Clinical manifestations

Presentation and course

In males, symptoms typically begin between 5 and 25 years of age; in many cases, patients are symptomatic by early adolescence. A retrospective study in a large cohort of patients with CMTX1 reported a mean age at onset of lower extremity symptoms of 16.2 +/- 11.4 years in males (79). The initial symptoms include difficulty running and frequently sprained ankles; foot drop and sensory loss in the legs develop later. Depending on the tempo of the disease, the distal weakness may progress to involve the gastrocnemius and soleus muscles, even to the point where assistive devices are required for ambulation.

Weakness, atrophy, and sensory loss also develop in the hands, particularly in the thenar muscles (02.) These clinical manifestations result from chronic, length-dependent axonal loss and are nearly indistinguishable from those seen in patients with CMT1A or CMT1B, although atrophy, particularly of intrinsic hand muscles, positive sensory phenomena, and sensory loss may be more prominent in patients with CMTX1 (85). Neurologic examination also reveals diminished to absent reflexes and sensory impairment, all of which are length-dependent and worsen insidiously over time but to varying degrees in different patients. Pes cavus, varus deformities, and "hammer toes" are frequently present.

Female carriers may be asymptomatic; if affected, they usually have a later onset and are less affected than males of the same age (70; 24). In the large retrospective study already referenced, female carriers presented with lower extremity symptoms at a mean age of 24.7 +/- 16.2 years (79). The reason that female carriers are less affected probably owes to X-inactivation; only a fraction of their myelinating Schwann cells express the mutant GJB1 allele (62).

GJB1 mutations cause numerous CNS manifestations. These are reviewed in detail in two reviews by Abrams (01; 02):

(1) Most mutations studied cause asymptomatic delayed brainstem auditory evoked responses, even without clinical involvement (64). Central visual, motor (09), and sensory (44) pathways may also be affected.

(2) Some mutations are associated with mild, fixed CNS findings, such as extensor plantar responses or abnormal MRIs, when asymptomatic (p.Trp3Gly, p.Trp24Cys, p.Met34Val, p.Ala39Val, p.Asn54His, p.Met93Val, p.Ser128Leu, p.Arg143Pro, p.Trp157Cys, p. Val170Phe, p.Asn175Ser). p.Asn54His was associated with persistent CNS dysfunction in one child who also had an abnormal brain MRI (87). p.Cys64Tyr was found in several members of a family who also had white matter changes on MRI, one of whom had a multiple sclerosis-like illness.

(3) Some mutations are associated with dramatic, transient clinical neurologic dysfunction and brain MRI abnormalities, often initially diagnosed as acute disseminated encephalomyelitis or a stroke. The majority of these are collated in a review by Tian and colleagues (p.Met1Ile, p.Arg22Gln, p.Val27Ala, p.Ile33Asn, p.Asn54Ser, .Thr55Ile, p.Cy60Tyr, p.Asp66Asn, p. Arg75Trp, p.Leu76Pro, p.Pro87Leu, p.Val91Met p.Met93Arg p.Val95Met p.Gln99_His 100insGln, 100insGln, p.Glu102del, p.Arg107Trp, p.Ile127Met, p.Trp132stop, p.Trp133fs, p.Val139Met p.Arg142Trp, p.Arg142Gln, p.Phe149Leu p.Leu156Arg, p.Arg164Trp, p.Arg164Gln, p.Cys168Tyr, p.Val177Ala, p.Val181Ala, p.Glu186stop, p.Thr188Ile, and p.Glu208Lys) (92). Additionally, at least three other CNS mutations causing a similar episodic phenotype are mentioned in another report (55) but not specifically distinguished from mutations causing static CNS changes. The Ala39Val and Phe51Leu mutations (SS Scherer p. c.) have also been associated with this syndrome. Findings typically include “upper motor neuron” weakness and dysarthria. Ataxia, respiratory distress, dysphagia, and altered consciousness have also been described. Symptoms last between a few hours and a few weeks and are often recurrent. Some episodes appear to occur without provocation, but most are associated with stressors such as hyperventilation or exertion, re-acclimatization after return from high altitude, fever, or minor infections. MRI changes, seen on both diffusion-weighted and T2-weighted sequences, preferentially involve subcortical white matter and the splenium of the corpus callosum (01; 02; 92). Some cases have been reported in females. Furthermore, acute attacks can occur in very young patients without a prior diagnosis of CMTX1 and may be the precipitant of the diagnosis (07; 65).

