Hereditary Spastic Paraplegia

Hereditary spastic paraplegia (HSP) (also known as familial spastic paraparesis and Strumpell-Lorrain syndrome; McKusick number ^[1] 18260) is not a single disease entity but rather a group of clinically and genetically diverse disorders that share the primary feature of progressive, generally severe, lower-extremity spasticity. ^[2-20] These disorders are classified according to the mode of inheritance (autosomal dominant, autosomal recessive, and X-linked) and whether progressive spasticity occurs in isolation ("uncomplicated HSP'') or with other neurologic abnormalities ("complicated HSP''), including optic neuropathy, retinopathy, extrapyramidal disturbance, dementia, ataxia, ichthyosis, mental retardation, and deafness. HSP is not rare; during the past several years we have identified more than 80 unrelated HSP families.

Clinical features.

Uncomplicated autosomal dominant HSP has been reviewed recently. ^[8,19] Following normal gestation, delivery, and early childhood development, subjects develop leg stiffness and gait disturbance (stumbling and tripping) due to difficulty dorsiflexing the foot and weakness of hip flexion. Although the majority of patients experience symptom onset in the second through fourth decades, there is a wide range of age of symptom onset (from infancy through age 85). ^[19,21,22] Gait disturbance progresses insidiously without exacerbations, remissions, or saltatory worsening. Paresthesiae below the knees are not uncommon. Urinary urgency progressing to urinary incontinence is a frequent, although variable, late manifestation.

Neurologic examination of subjects with uncomplicated HSP reveals normal facial and extraocular movements and normal fundi. Although jaw jerk may be brisk in older subjects, there is no evidence of frank corticobulbar tract dysfunction. Upper-extremity muscle tone and strength are normal. In the lower extremities, muscle tone is increased at the hamstrings, quadriceps, and ankles and weakness is most notable at the iliopsoas, tibialis anterior, and to a lesser extent, hamstring muscles. Muscle wasting may occur in uncomplicated HSP ^[23-26] but in our experience is mild in patients with uncomplicated HSP and limited to atrophy of the shins in wheelchair-dependent elderly patients. Decreased perception of sharp stimulation below the knees is noted occasionally. Vibratory sense is often diminished mildly in the distal lower extremities. Slight terminal dysmetria is observed occasionally on finger-to-nose testing in older affected subjects. Deep tendon reflexes may be brisk (2 to 3+) in the upper extremities but are pathologically increased (3 to 4+) in the lower extremities. Gait demonstrates circumduction owing to difficulty with hip flexion and ankle dorsiflexion. Crossed adductor reflexes, ankle clonus, and extensor plantar responses are present uniformly. Hoffman's and Tromner's signs may be observed. Pes cavus is generally present and usually prominent in older affected subjects.

Currently, there is no specific treatment to prevent, retard, or reverse HSP's progressive disability. Nonetheless, treatment approaches used for chronic paraplegia from other causes are useful. Patients in relatively early stages of the illness have obtained symptomatic improvement with oral and intrathecal baclofen and oral dantrolene. Bladder spasticity has been improved with oxybutynin.

Electrophysiology.

Electrophysiologic evaluation of peripheral nerve, muscle, dorsal column, and corticospinal tracts is useful for characterizing the distribution of neurologic deficits in HSP. Although results of electrophysiologic studies are variable, a number of generalizations can be made. Most studies found nerve conduction studies to be normal (in contrast to Friedrich's ataxia and vitamin B12 deficiency). ^[27] One study, however, showed that subclinical sensory impairment was common in HSP, with involvement of peripheral nerves, spinal pathways, or both. ^[28] Lower-extremity somatosensory evoked potentials (SSEP) show conduction delay in dorsal column fibers. ^[29] Cortical evoked potentials (reviewed in Polo et al. ^[11]), used to measure neurotransmission in corticospinal tracts, show greatly reduced corticospinal tract conduction velocity and amplitude of evoked potential ^[11,29-31] in muscles innervated by lumbar spinal segments. In contrast, cortical evoked potentials of the arms are normal or show only mildly reduced conduction velocity. These findings indicate that there are decreased numbers of corticospinal tract axons reaching the lumbar spinal cord and that the remaining axons have reduced conduction velocity. Schady et al. ^[31] emphasized the variable results of cortical evoked potentials. Central motor conduction velocity in the upper extremities was normal except for all affected members of one HSP kindred for whom responses were considerably delayed. They conclude that measurement of central motor conduction velocity may be a useful way of identifying clinical subgroups of HSP. The autonomic nervous system has not been studied systematically in HSP.

