Clinical and Structural Characteristics of NEU1 Variants Causing Sialidosis Type 1

Patient description and clinical evaluation

A family with a history of nonconsanguineousism was included in this study. The patients’ clinical history was collected and documented by two experienced physicians. The study was approved by the Ethics Committee of the Chinese PLA General Hospital (approval number: S2022-687-01) and followed the principles of the Declaration of Helsinki. Informed consent was obtained from each participant. The patients underwent a series of ophthalmic examinations, including visual acuity and visual field measurements, color fundus photography, optical coherence tomography (OCT), visual evoked potential (VEP) measurements and fundus fluorescein angiography (FFA). The neurological evaluations were conducted and analyzed by two neurologists. Brain magnetic resonance imaging (MRI), electroencephalography (EEG), electromyography (EMG) and cerebral 18F-fluorodexyglucose-positron emission tomography (¹⁸F-FDG-PET) were performed according to standard procedures. General blood tests and urinary amino acid, organic acid and fatty acid tests were performed on the proband.

DNA extraction and whole-exome sequencing

Blood samples were collected from all family members. Genomic DNA was extracted using a QIAamp 96 DNA Blood Kit (Qiagen, Hilden, Germany). The DNA concentrations were measured with a microvolume spectrophotometer (Pultton P100/P100+; Pultton Technology Ltd., San Jose, CA, USA). The DNA libraries were constructed using the KAPA Library Preparation Kit (Kapa Biosystems, Wilmington, MA, USA). The exomes were captured by the Agilent SureSelectXT Target Enrichment System (Agilent Technologies, Inc., Santa Clara, CA, USA). The captured DNA samples were sequenced by Illumina NovaSeq (Illumina, San Diego, CA, USA). After the sequencing data were evaluated by Illumina Sequence Control Software (Illumina), the data were read.

The raw genome sequences were aligned to the GRCh37/hg19 human reference genome. The proband’s phenotypic traits were defined according to Human Phenotype Ontology terms, and the variants were confirmed using a disease database (OMIM, ClinVar, gnomAD, and an in-house database). Pathogenicity and disease associations were evaluated according to the American College of Medical Genetics and Genomics guidelines, and bioinformatics analysis was performed. The variants identified in this study were confirmed by Sanger sequencing. Family verification was conducted based on the obtained results.

Sialidase activity test

The complementary DNA (cDNA) of NUE1 (NM_000434) and PPCA (NM_000308) were cloned and inserted into the pIRES-hrGFP-1a and pcDNA3.1- Myc-HIS vectors, respectively. The sialidase is catalytically active only when NEU1 is bound to PPCA. The S182G (c.544A>G) and V143E (c.428T>A) variants were introduced into the NEU1 cDNA. COS-7 cells were cultured as previously described. Then, the cells were transfected with NEU1 (wild-type or mutant construct) and PPCA at a ratio of 1:1 using Lipofectamine 2000 (Cat. No. 11668030; Thermo Fisher Scientific, Inc., Shanghai, China) [9]. After 24 hours of transfection, the cells were harvested for further assays. Sialidase activity assays were performed according to the instructions of the Neuraminidase Assay Kit (Merck KGaA, Darmstadt, Germany). The experiments were performed for at least 6 separate clones.

Structural analysis of the NEU1 protein

Structural models of human wild-type sialidase and its mutants were built using the I-TASSER server (https://seq2fun.dcmb.med.umich.edu//I-TASSER/). This resource hierarchically constructs full-length models by iteratively reassembling structure fragments extracted from the threading templates. The parameters in I-TASSER were set to the default values. The conservation of multiple protein sequences was analyzed using the Constraint-based Multiple Alignment Tool. Three-dimensional models of sialidase activity and the polar contacts of mutant residues were analyzed by the PyMOL Molecular Graphics System (version 2.3.0; Schrödinger, Inc., Shanghai, China). The system energies before and after the variation were calculated using HyperChem (V8.0.8; Hypercube, Inc., Gainesville, FL, USA), and the minimum energy state was obtained. Furthermore, to compare the changes in the flexibility of key residues between wildtype and mutant proteins, the flexibility of each residue of interest was calculated from the Gaussian network model (GNM), which is based on a harmonic potential and has been proven to be a reliable method for reproducing the intrinsic dynamics of biomacromolecules. The total internal potential energy of a network of N nodes can be written as

where γ is the harmonic force constant of the springs, the column vector ∆R represents the fluctuation of the N nodes, the superscript T denotes the transpose, E is the 3 × 3 identity matrix, ⊗ is the matrix direct product and Γ is the N × N symmetric Kirchhoff matrix.

