GAA-FGF14 disease: defining its frequency, molecular basis, and 4-aminopyridine response in a large downbeat nystagmus cohort

Abstract

Background: GAA-FGF14 disease/spinocerebellar ataxia 27B is a recently described neurodegenerative disease caused by (GAA)_≥250 expansions in the fibroblast growth factor 14 (FGF14) gene, but its phenotypic spectrum, pathogenic threshold, and evidence-based treatability remain to be established. We report on the frequency of FGF14 (GAA)_≥250 and (GAA)_200-249 expansions in a large cohort of patients with idiopathic downbeat nystagmus (DBN) and their response to 4-aminopyridine.

Methods: Retrospective cohort study of 170 patients with idiopathic DBN, comprising in-depth phenotyping and assessment of 4-aminopyridine treatment response, including re-analysis of placebo-controlled video-oculography treatment response data from a previous randomised double-blind 4-aminopyridine trial.

Findings: Frequency of FGF14 (GAA)_≥250 expansions was 48% (82/170) in patients with idiopathic DBN. Additional cerebellar ocular motor signs were observed in 100% (82/82) and cerebellar ataxia in 43% (35/82) of patients carrying an FGF14 (GAA)_≥250 expansion. FGF14 (GAA)_200-249 alleles were enriched in patients with DBN (12%; 20/170) compared to controls (0.87%; 19/2191; OR, 15.20; 95% CI, 7.52-30.80; p < 0.0001). The phenotype of patients carrying a (GAA)_200-249 allele closely mirrored that of patients carrying a (GAA)_≥250 allele. Patients carrying a (GAA)_≥250 or a (GAA)_200-249 allele had a significantly greater clinician-reported (80%, 33/41 vs 31%, 5/16; RR, 2.58; 95% CI, 1.23-5.41; Fisher's exact test, p = 0.0011) and self-reported (59%, 32/54 vs 11%, 2/19; RR, 5.63; 95% CI, 1.49-21.27; Fisher's exact test, p = 0.00033) response to 4-aminopyridine treatment compared to patients carrying a (GAA)_<200 allele. Placebo-controlled video-oculography data, available for four patients carrying an FGF14 (GAA)_≥250 expansion, showed a significant decrease in slow phase velocity of DBN with 4-aminopyridine, but not placebo.

Interpretation: This study confirms that FGF14 GAA expansions are a frequent cause of DBN syndromes. It provides preliminary evidence that (GAA)_200-249 alleles might be pathogenic. Finally, it provides large real-world and preliminary piloting placebo-controlled evidence for the efficacy of 4-aminopyridine in GAA-FGF14 disease.

Funding: This work was supported by the Clinician Scientist program "PRECISE.net" funded by the Else Kröner-Fresenius-Stiftung (to CW, AT, and MSy), the grant 779257 "Solve-RD" from the European's Union Horizon 2020 research and innovation program (to MSy), and the grant 01EO 1401 by the German Federal Ministry of Education and Research (BMBF) (to MSt). This work was also supported by the Deutsche Forschungsgemeinschaft (DFG, German Research Foundation) N° 441409627, as part of the PROSPAX consortium under the frame of EJP RD, the European Joint Programme on Rare Diseases, under the EJP RD COFUND-EJP N° 825575 (to MSy, BB and-as associated partner-SZ), the NIH National Institute of Neurological Disorders and Stroke (grant 2R01NS072248-11A1 to SZ), the Fondation Groupe Monaco (to BB), and the Montreal General Hospital Foundation (grant PT79418 to BB). The Care4Rare Canada Consortium is funded in part by Genome Canada and the Ontario Genomics Institute (OGI-147 to KMB), the Canadian Institutes of Health Research (CIHR GP1-155867 to KMB), Ontario Research Foundation, Genome Quebec, and the Children's Hospital of Eastern Ontario Foundation. The funders had no role in the conduct of this study.

Keywords: 4-Aminopyridine; Downbeat nystagmus; GAA-FGF14 ataxia; SCA27B; Treatment; Trial.

