Neuroscience Vol. 37, No. 3, pp. 829-837, 1990 Printed in Great Britain
ELECTROGENIC
03064522/90 $3.00 + 0.00 Pergamon Press plc IBRO
PUMP (Na+/K+-ATPase) RAT OPTIC NERVE
ACTIVITY
IN
T. R. GORDON,* J. D. KOCSISand S. G. WAXMAN Department of Neurology, Yale University School of Medicine and PVA/EPVA Center for Neuroscience Research, West Haven VA Medical Center, West Haven, CT 06516, U.S.A. Abstract-Rat optic nerves were studied in a sucrose gap chamber in order to study the origin of a late afterhyperpolarization that follows repetitive activity. The results provide evidence for electrogenic pump (Na+/K+-ATPase) activity in central nervous system myelinated axons and demonstrate an effect on axonal excitability. Repetitive stimulation (25-200 Hz; 200-5000 ms) led to a prolonged, temperaturedependent post-train afterhyperpolarization with duration up to about 40 s. The post-train afterhyperpolarization was blocked by the Na+/K+-ATPase blockers strophanthidin and ouabain, and the substitution of Li+ for Na+ in the test solution, which also blocks Na+/K+-ATPase. The peak amplitude of the post-train afterhyperpoiarization was minimally changed by the potassium-channel blocker tetraethylammonium (10 mM), and the Ca2+-channel blocker CoCi, (4 mM). Hyperpolarizing constant current did not reverse the afterhyperpolarization. The amplitude of the hyperpolarization was increased in the presence of the potassium-channel blocker 4-aminopyridine (1 mM). In the presence of rl-aminopyridine, the post-train hyperpolarization was much reduced by strophanthidin, except for a residual early component lasting several hundred milliseconds which was blocked by the potassium-channel blocker tetraethylammonium. This finding indicates that after exposure to 4-aminopyridine, repetitive stimulation leads to activation of a tetraethylammonium-sensitive K+-channel that contributes during the first several hundred milliseconds to the post-train afterhyperpolarization. The amplitude of the compound action potential elicited by a single submaximal stimulus during the post-train hyperpolarization was smaller than that of the control response. The decrement in amplitude was not present under identical stimulation conditions when the post-train hy~rpolarization was blocked by strophanthidin, indicting that the hyperpolarization associated with repetitive stimulation reduced excitability. The results indicate that the post-train hyperpolarization in rat optic nerve, a central nervous system myelinated axon bundle, is primarily due to activation of the electrogenic pump (Na+/K+-ATPase); in the presence of 4-aminopyridine, activation of a tetraethylammonium-sensitive K+ channel also contributes to an early component. Activation of the electrogenic pump by repetitive stimulation can lead to axonal hy~rpolarization with consequent decreased excitability which can be reversed by cardiac
glycosides.
A series of afterhy~rpola~zat~ons (AHPs) follows action potential activity in a number of excitable tissues. In many cells, electrogenic pump activity, due to activation of Na+/K+-ATPase, contributes to the series of AHPs. Electrogenic pump activity, manifested by a hyperpolarization that follows trains of electrical impulses, has been demonstrated in a variety of axon bundles including frog peripheral nerve,4 rabbit sympathetic nerve,2s mudpuppy optic nerve,28 rat ventral root2 and the parallel fibers of the rat cerebellum. *’ Hyperpola~zation might be expected to interfere with conduction, especially at sites of reduced safety factor. Baylor and Nicholls’ showed that the hyperpolarization associated with activation of an electrogenic pump in leech sensory neurons can lead to conduction block at sites of inhomogeneity. In frog peripheral nerve, activity-de~ndent decrease in excitability can be prevented by the cardiac glycoside ouabain,24 which blocks the electrogenic pump.24 *To whom correspondence should be addressed. AHP, afterh~r~la~zation; 4-AP, 4-aminopyridine; PTHP, post-train hyperpolarization; TEA, tetraethylammonium.
~bbr~~~atio~:
Bostock and Grafe2 demonstrated in rat peripherai nerve that activation of an electrogenic pump by repetitive stimulation increases threshold, due to hyperpolarization. From findings in normal and focally demyelinated nerve, they suggested that conduction block at sites of demyelination might be improved by blockade of the electrogenic pump, Recently, Kaji and Sumneri used somatosensory evoked potentials to study central axonal conduction in rats with spinal cord demyelinating lesions. Their results suggest that conduction vetocity and high frequency impulse transmission can be improved in a model of CNS demyelination after systemic treatment with ouabain, which blocks the electrogenic pump. Subsequently they showed that ouabain can restore conduction in single demyelinated fibers of rat spinal cord. I3 In a preliminary study, Kaji and Sumner’6 have also demonstrated that the cardiac glycoside digoxin can improve symptoms in human patients with multiple sclerosis, with concomitant improvement in visual-evoked potentials, without unacceptable side effects. The rat optic nerve provides an excellent model for the study of mammalian CNS myelinated axons.
829
T. R. GORDONet al.
830
Action potentials in the optic nerve are followed by a series of AHPs but the ionic basis for these AHPs is not yet clear. Previous studies of mammalian CNS white matter have shown that activity-dependent rises in extracellular K+ are followed by undershoots in extracellular K+@) that are sensitive to strophanthidiq6 indicating the presence of an electrogenic pump. More recently, two pharmacologically distinct K+ channels have been demonstrated on rat optic nerve,“.‘Y manifested by two distinct AHPs.” The presence of a Na +/K +-dependent inwardly rectifying channel has also been shown.’ EXPERIMENTAL
PROCEDURES
The fibers of the rat optic nerve are virtually 100% myelinated and the nerve is easily dissected and can be maintained in vitro for several hours. However, the fibers are small (mean diameter = 0.7 pm)9 so intra-axonal impalements can be achieved only transiently” and reliable nodal or patch clamp recording is impractical. In order to study the electrophysiology of the optic nerve, a modified sucrose gap recording method was used, which allowed recording of changes in membrane potential, and current clamp of the nerve.“,” Young adult (five- to eight-weeks-old) Wistar rats were deeply anesthetized with sodium pentobarbital (60 mg/kg,
i.p.) and killed by decapitation. Their optic nerves (n = 39) were rapidly dissected and placed in modified Krebs’ solution (in mM): NaCI 124, KC1 3.0, NaH,PO, 1.3, MgCIZ 2.0, CaCl, 2.0, NaHCO, 26.0 and dextrose 10.0 saturated with 95% 0, and 5% CO,. In the sucrose gap (Fig. IA), the nerve was positioned across three compartments containing three continuously flowing solutions: modified Krebs’ solution in the test compartment, isotonic sucrose (320 mM) on the middle segment of nerve, and isotonic KC1 (in mM); KC1 120, NaCl 7.0, NaH,PO, 1.3, MgCl, 2.0, NaHCO, 26.0 and dextrose 10.0 in the third compartment. The compartments were separated by silicon vacuum sealant. A bipolar Teflon-coated stainless steel stimulating electrode positioned along the nerve segment in the test compartment was used for direct whole nerve stimulation. Current pulses (100 ps) were generated by a digital timer led to a stimulus isolation unit. Ag-AgCl wire electrodes were positioned in the moditied Krebs’ solution and KC1 compartments. The wire electrodes were embedded in 3.0% agar mixed in modified Krebs’ solution and isotonic KCI, respectively. The electrode in the KC1 compartment was led to the reference input of the amplifier headstage, and the other electrode to the headstage ground. The nerve segment in the test compartment was kept short, in order to minimize action potential propagation and temporal dispersion. Because sucrose is virtually non-conductive and isotonic KC1 inactivates the nerve, the nerve was active in only the test compartment. The electrical circuit is completed intraaxonally between the test and KCI compartments; the whole
-
F I
1OmV
lOm8
atrophanthidin Fig. I. (A) The optic nerve was positioned across three continuously flowing baths: (a) isotonic KCI, (b) isotonic sucrose, (c) normal solution, to which pharma~logical agents were added. The nerve was stimulated with a bipolar electrode in the modified Krebs’ solution bath and recorded differentially across the isotonic KC1 and modified Krebs’ solution baths. Constant current could be passed through the outer baths to achieve a current clamp. A representative compound action potential is shown on the right. With the convention employed throughout this study, depolarization is downward. (B) At room temperature (22°C) repetitive stimulation at 100 Hz for 200 ms led to a marked post-train hype~oIa~zation (PTHP). (C) The PTHP was blocked by strophanthidin leading to prolonged membrane depolarization. The stimulus trains in this and subsequent figures are cropped; their details are not interpretable because they are beyond the resolution of the recording oscilloscope at the sweeps depicted. They are shown only to indicate repetitive stimulation. The voltage and time scales in C are the same as in B.
