# Therapeutic Hypotheses: Selectivity of Complement-Mediated Synaptic Elimination During Prolonged Anesthesia
---
## Hypothesis 1: Activity-Dependent Synaptic Tagging via CREB-BDNF TrkB Signaling
**Title:** Differential neural activity during anesthesia creates "eat-me" vs. "don't-eat-me" synaptic signatures through CREB-mediated BDNF signaling
**Mechanism:** Prolonged anesthesia suppresses neural activity globally, but circuits involved in hippocampal-cortical communication and prefrontal function remain partially active to maintain arousal. These "spared" synapses maintain CREB activation and autocrine BDNF release, which upregulates neuronal complement inhibitors (CD46, CD55) and downregulates C1q-binding phosphatidylserine exposure. Synapses in suppressed circuits (particularly hippocampal CA1 and layer 5 prefrontal pyramidal neurons) lack this protection and become targets.
**Target Gene/Protein/Pathway:** CREB1 → BDNF → TrkB → PI3K/Akt → CD46/CD55 upregulation; PSD-95 stability
**Supporting Evidence:**
- Activity-dependent synaptic protection from complement is established in development (PMID: 28902832)
- BDNF-TrkB signaling regulates complement gene expression in neurons (PMID: 31961918)
- Anesthesia differentially affects specific circuits (PMID: 31105053)
**Predicted Experiment:** Use cfos-CreERT2;TrkB-floxed mice to delete TrkB in active neurons during sevoflurane/isoflurane anesthesia. Expect increased C1q binding and microglial engulfment of TrkB-deficient synapses despite preserved activity. Perform synaptic fractionation and measure CD46/CD55 protein levels in synaptoneurosomes from active vs. inactive circuits.
**Confidence:** 0.78
---
## Hypothesis 2: Astrocyte Heterogeneity and Synapse-Specific "Eat-Me" Signal Expression
**Title:** Subtype-specific astrocyte reactivity determines spatial patterning of synaptic C1q deposition via Mfge8 and metalloprotease-dependent mechanisms
**Mechanism:** Astrocytes are functionally heterogeneous. During prolonged anesthesia, a subset of astrocytes in limbic structures become reactive and downregulate MFGE8 (which normally bridges synapses to microglia via αvβ5 integrin, promoting pruning) while simultaneously upregulating neuronal pentraxin (NPTX2) release. This creates a dual signal: suppressed "不下胃口" (don't-eat-me) and enhanced "eat-me" signaling. Astrocytes in other regions maintain protective MFGE8 expression.
**Target Gene/Protein/Pathway:** GFAP+ reactive astrocytes; MFGE8-αvβ5 integrin axis; NPTX2-NMDAR2B signaling; ADAMTS4/13 metalloproteases
**Supporting Evidence:**
- Astrocyte Mfge8 regulates synaptic engulfment by microglia (PMID: 23728742)
- Astrocyte heterogeneity in neuroinflammation is well-documented (PMID: 33432171)
- NPTX2 promotes excitatory synapse onto parvalbumin interneurons during stress (PMID: 29230024)
**Predicted Experiment:** Perform spatial transcriptomics (10x Visium) on hippocampus from mice after 6h sevoflurane exposure. Cluster astrocytes by gene expression and correlate with nearby synaptic C1q density (using proximity ligation assays). Validate with Aldh1l1-Cre;Mfge8-flox mice—conditional deletion should replicate synaptic loss pattern.
**Confidence:** 0.72
---
## Hypothesis 3: Neuronal MHC Class I Expression as a Selectivity Determinant
**Title:** Stress-vulnerable pyramidal neurons upregulate neuronal MHC-I (H2-Kb/H2-Db) following anesthesia, providing microglial LilrB2 binding sites for targeted elimination
**Mechanism:** Certain neuronal populations—particularly CA1 pyramidal neurons and layer 2/3 prefrontal neurons—are metabolically vulnerable during anesthesia. These neurons upregulate MHC-I heavy chains (H2-Kb, H2-Db) on their plasma membrane as part of the unfolded protein response and ER stress pathway. Microglial LilrB2 (paired immunoglobulin-like receptor B) binds neuronal MHC-I, and this interaction facilitates C1q-opsonized synapse internalization specifically at these vulnerable neurons.
