- Email: [email protected]
Electrical activity of mouse motor endings during muscle reinnervation
N~uroshwcr Vol.
16. No. 4. pp. 1047-1056.
0306~4522/85
1985
Printedin Great Britain
ELECTRICAL
ENDINGS
$3.00+0.00
Pergamon Press Lid [’ 1985IBRO
ACTItiITY
DURING
OF MOUSE MOTOR
MUSCLE REINNERVATION
D. ANGAUT-PETIT and A. MALLART Unit6 de Physiologie Neuromusculaire, Laboratoire de Neurobiologie Cellulaire, C.N.R.S. 91190~Gif-sur-Yvette, France Abstract-An in uirro study of electrical activity of regenerating motor endings was performed I l-15 days after the crushing of one motor nerve supplying the triangularis sterni muscle in the adult mouse. For this purpose, presynaptic membrane currents elicited by electrical stimulation of the regenerating nerve were recorded by external electrodes. Ionic channel distribution along the length of the endings was deduced from wave form configuration in normal perfusing fluid together with changes produced by application of specific channel blocking agents. The sharp negative deflection which was shown to correspond to inward Na+ current by its sensitivity to tetrodotoxin application could be recorded along most of the length of the endings indicating a widespread distribution of Na channels. Frequent absence of the late wave form component which signals K’ current was taken to indicate an even K+ current density in the few last nodes, the heminode and the distal part of the endings. Therefore, it appears that regenerating motor endings are characterized by an overlap of Na and K conductances all along their length. In the courSe of regeneration, the heminode loses the sensitivity to K channel blocking agents and the remainder of the terminal becomes insensitive to tetrodotoxin, the former change occurring usually earlier than the latter.
Heterogeneous ionic channel distribution has been shown to exist in mature mammalian myelinated fibre
membranes. A high density of sodium channels exists exclusively at nodes of Ranvier27,32which contain few if any potassium channels.5~7~‘0The latter are located in the internodal membrane.8,9 Studies of development or regeneration of myelinated fibres have shoCn that this segregation does not occur in immature motor axons where a potassium conductance is involved in action potential electrogenesis.‘5.‘7~26 A study concerning voltage-sensitive conductances in adult mouse motor endings has recently been performed by Brigant and Mallart” who also showed an uneven distribution of membrane permeabilities: a high density of sodium channels is mainly, if not exclusively, located in the immediate vicinity of the myelin end (i.e. in the heminode) while K channels, together with Ca channels are located only in the remainder of the terminal. The purpose of the present study was to investigate channel distribution in regenerating motor endings to know whether or not Na and K channels are segregated from each other along terminal branches, as they are in the normal ones. New terminals formed by regenerating axons of one nerve supplying the triangularis sterni muscle of the adult mouse were studied immediately after resumption of neuromuscular transmission. Part of this work has already been reported briefly.’ EXPERIMENTAL PROCEDURES
The triangularis sterni muscle of the adult mouse was used.2’ One of its nerves was crushed under pentobarbital (Nembutal) anaesthesia. An in uirro study of presynaptic 3,4-DAP, 3,4_diaminopyridine; TEA, tetraethylammonium; TTX, tetrodotoxin.
Abbreviarions:
currents in regenerating endings was performed I l-15 days later. The preparation was superfused at a rate of I5 ml min-‘,gt ;oom temperature (2l”C), by an oxygenated solution containine (mM): NaCI. 154: KCI. 5: CaCI,. 2: glucose, II, and
Collagenase (Sigma Type III) at a concentration of 0.05 mg ml-’ was applied for 30 min to remove connective tissue which covered the denervated-reinnervated area. This procedure associated with the use of Nomarski interference contrast optics and high magnification (x 500) allowed visualization of the myelin end as well as the unmyelinated nerve terminal branches. Figure I shows examples of regenerating endings from a muscle processed for histology I3 days after nerve crush. The regenerating nerve was stimulated by a suction electrode and muscle contraction was blocked by curare (2 PM at the beginning of the experiment to identify newly formed synapses by the presence of evoked release, followed by 20 FM to completely suppress postsynaptic activity). The technique for focal current recording has been described previously:6 the heat-polished recording electrodes (I or 2pm inside diameter), were filled with standard Krebs solution and applied to the outer surface of the membrane under visual control, while a remote Ag-AgCl indifferent electrode was placed in the bath. Signals were fed to a voltage follower stage. Individual signals (2(tlOO) were averaged by means of a calculator Intertechnique (Didac 800), or a digital oscilloscope (Tektronix 7854). Time resolution was 20~s in both cases. Downward deflections correspond to inward currents (negative waves). The principle of focal current recording is based on the flow of longitudinal current between the last few nodes of Ranvier (including the terminal heminode) and the motor endings. In terms of the local circuit theory,” this current is generated by potential differences that exist between adjacent areas of the membrane upon invasion by nerve impulses. As a matter of fact, the current that enters the nodes of Ranvier carried by Na+ flows axially to the endings where it discharges membrane capacity, depolarizes the terminal and promotes an increase in conductance to Ca, K and, to some extent, to Na ions.6,‘RThe current returns to the nodes of origin along the extracellular space. Since the axoplasm and the extracellular space may be regarded as passive conduc-
1047
1048
D. ANGAUT-PETITand A. MALLART
tars, the configuration of the externally recorded current will depend mainly on the membrane ionic conductances to Na+. K+ and Cd’+ at the nodes and at the endings. In the same way, local circuit current may flow between adjacent areas of the motor ending membrane because of differences in potential generated by differences in their respective ionic conductances. The local circuit current leaving or entering the membrane under the electrode is, therefore, net current resulting from the sum of ionic and capacitive inward and outward membrane currents. This current is recorded as a voltage drop across the sealing resistance of the electrode. The recording electrode seemed to be very selective and able to record currents flowing through a surface of a few pm’ of the membrane, since small longitudinal displacements resulted in changes in wave form configu~tion. Use of specific channel blocking drugs applied in the bath or by ionophoresis allowed us to identify the nature and location
of ionic ~rmeabiiities involved as well as to djfl.ercntiate between ionic and passive components of the currents. Tetraethylammonium (TEA; Koch-Light, I mM) -3,4-djaminopyridine (3,4-DAP; Fluka, 0. I mM) were used in bath application. Micropipettes for ionophoretic injection were tilled with a 500 mM TEA f 50 mM 3,4-DAP solution or a 3 x lO-‘M tetrodotoxin (TTX: Sigma) solution Their resistance ranged between 400 and 800 MR.
RESULTS
The muscles undergoing reinnervation were examined 1l-15 days after nerve crush. Transmitting synapses were not found before the 13th day and by the 14th day the recorded electrical signals displayed the “mature” configuration in most of the examined
Fig. 2. Upper part. Typical wave forms recorded from a single ending are shown at their approximate locations on a drawing from a silver impregnated muscle. The recording indicated by the arrow corresponds to the heminode (referred to as preterminal or proximal part by Brigant and Mallart”). The other traces correspond to what we call the main part of the ending. Late negative waves (downward deflections) of large amplitude correspond to end-plate currents. Lower part. First row shows examples of the four types of wave forms that can be distinguished in a proximodistai sequence: A,, monophasic negative wave; B,, triphasic positive--negativ&positive wave; C,, diphasic positive-negative wave and D,, monophasic positive wave. Rows 2 and 3 show signals that exhibit some changes with respect to the above typical recordings. Horizontal bar = 2 ms (same bar = 8 pm for the drawing). Vertical bar = 50 )JV in A,, A,. B,. B,, D,, D, and IOOpV for the other traces.
