Developmental and degenerative features in a complicated spastic paraplegia

Abstract

Objective

We sought to explore the genetic and molecular causes of Troyer syndrome, one of several complicated hereditary spastic paraplegias (HSPs). Troyer syndrome had been thought to be restricted to the Amish; however, we identified two Omani families with HSP, short stature, dysarthria and developmental delay—core features of Troyer syndrome—and a novel mutation in the SPG20 gene, which is also mutated in the Amish. In addition, we analyzed SPG20 expression throughout development to infer how disruption of this gene might generate the constellation of developmental and degenerative Troyer syndrome phenotypes.

Methods

Clinical characterization of two non-Amish families with Troyer syndrome, followed by linkage and sequencing analysis. qPCR and in situ hybridization analysis of SPG20 expression in embryonic and adult human and mouse tissue.

Results

Two Omani families carrying a novel SPG20 mutation display clinical features remarkably similar to the Amish patients with Troyer syndrome. SPG20 mRNA is expressed broadly but at low relative levels in the adult brain; however, it is robustly and specifically expressed in the limbs, face and brain during early morphogenesis.

Interpretation

Null mutations in SPG20 cause Troyer syndrome, a specific clinical entity with developmental and degenerative features. Maximal expression of SPG20 in the limb buds and forebrain during embryogenesis may explain the developmental origin of the skeletal and cognitive defects observed in this disorder.

Introduction

Hereditary spastic paraplegias (HSPs) comprise several disorders commonly divided into two subgroups: “pure” HSPs characterized by progressive spasticity in the lower limbs due to pyramidal tract degeneration and “complicated” HSPs where lower limb spasticity is associated with a variety of other neurological signs and clinical features. Complicated HSPs are clinically heterogeneous, mainly autosomal recessive syndromes, frequently described and mapped in sporadic families within inbred populations^1-3. Because of this heterogeneity, diagnosis and recommendations for genetic testing in these disorders have been a daunting task.

Troyer syndrome [OMIM #275900] is a complicated HSP associated with short stature, skeletal abnormalities, dysarthria and developmental delay, first described in the Old Order Amish^{4, 5}. Since the original description in 1967, several Troyer-like syndromes have been reported^6-9, but they often differed from “classical” Troyer syndrome in their neurological or skeletal features. The Amish founder mutation is a single nucleotide deletion in the SPG20 gene¹⁰, leading to the loss of the spartin protein¹¹; however, no additional SPG20 mutations were subsequently identified⁵, and it was suspected that Troyer syndrome may be restricted to the Amish.

We identified an Omani kindred presenting with clinical features resembling Troyer syndrome. All affected Omani individuals had a novel homozygous null mutation in SPG20, and their clinical descriptions matched closely those of Amish Troyer syndrome individuals of comparable ages. Because Troyer syndrome is associated with developmental features, such as short stature, skeletal abnormalities, and global developmental delay, we investigated the sites of action of SPG20 during development in humans and mice. SPG20/Spg20 expression in the adult is relatively modest and widespread in the nervous system; however, maximal and focal expression is observed in the embryonic limb buds, face and forebrain during early morphogenesis, which might explain the developmental phenotypic changes involving the extremities, face and brain.

Subjects and Methods

Enrollment and clinical studies

The subjects belonged to two families residing in a remote region of Oman and were originally ascertained by a local clinical geneticist. Detailed family and medical histories were obtained by a genetic counselor, who is a native Arabic speaker, and a developmental pediatrician, also a native Arabic speaker, conducted a developmental assessment. A pediatric neurologist performed a standard neurological examination, and a second clinical geneticist obtained anthropometric measurements. The height of all subjects and head circumference of subjects under the age of 36 months were plotted on the CDC (Centers for Disease Control and Prevention, USA) 2000 growth charts¹², while head circumference of subjects above 36 months and all other anthropometric measurements were plotted on standard charts¹³. Written, informed consent was obtained from the subjects or their legal guardians. The Ministry of Health in Oman and the Institutional Review Board of Children's Hospital Boston approved this study.