(4) One mutation, p.Pro58Ser, has been associated with spinocerebellar degeneration (88; 17).

(5) At least four different GJB1 mutations (p.Val38Ala, p.Thr55Arg, p.Arg142Gln, and p.Thr191frameshift) have been associated with hearing loss; some of these also were found to have prolonged brainstem auditory evoked responses (89; 51; 42), and there may be others. Males, and even females, can be affected in childhood.

The peripheral nerve-related clinical manifestations, including their age-related progression, are described in the clinical vignette.

Prognosis and complications

CMTX1 typically does not affect longevity. Regardless of the mutation, most affected men have similar degrees of impairment (86), but mutations in the cytoplasmic loop of the Cx32 protein appear to cause milder disease. Among the five most common variants in a large series, a mutation in the 5’ UTR (c.-17G> A) was shown to cause a more severe phenotype than the other four (79). Because of the random nature of X-inactivation, one cannot predict the degree to which a presymptomatic woman will be affected (70; 24).

There are a few potential complications. Ionasescu and colleagues reported "breathing difficulty due to phrenic nerve involvement" in severe cases of CMTX1, but no details were provided (39; 40). Scoliosis has been reported in patients with Charcot-Marie-Tooth disease, including CMTX1 (37). Lower urinary tract dysfunction may be more common in men with CMTX1 than in the general population (49). Although skin breakdown and trophic ulcers do not appear to be as problematic as in other neuropathies, it is prudent to advise patients about foot care. Autonomic involvement is not a prominent feature of CMTX1, and pain is less common in patients with CMTX1 than in patients with CMT1A (73). Patients with CMTX1 have not been reported to develop a superimposed inflammatory demyelinating neuropathy.

Patients with any inherited neuropathy may be at increased risk of developing neuropathy if they are exposed to agents that can cause neuropathy. A systematic review concluded that most drugs are likely as toxic (or nontoxic) in patients with Charcot-Marie-Tooth disease as in those without (19). However, two drugs, vincristine and possibly paclitaxel, can cause greater neurotoxicity in patients with Charcot-Marie-Tooth disease. Nonetheless, when possible, other drugs that can cause neuropathy should be avoided; these include cisplatin, suramin, colchicine, metronidazole, amiodarone, disulfiram, nitrofurantoin, isoniazid, dapsone, perhexiline, thalidomide, and "mega-doses" of pyridoxine (vitamin B6). However, cisplatin (20) and even vincristine (06) have been given to patients with CMTX1 with no obvious complications, but there is a brief report that the combination of vincristine and voriconazole worsened the neuropathy in a 5-year-old girl with CMTX1 (76).

Clinical vignette

The following four paragraphs are excerpted from a description of a CMTX1 by Hahn and colleagues (35); this family has a Tyr211stop mutation in GJB1.

Patient A. On examination, this 61-year-old male reported that the onset of his symptoms dated back to early childhood, when he noted awkwardness and difficulty in running. He always had highly arched feet. In his teens, the muscles in his lower legs became thin, and he frequently twisted his ankles. His penmanship was always poor, and his fingers and hands were clumsy, particularly in cold weather. Over the years, his symptoms progressed insidiously, and he developed pain and sensations of pins and needles in his feet and intermittently in his fingers. The hand muscles became progressively wasted, and his fingers became so clawed that he had poor use of his hands and could not even sign his name. A triple arthrodesis, performed in both feet at age 38 years, improved his walking, yet in the past 10 years, walking had become more difficult and he had to use two sticks. On examination there was marked wasting and weakness of the hand muscles, in particular the thenar muscles, and of all muscles below the knee. There was pes cavus and varus deformity. Tendon reflexes were depressed in the arms and at the knees, and ankle jerks were absent. There was a graded sensory loss to just above the wrist and to the knee. His gait was unsteady with bilateral foot drop.

