Investigating ITM2B‐associated ataxia in a Taiwanese cerebellar ataxia cohort

Abstract

Objective

The genetic causes of a significant number of patients with cerebellar ataxia remain unsolved. Variations in the ITM2B gene, typically linked to dominantly inherited dementia, can sometimes present with cerebellar ataxia as an early symptom. This study aims to investigate the role of ITM2B variations in a Taiwanese cohort with unsolved cerebellar ataxia.

Methods

Genetic analysis of ITM2B was performed in 212 unrelated Taiwanese patients with unsolved cerebellar ataxia. Eight short tandem repeat markers flanking ITM2B were genotyped to analyze the associated haplotype. Affected carriers underwent comprehensive clinical evaluations.

Results

A heterozygous ITM2B variant, c.800G>T (p.(Ter267LeuextTer11)), was identified in three patients. Haplotype analysis demonstrated a shared haplotype linked to this variant in the three families, suggesting a founder effect. The three probands and additional three affected relatives presented with cerebellar ataxia and unsteady gait with an average onset age of 43.2 years. Most participants had no cognitive impairment at symptom onset but experienced memory decline, oculomotor disturbances, lower limb spasticity, and extensor plantar responses within 2–5 years. Magnetic resonance imaging and spectroscopy revealed progressive extension of white matter hyperintensity over periventricular and subcortical regions, subtle hippocampal atrophy, preserved cerebellar volumes, and decreased N‐acetylaspartate/creatine ratio over the vermis.

Interpretation

ITM2B mutations accounted for 1.4% of cerebellar ataxia cases in the Taiwanese cohort, with patients carrying ITM2B c.800G>T descending from a common ancestor. This study underscores the importance of considering ITM2B variations as a potential cause of cerebellar ataxia, even in the absence of dementia at the initial presentation.

Introduction

Inherited cerebellar ataxia comprises a clinically and genetically heterogeneous group of disorders characterized by slowly progressive gait ataxia, impaired coordination of upper limbs and eye movements, and difficulties in speech. ¹ Although many genes have been identified as the causes of cerebellar ataxia when mutated, genetic diagnosis remains unclear in a significant number of patients despite extensive testing. ² , ³ , ⁴ Some of these cases may involve novel pathogenic gene variations related to cerebellar ataxia that are not yet known to the scientific community. Additionally, some patients may have hereditary diseases affecting multiple neurological systems, with cerebellar dysfunction as an early symptom. Taken adrenoleukodystrophy (ALD) ⁵ and neuronal intranuclear inclusion disease (NIID) ⁶ as examples, these diseases are not traditionally classified as inherited cerebellar ataxias, but may present with cerebellar dysfunction as the main feature in the early disease stages. Understanding the prevalence and impact of these kinds of diseases in patients with cerebellar ataxia can improve diagnostic processes and treatment strategies.

Variations in the integral membrane protein 2B (ITM2B) gene are typically associated with autosomal dominantly inherited dementia, with cerebellar ataxia being one of the common concurrent symptoms. ⁷ , ⁸ , ⁹ , ¹⁰ , ¹¹ , ¹² ITM2B encodes a type II transmembrane protein, integral membrane protein 2B (BRI2), which is synthesized as an immature precursor protein (imBRI2) with 266 amino acids (Fig. 1A). After the cleavage at the C‐terminus by a furin‐like proprotein convertase, the mature form of N‐terminal BRI2 (mBRI2) and a soluble C‐terminal fragment of 23 amino acids (Bri2‐23) will be produced. ¹³ , ¹⁴ , ¹⁵ BRI2 is ubiquitously expressed, with higher levels in the cerebellum, subthalamic nucleus, substantia nigra, and hippocampus compared to the cerebral cortex. ¹³ , ¹⁴ , ¹⁵ , ¹⁶ Although not fully understood, ¹⁷ the physiological function of BRI2 may be involved in apoptosis, mitochondrial homeostasis, ¹⁴ , ¹⁸ regulation of excitatory synaptic glutamate transmission, ¹⁹ , ²⁰ inhibition of amyloid‐β protein precursor (APP) processing, and modulating of amyloid‐β aggregation. ²¹ Patients with pathogenic ITM2B variants may have dementia accompanied by hippocampal neurofibrillar degeneration and widespread amyloid deposits in parenchyma and vasculature. These amyloid fibrils are composed of a 34 amino acid subunit (ABri or ADan), derived from an 11‐amino acid elongated mutant C‐terminal fragment of the BRI2 protein. ¹⁰ , ²² , ²³

Figure 1.

