Abstract

Spastic paraplegia type 7 is an autosomal recessive neurodegenerative disorder mainly characterized by progressive bilateral lower limb spasticity and referred to as a form of hereditary spastic paraplegia. Additional disease features may also be observed as part of a more complex phenotype. Many different mutations have already been identified, but no genotype–phenotype correlations have been found so far. From a total of almost 800 patients referred for testing, we identified 60 patients with mutations in the SPG7 gene. We identified 14 previously unreported mutations and detected a high recurrence rate of several earlier reported mutations. We were able to collect detailed clinical data for 49 patients, who were ranked based on a pure versus complex phenotype, ataxia versus no ataxia and missense versus null mutations. A generally complex phenotype occurred in 69% of all patients and was associated with a younger age at onset (trend with P = 0.07). Ataxia was observed in 57% of all patients. We found that null mutations were associated with the co-occurrence of cerebellar ataxia (trend with P = 0.06). The c.1409 G > A (p.Arg470Gln) mutation, which was found homozygously in two sibs, was associated with a specific complex phenotype that included predominant visual loss due to optical nerve atrophy. Neuropathology in one of these cases showed severe degeneration of the optic system, with less severe degeneration of the ascending tracts of the spinal cord and cerebellum. Other disease features encountered in this cohort included cervical dystonia, vertical gaze palsy, ptosis and severe intellectual disability. In this large Dutch cohort, we seem to have identified the first genotype–phenotype correlation in spastic paraplegia type 7 by observing an association between the cerebellar phenotype of spastic paraplegia type 7 and SPG7 null alleles. An overlapping phenotypic presentation with its biological counterpart AFG3L2, which when mutated causes spinocerebellar ataxia type 28, is apparent and possibly suggests that abnormal levels of the SPG7 protein impact the function of the mitochondrial ATPases associated with diverse cellular activities–protease complex (formed by SPG7 and AFG3L2) in the cerebellum. In addition, a missense mutation in exon 10 resulted in predominant optical nerve atrophy, which might suggest deleterious interactions of this SPG7 variant with its substrate OPA1, the mutated gene product in optic atrophy type 1. Functional studies are required to further investigate these interactions.

Introduction

The hereditary spastic paraplegias constitute a genetically and clinically heterogeneous group of neurodegenerative disorders. The main clinical feature is progressive lower limb spasticity due to pyramidal tract dysfunction, sometimes accompanied by urinary urgency and mild impairment of deep sensory modalities (Strümpell, 1904; Harding, 1983), which is referred to as a pure hereditary spastic paraplegia form. Hereditary spastic paraplegia is classified as complicated if additional neurological signs, such as intellectual disability, extrapyramidal signs, visual dysfunction, epilepsy, peripheral neuropathy or MRI abnormalities, are present (Harding, 1983).

Spastic paraplegia type 7 (SPG7) is one of the more common autosomal recessive hereditary spastic paraplegias and mutations in the SPG7 gene lead to both pure and complicated phenotypes (Brugman et al., 2008). With regard to the latter, cerebellar ataxia, optic nerve involvement and peripheral neuropathy have mostly been observed (Casari and Marconi, 2010). The SPG7 gene comprises 17 exons and encodes the protein paraplegin. Paraplegin is a mitchochondrial metalloprotease, belonging to the AAA family (ATPases associated with diverse cellular activities), and is located in the mitochondrial inner membrane (Casari et al., 1998) as a complex together with the highly homologous AFG3L2 protein (Atorino et al., 2003). This complex plays an important role in various mitochondrial processes including mitochondrial protein quality surveillance, protein dislocation, enzyme processing, ribosome assembly, protein synthesis and mitochondrial fusion (Koppen and Langer, 2007).

Based on the abovementioned complex phenotype of SPG7, there seems to be some overlap with genetically closely related dominant disorders such as spinocerebellar ataxia type 28 (caused by AGF3L2 mutations) and optic atrophy type 1 (Di Bella et al., 2010; Yu-Wai-Man et al., 2010). A few studies have also described specific clinical phenotypes (optic atrophy and widespread fibre damage on DTI–MRI) in patients with private, rare homozygous SPG7 mutations (Casari et al., 1998; Warnecke et al., 2010), suggesting that some specific mutations have a more widespread deleterious effect on a multitude of neurons and pathways. Currently, there are no known genotype–phenotype correlations in SPG7. Here, we studied the mutations and disease characteristics in a large Dutch cohort of SPG7 patients and investigated whether a genotype–phenotype correlation could be identified by ranking SPG7 mutations based on their predicted protein effect.

Materials and methods

Patients

Patients were recruited from the DNA diagnostic laboratory database of the Radboud University Nijmegen Medical Centre. Since this laboratory is the only laboratory providing SPG7 mutation analysis in The Netherlands, the database comprises all known Dutch SPG7 patients and families. Most patients were clinically reviewed by neurologists and/or clinical geneticists from the University Medical Centres of Nijmegen (S.d.B., B.v.d.W.), Utrecht (J.H.V., L.H.v.d.B.) and Groningen (C.C.V-B., H.K.). For those who were not, we obtained detailed clinical information from the referring neurologists. All clinical data were analysed without prior knowledge of the specific SPG7 genotype. Age at onset was defined as the age at which the walking difficulties were first noticed by the patient.

Genetic analysis

Mutation analysis of the SPG7 gene was performed by sequencing of the coding sequences including flanking intronic sequences, using the methods described previously (Brugman et al., 2008). Patients in whom a presumed homozygous mutation or only a single heterozygous mutation was identified were subsequently analysed by multiplex ligation-dependent probe amplification (MLPA) to detect deletions or duplications within the SPG7 gene (Brugman et al., 2008). NM_003119.2 was used as reference sequence, with nucleotide 1 corresponding to the A of the start codon. Test results were confirmed by an independent test according to standard procedures. Determination of pathogenicity of missense mutations was obtained by an in silico approach using the prediction programs SIFT (Sorting Intolerance from Tolerance; http://sift.jcvi.org), Align GVGD (http://agvgd.iarc.fr/) and POLYPHEN (Polymorphism Phenotyping; http://genetics.bwh.harvard.edu/pph/).

Histopathology

At autopsy of one of our patients (Patient 1, Table 1), biopsies from the femoral nerve, from several brain areas and from different levels of the spinal cord were taken. These freshly obtained samples were immersion fixed in 2% buffered glutaraldehyde and post-fixed using osmium tetroxide. After embedding in Epon® resin, 0.5 µm thick sections were cut and stained using toluidine blue (Dawson et al., 2003).

