Quantitative brain morphometry identifies cerebellar, cortical, and subcortical gray and white matter atrophy in late-onset Tay-Sachs disease

Abstract

Background:

Cerebellar atrophy is a characteristic sign of late-onset Tay-Sachs disease (LOTS). Other structural neuroimaging abnormalities are inconsistently reported. Our study aimed to perform a detailed whole-brain analysis and quantitatively characterize morphometric changes in LOTS patients.

Methods:

14 patients (8M/6F) with LOTS from three centers were included in this retrospective study. For morphometric brain analyses, we used deformation-based morphometry (DBM), voxel-based morphometry (VBM), surface-based morphometry (SBM), and spatially unbiased cerebellar atlas template (SUIT).

Results:

The quantitative whole-brain morphometric analysis confirmed the finding of profound pontocerebellar atrophy with most affected cerebellar lobules V and VI in LOTS patients. Additionally, the atrophy of structures mainly involved in motor control, including bilateral ventral and lateral thalamic nuclei, primary motor and sensory cortex, supplementary motor area, and white matter regions containing corticospinal tract together, was present. The atrophy of the right amygdala, hippocampus, and regions of occipital, parietal and temporal white matter was also observed in LOTS patients in contrast with controls (p<0.05, FWE corrected). Patients with dysarthria and those initially presenting with ataxia had more severe cerebellar atrophy.

Conclusion:

The quantitative whole-brain morphometric analysis confirmed impairment of cerebellar regions responsible for speech and hand motor function in LOTS patients. Widespread morphological changes of motor cortical and subcortical regions and tracts in white matter were observed in LOTS patients indicating abnormalities mainly in central motor circuits.

Introduction

Late-onset Tay-Sachs disease (LOTS) is a rare neurodegenerative disease caused by biallelic pathogenic variants in the HEXA gene. Dysfunction of beta-hexosaminidase A deficiency leads to intralysosomal accumulation of GM2 gangliosides in neurons.¹

The main clinical symptoms of LOTS include ataxia, symptoms of lower motor neuron disease, and psychiatric impairment.² Cerebellar atrophy is well known characteristic sign of patients with LOTS according to neuroradiological and histopathological studies.^3–10

The lobar atrophy and corpus callosum thinning have been reported, but the findings to date have been inconsistent in the descriptions resulting from visual assessments of brain MRI.¹¹ In a recent brain MRI and spectroscopic study, 10 late-onset GM2-gangliosidosis (including 7 LOTS) patients were compared to 7 controls; volumetric and cortical thickness analysis showed reduced cerebellar volumes with decreased N-acetylaspartate, increased myo-inositol and greater thickness of the left fusiform gyrus in the patient group.¹² Another spectroscopic study with 9 late-onset GM2-gangliosidosis showed significantly reduced N-acetylaspartate in the thalamus and normal-appearing white matter, supposedly reflecting neuronal loss and atrophy of these structures.¹³ A quantitative brain MRI study showed progressive brain disease pattern affecting children with infantile and juvenile GM2-gangliosidoses. The infantile group showed a progressive increase of intracranial and brain volume over time, whereas juvenile group demonstrated mild-to-moderate disease progression of total brain atrophy. Both groups however, presented with progressive cerebellar, basal ganglia and corpus callosum atrophy.¹⁴ To date, no study examined the pattern of brain atrophy in LOTS using a voxel-wise approach.

Our study aimed to perform a detailed whole-brain analysis and quantitatively characterize the pattern of morphometric changes in LOTS patients. Such knowledge may define the potential brain MRI markers of the disease, which may be useful for the design of future treatment trials.