Patient B. On examination, this 14-year-old male reported that he had difficulty running even when in kindergarten, when his feet were already highly arched. Between the ages of 8 and 10 years, his gait became increasingly awkward. He tripped easily and tended to walk on the lateral edge of the foot. He had difficulty using his hands in cold weather and needed help with shoelaces and buttons. On examination, there was mild wasting and weakness of the hand muscles, in particular the thenar muscles. His feet were highly arched with marked hammer toe formation and extensive callosities on the lateral foot borders. Peroneal and anterior tibial muscles, and to a lesser extent the gastrocnemius muscles, were wasted, and he was unable to dorsiflex his ankle. His tendon reflexes were depressed, and the ankle jerks were absent. All modalities of sensation were reduced to wrist and midcalf levels. Cutaneous nerves were not enlarged.

Patient C. On examination, this 37-year-old mother of Patient B reported that she was clumsy when running as a child and was never able to ice skate. In her late teens, her fingers felt awkward in cold weather and she had trouble buttoning her clothes. Her feet were always highly arched, but recently she had developed increasing hammer toes, and she had a tendency to sprain her ankles. Her feet became numb. The examination showed moderate wasting of the intrinsic hand muscles and the thenar eminences, and moderate wasting and weakness of the peroneal muscles and small foot muscles, with pes cavus and hammer toe formation. Ankle jerks were absent, and sensation was reduced to the wrists and ankles. She walked with a mild steppage gait.

Patient D. On examination, this 29-year-old female reported that she had highly arched feet since childhood, was unable to ice skate, and had slight difficulty walking on uneven ground with a tendency to trip. The examination showed slight wasting and weakness of intrinsic hand muscles, highly arched feet, and hammer toe formation but only a little weakness in ankle dorsiflexion and eversion. Ankle jerks were absent, and there was a mild impairment of sensation in her toes and soles. Her gait was normal.

Thus, the affected men and women developed symptoms and signs of a progressive, length-dependent neuropathy. Most males had clinical onset in the first decade, developed a significant gait disturbance in the second decade, with progression to the point of requiring aids, but did not become wheelchair bound. Women, by comparison, first noted symptoms toward the end of the second decade, and at every age, were much more mildly affected. Similarly, electrophysiological testing demonstrated that affected men had more axonal loss, including several who had an absent extensor digitorum brevis motor responses. Both men and women; however, had mild to moderate slowing of the peroneal motor response. Sensory nerve biopsies of two affected men (Patient A and Patient B) revealed evidence of demyelination and remyelination, as well as axonal degeneration and regeneration.

Biological basis

Etiology and pathogenesis

CMTX1 is caused by mutations in GJB1, the gene that encodes the gap junction protein connexin32 (Cx32). In the 100 years following Herringham's paper, many workers reported other CMTX1 kindreds from Europe and North America. Although the existence of X-linked kindreds with inherited neuropathy fell into dispute by 1980, their existence was confirmed by subsequent linkage studies. Early linkage studies excluded the distal short arm and the distal long arm of the X chromosome (80). Using restriction fragment length polymorphisms, many kindreds were linked to the proximal long arm of X chromosome. Further recombination analyses in several large families refined the localization of CMTX1 to an approximately 1.5 megabase interval in Xq13.1, where three genes previously had been mapped, including GJB1. Because mutations in genes expressed by myelinating Schwann cells were known to cause other inherited demyelinating neuropathies, Bergoffen and colleagues tested whether these candidate genes were expressed in normal rat peripheral nerves (13). GJB1 was the only one expressed, and direct sequencing of GJB1 in eight families demonstrated seven different mutations. Subsequent reports, largely from patients from Europe and North America but also from Russia, Japan, China, Korea, Turkey, and China, confirmed these findings. More than 730 different variants have been described; most of them are listed on the following website:Inherited Neuropathy Variant Browser website.