The mean-square fluctuation of the ith residue can be expressed as

where k_B is the Boltzmann constant and T is the absolute temperature. In GNM, as slow modes contribute the majority to residue fluctuations, we calculated the relative residue fluctuations (with 3k_BT/γ ignored) from the first 10 motional modes as the residue flexibility.

Literature review and statistical analysis

We searched PubMed using the term “NEU1” and extracted all clinical information and the results of functional verification, if any. The last search date was January 10, 2023. Statistical tests and plotting were performed using R software 3.6 (R Foundation for Statistical Computing, Vienna, Austria). The chi-square test was used to analyze the genotype-phenotype correlation. The statistical analysis of clinical symptoms and age at onset was performed with the Kruskal–Wallis test. A significance criterion of p < 0.05 was applied.

Clinical description and genetic analysis

The data of nine consanguineous individuals, including two affected members and seven unaffected members, were collected (Figure 1A). The proband (III-2) developed involuntary shaking of her left hand at age 13 and then gradually developed myoclonus in both upper extremities and dystonia in both lower extremities and began to fall frequently. Since the onset of the disease, she has experienced several seizures, including generalized tonic‒clonic seizures. EEG monitoring after the attack revealed multiple spikes, slow-wave complexes, and slow waves. Neurological examination revealed myoclonus, dysarthria, dystonia and an ataxic gait. Ophthalmic examination revealed impaired vision. A color fundus photograph confirmed the presence of macular cherry-red spots. OCT indicated bilateral global thickening of the peripapillary retinal nerve fiber layer. FFA excluded central retinal artery occlusion. The brainstem auditory-evoked potential and somatosensory-evoked potential (SEP) were normal, while the VEP showed prolonged peak latency and a reduced P100 amplitude. The proband’s EMG and brain MRI results were normal. ¹⁸F-FDG-PET analysis revealed normal cerebral glucose metabolism. Blood and cerebrospinal fluid examinations were unremarkable, and the urine sialic acid concentration was within the normal range.

Figure 1.

Family pedigree and genetic analysis. A: Pedigree of the family with NEU1 gene variants. The blue or red bars represent paternal or maternal chromosomes, respectively. The black shading indicates the affected patients. All members underwent genetic testing, and patients with NEU1 variants were flagged. B: Sanger sequencing electropherograms. Patients III-1 and III-2 were confirmed to have sialidosis type 1 with compound heterozygous variants. C: Conservation analysis of the mutant protein. V143E and S182G are evolutionarily conserved.

The patient’s older brother (III-1) had gait instability and experienced falls starting at the age of 14 years and developed seizures one year later. Ocular fundus photography also revealed cherry-red spots. The other examination results were similar to those of the proband. Both III-1 and III-2 had normal births and developmental milestone histories. Over the years, the patients have been treated with several medications, but only mild and transient benefits have been reported. During hospitalization, the patients were treated with valproate, clonazepam, topiramate, and levetiracetam in various combinations and at various doses, resulting in a reduced convulsive seizure frequency and improved myoclonus. After a few days, the patients could walk independently and perform activities of daily living.

Blood samples from the patients and their parents were examined via whole exome sequencing. The genetic test results confirmed the diagnosis of sialidosis type 1. Sanger sequencing confirmed the presence of compound heterozygous c.544A>G (p.S182G) and c.428T>A (p.V143E) variants in the NEU1 gene (Figure 1B). Among these variants, p.S182G has previously been reported as a common pathogenic variant in the Chinese population, whereas p.V143E is a previously unreported variant newly discovered in this study. For the two missense variants, interspecific protein alignments were performed to evaluate the conservation of the altered residues (Figure 1C). We found that p.S182G and p.V143E were highly conserved among several representative vertebrate species, suggesting their important roles in protein function.

