Nine pedigrees of the Scottish families with autosomal dominant hereditary spastic paraplegia showing evidence of linkage to the locus of SPG4, hich encodes spastin. Solid symbols indicate affected individuals; circles, females; squares, males; and slashes, deceased.
Sequence analysis showing the normal sequence for SPG4,which encodes spastin, compared with that with the exon 8 mutation (arrow) detected in the family SCO-C03, as well as corresponding amino acid sequences. A indicates adenosine; C, cytidine; G, guanosine; T, thymidine; N, heterozygous condition (adenosine/guanosine); Asn, asparagine; Ser, serine; and Gly, glycine.
Extended haplotype analysis at the locus for SPG4,which encodes spastin. Identical haplotypes on chromosome 2p region were found in patients from the 9 families with the N386S mutation. bp Indicates base pairs.
Localization of spastin N386S to microtubules. Wild-type spastin and spastin N386S were expressed as myc or green fluorescent protein (GFP) fusion in Cos-7 cells and are visible as green signal (anti-myc or GFP); nuclei are stained with DAPI (4′,6-diamidine-2′-phenylindole dihydrochloride; Roche Diagnostics, Rotkreuz, Switzerland), blue signal; microtubules were demonstrated by means of a monoclonal antibody against α-tubulin, red signal. A, Wild-type spastin accumulates in cytosolic aggregates. B, Spastin N386S expression begins in correspondence with the microtubule organizing center and then accumulates in filamentous structures. C and D, These filaments colocalize with microtubules. E and F, Microtubule association was confirmed by nocodazole treatment, which disrupts both the filamentous pattern of spastin (E) and the microtubule network (F).
Sagittal spin-echo T1-weighted (A) and T2-weighted (B) magnetic resonance images of a patient carrying the N386S mutation in the family SCO-A01. Images demonstrate thinning of the corpus callosum, more evident in its truncus, and mild atrophy of the cerebellar vermis.
Orlacchio A, Kawarai T, Totaro A, Errico A, St George-Hyslop PH, Rugarli EI, Bernardi G. Hereditary Spastic ParaplegiaClinical Genetic Study of 15 Families. Arch Neurol. 2004;61(6):849-855. doi:10.1001/archneur.61.6.849
Copyright 2004 American Medical Association. All Rights Reserved. Applicable FARS/DFARS Restrictions Apply to Government Use.2004
Autosomal dominant hereditary spastic paraplegia (ADHSP) is mainly caused by mutations in the SPG4 gene, which encodes a new member of the AAA (adenosine triphosphatases associated with diverse cellular activities) protein family (spastin). Accumulation of genotype-phenotype correlation is important for better understanding of SPG4-linked hereditary spastic paraplegia.
To perform a clinical and genetic study of families with ADHSP and to perform the functional analysis of the founder mutation discovered in the SPG4 gene.
Genetic and clinical study.
Fifteen unrelated families with ADHSP originating from southern Scotland.
Main Outcome Measures
Clinical assessment, linkage analysis, haplotype study, expression of mutant spastin protein in cultured cells.
Nine families with ADHSP were linked to the SPG4 locus at 2p21-p24. Sequence analysis of SPG4showed a novel N386S mutation in all 9 of these families. Expression of mutant spastin showed aberrant distribution in cultured cells. Haplotype analysis suggested the existence of a common founder. Clinical examination of the affected members carrying the mutation showed phenotypic variations including broad range of age at onset and disease duration and additional neurologic features such as mental retardation. Magnetic resonance imaging demonstrated unique features, including thin corpus callosum and atrophy of the cerebellum in 2 patients. Linkage and sequence analyses showed no evidence of linkage to the currently known ADHSP loci in the remaining 6 families.
A founder SPG4 mutation N386S was identified in the families with ADHSP originating from southern Scotland. Clinical investigation showed intrafamilial and interfamilial phenotypic variations. The genetic study demonstrated evidence of further genetic heterogeneity in ADHSP.
