TwitterRSSFacebookDirectors Blog
  Disorders A - Z:   A    B   C    D    E    F    G    H    I    J    K    L    M    N    O    P    Q    R    S    T    U    V    W    X    Y    Z

You Are Here: Home  »  News From NINDS  »  Proceedings  » 

Skip secondary menu

NIH Peripheral Neuropathy Conference

Bethesda Marriott, Bethesda, MD

October 22-24, 2006

Rationale and Goals
Conference Summary


Sponsored by

: National Institute of Neurological Disorders and Stroke (NINDS) and the Juvenile Diabetes Research Foundation (JDRF)


: Steven Scherer (University of Pennsylvania)

Organizing Committee

: Steven Scherer (University of Pennsylvania), Gary Bennett (McGill University), Eva Feldman (University of Michigan), Jack Griffin (Johns Hopkins University), Michael Shy (Wayne State University), and John Porter (NINDS)

Rationale and Goals

The NIH Peripheral Neuropathy Conference was designed to examine the status of research in the peripheral neuropathies with the additional goal of optimizing the development of new therapeutics. The organization of research and funding by the presumed etiology of the neuropathy (inherited, diabetic, inflammatory, or toxic) has tended to fragment the field, diminishing the cross-talk between researchers who study different types of neuropathy. Rather than organizing around the types of peripheral neuropathy, which might reinforce existing fragmentation, this Conference took a novel approach designed to integrate concepts and researchers from different areas to facilitate coordination and collaboration across the various types of peripheral neuropathies, with the ultimate goal of facilitating the development of therapies.

Experts from basic and clinical science were recruited into Topic Groups to examine the phases of therapeutic development through pre-meeting activities and then to present consensus reports at the Conference. The Topic areas were: (a) identifying disease mechanisms to provide therapeutic targets, (b) examining the status of diagnosis and biomarkers, to ensure sufficiently sensitive diagnostics to detect peripheral neuropathy at an early enough stage to make therapeutics effective, to adequately stratify patients for clinical trials, and to provide effective surrogates to evaluate intervention efficacy in clinical trials, (c) examining the status of therapeutic development efforts leading to Investigational New Drug (IND) applications to the FDA, and (d) developing protocols, common data elements, endpoints, and infrastructure to facilitate the conduct and cross-comparison of clinical trials. The Conference outcomes consist of status reports on disease mechanisms in the axonal neuropathies, demyelinating neuropathies, and neuropathic pain, diagnostics and biomarkers, therapeutic development, and clinical trials, including specific recommendations for the way forward in achieving safe and effective therapies for the peripheral neuropathies. The ultimate goal is that this Conference initiates dialog to reduce the fragmentation, and increases collaborations and awareness of complementary strengths for the often less glamorous but absolutely essential activities that are required to bring new therapies to patients suffering with the burden of peripheral neuropathy.

Conference Summary

  1. Prelude
  2. The Structure and Function of Peripheral Nerves
  3. Causes and Classifications of Peripheral Neuropathy
  4. Disease Mechanisms: Axonal Neuropathies
  5. Disease Mechanisms: Demyelinating Neuropathies
  6. Disease Mechanisms: Neuropathic Pain
  7. Diagnosis and Biomarkers
  8. Therapeutic Development
  9. Clinical Trials

A. Prelude

The development of novel therapeutics for the peripheral neuropathies will require a change in mindset and behavior among academic researchers, from hypothesis-driven to goal- or milestone-driven approaches, a change in collaboration paradigms, bringing in the broader range of expertise needed for development, production, and commercialization of new drug or biologic therapeutics, and a change in funding paradigms, since no single advocacy organization, corporation, or governmental agency can support the full gamut of activities required for therapy development in rare diseases. The problem facing the peripheral neuropathy field is that there are few effective therapies and substantial corporate involvement is currently restricted to limited areas (e.g., neuropathic pain). For new therapies to be developed, academic researchers will need to broaden their perspective about what is needed to bring new drugs and biologics to clinical trials. Such efforts will lower the scientific, corporate, and regulatory risks and facilitate corporate interest in therapy development programs for the peripheral neuropathies, many of which are rare diseases not readily embraced by corporate structure. For peripheral neuropathies, there is well-founded optimism that many of the corporate keys to program success either have been or can be met:

  • Disease amenable to palliative or curative therapy;
  • Chronic diseases more tractable than acute;
  • Therapeutic linked to a validated disease mechanism;
  • Patients readily diagnosed at a sufficiently early stage of disease (and existence of markers for patient stratification and surrogate endpoints);
  • Clinical trials feasible in the target population;
  • Ensure that therapeutic traits are optimized by program design.

Furthermore, studying individual kinds of peripheral neuropathies may provide answers that are more generally applicable, as well as additional benefits such as reducing the duplication of effort and increasing the likelihood that specific therapies will be developed.

This Conference took the first step of bringing together academic researchers on October 22-24, 2006 to begin to assess what knowledge and tools are available and what is still needed to develop novel therapeutics in this field-a key goal was to assess the current status of research in this area and begin the change of mindset and behavior that is necessary to therapy development. Representatives of the major funding agencies-patient advocacy organizations and components of the NIH-also participated. The NINDS and JDRF have initiated programs that address the iterative steps needed for an FDA IND application for therapeutics in peripheral neuropathy, thereby addressing the necessary change in funding paradigms to support a type of research not supported by traditional, hypothesis-driven grant programs. The NINDS funding program for therapeutic development is described at: and the JDRF program is described at:

The next steps must include a change in collaboration paradigms, which will require another level of interaction to obtain input from biotechnology companies, large pharmaceutical companies, regulatory agencies, and experts in development of academic-corporate partnerships. Finally, patient resources, expertise in translational research, and funding are scare commodities. From experience in other diseases, it will be essential to value team versus individual approaches (crossing academic, corporate, and national boundaries to collaborate and share knowledge and resources) in order to achieve success in moving any drug or biologic into the clinic.

B. The Structure and Function of Peripheral Nerves

The peripheral nervous system (PNS) is composed of motor, sensory, autonomic, and enteric neurons, as well as the glial cells that ensheathe their axons (Schwann cells) and cell bodies (satellite cells). Motor neurons innervate skeletal muscle fibers; autonomic (sympathetic and parasympathetic) neurons innervate and regulate the function of smooth muscle and secretory cells in a wide number of tissues. Sensory neurons innervate a variety of specialized sensory appendages (e.g., muscle spindles, Golgi tendon organs, Pacinian corpuscles, Ruffini corpuscles, hair follicles, touch domes) or terminate in anatomically unspecified nerve endings; each kind has precise patterns of synaptic connections in the central nervous system. In addition to their anatomical and physiological specifications, different kinds of neurons require different trophic factors for their development and perhaps even their maintenance.

The peripheral nerves themselves are largely comprised of myelinated and unmyelinated axons, typically grouped in fascicles, each of which is surrounded by a cellular barrier, the perineurium. Myelinated axons range from 1 to 10 microns in diameter. Alpha motor axons and a subset of sensory axons (Ia afferents) are the largest; most of the intermediate and smaller myelinated axons are sensory. Unmyelinated axons (C fibers) are smaller yet (typically less than 1 micron in diameter); these are autonomic and sensory axons, including those subserving nociception. Multiple unmyelinated axons and their associated Schwann cells comprise Remak bundles.

In the electron microscope, the most obvious structures in axons are neurofilaments and bundles of microtubules. Neurofilaments regulate the axonal caliber, and are composed of three subunits, termed heavy, medium, and light. Microtubules are composed of tubulins, and form the scaffolds for kinesins and dynactin; the molecular motors for orthograde and retrograde axonal transport, respectively. In spite of their deceptively simple appearance in fixed material, imaging reveals that living axons are highly active, with mitochondria and vesicles in seemingly incessant motion. Because the cell body is the site of most protein synthesis, axonal proteins must traffic great distances. Similarly, signals originating from the nerve terminal or axon itself must travel the entire length of the axon to reach the cell body.

Myelin is a spiral of specialized cell membrane that ensheathes axons except for small gaps - the nodes of Ranvier. The myelin sheath itself can be divided into two domains - compact and non-compact myelin - each of which contains a non-overlapping set of proteins. Compact myelin forms the bulk of the myelin sheath; non-compact myelin is found in paranodes (the lateral borders of the myelin sheath) and in Schmidt-Lanterman incisures (the funnel-shaped interruptions in the compact myelin). Compact myelin is largely comprised of lipids, including specialized lipids and proteins that play essential roles. Non-compact myelin is distinguished by tight junctions, gap junctions, and adherens junctions, between the apposed cell membrane of the myelin sheath.

The function of peripheral nerves is to conduct action potentials. In unmyelinated axons, action potentials conduct continuously, and slowly, about 1 meter/second. In myelinated axons, action potentials jump from node to node; this is called saltatory conduction and is much faster (up to 80 meters/second) than continuous conduction. Myelin sheaths facilitate saltatory conduction by reducing the capacitance of the internode, and by organizing axonal ion channels. In the nodal region, molecular interactions between Schwann cell microvilli and the nodal axolemma cluster voltage-gated Na+ channels, which are the source of depolarizing current required for saltatory conduction.

C. Causes and Classifications of Peripheral Neuropathy

Any disease of peripheral nerves can be called peripheral neuropathy, or simply neuropathy. There are many causes, but all of them injure axons or myelinating Schwann cells. Clinically, this dichotomy is reflected in the common usage of the terms "axonal" or "demyelinating" as adjectives to characterize an individual patient's peripheral neuropathy. This dichotomy has its roots in the cellular and molecular biology of axons; the very specializations that make them unique make them vulnerable to diseases. In addition to the issue of whether they are axonal or demyelinating, neuropathies can be classified according to whether they are inherited or acquired, or part of a syndrome. Examples are shown in Table 1; some of these will be discussed later.