Mutational analysis of the ITM2B gene. (A) Schematic illustration of structures of the ITM2B gene (upper panel) and the BRI2 protein (lower panel). The currently known pathogenic ITM2B variants associated with familial dementia are labeled. The variant identified in this study is labeled in red. (B) Sanger sequence traces revealing the ITM2B c.800G>T variant, which putatively switches the stop codon to Leucine, and extends the open reading frame with additional 11 amino acids, p.(Ter267LeuextTer11). The altered nucleotide is highlighted with a gray background. BRICHOS, BRI2, chrondromodulin‐I and surfactant protein C domain; FBD, familial British dementia; FCD, familial Chinese dementia; FDD, familial Danish dementia; FKD, familial Korean dementia; ICD, intracellular domain; TCPs, C‐terminal peptides; TMD, transmembrane domain.

To date, three missense variants and one insertion/deletion variant of ITM2B have been linked to autosomal dominant inherited dementia. These include c.799T>A in familial British dementia (FBD), ⁸ , ²² , ²⁴ c.800G>T in familial Chinese dementia (FCD), ¹² c.800G>C in familial Korean dementia (FKD), ²⁵ and c.795_796insTTTAATTTGT in familial Danish dementia (FDD) (Fig. 1A). ⁷ , ⁹ , ¹⁰ , ¹¹ These variants all extend the original reading frame of ITM2B, resulting in a longer precursor protein with 277 amino acids and a larger C‐terminal fragment, which are considered to be pathogenic due to their amyloidogenic properties and their presence in brain aggregates. ⁸ , ⁹ , ¹² , ²⁵ Patients with these variants present with young‐onset amnesic dementia and progressive leukoencephalopathy, often with cerebellar ataxia manifesting after the onset of cognitive symptoms. ⁷ , ⁸ , ⁹ , ¹⁰ , ¹¹ , ¹² , ²⁵

Given that cerebellar ataxia has been reported as a common manifestation of ITM2B‐associated dementia, it is intriguing to know whether ataxia could be the major symptom, at least in the early disease stage, in some patients. In this study, we aimed to investigate the role of ITM2B variations in a Taiwanese cohort of unsolved cerebellar ataxia. We further characterized the clinical features of the ITM2B c.800G>T (p.(Ter267LeuextTer11)) variant in patients from three pedigrees and demonstrated the founder effect of this variant.

Methods

Study subjects

We enrolled 212 unrelated patients with unsolved cerebellar ataxia from the Neurology Service of Taipei Veterans General Hospital. This cohort included 51 patients with autosomal dominant inheritance, 60 with autosomal recessive inheritance, and 101 with apparently sporadic cerebellar ataxia. Patients were categorized into dominantly inherited group if they had one or more affected family members in two or more generations, or recessively inherited group if they had affected siblings or consanguineous parents. Patients with idiopathic slowly progressive cerebellar ataxia, negative family history, and no autonomic dysfunction were assigned to the apparently sporadic cerebellar ataxia group. The genetic cause of ataxia in these patients remained unknown after excluding mutations responsible for spinocerebellar ataxia (SCA) 1, 2, 3, 6, 7, 8, 10, 12, 17, 19/22, 23, 26, 27, 28, 31, 35, 36, 47, dentatorubral‐pallidoluysian atrophy (DRPLA), and Friedreich's ataxia (FRDA). We also excluded patients with other degenerative or acquired causes of cerebellar ataxia, including multiple system atrophy (MSA), viral infection, alcohol or drug intoxication, brain tumor, and paraneoplastic syndrome. Peripheral blood samples were from participants after obtaining written informed consent. This study was approved by the Institutional Review Board of Taipei Veterans General Hospital.

