NEUROLOGICAL PROGRESS Hereditary Motor and Sensory Neuropathy with Agenesis of the Corpus Callosum Nicolas Dupré, MD, MSc,1 Heidi C. Howard, PhD,1 Jean Mathieu, MD, MSc,2 George Karpati, MD,3 Michel Vanasse, MD,4 Jean-Pierre Bouchard, MD,5 Stirling Carpenter, MD,6 and Guy A. Rouleau, MD, PhD1 Hereditary motor and sensory neuropathy associated with agenesis of the corpus callosum (OMIM 218000) is an autosomal recessive disease of early onset characterized by a delay in developmental milestones, a severe sensory-motor polyneuropathy with areflexia, a variable degree of agenesis of the corpus callosum, amyotrophy, hypotonia, and cognitive impairment. Although this disorder has rarely been reported worldwide, it has a high prevalence in the SaguenayLac-St-Jean region of the province of Quebec (Canada) predominantly because of a founder effect. The gene defect responsible for this disorder recently has been identified, and it is a protein-truncating mutation in the SLC12A6 gene, which codes for a cotransporter protein known as KCC3. Herein, we provide the first extensive review of this disorder, covering epidemiological, clinical, and molecular genetic studies. Ann Neurol 2003;54:9 –18 Hereditary motor and sensory neuropathy associated with agenesis of the corpus callosum (HMSN/ACC) (OMIM 218000), also known as peripheral neuropathy associated with agenesis of the corpus callosum (ACCPN) or Andermann syndrome, is an autosomal recessive disease of early onset. It is characterized by a delay in developmental milestones, a severe progressive sensory-motor neuropathy with areflexia, a variable degree of agenesis of the corpus callosum, amyotrophy, hypotonia, cognitive impairment, an atypical psychosis, and occasional dysmorphic features (high arched palate, syndactyly). Howard and colleagues1 recently identified the gene responsible for HMSN/ACC, solute carrier family 12 member 6 (SLC12A6), which codes for K⫹-Cl⫺ cotransporter 3 (KCC3). In this review, we present a complete summary on the current state of knowledge of HMSN/ACC. Although rare, this disorder is particularly interesting because it represents a unique example of a gene defect causing both developmental and neurodegenerative problems involving both the central and the peripheral nervous systems. The callosal agenesis indicates a defect in axonal migration during embryogenesis, whereas the progressive neuropathy along with the cognitive involvement and the atypical psychotic occurrences suggest a degenerative From the 1Centre for Research in Neurosciences and the Department of Neurology and Neurosurgery, McGill University, Montreal; 2Hôtel-Dieu de Chicoution, Chicoutioni; 3Montreal Neurological Hospital and Institute, Department of Neurology and Neurosurgery, McGill University, Montreal; 4Hôpital Ste-Justine, Université de Montréal, Montreal; 5Department of Neurology, Hôpital de l’Enfant-Jésus, Laval University, Quebec City, Quebec, Canada; and 6Hospital São João, Porto, Portugal. Received Jan 2, 2003, and in revised form Mar 20. Accepted for publication Mar 20, 2003. process. The severe sensory-motor neuropathy is striking both for its distinctive pathological features (axonal swelling) and for the distribution of its abnormalities (cranial nerves, nerve roots, peripheral nerves). HMSN/ ACC is an ion transporter disorder that is quite different from other diseases caused by dysfunctional ion channels. The fact that the K-Cl cotransporter 3 protein is at least indirectly involved in early axonal migration across the corpus callosum and in maintenance of the peripheral nervous system (PNS) throughout life can lead the way to a generation of new hypotheses on ion transporter function. Epidemiological Studies HMSN/ACC is found mainly in the French Canadian (FC) population of Quebec, and specifically in two regions of northeastern Quebec, the Saguenay-LacSt-Jean (SLSJ) region, and the Charlevoix County.2,3 In a study based on 101 individuals distributed in 82 families, De Braekeleer and colleagues2 analyzed the geographical distribution of HMSN/ACC in northeastern Quebec. The overall incidence at birth was 1 in 2,117 live births and the carrier rate was 1 in 23 inhabitants. Genealogical reconstruction was attempted in 161 obligate carriers, and a set of 22 founders was Address correspondence to Dr Rouleau, Centre for Research in Neurosciences and the Department of Neurology and Neurosurgery, McGill University, 1650 Cedar Avenue, Montreal, Quebec, H3G IA4 Canada. E-mail: email@example.com © 2003 American Neurological Association Published by Wiley-Liss, Inc., through Wiley Subscription Services 9 common to all of them. All these founders originated from France, including 3 from Poitou, 11 from Perche, and 4 from unknown regions. The authors concluded that based on genealogical data, the founders who came from France in the late 17th century probably established themselves in the Charlevoix County where they lived until their descendants moved to SLSJ in the