Structural and functional analyses of variants in the NEU1 protein

To determine the structural changes in the S182G and V143E mutant proteins, we selected three positive controls (R78C, Y370C, and G328S) for our calculations. Both the structure of the mutant protein and the sialidase activity are related to the sialidosis phenotype. These representative variants were selected for the following reasons: 1) R78C was the binding site of the protein; 2) Y370C was the active site, with a sialidase activity less than 10% of that of the wild-type protein [3]; 3) G328S and S182G were located in the surface region and had 40%–50% and 20%–40% sialidase activity, respectively [10]; and 4) R78C and Y370C were associated with only sialidosis type 2, while G328S and S182G were associated with only sialidosis type 1. As previously described, S182G was considered a positive control for exploring structural changes and predicting the severity of pathogenesis in V143E patients.

First, the energies of the optimal structures were calculated for the wild-type protein and mutant forms. The energy of the wild-type protein was -4122.16 kcal/mol, and the energy values of the S182G and V143E variants were -4120.52 kcal/mol and -4155.18 kcal/mol, respectively. In other words, the S182G variant increased the protein energy by 1.64 kcal/mol, while the V143E variant decreased the protein energy by 33.02 kcal/mol. The calculated energies of R78C, Y370C, and G328S increased by 273.68 kcal/mol, 72.40 kcal/mol, and 71.83 kcal/mol, respectively. In general, for the same protein, the lower the energy of the system is, the more stable the structure is, indicating decreased pathogenicity.

In general, variants in key residues (binding sites or active sites) often affect proteins most significantly and directly, causing a gain or loss of function [11]. To explore the influence of amino acid residue changes before and after variation on protein structural stability and the flexibility of key residues, we performed structural prediction of wild-type NEU1 and the V143E, S182G, G328S, R78C and Y370C mutants by I-TASSER. The conjugate gradient method was used to optimize the modeled structures for energy minimization. Subsequently, the flexibility values of key residues of the wild-type and mutant proteins were calculated using the GNM model and compared (Figure 2). The key residues (R78, R97, D103, E264, R280, R341, Y370, and E394) of NEU1 were defined based on information in the UniProt database.

Figure 2.

Flexibility of key residues in wild-type (WT) NEU1 and different mutant forms. A: Key residues refer to the catalytic and binding sites of the NEU1 protein defined in the UniProt database. V143E showed flexibility between variants associated with sialidosis type 1 (S182G and G328S) and type 2 (R78C and Y370C). B: Locations of the variants that were analyzed in this study.

Compared with the wild-type protein, the R78C mutant protein had reduced flexibility at all key residues, with the flexibility of R97 being reduced by 91%. Except for R341 and Y370, which lost 40% and 17% of their flexibility, respectively, the flexibility of all the key residues decreased by more than 60%. Furthermore, the flexibility of key residues of the Y370C mutant was also reduced, with the flexibility of R97 being reduced by 74%. Except for R341 and Y370, which lost 20% and 17%, respectively, of their flexibility, the flexibility values of all the key residues decreased by more than 50%. However, although the flexibility of key residues of V143E was also reduced, R97 and D103 were the only key residues whose flexibility values were reduced by more than 50% (with reductions of 53% and 69%, respectively), and the flexibility of Y370 did not change. Moreover, for S182G, the flexibility values of R97, D103, and R341 increased by 22%, 69%, and 20%, respectively; those of R78, E264, R280, and E394 decreased by 11%, 61%, 31%, and 46%, respectively; and the flexibility of Y370 did not change. Similarly, for the G328S variant, the flexibility values of R78, R97, and R280 increased by 25%, 39%, and 14%, respectively; those of E264, R341, and Y370 decreased by 33%, 40%, and 17%, respectively; and those of D103 and E394 were almost unchanged (increased by 3% and 8%).

