Neurogenetics J Neurol Neurosurg Psychiatry: first published as 10.1136/jnnp-2018-318839 on 26 September 2018. Downloaded from RESEARCH PAPER Genetic profile and onset features of 1005 patients with Charcot-Marie-Tooth disease in Japan Akiko Yoshimura,1 Jun-Hui Yuan,1 Akihiro Hashiguchi,1 Masahiro Ando,1 Yujiro Higuchi,1 Tomonori Nakamura,1 Yuji Okamoto,1 Masanori Nakagawa,2 Hiroshi Takashima1 1Department of Neurology and ABSTRact (CMT2D and distal HMN5A), HSPB1 (CMT2F Geriatrics, Kagoshima University Objective To identify the genetic characteristics in a and distal HMN2B) or IGHMBP2 (CMT2S and Graduate School of Medical and Dental Sciences, Kagoshima, large-scale of patients with Charcot-Marie-Tooth disease spinal muscular atrophy with respiratory distress 2–4 Japan (CMT). type 1), often produce other IPN phenotypes. To 2North Medical Center, Kyoto Methods From May 2012 to August 2016, we collected date, approximately 100 different genes have been Prefectural University of 1005 cases with suspected CMT throughout Japan, linked to CMT-like phenotypes (https:// neuromus- Medicine, Kyoto, Japan whereas PMP22 duplication/deletion were excluded in cular. wustl. edu/). Owing to its clinical complexity advance for demyelinating CMT cases. We performed Correspondence to and genetic diversity, the clinical subtyping of CMT Dr Hiroshi Takashima, next-generation sequencing targeting CMT-related gene is always laborious and difficult. Department of Neurology and panels using Illumina MiSeq or Ion Proton, then analysed The development of next-generation sequencing Geriatrics, Kagoshima University the gene-specific onset age of the identified cases (NGS) technology allows us to conduct gene Graduate School of Medical and and geographical differences in terms of their genetic panel sequencing simultaneously the targeting of Dental Sciences, Kagoshima spectrum. 890-8520, Japan; thiroshi@ m3. numerous genes. Within approximately 4 years, kufm. kagoshima- u. ac. jp Results From 40 genes, we identified pathogenic using two NGS systems successively, we have or likely pathogenic variants in 301 cases (30.0%). completed genetic assessment in more than 1000 Received 16 May 2018 The most common causative genes were GJB1 (n=66, Japanese cases with suspected CMT, which enables Revised 19 August 2018 21.9%), MFN2 (n=66, 21.9%) and MPZ (n=51, Accepted 26 August 2018 us to describe the genetic and clinical features of 16.9%). In demyelinating CMT, variants were detected copyright. Published Online First 26 these cases. September 2018 in 45.7% cases, and the most common reasons were GJB1 (40.3%), MPZ (27.1%), PMP22 point mutations (6.2%) and NEFL (4.7%). Axonal CMT yielded a relatively MatERIALS AND METHODS lower detection rate (22.9%), and the leading causes, From May 2012 to August 2016, we collected occupying 72.4%, were MFN2 (37.2%), MPZ (9.0%), HSPB1 (8.3%), GJB1 (7.7%), GDAP1 (5.1%) and MME blood or DNA samples from 1005 apparently unre- (5.1%). First decade of life was found as the most lated patients throughout Japan with suspected common disease onset period, and early-onset CMT CMT. These cases were examined by their local cases were most likely to receive a molecular diagnosis. neurologists or paediatricians and were referred Geographical distribution analysis indicated distinctive to our genetic laboratory for diagnostic genetic genetic spectrums in different regions of Japan. test. Duplication/deletion mutation of PMP22 was Conclusions Our results updated the genetic profile pre-excluded in all cases suspected with demye- http://jnnp.bmj.com/ within a large-scale of Japanese CMT cases. Subsequent linating CMT, using fluorescence in situ hybri- analyses regarding onset age and geographical disation or multiplex ligation-dependent probe distribution advanced our understanding of CMT, which amplification. would be beneficial for clinicians. On the basis of their family history, the included cases were grouped into sporadic (n=570, 56.7%), autosomal dominant (AD) or X-linked (n=341, 33.9%), autosomal recessive (AR; n=72, 7.2%) on October 2, 2021 by guest. Protected INTRODUCTION or with an unknown inheritance pattern (no clin- Charcot-Marie-Tooth disease (CMT) is the most ical data, n=22). All cases were further categorised common phenotype of inherited peripheral neuropathy (IPN), the latter of which also encom- as demyelinating CMT (n=282), axonal CMT pass hereditary sensory and autonomic neuropathy, (n=682) or unclassified type with no MNCV data hereditary neuropathy with liability to pressure or MNCV=0 (n=41) referring to their records of © Author(s) (or their palsy and hereditary motor neuropathy (HMN).1 In electrophysiological examination. employer(s)) 2019. Re-use Genomic DNA was extracted from peripheral permitted under CC BY-NC. No terms of median motor nerve conduction velocity commercial re-use. See rights (MNCV), CMT can be further classified into demy- blood using a Gentra Puregene Blood kit (Qiagen, and permissions. Published elinating CMT (MNCV<38 m/s) and axonal CMT Valencia, California, USA), according to the manu- by