Clinical variability.

The age of symptom onset, rate of symptom progression, and extent of disability are variable both within and between HSP kindreds. ^[3,11,19,28] As expected, phenotype variation is related, at least in part, to symptom duration. ^[19] In contrast to variable age of symptom onset and extent of disability, the distribution of neurologic deficits is invariant. With the exception of a trace degree of terminal dysmetria occasionally observed in older subjects, deficits are limited to corticospinal and dorsal column tracts subserving the lower extremities. Additional deficits such as visual disturbance, marked amyotrophy, fasciculations, dementia, seizures, or peripheral neuropathy in subjects from uncomplicated HSP kindreds should not be attributed to variant presentations of uncomplicated HSP. Rather, such subjects should be evaluated thoroughly for concurrent or alternative neurologic disorders.

We and others ^[21] have observed autosomal dominant uncomplicated HSP kindreds that exhibit onset of progressive spastic paraplegia in childhood (before age 6 years) and relatively little progression of symptoms beyond adolescence. These patients often do not experience urinary bladder disturbances and generally remain ambulatory (with assistance). While clinical heterogeneity may be significant, we are not aware of kindreds in whom some family members had childhood onset, relatively nonprogressive HSP and other family members had adolescent/adult onset, progressive HSP.

Neuropathology.

The major neuropathologic ^[2,8,14] feature of autosomal dominant, uncomplicated HSP is axonal degeneration that is maximal in the terminal portions of the longest descending and ascending tracts (crossed and uncrossed corticospinal tracts to the legs, fasciculus gracilis fibers, and to a lesser extent, spinocerebellar fibers). Neuronal cell bodies of degenerating fibers are preserved. ^[2] Mild loss of anterior horn cells may occur. ^[8] Dorsal root ganglia, posterior roots, and peripheral nerves are normal. ^[8] There is no evidence of primary demyelination.

The regional pattern of axonal degeneration in HSP is quite different from that seen in "system degeneration'' diseases, such as amyotrophic lateral sclerosis (ALS), which is characterized by degeneration of functionally and physically related motor system elements including pyramidal neurons, corticospinal tracts, anterior horn cells, and skeletal muscle. In contrast, axonal degeneration in uncomplicated, autosomal dominant HSP involves different classes of neurons (corticospinal tract fibers from pyramidal neurons in the motor cortex; fasciculus gracilus [and cuneatus to a lesser extent] from dorsal root ganglia neurons). The obvious feature shared by these degenerating axons is their length. ^[32] These are the longest axons in the CNS. Degeneration is maximal in the distal portion of these axons.

Differential diagnosis.

Diagnosing HSP is straightforward when the family history indicates inheritance of progressive spastic paraparesis as an isolated symptom. Documenting associated deficits (such as ataxia, retinitis pigmentosa, peripheral neuropathy, amyotrophy, and extrapyramidal signs) is the basis for classifying HSP as uncomplicated or complicated and helps to distinguish HSP from other neurologic disorders, including spinocerebellar ataxias and Machado-Joseph disease.

It is important to consider HSP a diagnosis of exclusion. Alternative disorders that should be excluded (summarized in Table 1) are most relevant when examining patients with atypical features and those with no family history of a similar disorder. Note that the differential diagnosis of HSP includes disorders for which specific treatments are available (e.g., vitamin B12 deficiency, DOPA-responsive dystonia, cervical spondylosis) and those whose prognosis differs significantly from HSP (e.g., familial ALS).

Genetic analysis.

Genetic penetrance, the frequency with which obligate HSP gene carriers exhibit the disorder, is age-dependent and nearly complete. (The age by which symptoms develop is variable between and within families.) This is an important consideration when counseling subjects who are at risk of inheriting mutant HSP genes. Segregation ratios in large, autosomal dominant HSP kindreds approach 0.5, as predicted for completely penetrant, autosomal dominant disorders. Among HSP kindreds with at least three generations of affected subjects, we did not find an instance of incomplete penetrance (an asymptomatic subject, older than the maximal age of symptom onset in the family, who had one affected parent and at least one affected child). In contrast, Cooley et al. ^[21] reported a kindred in which subject IV-9 is apparently an asymptomatic carrier of autosomal dominant, uncomplicated HSP.

