Genetics of hereditary motor and sensory neuropathy and the Costa Rican contribution

Hereditary motor and sensory neuropathy (HMSN) or Charcot-Marie-Tooth disease (CMT) is the most common hereditary illness of the peripheral nervous system. The genetics and the physiopathological aspects of the disease clarified until know, are here summarized. More than twenty genes and ten additional loci have been related with HMSN. These findings contribute to understand the metabolism of peripheral nerves and give the basis for molecular diagnostics and future therapy. Several Costa Rican families with CMT have been identified, specially with axonal forms. Two families present mutations in the myelin protein zero gene (MPZ). In addition, linkage have been found between the disease and locus 19q13.3 in an extended family, and a mutation segregating with the disease is present in a candidate gene of the critical interval. Costa Rica has several advantages for genetical studies, that can contribute importantly in the generation of knowledge in the neurogenetical field. Rev. Biol. Trop. 52(3): 475-483. Epub 2004 Dic 15.

Hereditary motor and sensory neuropathy (HMSN), also known as Charcot-Marie-Tooth disease (CMT), is a heterogenous neuropathical group characterized by progressive muscular weakness and atrophy of the distal muscles, often associated with mild to moderate sensory loss, foot deformities including pes cavus and pes planus, gait disturbance, and depressed deep tendon reflexes (Lupski 1998).Prevalence of CMT is estimated to be 1 in 2500-5000 people, constituting thus, the most common hereditary disease of the peripheral nervous system (Combarros et al. 1987, Holmberg 1993, Lupski et al. 1993, Skre 1974).By means of electrophysiological and morphological studies, CMT can be divided into a predominantly demyelinating type (HMSN type I) and an axonal form (HMSN type II), the latter being less frequent (Dyck et al. 1993).Primary peripheral demyelinating neuropathy is characterized by demyelination with severely reduced motor nerve conduction velocity (NCV) due to a partial or complete loss of the myelin sheath.CMT1 (autosomal dominant) and CMT4 (autosomal recessive) are forms distinguished by a slow nerve conduction, typically 10 to 30 meters per second (less than the minimal normal threshold of 38 m/s) (Table 1).On the other hand, primary axonal neuropathy results in axonal loss and normal or mildly reduced NCV (Dyck andLambert 1968a 1968b).CMT2 (autosomal dominant) and ARCMT2 (autosomal recessive) refer to an hereditary motor and sensory neuropathy in which nerve conduction velocity is usually within the normal range, although occasionally falling, into the low normal or mildly abnormal range (35-50 m/s), but presenting reduced compound muscle action potential (CMAP) (Dyck et al. 1993).Although CMT1 and CMT2 are often clinically distinguishable, their phenotypic resemblance suggests some degree of similarity in the pathological process (Mersiyanova et al. 2000a), probably due to the close relationship between the axon and the myelin: damage in the axon leads to myelin lesion, and vice versa.This is particularly noticeable in nerve biopsies of CMT2 patients.
Around 80% of the CMT patients have a demyelinating autosomal dominant form of HMSN (Nelis et al. 1996).Clinical heterogeneity within the demyelinating forms include cases with pyramidal tract features, deafness or other signs (Dyck et al. 1993).The clinical features of CMT2 patients overlap extensively with CMT1, although, patients with CMT2 tend to be less disabled and have less sensory loss, whereas X-linked form of CMT may present an axonally pronounced form of CMT that could be confused with CMT2 (Young et al. 2001).As a matter of fact, it is not always easy to distinguish between HMSN and non-genetic neuropathy, acquired through alcoholism, vitamin 12 deficiency, thyroid disease, diabetes mellitus, HIV infection, vasculitis, leprosy, neurosyphilis, amyloid associated with chronic inflammation, occult neoplasm, or severe metal intoxication (Bird 2002).Sural nerve biopsy can help to properly diagnose CMT, but the HMSN subtype can only be identified accurately if the molecular origin of the pathology is known.

