Clinical manifestations

Presentation and course

Harding described two clinical classifications of hereditary spastic paraplegia: (1) pure hereditary spastic paraplegia; and (2) complicated hereditary spastic paraplegia with other neurologic abnormalities (126).

Pure hereditary spastic paraplegia. Pure (or uncomplicated) hereditary spastic paraplegia is characterized by progressive spasticity of the lower limbs, hyperreflexia, and extensor plantar responses and urinary urgency, without other significant neurologic abnormalities. The age of onset can vary from infancy to the eighth decade (127), and a majority of the patients give history of first symptoms in the period between the second and fourth decades.

Initial presentation of the disease begins with difficulty in walking and abnormal gait. Motor milestones, especially the time of walking, may be delayed in childhood-onset forms. Stiffness of the legs, easily worn-out shoes, toe-walking, and paresthesia below the knees are the other complaints of these patients.

Obligatory criteria suggested for the diagnosis of pure hereditary spastic paraplegia include positive family history, progressive gait disturbance, spasticity of the lower limbs, and hyperreflexia of the lower limbs (192). Gait disturbance progresses slowly without exacerbations, remissions, or worsening (97). In addition to hypertonicity causing spastic circumduction and toe walking, there is mild or absent muscle weakness. If muscle weakness is present, it usually involves iliopsoas, tibialis anterior, and, to a lesser extent, the hamstring muscles. Muscle wasting can occur, but it is uncommon; when present, it is found in distal muscles of the lower limbs, small muscles of the foot, and tibialis anterior, usually in patients who have had the disease for more than 10 years (127; 77).

Deep tendon reflexes may be brisk in the upper extremities and are pathologically increased in the lower extremities. Sometimes ankle jerk may be absent (127). Ankle clonus and extensor plantar responses occur in all patients. Pes cavus is also a common finding of the disease. Fink and colleagues describe pes cavus as usually occurring in older affected patients, whereas Harding found pes cavus in one third of patients (127; 97). Flat foot, rather than pes cavus, has been identified in a Chinese family with pure type of autosomal dominant hereditary spastic paraplegia and SPG3A mutation (165).

Urinary sphincter disturbances, such as urgency, frequency, and hesitancy occur in most patients. Although rare, anal sphincter involvement and sexual dysfunction can also be seen. Scheltens and colleagues described a Dutch family with urinary and anal sphincter disturbances (273).

Sensory impairment is evident in 10% to 65% of cases and is common in long-standing disease; incidence increases with disease duration (127; 192). Vibration sense is often affected in the lower extremities. Joint position sense also may be impaired, but cutaneous sensory loss is uncommon (127; 77; 97).

Restless legs syndrome of moderate and severe grade is also associated with hereditary spastic paraplegia. It has been found more frequently (20.5%) than the general population (11%) (292).

Upper limb involvement is not common. In addition to mild hyperreflexia, there may be terminal dysmetria or clumsiness without other cerebellar signs due to pyramidal tract involvement. Muscle strength and tone are normal in the upper extremities. Corticobulbar tract and cranial nerves are also preserved.

If paresis is more marked than spasticity, ataxia, prominent amyotrophy, or early-onset amyotrophy during the course of the disease, prominent involvement of upper limbs, asymmetry, retinal pigmentation, cranial nerve involvement, or extrapyramidal signs, diagnosis of hereditary spastic paraplegia should be cautious, and other causes within the differential diagnosis should be reconsidered (192).

Complicated hereditary spastic paraplegia. In addition to spastic paraparesis, patients with complicated hereditary spastic paraplegia have other associated neurologic abnormalities. These associated clinical findings include optic atrophy, retinopathy, extrapyramidal signs, amyotrophy, dementia, mental retardation, ataxia, nystagmus, dysarthria, deafness, epilepsy, ichthyosis, peripheral neuropathy, and neuropsychiatric symptoms (194). Syndromic forms are also included in the complicated hereditary spastic paraplegia (126; 192). An example of complicated hereditary spastic paraplegia is SPG47, where the clinical features begin in the neonatal period with hypotonia that progresses to spasticity, early-onset developmental delay with prominent motor delay and severely impaired or absent speech development, stereotypic laughter, seizures, including frequent febrile seizures, thinning of the corpus callosum, and delayed myelination or white matter loss (79).

A pediatric case series described the expanding clinical spectrum of childhood-onset hereditary spastic paraplegia. Seventy percent of subjects with childhood-onset hereditary spastic paraplegia had comorbid neurocognitive deficits, including developmental delay, autism, epilepsy, attention-deficit/hyperactivity disorder, polyneuropathy, and rare genetic etiologies (231).

Due to the variability in presentation, some genetic types associated with the uncomplicated type of spastic paraplegia may have some symptoms or signs of complicated spastic paraplegia. There may also be differences in symptoms amongst individuals with the same genetic type of hereditary spastic paraplegia (96).

Prognosis and complications

There is significant variability amongst individuals with hereditary spastic paraplegia with regard to disease course. Disease progression of hereditary spastic paraplegia overall is usually slow, although a later onset is associated with earlier loss of independent walking (127). Disease severity is related to genotype. For example, SPG11 is associated with a higher disease severity, as determined by the Spastic Paraparesis Rating Scale, relative to other types (281). The average rate of disease progression is observed to be slower for patients with SPG3A than SPG4 but did not differ when comparing individuals with onset before 20 years of age (174). In a large cohort study, Schule and colleagues noted that disease severity is also related to age of onset, disease duration, cognitive impairment, dysphagia, dysarthria, and extrapyramidal peripheral motor involvement (281). Overall, they found the median disease duration to loss of ambulation to be 22 years and approximately one quarter of patients will require a wheelchair regularly after 37 years of age.

In general, hereditary spastic paraplegia does not limit life expectancy. In her discussion, Harding cites the findings of Bell and Carmichael, who found the mean age of death to be 57.5 ± 3.02 years, with the mean duration of the disease as 27.21 ± 13.69 years; also, they reported one patient who had the disease for more than 80 years (127).

Clinical vignette

The patient, an 18-year-old man with gait difficulty, initially presented with symptoms of tip-toe gait at 3.5 years of age. Perinatal history was non-contributory: he walked at 11 months; other milestones were also normal. His parents were first cousins.

The patient showed slow but gradual worsening over 14 years of follow-up. He developed mild speech difficulty after 13 years of age. Examination at 18 years of age showed increased deep tendon reflexes in the legs.

Plantars were extensor, and gait was typically spastic. He exhibited normal muscle power and intellect. Mild dysarthria and dysmetria were evident. He was performing regular physiotherapy and taking baclofen 60 mg per day.

Biological basis

Etiology and pathogenesis

Hereditary spastic paraplegia occurs due to the dysfunction of the upper motor neurons of the corticospinal tracts. More than 80 causative genetic mutations have been identified. The genes encode proteins involved in normal neuronal function. Disease mechanisms implicated include those relating to vesicular trafficking and organelle shaping, axonal transport, axonal development and myelination, mitochondrial functions, and complex lipid metabolism (156).

The effects of unknown genetic factors or environment on the development or progression of the disease are still not known.

Pathology. The major pathological finding of hereditary spastic paraplegia is the degeneration of axons of the pyramidal tract and dorsal column pathways of the spinal cord. The axons of the corticospinal tract are the longest in the nervous system, and degeneration of the crossed pyramidal tracts increases toward their terminal portions, maximally at the lumbar spinal cord; it may be considered that a “dying back” axonopathy occurs toward the cell body. Demyelination and gliosis can occur with axonal loss. Lateral corticospinal axonal degradation is predominantly observed in hereditary spastic paraplegia postmortem studies with a higher depletion rate in the spinal cord’s thoracic zone distal end and cervical zone mainly due to axons degeneration in fasciculus gracilis fibres to demyelinating (95). Sometimes, degeneration can extend to the rostrum, internal capsule, peduncles of the cerebellum, pons, and medullary zone, with declined Betz cells concentration (pyramidal neurons). The central nervous system long axons are hotspots and the first site of hereditary spastic paraplegia axonopathy (268).

Spinocerebellar tract involvement is seen in 50% of cases. Peripheral nerves, dorsal roots, and dorsal ganglia are generally normal. However, peripheral neuropathy, a common complex hereditary spastic paraplegia symptom, is caused by the depletion of other neurons. Due to neuropathy in specific locations, shorter neurons in the basal ganglia, cerebellum, anterior horn cells, and Clarke column cause hereditary spastic paraplegia features (329). Neuronal region genetic mutations cause deformities predominantly in the myelin layer, and other studies revealed cerebellar atrophy and CNS myelination, along with corticospinal axonal degeneration and developmental disorders, smaller spinal cord diameter, and thin corpus callosum as classical developmental abnormality signs in hereditary spastic paraplegias (209).

Anterior horn cells are also usually normal, but sometimes loss of anterior horn cells is also reported (283; 18; 127; 128; 192). Knowledge about the molecular and genetic basis of pathogenic mechanisms of hereditary spastic paraplegia continues to increase.

Genetics. Hereditary spastic paraplegias are genetically determined. Autosomal dominant, autosomal recessive, X-linked, and mitochondrial forms of inheritance have been described (322; 98). De novo mutations should be considered in patients with childhood-onset, nonprogressive, spastic diplegia with no previous family history of hereditary spastic paraplegia (248). Infantile onset hereditary spastic paraplegia may also be caused by mutations in multiple genes (27).

Pertinent genotype-phenotype correlations of hereditary spastic paraplegia described in the literature are summarized below.

SPG4 is the most frequent form of autosomal dominant spastic paraplegia and accounts for 45% of all autosomal dominant kindreds (97). It encodes a protein named spastin. There are more than 150 SPG4 mutations, including point mutations, small insertions, and deletions; most reported families have a unique mutation (66; 179). Usually, age of onset is in adulthood and genetic anticipation can occur with SPAST mutations, but a novel de novo mutation has been found to cause an infantile-onset disease in a family in which most members were diagnosed with diplegic cerebral palsy (27). Chelban and colleagues found a significant percentage of cases (16%) presenting before the age of 10 in their cohort study (48). Some mutations like multi-exonic SPG4 duplication lead to a more severe and early onset disease in males (203), and another mutation in an Italian family caused a reduced penetrance of disease (09). Two polymorphisms were defined in SPG4 gene, S44L and P45Q, causing a phenotype modifier effect associated with early onset of disease (302). Newton and colleagues reported an epistatic interaction by which deletions of SPAST that also involved the contiguous gene, DYP30, resulted in earlier onset of symptoms in hereditary spastic paraplegia (213).

Although SPG4 is largely an uncomplicated form of hereditary spastic paraplegia, in some families with mental retardation, auditory impairment, thin corpus callosum, and cerebellar atrophy, a new mutation of SPG4 has been detected (220). In a cohort study of SPAST patients, 74% were found to have a pure phenotype and 26% had a complicated form with associated cerebellar syndrome, peripheral neuropathy, learning disability, memory problems, and psychiatric symptoms (48). Families with dementia, dysplastic corpus callosum, and a family (with four generations) with cerebellar ataxia, dysarthria, unipolar depression, epilepsy, migraine, and cognitive impairment were also found to have mutations in SPG4 gene (193; 214; 04). However, significant ophthalmological complications usually do not complicate SPG4 patients (121). A cross-sectional study observed that dysarthria in SPG4 is frequent and mild and does not evolve in conjunction with more advanced motor diseases (144). In SPG4 patients with normal MRI, magnetic resonance spectroscopy showed metabolic abnormalities in the motor cortex of those with age-related verbal learning and memory reduction (82). Some alterations in resting-state functional magnetic resonance imaging in patients with SPG4 were noted to correlate with disease severity (166). Preliminary studies demonstrate that neurofilament light chain in cerebrospinal fluid is a promising biomarker that may indicate disease activity in prodromal SPG4 but needs further evaluation (251).

By phenotypic expansion of SPG4, it is suggested that not only the length of the axon but also modifying factors play a role in the pathogenesis of SPG4 (92). Proukakis and colleagues have found the frequency of SPG4 mutations in pure autosomal dominant disease to be 33.3% (244). With the introduction of multiplex ligation-dependent probe amplification, families with strong linkages to SPG4 but negative for SPAST mutations were screened, and aberrant profiles representing partial deletions were identified. The frequency of these deletions was estimated at 11%, raising the overall involvement of SPAST in autosomal dominant hereditary spastic paraplegia to over 50% (16). Males were reported more than females, suggesting that penetrance or severity of disease caused by SPAST mutations may be sex-dependent (245). In Australian patients, frequency of SPAST gene mutations with combined molecular techniques was 53% in unrelated hereditary spastic paraplegia patients, recording a very high frequency (316). In Brazil, mutation screening in a cohort of 55 patients with autosomal dominant hereditary spastic paraplegia revealed 35% of frequency of mutations (100).

