Exonic trinucleotide repeat expansions in ZFHX3 cause spinocerebellar ataxia type 4: A poly-glycine disease

Abstract

Autosomal-dominant ataxia with sensory and autonomic neuropathy is a highly specific combined phenotype that we described in two Swedish kindreds in 2014; its genetic cause had remained unknown. Here, we report the discovery of exonic GGC trinucleotide repeat expansions, encoding poly-glycine, in zinc finger homeobox 3 (ZFHX3) in these families. The expansions were identified in whole-genome datasets within genomic segments that all affected family members shared. Non-expanded alleles carried one or more interruptions within the repeat. We also found ZFHX3 repeat expansions in three additional families, all from the region of Skåne in southern Sweden. Individuals with expanded repeats developed balance and gait disturbances at 15 to 60 years of age and had sensory neuropathy and slow saccades. Anticipation was observed in all families and correlated with different repeat lengths determined through long-read sequencing in two family members. The most severely affected individuals had marked autonomic dysfunction, with severe orthostatism as the most disabling clinical feature. Neuropathology revealed p62-positive intracytoplasmic and intranuclear inclusions in neurons of the central and enteric nervous system, as well as alpha-synuclein positivity. ZFHX3 is located within the 16q22 locus, to which spinocerebellar ataxia type 4 (SCA4) repeatedly had been mapped; the clinical phenotype in our families corresponded well with the unique phenotype described in SCA4, and the original SCA4 kindred originated from Sweden. ZFHX3 has known functions in neuronal development and differentiation n both the central and peripheral nervous system. Our findings demonstrate that SCA4 is caused by repeat expansions in ZFHX3.

Keywords: GGC repeats; SCA4; ZFHX3; anticipation; autosomal dominant; interruptions; p62-positive inclusions; poly-glycine; spinocerebellar ataxia 4; trinucelotide repeats.

Graphical abstract

Figure 1

Family pedigrees Standard symbols were used. Black symbols represent individuals with cerebellar ataxia with sensory and autonomic neuropathy. Probands are indicated with black triangles. Gray symbols show family members who had gait and balance problems according to family history. Gray symbols with a central black dot represent two siblings who were evaluated within this study but were considered to have a neurodevelopmental disorder that is different from cerebellar ataxia with sensory and autonomic neuropathy, and who do not have ZFHX3 repeat expansions. The black numbers above the symbols are individual identifiers. For those family members for whom age at symptom onset was available, it is shown below the symbols in blue. Sex and gender of non-affected individuals was disguised, and sibling order partly altered to protect confidentiality. For the same reason, a few unaffected family members are not shown. UFM, unaffected family member who was analyzed genetically within this study.

Figure 2

ZFHX3 locus harboring the repeat expansion encoding poly-glycine ZFHX3 is encoded on the negative strand in the hg19 reference genome. For clarity, the negative strand sequence is included in the figure. Unless otherwise specified, we discuss the repeat region from the perspective of the negative strand. (A) In the vast majority of non-expanded alleles, the repeat region consists of exactly 18 GGC units, two GGT units, and a single AGT interruption; we report the total repeat length as 21 for these alleles. The locations of the GGT units (which also code for glycine but interrupt the repetitive GGC pattern on the DNA level) are given relative to the AGT interruption (light gray numbers). Such a normal allele results in a protein with 20 glycine residues interrupted by a single serine at position 7. This poly-glycine is followed by a single serine residue and a C2C2-type zinc finger motif. Generally, zinc finger motifs are considered elements that can bind to DNA in a sequence-specific manner. See main text for the exact composition of this locus in the minority of non-expanded alleles. (B) There were no detectable interruptions in any of the expanded alleles from the affected members of the five families described here—the repeat region was entirely composed of GGC units.

Figure 3

ZFHX3 repeat lengths in non-expanded alleles Distribution of the total number of STR repeat units, including potential interruptions, as determined from the DNA sequence of 1,000 unaffected individuals from the Swedish SweGen WGS database, 150 in-house WES or WGS datasets without the ZFHX3 repeat expansion, and 27 PacBio datasets from the Human PanGenome Reference Consortium. Numbers above the bars indicate the number of alleles. By contrast, expanded alleles in affected individuals (not shown above) were 42–72 GGC repeats in length, as determined from short-read WGS data, and 57 or 74 GGC repeats as determined by long-read sequencing of two affected individuals (see Table 1 and text), offering clear delineation between non-expanded and expanded alleles.

Figure 4

Neuropathology (A–C) Hematoxylin and eosin (H & E)-stained tissue from the medulla oblongata with eosinophilic, intracytoplasmic (A), and intranuclear (B and C) inclusions in neurons. (D) These inclusions stained positively with antibodies against p62. (E and F) Similar immunoreactivity against p62 was seen in inclusions in neurons of the esophageal myenteric plexus. (G and H) Alpha-synuclein-positive neurites and fine granular intracytoplasmic immunoreactivity in the medulla oblongata.