Background

The hereditary ataxias are a highly genetically heterogeneous group of disorders phenotypically characterized by gait ataxia, incoordination of eye movements, speech, and hand movements, and usually associated with cerebellar atrophy. Autosomal dominant forms (Spinocerebellar ataxia, SCA) typically have adult-onset; conversely autosomal recessive ataxias (Spinocerebellar ataxias autosomal recessive, SCAR) usually have onset in childhood [1,2].

Genetic tests needed to define the subtype are often long and laborious, due to the size of many ataxia genes (e.g., SYNE1, SACS, ATM, and SETX).

As a possible alternative to the gene-by-gene mutation screening, the application of Next Generation Sequencing technologies (NGS), involving simultaneous analysis of hundreds of selected exons or the entire sequencing of human coding genes (exome) are increasingly being used as a diagnostic tool, because of affordable costs and the availability of efficient bioinformatics tools. This approach becomes even more efficient in the case of a heterogeneous phenotype or unspecific clinical symptoms, which may make difficult a correct genetic diagnosis.

Beyond the mutation detection which is important to define a diagnosis for genetic counselling, future prenatal diagnosis and eventually therapy, exome analysis may also reveal unexpected findings, such as possible disease modifier genes [3].

In this study, we analyzed a single family with three sisters affected by ataxia with an atypical postural and intentional myoclonus. Exome sequencing revealed a homozygous SETX gene mutation, and an unexpected c.346G > A (p.Gly116Arg) variant in the AFG3L2 gene -responsible of SCA28- predicted to alter protein function. Clinical re-evaluation of the patients confirmed some typical feature of AOA2 and identified a fourth sibling carrying p.Gly116Arg with initial signs of myoclonus, suggesting variants in AFG3L2 may be also associated with myoclonus.

Methods

Study subjects

We identify three subjects from family ATA-2-TO (II.1, II.2, II.3 in Figure 1) presented with unsteadiness and difficulties in walking accompanied by a large spectrum of involuntary movements. Due to the early death of the mother and the unavailability of the father, we did not initially ascertain the type of transmission. As an initial workout, we excluded ataxias due to polyglutamine expansion (SCA1-3, 6, 7) and Friedreich ataxia; alpha-fetoprotein was available only at the end of the study. Short Tandem Repeats (STR) typing excluded SCA14 and SACS loci.

Figure 1
figure 1

Family tree, mutation analysis, and yeast functional assay. A. ATA-2-TO family tree suggested an autosomal recessive transmission, but clinical data from the father and the mother (died at 37 yrs) were incomplete. Below the segregation of three microsatellite markers surrounding the SETX gene show a homozygous region in the three affected individuals. Subject II.7 had subclinical myoclonus (see text). The presence of SETX c.6269C > T and AFG3L2 c.346G > A mutations are indicated below each subject. B. Electropherograms of SETX c.6269C > T and AFG3L2 c.346G > A mutations with accompanying amino acid changes. C. Evolutionary conservation of glycine at position 116 in the AFG3L2 protein. The position is conserved in all vertebrates analyzed. D. Yeast functional assay of p.Gly116Arg mutation [4] m-AAA proteases are evolutionarily highly conserved from human through to S. cerevisiae where both homo-oligomeric Yta12 (AFG3L2 orthologue), and hetero-oligomeric Yta10 (SPG7 orthologue) - Yta12 subunits are present. These proteins can be substituted with the human orthologues in order to study specific missense changes. In Δyta10Δyta12 yeast cells (ΔΔ) the synthesis of mitochondrial-encoded respiratory chain subunits is impaired and cells are unable to perform aerobic respiration, growing only on carbon fermentable source [Yeast extract Peptone and Dextrose (glucose): YPD] [4-6] and not on a non-fermentable source (Yeast extract Peptone and Glycerol: YPG). Δyta10Δyta12 yeast transfected with human AFG3L2 is able to grow on YPG, whereas human AFG3L2G116R is unable. The co-transfection of SPG7 and AFG3L2G116R rescued the phenotype.

Full size image

To evaluate the possible consanguinity in family ATA-2-TO, after mutation discovery, we studied the following markers in the 9q34 region surrounding SETX: chr9_135.0 (5′-tgttatttgcactgggagga; 5′-agctcaatgtagctgttatttttct); chr9_135.3 (5′-atgggatacatgcacacgttta; 5′-caatgcagctcacttgaaaactta); D19S149 (5′-gattgacctgtgaatttgtacagc; 5′-tgttatgccttgctgttgct). Research was approved by the Internal Review Board of the Department of Medical Sciences, University of Torino. A written informed consent for publication was obtained from members participating in the study.

