Brain Research, 524 (1990) 303-312 Elsevier
303
BRES 15760
Remyelination of nerve fibers in the transected frog sciatic nerve Chaim T. Rubinstein and Peter Shrager Department of Physiology, University of Rochester Medical Center, Rochester, NY 14642 (U.S.A ) (Accepted 20 February 1990) Key words: Remyelination; Schwann cell; Demyelination; Myelinated nerve; Degeneration
This project tests an important aspect of the cellular events controlling the processes of recovery of function and remyelination that follow demyelination in the peripheral nervous system. Frog sciatic nerves have been shown to survive and remain functional for up to 10 days following transection29. We have utilized this property in order to dissociate the recovery process from possible control by the neuronal soma. Xenopus sciatic nerves were demyelinated in one branch by an intraneural injection of lysolecithin. The nerve was cut proximally to the injection site either immediately before, or several days after the lysolecithin injection. Recovery of function and remyelination were then followed by electrophysioiogical, optical, and ultrastructural techniques applied both to whole branches and single fibers. Controls included the cut but uninjected branch, and injected but uncut nerves. The progression of events during both demyelination and recovery in cut axons was indistinguishable from that in uncut fibers. This suggests that this process may be under local control and can be initiated and carried out in the absence of constant communication with the nerve cell body. INTRODUCTION When axons of the peripheral nervous system are demyelinated conduction is at first blocked, but later recovers. The sequence of events involved in this process has been studied extensively both morphologically and physiologically6-8'23'24'3°'31. Damaged myelin is removed via phagocytosis by macrophages and Schwann cells. Proliferating Schwann cells are then observed to associate with the axons and surround them with numerous processes, ultimately resulting in remyelination. This procedure results in unusually short internodes as new nodes of Ranvier are formed in regions that previously were internodal. Electrical signalling at very low conduction velocity is seen in only a very small percentage of fibers prior to remyelination 6"24"26. The fraction of axons conducting through the lesion increases rapidly as early remyelination progresses, though long delays in propagation remain. Much work has been done to elucidate the cellular processes controlling both myelin removal and the subsequent recovery through Schwann cell proliferation and remyelination, though many interesting questions remain. Conditioned medium from cultured peritoneal macrophages that have phagocytosed a myelin membrane fraction is mitogenic for cultured Schwann cells 3. A study of Wallerian degeneration, in which the degenerating nerve was isolated from interaction with non-local cells
showed no proliferation of Schwann cells and no evidence for their role in the removal of myelin 4. When macrophages were allowed to migrate into the experimental chamber the macrophages were shown to attack selectively the myelin sheath rejected by the Schwann cells but not the Schwann cells themselves. The activity of the phagocytes was mediated by a signal which diminished in intensity with time, originating, presumably from the neuron. Similarly, Schwann cell proliferation, migration and adhesion can be stimulated by extracellular fluid surrounding regenerating peripheral nerve 14. When normal Schwann cells grown in tissue culture are transplanted into the demyelinated peripheral nerves or the demyelinated spinal cords of mutant quaking mice they will ensheathe and remyelinate these axons 2'12. Pellegrino et al., suggested that Schwann cell mitosis during peripheral nerve degeneration is due to a signal coming from the axon 2°. These studies suggest that control of both myelin removal and remyelination are complex and indicate that under some circumstances there is a neuronal signal required for ensheathment and myelination. Does this signal arise from an endogenous factor in the axon or does it require the synthesis of some substance by the cell body? In this report we test one aspect of the control process, the possible involvement of neuronal factors derived from the soma. Wang 29 has shown that in the frog the distal stump of sciatic nerves can survive at room
Correspondence: C.T. Rubinstein, Department of Physiology, Box 642, University of Rochester Medical Center, 601 Elmwood Avenue, Rochester, NY 14642, U.S.A. 0006-8993/90/$03.50 © 1990 Elsevier Science Publishers B.V. (Biomedical Division)
304 temperature for up to 10 days following transection (and even longer at cooler temperatures). We have previously studied both structural and functional aspects of demyelination and remyelination in Xenopus peripheral nerve 2324"26. In the present work we examine these processes following both experimental demyelination and transection to separate the lesioned area from the n e u r o n a l cell body. It is shown that events in the distal nerve stump are similar in most respects to those in uncut control axons. A brief report of this work has appeared in abstract form 22.
for each record, stimulating the nerve at 10 Hz. During analysis signals were further filtered using a Gaussian routine 1° and are plotted as the fractional change in transmitted light intensity. The cut-off frequency given for each experiment represents the overall (analog+digital) filtering. Optical measurements were made both on single fibers and on small tracts of axons of uniform morphology. More complete details of this system are given in Shrager and Rubinstein26.
Electron microscopy
MATERIALS AND METHODS
Immediately following either electrophysiological or optical measurements nerves were fixed in 3% glutaraldehyde-H20 2 buffered to pH 7.4 with 0.1 M cacodylate~I. The preparation was then further fixed, stained, dehydrated, embedded, sectioned, and re-stained as detailed in Shragerz4. Nerves were examined in cross-section only which resulted in a very low probability of observing nodes. Thus, we did not address the question of possible changes in membrane specialization at these sites. Thin sections were examined on a Zeiss 10A electron microscope.
Surgical procedures and electrophysiological recording Nerves from the clawed frog Xenopus laevis were demyelinated
RESULTS
with an intraneural injection of lysolecithin~3. Animals were anesthetized in tricaine methanesulfonate (3.7 g/l H20 ). The sciatic nerve in one leg was surgically exposed and 1/~1 of 1% lysolecithin in sterile frog Ringer's was injected into either the tibial or peroneal branch in the thigh via a glass micropipette. The uninjected branch served as a control. The incision was sutured and the animal was allowed to recover. At various times post-surgery frogs were similarly anesthetized and the sciatic plexus in the injected leg was cut with fine dissection scissors about 5-7 mm from the spine. All branches of the sciatic were transected. The proximal end of the distal stump was ligated with thread and the wound was sutured. In one group of frogs the transection and the injection were both done during the same surgery, with the transection preceding the demyelination. Frogs were maintained in aquaria at room temperature and were fed ground beef twice weekly. Six days after cutting the sciatic nerve, frogs were anesthetized in ice water and euthanized by decapitation. The experimental sciatic nerve was dissected and mounted in a chamber containing platinum wires for external stimulation and recording. The nerve was positioned with the tibial and peroneal branches in contact with separate pairs of wires so that activity from each branch could be measured independently. Compound action potentials were recorded in response to a supramaximal shock, from several locations along each branch. The nerve was then desheathed and was dissociated into single fibers by coilagenase (3.5 mg/ml for 1.5 h). Axons could then be gently spread from the bundle over a few millimeters for observation and patch clamping. The loose patch clamp was designed using a modified version of the active bridge circuit of Stuhmer et al.28. Details of this system are similar to those reported previously23-25. The design allows for automatic compensation for both the pipette resistance and the seal resistance. Seal resistances were generally in the range 6-15 MI2. Linear leakage and capacitative currents were subtracted by both analog a0d digital means as described earlierz4. Pulse protocols and data analysis were carried out with the aid of a laboratory computer. The composition of frog Ringer solution was (raM): NaCI, 115; KCI, 2.5; CaCI2, 1.8; HEPES, 5; pH 7.4. All experiments were run at room temperature.
