Overview of Genetic Causes of Spasticity in Adults and Children

CHAPTER 22 Overview of Genetic Causes of Spasticity in Adults and Children

Rebecca Schüle and Stephan Züchner

INTRODUCTION

Spasticity is a common and etiologically unspecific neurologic symptom. The term describes an increased muscle tone as a consequence of an overactivity of stretch reflexes. The pathophysiology of spasticity is complex. One main contributor, however, is dysfunction of the upper motor neurons that compose the corticospinal tract. Accordingly, spasticity is often accompanied by other signs and symptoms of corticospinal tract dysfunction: reduced voluntary movements, weakness, increased stretch reflexes, and presence of “pyramidal signs” such as an extensor plantar response. Probably the only disorder where spasticity is viewed as the defining symptom is hereditary spastic paraplegia (HSP), also known as familial spastic paraplegia (FSP). Utilizing spasticity to define and categorize HSP has been extremely successful for the identification of phenotypically similar families and the identification of underlying genes (1,2). From this clinical and genetic work it became evident that HSP is indeed a very heterogeneous group of diseases (1,3). HSP can be accompanied by a number of additional neurologic and nonneurologic signs and symptoms. This brings some complicated HSP forms in close relationship to motor-neuron diseases, hereditary ataxias, and axonopathies. This clinical, pathological, and genetic overlap of diseases, such as amyotrophic lateral sclerosis (ALS), distal hereditary motor neuropathy (dHMN), and hereditary motor and sensory neuropathy type 2 (HMSN/CMT2), will ultimately be advantageous for the understanding of the biology of HSP and the development of new therapies (4). Ultimately, the underlying gene defects will provide the basis of classifying such conditions. This chapter focuses on the hereditary spastic paraplegias and describes the spectrum of genes identified thus far. The molecular genetic findings of each gene are summarized and the major cell biological pathways involved in HSP are discussed. Finally, a perspective is provided on new developments in this field and potential approaches to therapeutic intervention are laid out.

HEREDITARY SPASTIC PARAPLEGIAS

HSPs are orphan diseases, collectively affecting approximately 1.2 to 9.6 in every 100,000 individuals (5,6). Traditionally, HSPs are divided into “pure” and “complicated” forms (1). Spastic paraparesis, increased stretch reflexes, and presence of pyramidal signs, often accompanied by neurogenic bladder disturbances and impairment of vibrations sense in the lower limbs, are the clinical hallmarks of all HSPs. In complicated HSP, additional neurologic and nonneurologic symptoms occur, including cognitive impairment, seizures, cerebellar ataxia, optic atrophy, or peripheral nerve involvement (1). If not confirmed genetically, the diagnosis of HSP is usually made by exclusion after alternative causes (structural, inflammatory, metabolic, and hereditary) have been ruled out. The presence of a family history of similar conditions can confirm the familial nature of the condition and provide information of the Mendelian trait (eg, X-linked, autosomal recessive, or autosomal-dominant disorder) (7).

GENETIC HETEROGENEITY IN HSP

Genetic studies have revealed more than 80 different chromosomal HSP loci (7). Of the at least 68 identified genes, 13 cause autosomal-dominant disease, 52 recessive, and three X-linked forms. About 60% of HSP cases report a positive family history, indicating an autosomal-dominant inheritance pattern in about three quarters and autosomal recessive inheritance in the remaining one quarter of cases. The rest—a sizable 40% of HSP cases—present as apparently sporadic disease. Only a small portion of these (<25%) can be explained by mutations in known HSP genes, and it is currently unknown whether mutations in unknown Mendelian HSP genes, more complex (eg, digenic and oligogenic) inheritance patterns, or even nongenetic factors contribute to the diagnostic gap in sporadic HSPs.

Among the autosomal-dominant HSPs, SPG4 is by far the most common form, explaining about 40% to 50% of cases. SPG3, SPG10, and SPG31 are about equally common with varying frequencies between approximately 3% and 10% depending on the cohort selection (8,9). All other autosomal-dominant subtypes of HSP account for less than 1% to 2% of cases (10,11).

Recessive HSPs are a rare cause for disease in the outbred Caucasian population, but are more frequent in regions such as Northern Africa and the Near East (12). Mutations in recessive genes are often associated with complicated forms of HSP. For example, SPG21, also known as Mast syndrome, was defined in an Amish family and the underlying (founder-) mutation was later identified in the gene maspardin (ACP33) (13). The affected patients suffered from dementia, developmental delay, and pseudobulbar, cerebellar, and extrapyramidal abnormalities (14). SPG11 is the most common form of autosomal recessive HSP, followed by SPG5 and SPG7, which can both cause pure as well as complicated forms of HSP.

