Professional Documents
Culture Documents
SYNOPSIS The simplest model for explaining conduction defects in multiple sclerosis (MS) and
other demyelinating diseases assumes that the only abnormality present is loss of myelin. The conse-
quences of such an assumption have been investigated by numerical solution of a well-known set of
differential equations describing conduction in a model demyelinated axon. In agreement with
clinical findings, we show that this model predicts that the temperature at which conduction block
occurs is a steep function of the extent of demyelination, so that small temperature increases may
block large numbers of conducting fibres. Decreasing calcium concentration (or increasing pH) is
calculated markedly to improve the conduction velocity of conducting demyelinated fibres and will,
in addition, restore conduction in blocked fibres. The effects of other pharmacological agents have
also been computed. The presence of a demyelinating lesion in a nerve fibre is shown greatly to impair
the ability of the fibre to conduct repetitive impulses, conduction failing at much lower frequencies
than in normal fibres. These calculations provide some insight into the nature of conduction defects
in demyelinated nerve, demonstrate that many clinical features of MS are the expected consequence
of loss of myelin and do not require the presence of other defects for their explanation, and provide
a useful approach to the search for a symptomatic therapy.
Multiple sclerosis (MS) is a disease of the central fluctuate dramatically with small changes in the
nervous system in which the characteristic internal environment. Thus small changes in
pathological alteration consists of myelin loss body temperature produce dramatic reversible
with a relative preservation of axis cylinders. It alterations in the neurological signs and symp-
is generally believed that the clinical signs and toms of MS, with heat causing a worsening
symptoms in MS are the result of defects of im- (Nelson and McDowell, 1959; Namerow, 1968b)
pulse conduction in demyelinated axons of the and cooling an improvement (Watson, 1959).
central nervous system. A striking example of Experimental evidence suggests that this phenom-
this is the central visual field defect associated enon is caused by an effect of temperature on
with optic nerve lesions. Here there can be little conduction in demyelinated axons (Davis, 1970;
doubt that the observed scotoma is related to a Davis and Jacobson, 1971; Rasminsky, 1972).
disturbance of conduction in a specific bundle of Also, acute transient improvement in scotomas,
topographically oriented demyelinated fibres. nystagmus, and oculomotor paresis has been
A spectrum of physiological abnormalities is demonstrated in MS patients after procedures
seen in MS patients. In addition to apparent known to lower serum calcium concentration
slowing and block of conduction (Namerow, (Davis et al., 1970).
1968a), some fibres are unable faithfully to con- The simplest possible model for MS (and pre-
duct repetitive impulses and appear to fatigue sumably for other demyelinating diseases)
rapidly (Namerow, 1972). There is also a curious assumes that the only defect present is loss of
lability whereby clinical signs and symptoms can myelin without any concomitant pathological
152
Conduction in multiple sclerosis 153
alterations in the nerve membrane itself. appropriate for the degree of myelin loss (equations
Rasminsky and Sears (1972) have provided strong 2 and 3 of Goldman and Albus, 1968) and expressed
experimental evidence in rat ventral nerve root as a fraction of the normal resistance. We assumed
fibres demyelinated by injection of diphtheria demyelination to begin at node 4 and proceed to the
toxin that the axon membrane in the internode is end of the fibre, except in those cases where a single
not excitable. Therefore, previous mathematical
demyelinated internode was studied. All constants
and initial conditions are given in Goldman and
models demonstrated to describe conduction in Albus (1968, Appendix). Note that the experimental
myelinated nerve (Fitzhugh, 1962; Goldman and data from which these equations were derived were
Albus, 1968) are modified to simulate myelin loss obtained at 200 C, which is consequently the reference
by only altering the myelin resistance and capaci- temperature for our calculations.
tance. The value of 30 ,tm chosen for normal external
Smith and Koles (1970) and Koles and myelin diameter is at the upper range of real fibre
Rasminsky (1972) have investigated the effects diameters. This was used in order to have values for
of several morphological patterns of myelin loss conduction velocities which were large, so as to
(uniform demyelination of one or more inter- minimize running time for the computer programmes.
nodes and paranodal demyelination) on conduc- The work of Goldman and Albus (1968) shows that
as long as the relationships between normal external
tion velocity using such a model system. In all myelin diameter, axis cylinder diameter, and inter-
cases they found conduction velocity to decrease nodal distance remain constant, the only effect of
with increasing demyelination until finally con- changing fibre diameter is to scale the conduction
duction block occurs, though the amount of velocity by a constant factor proportional to diam-
myelin just sufficient to sustain conduction de- eter.
pends on the exact morphology.
