Results

Identification of MME Mutations
Figure 1 shows a flow chart for the selection of patients, genetic tests, and the filtering process for homozygous variants using the ESVD system. We extracted three kinds of homozygous mutations in the MME gene in 5 patients in the first case series of patients with CMT with no pathogenic mutations in known CMT or IPN genes. In P1 to P3, we identified the same homozygous mutation in the splice donor site [c.654+1G>A] in intron 7 of MME. A novel homozygous missense mutation [c.1861T>C, p.Cys621Arg] and a nonsense mutation [c.661C>T, p.Gln221X] were found in P4 and P5, respectively. On the other hand, we could not extract presumed compound heterozygous variants or other homozygous variants in MME from the WES data of 163 AR/sporadic cases in first case series, although we identified a novel heterozygous nonsense variant of unknown significant (c.264C>A, p.Cys88X) and a rare SNV (c.1489T>C, p.Typ497His). In the second and third case series of patients with CMT with no pathogenic mutations in known CMT or IPN genes, five additional recessive mutations in the MME gene were identified (Fig 2): [c.1231_1233delTGT, p.Cys411del], [c.439+2T>A], [c.439+2T>A and c.655−2A>G (compound heterozygous mutation)], [c.1817G>A, p.Trp606X], and [c.655−2A>G] mutations in P6 to P10, respectively (Table 1). Segregation analysis from the families of P1, P4, P6, P7, P8, and P10 revealed that the MME mutations were segregated with the disease in each family (Fig 3). Glutamine 221, cysteine 411, tryptophan 606, cysteine 621, and canonical GT‐AG nucleotides of the splice donor and acceptor junctions in the MME gene were highly conserved among species (Fig 4A–C). Notably, the p.Cys621Arg mutation was classified as “pathogenic” or “deleterious” using in silico analysis (Polyphen2 score = 1.00, SIFT score = 0.00, and PROVEAN score = −11.2 [cutoff = −2.5]), and the p.Cys411del mutation was also predicted as deleterious in the PROVEAN software (PROVEAN score = −17.4 [cutoff = −2.5]). MAF of all the validated mutations in public databases, including Exome Sequencing Project (ESP) and Exome Aggregation Consortium (ExAC), and in‐house database was ≤0.002 (Supplementary Table 2).
Table 1  Genetic, Clinical, and Laboratory Findings   a  ‘Motor’ indicates distal lower limbs weakness/atrophy or gait disturbance, and ‘Sensory’ indicates decreased superficial sensation or dysesthesia in lower limbs. b  Scores indicating manual muscle testing (MMT) grade in the distal lower limbs. c  Brain atrophy is evaluated by CT or MRI. d  Nerve conduction study of Median nerve. Ax = axonal; CH = compound heterozygous; CT computed tomography; dCMAP = distal compound muscle action potential (mV); De = demyelinating; DL = distal latency (ms); DTRs = deep tendon reflexes; Htz = heterozygous; Hmz = homozygous; MCV = motor conduction velocity (m/s); MMSE = Mini–Mental State Examination; MRI = magnetic resonance imaging; NA = not available (not scored or not examined); NCS = nerve conduction study.
Figure 3  Pedigrees and segregation analysis of MME mutations in 10 families. P1, P2, P4, P5, P7, and P10 have consanguineous parents. P3, P6, and P9 have unaffected parents with affected siblings, and P8 was sporadic. Autosomal‐recessive inheritance is presumed in all pedigrees. A DNA sample was available from individuals marked with an asterisk. Affected members have a homozygous or compound heterozygous mutation, whereas unaffected members were heterozygous or wild‐type carriers. Squares represent males and circles represent females. Filled symbols represent those affected with a similar phenotype. Oblique lines represent deceased family members. Black arrows indicate the proband (P1–P10). +/+ = homozygous for mutation; +/− = heterozygous; −/− = homozygous for wild type.
Figure 4  Localization and conservation of MME mutations, haplotype analysis, and RNA analysis. (A and B) Schematic representation of the MME gene and neprilysin. Red arrows indicate the location of mutations in the extracellular domain. N = N‐glycosylation sites; TM = transmembrane domain. (C) Conservation analysis. Glutamine 221, cysteine 411, tryptophan 606, cysteine 621, and canonical GT‐AG nucleotides (c.439+2t, c.654+1g, c.655‐2a) of the splice donor and acceptor junctions in the MME gene were highly conserved among species. (D) Agarose gel electrophoresis of cDNA fragments obtained from RT‐PCR of P1, his family member (IV‐4, V‐2, IV‐9, and V‐5), P3 with the c.654+1G>A mutation, P5 with c.661C>T mutation, and a normal control (NC). The P1, IV‐4 (affected), IV‐9 (affected), and P3 lanes showed a 231‐bp band, which is smaller than the 350‐bp band in the NC lane. The V‐2 and V‐5 (unaffected heterozygous carrier) lanes showed a 231‐bp band and a 350‐bp band. The P5 lane showed no band. (E) Agarose gel electrophoresis of cDNA fragments obtained from RT‐PCR of P8, P10 with the c.655‐2A>G mutation, and the NC. The P10 lane showed a 284‐bp band, which is smaller than the 350‐bp band in the NC lane. The P8 lane showed a 284‐bp band and a 350‐bp band. (F) Agarose gel electrophoresis of cDNA fragments obtained from RT‐PCR of P7, P8 with the c.439+2T>A mutation, and the NC. The P7 lane showed a 263‐bp band, which is smaller than the 344‐bp band in the NC lane. The P8 lane showed a 263‐bp band and a 344‐bp band. (G– I) Sequence chromatogram of the RT‐PCR product from P1, P10, and P7 showing exon 7 skipping and premature termination within exon 8 (G), exon 8 skipping (H), and exon 5 skipping (I) as schematically shown in the lower panel, respectively. (J) Haplotype analysis in P1 to P3. Shared haplotypes are shown in the gray box.