With the advent of population genetic databases such as gnomAD), it has become clear that not all variants in GJB1 cause CMTX1. In fact, some variants are simply too common to be disease-causing, and it is likely that some of the less common variants are also not pathogenic. Most reported variants in gnomAD are missense mutations (amino acid substitutions) that collectively affect most of the 283 amino acids of Cx32. As of October 2023, gnomAD reports 66 missense variants within the coding sequence of GJB1. The phenotype of most patients with pathogenic missense mutations is similar to that seen in patients with nonsense mutations (premature stop codons), insertions and deletions (many of which lead to frameshifts with premature truncation), and deletion of the entire gene. Thus, we should accept variants that segregate with a “CMTX1 phenotype” in individual families, especially if they have been reported more than once.

ClinVar has begun to collate variants, but the database is incomplete. More information about ClinVar can be accessed at the following website: https://www.ncbi.nlm.nih.gov/clinvar/. Of the 471 missense GJB1 variants classified by ClinVar as of October 2023, 131 are considered “pathogenic” or “likely pathogenic,” 302 are considered “uncertain,” 34 are considered “conflicting,” and four are “benign” or “likely benign.” Many variants have not yet been properly classified by ClinVar; many “uncertain” variants are “pathogenic” or “likely pathogenic” (they have been reported in people with CMTX1) or “benign” (GJB1 variants with allele counts in gnomAD greater than two are probably too common to cause CMTX1). There are also hundreds of additional missense mutations that are not listed in ClinVar.

The large-scale DNA sequencing from cross-sections of people has now revealed many benign polymorphisms in GJB1 and even more variants of uncertain significance. Of the 70 missense variants found in gnomAD, 10 have an established relationship with CMTX1 (patients with these mutations have been seen by a Charcot-Marie-Tooth disease expert), 41 have an unknown relationship with CMTX1 (the variant is present in gnomAD, but a patient with this variant has not been reported to have CMTX1), and 19 have a doubtful association or no association with CMTX1 because the variants are too common in gnomAD to be a cause of CMTX1 (ie, they are more prevalent than the most common disease-associated variants) (74). More information can be found on the following website: https://gnomad.broadinstitute.org/. Basic statistical analysis using the population frequency of the most common variants of CMTX1 suggests that most of the GJB1 variants listed in gnomAD that are currently of unknown relationship to CMTX1 will turn out to be benign.

Properly classifying as many GJB1 variants as possible is important to properly diagnose patients and to provide accurate genetic counseling. To accomplish this goal, one must consider the patient’s phenotype, the variant segregation in that family, whether the variant has previously been well-documented to cause CMTX1, and the variant prevalence in the population. In addition, because Cx32 forms a channel with well-described electrophysiological characteristics, Cx32 mutants that do not form normal channels are almost certain to cause CMTX1. Furthermore, because the mouse and human Cx32 proteins have nearly identical amino acid sequences, one can make a given GJB1 variant in mice and determine whether this causes neuropathy. Until a variant is conclusively classified, merely finding a GJB1 variant in a patient with neuropathy is insufficient evidence for concluding that the affected person has CMTX1.

Some CMTX1 kindreds (perhaps more than 10%) do not have mutations in the open reading frame (93). In these families, mutations might affect the promoter, splice sites, or the untranslated portions of the mRNA. In mammals, GJB1 comprises two exons and a large intron (6 to 8 kb). The second exon contains the entire open reading frame for Cx32. Cx32 transcripts in peripheral nerves are initiated at an alternative promoter, termed P2, which is located close to exon 2. Transcripts in the liver, brain, spinal cord, and pancreas, on the other hand, are initiated at the P1 and the P2 promoters. Several mutations are just proximal to the start site of transcription, likely affecting the EGR2 and SOX10 binding sites; these are important transcription factors for developing and maintaining myelinating Schwann cells (90). Other mutations, c.-17G>A (63) and likely c-17+1G>T (93), affect the splicing of exons 1A and 2 or abolish an internal ribosome entry site (-459C> T) in the 5' UTR that is essential for the translation of Cx32 mRNA (38). Similar to what was described above, there are variants in the 5’ and 3’ UTR that cause CMTX1 and others that are probably benign (93).

