What determines the temporal sequence of cholinergic dysfunction versus amyloid/tau pathology in human AD?

Basal forebrain NGF/TrkA failure is an upstream trigger that makes cholinergic neurons permissive to later amyloid and tau spread

Mechanism: Early loss of retrograde NGF signaling from cortex/hippocampus to nucleus basalis cholinergic neurons reduces `NTRK1 (TrkA)` survival signaling, impairs axonal transport, and lowers cortical acetylcholine release. This produces synaptic inactivity, endosomal stress, and impaired APP trafficking, which then biases vulnerable projection fields toward increased amyloidogenic processing and later tau propagation. Target gene/protein/pathway: `NGF`, `NTRK1`, `BFCN` trophic support, axonal transport, APP endosomal processing. Supporting evidence: Human AD shows early basal forebrain cholinergic vulnerability and disrupted NGF-TrkA signaling; cholinergic denervation correlates with cortical dysfunction before severe neuron loss. Falsifiable experiment: In iPSC-derived human basal forebrain cholinergic neurons co-cultured with cortical neurons, selectively reduce `NTRK1` or retrograde NGF transport and test whether this increases APP beta-processing, neuronal hyperexcitability/inactivity transitions, and tau seeding susceptibility in cortical partner neurons. Rescue with TrkA agonism should reverse the effect. Confidence: 0.72

Amyloid first impairs septo-hippocampal cholinergic terminals through alpha7 nicotinic receptor-dependent synaptotoxicity, with cholinergic failure as an early downstream consequence rather than a cause

Mechanism: Soluble Aβ oligomers bind `CHRNA7`-containing nicotinic receptors on cholinergic and glutamatergic terminals, producing calcium dysregulation, presynaptic failure, and local inflammatory signaling. In this model, cholinergic dysfunction is an early readout of soluble amyloid toxicity, preceding major plaque burden but not causally upstream of amyloid generation. Target gene/protein/pathway: `APP`, Aβ oligomers, `CHRNA7`, calcium signaling, presynaptic vesicle cycling. Supporting evidence: There is longstanding evidence for high-affinity interaction between Aβ species and alpha7 nicotinic receptors, plus strong sensitivity of cholinergic terminals to soluble Aβ. Falsifiable experiment: Apply patient-derived soluble Aβ fractions to human cholinergic neuron-hippocampal organoid systems with and without `CHRNA7` knockout or alpha7 antagonism. If this hypothesis is correct, alpha7 blockade should preserve cholinergic synaptic release without reducing total Aβ exposure. Confidence: 0.68

Tau pathology reaches cholinergic projection neurons early because APOE4-microglial complement signaling selectively destabilizes cholinergic synapses before overt amyloid burden

Mechanism: `APOE4` shifts microglia toward a complement-active state (`C1q`, `C3`, `CR3`) that preferentially tags low-activity, long-range cholinergic synapses for elimination. Synapse stripping reduces cortical acetylcholine tone, increases network desynchronization, and creates conditions favoring tau spread along vulnerable corticobasal forebrain circuits. Target gene/protein/pathway: `APOE`, complement cascade (`C1QA/B/C`, `C3`, `ITGAM/CR3`), microglia-cholinergic synapse interactions. Supporting evidence: APOE4 is strongly linked to earlier AD pathobiology; complement-mediated synapse loss is well supported in AD; cholinergic projections are anatomically diffuse and potentially vulnerable to activity-dependent pruning. Falsifiable experiment: In APOE3 vs APOE4 human tri-cultures containing microglia, basal forebrain cholinergic neurons, and cortical neurons, quantify complement deposition on cholinergic boutons before amyloid plaque-like aggregation. `C1q` or `C3` blockade should preserve cholinergic terminals and reduce subsequent tau uptake/spread. Confidence: 0.64

Locus coeruleus degeneration temporally gates whether cholinergic dysfunction or amyloid/tau appears first

