Genetic and clinical spectrums in Korean Charcot‐Marie‐Tooth disease patients with myelin protein zero mutations

Abstract Background Charcot‐Marie‐Tooth disease (CMT) is the most common disorder of inherited peripheral neuropathies characterized by distal muscle weakness and sensory loss. CMT is usually classified into three types, demyelinating, axonal, and intermediate neuropathies. Mutations in myelin protein zero (MPZ) gene which encodes a transmembrane protein of the Schwann cells as a major component of peripheral myelin have been reported to cause various type of CMT. Methods This study screened MPZ mutations in Korean CMT patients (1,121 families) by whole exome sequencing and targeted sequencing. Results We identified 22 pathogenic or likely pathogenic MPZ mutations in 36 families as the underlying cause of the CMT1B, CMTDID, or CMT2I subtypes. Among them, five mutations were novel. The frequency of CMT patients with the MPZ mutations was similar or slightly lower compared to other ethnic groups. Conclusions We showed that the median onset ages and clinical phenotypes varied by subtypes: the most severe in the CMT1B group, and the mildest in the CMT2I group. This study also observed a clear correlation that earlier onsets cause more severe symptoms. We believe that this study will provide useful reference data for genetic and clinical information on CMT patients with MPZ mutations in Korea.


| INTRODUCTION
Charcot-Marie-Tooth disease (CMT), also called hereditary motor and sensory neuropathy (HMSN), is a genetically and clinically heterogeneous group of progressive peripheral neuropathies characterized by distal muscle atrophy, weakness, and sensory loss. Through advances in next generation sequencing technology including whole exome sequencing and targeted gene panel sequencing, more than 130 genes have been reported to be implicated in CMT and other related disorders (Cortese et al., 2020;Gonzaga-Jauregui et al., 2015). CMT is commonly classified into three types: demyelinating type (called CMT1) with a reduced median motor nerve conduction velocity (MNCV) of <38 m/s), axonal neuropathy (called CMT2) with a preserved or slightly reduced MNCV of >38 m/s, and intermediate type neuropathy with a MNCV of 25-45 m/s (Saporta et al., 2011). It is known that CMT exhibits a loose genotype-phenotype correlation.
MPZ, which is strictly expressed in myelinated Schwann cells, encodes a transmembrane protein as a major component of peripheral myelin (Lemke & Axel, 1985). MPZ protein has an important role in cell adhesion and holding multiple layers of myelin sheets together tightly (Wong & Filbin, 1994;Xu et al., 2001). Several studies have reported that abnormal MPZ proteins are retained in the endoplasmic reticulum (ER) instead of being transported to the cell membrane or myelin sheath (Bai et al., 2018;Chang et al., 2019;Saporta et al., 2012;Scapin et al., 2020). Altered Mpz proteins were retained in the ER and arrested Schwann cell development in the CMT1B mouse models (Saporta et al., 2012;Scapin et al., 2020). Transgenic mice with extra copies of Mpz exhibited congenital demyelinating neuropathy in a dose-dependent manner (Wrabetz et al., 2000).
MPZ mutations have shown wide phenotypic variation in many studied populations. We identified 22 pathogenic or likely pathogenic MPZ mutations in 36 families from the Korean CMT cohort study. This study grouped the patients with the MPZ mutations according to phenotypes and then compared the clinical characteristics among the subtypes.

| Ethical compliance
This study was approved by the Institutional Review Boards of Sungkyunkwan University, Samsung Medical Center (2014-08-057-002), and Kongju National University (KNU-IRB-2018-62). Written informed consent was obtained from all the participants.

| Patients
This study was conducted with a cohort of 1,121 unrelated Korean CMT families. From the analysis of the copy number variation in the 17p12 region, 353 families were determined to be CMT1A (MIM 118220) with PMP22 (MIM 601097) duplication, and the remaining 768 families were further investigated to find the MPZ mutations.

| Clinical and electrophysiological examinations
Clinical and electrophysiological examinations were performed by the methods of Lee, Nam, Choi, Noh, et al. (2020). The strengths of the flexor and extensor muscles were measured using the standard Medical Research Council (MRC) scale. Physical disability condition was determined by the CMT neuropathy score (CMTNS) version 2 and the functional disability scale (FDS). Patients were divided into three categories: mild (CMTNS ≤10), moderate , and severe (CMTNS ≥21) according to the CMTNS values.
Motor and sensory NCVs were determined by surface stimulation and recording electrodes. MNCVs and compound muscle action potentials (CMAPs) of the median and ulnar nerves were measured by stimulating the elbow and wrist and the abductor pollicis brevis and adductor digiti quinti, respectively. The MNCVs and CMAPs of the peroneal and tibial nerves were measured by stimulating the knee and ankle and the extensor digitorum brevis and adductor hallucis, respectively. CMAP amplitudes were measured from the baseline to the negative peak values. Sensory nerve action potentials (SNAPs) were measured from the positive peaks to the negative peaks. Sensory nerve conduction velocities (SNCVs) were determined over a finger-wrist segment from the median and ulnar nerves by orthodromic scoring.

