Neurotoxicologyand Teratology, Vol. 12, pp. 313-318. ©Pergamon Press plc, 1990. Printed in the U.S.A.
0892-0362/90 $3.00 + .00
The Action of Cobalt, Cadmium and Thallium on Presynaptic Currents in Mouse Motor Nerve Endings HERBERT WlEGAND, STEFAN UHLIG, URSULA GOTZSCH AND HORST LOHMANN ~
Medical Institute o f Environmental Hygiene at the Heinrich-Heine-University Diisseldorf Department o f Neurotoxicology, Auf'm Hennekamp 50, D-4000 Diisseldorf 1, Federal Republic of Germany R e c e i v e d 3 O c t o b e r 1989
WIEGAND, H., S. UHLIG, U. GOTZSCH AND H. LOHMANN. The action of cobalt, cadmium and thallium on presynaptic currents in mouse motor nerve endings. NEUROTOXICOL TERATOL 12(4) 313-318, 1990.--The action of cobalt, cadmium and thallium on presynaptic currents was investigated by recording extracellular potentials from mouse motor nerve terminals. Recorded waveforms consisted of two negative deflections preceded by a small positivity. The second negative deflection could be blocked by the potassium channel blockers tetraethylammonium (TEA) and 3,4-diaminopyridine (3,4-DAP). Application of either divalent cations (cobalt, cadmium) or monovalent thallous ions to the bath, even in mM concentrations, did not change these waveforms significantly. After application of high TEA and 3,4-DAP concentrations (10 mM and 250 p,M, respectively), a prolonged positive-going wave arose, which could be blocked reversibly by bath application of cobalt and cadmium, but not thallium. The concentration-inhibition curve for cadmium suggested two apparent dissociation constants, whereas cobalt seemed to have only one apparent dissociation constant. It was concluded that the long-lasting positive wave is driven by calcium influx, since it was competitively antagonized by the application of cobalt and cadmium. A different, short-lasting positive wave arose using lower concentrations of potassium channel blockers, and this wave could not be blocked by cobalt, cadmium, or thallium. Presynaptic currents
Cobalt
Cadmium
Thallium
Synaptic transmission
METHOD
THE release of neurotransmitter substances elicited by membrane depolarization is calcium dependent. The release of acetylcholine (ACh) in response to a nerve impulse is governed by the calcium concentration in the extracellular space [for review, see (18)], whereas the spontaneous release of ACh is presumed to depend on intracellular, presynaptic calcium concentration. Divalent metal cations (e.g., cadmium and cobalt) block impulse-evoked release of ACh from motor nerve endings by competitively antagonizing the depolarization-induced calcium influx (5). In a recent series of papers (20-22) we demonstrated that thallium appeared to impede synaptic transmission in a manner similar to divalent metal cations. In particular, we found evidence that thallium can antagonize presynaptic transmitter release (20). However, thallium seemed to disturb the utilization of calcium by mitochondria and/or endoplasmic reticulum inside the presynaptic terminals, thus acting on a stage of transmitter release different from that affected by the divalent metal cations. In this report, we describe studies which examine more directly the relationship between thallium and calcium currents by using perineural recording techniques and blocking potassium currents with tetraethylammonium (TEA) or 3,4-diaminopyridine (3,4-DAP) (3, 13, 14, 17). Some details of these studies were presented previously (23,24).
We performed the experiments on the M. triangularis sterni of adult mice (12) under visual control of a Zeiss microscope equipped with Nomarski interference contrast optics [(6); see Fig. la]. We perfused the muscle preparations with either Krebs solution as described previously by Penner and Dreyer (17), or a modified solution for investigations of cadmium's effect with the following composition (mM): NaC1 145; KC1 5; CaC12 2.5; MgSO4 l; HEPES 5; glucose 11; aerated with 02. The experiments were performed at room temperature (21 - 2°C). Upon nerve stimulation through a suction electrode applied to the N. intercostalis, we recorded signals inside the perineuronal sheath from nerve bundles containing usually 2 to 4 nerve fibers. Glass microelectrodes filled with 0.5 M NaC1 had resistances between 5 and 15 MOhm. The reference electrode was a chlorided silver wire in the recording chamber. Neuromuscular transmission was blocked by d-tubocurarine (50 IxM) added to the bath solution in order to suppress postsynaptic muscle responses. Investigations on presynaptic "calcium-potentials" were performed in the presence ot TEA (1-10 mM) and 3,4-DAP (50-250 IxM). To prevent repetitive firing of the nerves occurring in the presence of these
1Present address: Faculty of Biology, Gener. Zoology and Neurobiology, Geb. ND7, POB 102148, D-4630 Bochum, 1, FRG.
