Electrically Active Axons Degenerate When Exposed to Nitric Oxide Kenneth J. Smith, PhD,1,2 Raju Kapoor, DM, FRCP,1,3 Susan M. Hall, DSc,1,4 and Meirion Davies, BSc,1,2 Axonal degeneration is a major cause of permanent deficit in inflammatory neurological diseases such as multiple sclerosis. Axons undergo degeneration specifically at the site of the inflammatory lesions, suggesting that locally produced inflammatory factors mediate the phenomenon. One such factor is nitric oxide (NO), which we have previously reported can cause reversible conduction block in axons. Here we confirm these observations and extend them to show that axons exhibit the early stages of wallerian degeneration if they are conducting impulses at physiological frequencies while they are exposed to the low micromolar concentrations of NO that are likely to occur at sites of inflammation. Rat dorsal roots were concurrently exposed in vivo to both NO and sustained impulse activity at 1, 50, or 100 Hz. Although our in vivo observations necessarily focused on the more acute responses, morphological examination of exposed roots at the end of the recording period revealed nodal and paranodal changes consistent with acute wallerian degeneration in roots stimulated at 50 or 100 Hz. This interpretation was confirmed in a few experiments that were prolonged to permit more obvious indicators of degeneration to develop. In these experiments the formation of myelin ovoids and frank axonolysis occurred in more than 95% of fibers. Roots stimulated at only 1 Hz appeared normal. We propose that the combination of normal impulse traffic and NO at sites of inflammation may cause axonal degeneration and that electrical activity may therefore be an important factor in causing permanent disability in patients with neuroinflammatory disorders. Ann Neurol 2001;49:470 – 476 Axonal degeneration is an important cause of permanent disability in several neurological disorders in which inflammation plays a role, including multiple sclerosis (MS), inflammatory neuropathies such as Guillain-Barré syndrome, and stroke.1–3 In MS, at least, progressive axonal loss is the major cause of the gradual accumulation of permanent deficit in progressive disease.2,4 –7 Although it is recognised that axonal degeneration is an important cause of morbidity, the underlying pathogenetic mechanisms have not been identified. Pathological studies have revealed that in MS the axons appear to be transected specifically within the lesions and that the number of axons transected correlates with the severity of the inflammation.8,9 There is little evidence of a direct immune attack on axons, and it is assumed that axonal loss may, for example, be a consequence of exposure to the cocktail of deleterious factors produced at inflammatory sites, such as proteases, cytokines, and free radicals, including nitric oxide (NO). NO is known to be produced in raised concentrations at sites of inflammation, such as MS lesions,10 –15 as a result of the activity of the inducible form of nitric oxide synthase (iNOS), which is capable of releasing NO in low micromolar concentrations. At such concentrations, NO can impair mitochondrial metabolism.16,17 Indeed, mitochondrial function is known to be impaired in an animal model of MS.18 We have therefore reasoned that, not only may exposure to NO cause irreversible axonal damage, but axons in which metabolic activity is increased, for example, by sustained impulse activity, might be particularly vulnerable to its effects. To test these hypotheses, we have exposed axons in vivo to NO at low micromolar concentrations while simultaneously inducing a sustained train of impulses at physiological frequency along the axons. From the 1Neuroinflammation Research Group and 2Department of Neuroimmunology, Guy’s, King’s and St Thomas’ School of Medicine; 3National Hospital for Neurology and Neurosurgery; and 4 Guy’s, King’s and St Thomas’ School of Biomedical Sciences, London, United Kingdom. Address correspondence to Prof Smith, Department of Neuroimmunology, Guy’s, King’s and St Thomas’ School of Medicine, Guy’s Campus, St Thomas Street, London, SE1 9RT, UK. E-mail: email@example.com Received Aug 8, 2000, and in revised form Nov 1. Accepted for publication Nov 5, 2000. 