Journal o f Neurocytology 7 , 2 6 5 - 2 8 2 (1978)
Peripheral nervous system pathology in relapsing experimental allergic encephalomyelitis R. E. M A D R I D
1 and H. M. W I S N I E W S K I
2
1M.R.C. Demyelinating Diseases Unit, Newcastle-upon-Tyne, England z New York State Institute for Basic Research in Mental Retardation, Staten Island, New York, U.S.A.
Received 22 July 1977; revised 5 October and 16 November 1977; accepted 18 November 1977
Summary A recurrent type of primary demyelination was found in the anterior roots and dorsal root ganglia of strain 13 guinea-pigs with relapsing experimental allergic encephalomyelitis. Single nerve fibre studies revealed a predominant n o d a l - p a r a n o d a l distribution of the destruction of myelin in these nerves. Damage of the Schwann c e l l - a x o l e m m a l junction was found in the majority of the abnormal nodes of Ranvier. This was accompanied by the formation-of supernumerary adaxonal Schwann cell processes which further altered the normal myelinating c e l l - a x o n relationship. It is concluded that variations in antigenic composition may play a role in the selective involvement of the anterior roots. The pathogenesis of the observed nodal changes is discussed.
Introduction Multiple sclerosis (MS) is a disease characterised by C.N.S. myelin pathology in which recent and old demyelinating lesions usually coexist (Dawson, 1915). In the search for the cause of this disease an animal model has been developed which bears much closer similarities to the human disease than acute experimental allergic encephalomyelitis (EAE) (Stone and Lerner, 1965; Snyder et al., 1975;Wis'niewski and Keith, 1977). The observed changes in the P.N.S. in MS (Schob, 1915) have, up to now, been considered to be non-specific (Hasson et al., 1958) or merely to reflect changes in the glial portion of the spinal roots (see Dawson, 1915). However, recent studies (Calder and Pollock, 1976; Schoene et al., 1976) suggest that the P.N.S. may itself be abnormal in MS. A consistent involvement of the P.N.S. was not observed in previous studies on chronic EAE (Raine et al., 1974), although Schwann cell invasion of the spinal cord was frequently found in long-term animals (Snyder et al., 1975). The latter suggested to us the presence of a concomitant P.N.S. patholog3z. The present report is part of a comprehensive study on the ultrastructural pathology of the recurrent EAE model developed in this laboratory (Wigniewski and 9 1978 Chapman and Hall Ltd. Printed in Great Britain
266
MADRID and WISNIEWSKI
Keith, 1977) and focuses on the P.N.S. lesions, a h i t h e r t o u n r e c o g n i z e d aspect o f the disease process.
Materials and m e t h o d s Sixteen juvenile Strain-13 guinea-pigs were injected in the dorsum of both hind feet with a total of 0.2 ml of an emulsion containing equal volumes of a homogenate of guinea-pig spinal cord (1:1, w/v in saline) and Freund's complete adjuvant containing 10 mg/ml tubercle bacilli (H37Ra). Animals were periodically assessed for the presence of clinical signs which were evaluated under a scoring system described elsewhere (Eastman et al., 1978). At different times after the first, second and third clinical relapses (see Table 1) the animals were heparinized and killed by intracardiac perfusion fixation with 4% glutaraldehyde in 0.1 M phosphate buffer, pH 7.4 for 15 min. Immediately after perfusion, the spinal cord (SC) and associated dorsal root ganglia (DRG) and roots were dissected out. With the aid of a dissecting microscope, the dura was excised and the L7 and $1 spinal roots and DRG removed. Both anterior and posterior roots were divided into 3 segments in which the middle portion of each, about 1.5 cm in length, was prepared for teasing_ in fluid Epon 812, as previously described (Madrid and Wigniewski, 1977). Samples from the following regions were taken: the remaining 3 mm segments on each end of the roots (near the spinal cord and near the DRG)~ L7 and Sx spinal cord with corresponding DRG; and T~ 0 - T 12 spinal cord. Osmication was carried out in 1% Dalton's chrome osmium solution for 2 h. Specimens were then washed briefly in 0.1 M phosphate buffer with sucrose, dehydrated in graded concentrations of ethanol, cleared in propylene oxide, and embedded in Epon. 1/Jm thick sections from all the specimens mentioned above were cut in an LKB,III ultramicrotome, stained with toluidine blue and studied by light microscopy. From these blocks, selected areas were trimmed and thin sections produced with a diamond knife. Thin sections were collected on copper grids without support film and contrasted with ethanolic uranyl acetate and lead citrate. They were examined with Philips 201 or 300 electron microscopes operating at 60 KV. Approximately 500 nerve fibres from both anterior and posterior L7 and $1 spinal roots were teased out in each of the experimental animals. Selected fibres were transferred to toluene-cleaned glass slides and mounted in Epon for photographic purposes (Spencer and Thomas, 1970). Fibres of special interest were mounted on previously hardened Epon blocks and subsequently oriented for transverse or longitudinal thin sectioning (Ochoa, 1972).
Results T h e n u m b e r o f clinical episodes, the distribution o f the lesions in the p r o x i m a l P.N.S. and in the spinal cord, and the times after i m m u n i z a t i o n w h e n the guinea-pigs were killed, are shown in T a b l e 1. No positive correlation was f o u n d b e t w e e n the location a n d / o r the e x t e n t o f the lesions in the L 7 and S 1 spinal cord and t h o s e in the spinal roots. D e m y e l i n a t i n g plaques were m o r e extensive in the lower t h o r a c i c t h a n in the lumbosacral spinal cord in these long-term animals. In animals e x a m i n e d at intervals after the first and second clinical episode, the spinal r o o t s s h o w e d a small percentage of nerve fibres (usually less than 4%) u n d e r g o i n g axonal degeneration, with no evidence of p r i m a r y d e m y e l i n a t i o n and r e m y e l i n a t i o n . A characteristic p a t t e r n o f i n v o l v e m e n t o f the P.N.S. emerged o n l y after the third clinical episode, or when, after a series o f relapses, the animal had e n t e r e d a c h r o n i c progressive course of neurological deficit. T h e P.N.S. lesions were restricted exclusively t o the a n t e r i o r
267
P.N.S. P a t h o l o g y in relapsing EAE Table 1. Distribution of the lesions in the P.N.S. of inbred guinea-pigs with chronic relapsing-EAE. L 7 S1 ventral roots
GP number
Number of clinical episodes
Days postimmunization (DPI) at time of sacrifice
Near spinal cord
Teased fibres
Near dorsal root ganglia
2781/6 1805/6
1 1
105 107
(-) (-)
(--) (-)
(-) (--)
1897/6 1819/6 2196/6 1910/6
2 2 2 2
60 74 83 84
(+) (+) (--) (-)
(--) (--) (--) (--)
(--) (--) (-) (+)OB
1815/6 1807/6 1953/6 1809/6 1806/6 1027/6
3 3 3 3 3P 3 3 4 4 P
(+)OB (+)OB,MB (+)MB (+)OB,MB (+)OB (+)OB (+)OB (+)OB, MB (+)OB,MB (+)OB,MB
(-) (-) (+) (+) (+) (+) (+) (+) (+) (+)
(-) (-) ( ) ( ) (+)OB (+)OB (+)O B, MB (+) (+) (+)
1177/5
907/5 714/5 1006/5
112 156 197 200 213 223 370 240 635 383
No attempt was made at scoring the histological changes in the present series; GP, guineapig; P, animals showing a progressive type of neurological deficit after the third clinical episode (1806) or from the beginning (1006/5); (+) denotes the presence of demyelination and/or remyelination of P.N.S. fibres at the locations indicated; OB indicates the presence of onion-bulb formations; MB denotes the finding of paranodal ballooning of the myelin sheath; (-) denotes cases with no evidence of de- or remyelination of P.N.S. fibres.