313
314
WIEGAND, UHLIG, GOTZSCH AND LOHMANN
FIG. 1. (a) Neuromuscular junctions as viewed by Nomarski optics. An endplate with a nonmyelinated nerve terminal (arrow) is clear to identify. The perineuronal recording site is indicated by a microelectrode symbol. Magnification: × 400, Normarski interference contrast. (b) External current recorded perineuronally upon supramaximal stimulation from preterminal nerve bundles as indicated for a typical example in (a). First positive (upward) peak: stimulation artifact; second positive peak (single arrow): sodium-influx peak at the preterminal heminode; second negative peak (double arrow): potassium-influx peak at the terminal membrane [see Brigant and Mallart (3)].
potassium-channel blockers, procaine (100 IxM) was added to the bath solution. This technique allowed us to record potential changes between recording electrode and reference electrode through the open end of the nerve sheath where the synaptic terminals emerge. Under these special experimental conditions the potential changes reflect the sum of the longitudinal currents inside the perineuronal space flowing along the extracellular space of the axons, the block of
potassium channels unmasking a short- and a long-lasting, positive voltage deflection, which Penner and Dreyer (17) proved to be carried by influx of calcium via presynaptic voltage-gated ion channels. Because of the dependence of the calcium signal amplitude on stimulation frequency (17) we stimulated the intercostal nerves usually every 2 to 3 minutes with pulses of 50-1xsec duration and supramaximal voltage and thus obtained signals of sufficient signal to noise ratio without averaging.
PRESYNAPTIC CURRENTS AS INFLUENCED BY METAL IONS
315
In order to test the mode of action of either divalent cobalt and cadmium ions or monovalent thallium ions in the blockade of these calcium signals, we recorded from the perineurium of presynaptic axon bundles during superfusion either with Ringer's solution for a control period of usually half an hour, after which we treated the preparation with metal ion-containingRinger's solutions, until the perineuronal signals were reduced to a constant plateau. Subsequently untreated Ringer's solution washed out the metal ions from the preparations. Each muscle was treated only once, using a single concentration of metal. Signal analysis was performed using an LSI 11/03 computer with a modified software as described previously (22).
RESULTS
~kJ
2ms
Perineuronal Signals Without Potassium Channel Blockade After application of d-tubocurarine to the muscle preparation, the penetration of the perineural sheath was indicated by the recording of a negative potential drop in the range of 2 to 6 mV as described by Penner and Dreyer (17). Applying supramaximal stimulation to the intercostal nerve we recorded a nerve signal as shown in Fig. lb. It comprised a double-peaked negative (downward) component, often preceded by a small positivity (upward; see double arrow and single arrow in Fig. lb, respectively). The latter is due to a passive current flux evoked by sodium entrance within the next proximal node of Ranvier, whereas the first and the second negativity peaks correspond to the sodium influx at the heminode producing an outward current at the nerve endings and a potassium efflux in the nerve terminals, respectively (2, 3, 17). Neither the first nor the second negativity of the presynaptic nerve signal in this stage of the experiment could be changed even by superfusion with concentrations of heavy metal ions (cobalt, cadmium, thallium) up to 1 mM.
