hippocampus
  genes:
  - preferred_term: MAPT
    term:
      id: hgnc:6893
      label: MAPT
  - preferred_term: NUP98
    term:
      id: hgnc:8068
      label: NUP98
  mechanisms:
  - Direct binding of tau to nuclear pore complex nucleoporins
  - Impairment of nuclear import and export in tangle-bearing neurons
  - Cytoplasmic mislocalization of Nup98 and Nup98-facilitated tau aggregation
  consequences:
  - Neurofibrillary Tangle Formation
  evidence:
  - reference: PMID:30189209
    reference_title: "Tau Protein Disrupts Nucleocytoplasmic Transport in Alzheimer's Disease."
    supports: SUPPORT
    evidence_source: IN_VITRO
    snippet: "Here, we show that tau can directly interact with nucleoporins of the nuclear pore complex (NPC) and affect their structural and functional integrity."
    explanation: >-
      Establishes the direct physical interaction between tau and nuclear pore
      complex components.
  - reference: PMID:30189209
    reference_title: "Tau Protein Disrupts Nucleocytoplasmic Transport in Alzheimer's Disease."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "Pathological tau impairs nuclear import and export in tau-overexpressing transgenic mice and in human AD brain tissue."
    explanation: >-
      Demonstrates the functional transport deficit in human Alzheimer brain tissue
      as well as in the mouse model.
  - reference: PMID:36601621
    reference_title: "Altered localization of nucleoporin 98 in primary tauopathies."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "In the occipital cortex, which is relatively spared from pathological tau accumulations in these primary tauopathies, the localization of nucleoporin 98 was not significantly altered."
    explanation: >-
      Regional internal control in human tissue: the nucleoporin lesion is absent
      where tau pathology is absent, tying it specifically to tau burden rather
      than to neurodegeneration generally.
  downstream:
  - target: Neurofibrillary Tangle Formation
    description: >-
      Nup98 mislocalized into the neuronal cytoplasm facilitates tau aggregation,
      closing a feed-forward loop in which tau-driven transport failure promotes
      further tau aggregation.
    hypothesis_groups:
    - tau_neurodegeneration_model
    causal_link_type: INDIRECT_KNOWN_INTERMEDIATES
    intermediate_mechanisms:
    - Cytoplasmic Nup98 acting as an aggregation-promoting factor.
    evidence:
    - reference: PMID:30189209
      reference_title: "Tau Protein Disrupts Nucleocytoplasmic Transport in Alzheimer's Disease."
      supports: SUPPORT
      evidence_source: IN_VITRO
      snippet: "Furthermore, the nucleoporin Nup98 accumulates in the cell bodies of some tangle-bearing neurons and can facilitate tau aggregation in vitro."
      explanation: >-
        Supports the feed-forward edge from mislocalized Nup98 back onto tau
        aggregation.
  notes: >-
    PROVISIONAL. Unusually for a mechanism at this level of cell-biological
    detail, part of the primary dataset is human Alzheimer brain tissue rather
    than model systems alone, and the tauopathy follow-up is entirely human with a
    genuine regional negative control. What is not established is the causal
    contribution to neurodegeneration: the authors state only that the findings
    raise the possibility that nuclear pore dysfunction contributes to
    tau-induced neurotoxicity, and the mouse arm relies on tau overexpression. The
    two studies share a senior author, so they are not independent laboratories.
- name: Brain N-Glycan Hyperglycosylation
  biological_scale: CELLULAR
  role: amplifier
  mechanism_confidence: PROVISIONAL
  description: >-
    Net abundance of protein N-linked glycans rises in the Alzheimer brain,
    across grey and white matter in human frontal cortex and most markedly in
    cortex, hippocampus and thalamus in amyloid (5xFAD) and tau (PS19) mouse
    models. In human grey matter the increase tracks Braak stage, whereas in
    white matter it appears early and does not progress. Stable-isotope
    pulse-chase tracing in 5xFAD mice attributes it to increased glycan
    biosynthesis rather than impaired lysosomal degradation or salvage, with
    more label entering the nucleotide-sugar donors UDP-hexose and UDP-HexNAc.
    The extra UDP-GlcNAc is channelled selectively into N-glycosylation:
    O-GlcNAcylation and hyaluronan show a mild decrease in the same mouse
    brains, and the
    free N-glycan precursors N-acetylglucosamine and glucosamine-6-phosphate are
    depleted in human AD grey matter, as expected if they are being consumed.
    Lowering the flux improves memory in mice without changing amyloid-beta or
    GFAP.
  cell_types:
  - preferred_term: Neuron
    term:
      id: CL:0000540
      label: neuron
  biological_processes:
  - preferred_term: Protein N-linked glycosylation
    modifier: INCREASED
    term:
      id: GO:0006487
      label: protein N-linked glycosylation
  - preferred_term: UDP-N-acetylglucosamine biosynthesis (hexosamine pathway flux)
    modifier: INCREASED
    term:
      id: GO:0006048
      label: UDP-N-acetylglucosamine biosynthetic process
  - preferred_term: Protein O-GlcNAcylation
    modifier: DECREASED
    term:
      id: GO:7770074
      label: protein O-linked glycosylation via N-acetylglucosamine
  - preferred_term: Hyaluronan biosynthesis
    modifier: DECREASED
    term:
      id: GO:0030213
      label: hyaluronan biosynthetic process
  chemical_entities:
  - preferred_term: N-glycan (high-mannose and bisecting species)
    term:
      id: CHEBI:59520
      label: N-glycan
  locations:
  - preferred_term: Cerebral Cortex
    term:
      id: UBERON:0000956
      label: cerebral cortex
  - preferred_term: Hippocampus
    term:
      id: UBERON:0002421
      label: hippocampal formation
  - preferred_term: Thalamus
    term:
      id: UBERON:0001897
      label: dorsal plus ventral thalamus
  genes:
  - preferred_term: PGM3
    term:
      id: hgnc:8907
      label: PGM3
  mechanisms:
  - Increased de novo N-glycan biosynthesis, not reduced glycan degradation or salvage
  - Selective channelling of UDP-GlcNAc into N-glycosylation at the expense of O-GlcNAcylation and hyaluronan
  - Increased glycosylation of pre-existing glycoproteins rather than glycosylation of new protein substrates
  consequences:
  - Memory Loss
  evidence:
  - reference: PMID:42265388
    reference_title: "Hyperglycosylation is a metabolic driver of Alzheimer's disease."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "However, spatial glycomics analysis uncovered a particularly striking pattern: glycan abundance was dramatically elevated across both white and grey matter regions in AD samples"
    explanation: >-
      Post-mortem human frontal cortex, three AD cases and three age-, sex- and
      post-mortem-interval-matched controls, imaged by MALDI spatial glycomics.
  - reference: PMID:42265388
    reference_title: "Hyperglycosylation is a metabolic driver of Alzheimer's disease."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "In grey matter, a heat map of glycan abundance demonstrated a progressive increase in glycosylation across Braak stages, with the most pronounced changes observed in later stages"
    explanation: >-
      Stage dependence in a second, Braak-stratified human series: the grey
      matter increase grows with pathological severity.
  - reference: PMID:42265388
    reference_title: "Hyperglycosylation is a metabolic driver of Alzheimer's disease."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "suggesting that hyperglycosylation in white matter may be an early but non-progressive event."
    explanation: >-
      Qualifies the regional picture: the white matter increase is transient,
      so the progressive signal is a grey matter one.
  - reference: PMID:42265388
    reference_title: "Hyperglycosylation is a metabolic driver of Alzheimer's disease."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "Additional spatial metabolomics analysis revealed a significant reduction in N-acetylglucosamine and glucosamine-6-phosphate, both critical precursors in N-glycan biosynthesis"
    explanation: >-
      Depletion of free glycan precursors in human AD grey matter, consistent
      with their increased consumption by glycan biosynthesis.
  - reference: PMID:42265388
    reference_title: "Hyperglycosylation is a metabolic driver of Alzheimer's disease."
    supports: SUPPORT
    evidence_source: MODEL_ORGANISM
    snippet: "Notably, N-glycan profiling revealed the same hyperglycosylation phenotype both in the 5xFAD mouse model as well as the PS19 model, similar to those observed in human AD specimens"
    explanation: >-
      The phenotype is reproduced in both an amyloid-driven and a tau-driven
      mouse model, so it is not specific to one pathological driver.
  - reference: PMID:42265388
    reference_title: "Hyperglycosylation is a metabolic driver of Alzheimer's disease."
    supports: SUPPORT
    evidence_source: MODEL_ORGANISM
    snippet: "These data revealed that increased glycan biosynthesis, rather than impaired glycan degradation or salvage, is the predominant driver of hyperglycosylation in AD pathology."
    explanation: >-
      13C-glucose pulse-chase tracing in 5xFAD versus wild-type mice: labelling
      rises in the pulse phase while turnover in the chase phase is unchanged,
      locating the defect in synthesis.
  - reference: PMID:42265388
    reference_title: "Hyperglycosylation is a metabolic driver of Alzheimer's disease."
    supports: SUPPORT
    evidence_source: MODEL_ORGANISM
    snippet: "These findings demonstrate that despite elevated UDP-GlcNAc availability, its utilization is not globally upregulated across all downstream pathways."
    explanation: >-
      The additional UDP-GlcNAc is not used by every downstream branch. On its
      own this does not establish a decrease; the next item states it.
  - reference: PMID:42265388
    reference_title: "Hyperglycosylation is a metabolic driver of Alzheimer's disease."
    supports: SUPPORT
    evidence_source: MODEL_ORGANISM
    snippet: "Interestingly, we examined O-GlcNAcylation and hyaluronan production, both of which draw from the same UDP-GlcNAc pool as N-linked glycosylation, and observed a mild decrease in the brain for both pathways."
    explanation: >-
      Direct support for the DECREASED O-GlcNAcylation and hyaluronan
      annotations, measured in 5xFAD and PS19 mouse brain. The authors call the
      decrease mild.
  - reference: PMID:41042700
    reference_title: "Comprehensive and Site-Specific Characterization of Protein N-Glycosylation in AD Samples Reveals Its Potential Roles in Protein Aggregation and Synaptic Dysfunction."
    supports: REFUTE
    directness: INDIRECT
    evidence_source: HUMAN_CLINICAL
    snippet: "Additionally, downregulated N-glycosylation sites were enriched in synaptic membrane proteins, such as Ca2+ ion channels, GABA-gated chloride channels, and glutamate receptors, implicating glycosylation loss in synaptic dysfunction."
    explanation: >-
      Independent human post-mortem data showing reduced glycosylation at
      specific sites on synaptic proteins. INDIRECT against this node, which
      claims an increase in net released N-glycan abundance; a net increase and
      site-specific losses can coexist, but no study has measured both on the
      same tissue.
  downstream:
  - target: Memory Loss
    description: >-
      Reducing N-glycan biosynthesis restores social recognition memory in
      5xFAD and PS19 mice, and increasing it with oral glucosamine worsens
      memory in 5xFAD mice, without changes in amyloid-beta or astrocytic GFAP.
      The intermediate steps between excess glycosylation and impaired memory
      are not identified; the glycoproteomic data point to neuronal membrane
      proteins but no specific glycoprotein has been shown to carry the effect.
    hypothesis_groups:
    - n_glycan_hyperglycosylation_model
    causal_link_type: INDIRECT_UNKNOWN_INTERMEDIATES
    evidence:
    - reference: PMID:42265388
      reference_title: "Hyperglycosylation is a metabolic driver of Alzheimer's disease."
      supports: SUPPORT
      evidence_source: MODEL_ORGANISM
      snippet: "Notably, shPGM3-treated 5xFAD as well as PS19 mice exhibited significant improvement in social memory interaction patterns when compared with the shScr-treated mice"
      explanation: >-
        Knocking down PGM3, the hexosamine-pathway enzyme upstream of
        UDP-GlcNAc, lowers brain N-glycans and improves memory in both an
        amyloid and a tau model.
    - reference: PMID:42265388
      reference_title: "Hyperglycosylation is a metabolic driver of Alzheimer's disease."
      supports: SUPPORT
      evidence_source: MODEL_ORGANISM
      snippet: "These data reinforce that lowering excessive N-glycosylation is sufficient to enhance cognitive outcomes and highlight hyperglycosylation as an actionable target for dementia-related memory decline in mice."
      explanation: >-
        A mechanistically different intervention, the oligosaccharyltransferase
        inhibitor NGI-1, reaches the same endpoint. The authors describe the
        NGI-1 behavioural rescue as partial.
    - reference: PMID:42265388
      reference_title: "Hyperglycosylation is a metabolic driver of Alzheimer's disease."
      supports: SUPPORT
      evidence_source: MODEL_ORGANISM
      snippet: "By contrast, glucosamine-treated 5xFAD mice exhibited a further exacerbation of social memory deficits, showing no recognition of previously encountered conspecifics across trials"
      explanation: >-
        The opposite perturbation: raising glycan precursor supply worsens the
        same memory readout, so the relationship is bidirectional.
    - reference: PMID:42265388
      reference_title: "Hyperglycosylation is a metabolic driver of Alzheimer's disease."
      supports: SUPPORT
      evidence_source: MODEL_ORGANISM
      snippet: "Notably, our data indicate that neither OST inhibition nor shRNA-mediated inhibition of PGM3 significantly impacted neuroinflammatory markers such as astrocytosis or GFAP levels, nor did they substantially alter amyloid-beta formation in the mouse model."
      explanation: >-
        Places the memory effect outside the amyloid and astrocytic pathways,
        which is why this edge goes directly to the phenotype rather than
        through Amyloid Plaque Formation or Neuroinflammation.
  notes: >-
    PROVISIONAL. Human evidence is post-mortem and small (three cases and three
    controls in the discovery set); causal evidence is mouse-only, female-only,
    short-term, and rests on a single behavioural test. The Neuron cell type is
    inferred from a cell-type enrichment analysis of the identified
    glycopeptides, which the authors describe as a suggestion rather than a
    measurement. PGM3 is bound because it is the enzyme whose knockdown
    reverses the phenotype, not because its expression is shown to be altered.
    What produces the increased flux is unresolved: the phenotype appears in
    both amyloid-driven and tau-driven mice, so no inbound causal edge from
    either pathology is curated here. The direction of the change is contested
    at the level of individual glycosylation sites (PMID:41042700); see the
    attached discussion.
- name: Micro-MKKS63 Microprotein Depletion
  biological_scale: MOLECULAR
  role: trigger
  mechanism_confidence: PROVISIONAL
  description: >-
    The MKKS locus in human dorsolateral prefrontal cortex translates two
    products, and the dominant one is not the annotated one. A 63-amino-acid
    microprotein encoded by a small ORF at the locus - designated micro-MKKS63,
    catalogued only in unreviewed TrEMBL as Q9HB66 - is the predominant
    detectable protein product there, yielding over a thousand peptide-spectrum
    matches across the ROSMAP tandem-mass-tag cohort, while no peptide from the
    canonical 570-amino-acid MKKS protein (Q9NPJ1, the McKusick-Kaufman syndrome
    entry) was recovered at all. The same imbalance holds in a search of a
    public meta-database of 66,700 mass spectrometry runs, where canonical MKKS
    accounts for under 0.1% of the locus's peptide-spectrum matches.
    Micro-MKKS63 abundance is significantly lower in symptomatic Alzheimer
    disease than in non-Alzheimer donors (111 versus 187 donors, log2 fold change
    -0.063, Cohen's d -0.596, FDR 2.42 x 10-5), and RNA reads mapping uniquely to
    its smORF coding sequence anchor a coexpression network enriched for
    mitochondrial biology across all three Gene Ontology domains - whereas the
    network anchored to the canonical ORF shows no enrichment in either
    condition. The depletion is therefore a microprotein-level change at a locus
    whose annotated product carries none of the signal, which is why no previous
    Alzheimer proteomic study could have found it.
  genes:
  - preferred_term: MKKS locus (micro-MKKS63 smORF)
    description: >-
      The locus, not the annotated main ORF. hgnc:7108 is the only HGNC record at
      this locus and UniProt assigns the microprotein to gene MKKS, so it is the
      correct identifier for where micro-MKKS63 is encoded - but the name it
      carries, MKKS, denotes the McKusick-Kaufman syndrome gene whose product is
      the 570-amino-acid protein this node reports as undetectable. The
      preferred_term disambiguates which ORF at the locus is meant.
    term:
      id: hgnc:7108
      label: MKKS
  gene_products:
  - preferred_term: micro-MKKS63 microprotein (UniProt Q9HB66, 63 aa)
    description: >-
      The product this node is about. Left unbound: GeneProductDescriptor binds
      to the NCIT Gene Product hierarchy (NCIT:C26548), which has no term for
      this microprotein, and binding a generic protein term would say less than
      the name does. UniProt carries it only in unreviewed TrEMBL, where it is
      auto-annotated from a 2000-era submission as "McKusick-Kaufman syndrome,
      isoform CRA_a" and "PNAS-117" - an isoform of the canonical protein rather
      than a distinct product, which is the annotation error the source corrects.
  cellular_components:
  - preferred_term: Mitochondrion
    term:
      id: GO:0005739
      label: mitochondrion
  - preferred_term: Mitochondrial outer membrane
    term:
      id: GO:0005741
      label: mitochondrial outer membrane
  locations:
  - preferred_term: Dorsolateral Prefrontal Cortex
  mechanisms:
  - Predominant translation of a 63-amino-acid smORF product at the MKKS locus in place of the annotated 570-amino-acid ORF.
  - Reduced micro-MKKS63 abundance in symptomatic Alzheimer disease cortex, decoupled from the canonical MKKS ORF.
  - Localization of the microprotein to mitochondria, concentrated at peripheral mitochondrial regions and the cytosolic face of the outer membrane.
  consequences:
  - Microglial Oxidative Phosphorylation Failure
  evidence:
  - reference: PMID:42736449
    reference_title: A microprotein atlas of the human frontal cortex in Alzheimer's disease.
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    directness: DIRECT
    quote_role: PRIMARY_RESULT
    snippet: "the predominant detectable protein is a 63-aa smORF-encoded MP cataloged in unreviewed TrEMBL (Q9HB66), which we designated micro-MKKS63."
    explanation: >-
      Identifies the microprotein this node is about and establishes that it,
      rather than the annotated protein, is what the locus predominantly yields
      as protein in human cortex.
  - reference: PMID:42736449
    reference_title: A microprotein atlas of the human frontal cortex in Alzheimer's disease.
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    directness: DIRECT
    quote_role: PRIMARY_RESULT
    snippet: "No peptides from the canonical 570-aa MKKS sequence were detected across the ROSMAP TMT-MS samples, while micro-MKKS63 yielded over 1,000 PSMs"
    explanation: >-
      Quantifies the translational imbalance at the locus. The absence of
      canonical peptides is what makes the microprotein the only thing at this
      locus an Alzheimer proteomic study could have been measuring.
  - reference: PMID:42736449
    reference_title: A microprotein atlas of the human frontal cortex in Alzheimer's disease.
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    directness: DIRECT
    quote_role: PRIMARY_RESULT
    snippet: "TMT quantitation showed significantly decreased micro-MKKS63 abundance in sym-AD versus non-AD"
    explanation: >-
      The depletion itself, measured in human postmortem cortex across the ROSMAP
      diagnostic contrast.
  - reference: PMID:42736449
    reference_title: A microprotein atlas of the human frontal cortex in Alzheimer's disease.
    supports: SUPPORT
    evidence_source: COMPUTATIONAL
    directness: INDIRECT
    quote_role: PRIMARY_RESULT
    snippet: "This network was strongly enriched for mitochondrial biology across all three Gene Ontology (GO) domains"
    explanation: >-
      The link from the microprotein to mitochondrial function, and the reason
      the functional experiments were run at all. Graded COMPUTATIONAL because it
      is coexpression-network inference with Gene Ontology overrepresentation
      testing over human RNA-seq, not a measurement of mitochondrial function;
      INDIRECT because a coexpression signature implies a functional role only
      through an inference step. The same analysis anchored to the canonical ORF
      showed no significant enrichment in either diagnostic group.
  - reference: PMID:42736449
    reference_title: A microprotein atlas of the human frontal cortex in Alzheimer's disease.
    supports: SUPPORT
    evidence_source: IN_VITRO
    directness: DIRECT
    quote_role: PRIMARY_RESULT
    snippet: "confirmed that endogenous micro-MKKS63 ranked among the most highly enriched proteins in that fraction"
    explanation: >-
      Endogenous support for the mitochondrial cellular components bound on this
      node, from a reanalysis of published proximity-labeling proteomics of the
      cytosolic face of the outer mitochondrial membrane. This is the
      localization evidence that does not depend on overexpression.
  downstream:
  - target: Microglial Oxidative Phosphorylation Failure
    description: >-
      Loss of the microprotein lowers microglial oxidative phosphorylation.
      Curated as DIRECT because the knockout removes the microprotein itself and
      the respiratory deficit follows, with no intermediate proposed - but the
      demonstration is in an immortalized human microglial line, not in cortex,
      so the edge is a cell-biological result imported into a human pathograph
      rather than a human observation.
    causal_link_type: DIRECT
    hypothesis_groups:
    - smorf_microprotein_model
    evidence:
    - reference: PMID:42736449
      reference_title: A microprotein atlas of the human frontal cortex in Alzheimer's disease.
      supports: SUPPORT
      evidence_source: IN_VITRO
      directness: DIRECT
      quote_role: PRIMARY_RESULT
      snippet: "Seahorse Mito Stress Test analysis showed significant reductions in basal oxygen consumption rate (OCR), ATP-linked OCR and maximal respiration across KO lines relative to nontargeting controls"
      explanation: >-
        The loss-of-function arm: removing micro-MKKS63 reduces respiration in
        human microglial cells, which is the causal claim this edge makes.
    - reference: PMID:42736449
      reference_title: A microprotein atlas of the human frontal cortex in Alzheimer's disease.
      supports: SUPPORT
      evidence_source: IN_VITRO
      directness: DIRECT
      quote_role: PRIMARY_RESULT
      snippet: "we generated two independent polyclonal CRISPR–Cas9 KO HMC3 lines using single-guide RNAs (sgRNAs) disrupting the micro-MKKS63 smORF while preserving the mORF"
      explanation: >-
        The design that makes the edge attributable to the microprotein rather
        than to the locus: the main ORF is left intact, so the respiratory
        deficit cannot be loss of the 570-amino-acid protein.
  notes: >-
    No cell_types are bound on this node deliberately. The depletion was measured
    in bulk postmortem cortex, which cannot say which cell type lost the
    microprotein; the microglial attribution enters only downstream, through the
    HMC3 knockout, and binding microglia here would smuggle that attribution into
    a human tissue measurement that does not support it. The mitochondrial
    cellular components are bound on the strength of the localization experiments
    rather than the tissue measurement: a C-terminally FLAG-tagged construct
    overexpressed in HMC3 cells concentrates at peripheral mitochondrial regions
    alongside the IMMT marker, and a reanalysis of a published proximity-labeling
    dataset targeting the cytosolic face of the outer mitochondrial membrane
    (PRIDE PXD060912) ranks endogenous micro-MKKS63 among the most enriched
    proteins in that fraction. The second of those is endogenous and is what makes
    the binding more than an overexpression artifact, which is why it rather than
    the immunofluorescence is the evidence item cited above.
- name: Microglial Oxidative Phosphorylation Failure
  biological_scale: CELLULAR
  role: effector
  mechanism_confidence: HYPOTHETICAL
  description: >-
    Human microglia lacking micro-MKKS63 respire less. Two independent polyclonal
    CRISPR-Cas9 knockout HMC3 lines, each using single-guide RNAs that disrupt the
    smORF while leaving the main MKKS ORF intact, showed significant reductions in
    basal oxygen consumption rate, ATP-linked oxygen consumption and maximal
    respiration against non-targeting controls on the Seahorse XF Mito Stress
    Test, with both lines reducing oxygen consumption independently. The node
    records the bioenergetic deficit itself. Whether microglia in the Alzheimer
    brain carry it, and whether it contributes to the disease rather than
    accompanying it, are both open - which is why the node is curated as
    HYPOTHETICAL and left without downstream edges into the established Alzheimer
    chain.
  cell_types:
  - preferred_term: Microglia
    term:
      id: CL:0000129
      label: microglial cell
  biological_processes:
  - preferred_term: Oxidative phosphorylation
    modifier: DECREASED
    term:
      id: GO:0006119
      label: oxidative phosphorylation
  - preferred_term: Mitochondrial ATP synthesis
    modifier: DECREASED
    term:
      id: GO:0042776
      label: proton motive force-driven mitochondrial ATP synthesis
  cellular_components:
  - preferred_term: Mitochondrion
    term:
      id: GO:0005739
      label: mitochondrion
  mechanisms:
  - Loss of a mitochondrially localized microprotein required for normal microglial oxidative phosphorylation.
  evidence:
  - reference: PMID:42736449
    reference_title: A microprotein atlas of the human frontal cortex in Alzheimer's disease.
    supports: SUPPORT
    evidence_source: IN_VITRO
    directness: DIRECT
    quote_role: PRIMARY_RESULT
    snippet: "with consistent effects across two independent sgRNA lines, supporting a role for micro-MKKS63 in microglial oxidative phosphorylation."
    explanation: >-
      Establishes the node and its reproducibility: the respiratory deficit
      appears in two separately derived knockout lines, which is what
      distinguishes it from a single-clone artifact.
  notes: >-
    Deliberately terminal. Microglial bioenergetic failure has obvious candidate
    consequences in this entry - impaired amyloid phagocytosis, a shift toward a
    proinflammatory state, convergence on Neuroinflammation - and none of them is
    tested in the source. Drawing those edges would turn a plausible story into a
    curated causal claim, so the connection is recorded as an open question in
    gap_micro_mkks63_causal_role_in_alzheimer_disease instead. The node also does
    not declare conforms_to against
    mitochondrial_dysfunction#Bioenergetic Decline and Oxidative Stress, although
    it is plainly the same kind of state: that module node pairs decreased
    oxidative phosphorylation with increased reactive oxygen species metabolism,
    and the reactive-oxygen arm was not measured here. Conforming on half the node
    would assert the other half.
phenotypes:
- category: Cognitive
  name: Memory Loss
  frequency: VERY_FREQUENT
  diagnostic: true
  notes: The earliest and most prominent symptom.
  evidence:
  - reference: PMID:31724515
    reference_title: "Assessment of Memory Impairment in Early Diagnosis of Alzheimer's Disease."
    supports: SUPPORT
    snippet: Memory impairment has been considered as one of the earliest clinical hallmarks of Alzheimer's disease.
    explanation: This reference supports the statement that memory loss is a very frequent and early diagnostic symptom of Alzheimer's disease.
  phenotype_term:
    preferred_term: Memory impairment
    term:
      id: HP:0002354
      label: Memory impairment
- category: Cognitive
  name: Executive Dysfunction
  frequency: VERY_FREQUENT
  notes: Difficulty in planning, decision-making, and judgment.
  evidence:
  - reference: PMID:24011643
    reference_title: "Executive functions in clinical and preclinical Alzheimer's disease."
    supports: SUPPORT
    snippet: Impairment of executive functions is common in neurodegenerative disorders such as Alzheimer's disease.
    explanation: The literature states that executive dysfunction, which includes difficulties in planning, decision-making, and judgment, is common in Alzheimer's disease.
  phenotype_term:
    preferred_term: Impaired executive functioning
    term:
      id: HP:0033051
      label: Impaired executive functioning
- category: Psychiatric
  name: Behavioral Changes
  frequency: FREQUENT
  notes: Includes irritability, depression, and apathy.
  evidence:
  - reference: PMID:38157881
    reference_title: "Behavioral or neuropsychiatric symptoms of Alzheimer's disease: from psychopathology to pharmacological management."
    supports: SUPPORT
    snippet: Neuropsychiatric or behavioral symptoms of dementia encompass a series of disorders, such as anxiety, depression, apathy, psychosis, and agitation, all commonly present in individuals living with dementia.
    explanation: The abstract confirms that behavioral changes, including irritability, depression, and apathy, are common in Alzheimer's disease.
  - reference: PMID:28413709
    reference_title: "Awareness, apathy, and depression in Alzheimer's disease and mild cognitive impairment."
    supports: SUPPORT
    snippet: 'Using the median split approach, greater apathy and lower depression were associated with poorer awareness on the Self-Consciousness Scale (respectively: odds ratio ... = 4.8, p = .03; OR = 4.84, p = .04), and the PCRS (only apathy: OR = 9.3, p = .003).'
    explanation: This study indicates that apathy and depression are significant behavioral symptoms in Alzheimer's disease.
  phenotype_term:
    preferred_term: Behavioral abnormality
    term:
      id: HP:0000708
      label: Atypical behavior
- category: Neurologic
  name: Aphasia
  frequency: FREQUENT
  notes: Difficulty with speech and understanding language.
  evidence:
  - reference: PMID:24035593
    reference_title: "Aphasia(s) in Alzheimer."
    supports: SUPPORT
    snippet: Language disorders of degenerative origin are frequently tied to Alzheimer disease (AD) the different variants of which can result in primary and secondary aphasia syndromes.
    explanation: The literature clearly supports that aphasia, which involves difficulty with speech and understanding language, is a frequent neurologic symptom in Alzheimer's Disease.
  phenotype_term:
    preferred_term: Aphasia
    term:
      id: HP:0002381
      label: Aphasia
- category: Neurologic
  name: Apraxia
  frequency: OCCASIONAL
  notes: Difficulty with motor tasks despite intact motor function.
  evidence:
  - reference: PMID:36375032
    reference_title: "The dementia apraxia test can detect early-onset Alzheimer's disease."
    supports: SUPPORT
    snippet: Limb apraxia is a common early sign of Alzheimer's disease (AD) and is thought to occur specifically in early-onset (before the age of 65) AD.
    explanation: The literature indicates that limb apraxia is a common early sign of Alzheimer's disease, supporting the statement that apraxia can occur in Alzheimer's disease.
  phenotype_term:
    preferred_term: Apraxia
    term:
      id: HP:0002186
      label: Apraxia
- category: Neurologic
  name: Agnosia
  phenotype_term:
    preferred_term: Agnosia
    term:
      id: HP:0010524
      label: Disturbed sensory perception
  frequency: OCCASIONAL
  notes: Inability to recognize objects or people.
  evidence:
  - reference: PMID:31449049
    reference_title: "[Visual perceptual disorders in Alzheimer's disease]."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "both low-level and high-level visual perception disorders seem quite common in Alzheimer's disease, including, on a low-level, loss of visual field, decreased acuity and contrast sensitivity, and impaired color vision, and on a high-level, impaired color vision, motion perception, visuospatial deficits, object agnosia, prosopagnosia and impaired recognition of facial emotional expressions."
    explanation: Review of visual-perceptual deficits in Alzheimer's disease reporting that high-level disorders including object agnosia and prosopagnosia are common, directly supporting agnosia as a recognized (occasional) neurologic feature of the disease.
- category: Neurologic
  name: Seizures
  phenotype_term:
    preferred_term: Seizure
    term:
      id: HP:0001250
      label: Seizure
  frequency: OCCASIONAL
  notes: >-
    Usually non-motor and therefore easily missed. Clinically overt epilepsy is
    the visible fraction of a much larger network-hyperexcitability problem:
    extended EEG and MEG monitoring detects subclinical epileptiform activity in
    roughly 42% of patients with no seizure history, against about 10% of
    controls (PMID:27696483). Mechanistically linked to the Interneuron
    Dysfunction and Network Hypersynchrony node.
  evidence:
  - reference: PMID:23835471
    reference_title: "Seizures and epileptiform activity in the early stages of Alzheimer disease."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "Epilepsies were most often complex partial seizures (47%) and more than half were nonconvulsive (55%)."
    explanation: >-
      Characterizes the seizure semiology in Alzheimer disease as predominantly
      complex partial and nonconvulsive, which is why the phenotype is
      under-recognized clinically.
  - reference: PMID:23835471
    reference_title: "Seizures and epileptiform activity in the early stages of Alzheimer disease."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "Patients with AD who had epilepsy presented with cognitive decline 5.5 years earlier than patients with AD who did not have epilepsy (64.8 vs 70.3 years; P = .001)."
    explanation: >-
      Associates the presence of epilepsy with earlier onset of cognitive decline.
      The cohort is retrospective and referral-based, so the interval should not be
      read as a population estimate.
  - reference: PMID:27696483
    reference_title: "Incidence and impact of subclinical epileptiform activity in Alzheimer's disease."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "Extended monitoring detects subclinical epileptiform activity in a substantial proportion of patients with AD."
    explanation: >-
      Establishes that the epileptiform phenotype is far more common than overt
      seizures once monitoring is sensitive enough to detect it.
biochemical:
- name: Amyloid Beta (Aβ42)
  presence: Elevated
  context: Found in cerebrospinal fluid and brain tissue.
  evidence:
  - reference: PMID:19661632
    reference_title: "Cerebrospinal fluid biomarkers for Alzheimer's disease."
    supports: SUPPORT
    snippet: The core candidate CSF biomarkers Abeta42, total tau (T-tau), and phosphorylated tau (P-tau) have been shown to have a high diagnostic performance to identify AD also in the early phase of the disease.
    explanation: The literature confirms that Abeta42 is a core biomarker found in cerebrospinal fluid (CSF) and is elevated in Alzheimer's disease.
  - reference: PMID:31958088
    reference_title: "Aluminum and Amyloid-β in Familial Alzheimer's Disease."
    supports: SUPPORT
    snippet: Genetic predispositions associated with metabolism of the amyloid-beta protein precursor underlie familial Alzheimer's disease; a form of dementia characterized by early disease onset and elevated levels of cortical amyloid-beta.
    explanation: The literature confirms that elevated levels of amyloid-beta, including Abeta42, are found in the brain tissue of individuals with Alzheimer's disease.
  - reference: PMID:38987603
    reference_title: "CryoET of β-amyloid and tau within postmortem Alzheimer's disease brain."
    supports: SUPPORT
    snippet: beta-amyloid plaques contained a mixture of fibrils, some of which were branched, and protofilaments, arranged in parallel arrays and lattice-like structures.
    explanation: This reference supports the presence of beta-amyloid in brain tissue, which includes Abeta42, in Alzheimer's disease.
  - reference: PMID:33080124
    reference_title: "Amyloid-β PET and CSF in an autopsy-confirmed cohort."
    supports: SUPPORT
    snippet: Accumulation of amyloid-β is among the earliest changes in Alzheimer's disease (AD). Amyloid-β positron emission tomography (PET) and Aβ42 in cerebrospinal fluid (CSF) both assess amyloid-β pathology in-vivo
    explanation: This literature supports the presence and elevation of Abeta42 in both cerebrospinal fluid and brain tissue in Alzheimer's disease.
- name: Astrocyte Plasma Proteomic Age Gap
  biomarker_term:
    preferred_term: astrocyte plasma proteomic aging clock (age gap)
    term:
      id: NCIT:C97139
      label: Proteomic Profile
  presence: Elevated
  context: >-
    Blood-based cell-type-specific aging clock. Plasma proteins are mapped to
    their putative cell of origin using Human Protein Atlas single-cell
    transcriptomic data, and a machine-learning model estimates astrocyte
    biological age; the age gap is the difference between that estimate and
    chronological age. "Extreme" agers are those in the upper tail of the age-gap
    distribution and "youthful" agers the lower tail.
  cell_types:
  - preferred_term: Astrocyte
    term:
      id: CL:0000127
      label: astrocyte
  notes: >-
    Astrocyte age gap was the strongest cellular-aging predictor of incident
    Alzheimer disease in UK Biobank over 15 years of follow-up, and stratified
    risk within every APOE genotype stratum, including APOE4 homozygotes. Note
    that the headline 12.59-fold figure is an extreme-versus-youthful contrast;
    in the separate comparison against established risk factors the excess risk
    of extreme astrocyte aging was HR 5.16 (95% CI 4.06-6.56), comparable to
    APOE4 carrier status (HR 5.30) and exceeding the AD polygenic risk score
    (HR 2.14). This is a prognostic association from observational cohorts, not
    a demonstration that astrocyte aging causes Alzheimer disease.
  evidence:
  - reference: PMID:42297981
    reference_title: "Plasma proteomic signatures of cellular aging predict human disease."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: Individuals with extreme astrocyte aging demonstrated a 12.59-fold increased risk of incident AD compared to those with youthful aging.
    explanation: >-
      Quantifies the prognostic association between the astrocyte plasma
      proteomic age gap and incident Alzheimer disease over 15 years of UK
      Biobank follow-up.
  - reference: PMID:42297981
    reference_title: "Plasma proteomic signatures of cellular aging predict human disease."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: Individuals who were homozygous for APOE4 and had extreme astrocyte aging showed the highest cumulative incidence of 38.3% over 15 years of follow-up, compared to 12.6% for homozygotes with normal astrocyte aging.
    explanation: >-
      Shows that astrocyte aging status stratifies absolute AD incidence within
      the highest-risk APOE genotype, supporting use of the age gap as a risk
      modifier on top of genotype.
  - reference: PMID:42297981
    reference_title: "Plasma proteomic signatures of cellular aging predict human disease."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: extreme astrocyte aging consistently identified individuals at elevated AD risk compared to those with normal and youthful astrocyte aging, demonstrating that astrocyte aging provides independent risk stratification beyond APOE genotype
    explanation: >-
      States the authors' conclusion that the astrocyte age gap adds risk
      information independent of APOE genotype.
  - reference: PMID:42297981
    reference_title: "Plasma proteomic signatures of cellular aging predict human disease."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: Possessing both APOE4 and astrocyte extreme aging conferred a greater increase in AD risk for women
    explanation: >-
      Records the sex difference in the APOE4-by-astrocyte-aging interaction
      (HR 14.23, 95% CI 9.86-20.54 in women versus 10.95 in men), relevant to how
      the biomarker would be applied in stratification.
  readouts:
  - target: Neuroinflammation
    relationship: PREDICTS
    direction: POSITIVE
    endpoint_context: PROGNOSTIC
    interpretation: >-
      An elevated astrocyte age gap reports an aged astrocyte proteomic state and
      predicts incident Alzheimer disease years before diagnosis. It is linked to
      the astrocytic arm of the neuroinflammation node, but the study measures
      plasma protein signatures rather than brain astrocyte reactivity directly.
    evidence:
    - reference: PMID:42297981
      reference_title: "Plasma proteomic signatures of cellular aging predict human disease."
      supports: SUPPORT
      evidence_source: HUMAN_CLINICAL
      snippet: Individuals with extreme astrocyte aging demonstrated a 12.59-fold increased risk of incident AD compared to those with youthful aging.
      explanation: >-
        Supports treating the astrocyte age gap as a prognostic readout that
        anticipates clinical Alzheimer disease.
  - target: Neurofibrillary Tangle Formation
    relationship: CORRELATES_WITH
    direction: POSITIVE
    endpoint_context: DIAGNOSTIC
    interpretation: >-
      In the NSHD Insight-46 substudy the astrocyte age gap tracked plasma
      pTau-217 burden, an orthogonal blood measure of amyloid and tau pathology,
      linking the clock to tau pathology rather than to cognition alone.
    evidence:
    - reference: PMID:42297981
      reference_title: "Plasma proteomic signatures of cellular aging predict human disease."
      supports: SUPPORT
      evidence_source: HUMAN_CLINICAL
      snippet: we found significant associations between plasma phosphorylated tau-217 (pTau-217) burden and several neuronal and glial cell types, including astrocytes
      explanation: >-
        Provides the orthogonal pathology correlate for the astrocyte age gap,
        supporting a correlational readout against the tau pathology node.
- name: Oligodendrocyte Precursor Cell Plasma Proteomic Age Gap
  biomarker_term:
    preferred_term: oligodendrocyte precursor cell plasma proteomic aging clock (age gap)
    term:
      id: NCIT:C97139
      label: Proteomic Profile
  presence: Elevated
  context: >-
    Blood-based cell-type-specific aging clock (see Astrocyte Plasma Proteomic
    Age Gap for the derivation). Oligodendrocyte precursor cells were the most
    prominent cross-sectional association with clinical Alzheimer disease in the
    Global Neurodegeneration Proteomics Consortium cohort, though pancreatic
    endocrine cells and inhibitory neurons showed a comparable effect size.
  cell_types:
  - preferred_term: Oligodendrocyte precursor cell
    term:
      id: CL:0002453
      label: oligodendrocyte precursor cell
  notes: >-
    Unlike the astrocyte clock, whose value in this study is prognostic for
    incident disease, the oligodendrocyte precursor cell age gap behaves as a
    severity-tracking marker: it rises stepwise with Clinical Dementia Rating
    score and is among the strongest correlates of a worse Preclinical Alzheimer
    Cognitive Composite.
  evidence:
  - reference: PMID:42297981
    reference_title: "Plasma proteomic signatures of cellular aging predict human disease."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: AD was associated with accelerated aging across a wide range of cell types, most prominently oligodendrocyte precursor cells
    explanation: >-
      Establishes the cross-sectional case/control association in the GNPC cohort
      and the relative prominence of the oligodendrocyte precursor cell signal.
  - reference: PMID:42297981
    reference_title: "Plasma proteomic signatures of cellular aging predict human disease."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: oligodendrocyte precursor cell aging showed a consistent stepwise increase with worsening cognitive impairment
    explanation: >-
      Establishes the graded relationship between the oligodendrocyte precursor
      cell age gap and Clinical Dementia Rating stage.
  - reference: PMID:42297981
    reference_title: "Plasma proteomic signatures of cellular aging predict human disease."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: we observed an association between worsened cognition and elevated age gaps across a variety of cell types, with oligodendrocyte precursor cells demonstrating one of the strongest associations
    explanation: >-
      Independent cohort (NSHD Insight-46) association with the Preclinical
