Background
Slowly progressive distal weakness, muscle atrophy, and sensory loss due to an inherited peripheral neuropathy was described independently in 1886 by Charcot and Marie in France and by Tooth in England. [1, 2] A few years later, Dejerine and Sottas recognized and described a more severe, infantile form of inherited neuropathy. [3]
As more heterogeneity and overlap in clinical appearance, pathological features, and forms of inheritance were recognized in the following decades, an improved classification system was needed to avoid confusion. Starting in the 1950s, the clinical use of nerve conduction studies combined with pathological information allowed patients to be divided into 2 major groups.
-
Group 1 was characterized by slow nerve conduction velocities and evidence of hypertrophic demyelinating neuropathy.
-
Group 2 was characterized by relatively normal nerve conduction velocities and axonal degeneration.
Most of the families demonstrated autosomal dominant inheritance, and affected relatives in each family could all be categorized in the same group. [4, 5, 6]
In 1975, Dyck expanded the classification system of what was now known as hereditary motor and sensory neuropathy (HMSN) to include forms with additional features. [7]
-
HMSN types 1A and 1B (dominantly inherited hypertrophic demyelinating neuropathies)
-
HMSN type 2 (dominantly inherited neuronal neuropathies)
-
HMSN type 3 (hypertrophic neuropathy of infancy [Dejerine-Sottas])
-
HMSN type 4 (hypertrophic neuropathy [Refsum] associated with phytanic acid excess)
-
HMSN type 5 (associated with spastic paraplegia)
-
HMSN type 6 (with optic atrophy)
-
HMSN type 7 (with retinitis pigmentosa)
In the 1980s it became clear that this revised classification system, based on clinical and electrophysiologic characteristics, was inadequate to describe the genetic heterogeneity within each of these categories. Linkage studies revealed Charcot-Marie-Tooth type 1 loci on both chromosome 1 [8] and chromosome 17 [9] , and X-linked and recessively inherited forms were increasingly recognized.
In 1991, 2 groups showed that the most common form of CMT1, known as CMT1A, was associated with a duplication within chromosome 17p11.2. [10, 11] This duplication is believed to be in the peripheral myelin protein 22kD (PMP22) gene, and overexpression of the gene product appears to be causative, since a gene dosage effect has been demonstrated. [12] It is estimated that this duplication is responsible for about 70% of CMT1 cases [13] and the vast majority of CMT1A cases, with rare exceptions such as partial 17p trisomy. [12, 14]
Other Charcot-Marie-Tooth genes were discovered in the 1990s. The second most common form of CMT1 (CMT1B) and some cases of Déjerine-Sottas syndrome were found to be associated with mutations in the myelin protein zero (MPZ) gene on chromosome 1. [15, 16, 17] The most common form of CMTX (CMTX1), was found to be due to mutations in the gap junction protein beta 1/connexin 32 (Cx32) on chromosome Xq13.1. [18] Interestingly, hereditary neuropathy with liability to pressure palsies (HNPP) was found to be associated with a deletion in the PMP22 gene [19] , but this syndrome will not be reviewed here.
The difficulty classifying Charcot-Marie-Tooth subtypes is made more evident by the fact that mutations of each of these genes have been associated with multiple, overlapping phenotypes. For instance, myelin protein zero mutations are associated CMT1B, Déjerine-Sottas syndrome, and the axonal CMT2 phenotype. [20, 21, 22] Déjerine-Sottas syndrome has been associated with mutations or deletions in PMP22, myelin protein zero, and early growth response 2 (EGR2) genes. [15, 23, 24]
In addition, the boundaries between even the major types of CMT are not always as clear as the original series suggested. In the mid 1970s, Bradley, Davis, and Madrid performed a similar study to those performed by Dyck and Lambert, Thomas and Calne, and Buchthal and Behse, and proposed a CMT classification that included an intermediate group characterized by median motor nerve conduction velocities of 25-45 m/sec and intermediate pathological changes. [25, 26, 27] A study performed by Brust, Lovelace, and Devi suggested a bimodality in nerve conduction velocities among type 1 patients, again raising the possibility of an intermediate form [28] , but these were considered exceptions and were never incorporated into the major classification systems. In 1985, a large kinship with an intermediate form was described.