| Genetic study
Genomic DNA was extracted from whole blood samples using the QIAamp DNA Mini Kit (Qiagen, Hilden, Germany). For the patients who were negative for 17p12 (PMP22) duplication and deletion, the MPZ mutations were screened using whole exome sequencing (WES) or targeted sequencing of inherited peripheral neuropathy genes . The exome was captured using the SeqCap EZ v2.0 (Roche-NimbleGen, Madison, WI, USA) or the SureSelect Human All Exon 50 M Kit (Agilent Technologies, Santa Clara, CA, USA), and sequencing was performed by the HiSeq2000 or HiSeq2500 Genome Analyzer (Illumina, San Diego, CA, USA).

| Statistical analysis
All values are expressed as the median (interquartile range) or mean ± standard deviation (SD) in clinical and electrophysiological features respectively. The pairwise comparisons among the subgroups were evaluated by Student's t-test and one-way analysis of variance. Correlations were determined using Pearson's correlation coefficient (r). The statistical significance was determined at the level of p < 0.05.
Four of the five novel mutations are located at the extracellular domain of MPZ, and their crystal structures have been resolved . MPZ is expected to function as a homotetramer (Shapiro et al., 1996), and several clinically important mutations (p.R98C, p.R98H, p.T114M, and p.H81R) are putatively associated with the disruption of the tetramer assembly . According to the protein structure, these novel mutants might influence inter-and intra-tetramer interactions. According to the homotetramer model, F52 and S120 are located at the intermolecular interfaces; therefore, the mutations at these positions are presumed to weaken the tetramer assembly. The location of the mutation p.F52C is not desirable either because it adds another cysteine residue in the vicinity of the intramolecular disulfide bond (C50-C127, Figure 2a) that has a crucial role in the stability of the protein. Furthermore, the removal of the benzene ring that stabilizes the interaction between two β strands will be unfavorable ( Figure 2b). As for mutation p.S120del, 9 hydrogen bonds will be lost with that deletion (Figure 2c). Finally, mutations p.P132S and p.P133L are located at the proline-hinge; therefore, these mutations will alter the structure of the protein (Figure 2d,e).
Of the 20 trio families with father-mother-sibling(s), de novo mutations were observed in 9 families (p.C50Y in FC619, p.F52C in FC611, p.H81Q in FC1072, p.R98C in FC508, p.R98L in FC987, p.P132S in FC203 and FC572, p.L175del in FC455, and p.X249E*64 in FC1027) at a rate of 0.45. The frequency of CMT patients with the MPZ mutations was determined to be 3.2% in the total independent patients and 4.7% in the patients negative for PMP22 duplication (Table 3). These frequencies were similar with those of the China (3.3% and 6.4%) (Hsu et al., 2019) and Britain (3.1% and 5.1%) (Murphy et al., 2012), but were relatively lower than those of most other examined countries.
When the disease disabilities were determined by the measurement of CMTNS and FDS, the CMT1B group showed the most severe symptom, while CMT2I showed the mildest symptom with significant differences (CMTNS: p = 0.004, and FDS: p = 0.022). Although the CMTNS and FDS values of the CMT1B group were higher than those of the CMTDID group, no significant differences were found. In the CMT1B group, based on the CMTNS, patients with moderate disabilities were the most common (53%) and then severe patients (31%) and mild patients (17%). In the CMTDID and CMT2I groups, mild patients were the most common with frequencies of 75% and 80%, respectively. When the correlation between onset ages and clinical disability were examined, earlier onset was significantly correlated with severe disability in both comparisons of onset vs. CMTNS (p = 0.002, Figure 3a) and onset vs. FDS (p = 0.015, Figure 3b).

