SPG7 mutations are a common cause of undiagnosed ataxia

Exome sequencing in 2 undiagnosed ataxia patients identified compound heterozygous SPG7 mutations, not previously considered likely in the absence of pyramidal signs. This prompted us to prospectively study SPG7 in 70 other probands with undiagnosed ataxia and pyramidal signs attending routine follow-up over a 12-month period. Other sporadic and inherited causes of ataxia were excluded, including inflammatory, metabolic, neoplastic, and sporadic degenerative ataxia; spinocerebellar ataxia 1, 2, 3, 6, 7, 10, 12, and 17; dentatorubral-pallidoluysian atrophy; and Friedreich ataxia. Mutations in SPG7 were detected by Sanger sequencing of all 17 coding exons and multiplex ligation-dependent probe amplification analysis (MRC-Holland kit P213-B1, Amsterdam, the Netherlands). All patients provided written informed consent.

Exome sequencing identified 2 SPG7 mutations in patient 1 (c.1529C>T/p.Ala510Val and c.1715C>T/p.Ala572Val) and 2 in patient 3 (c.1529C>T/p.Ala510Val and c.1192C>T/p.Arg398*). These were confirmed by Sanger sequencing, and present in their affected siblings (patients 2 and P4, respectively). The variants were heterozygous in the unaffected parents, and were previously reported as pathogenic.¹ No other recessive mutations in relevant disease genes were identified (tables e-1 to e-4 on the Neurology® Web site at Neurology.org).

Of the 70 patients subsequently studied, 13 had likely recessive mutations (4 homozygous and 9 compound heterozygous). Two patients had novel mutations (c.1225_1229del/p.Glu409Arg_fs49* and c.2228T>C/p.Ile743Thr). All patients had the c.1529C>T/p.Ala510Val mutation on at least one allele. All patients were of British descent. No rearrangements were detected. The clinical features are summarized in table 1.

At initial presentation, all the patients presented with symptoms of ataxia or gait disturbance (mean age 36.3 years, SD 12.5). Midline ataxia was present in all patients at onset, with gait ataxia present in all patients and ocular signs in 5. Eleven patients (65%) had no pyramidal signs (normal reflexes, and no spasticity). Ocular findings were present in 5 (29%) patients.

On follow-up examination, most patients (76%) developed appendicular ataxia. All developed clear pyramidal signs: 12 had overt spasticity, and the remainder developed brisk tendon reflexes or extensor plantar responses. Ocular findings were present in 11 (65%), with nystagmus being the most common finding and partial ophthalmoparesis or slow saccades in 7 (41%). Twelve patients had brain MRI, with 11 (93%) showing cerebellar atrophy. Cervical spine MRI (n = 5) was normal in 4, with 1 patient having a likely incidental thoracic syrinx. Muscle biopsy identified cytochrome c oxidase–negative fibers in 2 patients, multiple mtDNA deletions in 1 patient, and coenzyme Q10 deficiency in a single patient.

We were surprised to find likely pathogenic SPG7 mutations in 18.6% of patients with unexplained ataxia. Although these patients did not have pure ataxia on follow-up, it was the predominating feature, and the patients had been clinically diagnosed and investigated for ataxic disorders. In our SPG7 patients, even after an average follow-up of 16.8 years, the pyramidal signs were subtle in many, endorsing our conclusion that a gene identified in patients with autosomal recessive hereditary spastic paraplegia should be considered in adults with unexplained ataxia. Eight (57%) of the probands had no relevant family history, so SPG7 should even be considered in sporadic cases. Combining these findings with another novel clinical presentation of SPG7,² we provide the first minimum prevalence of SPG7-related disease at 0.72/100,000, making this a common cause of inherited ataxia, comparable with both autosomal dominant spinocerebellar ataxia (1.59/100,000)³ and Friedrich ataxia (1.8/100,000).⁴ This is probably an underestimate given the late presentation of some cases. All of our patients had p.Ala510Val, reaffirming the pathogenicity of this allele, which has been considered a low frequency polymorphism (0.3463% in EVS, dbSNP rs61755320). This study illustrates the advantage of exome sequencing in neurogenetic disorders, where genes initially shown to cause one classical phenotype (such as hereditary spastic paraplegia) can also cause other phenotypes in a subgroup of patients (such as ataxia).

SPG7 encodes paraplegin, which is a component of the mitochondrial AAA protease, and the binding partner of AFG3L2.⁵ Both paraplegin and AFG3L2 are highly expressed in Purkinje neurons,⁶ and mutations in AFG3L2 cause spinocerebellar ataxia type 28.⁷ This explains why the phenotypic spectrum of SPG7 includes a predominantly ataxic presentation. It will be interesting to see whether specific mutations predispose to an ataxic or spastic presentation when cohorts increase in size. However, given the diverse nature of the clinical presentations, a more inclusive disease name such as parapleginopathy may avoid the misleading expectation that spasticity always predominates in this condition. Future study should address the relative likelihoods of these various presentations along the phenotypic spectrum of paraplegin-related diseases.

Four tables; Microsoft Word file.

Editorial, page 1070