Clinical Heterogeneity of Autosomal Recessive Spastic Paraplegias: Title and subTitle BreakAnalysis of 106 Patients in 46 Families FREE

HEREDITARY spastic paraplegias (HSPs) are a heterogeneous group of genetic diseases characterized by slowly progressive spasticity of the lower limbs. Classically they are divided into pure forms (comprising patients whose manifestations are limited to corticospinal signs until later stages of the disease) and complicated forms (comprising patients whose manifestations are associated with signs of involvement of other systems, such as mental retardation; epilepsy; retinal degeneration or optic atrophy; cerebellar, extrapyramidal, or peripheral signs; and cutaneous lesions). We believe that "complicated" is not an appropriate designation and will therefore use the term "complex" for them. Dominant, recessive, and X-linked modes of inheritance have been described. For the dominant forms, which include mostly pure forms, 4 different loci have been identified: 14q11.2-24.3,¹ 15q11-12,² 2p21-24,^{3 - 4} and 8q23-24.⁵ The first 3 loci account for 40% to 45% of the families.^{6 - 7} CAG repeat expansions were reported⁸ in 6 Danish families linked to chromosome 2. For recessive forms, a linkage to the centromeric region of chromosome 8 was found in 4 Tunisian families.⁹ In 1998, an additional locus was identified on chromosome 16 in 3 late-onset families, 2 of which had a pure form and the other a complex form of the disease. All the affected members carry a deletion or mutations, which involve part of a gene encoding a nuclear metalloprotease.¹⁰ Recently, a new locus on chromosome 15q13-15 was suggested in 8 (of 9) American and European kindreds.¹¹ X-linked forms, although rare, have been linked to 3 different loci: Xq28,¹² Xq21,¹³ and Xq11.¹⁴

Recessive HSP can then be considered a rare and heterogeneous condition on clinical and molecular grounds. In this study, we analyze the clinical features of 106 patients belonging to 46 Portuguese and Algerian kindreds affected by autosomal recessive HSP, either pure or complex, with the goal of defining homogeneous groups that could serve as a guide for future molecular studies. Among the 46 families, 23 were informative and were analyzed with 10 markers (Généthon, Paris, France)¹⁵ spanning the candidate regions on chromosomes 8 and 16.

The 106 patients reviewed were personally examined by one of us (P.C., J.G., C.A., R.C., J.B., A.H., or R.Z.). They belong to 36 Portuguese and 10 Algerian families. The Portuguese kindreds were identified through a population survey of hereditary ataxias and HSPs,¹⁶ which was initiated in 1993. Informed consent was obtained from each family member.

The inclusion criterion was the existence, and predominance in the clinical expression, of progressive spasticity of the lower limbs. To ascertain a recessive mode of transmission, the inclusion criteria were as follows: (1) more than 1 individual affected in the same sibship, with or without parental consanguinity or isolated patients with parental consanguinity; and (2) normal parents (examined whenever available or not known to be affected from clinical history if death occurred at an advanced age). One Sjögren-Larsson kindred and 2 families with early-onset juvenile autosomal recessive amyotrophic lateral sclerosis as described by Hentati et al¹⁷ were excluded from the study. For pure forms, the diagnostic criteria were strictly clinical and adapted from Harding¹⁸ : progressive, spastic paraparesis with pyramidal signs in the lower limbs. The patients were not excluded from this group if they had muscular wasting of the legs when symptoms had been present for more than 10 years, since this could be due to long-standing immobility. The presence of upper limb hyperreflexia and of pseudobulbar signs (spastic dysarthria, dysphagia, and brisk jaw jerk) did not exclude the diagnosis of a pure form, since these symptoms correspond to the involvement of the same corticospinal tract at higher levels. Sensory and sphincter disturbances were only validated in patients without mental retardation. In multiplex families in which only one affected individual had signs of multisystem involvement, the whole family was classified as complex.

All affected persons in each family and other available and cooperative members of the kindreds were examined at the health centers or at home. Whenever possible, a proband from each family was fully investigated at the hospital, using a protocol that was designed to exclude other possible causes of the syndrome.⁷ Only 36% of the patients have been seen regularly (follow-up, 1-29 years). For most of them, however, only one examination was performed, at varying stages of the disease.