Connexins are a family of approximately 20 homologous integral membrane proteins in mammals. The high degree of homology suggests that their structure and function were conserved as they evolved from a common ancestral gene (98). Connexins form hexameric hemichannels (connexons), which dock with other hemichannels across apposed plasma membranes to form cell-cell channels. Cell-cell channels, in turn, are organized into large two-dimensional arrays called gap junctions. Almost all connexins have been shown to form gap junctions that transfer metabolites, signaling molecules and nucleic acids between cells (50; 68; 94). Most connexins are expressed in more than one tissue, and most tissues express more than one connexin. The possible interactions of different connexins are potentially complex, as hemichannels may be composed of a single connexin (homomeric hemichannels) or different connexins (heteromeric hemichannels). Furthermore, connexons can couple with hemichannels composed of the same connexin (homotypic junctions), with connexons composed of different connexins (heterotypic junctions), or even with different combinations of heteromeric connexons. Not all combinations of hemichannels, however, can form heterotypic junctions.

Cx32 was the first connexin to be cloned and was named according to the predicted molecular mass of the protein, 32 Kda (72). It is highly conserved; the amino acid sequence of human Cx32 protein is 98% identical to those of the mouse and rat. Despite the broad expression pattern of Cx32, peripheral neuropathy is usually the sole clinical manifestation of GJB1 mutations, although CNS abnormalities have been seen with certain mutations. Why these other tissues are not affected is unclear. One reason may be the co-expression of other connexins, which could protect against the loss of Cx32.

Until GJB1 mutations were found to cause CMTX1, little attention had been paid to the finding that myelin sheaths may contain gap junctions. Putative gap junctions were seen within paranodes and incisures by freeze-fracture electron microscopy, exactly where Cx32-immunoreactivity is found (13; 58). Paranodes and incisures are regions of non-compact myelin in the myelin sheath and contain different intrinsic membrane proteins from those of compact myelin.

Although there is no evidence for gap junction-mediated coupling between myelinating Schwann cells, the canonical view of the role of Cx32 in Schwann cells, based on antibody-based localization of Cx32 to the paranodes and Schmidt-Lantermann incisures, has been that it forms intracellular, gap junctions between adjacent layers of noncompact myelin, creating a “short circuit” reflexive pathway between adjacent layers of noncompact myelin (84). This, in turn, would shorten the distance between the abaxonally located nucleus and the adaxonal membrane 300- to 1000-fold (84; 68). If Cx32 mutants interrupt the function of these gap junctions, myelinating Schwann cells and their axons could be damaged, leading to demyelination and axonal loss. Dye transfer studies, a standard way of demonstrating dye-coupling between cells, have demonstrated functional gap junctions in the myelin sheath (10). Ex-vivo myelinated fibers were injected with dyes of differing molecular mass. Dyes of low molecular mass passed from the outer (abaxonal) collar of Schwann cell cytoplasm to the inner (adaxonal) collar of cytoplasm, whereas high molecular mass dyes did not reach the adaxonal cytoplasm. Furthermore, pre-incubating the teased fibers in an agent known to uncouple gap junctions prevented low molecular mass dyes from diffusing into the cytoplasm adjacent to the axon. These results indicate a gap junction-mediated pathway for diffusion of small molecules directly across the myelin sheath, probably located within incisures. However, this pathway did not appear to be disrupted in the myelinating Schwann cells from Gjb1-null mice, indicating that other connexins may compensate for the absence of Cx32. Cx31.3, the human ortholog of rodent Cx29, would be a strong candidate, except that it does not appear to form functional channels (05; 83).

Although there is no evidence for Schwann cell-to-Schwann cell coupling in the peripheral nervous system, in the central nervous system, Cx32 does appear to form functional homotypic gap junction channels between oligodendrocytes (56; 97). In vivo data also suggest that oligodendrocytes can couple heterotypically with astrocytes to form gap junctions (02). Thus, in the central nervous system, mutations in GJB1 could potentially interfere with both oligodendrocyte-oligodendrocyte and oligodendrocyte-astrocyte coupling.