Optical recording of action potentials Nerves were prepared exactly as for loose patch clamp recording except that the potential sensitive dye RH15516 (2 mg/ml) was added to the collagenase solution. A Nikon Diaphot microscope was modified to increase the intensity of illumination, and the incident beam was filtered to 705 + 25 rim. Light transmitted through the axon was focussed on a photodiode connected to a currentto-voltage converter of variable gain. The amplified signals were filtered and digitally sampled. Typically, 64 sweeps were averaged
Electron microscopy The morphology of lysolecithin-induced demyelination and remyelination in otherwise intact nerve was first described by Hall and Gregson 13 in the mouse and more recently in Xenopus by Shrager 23'24. Within the first 3 days after injection debris-covered breaks appear in the myelin. Over the next few days most of this myelin debris is cleared by macrophages and Schwann cells via phagocytosis. Proliferating Schwann cells then bind to the demyelinated regions, extend n u m e r o u s processes, and remyelinate the fiber with the first few lamellae appearing at about 12 days post-injection26. The sequence of events in transected nerves was found to be very similar to that found in uncut nerves. Fig. 1 shows light micrographs of axons from transected sciatic nerves. Fig 1B illustrates a debris-covered break in the myelin of an axon that had been both cut and injected 6 days earlier. In Fig. 1A the damaged myelin has been removed over a short length exposing a bare stretch of axon, also 6 days post-injection. A t 14 days long demyelinated zones can be seen, and in some axons (Fig. 1C) both macrophages (large dark cells) and Schwann cells (smaller clear cells) are associated with the axolemma. In other fibers (Fig. 1D) macrophages are largely gone and Schwann cells are more a b u n d a n t and are beginning to extend processes and the first few lamellae of myelin around the axon (see below). In Fig. 1F we see an axon with small protrusions characteristic of new, thin myelin. This cell is 24 days post-injection. At 31 days the new myelin sheath is further developed, and a new node of R a n v i e r is clearly visible (arrowhead - Fig. 1E). New nodes are characteristic of remyelinating axons, forming short internodes along the remyelinated area 11A4"23. The process of remyelination was examined in transected axons up to 31 days from the time of the initial
305 d e m y e l i n a t i n g injection, We found no significant difference in the time course o f events during the demyelination and r e m y e l i n a t i o n o f the cut nerves as c o m p a r e d to control uncut axons. In o r d e r to obtain a m o r e detailed c o m p a r i s o n b e t w e e n the cut and uncut axons we examined thin sections of cut axons with an electron microscope, c o u n t e d the n u m b e r of myelin layers and com-
p a r e d them with d a t a of uncut remyelinating nerves 23' 24,26. Some r e p r e s e n t a t i v e fibers are shown in Fig. 2, and the results are s u m m a r i z e d in the histograms of Fig. 3. In Fig. 2 A and B we see axons at two stages of demyelination, 6 days post-surgery (and post-cut). In Fig, 2 A the myelin has l o o s e n e d and a m a c r o p h a g e (top) is removing it by phagocytosis. T h e fiber in Fig. 2B has
Fig 1. Light micrographs of unstained living axons. All axons have been cut 6 days prior to the experiment. A: a short break in the myelin sheath exposing a bare segment of an axon with no proliferating Schwann cells, 6 days post-injection. B: a debris-covered break in the myelin sheath, 6 days post-injection. C: macrophages (arrows) attached to a demyelinated segment of an axon at scattered sites, 14 days post-injection. D: a long demyelinated axon segment with adherent Schwann cells (arrowheads), 14 days post-injection. E: a new node of Ranvier (arrowhead) in a remyelinating fiber 31 days post-injection. F: a remyelinating axon 24 days post-surgery. Some attached Schwann cells (arrowheads) can be seen. Scale bars, 10/~m.
306
Fig. 2. Electron micrographs of thin sections of cut demyelinated axons. A: a macrophage (top of figure) associated with an axon that is surrounded by loose myelin, 6 days post-injection. B: a completely demyelinated axon with all myelin debris cleared away, 6 days post-injection. C: a remyelinating axon with one and a half turns of the Schwann cell, 14 days post-surgery. Inset: enlarged view of the outer Sehwann cell lip. D: a remyelinating axon with 5 layers of myelin, 14 days post-injection. Inset: The new, uncompacted myelin at higher magnification. E: 17 days post-injection. A remyelinating axon with 11 layers of new myelin. Inset: An enlarged view of the myelin layers. Scales, 1/tm except for the insets which are 0.1 ~m.
307 been totally demyelinated. Demyelination is largely complete in many axons at 6 days in injected but uncut fibers as well 24. During the second week post-injection proliferating Schwann cells appear throughout the lesion, adhere to the demyelinated axolemma, and begin the process of remyelination. In Fig. 2C a Schwann cell has completed one full turn of new myelin around a cut axon, 14 days post-injection. The cut axon in Fig. 2D, also 14 days after lysolecithin exposure, has 5 lines of new myelin. Fig. 2E shows a cut fiber 20 days post-injection with 11 lines of new myelin. The ultrastructural results are summarized in the histograms of Fig. 3. The label 'N' on the abscissa denotes naked axons. The label 'L' represents loose myelin. The numbers on the abscissa give the number of turns of Schwann cells and myelin counted, '1' denoting one complete turn of the Schwann ceil. A total of 382 cells were counted. The results indicate that the first few turns of new myelin appear between 1 and 2 weeks post-injection. The histograms show a gradual shift to fibers with > 6 turns over the third week. These results are very similar to those observed earlier in injected but uncut axons 26. In these latter experiments the first significant remyelination was seen 12 days post-injection, and the degree of remyelination
then increased steadily over the next several days.
Electrophysiology Compound action potentials were recorded separately from the injected and uninjected branches of the transected sciatic nerve (we varied the lysolecithin injection between the tibial and peroneal branches in different experiments). The uninjected branch served as a control for Wallerian degeneration due to the transection. In addition, following recording, nerves were desheathed, dissociated with collagenase and viewed with a light microscope for signs of degeneration. The criterion for a degenerated nerve was a recorded compound action potential from a control branch smaller than 500/~V, and in nerves failing this test most of the axons invariably had the characteristic appearance of Wallerian degeneration. These preparations were discarded. Over the course of this work we found that in about 55% of the frogs used, excessive degeneration had occurred. In principle this could have been improved by maintaining the animals at cooler temperatures 29 but since we had extensive studies on uncut Xenopus at 20-25 °C, we continued with the same conditions. Loss due to degeneration was greatest between January and May and thus may have had a
Z I00 6 days
14 days
n = 162
n = 72
i°°f
50
0
~
L
N
t
i
i
i
t
i
i
2
3
4
5
6
>6
IO0 17
L
N
t
2
3
4
100
days
20
n = 69
5
6
>6
6
>6
days
n = 79
50
50
L
N
i
2
MYELIN
3
4
LAYERS
5
6
>6
L
N
i
2
MYELIN
3
4
5
LAYERS
Fig. 3. Histograms illustrating the development of remyelination from 6 to 20 days post-injection. The abscissa shows the degree of remyelination. 'L' indicates loose myelin. 'N' indicates naked (completely demyelinated) axons. The numbers give the number of turns of Schwann cell and myelin. The ordinate plots the percentage of cells in each state. A total of 382 cells were counted. A: 6 days post-injection (162 ceils). B: 14 days post-injection (72 cells). C: 17 days post-injection (69 cells). D: 20 days post-injection (79 cells).
308 seasonal component. Extensive degeneration appeared to occur in an approximately all-or-none manner. Of the 15 frogs that passed the test and were studied extensively, one had a compound action potential of 700 #V, and the others were all above 1500 #V. However, even in those preparations that retained strong electrical activity, stimulating the transected nerve close to the cut failed to elicit a compound action potential due to degeneration, forcing us to stimulate the axon 3 - 4 mm more distal. In addition, a small percentage of fibers within a given nerve were seen to be degenerated. C o m p o u n d action potentials recorded from transected nerves over a range of times following injection of lysolecithin are shown in Fig. 4. The numbers at the right give the time post-injection. The top trace is from a cut but uninjected nerve. The stimulus artifact has been blanked and the large signal recorded with a short conduction time is close to that expected for a completely normal nerve. The second trace shows an amplified view of the former record. The vertical lines at the left indicate a scale of 1 0 0 # V at the recording wires. Note that signals due to very slowly conducting fibers are absent in this uninjected nerve. Slow signals are likewise not seen at 6 or 12 days post-injection. In these 2 records both the stimulus artifact and the comparatively large signal from
I
0O
HV UNINJ. ~
"~
UNINJ.
~
6 DAYS
~
12
DAYS
17
DAYS
24
DAYS
1
MS
Fig. 4. Compound action potentials recorded from cut, uninjected nerves (control) and from cut, injected (demyelinated) nerves over a range of times following injection of lysolecithin. All nerves were transected 6 days prior to the experimental day. The scale bar at the left of each trace represents 100 pV at the recording electrodes. Most of the stimulus artifact has been blanked from all sweeps and the initial downward deflection indicates the time of stimulation. The top 2 sweeps show control records from a cut, uninjected nerve. The numbers at the right of each of the lower 4 traces give the days post-injection of lysolecithin. In all but the top trace the fast components of the action potentials (representing the fraction of axons that remained myelinated following the lysolecithin injection) have been blanked.
axons that were never demyelinated by the lysolecithin were blanked before plotting. In general, a small fraction of axons in the bundle remains myelinated with the injection technique used. At 17 days the compound action potential has a significant contribution from fibers conducting with a wide range of very slow velocities. Note that the conduction delays seen are generated in a lesion that is just 2-3 mm in length. At 24 days the number of slowly conducting axons has increased and their average velocity has also increased. In nerves that were injected with lysolecithin but were not cut the profile of conduction velocities seen in compound action potentials was very similar to these results 24. In this earlier work up to 12 days post-injection almost no demyelinated fibers conducted successfully through the lesion. At 17 days post-injection numerous axons were found to conduct, but with a wide spectrum of propagation times, involving delays of about 4 ms. The percentage of remyelinating axons that conduct successfully through the 2-3 mm lesion increased significantly by 30 days and the pattern of a single broad peak was very similar to that seen here at 24 days. In earlier work we have measured ionic currents from the demyelinated axolemma using the loose patch clamp technique 23-25. At internodal sites currents due to voltage-dependent Na ÷ and K ÷ channels were recorded. In Xenopus axons, the average peak inward Na ÷ current was about 750 p A in the range of 6 to 54 days post-injection. In the present experiments these measurements were more difficult due to the apparent increased fragility of the transected plus demyelinated axons. Nonetheless, in some cases adequate sealing of the pipette to the surface membrane was possible. Fig. 5 shows internodal ionic currents recorded at 6 (left) and 16 (right) days post-surgery. The transient inward currents have kinetics and voltage dependence similar to those previously measured at frog nodes and internodes and identified as Na + currents 9'23-25. The peak amplitudes of these currents are within the range of values previously obtained in uncut axons (225-1450 pA). The delayed outward currents are similar to those identified earlier as originating in K ÷ channels on the basis of pharmacological block by tetraethylammonium ion and 4-aminopyridine. Thus these results demonstrate the existence of Na ÷ and K ÷ channels in the internodal membrane of cut axons. The source of the channels in the axon at 6 days post-surgery could not be from proteins newly synthesized in the cell soma in response to the demyelination, since the axon was severed from the soma prior to the injection of lysolecithin.