early 19th century. The Charlevoix County contributed 70% of the immigrants to SLSJ during the first 30 years of settlement, and, because migration was mainly of familial type,4 many child carriers of the HMSN/ACC gene may have entered the SLSJ region with their parents, contributing to the high gene frequency in the region. The most represented founder couple in SLSJ was found in only 8 of the 75 complete genealogies, indicating that quite a few carriers must have introduced the HMSN/ACC gene in this region. We recently performed an updated survey of the geographical distribution of HMSN/ACC cases across Quebec, by collecting data from each major pediatric hospital and rehabilitation center of the province (Fig 1). These results show that despite important migration within the province over the past 20 years, the disease remains concentrated in the SLSJ region. Clinical Studies Clinical Description In 1966, Leblanc and colleagues5 first reported the existence of HMSN/ACC. He recognized that this “syndrome of callosal agenesis” had a familial component and was particularly frequent in the SLSJ region of Quebec. Subsequently, Andermann and colleagues6 described HMSN/ACC as an autosomal recessive disease characterized by agenesis of the corpus callosum, a motor neuropathy (which they thought was compatible with anterior horn cell disease), mental retardation, areflexia, amyotrophy, hypotonia, and dysmorphic features. Subsequent descriptions of the disease recognized that it involved a sensory-motor neuropathy rather than an anterior horn cell disease.7 Mathieu and colleagues8 studied 64 patients between the ages of 2 and 34 years old. Forty-five (70%) of the 64 patients from this phenotypic study were later tested for the HMSN/ ACC gene mutation. They were all found to be homozygous for a single nucleotide deletion in exon 18 of the SLC12A6 gene (c.2436delG, Thr813fsX813), thereby providing molecular confirmation of the diagnosis and confirming the genetic homogeneity of this population (N. Dupré, unpublished data). Mathieu and colleagues8 found the following clinical characteristics in these patients: ptosis (59%), upper gaze palsy (30%), facial asymmetry (34%), areflexia (100%), hypotonia (36%), amyotrophy (86%), tremor (25%), seizures (17%), early Achilles’ tendon retraction (47%), scoliosis (86%). Dysmorphism also was evaluated in a 10 Annals of Neurology Vol 54 No 1 July 2003 systematic way, with the following characteristics among their 64 patients: brachycephaly (16%), higharched palate (39%), overriding of the first toe (16%), and partial syndactyly of second to third toes (8%). Clinical evolution was quantified with the following measures: average age of onset of walking 3.8 years (47 patients), average age of end of walking capacity 13.8 years (35 patients), average age of appearance of scoliosis 10.4 years (22 patients), and average age of death 24.8 years (6 patients). We recently reviewed 49 FC HMSN/ ACC cases (subjects who are now deceased) and found the average age of death to be 33 years; this is a longer survival period than previously estimated (N. Dupré, unpublished data). The main features of ACCPN in the FC population can be found in Table 1. Cognitive Function The cognitive function of individuals with HMSN/ ACC has been addressed in relatively few studies. In their clinical study of 64 cases, Mathieu and colleagues8 evaluated 53 subjects for cognitive dysfunction. Using the clinical classification of Taft9 to stratify the degree of mental retardation, they found that 8% of cases had normal intelligence, 49% had mild mental retardation, 40% had moderate mental retardation, and 4% had severe mental retardation. They did not, however, indicate the age at which the patients were evaluated, which, in light of the progressive nature of the disease, would be relevant information. They also reported that 39% of patients developed “psychotic episodes” after the age of 15 years. Paranoid delusions, depressive states, visual hallucinations, auditory hallucinations, or “autistic-like” features characterized these episodes. Filteau and colleagues10 studied a cohort of 62 patients to look at the relationship between imaging features and the appearance of “psychotic episodes” in HMSN/ACC. They identified 20 patients (32%) who had presented such episodes, with a mean age at the time of evaluation of 27.6 years and a mean age of onset of psychosis of 19.5 years. The psychiatric evaluation was performed using a semistructured interview according to Diagnostic and Statistical Manual III-R criteria. They found that all 20 identified patients had paranoid delusions often accompanied by visual or auditory hallucinations described as “monsters” or “persecutors.” Interestingly, it appears that psychosis was more common in HMSN/ACC patients (32%) than in adult populations of mentally retarded individuals (18%).11 Electrophysiology We recently reviewed the electrophysiological studies of 48 patients (aged 0 to 