Plaque Origin What is Established

An epistemic guardrail for the corpus. The Inside Out Plaque reading — the plaque as the gravestone of a neuron that failed from within — is the corpus's organizing image, and it is directionally well-supported. But most of the theses state it as settled fact, and it is not settled as a quantity. This note fixes the line between what the primary literature establishes and what remains hypothesis, so the confident chapters can defer to one graded statement rather than each overreaching on its own.

The common question — "how much of a plaque forms inside the neuron versus outside?" — has no validated numerical answer in human brain. The honest reason is that "formed" bundles three separable questions, which sit at three different levels of certainty.

The three questions, graded

1. Where is the Aβ peptide generated? — Intracellular. (Established.) Aβ is cleaved from APP by β- and γ-secretase within the endosomal/secretory system. The molecule is made mostly inside the neuron and then partly secreted. This is not controversial and is not the point in dispute.

2. Does intraneuronal Aβ42 accumulate before plaques? — Yes. (Established, human tissue.) Gouras et al. (2000) found γ-cleaved Aβ42 accumulating in AD-vulnerable human neurons in a pattern that "appears to precede both NFT and Aβ plaque deposition." Takahashi et al. (2002) localized it to multivesicular bodies in synaptic compartments, associated with abnormal synaptic morphology before plaque pathology. Note the original authors' language is deliberately cautious ("suggest," "appears to"), and immunodetection of intraneuronal Aβ has a live antibody-specificity caveat (see Plaque IgG Sequestration).

3. What is the deposit made of, and does it seed inside-out or outside-in? — Both mechanisms are demonstrated; the fraction is unresolved. This is the actual contested question, and the fair reading is that it is not either/or:

Inside-out is demonstrated — but the quantification is model-specific. Lee, Nixon et al. (2022) showed that de-acidified autolysosomes build up Aβ inside neurons, producing the PANTHOS "poisonous-flower" morphology, and that "individual neurons exhibiting PANTHOS are the principal source of senile plaques in amyloid precursor protein AD models." The load-bearing word is models: the quantitative "principal source" claim is from APP transgenic mice. PANTHOS is observed in human AD brain but its fractional contribution to human plaque burden has not been quantified.
Outside-in (extracellular seeding) is independently demonstrated. Meyer-Luehmann et al. (2006) showed that dilute Aβ-containing brain extract injected into the extracellular space nucleates cerebral β-amyloidosis in a time- and dose-dependent way, abolished by Aβ immunodepletion or denaturation — i.e. extracellular Aβ seeds can start deposits, prion-strain-like. Jucker & Walker (2013) generalized this to template-directed self-propagation. This route is real and drives spread and diffuse deposition; the corpus should not treat it merely as "the view to beat."

The defensible statement

Aβ is generated intracellularly; intraneuronal Aβ42 accumulates before plaques; a large share of dense-core plaques plausibly forms inside-out from failing neurons (established in mouse models, observed but unquantified in humans); and extracellular seeding independently nucleates and propagates deposits. The fraction of human plaque burden attributable to each route is not established. Any hard percentage overreaches the data.

Why the fraction is unresolved

The gap is not neglect — it is roughly 80% intrinsic measurement hardness, 20% a field that built its instruments for a different question. A conclusive human count is blocked by a stack of barriers, most fundamental first:

