Hypothesis

The aged brain actively evicts neurons whose metabolic inefficiency is amplified by proximity to senescent glial cells, coupling local SASP‑induced metabolic stress to complement‑tagging and JAK2‑dependent synaptic stripping, ultimately leading to neuronal removal.

Mechanistic Model

1. Senescent glial niche formation – With age, microglia and astrocytes acquire a senescent‑like phenotype, upregulating p16 and secreting SASP factors (IL‑1β, TNF‑α, C3) 4. These factors alter the extracellular microenvironment, reducing mitochondrial oxidative phosphorylation in nearby neurons.

2. Metabolic inefficiency sensor – Neurons within ~50 µm of SASP‑rich niches show decreased ATP/ADP ratios and elevated AMP‑activated protein kinase (AMPK) activity, flagging them as low‑efficiency units. This metabolic signature can be captured by spatial transcriptomics of genes such as Atp5a1, Ucp2, and Ppargc1a.

3. Complement tagging – SASP‑induced C1q and C3 deposition preferentially labels metabolically stressed neurons. The complement cascade converges on the neuronal soma, marking them for microglial phagocytosis, mirroring developmental synaptic pruning but acting on whole cells 2.

4. JAK2‑mediated synaptic competition – Prior to somatic elimination, inefficient neurons exhibit reduced synaptic activity, lowering JAK2‑STAT3 signaling in afferent axons. According to activity‑dependent competition models, these weak synapses are eliminated via JAK2 signaling, further isolating the neuron 2.

5. Feedback loop – Removed neurons release DAMPs that reinforce local senescence, amplifying the niche and creating a spatially propagating wave of eviction.

Testable Predictions

• Prediction 1: In aged mouse brain, neurons expressing low Atp5a1/Ucp2 will be significantly closer (p<0.01, permutation test) to p16^+ microglia than high‑expressing neurons, measurable by multiplexed ion beam imaging (MIBI) or CODEX.
• Prediction 2: Pharmacological blockade of complement C3 (using Cp40) will reduce the elimination of low‑efficiency neurons without affecting overall neuronal number, assessed by longitudinal two‑photon imaging of Thy1‑YFP mice.
• Prediction 3: Chemogenetic suppression of JAK2 in excitatory neurons will preserve synapses of metabolically inefficient cells, delaying their somatic removal, evident as preserved dendritic spine density despite low ATP readings.
• Prediction 4: Inducing senescence locally via p16‑INKB‑mediated expression in astrocytes will expand the radius of metabolic inefficiency tagging by ~30 µm, visualized by a FRET‑based ATP sensor.

Falsification

If spatial mapping shows no correlation between neuronal metabolic gene expression and distance to senescent cells, or if complement inhibition fails to alter neuronal loss rates, the hypothesis would be refuted, suggesting that neuronal eviction is driven primarily by cell‑intrinsic damage rather than microenvironmental efficiency sensing.

Broader Implications

Reframing age‑related neuronal loss as an adaptive, energetically driven process shifts therapeutic goals from preventing death to modulating the sensory‑metabolic niche—potentially preserving cognitive function by enhancing neuronal efficiency rather than blocking cell death.

References

[1] https://pmc.ncbi.nlm.nih.gov/articles/PMC11798877/ [2] https://pmc.ncbi.nlm.nih.gov/articles/PMC8068677/ [3] https://www.news-medical.net/news/20250213/Brain-aging-linked-to-neuronal-hyperactivation-not-decline-study-finds.aspx [4] https://doi.org/10.1101/2024.05.23.595605