Mechanism: Early noradrenergic loss from locus coeruleus reduces anti-inflammatory tone and impairs astrocytic/microglial Aβ clearance. In some patients this causes amyloid-first disease; in others, noradrenergic failure also deprives basal forebrain cholinergic neurons of modulatory support, causing an apparently cholinergic-first phenotype. The ordering is therefore determined by the severity and regional distribution of early catecholaminergic denervation. Target gene/protein/pathway: noradrenergic signaling, `DBH`, beta-adrenergic signaling, microglial clearance pathways, cholinergic-noradrenergic coupling. Supporting evidence: Human AD often shows early locus coeruleus tau pathology and noradrenergic dysfunction; noradrenaline regulates neuroinflammation and clearance programs relevant to amyloid. Falsifiable experiment: Stratify prodromal human subjects by PET/MRI proxies of locus coeruleus integrity, cholinergic basal forebrain volume, and amyloid/tau biomarkers longitudinally. The hypothesis predicts that low LC integrity will explain distinct ordering patterns between cholinergic decline and amyloid/tau positivity. Confidence: 0.61

Endosomal trafficking defects in basal forebrain cholinergic neurons are the common upstream lesion linking APP processing and cholinergic degeneration

Mechanism: Cholinergic neurons have extreme dependence on long-distance endosomal transport for neurotrophin signaling. Genetic or age-related impairment in `SORL1`, `BIN1`, `PICALM`, or retromer function causes both abnormal APP sorting toward beta-secretase compartments and failure of trophic signaling in cholinergic axons. This predicts that amyloidogenic processing and cholinergic dysfunction are parallel outputs of one earlier trafficking defect, not a simple linear sequence. Target gene/protein/pathway: `SORL1`, `BIN1`, `PICALM`, retromer (`VPS35`), endosome-lysosome trafficking, APP sorting. Supporting evidence: Multiple AD risk genes converge on endosomal biology; cholinergic neurons are especially vulnerable to transport defects because of their large arborization and NGF dependence. Falsifiable experiment: Introduce AD-risk `SORL1` or `BIN1` variants into human basal forebrain cholinergic neurons and matched cortical neurons, then compare APP processing, retrograde NGF signaling, and tau uptake. A cholinergic-selective transport phenotype would support this model. Confidence: 0.77

Astrocytic cholinesterase and reactive astrocyte programs create a self-reinforcing low-acetylcholine niche that accelerates tau more than amyloid

Mechanism: Reactive astrocytes in vulnerable cortex increase acetylcholine degradation and reduce cholinergic synapse support, pushing local circuits into impaired gamma/theta coupling and metabolic stress. This state may be particularly permissive for tau phosphorylation and trans-synaptic spread, making cholinergic dysfunction appear earlier in tau-predominant AD trajectories. Target gene/protein/pathway: astrocyte reactivity, `ACHE`, `BCHE`, cholinergic synapse maintenance, tau kinases (`GSK3B`, `CDK5`). Supporting evidence: Butyrylcholinesterase and astrocyte changes increase in AD tissue; cholinergic tone strongly shapes cortical oscillations and glial state. Falsifiable experiment: In human cortical organoids with reactive astrocytes, overexpress `BCHE` or induce astrocyte inflammatory programs and measure acetylcholine tone, tau phosphorylation, and tau seed propagation. Selective `BCHE` inhibition should blunt tau-related changes if this mechanism is correct. Confidence: 0.58

The temporal order is subtype-specific: APOE4/amyloid-endosomal AD is amyloid-first, while trophic-transport/cholinergic AD is cholinergic-first

Mechanism: There may not be one universal sequence. One subtype is driven by `APOE4`-biased amyloid clearance failure and cortical Aβ accumulation that secondarily injures cholinergic projections. Another is driven by early basal forebrain trophic transport failure (`NGF/TrkA`, retromer, mitochondrial stress), with cholinergic dysfunction preceding measurable amyloid or tau. Target gene/protein/pathway: `APOE`, `SORL1`, `NTRK1`, mitochondrial stress, complement, Aβ clearance. Supporting evidence: Human AD is biologically heterogeneous, and mixed imaging/biomarker studies already suggest non-identical prodromal trajectories. Falsifiable experiment: Perform multimodal longitudinal clustering in preclinical and MCI patients using basal forebrain MRI, cholinergic PET, plasma/CSF Aβ42/40, p-tau species, APOE genotype, and inflammatory markers. The hypothesis predicts at least two reproducible trajectory classes with different therapeutic response profiles. Confidence: 0.74

