OUP user menu

Phenotypic clustering in MPZ mutations

Michael E. Shy, Agnes Jáni, Karen Krajewski, Marina Grandis, Richard A. Lewis, Jun Li, Rosemary R. Shy, Janne Balsamo, Jack Lilien, James Y. Garbern, John Kamholz
DOI: http://dx.doi.org/10.1093/brain/awh048 371-384 First published online: 7 January 2004

Summary

Myelin protein zero (MPZ) is a member of the immunoglobulin gene superfamily with single extracellular, transmembrane and cytoplasmic domains. Homotypic interactions between extracellular domains of MPZ adhere adjacent myelin wraps to each other. MPZ is also necessary for myelin compaction since mice which lack MPZ develop severe dysmyelinating neuropathies in which compaction is dramatically disrupted. MPZ mutations in humans cause the inherited demyelinating neuropathy CMT1B. Some mutations cause the severe neuropathies of infancy designated as Dejerine‐Sottas disease, while others cause a ‘classical’ Charcot‐Marie‐Tooth (CMT) disease Type 1B (CMT1B) phenotype with normal early milestones but development of disability during the first two decades of life. Still other mutations cause a neuropathy that presents in adults, with normal nerve conduction velocities, designated as a ‘CMT2’ form of CMT1B. To correlate the phenotype of patients with MPZ mutations with their genotype, we identified and evaluated 13 patients from 12 different families with eight different MPZ mutations. In addition, we re‐analysed the clinical data from 64 cases of CMT1B from the literature. Contrary to our expectations, we found that most patients presented with either an early onset neuropathy with signs and symptoms prior to the onset of walking or a late onset neuropathy with signs and symptoms at around age 40 years. Only occasional patients presented with a ‘classical’ CMT phenotype. Correlation of specific MPZ mutations with their phenotypes demonstrated that addition of either a charged amino acid or altering a cysteine residue in the extracellular domain caused a severe early onset neuropathy. Severe neuropathy was also caused by truncation of the cytoplasmic domain or alteration of an evolutionarily conserved amino acid. Taken together, these data suggest that early onset neuropathy is caused by MPZ mutations that significantly disrupt the tertiary structure of MPZ and thus interfere with MPZ‐mediated adhesion and myelin compaction. In contrast, late onset neuropathy is caused by mutations that more subtly alter myelin structure and which probably disrupt Schwann cell‐axonal interactions.

  • MPZ; CMT1B; phenotype; myelin; PNS
  • CMAP = compound muscle action potential; CMT = Charcot‐Marie‐Tooth; CMTNS = CMT Neuropathy Score; MNCV = motor nerve conduction velocity; MPZ = myelin protein zero; NCV = nerve conduction velocity; PKC = protein kinase C; SNAP = sensory nerve action potential; PLP1 = proteolipid protein 1; TNS = total neuropathy score

Received April 22 2003. Second revision September 23, 2003. Accepted September 25, 2003.

Introduction

Mutations in the gene encoding the major PNS myelin protein, myelin protein zero (MPZ), cause an inherited demyelinating neuropathy, one of a group of neuropathies collectively called Charcot‐Marie‐Tooth disease (CMT). Individuals with MPZ mutations have a variety of clinical phenotypes, from a severe disease with onset of weakness and sensory loss in the neonatal period associated with very slow nerve conduction velocities (NCVs) called Dejerine‐Sottas syndrome, to a much milder disease with onset of symptoms in the 4th decade with only minimally slowed NCVs, which is sometimes referred to as CMT Type 2 (Warner et al., 1996; De Jonghe et al., 1999). MPZ mutations account for ∼5% of patients with CMT (Nelis et al., 1996).

MPZ, a transmembrane protein of 219 amino acids, is a member of the immunoglobulin gene superfamily. It has a single immunoglobulin‐like extracellular domain of 124 amino acids, a single transmembrane domain of 25 amino acids, and a single cytoplasmic domain of 69 amino acids (Lemke and Axel, 1985; Uyemura et al., 1992). In addition, MPZ is post‐translationally modified by the addition of an N‐linked oligosaccharide at a single asparagine residue in the extracellular domain as well by the addition of multiple sulphate, acyl and phosphate groups (D’Urso et al., 1990; Eichberg and Iyer, 1996). To date there are >95 mutations in MPZ known to cause peripheral neuropathy in patients, most of which are localized within the extracellular domain of the protein. A schematic diagram of the amino acid sequence of MPZ and its putative secondary structure with the mutations known to cause neuropathy is shown in Fig. 1. The phenotypes shown in Fig. 1 are those given by the reporting authors.

Fig. 1 Mutations in the open reading frame of MPZ associated with inherited neuropathies. Adhesive interface, four‐fold interface and head‐to‐head interface refer to amino acid residues deemed essential for cis and trans adhesion between adjacent myelin wraps. The numbering system for MPZ mutations does not include the 29 amino acid leader peptide cleaved before insertion in the myelin sheath. This figure is a modification from Shyet al., 2002. Clinical phenotypes are those utilized by the authors reporting the cases.

