Huntington disease (HD) is a neurodegenerative disorder with motor, cognitive, and psychiatric deficits caused by single gene mutation.¹ The YAC128 mouse model of HD provides an opportunity to evaluate candidate therapeutics in preclinical studies.^2,3 Memantine, a NMDA antagonist, may protect medium spiny neurons by dampening detrimental extrasynaptic NMDA receptor signaling in YAC128 mice.⁴ Its potential as a protective agent in patients with HD is supported by a recent open-label study.⁵ In this work we investigated the effect of memantine treatment on brain structural changes in YAC128 mice using structural MRI and DTI.

Wild-type (WT) and YAC128 mice of the FVB strain were randomly assigned to 3 treatment groups for each genotype: vehicle, low dose (2mg/kg), and high dose memantine (10mg/kg). Treatment started from 4-week of age via drinking water. A batch of cognitive and behavioral tests were conducted at 2, 4, 6, and 8 months of age, including novel object localization (NOL) test, rotarod test, and climbing test. After 8-month of treatment, 16 WT and 23 YAC128 mice were scanned on a 7T scanner (ClinScan, Bruker BioSpin, Germany) using 4 channel array coils. The structural image was acquired by a fast-spin-echo T2-weighted sequence with TR=3080ms, TE=43ms and 0.1x0.1x0.3mm³ voxel resolution with coil inhomogeneity normalization. Diffusion images were acquired using a spin echo EPI sequence of 30 diffusion sensitizing directions, b=1500s/mm², TR=11000ms, TE=41ms, voxel size=0.25x0.25x0.3mm³ and 9 averages. Reversed phase-encoding EPI images were collected for distortion correction.

Brain extraction was performed using 3D-PCNN ⁶ with manual editing. Structural images were linearly registered to a mouse brain template of the same strain of the same age ⁷ and then averaged to create a study specific template. Images were then nonlinearly registered to the template using FSL (v5.0).⁸ Tensor-based morphometry (TBM) was conducted to compare the volume differences between groups. Volume ratio of the caudate putamen (CPu) over the whole brain was also compared. For DTI, eddy current, motion, and EPI distortion were corrected before Fractional anisotropy (FA), radial diffusivity (Dr), and parallel diffusivity (Dp) were calculated (FSL). Two rounds of nonlinear registration based on the T2-w and FA images, respectively brought the FA, Dr, and Dp maps to a study-specific template. After Gaussian smoothing of 0.45 mm, voxel-wise comparisons were conducted using SPM8.⁹ Regions of interest (ROIs) defined based on the voxel-wise results and the white matter atlas¹⁰ were also compared.

Memantine-treated YAC128 mice still showed atrophy in the CPu, thalamus, amygdala, hippocampus, and some cortical regions, as well as enlarged cerebellum compared to WT mice. No significant difference in volume was found between low and high-dose groups. Relative CPu volume increased with memantine dosage dependently though the increase did not reach significance (Fig. 1). Compared to vehicle-treated WT mice, low dose-treated YAC mice showed no difference in FA, while vehicle and high dose treated YAC mice had FA reduction in bilateral anterior corpus callosum, anterior commissure, and internal capsule (Fig. 2), as well as right optic tract, left stria terminalis and left stria medularis. No difference in Dr or Dp was found. Memantine treatment resulted in significant cognitive improvements at 4 month of age as assessed by the rotarod motor learning task (Fig.3). Furthermore, FA values in major fiber tracts correlates with cognitive and behavioral scores at the end of treatment (Fig. 4).Our results suggest that memantine results in dosage dependent change in gray matter volume and especially white matter integrity measured by FA in YAC128 mice. Low dose memantine improved FA in many fiber bundles while high dose had no effect.