PATHOPHYSIOLOGY OF MULTIPLE SCLEROSIS: DEMYELINATION AND AXONAL INJURY

Published on 10/04/2015 by admin

Last modified 22/04/2025

Print this page

This article have been viewed 3159 times

CHAPTER 75 PATHOPHYSIOLOGY OF MULTIPLE SCLEROSIS: DEMYELINATION AND AXONAL INJURY

The inflammatory demyelinating lesions of multiple sclerosis cause a range of axonal conduction abnormalities that, in turn, produce the major clinical features of the illness. The clinical effects of these lesions were previously thought to arise directly from the consequences of the demyelination alone, but more recent findings have substantially broadened this understanding. It is now recognized that (1) conduction abnormalities reflect a reorganization of axonal membrane currents after demyelination, (2) inflammation may have direct effects on axonal function, and (3) neuroaxonal degeneration plays a significant role in disease progression.

DEMYELINATION

Action potentials generated by inward sodium currents at the nodes of Ranvier propagate successfully along normal myelinated axons because of the insulating layer of myelin covering the internode. The thickness of the myelin reduces the leakage of current between the nodes by increasing the resistance of the membrane and, of more importance, by reducing its capacitance. These properties have the effect of reducing the leakage of current between nodes and thus reducing the current that needs to be generated at the node for secure conduction to occur. In normal myelinated axons, the action potential at an individual node generates three to seven times more current than required to discharge the next node along the axon; this is referred to as a safety factor of three to seven.¹

It is now clear that the membrane of the adult myelinated axon has a complicated molecular structure (Fig. 75-1).^2,³ Apart from sodium channels, which are clustered at the nodes and expressed at a much lower density in the internodal membrane, there are also a number of potassium channels, including so-called slow channels, located mainly at the nodes, and faster channels, located in the juxtaparanodal region. These serve to limit the excitability of the fiber.⁴ There are a number of other potassium channels, of which some are inhibited by adenosine triphosphate (ATP) and others are activated by intracellular calcium ions. The axonal membrane is also known to contain calcium channels, but these do not appear to contribute significantly to action potential generation or propagation, at least under normal conditions. The ionic balance across the axonal membrane is maintained by a number of ATP-dependent pumps, including the sodium and calcium pumps, and calcium homeostasis also depends on the sodium/calcium exchanger.

Figure 75-1 Immunohistochemical labeling of sodium channels (red) and potassium channels (Kv1.2; green) at a node of Ranvier located on a rat sciatic nerve fiber. The sodium channels are restricted to the nodal gap (double arrowheads), whereas the potassium channels are segregated beneath the juxtaparanodal myelin (V-shaped arrowheads) and apposed to Schmidt-Lanterman incisures (triangular arrowhead).

(Reproduced from Arroyo EJ, Scherer SS: On the molecular architecture of myelinated fibers. Histochemistry 2000; 113:1-18.)

Demyelination reduces the safety factor for conduction, and conduction fails, at least initially, if an internode or more of myelin is removed (Fig. 75-2). The failure occurs partly because the local excitatory sodium current generated by the last node of Ranvier before the demyelinated segment is now dispersed across a much wider area of naked axonal membrane than just the nodal membrane. Even partial-thickness demyelination increases the effective capacitance of the axolemma as the thickness of the myelin is reduced, increasing the amount of local current necessary to depolarize the nodal membrane to its firing threshold and perhaps resulting in conduction block. Conduction is also likely to be blocked, or severely impaired, if the myelin loss occurs preferentially at the paranodes, which causes nodal widening and increases the capacitance of the “nodal” axolemma. Although conduction may later be restored (see later discussion), the density of sodium channels in the newly exposed axolemma is initially too low to support conduction along the demyelinated axolemma.^5,⁶

Figure 75-2 Records showing the slowing changes in conduction occurring over a 5-month period as a result of the induction and spontaneous repair of an experimental central demyelinating lesion in the dorsal columns. Records obtained when the conduction pathway excluded the lesion (left) were quite stable throughout, as were those obtained when the site of the lesion was included but before the lesion was induced (right, top three records). The intraspinal injection of lysolecithin resulted in axonal demyelination and conduction block, but conduction was restored progressively during the period of repair by remyelination. Remyelination also restored the ability of the axons to conduct closely spaced impulses (not shown).