470 © 2001 Wiley-Liss, Inc. Methods Rats (250 to 350 gm, male and female, Sprague-Dawley and Lewis) were anesthetized (halothane 1–1.5% in nitrous oxide:oxygen [2:1]) and prepared for a terminal electrophysiological examination. The spinal cord and spinal roots were exposed and the dura cut longitudinally beneath a mineral oil recording pool maintained at 37 ⫾ 0.1°C with radiant heat. The end-tidal CO2 and blood pressure were continuously monitored, and artificial ventilation was applied during prolonged recording sessions, if required. Two pairs of dorsal roots (usually left and right S1 and S2) were left in continuity at either end but separately laid across a pair of platinum wire stimulating electrodes, through an electrically grounded bath 7 mm in length in which the root could be exposed to media of various compositions, and across a pair of platinum wire recording electrodes (inset, Fig 1). Supramaximal electrical stimuli were applied to each root via an isolated stimulator (Digitimer DS2, Digitimer, Welwyn Garden City, UK) and the resulting antidromically conducted compound action potentials recorded differentially and stored on computer disc. Initially the bath contained tissue culture medium (Leibovitz), and control recordings were made for 1 to 2.5 hours to demonstrate the stability of the preparation. The medium in the bath was then exchanged for one containing the NO donor DETA NONOate (Alexis, Nottingham, UK). The concentration of NO was measured using an amperometric meter (ISO-NO II, WPI, Stevenage, UK). After 2 hours, the donor medium was replaced with Leibovitz’s medium for the remainder of the experiment. Each set of recordings in one animal included a comparison of the effects of NO at various frequencies of stimulation. Some roots were stimulated at 1 Hz, and others were stimulated at higher frequency. Each experiment also included at least 1 control root, which was exposed to a NO donor solution that was depleted of its NO content either by prior temporary acidification or by the inclusion of an NO scavenger such as hemoglobin or 2-phenyl-4,4,5,5-tetramethylimidazoline-1-oxyl-3-oxide (PTIO). At the termination of each experiment the rat was killed by anesthetic overdose, and the portion of the dorsal roots contained within the 7-mm bath was removed and fixed by immersion in glutaraldehyde (3% in 0.1 M PO4 buffer). The fixed roots were prepared for light microscopic examination in resin sections using standard protocols. The appearance of myelinated fibers was scored in 1-m resin transverse and longitudinal sections stained with thionin acridine orange by an observer who was blind to the experimental protocol used. Fibers were scored according to the following criteria: 0, no change from baseline conditions (ie, nodal gaps ⬍1m; compact paranodal myelin; Schmidt-Lanterman incisures “closed”; no evidence of widened periaxonal space at any point along internodal, paranodal, or nodal axolemmae; normal textured axoplasm containing the typical complement of scattered organelles, including elongated mitochondria); 1, at least 50% of axons within the root showing widened nodal gaps (⬍5m) produced by apparent retraction of paranodal myelin (which was often unilateral) but with no expansion of the periaxonal space; 2, at least 50% of axons within the root showing bilateral disruption and retraction of paranodal myelin (which typically exposed ⬎10 m of paranodal axolemma), accumulations of organelles in paranodal axoplasm, and short regions of expanded periaxonal space; 3, at least 50% of the axons within the root showing nodal and paranodal changes as in category 2 (sometimes accompanied by obvious swelling of paranodal Schwann cell cytoplasm) and extensive expansion of periaxonal spaces. Results The electrophysiological and morphological consequences of NO exposure were dependent on the frequency of impulse conduction. At 1-Hz stimulation, exposure to NO (1–10 M) for 2 hours resulted in a concentration-dependent block of conduction, and the block was reversed fully or almost fully upon removal of the NO (12 out of 12 roots). This effect of NO has been previously described by us and another group.19,20 In fact, with such low-frequency stimulation, the conduction block during a 2-hour exposure to NO was typically reversible unless very high concentrations of NO were used (⬎15 M); such concentrations are probably outside the pathophysiological range. However, persistent conduction block was produced by exposure to lower NO concentrations if exposure occurred in conjunction with stimulation of the axons at physiological frequencies. Thus, persistent conduction block occurred in many axons in 47 out of 50 roots exposed to NO at 1 to 10 M during stimulation at 100 Hz. Stimulation at the intermediate rate of 50 Hz during NO exposure resulted in persistent conduction block in many axons in 7 out of 15 roots. No persistent block occurred in roots stimulated for 5 to 6 hours at either 50 Hz (6 out of 6 roots) or 100 Hz (7 out of 7 roots) in the absence of NO. NOdepleted solutions had few if any effects on either the electrophysiological or the morphological properties of the axons. The effectiveness of NO-scavenged