Brain Research, 556 (1991) 311-316
(~) 1991 Elsevier Science Publishers B.V. All rights reserved. 0006-8993/91/$03.50
ADONIS 000689939124769Y
311
BRES 24769
Increase of sodium channels in demyelinated lesions of multiple sclerosis
C a r l M o l l 1'2, C h r i s t i a n e
M o u r r e 3, M i c h e l L a z d u n s k i 3 a n d J i i r g U l r i c h 1
llnstitute of Pathology, Department of Neuropathology, University Hospital of Basel, 2Instituteof Pathology, Department of Clinical
Pathology, University Hospital of Zf~rich (Switzerland) and 31nstitutde Pharmacologiemol~culaire et cellulaire, Valbonne (France)
(Accepted 30 April 1991)
Key words: Sodium channel; Demyelination; Multiple sclerosis; Human disease; Saxitoxin
Redistribution of sodium channels along demyelinated pathways in multiple sclerosis (MS) could be an important event in restoring
conduction prior to other reparative mechanisms such as remyelination. Sodium channels in human multiple sclerosis lesions were identified
by quantitative fight microscopic autoradiography using tritiated saxitoxin (STX), a highly specific sodium channel ligand. Demyelinated areas
in various central nervous system regions containing denuded but vital axons exhibited a high increase of STX-binding sites by up to a factor
of 4 as compared to normal human white matter. This important finding could explain aspects of fast clinical remissions and 'silent' MS lesions
on functional and morphological properties. Demyelinated axons may functionally reorganize their membranes and adapt properties similar
to those of slow conducting unmyelinated nerve fibres which have a higher amount and a more diffuse distribution of STX binding sites. This
report is the first description of an altered distribution of voltage-sensitive sodium channels in human multiple sclerosis lesions.
Multiple sclerosis (MS) is a relapsing chronic demyelinating disease of the central nervous system (CNS)
which often leads to conduction block of affected
myelinated nerve fiber tracts and consequently to paresis
during attacks. It is a common observation that clinical
remission from these attacks does not depend solely on
facultative remyelination. Therefore, more subtle and
faster structural adaptations might be responsible for
regaining conductivity, a hypothesis that is contradictory
in the literature 24. Conduction of nerve fibers is generated and propagated by activation of sodium channels
during the rising phase of the action potential. In
myelinated fibers these Na + channels are located mainly
in nodes of Ranvier enabling fast saltatory conduction.
Conversely, in unmyelinated nerve fibers, the distribution of Na + channels is more homogeneous along the
axonal membrane, conduction is slower and continuous.
Early morphological evidence of a rearrangement and
continuous distribution of sodium channels in experimentally demyelinated axons came from Foster et al. using
ferric ion-ferrocyanide (FeFCN) staining and electron
microscopy9. However, the FeFCN binding mechanism
to the sodium channel protein is not thoroughly understood and sodium channel redistribution has not been
confirmed in a human demyelinating disease so far.
The neurotoxin saxitoxin (STX) has been widely used
to analyze structure and function of Na + channels 12'13.
STX specifically blocks the Na + channel by binding at a
site associated with the selectivity filter. In a previous
study the distribution of Na + channels has been characterized in rat and normal human brain structures ls-'2°.
Generally high densities were found in neural layers of
neocortex, hippocampal formation, cerebellum and substantia nigra. Other grey areas presented intermediate
levels. Unaltered white matter was essentially devoid of
STX binding sites. To evaluate a possible effect of
demyelination on the appearance of detectable Na +"
channels in plaques of multiple sclerosis, we used
tritiated saxitoxin ([3H]STX) as a specific ligand in a
quantitative autoradiographic investigation.
Postmortem tissue specimens from numerous brain
and spinal cord regions containing demyelinated plaques
and normal white matter were obtained at autopsy from
4 multiple sclerosis patients (aged 48, 66, 69 and 78
years). Postmortem delay varied between 20 and 48 h.
Observed brain regions were optic tract, forebrain white
matter and cerebellum as well as cervical, thoracic and
lumbar spinal cord in longitudinal and horizontal sections. The tissue was quick-frozen in contact with blocks
of dry ice, cut in 15-/~m sections on a cryostat-microtome,
mounted on chrome-gelatine coated glass slides and
stored at -80 °C until use for autoradiography as previously
described 18-2°.