35 years) evaluated in Chicoutimi by one of the authors (J.M) over the last 20 years (Fig 2). We found that the median motor nerve conduction velocities ranged from 16 to 57m/sec Fig 1. Distribution of all hereditary motor and sensory neuropathy associated with agenesis of the corpus callosum cases seen by a neurologist in the province of Quebec as of July 2002. Each case is represented by a black dot. (A) Distribution throughout the entire province. (B) Distribution within the Saguenay-Lac-St-Jean region. (mean, 33m/sec) between 0 and 2 years; from 26 to 45m/sec (mean, 34m/sec) between 2 and 15 years; from 11 to 39m/sec (mean, 23m/sec) between 15 and 35 years. These conduction velocities show a wide rage of variability that are unlikely to result from technical factors because they all were performed by the same individual using standardized protocols. It is difficult to classify the neuropathy as primarily “axonal” or primarily “demyelinating,” because the range of conduc- tion velocities straddles the usual cutoff values used to make this distinction.12 The sensory nerve action potentials recorded either at the median, ulnar, or sural nerves were never obtainable even in patients at a young age. On needle electromyogram, fibrillations were found in 11 patients (23%) and were, in general, not abundant. Abnormal motor unit potentials were found in 12 patients (25%) and ranged from polyphasic to, rarely, giant potentials. Dupré et al: Motor and Hereditary Polyneuropathy 11 Table 1. Features of HMSN/ACC in the French Canadian Patients Clinical features Dysmorphic traits Hypertelorism (usually mild) Syndactyly of second and third toes Brachycephaly High-arched palate Overriding of the first toe Cognitive Function Mild to severe mental retardation “Psychotic episodes” appearing usually during adolescence Cranial Nerves Eyelid ptosis (symmetrical or asymmetrical) Facial plegia (symmetrical or asymmetrical; sometimes associated with hemifacial atrophy) Esotropia or exotropia (due to variable combinations of oculomotor nerve palsies) Upper gaze palsy Horizontal nystagmus Motor Progressive distal and proximal symmetrical limb weakness with muscle atrophy (the infant is invariably hypotonic in the first year of life, learns to walk with a device between 2 and 5 yr old and eventually becomes confined to a wheelchair) Diffuse limb tremor (probably due to the polyneuropathy) Reflexes Invariably absent from infancy Sensory All modalities are moderately to severely affected from infancy Other Scoliosis Pulmonary restrictive syndrome Seizures (infrequent) Investigation Lumbar puncture Usually mild elevation of proteins Nerve conductions and EMG in young children Invariably absent sensory potentials (sural, median, cubital) from infancy Variable motor nerve conduction velocities (median, cubital, tibial) Fibrillations EEG Usually normal Imaging May have no ACC, partial ACC, or complete ACC May see mild cortical or cerebellar atrophy in older subjects Sural nerve biopsy and muscle biopsy Sural nerve biopsies show a lack of large myelinated fibers, signs of axonal loss (ovoids of Wallerian degeneration), and some enlarged axons that on electron microscopy show a decreased density of neurofilaments Muscle biopsies show nonspecific signs of chronic denervation atrophy Autopsy The hallmark of autopsies is the swollen axons demonstrated in cranial nerve samples (third and seventh especially) as well as in dorsal and ventral nerve roots. Swollen axons can also be found scattered in the white matter The brain shows either no ACC, partial ACC, or complete ACC with preservation of Probst bundle HMSN/ACC ⫽ hereditary motor and sensory neuropathy with agenesis of the corpus callosum; EMG ⫽ electromyogram; EEG ⫽ electroencephalogram. 12 Annals of Neurology Vol 54 No 1 July 2003 Fig 2. Graph representing the median motor nerve conduction velocity (m/sec) as a function of the age in 48 French Canadian patients with hereditary motor and sensory neuropathy associated with agenesis of the corpus callosum. Most patients were evaluated at a young age when they presented with delayed motor development and hypotonia. MNCV ⫽ motor nerve conduction velocity. Imaging In their first report of the disease, Leblanc and colleagues5 had recognized the callosal agenesis based on pneumoencephalography. This was confirmed in later studies with the use of computed tomography (CT) imaging, and then with magnetic resonance imaging (Fig 3). Unfortunately, because few magnetic resonance images have been performed on these patients, we must rely mainly on the data provided by CT imaging to assess the extent of anatomical abnormalities in the brain. Mathieu and colleagues8 performed CT imaging on 64 patients with HMSN/ACC and found no evidence of ACC in 33% of cases, partial ACC in 9%, and complete ACC in 58%. They also found 4.7% of patients with diffuse cerebral atrophy and 17% with cerebellar atrophy. However, CT criteria are not ideal for assessing the