The reduced flexibility of the protein structure reflected the instability of the system, which may be related to the polar contacts of the mutant residues (Figure 3). As we found by PyMOL, S182 formed four polar contacts with T183 and W173, while the S182G mutant formed only one polar contact with W173. The V143E variant did not change the number of polar contacts with F152. Regarding the active site, R78 formed a total of nine polar contacts with five adjacent amino acids (E95, S202, N398, E402, and E394), while R78C interacted only with E95, forming two polar contacts. Regarding the binding site, Y370 interacted with both E264 and E394 through three polar contacts, while Y370C interacted only with E394 through one polar contact. Notably, E394 and E264 were also active sites. G328 interacted with no residues in the wild-type protein and formed two polar contacts with S327 and F350. Therefore, S182G, V143E, and G328S may cause milder protein function damage than R78C and Y370C. Interestingly, both R78C and Y370C were associated with sialidosis type 2, while S182G, V143E, and G328S were pathogenic variants of type 1 sialidosis, suggesting that local polar contacts and the altered flexibility of variants may be associated with the clinical phenotype of sialidosis.

Figure 3.

Conformational changes caused by NEU1 variants. Details of the regions adjacent to the S182 and V143 residues (this study), as well as R78 (binding site), Y370 (active site) and G328S. Hydrogen bond changes are shown as yellow dashed lines. The changes in the energy (kcal/mol) of the system before and after the variation are shown at the lower right. The calculated energies of S182G, R78C, Y370C and G328S increased, indicating an unstable structure of the mutant protein.

In summary, we found that the newly identified V143E variant showed decreased protein energy, an unchanged number of polar contacts, and flexibility between variants associated with sialidosis type 1 (S182G and G328S) and type 2 (R78C and Y370C). These results indicated the mild pathogenicity of V143E, which may be associated with increased sialidase activity. To confirm our hypothesis, the NUE1 (wild-type or mutant) construct and PPCA cDNA were cotransfected into COS-7 cells. The calculated sialidase activities of S182G and V143E were 90% and 80%, respectively, of that of the wild-type protein (Figure 4A). Therefore, we confirmed that both S182G and V143E were mild variants of NEU1.

Figure 4.

Distribution of identified missense variants in NEU1. A: Sialidase activity test. With the cotransfection of PPCA, the calculated sialidase activities of S182G and V143E were not significantly different from those of the wild-type (WT) protein (p > 0.05). B: Lateral view of the structural model of the NEU1 protein. The mutant residues of NEU1 are shown as spheres. Mild variants are shown in green, and severe variants are shown in red. C: Schematic representation of the NEU1 protein. The missense variants already identified in sialidosis are represented as vertical lines. The signal peptide is indicated by a blue box. Yellow boxes represent the Asp-box elements. Mild variants are shown as green lines and severe variants are shown in orange. Essential catalytic residues are indicated with red lines.

Clinical characteristics of patients with sialidosis type 1

A previous study confirmed that patients with sialidosis type 2 had variants that were catalytically inactive, and those with sialidosis type 1 had at least one variant that was catalytically active [3]. In this study, we reported a family with sialidosis type 1 harboring the compound heterozygous variants S182G and V143E. The two variants were confirmed to be catalytically active and were thus classified as “mild” variants. In contrast, variants that presented with undetectable sialidase activity and correlated with sialidosis type 2 were classified as “severe.” Interestingly, we found that in sialidosis type 2, both the homozygous and heterozygous variants were “severe,” while in sialidosis type 1, the homozygous variants were both “mild,” and the heterozygous variants could be divided into two groups: a “severe-mild” variant group and a “mild-mild” variant group. We subsequently reviewed previously reported missense variants in NEU1, and the sialidase activity of the NEU1 protein in transfected cells is summarized in Table 1. The enzyme activity levels of severe variants were mostly not determined or were less than 10% that of the wild type, while for mild variants, the activity levels reached approximately 90%. We mapped the distribution of all missense variants in the schematic and tertiary structures of the NEU1 protein (Figure 4B and C).

Table 1.