Disability was assessed on a 7-point scale: 1 indicates minimal disability (slight stiffness of the legs); 2, moderate disability (unable to run, but full autonomy); 3, severe disability in walking (reduced perimeter, frequent falls); 4, unilateral assistance required to walk; 5, bilateral assistance required to walk; 6, wheelchair bound; and 7, bedridden.

All probands underwent the following biochemical screening tests, based on the extensive list of differential diagnoses proposed by the HSP Working Group⁷ : blood and urine amino acids; chromatography of organic acids in the urine; serum lactate-pyruvate ratio; and arylsulfatases, galactocerebrosidase, long-chain fatty acids, human immunodeficiency virus, and human T-lymphotropic virus 1 antibody tests. Electromyographic and neurographic studies were performed whenever possible. Five patients with signs of lower motor neuron involvement underwent a muscle and nerve biopsy. Cranial magnetic resonance imaging (MRI) was performed to exclude white matter diseases, and spinal MRI was done in isolated forms without pseudobulbar signs.

The frequency of signs and symptoms in the different groups of patients was compared using the χ² test; for the identification of groups responsible for a significant χ² value, the adjusted residuals were analyzed. Since the sample sizes were not balanced for the different groups and most of the distributions for continuous variables (age at onset and age at examination) deviated significantly from the Gaussian distribution (P<.001 and P=.003, respectively, Wilks λ statistic), the Kruskal-Wallis test was used to study distributions across clinical types and multiple comparisons were performed using the Mann-Whitney test with adjustment of significance levels.

In 23 families, a total of 160 blood samples were collected for molecular analysis. DNA was extracted from whole blood using standard procedures. Polymerase chain reactions were carried out as previously described.¹⁹ Four amplification products, generated with separate primer sets on identical DNA samples, were pooled together and comigrated in a single lane of a 6% polyacrylamide denaturing gel. Separated products were then transferred to positively charged nylon membranes (Pall; ) and hybridized successively with nonradiolabeled (ECL; Amersham Corp, Arlington Heights, Ill) polymerase chain reaction primers as previously reported.²⁰

Two-point and multipoint lod scores were calculated using a package(LINKAGE²¹ ; or HOMOZ²² ), assuming an autosomal recessive FSP gene with a frequency of 10⁻⁵, equal female and male recombination rates, and an appropriate degree of consanguinity when required. Based on the clinical evaluation, we chose to estimate the pairwise and multipoint lod scores using a penetrance value of 100%. Marker allele frequencies were assumed to be equal.

The study includes 106 patients (60 males and 46 females). The median age at examination was 29 years (range, 4-77 years), with a median of 17 years' disease progression (range, 1-49 years). Table 1 summarizes the frequency of neurologic symptoms and signs for all patients. The median age at onset was 10 years (range, 1-58 years). The most frequent presenting symptoms were gait abnormalities, with leg stiffness in walking and abnormal wearing of the shoes. The family often noticed these problems before the patient was aware of them. For early-onset patients, a delay in walking and a tendency to walk on the toes were noticed first. In older patients, and after a long evolution, it was difficult to determine the precise age at onset; different ages were often reported in different interviews.

All patients had, by definition, hyperreflexia and spasticity of the lower limbs, while weakness was present in only 53.8%. Upper limb hyperreflexia, often associated with Hoffmann and Trömner signs, was present in 61.9% of the patients regardless of disease duration, and pseudobulbar signs in 30.5%; their frequency increased with the evolution of the disease. Pes cavus was a common finding (61.4%), especially in early-onset forms. Mental retardation was a less frequent finding (31.1%). In a few patients, particularly in those with long-standing motor disability, including dysarthria, and poor socioeconomic conditions, it was difficult to ascertain the true nature of the mental limitation. In other patients, profound mental retardation was evident and, therefore, taken into account in subsequent grouping of the families. Muscular wasting without fasciculations was observed in the upper limbs in 14 patients and in the lower limbs in 27 patients. Only upper limb amyotrophies were considered as indicative of peripheral motor neuron involvement. Sensory disturbances were rare, except for mild reduction of vibratory sense in the lower extremities; only one family (P31) had definite signs of posterior columns' involvement with loss of postural sensation in the lower limbs. The same was true for sphincter disturbances (urinary urgency and frequency), which were referred in 3 (4.1%) of the 73 patients without mental retardation.