Table 1. Classifying causes of neuropathies.
Neuropathies are classified by whether they are inherited or acquired, part of a syndrome, and by their primary pathological cause (axonal or demyelinating). Examples of each are provided. Axonal neuropathies can be further subdivided according to whether they chiefly large myelinated axons and/or small unmyelinated axons (so-called "small fiber" neuropathies). A few axonal neuropathies differentially affect sensory and motor axons.

inherited demyelinating
Waardenburg IVCMT1
MLD (recessive)CMT4
inherited axonal
GAN (recessive)CMT2B
acquired demyelinating
osteosclerotic myelomaAIDP, CIDP
acquired axonal
immunoglobulin amyloidosis; vasculitisAMAN; many toxins

aMLD: metachromatic leukodystrophy; FAP: familial amyloid polyneuropathy; GAN: giant axonal neuropathy; AIDP/CIDP: acute/chronic inflammatory demyelinating polyneuropathy; AMAN: acute motor axonal neuropathy.

The classification of non-syndromic inherited neuropathies is more elaborate. These are called Charcot-Marie-Tooth disease (CMT) or Hereditary Motor and Sensory Neuropathy (HMSN). Different kinds are recognized clinically, aided by electrophysiological testing of peripheral nerves. For dominantly inherited forms, if the forearm motor nerve conduction velocities (NCVs) are greater or less than 38 m/s, the neuropathy is traditionally considered to be axonal (CMT2/HMSN II) or demyelinating (CMT1/HMSN I), respectively, although "intermediate" forms have been recognized. CMT1 is more common, and nerve biopsies show segmental demyelination and remyelination as well as axonal loss. CMT2 typically has a later onset and is associated with loss of myelinated axons, without much demyelination. Whereas CMT1 and CMT2 are relatively common, dominantly inherited disorders, there are recessively inherited neuropathies; these are rarer, typically more severe, and the demyelinating forms are usually called CMT4. Besides CMT, other inherited neuropathies have been traditionally given different names. Hereditary neuropathy with liability to pressure palsies (HNPP) is a milder neuropathy and often has distinct episodes of focal neuropathies. Congenital hypomyelinating neuropathy and Dejerine-Sottas neuropathy are severe neuropathies (named according to whether they were clinically recognized around birth or during infancy, respectively); affected patients typically have NCVs less than 10 m/s, and dysmyelinated axons characterized by improperly formed myelin sheaths. In addition to these sensory and motor neuropathies, there are neuropathies that chiefly if not exclusively affect sensory or motor axons. Hereditary sensory and autonomic neuropathies are a group of disorders that affect sensory neurons and/or axons, with variably involvement of autonomic neurons/axons. Hereditary motor neuropathies are a group of disorders that chiefly affect motor axons, although some show a variable degree of CNS involvement.

Table 2. Non-syndromic inherited neuropathies (some may be neuronopathies).
Only the disorders in which the mutant gene has been identified are listed. For more discussion, see:,


CMT1 (autosomal dominant demyelinating or X-linked)




Putative mechanism of disease




haplotype insufficiency/underexpression




overexpression; other gain-of-function




haplotype insufficiency; gain-of-function








altered transcription of myelin genes




loss of gap junctions

Dominant intermediate CMT




impaired formation of endosomes




mischarged tyrosine tRNAs

CMT2 (autosomal dominant axonal/neuronal)




impaired orthograde axonal transport




impaired mitochondria fusion




impaired lysosomal formation/trafficking




mischarged glycine tRNAs




altered axonal cytoskeleton




impaired chaperone function




impaired myelination causing axon loss




impaired chaperone function





Congenital hypomyelinating neuropathy/Dejerine-Sottas neuropathy

Dominant mutations in PMP22, MPZ, EGR2, NEFL

Recessive mutations in GDAP1, MTMR2, MTMR13, PRX, KIAA1985

Autosomal recessive axonal neuropathy








impaired mitochondria

CMT4 (autosomal recessive demyelinating)




diminished mitochondrial function




loss of phosphatase activity




loss of phosphatase activity












altered linkage to dystroglycan




altered transcription of myelin genes

Hereditary sensory and autonomic neuropathies (HSAN); dominant or recessive




altered neuronal synthesis of ceramide








altered transcription




loss of nerve growth factor signaling




loss of nerve growth factor signaling

Primary erythermalgia



altered activity of voltage-gated Na+ channel Nav1.7

Hereditary motor neuropathies (HMN); dominant or recessive







impaired chaperone function

impaired chaperone function







mischarged glycine tRNAs

altered ER stress response




altered RNA processing?




impaired retrograde axonal transport




altered RNA processing?

HNA: hereditary neuralgic amyotrophy

Most of the symptoms (reported by patients) and signs (observations of clinicians) of neuropathy owe to a loss of function of the affected axons. Thus, loss of motor axons diminishes strength, loss of large myelinated sensory axons diminishes balance and vibratory sensation, loss of small myelinated and unmyelinated axons diminishes temperature and nociceptive sensation. In addition, spontaneous activity in affected axons may produce positive symptoms such as fasciculations (motor axons), paresthesias (myelinated sensory axons), and, most importantly, pain (unmyelinated axons).

D. Disease Mechanisms: Axonal Neuropathies

In many neuropathies, the clinical features tend to have a distal predilection, both in terms of first appearance and in ultimate severity. This suggests that axonal length is a factor in determining which neural elements are at risk. But distal distribution does not mean that the defect necessarily lies in the axon; it could just as well represent a primary neuron cell body abnormality. For instance, large doses of pyridoxine (vitamin B6) promptly kill large primary sensory neurons, whereas smaller doses cause only subtle shrinkage of these neurons and indolent, distal axonal degeneration. Thus, a modest neuronal abnormality may result in distal axonopathy, but a more severe insult of the same type may cause the neuron itself to degenerate as the primary event.

The selective vulnerability of PNS neurons that leads to neuropathy may be the axons themselves, whose length makes them the longest cells in the body. In addition, the volume of the axon is vastly greater than the cell body, which must synthesis its various molecular components and delivery them via axonal transport. The ability of several useful medications that affect microtubules - taxol, vincristine, and colchicine - to cause an axonal neuropathy may reflect their effects on axonal transport. Several other toxins, such as n-hexane and IDPN, cause a peripheral neuropathy that is characterized by massive accumulations of neurofilaments.

Hereditary axonal neuropathies also highlight the molecular vulnerability of axons. Dominant mutations in NEFL, the gene encoding the light subunit of neurofilament, cause an axonal neuropathy. Recessive mutations in the gene encoding gigaxonin cause giant axonal neuropathy, a syndrome affecting both PNS and CNS neurons. Gigaxonin binds to a microtubule-associated protein (MAP-1B-LC), and stabilizes microtubules, thereby promoting their axonal transport. A dominant mutation in the gene encoding kinesin KIF1Bb, which transports synaptic vesicles to the axon terminals, causes an inherited axonal neuropathy. A dominant mutation in the gene that encodes KIF5A likely causes a length-dependent axonal neuropathy of CNS axons, and thus a different clinical phenotype, hereditary spastic paraparesis. Finally, dominant mutations in p150Glued cause a length-dependent motor neuropathy with an unexplained predilection for the larynx and arms. p150Glued is a component of the dynactin/dynein complex - the motor for retrograde axonal transport.

In addition to disorders that affect the axonal cytoskeleton and axonal transport, other mutations that cause inherited axonal neuropathies underscore the importance of mitochondria. In particular, dominant mutations in MFN2, the gene that encodes mitofusin 2, are a common cause of CMT2. These mutations probably interfere with the ability of mitochondria to fuse, and possibly their ability to move. GDAP1 is also localized to mitochondria, and recessive mutations also cause peripheral neuropathy. Other syndromic mitochondrial diseases also cause neuropathy, and some of the spastic paraplegias are caused by mutations in genes that encode components of mitochondria (SPG7/parapleglin; SPG20/spartin; SGP31/REEP1; Friedrich ataxia/frataxin); neuropathy is also present in most of them. The effects of the antibiotic linezolid on mitochondria may be the reason that it causes an axonal neuropathy. The emerging concept is that axons may be uniquely vulnerable to mitochondrial diseases because every point along their length needs its own source of energy.

Neuropathy as part of the disease produced by immunoglobulin (AL) amyloidosis and the familial amyloidotic polyneuropathy associated with mutants of plasma transthyretin (TTR), gelsolin, and apolipoprotein AI (ApoAI) genes. These are systemic diseases in which soluble plasma proteins are transformed into beta-structured fibrils that are deposited in various organs and apparently cause dysfunction by their presence and magnitude. While each kind of amyloidosis is characterized by the specific protein that contributes the subunit of the fibrils, the pathologic processes leading to amyloid deposition are similar. A normally soluble protein exits the circulation, is partially proteolysed by cell associated mechanisms, and then assembled into proteolytically resistant beta-structured fibrils in extravascular spaces. These fibril deposits progressively enlarge, displacing normal structures including cells, basement membrane, and connective tissue. The degree and distribution of organ function impairment dictates the clinical manifestations. TTR mutations can be detected by gene sequencing, and liver transplantation can stabilize the disease.

