Hypothesis

The progressive decline of the CDP-choline (Kennedy) pathway in aging brain is not merely a loss of biosynthetic capacity but an adaptive mechanism that spares NAD+ by reducing PARP‑1 activation. Sustained phosphatidylcholine synthesis generates oxidizable phospholipid substrates that, upon peroxidation, trigger DNA strand breaks and PARP‑1–mediated NAD+ consumption. When NAD+ becomes limiting, continued pathway activity would exacerbate bioenergetic failure and accelerate neuronal senescence. Thus, the observed caspase‑3–mediated cleavage of CCTα and downregulation of GPCPD1/CHKA represent a programmed downregulation that preserves NAD+ pools for essential repair processes.

Mechanistic Rationale

1. Link between phospholipid turnover and PARP‑1
   • Phosphatidylcholine (PC) enriched membranes are susceptible to lipid peroxidation, producing aldehydes such as malondialdehyde that can induce DNA damage.[2]
   • DNA breaks activate PARP‑1, which consumes NAD+ to synthesize poly(ADP‑ribose) chains.[4]
   • In Huntington’s disease models, GPCPD1 deficiency elevates oxidative stress markers, suggesting that impaired PC remodeling increases peroxidation susceptibility.[2]
2. NAD+ allocation trade‑off
   • NAD+ fuels sirtuins, CD38, and PARPs; its decline with age forces cells to prioritize essential functions.[4]
   • Maintaining high PC synthesis rates would continuously generate peroxidizable lipids, creating a futile cycle of DNA damage and NAD+ drain.
   • Downregulating the Kennedy pathway reduces substrate supply for peroxidation, thereby lowering PARP‑1 activation and conserving NAD+ for mitochondrial maintenance and sirtuin‑mediated deacetylation.
3. Caspase‑3 as a sensor of metabolic stress
   • Caspase‑3 activation during cellular stress cleaves CCTα, suppressing PC synthesis by >70%.[1]
   • This proteolytic event could be interpreted as a metabolic checkpoint: when ATP/CTP or NAD+ levels fall, caspase‑3 mediated CCTα cleavage attenuates phospholipid biosynthesis to prevent NAD+ over‑consumption.

Testable Predictions

• Prediction 1: In aged neurons, pharmacological or genetic inhibition of CCTα cleavage will increase PC synthesis rates, elevate lipid peroxidation products (e.g., 4‑HNE), and augment PARP‑1 activity, resulting in measurable NAD+ depletion compared with controls.
  • Experiment: Use CRISPR‑knockin of a cleavage‑resistant CCTα allele in primary mouse cortical neurons; measure PC synthesis (choline incorporation assay), MDA/4‑HNE levels (ELISA or HPLC), PARP‑1 activity (PAR immunoblot), and NAD+ concentration (enzymatic cycling assay) in young vs. aged cultures.
• Prediction 2: Preventing age‑related Kennedy pathway downregulation will exacerbate NAD+‑dependent phenotypes such as reduced SIRT1 activity and increased mitochondrial ROS, whereas enhancing NAD+ availability (e.g., with NR) will rescue these effects only when pathway activity is restrained.
  • Experiment: Treat aged neurons with NR alongside CCTα overexpression; assess SIRT1 deacetylase activity (fluorometric assay), mitochondrial membrane potential (JC‑1), and ROS (MitoSOX). Expect that NR rescues phenotypes only in the presence of CCTα suppression.
• Prediction 3: In vivo, mice with brain‑specific expression of a non‑cleavable CCTα will show accelerated cognitive decline and greater NAD+ loss in the hippocampus relative to littermate controls, despite having higher PC levels.
  • Experiment: Generate CamKII‑Cre; CCTα^D/E (aspartate/glutamate mutant resistant to caspase‑3) mice; perform Morris water maze at 12 and 18 months, quantify hippocampal NAD+ and PC species via LC‑MS, and assess neurodegeneration markers (NeuN loss, GFAP astrocytosis).

Falsifiability

If experiments reveal that blocking CCTα cleavage does not increase lipid peroxidation, PARP‑1 activation, or NAD+ consumption, or that forced PC synthesis improves NAD+‑dependent functions without adverse effects, the hypothesis would be falsified. Conversely, consistent evidence that sustained Kennedy pathway activity accelerates NAD+ decline and neuronal dysfunction would support the adaptive downregulation model.