(Reproduced from Smith KJ, Blakemore WF, McDonald WI: Central remyelination restores secure conduction. Nature 1979; 280:395-396. Reprinted by permission from Macmillan Publishers Ltd. Copyright 1979. http://www.nature.com/index.html.)

Conduction block in clinically eloquent areas is thought to cause all the major negative features of relapse in multiple sclerosis, including sensory and motor deficits and visual failure. The restoration of conduction by adaptive mechanisms is associated with remission, although conduction may remain insecure, and the recovery may be incomplete. Of course, conduction block occurring in clinically silent areas may have no detectable clinical consequences, although the total volume of deep white matter lesions is correlated with the extent of neuropsychological dysfunction.

Effects of Temperature

Small changes in temperature can have dramatic effects on conduction in demyelinated axons ⁷^–¹⁰ (Fig. 75-3), because the safety factor in demyelinated axons is typically reduced near to unity. The success of conduction in many demyelinated axons is therefore finely balanced, inasmuch as small changes can tip the safety factor either to just below unity, in which case conduction is blocked, or to just above unity, in which case conduction will succeed. Temperature changes affect the safety factor by altering the kinetics of the sodium channels.¹¹ Warming shortens the action potential,¹² thereby reducing the time that current is able to flow to depolarize the demyelinated region to its firing threshold, and cooling has the opposite effect.

Figure 75-3 Records showing the sensitivity of conduction in demyelinated axons to changes in temperature. Two families of superimposed records are shown, obtained from two excised dorsal columns examined in vitro, with the recording arrangement shown. The records on the left show conduction along normal tissue, whereas those on the right show conduction along a pathway that includes an experimental demyelinating lesion. Temperature changes (from 25° to 37°C in 1°C intervals) applied selectively to the tissue in the central recording lane had little effect on conduction in the normal tissue but profoundly affected conduction in the tissue with lesions. Here, the compound action potential included a second, delayed peak resulting from conduction in demyelinated axons. Conduction in these axons was markedly temperature sensitive, so that very few were able to conduct at normal body temperature, but many were able to conduct when cooled. STIM, stimulation; REC, recording.

(Reproduced from Smith KJ: Conduction properties of central demyelinated and remyelinated axons, and their relation to symptom production in demyelinating disorders. Eye 1994; 8:224-237. Reprinted by permission from Macmillan Publishers Ltd. Copyright 1979. http://www.nature.com/eye/index.html.)

Even subtle changes in body temperature can profoundly alter the expression of symptoms, such as in the visual system (Uhthoff’s phenomenon).¹³ Thus, vision can improve simply by drinking a glass of cold water,¹⁴ whereas a hot shower or bath, or sunbathing, can aggravate disability significantly.¹⁵ Indeed, in earlier years, the deleterious effects of body warming underpinned the “hot bath test” for multiple sclerosis. The therapeutic benefits both of body cooling^11,¹⁶ and of prolonging the action potential^17,¹⁸ have been considered. Thus, the potassium channel-blocking drugs 4-aminopyridine and 3,4-diaminopyridine can prolong the action potential, favoring successful conduction in demyelinated axons.¹⁹ These drugs have been found to improve disability in clinical trials in multiple sclerosis,²⁰ although it has also been suggested that this effect could arise from the potentiation of synaptic transmission.²¹ The widespread use of aminopyridines has been inhibited by the fact that they are markedly proconvulsant at doses that are not much higher than the therapeutic dose.²²

