Hypothesis

Chronic activation of the JNK/AP-1 pathway in aging glia promotes a state of synaptic over-consolidation that manifests as impaired depotentiation and resistance to long-term depression (LTD), rather than outright synapse loss. This mechanistic shift explains why aged brains appear 'rigid': they have become overly confident in existing predictive models because plasticity is biased toward stabilization.

Mechanistic Reasoning

1. Glial JNK/AP-1 → extracellular matrix remodeling – Sustained c-Jun phosphorylation in astrocytes and microglia up-regulates genes for hyaluronan synthases and tenascins, leading to thicker perineuronal nets (PNNs) that physically constrain spine motility. PNN enrichment correlates with reduced LTD in hippocampal slices [https://www.frontiersin.org/journals/aging-neuroscience/articles/10.3389/fnagi.2024.1453710/full].
2. Direct neuronal effect – JNK phosphorylates the AMPA-receptor subunit GluA1 at Ser845, increasing its open probability and surface stability. This biases AMPA-receptor trafficking toward insertion and away from activity-dependent removal, weakening the cellular machinery for LTD [https://pmc.ncbi.nlm.nih.gov/articles/PMC8070500/].
3. Feedback loop – Stabilized synapses generate less calcium influx during low-frequency stimulation, further diminishing calcineurin-dependent dephosphorylation pathways that normally counteract JNK signaling, creating a self-reinforcing over-consolidation loop.

Testable Predictions

• Prediction 1: In aged mice, hippocampal slices will show (a) normal or slightly increased baseline LTP magnitude, (b) significantly impaired LTD and depotentiation after high-frequency stimulation, and (c) elevated PNN density and phospho-c-Jun colocalization with GFAP/Iba1 compared with young controls.
• Prediction 2: Pharmacological or genetic inhibition of JNK (e.g., SP600125 or astrocyte-specific c-Jun knockout) in aged animals will restore LTD/depotentiation to youthful levels without altering total synapse number (measured by synaptophysin or VGLUT1 puncta).
• Prediction 3: Behavioral assays that require updating of learned associations (e.g., reversal learning or surprise-based novelty preference) will improve following JNK inhibition, whereas simple acquisition tasks remain unchanged.

Experimental Design (brief)

1. Subjects: Young (3 mo) and aged (20 mo) C57BL/6J mice; subgroups receive JNK inhibitor or vehicle.
2. Electrophysiology: Field recordings in CA1 to quantify LTP, LTD, and depotentiation; paired-pulse ratio to rule out presynaptic changes.
3. Immunohistochemistry: Staining for phospho-c-Jun, GFAP, Iba1, WFA (PNN marker), and synaptophysin; quantify co-localization and net thickness.
4. Behavior: Morris water maze reversal, auditory fear conditioning extinction, and surprise-oddball task.
5. Analysis: Two-way ANOVA (age × treatment) with post-hoc tests; correlation of PNN thickness with LTD magnitude.

Falsifiability

If JNK inhibition fails to rescue LTD/depotentiation or does not reduce PNN accumulation, or if synapse loss accompanies any rescue, the hypothesis that JNK/AP-1 drives over-consolidation rather than degeneration would be refuted. Conversely, a rescue of bidirectional plasticity without synapse loss would support the over-consolidation model.

Broader Impact

Reframing cognitive aging as a plasticity-bias problem shifts therapeutic goals from neuroprotection to dynamic re-tuning of excitatory/inhibitory balance via glial signaling pathways.