Patients and methods

Study participants

14 LOTS patients (8M/6F) were included in this retrospective study. Patients were recruited from three centers - the Department of Pediatrics and Inherited Metabolic Disorders, First Faculty of Medicine, Charles University, Prague, Czech Republic (n=9), Department of Pediatrics, University of Minnesota, Minneapolis, USA (n=4) under the study protocol NCT 00668187, and the Department of Neurology Ludwig-Maximilians University, Munich, Germany (n=1). To increase the power of the study, a control-to-case ratio of 2:1 was chosen. Twenty-eight sex- and age-matched healthy control participants were examined at the Department of Radiology, First Faculty of Medicine, Charles University. They were recruited from the general community through advertisements as part of a project assessing variation in longitudinal changes of quantitative MRI parameters in healthy population under the protocol NCT03706118. The controls had to be free of neurologic or other medical disorders affecting brain health and free of structural brain abnormalities influencing morphometric measurements such as arachnoid cysts, white matter lesions etc.

Clinical findings

The following clinical data were reported: disease onset, first presenting symptom and individual neurological and psychiatric symptoms at the time of MRI examination, and HEXA pathogenic variants.

Imaging protocol

MRI examinations of all participants were performed on Siemens MRI scanners and included a Magnetization Prepared Rapid Gradient‐Echo (MPRAGE) image with the spatial resolution 1×1×1 mm³ covering the whole brain.

In the Prague and Minneapolis centers, MR images were acquired on a Siemens Skyra 3T scanner (Siemens Healthcare, Erlangen, Germany) equipped with a 20-channel head coil. The following MPRAGE pulse sequence parameters were used in Prague: echo time (TE) 2.96 ms, inversion time (TI) 900 ms, repetition time (TR) 2300 ms, flip angle (FA) 9°; field of view (FOV) 176×256×256 mm and in Minneapolis: TE 3.65 ms, TI 1100 ms, TR 2530 ms, FA 7°, FOV 256×256×256 mm. In the Munich center, MR images were acquired on a Siemens Aera 1.5T scanner (Siemens Healthcare, Erlangen, Germany) equipped with a 20-channel head coil and the following MPRAGE parameters: TE 2.49 ms, TI 1100 ms, TR 1670 ms, FA 15°, and FOV 256×223×280 mm.

Whole brain voxel-based, deformation-based, and surface-based morphometry

For voxel-based morphometry (VBM), deformation-based morphometry (DBM), and surface-based morphometry (SBM), the MRI preprocessing was done using the Statistical Parameter Mapping 12 (SPM12; version 7771) and the Computational Anatomy Toolbox 12 (CAT12; version 12.8.2, r2159).

For VBM, MPRAGE images were normalized and segmented into white matter (WM), gray matter (GM), and cerebrospinal fluid (CSF). These volumes were used to calculate brain parenchymal fraction (BPF), i.e., the ratio of the sum of GM and WM volume to the total intracranial volume (TIV). The segmented GM images were smoothed with a Gaussian kernel of 8 mm full width at half maximum (FWHM). The smoothed GM images were then used in statistical analysis to assess the GM atrophy.

DBM uses the Jacobian determinant of non-linear deformation of MRI image in registration to the template. It enables sensitive detection of shape differences and regional size alterations irrespective of GM, WM, or CSF tissue classification.¹⁵ For the analysis, Jacobians were smoothed using a Gaussian kernel of 8 mm FWHM.

Cortical thickness was estimated according to the earlier published method by Dahnke et al., 2013.¹⁶ The surface data were resampled to a 32,000 vertices mesh and smoothed with a Gaussian kernel of 12 mm FWMH.

Cerebellar morphometry

SUIT (a spatially unbiased atlas template of the cerebellum and brainstem) is a MATLAB toolbox dedicated to the analysis of imaging data of the human cerebellum and contains a high-resolution atlas template of the human cerebellum and brainstem.¹⁷ SUIT was used for detailed analysis of cerebellar atrophy.