Spastin is a member of group of proteins known as AAA proteins (ATPases associated with diverse cellular activities) (131). These proteins have several functions in protein degradation, trafficking, and organelle biogenesis. It is shown that spastin works in microtubule dynamics. Spastin interacts transiently with microtubules, and its microtubule-binding ability comes from the AAA homologous region that provides the structural framework for microtubule interaction (84). It was shown that endogenous spastin is located at the centrosome and mitotic spindle poles, and it persists throughout the cell cycle supporting its role in microtubule dynamics (302). Interaction of ATP-bound spastin with microtubules stimulates the hydrolysis of ATP. After hydrolysis it dissociates from the microtubules and exchanges ADP for ATP, regenerating energy for microtubule severing (328). Wild-type spastin promotes microtubule disassembly in transfected cells, and this property is decreased with mutant spastin, so spastin is described as microtubule severing protein. This property is closely related with katanin in sequence and function. Katanin is another microtubule severing protein that regulates axon growth and generates short microtubules that can be efficiently transported and remodeled. Studies have shown that, like katanin, spastin promotes the formation of microtubule networks that are essential for axon outgrowth, and reduced spastin function severely impairs motor axon outgrowth (334). It was demonstrated that spastin is necessary for regulation of intracellular calcium homeostasis, through its effects at the interface of endoplasmic reticulum and microtubule function. Thus, calcium dysregulation might be pivotal to a class of mechanisms that disrupt neuronal function in hereditary spastic paraplegia (258).

Altered microtubule dynamics may also result in abnormal mitochondrial distribution within the axon, which may lead to defective oxidative phosphorylation (92). By discovery of endosomal proteins as binding partners for spastin, it has been suggested that, besides microtubule regulation function, spastin may also be involved in membrane trafficking events (269). Spastin has been shown to affect axonal transport. Mutant spastin was found to perturb both anterograde and retrograde transport in mice (151). In a patient-derived stem cell model of hereditary spastic paraplegia with SPAST mutations, compared to control cells it showed less spastin and alfa tubulin, but more stathmin, a microtubule destabilizing enzyme to compensate for less spastin. Patient cells showed altered distribution of peroxisomes and mitochondria. Because of greater amount of stathmin, cells also showed reduced microtubule stabilization and altered organelle trafficking (02). Microtubule binding drugs vinblastine and paclitaxel increased acetylated alpha tubulin levels, indicating a new target of therapy (02).

SPG3A is another autosomal dominant hereditary spastic paraplegia that results from mutations in ATL1 (atlastin) and constitutes 10% of autosomal hereditary spastic paraplegias (97). It is the most common type in patients with onset before 10 years of age (210). Novel mutations in this gene showed that it can lead to both an early-onset and an adult-onset disease. Dalpozzo and colleagues reported a novel mutation in the atlastin gene that caused spastic paraparesis in infancy, resembling cerebral palsy, with upper limb amyotrophy (62). De novo occurrence of atlastin mutation has been reported in very early-onset disease, and investigation of de novo mutations might be considered in patients with cerebral palsy for which no other cause is identified (248). Fusco and colleagues have reported a very severe phenotype beginning at 3 months of age with severe hypotonia in a mutation in exon 10 (104). 18-Fluorodeoxyglucose positron emission tomography (FDG-PET) studies of two Japanese sisters showed cerebral hypometabolism in the frontal cortex associated with cognitive impairment (305).

SPG4 and SPG3A mutations together cause 50% of dominantly inherited hereditary spastic paraplegias. Genetic penetrance in these diseases is high, so prenatal testing is commercially available for these two hereditary spastic paraplegias (131; 91). Atlastin has been discovered in a large family with hereditary sensory neuropathy type 1 (HSN1). HSN1 and SPG3A are allelic disorders (120). Atlastin is a member of the dynamin family of large GTPases that play essential roles in vesicle trafficking events important for neurotransmission and action of neurotrophic factors. Also, atlastin was found to increase dendritic growth in mouse cerebral cortex via GTPase activity (106). Dynamins take place in vesicle formation and liberation in both endocytic and secretory pathways and have been implicated in the maintenance and distribution of the mitochondria (343; 93). Atlastin has been found in high concentration in vesicular structures of growth cones, implicating a possible role in axonal development other than trafficking (344). Three forms of atlastin have been discovered; atlastin-1 is mostly found in the brain whereas -2 and -3 are localized to endoplasmic reticulum and found mostly other tissues. In contrast to previous suggestions about their roles in vesicular trafficking, they are thought to function in both endoplasmic reticulum and Golgi morphogenesis. Morphologically abnormal Golgi and endoplasmic reticulum are reported to interfere with proper distribution or polarity of long corticospinal tract motor neurons (256). Atlastin interacts with many proteins in cellular processing and transport, proteins with roles in mRNA binding, and metabolism and mitochondrial proteins. The ubiquitin-selective AAA-ATPase valosin-containing protein (VCP) is one of the interactors of atlastin and also acts as a genetic modifier of atlastin (225). Atlastin and NIPA1 are both found to be related to bone morphogenic protein (BMP)-linked signalling pathways (342).

SPG6 is caused by NIPA1 gene mutations. Although allelic heterogeneity of the mutations is very limited, recurrent de novo c.316G >A mutation in the NIPA1 hotspot has been reported (134). The disease is an adolescent- or adult-onset form of a typical slowly progressive uncomplicated hereditary spastic paraplegia. However, idiopathic generalized epilepsy and peripheral neuropathy have been described in families with SPG6 (76; 303). The function of this gene is not yet known, but it is widely expressed in the central nervous system, and it is suggested that NIPA1 encodes a membrane protein. Patients totally lacking this gene (nonimprinted in Prader-Willi and Angelman loci) do not develop hereditary spastic paraplegia, so NIPA1 is said to act through “dominant negative” gain of function (94; 247). In a Chinese family with SPG6 mutations, severe spinal cord atrophy, especially in the cervical and upper thoracic cord, was found as a specific symptom for SPG6 (170). It was discovered that atlastin-1 and NIPA1 are direct partners with coexpression to support axonal maintenance. Mutations in one of the proteins caused protein sequestration either in Golgi complex or endoplasmic reticulum. Also, in cultured rat cortical neurons axonal and dendritic sprouting was reduced, showing the common biochemical pathway that atlastin and NIPA1 shared (31).

SPG8 is a severe, pure autosomal dominant hereditary spastic paraplegia on chromosome 8q24.13. Strumpellin protein of KIAA0196 gene is mutated in SPG8. The mutation p.(Gly69Ala), detected in a Dutch family, had less severe clinical course than previous families described (67). The exact function of this protein is still unknown, but motor-axon-outgrowth impairment is suggested (311). Strumpellin mutations are loss-of-function mutations and cause a reduction in axonal outgrowth in human neuroblastoma cells. Strumpellin was identified as a novel valosin-containing, protein-binding partner (51). Valosin-containing protein mutations cause autosomal dominant inclusion body myopathy associated with Paget disease of bone and frontotemporal dementia. Strumpellin has been shown to be a component of WASH complex, a multimeric protein complex that acts at the interface between actin regulation and endosomal membrane dynamics. However, mutations do not affect the assembly of the WASH complex (102).

SPG9 is a complicated dominant type of hereditary spastic paraplegia with bilateral cataracts and gastroesophageal reflux with persistent vomiting amyotrophy (232). The ALDH18A1 gene is located at 10q24.1 and encodes delta-1-pyrroline-5-carboxylate synthetase (P5CS), a mitochondrial bifunctional enzyme that catalyzes the first two steps in de novo biosynthesis of proline, ornithine, citrulline, and arginine. ALDH18A1-related disorders have been classified into groups, including autosomal dominant and recessive hereditary spastic paraplegia (SPG9A and SPG9B), respectively. Early-onset tremor, prior to spasticity, is unique to SPG9B (147).

A mutation in a kinesin heavy chain gene, KIF5A, results in chromosome 12q-linked autosomal dominant hereditary spastic paraplegia, SPG10 (253). It is thought to be a rare form of disease at least in Europe; only 3% of patients were found to have SPG10 in European autosomal dominant hereditary spastic paraplegia families in one report (279). In another report by Goizet in 2009, SPG10 mutations were found in 10% of patients with complicated autosomal dominant disease. Patients had variable complex features, such as peripheral neuropathy, severe limb amyotrophy (Silver syndrome-like), mental impairment, parkinsonism, and retinitis pigmentosa (113). A single KIF5A mutation was found in an Italian patient with Charcot-Marie-Tooth disease (CMT2) (58). A family from France with a novel missense mutation presented with dysautonomia, severe axonal neuropathy, symmetric cerebral demyelination, and spinal cord atrophy on MRI in one of the siblings. Skin biopsy showed a defect in the number of vesicles and synaptophysin density at the presynaptic membrane by electron microscopy, suggesting a disease resembling Charcot-Marie-Tooth disease (53). Mutations can cause both early- and adult-onset disease (26). Kinesin is part of a protein motor complex that functions in the microtubule transport system in both anterograde and retrograde direction in both slow and fast transport of cytoplasmic proteins. Kinesin mutation studies support a defect in microtubule-mediated trafficking, which leads to axonal degeneration (60; 152). There are three isoforms of kinesin. Kinesin-1 is responsible for the transport of cytoplasmic dynein and mediates retrograde cargo transport. Neurofilaments and mitochondria are the important cargo for KIF5A in neurons. Mutations can cause abnormal localization and transport of mitochondria. Frequency and velocity of neurofilament movements in both directions are also affected by mutations disturbing the neuronal homeostasis (152).

SPG12 is a pure autosomal dominant hereditary spastic paraplegia subtype. The causative gene is on 19q13 and the gene product is reticulon. This interacts with spastin and is involved in ER shaping (204).

Heat shock protein 60 (Hsp60), also known as “chaperonin 60,” is encoded by HSPD1 gene. It is a mitochondrial chaperone and has been identified in uncomplicated autosomal dominant form SPG13. Chaperones are the phenotype modulators. Hsp60 is part of a mitochondrial complex that regulates correct protein folding in mitochondria (125; 252). Hsp60 mutations cause increased oxidative stress in neuronal tissue. Manganese superoxide dismutase (MnSOD), a key antioxidant enzyme, interacts with Hsp60 and is a substrate of Hsp60 folding machinery. It is proposed that due to impaired folding of MnSOD, reaching the native state is impaired (181).

SPG13 is a rare cause of hereditary spastic paraplegia, but information shows the importance of this mitochondrial chaperone protein. A HSP60 polymorphism p.[Gly563Ala] was found in a spastin mutation-positive family suggesting that this polymorphism may result in earlier onset of phenotype (139). Also p.[Ser44Leu] polymorphism of spastin has been shown to have a phenotype modifier effect. Interactions between different genes of hereditary spastic paraplegia open new gateways to the pathogenetic mechanisms.

An allelic disorder of SPG13 caused by a missense mutation in HSP60, termed MitCHAP60, was defined in an Israeli Bedouin kindred with hypomyelinating leukodystrophy. The patient was negative for PLP1 and connexin 47 mutations, and inheritance was autosomal recessive (180).

Silver syndrome (SPG17) is an autosomal dominant hereditary spastic paraplegia associated with amyotrophy of hand muscles and weakness of small muscles. Although there is genetic heterogeneity in this locus, BSCL2 gene, seipin mutations were shown both in Silver syndrome and distal hereditary motor neuropathy type V (333). It is the gene of Berardinelli-Seip congenital lipodystrophy Type 2. Two heterozygous missense mutations resulting in amino acid substitutions N88S and S90L were shown in Silver syndrome and distal hereditary motor neuropathy type V. It is an integral membrane protein of the endoplasmic reticulum, and mutations cause aggregate formation leading to neurodegeneration (333). Seipin is expressed widely in the central nervous system, spinal cord, pituitary gland, and testis. Mutations result in proteins that are improperly folded and cause accumulation of some kind of inclusion bodies in endoplasmic reticulum (leading to activation of unfolded protein response pathway) and induce endoplasmic reticulum stress-mediated cell death. Because of a large group of both upper and lower motor neuron diseases caused by N88S and S90L mutations associated with endoplasmic reticulum stress, Ito and colleagues suggested seipin-related diseases be collectively called “seipinopathies’’ (142; 143).