Optical recording of action potential conduction In uncut axons, at 11 days post-injection proliferating
309 Schwann cells associate with demyelinated axons, but no new myelin is yet present. Conduction across long stretches of the lesion is seen in only a very small fraction of the fibers since the internodal Na + channels alone are insufficient to bridge the impedance mismatch at the transition point between myelinated and demyelinated regions. By 17 days post-injection axons remain covered by numerous Schwann cells and while in electron micrographs regions can be seen that are covered by several lamellae of new myelin, it is not possible to identify presumptive new nodes of Ranvier in the living cell. At this early stage of remyelination the probability of successful conduction increases markedly, though at low velocity26. This pattern is unaltered if the sciatic nerve is cut at the 11 day point. Fig. 6 shows conduction through a tract of just a few demyelinated fibers of uniform appearance measured optically in a preparation 17 days post-injection, cut at 11 days. The records represent the fractional change in transmitted light intensity. The amplitude of the signals is proportional to the change in membrane potential, but also depends on the area of excitable membrane seen by the photodiode and on the degree of dye binding. Thus, the primary information is in the shape and timing of the signals. The graph in Fig. 6B plots conduction time versus distance for these signals. The average conduction velocity (inverse slope) was about 1 m/s through the lesion, compared with 15-20 m/s in normal axons (dashed line). Similar results were obtained at 20 and 22 days post-injection (cut at 14 and 16 days, respectively). Thus, despite transection, axons
A
B
mm
A
B O.OO
2
0.25
0.51 H 1 1.52
,~._/v,,~,,,..~,.~ 2.o3 i
x 1O - 5
O0
i
DISTANCE
4 ms
i
t
i
2 (mm}
Fig. 6. Optical measurement of conduction through a small group of demyelinated axons. The nerve was injected with lysolecithin 17 days earlier and was cut 11 days after the injection. A: optical records taken at 5 successive sites, with the most proximal site as the origin, and distances indicated at the right of each sweep. The vertical calibration represents the fractional change in transmitted light intensity. Overall cut-off frequency, 0.22 kHz. B: conduction time (time to 50% peak) plotted versus distance for the records of (A). The average conduction velocity is 1 m/s. the dashed line gives the result that would be expected for uniform conduction in a normal axon of 15 m/s.
not only undergo the early stages of remyelination as demonstrated in electron micrographs, but also retain the ability to conduct at low velocity over considerable lengths. During remyelination of demyelinated but uncut nerves new nodes of Ranvier form in regions that formerly were internodal. At about 20 days post-injection these new nodes can be detected by light microscopy as gaps in the new myelin typically 5-30 /~m long. This sequence of events continued even if the nerve was transected at 14 days post-injection. In Fig. 7 we illustrate optical recording of conduction through one of these sites. The top trace represents the optical signal from a single new node and the lower record is the compound action potential (CAP). The first peak in the CAP results from axons that were not demyelinated by the lysolecithin. The remaining peaks represent remyelinating axons with slower conduction velocities through the lesion. The
OPT l
ms ]
CAP
ms
Fig. 5. Internodai ionic currents recorded from transected axons. A: an axon cut and then injected with lysolecithin 6 days prior to recording. The holding potential was hyperpolarized 23 mV from the resting level. Test depolarizations of 40, 60, 80, 100, 120, and 160 mV from the holding potential. B: an axon injected 16 days and cut 6 days prior to recording. The holding potential was hyperpolarized by 56 mV from rest. Test depolarizations of 80, 100, 120, 140, and 160 mV from the holding potential. In all records the uncompensated portions of the capacitative transient has been blanked. Pipette and bath solutions were frog Ringer's.
k
x t0 -5
4 ms
Fig. 7. Activity in a new node of Ranvier, formed in a transected axon. The nerve was 20 days post-injection and was cut 6 days before the experiment. The upper trace shows an optical signal from the node, and the lower trace is the compound action potential recorded from this branch. The overall cut-off frequency for both traces was 0.35 kHz.
310 optical signal falls within this latter class of remyelinating axons. In the experiment at 17 days (Fig. 6) the optical signal came from a group of fibers and originated both from regions that were destined to be new nodes and from internodal sites that were not yet significantly remyelinated. In Fig. 7 on the other hand, the recording was made from a single site that could be identified as a developing new node. Thus, new nodes form and become functional despite the transection. DISCUSSION The experiments presented in this paper were aimed at assessing the degree of participation of the neuronal soma in the events following myelin damage by lysolecithin. In one series of experiments nerves were transected just prior to lysolecithin injection and were examined 6 days later. We found that the appearance of macrophages within the lesion and the ensuing removal of myelin debris via phagocytosis were each comparable to that in uncut fibers. Demyelinated axons were mechanically more fragile in cut nerves than in uncut controls, but appeared similar under light and electron microscopy. In these experiments since the sciatic was cut prior to demyelination it would seem reasonable to conclude that the activation of macrophages was not dependent on signalling or synthesis originating in the soma. Further, we were able to record voltage-dependent Na + and K ÷ currents from the internodal axolemma of cut demyelinated axons, and the amplitude and kinetics of these currents were similar to those of earlier experiments on uncut fibers. Thus, these internodal channels did not arise as a result of somal synthesis stimulated by demyelination. We followed the events of functional recovery and remyelination over the 2 weeks following myelin removal (up to 22 days post-injection). In each case the axons were transected 6 days prior to the observations. By examining overlapping periods and comparing them to earlier results on uncut axons we are able to reach some conclusions about the control of the processes involved. At 6 to 8 days post-injection demyelinated zones are largely bare and conduction is blocked in almost all fibers. If nerves remain uncut, Schwann cells proliferate and bind to axons and, by about 12 days post-injection, the first few turns of new myelin are visible 26. Nerves transected at 6-8 days and examined 6 days later (12-14 days post-injection) followed a pattern indistinguishable from the uncut controls. Further, in uncut nerves up to 11 days post-injection, just prior to the earliest remyelination, conduction through the lesion is successful in only a small fraction of fibers, requiring some improvement in the extracellular insulation particularly at the transition
point between myelinated and demyelinated zones 26. Previous work from this laboratory on uncut axons has shown that in those fibers in which propagation is successful conduction through demyelinated segments is 'continuous' i.e. propagation occurred at constant velocity and involved internodal Na + channel activation 24'26. The similarity of the electrical and optical measurements made here to the earlier results suggests that the same process is active in transected fibers. By 17 days the probability of successful conduction has increased significantly and this can be demonstrated both electrically and optically in a large percentage of remyelinating axons 26. Nerves transected at 11 days and examined at 17 days again were identical in this respect to uncut controls. In the lysolecithin demyelinated lesion while the temporal progression of events is quite constant among all axons in the bundle, fibers vary considerably in the spatial extent of demyelination. Thus, while the compound action potential is useful in obtaining a profile of propagation in the cord, from that measurement alone one cannot rule out the possibility that conduction at 17 days is restored only in axons with a demyelinated region extending for only a few tens of micrometers. The optical techniques we have used allow a measurement of conduction in a single fiber or a group of just a few axons of uniform and observable morphology. Thus, we can judge in both control and cut axons that electrical signalling is restored over long lengths ( > 1 internode) in a high percentage of remyelinating axons at this time. Further, the conduction velocity within the lesioned zone is about 1 m/s in both cases. Thus, we conclude that separation of demyelinated axons from the neuronal soma at critical stages of the restorative process does not impede either structural or functional recovery. In uncut axons at about 3 weeks post-injection gaps in the new myelin can be seen in light microscopy. These regions represent the earliest stages of formation of new nodes of Ranvier, which develop at sites that previously were internodal. Following transection at 14 days, the appearance of these new nodes followed the same schedule, and were visible within the lesion 6 days later. Further, they were functional as judged by optically detected action potentials. We have previously shown that the gradient in Na + channel density at these new nodes is significantly less than that at the original nodes. Peak Na ÷ currents measured at new nodes were only about 3 times those at random internodal sites 24. Thus, the Na + channels populating these new nodes may arise as a result of a modest aggregation of internodal Na + channels. This hypothesis is strengthened by the results in this paper which suggest that at least during the 6 days just prior to node formation de novo synthesis of Na + channels is not required.