Whereas toxins and genetic diseases illuminate how axons can be damaged, relatively little is known about the pathogenesis of the most common axonal neuropathies. Chronic diabetes inevitably causes a length-dependent axonal neuropathy, and rigorous control of the blood glucose is the only known way to prevent or forestall this consequence. More perplexing yet are the large number of patients who have a neuropathy without an apparent cause. "Borderline" diabetes may be the cause sometimes, because an excessive number of these patients have abnormal results on glucose tolerance tests, but that do not reach the accepted criteria for diabetes.

Axonal neuropathies are associated with several autoimmune syndromes. Of these, Sjogren's syndrome stands out because it commonly associated with a small fiber neuropathy; sensory neuronopathies may rarely occur. Paraneoplastic autoantibodies that are associated with cancer. Of these, antibodies against Hu/ANNA1 cause a sensory neuronopathy that is typically associated with small cell lung cancer. In spite of their commercial availability, the role of other autoantibodies against a variety of other antigens in evaluating axonal neuropathies remains to be rigorously shown.

Vasculitic neuropathies are a special case of inflammatory neuropathies. Vasculitis causes ischemic injury, commonly resulting in the clinical picture of mononeuritis multiplex. The predilection of certain forms for medium-sized arterioles is thought to account for the propensity of peripheral nerve involvement; this is seen in polyarteritis nodosa (periarteritis), Churg-Strauss syndrome, rheumatoid vasculitis, Wegener's granulomatosis, cryoglobulinemia, and Sjogren's syndrome.

Axonal Degeneration and Regeneration

Wallerian degeneration refers to a series of dramatic events that occur distal to the site of axotomy. Although usually induced by mechanical injury in experimental studies, Wallerian degeneration is a feature of any insult that causes axonal degeneration, including peripheral neuropathies. During the first week post-axotomy, axons fragment and disappear, and the myelin sheaths are phagocytosed in part by Schwann cells but mainly by macrophages that invade the degenerating nerve. The clearance of myelin debris by macrophages promotes axonal regeneration, perhaps owing to the ability of myelin to inhibit axonal regeneration. Schwann cells undergo extensive proliferation beginning between three and five days post-axotomy. The basal lamina remains intact and continues to surround these expanded columns of "denervated" Schwann cells (traditionally called "bands of Büngner"). Thus, while axons and myelin sheaths degenerate during Wallerian degeneration, the Schwann cells and their basal laminae persist.

The signal that initiates Wallerian degeneration is unknown. One possibility is that axotomy interrupts the supply of a neuronal factor that maintains the axon (and thus the myelinating Schwann cells). This idea is supported by studies demonstrating that the disruption of fast axonal transport by a cold block, which would be expected to deplete such a factor, can initiate Wallerian degeneration. Axonal degeneration is regulated by cellular events, however, as it is markedly delayed in Wlds mice. These mice have a dominant mutation that encodes a novel transcript that likely leads to increased production of NADH, but how this protects axons from degeneration is not clear. Nevertheless, the expression of the Wlds gene protects axons from undergoing Wallerian-like degeneration in several models of experimental neuropathy.

Axonal degeneration also signals the nerve cell bodies - usually called chromatolysis or the axon reaction - initiating dramatic changes in the patterns of gene expression. For sympathetic neurons, one of the most salient features of chromatolysis - the changes in peptide expression - appear to be regulated both by decreased delivery in nerve growth factor (NGF) and by the increased expression of leukemia inhibitor factor (LIF). NGF and LIF are unlikely to serve comparable roles for motor or most sensory neurons, but chromatolysis could be modulated by different trophic factors in a cell type-specific manner. Alternatively, the signal for chromatolysis may be the intra-axonal synthesis of importins, followed by their retrograde transport to the neuronal cell body.

The changes in gene expression probably enable PNS neurons to regenerate their axons. At the site of injury, the proximal portion of a myelinated axon gives rise to one or more sprouts, each of which is tipped by a growth cone. These growth cones typically stay within their "Schwann tube", comprised of the now denervated Schwann cells and the basal laminae of the original axon. The Schwann tube is thought to provide an adhesive substrate as well as trophic support for regenerating axons. Subsequent axon-Schwann cell interactions during nerve regeneration are fundamentally similar to those that occur during development. Initially, Schwann cells surround bundles of regenerating axons. Eventually Schwann cells segregate the larger fibers into a 1:1 relationship, elaborate a new basal lamina and form myelin sheaths. With time, remyelinated axons may enlarge to nearly normal diameters, but the thickness of myelin sheath and length of the myelin internodes do not attain those of normal nerves. Axolemmal specializations at node of Ranvier are also reestablished restoring saltatory conduction of myelinated nerve fibers. As Schwann cells reform myelin sheaths around regenerating axons, the cease expressing the genes that characterize "denervated" Schwann cells, and reexpress myelin-related genes. Axonal regeneration in the PNS is robust, but its effectiveness diminishes over time. Thus, in clinical practice, the expectation is that axons can regenerate about 1 inch per month for about 12 months.

Future Research Directions

Peripheral neuropathy is a consequence of many specific causes, but likely fewer final common pathways. Elucidating the specific causes or the final common pathways will provide new targets for drug development. In addition to the ones discussed in Therapeutics Development, the following areas warrant emphasis for future research studies relevant to the axonal neuropathies:

  1. Can growth factors be used as therapeutic drugs for preventing axonal degeneration? Several medications, including several important chemotherapeutic drugs, inevitably cause neuropathy (which is typically its dose-limiting toxicity), but there is evidence that appropriate trophic factors can largely prevent the development of neuropathy. In spite of the apparent lack of efficacy of both ciliary neurotrophic factor (CNTF) and insulin-like growth factor I (IGF-I) in ALS and NGF in diabetic neuropathy, there is a strong conceptual basis for using growth factors to prevent neuropathy. In particular, neurotrophin-3 has been shown to reduce the neuropathy caused by cis-platinin. This is a top priority.
  2. What are the key molecular events of Wallerian degeneration? Can they be manipulated to diminish the degree of axonal loss in peripheral neuropathies? The data from the Wlds mouse provides a strong conceptual model for this possibility.
  3. What are the key molecular events of the axon reaction? Can they be manipulated to diminish the degree of axonal loss in peripheral neuropathies or improve axonal regeneration?
  4. Understand the role of axonal transport in the pathogenesis of neuropathies.

E. Disease Mechanisms: Demyelinating Neuropathies

The Structure and Function of Myelinated Axons

Myelin is mostly comprised of specialized cell membrane, which is relatively enriched in lipids, including cholesterol and sphingolipids, as well as the specialized glycolipids galactocerebroside and sulfatide. The PNS myelin sheath contains two regions of specialized membranes, compact and non-compact myelin, each of which contains a unique, non-overlapping set of proteins. Compact myelin constitutes the bulk of the myelin sheath and contains myelin protein zero (MPZ; also called P0), peripheral myelin protein 22kDa (PMP22), and myelin basic protein (MBP). The precise amount of these lipids and proteins appears to be crucial for the stability of the myelin sheath, as perturbing the level of any one component can result in demyelination. Compact myelin is nearly devoid of cytoplasm; it is comprised of layer upon layer of closely apposed cell membranes.

Two of these proteins play roles in demyelinating neuropathies. MPZ comprises ~50% of myelin protein. MPZ is a member of the immunoglobulin gene superfamily and a type 1 transmembrane protein; the first 29 amino acids are cleaved during synthesis. MPZ is also post-translationally modified (N-linked glycosylation, as well by the addition of sulfate, acyl and phosphate groups). MPZ is a homophilic adhesion molecule, and crystallographic analysis of the extracellular domain demonstrates that it forms a compact sandwich of beta-sheets held together by a disulfide bridge, similar to that of other members of the Ig-superfamily. In addition, these Ig-domain monomers interact to form a doughnut-like homotetramer; the homotetramers from one membrane form a lattice that interacts the those on the opposing membrane, resulting in stable membrane apposition. The cytoplasmic domain, including its PKC substrate motif (RSTK), is necessary for MPZ-mediated adhesion.

PMP22 is an intrinsic membrane protein, and accounts for about 5% of the PNS myelin proteins. Even though its function is unknown, the exact amount of PMP22 in compact myelin appears to be critical, as an extra copy of the PMP22 gene results in CMT1A. Conversely, a missing copy of the PMP22 gene results in hereditary neuropathy with liability to pressure palsies (HNPP).

Noncompact myelin is found in paranodal region, which is membrane that is adjacent to the node of Ranvier, and in Schmidt-Lanterman incisures, which are funnel-shaped interruptions in the compact myelin. Non-compact myelin contains gap junctions, tight junctions, and adherens junctions, which are found between adjacent layers of the myelin sheath. Connexin32 (Cx32), claudin-19, and E-cadherin and associated proteins form the gap, tight, and adherens junctions, respectively. Each of these proteins belongs to a family of similar proteins, members of which are expressed in other cell types. Mutations of GJB1, the gene that encodes Cx32, cause the X-linked form of Charcot-Marie-Tooth disease (CMT1X), a demyelinating neuropathy, indicating that the radial gap junction pathway (directly across the myelin sheath) plays an essential role in myelinating Schwann cells.

The node of Ranvier is enriched in voltage-gated Na+ (Nav) channels, which provides the depolarizing current for salutatory conduction. Although 10 different Nav channels have been identified, Nav1.6 is the predominant channel in adult nodes. Two K+ channels, KCNQ2 and KCNQ3, are found at nodes; these likely form the slow K+ current of myelinated axons that likely contributes to repolarization. An adaptor molecule, ankyrinG, links these ion channels to the betaIV spectrin cytoskeleton. The nodal axolemma also contains Nr-CAM and neurofascin-186, both of which are cell adhesion molecules that appear to be required for the clustering of ankyrinG and Nav channels, as described below.

