Introduction

The spinocerebellar ataxias (SCAs) are a heterogeneous group of rare neurodegenerative disorders. Spinocerebellar ataxia 27B (SCA27B OMIM:620174, GAA-FGF14 ataxia) is a late-onset (onset >30 years of age), slowly progressive autosomal dominant disease characterized by episodic vertigo and ataxia [1,2,3]. Expansion of intronic pure GAA repeats (expansion ≥250 triples) in FGF14 (GAA-FGF14) has been demonstrated to interfere with FGF14 transcription, resulting in GAA-FGF14 ataxia [1]. Although GAA-FGF14 ataxia was shown to be a common cause of SCA in French Canadian ataxia cohort (61%) [1] and European ataxia cohorts (5.4-51%) [3,4,5,6,7], it appeared relatively uncommon in Chinese (1.3%) and Japanese (1.1%) populations [8,9,10].

The burden of autonomic dysfunction in GAA-FGF14 patients is associated with disease course [2]. Previous studies reported that symptoms of autonomic dysfunction in GAA-FGF14 ataxia were rare in the early stage but became increasingly frequent in the late stage [2]. Nevertheless, age-related degeneration of the sympathetic nerve can interfere with evaluation of autonomic dysfunction in late-onset diseases [2]. A standardized and quantitative test of the autonomic nerve is therefore needed to determine whether autonomic dysfunctions are due to GAA-FGF14 ataxia, rather than age-related processes.

Here, we reported on 18 GAA-FGF14 patients from two unrelated families identified in a Chinese undiagnosed LOCA cohort. Agarose gel electrophoresis and sequencing analysis showed that the number of GAA repeats expanded in maternal transmissions and contracted in paternal transmissions. Brain MRI showed slight to moderate cerebellar atrophy. Furthermore, the amplitude of SSR was reduced in GAA-FGF14 patients compared with that of healthy controls.

Materials and methods

Participants and clinical assessments

A total of 18 GAA-FGF14 patients from two unrelated families were recruited from a cohort of 664 unrelated and undiagnosed Chinese individuals with LOCA (370 male/294 female; onset >30 years of age [11]) screened at The First Affiliated Hospital of Fujian Medical University (ClinicalTrials.gov number, NCT04010214). 45 (45/664, 6.67%) of individuals had family history, including 38 (38/45, 84.44%) individuals with an autosomal dominant family history and 7 (7/45, 15.56%) individuals with an autosomal recessive family history. These individuals, who were undiagnosed before GAA-FGF14 screening, were screened for expansions in ATXN 1, 2, 3, 7, 8, 10, CACNA1A, PPP2R2B, TBP, ATN1, FXN, NOP56, THAP11 and HTT genes and excluded multiple system atrophy (MSA) utilizing the second consensus statement on the diagnosis of MSA [12]. Whole exome sequencing (WES) was conducted in these patients to exclude previously identified causal pathogenic variants.

GAA-FGF14 patients underwent follow-up examinations to assess symptom, scale, and SSR. Disease severity was determined clinically using the Scale for the Assessment and Rating of Ataxia (SARA) [13] and The International Cooperative Ataxia Rating Scale (ICARS) [14]. Using a SARA cut-off value of 3, GAA-FGF14 patients were divided into pre-symptomatic stages (SARA < 3) and symptomatic stages (SARA ≥ 3) [15, 16]. SSR was used to test for involvement of the sympathetic nervous system in GAA-FGF14 patients by recording SSR amplitude and latency. SSR evoked by a square wave pulse delivered at the volar wrist was recorded at both the lower and upper limbs [17, 18]. The control cohort comprised 16 unrelated individuals who were healthy at the time of SSR examination, and whose age and gender matched the 16 GAA-FGF14 patients under follow-up. All individuals gave informed consent, and the study was approved by the Medical Ethical Committee of The First Affiliated Hospital of Fujian Medical University (MRCTA, ECFAH of FMU 20191195).