Activity-Dependent Conduction Block

Apart from fixed conduction block, demyelinated axons with a safety factor just above unity have a prolonged refractory period for conduction through the lesion ²³ (Fig. 75-4) and can also exhibit intermittent conduction failure as a result of periods of sustained electrical activity. The rise in the intra-axonal sodium concentration with repetitive firing stimulates the activity of the electrogenic Na⁺/K⁺-ATPase (sodium pump) and leads to axonal hyperpolarization.²⁴ Even a small hyperpolarization can effectively, and abruptly, block conduction in such demyelinated axons for periods of approximately 0.2 to 2 seconds.²⁵ Conduction resumes as the intra-axonal sodium concentration recovers and pump activity falls, and this cycle may then be repeated, causing a continuous train of impulses to be divided into short bursts separated by silent periods. The intermittent failure of impulse transmission on sustained activation may contribute significantly to the fatigability experienced by many patients with multiple sclerosis: for example, those with spinal cord lesions whose gait becomes impaired on walking short distances, or those whose vision “fades” or becomes blurred on fixated gaze, particularly in bright surroundings.^26,²⁷ There is therefore a potential therapeutic role for partial blockade of Na⁺/K⁺-ATPase, and indeed some benefit with such a therapy has been demonstrated in a small number of patients.²⁸

Figure 75-4 Records showing activity in a single unit isolated from a dorsal root caudal to a demyelinating lesion (hatched area) induced by the injection of diphtheria toxin into the cat dorsal column. The stimulus artifacts appear as dotted lines, and the action potentials as solid lines. Stimulation at S₁ includes the lesion in the conduction pathway, whereas the lesion is excluded by stimulation at S₂. The axon can conduct faithfully at 1000 Hz when the lesion is excluded from stimulation, but when it is included, and even when the stimulation frequency is reduced to 410 Hz, only three impulses are successfully propagated through the lesion before alternate impulses fail.

(Reproduced from McDonald WI, Sears TA: The effects of experimental demyelination on conduction in the central nervous system. Brain 1970; 93:583-598. Reprinted by permission of Oxford University Press.)

REMYELINATION

There is ample evidence that demyelinated axons can undergo repair by remyelination in multiple sclerosis ^29,³⁰ and that remyelination is likely to play a major role in restoring conduction and in the remission of neurological deficits after relapses.³¹ New internodes are thinner and shorter than normal, but remyelinated axons can nonetheless conduct, even when invested with only thin new myelin sheaths composed of as few as five myelin lamellae.³² New nodes formed by remyelination show aggregations of sodium channel immunoreactivity,³³ and, at least in the peripheral nervous system, the new nodes are excitable.³⁴ Not only do axons repaired by remyelination conduct (see Fig. 75-2) but also remyelination restores the security^5,³⁵ and velocity of conduction to near-normal values regardless of whether the repair is effected primarily by oligodendrocytes,³⁵ Schwann cells (even within the central nervous system),^36,³⁷ or transplanted olfactory ensheathing cells.³⁸ Although the role of axonal impulse activity in promoting myelination and remyelination remains controversial, it is possible that because demyelinated axons often exhibit conduction block, the absence of electrical activity could contribute to the failure of remyelination in some multiple sclerosis lesions.^39,⁴⁰ Remyelination is likely to confer not only secure conduction to repaired axons but also protection from degeneration, and developing strategies to promote remyelination is an active area of research.

ADAPTIVE CHANGES IN THE EXPRESSION OF ION CHANNELS IN DEMYELINATION

The clustering of voltage-sensitive ion channels and other axolemmal proteins, such as the L1 cell adhesion molecule family members neurofascin and NrCAM, and 480/270kD ankyrin at nodes of Ranvier,⁴¹ requires contact with, or at least close proximity to, myelinating Schwann cells in the peripheral nervous system and with oligodendrocytes in the central nervous system.⁴²^–⁴⁶ Because maintenance of ion channel distribution also seems to depend, at least in part, on the presence of compact myelin and the formation of normal, tight paranodal axoglial junctions,⁴⁷^–⁵⁰ demyelination can produce changes in the patterns of expression of these ion channels. In fact, the restoration of conduction in demyelinated axons involves substantial remodeling of the axonal membrane at a molecular level. Immunocytochemical studies have revealed an increased expression of two different isoforms of voltagegated sodium channels, Nav 1.6 and Nav 1.2, along demyelinated axons.^47,^48,⁵¹ Both of these sodium channels, particularly Nav 1.6, can generate persistent sodium currents, which may be important in axonal degeneration (see later discussion). The Nav 1.8 channel, which is normally expressed only in adult spinal sensory and trigeminal neurons, has also been shown to occur aberrantly in Purkinje cells in multiple sclerosis and in experimental autoimmune encephalomyelitis and potentially to contribute to cerebellar dysfunction.⁵²^–⁵⁵