degree of callosal agenesis. Although signs of complete or near complete ACC are unlikely to be missed, it is much more difficult to assess partial ACC. In addition, even in cases where the corpus callosum is anatomically well preserved, axonal loss is seen on pathology (see autopsy findings below), which suggests that most, if not all cases of HMSN/ACC have some degree of callosal abnormality. Pathology We reviewed five autopsies of patients who died between 26 and 37 years of age, as well as six sural nerves of patients biopsied between the ages of 2 and 12 years. Two of the cases autopsied have had molecular testing performed and were found to be homozygous for the typical FC mutation, c.2436delG mutation, in exon 18 of SLC12A6. Fig 3. Brain magnetic resonance image of a 23-year-old French Canadian patient with hereditary motor and sensory neuropathy associated with agenesis of the corpus callosum (ACC), with exon 18 c.2436delG showing complete ACC. (A) Sagital T1 section. (B) Axial T1 section. the medial side of the hemisphere corresponded to that usually associated with callosal agenesis. Probst’s bundles were prominent in these brains. In the last two cases, the corpus callosum was present in its entirety, although the posterior portion was thinner than the anterior (Fig 4A). One of these cases had had molecular testing and was found to have the typical FC mutation. The neocortex appeared grossly normal in all patients, as did the hemispheric white matter. Microscopic examination showed scattered small oval vacuoles in the white matter of the brain in all cases. They could be mistaken for Buscaino bodies, except that they never contained basophilic or metachromatic material, but were optically empty or contained an eosinophilic web. The evidence suggested that they had resulted from axonal swelling. The corpus callosum of the two patients in which it was completely present appeared normal in its anterior portion except for occasional vacuoles. In the posterior part, there was progressively more evidence of axonal loss, with increased packing density of oligodendroglial cells and proliferation of astrocytes. Where axonal loss was greatest, vacuoles were absent. Occasional pinkish, slightly enlarged axons could be identified on hematoxylin and eosin stains. In contrast with the subtle microscopic changes in the cerebral white matter, there were striking abnormalities in the spinal nerve roots and cranial nerves. Massively enlarged axons up to 120m in diameter could be identified on paraffin sections (see Fig 4B). On stains for axons, such as Bodian’s (see Fig 4C), these immense axons were a pale gray, but numerous axons with a lesser degree of enlargement showed darker staining that approached that of normal axons. In some sections of roots, there were numerous tiny axons, sometimes with a random or recurrent course that suggested aberrant regeneration. Moderate axonal loss per unit area was noted. No myelin could be stained around the largest axons, whereas moderately enlarged ones had abnormally thin sheaths. Epoxy resin sections confirmed the findings of paraffin sections on the roots and showed numerous onion bulb formations. Teased fibers from motor roots showed the greatest degree of abnormality at approximately 1cm from the spinal cord. Axonal enlargement tended to persist through several internodes, although focal accentuation of enlargement occurred. Balls of Schwann cells adhered to teased fiber bundles. FINDINGS IN SURAL NERVES. All biopsies showed an almost total lack of large myelinated fibers. Small myelinated fibers from the three oldest patients (6, 7, and 12 years old) were somewhat reduced in numbers, whereas in two patients (2 and 3 years old) small myelinated fibers were increased compared with controls (see Fig 4D). A few Wallerian ovoids were BIOPSY In two cases, the corpus callosum was entirely absent. One of these tested positive for the c.2436delG mutation in SLC12A6; the other was not tested. In a third case, only a small rounded portion of corpus callosum was present anteriorly and was in contact with a thin lamina terminalis. The gyral pattern on AUTOPSY FINDINGS. Dupré et al: Motor and Hereditary Polyneuropathy 13 Fig 4. (A) Medial view of one hemisphere of a hereditary motor and sensory neuropathy associated with agenesis of the corpus callosum patient who has a corpus callosum. The anterior portion of the corpus callosum is thicker and whiter than the posterior part. (B) Photomicrograph showing the longitudinal view of a lumbar motor root with three large swollen axons. Nuclei, probably of Schwann cells, border these axons. Hematoxylin and eosin, ⫻560. (C) A swollen axon colored with Bodian’s stain is pale. The arrow indicates the position where it is connected to a less swollen axon. (D) A sural nerve biopsy at the age of 2 years shows numerous small myelinated fibers and an absence of large fibers. Epoxy resin section, paraphenylene diamine, phase optics, ⫻250. found in