Two divergent types of nerve pathology in patients with different P sub 0 mutations in Charcot-Marie-Tooth disease

Charcot-Marie-Tooth disease type 1 (CMT1), or hereditary motor and sensory neuropathy type I (HMSN type I), is a demyelinating polyneuropathy, with prevailing autosomal dominant inheritance. Most autosomal dominant cases map to chromosome 17 (CMT1A); a minority map to chromosome 1 (CMT1B) or to another unknown autosome (CMT1C). CMT1A usually results from a duplication of a 1.5-Mb region on chromosome 17p12 encompassing the gene for the peripheral myelin protein PMP22; a few patients show point mutations in this gene (see review in Patel and Lupski ^[1]). Recent analysis of CMT1B families did reveal different point mutations in the gene for the major protein of peripheral myelin, protein zero (P₀). ^[2-9] In some sporadic patients with a more severe phenotype, formerly known as Dejerine-Sottas disease (DS), de novo P₀ mutations have been identified. ^[10,11] Nerve morphology in most CMT1B families is designated as ``hypertrophic demyelinating neuropathy.'' One report described a demyelinating polyneuropathy with an abundant occurrence of focally folded myelin in a father and son with a P₀ mutation. ^[12] Hypomyelination with extensive onion bulb formation dominated the pathology in two severely affected de novo mutation cases. ^[10,13,14]

We describe, for the first time, two divergent morphologic phenotypes in seven patients with different P₀ mutations. ^[15]

Patients and methods.

Mutation screening of P₀ exons was performed in seven unrelated patients without a duplication on chromosome 17p12, by single-stranded conformation polymorphism (SSCP) analysis, followed by sequencing of the relevant DNA region, according to previously described protocols. ^{[2,3,8,16-18]} The reported P₀ mutations were not found in 100 unrelated healthy controls. All patients suffered from a moderately to prominently disabling demyelinating motor and sensory neuropathy with marked conduction slowing Table 1 and had formerly undergone a sural nerve biopsy. None of the patients had experienced pressure palsies. Patients 1, 5, and 6 used wheelchairs from the ages of 35 years, 10 years, and 7 years, respectively. Patient 3 showed a marked deterioration from the age of 10 months onward and died unexpectedly at the age of 22 months.

Data of patients 1, 4, 5, and 7 are included in previous reports. ^[19-21] Patient 2 is a member of family NL-47. ^[3] Sural nerve biopsy specimens were prepared for light and electron microscopic studies and morphometric examinations using previously described methods. ^[20]

Results.

Patient 2, member of a CMT1B family, showed a three-base-pair deletion in the P₀ gene, resulting in a deletion of the codon for serine 34. ^[3] In the six other patients we identified a single base change, leading to an amino acid substitution in codons 5, 69, 101, and 106 of the P₀ gene Table 1. Amino acid numbering was started at the first amino acid (Ile) of the mature P₀ protein, after the 29 amino acids of the signal peptide. ^[22] Different codon 69 substitutions were identified in the patients 3 and 4. The Arg69His substitution of patient 4 was found in two other CMT1B families. ^[5,9] The unrelated patients 5 and 6 showed an identical mutation in codon 101. These six patients were isolated cases; parents of five patients were normal by clinical and electrophysiologic examination and analysis of parental DNA demonstrated that the amino acid substitutions were de novo mutations. For patient 1 neither the parents nor the nine siblings showed clinical signs of CMT. Chromosomal DNA was available from five siblings only and these did not carry the mutation. Therefore, it is likely but not proven that patient 1 also had a de novo mutation. The patients were heterozygous carriers of the mutation, concordant with autosomal dominant inheritance. A child of patient 4, born in 1992, showed at the age of 2 years clumsy motor performances, absent ankle jerks, and a median motor nerve conduction velocity of 21 m/s. DNA investigation revealed the same Arg69His mutation as in the mother.

Light microscopic examination showed in all cases a chronic demyelinating process with onion bulb formation. Myelinated fiber density was markedly decreased and myelinated fiber diameter histograms showed a preferential loss of large-diameter fibers. Total transverse fascicular area was moderately increased.

