Hypothesis

The aging brain does not passively lose neurons to glucocorticoid toxicity; instead, an elevated cortisol/DHEA ratio aberrantly reactivates the complement-mediated synaptic pruning cascade that normally shapes developing circuits. This reactivation tags metabolically expensive, weakly active neurons for removal, mimicking the 'eviction for inefficiency' described in the seed idea. Unlike developmental pruning, this process is driven by endocrine imbalance and results in maladaptive loss of hippocampal neurons that accelerates cognitive decline.

Mechanistic Rationale

1. Endocrine trigger – With age, zona reticularis atrophy cuts DHEAS production (1‑2% per year) while cortisol rises 20‑50%, raising the cortisol/DHEA ratio [https://pmc.ncbi.nlm.nih.gov/articles/PMC12357812/]. High cortisol reduces hippocampal glucocorticoid receptor feedback, prolonging exposure [https://www.frontiersin.org/journals/endocrinology/articles/10.3389/fendo.2019.00054/full]. Low DHEAS removes its neuroprotective actions on LTP and glial survival [https://pmc.ncbi.nlm.nih.gov/articles/PMC2829637/].

2. Complement re‑activation – Chronic glucocorticoid excess and DHEA deficit up‑regulate astrocytic C1q and microglial C3 expression, a pattern seen in developmental synaptic pruning and in Alzheimer models [https://pmc.ncbi.nlm.nih.gov/articles/PMC5621604/]. The inflammatory milieu from adrenal senescence (elevated IL‑6, TNF‑α) further primes microglia [https://pmc.ncbi.nlm.nih.gov/articles/PMC12313722/].

3. Selective targeting – Neurons with high basal firing rates consume more ATP; under limited energy supply (worsened by mitochondrial aging), they generate higher lactate and oxidative stress, exposing phosphatidylserine and “eat‑me” signals. Complement opsonization (C3b) preferentially marks these energetically costly, weakly connected synapses, steering microglia to prune the parent neuron [https://pmc.ncbi.nlm.nih.gov/articles/PMC12357812/]. This mirrors developmental logic where under‑active, high‑cost circuits are eliminated.

4. Outcome – Selective removal of these neurons reduces overall metabolic load, creating the illusion of increased efficiency per watt, yet it erodes memory encoding capacity and HPA‑axis regulation, producing the cognitive and hormonal phenotypes of aging.

Testable Predictions

• Prediction 1: In aged mice, hippocampal C3 deposition will colocalize with neurons expressing high metabolic markers (e.g., GLUT3, cytochrome c oxidase) and low activity markers (c‑Fos−). Blocking C3 with antibodies or genetic knock‑out will prevent the loss of these specific neurons without altering total neuronal counts.
• Prediction 2: Restoring DHEAS to youthful levels in aged animals will lower the cortisol/DHEA ratio, decrease astrocytic C1q/microglial C3 up‑regulation, and preserve the high‑metabolic neuronal subset.
• Prediction 3: Humans with elevated cortisol/DHEA ratios (top quartile) will show greater hippocampal C3‑positive puncta on post‑mortem tissue, correlated with selective atrophy of CA3 pyramidal neurons identified by metabolic gene expression profiles.

Falsification

If experimental elevation of cortisol/DHEA ratio fails to increase complement tagging or selective neuronal loss, or if complement inhibition does not rescue the metabolic‑neuron subset, the hypothesis would be refuted, supporting the view that neuronal loss in aging is purely passive glucocorticoid toxicity.

Implications

Reframing age‑related hippocampal shrinkage as an endocrine‑driven pruning process shifts therapeutic focus from generic neuroprotection to modulating the cortisol/DHEA axis and complement signaling. Interventions that lower the ratio or block complement could preserve energetically demanding neurons, maintaining cognitive resilience despite an aging brain.