Data from: Selective removal of astrocytic PERK protects against glymphatic impairment and decreases toxic aggregation of β-amyloid and tau
Data files
Dec 05, 2025 version files 150.29 KB
-
README.md
6.84 KB
-
SourceData_Fig_1.xlsx
11.73 KB
-
SourceData_Fig_2.xlsx
20.92 KB
-
SourceData_Fig_3.xlsx
27.09 KB
-
SourceData_Fig_4.xlsx
18.90 KB
-
SourceData_Fig_5.xlsx
19.58 KB
-
SourceData_Fig_6.xlsx
22.62 KB
-
SourceData_Fig_7.xlsx
22.60 KB
Abstract
Dysfunction of the glymphatic system, a brain-wide waste clearance network, is strongly linked to Alzheimer’s disease (AD) and the accumulation of β-amyloid (Aβ) and tau proteins. Here, we identify an astrocytic signaling pathway that can be targeted to preserve glymphatic function and mitigate neurotoxic protein buildup. Analysis of astrocytes from both human AD brains and two transgenic mouse models (5XFAD and PS19) reveals robust activation of the protein kinase R-like endoplasmic reticulum (ER) kinase (PERK)-α subunit of eukaryotic initiation factor 2 (eIF2α) branch of the unfolded protein response. Chronic PERK activation suppresses astrocytic protein synthesis and, through casein kinase 2 (CK2)-dependent mechanisms, disrupts the perivascular localization of aquaporin-4 (AQP4), a water channel essential for glymphatic flow. Importantly, astrocyte-specific PERK deletion or pharmacological inhibition restores AQP4 localization, enhances glymphatic clearance, reduces Aβ and tau pathology, and improves cognitive performance in mice. These findings highlight the critical role of the astrocytic PERK-CK2-AQP4 axis in glymphatic dysfunction and AD pathogenesis, positioning this pathway as a promising therapeutic target.
Dataset DOI: 10.5061/dryad.gb5mkkx43
Description of the data and file structure
This dataset contains the individual source data supporting figures and analyses in the associated article (doi: 10.1016/j.neuron.2025.04.027).
Files and variables
Description: Source data for Figure 1, panels C–F.
Sheet Fig. 1C: quantification of Eif2ak3 puncta number per S100β+ astrocyte in the cortex of 6-month-old WT and 5XFAD mice (n = 5 mice per group).
Sheet Fig. 1D: quantification of indicated proteins (fold change) in cortical astrocytes isolated from WT and 5XFAD mice (n = 5 mice per group).
Sheet Fig. 1E: quantification of p-PERK fluorescence intensity (fold change) in GFAP+ astrocytes from WT and 5XFAD mice (n = 5 mice per group).
Sheet Fig. 1F: quantification of p-PERK fluorescence intensity (fold change) in GFAP+ astrocytes from control and AD subjects (n = 5 samples per group).
Description: Source data for Figure 2, panels B, C, D, F–H.
Sheet Fig. 2B: Western blot analysis of puromycin-labeled proteins (fold change) in isolated cortical astrocytes from WT, cKO, 5XFAD, and 5XFAD;cKO mice.
Sheet Fig. 2C: quantification of perivascular AQP4 expression (arbitrary units) and AQP4 polarization.
Sheet Fig. 2D: Western blot analysis of AQP4 protein levels (fold change) in isolated cerebral vessels from WT, cKO, 5XFAD, and 5XFAD;cKO mice.
Sheet Fig. 2F: quantification of average fluorescence intensity (arbitrary units) in periarteriolar spaces in WT, cKO, 5XFAD, and 5XFAD;cKO mice.
Sheet Fig. 2G: quantification of fluorescence intensity (arbitrary units) in the somatosensory cortex of WT, cKO, 5XFAD, and 5XFAD;cKO mice.
Sheet Fig. 2H: quantification of tracer fluorescence intensity (arbitrary units) in dCLNs of WT, cKO, 5XFAD, and 5XFAD;cKO mice.
Description: Source data for Figure 3, panels B–E, G, and I.
Sheet Fig. 3B: ELISA quantification of Aβ40 and Aβ42 levels (ng/mL) in cortical lysates (n = 6 mice per group), measured in TBS, TBSX, and GDN fractions.
Sheet Fig. 3C: ELISA quantification of Aβ40 and Aβ42 levels (ng/mL) in hippocampal lysates (n = 6 mice per group), measured in TBS, TBSX, and GND fractions.
Sheet Fig. 3D: quantification of Aβ burden (fold change) in the cortex and hippocampus of 5XFAD and 5XFAD;cKO mice.