Residual sialidase activity reported in previous studies

We then further explored the differences in clinical features between the severe-mild and mild-mild groups and summarized all the NEU1 variants associated with sialidosis type 1 (Supplementary Table 1 in the online-only Data Supplement). First, we compared the proportions of patients with various clinical symptoms between the two groups and found that the number of patients with macular cherry-red spots in the severe-mild group was significantly greater than that in the mild-mild group (p = 0.0017), with prevalence rates of 73.7% and 25.0%, respectively (Figure 5A). Impaired cognition was found exclusively in the severe-mild group (p = 0.0072) (Figure 5B). However, patients with ataxia were mostly in the “mild-mild” group (p = 0.0347) (Figure 5C). Other frequent symptoms, such as myoclonus, dysarthria and epilepsy, were not significantly different between the two groups (p > 0.05) (Figure 5E-K). For clinical examinations, abnormal EEG, SEP, VEP, and MRI signals were found at similar rates (p > 0.05). In addition, there were no significant differences in sex distribution or age of onset between the two groups (p > 0.05) (Figure 5D and L).

Figure 5.

Genotype‒phenotype analysis of sialidosis type 1. A-K: Proportions of patients with clinical symptoms in the severe-mild and mild-mild sialidosis type 1 groups. The number of included patients was mapped on the bar plot. Patients in the severe-mild group were more likely to develop macular cherry-red spots (p = 0.0017) and impaired cognition (p = 0.0072). Patients with ataxia were mostly in the mild-mild group (p = 0.0347). Other variables were not significantly different between the two groups (p > 0.05). L: Association of genotype with age at onset (severe-mild group: n = 19; mild-mild group: n = 25; p = 0.1284).^*p < 0.05; ^**p < 0.01, Kruskal‒Wallis test. N, no; Y, yes; SEP, somatosensory-evoked potential; VEP, visual evoked potential; EEG, electroencephalography; MRI, magnetic resonance imaging.

Subsequently, we investigated the relationship between the appearance of different clinical symptoms in patients with sialidosis type 1 and the age of onset. Patients with myoclonus had late onset (p = 0.0397), while patients with cherry-red spots, abnormal EEG signals, or abnormal VEPs had early onset (p < 0.05) (Figure 6A-D). Symptoms such as ataxia, dysarthria, seizures, impaired cognition, abnormal SEPs, and abnormalities on MRI did not correlate with the age of onset (p > 0.05) (Figure 6E-J).

Figure 6.

Association of clinical symptoms with age at onset. A-D: Patients with myoclonus had a late onset (p = 0.0397), while patients with cherry-red spots (p = 0.0003), abnormal EEG (p = 0.0075), or abnormal VEP (p = 0.0138) showed an early age of onset. The numbers of patients included in the analysis were as follows: myoclonus: Y: 59, N: 9; cherry-red spots: Y: 38, N: 28; abnormal EEG: Y: 16, N: 24; abnormal VEP: Y: 14, N: 4. E-J: Symptoms such as ataxia, dysarthria, seizures, impaired cognition, abnormal SEPs, and abnormalities detectable by MRI did not correlate with the age of onset of the disease (p > 0.05). The numbers of patients included in the analysis were as follows: ataxia: Y: 60, N: 8; dysarthria: Y: 31, N: 21; seizure: Y: 43, N: 25; impaired cognition: Y: 5, N: 57; abnormal SEP: Y: 16, N: 3; abnormal MRI: Y: 23, N: 27. ^*p < 0.05; ^**p < 0.01; ^***p < 0.001, Kruskal‒Wallis test. N, no; Y, yes; SEP, somatosensory-evoked potential; VEP, visual evoked potential; EEG, electroencephalography; MRI, magnetic resonance imaging.

Structural and functional consequences of NEU1 variants

Studies have shown that patients with sialidosis type 2 have variants that are catalytically inactive (severe), whereas patients with sialidosis type 1 have at least one variant that is catalytically active (mild) [3]. Thus, the sialidosis type 1 genotype consists of two combinations: severe-mild or mild-mild. We performed a structural analysis of the newly identified V143E variant on the basis of four aspects: 1) the energy of the optimal structure, 2) the stability and flexibility of key residues, 3) the polar contacts of mutant residues, and 4) the tertiary structure of the NEU1 protein. Two severe variants (R78C and Y370C) and two mild variants (G328S and S182G) were used as positive controls. To investigate the relationship between protein energy changes and disease phenotype, we calculated the energies of the binding site and active site mutants Y370C and R78C, both of which are associated with sialidosis type 2. The results showed that the energies of both variants were greater than those of S182G and V143E. Unfortunately, further calculations of energy for severe and mild variants did not yield meaningful conclusions (Supplementary Table 2 in the online-only Data Supplement).