four nerves. Clusters of small myelinated fibers suggestive of axonal regeneration were seen in all biopsies, and a few thin onion bulbs. Unmyelinated fibers by phase microscopy and oil immersion appeared normal. Rare large fibers with an intact sheath but a crumpled or star-shaped outline of the axon and slightly dark axoplasm were found in five biopsies. These were thought to reflect more proximal damage. In biopsies of two patients (at 2 and 4 years of age), isolated fibers whose axons were the largest in the nerve had disproportionately thin myelin sheaths, suggesting that the axoplasm was swollen and the sheath had un- 14 Annals of Neurology Vol 54 No 1 July 2003 dergone either slippage or partial remyelination. Electron microscopy of one of these unusually large axons with a thin sheath showed decreased packing density of neurofilaments without signs of their degradation. This was interpreted as a sign of increased fluid in the axoplasm. Genetic Studies Haplotype Analysis and Gene Mapping In 1996, Casaubon and colleagues13 performed a genomewide scan and mapped the HMNS/ACC locus to a 5cM region on chromosome 15. The FC founder effect suggested by previous epidemiological studies subsequently was confirmed by haplotype and linkage disequilibrium analyses.13,14 Furthermore, Howard and colleagues1 established that the most common diseaseassociated haplotype covering 4cM on chromosome 15 is present on approximately 64% of disease chromosomes of FC HMSN/ACC patients. In addition, marker D15S1232, which shared the greatest linkage disequilibrium with the HMNS/ACC status, was present in more than 97% of affected chromosomes.1 Howard and colleagues1 reduced the HMNS/ACC critical region to approximately 1,000kb and proceeded to test genes within this interval. The SLC12A6 gene, which codes for a potassium chloride cotransporter (KCC3) protein, was found to have a guanine deletion in exon 18 (c.2436delG, Thr813fsX813) on both homologs in all but one FC HMNS/ACC patient.1 This FC HMNS/ACC patient was found to be a compound heterozygote with one chromosome 15 homolog bearing the c.2436delG mutation, whereas the other homolog has adjacent cytosine and thymine deleted with a guanine insertion (c.1584_1585delCTinsG, Phe529fs532 mutation) in exon 11 of the SLC12A6 gene.1 This patient’s phenotype does not differ significantly from the classic HMNS/ACC phenotype. Furthermore, as expected, all patients homozygous for c.2436delG mutation in SLC12A6 were homozygous for at least a portion of the founder haplotype, whereas the compound heterozygote had one founder haplotype and one nonfounder haplotype at flanking markers. A total of 110 unaffected individuals from the SLSJ region were screened for the exon 18 (c.2436delG) mutation. Four individuals (3.6%) were found to be carriers of the mutation, therefore confirming the carrier rate reported in this region by previous epidemiological studies.2 Although it is clear that a geographical cluster exists in the SLSJ region of the Province of Quebec, very few similar cases have been described outside Quebec (Table 2). Although the genealogical studies suggest an ancestral couple coming from France in the late 17th century, no cases of HMSN/ACC have so far been reported in France. Various centers in France have been contacted, and a few cases have been identified with atypical yet mildly comparable phenotypes, but no mutations in SLC12A6 were detected in these individuals suggesting that these are not authentic HMSN/ACC cases. As for other cases described in the literature worldwide, we found only six15–17 that have clinical characteristics completely compatible with those described in our large sample of FC HMSN/ACC patients.8 These reported non-FC cases have yet to be tested for the HMSN/ACC gene mutation. We have, however, formed collaborations with clinicians following up patients with typical characteristics of HMSN/ ACC (unreported in the literature) and were able to identify mutations in the HMSN/ACC gene in two non-FC families. A brother (4 years old) and his sister (5 years old) from the region of Verona (Italy) born from unrelated, unaffected parents presented with developmental delay, a sensory-motor axonal polyneuropathy, and callosal agenesis. Both have the same mutation in exon 15 of SLC12A6 (c.2023C-T, Arg675X). Two boys of Turkish origin born from unaffected parents who are second-degree cousins presented with developmental delay, areflexia, hypotonia, a sensorymotor polyneuropathy, and complete callosal agenesis. Both have the same nonsense mutation in exon 22 of SLC12A6 (c.3031C-T, Arg1011X). The finding of non-FC cases is significant because it confirms the existence of HSMN/ACC outside of FC and predicts the existence of other non-FC cases. Table 2. Published Cases with a Phenotype Similar to That of French Canadian HMSN/ACC Ethnic Origin No. of Siblings Italian (Veneto) 1 Austrian 2 Tanzanian (Mwanza) 3 Italian (Veneto) Turkish 2 2 NCS/EMG Nerve Biopsy