roots at various points in their subarachnoid course. Posterior root changes were f o u n d only in the DRG. In spite of extensive sampling, massive aggregates of i n f l a m m a t o r y ceils (as seen in the dorsal roots in acute EAN in the guinea-pig) could n o t be found. However, teasing of the anterior roots revealed the presence of fibres undergoing paranodal, or sometimes partial internodal d e m y e l i n a t i o n . These were easily identified b y the close apposition of myelin-laden macrophages to the bare segments of the axon (Fig. 1). The coexistence of d e m y e l i n a t i o n with varying degrees of r e m y e l i n a t i o n was d e m o n s t r a t e d in the same nerve fibre b y the variations in sheath thickness f r o m one segment to another. In addition, striking nodal and paranodal abnormalities were noticed in nerve fibres with otherwise normally m y e l i n a t e d internodes. The latter consisted of: (a) paranodal 'repair' and paranodal r e m y e l i n a t i o n (Fig. 2); (b) widening of the nodal gap; ( c ) t h e so-called 'paranodal
268
MADRID and W I S N I E W S K I
overriding' of the myelin sheath (Ballin and Thomas, 1969; King et al., 1975); and (d) myelin 'ballooning'. Nerve fibres in advanced stages of axonal degeneration were also occasionally observed. The uhrastructural features of nerve fibres undergoing paranodal demyelination will be described elsewhere. Onion bulb formations were often found in the anterior root exit zone (Figs. 3 - 7 ) and the anterior root itself (see Table 1)
Paranodal 'repair' and paranodal remyelination Thinning of the myelin sheath, assumed to be associated with the repair of partial damage of the paranodal myelin sheath (Ailt, 1969) was frequently found in 2 or more paranodes of a single nerve fibre (Fig. 2). On occasions, this was an isolated feature through 6 or more otherwise normal internodes, but more frequently, this paranodal thinning was seen in nerve fibres also showing short intercalated internodes (Fig. 2). The latter are also known to occur during repair of localized paranodal demyelination (Ochoa et al., 1972). Paranodal thinning of the myelin sheath was extensive, often starting 50 #m or more away from the nodal region (Fig. 2). Short intercalated internodes circa 50/am in length were also a common finding. These light microscopical indications of nodal damage and repair were confirmed by electron microscopy. In fibres displaying paranodal thinning of the myelin sheath Fig. 1. Teased fibre appearance of recurrent demyelination and remyelination. 240 days post-immunization (DPI). Segmental loss of myelin and paranodal concentration of myelin-laden macrophages (row 4) coexist with short remyelinated internodes of varying myelin thickness in the same nerve fibre. Fig. 2. Teased fibre from the same animal shown in Figure 1. A short (62/lm) internode (ic) is intercalated between segments with normal myelin thickness. Paranodal thinning of the myelin sheath (t) is seen in the next 2 internodes. Fig. 3. L7 spinal cord, 370 DPI. Low power electron micrograph of the anterior root exit zone. Concentric arrangements of attenuated Schwann and pial cell processes surround the majority of the P.N.S. fibres in this area. The anterior column is to the upper left in this figure. Fig. 4. Detail of the surrounding Schwann (S) and pial processes (P) from Figure 3. The latter are not covered by a basal lamina. Fig. 5. Extremely attenuated Schwann cell processes from the lower right onion bulb in Figure 3. Horizontally running collagen fibres are also evident (c). Fig. 6. Abaxonal origin of an attenuated Schwann profile. Fig. 7. A Schwann cell process from Figure 3 showing neurofilamentous hyperplasia. Note encompassing basal lamina (bl). Fig. 8. Paranodal thinning of the myelin sheath. The area of transition from thick to thin myelin is indicated (arrows). Fig. 9. Detail from figure 8. At least 5 reversing myelin lamellae are apposed to the axolemma. Fig. 10. Teased anterior root nerve fibre showing a 47/lm long intercalated internode. Compare with Figure 2. Fig. 11. Detail of lower node of Ranvier from the previous figure. The internode with normal myelin thickness shows dense terminal loops which are separated from the axolemma by interposed processes of Schwann cell origin (arrows). A mononuclear cell (M) is seen in the nearby endoneurial space.
272
MADRID and WISNIEWSKI
the area o f transition b e t w e e n thick and thin m y e l i n revealed an irregular t e r m i n a t i o n o f m y e l i n lateral loops some distance away f r o m the n o d e o f Ranvier. M a n y terminal loops failed to m a k e c o n t a c t with the a x o l e m m a while o t h e r s were apposed to the a x o n as reversing loops (Figs. 8 and 9). In contrast, the lateral loops o f the i n t e r n o d e s with n o r m a l m y e l i n thickness o n b o t h sides o f an intercalated segment shows a peculiarly increased e l e c t r o n d e n s i t y , and were generally smaller t h a n n o r m a l terminal loops (Fig. 10). In addition, these a b n o r m a l terminal loops were separated f r o m the a x o l e m m a b y interdigitating processes o f S c h w a n n cell origin which had emerged f r o m the a d a x o n a l region at varying distances f r o m the n o d e o f Ranvier (Fig. 11). These were long, flat processes which lacked the familiar club-shaped a p p e a r a n c e o f m y e l i n terminal loops. U p o n arrival at the node, these adaxonal Schwann cell processes a p p e a r e d to m o v e a w a y f r o m the a x o n m e m b r a n e t o w a r d the e n d o n e u r i a l space. Schwann cell microvilli were missing in these transitional nodes o f Ranvier, indicating that in spite o f advanced p a r a n o d a l r e m y e l i n a t i o n a n o r m a l c o n t a c t b e t w e e n the 2 S c h w a n n cells had n o t b e e n reestablished. M o n o n u c l e a r cells were f o u n d in the n e a r b y e n d o n e u r i a l space (Fig. 11). T h e m y e l i n terminal loops o f the intercalated segment a p p e a r e d normal.
Nodal widening Widening o f the n o d e o f Ranvier was a c o n s p i c u o u s f e a t u r e o f r e m y e l i n a t i n g m o t o r axons in the anterior r o o t s (Figs. 12 a n d 13). Occasionally, a s u p e r n u m e r a r y Fig. 12. 370 DPI. Nodal widening at the junction of 2 remyelinated segments. The 12/2m long gap (arrows) is filled here by a Schwann cell. Normal terminal loops are evident in the remyelinated internode on the right but a few myelin lamellae are seen ending beyond these terminal loops (right arrow). Fig. 13. A 29 ;lm long nodal gap between remyelinated internodes of a teased anterior root fibre. x 750. Fig. 14. Skip-serial thin section of the upper half of the nodal region from Fig. 12. The Schwann cell nucleus is now in the plane of the section. Note tortuous course of some of the adjacent myelin lamellae (right). The length of axolemma covered by the supernumerary Schwann cell is indicated (arrows). Fig. 15. Nodal widening at the junction of internodes with normal myelin thickness. 223 DPI. On both sides of the node of Ranvier the myelin sheath tapers off. The perinodal space is almost saturated with cellular processes. 1/2m toluidine blue section, x 750. Fig. 16. Ultrastructural appearance of a similarly damaged nerve fibre, 240 DPI. Dense terminal loops separated from the axolemma by interdigitating Schwann cell processes are seen on both myelinated segments (arrows). Figs. 17 and 18. High power view of the zones arrowed on the bottom half of the right paranodal segment in the previous figure. Fig. 17. A narrow Schwann cell process emerges from one of the major dense lines. Fig. 18. A seemingly independent process is seen at this level. Note the gradual change in size and electron density of the terminal loops as they approach the nodal gap. a, axolemma. Scale: 0.1 gtm. Fig. 19. Detail from a widened node similar to those shown in Figs. 15 and 16. A fine Schwann cell process extends from the paranodal area toward the node of Ranvier (arrows) and intervenes between the dense terminal loops and the axolemma. Accumulation of mitochondria (M) and the base of a narrow process (p) are seen in the abaxonal Schwann cell cytoplasm.
274
MADRID and WISNIEWSKI
Schwann cell was seen filling the widening gap (Fig. 14). Usually, one or b o t h remyelinated segments showed focal or generalized a b n o r m a l i t y in the a b u t m e n t o f myelin terminal loops, b u t the m o r p h o l o g y of the terminal loops was normal (Fig. 14). Nodal widening was also f o u n d in otherwise normal internodes (Fig. 15) usually accompanied b y a characteristic tapering of the paranodal m y e l i n sheath. A profusion of Schwann and m o n o n u c l e a r cell processes usually surrounded these abnormal nodal regions (Fig. 16). The latter were characterized b y the absence of a basal lamina. Electron dense terminal loops made up the regular ends of m o s t of the myelin lamellae at these nodes. A few 'clear' terminal loops closely associated with elongated Schwann cell processes were usually f o u n d at points farthest f r o m the node. These narrow Schwann cell processes spread t o w a r d the nodal gap, separating the electron dense terminal loops f r o m the a x o l e m m a (Figs. 17 and 18). A l t h o u g h the Schwann cell origin of m a n y of the elongated cellular processes could easily be d e m o n s t r a t e d (Fig. 19), some profiles appeared to be i n d e p e n d e n t of the p a r e n t myelinating cell. Myelin terminal loops which were in close c o n t a c t w i t h the latter showed an increased electron density. Schwann cell microvilli were absent in these abnormal nodes of Ranvier indicating t h a t d e m y e l i n a t i o n had previously occurred at this level.