MPZ, like other members of the immunoglobulin gene superfamily, is a homophilic adhesion molecule (Filbin et al., 1990). Heterologous cells expressing MPZ adhere to each other in an in vitro cell interaction assay (Xu et al., 2001). In addition, crystallographic analysis of the MPZ extracellular domain demonstrates that it forms homotetramers within the plane of the membrane: each homotetramer, a doughnut‐like structure with a large central hole interacts in trans with a similar homotetramer on the opposing membrane surface (Shapiro et al., 1996). Furthermore, absence of MPZ expression in animals, such as mpz knockout mice, causes myelin to be uncompacted (Giese et al., 1992). Taken together, these data demonstrate that MPZ plays an essential role in myelination, probably by holding together adjacent wraps of myelin membrane through MPZ‐mediated homotypic interactions.

The cytoplasmic domain, as well as the extracellular domain of MPZ, is necessary for MPZ‐mediated homotypic adhesion. Deletion of 28 amino acids from the carboxy‐terminus of the protein abolishes MPZ‐mediated adhesion in vitro (Filbin et al., 1999) and nonsense mutations within the cytoplasmic domain cause particularly severe forms of demyelinating peripheral neuropathy in patients (Warner et al., 1996; Mandich et al., 1999). Interestingly, a protein kinase C (PKC) substrate motif, RSTK, is located between amino acids 198 and 201 of the cytoplasmic domain. Mutation of this sequence abolishes MPZ‐mediated adhesion in vitro and causes CMT in patients (Xu et al., 2001). Although the mechanisms by which the cytoplasmic domain is involved in homotypic adhesion are not known, MPZ probably participates in an adhesion‐mediated signal transduction cascade—similar to that of other adhesion molecules such as the cadherins and the integrins—by interacting with the cell cytoskeleton by way of its cytoplasmic domain.

MPZ also has a regulatory role in myelination, which is probably a consequence of the MPZ‐mediated signal transduction cascade. Consistent with this function, we have previously shown that the absence of MPZ expression leads to both the dysregulation of myelin‐specific gene expression and abnormalities of myelin protein localization in Schwann cells (Xu et al., 2000; Menichella et al., 2001).

In this study, we have analysed the clinical and electrophysiological findings of 13 patients from 12 kindreds with eight MPZ mutations seen in our CMT clinic, and have compared these data to those from 64 patients with MPZ mutations reported in the literature for which clinical and electrophysiological data were available. Consistent with the two functions of the protein, we find that the large majority of patients with MPZ mutations have one of two distinct phenotypes: (i) an early (prior to walking) onset disease with very slow nerve conductions in which dysmyelination is the major pathological feature, as suggested by Bird et al. (1997); and (ii) a late (adult) onset disease with only moderately slow nerve conductions in which axonal degeneration is the major pathological feature. Only occasional patients have the ‘classical CMT’ phenotype seen in most patients with CMT Type 1A (CMT1A) (Thomas et al., 1997; Krajewski et al., 2000). These results suggest that mutations which disrupt MPZ structure cause early onset neuropathy, while those which disrupt MPZ‐mediated signal transduction and Schwann cell axonal interactions cause late onset neuropathy. This clinical study of patients with MPZ mutations thus provides an important insight into the biological functions of the protein.

Methods

Patient ascertainment and evaluation

Patients with various forms of CMT were recruited from within and outside the Detroit area to be evaluated at the CMT clinic at Wayne State University. The patients gave informed consent to participate in the study which was approved by the Wayne State University Institutional Review Board. All patients with MPZ mutations, paediatric and adult, seen in the clinic are included in this study. Evaluations consisted of a neurological history and examination, and nerve conduction study. Genetic testing for all but one patient evaluated at the Wayne State University was performed by Athena diagnostics, utilizing PCR amplification and automated sequencing of both genomic DNA strands for exons 2 through 6 coding for the entire mature protein. The mutation in the remaining case was diagnosed previously and reported by the Lupski group at Baylor College of Medicine (Warner et al., 1996).

CMT neuropathy score

To evaluate the severity of the peripheral neuropathy in our patients, we developed a CMT neuropathy score (CMTNS) (Krajewski et al., 2003), based on the Total Neuropathy Score (TNS) developed by Cornblath et al. (1999). The score is based on a combination of symptoms, abnormalities on neurological examination and abnormalities of nerve conduction. In general, patients with severe disability (unable to walk for example) have scores >20, patients with moderate disability have scores between 10 and 20, and those with mild disability scores between 0 and 10.

NCVs

NCVs were performed by standard techniques utilizing either Nicolet Viking or Quest machines (Madison, WI, USA). Temperature was maintained at 34°C. Surface electrodes were used in all studies. Sensory conduction velocities were antidromic. NCVs were calculated by standard techniques.

Literature evaluation

Reports of patients with 95 MPZ mutations were obtained from our own records, from a search of PubMed, and from a search from the website, http://molgen‐www.uia.ac.be/CMTMutations/. All 64 reports which provided adequate clinical and neurophysiolgical data to characterize the phenotypes were included in our evaluation.