Sheet Fig. 3E: quantification of Aβ plaque load (fold change) in the cortex and hippocampus of 5XFAD and 5XFAD;cKO mice.
Sheet Fig. 3G: escape latency (s) across four training days (days 1–4) in the Barnes maze and percentage of time spent in the target quadrant during the probe test (day 5).
Sheet Fig. 3I: fear response (%) on day 0, day 1, and day 7 in WT and 5XFAD mice with or without astrocytic PERK deletion.
Description: Source data for Figure 4, panels B–F.
Sheet Fig. 4B: Western blot analysis of puromycin-labeled proteins (fold change) in isolated cortical astrocytes from vehicle- or GSK-treated 5XFAD mice (n = 6 mice per group).
Sheet Fig. 4C: quantification of perivascular AQP4 expression (arbitrary units) and AQP4 polarization in GSK- versus vehicle-treated 5XFAD mice (n = 6 mice per group).
Sheet Fig. 4D: quantification of fluorescence intensity (arbitrary units) in periarteriolar spaces in 5XFAD mice treated with vehicle or GSK.
Sheet Fig. 4E: quantification of Aβ burden (fold change) in the cortex and hippocampus of 5XFAD mice treated with vehicle or GSK.
Sheet Fig. 4F: freezing response (%) on day 1 and day 7 post-conditioning in GSK- or vehicle-treated 5XFAD mice.
Description: Source data for Figure 5, panels A, C–G.
Sheet Fig. 5A: Western blot analysis of indicated proteins (fold change) in cortical astrocytes isolated from 8-month-old WT and PS19 mice (n = 5 mice per group).
Sheet Fig. 5C: Western blot analysis of puromycin-labeled proteins (fold change) in cortical astrocytes isolated from WT, PS19, and PERK-deleted PS19 mice (n = 5 mice per group).
Sheet Fig. 5D: quantification of perivascular AQP4 expression (arbitrary units) and AQP4 polarization.
Sheet Fig. 5E: Western blot analysis of AQP4 protein levels (fold change) in cerebral vessels isolated from WT, PS19, and PERK-deleted PS19 mice (n = 5 mice per group).
Sheet Fig. 5F: quantification of tracer fluorescence intensity (arbitrary units) in periarteriolar spaces.
Sheet Fig. 5G: quantification of tracer fluorescence intensity (arbitrary units) in the somatosensory cortex.
Description: Source data for Figure 6, panels A–F.
Sheet Fig. 6A: quantification of AT8-positive area (fold change) in the piriform cortex and hippocampus.
Sheet Fig. 6B: quantification of MC1-positive area (fold change) in the piriform cortex and hippocampus.
Sheet Fig. 6C: quantification of ThioS-positive area (fold change).
Sheet Fig. 6D: ELISA analysis of human tau (ng/mL) in cortical lysates from PS19 mice with or without astrocytic PERK deletion (n = 6 mice per group), measured in RAB, RIPA, and formic acid fractions.
Sheet Fig. 6E: escape latency (s) across four training days (days 1–4) in the Barnes maze and percentage of time spent in the target quadrant during the probe test (day 5).
Sheet Fig. 6F: freezing response (%) on day 1 and day 7 post-conditioning in WT, PS19, and PS19;cKO mice.
Description: Source data for Figure 7, panels A–C, E, and G–J.
Sheet Fig. 7A: quantification of CK2α immunoreactivity (fold change) in cortical astrocytes of WT, 5XFAD, and 5XFAD;cKO mice.
Sheet Fig. 7B: quantification of CK2α immunoreactivity (fold change) in cortical astrocytes of WT, PS19, and PS19;cKO mice.
Sheet Fig. 7C: quantification of CK2α immunoreactivity (fold change) in GFAP-positive astrocytes from control and AD subjects.
Sheet Fig. 7E: quantification of perivascular AQP4 expression (arbitrary units) and polarization following vehicle or CK2 administration.
Sheet Fig. 7G: quantification of perivascular AQP4 expression (arbitrary units) and polarization in vehicle- versus CX-4945-treated 5XFAD mice.
Sheet Fig. 7H: quantification of FITC-dextran tracer intensity in periarteriolar spaces following intracisternal injection of vehicle or CX-4945.
Sheet Fig. 7I: quantification of Aβ plaque burden in the cortex and hippocampus of vehicle- and CX-4945-treated 5XFAD mice (n = 6 mice per group).
Sheet Fig. 7J: freezing response (%) on day 1 and day 7 post-conditioning in vehicle- and CX-4945-treated 5XFAD mice.