The New England Journal of Medicine - January 29, 1998 - Volume 338, Number 5
Demyelinating Diseases - New pathological Insights, New Therapeutic Targets
Virtually every textbook of neurology or general medicine includes chapters on demyelinating diseases, with most of the attention devoted to multiple sclerosis. The concept of multiple sclerosis as a demyelinating disease is deeply ingrained. The early description of multiple sclerosis by Charcot stressed the loss of myelin. The diagnosis of multiple sclerosis rests in part on the demonstration, by measurement of evoked potentials, of slowed action-potential conduction, a physiologic hallmark of demyelination.
The lipid-rich myelin sheath is produced by Schwann cells in peripheral nerves and by oligodendrocytes in the brain and spinal cord, and myelin possesses high electrical resistance and low capacitance and thus acts as an insulator around axons. The myelin is arranged in segments separated by nodes of Ranvier, where sodium channels are clustered in high density in the axon membrane so that they can produce action potentials (Figure 1A). Myelin covers and masks the internodal parts of the axon, which contain fewer sodium channels and a higher density of potassium channels, which tend to oppose the generation of action potentials. Myelination increases the speed of conduction and improves its metabolic efficiency. Damage to the myelin is accompanied by a decrease in conduction velocity and, when severe, by conduction block ( Figure 1B). (1) In addition, cell-cell interactions between myelin-forming glial cells and their underlying axons actively influence the biochemical properties of the axons; a loss of the myelin is associated with destabilizing changes in the molecular structure of the axonal cytoskeleton. (2)
Multiple sclerosis can take several forms. In the relapsing-remitting type of multiple sclerosis, the patient's course is punctuated by exacerbations or relapses in which there is clinical worsening, but these are followed (within weeks to a few months) by remissions with partial or full recovery from the deficits. The molecular substrate for remissions appears to be provided by a remodeling of the demyelinated (formerly internodal) axonal membrane so that it acquires a higher-than-normal sodium-channel density, which permits conduction of action potentials despite the loss of myelin (Figure 1C). (3,4,5) Progressive forms of multiple sclerosis are characterized by a downhill course without remissions, so that the patient acquires more and more clinical deficits, either beginning at presentation (primary progressive form) or after a period of relapsing-remitting disease (secondary progressive form).
Why do the patients with progressive multiple sclerosis not have remissions? An answer to this question might help us to approach the important objective of limiting the development of permanent deficits. In a study reported in this issue of the Journal, using confocal microscopy and computer-based. Three-dimensional reconstructions, Trapp et al. (6) have provided an elegant demonstration of substantial damage to axons, as well as myelin, in the brains of patients with multiple sclerosis. Using antibodies to nonphosphorylated neurofilaments (a marker of axonal regions that lack myelin), they demonstrate axonal transection throughout active lesions (including acute lesions early in the course of the disease) and within chronic active lesions, particularly at the edges of actively demyelinating lesions, where major-histocompatibility-complex class II-positive inflammatory cells are abundant. They postulate that axonal degeneration (Figure 1D) is a pathologic correlate of irreversible neurologic impairment in multiple sclerosis.
This demonstration of axonal pathology in multiple sclerosis builds on a history of hints throughout the literature that it is more than just a demyelinating disease. Even Charcot's early description of multiple sclerosis mentioned axonal pathology. In the 1970's, on the basis of more contemporary molecular dissection of myelinated axons, which demonstrated a mutual interdependence of axons and myelin-forming glial cells, some of us speculated that there could be no "pure" demyelinating diseases. McDonald et al. (7) suggested in 1992, on the basis of magnetic resonance imaging and electron microscopy of brain tissue from patients with multiple sclerosis, that as lesions age, there is progressive axonal loss. This group also used magnetic resonance spectroscopy to study cerebellar white matter and found a significant reduction in a neuronal marker (N-acetyl aspartate) that was correlated with the presence of cerebellar deficits; these observations provided evidence that axonal loss is important in the development of persistent neurologic disability in multiple sclerosis.
The studies by Davie et al. (8) and Narayanan et al. (9) also found evidence of axonal degeneration in normal-appearing white matter outside of demyelinating lesions, possibly due to wallerian degeneration of axonal projections that had been disconnected from their origins as a result of transection. Using amyloid precursor protein as a histopathological marker of damaged axons in the brains of patients with multiple sclerosis, Ferguson et al. (10) recently found evidence of axonal injury throughout acute lesions and at the margins of active chronic lesions; like Trapp et al., (6) they interpret their results as suggesting that axonal damage may be associated with inflammation.
If axonal injury early in the course of disease contributes to the development of irreversible neurologic deficits, the prevention of axonal loss might be expected to prevent persistent disability. Trapp et al. (6) and Ferguson et al. (10) mote a relation between axonal injury and inflammation, suggesting that a reduction in the inflammatory response might result in the loss of fewer axons and thus in less clinical deficit. An alternative approach is suggested by studies that have demonstrated that ion channels and exchangers (11, 12, 13) together form a "final common pathway," subject to modulation by neurotransmitters, (14) that underlies axonal degeneration after various injuries. Axonal function and integrity can be preserved after acute insults by means of neuroprotective interventions that block or modulate injurious ion fluxes at several stages within this molecular death cascade (12, 14, 15) or that interfere with "downstream" degenerative events such as activation of calpains and other destructive enzymes. (13) Further studies will be needed to determine whether the reduction in inflammatory responses or the neuroprotection of axons can limit or prevent axonal degeneration in multiple sclerosis and, if so, whether this will reduce or prevent the acquisition of persistent neurologic deficits.
Chapters on "demyelinating" diseases will not necessarily become shorter as a result of the reclassification of multiple sclerosis as an "axonal" disorder. It has recently been recognized that in some patients with traumatic (nonpenetrating) spinal cord injury there are residual axons that maintain continuity through the lesion but fail to conduct impulses as a result of demyelination. These findings have been reported in some patients with "clinically complete" lesions (i.e., those with no function below the level of the lesion), which are classically considered to be due to transection of the spinal cord and its constituent axons within the lesion. (16, 17, 18) The demonstration of these preserved, but demyelinated, axons suggests that in spinal cord injury, at least some degree of functional recovery might be achieved by strategies that restore impulse conduction along demyelinated axons.
Recognition that multiple sclerosis is, in part, an axonal disease and that spinal cord injury is, in part, a disorder of myelin should trigger a critical rethinking of these disorders and provides us with new targets for therapy. Ideally, future studies will tell us whether the protection of axons from injury in multiple sclerosis and the repair of demyelinated axons in spinal cord injury are therapeutic strategies that will help preserve neurologic function in patients with these disorders.
Stephen G. Waxman, MD, Ph.D.
Yale School of Medicine