Fig. 20. 1 #m toluidine blue section of a teased anterior root fibre showing paranodal overriding, 223 DPI. A number of cells and their processes are seen in the adaxonal space of the myelinated segment. Demyelination has occurred on the right and a myelin-laden macrophage is still in the area (md). (R) denotes the site of the former node of Ranvier. Fig. 21. Serial thin section of the nerve fibre from Fig. 20. The axon is now out of the plane of section. Schwann (S) and mononuclear cells (M) and their processes (p) are present inside the myelin sheath. The (*) identifies the cells and processes shown at higher magnification in Figs. 22-24. Fig. 22. A minute Schwann cell with its basal lamina (bl) and a mononuclear celt process (mp) are seen in this area. Fig. 23. Mononuclear cell process (mp) near the previously adaxonal portion of the myelin sheath (m). 2 Schwann cell processes covered by a basal lamina are also seen. Fig. 24. Detail of a mononuclear cell process from the bottom half of Fig. 20, which is in contact with the inner aspect of the myelin sheath. Fig. 25. 223 DPI. Schwann and mononuclear cell infiltration around a myelinated fibre over a length of at least 100/~m. A gradual thinning of the outer myelin sheath occurs from left to right. The arrows point to the areas in which vesicular disruption of single myelin lamellae was observed. These are shown at higher magnification in Figs. 28 and 29. The location of the original node of Ranvier is indicated (R). Fig. 26. Nodal region from Fig. 25. A mononuclear cell process is in contact with the terminal loops some of which are undergoing vesiculation. Schwann profiles intervene between the inner (im) and outer (om) myelin sheaths. Fig. 27. 'Collagen pockets' in the extracellular gap of the myelinated fibre from Fig. 25. A mononuclear cell process is apposed to the outer myelin sheath. Fig. 28 and 29. Detail of the vesicular disruption of myelin lamellae from the outer myelin sheath in Fig. 25.
276
MADRID and W I S N I E W S K I
Paranodal overriding of the myelin sheath Longitudinal sections of teased anterior root fibres proved particularly informative in the study of the mechanism of formation of these bizarre Schwann cell-axon relationships. For example, Fig. 20 shows an axon partially encompassed by Schwann cell processes but not yet remyelinated. A combination of Schwann and mononuclear cells were found inside the myelin sheath (Figs. 21--24). A gradual attenuation in the thickness of the outer, 'overriding', myelin sheath was constantly found in this material. In the area of transition from thick to thin outer myelin sheath, a node of Ranvier or a Schmidt-Lanterman cleft was found. In the former, grossly enlarged processes of Schwann cell origin were in contact with the axolemma, which lifted away the dense myelin loops at that level. Invariably, mononuclear cells or their processes showed a tendency to lie in contact with the innermost myelin lamellae of the 'overriding' myelin sheath (Figs. 23 and 24). In well-developed cases, the space between the 2 myelin sheaths (the periaxonal space of the parent and now outer Schwann cell) was filled with minute, uncommitted Schwann cells, mononuclear cells, newly formed collagan fibrils and flattened Schwann cell processes (Fig. 25). Electron dense terminal loops, some of which were undergoing vesicular disruption, were frequently seen in this 'outer' myelin sheath in the presence of mononuclear cell processes for considerable distances from the node of Ranvier (Figs. 25, 28 and 29). This progressive peeling-off and vesiculation of the inner aspect of the 'overriding' myelin sheath, occurring in the presence of these processes, seemed to be the main cause of the gradual thinning of the paranodal myelin in these areas. No morphological indication of active growth or extension of the myelin sheath could be found at the level of the original node of Ranvier (Fig. 26). On the contrary, the lateral endings of the myelin sheath sometimes seemed to turn backward instead of projecting forward. Vesicular disruption near mononuclear cell processes also occurred at this level (Fig. 26). Very thin Schwann cell processes surrounded by a basal lamina were frequently seen emerging from the external surface of the original Schwann cell at its paranodal level, but these did not usually come into contact with each other. However in the dorsal root ganglia these processes were occasionally juxtaposed. All these examples of 'paranodal overriding' took place at the junction of internodes which had a myelin sheath of normal thickness. On occasions, however, this abnormality was detected in the nodal region of remyelinated segments of a nerve fibre.
Myelin ballooning Bizarre and extensive paranodal dilatations of the myelin sheath were frequently observed adjacent to a demyelinated segment of the axon in anterior root fibres.
Fig. 30. Single teased anterior root nerve fibre showing paranodal ballooning of the myelin sheath in the vicinity of a demyelinated segment. 1 0 - 2 0 / 2 m step serial 1 /lm Epon sections taken from the sectors indicated in the first row.
278
MADRID and WISNIEWSKI
These fibres were particularly difficult to tease apart, due to their frequent close relationship with endoneurial blood vessels, and to a peculiar 'stickiness' which firmly attached these myelin 'balloons' to neighbouring myelinated fibres. Skipserial transverse sections of the demyelinated area and the dilated portion of the myelin sheath showed that the axon lay outside the myelin bleb formation for a considerable distance (Fig. 30). A thin myelin sheath surrounded the axon at the level of its penetration into the dilated myelin sheath. Macrophages were seen at the level of penetration of the axon into the bleb formation but not inside the myelin balloon. Ultrastructural examination (levels 9 - 1 1 in Fig. 30) showed that there was no basal lamina over the thin myelin covering this axon, indicating that the thin adaxonal and the dilated myelin sheaths were derived from the same Schwann cell. Discussion
Although P.N.S. lesions have occasionally been reported in acute EAE in several species, the distribution of the lesions in the proximal P.N.S. has seldom been the subject of a detailed morphological analysis. Waksman and Adams (1956) found posterior root entry zone lesions in only 13% of 30 guinea-pigs immunized with C.N.S. tissue. They showed that the anterior roots were not significantly involved, in contrast to studies on EAE in rabbits (Wi~niewski et al., 1969). The P.N.S. lesions of the recurrent-EAE animals (R-EAE) described here: (a) showed a remarkable nodal-paranodal distribution; (b) occurred in the absence of perivascular aggregates of inflammatory Cells; and (c) were limited to the anterior roots and DRG. Some of the findings of this study are similar to those of King et al. (1975) in relapsing experimental allergic neuritis (EAN) in outbred guinea-pigs. These include the paranodal thinning of the myelin sheath and paranodal overriding of myelin lamellae. However, significant differences are found in the configuration of the lesions. Firstly, the paranodal attenuation of the myelin sheath was frequently observed in otherwise normally myelinated internodes as shown by single teased fibre preparations. King et aL (1975) showed paranodal thinning of remyelinated axons. It is possible that the latter was the consequence of unfinished lengthening of a recently formed myelin sheath while a partial repair and/or re-apposition of a partially damaged myelin sheath seems more likely in the present study. Secondly, the extracellular gap in the examples of paranodal overriding reported by King et al. (1975) appeared almost empty while it was filled with a number of Schwann and mononuclear cells in our material. A gradual attenuation of the outer myelin sheath in the presence of mononuclear cell processes was always found in this study but was not reported by those authors. 'Perinodal' overriding as described by King et aL (1975) was not observed by us. Therefore we conclude that in contrast to the situation reported in the relapsing EAN animals, paranodal 'overriding' in this material was due to Schwann and mononuclear cell infiltration of the adaxonal space. Whether or not these differences are due to our using inbred guinea-pigs immunized with a single injection of homologous C.N.S. tissue is not clear. That
P.N.S. pathology in relapsing EAE
279
presumably additional mechanisms of myelin damage are operating in this animal model is suggested by the frequent occurrence of nodal widening, a lesion which was not detected by King et al. (1975). Unlike the C.N.S. lesions found in these animals, the observed P.N.S. myelin pathology occurred in the absence of perivascular aggregates of inflammatory cells. That this could merely reflect a faster rate of clearance of inflammatory cells from the endoneurial space of the anterior roots is supported by a similar finding in relapsing EAN animals (King et al., 1975). The participation of inflammatory cells in the pathogenesis of the described myelin lesions was demonstrated here by single fibre microdissection. Teased fibre preparations also proved the recurrent nature of the P.N.S. disease process by disclosing the coexistence of varTing degrees of both paranodal demyelination and remyelination in consecutive internodal lengths of the same nerve fibre. Since the myelin balloons described here are in all respects similar to those previously described by Spencer and Thomas (1970), they are also evidence for demyelination of already remyelinated segments of those fibres. In contrast to a recent report (Snyder et al., 1977) our results indicate that 'onion bulbs' were the consequence of a recurrent process of demyelination and remyelination, rather than a single episode of demyelination. The restriction of P.N.S. lesions to the anterior roots and DRG in R-EAE agrees with the finding of Pollard et al. (1975) in outbred guinea-pigs induced to relapse by repeated injections of P.N.S. tissue. Acute EAN is known to be a predominantly posterior root disease (Waksman and Adams, 1956; Allt, 1976; Madrid and Wi~niewski, 1977). The reason for the shift in the location of the pathological changes in relapsing EAN (Pollard et al., 1975) and for the anterior root distribution of the lesions in R-EAE is not clear. Since the lesions seen in these long-term animals increased in frequency with time after immunization, the possibility that they could be age relate d should be considered. Age-related changes in rats (Berg et al., 1962; Gilmore, 1972) and dogs (Griffiths and Duncan, 1975) usually appear near the end of their mean lifespan, are apparently demyelinating in nature and involve both anterior and posterior roots. It is unlikely that this mechanism is operating in the present situation, because throughout this study the posterior roots were not involved. Alternatively, a hitherto unrecognized difference in the antigenic composition of the anterior roots could explain their preferential involvement in this series. Studies under progress in this laboratory are aimed at elucidating this problem. It is possible that the nodal widening found at the junction of remyelinated segments was due to an unfinished process of remyelination, or to the interfering effect of intercalated Schwann cells. However, the nodal widening of internodes with normal myelin thickness could be the consequence of a direct effect of antimyelin or antiaxonal antibodies. Since nodal widening and aberrant termination of paranodal myelin loops have been reported in the C.N.S. in MS (Suzuki et al., 1969; Andrews, 1972), the present study further emphasizes the relevance of the R-EAE model to MS. Although patho-physiological studies of the abnormal nodes were not done it is
280
MADRID and WISNIEWSKI
conceivable t h a t the n o d a l - p a r a n o d a l changes described here could significantly disrupt the propagation of the nerve action potential in the anterior roots of these animals.