Hereditary spastic paraplegia (HSP) is a clinically and genetically heterogeneous neurodegenerative disease characterized by progressive spasticity and weakness of the lower limbs. To date, 10 loci have been identified in autosomal dominant HSP (ADHSP) and mapped to chromosomes 14q11.2-q24.3 (SPG3A), 2p21-p24 (SPG4), 15q11.1 (SPG6), 16q24.3 (SPG7), 8q23-q24 (SPG8), 10q23.3-q24.1 (SPG9), 12q13 (SPG10), 19q13 (SPG12), 2q24-q34 (SPG13), and 9q33-q34 (SPG19).1 Five causative genes, which encode the proteins atlastin (SPG3A), spastin (SPG4), paraplegin (SPG7), kinesin heavy chain (KIF5A) (SPG10), and mitochondrial chaperonin HSP60 (SPG13), have been identified for 5 of the 10 ADHSP loci.2- 6 Spastin, the most common protein mutated in ADHSP, belongs to the so-called AAA family (adenosine triphosphatases associated with diverse cellular activities).3
Herein we present a novel SPG4 mutation that displays a founder effect in the Scottish population. Patients carrying the mutation demonstrated intrafamilial and interfamilial phenotypic variations, including novel magnetic resonance imaging findings such as thin corpus callosum and cerebellar atrophy.
The study was performed according to a protocol reviewed and approved by the Ethics Committee of the Istituto di Ricovero e Cura a Carattere Scientifico (IRCCS) Santa Lucia, Rome, Italy.
Included in the study were 14 families that emigrated to Italy and 1 family that emigrated to Canada from the south of Scotland between 1996 and 2003, with a history of HSP and at least 2 living first-degree relatives. Participants were recruited as Scottish if the proband or his or her family members were born in, or originated from, Scotland. After informed consent was obtained, a detailed clinical assessment of the pedigrees was undertaken, and a blood sample was taken for DNA extraction. Individuals of all ages from the 15 families were seen by at least 2 neurologists (A.O. and G.B.) and underwent full general and detailed neurologic examination. Information regarding family members who had died before the study began was taken from living relatives and corroborated, when possible, by medical records. All clinical evaluations, including the age at onset and the disease duration, were performed as previously described.7 The presence or absence of additional symptoms associated with HSP and the age at which these abnormalities appeared was carefully evaluated according to a rigorous protocol that included a full medical history and examination, with magnetic resonance imaging of the brain and spinal cord, esophageal endoscopy, Wechsler Adult Intelligence Scale–Revised, and neurophysiologic assessment with electromyography of the upper and lower extremities and motor nerve conduction studies of the tibial nerve including F-wave analysis. Audiologic studies were also performed, including impedance audiometry, otoacoustic emissions, and evoked brainstem response tests.
We initially performed a linkage study of all 15 families with ADHSP by means of 2 genetic markers flanking the currently known ADHSP loci (SPG3A, SPG4, SPG6, SPG7, SPG8, SPG10, SPG12, SPG13, and SPG19), which are informative microsatellite markers used in previous studies.7 Because clinical features, especially the additional neurologic features of family SCO-B02, overlapped with those of 2 families with ADHSP linked to the SPG9 locus, we included 2 microsatellite markers flanking the SPG9 locus previously described.8,9 To extend the haplotype analysis at the SPG4 locus, we included 6 other genetic markers (D2S2247, D2S365, D2S390, D2S367, D2S2230, and D2S2186). Calculation of 2-point logarithm of odds (LOD) scores under a genetic model based on clinical information was performed as previously described.7 No phenocopies were allowed. Marker allele frequencies were calculated by genotyping a panel of 200 unrelated Scottish individuals. At least 1 definite affected member from each family was subjected to the sequence analysis of 4 known causative genes for ADHSP (SPG3A, SPG4, SPG7, and SPG13) as described elsewhere.2- 4,6 Haplotype reconstruction was carried out over an approximately 9.8-megabase (National Center for Biotechnology Information sequence map) region by using the microsatellite markers cen-D2S2247-D2S365-D2S390-D2S352-N386S-D2S2347-D2S367-D2S2230-D2S2186-tel, as described elsewhere.7
Wild-type spastin and spastin N386S were transfected and the distribution of the wild and mutant spastin protein and α-tubulin was detected, as previously described.10 To identify the nuclei, cells were stained with the DNA-specific stain 4′,6-diamidine-2′-phenylindole dihydrochloride (DAPI; Roche Diagnostics, Rotkreuz, Switzerland).