In equilibrium binding studies, the brain tissue sections
Correspondence: C. Moll. Present address: Molecular Neurobiology Laboratory, The Salk Institute, 10010 North Torrey Pines Road, La Jolla,
CA 92037, U.S.A. Permanent address: Institute of Pathology, University of Ziirich, 8091 Ziirich, Switzerland.
312
were incubated at 4 °C with 3.6 nM [3H]STX dissolved in
20 m M Tris-HCl buffer at p H 7.4 containing (in mM):
N-methyl-glucosamine 140, KCI 5.4, CaCI 2 2.8 and
MgSO 4 1.3. The non-specific binding component was
determined in the presence of 1 /~M unlabeled STX.
After 60 min of incubation, tissue sections were rinsed
twice for 5 sec in the same buffer and twice for 5 sec in
distilled water. Then the slices were dried and exposed to
tritium-sensitive film (3H Hyperfilm) with tritium labeled
standards (Amersham) and exposed for 4 months. Autoradiograms were then analyzed and quantified using a
computerized image-analysis system (Numelec). The
!!iil
Fig. 1. a-c: recent demyelinated lesion in periventricular forebrain white matter (bright zone), characteristically located around a small vein
that shows little perivascular inflammation (a; right from center; Luxol stain; 15:<). Corresponding strong [3H]STX-binding signal in adjacent
section (b; autoradiograph; 15x) that leaves vein blank (right from center). Intact axons can be found in this signal enhancing lesion at higher
magnification (c; arrow; silver impregnation; 185×). d-f: old scarred demyelinated lesion of the medulla oblongata. Ventral pyramidal tract
(2 arrowheads) is affected (a; Luxol stain; 11 x). Undetectable corresponding [3H]STX-binding signal (e; autoradiograph; 11 x) due to severe
astrocytic scar and lack of surviving axons at higher resolution from within the lesion (f; silver impregnation; 185x).
Fig. 2. a-c: sections of an extensive demyelination in upper thoracic spinal cord. Luxol-stained cryostat section with demyelination of the left
half of spinal cord tracts (a; 8~) and moderate enhancement of [3H]STX-bindingsignal in the [3H]STX autoradiogram (b; 8~) of the
correspondingarea. Gross pathology is shown in c. Clinically the patient showed left-dominant motor and sensory deficits but some residual
function.
specific binding value was determined as the difference
between the total binding for a given area and the
non-specific binding. The latter was homogeneous and
weak (5% of total binding) over all the examined regions.
The mean value of specific binding in a given area was
calculated from 6 to 8 measurements.
Adjacent tissue,
that has been fixed in 7% neutral formaldehyde,
was
sectioned and stained for myelin and axons for conventional light microscopy using standard Luxol dye/silver
impregnation
procedures.
Additionally,
original tissue
sections used for 3H-autoradiography
were stained the
same way with good results after 4 months of exposure to
tritium-sensitive
film.
Binding experiments in MS human CNS tissue sections
with [3H]STX, which is highly specific for the voltagesensitive Na+ channel, have revealed high densities of
binding sites in neocortical and spinal cord grey matter
whereas in unchanged
white matter [3H]STX binding
sites were undetectable.
This localization is in agreement
with the STX binding
site distribution
in normal
human
brain and spinal cord2’. Fig. la-c shows a plaque of
complete demyelination
in subcortical white matter demonstrated by a loss of Luxol myelin stain. A strong local
increase of STX binding sites is found in the corresponding autoradiogram
(Table I). Lesions of the optic nerve
and of spinal white matter presented a lower, but still
increased labeling (Table I and Figs. 2 and 3). These
plaques revealed only partial demyelination,
An increase of t3H]STX binding site densities could
correspond
to an adaptive augmentation
of detectable
Na+ channels or only reflect a decrease of quenching.
The presence of myelin lipids significantly quenches the
autoradiographic
signal by autoabsorption
of tritium B
emission”.
Therefore, we corrected the measured densities of STX binding sites using published quenching
coefficients
for different
brain regions”.
