Version history

{
  "session_id": "sess_SDA-2026-04-21-gap-debate-20260417-033037-c43d12c2_20260421-021018",
  "round_number": 1,
  "agent_persona": "persona-theorist",
  "agent_backend": "minimax/MiniMax-M2.7",
  "action": "propose",
  "content": "1. **Glyco-C1q Lectin Bridging**\n\n**Mechanism:** C1q binds directly to disease-altered N-glycans or O-glycans on myelin debris, apoptotic neurites, or synaptic membranes, while a separate C1q domain engages microglial lectin receptors such as CLEC7A, SIGLEC11, or SIGLEC3/CD33. In this model, C1q effects that appear receptor-specific are actually ternary-complex effects requiring both purified C1q and a glycosylated ligand surface.\n\n**Key Evidence:** C1q recognizes diverse altered-self ligands on apoptotic cells and immune complexes through its globular heads, supporting multivalent pattern-recognition behavior (PMID: 20142073). CD33/SIGLEC3 suppresses microglial uptake of amyloid-beta and genetic variation at CD33 modifies AD risk (PMID: 23623698).\n\n**Testable Prediction:** In purified-protein SPR/BLI or glycan-array assays, recombinant C1q should bind AD-relevant sialylated or desialylated glycan motifs, and this binding should be abolished by enzymatic deglycosylation or C1q globular-head mutation. If C1q shows no binding to purified glycans or glycoprotein substrates under calcium-controlled conditions, the hypothesis is falsified.\n\n**Target Gene/Protein:** `C1QA/C1QB/C1QC` and `CD33`\n\n\n2. **ApoE-Isoform C1q Scaffold**\n\n**Mechanism:** C1q does not primarily bind microglial receptors directly; instead, it binds lipidated APOE-containing particles, forming an opsonic scaffold that secondarily alters engagement of TREM2, LDLR, LRP1, and complement receptors. APOE4 may stabilize or misorient this complex, shifting C1q from debris-clearance signaling toward inflammatory microglial activation.\n\n**Key Evidence:** C1q colocalizes with amyloid plaques and synapses in neurodegeneration models, and complement activation contributes to synapse elimination (PMID: 23525040; PMID: 27662259). APOE genotype strongly affects amyloid deposition, lipid handling, and microglial state in AD, with APOE4 increasing disease risk (PMID: 23410786).\n\n**Testable Prediction:** Purified lipidated APOE2, APOE3, and APOE4 particles should show isoform-dependent direct binding to purified C1q by BLI/SEC-MALS/crosslink-MS. If C1q binds equally weakly to all lipidated APOE isoforms and APOE depletion does not reduce C1q-dependent microglial responses in reconstituted assays, the hypothesis is falsified.\n\n**Target Gene/Protein:** `APOE`\n\n\n3. **C1q-TREM2 Co-Opsonin Complex**\n\n**Mechanism:** C1q directly binds the extracellular Ig-like domain of TREM2 only when TREM2 is presented with an anionic lipid or amyloid-associated ligand, generating a composite binding surface rather than a simple binary receptor-ligand interaction. This would explain why purified single-protein assays may miss the effect unless C1q, TREM2, and lipidated/amyloid substrate are tested together.\n\n**Key Evidence:** TREM2 binds anionic lipids, APOE, and amyloid-associated ligands and controls disease-associated microglia in AD models (PMID: 25728668; PMID: 26675745). C1q-mediated complement signaling contributes to synapse loss in amyloid models before extensive plaque pathology (PMID: 27662259).\n\n**Testable Prediction:** Purified soluble TREM2 should show little or no binding to purified C1q alone, but addition of phosphatidylserine liposomes or fibrillar amyloid-beta should create a measurable ternary complex by BLI or native PAGE. Failure to detect ligand-dependent ternary complex formation across multiple orthogonal assays would falsify the hypothesis.\n\n**Target Gene/Protein:** `TREM2`\n\n\n4. **C1q-Matrix Sequestration Gate**\n\n**Mechanism:** C1q binds extracellular matrix proteins enriched around plaques and dystrophic neurites, especially HSPG core proteins/perineuronal-net components such as HSPG2/perlecan, AGRN, VCAN, or BCAN. This matrix-bound C1q then presents a high-avidity surface for complement activation or microglial adhesion, making apparent “direct C1q effects” dependent on ECM immobilization.\n\n**Key Evidence:** Heparan sulfate proteoglycans accumulate with amyloid plaques and modulate amyloid aggregation and clearance in AD tissue and models (PMID: 14709550). C1q is a collagen-like, multivalent pattern-recognition molecule capable of binding immobilized ligands and initiating classical complement activation (PMID: 20142073).\n\n**Testable Prediction:** Purified C1q should bind heparin/HSPG fragments or purified perlecan/aggrecan-family proteins with salt- and sulfation-dependent kinetics by SPR. If desulfated glycosaminoglycans bind equivalently to sulfated ligands, or purified ECM components fail to enhance C1q-dependent microglial adhesion/phagocytic blockade, the model is falsified.\n\n**Target Gene/Protein:** `HSPG2`\n\n\n5. **Calreticulin-Dependent C1q Signaling**\n\n**Mechanism:** C1q effects on microglia may be mediated through direct binding of its collagen-like tails to CALR/calreticulin, with LRP1 acting as the signaling/endocytic co-receptor. This would create C1q-triggered changes in phagocytosis or cytokine tone without requiring direct engagement of ITGAM/CD11b.\n\n**Key Evidence:** The C1q collagen tail binds calreticulin/CD91-LRP1 complexes on phagocytes and can mediate uptake of apoptotic material (PMID: 14597714). LRP1 is highly expressed in brain innate immune and vascular compartments and regulates amyloid-beta clearance pathways relevant to AD (PMID: 16174740).\n\n**Testable Prediction:** Purified C1q collagen-like tail fragments should bind recombinant CALR by SPR/BLI, and blocking CALR or LRP1 should eliminate C1q-dependent microglial responses even when ITGAM/CD11b is genetically disrupted. If CALR/LRP1 blockade has no effect in a purified C1q add-back assay, this hypothesis is falsified.\n\n**Target Gene/Protein:** `CALR`",
  "tokens_used": "1422"
}