On physiological examination, both a seemingly continuous distribution of sodium channels ⁵⁶ and, in contrast, an aggregation of sodium channels into “φ-nodes,”³⁴ have been described. φ-Nodes appear to be the precursors of the new nodes of Ranvier formed during remyelination, and it may be significant that the initial observation of a continuous distribution of sodium channels was made in a lesion (diphtheria toxin) in which repair by remyelination is only slowly achieved. It is still not clear how sodium channels aggregate during remyelination to form the new nodes of Ranvier. The formation of the channel aggregations may be driven by the axon and may also depend on the adjacent myelinating cells.^33,^44,^46,^50,^57,⁵⁸

These adaptive changes can, in some axons, lead to the restoration of conduction, even in the absence of repair by remyelination. Factors favoring the restoration of conduction in a particular axon include a small diameter ^59,⁶⁰ and either a short internode^59,^61,⁶² or a widened node before the demyelinated region. The first two of these factors characterize axons in the optic nerve, and they may contribute to the excellent recovery of vision that can follow optic neuritis. However, axons as large as 5.5μm in diameter have been shown to conduct in the absence of remyelination, which raises the possibility that most central axons might be able to conduct under ideal conditions. This possibility may help explain the presence in some patients of clinically silent demyelinating lesions in pathways that usually manifest in symptom production.⁶³^–⁶⁷

The restoration of conduction in segmentally demyelinated axons was first demonstrated by a sophisticated examination of conduction in spinal root axons demyelinated by the intrathecal injection of diphtheria toxin ⁵⁶; nearly 20 years were to pass before there was proof that conduction could also occur in demyelinated central axons,³² although circumstantial evidence had accumulated.⁶⁸^–⁷¹ In the peripheral nervous system, conduction can resume as soon as 6 days after the induction of a demyelinating lesion, but in the central nervous system, longer intervals (e.g., 2 to 3 weeks) appear to be required. Once restored, conduction is continuous, or microsaltatory, rather than saltatory, and it is therefore much slower than normal and also less secure. The velocity of conduction along a demyelinated segment of rat ventral root axons is reduced from normal, to the range of 0.7 to 2.3m/second^34,⁵⁶ and similar values are likely to apply in the central nervous system. Even though this degree of slowing occurs across only the demyelinated portion of an axon, it is still sufficient to disperse and delay the compound action potential and, perhaps coupled with a reduction in the total number of axons conducting, to result in diagnostically valuable delays in the visual,⁷² somatosensory, and brainstem auditory evoked potentials. Slowing of conduction (as opposed to conduction block) generally causes few symptoms, although sensations dependent on the precise timing of impulses in different axons may be affected. For example, patients with unilateral optic neuritis may perceive the Pulfrich phenomenon (the tendency of a pendulum to appear to trace a circle when a neutral density filter is placed over one eye, because of delayed transmission of information in the filtered pathway) without the normal need of the filter,⁷³ and auditory functions dependent on precisely coordinated information may be distorted by a unilateral lesion in the auditory pathway.

The reversal of neurological deficit during remissions in multiple sclerosis can be attributed to the restoration of conduction to demyelinated axons by a combination of remyelination and adaptive changes in the axonal membrane. It seems likely that the relative importance of each mechanism varies between patients and in individual patients at different times.

PLASTIC CHANGES

Recovery from a relapse of multiple sclerosis is also likely to involve “plastic” cortical changes; functional magnetic resonance imaging and other evidence support this view.⁷⁴^–⁸¹ Thus, to accomplish even minor tasks, patients with multiple sclerosis activate a larger part of the cortex than normal, including cortical areas not normally associated with the particular activity. Whether this increased cortical activation contributes to the sense of fatigue experienced by many patients remains unresolved.