Electrogenic pump in rat optic nerve nerve responses reflect transmembrane d.c. changes.‘O Figure IA shows the response to a single stimulus in the sucrose gap. An upward deflection indicates intra-axonal hyperpolarization with the recording convention utilized.” Constant current generated by the amplifier could be passed through the Ag-AgCl/agar electrodes leading to changes in membrane potential and providing for current clamp of the nerve. Only nerves with stable compound action potentials greater than 20mV with supramaximal stimulation were used for analysis. Solution flows were maintained by gravity. Pharmacological agents [0.05-50 p M strophanthidin; IO mM tetraethylammonium (TEA); 1.OmM 4-aminopyridine (4-AP); 4.0 mM CoCl,; 25 PM ouabain] were added to modified Krebs’ solution and solution changes made with a siphon. For experiments in which Li+ was substituted for Na+ in modified Krebs’ solution, choline phosphate replaced NaH,PO,. and LiCl reulaced NaCl on a millieauivalent basis: Temperature was maintained by a heating coil in a water jacket in which the tubing to the test compartment was immersed. A thermister in the test compartment was led to a temperature control unit that regulated the output of the heating coil. The nerves were allowed to stabilize for 30 min in the recording chamber before data were collected. Unless otherwise indicated, standard repetitive stimulation consisted of IOO-ps pulses delivered at 100 Hz for 2000 ms, with a 60-s rest period after each stimulus train. Individual action potentials in the stimulus trains depicted were beyond the resolution of the oscilloscope at slow sweep speeds, but could be resolved with faster sweep speeds.
3
1
I
P ii
831
1
1
I 00
I
l
10
20
30
40
.
50
Strophanthtdfn (1LNlf
Fig. 2. The PTHP amplitude of a representative nerve at 32°C is plotted against strophanthidin concentration, Note that the post-train potential amplitude is negative at 50 PM, indicating depolarization. A strophanthidin concentration of 50 PM was chosen for subsequent experiments in order to reach a near-maximal effect.
RESULTS Repetitive stimulation of the room temperature (about 22°C) hyperpolarization (PTHP) (Fig. 40 s. The PTHP was blocked (50 PM), a cardiac glycoside that
rat optic
nerve
at
led to a post-train 1B) lasting up to by strophanthidin
blocks the electrogenie pump2’ (Fig. IC). Figure 2 shows that the effect of strophanthidin on the PTHP is dose-dependent. Note that in 50/1M strophanthidin, the post-train potential has a slightly negative value, indicating depolarization. The blocking effect of each strophanthidin concentration was maximal after 5 min, with little effect on the compound action potential. The amplitude of the PTHP was dependent on the stimulus frequency. Figure 3A shows an increase in the amplitude of the PTHP when the stimulus frequency is increased from 25 to 100 Hz and the duration of the stimulus train is kept constant (2 s). The relationship between stimulus frequency and PTHP peak amplitude is shown in Fig. 3B. Up to 200 Hz, an increase in stimulus frequency leads to an increase in amplitude of the PTHP. Note that at 500 Hz, PTHP amplitude decreases, rather than increases, probably reflecting refractoriness of some fibers in the nerve with such a high stimulus frequency. The amplitude of the PTHP was also dependent on the duration of stimulation. Figure 3C shows an increase in amplitude of the response to 100 Hz stimulation when the stimulus train duration is changed over a range of 500 ms to 5 s. The relationship between PTHP peak amplitude and duration of stimulation for a representative nerve is illustrated in
Fig. 3D. A longer duration stimulus train led to a larger PTHP. The effects of stimulus frequency and duration on the PTHP were independent of the order in which the stimulus protocol was carried out, indicating that the effects are reversible and not accumulative. It is well established that the electrogenic pump is temperature-dependent. 26 The PTHP of the optic nerve is also temperature-dependent as shown in Fig. 4Al where a larger peak amplitude PTHP is present at 34°C than at 26°C. A graph showing the temperature-dependency of the PTHP peak amplitude for a different nerve is plotted in Fig. 4A2, showing that the amplitude increases with temperature. The substitution of Li+ for Na+ does not appreciably affect the inward conductance responsible for the action potential, i* but is known to block the electrogenic pump.‘R*23In rat optic nerve, the substitution of Li+ for Na+ in the test solution led to reversible partial block of PTHP (Fig. 4Bl). Note that the PTHP is prolonged after washout of Li+ in NS. Li+ had only a small decremental effect on the optic nerve compound action potential amplitude after 5 min (Fig. 4B2) despite its effect on the PTHP. Longer application of Li+ led to complete block of the PTHP but also led to marked decrement in the compound action potential amplitude. The effect of another cardiac glycoside, ouabain, on the PTHP is shown in Fig. 5A. The PTHP was completely blocked by ouabain. The effect was partially reversible after 30min washout in modified
832
T. R.
1
GORDON
et
al.
*w 0..
3. 0
100
I. 200
Stimulus
I. 300 Frequency
I, 400
I 500
(Hz)
.
.
0 0
1
2 Stimulus
3
4
5
Duration (see)
Fig. 3. (A) Repetitive stimulation for 2000 ms led to a larger amplitude PTHP at 100 Hz than at 25 Hz. (B) PTHP amplitude is plotted against stimulus frequency for a representative nerve. Stimulus duration is 2000 ms. Note the decrement in amplitude at 500 Hz, indicating refractoriness. (C) At a stimulus frequency of 100 Hz, the PTHP amplitude was larger after a 5000-ms train than after 500 ms. The voltage scale is the same as in A. The two PTHPs have been superimposed and the timebase of the stimulus trams truncated. (D) PTHP amplitude after repetitive stimulation is plotted against stimulus duration. Stimulus frequency is 100 Hz. Temperature was 32°C.
solution. Ouabain had no appreciable effect on compound action potential amplitude to a single supramaximal stimulus (Fig. 5B) despite blockade of the PTHP. As in normal solution, in ouabain the optic nerve was able to reliably follow a lO.O-Hz, 2-s stimulus train, while the PTHP was completely blocked by ouabain. The last four action potentials elicited by such a train is shown in Fig. 5C. Note the decrement in the amplitude of the compound action potentials compared with that in Fig. 5B, indicating relative refractoriness, which also occurs in modified Krebs’ sofution. Because there are two distinct AHPs in rat optic nerve that are blocked by 4-AP and TEA, respectively,ic.1.19 the effects of potassium-channel blockers on the PTHP were tested. TEA alone had little effect on the maximal amplitude of the PTHP but the duration of the PTHP was decreased (Fig. 6AI). The combination of 4-AP and TEA ted to a marked increase in the amplitude of the PTHP (Fig. 6A2); subsequent addition of strophanthidin led to a complete block of the PTHP (Fig. 6B). 4-AP alone also led to an increase in PTHP amplitude (Fig. 6C). The addition of strophanthidin to 4-AP blocked the PTHP, except for an initial component that was sensitive to subsequent addition of TEA (Fig. 6D), indicating that in 4-AP, a TEA-sensitive voltage Krebs’
response contributes to the initial component of the PTHP. Co*+ did not decrease the amplitude of the PTHP, suggesting that a calcium-activated K” current conductance does not contribute to the PTHP (Fig. 6E). The passage of hyperpolarizing or depolarizing constant current, which can reverse or increase, respectively, conductance through potassium channels, had little effect on the PTHP (Fig. 6F). The constant current levels used were sufficient to reverse or increase TEA- and 4-AP-sensitive K+ currents in optic nerve,” although the absolute values of the current could not be determined, because of the limitations of the sucrose gap technique. The effect of the PTHP on whole nerve excitability in modified Krebs’ solution was tested by comparing a control response with a single submaximal stimulus to that delivered during PTHP activation (Fig. 7A2). The amplitude of the compound action potential was decreased during the PTHP, reflecting an increase in the number of fibers that are subthreshold due to hy~rpolarization. This decrease in excitability was reversed by strophanthidin as shown in Fig. 7B. In Fig. 8, the compound action potential amplitude of a single response elicited by a submaximal stimulus is plotted against the time after repetitive stimulation for modified Krebs’ solution and strophanthidin, in
833
Electrogenic pump in rat optic nerve A
2
1
4-
. .