Fig. 1. Examples of regenerating axons and newly-formed terminals (Bielschowsky silver staining). They are characterized by the small number of their terminal branches. Unbranched and singly branched terminals are shown in A (arrow) and B, respectively. Poorly myelinated last internode is indicated by an arrow in C. Double arrow in A shows an end-plate site which is innervated by terminals emerging from two independent axons. Calibration = 15pm. 1049
Electrical activity of regenerating motor endings end-plates. We, therefore, concentrated our study to the 13th day post-denervation period. As previously shown in a morphological study of triangularis sterni muscles fixed at this period of time after a similar nerve crush, the regenerating endings displayed a variety of branching patterns, whose complexity decreased in a proximodistal sequence, suggesting a decreasing degree of growth (see Angaut-Petit et al.,* Fig. 5). In the present experiments, no precise correlation could be established between the electrical activity and the morphology of individual endings because in many instances we could not distinguish the whole terminal arborization. By contrast, the myelin end of regenerating fibres entering the endplate area was visible despite incomplete myelination at this stage (see Fig. IC). The synapses formed by regenerating axons of the crushed motor nerve exhibited features classically described in muscles undergoing reinnervation: low quanta1 content, long latency of end-plate potentials and rapid fatigue in response to repetitive stimulation.4.1’ Figure 2 (upper part) shows wave forms recorded along a singly branched motor ending. The drawing has been made from a silver impregnation of a similar terminal branch on which we indicated the approximate location of wave forms. These recordings of pre- and postsynaptic activities differed widely from each other. Three types of presynaptic signals could be recognized. Starting at the myelin end, the presynaptic responses were successively monophasic negative, triphasic positive-negative-positive at the next point, then diphasic positive-negative up to the most distal part of the branch. Variations from these three types of signal configuration were often encountered. The most frequently found in the 74 regenerating endings that we investigated are shown in Fig. 2 (lower part). Monophasic negative signals typical of the heminode were sometimes followed by a second negative wave (A2 and A3) of smaller amplitude. In this case, the response closely resembled the wave form recorded in similar regions of mature motor endings.6 Triphasic wave forms showed frequently a late component of complex configuration (for instance B, seems to be an intermediate pattern between signals shown in A, and B,). Triphasic responses as well as diphasic ones could be recorded in many cases from the whole length of the endings. Another type of presynaptic wave form, absent in the end-plate illustrated in Fig. 2 (upper part), but often observed at the most distal part of the endings, consisted of a monophasic positive wave (see D, and D,). Striking variations were observed in the decay time of this wave which in some cases was extremely fast (100 ps in D2). Sometimes the positive signal had two components (DJ in which cases the response resembled the typical response of distal regions in normal endings.6 It appears that a large variability was present among the recorded signals. When two different
1051
branches of a single ending were visible clearly enough to be explored with accuracy, the overall pattern of electrical activity differed frequently from one branch to the other. To better understand the significance of the different types of wave form configuration found here, it is important to recall the nature of the electrical signals recorded from normal endings. In normal adult mouse motor terminals, the capacity current is seen as a positive spike which is followed by a positive wave corresponding to outward K+ current. The currents entering the heminode are almost mirror images of those leaving the terminal: the first negative spike signals inward Na+ current, while the second negative component corresponds to the sink of current promoted by K+ outflux at the terminal part of the endings6 The recordings that characterize regenerating terminals differ from normal ones by two main features. First, the second positive wave in the distal part, which signals outward K+ current, or the second negative deflection at the heminode which corresponds to local circuit current entering this area in fully developed endings, was missing in most of our recordings. A second important difference between responses of regenerating terminals and normal ones is the presence in the former of large negative deflections in diphasic or triphasic signals recorded from most of the length of the terminals. These features may depend on differences either in membrane ionic properties or in nerve terminal geometry between regenerating and normal endings. This question was examined by using specific K or Na channel blockers in bath or in ionophoretic application. Application of K channel blockers
During bath application of TEA + 3,4-DAP an increase in the latency of the presynaptic signal usually occurred (see Brigant and Mallart, 1982) together with changes in its configuration. As shown in Fig. 3A, the late positive deflection of the triphasic wave was readily suppressed by bath application of TEA + 3,4-DAP. In this particular case the recording electrode was probably positioned at a point near the boundary between low and high K channel density areas. Thus, the electrode presumably picked up ionic outward K+ current together with the corresponding sink of local circuit current which caused a broadening of the inward Na spike. A similar situation is shown in Fig. 2B, (see also Mallart, 1984,*’ Fig. IH). Potassium channel blockers suppressed both the late outward current and the broadening of the Na spike. When present, the second positive component of distal responses similarly disappeared (Fig. 3B and C). In some cases the recorded signal was diphasic consisting in a positive capacity spike followed by a late negativity (Fig. 2C,, C,; Fig. 3D, E). The latter may be due either to inward Na+ current or to a sink of current corresponding to a source of outward current in adjacent areas of the membrane. As shown
D. ANGAUT-PETITand A. MALLART
1052
TEA
+DAP
A
8
C
D
E
F Fig. 3. Late wave form components are modified by bath application of K channel blockers. When positive, they are supressed (A, B) or replaced by a negative one (C). When negative, they are suppressed (D) or greatly reduced in ampiitude and duration (E). When no late component is present, no conspicuous change occurs in the presynaptic signal during bath application of the blocking agents (F). Calibration = I ms and 25 pV.
in Fig. 3D and E, bath application of K channel blockers suppressed more or less completely the late negative component, suggesting the presence of outward K + current in neighbouring areas of the membrane. However, a residual negative component may persist (Fig. 3E) or appear (Fig. 3C) during drug action, whose nature will be discussed later. In contrast to the conspicuous changes induced by K channel blockers just described, no obvious changes occurred when the distal response was composed of a singie positive wave, i.e. when late membrane current was missing (Fig. 3F) although in these cases bath application of K channel blockers was able to increase the release of transmitter by nerve impulses. This means that the absence of a late cornponent in presynaptic signals does not necessarily imply an absence of K channels. Rather, uniform
distribution of K channels over the main part of the endings, the preterminal nodes of Ranvier and the heminode will give no differences between falling phases of action potentials generated at these parts of the membrane and thus no late net current will flow along the nerve terminal. The view that a widespread distrjbut~on of K channels existed aiong regenerating terminals was tested by ionophoretic apphcation of specific K channet blocking drugs as close as possible to the recording electrode, as shown in the diagram in Fig. 4. During ionophoresis of the drug at the heminode, a late negative wave either appeared or, when present with a small amplitude, increased in amplitude and duration (Fig. 4A,, A?, arrows). Similarly TEA -I- 3,4-DAP ionophoresis onto the main part of an ending induced the appearance of a fate negative
Electrical activity of regenerating motor endings
1053
2
Fig. 4. Ionophoretic application of K channel blocking agents at the heminode (A) or at the distal part (B) of two regenerating endings (1 and 2). The upper trace in each pair is average of control recordings while the lower one is during drug ionophoresis. The application produces an enhancement in amplitude and duration of late negative deflections (A, and A,, arrows) or their appearance (B, and B,, arrows). The small and prolonged late negative deflection which appears in B, during drug ionophoresis corresponds probably to postsynaptic activity. Horizontal calibration = 1 ms. Vertical calibration is IOOpV in 1, 5OpV in 2.