Mutational analysis

Genomic DNA was extracted from peripheral blood. The coding exons and their flanking regions of ITM2B were analyzed by PCR amplification and Sanger sequencing with intronic primers using the Big Dye 3.1 dideoxy terminator method (Applied Biosystems, Foster City, CA, USA) on an ABI Prism 3700 Genetic Analyzer (Applied Biosystems). The amplicon sequences were compared with the reference ITM2B coding sequence (NM_021999.5).

Clinical evaluations

The probands and their affected family members were thoroughly evaluated by history taking and neurological examinations. Age at onset was defined as the age when the earliest symptoms, either cognitive impairment or truncal or appendicular ataxia, first appeared. The clinical severity of ataxia was measured using Scale for the Assessment and Rating of Ataxia (SARA, score range 0–40), a validated severity scale for cerebellar ataxia, measuring eight items designed to rate functional impairment of gait, stance, sitting, speech, finger‐chase test, nose‐finger test, fast alternating movements and heel‐shin test. ²⁶ Cognitive function was further assessed with the Mini‐Mental Status Examination (MMSE, score range 0–30). ²⁷ Brain and spine magnetic resonance imaging (MRI) were performed with a 1.5‐T system (Signa EXCITE, GE Medical Systems, Milwaukee, WI). For the single‐voxel magnetic resonance spectroscopy (MRS), the volume of interest was set in bilateral cerebellar hemispheres and cerebellar vermis. We calculated peak areas for N‐acetylaspartate (NAA) at 2.02 parts per million (ppm), creatine (Cr) at 3.03 ppm, and the metabolite intensity ratio (NAA/Cr ratio) of both cerebellar hemispheres and vermis using FuncTool software (GE Healthcare, Milwaukee, WI). ²⁸ We also conducted nerve conduction studies (NCS) and electromyography (EMG), and performed motor evoked potentials (MEPs) and somatosensory evoked potentials (SSEPs).

Haplotype analysis of the ITM2B c.800G>T variant

To explore the possible founder effect of the ITM2B c.800G>T variant, haplotype analysis was performed in the three pedigrees harboring the c.800G>T variant. We selected polymorphic short tandem repeat (STR) markers flanking the ITM2B gene on chromosome 13 by performing a targeted search through 14 in‐house PCR‐free whole genome sequencing (WGS) data using ExpansionHunter (v5.0.0). These 14 WGS datasets are derived from individuals of four Taiwanese families without disease‐associated variants in ITM2B. Targeted STR markers were selected based on the STR catalogs released by Illumina (https://github.com/Illumina/RepeatCatalogs). After prioritizing 2‐mer STRs with diverse alleles among these 14 individuals, we selected eight STR markers in the region of chr13:48140856‐48499111 (GRCh38), namely MK1F to MK8F in numerical order. The first four markers are centromeric to ITM2B, and the rest fours are telomeric (Table S1).

Results

Identification of ITM2B variants in the cohort of unsolved cerebellar ataxia

Mutational analysis of ITM2B in the cerebellar ataxia cohort identified a heterozygous extension variant, c.800G>T (p.(Ter267LeuextTer11)) (Fig. 1B). This ITM2B variant was present in three unrelated patients (Fig. 2), accounting for 1.4% (3/212) of the Taiwanese cerebellar ataxia cohort. The variant alters the second nucleotide of the ITM2B stop codon, changes it to leucine, and results in an extended protein product of 277 amino acids.

Figure 2.

The three pedigrees carrying the ITM2B c.800G>T variant in this study. (A) The pedigree A, (B) pedigree B, and (C) pedigree C. The probands are indicated by arrows. The “M” represents a mutant ITM2B allele and the “W” means a wild type allele. The squares and circles denote males and females. The filled and open symbols represent affected and unaffected members, respectively. The colors filled represent the clinical features, in which black color suggestive of cerebellar ataxia, light brown representing cognitive impairment, dark brown for spasticity or extensor plantar reflex, and blue color standing for visual or auditory impairment. The dotted symbols indicate the asymptomatic carriers of the mutant ITM2B allele. A slash indicates a deceased individual.