      Alzheimer Cognitive Composite, supporting the marker as a correlate of
      early cognitive decline.
  readouts:
  - target: Neurofibrillary Tangle Formation
    relationship: CORRELATES_WITH
    direction: POSITIVE
    endpoint_context: DIAGNOSTIC
    interpretation: >-
      The oligodendrocyte precursor cell age gap tracked plasma pTau-217 burden
      in the Insight-46 substudy, tying this glial clock to measured tau
      pathology as well as to cognitive stage.
    evidence:
    - reference: PMID:42297981
      reference_title: "Plasma proteomic signatures of cellular aging predict human disease."
      supports: SUPPORT
      evidence_source: HUMAN_CLINICAL
      snippet: we found significant associations between plasma phosphorylated tau-217 (pTau-217) burden and several neuronal and glial cell types, including astrocytes (coefficient beta (ꞵ) = 1.08, adjusted P = 3.92 × 10−7), oligodendrocyte precursor cells (ꞵ = 1.02, adjusted P = 1.19 × 10−6)
      explanation: >-
        Names oligodendrocyte precursor cells explicitly, with a pTau-217
        regression coefficient comparable to that of astrocytes, in the NSHD
        Insight-46 substudy.
- name: Inhibitory Neuron Plasma Proteomic Age Gap
  biomarker_term:
    preferred_term: inhibitory neuron plasma proteomic aging clock (age gap)
    term:
      id: NCIT:C97139
      label: Proteomic Profile
  presence: Elevated
  context: >-
    Blood-based cell-type-specific aging clock (see Astrocyte Plasma Proteomic
    Age Gap for the derivation). The finding is notable for its selectivity:
    inhibitory neuron aging associated with Alzheimer disease while excitatory
    neuron aging did not.
  cell_types:
  - preferred_term: Inhibitory (GABAergic) neuron
    term:
      id: CL:0000617
      label: GABAergic neuron
  notes: >-
    The inhibitory-but-not-excitatory pattern is concordant with the selective
    interneuron vulnerability and excitation/inhibition imbalance literature in
    Alzheimer disease, though this study measures a plasma proteomic signature
    and does not itself demonstrate interneuron loss.
  evidence:
  - reference: PMID:42297981
    reference_title: "Plasma proteomic signatures of cellular aging predict human disease."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: while aging of inhibitory neurons was linked to AD, aging of excitatory neurons was not
    explanation: >-
      Establishes the cell-type selectivity that distinguishes this marker from a
      generic neuronal aging signal.
  - reference: PMID:42297981
    reference_title: "Plasma proteomic signatures of cellular aging predict human disease."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: inhibitory neuron age gaps increased with higher CDR scores across cohorts
    explanation: >-
      Shows the inhibitory neuron age gap tracks dementia severity stage, not
      only case/control status.
  readouts:
  - target: Synaptic Dysfunction
    relationship: CORRELATES_WITH
    direction: POSITIVE
    endpoint_context: DIAGNOSTIC
    interpretation: >-
      An elevated inhibitory neuron age gap, in the absence of a matching
      excitatory neuron signal, is read as a plasma correlate of the selective
      interneuron vulnerability that contributes to excitation/inhibition
      imbalance and synaptic dysfunction in Alzheimer disease.
    evidence:
    - reference: PMID:42297981
      reference_title: "Plasma proteomic signatures of cellular aging predict human disease."
      supports: SUPPORT
      evidence_source: HUMAN_CLINICAL
      snippet: while aging of inhibitory neurons was linked to AD, aging of excitatory neurons was not
      explanation: >-
        The selectivity of the association is what licenses reading this marker
        against the interneuron/synaptic arm rather than neurodegeneration
        generally.
genetic:
- name: APP
  gene_term:
    preferred_term: APP
    term:
      id: hgnc:620
      label: APP
  association: Genetic Mutation
  subtype: Early-Onset Alzheimer's Disease
  evidence:
  - reference: PMID:1365885
    reference_title: "Alzheimer's disease untangled."
    supports: SUPPORT
    snippet: Four mutations involving amino acid substitutions in exons 16 and 17 of the amyloid precursor protein (APP) gene, have been identified which co-segregate with the disease in some families multiply affected by early onset Alzheimer's disease.
    explanation: This reference supports the association of APP genetic mutations with early-onset Alzheimer's disease.
  - reference: PMID:33789815
    reference_title: "Low-degree trisomy 21 mosaicism promotes early-onset Alzheimer disease."
    supports: SUPPORT
    snippet: This is the first case demonstrating that a low-degree APP gene-dose increase suffices to cause EOAD with prominent amyloid-beta/tau pathology.
    explanation: This reference supports the role of APP gene mutations in causing early-onset Alzheimer's disease.
  - reference: PMID:36306459
    reference_title: "Screening for Genetic Mutations Associated with Early-Onset Alzheimer's Disease in Han Chinese."
    supports: SUPPORT
    snippet: Numerous mutations in amyloid precursor protein (APP) and presenilin 1 and 2 (PSEN1 and PSEN2) have been identified for EOAD, but they can only account for a small proportion of EOAD cases.
    explanation: This reference supports the association of APP genetic mutations with early-onset Alzheimer's disease.
  - reference: PMID:31623876
    reference_title: "Novel amyloid precursor protein mutation, Val669Leu (\"Seoul APP\"), in a Korean patient with early-onset Alzheimer's disease."
    supports: SUPPORT
    snippet: In this study, a novel mutation in APP gene, Val669Leu ('Seoul APP'), was reported in a Korean female patient with Alzheimer's disease.
    explanation: This reference supports the association of APP genetic mutations with early-onset Alzheimer's disease.
  - reference: PMID:26243569
    reference_title: "A genetic cause of Alzheimer disease: mechanistic insights from Down syndrome."
    supports: SUPPORT
    snippet: It is thought that this risk is conferred by the presence of three copies of the gene encoding amyloid precursor protein (APP)--an Alzheimer disease risk factor.
    explanation: This reference supports the association of APP genetic mutations with early-onset Alzheimer's disease.
  - reference: PMID:22727994
    reference_title: "Good gene, bad gene: new APP variant may be both."
    supports: SUPPORT
    snippet: APP mutations cause Alzheimer disease (AD) with virtually complete penetrance.
    explanation: This reference supports the association of APP genetic mutations with early-onset Alzheimer's disease.
- name: PSEN1
  gene_term:
    preferred_term: PSEN1
    term:
      id: hgnc:9508
      label: PSEN1
  association: Genetic Mutation
  subtype: Early-Onset Alzheimer's Disease
  evidence:
  - reference: PMID:31296348
    reference_title: "A patient with early-onset Alzheimer's disease with a novel PSEN1 p.Leu424Pro mutation."
    supports: SUPPORT
    snippet: Presenilin 1 (PSEN1) gene mutations are the major known genetic cause of early-onset Alzheimer's disease.
    explanation: The study reports a novel PSEN1 mutation associated with early-onset Alzheimer's disease, supporting the genetic association.
  - reference: PMID:35487021
    reference_title: "A review ofimaging genetics in Alzheimer's disease."
    supports: SUPPORT
    snippet: At present, three early-onset AD genes (APP, PSEN1, PSEN2) and one late-onset AD susceptibility gene apolipoprotein E (APOE) have been determined.
    explanation: The reference confirms that PSEN1 is one of the genes associated with early-onset Alzheimer's disease.
  - reference: PMID:36951251
    reference_title: "Genetic associations with age at dementia onset in the PSEN1 E280A Colombian kindred."
    supports: SUPPORT
    snippet: A genetic association study was conducted to examine ADAD AAO in 340 individuals with the PSEN1 E280A mutation.
    explanation: The study discusses a large cohort with a specific PSEN1 mutation associated with early-onset Alzheimer's disease, reinforcing the genetic link.
  - reference: PMID:36641620
    reference_title: "Embryo Selection for a Carrier of an Early-Onset Alzheimer's Disease-Associated Mutation in the PSEN1 Gene."
    supports: SUPPORT
    snippet: We present two asymptomatic siblings, a 30-year-old male and a 34- year-old female, who are heterozygous carriers of a pathogenic variant c.428T>C (Ile143Thr) in the presenilin 1 (PSEN1) gene.
    explanation: Reports carriers of PSEN1 Ile143Thr, a variant the authors classify as pathogenic for early-onset Alzheimer disease, in a family referred for reproductive genetic counseling.
  - reference: PMID:35430993
    reference_title: "Clinical and Molecular Findings in a Turkish Family Who Had a (c.869- 1G>A) Splicing Variant in PSEN1 Gene with A Rare Condition: The Variant Alzheimer's Disease with Spastic Paraparesis."
    supports: SUPPORT
    snippet: We identified a heterozygous splicing variant (c.869-1G>A) in the PSEN1 gene, in a family with two affected individuals who present with varAD.
    explanation: A PSEN1 splice-site variant segregating in a family with early-onset Alzheimer disease with spastic paraparesis, a PSEN1-associated clinical variant of familial early-onset disease.
  - reference: PMID:30549411
    reference_title: "Gene mutations in a Han Chinese Alzheimer's disease cohort."
    supports: SUPPORT
    snippet: The gene mutations in the amyloid precursor protein (APP), presenilin 1 (PSEN1), and presenilin 2 (PSEN2) are the frequent causes of AD.
    explanation: The reference confirms that PSEN1 mutations are a frequent cause of early-onset Alzheimer's disease.
- name: PSEN2
  gene_term:
    preferred_term: PSEN2
    term:
      id: hgnc:9509
      label: PSEN2
  association: Genetic Mutation
  subtype: Early-Onset Alzheimer's Disease
  evidence:
  - reference: PMID:35491795
    reference_title: "PSEN2 Mutation Spectrum and Novel Functionally Validated Mutations in Alzheimer's Disease: Data from PUMCH Dementia Cohort."
    supports: SUPPORT
    snippet: The established causative mutations in the APP, PSEN1, and PSEN2 can explain less than 1%, Alzheimer's disease (AD) patients. Of the identified variants, the PSEN2 mutations are even less common.
    explanation: The reference confirms that PSEN2 mutations are among the causative factors for early-onset Alzheimer's disease.
  - reference: PMID:36701017
    reference_title: "PSEN2 and ABCA7 variants causing early-onset preclinical pathological changes in Alzheimer's disease: a case report and literature review."
    supports: SUPPORT
    snippet: Early-onset AD (EOAD) was defined as AD occurring before age 65. Although it has a high genetic risk, EOAD due to PSEN2 variation is very rare.
    explanation: The reference supports the association of PSEN2 with early-onset Alzheimer's disease, although it notes that such cases are rare.
  - reference: PMID:32741831
    reference_title: "Early-Onset Familial Alzheimer Disease Variant PSEN2 N141I Heterozygosity is Associated with Altered Microglia Phenotype."
    supports: SUPPORT
    snippet: Early-onset familial Alzheimer disease (EOFAD) is caused by heterozygous variants in the presenilin 1 (PSEN1), presenilin 2 (PSEN2), and APP genes.
    explanation: The reference explicitly states that early-onset familial Alzheimer's disease can be caused by PSEN2 mutations.
- name: APOE
  gene_term:
    preferred_term: APOE
    term:
      id: hgnc:613
      label: APOE
  association: Risk Factor
  subtype: Late-Onset Alzheimer's Disease
  notes: >-
    APOE is the strongest common-variant genetic risk factor for late-onset
    Alzheimer's disease. The APOE4 allele increases risk and accelerates onset in
    a gene-dose-dependent manner (relative to the common APOE3), while APOE2 is
    protective. Beyond amyloid, APOE4 exacerbates tau-mediated neurodegeneration,
    neuroinflammation, glial lipid/cholesterol accumulation, and dysfunction of
    the blood-brain barrier. Rare protective coding variants illuminate mechanism
    and
    are therapeutic templates: APOE3 Christchurch (R136S) reduces receptor-binding
    avidity and limits tau seeding/spreading (a homozygous carrier resisted
    autosomal-dominant AD for ~3 decades despite heavy amyloid), and APOE3
    Jacksonville (V236E) reduces APOE self-aggregation and amyloid pathology.
    Preclinically, lowering APOE4 (e.g., with antisense oligonucleotides) reduces
    tau pathology, neurodegeneration, and neuroinflammation, motivating
    APOE-directed therapeutics. See the apoe_risk_modulation_model mechanistic
    hypothesis.
  evidence:
  - reference: PMID:28959956
    reference_title: "ApoE4 markedly exacerbates tau-mediated neurodegeneration in a mouse model of tauopathy."
    supports: SUPPORT
    evidence_source: MODEL_ORGANISM
    snippet: "APOE4 is the strongest genetic risk factor for late-onset Alzheimer disease."
    explanation: Establishes APOE4 as the leading genetic risk factor for late-onset AD.
  - reference: PMID:31686034
    reference_title: "Resistance to autosomal dominant Alzheimer's disease in an APOE3 Christchurch homozygote: a case report."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "The individual had two copies of the APOE3 Christchurch (R136S) mutation, unusually high brain amyloid levels and limited tau and neurodegenerative measurements."
    explanation: >-
      The homozygous APOE3 Christchurch (R136S) variant is associated with
      resistance to autosomal-dominant AD despite high amyloid, evidence for a
      protective APOE variant acting on tau/neurodegeneration.
  - reference: PMID:34586832
    reference_title: "APOE3-Jacksonville (V236E) variant reduces self-aggregation and risk of dementia."
    supports: SUPPORT
    evidence_source: MODEL_ORGANISM
    snippet: "Mice expressing APOE3-Jac have reduced amyloid pathology, plaque-associated immune responses, and neuritic dystrophy."
    explanation: >-
      The APOE3-Jacksonville (V236E) variant reduces APOE aggregation and amyloid
      pathology, a second protective APOE variant supporting APOE aggregation as a
      therapeutic target.
  - reference: PMID:33550655
    reference_title: "Apolipoprotein E4 Reduction with Antisense Oligonucleotides Decreases Neurodegeneration in a Tauopathy Model."
    supports: SUPPORT
    evidence_source: MODEL_ORGANISM
    snippet: "We conclude that reducing ApoE4 levels should be explored further as a therapeutic approach for APOE4 carriers with tauopathy including Alzheimer's disease."
    explanation: >-
      APOE4-lowering antisense oligonucleotides reduce tau pathology,
      neurodegeneration, and neuroinflammation, supporting APOE as an actionable
      therapeutic node.
- name: BIN1
  gene_term:
    preferred_term: BIN1
    term:
      id: hgnc:1052
      label: BIN1
  association: Risk Factor
  subtype: Late-Onset Alzheimer's Disease
  notes: >-
    BIN1 encodes bridging integrator 1, a phosphoinositide-binding adaptor protein involved in membrane dynamics, endocytosis, and synaptic function. Identified as the strongest broad AD-risk anchor across blood and brain-region eQTL layers (whole blood, cortex, hippocampus, cerebellum). The locus shows distinct causal variants (H3-dominated colocalization) across tissues, indicating region-specific regulatory mechanisms rather than a single shared causal variant. Disease-state brain-expression support shows reduced BIN1 expression in Alzheimer disease versus control tissue.
  evidence:
  - reference: PPR:PPR1263744
    reference_title: "Brain-region-aware genetic prioritization separates Alzheimer disease risk from APOE-sensitive β-amyloid burden in public genetic and expression quantitative trait locus resources"
    supports: SUPPORT
    evidence_source: COMPUTATIONAL
    snippet: "eQTLGen tested 16,875 genes for AD risk and identiﬁed 204 FDR-signiﬁcant genes, led by BIN1."
    explanation: eQTLGen blood eQTL analysis identified BIN1 as the lead AD-risk gene among 204 FDR-significant hits, establishing it as a primary genetic risk factor distinct from amyloid-burden genetics.
  - reference: PPR:PPR1263744
    reference_title: "Brain-region-aware genetic prioritization separates Alzheimer disease risk from APOE-sensitive β-amyloid burden in public genetic and expression quantitative trait locus resources"
    supports: SUPPORT
    evidence_source: COMPUTATIONAL
    snippet: "CEACAM16-AS1 led cortex, frontal cortex BA9, anterior cingulate BA24, and hippocampus, whereas BIN1 led cerebellum"
    explanation: Multi-region eQTL analysis (cortex, frontal cortex BA9, anterior cingulate BA24, hippocampus, cerebellum) confirms BIN1 as the lead gene in cerebellum, establishing brain-region-specific eQTL support.
  - reference: PPR:PPR1263744
    reference_title: "Brain-region-aware genetic prioritization separates Alzheimer disease risk from APOE-sensitive β-amyloid burden in public genetic and expression quantitative trait locus resources"
    supports: SUPPORT
    evidence_source: COMPUTATIONAL
    snippet: "BIN1 is the strongest broad AD-risk anchor and has external disease-state expression support."
    explanation: Integrated analysis across eQTL, colocalization, MetaBrain, and disease-state expression layers converges on BIN1 as the primary non-APOE AD-risk architecture anchor.
  - reference: PPR:PPR1263744
    reference_title: "Brain-region-aware genetic prioritization separates Alzheimer disease risk from APOE-sensitive β-amyloid burden in public genetic and expression quantitative trait locus resources"
    supports: SUPPORT
    evidence_source: COMPUTATIONAL
    snippet: "BIN1 was testable in 10 region-dataset pairs across GSE5281 and GSE48350 and showed four nominally signiﬁcant AD–control differences, all of which survived FDR correction."
    explanation: Disease-state expression analysis supports altered BIN1 expression in Alzheimer disease post-mortem brain tissue, validating the genetic risk association with disease-relevant molecular change.
  - reference: PMID:24162737
    reference_title: "Meta-analysis of 74,046 individuals identifies 11 new susceptibility loci for Alzheimer's disease."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "the already known, GWAS-defined genes (ABCA7, BIN1, CD33, CLU, CR1, CD2AP, EPHA1"
    explanation: The IGAP genome-wide association meta-analysis (74,046 individuals) lists BIN1 among the established GWAS-defined late-onset Alzheimer disease susceptibility genes, providing peer-reviewed foundational support for the BIN1 risk association beyond the eQTL preprint.
- name: TREM2
  gene_term:
    preferred_term: TREM2
    term:
      id: hgnc:17761
      label: TREM2
  association: Risk Factor
  subtype: Late-Onset Alzheimer's Disease
  notes: Microglial receptor that influences disease-associated microglial phenotypes, phagocytosis, and immune responses to Aβ pathology.
- name: NLRP3
  gene_term:
    preferred_term: NLRP3
    term:
      id: hgnc:16400
      label: NLRP3
  association: Genetic Variant
  notes: Encodes the NLRP3 inflammasome sensor in microglia that drives IL-1β production, pyroptosis, and neuroinflammation in response to Aβ and tau aggregates.
- name: PYCARD
  gene_term:
    preferred_term: PYCARD
    term:
      id: hgnc:16608
      label: PYCARD
  association: Genetic Variant
  notes: Encodes ASC (apoptosis-associated speck-like protein containing a CARD), the inflammasome adaptor required for NLRP3 signaling and IL-1β maturation.
- name: LRP1
  gene_term:
    preferred_term: LRP1
    term:
      id: hgnc:6692
      label: LRP1
  association: Genetic Variant
  notes: Low-density lipoprotein receptor-related protein 1, mediates Aβ clearance across the blood-brain barrier. Variants and reduced expression impair Aβ efflux.
  evidence:
  - reference: PMID:19098903
    reference_title: SRF and myocardin regulate LRP-mediated amyloid-beta clearance in brain vascular cells.
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: Western blot analysis of small cerebral leptomeningeal arterial blood vessel homogenates from AD individuals with CAA showed elevated levels of SRF and MYOCD, Aβ accumulation and significantly lower levels of LRP, compared with age-matched, non-demented controls without CAA
    explanation: LRP1 protein is reduced in post-mortem cerebral arterial vessels from AD patients with CAA, alongside vessel-wall amyloid-beta accumulation, supporting the reduced-expression part of this record.
computational_models:
- name: Digital Alzheimer's Disease Diagnosis (DADD) patient-specific brain model
  description: >-
    A 76-region cortical neural-mass model in The Virtual Brain whose synaptic
    degeneration, structural-connectivity degeneration, and compensatory
    rewiring parameters are personalized by inversion against task EEG. The
    inferred parameters serve as interpretable digital biomarkers for early
    Alzheimer disease risk and progression.
  model_type: DIGITAL_TWIN
  repository_url: https://github.com/LoreAma/Code_AD_simulations
  model_id: DADD_Code_AD_simulations
  model_software: Python 3.10; The Virtual Brain; NumPy; SciPy; scikit-learn
  model_format: Jupyter notebooks and NumPy arrays
  publication: PMID:40450374
  modeled_mechanisms:
  - target: Cortical Structural Connectivity Degeneration
    relationship: MEASURES
    fidelity: MODERATE
    description: >-
      Model inversion estimates patient-specific connectivity degeneration and
      neuroplastic rewiring from the similarity between simulated and observed
      EEG signals.
    limitations: >-
      Personalization was evaluated in a single-center cohort enriched for
      subjective cognitive decline; the inferred parameters are model-dependent
      latent biomarkers rather than direct measurements of white-matter loss,
      and external multicenter validation was not reported.
    readouts:
    - name: Connectivity degeneration parameter
      target: Cortical Structural Connectivity Degeneration
      direction: INCREASED
      interpretation: >-
        The inferred cp parameter was higher in the subjective-cognitive-decline
        group than in controls, but it remains an indirect, model-derived measure.
      evidence:
      - reference: PMID:40450374
        reference_title: "Digital twins and non-invasive recordings enable early diagnosis of Alzheimer's disease."
        supports: SUPPORT
        evidence_source: COMPUTATIONAL
        snippet: "Statistical analysis (Fig. 2D) confirmed significant differences across groups in connectivity degeneration (U = 1078.5, p = 0.029, d = −0.6) while differences in synaptic degeneration were non-statistically significant (U = 880.0, p = 0.60)."
        explanation: Reports the group difference for the model-inferred connectivity-degeneration parameter while preserving the negative synaptic-degeneration result.
    - name: Neuroplastic connectivity rewiring parameter
      target: Cortical Structural Connectivity Degeneration
      direction: DECREASED
      interpretation: >-
        The inferred np parameter was lower in the subjective-cognitive-decline
        group than in controls, consistent with reduced compensatory rewiring;
        it remains an indirect model-derived quantity.
      evidence:
      - reference: PMID:40450374
        reference_title: "Digital twins and non-invasive recordings enable early diagnosis of Alzheimer's disease."
        supports: SUPPORT
        evidence_source: COMPUTATIONAL
        snippet: "Additionally, neuroplastic rewiring levels were significantly higher in CTR subjects (U = 474.0, p = 0.004, d = 0.8)."
        explanation: Reports the largest standardized group difference among the personalized DADD parameters and establishes lower rewiring in the SCD group.
    evidence:
    - reference: PMID:40450374
      reference_title: "Digital twins and non-invasive recordings enable early diagnosis of Alzheimer's disease."
      supports: SUPPORT
      evidence_source: COMPUTATIONAL
      snippet: "Personalized digital biomarkers reflecting individual neurodegeneration levels were reconstructed through model inversion by maximizing the similarity between experimental and simulated signals."
      explanation: Establishes how the patient-specific connectivity and degeneration parameters were inferred.
  - target: Synaptic Dysfunction
    relationship: MEASURES
    fidelity: LOW
    description: The personalized lp parameter represents synaptic degeneration through modeled cortical hypo-inhibition and hyperexcitation.
    limitations: >-
      The lp parameter was not significantly different between the
      subjective-cognitive-decline and control groups, and a neural-mass parameter cannot
      distinguish synapse loss from other causes of altered excitation-inhibition balance.
    readouts:
    - name: Synaptic degeneration parameter
      target: Synaptic Dysfunction
      direction: UNCHANGED
      interpretation: Patient-specific lp inferred by EEG-constrained model inversion; no consistent group-level direction was established.
      evidence:
      - reference: PMID:40450374
        reference_title: "Digital twins and non-invasive recordings enable early diagnosis of Alzheimer's disease."
        supports: SUPPORT
        evidence_source: COMPUTATIONAL
        snippet: "Simulated neurodegeneration is regulated by parameters cp and lp, governing connectivity and synaptic degeneration, respectively."
        explanation: Defines lp as the model's synaptic-degeneration readout; the null group comparison is encoded by direction and interpretation.
    evidence:
    - reference: PMID:40450374
      reference_title: "Digital twins and non-invasive recordings enable early diagnosis of Alzheimer's disease."
      supports: SUPPORT
      evidence_source: COMPUTATIONAL
      snippet: "Biophysically, lp reflects cortical hypo-inhibition and hyperexcitation"
      explanation: Defines the physiological interpretation of the modeled synaptic-degeneration parameter.
  variables:
  - name: Connectivity degeneration
    dataset_identifier: cp
    description: Severity of large-scale structural-connectivity loss, ranging from 0 to 1.
    unit: dimensionless
  - name: Synaptic degeneration
    dataset_identifier: lp
    description: Severity of local synaptic degeneration and cortical excitation-inhibition alteration, ranging from 0 to 1.
    unit: dimensionless
  - name: Neuroplastic connectivity rewiring
    dataset_identifier: np
    description: Compensatory short-range connectivity rewiring parameter.
    unit: dimensionless
  findings:
  - statement: DADD digital biomarkers distinguished subjective cognitive decline from healthy controls with seven percentage points greater accuracy than standard EEG biomarkers.
    evidence:
    - reference: PMID:40450374
      reference_title: "Digital twins and non-invasive recordings enable early diagnosis of Alzheimer's disease."
      supports: SUPPORT
      evidence_source: COMPUTATIONAL
      snippet: "Robustly distinguished SCD from healthy participants, improving classification accuracy by 7% compared to standard EEG biomarkers"
      explanation: Reports the publication's headline comparison between the DADD-derived and conventional EEG biomarker classifiers.
  - statement: DADD-derived digital biomarkers identified participants positive for Alzheimer CSF biomarkers with 88% accuracy and predicted clinical cognitive-decline conversion with 87% accuracy.
    evidence:
    - reference: PMID:40450374
      reference_title: "Digital twins and non-invasive recordings enable early diagnosis of Alzheimer's disease."
      supports: SUPPORT
      evidence_source: COMPUTATIONAL
      snippet: "Identified patients positive for CSF biomarkers of AD with 88% accuracy (significantly outperforming standard EEG biomarkers, which achieved 58% accuracy); iii) Predicted follow-up conversions to clinical cognitive decline with 87% accuracy (compared to 54% accuracy for standard EEG biomarkers)."
      explanation: Reports the resampling performance stated in the publication abstract; it is not external prospective validation.
  evidence:
  - reference: PMID:40450374
    reference_title: "Digital twins and non-invasive recordings enable early diagnosis of Alzheimer's disease."
    supports: SUPPORT
    evidence_source: COMPUTATIONAL
    snippet: "We introduce an early diagnostic pipeline based on the Digital Alzheimer's Disease Diagnosis (DADD) digital twin model, which derives personalized AD biomarkers from non-invasive electroencephalographic (EEG) recordings."
    explanation: Establishes the model as an EEG-personalized Alzheimer disease digital twin.
  notes: >-
    Repository inspected at commit bf480f5cbba429639fcd6f1cb8dc92081a4f5975
    (2025-11-28). It contains three notebooks and a structural-connectivity NumPy
    array under GPL-2.0, but no pinned dependency manifest, release, or automated
    tests. The notebooks import The Virtual Brain and scientific-Python packages;
    execution was not attempted because the repository does not specify compatible
    package versions. The public notebooks support simulations and model inversion,
    while patient EEG data are not deposited in the repository.
environmental:
- name: Cumulative exposure to strong central anticholinergic medication
  exposure_term:
    preferred_term: exposure to muscarinic antagonist
    term:
      id: ECTO:9001824
      label: exposure to muscarinic antagonist
  description: >-
    Years of cumulative exposure to CNS-penetrant muscarinic antagonists —
    first-generation antihistamines such as diphenhydramine, tricyclic
    antidepressants, bladder antimuscarinics, and antiparkinson
    anticholinergics — is associated with incident dementia and Alzheimer
    disease in a dose-dependent manner across several large observational
    cohorts. Curated as a candidate modifiable risk factor, NOT as an
    established cause; see notes and the attached discussions.
  evidence:
  - reference: PMID:25621434
    reference_title: "Cumulative use of strong anticholinergics and incident dementia: a prospective cohort study."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "The most common anticholinergic classes used were tricyclic antidepressants, first-generation antihistamines, and bladder antimuscarinics."
    explanation: >-
      Identifies the drug classes constituting the exposure, including the
      first-generation antihistamines of which diphenhydramine is the exemplar.
  - reference: PMID:40707785
    reference_title: "Systemic medications and dementia risk: a systematic umbrella review."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "moderate certainty of increased risk with anticholinergics"
    explanation: >-
      A GRADE-rated umbrella review across 68 meta-analyses places the
      anticholinergic-dementia association at moderate certainty, the current
      ceiling for this exposure.
  - reference: PMID:11570937
    reference_title: Cognitive and other adverse effects of diphenhydramine use in hospitalized older patients.
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "Diphenhydramine administration in older hospitalized patients is associated with an increased risk of cognitive decline and other adverse effects with a dose-response relationship."
    explanation: >-
      Establishes drug-specific, dose-responsive acute cognitive harm from
      diphenhydramine — the well-supported near-term effect, distinct from the
      contested long-term dementia claim.
  - reference: PMID:42250644
    reference_title: "Second-Generation H1-Antihistamines Do Not Alter Dementia Risk in Type 2 Inflammatory Diseases: A Target Trial Emulation Using Real-World Data."
    supports: REFUTE
    directness: INDIRECT
    evidence_source: HUMAN_CLINICAL
    snippet: "Across all comparisons, no increased dementia risk was observed for users of any AH versus non-AH controls."
    explanation: >-
      A propensity-matched target trial emulation finds no dementia signal for
      antihistamines, including first-generation agents, which argues against
      the antihistamine arm of this exposure. Indirect rather than a direct
      refutation of the whole exposure for three reasons: it is a null in an indication cohort (chronic urticaria,
      chronic sinusitis, allergic rhinitis) rather than in the general older
      population this exposure describes; the exposure covers all strong
      central anticholinergics, not antihistamines alone; and TriNetX records
      dispensing-linked EHR data, so it shares the over-the-counter blind spot
      this entry's notes hold against the prescription databases.
  influences_mechanisms:
  - target: Muscarinic M1 Receptor Signaling Loss
    environmental_effect: PREDISPOSES
    causal_link_type: DIRECT
    description: >-
      Chronic anticholinergic exposure is the pharmacologic route by which M1
      muscarinic signaling is persistently withdrawn, the state the mouse work
      links to accelerated amyloid and tau pathology. PREDISPOSES rather than
      TRIGGERS: the human evidence is observational, dose-dependent, and does
      not establish that the exposure initiates disease.
    evidence:
    - reference: PMID:25621434
      reference_title: "Cumulative use of strong anticholinergics and incident dementia: a prospective cohort study."
      supports: SUPPORT
      evidence_source: HUMAN_CLINICAL
      snippet: "A 10-year cumulative dose-response relationship was observed for dementia and Alzheimer disease"
      explanation: >-
        Cumulative-dose dependence is the signature expected if sustained
        receptor blockade, rather than transient pharmacologic effect, is what
        matters.
    - reference: PMID:32878992
      reference_title: Association of anticholinergic medications and AD biomarkers with incidence of MCI among cognitively normal older adults.
      supports: SUPPORT
      evidence_source: HUMAN_CLINICAL
      snippet: "aCH+ participants had increased risk of progression to MCI"
      explanation: >-
        Anticholinergic exposure predicts incident MCI in cognitively normal
        adults, with the effect amplified by APOE4 and CSF AD pathology —
        interaction with AD biology rather than an independent cognitive effect.
  notes: >-
    Curated deliberately as PREDISPOSES, not as an established cause. Three
    findings keep it there. First, no randomized evidence: the Cochrane review
    of anticholinergic deprescribing found very low certainty evidence that can
    neither support nor refute a cognitive benefit of stopping these drugs
    (PMID:38063254). Second, class specificity: the large UK studies
    concentrate the signal in antidepressant, urological and antiparkinson
    drugs (PMID:29695481, PMID:31233095), and a propensity-matched target trial
    emulation finds no antihistamine signal (PMID:42250644). Third, the one
    large cohort reporting an antihistamine dose-response reports it for
    second-generation agents too (PMID:38935035), which do not meaningfully
    enter the CNS. That third point cuts both ways and should not be quoted as
    a failed negative control: the gradient is steeper for the
    first-generation, CNS-penetrant class (1.13/1.29/1.51 across rising
    cumulative dose versus 1.11/1.19/1.26 for second-generation), which is
    concordant with a central mechanism; it is the residual non-null
    second-generation signal, not the comparison between the classes, that the
    muscarinic mechanism fails to account for.
    IMPORTANT COUNTERWEIGHT, from the CPRD full text rather than its abstract:
    the antihistamine null there is weaker than it looks. Antihistamines were
    tested and were not significant on any-prescription exposure, but the
    authors report a tentative effect above 365 defined daily doses, describe
    the overall antihistamine association as small and short of their
    significance threshold, state that over-the-counter supply is not captured
    in the database so antihistamine use is underestimated, and call for
    independent confirmation specifically for antihistamines. Since
    diphenhydramine is predominantly an over-the-counter drug, exposure
    measures built from prescriptions systematically undercount precisely the
    exposure of interest, so the antihistamine nulls are substantially nulls of
    exposure misclassification rather than clean refutations. The claim that
    survives all of this unambiguously is the acute one (PMID:11570937), which
    is why diphenhydramine is listed as potentially inappropriate in older
    adults regardless of how the dementia question resolves.
- name: Oral glucosamine supplementation
  exposure_term:
    preferred_term: exposure to glucosamine supplement via ingestion
    term:
      id: ECTO:0070168
      label: exposure to glucosamine supplement via ingestion
  description: >-
    Glucosamine, taken widely as a joint supplement, enters the hexosamine
    pathway directly and so feeds N-glycan biosynthesis. In 5xFAD mice it
    raises brain N-glycans and worsens memory, while in wild-type mice it does
    neither. In a single-health-system electronic health record analysis,
    documented glucosamine use was associated with higher mortality in AD and
    related dementias and with more frequent conversion from mild cognitive
    impairment to dementia, but not with mortality in mild cognitive impairment.
    Curated as a candidate disease-exacerbating exposure in established
    disease, NOT as a risk factor for developing it: several large UK Biobank
    cohorts find glucosamine use associated with lower or unchanged incidence of
    dementia and Alzheimer disease.
  evidence:
  - reference: PMID:42265388
    reference_title: "Hyperglycosylation is a metabolic driver of Alzheimer's disease."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "Our analysis revealed that glucosamine usage was associated with a 25% increase in mortality risk among ADRD patients"
    explanation: >-
      The survival signal in the dementia cohort. ADRD here pools Alzheimer
      disease with vascular, Lewy body and frontotemporal dementia, so it is
      not an Alzheimer-specific estimate.
  - reference: PMID:42265388
    reference_title: "Hyperglycosylation is a metabolic driver of Alzheimer's disease."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "In this analysis, we observed a 25% increase in the proportion of patients with MCI transitioning to ADRD in the glucosamine user group"
    explanation: >-
      The progression signal: more conversion from mild cognitive impairment to
      dementia among documented users.
  - reference: PMID:42265388
    reference_title: "Hyperglycosylation is a metabolic driver of Alzheimer's disease."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "By contrast, glucosamine usage did not significantly impact mortality risk in the MCI cohort"
    explanation: >-
      The within-study contrast the authors use to argue that harm depends on
      established disease; it is what restricts this entry to EXACERBATES.
  - reference: PMID:42265388
    reference_title: "Hyperglycosylation is a metabolic driver of Alzheimer's disease."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "Glucosamine users were identified through a natural language processing pipeline that extracted glucosamine-related records from clinical notes using a regex-based keyword matching approach."
    explanation: >-
      Records how exposure was ascertained. Keyword matching on notes cannot
      distinguish current from past use, dose, or duration, and captures only
      use that a clinician wrote down.
  - reference: PMID:36978077
    reference_title: "Association of regular glucosamine use with incident dementia: evidence from a longitudinal cohort and Mendelian randomization study."
    supports: REFUTE
    directness: INDIRECT
    evidence_source: HUMAN_CLINICAL
    snippet: "In multivariable analysis, the hazard ratios (HR) of glucosamine users for all-cause dementia, AD, and vascular dementia were 0.84 (95% CI 0.75-0.93), 0.83 (95% CI 0.71-0.98), and 0.74 (95% CI 0.58-0.95), respectively."
    explanation: >-
      UK Biobank, dementia-free at baseline: glucosamine use is associated with
      lower incident Alzheimer disease. INDIRECT because it tests incidence in
      the general population, not progression or survival once disease is
      present, which is the claim this entry makes.
  - reference: PMID:36514123
    reference_title: "The role of type 2 diabetes in the association between habitual glucosamine use and dementia: a prospective cohort study."
    supports: REFUTE
    directness: INDIRECT
    evidence_source: HUMAN_CLINICAL
    snippet: "In fully adjusted models, habitual glucosamine use was associated with a significantly lower risk of dementia (HR = 0.87, 95% CI: 0.82-0.93)."
    explanation: >-
      A second UK Biobank analysis, near-complete cohort, all-cause dementia
      incidence. Same limitation as above: incidence rather than progression.
  - reference: PMID:37689747
    reference_title: "Habitual glucosamine use, APOE genotypes, and risk of incident cause-specific dementia in the older population."
    supports: REFUTE
    directness: INDIRECT
    evidence_source: HUMAN_CLINICAL
    snippet: "However, APOE ε4 dosage and baseline cognitive function did not significantly modify the relationships of glucosamine use with incident vascular dementia or Alzheimer's disease"
    explanation: >-
      In adults over 60 the association with incident Alzheimer disease is
      null, and baseline cognitive function does not modify it. That bears on
      the originating paper's argument that harm appears only once cognition is
      already affected, although lower baseline cognition in a dementia-free
      cohort is not the same state as diagnosed disease.
  - reference: PMID:37689747
    reference_title: "Habitual glucosamine use, APOE genotypes, and risk of incident cause-specific dementia in the older population."
    supports: REFUTE
    directness: INDIRECT
    evidence_source: HUMAN_CLINICAL
    snippet: "but not significantly associated with incident Alzheimer's disease (adjusted HR, 1.02; 95%CI, 0.92-1.14)"
    explanation: >-
      The direct result for Alzheimer disease in adults over 60: no association
      between habitual glucosamine use and incidence. INDIRECT for the same
      reason as the other UK Biobank items, since it concerns incidence rather
      than progression after diagnosis.
  influences_mechanisms:
  - target: Brain N-Glycan Hyperglycosylation
    environmental_effect: EXACERBATES
    causal_link_type: DIRECT
    description: >-
      Glucosamine bypasses the rate-limiting step of the hexosamine pathway and
      supplies precursor for N-glycan biosynthesis. EXACERBATES rather than
      TRIGGERS: in mice it raises brain N-glycans only on an AD background and
      leaves wild-type brains unchanged.
    evidence:
    - reference: PMID:42265388
      reference_title: "Hyperglycosylation is a metabolic driver of Alzheimer's disease."
      supports: SUPPORT
      evidence_source: MODEL_ORGANISM
      snippet: "Conversely, glucosamine supplementation significantly increased glycosylation across the brain, as shown by global heat map analysis"
      explanation: >-
        Oral glucosamine for two weeks raises brain N-glycans in 5xFAD mice, at
        a dose scaled from a human intake of about 2.5 g per day.
    - reference: PMID:42265388
      reference_title: "Hyperglycosylation is a metabolic driver of Alzheimer's disease."
      supports: SUPPORT
      evidence_source: MODEL_ORGANISM
      snippet: "Strikingly, glucosamine supplementation did not result in hyperglycosylation in WT mouse brains"
      explanation: >-
        The same regimen has no effect on wild-type brains, which is why the
        effect is curated as exacerbation of an existing state.
  notes: >-
    The human evidence for harm and the human evidence against it answer
    different questions and should not be netted against each other. The
    originating EHR analysis (PMID:42265388) is about outcomes after diagnosis,
    in one health system, with exposure from keyword matching of clinical notes
    in roughly 8% of patients, and with ADRD pooling several dementias. It is
    exposed to confounding by indication and frailty: glucosamine is taken for
    osteoarthritis and joint pain, which track age, immobility and
    multimorbidity, and the reported adjustment covers age, sex and
    demographics. The UK Biobank cohorts (PMID:36978077, PMID:36514123,
    PMID:37689747) are about incidence in people without dementia at baseline
    and find inverse or null associations; a fourth UK Biobank analysis
    (PMID:37158699) also finds no association with incident dementia (HR 1.06,
    95% CI 0.99-1.14). None of these is randomized. The originating authors call
    for a double-blind trial, which is the only design that separates a
    glucosamine effect from the reasons people take it.
diagnosis:
- name: Neuropsychological Tests
  notes: Assess cognitive function including memory, language, and problem-solving.
  evidence:
  - reference: PMID:23072720
    reference_title: "Neuropsychology of Alzheimer's disease."
    supports: SUPPORT
    snippet: A comprehensive neuropsychological examination encompassing several cognitive domains can provide a pattern of altered and preserved functions that is helpful to early detection, differential diagnosis and even prognosis of progression in predementia stages.
    explanation: The article mentions that neuropsychological tests encompass several cognitive domains, which include memory, language, and problem-solving.
  - reference: PMID:29851873
    reference_title: "Bedside Approach to the Mental Status Assessment."
    supports: SUPPORT