in the recordings (Fig. 4B,, B,, arrows). Changes in late component configuration can be explained by supposing that the ionophoretic apphcation of K channel blockers suppressed an outward current at each location of the electrode. At the heminode (A,, A,), drug action led to an increase in the amplitude of the late net inward current, while at the terminal part (B,, B2) it unmasked a late net inward current. In cases of endings reaching maturity, where the signals displayed two distinct negative or positive peaks at the heminode and terminal portions respectively (Fig. 2A, and D,, Fig. 3B and C), drug ionophoresis was effective only at the latter.
component
Effect of tetrodotoxin
As described above, (TEA + 3,4-DAP)-resistant diphasic wave forms (see Fig. 3) could be recorded from most of the length of some endings and triphasic signals were found at considerable distance from the heminode, which would indicate active impulse propagation.i4 However, since diphasicity is not an absolute criterium to distinguish between active propagation and electrotonic spread,” we used TTX in ionophoretic application close to the recording electrode to obtain a more precise picture of Na channel distribution in regenerating endings. The large negative wave that characterized signals
from the heminode was immediately abolished by TTX application (Fig. 5A) indicating the presence at this point of an important Na+ influx. TTX also affected responses in more distal areas of some endings. The small negative component of diphasic signals present in these regions (Fig. 5, B,, upper trace) was suppressed within a few seconds of drug application (Fig. 5, B,, lower trace). In other endings, the effect of TTX ionophoresis in distal areas was to enhance the amplitude of outward current or to increase the decay time of the monophasic positive response (Fig. 5B,). Our interpretation of the discrete changes in the outward current elicited by TTX application is that localized suppression of Na current enhances outward capacity current at that point. Either slight or conspicuous the changes promoted by TTX indicate that Na channels are present, perhaps with a low density, up to the most distal part of some regenerating endings. However, in some endplates where wave forms displayed an “immature” look regarding the K component, TTX application was without effect, suggesting that Na channels have withdrawn from the main part of the endings. One can say that in these cases, factors affecting Na channel redistribution precede those controlling K channel availability.
D. ANGAUT-PETIT
and A.
MALLART
Fig. 5. Effect of ionophoretic TTX application on signals recorded from the heminode (A) and from the distal part of regenerating endings (B, and Br). The upper trace in each pair corresponds to control recordings while the lower one is during ‘ITX application. In A: TTX induces the disappearance of the sharp negative deflection. Similarly, in B, the small negative deflection disappears during TTX ionophoresis. A diphasic potential with a small amplitude negative wave persists after TTX ionophoresis in these two examples: it probably corresponds to current flowing between actively and electrotonically depolarized membrane areas (see Ref. 1I). In B,: TTX application induces an increase in the duration of the monophasic positive wave. Calibration: I ms and 100 {tV.
Ca current
The inward current was proved to be promoted by a calcium influx, firstly by its increase in amplitude when the usual calcium concentration of 2 mM was increased to 6mM, and secondly by its complete disappearance when IOmM cobalt were added to the perfusing medium (Fig. 6Br). Since Ca?+ current is recorded with a negative polarity only in the main part of the endings, it follows that Ca channels exist with higher density there, as compared to the heminode. endings.
During bath application of TEA + 3,4-DAP, a negative wave of long duration and low ampI~tude followed in some cases the positive spike recorded at the main part of the endings (Fig. 6A, and 3,). A positive wave which was its mirror image was recorded at the heminode (Fig. 6A,). The time course of these waves was similar to that of the calcium current shown by Mallart and Brigant’j in normal
BI
__ kb?aJL
* A2
.-we.
I
‘I
Fig. 6. A, and A, correspond respectively to recordings from the main part of an ending and its heminode during bath application of K channel blockers + 6 mM calcium. A late component is visible which is negative in A,, positive in A,. The wave form shown in B, was recorded from the main part of another ending in the same experimental conditions. The negative wave in B, disappeared in &?by supplementary addition of IO mM cobalt to the perfusing fluid. Calibration: I ms and 25 ,uV.
Electrical activity of regenerating motor endings DlSCUSSlON
The present study shows that the electrical activity in regenerating motor endings differs in two main features from that of mature terminals. One is the considerable variation in shape and amplitude of the K-dependent wave form component. The other is the length over which a Na-dependent wave form component can be recorded. The fact that the normal type presynaptic wave form coexists in the same muscle with signals displaying features observed only in regenerating endings suggests that the regenerating process is very rapid. A typical feature of presynaptic currents in regenerating nerve endings revealed by the present study is the presence of a K component whose polarity was either negative, indicating an absence or a lower density of K channels and consequently a longer action potential in this area as compared to the heminode and preterminal nodes, or positive, indicating the reverse situation. In some cases no visible K+ current was present, indicating the presence of K channels with a similar density, or accessibility, in the terminals and in the parent axon. Differences in K-dependent components between regenerating and mature endings can be explained by the process of myelin formation which, by insulating inter- and paranodal regions of preterminal axons, suppresses sources for outward K+ current.‘6,26 Our results do not allow us to tell whether this mechanism suffices to account for changes in K component configuration which occur in motor endings as they reach maturation. Nonetheless, to explain the change from a K component of negative polarity observed in some regenerating endings into a positive one which characterizes the normal state, one has to postulate that, in addition to axonal myelination, a real increase in K channel density occurs at the terminals during regeneration. In our experiments, Na channel activity was signalled by the presence along most of the length of regenerating endings of triphasic as well as diphasic waves whose large negative components were reduced or suppressed by ionophoretic TTX application. Nevertheless, the amplitude of inward current at the main part of the terminals was never as large as at the heminode, as shown by the inverse relationship that exists between outward capacitive and inward ionic Na+ current at these two recording sites. That triphasic waves do not originate in growing axons passing nearby is shown by the fact that these signals were only recorded from end-plate areas (as defined by the presence of end-plate nuclei) at electrode emplacements within IO-20pm of the myelin end, and that they changed usually to a diphasic or monophasic positive configuration within the endplate limits. TTX was effective in changing wave form configuration also in cases where no net inward current was visible, indicating the presence of Na channels, however they were insufficient to outweigh
1055
the ionic or passive outward currents occurring at the same place. Finally, cases where TTX ionophoresis failed to reveal the presence of Na channels were also observed. Why does the main part of some regenerating endings exhibit detectable Na+ conductance while others do not, the latter being in this respect comparable to mature endings? Recordings from normal endings failed to show any Na-dependent inward current, while outward K+ and inward Ca*+ currents could be readily demonstrated.6 A possible explanation for our failure to record net inward Na+ current is that Na channels exist in an exceedingly low density at the endings, as compared to the nodes of Ranvier or the heminode. As a matter of fact, binding studies revealed a Na channel density ranging from 25 to 110 sites km -* in unmyelinated fibres, while Na channel density was as high as 12 000 pm-* at the nodes of Ranvier (see Ref. 28). By assuming a similar Na channel density in unmyelinated axons and in nerve endings, one would expect an inward Na+ current at the terminals of negligible intensity as compared to that in the nodes. Under these circumstances, the current that enters the membrane at the nodes carried by Na ions, and leaves it at the endings by discharging membrane capacity would easily outweigh the inward Na+ current at the terminals.‘8 One can therefore conclude that the absence of net inward current which we observed at some regenerating terminals is related to an extremely low density of Na channels. On the contrary, the fact that in many instances we were able to record Na-dependent inward currents of large amplitude, together with the presence of triphasic wave forms over a considerable length of the terminals, which indicates active impulse conduction,‘4 strongly suggests that, unlike in normal endings, Na channels occur with a nonnegligible density in regenerating terminals. A small amplitude wave form component which corresponds to inward Ca*+ current was recorded from regenerating terminal branches in 3,4-DAP + TEA-treated preparations. The use of both K channel blocking agents was required because, in mouse motor endings, voltage-dependent K channels are more readily blocked by 3,4-DAP, while Ca*+-activated K channels are sensitive only to TEA.** The small amplitude of the Ca wave form was an unexpected finding since Ca*+ currents have been reported to be predominant during early stages of development. ‘2,29Furthermore, calcium spikes are present in growing dendrites,” cultured cells3 and regenerating axons.20~25~30 The small amplitude of Ca responses in our preparations can be explained by assuming that a non-negligible amount of inward Ca*+ current is present at the preterminal part of the axon which would cancel, in part, external current promoted by Ca*+ entry at the terminals. Alternatively, the small amplitude of Ca responses would reveal a genuine lower density of Ca channels in regenerating endings as compared to mature ones,
