Brain 2006 129(12):3141-3146; doi:10.1093/brain/awl328
© The Author (2006). Published by Oxford University Press on behalf of the Guarantors of Brain. All rights reserved. For Permissions, please email: journals.permissions@oxfordjournals.org
From the Archives
Alastair Compston
Cambridge
Focal experimental demyelination in the central nervous system. By W.I. McDonald and T.A. Sears (From the University Department of Clinical Neurology, and the Department of Neurophysiology, Institute of Neurology, Queen Square, London). Brain 1970: 93; 575–582. with The effects of experimental demyelination on conduction in the central nervous system. By W.I. McDonald and T.A. Sears (From the University Department of Clinical Neurology, and the Department of Neurophysiology, Institute of Neurology, Queen Square, London). Brain 1970: 93; 583–598. with The restoration of conduction by central remyelination. By K.J. Smith, W.F. Blakemore and W.I. McDonald. (From the Institute of Neurology, Queen Square, London and the Department of Clinical Veterinary Medicine, Madingley Road, Cambridge). Brain 1981: 104; 383–404.
‘It is obvious that when a nerve fibre disintegrates completely, conduction must cease’. But Ian McDonald and Tom Sears ask: what happens to nerve impulses in the central nervous system in the context of demeyelination and preservation of axon continuity; and is the nerve fibre still connected to its cell body capable of transmitting an impulse beyond the focal demyelinating lesion? Much has been learned concerning differences in the axon-glial arrangements of myelinated fibres in the central and peripheral nervous systems in the 7 years since Dr McDonald himself illuminated the nature of conduction in demyelinated peripheral nerve (McDonald WI. The effects of experimental demyelination on conduction in peripheral nerve: a histological and electrophysiological study. I. Clinical and histological observations. II Electrophysiological observations. Brain 1963: 86; 481–500 and 501–524). Therefore, the electrical properties of central conduction cannot be assumed but must be studied directly. But how to create a suitable model? Sixteen cats received focal injections of diphtheria toxin directly into the dorsal columns of the spinal cord: in two others the toxin was first heat inactivated; and eight animals had a complete surgical lesion of the dorsal half of the spinal cord to produce Wallerian degeneration. At 12–28 days, diphtheria toxin (0.005 ml) produces a 4 mm lesion with a small central core of disintegrating axons surrounded by an area in which the myelin is broken into balls and granules leaving surviving axons displaced and irregularly grouped (see Fig. 1); and there is a cellular infiltrate of foamy cells with ingested myelin debris. For up to four weeks, depending on the volume of injected toxin, animals have an adducted hind limb that extends when the cat is placed on its haunches; resistance to passive movement of that limb with an increased knee jerk; a delayed or abolished hopping reaction; diminished withdrawal responses; and retained urine. Cats given heat-inactivated toxin remain functionally intact and show no histological abnormalities. The mechanism of this primary demyelination is a matter of speculation, and the similarity of the histology to the lesions of multiple sclerosis, as described in the classical literature [Babinski J. Recherches sur l'anatomie pathologique de la sclérose en plaques et etude comparative des diverses varietes de la scléroses de la moelle. Arch Physiol (Paris) 1885: 2–6; 186–207; and Dawson J The histology of disseminated sclerosis. Transactions of the Royal Society of Edinburgh 1916: 50; 517–740], of some interest but, crucially, here is a model ripe for electrophysiological exploration.
View larger version (100K):
[in this window]
[in a new window]
[Download PowerPoint slide]
|
Fig. 1 Longitudinal sections through the posterior columns. (A) Normal cat (x200). (B) Injected cat, with a lesion made 15 days previously (x200). (C) Just rostral to a cut through the posterior columns made 14 days previously (x200). Holmes' silver-luxol fast blue-cresyl fast violet.
