Clinical and Molecular Characterization of BSCL2 Mutations in a Taiwanese Cohort with Hereditary Neuropathy

Background A small group of patients with inherited neuropathy that has been shown to be caused by mutations in the BSCL2 gene. However, little information is available about the role of BSCL2 mutations in inherited neuropathies in Taiwan. Methodology and Principal Findings Utilizing targeted sequencing, 76 patients with molecularly unassigned Charcot-Marie-Tooth disease type 2 (CMT2) and 8 with distal hereditary motor neuropathy (dHMN), who were selected from 348 unrelated patients with inherited neuropathies, were screened for mutations in the coding regions of BSCL2. Two heterozygous BSCL2 mutations, p.S90L and p.R96H, were identified, of which the p.R96H mutation is novel. The p.S90L was identified in a pedigree with CMT2 while the p.R96H was identified in a patient with apparently sporadic dHMN. In vitro studies demonstrated that the p.R96H mutation results in a remarkably low seipin expression and reduced cell viability. Conclusion BSCL2 mutations account for a small number of patients with inherited neuropathies in Taiwan. The p.R96H mutation is associated with dHMN. This study expands the molecular spectrum of BSCL2 mutations and also emphasizes the pathogenic role of BSCL2 mutations in molecularly unassigned hereditary neuropathies.

Seipinopathy is designated as a spectrum of neurologic diseases caused by BSCL2 mutations [15]. Although a number of cases of seipinopathy have been reported worldwide, only three mutations in BSCL2 were identified as cause for inherited neuropathy, including the p.N88S, p. S90L and p.S90W mutations [2][3][4]16]. The p.N88S and p.S90L mutations were demonstrated to compromise protein folding of seipin, provoke endoplasmic reticulum (ER) stress, induce cell toxicity, and cause dHMNV, Silver syndrome, or CMT2 [15,17]. The p.S90W mutation was recently identified in a Korean pedigree with CMT2 by genome-wide linkage analysis and whole exome sequencing [16]. The small number of BSCL2 mutations associated with inherited neuropathies implicates that the frequency of BSCL2 mutations in patients with inherited neuropathies may be underestimated. This point is further supported by two recent studies which utilized whole exome sequencing to investigate the genetic cause of inherited neuropathies with earlier unsuccessful candidate gene testing and both identified mutations in BSCL2 in their cases [18,19]. Currently, studies about seipinopathy in Han Chinese populations remain sparse, and BSCL2 mutations have rarely been screened in inherited neuropathy cohorts before. The aim of this study is to investigate the frequency and spectrum of BSCL2 mutations in a cohort of Taiwanese patients with genetically unassigned hereditary neuropathy. The clinical and molecular features of the identified BSCL2 mutations were also characterized.

Patients
Eighty-four unrelated patients of Han-Chinese origin with inherited neuropathies of unknown genetic cause, including 76 with CMT2 and 8 with dHMN, were recruited into this study. These patients were chosen from a continuous series of 340 unrelated patients with CMT and 8 with dHMN recruited from the Neurology Clinics of Taipei Veterans General Hospital. Among these patients, CMT2 and dHMN were diagnosed in 103 and 8 patients, respectively, according to the guidelines described in the report of the 2nd Workshop of the European CMT consortium [20]. Among the CMT2 patients, 27 have been already proven as having a pathogenic mutation [21,22,23]. The remaining 76 CMT2 patients with unknown genetic cause and the 8 dHMN patients received mutational analyses of BSCL2 in this study. A flow chart to clarify which patients were selected for sequencing analyses is available in the Supporting Information (S1 Fig). Peripheral blood samples were collected after obtaining the written informed consent from the patients or one of their parents for those younger than 18 years. The protocols for this study were approved by the Institutional Review Board of Taipei Veterans General Hospital. The individuals have provided written consent for the use of their information and videos as per the consent form for publication in a PLOS Journal.

