Introduction

Until recently, the cause of downbeat nystagmus (DBN) has remained unknown (“idiopathic”) in approximately 30% of cases [1]. However, intronic FGF14 (GAA)≥250 repeat expansions, known to cause spinocerebellar ataxia 27B/GAA-FGF14 disease [2, 3], were lately shown to account for almost 50% of previously unexplained DBN cases [4], suggesting that monogenic causes may be a recurrent cause of what has so far been considered “idiopathic” DBN.

In particular, biallelic RFC1 repeat expansions may represent a common cause of “idiopathic” DBN syndrome given the anecdotal reports of DBN in RFC1-related disorder [5,6,7,8,9]. To test this hypothesis, we studied the frequency of RFC1 repeat expansions in a cohort of patients with “idiopathic” DBN, characterized the phenotypic profile of the RFC1-related DBN syndrome, and compared it to that of the GAA-FGF14-related DBN syndrome.

Methods

We studied a series of 219 patients with suspected DBN of unknown etiology referred to the Department of Neurology or the German Center for Vertigo and Balance Disorders at the LMU Hospital in Munich, Germany, between 2012 and 2020. Patients underwent comprehensive etiologic evaluation of DBN syndrome and in-depth phenotyping as described previously [4]. Patients were excluded from the study if no or insufficient DNA was available for genetic screening (n = 5) or if DBN was not objectified on examination (n = 11) (Fig. 1). Of the remaining 203 patients with DBN, a presumed secondary cause of DBN—either acquired or genetic, but excluding GAA-FGF14 disease—had previously been identified in 36 patients during clinical and paraclinical evaluation, an FGF14 (GAA)≥250 allele in 82 patients, and an FGF14 (GAA)200–249 allele in 20 patients, yielding 65 patients with “idiopathic” DBN (Fig. 1). Patients carrying an FGF14 (GAA)≥250 allele and a (GAA)200–249 allele were analyzed together due to recent evidence suggesting that (GAA)200–249 alleles may be associated with DBN, given their significant enrichment in patients with DBN and that the phenotype of (GAA)200–249-FGF14 patients closely mirrored that of (GAA)≥250-FGF14 patients [4]. All 203 patients with DBN were screened for RFC1 AAGGG expansions as described previously [10]. Patients with a presumed secondary cause of DBN and GAA-FGF14-related DBN were not excluded from RFC1 screening to explore the possibility of co-occurring diseases. Two patients with a presumed secondary cause of DBN (chronic alcohol use) who were found to carry biallelic RFC1 repeat expansions were not included in the phenotypic analysis of the RFC1-related DBN syndrome cohort given the difficulty in determining the relative phenotypic contributions of chronic alcohol use and RFC1 repeat expansions.

Fig. 1
figure 1

Study flowchart of the recruitment of patients with DBN. DBN downbeat nystagmus

Full size image

Deep phenotyping was performed by reassessing medical records using a standardized data sheet. Bilateral vestibulopathy (BVP) was diagnosed as per the consensus criteria of the Bárány Society requiring the documentation of bilaterally reduced or absent angular vestibular ocular reflex (VOR) function by caloric stimulation, video head impulse test (vHIT), or rotatory chair [11]. Polyneuropathy was diagnosed on nerve conduction studies (NCS), excluding focal entrapment neuropathies, or clinically defined by the combination of significantly decreased vibration sense at the ankles (≤ 3/8 on the Rydel–Seiffer scale) and decreased ankle reflexes [12].

This study was approved by the ethics committee of the LMU Munich and we obtained written informed consent from all the participants.

Results

Frequency of biallelic RFC1 expansions

Biallelic RFC1 AAGGG repeat expansions were identified in 15 of 65 (23%) patients with “idiopathic” DBN (Figs.1 and 2A). Moreover, a high frequency of heterozygous RFC1 repeat expansion carriers was observed in the “idiopathic” DBN cohort (12%, 8/65 patients / 6.2%, 8/130 allele frequency; compared to 0.7–6.5% carrier frequency in the general population [13]). A total of 50 patients remained unsolved after RFC1 screening, and will be referred to onward as “genetically unexplained”. In addition, 2 of the 36 (6%) patients with presumed secondary DBN were found to carry biallelic RFC1 AAGGG repeat expansions, while none of the 102 (GAA)≥200-FGF14-positive patients did (Fig. 1).