311 What can be said about the signals triggering early remyelination? In principle, proteins synthesized after the transection should be prevented from reaching the lesion. However, since in most of the above experiments the cut was not made earlier than 6 days post injection, it is possible that synthesis of some trophic substance was already complete. Cytosolic and cytoskeletal elements are carried via the slow axonal transport system, which has been measured in a variety of species at about 1 mm/day 18. Thus, molecules following this pathway require 20-30 days to reach the lesion and are therefore unlikely to constitute the trophic message. Fast axoplasmic transport in the frog occurs at 15 cm/day at 25 °C and can continue for at least a few hours following transection 1"17-19. This fast system is utilized by noncytosolic molecules such as membrane proteins and secreted compounds. Thus, these substances would reach the lesion in a few hours following synthesis. If trophic molecules of axonal origin following this path are active in promoting remyelination they must reach the lesion prior to Schwann cell proliferation and must remain active for up to 6 days. No replenishment would occur beyond a few hours following transection. Thus, while this scheme cannot be ruled out, for example, in the development of new nodes of Ranvier 3 weeks post-
injection, a simpler possibility is that the process of remyelination may be entirely under local control. Schwann cell proliferation may be stimulated by a variety of mitogens under different conditions 3,4,15. Of particular interest here are the recent experiments of Baichwal et al. 3 which suggest that a mitogenic factor may be released by macrophages that are actively phagocytizing myelin. Since in our 6 day nerves axons were transected prior to lysolecithin injection involvement of the soma in macrophage stimulation is unlikely. It has been suggested that Schwann cells may provide Na ÷ channels to the axolemma at nodes of Ranvier 27. More recently, evidence both supportive 5 and contradictory 25 to this hypothesis has appeared. There is, on the other hand, no evidence for Na ÷ channels in Schwann cells of X e n o p u s 24. The results of these experiments suggest that both an initial stimulus for demyelination and further events in the process of remyelination, including the formation of new nodes of Ranvier, can proceed in the absence of continuous communication with the neuronal cell body. Acknowledgements. Mrs. Ellen Brunschweiger provided excellent technical assistance during this study. This work was supported by Grants NS17965 from the National Institutes of Health and RG-1774 from the National Multiple Sclerosis Society.
REFERENCES 1 Abe, T., Haga, T. and Kurokawa, M., Rapid transport of phosphatidylecholine occurring simultaneously with protein transport in the frog sciatic nerve, Biochem. J., 136 (1973) 731-740. 2 Aguayo, A.T., Bunge, R.P., Duncan, I.D., Wood, P.M., Bray, G.M., Rat Schwann cells, cultured in vitro can ensheath axons regenerating in mouse nerves, Neurology, 29 (1979) 589. 3 Baichwal, R.R., Bigbee, J.W. and DeVries, G.H., Macrophagemediated myelin-related mitogenic factor for cultured Schwann cells, Proc. Natl. Acad. Sci. U.S.A., 85 (1988) 1701-1705. 4 Beuche, W.J. and Friede, R.L., The role of non-resident cells in Wallerian degeneration, J. Neurocytol., 13 (1984) 767-796. 5 Black, J.A., Waxman, S.G., Friedman, B., Elmer, L.W. and Angelides, K.J., Sodium channels in astrocytes of rat optic nerve in situ: immuno-electron microscopic studies, Glia, 2 (1989) 353-369. 6 Bostock, H. and Sears, T,A., The internodal axon membrane: electrical excitability and continuous conduction in segmental demyelination, J. Physiol., 280 (1978) 273-301. 7 Bray, G.M., Raminsky, M. and Aguayo, A.J., Interactions between axons and their sheath cells, Annu. Rev. Neurosci., 4 (1981) 127-162. 8 Bunge, R.P., Bunge, M.B. and Eldridge, C.F., Linkage between axonal ensheathment and basal lamina production by Schwann ceils, Annu. Rev. Neurosci., 9 (1986) 305-328. 9 Chiu, S.Y., Shrager, S. and Ritchie, J.M., Loose patch clamp recording of ionic currents in demyelinated frog nerve fibers, Brain Research, 359 (1985) 338-342. 10 Colquhoun, D. and Sigworth, E J. Fitting and statistical analysis of single-channel records. In B. A. Sakman and E. Neher (Eds.), Single Channel Recording, Plenum Press, New York, 1983, pp. 191-263. 11 Cragg, B.G. and Thomas, P. C., Changes in nerve conduction
12 13
14 15
16
17 18 19 20
21 22
in experimental allergic neuritis, J. Neurol. Neurosurg. Psychiatry, 27 (1964) 106-115. Duncan I.D., Aguayo A.J., Bunge, R.P. and Wood, P.M., Transplantation of rat Schwann cells grown in tissue culture into the mouse spinal cord, J. Neurosci., 49 (1980) 241-252. Hall, S.M. and Gregson, N.A., The in vivo and ultrastructural effects of injection of lysophosphatidyle choline into myelinated peripheral nerve fibers of the adult mouse, J. Cell Sci., 9 (1971) 769-789. Jacobs, J.M. and Cavanagh, J.B. Species differences in internode formation following two type of peripheral nerve injury, J. Anat., 105 (1969) 295-306. Le Beau, J.M., LaCarbiere, M. PoweU, H.C. Ellisman, M.H. and Schubert, D., Extracellular fluid conditioned during peripheral nerve regeneration stimulates Schwann cell adhesion, migration and proliferation, Brain Research, 459 (1988) 93-104. Lev-Ram, V. and Grinvald A., Ca ++- and K÷-dependent communication between central nervous system myelinated axons and oligodendrocytes revealed by voltage-sensitive dyes, Proc. Natl. Acad. Sci. U.S.A., 83 (1986) 6651-6655. Ochs, S., Retention and redistribution of proteins in mammalian nerve fibers, J. Physiol., 253 (1975) 459-475. Ochs, S., Axoplasmatic Transport and its Relation to Other Nerve Functions, Wiley, 1982, pp. 11-47. Ochs, S. and Ramish, S., Characteristics of the fast transport system in mammalian nerve fibers, J. Neurobiol., 1 (1969) 247-261. Pellegrino, R.G., Politis, M.J., Ritchie, J.M.and Spencer, P.S., Events in degenerating cat peripheral nerve: induction of Schwann cell S phage and its relation to nerve fiber degeneration, J. Neurocytol., 15 (1986) 17-28. Peracchia, C. and Mittler, B.S., Fixation by means of glutaraldehyde-hydrogen peroxide reaction products, J. Cell. Biol., 53 (1972) 234-238. Rubinstein, C.T. and Shrager, P., A study of remyelination in
312
23 24 25 26 27
the transected frog peripheral nerve, Biophys. J., 57, 2 (1990) 131a. Shrager, P., The distribution of sodium and potassium channels in single demyelinated axons of the frog, J. Physiol., 392 (1987) 587-602. Shrager, P., Ionic channels and signals conduction in single remyelinating frog nerve fibers, J. Physiol., 404 (1988) 695-712. Shrager, P., Sodium Channels in single demyelinated mammalian axons, Brain Research, 483 (1989) 149-154. Shrager, P. and Rubinstein C.T., Optical measurements of conduction in single demyelinating axons, J. Gen. Physiol., (1990) in press. Shrager, S. Chiu, S.Y. and Ritchie, J.M., Voltage-dependent sodium and potassium channels in mammalian cultured Schwann
cells, Proc. Natl. Acad. Sci. U.S.A., 82 (1985) 948-952. 28 Stuhmer, W., Roberts, W.M. and Almers W., The loose patch clamp. In B. Sakman and E. Neher (Eds.), Single Channel Recording, Plenum, New York 1983, pp 123-132. 29 Wang G., The long-term excitability of myelinated nerve fibers in the transected frog sciatic nerve, J. Physiol., 368 (1985) 309-321. 30 Waxman, S.G. Biophysical mechanisms of impulse conduction in demyelinating axons, Adv. Neurol., 47 (1988) 185-213. 31 Waxman, S.G. Rules governing membrane reorganization and axon-glial interactions during the development of myelinated fibers. In F.J. Seil, E. Herbert and B.M. Carlson (Eds.), Progress in Brain Research, Vol. 71, Elsevier, Amsterdam, 1987, pp. 121-141.