A paranode is formed by the lateral edge of the myelin sheath spiraling around the axon adjacent to a node. At paranodes, the loops of Schwann cell membrane closely appose the axolemma, and appear to be joined to it by septate-like junctions, which not only appear similar to invertebrate septate junctions, but also have similar molecular components. The resulting structure has been compared to a bolt (the axon) threaded into a nut (the paranodal loops). Septate-like junctions contain contactin and contactin-associated protein (Caspr), which form heterodimers on the axonal membrane; the glial membrane contains neurofascin-155. Gene knockouts in mice demonstrate that all three proteins are required to form septate-like junctions. Paranodes are thought to sequester the intrinsic membrane components of juxataparanodes away from nodes, and to provide a diffusion barrier between the periaxonal space (between the axon and the myelin sheath) and the extracellular space at the node. Analysis of various mutant mice confirms that the lack of septate-like junctions allows Kv1.1 and Kv1.2 K+ channels to become localized to the paranodal axonal membrane, resulting in the shunting of current during the propagation of action potentials.

Juxtaparanodes are the regions of the internodes that are closest to paranodes. The juxtaparanodal axolemma is enriched in heterotetramers of Kv1.1 and Kv1.2, which form a complex along with Caspr2. The exact role of these potassium channels is not known but they may prevent repetitive activity, as dominant mutations in the gene encoding Kv1.1 as well as autoantibodies against Kv1.1 and/or Kv1.2 result in repetitive activity of motor axons (neuromyokymia or neuromyotonia). Homophilic interactions of the cell adhesion molecule TAG-1 may join the Schwann cell and axonal membranes at juxtaparanodes.

The Development of Myelinated Axons

The development of Schwann cells has mainly been investigated in rodents, as described below; incomplete but comparable findings have been made in humans. The neural crest gives rise to Schwann cell precursors, which migrate on developing peripheral axons and appear to be essential for their normal fasciculation. Precursors develop into immature Schwann cells, which ensheathe large bundles of developing axons, and, in a process termed "radial sorting", subsequently separate axons into smaller and smaller bundles. Axons destined to be myelinated establish a one-to-one association with a Schwann cell. The so-called "promyelinating" Schwann cells cells express the transcription factors Oct-6, then Egr2/Krox20 and Brn-2. These transcription factors, as well as Sox10, are required for the normal development of myelinating Schwann cells. They upregulate the transcription of many myelin-related genes, including those required for the synthesis of lipids as well as structural proteins of the myelin sheath.

The decision to differentiate into a myelinating Schwann cell appears to be driven by the axons. Classical studies indicate that "signal for myelination" is related to the axonal caliber, and recent studies have implicated neuregulin-1 as the signal. Conversely, as myelinating Schwann cells differentiate, they alter the molecular organization of the axonal membrane, particularly at nodes of Ranvier. Schwann cell processes cluster ion channels on the axonal membrane. Schwann cells express gliomedin, which binds to Nr-CAM and neurofascin-186, cell adhesion molecules on the nodal axolemma. Gliomedin clusters Nr-CAM and neurofascin-186, which in turn, bind to ankyrinG, which in turn, clusters Nav channels at nodes. Disrupting gliomedin, Nr-CAM, or neurofascin 186 disrupts or ankyrinG disrupts the nodal clustering to voltage-gated ion channels.

After Schwann cells have established a one-to-one association with an axon, the innermost sheet of membrane spirals around and around the axon, forming the myelin sheath. Concurrently, gliomedin is expressed at the sides of the myelin sheath, clustering Nr-CAM, neurofascin186, ankyrinG, and Nav channels. As myelinating Schwann cells elongate along their axon, each hemi-node comes to appose the hemi-node of the adjacent myelin, and fuses to form a proper node of Ranvier. It is plausible that newly formed nodes initially express other Nav channels, but Nav1.6 is the predominant one at mature nodes, and is essential for motor axon function, as demonstrated by the lethal paralytic phenotype of motor endplate deficient (med) mice. Paranodal specializations develop after nodes are formed, and subsequently sequester Kv1.1 and Kv1.2 channels to the juxtaparanode.

The Genetics of Inherited Demyelinating Neuropathies

As shown in Table 1, each of these disorders shows remarkable genetic heterogeneity - CMT1 (mutations in 5 different genes); Dejerine-Sottas neuropathy and congenital hypomyelinating neuropathy (9); recessive demyelinating neuropathies (9). Further, different mutations in the same gene produce different phenotypes. In addition to CMT, de/dysmyelination neuropathy is a feature of metachromatic and globoid cell leukodystrophy, as well as Cockayne's syndrome, syndromes associated with SOX10 mutations, congenital disorder of glycosylation type Ia, Refsum's disease, congenital cataracts, facial dysmorphism, and neuropathy, and minifascicular neuropathy. In these diseases, the neuropathy is typically overshadowed by other, more severe manifestations.

Although there are many different demyelinating forms of CMT, mutations in just 3 genes - PMP22, MPZ, and GJB1 - account for the majority. That said, different mutations produce diverse phenotypes, indicating that there are multiple mechanisms that produce demyelination.

PMP22. Unequal crossing over between two homologous regions on chromosome 17 that flank the PMP22 gene cause HNPP and CMT1A. The deletion of the 1.4 Mb region causes HNPP; duplication causes CMT1A. Deletion results in haplotype insufficiency, resulting in too little PMP22 in compact myelin; some loss-of-function PMP22 mutations produce the same phenotype, demonstrating that PMP22 is the critical gene in the deleted region. Duplication results in too much PMP22 in compact myelin, which results in demyelination. In addition to PMP22 duplication, some missense mutations (amino acid substitutions) are said to cause CMT1A, but these typically cause a more severe phenotype, ranging from congenital hypomyelinating neuropathy to a CMT1-like phenotype. PMP22 duplications are the most common cause of CMT, and one suspects that HNPP may be even more common, but is under-recognized because it produced a much milder neuropathy.

MPZ. Nearly 100 different mutations in MPZ have been identified that cause CMT1B. Most of these patients fall into two distinct phenotypic groups: one with extremely slow nerve conduction velocities and onset of symptoms during the period of motor development; and a second with essentially normal nerve conduction velocities and the onset of symptoms as adults. Other dominant mutations result in congenital hypomyelinating neuropathy or Dejerine-Sottas neuropathy.

GJB1. More that 300 different mutations in the GJB1 gene cause CMT1X. Most mutations result in a similar phenotype, which leads to the idea that all GJB1 mutations cause loss-of-function. A few GJB1 mutations are also associated with transient disturbances of CNS myelin, indicating that Cx32 also plays a role in the biology of oligodendrocytes.

The Molecular Pathogenesis of Inherited Demyelinating Neuropathies

The great advances in human genetics have led to the identification of many genes that result in this clinical picture. As anticipated from the nerve transplantation studies of Aguayo and Bray 30 years ago, demyelinating is a Schwann cell autonomous phenotype. In nearly every case, myelinating Schwann cells express the gene that causes demyelination. Different kinds of perturbations have been identified:

  • Altered transcription. Dominant mutations in EGR2 and SOX10 cause severe demyelinating/dysmyelinating neuropathies. These transcription factors regulate the expression of MPZ, MBP, and Cx32 as well as proteins involved in the biosynthesis of lipids. Studies in mice demonstrated that Egr2 is required for the development and even the maintenance of myelinating Schwann cells. Dominant mutations in these transcription factors appear to cause gain-of-function, perhaps by profoundly affecting the expression of many myelin-related genes.
  • Altered protein trafficking causes gain-of-function. Dominant missense mutations in both MPZ and PMP22 appear to disrupting the trafficking of the mutant proteins to the myelin sheath. Many of these mutant proteins are retained in the ER, probably because they misfold, and this could result in several adverse effects - sequestration of chaperone proteins in the ER, produce an unfolded protein response, and overwhelming the ability of the cell to degrade them. Altered protein degradation may be the reason that dominant mutations in SIMPLE cause demyelination. Its murine ortholog interacts with Nedd4, an E3 ubiquitin ligase that targets membrane proteins to lysosomes. In some cases, nonsense-mediated decay ameliorates the phenotype of certain mutations, because the premature stop codon enhances the degradation of the mutant mRNA.
  • Altered lipid phosphorylation. Recessive mutations in MTMR2 and MTMR13 cause demyelinating neuropathies, with prominent myelin outfoldings. These two proteins interact to form a phosphatase, and MTMR2 mutants have significantly reduced phosphatase activity for phosphotidylinositol.

The Pathogenesis of Acquired Demyelinating Neuropathies

By comparison to the inherited demyelinating neuropathies, the pathogenesis of acquired demyelinating neuropathies is less well known. Although a few toxins are known to cause demyelination, most acquired demyelinating neuropathies are believed to be autoimmune in nature. These fall into a few groups: some forms of Guillain-Barre syndrome (GBS), chronic inflammatory demyelinating polyneuropathy (CIDP), and demyelinating neuropathies associated with paraproteins, either osteosclerotic myeloma (usually a lamba light chain) or an IgM that may or may not bind to a carbohydrate epitope that is found on myelin-associated glycoprotein (MAG; hence the misnomer "anti-MAG neuropathy") as well as other proteins and even glycolipids.