Clinical information of the patients carrying the ITM2B c.800G>T variant

The probands of the three families and another three symptomatic carriers from family A (A‐II‐6, A‐III‐6 and A‐III‐7) were evaluated (Fig. 2). The average age at onset was 43.2 ± 12.2 years (range 23–60 years). All the six patients carrying the ITM2B c.800G>T variant initially presented with unsteady gait, and four of the six patients developed impaired short‐term memory 2–5 years later. Other manifestations included dysarthria or dysphagia (5 of 6), lower limbs spasticity or other upper motor neuron signs (5 of 6), eye movement disturbance (4 of 6), hearing impairment (2 of 6), and decreased visual acuity (1 of 6). The clinical characteristics of these six patients, along with one historical case carrying the ITM2B c.800G>T variant, ¹² are summarized in Table 1.

Table 1.

Clinical characteristics of patients with ITM2B c.800G>T variant from the present study and literature.

Patient	A‐II‐6	A‐III‐1	A‐III‐6	A‐III‐7	B‐II‐2	C‐II‐1	FCD12
Sex	Female	Female	Male	Male	Female	Female	Male
Age at onset, years	60	45	23	46	47	38	40
Age at test, years	64	48	48	47	49	40	44
Initial symptoms	Unsteady gait	Unsteady gait	Unsteady gait	Unsteady gait, dysarthria	Unsteady gait	Unsteady gait, poor memory	Unsteady gait, limbs stiffness
Ataxia	+	+	+	+	+	+	+
Saccadic pursuit	+	−	+	−	+	+	−
Ataxic dysarthria	+	+	+	+	+	−	+
DTR (knee) a	+++	++	+++	+++	+++	+++	+++
Plantar response	Flexor	Flexor	Flexor	Extensor	Flexor	Extensor	Extensor
Legs spasticity	−	−	−	−	+	+	+
Urinary incontinence	+	−	−	−	+	−	−
Hearing or visual impairment	−	+	−	−	+	−	+
Memory impairment	+	+	−	−	+	+	+
SARA	25	8	7	14.5	10	8	NA
MMSE	NA	24	NA	NA	26	28	26

+: presence, −: absence.

DTR, deep tendon reflex; FCD, familial Chinese dementia; MMSE, mini‐mental state examination; NA, not available; SARA, Scale for the Assessment and Rating of Ataxia.

^a National Institute of Neurological Disorders and Stroke (NINDS) Scale for tendon reflex: 0−++++.

Pedigree A

The proband (A‐III‐1) experienced progressive gait unsteadiness since age of 45 years. She later developed difficulty in speech and handwriting, poor short‐term memory, and hearing impairment in the right ear. At age 48, neurological examination revealed ataxic dysarthria, limbs and truncal ataxia, and impaired right ear hearing, but normal muscle strength and deep tendon reflexes. Her SARA score was 8 and MMSE score was 24. Brain MRI revealed preserved cerebellar volume (Fig. 3A,B), mild hippocampal atrophy with medial temporal lobe atrophy score of 1 (Fig. 3C), and multiple small white matter hyperintensity (WMH) dispersed in the bilateral subcortical regions (Fig. 3D). MRS showed a reduced NAA/Cr ratio in the cerebellar vermis (0.68 compared to 0.9 ± 0.11 in healthy controls), ²⁸ indicating mild neuronal dysfunction (Fig. 3E,F). The clinical manifestations of patients A‐II‐6, A‐III‐6, and A‐III‐7 from pedigree A were similar to the proband. However, hyperreflexia of knee jerks and extensor plantar responses in patient A‐III‐7, as well as urinary incontinence and dysphagia in patient A‐II‐6, were absent in the proband. Brain MRI and MRS of patients A‐II‐6 and A‐III‐6 also showed multiple small focal WMH and reduced NAA/Cr ratio in the cerebellar vermis.

Figure 3.

The neuroimages of proband A‐III‐1 in pedigree A. (A and B) The T1‐weighted image of MRI denotes no apparent atrophy of the cerebellum. (C) The T1‐weighted coronal view image show mild degree of hippocampal atrophy. (D) The fluid‐attenuated inversion recovery (FLAIR) axial view image demonstrates multiple white matter hyperintensity scattered over the subcortical regions. (E and F) The MR spectroscopy of the vermis and right cerebellar hemisphere show decreased metabolite intensity ratio (NAA/Cr ratio) in the vermis, but preserved in the cerebellar hemisphere. MR spectra is labeled with arrow (Creatine) and dashed arrow (N‐acetylaspartate) arrow. Cr, Creatine; NAA, N‐acetylaspartate.