303
BRES 15760
Remyelination of nerve fibers in the transected frog sciatic nerve Chaim T. Rubinstein and Peter Shrager Department of Physiology, University of Rochester Medical Center, Rochester, NY 14642 (U.S.A ) (Accepted 20 February 1990) Key words: Remyelination; Schwann cell; Demyelination; Myelinated nerve; Degeneration
This project tests an important aspect of the cellular events controlling the processes of recovery of function and remyelination that follow demyelination in the peripheral nervous system. Frog sciatic nerves have been shown to survive and remain functional for up to 10 days following transection29. We have utilized this property in order to dissociate the recovery process from possible control by the neuronal soma. Xenopus sciatic nerves were demyelinated in one branch by an intraneural injection of lysolecithin. The nerve was cut proximally to the injection site either immediately before, or several days after the lysolecithin injection. Recovery of function and remyelination were then followed by electrophysioiogical, optical, and ultrastructural techniques applied both to whole branches and single fibers. Controls included the cut but uninjected branch, and injected but uncut nerves. The progression of events during both demyelination and recovery in cut axons was indistinguishable from that in uncut fibers. This suggests that this process may be under local control and can be initiated and carried out in the absence of constant communication with the nerve cell body. INTRODUCTION When axons of the peripheral nervous system are demyelinated conduction is at first blocked, but later recovers. The sequence of events involved in this process has been studied extensively both morphologically and physiologically6-8'23'24'3°'31. Damaged myelin is removed via phagocytosis by macrophages and Schwann cells. Proliferating Schwann cells are then observed to associate with the axons and surround them with numerous processes, ultimately resulting in remyelination. This procedure results in unusually short internodes as new nodes of Ranvier are formed in regions that previously were internodal. Electrical signalling at very low conduction velocity is seen in only a very small percentage of fibers prior to remyelination 6"24"26. The fraction of axons conducting through the lesion increases rapidly as early remyelination progresses, though long delays in propagation remain. Much work has been done to elucidate the cellular processes controlling both myelin removal and the subsequent recovery through Schwann cell proliferation and remyelination, though many interesting questions remain. Conditioned medium from cultured peritoneal macrophages that have phagocytosed a myelin membrane fraction is mitogenic for cultured Schwann cells 3. A study of Wallerian degeneration, in which the degenerating nerve was isolated from interaction with non-local cells
showed no proliferation of Schwann cells and no evidence for their role in the removal of myelin 4. When macrophages were allowed to migrate into the experimental chamber the macrophages were shown to attack selectively the myelin sheath rejected by the Schwann cells but not the Schwann cells themselves. The activity of the phagocytes was mediated by a signal which diminished in intensity with time, originating, presumably from the neuron. Similarly, Schwann cell proliferation, migration and adhesion can be stimulated by extracellular fluid surrounding regenerating peripheral nerve 14. When normal Schwann cells grown in tissue culture are transplanted into the demyelinated peripheral nerves or the demyelinated spinal cords of mutant quaking mice they will ensheathe and remyelinate these axons 2'12. Pellegrino et al., suggested that Schwann cell mitosis during peripheral nerve degeneration is due to a signal coming from the axon 2°. These studies suggest that control of both myelin removal and remyelination are complex and indicate that under some circumstances there is a neuronal signal required for ensheathment and myelination. Does this signal arise from an endogenous factor in the axon or does it require the synthesis of some substance by the cell body? In this report we test one aspect of the control process, the possible involvement of neuronal factors derived from the soma. Wang 29 has shown that in the frog the distal stump of sciatic nerves can survive at room
Correspondence: C.T. Rubinstein, Department of Physiology, Box 642, University of Rochester Medical Center, 601 Elmwood Avenue, Rochester, NY 14642, U.S.A. 0006-8993/90/$03.50 © 1990 Elsevier Science Publishers B.V. (Biomedical Division)
304 temperature for up to 10 days following transection (and even longer at cooler temperatures). We have previously studied both structural and functional aspects of demyelination and remyelination in Xenopus peripheral nerve 2324"26. In the present work we examine these processes following both experimental demyelination and transection to separate the lesioned area from the n e u r o n a l cell body. It is shown that events in the distal nerve stump are similar in most respects to those in uncut control axons. A brief report of this work has appeared in abstract form 22.
for each record, stimulating the nerve at 10 Hz. During analysis signals were further filtered using a Gaussian routine 1° and are plotted as the fractional change in transmitted light intensity. The cut-off frequency given for each experiment represents the overall (analog+digital) filtering. Optical measurements were made both on single fibers and on small tracts of axons of uniform morphology. More complete details of this system are given in Shrager and Rubinstein26.
Electron microscopy
MATERIALS AND METHODS
Immediately following either electrophysiological or optical measurements nerves were fixed in 3% glutaraldehyde-H20 2 buffered to pH 7.4 with 0.1 M cacodylate~I. The preparation was then further fixed, stained, dehydrated, embedded, sectioned, and re-stained as detailed in Shragerz4. Nerves were examined in cross-section only which resulted in a very low probability of observing nodes. Thus, we did not address the question of possible changes in membrane specialization at these sites. Thin sections were examined on a Zeiss 10A electron microscope.
Surgical procedures and electrophysiological recording Nerves from the clawed frog Xenopus laevis were demyelinated
RESULTS
with an intraneural injection of lysolecithin~3. Animals were anesthetized in tricaine methanesulfonate (3.7 g/l H20 ). The sciatic nerve in one leg was surgically exposed and 1/~1 of 1% lysolecithin in sterile frog Ringer's was injected into either the tibial or peroneal branch in the thigh via a glass micropipette. The uninjected branch served as a control. The incision was sutured and the animal was allowed to recover. At various times post-surgery frogs were similarly anesthetized and the sciatic plexus in the injected leg was cut with fine dissection scissors about 5-7 mm from the spine. All branches of the sciatic were transected. The proximal end of the distal stump was ligated with thread and the wound was sutured. In one group of frogs the transection and the injection were both done during the same surgery, with the transection preceding the demyelination. Frogs were maintained in aquaria at room temperature and were fed ground beef twice weekly. Six days after cutting the sciatic nerve, frogs were anesthetized in ice water and euthanized by decapitation. The experimental sciatic nerve was dissected and mounted in a chamber containing platinum wires for external stimulation and recording. The nerve was positioned with the tibial and peroneal branches in contact with separate pairs of wires so that activity from each branch could be measured independently. Compound action potentials were recorded in response to a supramaximal shock, from several locations along each branch. The nerve was then desheathed and was dissociated into single fibers by coilagenase (3.5 mg/ml for 1.5 h). Axons could then be gently spread from the bundle over a few millimeters for observation and patch clamping. The loose patch clamp was designed using a modified version of the active bridge circuit of Stuhmer et al.28. Details of this system are similar to those reported previously23-25. The design allows for automatic compensation for both the pipette resistance and the seal resistance. Seal resistances were generally in the range 6-15 MI2. Linear leakage and capacitative currents were subtracted by both analog a0d digital means as described earlierz4. Pulse protocols and data analysis were carried out with the aid of a laboratory computer. The composition of frog Ringer solution was (raM): NaCI, 115; KCI, 2.5; CaCI2, 1.8; HEPES, 5; pH 7.4. All experiments were run at room temperature.