Results

Evaluation of patients with CMT1B

We identified and evaluated 13 individuals in our clinic from 12 kindreds with eight distinct MPZ mutations. Nine of the 12 kindreds had missense mutations within the extracellular domain of the molecule, while three of the kindreds had mutations within the cytoplasmic domain (one missense mutation, one truncation and one single amino acid deletion). The ages of our patients ranged from 5 to 71 years, and included five males and eight females. Three patients required the use of a wheelchair, but the remainder of the group were ambulatory and had not undergone surgery for foot deformities. The average CMTNS for the group was within the moderate range (16; moderate = 10–20), although four patients had mild disease (<10) and three had severe disease (∼30). Three of these patients have been briefly reported previously (Warner et al., 1996; Xu et al., 2001; Boerkoel et al., 2002). The His10Pro and Del 21–29 mutations have not been described previously. Because the initial 29 amino acids of MZP (encoded by exon 1) are cleaved prior to insertion into the myelin sheath, these 29 amino acids are not used in our numbering system. A summary of the epidemiological, genetic, neurophysiological and clinical data from our patients is shown in Table 1AA and B.

View this table:
Table 1A

Features of early onset neuropathy (CMT Clinic, Wayne State University)

PatientAge (years)Amino acid changeAge walkedNCVFamily historyNeuropathy score
123CYS21DEL4 yearsm2.6, u2.9, pNRDaughter32
250GLY94CYS3 yearsm4.3, uNR, pNRBrother27
35LYS101AEG22 monthsm16, uNO, pNONo9
417GLN186stop4 yearsm9, u9, p5No27

Two distinct phenotypes result from MPZ mutations

The clinical and electrophysiological data from this group of patients with MPZ mutations suggest they have two relatively distinct phenotypes: (i) an early (infantile) onset disease with very slow nerve conductions; and (ii) a late (adult) onset disease with only minimally to moderately slow nerve conductions. All three patients with severe neuropathy, for example, had delayed motor milestones during infancy and did not walk independently until after 3 years of age. Their NCVs, where they could be measured, were <10 m/s. A fourth patient, who was 5 years old, did not walk until 22 months of age and is currently unable to run. Her median motor nerve conduction velocity (MNCV) is 16 m/s. In contrast, the other nine patients all walked prior to 1 year of age and were able to run independently as children. In addition, their neurological signs and symptoms did not appear until after age 35 years, and their NCVs were only mildly slow at 30–40 m/s in the upper extremities.

To determine whether the onset and clinical course of the neuropathy in our 13 patients is similar to that of other patients with CMT Type 1B (CMT1B), we collected and analysed the data from the additional 95 published reports of patients with MPZ mutations. As shown in Fig. 1, many of these patients have been classified as having CMT1B, Dejerine‐Sottas disease, congenital hypomyelination or ‘CMT2’‐like neuropathy. However, authors have used variable criteria to make these diagnoses and, in many cases, provide no clinical information to support them. There were, however, adequate clinical and physiological data to interpret the phenotypes of individuals with 64 of the 95 known MPZ mutations. As shown in Table 2A and Fig. 2, 58 of these 64 patients (90%) could also be placed into one of our two phenotypic groups with respect to the age of onset of their symptoms and their NCVs.

Fig. 2 Mutations in the open reading frame of MPZ that are associated with early onset or late onset symptoms of neuropathy. Early onset cases are defined as those in which there was a delayed onset of walking or other early milestones. Late onset cases are defined as those in which the first symptoms of neuropathy occurred after the age of 21 years.

View this table:
Table 1B

Clinical features of late onset neuropathy (CMT Clinic, Wayne State University)

Patient Age(years)Amino acidchangeAge walked(years)FirstlimitationsNCVFamily historyNeuropathyscore
544ARG198SERNormal40–50 yearsm32, u30, p22Suspect sister, father, grandmother6
657HIS10PRONormal30–40 yearsm49, u52, pNRSuspect mother, cousins; sister of Patient 717
756Normal30–40 yearsm37, u37, p22Brother of Patient 612
870HIS10PRONormal40–50 yearsm26, u40, p28No17
949SER15PHENormal40–50 yearsm44, uNO, pNOSuspect father, sister, brother13
1051HIS10PRONormal40–50 yearsm24, u44, p34Suspect brother, daughter, father, mother,grandparents9
1157ARG69HISNormal50–60 yearsm19, u20, pNRNo15
1270HIS10PRONormal60–70 yearsm48, u46, pNRNo16
1332SER111THRNormal30–40 yearsm36, u38, p27No8

m = median; p = peroneal; u = ulnar.

