Hypothesis

The aging brain does not merely prune weak synapses; it actively evicts whole neurons whose metabolic output falls below a threshold sensed by microglia via lactate sensing. Neurons that sustain balanced excitatory/inhibitory activity produce sufficient lactate to activate the microglial lactate receptor HCAR1 (also known as GPR81), which suppresses phagocytic signaling. When activity‑driven lactate production declines, HCAR1 signaling wanes, exposing phosphatidylserine and triggering microglial phagocytosis. This links the energy‑budget argument directly to a metabolite‑dependent checkpoint, extending the inefficiency‑based eviction model.

Mechanistic Model

1. Activity‑dependent lactate production – Balanced neuronal firing drives aerobic glycolysis, exporting lactate via monocarboxylate transporters (MCT1/4).
2. Microglial HCAR1 signaling – Lactate binds HCAR1 on microglia, raising intracellular cAMP and inhibiting the NADPH‑oxidase–ROS pathway that promotes phagocytosis.
3. Energy‑sensing checkpoint – In aging, mitochondrial efficiency declines, reducing lactate output despite preserved firing rates; the lactate signal falls below the HCAR1 activation threshold.
4. Phosphatidylserine exposure – Low HCAR1 signaling disinhibits TAM receptor pathways, leading to phosphatidylserine externalization on the neuronal surface.
5. Microglial phagocytosis – PS‑recognizing receptors (TIM4, BAI1) engage, leading to neuronal removal.

This model explains why interventions that restore activity balance (e.g., GABA modulation) can rescue neurons: they restore lactate production, re‑engage HCAR1, and block the eat‑me signal.

Testable Predictions

• Prediction 1: In aged mice, extracellular lactate concentration in the hippocampal interstitial fluid will correlate negatively with neuronal loss; lactate supplementation will reduce age‑dependent neuronal eviction without altering developmental pruning.
  • Experiment: Microdialysis to measure lactate; chronic lactate infusion via osmotic pump; stereological neuron counts.
• Prediction 2: Microglial-specific knockout of Hcar1 will accelerate age‑related neuronal loss, whereas overexpression will protect neurons even when mitochondrial complex I activity is inhibited.
  • Experiment: Cx3cr1‑CreER;Hcar1^fl/fl mice treated with tamoxifen at middle age; assess neuron numbers and microglial phagocytic markers.
• Prediction 3: Neurons genetically engineered to express a constitutively active lactate dehydrogenase (LDHA) will resist eviction despite hyper‑ or hypo‑activity, uncoupling activity from lactate output.
  • Experiment: AAV‑LDHA expression in excitatory neurons; two‑photon imaging of dendritic spines and caspase‑3 activation over months.
• Prediction 4: Blocking phosphatidylserine exposure with annexin V‑derived peptides will attenuate microglial phagocytosis of lactate‑low neurons, providing a pharmacological avenue.
  • Experiment: Peptide infusion in aged rats; assess microglial PS‑binding and neuron survival.

Implications

If lactate‑HCAR1 signaling constitutes a metabolic checkpoint for neuronal eviction, then aging‑related cognitive decline may be exacerbated not by excess pruning per se, but by a mismatch between neuronal activity and metabolic output. Therapeutic strategies that boost neuronal lactate production (e.g., pyruvate supplementation, MCT4 overexpression) or mimic HCAR1 signaling could preserve neuronal numbers without interfering with the beneficial synaptic pruning that occurs during development.

References

• [1] https://elifesciences.org/articles/72135
• [2] https://pmc.ncbi.nlm.nih.gov/articles/PMC12500284/
• [3] https://doi.org/10.1126/science.286.5440.785
• [4] https://pmc.ncbi.nlm.nih.gov/articles/PMC8003660/