Fig. 2
figure 2

Frequency of RFC1 repeat expansions in DBN syndromes and progression of functional disability in the RFC1-related DBN syndrome. A Percentage of patients carrying biallelic RFC1 AAGGG repeat expansions in a cohort of 65 patients with idiopathic downbeat nystagmus (DBN). B Percentage of patients carrying biallelic RFC1 AAGGG repeat expansions in the phenotypic subgroups with (1) pure DBN, (2) DBN plus cerebellar impairment (DBN + C), (3) DBN plus cerebellar impairment and bilateral vestibulopathy (DBN + C + BVP), (4) DBN plus cerebellar impairment and polyneuropathy (DBN + C + N), and (5) DBN plus cerebellar impairment, BVP, and polyneuropathy (DBN + C + BVP + N). No patient with DBN plus isolated BVP or isolated neuropathy was identified among the idiopathic DBN cohort. C Disease duration at time of onset of regular use of walking aid and regular falls in the RFC1-positive patients with DBN. D Longitudinal intra-individual progression of functional impairment as assessed by the FARS functional disability stage relative to disease duration (35 observations from 11 patients with DBN carrying biallelic RFC1 repeat expansions). Observations from the same patient are connected by a dotted line. The FARS functional stage assesses disability through a 7-point ordinal scale: 0 = normal; 1 = minimal signs on examination; 2 = minimal disability; 3 = mild disability; 4 = moderate disability, requires a walker; 5 = severe disability, confined but can navigate a wheelchair; 6 = total disability

Full size image

Phenotypic characterization of the RFC1-related DBN syndrome

The frequency of biallelic RFC1 repeat expansions stratified by DBN subgroups was 50% (3/6) for DBN plus cerebellar impairment and BVP, and 86% (12/14) for DBN plus cerebellar impairment, BVP, and polyneuropathy (Fig. 2B). None of the patients in the other DBN subgroups carried biallelic RFC1 repeat expansions (Fig. 2B). Table 1 presents the clinical features of the RFC1-positive, (GAA)≥200-FGF14-positive, and genetically unexplained DBN cohorts.

Table 1 Characteristics and discriminative features of the RFC1-related downbeat nystagmus syndrome
Full size table

DBN occurred with cerebellar impairment in all 15 RFC1-positive patients, which was limited to the ocular motor system with typical cerebellar ocular motor signs (i.e., saccadic pursuit, dysmetric saccades, gaze-evoked nystagmus) in 5 patients (33%). Additional cerebellar ocular motor signs were observed in all RFC1-positive patients. Brain MRI showed global cerebellar atrophy in 27% (3/11) and isolated vermis atrophy in 9% (1/11) of patients. BVP was documented in all RFC1-positive patients by vHIT (n = 10) or caloric stimulation (n = 5). Polyneuropathy was identified in 12 of 15 (80%) RFC1-positive patients, and was diagnosed on NCS in six patients and clinically in six patients. Three patients had no evidence of neuropathic features on examination, though NCS were not available for these patients. The presence of chronic cough could not be reliably extracted from medical records, although it was documented in two patients in whom it developed more than 10 years before the onset of gait impairment.

Progression of functional disability in the RFC1-related DBN syndrome

A substantial proportion of RFC1-positive patients experienced regular falls (89%, 8/9), some of them as early as 2.5 years after disease onset (median disease duration at onset of regular falls, 4 years; range, 2.5–7.5). Furthermore, walking aids were used by 50% of patients (7/14) after a median disease duration of 6.5 years (range, 2.5–18) (Fig. 2C). At time of last examination, the median Friedreich Ataxia Rating Scale (FARS) functional stage was 3.25 (range, 1.5–4), indicating a mild-to-moderate disability (Fig. 2D).