We have extended calculations based on this
model system to attempt to account for the TABLE
complex clinical phenomena described earlier. EXPERIMENTAL EFFECTS OF ALTERATIONS IN TEMPERATURE
These studies provide a unified account of AND CALCIUM CONCENTRATION
temperature and calcium effects and suggest one
possible cause of observed impairment in re- Parameter Q1o* m V shift pert
e-fold change
petitive impulse conduction. In addition, they in [Ca+ +]
indicate a sound basis for the possible develop- PNa 1-3
ment of a symptomatic therapy for demyelina- P'K 12 -
ting diseases. Pp 12 -
am 1-8 8-6
Pm 1-7 8-6
ah 2-8 7-4
METHODS Ph 2-9 7-4
an 3-2 0-8
The system of equations describing the model system On 2-8 0-8
is identical with that first presented by Goldman and ap 3 0$ 0°°:1
,Bp 3t0: 00:
Albus (1968) and modified to account for demyelina-
tion by Koles and Rasminsky (1972), except for * Data from Frankenhaeuser (1963).
changes in the voltage-dependent rate constants and t Data from Hille (1968). Note that Hille
permeability constants (equations 9-16 of Goldman presented his results as shifts in time con-
stants and steady-state levels for the
and Albus, 1968) to account for the effects of changes various parameters, not directly as rate
in temperature, calcium concentration, and pH, and constants. The values here are derived
from his data.
the application of pharmacological agents. t Values assumed. Experimental data not
The model nerve consisted of 20 nodes with available.
stimulating current applied to node 1 and with node
20 shorted to ground. The geometry was identical
with that of Goldman and Albus (1968), with normal The effects of temperature variation were allowed
external myelin diameter chosen to be 30 Vm. Para- for by scaling all rate constants and permeability
nodal demyelination is not a characteristic feature of constants by the experimentally determined Qlo's
MS, and hence, to account for demyelination, (Frankenhaeuser and Moore, 1963; Frankenhaueser,
myelin resistance was usually uniformly decreased 1965) (Table). This method of incorporating the
and myelin capacitance increased by an amount effect of temperature is justified by the experimental
154 C. L. Schauf and Floyd A. Davis
data and has been shown to produce in model nerve 16/65 digital computers. In most cases the fibre was
fibres calculated responses which are in agreement sampled at each node and at space increments of 1/10
with observations on normal living nerve (Hodgkin the internodal distance. Time increments were auto-
and Huxley, 1952; Huxley, 1959; Frankenhaeuser, matically adjusted to maintain errors in voltage at
1965; Cooley and Dodge, 1966; Smith and Koles, less than 10 ,uV.
1970). The effect of temperature on the myelin
resistance and capacitance and axoplasmic resistivity NUMERICAL COMPUTATIONS
is not likely to be a significant factor, but in any case
could not be taken into account for lack of adequate EFFECTS OF TEMPERATURE Figures 1 and 2 show
experimental data. the results of the computations we have carried
The effects of calcium and pH were calculated by out to examine the combined effects of tempera-
translating the functions describing the voltage- ture and loss of myelin on conduction. Figure 1
dependent rate constants along the voltage axis gives the calculated conduction velocity as a
according to experimental data (Hille, 1968-see function of temperature for a normal nerve fibre
Table). This general procedure has also been justified (top curve) and a series of fibres subjected to in-
by computations of the effect of calcium on conduc- creasing amounts of demyelination (0O5, 033,
tion in normal, unmyelinated nerve (Huxley, 1959). 0-2, 0-1, and 0-06 times the normal myelin
The differential equations were converted to
difference form and integrated using a fourth-order resistance). In each curve the last point at the
Runge-Kutta procedure on IBM 370/155 and SPC right is at the highest temperature at which con-
duction occurs. A temperature 30 C higher does
not permit conduction, as indicated by the
*-Normal Myelin
A-0.5X Normal dashed lines.