POSITIVE SYMPTOMS: AXONAL HYPEREXCITABILITY

The recovery processes in demyelinated axons can lead not only to the restoration of conduction but also to axonal hyperexcitability.^82,⁸³ Ectopic impulses can be generated in regular or bursting discharges, and they conduct away from their site of initiation in both directions along the axon.^84,⁸⁵ Discharges in sensory pathways may give rise to tingling sensations referred to the body parts normally innervated by those axons. Similar discharges are likely to underlie other paroxysmal phenomena, including trigeminal neuralgia, tonic spasms, hemifacial spasm, and possibly episodic dysarthria and limb ataxia. Positive phenomena can be enhanced in patients by measures that increase axonal excitability, such as hyperventilation, and treatment with bicarbonate or calcium chelators.^83,⁸⁶^–⁸⁸

It appears that spontaneous discharges can arise from the excitatory effects of inward sodium currents that can develop in the demyelinated axolemma,^83,⁸⁹ and sodium channel blockers such as carbamazepine and phenytoin are effective treatments for positive symptoms. Spontaneous discharges could also arise from inward potassium currents that can result from the accumulation of potassium ions in limited extracellular compartments around axons.⁹⁰ Finally, it has been proposed that impulses can sometimes be “reflected” from sites of demyelination⁹¹^–⁹³; that is, an impulse propagating through a demyelinated site can induce the formation of a second impulse that travels back along the same axon in the opposite direction. A pair of reflecting sites could generate a train of apparently spontaneous impulses. Single, normally evoked impulses propagating through generator sites where any of these mechanisms of hyperexcitability have arisen could also trigger episodes of ectopic activity.^25,^91,^94,⁹⁵

Sensory axons are more prone to develop ectopic discharges than are motor axons, and, accordingly, positive sensory phenomena are more common in demyelinating disease than are their motor counterparts.^96,⁹⁷ These differences may occur partly because of the greater expression of persistent sodium currents along sensory axons.⁹⁸ The application of tumor necrosis factor directly to the sciatic nerve has been found to evoke ectopic discharges in nociceptive axons, including unmyelinated axons.⁹⁹ It is therefore possible that in inflammatory demyelinating diseases such as multiple sclerosis, tumor necrosis factor is involved in promoting the generation of ectopic discharges along demyelinated axons within inflammatory foci. Other inflammatory mediators may also favor ectopic activity. For example, exposure to nitric oxide is known to be conducive to persistent opening of sodium channels.^100,¹⁰¹

Ephaptic Transmission between Axons

Ephaptic interactions involving electrical crosstalk between demyelinated axons can explain a number of clinical observations in which paroxysmal, positive phenomena affect anatomically adjacent axonal pathways.¹⁰²^–¹⁰⁴ For example, tonic spasms may involve sensory symptoms along with contralateral limb contraction, as expected from a lateral spread of excitation from the spinothalamic to the corticospinal tract within the spinal cord. Hartmann¹⁰² described a patient in whom trigeminal neuralgia could be triggered by auditory stimuli and who had a pontine lesion involving the adjacent lateral lemniscus and trigeminal sensory pathway on magnetic resonance imaging. That said, ephaptic transmission has rarely been demonstrated electrophysiologically. Indeed, the best documented interaction occurs between normal and amyelinated (i.e., never myelinated, rather than demyelinated) axons in the spinal roots of the dystrophic mouse.^105,¹⁰⁶

The authors have occasionally observed massed synchronous discharges arising from the spinal cord in animals with experimental demyelinating lesions in the dorsal columns.¹⁰⁷ The discharges take the form of repeated bouts of high-frequency bursts of compound action potentials, but their relevance to multiple sclerosis remains uncertain.

Mechanosensitivity

Demyelinated axons can become markedly mechanosensitive, so that even mild deformation can result in the generation of bursts of impulses or change the firing frequency of axons that are already spontaneously active.¹⁰⁸ Stretch-sensitive ion channels, if they appear in the demyelinated membrane, might underlie mechanosensitivity, but this possibility remains speculative. It is reasonable to suppose that the mechanosensitivity of demyelinated axons in the cervical dorsal columns is responsible for Lhermitte’s phenomenon¹⁰⁹^–¹¹¹ and for the perception of flashes of light (phosphenes) upon eye movements in patients with demyelinating lesions of the optic nerve.