Table 1

Overview of individual SPG7 patients

Patient Sex Family Age at onset (years) Disease duration (years) Ataxia (yes/no) Phenotypic description Mutation 1 Mutation 2 Mutation category (1:other, 2:null) 
10 70 No Complex phenotype with optical nerve atrophy c.1409 G > A (p.Arg470Gln) c.1409 G > A (p.Arg470Gln) 
27 55 ND Complex phenotype with optical nerve atrophy c.1409 G > A (p.Arg470Gln) c.1409 G > A (p.Arg470Gln) 
43 25 No Complex phenotype with eye movement abnormalities c.1454_1462del (p.Arg485_Glu487del) c.1454_1462del (p.Arg485_Glu487del) 
47 No Complex phenotype with eye movement abnormalities c.1454_1462del (p.Arg485_Glu487del) c.1454_1462del (p.Arg485_Glu487del) 
ND ND ND Complex phenotype with severe intellectual disability c.1454_1462del (p.Arg485_Glu487del) c.1454_1462del (p.Arg485_Glu487del) 
37 13 No Pure c.1454_1462del (p.Arg485_Glu487del) c.1454_1462del (p.Arg485_Glu487del) 
36 11 Yes Complex phenotype with ataxia c.1454_1462del (p.Arg485_Glu487del) c.1672 A > T (p.Lys558X) 
28 19 Yes Complex phenotype with ataxia c.1454_1462del (p.Arg485_Glu487del) c.1672 A > T (p.Lys558X) 
34 25 Yes Complex phenotype with ataxia and optical nerve atrophy c.1454_1462del (p.Arg485_Glu487del) c.2115_2131del (p.Leu706fs) 
10 38 26 Yes Complex phenotype with ataxia c.1454_1462del (p.Arg485_Glu487del) c.2115_2131del (p.Leu706fs) 
11 53 No Pure c.1529 C > T (p.Ala510Val) c.1454_1462del (p.Arg485_Glu487del) 
12 36 13 No Pure c.1529 C > T (p.Ala510Val) c.1454_1462del (p.Arg485_Glu487del) 
13 47 Yes Complex phenotype with ataxia c.1529 C > T (p.Ala510Val) c.1529 C > T (p.Ala510Val) 
14 44 12 Yes Complex phenotype with ataxia c.1529 C > T (p.Ala510Val) c.1529 C > T (p.Ala510Val) 
15 52 27 Yes Complex phenotype with ataxia c.1529 C > T (p.Ala510Val) c.1937-2del (p.?) 
16 45 29 Yes Complex phenotype with ataxia c.1529 C > T (p.Ala510Val) c.1937-2del (p.?) 
17 21 10 No Pure c.1529 C > T (p.Ala510Val) c.861 + 2dup (p.?) 
18 24 Yes Complex phenotype with ataxia c.1529 C > T (p.Ala510Val) c.861 + 2dup (p.?) 
19 10 46 14 Yes Complex phenotype with ataxia c.861dup (p.Asn288X) c.2228 T > C (p.Ile743Thr) 
20 10 48 Yes Complex phenotype with ataxia, parkinsonisma and cervical dystoniaa c.861dup (p.Asn288X) c.2228 T > C (p.Ile743Thr) 
21 11 33 22 Yes Complex phenotype with ataxia c.861dup (p.Asn288X) c.861dup (p.Asn288X) 
22 11 33 20 Yes Complex phenotype with ataxia c.861dup (p.Asn288X) c.861dup (p.Asn288X) 
23 12 ND ND ND ND c.1529 C > T (p.Ala510Val) c.1045 G > A (p.Gly349Ser) 
24 12 ND ND ND ND c.1529 C > T (p.Ala510Val) c.1529 C > T (p.Ala510Val) 
25 13 ND ND ND ND c.1529 C > T (p.Ala510Val) c.1529 C > T (p.Ala510Val) 
26 13 ND ND ND ND c.1529 C > T (p.Ala510Val) c.1529 C > T (p.Ala510Val) 
27 13 ND ND ND ND c.1529 C > T (p.Ala510Val) c.1529 C > T (p.Ala510Val) 
28 – 42 14 Yes Complex phenotype with ataxia c.1045 G > A (p.Gly349Ser) c.2069 C > T (p.Pro690Leu) 
29 – 40 No Pure c.1045 G > A (p.Gly349Ser) c.2090 A > C (p.Gln697Pro) 
30 – 42 No Pure c.1045 G > A (p.Gly349Ser) c.2115_2131del (p.Leu706fs) 
31 – 60 No Pure c.1147 G > T (p.Gly383X) c.1822 C > T (p.Gln608X) 
32 – 34 26 Yes Complex phenotype with ataxia c.1454_1462del (p.Arg485_Glu487del) c.2228 T > C (p.Ile743Thr) 
33 – 47 Yes Complex phenotype with ataxia c.1454_1462del (p.Arg485_Glu487del) c.233 T > A (p.Leu78X) 
34 – 38 12 Yes Complex phenotype with ataxia c.1529 C > T (p.Ala510Val) c.1454_1462del (p.Arg485_Glu487del) 
35 – 22 11 Yes Complex phenotype with ataxia c.1529 C > T (p.Ala510Val) c.1454_1462del (p.Arg485_Glu487del) 
36 – 43 10 No Pure c.1529 C > T (p.Ala510Val) c.1454_1462del (p.Arg485_Glu487del) 
37 – 33 Yes Complex phenotype with ataxia c.1529 C > T (p.Ala510Val) c.1454_1462del (p.Arg485_Glu487del) 
38 – 57 No Pure c.1529 C > T (p.Ala510Val) c.1454_1462del (p.Arg485_Glu487del) 
39 – 47 13 Yes Complex phenotype with ataxia c.1529 C > T (p.Ala510Val) c.1529 C > T (p.Ala510Val) 
40 – 53 No Pure c.1529 C > T (p.Ala510Val) c.1529 C > T (p.Ala510Val) 
41 – 60 Yes Complex phenotype with ataxia c.1529 C > T (p.Ala510Val) c.1529 C > T (p.Ala510Val) 
42 – 58 No Pure c.1529 C > T (p.Ala510Val) c.1529 C > T (p.Ala510Val) 
43 – 44 No Pure c.1529 C > T (p.Ala510Val) c.1529 C > T (p.Ala510Val) 
44 – 37 15 No Pure c.1529 C > T (p.Ala510Val) c.2014 G > A (p.Gly672Arg) 
45 – 50 No Pure c.1529 C > T (p.Ala510Val) c.2115_2131del (p.Leu706fs) 
46 – 27 27 Yes Complex phenotype with ataxia c.1529 C > T (p.Ala510Val) c.2115_2131del (p.Leu706fs) 
47 – 20 26 Yes Complex phenotype with ataxia c.1529 C > T (p.Ala510Val) c.2115_2131del (p.Leu706fs) 
48 – 55 15 Yes Complex phenotype with ataxia c.1529 C > T (p.Ala510Val) c.759 -?_2181 + ?del (p.?) 
49 – 51 No Pure c.1529 C > T (p.Ala510Val) c.759 -?_2181 + ?del (p.?) 
50 – 63 ND ND c.1672 A > T (p.Lys558X) c.1672 A > T (p.Lys558X) 
51 – 16 22 No Complex phenotype with eye movement abnormalities c.1894 G > A (p.Gly632Arg) c.1984delinsTTC (p.Val662fs) 
52 – 20 Yes Complex phenotype with ataxia c.2115_2131del (p.Leu706fs) c.2115_2131del (p.Leu706fs) 
53 – 16 52 Yes Complex phenotype with ataxia and cervical dystonia c.2115_2131del (p.Leu706fs) c.2115_2131del (p.Leu706fs) 
54 – 10 40 Yes Complex phenotype with ataxia c.3 G > A (p.Met1?) c.3 G > A (p.Met1?) 
55 – 51 10 No Pure c.618 + 11_618 + 68del (p.?) c.2219 A > G (p.Tyr740Cys) 
56 – ND ND ND ND c.1045 G > A (p.Gly349Ser) c.1454_1462del (p.Arg485_Glu487del) 
57 – ND ND ND ND c.1529 C > T (p.Ala510Val) c.1045 G > A (p.Gly349Ser) 
58 – ND ND ND ND c.1529 C > T (p.Ala510Val) c.1045 G > A (p.Gly349Ser) 
59 – ND ND ND ND c.1529 C > T (p.Ala510Val) c.1053dup (p.Gly352fs) 
60 – ND ND ND ND c.1529 C > T (p.Ala510Val) c.1454_1462del (p.Arg485_Glu487del) 
Patient Sex Family Age at onset (years) Disease duration (years) Ataxia (yes/no) Phenotypic description Mutation 1 Mutation 2 Mutation category (1:other, 2:null) 
10 70 No Complex phenotype with optical nerve atrophy c.1409 G > A (p.Arg470Gln) c.1409 G > A (p.Arg470Gln) 
27 55 ND Complex phenotype with optical nerve atrophy c.1409 G > A (p.Arg470Gln) c.1409 G > A (p.Arg470Gln) 
43 25 No Complex phenotype with eye movement abnormalities c.1454_1462del (p.Arg485_Glu487del) c.1454_1462del (p.Arg485_Glu487del) 
47 No Complex phenotype with eye movement abnormalities c.1454_1462del (p.Arg485_Glu487del) c.1454_1462del (p.Arg485_Glu487del) 
ND ND ND Complex phenotype with severe intellectual disability c.1454_1462del (p.Arg485_Glu487del) c.1454_1462del (p.Arg485_Glu487del) 
37 13 No Pure c.1454_1462del (p.Arg485_Glu487del) c.1454_1462del (p.Arg485_Glu487del) 
36 11 Yes Complex phenotype with ataxia c.1454_1462del (p.Arg485_Glu487del) c.1672 A > T (p.Lys558X) 
28 19 Yes Complex phenotype with ataxia c.1454_1462del (p.Arg485_Glu487del) c.1672 A > T (p.Lys558X) 
34 25 Yes Complex phenotype with ataxia and optical nerve atrophy c.1454_1462del (p.Arg485_Glu487del) c.2115_2131del (p.Leu706fs) 
10 38 26 Yes Complex phenotype with ataxia c.1454_1462del (p.Arg485_Glu487del) c.2115_2131del (p.Leu706fs) 
11 53 No Pure c.1529 C > T (p.Ala510Val) c.1454_1462del (p.Arg485_Glu487del) 
12 36 13 No Pure c.1529 C > T (p.Ala510Val) c.1454_1462del (p.Arg485_Glu487del) 
13 47 Yes Complex phenotype with ataxia c.1529 C > T (p.Ala510Val) c.1529 C > T (p.Ala510Val) 
14 44 12 Yes Complex phenotype with ataxia c.1529 C > T (p.Ala510Val) c.1529 C > T (p.Ala510Val) 
15 52 27 Yes Complex phenotype with ataxia c.1529 C > T (p.Ala510Val) c.1937-2del (p.?) 
16 45 29 Yes Complex phenotype with ataxia c.1529 C > T (p.Ala510Val) c.1937-2del (p.?) 
17 21 10 No Pure c.1529 C > T (p.Ala510Val) c.861 + 2dup (p.?) 
18 24 Yes Complex phenotype with ataxia c.1529 C > T (p.Ala510Val) c.861 + 2dup (p.?) 
19 10 46 14 Yes Complex phenotype with ataxia c.861dup (p.Asn288X) c.2228 T > C (p.Ile743Thr) 
20 10 48 Yes Complex phenotype with ataxia, parkinsonisma and cervical dystoniaa c.861dup (p.Asn288X) c.2228 T > C (p.Ile743Thr) 
21 11 33 22 Yes Complex phenotype with ataxia c.861dup (p.Asn288X) c.861dup (p.Asn288X) 
22 11 33 20 Yes Complex phenotype with ataxia c.861dup (p.Asn288X) c.861dup (p.Asn288X) 
23 12 ND ND ND ND c.1529 C > T (p.Ala510Val) c.1045 G > A (p.Gly349Ser) 
24 12 ND ND ND ND c.1529 C > T (p.Ala510Val) c.1529 C > T (p.Ala510Val) 
25 13 ND ND ND ND c.1529 C > T (p.Ala510Val) c.1529 C > T (p.Ala510Val) 
26 13 ND ND ND ND c.1529 C > T (p.Ala510Val) c.1529 C > T (p.Ala510Val) 
27 13 ND ND ND ND c.1529 C > T (p.Ala510Val) c.1529 C > T (p.Ala510Val) 
28 – 42 14 Yes Complex phenotype with ataxia c.1045 G > A (p.Gly349Ser) c.2069 C > T (p.Pro690Leu) 
29 – 40 No Pure c.1045 G > A (p.Gly349Ser) c.2090 A > C (p.Gln697Pro) 
30 – 42 No Pure c.1045 G > A (p.Gly349Ser) c.2115_2131del (p.Leu706fs) 
31 – 60 No Pure c.1147 G > T (p.Gly383X) c.1822 C > T (p.Gln608X) 
32 – 34 26 Yes Complex phenotype with ataxia c.1454_1462del (p.Arg485_Glu487del) c.2228 T > C (p.Ile743Thr) 
33 – 47 Yes Complex phenotype with ataxia c.1454_1462del (p.Arg485_Glu487del) c.233 T > A (p.Leu78X) 
34 – 38 12 Yes Complex phenotype with ataxia c.1529 C > T (p.Ala510Val) c.1454_1462del (p.Arg485_Glu487del) 
35 – 22 11 Yes Complex phenotype with ataxia c.1529 C > T (p.Ala510Val) c.1454_1462del (p.Arg485_Glu487del) 
36 – 43 10 No Pure c.1529 C > T (p.Ala510Val) c.1454_1462del (p.Arg485_Glu487del) 
37 – 33 Yes Complex phenotype with ataxia c.1529 C > T (p.Ala510Val) c.1454_1462del (p.Arg485_Glu487del) 
38 – 57 No Pure c.1529 C > T (p.Ala510Val) c.1454_1462del (p.Arg485_Glu487del) 
39 – 47 13 Yes Complex phenotype with ataxia c.1529 C > T (p.Ala510Val) c.1529 C > T (p.Ala510Val) 
40 – 53 No Pure c.1529 C > T (p.Ala510Val) c.1529 C > T (p.Ala510Val) 
41 – 60 Yes Complex phenotype with ataxia c.1529 C > T (p.Ala510Val) c.1529 C > T (p.Ala510Val) 
42 – 58 No Pure c.1529 C > T (p.Ala510Val) c.1529 C > T (p.Ala510Val) 
43 – 44 No Pure c.1529 C > T (p.Ala510Val) c.1529 C > T (p.Ala510Val) 
44 – 37 15 No Pure c.1529 C > T (p.Ala510Val) c.2014 G > A (p.Gly672Arg) 
45 – 50 No Pure c.1529 C > T (p.Ala510Val) c.2115_2131del (p.Leu706fs) 
46 – 27 27 Yes Complex phenotype with ataxia c.1529 C > T (p.Ala510Val) c.2115_2131del (p.Leu706fs) 
47 – 20 26 Yes Complex phenotype with ataxia c.1529 C > T (p.Ala510Val) c.2115_2131del (p.Leu706fs) 
48 – 55 15 Yes Complex phenotype with ataxia c.1529 C > T (p.Ala510Val) c.759 -?_2181 + ?del (p.?) 
49 – 51 No Pure c.1529 C > T (p.Ala510Val) c.759 -?_2181 + ?del (p.?) 
50 – 63 ND ND c.1672 A > T (p.Lys558X) c.1672 A > T (p.Lys558X) 
51 – 16 22 No Complex phenotype with eye movement abnormalities c.1894 G > A (p.Gly632Arg) c.1984delinsTTC (p.Val662fs) 
52 – 20 Yes Complex phenotype with ataxia c.2115_2131del (p.Leu706fs) c.2115_2131del (p.Leu706fs) 
53 – 16 52 Yes Complex phenotype with ataxia and cervical dystonia c.2115_2131del (p.Leu706fs) c.2115_2131del (p.Leu706fs) 
54 – 10 40 Yes Complex phenotype with ataxia c.3 G > A (p.Met1?) c.3 G > A (p.Met1?) 
55 – 51 10 No Pure c.618 + 11_618 + 68del (p.?) c.2219 A > G (p.Tyr740Cys) 
56 – ND ND ND ND c.1045 G > A (p.Gly349Ser) c.1454_1462del (p.Arg485_Glu487del) 
57 – ND ND ND ND c.1529 C > T (p.Ala510Val) c.1045 G > A (p.Gly349Ser) 
58 – ND ND ND ND c.1529 C > T (p.Ala510Val) c.1045 G > A (p.Gly349Ser) 
59 – ND ND ND ND c.1529 C > T (p.Ala510Val) c.1053dup (p.Gly352fs) 
60 – ND ND ND ND c.1529 C > T (p.Ala510Val) c.1454_1462del (p.Arg485_Glu487del) 