Clinical Features of 10 Unrelated Patients With MME Mutations
P1, a 67‐year‐old man born to healthy consanguineous parents, had no relevant medical history. At aged 54 years, he first noticed that his flip‐flops slipped off easily because of foot drop and subsequently developed a slowly progressive gait disturbance and dysesthesia of the lower limbs. Neurological examination revealed severe weakness and atrophy of the distal limb muscles, especially the bilateral tibialis anterior muscles, which was grade 1 on the Medical Research Council scale.26 Superficial and deep sensations were decreased in the lower limbs. The cranial nerves were normal. Nerve conduction studies (NCSs) showed an axonal type of motor and sensory neuropathy: mild slowing of conduction velocity (> 38 m/s) in the median and ulnar nerve, and absent motor and sensory responses in the lower extremities (Table). On the basis of these findings, he was diagnosed with an autosomal recessively inherited axonal form of CMT. He had no obvious cognitive impairment; his MMSE score was 29/30 and his cognitive subscale of the Japanese version of the Alzheimer's disease assessment scale (ADAS‐J‐cog) score was 5/70. His brain MRI was almost normal (Fig 5A), and a PiB‐PET scan did not show a significant amount of amyloid deposition (Fig 5B).
Figure 5  Magnetic resonance imaging and Pittsburgh compound‐B (PiB) positron emission tomography imaging. (A) T2‐weighted images of P1 showing no significant brain atrophy at 67 years of age. (B) PiB standardized uptake value (SUV) images of PiB retention in P1 at 67‐years of age and a 11C‐PiB‐positive Alzheimer's disease patient. Images from P1 show a lack of PiB retention throughout the gray matter and nonspecific PiB retention in the white matter compared with that of the patient with Alzheimer's disease, which shows a high retention of PiB throughout the gray matter. SUVR = standardized uptake value retention.   We summarized the clinical features and electrophysiological findings of all 10 patients in the Table. Mean age of onset of disease was 47.2 years (range, 36–56). Six of them were born to consanguineous parents (Fig 3). Clinically, all patients had slowly progressive weakness and atrophy of distal limb muscles, gait disturbance (but not yet become wheelchair dependent), sensory disturbance of the distal limbs, and decreased or absent tendon reflexes, all of which were the typical CMT phenotype. No patients showed additional neurological findings, such as pyramidal signs, cerebellar ataxia, and other CNS symptoms. NCSs showed an axonal sensorimotor neuropathy in all patients, except for P5 who was electrophysiologically diagnosed with a demyelinating/intermediate form based on delayed median nerve conduction velocities (37.4 m/s). Nine of the patients evaluated had no apparent cognitive impairment as assessed by the MMSE; all scored 26 or higher. The remaining patient had no subjective memory complaints and no clinically apparent cognitive impairment. MRI or CT scans did not show cerebral atrophy in the 5 patients evaluated.