THE ROLE OF INFLAMMATION

It is now accepted that the inflammatory response in multiple sclerosis plays a significant role in symptom production and clinical relapse.¹¹² Clinical and magnetic resonance imaging studies have shown a correlation between inflammation in the optic nerve and visual loss in patients with acute optic neuritis,¹¹³ and administration of the antilymphocyte antibody alemtuzumab (Campath-1H) to patients with multiple sclerosis provokes acute exacerbations of previously expressed symptoms, presumably because conduction in damaged axons had been blocked.¹¹⁴

Inflammation may impair conduction indirectly by opening the blood-brain barrier, thereby exposing axons to potentially deleterious factors in the vasculature, or directly, as a result of the local action of soluble agents such as nitric oxide, cytokines, peptides, and antibodies produced during the immune response. Demyelinated axons have an inherently low safety factor for conduction and are therefore likely to be particularly susceptible to factors impairing conduction, even when the factors may exert only subtle effects on normal axons. Clinically, the time course of relapse and remission in multiple sclerosis tends to coincide with the onset and resolution of inflammatory activity within plaques, which supports the beliefs that inflammation contributes to the deficit and that recovery is associated with the restoration of conduction as inflammation subsides.¹¹³

Cytokines

Cytokines have been implicated as a cause of conduction block in some clinical studies, particularly in axons already compromised by disease processes. For example, the symptoms that occur after the administration of alemtuzumab are associated with a transient elevation of circulating proinflammatory cytokines,¹¹⁵ and they can be prevented by pretreatment with steroids. These neurological effects could be mediated by direct effects of cytokines on axonal ion channels and/or indirectly, because cytokines such as interleukin-1β and interferon γ are potent inducers of the inducible form of nitric oxide synthase (iNOS) in human astrocytes and also probably in microglia. It may therefore be clinically significant that interferon β inhibits both the expression of iNOS ¹¹⁶ by astrocytes and the production of reactive oxygen species by monocytes from patients with relapsing-remitting multiple sclerosis.

Nitric Oxide

Levels of nitric oxide metabolites and of iNOS are increased in multiple sclerosis,^117,¹¹⁸ and low micromolar concentrations of nitric oxide can block axonal conduction,^119,¹²⁰ particularly in demyelinated axons.¹¹⁹ The blockade is reversible, which raises the possibility that nitric oxide may be an important contributor to the reversible clinical deficits in the relapses and remissions of multiple sclerosis. The mechanisms underlying conduction block have not been established but may include direct effects of nitric oxide on ion channels,^100,¹²¹^–¹²³ inhibition of mitochondrial respiration by nitric oxide,^124,¹²⁵ and effects on the levels of axonal cyclic guanosine monophosphate.¹²⁶ If nitric oxide does impair conduction in inflammatory demyelinating disease, then agents that inhibit its production should ameliorate symptoms. However, this rationale may have unpredictable consequences in the context of an autoimmune demyelinating disease, in which nitric oxide may have protective as well as damaging effects. There have been more than 40 studies of the role of nitric oxide and its related compounds in experimental autoimmune encephalomyelitis, but there is little consensus regarding its effects or the mechanisms through which they are mediated.¹²⁷

Antibodies

Antiganglioside antibodies have been identified in patients with progressive forms of multiple sclerosis.^128,¹²⁹ These antibodies may contribute to the neurological deficit because some (but not all) studies have revealed that they can have electrophysiological effects on peripheral axons.¹³⁰

DISABILITY AND AXONAL DEGENERATION

It has long been believed that the persistence and progression of disability in multiple sclerosis can be explained by persistent conduction block in demyelinated axons caused by a failure of the repair mechanisms discussed earlier. However, it is now clear that axonal degeneration also plays an important and perhaps primary role.¹³¹^–¹³⁶ In pathological studies, there is prominent axonal degeneration at an early stage of multiple sclerosis,^137,¹³⁸ and in magnetic resonance studies, there is atrophy of the whole brain and regions that include the cerebellum, spinal cord and thalamus, as well as a reduction of the neuronal marker N-acetyl aspartate.^131,^133,¹³⁹ Axonal injury is correlated with disability in both the pathological and magnetic resonance studies.

