Glymphatic influx and clearance are perturbed in Huntington's disease

Abstract

The accumulation of mutant huntingtin protein aggregates in neurons is a pathological hallmark of Huntington's disease (HD). The glymphatic system, a brain-wide perivascular network, facilitates the exchange of interstitial fluid and cerebrospinal fluid (CSF), supporting interstitial solute clearance of brain wastes. In this study, we employed dynamic glucose-enhanced (DGE) MRI to measure d-glucose clearance from CSF as a tool to predict glymphatic function in a mouse model of HD. We found significantly diminished CSF clearance efficiency in HD mice before phenotypic onset. The impairment of CSF clearance efficiency worsened with disease progression. These DGE MRI findings in compromised glymphatic function were further verified with fluorescence-based imaging of CSF tracer influx, suggesting an impaired glymphatic function in premanifest HD. Moreover, expression of the astroglial water channel aquaporin-4 in the perivascular compartment, a key mediator of glymphatic function, was significantly diminished in both HD mouse brain and human HD brain. Our data, acquired using a clinically translatable MRI, indicate a perturbed glymphatic network in the HD brain. Further validation of these findings in clinical studies will provide insights into the potential of glymphatic clearance as a therapeutic target as well as an early biomarker in HD.

Keywords: Neurodegeneration; Neuroimaging; Neuroscience.

Figure 1. Impaired glymphatic system as revealed by DGE MRI and fluorescence-based imaging in premanifest zQ175 HD mice.

(A) Illustration of DGE MRI scan timeline (upper panel) and representative DGE images for the third ventricle (lower panel) in a WT mouse and a zQ175 mouse at 4 months of age. (B) Average CSF-based DGE signal changes during the entire scan period from male WT and zQ175 mice. n = 5 mice/genotype. (C) Comparison of fitted clearance parameter μ_out for CSF. *P < 0.05 vs. WT by standard Student’s t test. (D) Comparison of fitted uptake parameter μ_in for CSF. (E) Representative images of BSA-647 fluorescent dye distribution in the brain parenchyma at 60 minutes after intra-CM injection. Note the wide distribution of fluorescent dye along the glymphatic pathway in WT mice, while very limited fluorescence distribution was seen in the HD mouse brain. Scale bar = 1 cm. The left panel shows the BSA-647 fluorescence images, and the right panel shows BSA-647 fluorescence images merged with DAPI staining images. (F) Quantification of the fluorescent dye distribution at 60 minutes after CSF tracer injection. *P < 0.05 vs. WT by standard Student’s t test. (G) Representative images of BSA-647 fluorescent dye distribution in the brain parenchyma at 180 minutes after intra-CM injection. Scale bar = 1 cm. (H) Quantification of the fluorescent dye distribution at 180 minutes after tracer injection. *P < 0.05 vs. WT by standard Student’s t test.

Figure 2. Decreased AQP4 perivascular localization (polarization) and levels in the brain of premanifest zQ175 HD mice.

(A) Representative Z-stack images of coimmunofluorescence staining of AQP4 and collagen IV in the striatum and cortex of 4-month-old zQ175 and WT controls. Scale bar = 100 μm. (B and C) Quantification of colocalized pixels of AQP4 (green in A) and collagen IV (red in A) in the striatum (B) and cortex (C) of 4-month-old male zQ175 mice and WT controls. *P < 0.05, **P < 0.01 vs. WT by standard Student’s t test. (D) Illustration of AQP4 and its astrocytic endfeet–anchoring protein complex. (E–J) Western blots (E) and quantification of Aqp4 (F); Aqp4-anchoring protein complex components Snta1 (G), Dtna (H), and Dag1 (I); and activated astrocyte marker Gfap (J). **P < 0.01 vs. WT by standard Student’s t test. Aqp4, aquaporin-4; Snta1, syntrophin alpha 1; Dtna, dystrobrevin alpha; Dag1, dystroglycan 1; Gfap, glial fibrillary acidic protein.

Figure 3. Perturbed glymphatic function and exacerbated AQP4 perivascular localization (polarization) in the manifest zQ175 HD mouse brain.

(A) Representative images of DGE signals (lower panel) in a WT mouse and a zQ175 mouse at 10 months of age. (B) Average dynamic d-glucose signals during the entire scan period from male WT and zQ75 mice. n = 5 mice/genotype. (C) Comparison of fitted clearance parameter μ_out d-glucose clearance rate in CSF of the third ventricle. **P < 0.01 vs. WT by standard Student’s t test. (D) Comparison of fitted clearance parameter μ_in d-glucose uptake rate in CSF of the third ventricle. (E) Representative images of coimmunofluorescence staining of AQP4 and collagen IV in the striatum (upper panel) and cortex (bottom panel) of 10-month-old male zQ175 and WT controls. Scale bar = 100 μm. (F) Quantification of colocalized pixels of AQP4 (green in E) and collagen IV (red in E) in the striatum of 10-month-old male zQ175 mice and WT controls. *P < 0.05 vs. WT by standard Student’s t test. (G) Quantification of colocalized pixels of AQP4 (green in E) and collagen IV (red in E) in the cortex of 10-month-old male zQ175 mice and WT controls. **P < 0.01 vs. WT by standard Student’s t test. (H) Representative images of BSA-647 fluorescent dye distribution in the brain parenchyma at 60 minutes after intra-CM injection in 10-month-old mice. Scale bar = 1 cm. The left panel shows the BSA-647 fluorescence images, and the right panel shows BSA-647 fluorescence images merged with DAPI staining images. (I) Quantification of the fluorescent dye distribution at 60 minutes after CSF tracer injection. *P < 0.05 vs. WT by standard Student’s t test.

Figure 4. Reduced perivascular AQP4 localization accompanying astrogliosis in the human HD brain.

(A) Representative images of coimmunofluorescence staining of AQP4 (red) and collagen IV (green) in the caudate putamen of a patient with HD and age-matched control. Scale bar = 100 μm. Insets in A are 3 times enlarged from the original images. (B) Quantification of colocalized pixels of AQP4 (red in A) and collagen IV (green in A) in the caudate putamen of HD patients (n = 6) and age-matched controls (n = 4). *P < 0.05 vs. control by standard Student’s t test. (C) Western blots of AQP4, SNTA1, and GFAP in the human caudate samples from 13 HD brains and 10 control brains. (D) Quantification of AQP4 (both isoforms), SNTA1, and GFAP protein levels (ratio to the loading control β-actin) in the caudate samples. **P < 0.01 vs. control by standard Student’s t test.