Linkage analysis

Genomic DNA was purified from lymphocytes separated from peripheral blood using commercial kits (Qiagen). 600 ng of genomic DNA were used to hybridize Affymetrix Human SNP Array 6.0 at the Genomic Analysis Core of the UNC Neuroscience Center of the University of North Carolina School of Medicine, Chapel Hill, NC. After removal of low quality calls and of Mendelian and non-Mendelian errors using Merlin software¹⁴, the SNP data was analyzed using Allegro software¹⁵ under a fully penetrant autosomal recessive model. For microsatellite analysis, highly polymorphic microsatellite markers were chosen from the Marshfield database in the UCSC Genome Browser. Fluorescently labelled PCR primers (ABI) were used to amplify DNA samples using standard conditions and PCR products were resolved on an ABI 3130xl Genetic analyzer.

SPG20 sequencing analysis

Sequencing of SPG20 (NM_015087; NM_001142294-6) coding region was performed by SeqWright (Houston, TX) on PCR products after amplification of genomic DNA. PCR primers were designed for each exon including at least 50 base pairs of flanking intronic sequences. Primer sequences are available upon request. Once a putative change was identified at least 96 control DNAs (192 alleles) were tested to exclude the possibility of a benign polymorphic change.

Patient cell lines and mutated protein/mRNA detection

Transformed lymphoblastoid cell lines were established from peripheral blood of affected and unaffected individuals at the Partners Center for Personalized Genetic Medicine, Cambridge, MA. Cell lines were maintained in RPMI medium supplemented with 10% fetal bovine serum, glutamine (2mM) and penicillin/streptomycin (100U/100μg) (all from Gibco). Protein lysates were obtained by boiling a cell pellet in Laemmli sample buffer (Bio-Rad). Western blotting was performed using standard protocols on Bio-Rad equipment. Presence of spartin was assessed using a rabbit anti-SPG20 antibody (Proteintech Group, Inc) and a goat anti-rabbit IRDye-800CW secondary antibody (Li-Cor Biosciences) detected on an Odyssey Imager (Li-Cor Biosciences).

Total RNA was purified from patient cell lines using the mirVana RNA Isolation Kit (Ambion). Random primed cDNA was generated using reverse transcriptase (Promega) and analyzed by quantitative PCR (qPCR) using SYBR Green reagents (ABI) on an ABI 7500 qPCR platform as previously described¹⁶. Primer sequences are available in Suppl. Table 1.

SPG20/Spg20 expression analysis

cDNA samples from human fetal and adult post-mortem brain tissue were obtained commercially (BioChain and Clontech). cDNA from embryonic and post-natal mouse brains was generated as described above. qPCR was performed as described above (Suppl. Table 1 for primer sequences). DNA templates for in situ hybridization probes were cloned using a 550bp fragment of the Spg20 coding sequence using the following primers: Spg20-BamHI 5′-ctacatggatccgatatcaaccggaggagcagccaaagtcagc-3′ and Spg20-SalI 5′-gtccttgtcgactgcccctggcttctcttcctccacctg-3′. Digoxigenin-labeled riboprobes were synthesized and in situ hybridization was performed as previously described¹⁷. Video-images of the hybridized sections were obtained on a Wild photomicroscope or Leica DMR microscope under consistent illumination conditions.

Results

Identification of a novel SPG20 mutation in an Omani kindred

We examined two related Omani families presenting with short stature, dysarthria, motor and cognitive developmental delay and increased muscle tone (Families 1 and 2 in Fig.1A). The fathers (individuals 1-2 and 2-1) were brothers and the mothers (1-1 and 2-2) were aunt and niece. Family 1 had nine children: four reportedly unaffected females, who were married and lived elsewhere, one male (1-4), who was unaffected by clinical examination, three affected males (1-3, 1-6 and 1-7) and one affected female (1-5). Family 2 had eight children: four unaffected girls and one unaffected boy, three of whom were present at the time of our evaluation (2-4, 2-5 and 2-6), one affected male (2-3), one affected female (2-7). The youngest child (2-8) was small for her age and could neither talk nor walk, but did not display overt neurological symptoms; we were not able to obtain DNA from her for genotyping or sequencing.

Figure 1. Affected individuals carry a homozygous null mutation in the SPG20 gene.