Investigation of two Italian families with characteristics of Silver syndrome resulted in a novel locus at 4p16-p15 (SPG38) for Silver syndrome and a mutation of SPG4 (221). Among the 15 members of the family with a mutation in SPG4, all presented with cardinal features of Silver syndrome, and 14 presented with temporal lobe epilepsy and cognitive dysfunctions; this was the first report of a family with temporal lobe epilepsy in hereditary spastic paraplegia. Silver syndrome also shows genetic heterogeneity like SPG4, and the relation with SPG4 demonstrates that hereditary spastic paraplegia genes might interact with each other.

SPG29 is a complicated autosomal dominant form associated with sensorineural hearing impairment and vomiting due to hiatal hernia. Most of the patients had hyperbilirubinemia without kernicterus in the neonatal period (219).

The causative protein in SPG31 is receptor expression-enhancing protein 1 (REEP1), which is related with odorant receptors in mouse tissue and promotes their expression in various brain tissues. It is located in mitochondria. Hereditary spastic paraplegia caused by REEP1 mutations is suggested to be the third most common type after spastin and atlastin mutations (348). In a study composed of 535 hereditary spastic paraplegia patients, overall mutation rate for REEP1 mutations was 3%. When the study group was narrowed to only pure hereditary spastic paraplegia, the mutation rate was 8%. So, it is suggested that SPG31 is a relatively frequent type, and genetic testing should be done for REEP1 as well (17; 191; 65). Patients showed pure-type hereditary spastic paraplegia with bimodal pattern of onset of age--either before 20 years of age or after 30. A 5-year-old patient with slowly progressive disease has been reported with a multiexonic deletion in REEP1 gene (13). With the availability of genetic testing, complicated phenotypes are also described. Peripheral neuropathy, spastic tetraparesis, and pseudobulbar palsy have been found in two patients with REEP1 mutations (138). In a Chinese family with a novel mutation in the REEP1 gene, clinical features and electrophysiological studies showed pure phenotype, but MRI revealed atrophy of the thoracic spinal cord (169). A closely related gene, REEP2, is mutated in SPG72. One study found that inhibition of endoplasmic reticulum stress improves progressive motor deficits in a REEP1-null mouse model of hereditary spastic paraplegia (325). Another study found that naringenin, a flavonoid that possesses strong antioxidant and neuroprotective activity, can rescue the cellular phenotypes, the lifespan, and locomotor disability associated with ReepA loss of function (211).

SPG33 is another pure type of hereditary spastic paraplegia in which the causative mutation is in zinc finger FYVE domain-containing protein 27/protrudin (ZFYVE27). It is an intracellular protein with close interaction with spastin through its microtubule interacting and trafficking motif. Mutations severely affect the interaction with spastin (183). Together with spastin, protrudin takes a role in neurite outgrowth (339).

SPG36 is on 12q23-24 and is complicated by peripheral neuropathy (278).

SPG37 is a “pure” autosomal dominant hereditary spastic paraplegia on chromosome 8p21.1-q13.3 (122).

SPG41 is a pure form described in a Chinese family and linked to 11p14.1-p11.2 (341).

SPG42 is caused by a mutation in the gene for acetyl-CoA transporter (SLC33A1), which is on 3q24-q26 (167). It is suggested that this is a loss-of-function mutation important in the outgrowth and maintenance of the motor neuron axons.

SPG72 is secondary to mutations in the REEP2 gene. Esteves and colleagues reported three REEP2 mutations amongst two families resulting in a childhood onset pure form of spastic paraplegia believed to be secondary to impact on the cellular binding of REEP2 (85). A novel de novo missense variant has been described, emphasizing the importance of considering sporadic cases of dominant hereditary spastic paraplegia (260).

Rinaldi and colleagues reported on an Italian family with a heterozygous mutation of CPT1C resulting in a pure spastic paraplegia, which was postulated to be the result of altered lipid-mediated signal transduction (255). This form has been designated SPG73. Cooper and colleagues described a dominantly inherited heterozygous variant in ATAD3A in a family with hereditary spastic paraplegia (54). Novel genetic causes of hereditary spastic paraplegia are continuously being discovered. Some examples include novel variants in the ZFYVE26 gene, VPS13D gene, ATP1A1 gene, a recurrent 2‐base pair deletion in the UBAP1 gene and biallelic truncating variants in the RNF170 gene (227; 34; 119; 157; 299; 71). A mutation in the LYST gene has been found to be causative of isolated adult‑onset pure spastic paraplegia (182). Mutations in SPTAN are associated with neurodevelopmental disorders and epilepsy but can also be a genetic cause of pure or complex hereditary spastic paraplegia (206).

Table 1. Autosomal Dominant Hereditary Spastic Paraplegias

Autosomal recessive spastic paraplegias. SPG5 was the first autosomal recessive locus mapped in 1994 in a group of Tunisian families; it was designated as SPG5B. Muscle biopsies of a family linked to 8q11.1-q21.2 did not show oxidative phosphorylation system defects, suggesting that gene probably was not coding for a mitochondrial protein (330). Locus was narrowed down to 8q12.3, and a mutation in cytochrome P450-7B1 (CYP7B1) was found to cause SPG5A. CYP7B1 is involved in an alternative pathway for cholesterol degradation in the liver, and because cholesterol cannot cross the brain-blood barrier and is produced locally in brain, CYP7B1 provides the primary means of modification of dehydroepiandrosterone neurosteroids in the brain (309). Mutations in this alternative pathway might cause neurosteroid toxicity because of increased oxysterol levels or because impaired DHEA-related neurosteroids lose neuroprotection function. This interesting finding brings insight into hereditary spastic paraplegia because defective lipid metabolism might be one of the underlying pathogenic causes and might identify a new therapeutic strategy. Mutations in CYP7B1 are found in patients with both pure and complex clinical features (112). The frequency of mutations in autosomal recessive families was 7.3% and 3.3% in sporadic pure spastic paraplegias (112). Age of onset ranges from early childhood to the fifth decade. Cerebellar signs, distal amyotrophy, and white matter hyperintensities are common in patients with complex phenotype (24; 59; 112). MRI investigations have showed variable lesions like diffuse cerebral atrophy, punctate hyperintensities in white matter and cerebellum, and enlargement of the fourth ventricle. One patient had autoimmune hepatitis, and it is not clear if mutation pathogenesis has a role in hepatic disorder (112). Magnetic resonance spectroscopy (MRS) showed elevated mI/Cr ratio in patients with and without white matter hyperintensities. One patient presented with saccadic pursuit on examination of eye movements without any other significant cerebellar involvement (263).

In autosomal recessive hereditary spastic paraplegia, the first gene identified was paraplegin. It can either cause a pure or a complicated form of disease with dysarthria, dysphagia, optic disc pallor, axonal neuropathy, cerebellar atrophy, and cerebral atrophy. Even if the clinical picture shows a pure form, cranial imaging may show cerebellar atrophy (81). A novel SPG7 mutation has been defined in six members of a family with supranuclear palsy, cerebellar syndrome, and cognitive impairment. Diffusion tensor imaging studies have shown bilateral disturbance of white matter integrity in corticospinal tracts, frontal lobes, and midbrain, widening the different brain systems involved in hereditary spastic paraplegias (327). In a large cohort of Dutch hereditary spastic paraplegia patients, researchers have identified a correlation between the cerebellar phenotype of SPG7 and SPG7 null alleles. They have also identified four frequent mutations, and the most frequent, the c.1529 C4T (p.Ala510Val) mutation in exon 11, was identified in 26 separate families (315).

Klebe and colleagues have shown in a study of 135 patients, that all SPG7-positive patients tested had optic neuropathy or abnormalities revealed by optical coherence tomography, indicating that abnormalities in optical coherence tomography could be a clinical biomarker for SPG7 testing. They proposed that the presence of late-onset, very slowly progressive spastic gait associated with cerebellar ataxia or cerebellar atrophy constitute, with abnormal optical coherence tomography, key features pointing towards SPG7-testing. In one large family, novel missense SPG7 mutation at the heterozygous state (Asp411Ala) was identified as the cause of autosomal dominant optic neuropathy, indicating that some SPG7 mutations can occasionally be dominantly inherited and be an uncommon cause of isolated optic neuropathy. SPG7 mutation carriers also showed slight cerebellar syndrome and cerebellar atrophy on MRI or at least minor abnormalities on diffusion tensor imaging (155). In a Japanese patient with spastic ataxia who presented with arrhythmia, stiffness of legs, slurred speech, and cerebellar signs, an SPG7 mutation was detected for the first time in Japan by exome sequencing (75). Pfeffer and associates identified previously described mutations in the SPG7 gene in two patients presenting with the predominant finding of cerebellar ataxia (240). Follow-up showed pyramidal tract involvement as well. Prospectively, the authors then identified SPG7 mutations in 13 of 70 patients who presented with ataxia, most of whom had cerebellar atrophy documented by MRI scan. Each of them had at least one known pathogenic mutation, p.Ala510Val, at one allele and the same or other mutations at the other allele. All of these patients additionally manifested signs of pyramidal involvement. There are reports of SPG7 associated with macrocephaly and palatal tremor (108).

Paraplegin is the product of SPG7 gene. It is a member of AAA proteins and located in mitochondria. Mitochondrial AAA proteins have chaperone-like activity, and they have a role in the activation of respiratory chain complexes; furthermore, they function in protein quality control by binding unfolded peptides and ensuring the specificity of the proteins (164; 45). Although muscle involvement is unexpected in hereditary spastic paraplegia, the original family in which the paraplegin mutation was identified had muscle biopsies resembling the mitochondrial diseases with ragged-red fibers, with intense succinate dehydrogenase–stained areas and cytochrome oxidase negative fibers. But the other families having paraplegin mutation did not show these types of changes in their muscles. A study from the UK showed that in two patients with SPG7 mutation, although muscle biopsies did not show any histological phosphorylation defect histologically, biochemical studies showed a reduction in citrate-synthase corrected complex I and complex II/III activities in muscle samples and complex I activity in cultured myoblasts, which may suggest a respiratory chain defect in pathogenesis of the disease (331). Location of paraplegin was shown in the inner mitochondrial membrane along with its binding partner. This complex is lacking in mutated cells, and they show diminished activity of complex-1 of the respiratory chain leading to increased susceptibility to oxidative stress. Paraplegin shows both proteolytic and chaperone-like activities (10; 168). Mutations can lead to a dysfunction of mitochondria causing an accumulation of misfolded proteins within the mitochondrial matrix. Although all tissues may have this impaired quality control and poor proof-reading, the longest tract with high energy requirements, corticospinal tract axons, might show this kind of mitochondrial neurodegeneration (45). Experimental studies in SPG7-/- mice have shown an interesting finding: enlarged and structurally abnormal mitochondria accumulated in the synaptic terminals of long motor axons, not shorter axons and glial cells. Dynamic changes in the structure of the mitochondria seem to affect the axonal transport of mitochondria. Abnormal mitochondria might be less efficiently transported at the distal parts of very long axons, which may contribute to the axonal degeneration (265). This process of neurodegeneration is blocked by intramuscular injection of adeno-associated viral vector that encodes paraplegin to the hind limb muscles of SPG7-/- mice. After injection, transportation of the vector along the axons leads synthesis of new paraplegin, which is correctly imported into mitochondria and reduces numbers of swollen axons (241).

SPG11 causes complicated spastic paraplegia characterized by slowly progressive spastic paraparesis, mental retardation, and thin corpus callosum. Sometimes upper extremity weakness, dysarthria, and nystagmus can be seen. Tremor might be the initial presenting sign (275). Attention/calculation deficits and being overweight as well as pseudobulbar dysarthria were common additional symptoms in patients with SPG11 hereditary spastic paraplegia (149). The locus of SPG11 has been refined to 2.93 cM with a maximum lod score of 11.84 at marker D15S659 (217), and the causative gene KIAA1840 (also known as FLJ21439) has been found (295). A product of the gene is spatacsin, which has a subcellular localization expressed in all tissues associated with mitochondria and endoplasmic reticulum, but not with Golgi structures or cytoskeleton. Spatacsin may be involved in gene expression, or, like many other hereditary spastic paraplegia proteins, it may be involved in cellular trafficking events (297; 228). A role for spatacstin in lysosome membrane recycling and the clearance of gangliosides from lysosomes, preventing their accumulation, was demonstrated (36). SPG11 is allelic to juvenile amyotrophic lateral sclerosis and availability of genetic tests showed a mutation causing both juvenile amyotrophic lateral sclerosis and SPG11 in the same family. By exome sequencing, two affected patients showed two compound heterozygous deletions (63).