For some patients, no clinical data were available (ND).

aConcurrent use of neuroleptics.

After 2 weeks of immersion fixation in 4% phosphate-buffered formalin, the brain and spinal cord were sectioned as follows: the cerebrum in the frontal plane, the cerebellum in the sagittal plane, and the brainstem and spinal cord in the transverse plane. Tissue blocks were sampled for paraffin embedding; 5 - to 20 -µm-thick paraffin sections were prepared. Routine histochemical staining was performed using haematoxylin and eosin, Luxol fast blue/Klüver-Barrera (for myelin) and Sevier-Munger (silver staining for axons) (Dawson et al., 2003). A list of primary antibodies is supplied in the online Supplementary material. For more detailed methods, we refer to other publications (Seidel et al., 2010, 2012; Olah et al., 2012).

A muscle biopsy was also taken from the quadriceps muscle (vastus lateralis on the right side). This biopsy was snap frozen according to routine procedures (Dubowitz and Sewry, 2007). On frozen sections, routine histochemical staining (haematoxylin and eosin, periodic acid–Schiff, periodic acid–Schiff after diastase, oil red O, Gomori) and enzyme histochemical staining (ATPases at different pH’s, NADH, SDH, COX, combined COX–SDH) were performed (Dubowitz and Sewry, 2007).

Statistical analysis

Statistical analysis was performed using SPSS (v16.0). Patients were ranked based on a pure versus complex phenotype, presence versus absence of cerebellar ataxia, and missense mutations (or small in-frame deletions) versus null mutations, the latter defined as at least one allele that probably does not lead to a protein product or to a severely truncated protein product. Ranked data were compared using Fisher’s exact test. The Mann–Whitney test was used to evaluate correlations between continuous data (age, disease duration) and ranked data. In all analyses, P < 0.05 was considered statistically significant.

Results

Mutation analysis

From a total of 791 patients who were referred for complete SPG7 mutation analysis, we identified homozygous or compound heterozygous mutations in 46 families (totalling 60 patients, Table 1). We identified 23 different (likely to be pathogenic) mutations (Table 2, Fig. 1). Nine were missense mutations (totalling 63 alleles), eight were nonsense mutations (totalling 25 alleles), two were splice-site mutations (totalling four alleles), one was a start codon mutation (homozygous), one was an in-frame deletion of three amino acids (totalling 23 alleles), one was a large multi-exon deletion (totalling two alleles) and one was an intronic deletion of unknown significance (one allele).