In our study, we used I-TASSER to predict the structures of wild-type and mutant NEU1 proteins for the first time, and we calculated the energies of the proteins and the flexibility of their key residues. The analysis of flexibility changes at key residues showed that the R78C and Y370C variants at the binding and active sites, respectively, reduce the flexibility of key residues. As a result, these key sites fail to bind effectively to the substrate, thus weakening the function of NEU1. However, the flexibility values of the key residues in the G328S and S182G mutants were not significantly reduced, and the flexibility values of three sites were increased; these findings also indicate that G328S and S182G are mild variants, resulting in a type 1 clinical phenotype. For V143E, the curve of residue flexibility was between that of severe variants and that of mild variants. We then analyzed the local polarity of each variant site, and the results were consistent with the changes in flexibility. That is, variants that cause large changes in polar contacts, such as R78C and Y370C, reduce the flexibility of key residues, leading to the instability of protein structures. In particular, both R78 and Y370 interact with the E394 residue, which is also a binding site. The R78C variant eliminates the interaction between R78C and E394, and the Y370C variant weakens the interaction between Y370C and E394, which may explain the decreased flexibility caused by the R78C variant. However, V143E, S182G, and G328S did not change the polar contacts that interacted with the catalytic residues. Herein, we speculated that V143E was more likely a mild variant. Our findings suggested that changes in the flexibility and local polar contacts of the mutant protein may be associated with the severity of the clinical sialidosis phenotype.

The locations of the variant sites within the protein were divided into two groups: 1) those that caused large conformational changes, which included active-site and binding-site variants such as P80L and P316S; and 2) those that caused local conformational changes in molecular surface regions, such as W240R, G243R, and V217M [8]. However, regarding clinical phenotypes, P80L, W240R, and G243R are associated with sialidosis type 2, while P316S and V217M are associated with type 1. Another study included five NEU1 variants that were concentrated in the same region on the molecular surface of sialidase, and all of these variants resulted in decreased enzyme activity [12]. Among these variants, G227R, A298V, G68V, and L270P were considered “severe,” while S182G and G328S were considered “mild.” This result suggested that the location of a NEU1 variant could not fully explain the correlation between structure and phenotype.

The residual enzyme activity of the mutant protein is closely related to the severity of clinical symptoms [3]. We further tested the sialidase activity of S182G and V143E and reviewed all the NEU1 missense variants and studies on residual sialidase activity. Our results confirmed that V143E was a mild variant. The changes in sialidase activity may be related to structural changes in the mutant proteins. In a previous study, F260Y and A298V were shown to cluster in the same region on the surface of the sialidase molecule, significantly reducing enzyme activity, and they are both associated with sialidosis type 2 [10]. As mentioned, both V217M and G234R can cause local structural changes on the protein surface [8], but the predicted changes caused by V217M are smaller than those caused by G243R, resulting in drastic, widespread changes. The expressed neuraminidase containing the V217M variant showed some residual enzyme activity, while the G243R mutant lost all of its enzyme activity [13]. Therefore, we speculate that both the residual enzyme activity and the conformational changes caused by the mutant protein determine the clinical phenotype of the patient.

Clinical characteristics of patients with sialidosis type 1

Sialidosis type 2 can be classified into two subtypes, juvenile and infant, but no related study has been conducted on sialidosis type 1. In this study, for the first time, sialidosis type 1 was classified by genotype, and meaningful results were obtained. Analysis of clinical characteristics revealed that patients with severe-mild sialidosis type 1 had a high incidence of cherry-red spot syndrome and impaired cognition but a low rate of ataxia. Patients with myoclonus developed symptoms at a late age, and those with cherry-red spots, abnormal EEGs and VEP had an early onset of symptoms.