Nine families with ADHSP (SCO-A01, SCO-C03, SCO-D04, SCO-E05, SCO-F06, SCO-I09, SCO-L12, SCO-M13, and SCO-O15) showed evidence of linkage to the SPG4 locus (Figure 1). Sequence analysis of SPG4 in affected members from 9 families, showing cumulative 2-point LOD scores of 10.12 at the recombination fraction θ = 0.0 for the SPG4 locus, demonstrated a nucleotide substitution A to G at 1157, resulting in a novel missense mutation, N386S (Figure 2). The polymerase chain reaction–restriction fragment length polymorphism analysis showed a heterozygous mutant allele in all affected members (data not shown). The missense mutation N386S cosegregated with the affected individuals of each family with ADHSP and could not be detected in the genomic DNA sequence of 100 normal Scottish controls. Neither phenocopy nor incomplete penetrance was identified in these 9 families with ADHSP. All affected members carrying the N386S missense mutation from the previously mentioned 9 kindreds had the same haplotypes (identical alleles for each different genetic marker) over an interval of at least 8.2 centimorgans (approximately 8 megabases, National Center for Biotechnology Information sequence map), since the alleles for all 6 markers from D2S365 to D2S2230 were identical (Figure 3).
The expression experiment of spastin N386S mutation in cultured cells demonstrated an aberrant subcellular localization, as previously observed for different mutations within the AAA cassette (Figure 4).10
Examination of the 10 ADHSP loci in 6 families (SCO-B02, SCO-G07, SCO-H08, SCO-J10, SCO-K11, and SCO-N14) demonstrated a negative LOD score between recombination fractions of 0.0 and 0.05. Cumulative 2-point LOD scores at each microsatellite marker showed evidence of exclusion of linkage within an interval of 0.1 centimorgan (LOD scores <−2.0) (data not shown). Moreover, sequence analysis of the open reading frame of the other known causative genes, except KIF5A, demonstrated no mutation in the affected members from the 6 families.
Detailed clinical examinations of each family are summarized in Table 1 and Table 2. In the 9 families carrying the N386S mutation, a thin corpus callosum was observed in 1 of the 3 affected members in the SCO-A01 family and in 1 of the 2 affected members of the SCO-O15 pedigree. In the patients with thin corpus callosum (SCO-A01 and SCO-O15), ataxia was not apparent, but the magnetic resonance images showed mild atrophy of the cerebellar vermis (Figure 5). Mental retardation was observed in 1 of the 2 affected members of the SCO-D04 family and in 1 of the 3 affected members of the SCO-F06 family. Moreover, the age at onset of spastic gait in the affected individuals from the pedigrees with the same mutation varied from 11 years (family SCO-D04) to 53 years (families SCO-I09 and SCO-M13). The duration of the disease also varied widely among patients.
Linkage to any of the 10 known ADHSP loci was unlikely in the remaining 6 families with ADHSP, as shown by sequence analysis and linkage study. Genealogic study of these families showed no relationship to each other. Clinical investigation demonstrated distinctive features in family SCO-B02, including hearing impairment, pes cavus, and persistent vomiting. Audiovestibular examinations showed that hearing impairment was due to auditory neuropathy. Brainstem-evoked potentials showed delayed bilateral I-III latency in the members examined. Esophageal endoscopy showed that persistent vomiting was due to hiatal hernia. Hiatal hernia was identified in 12 of 18 affected members. An impression of genetic anticipation (an earlier age at onset in succeeding generations) was obtained from the clinical records in family SCO-B02. Additional neurologic features observed in family SCO-B02 overlapped with those found in 2 families linked to the SPG9 locus8,9; however, linkage to the locus was excluded by 2-point analysis (data not shown). Genome-wide survey of family SCO-B02 showed evidence of linkage to a new locus on chromosome 1 (A.O., unpublished data, 2004). The other 5 families showed no distinct additional neurologic features.
In this study, we report genetic analyses of 15 Scottish families with ADHSP. In 9 of these families, the trait was linked to the SPG4 locus and arose from a novel mutation in the SPG4 gene, N386S. Haplotype analysis of affected members strongly suggests the presence of a founder effect in the southern Scottish population, although no ancestor could be identified by genealogic studies. Some affected members carrying the SPG4 mutation emigrated from Scotland to Italy and to the Canadian provinces of Nova Scotia and Ontario. Further genetic studies of other Scottish families with ADHSP in Scotland, Italy, and Canada could disclose the same SPG4 mutation.