A-0.33X Normal The conduction velocity of normal nerve in-
°D-0.2X Normal
O-O.IX Normal creases with increasing temperature over a broad
0-0.06X Normal range, reaches a maximum at about 42° C, and
then decreases before block. This result, includ-
LId
i
-
LU
LLI
50 r
Cf) 4 C)
0- 40
z
0
0i cr-
F
/ es $ ~~30
0 4 ~~~~ii
z
20
ID
4
I
~~~~z
Y
0 10 _
10 20 30 40 50 0
TEMPERATURE (°C) m
FIG. 1. Calculated values of conduction velocity as a 0
0 0
function of temperature for a normal nerve fibre 0.2 0.4 0.6 0.8 1.0
(upper curve) and fibres with increasing loss of myelin. RELATIVE DEGREE OF MYEL INAT ION
For each curve the last datum point at the right is for
the highest temperature at which continuous conduc- AS FRACTION OF NORMAL
tion is possible, the dashed lines and arrows simply FIG. 2. Blocking temperature as a function of the
being a convenient means of indicating block of con- amount ofmyelin loss. The error bars correspond to an
duction at a temperature 30 C greater. The amount of uncertainty of 20 C. The amount of demyelination is
demyelination is expressed in terms of the myelin expressed in terms of the myelin resistance relative to
resistance relative to normal. normal.
Conduction in multiple sclerosis 155
blocking temperature to increase toward nor- tion will restore conduction in a fibre blocked by
mal, and this effect is most predominant in demyelination of a single internode or by para-
severely demyelinated fibres. nodal demyelination.
Figure 3b shows that both the blocking
temperature and the conduction velocity are RESPONSES TO REPETITIVE STIMULATION The
linear functions of the logarithm of the calcium response of the model axon to a second stimulus
concentration. This figure is for a fibre with 10% delivered at a variable time after an initial
normal myelination. If a larger amount of stimulus is shown in Fig. 4 for both a normal
myelin is present, the slope of the curve relating nerve and a nerve which has lost all but 10% of
blocking temperature to calcium concentration its myelin. The action potential amplitude and
is decreased. conduction velocity resulting from the initial
Thus, the result of decreasing the calcium stimulus were 110-4 mV and 37-8 m/sec for nor-
concentration in a population of fibres having mal nerve, and 90-0 mV and 8-7 m/sec for the
different amounts of myelin lost is predicted to 9000 demyelinated fibre. To prepare Fig. 4, the
be an increase in the number of conducting magnitude and conduction velocity of the
fibres, combined with a restoration of the con- response produced by a second impulse were
duction velocity toward more normal values in divided by the corresponding values for the
those fibres which were not blocked at normal initial response for both normal and demyelin-
calcium levels. This is in contrast with the effects ated nerve.
of temperature in which cooling increases the For normal nerve, a second stimulus at 1 85
number of conducting fibres but decreases con- msec after the initial stimulus is unable to pro-
duction velocity in those which were not blocked. duce a response, whereas at 2-04 msec a second
Increasing pH acts on the rate constants in the response with initial amplitude and conduction
same direction as decreases in calcium con- velocity of 69 mV and 14X6 m/sec is produced.
centration. Detailed calculations were not per- Intermediate times were not explored. With in-
formed, but since the rate constants shift by only creasing time between the stimuli the amplitude
1'3 mV per e-fold change in hydrogen ion con- and velocity of the second response increase
centration (3.0 mV per pH unit) at physiological rapidly toward initial values. During the interval
pH (Hille, 1968), we may conclude that this in which a second response is reduced, the second
alteration would be much less effective than response is continually lagging behind the initial
similar changes in calcium concentration. impulse, while gradually recovering its amplitude
Effects qualitatively similar to these have also and velocity. Whether substantial changes in
been calculated for the other morphological frequency occur depends on the conduction
patterns of myelin loss considered by Koles and distance involved.
