Hypothesis

Alternate-day fasting (ADF) induces periodic ketosis that triggers astrocytic GPR109A signaling, leading to lactate release and activation of matrix metalloproteinase‑9 (MMP‑9). This enzymatic activity degrades perineuronal nets (PNNs) around hippocampal parvalbumin interneurons and shifts the BDNF/proBDNF balance toward mature BDNF, thereby reducing p75^NTR‑mediated RhoA‑ROCK2‑LIMK1 signaling. The net effect is a reversible decrease in synaptic rigidity and a restoration of LTP‑dependent plasticity in the aged brain.

Mechanistic Rationale

Ketone bodies (β‑hydroxybutyrate) act as endogenous inhibitors of class I histone deacetylases (HDACs), increasing transcription of Mmp9 in astrocytes【6】. Astrocytic lactate, released via GPR109A‑dependent mechanisms, further activates neuronal NMDA receptors, boosting calcium‑dependent calpain activity that cleaves proBDNF to mature BDNF and reduces proBDNF‑driven p75^NTR activation【1】. Simultaneously, fasting‑induced low insulin diminishes Akt‑mediated inhibition of GSK‑3β, which phosphorylates RhoA and attenuates ROCK2‑LIMK1 driven actin stabilization【1】. Together, these shifts favor actin turnover, PNN remodeling, and a plastic state reminiscent of juvenile circuits.

Testable Predictions

1. PNN thickness – Aged mice subjected to 4 weeks of ADF will show a ≥30 % reduction in WFA‑labelled PNN density in the dorsal hippocampus compared with ad libitum fed controls (p<0.01).
2. p75^NTR signaling – Hippocampal p75^NTR protein levels and downstream phospho‑LIMK1 will decrease by ≥25 % in ADF mice, while the mature BDNF/proBDNF ratio will increase by ≥40 %【1】.
3. Electrophysiology – Acute hippocampal slices from ADF mice will recover LTP magnitude to ≥80 % of young adult levels, whereas control aged slices remain ≤50 %【4】.
4. MMP‑9 dependence – Pharmacological inhibition of MMP‑9 (SB‑3CT) during ADF will abolish the PNN thinning and LTP rescue, indicating mechanistic necessity.
5. GPR109A requirement – Mice lacking GPR109A in astrocytes (GFAP‑Cre;Gpr109a^fl/fl) will fail to exhibit PNN remodeling or cognitive improvement under ADF, confirming the astrocyte‑ketone axis.

Falsifiability

If ADF does not produce measurable PNN degradation, does not alter p75^NTR or BDNF/proBDNF ratios, and fails to improve LTP or memory performance, the hypothesis is falsified. Conversely, observation of any of the predicted changes without accompanying behavioral rescue would still support a mechanistic link but would require refinement of the behavioral readout.

Novel Insight

Unlike prior views that frame aging as loss of metabolic vigor, this hypothesis posits that the aged brain’s rigidity stems from a static metabolic state that suppresses extracellular matrix turnover. Introducing controlled metabolic oscillations—specifically, the intermittent rise of ketone bodies—re‑engages proteolytic and signaling pathways that actively dismantle the structural substrates of over‑consolidation, turning a supposedly degenerative process into a tunable, reversible state.