Ultrastructural examination in patients 1, 2, 3, and 4 revealed the presence of uncompacted myelin in 23 to 68% of the myelinated fibers. The uncompacted structure was commonly seen in the inner layers of the myelin sheath, but occasionally it was seen in the outer lamellae as well Figure 1A. At the transitional zone between compacted and uncompacted myelin, the major dense lines split to enclose thin layers of Schwann cell cytoplasm, sometimes with adherens junctions ^[23] Figure 1B. Usually, the intraperiod distance was slightly broadened and sometimes partial fusion of the extracellular membranes occurred over a short distance Figure 1C. Schmidt-Lanterman incisures were frequent and abnormally wide. Occasionally, they could not be distinguished from areas of uncompacted myelin. In patients 1 and 4 several fibers showed undulating major dense lines with dilatation of intraperiod lines only Figure 1D. The compacted parts of the sheaths showed a normal periodicity of myelin. Myelin-like figures were often seen between the uncompacted lamellae, suggestive of incipient myelin degradation (see Figure 1B). Several demyelinated fibers were present; tomacula occurred sporadically in the patients 1 to 4. Onion bulb formations occurred rather frequently in the older patients and were composed of thin Schwann cell lamellae. The myelin sheath of nearly all fibers was too thin for the axon diameter, witness the high g-ratios (axon diameter: fiber diameter) (see Table 1).

• Download figure
• Open in new tab
• Download powerpoint

Figure 1. Electron micrographs of myelinated fibers in patients with uncompacted myelin (patients 1 to 4). (a) Uncompacted lamellae of the inner myelin layers. Bar = 500 nm. (b) Uncompacted myelin at inner and outer parts of the sheath. Splitting of major dense lines with adherens junctions (closed arrow). Myelin figure in uncompacted part of myelin (open arrow). Bar = 500 nm. (c) Dilatation of major dense lines (closed arrow) and fusion of intraperiod lines (open arrow). Bar = 200 nm. (d) Undulated major dense lines and uncompacted myelin. Bar = 500 nm.

In patients 5, 6, and 7 the occurrence of many folded myelin loops (tomacula) dominated the pathology Figure 2. In longitudinal sections excessive folded myelin appeared to pass into thin myelin sheaths within the same internode. Fibers outside tomacula were frequently demyelinated or thinly (re)myelinated. Onion bulbs were composed of thin Schwann cell layers and contained many double basement membranes, especially in case 7. Uncompacted myelin was encountered only in a few fibers. Involvement of unmyelinated axons was not obvious in either morphologic type.

• Download figure
• Open in new tab
• Download powerpoint

Figure 2. Electron micrograph of sural nerve in patient with focal myelin foldings (patient 7). Two fibers with excessive folded myelin. Thinly myelinated fiber. Several demyelinated fibers surrounded by basal lamina and few thin Schwann cell lamellae. Bar = 2 micro meter.

Discussion.

We investigated nerve pathology in seven patients with different mutations in the P₀ gene. The clinical phenotypes ranged from a severe, infantile neuropathy with very low conduction velocities to a milder neuropathy with later onset. The severe, infantile cases might be classified as DS disease, although they do not show an autosomal recessive inheritance, which is part of the original concept. ^[24] In the future, it would be preferable to classify the inherited neuropathies not by clinical and electrophysiologic severity but by the underlying genetic defects.

P₀, the major peripheral myelin protein, is an integral membrane glycoprotein. Gene expression is almost restricted to the myelin-forming Schwann cells. Most human P₀ mutations are located in exons 2 and 3 of the gene, corresponding to the immunoglobulin-like extracellular domain of the protein. ^[25] Few mutations are known to occur in the cytoplasmic domain ^[8,11]; two nonsense mutations were located in exon 4, presumably at the margins of the transmembrane domain, ^[8] and only one patient showed a mutation in the transmembrane domain of P₀. ^[10] The extracellular domain behaves like a homophilic adhesion molecule and has a role in compaction of myelin at the intraperiod lines. ^[26-28] Studies in P sub 0-gene-deficient mice revealed a pivotal role for this molecule in the formation and maintenance of myelin. ^[29] P₀ probably contributes to the (early) myelin compaction at the major dense lines. ^[30,31] One study suggested that P₀ is also involved in heterophilic interactions as the protein served as promotor for neurite outgrowth. ^[28]