Rasminsky (1972). Reducing calcium concentra- In the case of a 9000 demyelinated fibre, a
PHARMACOLOGICAL EFFECTS This model system mental justification. Fitzhugh (1962) first de-
provides a convenient means of defining the scribed the procedure for combining the
manner in which the kinetic properties of the phenomenological description of the behaviour
nerve membrane should be altered in order to of excitable membranes contained in the
improve conduction in demyelinated nerve. Pre- Hodgkin-Huxley equations (Hodgkin and Hux-
sumably such effects occurring in MS patients ley, 1952) with the equations for a passive cable
would improve clinical signs and symptoms. We (Hodgkin and Rushton, 1946) to yield a descrip-
have already described the effects of tempera- tion of conducted action potentials in myelinated
ture, calcium, and pH. Conduction in demyelin- nerve. Goldman and Albus (1968) extended this
ated nerve is improved by factors which increase work by replacing the Hodgkin-Huxley equa-
the duration of the action potential without tions, which describe the behaviour of the squid
adversely affecting conduction velocity or action giant axon, with a similar set of relationships
potential amplitude, since the total membrane (the Frankenhaeuser-Huxley equations) which
current is increased. This may be accomplished were developed by the application of the voltage
by reducing the contribution of the steady-state clamp technique to the myelinated nerve fibre
conductance, or prolonging the sodium conduc- (Dodge and Frankenhaeuser, 1958, 1959;
tance by slowing the inactivation process. Frankenhaeuser, 1959, 1960, 1962a-c, 1963a, b;
Figures 5 and 6 illustrate the effect of a four-
Frankenhaeuser and Huxley, 1964).
fold increase in the sodium inactivation time The calculations of these and other investiga-
constant, rh. This alteration results in an actiontors concerning the behaviour of normal
potential in normal nerve which is about 2-5 myelinated fibres (Frankenhaeuser and Huxley,
times as long. Figure 5 shows that when applied 1964; Frankenhaeuser, 1965; Hutchinson et al.,
to a demyelinated fibre, this procedure does not 1970), combined with similar computations of
appreciably affect the conduction velocity but membrane action potentials (Huxley, 1959;
greatly increases the blocking temperature for Fitzhugh, 1960; Fishman, 1970) and propagated
any amount of demyelination. This is more activity (Cooley and Dodge, 1966) in non-
clearly seen in the upper curve of Fig. 6. The myelinated fibres, have provided compelling evi-
curve relating blocking temperature and amount dence for accepting these models as excellent
of myelin is shifted vertically by more than 150 C.
descriptions of living nerve fibres under a
Thus at any given temperature conduction is variety of environmental conditions.
possible with far less myelin. The calculations of Smith and Koles (1970)
Reduction of the maximum potassium per- and Koles and Rasminsky (1972), when com-
meability will also improve conduction in pared with the experimental data of Rasminsky
demyelinated nerve by increasing the blocking and Sears (1972), demonstrate the ability of this
temperature, but the effects are not as dramatic. model to simulate accurately the behaviour of
Thus, a 5000 decrease in P'k increases the block- real demyelinated fibres.
ing temperature of an 80% demyelinated fibre Koles and Rasminsky (1972 briefly reported
by only 20 C. This is due to the fact that the in-one calculation concerning the increased velocity
creasing potassium permeability does not con- and subsequent block produced by warming a
tribute as much to the falling phase of the actionparanodally demyelinated fibre barely conduct-
potential as the simultaneous inactivation of the ing at 200 C, and mentioned that this quali-
sodium permeability. A 5000 decrease in maxi- tatively agreed with experimental observations of
mum potassium permeability prolongs the nor- Rasminsky (1971) on rat ventral root fibres de-
mal action potential only by about 10% and myelinated by diphtheria toxin. The complete
does not affect conduction velocity at all. family of curves shown in Fig. 1 contains quite
complex behaviour, and it would be an excellent
DISCUSSION
test of the model's accuracy to see if there is
agreement with experimental observations over
It should be stressed that the procedure of using the entire range.
a mathematical model to study the behaviour of The calculated relationships of Figs 1 and 2,
a demyelinated axon has a great deal of experi- in addition to indicating that increased numbers
Conduction in multiple sclerosis 159
present. Also, the theoretical effects of a variety Frankenhaeuser, B. (1960). Quantitative description of
of pharmacological agents on conduction in de- sodium currents in myelinated nerve fibres of Xenopus
laevis. Journal of Physiology, 151, 491-501.
myelinated nerve can be predicted given a suffi- Frankenhaeuser, B. (1962a). Delayed currents in myelinated
ciently detailed understanding of their effects on nerve fibres of Xenopus laevis investigated with voltage
normal nerve fibres. Whether this approach will clamp technique. Journal of Physiology, 160, 40-45.