Inheritance (SLC12A6 mutation) Reference Clinical Features Imaging Developmental delay, hypotonia, areflexia, amyotrophy, bilateral ptosis, strabismus Developmental delay, areflexia, amyotrophy Developmental delay, hypertelorism, areflexia, amyotrophy Developmental delay Complete ACC on CT Axonal motor polyneuropathy Fewer myelinated fibers, “demyelination” AR (not tested) 15 Complete ACC on MRI Axonal sensory-motor polyneuropathy AR (not tested) 16 Complete or partial ACC on CT Axonal sensory-motor polyneuropathy Fewer myelinated fibers, thin myelin sheaths, onion bulbs Degeneration and vacuolation of axons AR (not tested) 17 ACC Not done Complete ACC on MRI AR (exon 15, c.2023C-T) AR (exon 22, c.3031C-T) 1 Developmental delay, hypotonia, areflexia Axonal sensory-motor polyneuropathy Sensory-motor polyneuropathy (axonal?) Not done 1 HMSN/ACC ⫽ hereditary motor and sensory neuropathy with a genesis of the corpus callosum; NCS ⫽ nerve conduction studies; EMG ⫽ electroencephalomyogram; CT ⫽ computed tomography; AR ⫽ autosomal recessive; MRI ⫽ magnetic resonance imaging. Dupré et al: Motor and Hereditary Polyneuropathy 15 Gene Expression, Gene Function, and Animal Models The gene involved in HMSN/ACC thus is SLC12A6, which codes for a K⫹-Cl⫺ cotransporter named KCC3. The data on the expression and function of this gene are preliminary, and the final data from animal models are yet to come. SLC12A6 initially was cloned and characterized by both Mount and colleagues18 and Hiki and colleagues.19 By analysis of multiple tissue Northern blots probed with a KCC3specific sequence, Hiki and colleagues19 showed a tissue-specific expression pattern in mice with the highest levels observed in kidney, skeletal muscle, heart, and brain. Race and colleagues20 further characterized SLC12A6 from human placenta tissue. The predicted KCC3 protein structure includes (Fig 5) 12 putative membrane-spanning helices with large NH2 and COOH termini; a large extracellular loop between transmembrane domains seven and eight with five potential sites for N-linked glycosylation; two consensus cAMP-dependent protein kinase phosphorylation sites; and four consensus protein kinase C phosphorylation sites in the COOH terminus. They speculated that KCC3 could be involved in volume regulation, in transepithelial transport of salt and water, or in regulation of K and Cl concentrations in cells and in the interstitial space. Shen and colleagues21 studied the function of KCC3 in mouse fibroblast–derived NIH/ 3T3 7-4 cells by comparing KCC3-transfected cells with mock-transfected cells. They observed that KCC3-transfected NIH/3T3 cells do not appear to function in cell volume regulation as is described for other KCC isoforms, including KCC3 in different cell types. Instead, they showed disturbances in ion homeostasis (Cl⫺ equilibrium), which they suggested might have broader roles in phenomena such as cell proliferation and cell death. Pearson and colleagues22 studied the localization of KCC3 in mouse nervous system using polyclonal antibodies. They observed that KCC3 was abundant in all regions of the central nervous system they examined (cerebral cortex, hippocampus, diencephalon, brainstem, cerebellum, white matter tracts of the internal capsule, corpus callosum, and spinal cord). In contrast, they found relatively low expression of KCC3 in dorsal root ganglia and trace amounts in sciatic and trigeminal nerves. In the brain and spinal cord, KCC3 seemed to colocalize with the oligodendrocyte markers CNPase and myeline basic protein, thereby suggesting expression in white matter tracts. KCC3 was also present in the cortex, hippocampus, and in Purkinje neurons. The pattern of KCC3 expression was also shown to change over time, being low at birth and increasing significantly throughout postnatal development. Finally, Li and colleagues23 recently showed that KCC3 mRNA levels are, in general, low in embryonic rodent brain (mouse and rat) and the level increases slightly at birth. Fig 5. Representation of the KCC3 protein. There are 12 transmembrane domains along with large intracellular amino and carboxy termini. Lines perpendicular to the protein represent the different protein-truncating mutations identified in the DNA of hereditary motor and sensory neuropathy associated with agenesis of the corpus callosum patients. The number in brackets indicates the number of amino acids deleted as a result of each mutation. 