GENETIC HETEROGENEITY AND MOLECULAR CELL BIOLOGY OF CMT
Multiple genes and loci related to this neuropathy have been identified.All of them follow Mendelian patterns of inheritance (Table 1).The high clinical and genetic heterogeneity in CMT is remarkable.Several genetic loci and genes have been related with primary demyelinating neuropathy, including altered dosage of the peripheral myelin protein 22 gene (PMP22).Seventy percent of patients with a demyelinating form carry a 1.4 Mb duplication in chromosome 17p, where PMP22 is located (Lupski et al. 1991, Patel and Lupski 1994, Timmerman et al. 1992).Additionally, CMT1 has been associated with mutations at the following genes: PMP22, connexin 32 gene or gap junction protein B1 gene (GJB1), the myelin protein zero gene (MPZ), the early growth response gene 2 (EGR2), the myotubularin-related protein 2 gene (MTMR2), the Nmyc downstream-regulated protein 2 gene (NDRG1) and the periaxin gene (PRX) (Nelis et al. 2004).
Intracellular retention in myelinating Schwann cells is probably the most common disease mechanism of PMP22 missense mutations (Naef and Suter 1999).It has been proposed that PMP22 may participate in cell proliferation, differentiation, apoptosis and cell shape definition (Jetten and Suter 2000).Delayed myelination in mice without PMP22 suggests some function in Schwann cell differentiation and the myelination process (Adlkofer et al. 1995).Both PMP22 and MPZ are components of the myelin membrane, where these two proteins form complexes (D`Urso et al. 1999).MPZ is the major component of the compact myelin sheath in the PNS (Suter 1997).Most MPZ mutations are associated with a CMT1B phenotype, but in some cases, primary axonal involvement has been found.GJB1 is a non-compact myelin protein which forms a pore unit through the myelin (Balice-Gordon et al. 1998).It has been shown, that some mutations do not disrupt the cell-cell channels, but increase the opening of hemi-channels in the non-junctional surface membrane.This increase, may damage the cells through the loss of ionic gradients and small metabolites and an increased influx of Ca2+, leading to Schwann cell damage (Abrams et al. 2002).Another gene related with demyelinating neuropathology is MTMR2, that encodes for a phosphatase that may lead to CMT4B1 through a malfunction of neural membrane recycling, membrane trafficking, endocytic or exocytotic processes, combined with altered axon-Schwann cell interaction (Berger et al. 2002).Like PMP22, NDRG1 possibly plays a role in growth arrest and cell differentiation and may be related to the complex developmental transition of the Schwann cells and Schwann cell-axonal interaction (Kalaydjieva et al. 2000).Recently, Senderek et al. (2003) found a positional candidate gene within the CMT4B2 interval, with 30% homology to MTMR2: the SET binding factor 2 gene (SBF2), encoding a member of the myotubularin family.They sequenced this gene in patients and controls and found mutations that cosegregate with the disease.Since SBF2 lacks phosphatase activity, it would act differently from MTMR2, but maybe they converge in a common pathway controlling the myelin formation in the PNS.Azzedine et al. (2003) also found mutations in this gene (that they called MTMR13) which causes demyelinating CMT and early-onset glaucoma.They proposed that this pseudophosphatase could be important in the development of both the peripheral nerve and the trabeculum trabecular meshwork, which permits the outflow of the aqueous humor from the eye; it is worth to remark that both of these tissues derive from the endoderm.EGR2 is a zinc-finger transcription factor that regulates the expression of myelin proteins.Mutations in the DNA-binding domain lead to dominantly inherited forms, but mutations in the NAB-protein-binding site lead to recessively inherited forms (Warner et al. 1999).Another protein, PRX, is a non-compact myelin cytoskeleton-associated protein, that links the cytoskeleton of the Schwann cell with the extra-cellular matrix (Boerkoel et al. 2002, Sherman et al. 2001).Mutations cause the CMT4F subtype or Dejerine-Sottas syndrome.
On the other hand, the CMT1C syndrome is caused by mutations in a lipopolysaccharideinduced transcription factor (LITAF), that regulates the tumor necrosis factor alpha (Myokai et al. 1999, Street et al. 2003).How these mutations lead to demyelinating neuropathy is unknown.In addition, Parkinson et al. (2001) reported that mutations in the murine beta-spectrin 4 gene (Spn4) are responsible for deafness and motor neuropathy.These mutations cause alterations in the ion channel localization in the myelinated nerves, however the human homologue has not been implicated as responsible for any neuropathy.