Fig. 1

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

Fig. 2

Frequency of FGF14 GAA repeat expansions in DBN syndromes. (a) Allele distribution of the FGF14 repeat locus in 170 patients with idiopathic DBN syndromes (340 chromosomes). The density plot shows allele size frequencies, with higher densities indicating greater frequencies. The box-and-whisker plot above the graph shows the allele distribution. The box indicates the 25th percentile (first quartile), the median, and the 75th percentile (third quartile), and the whiskers indicate the 25th percentile minus 1.5 x IQR and the 75th percentile plus 1.5 x IQR. Outliers are represented by black dots. The dashed gray line and the shaded gray area indicate the so-called “intermediate” allele range of (GAA)_200-249, and the dashed red line represents the pathogenic threshold of (GAA)_≥250 repeat units. (b) Percentage of patients carrying an FGF14 (GAA)_≥250 expansion (dark blue), an FGF14 (GAA)_200-249 allele (light blue), and an FGF14 (GAA)_<200 allele (gray) in the subgroups with (1) pure DBN (0/7, 0/7, and 7/7 patients, respectively), (2) DBN plus additional isolated cerebellar ocular motor signs (DBN + COM) (42/75, 10/75, and 23/75 patients), (3) DBN plus cerebellar ataxia (DBN + CA) (23/32, 2/32, and 7/32 patients), and (4) DBN plus cerebellar ocular motor signs and/or ataxia and extracerebellar features (DBN + EC) (17/56, 8/56, and 31/56 patients). (c) Frequency of presenting symptoms in 81 patients with DBN carrying an FGF14 (GAA)_≥250 expansion. Data on presenting symptoms were missing for one patient. Patients may present with multiple symptoms at disease onset. Visual disturbances include diplopia, oscillopsia, and visual blurring. (d) Weak inverse association between size of the repeat expansion and age at disease onset in 74 patients carrying an FGF14 (GAA)_≥250 expansion for whom information on age at onset was available (Spearman's r, −0.27; 95% CI, −0.47 to −0.03; p = 0.022). (e) UpSet plot showing co-occurrence of cerebellar ocular motor signs (COM), cerebellar ataxia (CA), bilateral vestibulopathy (BVP), and polyneuropathy among 82 patients with DBN carrying an FGF14 (GAA)_≥250 expansion.

Fig. 3

Phenotypic evolution of the DBN endophenotypic cluster of GAA-FGF14 disease. (a) Kaplan–Meier curve showing the probability of being disease free as a function of the age (years) for 74 patients carrying an FGF14 (GAA)_≥250 expansion for whom age at onset was available. Median survival before occurrence of disease was 67 years. For clarity, ages below 20 years were not included on the graph, as none of the patients developed GAA-FGF14 disease before the age of 30 years. The analysis commences upon the onset of GAA-FGF14 disease in the first patient and ends when all patients in the cohort have developed the disease. The shaded gray area displays the 95% confidence interval around the probability estimate. (b) Disease duration at last examination for patients carrying an FGF14 (GAA)_≥250 expansion with DBN plus additional isolated cerebellar ocular motor signs (DBN + COM, n = 36; median, 4.25 years; IQR, 3–8 years), DBN plus cerebellar ataxia (DBN + CA, n = 21; median, 6 years; IQR, 3–8 years), and DBN plus cerebellar ocular motor signs and/or ataxia and extracerebellar features (DBN + EC, n = 17; median, 9 years; IQR, 5.5–14). Disease duration was significantly longer for patients with extracerebellar features (Kruskal–Wallis test, p = 0.014). (c) Temporal evolution of select phenotypic features relative to the onset of gait impairment (dotted line). The solid black lines show the median time from onset of gait impairment for each individual feature. (d) Occurrence of cerebellar atrophy on brain MRI (n = 7), polyneuropathy on nerve conduction studies (n = 7), and BVP on caloric stimulation or vHIT (n = 7) in relation to disease onset. BVP occurred later in disease course compared to cerebellar ataxia and polyneuropathy (Kruskal–Wallis test, p = 0.0041). In panels b and d, the box indicates the 25th percentile (first quartile), the median, and the 75th percentile (third quartile), and the whiskers indicate the 25th percentile minus 1.5 x IQR and the 75th percentile plus 1.5 x IQR. Outliers are represented by black dots. In panels b and d, the individual between-group adjusted p values (q values) as per the Benjamini-Hochberg method are shown in the graphs. Adjusted p values < 0.05 indicate statistically significant difference.