Three genes are known to cause X-linked HSP: L1CAM, PLP1, and SLC16A2. All three genes most typically cause complicated forms of HSP although pure HSP has been described with some mutation types and in female carriers of PLP1 mutation. Other examples of rare, and not yet clarified, loci are SPG23 and SPG29 (15,16). SPG23 is a recessive HSP form with skin pigmentary abnormalities, and SPG29 is complicated by hearing impairment and persistent vomiting due to hiatal hernia (16).

Most genetic defects in HSP are point mutations or small insertions and deletions. However, Beetz et al have recently shown that small copy number variations in spastin at the subgene level are a relatively frequent cause of SPG4 (17). It remains to be seen whether this mutation mechanism does apply to other HSP genes. Another novel molecular mechanism causing HSP has been proposed that involves variations in highly conserved binding sites for micro-RNAs (9). These binding or target sites are generally present only in the three untranslated regions of genes. Micro-RNAs confer another means of gene expression by controlling translation and mRNA stability (18). Several SPG31 (REEP1) families have been reported that carried mutations in the target site of micro-RNA140 (miR140) in REEP1 (8,9).

CLINICAL HETEROGENEITY IN HSP

Although certain complicating features are more common in some genetic subtypes of HSP than in others, the clinical presentation of virtually all forms of HSP is highly variable and individual prediction of the genotype based on the phenotype is therefore not reliably possible with very few exceptions. Not a single so-called pure form of HSP such as, for example, SPG3 or SPG4 presents exclusively with the hallmarks of pure HSP—spastic paraparesis, bladder dysfunction, and impairment of vibrations sense. Instead, for all subtypes of HSP, complicated presentations have been described in at least some families. Age of onset, initially believed to discriminate SPG3 and SPG10 as so-called childhood onset forms of HSP, follows a wide distribution over several decades in all subtypes of HSP that are common enough that we can deduce the phenotypic spectrum from the published families. Mode of inheritance and frequency of specific genetic subtypes are therefore the best criteria we can use to predict the genotype and guide diagnostic testing.

AUTOSOMAL-DOMINANT HSP

At least 13 different genes are known to cause autosomal-dominant HSP: ATL1 (SPG3), SPAST (SPG4), NIPA1 (SPG6), KIAA0196 (SPG8), KIF5A (SPG10), RTN2 (SPG12), HSPD1 (SPG13), BSCL2 (SPG17), REEP1 (SPG31), SLC33A1 (SPG42), REEP2 (SPG72), BICD2, and VCP. Collectively, mutations in these genes explain about 70% of autosomal-dominant cases, indicating that some autosomal HSP genes remain to be identified. Frequently, autosomal-dominant HSPs are associated with a pure or predominantly pure phenotype, but all genetic subtypes can also produce complicated phenotypes.

SPG3, Atlastin-1 (ATL1). SPG3 is caused by mutations in the gene atlastin-1 (ATL1) (19). SPG3 is typically an uncomplicated HSP form with symptoms beginning in childhood (20) but onset until the fifth decade has been reported. Durr et al studied 12 SPG3 families and found that scoliosis was present in 22% of patients, mild pes cavus in 15%, and brisk upper limb reflexes in 10%; while sensation was not impaired, and only 13% of patients reported decreased vibration sense in the ankles (21). In addition to this pure presentation of SPG3, ATL1 mutations can cause complicated forms of HSP. In one study, 17% of 36 affected individuals exhibited an axonal, predominantly motor polyneuropathy (22). Very severe childhood onset cases with pseudobulbar palsy, severe tetraparesis, motor axonal neuropathy and variably cognitive impairment, TCC, or optic atrophy have also been described (23,24).

SPG3 mutations account for up to 10% of all autosomal cases and the characteristic early age at onset should guide the genetic testing efforts. Although sensory neuropathy is a rare finding in typical SPG3, ALT1 have been found to also cause hereditary sensory neuropathy (HSAN1D), an autosomal-dominant sensory-motor neuropathy leading to severe lower limb mutilations (25).