.
f”C)
35
.
2mv
,
I_
P 2t
0
!sooms
/’ 20
25 wnpem:re
40
wash
5ms
Fig. 4. (A) (1) A standard stimulus train (see Experimental Procedures) led to a larger PTHP at 34°C than at 26°C. The relationship between PTHP amplitude and temperature for a representative nerve is plotted in 2. (B) (1) The PTHP in modified Krebs’ solution was partially blocked by the substitution of Lit for Na+ in the test solution after Smin. Washout in modified Krebs’ solution led to subsequent increase in the PTHP. (2) The arrow points to the peak amplitude of the compound action potential in Lit after 5 min. The control is unlabeled. The scale in Bl is the same as in Al.
A
%
If--6nu
ISm
Fig. 5. (A) The PTHP in modified Krebs’ solution was blocked by ouabain. After washout in modified Krebs’ solution for 30min, this effect was somewhat reversible. (B) A compound action potential in ouabain is superimposed on one in modified Krebs’ solution; there is no appreciable effect. C shows the last four compound action potentials of a stimulus train in ouabain, indicating that ouabain does not prevent the optic nerve from following a stimulus train. The relative ~fracto~n~s in ouabain (compare C with B, the voltage scales are the same) is not sufficient to account for the block of the PTHP.
834
T. R. GORDONet al. TEA+4-AP TEA+I-AP
smv I
*
T~+4-AP+strophanthldin
4.AP+strophanthidin /.f
Fig. 6. (A) The duration of the PTHP was decreased somewhat in TEA alone. although its maximal amplitude was not decreased, compared with the PTHP in modified Krebs solution (unla~led trace). The addition of 4-AP to TEA led to a marked increase in the size of the PTHP. (B) The PTHP in TEA and 4-AP in combination was blocked by the addition of strophanthidin. Note that the voltage and time scales are different from those in A, although the nerve is the same. (C) The amplitude of the PTHP in 4-AP alone is greater than that in modified Krebs’ solution. (D) Subsequent addition of strophanthidin to the 4-AP-containing solution blocked the PTHP, except for a brief component. This residual component of the PTHP was sensitive to TEA. The scale in D is the same as in C. (E) The PTHP in modified Krebs’ solution was not decreased by the addition of Co’+. The scale in E is the same as in C. (F) The passage of hyperpolarizing (trace I) and depolarizing (trace 2) constant current had little effect on the PTHP in modified Krebs’ solution (unlabeled), although the currents used were sufficient to reverse or markedly increase, respectively, K+ currents.”
a reeresentative optic nerve. Excitability was consistently decreased during the PTHP and increased in strophanthidin (n = 5) relative to that in modified Krebs’ solution. DISCUSSION
Repetitive action potential activity in a variety of excitable cells is followed by a series of AHPs mediated by several ionic conductances and electrogenie pump activity. Repetitive stimulation is a strong stimulus for activation of the electrogenic pump because it increases intra-axonal [Na+] and leads to accumulation of extra-axonal K+, both of which activate the electrogenic pump.23 Two distinct AHPs, mediated by pharmacologically distinct potassium channels, are observed in rat optic nerve.” In this CNS tract it has also been shown that activity-dependent accumulation of extra-axonal K + is appreciable5,6,8 and might provide a powerful activator for the electrogenic pump. Undershoots in activity-dependent [K.‘f, in optic nerve have been attributed to electrogenic pump activity because of their sensitivity to strophanthidin.’ The present results indicate that the PTHP in rat optic nerve is due, at least in large part, to activation of the
electrogenic pump and not K+ currents. The PTHP is completely blocked by cardiac glycosides in a dose-dependent manner (Figs lB,C; 2 and 5), is increased with increased stimulus frequency and duration (Fig. 3), is temperature-sensitive (Fig. 4A), and is blocked by the substitution of Li+ for Na+ in the test solution (Fig. 4B). Axonal excitability is decreased during the PTHP and the decreased excitability is reversed by strophanthidin {Figs 7 and 8). The effects of cardiac glycosides on the PTHP cannot be attributed to a decrease in the amplitude of the compound action potential nor to the relative refractoriness that occurs during repetitive stimulation (Fig. 5). Although two pha~acolo~caIly distinct potassium channels are known to be present on rat optic nerve,10.‘l.19 the PTHP is not primarily due to the activation of potassium conductances. TEA alone decreased the duration of the PTHP (Fig. 6A) but did not decrease its peak amplitude. The TEA-sensitive potassium-channel that has been demonstrated in rat optic nerve, based on intracellular and whole nerve voltage recordings, inactivates more rapidly than the PTHP.‘0,“.‘9 A more likely explanation for the effect of TEA is that there is a decrease in extracellular potassium in the presence of TEA, due to block of a
835
Electrogenic pump in rat optic nerve A
1
2
I _v-....___
NS
NS 15 mV I
600 ma
-
strophanthidin
Fig. 7. (A) (1) To assess the effect of the PTHP on excitability, the response to a single submaximal stimulus delivered during the PTHP was compared with that of a single submaxim~ stimulus of the same intensity in modified Krebs’ solution. The arrow indicates the point of stimulation during the PTHP (the single response is beyond the resolution of the oscilloscope with the sweep speed). (2) The modified Krebs’ solution control response was larger than that obtained during the PTHP (post-train) indicating decreased excitability during the PTHP. (B) (1) In strophanthidin, a PTHP was not present, and 2, the response to a single submaximal stimulus delivered at the same time as in A after repetitive stimulation was greater in strophanthidin than it was in modified Krebs’ solution, indicating an increase in axonal excitability in strophanthidin, under the experimental conditions. The train in B was the same as that in A, but appears different because of the resolution limit of the recording oscilloscope. The scale in Bl is the same as in A I.
TEA-sensitive channel, with consequent decrease in the duration of K+-mediated pump activation. The combined effect of TEA and 4-AP was to increase PTHP amplitude (Fig. 6A), a result that might be due to the spontaneous bursting known to occur under these conditions in rat optic nerve.‘O Such bursting could lead to increased intra-axonal Na+ concen12.51P.O- .
Fig. 8. The amplitude of the compound action potential elicited by a single submaximal stimulus is plotted against the time of stimulation after a standard stimulus train. The compound action potential amplitude is greater in strophanthidin than in modified Krebs’ solution up until about 5 s after the stimulus train. The time course of the difference is similar to that of the PTHP (e.g. Fig. 6B).
tration23 and to increased extra-axonal K+ concentration, which is activity-dependent6.’ thereby increasing pump activation. Another possibility is that the increase in amplitude reflects increased membrane resistance due to block of potassium conductance. The increase in amplitude of the PTHP in the presence of 4-AP alone (Fig. 6C) may reflect increased membrane resistance or an increase in sodium loading, due to broadening of the action potential in the presence of 4-AP, which is known to occur,” and to a paradoxical rise in extracellular potassium, as occurs in rat cerebe11um.22 Many optic nerve fibers repetitively fire spontaneously in the presence of 4-AP with consequent activation of TEA-sensitive potassium channels.” A rise in extracellular K+ due to activation of these channels might also increase Nat/K+-ATPase activity. In 4-AP there is a component of the PTHP that can be attributed to the TEA-sensitive potassium conductance that is seen in the presence of 4-AP, “Jo but it lasts only several hundred milliseconds (Fig. 6D), whereas the PTHP lasts for seconds. Although rat retinal ganglion cells display Ca*+ and Ca-activated K+ channels,*’ evidence for the presence of such channels on rat optic nerve is not seen.” CoCl,, which blocks Ca*+ channels, had
836
T. R. GORDON~~
little effect on the PTHP, indicating that activation of a Ca*+-dependent K+-channel is unlikely to contribute to the PTHP (Fig. 6E). The small increase in amplitude may be secondary to an increase in input resistance of axonal membranes. Further evidence that the PTHP is not primarily due to activation of a K +-channel is the finding that depolarizing and hyperpolarizing current steps that are sufficient to increase or attenuate and reverse the amplitude of AHPs due to activation of K+ channels in rat optic nerve ” had only minor effects on the PTHP (Fig. 6F). The results indicate that a significant component of the PTHP in rat optic nerve is due to activation of the electrogenic pump and not to activation of K+ channels. In rat optic nerve, a 4-AP-sensitive K+channel that is rapidly activated and inactivated appears to be important for rapid repolarization of the action potential, whereas a more slowly activated and inactivated, TEA-sensitive K+-channel may attenuate repetitive firing. lo.19Evidence indicates that a Na+- and K+-dependent inwardly rectifying channel in rat optic nerve’ might act to attenuate hyperpolarization associated with activation of potassium channels or an electrogenic pump. In rat peripheral nerve, activation of an electrogenic pump during repetitive stimulation leads to a decrease in excitability.* The present study shows that activation of the electrogenic pump in rat optic nerve leads to a decrease in whole nerve excitability as shown by a decrease in the amplitude of the compound action elicited by a submaximal stimulus delivered during the PTHP (Fig. 7A). An explanation for this is that a subpopulation of fibers cannot reach threshold due to hyperpolarization by the electrogenic pump. By blocking the electrogenic pump with cardiac glycosides, excitability can be increased in rat optic nerve (Figs 7B and 8). The possible role of glial Na+/K+-ATPase in the PTHP in rat optic nerve is not clear. In mudpuppy optic nerve, activity-dependent electrogenic pump activity is thought to be axonal without significant contribution by glia, despite the presence of a
al.