A. Pedigree of two related Omani families affected with short stature, spasticity, dysarthria and developmental delay. Affected individuals are in black and unaffected in white. The status of individual 2-8 (in gray) could not be determined, since she is too young to properly assess neurological symptoms. All numbered individuals were examined, but genomic DNA was only collected from individuals for which microsatellite analysis is shown. Microsatellite analysis revealed common maternal (in yellow) and paternal (in blue) haplotypes in all affected. Microsatellite markers in the linkage region identified by the SNP analysis are in bold. One homozygous marker (D13S219) is common to all affected individuals (highlighted in orange). B. The homozygous region contains 6 genes, including SPG20 (in orange). C. The affected individuals carried a homozygous 2bp deletion in SPG20, which was present in heterozygosity in the carriers. D. Western blot analysis of patient cell lines showed that full length SPG20 protein is missing in the affected individuals (A) compared to a non-affected non-carrier sibling (NA). E. qPCR analysis of cDNA from the patient cell lines indicated that SPG20 mRNA is not present in the affected individual (A).

Genomic DNA was obtained from most individuals in the pedigree and hybridized to Affymetrix Human SNP Array 6.0 chips for genome-wide genotyping. We suspected the presence of a homozygous ancestral mutation because the families resided in an isolated mountain village and large stretches of homozygous DNA identified in all individuals suggested some degree of consanguinity; however, we could not ascertain any shared ancestry as far as five generations removed from the probands by pedigree analysis. Thus, the SNP data was first analyzed following an autosomal recessive model assuming no consanguinity. Linkage analysis identified one region of 7.0 Mb flanked by SNPs rs11619306 and rs9576104 on chromosome 13q12.3 (LOD=3.6). Microsatellite analysis confirmed a shared maternal and paternal haplotype in all affected individuals (Fig.1A). Individual 1-4, who was unaffected, was homozygous for the maternal haplotype throughout most of the linkage region, excluding maker D13S219, which was homozygous on all affected individuals. At this locus, the SNP data indicated an 844 kb region of homozygosity between markers rs1418987 and rs9315443, shared only by affected but not unaffected individuals. Among the six genes in this region (Fig.1B), SPG20 was the strongest candidate gene since individuals carrying an SPG20 mutation were affected with a remarkably similar phenotype, a complicated form of hereditary spastic paraplegia associated with short stature, dysarthria and developmental delay, called Troyer syndrome^{4, 10}.

Sequencing of the SPG20 coding regions revealed a homozygous 2bp deletion (c.364_365delAT), which resulted in an amino acid substitution followed by a stop codon in the first coding exon (p.M122VfsX1; Fig.1C). This mutation was confirmed in all affected individuals and the inheritance pattern was confirmed in the rest of the families. This mutation was not identified in control individuals indicating this is not a common polymorphism.

Since the SPG20 mutation in the Amish (c.1110delA) was reported to be a null mutation¹¹, we generated lymphoblastoid cell lines from two affected (1-3 and 1-5) and one unaffected individual (2-5). We found that the full length spartin protein was absent in the affected individuals (Fig.1D). We also carried out qPCR for SPG20 on cDNA samples from the two of the lymphoblastoid lines and could not detect any mRNA in the affected individual (Fig.1E), indicating that this mutation is a null allele.

Clinical characterization of patients carrying a novel SPG20 mutation

To ascertain phenotypic similarities or differences among individuals carrying SPG20 mutations, we compared the clinical presentations of the Omani individuals with the clinical description of twenty-one Amish patients by Proukakis et al.⁵ (Table 1). All affected individuals in the Omani cohort presented with short stature and dysarthria, and weredelayed in reaching motor and cognitive developmental milestones. They all had some difficulties walking with clumsy, mildly spastic gait, which their parents reported to be worsening over time. The most common physical features were relative hypertelorism and overgrowth of the maxilla leading to overbite (Fig.2A-D), as well as hand and feet anomalies such as brachydactyly (5/6 patients) (Fig.2E), hammer toes, and pes cavus (Fig.2G-H). Additional non-specific skeletal malformations observed in the hands were clynodactyly, camptodactyly and hypoplastic 5^th middle phalanges (Fig.2E-F). Most affected individuals had persistent cognitive deficits and poor performance in school, but no emotional lability was reported. Detailed neuropsychological testing was not available.

Table 1. Clinical features of the Omani Troyer syndrome individuals and comparison with the Amish cohort.