Acknowledgements This w o r k was supported b y the Multiple Sclerosis Society of Great Britain and N o r t h e r n Ireland (grant no. MNI-1/75). We t h a n k Mr A. B. Keith for dedicated care of the long-term animals used in this study, Miss Margaret Herron for invaluable secretarial assistance, and Dr J. R. M c D e r m o t t for reading the manuscript. References ALLT, G. (1969) Repair of segmental demyelination in peripheral nerves: An electron microscope study. Brain 92, 639-46. ALLT, G. (1976) Pathology of the peripheral nerve. In: The Peripheral Nerve (edited by LANDON, D. N.), pp. 666-740. London: Chapman and Hall. ANDREWS, J.M. (1972) The ultrastructural neuropathology of multiple sclerosis. In: Multiple Sclerosis, Immunology, Virology and Ultrastructure (edited by WOLFGRAM, F., ELLISON, G.W., STEVENS, J.G. and ANDREWS, J. M.), pp. 23--52. New York, London: Academic
Press. R.H.M. and THOMAS, P.K. (1968) Electron microscope observations on demyelination and remyelination in experimental allergic neuritis. I. Demyelination. Journal of the Neurological Sciences 8, 1-18. BALLIN, R.H.M. and THOMAS, P.K. (1969) Electron microscope observations on demyelination and remyelination in experimental allergic neuritis. 2. Remyelination. Journal of the Neurological Sciences 8, 225--37. BALLIN,
BERG, B. M., WOLF, A., SIMMS, H. S. (1962) Degenerative lesions of spinal roots and peripheral
nerves in aging rats. Gerontologia 6, 72-80. C.S. and POLLOCK, M. (1976) Peripheral nerve changes in multiple sclerosis. Proceedings of the University of Otago Medical School 54, 5-6. DAWSON, J.W. (1915) The histology of disseminated sclerosis. Transactions of the RoyalSociety of Edinburgh 50, (iii), 517--740.
CALDER,
EASTMAN, R., KEITH, A. B. and WISNIEWSKI, H.M. (1978) Chronic relapsing experimental
allergic encephalomyelitis: a clinico-pathological study. Neuropatbology and Applied Neurobiology, (submitted). GILMORE, S.A. (1972) Spinal nerve root degeneration in aging laboratory rats: a light microscopic study. Anatomical Record 174, 251-8. GRIFFITHS, I. R. and DUNCAN, I. D. (1975) Age changes in the dorsal and ventral lumbar nerve roots of dogs. Acta Neuropathologica 32, 75--85. HASSON, J., TERRY, R. D. and ZIMMERMAN, H. M. (1958) Peripheral neuropathy in multiple
sclerosis. Neurology 8,503--10. KING, R. H. M., POLLARD, J. D. and THOMAS, P. K. (1975) Aberrant remyelination in chronic
relapsing experimenta! allergic neuritis. Neuropathology and Applied Neurobiology 1, 367-78. MADRID, R. E. and WISNIEWSKI, H. M. (1977) Axonal degeneration in demyelinating disorders.
Journal of Neurocytology 6, 108-17. OCHOA, J. (1972) Ultrathin longitudinal sections of single myelinated fibres for electron microscopy. Journal of the Neurological Sciences 17, 103-6.
P.N.S. pathology in relapsing EAE
281
OCHOA, J., FOWLER, T. J. and GILLIAT, R. W. (1972) Anatomical changes in peripheral nerves compressed by a pneumatic tourniquet. Journal of A natomy 113, 433-55. POLLARD, J.D., KING, R. H. M. and THOMAS, P. K. (1975) Recurrent experimental allergic neuritis. An electron microscope study. Journal of the Neurological Sciences 24, 365-83. RAINE, C.S., SNYDER, D.H., VALSAMIS, M.P. and STONE, S.H. (1974) Chronic experimental allergic encephalomyelitis in inbred guinea pigs. An ultrastructural study. Laboratory Investigation 31, 3 6 9 - 3 8 0 . SCHOB, F. (1915) Ein beitrag zur pathologischen Anatomic der multiplen sklerose. Monatscbrift fur Psychiatric und Neurologie 22, 62-87. SCHOENE, W. C., GESCHWIND, N. O., BEHAN, P. O. and CARPENTER, S. (1976) Coexistence of multiple sclerosis and hypertrophic neuropathy: Light, electron microscopical, immunological and clinical studies. Neuropathology and Applied Neurobiology 2, 165 (abstract)~ SNYDER, D.H., VALSAMIS, M.P., STONE, S.H. and RAINE, C.S. (1975) Progressive demyelination and reparative phenomena in chronic experimental allergic encephalomyelitis. Journal of Neuropatbology and Experimental Neurology 34, 2 0 9 - 2 1 . SNYDER, D.H., STONE, S.H. and RAINE, C.S. (1977) Attempts to induce chronic experimental allergic neuritis in Strain 13 and Harttey guinea pigs. Journal of Neuropathology and Experimental Neurology 36, 4 8 8 - 9 8 . SPENCER, P. S. and THOMAS, P. K. (1970) The examination of isolated nerve fibres by light and electron microscopy with observations on demyelination proximal to neuromas. Acta Neuropatbologica 16, 177--86. STONE, S.H. and LERNER, E2 M. (1965) Chronic disseminated allergic encephalomyelitis in guinea pigs. Annals of the New York Academy of Sciences 122, 2 2 7 - 4 1 . SUZUKI, K., ANDREWS, J. M., WALTZ, J. M. and TERRY, R. D. (1969) Ultrastructural studies of multiple sclerosis. Laboratory Investigation 20, 444--54. WAKSMAN, B.H. and ADAMS, R.D. (1956) A comparative study of experimental allergic neuritis in the rabbit, guinea pig and mouse. Journal of Neuropathology and Experimental Neurology 19, 357-62. WlSNIEWSKI, H. M.. and KEITH, A. B. (1977) Chronic relapsing EAE - an experimental model of multiple sclerosi~s. Annals of Neurology 1, 144--8. WISN1EWSKI, H.M., PRINEAS, J. and RAINE, C.S. (1969) An ultrastructural study of experimental demyelination and remyelination. I. Acute experimental allergic encephalomyelitis in the PNS. Laboratory Investigation 2 1 , 1 0 5 - 1 8 . r
Peripheral nervous system pathology in relapsing experimental allergic encephalomyelitis R. E. M A D R I D
1 and H. M. W I S N I E W S K I
2
1M.R.C. Demyelinating Diseases Unit, Newcastle-upon-Tyne, England z New York State Institute for Basic Research in Mental Retardation, Staten Island, New York, U.S.A.