K+-dependent, strophanthidin-sensitive electrogenic pump.28 In mudpuppy optic nerve it seems likely that the pump-mediated hyperpolarization reflects axonal, and not glial, pump activity as intracellular Na+ loading in glial cells is not associated with trains of action potentials. Studies on experimentally demyelinated axons have shown that the potassium-channel blocker 4-AP can reverse conduction block after demyelination.3,29 In vivo studies show that doses of 4-AP adequate to reverse conduction block in rats with spinal cord demyelinating lesions can lead to convulsions.‘7 Experimental clinical trials of 4-AP to improve conduction in humans with CNS demyelinating disorders have been associated with side effects such as paresthesias and seizures.‘3,27 The effects of 4-AP on rat optic nerve include broadening of the action potential and action potential bursting after a single stimulus,” findings that could account for some of the side effects observed in humans. Kaji and SumnerI showed that ouabain can improve conduction through sites of demyelination in rat spinal cord. In a preliminary experimental trial of digoxin for the treatment of multiple sclerosis, they showed that physiological and symptomatic improvement can be achieved without unacceptable side effectsI The present study shows that, in addition to two AHPs mediated, respectively, by fast and slow potassium channels, in rat optic nerve, a central myelinated axon bundle, there is a PTHP that is due primarily to electrogenic pump activity. Furthermore, blockade of the electrogenic pump with cardiac glycosides can increase excitability to submaximal stimuli in normal CNS myelinated fibers without the action potential broadening and bursting associated with 4-AP blockade of potassium channels. Acknowledgements-This work was supported in part by grants from the National Institutes of Health and the National Multiple Sclerosis Society, and by the Medical Research Service, Veterans Administration. T.R.G. was supported in part by a grant from the Daniel Heumann Fund.
REFERENCES 1. Baylor D. A. and Nicholls J. G. (1969) After-effects of nerve impulses on signalling in the central nervous system of leech. J. Physiol., Lond. 203, 571-589. 2. Bostock H. and Grafe P. (1985) Activity-dependent excitability changes in normal and demyelinated rat spinal root axons. J. Physiol., Lond. 365, 239-257. 3. Bostock H., Sears T. A. and Sherratt R. M. (1981) The effects of 4-aminopyridine and tetraethylammonium ions on normal and demyelinated mammalian nerve fibers. J. Physiol., Land. 313, 301-315. 4. Connelly C. M. (1959) Recovery processes and metabolism of nerve. Rev. mod. Physiol. 31, 475-484. 5. Connors B. W. and Ransom B. R. (1984) Chloride conductance and extracellular potassium concentration interact to modify the excitability of rat optic nerve fibres. J. Physiol., Lond. 355, 619-633. 6. Connors B. W., Ransom B. R., Kunis D. M. and Gutnick M. J. (1982) Activity-dependent K+ accumulation in the developing rat optic nerve. Science 216, 1341-1343. 7. Ene D. L.. Gordon T. R.. Kocsis J. D. and Waxman S. G. (1990) Current-clamp analysis of a time-dependent rec&cation in rat optic nerve. J. Physiol., Land. 420, 185-202. 8. Forstl J., Galvan M. and Ten Bruggencate G. (1982) Extracellular K+ concentration during electrical stimulation of rat isolated sympathetic ganglia, vagus and optic nerves, Neuroscience 7, 3221-3229. 9. Foster R. E., Connors B. and Waxman S. G. (1982) Rat optic nerve: electrophysiological, pharmacological and anatomical studies during development. Deal Brain Res. 3, 371-386.
Eiectrogenic pump in rat optic nerve
837
10. Gordon T. R., Kocsis J. D. and Waxman S. G. (1988) Evidence for the presence of two types of potassium channels in rat optic nerve. Brain Res. 447, l-9. 11. Gordon T. R., Kocsis J. D. and Waxman S. G. (1989) Pha~a~logi~i ~nsitivities of two afterhy~~ola~~tions in rat optic nerve. Brain Res. 502, 252-257. 12. Hodgkin A. L. and Katz B. (1949) The effect of sodium ions on the electrical activity of giant axons of the squid. J. Pkysiof., Lond. 108, 37-X’. 13. Jones R. E., Heron J. R., Foster D. H., Snelgar R. S. and Mason R. J. (1983) Effects of 4-aminopyridine in patients with multiple sclerosis. f. Neuroi. Sci. 60, 353-362. 14. Kaji R. and Sumner A. J. (1989) Effect of digitalis on central demyelinative conduction block in oioo. Ann. Neurol. 25, 159-167. IS. Kaji R. and Sumner A. J. (1989) Ouabain reverses conduction disturbances in single demyelinated nerve fibers. Neurology 39, 136441366. 16. Kaji R. and Sumner A. J. (1989) Effect of digoxin on clinical symptoms in patients with multiple sclerosis. Ann. Neural. 26, 185. 17. Kaji R. and Sumner A. J. (1988) Effects of 4-aminopyridine in experimental CNS demyelination. Neurology 38, 1884-1887. 18. Keynes R. D. and Swan R. C. (1959) The effects of external sodium concentration on the sodium fluxes in from skeletal muscle. J. Pkysiol., Lund. 147, 591-625. 19. Kocsis J. D., Gordon T. R. and Waxman S. G. (1986) Mammalian optic nerve fibers display two pharmacologically distinct potassium channels. Bruin Res. 393, 357-366. 20. Kocsis J. D., Malenka R. C. and Waxman S. G. (1983) Effects of extracellular potassium concentration on the excitability of the parallel fibres of the rat cerebellum. J. Pkysiol., Land. 334, 225-244. 21. Lipton S. A. and Tauck D. L. (1987) Voltage-dependent conductances of solitary ganglion cells dissociated from the rat retina. J. Pkysiol., Land. 385, 361.-391. 22. Nicholson C., Ten Bruggencate G. and Senekowitsch R. (1976) Large potassium signals and slow potentials evoked during aminopyridine or barium superfusion in cat cerebellum. Brain Res. 113, 606-619. 23. Rang H. P. and Ritchie J. M. (1968) On the electrogenic sodium pump in mammalian non-myelinated nerve fibres and its activation by various external cations. J. Pkysiol., Lond. I%, 183-221. 24. Raymond S. A. and Lettvin J. Y. (1978) Afterelfects of activity in peripheral axons as a clue to nervous coding. In Physiology and Patkobiology of Axons fed. Waxman S. G. ), pp. 203-225. Raven Press, New York. 25. Ritchie J. M. and Straub R. W. (1957) The hy~~ola~~tion which follows activity in mammalian non-medulla&d fibres. J. Pkysiol., Lond. 136, 80-97. 26. Senft J.P. (1967) Effects of some inhibitors on the temperature dependent component of resting potential in lobster axon. J. gen. Pkysiol. 50, 1835-1948. 27. Stefoski D., Davis F. A., Faut M. and Schauf C. L. (1987) 4-Aminopyridine improves clinical signs in multiple sclerosis. Ann. Neuroi. 21, 71-77. 28. Tang C.-M., Cohen M. W. and Orkand R. K. (1980) Electrogenic pumps in axons and neurogha and extracellular potassium homeostasis. Brain Res. 194, 283-286. 29. Targ E. G. and Kocsis J. D. (1985) CAminopyridine leads to restoration of conduction in demyelinated rat sciatic nerve. Brain Res. 328, 358-361. (Accepted 27 March 1990)
ELECTROGENIC
03064522/90 $3.00 + 0.00 Pergamon Press plc IBRO
PUMP (Na+/K+-ATPase) RAT OPTIC NERVE
ACTIVITY
IN