Mutation analyses
Genomic DNA was isolated from peripheral blood leukocytes. Mutational analyses of BSCL2 was conducted by targeted sequencing in the same way as previously described [23]. In brief, a high-throughput targeted sequencing panel covering complete coding regions of BSCL2 was designed to screen the patients for BSCL2 mutations. This targeted sequencing panel also covers other 60 genes associated with inherited peripheral neuropathies (S1 Table). The enrichment of the targeted regions was performed using NimbleGen SeqCap EZ Choice Library system (Roche NimbleGen, Madison, WI). Enriched samples were sequenced on the HiSeq2000 platform (Illumina, San Diego, CA) using the paired-end 100bp protocol. All sequenced reads were mapped to the Human Genome version 19 (hg19/GRCh37). The Base-Space pipeline (https://basespace.illumina.com/) and the Illumina VariantStudio software (http://variantstudio.software.illumina.com/) were used to variant calling and annotate variants, respectively. Sanger sequencing was performed to confirm the identified BSCL2 variants. Amplicon sequences were aligned to the reference BSCL2 coding sequence (NM_032667.6). The pathogenic property of the novel BSCL2 variant was further assessed by in silico analysis with three programs, including Mutation Taster (http://www.mutationtaster.org) [24], Poly-Phen-2 (http://genetics.bwh.harvard.edu) [25] and SIFT (http://http://sift.jcvi.org) [26]. The putative pathogenic variant was further validated in 500 neurologically healthy individuals of Han Chinese origin recruited at our hospital and the two large genetic polymorphisms databases, Exome Aggregation Consortium (ExAC, version 0.3; http://exac.broadinstitute.org) and dbSNP (Build 144; https://www.ncbi.nlm.nih.gov/snp). Phylogenetic conservation of the mutated amino acid residue was analyzed by aligning the amino-acid sequences of seipin orthologs from several species using the UniProt website (http://www.uniprot.org) [27].
Western blot analysis and protein solubility and stability assay. HEK293 cells were transfected with WT BSCL2 or its mutant-expressing plasmids (p.S90L or p.R96H). Fortyeight hours post-transfection, cells were lysed in RIPA buffer supplemented with protease inhibitor cocktail (Merck Millipore, Darmstadt, Germany). Cell lysates were used for Western blotting and the protein concentration was determined using a Bradford protein assay kit (Bio-Rad, Hercules, CA).
To examine the protein solubility profile of seipin, sequential extractions were performed. Transfected cells were extracted with RIPA buffer and the cell lysates were centrifuged to generate the RIPA-soluble fraction. The resulting pellets were further re-solubilized with urea buffer (7M urea, 2M thiourea, 1% ASB-14, 40mM Tris, pH 8.5) to recover the most insoluble fraction. The protein from each fraction (40 μg) was separated on 8% SDS-PAGE, transferred to PVDF membranes, and immunoblotted with anti-BSCL2 antibodies (ab106793; Abcam, Cambridge, UK) followed by HRP-conjugated secondary antibodies. Detection was performed with a standard enhanced chemiluminescence assay (Perkin-Elmer Life Sciences, Boston, MA).
To determine the stability of the WT and mutant seipin proteins, cycloheximide-chase assays were conducted with cells transfected with different BSCL2 constructs. Twenty-four hours after transfection, transfected cells were trypsinized and re-seeded into 6-well culture plates. After a further 24 hours, cycloheximide (CHX; Sigma-Aldrich, St. Louis, MO) was added to a final concentration of 0.1 mg/ml. Cell lysates were harvested at the indicated time points and subjected to Western blotting with the anti-BSCL2 antibody. Actin was used as a loading control. The ratios of seipin to actin were calculated densitometrically.
Real-time quantitative PCR (RT-qPCR) and ER stress detection. RT-qPCR was performed to measure the mRNA levels of BSCL2 in cells transfected with different BSCL2 constructs. Total RNA of transfected cells was harvested 24 hours after transfection and extracted using the Qiagen RNeasy mini kit (Qiagen, Valencia, CA). The cDNA synthesis was carried out with the SuperScript1 III 1st Strand Synthesis Kit (Invitrogen). The RT-qPCR reactions were performed using a 7500 Fast Real-Time PCR System (Applied Biosystems, Thermo Fisher Scientific) in a 96-well format using the Fast SYBR1 Green Master Mix (Applied Biosystems). The relative gene expression was normalized against GAPDH expression. To investigate the effect of the BSCL2 mutations on ER stress, the mRNA expression levels of the ER stress biomarkers, BiP and CHOP, were examined in the transfected HEK293 cells by using RT-qPCR.
