Advancing molecular, phenotypic and mechanistic insights of FGF14 pathogenic expansions (SCA27B)

Abstract

Repeat expansions in the FGF14 gene have recently been identified as a frequent cause of autosomal dominant late-onset cerebellar ataxia (SCA27B). The threshold for pathogenicity was estimated to range from 250 (incomplete penetrance) to 300 AAG repeats (full penetrance) based on expansion sizes observed in patients and controls. However, the full sequences of pathogenic and non-pathogenic alleles remain largely unknown. In this study, we used STRling and ExpansionHunter/STRipy to detect short tandem repeat expansions in short-read genome data of 80 patients with unsolved neurological disorders, including 48 patients from 39 unrelated families with cerebellar ataxia. Significant outlier values indicating a possible FGF14 repeat expansion were detected in 49% of families with ataxia. We used long-range PCR and repeat-primed PCR to confirm repeat motifs and number, and to further screen for FGF14 repeat expansions in 106 additional patients. The distribution of FGF14 alleles was also analyzed in 802 control individuals. Long-range PCR amplicons from affected and control subjects with at least one expanded allele were sequenced by nanopore sequencing. Altogether, we report a total of 38 individuals from 31 families with a pure AAG expansion [≥]250 repeats in FGF14 (range: 254-937) out of 134 families with unsolved ataxia (23%). Four families had biallelic expansions. One individual showed mosaicism of the expansion, with two distinct large alleles detected. The overall distribution of FGF14 alleles significantly differs in patients and control subjects with pure expansions from 180 repeats being enriched in patients while AAGGAG and interrupted alleles were more frequent in controls. Expanded and non-expanded FGF14 alleles were associated with different 5-flanking regions correlating with repeat stability. Episodic symptoms of ataxia and downbeat nystagmus were frequent in patients with SCA27B independently of the repeat number. A family with a novel nonsense variant (NM_175929.3: c.239T>G; p.Leu80*; SCA27A) exhibited similar symptoms but earlier age at onset than patients with pathogenic expansions (SCA27B). In conclusion, this study reveals the complete sequence of pathogenic and non-pathogenic FGF14 expansions. We suggest that pure AAG expansions are pathogenic from a lower threshold than previously reported, comprised between 180 and 220 pure AAG repeats, while full penetrance would be above 320-335 repeats. Interrupted expansions and expansions composed of other hexanucleotide repeats motifs are non-pathogenic. Diagnostic tests should thus include a complete sequencing of the expansion in addition to repeat number assessment. Finally, we provide mechanistic insights into how pure AAG repeat expansions could lead to adult-onset cerebellar ataxia.