T. R. GORDON,* J. D. KOCSISand S. G. WAXMAN Department of Neurology, Yale University School of Medicine and PVA/EPVA Center for Neuroscience Research, West Haven VA Medical Center, West Haven, CT 06516, U.S.A. Abstract-Rat optic nerves were studied in a sucrose gap chamber in order to study the origin of a late afterhyperpolarization that follows repetitive activity. The results provide evidence for electrogenic pump (Na+/K+-ATPase) activity in central nervous system myelinated axons and demonstrate an effect on axonal excitability. Repetitive stimulation (25-200 Hz; 200-5000 ms) led to a prolonged, temperaturedependent post-train afterhyperpolarization with duration up to about 40 s. The post-train afterhyperpolarization was blocked by the Na+/K+-ATPase blockers strophanthidin and ouabain, and the substitution of Li+ for Na+ in the test solution, which also blocks Na+/K+-ATPase. The peak amplitude of the post-train afterhyperpoiarization was minimally changed by the potassium-channel blocker tetraethylammonium (10 mM), and the Ca2+-channel blocker CoCi, (4 mM). Hyperpolarizing constant current did not reverse the afterhyperpolarization. The amplitude of the hyperpolarization was increased in the presence of the potassium-channel blocker 4-aminopyridine (1 mM). In the presence of rl-aminopyridine, the post-train hyperpolarization was much reduced by strophanthidin, except for a residual early component lasting several hundred milliseconds which was blocked by the potassium-channel blocker tetraethylammonium. This finding indicates that after exposure to 4-aminopyridine, repetitive stimulation leads to activation of a tetraethylammonium-sensitive K+-channel that contributes during the first several hundred milliseconds to the post-train afterhyperpolarization. The amplitude of the compound action potential elicited by a single submaximal stimulus during the post-train hyperpolarization was smaller than that of the control response. The decrement in amplitude was not present under identical stimulation conditions when the post-train hy~rpolarization was blocked by strophanthidin, indicting that the hyperpolarization associated with repetitive stimulation reduced excitability. The results indicate that the post-train hyperpolarization in rat optic nerve, a central nervous system myelinated axon bundle, is primarily due to activation of the electrogenic pump (Na+/K+-ATPase); in the presence of 4-aminopyridine, activation of a tetraethylammonium-sensitive K+ channel also contributes to an early component. Activation of the electrogenic pump by repetitive stimulation can lead to axonal hy~rpolarization with consequent decreased excitability which can be reversed by cardiac
glycosides.
A series of afterhy~rpola~zat~ons (AHPs) follows action potential activity in a number of excitable tissues. In many cells, electrogenic pump activity, due to activation of Na+/K+-ATPase, contributes to the series of AHPs. Electrogenic pump activity, manifested by a hyperpolarization that follows trains of electrical impulses, has been demonstrated in a variety of axon bundles including frog peripheral nerve,4 rabbit sympathetic nerve,2s mudpuppy optic nerve,28 rat ventral root2 and the parallel fibers of the rat cerebellum. *’ Hyperpola~zation might be expected to interfere with conduction, especially at sites of reduced safety factor. Baylor and Nicholls’ showed that the hyperpolarization associated with activation of an electrogenic pump in leech sensory neurons can lead to conduction block at sites of inhomogeneity. In frog peripheral nerve, activity-de~ndent decrease in excitability can be prevented by the cardiac glycoside ouabain,24 which blocks the electrogenic pump.24 *To whom correspondence should be addressed. AHP, afterh~r~la~zation; 4-AP, 4-aminopyridine; PTHP, post-train hyperpolarization; TEA, tetraethylammonium.
~bbr~~~atio~:
Bostock and Grafe2 demonstrated in rat peripherai nerve that activation of an electrogenic pump by repetitive stimulation increases threshold, due to hyperpolarization. From findings in normal and focally demyelinated nerve, they suggested that conduction block at sites of demyelination might be improved by blockade of the electrogenic pump, Recently, Kaji and Sumneri used somatosensory evoked potentials to study central axonal conduction in rats with spinal cord demyelinating lesions. Their results suggest that conduction vetocity and high frequency impulse transmission can be improved in a model of CNS demyelination after systemic treatment with ouabain, which blocks the electrogenic pump. Subsequently they showed that ouabain can restore conduction in single demyelinated fibers of rat spinal cord. I3 In a preliminary study, Kaji and Sumner’6 have also demonstrated that the cardiac glycoside digoxin can improve symptoms in human patients with multiple sclerosis, with concomitant improvement in visual-evoked potentials, without unacceptable side effects. The rat optic nerve provides an excellent model for the study of mammalian CNS myelinated axons.
829
T. R. GORDONet al.
830
Action potentials in the optic nerve are followed by a series of AHPs but the ionic basis for these AHPs is not yet clear. Previous studies of mammalian CNS white matter have shown that activity-dependent rises in extracellular K+ are followed by undershoots in extracellular K+@) that are sensitive to strophanthidiq6 indicating the presence of an electrogenic pump. More recently, two pharmacologically distinct K+ channels have been demonstrated on rat optic nerve,“.‘Y manifested by two distinct AHPs.” The presence of a Na +/K +-dependent inwardly rectifying channel has also been shown.’ EXPERIMENTAL
PROCEDURES
The fibers of the rat optic nerve are virtually 100% myelinated and the nerve is easily dissected and can be maintained in vitro for several hours. However, the fibers are small (mean diameter = 0.7 pm)9 so intra-axonal impalements can be achieved only transiently” and reliable nodal or patch clamp recording is impractical. In order to study the electrophysiology of the optic nerve, a modified sucrose gap recording method was used, which allowed recording of changes in membrane potential, and current clamp of the nerve.“,” Young adult (five- to eight-weeks-old) Wistar rats were deeply anesthetized with sodium pentobarbital (60 mg/kg,
i.p.) and killed by decapitation. Their optic nerves (n = 39) were rapidly dissected and placed in modified Krebs’ solution (in mM): NaCI 124, KC1 3.0, NaH,PO, 1.3, MgCIZ 2.0, CaCl, 2.0, NaHCO, 26.0 and dextrose 10.0 saturated with 95% 0, and 5% CO,. In the sucrose gap (Fig. IA), the nerve was positioned across three compartments containing three continuously flowing solutions: modified Krebs’ solution in the test compartment, isotonic sucrose (320 mM) on the middle segment of nerve, and isotonic KC1 (in mM); KC1 120, NaCl 7.0, NaH,PO, 1.3, MgCl, 2.0, NaHCO, 26.0 and dextrose 10.0 in the third compartment. The compartments were separated by silicon vacuum sealant. A bipolar Teflon-coated stainless steel stimulating electrode positioned along the nerve segment in the test compartment was used for direct whole nerve stimulation. Current pulses (100 ps) were generated by a digital timer led to a stimulus isolation unit. Ag-AgCl wire electrodes were positioned in the moditied Krebs’ solution and KC1 compartments. The wire electrodes were embedded in 3.0% agar mixed in modified Krebs’ solution and isotonic KCI, respectively. The electrode in the KC1 compartment was led to the reference input of the amplifier headstage, and the other electrode to the headstage ground. The nerve segment in the test compartment was kept short, in order to minimize action potential propagation and temporal dispersion. Because sucrose is virtually non-conductive and isotonic KC1 inactivates the nerve, the nerve was active in only the test compartment. The electrical circuit is completed intraaxonally between the test and KCI compartments; the whole
-
F I
1OmV
lOm8
atrophanthidin Fig. I. (A) The optic nerve was positioned across three continuously flowing baths: (a) isotonic KCI, (b) isotonic sucrose, (c) normal solution, to which pharma~logical agents were added. The nerve was stimulated with a bipolar electrode in the modified Krebs’ solution bath and recorded differentially across the isotonic KC1 and modified Krebs’ solution baths. Constant current could be passed through the outer baths to achieve a current clamp. A representative compound action potential is shown on the right. With the convention employed throughout this study, depolarization is downward. (B) At room temperature (22°C) repetitive stimulation at 100 Hz for 200 ms led to a marked post-train hype~oIa~zation (PTHP). (C) The PTHP was blocked by strophanthidin leading to prolonged membrane depolarization. The stimulus trains in this and subsequent figures are cropped; their details are not interpretable because they are beyond the resolution of the recording oscilloscope at the sweeps depicted. They are shown only to indicate repetitive stimulation. The voltage and time scales in C are the same as in B.