Pedigree B

The proband (B‐II‐2) began to have progressive gait instability at age of 47 years, followed by slurred speech, hand tremors, impaired visual and hearing acuity, and slow response to the environment 1 year after the disease onset. Neurological examination at 49 revealed saccadic pursuit, ataxic dysarthria, appendicular ataxia, truncal titubation, lower limbs spasticity and hyperreflexia. She exhibited ataxic and spastic gait and could barely walk without assistance despite preserved muscle strength. Bilateral sensorineural hearing loss was diagnosed with an average hearing threshold of 70 dB in the left ear and 52 dB in the right ear. Visual impairment and optic nerve atrophy were noted, with visual acuity of 20/40 in the right eye and 20/50 in the left eye. Her SARA score was 10 and MMSE was 26. She later developed slowly progressive cognitive decline, leading to persistent confusion and bedridden status at age 56. Brain MRI conducted at three and 9 years after disease onset revealed a progressive extension of WMH from normal pattern to a more confluent and widespread involvement (Fig. 4A–D). However, no hippocampal atrophy was observed despite severe dementia (Fig. 4E). Her sister (B‐II‐1) also had unsteady gait at age 47, but the genetic status was unknown.

Figure 4.

The neuroimages of proband B‐II‐2 (A–E) and proband C‐II‐1 (F–J). The FLAIR images of proband B‐II‐2 performed at age 50 (disease duration of 3 years) (A andB) and 6 years later (C and D) reveal progressive extension of the white matter hyperintensity over subcortical and periventricular regions. (E) The T1‐weighted coronal view image with Gadolinium injection of proband B‐II‐2 conducted 9 years after the disease onset shows no hippocampal atrophy. (F and G) The FLAIR images of proband C‐II‐1 recorded after disease onset for 6 years shows multifocal small white matter hyperintensity. (H) The T1‐weighted axial view image shows no hippocampal atrophy. (I and J) The MR spectroscopy of the vermis and right cerebellar hemisphere shows decreased level of NAA/Cr ratio in the vermis. MR spectra is labeled with arrow (Creatine) and dashed arrow (N‐acetylaspartate) arrow. Cr, Creatine; NAA, N‐acetylaspartate.

Pedigree C

The proband (C‐II‐1) experienced progressively unsteady gait beginning at age 38 years, followed by difficulties in fine hands movements and a decline in short‐term memory. Neurological evaluation at age 40 showed saccadic pursuit, horizontal nystagmus, limbs dysmetria, hyperreflexia and extensor plantar responses, and difficulty in tandem gait. Between age 40 and 43, her SARA score increased from 8 to 10, and MMSE score declined from 28 to 23. Brain MRI showed multiple small WMH in bilateral subcortical and periventricular regions (Fig. 4F,G), while the volume of hippocampus remained preserved (Fig. 4H). MRS indicated mild neuronal dysfunction in the cerebellar vermis by reduced NAA/Cr ratio (Fig. 4I,J). Her father (C‐I‐1), whose genetic status was unknown, also suffered from unsteady gait starting at age 50.

Haplotype analysis

Haplotype analysis was conducted on nine individuals, including six affected patients, one asymptomatic carrier, and two unaffected individuals, from the three families harboring the ITM2B c.800G>T variant (Fig. 5). Pedigree A and B shared an identical haplotype across all eight STR loci linked to the ITM2B c.800G>T variant (2–2‐1‐1‐c.800G>T‐1‐5‐1‐1) (Table 2). For pedigree C, although the exact phase of the haplotype could not be definitively determined, the evidence strongly suggested that the patient in Family C also shared this haplotype (Table 2). These findings indicate that the ITM2B c.800G>T variant in these three families likely originates from a common ancestor.

Figure 5.

Haplotype analysis of the eight short tandem repeat (STR) markers flanking the ITM2B gene in three pedigrees carrying the ITM2B c.800G>T variant. The squares and circles denote males and females. The filled and open symbols represent affected and unaffected members. Dotted symbols indicate asymptomatic gene carriers and the numbers in the squares or circles are the numbers of unaffected individuals. A slash represents deceased members.

Table 2.

Haplotypes linked to ITM2B c.800G>T mutation in the three pedigrees.