Before its association with SPG3, ATL1 was called guanylate-binding protein 3 or GBP3 (26). GBP3 was found to interact with the hydrophobic leucine-rich CHN domain of mitogen-activated protein kinase 4 (MAPK4), a known activator of the stress-activated protein kinase/c-Jun NH2-terminal kinase (SAPK/JNK) cascade (26). Subsequent to the identification of mutations in ATL1 in SPG3, molecular studies revealed that ATL1 is an oligomeric integral membrane GTPase with two putative transmembrane domains projecting the N- and C-terminals into the cytoplasm (27). Integrity of the transmembrane domains is required for proper localization of ATL1 to the tubular endoplasmic reticulum (ER). Two additional members of the atlastin family were identified, atlastin-2 and atlastin-3, that have a similar membrane orientation and high sequence homology. Interestingly, the latter has been found to cause hereditary sensory neuropathy 1F that shares phenotypic similarities with the neuropathy caused by ATL1 mutations (28).

The atlastin family shares membrane topology and has high sequence homology with other dynamin-like GTPase, such as mitofusin, which causes axonal degeneration in peripheral nerves (27,29). Examination of atlastin in rat brain showed that it is enriched in the lamina V pyramidal neurons of the cerebral cortex, which is selectively affected in HSP (30,31). Two independent groups reported yeast two-hybrid screens using SPG4 (spastin) as bait, with subsequent identification of atlastin as a binding partner of spastin (30,31). These experiments thus linked the two most common HSP genes.

Atlastin GTPases mediate homotypic fusion of ER tubules and are therefore essential for the formation of three-way junctions of the polygonal ER network (32,33). ATL1, the only atlastin family member highly expressed in the central nervous system (CNS), also interacts with another HSP-related protein—REEP1—that is involved in membrane shaping at the tubular ER and interaction between the ER and the microtubule cytoskeleton (34). Knockdown of ATL1 in various model systems causes impairment of axon outgrowth and branching, deficits that can be rescued by the microtubule-destabilizing drug vinblastine in drosophila (35).

SPG4, Spastin (SPAST). SPG4 is caused by mutations in the SPAST gene (36). Typically, SPG4 presents with a pure phenotype. The age at onset spectrum is broad and stretches over eight decades with a mean age at onset of about 30 years. The age at onset can vary by several decades even within families. Reduced penetrance has been described (37). Neurogenic bladder disturbances and affection of the dorsal column are quite common in SPG4, present in approximately 70% and approximately 50% of patients, respectively. Rarely, complicated features are associated with SPG4; complicating features include upper limb involvement, atrophy of intrinsic hand muscles, thin corpus callosum, and—most commonly—cognitive deficits. The latter are usually mild but may be more common than usually acknowledged (38).

SPG4 is by far the most common subtype of HSP and accounts for about 50% of autosomal-dominant cases. Most mutations in SPG4 are thought to represent loss-of-function alleles. Mutations types include missense, nonsense and splice site mutations, small insertions and deletions as well as large genomic deletions. The latter mutation type, comprising about one fifth of all HSP cases, is commonly missed by most Sanger sequencing and next-generation sequencing techniques. Diagnostic testing in SPG4 should, therefore, include a method to screen for macro-deletions like multiplex ligation dependent probe amplification (MLPA) (17).

The identification of mutations in SPAST was accompanied by the immediate recognition of spastin as a member of the AAA protein family (such as SPG7, paraplegin). Other conserved protein domains in spastin suggested interaction with nuclear protein complexes (19,36). Although the molecular function of spastin was at the time unknown, unique splice site mutations were reported to be partially penetrant or “leaky,” as expressions of both mutant and wild type splice variants were observed (39). Such a phenomenon suggested that the function of spastin was highly dependent on expression levels (39). In experiments with a controlled spastin expression, it was confirmed that different levels of spastin have pleiotropic effects on neuronal morphology (40). Two alternative transcription start sites within spastin result in two spastin isoforms, of 68 and 60 kDa, while alternative splicing of both isoforms can include or skip exon 4, yielding a 64- and 55-kDa protein (41). In cultured neurons, overexpression of the long but not the shorter isoform of spastin impairs axon growth. The longer isoform is also particularly enriched in spinal cord (42). However, it is still debated as to which isoform is the predominant player in HSP (41,43).

Spastin directly interacts with several other HSP-related proteins, including atlastin 1, REEP1, and reticulon 1 (44). Cell models and molecular analysis revealed that spastin functions in a very similar way to the microtubule-severing protein, katanin (43,45,46). Overexpression of exogenous spastin in cultured cells resulted in the localizing of spastin to microtubule asters (45,46). Disease-associated mutations in the ATPase domain abrogate ATPase activity and are sufficient to ablate the microtubule-severing activity of spastin in mammalian cell models and drosophila (43,47,48). As a consequence, microtubule-dependent transport is disrupted as indicated by perinuclear clustering of mitochondria and peroxisomes (46). Spastin and spartin (SPG2) contain a conserved microtubule-interacting and trafficking domain (MIT), within which disease-causing mutations are found (49).