GBS is the name given to a group of acute, acquired neuropathies. AIDP (acute inflammatory demyelinating neuropathy) and Miller Fisher syndrome are the demyelinating forms. Miller Fisher syndrome is associated with antibodies against the ganglioside GQ1b, which is found in the paranodal region of the population of myelinated axons that are clinically affected (large sensory axons and the motor axons of certain cranial nerves). Autopsies from patients who died acutely with ADIP show evidence of complement-mediated damage to myelinating Schwann cells. Some chronic acquired demyelinating neuropathies are associated with antibodies against gangliosides or MPZ/P0. Some patients with lamba light chain or IgM paraproteins, the paraprotein appears to intercalate into the myelin sheath, but it remains to be shown that this causes demyelination.

In AIDP and CIDP, macrophages appear to mediate demyelination, in some cases, targeted to the affected region by antibodies and/or complement deposition. Macrophages may also mediate axon damage in other forms of GBS, including acute motor axonal neuropathy (AMAN), where they insert themselves into the periaxonal space. Activated macrophages release potentially injurious molecules such as oxygen radicals, nitric oxide metabolites, proteases, eicosanoids, cytokines and chemokines, all of which may contribute to myelin and axonal damage. In animal models of hereditary demyelinating neuropathies, an intact immune system contributes to the severity of the pathology. Macrophages can also play salutary roles in peripheral neuropathies, by phagocytosing myelin, thereby facilitating axonal regeneration.

The Cellular and Molecular Consequences of Demyelination

Demyelination is the final common pathway of a large number of genetic and environmental insults. Segmental and paranodal demyelination are the anatomical hallmarks of demyelinating neuropathies. Both decrease the capacitance of the axonal membrane. In addition, affected axons may show loss of NaV channels, and disorganized or absent paranodal septate-like junctions allow Kv1.1 and Kv1.2 channels to apposed nodal NaV channels. Neurofilaments are less phosphorylated and more tightly packed in demyelinated internodes, and axonal transport is locally affected. Finally, demyelination leads to axonal degeneration, which is the most important consequence of demyelination, as it is the major cause of disability. Children with CMT1A, for example, have slow nerve conduction velocities before the onset of symptoms, and the velocities do not change appreciably as the disease progresses; demyelination per se does not seem to be sufficient to cause the neurological signs and symptoms.

Thus, preventing axonal loss is an attractive approach for treating inherited demyelinating neuropathies, but this approach is hampered by the limited knowledge of the mechanism(s). Possibilities include the disruption of axon-Schwann cell interactions, which could affect signaling (including trophic factor signaling) between these two cell types, increased energy demands to maintain ion gradients in demyelinated segments, altered axonal transport of critical molecules that are required for axonal integrity, calpain-mediated axonal degeneration, or a hostile local milieu owing to inflammatory changes. Moreover, the molecular mechanisms of axonal degeneration are also unknown.

If axons remain intact, then they will be remyelinated by Schwann cells resident in the nerve, although remyelinated internodes are shorter and thinner than normal. As Schwann cells remyelinate axons, they differentiate along the lines described in developing nerves, including the formation of nodes of Ranvier by the fusion of adjacent hemi-nodes. Remyelination restores saltatory conduction. In animal models of inherited demyelinating neuropathies, Nav1.8 is found at an abnormally large proportion of nodes, and similar findings have been made in the sural nerve from one patient with an R69C MPZ mutation. The functional significance of aberrant nodal Nav1.8 expression remains to be shown.

Future Research Directions

In addition to the ones discussed in Therapeutics Development, the following areas warrant emphasis for future research studies relevant to the demyelinating neuropathies:

  1. Develop cellular assays for high-throughput screens for disease-related topics. Projected targets - screening for small molecules that reduce PMP22 expression, or modify the unfolded protein response. CMT1A appears to be caused by a modest overexpression of PMP22. Two different compounds (progesterone antagonists and ascorbic acid) that were selected because of their possible effects on myelination, decrease PMP22 expression and diminish the severity of demyelination in animal models. A high-throughput screen for compounds that decrease PMP22 expression in selected cell models of myelin gene expression is a plausible and highly promising approach that will lead to treatments for CMT1A - the most common form of CMT1. This is a top priority.
  2. Axonal loss in demyelinating neuropathies. Why axonal loss occurs in the demyelinating neuropathies is not yet known, potential explanations include disrupted molecular signaling between the axolemma and myelinating Schwann cells in the different domains (node, paranode, juxtaparanode, internode), ion channel abnormalities, increased energy demands on the neuron, altered axonal transport, or altered trophic effects. This is a fundamentally important problem because much of the disability in demyelinating neuropathies results from axonal loss, and has broader significance because axonal loss in appears to be an important cause of disability in multiple sclerosis - a demyelinating disease of the CNS. This is a top priority.
  3. 3. Investigating the effects of mutant proteins. The identification of mutant genes is but the beginning in the quest for how mutant proteins cause demyelination. It remains to be determined whether there are final common pathways that result in demyelination. The ability to do this work depends on understanding how Schwann develop, interact with axons, form and maintain myelin sheaths, and remyelinate axons. This is a high priority.
  4. Identifying additional genetic causes of demyelinating neuropathies. Discovering the genes responsible for demyelinating neuropathies has provided important new knowledge on the biology of myelinating Schwann cells, and even suggested therapeutic strategies to treat some of these disorders, such as CMT1A. Identifying additional genes that cause demyelination is crucial for progress in understanding the pathogenesis of these diseases.
  5. Determine how demyelination and remyelination affect the molecular organization and function of myelinated axons. Demyelination and remyelination alters the axons in fundamental ways - the organization of ion channels, altered axoglial functions, abnormal signaling are potential targets. Are the signals in remyelinating the same?
  6. Elucidating the synthesis myelin membranes. Because myelin is a complex biological structure, and sensitive to numerous genetic perturbations, it is essential to understand how individual myelin-related proteins are synthesized, transported to the various membrane domains, and degraded. Continued integrated research involving the generation of animal models, cell culture studies, and patient investigations will be necessary to further define these pathways, which in turn will be necessary to develop rational therapies for the demyelinating neuropathies.
  7. Developing animal models that mimic the human diseases. Many mouse models do not accurately reflect the disease severity of the patients they model. Whether this is because of dosage, life span or because of differences in myelin biology in the different organisms is not known. Particularly if animals are used to determine candidate therapies or to study the biology of human neuropathies, this knowledge will be critical.
  8. Investigations into the biological role of inflammatory cells in inherited demyelination. Although demyelination originates with the expression of the mutant gene in myelinating Schwann cells, immune cells modify the severity of disease in animal models. It is unclear whether macrophages and T-cells are part of a non-specific inflammatory response to clear myelin and axonal debris in many demyelinating neuropathies. It is also not well understood what signals trigger the inflammatory response and whether they are linked to demyelination, axonal injury or both.
  9. Investigate the biology of humans affected with of demyelinating neuropathies. Study of human patients has generated much of current understanding of these neuropathies, including their natural history and genotype-phenotype correlations. These data lead to the quantitative measurement of disease progression, which is key to determine the effects of treatments and to determine whether animal models accurately model the neuropathy in humans.
  10. Pathogenesis of acquired demyelinating neuropathies. Antigens of AIDP and CIDP. Develop an international repository of sera from affected patients. What are the effector mechanisms in AIDP? Could anti-complement strategies work? What is the role of T cells, especially in CIDP? This would likely require a repository of sera with high quality clinical data, and would have therapeutic implications.
  11. Genomic approaches to discover disease-modifying genes. Among patients with a comparable degree of hyperglycemia, why do some patients develop a severe diabetic neuropathy? Why do some patients have repeated attacks of GBS? Why does the severity of neuropathy vary even in patients with the same mutation (e.g., PMP22 duplication)? Such an effort will likely need a repository of DNA with high-quality clinical information.

F. Disease Mechanisms: Neuropathic Pain

Neuropathy often causes neuropathic pain - heightened sensitivity to noxious (hyperpathia) and even ordinarily non-noxious (allodynia) stimuli as well as spontaneous pain. Neuropathic pain is particularly associated with certain neuropathies ("small fiber" neuropathies) and kinds of nerve injury (median causalgia, postherpetic neuralgia). In spite of the uncertainties about the mechanisms that cause it, neuropathic pain is a difficult clinical problem, often more vexing than all the other manifestations of a given person's neuropathy. Recommended research objectives are listed for the following questions; all are intended to encourage clinical and basic science studies on different kinds of neuropathic pain, with the aim of identifying different pain mechanisms.