In 1998, sequencing of the PMP22 and MPZ genes in this kinship did not reveal mutations, underscoring that it is a distinct entity. [29, 30] Several associated mutations have been discovered since, and as with other forms of CMT, these studies have demonstrated genetic heterogeneity.
Given what is now known about clinical and genetic heterogeneity, a system that takes both clinical and genetic characteristics into account is not possible. This review will use the following system:
-
CMT1 is a dominantly inherited, hypertrophic, predominantly demyelinating form.
-
CMT2 is a dominantly inherited predominantly axonal form.
-
Dejerine-Sottas is a severe form with onset in infancy.
-
CMTX is inherited in an X-linked manner.
-
CMT4 includes the various demyelinating autosomal recessive forms of Charcot-Marie-Tooth disease.
Pathophysiology
CMT1A
The extra PMP22 gene copy within the 1.5 mB duplication on chromosome 17 is believed to cause most cases. [12] PMP22 is a 160 amino acid integral membrane protein that is expressed at high levels in myelinating Schwann cells, localizing to compact myelin and making up 2–5% of total myelin protein. [31, 32] PMP22 expression in CMT1A nerve biopsies is increased [33] , but the process by which protein overexpression actually causes the Charcot-Marie-Tooth phenotype remains unclear.
Abnormal expression of PMP22 seems to alter Schwann cell growth and differentiation both in vivo and in vitro, and may impair the ability of the Schwann cell to maintain normal myelin stability and turnover. A putative interaction between PMP22 and MPZ could play a role in maintaining myelin compaction and stability, and an imbalance in one could explain why changes in expression in either gene can lead to the clinically and pathologically indistinguishable CMT1A and CMT1B phenotypes. [34]
Despite the evidence of demyelination found on pathological and electrophysiological studies and the more recent implication of myelin proteins, the signs and symptoms of weakness and sensory loss are likely the result of axonal degeneration rather than demyelination. Anatomical evidence of progressive length-dependent axonal loss following demyelination in CMT1 exists. [35] Children with CMT1A have slow nerve conduction velocities in the first years of life, generally preceding the development of signs and symptoms. [36] Krajewski and colleagues demonstrated that compound motor action potential amplitudes correlate best with weakness in CMT1A rather than nerve conduction velocities. [37]
In vivo studies of CMT1A patients have demonstrated altered axonal excitability, in the form of elevated stimulation thresholds and a markedly abnormal threshold electrotonus. These findings suggested that demyelination causes exposure of fast K+ channels, which are normally sequestered under the myelin. [38]
CMT1B
Myelin protein zero is an integral type I membrane protein, and is the most abundant structural protein of compact peripheral nerve myelin. [39, 15]
More than 80 mutations in myelin protein zero have been described. Most are associated with the typical CMT1B phenotype, but correlations have been found with other clinical phenotypes including CMT2, Dejerine Sottas syndrome, and congenital hypomyelination neuropathy. [40, 41]
Predominantly axonal and predominantly demyelinating forms appear to be due to different mutations in the same gene. Mutations that introduce a charged amino acid, removed or added a cysteine residue, or altered an evolutionarily conserved amino acid alter the tertiary structure of the MPZ protein as a result and lead to the more severe early onset phenotype. As with diseases of PMP22 expression, the mechanism for either demyelinating or axonal forms remains unclear, but may be due to impaired myelin-axon interaction. [42, 43, 44]
CMTX
CMTX is caused by mutations in connexin32/Gap Junction Beta 1 (Cx32 or GJB1), which maps to chromosome Xq13. Connexins assemble to form intercellular gap junctions, which allow the diffusion of ions and small molecules across apposed cell membranes. Connexin 32 is expressed in many tissues, but is localized in the myelin sheath of large diameter fibers near the nodes of Ranvier. [18, 45]
Axonal and demyelinating changes are mixed in CMTX. Men are clinically and electrophysiologically more severely affected than affected women. Interestingly, nerve conduction velocities in affected women can vary markedly within the same limb, in contrast to men whose conduction velocities tend to resemble the diffusely slow velocities seen in CMT1. This may be partly explained by X-inactivation of Schwann cell precursors during development. The mechanisms by which different Cx32 mutations cause CMT1X are not fully understood. Phenotype-genotype correlations in CMT1X will be difficult to establish because of phenotypic variability both within and between kindreds. [46, 47, 48]
CMT2
The most common mutation responsible for CMT2 has been found in the mitochondrial GTPase mitofusin 2 (MFN2) gene. [49] This is referred to as CMT2A.
