Hereditary spastic paraplegias (HSPs) are characterized by various inherited disorders in which weakness and spasticity of the lower extremities are the predominant symptoms [1]. HSPs caused by ALDH18A1 mutations have been reported as hereditary spastic paraplegia type 9 (SPG9), i.e., SPG9A and SPG9B, which are distinguished by the mode of inheritance (OMIM 601162, 616586) [2, 3]. Patients with SPG9 present with a pure or complicated form of HSPs. To date, SPG9A and SPG9B have been reported in five and two families, respectively [2, 3]. Since the number of SPG9B cases is small, the clinical features in SPG9B has not been established sufficiently. Here we report the clinical and genetic findings in SPG9B patients with novel ALDH18A1 mutations in two Japanese families.

Clinical and genetic study

In this study, 172 HSP patients exhibiting autosomal recessive transmission or a sporadic status were recruited through the Japan Spastic Paraplegia Research Consortium (JASPAC). This study was approved by our institutional review board, and informed consent was obtained from all participants. We performed whole-exome analysis on 172 HSP patients to find ALDH18A1 mutations and confirmed the mutations by Sanger sequencing. Exon capture was performed using Sureselect Human All Exon V4 + UTRs Kit, followed by massively parallel sequencing using Illumina HiSeq 2000 (100 bp paired end). We aligned the exome data with Burrows-Wheeler Aligner and extracted single-nucleotide variations using SAMtools. We checked known HSP genes in whole-exome analysis (supplement 1). Then, we conducted a co-segregation study in one family with ALDH18A1 mutations. We referred NCBI Reference Sequence (NM_002860.3) for ALDH18A1. We evaluated the functional prediction of ALDH18A1 mutations by means of in silico algorithms using the PHRED-like CADD score [4]. Furthermore, we performed blood amino-acid chromatography for two patients and the unaffected father in one family.

We found two families with ALDH18A1 mutations exhibiting autosomal recessive inheritance through JASPAC. The pedigrees of the two Japanese families are shown in Fig. 1a.

Fig. 1
figure 1

a Pedigree charts. Squares and circles indicate males and females, respectively. Filled symbols indicate affected individuals, whereas open symbols indicate unaffected individuals. Slash and double lines indicate diseased and consanguineous marriage, respectively. Red dots: individuals from whom genomic DNA were collected. b Brain MRI of patients (HSP48-II-1, HSP190-II-1, and HSP190-II-2). Two patients (HSP190-II-1 and HSP190-II-2) showed mild cerebellar atrophy

The HSP190 family contained two patients (HSP190-II-1 and HSP190-II-2) carrying two heterozygous ALDH18A1 mutations (c.1321 C > T, p.R441* and c.1994G > A, p.R665Q) (Fig. 1a). The father and mother carried a heterozygous ALDH18A1 mutation of p.R441* and p.R665Q, respectively (Fig. 1a). Therefore, we concluded that the two patients carried a compound heterozygous mutation in the ALDH18A1 gene.

The HSP48 family included two patients whose parents are related. The proband (HSP48-II-1) carried two homozygous ALDH18A1 mutations of c. 30 C > A/c.30 C > A, p.F10L/p.F10L and c.383 G > A/c.383 G > A, p.R128H/p.R128H (Fig. 1a). A DNA sample from his affected sister (HSP48-II-4) was unavailable.

These mutations were absent in an in-house control database consisting 1261 subjects and registered in ExAC (accessed on 3 April 2018) with extremely low-allele frequencies (c.30 C > A, 0%; c.383 G > A, 0.0033%; c.1321 C > T, 0.0016%; and c.1994G > A, 0.0008%). The amino acids p.R128 and p.R665 are evolutionally highly conserved among species. The PHRED-like CADD scores for three of these four mutations are more than 20 (Supplement 2) [4]. The mutations of p.R128H and p.R441* have been described as causing SPG9B and autosomal recessive cutis laxa type 3 A [OMIM 219150], respectively [2, 5]. The mutation of p.R665L, that affects the same amino acid as the mutation identified in this study, has been reported to be a pathogenic mutation in autosomal dominant SPG9A [2]. Thus, we concluded that the compound heterozygous mutation (p.R441*/p.R665Q) and the homozygous one (p.R128H/p.R128H) in ALDH18A1 are responsible in the SPG9B families, respectively.

