Why have anti-Aβ clinical trials failed despite the established role of Aβ in AD pathogenesis?

Mechanistically-Specific Hypotheses: Aβ-Tau Synergy in Alzheimer's Disease

Hypothesis 1: Prion-Like Cross-Seeding via Toxic Oligomer Interfaces

Title: Aβ oligomers template tau nucleation

Mechanism: Amyloid-β oligomers (particularly dodecameric Aβ*56) expose cryptic β-strand domains that directly interact with tau's R2/R3 microtubule-binding repeat domain, catalyzing conformational conversion into β-sheet-rich nuclei. This cross-seeding bypasses the slow homogeneous nucleation barrier, explaining why Aβ precedes and accelerates tau spreading. The formed tau seeds are then transmitted transsynaptically regardless of ongoing Aβ burden.

Key Evidence:

• Aβ-fibrils accelerate recombinant tau aggregation in vitro (PMID 28504677)

-Human PET imaging demonstrates spatial correlation between early Aβ deposition and subsequent tau spread pattern (PMID 29044071)

Testable Prediction: In a mouse model where Aβ is turned off after plaque formation, tau pathology will continue to propagate and spread if Aβ*56 oligomers are removed early—but will plateau if tau seeds have already invaded downstream brain regions.

Target Gene/Protein: Aβ*56 oligomer / tau R2 domain interface

Hypothesis 2: Kinase-Phosphatase Set-Point Disruption

Title: Aβ permanently raises GSK3β activity

Mechanism: Aβ42 oligomers trigger persistent activation of GSK3β (via PP2A inhibition and FYN-mediated tyrosine phosphorylation of GSK3β-Y216) while simultaneously inhibiting PP2A through POSTNSA-mediated recruitment. This creates a new kinase-phosphatase equilibrium biased toward tau hyperphosphorylation. Because GSK3β also phosphorylates Aβ-metabolizing enzymes (BACE1, IDE), this circuit generates more Aβ and more phospho-tau in a feed-forward loop.

Key Evidence:

• Post-mortem AD brain shows reduced PP2A activity and elevated active GSK3β (PMID 12431986)
• GSK3β inhibition reduces both Aβ and tau pathology in 3xTg mice (PMID 18930951)

Testable Prediction: Acute GSK3β inhibition in late-stage 3xTg mice will reduce tau phosphorylation but NOT restore PP2A activity or reduce Aβ—indicating the set-point is locked and requires earlier intervention.

Target Gene/Protein: GSK

Skeptic's Evaluation: Aβ-Tau Synergy Hypotheses

Hypothesis 1: Prion-Like Cross-Seeding via Toxic Oligomer Interfaces

Strongest Weakness

Unresolved structural basis for the claimed interface. The hypothesis asserts that Aβ*56 oligomers "expose cryptic β-strand domains" that interact specifically with tau's R2/R3 repeat domain, but no study has characterized the actual structural interaction. Cross-seeding in vitro typically requires direct mixing of recombinantly aggregated proteins under non-physiological concentrations—this does not establish that membrane-associated oligomers in the human brain template tau conformation with meaningful efficiency.

Counter-Evidence

• Reproducibility concerns with Aβ56: The original Aβ56 characterization (Lesné et al., 2006) has not been robustly replicated by independent groups. Many studies now focus on Aβ*56 as an artifact or minor species in vivo. The reliance on this specific oligomer is a fragile foundation (see PMID 23254930 for review of replication issues).
• Tau propagation can occur independently: In mouse models where tau is genetically deleted or silenced, Aβ deposition still occurs normally—but critically, tau pathology can still spread in the absence of Aβ in models using human tau (Busche et al., 2019, PMID 30643256). This suggests tau seeding has autonomous capacity.

Pointed Question

If the cryptic β-strand domains are truly "cryptic" (not exposed in mature fibrils but only in oligomers), how do you propose to specifically detect and quantify these transient interfacial surfaces in vivo in human brain? Without a method to verify the presence and abundance of these interfaces during life, the mechanism remains inferential.

Confidence: Weak → Moderate (conditional)

The concept of cross-seeding is mechanistically plausible and in vitro evidence supports the possibility. However, reliance on Aβ56 (reproducibility concerns) and absence of structural characterization of the proposed interface substantially weaken the specific claim. Confidence would increase substantially if validated structural data on the Aβ-tau interface became available, or if Aβ56 were replaced with a better-evidenced oligomeric species.