Figure 1

Graphical overview of mutations found in this and other studies (Human Gene Mutation database and Schlipf et al., 2011). Mutations with a presumed null-effect are graphed as squares; other mutations are graphed as triangles. New, previously unreported mutations are denoted in red. The asterisk denotes mutations that have been found homozygously. The arrow denotes the missense mutation that is associated with optical atrophy. The ‘?’ denotes the intronic deletion of unknown significance. AA = amino acid; HGMD = Human Gene Mutation database; MTS = mitochondrial targeting sequence; TM = transmembrane domain.

Figure 1

Graphical overview of mutations found in this and other studies (Human Gene Mutation database and Schlipf et al., 2011). Mutations with a presumed null-effect are graphed as squares; other mutations are graphed as triangles. New, previously unreported mutations are denoted in red. The asterisk denotes mutations that have been found homozygously. The arrow denotes the missense mutation that is associated with optical atrophy. The ‘?’ denotes the intronic deletion of unknown significance. AA = amino acid; HGMD = Human Gene Mutation database; MTS = mitochondrial targeting sequence; TM = transmembrane domain.

Table 2

Identified mutations

Patient Location Nucleotide change Amino acid change Number of patients Number of alleles Number of families Allele Fq controls Allele Fq NHLBI-ESP study (EA;AA) Predicted effect Mutation category % Patients with ataxia PhyloP Conservation score Pathogenicity 
Exon 1 c.3 G > A p.Met1? ND ND Deleterious Null 100 2.55 Likely pathogenic 
Exon 2 c.233 T > A p.Leu78X ND 0/10758 Nonsense Null 100 −0.04 Pathogenic 
Intron 4 c.618 + 11_618 + 68del p.? ND ND Splicing  NA Unknown 
Exon 6-16 c.759 -?_2181 + ?del p.? ND ND Deletion exon 6-16 Null 50 NA Pathogenic 
Exon 6 c.861dup p.Asn288X ND ND Nonsense Null 100 NA Pathogenic 
Intron 6 c.861 + 2dup p.? ND ND Splicing Null 50 NA Likely pathogenic 
Exon 8 c.1045 G > A p.Gly349Ser 2/728 15/7020;0/3738 Missense Other 33 6.26 Pathogenica 
Exon 8 c.1053dup p.Gly352fs 0/728 ND Frame shift Null  NA Pathogenic 
Exon 8 c.1147 G > T p.Gly383X 0/728 0/10742 Nonsense Null 6.26 Pathogenic 
10 Exon 10 c.1409 G > A p.Arg470Gln ND 0/10758 Missense Other 100 5.86 Likely pathogenic 
11 Exon 11 c.1454-1462del p.Arg485_Glu487del 19 23 14 0/696 ND In-frame deletion Other 56 NA Pathogenica 
12 Exon 11 c.1529 C > T p.Ala510Val 33 44 26 1/696 33/7020;6/3738 Missense Other 54 6.02 Pathogenica 
13 Exon 13 c.1672 A > T p.Lys558X ND 1/7020;0/3738 Nonsense Null 100 5.05 Pathogenic 
14 Exon 14 c.1822 C > T p.Gln608X ND 0/10758 Nonsense Null 4.56 Pathogenic 
15 Exon 14 c.1894 G > A p.Gly632Arg ND 0/10758 Missense Other 6.1 Likely pathogenic 
16 Intron 14 c.1937-2del p.? ND ND Splicing Null 100 4.24 Pathogenic 
17 Exon 15 c.1984delinsTTC p.Val662fs ND ND Frame shift Null NA Pathogenic 
18 Exon 15 c.2014 G > A p.Gly672Arg ND 0/7020;1/3738 Missense Other 5.69 Likely pathogenic 
19 Exon 15 c.2069 C > T p.Pro690Leu ND 0/10758 Missense Other 100 5.69 Likely pathogenic 
20 Exon 15 c.2090 A > C p.Gln697Pro ND 0/10758 Missense Other 4.73 Likely pathogenic 
21 Exon 16 c.2115-2131del p.Leu706fs 10 ND ND Frame shift Null 75 NA Pathogenic 
22 Exon 17 c.2219 A > G p.Tyr740Cys ND 0/7020;1/3738 Missense Other 2.55 Likely pathogenic 
23 Exon 17 c.2228 T > C p.Ile743Thr ND 1/7019;0/3738 Missense Other 100 3.43 Likely pathogenic 
Patient Location Nucleotide change Amino acid change Number of patients Number of alleles Number of families Allele Fq controls Allele Fq NHLBI-ESP study (EA;AA) Predicted effect Mutation category % Patients with ataxia PhyloP Conservation score Pathogenicity 
Exon 1 c.3 G > A p.Met1? ND ND Deleterious Null 100 2.55 Likely pathogenic 
Exon 2 c.233 T > A p.Leu78X ND 0/10758 Nonsense Null 100 −0.04 Pathogenic 
Intron 4 c.618 + 11_618 + 68del p.? ND ND Splicing  NA Unknown 
Exon 6-16 c.759 -?_2181 + ?del p.? ND ND Deletion exon 6-16 Null 50 NA Pathogenic 
Exon 6 c.861dup p.Asn288X ND ND Nonsense Null 100 NA Pathogenic 
Intron 6 c.861 + 2dup p.? ND ND Splicing Null 50 NA Likely pathogenic 
Exon 8 c.1045 G > A p.Gly349Ser 2/728 15/7020;0/3738 Missense Other 33 6.26 Pathogenica 
Exon 8 c.1053dup p.Gly352fs 0/728 ND Frame shift Null  NA Pathogenic 
Exon 8 c.1147 G > T p.Gly383X 0/728 0/10742 Nonsense Null 6.26 Pathogenic 
10 Exon 10 c.1409 G > A p.Arg470Gln ND 0/10758 Missense Other 100 5.86 Likely pathogenic 
11 Exon 11 c.1454-1462del p.Arg485_Glu487del 19 23 14 0/696 ND In-frame deletion Other 56 NA Pathogenica 
12 Exon 11 c.1529 C > T p.Ala510Val 33 44 26 1/696 33/7020;6/3738 Missense Other 54 6.02 Pathogenica 
13 Exon 13 c.1672 A > T p.Lys558X ND 1/7020;0/3738 Nonsense Null 100 5.05 Pathogenic 
14 Exon 14 c.1822 C > T p.Gln608X ND 0/10758 Nonsense Null 4.56 Pathogenic 
15 Exon 14 c.1894 G > A p.Gly632Arg ND 0/10758 Missense Other 6.1 Likely pathogenic 
16 Intron 14 c.1937-2del p.? ND ND Splicing Null 100 4.24 Pathogenic 
17 Exon 15 c.1984delinsTTC p.Val662fs ND ND Frame shift Null NA Pathogenic 
18 Exon 15 c.2014 G > A p.Gly672Arg ND 0/7020;1/3738 Missense Other 5.69 Likely pathogenic 
19 Exon 15 c.2069 C > T p.Pro690Leu ND 0/10758 Missense Other 100 5.69 Likely pathogenic 
20 Exon 15 c.2090 A > C p.Gln697Pro ND 0/10758 Missense Other 4.73 Likely pathogenic 
21 Exon 16 c.2115-2131del p.Leu706fs 10 ND ND Frame shift Null 75 NA Pathogenic 
22 Exon 17 c.2219 A > G p.Tyr740Cys ND 0/7020;1/3738 Missense Other 2.55 Likely pathogenic 
23 Exon 17 c.2228 T > C p.Ile743Thr ND 1/7019;0/3738 Missense Other 100 3.43 Likely pathogenic 

Mutations are listed based on nucleotide numbering. Allele frequencies (Fq) are listed as number of of alleles found/number of alleles tested.

AA = African descent; EA = European descent; NA = not applicable; ND = not determined.

aBased on this study.

