IFN-I–Dependent Neuronal ISG Tagging Converts Metabolic Inefficiency into a Microglial 'Eat Me' Signal in Aging

Hypothesis Chronic low‑grade type I interferon (IFN‑I) signaling in the aging brain directly tags metabolically inefficient neurons for microglial phagocytosis by upregulating neuronal interferon‑stimulated genes (ISGs) that expose complement‑like ‘eat me’ motifs on the soma, while simultaneously downregulating protective ‘don’t eat me’ signals such as CD47. This converts intrinsic metabolic stress into an extracellular cue that drives complement‑dependent, microglia‑mediated neuronal eviction, mirroring the activity‑based synaptic pruning pathway.

Mechanistic Rationale

1. Neuronal ISG induction – Neurons express ISGs in interferonopathic contexts [4]; tonic IFN‑I (e.g., from STING activation) can induce neuronal expression of ISG15, ISG56 (MX1), and the atypical complement component C1q‑like protein 4 (C1QL4). Preliminary data show ISG15 conjugation to lysosomal‑associated membrane proteins increases surface exposure of phosphatidylserine (PS) under metabolic stress.
2. Metabolic stress sensor – Neurons with high ATP cost per spike (e.g., fast‑spiking interneurons) exhibit elevated mitochondrial ROS and reduced NAD⁺/NADH ratios, which activate the cytosolic DNA‑sensing cGAS‑STING pathway, amplifying local IFN‑β production. This creates a feed‑forward loop where inefficient neurons generate more IFN‑I, increasing their own ISG load.
3. Conversion to phagocytic signal – ISG15‑ylated PS and neuronal C1QL4 can bind microglial complement receptors (CR3/CR4) and TREM2, respectively, mirroring the synaptic C3‑CR3 axis [1][3]. Concurrently, IFN‑I suppresses neuronal CD47 transcription via STAT1‑mediated repression, diminishing the ‘don’t eat me’ brake.
4. Microglial execution – Activated microglia, already primed by tonic IFN‑I to a phagocytic phenotype [2], engulf ISG‑tagged somata via complement‑dependent phagocytosis, leading to selective neuronal loss that correlates with metabolic inefficiency rather than random damage.

Testable Predictions

• In aged mice, neurons with the highest basal ATP consumption per action potential (measured via in vivo two‑photon calcium imaging combined with fluorescent ATP sensors) will show greater neuronal ISG15‑PS colocalization and lower CD47 surface levels than low‑firing neurons.
• Genetic or pharmacological blockade of IFNAR specifically in neurons (Nestin‑Cre;Ifnar1^fl/fl) will reduce neuronal ISG15‑PS exposure and preserve high‑firing neuron numbers despite unchanged systemic IFN‑I levels.
• Conversely, neuronal overexpression of a non‑cleavable ISG15‑PS fusion protein will accelerate microglial engraftment of those neurons in young mice, an effect blocked by TREM2 or CR3 inhibition.
• Electrophysiological recordings from brain slices of aged Ifnar1^−/− mice will reveal preserved firing rates and reduced paired‑pulse depression in high‑energy‑demanding circuits compared with wild‑type littermates.

Experimental Approach

1. Metabolic profiling – Use Seahorse assays on FACS‑sorted NeuN^+ neurons from young (3 mo) and aged (24 mo) mice to quantify basal and maximal respiration, ATP production, and proton leak. Correlate these metrics with intracellular ISG15‑PS staining (flow cytometry) and CD47 levels.
2. Loss‑of‑function – Generate neuron‑specific Ifnar1 knockout mice; assess neuronal density in cortex and hippocampus via unbiased stereology at 24 mo, with and without systemic IFN‑β neutralization.
3. Gain‑of‑function – Deliver AAV‑Syn‑ISG15‑PS‑mCherry to the motor cortex of 3‑mo mice; monitor microglial engulfment via live‑imaging of CX3CR1^GFP microglia and quantify neuron loss over 8 weeks.
4. Functional readout – Perform in‑vivo patch‑clamp or Neuropixels recordings to assess whether preserved neurons in Ifnar1^−/− animals maintain higher firing fidelity and lower spike‑failure rates during sustained sensory stimulation.

Falsifiability If neuronal ISG15‑PS upregulation does not correlate with metabolic inefficiency, or if neuronal IFNAR blockade fails to rescue high‑firing neuron loss despite reducing ISG signals, the hypothesis would be refuted. Likewise, if blocking microglial TREM2 or CR3 does not prevent ISG15‑mediated neuronal engulfment, the proposed phagocytic link would be unsupported.

Implications This reframes age‑related neuronal loss not as passive decay but as an IFN‑I‑driven quality‑control mechanism that trims energetically costly circuits when systemic resources decline. It suggests that modulating neuronal ISG expression—or the IFN‑I‑microglia axis—could preserve functional networks in neurodegeneration without globally suppressing beneficial immune surveillance.