Frankenhaeuser, B. (1962b). Instantaneous potassium
lead to any clinically useful therapeutic agent is currents in myelinated nerve fibres of Xenopus laevis.
unknown, but the basic idea seems scientifically Journal ofPhysiology, 160, 46-53.
sound and worthy of further investigation. Frankenhaeuser, B. (1962c). Potassium permeability in
myelinated nerve fibres of Xenopus laevis. Journal of
Physiology, 160, 54-61.
We thank Dr. T. Hoeppner, Dr. J. A. Michael, and Frankenhaeuser, B. (1963a). A quantitative description of
Dr. F. Morrell for their helpful comments about this potassium currents in myelinated nerve fibres of Xenopus
laevis. Journal of Physiology, 169, 424-430.
work, and Mrs. Valerie Meineke for her effort in Frankenhaeuser, B. (1963b). Inactivation of the sodium-
preparing the manuscript. This investigation was carrying mechanism in myelinated nerve fibres of Xenopus
supported by the Morris Multiple Sclerosis Research laevis. Journal of Physiology, 169, 445-451.
Fund. Frankenhaeuser, B. (1965). Computed action potential in
nerve from Xenopus laevis. Journal of Physiology, 180,
780-787.
Frankenhaeuser, B., and Huxley, A. F. (1964). The action
REFERENCES potential in the myelinated nerve fibre of Xenopus laevis as
computed on the basis of voltage clamp data. Journal of
Becker, F. O., Michael, J. A., and Davis, F. A. (1973). Acute Physiology, 171, 302-315.
effect of oral phosphate on visual function in multiple Frankenhaeuser, B., and Moore, L. E. (1963). The effect of
sclerosis. Abstracts of the 1973 Annual Meeting, American temperature on the sodium and potassium permeability
Academy of Neurology. Neurology, 23, 431. changes in myelinated nerve fibres of Xenopus laevis.
Cooley, J. W., and Dodge, F. A., Jr. (1966). Digital computer Journal of Physiology, 169, 431-437.
solutions for excitation and propagation of the nerve Goldman, L., and Albus, J. S. (1968). Computation of im-
impulse. Biophysical Journal, 6, 583-599. pulse conduction in myelinated fibers: theoretical basis of
Cragg, B. G., and Thomas, P. K. (1964). Changes in nerve the velocity-diameter relation. Biophysical Journal, 8, 596-
conduction in experimental allergic neuritis. Journal of 607.
Neurology, Neurosurgery, and Psychiatry, 27, 106-115.
Davis, F. A. (1970). Axonal conduction studies based on Greenfield, J. G. (1958). Demyelinating diseases. In Green-
some considerations of temperature effects in multiple field's Neuropathology, pp. 441-474. Arnold: London.
sclerosis. Electroencephalography and Clinical Neuro- Hille, B. (1967). The selective inhibition of delayed potassium
physiology, 28, 281-286. currents in nerve by tetraethylammonium ion. Journal of
Davis, F. A. (1972). Impairment of repetitive impulse con- General Physiology, 50, 1287-1302.
duction in experimentally demyelinated and pressure- Hille, B. (1968). Charges and potentials at the nerve surface.
injured nerves. Journal of Neurology, Neurosurgery, and Divalent ions and pH. Journal of General Physiology, 51,
Psychiatry, 35, 537-544. 221-236.
Davis, F. A., Becker, F. O., Michael, J. A., and Sorensen, E. Hodgkin, A. L., and Huxley, A. F. (1952). A quantitative
(1970). Effect of intravenous sodium bicarbonate, disodium description of membrane current and its application to
edetate (Na2EDTA), and hyperventilation on visual and conduction and excitation in nerve. Journal of Physiology,
oculomotor signs in multiple sclerosis. Journal of Neurology, 117, 500-544.
Neurosurgery, and Psychiatry, 33, 723-732. Hodgkin, A. L., and Rushton, W. A. H. (1946). The electrical
Davis, F. A., and Jacobson, S. (1971). Altered thermal constants of a crustacean nerve fibre. Proceedings of the
sensitivity in injured and demyelinated nerve. Journal of Royal Society of London B., 133, 444-479.