Available clinical data, including the similarity of PNS phenotypes in patients with most point mutations, promotor mutations, upstream nonsense mutations, and deletion of the entire coding sequence of Cx32, support the notion that loss of function is sufficient to cause CMTX1 (86; 70; 93; 79); however, the exact nature of that loss of function is unclear. Exogenous expression studies in mammalian cell lines have shown that some mutations lead to disruption of the normal localization of connexin protein to gap junction plaques (22) and that this mislocalization may lead to degradation in one of two distinct pathways involving either proteasomes or lysosomes (95). Experiments in both Xenopus oocytes and mammalian cell lines have shown that in many cases, gap junction plaques form, but the resulting channels show functional alterations in either voltage-dependent gating or permeability, which disrupt the ability of Cx32 to form functional cell-cell channels (02). Insights from animal models of CMTX1 also support the notion that CMTX1 is a loss-of-function disorder. Our laboratory has confirmed that the phenotype of the mouse with targeted ablation of Gjb1 (a clear loss-of-function model) has a behavioral, electrical, and pathological phenotype that suggests it is a good model for CMTX1 (04). Furthermore, the phenotype is identical to that for two knock-in models expressing either the p.Thr55Ile or p.Arg75Trp mutations, suggesting these are all loss-of-function mutations. Interestingly, a different mutation, p.Arg15Gln, has also been characterized (91) and has a milder phenotype than the knockout, suggesting it may result in partial loss of function.

Notwithstanding the above-summarized studies, work from our laboratory and others suggests that cell-cell communication through gap junctions may not account for the full extent of the role played by Cx32 in glial cells or by connexins in other cell types (61; 67; 54; 28). Our finding particularly highlights that retention of the canonical connexin functions (ie, the ability to form morphologic plaques and functional gap junction channels), although protecting from CNS dysfunction, appears to be insufficient to protect against development of peripheral neuropathy (03). Although this may reflect differences in the behavior of connexin proteins in transfected cells versus the in vivo myelinating Schwann cell, it does suggest the possibility that Cx32 may have other non-gap junction-dependent functions in the Schwann cell. Furthermore, data from our lab and others suggest that dysfunction of hemichannels formed by connexin 32 may be involved in the development of Cx32-mediated peripheral neuropathy (28; 66; 18; 32; 78).

Schwann cells and oligodendroglial connexins may play a role in spatial buffering of potassium (K+) during neuronal activity. Intraocular tetrodotoxin, which suppresses retinal ganglion cell activity, reduced optic nerve vacuoles in mice lacking both oligodendroglial Cx32 and Cx47. Conversely, cholera toxin increases retinal ganglion cell activity and led to increased vacuolization (59). An alternative explanation for the findings of activity-dependent pathology is that loss of Cx32 may lead to energy failure. Connexins, particularly Cx43, mediate resistance to injury through expression in cardiomyocyte mitochondria (15). An unbiased approach to studying the interactome of Cx32 in the liver definitively identified Cx32 as a component of the inner mitochondrial membrane (26). At this point, however, no specific role for Cx32 has been identified in mitochondria. As noted above, the DWI abnormalities on MRI of patients with florid CNS syndromes may arise due to energy failure; these abnormalities virtually always involve the corpus callosum, a metabolically highly active brain region.

Immune mechanisms may contribute to the pathogenesis of CMTX1. Mice with targeted ablation of Gjb1 and recombination activating gene-1 (Rag1 mice lack mature T- and B-lymphocytes) showed reduced endoneurial macrophages and reduced nerve pathology (47). Blocking the colony-stimulating factor 1 (CSF-1) receptor kinase (c-FMS) in the Cx32 KO mouse reduced endoneurial macrophages and improved nerve function and morphology (45). Transcriptomic analysis has also identified a key role of immune responses in CMTX1. Freidin and colleagues compared gene expression in the sciatic nerve from 35-day-old WT and Cx32 KO animals (27). Four of the five top pathways represented in the differentially expressed genes at baseline are related to immune or inflammatory responses.