Optical recording of action potentials Nerves were prepared exactly as for loose patch clamp recording except that the potential sensitive dye RH15516 (2 mg/ml) was added to the collagenase solution. A Nikon Diaphot microscope was modified to increase the intensity of illumination, and the incident beam was filtered to 705 + 25 rim. Light transmitted through the axon was focussed on a photodiode connected to a currentto-voltage converter of variable gain. The amplified signals were filtered and digitally sampled. Typically, 64 sweeps were averaged
Electron microscopy The morphology of lysolecithin-induced demyelination and remyelination in otherwise intact nerve was first described by Hall and Gregson 13 in the mouse and more recently in Xenopus by Shrager 23'24. Within the first 3 days after injection debris-covered breaks appear in the myelin. Over the next few days most of this myelin debris is cleared by macrophages and Schwann cells via phagocytosis. Proliferating Schwann cells then bind to the demyelinated regions, extend n u m e r o u s processes, and remyelinate the fiber with the first few lamellae appearing at about 12 days post-injection26. The sequence of events in transected nerves was found to be very similar to that found in uncut nerves. Fig. 1 shows light micrographs of axons from transected sciatic nerves. Fig 1B illustrates a debris-covered break in the myelin of an axon that had been both cut and injected 6 days earlier. In Fig. 1A the damaged myelin has been removed over a short length exposing a bare stretch of axon, also 6 days post-injection. A t 14 days long demyelinated zones can be seen, and in some axons (Fig. 1C) both macrophages (large dark cells) and Schwann cells (smaller clear cells) are associated with the axolemma. In other fibers (Fig. 1D) macrophages are largely gone and Schwann cells are more a b u n d a n t and are beginning to extend processes and the first few lamellae of myelin around the axon (see below). In Fig. 1F we see an axon with small protrusions characteristic of new, thin myelin. This cell is 24 days post-injection. At 31 days the new myelin sheath is further developed, and a new node of R a n v i e r is clearly visible (arrowhead - Fig. 1E). New nodes are characteristic of remyelinating axons, forming short internodes along the remyelinated area 11A4"23. The process of remyelination was examined in transected axons up to 31 days from the time of the initial
305 d e m y e l i n a t i n g injection, We found no significant difference in the time course o f events during the demyelination and r e m y e l i n a t i o n o f the cut nerves as c o m p a r e d to control uncut axons. In o r d e r to obtain a m o r e detailed c o m p a r i s o n b e t w e e n the cut and uncut axons we examined thin sections of cut axons with an electron microscope, c o u n t e d the n u m b e r of myelin layers and com-
p a r e d them with d a t a of uncut remyelinating nerves 23' 24,26. Some r e p r e s e n t a t i v e fibers are shown in Fig. 2, and the results are s u m m a r i z e d in the histograms of Fig. 3. In Fig. 2 A and B we see axons at two stages of demyelination, 6 days post-surgery (and post-cut). In Fig, 2 A the myelin has l o o s e n e d and a m a c r o p h a g e (top) is removing it by phagocytosis. T h e fiber in Fig. 2B has
Fig 1. Light micrographs of unstained living axons. All axons have been cut 6 days prior to the experiment. A: a short break in the myelin sheath exposing a bare segment of an axon with no proliferating Schwann cells, 6 days post-injection. B: a debris-covered break in the myelin sheath, 6 days post-injection. C: macrophages (arrows) attached to a demyelinated segment of an axon at scattered sites, 14 days post-injection. D: a long demyelinated axon segment with adherent Schwann cells (arrowheads), 14 days post-injection. E: a new node of Ranvier (arrowhead) in a remyelinating fiber 31 days post-injection. F: a remyelinating axon 24 days post-surgery. Some attached Schwann cells (arrowheads) can be seen. Scale bars, 10/~m.
306
Fig. 2. Electron micrographs of thin sections of cut demyelinated axons. A: a macrophage (top of figure) associated with an axon that is surrounded by loose myelin, 6 days post-injection. B: a completely demyelinated axon with all myelin debris cleared away, 6 days post-injection. C: a remyelinating axon with one and a half turns of the Schwann cell, 14 days post-surgery. Inset: enlarged view of the outer Sehwann cell lip. D: a remyelinating axon with 5 layers of myelin, 14 days post-injection. Inset: The new, uncompacted myelin at higher magnification. E: 17 days post-injection. A remyelinating axon with 11 layers of new myelin. Inset: An enlarged view of the myelin layers. Scales, 1/tm except for the insets which are 0.1 ~m.
307 been totally demyelinated. Demyelination is largely complete in many axons at 6 days in injected but uncut fibers as well 24. During the second week post-injection proliferating Schwann cells appear throughout the lesion, adhere to the demyelinated axolemma, and begin the process of remyelination. In Fig. 2C a Schwann cell has completed one full turn of new myelin around a cut axon, 14 days post-injection. The cut axon in Fig. 2D, also 14 days after lysolecithin exposure, has 5 lines of new myelin. Fig. 2E shows a cut fiber 20 days post-injection with 11 lines of new myelin. The ultrastructural results are summarized in the histograms of Fig. 3. The label 'N' on the abscissa denotes naked axons. The label 'L' represents loose myelin. The numbers on the abscissa give the number of turns of Schwann cells and myelin counted, '1' denoting one complete turn of the Schwann ceil. A total of 382 cells were counted. The results indicate that the first few turns of new myelin appear between 1 and 2 weeks post-injection. The histograms show a gradual shift to fibers with > 6 turns over the third week. These results are very similar to those observed earlier in injected but uncut axons 26. In these latter experiments the first significant remyelination was seen 12 days post-injection, and the degree of remyelination
then increased steadily over the next several days.
Electrophysiology Compound action potentials were recorded separately from the injected and uninjected branches of the transected sciatic nerve (we varied the lysolecithin injection between the tibial and peroneal branches in different experiments). The uninjected branch served as a control for Wallerian degeneration due to the transection. In addition, following recording, nerves were desheathed, dissociated with collagenase and viewed with a light microscope for signs of degeneration. The criterion for a degenerated nerve was a recorded compound action potential from a control branch smaller than 500/~V, and in nerves failing this test most of the axons invariably had the characteristic appearance of Wallerian degeneration. These preparations were discarded. Over the course of this work we found that in about 55% of the frogs used, excessive degeneration had occurred. In principle this could have been improved by maintaining the animals at cooler temperatures 29 but since we had extensive studies on uncut Xenopus at 20-25 °C, we continued with the same conditions. Loss due to degeneration was greatest between January and May and thus may have had a
Z I00 6 days
14 days
n = 162
n = 72
i°°f
50
0
~
L
N
t
i
i
i
t
i
i
2
3
4
5
6
>6
IO0 17
L
N
t
2
3
4
100
days
20
n = 69
5
6
>6
6
>6
days
n = 79
50
50
L
N
i
2
MYELIN
3
4
LAYERS
5
6
>6
L
N
i
2
MYELIN
3
4
5
LAYERS
Fig. 3. Histograms illustrating the development of remyelination from 6 to 20 days post-injection. The abscissa shows the degree of remyelination. 'L' indicates loose myelin. 'N' indicates naked (completely demyelinated) axons. The numbers give the number of turns of Schwann cell and myelin. The ordinate plots the percentage of cells in each state. A total of 382 cells were counted. A: 6 days post-injection (162 ceils). B: 14 days post-injection (72 cells). C: 17 days post-injection (69 cells). D: 20 days post-injection (79 cells).
308 seasonal component. Extensive degeneration appeared to occur in an approximately all-or-none manner. Of the 15 frogs that passed the test and were studied extensively, one had a compound action potential of 700 #V, and the others were all above 1500 #V. However, even in those preparations that retained strong electrical activity, stimulating the transected nerve close to the cut failed to elicit a compound action potential due to degeneration, forcing us to stimulate the axon 3 - 4 mm more distal. In addition, a small percentage of fibers within a given nerve were seen to be degenerated. C o m p o u n d action potentials recorded from transected nerves over a range of times following injection of lysolecithin are shown in Fig. 4. The numbers at the right give the time post-injection. The top trace is from a cut but uninjected nerve. The stimulus artifact has been blanked and the large signal recorded with a short conduction time is close to that expected for a completely normal nerve. The second trace shows an amplified view of the former record. The vertical lines at the left indicate a scale of 1 0 0 # V at the recording wires. Note that signals due to very slowly conducting fibers are absent in this uninjected nerve. Slow signals are likewise not seen at 6 or 12 days post-injection. In these 2 records both the stimulus artifact and the comparatively large signal from
I
0O
HV UNINJ. ~
"~
UNINJ.
~
6 DAYS
~
12
DAYS
17
DAYS
24
DAYS
1
MS
Fig. 4. Compound action potentials recorded from cut, uninjected nerves (control) and from cut, injected (demyelinated) nerves over a range of times following injection of lysolecithin. All nerves were transected 6 days prior to the experimental day. The scale bar at the left of each trace represents 100 pV at the recording electrodes. Most of the stimulus artifact has been blanked from all sweeps and the initial downward deflection indicates the time of stimulation. The top 2 sweeps show control records from a cut, uninjected nerve. The numbers at the right of each of the lower 4 traces give the days post-injection of lysolecithin. In all but the top trace the fast components of the action potentials (representing the fraction of axons that remained myelinated following the lysolecithin injection) have been blanked.
axons that were never demyelinated by the lysolecithin were blanked before plotting. In general, a small fraction of axons in the bundle remains myelinated with the injection technique used. At 17 days the compound action potential has a significant contribution from fibers conducting with a wide range of very slow velocities. Note that the conduction delays seen are generated in a lesion that is just 2-3 mm in length. At 24 days the number of slowly conducting axons has increased and their average velocity has also increased. In nerves that were injected with lysolecithin but were not cut the profile of conduction velocities seen in compound action potentials was very similar to these results 24. In this earlier work up to 12 days post-injection almost no demyelinated fibers conducted successfully through the lesion. At 17 days post-injection numerous axons were found to conduct, but with a wide spectrum of propagation times, involving delays of about 4 ms. The percentage of remyelinating axons that conduct successfully through the 2-3 mm lesion increased significantly by 30 days and the pattern of a single broad peak was very similar to that seen here at 24 days. In earlier work we have measured ionic currents from the demyelinated axolemma using the loose patch clamp technique 23-25. At internodal sites currents due to voltage-dependent Na ÷ and K ÷ channels were recorded. In Xenopus axons, the average peak inward Na ÷ current was about 750 p A in the range of 6 to 54 days post-injection. In the present experiments these measurements were more difficult due to the apparent increased fragility of the transected plus demyelinated axons. Nonetheless, in some cases adequate sealing of the pipette to the surface membrane was possible. Fig. 5 shows internodal ionic currents recorded at 6 (left) and 16 (right) days post-surgery. The transient inward currents have kinetics and voltage dependence similar to those previously measured at frog nodes and internodes and identified as Na + currents 9'23-25. The peak amplitudes of these currents are within the range of values previously obtained in uncut axons (225-1450 pA). The delayed outward currents are similar to those identified earlier as originating in K ÷ channels on the basis of pharmacological block by tetraethylammonium ion and 4-aminopyridine. Thus these results demonstrate the existence of Na ÷ and K ÷ channels in the internodal membrane of cut axons. The source of the channels in the axon at 6 days post-surgery could not be from proteins newly synthesized in the cell soma in response to the demyelination, since the axon was severed from the soma prior to the injection of lysolecithin.