Hypothesis 2: Kinase-Phosphatase Set-Point Disruption

Strongest Weakness

The causal direction is unproven. The hypothesis proposes that Aβ oligomers "permanently" raise GSK3β activity, creating a feed-forward loop. However, all cited evidence is correlative—post-mortem brains showing reduced PP2A and elevated GSK3β are end-stage observations that cannot establish causation. GSK3β elevation could be a consequence of neurodegeneration, inflammation, or upstream metabolic dysfunction rather than a driver of the Aβ-tau

Domain Expert Evaluation: Aβ-Tau Synergy Hypotheses

1. Translational Potential Assessment

Ranking by Clinical Readiness

| Hypothesis | Translational Potential | Rationale |
|------------|------------------------|-----------|
| Hypothesis 2 (Kinase-Phosphatase Set-Point Disruption) | Highest | Targets downstream effector mechanism; multiple kinase inhibitor candidates already in pharmaceutical development pipelines |
| Hypothesis 1 (Prion-Like Cross-Seeding) | Moderate | Mechanistically compelling but structural basis unresolved; timing windows may be too narrow for clinical intervention |
| Unspecified/Implied mechanisms from H2 | Highest overall | FYN and GSK3β inhibitors have existing safety profiles from other indications; combination with existing anti-Aβ antibodies is feasible |

Clinical Landscape Fit Analysis

Current context: Post-lecanemab/donanemab approvals, we now have anti-Aβ agents with modest but real efficacy. The field is grappling with how to improve on these results. The key question becomes: can we enhance anti-Aβ efficacy by simultaneously modulating tau-relevant downstream pathways?

Hypothesis 2 advantage: Unlike directly targeting Aβ (which has now been "validated" but with limited effect size), kinase-phosphatase modulation represents an adjunctive strategy that could be layered onto existing approved treatments. This is immediately actionable.

2. Clinical Evidence and Patient Population Fit

Hypothesis 2: Kinase-Phosphatase Set-Point Disruption

Current Clinical Evidence:

• Tideglusib (GSK3β inhibitor): Completed Phase II trials in AD and Down syndrome (NCT00948259, NCT01689246). Showed acceptable safety but failed primary cognitive endpoints. Importantly, CSF p-tau181 reduction was observed—suggesting target engagement was achievable.
• FYN inhibitors: Being developed for cancer with excellent safety profiles (e.g., dasatinib, though not selective for FYN). Saracatinib (AZD0530) showed promise in preclinical AD models (PMID 26324930) and has been repurposed—Phase Ib trial showed safety but limited efficacy as single agent.
• PP2A activators: Eryptosine analogs in early development; no human data yet.

Patient Population Fit:

• Optimal: Patients with established Aβ pathology but early tau (pre-MCI to mild dementia, A+T+ by AT(N) framework)
• Rationale: The "set-point" disruption model implies a therapeutic window before irreversible tau-mediated neurodegeneration
• Concern: Patients with advanced tau Braak staging (IV-VI) may have passed the point of kinase-mediated modulation benefit

Safety Considerations:

• GSK3β inhibition carries theoretical cancer risk (GSK3β phosphorylates β-catenin degradation complex; constitutive activation could theoretically promote proliferation in rare cell populations)
• FYN inhibition has cytopenias and fluid retention concerns from oncology experience
• Mitigation: Localized CNS delivery or blood-brain barrier-penetrant selective inhibitors currently in development

Hypothesis 1: Prion-Like Cross-Seeding (If pursued despite concerns)

Current Clinical Evidence:

• Anti-tau antibodies targeting seeding-competent tau species (e.g., semorinemab, gosuranemab) have failed in Phase II/III trials (TAURIAN, NCT02824150)
• This undermines the premise that tau seeds are the primary therapeutic target
• However, these trials did not stratify by Aβ burden—a critical missed opportunity

Patient Population Fit:

• Theoretical optimal: Very early Aβ-positive, tau-negative individuals (A+T- stage)
• Problem: This population is difficult to identify prospectively and has not been the focus of registration trials
• Registry-enriched screening using plasma p-tau217 could now enable such selection

3. Response to Skeptic's Challenges

Primary Challenge Addressed: Aβ*56 Reproducibility Crisis

The Skeptic correctly identifies the Aβ*56 replication failure as a foundational