Individual	1-3	1-5	2-3	1-6	1-7	2-7	Omani	Amish I	Amish II
Age (years)	27	19	19	17	14	4	mean 16.6	mean 15	mean 43.5
Sex	M	F	M	M	M	F
Height (cm)	149	133	145	146	n/a	83	short	short	short
%	≪3rd	<3rd	≪3rd	≪3rd		≪3rd
OFC (cm)	52.4	51.6	52.8	54.8	53.0	48.4	normal	normal	normal
%	3rd-10th	<3rd	10th	25th-50th	10th-25th	3rd-10th
Developmental milestones and cognition
Talked at	2y	2y	3y	late	late	late	late(100%)	late(80%)	late(87%)
Walked at	2y	3y	1.5y	2.5y	3y	1.5y	late(67%)	late(80%)	late(87%)
Poor school performance	+	+	+	-	+	n/a	+(80%)	+(100%)	+(80%)
Emotional lability	-	-	-	-	-	n/a	no	+(70%)	+(91%)
Neurological exam
Dysarthria	+	+	+	+	+	n/a	+(100%)	+(100%)	+(100%)
Tongue dyspraxia	+	-	+	+	+	n/a	+(80%)	often	often
Distal amyotrophy	+	-	+	+	+	-	+(67%)	+(70%)	+(100%)
Hyperreflexia
upper	-	-	+	+	-	-	+(33%)	+(30%)	+(90%)
lower	+	-	+	+	+	-	+(67%)	+(100%)	+(100%)
Ankle clonus	-	-	+	+	+	-	+(50%)	+(30%)	+(36%)
Clumsy, spastic gait	+	+	+	+	+	+	+(100%)	+(100%)	+(100%)
Dysmetria	-	-	+	+	+	-	+(50%)	+(70%)	+(100%)
Skeletal abnormalities
Overbite	+	+	+	+	+	?	+(100%)	no	no
Hypertelorism	+	+	+	-	+	+	+(83%)	no	no
Hand/foot abnormalities	+	+	+	+	+	+	+(100%)	+(100%)	+(100%)

Figure 2. Morphological and radiological features of Omani individuals with Troyer syndrome.

Comparison between an unaffected (A, C) and an affected individual (B, D) revealed relative hypertelorism (A-B) and pronounced overbite (C-D). Examples of skeletal anomalies in the extremities: brachydactyly (short digits; E-F), camptodactyly (flexed digit; asterisks in E-F), hammer toes (asterisks in G-H) and clinodactyly (curved digit; arrowheads in G-H). Brain MRI shows atrophy of the cerebellar vermis (arrowhead in I) and white matter hyperintensity in T2-weighted images, which is more prominent posteriorly (arrowheads in J).

The neurological examination revealed distal amyotrophy (4/6 patients), dysmetria in the upper extremities (3/6 patients), hyperreflexia, which was more severe in the lower limbs (4/6 patients), and ankle clonus (3/6 patients). Heel cords were tight in the four older affected individuals. Brain MRI was performed in two individuals (1-7 at 14y and 2-3 at 19y). Both individuals had mild atrophy of the cerebellar vermis, mild white matter volume loss, and revealed periventricular white matter hyperintensity in T2-weighted images, consistent with gliosis (Fig.2I-K).

The mean age of our cohort (16.6y) was younger than the Amish cohort, which might explain the milder and more variable phenotype observed in the Omani individuals; therefore, we subdivided the Amish cohort into a younger subset (Amish I) comparable in age to our cohort (<27y; mean age 15y) and an older subset (Amish II) (>28y; mean age 43.5y). Our cohort closely matched the description of the Amish I group, while the Amish II group was more severely affected, consistent with the progressive, degenerative aspects of Troyer syndrome (Table 1).

Expression of Spg20 in the developing and adult nervous system

There appears to be two major components in Troyer syndrome: an early developmental aspect and a neurodegenerative process. The combination of phenotypes including limb, craniofacial and behavioral anomalies is common in complex genetic disorders affecting early morphogenetic events¹⁸, and the relevant genes are often specifically expressed in the embryonic rudiments of the affected adult structures. We first analyzed SPG20 expression in the developing and adult human brain using quantitative PCR (qPCR). SPG20 was expressed at modest levels in the fetal and adult human brain compared to a highly expressed neural specific gene, synaptophysin (SYP; Fig.3A-B). Levels of SPG20 were highest in the amygdala, cortex and thalamus, and lowest in the hippocampus and cerebellum. Though clearly detectable, SPG20 was expressed at substantially lower levels than SYP. Expression levels and distribution of the mouse orthologue of SPG20, Spg20, in the adult mouse brain parallel that of the human, with some divergence; particularly, relative expression of Spg20 in the hippocampus was slightly elevated compared to humans (Fig.3C). In addition, we analyzed Spg20 expression in mouse spinal cord and brain stem and found that it was highest in the spinal cord of all brain regions measured (Fig.3C). As in the human, Spg20 was expressed at comparatively lower levels than Syp (Fig.3D). Spg20 is not a brain-specific gene, since it is also expressed in several other organs^{10, 19} (Fig.3E). It is, however, developmentally regulated, as expression was maximal at mid-gestation, embryonic day (E)10, and declined precipitously thereafter (Fig.3F).

