Version history

{
  "session_id": "sess_SDA-2026-04-02-gap-ev-ad-biomarkers",
  "round_number": 1,
  "agent_persona": "persona-theorist",
  "agent_backend": "claude-haiku-4-5",
  "action": "propose",
  "content": "# Mechanistic Hypotheses for EVs as Early AD Biomarkers\n\n## The \"Prion-Like Templating Cascade\" Hypothesis\n\nI propose that disease-specific EV subtypes function as nucleation templates that accelerate pathological protein conversion in a feed-forward manner long before plaque/tangle deposition becomes extensive. Specifically, neurons experiencing early metabolic stress release EVs enriched in oligomeric amyloid-β and phosphorylated tau conformers that seed conversion of naive proteins in recipient cells—including peripheral immune cells accessible via blood sampling. This predicts that: (1) plasma EVs from preclinical AD should show >3-fold enrichment in β-sheet-prone tau epitopes compared to age-matched controls, detectable by conformer-specific biosensors; (2) EV-to-neuron transfer in microfluidic co-cultures should produce 5-10x amplification of pathological tau/Aβ in recipient neurons within 48 hours, quantifiable by high-resolution live imaging; and (3) EV composition should predict cognitive decline trajectory better than static CSF biomarkers, even among cognitively normal APOE4+ individuals. The testable prediction is that blocking EV uptake via heparin sulfate antagonists halts this propagation *in vitro*, providing mechanistic validation.\n\n## The \"Glial Licensing\" Alternative\n\nRather than assuming neuronal EVs are primary drivers, I hypothesize that early AD involves a *licensing phase* where astrocytes and microglia release EV populations carrying DAMPs (damage-associated molecular patterns) and pro-inflammatory lipids that prime peripheral immune cells for CNS tropism. These glial EVs carry specific phospholipid signatures (oxidized phosphatidylserine, phosphatidic acid) and cytokine cargo (IL-18, TNF-α) that systemically reshape immune surveillance. This predicts: (1) glial-derived EVs should precede neuronal markers by 3-5 years in longitudinal plasma biobanks; (2) EV phospholipid composition (via lipidomics) should distinguish preclinical AD from healthy aging with >90% sensitivity; and (3) transfer of such EVs to primary monocytes should reprogram them toward an AD-associated phenotype (upregulation of CD11c, IL-1β) without exposure to actual amyloid—a pure EV-driven effect.\n\n## Testable Predictions & Experimental Design\n\nBoth hypotheses converge on a critical prediction: **EV biomarkers should outperform current static markers in predicting conversion to MCI specifically in cognitively normal individuals with normal amyloi",
  "tokens_used": "656"
}