Disability varied according to the duration of the disease and the family. It was also correlated with the presence of mental retardation and pseudobulbar signs. Another indicator of disability was the social handicap, as shown by the fact that most patients were never able to work, and only 18 were married and had children, mostly in the group with the late-onset pure form.

Most biochemical tests were used for exclusion of other disorders and consequently produced negative results in the patients described herein. Electromyography was normal in 23 (56%) of the 41 patients tested. Minor signs of anterior horn involvement were found in 10 patients, and peripheral neuropathy of axonal type was found in 8 patients. Neuromuscular biopsy specimens in 5 of these patients (P8, P9, P10, P18, and P29) confirmed the axonal neuropathy. There was no evidence of mitochondrial involvement in the muscle. Cranial MRI revealed leukodystrophy in 4 families and led to their exclusion from this study. The results were normal in all other patients, except for 2 belonging to the same family (P11), in whom hypoplasia of the corpus callosum was evident. Spinal MRI showed spinal cord atrophy, particularly at the cervical level, in 7 of the 12 patients studied.

The most relevant features of each family are summarized in Table 2 and Table 3. There was evidence of consanguinity in 27 kindreds (58%), of the first degree in 15. Twelve patients were isolated cases. As shown in Table 2 and Table 3, seven of them were males, while in 11 other kindreds, the 2 affected siblings were also males. The possibility that they could represent potential X-linked recessive forms was not excluded.

We considered that there was homochrony in a kindred whenever the variation of age at onset among siblings was less than 6 years. This homogeneous age at onset was verified in 70.6% of the kindreds, despite the tendency of the eldest affected member of the sibship to have a later onset. The family's awareness of the disease obviously leads to earlier detection of the first symptoms. For homotypy, most families showed a considerable homogeneity of the neurologic expression of the disease, but in several kindreds there was a variation among siblings, not related to the duration of the disease.

Fifty-three percent of the families had pure forms of HSP, whereas 47% had complex forms (Table 1). Furthermore, some subgroupings could be defined within the classic divisions (Table 4):

Ten markers from the map (Généthon), D8S268, D8S1828, D8S260, D8S1797, D8S543, D8S279, D16S3023, D16S3026, D16S413, and D16S3048,¹⁸ were initially selected and tested for linkage to the loci on chromosomes 8 and 16 in 23 of the families with autosomal recessive spastic paraplegia. Homozygosity mapping was performed in the consanguineous families using the HOMOZ package,²² and pairwise linkage analyses were carried out in the nonconsanguineous kindreds.

The locus on chromosome 8q was excluded in 22 pedigrees. A maximum 2-point lod score of 1.96 at θ=0 with marker D8S1797 was obtained in the consanguineous family A8 (Table 2) of Algerian descent. Since the maximum lod score was inferior to the threshold value of 3, a haplotype was then constructed using 11 markers spanning the candidate region: a homozygous region of 30 centimorgans encompassing but not reducing the locus previously defined in family A8. Likewise, the locus previously defined on chromosome 16q was excluded in 21 kindreds, and linkage to this locus was demonstrated for the remaining 2 families of Portuguese descent (P11 and P23, Table 2 and Table 3), both with complex forms of the disease. The localization recently reported on chromosome 15¹¹ was not excluded.

This study is the result of collaboration between 2 groups working with populations of different ethnic origins. An unusual feature is that most of the Portuguese cases were identified through a population-based survey.¹⁶ The fact that they were actively looked for probably explains the relative frequency of this syndrome, which is considered to be rare. In Portugal, the overall estimated prevalence of HSPs is 2.8 per 100,000 inhabitants (1.3 per 100,000 for dominant forms). These rates are in disagreement with the findings of Harding¹⁸ and Polo et al,²³ who considered autosomal dominant HSP to be more frequent (70%-80%) than autosomal recessive HSP. The highest prevalence for autosomal recessive HSP found so far in Portugal (9 per 100,000) occurs in a northeastern district.

