Background

Waardenburg syndrome (WS) is a hereditary disease characterized by sensorineural hearing loss and pigmentation abnormalities. WS has been classified into four subtypes on the basis of the clinical symptoms, namely, WS1 to WS4 [1]. Mutations in six genes (PAX3, MITF, EDN3, EDNRB, SOX10, and SNAI2) have been reported as the cause of WS [1].

Mutations in SOX10 were first reported to associate with WS4 and then later with WS2. WS4 can be accompanied by neurological symptoms, in which case it is called PCWH (Peripheral demyelinating neuropathy, Central dysmyelinating leukodystrophy, Waardenburg syndrome, Hirschsprung disease, OMIM: #609136) [2]. While WS2 can also be accompanied by neurological symptoms, only six cases have been reported to date [3,4,5]. Those reports suggest that the neurological symptoms that associate with WS are congenital and irreversible.

In this report, we describe the case of a patient with WS1 who harbored a de novo SOX10 heterozygous nonsense mutation. The patient had mild prodromal neurological symptoms that were followed by an attack of generalized seizures that associated with delayed myelination of the brain which recovered later.

Case presentation

A 5-month-old boy was referred to our hospital with the chief complaint of congenital deafness. No abnormalities were observed during his perinatal course. However, at presentation, his motor development was delayed: he was still unable to hold up his head. He also showed bilateral congenital horizontally narrow eyes with drooping eyelids (blepharoptosis), heterochromia iridis, and dystopia canthorum (W-index: 2.24) (Fig. 1a). Interview and visual inspection did not find musculoskeletal abnormalities or limits of mobility in the limbs. His defecation was normal and no intestinal signs indicating constipation, disorder of peristalsis, or obstruction were apparent by interview and auscultation. Oto-acoustic emission, conditioned orientation response audiometry, and auditory brainstem response revealed severe hearing loss (Additional file 1). Computed tomography of the temporal bone revealed hypoplasia of the semicircular canals and cochlea in the bilateral ears (Additional file 2). None of the members of his family had any of the symptoms found in the proband (Additional file 3).

Fig. 1
figure 1

Clinical features of the patient. a The face at 5 months of age. The patient had blue irides, blepharoptosis, and dystopia canthorum (W-index: 2.24). b-e Brain MRI findings at 8 and 18 months of age. At 8 months of age, namely, 3 months after the proband had a 4-min-long systemic clonic seizure, T1- b and T2-weighted images c demonstrated delayed myelination of the frontal lobe. In normal development, myelination of frontal lobe is observed at 3–4 months of age on T1-weighted imaging, and at 7–8 months of age on T2-weighted imaging. Follow-up MRI performed at 18 months of age demonstrated almost normal myelination on T1-weighted imaging d and T2-weighted imaging e

At 5 months of age, the proband developed an attack of generalized seizures. Initially, the seizures lasted for a few seconds and occurred once a day. In the following 2 weeks, both the frequency and duration of the seizures increased. 2 weeks after the seizures started, the proband had a systemic clonic seizure that lasted for about 4 min and the next day he had a longer clonic seizure with eye deviation to the right for about 12 min. Although obvious abnormal findings were not observed in the electroencephalogram, epilepsy was suspected. Sodium valproate was started 3 days after the 4-min-long seizure. Since then, the seizures have stopped for 3 years even after sodium valproate was stopped 2 years after the start of prescription. At the age of 12 months, he started holding up his head on his own, and, at 16 and 20 months, he started pulling up on things and walking along against a wall, respectively. Magnetic resonance imaging (MRI) performed at 8 months of age revealed delayed myelination of the frontal lobe on both T1- and T2-weighted imaging (Fig. 1b and c). However, follow-up MRI performed at 18 months of age demonstrated almost normal myelination (Fig. 1d and e).

Genomic DNA was extracted from whole blood and subjected to whole exome sequencing analysis using SureSelect Human All Exon V5 (Agilent Technologies, CA, USA). The captured DNA was subjected to the massively parallel sequencing system (HiSeq2000, Illumina, CA, USA). Paired-end read sequences were mapped and quality-checked with StrandNGS (Strand Genomics, CA, USA) using the reference sequence hg19/GRCh37. Among 88,273 variants detected in the patient and parents, 263 variants predicted to affect amino acid residues and detected in the patient with minor allele frequency (MAF) < 0.5% in public databases [6,7,8,9] were extracted and proceeded for segregation analysis within the family. Finally, eight variants of seven genes were selected as possible pathogenic mutations (Additional file 4). Of these, the SOX10 mutation (c.652G > T, p.G218*) was absent from public databases, i.e. novel. SOX10 has been previously reported as a responsible gene for PCWH, WS 2E (OMIM: #611584), and 4C (#613266). Since multiple nonsense mutations of SOX10 have already been reported to cause PCWH or WS2/4 [2], the mutation detected in this case was also considered a pathogenic mutation. The heterozygous nonsense mutation in SOX10 was in exon 3 (Additional file 5). Sanger sequencing validated the presence of the mutation in the proband, and the mutation was not present in his parents (Additional file 6).

Discussion

WS1 can be distinguished from WS2 by the presence of dystopia canthorum [1]. About 96% of cases of WS1 are due to PAX3 mutations [2, 3]. It has been suggested that about 3% of cases who were not associated with mutations in PAX3 had W-index scores that were diagnostic of WS1 [10]. Indeed, although our proband was clinically classified as having WS1, whole exome sequencing did not reveal any mutation in PAX3; instead, a novel SOX10 mutation was found. These findings suggest that SOX10 mutations are one of the causes of WS1.

Truncating mutations often inactivate gene function, either because they produce truncated protein products or because there is a significant decrease in cytoplasmic mRNA abundance by nonsense-mediated mRNA decay (NMD). When a truncating mutation occurs at a nucleotide located in the last coding exon or less than 50–55 nucleotides upstream of the last coding exon, the truncated mRNA is not recognized and the NMD is escaped [11]. Then, the mutant protein is synthesized and acts as a dominant negative protein that impairs the function of the wild-type SOX10, resulting in the neurological phenotypes [12, 13]. To date, the neurological symptoms associated with SOX10 have been reported only in the patients with truncating mutations in the last exon (exon 4 in NM_006941) of SOX10 [9]. The present patient had a novel nonsense mutation in exon 3, 45 nucleotides upstream from the 5′ end of exon 4. Thus, this is the first case to show the neurological phenotypes due to the nonsense mutation in exon 3 of SOX10 (Additional file 5).

Previous reports stated that the neurological symptoms of WS due to delayed myelination are congenital and irreversible [3,4,5]. However, the present report suggests that, while patients with a nonsense mutation in SOX10 that is proximal to the 3′ end of exon 3 can also exhibit delayed myelination, they may develop severe attacks of generalized seizures that can be improved because the immature myelination can recover later. Long-term follow-up is required for these patients because recurrence of seizure attacks may occur in the future. These findings suggest that they may be clinical characteristics of WS that arise from a nonsense mutation in exon 3 of SOX10. It is also possible to speculate that modifier genes that complement or repair immature myelination played a role. In conclusion, when patients of WS show mild prodromal neurological symptoms, such as delayed motor development and/or blepharoptosis, the clinician should be aware of the possibility that severe attacks of generalized seizures may follow, which may be associated with the truncating mutation in exon 3 of SOX10.