Hypothesis

Aging-related decline in base excision repair (BER) leads to progressive accumulation of 8‑oxoguanine (8-oxoG) in neurons. Rather than causing immediate cell death, this oxidative lesion triggers a retrograde signaling cascade that reduces neuronal excitability and mitochondrial output. Low‑activity, high‑damage neurons are then recognized by microglia via cGAS‑STING‑dependent sensing of extracellular mtDNA and are preferentially phagocytosed. This couples BER insufficiency to an activity‑dependent pruning mechanism that optimizes cortical networks under falling energy budgets, preserving overall circuit efficiency while allowing damaged cells to be removed before they become sources of chronic inflammation.

Mechanistic Rationale

1. BER loss → 8-oxoG ↑ → transcriptional repression of activity‑genes

   • 8-oxoG in promoter regions can impair binding of CREB and NRF2, decreasing expression of activity‑dependent genes (e.g., Fos, Arc, Pgc‑1α).
   • Reduced CREB signaling lowers BDNF transcription, diminishing synaptic strength and firing rates.

2. Hypo‑active neurons release damaged mtDNA

   • Mitochondrial stress from unrepaired oxidative bases increases mtDNA fragments that escape into the cytosol.
   • Cytosolic mtDNA activates cGAS‑STING, driving type‑I interferon production and the release of chemokines (CCL2, CXCL10) that attract microglia.

3. Microglial phagocytosis thresholds set by neuronal activity

   • Microglia express purinergic receptors (P2Y12) that detect ATP release; active neurons secrete more ATP, raising the phagocytic threshold.
   • When activity falls below a set point (due to 8-oxoG‑mediated transcriptional repression), the same microglial surveillance interprets the neuron as “low‑value” and initiates engulfment via complement tagging (C1q, C3).

4. Functional outcome: network sharpening

   • Removal of the least active, most damaged nodes reduces noisy signaling and lowers energetic demand, yielding a higher signal‑to‑noise ratio per watt consumed—consistent with the observation that aging brains can maintain or even improve certain cognitive metrics despite cell loss.

Testable Predictions

• Prediction 1: In aged mice, neurons with high 8-oxoG levels will show reduced c‑Fos immunoreactivity and lower spontaneous firing rates compared with low‑8-oxoG counterparts.
• Prediction 2: Pharmacological elevation of neuronal activity (e.g., chemogenetic hM3Dq activation) in 8-oxoG‑rich neurons will decrease their likelihood of microglial engraftment, without altering global 8-oxoG burden.
• Prediction 3: Neuron‑specific overexpression of OGG1 or APE1 in aged mice will preserve activity‑dependent gene expression, sustain ATP release, and reduce complement‑mediated microglial phagocytosis despite age‑related BER decline.
• Prediction 4: Blocking cGAS‑STING signaling (using H‑151 or STING‑KO) will uncouple 8-oxoG accumulation from microglial phagocytosis, leading to accumulation of hyper‑damaged but still electrically silent neurons and increased cortical inflammation.

Experimental Approach

1. In vivo imaging & electrophysiology

   • Use Thy1‑GCaMP6f mice crossed with a reporter for 8-oxoG (e.g., GFP‑OGG1 fusion that fluoresces upon binding). Perform two‑photon calcium imaging in layer 2/3 somatosensory cortex of young (3 mo) vs aged (24 mo) mice to correlate 8-oxoG signal with event‑rate and calcium transient amplitude.

2. Manipulating neuronal activity

   • Inject AAV‑DIO‑hM3Dq‑mCherry into excitatory neurons of aged OGG1‑heterozygous mice; administer CNO to elevate firing for 2 weeks. Assess microglial proximity (Iba1 staining) and phagocytic cups (CD68+ / NeuN+ overlaps) via confocal microscopy.

3. Genetic rescue of BER

   • Generate Camk2a‑Cre; OGG1^fl/fl; APE1^fl/fl mice with inducible overexpression vectors. Measure 8-oxoG (dot‑blot), c‑Fos, synaptic protein levels (PSD‑95, Synaptophysin), and microglial complement deposition (C1q, C3).

4. Pharmacological STING inhibition

   • Treat aged wild‑type mice with H‑151 for 4 weeks; quantify neuronal loss (NeuN+ density), inflammatory cytokines (IL‑1β, TNF‑α), and behavioral performance on rotor‑rod and novel‑object recognition.

Potential Implications

If validated, this hypothesis reframes age‑related neuronal loss as a synergy between metabolic constraint and quality‑control surveillance, suggesting that therapeutic strategies aimed solely at boosting DNA repair may be insufficient without concurrently modulating neuronal activity or microglial activation states. It also opens avenues for detecting early network “pruning” via activity‑dependent biomarkers (e.g., lowered resting‑state fMRI variance combined with peripheral mtDNA levels) before irreversible cognitive decline emerges.