16 Annals of Neurology Vol 54 No 1 July 2003 Howard and colleagues1 studied the functional consequence of the predominant FC mutation (c.2436delG) of SLC12A6 in Xenopus laevis oocytes. They observed that the truncated mutant was appropriately glycosylated and expressed at the cellular membrane, but that it was nonfunctional. A KCC3 knockout mouse was generated in which exon 3 of SLC12a6 was removed and substituted with a ␤-galactosidase/ neomycin cassette, therefore deleting the remainder of the protein. KCC3 expression was significantly reduced in heterozygous mice and was absent in homozygous mice. Starting at 2 weeks of age, KCC3-null mice showed an abnormal locomotor function compared with controls and heterozygotes. They had a low posture with limb weakness and disorganized limb movements. Behavioral tests (exploratory locomotion and prepulse inhibition) were conducted to assess central nervous system function: both heterozygous and homozygous mice exhibited a significantly lower level of exploratory behavior; prepulse inhibition was abnormal both in heterozygous and in homozygous mice, but more so in the later. The researchers performed detailed anatomical studies on the brain and spinal cord of three homozygous mice and three control mice: homozygous mice showed no specific abnormalities from morphometric analysis (cortex, corpus callosum, cerebellum) and immunostaining. Interestingly, however, sciatic nerves of homozygous mice showed many very large axons with thin myelin sheaths, reminiscent of those seen in HMSN/ACC patient’s biopsies, as well as extensive Wallerian degeneration. Discussion In light of the HMSN/ACC gene discovery, many questions need to be addressed. What is the role of KCC3 in the development and maintenance of the central nervous system and PNS? Is HMSN/ACC primarily a disease of the axon or the myelin sheath? Can all the features of HMSN/ACC be explained by a single gene defect, or are other genes involved, either within the same chromosomal region or on other chromosomes as modifier genes? Can environmental factors influence the phenotypic expression of HMSN/ACC? Lack of KCC3 in the developing nervous system may increase the susceptibility of damaging the fibers migrating across the midline close to the subarachnoid space to form the corpus callosum. As this structure forms during embryogenesis, absence of KCC3 must have phenotypic effects early during neuronal development. One possible mechanism was proposed by Shen and colleagues,21 who showed that KCC3 may be involved in ion homeostasis (Cl⫺ equilibrium) with a possible role in cell proliferation via ion-sensitive kinases. The occurrence of cases without callosal agenesis leaves open the question of interacting genetic and environmental factors, such as prenatal tobacco or alcohol exposure.24 The findings in the PNS, on the other hand, are progressive and do not suggest any migratory abnormality. The site of maximum damage in the PNS appears to be in the nerve roots, where nerve fibers are bathed in cerebrospinal fluid. This is where the great majority of swollen axons are encountered, along with aberrant regeneration and Schwann cell proliferation. The presence of some onion bulbs in peripheral nerve biopsies25 and in nerve roots raises the question as to whether Schwann cell or axon is the primary site of damage. ␤-Iminodiproprionitrile is a toxin that leads to the formation of neurofilamentous masses that focally distend axons.26 Animals treated continuously with this toxin have shown axonal swellings in nerve roots along with segments of demyelination and accumulation of onion bulbs.26 Axonal swelling has not been described as a result of demyelination, and acutely demyelinated axons tend to show reduced cross sectional area. The pathological evidence in HMSN/ACC points to the axon as the primary structure damaged and suggests that watery swelling of the axon is the earliest morphological abnormality. This would be consistent with abnormal electrolyte gradients, which may occur with dysfunction of an electrolyte channel in the axonal membrane. Acute swelling of axons at nodes of Ranvier has been reported to occur with toxins that inactivate sodium channels.27 Selective large myelinated fiber loss as seen in the nerve biopsies could be a useful diagnostic clue in cases where the corpus callosum is present, although it is seen also in Friedreich’s ataxia and some children considered to have the neuronal type of hereditary motor and sensory neuropathy, especially with onset in early childhood.28 In patients with HMSN/ACC, it almost certainly reflects greater susceptibility of large axons in the roots with their more extensive axolemma. Various secondary changes in axons, such as accumulation of damaged organelles, probably distal to the site of maximal damage, could be observed, but this could be easily distinguished from the accumulation of neurofilaments in giant axonal neuropathy and of various abnormal organelles in neuroaxonal dystrophy. Now that we know the gene defect responsible for HMSN/ACC, it will be important to continue detailed clinical studies as well as pursue detailed molecular studies to better understand the pathophysiology of this disease. In addition, it will now be possible to determine the prevalence of this disorder outside of Quebec. This work was supported by “La Fondation des Jumelles Coudées” and the Canadian Institute of Health Research (#15459, G.A.R.). Dupré et al: Motor and Hereditary Polyneuropathy 17 References 1. Howard H, Mount D, Rochefort D, et al. Mutations in the K-Cl cotransporter KCC3 cause a severe peripheral neuropathy associated with agenesis of the corpus callosum. Nat Genet 2002;32:384 –392. 