| More decreased NCVs and action potentials in CMT1B
Motor and sensory nerve conduction studies were performed on the 48 CMT patients with MPZ mutations (Table  4). The mean median MNCV of the CMT1B patients was 12.2 ± 11.0 m/s, which was significantly lower than those of CMTDID (41.3 ± 3.1 m/s, p < 0.001) and CMT2I (46.0 ± 6.6 m/s, p < 0.001). The mean median SNCV (7.4 ± 11.7 m/s) of the CMT1B patients was also significantly lower than those of CMTDID (18.0 ± 25.5 m/s, p = 0.021) and CMT2I (34.4 ± 4.3 m/s, p < 0.001). Moreover, the peroneal MNCV and sural SNCV were significantly decreased in the CMT1B patients compared to those in the CMTDID or CMT2I patients. The median motor nerve CMAP (6.0 ± 5.6 mV) of the CMT1B group was largely decreased compared to those of the CMTDID (12.8 ± 1.9 mV, p = 0.033) and CMT2I patients (13.0 ± 4.3 mV, p = 0.011). The CMAP of the peroneal nerve and the SNAPs of the median and sural nerves were also significantly decreased in the CMT1B patients than those in the CMTDID and CMT2I patients. However, no significant difference was observed in the nerve conduction velocities and action potentials between CMTDID and CMT2I. The analysis of the Pearson's correlation between onset ages and electrophysiological values showed that earlier onset was significantly correlated with more severely decreased NCVs and action potentials, i.e., onset vs. median MNCV (p < 0.001, Figure 3c), onset vs. median CMAP (p < 0.001, Figure 3d), onset vs. sural SNCV (p < 0.001, Figure  3e), and onset vs. sural SNAP (p = 0.006, Figure 3f). These results are consistent with the result that when the onset age is earlier, the clinical disability is more severe.
CMT1B patients with MRI consisted of five males and ten females who were mostly in their first to third decades of life. Most patients demonstrated mild (grade 1 and 2) fat infiltrations in the calf leg and thigh muscles. However, among the impaired muscles, the anterior, lateral, and superficial posterior compartment muscles of the distal lower legs showed severe (grades 3 and 4) fat infiltrations in some patients. A 56 year-old female patient (FC1159) with p.H81R mutation showed almost total fat replacement of anterior and lateral compartment muscles of the lower leg along their entire length and anterior and posterior compartment muscles in the distal thigh also showed severe fat infiltration (Figure 4a). In a CMTDID patient who was a 77-year-old male (FC943) with the c.449-1G>T mutation, he had severe (grade 3 and 4) fat infiltration in nearly all compartment muscles of the distal lower leg and in the superficial posterior compartment muscles of the proximal lower leg (Figure 4b). However, the thigh muscles showed a relatively mild fat infiltration. In a CMT2I patient, a 53-year-old male (FC658) with the p.T124M mutation showed severe (grades 3 and 4) fat infiltration in the anterior, superficial, and deep posterior compartment muscles of the distal lower leg, while mild fat infiltrations were seen in the thigh muscles (Figure 4c). in five independent families, and the p.S78L and p.R98C mutations were observed in each of the three families. Since these mutations have also been reported several times in different ethnic groups (Nelis et al., 1994;Rouger et al., 1996;Song et al., 2006), these sites are suggested as mutational hot spots. The c.233C and c.292C are located at the CpG sites, which may be partly related to the frequent mutations. The c.449-1G>T and p.F52V mutations were particularly identified in two types of affected individuals (CMT1B and CMTDID for c.449-1G>T and CMT1B and CMT2I for p.F52V). Some previous studies have reported phenotypic heterogeneities for the same MPZ mutations (Mazzeo et al., 2008;Senderek et al., 2001). The rate of de novo mutations was determined to be 45.0% of the trio CMT families with the MPZ mutations. This rate is significantly higher compared to the de novo CMT1A cases in Korea (18.7%) (Lee, Nam, Choi, Noh, et al., 2020). Of the 22 MPZ mutations, five mutations were novel: a nonframeshift deletion (p.S120del), and four were missense (p.F52C, p.P132S, p.P133L, and p.Y220C) mutations. These novel mutations were cosegregated with the affected individual(s) within each family and were not reported in the Korean genome databases (KRGDB) or in global public human genome databases (such as 1000 Genomes project, gnomAD, and EVS). The novel mutation sites are well conserved among different animal species, and several in silico analyses predicted that all the missense mutations affect the protein structure. Furthermore, inspection of the mutations in the crystal structure suggests disruption of intra-molecular interactions and impaired MPZ tetramer assembly.
The frequency of the MPZ mutations was determined to be 3.2% in the total independent patients diagnosed with CMT and 4.7% in the patients negative for PMP22 duplication. The frequencies of CMT patients with MPZ mutations are different depending on the ethnic groups. The frequency of CMT patients with the MPZ mutations from the total Korean cases was similar with Chinese (3.3%), British (3.1%), and Russian patients (3.5%) (Hsu et al.,  (  2019; Mersiyanova et al., 2000;Murphy et al., 2012). However, the Korean frequency was somewhat lower than those in most of the other examined countries: from 4.0% in Austrians to 6.0% in Norwegians (Fridman et al., 2015;Gess et al., 2013;Manganelli et al., 2014;Milley et al., 2018;Miltenberger-Miltenyi et al., 2009;Østern et al., 2013;Silander et al., 1998;Sivera et al., 2013;Yoshimura et al., 2019). When we compared the frequencies of patients with the MPZ mutations among patients excluding CMT1A, the frequency of Koreans was also lower than those of most other populations. When disease disability was compared among different groups, the degree of severity was determined to be in the order of CMT1B, CMTDID, and CMT2I, based on the CMTNS and FDS. Similarly, the average onset ages were also earlier in the same order. In most of the nerves examined, the NCVs and action potentials were significantly decreased in the CMT1B group compared to the other two groups. Those results were similar in previous reports (Hattori et al., 2003). The CMTDID group showed slightly lower electrophysiological values compared to the CMT2I group, but there was no significant difference.
In the affected individuals with the MPZ mutations, significant correlations were found between the onset ages and clinical phenotypes. When the correlation between onset ages and clinical disability were examined, an earlier onset was positively correlated with worse symptoms in both CMTNS and FDS. Significant correlations were also found in motor and sensory NCVs and action potentials: when the onset was earlier, the values were more decreased.
MRI analyses revealed varying degrees of intramuscular fat infiltration in the lower extremity muscles of MPZ patients. CMT1B-type patients mostly showed mild (grades 1 and 2) fat infiltration in both the thigh and lower leg. In comparison of a 56 year-old CMT1B patient with CMTDID patient (77 year-old) and CMT2I patient (53 year-old), the CMT1B patient showed more severe fat infiltration involving the distal thigh muscles. While CMTDID and CMT2I patients showed only mild fat infiltration. In the lower leg, the CMT1B patient showed prominent predilection for anterior and lateral compartment muscles, demonstrating total fat replacement of the muscles in their entire length. In contrast, the CMT2I patient showed severe fat infiltrations predominantly involving superficial and deep posterior compartment muscles. The CMTDID showed severe fat infiltration in nearly all compartment muscles of the distal lower leg. These difference in MRI findings may suggest difference of primarily affected muscle group among the three groups. Yet, further study including larger number of patients with diverse age is needed to clarify this difference.
In conclusion, our cohort study identified 22 MPZ mutations including five novel mutations as the underlying cause of the CMT1B, CMTDID and CMT2I subtypes. It seems that the frequency of the MPZ mutations was similar or slightly low compared to other ethnic groups. This study found that the median onset ages and clinical phenotypes were different according to the subtypes. We also observed a clear correlation that earlier onsets cause more severe symptoms. We believe that this study will provide useful reference data for the genetic and clinical information on CMT patients with MPZ mutations in Korea.