solutions depended on the concentration of scavenger; occasionally delayed effects attributable to NO were seen once the scavenger had been exhausted. Impulse activity rendered the axons more sensitive to the conduction-blocking effects of NO. Thus, whereas a 2-hour exposure to a relatively low concentration of NO (⬍0.1–1.0 M) had little effect on conduction at 1 Hz stimulation (5 out of 5 roots), such low concentrations were sufficient to block conduction in many axons (50 –100%) if they were stimulated at 100 Hz during NO exposure. A similar effect of impulse frequency was observed at higher NO concentrations (1– 6 M). Figure 1 shows four series of records obtained simultaneously in one animal from 4 spinal roots exposed to 0 or 4 M NO and to 1-, 50-, or 100-Hz stimulation. For each root the same frequency of stimulation was applied continuously throughout the 5-hour duration of the recording. The first two plots show control data and establish that the axons can conduct impulses continuously at 50 Hz for 5 hours in the absence of NO or conduct at 1 Hz when exposed for 2 hours to a medium liberating NO at 4 M concentration. However, the third plot shows that the combination of 50-Hz stimulation and exposure to 4 M NO results in a prompt reduction in the amplitude of the compound action potential, indicating the appearance of Smith et al: Impulse Activity, NO, and Degeneration 471 Fig 1. Four series of averaged (n ⫽ 8) compound action potentials recorded in parallel from 4 separate dorsal roots in a rat (female, Lewis) using the arrangement indicated (see inset). The data are shown in three-dimensional perspective, with the earliest records at the front and a 2-minute interval between adjacent records; each plot therefore shows approximately 5 hours of recorded data. The first plot shows control data obtained in response to continuous supramaximal stimulation of the root at 50 Hz. The stimulus artifact (indicated) can easily be distinguished from the compound action potential, which remains constant in configuration and amplitude throughout the recording period. During a 2-hour period (indicated) the culture medium in a 7-mm long bath surrounding the middle portion of the root was exchanged for one nominally containing the NO donor DETA NONOate. The concentration of donor was the same as that applied to the other spinal roots, but in this case the donor had been entirely depleted of its NO content by prior, temporary acidification. The second plot, also for a control root, shows the effect at 1-Hz stimulation of a medium liberating NO at 4-M concentration. Nearly all the axons continue to conduct, although there is a gradual appearance of conduction block in a few axons as time progresses; this block was reversed when the NO donor was removed at the end of the 2-hour exposure period. The third plot shows the effect of 50-Hz stimulation in conjunction with exposure to 4 M NO. During the first hour of recording, the compound action potential is quite stable, but when the root is exposed to NO at 4 M there is a prompt reduction in the amplitude of the small, second negative (ie, upward) peak, indicating the appearance of conduction block in the more slowly conducting axons. (These axons are the most susceptible to conduction block by NO, as can most clearly be seen in monophasic recordings [see Fig 4 in Ref 19]. However, monophasic recordings typically require the roots to be cut or crushed, and we have therefore avoided their use in these experiments because damaging the roots would affect their histological appearance and interfere especially with the crucial determination of the presence or absence of early wallerian-type degeneration.) After about 15 minutes of NO exposure, the amplitude of the main peak of the compound action potential also gradually declines, indicating the appearance of conduction block in the majority of the axons in the root. The fastest axons are relatively resistant to the effects of NO, and so a peak of the compound action potential with short latency persists. Washing the root after 2 hours of NO exposure results in the recovery of conduction in some axons. However, the recovery is only temporary. After 1 to 2 hours conduction again diminishes, and the appearance at 5 hours is one of persistent conduction block in most of the axons. The final plot shows the effect of the same NO concentration with stimulation at 100 Hz. The onset of exposure to NO is associated with a relatively prompt reduction in the amplitude of the compound action potential, indicating that conduction is blocked in most axons within 10 minutes. The fastest axons are again relatively resistant to the effects of NO, but they also fail to conduct after about 1 hour of exposure. Only a few axons recover the ability to conduct upon removal of the NO, and in some of them the recovery is only temporary. As at 50 Hz, the combination of sustained impulse activity and the presence of NO results in persistent conduction block in most axons. 