D. ANGAUT-PETIIand A.
1056
which is in agreement with the lower quanta1 content of postsynaptic responses elicited by the former. Our results do not allow us to decide between hypotheses.
these two
MALLAKT
Acknon,/edgmzents-We wish to thank L. Fmlle fur help with histological techniques. and D. Chaslard for typing rhc manuscript. This work was supported by a research ptxnt from INSERM no. 816003.
REFERENCES
I. Angaut-Petit 345, 57P.
D. and Mallart A. (1983) Presynaptic currents in regenerating mouse motor endings. J. Physiol., Lend.
2. Angaut-Petit D., Mallart A. and Faille L. (1982) Role of denervated sheaths and end-plates in muscle reinnervation by collateral sprouting in the mouse. Bioi. Cell. 46, 2777290. 3. Anglister L., Farber I. C., Shahar A. and Grinvald A. (1982) Localization of voltage-sensitive calcium channels along developing neurites: their possible role in regulating neurite elongation. Dec. Biol. 94, 351-365. 4. Bennett M. R. and Florin T. (1974) A statistical analysis of the release of acetylcholine at newly formed synapses m striated muscle. J. Physiol., Land. 238, 93-107. 5. Bostock H., Sears T. A. and Sherratt R. M. (1981) The effects of 4-aminopyridine and tetraethylammonium ions on normal and demyelinated mammalian nerve fibres. J. Physiol., Land. 313, 301-315. 6. Brigant J. L. and Mallart A. (1982) Presynaptic currents in mouse motor endings. /. Physiol., Lend. 333, 619-636. 7. Brismar T. (1980) Potential clamp analysis of membrane currents in rat myelinated nerve fibres. J. Physiol., Lond. 298, 171-184.
8. Chiu S. Y. and Ritchie J. M. (1980) Potassium channels in nodal and internodal axonal membrane of mammalian myelinated fibres. Nature 2B4, 170-171. 9. Chiu S. Y. and Ritchie J. M. (1984) On the physiological role of internodal potassium channels and the security of conduction in myelinated nerve fibres. Proc. R. Sot. Lond. B 220, 415-422. 10. Chiu S. Y., Ritchie J. M., Rogart R. B. and Stagg D. (1979) A quantitative description of membrane currents in rabbit myelinated nerve. J. Physiol., Lond. 292, 149-166. 11. Eccles J. C. (1964) The Physiology of Synapses, p. 124. Academic Press, New York. 12. Hagiwara S. and Byerly L. (1981) Calcium channel. Ann. Rev. Neurosci. 4, 69-125. 13. Hodgkin A. L. (1937) Evidence for electrical transmission in nerve. Part 1. J. Physiol., Land. 90, 183-210. 14. Katz B. and Miledi R. (1965) Propagation of electric activity in motor nerve terminals. Proc. R. Sot. Lond. B 161, 453-482. 15. Kocsis J. D., Ruiz J. A. and Waxman S. G. (1983) Maturation of mammalian myehnated fibers: changes in action-potential characteristics following 4-aminopyridine application. J. Neurophysiol. 50, 449-463. 16. Kocsis J. D. and Waxman S. G. (1983) Electrophysiology of conduction in mammalian regenerating nerves. in Nerre, Organ and Tissue Regeneration: Research Perpecfiues (ed. Seil F. J.), pp. 89-107. Academic Press, New York. 17. Kocsis J. D., Waxman S. G., Hildebrand C. and Ruiz J. A. (1982) Regenerating mammalian nerve fibres: changes in action potential wave form and firing characteristics following blockage of potassium conductance. Proc. R. Sor. Land. B 217, 77-87. 18. Konishi T. and Sears T. A. (1984) Electrical activity of mouse motor nerve terminals. Proc. R. Sot. Lond. t3 222, 115-120.
19. Llinas R. and Sugimori M., (1979) Calcium conductances in Purkinje cell dendrites: their role in development and integration. In Deoelopment and Chemical Specificity of Neurons. Progr. Brain Research (eds Cuinot M.. Kreutzberg G. W. and Bloom F. E.), 51, pp. 323-334. Elsevier, Amsterdam. 20. MacVicar B. A. and Llinas R. (1982) Calcium spikes in regenerating giant axons of the lamprey spinal cord. Sot. Neurosci. Abst 8, 914.
21. Mallart A. (1984) Presynaptic currents in frog motor endings. Pfltigers Arch. 400, 8-13. 22. Mallart A. (1984) Calcium-activated potassium current in presynaptic terminals. Biomed. Res. 5, 287-290. 23. Mallart A. and Brigant J. L. (1982) Electrical activity at motor nerve terminals of the mouse. J. Physiol., Paris 78, 407-41
I.
24. McArdle J. J., Angaut-Petit D., Mallart A., Bournaud R., Faille L. and Brigant J. L. (1981) Advantages of the Triangularis sterni muscle of the mouse for investigations of synaptic phenomena. J. Neurosci. Merh. 4, 109-I 15. 25. Meiri H., Spira M. E. and Parnas I. (1981) Membrane conductance and action potential of a regenerating axonal tip. Science 211, 709-712.
26. Ritchie J. M. (1982) Sodium and potassium channels in regenerating and developing mammalian myelinated nerves. Proc. R. Sot. Lond. B 215, 273-287.