The segmented cerebellar GM data were transformed into the Diffeomorphic Anatomical Registration Through Exponentiated Lie Algebra (DARTEL) atlas space. The images were smoothed with a Gaussian kernel of 3mm FWHM. Results were visualized as cerebellar flatmaps using the SUIT toolbox.¹⁸ Total volume of the cerebellum as well as individual volumes of cerebellar lobes (including both, GM and WM) were extracted using the SUIT atlas.

Corpus callosum assessment

The cross-sectional area of the corpus callosum (CC) was measured on T1w MPRAGE images and averaged in seven sagittal slices (slice thickness = 1 mm) centered around the mid-sagittal slice using an automated procedure as described previously¹⁹ and also used in Nestrasil et al.¹⁴

Statistical analysis

Volumetric and morphometric data were compared between LOTS and controls using the general linear model with sex, age, and TIV as covariates of no interest. Effect size was estimated using Partial Eta Squared which indicates the proportion of the total variance in the dependent variable that is attributed to an independent variable while controlling for covariates.

Statistical analysis of voxel-wise methods was performed using the non-parametric TFCE test (threshold-free cluster enhancement) with 10 000 permutations.²⁰ Age, sex, and in the case of VBM also TIV were used as covariates of no interest. Statistical maps were generated using a family-wise error–corrected threshold of P <0.05. Visualization of statistical maps was performed with MRICroGL software (https://www.nitrc.org/projects/mricrogl).

Additionally, we have analyzed the associations between morphometric measures (cerebellar regional volumes, CC area, and BPF) and clinical parameters (disease duration, ataxia as a presenting symptom, presence of tremor, dysarthria, psychiatric disturbances, and upper motor neuron signs at the time of MRI examination). For these analyses, cerebellar volumes and CC area were adjusted according to TIV using the residual method.²¹ The effect of disease duration on morphometric measures was assessed using partial correlation controlling for age. Mann-Whitney U test was used to compare morphometric measures between patient groups with and without a particular clinical symptom. P values <0.05 were considered statistically significant.

Ethics

All information was accessed in accordance with the applicable laws and ethical requirements for the study period concerned in compliance with the Declaration of Helsinki, revised in 2013. All examined patients signed an informed consent form for genetic testing. The study was approved by the Institutional Review Board (IRB) of the General University Hospital in Prague (Ethics Committee Approval Number: 1471/19) and the University of Minnesota (code number: 1007M85712). The data sharing was approved by the University of Minnesota IRB (code number: 1504E69081).

Results

Patient characteristics

The demographic and clinical data of 14 patients (8 males, 6 females), including three sib pairs, are summarized in Table 1. The median age of disease onset was 21.5 years (range 12–46 years). The most common presenting symptoms were leg weakness in 8/14 (57%) and gait ataxia in 5/14 (36%). The median age at diagnosis was 37.5 years (range 20–69 years). The median age at the MRI exam used for the analysis in this study was 42.5 years (range 20–70 years). Six patients included in this study were also assessed in a previous imaging study focused on semiquantitative parameters of cerebellar and brainstem atrophy.¹¹

Table 1:

Clinical findings in patients with adult form of Tay-Sachs disease

Patient	Gender	Age at onset	Age at diagnosis	Age at MRI	HEXA mutations	Neurologic symptoms	Psychiatric symptoms	Cerebellar volume [cm3]
1	M	20	30	42	c.[805G>A];[1274_1277dup] p.[G269S];[Y427Ifs*5]	proximal leg weakness, gait and hand ataxia, action and resting hand tremor, dysarthria, Babinski+, hyperreflexia	depression, anxiety	44
2	M	46	69	70	c.[805G>A];[805G>A] p.[G269S];[G269S]	proximal leg weakness, action hand tremor	-	101
3	F	12	42	43	c.[805G>A];[1330+1G>A] p.[G269S]/missplicing	stutter (dysarthria), proximal leg and arm weakness, gait and hand ataxia, hyperreflexia	-	37
4	M	25	54	52	c.[805G>A];[754C>T] p.[G269S];[R252C]	gait ataxia, hand ataxia, proximal leg and arm weakness, dysarthria	-	50
5	M	25	46	57	c.[805G>A].;[754C>T] p.[G269S];[R252C]	gait ataxia, proximal leg weakness, dysarthria, action hand tremor	cognitive impairment	40
6	F	17	20	24	c.[805G>A];[1274_1277dup] p.[G269S];[Y427Ifs*5]	proximal leg weakness, gait and hand ataxia	depression	66
7	F	17	24	28	c.[805G>A];[1274_1277dup] p.[G269S];[Y427Ifs*5]	gait ataxia, hand ataxia, proximal leg weakness	depression, cognitive impairment	58
8	M	16	25	28	c.[805G>A];[947dupA] p.[G269S];[T316*]	proximal leg weakness, gait ataxia, proximal arm weakness, action hand tremor, dysarthria, dystonia, Babinski+, hyperreflexia	-	68
9	M	30	46	47	c.[805G>A];[805G>A] p.[G269S];[G269S]	distal leg weakness followed by proximal weakness, action hand tremor, gait ataxia	-	90
10	M	19	47	50	c.[806–7 G>A];[1073+1 G>A] missplicing/missplicing	proximal leg weakness, proximal arm weakness, fasciculations of the abdomen and back muscles, gait ataxia, hyperreflexia	depression, anxiety	126
11	F	34	34	38	c.[806–7 G>A];[1073+1 G>A] missplicing/missplicing	proximal leg weakness, action and resting hand tremor, headache, tongue atrophy and fasciculations, gait ataxia, dysarthria, hyperreflexia, Hoffmann+	adjustment disorder, anxiety, cognitive impairment	119
12	M	35	41	43	c.[805G>A];[1274_1277dup] p.[G269S];[Y427Ifs*5]	gait ataxia, hand ataxia, proximal leg weakness, resting hand tremor, dysarthria, headache	anxiety, depression	52
13	F	22	23	23	c.[805G>A];[1510C>T] p.[G269S];[R504C]	proximal leg weakness, hand and gait ataxia	-	58
14	F	21	21	20	c.[805G>A];[1510C>T] p.[G269S];[R504C]	hand ataxia, action hand tremor, proximal leg weakness	-	83

Patient 4 and 5; patient 6 and 7, patient 13 and 14 are siblings. The initial presenting symptom is highlighted in bold font; other symptoms are listed without particular order.

The most common neurological findings at the time of MRI examination included proximal leg weakness 14/14 (100%), gait and/or hand ataxia 13/14 (93%), tremor in hands 8/14 (57%), and dysarthria 7/14 (50%). Other symptoms included dystonia 1/14 (7%) and headache 2/14 (14%). Symptoms of upper motor neuron lesion were present in 5/14 (36%); hyperreflexia in five and positive Babinski or Hoffmann signs in three patients. Depression and/or anxiety 6/14 (43%), and cognitive impairment 3/14 (21%) were the most frequent psychiatric symptoms. The most frequent genetic mutation was 805G>A, with an allelic frequency of 50%. Notably, two homozygotes with this mutation had a relatively late symptom onset and mild cerebellar atrophy. Additionally, two unrelated compound heterozygotes with mutations 806–7 G>A and 1073+1 G>A had remarkably mild cerebellar atrophy.

MRI datasets of 28 age- and sex-matched controls were available (16 males, 12 females). The median age at their MRI examination was 43 years (range 21–73 years).

Voxel-based, deformation-based, and surface-based morphometry

Adjusted for age and sex, LOTS patients had significantly lower BPF compared to controls (0.82±0.03 vs. 0.77±0.04, p<0.001; Fig.1). VBM analysis showed the cerebellar atrophy, the atrophy of right amygdala, hippocampus, and adjacent insular cortex, bilateral anterior thalamic nuclei, primary motor and sensory cortex, supplementary motor area, and regions of occipital (bilateral lingual gyri and right calcarine sulcus) and temporal (bilateral fusiform gyri) lobes in LOTS patients in contrast with controls (p<0.05, FWE corrected; Fig.2).