Immunofluorescence analyses. Forty-eight hours after transfection with the WT or mutant BSCL2 constructs, cells were fixed with 4% paraformaldehyde and then permeabilized in 0.2% tween-20 (Sigma-Aldrich, St. Louis, MO). After blocking of non-specific binding with 5% bovine serum albumin, the cells were stained for seipin using the anti-BSCL2 antibody conjugated to Alexa 488 (Thermo Fisher Scientific) together with DAPI counter staining of cell nuclei. Immunofluorescent staining was examined under an Olympus FluoView FV10i confocal laser scanning fluorescence microscopy system with a 60X oil immersion objective (Olympus, Tokyo, Japan).
Cell viability assay. Cell counting kit-8 (CCK-8) colorimetric assay (Dojindo Molecular Technologies, Rockville, MD) was used to assess the cell viability. HEK293 cells (8 × 10 3 cells per well) were grown in 96-well plates and transfected with different BSCL2 constructs. Fortyeight hours post-transfection, CCK-8 solution (10 μl) was added to growing cultures and incubated at 37°C for 2 h. The absorbance at 450 nm was determined by Multiskan™ FC Microplate Photometer (Thermo Fisher Scientific). Cell viability was expressed as a percentage of that of the cells transfected with WT BSCL2 expression plasmid.

Identification of the BSCL2 mutations
The BSCL2-targeted sequencing of the 84 patients with CMT2 or dHMN has an average coverage of 98.7% and read depth of 597X per targeted base. Two heterozygous missense variants in BSCL2 were identified, including p.S90L (c.269C>T; ch11:62469965G>A) and p.R96H (c.287G>A; ch11:62469947C>T) (Fig 1A and 1B). The first variant, p.S90L, which was identified in a son and his father with CMT2, has been recognized as a disease-causative mutation in multiple previous reports [2,3,18]. The p.R96H is a novel BSCL2 variant. It was identified in an apparently sporadic patient. Although the genomic DNAs of the patient's families were unavailable, the p.R96H is not found in 500 ethnically matched healthy controls and is present in only one African out of the 60,067 individuals of diverse ethnicities in the databases of ExAC. The p.R96H is also absent in dbSNP database. The 90th and 96th amino acid residues of the human seipin protein are highly evolutionarily conserved (Fig 1C). The BSCL2 p.R96H is also predicted as a pathogenic mutation by Mutation Taster, Polyphen-2 and SIFT programs ( Table 1).

Clinical information of the patient carrying the BSCL2 mutations
The clinical and electrophysiological features of the patients with BSCL2 mutations in this study were summarized in Table 2. Patient A-III.1 is heterozygous for the BSCL2 p.S90L

Remarks
All are predicted to be disease-causing All are predicted as probably damaging mutation ( Fig 1A). He had a slowly progressive distal muscle weakness and atrophy in the lower limbs since age 5 years and then in the hands since age 12. Neurological examination at age 23 revealed a pes cavus, generalized hypoactive deep tendon reflexes (DTRs) except for brisk knee jerks, symmetrical weakness and atrophy of the muscles in distal extremities, flexor plantar responses and a steppage gait (S1 Video). The Medical Research Council (MRC) scale scores were grade 4 for the intrinsic hand and feet muscles and dorsiflexors of his feet. Strength of other muscle groups was intact and all modalities of sensation were normal. The nerve conduction studies (NCS) demonstrated an axonal sensorimotor polyneuropathy (Table 1). His father, Patient A-II.4, also has a similar clinical phenotype since childhood and a BSCL2 p.S90L mutation (Fig 1A). Neurological examinations at age 49 revealed a pes cavus, generalized areflexia except for brisk knee jerks, bilateral flexor plantar responses, symmetrical weakness and atrophy of the muscles in distal extremities, and a steppage gait with mild spastic feature (S2 Video). The MRC scale scores were grade 4 for finger extensor and flexor muscles and muscles in the legs, and grade 2-3 for the intrinsic hand and feet muscles. The NCS demonstrated an axonal sensorimotor polyneuropathy ( Table 2). Patient B-III.1 is heterozygous for the BSCL2 p.R96H mutation. She is the only child of her family and had an insidious onset and gradually progressive foot drop since age 10. Neurological examination at age 19 revealed a pes cavus, generalized hypoactive DTRs except for normal knee jerks, flexor plantar responses, weakness and atrophy of the muscles in the legs, and a steppage gait. The MRC scale scores were grade 2-3 for the dorsiflexors of her feet and toes, and grade 5 for her plantar flexor muscles. Strength of other muscle groups was intact and all modalities of sensation were normal. The NCS revealed a motor neuropathy without sensory involvement ( Table 2).