Future Research Directions

  1. What is the extent of the clinical problem and exactly what symptoms do the patients have?
    1. Participate in the research program initiated by a consortium of German pain specialists. Collect data on specifics of symptom presentation with expanded data collection (e.g., plasma, DNA, and skin samples) and emphasis on conditions under-represented in German effort (especially diabetes).
    2. Determine the prevalence of peripheral neuropathy in the USA (this could be part of effort related to other diseases), including the prevalence of neuropathy in patients with cancer or HIV infection. Such a survey would include the incidence of pain, autonomic dysfunction, and other features.
    3. Develop a practical means for assessing different kinds of neuropathic pain abnormalities in clinical settings.
  2. Develop animal models of clinically relevant neuropathic pain.
    1. Validate a clinically relevant animal model of painful diabetic neuropathy.
    2. Validate a clinically relevant model of postherpetic neuralgia.
    3. Investigate pain phenotypes in mouse models of demyelinating neuropathies and axonopathies,
  3. Why do only a minority of patients with peripheral neuropathy have pain as a symptom?
    1. Conduct whole genome haplotype analyses of patients with various pain conditions and symptoms (e.g., patients who have diabetic neuropathy or postherpetic neuralgia, with and without allodynia).
    2. Conduct a genomic analysis as part of a prospective study of patients at risk to develop neuropathic pain from postherpetic neuralgia, chemotherapy, and HIV infection, including their response to treatment.
  4. What are the mechanisms that generate spontaneous discharge in injured somatosensory primary afferent neurons?
    1. Do transcriptional, translational, and post-translational analyses of axonal pathology-evoked molecular changes in DRG neurons that lead to afferent hyper-excitability.
    2. Determine the origin and time course of ectopic discharges in select animal models and correlate this with pain phenotype (spontaneous pain and stimulus-evoked pain).
  5. It takes a village to study C-fibers?
    1. Investigate status of C-fibers and their associated Schwann cells in animal and human models of neuropathic pain.
  6. What are the key changes in the spinal cord dorsal horn for the genesis and maintenance of painful peripheral neuropathy?
    1. Identify the responses in primary afferent and spinal cord dorsal horn neurons that contribute to the genesis and maintenance of neuropathic pain.
    2. Identify changes in descending excitatory and inhibitory pain modulating pathways that contribute to the genesis and maintenance of neuropathic pain.
  7. How do cytokines, chemokines and inflammatory mediators contribute to neuropathic pain in the PNS and spinal cord?
    1. Identify mechanisms whereby pro- and anti-inflammatory cytokines and chemokines interact with PNS and spinal cord neurons to engage and maintain pain symptoms.
    2. Determine the role of growth factors in the genesis of the pain phenotype.
  8. High-throughput screens for primary afferent neuron plasticity.
    1. Develop and validate of in vitro assays of primary afferent neurons for analgesics and pathology modifiers

G. Diagnosis and Biomarkers

Methods to identify abnormalities in large, myelinated motor and sensory nerve fibers are well established. Electrodiagnostic studies provide both quantitative and semi-quantitative measures that can be followed over time, and certain peripheral nerves can be routinely biopsied. These methods form the foundation to our current understanding of peripheral neuropathies, supported by genetic tests for hereditary neuropathies and serologic tests for some of the immune-mediated neuropathies.

By comparison, assessing the function of small myelinated unmyelinated axons is less well developed, even though they subserve important roles such as pain and autonomic function. Diseases involving unmyelinated (C) fibers include distal small fiber neuropathy, various neuropathic pain states, and the autonomic neuropathies. In the past two decades, a number of advances have improved our ability to diagnose involvement of somatic and autonomic C fiber earlier and more quantitatively. These tools have enhanced the ability of clinicians to diagnose such disorders as distal small fiber neuropathy, autonomic neuropathies, and the early neuropathy associated with diabetes. Unfortunately, these tools are not widely available or uniformly utilized, and the clinical diagnosis of small fiber neuropathy needs additional refinement.

The goal is to develop tests and use them efficiently in order to establish the cause of neuropathy in every affected patient. This will enable clinicians to make an accurate diagnosis as early as possible, when the potential for reversibility is greatest. In addition, novel outcome measures and biomarkers are needed to improve the ability to assess promising agents. We propose a number of research goals and objectives that could further improve the field. New tests need to be developed and some currently used tests need to be critically evaluated.

Future Research Directions

  1. Identify the causes of what is currently called idiopathic neuropathy.
  2. Identify additional causes of hereditary neuropathy. Despite of the great advances to date, the cause of inherited demyelinating neuropathy remains unknown in one-quarter of patients, and three-quarters of those with an inherited axonal neuropathy.
  3. There is growing evidence that impaired glucose tolerance alone can cause neuropathy, but there is currently no way to identifying these patients.
    1. Define criteria for neuropathy associated with glucose intolerance, its relationship to diabetic neuropathy, and its natural history.
    2. Define sensitive measures for diagnosis of early diabetic neuropathy.
  4. Certain patterns of neuropathy/neuronopathy, including sensory ganglionopathy, subacute autonomic ganglionopathy, and multifocal motor neuropathy, are more likely to have an autoimmune pathogenesis. Some may develop spontaneously; others often occur in the context of a systemic malignancy (paraneoplastic neuropathy/neuronopathy). The diagnosis is often made indirectly through detection of specific serum autoantibodies, but their prevalence is unknown, and extensive autoantibody testing is not indicated in every patient who has neuropathy.
    1. Define diagnostic criteria and better assess the incidence of paraneoplastic neuropathy.
    2. Define the utility of serological tests in the evaluation of different kinds of autoimmune neuropathy.
    3. Define autoimmune autonomic ganglionopathy and the role of specific autonomic function tests, serum antibodies and constituents in the diagnosis of these neuropathies.
    4. Determine the utility of autoantibodies against sulfatide, markers of celiac disease (e.g., tissue transglutaminase), and various gangliosides.
  5. Define criteria for the diagnosis of the small fiber predominant neuropathies, including distal small fiber neuropathy, distal autonomic neuropathies and generalized autonomic neuropathies.
  6. Identify robust tests for small fiber function including skin biopsy and QSART, and establish national normative databases for these small fiber function tests. Longitudinal data defining the rate of functional decline are needed and clinically meaningful changes for these measures need to be defined.
  7. Define diagnostic criteria for specific neuropathic phenotypes including sensory neuronopathy, radiculoplexus neuropathy, multifocal motor neuropathy, and proximal sensory neuropathy/radiculopathy.
  8. Identify potential new biomarkers of nerve disease and improve existing measures for use in trials of interventions for peripheral nerve disease (e.g. diabetic neuropathy).
  9. Validate neuropathic scales for the evaluation, and monitoring of neuropathic outcomes.
  10. Develop evidence-based algorithms for evaluating patients with neuropathy to facilitate diagnosis and minimize unnecessary testing.

H. Therapeutic Development

In addition to the specific recommendations that are found in other sections of this document, the research objectives for developing new therapies should include the following concepts:

  • Treatments should ideally address multiple kinds of neuropathy. The causes of neuropathy are diverse, and current treatments depend on identifying the underlying cause and specifically addressing it (e.g., neuropathy caused by B12 deficiency). Finding effective therapies for axonal neuropathies in general would enable one to treat an enormous number of patients, even if the cause was not known. It is recognized that these opportunities for a broader impact may be derived from treatments developed for individual disease states.
  • Identify gaps that prevent efficient Therapeutics Development.
  • Establish a process for regular evaluation of therapeutics development, with the aim of regularly revising the priorities.
  • Building a "Bench to Bedside" process for Therapeutics Development. If a uniform process can be applied to translate research from bench to bedside, as an aid to developing neuropathy therapies, then the steps include:
    1. Identifying and measuring physiological mechanisms relevant to disease initiation and progression.
    2. Validating the molecular targets that regulate this physiopathology and establishing efficient tests to measure a drug-target interaction.
    3. Identifying possible candidate drugs and establishing a proof-of-principle of their disease altering capability in animal disease models.
    The peripheral neuropathies represent a challenge because the disease populations and causative agents are fragmented. Putative physiological mechanisms (and targets) identified at the bench need to be matched with disease states and interventions (for example; de novo drug design, drug repositioning).

Future Research Directions

  1. Therapies that target sensory and/or motor neurons in a mechanistic manner. Some neuropathies are directly due to damage to the cell body, including some chemotherapeutic drugs and autoimmune sensory neuronopathies, motivating treatment strategies directed at this feature. A much more common problem, the role of the cell body in length-dependent neuropathies, is a major gap in our knowledge. Understanding this problem will provide a platform for developing strategies to treat axonal neuropathies.
  2. Distal axonopathies should be more amenable to therapy. It is important to bear in mind that most neuropathies are distal axonopathies. In terms of function, losing the last 5 microns of an axon is equivalent to losing the entire axon or even the neuron itself. In terms of therapy, however, these are important distinctions because a more distal lesion is more amenable to therapy. Drug development efforts using cell-based models need to keep that in mind.
  3. Develop high-throughput screens (HTS) for axonal protection or regeneration. There is an urgent need to develop HTS for these purposes. Most drug screens are done with non-neuronal cell lines or neuronal cells lines that may not accurately represent peripheral neurons. Additional refinements may be required, such as developing neuronal cell lines with long axons, cell lines for different kinds of neurons (motor vs sensory) and different axonal calibers. In addition to the available automated imaging stations for quantifying axonal length, industry has developed tools for measuring degenerating axons. Another useful tool would be means of measuring axonal "health".
  4. Translate findings in animal models. Cell culture-based models provide tractable systems for exploring specific causes of neuropathy. Animal models incorporate the complexities of the PNS. In both models, the key features of the human disease should be present and guide the interrogation of the disease target. There are genetically authentic animal models of many kinds of inherited neuropathies, as well as animal models toxic, diabetic, and inflammatory neuropathies. The strengths and weaknesses of these animal models, including their suitable for drug screening, need to be clarified. These uncertainties raise the questions about whether potential drug candidates should be tested in multiple models of the same disease (to overcome weakness of current animal models) as well as in multiple disease models (to find drugs that may be useful in multiple diseases) before moving them into clinical trials.

    Nevertheless, given the shared anatomical, physiological, and molecular features of mammalian peripheral nerves, one anticipates that many animal models of neuropathy will share a common pathophysiology with their human counterparts, so that treatments that work in mice will also work in humans.