SPG11 is the most common form of autosomal recessive hereditary spastic paraplegia worldwide from China to Europe (163; 228; 294). In 76 patients, mutation frequency was 41% in those who presented with thin corpus callosum and 4.5% in those without thin corpus callosum. Another study showed that overall frequency of SPG11 was 14% and 42% in patients with thin corpus callosum (278b). It is highly found in patients without thin corpus callosum in both familial and nonfamilial cases (57). Additional features include lower motor degeneration with wasting and slight ocular cerebellar signs with long duration; on MRI, white matter changes and cortical (especially frontal) atrophy may be seen in addition to thin corpus callosum (294). Diffusion tensor imaging revealed increased mean diffusion and decreased fractional anisotropy in the corpus callosum and subcortical white matter of frontal and temporal lobes, peritrigonal white matter, posterior limb of the internal capsule, and in the semioval centers (44; 107). Diffusion spectrum imaging (DSI) tractography showed a significant reduction of the microstructural in all neural fiber types, whereas prefrontal and motor portions of the corpus callosum were the most severely affected among all the fiber tracts (230). Extent of corpus callosum thinning and cerebellar atrophy might be correlated with disease severity (249). Long-term follow-up of nine SPG11 Dutch patients showed that death occurs between 30 to 48 years of age, 3 to 4 decades after onset of gait impairment. Onset of age was between 4 months and 14 years, and learning difficulties were the second main complaint of onset after gait impairment. At the end stage of the disease loss of speech, severe dysphagia, and spastic tetraplegia with contractures were common (65). Clinical heterogeneity is common in this form, and although onset is in the first and second decades, a female athlete was reported to have first symptoms at 27 years of age and had a mild course (229). Involuntary movements and extrapyramidal signs can complicate SPG11, and L-dopa-responsive parkinsonism has been related to this form (87; 317). In three of four SPG11 patients with extrapyramidal signs like rigidity, dystonia, and bradykinesia, cerebrospinal fluid analysis showed low levels of neurotransmitters. After supplementation with L-dopa/carbidopa and sapropterin, all 4, including the one with normal neurotransmitter levels, responded to treatment, although the etiology of abnormal levels of neurotransmitters is not known (317).

Two patients with early-onset parkinsonism with resting tremor, akinesia, and rigidity who were initially responsive to levodopa developed progressive spastic paraparesis in the second decade. Their mutation analysis showed a mutation in SPG11, indicating that juvenile parkinsonism may also be related to SPG11 (06). A higher obesity rate in SPG11 patients has been observed, possibly associated with evidence of hypothalamic atrophy (68).

Spatacsin is also related to autosomal recessive juvenile amyotrophic lateral sclerosis (218). According to El-Escorial criterion from Italy, Brazil, Canada, Japan, and Turkey, spatacsin mutations were present in 10 of 25 families with definite autosomal recessive juvenile amyotrophic lateral sclerosis. SPG11 and ALS5 share the same chromosomal segment, and mutations found in ALS5 patients extended the spectrum of spatacsin-causative diseases (218).

SPG14 causes complicated spastic paraplegia with intellectual disability and distal motor neuropathy.

SPG15 (Kjellin syndrome) results in a complicated spastic paraplegia associated with pigmented maculopathy, distal amyotrophy, dysarthria, mental retardation, and further intellectual deterioration. Childhood-onset writer’s cramp has also been described in a patient with SPG15 (207). Genetic heterogeneity is also evident in this form of hereditary spastic paraplegia. Families with signs of cerebellar dysfunction, axonal neuropathy, mental decline, and even thin corpus callosum, but not the other signs of Kjellin syndrome, were linked to SPG15 (80). Disease is caused by a mutation in zinc finger protein called spastizin encoded ZFYVE26 (123). It is widely distributed in human tissues and in rat embryos and colocalized with endoplasmic reticulum and endosomes, suggesting a possible role in trafficking (123). The SPG15 gene is thought to code for a protein that interacts with Beclin-1 and aid in autophagy initiation. Beclin-1 is an orthologue protein used in the pathway for apoptosis. Interruption in this process has shown to lead to neurodegeneration (335).

Published reports showed that patients with thin corpus callosum and mental decline can have mutations in ZFYVE26/SPG15; this mutation is found in roughly 5% of patients with SPG11 in Italy (72). Patients with a clinical phenotype of Kjellin syndrome with central retinal degeneration also showed mutations in SPG11 (222). Orlen and her colleagues reported that central retinal degeneration, unreported in SPG11 before, is also a feature of SPG11 and that it develops later in the course of the disease and should be overlooked in SPG11 patients. SPG11 and SPG15 make up most of the autosomal recessive hereditary spastic paraplegias, and after reports with clinical overlap, genetic testing is required for both in autosomal recessive patterns. EFNS guidelines suggest molecular testing for both SPG11 and SPG15 when thin corpus callosum with an autosomal recessive inheritance pattern is present, and molecular diagnosis for SPG7 is suggested if cerebellar features are positive (109).

Both products of SPG11 and SPG15 are large proteins, and they are both associated with the AP-5 complex, a member of the family of heterotetrameric adaptor proteins (AP complexes). AP-5, SPG11, and SPG15 colocalize on a late endosomal/lysosomal compartment. Hirst and colleagues propose that AP-5, SPG15, and SPG11 form a coat-like complex, with AP-5 involved in protein sorting, SPG15 facilitating the docking of the coat onto membranes, and SPG11 (possibly together with SPG15) forming a scaffold (140).

SPG18 is a complicated form of hereditary spastic paraplegia that has been described in a large Saudi family. It is associated with ERLIN2 mutations, which is a component of endoplasmic reticulum-associated degradation pathway, and it is localized to endoplasmic reticulum (03). Wakil and colleagues have found a new mutation in a Saudi family. The two affected siblings had early-onset, cognitive, speech, and motor involvement with spasticity of the lower extremities (324). A novel ERLIN2 variant was reported in two families with an autosomal dominant, pure form of hereditary spastic paraplegia; therefore, ERLIN2 variants should be considered in both autosomal recessive and autosomal dominant forms of hereditary spastic paraplegia (267).

Troyer syndrome, one of the syndromic forms of hereditary spastic paraplegia (SPG20), is caused by a mutation in SPARTIN, located on chromosome 13q12.3. Troyer syndrome is associated with dysarthria and distal amyotrophy and is seen with high frequency in the Old Order Amish population; however, it has been detected in two Omani families also (184). This was followed by two more publications on Turkish and Filipino cases, respectively (304; 41). Spartin is thought to have an important role in intracellular protein trafficking and trafficking through the endocytic pathway (between plasma membranes and trans-Golgi network), like spastin molecule by its microtubule-interacting and trafficking molecules domain and through an interaction with Eps15 protein (60; 259). Studies have shown its subcellular location and localized it to mitochondria. Spartin, with its plant-related senescence domain, association with cardiolipin, a major mitochondrial phospholipid. Spartin knock-out mice and spartin-depleted cells have depolarized mitochondrial membrane and reduced mitochondrial calcium influx (145). Mutant protein loses its mitochondrial localization. Lu and colleagues suggest that, in addition to defective trafficking events, mitochondrial dysfunction may also be involved in the pathogenesis of Troyer syndrome (175). However, in cultured fibroblasts and lymphoblasts derived from affected individuals, no spartin protein could be detected, and it was suggested that Troyer syndrome results from complete loss of spartin protein (11). Manzini and colleagues showed that in vitro SPG20 expression is highest and more specific to limbs, face, and forebrain during early morphogenesis, supporting the phenotypic congenital abnormalities seen in Troyer syndrome (184).

Mast syndrome, SPG21, is a complicated form of hereditary spastic paraplegia that is also seen in high frequency in the Old Order Amish. It is usually a disease of early adulthood and is associated with slow progression to dementia, cerebellar and extrapyramidal signs, thin corpus callosum, and white matter abnormalities. Childhood-onset diseases with hypoplastic corpus callosum have been defined in Caucasian patients (40). It was mapped to 15q22.31, and the gene encodes a protein named maspardin (Mast syndrome, spastic paraplegia, autosomal recessive with dementia). Exact function of maspardin is not yet clear. It may be involved in sorting or trafficking of molecules between the endosomal/trans-Golgi network (289). Hanna and Blackstone showed that maspardin localizes to cytoplasm and membranes, supporting a functional role in the trans-Golgi network/endosomal pathway. It interacts with aldehyde dehydrogenase ALDH16A1 as a binding partner in this network. How mutations cause neurodegeneration is unknown, but mutations in aldehyde dehydrogenase ALDH3A2 result in Sjögren-Larsson syndrome, which is characterized by spastic paraparesis and ichthyosis; this shows the importance of aldehyde dehydrogenases in the pathogenesis of neurologic diseases (124).

SPG23 is associated with prematurely aged facial appearance, hypopigmentation, microcephaly, and cognitive impairment (28).

SPG24 is a pure autosomal recessive form (141).

SPG25 is complicated with disc hernia and linked to 6q23.3-q24.1 (347).

In SPG26, Wilkinson and colleagues reported wasting of the small muscles of the hands and feet and mild intellectual impairment in a Kuwaiti family (332). In families from Kuwait, Italy, and the Old Order Amish (129), and in families from Spain, Tunisia, Brasil, Germany, and Portugal, Boukhris and colleagues found mutations in the B4GALNT1 gene encoding GM2 synthase (33; 129). Patients presented with early-onset disease, intellectual disability, cerebellar ataxia, and peripheral neuropathy (33). GM2 synthase is involved in the biosynthesis of complex gangliosides (G), which are mono- (M), di- (D), and tri- (T) sialic acid-containing glycosphingolipids generated by sequential glycosylations. Gangliosides are part of the larger family of glycosphingolipids, are components of the synaptic plasma membrane involved in synaptic plasticity, signal transduction, and endocytosis, and are critical for central nervous system development. B4GALNT1 catalyzes the transfer of N-acetylgalactosamine into GM3, GD3, and globotriaosylceramide by a beta-1,4 linkage (33). Harlalka and colleagues have found that biochemical profiling of glycosphingolipid biosynthesis confirmed a lack of GM2 in affected subjects in association with a predictable increase in levels of its precursor, GM3, a finding that will greatly facilitate diagnosis of this condition (129). It has been discovered that this phenotype is also associated with transient febrile ataxia and myokymia (61).

SPG27 is linked to 10q22.1-10q24.1, which overlaps the locus of SPG9. It is not known whether it is a new locus or an allelic disorder with different mode of inheritance (197), but a new family different than SPG9 with early-onset spastic paraparesis, mental retardation, sensorimotor neuropathy, cerebellar syndrome, and dysmorphic face in one child has reduced the SPG27 locus, suggesting that it is a new type (254).

Bouslam and colleagues mapped SPG28 to 14q21.3-q22.3 with pure characteristics (35). A consanguineous Moroccan family was later studied by Tesson and colleagues with 99 other autosomal recessive hereditary spastic paraplegia patients without identified mutations, and a homozygous mutation in the DDHD1 gene was identified (307). The same mutation was found in two Turkish brothers, one of whom had cerebellar oculomotor disturbance with saccadic eye movements, as well as in a French woman with axonal neuropathy; extending the phenotype from pure to complicated (307). By whole-genome linkage mapping and next-generation sequencing of 99 patients, CYP2U1 mutation was detected in three patients with thin corpus callosum, white matter abnormalities, or calcification of the basal ganglia. Although the authors have called this type SPG49, the article by Oz-Levi and colleagues was sent for review before, so it was designated as SPG49 and this new type caused by CYP2U1 mutations was designated as SPG56 by the genome database (226; 307). Both genes encode for enzymes involved possibly in the same pathway related to phospholipid degradation and fatty acid metabolism. They were both expressed concomitantly in the developing mouse brain, and both showed partial mitochondrial localization. Mutant cells from SPG28 and SPG56 patients showed significantly lower mitochondrial respiration activity, lower ATP levels, and increased cytosolic hydrogen peroxide compared to controls. Alteration of mitochondrial architecture and bioenergetics with increased stress due to alteration of lipid metabolism has been shown on another pathway of hereditary spastic paraplegia pathophysiology (307).