472 Annals of Neurology Vol 49 No 4 April 2001 conduction block in many axons. The block is persistent in most of the axons after a temporary recovery of conduction in some axons (see Fig 1 legend for a more complete description). The persistent conduction block occurs in nearly all the axons if they are stimulated at 100 Hz (fourth plot). Histological examination of the roots at the end of the recording period confirmed that many of the axons in the roots stimulated at 50 or 100 Hz showed evidence of paranodal and nodal damage, myelin splitting, and disruption of Schmidt-Lanterman incisures, whereas these changes were rare in the 2 control roots (left of plot). Where present, the effects were most marked in the smallest-caliber fibers. It may be significant that small-diameter axons are selectively lost in the spinal cords of MS patients.21,22 In the experiment shown in Fig 1, the recovery period following NO exposure was only 2 hours. We recognize that this is unlikely to be a sufficient time for unequivocal morphological evidence of axonal degeneration to develop, and so the recording period was prolonged as much as was technically feasible in several experiments. Figure 2 shows an experiment in which recording persisted for 12 hours. Exposure to 10 m NO for 2 hours caused reversible conduction block in almost all axons at 1-Hz stimulation, but at 100-Hz stimulation exposure to the same NO concentration resulted in persistent conduction block in all axons following a temporary period when conduction was restored in some axons. Histologically, roots stimulated at 1 Hz during NO exposure appeared normal, whereas all or nearly all fibers in roots stimulated at 100 Hz were undergoing ovoid formation and axonolysis typical of acute wallerian-type degeneration (Fig 2B). Morphological data were systematically studied from 11 animals (44 spinal roots) with survival times of between 2 and 9 hours following NO exposure and in which electrophysiological data were available for comparison among all 4 roots. Sixteen of the 44 roots were exposed to NO in conjunction with stimulation at 50 or 100 Hz, and the remaining roots acted as control roots exposed either to the same NO concentrations at 1 Hz stimulation or to stimulation at 50 or 100 Hz in the absence of NO. Control roots contained few, if any, fibers scoring greater than 1. In marked contrast, with the exception of roots exposed for the shortest time periods to NO, all affected fibers (13 out of 16) within roots exposed to both insults (ie, with 50 – 100-Hz stimulation in conjunction with NO concentrations from as low as 1 M) had scores of 2 or 3, that is, they exhibited the morphological correlates of early wallerian degeneration. In longitudinal sections, many axons exhibited an extensive expansion of what appeared to be the periaxonal space, indicating significant and possibly irreversible alterations in axon– Schwann cell relationships. It seems reasonable to as- sume that disruption of nodal and paranodal morphology preceded the loss of contact between axons and their ensheathing Schwann cells. That the two phenomena were causally linked is supported by our consistent finding that widening of the periaxonal space was most typically observed within the first cylindricoconical segment up- or downstream from a node and was less common in mid-internode. In two of the longer-term experiments, degenerative changes in more than 95% of the fibers within the roots had progressed to the stage of ovoid formation and axoplasmic dissolution, previously described as granular disintegration of the cytoskeleton. Discussion We have demonstrated that, at the light microscopic level, fibers undergo morphological changes consistent with acute wallerian degeneration when exposed to a combination of impulse activity and NO. Neither impulse activity nor NO exposure caused similar changes when applied independently at similar levels. Because these experiments have revealed incipient fiber degeneration after NO exposure for only 2 hours, we speculate that more prolonged exposure, which would occur at a site of inflammation, might result in axonal degeneration at lower frequencies of impulse activity and/or at lower concentrations of NO. In MS, there are likely to be several mechanisms, acting over different time courses, by which axons undergo degeneration. However, significant amounts of axonal degeneration or transection have been reported in inflammatory lesions,23 