  5. Develop mechanism-based therapies for immune-mediated neuropathies. In these neuropathies, the specific molecular target of the inappropriate immune response at least partially determines clinical manifestations. Immunosuppression using acutely active (e.g., IVIg, aphoresis, corticosteroids) and chronic agents (e.g., azathioprine, cyclophosphamide, mycophenolate mofetil, rituximab) are recognized therapies, but their effectiveness has not been related to the specific molecular target of immune response. Thus, refining immunosuppressive regimes and broadening treatment options beyond immunosuppression are important goals.
    • Identifying specific molecular targets to autoimmune attack may permit more focused therapies;
    • Preventing secondary axonal injury in demyelinating neuropathies;
    • Protecting myelinating Schwann cells, including the specializations at nodes of Ranvier.
  6. Develop therapies for inherited demyelinating neuropathies. Different mutations in many different genes produce a final common outcome, demyelination. Because the Schwann cells themselves express these genes, remyelination can also be affected. Demyelination and incomplete remyelination leave the axon vulnerable to degenerative changes and possibly inflammatory injury. The demyelinating phenotype can be reproduced in genetically authentic animal models, enabling investigations of altered transcription (e.g., ascorbic acid and progesterone in CMT1A), gene therapy by viral vectors, correction of protein misfolding, and neurotrophic supplementation for axons/neurons as well as Schwann cells

I. Clinical Trials

The process of drug development and discovery is complex. Selecting the compounds that are most likely to make it into the clinic is a fundamental problem, for the pharmaceutical industry and for physicians. Early in the pipeline, there are thousands of compounds; the number gets winnowed down for several reasons - efficacy, safety, tolerability, and financial. After identifying targets and treatments from preclinical data, drugs are subjected to clinical trials.

  • Phase I studies the pharmacology and toxicity in humans;
  • Phase II tests proof of concept in humans;
  • Phase III tests efficacy in humans (sometimes called proof of principle trials), and are submitted to regulatory agencies;
  • Phase IV trials designed and intended to inform clinical practice in the community.
 symptomatic treatmentsdisease-modifying treatmentsregenerative treatmentspreventative treatments
Phase I pharmacology & toxicology    
Phase II proof of concept    
Phase III proof of principle    
Phase IV effectiveness in clinical practice    

The clinical trials working group considered research objectives for four different types of interventions for patients with peripheral neuropathy:

  1. Symptomatic. Neuropathic pain is an example.
  2. Disease-modifying. Treating inflammatory neuropathies such as Guillain-Barre syndrome with immune-modulating medications is an example.
  3. Preventative. Preventing or reducing neuropathy caused by chemotherapy is an example.
  4. Regenerative. Enhancing the restoration of function (e.g., by promoting axonal regeneration) is an example.
  5. At least 12 different types of trials can therefore be identified for peripheral nerve disease. In the following list of research objectives, however, no attempt has been made to identify research objectives for each of these different situations.

Future Research Directions

  1. To conduct critical reviews to identify targets and treatments:
    • Preclinical data
    • Phase I pharmacology and toxicity studies
    • Human experimental models
    • In this process, intellectual property and regulatory issues must be considered, and input must be obtained from industry, regulatory agencies, and patient advocacy groups.
  2. To maintain a continuously updated set of systematic reviews of randomized clinical trials of interventions for peripheral nerve disease and set priorities for future systematic reviews.
    • To work closely with existing groups currently providing such reviews, such as the Cochrane Group, American Academy of Neurology (AAN), European Federation of Neurological Societies (EFNS), and Peripheral Nerve Society (PNS).
    • To ensure that reports of negative trials are disseminated and provide adequate information about methods and results to inform the design of future studies, especially patient characteristics, side effects, and placebo responses.
  3. To develop better approaches for selecting the most appropriate patients for symptomatic, disease-modifying, regenerative, and preventive clinical trials.
  4. To systematically collect samples, especially DNA, that can be used for future research, in particular, identifying treatment responders and toxicities.
  5. To identify the role of "surrogate" measures in establishing efficacy and in refining recruitment strategies (e.g., nerve conduction velocity; epidermal fiber density; profiles of protein and gene expression; imaging of nerves and denervated muscles; quantitative sensory testing; testing of autonomic function).
  6. To conduct prospective studies to define better the natural history in concert with pathophysiologic mechanisms of different kinds of peripheral neuropathy - including but not limited to diabetic, impaired glucose tolerance (IGT)-associated, HIV, chemotherapy, idiopathic, Guillain-Barre syndrome (GBS), CMT, chronic inflammatory demyelinating neuropathy (CIDP), post-surgical neuralgias, and postherpetic neuralgia.
    • Especially in conditions where the condition or clinical practice is changing (e.g., diabetes, HIV infection, chemotherapy).
    • When possible, existing data from placebo groups of industry-sponsored and other trials should be exploited.
  7. To identify therapeutic targets and develop optimal research designs for proof of concept studies evaluating symptomatic, disease-modifying, regenerative, and preventive interventions.
    • To evaluate the role of single-dose studies of symptomatic treatments for chronic painful neuropathies.
    • To evaluate treatments and trial designs for the prevention of chemotherapy neuropathy.
    • To determine whether peri-operative analgesia prevents either acute post-operative pain or the development of chronic post-surgical neuralgia or both.
  8. To conduct Phase III studies of the efficacy, tolerability, and safety of treatments for peripheral neuropathies.
    • To evaluate the efficacy of existing symptomatic treatments in conditions for which they have not been adequately studied, for example, gabapentin for treating pain in neuropathy caused by chemotherapy or HIV.
    • To evaluate add-on therapy for treatment refractory GBS patients.
    • To evaluate new agents for CMT and related neuropathies.
  9. To conduct studies that improve treatment effectiveness in clinical practice for existing symptomatic and disease-modifying treatments.
    • To perform dose ranging studies of IVIg for GBS and other acute peripheral neuropathies and for CIDP and related disorders.
    • To conduct head-to-head comparisons of existing symptomatic and disease-modifying treatments. For example, to compare the efficacy, tolerability, and safety of existing treatments for painful diabetic neuropathy-duloxetine, gabapentin, oxycodone, pregabalin, and tramadol.
    • To conduct trials in which the efficacy, safety, and tolerability of existing symptomatic and disease-modifying treatments administered in various combinations are compared with these treatments administered alone.
    • To perform studies of the effectiveness and safety of long-term treatment of chronic peripheral and slowly progressive neuropathies, for example, long-term immunosuppression in CIDP.
  10. To develop new outcome measures or improve existing outcome measures for use in trials of interventions for peripheral nerve disease.
    • To prepare and maintain systematic reviews of outcome measures.
    • To clarify clinimetric issues related to outcome measures for peripheral neuropathy trials (e.g., reliability, validity, responsiveness to change, standardization, normal values, units of measurement).
    • To identify the strengths and limitations of the use of reference populations and the reporting of outcome measures in terms of standard deviation units.
    • To further clarify the role of quality of life outcomes in trials of peripheral neuropathy.
    • To evaluate the strengths and limitations of combining endpoints into composite measures.
    • To develop outcome measures specific for assessment of neuropathy in infancy and childhood.
    • To work closely with regulatory agencies and industry in determining requirements for outcome measures used in obtaining approval.
  11. To develop new designs and strategies for peripheral neuropathy clinical trials.
    • To investigate alternatives to the parallel group superiority trial and their relative advantages and disadvantages for both initial proof of concept studies and pivotal (Phase III) trials; for example, Bayesian methods and adaptive allocation, futility designs, crossover trials, enrichment strategies, and randomized withdrawal designs.
    • To develop novel clinical trial methods for interventions for peripheral nerve disease that target shared underlying mechanisms and pathways rather than specific disease entities.
    • To develop optimal approaches to recording, reporting, and analyzing adverse events in peripheral neuropathy clinical trials.
    • To identify strategies to minimize missing data in clinical trials and statistical approaches for missing data in efficacy and safety analyses that can be used in peripheral neuropathy trials.
    • To identify strategies for the analysis of multiple primary and/or secondary endpoints that can be used in peripheral neuropathy trials.
    • To investigate placebo group response in trials of interventions for peripheral nerve disease, including whether efforts should be made to minimize placebo response if methods for doing so can be identified.
  12. To develop an international consortium of clinical researchers and basic scientists with the following objectives:
    • To identify priorities for randomized clinical trials in each of the major peripheral nerve diseases, taking into account existing preclinical and clinical knowledge.
    • To establish guidelines for designing and performing an ongoing sequence of multicenter trials that will accelerate the development of new therapies and address the major unmet needs in peripheral nerve diseases.
    • To recruit experienced centers with established patient samples and provide training for new centers with the capability to both participate in these multicenter trials and perform observational studies defining the natural history of disease with validated outcome measures.
    • To centralize statistical support and other core facilities, resources, and services for trials in peripheral neuropathy.
    • To collaborate with regulatory agencies, industry, and patient advocacy groups.