Electrogenic pump in rat optic nerve nerve responses reflect transmembrane d.c. changes.‘O Figure IA shows the response to a single stimulus in the sucrose gap. An upward deflection indicates intra-axonal hyperpolarization with the recording convention utilized.” Constant current generated by the amplifier could be passed through the Ag-AgCl/agar electrodes leading to changes in membrane potential and providing for current clamp of the nerve. Only nerves with stable compound action potentials greater than 20mV with supramaximal stimulation were used for analysis. Solution flows were maintained by gravity. Pharmacological agents [0.05-50 p M strophanthidin; IO mM tetraethylammonium (TEA); 1.OmM 4-aminopyridine (4-AP); 4.0 mM CoCl,; 25 PM ouabain] were added to modified Krebs’ solution and solution changes made with a siphon. For experiments in which Li+ was substituted for Na+ in modified Krebs’ solution, choline phosphate replaced NaH,PO,. and LiCl reulaced NaCl on a millieauivalent basis: Temperature was maintained by a heating coil in a water jacket in which the tubing to the test compartment was immersed. A thermister in the test compartment was led to a temperature control unit that regulated the output of the heating coil. The nerves were allowed to stabilize for 30 min in the recording chamber before data were collected. Unless otherwise indicated, standard repetitive stimulation consisted of IOO-ps pulses delivered at 100 Hz for 2000 ms, with a 60-s rest period after each stimulus train. Individual action potentials in the stimulus trains depicted were beyond the resolution of the oscilloscope at slow sweep speeds, but could be resolved with faster sweep speeds.
3
1
I
P ii
831
1
1
I 00
I
l
10
20
30
40
.
50
Strophanthtdfn (1LNlf
Fig. 2. The PTHP amplitude of a representative nerve at 32°C is plotted against strophanthidin concentration, Note that the post-train potential amplitude is negative at 50 PM, indicating depolarization. A strophanthidin concentration of 50 PM was chosen for subsequent experiments in order to reach a near-maximal effect.
RESULTS Repetitive stimulation of the room temperature (about 22°C) hyperpolarization (PTHP) (Fig. 40 s. The PTHP was blocked (50 PM), a cardiac glycoside that
rat optic
nerve
at
led to a post-train 1B) lasting up to by strophanthidin
blocks the electrogenie pump2’ (Fig. IC). Figure 2 shows that the effect of strophanthidin on the PTHP is dose-dependent. Note that in 50/1M strophanthidin, the post-train potential has a slightly negative value, indicating depolarization. The blocking effect of each strophanthidin concentration was maximal after 5 min, with little effect on the compound action potential. The amplitude of the PTHP was dependent on the stimulus frequency. Figure 3A shows an increase in the amplitude of the PTHP when the stimulus frequency is increased from 25 to 100 Hz and the duration of the stimulus train is kept constant (2 s). The relationship between stimulus frequency and PTHP peak amplitude is shown in Fig. 3B. Up to 200 Hz, an increase in stimulus frequency leads to an increase in amplitude of the PTHP. Note that at 500 Hz, PTHP amplitude decreases, rather than increases, probably reflecting refractoriness of some fibers in the nerve with such a high stimulus frequency. The amplitude of the PTHP was also dependent on the duration of stimulation. Figure 3C shows an increase in amplitude of the response to 100 Hz stimulation when the stimulus train duration is changed over a range of 500 ms to 5 s. The relationship between PTHP peak amplitude and duration of stimulation for a representative nerve is illustrated in
Fig. 3D. A longer duration stimulus train led to a larger PTHP. The effects of stimulus frequency and duration on the PTHP were independent of the order in which the stimulus protocol was carried out, indicating that the effects are reversible and not accumulative. It is well established that the electrogenic pump is temperature-dependent. 26 The PTHP of the optic nerve is also temperature-dependent as shown in Fig. 4Al where a larger peak amplitude PTHP is present at 34°C than at 26°C. A graph showing the temperature-dependency of the PTHP peak amplitude for a different nerve is plotted in Fig. 4A2, showing that the amplitude increases with temperature. The substitution of Li+ for Na+ does not appreciably affect the inward conductance responsible for the action potential, i* but is known to block the electrogenic pump.‘R*23In rat optic nerve, the substitution of Li+ for Na+ in the test solution led to reversible partial block of PTHP (Fig. 4Bl). Note that the PTHP is prolonged after washout of Li+ in NS. Li+ had only a small decremental effect on the optic nerve compound action potential amplitude after 5 min (Fig. 4B2) despite its effect on the PTHP. Longer application of Li+ led to complete block of the PTHP but also led to marked decrement in the compound action potential amplitude. The effect of another cardiac glycoside, ouabain, on the PTHP is shown in Fig. 5A. The PTHP was completely blocked by ouabain. The effect was partially reversible after 30min washout in modified
832
T. R.
1
GORDON
et
al.
*w 0..
3. 0
100
I. 200
Stimulus
I. 300 Frequency
I, 400
I 500
(Hz)
.
.
0 0
1
2 Stimulus
3
4
5
Duration (see)
Fig. 3. (A) Repetitive stimulation for 2000 ms led to a larger amplitude PTHP at 100 Hz than at 25 Hz. (B) PTHP amplitude is plotted against stimulus frequency for a representative nerve. Stimulus duration is 2000 ms. Note the decrement in amplitude at 500 Hz, indicating refractoriness. (C) At a stimulus frequency of 100 Hz, the PTHP amplitude was larger after a 5000-ms train than after 500 ms. The voltage scale is the same as in A. The two PTHPs have been superimposed and the timebase of the stimulus trams truncated. (D) PTHP amplitude after repetitive stimulation is plotted against stimulus duration. Stimulus frequency is 100 Hz. Temperature was 32°C.
solution. Ouabain had no appreciable effect on compound action potential amplitude to a single supramaximal stimulus (Fig. 5B) despite blockade of the PTHP. As in normal solution, in ouabain the optic nerve was able to reliably follow a lO.O-Hz, 2-s stimulus train, while the PTHP was completely blocked by ouabain. The last four action potentials elicited by such a train is shown in Fig. 5C. Note the decrement in the amplitude of the compound action potentials compared with that in Fig. 5B, indicating relative refractoriness, which also occurs in modified Krebs’ sofution. Because there are two distinct AHPs in rat optic nerve that are blocked by 4-AP and TEA, respectively,ic.1.19 the effects of potassium-channel blockers on the PTHP were tested. TEA alone had little effect on the maximal amplitude of the PTHP but the duration of the PTHP was decreased (Fig. 6AI). The combination of 4-AP and TEA ted to a marked increase in the amplitude of the PTHP (Fig. 6A2); subsequent addition of strophanthidin led to a complete block of the PTHP (Fig. 6B). 4-AP alone also led to an increase in PTHP amplitude (Fig. 6C). The addition of strophanthidin to 4-AP blocked the PTHP, except for an initial component that was sensitive to subsequent addition of TEA (Fig. 6D), indicating that in 4-AP, a TEA-sensitive voltage Krebs’
response contributes to the initial component of the PTHP. Co*+ did not decrease the amplitude of the PTHP, suggesting that a calcium-activated K” current conductance does not contribute to the PTHP (Fig. 6E). The passage of hyperpolarizing or depolarizing constant current, which can reverse or increase, respectively, conductance through potassium channels, had little effect on the PTHP (Fig. 6F). The constant current levels used were sufficient to reverse or increase TEA- and 4-AP-sensitive K+ currents in optic nerve,” although the absolute values of the current could not be determined, because of the limitations of the sucrose gap technique. The effect of the PTHP on whole nerve excitability in modified Krebs’ solution was tested by comparing a control response with a single submaximal stimulus to that delivered during PTHP activation (Fig. 7A2). The amplitude of the compound action potential was decreased during the PTHP, reflecting an increase in the number of fibers that are subthreshold due to hy~rpolarization. This decrease in excitability was reversed by strophanthidin as shown in Fig. 7B. In Fig. 8, the compound action potential amplitude of a single response elicited by a submaximal stimulus is plotted against the time after repetitive stimulation for modified Krebs’ solution and strophanthidin, in
833
Electrogenic pump in rat optic nerve A
2
1
4-
. .