suggesting that individual axons can degenerate acutely at inflammatory sites. Our findings may be pertinent to the mechanism of this damage, which, occurring at different sites over time, could explain the apparently gradual accumulation of axonal loss in MS. To mimic the effect of physiological impulse activity in the axons, we stimulated them at either 50 or 100 Hz. Are these frequencies representative of physiological levels of activity? Axons of different modalities normally fire at different frequencies, and therefore no single frequency can be considered “physiological” for all groups of axons. However, some axons, such as the primary afferents from muscle spindles, are normally active for protracted periods of time, and recordings from hamstring afferents in freely moving preparations have revealed average firing rates of 100 Hz or more during normal stepping.24 For these axons, at least, the 100-Hz stimulation imposed during NO exposure in the present study may be analogous to a 2-hour walk. The fact that 50-Hz stimulation also resulted in degeneration implies that our findings may be relevant to a range of various types of nerve fiber in patients. Indeed, the level of axonal impulse activity can be significantly raised in patients with demyelinating dis- Smith et al: Impulse Activity, NO, and Degeneration 473 Fig 2. (A) Two series of averaged (n ⫽ 64) compound action potentials recorded in parallel from 2 dorsal roots in a rat (female, Lewis) using the arrangement indicated (see inset). The data are illustrated in the same way as in Fig 1, but here each plot shows approximately 12 hours of recorded data. The left plot shows records obtained at continuous 1-Hz stimulation. 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 NO at 10-M concentration. This block was released upon removal of the NO, and the axons continued to conduct for the remaining 7.5 hours of the experiment. In contrast, the right plot shows the effect of the same exposure to NO in axons stimulated continuously at 100 Hz for the first 6 hours of recording. The onset of NO exposure was marked by the transient emergence of an early peak, representing conduction in the fastest axons within the root. The emergence of this peak was permitted when the contribution of the fastest axons to the compound potential was released from phase cancellation when conduction in the medium and slowly conducting fibers became blocked. Eventually, conduction in the fastest axons was blocked, and all the axons remained blocked for the duration of exposure to NO. Conduction was restored to some axons upon removal of the NO, but the recovery was only temporary and after approximately 2 hours a total conduction block ensued. This block was persistent, and there was no recovery of function, even when the frequency of stimulation was reduced after 6 hours to only 1 Hz. (B) Histological examination of the roots illustrated in A at the end of the recording period revealed that the root stimulated at only 1 Hz during the period of NO exposure (left) was “quite” or “relatively” normal in appearance. The integrity of the axons and their myelin sheaths was maintained, nodal integrity was maintained (arrows), and the Schmidt-Lanterman incisures were closed; nodes and incisures are normally sensitive indicators of fiber pathology.38 In contrast, in the root stimulated at 100 Hz during exposure to NO (right), almost all fibers were undergoing ovoid formation and axonolysis, changes characteristic of acute wallerian-type degeneration. Thionin acridine orange stain. ease,25 because “spontaneous” ectopic activity is generated at the lesion in axons rendered hyperexcitable by their response to demyelination. In experimental animals, demyelinated axons can generate continuous trains of impulses at frequencies of up to 45 Hz that persist for many hours (⬎10 hr)26 and conduct away from the lesion in both directions along the axon.27–29 The addition of such trains to the normal impulse traf- 474 Annals of Neurology Vol 49 No 4 April 2001 fic may significantly increase the impulse load of individual axons. It is therefore possible that pathological axons may themselves generate impulse loads that are sufficient to cause their degeneration within inflamed tissue. In a multifocal disease such as MS, it is possible that the impulses could be generated at a site of demyelination but cause degeneration at a (possibly remote) site of inflammation. These experiments have sought to examine the effects on axons of exposure to NO at concentrations representative of those that may be expected at sites of inflammation. The precise concentration of NO within neuroinflammatory lesions is not known, but NO concentrations of up to 100 nM