Two frequent mutations in exon 11 (p.Arg485_Glu487del and p.Ala510Val) were responsible for more than half of the disease alleles (67/120) and were frequently detected in a homozygous state. Because of the many homozygous cases, the frequencies of these variants in our patient cohort strongly deviated from the Hardy–Weinberg equilibrium (p.Arg485_Glu487del: χ2 = 89, P < 0.000001; p.Ala510Val; χ2 = 187, P < 0.000001). These two mutations have been described previously (McDermott et al., 2001; Brugman et al., 2008), but since the carrier frequency in the general population, especially for the p.Ala510Val mutation, is suggested to be as high as 3% (McDermott et al., 2001), the pathogenicity of these variants remains unclear. The p.Gly349Ser mutation has also been described previously (Brugman et al., 2008). The carrier frequency for this mutation is unknown, but this mutation was also identified as a single heterozygous mutation in five patients (out of the 791 patients for whom SPG7 analysis had been requested; details not included in this study), which suggests a relatively high carrier frequency.

To investigate the carrier frequency of these three mutations in the Dutch population, we performed carrier analysis in an anonymous control cohort comprising ∼350 randomly selected healthy individuals, who were tested negative for a known familial mutation in any gene but SPG7. The p.Gly349Ser mutation was identified heterozygously in 2 out of 364 controls, the p.Ala510Val mutation was identified heterozygously in 1 out of 348 controls and the p.Arg485_Glu487del mutation was not detected in 348 controls (Table 2). Recent exome sequencing data of a large American cohort (of European and African descent) confirms the carrier frequency of these mutations [Exome Variant Server, NHLBI Exome Sequencing Project (ESP), Seattle, WA (http://evs.gs.washington.edu/EVS/) (01-2012); Table 2].

The SPG7 phenotype

The phenotype could be defined for 49 confirmed homozygous or compound heterozygous SPG7 mutation carriers (from 37 families), of whom the clinical characteristics are summarized in Table 3. Inheritance was most often compatible with an autosomal recessive mode. In one family, a seemingly autosomal dominant mode of inheritance was encountered (affected father and son; Family 12 in Table 1). This was explained by a segregating third pathogenic allele (pseudo-dominant inheritance). The age at onset varied from 10 to 63 years of age [mean 39 years, standard deviation (SD) 13.7], showing a unimodal distribution with a peak in the fifth decade. However, according to the available information, at least six patients experienced (minor) motor problems in the legs in their first decade, which were not considered to be clearly indicative for hereditary spastic paraplegia. In these patients progressive motor disturbances manifested in their second, third or fourth decade of life.

Table 3

Phenotypic description of patient cohort

Male:Female 32:28 
Mean age at onset (years) 39 (SD 13.7, range 10–63) 
Mean disease duration at last examination (years) 16 (SD 14.1, range 1–70) 
Family history positive (%) 30/46 (65) 
Handicap (%) 
    Mild 36/47 (77) 
        No walking aid 23/47 (49) 
        Walking aid 13/47 (28) 
    Severea 11/47 (23) 
Muscle weakness 25/37 (68) 
    Upper limb (%) 1/37 (3) 
    Lower limb (%) 24/37 (65) 
Muscle tone increased (%) 
    Upper limb 6/42 (14) 
    Lower limb 42/43 (98) 
        Mild 15/42 (36) 
        Severe 13/42 (31) 
        Unspecified 14/42 (33) 
Amyotrophy (%) 4/35 (11) 
Tendon reflexes (%) 
 Upper Limbs 
        Normal 11/34 (32) 
        Brisk 23/34 (68) 
 Lower limbs 
        Normal 3/47 (6) 
        Brisk 44/47 (94) 
    Babinski sign 43/47 (91) 
Sensory abnormalities (%) 15/37 (41) 
    Superficial 5/37 (14) 
    Deep 15/37 (41) 
Sphincter disturbances (%) 22/42 (52) 
Bladder 22/42 (52) 
Anal 6/41 (15) 
Cerebellar ataxiab (%) 27/47 (57) 
Neuroimaging, MRI and CT (%) 
    Normal 17/36 (47) 
    Cerebellar atrophy 14/36 (39) 
    Spinal cord atrophy 2/36 (6) 
    Unspecific white matter lesions 2/36 (6) 
    Previous lacunar infarction 2/36 (6) 
Electromyography (%) 
    Normal 9/14 (64) 
    Motor neuropathy 5/14 (36) 
    Sensory polyneuropathy 0/14 (0) 
Abnormalities of cranial nerves/brainstem (%) 
    Eye movement abnormalities 15/46 (33) 
        Dysmetric saccades 11/46 (24) 
        Jerky pursuit 5/46 (11) 
        Limited range of eye movements 3/46 (7) 
        Diplopia 4/46 (9) 
    Nystagmus 6/45 (13) 
    Dysartria 16/43 (37) 
        Cerebellar 6/43 (14) 
        Spastic 1/43 (2) 
        Mixed 2/43 (5) 
        Unspecified 7/43 (16) 
    Ptosis 3/41 (7) 
    Optical nerve atrophy 3/41 (7) 
Extrapyramidal symptoms 
    Parkinsonism 1 c/47 (2) 
    Cervical dystonia 2/47 (4) 
Cognitive impairment (%) 
    Severe intellectual disability 1/45 (2) 
Male:Female 32:28 
Mean age at onset (years) 39 (SD 13.7, range 10–63) 
Mean disease duration at last examination (years) 16 (SD 14.1, range 1–70) 
Family history positive (%) 30/46 (65) 
Handicap (%) 
    Mild 36/47 (77) 
        No walking aid 23/47 (49) 
        Walking aid 13/47 (28) 
    Severea 11/47 (23) 
Muscle weakness 25/37 (68) 
    Upper limb (%) 1/37 (3) 
    Lower limb (%) 24/37 (65) 
Muscle tone increased (%) 
    Upper limb 6/42 (14) 
    Lower limb 42/43 (98) 
        Mild 15/42 (36) 
        Severe 13/42 (31) 
        Unspecified 14/42 (33) 
Amyotrophy (%) 4/35 (11) 
Tendon reflexes (%) 
 Upper Limbs 
        Normal 11/34 (32) 
        Brisk 23/34 (68) 
 Lower limbs 
        Normal 3/47 (6) 
        Brisk 44/47 (94) 
    Babinski sign 43/47 (91) 
Sensory abnormalities (%) 15/37 (41) 
    Superficial 5/37 (14) 
    Deep 15/37 (41) 
Sphincter disturbances (%) 22/42 (52) 
Bladder 22/42 (52) 
Anal 6/41 (15) 
Cerebellar ataxiab (%) 27/47 (57) 
Neuroimaging, MRI and CT (%) 
    Normal 17/36 (47) 
    Cerebellar atrophy 14/36 (39) 
    Spinal cord atrophy 2/36 (6) 
    Unspecific white matter lesions 2/36 (6) 
    Previous lacunar infarction 2/36 (6) 
Electromyography (%) 
    Normal 9/14 (64) 
    Motor neuropathy 5/14 (36) 
    Sensory polyneuropathy 0/14 (0) 
Abnormalities of cranial nerves/brainstem (%) 
    Eye movement abnormalities 15/46 (33) 
        Dysmetric saccades 11/46 (24) 
        Jerky pursuit 5/46 (11) 
        Limited range of eye movements 3/46 (7) 
        Diplopia 4/46 (9) 
    Nystagmus 6/45 (13) 
    Dysartria 16/43 (37) 
        Cerebellar 6/43 (14) 
        Spastic 1/43 (2) 
        Mixed 2/43 (5) 
        Unspecified 7/43 (16) 
    Ptosis 3/41 (7) 
    Optical nerve atrophy 3/41 (7) 
Extrapyramidal symptoms 
    Parkinsonism 1 c/47 (2) 
    Cervical dystonia 2/47 (4) 
Cognitive impairment (%) 
    Severe intellectual disability 1/45 (2) 

aDefined as using a wheelchair (most commonly used device).

bDefined as gait ataxia and/or upper limb ataxia and/or lower limb ataxia.

cConcurrent use of neuroleptics.

The presenting symptoms were gait difficulties in all but two cases; these two presented with optical nerve atrophy (see below). After a mean disease duration of 16 years (range 1–70), half of the patients used some kind of walking aid. Twenty percent of patients were severely handicapped, defined as being predominantly wheelchair bound.

Almost all patients were reported to have an increased muscle tone of the lower limbs. Upper limb muscle tone was normal in >80% of the cases. Brisk reflexes of the lower limbs and Babinski sign were present in 94% and 91%, respectively. In over two-thirds of our study population, mild impairment of muscle strength was present in the lower limbs, and more specifically of the hip flexors and ankle dorsiflexors. Only one patient had severe leg muscle weakness (with a score on the Medical Research Council scale of 2 out of 5). Muscle strength in the arms was mostly normal. Deep sensory modalities, predominantly vibration sense, were disturbed in >40% of cases. In five patients, there was an additional impairment of superficial sensory abnormalities, which in two of them were present in a stock-and-glove distribution, suggesting a peripheral neuropathy. Bladder and anal sphincter disturbances, mainly urinary and faecal urgency, were reported in half of the patients. Foot deformities, most frequently pes cavus, were observed in 11 out of 44 patients.