Figure 1: Basic morphometric measures and comparisons between LOTS and controls.

Dot graphs comparing brain parenchymal fraction (BPF), mean cortical thickness (CT), and corpus callosum (CC) sagittal area; control values are represented by black circles, late-onset Tay-Sachs (LOTS) patient values by grey rectangles. Results of statistical comparisons adjusted for age, sex, and total intracranial volume (for CC sagittal area) are shown; ns = not significant.

Figure 2: DBM and VBM results for contrast controls > LOTS in the axial plane.

Significant atrophy (thresholded at P_FWE < 0.05) is displayed in blue–green scale for DBM and red – yellow scale for VBM. Z-coordinates in mm in the Montreal Neurological Institute (MNI) space are shown for each slice. Results of the cerebellar atrophy are shown only for DBM due to noticeable overlap. VBM results for the cerebellum are not shown.

DBM analysis revealed profound pontocerebellar atrophy and volume loss in bilateral anterior, ventral, and lateral thalamic nuclei, white matter regions containing the corticospinal tract (deep lobar white matter, capsula interna, and cerebral peduncles), and in the right parietal (angular and inferior parietal gyri) and left temporal (middle temporal gyrus) lobes (p<0.05, FWE corrected; Fig.2)

SBM analysis returned no significant differences in regional cortical thickness between LOTS patients and age- and sex-matched controls. The average cortical thickness in LOTS patients was 2.50±0.12 mm compared to 2.49±0.08 mm in controls (p = 0.70; Fig.1).

Cerebellar morphometry

Targeted cerebellar VBM using the SUIT software showed diffuse cerebellar atrophy with more involvement of the right hemisphere and the most noticeable volume loss in lobuli V and VI, followed by lobuli VIIb and Crus I (Fig.3, Table 2). Individual total cerebellar volumes for each patient are shown in Table 1.

Figure 3: SUIT-VBM results.

Cerebellar FLATMAP (a flat representation of the human cerebellum) thresholded at P_FWE < 0.05 shows diffuse cerebellar atrophy in LOTS patients with more involvement of the right hemisphere. The most noticeable volume loss is in the lobules V and VI, followed by the lobules VII and VIII.

L – left, R - right

Table 2:

Volumetric comparison of cerebellar structures between late-onset Tay-Sachs patients and healthy controls

Cerebellar structure	Volume [cm3]		P-value	Partial Eta Squared
	LOTS patients	Healthy controls
Total cerebellum	70.84 (28.81)dys	114.94 (9.54)	<0.001	0.57
Right hemisphere	33.55 (13.68)dys	54.94 (4.56)	<0.001	0.58
Left hemisphere	34.01 (14.16)dys, atax	55.24 (4.79)	<0.001	0.56
Vermis	3.28 (0.99)dys	4.75 (0.44)	<0.001	0.52
Regional divisions
Lobule I-IV	4.27 (1.21)dys	6.03 (0.58)	<0.001	0.51
Lobule V	5.08 (1.94)dys, atax	8.12 (0.68)	<0.001	0.58
Lobule VI	10.04 (4.84)dys, atax	17.57 (1.39)	<0.001	0.58
Crus I	15.09 (6.66)	25.22 (2.54)	<0.001	0.55
Crus II	11.95 (4.54)	18.75 (1.86)	<0.001	0.53
Lobule VIIb	6.19 (2.77)dys	10.29 (1.01)	<0.001	0.55
Lobule VIIIa	5.94 (2.57)dys, atax	9.58 (1.08)	<0.001	0.51
Lobule VIIIb	4.76 (2.07)dys, atax	7.56 (1.03)	<0.001	0.46
Lobule IX	3.37 (1.61)atax	5.94 (0.96)	<0.001	0.49
Lobule X	0.86 (0.16)	1.12 (0.12)	<0.001	0.46
Vermis VI	1.05 (0.39)dys	1.59 (0.13)	<0.001	0.52
Vermis Crus I	0.01 (0.00)atax	0.01 (0.00)	0.001	0.25
Vermis Crus II	0.23 (0.09)dys, atax	0.36 (0.04)	<0.001	0.50
Vermis VIIb	0.10 (0.04)	0.18 (0.02)	<0.001	0.55
Vermis VIIIa	0.88 (0.19)dys	1.17 (0.13)	<0.001	0.48
Vermis VIIIb	0.37 (0.11)dys	0.52 (0.06)	<0.001	0.44
Vermis IX	0.49 (0.15)dys	0.70 (0.09)	<0.001	0.42
Vermis X	0.16 (0.05)	0.22 (0.04)	<0.001	0.33