Genetic screening for pure GAA repeat expansion in FGF14

Genetic screening for GAA-FGF14 repeat expansion ([GAA] ≥250) was performed in 664 undiagnosed individuals with LOCA and 39 family members of the two GAA-FGF14 probands. All blood samples used for sequencing were collected at The First Affiliated Hospital of Fujian Medical University. The FGF14 repeat locus was screened by long-range polymerase chain reaction (LR-PCR) and repeated-primed polymerase chain reaction (RP-PCR) [1]. Agarose gel electrophoresis of LR-PCR products was used to calculate the number of GAA repeats [1, 19]. The LR-PCR products were separated on a 1.5% agarose gel (180 V, 30 min). The size of the alleles was calculated with Image Lab Software V6.1 (Bio-Rad) with a DNA size marker (D2000, MD114, TIANGEN; D15000, MD110, TIANGEN) simultaneously electrophoresed as a reference. The trinucleotide repeat number was estimated after subtraction of the flanking fragment size. Sanger sequencing was used to distinguish between alternate and pure intronic GAA expansion ([GAA] ≥250) alleles in FGF14 [3]. Utilizing the primers in Supplementary Data S1 respectively sequenced from upstream and downstream of trinucleotide repeat region to ensure this region had been full sequenced and exclude small interruptions or variations which existed in repeat motif. The RP-PCR products was mixed with GeneScan™ 500 LIZ (4322682, Applied Biosystems) and performed on an ABI 3730xl Genetic Analyzer (Applied Biosystems). The data were analyzed with the GeneMapper software V4.0 (Applied Biosystems). The LR-PCR and RP-PCR protocols are summarized in Supplementary Data S1.

Long-read sequencing

Libraries for long-read sequencing were created utilizing DNA ligation kits (SQK-LSK109, Oxford Nanopore Technologies, Oxford, UK) and sequenced on one R9.4.1 flowcell per individual utilizing a PromethION sequencer (Oxford Nanopore Technologies, Oxford, UK). Long reads were aligned to the hg38 using minimap2 [20], and the results were visualized in the Integrative Genomics Viewer [21]. The counts of trinucleotide repeat in FGF14 were estimated using Straglr [22].

Statistical analysis

For continuous variable data, individual values were presented with the median and standard error of the mean (SEM). Continuous variable data were compared by unpaired t-test, Wilcoxon matched-pairs signed rank test or Mann–Whitney test, as applicable. Spearman’s correlation coefficient was calculated to test for possible associations between SSR amplitude and ICARS scores. The significance level was set at P < 0.05.

Results

Genetic screening for GAA-FGF14 repeat expansion

In the cohort of 664 unrelated and undiagnosed individuals with LOCA, GAA-FGF14 repeat expansion was identified in two probands from 38 individuals with an autosomal dominant family history utilizing Sanger sequencing, while it wasn’t identified in sporadic cases or individuals with an autosomal recessive family history. Genetic screening for GAA-FGF14 repeat expansion was performed in 39 family members, 16 of whom were genetically diagnosed with GAA-FGF14 ataxia. Sanger sequencing was performed upstream and downstream of trinucleotide repeat region, confirming that this region had been fully sequenced and identified as containing pure GAA repeats (expansion ≥ 250 triples) in FGF14. The Sanger sequencing results of intronic GAA expansion alleles in GAA-FGF14 patients was shown in Supplementary Data S2. Alignment of long-reads to the FGF14 intronic locus (chr13:102,161,544-102,161,749; GRCh38) showed the presence of a heterozygous GAA repeat expansion in two probands who underwent long-read genome sequencing (Fig. 2A). The results of RP-PCR from two probands also supported pure GAA repeats (expansion ≥ 250 triples) in FGF14 (Fig. 2B). Because many expanded alleles in our patients were too large to be detected by capillary electrophoresis, agarose gel electrophoresis of LR-PCR products was used to calculate the number of GAA repeats (Supplementary Data S3) [19]. Among the 18 total GAA-FGF14 patients (median repeat number: 342; range: 259–402), 7 (39%) were in the symptomatic stage and 11 (61%) remained in the pre-symptomatic stage. Family pedigrees and GAA-FGF14 repeat numbers of each individual are shown in Fig. 1. In two probands, their GAA repeat numbers were also evaluated using long-read sequencing (Supplementary Data S4). Compared with agarose gel electrophoresis, the GAA-FGF14 repeat number evaluated for patient II.1 in Family 1 by long-read sequencing was smaller (agarose gel electrophoresis: 343 repeats; long-read sequencing 311.3 repeats). The GAA-FGF14 repeat number for patient II.2 in Family 2, as evaluated by long-read sequencing, was close to that obtained by agarose gel electrophoresis (agarose gel electrophoresis: 373 repeats; long-read sequencing 376.7 repeats).