There are probably numerous mechanisms of axonal degeneration. However, the magnitude of axonal degeneration in pathological studies is correlated with the intensity of the inflammatory response, which raises the possibility that some mediators of inflammation may have a damaging effect on adjacent axons. Experimentally sustained impulse activity at physiological frequencies has been shown to result in axonal degeneration, even in normal axons, if the activity occurs in conjunction with exposure to nitric oxide (Fig. 75-5).¹⁴⁰ The combination of impulse activity and nitric oxide exposure may be expected to occur within inflammatory lesions; therefore, this mechanism could contribute to axonal degeneration in multiple sclerosis. In this respect, it is interesting that the smallest axons are the most susceptible to degeneration caused by impulse activity during nitric oxide exposure and that they are also the axons that are preferentially lost in multiple sclerosis lesions.^141,¹⁴²

Figure 75-5 Records showing the vulnerability of nerve fibers to degeneration as a result of sustained impulse activity during exposure to nitric oxide (NO). Two series of averaged compound action potentials can be seen, recorded in parallel from two dorsal roots in an anesthetized rat, in the arrangement indicated (see inset). The data are shown in three-dimensional perspective, with the earliest records at the front and with a 2-minute interval between adjacent records; each plot therefore shows approximately 12 hours of recorded data. The left plot shows records obtained at continuous 1-Hz stimulation, whereas in the right plot, the nerve fibers were stimulated at 100 Hz for the first 6 hours of recording. On the left, the records were stable for the first 2.5 hours, but conduction block was imposed on nearly all the axons by a 2-hour exposure to nitric oxide. Such block has been described by Redford and colleagues,¹¹⁹ and it might help explain how inflammation can cause negative symptoms in multiple sclerosis (see text). The block was released on removal of nitric oxide, and the axons continued to conduct for the remaining 7.5 hours of the experiment. On the right, conduction was also blocked during exposure to nitric oxide, but recovery after washing was inconsistent and only temporary, so that within 2 hours, conduction had ceased in all axons. Histological examination of the roots at the end of the experiment revealed that all the axons exposed to nitric oxide in conjunction with stimulation at 100 Hz had undergone degeneration, whereas the root stimulated at only 1 Hz during the period of nitric oxide exposure was quite normal in appearance.

(Redrawn from Smith KJ, Kapoor R, Hall SM, et al: Electrically active axons degenerate when exposed to nitric oxide. Ann Neurol 2001; 49:470-476.)

The mechanisms responsible for the effects of nitric oxide are not yet certain, but because nitric oxide is known to inhibit mitochondrial energy production, and because demyelinated axons express more sodium channels, it seems reasonable to hypothesize that axonal degeneration might result from the combination of a reduced capacity for energy production as a result of nitric oxide exposure and an increased energy demand from sodium loading of the axon as a result of impulse activity.¹⁴³ The consequent failure of Na⁺/K⁺-ATPase to maintain ionic homeostasis would result in an elevation of the intra-axonal sodium ion concentration, perhaps to a level capable of driving the sodium-calcium exchanger into reverse activity.¹⁴⁴ This sequence of events could culminate in a pathological accumulation of calcium ions within the axon and the activation of intra-axonal degrading enzymes.

Demyelinated axons in multiple sclerosis or experimental autoimmune encephalomyelitis can also express immunoreactivity for subunits of the N-type calcium channel:¹⁴⁵ If such channels are functional, they could add to an influx of calcium ions during impulse activity. Furthermore, there is increasing evidence of mitochondrial genetic dysfunction in multiple sclerosis,^145a which might render subgroups of patients more vulnerable to axonal injury that depends on energy failure.

This hypothesis, which links inflammation, increased expression of sodium channels, metabolic failure, and ultimately calcium overloading of the axon, has received support from studies showing that sodium channel blockade with drugs such as phenytoin,¹⁴⁶ flecainide (Fig. 75-6),¹⁴⁷^–¹⁵⁰ and lamotrigine (unpublished observations) can indeed protect axons from degeneration, both on exposure to nitric oxide and in animals with experimental autoimmune encephalomyelitis and experimental autoimmune neuritis.