    snippet: The focused history and mental status examination remain essential tools for the evaluation and diagnosis of neurologic disorders affecting cognition, language, and behavior.
    explanation: The article highlights the importance of mental status examinations, which are part of neuropsychological tests, in assessing cognitive functions including memory, language, and behavior.
  - reference: PMID:37244373
    reference_title: "Association between dual-task function and neuropsychological testing in older adults with cognitive impairment."
    supports: SUPPORT
    snippet: The results demonstrate significant correlations between UEF cognitive score and mini-mental state examination (MMSE), Mini-Cog, Category fluency, Benson complex figure copy, Trail making test, and Montreal cognitive assessment (MOCA).
    explanation: The study shows that neuropsychological tests like MMSE, Mini-Cog, and others are used to assess cognitive functions, which include memory, language, and problem-solving.
  - reference: PMID:35308911
    reference_title: "Cognitive Function Characterization Using Electronic Health Records Notes."
    supports: SUPPORT
    snippet: Assessing cognitive impairment is important for diagnostic, clinical management, and research purposes. The Folstein Mini-Mental State Examination (MMSE) is the most common screening measure of cognitive function.
    explanation: The article supports the use of neuropsychological tests, such as the MMSE, for assessing cognitive functions including memory, language, and problem-solving.
- name: Brain Imaging
  notes: MRI and PET scans to detect brain changes such as shrinkage and amyloid plaques.
  evidence:
  - reference: PMID:22173295
    reference_title: "Brain imaging in the study of Alzheimer's disease."
    supports: SUPPORT
    snippet: Brain imaging researchers have contributed to the scientific understanding, early detection and tracking of AD. ... They have developed ground-breaking methods, including positron emission tomography (PET) ligands to measure fibrillar amyloid-β (Aβ) deposition, new magnetic resonance imaging (MRI) pulse sequences, and powerful image analysis techniques
    explanation: The literature supports the use of MRI and PET scans for detecting brain changes such as shrinkage and amyloid plaques in Alzheimer's Disease.
  - reference: PMID:33640881
    reference_title: "Evaluating the association between brain atrophy, hypometabolism, and cognitive decline in Alzheimer's disease: a PET/MRI study."
    supports: SUPPORT
    snippet: Glucose metabolism reduction and brain volume losses are widely reported in Alzheimer's disease (AD). ... The AD group had significantly reduced volume in the hippocampus and DMN regions (P < 0.001) relative to that of normal controls determined by using ROI analysis.
    explanation: The literature supports the use of brain imaging techniques like MRI and PET to detect brain changes such as shrinkage and amyloid plaques in Alzheimer's Disease.
  - reference: PMID:34127752
    reference_title: "In vivo multi-parametric manganese-enhanced MRI for detecting amyloid plaques in rodent models of Alzheimer's disease."
    supports: SUPPORT
    evidence_source: MODEL_ORGANISM
    snippet: Amyloid plaques are a hallmark of Alzheimer's disease (AD) that develop in its earliest stages. Thus, non-invasive detection of these plaques would be invaluable for diagnosis and the development and monitoring of treatments.
    explanation: The literature supports the use of MRI and PET scans for detecting amyloid plaques in Alzheimer's Disease.
  - reference: PMID:18694837
    reference_title: "Multitracer PET imaging of amyloid plaques and neurofibrillary tangles in Alzheimer's disease."
    supports: SUPPORT
    snippet: Recently developed positron emission tomography (PET) tracers, such as PIB and FDDNP, help to visualize amyloid plaques and neurofibrillary tangles in living subjects.
    explanation: The literature supports the use of PET scans to detect amyloid plaques and neurofibrillary tangles in Alzheimer's Disease.
- name: Cerebrospinal Fluid Analysis
  notes: Measurement of amyloid-beta, total tau, and phosphorylated tau levels.
  evidence:
  - reference: PMID:19661632
    reference_title: "Cerebrospinal fluid biomarkers for Alzheimer's disease."
    supports: SUPPORT
    snippet: The core candidate CSF biomarkers Abeta42, total tau (T-tau), and phosphorylated tau (P-tau) have been shown to have a high diagnostic performance to identify AD also in the early phase of the disease.
    explanation: This reference explicitly mentions the use of amyloid-beta, total tau, and phosphorylated tau levels in cerebrospinal fluid for diagnosing Alzheimer's disease.
  - reference: PMID:12975285
    reference_title: "Total tau and phosphorylated tau 181 levels in the cerebrospinal fluid of patients with frontotemporal dementia due to P301L and G272V tau mutations."
    supports: SUPPORT
    snippet: Total tau, Ptau-181, and amyloid-beta1-42 levels in CSF, obtained by lumbar puncture, were determined by sandwich enzyme-linked immunosorbent assay.
    explanation: This study evaluates levels of total tau, phosphorylated tau 181, and amyloid-beta in the cerebrospinal fluid of patients with Alzheimer's disease, supporting the statement.
  - reference: PMID:36510321
    reference_title: "Antibody-free measurement of cerebrospinal fluid tau phosphorylation across the Alzheimer's disease continuum."
    supports: SUPPORT
    snippet: Alzheimer's disease is characterized by an abnormal increase of phosphorylated tau (pTau) species in the CSF.
    explanation: This reference supports the statement by discussing the increase of phosphorylated tau in cerebrospinal fluid in Alzheimer's disease.
  - reference: PMID:35841250
    reference_title: "Elecsys Cerebrospinal Fluid Assays Accurately Distinguish Alzheimer's Disease from Frontotemporal Lobar Degeneration."
    supports: SUPPORT
    snippet: CSF diagnostic assays for the differentiation of AD and FTLD may increase diagnostic accuracy.
    explanation: This reference supports the statement by discussing the use of cerebrospinal fluid diagnostic assays for Alzheimer's disease, which include measurements of amyloid-beta and tau proteins.
  - reference: PMID:38431278
    reference_title: "CSF Biomarkers in Longitudinal Alzheimer Disease Cohorts: Pre-Analytic Challenges."
    supports: SUPPORT
    snippet: Levels of amyloid beta 1-42 (Abeta42), phosphorylated tau 181 (pTau181), and total tau (tTau) were obtained using an Elecsys cobas e 601 platform.
    explanation: This reference supports the statement by discussing the measurement of amyloid-beta, total tau, and phosphorylated tau levels in cerebrospinal fluid.
- name: Blood-Based Biomarkers (plasma p-tau217)
  notes: >-
    Minimally invasive plasma assays now approach the accuracy of amyloid-PET and
    CSF for detecting AD pathology, greatly expanding access to biomarker-based
    diagnosis. Plasma phosphorylated-tau 217 (p-tau217) and the p-tau217/Aβ42
    ratio are the strongest performers for classifying amyloid status and
    predicting progression; the plasma Aβ42/40 ratio shows the earliest
    detectable abnormalities. In May 2025 the first blood test for brain amyloid
    (a p-tau217/β-amyloid 1-42 ratio) received FDA approval. Blood biomarkers
    complement — and are increasingly used to triage before — CSF and PET testing,
    which remains relevant given that anti-amyloid immunotherapies require
    confirmed amyloid positivity.
  evidence:
  - reference: PMID:32722745
    reference_title: "Discriminative Accuracy of Plasma Phospho-tau217 for Alzheimer Disease vs Other Neurodegenerative Disorders."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "Among 1402 participants from 3 selected cohorts, plasma P-tau217 discriminated AD from other neurodegenerative diseases, with significantly higher accuracy than established plasma- and MRI-based biomarkers, and its performance was not significantly different from key CSF- or PET-based measures."
    explanation: >-
      Plasma p-tau217 discriminates AD from other neurodegenerative diseases with
      accuracy exceeding other blood/MRI biomarkers and comparable to CSF/PET,
      supporting its use as a blood-based diagnostic biomarker.
  - reference: PMID:32722745
    reference_title: "Discriminative Accuracy of Plasma Phospho-tau217 for Alzheimer Disease vs Other Neurodegenerative Disorders."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "To examine plasma tau phosphorylated at threonine 217 (P-tau217) as a diagnostic biomarker for AD."
    explanation: >-
      Establishes plasma p-tau217 (tau phosphorylated at threonine 217) as the
      diagnostic biomarker evaluated in this study.
- name: Montreal Cognitive Assessment
  description: >-
    Brief cognitive screening instrument developed to help first-line physicians
    detect mild cognitive impairment. Its validation study administered it,
    alongside the MMSE, to patients with mild Alzheimer's disease, patients with
    mild cognitive impairment and healthy elderly controls, and it detected every
    patient in the mild Alzheimer's disease group at the published cut-off.
  notes: >-
    Copyright © Tina Brosseau for Center for Diagnosis & Research on Alzheimer's disease (CEDRA). Used with permission.
    The instrument is bound at its LOINC panel code and the cut-offs at its
    total-score code. The instrument's items are not reproduced in this entry;
    only published thresholds are recorded.
  measurements:
  - preferred_term: MoCA
    term:
      id: LOINC:72133-2
      label: Montreal Cognitive Assessment [MoCA]
  reference_ranges:
  - loinc_term:
      id: LOINC:72172-0
      label: Total score [MoCA]
    unit: '{score}'
    population: >-
      MoCA validation study, 2005 (mild Alzheimer's disease with MMSE score of 17
      or more, mild cognitive impairment, and healthy elderly controls; a
      community clinic and an academic center)
    notes: >-
      The abstract states the cut-off as "Using a cutoff score 26" and does not
      say in its own text on which side a score of exactly 26 falls; the bands
      put 26 in the unimpaired tier and 25 and below in the impaired tier, a
      placement the cited abstract does not itself state. A later meta-analysis
      of MCI validation studies (PMID:28731508) reports that "the originally
      suggested cutoff score of 26/30 leads to an inflated rate of false
      positives, particularly for those of older age and/or lower education" and
      recommends 23; it concerns MCI rather than Alzheimer's disease, so it is
      not recorded as a separate range here.
    evidence:
    - reference: PMID:15817019
      reference_title: "The Montreal Cognitive Assessment, MoCA: a brief screening tool for mild cognitive impairment."
      supports: SUPPORT
      evidence_source: HUMAN_CLINICAL
      quote_role: PRIMARY_RESULT
      snippet: "Using a cutoff score 26, the MMSE had a sensitivity of 18% to detect MCI, whereas the MoCA detected 90% of MCI subjects. In the mild AD group, the MMSE had a sensitivity of 78%, whereas the MoCA detected 100%."
      explanation: States the cut-off of 26 and the MoCA's sensitivity for mild Alzheimer's disease at that cut-off.
    - reference: PMID:28731508
      reference_title: "A re-examination of Montreal Cognitive Assessment (MoCA) cutoff scores."
      supports: SUPPORT
      evidence_source: HUMAN_CLINICAL
      quote_role: BACKGROUND
      snippet: "Some studies have indicated that the originally suggested cutoff score of 26/30 leads to an inflated rate of false positives, particularly for those of older age and/or lower education."
      explanation: Confirms that 26 is the originally suggested cut-off and records the known false-positive caveat in older and less-educated people.
    interpretation_bands:
    - name: Below the cut-off (cognitive impairment)
      upper_bound: 26
      unit: '{score}'
      abnormal_flag: LOW
    - name: At or above the cut-off
      lower_bound: 26
      unit: '{score}'
      abnormal_flag: NORMAL
  - loinc_term:
      id: LOINC:72172-0
      label: Total score [MoCA]
    unit: '{score}'
    population: >-
      Taiwanese cohort of 67 patients with mild-to-moderate Alzheimer's disease,
      36 with dementia with Lewy bodies and 62 healthy participants without
      dementia, aged 60 to 90 (Alzheimer's disease versus controls)
    notes: >-
      The abstract states "The cut-off point of the MoCA for AD was 21/22"; the
      bands read the x/y notation as 21 or below positive and 22 or above
      negative, which the abstract does not spell out.
    evidence:
    - reference: PMID:23919950
      reference_title: "Montreal Cognitive Assessment and Mini-Mental State Examination performance in patients with mild-to-moderate dementia with Lewy bodies, Alzheimer's disease, and normal participants in Taiwan."
      supports: SUPPORT
      evidence_source: HUMAN_CLINICAL
      quote_role: PRIMARY_RESULT
      snippet: "The cut-off point of the MoCA for AD was 21/22 with a sensitivity of 95.5% and a specificity of 82.3% (area under the curve (AUC): 0.945)"
      explanation: States the MoCA cut-off separating Alzheimer's disease from controls in this cohort.
    interpretation_bands:
    - name: At or below 21 (screen positive for Alzheimer's disease)
      upper_bound: 22
      unit: '{score}'
      abnormal_flag: LOW
    - name: 22 or above
      lower_bound: 22
      unit: '{score}'
      abnormal_flag: NORMAL
  evidence:
  - reference: PMID:15817019
    reference_title: "The Montreal Cognitive Assessment, MoCA: a brief screening tool for mild cognitive impairment."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    quote_role: PRIMARY_RESULT
    snippet: "The MoCA and MMSE were administered to all participants, and sensitivity and specificity of both measures were assessed for detection of MCI and mild AD."
    explanation: The instrument was validated for detection of mild Alzheimer's disease as well as MCI.
- name: Mini-Mental State Examination
  description: >-
    The most common screening measure of cognitive function. MMSE score ranges
    have been mapped onto Clinical Dementia Rating stages so that the MMSE can
    stand in for the CDR when staging dementia in Alzheimer's disease.
  notes: >-
    Copyright © Vicki King for Psychological Assessment Resources. Used with permission.
    The instrument is bound at its LOINC panel code and the cut-offs at its
    total-score code. The instrument's items are not reproduced in this entry;
    only published thresholds are recorded.
  measurements:
  - preferred_term: MMSE
    term:
      id: LOINC:72107-6
      label: Mini-Mental State Examination [MMSE]
  reference_ranges:
  - loinc_term:
      id: LOINC:72106-8
      label: Total score [MMSE]
    unit: '{score}'
    population: >-
      863 probands (524 with probable Alzheimer's disease, 92 with questionable
      dementia, 247 with memory complaints but no objective cognitive
      impairment), MMSE ranges mapped onto CDR stages
    notes: >-
      The abstract states "The MMSE ranges were 30 for no, 26-29 for
      questionable, 21-25 for mild, 11-20 for moderate, and 0-10 for severe
      dementia." The integer ranges are written as half-open bands. The same
      abstract reports that the MMSE "performed poorly in the separation between
      CDR stages zero and 0.5", with only moderate agreement for no dementia
      (kappa=0.44) and fair agreement for questionable dementia (kappa=0.28), so
      the two upper bands are the least reliable.
    evidence:
    - reference: PMID:16473978
      reference_title: "Mapping scores onto stages: mini-mental state examination and clinical dementia rating."
      supports: SUPPORT
      evidence_source: HUMAN_CLINICAL
      quote_role: PRIMARY_RESULT
      snippet: "The MMSE ranges were 30 for no, 26-29 for questionable, 21-25 for mild, 11-20 for moderate, and 0-10 for severe dementia."
      explanation: States the MMSE severity ranges recorded as bands.
    - reference: PMID:16473978
      reference_title: "Mapping scores onto stages: mini-mental state examination and clinical dementia rating."
      supports: SUPPORT
      evidence_source: HUMAN_CLINICAL
      quote_role: PRIMARY_RESULT
      snippet: "The MMSE can be used as a surrogate measure for the CDR for the staging of dementia in AD."
      explanation: The ranges are a staging scheme for dementia in Alzheimer's disease.
    interpretation_bands:
    - name: Severe dementia
      lower_bound: 0
      upper_bound: 11
      unit: '{score}'
      abnormal_flag: LOW
      severity: SEVERE
    - name: Moderate dementia
      lower_bound: 11
      upper_bound: 21
      unit: '{score}'
      abnormal_flag: LOW
      severity: MODERATE
    - name: Mild dementia
      lower_bound: 21
      upper_bound: 26
      unit: '{score}'
      abnormal_flag: LOW
      severity: MILD
    - name: Questionable dementia
      lower_bound: 26
      upper_bound: 30
      unit: '{score}'
      abnormal_flag: LOW
    - name: No dementia
      lower_bound: 30
      unit: '{score}'
      abnormal_flag: NORMAL
  - loinc_term:
      id: LOINC:72106-8
      label: Total score [MMSE]
    unit: '{score}'
    population: >-
      Taiwanese cohort of 67 patients with mild-to-moderate Alzheimer's disease,
      36 with dementia with Lewy bodies and 62 healthy participants without
      dementia, aged 60 to 90 (Alzheimer's disease versus controls)
    notes: >-
      The abstract states "For the MMSE, the cut-off points for AD and for DLB
      from NC were all 24/25"; the bands read the x/y notation as 24 or below
      positive and 25 or above negative, which the abstract does not spell out.
    evidence:
    - reference: PMID:23919950
      reference_title: "Montreal Cognitive Assessment and Mini-Mental State Examination performance in patients with mild-to-moderate dementia with Lewy bodies, Alzheimer's disease, and normal participants in Taiwan."
      supports: SUPPORT
      evidence_source: HUMAN_CLINICAL
      quote_role: PRIMARY_RESULT
      snippet: "For the MMSE, the cut-off points for AD and for DLB from NC were all 24/25, with a sensitivity of 88.1% and a specificity of 85.5% for AD (AUC: 0.92)"
      explanation: States the MMSE cut-off separating Alzheimer's disease from controls in this cohort.
    interpretation_bands:
    - name: At or below 24 (screen positive for Alzheimer's disease)
      upper_bound: 25
      unit: '{score}'
      abnormal_flag: LOW
    - name: 25 or above
      lower_bound: 25
      unit: '{score}'
      abnormal_flag: NORMAL
  evidence:
  - reference: PMID:35308911
    reference_title: "Cognitive Function Characterization Using Electronic Health Records Notes."
    supports: SUPPORT
    snippet: Assessing cognitive impairment is important for diagnostic, clinical management, and research purposes. The Folstein Mini-Mental State Examination (MMSE) is the most common screening measure of cognitive function.
    explanation: The MMSE is the most commonly used cognitive screening measure, which is why it is bound as a measurement here.
- name: Clinical Dementia Rating
  description: >-
    Global dementia staging instrument rated from information collected from
    both the patient and an informant. Its stages (0, 0.5, 1, 2 and 3) grade
    Alzheimer's disease from no dementia through questionable, mild, moderate
    and severe dementia. Its sum of boxes is the primary endpoint of several
    trials recorded in this entry's treatments.
  notes: >-
    Copyright © Washington University Charles F. and Joanne Knight Alzheimer's Disease Research Center (Knight ADRC). Used with permission.
    LOINC carries the CDR only as a panel. A search of the LOINC 2.83 table
    (the Tuva Project extract pinned in data/loinc/MANIFEST.yaml) for "CDR",
    "dementia rating" and "sum of boxes", and of the Monarch KG for "Clinical
    Dementia Rating", "CDR global", "Sum of boxes score [CDR]" and "Total score
    [CDR]", found no global-score or sum-of-boxes code, so the CDR stages are not
    recorded as interpretation bands: a numeric band on a panel code would put a
    score threshold on the instrument's identity. The instrument's items are not
    reproduced in this entry.
  measurements:
  - preferred_term: CDR
    term:
      id: LOINC:72088-8
      label: Clinical Dementia Rating scale [CDR]
  evidence:
  - reference: PMID:16473978
    reference_title: "Mapping scores onto stages: mini-mental state examination and clinical dementia rating."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    quote_role: BACKGROUND
    snippet: "The Clinical Dementia Rating (CDR) has demonstrated high validity and reliability for this purpose, but it requires a considerable amount of data to be collected both from the patient and from an informant."
    explanation: The CDR is the validated instrument for distinguishing severity levels of Alzheimer's disease, rated from patient and informant data.
  - reference: PMID:16473978
    reference_title: "Mapping scores onto stages: mini-mental state examination and clinical dementia rating."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    quote_role: PRIMARY_RESULT
    snippet: "The MMSE discriminated well between CDR stages 0.5, 1, 2, and 3 but performed poorly in the separation between CDR stages zero and 0.5."
    explanation: Names the CDR stages used to stage dementia in Alzheimer's disease.
treatments:
- name: Cholinesterase Inhibitors
  description: Medications that slow the breakdown of acetylcholine to help with memory and cognitive function (e.g., donepezil, rivastigmine).
  evidence:
  - reference: PMID:24807367
    reference_title: "[Acetylcholinesterase inhibitors for treatment of Alzheimer's disease]."
    supports: SUPPORT
    snippet: Donepezil, galantamine and rivastigmine are commonly used AChEIs in pharmacotherapy for AD, slowing the progression and controlling the symptoms of AD.
    explanation: The literature supports that cholinesterase inhibitors, including donepezil and rivastigmine, are used to help with symptoms of Alzheimer's disease by slowing the progression and controlling cognitive symptoms.
  - reference: PMID:28671413
    reference_title: "Alzheimer Disease: Pharmacologic and Nonpharmacologic Therapies for Cognitive and Functional Symptoms."
    supports: SUPPORT
    snippet: Cholinesterase inhibitors, memantine, and a combination of a cholinesterase inhibitor and memantine have produced statistically significant but clinically small delays in various domains of cognitive and functional decline in select patients with Alzheimer disease.
    explanation: The literature supports the use of cholinesterase inhibitors in delaying cognitive decline in Alzheimer's disease.
  - reference: PMID:35608903
    reference_title: "Withdrawal or continuation of cholinesterase inhibitors or memantine or both, in people with dementia."
    supports: SUPPORT
    snippet: Two classes of drug - cholinesterase inhibitors (donepezil, galantamine and rivastigmine) and memantine - are widely licensed for dementia due to Alzheimer's disease.
    explanation: The literature confirms that cholinesterase inhibitors are licensed for use in Alzheimer's disease to alleviate symptoms and delay disease progression.
  - reference: PMID:36412156
    reference_title: "[Capabilities of combined therapy of Alzheimer's disease]."
    supports: SUPPORT
    snippet: 'Currently approved medications are symptomatic and include two classes: cholinesterase inhibitors, such as donepezil, and NMDA receptor antagonist memantine.'
    explanation: The literature supports that cholinesterase inhibitors are approved symptomatic treatments for Alzheimer's disease.
  - reference: PMID:9108896
    reference_title: "Donepezil."
    supports: SUPPORT
    snippet: Donepezil is a specific and potent acetylcholinesterase inhibitor... Donepezil 5 or 10 mg/day was associated with significant improvements in cognitive function.
    explanation: The literature supports that donepezil, a cholinesterase inhibitor, helps improve cognitive function in Alzheimer's disease.
  treatment_term:
    preferred_term: cholinesterase inhibitor therapy
    term:
      id: NCIT:C15986
      label: Pharmacotherapy
- name: Anti-Amyloid Monoclonal Antibody Therapy
  description: >-
    Amyloid-beta-directed monoclonal antibodies (lecanemab, donanemab) are
    disease-modifying immunotherapies for early Alzheimer disease (mild cognitive
    impairment or mild dementia with confirmed amyloid pathology). They clear
    brain amyloid-beta and produce a modest slowing of cognitive and functional
    decline; use is limited by amyloid-related imaging abnormalities (ARIA) and
    infusion reactions, and requires amyloid confirmation plus MRI monitoring.
  therapeutic_modality: MONOCLONAL_ANTIBODY
  evidence:
  - reference: PMID:36449413
    reference_title: "Lecanemab in Early Alzheimer's Disease."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "Lecanemab, a humanized IgG1 monoclonal antibody that binds with high affinity to Aβ soluble protofibrils, is being tested in persons with early Alzheimer's disease."
    explanation: Identifies lecanemab as an amyloid-beta protofibril-directed monoclonal antibody developed for early Alzheimer disease.
  - reference: PMID:36449413
    reference_title: "Lecanemab in Early Alzheimer's Disease."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "Lecanemab reduced markers of amyloid in early Alzheimer's disease and resulted in moderately less decline on measures of cognition and function than placebo at 18 months but was associated with adverse events."
    explanation: Phase 3 CLARITY-AD trial (NCT03887455) shows lecanemab lowers amyloid burden and modestly slows cognitive/functional decline versus placebo, supporting the disease-modifying anti-amyloid immunotherapy class.
  - reference: PMID:37459141
    reference_title: "Donanemab in Early Symptomatic Alzheimer Disease: The TRAILBLAZER-ALZ 2 Randomized Clinical Trial."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "Among participants with early symptomatic Alzheimer disease and amyloid and tau pathology, donanemab significantly slowed clinical progression at 76 weeks in those with low/medium tau and in the combined low/medium and high tau pathology population."
    explanation: Phase 3 TRAILBLAZER-ALZ 2 trial (NCT04437511) demonstrates that donanemab significantly slows clinical progression in early symptomatic Alzheimer disease, anchoring the second FDA-reviewed anti-amyloid immunotherapy in this class.
  - reference: PMID:41390238
    reference_title: Lecanemab appropriate use recommendations for clinical practice in the UK.
    supports: SUPPORT
    evidence_source: OTHER
    snippet: "Lecanemab is an anti-amyloid monoclonal antibody, recently approved in the UK as a treatment for mild cognitive impairment (MCI) and mild dementia due to Alzheimer's disease (AD)"
    explanation: UK appropriate use recommendations position lecanemab as an approved anti-amyloid monoclonal antibody for mild cognitive impairment and mild dementia due to Alzheimer disease.
  treatment_term:
    preferred_term: Pharmacotherapy
    term:
      id: NCIT:C15986
      label: Pharmacotherapy
    therapeutic_agent:
    - preferred_term: lecanemab
      term:
        id: NCIT:C175105
        label: Lecanemab
    - preferred_term: donanemab
      term:
        id: NCIT:C166484
        label: Donanemab
    - preferred_term: monoclonal antibody
      term:
        id: NCIT:C20401
        label: Monoclonal Antibody
- name: Tau-Directed Antisense Oligonucleotide Therapy
  description: >-
    Diranersen (BIIB080 / IONIS-MAPTRx) is an investigational, intrathecally
    administered antisense oligonucleotide that binds MAPT mRNA and recruits
    RNase H1 to degrade it, lowering production of both intracellular and
    extracellular tau. It is the first tau-lowering strategy to demonstrate
    target engagement in Alzheimer disease and represents a mechanistically
    distinct disease-modifying approach from the anti-amyloid antibodies:
    rather than clearing extracellular amyloid-beta, it reduces the tau
    substrate that hyperphosphorylates and aggregates into neurofibrillary
    tangles, the pathology most tightly correlated with cognitive decline. The
    phase 1b MAPTRx trial (NCT03186989) showed dose-dependent >50% mean
    reductions in CSF total-tau, and the phase 2 CELIA study (NCT05399888) in
    mild cognitive impairment / mild AD dementia used the Clinical Dementia
    Rating-Sum of Boxes (CDR-SB) as its primary endpoint. Per Biogen/Ionis
    topline results reported at AAIC 2026, CELIA reduced CSF total-tau by
    ~50-65% across doses and showed a slowing of clinical decline versus
    placebo, although it did not meet its primary endpoint; the FDA granted
    diranersen Fast Track designation in 2025.
  therapeutic_modality: ANTISENSE_OLIGONUCLEOTIDE
  oligonucleotide_details:
    oligonucleotide_mechanism: RNASE_H_KNOCKDOWN
    target_gene:
      preferred_term: MAPT
      term:
        id: hgnc:6893
        label: MAPT
    target_transcript: MAPT mRNA
    oligonucleotide_chemistry: TWO_PRIME_O_METHOXYETHYL
    conjugation: UNCONJUGATED
    delivery_platform: UNFORMULATED
  treatment_term:
    preferred_term: Pharmacotherapy
    term:
      id: NCIT:C15986
      label: Pharmacotherapy
    therapeutic_agent:
    - preferred_term: diranersen
      term:
        id: NCIT:C1291
        label: Antisense Oligonucleotides
  evidence:
  - reference: PMID:37095250
    reference_title: "Tau-targeting antisense oligonucleotide MAPT(Rx) in mild Alzheimer's disease: a phase 1b, randomized, placebo-controlled trial."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "We sought to inhibit MAPT expression with a tau-targeting antisense oligonucleotide (MAPTRx) and reduce tau levels in patients with mild AD."
    explanation: Phase 1b MAPTRx trial establishes diranersen as an antisense oligonucleotide that lowers tau by inhibiting MAPT expression in mild Alzheimer disease.
  - reference: PMID:37095250
    reference_title: "Tau-targeting antisense oligonucleotide MAPT(Rx) in mild Alzheimer's disease: a phase 1b, randomized, placebo-controlled trial."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "Dose-dependent reduction in the CSF total-tau concentration was observed with greater than 50% mean reduction from baseline at 24 weeks post-last dose in the 60 mg (four doses) and 115 mg (two doses) MAPTRx groups."
    explanation: Demonstrates dose-dependent CSF total-tau lowering (target engagement) with MAPTRx, the pharmacodynamic basis for the tau-directed ASO approach.
  - reference: PMID:37095250
    reference_title: "Tau-targeting antisense oligonucleotide MAPT(Rx) in mild Alzheimer's disease: a phase 1b, randomized, placebo-controlled trial."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "randomized 3:1 to intrathecal bolus administrations of MAPTRx or placebo every 4 or 12 weeks during the 13-week treatment period"
    explanation: Documents the intrathecal (CSF) route of administration required for CNS delivery of the unconjugated ASO.
  target_mechanisms:
  - target: Neurofibrillary Tangle Formation
    treatment_effect: INHIBITS
    description: >-
      Diranersen indirectly inhibits this mechanism upstream rather than
      blocking aggregate assembly directly: RNase H1-mediated MAPT mRNA
      knockdown lowers the total tau available to become hyperphosphorylated and
      aggregate into neurofibrillary tangles.
- name: NMDA Receptor Antagonist
  description: Medication that regulates glutamate activity to improve symptoms (e.g., memantine).
  evidence:
  - reference: PMID:12768511
    reference_title: "[Memantine]."
    supports: SUPPORT
    snippet: Memantine, an antagonist of the glutamatergic NMDA receptor, has been recently approved for the treatment of advanced AD. Due to its action mechanism, memantine is considered a neuroprotective drug, whose utility has been demonstrated in preclinical studies, and a useful symptomatic treatment for AD and vascular dementia.
    explanation: The abstract confirms that memantine, an NMDA receptor antagonist, is used to treat Alzheimer's Disease by regulating glutamate activity.
  - reference: PMID:27662322
    reference_title: "Role of Glutamate and NMDA Receptors in Alzheimer's Disease."
    supports: SUPPORT
    snippet: Studies indicate that the distinct outcomes of NMDAR-mediated responses are induced by regionalized receptor activities, followed by different downstream signaling pathways. The activation of synaptic NMDARs initiates plasticity and stimulates cell survival. In contrast, the activation of extrasynaptic NMDARs promotes cell death and thus contributes to the etiology of AD, which can be blocked by an AD drug, memantine, an NMDAR antagonist that selectively blocks the function of extrasynaptic NMDARs.
    explanation: This abstract highlights that memantine, an NMDA receptor antagonist, helps in blocking the negative effects of excessive NMDAR activity in Alzheimer's Disease.
  - reference: PMID:20943326
    reference_title: "Targeting glutamate mediated excitotoxicity in Huntington's disease: neural progenitors and partial glutamate antagonist--memantine."
    supports: SUPPORT
    directness: INDIRECT
    snippet: The hypothesis proposed is restoration of medium spiny neurons in Huntington's disease using neural progenitor cell implantation and attenuation of glutamate mediated excitotoxicity using a partial glutamate antagonist - Memantine. Memantine can block the NMDA receptors and will prevent excess calcium influx into the neurons decreases the vulnerability of medium spiny neurons to glutamate mediated excitotoxicity.
    explanation: Although primarily discussing Huntington's Disease, this abstract supports the idea that memantine, an NMDA receptor antagonist, regulates glutamate activity, which is relevant to Alzheimer's Disease treatment.
  treatment_term:
    preferred_term: Pharmacotherapy
    term:
      id: NCIT:C15986
      label: Pharmacotherapy
    therapeutic_agent:
    - preferred_term: memantine
      term:
        id: CHEBI:64312
        label: memantine
- name: Cognitive Therapy
  description: Non-pharmacological interventions to maintain cognitive function.
  evidence:
  - reference: PMID:35621327
    reference_title: "Cognitive Stimulation in Moderate Alzheimer's Disease."
    supports: SUPPORT
    snippet: Cognitive stimulation was found to be an effective intervention for people with moderate Alzheimer's disease because it helped to maintain memory function, executive functions, and attention.
    explanation: This study specifically highlights the effectiveness of cognitive stimulation in maintaining cognitive functions in patients with moderate Alzheimer's disease.
  - reference: PMID:28671413
    reference_title: "Alzheimer Disease: Pharmacologic and Nonpharmacologic Therapies for Cognitive and Functional Symptoms."
    supports: SUPPORT
    snippet: Cognitive stimulation programs show benefit in maintenance of cognitive function and improved self-reported quality of life in patients with mild to moderate Alzheimer disease.
    explanation: This reference supports the statement by indicating that cognitive stimulation programs help maintain cognitive function in patients with mild to moderate Alzheimer's disease.
  - reference: PMID:37428401
    reference_title: "Cognitive Interventions for Neurodegenerative Disease."
    supports: SUPPORT
    snippet: CT can improve discrete cognitive functions, but durability is limited and real-world utility is unclear. CR treatments are holistic and flexible and, therefore, most promising but are difficult to simulate and study under rigorous experimental conditions.
    explanation: A review of cognitive interventions in Alzheimer and Parkinson disease. Cognitive training improves specific functions with limited durability, and cognitive rehabilitation is the most promising but least rigorously tested approach, so the support for maintaining cognitive function is real but modest.
  - reference: PMID:21643921
    reference_title: "Cognitive reserve and its implications for rehabilitation and Alzheimer's disease."
    supports: SUPPORT
    snippet: Mental training and cognitive stimulation interventions in AD have been shown to be useful in increasing patients' ability in performing activities of daily living (ADL), allowing them to maintain relative independence.
    explanation: This reference supports the utility of cognitive interventions in maintaining daily functioning, which is related to cognitive function, but does not directly address cognitive maintenance alone.
  - reference: PMID:27159433
    reference_title: "Healthy cognitive aging and dementia prevention."
    supports: SUPPORT
    snippet: Interventions involving physical exercise and cognitive training have consistently shown positive effects on cognition in older adults.
    explanation: This reference supports the statement by indicating that cognitive training, a form of cognitive therapy, has positive effects on cognition in older adults, including those with Alzheimer's disease.
  therapeutic_modality: BEHAVIORAL
  treatment_term:
    preferred_term: behavioral counseling
    term:
      id: NCIT:C181743
      label: Behavioral Counseling
- name: Supportive Care
  description: Includes occupational therapy, speech therapy, and caregiver support.
  evidence:
  - reference: PMID:28809650
    reference_title: "Supporting Adults With Alzheimer's Disease and Related Major Neurocognitive Disorders and Their Caregivers: Effective Occupational Therapy Interventions."
    supports: SUPPORT
    snippet: Occupational therapy practitioners play a significant role in supporting adults with Alzheimer's disease and related major neurocognitive disorders, as well as their caregivers, through all phases of the disease process.
    explanation: The reference highlights the role of occupational therapy in supporting individuals with Alzheimer's disease, which aligns with the statement that includes occupational therapy as part of supportive care.
  - reference: PMID:38883339
    reference_title: "Self-help support: The Alzheimer's telephone from the user's perspective."
    supports: SUPPORT
    snippet: The telephone hotline is a useful component of dementia care in Germany and an important contribution to the National Dementia Strategy.
    explanation: The reference discusses the importance of caregiver support through telephone counseling, aligning with the statement that includes caregiver support as part of supportive care.
  - reference: PMID:29361068
    reference_title: "Meeting the Informational, Educational, and Psychosocial Support Needs of Persons Living With Dementia and Their Family Caregivers."
    supports: SUPPORT
    snippet: This article first describes the educational, information, and support needs of individuals living dementia and their family caregivers across all stages of Alzheimer's.
    explanation: The reference discusses the support needs of individuals with Alzheimer's disease and their caregivers, which aligns with the statement that includes caregiver support as part of supportive care.
  treatment_term:
    preferred_term: Supportive Care
    term:
      id: NCIT:C15747
      label: Supportive Care
- name: Lifestyle Modifications
  description: Physical exercise, mental stimulation, and healthy diet to potentially slow disease progression.
  evidence:
  - reference: PMID:35503939
    reference_title: "Could Mental and Physical Exercise Alleviate Alzheimer's Disease?"
    supports: SUPPORT
    snippet: The aim of this review is to emphasize the importance of mental activity and aerobic physical exercise as one of the most important health-related activities which may delay the onset or slow down the progression of Alzheimer's dementia.
    explanation: The review highlights the importance of mental and physical exercise in potentially slowing the progression of Alzheimer's disease.
  - reference: PMID:32579499
    reference_title: "Risk Reduction and Prevention of Alzheimer's Disease: Biological Mechanisms of Diet."
    supports: SUPPORT
    snippet: Combined with the prevention of AD risk factors such as heart disease, diabetes, and with more recent evidence, microbiome dysfunction, there is a substantial foundation for diet as a modifiable risk factor and preventative measure for AD.
    explanation: The review suggests that a healthy diet can be a preventative measure for Alzheimer's disease, supporting the role of lifestyle modifications.
  - reference: PMID:37321363
    reference_title: "The role of lifestyle factors in cognitive health and dementia in oldest-old: A systematic review."
    supports: SUPPORT
    snippet: Results showed that eating a healthy diet with plenty of fruits and vegetables, and participation in leisure and physical activities may protect against cognitive decline and cognitive impairment among oldest-old regardless of the APOE genotype.
    explanation: The systematic review indicates that lifestyle factors such as diet and physical activities may protect against cognitive decline, supporting the statement.
  - reference: PMID:38129775
    reference_title: "Memory support training and lifestyle modifications to promote healthy aging in persons at risk for Alzheimer's disease: a digital application supported intervention (Brain Boosters)."
    supports: SUPPORT
    snippet: Rehabilitation approaches to support memory and behavioral/lifestyle interventions are recognized as promising strategies for preserving or improving cognitive health.
    explanation: The intervention described combines lifestyle modifications and memory support to improve cognitive health, supporting the statement.
  treatment_term:
    preferred_term: Lifestyle Therapy
    term:
      id: NCIT:C15900
      label: Lifestyle Therapy
clinical_trials:
- name: NCT03486938
  phase: PHASE_III
  status: COMPLETED
  description: >-
    HOPE4MCI: a randomized, double-blind, 78-week trial of AGB101, an
    extended-release low-dose (220 mg) levetiracetam formulation selected for its
    ability to normalize heightened hippocampal activity, in 164 participants with
    mild cognitive impairment due to Alzheimer disease. The largest test to date of
    the network-hyperexcitability model's therapeutic prediction. ClinicalTrials.gov
    registers the study under the combined designation PHASE2/PHASE3; the schema
    enum has no combined value, so `phase` is recorded here as PHASE_III, the
    pivotal-efficacy designation under which the trial is reported.
  target_phenotypes:
  - preferred_term: Memory impairment
    term:
      id: HP:0002354
      label: Memory impairment
  evidence:
  - reference: PMID:38356475
    reference_title: "The HOPE4MCI study: A randomized double-blind assessment of AGB101 for the treatment of MCI due to AD."
    supports: REFUTE
    evidence_source: HUMAN_CLINICAL
    snippet: "The estimated difference between arms is -0.10 (95% CI: -0.85, 0.58), which was not statistically significant."
    explanation: >-
      Negative on the primary endpoint (Clinical Dementia Rating Sum of Boxes) in
      an unselected population, which is the strongest single constraint on the
      claim that normalizing network activity is therapeutically actionable.
  - reference: PMID:38356475
    reference_title: "The HOPE4MCI study: A randomized double-blind assessment of AGB101 for the treatment of MCI due to AD."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "One hundred and sixty-four participants were randomized to placebo (n = 83) or AGB101 (n = 81), an extended-release formulation of low dose (220 mg) levetiracetam."
    explanation: Establishes the trial's size, randomization and intervention.
  - reference: clinicaltrials:NCT03486938
    reference_title: "A Multicenter, Randomized, Double-blind, Placebo-controlled Study Evaluating the Efficacy and Safety of AGB101on Slowing Progression of Mild Cognitive Impairment Due to Alzheimer's Disease"
    supports: SUPPORT
    evidence_source: OTHER
    snippet: "Participants will be randomized to receive placebo or AGB101 (220 mg), once daily for 78 weeks."
    explanation: >-
      Trial registration record establishing the intervention, dose and duration.
- name: NCT04063124
  phase: PHASE_I
  status: COMPLETED
  description: >-
    SToMP-AD: an open-label, proof-of-concept phase 1 trial of oral senolytic
    therapy (dasatinib plus quercetin) in five participants with early
    symptomatic Alzheimer disease, testing central nervous system penetrance,
    safety and feasibility rather than efficacy. The first clinical test of the
    cellular senescence model in Alzheimer disease.
  target_phenotypes:
  - preferred_term: Memory impairment
    term:
      id: HP:0002354
      label: Memory impairment
  - preferred_term: Impaired executive functioning
    term:
      id: HP:0033051
      label: Impaired executive functioning
  evidence:
  - reference: PMID:37679434
    reference_title: "Senolytic therapy in mild Alzheimer's disease: a phase 1 feasibility trial."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "In cerebrospinal fluid (CSF), D levels were detected in four participants (80%) ranging from 0.281 to 0.536 ml-1 with a CSF to plasma ratio of 0.422-0.919%; Q was not detected."
    explanation: >-
      Establishes central nervous system penetrance for dasatinib but not
      quercetin, which is the trial's primary question and a precondition for any
      brain-directed senolytic strategy.
  - reference: clinicaltrials:NCT04063124
    reference_title: "Pilot Study to Investigate the Safety and Feasibility of Senolytic Therapy to Modulate Progression of Alzheimer's Disease (SToMP-AD)"
    supports: SUPPORT
    evidence_source: OTHER
    snippet: "The purpose of this pilot study is to evaluate whether a combination of two drugs, dasatinib (D) and quercetin (Q)"