Although the prevalence of primary peripheral axonal neuropathy is lower than demyelinating neuropathy, a considerable number of loci and genes have been related with this form.The neurofilament light chain gene (NEFL) (Mersiyanova et al. 2000b), the kinesin 1B-beta gene (KIF1Bβ) (Zhao et al. 2001), the mitochondrial GTPase mitofusin 2 gene (MFN2) (Zuechner et al. 2004); the gigaxonin gene (GAN1) (Bomont et al. 2000), the ganglioside-induced differentiation-associated protein 1 gene (GDAP1) (Baxter et al. 2002;Cuesta et al. 2002), the RAS-associated protein 7 gene (RAB7) (Verhoeven et al. 2003), the Lamin A/C gene (LMNA) (De Sandre-Giovannoli et al. 2002), and the glycyl tRNA synthetase gene (GARS) (Antonellis et al. 2003) have been associated with HMSN type II.Mutations in NEFL were found to be a cause of CMT2E.It encodes for one of the three neurofilament proteins in higher eukaryotes (together with medium and high neurofilaments).Brownlees et al. (2002) demonstrated that NEFL mutant neurofilaments assembly and their axonal transport in mammalian cells and neurons is disrupted.Disorganization of the neurofilament network is also present in patients with giant axonal neuropathy (affecting both CNS and PNS), due to mutations in the GAN1 gene, a member of the cytoskeletal BTB/kelch repeat family, which is involved in the organization of the actin network (Bomont et al. 2000).Zhao et al. (2001) obtained evidence that the KIF1Bβ-a protein related with CMT2A (chr.1p36) -transports synaptic vesicle precursors along the microtubules of the neurons.However, an isoform, KIF1Bα, is translated in other tissues and participates in mitochondrial transportation.Recently, Zuechner et al. (2004) found that patients with CMT2A did not present any mutation in KIF1Bβ, but in the MFN2A located at 1p36 too, a gene involved in the control of the mitochondrial morphology.After this finding, several families have been found with CMT2A and mutations in MFN2A.Additionally, two research groups found mutations in the GDAP1 gene, related with both axonal and demyelinating phenotypes.GDAP1 presents a strong amino acid sequence similarity to glutathione S-transferases (GSTs) which are enzymes that play a role in cell detoxification.Cuesta et al. (2002) proposed that the mutated GDAP1 might prevent the correct catalysing S-conjugation of reduced glutathione (antioxidant), resulting in a progressive attrition of both axons and Schwann cells.Mutations in the small GTP-ase late endosomal protein RAB7 have also been related with axonal neuropathy.RAB7 belongs to the Rab family of Ras-related GTP-ases involed in the regulation of intra-cellular membrane trafficking and may play a role in linking vesicles and target membranes to the cytoskeleton.In fact, the Rab7 effector protein RILP controls lysosomal transport by inducing the recruitment of dynein-dynactin motors (Jordens et al. 2001), although it is still unknown as to how mutations in a universally expressed protein like this cause axonal CMT (Verhoeven et al. 2003).Recently, Antonellis et al. (2003) found that mutations in the glycil tRNA synthetase, an ubiquitously expressed gene, cause CMT2D and distal spinal muscular atrophy type V. Finally, the relationship between LMNA and CMT is of particular interest, since this protein was previously associated with limb-girdle muscular dystrophy type 1B, autosomal dominant Emery-Dreifuss muscular dystrophy, dilated cardiomyopathy type 1A, and autosomal dominant partial lipodystrophy.The relationship of this gene with autosomal recessive CMT type 2 (ARCMT2A) suggests that mutations in different protein domains lead to very different pathologies (De Sandre-Giovannoli et al. 2002).

THE COSTA RICAN CONTRIBUTION
In Costa Rica, epidemiological data for CMT are lacking.Nevertheless, since the public national health system is centralized, with four reference Hospitals for neuropathical cases, it is relatively easy to find patients affected by CMT and to approach their families.We have localized several families with the disease, and remarkably, most of them present an axonal polyneuropathy.Until now, our group has extensively studied five of these CMT families.
Among them, two mutations in the myelin protein zero gene have been identified as responsible for the disease in two families with both demielinating and axonal peripheral neuropathy.A Thr124Met mutation is responsible for the disease in the CR-C family with five affected individuals, all of them heterozygous for the mutation (unpublished data).In this family, affected individuals also show pupillary abnormalities and deafness.The molecular pathology of this mutation remains unclear.There may be a relationship with the role of the changed amino acid, since the variation is part of the extracellular loop close to the glycosylation site at codon 122 and the cysteine residue at codon 127, which is involved in the formation of a disulphide bridge, but it is currently unknown whether and how this mutation disrupts the normal adhesion function of the MPZ molecule (De Jonghe et al. 1999).