The mean
quenching
coefficient
corresponds
to a 35.6% and a
110.7% increase of the measured density of sites in grey
314
TABLE I
Distribution of STX binding sites (finol/mg grey matter) and
quenching-corrected valuesw in human multiple sclerosis lesions
S.E.M. = Standard error of the mean.
Localization
Specific + S. E.M.
binding
(fmollmg
grey
matter)
Specific
binding
corrected
for quenching
Density
ratio
demyelinated/myeiinated
Forebrain
grey matter
(cortex)
white matter
demyelinated
plaque
139.5
7.4
+16.2
+0.5
188.3
15.5
4.02
47.0
+6.3
63.5
Spinal cord
grey matter
white matter
demyelinated
plaque
25.8
3.8
+0.9
+0.6
34.9
8.0
12.2
+0.2
16.5
1.7
+0.3
3.6
5.3
+0.9
7.2
Optic nerve
white matter
demyelinated
plaque
2.06
2.0
matter and white matter regions respectively. Because
demyelinated MS plaques lack myelin lipids, we used the
correcting factor for grey matter.
The analysis of the corrected values (Table I) indicates
a large to moderate increase of [3H]STX binding sites in
demyelinated plaques of subcortical white matter (a
factor of 4), optic nerve (a factor of 2) and spinal cord (a
factor of 2). It should be stressed that the calculated
values presented in Table I are minimum values, since, if
demyelination was only partial, the quenching coefficient
that was used was too low and corrected values would be
higher. The difference between normal white matter and
demyelinated plaques would then be even more pronounced.
The demonstrated increase of Na ÷ channels in demyelinated lesions could be of glial or axonal origin.
Astrocytic proliferation is a common finding in MS
plaques and astrocytes are able to generate sodium
currents which are inhibited by tetrodotoxin 1'2. However,
the maximal density of STX binding sites in rat astrocytes
is 10 times lower compared to neuronal elements 4,s. In
our study, the presence of intact axons inside demyelinated plaques seems to be required for a higher level of
STX signal (Fig. lb,c). Heavily scarred plaques with a
complete absence of axons were devoid of detectable
STX binding sites (Fig. le,f), favoring a neuronal origin.
The pattern of Na ÷ channel distribution changes
drastically with myelination during ontogenesis of the rat
brain. In premyelinated fiber tracts STX binding sites are
Fig. 3. Subdural triangular shaped incomplete demyelinated area in
the optic nerve (a; Luxol stain; 23x). Moderate enhancement of
[3H]STX-binding sites (b; autoradiogram; 23x).
present at an intermediate density but the specific
binding signal becomes weaker and eventually undetectable as myelination proceeds ~9. The process of demyelination seems to induce an adaptive reverse phenomenon leading to the reappearance of Na ÷ channels in white
matter areas, comparable to premyelination stages. This
interpretation would be in agreement with clinical electrophysiological data from MS patients, presenting complete conduction block in acute demyelinated pathways 26'
27 During clinical remission after a few days - - an
interval too short for remyelination to occur - - evoked
potentials with slowed conduction velocities reappear, as
could be nicely demonstrated in the optic system 15.
Again, an increase of Na ÷ channel density along demyelinated segments might cause changes of conduction
properties which then resemble properties of unmyelinated or premyelinated nerve fibers 9"22,26.
An increase of Na ÷ channel density could be the result
of a de novo synthesis of the sodium channel protein
or/and of changes of axonal transport 14 or of a transfer
from astrocytes ~'2. An adaptation of sodium channel
density similar to that described in the present paper for
multiple sclerosis plaques has been seen in demyelinated
axons of the sciatic nerve in the med/rned mouse 2~ and in
315
experimentally demyelinated axons using doxorubicin to
kill Schwann cells 8. These studies suggest that an upregulation of sodium channel protein in experimentally
demyelinated axons can occur without glial participation
and is basically of neuronal origin. O u r data indirectly
support this and contrast the conclusion reached in a
recent paper by Ritchie et al. 23 studying sodium channels
in Schwann cells: we could not detect STX binding sites
in axon-free astrocytic scars of MS patients. However,
recent data from cultured type-1 astrocytes 3 and from the
rat optic nerve after experimental enucleation ~5, indicate
that astrocytes lose their ability to carry sodium current
when deprived of axonal contact. These findings point to
an interdependently regulated axo-glial sodium channel
expression which could, at least in part, also be responsible for undetectable STX-binding sites in axon-free
astrocytic scars of MS patients. The reverse is not true:
it has not been shown so far, that neurons and their
processes depend on astrocytes in expressing the sodium
channel protein.