SPG30 is linked to chromosome 2q37.3 and caused by a mutation in the motor domain of KIF1A, a protein that functions in anterograde axonal transportation. This leads to an autosomal dominant or recessive form of pure hereditary spastic paraplegia. One study found KIF1A variants as a frequent cause of autosomal dominant hereditary spastic paraplegia in 6% to 7% of their cohort. The identification of KIF1A loss-of-function variants led the authors to suggest haploinsufficiency as a possible mechanism in autosomal dominant spastic paraplegia (237). A report demonstrated this gene mutation leading to a subtype of complicated hereditary spastic paraplegia. This complex form consists of mild intellectual disability with language delay, epilepsy, optic nerve atrophy, thinning of corpus callosum, periventricular white matter lesion, and microcephaly (49). Erlich and colleagues discovered the gene in a pure Palestinian hereditary spastic paraplegia family; however, patients reported in 2006 showed progressive spastic paraparesis, neuropathy, cerebellar signs like ataxia, dysarthria, and nystagmus, and cerebellar atrophy on cranial imaging studies (154; 83). Together with SPG7, SPG21, and SPG27, SPG30 shows that cerebellar atrophy might not be an infrequent sign in hereditary spastic paraplegias and should not be overlooked even in patients with the pure form of the disease (154).

SPG32 is a new locus at 14q12-q21 with mild mental retardation, brainstem dysraphia, cerebellar atrophy, and corpus callosum thinning. It is an early-onset type of disease with very slow progression, and a marked feature is malformation of the pons (296).

SPG35 is a complicated form defined in an Omani family, and it is linked to 16q21-q23.1. Clinical features were complicated with seizures, dysarthria, and intellectual disability. MRI was normal (73). In another Omani family linked to the same locus, patients had ophthalmoplegia, optic atrophy, and dystonic upper limb movements with cognitive decline (74). Hyperintense lesions in cerebral and cerebellar white matter, cerebral and cerebellar atrophy, in addition to brainstem atrophy were detected on MRI. Mutation analysis of both families showed homozygous mutations in fatty acid-2 hydroxylase gene (FA2H) (74). Some patients with mutations in the FA2H gene have radiographic evidence of neurodegeneration with brain iron accumulation (NBIA), thus, expanding the phenotype. Kruer and colleagues referred to this phenotypic spectrum of disorders as fatty acid hydrolase-associated neurodegeneration (FAHN) (159). Rupps and colleagues have shown that a 5-year-old girl with spastic paraparesis and periventricular leucomalacia showed compound heterozygosity for FA2H (266). Cao and colleagues showed novel triple mutations in two Han siblings from nonconsanguineous parents with severe clinical features (43).

Like other motor neuron diseases like amyotrophic lateral sclerosis, causative mutation of SPG39 supported the role of neurotoxic substances in motor neuron disorders. Organophosphate compound-induced delayed neuropathy occurring after exposure to toxic substances may cause ataxia, weakness, muscle fasciculation, and flaccid paralysis followed by progressive spastic paraplegia (246; 46). It causes degeneration of the longest axons of nervous system and involves neuropathy target esterase protein. In a family with clinical features resembling either Troyer syndrome or organophosphate compound-induced delayed neuropathy, Rainier and her colleagues demonstrated a mutation in neuropathy target esterase protein, and it is designated as SPG39. They suggested that polymorphisms of neuropathy target esterase may play a role in motor neuron disorders like amyotrophic lateral sclerosis, even in the absence of exposure to toxic substances (246). NTE protein is found mainly in the large neurons of the central nervous system and regulates phospholipid metabolism. Lethality in knockout mice suggested its role in embryonic survival (47).

SPG43 was described in two sisters with hereditary spastic paraplegia and distal amyotrophy (198). Landoure and colleagues found the causative mutation in C19orf12 in the same family (162). C19orf12 is involved in neurodegeneration with brain iron accumulation-4 (NBIA-4), but the original family did not show iron deposition on MRI. Neurodegeneration with brain iron accumulation disorders are a group of disorders caused by different mutations. The same homozygous A63P mutation was detected in two siblings from a consanguineous Brazilian family with neurodegeneration with brain iron accumulation-4, and haplotype analysis indicated a founder effect between the Malian and Brazilian families. Although the cellular functions of C19orf12 are unclear, SPG35, SPG43, and NBIA can all share the common pathway (162).

SPG44 causes a complicated phenotype with pes cavus, scoliosis, and hypomyelinating leukoencephalopathy on MRI and MR spectroscopy imaging. The gene product is connexin47, and it is allelic to Pelizaeus-Merzbacher-like disease, an early-onset dysmyelinating disorder of the CNS (224). A 38-year-old patient with segmental dystonia, leg spasticity, and mild intellectual disability with novel mutations in the GJC2 gene shows the variability of the phenotypes (345).

SPG45 is an autosomal recessive inherited form of hereditary spastic paraplegia and is complicated with mental retardation and ocular abnormalities like myopia, hyperopia, optic atrophy, and nystagmus (78). This specific form of hereditary spastic paraplegia is now linked to a missense mutation in NT5C2 (5’- Nucleotidase Cytosolic II protein), which is responsible for regular maintenance of purine and pyrimidine composition intracellularly (298).

SPG46 is a new type with thin corpus callosum and mental impairment. In addition to spastic paraplegia, thin corpus callosum, and mental impairment, patients all have congenital cataract and cerebellar ataxia (32). Four different mutations were found in glucocerebrosidase GBA2 in SPG46, which encodes a microsomal nonlysosomal glucosylceramidase that catalyzes the conversion of glucosylceramide to free glucose, ceramide, and hydrolysis (188). In addition to cataract and mental impairment, hypogonadism was present in males, drawing attention to the importance of enzymes in different tissues. Like glucocerebrosidase GBA1, defective in Gaucher disease, Martin and colleagues assessed the enzymatic activity, and no detectable activity of GBA2 could be observed (188). Martin and colleagues suggest that enzyme replacement, like in Gaucher and Niemann Pick type C, are not suitable for SPG46 patients, but isofagomine used in Gaucher disease might be worth a try in SPG46 (188).

SPG47 is inherited in an autosomal recessive manner. It has been described in two siblings from an Arabic consanguineous family with slowly progressive spastic paraparesis, mental retardation, seizures, thin corpus callosum, and periventricular white matter abnormalities (29). The mutation is in the AP4B1 gene that encodes a member of the AP4 complex involved in forming and sorting of the vesicles. The patient initially presented with severe hypotonia and later developed progressive spastic paraparesis and convulsions (14). Kong and colleagues have identified a novel mutation in identical twins with hereditary spastic paraplegia and lymphadenitis caused by the live Bacillus Calmette-Guerin (BCG) vaccine. Although no pathological defect has been identified in immunological tests, AP4 has been suggested to be important in immunological pathways (158).

SPG48 is caused by a mutation in the KIAA0415/SPG48 gene. It encodes a helicase involved in DNA double-strand break repair pathway. By immunoprecipitation analysis of HeLa cells, Slabicki and colleagues showed that KIAA0415 exists in a core protein complex containing SPG11, SPG15 (ZFYVE26), C20ORF29 (AP5S1), and DKFZp761E198 (290). The Mu-2-related death-inducing (MuD) gene, which has a role in cell death and apoptosis, has been shown to be a subunit of AP-5 complex together with KIAA0415. Wagley and colleagues have reported a mouse monoclonal antibody against the middle domain of MuD protein (323). This might cause a new diagnostic option to determine prognosis of patients with hereditary spastic paraplegia by studying MuD expression using anti-Mud antibody (323).

SPG49 is a complicated type of hereditary spastic paraplegia caused by a homozygous mutation in the TECPR2 gene on chromosome 14q32 (226). Patients presented with severe intellectual disability, fluctuating central hypoventilation, gastroesophageal reflux disease, wake apnea, areflexia, and dysmorphic features. Thin corpus callosum and cerebral and cerebellar atrophy were present on MRI. TECPR2 has been reported as a positive regulator of autophagy, which is the main protein degradation system responsible for the turnover of bulky cellular constituents. Defective protein caused by a mutation in TECPR2 results in modification of autophagy and cytotoxicity, and dysfunction of this system might be the underlying mechanism of neurodegeneration (226).

SPG50 was defined in one Moroccan family with early infantile hypotonia, delayed psychomotor development, strabismus, lack of independent walking, and severe mental retardation. The affected family members developed spastic quadriparesis and speech impairment (321). The researchers postulated that the genetic defect results in abnormal cycling of glutamate receptors, mimicking glutamate-mediated perinatal white matter injury.

SPG51 is another autosomal recessive neurodevelopmental disorder characterized by neonatal hypotonia, microcephaly that progresses to hypertonia and spasticity, and severe mental retardation with poor or absent speech development (205).

SPG52 is defined by Abou Jamra and colleagues. Patients showed muscular hypertonia, contractures, talipes equinovarus, decreased shank muscle mass, short stature and microcephaly, severe cognitive deficit, and absent speech. The gene is AP4S1 and is located on 14q12 (01).

SPG53 is caused by a mutation in the VPS37A gene on chromosome 8p22. It is a complicated type with early-onset disease, kyphosis, hypertrichosis, and cognitive impairment (346).

SPG54 is a complicated type of hereditary spastic paraplegia caused by a mutation in the DDHD2 gene and is defined in two unrelated families (282). One of the families with epilepsy and thin corpus callosum was described by Al-Yahyaee and colleagues and was accepted as SPG18 (05). However, by exome sequencing, this large Omani and another Iranian family have been found to have a mutation in the DDHD2 gene. All of the patients had psychomotor delay and a very early onset progressive spasticity. Foot contractures, strabismus, dysarthria, dysphagia, and optic nerve hypoplasia were also present in most of the patients. MRI showed thin corpus callosum and white matter hyperintensities. Cerebral proton MR spectroscopy showed an abnormal lipid peak around the basal ganglia and thalamus that was very similar to the abnormal lipid peak seen in Sjogren-Larsson syndrome due to abnormal brain lipid accumulation (282). In other families with DDHD2 mutations, mild facial dysmorphism, short stature, and dysgenesis of corpus callosum were present, in addition to spastic paraplegia and mental retardation (114). DDHD2 ablation induces reactive oxygen species production in mitochondria, which was noted to be reversed with wild-type DDHD2, thus, suggesting this as a possible underlying mechanism for SPG54 (190).

DDHD2 belongs to the mammalian intracellular phospholipase A1 family, together with DDHD1 (in SPG28) and SEC231P. They hydrolyze acyl groups on phospholipids, and all three have roles in organelle biogenesis and membrane trafficking. DDHD2 was suggested to have a function in membrane curvature and membrane and vesicle fusion by modification of membranes through phospholipid hydrolysis. It is also involved in lipid signaling (282; 114).

SPG57 was reserved by Beetz and colleagues for a new type of complicated autosomal recessive hereditary spastic paraplegia in two siblings with early-onset spastic paraplegia, optic atrophy, and neuropathy (15). Exome sequencing revealed a homozygous c.316C>T (p.R106C) variant in the Trk-fused gene (TFG). TFG-depleted cells showed altered endoplasmic reticulum network, collapsed endoplasmic reticulum tubules on cytoskeleton, and abnormal clustering of mitochondria, providing a link between altered endoplasmic reticulum architecture and neurodegeneration (15). Proximal hereditary motor and sensory neuropathy (HMSNP) is also caused by a heterozygous mutation in the TFG gene.

SPG58 is secondary to KIF1C mutations resulting in alterations in the domains involved in adenosine triphosphate hydrolysis and microtubule binding (42). Cervical dystonia and cerebellar ataxia were noted to be complicating features of this disorder.

Novarino and colleagues identified previously unknown candidate genes involved in cellular transport, nucleotide metabolism, and synapse and axon development through whole-exome sequencing performed on individuals from 55 families with autosomal dominant hereditary spastic paraplegia (216). These were designated SPG54, SPG58, SPG59, SPG60, SPG61, SPG62, SPG63, SPG64, SPG65, SPG66, SPG67, SPG68, SPG69, SPG70, and SPG 71. Myelin-associated glycoprotein is a membrane-bound adhesion protein required for myelin function. MAG was implicated as a potential hereditary spastic paraplegia candidate resulting in a complicated form of the disorder (216).