Received 22 July 1977; revised 5 October and 16 November 1977; accepted 18 November 1977
Summary A recurrent type of primary demyelination was found in the anterior roots and dorsal root ganglia of strain 13 guinea-pigs with relapsing experimental allergic encephalomyelitis. Single nerve fibre studies revealed a predominant n o d a l - p a r a n o d a l distribution of the destruction of myelin in these nerves. Damage of the Schwann c e l l - a x o l e m m a l junction was found in the majority of the abnormal nodes of Ranvier. This was accompanied by the formation-of supernumerary adaxonal Schwann cell processes which further altered the normal myelinating c e l l - a x o n relationship. It is concluded that variations in antigenic composition may play a role in the selective involvement of the anterior roots. The pathogenesis of the observed nodal changes is discussed.
Introduction Multiple sclerosis (MS) is a disease characterised by C.N.S. myelin pathology in which recent and old demyelinating lesions usually coexist (Dawson, 1915). In the search for the cause of this disease an animal model has been developed which bears much closer similarities to the human disease than acute experimental allergic encephalomyelitis (EAE) (Stone and Lerner, 1965; Snyder et al., 1975;Wis'niewski and Keith, 1977). The observed changes in the P.N.S. in MS (Schob, 1915) have, up to now, been considered to be non-specific (Hasson et al., 1958) or merely to reflect changes in the glial portion of the spinal roots (see Dawson, 1915). However, recent studies (Calder and Pollock, 1976; Schoene et al., 1976) suggest that the P.N.S. may itself be abnormal in MS. A consistent involvement of the P.N.S. was not observed in previous studies on chronic EAE (Raine et al., 1974), although Schwann cell invasion of the spinal cord was frequently found in long-term animals (Snyder et al., 1975). The latter suggested to us the presence of a concomitant P.N.S. patholog3z. The present report is part of a comprehensive study on the ultrastructural pathology of the recurrent EAE model developed in this laboratory (Wigniewski and 9 1978 Chapman and Hall Ltd. Printed in Great Britain
266
MADRID and WISNIEWSKI
Keith, 1977) and focuses on the P.N.S. lesions, a h i t h e r t o u n r e c o g n i z e d aspect o f the disease process.
Materials and m e t h o d s Sixteen juvenile Strain-13 guinea-pigs were injected in the dorsum of both hind feet with a total of 0.2 ml of an emulsion containing equal volumes of a homogenate of guinea-pig spinal cord (1:1, w/v in saline) and Freund's complete adjuvant containing 10 mg/ml tubercle bacilli (H37Ra). Animals were periodically assessed for the presence of clinical signs which were evaluated under a scoring system described elsewhere (Eastman et al., 1978). At different times after the first, second and third clinical relapses (see Table 1) the animals were heparinized and killed by intracardiac perfusion fixation with 4% glutaraldehyde in 0.1 M phosphate buffer, pH 7.4 for 15 min. Immediately after perfusion, the spinal cord (SC) and associated dorsal root ganglia (DRG) and roots were dissected out. With the aid of a dissecting microscope, the dura was excised and the L7 and $1 spinal roots and DRG removed. Both anterior and posterior roots were divided into 3 segments in which the middle portion of each, about 1.5 cm in length, was prepared for teasing_ in fluid Epon 812, as previously described (Madrid and Wigniewski, 1977). Samples from the following regions were taken: the remaining 3 mm segments on each end of the roots (near the spinal cord and near the DRG)~ L7 and Sx spinal cord with corresponding DRG; and T~ 0 - T 12 spinal cord. Osmication was carried out in 1% Dalton's chrome osmium solution for 2 h. Specimens were then washed briefly in 0.1 M phosphate buffer with sucrose, dehydrated in graded concentrations of ethanol, cleared in propylene oxide, and embedded in Epon. 1/Jm thick sections from all the specimens mentioned above were cut in an LKB,III ultramicrotome, stained with toluidine blue and studied by light microscopy. From these blocks, selected areas were trimmed and thin sections produced with a diamond knife. Thin sections were collected on copper grids without support film and contrasted with ethanolic uranyl acetate and lead citrate. They were examined with Philips 201 or 300 electron microscopes operating at 60 KV. Approximately 500 nerve fibres from both anterior and posterior L7 and $1 spinal roots were teased out in each of the experimental animals. Selected fibres were transferred to toluene-cleaned glass slides and mounted in Epon for photographic purposes (Spencer and Thomas, 1970). Fibres of special interest were mounted on previously hardened Epon blocks and subsequently oriented for transverse or longitudinal thin sectioning (Ochoa, 1972).
Results T h e n u m b e r o f clinical episodes, the distribution o f the lesions in the p r o x i m a l P.N.S. and in the spinal cord, and the times after i m m u n i z a t i o n w h e n the guinea-pigs were killed, are shown in T a b l e 1. No positive correlation was f o u n d b e t w e e n the location a n d / o r the e x t e n t o f the lesions in the L 7 and S 1 spinal cord and t h o s e in the spinal roots. D e m y e l i n a t i n g plaques were m o r e extensive in the lower t h o r a c i c t h a n in the lumbosacral spinal cord in these long-term animals. In animals e x a m i n e d at intervals after the first and second clinical episode, the spinal r o o t s s h o w e d a small percentage of nerve fibres (usually less than 4%) u n d e r g o i n g axonal degeneration, with no evidence of p r i m a r y d e m y e l i n a t i o n and r e m y e l i n a t i o n . A characteristic p a t t e r n o f i n v o l v e m e n t o f the P.N.S. emerged o n l y after the third clinical episode, or when, after a series o f relapses, the animal had e n t e r e d a c h r o n i c progressive course of neurological deficit. T h e P.N.S. lesions were restricted exclusively t o the a n t e r i o r
267
P.N.S. P a t h o l o g y in relapsing EAE Table 1. Distribution of the lesions in the P.N.S. of inbred guinea-pigs with chronic relapsing-EAE. L 7 S1 ventral roots
GP number
Number of clinical episodes
Days postimmunization (DPI) at time of sacrifice
Near spinal cord
Teased fibres
Near dorsal root ganglia
2781/6 1805/6
1 1
105 107
(-) (-)
(--) (-)
(-) (--)
1897/6 1819/6 2196/6 1910/6
2 2 2 2
60 74 83 84
(+) (+) (--) (-)
(--) (--) (--) (--)
(--) (--) (-) (+)OB
1815/6 1807/6 1953/6 1809/6 1806/6 1027/6
3 3 3 3 3P 3 3 4 4 P
(+)OB (+)OB,MB (+)MB (+)OB,MB (+)OB (+)OB (+)OB (+)OB, MB (+)OB,MB (+)OB,MB
(-) (-) (+) (+) (+) (+) (+) (+) (+) (+)
(-) (-) ( ) ( ) (+)OB (+)OB (+)O B, MB (+) (+) (+)
1177/5
907/5 714/5 1006/5
112 156 197 200 213 223 370 240 635 383
No attempt was made at scoring the histological changes in the present series; GP, guineapig; P, animals showing a progressive type of neurological deficit after the third clinical episode (1806) or from the beginning (1006/5); (+) denotes the presence of demyelination and/or remyelination of P.N.S. fibres at the locations indicated; OB indicates the presence of onion-bulb formations; MB denotes the finding of paranodal ballooning of the myelin sheath; (-) denotes cases with no evidence of de- or remyelination of P.N.S. fibres.