dys - regions with more severe atrophy in patients with dysarthria;

atax – regions with more severe atrophy in patients with primary atactic presentation

Corpus callosum sagittal area

Adjusted for age, sex, and TIV, the mean sagittal area of the corpus callosum was numerically smaller in LOTS patients compared to the controls; however, there was only a trend for statistical significance of this finding (6.0±0.9 vs. 6.7±0.8 cm², p = 0.08; Fig.1).

Associations with clinical parameters

Controlling for age, disease duration was not associated with any morphometric measure, including BPF, CC area, and regional cerebellar volumes. Patients with dysarthria (n=7) had significantly more severe cerebellar atrophy, particularly affecting lobules I-IV, V, VI, VIIb, VIIIa, VIIIb, and vermis VI, crus II, VIIIa, VIIIb, and IX compared to patients without it. Patients with ataxia as the initial presenting symptom (n=5) had significantly more severe atrophy of the left cerebellar hemisphere, lobule V, VI, VIIIa, VIIIb, IX, and vermis crus I and II compared to patients with other presentations (Table 2). No significant morphometric differences were found in patients with tremor, psychiatric symptoms, and upper motor neuron signs.

Discussion

The morphometric analysis confirmed the finding of pontocerebellar atrophy in accordance with our earlier published results based on manual semiquantitative measurements.¹¹ With almost 60% of new patients, the present dataset can be considered to mostly not overlap with the previous study, so the finding of pontocerebellar atrophy was confirmed in a largely independent cohort. The morphometric analysis with DBM revealed the volume loss of the anterior, ventral and lateral thalamic nuclei, capsula interna, corticospinal tract, and parietal lobe. Complementary VBM analysis revealed the atrophy of anterior thalamic nuclei, primary sensorimotor cortex, and regions in the occipital and temporal lobes. More severe cerebellar atrophy was found in patients with dysarthria and in those initially presenting with ataxia.

The analysis of morphologic changes of the cerebellum via SUIT confirmed the finding of diffuse atrophy with more extensive impairment of the right hemisphere. The atrophy was also found in the pons, mesencephalon, and cerebral peduncles, presumably due to the atrophy of the corticospinal tract. The most affected regions were lobuli V and VI, responsible for the tongue, lips, and hand motor functions according to the somatotopic cerebellar map.¹⁸ This finding is in accordance with the dominant CNS clinical presentation in LOTS patients, i.e., cerebellar ataxia, dysarthria, and tremor.

The predominant involvement of lobulus VI is similar to the findings in patients with advanced Friedreich ataxia.²² In contrast, Cocozza S. et al. described a different cerebellar atrophy pattern in Friedreich ataxia patients, i.e., major involvement of posterior lobules, predominantly the lobule IX.²³

Patients with spinocerebellar ataxia (SCA) type 3 demonstrated atrophy involving almost all cerebellar regions except parts of lobules VIIb, VIIIa, and VIIIb bilaterally.²⁴ A morphometric study in patients with SCA type 17 revealed cerebellar atrophy predominantly localized in left lobule IX extending to the lobules VIIIA and VIIIb as well as the lobule VIIA and in the bilateral lobule V.²⁵

Yet, SCA type 10 showed a different atrophy pattern. The cerebellar degeneration was suggested to start in the posterior lobule and, in later stages, expand to the anterior lobe.²⁶ Overall, these findings indicate that specific patterns of cerebellar degeneration exist across different genetic disorders and are indicative of disease-specific pathological changes in the cerebellar structure.

