Hypothesis

During sleep, reduced PERK signaling restores eIF2α‑dependent translation of aquaporin‑4 (AQP4) and other glymphatic components, enabling efficient clearance of misfolded proteins that would otherwise sustain PERK‑CHOP activation. In aged tissue, sleep loss prevents this PERK downturn, locking the UPR into a PERK‑dominant state that blocks glymphatic flux and accelerates proteotoxic collapse.

Mechanistic reasoning

1. PERK‑eIF2α axis controls AQP4 synthesis – PERK activation phosphorylates eIF2α, globally attenuating cap‑dependent translation while favoring ATF4‑driven stress genes. AQP4 mRNA is highly dependent on cap‑dependent translation; thus, sustained PERK activity reduces AQP4 protein levels, impairing perivascular water flux and glymphatic clearance (1).
2. Sleep triggers PERK dephosphorylation – Neuronal activity drops during slow‑wave sleep, lowering calcium‑dependent kinase inputs (e.g., PERK‑activating PKR‑like ER kinase) and activating phosphatases such as PP1/GADD34 complexes that dephosphorylate eIF2α (2). This transiently lifts translational repression, allowing AQP4 and other glymphatic effectors to be replenished.
3. Aged brains exhibit blunted phosphatase response – Aging increases basal PP1 inhibitors (e.g., CIP2A) and oxidative modification of phosphatases, dampening eIF2α dephosphorylation after sleep onset (3). Consequently, PERK remains active, AQP4 stays low, and glymphatic flow fails to clear wake‑accumulated misfolded proteins.
4. Feedback loop – Persistent misfolded proteins keep BiP dissociated from PERK, maintaining its activation and further suppressing AQP4, creating a vicious cycle that shifts the UPR balance toward PERK‑CHOP driven apoptosis (4).

Testable predictions

• Prediction 1: In young mice, sleep deprivation will increase p‑eIF2α and decrease AQP4 levels in cortical astrocytes; recovery sleep will reverse both changes. In aged mice, the same deprivation will produce a larger and persistent p‑eIF2α elevation with incomplete AQP4 recovery after sleep.
• Prediction 2: Pharmacological inhibition of PERK (GSK2606414) during sleep deprivation in aged mice will restore AQP4 expression and glymphatic influx of fluorescent tracer (e.g., CSF‑injected Evans blue) to young‑like levels.
• Prediction 3: Genetic overexpression of a constitutively active PP1 catalytic subunit specifically in astrocytes will normalize p‑eIF2α during sleep deprivation in aged mice, rescuing glymphatic clearance and reducing CHOP‑positive cells.

Experimental approach

1. Animal cohorts: Young (3 mo) and aged (20 mo) C57BL/6 mice subjected to 6 h sleep deprivation via gentle handling, followed by 6 h recovery sleep or continued wake.
2. Readouts: Western blot of cortical lysates for p‑eIF2α, total eIF2α, ATF4, CHOP, AQP4, and BiP; immunofluorescence for AQP4 polarization at endfeet; glymphatic function quantified by MRI‑based intrathecal CSF tracer clearance.
3. Interventions: Subcutaneous GSK2606414 (PERK inhibitor) or AAV‑mediated astrocyte‑specific PP1α overexpression administered prior to deprivation.
4. Analysis: Two‑way ANOVA (age × treatment) with post‑hoc tests; significance set at p<0.05.

If the data show that aged mice fail to lower p‑eIF2α and restore AQP4 after sleep, and that PERK inhibition or PP1 activation rescues glymphatic clearance, the hypothesis will be supported. Conversely, if sleep recovery normalizes p‑eIF2α and AQP4 equally in young and aged animals, the hypothesis would be falsified, indicating that PERK‑eIF2α dysregulation is not the primary sleep‑dependent gate of glymphatic function in aging.