Fig. 1
figure 1

Pedigrees of two families with GAA-FGF14 ataxia. A four-generation family (Family 1) had 69 individuals, including 16 individuals with GAA-FGF14 repeat expansion. Another affected family (Family 2) included the proband and his mother with ataxia. The labels were as follows: square for male, circle for female, a diagonal line through a symbol for deceased individual, filled symbol with black for individual with paroxysmal dizziness or gait instability, and the numbers below the member code represented GAA-FGF14 repeat sizes

Full size image

Genetic instability of GAA-FGF14 repeat expansion

Agarose gel electrophoresis for LR-PCR products from the proband’s families showed that GAA-FGF14 repeat number expanded in 10 individuals with maternal transmission (expansion range from 1 to 35 triples), but contracted upon paternal transmission in 3 individuals (contraction range from 93 to 99 triples), across 13 meiotic events in two families (Fig. 3A). For instance, two meiotic events resulted in two different fathers transmitting contractive alleles (Family 1, patient II.11 [352 repeats] transmitted to patient III.26 [259 repeats]; patient III.7 [371 repeats] transmitted to patient IV.6 [278 repeats]) with incomplete penetrance in offspring (Fig. 1). Another meiotic event was detected in which a benign allele [221 repeats] was inherited from a father with a pathogenic allele (Family 1, patient III.22 [320 repeats]; Fig. 1). The genetic and clinical characteristics of GAA-FGF14 patients are summarized in Table 1.

Table 1 Genetic and clinical characteristics of GAA-FGF14 patients
Full size table

Radiological characteristics of GAA-FGF14 patients

Five GAA-FGF14 patients underwent brain magnetic resonance imaging (MRI) (Fig. 2C). Brain MRI revealed that moderate cerebellar atrophy in the patient II.2 in Family 2 (SARA score = 8), mild cerebellar atrophy in the patient II.7 in Family 1 (SARA score = 15.5), slight cerebellar atrophy in the patient II.1 in Family 1 (SARA score = 6.5), II.3 in Family 1 (SARA score = 9.5) and II.9 in Family 1 (SARA score = 3). None of the patients exhibited brainstem atrophy.

Fig. 2
figure 2

Alignment of long-reads to the FGF14 intronic locus, Repeat-primed PCR and brain MRI in GAA-FGF14 patients. A The Integrative Genomics Viewer showed the GAA repeat expansion in FGF14 (chr13: 102,161,544-102,161,749, hg38 version), as opposed to the normal allele, in Family 1, patient II.1 and Family 2, patient II.2. B Repeat-primed PCR results for Family 1, patient II.1 and Family 2, patient II.2. C Brain MRI revealed that moderate cerebellar atrophy in the patient II.2 in Family 2(P-5), mild cerebellar atrophy in the patient II.1 in Family 7 (P-3), slight cerebellar atrophy in the patient II.1 in Family 1 (P-1), II.3 in Family 1 (P-2) and II.9 in Family 1 (P-4)

Full size image

Sympathetic nerve involvement in GAA-FGF14 patients

A previous study reported that the incidence of autonomic dysfunction symptoms in GAA-FGF14 ataxia was correlated with disease stage [2]. To determine whether GAA-FGF14 patients in the pre-symptomatic stage suffered from sympathetic nerve involvement and to evaluate the extent of involvement in different disease stages, clinical examinations and SSR recordings were conducted in 16 GAA-FGF14 patients (6 symptomatic, 10 pre-symptomatic) and 16 age- and sex-matched controls.

Although frequent micturition (6.25%; 1 of 16), urgent urination (6.25%; 1 of 16) and uracratia (12.5%; 2 of 16) were observed in some of these GAA-FGF14 patients and could be symptoms of autonomic dysfunction, these symptoms could also be attributed to chronic cystitis and/or pelvic floor dysfunction. In addition, no evidence of postural hypotension (minimum SBP variations: -7 mmHg; minimum DBP variations: -8 mmHg) was observed in any of the 16 GAA-FGF14 patients (Fig. 3G).

Fig. 3
figure 3

Characteristics of GAA repeat intergenerational instability and sympathetic nerve involvement in GAA-FGF14 ataxia. A Relationship between intergenerational GAA repeats number variation and parental gender. B Comparison with SSR amplitudes between GAA-FGF14 patients and healthy controls in each limb. C Comparison with SSR amplitudes between GAA-FGF14 patients in pre-symptomatic stages (SARA score <3, n = 10) and healthy controls (n = 10) in lower (n = 20) and upper (n = 20) limbs. D Comparison with SSR amplitudes between GAA-FGF14 patients in symptomatic (n = 6) and pre-symptomatic (n = 10) stages in lower and upper limbs. E Comparison with SSR amplitudes between GAA-FGF14 patients in symptomatic stages (SARA score ≥3, n = 6) and healthy controls (n = 6) in lower (n = 12) and upper (n = 12) limbs. F Correlation between SSR amplitudes and ICARS scores in GAA-FGF14 patients (n = 16). G Postural blood pressure variations in GAA-FGF14 patients (n = 16). Red area (SBP variation ≤ −20 mmHg) and blue area (DBP variation ≤ −10 mmHg) represent diagnostic range for postural hypotension. ns P ≥ 0.05, *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001. mV millivolt, SBP Systolic Blood Pressure, DBP Diastolic Blood Pressure