27. Ritchie J. M. and Rogart R. B. (1977) Density of sodium channels in mammalian myelinated nerve fibers and nature of the axonal membrane under the myelin sheath. Proc. narn. Acad. Sci. USA 74, 21 I-215. 28. Rogart R. (1981) Sodium channels in nerve and muscle membrane. Ann. Rev. Physiol. 43, 71 l-725. 29. Spitzer N. C. (_l979) Ion channels in development. Ann. Reo. Neurosci. 2, 363-397. 30. Strichartz G., Small R., Nicholson C., Pfenninger K. H. and Llinls R. (1980) Ionic mechanisms for impulse propagation in growing nonmyelinated axons:saxitoxin binding and electrophysiology. Sot. Neurosci. Abst 6, 660. 31. Tonge D. A. (1974) Physiological characteristics of re-innervation of skeletal muscle in the mouse. J. Physiol., Land. 241, 141-153. 32. Waxman S. G. and Foster R. E. (1980) Ionic channel distribution and heterogeneity of the axon membrane in myelinated fibers. Brain Res. Rev. 2, 205-234. (Accepred 21 June 1985)
16. No. 4. pp. 1047-1056.
0306~4522/85
1985
Printedin Great Britain
ELECTRICAL
ENDINGS
$3.00+0.00
Pergamon Press Lid [’ 1985IBRO
ACTItiITY
DURING
OF MOUSE MOTOR
MUSCLE REINNERVATION
D. ANGAUT-PETIT and A. MALLART Unit6 de Physiologie Neuromusculaire, Laboratoire de Neurobiologie Cellulaire, C.N.R.S. 91190~Gif-sur-Yvette, France Abstract-An in uirro study of electrical activity of regenerating motor endings was performed I l-15 days after the crushing of one motor nerve supplying the triangularis sterni muscle in the adult mouse. For this purpose, presynaptic membrane currents elicited by electrical stimulation of the regenerating nerve were recorded by external electrodes. Ionic channel distribution along the length of the endings was deduced from wave form configuration in normal perfusing fluid together with changes produced by application of specific channel blocking agents. The sharp negative deflection which was shown to correspond to inward Na+ current by its sensitivity to tetrodotoxin application could be recorded along most of the length of the endings indicating a widespread distribution of Na channels. Frequent absence of the late wave form component which signals K’ current was taken to indicate an even K+ current density in the few last nodes, the heminode and the distal part of the endings. Therefore, it appears that regenerating motor endings are characterized by an overlap of Na and K conductances all along their length. In the courSe of regeneration, the heminode loses the sensitivity to K channel blocking agents and the remainder of the terminal becomes insensitive to tetrodotoxin, the former change occurring usually earlier than the latter.
Heterogeneous ionic channel distribution has been shown to exist in mature mammalian myelinated fibre
membranes. A high density of sodium channels exists exclusively at nodes of Ranvier27,32which contain few if any potassium channels.5~7~‘0The latter are located in the internodal membrane.8,9 Studies of development or regeneration of myelinated fibres have shoCn that this segregation does not occur in immature motor axons where a potassium conductance is involved in action potential electrogenesis.‘5.‘7~26 A study concerning voltage-sensitive conductances in adult mouse motor endings has recently been performed by Brigant and Mallart” who also showed an uneven distribution of membrane permeabilities: a high density of sodium channels is mainly, if not exclusively, located in the immediate vicinity of the myelin end (i.e. in the heminode) while K channels, together with Ca channels are located only in the remainder of the terminal. The purpose of the present study was to investigate channel distribution in regenerating motor endings to know whether or not Na and K channels are segregated from each other along terminal branches, as they are in the normal ones. New terminals formed by regenerating axons of one nerve supplying the triangularis sterni muscle of the adult mouse were studied immediately after resumption of neuromuscular transmission. Part of this work has already been reported briefly.’ EXPERIMENTAL PROCEDURES
The triangularis sterni muscle of the adult mouse was used.2’ One of its nerves was crushed under pentobarbital (Nembutal) anaesthesia. An in uirro study of presynaptic 3,4-DAP, 3,4_diaminopyridine; TEA, tetraethylammonium; TTX, tetrodotoxin.
Abbreviarions:
currents in regenerating endings was performed I l-15 days later. The preparation was superfused at a rate of I5 ml min-‘,gt ;oom temperature (2l”C), by an oxygenated solution containine (mM): NaCI. 154: KCI. 5: CaCI,. 2: glucose, II, and
Collagenase (Sigma Type III) at a concentration of 0.05 mg ml-’ was applied for 30 min to remove connective tissue which covered the denervated-reinnervated area. This procedure associated with the use of Nomarski interference contrast optics and high magnification (x 500) allowed visualization of the myelin end as well as the unmyelinated nerve terminal branches. Figure I shows examples of regenerating endings from a muscle processed for histology I3 days after nerve crush. The regenerating nerve was stimulated by a suction electrode and muscle contraction was blocked by curare (2 PM at the beginning of the experiment to identify newly formed synapses by the presence of evoked release, followed by 20 FM to completely suppress postsynaptic activity). The technique for focal current recording has been described previously:6 the heat-polished recording electrodes (I or 2pm inside diameter), were filled with standard Krebs solution and applied to the outer surface of the membrane under visual control, while a remote Ag-AgCl indifferent electrode was placed in the bath. Signals were fed to a voltage follower stage. Individual signals (2(tlOO) were averaged by means of a calculator Intertechnique (Didac 800), or a digital oscilloscope (Tektronix 7854). Time resolution was 20~s in both cases. Downward deflections correspond to inward currents (negative waves). The principle of focal current recording is based on the flow of longitudinal current between the last few nodes of Ranvier (including the terminal heminode) and the motor endings. In terms of the local circuit theory,” this current is generated by potential differences that exist between adjacent areas of the membrane upon invasion by nerve impulses. As a matter of fact, the current that enters the nodes of Ranvier carried by Na+ flows axially to the endings where it discharges membrane capacity, depolarizes the terminal and promotes an increase in conductance to Ca, K and, to some extent, to Na ions.6,‘RThe current returns to the nodes of origin along the extracellular space. Since the axoplasm and the extracellular space may be regarded as passive conduc-
1047
1048
D. ANGAUT-PETITand A. MALLART
tars, the configuration of the externally recorded current will depend mainly on the membrane ionic conductances to Na+. K+ and Cd’+ at the nodes and at the endings. In the same way, local circuit current may flow between adjacent areas of the motor ending membrane because of differences in potential generated by differences in their respective ionic conductances. The local circuit current leaving or entering the membrane under the electrode is, therefore, net current resulting from the sum of ionic and capacitive inward and outward membrane currents. This current is recorded as a voltage drop across the sealing resistance of the electrode. The recording electrode seemed to be very selective and able to record currents flowing through a surface of a few pm’ of the membrane, since small longitudinal displacements resulted in changes in wave form configu~tion. Use of specific channel blocking drugs applied in the bath or by ionophoresis allowed us to identify the nature and location
of ionic ~rmeabiiities involved as well as to djfl.ercntiate between ionic and passive components of the currents. Tetraethylammonium (TEA; Koch-Light, I mM) -3,4-djaminopyridine (3,4-DAP; Fluka, 0. I mM) were used in bath application. Micropipettes for ionophoretic injection were tilled with a 500 mM TEA f 50 mM 3,4-DAP solution or a 3 x lO-‘M tetrodotoxin (TTX: Sigma) solution Their resistance ranged between 400 and 800 MR.
RESULTS
The muscles undergoing reinnervation were examined 1l-15 days after nerve crush. Transmitting synapses were not found before the 13th day and by the 14th day the recorded electrical signals displayed the “mature” configuration in most of the examined
Fig. 2. Upper part. Typical wave forms recorded from a single ending are shown at their approximate locations on a drawing from a silver impregnated muscle. The recording indicated by the arrow corresponds to the heminode (referred to as preterminal or proximal part by Brigant and Mallart”). The other traces correspond to what we call the main part of the ending. Late negative waves (downward deflections) of large amplitude correspond to end-plate currents. Lower part. First row shows examples of the four types of wave forms that can be distinguished in a proximodistai sequence: A,, monophasic negative wave; B,, triphasic positive--negativ&positive wave; C,, diphasic positive-negative wave and D,, monophasic positive wave. Rows 2 and 3 show signals that exhibit some changes with respect to the above typical recordings. Horizontal bar = 2 ms (same bar = 8 pm for the drawing). Vertical bar = 50 )JV in A,, A,. B,. B,, D,, D, and IOOpV for the other traces.