occur in electrically stimulated cerebellar slices,30 and concentrations within the central nervous system of animals with experimental autoimmune encephalomyelitis may be up to 30 times greater than in control animals.31 A steady-state concentration of 1 m has been reported in cultures of iNOS-expressing astrocytes at only 1/100 of their density within the brain,17 and a maximum concentration of 4 M has been observed in the rat brain following transient occlusion of the middle cerebral artery.32 Much higher concentrations have been described in an inflammatory pleural exudate.33 Therefore, we have examined the effects of NO at the low micromolar concentrations that may be anticipated at sites of inflammation within the brain. We have not examined the mechanisms responsible for inducing these changes. However, it has long been assumed that axonal breakdown is the result of a block in energy-dependent calcium-excluding mechanisms, producing an increase in the intra-axonal ionic calcium level that activates the calcium-sensitive protease calpain II. In addition to the usual routes of calcium entry, we suggest that, in the experiments we have described, calcium may also enter axons as an indirect consequence of a rise in the intra-axonal sodium level. Low micromolar concentrations of NO impair mitochondrial metabolism,16,34 and so the extrusion of sodium ions that have entered axons during sustained impulse activity will also be impaired owing to an insufficient ATP concentration. An elevated intra-axonal sodium level is known to cause the Na⫹-Ca2⫹ exchanger to run in reverse, importing calcium to the axon.35,36 Because the constitutive mechanisms of calcium sequestration and elimination are impaired by the action of NO on mitochondria,16,17,37 calcium may be expected to accumulate intra-axonally, leading to the activation of various degradative proteases and phospholipases, resulting in granular disintegration of the cytoskeleton and ultimately to wallerian degeneration. In summary, we have produced a model of inflammatory neurological disease that contains two components that may be expected to occur within lesions in patients: a raised concentration of NO and sustained impulse activity in the axons. We suggest that the combination of an increased metabolic demand resulting from sustained impulse activity and a diminished metabolic capacity resulting from NO exposure may result in loss of ionic homeostasis and a calcium-mediated degeneration of the axonal cytoskeleton. We further suggest that a similar mechanism may contribute to axonal loss in inflammatory neurological disease. In cer- tain pathways, such as motor tracts, axonal loss may result in a compensatory increase in the impulse traffic in remaining axons, which would accelerate their degeneration. If this is the case, then a need for early treatment may be highlighted. Measures to reduce NO production within lesions or to reduce the sodium and calcium load imposed on active axons may provide an effective strategy for axonal protection and thereby prevent persistent neurological disability in patients. This study was supported by the Multiple Sclerosis Society of Great Britain and Northern Ireland, the Medical Research Council, SmithKline Beecham, and Guy’s and St Thomas’ Charitable Foundation. We thank Stephen Asquith and Matthew Purcell for their expert technical assistance. References 1. Feasby TE, Hahn AF, Brown WF, et al. Severe axonal degeneration in acute Guillain-Barré syndrome: evidence of two different mechanisms? J Neurol Sci 1993;116:185–192. 2. Scolding N, Franklin R. Axon loss in multiple sclerosis. Lancet 1998;352:340 –341. 3. Prineas JW. Pathology of inflammatory demyelinating neuropathies. Bailliere’s Clin Neurol 1994;3:1–24. 4. Antel J. Multiple sclerosis: emerging concepts of disease pathogenesis. J Neuroimmunol 1999;98:45– 48. 5. Coles AJ, Wing MG, Molyneux P, et al. Monoclonal antibody treatment exposes three mechanisms underlying the clinical course of multiple sclerosis. Ann Neurol 1999;46:296 –304. 6. Davie CA, Silver NC, Barker GJ, et al. Does the extent of axonal loss and demyelination from chronic lesions in multiple sclerosis correlate with the clinical subgroup? J Neurol Neurosurg Psychiatry 1999;67:710 –715. 7. Trapp BD, Bo L, Mork S, Chang A. Pathogenesis of tissue injury in MS lesions. J Neuroimmunol 1999;98:49 –56. 8. Trapp BD, Peterson J, Ransohoff RM, et al. Axonal transection in the lesions of multiple sclerosis. N Engl J Med 1998;338: 278 –285. 9. Kornek B, Lassmann H. Axonal pathology in multiple sclerosis: a historical note. Brain Pathol 1999;9:651– 656. 10. Bo L, Dawson TM, Wesselingh S, et al. Induction of nitric oxide synthase in demyelinating regions of multiple sclerosis brains. Ann Neurol 1994;36:778 –786. 