    explanation: >-
      Trial registration record establishing the intervention and pilot design.
- name: NCT05399888
  phase: PHASE_II
  status: ACTIVE_NOT_RECRUITING
  description: >-
    CELIA — a randomized, double-blind, placebo-controlled, parallel-group study
    of diranersen (BIIB080), a tau-targeting antisense oligonucleotide, in
    subjects with mild cognitive impairment due to Alzheimer disease or mild
    Alzheimer disease dementia. Diranersen or placebo is given by intrathecal
    injection every 12 or 24 weeks over a 76-week placebo-controlled period
    followed by a long-term extension, with the Clinical Dementia Rating-Sum of
    Boxes (CDR-SB) as the primary efficacy measure.
  target_phenotypes:
  - preferred_term: Dementia
    term:
      id: HP:0000726
      label: Dementia
  - preferred_term: Memory impairment
    term:
      id: HP:0002354
      label: Memory impairment
  evidence:
  - reference: clinicaltrials:NCT05399888
    reference_title: "A Randomized, Double-Blind, Placebo-Controlled, Parallel-Group Study to Assess the Efficacy, Safety, and Tolerability of BIIB080 in Subjects With Mild Cognitive Impairment Due to Alzheimer's Disease or Mild Alzheimer's Disease Dementia"
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "The main question researchers are trying to answer is if diranersen can slow the worsening of AD more than placebo."
    explanation: Defines CELIA's primary objective — whether the tau-targeting ASO diranersen slows Alzheimer disease progression versus placebo.
  - reference: clinicaltrials:NCT05399888
    reference_title: "A Randomized, Double-Blind, Placebo-Controlled, Parallel-Group Study to Assess the Efficacy, Safety, and Tolerability of BIIB080 in Subjects With Mild Cognitive Impairment Due to Alzheimer's Disease or Mild Alzheimer's Disease Dementia"
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "an injection into the fluid around the spinal cord (cerebrospinal fluid)"
    explanation: Confirms the intrathecal route of administration used to deliver diranersen to the CNS.
experimental_models:
- name: HMC3 human microglial micro-MKKS63 CRISPR-Cas9 knockout lines
  experimental_model_type: CELL_LINE
  publication: PMID:42736449
  organism:
    preferred_term: human
    term:
      id: NCBITaxon:9606
      label: Homo sapiens
  cell_types:
  - preferred_term: Microglia
    term:
      id: CL:0000129
      label: microglial cell
  cell_source: >-
    HMC3, an SV40-immortalized human microglial line (ATCC CRL-3304), transduced
    with lentiCas9-Blast and then with either of two single-guide RNAs disrupting
    the micro-MKKS63 smORF, or a non-targeting control guide. Marker-positive
    populations were recovered by fluorescence-activated cell sorting and editing
    was confirmed by PCR across the cut site followed by Sanger sequencing and
    TIDE analysis, so the lines are polyclonal edited populations rather than
    clonal knockouts.
  description: >-
    The design is what makes this model informative rather than merely
    convenient. The guides disrupt the smORF while preserving the main MKKS ORF,
    so a respiratory deficit in these lines is attributable to loss of the
    63-amino-acid microprotein and not to loss of the 570-amino-acid MKKS
    protein - a separation that no whole-locus knockout could make, and the
    reason this is the experiment that turns a coexpression correlation into a
    functional claim. Endogenous micro-MKKS63 is detectable by mass spectrometry
    in unedited HMC3 extracts with no peptide recovered from the canonical MKKS
    protein, mirroring the pattern in postmortem cortex, so the line reproduces
    the locus's translational imbalance before anything is edited.
  modeled_mechanisms:
  - target: Microglial Oxidative Phosphorylation Failure
    relationship: PERTURBS
    fidelity: LOW
    model_scale: CELLULAR
    description: >-
      Removes the microprotein and measures respiration, establishing that
      micro-MKKS63 is required for normal oxidative phosphorylation in human
      microglial cells.
    limitations: >-
      Fidelity is LOW because the model establishes the requirement without
      modeling the disease state. HMC3 is an immortalized line under standard
      culture oxygen and glucose, not aged cortical microglia in situ; the
      knockout abolishes the microprotein where Alzheimer cortex shows a partial
      decrease of roughly 4% (log2 fold change -0.063); and nothing downstream of
      respiration - phagocytosis, cytokine output, amyloid handling - was
      measured, so the model says nothing about whether the deficit matters for
      the disease.
    divergences:
    - divergence_type: POPULATION_MISMATCH
      materiality: QUALIFYING
      description: >-
        HMC3 is an immortalized fetal-derived microglial line maintained in
        culture. The node describes microglia in the aged human cortex, whose
        transcriptional state, metabolic substrate availability and disease
        context differ; cultured microglial lines are known to drift from the in
        situ microglial phenotype.
    - divergence_type: CAUSE_UNREPRESENTED
      materiality: QUALIFYING
      description: >-
        The knockout deletes the smORF outright. In Alzheimer cortex
        micro-MKKS63 is reduced rather than abolished, and whatever process
        drives that reduction - transcriptional, splicing-level, or
        translational - is absent from the model. The model therefore represents
        a complete loss standing in for a partial one, and represents none of
        its cause.
    readouts:
    - name: Basal oxygen consumption rate
      target: Microglial Oxidative Phosphorylation Failure
      direction: DECREASED
      description: >-
        Seahorse XF Mito Stress Test basal oxygen consumption rate, normalized to
        total protein per well by bicinchoninic acid assay, in each knockout line
        against the non-targeting control.
      biological_processes:
      - preferred_term: Oxidative phosphorylation
        term:
          id: GO:0006119
          label: oxidative phosphorylation
      interpretation: >-
        A reduction in resting respiratory flux is the most direct available
        measure of the node; both knockout lines reduced it independently, which
        is what distinguishes the result from a single-clone artifact.
      evidence:
      - reference: PMID:42736449
        reference_title: A microprotein atlas of the human frontal cortex in Alzheimer's disease.
        supports: SUPPORT
        evidence_source: IN_VITRO
        directness: DIRECT
        quote_role: PRIMARY_RESULT
        snippet: "Seahorse Mito Stress Test analysis showed significant reductions in basal oxygen consumption rate (OCR), ATP-linked OCR and maximal respiration across KO lines relative to nontargeting controls"
        explanation: >-
          Reports this measurement as the first of the three named reductions.
    - name: ATP-linked oxygen consumption rate
      target: Microglial Oxidative Phosphorylation Failure
      direction: DECREASED
      description: >-
        Oligomycin-sensitive fraction of oxygen consumption in the same Mito
        Stress Test.
      biological_processes:
      - preferred_term: Mitochondrial ATP synthesis
        term:
          id: GO:0042776
          label: proton motive force-driven mitochondrial ATP synthesis
      interpretation: >-
        Separates respiration coupled to ATP synthesis from proton leak, so a
        decrease here locates the deficit in useful energy output rather than in
        membrane integrity alone.
      evidence:
      - reference: PMID:42736449
        reference_title: A microprotein atlas of the human frontal cortex in Alzheimer's disease.
        supports: SUPPORT
        evidence_source: IN_VITRO
        directness: DIRECT
        quote_role: PRIMARY_RESULT
        snippet: "Seahorse Mito Stress Test analysis showed significant reductions in basal oxygen consumption rate (OCR), ATP-linked OCR and maximal respiration across KO lines relative to nontargeting controls"
        explanation: >-
          Reports this measurement as the second of the three named reductions.
    - name: Maximal respiration
      target: Microglial Oxidative Phosphorylation Failure
      direction: DECREASED
      description: >-
        Oxygen consumption after uncoupling, reporting the ceiling of the
        electron transport chain rather than its resting rate.
      biological_processes:
      - preferred_term: Oxidative phosphorylation
        term:
          id: GO:0006119
          label: oxidative phosphorylation
      interpretation: >-
        A lower ceiling indicates reduced respiratory capacity rather than
        reduced demand, which is the distinction that makes the basal deficit
        interpretable as impairment.
      evidence:
      - reference: PMID:42736449
        reference_title: A microprotein atlas of the human frontal cortex in Alzheimer's disease.
        supports: SUPPORT
        evidence_source: IN_VITRO
        directness: DIRECT
        quote_role: PRIMARY_RESULT
        snippet: "Seahorse Mito Stress Test analysis showed significant reductions in basal oxygen consumption rate (OCR), ATP-linked OCR and maximal respiration across KO lines relative to nontargeting controls"
        explanation: >-
          Reports this measurement as the third of the three named reductions.
    - name: Reserve respiratory capacity
      target: Microglial Oxidative Phosphorylation Failure
      direction: DECREASED
      description: >-
        Spare capacity between basal and maximal respiration, measured in the
        same assay but reported only in the figure, not in the results text.
      biological_processes:
      - preferred_term: Oxidative phosphorylation
        term:
          id: GO:0006119
          label: oxidative phosphorylation
      interpretation: >-
        The one readout in this panel where the two knockout lines disagree:
        reserve capacity falls significantly in the sgRNA1 line but not in the
        sgRNA2 line, and the results text names only the other three parameters.
        Recorded because a partially discordant readout is part of the result,
        and omitting it would make the panel look more uniform than it is.
      notes: >-
        Carries no evidence item. The discordance is legible only in the figure
        legend's per-line statistics, which is a block of test values rather than
        a sentence that asserts anything, so quoting it would attach a
        non-propositional snippet to a substantive claim.
    evidence:
    - reference: PMID:42736449
      reference_title: A microprotein atlas of the human frontal cortex in Alzheimer's disease.
      supports: SUPPORT
      evidence_source: IN_VITRO
      directness: DIRECT
      quote_role: PRIMARY_RESULT
      snippet: "we generated two independent polyclonal CRISPR–Cas9 KO HMC3 lines using single-guide RNAs (sgRNAs) disrupting the micro-MKKS63 smORF while preserving the mORF"
      explanation: >-
        Attests that this model is informative for the linked node: it is a
        microprotein-specific loss-of-function in the cell type the node names,
        with the main ORF left intact so the manipulation is attributable.
  evidence:
  - reference: PMID:42736449
    reference_title: A microprotein atlas of the human frontal cortex in Alzheimer's disease.
    supports: SUPPORT
    evidence_source: IN_VITRO
    directness: DIRECT
    quote_role: PRIMARY_RESULT
    snippet: "MS on HMC3 protein extracts identified robust micro-MKKS63 detection (PXD071500), with no peptides recovered from the canonical 570-aa MKKS protein, mirroring the pattern seen across ROSMAP brain tissue."
    explanation: >-
      Establishes that the unedited line reproduces the translational imbalance
      seen in human cortex, which is what makes it an appropriate background for
      a microprotein-specific knockout rather than an arbitrary cell line.
  notes: >-
    Raw mass spectrometry for the HMC3 work is deposited under pride:PXD071500,
    curated in this entry's datasets block.
datasets:
- accession: geo:GSE147528
  title: Molecular characterization of selectively vulnerable neurons in Alzheimer's Disease
  description: >-
    Single-nucleus RNA-seq of caudal entorhinal cortex and superior frontal gyrus
    from post-mortem brains spanning the progression of tau neurofibrillary
    pathology. The dataset from which RORB was identified as a marker of
    selectively vulnerable excitatory neurons.
  organism:
    preferred_term: Homo sapiens
    term:
      id: NCBITaxon:9606
      label: Homo sapiens
  data_type: SINGLE_CELL_RNA_SEQ
  sample_types:
  - preferred_term: entorhinal cortex
    term:
      id: UBERON:0002728
      label: entorhinal cortex
  - preferred_term: frontal cortex
    term:
      id: UBERON:0001870
      label: frontal cortex
  sample_count: 20
  publication: PMID:33432193
  notes: >-
    Underlies the Selective Vulnerability of RORB+ Entorhinal Excitatory Neurons
    node. Accession resolved against the GEO API with `just verify-datasets`.
    Re-deriving the RORB result from this dataset is not an independent test of it.
- accession: geo:GSE129308
  title: Molecular signatures underlying neurofibrillary tangle susceptibility in Alzheimer's disease
  description: >-
    Transcriptomes of single somas bearing neurofibrillary tangles compared with
    tangle-free somas isolated from the same human Alzheimer brains. The
    within-donor design controls for donor-level confounding when asking what
    distinguishes an aggregation-prone from an aggregation-resistant neuron.
  organism:
    preferred_term: Homo sapiens
    term:
      id: NCBITaxon:9606
      label: Homo sapiens
  data_type: SINGLE_CELL_RNA_SEQ
  sample_types:
  - preferred_term: cerebral cortex
    term:
      id: UBERON:0000956
      label: cerebral cortex
  sample_count: 27
  publication: PMID:35882228
  notes: >-
    The sharpest available human contrast for three separate questions curated in
    this entry: which neurons are selectively vulnerable, whether senescent cells
    are the tangle-bearing excitatory neurons reported in PMID:35531351, and
    whether necroptosis effectors are enriched in pathology-bearing neurons.
    Accession resolved against the GEO API with `just verify-datasets`. Note that
    the GEO record's own linked citation is PMID:41620473, a later reuse of the
    series (Dharshini et al. 2026 on layer 4 resilience); the originating study is
    Otero-Garcia et al. 2022 (PMID:35882228), whose title matches the series title.
    Do not take the GEO-linked PMID as the source publication here.
- accession: geo:GSE254205
  title: APOE4/4 is linked to damaging lipid droplets in Alzheimer's microglia
  description: >-
    Single-nucleus RNA-seq of human Alzheimer brain stratified by APOE genotype,
    defining the ACSL1-positive lipid-droplet microglial state that is most
    abundant in APOE4/4 carriers.
  organism:
    preferred_term: Homo sapiens
    term:
      id: NCBITaxon:9606
      label: Homo sapiens
  data_type: SINGLE_CELL_RNA_SEQ
  sample_types:
  - preferred_term: cerebral cortex
    term:
      id: UBERON:0000956
      label: cerebral cortex
  sample_count: 102
  publication: PMID:38480892
  notes: >-
    Source data for the Microglial Lipid Droplet Accumulation node. Because the
    donors are APOE-genotyped, the same nuclei also bear on whether the glial
    lipid phenotype is microglia-specific or shared with the oligodendrocyte
    lineage. Accession resolved against the GEO API with `just verify-datasets`.
- accession: geo:GSE174367
  title: Single-nucleus chromatin accessibility and transcriptomic characterization of Alzheimer's Disease
  description: >-
    Paired single-nucleus RNA-seq and ATAC-seq of human Alzheimer cortex, so
    expression changes can be checked against chromatin accessibility in the same
    cell types.
  organism:
    preferred_term: Homo sapiens
    term:
      id: NCBITaxon:9606
      label: Homo sapiens
  data_type: MULTI_OMICS
  sample_types:
  - preferred_term: cerebral cortex
    term:
      id: UBERON:0000956
      label: cerebral cortex
  sample_count: 230
  publication: PMID:34239132
  notes: >-
    The paired assay is what makes this dataset useful beyond replication: it
    allows a senescence or necroptosis expression claim to be tested for
    regulatory support rather than transcript abundance alone. Accession resolved
    against the GEO API with `just verify-datasets`.
- accession: geo:GSE157827
  title: Single-nucleus transcriptome analysis reveals dysregulation of angiogenic endothelial cells and neuroprotective glia in Alzheimer's disease
  description: >-
    Independent human cortical single-nucleus RNA-seq cohort with endothelial and
    glial coverage.
  organism:
    preferred_term: Homo sapiens
    term:
      id: NCBITaxon:9606
      label: Homo sapiens
  data_type: SINGLE_CELL_RNA_SEQ
  sample_types:
  - preferred_term: cerebral cortex
    term:
      id: UBERON:0000956
      label: cerebral cortex
  sample_count: 21
  publication: PMID:32989152
  notes: >-
    Curated as a replication cohort rather than for a specific node. Accession
    resolved against the GEO API with `just verify-datasets`.
- accession: geo:GSE138852
  title: A single-cell atlas of the human cortex reveals drivers of transcriptional changes in Alzheimer's disease in specific cell subpopulations
  description: >-
    Human entorhinal cortex single-nucleus atlas resolving cell-subpopulation
    transcriptional change in Alzheimer disease.
  organism:
    preferred_term: Homo sapiens
    term:
      id: NCBITaxon:9606
      label: Homo sapiens
  data_type: SINGLE_CELL_RNA_SEQ
  sample_types:
  - preferred_term: entorhinal cortex
    term:
      id: UBERON:0002728
      label: entorhinal cortex
  sample_count: 8
  publication: PMID:31768052
  notes: >-
    Small donor count (8); curated as a replication cohort, not as a primary
    discovery set. Accession resolved against the GEO API with
    `just verify-datasets`.
- accession: geo:GSE148822
  title: Distinct amyloid-b and tau associated microglia profiles in Alzheimer's disease
  description: >-
    Human single-nucleus microglial profiling separating amyloid-associated from
    tau-associated microglial states.
  organism:
    preferred_term: Homo sapiens
    term:
      id: NCBITaxon:9606
      label: Homo sapiens
  data_type: SINGLE_CELL_RNA_SEQ
  sample_types:
  - preferred_term: cerebral cortex
    term:
      id: UBERON:0000956
      label: cerebral cortex
  sample_count: 95
  publication: PMID:33609158
  notes: >-
    Directly relevant to the entry's repeated claim that glial responses are
    pathology-context-dependent rather than uniform, including the
    amyloid-versus-tau split curated on the Neuroinflammation node. Accession
    resolved against the GEO API with `just verify-datasets`.
- accession: geo:GSE160936
  title: Diverse human astrocyte and microglial transcriptional responses to Alzheimer's pathology
  description: >-
    Human astrocyte and microglial transcriptional responses to Alzheimer
    pathology, profiled across donors.
  organism:
    preferred_term: Homo sapiens
    term:
      id: NCBITaxon:9606
      label: Homo sapiens
  data_type: SINGLE_CELL_RNA_SEQ
  sample_types:
  - preferred_term: cerebral cortex
    term:
      id: UBERON:0000956
      label: cerebral cortex
  sample_count: 24
  publication: PMID:34767070
  notes: >-
    Glia-focused, and so a natural place to characterize astrocyte and microglial
    responses to Alzheimer pathology. It is not a place to settle the
    cell-identity dispute curated on the Senescent Cell Accumulation node: it
    profiles glia only, so it carries no neuronal comparator and cannot
    adjudicate a glia-versus-neuron contrast, which needs both cell classes
    measured in the same assay. It can test whether a glial senescence signature
    is present; it cannot test whether glia rather than excitatory neurons are
    the senescent population. Accession resolved against the GEO API with
    `just verify-datasets`.
- accession: synapse:syn26223298
  title: SEA-AD (Seattle Alzheimer's Disease Brain Cell Atlas)
  description: >-
    Multimodal atlas of 84 aged donors spanning the full range of Alzheimer
    neuropathology, with single-nucleus RNA-seq, ATAC-seq and multiome, MERFISH
    spatial transcriptomics, and quantitative neuropathology, resolved to 139
    molecular cell types across middle temporal gyrus and Brodmann area 9.
  organism:
    preferred_term: Homo sapiens
    term:
      id: NCBITaxon:9606
      label: Homo sapiens
  data_type: MULTI_OMICS
  sample_types:
  - preferred_term: temporal cortex
    term:
      id: UBERON:0016538
      label: temporal cortex
  - preferred_term: frontal cortex
    term:
      id: UBERON:0001870
      label: frontal cortex
  sample_count: 84
  publication: PMID:39402379
  notes: >-
    Access is split, and the distinction matters for anyone planning an analysis:
    the raw FASTQs are controlled (Sage/Synapse data use agreement), but the
    nuclei-by-gene matrices are openly downloadable and are also served through
    CELLxGENE and the Allen Brain Cell Atlas. `synapse:` accessions have no open
    per-record metadata API, so `just verify-datasets` reports this one as
    UNSUPPORTED rather than OK — that is the expected result, not a failure.
    This is the only human dataset here with enough nuclei per glial class to give
    the astrocyte, microglia and oligodendrocyte-progenitor senescence hypotheses a
    fair test against the excitatory-neuron claim.
- accession: synapse:syn60084804
  title: PsychAD_NPS
  description: >-
    PsychAD single-nucleus RNA-seq of dorsolateral prefrontal cortex from 1,494
    donors (about 6.3 million nuclei, 27 cell subclasses and 65 subtypes), with
    donor-level annotation of Alzheimer diagnosis, Braak stage, clinical dementia
    rating, cognitive resilience and neuropsychiatric symptoms including
    depression and agitation. PMID:42778763 analysed the 584 donors carrying
    Alzheimer-related phenotypes to score phenotype-associated cells.
  organism:
    preferred_term: Homo sapiens
    term:
      id: NCBITaxon:9606
      label: Homo sapiens
  data_type: SINGLE_CELL_RNA_SEQ
  sample_types:
  - preferred_term: dorsolateral prefrontal cortex
    term:
      id: UBERON:0009834
      label: dorsolateral prefrontal cortex
  sample_count: 1494
  publication: PMID:42778763
  notes: >-
    Distributed through the AD Knowledge Portal under its data use requirements.
    The title is the Synapse project name as registered (DataCite
    10.7303/syn60084804). As with SEA-AD, `synapse:` accessions have no open
    per-record metadata API, so `just verify-datasets` reports this one as
    UNSUPPORTED. Unlike SEA-AD it carries neuropsychiatric-symptom annotation,
    which makes it the dataset here that can separate depression-associated
    cell states from the Alzheimer signal within the same donors.
- accession: geo:GSE195455
  title: MEG3 activates necroptosis in human neuron xenografts modeling Alzheimer's disease
  description: >-
    RNA-seq of human H9-derived neurons seven days after transduction with either
    a control virus or a MEG3 lentiviral vector, isolating the transcriptional
    consequence of MEG3 gain of function in human neurons.
  organism:
    preferred_term: Homo sapiens
    term:
      id: NCBITaxon:9606
      label: Homo sapiens
  data_type: BULK_RNA_SEQ
  sample_count: 12
  publication: PMID:37708272
  notes: >-
    The causal arm of the necroptosis model, and open access. Its value is as a
    reference signature: the programme MEG3 overexpression induces in human neurons
    can be scored against human Alzheimer neurons in the atlases above, which is
    the actual discriminating test. Note the parent SuperSeries geo:GSE195458 also
    contains xenograft bulk RNA-seq (geo:GSE195457); the study deposited no human
    post-mortem cohort of its own, so its claim that MEG3 is raised in patients is
    not backed by a deposited human dataset.
- accession: geo:GSE178304
  title: Ageing-associated myelin dysfunction drives amyloid deposition in mouse models of Alzheimer's disease
  description: >-
    SuperSeries of mouse single-nucleus and bulk transcriptomics testing whether
    myelin dysfunction drives amyloid deposition, including grey versus white
    matter under myelin abnormality and microglial profiling on a 5xFAD background.
  organism:
    preferred_term: Mus musculus
    term:
      id: NCBITaxon:10090
      label: Mus musculus
  data_type: MULTI_OMICS
  sample_count: 30
  publication: PMID:37258678
  notes: >-
    Mouse only, and deliberately curated as such: this is the data establishing the
    causal direction (myelin dysfunction to amyloid) that no human observational
    dataset can establish, and it is the pairing partner for the human
    oligodendrocyte arm. The human counterpart from the APOE4-myelination work
    (PMID:36385529) is Synapse syn38120890 and is fully controlled-access with no
    GEO deposit, so SEA-AD's open oligodendrocyte matrices are the practical human
    arm instead. Accession resolved against the GEO API with `just verify-datasets`.
- accession: pride:PXD071500
  title: Identification of Microproteins in AD Using DIA Mass Spectrometry
  description: >-
    Raw mass spectrometry underlying the brain microprotein atlas: the HMC3 human
    microglial experiments and a small data-independent-acquisition dorsolateral
    prefrontal cortex pilot cohort. The HMC3 runs are the ones this entry depends
    on - they detect endogenous micro-MKKS63 in a human microglial line while
    recovering no peptide from the canonical 570-amino-acid MKKS protein, which is
    the in-cell counterpart of the translational imbalance seen in postmortem
    cortex.
  organism:
    preferred_term: Homo sapiens
    term:
      id: NCBITaxon:9606
      label: Homo sapiens
  data_type: PROTEOMICS
  sample_types:
  - preferred_term: dorsolateral prefrontal cortex
    term:
      id: UBERON:0009834
      label: dorsolateral prefrontal cortex
  publication: PMID:42736449
  notes: >-
    Underlies the Micro-MKKS63 Microprotein Depletion and Microglial Oxidative
    Phosphorylation Failure nodes. Accession resolved against the PRIDE API with
    `just verify-datasets`; PRIDE is not a reference-cached prefix, so there is no
    committed cache file and the check makes a live call on every run. The title
    is copied verbatim from the PRIDE record and names only the DIA arm, although
    the deposition also covers the HMC3 runs - the repository's own title is what
    belongs in this slot, so it is not corrected here. The large ROSMAP and MSBB
    cohorts that the atlas reanalyzed are not curated as separate dataset records
    on this entry: they are AMP-AD controlled-access Synapse resources
    (syn3219045, syn3159438) rather than open accessions, and ROSMAP single-nucleus
    data are already represented here by synapse:syn26223298.
discussions:
- discussion_id: gap_ad_amyloid_tau_glia_mitochondria_ordering_resilience
  prompt: >-
    Which temporal causal ordering among amyloid-beta accumulation, tau spread,
    microglial/astrocytic activation, mitochondrial quality-control failure, and
    synaptic/cognitive decline best explains stage-specific Alzheimer disease
    progression and cognitive resilience?
  kind: KNOWLEDGE_GAP
  status: OPEN
  attaches_to:
  - pathophysiology#Amyloid Plaque Formation
  - pathophysiology#Neurofibrillary Tangle Formation
  - pathophysiology#Neuroinflammation
  - pathophysiology#Mitochondrial Quality-Control Failure
  - pathophysiology#Synaptic Dysfunction
  rationale: >-
    Amyloid plaques, tau tangles, glial activation, mitochondrial stress, and
    cognitive decline are all established Alzheimer disease features, but the
    disease entry should not imply a single fixed ordering across all stages.
    Current evidence supports amyloid-triggered tau and glial pathways, reciprocal
    inflammasome-to-amyloid amplification, protective TREM2-dependent microglial
    containment of amyloid-associated tau spread, mitochondrial quality-control
    effects on amyloid/tau/inflammation, and cognitive reserve that can decouple
    pathology burden from symptoms. The open curation question is which edges are
    upstream drivers, downstream amplifiers, parallel state changes, or resilience
    modifiers in preclinical, prodromal, and dementia-stage Alzheimer disease.
  proposed_experiments:
  - experiment_id: exp_ad_longitudinal_pathology_glia_mitophagy_resilience
    name: Longitudinal amyloid-tau-glia-mitophagy-resilience ordering cohort
    description: >-
      Follow amyloid-negative, amyloid-positive tau-negative, and amyloid-positive
      tau-positive older adults across preclinical, prodromal, and dementia
      stages with amyloid PET, tau PET, plasma and CSF glial biomarkers, candidate
      mitochondrial/mitophagy biomarkers, structural and functional imaging, and
      cognitive reserve measures. The study should test whether glial and
      mitochondrial changes precede tau spread, follow amyloid/tau deposition,
      amplify clinical decline, or mark resilient compensation.
    experiment_type:
      preferred_term: longitudinal multimodal cohort study
    assays:
    - preferred_term: amyloid positron emission tomography
    - preferred_term: tau positron emission tomography
    - preferred_term: cerebrospinal fluid biomarker profiling
    - preferred_term: plasma biomarker profiling
    - preferred_term: neuropsychological assessment
    readouts:
    - name: Glial activation temporal precedence
      target: pathophysiology#Neuroinflammation
      description: >-
        Plasma GFAP, CSF sTREM2, inflammasome, cytokine, and complement readouts
        modeled relative to amyloid PET conversion, tau PET spread, and cognitive
        change.
      assays:
      - preferred_term: plasma GFAP measurement
      - preferred_term: CSF soluble TREM2 measurement
      - preferred_term: complement biomarker profiling
      direction: POSITIVE
      interpretation: >-
        Glial-marker elevation before regional tau spread would support an
        upstream amplifier role; elevation only after amyloid/tau burden would
        support downstream response.
    - name: Mitochondrial quality-control state
      target: pathophysiology#Mitochondrial Quality-Control Failure
      description: >-
        Mitophagy and mitochondrial stress markers assessed against amyloid, tau,
        glial, and cognitive trajectories.
      biological_processes:
      - preferred_term: Mitophagy
        term:
          id: GO:0000422
          label: autophagy of mitochondrion
      direction: NEGATIVE
      interpretation: >-
        Declining mitophagy markers before amyloid/tau or glial acceleration
        would support mitochondrial quality-control failure as an upstream or
        parallel driver rather than only a late downstream injury marker.
    - name: Pathology-symptom decoupling
      target: phenotypes#Memory Loss
      description: >-
        Cognitive trajectories stratified by education, occupation, engagement,
        social-network, and functional-network measures among participants with
        comparable amyloid, tau, glial, and mitochondrial marker burden.
      assays:
      - preferred_term: neuropsychological assessment
      - preferred_term: functional connectivity imaging
      direction: NEGATIVE
      interpretation: >-
        Preserved cognition despite high pathology burden would support cognitive
        resilience as a modifier that should be curated separately from the core
        amyloid-tau-glia pathophysiology chain.
    controls:
    - name: Amyloid-negative age-matched controls
      description: Older adult participants without amyloid PET positivity at baseline.
    - name: Pathology-burden matched resilience strata
      description: >-
        Participants matched for amyloid/tau/glial burden but differing in
        cognitive reserve and longitudinal cognitive decline.
    decision_criterion: >-
      The ordering model should be revised according to which biomarker changes
      temporally precede tau spread and cognitive decline after adjustment for
      pathology burden and resilience variables; a resilience discussion should
      remain separate if cognitive trajectories decouple from amyloid, tau, glial,
      and mitochondrial burden.
    would_support:
    - pathophysiology#Neuroinflammation
    - pathophysiology#Mitochondrial Quality-Control Failure
    - mechanistic_hypotheses#neuroimmune_glial_amplification_model
    - mechanistic_hypotheses#autophagy_lysosomal_clearance_model
    evidence:
    - reference: PMID:37924152
      reference_title: "Clinical and biological relevance of glial fibrillary acidic protein in Alzheimer's disease."
      supports: SUPPORT
      evidence_source: HUMAN_CLINICAL
      snippet: "Brain amyloid was strongly associated with plasma GFAP and ptau-181 and to a lesser extent with plasma NfL."
      explanation: >-
        Demonstrates feasibility and relevance of measuring astrocytic GFAP
        alongside amyloid, tau, and neurodegeneration markers in human cohorts.
  evidence:
  - reference: PMID:36911732
    reference_title: "The effects of microglia-associated neuroinflammation on Alzheimer's disease."
    supports: SUPPORT
    evidence_source: OTHER
    snippet: "the temporal and spatial changes in microglial phenotype, the interactions among microglia, Aβ, tau, and neurons"
    explanation: >-
      Review-level support that the unresolved issue is temporal and spatial
      ordering among microglia, amyloid-beta, tau, and neuronal injury rather than
      the mere presence of those processes.
  - reference: PMID:37308616
    reference_title: "Functional roles of reactive astrocytes in neuroinflammation and neurodegeneration."
    supports: SUPPORT
    evidence_source: OTHER
    snippet: "these proteins do not act in isolation but form part of a pathological network."
    explanation: >-
      Supports representing amyloid-beta and tau as embedded in a multi-cellular
      pathological network that includes reactive astrocyte states.
  - reference: PMID:30742114
    reference_title: "Mitophagy inhibits amyloid-β and tau pathology and reverses cognitive deficits in models of Alzheimer's disease."
    supports: SUPPORT
    evidence_source: OTHER
    snippet: "Our findings suggest that impaired removal of defective mitochondria is a pivotal event in AD pathogenesis and that mitophagy represents a potential therapeutic intervention."
    explanation: >-
      Supports including mitochondrial quality-control failure in the ordering
      gap rather than treating mitochondrial stress only as a generic oxidative
      consequence.
  - reference: PMID:21643921
    reference_title: "Cognitive reserve and its implications for rehabilitation and Alzheimer's disease."
    supports: SUPPORT
    evidence_source: OTHER
    snippet: "cognitive reserve is not a fixed factor, but can be continuously modified by life experiences, even when the brain is already affected by neuropathology."
    explanation: >-
      Supports tracking cognitive resilience as a separate modifier because
      clinical status can be shaped by reserve even in the presence of
      neuropathology.
  - reference: PMID:42778763
    reference_title: "AI-based characterization of Alzheimer's disease phenotypes from population-scale single-cell data."
    supports: SUPPORT
    evidence_source: COMPUTATIONAL
    quote_role: PRIMARY_RESULT
    snippet: "reactive astrocyte subtypes with altered neuroprotective and neurotoxic gene expression that likely confer cognitive resilience"
    explanation: >-
      Population-scale single-nucleus evidence that cognitive resilience has a
      glial as well as a neuronal cellular correlate. In the full text, among
      late-Braak donors the Astro_PLSCR1 subtype of cognitively resilient donors
      upregulated the neuroprotective reactive-astrocyte marker EMP1, while
      cognitively impaired donors showed higher FKBP5. The authors' own wording
      ("likely confer") marks this as an association from computational
      phenotype scoring, not a tested protective mechanism, so it bears on the
      resilience-modifier half of this gap rather than settling causal order.
  posed_date: "2026-06-03T00:00:00Z"
  notes: >-
    Seeded for issue 3661. This discussion deliberately separates established
    pathologies from the unresolved causal ordering among amyloid, tau,
    microglial/astrocytic inflammation, mitochondrial quality control, and
    cognitive resilience.
- discussion_id: disc_hsv1_causality
  prompt: >-
    Is HSV-1 reactivation in RORB+ glutamatergic neurons CAUSAL for their
    selective loss in Alzheimer's disease, or a marker of cells already
    destined to die?
  kind: KNOWLEDGE_GAP
  status: OPEN
  attaches_to:
  - pathophysiology#HSV-1 Reactivation in RORB+ Glutamatergic Neurons
  rationale: >-
    Cross-sectional post-mortem evidence establishes the association but
    cannot resolve causal direction. The answer determines whether antivirals
    (or pre-emptive HSV-1 suppression) are candidate disease-modifying
    therapies, or merely a downstream readout of cells already committed to
    neurodegeneration.
  proposed_experiments:
  - experiment_id: exp_hsv1_organoid_reactivation_causality
    name: HSV-1 reactivation perturbation in human cortical neuron-glia organoids
    description: >-
      Introduce controlled HSV-1 latency/reactivation into human cortical
      neuron-glia organoids enriched for glutamatergic neurons, then compare
      RORB+ neuron survival and neuroinflammatory state against matched mock
      and antiviral-rescue controls.
    experiment_type:
      preferred_term: controlled perturbation experiment
    model_systems:
    - name: Human cortical neuron-glia organoid
      description: >-
        Human pluripotent-stem-cell-derived cortical organoid with
        glutamatergic neurons and glial support cells, used to test whether
        viral reactivation precedes neuronal vulnerability in a disease-relevant
        human cellular context.
      experimental_model_type: ORGANOID
      namo_type: namo:Organoid
      organism:
        preferred_term: human
        term:
          id: NCBITaxon:9606
          label: Homo sapiens
      cell_types:
      - preferred_term: RORB+ glutamatergic neuron
        term:
          id: CL:0000679
          label: glutamatergic neuron
      - preferred_term: microglial cell
        term:
          id: CL:0000129
          label: microglial cell
    perturbations:
    - name: Induced HSV-1 reactivation
      target: pathophysiology#HSV-1 Reactivation in RORB+ Glutamatergic Neurons
      description: >-
        Establish latent HSV-1 infection and trigger controlled reactivation to
        test whether reactivation is sufficient to drive selective RORB+
        glutamatergic-neuron loss.
      biological_processes:
      - preferred_term: viral process
        term:
          id: GO:0016032
          label: viral process
    readouts:
    - name: RORB+ glutamatergic neuron survival
      target: pathophysiology#HSV-1 Reactivation in RORB+ Glutamatergic Neurons
      description: >
        Single-cell and imaging readout of whether RORB+ glutamatergic neurons
        are selectively depleted after reactivation.
      assays:
      - preferred_term: single-cell transcriptomic profiling
      - preferred_term: high-content imaging
      direction: NEGATIVE
      interpretation: >-
        Selective loss after reactivation, reduced by antiviral rescue, would
        support HSV-1 reactivation as a causal injury mechanism.
    - name: Neuroinflammatory activation
      target: pathophysiology#Neuroinflammation
      description: >
        Cytokine, glial activation, and stress-response measurements to test
        whether viral reactivation creates an inflammatory state upstream of
        neuron loss.
      biological_processes:
      - preferred_term: inflammatory response
        term:
          id: GO:0006954
          label: inflammatory response
      assays:
      - preferred_term: multiplex cytokine profiling
      - preferred_term: single-cell transcriptomic profiling
      direction: POSITIVE
    controls:
    - name: Mock-reactivated isogenic organoids
      description: Matched organoids handled identically without HSV-1 reactivation.
    - name: Antiviral rescue arm
      description: Reactivated organoids treated with antiviral suppression before readout.
    decision_criterion: >-
      HSV-1 reactivation should temporally precede selective RORB+ neuron loss,
      and antiviral rescue should attenuate both viral signal and neuronal loss.
    would_support:
    - pathophysiology#HSV-1 Reactivation in RORB+ Glutamatergic Neurons
    would_refute:
    - pathophysiology#HSV-1 Reactivation in RORB+ Glutamatergic Neurons
    evidence:
    - reference: PMID:42094473
      reference_title: "Resolving human neuronal herpesvirus reactivation via petabase-scale association studies."
      supports: SUPPORT
      evidence_source: HUMAN_CLINICAL
      snippet: "Integrative single-nucleus analyses resolve direct evidence of HSV-1 expression in RORB+ glutamatergic neurons, implicating viral reactivation in a neuronal population progressively lost during dementia."
      explanation: >-
        Human post-mortem evidence motivates a perturbational organoid test
        that can distinguish causal reactivation from end-stage association.
  evidence:
  - reference: PMID:42094473
    reference_title: "Resolving human neuronal herpesvirus reactivation via petabase-scale association studies."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "Integrative single-nucleus analyses resolve direct evidence of HSV-1 expression in RORB+ glutamatergic neurons, implicating viral reactivation in a neuronal population progressively lost during dementia."
    explanation: >-
      Establishes the association in post-mortem human brain but the
      cross-sectional design cannot distinguish whether HSV-1 reactivation
      drives the loss of RORB+ neurons or merely marks neurons destined to
      die.
  - reference: PMID:41405855
    reference_title: "Valacyclovir Treatment of Early Symptomatic Alzheimer Disease: The VALAD Randomized Clinical Trial."
    supports: REFUTE
    evidence_source: HUMAN_CLINICAL
    snippet: "greater cognitive worsening with valacyclovir than placebo"
    explanation: >-
      The phase II VALAD randomized controlled trial (JAMA 2026) of
      valacyclovir 4 g/day vs placebo over 78 weeks in HSV-seropositive early
      symptomatic AD found no cognitive benefit and, if anything, greater
      decline in the treatment arm (between-group ADAS-Cog difference 3.93, 95%
      CI 1.03-6.83, P=.01). Direct interventional evidence that anti-HSV
      therapy does not help ESTABLISHED symptomatic AD — refuting the treatment
      corollary while leaving the upstream reactivation hypothesis and the
      pre-symptomatic-prevention question open (the same treatment-vs-prevention
      split seen in the zoster-vaccine natural experiments).
  - reference: PMID:41467972
    reference_title: "Anti-herpetic treatment reduces dementia risk: A systematic review and meta-analysis."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "diagnosed and treated versus diagnosed but untreated (aHR=0.77, 95%"
    explanation: >-
      Meta-analysis of 14 cohort studies in non-demented adults >=50 covering
      HSV and VZV finds anti-herpetic treatment associated with lower incident
      dementia (aHR 0.77 for diagnosed-and-treated vs diagnosed-but-untreated).
      Observational prevention-side signal complementing the null VALAD
      treatment trial and bridging to the zoster-vaccine evidence (VZV
      included).
  - reference: PMID:41779765
    reference_title: "Antiviral Treatment of Herpes Simplex Virus Decreases the Risk of Alzheimer's Disease and Dementia."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "significantly reduced the risk of AD (RR 0.87; 95% CI 0.73-0.92) and dementia"
    explanation: >-
      Propensity-matched TriNetX retrospective cohort (231,277 matched pairs)
      finds antiviral therapy for oral/mucocutaneous HSV associated with lower
      subsequent risk of AD (RR 0.87) and dementia (RR 0.83). A second
      observational prevention-side signal; confounding by indication and
      healthy-user bias cannot be excluded.
  posed_date: "2026-05-16T00:00:00Z"
  notes: >-
    Seeded from PR 2789 alongside the new HSV-1 pathophysiology node. See
    the broader discussion of how `discussions:` (this layer) relates to a
    proposed structural `knowledge_gaps:` layer in
    https://github.com/monarch-initiative/dismech/issues/2617#issuecomment-4467637580
    The interventional VALAD (PMID:41405855) and observational anti-herpetic
    prevention evidence (PMID:41467972, PMID:41779765) were surfaced by mining
    the existing OpenScientist deep-research report for this hypothesis
    (kb/hypotheses/Alzheimer_Disease/hsv1_reactivation_model/openscientist.md;
    two fresh OpenScientist re-runs both hit the provider's 2 h job ceiling and
    produced no output). Each PMID and snippet was independently verified
    against the fetched abstract before curation, per the deep-research
    anti-hallucination SOP; in the process one DR-reported figure was found
    misattributed (the report cited OR=1.46 for the HSV-AD meta-analysis
    PMID:40934136, whose actual abstract reports OR=1.32 case-control / 20%
    cohort), so that number was not used.
- discussion_id: gap_ad_tcell_antigen_identity
  prompt: >-