In family CR-D, a novel Tyr145Ser MPZ mutation was found in four affected members who also are heterozygous.In addition, two siblings were homozygous for this mutation.On neurological examination, the heterozygous parents as well as their homozygous children showed distal sensory deficits.The mother and the siblings were affected by impaired deep tendon reflexes and mild sensory ataxia; however, the homozygous individuals had an earlier age of onset and more severe distal motor weakness and pupillary abnormalities (Leal et al. 2003).It is possible to hypothesize that the Tyr145Ser MPZ mutation could cause a mild dominant form of CMT, because of a less aggressive dominant-negative effect in heterozygous patients.In the case of homozygosity, the fundamental role of P0 in adhesion and compaction would be only carried out by subnormal proteins, resulting in more severe neuropathy.
Linkage analysis was carried out in three other families affected with autosomal recessive CMT2: CR-A, CR-SR and CR-P, but in the first two families, it did not show any significant statistical relationship between the disease and a specific chromosomal region.However, after genome-wide search in the inbred family CR-P, linkage analysis of a subgroup comprising three branches of the pedigree detected linkage to the 19q13.3region and subsequent homozygosity mapping defined shared haplotypes between marker D19S902 and D19S907 (Leal et al. 2001).A maximum two-point LOD score was obtained for marker D19S867 (Z max = 9.08) at a recombination fraction of 0.00.Clinically, the age of onset of chronic symmetric sensorymotor polyneuropathy was 28 to 42 (mean 33.8), and the electrophysiologic data clearly reflect an axonal degenerative process.Thus, phenotypically and genotypically this is a different entity from the demyelinating CMT4F mapped to 19q13.1-13.3(Delague et al. 2000), hence the disease affecting the CR-P family constitutes an axonal, autosomal recessive CMT subtype (ARCMT2B).After the definition of the critical interval in chromosome 19q13.3for the CR-P family, a genomic sequence contig was built up of the critical region giving some structure to the information of human draft genome sequence databases.All the genes present in this interval were analyzed, and a mutation in one gene was identified, that probably cause ARCMT2B (unpublished data).The mutation in this gene encoded protein occurred in a evolutionary domain which is highly conserved and the domain indicate close homology to a protein that plays a role in neurons, in a region with adhesion function.This domain also presents homology to a gene responsible for a neuropathy in mice.Functional analysis and screening in additional families will be performed for confirmation purposes.If this polymorphism is causative of HMSN type II, it would be a very valuable contribution in the understanding of CMT, the physiopathology of hereditary neuropathy and the function of the peripheral nerves.
Following these results, research in Costa Rican families can provide also a useful instrument for the diagnosis, prevention, and possible future therapy.Since in the near future a molecular diagnostics system for HMSN will be established in Costa Rica, both carrier's test (with counseling) and contacts with interesting cases can improve.Candidate genes are going to be tested according to the phenotype; the duplication in chromosome 17, and mutations in PMP22, Cx32 and MPZ, as well as in other genes will be searched.This system will be available for neurologists and other physicians at the University of Costa Rica.Also orthopaedical and pharmacological therapies are going to be offered to patients in Costa Rica soon.All these advances announce a development in the study of neurogenetics of HMSN in Costa Rica, and also a in the medical care offered to patients.In addition, in our proteomics age, expression patterns and interactions between CMT-related proteins, and the role of these phenomena in the phenotype determination will be studied, as well as the contribution to the peripheral nerve metabolism of interactions between these proteins and the environment.

ACKNOWLEDGMENTS
The author would like to thank Bernd Rautenstrauss, André Reis and Jorge Azofeifa for their advices and suggestions for this manuscript.Support from the Vicerrectoría de Investigación, University of Costa Rica (projects 742-98-241 and 742-A4-334) and Institute of Human Genetics, University of Erlangen-Nuremberg, is gratefully acknowledged.

TABLE 1
(Pestronk 2004)netic classification of hereditary motor and sensory neuropathy, with the linked locus and gene if known(Pestronk 2004)