Therapeutic implication across these hypotheses: the most informative interventions are not generic cholinesterase inhibitors, but subtype-matched trials of `TrkA/NGF` restoration, complement blockade, alpha7-nAChR modulation, retromer enhancement, or early noradrenergic support, with longitudinal biomarker readouts to determine ordering rather than assuming it.

NGF/TrkA failure is upstream

Weak evidence: Most human support is correlational and late-stage. Reduced `NTRK1`/NGF signaling could be a consequence of early tau, endosomal stress, or synapse loss rather than the initiating lesion. “Before severe neuron loss” does not establish before soluble Aβ or seed-competent tau.
Alternative mechanisms: Early tau in entorhinal-limbic circuits, APP/endosomal defects, mitochondrial failure, or vascular hypoperfusion could independently cause both cholinergic dysfunction and apparent NGF signaling failure.
Translational risks: iPSC cholinergic neurons do not reproduce decades of retrograde trophic dependence, long axons, aging, or human basal forebrain circuit architecture. TrkA agonism may improve survival markers without proving disease-ordering relevance.
Falsifying experiment: In longitudinal human cohorts, show that sensitive amyloid/tau biomarkers become abnormal before any cholinergic PET/MRI deficit in people with intact NGF/TrkA proxies. In vitro, if selective `NTRK1` reduction does not increase APP beta-processing or tau seeding susceptibility beyond nonspecific stress, the mechanism weakens substantially.

Amyloid first via alpha7 nicotinic receptor synaptotoxicity

Weak evidence: The Aβ-`CHRNA7` interaction literature is mixed and heavily model-dependent; binding affinity and functional relevance vary by Aβ preparation. Presynaptic toxicity is plausible, but specificity to cholinergic terminals is not well established in humans.
Alternative mechanisms: Aβ may impair synapses through NMDA receptor dysregulation, membrane pore effects, oxidative stress, microglial signaling, or generic calcium overload rather than a primary `CHRNA7` route.
Translational risks: Patient-derived soluble Aβ fractions are heterogeneous and unstable. `CHRNA7` knockout may alter baseline synapse maturation and excitability, confounding interpretation. Alpha7 antagonism could protect acutely while worsening cognition in vivo.
Falsifying experiment: Use well-characterized human organoid/co-culture systems and demonstrate that alpha7 blockade or `CHRNA7` deletion fails to rescue cholinergic release despite clear Aβ-induced toxicity. Strong falsification would be equal toxicity in `CHRNA7`-null and wild-type conditions.

APOE4-microglial complement selectively hits cholinergic synapses first

Weak evidence: Complement-mediated pruning is supported broadly, but “selective” vulnerability of cholinergic synapses is mostly inferred from anatomy, not demonstrated. APOE4 effects may amplify many injury pathways, not uniquely cholinergic pruning.
Alternative mechanisms: APOE4 could act primarily through impaired lipid handling, amyloid clearance, BBB dysfunction, or astrocyte-mediated inflammation, with complement as a downstream amplifier rather than the ordering determinant.
Translational risks: Human tri-cultures poorly model region-specific microglial states, complement gradients, and long-range cholinergic projections. Complement blockade may preserve synapses in vitro but fail clinically because it misses upstream tau/amyloid drivers.
Falsifying experiment: In APOE4 human systems and longitudinal human tissue/imaging, test whether complement deposition is not enriched on cholinergic boutons before overt amyloid/tau changes. If `C1q`/`C3` inhibition preserves terminals but does not alter downstream tau spread, the causal claim fails.

