This single-center retrospective study analyzed data from children (aged ≤18 years) diagnosed with HSP in Kanagawa Children’s Medical Center, a tertiary pediatric hospital in Japan between January 1, 2012, and December 31, 2021. Patients with bilateral lower-extremity spasticity and an HSP diagnosis were included using a search for “hereditary spastic paraplegia” or “spastic paraplegia” against the institution’s medical records. Those with only lower-extremity spasticity were classified as pure-type, and those with lower-extremity spasticity and other neurological symptoms, such as seizures, cognitive impairment, ataxia, extrapyramidal symptoms, and hearing loss, were classified as complex-type. We excluded patients with systemic diseases, such as metabolic disorders and acquired central nervous system disorders. Patient medical data, including genetic analysis results, family medical history, clinical information (time of onset, modified Rankin Scale for Neurologic Disability [mRS] scores, neurological symptoms, and treatment and management), brain and spinal cord MRI findings, and electrophysiologic study results, were obtained from the institution’s medical records.

This study was conducted as per the tenets of the 2013 revision of the Declaration of Helsinki. It was approved by the Ethics Committee of Kanagawa Children’s Medical Center (approval number: 2101-5). The parents or guardians of the patients provided written informed consent to the participation of their children and the publication of this study.

Results of genetic analysis were obtained from the patient records. We used exome analysis, direct sequencing, or whole-exome sequencing for peripheral blood. For the pure-type with a family history, SPAST sequencing was performed first, and exome panel analysis or whole exome analysis was performed for negative cases. For the pure-type without a family history and all complex-type, exome panel analysis or whole exome analysis was performed. Trio analysis including both parents was performed in all cases. No patient underwent mitochondrial sequencing. Exome sequencing was performed using a TruSight One Sequencing Panel and the MiSeq platform (Illumina Inc., San Diego, CA, United States). Exome data alignment, variant calling, and variant annotations were assessed as previously described (14). Whole exome analysis was performed as previously described (15). Sequencing data were analyzed with the Burrows-Wheeler Alignment tool for mapping to a reference sequence, SAMTools for converting a SAM file to a BAM file, Picard for eliminating duplicate sequenced regions, the Genome Analysis Toolkit for variant calling, and ANNOVAR4 or SnpEff5 for variant annotation. Furthermore, to exclude common variants, we used data from the NHLBI GO Exome Sequencing Project,¹ Genome Aggregation Database,² 1,000 Genomes Project,³ Exome Aggregation Consortium, Human Genetic Variation Database,⁴ and jMorp.⁵ Out of all the called variants within exons or ± 10 bp from exon–intron, those registered in the above-mentioned database of common variants were eliminated with criteria of MAF 0.01 or higher. Variants were confirmed as true positives by Sanger sequencing. We used the Integrative Genomics Viewer for data visualization. Copy number variation analysis was performed using the log2-ratio of read depth on each exon as described previously and the eXome-Hidden Markov Model method.

The pathogenicity of the identified variants was evaluated according to the guidelines of the American College of Medical Genetics and Genomics (16). Variants were annotated with reference to the following transcripts: SPAST (NM_014946.4), KIF1A (NM_001244008.2), ALS2 (NM_020919.4), SACS (NM_014363.6), L1CAM (NM_001278116.2), and ATL1 (NM_015915.5). Microarray testing was performed with an Agilent SurePrint G3 Human CGH Microarray Kit 8 60 K (Agilent Technologies, Inc., Santa Clara, CA, United States), per the manufacturer’s instructions, to identify deletions and duplications (17).

Statistical analyses were performed using v.24 of SPSS for Windows (IBM Corp., Armonk, NY, United States) software. Comparisons of patients’ ages at the time of investigation and time of onset, and the mRS scores of children with pure- and complex-type HSPs were made with the Mann–Whitney U test. Fisher’s exact test was used to compare genetic diagnostic yields, family histories, brain and spinal cord MRI results, peripheral neuropathies, and intellectual disabilities. All the statistical tests were two-sided. p-values of <0.05 were considered statistically significant.

Of the 37 patients included in this study, 14 had a family history of HSPs, and 23 were without a clear family history. There were 20 pure-type patients with HSP in the cohort and 17 complex-type. Genetic data were available for 11 of the pure-type and 16 of the complex-type patients. Of these, five (45%) and 13 (81%) patients, respectively, had genetic diagnoses (p = 0.11) (Figure 1).