Clinical information on our patients and reported SPG9B patients is presented in Table 1. The age at onset of patients with ALDH18A1 mutations ranged widely from birth to 32 years old. All of our patients with ALDH18A1 mutations exhibited intellectual disability. Cerebellar ataxia was observed in two patients (HSP190-II-1 and HSP190-II-2) showing mild cerebellar atrophy (Fig. 1b). None of the patients presented any symptom of cutis laxa (OMIM 123700, 219100, 219200, 612940, 613177, 614434, 614437, 614438, 616603, 617402, and 617403). The two patients in HSP190 family and their unaffected father who had a heterozygous nonsense mutation, showed normal serum amino-acid levels (Table 2).

Table 1 Clinical and genetic findings in patients in the present study and reported ones with SPG9B
Table 2 Amino-acid profiles

Furthermore, one patient (HSP190-I-2), who carried a heterozygous c.1994G > A mutation, suffered from not HSP but amyotrophic lateral sclerosis (ALS) and died owing to respiratory failure within 1 year from onset. She had dysarthria, tongue atrophy, positive head retraction reflex, exaggerated deep tendon reflexes, and amyotrophy at age 66. She exhibited giant motor unit potentials and chronic denervation potentials in the cranial nerve region, upper limbs, and lower limbs on electromyography. These results indicated the presence of upper and lower motor neuron degeneration in bulbar region and two spinal regions. Therefore, we diagnosed her as having definite ALS by the Awaji criteria. Her brain magnetic resonance imaging revealed unremarkable findings except for bilateral subdural hygromas and age-related atrophy. The clinical presentation of the mother was clearly distinct from that of HSPs.


In this study, we found two Japanese SPG9B families, this being the first report of SPG 9B in non-Caucasians. Mutations in ALDH18A1 have been reported to cause cutis laxa and pure or complicated spastic paraplegia [6, 7] (Supplement 3).

All three patients described in this study and two previously reported families with SPG9B presented intellectual disability, whereas only three of 15 patients with SPG9A presented it. We found cerebellar ataxia and cerebellar atrophy in one family, although cerebellar ataxia has not been reported in SPG9B previously. There was, however, no difference in the frequency of ataxia between SPG9A (n = 4/36) and SPG9B (n = 2/9) (P = 0.58, Fisher exact test).

ALDH18A1 encodes delta-5-carboxylate synthetase, which catalyzes the first two and common steps of de novo biosynthesis of proline and ornithine. The patients in the HSP190 family, however, showed normal amino-acid levels. According to earlier reports, patients with SPG9B do not show serum amino-acid abnormalities and ones with SPG9A tend to have serum amino-acid abnormalities (100%) [2, 3]. Meanwhile, as to patients with ALDH18A1-related disorders including cutis laxa and hypotonia, all 11 patients with heterozygous ALDH18A1 mutations and eight of 15 patients with homozygous or compound heterozygous mutations exhibited decreasing amino-acid profiles [2, 3, 6,7,8,9,10,11,12,13,14,15]. Furthermore, fluctuations in amino-acid profiles have also been reported [9]. Thus, although further investigation is needed to clarify the association between amino-acid profiles and ALDH18A-related diseases, amino-acid profiles seem to be not always useful for diagnosis of ALDH18A1-related disorders, especially the autosomal recessive form.

Interestingly, our SPG9B patients in the HSP190 family had a history of ALS. The mutation position (c.1994G) in the patient with ALS (HSP190-I-2) might be important, because a heterozygous mutation (c.1994G > T) has been reported to exhibit autosomal dominant inheritance [2]. Emerging evidence indicates that ALS and HSP share some causative or susceptible genes such as ALS2, SPG11, ERLIN1, VCP, KIF5A, and NIPA1 [16,17,18,19,20]. Thus, this family might suggest an association between mutations in ALDH18A1 and ALS. Further examination is required to elucidate the molecular mechanisms underlying ALDH18A1-related disorders [21, 22].