|
|
The questions are these: is conduction blocked at the site of
a focal demyelinating lesion; can conduction occur in the segment
of nerve distal to such a block, with respect to the cell body;
and if conduction occurs through a region of demyelination,
what are its characteristics? Answers that, to us, support concepts
so fundamental as seemingly always to have been part of our
general understanding of demyelinating disease are now revealed
by the characterization of compound action potentials and single
fibre studies in the posterior columns below, at and above the
site of these diphtheritic focal feline experimental demyelinating
lesions. In their second paper from 1970, Drs McDonald and Sears
compare 9 normal cats, 15 with demyelinating lesions of the
posterior columns, and 8 with Wallerian degeneration induced
by surgical section. In order to isolate the lesion for electrophysiological
studies, the thoracic spinal cord is sectioned over several
segments at a level above or below the toxin induced lesion
(to allow the study of orthodromic or antidromic conduction,
respectively) leaving the posterior columns forming an exclusive
bridge across rostral—caudal sections of the lesioned
spinal cord. Single fibres are studied by recording responses
in isolated intercostals nerve fibres from muscle spindles following
antidromic stimulation of their central projections above or
below the diphtheria toxin lesion in the spinal cord. Normal
fibres conduct at
41 m/s producing a triphasic action potential—the
overall size normally decreasing with distance from the stimulus
due to dispersion and loss of fibres from the posterior columns.
In the demyelinated cord, there is an abrupt reduction in the
negative component of the action potential and increased positivity,
features that decay thereafter as conduction is traced through
the lesion (see
Fig. 2). It is evident that the recording electrode,
sensing the approach of the compound action potential (the positive
wave) has failed to register its successful transmission (the
negative deflection). In short, complete conduction block has
developed in a substantial number of fibres. But poor discrimination
with respect to demyelination and Wallerian degeneration seen
with orthodromic caudal stimulation and rostral recording (not
illustrated) now make it necessary for Ian McDonald and Tom
Sears to confine their further experiments to antidromic volleys.
Even so, a nagging doubt that the demonstration of a compound
action potential following antidromic stimulation—suggesting
that caudal fibes are still capable of transmitting the nerve
impulse—might yet be confounded by orthodromic conduction
in afferent fibres, described by Cajal (Ramón y Cajal
S. Textura del sistema nervioso del hombre y de los vertebrados.
Madrid 1899–1904 [volume 1], cited here in the 1952 reprint
of the French translation from 1909 to 1911), which enter the
posterior columns and send collaterals in a caudal as well as
rostral direction, necessitates a series of additional experiments.
But all is well: ‘action potentials in descending collaterals
did not contribute significantly to the antidromic compound
action potential recorded just rostral to the lesion produced
by diphtheria toxin ... the fibres contributing to the compound
action potential must have passed through the lesion from cell
bodies more caudally situated. Thus, the lesion has blocked
conduction in the fibres without destroying axonal continuity,
or interfering with conduction in the fibres distal to the lesion
(with respect to their cell bodies)’ (see
Fig. 3). But
now a new difficulty arises. It is not possible to decide whether
the demonstrable slowing of conduction, sometimes down to 8
m/s, through a lesion is explained by preserved conduction in
normal but slowly conducting fibres that have been spared, or
to local slowing in unhappy but previously fast-conducting axons
(see
Fig. 4). Single fibre studies are needed.
View larger version (12K):
[in this window]
[in a new window]
[Download PowerPoint slide]
|
Fig. 2 Injected cat 14 days' survival, L1 lesion. Conduction distances measured from first recording site. Records made with averaging computer (Biomac). Note between 7 and 8 mm, abrupt transition in the form of action potential with differential reduction in amplitude of negative component.
|
|
View larger version (9K):
[in this window]
[in a new window]
[Download PowerPoint slide]
|
Fig. 3 Comparison of antidromic compound action potentials recorded under standard conditions in a normal cat, rostral to a demyelinating lesion 14 days after injection, and rostral to a complete transaction of the dorsal columns (Wallerian degeneration) 14 days previously.
|
|
These inform three aspects of the diphtheria toxin lesion: conduction
velocity; the refractory period of transmission; and the ability
to transmit trains of impulses at high frequency. On the issue
of velocity, whereas in normal cats intraspinal conduction does
not vary by >15 m/s, across a lesion there is a reduction
to
30% of that achieved by the same fibres conducting through
normal sections of their trajectories (see
Fig. 5). Using a
stimulus at 120% of that needed to ensure certain firing, McDonald
and Sears reduce the interval between paired shocks to determine
the minimum refractory period of transmission: the normal interval
of 0.5–1.1 (mean 0.85) ms increases through the lesion
to 0.6–4.2 (mean 1.7) ms. And, whereas the intact fibre
caudal to a lesion can faithfully transmit rapid trains of impulses,
across the lesion there is often 50% fallout at stimulation
rates (
500 Hz) that the normal cord can easily sustain—the
lowest rate at which failure occurs being 290 Hz (see
Fig. 6).