The endpoint erases its own origin. A finished plaque is the same static object whether it formed inside-out (a neuron ruptured and spilled its Aβ) or outside-in (secreted Aβ nucleated in the neuropil). Counting origins is therefore forensic reconstruction, and the inside-out route specifically destroys the neuron whose contents you would need to attribute.
Humans give only a single frozen frame. Plaque formation runs 10–20 years; autopsy is one snapshot. All human "temporal ordering" is stitched across different brains at different stages, never one cortical location over time.
The Aβ carries no provenance barcode. Secreted Aβ and lysis-released Aβ are the identical peptide; nothing tags a molecule as exocytosed versus spilled from a ruptured autolysosome. The Aβ in a plaque cannot be sorted by origin.
The one system you can watch is biased and non-human. In vivo two-photon microscopy catches plaque birth in mice — plaques appear in ~24 h, with microglia arriving within 1–2 days and dystrophic neurites forming over weeks (Meyer-Luehmann et al. 2008). But it resolves when, not from what, and the models are APP-overexpressors that force intracellular Aβ loads, arguably biasing toward the inside-out route in an animal lacking full human tauopathy.
"The fraction" may be ill-posed. The routes interconvert: a plaque can nucleate inside-out and then grow outside-in by recruiting secreted Aβ (Meyer-Luehmann et al. 2006). Dense-core and diffuse plaques may differ. If a single plaque is both, no clean percentage exists to measure.
Even the precursor state is contested. Immunodetection of intraneuronal Aβ leans on antibodies (6E10 and kin) that cross-react with APP and its C-terminal fragments; if the "before" cannot be measured cleanly, the endpoint cannot be attributed cleanly (see Plaque IgG Sequestration).

What would actually settle it: fate-mapping in knock-in (non-overexpressing) mice — a neuronal-content label that survives into the deposit, imaged longitudinally to see whether nucleation occurs at a dying tagged neuron or de novo in the neuropil; molecular-provenance signatures (does the plaque core carry autolysosomal cargo — cathepsins, LC3, lysosomal membrane proteins?); and spatial-omics on staged human brains to tally "plaques-in-the-making." Knock-in models, fate-mapping, and spatial omics are only now the right tools, so the mouse fraction is plausibly answerable within years; the human fraction may remain permanently an inference. That asymmetry is the whole reason the corpus should state inside-out as the better-supported initiating picture, not a proven quantitative monopoly.

How the corpus should use this

The one document already holding this line is the Bioenergetic Collapse (Revised) thesis, which states that "the fractional contribution of PANTHOS versus extracellular seeding to total plaque burden in humans" is unresolved and that "both mechanisms are likely contributors." Every chapter that asserts "the plaque is a gravestone" as flat fact — the Inside Out Plaque concept, and the Clearance, Synaptic, Bioenergetic (original), Temporal- and Cellular-Architecture theses — should be read as carrying this hedge, not as having proven the sole mechanism. Inside-out is the better-supported initiating picture; it is not an established quantitative monopoly on plaque formation.

Primary sources (all PubMed-verified)

Gouras GK, Tsai J, Näslund J, et al. (2000). Intraneuronal Aβ42 accumulation in human brain. Am J Pathol 156(1):15–20. PMID 10623648 · doi:10.1016/S0002-9440(10)64700-1
Takahashi RH, Milner TA, Li F, et al. (2002). Intraneuronal Alzheimer Aβ42 accumulates in multivesicular bodies and is associated with synaptic pathology. Am J Pathol 161(5):1869–1879. PMID 12414533 · doi:10.1016/S0002-9440(10)64463-X
Lee J-H, Yang D-S, Goulbourne CN, … Nixon RA. (2022). Faulty autolysosome acidification in Alzheimer's disease mouse models induces autophagic build-up of Aβ in neurons, yielding senile plaques. Nat Neurosci 25(6):688–701. PMID 35654956 · doi:10.1038/s41593-022-01084-8
Meyer-Luehmann M, Coomaraswamy J, Bolmont T, et al. (2006). Exogenous induction of cerebral β-amyloidogenesis is governed by agent and host. Science 313(5794):1781–1784. PMID 16990547 · doi:10.1126/science.1131864
Meyer-Luehmann M, Spires-Jones TL, Prada C, et al. (2008). Rapid appearance and local toxicity of amyloid-β plaques in a mouse model of Alzheimer's disease. Nature 451(7179):720–724. PMID 18256671 · doi:10.1038/nature06616
Jucker M, Walker LC. (2013). Self-propagation of pathogenic protein aggregates in neurodegenerative diseases. Nature 501(7465):45–51. PMID 24005412 · doi:10.1038/nature12481

Converges on

Inside Out Plaque · Intracellular Amyloid · PANTHOS · Amyloid Nucleation · Protein templating and seeding · Neuronal lysis forming extracellular plaques · Endosomal Nexus · Plaque IgG Sequestration

Studied by

Gunnar Gouras · Ralph Nixon