Hypothesis: Age-related decline in neuronal ATP production elevates extracellular adenosine, which activates microglial A2A receptors to increase complement C3 synthesis and synapse opsonization, thereby coupling metabolic inefficiency to complement-mediated synapse pruning.

Rationale: The seed idea frames neuronal loss in aging as an optimization process under falling energy budgets. Complement-dependent pruning is known to be overactive in aged brains, yet its selection criteria remain unclear. Extracellular adenosine rises when neuronal ATP falls (e.g., during hypoxia or metabolic stress) and is a potent modulator of microglial phenotype via A2A receptors, which can stimulate C3 transcription. This provides a plausible molecular bridge: metabolically weak neurons release more adenosine, tagging their synapses for microglial removal.

Mechanistic steps:

1. Aging neurons exhibit reduced mitochondrial oxidative phosphorylation, lowering intracellular ATP.
2. ATP depletion triggers increased extracellular ATP release through pannexin‑1/connexin‑43 hemichannels; extracellular ATP is rapidly hydrolyzed to adenosine by CD39/CD73 ectoenzymatic activity.
3. Elevated adenosine binds A2A receptors on microglia, raising cAMP and activating CREB‑dependent transcription of C3 and C1q.
4. Neuronal synapses with higher adenosine exposure acquire more C3b/iC3b opsonins, engaging microglial CR3 and leading to phagocytosis.
5. Synapses that maintain robust ATP output release less adenosine, escape tagging, and are preserved.

Testable predictions:

• In aged mice, extracellular adenosine levels in the hippocampal interstitial fluid will correlate positively with synaptic C3 deposition and inversely with local ATP markers (e.g., luciferase‑based ATP sensors).
• Genetic or pharmacological reduction of neuronal ATP (e.g., heterozygous knockout of mitochondrial complex I subunit Ndufs4) will increase adenosine release, microglial A2A signaling, and C3‑dependent synapse loss, even in young animals.
• Blocking adenosine production (CD73 knockout) or A2A signaling (A2A receptor antagonist) in aged mice will attenuate synaptic C3 tagging and preserve synapse density without altering amyloid burden.
• Cerebellar resistance to complement pruning will be associated with higher extracellular ATP‑to‑adenosine ratios, possibly due to elevated ecto‑ATPase activity or lower CD73 expression.

Experimental approach:

1. Use in vivo microdialysis coupled with HPLC‑MS to quantify adenosine and ATP in hippocampal CA3 of young (3 mo), aged (18–24 mo), and ATP‑deficient (Ndufs4^+/-) mice.
2. Perform immunofluorescence for C3b/iC3b and synaptic markers (PSD‑95, Synaptophysin) to assess pruning levels.
3. Manipulate adenosine pathways: CD73^−/− mice, A2A antagonist (SCH58261) treatment, or microglia‑specific A2A knockdown via AAV‑Cre in A2A^fl/fl mice.
4. Measure cognitive performance (Morris water maze, novel object recognition) to link synaptic preservation with function.
5. Apply two‑photon imaging of synaptic turnover in Thy1‑YFP mice complemented with a genetically encoded ATP sensor (ATeam) to directly test whether low‑ATP synapses acquire more C3 tags.

Falsifiability: If extracellular adenosine does not rise with neuronal ATP loss, or if blocking A2A signaling fails to reduce C3 deposition and synapse loss despite confirmed target engagement, the hypothesis would be refuted. Conversely, demonstrating that adenosine manipulation rescues synapses without affecting amyloid plaques would support the proposed metabolic‑immune coupling mechanism.

This hypothesis extends the complement‑pruning framework by proposing a specific, activity‑independent metabolic signal (adenosine) that tags inefficient synapses for removal, thereby offering a testable link between cellular energetics and immune‑mediated synapse elimination in aging.