Perineuronal Signals During Potassium Channel Blockade Blockade by TEA (10 mM) and 3,4-DAP (250 0.M) of the potassium channels of the terminals diminished the second negative peak and resulted in development of a long-lasting positivity, which seemed to consist of two components as shown in Fig. 2. How these signals were generated is difficult to interpret. Nevertheless Penner and Dreyer (17) identified these potentials by application of channel-blocking agents to the nerve terminals as related to potassium and calcium currents, respectively: As considerable longitudinal potential gradients are generated within the endoneuronal space because of the insulating properties of the perineuronal sheath, the potentials recorded at some distance from the neuromuscular junctions are driven by passive currents resulting from ion-flux changes within the presynaptic terminals. This membrane region of the presynaptic, mammalian terminals is devoid of sodium channels, but calcium channels are abundant (4). Penner and Dreyer (17) recorded perineuronal potentials from the nerve bundles and by iontophoretically applied calcium channel blockers evidenced that the positive potentials were carried by calcium ion-influx into the presynaptic nerve terminals. Therefore, the term "calcium plateau" [see (14)] for the long-lasting component seems to be justified. The waveforms recorded critically depended on the electrode position within the perineuronal space. Therefore we continuously monitored the waves obtained from the same recording site before and throughout application of metal ions. In control experiments we succeeded in recording the same waveforms without a variation of more than about 5% during a three-hour recording period. As a criterion of recording stability, we monitored the amplitudes of the first, sodium-carried negativity, and required less than about 5% variability during the recording period. This was justified because
FIG. 2. Externalcurrents from preterminalendoneuriumupon supramaximal stimulationduring superfusionwith potassium-channelblockers TEA (10 mM) and 3,4-DAP (250 ~M). The second negative peak begins to vanish shortly after applicationof TEA; in combinationwith 3,4-DAP a prolonged positive voltage deflection arises, carried by terminal calcium influx, i.e., calciumplateau signal.
heavy metal ions did not change the sodium-carried wave, even with the highest metal concentrations used.
Effects of Metal Ions on Perineuronal Calcium Signals Application of cobalt to presynaptic terminals. In order to investigate the influence of the calcium channel blocker cobalt on the positive perineuronal signals, we superfused triangularis preparations, pretreated with low TEA and 3,4-DAP concentrations, d-tubocurarine and procaine as described in the Method section, with cobalt-containing Ringer's solutions. The result of a typical experiment is shown in Fig. 3. The top inset shows the waveform which we obtained using this moderate potassium channel blockade by application of 1 mM TEA and 50 IxM 3,4-DAP during a control superfusion. It consists of a brief positive signal (about 6-msec duration), followed by a prolonged positive signal (about 50-msec duration) at a stimulation rate of 1 per 2 or 3 minutes. The middle inset shows the signal 10 minutes after superfusion of 50 p,M cobalt-containingRinger's solution. The long-lasting positive waveform is fully blocked, whereas the short-lasting positivity is uninfluenced by cobalt application. The bottom inset shows recovery of the long-lasting positivity 12 minutes after resuperfusion with Ringer's solution without cobalt. Different cobalt concentrations were applied to different muscle preparations; the effect of cobalt on the positive-going signals was recorded to evaluate the area between zero-line and positivegoing waveform. From these values we calculated the percent inhibition by cobalt of the long-lasting positive wave and plotted the concentration inhibition-curve as shown in Fig. 4. The inhibition by cobalt of the long-lasting calcium influx into the presynaptic terminals was maximal in the range of 100 ~M, minimal inhibition being in the submicromolar range. Fifty percent inhibition of the long-lasting positive signal occurred at about 20 ixmolar. Application of cadmium to presynaptic terminals. If we conducted the same procedure using different concentrations of cadmium-containingRinger's solution for superfusion of triangularis muscles, an effect similar to that produced by cobalt was observed. Figure 5 shows a typical example with a long-lasting positivity produced by potassium channel blockade with higher TEA/3,4-DAP concentrations (10 mM TEA; 250 I~M 3,4-DAP) in the top inset as control. Analogous to the findings of Penner and
316
WIEGAND, UHLIG, GOTZSCH AND LOHMANN
O.5mV[ FIG. 3. Inhibition by cobalt (50 ~xM) of the calcium plateau of perineuronally recorded external currents. Top: control; middle: 10 rain after cobalt superfusion; bottom: 12 rain after wash. The recordings are from one and the same preparation.
Dreyer (17), these higher concentrations of potassium channel blockers TEA and 3,4-DAP resulted in a positive-going waveform, in which short-lasting and long-lasting signals could not be differentiated. Application of cadmium (10 IxM) fully blocked the long-lasting positive waveform leaving the short-lasting positive signal more or less unaffected. Resuperfusion with modified control-Ringer's solution resulted in an at least partially reversible long-lasting positivity. The areas between zero-line and positive waveforms were measured and percent inhibition produced by different concentrations of bath-applied cadmium was calculated.
100-
50. 10/.