NO=not obtained; NR=not recordable

View this table:
Table 2A

Features of early onset neuropathy (literature review)

Amino acid changeAuthordiagnosisFirst limitationsAge walkedClinical courseNCVFamily historyAuthors
Ile1MetDSSPrior to age 3 years3 yearsAFO at 5 yearsHayasakaet al., 1993
Thr5IleCMT1B3 yearsm8Gabreels‐Festen et al., 1996
Ile33PheDSS18–24 monthsSevere at age >26 yearsSiblingsNakagawa et al., 1999
Ser34PheCMT1BAge of walkingm13Mother (didn’t walkuntil 12 years old)Blanquet‐Grossard et al., 1995
Ser34CysDSS/CH3 yearsWorst possible clinical scoreOuvrieret al., 1987 andHayasaka et al., 1993
**Phe35del1B/DSS18 monthsSevereWarneret al., 1996
Tyr39CysEO–CMT1Age of walkingSevereSorour and Upadhyaya, 1998
Ser49LeuSevere 1BPrior to age 1 yearDelayedFatherSilanderet al., 1998
His52Arg1B/DSSDelayed20 monthsNever ranNRShizukaet al., 1999
Tyr53CysDSSDelayed at 9 months20 monthsSevere at 25 yearsu4.7, m6.2,t3.3
Gln56_Ile60del/insHisLeuPheDSSInfantile12 months but fallsfrequentlym 10SporadicSilanderet al., 1996
Asp61GluCMT1BDelayed in 6 patientsModerate5–15 m/sFour generationsBirdet al., 1997
Gly64GluCMT1BAbnormal gait at age7 years18 monthsm 6, u 5DominantIkegamiet al., 1997
Arg69SerCMT1BToe walking3 yearsAbnormal gaitu13, m11Warneret al., 1996
Arg69CysCMT1BDelayed2 yearsAbnormal gaitm7Warneret al., 1996
Arg69Cys1B/DSSHypotonia2 yearsAbnormal gaitp7.5, s5.1,m8, t4.9Komiyamaet al., 1997
Arg69CysCMT1BSevere at 10 monthsNeverDied at age 22 monthsu8.5Gabreels‐Festenet al., 1996
Arg69HisDelayedMeijerinket al., 1996
Try72CysCMT1BAge of walkingM 10Daughter similarLatour et al., 1995
**Val73fsDSSDelayed sitting prior to1 yearDelayedSevere, WCBNRBrother, parentsPareysonet al., 1999
DSSDelayed sitting priorto 1 year3.5 yearsSevere, WCBNRSisterWarneret al., 1996
Gly74GluHypotonia4–5 yearsWCB at age 10 yearsSlowedMother (germ linemosaicism)Fabriziet al., 2001
*Ile85Thr,DSSFloppy infant, sat at2 years>2 yearsWarneret al., 1997
*Asn87His,DSSWarneret al., 1997
Asp89Tyr90del/insPheTyrDSSDelayed at 7 months,DysphagiaNeverDied at 10 monthspNR, m‘delayed’Ikegamiet al., 1998
Thr95Phe96delDSSProximal distal weaknessat 13 yearsm 6.t 6SporadicSchiavonet al., 1998
Cys98TyrDSS/CH10 yearsu 3.8, p NR,s NRFabriziet al., 1999
*Asp99AsnDSSWarneret al., 1997
Lys101ArgDSS2.5 yearsWCB at age 10 yearsm11Gabreels‐Festenet al., 1996
Lys101ArgDSSWCB at age 7 yearsm10Gabreels‐Festenet al., 1996
Lys101ArgCMT1Sat at 18 months>20 monthsSevere<10Sister, mother,grandmotherTachiet al., 1996
Asn102LysRLSDelayed development5 yearsSevere at later age<16Five out of seven familymembersPlante‐Bordeneuveet al., 1999
Ile106Leu1B/DSS2.5 yearsm7Gabreels‐Festenet al., 1996
Ile106ThrDSS2 yearsAge 6 years walks shortdistance u 8. p NRSporadicTysonet al., 1997
Val117PheCMT1BUnsteady gait 18 monthsSevere at 15 yearsm 5Ohnishi, 2000
Leu137fs222XDSSDelayed motormilestonesWCB at age 25 yearsu 6SporadicTysonet al., 1997
Gly138AlaEO–CMT1Early onsetSevereSporadicSorouret al., 1998
Gly138ArgDSSDelayed motordevelopment30 monthsWCB at age 18 years,respiratory supportGerm line mosaicismTakashimaet al., 1999
3 yearsWCB at age 20 years,respiratory distressSister of aboveTakashimaet al., 1999
Leu145fsDSSDelayed at 6 months2–3 yearsHead control at 1 year,AbnormalSporadicWarneret al., 1996
Ala159fs206XDSSFloppy infant at6 months,delayed4 yearsSevere at 8 years(neonatal hypoxia)NRTachiet al., 1998
Glu186XCHDelayed at age 1 year,hypotonia3.5 yearsm3.6, p4.2Mandichet al., 1999
Ala192fsDSS15 m/sSporadicRautenstrausset al., 1994

AFO = ankle–foot orthosis; CH = congential hypomyelination; DSS = Dejerine–Sottas disease; EO–CMT = early onset CMT; m = median; p = peroneal; NR = non‐recordable; RLS = Rossy–Levy disease; s = sural; t = tibial; WCB = wheel chair bound; u = ulnar. *All three mutations in the same patient; **homozygous mutations.