Two-thirds of the patients manifested a complex phenotype based on clinical and/or imaging features, although the complex features were often relatively mild. Patients with a complex phenotype tended to be younger at onset (36 versus 44 years, P = 0.07). Disease duration, however, was longer in patients with a complex phenotype (20 versus 8 years, P = 0.005). This could be attributable to diagnostic delay in complex cases, or to the possibility that additional disease features arise later in the course of the disease. Complexity was mostly based on concomitant cerebellar ataxia, which was present in 57% of all cases and should thus be regarded as a second core clinical disease feature in SPG7. In some, cerebellar ataxia was actually the presenting clinical syndrome. Dysarthria, most frequently of the cerebellar type, was reported in 37% of cases. Nystagmus was present in 13% of cases and other eye movement abnormalities were observed in one-third of cases. Regarding the latter, vertical gaze palsy was seen in two patients. Other rare disease features of note were ptosis (n = 3), cervical dystonia (n = 2) and severe intellectual disability (n = 1).

Imaging of the brain was performed in 36 patients and was found to be normal in about half of them. Cerebellar atrophy, of various degrees, was evident in one-third of all cases and was as such the most frequently reported abnormality. The presence of cerebellar atrophy on MRI and clinical signs of cerebellar ataxia were, however, not always correlated. Three patients, suffering from mild to moderate cerebellar ataxia, showed no cerebellar atrophy on MRI. In contrast, in two other patients there was marked cerebellar atrophy, while they only suffered from mild ataxia. Two patients showed spinal cord atrophy. In four patients, aged 60–80 years, non-specific white matter lesions were found. Two patients, aged 62 and 70 years, showed signs of previous lacunar infarction. Electromyography was performed in 14 cases and showed evidence of a mainly motor neuropathy in five.

Genotype–phenotype correlations

To explore possible genotype–phenotype correlations, mutations were first ranked according to their predicted effect. Since SPG7 forms a protein complex with AFG3L2, we hypothesized that mutations that lead to a defective protein product might have a different phenotypic effect in comparison with mutations that lead to the absence of a protein product. Mutations that probably still result in a protein product (although defective) were classified as Class 1 (mostly missense mutations), while mutations that probably do not lead to a protein product or to a severely truncated protein product were classified as Class 2 (null; Table 2). Then, patients were ranked based on mutations. Patients with only Class 1 mutations were ranked in Group 1, while patients with at least one Class 2 mutation were ranked in Group 2.

We found that null mutations were more frequently (16 out of 22 patients, 73%) associated with the co-occurrence of cerebellar ataxia compared with Class 1 mutations (11 out of 24 patients, 46%) (P = 0.06). In patients with null mutations on both alleles, cerebellar ataxia was present in five out of six patients (83%), which supports the observed correlation between null mutations and cerebellar ataxia. Also, when evaluating specific mutations for the co-occurrence with ataxia, patients with null mutations more often had ataxia compared with those with missense mutations (Table 2).

In five cases (Patients 1, 2, 5, 20 and 53), a more severe and previously unreported complex phenotype was seen (Table 1), encompassing early and severe visual loss (optical atrophy; Patients 1 and 2), cervical dystonia (Patients 20 and 53) and severe intellectual disability (Patient 5). In Patients 5, 20 and 53, these additional features were not genotype-specific, since either siblings or unrelated patients with identical mutations suffered from the more pure forms of SPG7 (Table 2). In families, the clinical phenotype was mostly concordant (except Families 3 and 9). However, when patients with identical mutations from different families were compared, this concordance was less (Table 1). This suggests that other factors (genetic and/or environmental) might also contribute to the phenotypic expression of SPG7 mutations.

Patients 1 and 2 (sibs) both presented initially with loss of vision years before gait difficulties evolved. Loss of vision was caused by optical nerve atrophy, starting at age 7 and 17 years, respectively. During life, they both became fully blind. A private, homozygous missense mutation in exon 10 of the SPG7 gene (p.Arg470Gln) was detected in these patients. Since these patients were siblings, mutations in other genes responsible for this visual predominant phenotype cannot be fully excluded. Optic nerve atrophy was also observed in Patient 9, who carries the compound heterozygous mutations p.Arg485_Glu487del and p.Leu706fs.

Pathological evaluation and histochemistry

Autopsy was performed on Patient 1 (Table 1). The brain weight was 1214 g. The cerebellum and brainstem weighed 154 g, which is normal for the total brain weight (12.7%). Externally and on palpation, no abnormalities were observed on the cerebrum, cerebellum, brainstem or spinal cord. The ventricular system and basal ganglia were symmetrical without macroscopic changes.

Muscle biopsy showed a clear increase in the variation of the muscle fibre diameter with several atrophic fibres (Fig. 2A and B). Diffusely spread in this biopsy, clumps of pyknotic nuclei were observed, which are indicative of chronic atrophy. Fibre typing (among others with ATPase enzyme-histochemical staining) showed that there was preferential type 2 muscle fibre atrophy, which is often observed in the case of disuse (Fig. 2A).

Figure 2

Histochemistry. (A and B) The muscle biopsy from the vastus lateralis muscle (A = ATPase pH 4.3, B = COX–SDH). There is a preference of atrophy of type 2 fibres (pink in A). Furthermore, COX-negative fibres (arrows in B) can be observed. (C) A semi-thin section (osmium-tetroxide) of the femoral nerve with some swollen (arrow) and degenerating (arrowhead) nerve fibres. (D) A sensory ganglion (haematoxylin and eosin staining) without obvious pathology. (E) A semi-thin section (osmium tetroxide + toluidine blue) from the cervical spinal cord with swollen (arrow) and degenerating (arrowheads) axons in the spinocerebellar tracts. The same type of pathology could be observed in the medial portion of the dorsal tracts. (F) Detail from the cerebellar cortex (Sevier-Munger silver stain) with an axonal torpedo (arrow). Scale bars: A, B and D = 100 µm; C, E and F = 50 µm.

Figure 2

Histochemistry. (A and B) The muscle biopsy from the vastus lateralis muscle (A = ATPase pH 4.3, B = COX–SDH). There is a preference of atrophy of type 2 fibres (pink in A). Furthermore, COX-negative fibres (arrows in B) can be observed. (C) A semi-thin section (osmium-tetroxide) of the femoral nerve with some swollen (arrow) and degenerating (arrowhead) nerve fibres. (D) A sensory ganglion (haematoxylin and eosin staining) without obvious pathology. (E) A semi-thin section (osmium tetroxide + toluidine blue) from the cervical spinal cord with swollen (arrow) and degenerating (arrowheads) axons in the spinocerebellar tracts. The same type of pathology could be observed in the medial portion of the dorsal tracts. (F) Detail from the cerebellar cortex (Sevier-Munger silver stain) with an axonal torpedo (arrow). Scale bars: A, B and D = 100 µm; C, E and F = 50 µm.

At different levels of the spinal cord, the anterior and dorsal nerve roots had a normal appearance. In the dorsal ganglia, the cellularity was normal. Also the number of motor neurons in the anterior horn was normal. Using thick sections with a Kluver-Barrera staining, it was observed that there was more loss of myelin in the fasciculus gracilis than in the fasciculus cuneatus and that this was most obvious in the cervical spinal cord. Epon resin slides showed swollen axons in these tracts. Other ascending tracts such as the spinocerebellar tracts showed some spongiosis and some swollen axons as well (Fig. 2E). On the contrary, clear loss of fibres from the corticospinal tract was not seen, but in the semi-thin Epon® resin slides some swollen axons could be observed. Clear astrogliosis (glial fibrillary acidic protein staining) or microgliosis (CD68, IBA-1 stainings) was not observed in the spinal cord.

In the anterior vermis, there was limited loss of Purkinje cells. This was slightly more severe in the posterior vermis and the cerebellar hemispheres, where empty baskets as well as axonal torpedoes were observed (Fig. 2F).

In this patient with a very specific phenotype that consisted of concomitant optical nerve atrophy, the optic nerves, chiasm and optic tracts showed severe atrophy with an almost complete loss of axons and myelin sheaths in combination with severe astrogliosis (Fig. 3). Microgliosis was not observed.

Figure 3

Comparison between the optic nerve of the SPG7 patient and an age-matched control. Note the severe atrophy (A versus B), hypomyelination (C versus D, Luxol fast blue staining) and severe loss of nerve fibres (E versus F, neurofilament staining). Scale bars: A and B = 500 µm; CF = 50 µm.