In a preliminary immunohistochemical experiment we
used anti-neurofilament antibodies to show a significant
increase of neurofilament components (triplet) in axons
passing demyelinated areas (data not shown). Whether
this increase reflects a general upregulation of neuronal
protein synthesis or is a sign of a disturbed axonal
transport leading to protein accumulation cannot be
decided so far.
If demyelinated segments in nerve fibers could adopt
conduction properties of unmyelinated nerve fibers by an
increase and/or a redistribution of sodium channels along
their axonal membrane, this would be an interesting
example of neuroaxonal plasticity. Bostock and Sears 6'7
demonstrated continuous conduction along short segments of axons after demyelination by diphtheria toxin.
Clinically, in certain types of chronic demyelinating
peripheral neuropathies, such as the hereditary neuropathy with liability to pressure palsies, a slowed conduction velocity is generally well tolerated by affected family
members. Often they are free of any symptoms 16. In
analogy, it may be that clinically asymptomatic or 'silent'
demyelinated MS lesions, as have been described in the
spinal cord 11, do not result in apparent paresis because of
a functional adaptation of axonal membranes without
necessary remyelination. The latter mechanism might
inherit a significant change of sodium ion channel
distribution, as shown here. Our observation could also
help to explain a reasonable visual function in conditions
of complete demyelination, as described in the optic
nerve 25.
1 Barres, B.A., Chun, E L Y . and Corey, D.P., Ion channel
expression by white matter glia: I. Type 2 astrocytes and
oligodendrocytes, Glia, 1 (1988) 10-30.
2 Barres, B.A., Chun, E L Y . and Corey, D.P., Glial and
neuronal forms of voltage-dependent sodium channel: characteristics and cell-type distribution, Neuron, 2 (1989) 1375-1388.
3 Barres, B.A., Koroshetz, W.J., Chun, E L Y . and Corey, D.P.,
Ion channel expression by white matter glia: type-1 astrocyte,
Neuron, 5 (1990) 527-544.
4 Bevan, S., Chiu, S.Y., Gray, P.T.A. and Ritchie, J.M., The
presence of voltage-gated sodium, potassium and chloride
channels in rat cultured astrocytes, Proc. R. Soc. London, B225
(1985) 299-313.
5 Bevan, S., Chin, S.Y., Gray, P.T.A. and Ritchie, J.M., Voltagegated ion channels in rat cultured astrocytes. In J.H. Ritchie and
R.D. Keynes (Eds.), Ion Channels in Neural Membranes, Liss,
New York, 1986, pp. 159-174.
6 Bostock, H. and Sears, T.A., Continuous conduction in demyelinated mammalian nerve fibres, Nature, 263 (1976) 786-787.
7 Bostock, H. and Sears, T.A., The internodal axon membrane:
electrical excitability and continuous conduction in segmental
demyelination, J. Physiol., 280 (1978) 273-301.
8 England, J.D., Gamboni, F., Levinson, S.R. and Finger, E.,
Changed distribution of sodium channels along demyelinated
axons, Proc. Natl. Acad. Sci. U.S.A., 87 (1990) 6777-6780.
9 Foster, R.E., Whalen, C.C. and Waxman, S.G., Reorganization
of the axon membrane in demyelinated peripheral nerve fibres:
morphological evidence, Science, 210 (1980) 661-663.
10 Geary, W.A. and Wooten, G.E, Regional tritium quenching in
quantitative autoradiography of the central nervous system,
Brain Research, 336 (1985) 334-336.
11 Ghatak, N.R., Hirano, A., Lijtmaer, H. and Zimmerman,
H.M., Asymptomatic demyelinated plaque in the spinal cord,
Arch. Neurol., 30 (1974) 484-486.