The combination of spastic paraplegia with optic atrophy and peripheral neuropathy is rare. Lossos and colleagues described a family with this form with a mutation in the IBA57 gene, which is essentially involved in the biogenesis of mitochondrial 4Fe-4S proteins (173). This has been designated SPG74.

Gan-Or and colleagues identified CAPN1 mutations as a cause for an autosomal recessive form of hereditary spastic paraplegia, SPG76 (105). Calpain 1, the protease encoded by CAPN1, has a role in synaptic plasticity, synaptic restructuring, and axon maturation and maintenance. In addition to lower limb spasticity, other abnormalities reported in patients included upper limb hyperreflexia, dysarthria, ataxia, foot deformities, abnormalities of bladder function, and distal sensory impairment.

Yang and associates reported on FARS2 as a causative gene for a pure autosomal recessive form of hereditary spastic paraplegia (337). This gene encodes a mitochondrial phenylalanyl tRNA synthase.

Mutations in the ATP13A2 gene associated with hereditary spastic paraplegia are thought to result in impairment in lysosomal and mitochondrial function resulting in a complicated form of the disease, SPG78 (86).

SPG79 is the result of mutation in UCHL1. Bilguvar and colleagues reported a homozygous mutation in UCHL1 in three siblings with consanguineous parents with early-onset blindness secondary to optic atrophy, ataxia, nystagmus, dorsal column dysfunction, and spasticity (25).

Bouwkamp and colleagues reported a mutation in the ACO2 gene, a gene that encodes for mitochondrial aconitase, in a family with autosomal recessive complicated hereditary spastic paraplegia and suggested this should be considered as a disease-causing gene (37).

Another type of autosomal recessive hereditary spastic paraplegia is infantile-onset spastic paraplegia, which is allelic to the condition reported as juvenile amyotrophic lateral sclerosis. This type presents with infantile-onset severe spastic paraparesis with progression to tetraplegia and anarthria during the early teenage years; dysphagia and slow eye movements develop in the second decade. Eymard-Pierre and colleagues reported that this type of hereditary spastic paraplegia is associated with alsin mutation (88). Alsin is thought to have guanine nucleotide exchange factor activities. Guanine nucleotide exchange factors are known to associate with the GDP-bound form of GTPases. They are thought to have function in GTP binding and dissociation, thus, activating the GTPases. Alsin might have a key function in regulating these guanine nucleotide exchange factors and, thus, a role in many functions of endocytic trafficking, microtubule organization, and signaling cascades as well as neuronal development (88). It has been shown that alsin has a role also in neurodegeneration by enhancing apoptosis, and a missense mutation of alsin is cytotoxic (233). Cobanoglu and colleagues have found that alsin and spartin are colocalized and suggest that they may interact with each other giving another role to alsin (52).

Many alsin mRNA transcripts were found in various tissues when Als2 knockout mice were examined. It is suggested that different isoforms produced by different types of mutations are responsible for phenotypes of alsin mutations (116). A new mutation caused by maternal uniparental heterodisomy with partial isodisomy of chromosome 2 was detected in an infantile ascending hereditary spastic paraplegia patient, extending the necessity of genetic testing in patients without consanguinity (137).

GCH1 mutations have been associated with dopa-responsive dystonia, Parkinson disease, and tetrahydrobiopterin (BH4)-deficient hyperphenylalaninemia B. GCH1 mutations have been reported in five patients with hereditary spastic paraplegia. Pathway enrichment analysis suggested that GCH1 shares processes and pathways with other hereditary spastic paraplegia-associated genes, and structural analysis of the variants indicated a disruptive effect. Therefore, it has been suggested to try levodopa in patients with hereditary spastic paraplegia and include GCH1 in the screening panels of hereditary spastic paraplegia genes (318).

Table 2. Autosomal Recessive Hereditary Spastic Paraplegias

X-linked hereditary spastic paraplegias. SPG1 results from mutations in L1CAM, which is one of a subgroup of integral membrane glycoproteins belonging to a large class of immunoglobulin superfamily cellular adhesion molecules that mediate cell-to-cell adhesion at the cell surface. They guide neurite outgrowth during development, neuronal cell migration, axon bundling, synaptogenesis, myelination, and neuronal cell survival, and they have a long-term effect on signal transduction pathways (153). Gene locus is Xq28. This adhesion molecule is required for normal development of the corticospinal tract. In patients with altered L1CAM function, pyramids were found to be absent or reduced in size, and spastic paraplegia is a common finding (336). L1CAM mutations can also cause X-linked hydrocephalus, MASA syndrome (intellectual disability, aphasia, shuffling gait, adducted thumbs), and agenesis of corpus callosum. Because there is interfamilial and intrafamilial variability with this mutation, Fransen and colleagues refer to the syndrome as CRASH syndrome (corpus callosum hypoplasia, retardation, adducted thumbs, spastic paraplegia, and hydrocephalus) (101).

SPG2 can give rise to both pure and complicated types of hereditary spastic paraplegia. Gene map locus is Xq22, and it results from the mutations in proteolipid protein. This locus is the same with the Pelizaeus-Merzbacher disease; therefore, SPG2 and Pelizaeus-Merzbacher disease are allelic disorders at the same locus (272). A high number of mutations have been reported in proteolipid protein-related disorders. Gene duplications of variable size are present in 60% to 70% of patients, and intragenic lesions are present in 15% to 20% of patients. Clinical severity is independent of the size of the duplication in most of the patients, and peripheral neuropathy never occurs in patients with PLP1 gene duplications. Most severe phenotypes are usually associated with PLP1 missense mutations (117). A family from Argentina consisting of a normal phenotype mother and two severely affected sons had shown a novel missense mutation in PLP1. MRI of the patients showed patchy leukodystrophy areas of hypomyelination (215). A novel missense mutation was found in an autopsy case of a 67-year-old SPG2 patient; this case represents one of the oldest ages of onset at 35 years of age (301). Proteolipid protein is the primary constituent of myelin, and it has two isoforms, namely proteolipid protein and DM20. Altered proteolipid protein structures are thought to compromise oligodendrocyte function. Experimental studies in knockout animals that do not express any proteolipid protein have shown that these animals have normal CNS function with normal life expectancy, but they have axonal swelling and degeneration due to impaired axonal transport (115). These data may help to understand that proteolipid protein mutation causes dysfunction in local support of oligodendrogliocytes for myelinated axons because its absence does not cause dysmyelination as in Pelizaeus-Merzbacher disease (45).

SPG16 is another X-linked hereditary spastic paraplegia that causes both pure and complicated forms of the disease. Its gene locus is Xq11.2. Motor aphasia, reduced vision, mild intellectual disability, and bowel and bladder dysfunction are the symptoms of the complicated form (93).

Some families with X-linked pedigrees do not show linkage to the known loci. In 2002, Starling and colleagues reported a large Brazilian family with pure hereditary spastic paraplegia showing linkage to the proteolipid protein locus, but no mutation could be found. They suggested that their results may show either a fourth X-linked gene with close proximity to proteolipid protein or a novel mutation in the noncoding regions of the proteolipid protein gene on Xq24-q25 (293). Macedo-Souza and colleagues reported a Brazilian family linked to the same region of Xq24-q25 with 11 affected men between 12 to 25 years of age who presented mainly with pure hereditary spastic paraplegia, and they definitely excluded Xq22.2. This new locus was the fourth locus of X-linked hereditary spastic paraplegias and was designated as SPG34 (177).

Table 3. X-Linked Recessive Hereditary Spastic Paraplegias

Mitochondrial inherited hereditary spastic paraplegias. Mitochondrial or maternally inherited hereditary spastic paraplegias are predominantly complicated hereditary spastic paraplegia and are associated with approximately 1% to 2% of hereditary spastic paraplegia cases (200). Mitochondrial mutations in MT-CO3 gene encoding for cytochrome c oxidase III/complex IV—respiratory chain complex IV subunit; MT-T1 gene associated with Isoleucine transfer RNA; and MT-ND4 and MT-ATP6 genes encoding for Complex V, ATP synthase, and subunit ATPase 6—respiratory chain complex V subunit are associated with hereditary spastic paraplegia (98; 209). Other symptoms include intellectual impairment, cerebellar ataxia, hearing loss, progressive external ophthalmoplegia, and neuropathy (322; 209).

There are also unclassified hereditary spastic paraplegias that are not in the present classification but may become a part of this classification in the future. Maternally inherited spastic paraplegias; or spastic paraplegia, optic atrophy, and neuropathy syndrome (SPOAN); and oligodendropathy are some of the candidates (99). SPOAN is characterized by congenital optic atrophy, progressive spastic paraplegia with onset in infancy, and progressive motor and sensory axonal neuropathy (178). Out of 68 Brazilian patients, 64 were genetically investigated and all were homozygotes for D11S1889 at 11q13 (178).

MEGDEL syndrome (methylglutaconic aciduria, deafness, encephalopathy, Leigh-like syndrome) is a syndrome associated with the SERAC1 mutation. SERAC1 is a phosphatidylglycerol remodeling enzyme that is used both in mitochondrial functioning and intracellular cholesterol trafficking. MEGDEL syndrome presents in infancy and is characterized by feeding problems, liver failure, spasticity, dystonia, hearing loss, and axial hypotonia. It has now been discovered that SERAC1 mutation can lead to a SERAC1 deficiency syndrome leading to new forms of complex hereditary spastic paraplegia that present in late adolescence or adulthood. The milder deficiency leads to an oligo-systemic juvenile complex hereditary spastic paraplegia, whereas the moderate deficiency leads to multisystemic juvenile complex hereditary spastic paraplegia. The additional phenotype to this subgroup presentation includes seizures, cognitive impairment, dystonia, bulbar signs, and a sensorimotor axonal neuropathy (261).

Sporadic hereditary spastic paraplegia. Thirteen percent to 40% of hereditary spastic paraplegias are sporadic (ie, with no family history) (281; 50).

Epidemiology

Hereditary spastic paraplegia has a global prevalence of 1.8 per 100,000 in their systemic review and meta-analysis of prevalence studies performed across various countries (264). A large, nationwide, epidemiological survey in Portugal reported the overall prevalence of hereditary spastic paraplegia at 4.1 per 100,000. In this survey, the prevalence of autosomal dominant hereditary spastic paraplegia was 5.6 per 100,000 and autosomal recessive hereditary spastic paraplegia was 3.3 per 100,000 population (56). However, large areas of the world remain without prevalence studies. Autosomal dominant, autosomal recessive, or X-linked modes of inheritance are reported, with 13% to 40% of cases being sporadic (ie, with no family history) (281; 50).

There is significant variation in the reported prevalence of hereditary spastic paraplegia. This is likely due to the different genetic makeup of the populations and also methodical heterogeneity. In North American and north European hereditary spastic paraplegia populations, SPG4/SPAST mutations are evident in 40%, SPG3A/ATL1 mutations in 10%, SPG31/REEP1 mutations in 10%, and SPG10/KIF5A mutations in 3% (306; 200). Autosomal recessive hereditary spastic paraplegias are more complex and seen in a high degree of the consanguineous marriage population, with nearly 30% registered hereditary spastic paraplegias from the Middle East and Northern Africa. SPG11 and SPG15 constitute a significant proportion of autosomal recessive forms of hereditary spastic paraplegias (264). Approximately 1% to 2% of cases show X and mitochondrial chromosomes mutations (201).

Prevention

Pattern recognition and pedigree information is essential. Genetic counseling could be provided for this heterogenous disorder.

Differential diagnosis

Confusing conditions

Consideration needs to be given to performing investigations that rule out other pathological processes (compressive, inflammatory, infectious, vascular) that may mimic the presentation of hereditary spastic paraplegia. Hereditary spastic paraplegia is a diagnosis exclusion.

Structural spinal cord abnormalities such as Arnold-Chiari malformation, cervical or lumbar spondylolysis, arteriovenous malformation, tethered cord syndrome, and syringomyelia should be excluded. Granulomas like tuberculosis and neoplasms involving the spinal cord also should be considered in the differential diagnosis. Degenerative diseases such as multiple sclerosis, amyotrophic lateral sclerosis, or spinocerebellar ataxias should be distinguished. The most important type of spinocerebellar ataxias that mimics hereditary spastic paraplegia is SCA3 (Machado-Joseph disease), and definite diagnosis requires genetic testing to differentiate SCA3 from hereditary spastic paraplegias (326).