3-tog
FIG. 4. Percent inhibition of calcium plateau signals 20 min after start of cobalt superfusion as normalized to controls is plotted versus molar concentrations of cobalt in the superfusion medium. Each point is mean "4-SEMof at least n=4 experiments. The curve was fitted by eye.
Figure 6 shows the resulting concentration-inhibition curve for cadmium. This inhibition curve for the slow positive signal shows evidence for the existence of two apparent dissociation constants for cadmium; the lower one being in the nanomolar range, the higher one being in the micromolar range. Application of thallium to presynaptic terminals. When TEA/ 3,4-DAP-treated muscle preparations were superfused with thallium-containing Ringer's solutions, we observed neither an effect on the sodium-influx-carried negative signal of presynaptic terminals nor an effect on the short-lasting or the long-lasting positive waveforms obtained after potassium channel blockade. This is demonstrated in Fig. 7. The top inset shows the control signal obtained after potassium blockade by application of the high blockade version (10 mM TEA; 250 txM 3,4-DAP). In our hands this kind of potassium blockade seems to be the maximum effect, because elevation of either TEA or 3,4-DAP concentration in the bath solution did not affect the signal form substantially. The middle inset shows the signal 21 minutes after application of a Ringer's solution containing thallium in a concentration of I mM. Whereas the inhibition of the long-lasting positive potentials was fully developed by this time in the experiments using either cobalt or cadmium, even this relatively high concentration of thallium affected neither the negative-going sodium signal nor the positivegoing long-lasting calcium waveform. If the duration of exposure to this high thallium concentration was extended to 1 hour or more (bottom record), the amplitudes of both the positive- and negativegoing waves declined, fading to zero during the next 2 or 3 hours of superfusion. However, we believe that this was the sign of nonspecific toxicity due to the extreme thallium concentration used since the effects were not reversible by resuperfusion with control Ringer's once the decline of signal had started. DISCUSSION
The aim of this paper was to analyze presynaptic events as influenced by heavy metals using the recording of typical nerve signals from inside the perineuronal space close to the nerve terminals. After application of d-tubocurarine to the neuromuscular preparations in order to block neuromuscular transmission the typical double negativity of the passing nerve action potential was recorded, using intraperineural microelectrodes (Fig. la). Preliminary experiments revealed that the amplitudes of the first negativity, which is carried by sodium influx in the next proximal node of Ranvier (17), did not change during superfusions with heavy metals. During these control experiments we further recognized that the amplitude of this sodium peak would serve as a measure of quality of recording conditions. Because of the different number of motor axons within the perineural sheath of different recording sites, amplitudes differed from one preparation to the other. The higher the number of axons, the larger the amplitude of the sodium peak was. The greater the distance of the recording site from the contributing neuromuscular junctions (this could be exactly measured by the use of Nomarski interference contrast optics), the lower the amplitudes of the recorded signals were. In analogy to the lack of influence of metal ions on the first negativity carried by sodium influx we never saw an influence on the second negativity, the so-called fast potassium current as described by Penner and Dreyer (17). This fast potassium negativity could be blocked by application of TEA and 3,4-DAP to the bath solution, the blockade developing within several minutes as shown in Fig. 2. No effects of the metal ions under investigation were seen on nodal or presynaptic membrane currents recorded in the absence of any channel blocker. In accordance with the report of Dreyer and Penner (7), high
PRESYNAPTIC CURRENTS AS INFLUENCED BY METAL IONS
C-----------0.5mVI lOms
FIG. 5. Inhibition by cadmium (10 I~M) of the calcium plateau of perineuronally recorded external currents. Top: control; middle: 20 min after cadmium superfusion; bottom: 20 min after wash. The recordings are from one and the same preparation. concentrations of 3,4-DAP consistently replaced the potassium negativity with a short positive-going wave due to the prolonged depolarization of the nerve terminals caused by blockade of the fast repolarizing potassium efflux. If in addition TEA was applied in high concentrations, a more or less pronounced slow positivity appeared which was evidenced pharmacologically to be due to a distinct set of slow calcium channels (17). Mallart (14), in contrast, suggested that the fast and the slow positive signal components are the mirror of one and the same calcium current becoming regenerative. In accordance with the findings of Penner and Dreyer (17) our results concerning the application of divalent heavy metal ions
~100.£ ..£ 5o-
10Concentration (M) FIG. 6. Percent inhibition of the calcium plateau signals 20 min after start of cadmium superfusion as normalized to controls is plotted versus molar concentration of cadmium in the superfusion medium. Each point is mean -+SEM of at least 4 experiments. The curve was fitted by eye.
