Reduced myelin contributes to cognitive impairment in patients with monogenic small vessel disease

Abstract

Introduction: Myelin is pivotal for signal transfer and thus cognition. Cerebral small vessel disease (cSVD) is primarily associated with white matter (WM) lesions and diffusion changes; however, myelin alterations and related cognitive impairments in cSVD remain unclear.

Methods: We included 64 patients with familial cSVD (i.e., cerebral autosomal dominant arteriopathy with subcortical infarcts and leukoencephalopathy [CADASIL]) and 20 cognitively unimpaired individuals. χ separation applied to susceptibility weighted imaging was used to assess myelin and iron within WM hyperintensities, normal appearing WM, and two strategic fiber tracts. Diffusion-based mean diffusivity and free water were analyzed for comparisons. Cognitive impairment was assessed by the Trail Making Test.

Results: CADASIL patients showed reduced myelin within WM hyperintensities and its penumbra in the normal appearing WM. Myelin was moderately correlated with diffusion and iron changes and associated with slower processing speed controlled for diffusion and iron alterations.

Discussion: Myelin constitutes WM alterations distinct from diffusion changes and substantially contributes to explaining cognitive impairment in cSVD.

Highlights: χ-negative magnetic resonance signal was reduced within white matter hyperintensities and normal appearing white matter in patients with cerebral autosomal dominant arteriopathy with subcortical infarcts and leukoencephalopathy, suggesting widespread myelin decreases due to cerebral small vessel disease (cSVD). χ-negative values were only moderately associated with diffusion tensor imaging derived indices including free water and mean diffusivity, suggesting that χ separation depicts distinct microstructural changes in cSVD. Alterations in χ-negative values made a unique contribution to explain processing speed impairment, even when controlled for diffusion and iron changes.

Keywords: cerebral autosomal dominant arteriopathy with subcortical infarcts and leukoencephalopathy; chi separation; diffusion magnetic resonance imaging; susceptibility mapping; white matter hyperintensities.

FIGURE 1

Spatial frequency of white matter hyperintensities and lacunar lesions. A, Frequency mapping of WMHs: the percentage of subjects with WMHs are mapped voxel‐wise onto axial slices for the CN group (upper row) and the CADASIL group (lower row). B, The frequency mapping for lacunes is shown in a corresponding way. Only voxels with a percentage > 0 are mapped. Left side is left hemisphere. CADASIL, cerebral autosomal dominant arteriopathy with subcortical infarcts and leukoencephalopathy; CN, cognitively normal; WM, white matter; WMH, white matter hyperintensity.

FIGURE 2

Voxel‐wise and ROI level descriptive results for χ negative showing group difference in CADASIL and CN controls. A, Significant voxel‐wise difference between CADASIL patients and CN of χ negative (CADASIL < CN) in the white matter mapped onto axial and coronal brain slices. Only voxels are plotted which survived false discovery rate correction on a level of α = 0.05 and a voxel extent threshold of 10. Testing for the opposite contrasts did not yield any significant cluster. Left side is left hemisphere. B, Violin plots with inserted box plots for χ‐negative difference scores extracted from NAWM and WMH areas. Each line represents the values for a CADASIL patient. C, Regular mean scores extracted from the global white matter, left ATR, and FM. D, Differences scores in χ‐negative values as a function of distance from WMH areas in the CADASIL group. Difference scores have been calculated in CADASIL from each subject's specific WMH/NAWM mask after they were corrected for age, sex, education, χ positive, and FW such that difference scores < 0 indicates lower χ‐negative scores (reduced myelin) in CADASIL compared to the average score in the CN group. Regular scores for the white matter and the two tracts are not given as differences scores, because these areas don't vary between individuals. For NAWM, WMH, and the WMH penumbra, one‐sample Welch t tests (μ = 0) were conducted and p values are plotted as asterisks (* p < 0.05, ** p < 0.01, *** p < 0.001). For the comparison of the white matter and tract scores, a regular linear model with a grouping factor was used, with age, sex, education, χ positive, and FW added as covariates. ATR, anterior thalamic radiation; CADASIL, cerebral autosomal dominant arteriopathy with subcortical infarcts and leukoencephalopathy; FM, forceps minor; NAWM, normal appearing white matter; ROI region of interest; WM, white matter; WMH, white matter hyperintensity.

FIGURE 3

Association between χ negative and each MD, RD, FW, and χ positive in NAWM and WMH. Scatterplot for the association between χ negative and MD (A), RD (B), FW (C), χ positive (D) for the CADASIL group in NAWM (left column) and WMH (middle column) and the CN group in NAWM (right column). Regression line and 95% CI (shaded area) are shown. ATR, anterior thalamic radiation; CADASIL, cerebral autosomal dominant arteriopathy with subcortical infarcts and leukoencephalopathy; CI, confidence interval; FM, forceps minor; FW, free water; MD, mean diffusivity; NAWM, normal appearing white matter; RD, radial diffusivity; ROI region of interest; WM, white matter; WMH, white matter hyperintensity.

FIGURE 4

Association between χ‐negative scores in white matter regions and processing speed. Residual regression plots for different ROIs (A–E) from ridge regression models for processing speed scores as a function of χ‐negative scores in patients with CADASIL. Regression line and 95% confidence level (shaded area) are shown. ATR, anterior thalamic radiation; CADASIL, cerebral autosomal dominant arteriopathy with subcortical infarcts and leukoencephalopathy; FM, forceps minor; NAWM, normal appearing white matter; ROI region of interest; WM, white matter; WMH, white matter hyperintensity.