    Do the clonally expanded, cytotoxic CD8+ T cells that drive tau-associated
    neurodegeneration recognize a tau-derived (or other brain) antigen, or are
    they antigen-independent (e.g., age-expanded virtual-memory CD8 T cells)?
  kind: KNOWLEDGE_GAP
  status: OPEN
  attaches_to:
  - pathophysiology#Adaptive Immune T Cell Response to Tau Pathology
  rationale: >-
    The adaptive_immune_tcell_model reads "restricted TCR clonality" as evidence
    of an antigen-driven CD8 response, and cDC1 cross-presentation is now
    established as the priming route (PMID:41890046) — but the recognized
    antigen(s) remain unidentified. Two findings keep the antigen-driven
    interpretation open: aging independently expands antigen-independent
    virtual-memory CD8 T cells that are clonal and cytotoxic without foreign-antigen
    priming (PMID:42432776); and the clonally expanded CD8 TCRs recovered from
    human AD CSF were shown to be specific to Epstein-Barr virus antigens rather
    than tau (PMID:31915375). Distinguishing a tau-antigen-driven response from an
    antigen-independent (or virally-primed) bystander clonal expansion is the
    single most important unresolved edge in this model and gates whether
    tau-antigen-directed tolerization is even a coherent therapeutic strategy.
  proposed_experiments:
  - experiment_id: exp_ad_tcell_antigen_discovery
    name: Brain-infiltrating CD8 clone antigen-discovery screen
    description: >-
      Isolate brain- and CSF-infiltrating CD8+ T cell clones from tauopathy mice
      and human AD/primary-tauopathy tissue, pair single-cell TCR sequencing with
      peptide-MHC tetramer and pMHC-screen reactivity assays against tau-derived
      peptides, candidate self-antigens, and viral (EBV) antigens, and fate-map
      virtual-memory versus conventional-memory origin.
    experiment_type:
      preferred_term: antigen-specificity T cell reactivity screen
    assays:
    - preferred_term: single-cell T cell receptor sequencing
    - preferred_term: peptide-MHC tetramer binding assay
    readouts:
    - name: Dominant-clone antigen specificity
      target: pathophysiology#Adaptive Immune T Cell Response to Tau Pathology
      description: >-
        Fraction of dominant brain-infiltrating CD8 clones with demonstrable
        reactivity to tau-derived peptides versus viral/self antigens versus no
        detectable cognate antigen (virtual-memory phenotype).
      direction: POSITIVE
      interpretation: >-
        Dominant clones reactive to tau peptides would support a tau-antigen-driven
        response; predominance of viral-reactive or antigen-inexperienced
        virtual-memory clones would reframe the arm as cytotoxic-lymphocyte- rather
        than tau-antigen-driven.
  posed_date: "2026-07-11T00:00:00Z"
  notes: >-
    Seeded from an OpenScientist hypothesis deep-research run on
    adaptive_immune_tcell_model; all cited PMIDs independently verified against
    their abstracts before curation.
- discussion_id: gap_ad_tcell_ifng_pd1_valence_and_source
  prompt: >-
    Is IFN-gamma/PD-1 signaling net-neurotoxic or net-protective in Alzheimer
    disease, and how much of the IFN-gamma effect attributed to clonal CD8 T cells
    is actually contributed by NK cells?
  kind: KNOWLEDGE_GAP
  status: OPEN
  attaches_to:
  - pathophysiology#Adaptive Immune T Cell Response to Tau Pathology
  - pathophysiology#Neuroinflammation
  rationale: >-
    The model predicts IFN-gamma/PD-1 blockade is neuroprotective, based on mouse
    tauopathy where inhibiting IFN-gamma or PD-1 ameliorates brain atrophy
    (PMID:36890231). But the therapeutic valence reverses by pathology: in an
    amyloid context, PD-1 checkpoint blockade drives an IFN-gamma-dependent myeloid
    response that clears Aβ and improves cognition (PMID:26779813), so a universal
    "block IFN-gamma/PD-1" prescription is not supported. Separately, IFN-gamma is
    not CD8-exclusive — NK cells produce IFN-gamma/TNF-alpha that negatively
    correlates with cognition in AD (PMID:11268360) — so some effects attributed to
    clonal CD8 may be innate-lymphoid in origin. Resolving the context-dependent
    valence and the true cellular source of pathogenic IFN-gamma is required before
    the node's therapeutic predictions can be curated as directional.
  proposed_experiments:
  - experiment_id: exp_ad_ifng_source_valence_matched_models
    name: Cell-type-resolved IFN-gamma modulation across tau vs amyloid models
    description: >-
      Apply matched IFN-gamma/PD-1 modulation and cell-type-specific conditional
      Ifng deletion (CD8-restricted vs NK-restricted vs total) across tau-only,
      amyloid-only, and mixed mouse models, reading neurodegeneration and cognition
      to separate the direction of the IFN-gamma effect from its cellular source.
    experiment_type:
      preferred_term: comparative conditional-knockout intervention study
    assays:
    - preferred_term: conditional gene knockout
    - preferred_term: neuropsychological assessment
    readouts:
    - name: IFN-gamma source-and-valence effect on neurodegeneration
      target: pathophysiology#Adaptive Immune T Cell Response to Tau Pathology
      description: >-
        Change in atrophy/neuronal loss when IFN-gamma is deleted from CD8 versus
        NK cells, and when IFN-gamma/PD-1 is blocked, in tau-only versus amyloid-only
        contexts.
      direction: NEGATIVE
      interpretation: >-
        Protection only with CD8-restricted IFN-gamma loss in the tau context (with
        harm or no benefit in amyloid) would confirm a tau-scoped, CD8-driven
        pathogenic axis; protection with NK-restricted loss would reassign part of
        the effect to innate lymphocytes.
  posed_date: "2026-07-11T00:00:00Z"
  notes: >-
    Seeded from an OpenScientist hypothesis deep-research run on
    adaptive_immune_tcell_model; all cited PMIDs independently verified against
    their abstracts before curation.
- discussion_id: mismatch_ad_tcell_arm_not_inherited_risk
  prompt: >-
    Does the cytotoxic CD8+ T-cell arm that is functionally required for
    tau-mediated neurodegeneration in mouse tauopathy actually contribute to human
    Alzheimer disease, given that adaptive-immune/CD8/IFN-gamma/MHC-I genes carry
    no detectable inherited AD risk while innate-microglial genes dominate the
    human genetic-association signal?
  kind: HUMAN_MODEL_MISMATCH
  status: OPEN
  attaches_to:
  - pathophysiology#Adaptive Immune T Cell Response to Tau Pathology
  rationale: >-
    In mouse tauopathy the adaptive CD8+ T-cell arm is functionally required —
    T-cell depletion, IFN-gamma/PD-1 inhibition, and cDC1 ablation are each
    neuroprotective (PMID:36890231, PMID:41890046) — yet human inherited AD risk
    shows no adaptive-immune signal. A gene-set enrichment of Alzheimer disease
    genetic-association targets from the Open Targets Platform (disease
    MONDO:0004975; the top 200 of 13,367 targets ranked by the genetic_association
    datatype score; retrieved and independently reproduced on 2026-07-11) finds a
    curated 35-gene innate-microglial set strongly over-represented (18/35 overlap;
    odds ratio 115.1; Fisher one-sided p = 1.79e-27; the overlap includes TREM2,
    PLCG2, INPP5D, ABCA7, BIN1, SORL1, CR1, CD33, APOE, CLU, MS4A6A, PICALM),
    while a 46-gene adaptive-immune/CD8/IFN-gamma/MHC-I set is not enriched at all
    (0/46 overlap; odds ratio 0.0; p = 1.0). The null adaptive signal is robust
    across top-N cutoffs (100-1000) and gene-background choices (19,000/20,000),
    and adaptive-immune genes are absent even from the top 2,000 ranked targets.
    This coheres with the published human-genetics consensus that microglia and
    innate immunity are the central inherited determinants of AD susceptibility
    (PMID:41888907) and that AD risk loci act specifically through microglial
    eQTLs (PMID:40670704). The mismatch is mechanistically meaningful: it implies
    the pathogenic adaptive T-cell arm — potent in mouse tauopathy — is most
    plausibly an acquired/downstream response to tau pathology rather than an
    upstream heritable driver of AD, reconciling the mouse functional data with the
    absence of adaptive-immune inherited risk and cohering with the model's own
    caveats (EBV-specific CSF TCRs, age-expanded antigen-independent virtual-memory
    CD8). Caveats: this is a coarse gene-set enrichment (not colocalization or
    fine-mapping); most AD GWAS are amyloid-weighted diagnosis studies whereas the
    model is tau-scoped; and the LD-complex MHC region may be under-represented at
    target-level association scores.
  proposed_experiments:
  - experiment_id: exp_ad_celltype_coloc_adaptive_vs_microglial
    name: Cell-type-resolved colocalization of AD GWAS loci with CD8 T-cell vs microglial QTLs
    description: >-
      Move beyond coarse enrichment to statistically colocalize Alzheimer disease
      GWAS loci with CD8+ T-cell versus microglial eQTL/caQTL maps, directly testing
      whether any adaptive-immune regulatory variants influence inherited AD risk
      rather than acting only downstream of established tau pathology.
    experiment_type:
      preferred_term: GWAS-QTL colocalization analysis
    assays:
    - preferred_term: expression quantitative trait locus colocalization
    - preferred_term: chromatin accessibility quantitative trait locus mapping
    readouts:
    - name: Adaptive-immune vs microglial colocalization signal
      target: pathophysiology#Adaptive Immune T Cell Response to Tau Pathology
      description: >-
        Number and posterior probability of AD loci colocalizing with CD8+ T-cell
        QTLs versus microglial QTLs.
      direction: POSITIVE
      interpretation: >-
        Robust colocalization of AD loci with CD8 T-cell QTLs would revise the
        downstream-only reading and support an inherited adaptive-immune
        contribution; continued microglia-exclusive colocalization would confirm the
        acquired/downstream interpretation of the T-cell arm.
  - experiment_id: exp_ad_tau_stratified_genetic_enrichment
    name: Tau-endophenotype-stratified genetic enrichment
    description: >-
      Repeat the adaptive-immune-vs-microglial enrichment against genetic-association
      scores derived from tau-PET or CSF-tau endophenotype GWAS rather than
      amyloid-weighted clinical-diagnosis GWAS, the fairer test of the tau-scoped
      model.
    experiment_type:
      preferred_term: gene-set enrichment analysis
    assays:
    - preferred_term: genome-wide association study
    readouts:
    - name: Adaptive-immune enrichment under tau-endophenotype GWAS
      target: pathophysiology#Adaptive Immune T Cell Response to Tau Pathology
      description: >-
        Odds ratio and significance of adaptive-immune gene-set enrichment when
        targets are ranked by tau-endophenotype genetic association.
      direction: POSITIVE
      interpretation: >-
        Emergence of adaptive-immune enrichment specifically under tau-endophenotype
        GWAS (absent under amyloid-weighted diagnosis GWAS) would indicate the
        coarse diagnosis-GWAS null understated a tau-specific inherited contribution.
  posed_date: "2026-07-11T00:00:00Z"
  notes: >-
    Seeded from a scoped OpenScientist code-execution deep-research run on
    adaptive_immune_tcell_model that queried the Open Targets Platform GraphQL API
    and computed the enrichment. The reported numbers (13,367 total targets; top-200
    genetic_association score range 0.461-0.954; first-15 targets PSEN1, APP, CR1,
    PSEN2, APOE, TREM2, PLCG2, MS4A6A, HFE, EPHA1, APH1B, ACE, H2BC4, BIN1, MPO;
    innate OR 115.1 p 1.79e-27; adaptive OR 0.0 p 1.0) were independently reproduced
    to the digit by re-running the same GraphQL query and Fisher exact test before
    curation. Query: disease(efoId ->MONDO_0004975) associatedTargets
    orderByScore:"genetic_association", top 200 with score > 0; background 20,000
    protein-coding genes; one-sided Fisher exact per gene set.
- discussion_id: mismatch_ad_ldir_dose_regimen_translation
  prompt: >-
    Do preclinical reports that low-dose ionizing radiation below 100 mSv, at
    selected doses and regimens, reduce Alzheimer-relevant amyloid burden and
    neuroinflammation or improve cognition in humans without delayed harm, and
    where does the response change from potentially protective to neurotoxic?
  kind: HUMAN_MODEL_MISMATCH
  status: OPEN
  attaches_to:
  - pathophysiology#Amyloid Plaque Formation
  - pathophysiology#Neuroinflammation
  - pathophysiology#Synaptic Dysfunction
  - phenotypes#Memory Loss
  rationale: >-
    A 2026 review describes potentially protective effects only at specific
    low-dose regimens and explicitly reports no consensus on the dual mechanisms,
    safety threshold, or long-term outcomes because of small samples, ethical
    constraints, and inter-model variability. The abstract synthesizes
    preclinical evidence and prospective applications rather than establishing a
    reproducible human Alzheimer disease response. This entry therefore does not
    assert a protective environmental edge or add low-dose-radiation mechanism
    nodes. A single descriptor modifier would also be misleading: modifiers encode
    one biological state, not a dose- and regimen-conditioned reversal of effect.
    If the response is later established, detrimental and protective arms should
    be represented separately with explicit exposure context and arm-specific
    evidence. The question is kept at entry level because the reported reduction
    in amyloid-beta deposition maps directly to the existing Alzheimer pathograph;
    a reusable hormesis module would require a conserved causal chain, a key
    conformance target, and validated recurrence across at least two disorders,
    none of which this single review establishes.
  evidence:
  - reference: PMID:42551708
    reference_title: "Low-dose ionizing radiation and cognitive function: evidence, mechanisms, and controversies."
    supports: SUPPORT
    evidence_source: OTHER
    snippet: >-
      Nevertheless, emerging evidence indicates that LDIR administered at specific doses and regimens may exert neuroprotective effects via hormesis: it reduces Aβ deposition, promotes anti-inflammatory microglial polarization, and enhances hippocampal neurogenesis, thereby ameliorating cognitive impairment.
    explanation: >-
      The review motivates the protective arm of the question but limits it to
      selected doses and regimens; it does not establish a human Alzheimer disease
      mechanism or a safe intervention threshold.
  - reference: PMID:42551708
    reference_title: "Low-dose ionizing radiation and cognitive function: evidence, mechanisms, and controversies."
    supports: SUPPORT
    evidence_source: OTHER
    snippet: >-
      Constrained by ethical considerations, sample size limitations, and inter-model variability, no consensus has been reached regarding its dual mechanisms, safety thresholds, and long-term outcomes.
    explanation: >-
      Directly supports retaining this as an open model-to-human mismatch rather
      than asserting either a protective pathograph or a general hormesis module.
  proposed_experiments:
  - experiment_id: exp_ad_ldir_dose_regimen_translation
    name: Harmonized low-dose ionizing-radiation dose-response study
    description: >-
      Apply a harmonized absorbed-dose, dose-rate, radiation-quality, and
      fractionation matrix across human iPSC-derived hippocampal
      neuron-astrocyte-microglia cultures and an Alzheimer-relevant in vivo model,
      then compare the prespecified biomarker directions with longitudinal data
      from consented cohorts receiving clinically indicated imaging or occupational
      monitoring. Report absorbed dose in mGy for experimental systems and
      effective dose in mSv for human cohorts rather than treating the units as
      interchangeable. No participant should receive radiation solely for this study.
      Follow-up must be long enough to detect delayed neurocognitive or genotoxic
      harm rather than stopping at an acute adaptive response.
    experiment_type:
      preferred_term: harmonized cross-model dose-response study
    perturbations:
    - name: Low-dose ionizing-radiation dose and regimen matrix
      target: pathophysiology#Amyloid Plaque Formation
      exposure_term:
        preferred_term: exposure to ionizing radiation
        term:
          id: ECTO:7000047
          label: exposure to ionizing radiation
      description: >-
        Compare sham exposure with prespecified single and fractionated doses
        spanning an absorbed-dose range chosen to correspond to the review's
        below-100-mSv human exposure definition while holding radiation quality
        and dosimetry constant; include a higher-dose positive-control arm only in
        non-participant experimental systems, never in observational cohorts.
    readouts:
    - name: Amyloid-beta deposition response
      target: pathophysiology#Amyloid Plaque Formation
      description: >-
        Quantify soluble and deposited amyloid-beta across dose, dose-rate, and
        fractionation conditions and in longitudinal human biomarkers.
      assays:
      - preferred_term: amyloid-beta quantification
      direction: NEGATIVE
      interpretation: >-
        A replicated decrease confined to a defined low-dose window would support
        the proposed protective amyloid arm; a monotonic increase or failure to
        replicate would not.
    - name: Microglial inflammatory-state response
      target: pathophysiology#Neuroinflammation
      description: >-
        Measure microglial inflammatory-state markers, cytokines, and phagocytic
        function across the harmonized exposure matrix.
      assays:
      - preferred_term: single-cell transcriptomic profiling
      - preferred_term: cytokine profiling
      direction: NEGATIVE
      interpretation: >-
        Reduced inflammatory signaling together with preserved phagocytic
        function would support an anti-inflammatory arm; a transient marker shift
        without functional benefit would not.
    - name: Synaptic and cognitive outcome
      target: pathophysiology#Synaptic Dysfunction
      description: >-
        Measure synaptic plasticity and network activity in experimental systems
        and longitudinal cognitive change in the observational human cohorts.
      assays:
      - preferred_term: electrophysiological recording
      - preferred_term: neuropsychological assessment
      direction: NEGATIVE
      interpretation: >-
        Molecular changes count as neuroprotective only if they accompany improved
        or preserved synaptic function and a concordant human cognitive direction.
    - name: Delayed oxidative and genotoxic injury
      target: pathophysiology#Oxidative Stress
      description: >-
        Measure persistent oxidative stress, DNA damage, and delayed neuronal or
        cognitive injury after the acute response window.
      assays:
      - preferred_term: DNA damage assay
      - preferred_term: oxidative stress biomarker profiling
      direction: POSITIVE
      interpretation: >-
        Delayed injury at a nominally protective dose would refute a net-benefit
        interpretation even if acute amyloid or inflammatory markers improve.
    controls:
    - name: Sham-exposed matched controls
      description: >-
        Isogenic cultures and matched animals handled identically without
        ionizing-radiation exposure, plus unexposed comparison participants for
        the observational cohorts.
    - name: Dosimetry and regimen controls
      description: >-
        Match radiation quality, absorbed dose, dose rate, fractionation, age, and
        disease stage as closely as each system permits, and publish the conversion
        assumptions, so a unit or regimen difference is not misread as a species
        effect.
    decision_criterion: >-
      Promote an Alzheimer-specific protective arm only if a preregistered
      non-monotonic dose-response is independently replicated, improves amyloid,
      inflammatory, and synaptic outcomes without delayed harm, and shows a
      directionally concordant human signal. Define separate harmful and
      protective exposure-mechanism links only after their dose and regimen
      boundaries are supported; do not infer a clinical safety threshold from
      model-system data alone.
  posed_date: "2026-08-08T00:00:00Z"
  notes: >-
    Scope resolution for GitHub issue 8135: entry-level structured uncertainty was
    chosen over asserted mechanism nodes or a cross-disease hormesis module.
- discussion_id: mismatch_m1_antagonism_mouse_vs_human_anticholinergic_exposure
  prompt: >-
    Does chronic exposure to the non-selective anticholinergic drugs used by
    older adults reproduce, in the human brain, the acceleration of amyloid and
    tau pathology that M1-selective antagonism and M1 ablation produce in
    transgenic mouse models?
  kind: HUMAN_MODEL_MISMATCH
  status: OPEN
  attaches_to:
  - pathophysiology#Muscarinic M1 Receptor Signaling Loss
  rationale: >-
    The mechanistic case that anticholinergic burden is disease-modifying rather
    than merely symptom-masking rests on mouse work that differs from human
    exposure in three ways. (1) Receptor selectivity: dicyclomine and the
    M1(-/-) allele isolate M1, whereas diphenhydramine, tricyclics and
    oxybutynin block M1-M5 and, in the antihistamine case, H1 as well — so the
    human exposure engages targets whose net effect on APP processing is not
    established, and the mouse result cannot simply be scaled up. (2) Baseline
    pathology: the mouse experiments manipulate M1 against a pre-existing
    transgenic amyloid/tau burden, so they demonstrate acceleration of ongoing
    pathology, not initiation in a healthy brain, which is what the
    population-level risk claim requires. (3) CNS exposure: the human
    epidemiology assigns risk by dispensing records, not by brain receptor
    occupancy, and the one large antihistamine cohort finds a dose-response for
    second-generation agents that do not appreciably cross the blood-brain
    barrier, alongside a steeper one for first-generation agents that do
    (1.13/1.29/1.51 versus 1.11/1.19/1.26 across rising cumulative dose). The
    between-class ordering is what a CNS-penetration-dependent mechanism
    predicts; what it does not predict is that the second-generation arm is
    non-null at all, and distinguishing residual confounding by indication from
    a genuine peripheral or non-muscarinic contribution needs measured central
    occupancy rather than dispensing proxies. Resolving this determines whether
    reducing anticholinergic burden is a prevention strategy or only good
    symptomatic geriatric practice.
  proposed_experiments:
  - experiment_id: exp_anticholinergic_occupancy_biomarker_cohort
    name: Muscarinic occupancy and AD biomarker trajectory under chronic anticholinergic exposure
    description: >-
      Follow cognitively normal older adults initiating chronic CNS-penetrant
      anticholinergics, matched initiators of peripherally restricted or
      second-generation comparators, and non-users, with serial plasma and CSF
      amyloid and phospho-tau measures and amyloid/tau PET. Quantify actual
      central muscarinic receptor occupancy in a subset by PET rather than
      inferring CNS exposure from dispensing data, and test whether biomarker
      trajectory tracks occupancy rather than prescription class.
    experiment_type:
      preferred_term: prospective biomarker cohort study with receptor occupancy imaging
    assays:
    - preferred_term: muscarinic receptor positron emission tomography
    - preferred_term: amyloid positron emission tomography
    - preferred_term: tau positron emission tomography
    - preferred_term: cerebrospinal fluid biomarker profiling
    - preferred_term: plasma biomarker profiling
    readouts:
    - name: Amyloid trajectory versus central muscarinic occupancy
      target: pathophysiology#Muscarinic M1 Receptor Signaling Loss
      description: >-
        Rate of change in amyloid PET signal and plasma amyloid-beta 42/40
        modeled against measured central muscarinic occupancy, with prescription
        class as a competing predictor.
      assays:
      - preferred_term: amyloid positron emission tomography
      - preferred_term: plasma amyloid-beta 42/40 measurement
      direction: POSITIVE
      interpretation: >-
        Amyloid accumulation scaling with measured central occupancy would
        support the muscarinic mechanism in humans; accumulation tracking
        prescription class irrespective of occupancy would instead implicate
        confounding by indication.
- discussion_id: controversy_anticholinergic_dementia_reverse_causation_class_specificity
  prompt: >-
    Is the observed association between cumulative anticholinergic exposure and
    incident dementia causal, or is it produced by reverse causation and
    confounding by indication — and why is the signal concentrated in drug
    classes other than the antihistamines?
  kind: CONTROVERSY
  status: OPEN
  attaches_to:
  - pathophysiology#Muscarinic M1 Receptor Signaling Loss
  rationale: >-
    The evidence is genuinely conflicting rather than merely thin. Supporting a
    causal reading: a cumulative dose-response in a prospective cohort with the
    most recent 12 months of exposure excluded (PMID:25621434); persistence of
    the association in windows 5-20 years before diagnosis (PMID:31233095); a
    dose-response that survives an active comparator, with bladder
    anticholinergics carrying higher dementia risk than mirabegron in patients
    who all have overactive bladder (PMID:41066055); and interaction with APOE4
    and CSF AD pathology rather than an APOE-independent effect (PMID:32878992).
    Against it: prodromal dementia causes insomnia, depression and urinary
    urgency, which are these drugs' indications, and the Richardson analysis
    explicitly names this as an alternative reading of its own result
    (PMID:29695481); the class distribution is wrong for a shared muscarinic
    mechanism, since gastrointestinal anticholinergics were null while
    antidepressant, urological and antiparkinson drugs were not; the largest
    antihistamine-specific cohort finds a non-null dose-response even for
    second-generation agents that lack meaningful CNS penetration
    (PMID:38935035, though its gradient is steeper for the first-generation
    class, 1.13/1.29/1.51 versus 1.11/1.19/1.26, so the between-class ordering
    is concordant with a central mechanism and only the non-null
    second-generation arm is anomalous); a propensity-matched target trial emulation finds no
    antihistamine signal (PMID:42250644); an inhaled peripherally restricted
    antimuscarinic shows an effect of questionable clinical importance and a
    null as-treated estimate (PMID:40388132); and the only randomized test,
    deprescribing, is uninformative (PMID:38063254). A third position is that
    the antihistamine-specific nulls are not informative either way. The CPRD
    full text reports that antihistamines were non-significant on
    any-prescription exposure but showed a tentative effect above 365 defined
    daily doses, that only about 0.3% of the sample reached that exposure on
    prescription, that over-the-counter supply is invisible to the database so
    antihistamine use is underestimated, and that the antihistamine result
    needs independent confirmation. Because diphenhydramine is largely bought
    over the counter, every prescription-database study in this literature is
    weakest exactly where the public question is strongest, and the two studies
    that do capture antihistamine exposure directly disagree with each other
    (PMID:38935035 positive and dose-dependent, PMID:42250644 null). Curation
    consequence: the exposure is recorded as PREDISPOSES, and the entry should
    not be strengthened to a causal claim on observational evidence alone —
    nor weakened to "refuted for antihistamines" on the strength of nulls drawn
    from data that cannot see the exposure.
  proposed_experiments:
  - experiment_id: exp_anticholinergic_deprescribing_dementia_rct
    name: Long-horizon anticholinergic deprescribing trial with dementia endpoints
    description: >-
      Randomize older adults with high anticholinergic burden and no cognitive
      impairment to structured deprescribing with non-anticholinergic
      substitution versus usual care, powered for incident MCI and dementia over
      at least five years rather than the one-to-three-month cognitive test
      outcomes of existing trials. Stratify by APOE4 and baseline amyloid status,
      and include a prespecified negative-control arm substituting a
      peripherally restricted agent, so that a null result can distinguish "no
      causal effect" from "insufficient CNS exposure contrast".
    experiment_type:
      preferred_term: randomized controlled deprescribing trial
    assays:
    - preferred_term: neuropsychological assessment
    - preferred_term: amyloid positron emission tomography
    - preferred_term: plasma biomarker profiling
    readouts:
    - name: Incident dementia after burden reduction
      target: pathophysiology#Muscarinic M1 Receptor Signaling Loss
      description: >-
        Incidence of adjudicated MCI and dementia in the deprescribing arm
        relative to usual care over five or more years.
      assays:
      - preferred_term: neuropsychological assessment
      direction: NEGATIVE
      interpretation: >-
        Reduced incidence after burden reduction would support causation;
        unchanged incidence despite a verified reduction in central
        anticholinergic exposure would favor reverse causation and confounding
        by indication.
- discussion_id: gap_cellular_aging_clock_generalizability_and_cell_assignment
  prompt: >-
    Do plasma proteomic cellular aging clocks retain their Alzheimer disease risk
    stratification in younger and non-European-ancestry populations, and do the
    probabilistic protein-to-cell-type assignments that define them reflect true
    cell of origin?
  kind: KNOWLEDGE_GAP
  status: OPEN
  attaches_to:
  - pathophysiology#Neuroinflammation
  - pathophysiology#Neurofibrillary Tangle Formation
  rationale: >-
    The astrocyte, oligodendrocyte precursor cell and inhibitory neuron age gaps
    curated in this entry rest on two assumptions that the source study itself
    flags as unresolved. First, the derivation cohorts (GNPC, UK Biobank, NSHD)
    are predominantly older and of European ancestry, so the reported hazard
    ratios and the APOE-by-astrocyte-aging interaction may not transfer to
    younger or more diverse populations - a consequential gap for a marker
    proposed as an instrument of personalized risk stratification. Second, the
    protein-to-cell-type mapping is probabilistic: it is bounded by the cell
    types catalogued in the Human Protein Atlas, rests on a twofold
    transcriptomic enrichment threshold, and cannot fully disentangle plasma
    proteins released from more than one source. A "cell-type-specific" clock
    whose inputs are only enriched rather than exclusive is a weaker claim than
    the name implies, and orthogonal validation against neuropathology, CSF glial
    markers or neuroinflammation imaging has not yet been done. Until both are
    addressed, these age gaps should be curated as prognostic associations rather
    than as measurements of the aging state of a particular brain cell.
  evidence:
  - reference: PMID:42297981
    reference_title: "Plasma proteomic signatures of cellular aging predict human disease."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: Due to the fact cohorts were predominantly older and Caucasian, broader validation in younger and more diverse populations is essential to extend the generalizability of results reported in this study.
    explanation: >-
      The authors' own statement of the generalizability limitation underlying
      the first half of this gap.
  - reference: PMID:42297981
    reference_title: "Plasma proteomic signatures of cellular aging predict human disease."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: Analysis was based on and is thus restricted to cell types cataloged in the Human Protein Atlas, leaving certain specialized cell populations underrepresented.
    explanation: >-
      Documents the coverage bound on which cell types can have a clock at all.
  - reference: PMID:42297981
    reference_title: "Plasma proteomic signatures of cellular aging predict human disease."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: Accordingly, plasma proteins arise from multiple cellular processes not explicitly disentangled here
    explanation: >-
      The authors concede that a plasma protein assigned to one cell type may
      have multiple sources, which is the substance of the cell-assignment half
      of this gap.
- discussion_id: gap_ad_herpesvirus_zoster_vaccination_dementia_prevention
  prompt: >-
    Does live-attenuated herpes zoster (VZV) vaccination causally reduce
    dementia incidence at the population level — i.e., is the
    neurotropic-herpesvirus contribution to Alzheimer disease causal and
    preventable by vaccination — given that two regression-discontinuity
    analyses of the same age-based vaccine rollout reach opposite conclusions?
  kind: KNOWLEDGE_GAP
  status: OPEN
  attaches_to:
  - pathophysiology#HSV-1 Reactivation in RORB+ Glutamatergic Neurons
  - pathophysiology#Neuroinflammation
  rationale: >-
    The neurotropic-herpesvirus hypothesis predicts that reducing herpesvirus
    reactivation (e.g., by zoster vaccination) should lower dementia risk.
    Multiple independent quasi-experimental studies now bear on this: three
    report a protective association and one eligibility-based analysis is null.
    On the positive side: the Welsh regression-discontinuity (RD) analysis of
    the date-of-birth eligibility cutoff (PMID:40175543) found that receiving
    the live vaccine reduced new dementia diagnoses by 3.5 percentage points
    over 7 years (~20% relative reduction, stronger in women); an independent
    Canadian analysis of natural experiments in Ontario (PMID:41579903) found
    that birth just before versus just after the vaccine-eligibility cutoff
    decreased new dementia diagnoses by 2.0 percentage points over 5.5 years,
    corroborated by synthetic-control comparisons against the same birth
    cohorts in provinces with no vaccination programme; and a vaccine-type
    natural experiment
    (DOI:10.1038/s41591-024-03201-5) found the recombinant (Shingrix) vaccine
    associated with even lower dementia risk than the live vaccine, again
    stronger in women. Against a detectable effect: an English Hospital Episode
    Statistics RD analysis by an overlapping author group
    (DOI:10.64898/2026.07.20.26358345) reproduced the expected sharp drop in
    hospital-coded shingles but found no detectable intention-to-treat effect
    of vaccine ELIGIBILITY on hospital-coded dementia, robust across
    specifications, placebo cutoffs, negative controls, and a separately held
    HES extract. A methodological reconciliation has been proposed — that the
    null result reflects ITT dilution by imperfect uptake plus insensitive
    hospital-coded ascertainment, given that dementia is largely coded in
    primary care and on death certificates — but it does not currently settle
    the disagreement, for two reasons. First, the null study tested that
    explanation itself: comparator analyses in Welsh data with linked
    primary-care and death data did not suggest its results were driven by
    reliance on hospital data. Second, the receipt-versus-eligibility
    distinction does not separate the positive studies from the null one: the
    Canadian headline estimate (PMID:41579903) is itself an eligibility
    contrast at the birth-date cutoff, like the English one, so ITT dilution
    alone cannot explain why the Canadian RD detects an effect where the
    English HES RD does not. Only the Welsh headline estimate is scaled to
    vaccine receipt. The designs do still differ in dementia ascertainment
    (primary-care records and death certificates versus hospital-coded
    diagnoses) and in country, health system, and follow-up length. Whether the
    disagreement reflects outcome ascertainment, other residual design
    differences, or a genuinely absent or much smaller population-level effect
    is unresolved — which determines how strongly the herpesvirus arm should be
    weighted as a causal, modifiable contributor to Alzheimer disease, and
    motivates a randomized controlled trial of zoster vaccination for dementia
    prevention.
  evidence:
  - reference: PMID:40175543
    reference_title: "A natural experiment on the effect of herpes zoster vaccination on dementia."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "reduced the probability of a new dementia diagnosis"
    explanation: >-
      Welsh regression-discontinuity natural experiment supporting a
      dementia-preventing/-delaying effect of zoster-vaccine receipt (reported
      as a 3.5-percentage-point / ~20% relative reduction over 7 years), the
      positive pole of this open question.
  - reference: PMID:41579903
    reference_title: "Herpes zoster vaccination and incident dementia in Canada: an analysis of natural experiments."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "decreased the probability of receiving a new dementia diagnosis by an absolute difference of 2·0 percentage points (95% CI 0·4-3·5, p=0·012) over a 5·5-year follow-up"
    explanation: >-
      The Canadian study's own primary regression-discontinuity FINDING (Ontario
      Jan 1, 1946 eligibility threshold), reproducing the protective signal in a
      different country and health system. Note this is an eligibility contrast
      rather than a receipt-scaled estimate, which is why it also bears on the
      ITT-dilution explanation for the null English analysis.
  - reference: DOI:10.1038/s41591-024-03201-5
    reference_title: "The recombinant shingles vaccine is associated with lower risk of dementia"
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "the recombinant vaccine is associated with a significantly lower risk of dementia in the 6 years post-vaccination"
    explanation: >-
      Vaccine-type natural experiment (transition from live to recombinant
      zoster vaccine) showing the recombinant (Shingrix) vaccine is associated
      with even lower dementia risk than the live vaccine, adding a
      vaccine-platform dimension the eligibility-based null design cannot
      capture and motivating a confirmatory randomized trial.
  - reference: DOI:10.64898/2026.07.20.26358345
    reference_title: "Eligibility for shingles vaccination and hospital-coded dementia in England and Wales: a regression discontinuity analysis in England"
    supports: REFUTE
    evidence_source: HUMAN_CLINICAL
    snippet: "These findings do not support a detectable intention-to-treat effect of live-attenuated shingles vaccine eligibility on hospital-coded dementia in England."
    explanation: >-
      Independent English HES regression-discontinuity analysis (medRxiv
      preprint, overlapping author group) finding no detectable ITT effect of
      zoster-vaccine eligibility on hospital-coded dementia despite a clear
      effect on shingles — the null pole that puts the preventability corollary
      of the herpesvirus hypothesis in genuine equipoise.
  - reference: DOI:10.64898/2026.07.20.26358345
    reference_title: "Eligibility for shingles vaccination and hospital-coded dementia in England and Wales: a regression discontinuity analysis in England"
    supports: REFUTE
    evidence_source: HUMAN_CLINICAL
    snippet: "Comparator analyses in Welsh data with linked primary care and death data did not suggest these results were driven by our reliance on hospital data."
    explanation: >-
      The null study's own pre-rebuttal of the hospital-coded-ascertainment
      explanation for its null result, curated separately because it is a
      distinct claim from the null estimate itself: it is why the
      insensitive-outcome reconciliation cannot be asserted as the likely
      resolution of this gap, and it keeps the gap genuinely open rather than
      self-resolving in prose.
  posed_date: "2026-07-23T00:00:00Z"
  notes: >-
    Seeded from a medRxiv paper review of DOI:10.64898/2026.07.20.26358345
    (Hamilton, Geldsetzer, Davey Smith et al., 2026), then extended with a
    literature investigation of the wider zoster-vaccine/dementia
    natural-experiment landscape (Canadian Lancet Neurology RD, PMID:41579903;
    recombinant-vs-live vaccine-type natural experiment, Nature Medicine 2024,
    DOI:10.1038/s41591-024-03201-5). All four cited references were fetched into
    references_cache/ and their snippets verified as exact substrings of the
    cached abstracts before curation; the medRxiv item is a non-peer-reviewed
    preprint (is_preprint: true) and is curated as an epidemiological ITT
    counterweight, not as mechanistic proof. The residual tension is left open
    rather than adjudicated here: the ascertainment/dilution reconciliation is
    recorded as a proposed but unsettled explanation, since the null study's own
    Welsh comparator analyses with linked primary-care and death data did not
    support it, and the eligibility-versus-receipt contrast does not cleanly
    separate the positive studies from the null one. Complements the
    HSV-1-specific antiviral-suppression knowledge gap attached to the same
    reactivation node.
- discussion_id: controversy_senescent_cell_identity_glia_versus_neurons
  prompt: >-
    Which cells are senescent in the human Alzheimer brain — astrocytes and
    microglia, plaque-associated oligodendrocyte progenitor cells, or excitatory
    neurons — and does the answer invalidate the therapeutic rationale for
    senolytics in Alzheimer disease?
  kind: CONTROVERSY
  status: OPEN
  attaches_to:
  - pathophysiology#Senescent Cell Accumulation
  rationale: >-
    The evidence conflicts on cell identity, and the conflict is consequential
    rather than academic, because senolytics work by killing the cells they
    target. Note first what the conflict is NOT: it is not a species split. A
    fourth study (PMID:30126037) reports both that microdissected tangle-bearing
    neurons from human post-mortem Alzheimer brain carry a senescence expression
    profile, and that in four transgenic mouse models it is neurofibrillary
    tangles rather than amyloid plaques that show the senescence-like phenotype.
    That is mouse and human evidence agreeing on NEURONS. The disagreement is
    between studies, not between model systems. Four studies give three answers. Genetic clearance of p16INK4A-positive
    cells in a tau mouse model attributes the pathology to senescent astrocytes
    and microglia and shows that removing them prevents tangle deposition and
    neuronal loss (PMID:30232451). A senolytic study in an amyloid model finds the
    senescent cell to be the plaque-associated oligodendrocyte progenitor cell and
    states explicitly that astrocytes, microglia and oligodendrocytes do not show
    the phenotype (PMID:30936558) — a direct contradiction of the first study's
    assignment, not merely a different emphasis. The largest human dataset, a
    senescence-eigengene analysis of roughly 140,000 nuclei from 76 postmortem
    brains, assigns more than 97% of senescent cells to excitatory neurons
    overlapping with neurofibrillary tangles (PMID:35531351).
    If the human result is right, a senolytic given for Alzheimer disease would be
    directed at postmitotic neurons rather than at dispensable glia, which is a
    different risk-benefit calculation from the one the mouse studies motivate.
    Both mouse results are causal and both human-relevant readings cannot be
    correct as stated. The main methodological objection to the human study is
    that it identifies senescence by a derived eigengene rather than a
    gold-standard marker; the main objection to the mouse studies is that the
    markers and models differ between them.
    Two further facts about the human study bear on how much weight it can carry,
    both checkable in its own methods. First, it is not an independent cohort: it
    generated no new sequencing data and re-scored two existing dorsolateral
    prefrontal cortex datasets, stating that "Datasets were accessed through
    Accelerating Medicines Partnership - AD (AMP-AD41) with Synapse IDs
    syn18485175 and syn21126462" — Mathys and Zhou respectively. Its cell-type
    call therefore inherits those studies' cell-type annotations and the
    neuron-versus-glia capture characteristics of their single-nucleus
    preparations. Second, to compare cell types it oversampled to equalize them,
    reporting that "we repeated the data of each astrocyte 104, endothelial cell
    2,919, excitatory neuron 10". Nominal counts are then balanced, but the
    effective number of independent observations behind a rare glial or vascular
    cell type is far smaller than behind an excitatory neuron. That is a reason to
    treat a negative result in the rarer types cautiously, rather than a
    demonstration that the positive result in neurons is an artifact.
    A further consideration now bears on the therapeutic stake: the first human
    senolytic trial in Alzheimer disease has reported (SToMP-AD, NCT04063124;
    PMID:37679434, PMID:40274471). It was open-label and uncontrolled in five
    participants, so it cannot test efficacy, but it establishes that the question