Patient	Centromeric haplotype markers				ITM2B	Telomeric haplotype markers
	MK1F	MK2F	MK3F	MK4F	Mutation	MK5F	MK6F	MK7F	MK7F
A‐II‐6	2	2	1	1	c.800G>T	1	5	1	1
A‐III‐1	2	2	1	1	c.800G>T	1	5	1	1
A‐III‐6	2	2	1	1	c.800G>T	1	5	1	1
A‐III‐7	2	2	1	1	c.800G>T	1	5	1	1
B‐II‐2	2	2	1	1	c.800G>T	1	5	1	1
B‐III‐2	2	2	1	1	c.800G>T	1	5	1	1
C‐II‐1	2	1/2 a	1	1	c.800G>T	1	5/7 a	1	1

^a Alleles with an unknown phase are separated with a slash, indicating the two possible haplotypes of the markers.

Discussion

In this study, we identified a rare, heterozygous ITM2B variant, c.800G>T (p.(Ter267LeuextTer11)), in three Taiwanese families from a cohort of unsolved cerebellar ataxia. This variant has been previously reported in FCD. ¹² Although the genetic variations slightly differ, the protein alteration caused by the ITM2B c.800G>T in the three Taiwanese families is almost identical to those seen in FBD, FDD, and FKD (Fig. 1A), supporting its pathogenicity. ⁸ , ¹⁰ , ¹¹ , ²⁵ Our findings have several important implications.

First, ITM2B variations are associated with a wide phenotypic spectrum, extending beyond dementia to include cerebellar ataxia as an early disease feature. While ITM2B variations have been linked to autosomal dominant dementia in British, Danish, Chinese, and Korean families, ⁷ , ⁸ , ⁹ , ¹⁰ , ¹¹ , ¹² , ²⁵ our study is the first to point out that patients carrying an ITM2B variant could primarily manifest with cerebellar ataxia. The patients in our study initially presented with symptoms such as unsteady gait, truncal ataxia, dysarthria, and restricted eye movement, with appendicular ataxia developing later. Upper motor neuron signs, such as lower limb spasticity and extensor plantar reflex, were also common. Although subjective memory decline was reported early in some patients, there was no significant evidence of cognitive impairment at recruitment based on the MMSE scores. Interestingly, all six patients in our study, along with the FCD proband, shared the same ITM2B variant and presented similar phenotypes, including a variable combination of cerebellar ataxia, memory impairment, and spastic paraparesis (Table 1). ¹² This suggests a genotype–phenotype correlation in ITM2B variations.

Second, ITM2B variations appear to be an uncommon cause of hereditary cerebellar ataxia in the Taiwanese cohort, occurring in 1.4% of cases with unsolved cerebellar ataxia. To date, no other studies have directly linked ITM2B variations to hereditary cerebellar ataxia, making its overall prevalence unknown. ITM2B variations, typically associated with familial dementia, might be overlooked in patients presenting solely with hereditary ataxia, especially in the early stages of the disease. Previous reports have suggested that the combination of young‐onset amnesic dementia and cerebellar ataxia, sometimes accompanied by spastic paraparesis is characteristic for ITM2B‐associated diseases. ⁸ , ⁹ , ¹⁰ , ¹¹ , ¹² , ²⁵ Therefore, young‐onset dementia and spasticity could serve as important diagnostic clues to distinguish ITM2B‐associated disease from other unsolved cerebellar ataxia. However, it is crucial to consider other differential diagnoses. For instance, among inherited ataxias, dementia has been reported in SCA2, SCA17, and DRPLA, ²⁹ while spasticity may occur in SCA1, SCA3, and late‐onset FRDA. ³⁰ Additionally, mutations in other genes can also present with various combinations of ataxia, dementia, and spastic paraparesis, such as PSEN1 (young‐onset Alzheimer's disease), ³¹ , ³² C9orf72 (frontotemporal dementia), ³³ PRNP (Gerstmann‐Sträussler‐Scheinker syndrome), ³⁴ CTP27A1 (cerebrotendinous xanthomatosis), ³⁵ TTPA (ataxia with vitamin E deficiency), ³⁰ ANO10 (autosomal recessive cerebellar ataxia type 3), ³⁰ SPG7 (hereditary spastic paraplegia type 7), ³⁶ and FMR1 (fragile X tremor ataxia syndrome). ³⁷ Our findings suggest that ITM2B should be considered as a potential causal gene in patients with hereditary cerebellar ataxia, especially when common hereditary ataxias have been ruled out. This is particularly relevant in cases accompanied by limb spasticity or where young‐onset dementia is present.