The N386S missense mutation is located within the spastin Walker motif A. Asparagin at codon 386 is conserved in the Mus musculus (XP_128755) and Drosophila melanogaster (AAN71010) spastin orthologues. A serine at codon 386 has never been reported in public databases. A pathologic role of the N386S substitution is supported by the effect of expressing this mutation in eukaryotic cells. The N386S spastin showed constitutive binding to a subset of microtubules, which were reorganized in thick perinuclear bundles with disappearance of the aster. Although binding of endogenous spastin to microtubules has not been demonstrated so far,11 this result is in agreement with previous data obtained in vitro with other missense mutations in the AAA domain.10
The broad range of age at onset of spastic gait and the variable presence of other neurologic features such as mental retardation suggest that other genetic or nongenetic factors may modify the phenotype of the mutation. Thinning of the corpus callosum has been reported in autosomal recessive HSP,12 where the genetic defect remains unknown. Thin corpus callosum was observed in 2 patients carrying the N386S mutation. Cerebellar atrophy was also demonstrated in the same 2 patients, although ataxia was not prominent. Similar magnetic resonance imaging findings have been reported in patients with autosomal recessive HSP, but clinically apparent ataxia was present.13 It remains unknown whether the N386S mutation and the genetic defect in autosomal recessive HSP with thin corpus callosum share a common biological pathway leading to maldevelopment or neurodegeneration, or both, in the corpus callosum, cerebellum, and corticospinal tract. Identification of the causative gene in autosomal recessive HSP with thin corpus callosum and further investigation of the effect of SPG4 mutations may clarify the underlying mechanism of these features.
The results of the genetic study also suggested further genetic heterogeneity in ADHSP. Family SCO-B02 may have an independent genetic disorder from the viewpoint of unique clinical features such as sensorial hearing impairment, pes cavus, and hiatal hernia. Pes cavus and persistent vomiting due to hiatal hernia were also described in one large Italian family and in one smaller British pedigree with HSP linked to the SPG9 locus.8,9 In addition, in the SPG9 pedigrees, cataract and muscle wasting were also present.8,9 Neurophysiologic examinations demonstrated that muscle wasting observed by Seri et al8 and Lo Nigro et al9 was due to axonal motor neuropathy, but no evidence of involvement of the lower motor neuron was obtained in family SCO-B02. Further genetic study of the family may show gene and locus identification and allow determination of genotype and locus–phenotype correlations.
Corresponding author and reprints: Antonio Orlacchio, MD, PhD, Laboratorio di Neurogenetica, IRCCS Santa Lucia, Via Ardeatina 354, Rome 00179, Italy (e-mail: email@example.com).
Accepted for publication September 30, 2003.
Author contributions: Study concept and design (Drs Orlacchio, Kawarai, St George-Hyslop, Rugarli, and Bernardi); acquisition of data (Drs Orlacchio, Kawarai, St George-Hyslop, Rugarli, and Bernardi); analysis and interpretation of data (Drs Orlacchio, Kawarai, St George-Hyslop, Rugarli, and Bernardi, Mr Totaro, and Ms Errico); drafting of the manuscript (Drs Orlacchio, Kawarai, St George-Hyslop, Rugarli, and Bernardi, Mr Totaro, and Ms Errico); critical revision of the manuscript for important intellectual content (Drs Orlacchio, Kawarai, St George-Hyslop, Rugarli, and Bernardi); statistical expertise (Drs Orlacchio and Kawarai); obtained funding (Drs Orlacchio, Kawarai, St George-Hyslop, Rugarli, and Bernardi); administrative, technical, and material support (Mr Totaro and Ms Errico); study supervision (Drs Orlacchio, Kawarai, St George-Hyslop, Rugarli, and Bernardi).
This work was supported by a grant from the Ministero della Salute and grant GGP030368 from the Comitato Telethon Fondazione Onlus (Rome, Italy); the Canadian Institute of Health Research, the Howard Hughes Medical Research Foundation, the Canadian Genetic Disease Network, and the Alzheimer Society of Ontario (Toronto); and the Nakabayashi Trust for ALS Research, 2003 (Tokyo, Japan).
We thank all the patients and family members involved in this study. We also thank Peter Bross, PhD, and Susan Ling, BSc Hons, for their assistance. We are extremely grateful to the Genetic Bank of the Laboratorio di Neurogenetica, IRCCS Santa Lucia, Rome, Italy.