There are not many large autosomal recessive HSP series published; most articles have described families with particular features. Several reviews of the literature are available, but they usually deal with dominant and recessive forms and are frequently limited to pure forms. In the 104 families reviewed by Holmes and Shaywitz,²⁴ the recessive forms represented 30% of all cases. Bell and Carmichael²⁵ reviewed 74 families with pure and complex forms, including 25 dominant and 49 recessive, which is closer to our findings. The series by Topaloglu et al²⁶ is interesting because it represents the neuropediatric point of view on autosomal recessive HSP, in opposition to ours, in which most patients were examined as adults. Because both series have short follow-ups, it is impossible to ascertain whether our findings represent the end point of what Topaloglu and coauthors observed at an earlier stage in the disease.

In the present series, pure forms are more frequent (58.7%) than complex forms (41.3%). This ratio is similar to that reported for dominant forms,^{18 ,24} and contrasts with the series of Topaloglu et al,²⁶ in which only 9 of 23 families were considered pure. Polo et al²⁷ reviewed 46 patients in 9 kindreds with pure HSP and found them "monomorphic and stereotyped." However, his series comprised only 2 autosomal recessive families including 6 patients, 2 with decreased vibration sense and 1 with absent ankle jerks.

Two groups can be defined within the pure form: an early-onset group (median, 7 years), the most frequent, and a late-onset group (median, 32 years), similar to autosomal dominant HSP, but with a higher rate of progression. They may represent distinct entities: besides the nonoverlapping distribution of the age at onset, the degree of disability is clearly worse in the second group for the same disease duration, and there is evidence of subclinical peripheral involvement in the late-onset group. This late-onset group had already been described by Harding¹⁸ and Meierkord et al.²⁸ In contrast to the pure forms, complex forms always begin before adulthood, as reported in previous publications.^{23 - 25} The functional prognosis is poorer in complex forms, particularly in the syndrome associated with mental retardation and neuropathy. The form with mental retardation has an earlier onset than the pure forms. The complex form with cerebellar ataxia was seen in 3 families; none of these affected individuals had mental retardation. Finally, there is a "miscellaneous" group, comprising several families with special features, such as macular degeneration, dystonia, parkinsonism, or position sense disturbances. One kindred had mental retardation, epilepsy, and peripheral neuropathy, and another, showing linkage to the chromosome 16 locus, corresponds to an entity described in 1994 by Iwabuchi et al,²⁹ associating HSP, mental retardation, and hypoplasia of the corpus callosum.

Autosomal recessive HSP is an exclusion diagnosis. Conditions to be ruled out are mainly white matter diseases, namely, the leukodystrophies; consequently, cranial MRI was the most useful diagnostic tool.

In the present series, there is a marked heterogeneity in the clinical presentation, as shown by differences in age at onset, rate of evolution, disability, and complexity of the neurologic picture. Even within the same phenotype, there are kindreds with different genotypes. The same happens with autosomal dominant HSP, in which kindreds with strong phenotype similarity showed linkage to different loci.⁷ The genetic heterogeneity demonstrated for dominant and X-linked forms of HSP will certainly be even greater in autosomal recessive forms.

Accepted for publication March 23, 1999.

This study was supported by Généthon, Paris, France; and 2 grants from the Portuguese Foundation for Science and Technology and the Portuguese Health Administration (projects STRDA/C/SAU/277/92 and PECS/C/SAU/219/95).

We thank the patients and families for participating in this study, their physicians for referring them to the hereditary ataxias and spastic paraplegias survey in Portugal, and Susan Cure for help in editing the manuscript.

Reprints: Paula Coutinho, MD, PhD, Division of Neurology, Department of Medicine, Hospital S. Sebastião, 4520 Santa Maria de Feira, Portugal.