MFN2 is a dynamin family GTPase that spans the outer mitochondrial membrane and is believed to primarily be involved in mitochondrial docking, tethering, and fusion. Recent data were supportive of a mitochondrial trafficking insult as a possible mechanism of length-dependent axonal neuropathy. [50, 51]
CMT4
The causes of CMT4 are diverse and very rare and will not be reviewed here.
Epidemiology
Frequency
International
CMT is the most common inherited neuromuscular disorder. Estimates of the frequency of Charcot-Marie-Tooth disease vary widely. In 1974, Skre and colleagues reported a prevalence of 1 case per 2,500 individuals. A worldwide meta-analysis estimated a prevalence of 1 per 10,000 individuals, and a prevalence of 10.8 per 100,000 was found in western Japan. [52, 53]
Charcot-Marie-Tooth disease type 1 accounts for about two thirds of CMT cases. Of these, 70% are due to duplications in PMP22. Duplications in PMP22 therefore account for roughly half of all combined CMT subtypes. [13, 54] CMT1B makes up about 5% of CMT1 and about 1.6% of all CMT cases. [54, 55]
Charcot-Marie-Tooth disease type 2 accounts for about 22% of CMT cases. [54] CMT2A (MFN2 mutation) has been estimated to account for 11-23% of all CMT2 cases. [56]
Charcot-Marie-Tooth disease type X accounts for about 16% of CMT cases. [54] CMTX1 (Cx32 mutation) accounts for most CMTX cases and may be the second most common cause of CMT overall, at 7-20%. [42, 43]
A 2016 systematic review of previously published epidemiologic Charcot-Marie-Tooth disease studies showed significant variation in the worldwide prevalence of CMT dependent on country, region, and/or ethnicity. [57]
Mortality/Morbidity
Most people with CMT have a normal life expectancy, the exceptions being patients with respiratory involvement or severe disability. Disability varies greatly both between and within families, and can range from asymptomatic with minimal examination findings to severe. Marked differences in clinical severity have even been reported in monozygotic twins with CMT1A (despite similar nerve conduction velocities) [58] and myelin protein zero mutations [59] .
A 2001 study showed that 44% of CMT1A patients were significantly disabled, and 18% were depressed. It was estimated that CMT1 patients suffer emotional stress similar to patients with stroke and comparable disability. [60]
CMT is nearly always slowly progressive. A longitudinal study showed steady progression in nerve conduction velocity and disability, and severity could be predicted based on nerve conduction velocity abnormalities in childhood. [61] Killian and colleagues studied nerve conduction velocities and neurologic examinations in 8 members of a single family over 22 years and found minimal progression of disability. [62] Another study of 43 patients with CMT2 over 5 years documented slow progression of weakness and disability, and most patients remained ambulatory. [63]
As CMT is relatively common, it can occur with other inherited or acquired neuromuscular conditions. If progression accelerates, these possibilities should be pursued. [64]
Race
No racial predilection is recognized in the common forms of CMT. Rare, recessively inherited forms (CMT4) cluster in particular ethnic groups.
Sex
As expected, CMTX affects males earlier, more frequently, and more severely than females. Other forms of CMT do not show predilection for either sex.
Age
Onset is usually in childhood. Thomas and colleagues found that 75% of patients with CMT1A developed clinical evidence of disease before age 10, and 85% before age 20. [65] Some patients experience very slow progression of mild, lifelong symptoms, and may seek medical attention only late in life.
There are patients and families in which CMT has a late onset, as late as the fifth or sixth decade in some MPZ mutations, and even as late as the seventh decade in CMT2. [41, 66]
Prognosis
Life expectancy is not altered in all but the most severe cases. Disability is highly variable within and between kindreds and cannot be predicted with any certainty, even among siblings. More marked disability does seem to be linked to earlier onset.
Patient Education
The CMT Association is a nonprofit organization concerned with patient support, public education, and promotion and support of research into the cause and treatment of CMT.