Discriminative features of the RFC1-related DBN syndrome

Compared to (GAA)≥200-FGF14-positive and genetically unexplained patients with DBN, RFC1-positive patients with DBN appeared more functionally impaired, as assessed by the FARS functional stage, history of regular falls and use of walking aids, and gait impairment on examination (Table 1). However, the RFC1-positive DBN group also had a significantly longer disease duration compared to the genetically unexplained group (median, 7 vs 4 years; p = 0.007) and a trend toward longer disease duration compared to the (GAA)≥200-FGF14-positive group (median, 7 vs 6 years; p = 0.070), which may account in part for the higher degree of functional impairment in the RFC1-positive group. Neuropathic features and proprioceptive dysfunction on examination were significantly more common in the RFC1-positive group (Table 1), in keeping with early and preferential involvement of dorsal root ganglia in that disease [14]. Vestibular impairment was also significantly more common and severe, as measured by VOR gains on vHIT, in RFC1-positive patients (n = 9) compared to (GAA)≥200-FGF14-positive (n = 9) and genetically unexplained patients (n = 2) (Table 1).

Discussion

This study showed that biallelic RFC1 AAGGG repeat expansions are a common monogenic cause of DBN syndrome, accounting for 23% of previously “idiopathic” DBN cases. Given this high frequency, genetic testing for RFC1 repeat expansions may now become part of the diagnostic workup of patients with “idiopathic” DBN. Of note, since biallelic RFC1 repeat expansions were also identified in 6% of patients who had a presumed secondary cause of DBN, genetic testing might need to be extended to this population as well—especially in the presence of other cerebellar signs, vestibular hypofunction, and/or polyneuropathy—given the implications for clinical management and eligibility for future clinical trials.

Our findings provide a deeper phenotypic characterization of the RFC1-related DBN syndrome by showing that they present along a continuum of involvement of the cerebellar, sensory, and vestibular systems. This confirms and extends previous notions of widespread neurodegeneration occurring in RFC1-related disorder [15]. Accordingly, no patient with pure DBN or DBN plus cerebellar impairment (without BVP and/or polyneuropathy) was found to carry biallelic RFC1 repeat expansions, strengthening the observation that RFC1-related disorder is unlikely in presence of isolated cerebellar ataxia without sensory neuropathy [16]. The multisystemic involvement in the RFC1-positive DBN syndrome was further reflected by the significantly more common neuropathic features and proprioceptive dysfunction on examination as well as vestibular impairment—which was comparatively more severe—in this group compared to the (GAA)≥200-FGF14-positive and genetically unexplained groups. These phenotypic findings might help to raise clinical suspicion for RFC1-related disease over other monogenic causes of DBN, such as GAA-FGF14 disease [4].

Our study also provides preliminary insights into the natural evolution of the RFC1-related DBN syndrome. A significant proportion of patients experienced regular falls and needed walking aids relatively early in the disease course, which is of importance for clinical management given the relevance for everyday living and as potentially highly meaningful outcomes in future treatment trials [17]. However, it remains to be established in larger, prospective cohort series if functional impairment progresses more rapidly in the RFC1-related DBN syndrome compared to the GAA-FGF14-related DBN syndrome, which would be in line with the higher degree of underlying multisystemic neurodegeneration in RFC1-related disease [15, 18].

Our study has several limitations. First, it is a single-centre retrospective study, which limited our ability to assess the evolution of multisystemic damage in the RFC1-positive DBN syndrome. Second, our study provides a conservative estimate of the real frequency of RFC1-related disorder in “idiopathic” DBN as it only screened for pathogenic AAGGG repeat motifs and not for truncating variants and other non-reference pathogenic motifs that have recently been shown to cause RFC1-related disorder [7, 19]. The elevated frequency of heterozygous RFC1 repeat expansion carriers in our cohort (12%) raises the possibility that some of these patients may carry a novel variant in trans with the AAGGG expansion. Third, we were unable to objectify the presence of polyneuropathy—a universal feature of RFC1-related disorder [14, 15]—in all RFC1-positive patients as only 40% underwent NCS.

In conclusion, we showed that biallelic RFC1 AAGGG repeat expansions are a recurrent monogenic cause of DBN syndrome.