2. De Braekeleer M, Dallaire A, Mathieu J. Genetic epidemiology of sensorimotor polyneuropathy with or without agenesis of the corpus callosum in northeastern Quebec. Hum Genet 1993;91: 223–227. 3. Andermann E, Andermann F, Bergeron D, et al. Familial agenesis of the corpus callosum with sensorimotor neuronopathy: genetic and epidemiological studies of over 170 patients. Can J Neurol Sci 1979;6:400. 4. Gauvreau D, Guérin M, Hamel M. De Charlevoix Saguenay: mesure et charactéristiques du mouvement migratoire avant 1991. In: Bouchard D, De Braekeleer M, eds. Histoire d’un génôme: population, société et génétique dans l’est du Québec. Sillery: Presses de l’Université du Québec, 1991:76 –106. 5. Leblanc G, Mortezai M, Popez-Pinto C. Agénésie du corps calleux (12 cas). Neurochirurgie 1966;7:789. 6. Andermann E, Andermann F, Joubert M, et al. Familial agenesis of the corpus callosum with anterior horn cell disease. A syndrome of mental retardation, areflexia and paraplegia. Trans Am Neurol Assoc 1972;97:242–244. 7. Andermann E. Sensorimotor neuronopathy with agenesis of the corpus callosum. In: Teopoulos M, ed. Handbook of clinical neurology. Vol 42. Amsterdam: North-Holland, 1981: 100 –103. 8. Mathieu J, Bédard F, Prévost C, Langevin P. Neuropathie Sensitivo-Motrice Héréditaire avec ou sans Agénésie du Corps Calleux : Étude Radiologique et Cinique de 64 cas. Can J Neurol Sci 1990;17:103–108. 9. Taft L. Mental retardation: an overview. Pediatr Ann 1973;2: 10 –24. 10. Filteau M, Pourche E, Bouchard R, et al. Corpus callosum agenesis and psychosis in Andermann syndrome. Arch Neurol 1991;48:1275–1280. 11. Linaker O, Nitter R. Psychopathology in institutionalized mentally retarded adults. Br J Psychiat 1990;156. 12. Dyck P, Chance P, Leno R, Carney J. Hereditary motor and sensory neuropathies. In: Dyck P, Thomas P, Griffin J, eds. Peripheral neuropathy. Philadelphia: Saunders, 1993: 1094 –1136. 13. Casaubon L, Melançon M, Lopes-Cendes I, et al. The gene responsible for a severe form of peripheral neuropathy and agenesis of the corpus callosum maps to chromosome 15q. Am J Hum Genet 1996;58:28 –34. 18 Annals of Neurology Vol 54 No 1 July 2003 14. Howard H, Dubé M-P, Prevost C, et al. Fine mapping the candidate region for peripheral neuropathy with or without agenesis of the corpus callosum in the French Canadian population. Eur J Hum Genet 2002;10:406 – 412. 15. Battistella P, Drigo P, Laverda A, et al. The Andermann syndrome. Progressive neuropathy, mental retardation with agenesis of the corpus callosum. Ital J Pediatrics 1987;13:200 –202. 16. Hauser E, Bittner P, Liegl C, et al. Occurrence of Andermann syndrome out of French Canada—agenesis of the corpus callosum with neuronopathy. Neuropediatrics 1993;24:107–110. 17. Deleu D, Bamanikar S, Muirhead D, Louon A. Familial progressive sensorimotor neuropathy with agenesis of the corpus callosum (Andermann syndrome): a clinical, neuroradiological and histopathological study. Eur Neurol 1997;37:104 –109. 18. Mount D, Mercado A, Song L, et al. Cloning and characterization of KCC3 and KCC4, new members of the cationchloride cotransporter gene family. J Biol Chem 1999;274: 16355–16362. 19. Hiki K, D’Andrea R, Furze J, et al. Cloning, characterization, and chromosomal location of a novel human K-Cl cotransporter. J Biol Chem 1999;274:10661–10667. 20. Race J, Makhlouf F, Logue P, et al. Molecular cloning and functional characterization of KCC3, a new K-Cl cotransporter. Am J Physiol 1999;277:C1210 –C1219. 21. Shen M, Chou C, Hsu K, et al. The KCl cotransporter isoform KCC3 can play an important role in cell growth regulation. Proc Natl Acad Sci USA 2001;98:14714 –14719. 22. Pearson M, Lu J, Mount D, Delpire E. Localization of the K-Cl cotransporter, KCC3, in the central and peripheral nervous system: expression in the choroid plexus, large neurons and white matter tracts. Neuroscience 2001;103:481– 491. 23. Li H, Tornberg J, Kaila K, et al. Patterns of cation-chloride cotransporter expression during embryonic rodent CNS development. Eur J Neurosci 2002;16:2358 –2370. 24. Sowell E, Mattson S, Thompson P, et al. Mapping callosal morphology and cognitive correlates: effects of heavy prenatal alcohol exposure. Neurology 2001;57:235–244. 25. Carpenter S. The pathology of the Andermann syndrome. In: Lassonde M, Jeeves M, eds. Callosal agenesis: a natural split brain? New York: Plenum, 1994:27–30. 26. Griffin J, Price D. beta, beta⬘-iminodiproprionitrile and hexacarbon neuropathies. Evidence for an axonal influence. Lab Invest 1981;45:130 –141. 27. Love S, Cruz-Hofling M. Acute swelling of nodes of Ranvier caused by venoms which slow inactivation of sodium channels. Acta Neuropathol 1986;70:1–9. 28. Ouvrier R, McLeod J, Pollard J. Peripheral neuropathy in childhood. New York: Raven Press, 1990:91–97.