HSP's apparently high degree of genetic penetrance may include an ascertainment bias. Autosomal dominant HSP is more readily diagnosed in large kindreds for whom the disorder is highly penetrant. This could contribute to overrepresentation of highly penetrant HSP kindreds in published studies. Genetic penetrance could not be determined in approximately 30% of the HSP kindreds we examined. For example, it is not possible to estimate genetic penetrance from kindreds consisting of affected siblings (with unaffected parents) since this could represent an autosomal recessive disorder or an autosomal dominant disorder with incomplete penetrance. Similarly, it is not possible to estimate penetrance from kindreds consisting of an affected child, an affected parent, and unaffected grandparents since this could represent either incomplete penetrance (in a grandparent) or a new mutation (in a parent).

Occasionally, HSP kindreds exhibit progressively younger age of symptom onset or increased disease severity in succeeding generations. Examples of such genetic anticipation has been reported in German ^[33] and Dutch ^[34] HSP kindreds. The possibility that ascertainment bias may contribute to apparent anticipation must also be considered. The infrequent occurrence of genetic anticipation among HSP kindreds suggests that trinucleotide repeat mutations probably are not responsible for HSP in the majority of kindreds. Disorders due to such mutations, including Huntington's chorea, Machado-Joseph's disease, fragile X, myotonic dystrophy, Kennedy syndrome, spinocerebellar ataxia type I, and dentato-rubro-pallidoluysian atrophy often exhibit genetic anticipation (see Plassart and Fontaine ^[35] for a recent review).

Genetic linkage analysis of uncomplicated HSP.

Genetic linkage analysis is making important contributions to our understanding of HSP. Hentati et al. ^[36] showed that uncomplicated, autosomal recessive HSP is mapped to chromosome 8q12-13; they showed that uncomplicated autosomal recessive HSP is genetically heterogeneous by excluding this locus in other autosomal recessive HSP families. ^[36] Nonallelic genetic heterogeneity indicates that the syndrome of HSP can be caused by mutations in different genes. Three loci for autosomal dominant HSP have been discovered in the past 18 months. Hazan et al. ^[22] found the first uncomplicated autosomal dominant HSP locus on chromosome 14q in a large French kindred. Subsequently, independent investigators identified linkage to this locus in unrelated German ^[33] and North American ^[37] HSP kindreds. Two additional loci for autosomal dominant HSP have been discovered thus far. Several groups working independently identified a locus on chromosome 2p that is linked to HSP in unrelated North American, French-Canadian, and Belgian kindreds. ^[38-40] Finally, Fink et al. ^[20] identified tight linkage between a locus on chromosome 15q and autosomal dominant HSP in a large North American HSP kindred.

X-linked uncomplicated HSP is rare and reported in only four pedigrees. ^[16,41-43] The disorder in two kindreds studied by genetic linkage analysis was shown to map to a locus ^[16,43] on Xq22. Cambi et al. ^[43] found a missense mutation in the proteolipoprotein (PLP) gene coding sequence in affected males of one family (K313), but not in the other family studied (K101). ^[43] Thus in one family uncomplicated X-linked HSP is allelic to complicated X-linked HSP ^[42,44] and Pelizaeus-Merzbacher disease, as they result from PLP gene mutations.

(Table 2) summarizes genetic linkage analysis of 22 unrelated HSP kindreds. We used the following microsatellite polymorphism to evaluate HSP loci on chromosome 2: D2S174, D2S400, D2S352, D2S367, D2S177; chromosome 8: D8S268, D8S166, D8S285, D8S279; chromosome 14: D14S75, D14S70, D14S306, D14S69, D14S269, D14S587, D14S66; chromosome 15: D15S128, D15S122, D15S156, D15S165. Criteria for assigning linkage (two point lod score <or=to-2.0) were used. In the case of two rather small kindreds, evidence of linkage to chromosome 2p and chromosome 14 was based on the test of heterogeneity showing a posterior probability of linkage >or=to0.90. A lod score of +2.41 was taken as evidence of probable linkage to Xq22 in another small kindred. Linkage to chromosome 2p was the most frequent among autosomal dominant HSP families linked to any chromosomal locus. Linkage to chromosome 2p has been observed in 15 of 33 kindreds (45%) we analyzed to date. Two kindreds (6%) show linkage to chromosome 14q. Linkage to chromosome 15q has been observed in a single kindred. Known HSP loci on chromosomes 2p, 14q, and 15q have been excluded in 15 autosomal dominant HSP kindreds. Thus the disorder is linked to additional, as yet unidentified, HSP loci in approximately 45% of autosomal dominant HSP kindreds.