From the available data, patients with 37 mutations could be classified as having an early onset neuropathy. Walking was either delayed in these individuals (30 cases) or the diagnosis of their neuropathy was made in infancy (seven cases). Five children in this group were described as floppy or hypotonic prior to 9 months of age. All these patients, however, with the exception of two children who died prior to age two years, were eventually able to ambulate, although they rarely walked normally and often could not run. Two patients in this group did not walk until 10 and 12 years of age, respectively, and at least six patients have required the use of a wheelchair. Wheelchair use was not inevitable, however, and several patients continued to walk with a cane into their 80s (Bird et al., 1997). Despite the severity of the neuropathy, life expectancy was not generally reduced.

Electrophysiological data from the early onset group were consistent with a severe demyelinating neuropathy with secondary axonal degeneration. As shown in Table 2A, MNCVs were quite slow in these patients. In all but two patients, NCVs were <15 m/s in the upper extremities and were often <10 m/s. Despite severe disability in lower limbs, several patients had obtainable peroneal, tibial, or sural conduction velocities, all of which were similar to the NCVs in their upper extremities. Compound muscle action potentials (CMAP) and sensory nerve action potentials (SNAP) were also reduced. Neither temporal dispersion nor conduction block was reported in any patient.

From the available data, patients with 21 distinct mutations could be classified as having an adult onset neuropathy. Most of these patients developed symptoms after age 30 years, but several were asymptomatic at the time of their diagnosis. Although in many instances the clinical signs and symptoms of neuropathy remained mild after onset, this was not inevitably so. At least seven patients with a Thr95Met mutation, for example, became wheelchair bound (De Jonghe et al., 1999), in one instance by age 49 years; this patient subsequently died at 52 years because of an aspiration pneumonia related to swallowing dysfunction.

Electrophysiological analysis of patients with adult onset neuropathy typically demonstrated at least mild evidence of demyelination, but effects on normal salutatory conduction were often minimal with more prominent axonal degeneration. All but one of these patients, for example, had MNCVs >20 m/s in the upper extremities and, in most cases, their conduction speeds were >30 m/s. In fact, several patients had nerve conductions >50 m/s and were diagnosed with CMT2. CMAPs and SNAPs, however, were typically reduced, even in patients with normal or near NCVs, demonstrating the presence of prominent axonal dysfunction.

Neither of the two phenotypes associated with MPZ mutations is typical of patients with CMT1A caused by a duplication of chromosome 17p. In CMT1A patients, for example, early milestones are usually normal and disability develops gradually during the first two decades of life (Thomas et al., 1997). Eighty‐three of the 90 CMT1A patients seen in our clinic began walking prior to or around 1 year of age, and clinical abnormalities developed within the first two decades of life in most of them (Krajewski et al., 2000). In addition, the NCV for patients with CMT1A is rarely <15 m/s or >30 m/s. In a combined series of 141 CMT1A patients seen at the National Hospital of Nervous Diseases in London and the CMT clinic at Wayne State University, the mean MNCV was 21 m/s, with an SD of 5 m/s. None of the patients in this group had a NCV >34 m/s, only five had NCV >30 m/s, and three had NCV <15 m/s (Laurá et al., 2003). A comparison of the distribution of NCVs from patients with CMT1B analysed in this work with the 141 patients with CMT1A analysed by Laurá and co‐workers is shown in Fig. 3. As can be seen in this figure, the distribution of NCVs from patients with MPZ mutations forms two separate groups, both of which are distinct from the distribution of NCV from patients with CMT1A.

Fig. 3 Correlation between early and late onset phenotypes and NCV for patients with MPZ mutations. Areas in grey represent the mean MNCV ± SD for the CMT1A patients as described by Laurá et al. (2003). Note that very few patients with MPZ mutations have NCV in the range of CMT1A patients.

The two MPZ phenotypes are associated with distinct neuropathological changes

Although we have not performed pathological studies on our own patients, we extended our literature review to determine whether pathological features might provide clues as to why certain mutations produce early or late onset neuropathies.

Abnormalities of myelin are the prominent feature of nerve biopsies from early onset patients. A loss of myelinated axons was frequent in sural nerve biopsies (Ikegami et al., 1996; Meijerink et al., 1996; Tachi et al., 1996, 1998; Bird et al., 1997; Komiyama et al., 1997; Plante‐Bordeneuve et al., 1999), and teased fibre analysis has demonstrated demyelination and remyelination in all the patients studied (Ikegami et al., 1996; Warner et al., 1996; Fabrizi et al., 1999; Plante‐Bordeneuve et al., 1999). In addition, there was onion bulb formation (Latour et al., 1995; Gabreels‐Festen et al., 1996; Ikegami et al., 1996; Meijerink et al., 1996; Tachi et al., 1996, 1998; Bird et al., 1997; Komiyama et al., 1997; Tyson et al., 1997; Warner et al., 1997; Silander et al., 1998; Mandich et al., 1999; Young et al., 2001) and abnormally thinly myelinated axons (Ikegami et al., 1996; Meijerink et al., 1996; Warner et al., 1996; Bird et al., 1997; Komiyama et al., 1997; Fabrizi et al., 1999; Mandich et al., 1999; Plante‐Bordeneuve et al., 1999) in most patients. Electron microscope studies, however, have identified two types of ultrastructural abnormalities in these individuals: one in which there are extensive areas of uncompacted myelin and a second in which there are numerous areas of focally folded myelin (tomacula) (Gabreels‐Festen et al., 1996). Interestingly, these two abnormalities appear to be mutually exclusive (Gabreels‐Festen et al., 1996).