Silencing of spg4 in zebra fish caused widespread defects in neuronal connectivity and extensive neuronal apoptosis, with specific defects in motor-neuron outgrowth and disorganized axonal microtubule arrays (50). Drosophila models of spastin revealed irregularities in microtubules of neurons, severe reduction of the synaptic terminal area at the neuromuscular junction, and subsequent neurotransmission defects (51–53). The so-called cut and run hypothesis describes a system of mobile short microtubule segments, severed by spastin from intact microtubule lattices, that are transported, anchored, and immobilized again as they elongate microtubules in a treadmill-like process (54,55). In this model, the mobility of a severed microtubule is inversely proportional to its length providing a potential explanation for the length-dependent defects in HSP (54,55). Interestingly, in spastin-null flies, administration of the vinca alkaloid vinblastin, which inhibits microtubule assembly, diminishes phenotypes associated with loss of spastin in drosophila (52). This effect was confirmed in cortical neurons derived of a spastin knockout mouse model as well as from human-induced pluripotent stem cells; in these cells, treatment with vinblastine reduced axonal swellings (56,57). Although this observation suggests a potential therapeutic mechanism for HSP, it directly conflicts with important published reports of vincristine administration exacerbating axonal neuropathies in peripheral nerves (58–63).

Mice deficient of spastin are viable and show no motor deficiencies until 2 years of age (64). However, examination of corticospinal neurons demonstrated an age-dependent progressive increase in focal axonal swellings, accumulations of organelles and intermediate filaments within the regions where stable and dynamic microtubule lattices meet (64). Conversely, spastin overexpression resulted in more microtubule branching and shorter severed microtubule segments (65).

SPG10, Kinesin 5A (KIF5A). SPG10 is caused by mutations in the gene KIF5A (10). Only few families with KIF5A mutations have been reported to date (66,67). Recently, Goizet et al have reported several new KIF5A families with additional clinical features including peripheral neuropathy, severe upper limb amyotrophy (Silver syndrome-like), mental impairment, parkinsonism, deafness, and retinitis pigmentosa (68). They concluded that KIF5A accounts for 10% of complicated HSP forms in their sample, making KIF5A the most important form of complicated autosomal-dominant HSP (68). If confirmed, this would be important for future decisions on medical genetic testing. The age of onset in uncomplicated SPG10 ranges from 8 to 40 years (67).

Kinesin motor proteins are microtubule-dependent trafficking machines. Axonal transport genes are very good candidates for spastic paraplegia as some of the longest axons in the human body are involved in the disease. Studies found that KIF5A−/− mice were neonatal lethal (69). In a conditional knockout model, the postnatal loss of KIF5A caused sensory neuron degeneration and seizures, with no morphological abnormalities in the spinal cord (69). Fast axonal transport appeared to be normal in the absence of KIF5A (69). However, slow axonal transport, and more specifically, the transport of neurofilaments, was profoundly disrupted in KIF5A−/− mice (69). Neurofilaments are thought to play an important role in the increase in caliber of large axons during development, as heavy neurofilament (NF-H)−/− mice exhibit a severe axonal hypotrophy (70). Mutations in neurofilaments are also associated with forms of axonal degeneration in the peripheral nerve, suggesting that these disease pathways converge (71).

Two models of how mutations in KIF5A cause HSP have been suggested. The first model proposes that mutant KIF5A are slower motor proteins, with slower transport rates having retarding effects on axonal transport (72). Alternatively, mutations may abrogate microtubule binding, allowing free mutant KIF5A to compete with microtubule-bound wild type KIF5A for cargo (72). Using in vitro homodimeric motor gliding assays, it was determined that both these models may be true depending on the exact mutation. The K253N and R280S mutations diminished the fraction of transported cargo, suggesting they competed with wild type KIF5A for cargo, statistically sequestering it from transport (72). The N256S mutant still binds to microtubules but slows down the transport of cargo (72).

SPG17, Seipin (BSCL2). SPG17 or Silver syndrome is caused by mutations in BSCL2 (73). Silver et al reported two families with spastic paraplegia and amyotrophy of the hands inherited in an autosomal-dominant pattern (74). But mutations in the BSCL2 gene are associated with a broader array of diseases, including the recessive Berardinelli–Seip congenital lipodystrophy type 2 (BSCL2 or CGL2), dHMN type V (dHMN-V), and variants of axonal Charcot–Marie–Tooth disease (75–77). Additional symptoms of SPG17 include distal motor neuropathy and pes cavus (78–81). The age of onset varies from childhood to the fourth decade of life (74). Few families have been reported so far with SPG17 and BSCL2 mutations.