In vitro analyses of the functional significance of the BSCL2 p.R96H mutations BSCL2 p.S90L has been well characterized in previous in vitro studies and was used as a reference here [15,17]. To investigate the impact of p.R96H mutation on the seipin function, we first utilized western blotting to evaluate whether the mutation influenced protein expression, and found that the steady-state level of the mutant R96H seipin was much lower than that of the WT protein (Fig 2A). Then, cycloheximide-chase assay was conducted and revealed that the p.R96H mutation did not compromise the seipin protein stability (Fig 2B). Further analysis of BSCL2 mRNA levels in HEK293 cells transfected with different BSCL2 constructs by RT-qPCR demonstrated that the steady-state mRNA level of p.R96H BSCL2 was significantly lower than that of WT BSCL2 (Fig 2C). Since both N88S and S90L mutants were shown to form intracellular inclusions, which were nearly undetectable in cells expressing the WT seipin [17], it was interesting to define the intracellular features of the mutant R96H seipin. We performed immunofluorescence analysis to visualize the intracellular patterns of WT or the mutant seipin proteins in HEK293 cells. Cells expressing WT seipin showed a seipin-specific reticular staining throughout the cytoplasm, whereas cells expressing the S90L or R96H mutants demonstrated large cytoplasmic aggregates with seipin-specific staining (Fig 3A).
Because the BSCL2 p.R96H mutation increases the aggregation tendency of seipin, we next examined the biochemical solubility of WT seipin and the seipin mutants expressed in HEK293 cells. The cells transfected with different BSCL2 constructs were extracted with RIPA buffer (soluble fraction) first and subsequently extracted with urea buffer to recover insoluble seipin proteins. Western blot analysis revealed that more than half of the S90L and R96H mutant proteins were found in the insoluble fraction of cell lysates, whereas the WT protein was detected predominantly in the soluble fraction (Fig 3B). These results indicate an obvious tendency of the S90L and R96H mutant seipin proteins to form detergent-insoluble aggregates, leading to a significant depletion of soluble seipin in the cytosol. To determine whether expression of R96H seipin causes cell toxicity, HEK293 cells were transfected with the WT or mutant BSCL2 constructs for 48 hours and then assessed the cell viability by CCK-8 colorimetric assay. As shown in Fig 4A, compared with WT seipin, both S90L and R96H seipin have a significant cellular toxicity. Moreover, we also investigated the effect of R96H seipin on the ER stress because both N88S and S90L mutants were reported to induce unfolded protein response and ER stress [17]. We analyzed the mRNA expression of two ER stress markers, BiP and CHOP, in HEK293 cells transfected with the WT or mutant BSCL2 constructs for 48 hours by RT-qPCR. HEK293 cells exposed to an ER stress stimulator dithiothreitol (DTT) or transfected to express S90L seipin were used as positive controls. As shown in Fig 4B, cells expressing R96H seipin did not have a significantly higher expression level of Bip or CHOP as compared with cells expressing WT seipin, indicating that the R96H seipin did not induce ER stress.