Other genetic diseases such as adrenoleukodystrophy and adrenomyeloneuropathy, Krabbe leukodystrophy, metachromatic leukodystrophy, and metabolic causes such as subacute combined degeneration, abetalipoproteinemia, arginase deficiency, and mitochondrial encephalopathies should also be considered before hereditary spastic paraplegia is diagnosed. Nutritional deficiencies (eg, vitamin B12, vitamin E, and copper deficiency) are also considered in the differential diagnosis.

Infectious causes like neurosyphilis, AIDS, and HTLV-1 infections should be excluded. Spastic diplegic form of cerebral palsy or dopa-responsive dystonias may present in a similar manner as hereditary spastic paraplegia. The perinatal history will be important for cerebral palsy and the diurnal fluctuation history can indicate dopa-responsive dystonias. Refer to Tables 4 and 5. On physical examination, knee hyperextension, nonsagittal pelvic movements, and reduced range of motion at the knee, ankle, and hip represent the most peculiar patterns in hereditary spastic paraplegia, compared to diplegic cerebral palsy and stroke (89).

Diagnostic workup

Diagnostic workup can help support the diagnosis of hereditary spastic paraplegia and exclude some of the diseases in the differential diagnosis. Consider hereditary spastic paraplegia in children whom no acquired cause for their spastic paraplegia is identified.

Plasma levels of 25 hydroxycholesterol and 27 hydroxycholesterol, plus their ratio to cholesterol, can be considered as first-line investigations for patients with unexplained spastic paraplegia (187). The levels will be elevated and, thus, considered as diagnostic biomarkers for SPG5 hereditary spastic paraplegia with 100% sensitivity and specificity. This guides further testing to identify the specific CYP7B1 mutation.

Investigations to exclude acquired causes of progressive spastic paraplegia should be considered in the initial work-up. These include:

Ophthalmological abnormalities may occur in specific subtypes of hereditary spastic paraplegia and in genetic diseases that present with spastic paraplegia mimicking hereditary spastic paraplegia. These ophthalmological changes may precede the motor symptoms and include pigmentary retinal degeneration, ophthalmoplegia, optic atrophy, cataracts, and nystagmus (69). Some ophthalmological abnormalities are more prevalent in specific forms of hereditary spastic paraplegia. Considering that the diagnosis of hereditary spastic paraplegia is usually difficult and complex, specific ophthalmological changes may guide the genetic testing. For example, optic atrophy is notable in some forms of hereditary spastic paraplegia, such as SPG7 (155). SPG4 mutations were, however, not associated with the development of significant ophthalmological complications (121). Temporal thinning of the retinal nerve fiber layer thickness shown by spectral domain optical coherence tomography has been reported in association with the complex forms in hereditary spastic paraplegia, but it might not be suitable for use as a biomarker in hereditary spastic paraplegia as it appears not to be specific to this condition and can be a feature of aging (319).

A new cellular imaging–based method to distinguish the SPG4 subtype has been described (270). It was observed that lymphoblastoid cells and peripheral blood mononuclear cells from individuals affected by SPG4 hereditary spastic paraplegia show a polarized microtubule cytoskeleton organization. This finding could discriminate SPG4 hereditary spastic paraplegia from healthy donors and other hereditary spastic paraplegia subtypes and can detect the effects of spastin protein level changes.

Neurophysiology. Electrophysiologic studies in hereditary spastic paraplegia have demonstrated a broad spectrum of involvement of the motor and sensory tracts with axonal or demyelinating patterns. This heterogeneity can be partly explained by the different genetic subtypes of hereditary spastic paraplegia (150).

Motor and somatosensory evoked potentials (SSEP) can distinguish hereditary spastic paraplegia subjects from controls. In a cross-sectional case-control study, motor evoked potentials (MEPs) were severely affected and somatosensory evoked potentials latencies were prolonged, with longer latencies being related to more severe disease, indicating that somatosensory evoked potentials are candidate disease biomarkers for hereditary spastic paraplegia (189; 39). Electromyography and nerve conduction studies identified distal axonal motor neuropathy and neurogenic changes in almost half of the patients. Most of these patients had SPG7, SPG11, SPG15, and SPG35 mutations. Abnormalities in the somatosensory evoked potentials were identified in two thirds of the patients with hereditary spastic paraplegia subtypes other than SPG10, SPG31, and SPG35.

Neuroimaging. Magnetic resonance imaging is an important tool that helps to characterize the phenotypes of hereditary spastic paraplegia and also to rule out important mimics of hereditary spastic paraplegia. This will assist in the approach to genetic testing. Spinal MRI is generally normal in uncomplicated hereditary spastic paraplegia and in most cases of complicated hereditary spastic paraplegia. The incidence of spinal atrophy, especially in the cervical and upper thoracic regions, is, however, higher than in the normal population. Spinal cord atrophy may be seen in SPG6 and SPG8 more severely than other types of hereditary spastic paraplegia. Such atrophy is, however, not exclusive to hereditary spastic paraplegia, as it can be found in other acquired motor neuron diseases (135; 291).

Brain MRI may provide clues to specific genetic types of hereditary spastic paraplegia. Sometimes cortical atrophy and hypoplasia of corpus callosum may be detected in cranial MRI (128; 192). One study noted that brain abnormalities in hereditary spastic paraplegia are gene-specific: basal ganglia, thalamic, and posterior white matter involvement in SPG4; diffuse white matter and cerebellar involvement in SPG7; cortical thinning at the motor cortices and pallidal atrophy in SPG8; and widespread grey matter, white matter, and deep cerebellar nuclei damage in SPG11. Abnormal regions in SPG4 and SPG8 matched those with higher SPAST and WASHC5 expression, whereas in SPG7 and SPG11 this concordance was only noticed in the cerebellum (287). Pensato and colleagues found that most of the patients in their cohort (n=61) that presented with thin corpus callosum on neuroimaging had mutations in the SPG11 and SPG15 genes (238). Using automated segmentation of the subcortical structures, one group found a significant reduction in thalamic volume bilaterally as well as an inward deformation, mainly in the sensory-motor thalamic regions in patients with pure hereditary spastic paraplegia and a mutation in SPG4. A significant negative correlation between the shape of the thalamus and clinical scores was also observed (212).

The “ear of the lynx” sign was noted to be a characteristic early imaging finding of hereditary spastic paraplegia-thin corpus callosum in patients with SPG11 mutations. This finding is due to abnormal MRI signal at the forceps minor of the corpus callosum (257). This may also be seen with SPG15 mutations. In a patient with hydrocephalus, gait abnormalities, and X-linked inheritance abnormalities, SPG1 should be highly considered (277). Some hereditary spastic paraplegia subtypes present with basal ganglia iron deposition seen as hypointense signal lesions on T2 or susceptibility weighted images (64). Cerebellar atrophy, white matter lesions, and white matter abnormalities may also be seen, particularly in complicated forms of the disease (156). If marked atrophy or marked changes in cerebral white matter are seen, other diseases should be investigated and excluded before a definite diagnosis of hereditary spastic paraplegia can be made.

In patients with a predominant upper motor neuron (UMN) phenotype, the major differential diagnosis includes hereditary spastic paraplegia (HSP), upper motor neuron predominant amyotrophic lateral sclerosis (ALS), and primary lateral sclerosis (PLS). Cosottini and colleagues explored the iron-related signal changes of the primary motor cortex in patients with hereditary spastic paraplegia, primary lateral sclerosis, and amyotrophic lateral sclerosis using targeted T2- weighted imaging (55). Notably, most patients with hereditary spastic paraplegia had normal signal intensity in the primary motor cortex, whereas the majority of patients with primary lateral sclerosis and amyotrophic lateral sclerosis had signal hypointensity, suggestive of microglial iron accumulation. Therefore, T2-weighted imaging of the primary motor cortex could provide useful information in the differential diagnosis of sporadic adult-onset upper motor neuron syndromes (55).

Table 4. Neuroimaging Findings in Hereditary Spastic Paraplegia

Genetic testing. Genetic studies are helpful to confirm the clinical diagnosis of hereditary spastic paraplegia. Despite the delineation of numerous genetic causes of hereditary spastic paraplegia, a significant portion of individuals with hereditary spastic paraplegia remain molecularly undiagnosed.

There are different approaches to the genetic workup of these disorders that continues to evolve with increasing accessibility of comprehensive genomic testing. The approaches include single-gene testing, next-generation sequencing-based gene panel, or whole-exome sequencing.

Next-generation sequencing-based gene panels for hereditary spastic paraplegia are increasingly cost-efficient and now widely available. These panels are most commonly used to screen the exons of a large number of genes associated with hereditary spastic paraplegias. However, there are limitations, such as they will generally not identify copy number variants (ie, large deletions or duplications, including exon deletions), mutations in promotor or deep intronic regions, and triplet repeat disorders (136). This technical limitation is a particular problem for SPAST, where exon deletions are relatively common. Multiplex ligation-dependent probe amplification is recommended for this gene if sequencing approaches give normal results (30).

Exome sequencing has been shown to assist with the diagnosis of patients after many years of follow-up and after time-consuming and expensive tests. Two Spanish siblings with upper and lower motor neuron findings and ocular and cerebellar signs were diagnosed as SPG11, and a novel nonsense mutation was detected in exome sequencing after a long follow-up since 1991 (23).

Kumar and colleagues have identified 17 SPAST mutations by targeted next-generation sequencing in 27 previously studied SPAST-negative patients, suggesting further studies to define the most cost-effective and practical method of molecular diagnosis of hereditary spastic paraplegia (160). Marais and colleagues suggested that gene panel studies similar to that used in their study (incorporating 34 genes) could provide a diagnosis in nearly half of all patients if used as a first approach (185). They noted that a more extensive panel did not significantly increase yield. Luo and colleagues have designed a gene chip for the most frequent genes for a cheaper way of genetic diagnosis and suggested microarray technology as an alternative for whole-exome sequencing and whole gene sequencing (176).

Table 5. Differential Diagnosis and Diagnostic Workup

Management

Currently, there are no specific treatments that can delay, prevent, or reverse the progressive disability that is seen in hereditary spastic paraplegia. Management is mainly symptomatic, in a multidisciplinary clinic with the neurologist, primary care physician, physiotherapist, and occupational therapist. The aim is to reduce spasticity, improve function, and enhance quality of life.

Symptomatic treatment. Antispastic drugs such as baclofen, tizanidine, gabapentin, and dantrolene sodium can be used. Oral baclofen is commenced at a low dose and is gradually titrated up in order to reach a level that is sufficient enough to cross the blood-brain barrier and reach the spinal cord. For severe cases, intrathecal baclofen treatment can be considered and has previously been found to be effective in relieving spasticity (199) but not in sleep and respiratory functions (20). Margetis and colleagues demonstrated that intrathecal baclofen improved the spasticity and walking performance of patients with hereditary spastic paraplegia who had initially failed oral antispasticity medication (186).

Gabapentin is commonly used to relieve spasticity. Its efficacy was tested in a double-blind, prospective, crossover placebo-controlled trial in 10 patients with SPG4 spastic paraplegia. Gabapentin was given for 2 months, with a 10-day wash-out period, followed by placebo for 2 months. The study found no statistically significant clinical difference between the gabapentin arm and the placebo arm when measured against objective clinical measures and self-reported measures (274).

Botulinum toxin is a well-established therapy for focal spasticity. Investigators continue to evaluate its efficacy in hereditary spastic paraplegia. It has been found effective in reduction of muscle tone and improvement of gait velocity (70). In a series of 19 patients, botulinum toxin was effective in reduction of spasticity, especially in mild and moderate degrees of spasticity, without many side effects. Its effect was thought to be higher when combined with physical therapy (132; 70). In a clinical trial of 12 children with hereditary spastic paraplegia, botulinum toxin improved motor function and muscle tone, even in a very severe case of infantile-onset ascending hereditary spastic paraplegia. The effect lasted for a mean of 6 months, and botulinum toxin was suggested for a prolonged functional improvement (111). A study of 33 adult patients suggested only limited effect with regard to motor measures; however, significant improvement in fatigue scores were noted (286). Combined treatment of botulinum toxin type A and physiotherapy led to improvement of spasticity and quality of life in patients with hereditary spastic paraplegia at 1- and 3-months follow-up (234).

