NAD+‑Dependent CD47 Upregulation Protects Active Neurons from Age‑Related Microglial Pruning

Hypothesis

Age‑related synaptic loss stems not only from dysregulated complement signaling but also from a failure to sustain CD47 "don’t eat me" signals on metabolically active synapses. Boosting neuronal NAD+ restores CD47 expression via SIRT1‑mediated deacetylation, thereby shifting the C1q/CD47 balance toward preservation of high‑activity circuits and maintaining cognitive function despite ongoing neural firing.

Mechanistic Rationale

1. Neuronal activity and NAD+ consumption – Sustained firing increases ATP demand, draining NAD+ pools and reducing SIRT1 activity [1]. Low SIRT1 leads to hyperacetylation of histones at the Cd47 promoter, decreasing its transcription.
2. CD47 loss tips microglial decision‑making – Microglia interpret reduced CD47 as an “eat me” cue, amplifying C1q‑driven phagocytosis of synapses that are still functionally relevant [2].
3. SIRT1 activation restores CD47 – NAD+ precursors activate SIRT1, which deacetylates NF‑κB p65, reducing inflammatory transcription and directly increasing Cd47 gene expression [3]. Elevated CD47 reinstates the inhibitory signal, protecting active synapses from complement‑tagged removal.
4. Preservation of plasticity reserve – Neurons that retain CD47 maintain synaptic integrity, preserving the plasticity reserve observed in low‑activity cells during aging [1]. This creates a selective advantage for circuits that continue to fire, counteracting the hypothesis that all high‑activity neurons are indiscriminately culled.

Predictions and Experimental Design

• Prediction 1: In aged mice, nicotinamide riboside (NR) supplementation will raise hippocampal NAD+ levels by ~30% and increase SIRT1 activity relative to vehicle controls.
• Prediction 2: NR‑treated aged mice will show higher CD47 immunoreactivity on presynaptic terminals of CaMKII‑positive excitatory neurons, without altering overall C1q levels.
• Prediction 3: Synaptic density (measured by synaptophysin and PSD‑95 puncta) in the CA1 region will be significantly greater in NR‑treated mice compared with controls, despite equivalent or higher baseline neuronal activity (assessed via in vivo calcium imaging).
• Prediction 4: Behavioral assays (Morris water maze, novel object recognition) will reveal improved spatial and recognition memory in NR‑treated aged mice, correlating with synaptic preservation.

Experimental groups (n=12 per group): young (3‑mo) vehicle, aged (24‑mo) vehicle, aged + NR (400 mg/kg/day, oral, 12 weeks). Measures: NAD+ quantification (enzymatic assay), SIRT1 activity (fluorometric deacetylation assay), flow‑cytometric CD47 surface labeling on synaptosomes, immunostaining for C1q, microglial phagocytosis markers (CD68+ synaptosomes), longitudinal two‑photon calcium imaging, and cognitive testing.

Statistical analysis: two‑way ANOVA with factors age and treatment, followed by Tukey post‑hoc tests. Falsification occurs if NR fails to elevate CD47 or synaptic density, or if cognitive performance does not improve despite biochemical changes.

Potential Implications

If validated, this hypothesis reframes age‑related cognitive decline as a correctable signaling imbalance rather than inevitable neuronal loss. It suggests that metabolic interventions targeting NAD+ can synergize with activity‑based cognitive enrichment strategies, preserving the brain’s most energetically costly yet functionally vital networks. Conversely, a negative result would support the view that synaptic pruning in aging is primarily driven by irreversible complement dysregulation, redirecting therapeutic focus toward upstream inhibitors of C1q activation or microglial reprogramming.