Discussion
In this study, we identified two heterozygous missense mutations in the BSCL2 gene after screening 84 unrelated patients with molecularly-unassigned CMT2 or dHMN for BSCL2 mutations. One is p.S90L in a family with CMT2, and the other is p.R96H in an apparently sporadic patient with dHMN. The p.S90L has been well documented as pathogenic before [2,3,18], whereas the p.R96H is a novel mutation. The pathogenicity of the BSCL2 p.R96H mutation is supported by the following facts. First, the BSCL2 p.R96H mutation is not present in the 1,000 control chromosomes from 500 neurologically healthy Taiwanese individuals. Second, the mutation is absent in the large genetic polymorphisms databases, dbSNP, and is present with an extremely rare allele frequency (1/ 120,134) in the ExAC database. Third, the BSCL2 p. R96H mutation occurs at an evolutionarily conserved amino acid residue of the human seipin and multiple computational predictive programs, including Mutation Taster, Polyphen-2 and SIFT, support its deleterious effect. Four, in vitro functional studies demonstrated that the p. R96H mutation results in a low protein expression level, increases the aggregation tendency of colorimetric assay revealed that expression of S90L or R96H seipin led to a significant reduction in viability of HEK293 cells. The percentage of the viable cells in HEK293 cells transfected to express wild-type seipin was set as 100% and used for comparison. Values are shown as means ± SEM of 24 independent transfections (**, p < 0.01). (B) The activation of ER stress was evaluated by quantitative real-time PCR (qRT-PCR), measuring the expression levels of BIP and CHOP mRNA. Cells exposed to DTT or transfected with S90L seipin expression plasmid represented the positive controls. All values (mean ± SEM, n = 4) were normalized to GAPDH mRNA levels (**, p < 0.01). Cells expressing R96H seipin did not have a significantly higher expression levels of BIP or CHOP mRNA compared with cells expressing WT seipin, suggesting that the R96H seipin did not induce ER stress.
Because the first two seipinopathy-related mutations, BSCL2 p.N88S and p.S90L, were originally identified in patients with dHMN or Silver syndrome [2,3], seipinopathy is commonly recognized as a spectrum of motor neuron disorders presenting with variable combination of upper and lower motor neuron signs. However, CMT2 was also found in a minor portion of patients with seipinopathy later. In a study of 90 patients from three families with the p.N88S mutation, 20% of the patients had CMT2 [4]. To date, the p.S90W mutation was only found in patients with CMT2 [16]. In our study, the p.S90L mutation was also identified in a father and son with CMT2. These findings emphasize the importance of considering the role of peripheral sensory involvement in seipinopathy and also underlie the contribution of BSCL2 mutations to CMT2.
Although the BSCL2 p.R96H mutation results in dHMN, a typical phenotype of seipinopathy, its molecular pathogenic mechanisms seem to be different from those of the two most common seipinopathy-related mutations, BSCL2 p.N88S and p.S90L mutations. Previous studies have shown that the p.N88S and p.S90L mutations enhance ubiquitination and degradation of seipin by the ubiquitin-proteasome system (UPS) and appear to induce seipin proteins misfolding [17]. These misfolded mutant proteins accumulate in the ER and then provoke ER stress leading to cell toxicity [17]. This study has demonstrated that the p.R96H mutation results in a profoundly decreased level of soluble seipin protein due to decreased BSCL2 mRNA expression and increased propensity to form insoluble aggregates. Expression of the R96H mutant seipin also reduced cellular viability without provoking ER stress. The discrepancies between the molecular pathogenic mechanisms of different mutations within the same gene are not rare in inherited neuropathy. For example, different mutations in GJB1 or MPZ may result in disparate intracellular trafficking patterns of the GJB1 or P 0 proteins, although they all cause CMT [28,29].
The prevalence of BSCL2 mutations in inherited neuropathy seems to be low. In previous studies, Murphy et al. identified BSCL2 mutations in one out of 425 patients with CMT and another one patient out of 61 HMN patients in their inherited neuropathy clinic [30]. Fridman et al. found that five out of 1,652 patients with inherited neuropathy had BSCL2 mutations (0.3%; 5/1652) [31]. Our study demonstrated that the frequency of BSCL2 mutations in Taiwanese patients with inherited neuropathy is low (0.57%; 2/348). Pooling the data from the above three studies shows that the frequency of BSCL2 mutations in patients with inherited neuropathy is approximately 0.36% (9/2486), indicating that BSCL2 mutations are an uncommon cause of inherited neuropathy.
In conclusion, BSCL2 mutations account for a small number of patients with inherited neuropathies in Taiwan. The BSCL2 p.R96H mutation is a novel cause of dHMN. In vitro functional study revealed that the p.R96H mutation results in a remarkably low seipin expression and reduced cellular viability. This study expands the molecular spectrum of BSCL2 mutations and also emphasizes the pathogenic role of BSCL2 mutations in molecularly unassigned hereditary neuropathies.