12 Kao, C.Y. and Levinson, S.R. (Eds.), Tetrodotoxin, Saxitoxin
We wish to thank Prof. M. Mumenthaler, Bern, and Dr. J.
Kesselring, Valens. This work was generously supported by a grant
from The Swiss Multiple Sclerosis Society.
and the Molecular Biology of the Sodium Channel, VoL 479,
Ann. N.Y. Acad. Sci., 1986.
13 Lazdunski M., Frelin, C., Barhanin, J., Lombet, A., Meiri, H.,
Pauron, D., Romey, G., Schmid, A., Schweitz, H., Vigne, P.
and Vijverberg, H.P.M., Polypeptide toxins as tools to study
voltage-sensitive Na + channels. In Tetrodotoxin, Saxitoxin and
the Molecular Biology of the Sodium Channel, Vol. 479, Ann.
N.Y. Acad. SCI., 1986, pp. 204-220.
14 Lombet, A., Laduron, P., Mourre, C., Jacomet., Y. and
Lazdunski, M., Axonal transport of the voltage-dependent Na +
channel protein identified by its tetrodotoxin binding site on rat
sciatic nerve, Brain Research, 345 (1985) 153-158.
15 McDonald, W.I., Acute optic neuritis, Br. J. Hosp. Med., 18
(1977) 42-48.
16 Meier, C. and Moll, C., Hereditary neuropathy with liability to
pressure palsies: report of two families and review of the
literature, J. Neurol., 228 (1982) 73-95.
17 Minturn, J.E., Black, J.A., Sontheimer, H., Emanuel, J.R.,
Ransom, B.R. and Waxman, S.G., Voltage-sensitive sodium
channels in astrocytes deprived of axonal contact: peptides,
mRNAs and sodium current expression, J. Cell BioL, 111 (1990)
489 (Abstract no. 2745).
18 Mourre, C., Lombet, A. and Lazdunski, M., Autoradiographic
localization of tetrodotoxin-sensitive Na + channels in rat brain,
Neurosci. Lett., 52 (1984) 31-35.
19 Mourre, C., Cervera, P. and Lazdunski, M., Autoradiographic
analysis in rat brain of the post-natal ontogeny of voltagedependent Na + channels, Ca2+-dependent K + channels and
slow Caz+ channels identified as receptors for tetrodotoxin,
apamin and (-)desmethoxyverapamil, Brain Research, 417
(1987) 21-32.
316
20 Mourre, C., Moll, C., Lombet, A. and Lazdunski, M., Distribution of voltage-dependant Na + channels identified by high
affinity receptors for tetrodotoxin and saxitoxin in rat and human
brains: quantitative autoradiographic analysis, Brain Research,
448 (1988) 128-139.
21 Rieger, E, Pin~on-Raymond, M., Lombet, A., Ponzio, G.,
Lazdunski, M. and Sidman, R.L., Paranodal dysmyelination and
increase in tetrodotoxin binding sites in the sciatic nerve of the
motor end-plate disease (med/med) mouse during postnatal
development, Dev. Biol., 101 (1984) 401-409.
22 Ritchie, J.M., Rang, H.P. and Pellegdno, R., Sodium and
potassium channels in demyelinated and remyelinated mammalian nerve, Nature, 294 (1981) 257-259.
23 Ritchie, J.M., Black, J.A., Waxman, S.G. and Angelides, K.J.,
24
25
26
27
Sodium channels in the cytoplasm of Schwann cells, Proc. Natl.
Acad. Sci. U.S.A., 87 (1990) 9290--9294.
Rubinstein, C.T. and Shrager, P., Remyclination of nerve fibers
in the transectcd frog sciatic nerve, Brain Research, 524 (1990)
303-312.
Ulrich, J. and Groebke-Lorenz, W., The optic nerve in multiple
sclerosis. A morphological study with retrospective clinicopathological correlations, Neurol. Ophthalmol., 3 (1983) 149-159.
Waxman, S.G., Conduction in myelinated, unmyelinated and
demyelinated fibres, Arch. Neurol., 34 (1977) 585-589.
Waxman, S.G., Conduction through demyelinated plaques in
multiple sclerosis: computer simulations of facilitation by short
internodes, J. Neurol. Neurosurg. Psychiatry, 41 (1978) 408-416.