BMJ. (MNCV ≥38 m/s). facturer’s instructions. The protocol was reviewed To cite: Yoshimura A, Yuan CMT is typically characterised by progressive and approved by the Institutional Review Board of J-H, Hashiguchi A, et al. J motor and sensory polyneuropathy, but it may Kagoshima University. All cases and their family Neurol Neurosurg Psychiatry also present with significant clinical heteroge- members provided written informed consent to 2019;90:195–202. neity. CMT disease-causing genes, such as GARS participate in this study. Yoshimura A, et al. J Neurol Neurosurg Psychiatry 2019;90:195–202. doi:10.1136/jnnp-2018-318839 195 Neurogenetics J Neurol Neurosurg Psychiatry: first published as 10.1136/jnnp-2018-318839 on 26 September 2018. Downloaded from Targeted gene panel sequencing jp). We also checked the variants against our in-house whole- Primers were designed to cover the coding regions and exon/ exome sequencing database of individuals with non-IPNs. In intron junctions of genes in our CMT panel. Beginning in May silico analyses of variants were performed using SIFT (http:// 2012, we conducted mutation screening targeting 60 genes sift. jcvi. org), PolyPhen2 (http:// genetics. bwh. harvard. edu/ pph2), (online supplementary table 1) with the Illumina MiSeq platform PROVEAN (http:// provean. jcvi. org/ index. php), Mutation (Illumina, San Diego, California, USA). We used the same meth- Assessor (http:// mutationassessor. org) and Condel (http:// bg. upf. odology as the one employed in a previous study.5 We completed edu/ fannsdb). We completed the annotation process using the genetic analysis in 437 cases with this system, concluding in July CLC Genomic Workbench software and an in-house R script. 2014. All suspected variants were validated using Sanger sequencing In September 2014, a custom Ion AmpliSeq gene panel and interpreted according to the American College of Medical targeting 72 IPNs disease-causing or candidate genes (online Genetics and Genomics standards and guidelines.7 supplementary table 1) was designed and introduced. This panel consisted of 1800 amplicons divided into two primer pools. Library and template preparation was performed according to RESULTS the manufacturer’s instructions, and then run on the Ion Proton Genetic profile (Thermo Fisher Scientific, Waltham, Massachusetts, USA) Among the 1005 cases with suspected CMT, we detected patho- applying the Ion PI Chip kit v2/v3 BC (Thermo Fisher Scientific, genic or likely pathogenic variants in 301 cases (30.0%). The Carlsbad, California, USA). We used the same methodology as most common genetic causes in the mutation-positive cases were the one employed in a previous study.6 Using this platform, we GJB1 and MFN2, and each accounted for 21.9% (66 cases). executed genetic assessment in 568 cases until August 2016. Within MFN2, 40 types of reported and three novel variants (two pathogenic and one likely pathogenic) were identified. Data analysis and variant interpretation The following genetic causes were MPZ (n=51, 16.9%), HSPB1 We confirmed all previously reported pathogenic mutations with (n=14, 4.6%), PMP22 point mutations (n=13, 4.3%), GDAP1 reference to the Human Gene Mutation Database Professional (n=9, 3.0%), NEFL (n=9, 3.0%), MME (n=8, 2.7%), BSCL2 2017.3 (https:// portal. biobase-international. com/ hgmd/ pro). (n=6, 2.0%), MARS (n=6, 2.0%), DNM2 (n=5, 1.7%), SETX Moreover, we checked all variants against global databases, (n=5, 1.7%), SH3TC2 (n=5, 1.7%), PRX (n=4), GARS (n=3), including the 1000 Genomes (http://www.internationalgenome. IGHMBP2 (n=3), LRSAM1 (n=3), AARS (n=2), ARHGEF10 org), the Exome Sequencing Project (http:// evs. gs. washington. (n=2), FGD4 (n=2), SACS (n=2), SBF2 (n=2), TRPV4 (n=2) edu/ EVS) and the Exome Aggregation Consortium (http:// and TTR (n=2). Pathogenic or likely pathogenic variants were exac.broadinstitute. org/), as well as against Japanese databases, also detected in COA7, DCTN1, DHTKD1, EGR2, FBLN5, copyright. including the integrative Japanese Genome Variation Database GALC, GAN, HARS, HSPB3, HSPB8, INF2, KARS, MTMR2, (https:// ijgvd. megabank. tohoku. ac.jp) and the Human Genetic PRPS1, RAB7A and SOX10 in single cases (figure 1). Additionally, Variation Database (http://www. hgvd. genome. med. kyoto- u. ac. digenic variants were identified in five cases, which were variants http://jnnp.bmj.com/ on October 2, 2021 by guest. Protected Figure 1 Genetic spectrum of 301 cases with pathogenic or likely pathogenic variants. The following genes are indicated: GJB1 (21.9%), MFN2 (21.9%), MPZ (16.9%), HSPB1 (4.6%), PMP22 point mutation (4.3%), GDAP1 (3.0%), NEFL (3.0%), MME (2.7%), BSCL2 (2.0%), MARS (2.0%), DNM2 (1.7%), SETX (1.7%), SH3TC2 (1.7%), PRX (1.3%), GARS (1.0%), IGHMBP2 (1.0%), LRSAM1 (1.0%), AARS (0.7%), ARHGEF10 (0.7%), FGD4 (0.7%), SACS (0.7%), SBF2 (0.7%), TRPV4 (0.7%), TTR (0.7%), COA7 (0.3%), DCTN1 (0.3%), DHTKD1 (0.3%), EGR2 (0.3%), FBLN5 (0.3%), GALC (0.3%), GAN (0.3%), HARS (0.3%), HSPB3 (0.3%), HSPB8 (0.3%), INF2 (0.3%), KARS (0.3%), MTMR2 (0.3%), PRPS1 (0.3%), RAB7A (0.3%) and SOX10 (0.3%).