The involvement of upper motor neurons, lower motor neurons, and peripheral nerves has led Ita and Suzuki to propose the term “seipinopathies” to describe this collection of related diseases (82). Seipin has no closely related homolog genes. The highly conserved amino acids 1 to 280 encode a leucine zipper domain motif. The predicted secondary structure of the zipper domain is similar to sterol regulatory element binding proteins (SREBPs), which regulate cholesterol and lipid metabolism (82). Seipin also contains two transmembrane domains and has been shown in multiple studies to be an ER membrane protein (73,82,83). Two seipin transcripts (1.8 and 2.4 kb) are ubiquitously expressed, while the 2.0-kb transcript is expressed at high levels in the CNS and testis (73). Molecular studies have demonstrated that seipinopathies associated with mutations in BSCL2 are diseases resulting from problems with dysfunctional protein folding within the ER (84). Eventually, misfolded mutant proteins undergo a conformational change that leads to aggregation, a phenomenon common to other neurodegenerative diseases of the CNS (84,85). The N88S and S90L BSCL2 mutations, which are associated with “seipinopathic” motor-neuron diseases, alter the N-glycosylation site of seipin, and cause accumulation of unfolded protein in the ER (84). It was demonstrated that the expression of mutant seipin in cultured cells activates UPR stress and induces ER stress-mediated apoptosis (85).

Recent reports suggest a function for seipin in the interface between the ER and lipid droplets, or adiposomes. Deletion of yeast seipin results in irregular adiposome formation (86). Fibroblasts from patients with congenital lipodystrophy type 2 also form irregular lipid droplets (86). Other reports have revealed a role of seipin in the adipogenesis, with obvious relevance to the lack of adipose tissue observed in Berardinelli–Seip congenital lipodystrophy type 2 (87).

SPG31, Receptor expression enhancing protein 1 (REEP1). SPG31 is caused by mutations in REEP1 (9). An increasing number of families have recently been reported with SPG31 and it is suggested that this HSP form accounts for up to 8.2% of all autosomal-dominant cases (8,88–90). Most cases are clinically uncomplicated but additional symptoms may include scoliosis, peripheral neuropathy, spastic tetraparesis, and bulbar dysfunction (8,89). The age of onset appears to be bi-modal with a subgroup of patients showing early onset in childhood and a second peak in the third and fourth decades of life (8).

REEP1 was first reported as a protein, which associated with odorant receptor proteins and enhanced the receptor response to odorant ligands (91). Along with its identification as the gene mapping to SPG31, REEP1 was shown to be ubiquitously expressed and localized to mitochondria in a number of cell types (9). Interestingly, mutation analysis has revealed mutations in highly conserved predicted micro-RNA binding sites (8). REEP1 also contains a TB2/DP1/HVA22 domain common to heat-shock proteins. Recent work in our laboratory suggests that REEP1 may have alternative splice isoforms, which are expressed and have contrasting subcellular localization patterns (unpublished data). Along with HSP60 (SPG13) and paraplegin (SPG7), mutations in REEP1 point to the importance of mitochondria in the pathogenesis of HSP.

AUTOSOMAL RECESSIVE HSP

SPG5, Cytochrome P450, family 7, subfamily B, polypeptide 1 (CYP7B1). SPG5 is caused by mutations in CYP7B1 (92). All patients reported had a pure form of motor-neuron degeneration with progressive spastic paraplegia and variable bladder and sensory impairment (92). The findings indicated a pivotal role of altered cholesterol metabolism in the pathogenesis of motor-neuron degenerative disease. Recent genetic screens have identified additional mutations in CYP7B1 with some families expressing symptoms of the complicated HSP spectrum, including optic atrophy, cerebellar abnormalities, and white matter lesions (93,94). However, other reports suggested screening for mutations in CYP7B1 in sporadic and familial cases are of low diagnostic yield (94).

The majority of work on CYP7B1 relates to its function as the first catabolic enzyme in the degradation of cholesterol to bile acids in extrahepatic tissue. CYP7B1 is also thought to function in the pathogenic cholesterol pathways involved in atherosclerosis and act as a metabolic enzyme of neurosteroids and sex hormones. How cholesterol metabolism is related to the pathogenesis of HSP remains an intriguing and unanswered question.

SPG7, Paraplegin

Only gold members can continue reading. Log In or Register to continue