Third, the early dysfunction of the cerebellar vermis may be a distinguished feature in patients with ITM2B‐associated cerebellar ataxia. In our study, gait and truncal ataxia developed prior to limbs ataxia in the patients carrying the ITM2B variant. MRI studies showed preserved cerebellum volume without atrophy, but MRS studies revealed neuronal dysfunction at the cerebellar vermis evident by a reduced NAA/Cr ratio. These findings, along with the clinical profiles of our cases and a previously reported Chinese proband, ¹² suggest early involvement of the spinocerebellum and vestibulocerebellum. Unlike degenerative cerebellar ataxias like SCA and MSA that are typically influencing the entire cerebellum, ³⁸ , ³⁹ predominant vermis involvement is more common in acquired ataxias such as immune‐mediated cerebellar ataxias and alcoholic cerebellar degeneration, ⁴⁰ , ⁴¹ , ⁴² as well as some rare autosomal recessive cerebellar ataxias. ⁴³ , ⁴⁴ MRS data of ITM2B‐associated ataxia has never been investigated before, and our study is the first to show earlier dysfunction of cerebellar vermis compared to the hemispheres in affected patients. However, due to the rarity of the cases, further studies are necessary to confirm whether this is a universal feature of ITM2B‐associated ataxias.

Additionally, we identified a common haplotype linked to the c.800G>T variant, spanning approximately 1 Mb across all eight STR loci associated with ITM2B in the three Taiwanese families. This finding suggests that carriers of the c.800G>T variant in Taiwan may share a common ancestor. While the clinical severity is not mild, the later onset of the disease, often occurring after reproductive age, may have allowed the variant to be transmitted across generations.

In conclusion, this study delineated the clinical features of ITM2B‐associated cerebellar ataxia in three unrelated families, accounting for 1.4% of the Taiwanese cohort of unsolved cerebellar ataxia. ITM2B‐associated cerebellar ataxia typically begins with gait and truncal ataxia, progressing to limbs ataxia, young‐onset amnestic dementia, and spastic paraparesis. The three families shared a common ITM2B variant, c.800G>T (p.(Ter267LeuextTer11)), with evidence of a founder effect in Taiwan. Our findings highlight the clinical heterogeneity of ITM2B variations and underscore the importance of considering ITM2B variations in patients with cerebellar ataxia, even in the absence of dementia at the initial presentation.

Author Contributions

Shih‐Yu Fang, Cheng‐Tsung Hsiao, Yi‐Chu Liao, Yi‐Chung Lee contributed to conception and design of the study. Shih‐Yu Fang, Cheng‐Tsung Hsiao, Kang‐Yang Jih, Yu‐Sheun Tsai, Kuan‐Lin Lai, Cheng‐Ta Chou contributed to acquisition and analysis of data. Shih‐Yu Fang, Kuan‐Lin Lai, Yi‐Chu Liao, Yi‐Chung Lee contributed to drafting of the manuscript.

Funding Information

This study was supported by the grants from National Science and Technology Council, Taiwan (112‐2314‐B‐075‐034‐MY3;113‐2634‐F‐039‐001), “Center for Intelligent Drug Systems and Smart Bio‐devices (IDS²B)” and “Brain Research Center”, National Yang Ming Chiao Tung University from The Featured Areas Research Center Program within the framework of the Higher Education Sprout Project by the Ministry of Education in Taiwan.

Conflict of Interest

The authors disclose no conflicts of interest.

Supporting information

Table S1. Short tandem repeats markers for the haplotype analysis.

Funding Statement

This work was funded by National Science and Technology Council, Taiwan grants 112‐2314‐B‐075‐034‐MY3 and 113‐2634‐F‐039‐001; Center for Intelligent Drug Systems and Smart Bio‐devices (IDS2B); Brain Research Center.

Data Availability Statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.