    is no longer purely preclinical and that a trial can be run before the
    cell-identity question is settled. Its biomarker result — stable CSF
    amyloid-beta and tau alongside a rise in CSF interleukin-6 — is not what a
    simple "clear the senescent cells, lower the SASP, reduce pathology" reading
    predicts, and is a further reason not to treat the mouse clearance results as
    directly transferable.
    Resolving this needs marker-agnostic
    single-cell senescence profiling in human tissue with orthogonal confirmation,
    and ideally a cell-type-restricted clearance experiment. The most direct
    available test is to re-score senescence signatures per cell type in a cohort
    with enough nuclei per glial class to give the glial hypotheses a fair chance —
    synapse:syn26223298 (SEA-AD, 84 donors, 139 molecular cell types, openly
    downloadable matrices) — and to check the neuron arm against geo:GSE129308,
    where tangle-bearing and tangle-free somas from the same donors make the
    senescence-overlaps-tangles claim testable against measured rather than
    inferred tangle status. One caveat governs how that second test must be run:
    PMID:30126037 derived its senescence expression profile FROM microdissected
    tangle-bearing neuron transcriptomes, so re-scoring that same signature in a
    design that contrasts tangle-bearing with tangle-free somas is partly
    circular. Making the test informative requires a senescence signature defined
    independently of tangle status. Both are curated in this entry's `datasets:`
    block.
    Until then the
    Senescent Cell Accumulation node deliberately lists all four candidate cell
    types rather than choosing one.
  notes: >-
    Raised during a cell-biology review of this entry. All three references were
    fetched into references_cache/ and their snippets verified as exact
    substrings before curation; the two quoted methods statements were read from
    the cached full text of PMID:35531351. Note the entry separately carries an
    oligodendrocyte precursor cell plasma proteomic age gap biomarker, which is
    consistent with — but does not adjudicate — the OPC arm of this dispute.
- discussion_id: controversy_plaque_origin_intraneuronal_versus_secreted
  prompt: >-
    Does the senile plaque form from amyloid-beta secreted into the extracellular
    space, or is it the residue of a single neuron that accumulated amyloid-beta
    in de-acidified autolysosomes and then ruptured?
  kind: CONTROVERSY
  status: OPEN
  attaches_to:
  - pathophysiology#Neuronal Endosomal-Lysosomal Acidification Failure
  - pathophysiology#Amyloid Plaque Formation
  rationale: >-
    The two accounts make incompatible claims about where amyloid pathology
    begins, and they imply different therapeutic targets. The extracellular
    account, which the Amyloid Plaque Formation node and the amyloid_cascade_model
    encode, treats the plaque as a deposit of secreted peptide produced by
    sequential beta- and gamma-secretase cleavage of amyloid precursor protein;
    it is the premise of amyloid-directed monoclonal antibody therapy, which acts
    in the extracellular compartment. The inside-out account holds that
    autolysosome acidification declines in neurons well before extracellular
    deposition, that amyloid-beta and APP-βCTF build up inside enlarged
    de-acidified autolysosomes, and that individual neurons showing the resulting
    PANTHOS pattern are the principal source of senile plaques (PMID:35654956).
    Its human support is genuine but narrower than its mouse support: enlarged
    Rab5-positive endosomes are the earliest known intraneuronal change in
    sporadic disease and are present at preclinical stages (PMID:10880397), and
    intraneuronal amyloid-beta localizes principally to those endosomes
    (PMID:15465622), but the only verified human claim about PANTHOS itself is
    that the pattern is present in Alzheimer brains — not that it is the source of
    human plaques, a claim its authors scope to amyloid precursor protein
    transgenic models where APP is overexpressed.
    A further complication sits inside the human data: endosomes were of normal
    size in advanced presenilin-mutation familial disease (PMID:10880397), so the
    endosomal route appeared not to generalize across Alzheimer genetic subtypes,
    even though PSEN1 is independently required for lysosomal acidification
    (PMID:20541250). That tension is now partly resolved rather than merely
    recorded: in isogenic human iPSC neurons, PSEN1 mutants do accumulate β-CTFs
    and do enlarge Rab5-positive endosomes (PMID:31416668), so the postmortem
    negative may reflect end-stage tissue or the limits of an endosome-size
    readout rather than absence of the mechanism in that subtype.
    Two things sharpen how this could be settled. First, the discriminating
    question is not transcriptomic: organelle pH, v-ATPase subunit assembly,
    peptide sub-cellular location and plaque-to-neuron correspondence are
    imaging, proteomics and quantitative-neuropathology phenotypes, so none of the
    single-nucleus datasets curated in this entry can adjudicate them. Second, and
    more usefully, the relevant human data type already exists: laser-capture
    proteomics of human amyloid plaques against neighbouring tissue reports that
    endosomal and lysosomal proteins are particularly highly enriched within
    plaques (PMID:35418158). That is qualitatively what an intraneuronal origin
    predicts. It does not exclude glial or dystrophic-neurite lysosomal
    contributions, so it does not settle the question — but it does mean the crux
    is re-analyzable against existing human proteomics rather than blocked on data
    that do not exist. Deciding it still needs quantitative human neuropathology
    establishing what fraction of plaques have a neuronal origin, in brains not
    overexpressing APP.
  notes: >-
    Raised during a cell-biology review of this entry. Curated as a CONTROVERSY
    rather than a KNOWLEDGE_GAP because the two accounts are both actively
    defended and make opposing claims, rather than one of them being simply
    unstudied. The PANTHOS phenotype and the APP-βCTF/v-ATPase mechanism come from
    the same laboratory and are not independent replications of one another.
- discussion_id: mismatch_complement_synapse_pruning_mouse_versus_human_opsonin
  prompt: >-
    Does the C1q/C3/CR3 complement pathway that mediates microglial synapse
    elimination in Alzheimer mouse models operate in the human Alzheimer brain,
    where the only direct evidence of glial synapse ingestion implicates MFG-E8
    and involves astrocytes as well as microglia?
  kind: HUMAN_MODEL_MISMATCH
  status: OPEN
  attaches_to:
  - pathophysiology#Complement-Mediated Microglial Synapse Elimination
  rationale: >-
    The necessity evidence for the complement route is entirely murine, and it is
    strong: blocking C1q, C3, or the microglial receptor CR3 each reduces early
    synapse loss (PMID:27033548), and a C1q-blocking antibody rescues synapse
    density in a tau model (PMID:30392797). The human evidence establishes the
    phenomenon but not the pathway. The study that directly demonstrates glial
    synapse ingestion in human Alzheimer tissue states that direct human evidence
    for glial involvement in synapse removal remained to be established, and finds
    that blocking MFG-E8 — not complement — reverses the elevated engulfment, with
    astrocytes participating alongside microglia (PMID:37652017).
    Human corroboration of complement specifically is better than this discussion
    originally recorded, and the correction is worth stating precisely. Beyond C1q
    in postsynaptic densities, C3 protein is elevated in Alzheimer patient brains
    including at synapses, with cerebrospinal fluid C3 tracking tau
    (PMID:31433986); complement dysregulation with C1q deposited on tau aggregates
    extends to human tauopathies where amyloid is absent (PMID:42271460); and
    common-variant genetics implicates two complement loci, CR1 and CLU
    (PMID:29504051). The genetic association matters most, because unlike every
    postmortem correlation it cannot be a downstream consequence of established
    pathology.
    What none of this establishes is the executing step. No human study shows glia
    phagocytosing structurally intact synapses via complement opsonization. The
    phenomenon is human-confirmed and the pathway is human-implicated; the join
    between them is not.
    This matters because anti-C1q antibody therapy is in clinical development on
    the strength of the mouse pathway. If the dominant human opsonin is MFG-E8, or
    if astrocytes rather than microglia are the principal effector, the mouse
    result could be real and the therapy still miss its target. A second, separate
    caution comes from within the mouse literature itself: C3 deletion protects
    synapses and cognition while increasing plaque burden (PMID:28566429), so
    complement inhibition trades amyloid clearance against synapse preservation
    rather than being unambiguously beneficial.
    Resolving this needs human tissue or human cellular models in which the
    complement and MFG-E8 routes are blocked separately and compared, and ideally
    a synaptic-density readout from the ongoing anti-C1q trials. Note that the
    single-nucleus datasets curated in this entry cannot close this gap: they can
    show where and when complement genes are induced and whether the complement
    and MFG-E8 routes are expressed in the same cells, but engulfment is a
    phagocytic event and opsonin identity is a protein-level question, neither of
    which is visible in transcript abundance.
  notes: >-
    Raised during a cell-biology review of this entry. All four references were
    fetched into references_cache/ and their snippets verified as exact substrings
    before curation. The MFGE8 gene is listed on the pathophysiology node
    alongside the complement genes to keep this alternative visible at the point
    of use.
- discussion_id: gap_ad_tau_toxic_species_and_glial_cofactor
  prompt: >-
    Which tau species is the proximate neurotoxic entity in Alzheimer disease —
    soluble pretangle assemblies or mature neurofibrillary tangles — and is the
    glial state that separates demented from cognitively resilient people at
    matched tangle burden a required intermediary or a parallel process?
  kind: KNOWLEDGE_GAP
  status: OPEN
  attaches_to:
  - mechanistic_hypotheses#tau_neurodegeneration_model
  - pathophysiology#Neurofibrillary Tangle Formation
  rationale: >-
    Two claims that would materially change how this entry reads are currently
    unsupported by anything citable, and both came out of the OpenScientist
    tau_neurodegeneration_model report with citations that do not carry them.
    The first is that soluble pretangle tau rather than the mature tangle is the
    primary toxic species, which would relocate the toxicity of this hypothesis
    upstream of the lesion it is named for. The sentence the report quotes for it
    is background framing in PMID:41695614, not that paper's result; what the
    paper does establish is that pretangle moieties marked by pS422, TOC1 and TNT2
    are present as neuropil threads and neuronal inclusions in default-mode-network
    hubs from Braak stage III, with stepwise increases through Braak IV to V.
    The mirror claim, that tangle formation is a protective sequestration step, is
    worse served: PMID:31456657 is an in vitro study of how EFhd2 alters tau liquid
    liquid phase separation and measures no toxicity, no neurons and no protection.
    The oligomer-versus-tangle question is a real and live controversy; neither
    citation can settle it, and it should not be curated as evidence.
    The second claim is that glial activation is a required intermediary
    converting tau pathology into clinical disease. The observation behind it is
    solid and is curated on the hypothesis block: at matched Braak III-IV burden,
    demented and resilient brains differ in glial phenotype, including in visual
    cortex carrying no overt tau (PMID:35534858, 55 brains). But the design is
    cross-sectional post-mortem comparison. Nothing cited perturbs glia and
    measures the degenerative or clinical consequence at a fixed tau burden, which
    is what necessity would require. Resolving both needs the same kind of study:
    a perturbation with a tau-species-resolved and a glia-resolved readout in the
    same system, ideally human tissue or human cellular models, with tangle burden
    held constant.
  notes: >-
    Raised from the OpenScientist tau_neurodegeneration_model assessment. Recorded
    as a gap rather than as evidence precisely because the report's two supporting
    citations are framing sentences rather than results; see the assessment sidecar
    for the per-claim reasoning.
- discussion_id: gap_ad_synaptic_convergence_relative_stressor_weights
  prompt: >-
    What is the relative contribution of each upstream stressor — amyloid, tau,
    complement-mediated pruning, vascular and glymphatic failure, oxidative
    injury, gut-derived inflammation, infection — to synaptic failure in Alzheimer
    disease, and what observation would falsify the convergence model as stated?
  kind: KNOWLEDGE_GAP
  status: OPEN
  attaches_to:
  - mechanistic_hypotheses#synaptic_failure_convergence_model
  - pathophysiology#Synaptic Dysfunction
  rationale: >-
    The convergence model's descriptive breadth is also its methodological
    weakness, and the OpenScientist synaptic_failure_convergence_model report says
    so about its own hypothesis: by accommodating nearly every known Alzheimer
    mechanism as an upstream stressor, the model risks becoming unfalsifiable. The
    concrete form of that critique is that no study quantifies how much of the
    synaptic loss each stressor accounts for, so no arm of the model can be
    weakened by evidence without weakening all of them equally.
    The arm-by-arm support is also markedly uneven, which the summary verdict hides.
    The report's own link-strength table scores five of eleven causal links as
    established, five as emerging, and the infectious arm as speculative against a
    negative randomised trial of valacyclovir. The amyloid, tau and complement arms
    rest on multiple independent lines; the gut-vesicle route into complement
    pruning (PMID:40731189) and the glymphatic-to-synapse association
    (PMID:41454064) each rest on one study; and the infectious arm belongs in the
    separate hsv1_reactivation_model block rather than here.
    What would move this is a formalisation rather than more association studies:
    a weighted causal graph fitted to multimodal human data in which stressor
    contributions are estimated jointly and can be driven to zero. Note also that
    synaptic loss appears to sit at the pathology-to-symptom step rather than at
    pathology onset — hippocampal synaptic density is similar in controls and in
    cognitively unimpaired people with biological Alzheimer disease, and falls once
    impairment appears (PMID:40465636) — so any such model has to explain the
    threshold, not only the slope.
  notes: >-
    Raised from the OpenScientist synaptic_failure_convergence_model assessment.
    The report recommended no status change and the CANONICAL status is unchanged;
    this records the self-critique, which was the most useful thing in the report
    and would otherwise have been lost. The report's ontology lead binding
    complement-mediated synapse elimination to GO:0150076 was rejected in
    assessment and not curated: that term is the neuroinflammatory response.
- discussion_id: gap_ad_vascular_to_tau_causal_link
  prompt: >-
    Does neurovascular and blood-brain barrier dysfunction cause or accelerate tau
    pathology in Alzheimer disease, or does its explanatory reach stop at amyloid
    accumulation while tau follows its own propagation logic?
  kind: KNOWLEDGE_GAP
  status: OPEN
  attaches_to:
  - mechanistic_hypotheses#vascular_bbb_clearance_model
  - pathophysiology#Vascular Dysfunction
  rationale: >-
    This is the boundary that keeps the vascular model at ALTERNATIVE, and the
    entry did not previously record it. The evidence that vascular injury acts on
    amyloid is now substantial and partly interventional: hypoperfusion raises
    soluble and then insoluble amyloid in a transgenic model within one to three
    months (PMID:28667120), APOE4-graded pericyte degeneration tracks barrier
    breakdown in human post-mortem tissue (PMID:25757756), and a common risk locus
    routes into barrier efflux through PICALM (PMID:30041013). Nothing comparable
    connects the vascular arm to tau, which propagates by templated, prion-like
    spread along connected circuits. Because tau burden rather than amyloid is the
    stronger correlate of cognitive decline, a model that explains amyloid
    accumulation but not tau has bounded reach even where it is correct.
    Two further observations sharpen the gap rather than closing it. Cerebral blood
    flow associations with tau invert by disease stage — negative in mild cognitive
    impairment, positive in established disease (PMID:40884970) — so the vascular
    arm does not even relate to tau monotonically. And no human interventional
    trial has shown that correcting vascular dysfunction prevents or slows amyloid
    accumulation. The rrAD trial (NCT02913664) randomises 640 cognitively normal
    hypertensive adults to intensive vascular risk reduction, exercise, or both,
    but its primary outcomes are neurocognitive with MRI at baseline and 24 months,
    so it will not deliver the serial amyloid PET endpoint this question needs.
  notes: >-
    Raised from the OpenScientist vascular_bbb_clearance_model assessment, which
    identified this as the hypothesis's most important limitation. The trial
    identifier was checked against the cached rationale-and-methods publication
    (PMID:30826452), which carries NCT02913664.
- discussion_id: gap_ad_arc_ev_tau_replication_and_genetic_support
  prompt: >-
    Does the Arc-dependent extracellular-vesicle route for tau propagation
    replicate outside the founding laboratory, and does the complete absence of
    human genetic association for ARC argue against it carrying a material share
    of tau spread in Alzheimer disease?
  kind: KNOWLEDGE_GAP
  status: OPEN
  attaches_to:
  - mechanistic_hypotheses#ev_mediated_tau_propagation_model
  - pathophysiology#Intercellular Tau Transmission via Extracellular Vesicles
  rationale: >-
    Two independent reasons to hold this mechanism at EMERGING, neither of which is
    a criticism of the founding result.
    The first is single-source dependency with a disclosed commercial interest. Arc
    binding tau, and the near-abolition of intercellular tau transmission in
    Arc-knockout animals, both come from PMID:42372723 and from nowhere else; the
    paper appeared online only shortly before the provider search, so the absence
    of replication is expected rather than suspicious. What makes it worth
    recording is that the senior author discloses co-founding VNV, LLC and
    consulting for Aera Therapeutics, which licenses patents covering Arc capsids.
    Single-source results with commercial interest in the mechanism are exactly
    where independent replication carries the most weight. A replication should
    also use authentic Alzheimer-type 3R/4R paired helical filaments rather than
    the P301L 4R tau of the original model.
    The second is the absence of any human genetic footprint. A provider survey of
    UniProt, Open Targets, the GWAS Catalog and GenCC found no genetic, GWAS,
    somatic-mutation or drug-target association between ARC and Alzheimer disease.
    That absence is a legitimate structural argument, and it contrasts with the
    competing tunnelling-nanotube route, which does connect to Alzheimer risk genes.
    It is recorded here as a gap rather than as evidence because none of the four
    surveys was committed with a query, release version, timestamp or saved
    response, and an Open Targets score quoted to four decimal places is not
    reproducible without a release version. Re-derive before relying on it.
    A third, separate question the same report raised should not be mistaken for a
    finding: whether hyperexcitability induces Arc, which raises vesicle tau
    release, which raises excitability in recipient neurons, closing a
    self-amplifying loop. Each link has separate support but no experiment closes
    the circuit, and the amyloid-Arc suppression already curated on this hypothesis
    cuts against the loop operating where Alzheimer pathology is heaviest. It
    belongs here as a question, never as pathophysiology edges.
  notes: >-
    Raised from the OpenScientist ev_mediated_tau_propagation_model assessment. The
    report's proposed tau-strain snippet for PMID:33693528 was rejected as a
    lightly reworded paper title and not curated; its 8-month pathology-convergence
    claim is not in the cached abstract and no early-stage restriction of the model
    was made on the strength of it.
- discussion_id: gap_glucosamine_in_established_ad
  prompt: >-
    Does glucosamine supplementation worsen cognitive decline or survival in
    people who already have Alzheimer disease, and is any harm confined to
    brains that are already hyperglycosylated?
  kind: KNOWLEDGE_GAP
  status: OPEN
  attaches_to:
  - environmental#Oral glucosamine supplementation
  - pathophysiology#Brain N-Glycan Hyperglycosylation
  - mechanistic_hypotheses#n_glycan_hyperglycosylation_model
  rationale: >-
    Mouse data show glucosamine raising brain N-glycans and worsening memory in
    5xFAD mice but not in wild-type mice, and one health-system EHR analysis
    associates documented use with worse survival in dementia. Large
    population cohorts associate the same supplement with lower or unchanged
    dementia incidence. The two bodies of evidence ask different questions
    (progression after diagnosis versus incidence), and all of the human data
    are observational, so whether a widely used over-the-counter supplement is
    harmful in established disease is unresolved.
  proposed_experiments:
  - experiment_id: exp_glucosamine_rct_established_ad
    name: Randomized placebo-controlled glucosamine trial in biomarker-confirmed Alzheimer disease
    description: >-
      Randomize people with biomarker-confirmed mild cognitive impairment or
      mild dementia due to Alzheimer disease who already take glucosamine to
      continuation or blinded placebo substitution, with cognitive, functional
      and survival endpoints and, where feasible, CSF or imaging readouts of
      glycosylation. Stratify by osteoarthritis severity to separate the
      supplement from its indication.
    experiment_type:
      preferred_term: randomized double-blind placebo-controlled trial
    assays:
    - preferred_term: neuropsychological assessment
    readouts:
    - name: Rate of cognitive decline
      target: phenotypes#Memory Loss
      direction: INCREASED
      interpretation: >-
        Faster decline on continued glucosamine than on placebo would support
        the exacerbation claim; no difference would argue that the EHR signal
        reflected confounding.
    decision_criterion: >-
      The environmental entry should move from EXACERBATES to a documented null
      if a trial powered for cognitive decline finds no difference, and the
      hypothesis group can advance past EMERGING only if the trial shows harm
      together with evidence that glucosamine alters brain glycosylation in
      humans.
    would_support:
    - mechanistic_hypotheses#n_glycan_hyperglycosylation_model
    would_refute:
    - mechanistic_hypotheses#n_glycan_hyperglycosylation_model
    evidence:
    - reference: PMID:42265388
      reference_title: "Hyperglycosylation is a metabolic driver of Alzheimer's disease."
      supports: SUPPORT
      evidence_source: HUMAN_CLINICAL
      snippet: "These findings emphasize the urgent need for a well-designed, double-blind clinical trial to systematically evaluate the impact of glucosamine on cognitive decline in patients with dementia."
      explanation: >-
        The originating authors identify a randomized trial as the missing
        test of their EHR association.
  posed_date: "2026-09-27T00:00:00Z"
- discussion_id: gap_n_glycan_direction_in_ad_brain
  prompt: >-
    Is protein N-glycosylation increased or decreased in the Alzheimer brain,
    and can a net rise in released N-glycans coexist with site-specific loss of
    glycosylation on synaptic proteins?
  kind: KNOWLEDGE_GAP
  status: OPEN
  attaches_to:
  - pathophysiology#Brain N-Glycan Hyperglycosylation
  - pathophysiology#Synaptic Dysfunction
  rationale: >-
    MALDI imaging of released N-glycans finds a net increase in human and mouse
    AD brain, while site-specific glycoproteomics of post-mortem AD brain finds
    reduced glycosylation at many sites, enriched on synaptic channels and
    receptors, and proposes that loss of glycosylation drives synaptic
    dysfunction. A third glycoproteomic study finds stage-associated changes in
    glycan structure at individual sites without a single direction. The
    methods measure different quantities, so the findings need not conflict,
    but they point therapy in opposite directions: reducing N-glycan
    biosynthesis is protective under one reading and harmful under the other.
  proposed_experiments:
  - experiment_id: exp_paired_glycan_imaging_and_site_glycoproteomics
    name: Paired released-glycan imaging and site-specific glycoproteomics on the same AD tissue
    description: >-
      On matched sections from the same post-mortem AD and control brains,
      quantify released N-glycans by MALDI imaging and site occupancy by
      site-specific glycoproteomics, then test whether net glycan abundance and
      synaptic site occupancy move in opposite directions within the same
      samples and brain regions.
    experiment_type:
      preferred_term: cross-sectional post-mortem tissue study
    assays:
    - preferred_term: MALDI mass spectrometry imaging of released N-glycans
    - preferred_term: site-specific N-glycoproteomics
    readouts:
    - name: Net released N-glycan abundance
      target: pathophysiology#Brain N-Glycan Hyperglycosylation
      biological_processes:
      - preferred_term: Protein N-linked glycosylation
        term:
          id: GO:0006487
          label: protein N-linked glycosylation
      direction: INCREASED
      interpretation: >-
        An increase alongside reduced synaptic site occupancy in the same
        samples would reconcile the two literatures; a decrease would refute
        the hyperglycosylation node in that tissue.
    decision_criterion: >-
      If released-glycan abundance is not increased when measured side by side
      with site occupancy, the hyperglycosylation node should be downgraded to
      HYPOTHETICAL.
    would_support:
    - pathophysiology#Brain N-Glycan Hyperglycosylation
    would_refute:
    - pathophysiology#Brain N-Glycan Hyperglycosylation
  evidence:
  - reference: PMID:42265388
    reference_title: "Hyperglycosylation is a metabolic driver of Alzheimer's disease."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "Our work complements these findings by measuring the net abundance of released N-glycans across brain regions rather than peptide-level occupancy."
    explanation: >-
      The originating authors' own statement that they and the site-specific
      studies measure different quantities.
  - reference: PMID:41042700
    reference_title: "Comprehensive and Site-Specific Characterization of Protein N-Glycosylation in AD Samples Reveals Its Potential Roles in Protein Aggregation and Synaptic Dysfunction."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "Our findings suggest that the loss of N-glycosylation may contribute to the pathogenesis of AD through impairing synaptic transmission and promoting protein aggregation."
    explanation: >-
      The competing directional claim this discussion asks to be resolved.
  - reference: PMID:36309312
    reference_title: "Glycoproteomics Landscape of Asymptomatic and Symptomatic Human Alzheimer's Disease Brain."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "We observed AD-associated differences in the number of antennae, frequency of fucosylation, bisection, and other monosaccharides at individual glycosylation sites among samples from our three groups."
    explanation: >-
      A third, independent dataset showing that N-glycan structure changes
      across asymptomatic and symptomatic AD at individual sites, without
      reducing to a single direction.
  posed_date: "2026-09-27T00:00:00Z"
- discussion_id: mismatch_micro_mkks63_hmc3_versus_human_microglia
  prompt: >-
    Does the respiratory deficit produced by micro-MKKS63 knockout in an
    immortalized human microglial line occur in microglia in the Alzheimer brain,
    where the microprotein is reduced rather than absent and where the depletion
    was measured in bulk tissue that cannot say which cell type lost it?
  kind: HUMAN_MODEL_MISMATCH
  status: OPEN
  attaches_to:
  - pathophysiology#Micro-MKKS63 Microprotein Depletion
  - pathophysiology#Microglial Oxidative Phosphorylation Failure
  - experimental_models#HMC3 human microglial micro-MKKS63 CRISPR-Cas9 knockout lines
  rationale: >-
    Two claims are being joined here, and each is sound on its own. The human
    claim is that micro-MKKS63 falls in symptomatic Alzheimer cortex; it comes
    from tandem-mass-tag quantitation across hundreds of ROSMAP donors. The
    cellular claim is that losing the microprotein impairs oxidative
    phosphorylation; it comes from two independent CRISPR knockout lines in HMC3
    cells. The join between them is the weak point, and it fails in two distinct
    ways rather than one.
    First, cell type. The depletion is a bulk-tissue measurement. Microglia are a
    small minority of cortical cells, so a bulk decrease is more likely to reflect
    the majority populations than the microglial one, and cell-type attribution
    cannot be recovered from the tandem-mass-tag data at all. It is entirely
    possible that micro-MKKS63 falls in neurons while microglial levels are
    unchanged, in which case the human observation and the cellular experiment
    would be about different cells. The microprotein's coexpression network is
    enriched for mitochondrial biology, which is consistent with a role in any
    high-demand cell type and does not adjudicate this.
    Second, dose and context. The knockout abolishes the microprotein; the disease
    shows a partial decrease. For micro-MKKS63 specifically the log2 fold change is
    -0.063 - about a 4% reduction - which is typical rather than unusually small:
    the median absolute log2 fold change across all significant microproteins in
    the stringent model is 0.0969, which the authors attribute to cellular
    heterogeneity in bulk postmortem tissue. Whether a reduction of that size
    produces any respiratory change is untested, and there is no dose-response
    series. HMC3 is additionally an SV40-immortalized fetal-derived line grown in
    standard culture, a system whose metabolic setpoint and transcriptional state
    are known to diverge from microglia in aged tissue.
    Resolving this needs the cell-type question answered first, because the dose
    question is only worth asking in whichever cell type actually loses the
    microprotein. Nothing currently curated in this entry can answer it:
    single-nucleus datasets report transcripts, and a smORF's transcript is
    usually shared with the host locus, so transcript-level data cannot separate
    micro-MKKS63 from canonical MKKS - which is precisely the decoupling the
    finding rests on.
  evidence:
  - reference: PMID:42736449
    reference_title: A microprotein atlas of the human frontal cortex in Alzheimer's disease.
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    directness: DIRECT
    quote_role: PRIMARY_RESULT
    snippet: "Effect sizes reflect directional associations rather than large shifts, as median absolute log2(fold change) among significant stringent-model MPs was 0.0969, consistent with cellular heterogeneity in bulk postmortem tissue"
    explanation: >-
      Supports the dose half of the mismatch, and does so in the authors' own
      words: the disease-associated microprotein changes are small directional
      shifts in bulk tissue, not the complete loss the knockout imposes. The
      cellular-heterogeneity attribution is also the cell-type half of the
      mismatch stated from the other direction.
  notes: >-
    Raised while curating the microprotein atlas into this entry. Recorded as
    HUMAN_MODEL_MISMATCH rather than KNOWLEDGE_GAP because the evidence is not
    absent: it exists in a model system, and its translational validity is the
    open question. The companion discussion
    gap_micro_mkks63_causal_role_in_alzheimer_disease covers the separate
    question of causal direction.
- discussion_id: gap_micro_mkks63_causal_role_in_alzheimer_disease
  prompt: >-
    Is micro-MKKS63 depletion a contributor to Alzheimer disease, a consequence
    of it, or an incidental correlate - and if it contributes, through what
    downstream step?
  kind: KNOWLEDGE_GAP
  status: OPEN
  attaches_to:
  - pathophysiology#Micro-MKKS63 Microprotein Depletion
  - pathophysiology#Microglial Oxidative Phosphorylation Failure
  - mechanistic_hypotheses#smorf_microprotein_model
  rationale: >-
    The Microglial Oxidative Phosphorylation Failure node is deliberately
    terminal, and this discussion records why. Microglial bioenergetic failure has
    obvious candidate consequences in this entry - impaired phagocytic clearance
    of amyloid, a shift toward a proinflammatory state feeding Neuroinflammation,
    convergence on the mitochondrial nodes already curated - and not one of them
    was tested. Drawing any of those edges would convert a plausible narrative
    into a curated causal claim, which is the failure mode this entry's other
    discussions were written to avoid.
    Direction is equally unsettled. Postmortem tissue in advanced disease cannot
    distinguish a change that drove pathology from one produced by it, and
    Alzheimer cortex is a tissue in which mitochondrial function, cell
    composition, and translation are all already disturbed - so a microprotein
    that is mitochondria-associated has several routes to falling as a bystander.
    There is no genetic evidence to break the tie: the MKKS locus carries no known
    Alzheimer association, and the canonical protein's disease association is with
    McKusick-Kaufman and Bardet-Biedl syndromes, which the microprotein result
    does not touch and which are not evidence about this ORF.
    What would move it: a cell-type-resolved measurement of the microprotein in
    Alzheimer and control cortex, to establish which cell loses it; a partial
    knockdown series in human microglia matched to the disease-scale reduction,
    with phagocytosis and cytokine readouts rather than respirometry alone, to
    establish whether the deficit propagates; and a variant-level lookup at the
    smORF in Alzheimer genome-wide association data, which is the only design
    among these that could establish direction rather than association.
  proposed_experiments:
  - experiment_id: exp_micro_mkks63_celltype_resolved_depletion
    name: Cell-type-resolved micro-MKKS63 quantitation in Alzheimer and control cortex
    description: >-
      Quantify micro-MKKS63 in microglia, neurons, astrocytes and oligodendrocytes
      separately from Alzheimer and non-Alzheimer dorsolateral prefrontal cortex,
      using cell-type-sorted or laser-capture material with a microprotein-aware
      search database and a targeted assay for the smORF's unique tryptic
      peptides. The experiment exists to answer which population loses the
      microprotein, which is a precondition for interpreting the knockout result
      at all.
    experiment_type:
      preferred_term: cell-type-resolved quantitative proteomics of postmortem tissue
    assays:
    - preferred_term: fluorescence-activated nuclei or cell sorting
    - preferred_term: targeted parallel reaction monitoring mass spectrometry
    perturbations:
    - name: Alzheimer disease status
      description: >-
        Not an applied intervention but the contrast under study: symptomatic
        Alzheimer donors against non-Alzheimer donors, matched for age, sex and
        postmortem interval.
      target: pathophysiology#Micro-MKKS63 Microprotein Depletion
    readouts:
    - name: Micro-MKKS63 abundance per cell type
      target: pathophysiology#Micro-MKKS63 Microprotein Depletion
      direction: DECREASED
      description: >-
        Abundance of micro-MKKS63 unique peptides in each sorted population,
        expressed against the same population in control donors, with canonical
        MKKS peptides quantified alongside as the decoupling control.
      interpretation: >-
        A microglia-selective decrease would join the human observation to the
        cellular experiment. A decrease confined to neurons would mean the two
        are about different cells and that the knockout result, while real,
        models something the disease does not do.
    would_support:
    - pathophysiology#Microglial Oxidative Phosphorylation Failure
    supporting_outcome:
    - >-
      Micro-MKKS63 is reduced in Alzheimer microglia specifically, while canonical
      MKKS is unchanged in the same cells.
    refuting_outcome:
    - >-
      Micro-MKKS63 is unchanged in Alzheimer microglia, with the bulk-tissue
      decrease accounted for by neurons or another non-microglial population.
    notes: >-
      Listed first because it gates the others: a dose-response knockdown in
      microglia is not worth running if microglia are not the cells that lose the
      microprotein.
  - experiment_id: exp_micro_mkks63_graded_knockdown_microglial_function
    name: Disease-scale graded knockdown of micro-MKKS63 in human microglia
    description: >-
      Titrate micro-MKKS63 loss in human induced-pluripotent-stem-cell-derived
      microglia across a range spanning the reduction seen in Alzheimer cortex up
      to complete loss, leaving the canonical MKKS ORF intact, and measure
      respiration alongside the microglial functions that would connect a
      bioenergetic deficit to this entry's existing pathograph.
    experiment_type:
      preferred_term: graded loss-of-function series in human iPSC-derived microglia
    assays:
    - preferred_term: extracellular flux respirometry
    - preferred_term: amyloid-beta phagocytosis assay
    - preferred_term: multiplexed cytokine profiling
    perturbations:
    - name: Graded micro-MKKS63 depletion
      description: >-
        An allelic or knockdown series producing partial reductions matched to the
        disease-scale effect size as well as complete loss, with the main MKKS ORF
        preserved throughout.
      target: pathophysiology#Micro-MKKS63 Microprotein Depletion
      genes:
      - preferred_term: MKKS locus (micro-MKKS63 smORF)
        description: >-
          The series targets the smORF and leaves the main ORF intact, so the
          binding names where the smORF sits, not the annotated gene product.
        term:
          id: hgnc:7108
          label: MKKS
    readouts:
    - name: Respiratory capacity across the depletion series
      target: pathophysiology#Microglial Oxidative Phosphorylation Failure
      direction: DECREASED
      description: >-
        Basal, ATP-linked and maximal oxygen consumption at each depletion level,
        in a human microglial system rather than an immortalized line.
      interpretation: >-
        A deficit appearing only at complete loss would mean the published
        knockout overshoots the disease; a graded deficit reaching down to the
        disease-scale reduction would make the bioenergetic claim relevant to
        Alzheimer cortex.
    - name: Amyloid-beta phagocytic capacity
      target: pathophysiology#Microglial Oxidative Phosphorylation Failure
      direction: DECREASED
      description: >-
        Uptake and degradation of fibrillar amyloid-beta by microglia at each
        depletion level.
      interpretation: >-
        This is the readout that would justify an edge from the bioenergetic node
        into the amyloid arm of this entry. Without it the node stays terminal,
        which is its current state.
    would_support:
    - pathophysiology#Microglial Oxidative Phosphorylation Failure
    supporting_outcome:
    - >-
      Disease-scale depletion reproduces the respiratory deficit and impairs
      amyloid phagocytosis or shifts cytokine output.
    refuting_outcome:
    - >-
      Only complete loss affects respiration, and microglial function is
      unchanged at disease-scale depletion.
  notes: >-
    Raised while curating the microprotein atlas into this entry, alongside
    mismatch_micro_mkks63_hmc3_versus_human_microglia, which covers the separate
    translational-validity question. The proposed experiments are ordered: the
    cell-type-resolved measurement gates the knockdown series.
tracked_issues:
- url: https://github.com/monarch-initiative/dismech/issues/2617
  title: "Add explicit representation of knowledge gaps"
  tracked_issue_role: schema_followup
  notes: >-
    Meta-issue tracking the design of a `knowledge_gaps:` (structural) layer
    that would complement the `discussions:` (discourse) layer demonstrated
    on this entry by the HSV-1 reactivation discussion.
references:
- reference: DOI:10.3390/cimb47080580
  title: 'From Amyloid to Synaptic Dysfunction: Biomarker-Driven Insights into Alzheimer’s Disease'
  findings: []
- reference: DOI:10.3390/ijms25136901
  title: 'Comprehensive Overview of Alzheimer’s Disease: Etiological Insights and Degradation Strategies'
  findings: []
- reference: DOI:10.3390/ijms252212311
  title: 'From Fundamentals to Innovation in Alzheimer’s Disease: Molecular Findings and Revolutionary Therapies'
  findings: []
- reference: DOI:10.1002/alz.13859
  title: "Revised criteria for diagnosis and staging of Alzheimer's disease: Alzheimer's Association Workgroup"
  found_in:
  - Alzheimer_Disease-deep-research-falcon.md
  findings:
  - statement: "Revised criteria for diagnosis and staging of Alzheimer's disease: Alzheimer's Association Workgroup"
    supporting_text: The National Institute on Aging and the Alzheimer's Association convened three separate work groups in 2011 and single work groups in 2012 and 2018 to create recommendations for the diagnosis and characterization of Alzheimer's disease (AD).
    evidence:
    - reference: DOI:10.1002/alz.13859
      reference_title: "Revised criteria for diagnosis and staging of Alzheimer's disease: Alzheimer's Association Workgroup"
      supports: SUPPORT
      evidence_source: HUMAN_CLINICAL
      snippet: The National Institute on Aging and the Alzheimer's Association convened three separate work groups in 2011 and single work groups in 2012 and 2018 to create recommendations for the diagnosis and characterization of Alzheimer's disease (AD).
      explanation: Deep research cited this publication as relevant literature for Alzheimer Disease.
- reference: DOI:10.1038/s41467-023-37437-5
  title: Single-nucleus RNA-sequencing of autosomal dominant Alzheimer disease and risk variant carriers
  found_in:
  - Alzheimer_Disease-deep-research-falcon.md
  findings:
  - statement: Genetic studies of Alzheimer disease (AD) have prioritized variants in genes related to the amyloid cascade, lipid metabolism, and neuroimmune modulation.
    supporting_text: Genetic studies of Alzheimer disease (AD) have prioritized variants in genes related to the amyloid cascade, lipid metabolism, and neuroimmune modulation.
    evidence:
    - reference: DOI:10.1038/s41467-023-37437-5
      reference_title: Single-nucleus RNA-sequencing of autosomal dominant Alzheimer disease and risk variant carriers
      supports: SUPPORT
      evidence_source: HUMAN_CLINICAL
      snippet: Genetic studies of Alzheimer disease (AD) have prioritized variants in genes related to the amyloid cascade, lipid metabolism, and neuroimmune modulation.
      explanation: Deep research cited this publication as relevant literature for Alzheimer Disease.
- reference: DOI:10.1186/s13195-024-01469-w
  title: Combining plasma Aβ and p-tau217 improves detection of brain amyloid in non-demented elderly
  found_in:
  - Alzheimer_Disease-deep-research-falcon.md
  findings:
  - statement: Combining plasma Aβ and p-tau217 improves detection of brain amyloid in non-demented elderly
    supporting_text: Maximizing the efficiency to screen amyloid-positive individuals in asymptomatic and non-demented aged population using blood-based biomarkers is essential for future success of clinical trials in the early stage of Alzheimer’s disease (AD).
    evidence:
    - reference: DOI:10.1186/s13195-024-01469-w
      reference_title: Combining plasma Aβ and p-tau217 improves detection of brain amyloid in non-demented elderly
      supports: SUPPORT
      evidence_source: HUMAN_CLINICAL
      snippet: Maximizing the efficiency to screen amyloid-positive individuals in asymptomatic and non-demented aged population using blood-based biomarkers is essential for future success of clinical trials in the early stage of Alzheimer’s disease (AD).
      explanation: Deep research cited this publication as relevant literature for Alzheimer Disease.