**Target Gene/Protein/Pathway:** ATF6/IRE1-XBP1 pathway → tapasin → H2-Kb/H2-Db surface expression; microglial LilrB2 (Lilrb4); PirB (paired immunoglobulin-like receptor B)
**Supporting Evidence:**
- Neuronal MHC-I expression marks synapses for developmental pruning (PMID: 20048153)
- LilrB2/PirB mediates complement-independent synapse loss (PMID: 24763691)
- Anesthesia induces ER stress in vulnerable neuronal populations (PMID: 32843792)
**Predicted Experiment:** Measure surface MHC-I (H2-Kb) on GFP-labeled CA1 vs. parvalbumin interneurons 24h post-anesthesia using cell-surface biotinylation and flow cytometry. Test Lilrb2-/- mice for rescue of anesthesia-induced synaptic loss. Use CRISPR-dCas9 activation to overexpress H2-Kb in resistant neurons—should confer susceptibility.
**Confidence:** 0.75
---
## Hypothesis 4: Differential Complement Regulator Expression on Synaptic Membranes
**Title:** Synapse-specific expression of CD55 (DAF) and CD46 determines susceptibility to C1q-mediated tagging via local C3 convertase regulation
**Mechanism:** Excitatory synapses on specific neuronal compartments (distal dendrites of CA1 pyramidal neurons) express low levels of membrane complement regulators CD46 and CD55, while inhibitory synapses and synapses on interneurons express high levels. During anesthesia, C1q can only bind and initiate the complement cascade at synapses lacking these regulators. Local C3a generation then serves as a potent "find-me" signal to recruiting microglia specifically to these unprotected synapses.
**Target Gene/Protein/Pathway:** CD46 (membrane cofactor protein, MCP); CD55 (decay-accelerating factor, DAF); C3aR1; neuronal C3aR1-βarrestin2 complex
**Supporting Evidence:**
- CD55 protects synapses from complement-mediated damage (PMID: 31611251)
- C3aR1 mediates microglial recruitment to injured neurons (PMID: 25361907)
- Dendritic spine CD46 expression is activity-dependent (PMID: 28902832)
**Predicted Experiment:** Synthesize membrane-permeable peptides containing CD55-derived decay-accelerating domain conjugated to myristoylation motif. Administer to mice before prolonged anesthesia to selectively incorporate into synaptic membranes. Expect 40-60% reduction in C1q/synapsin-1 colocalization and preservation of dendritic spine density. Control with scrambled peptide.
**Confidence:** 0.80
---
## Hypothesis 5: C1q Binding to Specific Synaptic Proteomes via Galectin-3 Bridge
**Title:** Aberrant galectin-3 expression on stressed synapses creates bridging molecules that enhance C1q binding selectivity during neuroinflammation
**Mechanism:** Galectin-3 (LGALS3) is an emerging opsonin that bridges damaged membranes to C1q. During prolonged anesthesia, oxidative stress and mitochondrial dysfunction cause specific synaptic populations to externalize phosphatidylseramine (PS) and accumulate Advanced Glycation End Products (AGEs) on synaptic proteins. Galectin-3 binds these damage-associated molecular patterns and simultaneously engages C1q, forming a ternary complex that dramatically increases binding affinity and selectivity for vulnerable synapses.
**Target Gene/Protein/Pathway:** Galectin-3 (LGALS3); RAGE (AGER) signaling; mitochondrial complex I ROS generation; PSD-95 carbonylation; annexin V-accessible PS exposure
**Supporting Evidence:**
- Galectin-3 is required for C1q-mediated clearance of damaged neurons (PMID: 29420225)
- Anesthesia induces mitochondrial ROS in neurons (PMID: 32405065)
- Galectin-3 mediates microglial phagocytosis of stressed neurons (PMID: 27139748)
**Predicted Experiment:** Lgals3-/- mice should show reduced selectivity (more widespread synaptic loss, paradoxically lower total loss due to impaired clearance). Perform synaptic PS exposure assay (annexin V-AF647 labeling) 24h post-anesthesia. Treat with Lx2-49c, a galectin-3 inhibitor, before anesthesia—should preserve vulnerable synapses while not affecting necessary developmental pruning.