Figure 3. qPCR analysis of SPG20/Spg20 expression in human and mouse tissues.

A. There is modest variation in SPG20 local expression in the adult human brain when distinct regions are compared to whole brain samples. B. Human SPG20 expression is substantially lower than that the brain-specific gene, synaptophysin (SYP) across all brain regions with the exception of the amygdala. C. A similar profile of regionally variable Spg20 expression, with some modest differences, is seen in the mouse brain. D. Lower Spg20 expression is observed when compared to Syp. E. Spg20 is expressed in several murine tissues beside the brain. F. Spg20 is developmentally regulated in the mouse embryo with highest expression levels in the whole embryo at E10.5.

We then used in situ hybridization to localize Spg20 in adult, fetal and embryonic mouse brain. Spg20 was expressed at relatively low levels in neurons and glia throughout the adult brain (Fig.4A-C), including glia in fiber tracts (Fig.4C). Modestly elevated expression was seen throughout hippocampal stratum pyramidale of the CA fields and dentate gyrus (Fig.4B) and the transitional parahippocampal/entorhinal cortex (Fig.4D). Low expression was seen in the cerebral cortex with no apparent laminar or cell class specificity (Fig.4E). The only additional site of elevated expression in the forebrain was the habenular complex, including the habenular recess appended to the corpus callosum (Fig.4F-G). Spg20 is expressed throughout the cerebellum in Purkinje cells, granule cells, and scattered cells in the molecular layer (Fig.4H). In the brainstem, there is robust expression in large neurons, probably motor neurons, and in the facial nucleus (Fig.4I). Spg20 is also expressed in cells distributed throughout the spinal cord with no noticeable discontinuities (Fig.4J). As predicted from the qPCR data, expression in the fetal brain (Fig.4K) is relatively low, with no apparent regional distinctions. Aside from elevated expression in the lens placode and pigment epithelium (Fig.4L), cochlear epithelium (Fig.4M) and condensing mesenchyme at sites of myogenic or cartilage formation (Fig.4N), there is little focal expression at fetal stages (E12.5 and E14.5 shown here).

Figure 4. In situ analysis of Spg20 expression in mouse.

A-J. Spg20 expression in the adult brain is modest and focal. Cortical expression is moderate and not layer-specific (E). Relatively higher expression is observed in the hippocampus (B), glial cells in the corpus callosum (C), entorhinal cortex (D), habenular complex (F-G), cerebellum (m=molecular layer, P=Purkinje cells payer, g=granule cell layer; H), facial nucleus (fN, I) and spinal cord (J). K-N. Fetal expression is also modest and distributed in the brain (Ctx=cortex, LGE=lateral ganglionic eminence, MGE=medial ganglionic eminence;K) with higher expression in the lens placode (le, L), the cochlear epithelium (M) and condensing mesenchyme at sites of myogenic and cartilage formation (asterisks in N). O-Q. Early embryonic expression (E10.5) is highly patterned and enhanced in the forebrain (fb), frontonasal mass (fnm), maxilla (mx), branchial arches (ba1, ba2), heart (h) and limb buds (flb) (O). The antisense control is completely unlabeled (P). Selective, focally elevated expression in the fnm/fb, flb, and ba1/2 is confirmed by qPCR after microdissection compared to the heart (h) and whole embryo (Wh) (Q).