11. Cross AH, Manning PT, Keeling RM, et al. Peroxynitrite formation within the central nervous system in active multiple sclerosis. J Neuroimmunol 1998;88:45–56. 12. Giovannoni G, Heales SJ, Silver NC, et al. Raised serum nitrate and nitrite levels in patients with multiple sclerosis. J Neurol Sci 1997;145:77– 81. 13. Johnson AW, Land JM, Thompson EJ, et al. Evidence for increased nitric oxide production in multiple sclerosis. J Neurol Neurosurg Psychiatry 1995;58:107. 14. Yamashita T, Ando Y, Obayashi K, et al. Changes in nitrite and nitrate (NO2⫺/NO3⫺) levels in cerebrospinal fluid of patients with multiple sclerosis. J Neurol Sci 1997;153:32–34. 15. Bagasra O, Michaels FH, Zheng YM, et al. Activation of the inducible form of nitric oxide synthase in the brains of patients with multiple sclerosis. Proc Natl Acad Sci USA 1995;92: 12041–12045. 16. Bolanos JP, Almeida A, Stewart V, et al. Nitric oxide-mediated mitochondrial damage in the brain: mechanisms and implica- Smith et al: Impulse Activity, NO, and Degeneration 475 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 476 tions for neurodegenerative diseases. J Neurochem 1997;68: 2227–2240. Brown GC, Bolanos JP, Heales SJ, Clark JB. Nitric oxide produced by activated astrocytes rapidly and reversibly inhibits cellular respiration. Neurosci Lett 1995;193:201–204. Zielasek J, Reichmann H, Kunzig H, et al. Inhibition of brain macrophage/microgial respiratory chain enzyme activity in experimental autoimmune encephalomyelitis of the Lewis rat. Neurosci Lett 1995;184:129 –132. Redford EJ, Kapoor R, Smith KJ. Nitric oxide donors reversibly block axonal conduction: demyelinated axons are especially susceptible. Brain 1997;120:2149 –2157. Shrager P, Custer AW, Kazarinova K, et al. Nerve conduction block by nitric oxide that is mediated by the axonal environment. J Neurophysiol 1998;79:529 –536. Ganter P, Prince C, Esiri MM. Spinal cord axonal loss in multiple sclerosis: a post-mortem study. Neuropathol Appl Neurobiol 1999;25:459 – 467. Lovas G, Szilagyi N, Majtenyi K, et al. Axonal changes in chronic demyelinated cervical spinal cord plaques. Brain 2000; 123:308 –317. Trapp BD, Peterson J, Ransohoff RM, et al. Axonal transection in the lesions of multiple sclerosis. N Engl J Med 1998;338: 278 –285. Prochazka A, Gorassini M. Models of ensemble firing of muscle spindle afferents recorded during normal locomotion in cats. J Physiol (Lond) 1998;507:277–291. Smith KJ, Felts PA, Kapoor R. Axonal hyperexcitability: mechanisms and role in symptom production in demyelinating diseases. Neuroscientist 1997;3:237–246. Smith KJ, McDonald WI. Spontaneous and evoked electrical discharges from a central demyelinating lesion. J Neurol Sci 1982;55:39 – 47. Smith KJ, McDonald WI. Spontaneous and mechanically evoked activity due to central demyelinating lesion. Nature 1980;286:154 –155. Annals of Neurology Vol 49 No 4 April 2001 28. Baker M, Bostock H. Ectopic activity in demyelinated spinal root axons of the rat. J Physiol (Lond) 1992;451:539 –552. 29. Kapoor R, Li YG, Smith KJ. Slow sodium-dependent potential oscillations contribute to ectopic firing in mammalian demyelinated axons. Brain 1997;120:647– 652. 30. Shibuki K, Okada D. Endogenous nitric oxide release required for long-term synaptic depression in the cerebellum. Nature 1991;349:326 –328. 31. Hooper DC, Ohnishi ST, Kean R, et al. Local nitric oxide production in viral and autoimmune diseases of the central nervous system. Proc Natl Acad Sci USA 1995;92:5312–5316. 32. Malinski T, Bailey F, Zhang ZG, Chopp M. Nitric oxide measured by a porphyrinic microsensor in rat brain after transient middle cerebral artery occlusion. J Cereb Blood Flow Metab 1993;13:355–358. 33. Regnault C, Roch-Arveiller M, Florentin I, et al. Kinetic evaluation of nitric oxide production in pleural exudate after induction of two inflammatory reactions in the rat. Inflammation 1996;20:613– 622. 34. Borutaite V, Brown GC. Rapid reduction of nitric oxide by mitochondria, and reversible inhibition of mitochondrial respiration by nitric oxide. Biochem J 1996;315:295–299. 35. LoPachin RM, Lehning EJ. Mechanism of calcium entry during axon injury and degeneration. Toxicol Appl Pharmacol 1997; 143:233–244. 36. Stys PK. Anoxic and ischemic injury of myelinated axons in CNS white matter: from mechanistic concepts to therapeutics. J Cereb Blood Flow Metab 1998;18:2–25. 37. Brorson JR, Sulit RA, Zhang H. Nitric oxide disrupts Ca2⫹ homeostasis in hippocampal neurons. J Neurochem 1997;68: 95–105. 38. Williams PL, Hall SM. Prolonged in vivo observations of normal peripheral nerve fibres and their acute reactions to crush and deliberate trauma. J Anat 1971;108:397– 408.