Optical recording of action potential conduction In uncut axons, at 11 days post-injection proliferating
309 Schwann cells associate with demyelinated axons, but no new myelin is yet present. Conduction across long stretches of the lesion is seen in only a very small fraction of the fibers since the internodal Na + channels alone are insufficient to bridge the impedance mismatch at the transition point between myelinated and demyelinated regions. By 17 days post-injection axons remain covered by numerous Schwann cells and while in electron micrographs regions can be seen that are covered by several lamellae of new myelin, it is not possible to identify presumptive new nodes of Ranvier in the living cell. At this early stage of remyelination the probability of successful conduction increases markedly, though at low velocity26. This pattern is unaltered if the sciatic nerve is cut at the 11 day point. Fig. 6 shows conduction through a tract of just a few demyelinated fibers of uniform appearance measured optically in a preparation 17 days post-injection, cut at 11 days. The records represent the fractional change in transmitted light intensity. The amplitude of the signals is proportional to the change in membrane potential, but also depends on the area of excitable membrane seen by the photodiode and on the degree of dye binding. Thus, the primary information is in the shape and timing of the signals. The graph in Fig. 6B plots conduction time versus distance for these signals. The average conduction velocity (inverse slope) was about 1 m/s through the lesion, compared with 15-20 m/s in normal axons (dashed line). Similar results were obtained at 20 and 22 days post-injection (cut at 14 and 16 days, respectively). Thus, despite transection, axons
A
B
mm
A
B O.OO
2
0.25
0.51 H 1 1.52
,~._/v,,~,,,..~,.~ 2.o3 i
x 1O - 5
O0
i
DISTANCE
4 ms
i
t
i
2 (mm}
Fig. 6. Optical measurement of conduction through a small group of demyelinated axons. The nerve was injected with lysolecithin 17 days earlier and was cut 11 days after the injection. A: optical records taken at 5 successive sites, with the most proximal site as the origin, and distances indicated at the right of each sweep. The vertical calibration represents the fractional change in transmitted light intensity. Overall cut-off frequency, 0.22 kHz. B: conduction time (time to 50% peak) plotted versus distance for the records of (A). The average conduction velocity is 1 m/s. the dashed line gives the result that would be expected for uniform conduction in a normal axon of 15 m/s.
not only undergo the early stages of remyelination as demonstrated in electron micrographs, but also retain the ability to conduct at low velocity over considerable lengths. During remyelination of demyelinated but uncut nerves new nodes of Ranvier form in regions that formerly were internodal. At about 20 days post-injection these new nodes can be detected by light microscopy as gaps in the new myelin typically 5-30 /~m long. This sequence of events continued even if the nerve was transected at 14 days post-injection. In Fig. 7 we illustrate optical recording of conduction through one of these sites. The top trace represents the optical signal from a single new node and the lower record is the compound action potential (CAP). The first peak in the CAP results from axons that were not demyelinated by the lysolecithin. The remaining peaks represent remyelinating axons with slower conduction velocities through the lesion. The
OPT l
ms ]
CAP
ms
Fig. 5. Internodai ionic currents recorded from transected axons. A: an axon cut and then injected with lysolecithin 6 days prior to recording. The holding potential was hyperpolarized 23 mV from the resting level. Test depolarizations of 40, 60, 80, 100, 120, and 160 mV from the holding potential. B: an axon injected 16 days and cut 6 days prior to recording. The holding potential was hyperpolarized by 56 mV from rest. Test depolarizations of 80, 100, 120, 140, and 160 mV from the holding potential. In all records the uncompensated portions of the capacitative transient has been blanked. Pipette and bath solutions were frog Ringer's.
k
x t0 -5
4 ms
Fig. 7. Activity in a new node of Ranvier, formed in a transected axon. The nerve was 20 days post-injection and was cut 6 days before the experiment. The upper trace shows an optical signal from the node, and the lower trace is the compound action potential recorded from this branch. The overall cut-off frequency for both traces was 0.35 kHz.
310 optical signal falls within this latter class of remyelinating axons. In the experiment at 17 days (Fig. 6) the optical signal came from a group of fibers and originated both from regions that were destined to be new nodes and from internodal sites that were not yet significantly remyelinated. In Fig. 7 on the other hand, the recording was made from a single site that could be identified as a developing new node. Thus, new nodes form and become functional despite the transection. DISCUSSION The experiments presented in this paper were aimed at assessing the degree of participation of the neuronal soma in the events following myelin damage by lysolecithin. In one series of experiments nerves were transected just prior to lysolecithin injection and were examined 6 days later. We found that the appearance of macrophages within the lesion and the ensuing removal of myelin debris via phagocytosis were each comparable to that in uncut fibers. Demyelinated axons were mechanically more fragile in cut nerves than in uncut controls, but appeared similar under light and electron microscopy. In these experiments since the sciatic was cut prior to demyelination it would seem reasonable to conclude that the activation of macrophages was not dependent on signalling or synthesis originating in the soma. Further, we were able to record voltage-dependent Na + and K ÷ currents from the internodal axolemma of cut demyelinated axons, and the amplitude and kinetics of these currents were similar to those of earlier experiments on uncut fibers. Thus, these internodal channels did not arise as a result of somal synthesis stimulated by demyelination. We followed the events of functional recovery and remyelination over the 2 weeks following myelin removal (up to 22 days post-injection). In each case the axons were transected 6 days prior to the observations. By examining overlapping periods and comparing them to earlier results on uncut axons we are able to reach some conclusions about the control of the processes involved. At 6 to 8 days post-injection demyelinated zones are largely bare and conduction is blocked in almost all fibers. If nerves remain uncut, Schwann cells proliferate and bind to axons and, by about 12 days post-injection, the first few turns of new myelin are visible 26. Nerves transected at 6-8 days and examined 6 days later (12-14 days post-injection) followed a pattern indistinguishable from the uncut controls. Further, in uncut nerves up to 11 days post-injection, just prior to the earliest remyelination, conduction through the lesion is successful in only a small fraction of fibers, requiring some improvement in the extracellular insulation particularly at the transition
point between myelinated and demyelinated zones 26. Previous work from this laboratory on uncut axons has shown that in those fibers in which propagation is successful conduction through demyelinated segments is 'continuous' i.e. propagation occurred at constant velocity and involved internodal Na + channel activation 24'26. The similarity of the electrical and optical measurements made here to the earlier results suggests that the same process is active in transected fibers. By 17 days the probability of successful conduction has increased significantly and this can be demonstrated both electrically and optically in a large percentage of remyelinating axons 26. Nerves transected at 11 days and examined at 17 days again were identical in this respect to uncut controls. In the lysolecithin demyelinated lesion while the temporal progression of events is quite constant among all axons in the bundle, fibers vary considerably in the spatial extent of demyelination. Thus, while the compound action potential is useful in obtaining a profile of propagation in the cord, from that measurement alone one cannot rule out the possibility that conduction at 17 days is restored only in axons with a demyelinated region extending for only a few tens of micrometers. The optical techniques we have used allow a measurement of conduction in a single fiber or a group of just a few axons of uniform and observable morphology. Thus, we can judge in both control and cut axons that electrical signalling is restored over long lengths ( > 1 internode) in a high percentage of remyelinating axons at this time. Further, the conduction velocity within the lesioned zone is about 1 m/s in both cases. Thus, we conclude that separation of demyelinated axons from the neuronal soma at critical stages of the restorative process does not impede either structural or functional recovery. In uncut axons at about 3 weeks post-injection gaps in the new myelin can be seen in light microscopy. These regions represent the earliest stages of formation of new nodes of Ranvier, which develop at sites that previously were internodal. Following transection at 14 days, the appearance of these new nodes followed the same schedule, and were visible within the lesion 6 days later. Further, they were functional as judged by optically detected action potentials. We have previously shown that the gradient in Na + channel density at these new nodes is significantly less than that at the original nodes. Peak Na ÷ currents measured at new nodes were only about 3 times those at random internodal sites 24. Thus, the Na + channels populating these new nodes may arise as a result of a modest aggregation of internodal Na + channels. This hypothesis is strengthened by the results in this paper which suggest that at least during the 6 days just prior to node formation de novo synthesis of Na + channels is not required.