In contrast to the later stages, and, as predicted from the qPCR analysis, Spg20 expression in the mid-gestation embryo is elevated and highly patterned. In the E10.5 embryos, shortly after neural tube closure, Spg20 is specifically expressed in the initial frontonasal mass/forebrain, craniofacial structures, aortic arch/heart primordium, and limb buds during morphogenesis, with lowest expression in the heart (Fig.4O-P). qPCR analysis confirms this distribution; expression levels in limb buds, branchial arches and frontonasal mass/forebrain are substantially elevated compared to the heart as well as to whole E10.5 embryo (Fig.4R). Our observations suggest that Spg20 has its most specific and maximal activity in the limbs, face and forebrain during early morphogenesis. This is remarkable since the sites of phenotypic manifestations in the affected members of the Omani Troyer syndrome kindred described here include anomalies of the limb extremities, face, and brain.

Discussion

Our identification of a novel, disease-associated SPG20 mutation in an Omani kindred, indicates for the first time that Troyer syndrome is not restricted to the Amish, as previously proposed². Complicated autosomal recessive HSPs are heterogeneous disorders often lacking clear clinical and molecular diagnostic guidelines. Of the seventeen distinct loci for complicated HSP identified to date, more than half have been described in single families or isolated populations^1,2. In the past four decades, several HSP syndromes with Troyer-like features have been reported, but none matched the clinical presentation in the Amish cases⁵. Some of these disorders were later mapped to different loci such as: SPG26^{6, 20}, SPG39²¹ and ARSACS^{22, 23}. Even within the isolated region where these Omani families reside, we had described a similar extended pedigree with dysarthria, mental retardation, cerebral palsy (CP) and microcephaly²⁴. Linkage analysis ruled out the SPG20 locus in the second kindred, confirming that the two disorders are separate clinical and genetic entities (A.R., R.S.H., C.A.W unpublished data). Four individuals in this extended family (N=45 individuals) also were heterozygous carriers for the SPG20 mutation, highlighting its presence in this isolated population. Thus, this novel mutation in SPG20 is geographically and genetically distinct from that in the Amish population; nevertheless, both mutations result in the phenotypic spectrum associated with Troyer syndrome.

Though a clear Troyer syndrome diagnosis is difficult to reach in young subjects due to the variability and mildness of their symptoms, direct comparison of clinical features showed that the Omani cohort closely resembled the age-matched Amish Troyer syndrome group (Amish I), suggesting that SPG20 null mutations cause a well-defined phenotype. As in the Amish, we observed short stature, skeletal abnormalities in the extremities, developmental delay and dysarthria of possible cerebellar origin from an early age. Therefore, a diagnosis of Troyer syndrome must be considered when this constellation of phenotypes is present in young children. Facial dysmorphism and skeletal features are subtle and in single cases in non-consanguineous populations this disorder could be initially diagnosed simply as cerebral palsy. Spasticity and distal amyotrophy appeared in the teen-age years and worsened slowly over time. In the Amish, brain MRI revealed white matter abnormalities, which were less severe in the youngest patient examined (age 15y), suggesting a progressive worsening of the condition⁵. MRI analysis indicates that the Omani Troyer syndrome individuals of comparable ages also have mild white matter abnormalities, supporting the hypothesis that white matter degeneration accompanies disease progression, although direct neuronal degeneration cannot be ruled out. In addition, we identified mild cerebellar atrophy, which is consistent with the cerebellar signs. Serial MRI scans on the Omani individuals might resolve whether progressive neurological symptoms are correlated with increased white or gray matter degeneration. Despite many similarities to the Amish cohort, a few specific differences were observed in the affected Omani individuals, including the presence of hypertelorism, and a pronounced overbite, and the absence of inappropriate emotional responses. As new cases are identified it will be interesting to assess whether these traits have variable penetrance or are population-specific.

Functional studies on spartin, the protein encoded by SPG20, have identified multiple roles in protein ubiquitination^{25, 26, 27}, lipid droplet formation^{19, 25, 26} and a possible link to EGF receptor signaling²⁷. Previous expression analyses by northern blot or qPCR identified widespread SPG20/spartin localization in the brain and in other tissues^{10, 19}; however, it was unclear whether regional and cellular dynamism in SPG20 expression could explain the phenotypes observed in the affected individuals. Our study shows that while SPG20/Spg20 expression is modest and virtually ubiquitous in the adult and developing brain, early stages of embryonic development show maximal levels of Spg20 in the limbs, face and forebrain primordia. This localized expression suggests a parallel role for SPG20/Spg20 in morphogenesis and differentiation at these phenotypic sites; accordingly, a loss of function mutation may contribute to the phenotypic spectrum of Troyer syndrome.