Berberine’s weaker activation of AMPK relative to metformin places neuronal autophagy in a ‘selective‑clearance’ mode that removes damaged mitochondria and protein aggregates while sparing synaptic components, thereby opposing the age‑related eviction of weakly connected neurons. In contrast, stronger AMPK activation drives excessive autophagic flux that leads to synapse loss and network dysfunction.

Mechanistic Rationale

AMPK activates the ULK1 complex to initiate autophagy, but the downstream outcome depends on the balance between ULK1‑driven phagophore formation and mTORC1‑mediated inhibition of lysosomal maturation. Berberine, by only partially elevating AMPK activity, may sustain enough ULK1 phosphorylation to initiate autophagosome formation yet fail to fully suppress mTORC1 in neurons with high energetic demand. This creates a bottleneck where autophagosomes accumulate but lysosomal degradation is rate‑limited, favoring the turnover of long‑lived, damaged organelles (e.g., via mitophagy) over short‑lived synaptic proteins that require rapid recycling. Additionally, berberine’s ability to inhibit PCSK9 in hepatocytes reduces circulating LDL‑derived oxidative lipids, lowering peripheral inflammatory cues that would otherwise amplify microglial complement signaling—a key driver of synaptic pruning.

Predictions

1. In aged mice, berberine treatment will increase LC3‑II puncta (indicating autophagosome formation) without a proportional rise in lysosomal cathepsin activity, whereas metformin will elevate both markers.
2. Synaptic protein levels (synaptophysin, PSD‑95) will be preserved or slightly increased with berberine but reduced with metformin despite similar reductions in neuroinflammatory cytokines (IL‑1β, TNF‑α).
3. Microglial phagocytic activity, measured by colocalization of Iba1 with synaptic markers, will be significantly lower in berberine‑treated brains compared with metformin‑treated or untreated aged controls.
4. Mitochondrial health assays (MitoTracker red/green ratio, ROS production) will show improved mitophagy flux with berberine, reflected by increased PINK1/Parkin colocalization with mitochondria.

Experimental Design

• Animals: 20‑month‑old C57BL/6J mice, n=10 per group (vehicle, berberine 5 mg/kg/day, metformin 100 mg/kg/day) for 12 weeks.
• Readouts: Western blot and immunofluorescence for LC3‑II, p62, LAMP1, cathepsin B; synaptic proteins; mitochondrial markers; cytokine ELISA; microglial phagocytosis assay using synaptosome‑bound pHrodo.
• Functional tests: Rotarod, Y‑maze, and in vivo two‑photon imaging of dendritic spine density in motor cortex.
• Statistical plan: Two‑way ANOVA with treatment and time as factors; post‑hoc Tukey; significance set at p<0.05.

Potential Implications

If confirmed, this hypothesis reframes berberine not merely as a weaker metformin analog but as a fine‑tuner of neuronal quality control that aligns with the brain’s energy‑saving pruning program during aging. It would suggest that therapeutic strategies aiming to blunt AMPK activation—rather than maximize it—could protect synaptic integrity while still mitigating oxidative stress and inflammation. Such an approach may be especially relevant for neurodegenerative conditions where synapse loss precedes overt neuronal death.