Hypothesis

Aging-related synaptic loss results from a decline in microglial NAD+-dependent metabolism that weakens the CD47 'don't eat me' signal, causing indiscriminate phagocytosis of synapses rather than a selective, efficiency‑driven pruning process.

Proposed Mechanism

In young microglia, NAD+ fuels SIRT1 activity, which deacetylates NF‑κB and promotes transcription of Cd47 and other inhibitory receptors, preserving the 'don't eat me' cue while allowing activity‑dependent complement tagging (C1q) to mark weak synapses for removal 2 3. With age, mitochondrial dysfunction lowers NAD+ levels, reducing SIRT1 activity and leading to hyperacetylated NF‑κB that drives excess C1q expression and suppresses Cd47 transcription 1 4. Consequently, microglia lose the ability to discriminate between active and inactive synapses, shifting from apoptosis‑mediated refinement in early life to necrosis‑like, uncontrolled clearance in the aged brain 5 6. This metabolic shift explains why surviving neurons are not functionally superior; the pruning machinery becomes noisy rather than optimized.

Testable Predictions

1. Metabolic rescue – Chronic supplementation with NAD+ precursors (e.g., nicotinamide riboside) in aged mice will restore microglial NAD+ and SIRT1 activity, increase Cd47 mRNA and protein levels, decrease C1q deposition on synapses, and preserve synapse density compared with vehicle controls.
2. Phagocytic specificity – In vivo two‑photon imaging of microglial-synapse interactions will show that, after NAD+ boosting, microglia preferentially engulf low‑activity dendritic spines (identified by reduced calcium signaling) while sparing high‑activity spines, whereas untreated aged mice display non‑selective engulfment.
3. Cognitive correlation – Improvements in synaptic selectivity will correlate with better performance on hippocampal‑dependent tasks (e.g., Morris water maze), establishing a causal link between microglial metabolic state, precise pruning, and cognitive outcomes.

These experiments directly test whether age‑related synaptic loss stems from a metabolic failure of the microglial checkpoint rather than an adaptive, efficiency‑driven eviction of neurons.