roots at various points in their subarachnoid course. Posterior root changes were f o u n d only in the DRG. In spite of extensive sampling, massive aggregates of i n f l a m m a t o r y ceils (as seen in the dorsal roots in acute EAN in the guinea-pig) could n o t be found. However, teasing of the anterior roots revealed the presence of fibres undergoing paranodal, or sometimes partial internodal d e m y e l i n a t i o n . These were easily identified b y the close apposition of myelin-laden macrophages to the bare segments of the axon (Fig. 1). The coexistence of d e m y e l i n a t i o n with varying degrees of r e m y e l i n a t i o n was d e m o n s t r a t e d in the same nerve fibre b y the variations in sheath thickness f r o m one segment to another. In addition, striking nodal and paranodal abnormalities were noticed in nerve fibres with otherwise normally m y e l i n a t e d internodes. The latter consisted of: (a) paranodal 'repair' and paranodal r e m y e l i n a t i o n (Fig. 2); (b) widening of the nodal gap; ( c ) t h e so-called 'paranodal
268
MADRID and W I S N I E W S K I
overriding' of the myelin sheath (Ballin and Thomas, 1969; King et al., 1975); and (d) myelin 'ballooning'. Nerve fibres in advanced stages of axonal degeneration were also occasionally observed. The uhrastructural features of nerve fibres undergoing paranodal demyelination will be described elsewhere. Onion bulb formations were often found in the anterior root exit zone (Figs. 3 - 7 ) and the anterior root itself (see Table 1)
Paranodal 'repair' and paranodal remyelination Thinning of the myelin sheath, assumed to be associated with the repair of partial damage of the paranodal myelin sheath (Ailt, 1969) was frequently found in 2 or more paranodes of a single nerve fibre (Fig. 2). On occasions, this was an isolated feature through 6 or more otherwise normal internodes, but more frequently, this paranodal thinning was seen in nerve fibres also showing short intercalated internodes (Fig. 2). The latter are also known to occur during repair of localized paranodal demyelination (Ochoa et al., 1972). Paranodal thinning of the myelin sheath was extensive, often starting 50 #m or more away from the nodal region (Fig. 2). Short intercalated internodes circa 50/am in length were also a common finding. These light microscopical indications of nodal damage and repair were confirmed by electron microscopy. In fibres displaying paranodal thinning of the myelin sheath Fig. 1. Teased fibre appearance of recurrent demyelination and remyelination. 240 days post-immunization (DPI). Segmental loss of myelin and paranodal concentration of myelin-laden macrophages (row 4) coexist with short remyelinated internodes of varying myelin thickness in the same nerve fibre. Fig. 2. Teased fibre from the same animal shown in Figure 1. A short (62/lm) internode (ic) is intercalated between segments with normal myelin thickness. Paranodal thinning of the myelin sheath (t) is seen in the next 2 internodes. Fig. 3. L7 spinal cord, 370 DPI. Low power electron micrograph of the anterior root exit zone. Concentric arrangements of attenuated Schwann and pial cell processes surround the majority of the P.N.S. fibres in this area. The anterior column is to the upper left in this figure. Fig. 4. Detail of the surrounding Schwann (S) and pial processes (P) from Figure 3. The latter are not covered by a basal lamina. Fig. 5. Extremely attenuated Schwann cell processes from the lower right onion bulb in Figure 3. Horizontally running collagen fibres are also evident (c). Fig. 6. Abaxonal origin of an attenuated Schwann profile. Fig. 7. A Schwann cell process from Figure 3 showing neurofilamentous hyperplasia. Note encompassing basal lamina (bl). Fig. 8. Paranodal thinning of the myelin sheath. The area of transition from thick to thin myelin is indicated (arrows). Fig. 9. Detail from figure 8. At least 5 reversing myelin lamellae are apposed to the axolemma. Fig. 10. Teased anterior root nerve fibre showing a 47/lm long intercalated internode. Compare with Figure 2. Fig. 11. Detail of lower node of Ranvier from the previous figure. The internode with normal myelin thickness shows dense terminal loops which are separated from the axolemma by interposed processes of Schwann cell origin (arrows). A mononuclear cell (M) is seen in the nearby endoneurial space.
272
MADRID and WISNIEWSKI
the area o f transition b e t w e e n thick and thin m y e l i n revealed an irregular t e r m i n a t i o n o f m y e l i n lateral loops some distance away f r o m the n o d e o f Ranvier. M a n y terminal loops failed to m a k e c o n t a c t with the a x o l e m m a while o t h e r s were apposed to the a x o n as reversing loops (Figs. 8 and 9). In contrast, the lateral loops o f the i n t e r n o d e s with n o r m a l m y e l i n thickness o n b o t h sides o f an intercalated segment shows a peculiarly increased e l e c t r o n d e n s i t y , and were generally smaller t h a n n o r m a l terminal loops (Fig. 10). In addition, these a b n o r m a l terminal loops were separated f r o m the a x o l e m m a b y interdigitating processes o f S c h w a n n cell origin which had emerged f r o m the a d a x o n a l region at varying distances f r o m the n o d e o f Ranvier (Fig. 11). These were long, flat processes which lacked the familiar club-shaped a p p e a r a n c e o f m y e l i n terminal loops. U p o n arrival at the node, these adaxonal Schwann cell processes a p p e a r e d to m o v e a w a y f r o m the a x o n m e m b r a n e t o w a r d the e n d o n e u r i a l space. Schwann cell microvilli were missing in these transitional nodes o f Ranvier, indicating that in spite o f advanced p a r a n o d a l r e m y e l i n a t i o n a n o r m a l c o n t a c t b e t w e e n the 2 S c h w a n n cells had n o t b e e n reestablished. M o n o n u c l e a r cells were f o u n d in the n e a r b y e n d o n e u r i a l space (Fig. 11). T h e m y e l i n terminal loops o f the intercalated segment a p p e a r e d normal.
Nodal widening Widening o f the n o d e o f Ranvier was a c o n s p i c u o u s f e a t u r e o f r e m y e l i n a t i n g m o t o r axons in the anterior r o o t s (Figs. 12 a n d 13). Occasionally, a s u p e r n u m e r a r y Fig. 12. 370 DPI. Nodal widening at the junction of 2 remyelinated segments. The 12/2m long gap (arrows) is filled here by a Schwann cell. Normal terminal loops are evident in the remyelinated internode on the right but a few myelin lamellae are seen ending beyond these terminal loops (right arrow). Fig. 13. A 29 ;lm long nodal gap between remyelinated internodes of a teased anterior root fibre. x 750. Fig. 14. Skip-serial thin section of the upper half of the nodal region from Fig. 12. The Schwann cell nucleus is now in the plane of the section. Note tortuous course of some of the adjacent myelin lamellae (right). The length of axolemma covered by the supernumerary Schwann cell is indicated (arrows). Fig. 15. Nodal widening at the junction of internodes with normal myelin thickness. 223 DPI. On both sides of the node of Ranvier the myelin sheath tapers off. The perinodal space is almost saturated with cellular processes. 1/2m toluidine blue section, x 750. Fig. 16. Ultrastructural appearance of a similarly damaged nerve fibre, 240 DPI. Dense terminal loops separated from the axolemma by interdigitating Schwann cell processes are seen on both myelinated segments (arrows). Figs. 17 and 18. High power view of the zones arrowed on the bottom half of the right paranodal segment in the previous figure. Fig. 17. A narrow Schwann cell process emerges from one of the major dense lines. Fig. 18. A seemingly independent process is seen at this level. Note the gradual change in size and electron density of the terminal loops as they approach the nodal gap. a, axolemma. Scale: 0.1 gtm. Fig. 19. Detail from a widened node similar to those shown in Figs. 15 and 16. A fine Schwann cell process extends from the paranodal area toward the node of Ranvier (arrows) and intervenes between the dense terminal loops and the axolemma. Accumulation of mitochondria (M) and the base of a narrow process (p) are seen in the abaxonal Schwann cell cytoplasm.
274
MADRID and WISNIEWSKI
Schwann cell was seen filling the widening gap (Fig. 14). Usually, one or b o t h remyelinated segments showed focal or generalized a b n o r m a l i t y in the a b u t m e n t o f myelin terminal loops, b u t the m o r p h o l o g y of the terminal loops was normal (Fig. 14). Nodal widening was also f o u n d in otherwise normal internodes (Fig. 15) usually accompanied b y a characteristic tapering of the paranodal m y e l i n sheath. A profusion of Schwann and m o n o n u c l e a r cell processes usually surrounded these abnormal nodal regions (Fig. 16). The latter were characterized b y the absence of a basal lamina. Electron dense terminal loops made up the regular ends of m o s t of the myelin lamellae at these nodes. A few 'clear' terminal loops closely associated with elongated Schwann cell processes were usually f o u n d at points farthest f r o m the node. These narrow Schwann cell processes spread t o w a r d the nodal gap, separating the electron dense terminal loops f r o m the a x o l e m m a (Figs. 17 and 18). A l t h o u g h the Schwann cell origin of m a n y of the elongated cellular processes could easily be d e m o n s t r a t e d (Fig. 19), some profiles appeared to be i n d e p e n d e n t of the p a r e n t myelinating cell. Myelin terminal loops which were in close c o n t a c t w i t h the latter showed an increased electron density. Schwann cell microvilli were absent in these abnormal nodes of Ranvier indicating t h a t d e m y e l i n a t i o n had previously occurred at this level.