Figure 3

Comparison between the optic nerve of the SPG7 patient and an age-matched control. Note the severe atrophy (A versus B), hypomyelination (C versus D, Luxol fast blue staining) and severe loss of nerve fibres (E versus F, neurofilament staining). Scale bars: A and B = 500 µm; CF = 50 µm.

In select areas in the cerebrum, cerebellum and brainstem, no protein aggregations were found after staining for hyperphosphorylated tau, α-synuclein, β-amyloid, ubiquitin and p62. Perforating arteries in the basal ganglia showed some mineralization, but apart from that no obvious vascular pathology was observed. Mineralization of blood vessels was also observed in the hippocampal complex. The primary motor cortex did not show an obvious loss of Betz neurons or signs of micro- or astrogliosis.

Discussion

We present a comprehensive overview of the phenotypic and genotypic spectrum of all identified SPG7 patients in The Netherlands, including the first histopathological examination. We screened the SPG7 gene in a large mixed cohort of 791 patients with (suspected) hereditary spastic paraplegia, mostly of Dutch origin. We detected homozygous or compound heterozygous mutations in 60 patients (46 families), and this therefore is the largest reported collection of SPG7 patients to date. We detected 23 different mutations of which 14 were novel (Table 2).

The mean age at onset was 39 years (range 10–63 years), which is similar to previous studies (Brugman et al., 2008; Casari and Marconi, 2010) and underscores the possibility of a very late onset of this disease. The overall clinical phenotype was that of a slowly progressive, spastic paraparesis, with a more complex phenotype in two-thirds of the patients although the complexity was often mild. Interestingly, cerebellar ataxia was present in 57% of the SPG7 patients, and in some of them the cerebellar syndrome was the presenting or more dominating disease feature. This percentage might even be an underestimation since such cases might not all be referred for SPG7 testing. Cerebellar ataxia is thus a very common element of the SPG7 phenotype, compatible with previous data (Elleuch et al., 2006), and SPG7 should therefore be included in the differential diagnosis of spastic ataxia. Interestingly, in our exploration of genotype–phenotype correlations in this large SPG7 cohort, we seem to have identified such a correlation as the cerebellar phenotype of SPG7 was associated with SPG7 null alleles (P = 0.06).

There were some patients with a clinical suspicion of peripheral neuropathy, but this was not verified by electromyography. Worth mentioning are ptosis, vertical gaze palsy, blindness due to optic nerve pathology, cervical dystonia and severe intellectual disability as other complex disease features, but these were all observed in only a few patients.

The genotypic spectrum comprises a few frequent recurrent mutations and many private familial mutations. Four frequent mutations were identified (mutations 7, 11, 12 and 21; Table 2), and the most frequent one, the c.1529 C > T (p.Ala510Val) mutation in exon 11, was identified in 26 separate families, comprising a total of 44 alleles in 33 patients (Table 2). The relatively high frequency of homozygous c.1529C > T (p.Ala510Val) mutations is expected given the mutation allele distribution, i.e. 37% of disease alleles in this study are caused by the c.1529 C > T (p.Ala510Val) mutation. The second most frequent mutation, also in exon 11, c.1454_1462del (p.Arg485_Glu487del), was identified in 14 families, comprising a total of 23 alleles in 19 patients. Both mutations have been described previously (McDermott et al., 2001; Brugman et al., 2008) and are of uncertain pathogenicity. Functional studies of the p.Ala510Val mutation suggest a pathogenic effect (Bonn et al., 2010). Based on the low control versus high patient frequencies, the co-occurrence with other proven pathogenic mutations and the deviations from the Hardy–Weinberg equilibrium in our patient population, we do consider both these mutations to be pathogenic. Mutation c.1045 G > A (p.Gly349Ser), which was identified in seven separate families comprising a total of seven alleles in seven patients, was also identified in our control cohort with a carrier frequency similar to that of the p.Ala510Val mutation (0.5% versus 0.3–1%). Therefore, the relative rarity of this mutation in the patient cohort compared with the p.Ala510Val mutation is unexpected. This mutation co-occurred with other mutations in patients (Table 1) and is therefore also most likely to be pathogenic. Moreover, this mutation was detected in a family with a dominant inheritance pattern (affected father and son) as the third pathogenic allele (pseudo-dominant inheritance; Family 12 in Table 1). We, however, have not detected this mutation in more than one patient per family and no homozygous patients were identified. This mutation could represent a hypomorphic allele, with a low context-specific penetrance (e.g. co-occurrence with a specific second mutation). This hypothesis is supported by a functional study on the effect of this mutation that suggested a context-dependent pathogenic effect, which depended on additional defects in AFG3L2 (Bonn et al., 2010).

We identified several novel missense mutations (Table 2). They are suggested to be pathogenic by in silico predictions, co-occurrence of these mutations with known pathogenic mutations (Table 1; the probability of this to occur by chance alone is very low) and/or co-segregation in the family. Additionally, these mutations were not found in >5000 controls [Exome Variant Server, NHLBI ESP]. We, however, did not conduct functional studies and have therefore not formally proven the pathogenicity of these mutations.

In ∼1.5% of patients (from the cohort of 791), a single heterozygous pathogenic or likely pathogenic mutation was identified (patients not included in this study). The great majority of mutations identified in this group are mutations of which the allele frequency in the general population is relatively high (p.Gly349Ser, p.Arg485_Glu487del and p.Ala510Val). Haplotype analysis did not reveal a frequent second allele (which could have pointed to a frequently missed mutation) within this group. Cumulative carriership of the p.Ala510Val and p.Gly349Ser mutations alone is 1.4% according to the NHLBI ESP study (Table 2). Thus, the frequency of carriership found in our cohort is in line with the carriership in the general population (1.5% versus 1.4%) and it is therefore unlikely that we have missed a second mutation in these cases, although it remains a possibility.

There seems to be an overlapping phenotypic presentation of SPG7 and its biological counterpart AFG3L2, which when heterozygously mutated causes spinal cerebellar ataxia type 28 (Di Bella et al., 2010). Genetic interaction has been shown in SPG7-deficient mice that were also heterozygously deficient for AFG3L3. These mice displayed acceleration and worsening of the axonopathy compared to SPG7-deficient mice (Martinelli et al., 2009). Our results suggest that abnormal levels of the SPG7 protein can also impact the function of the mitochondrial AAA–protease complex in the cerebellum. In addition, two patients carrying a homozygous missense mutation in exon 10, c.1409G > A (p.Arg470Gln), were suffering from concomitant severe optical nerve pathology, which has been described in SPG7 previously (Casari et al., 1998; McDermott et al., 2001) This missense mutation in exon 10 is the first one identified in exon 10 and is located in the ATPase AAA core domain (Fig. 1), which is highly conserved up to Caenorhabditis elegans. Possibly deleterious interactions of this SPG7 gene variant occur with its substrate OPA1, the mutated gene product in autosomal dominant optic atrophy type 1 (Alexander et al., 2000). As we have not addressed these interactions directly but only made inferences based on the clinical and genetic observations, functional studies are required to investigate these interactions further.

Neuropathological examinations of SPG7 cases are very rare and concern only muscle and/or nerve biopsies (Casari et al., 1998; Arnoldi et al., 2008). We were able to include a detailed histoneuropathological evaluation of one of the SPG7 patients with a deviant phenotype of severe visual loss who carried the p.Arg470Gln homozygous mutation. In this patient (Patient 1, Table 1), some COX-negative muscle fibres were found, as described by others (McDermott et al., 2001) (Fig. 2); however, the patient described here was 70 years of age and in normal ageing some COX-negative muscle fibres may be encountered. In the muscle biopsy of our patient, atrophic fibres were found but without clear signs of type-grouping or large group atrophy, which would point in the direction of an ongoing sequence of denervation, re-innervation and again denervation, as was seen in a paraplegin-deficient mouse model (Ferreirinha et al., 2004). Furthermore, we found more selective type 2 fibre atrophy, which could be a consequence of disuse. In the femoral nerve in our patient, some swollen and degenerating fibres were found. This peripheral neuropathy was less severe than documented in the mouse model (Ferreirinha et al., 2004). On the contrary, we found a severely affected optic system and affected fibres in the ascending spinal tracts (Figs 2 and 3). In the paraplegin-deficient mice, defects of the long tracks in the spinal cord preceded defects of the optic system, a difference that is difficult to explain without further studies in humans and mice. Additionally, we found some Purkinje cell axonal swellings and cell loss, without the presence of ataxia clinically. At the cervical level of the ascending tracts swollen axons could be observed. To a lesser extent some swollen axons could be observed at lumbosacral levels in the corticospinal tracts, although there was no obvious degeneration of the corticospinal tracts and no obvious loss of Betz neurons in the primary motor cortex. It seems that in the case of this patient, the distal parts of axons in both ascending and descending tracts were more affected than the proximal parts/cell bodies.