📚

References & Deep Research

References

DOI:10.3390/cimb47080580

From Amyloid to Synaptic Dysfunction: Biomarker-Driven Insights into Alzheimer’s Disease

No top-level findings curated for this source.

DOI:10.3390/ijms25136901

Comprehensive Overview of Alzheimer’s Disease: Etiological Insights and Degradation Strategies

No top-level findings curated for this source.

DOI:10.3390/ijms252212311

From Fundamentals to Innovation in Alzheimer’s Disease: Molecular Findings and Revolutionary Therapies

No top-level findings curated for this source.

DOI:10.1002/alz.13859

Revised criteria for diagnosis and staging of Alzheimer's disease: Alzheimer's Association Workgroup

1 finding

Revised criteria for diagnosis and staging of Alzheimer's disease: Alzheimer's Association Workgroup

"The National Institute on Aging and the Alzheimer's Association convened three separate work groups in 2011 and single work groups in 2012 and 2018 to create recommendations for the diagnosis and characterization of Alzheimer's disease (AD)."

Show evidence (1 reference)

DOI:10.1002/alz.13859 SUPPORT Human Clinical

Deep research cited this publication as relevant literature for Alzheimer Disease.

DOI:10.1038/s41467-023-37437-5

Single-nucleus RNA-sequencing of autosomal dominant Alzheimer disease and risk variant carriers

1 finding

Genetic studies of Alzheimer disease (AD) have prioritized variants in genes related to the amyloid cascade, lipid metabolism, and neuroimmune modulation.

"Genetic studies of Alzheimer disease (AD) have prioritized variants in genes related to the amyloid cascade, lipid metabolism, and neuroimmune modulation."

Show evidence (1 reference)

DOI:10.1038/s41467-023-37437-5 SUPPORT Human Clinical

"Genetic studies of Alzheimer disease (AD) have prioritized variants in genes related to the amyloid cascade, lipid metabolism, and neuroimmune modulation."

Deep research cited this publication as relevant literature for Alzheimer Disease.

DOI:10.1186/s13195-024-01469-w

Combining plasma Aβ and p-tau217 improves detection of brain amyloid in non-demented elderly

1 finding

Combining plasma Aβ and p-tau217 improves detection of brain amyloid in non-demented elderly

"Maximizing the efficiency to screen amyloid-positive individuals in asymptomatic and non-demented aged population using blood-based biomarkers is essential for future success of clinical trials in the early stage of Alzheimer’s disease (AD)."

Show evidence (1 reference)

DOI:10.1186/s13195-024-01469-w SUPPORT Human Clinical

Deep research cited this publication as relevant literature for Alzheimer Disease.

Deep Research

Deep research results are used as seeds for research; they do not undergo the same validation as the main records and may contain errors. How we use deep research.

Evaluations and curation notes (2)

Add verified antiviral prevention-vs-treatment evidence (VALAD RCT + meta-analyses) from OpenScientist mining · 2026-07-24T07:30:16Z · View source

Attempted fresh OpenScientist deep-research runs on the hsv1_reactivation_model hypothesis (twice); both hit the provider's 2h job ceiling and produced no output. Mined the existing OpenScientist report (kb/hypotheses/Alzheimer_Disease/hsv1_reactivation_model/openscientist.md) instead. Added three independently verified evidence items to the existing antiviral-suppression knowledge-gap discussion: VALAD phase II RCT (PMID:41405855, JAMA 2026; valacyclovir no benefit/possible harm in early symptomatic AD, ADAS-Cog +3.93 P=.01) as REFUTE for the treatment corollary; anti-herpetic meta-analysis (PMID:41467972; HSV+VZV, aHR 0.77) and TriNetX antiviral cohort (PMID:41779765; RR 0.87 AD / 0.83 dementia) as SUPPORT for the prevention corollary. This mirrors the zoster-vaccine prevention-vs-treatment split. Per the DR anti-hallucination SOP, every PMID and snippet was verified against the fetched abstract; one DR-reported number was caught misattributed (report cited OR=1.46 for PMID:40934136 whose real abstract reports OR=1.32/20%) and was not used. No mechanistic nodes or ontology terms changed. Schema validation passed.

Refine adaptive_immune_tcell_model from verified OpenScientist DR leads · 2026-07-11T14:33:15Z · View source

Folded verified leads from an OpenScientist hypothesis deep-research run into the adaptive-immune T-cell model. (1) Tau-scoped the IFN-gamma/PD-1 therapeutic prediction and recorded the amyloid-context valence reversal (added PMID:41890046 SUPPORT cDC1 priming; PMID:26779813 PARTIAL amyloid PD-1 benefit). (2) Enriched the 'Adaptive Immune T Cell Response to Tau Pathology' node with cDC1-priming and granzyme K-PAR-1 mechanisms and evidence PMID:41890046, PMID:40993111 (CD103-negative granzyme-K CD8; sign corrected from DR report), and human PMID:31915375 (clonal CD8 TEMRA in AD CSF). (3) Added two KNOWLEDGE_GAP discussions: antigen identity vs antigen-independent virtual-memory CD8 (PMID:42432776) and EBV-specific CSF TCRs (PMID:31915375); and IFN-gamma/PD-1 valence reversal plus NK-cell IFN-gamma source (PMID:11268360). All evidence snippets independently verified as exact abstract substrings; schema/term/reference/folded-hyphen checks pass.

Falcon ▸

Disease Characteristics Research Template

Edison Scientific Literature 20 citations 2026-05-08T08:50:57.133618

Question: You are an expert researcher providing comprehensive, well-cited information.

Provide detailed information focusing on: 1. Key concepts and definitions with current understanding 2. Recent developments and latest research (prioritize 2023-2024 sources) 3. Current applications and real-world implementations 4. Expert opinions and analysis from authoritative sources 5. Relevant statistics and data from recent studies

Format as a comprehensive research report with proper citations. Include URLs and publication dates where available. Always prioritize recent, authoritative sources and provide specific citations for all major claims.

Disease Characteristics Research Template

Target Disease

Disease Name: Alzheimer Disease
MONDO ID: (if available)
Category: Neurodegenerative Disorder

Research Objectives

Please provide a comprehensive research report on Alzheimer Disease covering all of the disease characteristics listed below. This report will be used to populate a disease knowledge base entry. Be thorough and cite primary literature (PMID preferred) for all claims.

For each section, suggested databases/resources are listed. These are the first places you should search for information on each topic.

1. Disease Information

Search first: OMIM, Orphanet, ICD-10/ICD-11, MeSH, PubMed

What is the disease? Provide a concise overview.
What are the key identifiers? (OMIM, Orphanet, ICD-10/ICD-11, MeSH, Mondo)
What are the common synonyms and alternative names?
Is the information derived from individual patients (e.g., EHR) or aggregated disease-level resources?

2. Etiology

Disease Causal Factors: What are the primary causes? (genetic, environmental, infectious, mechanistic)
Risk Factors:

Search first: PubMed, Cochrane Library, UpToDate, clinical guidelines, ClinVar, ClinGen, GWAS Catalog, PheGenI, CTD, CDC, WHO, epidemiological databases
Genetic risk factors (causal variants, susceptibility loci, modifier genes)
Environmental risk factors (toxins, lifestyle, occupational exposures, age, sex, family history)
Protective Factors:

Search first: PubMed, Cochrane Library, clinical trial databases, GWAS Catalog, gnomAD, WHO, CDC, nutrition databases
Genetic protective factors (protective variants, modifier alleles)
Environmental protective factors (diet, lifestyle, exposures that reduce risk)
Gene-Environment Interactions: How do genetic and environmental factors interact to influence disease?

Search first: CTD, PubMed, PheGenI, GxE databases

3. Phenotypes

Search first: HPO (Human Phenotype Ontology), OMIM, Orphanet, PubMed, clinicaltrials.gov, MedDRA, SNOMED CT, DECIPHER, LOINC

For each phenotype, provide: - Phenotype type: symptoms, clinical signs, physical manifestations, behavioral changes, or laboratory abnormalities

For symptoms/signs: HPO, OMIM, Orphanet, PubMed For behavioral changes: HPO, DSM, RDoC (Research Domain Criteria), PubMed For laboratory abnormalities: LOINC, SNOMED CT, LabTests Online, PubMed - Phenotype characteristics: Search first: OMIM, Orphanet, HPO, PubMed - Age of symptom onset (neonatal, childhood, adult-onset, late-onset) - Symptom severity (mild, moderate, severe, variable) - Symptom progression (stable, progressive, episodic, fluctuating) - Frequency among affected individuals (percentage or qualitative) - Quality of life impact: Effects on daily functioning and well-being (per-phenotype when possible) Search first: EQ-5D database, SF-36, WHO QOL databases, PubMed - Suggest HPO (Human Phenotype Ontology) terms for each phenotype

4. Genetic/Molecular Information

Causal Genes: Gene mutations or chromosomal abnormalities responsible for disease (gene symbols, OMIM IDs)

Search first: OMIM, ClinVar, HGMD, Ensembl, NCBI Gene
Pathogenic Variants:
Affected genes (gene symbols, HGNC IDs) > Search first: OMIM, NCBI Gene, Ensembl, HGNC, UniProt, GeneCards
Variant classification (pathogenic, likely pathogenic, VUS per ACMG/AMP guidelines) > Search first: ClinVar, ClinGen, ACMG/AMP guidelines, VarSome
Variant type/class (missense, frameshift, nonsense, splice-site, structural)
Allele frequency in population databases > Search first: gnomAD, 1000 Genomes, ExAC, TOPMed, dbSNP
Somatic vs germline origin > Search first: COSMIC (somatic), ClinVar, ICGC, TCGA
Functional consequences (loss of function, gain of function, dominant negative)
Modifier Genes: Genes that modify disease severity or expression
Epigenetic Information: DNA methylation, histone modifications, chromatin changes affecting disease

Search first: ENCODE, Roadmap Epigenomics, MethBase, DiseaseMeth
Chromosomal Abnormalities: Large-scale genetic changes (aneuploidy, translocations, inversions)

Search first: DECIPHER, ClinVar, ECARUCA, UCSC Genome Browser

5. Environmental Information

Environmental Factors: Non-genetic contributing factors (toxins, radiation, pollution, occupational exposure)

Search first: CTD (Comparative Toxicogenomics Database), TOXNET, PubMed, EPA databases
Lifestyle Factors: Behavioral factors (smoking, diet, exercise, alcohol consumption)

Search first: CDC databases, WHO, PubMed, NHANES
Infectious Agents: If applicable, pathogens causing or triggering disease (bacteria, viruses, fungi, parasites)

Search first: NCBI Taxonomy, ViPR, BV-BRC, MicrobeDB, GIDEON

6. Mechanism / Pathophysiology

Molecular Pathways: Specific signaling cascades or biochemical pathways involved (Wnt, MAPK, mTOR, PI3K-AKT, etc.)

Search first: KEGG, Reactome, WikiPathways, PathBank, BioCyc
Cellular Processes: Cell-level mechanisms (apoptosis, autophagy, cell cycle dysregulation, inflammation, etc.)

Search first: Gene Ontology (GO), Reactome, KEGG, PubMed
Protein Dysfunction: How protein structure or function is altered (misfolding, aggregation, loss of function, gain of function)

Search first: UniProt, PDB (Protein Data Bank), InterPro, Pfam, AlphaFold
Metabolic Changes: Alterations in metabolic processes (energy metabolism, lipid metabolism, amino acid metabolism)

Search first: KEGG, BioCyc, HMDB (Human Metabolome Database), BRENDA
Immune System Involvement: Role of immune response (autoimmunity, immunodeficiency, chronic inflammation)

Search first: ImmPort, Immunome Database, IEDB, Gene Ontology
Tissue Damage Mechanisms: How tissues/ are injured (oxidative stress, ischemia, fibrosis, necrosis)

Search first: PubMed, Gene Ontology, Reactome
Biochemical Abnormalities: Specific molecular defects (enzyme deficiencies, receptor dysfunction, ion channel defects)

Search first: BRENDA, UniProt, KEGG, OMIM, PubMed
Epigenetic Changes: DNA methylation, histone modifications affecting gene expression in disease

Search first: ENCODE, Roadmap Epigenomics, MethBase, DiseaseMeth
Molecular Profiling (if available):
Transcriptomics/gene expression changes > Search first: GEO (Gene Expression Omnibus), ArrayExpress, GTEx, Human Cell Atlas, SRA
Proteomics findings > Search first: PRIDE, ProteomeXchange, Human Protein Atlas, STRING, BioGRID
Metabolomics signatures > Search first: MetaboLights, Metabolomics Workbench, HMDB, METLIN
Lipidomics alterations > Search first: LIPID MAPS, SwissLipids, LipidHome, Metabolomics Workbench
Genomic structural features > Search first: UCSC Genome Browser, Ensembl, NCBI, dbVar, DGV
Advanced Technologies (if applicable):
Single-cell analysis findings (cell-type specific mechanisms, cellular heterogeneity) > Search first: Human Cell Atlas, Single Cell Portal, GEO, CELLxGENE
Spatial transcriptomics findings > Search first: GEO, Spatial Research, Vizgen, 10x Genomics data
Multi-omics integration results > Search first: TCGA, ICGC, cBioPortal, LinkedOmics, PubMed
Functional genomics screens (CRISPR, RNAi) > Search first: DepMap, GenomeRNAi, PubMed, BioGRID ORCS

For each mechanism, describe: - The causal chain from initial trigger to clinical manifestation - Which mechanisms are upstream vs downstream - What cell types and biological processes are involved - Suggest GO terms for biological processes and CL terms for cell types

7. Anatomical Structures Affected

Organ Level:
Primary organs directly affected
Secondary organ involvement (complications, secondary effects)
Body systems involved (cardiovascular, nervous, digestive, respiratory, endocrine, etc.)

Search first: Uberon, FMA (Foundational Model of Anatomy), OMIM, HPO, ICD-11, MeSH, SNOMED CT
Tissue and Cell Level:
Specific tissue types affected (epithelial, connective, muscle, nervous)
Specific cell populations targeted (with Cell Ontology terms)

Search first: Uberon, Human Protein Atlas, Cell Ontology, Human Cell Atlas, CellMarker, PanglaoDB
Subcellular Level:
Cellular compartments involved (mitochondria, nucleus, ER, lysosomes) (with GO Cellular Component terms)

Search first: Gene Ontology (Cellular Component), UniProt, Human Protein Atlas
Localization:
Specific anatomical sites (with UBERON terms) > Search first: FMA, Uberon, NeuroNames (for brain), SNOMED CT
Lateralization (unilateral, bilateral, asymmetric) > Search first: HPO, clinical literature, imaging databases

8. Temporal Development

Onset:
Typical age of onset (congenital, pediatric, adult, geriatric)
Onset pattern (acute, subacute, chronic, insidious)

Search first: OMIM, Orphanet, HPO, PubMed
Progression:
Disease stages (early, intermediate, advanced, end-stage) > Search first: Cancer Staging Manual (AJCC), WHO classifications, PubMed
Progression rate (rapid, slow, variable)
Disease course pattern (episodic, relapsing-remitting, progressive, stable)
Disease duration (self-limited, chronic lifelong)

Search first: Disease registries, longitudinal cohort databases, natural history studies, PubMed, Orphanet, OMIM
Patterns:
Remission patterns (spontaneous, treatment-induced) > Search first: Clinical trial databases, disease registries, PubMed
Critical periods (time windows of vulnerability or opportunity for intervention) > Search first: PubMed, developmental biology databases, clinical guidelines

9. Inheritance and Population

Epidemiology:
Prevalence (cases per 100,000 at given time)
Incidence (new cases per 100,000 per year)

Search first: Orphanet, CDC, WHO, GBD (Global Burden of Disease), national registries, SEER, disease registries
For Genetic Etiology:
Inheritance pattern (AD, AR, X-linked, mitochondrial, multifactorial, polygenic) > Search first: OMIM, Orphanet, ClinVar, GTR (Genetic Testing Registry)
Penetrance (complete, incomplete, age-dependent) > Search first: ClinVar, OMIM, PubMed, ClinGen
Expressivity (variable, consistent) > Search first: OMIM, ClinVar, PubMed
Genetic anticipation (increasing severity in successive generations) > Search first: OMIM, PubMed (especially for repeat expansion disorders)
Germline mosaicism > Search first: ClinVar, OMIM, genetic counseling literature, PubMed
Founder effects (population-specific mutations) > Search first: gnomAD, population genetics databases, PubMed
Consanguinity role > Search first: OMIM, population studies, genetic counseling resources
Carrier frequency > Search first: gnomAD, carrier screening databases, GeneReviews, GTR
Population Demographics:
Affected populations (ethnic or demographic groups with higher prevalence) > Search first: gnomAD, 1000 Genomes, PAGE Study, PubMed, population registries
Geographic distribution (endemic areas, regional variation) > Search first: WHO, CDC, GBD, Orphanet, geographic epidemiology databases
Geographic distribution of specific variants
Sex ratio (male:female) > Search first: Disease registries, OMIM, PubMed, epidemiological databases
Age distribution of affected individuals > Search first: CDC, disease registries, SEER, Orphanet

10. Diagnostics

Clinical Tests:
Laboratory tests (blood, urine, tissue chemistry, specific enzyme assays) > Search first: LOINC, LabTests Online, PubMed
Biomarkers (proteins, metabolites, genetic markers, circulating biomarkers) > Search first: FDA Biomarker List, BEST (Biomarkers, EndpointS, and other Tools), PubMed
Imaging studies (X-ray, CT, MRI, PET, ultrasound) > Search first: RadLex, DICOM, Radiopaedia, imaging databases
Functional tests (pulmonary function, cardiac stress tests) > Search first: LOINC, clinical guidelines, PubMed
Electrophysiology (EEG, EMG, ECG, nerve conduction studies) > Search first: LOINC, clinical neurophysiology databases, PubMed
Biopsy findings (histopathology, immunohistochemistry) > Search first: SNOMED CT, College of American Pathologists resources, PubMed
Pathology findings (microscopic examination) > Search first: SNOMED CT, Digital Pathology databases, PubMed
Genetic Testing:

Search first: GTR (Genetic Testing Registry), GeneReviews, ClinGen
Overview of recommended genetic testing approach
Whole genome sequencing (WGS) utility > Search first: GTR, ClinVar, GEL (Genomics England), gnomAD
Whole exome sequencing (WES) utility > Search first: GTR, ClinVar, OMIM, GeneMatcher
Gene panels (which panels, which genes) > Search first: GTR, ClinVar, laboratory-specific databases
Single gene testing > Search first: GTR, ClinVar, OMIM, GeneReviews
Chromosomal microarray (CMA) > Search first: DECIPHER, ClinVar, dbVar, ECARUCA
Karyotyping > Search first: Chromosome Abnormality Database, ClinVar, cytogenetics resources
FISH > Search first: ClinVar, cytogenetics databases, PubMed
Mitochondrial DNA testing > Search first: MITOMAP, MSeqDR, ClinVar, GTR
Repeat expansion testing > Search first: GTR, ClinVar, repeat expansion databases, PubMed
Omics-Based Diagnostics (if applicable):
RNA sequencing / transcriptomics > Search first: GEO, ArrayExpress, GTEx, RNA-seq databases
Proteomics > Search first: PRIDE, ProteomeXchange, FDA Biomarker database
Metabolomics > Search first: MetaboLights, Metabolomics Workbench, HMDB
Epigenomics > Search first: GEO, ENCODE, Roadmap Epigenomics, MethBase
Liquid biopsy > Search first: COSMIC, ClinVar, liquid biopsy databases, PubMed
Clinical Criteria:
Standardized diagnostic criteria (DSM, ICD, society guidelines) > Search first: DSM-5, ICD-11, clinical society guidelines, UpToDate
Differential diagnosis (other conditions to rule out, with distinguishing features) > Search first: DynaMed, UpToDate, clinical decision support systems
Screening:
Screening methods for asymptomatic individuals (newborn screening, carrier screening, cascade screening) > Search first: ACMG recommendations, CDC newborn screening, GTR

11. Outcome/Prognosis

Survival and Mortality:
Survival rate (5-year, 10-year, overall) > Search first: SEER, cancer registries, disease-specific registries, PubMed
Life expectancy (with and without treatment if applicable) > Search first: Orphanet, disease registries, actuarial databases, PubMed
Mortality rate > Search first: CDC, WHO, GBD, national mortality databases
Disease-specific mortality (deaths directly attributable to disease) > Search first: Disease registries, CDC Wonder, GBD, PubMed
Morbidity and Function:
Morbidity (disease-related disability and health impacts) > Search first: GBD, WHO, disability databases, PubMed
Disability outcomes (long-term functional impairments) > Search first: ICF (International Classification of Functioning), disability registries
Quality of life measures (EQ-5D, SF-36, PROMIS, disease-specific tools) > Search first: EQ-5D database, SF-36, PROMIS, PubMed
Disease Course:
Complications (secondary problems: infections, organ failure, etc.) > Search first: ICD codes, disease registries, clinical databases, PubMed
Recovery potential (likelihood and extent of recovery, with vs without treatment) > Search first: Natural history studies, rehabilitation databases, PubMed
Prediction:
Prognostic factors (age, disease severity, biomarkers, treatment response) > Search first: Prognostic models databases, clinical calculators, PubMed
Prognostic biomarkers (molecular markers predicting disease course) > Search first: FDA Biomarker database, PubMed, cancer prognostic databases

12. Treatment

Pharmacotherapy:
Pharmacological treatments (drug names, drug classes, mechanisms of action) > Search first: DrugBank, RxNorm, ATC classification, DailyMed, FDA databases
Pharmacogenomics (how genetic variants affect drug metabolism, efficacy, toxicity) > Search first: PharmGKB, CPIC (Clinical Pharmacogenetics), FDA Table of PGx Biomarkers
Advanced Therapeutics:
Gene therapy (viral vectors, CRISPR, gene replacement, gene editing) > Search first: ClinicalTrials.gov, FDA gene therapy database, ASGCT resources
Cell therapy (stem cell transplant, CAR-T, cellular therapeutics) > Search first: ClinicalTrials.gov, FDA cell therapy database, FACT standards
RNA-based therapies (ASOs, siRNA, mRNA therapies) > Search first: ClinicalTrials.gov, FDA approvals, PubMed
Targeted therapies (treatments directed at specific molecular targets) > Search first: My Cancer Genome, OncoKB, ClinicalTrials.gov, FDA approvals
Immunotherapies (checkpoint inhibitors, monoclonal antibodies) > Search first: Cancer Immunotherapy Database, FDA approvals, ClinicalTrials.gov
Surgical and Interventional:
Surgical interventions (types of surgery, timing, outcomes) > Search first: CPT codes, surgical registries, clinical guidelines, PubMed
Supportive and Rehabilitative:
Supportive care (symptom management, pain control, nutrition) > Search first: Clinical guidelines, Cochrane Library, PubMed
Rehabilitation (physical therapy, occupational therapy, speech therapy) > Search first: Rehabilitation medicine databases, clinical guidelines, PubMed
Experimental:
Experimental treatments in clinical trials (with NCT identifiers if available) > Search first: ClinicalTrials.gov, EU Clinical Trials Register, WHO ICTRP
Treatment Outcomes:
Treatment response rates > Search first: Clinical trial databases, FDA reviews, systematic reviews, PubMed
Side effects and adverse events > Search first: FDA Adverse Event Reporting System (FAERS), MedWatch, PubMed
Treatment Strategy:
Treatment algorithms (clinical pathways, decision trees) > Search first: Clinical practice guidelines, NCCN Guidelines, UpToDate
Combination therapies > Search first: ClinicalTrials.gov, treatment guidelines, PubMed
Personalized medicine approaches (genotype-guided treatment) > Search first: My Cancer Genome, CIViC, PharmGKB, precision medicine databases

For each treatment, suggest MAXO (Medical Action Ontology) terms where applicable.

