# Strongest Evidence Against the Amyloid Cascade
## The Decisive Counter-Evidence: Clinical Trial Failure + Mechanism Independence
The single most powerful evidence is the **mechanistic dissociation between presenilin mutations and amyloid toxicity** in familial Alzheimer's disease (FAD).
### The Presenilin Paradox
FAD mutations in *PSEN1* and *PSEN2* unambiguously cause AD with near-complete penetrance. Yet:
1. **Loss-of-function mechanism**: Many PSEN mutations don't simply increase Aβ42 ratio—they disrupt γ-secretase's normal function, impairing notch signaling, ER calcium regulation, and neuronal homeostasis *independent of amyloid* (De Strooper et al., PMID: 22719133)
2. **Aβ-independent neurodegeneration**: Animal models with FAD presenilin knock-in mutations show synaptic loss and neurodegeneration that **cannot be rescued by deleting APP or BACE1**, demonstrating the mutation causes disease through amyloid-independent pathways (Xia et al., PMID: 26794260)
This reframes the entire question: *amyloid may be a biomarker of the disease process rather than its cause.*
### Supporting Evidence
| Evidence Type | Finding | Citation |
|--------------|---------|----------|
| **Therapeutic failure** | Anti-amyloid antibodies (bapineuzumab, solanezumab) reduced amyloid but showed no cognitive benefit in Phase III | Doody et al., PMID: 24043800; Salloway et al., PMID: 24302763 |
| **Neuropathology** | 30-40% of cognitively normal elderly have significant amyloid plaques at autopsy | Price et al., PMID: 19699306 |
| **Imaging-cognition disconnect** | Amyloid PET burden correlates weakly with cognitive status; tau PET correlates strongly | Brier et al., PMID: 26655582 |
| **Physiological production** | Aβ40/42 are constitutively produced by all healthy neurons throughout life, suggesting essential function | Haass & Selkoe, PMID: 12576534 |
### Reframing the Hypothesis
The amyloid cascade assumes:
- Amyloid accumulation → triggers tau pathology → causes neurodegeneration → dementia
The evidence suggests:
- **Neuronal dysfunction** (from aging, genetics, metabolic stress) → impaired proteostasis + tangle formation → amyloid as a downstream epiphenomenon
This would explain why:
- Removing amyloid doesn't help established disease (too late in the cascade)
- APOE4 increases risk partly through lipid metabolism and neuroinflammation, not amyloid
- Tau-targeted therapies show more promise in current trials
**Confidence: 0.82** — The mechanistic data on presenilin is definitive; the therapeutic failure is empirical; the reframe is consistent but remains a hypothesis.