RNA Expression Analysis and Haplotype Analysis of MME 
RNA analysis obtained from 6 patients (P1, P3, P5, P7, P8, and P10) and a P1 family member revealed abnormal transcripts. In P1 and P3 with the c.654+1G>A mutation at the splicing donor site of intron 7 of the MME gene, agarose gel electrophoresis of the RT‐PCR products obtained with primer pair 1 showed a band smaller than the 350‐bp band observed in the healthy normal control (NC; Fig 4D). Sequences of the RT‐PCR products in P1 and P3 revealed aberrantly spliced mRNA lacking exon 7, resulting in a 231‐bp fragment (Fig 4G). The lack of exon 7 creates a frameshift and, consequently, premature termination within exon 8; therefore, the c.654+1G>A mutation was designed as p.Gly179AspfsX2 at the protein level. In P5 with the c.661C>T nonsense mutation, RT‐PCR products were not detected (Fig 4D), suggesting the occurrence of nonsense‐mediated decay of MME mRNA. In P8 and P10 with the c.655−2A>G mutation at the splicing acceptor site of intron 7, the RT‐PCR products showed a smaller band than the 350‐bp band of the NC (Fig 4E). The sequences of the RT‐PCR products in P10 revealed aberrantly spliced mRNA lacking exon 8, resulting in a 284‐bp fragment (Fig 4H). The lack of exon 8 lead to an in‐frame deletion of 22 amino acids from the catalytic (extracellular) domain of NEP (p.Ile219_Glu240del). In P7 and P8 with the c.439+2T>A mutation at the splicing donor site of intron 5, the RT‐PCR products obtained with primer pair 2 showed a band smaller than the 344‐bp band of the NC (Fig 4F). Sequences of the RT‐PCR products revealed aberrantly spliced mRNA lacking exon 5, resulting in a 263‐bp fragment (Fig 4I). The lack of exon 5 lead to an in‐frame deletion of 27 amino acids from the catalytic domain of NEP (p.Asp120_Glu146del; Asp120Ala).
Haplotype analysis in 3 patients (P1–P3) with the same homozygous c.654+1G>A mutation revealed that P1 and P3, but not P2, shared the same haplotype for the two closest markers (D3S3509 and D3S1275) and four SNPs (rs12493885, rs12497267, rs9438, and rs358733; Fig 4J).

Histopathological Findings and Expression of NEP
An immunohistochemical assay with an anti‐NEP/CD10 antibody showed expression of NEP in the germinal center of human tonsil tissue, in which NEP is highly expressed on proliferating B cells, confirming the specificity of the antibody against NEP (Fig 6A). Using this antibody, we confirmed the obvious expression of NEP in the myelin sheath of the sural nerve from the control patient, although it was also slightly expressed in the axon (Fig 6B). We obtained sural nerve biopsies from P1 and P4. Both showed a remarkable decrease in the density of large myelinated fibers with thin myelin sheaths and clusters of myelinated fibers (Fig 6C,E). Inflammatory cells and onion‐bulb formation were absent. Immunohistochemical staining with an anti‐NEP/CD10 antibody showed a negative and a partial positive reaction in P1 and P4, respectively, compared to myelin of the sural nerve from the control patient (Fig 6D,F). In addition, the antibody detected a band migrating at approximately 90kDa, corresponding to NEP, in the homogenates from the sural nerve of the control patient by western blotting, but the band was not detected in P1 (Fig 6G).
Figure 6  Neprilysin immunohistochemistry and western blot analysis. (A) Immunohistochemical staining with an anti‐NEP/CD10 antibody revealed the expression of NEP/CD10 (brown) in the marginal center of human tonsil. (B) Expression of NEP/CD10 (brown) was observed in myelin and axons, especially at the outer surface of myelin of the sural nerve from a control patient (inset). (C and E) Toluidine blue staining of a sural nerve. Densities of large myelinated fibers are markedly decreased in both patient 1 (C) and 4 (E), who harbor the c.654+1G>A (p.Gly179fs) and c.1861T>C (p.Cys621Arg) mutation, respectively. Clusters of small myelinated fibers are occasionally noted. (D and F) Expression of NEP/CD10 (brown) was not detected in the nerve from P1 (D), but partially detected in the nerve from P4 (F). (G) Western blot analysis of NEP/CD10. Blotted bands were detected with a rabbit anti‐NEP/CD10 antibody (arrowhead). NEP/CD10 was detected in homogenates from the sural nerve of the control patient, but not P1. White bar, 100 μm; black bar, 20 μm. CP = control patient; MW = molecular weight marker; P1 = patient 1; Ton = human tonsil.

Di