Axonal degeneration was the prominent feature of nerve biopsies from patients with late onset disease, and demyelination was less evident, though often present. Most biopsies had marked loss of myelinated fibres of all calibres, with numerous clusters of regenerating axons (Chapon et al., 1999; De Jonghe et al., 1999; Senderek et al., 2000). In one biopsy, for example, 73% of the myelinated axons had been lost (De Jonghe et al., 1999); in another, there was a selective loss of predominantly large diameter fibres with extensive sprouting and atrophy of remaining fibres. However, teased fibre analysis from this same patient demonstrated axonal loss but with little evidence of demyelination (Misu et al., 2000) and similar findings were noted in several other biopsies. Electron microscope studies, when performed, revealed little segmental demyelination of internodes or paranodal retraction (Senderek et al., 2000). Although evidence of abnormal compaction and tomacula formation was occasionally found in late onset cases, it was much less frequent than in early onset cases.

MPZ mutations that disrupt protein structure or truncate the cytoplasmic domain are associated with early onset phenotype

In order to understand further how the MPZ mutations cause these two phenotypes, we mapped the known mutations onto a model of the secondary structure of the protein. These results are shown in Fig. 2. Mutations causing the early onset phenotype are in red, while those causing the late onset phenotype are in blue. A list of these same mutations and their effect on the amino acid sequence of the protein are shown in Table 2A.

Mutations that introduced a charged amino acid, removed or added a cysteine residue, or altered an evolutionarily conserved amino acid were more likely to cause early onset neuropathy. All five mutations in this series, for example, which introduced a charge change within the extracellular domain of MPZ resulted in early onset disease as did eight out of the nine mutations that disrupted or added a cysteine residue, as well as 18 of the 23 mutations which altered an amino acid conserved throughout vertebrate evolution. These data suggest that mutations that disrupt the tertiary structure of MPZ are more likely to cause severe, early onset disease (Table 3A).

View this table:
Table 2B

Features of late onset neuropathy (literature review)

Amino acidchangeAuthordiagnosisAgewalkedFirst limitationsClinical courseNCVFamily historyOther featuresAuthors
Asp6TyrCMT1BNormalSevere in twopatientsm24–4010 patientsMastaglia et al., 1999
Ser15PheCMT2Normal16 patientsMarrosu et al., 1998
Ser22PheCMT1BNormalAge 45 yearsMildu26, p31SiblingsYoung et al., 2001
Age 42 yearsu16.4
Asp32GlyCMT2Age 50–59 yearsModerate to severem36–43, t27–31DominantSenderek et al., 2000
Phe35delCMT1BNormalNonem31, p31FatherSevere crampsIkegami et al., 1996
CMT1BNormalNonem36, p29Mother
Glu42stopCMT1BNormalAge 32–36 yearsMild31–53, manyLagueny et al., 2001
Asp46ValCMT2NormalAge 60–61 yearsMildm44–523 familiesDeafnessMisu et al., 2000
t42
Ser49LeuCMT1BNormalAge 53 yearsMildm28, p232 separate pedigreesFabrizi et al., 2000
Age 33 yearsMildm29
Ser49LeuCMT1BNormalAge 42 yearsSevere at 72 yearsp21.6DominantYoung et al., 2001
*His52TyrCMT1BNormalAge 35 yearsModeratem35, u37UnclearAdies pupilBienfait et al., 2002
*Val84PheCMT1B
Arg69HisCMT1BNormalAge 23 yearsSevereYoung et al., 2001
Arg69HisCMT1BNormalAge 47 yearsSeverem24DominantLagueny et al., 2001
Arg69HisCMT1BAge 29 yearsm16, u19, t14Steroid responsiveWatanabe et al., 2002
Ile70ThrCMT1BNormalAge 58 yearsSeverem 33, u 349 patientsSteroid responsiveDonaghy et al., 2000
Val73fsCMT1BNormalAge 41 yearsNormalp29, u44, s30, m34FatherWarner et al., 1996 andPareyson et al., 1999
Tyr90CysCMT2MildDominantSenderek et al., 2000
Asn93SerCMT1BAge 44 yearsMild at 53 yearsm 32Blanquet‐Grossard et al., 1996
Thr95MetCMT2NormalAge 37–56 yearsm47–49, t32–444 familiesAdies pupilMisu et al., 2000
CMT2Age 18–50 yearsm34–58, u36–523 generations,7 patientsArg. Rob. Pupil, deafChapon et al., 1999
CMT2Age 40+ yearsm34–58, u36–529 familiesAdies pupil, deafDe Jonghe et al., 1999
Ser111ThrCMTAge 30 yearsMild to moderatem37, t26DominantStreet et al., 2002
Glu112XCMT2m 44.5DominantYoung et al., 2001
Gly134ArgCMTAge 27 yearsMildm 33, u35, p21Dominant, similarphenotypeConduction blockStreet et al., 2002
Arg198SerCMT1BNormalAge 38 yearsMildm 32, u30SporadicXu et al., 2001
Lys207DelCMTAge 45 yearsMild at 49 yearsm38, u40, p27, t 39Dominant, similarphenotypeStreet et al., 2002