Genetic analyses

Genomic DNA was extracted from peripheral blood (Qiagen, Hilden, Germany) following the manufacturer instructions.

Exome capture was performed with Agilent SureSelect Human All Exon kit (Agilent Technologies, Santa Clara, CA, USA) according to the manufacturer protocols.

Sequencing was performed on an Illumina HiSeq 2000 machine using a standard pair-end read sequencing protocol (Illumina, San Diego, CA, USA) to generate sequencing reads of up to 90 cycles. We applied default settings in the Illumina pipeline to call bases from raw images, generating raw sequencing reads in the format of fastq files. We subsequently applied two independent analysis methods to perform alignment, variant calling and annotation. Pipeline 1 (P1) aligned fastq files to the human reference genome (UCSC hg19) with BWA then used Genome Analysis Tool Kit (version 1.4) to call variants, followed by utilizing Annovar and SnpEff for the variant functional annotation step. Pipeline 2 (P2) mapped fastq files to the human reference genome (UCSC hg19) with SOAP (version 2.21) and used SOAPsnp (version 1.05) for single nucleotide variant (SNV) detection, and used GATK for small insertion-deletion (indel) detection, followed by BGI self-developed programs to perform variant functional annotation.

We excluded variants that were: 1) out of exonic regions, 2) synonymous changes and 3) with minor allele frequency (MAF) > 0.5% in either the 1000 Genomes Project (http://www.1000genomes.org/), the NHLBI Exome Sequencing Project (ESP6500; http://evs.gs.washington.edu/EVS/), or our internal exome datasets. Variants near splicing donor/recipient sites and frameshift indels were given particular attention as they could cause pathogenic changes like exon-skipping or frameshifts. We filtered variants based on 1) evolutionary conservation, ie variants of PhyloP value <0.95 were considered to be in non-conserved regions thus discarded; 2) prediction of pathogenicity by PolyPhen and SIFT software.

Validation of the mutation was performed by Sanger sequencing in all members of the family with primers designed to amplify the exon 17 region of SETX (Reference sequence: NM_015046.5)(5′-tgggaactaattttccttctcattg and 5′-agcctggaagacagagcaagact).

AFG3L2 mutation analysis and functional assay

Clustal omega (http://www.clustal.org/omega/) was used for evolutionary conservation analysis. AFG3L2 (NM_006796) exon 4 was amplified and sequenced accordingly to previous published protocol [7]. For the yeast complementation assay a combination of human wild-type or mutant human AFG3L2 was expressed with or without wild-type paraplegin in a previously generated yeast strain Δyta10Δyta1.[5] AFG3L2 was mutagenized in the yeast expression constructs using the QuickChange XL Site-Directed Mutagenesis Kit (Stratagene, LaJolla, CA). Yeast cells were grown according to standard procedures at 30°C in YP medium containing 2% (w/v) dextrose (glucose) (YPD). For testing the respiratory activity, yeast cells were grown at 30°C on solid state or in the liquid YP media containing 3% (w/v) glycerol as the sole carbon source (YPG). Experiments were repeated three times with overlapping results.

Results

Genetic analysis

Whole exome sequencing on three family members (II.1, II.3, II.7) was performed with at least 57X coverage depth (Additional file 1: Table S1 and S2). Two analysis pipelines were used for each exome: P1 called ~45,000 SNVs and ~7,500 indels, P2 ~ 74,000 SNVs and ~5,000 indels. Further analysis was restricted to variants occurring in coding regions or at splicing sites, reducing SNVs to ~17,000 and indels to ~700 in P1, and ~17,000 SNVs and ~630 indels in P2 (Additional file 1: Table S1).