ETHICAL COMPLIANCE
This study was performed in accordance with the protocol approved by the Institutional Review Boards of Sungkyunkwan University, Samsung Medical Center (2014-08-057-002), and Kongju National University (KNU-IRB-2018-62). Written informed consent was obtained from all the participants.

ACKNOWLEDGMENTS
We would like to thank all the affected individuals and their family members who participated in this study. The crystal structure of the human MPZ protein was acquired from Protein Data Bank (http://www.rcsb.org), and the structures were visualized using its 3D view feature, Mol*. This work was supported by grants from the National Research Foundation (2019R1A2C1087547, 2020M3H4A1A03084600, and 2021R1A4A2001389) and the Korean Health Technology R&D Project, Ministry of Health and Welfare (HI14C3484 and HI20C0039), Republic of Korea.

CONFLICT OF INTEREST
The authors declare no competing interests.

AUTHOR CONTRIBUTIONS
Byung-Ok Choi and Ki Wha Chung planed and supervised this study. Hye Jin Kim, Soo Hyun Nam, Hye Mi Kwon, and Si On Lim performed molecular genetic works. Soo Hyun Nam, and Jae Hong Park performed clinical works. Hye Jin Kim, Kyung Suk Lee, Ji Eun Lee, and Ki Wha Chung interpreted genetic data and statistical analysis. Sang Beom Kim, and Byung-Ok Choi collected the participants' samples and information. Hye Jin Kim, Byung-Ok Choi, and Ki Wha Chung wrote the manuscript. All the co-authors read and approved the final version of the manuscript.

DATA AVAILABILITY STATEMENT
All raw genetic and clinical data generated or analyzed during this study are available upon request to the corresponding author.