Full size image

We next evaluated differences in sympathetic nerve system involvement between GAA-FGF14 patients at different disease stages (6 symptomatic; 10 pre-symptomatic) through SSR recordings. In these 16 total GAA-FGF14 patients, SSR amplitude was lower in each limb compared to that in controls (n = 16) (left foot: P = 0.0131, right foot: P = 0.0092, left hand: P = 0.0076, right hand: P = 0.0027, Wilcoxon matched-pairs signed rank tests) (Fig. 3B). Subsequent comparison of SSR amplitude between different stages of GAA-FGF14 ataxia showed that SSR amplitude was significantly lower in the GAA-FGF14 patients in pre-symptomatic stage than in healthy controls (lower limbs: P = 0.0400; upper limbs: P = 0.0400, Wilcoxon matched-pairs signed rank tests) (Fig. 3C), while SSR amplitude was lower in the GAA-FGF14 patients in symptomatic stage compared to those in pre-symptomatic stage (lower limbs: P = 0.0072, upper limbs: P = 0.0010; Mann-Whitney tests) (Fig. 3D), as well as the healthy controls (lower limbs: P = 0.0005, upper limbs: P = 0.0005; Wilcoxon matched-pairs signed rank tests) (Fig. 3E). The amplitude of SSR in the upper limbs was negatively correlated with ICARS score (upper limbs: P = 0.0054; Spearman correlation) (Fig. 3F). In contrast, no significant correlation was found between SSR amplitude in lower limbs and ICARS scores (Spearman correlation, lower limbs: P = 0.2790) (Fig. 3F).

Discussion

Intronic GAA repeat expansion in FGF14 (GAA-FGF14) has been associated with SCA27B, a late-onset, slowly progressive autosomal dominant disease [1, 3]. Typical clinical manifestations of GAA-FGF14 ataxia are episodic vertigo and ataxia, while imaging and neuropathological findings show cerebellar vermis atrophy [1, 2]. In this study, we identified two unrelated individuals (0.3%) with GAA-FGF14 ataxia through a screen of 664 unrelated individuals with undiagnosed LOCA. GAA-FGF14 repeat expansion was also present in 16 affected relatives of the two probands. Our analyses suggested that intergenerational variation in GAA repeat number could be related to the gender of parent transmitting the pathogenic allele. Brain MRI showed slight to moderate cerebellar atrophy. Clinical assessments and SSR recordings in the current study indicated that even pre-symptomatic GAA-FGF14 patients could exhibit sympathetic nerve involvement, although the involvement was not sufficiently severe to induce clinical symptoms.

The morbidity rates of repeat expansion disorders have been reported to vary among geographic regions. For instance, GGC repeat expansion in NOTCH2NLC, which causes neuronal intranuclear inclusion disease, is rare in European cohorts [23, 24]. Similarly, GAA repeat expansion in FXN is rarely observed among sub-Saharan Africans, Amerindians, and people from China, Japan, and Southeast Asia [25]. Other studies suggest that high variability in the morbidity rates of repeat expansion disorders is associated with founder effects [26]. Moreover, expanded alleles can be linked to a single or predominant haplotype [25, 26]. Previous studies had shown that GAA-FGF14 ataxia may not be as common in the Japanese ataxia cohort (1.1%) [10]. In this study, GAA-FGF14 ataxia comprised only 0.3% (2/664) of our Chinese undiagnosed LOCA cohort, which was substantially lower than the French Canadian ataxia cohort (61%), European ataxia cohorts (5.4–51%), Indian ataxia cohort (10%), Brazilian ataxia cohorts (9%) and Australian ataxia cohorts (9.5%) [1, 3,4,5,6,7]. As we were preparing this report, we noticed that another research group found the low frequency (1.3%) of GAA-FGF14 ataxia in Chinese LOCA cohort [9]. This striking difference further illustrates the regional variability of GAA-FGF14 ataxia incidence. However, the high proportion (619/664, 93.2%) of sporadic patients in our cohort might have contributed to the observed low frequency of GAA‐FGF14 ataxia in our study. The sporadic patients with nonhereditary ataxia, such as immune‐associated cerebellar ataxia, might have been included in our cohort. On the other hand, the higher proportion (2/38, 5.3%) of GAA‐FGF14 ataxia in the LOCA cohort with an autosomal dominant family history indicates that the repeat expansion analysis of FGF14 in Chinese LOCA patients is needed, especially when the proband has an autosomal dominant family history. As with other repeat expansion disorders, future studies in larger cohorts are needed to more rigorously compare GAA-FGF14 frequency between regions.