Locus coeruleus degeneration gates sequence

Weak evidence: This is plausible but underspecified. LC tau pathology is early, yet proving it determines whether cholinergic or amyloid/tau changes appear first is much stronger than showing association. LC integrity measures are noisy proxies.
Alternative mechanisms: LC degeneration may simply be another parallel early lesion caused by tau vulnerability, aging, sleep disruption, or vascular disease. The observed ordering could instead reflect subtype biology or measurement sensitivity.
Translational risks: Human stratification studies are vulnerable to confounding by arousal state, antidepressants, vascular burden, and imaging resolution limits. LC-targeted support may have broad neuromodulatory effects without changing core AD progression.
Falsifying experiment: In a longitudinal prodromal cohort, test whether LC metrics fail to predict subsequent ordering once age, APOE, baseline amyloid/tau burden, vascular disease, and sleep variables are controlled. If predictive value disappears, the gating hypothesis is weak.

Endosomal trafficking defects are the common upstream lesion

Weak evidence: This is the strongest mechanistically because AD genetics converges on trafficking, but evidence for basal-forebrain selectivity remains incomplete. Parallel outputs from one lesion are plausible, yet direct proof that cholinergic dysfunction and amyloid arise from the same earliest defect in humans is lacking.
Alternative mechanisms: Trafficking defects may be one branch of a broader aging network that includes lysosomal failure, mitochondrial stress, and proteostasis collapse. Amyloid and cholinergic dysfunction may still be partially sequential, not merely parallel outputs.
Translational risks: Editing `SORL1`/`BIN1` in culture can generate artificial large-effect phenotypes unlike human heterozygous aging. Cell-autonomous models miss glia, circuit activity, and decades-long compensation.
Falsifying experiment: In isogenic human neurons, if AD-risk trafficking variants change APP processing similarly in cortical and cholinergic neurons without a cholinergic-selective trophic transport deficit, the proposed selectivity is undermined. Human falsification would be normal early trafficking markers in individuals who still develop cholinergic-first trajectories.

Reactive astrocytes and cholinesterases create a low-ACh niche that accelerates tau

Weak evidence: This is the weakest of the set. Increased `BCHE`/astrocyte reactivity in AD tissue is real, but causality and timing are unclear, and the link to tau-first progression is indirect. Astrocyte cholinesterase changes may be compensatory or late-stage.
Alternative mechanisms: Reactive astrocytes may primarily reflect response to existing amyloid/tau, cytokines, or neuronal injury. Tau acceleration could come from inflammatory kinases, lipid dysregulation, or impaired glutamate handling independent of cholinergic tone.
Translational risks: Organoids poorly model extracellular acetylcholine dynamics and mature astrocyte states. Cholinesterase inhibition may alter network activity broadly, creating false-positive “rescue” effects unrelated to disease mechanism.
Falsifying experiment: Induce reactive astrocytes or `BCHE` overexpression in human cortical systems and show no increase in tau phosphorylation/propagation after controlling for general inflammation and excitability. If selective `BCHE` inhibition does not rescue tau phenotypes, the hypothesis largely collapses.

Temporal order is subtype-specific

Weak evidence: Heterogeneity is highly plausible, but this formulation risks becoming unfalsifiable because any ordering can be assigned to a subtype after the fact. Existing human trajectory studies are sensitive to biomarker choice and clustering method.
Alternative mechanisms: Apparent subtypes may reflect staging, measurement thresholds, ascertainment bias, co-pathologies, or continuous variation rather than discrete disease classes.
Translational risks: Longitudinal clustering often yields unstable classes across cohorts. Over-stratification can generate nonreproducible therapeutic niches and underpowered trials.
Falsifying experiment: Pre-register clustering features and replication criteria across independent cohorts. If the same two or more trajectory classes do not reproduce with stable assignments and distinct outcomes, the subtype-ordering model is not robust.

Cross-cutting skeptical points

• The main weakness across all seven is temporal inference from cross-sectional or reductionist systems.
• Measurement asymmetry matters: amyloid/tau biomarkers are currently more sensitive and standardized than cholinergic ones, so “amyloid-first” may partly reflect assay availability.
• Co-pathologies, vascular injury, sleep dysfunction, medications, and age-related neuromodulatory decline are major confounds that can mimic cholinergic-first trajectories.
• A strong discriminating program would require longitudinal multimodal human data with repeated amyloid, tau, cholinergic, LC, inflammatory, and vascular measures, plus prespecified causal models rather than post hoc narrative fitting.