311 What can be said about the signals triggering early remyelination? In principle, proteins synthesized after the transection should be prevented from reaching the lesion. However, since in most of the above experiments the cut was not made earlier than 6 days post injection, it is possible that synthesis of some trophic substance was already complete. Cytosolic and cytoskeletal elements are carried via the slow axonal transport system, which has been measured in a variety of species at about 1 mm/day 18. Thus, molecules following this pathway require 20-30 days to reach the lesion and are therefore unlikely to constitute the trophic message. Fast axoplasmic transport in the frog occurs at 15 cm/day at 25 °C and can continue for at least a few hours following transection 1"17-19. This fast system is utilized by noncytosolic molecules such as membrane proteins and secreted compounds. Thus, these substances would reach the lesion in a few hours following synthesis. If trophic molecules of axonal origin following this path are active in promoting remyelination they must reach the lesion prior to Schwann cell proliferation and must remain active for up to 6 days. No replenishment would occur beyond a few hours following transection. Thus, while this scheme cannot be ruled out, for example, in the development of new nodes of Ranvier 3 weeks post-
injection, a simpler possibility is that the process of remyelination may be entirely under local control. Schwann cell proliferation may be stimulated by a variety of mitogens under different conditions 3,4,15. Of particular interest here are the recent experiments of Baichwal et al. 3 which suggest that a mitogenic factor may be released by macrophages that are actively phagocytizing myelin. Since in our 6 day nerves axons were transected prior to lysolecithin injection involvement of the soma in macrophage stimulation is unlikely. It has been suggested that Schwann cells may provide Na ÷ channels to the axolemma at nodes of Ranvier 27. More recently, evidence both supportive 5 and contradictory 25 to this hypothesis has appeared. There is, on the other hand, no evidence for Na ÷ channels in Schwann cells of X e n o p u s 24. The results of these experiments suggest that both an initial stimulus for demyelination and further events in the process of remyelination, including the formation of new nodes of Ranvier, can proceed in the absence of continuous communication with the neuronal cell body. Acknowledgements. Mrs. Ellen Brunschweiger provided excellent technical assistance during this study. This work was supported by Grants NS17965 from the National Institutes of Health and RG-1774 from the National Multiple Sclerosis Society.
REFERENCES 1 Abe, T., Haga, T. and Kurokawa, M., Rapid transport of phosphatidylecholine occurring simultaneously with protein transport in the frog sciatic nerve, Biochem. J., 136 (1973) 731-740. 2 Aguayo, A.T., Bunge, R.P., Duncan, I.D., Wood, P.M., Bray, G.M., Rat Schwann cells, cultured in vitro can ensheath axons regenerating in mouse nerves, Neurology, 29 (1979) 589. 3 Baichwal, R.R., Bigbee, J.W. and DeVries, G.H., Macrophagemediated myelin-related mitogenic factor for cultured Schwann cells, Proc. Natl. Acad. Sci. U.S.A., 85 (1988) 1701-1705. 4 Beuche, W.J. and Friede, R.L., The role of non-resident cells in Wallerian degeneration, J. Neurocytol., 13 (1984) 767-796. 5 Black, J.A., Waxman, S.G., Friedman, B., Elmer, L.W. and Angelides, K.J., Sodium channels in astrocytes of rat optic nerve in situ: immuno-electron microscopic studies, Glia, 2 (1989) 353-369. 6 Bostock, H. and Sears, T,A., The internodal axon membrane: electrical excitability and continuous conduction in segmental demyelination, J. Physiol., 280 (1978) 273-301. 7 Bray, G.M., Raminsky, M. and Aguayo, A.J., Interactions between axons and their sheath cells, Annu. Rev. Neurosci., 4 (1981) 127-162. 8 Bunge, R.P., Bunge, M.B. and Eldridge, C.F., Linkage between axonal ensheathment and basal lamina production by Schwann ceils, Annu. Rev. Neurosci., 9 (1986) 305-328. 9 Chiu, S.Y., Shrager, S. and Ritchie, J.M., Loose patch clamp recording of ionic currents in demyelinated frog nerve fibers, Brain Research, 359 (1985) 338-342. 10 Colquhoun, D. and Sigworth, E J. Fitting and statistical analysis of single-channel records. In B. A. Sakman and E. Neher (Eds.), Single Channel Recording, Plenum Press, New York, 1983, pp. 191-263. 11 Cragg, B.G. and Thomas, P. C., Changes in nerve conduction
12 13
14 15
16
17 18 19 20
21 22
in experimental allergic neuritis, J. Neurol. Neurosurg. Psychiatry, 27 (1964) 106-115. Duncan I.D., Aguayo A.J., Bunge, R.P. and Wood, P.M., Transplantation of rat Schwann cells grown in tissue culture into the mouse spinal cord, J. Neurosci., 49 (1980) 241-252. Hall, S.M. and Gregson, N.A., The in vivo and ultrastructural effects of injection of lysophosphatidyle choline into myelinated peripheral nerve fibers of the adult mouse, J. Cell Sci., 9 (1971) 769-789. Jacobs, J.M. and Cavanagh, J.B. Species differences in internode formation following two type of peripheral nerve injury, J. Anat., 105 (1969) 295-306. Le Beau, J.M., LaCarbiere, M. PoweU, H.C. Ellisman, M.H. and Schubert, D., Extracellular fluid conditioned during peripheral nerve regeneration stimulates Schwann cell adhesion, migration and proliferation, Brain Research, 459 (1988) 93-104. Lev-Ram, V. and Grinvald A., Ca ++- and K÷-dependent communication between central nervous system myelinated axons and oligodendrocytes revealed by voltage-sensitive dyes, Proc. Natl. Acad. Sci. U.S.A., 83 (1986) 6651-6655. Ochs, S., Retention and redistribution of proteins in mammalian nerve fibers, J. Physiol., 253 (1975) 459-475. Ochs, S., Axoplasmatic Transport and its Relation to Other Nerve Functions, Wiley, 1982, pp. 11-47. Ochs, S. and Ramish, S., Characteristics of the fast transport system in mammalian nerve fibers, J. Neurobiol., 1 (1969) 247-261. Pellegrino, R.G., Politis, M.J., Ritchie, J.M.and Spencer, P.S., Events in degenerating cat peripheral nerve: induction of Schwann cell S phage and its relation to nerve fiber degeneration, J. Neurocytol., 15 (1986) 17-28. Peracchia, C. and Mittler, B.S., Fixation by means of glutaraldehyde-hydrogen peroxide reaction products, J. Cell. Biol., 53 (1972) 234-238. Rubinstein, C.T. and Shrager, P., A study of remyelination in
312
23 24 25 26 27
the transected frog peripheral nerve, Biophys. J., 57, 2 (1990) 131a. Shrager, P., The distribution of sodium and potassium channels in single demyelinated axons of the frog, J. Physiol., 392 (1987) 587-602. Shrager, P., Ionic channels and signals conduction in single remyelinating frog nerve fibers, J. Physiol., 404 (1988) 695-712. Shrager, P., Sodium Channels in single demyelinated mammalian axons, Brain Research, 483 (1989) 149-154. Shrager, P. and Rubinstein C.T., Optical measurements of conduction in single demyelinating axons, J. Gen. Physiol., (1990) in press. Shrager, S. Chiu, S.Y. and Ritchie, J.M., Voltage-dependent sodium and potassium channels in mammalian cultured Schwann
cells, Proc. Natl. Acad. Sci. U.S.A., 82 (1985) 948-952. 28 Stuhmer, W., Roberts, W.M. and Almers W., The loose patch clamp. In B. Sakman and E. Neher (Eds.), Single Channel Recording, Plenum, New York 1983, pp 123-132. 29 Wang G., The long-term excitability of myelinated nerve fibers in the transected frog sciatic nerve, J. Physiol., 368 (1985) 309-321. 30 Waxman, S.G. Biophysical mechanisms of impulse conduction in demyelinating axons, Adv. Neurol., 47 (1988) 185-213. 31 Waxman, S.G. Rules governing membrane reorganization and axon-glial interactions during the development of myelinated fibers. In F.J. Seil, E. Herbert and B.M. Carlson (Eds.), Progress in Brain Research, Vol. 71, Elsevier, Amsterdam, 1987, pp. 121-141.