The administration of GABA agonist progabide with a median dose of 24.3 mg/kg significantly reduces hypertonia spasticity, flexor spasms, and reflex responses compared with placebo (19). Complicated hereditary spastic paraplegia patients with methylenetetrahydrofolate reductase inadequacy treated with vitamins B12 (1000 mg per month), folinic acid (45 mg per day), folic acid (15 mg per day), and betaine anhydrous (10 g per day) resulted in the reduction in homocysteine concentration and improved the conditions. However, more studies are required to obtain conclusive data (19; 201).

Dalfampridine (4-aminopyridine) was trialed in an open, proof-of-concept study in a small group of patients with confirmed hereditary spastic paraplegia. Bereau and colleagues demonstrated its efficacy with significant improvements in walking ability after a 2-week course (21). Subsequently, a second study to explore its efficacy found significant improvement in the walking ability, spasticity, and hand dexterity in the patients who used dalfampridine (310). In a triple-blinded, randomized, placebo-controlled pilot trial, four patients with hereditary spastic paraplegia received dalfampridine (10 mg twice daily) in addition to physiotherapy (twice a week), and four patients received placebo in addition to physiotherapy for 8 weeks (208). In the experimental group, the improvements were significantly higher for walking speed and muscle parameters, such as muscle length and spasticity assessments and functional muscle strength.

Earlier studies on the effect of L-threonine for the treatment of spasticity in patients with familial spastic paraparesis concluded that although it significantly suppressed the signs of spasticity, there were no clinically significant benefits on clinical assessment and patient self-evaluation (118). L-threonine was used to increase the inhibitory glycine levels in the spinal cord, hence, possibly reducing spasticity and signs of upper motor syndrome. Selective dorsal rhizotomy is a feasible treatment option in carefully selected children with hereditary spastic paraplegia, especially in walking patients. The majority of patients benefit with respect to gross motor function, and the complication risk is low (312). However, clinicians should be aware that interventions reducing spasticity might result in ineffective functional outcomes unveiling weakness. Intensive active physical therapy and functional electrical stimulation might improve gait velocity in the very short term (89).

An early and regular physical rehabilitation program may be beneficial for patients to help maintain flexibility, strength, range of motion, and ambulation. Assistive devices may be utilized to maximize function. Monitoring gait parameters are meaningful measures of hereditary spastic paraplegia progression as it has been shown that gait speed is significantly associated with progression on The Spastic Paraplegia Rating (SPR) Scale (171). Identifying the specific gait patterns may help to individualize rehabilitative therapies and evaluate their effects over time (285). Orthotics, such as ankle-foot orthotics and heel raises, are used as walking aids to improve mobility and balance. Botulinum toxin treatment and subsequent muscle stretching of the calves improved comfortable gait velocity and reduced muscle tone in 15 patients with pure hereditary spastic paraplegia while preserving muscle strength (70). In a randomized control trial, there was insufficient evidence to conclude that 5 weeks of gait-adaptability training (10 hours of C-Mill training—a treadmill equipped with augmented reality) leads to greater improvement of gait adaptability in people with pure hereditary spastic paraplegia (314).

Gastrointestinal complaints (including constipation and fecal incontinence) and voiding dysfunction are common in hereditary spastic paraplegia and should be explored and addressed by clinicians (148). Oropharyngeal dysphagia can be present, and surveillance of swallowing function should be part of the management of these patients.

Myocardial involvement may occur as cardiomyopathy in hereditary spastic paraplegia. A routine cardiological evaluation like the other neuromuscular disorders in the follow-up of patients is suggested (110).

Patients with hereditary spastic paraplegia may present with different forms of movement disorders, such as parkinsonism, dystonia, tremor, myoclonus, and ataxia. Correct identification of the movement disorder may guide specific treatments, such as levodopa use, botulinum toxin, and others (236).

Although in general no disease-modifying treatment exists at present, experimental studies provide novel insights and hope into both molecular mechanisms and future therapeutic modalities. Examples include paraplegin synthesis after an adeno-associated virus vector injection into skeletal muscle of mice that leads to an improvement by reducing abnormal mitochondria and vinblastine, a vinca alkaloid treatment in Drosophila with spastin mutations (223; 241). Further research into cellular trafficking and autophagy contributing to neurodegeneration in hereditary spastic paraplegia has led to the use of rapamycin in Drosophila atlastin knockdown hereditary spastic paraplegia models (simulating the SPG3A gene variant). It works by inhibiting mTOR kinase, the enzyme that inhibits autophagy, and it has been promising in delaying or attenuating neurodegeneration, muscle degeneration, and protein deposition in muscle fibers. This treatment has not been applied past a laboratory stage but shows promise (335). Severe 5,10-methylenetetrahydrofolate reductase (MTHFR) deficiency has been reported as a rare cause of complicated hereditary spastic paraplegia enabling treatment with betaine and vitamins (239).

Repetitive transcranial magnetic stimulation (TMS) has been investigated as a therapeutic option for hereditary spastic paraplegias. Antczak and associates demonstrated an increase in strength of the distal and proximal muscles of the lower extremities and decreased spasticity of the proximal muscles following five sessions of bilateral 10 Hz repetitive transcranial magnetic stimulation over the primary motor areas of the muscles of the lower extremities (08). A randomized control trial demonstrated that repetitive transcranial magnetic stimulation is effective in reducing lower limb spasticity in patients with hereditary spastic paraplegia (12).

Electrical twitch-obtaining intramuscular stimulation (ETOIMS) has been shown effective for improving gait speed and stability by relaxing the muscles or alleviating the pain in the lower back and gluteal area in patients with hereditary spastic paraplegia (288). In addition, a randomized controlled trial, Move-HSP, will be commencing in patients with pure hereditary spastic paraplegia. In the trial, C-Mill, a treadmill equipped with augmented reality that enables visual projections to serve as stepping targets or obstacles, will be used, and effects on gait will be assessed (313).

Robotic locomotive systems are utilized to improve the gait in patients with neurologic disorders. One uncontrolled study demonstrated a beneficial effect of robotics assistance in intensive training in adults with pure hereditary spastic paraplegia. The berg balance and Meter walk test scores were significantly improved with the robotic aided training (22). A case report showed a slight improvement in walking speed and balance in a 28-year-old pure hereditary spastic paraplegia man after giving 25 training sessions for 6 weeks with physiotherapy and robotic training (284).

A small uncontrolled trial has assessed the efficacy of hydro-training in enhancing the locomotor function in patients with late-onset hereditary spastic paraplegia after 10 weeks of 45-minute sessions (340). Significant improvement was observed in the kinematics and spatiotemporal measures to improve gait. However, further studies are required to signify the efficacy of hydro-training.

A medical geneticist or genetic counsellor can assist in cases of individuals seeking genetic counseling for hereditary spastic paraplegia. Also, regularly screening for depression, pain, and fatigue and asking about bladder, sexual, and defecation problems and recognizing and treating nonmotor symptoms can improve quality of life in patients (250).

Gene-specific therapies. Despite continuing advances made in understanding the genetic basis of hereditary spastic paraplegia, there is still no effective treatment for hereditary spastic paraplegia. Potential therapeutic targeted treatments for some hereditary spastic paraplegia subtypes are summarized below.

SPG4 mutations (Loss of function mutations in the SPAST gene) lead to deficiency of spastin, which is then associated with lower concentration of dynamic microtubules in the cells and increased stable acetylated microtubules (262). Other studies show that SPAST modulates endoplasmic reticulum stress. Null SPAST homologs in Caenorhabditis elegans, fruit fly, and zebrafish were studied, and compounds known to modulate endoplasmic reticulum stress (such as phenazine, methylene blue, N-acetyl-cysteine, guanabenz, and salubrinal) partially rescued locomotor defects in the model organisms (146). This study provided insights to further novel therapeutic avenues for hereditary spastic paraplegia.

SPG5 is caused by mutations in the gene CYP7B1 that encodes for oxysterol-7α-hydroxylase implicated in cholesterol and bile acid metabolism. The deficiency of this enzyme has been shown to increase the levels of neurotoxic oxysterols, such as 27 hydroxycholesterol and 25 hydroxycholesterol. Schöls and colleagues performed a randomized, placebo-controlled trial with atorvastatin 40 mg/day given for 9 weeks to a small cohort of patients with SPG5 hereditary spastic paraplegia. They were able to demonstrate a reduction in the serum levels of 27 hydroxycholesterol and 25 hydroxycholesterol. There was an insignificant reduction of CSF 27 hydroxycholesterol when compared with the placebo and no effects were seen on the clinical parameters measured (276). This study, however, marked the first causal treatment study in hereditary spastic paraplegia.

Marelli and associates demonstrated that atorvastatin when combined with chenodeoxycholic acid may reduce the abnormal serum levels of 27 hydroxycholesterol and bile acid profile that is seen in patients with SPG5 hereditary spastic paraplegia (187). However, the neurologic benefit of these two interventions remains to be evaluated in phase 3 therapeutic trials.

SPG7 mutation is associated with deficiency of paraplegin, a protease found on the inner mitochondrial membrane. Paraplegin-deficient mice showed progressive degeneration in multiple axonal tracts and accumulation of abnormal mitochondria. Pirozzi and colleagues demonstrated that adeno-associated virus-mediated intramuscular delivery of paraplegin halted the progression of the neuropathological changes in paraplegin-deficient mice (241). This study shows that gene transfer may be an effective gene-specific therapy option for patients with SPG7 mutations and needs to be explored further.

Lipid metabolism is also a treatment target for SPG11, for which the goal is to decrease ganglioside accumulation. Miglustat improves the SPG11 zebrafish phenotype, but the estimated dose required to provide a benefit in humans is far too high to be ingested safely (38). New studies of Tideglusib, an FDA-approved GSK3ß inhibitor, on SPG11 induced pluripotent stem cells have tested its ability to preserve neuronal progenitors and reduce impaired neurogenesis and neuronal differentiation (202; 243). Preclinical and clinical trials are expected. Studies of microtubule-targeting drugs have also been conducted on drosophila SPG4 models and Spg4 knockout mouse neurons (90). The results suggest that such drugs (nocodazole, taxol, or vinblastine) could have a beneficial impact on microtubule dynamics.

Outcomes

In a large retrospective cohort, the median disease duration until loss of independent walking was 22 years (281). This German cohort of 608 patients included 198 (33%) SPG4, 28 (5%) SPG7, and 15 (2%) SPG11 cases, among numerous rarer SPG types. After 20 years of disease duration, 48% used a walking aid and 12% used a wheelchair. After 40 years of disease duration of 40 years, 72% used a walking aid and 29% used a wheelchair (281).

As hereditary spastic paraplegias are a disabling group of motor neuron diseases, with slow progression in most, efforts should be focused on certain hereditary spastic paraplegia genotypes in which work could be used to identify biomarkers (rating scales, blood, and CSF) that are sensitive to slow progression or change.

The following years will see the identification of yet more hereditary spastic paraplegia loci. However, we also hope that they will bring the beginning of rationally designed therapeutic approaches for hereditary spastic paraplegia.

Special considerations

Pregnancy

Nothing is known about the risks to the mother or the fetus regarding hereditary spastic paraplegia in pregnancy.

Anesthesia

Reports on anesthetic management of hereditary spastic paraplegias are few. McTiernan and Haagenvik reported a patient presenting for an emergency Cesarean section (196). Regional anesthesia was considered because the patient preferred to be awake and also because of the additional risks of aspiration and succinylcholine inducing hyperkalemia in patients with paraplegia. They used fentanyl for regional block. Although no reports support their theories, it is reported that epidural anesthesia is better in patients with chronic neurologic disease because they usually have poor respiratory functions, one single shot of spinal anesthesia provides less controllable anesthesia, and potential neurotoxicity of local anesthetics would be less hazardous in epidural anesthesia. Spinal anesthesia was used in an 18-year-old patient in an emergency condition after the labor to remove the placenta manually. Fentanyl and plain bupivacaine were used without any complication (308).

Another anecdotal report from Japan describes the use of general anesthesia for a 39-year-old male with hereditary spastic paraplegia (161). They avoided muscle relaxants and narcotics and used N2O, oxygen, and sevoflurane without any complication.