In the LOTS mouse model, positive ganglioside storage was found in almost all cerebellar lobules that is compatible with the cerebellar atrophy finding in our study.²⁷ In another study of Tay-Sachs mouse model, the accumulation of gangliosides was observed specifically in the cerebellar crus I (ansiform lobule) and lobules IV and V.²⁸

In Niemann-Pick disease type C, animal studies showed the most rapid cerebellar Purkinje (Purkyně in Czech) cell loss in the anterior lobe vermis as compared to the posterior lobe, while the lobule X was the most resistant.²⁹ LOTS patients in our study and other neurodegenerative disorders also showed an identical cerebellar atrophy pattern of Purkinje cell loss with longer survival of lobule X.³⁰ Such cerebellar atrophy pattern is explained by the zebrin II expression, a specific Purkinje cell molecular marker.³¹ Purkinje cells of the anterior lobe are zebrin II negative and are preferentially vulnerable to insult, which is in contrast with predominantly zebrin II positive Purkinje cells located in the posterior lobe.^29,32

Additionally, the morphometric analysis confirmed impairment of motor cortical and subcortical regions and associated white matter tracts in LOTS patients. This finding contrasts with the results of the MRI study by Rowe O. et al., which did not show any significant differences in the motor pathways.¹² However, with only 10 patients and 7 controls examined, the latter study was inherently limited by its low statistical power. Large interindividual differences in tissue loss also exist and lead to variable findings in studies with low patient numbers. In like manner, there have been variable findings in LOTS case studies where lobar atrophy was assessed visually.^10,33 Our findings also show that corpus callosum thinning is present only in a subgroup of LOTS patients.

Atrophy of motor regions may be directly related to the disease pathophysiology as GM2 gangliosides accumulation was observed in the primary and secondary motor cortex and in the striatum in the mouse disease model.²⁸ Our study additionally showed atrophy of ventral and lateral thalamic motor nuclei and corticospinal pathway at different levels, including deep lobar white matter, internal capsule, cerebral peduncles, and pontine tegmentum, indicating complex derangements of central motor pathways. Although the clinical presentation of lower limb weakness emerges from progressive loss of anterior horn cell motor neurons, spasticity was also described in some LOTS patients. That indicates a dysfunction at the level of the upper motor neuron and the descending tracts. Despite predominant peripheral weakness, the findings of hyperreflexia with positive Babinski and/or Hoffman signs are not rare in patients with GM2-gangliosidoses.^{3–6,10,34–38}

We observed clinical symptoms of upper motor neuron lesion (hyperreflexia and/or Babinski sign) only in a minority of patients (36%), suggesting that central motor pathway abnormalities might have been subclinical in our LOTS cohort. Alternatively, in the case of severe lower motor neuron lesion, pyramidal signs may be masked by muscle weakness.