Bottom Line

The ideas worth carrying forward are `#5 endosomal-trafficking-first`, `#7 subtype-specific ordering`, `#1 NGF/TrkA trophic failure`, and `#3 APOE4-complement pruning`. `#4 locus coeruleus gating` is useful mainly as a stratification axis, not as a primary drug program. I would drop `#2 alpha7-nAChR amyloid synaptotoxicity` and `#6 astrocytic cholinesterase niche` as lead translational bets.

Priority Order

`#5 Endosomal trafficking defects are the common upstream lesion`

Druggability is moderate now and potentially high later: `SORL1/retromer` is genetically anchored, and retromer-enhancing small molecules have rescued endosomal/amyloid/tau phenotypes in human `SORL1` neuronal models, but this is still preclinical rather than trial-ready. Biomarkers are better than they were a year ago: genotype enrichment (`SORL1/BIN1/PICALM`), standard amyloid/tau markers, cholinergic readouts (`[18F]FEOBV` VAChT PET, NBM MRI), plus emerging CSF `sSorLA` as a trafficking biomarker. Best models are isogenic human iPSC basal-forebrain cholinergic neuron plus cortical/microglia assembloids with microfluidic retrograde transport; rodent models are acceptable for PK/PD only. Main safety risk is that trafficking modulators are pleiotropic and can hit lysosomal, cardiac, or immune biology off-target; gene therapy adds neurosurgical/AAV risk. Realistic timeline/cost if starting now: `18-24 months / $8-15M` for decisive preclinical package, `3-4 years / $25-50M` to IND, `6-9 years / $120-250M` to a meaningful phase 2 signal.

`#7 Temporal order is subtype-specific`

This is the best working framework, but it is a stratification hypothesis, not a single drug target. Druggability comes from using it to place patients into `amyloid-clearance`, `microglia/complement`, or `trophic-transport` arms rather than treating “AD” as one biology. Biomarker readiness is relatively strong: plasma `Aβ42/40`, `p-tau217/181/231`, amyloid/tau PET, APOE genotype, NBM MRI, LC MRI, and FEOBV PET at selected centers; the weak link remains scalable cholinergic biomarkers. Model systems matter less than longitudinal human cohorts here. Safety risk is low for the hypothesis itself but high for over-stratification: unstable clusters will kill trial power. Realistic program: `2-4 years / $15-40M` for multi-cohort replication with prespecified classes; `5-8 years / $80-200M` if converted into a biomarker-stratified platform trial.

`#1 Basal forebrain NGF/TrkA failure is upstream`

Biology is plausible and clinically relevant, but druggability is hard because `NGF/TrkA` is a delivery problem, not a simple pill target. Prior NGF gene-therapy work showed degenerating human neurons can still mount trophic responses, but also exposed delivery limitations; the ongoing `AAV2-BDNF` phase 1 trial (`NCT05040217`, started February 7, 2022; estimated primary completion December 1, 2027) shows the field is still in early invasive gene-therapy mode rather than scalable AD therapeutics. Biomarkers: NBM MRI, FEOBV PET, amyloid/tau markers, possibly phospho-neurofilament/synaptic injury panels; there is no validated circulating TrkA/retrograde transport biomarker. Best models are aged human BFCN-cortical microfluidic systems with explicit retrograde transport assays. Safety risk is substantial: pain/autonomic effects for NGF-class biology, plus neurosurgical and durability risks for AAV delivery. Realistic timeline/cost: `18-30 months / $6-12M` to show causal transport-to-amyloid/tau linkage, `5-7 years / $80-180M` to an early clinical readout.