So, for a 5 mm lesion the evidence is clear: in the context
of pure demyelination, conduction is blocked but normal transmission
occurs above and below the lesion; and the distinction is only
made from the electrophysiological consequences of Wallerian
degeneration if reliance is placed exclusively on studies involving
antidromic stimulation. Aware from their experiments that these
abnormalities vary between fibres, some of which are electrically
normal, and noting that the histological changes in such lesions
combine demyelination and remyelination, alteration of intracellular
fluid composition and oedema, nodal widening, and axon degeneration,
McDonald and Sears do not yet feel comfortable in concluding
that the completely demyelinated mature internodal axon is electrically
excitable; nor are they sure that continuous conduction of an
impulse is possible in a fibre whose previous conduction was,
by definition, saltatory. Theoretically, any event that leads
to internodal loss of current will compromise saltatory events
at the next node of Ranvier; and, ultimately, a critical stage
is reached whereby this loss exceeds that needed to support
transmission from the last surviving node, and conduction fails.
But, despite these commendable restraints on what can reasonably
be concluded from the electrophysiological data, McDonald and
Sears do allow themselves to speculate on the pathophysiology
of symptom production in multiple sclerosis, especially symptoms
that fatigue; and they serve notice that their model is one
that might inform the electrophysiological study of recovery
of function attributable to remyelination, and hence promising
special significance for clinical interpretation, in due course.
View larger version (12K):
[in this window]
[in a new window]
[Download PowerPoint slide]
|
Fig. 5 Conduction velocity of two contiguous segments of individual nerve fibres. Below, schematic representation of electrode arrangements and conduction path; shaded area represents site of lesion. Two pairs of stimulating electrodes placed on the posterior column, S2 just rostral to the dorsal root entry zone of the segment in which the single unit being studied enters the cord, and S1 1–2 cm rostral to this. Intraspinal velocity measured by difference between conduction times from S1 and S2 to r, and the more peripheral velocity between S2 and r. Symbols are aligned above centre of conduction path. Full and dotted lines connect the pair of measurements from the same fibre. Open circles, normal cats; filled triangles, injected cats.
|
|
View larger version (11K):
[in this window]
[in a new window]
[Download PowerPoint slide]
|
Fig. 6 Injected cat, 12 days' survival, T9 lesion. Transmission failure through a lesion of a brief, high frequency train of impulses; dotted vertical line is stimulus artefact. Note with transmission through lesion, response only to alternate stimuli after the first three.
|
|
That study is published 11 years later based on work carried
out as part of his PhD by Kenneth Smith (see page 3147), supervised
by Professor McDonald (the intervening years have seen his elevation
to professorial status) and in collaboration with W.F (‘Bill’)
Blakemore. The logic goes thus: symptoms associated with demyelination
of the central nervous system get better; remyelination in peripheral
nerve results in reversal of conduction block; remyelination
occurs in the central nervous system; but peripheral and central
nerve fibres are not the same; and, in particular, thinly
remyelinated
central fibres may not sustain conduction. Now, diphtheria toxin
is put aside and their model uses direct lysophosphatidyl choline
(LPC) micro-injection into the spinal cord, since this produces
less Wallerian degeneration and is followed by more remyelination.
First, stimulus and recording electrodes are embedded around
the spinal cord. Cats are studied before and for up to 1 year
after the establishment of the demyelinated LPC lesion that
measures between 2 and 5 mm (in length) and 1.5 mm
2 in transverse
section. The histological features evolve from splitting of
the intraperiod line, to myelin breakdown and demyelination
with infiltration by macrophages (at 7 days); restoration of
axon-glial arrangements (at 10 days); extensive oligodendrocyte
remyelination (at 1 month); the appearance of short internodes
with thinly myelinated axons (at 2 months); and thickening of
the myelin sheaths (at 3 months: see
Fig. 7). The acute physiological
findings are familiar: reduction in amplitude of the compound
action potential with no detectable increase in the number of
slowly conducting fibres, consistent with conduction block through
the lesion, and with normal conduction at other sites in the
spinal cord. But now the amplitude of the compound action potential
is seen to increase from 14 days, peaking at 22–29 days
but showing no further increments after 81 days. At 14 days
the refractory period for transmission has increased from the
normal value of
0.6–0.8 ms to at least 1.6 ms with some
fibres having periods prolonged to 3 or 4 ms. Later, this refractory
period shortens as the compound action potential returns. The
loss of conduction and its restoration coincide with the onsets
of demyelination and remyelination, respectively (see
Fig. 8);
more precisely, the severity of the physiological changes at
their nadir, correlate with the size of the lesion, itself somewhat
variable between animals, in the dorsal column. But to understand
the basis for recovery, the pathophysiology of changes occurring
during the phase of intense demyelination must first be elucidated.
