Fig 2: AQP4 deletion increased intraneuronal apoE aggregation in 3-month-old AQP4−/−/APP/PS1 mice. a, b Representative bands of Western blot and densitometry analysis of apoE and LRP1. c Immunofluorescence of apoE. ApoE expression was significantly decreased in the cerebral cortex following local injection of AAV encoding apoE siRNAs. d Quantification of apoE-positive area fraction in the cerebral cortex. e Double immunofluorescence of NeuN and total-Aβ. Total-Aβ expression was also obviously decreased in the cerebral cortex following local injection of AAV encoding apoE siRNAs. f, g Quantification of total-Aβ and NeuN/total-Aβ-positive area fraction in the cerebral cortex, respectively. h, i Representative bands of Western blot and densitometry analysis of APP, CTF-β, and Aβ monomers/oligomers in the cortex after local injection of AAV encoding apoE siRNAs. Data were analyzed by two-way ANOVA with Newman-Keuls post hoc test. Data are mean ± SEM, n = 4 per group, *p < 0.05; ***p < 0.001

Fig 3: AQP4 deletion aggravates decreases of florescent tracer influx into, and efflux out of, the brain of 3-month-old APP/PS1 mice. a, b Schematic images for intracisternal injection and intrastriatal injection of TR-d3. c Representative images of serial coronal brain sections from 1.0 mm anterior to 1.8 mm posterior to bregma (upper panel) and high magnification micrographs of the sections at the level of 0.5 mm anterior to bregma (low panel) showing TR-d3 penetration into the brain parenchyma at 30 min after intracisternal injection. d High magnification micrographs of the cortex showing TR-d3 penetration into the perivascular space (star) and adjacent brain parenchyma (dotted line) at 30 min after intracisternal injection. e Representative images of serial coronal brain sections from 1.0 mm anterior to 1.8 mm posterior to bregma (upper panel) and high magnification micrographs of the sections at the level of 0.2 mm anterior to bregma (low panel) showing residue of TR-d3 within the striatum at 1 h after intracisternal injection. f Low (upper panel) and high (low panel) magnification micrographs of sections of the dcLNs showing TR-d3 distribution at 1 h after intracisternal injection. g Quantification of intracisternally injected TR-d3 area fraction in the brain. h Quantification of the mean fluorescence intensity of TR-d3 in the cerebral cortex after intracisternal injection. i Quantification of intrastriatally injected TR-d3 area in the brain. j Quantification of intrastriatal injected TR-d3 in the dcLNs. Data were analyzed by the two-way ANOVA with Tukey’s post hoc test or repeated two-way ANOVA with Tukey’s post hoc test. Data are mean ± SEM. n = 4 per group, *p < 0.05; **p < 0.01; ***p < 0.001

Fig 4: Schematic diagram of our working model for the glymphatic system mediated Aβ and apoE clearance and microglia mediated Aβ degradation synergistically preventing Aβ plague formation in the early stages of the AD mouse model. 3-month-old APP/PS1 mice show extracellular and intraneuronal accumulation of Aβ because of the long-term of expression transgenic APP/PS1. Excessive interstitial Aβ may result in activation of astrocytes with loss of perivascular AQP4 polarization, subsequently impairing glymphatic clearance of Aβ and apoE. On the other hand, microglia undergo activation and promote Aβ clearance, potentially preventing Aβ plaque formation. AQP4 deletion exacerbates glymphatic transport impairment and intraneuronal accumulation of Aβ and apoE, but Aβ plaques have not yet aggregated, which may be related to the increase of phagocytosis of reactive microglia

Fig 5: AQP4 deletion increased intraneuronal apoE aggregation in 3-month-old AQP4−/−/APP/PS1 mice. a, b Representative bands of Western blot and densitometry analysis of apoE. c–e Double immunofluorescence of apoE/GS, apoE/NeuN, and apoE/total-Aβ, respectively. Note that AQP4 deletion in APP/PS1 mice increased apoE immunoreactive intensity in GS-positive astrocytes and NeuN-positive neurons. ApoE and total-Aβ double immunoreactive products were also accumulated. f–i Quantification of apoE, apoE/GS, apoE/NeuN, and apoE/total-Aβ-positive area fraction in the cerebral cortex, respectively. Data in i were analyzed by Student’s t test; other data were analyzed by the two-way ANOVA with Newman-Keuls post hoc test. Data are mean ± SEM, n = 4 per group, *p < 0.05; **p < 0.01; ***p < 0.001