Figure 75-6 Representative photomicrographs of the dorsal columns of rats with chronic relapsing experimental allergic encephalomyelitis. The axons have been labeled immunohistochemically for neurofilaments. Control animals show substantial axonal loss (left; seen at higher magnification, bottom), whereas in animals treated with the sodium channel-blocking agent flecainide, most of the axons have survived (right).

(Modified from Bechtold DA, Kapoor R, Smith KJ: Axonal protection using flecainide in experimental autoimmune encephalomyelitis. Ann Neurol 2004; 55:607-616.)

Glutamate Excitotoxicity

Glutamate is likely to be present in active multiple sclerosis lesions after its liberation by activated microglia and leukocytes ¹⁵¹ and perhaps by axons,¹⁴⁴ and its levels may be amplified still further within inflammatory foci by the action of proinflammatory cytokines, such as interleukin-1β and tumor necrosis factor α, which have been reported to inhibit the glutamate uptake activity of astrocytes in vitro by a mechanism that involves nitric oxide. Glutamate-mediated excitotoxicity acting through Ca²⁺-permeable α-amino-3-hydroxy-5-methyl-4-isoxazole propionic acid (AMPA)/kainate receptors may contribute to the death of oligodendrocytes (and neurons) in multiple sclerosis, and indeed the blockade of AMPA/kainate receptors has been found to reduce both axon and oligodendrocyte loss in experimental allergic encephalomyelitis, with a corresponding amelioration of clinical outcome.^152,¹⁵³ Furthermore, excitotoxic damage to oligodendrocyte-like cells is prevented by agents that elevate cyclic adenosine monophosphate levels, such as rolipram, which has been shown to be protective in experimental allergic encephalomyelitis.^154,¹⁵⁵ The precise mechanisms of injury to neurons and to oligodendrocytes may differ, inasmuch as oligodendrocytes appear to possess only AMPA/kainate receptors, whereas neurons possess N-methyl-D-aspartate receptors as well. The situation is less clear for axons, although there is emerging evidence that they do express novel glutamate receptors after demyelination. Finally, research suggests that axonal degeneration in experimental allergic encephalomyelitis may be influenced by the activity of cannabinoid receptors, acting in part through their effects on glutaminergic systems.¹⁵⁶

Cortical Lesions and Neuronal Loss in Gray Matter

The lesions of multiple sclerosis are not restricted to the white matter, and different forms of cortical lesions are known to occur.^157,¹⁵⁸ Their contribution to the clinical features of multiple sclerosis is a subject of increasing interest. Magnetic resonance imaging and spectroscopy have demonstrated that atrophy and lowered N-acetyl aspartate levels are present in cortical and deep gray matter areas.¹⁵⁹^–¹⁶³ This work, which is in keeping with the results of histopathological studies, indicates that there is a significant component of degeneration of neuronal cell bodies in multiple sclerosis, not only in advanced secondary progressive disease but also in early relapsing disease. Gray matter neuroaxonal pathology may therefore be an important factor in the development of disability. The mechanisms of damage to neuronal cell bodies probably differ to some extent from those causing axonal degeneration, and it may be relevant that cortical lesions in multiple sclerosis contain a smaller leukocytic infiltrate and more microglial activity than do white matter lesions.

KEY POINTS

• The pathological hallmarks of multiple sclerosis are demyelination, inflammation, and neuroaxonal degeneration. The pathophysiological consequences of these events account for the key clinical features of the disorder, including the phenomena of relapse, remission, and the progressive accumulation of disability.

• Demyelination can lead to axonal conduction block, the main cause of neurological dysfunction in relapse, and can also be associated with temperature-dependent symptoms and fatigability of function.

• A significant cause of the conduction block and disability during relapse is believed to be associated with the inflammation that can be a prominent component of lesions, rather than with demyelination alone.

• Recovery from relapse involves remyelination and adaptive changes in the axonal membrane, including an increased expression of sodium channels. “Plastic” changes also contribute.

• Residual and progressive disability is related to the failure of these adaptive mechanisms and to neuroaxonal degeneration, which occurs at an early stage of the illness.

• The mechanisms of neurodegeneration are being clarified, and some research suggests that it is realistic to believe that a neuroprotective treatment may be achievable within the foreseeable future.