Fig. 20. 1 #m toluidine blue section of a teased anterior root fibre showing paranodal overriding, 223 DPI. A number of cells and their processes are seen in the adaxonal space of the myelinated segment. Demyelination has occurred on the right and a myelin-laden macrophage is still in the area (md). (R) denotes the site of the former node of Ranvier. Fig. 21. Serial thin section of the nerve fibre from Fig. 20. The axon is now out of the plane of section. Schwann (S) and mononuclear cells (M) and their processes (p) are present inside the myelin sheath. The (*) identifies the cells and processes shown at higher magnification in Figs. 22-24. Fig. 22. A minute Schwann cell with its basal lamina (bl) and a mononuclear celt process (mp) are seen in this area. Fig. 23. Mononuclear cell process (mp) near the previously adaxonal portion of the myelin sheath (m). 2 Schwann cell processes covered by a basal lamina are also seen. Fig. 24. Detail of a mononuclear cell process from the bottom half of Fig. 20, which is in contact with the inner aspect of the myelin sheath. Fig. 25. 223 DPI. Schwann and mononuclear cell infiltration around a myelinated fibre over a length of at least 100/~m. A gradual thinning of the outer myelin sheath occurs from left to right. The arrows point to the areas in which vesicular disruption of single myelin lamellae was observed. These are shown at higher magnification in Figs. 28 and 29. The location of the original node of Ranvier is indicated (R). Fig. 26. Nodal region from Fig. 25. A mononuclear cell process is in contact with the terminal loops some of which are undergoing vesiculation. Schwann profiles intervene between the inner (im) and outer (om) myelin sheaths. Fig. 27. 'Collagen pockets' in the extracellular gap of the myelinated fibre from Fig. 25. A mononuclear cell process is apposed to the outer myelin sheath. Fig. 28 and 29. Detail of the vesicular disruption of myelin lamellae from the outer myelin sheath in Fig. 25.
276
MADRID and W I S N I E W S K I
Paranodal overriding of the myelin sheath Longitudinal sections of teased anterior root fibres proved particularly informative in the study of the mechanism of formation of these bizarre Schwann cell-axon relationships. For example, Fig. 20 shows an axon partially encompassed by Schwann cell processes but not yet remyelinated. A combination of Schwann and mononuclear cells were found inside the myelin sheath (Figs. 21--24). A gradual attenuation in the thickness of the outer, 'overriding', myelin sheath was constantly found in this material. In the area of transition from thick to thin outer myelin sheath, a node of Ranvier or a Schmidt-Lanterman cleft was found. In the former, grossly enlarged processes of Schwann cell origin were in contact with the axolemma, which lifted away the dense myelin loops at that level. Invariably, mononuclear cells or their processes showed a tendency to lie in contact with the innermost myelin lamellae of the 'overriding' myelin sheath (Figs. 23 and 24). In well-developed cases, the space between the 2 myelin sheaths (the periaxonal space of the parent and now outer Schwann cell) was filled with minute, uncommitted Schwann cells, mononuclear cells, newly formed collagan fibrils and flattened Schwann cell processes (Fig. 25). Electron dense terminal loops, some of which were undergoing vesicular disruption, were frequently seen in this 'outer' myelin sheath in the presence of mononuclear cell processes for considerable distances from the node of Ranvier (Figs. 25, 28 and 29). This progressive peeling-off and vesiculation of the inner aspect of the 'overriding' myelin sheath, occurring in the presence of these processes, seemed to be the main cause of the gradual thinning of the paranodal myelin in these areas. No morphological indication of active growth or extension of the myelin sheath could be found at the level of the original node of Ranvier (Fig. 26). On the contrary, the lateral endings of the myelin sheath sometimes seemed to turn backward instead of projecting forward. Vesicular disruption near mononuclear cell processes also occurred at this level (Fig. 26). Very thin Schwann cell processes surrounded by a basal lamina were frequently seen emerging from the external surface of the original Schwann cell at its paranodal level, but these did not usually come into contact with each other. However in the dorsal root ganglia these processes were occasionally juxtaposed. All these examples of 'paranodal overriding' took place at the junction of internodes which had a myelin sheath of normal thickness. On occasions, however, this abnormality was detected in the nodal region of remyelinated segments of a nerve fibre.
Myelin ballooning Bizarre and extensive paranodal dilatations of the myelin sheath were frequently observed adjacent to a demyelinated segment of the axon in anterior root fibres.
Fig. 30. Single teased anterior root nerve fibre showing paranodal ballooning of the myelin sheath in the vicinity of a demyelinated segment. 1 0 - 2 0 / 2 m step serial 1 /lm Epon sections taken from the sectors indicated in the first row.
278
MADRID and WISNIEWSKI
These fibres were particularly difficult to tease apart, due to their frequent close relationship with endoneurial blood vessels, and to a peculiar 'stickiness' which firmly attached these myelin 'balloons' to neighbouring myelinated fibres. Skipserial transverse sections of the demyelinated area and the dilated portion of the myelin sheath showed that the axon lay outside the myelin bleb formation for a considerable distance (Fig. 30). A thin myelin sheath surrounded the axon at the level of its penetration into the dilated myelin sheath. Macrophages were seen at the level of penetration of the axon into the bleb formation but not inside the myelin balloon. Ultrastructural examination (levels 9 - 1 1 in Fig. 30) showed that there was no basal lamina over the thin myelin covering this axon, indicating that the thin adaxonal and the dilated myelin sheaths were derived from the same Schwann cell. Discussion
Although P.N.S. lesions have occasionally been reported in acute EAE in several species, the distribution of the lesions in the proximal P.N.S. has seldom been the subject of a detailed morphological analysis. Waksman and Adams (1956) found posterior root entry zone lesions in only 13% of 30 guinea-pigs immunized with C.N.S. tissue. They showed that the anterior roots were not significantly involved, in contrast to studies on EAE in rabbits (Wi~niewski et al., 1969). The P.N.S. lesions of the recurrent-EAE animals (R-EAE) described here: (a) showed a remarkable nodal-paranodal distribution; (b) occurred in the absence of perivascular aggregates of inflammatory Cells; and (c) were limited to the anterior roots and DRG. Some of the findings of this study are similar to those of King et al. (1975) in relapsing experimental allergic neuritis (EAN) in outbred guinea-pigs. These include the paranodal thinning of the myelin sheath and paranodal overriding of myelin lamellae. However, significant differences are found in the configuration of the lesions. Firstly, the paranodal attenuation of the myelin sheath was frequently observed in otherwise normally myelinated internodes as shown by single teased fibre preparations. King et aL (1975) showed paranodal thinning of remyelinated axons. It is possible that the latter was the consequence of unfinished lengthening of a recently formed myelin sheath while a partial repair and/or re-apposition of a partially damaged myelin sheath seems more likely in the present study. Secondly, the extracellular gap in the examples of paranodal overriding reported by King et al. (1975) appeared almost empty while it was filled with a number of Schwann and mononuclear cells in our material. A gradual attenuation of the outer myelin sheath in the presence of mononuclear cell processes was always found in this study but was not reported by those authors. 'Perinodal' overriding as described by King et aL (1975) was not observed by us. Therefore we conclude that in contrast to the situation reported in the relapsing EAN animals, paranodal 'overriding' in this material was due to Schwann and mononuclear cell infiltration of the adaxonal space. Whether or not these differences are due to our using inbred guinea-pigs immunized with a single injection of homologous C.N.S. tissue is not clear. That
P.N.S. pathology in relapsing EAE
279
presumably additional mechanisms of myelin damage are operating in this animal model is suggested by the frequent occurrence of nodal widening, a lesion which was not detected by King et al. (1975). Unlike the C.N.S. lesions found in these animals, the observed P.N.S. myelin pathology occurred in the absence of perivascular aggregates of inflammatory cells. That this could merely reflect a faster rate of clearance of inflammatory cells from the endoneurial space of the anterior roots is supported by a similar finding in relapsing EAN animals (King et al., 1975). The participation of inflammatory cells in the pathogenesis of the described myelin lesions was demonstrated here by single fibre microdissection. Teased fibre preparations also proved the recurrent nature of the P.N.S. disease process by disclosing the coexistence of varTing degrees of both paranodal demyelination and remyelination in consecutive internodal lengths of the same nerve fibre. Since the myelin balloons described here are in all respects similar to those previously described by Spencer and Thomas (1970), they are also evidence for demyelination of already remyelinated segments of those fibres. In contrast to a recent report (Snyder et al., 1977) our results indicate that 'onion bulbs' were the consequence of a recurrent process of demyelination and remyelination, rather than a single episode of demyelination. The restriction of P.N.S. lesions to the anterior roots and DRG in R-EAE agrees with the finding of Pollard et al. (1975) in outbred guinea-pigs induced to relapse by repeated injections of P.N.S. tissue. Acute EAN is known to be a predominantly posterior root disease (Waksman and Adams, 1956; Allt, 1976; Madrid and Wi~niewski, 1977). The reason for the shift in the location of the pathological changes in relapsing EAN (Pollard et al., 1975) and for the anterior root distribution of the lesions in R-EAE is not clear. Since the lesions seen in these long-term animals increased in frequency with time after immunization, the possibility that they could be age relate d should be considered. Age-related changes in rats (Berg et al., 1962; Gilmore, 1972) and dogs (Griffiths and Duncan, 1975) usually appear near the end of their mean lifespan, are apparently demyelinating in nature and involve both anterior and posterior roots. It is unlikely that this mechanism is operating in the present situation, because throughout this study the posterior roots were not involved. Alternatively, a hitherto unrecognized difference in the antigenic composition of the anterior roots could explain their preferential involvement in this series. Studies under progress in this laboratory are aimed at elucidating this problem. It is possible that the nodal widening found at the junction of remyelinated segments was due to an unfinished process of remyelination, or to the interfering effect of intercalated Schwann cells. However, the nodal widening of internodes with normal myelin thickness could be the consequence of a direct effect of antimyelin or antiaxonal antibodies. Since nodal widening and aberrant termination of paranodal myelin loops have been reported in the C.N.S. in MS (Suzuki et al., 1969; Andrews, 1972), the present study further emphasizes the relevance of the R-EAE model to MS. Although patho-physiological studies of the abnormal nodes were not done it is
280
MADRID and WISNIEWSKI
conceivable t h a t the n o d a l - p a r a n o d a l changes described here could significantly disrupt the propagation of the nerve action potential in the anterior roots of these animals.
