Late-onset GM2 gangliosidosis: magnetic resonance imaging, diffusion tensor imaging, and correlational fiber tractography differentiate Tay-Sachs and Sandhoff diseases

Abstract

GM2 gangliosidosis is lysosomal storage disorder caused by deficiency of the heterodimeric enzyme β-hexosaminidase A. Tay-Sachs disease is caused by variants in HEXA encoding the α-subunit and Sandhoff disease is caused by variants in HEXB encoding the β-subunit. Due to shared clinical and biochemical findings, the two have been considered indistinguishable. We applied T1-weighted volumetric analysis, diffusion tensor imaging (DTI), and correlational fiber tractography to assess phenotypic differences in these two diseases. 51 T1-weighted and 40 DTI scans from 19 Late-Onset GM2 patients with either late-onset Sandhoff disease (LOSD), or late-onset Tay-Sachs (LOTS) were included and compared to 1033 neurotypical control volumetric MRI scans. LOTS patients had significantly smaller cerebellum volume compared to neurotypical controls (p < 0.0001) and LOSD patients (p < 0.0001). There was no statistical difference for the volume of any structure between LOSD and neurotypical controls. DTI analysis showed LOTS patients had higher mean diffusivity (MD) in the left cerebellum (p = 0.003703), right cerebellum (p = 0.003435), superior cerebellar peduncle (p = 0.007332), and vermis (p = 0.01007) compared to LOSD. LOTS patients had lower fractional anisotropy (FA) in the left cerebellum (p = 0.005537), right cerebellum (p = 0.01905), SCP (p = 0.02844), and vermis (p = 0.02469) when compared to LOSD. Correlational fiber tractography identified fiber tracts in cerebellar pathways with higher FA and lower MD in LOSD patients compared to LOTS patients. Our study shows neurobiologic differences between these two related disorders. To our knowledge, this is the first study using correlational tractography in a lysosomal storage disorder. This result indicates a greater burden of cerebellar pathology in LOTS patients compared with LOSD patients.

Keywords: Correlational tractography; GM2 gangliosidosis; Late-onset Sandhoff; Late-onset Tay–Sachs; Neuroimaging biomarkers.

Fig. 1

Participant age at each T1-weighted MRI Scan. LOTS patients are shown in blue and LOSD Patients are shown in red. Each T1-weighted scan is represented as a circle for all 51 scans where each of the 19 participants is on a separate row

Fig. 2

T1-weighted MRI comparisons. One age matched neurotypical control (NC, 55-year-old male), one age matched LOSD patient (56-year-old male), and one age matched LOTS patient (56-year-old female) shown in the Sagittal, Axial, and Coronal planes. The blue arrows designate the 4 th ventricle for each participant and the red arrows designate the cerebellum for each participant. The LOTS patient shows significant cerebellar atrophy in all three views. The neurotypical control was from the NIMH dataset

Fig. 3

Age related changes in cerebellum volume. LOSD (red) participants demonstrate normal cerebellum volume when compared with neurotypical age matched controls. LOTS participants (blue) demonstrate significantly diminished cerebellum volume when compared with neurotypical controls. A Total cerebellum volume including bilateral cerebellum gray and white matter. B Cerebellum volumes were normalized to total intracranial volume and are expressed as a percentage. Statistical differences were observed between GM2 patients and neurotypical controls (p_corrected < 0.01), LOTS patients and neurotypical controls (p_corrected < 0.01), and between LOTS patients and LOSD patients (p_corrected < 0.01) for cerebellar volume

Fig. 4

Atlas Based Fiber Tractography of the whole brain demonstrating age related effects on A fractional anisotropy (FA), B mean diffusivity (MD), C radial diffusivity (RD), D axial diffusivity (AD) between LOTS patients (blue) and LOSD patients (red). LOTS patients demonstrated no difference in FA (χ²(1) = 0.70, p = 0.40), and increased MD (χ²(1) = 7.62, p < 0.01), RD (χ²(1) = 7.28, p < 0.01), and AD (χ²(1) = 7.86, p < 0.01) compared to LOSD patients in fiber tracts throughout the whole brain when age was accounted for

Fig. 5

Atlas Based Fiber Tractography of the left cerebellum demonstrating age related effects on A fractional anisotropy (FA), B mean diffusivity (MD), C radial diffusivity (RD), D axial diffusivity (AD) between LOTS patients (blue) and LOSD patients (red). LOTS patients demonstrated decreased FA (χ²(1) = 7.70, p < 0.01), and increased MD (χ²(1) = 8.42, p < 0.01), RD (χ²(1) = 8.51, p < 0.01), and AD (χ²(1) = 8.25, p < 0.01) compared to LOSD patients in left cerebellar fiber tracts when age was accounted for

Fig. 6

Atlas Based Fiber Tractography of the corpus callosum demonstrating age related effects on A fractional anisotropy (FA), B mean diffusivity (MD), C radial diffusivity (RD), D axial diffusivity (AD) between LOTS patients (blue) and LOSD patients (red). There were no significant differences in between LOTS and LOSD for FA, MD, RD, or AD in the corpus callosum when age was accounted for

Fig. 7

Correlational fiber tractography assessed differences in fractional anisotropy in LOSD and LOTS patients at varying length (voxels) and T thresholds. Fiber tracts shown in red were evaluated to have a higher fractional anisotropy in LOSD patients compared to LOTS patients and were observed primarily in the cerebellum (FDR < 0.05). No fiber tracts were evaluated to have a higher FA in LOTS patients compared to LOSD patients (blue)

Fig. 8

Correlational fiber tractography assessed differences in mean diffusivity in LOSD and LOTS patients at varying length (voxels) and T thresholds. Fiber tracts shown in blue were evaluated to have a higher mean diffusivity in LOTS patients compared to LOSD patients and were observed primarily in the cerebellum (FDR < 0.05). No fiber tracts were evaluated to have a higher mean diffusivity in LOSD patients compared to LOTS patients (red)