Fig. 1. Examples of regenerating axons and newly-formed terminals (Bielschowsky silver staining). They are characterized by the small number of their terminal branches. Unbranched and singly branched terminals are shown in A (arrow) and B, respectively. Poorly myelinated last internode is indicated by an arrow in C. Double arrow in A shows an end-plate site which is innervated by terminals emerging from two independent axons. Calibration = 15pm. 1049
Electrical activity of regenerating motor endings end-plates. We, therefore, concentrated our study to the 13th day post-denervation period. As previously shown in a morphological study of triangularis sterni muscles fixed at this period of time after a similar nerve crush, the regenerating endings displayed a variety of branching patterns, whose complexity decreased in a proximodistal sequence, suggesting a decreasing degree of growth (see Angaut-Petit et al.,* Fig. 5). In the present experiments, no precise correlation could be established between the electrical activity and the morphology of individual endings because in many instances we could not distinguish the whole terminal arborization. By contrast, the myelin end of regenerating fibres entering the endplate area was visible despite incomplete myelination at this stage (see Fig. IC). The synapses formed by regenerating axons of the crushed motor nerve exhibited features classically described in muscles undergoing reinnervation: low quanta1 content, long latency of end-plate potentials and rapid fatigue in response to repetitive stimulation.4.1’ Figure 2 (upper part) shows wave forms recorded along a singly branched motor ending. The drawing has been made from a silver impregnation of a similar terminal branch on which we indicated the approximate location of wave forms. These recordings of pre- and postsynaptic activities differed widely from each other. Three types of presynaptic signals could be recognized. Starting at the myelin end, the presynaptic responses were successively monophasic negative, triphasic positive-negative-positive at the next point, then diphasic positive-negative up to the most distal part of the branch. Variations from these three types of signal configuration were often encountered. The most frequently found in the 74 regenerating endings that we investigated are shown in Fig. 2 (lower part). Monophasic negative signals typical of the heminode were sometimes followed by a second negative wave (A2 and A3) of smaller amplitude. In this case, the response closely resembled the wave form recorded in similar regions of mature motor endings.6 Triphasic wave forms showed frequently a late component of complex configuration (for instance B, seems to be an intermediate pattern between signals shown in A, and B,). Triphasic responses as well as diphasic ones could be recorded in many cases from the whole length of the endings. Another type of presynaptic wave form, absent in the end-plate illustrated in Fig. 2 (upper part), but often observed at the most distal part of the endings, consisted of a monophasic positive wave (see D, and D,). Striking variations were observed in the decay time of this wave which in some cases was extremely fast (100 ps in D2). Sometimes the positive signal had two components (DJ in which cases the response resembled the typical response of distal regions in normal endings.6 It appears that a large variability was present among the recorded signals. When two different
1051
branches of a single ending were visible clearly enough to be explored with accuracy, the overall pattern of electrical activity differed frequently from one branch to the other. To better understand the significance of the different types of wave form configuration found here, it is important to recall the nature of the electrical signals recorded from normal endings. In normal adult mouse motor terminals, the capacity current is seen as a positive spike which is followed by a positive wave corresponding to outward K+ current. The currents entering the heminode are almost mirror images of those leaving the terminal: the first negative spike signals inward Na+ current, while the second negative component corresponds to the sink of current promoted by K+ outflux at the terminal part of the endings6 The recordings that characterize regenerating terminals differ from normal ones by two main features. First, the second positive wave in the distal part, which signals outward K+ current, or the second negative deflection at the heminode which corresponds to local circuit current entering this area in fully developed endings, was missing in most of our recordings. A second important difference between responses of regenerating terminals and normal ones is the presence in the former of large negative deflections in diphasic or triphasic signals recorded from most of the length of the terminals. These features may depend on differences either in membrane ionic properties or in nerve terminal geometry between regenerating and normal endings. This question was examined by using specific K or Na channel blockers in bath or in ionophoretic application. Application of K channel blockers
During bath application of TEA + 3,4-DAP an increase in the latency of the presynaptic signal usually occurred (see Brigant and Mallart, 1982) together with changes in its configuration. As shown in Fig. 3A, the late positive deflection of the triphasic wave was readily suppressed by bath application of TEA + 3,4-DAP. In this particular case the recording electrode was probably positioned at a point near the boundary between low and high K channel density areas. Thus, the electrode presumably picked up ionic outward K+ current together with the corresponding sink of local circuit current which caused a broadening of the inward Na spike. A similar situation is shown in Fig. 2B, (see also Mallart, 1984,*’ Fig. IH). Potassium channel blockers suppressed both the late outward current and the broadening of the Na spike. When present, the second positive component of distal responses similarly disappeared (Fig. 3B and C). In some cases the recorded signal was diphasic consisting in a positive capacity spike followed by a late negativity (Fig. 2C,, C,; Fig. 3D, E). The latter may be due either to inward Na+ current or to a sink of current corresponding to a source of outward current in adjacent areas of the membrane. As shown
D. ANGAUT-PETITand A. MALLART
1052
TEA
+DAP
A
8
C
D
E
F Fig. 3. Late wave form components are modified by bath application of K channel blockers. When positive, they are supressed (A, B) or replaced by a negative one (C). When negative, they are suppressed (D) or greatly reduced in ampiitude and duration (E). When no late component is present, no conspicuous change occurs in the presynaptic signal during bath application of the blocking agents (F). Calibration = I ms and 25 pV.
in Fig. 3D and E, bath application of K channel blockers suppressed more or less completely the late negative component, suggesting the presence of outward K + current in neighbouring areas of the membrane. However, a residual negative component may persist (Fig. 3E) or appear (Fig. 3C) during drug action, whose nature will be discussed later. In contrast to the conspicuous changes induced by K channel blockers just described, no obvious changes occurred when the distal response was composed of a singie positive wave, i.e. when late membrane current was missing (Fig. 3F) although in these cases bath application of K channel blockers was able to increase the release of transmitter by nerve impulses. This means that the absence of a late cornponent in presynaptic signals does not necessarily imply an absence of K channels. Rather, uniform
distribution of K channels over the main part of the endings, the preterminal nodes of Ranvier and the heminode will give no differences between falling phases of action potentials generated at these parts of the membrane and thus no late net current will flow along the nerve terminal. The view that a widespread distrjbut~on of K channels existed aiong regenerating terminals was tested by ionophoretic apphcation of specific K channet blocking drugs as close as possible to the recording electrode, as shown in the diagram in Fig. 4. During ionophoresis of the drug at the heminode, a late negative wave either appeared or, when present with a small amplitude, increased in amplitude and duration (Fig. 4A,, A?, arrows). Similarly TEA -I- 3,4-DAP ionophoresis onto the main part of an ending induced the appearance of a fate negative
Electrical activity of regenerating motor endings
1053
2
Fig. 4. Ionophoretic application of K channel blocking agents at the heminode (A) or at the distal part (B) of two regenerating endings (1 and 2). The upper trace in each pair is average of control recordings while the lower one is during drug ionophoresis. The application produces an enhancement in amplitude and duration of late negative deflections (A, and A,, arrows) or their appearance (B, and B,, arrows). The small and prolonged late negative deflection which appears in B, during drug ionophoresis corresponds probably to postsynaptic activity. Horizontal calibration = 1 ms. Vertical calibration is IOOpV in 1, 5OpV in 2.