Acknowledgements This w o r k was supported b y the Multiple Sclerosis Society of Great Britain and N o r t h e r n Ireland (grant no. MNI-1/75). We t h a n k Mr A. B. Keith for dedicated care of the long-term animals used in this study, Miss Margaret Herron for invaluable secretarial assistance, and Dr J. R. M c D e r m o t t for reading the manuscript. References ALLT, G. (1969) Repair of segmental demyelination in peripheral nerves: An electron microscope study. Brain 92, 639-46. ALLT, G. (1976) Pathology of the peripheral nerve. In: The Peripheral Nerve (edited by LANDON, D. N.), pp. 666-740. London: Chapman and Hall. ANDREWS, J.M. (1972) The ultrastructural neuropathology of multiple sclerosis. In: Multiple Sclerosis, Immunology, Virology and Ultrastructure (edited by WOLFGRAM, F., ELLISON, G.W., STEVENS, J.G. and ANDREWS, J. M.), pp. 23--52. New York, London: Academic
Press. R.H.M. and THOMAS, P.K. (1968) Electron microscope observations on demyelination and remyelination in experimental allergic neuritis. I. Demyelination. Journal of the Neurological Sciences 8, 1-18. BALLIN, R.H.M. and THOMAS, P.K. (1969) Electron microscope observations on demyelination and remyelination in experimental allergic neuritis. 2. Remyelination. Journal of the Neurological Sciences 8, 225--37. BALLIN,
BERG, B. M., WOLF, A., SIMMS, H. S. (1962) Degenerative lesions of spinal roots and peripheral
nerves in aging rats. Gerontologia 6, 72-80. C.S. and POLLOCK, M. (1976) Peripheral nerve changes in multiple sclerosis. Proceedings of the University of Otago Medical School 54, 5-6. DAWSON, J.W. (1915) The histology of disseminated sclerosis. Transactions of the RoyalSociety of Edinburgh 50, (iii), 517--740.
CALDER,
EASTMAN, R., KEITH, A. B. and WISNIEWSKI, H.M. (1978) Chronic relapsing experimental
allergic encephalomyelitis: a clinico-pathological study. Neuropatbology and Applied Neurobiology, (submitted). GILMORE, S.A. (1972) Spinal nerve root degeneration in aging laboratory rats: a light microscopic study. Anatomical Record 174, 251-8. GRIFFITHS, I. R. and DUNCAN, I. D. (1975) Age changes in the dorsal and ventral lumbar nerve roots of dogs. Acta Neuropathologica 32, 75--85. HASSON, J., TERRY, R. D. and ZIMMERMAN, H. M. (1958) Peripheral neuropathy in multiple
sclerosis. Neurology 8,503--10. KING, R. H. M., POLLARD, J. D. and THOMAS, P. K. (1975) Aberrant remyelination in chronic
relapsing experimenta! allergic neuritis. Neuropathology and Applied Neurobiology 1, 367-78. MADRID, R. E. and WISNIEWSKI, H. M. (1977) Axonal degeneration in demyelinating disorders.
Journal of Neurocytology 6, 108-17. OCHOA, J. (1972) Ultrathin longitudinal sections of single myelinated fibres for electron microscopy. Journal of the Neurological Sciences 17, 103-6.
P.N.S. pathology in relapsing EAE
281
OCHOA, J., FOWLER, T. J. and GILLIAT, R. W. (1972) Anatomical changes in peripheral nerves compressed by a pneumatic tourniquet. Journal of A natomy 113, 433-55. POLLARD, J.D., KING, R. H. M. and THOMAS, P. K. (1975) Recurrent experimental allergic neuritis. An electron microscope study. Journal of the Neurological Sciences 24, 365-83. RAINE, C.S., SNYDER, D.H., VALSAMIS, M.P. and STONE, S.H. (1974) Chronic experimental allergic encephalomyelitis in inbred guinea pigs. An ultrastructural study. Laboratory Investigation 31, 3 6 9 - 3 8 0 . SCHOB, F. (1915) Ein beitrag zur pathologischen Anatomic der multiplen sklerose. Monatscbrift fur Psychiatric und Neurologie 22, 62-87. SCHOENE, W. C., GESCHWIND, N. O., BEHAN, P. O. and CARPENTER, S. (1976) Coexistence of multiple sclerosis and hypertrophic neuropathy: Light, electron microscopical, immunological and clinical studies. Neuropathology and Applied Neurobiology 2, 165 (abstract)~ SNYDER, D.H., VALSAMIS, M.P., STONE, S.H. and RAINE, C.S. (1975) Progressive demyelination and reparative phenomena in chronic experimental allergic encephalomyelitis. Journal of Neuropatbology and Experimental Neurology 34, 2 0 9 - 2 1 . SNYDER, D.H., STONE, S.H. and RAINE, C.S. (1977) Attempts to induce chronic experimental allergic neuritis in Strain 13 and Harttey guinea pigs. Journal of Neuropathology and Experimental Neurology 36, 4 8 8 - 9 8 . SPENCER, P. S. and THOMAS, P. K. (1970) The examination of isolated nerve fibres by light and electron microscopy with observations on demyelination proximal to neuromas. Acta Neuropatbologica 16, 177--86. STONE, S.H. and LERNER, E2 M. (1965) Chronic disseminated allergic encephalomyelitis in guinea pigs. Annals of the New York Academy of Sciences 122, 2 2 7 - 4 1 . SUZUKI, K., ANDREWS, J. M., WALTZ, J. M. and TERRY, R. D. (1969) Ultrastructural studies of multiple sclerosis. Laboratory Investigation 20, 444--54. WAKSMAN, B.H. and ADAMS, R.D. (1956) A comparative study of experimental allergic neuritis in the rabbit, guinea pig and mouse. Journal of Neuropathology and Experimental Neurology 19, 357-62. WlSNIEWSKI, H. M.. and KEITH, A. B. (1977) Chronic relapsing EAE - an experimental model of multiple sclerosi~s. Annals of Neurology 1, 144--8. WISN1EWSKI, H.M., PRINEAS, J. and RAINE, C.S. (1969) An ultrastructural study of experimental demyelination and remyelination. I. Acute experimental allergic encephalomyelitis in the PNS. Laboratory Investigation 2 1 , 1 0 5 - 1 8 . r