The most frequent pathogenic variant in our cohort was 805G>A which has been previously shown to be associated with adult-onset disease.³⁹ Interestingly, two patients carrying this variant in homozygous constitution (Table 1, patient number 2 and 9) had a later onset of the disease, milder neurologic symptoms, no psychiatric symptoms and mild cerebellar atrophy. Together with an earlier report⁴⁰, our data indicate that biallelic 805G>A variant leads to LOTS with a milder phenotype. Additionally, we have observed remarkably mild cerebellar atrophy in two patients harboring the 806–7 G>A variant. This pathogenic variant was previously associated with adult-onset disease and predominantly neuromuscular symptoms and may thus be associated with milder cerebellar damage.⁴¹ These observations, together with the lack of association between cerebellar volume and disease duration, indicate that cerebellar degeneration in LOTS is driven predominantly by genetic factors. Patients number 13 and 14 had profound cerebellar atrophy despite having mild symptoms lasting for less than one year. Thus, cerebellar degeneration in LOTS likely occurs early in life, while cerebellar dysfunction may be ameliorated by compensatory mechanisms for some time. Accordingly, many LOTS patients retrospectively report being clumsy during childhood, long before they seek neurologic care.⁴² The relatively weak association between the severity of ataxia and cerebellar atrophy reported previously^11,42 and also shown in our current study suggests that the occurrence of prominent ataxia in the disease course may be linked not only to progression of cerebellar degeneration but also to the loss of compensatory mechanisms or lesion of ascending spinocerebellar pathways.

Interestingly, anomalous anatomy in the primary sensorimotor cortex, perisylvian speech and language areas, and cerebellar speech areas regions often affected in LOTS, was also demonstrated in adults with persistent developmental stutters.^43,44 Stutter is described in some LOTS patients since their childhood, similarly to patient 3 in our study⁴⁵, and speech abnormalities are often one of the initial symptoms of LOTS that is usually presented in the first two decades.^42,45–47 One may speculate that abnormal development of cerebellar and cortical motor circuits underlies developmental stutter in LOTS patients. It is also possible to speculate that the observed asymmetric cerebellar atrophy may be related to speech abnormalities, as studies on functional lateralization of the cerebellum indicate the role of the right cerebellar hemisphere in language and speech processing.⁴⁸

Several study limitations need to be mentioned. First, quantitative assessment of clinical severity was not available for most patients, precluding clinical phenotype and brain-morphometry correlation analyses. Second, we analyzed MR images from three centers. Due to the low patient numbers and lack of controls from the two centers, this potential bias could not be accounted for in the statistical analysis. Despite all scanners being from the same vendor and the majority of scans (41 out of 42) being performed on the same type of 3T scanner (Siemens Skyra), the coil, software version, or MPRAGE pulse sequence parameters related differences might still alter the grey-white matter contrast, bias the segmentation of grey and white matter and influence the volumetric measurements.⁴⁹ On the other side, the effect of such biases on VBM results are smaller compared to the effects of major disease-related neurodegenerative changes.⁵⁰ In the case of volumetric measurements, it was documented that between-scanner percentual volume differences for various brain regions are below 5%.⁵¹ In comparison, between-group cerebellar volumetric differences in our study were 50% on average, i.e., an order of magnitude higher. Thus, it is unlikely that scanner-related biases significantly influenced the results of our study.

Conclusion

The quantitative whole-brain morphometric study in LOTS patients confirmed profound pontocerebellar atrophy with predominant involvement of cerebellar lobules V and VI, i.e., regions responsible for speech and motor functions. Despite the absence of clinical upper motor neuron signs in most patients, widespread morphological changes of motor cortical and subcortical regions, including associated white matter tracts, were observed, indicating abnormalities in central motor circuits in LOTS. Patients initially presenting with ataxia and those manifesting with dysarthria had more severe cerebellar atrophy, specifically in lobules V, VI, VIIIa, and VIIIb. We suggest that objective acoustic speech analysis, together with quantitative cerebellar regional morphometry, should be further studied as potential parameters for disease monitoring.

Take-home message.

Quantitative whole-brain morphometric analysis not only confirmed profound pontocerebellar atrophy affecting predominantly lobuli V and VI in late-onset Tay-Sachs patients, but also revealed widespread morphological changes of motor cortical and subcortical regions and tracts in white matter, indicating abnormalities in central motor circuits likely co-responsible for impaired speech and motor function.

Data sharing statement

Data from this will be made available on reasonable request to the corresponding author