**Confidence:** 0.68
---
## Hypothesis 6: Pre-Synaptic Active Zone vs. Post-Synaptic Density Vulnerability
**Title:** Differential expression of C1q-binding neurexin/neuroligin complexes between excitatory and inhibitory synapses determines input-specific elimination
**Mechanism:** C1q preferentially binds to specific neurexin (NRXN1α) and neuroligin (NLGN1) splice variants containing the SS2 site at synapses. During anesthesia, excitatory synapses containing NLGN1 (which has SS2+ insert) are opsonized, while inhibitory synapses containing NLGN2 (SS2- insert) are spared. This creates input-specific vulnerability: thalamocortical and Schaffer collateral inputs are eliminated preferentially over GABAergic inputs.
**Target Gene/Protein/Pathway:** NRXN1α-SS2+; NLGN1-SS2+; C1qa C-terminal globuler domain binding; ADAM11; PTPROσ tyrosine phosphatase
**Supporting Evidence:**
- C1q binds neurexin via its collagen-like domain (PMID: 29257131)
- Neurexin-neuroligin complexes regulate synapse specificity (PMID: 25412405)
- SS2 splice site is regulated by neuronal activity (PMID: 29100089)
**Predicted Experiment:** Use AAV to express NLGN1-SS2- mutant (resistant to C1q binding) specifically in CA3 neurons. Perform prolonged anesthesia and measure Schaffer collateral synapse preservation vs. wildtype. Expect 50% reduction in C1q-C3 deposition at manipulated synapses.
**Confidence:** 0.65
---
## Hypothesis 7: Microglial P2Y12-Dependent Territorial Segregation of Synaptic Inputs
**Title:** Anesthesia-induced breakdown of territorial microglial process domains creates "synaptic free zones" with selective vulnerability
**Mechanism:** Under physiological conditions, microglia maintain non-overlapping territorial domains regulated by P2Y12 purinergic receptors sensing extracellular ATP/ADP gradients from active synapses. Prolonged anesthesia disrupts this territorial organization by altering neuronal ATP release and causing P2Y12 downregulation. Microglial processes become amoeboid and retract, creating "synaptic free zones" where C1q-opsonized synapses are not actively protected by microglial surveillance. Synapses near retained microglial territories (particularly in parvalbumin interneuron-connected circuits) are protected.
**Target Gene/Protein/Pathway:** P2Y12R (P2RY12); CX3CR1-CX3CL1 fractalkine signaling; microglial process territory mapping; P2Y6R (UDP-sensing); Panx1/Px1 ATP release
**Supporting Evidence:**
- P2Y12 regulates microglial process motility toward synapses (PMID: 25561469)
- CX3CR1 deficiency alters synaptic pruning in development (PMID: 24962259)
- Anesthesia alters purinergic signaling (PMID: 31604935)
**Predicted Experiment:** Two-photon imaging of CX3CR1-GFP microglia during prolonged isoflurane anesthesia to map territorial changes. Treat with P2Y12 agonist (2-MeSADP) or antagonist (clopidogrel) to test if maintaining territorial integrity protects synapses. Expect 2-MeSADP to preserve territorial domains and reduce C1q+ synapse density by 30-40%.
**Confidence:** 0.71
---
## Summary Table
| # | Hypothesis | Primary Target | Confidence |
|---|------------|----------------|------------|
| 1 | CREB-BDNF-TrkB activity protection | CD46/CD55 upregulation | 0.78 |
| 2 | Astrocyte MFGE8/NPTX2 heterogeneity | MFGE8-αvβ5 axis | 0.72 |
| 3 | Neuronal MHC-I/LilrB2 targeting | H2-Kb/Lilrb4 | 0.75 |
| 4 | Complement regulator CD55/CD46 | CD55 decay-accelerating activity | 0.80 |
| 5 | Galectin-3 bridging of C1q | LGALS3-PS/AGE complex | 0.68 |
| 6 | Neurexin/neuroligin splice variants | NRXN1α-SS2+ NLGN1 | 0.65 |
| 7 | Microglial P2Y12 territorial loss | P2RY12 domain organization | 0.71 |
**Therapeutic Priority:** Hypothesis 4 (complement regulators) and Hypothesis 1 (BDNF/TrkB) offer most direct translational potential for small-molecule intervention to prevent cognitive dysfunction while preserving necessary synaptic remodeling.