Preliminary genotype-phenotype correlations.

With identification of HSP loci on chromosomes X, 2p, 8q, 14q, and 15q, it is possible to compare phenotypes in families for whom the disorder is linked to one of these loci as well as HSP families for whom these loci are excluded. Thus far, genetically diverse types of autosomal dominant HSP (those linked to chromosomes 2p, 14q, and 15q) are clinically and electrophysiologically extremely similar. This suggests that the different abnormal gene products may interact in a common biochemical cascade that results in similar patterns of neuronal degeneration. The disorder may be more severe in the kindred with chromosome 15q linkage compared with kindreds with chromosome 14q linkage. Only one subject required a wheelchair in the kindred with chromosome 14q linkage reported by Hazan et al. ^[22] In contrast, nine of the affected subjects in a kindred with chromosome 15q linkage HSP required a wheelchair (beginning for some in their 40s). ^[20] Autosomal dominant HSP kindreds with linkage to chromosome 2p have exhibited both the prototypical adolescent/adultonset, progressive form and the less common childhood-onset, relatively nonprogressive form of HSP. These significant variations in age of symptom onset and degree of progression in HSP kindreds linked to chromosome 2p indicate that the complete phenotype is influenced either by different mutations in the same gene or by the effects of modifying genes.

Previously, age of symptom onset was one variable used to subclassify ^[23] autosomal dominant uncomplicated HSP as type I (onset occurred before age 35, spasticity exceeded weakness, and progression was slow) and as type II (onset occurred after age 35; patients had weakness in addition to spasticity, mild distal sensory loss, urinary bladder disturbance, and faster progression). ^[3,23,45] We observed that the range of ages at which symptoms begin overlaps in autosomal dominant HSP kindreds with linkage to chromosomes 2p, 14q, and 15q. Assuming that kindreds with linkage to these loci were represented among those Harding ^[23] classified as type I and type II, there does not appear to be a genetic basis for HSP classification based entirely on age of symptom onset.

Genetic counseling.

Genetic counseling must consider that although autosomal dominant HSP often exhibits complete genetic penetrance, age of symptom onset, rate of symptom progression, extent of urinary bladder involvement, and functional disability may vary significantly. Although the age of symptom onset, rate of symptom progression, and functional disability may be stereotyped within families, significant intra- and inter-family variability is well known. One cannot predict with certainty that a wheelchair will be required or that urinary incontinence will inevitably occur. A cautious wait-and-see approach is advised. Genetic counselors and physicians must bear in mind that in some HSP kindreds, the condition begins in childhood and is relatively nonprogressive after approximately age 10. Patients in these kindreds often remain ambulatory.

The availability of genetic markers tightly linked with HSP loci (listed above) greatly improves genetic counseling but is applicable only to counseling informative subjects from kindreds with linkage to these loci who are at risk of transmitting or inheriting this disorder. Inheritance of tightly linked markers alone should not be used for early diagnosis of HSP, however. Each patient suspected of having HSP should undergo thorough neurologic examination and often neuroimaging to exclude other neurologic disorders. Although limited to informative members of kindreds with linkage to known HSP loci, tightly linked markers can be used prenatally to assess the chance that the fetus has inherited HSP.

Extreme caution must be used in counseling unaffected parents of a child with suspected HSP about the risk of HSP in other progeny. Genetic markers that flank known HSP loci cannot be used in this circumstance unless the unaffected parents were members of an extended kindred for whom the disorder was linked to a known locus. Assuming the diagnosis of HSP is correct, the affected child could represent a new mutation or nonpaternity (in either case the chance of HSP in other siblings is low), or incomplete penetrance (in which case there would be higher chance of HSP in other siblings). Also, one of the parents may be young and still at risk of developing HSP.

Conclusions and recommendations.

Acknowledgments

The authors gratefully acknowledge the participation of HSP patients and their families, the encouragement of Dr. Giovanna M. Spinella, Health Scientist Administrator, Developmental Neurology Branch, Division of Convulsive Developmental and Neuromuscular Disorders, NINDS, and the expert secretarial assistance of Ms. Lynette Girbach.