Based on our results, there are some practical recommendations. Because cerebellar ataxia is a very common element of the SPG7 phenotype, and can be the first presenting phenotype, SPG7 should be included in the differential diagnosis of spastic ataxia or ataxia with pyramidal features, even when the onset age is as high as 60 years. These differential diagnosis criteria are included in a review about spastic ataxia (de Bot et al., in press). Since the carrier frequency in the general population is relatively high for some of the SPG7 mutations, pseudo-dominant inheritance is a real possibility (as illustrated by our Family 12). Therefore, SPG7 should not be discarded in case such a hereditary spastic paraplegia pedigree is encountered. Along these lines, one might also consider offering carrier analysis in partners of patients for a better risk prediction for future offspring.

Funding

J.H.V. is funded by the Thierry Latran Foundation. A part of the research leading to these results has received funding from the European Community’s Health Seventh Framework Programme (FP7/2007-2013) (grant agreement no 259867).

Supplementary material

Supplementary material is available at Brain online.

Acknowledgements

We are grateful to the patients and their clinicians for their contributions.

Abbreviations

    Abbreviations
  • AAA

    ATPases associated with diverse cellular activities

  • SPG7

    spastic paraplegia type 7

References

Alexander
C
Votruba
M
Pesch
UE
Thiselton
DL
Mayer
S
Moore
A
, et al.  . 
OPA1, encoding a dynamin-related GTPase, is mutated in autosomal dominant optic atrophy linked to chromosome 3q28
Nat Genet
 , 
2000
, vol. 
26
 (pg. 
211
-
5
)
Arnoldi
A
Tonelli
A
Crippa
F
Villani
G
Pacelli
C
Sironi
M
, et al.  . 
A clinical, genetic, and biochemical characterization of SPG7 mutations in a large cohort of patients with hereditary spastic paraplegia
Hum Mutat
 , 
2008
, vol. 
29
 (pg. 
522
-
31
)
Atorino
L
Silvestri
L
Koppen
M
Cassina
L
Ballabio
A
Marconi
R
, et al.  . 
Loss of m-AAA protease in mitochondria causes complex I deficiency and increased sensitivity to oxidative stress in hereditary spastic paraplegia
J Cell Biol
 , 
2003
, vol. 
163
 (pg. 
777
-
87
)
Bonn
F
Pantakani
K
Shoukier
M
Langer
T
Mannan
AU
Functional evaluation of paraplegin mutations by a yeast complementation assay
Hum Mutat
 , 
2010
, vol. 
31
 (pg. 
617
-
21
)
Brugman
F
Scheffer
H
Wokke
JH
Nillesen
WM
de Visser
M
Aronica
E
, et al.  . 
Paraplegin mutations in sporadic adult-onset upper motor neuron syndromes
Neurology
 , 
2008
, vol. 
71
 (pg. 
1500
-
5
)
Casari
G
De Fusco
M
Ciarmatori
S
Zeviani
M
Mora
M
Fernandez
P
, et al.  . 
Spastic paraplegia and OXPHOS impairment caused by mutations in paraplegin, a nuclear-encoded mitochondrial metalloprotease
Cell
 , 
1998
, vol. 
93
 (pg. 
973
-
83
)
Casari
G
Marconi
R
Pagon
RA
Bird
TD
Dolan
CR
Stephens
K
Spastic Paraplegia 7
GeneReviews
  
1993–2006, http://www.ncbi.nlm.nih.gov/books/NBK1107/ August 24 [updated December 23, 2010]
Dawson
TP
Neal
JW
Llewellyn
L
Thomas
C
Dawson
TP
Neal
JW
Llewellyn
L
Thomas
C
Chapters 6, 8 and 9
Neuropathology techniques
 , 
2003
London
Hodder Arnold Publishers
de Bot
ST
Willemsen
MAAP
Vermeer
S
Kremer
HPH
van de Warrenburg
BP
Reviewing the genetic causes of spastic ataxia
 
Neurology
Di Bella
D
Lazzaro
F
Brusco
A
Plumari
M
Battaglia
G
Pastore
A
, et al.  . 
Mutations in the mitochondrial protease gene AFG3L2 cause dominant hereditary ataxia SCA28
Nat Genet
 , 
2010
, vol. 
42
 (pg. 
313
-
21
)
Dubowitz
V
Sewry
CA
Chapters 1 and 2
Muscle Biopsy; a practical approach
 , 
2007
3rd edn
Oxford
Saunders Elsevier
Elleuch
N
Depienne
C
Benomar
A
Hernandez
AM
Ferrer
X
Fontaine
B
, et al.  . 
Mutation analysis of the paraplegin gene (SPG7) in patients with hereditary spastic paraplegia
Neurology
 , 
2006
, vol. 
66
 (pg. 
654
-
9
)
Ferreirinha
F
Quattrini
A
Pirozzi
M
Valsecchi
V
Dina
G
Broccoli
V
, et al.  . 
Axonal degeneration in paraplegin-deficient mice is associated with abnormal mitochondria and impairment of axonal transport
J Clin Invest
 , 
2004
, vol. 
113
 (pg. 
231
-
42
)
Harding
AE
Classification of the hereditary ataxias and paraplegias
Lancet
 , 
1983
, vol. 
1
 (pg. 
1151
-
5
)
Koppen
M
Langer
T
Protein degradation within mitochondria: versatile activities of AAA proteases and other peptidases
Crit Rev Biochem Mol Biol
 , 
2007
, vol. 
42
 (pg. 
221
-
42
)
Martinelli
P
La Mattina
V
Bernacchia
A
Magnoni
R
Cerri
F
Cox
G
, et al.  . 
Genetic interaction between the m-AAA protease isoenzymes reveals novel roles in cerebellar degeneration
Hum Mol Genet
 , 
2009
, vol. 
18
 (pg. 
2001
-
13
)
McDermott
CJ
Dayaratne
RK
Tomkins
J
Lusher
ME
Lindsey
JC
Johnson
MA
, et al.  . 
Paraplegin gene analysis in hereditary spastic paraparesis (HSP) pedigrees in northeast England
Neurology
 , 
2001
, vol. 
56
 (pg. 
467
-
71
)
Olah
M
Raj
D
Brouwer
N
De Haas
AH
Eggen
BJ
Den Dunnen
WF
, et al.  . 
An optimized protocol for the acute isolation of human microglia from autopsy brain samples
Glia
 , 
2012
, vol. 
60
 (pg. 
96
-
111
)
Schlipf
NA
Schüle
R
Klimpe
S
Karle
KN
Synofzik
M
Schicks
J
, et al.  . 
Amplicon based high-throughput pooled sequencing identifies mutations in CYP7B1 and SPG7 in sporadic spastic paraplegia patients
Clin Genet
 , 
2011
, vol. 
80
 (pg. 
148
-
60
)
Seidel
K
den Dunnen
WF
Schultz
C
Paulson
H
Frank
S
de Vos
RA
, et al.  . 
Axonal inclusions in spinocerebellar ataxia type 3
Acta Neuropathol
 , 
2010
, vol. 
120
 (pg. 
449
-
60
)
Seidel
K
Vinet
J
den Dunnen
WF
Brunt
ER
Meister
M
Boncoraglio
A
, et al.  . 
The HSPB8-BAG3 chaperone complex is upregulated in astrocytes in the human brain affected by protein aggregation diseases
Neuropathol Appl Neurobiol
 , 
2012
, vol. 
38
 (pg. 
39
-
53
)
Strümpell
A
Die primäre Seitenstrangsklerose (spastische Spinalparalyse)
J Neurol
 , 
1904
, vol. 
27
 (pg. 
291
-
339
)
Warnecke
T
Duning
T
Schirmacher
A
Mohammadi
S
Schwindt
W
Lohmann
H
, et al.  . 
A novel splice site mutation in the SPG7 gene causing widespread fiber damage in homozygous and heterozygous subjects
Mov Disord
 , 
2010
, vol. 
25
 (pg. 
413
-
20
)
Yu-Wai-Man
P
Griffiths
PG
Gorman
GS
Lourenco
CM
Wright
AF
Auer-Grumbach
M
, et al.  . 
Multi-system neurological disease is common in patients with OPA1 mutations
Brain
 , 
2010
, vol. 
133
 (pg. 
771
-
86
)