Fig. 4

Progression of disability of the DBN endophenotypic cluster of GAA-FGF14 disease. (a) Cross-sectional progression of the functional impairment as assessed by the FARS functional disability stage relative to disease duration (n = 73 patients carrying an FGF14 [GAA]_≥250 expansion). The FARS functional stage was not significantly associated with disease duration (Spearman's r, 0.194; 95% CI, −0.04–0.41; p = 0.10). (b) Longitudinal intra-individual progression of functional impairment as assessed by the FARS functional disability stage relative to disease duration (148 observations from 40 patients carrying an FGF14 [GAA]_≥250 expansion are shown). Observations from the same patient are connected by a dotted line. The solid black line shows the average progression of the FARS stage over disease duration across all patients as modelled by a linear mixed-effects model accounting for disease duration as fixed effect and with random intercepts and random slopes. In both panels, only FARS stages measured off 4-AP treatment are shown. Two patients who became wheelchair-dependent following a hip fracture and a prolonged medical admission, respectively, are represented by orange squares in panel a and by squares in panel b. These patients also carried alleles of 218 and 270 repeat units and 280 and 313 repeat units, respectively.

Fig. 5

Association of the rs72665334 variant with FGF14 GAA expansions. (a) Diagram of the FGF14 gene, isoform 1b showing the location of the rs72665334 variant (GRCh38, chr13:102,150,076) in relation to the (GAA)·(TTC) repeat locus in the first intron (GRCh38, chr13:102,161,575-102,161,726). The rs72665334 variant is located ∼11 kb 5′ of the intronic GAA short tandem repeat. (b) Percentage of patients with DBN carrying an FGF14 (GAA)_≥250 expansion (n = 35; dark blue), an FGF14 (GAA)_200-249 allele (n = 9; light blue), and an FGF14 (GAA)_<200 allele (n = 29; gray) with the C|C, C|T, and T|T rs72665334 genotypes. There was a significantly greater proportion of patients with a C|T or T|T genotype who carried an FGF14 (GAA)_≥250 expansion compared to an FGF14 (GAA)_<200 allele (54%, 19/35 vs 14%, 4/29; OR, 7.18; 95% CI, 1.91–34.43; Fisher's exact test p = 0.0014). (c) Frequency of the C|C, C|T, and T|T rs72665334 genotypes in 503 European controls from the 1000 Genomes Project and 72 patients with DBN carrying an FGF14 (GAA)_≥250 expansion (n = 35 from the idiopathic DBN cohort and n = 37 from the French-Canadian cohort). The C|T and T|T genotypes were significantly more frequent in patients with DBN carrying an FGF14 (GAA)_≥250 expansion compared to controls (46%, 33/72 vs 12%, 60/503; OR, 6.22; 95% CI, 3.51–11.02; Fisher's exact test, p < 0.0001).

Fig. 6

Treatment response to 4-aminopyridine. Open-label real-world treatment response data. Percentage of patients presenting (a) a clinician-reported response and (b) a patient-reported response to 4-AP treatment in the subgroups with an FGF14 (GAA)_≥250 expansion (29/36 and 29/49 patients, respectively), (GAA)_200-249 allele (4/5 and 3/5 patients), and (GAA)_<200 allele (5/16 and 2/19 patients). Double-blind placebo-controlled clinical trial data. (c) Effect of 4-AP on the slow phase velocity (SPV) of DBN in a randomised double-blind trial. Compared to baseline, the absolute value of the SPV significantly decreased under 4-AP 60 min after the first dose and under daily treatment, but not under placebo. Note that all four patients were randomised to receive active treatment first (=4-AP) and placebo second. SPV during all three measurements of the placebo phase was higher than at the baseline of the treatment phase. The improved values of the SPV under 4-AP were in the range of the placebo values. (d) Treatment effect of 4-AP, illustrated by relative reduction (improvement) of SPV as potential future trial outcome assessment. The genotype of the four patients was: COR4418, 65 and 298 repeat units; COR5001, 44 and 354 repeat units; COA7220, 16 and 265 repeat units; COR1375, 280 and 313 repeat units. In panels c and d, the black line shows the mean value for all four patients at each time point.

Fig. 7

Effect of 4-aminopyridine on functional impairment. Longitudinal intra-individual progression of functional impairment while on and off 4-AP treatment as assessed by the FARS functional disability stage relative to disease duration for (a–g) seven patients carrying an FGF14 (GAA)_≥250 expansion and (h and i) two patients carrying an FGF14 (GAA)_200-249 allele. FARS stages were lower while on 4-AP treatment for all nine patients, indicating an improvement in the level of disability. No patients carrying an FGF14 (GAA)_<200 allele had FARS stages recorded while on and off 4-AP treatment. The genotype of the nine patients was: (a) 29 and 263 repeat units; (b) 18 and 292 repeat units; (c) 9 and 348 repeat units; (d) 16 and 288 repeat units; (e) 8 and 500 repeat units; (f) 16 and 424 repeat units; (g) 16 and 336 repeat units; (h) 34 and 223 repeat units; (i) 9 and 242 repeat units.