Hypothesis

Chronic choline limitation reduces CDP‑choline–derived phosphatidylcholine, altering the lipid microenvironment of astrocytic AQP4 water channels. This biophysical shift preferentially impairs autophagic sequestration of tau while sparing—or even enhancing—glymphatic clearance of amyloid‑β during sleep, biasing the nightly "edit" toward retention of tau‑rich synapses.

Mechanistic Basis

• Phosphatidylcholine enrichment stabilizes AQP4 tetramers in orthogonal arrays, facilitating bulk CSF‑ISF exchange.
• CDP‑choline shortage leads to increased lysophosphatidylcholine and diacylglycerol, which increase membrane curvature and promote AQP4 mislocalization to non‑endfoot domains.
• Mislocalized AQP4 reduces perivascular convective flow but preserves microdomain‑scale fluid fluxes that support lysosomal‑autophagosome fusion in astrocytes.
• Consequently, amyloid‑β, which relies on bulk convective clearance, accumulates less efficiently, whereas tau, which is cleared via autophagy‑lysosomal pathways within astrocytes, suffers a greater clearance deficit.

Testable Predictions

1. In aged mice fed a choline‑deficient diet, nocturnal glymphatic influx of CSF‑derived fluorescent amyloid‑β analogues will be reduced by ~30 % relative to controls, while autophagic flux (LC3‑II/I ratio, p62 degradation) in cortical astrocytes will drop by >50 % during sleep.
2. Chronic citicoline supplementation will restore AQP4 orthogonal array organization (quantified by freeze‑fracture EM) and normalize autophagic flux without fully rescuing bulk glymphatic inflow, resulting in a selective decrease in insoluble tau but not amyloid‑β burden.
3. Pharmacological activation of autophagy (e.g., low‑dose rapamycin) in choline‑deficient mice will rescue tau clearance and ameliorate memory deficits, even when glymphatic flow remains impaired.

Experimental Approach

• Use aged wild‑type mice; assign to choline‑deficient, choline‑sufficient, and choline‑sufficient + citicoline groups for 8 weeks.
• Measure glymphatic function via intrathecal infusion of fluorescently labeled amyloid‑β monomer and two‑photon imaging of CSF‑ISF exchange during natural sleep.
• Assess astrocytic autophagy by immunobluct of LC3‑II, p62, and lysosomal cathepsin activity in FACS‑sorted astrocytes from cortical tissue harvested at ZT6 (mid‑sleep).
• Quantify AQP4 orthogonal array size using freeze‑fracture replica immunogold labeling and EM.
• Endpoint: insoluble tau and amyloid‑β levels by ELISA and immunohistochemistry; behavioral readout via Morris water maze.

Falsifiability: If choline deficiency reduces both glymphatic influx and autophagic flux proportionally, or if citicoline equally rescues both pathways, the hypothesis of a selective autophagic bias is falsified.