Neurology, Neurosurgery, and Psychiatry, 34, 551-561. Hutchinson, N. A., Koles, Z. J., and Smith, R. S. (1970).
Dodge, F. A., and Frankenhaeuser, B. (1958). Membrane Conduction velocity in myelinated nerve fibres of Xenopus
currents in isolated frog nerve fibre under voltage clamp laevis. Journal of Physiology, 208, 279-289.
conditions. Journal of Physiology, 143, 76-90. Huxley, A. F. (1959). Ion movements during nerve activity.
Dodge, F. A., and Frankenhaeuser, B. (1959). Sodium Annals of the New York Academy of Sciences, 81, 221-246.
currents in the myelinated nerve fibre of Xenopus laevis
investigated with the voltage clamp technique. Journal of Koles, Z. J., and Rasminsky, M. (1972). A computer simula-
Physiology, 148, 188-200. tion of conduction in demyelinated nerve fibres. Journal of
Fishman, H. M. (1970). Direct and rapid description of the Physiology, 227, 351-364.
individual ionic currents of squid axon membrane by ramp Lumsden, C. E. (1970). The neuropathology of multiple
potential control. Biophysical Journal, 10, 799-817. sclerosis. In Handbook of Clinical Neurology, Vol. 9, pp.
Fitzhugh, R. (1960). Thresholds and plateaus in the Hodgkin- 217-309. Edited by P. J. Vinken and G. W. Bruyn. North-
Huxley nerve equations. Journal of General Physiology, 43, Holland: Amsterdam.
867-896. McDonald, W. I., and Sears, T. A. (1970). Effect of a de-
Fitzhugh, R. (1962). Computation of impulse initiation and myelinating lesion on conduction in the central nervous
saltatory conduction in a myelinated nerve fiber. Bio- system studied in single nerve fibres. Journal of Physiology,
physical Journal, 2, 11-21. 207, 53-54P.
Frankenhaeuser, B. (1959). Steady state inactivation of Namerow, N. S. (1968a). Somatosensory evoked responses in
sodium permeability in myelinated nerve fibres of Xenopus multiple sclerosis. Bulletin of the Los Angeles Neurological
laevis. Journal of Physiology, 148, 671-676. Society, 33, 74-81.
Conduction in multiple sclerosis 161
Namerow, N. S. (1968b). Circadian temperature rhythm and Nelson, D. A., and McDowell, F. (1959). The effects of
vision in multiple sclerosis. Neurology (Minneap.), 18, 417- induced hyperthermia on patients with multiple sclerosis.
422. Journal of Neurology, Neurosurgery, and Psychiatry, 22,
Namerow, N. S. (1972). The pathophysiology of multiple 113-116.
sclerosis. In Multiple Sclerosis. Immunology, Virology, and Rasminsky, M. (1972). Internodal Conduction in Normal and
Ultrastructure, pp. 143-181. Edited by F. Wolfgram, G. W. Demyelinated Mammalian Nerve Fibres. Ph.D. Thesis,
Ellison, J. G. Stevens, and J. M. Andrews. Academic University of London.
Press: New York.
Narahashi, T., and Hass, H. G. (1968). Interaction of DDT Rasminsky, M., and Sears, T. A. (1972). Internodal conduc-
with the components of lobster nerve membrane conduc- tion in undissected demyelinated nerve fibres. Journal of
tance. Journal of General Physiology, 51, 177-198. Physiology, 227, 323-350.
Narahashi, T., Moore, J. W., and Shapiro, B. I. (1969). Smith, R. S., and Koles, Z. J. (1970). Myelinated nerve
Condylactis toxin: interaction with nerve membrane ionic fibres: computed effect of myelin thickness on conduction
conductances. Science, 163, 680-681. velocity. American Journal of Physiology, 219, 1256-1258.
Narahashi, T., Shapiro, B. I., Deguchi, T., Scuka, M., and
Wang, C. M. (1972). Effects of scorpion venom on squid Watson, C. W. (1959). Effect of lowering of body tempera-
axon membranes. American Journal of Physiology, 222, ture on the symptoms and signs of multiple sclerosis. New
850-857. England Journal of Medicine, 261, 1253-1259.