m = median; p = peroneal; s = sural; t = tibial; u = ulnar; Arg.=Argyll; Rob.=Robertson. *Double mutation.

View this table:
Table 3A

Features of mutations affecting amino acids critical for homotypic interactions (Shapiro et al., 1996)

InterfaceEarly onset neuropathyLate onset neuropathy
Four foldR69C, R69S, R69H, N87HS15F, R69H
AdhesiveI1M, C21del, Y39C, S49L, H52RD46V, S49L, H52Y
Head to headC21delD32G
Remaining aminoacidsT5I, I32F, Del35F (homo), Y53C, Del56QPYI/InsHLF D61E, G64E, W72C,V73fs (homo), G74E, I85T, Del89DY/InsFY, G94C, Del96FT, C98Y, K101R,I106T/L, N102K, V117F, G138RD6Y, H10P, S22F, F35del,I70T, V73fs (hetero), Y90C,N93S, T95M, S111N, G134R,R198S, Del207K
View this table:
Table 3B

Features of mutations affecting phenotypes

Early onsetIntroduction of new ± chargesG64E, G74E, N87H, N102K, G138R
Taking away ± chargesR69S, R69C, D99N
ExtracellularDeletion truncating the EC domainC21del, Q56del, D89del, F96del
Disruption involving CysC21del, S34C, Y39C, Y52C, R69C,W72C, G94C, C98Y
Changes in polarity and hydrophobicityT5I, S34F, S49L, T85I, I106T
TransmembraneFrame shift that truncates TM andIC domainG138fs, L145fs
CytoplasmicFrame shift that truncates IC domainA159fs, A192fs, Q186stop
Late onsetTaking away ± chargeD6Y, H10P, D32G, D42V, H52Y,
Change at/near glycosylation siteN93S, Y90C, T95M
ExtracellularIntroduction of CysY90C
Changes in polarity and hydrophobicityD6Y, H10P, S15F, S22F, D32G, D46V,H52Y, I69T, T95M
TransmembraneG134R
CytoplasmicChange at RSTK site (PKC)R198S
DeletionK207del

EC=extracellular; TM=transmembrane; IC=intracellular; PKC=protein kinase C.

In addition, five of the six mutations that truncate the cytoplamic domain of MPZ, eliminating amino acids 198–201 (RSTK) have caused severe early onset disease, as has a point mutation altering Ser204, which has been hypothesized to be phosphorylated via PKC binding to the RSTK domain (Table 3A).

Discussion

Analysis of 13 patients with eight MPZ mutations from our clinic and 64 additional cases from the literature demonstrate that most individuals with CMT1B develop one of two phenotypes: an early (childhood) onset neuropathy with very slow NCVs and predominant dysmyelination on nerve biopsy, or a late (adult) onset neuropathy with minimal to moderately slowed NCVs and a predominant axonal neuropathy on nerve biopsy. Very few patients had a classic CMT1A phenotype with onset during the first two decades and moderately slowed NCVs. These data thus suggest that MPZ mutations also fall into two classes: those that predominantly affect myelination during development and cause early onset disease, and those that predominantly affect axons but not myelination and cause late onset disease. Correlation of the genotype with the phenotype for patients with CMT1B has thus provided important insights into the possible function of MPZ in the PNS. Potential molecular mechanisms for each class of mutations will be discussed further below.

One potential problem with classifying patients with MPZ mutations is that of phenotypic variability, in which some patients with the same mutation have different clinical phenotypes. Although there is some phenotypic variability among patients with MPZ mutations (Marques et al., 1999; Senderek et al., 2001), the majority of patients fit clearly into one of the two distinct clinical phenotypes described in this work. All four patients from the three families we have evaluated with a His10Pro mutation, for example, have an unambiguous late onset phenotype. In addition, several separate groups have identified >20 patients with a Thr95Met mutation, all of whom have a similar late onset neuropathy associated with pupillary abnormalities and hearing loss (Chapon et al., 1999; De Jonghe et al., 1999; Misu et al., 2000). In one of the largest single families with CMT1B, Bird et al. (1997) evaluated 13 affected individuals with the same MPZ mutation and found a similar early onset phenotype among all but one of them. Six of these 13 individuals had a delayed onset of walking, 10 could not keep up with their peers by age 6 years, and most were clumsy runners during childhood. Only one of the 13 developed symptoms after the age of 21 years. NCVs were <15 m/s in all nine patients who were evaluated (Bird et al., 1997). Only two of the 72 MPZ mutations analysed in this work, R69H and S49L, were difficult to classify, since they were associated with patients with both early and late onset neuropathy. Even in these two examples, the NCVs were similar and slower than those of most patients with late onset disease (Table 2A). It is thus our belief that the presence of phenotypic variability among patients with MPZ mutations has not significantly interfered with our ability to make genotype–phenotype correlations and thus does not alter the major conclusions of this study.