in the recordings (Fig. 4B,, B,, arrows). Changes in late component configuration can be explained by supposing that the ionophoretic apphcation of K channel blockers suppressed an outward current at each location of the electrode. At the heminode (A,, A,), drug action led to an increase in the amplitude of the late net inward current, while at the terminal part (B,, B2) it unmasked a late net inward current. In cases of endings reaching maturity, where the signals displayed two distinct negative or positive peaks at the heminode and terminal portions respectively (Fig. 2A, and D,, Fig. 3B and C), drug ionophoresis was effective only at the latter.
component
Effect of tetrodotoxin
As described above, (TEA + 3,4-DAP)-resistant diphasic wave forms (see Fig. 3) could be recorded from most of the length of some endings and triphasic signals were found at considerable distance from the heminode, which would indicate active impulse propagation.i4 However, since diphasicity is not an absolute criterium to distinguish between active propagation and electrotonic spread,” we used TTX in ionophoretic application close to the recording electrode to obtain a more precise picture of Na channel distribution in regenerating endings. The large negative wave that characterized signals
from the heminode was immediately abolished by TTX application (Fig. 5A) indicating the presence at this point of an important Na+ influx. TTX also affected responses in more distal areas of some endings. The small negative component of diphasic signals present in these regions (Fig. 5, B,, upper trace) was suppressed within a few seconds of drug application (Fig. 5, B,, lower trace). In other endings, the effect of TTX ionophoresis in distal areas was to enhance the amplitude of outward current or to increase the decay time of the monophasic positive response (Fig. 5B,). Our interpretation of the discrete changes in the outward current elicited by TTX application is that localized suppression of Na current enhances outward capacity current at that point. Either slight or conspicuous the changes promoted by TTX indicate that Na channels are present, perhaps with a low density, up to the most distal part of some regenerating endings. However, in some endplates where wave forms displayed an “immature” look regarding the K component, TTX application was without effect, suggesting that Na channels have withdrawn from the main part of the endings. One can say that in these cases, factors affecting Na channel redistribution precede those controlling K channel availability.
D. ANGAUT-PETIT
and A.
MALLART
Fig. 5. Effect of ionophoretic TTX application on signals recorded from the heminode (A) and from the distal part of regenerating endings (B, and Br). The upper trace in each pair corresponds to control recordings while the lower one is during ‘ITX application. In A: TTX induces the disappearance of the sharp negative deflection. Similarly, in B, the small negative deflection disappears during TTX ionophoresis. A diphasic potential with a small amplitude negative wave persists after TTX ionophoresis in these two examples: it probably corresponds to current flowing between actively and electrotonically depolarized membrane areas (see Ref. 1I). In B,: TTX application induces an increase in the duration of the monophasic positive wave. Calibration: I ms and 100 {tV.
Ca current
The inward current was proved to be promoted by a calcium influx, firstly by its increase in amplitude when the usual calcium concentration of 2 mM was increased to 6mM, and secondly by its complete disappearance when IOmM cobalt were added to the perfusing medium (Fig. 6Br). Since Ca?+ current is recorded with a negative polarity only in the main part of the endings, it follows that Ca channels exist with higher density there, as compared to the heminode. endings.
During bath application of TEA + 3,4-DAP, a negative wave of long duration and low ampI~tude followed in some cases the positive spike recorded at the main part of the endings (Fig. 6A, and 3,). A positive wave which was its mirror image was recorded at the heminode (Fig. 6A,). The time course of these waves was similar to that of the calcium current shown by Mallart and Brigant’j in normal
BI
__ kb?aJL
* A2
.-we.
I
‘I
Fig. 6. A, and A, correspond respectively to recordings from the main part of an ending and its heminode during bath application of K channel blockers + 6 mM calcium. A late component is visible which is negative in A,, positive in A,. The wave form shown in B, was recorded from the main part of another ending in the same experimental conditions. The negative wave in B, disappeared in &?by supplementary addition of IO mM cobalt to the perfusing fluid. Calibration: I ms and 25 ,uV.
Electrical activity of regenerating motor endings DlSCUSSlON
The present study shows that the electrical activity in regenerating motor endings differs in two main features from that of mature terminals. One is the considerable variation in shape and amplitude of the K-dependent wave form component. The other is the length over which a Na-dependent wave form component can be recorded. The fact that the normal type presynaptic wave form coexists in the same muscle with signals displaying features observed only in regenerating endings suggests that the regenerating process is very rapid. A typical feature of presynaptic currents in regenerating nerve endings revealed by the present study is the presence of a K component whose polarity was either negative, indicating an absence or a lower density of K channels and consequently a longer action potential in this area as compared to the heminode and preterminal nodes, or positive, indicating the reverse situation. In some cases no visible K+ current was present, indicating the presence of K channels with a similar density, or accessibility, in the terminals and in the parent axon. Differences in K-dependent components between regenerating and mature endings can be explained by the process of myelin formation which, by insulating inter- and paranodal regions of preterminal axons, suppresses sources for outward K+ current.‘6,26 Our results do not allow us to tell whether this mechanism suffices to account for changes in K component configuration which occur in motor endings as they reach maturation. Nonetheless, to explain the change from a K component of negative polarity observed in some regenerating endings into a positive one which characterizes the normal state, one has to postulate that, in addition to axonal myelination, a real increase in K channel density occurs at the terminals during regeneration. In our experiments, Na channel activity was signalled by the presence along most of the length of regenerating endings of triphasic as well as diphasic waves whose large negative components were reduced or suppressed by ionophoretic TTX application. Nevertheless, the amplitude of inward current at the main part of the terminals was never as large as at the heminode, as shown by the inverse relationship that exists between outward capacitive and inward ionic Na+ current at these two recording sites. That triphasic waves do not originate in growing axons passing nearby is shown by the fact that these signals were only recorded from end-plate areas (as defined by the presence of end-plate nuclei) at electrode emplacements within IO-20pm of the myelin end, and that they changed usually to a diphasic or monophasic positive configuration within the endplate limits. TTX was effective in changing wave form configuration also in cases where no net inward current was visible, indicating the presence of Na channels, however they were insufficient to outweigh
1055
the ionic or passive outward currents occurring at the same place. Finally, cases where TTX ionophoresis failed to reveal the presence of Na channels were also observed. Why does the main part of some regenerating endings exhibit detectable Na+ conductance while others do not, the latter being in this respect comparable to mature endings? Recordings from normal endings failed to show any Na-dependent inward current, while outward K+ and inward Ca*+ currents could be readily demonstrated.6 A possible explanation for our failure to record net inward Na+ current is that Na channels exist in an exceedingly low density at the endings, as compared to the nodes of Ranvier or the heminode. As a matter of fact, binding studies revealed a Na channel density ranging from 25 to 110 sites km -* in unmyelinated fibres, while Na channel density was as high as 12 000 pm-* at the nodes of Ranvier (see Ref. 28). By assuming a similar Na channel density in unmyelinated axons and in nerve endings, one would expect an inward Na+ current at the terminals of negligible intensity as compared to that in the nodes. Under these circumstances, the current that enters the membrane at the nodes carried by Na ions, and leaves it at the endings by discharging membrane capacity would easily outweigh the inward Na+ current at the terminals.‘8 One can therefore conclude that the absence of net inward current which we observed at some regenerating terminals is related to an extremely low density of Na channels. On the contrary, the fact that in many instances we were able to record Na-dependent inward currents of large amplitude, together with the presence of triphasic wave forms over a considerable length of the terminals, which indicates active impulse conduction,‘4 strongly suggests that, unlike in normal endings, Na channels occur with a nonnegligible density in regenerating terminals. A small amplitude wave form component which corresponds to inward Ca*+ current was recorded from regenerating terminal branches in 3,4-DAP + TEA-treated preparations. The use of both K channel blocking agents was required because, in mouse motor endings, voltage-dependent K channels are more readily blocked by 3,4-DAP, while Ca*+-activated K channels are sensitive only to TEA.** The small amplitude of the Ca wave form was an unexpected finding since Ca*+ currents have been reported to be predominant during early stages of development. ‘2,29Furthermore, calcium spikes are present in growing dendrites,” cultured cells3 and regenerating axons.20~25~30 The small amplitude of Ca responses in our preparations can be explained by assuming that a non-negligible amount of inward Ca*+ current is present at the preterminal part of the axon which would cancel, in part, external current promoted by Ca*+ entry at the terminals. Alternatively, the small amplitude of Ca responses would reveal a genuine lower density of Ca channels in regenerating endings as compared to mature ones,
