Hypothesis

Aging-associated expansion of TNF‑producing age‑associated B cells (ABCs) primes microglia to phagocytose neurons that exhibit low metabolic activity, thereby coupling peripheral immune signals to central synaptic and neuronal pruning. This mechanism provides a testable basis for the “eviction for inefficiency” concept: neurons are not randomly lost but are selectively removed when their energetic output falls below a threshold set by TNF‑sensitized microglial phagocytic receptors.

Mechanistic Basis

• ABCs accumulate with age and secrete TNF‑α, which upregulates microglial expression of complement receptors (CR3) and scavenger receptors that recognize neuronal “eat‑me” signals such as phosphatidylserine exposed by depolarized mitochondria (TNF‑α upregulates microglial complement receptors).
• Neurons with reduced oxidative phosphorylation release less ATP and more mitochondrial reactive oxygen species, leading to externalization of phosphatidylserine and decreased CD47 “don’t‑eat‑me” signaling (mitochondrial ROS phosphatidylserine exposure).
• In aged female microglia, estrogen‑dependent downregulation of TNF‑receptor‑associated signaling blunts the ability to adjust phagocytosis in response to IFN‑γ, creating a sex‑biased shift toward constitutive TNF‑driven phagocytosis (estrogen‑TNF‑receptor signaling in female microglia) (B cell somatic hypermutation sex differences).
• Thus, when TNF levels are high, microglia become primed to clear any neuron that fails to maintain sufficient metabolic vigor, irrespective of actual damage.

Predictions & Experimental Tests

1. Correlation – In aged mice, neuronal loss in cortical layers should inversely correlate with local mitochondrial activity measured by Seahorse assay or NADH fluorescence, and this correlation should be stronger in females than males.
2. TNF blockade – Genetic deletion of TNF in ABCs or systemic anti‑TNF treatment will reduce age‑related neuronal loss without affecting microglial baseline surveillance, and will rescue the sex‑specific difference in phagocytic adaptation to IFN‑γ.
3. ABC depletion – Targeted ablation of ABCs (e.g., using CD11c‑DTR mice) will lower microglial TNF priming and preserve neuronal numbers, especially in females, while leaving pathogen clearance intact.
4. Neuronal activity rescue – Chemogenetic activation of hypoactive neurons (hM3Dq DREADD) should decrease their phosphatidylserine exposure and protect them from microglial phagocytosis even in the presence of high TNF.
5. Phagocytic specificity – Aged microglia from TNF‑blocked animals will show normal uptake of apoptotic beads but reduced uptake of live neurons with low mitochondrial membrane potential, demonstrating a shift from indiscriminate to activity‑dependent clearance.

Potential Confounds & Alternatives

• Increased microglial phagocytosis could reflect a general response to heightened neuronal death from other causes; therefore, experiments must distinguish cause (low activity) from consequence (cell death) using temporal labeling (e.g., birth‑dating with EdU) and real‑time imaging of microglial‑neuron contacts.
• Complement cascade activation (C1q, C3) also tags synapses for pruning; blocking complement should not affect the predicted neuronal loss if the primary driver is TNF‑mediated phagocytosis of whole cells, allowing dissection of parallel pathways.

By linking peripheral ABC‑derived TNF to microglial metabolic sensing of neurons, this hypothesis turns the “eviction for inefficiency” idea into a concrete, falsifiable pathway that can be manipulated genetically, pharmacologically, or physiologically.