Epidemiology

Charcot-Marie-Tooth is a common genetic disease, with an estimated prevalence that ranges from ten to 82 per 100,000, depending on the geographic location (11). CMTX1 is the third most common form of Charcot-Marie-Tooth disease, after CMT1A and hereditary neuropathy with liability to pressure palsies, accounting for about 11% of all patients with Charcot-Marie-Tooth disease in a large CMT database (29).

Prevention

The manifestations of CMTX1 cannot be prevented by any known measures. Prenatal and preimplantation testing can be done.

Differential diagnosis

Confusing conditions

Inherited neuropathy should be considered in the differential diagnosis of every patient who has a chronic, length-dependent neuropathy of unknown etiology. A clear family history and electrophysiological evidence of a chronic, length-dependent neuropathy are the hallmarks of all inherited neuropathies. However, simplex cases of CMTX1 likely represent 5% to 10% of all CMTX1 cases (01). “De novo” mutations likely occur, though their frequency is unclear (25; 81); therefore, the absence of a family history does not exclude CMTX1 (or any other inherited neuropathy).

A diagnosis of CMTX1 should be considered for patients who appear to have CMT1, especially if the family history indicates that men are more affected than women and no male-to-male transmission is documented. The degree of slowing in motor responses helps to distinguish CMT1 from CMTX1, as forearm motor velocities in CMT1 (10 to 40 m/sec) are often slower than in CMTX1 (30 to 40 m/sec). This so-called "intermediate slowing" is typical for affected men but is also seen in affected women (70). The nerve biopsies from patients with CMTX1 show more axonal loss and “regenerated clusters” of myelinated axons and fewer "onion bulbs" as opposed to CMT1A (82; 34; 96).

Distinguishing CMTX1 from CMT2 or "dominant intermediate" forms of Charcot-Marie-Tooth disease is more problematic, especially in small kindreds where the lack of male-to-male transmission is not evident. Further, the clinical phenotypes of CMTX1 and CMT2 can be similar, and the ranges of motor nerve conduction velocities overlap.

Finally, it should be emphasized that atypical cases of CMTX1 have also been described, largely owing to the availability of genetic testing: (1) GJB1 mutations have been found in patients who are suspected of having inflammatory demyelinating neuropathies not responding to treatment (60). (2) Hearing loss and fixed abnormalities in the central nervous system have been described. (3) CMTX1 can affect young women and children of either sex. In particular, stroke-like episodes of CNS dysfunction that are accompanied by MRI abnormalities have been described in patients with no history of Charcot-Marie-Tooth disease as well as in patients with a history of the disease (07; 65). These episodes may recur.

The differential diagnosis of stroke-like episodes and peripheral neuropathy is extensive but includes MELAS (mitochondrial encephalopathy lactic acidosis and stroke) or other mitochondrial diseases, POLG-related disorders, CADASIL (cerebral autosomal dominant arteriopathy with subcortical infarcts and leukoencephalopathy), neurosarcoid, collagen vascular diseases, and vasculitis, though most of these will be distinguishable with appropriate physical examination and laboratory work-up. The stroke-like episodes seen in acute reversible leukoencephalopathy with increased urinary alpha-ketoglutarate (ARLIAK) due to mutations in SLC13A3 are very similar to those seen in CMTX1 (23). Other metabolic conditions may also present with stroke-like episodes (57). In the absence of clear evidence of neuropathy, acute disseminated encephalomyelitis, or new-onset multiple sclerosis, especially in children, may also be diagnostic considerations.

Diagnostic workup

A family history of neuropathy and a physical exam that reveals features of a length-dependent peripheral neuropathy are strongly suggestive of an inherited neuropathy. Electrophysiological evaluation of multiple peripheral nerves along with electromyography will confirm whether the patient has a length-dependent neuropathy and should also establish whether this is associated with slow (less than 30 m/sec), intermediate (30 to 40 m/sec), or normal (greater than 45 m/sec) motor conduction velocities in the arms. Symmetrical slowing indicates an inherited demyelinating neuropathy; asymmetrical slowing is characteristic of an acquired, demyelinating neuropathy. In CMTX1, however, the nerve conduction velocities are less uniform than in other forms of CMT1 (33; 53), to the point that some patients were treated (unsuccessfully) for CIDP (60). The cerebrospinal fluid from CMTX1 patients has not been systematically characterized, but reports of mild elevations (less than 100 mg/dl) may distinguish CMTX1 from chronic, acquired demyelinating neuropathies.