D. ANGAUT-PETIIand A.
1056
which is in agreement with the lower quanta1 content of postsynaptic responses elicited by the former. Our results do not allow us to decide between hypotheses.
these two
MALLAKT
Acknon,/edgmzents-We wish to thank L. Fmlle fur help with histological techniques. and D. Chaslard for typing rhc manuscript. This work was supported by a research ptxnt from INSERM no. 816003.
REFERENCES
I. Angaut-Petit 345, 57P.
D. and Mallart A. (1983) Presynaptic currents in regenerating mouse motor endings. J. Physiol., Lend.
2. Angaut-Petit D., Mallart A. and Faille L. (1982) Role of denervated sheaths and end-plates in muscle reinnervation by collateral sprouting in the mouse. Bioi. Cell. 46, 2777290. 3. Anglister L., Farber I. C., Shahar A. and Grinvald A. (1982) Localization of voltage-sensitive calcium channels along developing neurites: their possible role in regulating neurite elongation. Dec. Biol. 94, 351-365. 4. Bennett M. R. and Florin T. (1974) A statistical analysis of the release of acetylcholine at newly formed synapses m striated muscle. J. Physiol., Land. 238, 93-107. 5. Bostock H., Sears T. A. and Sherratt R. M. (1981) The effects of 4-aminopyridine and tetraethylammonium ions on normal and demyelinated mammalian nerve fibres. J. Physiol., Land. 313, 301-315. 6. Brigant J. L. and Mallart A. (1982) Presynaptic currents in mouse motor endings. /. Physiol., Lend. 333, 619-636. 7. Brismar T. (1980) Potential clamp analysis of membrane currents in rat myelinated nerve fibres. J. Physiol., Lond. 298, 171-184.
8. Chiu S. Y. and Ritchie J. M. (1980) Potassium channels in nodal and internodal axonal membrane of mammalian myelinated fibres. Nature 2B4, 170-171. 9. Chiu S. Y. and Ritchie J. M. (1984) On the physiological role of internodal potassium channels and the security of conduction in myelinated nerve fibres. Proc. R. Sot. Lond. B 220, 415-422. 10. Chiu S. Y., Ritchie J. M., Rogart R. B. and Stagg D. (1979) A quantitative description of membrane currents in rabbit myelinated nerve. J. Physiol., Lond. 292, 149-166. 11. Eccles J. C. (1964) The Physiology of Synapses, p. 124. Academic Press, New York. 12. Hagiwara S. and Byerly L. (1981) Calcium channel. Ann. Rev. Neurosci. 4, 69-125. 13. Hodgkin A. L. (1937) Evidence for electrical transmission in nerve. Part 1. J. Physiol., Land. 90, 183-210. 14. Katz B. and Miledi R. (1965) Propagation of electric activity in motor nerve terminals. Proc. R. Sot. Lond. B 161, 453-482. 15. Kocsis J. D., Ruiz J. A. and Waxman S. G. (1983) Maturation of mammalian myehnated fibers: changes in action-potential characteristics following 4-aminopyridine application. J. Neurophysiol. 50, 449-463. 16. Kocsis J. D. and Waxman S. G. (1983) Electrophysiology of conduction in mammalian regenerating nerves. in Nerre, Organ and Tissue Regeneration: Research Perpecfiues (ed. Seil F. J.), pp. 89-107. Academic Press, New York. 17. Kocsis J. D., Waxman S. G., Hildebrand C. and Ruiz J. A. (1982) Regenerating mammalian nerve fibres: changes in action potential wave form and firing characteristics following blockage of potassium conductance. Proc. R. Sor. Land. B 217, 77-87. 18. Konishi T. and Sears T. A. (1984) Electrical activity of mouse motor nerve terminals. Proc. R. Sot. Lond. t3 222, 115-120.
19. Llinas R. and Sugimori M., (1979) Calcium conductances in Purkinje cell dendrites: their role in development and integration. In Deoelopment and Chemical Specificity of Neurons. Progr. Brain Research (eds Cuinot M.. Kreutzberg G. W. and Bloom F. E.), 51, pp. 323-334. Elsevier, Amsterdam. 20. MacVicar B. A. and Llinas R. (1982) Calcium spikes in regenerating giant axons of the lamprey spinal cord. Sot. Neurosci. Abst 8, 914.
21. Mallart A. (1984) Presynaptic currents in frog motor endings. Pfltigers Arch. 400, 8-13. 22. Mallart A. (1984) Calcium-activated potassium current in presynaptic terminals. Biomed. Res. 5, 287-290. 23. Mallart A. and Brigant J. L. (1982) Electrical activity at motor nerve terminals of the mouse. J. Physiol., Paris 78, 407-41
I.
24. McArdle J. J., Angaut-Petit D., Mallart A., Bournaud R., Faille L. and Brigant J. L. (1981) Advantages of the Triangularis sterni muscle of the mouse for investigations of synaptic phenomena. J. Neurosci. Merh. 4, 109-I 15. 25. Meiri H., Spira M. E. and Parnas I. (1981) Membrane conductance and action potential of a regenerating axonal tip. Science 211, 709-712.
26. Ritchie J. M. (1982) Sodium and potassium channels in regenerating and developing mammalian myelinated nerves. Proc. R. Sot. Lond. B 215, 273-287.
27. Ritchie J. M. and Rogart R. B. (1977) Density of sodium channels in mammalian myelinated nerve fibers and nature of the axonal membrane under the myelin sheath. Proc. narn. Acad. Sci. USA 74, 21 I-215. 28. Rogart R. (1981) Sodium channels in nerve and muscle membrane. Ann. Rev. Physiol. 43, 71 l-725. 29. Spitzer N. C. (_l979) Ion channels in development. Ann. Reo. Neurosci. 2, 363-397. 30. Strichartz G., Small R., Nicholson C., Pfenninger K. H. and Llinls R. (1980) Ionic mechanisms for impulse propagation in growing nonmyelinated axons:saxitoxin binding and electrophysiology. Sot. Neurosci. Abst 6, 660. 31. Tonge D. A. (1974) Physiological characteristics of re-innervation of skeletal muscle in the mouse. J. Physiol., Land. 241, 141-153. 32. Waxman S. G. and Foster R. E. (1980) Ionic channel distribution and heterogeneity of the axon membrane in myelinated fibers. Brain Res. Rev. 2, 205-234. (Accepred 21 June 1985)