13. Prevention

Prevention Levels:
Primary prevention (preventing disease occurrence: vaccination, risk factor modification) > Search first: CDC, WHO, USPSTF recommendations, Cochrane Library
Secondary prevention (early detection and treatment: screening programs, early intervention) > Search first: USPSTF, CDC screening guidelines, WHO
Tertiary prevention (preventing complications in those with disease) > Search first: Clinical guidelines, disease management protocols, PubMed
Immunization: Vaccine strategies (if applicable)

Search first: CDC vaccine schedules, WHO immunization, FDA vaccine database
Screening and Early Detection:
Screening programs (population-based: newborn screening, cancer screening) > Search first: CDC screening programs, USPSTF, cancer screening databases
Genetic screening (carrier screening, preimplantation genetic diagnosis, prenatal testing) > Search first: ACMG recommendations, ACOG guidelines, GTR
Risk stratification (identifying high-risk individuals for targeted prevention) > Search first: Risk prediction models, clinical calculators, PubMed
Behavioral Interventions: Lifestyle modifications to reduce risk

Search first: CDC, WHO, behavioral intervention databases, Cochrane Library
Counseling: Genetic counseling (risk assessment, family planning guidance)

Search first: NSGC resources, ACMG guidelines, GeneReviews
Public Health:
Public health interventions (sanitation, vector control, health education) > Search first: CDC, WHO, public health databases, PubMed
Environmental interventions (reducing environmental risk factors) > Search first: EPA databases, WHO environmental health, PubMed
Prophylaxis: Preventive medications or procedures

Search first: Clinical guidelines, FDA approvals, PubMed

14. Other Species / Natural Disease

Taxonomy: Species affected (with NCBI Taxon identifiers)

Search first: NCBI Taxonomy
Breed: Specific breeds affected (with VBO identifiers if applicable)

Search first: VBO (Vertebrate Breed Ontology)
Gene: Orthologous genes in other species (with NCBI Gene IDs)

Search first: NCBI Gene
Natural Disease:
Naturally occurring disease in other species (companion animals, wildlife) > Search first: OMIA (Online Mendelian Inheritance in Animals), VetCompass, PubMed
Veterinary relevance and importance in animal health > Search first: OMIA, veterinary databases, PubMed
Comparative Biology:
Comparative pathology (similarities and differences across species) > Search first: OMIA, comparative pathology databases, PubMed
Evolutionary conservation of disease mechanisms > Search first: HomoloGene, OrthoMCL, Alliance of Genome Resources
Transmission (if applicable):
Zoonotic potential > Search first: CDC zoonotic diseases, WHO zoonoses, GIDEON
Cross-species susceptibility > Search first: NCBI Taxonomy, veterinary databases, PubMed

15. Model Organisms

Model Types:
Model organism type (mammalian, invertebrate, cellular, in vitro) > Search first: Alliance of Genome Resources, model organism databases
Specific model systems (mouse, rat, zebrafish, Drosophila, C. elegans, yeast, cell lines, organoids, iPSCs) > Search first: MGI, RGD, ZFIN, FlyBase, WormBase, SGD, ATCC, Cellosaurus
Induced models (drug treatment, surgical intervention, environmental manipulation) > Search first: MGI, model organism databases, PubMed
Genetic Models:
Types available (knockout, knock-in, transgenic, conditional, humanized) > Search first: MGI, IMPC, KOMP, EuMMCR, IMSR
Model Characteristics:
Phenotype recapitulation (how well model reproduces human disease features) > Search first: Model organism databases, comparative studies, PubMed
Model limitations (aspects of human disease not captured) > Search first: Model organism databases, PubMed, review articles
Applications:
Research applications (what aspects of disease can be studied) > Search first: Model organism databases, PubMed
Resources:
Model databases > Search first: MGI, RGD, ZFIN, FlyBase, WormBase, IMSR, EMMA, MMRRC

Citation Requirements

Cite primary literature (PMID preferred) for all mechanistic and clinical claims
Prioritize recent reviews and landmark papers
Include direct quotes from abstracts where possible to support key statements
Distinguish evidence source types: human clinical, model organism, in vitro, computational

Output Format

Structure your response as a comprehensive narrative organized by the sections above. For each section, provide: - Factual content with specific details (numbers, percentages, gene names, variant nomenclature) - Ontology term suggestions (HPO, GO, CL, UBERON, CHEBI, MAXO, MONDO) where applicable - Evidence citations with PMIDs - Direct quotes from abstracts to support key claims - Clear indication when information is not available or not applicable for this disease

This report will be used to populate a disease knowledge base entry with: - Pathophysiology descriptions with causal chains - Gene/protein annotations (HGNC, GO terms) - Phenotype associations (HP terms) with frequencies - Cell type involvement (CL terms) - Anatomical locations (UBERON terms) - Chemical entities (CHEBI terms) - Treatment annotations (MAXO terms) - Evidence items with PMIDs and exact abstract quotes - Epidemiology, prognosis, diagnostic, and prevention information - Animal model descriptions with phenotype recapitulation details

Alzheimer Disease (AD) — Comprehensive Disease Characteristics Research Report (2023–2024 emphasis)

Executive summary

Alzheimer disease (AD) is increasingly framed as a biologically defined neurodegenerative disease characterized by amyloid-β (Aβ) plaque pathology and phosphorylated tau (p-tau) neurofibrillary tangle pathology, with neurodegeneration and clinical syndromes (e.g., mild cognitive impairment and dementia) occurring variably along this continuum. The most influential 2024 development is the Alzheimer’s Association (AA) Workgroup revised criteria proposing that an abnormal “Core 1” biomarker (amyloid PET, approved CSF assays, or sufficiently accurate plasma assays) can be sufficient to establish a biological diagnosis of AD, enabling treatment selection and staging across symptomatic and asymptomatic phases. (jack2024revisedcriteriafor pages 1-2)

A major applied 2023–2024 trend is the rapid maturation of blood-based biomarkers (BBMs), especially plasma p-tau217, with multiple studies reporting AUCs ≈0.9 or higher for detecting AD-related Aβ pathology and workflows that may reduce invasive testing in specialist memory services. (dyer2024performanceofplasma pages 10-11, niimi2024combiningplasmaaβ pages 7-8)

High-value recent sources (quick reference)

Topic	Source (first author, journal)	Publication date (month year)	URL/DOI	Key quantitative finding(s) extracted from evidence	Evidence type (human clinical / cohort / guideline / omics)
Biological definition and staging criteria for AD	Jack et al., Alzheimer's & Dementia	June 2024	https://doi.org/10.1002/alz.13859	2024 AA workgroup defines AD biologically; an abnormal Core 1 biomarker can establish AD. For standalone plasma use, Core 1 blood biomarkers should achieve ≥90% accuracy versus amyloid PET. Florbetapir PET visual reads showed 96% sensitivity / 100% specificity versus CERAD neuritic plaque reference; approved CSF assays showed about 88%/93% and 85%/94% sensitivity/specificity versus amyloid PET visual reads. Neuropathology concordance: among symptomatic individuals with moderate/frequent plaques, 4390/4637 (95%) were Braak III–VI; among cognitively unimpaired decedents with moderate/frequent plaques, 107/123 (87%) were Braak III–VI; in a larger NACC sample, 186/252 (74%) were Braak III–VI and 226/252 (91%) were Braak II–VI (jack2024revisedcriteriafor pages 8-9, jack2024revisedcriteriafor pages 1-2)	Guideline
Plasma p-tau217 in real-world memory clinic diagnosis	Dyer et al., Alzheimer's Research & Therapy	August 2024	https://doi.org/10.1186/s13195-024-01555-z	In a symptomatic memory-clinic cohort, plasma p-tau217 detected Aβ pathology with AUC 0.91 and outperformed plasma p-tau181 (AUC 0.73). A two-threshold strategy suggested confirmatory lumbar puncture could potentially be avoided in 68% of cases at 95% sensitivity / 95% specificity, or 58% at 97.5% / 97.5% (dyer2024performanceofplasma pages 10-11)	Human clinical / cohort
Combined plasma Aβ and p-tau217 models for Aβ-PET prediction	Niimi et al., Alzheimer's Research & Therapy	May 2024	https://doi.org/10.1186/s13195-024-01469-w	In non-demented Japanese J-TRC participants, the best total-cohort model reached AUC 0.936 for p-tau217/Aβ42 + APOE + age + sex; in CDR 0 the best model reached AUC 0.948 for p-tau217 + Aβ42/40 + APOE + age + sex; in CDR 0.5 the best model reached AUC 0.955 for p-tau217/Aβ42 + APOE + age + sex. Individual p-tau217 AUCs were 0.913 (total), 0.889 (CDR 0), and 0.925 (CDR 0.5) (niimi2024combiningplasmaaβ pages 7-8)	Cohort
Variant-aware single-nucleus transcriptomics in AD	Brase et al., Nature Communications	April 2023	https://doi.org/10.1038/s41467-023-37437-5	snRNA-seq of nearly 300,000 nuclei from parietal cortex of autosomal-dominant AD and risk-variant carriers identified variant-specific cell states: TREM2 oligodendrocytes showed dysregulated autophagy-lysosomal pathways; MS4A microglia showed dysregulated complement cascade genes; APOE ε4 inhibitory neurons showed signatures of ferroptosis. The paper also reports dose-dependent enrichment of an MS4A rs1582763-A pro-inflammatory microglial state and astrocyte activation trends in carriers (brase2023singlenucleusrnasequencingof pages 7-8)	Omics
Modifiable risk factors and prevention potential	Parums, Medical Science Monitor	May 2024	https://doi.org/10.12659/msm.945091	Summarizing Lancet Commission prevention evidence, the review states 12 modifiable risk factors (education, hypertension, hearing impairment, obesity, smoking, depression, physical inactivity, social isolation, diabetes, alcohol, traumatic brain injury, air pollution) may account for up to 40% of dementia cases worldwide (parums2024areviewof pages 4-6)	Review / public health synthesis
Real-world implementation of anti-amyloid therapy	Jessen et al., Journal of Prevention of Alzheimer's Disease	October 2024	https://doi.org/10.14283/jpad.2024.153	EADC position statement argues anti-amyloid antibodies have meaningful though modest effects and manageable adverse effects, and recommends clinical use in selected patients with treatment documentation in registries. It emphasizes that the number eventually treated will be only a fraction of all early AD patients because of narrow eligibility and access barriers; cited meta-analysis estimate indicates 40% of dementia population-attributable risk relates to modifiable factors (jessen2024progressinthe pages 1-2)	Guideline / expert consensus

Table: This table compiles high-yield 2023–2024 Alzheimer disease sources with the most decision-relevant quantitative findings for diagnosis, biomarkers, omics, prevention, and implementation. It is useful as a quick-reference evidence map for a disease knowledge base.

1. Disease information

1.1 Concise overview

The AA Workgroup (2024) defines AD as beginning with AD neuropathologic change (ADNPC) and emphasizes that AD pathology can be present before symptoms. The Workgroup’s intent is to provide objective diagnostic and staging criteria that bridge research and clinical care (not step-by-step clinical workflow protocols). (jack2024revisedcriteriafor pages 1-2)

1.2 Key identifiers and ontology codes

MONDO ID: Open Targets lists MONDO_0004975 for Alzheimer disease. (OpenTargets Search: Alzheimer disease)
Other identifiers (ICD-10/ICD-11/MeSH/OMIM/Orphanet): Not extracted from the retrieved primary sources in this run; should be populated from OMIM/MeSH/ICD registries directly.

1.3 Common synonyms / alternative names

Within the retrieved sources, AD is discussed both as a clinical syndrome (e.g., mild dementia) and as “biological AD” defined by biomarkers/pathology. (jack2024revisedcriteriafor pages 1-2, jessen2024progressinthe pages 1-2)

1.4 Evidence source type

The evidence in this report is derived from: * Aggregated disease-level resources and consensus criteria (AA Workgroup). (jack2024revisedcriteriafor pages 1-2) * Human observational cohorts and real-world memory clinic studies for plasma biomarkers. (dyer2024performanceofplasma pages 10-11, niimi2024combiningplasmaaβ pages 7-8) * Human multi-omic / single-nucleus transcriptomics in postmortem brain. (brase2023singlenucleusrnasequencingof pages 7-8) * Expert position statements / reviews for implementation and prevention. (jessen2024progressinthe pages 1-2, parums2024areviewof pages 4-6)

2. Etiology

2.1 Disease causal factors (current understanding)

Biological drivers emphasized across authoritative sources include Aβ deposition and tau aggregation. The EADC investigators summarize AD biologically as extracellular β-amyloid plaques plus intraneuronal phosphorylated tau with ensuing neurodegeneration, detectable by CSF and PET biomarkers (often already at the MCI stage). (jessen2024progressinthe pages 1-2)

2.2 Genetic risk factors (selected; recent mechanistic emphasis)

The 2023 single-nucleus atlas of autosomal dominant AD and risk-variant carriers highlights that AD genetic architecture maps to cell-type–specific transcriptional states and pathways (microglia, astrocytes, oligodendrocytes, neurons), supporting a view that genetic risk influences AD through neuroimmune, lysosomal/autophagy, complement, and neuronal stress-death programs. (brase2023singlenucleusrnasequencingof pages 7-8)

Ontology suggestions (genes): APP, PSEN1, PSEN2, APOE, TREM2, MS4A locus genes (HGNC symbols).

2.3 Protective factors

The centenarian resilience study was retrieved but the extractable quantitative protective-allele statistics were not captured in the evidence snippets returned by the tools in this run; this remains a gap for this report’s citation-backed content.

2.4 Gene–environment interactions

Not directly quantified in the retrieved evidence.

3. Phenotypes

3.1 Core clinical phenotypes (high-level)

The clinical syndrome is described in the retrieved literature as progressive cognitive decline with functional impairment, and trials/biomarker studies commonly focus on early symptomatic phases (MCI/mild dementia). (dyer2024performanceofplasma pages 10-11, jessen2024progressinthe pages 1-2, niimi2024combiningplasmaaβ pages 7-8)

3.2 HPO term suggestions (non-exhaustive; for knowledge base structuring)

Memory impairment: HP:0002354
Cognitive impairment: HP:0100543
Dementia: HP:0000726
Executive dysfunction: HP:0000726 (or more specific executive-function terms as needed)

Frequency/severity/progression statistics: Not systematically extracted from primary clinical phenotype cohorts in the current evidence set.

4. Genetic / molecular information

4.1 Molecular hallmarks

AD’s defining lesions include Aβ plaques and tau pathology; the AA Workgroup uses biomarker mapping to these lesions to define biological AD. (jack2024revisedcriteriafor pages 1-2)

4.2 Pathogenic variants and variant classes

Pathogenic-variant details (specific amino-acid changes, allele frequencies, ClinVar classifications) were not captured in the available evidence. Open Targets lists major AD-associated targets (e.g., APP, PSEN1, PSEN2, APOE, SORL1) but does not provide variant-level detail in the retrieved snapshot. (OpenTargets Search: Alzheimer disease)

4.3 Epigenetic information and chromosomal abnormalities

Not extracted in the retrieved evidence.

5. Environmental information

5.1 Lifestyle and environmental risk factors (modifiable)

A prevention-oriented synthesis reports the 2020 Lancet Commission’s 12 modifiable risk factors (education, hypertension, hearing impairment, obesity, smoking, depression, physical inactivity, social isolation, diabetes, alcohol, traumatic brain injury, air pollution) and states these can account for up to ~40% of dementia cases worldwide. (parums2024areviewof pages 4-6)

CHEBI suggestions (exposures): nitrogen dioxide (NO2), ethanol.

5.2 Infectious agents

No evidence in retrieved sources supporting a specific infectious etiology for typical AD.

6. Mechanism / pathophysiology

6.1 Updated causal chain (biomarker-informed)

The AA Workgroup positions AD as starting with ADNPC (biological disease), detectable by Core 1 biomarkers, with later biomarkers providing prognostic staging and increasing confidence that AD pathology contributes to symptoms. (jack2024revisedcriteriafor pages 1-2)

Core 1 biomarkers include amyloid PET, approved CSF biomarkers, and sufficiently accurate plasma biomarkers (notably plasma p-tau217 in some assays). (jack2024revisedcriteriafor pages 1-2)

6.2 Microglia/astrocyte/neuron/oligodendrocyte programs (single-nucleus evidence)

In a large snRNA-seq study of autosomal dominant AD and risk-variant carriers, variant-specific states were reported, including: * TREM2 oligodendrocytes with dysregulated autophagy–lysosomal pathway. * MS4A microglia with dysregulated complement cascade genes. * APOE ε4 inhibitory neurons showing signatures consistent with ferroptosis. These findings support a multi-cell-type mechanistic model in which glial activation states and neuronal stress-death programs interact with genetic risk architecture. (brase2023singlenucleusrnasequencingof pages 7-8)

6.3 Pathways / ontology suggestions

GO Biological Process (examples): * Amyloid-beta metabolic process * Tau protein binding / tau protein phosphorylation (as appropriate) * Microglial phagocytosis * Complement activation * Autophagy / lysosome organization * Ferroptosis

Cell Ontology (CL) suggestions: * Microglial cell * Astrocyte * Oligodendrocyte * Cortical inhibitory neuron

GO Cellular Component suggestions: * Lysosome * Endosome * Synapse * Mitochondrion

7. Anatomical structures affected

7.1 Organ/system level

Primary system: central nervous system. Biomarker and transcriptomic studies emphasize cortical involvement (e.g., parietal cortex; dorsolateral prefrontal cortex resources and repositories are referenced in omics work). (brase2023singlenucleusrnasequencingof pages 7-8, green2024cellularcommunitiesreveal pages 1-2)

7.2 UBERON suggestions

Brain (UBERON:0000955)
Cerebral cortex (UBERON:0000956)

8. Temporal development

The AA Workgroup explicitly incorporates asymptomatic biological AD and proposes an integrated biological–clinical staging approach across the continuum. (jack2024revisedcriteriafor pages 1-2)

Blood biomarkers are being positioned to enable earlier detection and staging—e.g., plasma biomarker performance for detecting brain Aβ pathology in non-demented cohorts. (niimi2024combiningplasmaaβ pages 7-8)

9. Inheritance and population

9.1 Epidemiology

This run did not retrieve a primary global epidemiology paper with prevalence/incidence rates for AD specifically; therefore, numeric prevalence/incidence statements are not provided here.

9.2 Genetic architecture / inheritance patterns

The retrieved omics study includes autosomal dominant AD carriers and late-onset risk-variant carriers, consistent with AD’s mixed architecture (rare Mendelian forms plus common polygenic susceptibility). (brase2023singlenucleusrnasequencingof pages 7-8)

10. Diagnostics

10.1 AA Workgroup revised diagnostic and staging criteria (2024)

Key definitions and performance anchors include: * AD can be defined biologically; ADNPC can exist without symptoms. (jack2024revisedcriteriafor pages 1-2) * Core 1 biomarkers (amyloid PET, approved CSF, accurate plasma biomarkers) are intended to map to plaques/tangles and support biological diagnosis. (jack2024revisedcriteriafor pages 1-2) * Core 1 plasma benchmark: the Workgroup proposes ≥90% accuracy vs amyloid PET for a standalone plasma biomarker used to establish amyloid pathology. (jack2024revisedcriteriafor pages 8-9) * Example reference performance: florbetapir PET visual reads showed 96% sensitivity / 100% specificity vs CERAD neuritic plaques; approved CSF assays showed approximately 88%/93% and 85%/94% sensitivity/specificity vs amyloid PET visual reads. (jack2024revisedcriteriafor pages 8-9)

10.2 Blood biomarkers (2023–2024 emphasis): plasma p-tau217

Real-world memory clinic performance: plasma p-tau217 (ECL immunoassay) detected CSF-defined Aβ pathology with AUC 0.91 and outperformed plasma p-tau181 (AUC 0.73). A two-threshold triage approach suggested that confirmatory lumbar puncture might be avoided in ~58–68% of cases depending on chosen sensitivity/specificity operating points. (dyer2024performanceofplasma pages 10-11)

Non-demented trial-ready / research cohorts: combining plasma Aβ measures and p-tau217 can yield AUCs approaching ~0.93–0.95 for predicting abnormal Aβ-PET, with best-performing models depending on clinical stage (CDR 0 vs 0.5) and inclusion of age/sex/APOE. (niimi2024combiningplasmaaβ pages 7-8)

10.3 Imaging and CSF biomarkers

The AA Workgroup uses amyloid PET and CSF assays as Core 1 standards and provides reference sensitivity/specificity anchors (see above). (jack2024revisedcriteriafor pages 8-9)

11. Outcome / prognosis

Evidence in the current retrieved set is insufficient to provide robust, citation-backed survival estimates, stage durations, or validated prognostic models.

12. Treatment

12.1 Current applications and real-world implementation (2023–2024)

Anti-amyloid monoclonal antibodies are described as the first generation of causal (pathology-targeting) therapies but with modest clinical effect sizes, safety/monitoring burdens, and eligibility constraints. The EADC investigators recommend selected patient use with treatment documentation in registries to inform real-world effectiveness and system readiness. (jessen2024progressinthe pages 1-2)

12.2 Safety considerations (ARIA)

A review summarizing aducanumab reports that while amyloid was reduced on PET, there was “no apparent improvement in cognitive function,” and ARIA occurred in ~40% of high-dose aducanumab-treated patients. (parums2024areviewof pages 4-6)

12.3 MAXO term suggestions

Anti-amyloid beta immunotherapy (monoclonal antibody therapy)
Magnetic resonance imaging monitoring
Lumbar puncture (confirmatory diagnostics)

Note: Detailed lecanemab/donanemab trial effect sizes and ARIA rates for those agents were not extractable from the currently captured evidence snippets.

13. Prevention

The prevention evidence emphasized in retrieved sources is consistent with a substantial preventable fraction of dementia via risk factor modification. A synthesis citing the Lancet Commission reports that a set of modifiable factors can account for up to ~40% of dementia cases. (parums2024areviewof pages 4-6)

14. Other species / natural disease

Not covered in the retrieved evidence.

15. Model organisms

Not covered in the retrieved evidence.

16. Notable 2023–2024 developments (concise analysis)

Shift toward biological diagnosis: 2024 AA criteria explicitly formalize “biological AD” as diagnosable via biomarkers even without symptoms, motivated by the need to select patients for disease-modifying therapies. (jack2024revisedcriteriafor pages 1-2)
Blood-based biomarker operationalization: Multiple 2024 studies support p-tau217 as a leading BBM with AUCs around 0.9 and plausible clinical triage workflows to reduce lumbar punctures and expand access. (dyer2024performanceofplasma pages 10-11, niimi2024combiningplasmaaβ pages 7-8)
Systems and cell-state view of AD mechanisms: snRNA-seq studies in 2023–2024 strengthen the view that variant-associated microglial/astrocytic/oligodendroglial/neuron states are central to disease heterogeneity, suggesting multi-target and combination-therapy rationales. (brase2023singlenucleusrnasequencingof pages 7-8)

17. Evidence gaps / limitations of this report

Despite extensive retrieval, several template-required elements were not available as citation-backed facts in the captured evidence snippets: * ICD/MeSH/OMIM/Orphanet identifiers (should be filled from authoritative registries). * Quantitative prevalence/incidence statistics and survival estimates from primary epidemiologic cohorts. * Comprehensive phenotype frequencies mapped to HPO with percentages. * Detailed variant nomenclature, allele frequencies, and ClinVar classifications. * Detailed lecanemab/donanemab trial efficacy/safety metrics (beyond class-level summaries and aducanumab ARIA estimate).

These can be completed by targeted retrieval of (i) epidemiology (GBD dementia/AD-specific tables), (ii) GeneReviews/OMIM/ClinVar summaries for causative variants, and (iii) primary phase 3 trial publications for lecanemab and donanemab.

References

(jack2024revisedcriteriafor pages 1-2): Clifford R. Jack, J. Scott Andrews, Thomas G. Beach, Teresa Buracchio, Billy Dunn, Ana Graf, Oskar Hansson, Carole Ho, William Jagust, Eric McDade, Jose Luis Molinuevo, Ozioma C. Okonkwo, Luca Pani, Michael S. Rafii, Philip Scheltens, Eric Siemers, Heather M. Snyder, Reisa Sperling, Charlotte E. Teunissen, and Maria C. Carrillo. Revised criteria for diagnosis and staging of alzheimer's disease: alzheimer's association workgroup. Alzheimer's & Dementia, 20:5143-5169, Jun 2024. URL: https://doi.org/10.1002/alz.13859, doi:10.1002/alz.13859. This article has 2357 citations and is from a highest quality peer-reviewed journal.
(dyer2024performanceofplasma pages 10-11): Adam H. Dyer, Helena Dolphin, Antoinette O’Connor, Laura Morrison, Gavin Sedgwick, Conor Young, Emily Killeen, Conal Gallagher, Aoife McFeely, Eimear Connolly, Naomi Davey, Paul Claffey, Paddy Doyle, Shane Lyons, Christine Gaffney, Ruth Ennis, Cathy McHale, Jasmine Joseph, Graham Knight, Emmet Kelly, Cliona O’Farrelly, Aoife Fallon, Sean O’Dowd, Nollaig M. Bourke, and Sean P. Kennelly. Performance of plasma p-tau217 for the detection of amyloid-β positivity in a memory clinic cohort using an electrochemiluminescence immunoassay. Alzheimer's Research & Therapy, Aug 2024. URL: https://doi.org/10.1186/s13195-024-01555-z, doi:10.1186/s13195-024-01555-z. This article has 33 citations and is from a domain leading peer-reviewed journal.
(niimi2024combiningplasmaaβ pages 7-8): Yoshiki Niimi, Shorena Janelidze, Kenichiro Sato, Naoki Tomita, Tadashi Tsukamoto, Takashi Kato, Kenji Yoshiyama, Hisatomo Kowa, Atsushi Iwata, Ryoko Ihara, Kazushi Suzuki, Kensaku Kasuga, Takeshi Ikeuchi, Kenji Ishii, Kengo Ito, Akinori Nakamura, Michio Senda, Theresa A. Day, Samantha C. Burnham, Leonardo Iaccarino, Michael J. Pontecorvo, Oskar Hansson, and Takeshi Iwatsubo. Combining plasma aβ and p-tau217 improves detection of brain amyloid in non-demented elderly. Alzheimer's Research & Therapy, May 2024. URL: https://doi.org/10.1186/s13195-024-01469-w, doi:10.1186/s13195-024-01469-w. This article has 68 citations and is from a domain leading peer-reviewed journal.
(jack2024revisedcriteriafor pages 8-9): Clifford R. Jack, J. Scott Andrews, Thomas G. Beach, Teresa Buracchio, Billy Dunn, Ana Graf, Oskar Hansson, Carole Ho, William Jagust, Eric McDade, Jose Luis Molinuevo, Ozioma C. Okonkwo, Luca Pani, Michael S. Rafii, Philip Scheltens, Eric Siemers, Heather M. Snyder, Reisa Sperling, Charlotte E. Teunissen, and Maria C. Carrillo. Revised criteria for diagnosis and staging of alzheimer's disease: alzheimer's association workgroup. Alzheimer's & Dementia, 20:5143-5169, Jun 2024. URL: https://doi.org/10.1002/alz.13859, doi:10.1002/alz.13859. This article has 2357 citations and is from a highest quality peer-reviewed journal.
(brase2023singlenucleusrnasequencingof pages 7-8): Logan Brase, Shih-Feng You, Ricardo D’Oliveira Albanus, Jorge L. Del-Aguila, Yaoyi Dai, Brenna C. Novotny, Carolina Soriano-Tarraga, Taitea Dykstra, Maria Victoria Fernandez, John P. Budde, Kristy Bergmann, John C. Morris, Randall J. Bateman, Richard J. Perrin, Eric McDade, Chengjie Xiong, Alison M. Goate, Martin Farlow, Greg T. Sutherland, Jonathan Kipnis, Celeste M. Karch, Bruno A. Benitez, and Oscar Harari. Single-nucleus rna-sequencing of autosomal dominant alzheimer disease and risk variant carriers. Nature Communications, Apr 2023. URL: https://doi.org/10.1038/s41467-023-37437-5, doi:10.1038/s41467-023-37437-5. This article has 95 citations and is from a highest quality peer-reviewed journal.
(parums2024areviewof pages 4-6): Dinah V. Parums. A review of the current status of disease-modifying therapies and prevention of alzheimer’s disease. Medical Science Monitor, 30:e945091-1-e945091-7, May 2024. URL: https://doi.org/10.12659/msm.945091, doi:10.12659/msm.945091. This article has 39 citations and is from a peer-reviewed journal.
(jessen2024progressinthe pages 1-2): Frank Jessen, M.G. Kramberger, D. Angioni, D. Aarsland, M. Balasa, K. Bennys, M. Boada, M. Boban, A. Chincarini, L. Exalto, A. Felbecker, K. Fliessbach, G.B. Frisoni, A.J. Garza-Martínez, T. Grimmer, B. Hanseeuw, J. Hort, A. Ivanoiu, S. Klöppel, L. Krajcovicova, B. McGuinness, P. Mecocci, A. de Mendonca, A. Nous, P.-J. Ousset, C. Paquet, R. Perneczky, O. Peters, M. Tabuas-Pereira, F. Piazza, D. Plantone, M. Riverol, A. Ruiz, G. Sacco, I. Santana, N. Scarmeas, E. Solje, E. Stefanova, S. Sutovsky, W. van der Flier, T. Welsh, A. Wimo, B. Winblad, L. Frölich, and S. Engelborghs. Progress in the treatment of alzheimer's disease is needed – position statement of european alzheimer's disease consortium (eadc) investigators. The Journal of Prevention of Alzheimer's Disease, 11:1212-1218, Oct 2024. URL: https://doi.org/10.14283/jpad.2024.153, doi:10.14283/jpad.2024.153. This article has 23 citations and is from a peer-reviewed journal.
(OpenTargets Search: Alzheimer disease): Open Targets Query (Alzheimer disease, 39 results). Buniello, A. et al. (2025). Open Targets Platform: facilitating therapeutic hypotheses building in drug discovery. Nucleic Acids Research.
(green2024cellularcommunitiesreveal pages 1-2): Gilad Sahar Green, Masashi Fujita, Hyun-Sik Yang, Mariko Taga, Anael Cain, Cristin McCabe, Natacha Comandante-Lou, Charles C. White, Anna K. Schmidtner, Lu Zeng, Alina Sigalov, Yangling Wang, Aviv Regev, Hans-Ulrich Klein, Vilas Menon, David A. Bennett, Naomi Habib, and Philip L. De Jager. Cellular communities reveal trajectories of brain ageing and alzheimer's disease. Nature, 633:634-645, Aug 2024. URL: https://doi.org/10.1038/s41586-024-07871-6, doi:10.1038/s41586-024-07871-6. This article has 177 citations and is from a highest quality peer-reviewed journal.

Ask OpenScientist

OpenScientist Report

Subtypes

Comorbidities

Mechanistic Hypotheses

Discussions and Knowledge Gaps

Pathophysiology

Pathograph

Phenotypes

Genetic Associations

Medical Actions

Environmental Factors

Biochemical Markers

Diagnosis

Prevalence

Related Datasets

Clinical Trials

Experimental Models

Computational Models

Source YAML

References & Deep Research

References

Deep Research

Disease Characteristics Research Template

Target Disease

Research Objectives

1. Disease Information

2. Etiology

3. Phenotypes

4. Genetic/Molecular Information

5. Environmental Information

6. Mechanism / Pathophysiology

7. Anatomical Structures Affected

8. Temporal Development

9. Inheritance and Population

10. Diagnostics

11. Outcome/Prognosis

12. Treatment

13. Prevention

14. Other Species / Natural Disease

15. Model Organisms

Citation Requirements

Output Format

Alzheimer Disease (AD) — Comprehensive Disease Characteristics Research Report (2023–2024 emphasis)

Executive summary

High-value recent sources (quick reference)

1. Disease information

1.1 Concise overview

1.2 Key identifiers and ontology codes

1.3 Common synonyms / alternative names

1.4 Evidence source type

2. Etiology

2.1 Disease causal factors (current understanding)

2.2 Genetic risk factors (selected; recent mechanistic emphasis)

2.3 Protective factors

2.4 Gene–environment interactions

3. Phenotypes

3.1 Core clinical phenotypes (high-level)

3.2 HPO term suggestions (non-exhaustive; for knowledge base structuring)

4. Genetic / molecular information

4.1 Molecular hallmarks

4.2 Pathogenic variants and variant classes

4.3 Epigenetic information and chromosomal abnormalities

5. Environmental information

5.1 Lifestyle and environmental risk factors (modifiable)

5.2 Infectious agents

6. Mechanism / pathophysiology

6.1 Updated causal chain (biomarker-informed)

6.2 Microglia/astrocyte/neuron/oligodendrocyte programs (single-nucleus evidence)

6.3 Pathways / ontology suggestions

7. Anatomical structures affected

7.1 Organ/system level

7.2 UBERON suggestions

8. Temporal development

9. Inheritance and population

9.1 Epidemiology

9.2 Genetic architecture / inheritance patterns

10. Diagnostics

10.1 AA Workgroup revised diagnostic and staging criteria (2024)

10.2 Blood biomarkers (2023–2024 emphasis): plasma p-tau217