Are these the result of dispersion or fibre degeneration? Since
no nerve fibre regeneration is seen, and no fibres show prolonged
latencies, conduction block in a large proportion of surviving
axons is the only reasonable explanation. The temporal association
of these electrophysiological changes and histological evidence
for demyelination make it churlish to reject a causative relationship.
It follows that the increase in amplitude of the compound action
potential from 14 days indicates either that more fibres are
conducting or the active ensemble is now better coordinated.
Discarding the latter as unlikely, Kenneth Smith and colleagues
conclude that conduction is restored in fibres that were previously
blocked and had a prolonged refractory period for transmission.
But is remyelination the most parsimonious explanation for this
recovery of function? Formal consideration of other possibilities
ensues. Gallamine, used to relax the anaesthetized cats during
physiological experimentation, might have a direct effect on
the compound action potential: loss of current across the internode
of thinly myelinated nerve fibres in overcoming the raised capacitance
and conductance reduces the safety factor and increases the
refractory period for transmission—features that may reverse
as the myelin sheaths thickens but not to a degree that would
confidently predict restoration of the refractory period to
normal, unless the shortening of internodal distance compensates
for this probable loss of current across the remyelinated sheath;
and failure of the compound action potential ever to reach its
pre-lesion amplitude is explained by associated nerve fibre
loss as part of the original toxic lesion—although the
latencies indicate that surviving fibres are conducting at normal
velocity. Taken together, Kenneth Smith, Bill Blakemore and
Ian McDonald cannot improve on the assumption that conduction
is restored directly as a result of remyelination in previously
demyelinated fibres with conduction block.
View larger version (58K):
[in this window]
[in a new window]
[Download PowerPoint slide]
|
Fig. 7 (A and B) Eleven-day lesion. An axon is covered by a thin layer of myelin. A node is present at n. One month lesion. Most of the axons are thinly remyelinated but a few demyelinated axons and some normally myelinated axons are also visible.
|
|
View larger version (13K):
[in this window]
[in a new window]
[Download PowerPoint slide]
|
Fig. 8 Diagram showing the temporal relationship between the compound action potentials recorded through the lesion and the sequence of histological events at the lesion. The block and subsequent restoration of conduction correlate with the periods of demyelination and remyelination, respectively.
|
|
Reflecting on these experiments 25 years later, Ian McDonald
recalls that—despite the advances—‘the story
was far from complete, not least because it could not account
for the striking recovery that follows some purely degenerative
lesions. Filling in the gaps took a long time, requiring as
it did not only technical advances in experimental and clinical
neurophysiology, but the revolution which led to the development
of methods for
in vivo monitoring of inflammation in focal lesions
(by Gadolinium enhanced MRI) and adaptive changes in cerebral
processing (by functional imaging). Our physiological experiments
rarely lasted <18 h; Tom Sears and I often watched dawn breaking
as we drove home, knowing that I would be back for a clinic
at 9.00 am. Now, the observations can be made comfortably during
the ordinary working day—a much more satisfactory arrangement
for all concerned.’
As for the clinical significance of these findings, less could be said at the time. Taking the contemporary position that remyelination is scanty in multiple sclerosis, Smith, Blakemore and McDonald doubt whether the pathophysiological mechanisms so neatly described have much to do with recovery of function in that disease: rather, they prefer this as a mechanism more relevant to compression and traumatic lesions of the nervous system. But as papers in the current issue endorse, the extent of reymelination in multiple sclerosis may then have been under-scored, and the ion channel modulations associated with loss and gain of the myelin membrane—bringing benefits and threats to the underlying axon—not fully appreciated. Now, it is clear that these physiological studies of demyelinated and remyelinated fibres make a classic contribution to our understanding of the pathophysiology in multiple sclerosis.
CiteULike 
Connotea 
Del.icio.us What's this?