We found that patients with sialidosis type 1 harboring severemild variants are more prone to cherry-red spots than those harboring mild-mild variants. However, a mild-mild genotype was confirmed in this new family, and cherry-red spots were found in both the proband and her brother, who developed symptoms at the ages of 13 and 14, respectively. The inconsistency of our results can be explained by the fact that patients with cherry-red spots are often diagnosed at an early age at onset, with an average age of 13 years. Studies have shown that cherry-red spots are less common in Asian [14] and Taiwanese populations [15] than in other populations worldwide. However, we discovered this was because almost none of the patients with a homozygous S182G variant presented cherry-red spots. S182G is a hot spot variant that is common in Chinese people [15,16], and most of the known cases of this variant were reported in Taiwan [4]. Therefore, we believe that the internal reason for the low occurrence rate of cherry-red spots was related to the mild-mild variant of sialidosis type 1 caused by S182G. Ataxia is another highly prevalent symptom in patients with sialidosis. Our study showed that patients carrying mild-mild variants are prone to ataxia. Moreover, ataxia can be a milder or prior manifestation of myoclonus. Therefore, ataxia was more common in the mild-mild group. In this study, both patients had ataxia, which was consistent with our systemic findings. The results from a previous study confirmed that patients with late-onset disease have slow progression and apparent mild clinical symptoms, such as the absence of cherry-red spots, seizures and ataxia [6]. Abnormal EEG and VEP signals were also found at an early age of onset and could play an important role in the clinical evaluation and early diagnosis of sialidosis.

Because we summarized the data from previous work, the included sample size of each group may not be consistent, and the following results need to be reconsidered. First, we found that impaired cognition occurred only in the severe-mild group, and the symptoms did not correlate with the age of onset. However, the average age of these patients was 14 years, ranging from 8–17 years. Due to the large number of patients without impaired cognition, the average age at onset was 14 years, ranging from 8–33 years. Thus, we speculated that older patients may not be prone to impaired cognition. Intellectual disability is a clinical manifestation of sialidosis type 2, while it is rare in type 1 sialidosis. In a 30-year follow-up study of treatment for type 1 sialidosis, severe motor and speech impairments were observed, but cognitive function was retained [17]. Second, multiple studies have shown that myoclonus is the most common symptom of type 1 sialidosis, occurring in 96%–100% of cases [14,15,18]. In this study, there was no difference in the prevalence of myoclonus between the severe-mild and mild-mild groups, while patients with myoclonus seemed to have a late age of onset according to our results. Although the sample size of patients with myoclonus was much larger than that of patients without myoclonus, patients without myoclonus had an earlier onset age, ranging from 8–17 years, than did those without myoclonus, ranging from 9–33 years. A previous study demonstrated that most patients develop myoclonus, ataxia, and seizures within 5 years of onset [4]. The computed tomography scans of a patient at 21 years of age showed enlargement of the fourth ventricle. At 40 years of age, an MRI revealed severe atrophy in the cerebellum and pontine regions, as well as atrophy in the hemispheres and corpus callosum [19]. These findings suggested that the appearance of abnormalities in some test results is closely related to the course of the disease. Thus, a proper explanation of this finding may be that these patients were too young to develop myoclonus. Third, few patients underwent SEP examination. However, SEP abnormalities were found in almost all patients with type 1 sialidosis, indicating that SEP abnormalities are useful indicators for early diagnosis. Unfortunately, there was no significant difference in the age of onset between the severe-mild and mild-mild groups due to the small sample sizes.

In conclusion, our study identified a novel family of patients with sialidosis type 1 harboring S182G and V143E variants, of which the V143E variant was newly reported. Through structural analysis of changes in the systematic energy, flexibility and local polar contacts of the mutant protein, V143E was predicted to be a mild pathogenic variant, which was confirmed by a sialidase activity assay. The sialidosis type 1 strains can be further divided into two groups. The severe-mild variant group was more likely to have cherry-red spots and impaired cognition than the mild-mild variant group, and the former had a lower rate of ataxia. Moreover, cherry-red spots and abnormal EEGs and VEPs were found in patients with early-onset disease, while patients with myoclonus had late-onset disease. Almost all patients had abnormal SEPs, which was helpful for the early diagnosis of sialidosis. Due to mild systematic impairment in sialidosis type 1 patients, the management of these patients is important for their prognosis. Therefore, these particular symptoms should be considered when patients are diagnosed with sialidosis type 1. Our future studies will examine the differences in treatment and prognosis between the two groups, hoping to identify effective interventions.