.
f”C)
35
.
2mv
,
I_
P 2t
0
!sooms
/’ 20
25 wnpem:re
40
wash
5ms
Fig. 4. (A) (1) A standard stimulus train (see Experimental Procedures) led to a larger PTHP at 34°C than at 26°C. The relationship between PTHP amplitude and temperature for a representative nerve is plotted in 2. (B) (1) The PTHP in modified Krebs’ solution was partially blocked by the substitution of Lit for Na+ in the test solution after Smin. Washout in modified Krebs’ solution led to subsequent increase in the PTHP. (2) The arrow points to the peak amplitude of the compound action potential in Lit after 5 min. The control is unlabeled. The scale in Bl is the same as in Al.
A
%
If--6nu
ISm
Fig. 5. (A) The PTHP in modified Krebs’ solution was blocked by ouabain. After washout in modified Krebs’ solution for 30min, this effect was somewhat reversible. (B) A compound action potential in ouabain is superimposed on one in modified Krebs’ solution; there is no appreciable effect. C shows the last four compound action potentials of a stimulus train in ouabain, indicating that ouabain does not prevent the optic nerve from following a stimulus train. The relative ~fracto~n~s in ouabain (compare C with B, the voltage scales are the same) is not sufficient to account for the block of the PTHP.
834
T. R. GORDONet al. TEA+4-AP TEA+I-AP
smv I
*
T~+4-AP+strophanthldin
4.AP+strophanthidin /.f
Fig. 6. (A) The duration of the PTHP was decreased somewhat in TEA alone. although its maximal amplitude was not decreased, compared with the PTHP in modified Krebs solution (unla~led trace). The addition of 4-AP to TEA led to a marked increase in the size of the PTHP. (B) The PTHP in TEA and 4-AP in combination was blocked by the addition of strophanthidin. Note that the voltage and time scales are different from those in A, although the nerve is the same. (C) The amplitude of the PTHP in 4-AP alone is greater than that in modified Krebs’ solution. (D) Subsequent addition of strophanthidin to the 4-AP-containing solution blocked the PTHP, except for a brief component. This residual component of the PTHP was sensitive to TEA. The scale in D is the same as in C. (E) The PTHP in modified Krebs’ solution was not decreased by the addition of Co’+. The scale in E is the same as in C. (F) The passage of hyperpolarizing (trace I) and depolarizing (trace 2) constant current had little effect on the PTHP in modified Krebs’ solution (unlabeled), although the currents used were sufficient to reverse or markedly increase, respectively, K+ currents.”
a reeresentative optic nerve. Excitability was consistently decreased during the PTHP and increased in strophanthidin (n = 5) relative to that in modified Krebs’ solution. DISCUSSION
Repetitive action potential activity in a variety of excitable cells is followed by a series of AHPs mediated by several ionic conductances and electrogenie pump activity. Repetitive stimulation is a strong stimulus for activation of the electrogenic pump because it increases intra-axonal [Na+] and leads to accumulation of extra-axonal K+, both of which activate the electrogenic pump.23 Two distinct AHPs, mediated by pharmacologically distinct potassium channels, are observed in rat optic nerve.” In this CNS tract it has also been shown that activity-dependent accumulation of extra-axonal K + is appreciable5,6,8 and might provide a powerful activator for the electrogenic pump. Undershoots in activity-dependent [K.‘f, in optic nerve have been attributed to electrogenic pump activity because of their sensitivity to strophanthidin.’ The present results indicate that the PTHP in rat optic nerve is due, at least in large part, to activation of the
electrogenic pump and not K+ currents. The PTHP is completely blocked by cardiac glycosides in a dose-dependent manner (Figs lB,C; 2 and 5), is increased with increased stimulus frequency and duration (Fig. 3), is temperature-sensitive (Fig. 4A), and is blocked by the substitution of Li+ for Na+ in the test solution (Fig. 4B). Axonal excitability is decreased during the PTHP and the decreased excitability is reversed by strophanthidin {Figs 7 and 8). The effects of cardiac glycosides on the PTHP cannot be attributed to a decrease in the amplitude of the compound action potential nor to the relative refractoriness that occurs during repetitive stimulation (Fig. 5). Although two pha~acolo~caIly distinct potassium channels are known to be present on rat optic nerve,10.‘l.19 the PTHP is not primarily due to the activation of potassium conductances. TEA alone decreased the duration of the PTHP (Fig. 6A) but did not decrease its peak amplitude. The TEA-sensitive potassium-channel that has been demonstrated in rat optic nerve, based on intracellular and whole nerve voltage recordings, inactivates more rapidly than the PTHP.‘0,“.‘9 A more likely explanation for the effect of TEA is that there is a decrease in extracellular potassium in the presence of TEA, due to block of a
835
Electrogenic pump in rat optic nerve A
1
2
I _v-....___
NS
NS 15 mV I
600 ma
-
strophanthidin
Fig. 7. (A) (1) To assess the effect of the PTHP on excitability, the response to a single submaximal stimulus delivered during the PTHP was compared with that of a single submaxim~ stimulus of the same intensity in modified Krebs’ solution. The arrow indicates the point of stimulation during the PTHP (the single response is beyond the resolution of the oscilloscope with the sweep speed). (2) The modified Krebs’ solution control response was larger than that obtained during the PTHP (post-train) indicating decreased excitability during the PTHP. (B) (1) In strophanthidin, a PTHP was not present, and 2, the response to a single submaximal stimulus delivered at the same time as in A after repetitive stimulation was greater in strophanthidin than it was in modified Krebs’ solution, indicating an increase in axonal excitability in strophanthidin, under the experimental conditions. The train in B was the same as that in A, but appears different because of the resolution limit of the recording oscilloscope. The scale in Bl is the same as in A I.
TEA-sensitive channel, with consequent decrease in the duration of K+-mediated pump activation. The combined effect of TEA and 4-AP was to increase PTHP amplitude (Fig. 6A), a result that might be due to the spontaneous bursting known to occur under these conditions in rat optic nerve.‘O Such bursting could lead to increased intra-axonal Na+ concen12.51P.O- .
Fig. 8. The amplitude of the compound action potential elicited by a single submaximal stimulus is plotted against the time of stimulation after a standard stimulus train. The compound action potential amplitude is greater in strophanthidin than in modified Krebs’ solution up until about 5 s after the stimulus train. The time course of the difference is similar to that of the PTHP (e.g. Fig. 6B).
tration23 and to increased extra-axonal K+ concentration, which is activity-dependent6.’ thereby increasing pump activation. Another possibility is that the increase in amplitude reflects increased membrane resistance due to block of potassium conductance. The increase in amplitude of the PTHP in the presence of 4-AP alone (Fig. 6C) may reflect increased membrane resistance or an increase in sodium loading, due to broadening of the action potential in the presence of 4-AP, which is known to occur,” and to a paradoxical rise in extracellular potassium, as occurs in rat cerebe11um.22 Many optic nerve fibers repetitively fire spontaneously in the presence of 4-AP with consequent activation of TEA-sensitive potassium channels.” A rise in extracellular K+ due to activation of these channels might also increase Nat/K+-ATPase activity. In 4-AP there is a component of the PTHP that can be attributed to the TEA-sensitive potassium conductance that is seen in the presence of 4-AP, “Jo but it lasts only several hundred milliseconds (Fig. 6D), whereas the PTHP lasts for seconds. Although rat retinal ganglion cells display Ca*+ and Ca-activated K+ channels,*’ evidence for the presence of such channels on rat optic nerve is not seen.” CoCl,, which blocks Ca*+ channels, had
836
T. R. GORDON~~
little effect on the PTHP, indicating that activation of a Ca*+-dependent K+-channel is unlikely to contribute to the PTHP (Fig. 6E). The small increase in amplitude may be secondary to an increase in input resistance of axonal membranes. Further evidence that the PTHP is not primarily due to activation of a K +-channel is the finding that depolarizing and hyperpolarizing current steps that are sufficient to increase or attenuate and reverse the amplitude of AHPs due to activation of K+ channels in rat optic nerve ” had only minor effects on the PTHP (Fig. 6F). The results indicate that a significant component of the PTHP in rat optic nerve is due to activation of the electrogenic pump and not to activation of K+ channels. In rat optic nerve, a 4-AP-sensitive K+channel that is rapidly activated and inactivated appears to be important for rapid repolarization of the action potential, whereas a more slowly activated and inactivated, TEA-sensitive K+-channel may attenuate repetitive firing. lo.19Evidence indicates that a Na+- and K+-dependent inwardly rectifying channel in rat optic nerve’ might act to attenuate hyperpolarization associated with activation of potassium channels or an electrogenic pump. In rat peripheral nerve, activation of an electrogenic pump during repetitive stimulation leads to a decrease in excitability.* The present study shows that activation of the electrogenic pump in rat optic nerve leads to a decrease in whole nerve excitability as shown by a decrease in the amplitude of the compound action elicited by a submaximal stimulus delivered during the PTHP (Fig. 7A). An explanation for this is that a subpopulation of fibers cannot reach threshold due to hyperpolarization by the electrogenic pump. By blocking the electrogenic pump with cardiac glycosides, excitability can be increased in rat optic nerve (Figs 7B and 8). The possible role of glial Na+/K+-ATPase in the PTHP in rat optic nerve is not clear. In mudpuppy optic nerve, activity-dependent electrogenic pump activity is thought to be axonal without significant contribution by glia, despite the presence of a
al.
