Hypothesis

Chronic choline deficiency impairs the Kennedy pathway, reducing phosphatidylcholine (PC) synthesis and causing compensatory remodeling of mitochondrial cardiolipin (CL) species. This altered CL becomes highly susceptible to peroxidation by mitochondrial ROS, triggering cytochrome c release, caspase‑3 activation, and synaptotoxic cascades that precede amyloid‑β plaque formation. Restoring PC synthesis via CDP‑choline supplementation normalizes CL composition, blocks peroxidation, and preserves synaptic integrity independent of amyloid load.

Mechanistic Rationale

1. Choline → PC → Mitochondrial Membrane Homeostasis

   • Choline is the rate‑limiting substrate for PC synthesis via the CDP‑choline (Kennedy) pathway [3].
   • PC supplies the phospholipid backbone for mitochondrial inner‑membrane remodeling, especially CL, which requires a saturated acyl chain profile for optimal electron‑transport‑chain (ETC) function.
   • When choline falls, PC synthesis drops, and mitochondria compensate by incorporating more polyunsaturated fatty acids into CL, increasing its peroxidation propensity.

2. Peroxidized Cardiolipin → Cytochrome c Release → Synaptic Loss

   • Oxidized CL loses its affinity for cytochrome c, promoting its detachment into the intermembrane space and subsequent caspase‑3 activation [2].
   • Caspase‑3 cleaves synaptic proteins (e.g., SNAP‑25, PSD‑95), leading to dendritic spine shrinkage and loss—a process observable before significant amyloid accumulation in AD models.
   • This mechanism aligns with post‑mortem findings of reduced choline acetyltransferase and plasmalogen choline in AD brains [4], reflecting a broader membrane‑lipid failure.

3. CDP‑Choline Rescue

   • Exogenous CDP‑choline bypasses the rate‑limiting step, boosting PC synthesis, restoring CL saturation, and decreasing ROS‑driven peroxidation [3].
   • Improved CL integrity enhances ETC efficiency, raises glutathione levels, and inhibits apoptosis, consistent with observed cognitive benefits in older adults after 12‑week supplementation [5].

Testable Predictions

| Prediction | Experimental Approach | Expected Outcome if Hypothesis True | |------------|----------------------|--------------------------------------| | Elevated peroxidized CL in preclinical AD | Isolate extracellular vesicles or CSF mitochondria from cognitively normal older adults with low plasma choline (<8 µmol/L) and assay CL peroxidation using LC‑MS/MS‑based hydroxy‑CL biomarkers. | Higher peroxidized‑CL/total‑CL ratio correlates with low choline and predicts future cognitive decline independent of amyloid‑β PET status. | | CDP‑choline normalizes CL peroxidation | Randomized, double‑blind trial in obese young adults (age 25‑35) with baseline low choline and elevated neurofilament light (NfL). Participants receive 500 mg/day CDP‑choline or placebo for 6 months; repeat CSF mitochondrial CL peroxidation and NfL measurements. | CDP‑choline group shows significant reduction in peroxidized CL and attenuated NfL rise vs. placebo. | | Blocking CL peroxidation prevents synaptotoxicity in choline‑deficient mice | Feed mice a choline‑deficient diet for 12 weeks, with/without the specific CL peroxidation inhibitor SS‑31 (elamipretide). Assess synaptic protein levels (synaptophysin, PSD‑95) and caspase‑3 activity in hippocampus. | SS‑31 rescues synaptic markers and reduces caspase‑3 activation despite choline deficiency, whereas vehicle‑treated deficient mice show synaptic loss. | | Peroxidized CL mediates amyloid‑independent toxicity | In APP/PS1 mice, cross with a CL‑peroxidation‑prone strain (e.g., cardiolipin synthase heterozygous). Compare amyloid load and cognition at 9 months. | Increased peroxidized CL exacerbates memory deficits without altering plaque burden, indicating a parallel, amyloid‑independent pathway. |

Falsifiability

If longitudinal human studies reveal no association between plasma choline, mitochondrial CL peroxidation biomarkers, and future cognitive trajectories—or if CDP‑choline supplementation fails to modify CL peroxidation or NfL levels—the hypothesis would be refuted. Similarly, if inhibiting CL peroxidation does not protect synapses in choline‑deficient animal models, the mechanistic link would be invalidated.

Implications

This hypothesis shifts focus from amyloid‑centric models to mitochondrial lipid redox biology as an early, actionable node in AD pathogenesis. It suggests that choline status, modulatable through diet or CDP‑choline, could be leveraged to preserve mitochondrial membrane integrity and stave off synaptic decline decades before clinical dementia emerges.