`#3 APOE4-microglial complement signaling selectively destabilizes cholinergic synapses`

This is druggable in principle but only if you relax the claim from “selectively cholinergic” to “APOE4-biased synaptic vulnerability with a cholinergic-enriched phenotype.” Complement and microglia are targetable, but the field has already learned that microglial modulation can show target engagement without efficacy: `AL002` reached phase 2 and, as published March 5, 2026, did not meet its primary endpoint while ARIA-like MRI abnormalities were common. Biomarkers are fairly good: `APOE` genotype, CSF/plasma complement proteins, `sTREM2`, osteopontin/SPP1, amyloid/tau markers, plus FEOBV PET or NBM MRI to test cholinergic selectivity. Best models are APOE3/4 isogenic human tri-cultures/assembloids with microglia and complement-competent media; standard organoids are not enough. Safety risk is meaningful: infection/immunomodulation tradeoffs for complement blockade, plus ARIA-like vascular/inflammatory liabilities for CNS immunotherapy. Realistic timeline/cost: `12-24 months / $5-10M` for human-system de-risking, `4-6 years / $60-140M` if piggybacking on an existing brain-penetrant immunology asset.

`#4 Locus coeruleus degeneration gates sequence`

This survives as a modifier, not a lead causal program. Druggability is modest: noradrenergic support may help network resilience or inflammation, but it is unlikely to settle the core sequence question alone. Biomarkers are the main value: neuromelanin-sensitive LC MRI is usable now, and combined LC plus NBM imaging has already shown both systems are abnormal early in AD; pupillometry and sleep/autonomic phenotyping can be added cheaply. Best use is as a prespecified covariate in longitudinal human studies and adaptive trials. Safety depends on mechanism; pro-noradrenergic drugs bring cardiovascular, sleep, and anxiety liabilities in older adults. Realistic timeline/cost: `2-3 years / $10-25M` for a strong observational biomarker study, `4-6 years / $40-100M` for an interventional proof-of-concept.

Not Worth Leading With

`#2 alpha7-nAChR amyloid synaptotoxicity` is too weak as a primary ordering hypothesis. The biology is mixed, specificity to human cholinergic terminals is unproven, and the target class has a poor clinical history in cognition programs. It is more a mechanistic sub-branch than a platform.

`#6 reactive astrocyte cholinesterase low-ACh niche` is the least mature. It may matter as a secondary amplifier of tau/network dysfunction, but it is not trial-ready as a disease-ordering thesis.

What I Would Actually Fund

Fund a single integrated program around `#5 + #7`, with `#1`, `#3`, and `#4` as subtype modifiers. The decisive package is a longitudinal human cohort plus a matched human-cell perturbation stack: FEOBV PET or equivalent cholinergic imaging at selected sites, NBM/LC MRI everywhere, plasma/CSF amyloid and p-tau, APOE and trafficking-genetics enrichment, and pre-registered causal models. That is the fastest way to turn this from narrative debate into a trial-enabling taxonomy.

Sources

• FEOBV cholinergic PET in AD/MCI: https://pubmed.ncbi.nlm.nih.gov/28894304/ , https://pmc.ncbi.nlm.nih.gov/articles/PMC8958543/
• Basal forebrain cholinergic review: https://pubmed.ncbi.nlm.nih.gov/37086935/
• LC/NBM imaging in early AD: https://alzres.biomedcentral.com/articles/10.1186/s13195-024-01466-z
• Retromer/SORL1 therapeutic and biomarker evidence: https://pubmed.ncbi.nlm.nih.gov/37949073/ , https://pubmed.ncbi.nlm.nih.gov/40336092/ , https://pmc.ncbi.nlm.nih.gov/articles/PMC11713639/
• NGF/BDNF clinical translation: https://journals.sagepub.com/doi/10.3233/JAD-240545 , https://pubmed.ncbi.nlm.nih.gov/32126838/ , https://clinicaltrials.gov/study/NCT05040217
• Microglial/TREM2 clinical benchmark: https://alzres.biomedcentral.com/articles/10.1186/s13195-024-01599-1 , https://www.nature.com/articles/s41591-026-04273-1 , https://clinicaltrials.gov/study/NCT04592874

If you want, I can turn this into a one-page go/no-go matrix with scores for `human evidence`, `druggability`, `biomarker readiness`, `model validity`, and `clinical feasibility`.