Short tandem repeats show relatively low stability between generations. For example, the number of GAA repeats in FXN tended to contract in paternal transmissions, but might expand or contract in maternal transmissions [27]. In our cohort, parental gender also appeared to affect GAA repeat number in subsequent generations. In particular, the number of GAA repeats in FGF14 typically expanded in maternal transmissions and contracted in paternal transmissions, which was consistent with observations reported by Pellerin and colleagues [1]. Moreover, the contractions of repeat number in FXN were discovered in the male germline by Pianese and co-workers [27]. Further study in larger cohorts will potentially help to clarify the patterns of variation in GAA-FGF14 repeat number in the male germline.

Five individuals who suffer from GAA-FGF14 ataxia in our families have undergone brain MRI which indicated slight to moderate cerebellar atrophy. None of the patients exhibits brainstem atrophy. In previous studies, the brain MRI showed that GAA-FGF14 patients had mild to moderate cerebellar atrophy [1, 2]. Our MRI results further support these findings. However, slight cerebellar atrophy was observed in two of our patients, despite having a SARA score ≥ 3. This indicates that some GAA-FGF14 patients have developed symptoms of ataxia without apparent cerebellar atrophy. It is important to note that the repeat expansion analysis of FGF14 is needed for patients who present symptoms of ataxia without apparent cerebellar atrophy on brain MRI.

Previous studies have described urinary urgency (28%) and erectile dysfunction (13%) as symptoms of autonomic dysfunction in GAA-FGF14 patients [2]. In our research, although two patients in the symptomatic stage suffered from frequent micturition, urgent urination, and uracratia, these symptoms could be related to chronic cystitis or pelvic floor dysfunction. It is worth noting that secondary factors, such as age and chronic disease, can lead to false-positive results of autonomic dysfunction in late-onset diseases. Moreover, the low frequency of occurrence of these symptoms suggests that sympathetic nerve involvement might not reach a severity level high enough to induce clinical symptoms in our GAA-FGF14 cohort. However, it is also possible that the relatively low rates of these symptoms may be an effect of the limited sample size. Thus, as with other unanswered questions in this rare disorder, systematic and quantitative evaluation of sympathetic nerve involvement in GAA-FGF14 ataxia requires study in a larger cohort.

The abnormal SSR responses can be considered evidence of sweating dysfunction [28]. The 2nd to 9th thoracic segments regulates the skin of the upper limbs and the 10th thoracic to 3rd lumbar segments regulates the skin of the lower extremities [28]. SSR alterations have been previously described in several spinocerebellar ataxia (SCA) subtypes, including spinocerebellar ataxia 2 (SCA2), spinocerebellar ataxia 3 (SCA3), and spinocerebellar ataxia 49 (SCA49) [29,30,31,32]. In our study, quantitative evaluation of sympathetic nerve involvement by SSR recordings revealed that SSR amplitude was lower in GAA-FGF14 patients in pre-symptomatic stage compared to healthy controls, and further decreased in the symptomatic stage. Additionally, SSR amplitude recording in the upper limbs exhibited a negative correlation with disease severity, whereas no significant correlation was observed for lower limb SSR amplitude. These data indicate that the sympathetic nerves arise from 2nd to 9th thoracic segments are preferentially involved in GAA-FGF14 ataxia. Our findings suggest that reduced amplitude of SSR could potentially serve as a clinical indicator of GAA-FGF14 ataxia. However, further studies, including additional clinical assessments such as heart rate variability [33], are needed to comprehensively evaluate sympathetic nerve involvement in GAA-FGF14 ataxia.

In conclusion, the current study identified two families with GAA-FGF14 ataxia, illustrating the low incidence of GAA-FGF14 ataxia among Chinese LOCA cases. Repeat expansion analysis of FGF14 is recommended for patients presenting with LOCA, especially when there is a history of autosomal dominant inheritance in the family. The gender of parents transmitting pathogenic GAA-FGF14 ataxia alleles also appeared to be related to variation in repeat number between generations. Reduced SSR amplitude also emerged as a salient feature in GAA-FGF14 patients warranting further exploration as a potential clinical indicator for diagnosis of GAA-FGF14 ataxia. Our findings can help guide future studies of the molecular mechanisms responsible for tandem repeat expansion and facilitate improved clinical diagnosis of GAA-FGF14 ataxia.