Patients who are suspected of having a genetic neuropathy with slow or intermediate motor conduction velocities should be advised that this could be a genetic neuropathy and that genetic testing is available. It would be reasonable to refer the patient to a neurologist specializing in neuromuscular diseases or a genetic counselor familiar with Charcot-Marie-Tooth disease. After explaining the genetic testing, cheek swab samples can be sent for testing. For CMTX1, the entire open reading frame of the GJB1 gene should sequenced. If no mutations are found, and CMTX1 is strongly suspected, the promoter and 5’ untranslated region could be sequenced. Updated information on the availability of genetic testing can be found on the following website: GeneTests Website.

A nerve biopsy is not required to diagnose CMTX1.

Management

Interventional approaches. The management issues are chiefly related to the orthopedic complications of the disease (71). These include the well-known foot deformities common to all types of Charcot-Marie-Tooth disease: pes cavus, varus, callosities on the lateral foot borders, and "hammer toes," which can compromise ambulation. Ankle instability can be treated with high-top shoes, boots, or orthoses, and foot drop can be treated with braces. In my opinion, surgical intervention should be reserved for cases where orthotics or bracing cannot adequately correct functional impairments. Scoliosis has been noted in more severely affected patients, but the literature does not indicate that patients with CMTX1 typically need intervention (37). Physical therapy, especially stretching for contractures, is advocated, and splinting, specific exercises, adaptive devices, and, in rare cases, surgery may help maintain hand function.

Although skin breakdown and trophic ulcers do not appear to be as problematic as in other neuropathies, it is prudent to advise patients about foot care. Similarly, pain is less common in patients with CMTX1 than in patients with CMT1A but should be addressed when present (73).

Patients with CMTX1 have not been reported to develop a superimposed inflammatory demyelinating neuropathy.

Patients with any inherited neuropathy may be at increased risk of developing neuropathy if they are exposed to agents that can cause neuropathy. However, the risks of a particular neurotoxic medication must be weighed against the potential benefits of the medication and the efficacy of less toxic alternatives.

Regarding CNS manifestations, it is difficult to predict with certainty which patients will develop stroke-like episodes. However, recurrences are not uncommon, and some mutations appear to be more commonly associated with this phenotype. Patients should be advised that these attacks are generally rare, but patients at higher risk should try to avoid precipitant stressors, such as hyperventilation or exertion, fever, or minor infections, and they should take care to re-acclimatize after return from extremely high altitude.

Treatments on the horizon. Work in animal models suggests several potential translatable avenues for treating CMTX1. These include gene replacement therapy via AAV9-mediated Schwann cell expression of wild-type Cx32 (41), virally-mediated expression of neurotropic factor NT-3 in skeletal muscle (69), and neuroprotection via modulation of the expression of HSP70 (43). A report by Klein and colleagues suggests that physical exercise may improve neuropathy in a CMTX1 model by reducing inflammation (46).

Special considerations

Pregnancy

Women with CMTX1 are fertile. Pregnancy, labor, and delivery have not been reported to alter the disease generally. There is a single report of recurrent postpartum episodes of neurologic dysfunction in a woman carrying the Ala39Val mutation in GJB1 (77).

Anesthesia

There is no specific information available about anesthesia and CMTX1. Although there are theoretical reasons for avoiding depolarizing neuromuscular agents such as succinylcholine during anesthesia because denervated muscle fibers might release potassium, thereby causing hyperkalemia, this has not been observed in patients with CMT1 (08). Patients with CMT1 may have increased sensitivity to thiopental, a drug widely used during the induction of anesthesia (48), and vecuronium, a neuromuscular blocker (75).