We initially analyzed the coding region of known ataxia causing genes assuming an autosomal recessive or dominant model of inheritance. We retained only rare (MAF < 0.5% in 1000 Genomes Project, the ESP6500 and CAG/BGI internal exome datasets) non-synonymous variants or splice site variants. A single homozygous C to T variant on chr 9:135,163,655 causing a c.6292C > T (p.Arg2098*) in the SETX gene (NM_015046.5, OMIM *608465, http://www.omim.org/) was left after both P1 and P2 lists (Additional file 1: Table S3). This homozygous mutation was confirmed by Sanger sequencing in all affected individuals (Figure 1B), and haplotype analysis showed that it was derived from a common ancestor (Figure 1A). Although consanguinity in patients parents was denied, they were both born in a small village in Sicily increasing the likelihood they were consanguineous. The three sisters (II.1, II.2 and II.7) also shared the c.346G > A variant in the AFG3L2 gene changing amino acid glycine 116 into arginine (Additional file 1: Table S4). AFG3L2 encodes for a subunit of the hetero-oligomeric m-AAA protease (ATPases associated with various cellular activities), a component of the mitochondrial ATP-dependent metalloprotease located on the inner mitochondrial membrane, and its mutations in heterozygosis are associated with Spinocerebellar ataxia type 28 (SCA28) [8].

Segregation analysis showed that this variant was present also in the third affected sister II.3, but it was absent in II.4, II.5 and II.6 (Figure 1A). This substitution is not reported as validated polymorphism in Human variation resources (dbSNP(138), 1000 genomes, EVS, ExAC or HapMap databases), hit a highly conserved residue through all vertebrates (Figure 1C), it was predicted to be highly pathogenic by Mutation taster software (probability: 0.999; www.mutationtaster.org/) and to affect protein function with a score of 0.03 by SIFT Alignment (Median sequence conservation: 3.34; sift.jcvi.org/). The substitution is between the two mitochondrial transmembrane domains of AFG3L2.

We introduced the c.346G > A (p.Gly116Arg) mutation in the human full length AFG3L2 cDNA. Expression in Δyta10Δyta12 cells did not restore mitochondrial respiratory activity as shown by the inhibition of growth on glycerol-containing media. Co-expression of SPG7 with AFG3L2G116R in Δyta10Δyta12 cells substantially restored respiratory growth to normal levels (Figure 1D).

Clinical findings

Between 9 and 12 yrs, three subjects from family ATA-2-TO (II.1, II.2, II.3 in Figure 1A) presented with unsteadiness and difficulties in walking accompanied by a large spectrum of involuntary movements such as myoclonic jerks at arms and hands, and/or choreic movements and facial dyskinesia. Myoclonic jerks were spontaneous and stimulus-insensitive, of small amplitude and arrhythmic. They were present at rest, on sustained posture and on action, involved proximal and distal segments of the upper limbs, and the perioral region. They attenuated over the course of the disease without disappearing.

Basal electroencephalography (EEG) and polygraphic EEG-EMG were normal and disease progression was slow. Patients developed truncal and limb ataxia with a progressive sensory-motor neuropathy that led them to difficulty walk without help after 5–8 years of disease. All patients were wheelchair-bound before 30 yrs.

Further signs included areflexia, distal absence of vibratory sensibility of the knees, and altered Somatosensory Evoked Potential.

Oculomotor abnormalities included nystagmus in all directions, hypometric saccades, saccadic pursuit, but no overt oculomotor apraxia. A mild limitation of the bilateral horizontal gaze was observed in II.3 (convergent strabismus) and a mild eyelid ptosis in II.1 was present after more than 30 years of disease.

One patient (II.3) showed a borderline intellectual disability (IQ = 77).

At last examination (~40 yrs of disease duration, summary in Table 1), subjects presented with severe ataxia. Notably, all cases had an early menopause.

Table 1 Neurological and biochemical features of AOA2 patients
Full size table

Skeletal abnormalities included a kyphoscoliosis (3/3) whereas pes cavus was not present.

Serum alpha-fetoprotein, obtained only at the last examination, was increased (Table 1).

Neuroradiology showed a marked cerebellar atrophy of hemispheres and vermis (Additional file 2: Figure S1).

Parents were reported healthy. The mother died at age 37 from pregnancy related complications, the father living at age 72 has been reported as unaffected, but he was unavailable for clinical examination.

Patient II.7

All siblings of the three affected individuals underwent a neurological investigation and did not show any neurologic phenotype except II.7. She reported occasional experience of imbalance, mainly after fatigue or stress at 36 yrs. At 42 yrs, neurological examination showed the presence of sporadic spontaneous myoclonus at proximal upper limbs and at the perioral district, confirmed by electromyography. Hyperreflexia at upper limbs, hyperactive patellar and ankle reflexes with a bilateral foot clonus and mild hypertonia at lower limbs with a weak Babinski sign, suggested a mild pyramidal involvement. No cerebellar ataxia, oculomotor abnormalities, and sensory deficits were present. MRI was not available.