K+-dependent, strophanthidin-sensitive electrogenic pump.28 In mudpuppy optic nerve it seems likely that the pump-mediated hyperpolarization reflects axonal, and not glial, pump activity as intracellular Na+ loading in glial cells is not associated with trains of action potentials. Studies on experimentally demyelinated axons have shown that the potassium-channel blocker 4-AP can reverse conduction block after demyelination.3,29 In vivo studies show that doses of 4-AP adequate to reverse conduction block in rats with spinal cord demyelinating lesions can lead to convulsions.‘7 Experimental clinical trials of 4-AP to improve conduction in humans with CNS demyelinating disorders have been associated with side effects such as paresthesias and seizures.‘3,27 The effects of 4-AP on rat optic nerve include broadening of the action potential and action potential bursting after a single stimulus,” findings that could account for some of the side effects observed in humans. Kaji and SumnerI showed that ouabain can improve conduction through sites of demyelination in rat spinal cord. In a preliminary experimental trial of digoxin for the treatment of multiple sclerosis, they showed that physiological and symptomatic improvement can be achieved without unacceptable side effectsI The present study shows that, in addition to two AHPs mediated, respectively, by fast and slow potassium channels, in rat optic nerve, a central myelinated axon bundle, there is a PTHP that is due primarily to electrogenic pump activity. Furthermore, blockade of the electrogenic pump with cardiac glycosides can increase excitability to submaximal stimuli in normal CNS myelinated fibers without the action potential broadening and bursting associated with 4-AP blockade of potassium channels. Acknowledgements-This work was supported in part by grants from the National Institutes of Health and the National Multiple Sclerosis Society, and by the Medical Research Service, Veterans Administration. T.R.G. was supported in part by a grant from the Daniel Heumann Fund.
REFERENCES 1. Baylor D. A. and Nicholls J. G. (1969) After-effects of nerve impulses on signalling in the central nervous system of leech. J. Physiol., Lond. 203, 571-589. 2. Bostock H. and Grafe P. (1985) Activity-dependent excitability changes in normal and demyelinated rat spinal root axons. J. Physiol., Lond. 365, 239-257. 3. Bostock H., Sears T. A. and Sherratt R. M. (1981) The effects of 4-aminopyridine and tetraethylammonium ions on normal and demyelinated mammalian nerve fibers. J. Physiol., Land. 313, 301-315. 4. Connelly C. M. (1959) Recovery processes and metabolism of nerve. Rev. mod. Physiol. 31, 475-484. 5. Connors B. W. and Ransom B. R. (1984) Chloride conductance and extracellular potassium concentration interact to modify the excitability of rat optic nerve fibres. J. Physiol., Lond. 355, 619-633. 6. Connors B. W., Ransom B. R., Kunis D. M. and Gutnick M. J. (1982) Activity-dependent K+ accumulation in the developing rat optic nerve. Science 216, 1341-1343. 7. Ene D. L.. Gordon T. R.. Kocsis J. D. and Waxman S. G. (1990) Current-clamp analysis of a time-dependent rec&cation in rat optic nerve. J. Physiol., Land. 420, 185-202. 8. Forstl J., Galvan M. and Ten Bruggencate G. (1982) Extracellular K+ concentration during electrical stimulation of rat isolated sympathetic ganglia, vagus and optic nerves, Neuroscience 7, 3221-3229. 9. Foster R. E., Connors B. and Waxman S. G. (1982) Rat optic nerve: electrophysiological, pharmacological and anatomical studies during development. Deal Brain Res. 3, 371-386.
Eiectrogenic pump in rat optic nerve
837
10. Gordon T. R., Kocsis J. D. and Waxman S. G. (1988) Evidence for the presence of two types of potassium channels in rat optic nerve. Brain Res. 447, l-9. 11. Gordon T. R., Kocsis J. D. and Waxman S. G. (1989) Pha~a~logi~i ~nsitivities of two afterhy~~ola~~tions in rat optic nerve. Brain Res. 502, 252-257. 12. Hodgkin A. L. and Katz B. (1949) The effect of sodium ions on the electrical activity of giant axons of the squid. J. Pkysiof., Lond. 108, 37-X’. 13. Jones R. E., Heron J. R., Foster D. H., Snelgar R. S. and Mason R. J. (1983) Effects of 4-aminopyridine in patients with multiple sclerosis. f. Neuroi. Sci. 60, 353-362. 14. Kaji R. and Sumner A. J. (1989) Effect of digitalis on central demyelinative conduction block in oioo. Ann. Neurol. 25, 159-167. IS. Kaji R. and Sumner A. J. (1989) Ouabain reverses conduction disturbances in single demyelinated nerve fibers. Neurology 39, 136441366. 16. Kaji R. and Sumner A. J. (1989) Effect of digoxin on clinical symptoms in patients with multiple sclerosis. Ann. Neural. 26, 185. 17. Kaji R. and Sumner A. J. (1988) Effects of 4-aminopyridine in experimental CNS demyelination. Neurology 38, 1884-1887. 18. Keynes R. D. and Swan R. C. (1959) The effects of external sodium concentration on the sodium fluxes in from skeletal muscle. J. Pkysiol., Lund. 147, 591-625. 19. Kocsis J. D., Gordon T. R. and Waxman S. G. (1986) Mammalian optic nerve fibers display two pharmacologically distinct potassium channels. Bruin Res. 393, 357-366. 20. Kocsis J. D., Malenka R. C. and Waxman S. G. (1983) Effects of extracellular potassium concentration on the excitability of the parallel fibres of the rat cerebellum. J. Pkysiol., Land. 334, 225-244. 21. Lipton S. A. and Tauck D. L. (1987) Voltage-dependent conductances of solitary ganglion cells dissociated from the rat retina. J. Pkysiol., Land. 385, 361.-391. 22. Nicholson C., Ten Bruggencate G. and Senekowitsch R. (1976) Large potassium signals and slow potentials evoked during aminopyridine or barium superfusion in cat cerebellum. Brain Res. 113, 606-619. 23. Rang H. P. and Ritchie J. M. (1968) On the electrogenic sodium pump in mammalian non-myelinated nerve fibres and its activation by various external cations. J. Pkysiol., Lond. I%, 183-221. 24. Raymond S. A. and Lettvin J. Y. (1978) Afterelfects of activity in peripheral axons as a clue to nervous coding. In Physiology and Patkobiology of Axons fed. Waxman S. G. ), pp. 203-225. Raven Press, New York. 25. Ritchie J. M. and Straub R. W. (1957) The hy~~ola~~tion which follows activity in mammalian non-medulla&d fibres. J. Pkysiol., Lond. 136, 80-97. 26. Senft J.P. (1967) Effects of some inhibitors on the temperature dependent component of resting potential in lobster axon. J. gen. Pkysiol. 50, 1835-1948. 27. Stefoski D., Davis F. A., Faut M. and Schauf C. L. (1987) 4-Aminopyridine improves clinical signs in multiple sclerosis. Ann. Neuroi. 21, 71-77. 28. Tang C.-M., Cohen M. W. and Orkand R. K. (1980) Electrogenic pumps in axons and neurogha and extracellular potassium homeostasis. Brain Res. 194, 283-286. 29. Targ E. G. and Kocsis J. D. (1985) CAminopyridine leads to restoration of conduction in demyelinated rat sciatic nerve. Brain Res. 328, 358-361. (Accepted 27 March 1990)