Sunday, October 22, 2006

6:00 p.m.
Reception and Dinner
(Sponsored by: The Juvenile Diabetes Research Foundation)
8:00 p.m.
Welcome and Introductions
John Porter (NINDS), Story Landis (NINDS), and Steve Scherer
8:15 p.m.
Case Studies in Therapeutic Development for Neuropathy
Mark Scherer: Introduction
Peter Dyck: Therapeutic Accomplishments in Peripheral Nerve Disease: What Next?
Stefano Previtali: Target Validation in Human Peripheral Neuropathies Baldomero Olivera: From Cone Snail Venoms to Therapeutics for Pain

Monday, October 23, 2006

8:00 - 9:15 a.m.
Disease Mechanisms: Axonal Neuropathies
Jack Griffin: Introduction
Zhigang He: Acute axonal degeneration: energy metabolism and lessons from the Wlds mouse
Peter Stys: Acute axonal degeneration: cation-based injury
Jack Griffin: Chronic axonal degenerations: growth actors and lessons from molecular genetics
Jack Griffin: Conclusions
9:15 - 10:30 a.m.
Disease Mechanisms: Demyelinating Neuropathies
Steve Scherer: PNS myelin sheaths and demyelinating neuropathies
Elior Peles: Molecular basis of myelination and demyelination
Jim Salzer: Signal transduction pathways in myelin
10:30 - 10:45 a.m.
10:45 a.m. - 12:00
Disease Mechanisms: Neuropathic Pain
Claudia Sommers: Clinical data on painful peripheral neuropathies
Gary Bennett: Animal models of painful peripheral neuropathies
Clifford Woolf: Genetic basis of neuropathic pain- animal and human data
12:00 - 12:30 p.m.
12:30 - 1:45 p.m.
Diagnosis and Biomarkers
Michael Polydefkis: Pathology
Philip Low: Electrophysiology
Steve Vernino: Serology
1:45 - 3:00 p.m.
Therapeutic Development
Douglas Zochodne: Axonal Neuropathies: The Role of the Cell Body--Gaps, Opportunities and Common Targets in the Arena of Therapeutics
Ahmet Hoke: Axonal Neuropathies: The Role of the Axon: Regenerative or Reparative? Gaps, Opportunities and Common Targets in the Arena of Therapeutics
Rob Singleton and Kurt Fischbach: Acquired and Hereditary Demyelinating Neuropathies: Gaps, Opportunities and Common Targets in the Arena of Therapeutics
Mark Scheidler: Building a "Bench to Bedside" process for Therapeutics Development
3:00 - 3:15 p.m.
3:15 - 4:30 p.m.
Clinical Trials
Richard Hughes: Research synthesis as the basis of randomized trials: the Cochrane Collaboration contribution
Joseph Arezzo: Criteria for the selection of appropriate end points in clinical studies of peripheral neuropathy
Steven Goodman: Novel designs for clinical trials
David Cornblath: Roadway to success (R2S)

Tuesday, October 24, 2006

8:30 - 8:45 a.m.
Introduction to the Research Objective Development Process
John Porter and Steve Scherer
8:45 - 10:00 a.m.
Topic Group Breakout Sessions
10:00 a.m. -1:00 p.m.
Research Objective Presentations
Clinical Trials
Therapeutic Development
Diagnosis and Biomarkers
Disease Mechanisms: Neuropathic Pain
Disease Mechanisms: Demyelinating Neuropathies
Disease Mechanisms: Axonal Neuropathies
1:00 p.m.
Next Steps/Meeting adjourns


Joseph Arezzo, Ph.D.
Departments of Neuroscience and Neurology
Albert Einstein College of Medicine

Gary Bennett, Ph.D.
Senior Research Chair
Department of Anesthesiology
McGill University

William Benzing, Ph.D.
Deputy Chief
Brain Disorders Clinical Neuroscience IRG
National Institutes of Health

Phillip Chance, M.D.
Professor and Division Chief
Departments of Genetics and Developmental Medicine
University of Washington School of Medicine

Robin Conwit, M.D.
Program Director
National Institute of Neurological Disorders and Stroke
National Institutes of Health
Tel: 301-496-9135

David Cornblath, M.D.
Department of Neurology
Johns Hopkins University

Ted Cummins, Ph.D.
Assistant Professor
Departments of Pharmacology and Toxicology
Indiana University School of Medicine

Marinos Dalakas, M.D.
Chief, Neuromuscular Diseases Section
National Institute of Neurological Disorders and Stroke

Patricia Dreibelbis, M.A.
Director of Program Services
Charcot-Marie-Tooth Association

Robert Dworkin, Ph.D.
University of Rochester

Peter Dyck, M.D.
Department of Neurology
Mayo Clinic

Marian Emr
Director, Office of Communications and Public Liaison
National Institute of Neurological Disorders and Stroke
National Institutes of Health
Tel: 301-496-5924

Eva Feldman, M.D., Ph.D.
Russell N. DeJong Professor
Department of Neurology
University of Michigan

Kenneth Fischbeck, M.D.
Chief, Neurogenetics Branch
National Institute of Neurological Disorders and Stroke

Susan Fleetwood-Walker, B.Sc., Ph.D.
Centre for Neuroscience Research
University of Edinburgh

Roy Freeman, M.D.
Department of Neurology
Harvard University School of Medicine

Jonathan Glass, M.D.
Department of Neurology
Emory University School of Medicine

Robert Goldstein
Juvenile Diabetes Research Foundation

Steven Goodman, M.D., M.H.S., Ph.D.
Associate Professor
Departments of Oncology and Biostatistics
Johns Hopkins Schools of Medicine and Public Health

Jack Griffin, M.D.
Department of Neurology
Johns Hopkins University School of Medicine

Laurie Gutmann, M.D.
Program Director
National Institute of Neurological Disorders and Stroke
National Institutes of Health

Katrina Gwinn-Hardy, M.D.
Program Director
National Institute of Neurological Disorders and Stroke
National Institutes of Health
Tel: 301-496-5745

Zhigang He, Ph.D.
Associate Professor
Departments of Neuroscience and Neurology
Children's Hospital Boston

Amet Hoke, M.D., Ph.D.
Associate Professor
Department of Neurology
Johns Hopkins University

Richard Hughes, M.D.
Department of Clinical Neuroscience
Kings College London School of Medicine

Teresa Jones, Ph.D.
Program Director
National Institute of Diabetes and Digestive, and Kidney Diseases
National Institutes of Health

Naomi Kleitman, Ph.D.
Program Director
National Institute of Neurological Disorders and Stroke
National Institutes of Health
Tel: 301-496-1447

Story Landis, Ph.D.
National Institute of Neurological Disorders and Stroke
National Institutes of Health
Tel: 301-496-9746

Jon Levine, M.D., Ph.D.
Department of Medicine
University of California San Francisco

Phillip Low, M.D.
Department of Neurology
Mayo Clinic

Justin McArthur, M.B.B.S., M.P.H.
Johns Hopkins University

William "Billy" Moore, M.D.
Director, Research Development
Muscular Dystrophy Association

Robert Myers, Ph.D.
University of California San Diego

Klaus-Armin Nave, Ph.D.
Department of Neurogenetics
Max-Planck Institute of Experimental Medicine

Helen Nickerson, Ph.D.
Scientific Program Manager
Juvenile Diabetes Research Foundation

Michael Nunn, Ph.D.
Program Director
National Institute of Neurological Disorders and Stroke
National Institutes of Health

Baldomero Olivera, Ph.D.
Distinguished Professor
Biology Department
University of Utah

Davide Pareyson, M.D.
Division of Biochemistry and Genetics
Carlo Besta National Neurological Institute

Audrey Penn, M.D.
Deputy Directory
National Institute of Neurological Disorders and Stroke
National Institutes of Health
Tel: 301-496-3167

Michael Polydefkis, M.D.
Assistant Professor
Department of Neurology
Johns Hopkins University

John Porter, Ph.D.
Program Director
National Institute of Neurological Disorders and Stroke
National Institutes of Health
Tel: 301-496-5739

Linda Porter, Ph.D.
Program Director
National Institute of Neurological Disorders and Stroke
National Institutes of Health
Tel: 301-496-9964

Stefano Previtali, M.D., Ph.D.
Senior Researcher
Department of Neurology
San Raffaele Scientific Institute

Angelo Quattrini, M.D.
Head, Neuropatholgy Institute
Department of Neurology
San Raffaele Scientific Institute

Lynn Rundhaugen, M.P.H.
Program Analyst
National Institute of Neurological Disorders and Stroke
National Institutes of Health

Domenic Ruscio
Washington Representative
Neuropathy Research Foundation

James Salzer, M.D., Ph.D.
Division of Molecular Neurobiology, Neurology, and Cell Biology
New York University School of Medicine

Mark Scheideler, Ph.D.
Program Director
National Institute of Neurological Disorders and Stroke
National Institutes of Health
Tel: 301-496-8909

Steve Scherer, M.D., Ph.D.
Department of Neurology
University of Pennsylvania School of Medicine

Michael Shy, M.D.
Department of Neurology
Wayne State University

J. Robinson Singleton, M.D.
Associate Professor
Department of Neurology
University of Utah

Gordon Smith, M.D.
Associate Professor
Department of Neurology
University of Utah

Claudia Sommer, Ph.D.
Department of Neurology
University of Wurzburg

Peter Stys, M.D., F.R.C.P.(C)
Professor and Senior Scientist
Departments of Neurosciences, Medicine, and Neurology
Ottawa Health Research Institute

Vincent Timmerman, Ph.D.
Molecular Genetics Department
University of Antwerp

Ursula Utz, Ph.D.
Program Director
National Institute of Neurological Disorders and Stroke
National Institutes of Health
Tel: 301-496-1431

Steven Vernino, M.D., Ph.D.
Associate Professor
University of Texas Southwestern Medical Center

Christina Vert, M.S.
Program Analyst
National Institute of Neurological Disorders and Stroke
National Institutes of Health
Tel: 301-496-1917

Clifford Woolf, Ph.D.
Departments of Anesthesia and Critical Care
Harvard University

Douglas Zochodne, M.D., F.R.C.P.C.
Department of Clinical Neurosciences
University of Calgary

Last Modified March 7, 2011