Discussion

Diagnosis for rare inherited ataxias is hampered by the clinical overlap of heterogeneous disorders, and routine tests are often limited to repeat expansions, with rarer genes interrogated only in a research based setting. Targeted resequencing and exome sequencing are rapidly changing this approach demonstrating their utility as a diagnostic tool [9].

Ataxia and sensory-motor neuropathy are typical symptoms of AOA2 that may be complicated by additional movement disorders during the disease course. Chorea/choreoathetosis, facial dyskinesia, dystonia, head/postural tremor are described in 14% of AOA2 patients, but they were never reported at onset [10-12]. Myoclonus is described in only one AOA2 case [13].

We studied an atypical form of AOA2 with involuntary movements/myoclonus at onset shared by three sisters. None had oculomotor apraxia (OMA), reported in half of the AOA2 cases. However, all had hypometric saccades with a staircase pattern, a sign recently suggested as more reliable for OMA [14].

A mild limitation of the lateral gaze and a mild convergent strabismus in the left eye occurred in II-3 after 26 yrs. of disease (12-38% of AOA2 patients) [10,11]. Ptosis, never reported in AOA2, may be a late sign of disease, and appeared unilaterally in the eldest patient at 51 yrs. Further uncommon features included a mild cognitive impairment in II.3 [11]. Skeletal abnormalities included a kyphoscoliosis (3/3) never reported in AOA2, whereas pes cavus, a finding present in almost all reported AOA2 cases was not present [10,11].

We noted that all three sisters presented a premature menopause between 32 and 34 yrs, a feature already occasionally reported in AOA2 [15,16]. This may suggest an association between SETX mutations and premature ovarian failure/premature menopause that deserves further studies.

Among the other four siblings, II.7 showed spontaneous mild myoclonus and hyperreflexia at upper and lower limbs.

Exome sequencing identified a homozygous c.6292C > T (p.Arg2098*) mutation in this gene. This same mutation in homozygosis had already been described in six AOA2 Tunisian patients (Family E) [17], who showed a typical form of disease without involuntary movements/myoclonus. This finding suggested that a second genetic determinant may have combined its effect with the SETX mutation to give the atypical phenotype in our family. Exome data also revealed the c.346G > A (p.Gly116Arg) missense change in AFG3L2, encoding for a mitochondrial protease. In the inner membrane of human mitochondria two different functionally active m-AAA isoenzymes are present: a homo-oligomeric AFG3L2 complex and a hetero-oligomeric complex where AFG3L2 subunits assemble with paraplegin. Mutations in this gene cause both an autosomal dominant spinocerebellar ataxia (SCA28) characterized by ophtalmoplegia/ptosis and oculomotor alterations [7,8] or an autosomal recessive form of early onset spastic ataxia with neuropathy and myoclonic epilepsy (SPAX5) [18].

Yeast experiments demonstrated that only homo-oligomeric m-AAA protease complexes composed by AFG3L2G116R subunits are functionally impaired. We concluded that AFG3L2G116R is likely pathogenic, but behaves differently from AFG3L2 mutations causing SCA28, that fail to be rescued by paraplegin co-expression [8]. AFG3L2 mutations reported in the literature mainly cluster in the protease C-terminal domain [7,8,19]. Localization of p.Gly116Arg at the N-terminal portion of the protein, suggests a different effect on the protein that may lead to a so far never reported phenotype.

Indeed, p.Gly116Arg was present in the three affected sisters and II.7 only. The latter had subclinical myoclonus at age 41. All four siblings without AOA2 (including II.7) were carriers of the c.6292C > T mutation in SETX.

Conclusion

Using exome sequencing, we identified an atypical form of AOA2 with myoclonus at onset associated with the c.6292C > T (p.Arg2098*) homozygous mutation. We hypothesize that the combination of AOA2 with the c.346G > A (p.Gly116Arg) mutation in AFG3L2, is responsible for the variant phenotype seen in three patients. The p.Gly116Arg variant in AFG3L2 may be a candidate for myoclonus, but we cannot exclude this is an initial symptom that will evolve into a more complex phenotype.

We suggest that variant phenotypes may be due to the combined effect of different mutated genes associated to ataxia or related disorders, that will become more apparent as the costs of exome sequencing progressively will reduce, amplifying its diagnostics use, and meanwhile proposing significant challenges in the interpretation of the data.