How might mutations causing early onset neuropathy interfere with myelination during development? One possibility is that MPZ mutations indirectly affect the myelinating Schwann cell due to abnormalities of protein misfolding and/or altered intracellular transport, as has been found for mutations in the major CNS myelin protein proteolipid protein (PLP1) (Gow et al., 1998, 2004) and the PNS myelin protein, PMP22 (Naef and Suter, 1999). In fact, two independent studies have demonstrated that the Ser34del mutation of MPZ, associated with early onset disease, prevents mutant MPZ from being transported out of the endoplasmic reticulum (Matsuyama et al., 2002; Shames et al., 2003). A second possibility, however, is that the MPZ mutations are incorporated into myelin and alter myelination directly. In a recent paper, for example, Shames and co‐workers have shown that several MPZ mutations which we have shown cause both early (Ser34Cys, Ser34Phe, Ser49Leu, Lys101Arg, Gly138Arg, Gln186stop) and late (Ser15Phe, Thr95Met) onset disease, are transported out of the endoplasmic reticulum to the plasma membrane (Shames et al. 2003). In addition, we have shown the same for the Arg198Ser mutation (Xu et al., 2001). Kirschner and colleagues have also demonstrated, using X‐ray diffraction (Kirschner et al., 1994) and high resolution electron microscopy (Kirschner et al., 1996), that several MPZ mutations disrupt both extracellular and cytoplasmic apposition of myelin wraps, presumably because the mutant protein has been transported to the cell surface and incorporated into myelin. Thus, most mutant MPZs probably disrupt myelination directly, although indirect mechanisms also occur with some mutations.

Crystal structure analysis of the extracellular domain of MPZ suggests that the protein forms homotetramers within the plane of the membrane that interact across the plane of the membrane to mediate homotypic adhesion. We hypothesize that mutations that interfere with these interactions would also be expected to disrupt myelination during development and cause early onset disease. Consistent with this interpretation, MPZ mutations that dramatically alter the secondary and tertiary structure of the protein by adding a charge or a cysteine residue are associated with early onset neuropathy. Interestingly, mutations at amino acid residues predicted from the crystal structure of the extracellular domain of MPZ to participate directly in either cis‐ and trans‐interactions did not consistently correlate with early onset neuropathy. This crystal structure analysis, however, did not include the cytoplasmic domain of MPZ that is known to participate in MPZ‐mediated homotypic adhesion (Wong and Filbin, 1996), and its absence might have altered the native MPZ structure. Analysis of the crystal structure of additional MPZ mutants, both with and without the cytoplasmic domain, will be necessary to resolve this issue.

How might MPZ mutations causing late onset neuropathy interfere with axonal function? One attractive possibility is that these mutations produce a subtly abnormal myelin sheath that causes alteration in Schwann cell axonal interactions, and thereby lead to axonal degeneration. Most patients with late onset neuropathy, for example, have relatively normal or only slightly slowed nerve conductions, and show predominant axonal loss with mild demyelination on nerve biopsy. Consistent with this hypothesis, axonal degeneration has been shown to occur in virtually all demyelinating neuropathies (for a review, see Shy et al., 2002), including patients with CMTX, who also often have subtle evidence of demyelination despite marked axonal loss. Furthermore, transplantation studies have shown that demyelinating Schwann cells directly alter axonal transport and neurofilament phosphorylation (Aguayo et al., 1977; de Waegh and Brady, 1990; Sahenk et al., 1999). Interestingly, the best characterized late onset mutation, Thr95Met, is a conservative amino acid substitution of one hydrophobic amino acid for another, located within one of the hydrophobic β‐sheet cores of the immunoglobulin‐like domain of MPZ (Shapiro et al., 1996), and would thus not be predicted to alter the MPZ structure significantly. The molecular and cellular mechanisms by which these MPZ mutations disrupt Schwann cell axonal interactions, however, are not known.

Whatever the mechanisms by which MPZ mutations cause neuropathy, our data demonstrate that careful analysis of the phenotypes of these patients is important to understand the molecular and cellular basis of CMT1B. These clinical and electrophysiological studies strongly suggest that some MPZ mutations cause a late onset, predominantly axonal neuropathy by altering Schwann cell axonal interactions, while others cause an early onset, predominantly dysmyelinating neuropathy by interfering with the myelination during development. Further insight into the molecular basis of these clinical phenotypes will thus be important for understanding the cell biology of myelination, the function of MPZ and the pathogenesis of CMT1B.

Acknowledgements

This work was supported by grants from the NIH (R01 NS41319A) and from the Charcot‐Marie‐Tooth Association.

References

View Abstract