Our patients showed two different types of pathology. The morphologic phenotypes are probably directly related to the mutations, because (1) patients with the same mutation display the same morphology, (2) neither morphologic expression is age-dependent, and (3) different amino acid substitutions in the same codon give rise to the same type of morphology, despite differences in severity (i.e., Arg69Cys versus Arg69His). It is plausible that the Arg69 substitution by Cys (patient 3) has a more deleterious effect on P₀ function than the substitution by His (patient 4), since cysteine might interact with the disulfide bond, which is essential for its function as an adhesion molecule. ^[32] The transmembrane mutation probably has the largest impact on P₀ function, since the patient with the Gly138Arg mutation had extremely hypomyelinated fibers and the lowest nerve conduction velocities of all reported patients with a P₀ mutation. ^[13] We suspect that differences in clinical severity (i.e., CMT versus DS) are largely determined by site and nature of the amino acid substitution. The morphology in patients 5, 6, and 7 is fully concordant with the findings in the family examined by Thomas et al. ^[12] However, the morphology in patients 1 to 4 is different in that it shows uncompacted myelin. There are previous descriptions of uncompacted myelin in human pathology, especially in a subset of plasma cell dyscrasias. ^[33-37] Three reports described this phenomenon in cases of (most probably) hereditary neuropathies; one report dealt with a severe infantile polyneuropathy, ^[38] two with cases of presumed hereditary neuropathy with liability to pressure palsies (HNPP). ^[39,40] In our experience, covering 21 cases with a proven 17p12 deletion, uncompacted myelin occurs rarely in HNPP. ^[41]

It is not known how the different mutations might interfere with P₀ functions. The disturbance of myelin compaction is in agreement with the widely accepted function of P₀ as a homophilic adhesion molecule. ^[26-28] The appearance of many tomacula is unexplained, but it might refer to an additional function of P₀. In this respect it is remarkable that Filbin and Tennekoon ^[42] tentatively suggested a function of P₀ in signal transduction, by which the nucleus of the Schwann cell is informed that sufficient myelin has been laid down. Alternatively, uncompacted myelin and tomacula could represent stages in different mechanisms of demyelination.

Recent experiments in transgenic mice homozygous for the inactivated P₀ gene showed thin myelin sheaths, uncompacted myelin, and degeneration of myelin and axons, partly comparable to the pathology in the patients 1 to 4. ^[29,43] However, the heterozygous null mutants showed normal myelination but developed progressive demyelination with onion bulbs after 4 months of age. ^[43] Although our patients with P₀ mutations are also heterozygotes and presumably express half the dose of the normal protein as in the heterozygous null mutant, ^[29] they probably also express the mutated protein. The formation of an aberrant protein may cause dominant-negative effects by forming homodimers of mutant type P₀ or heterodimers of wild and mutant type P₀, which may affect the normal P₀ functions. Alternatively, the aberrant P₀ protein might be poorly transported to the Schwann cell membrane or poorly inserted and is therefore metabolically toxic.

From the morphologic and molecular data we conclude that mutations in the extracellular domain of P₀ have at least two different effects on myelin sheaths: (1) mutations resulting in disturbed compaction of myelin, probably by interfering with the homophilic adhesion function of P₀, and (2) mutations resulting in focally folding of myelin. The two divergent pathologic phenotypes exemplify that some mutations act differently on P₀ protein formation or function than others, which is probably determined by site and nature of the mutation in the P₀ gene.

Acknowledgments

We thank L. Boender-van Rossum, L. Eshuis, S. van der Velde-Visser, and H. Veldman for technical assistance. We thank Dr. C.J. Howeler for providing DNA of patient 1.