The patients had a median age of 11 years (interquartile [IQR] range, 5–13 years) at the time of investigation, and the median age of onset was 18 months (IQR 13–33). These did not differ significantly between the children with pure-type and complex-type HSPs (Table 1). The incidence of brain MRI abnormalities was significantly higher in children with complex-type than pure-type HSPs. None of the children with pure-type, and only one child with complex-type, HSP had any spinal cord abnormality. Among the 24 patients who underwent electrophysiologic testing, peripheral neuropathy was confirmed in seven (29%) overall, with no significant difference between the pure- and complex-type patients. The mean overall mRS score was 2.8 (±1.2), and the mean in complex-type patients was higher than that of pure-type patients (3.5 [±1.0] vs. 2.1 [±0.9], p < 0.001). Of the 17 children with complex-type HSPs, 13 (76%) had an intellectual disability. The most frequent neurological manifestations of complex-type HSPs were dystonia and epilepsy.

The most common treatment was rehabilitation therapy, followed by botulinum toxin type A administration, and surgical treatment (including lower limb tendon lengthening and muscle dissection) (Table 2). Baclofen or levodopa was prescribed to four patients (11%) to treat spasticity and dystonia, respectively.

Figure 2 and Table 3 show the identified causative genes and pathogenic variants. SPAST variants were seen in four pure-type and one complex-type HSP. ALT1 occurred only in pure-type HSPs. KIF1A, ALS2, SACS, and L1CAM variants were found only in complex-type HSPs. The 10p15.3p13 duplication was found in one patient with complex-type HSP.

The inheritance of SPAST was autosomal dominant and all variants found have been previously reported. The variants were diverse and included missense, nonsense, and intronic mutations, and one silent mutation. Four patients had a family history of HSPs. The only complex-type patient with a SPAST variant also had a SOX10 variant, which is the causative gene in Waardenburg syndrome. The patient also had unilateral heterochromia iridis, severe deafness, and Hirschsprung disease, all of which are symptoms of Waardenburg syndrome (18). The age of onset and severity of HSPs varied among the patients. In all patients with missense mutations, HSP onset occurred in the first year of life and there was a high degree of spasticity (mRS score 3–4). In a patient with silent mutation, the onset of HSPs was at 18 months, with mild symptoms, and an mRS score of 1.

All KIFIA variants showed autosomal dominant inheritance and were de novo. None of the four patients with this variant had a family history of HSP. One patient had a previously unreported missense mutation (308A > G) and the other three had previously reported missense mutations. All patients with KIFIA variants had complex-type HSPs. Cerebellar atrophy was observed in the three patients with MRI results, with two of these having undergone serial evaluations that had shown their HSPs to be progressive. The age of onset in all patients was ≤1 year and all had severe mRS scores (3, 4), and intellectual disabilities.

The inheritance of ALS2 variants was autosomal recessive. Two of the three complex-type patients with ALS2 variants were monochorionic diamniotic twins and shared previously unreported missense (c.470G > T) and frameshift (2325_2326del) mutations associated with their HSPs (Table 3). The same codon mutation has been reported in a patient with infantile-onset ascending spastic paraplegia (19). In addition to the spasticity of the lower extremities, symptoms included spasticity in the upper extremities and dysarthria. There was no intellectual disability. The third patient had a previously unreported frameshift mutation for both alleles, spasticity of the upper extremities, eye movement disorder, and dysarthria. Both phenotypes matched those of infantile-onset ascending HSP caused by an ALS2 variant (20, 21). All three patients had dystonia.

Inheritance of variants in SACS was autosomal recessive. The two patients with these variants were siblings with compound heterozygous unreported nonsense and frameshift mutations (Table 3). Both had pyramidal and cerebellar symptoms, and peripheral neuropathy, which are the three hallmarks of Charlevoix–Saguenay autosomal recessive spastic ataxia caused by SACS variants (22). Both exhibited pontine linear low-intensity lesions on T2-weighted MRI. This is characteristic of patients with SACS variants (23). Neither patient exhibited any deterioration in their condition and both were able to walk unaided.

Inheritance of L1CAM variants was X-linked recessive. This is the causative gene in X-linked hereditary hydrocephalus and X-linked spastic paraplegias (24). The age of onset in the two affected patients was 10–11 months and both had congenital hydrocephalus and severe motor dysfunction. One of the two had undergone surgical periarticular muscle dissection of the hip for severe spasticity. Both patients had severe intellectual disabilities and epilepsy.

The ATL1 variant occurred in one patient and showed autosomal dominant inheritance. This was diagnosed after the mother was affected. The child had pure-type HSP with mild spasticity and an mRS score of 2.

The 10p15.3p13 duplication we found was a previously unreported 12.3 Mb duplication. The affected patient had a mild intellectual disability. The patient’s brain MRI results showed thinning of the corpus callosum. The age of onset was 33 months, which was late compared to the onset ages of the other patients in this study. The patient’s mRS score was two, and the spasticity was mild.