317
05mVI lOms FIG. 7. Failure of inhibition by thallium (1 mM) of the calcium plateau of perineuronally recorded external currents. Top: control; middle: 20 min after thallium superfusion; bottom: 60 min after thallium superfusion. The recordings are from one and the same preparation.
clearly distinguished two different presynaptic calcium currents that had different thresholds of activation in mouse motor nerve terminals. The method by which we recorded the presynaptic currents does not unambiguously characterize which type of current is actually influenced. All currents activated during the action potential contribute to the waveforms. With this precaution in mind, and with the findings on the pharmacological dissection of the fast and slow positivity (17), these positive-going signals represented an appropriate tool for studying the influence of metal ions on presynaptic calcium currents as directly as possible in mammalian terminals. Aside from Penner and Dreyer (17), the authors are not aware of recordings from mammalian nerve terminals which were sufficiently stable to allow long-term pharmacological analysis of dose-response relationships. The effects of cobalt and cadmium on both the fast and the long-lasting positive components of the signals during potassium channel blockade indicate that these were due to calcium channel conductances. Among the divalent metal cations, cobalt and cadmium are usually thought to be typical inorganic calcium channel blockers (5, 10, 19). Both heavy metals affected the long-lasting calcium signal component in the ranges of concentrations applied here. Penner and Dreyer (17) blocked the fast calcium signal by application of far higher cadmium concentrations (as high as 15 mM). However, we observed a fading depression of the nerve spikes after application of metal concentrations higher than 1 mM and therefore did not systematically investigate the effects of such high metal concentrations. The action of cobalt on the calcium signals showed a fundamental difference between the two calcium potentials. As is clearly shown in Fig. 3, cobalt had no effect whatsoever on the fast calcium component, whereas the slow calcium component was blocked completely. The same holds, more or less, for cadmium. It is interesting that according to Penner and Dreyer (17), verapamil as well as diltiazem showed similar effects with regard to the fast and slow calcium component.
318
WIEGAND, UHLIG, GOTZSCH AND LOHMANN
In our investigation the concentration-effect curves of cadmium and cobalt clearly indicated an effect on the long-lasting calcium signal in the range of txmolar concentrations. In addition, the shape of the cadmium concentration-effect relationship points to a second binding site in the range of nano- or subnanomolar concentrations. These findings are in accordance with those of Penner and Dreyer (17). The slow calcium potential tends to decline with increasing stimulation frequency. However, since we used stimulation frequencies as low as 0.5-1 per minute, we believe that recovery time was ample. These long-lasting recovery periods should have prevented occurrence of an increase in intracellular presynaptic calcium concentration resulting from repetitive stimulation (8). The failure of thallium to block presynaptic calcium currents needs a further comment. In a recent series of papers (20-22) we tried to characterize the action of thallium on spontaneous and stimulated transmitter release. From these investigations the action of thallium on presynaptic nerve terminal ion channels remained inconclusive. The present investigation supports the view that neither presynaptic potassium channels nor presynaptic calcium channels could be influenced by thallium in acute superfusion experiments. Thus, the thallium-induced enhancement of spontaneous release of acetylcholine and the thallium-induced reduction of phasic transmitter release at the neuromuscular junction must result from other mechanisms.
As was stated by Penner and Dreyer (17), the calcium plateau signal of mouse motor nerve terminals recorded using the methods described cannot be unequivocally attributed to any presynaptic physiological event. If one used the model of Miller (15), there might be two kinds of calcium channels within a nerve terminal, a dihydropyridine sensitive (or L-type) channel and a dihydropyridine insensitive (or N-type) channel. Both L as well as N channels were strongly blocked by cadmium and omega-conotoxin [for review, see (15)]. The calcium plateau signal of presynaptic motor terminals fit neither of these schemes. Though it was strongly blocked by cadmium [(17), this paper], it was insensitive to 1,4-dihydropyridine blockade [(17), Wiegand and Meis, unpublished results] as well as for omega-conotoxin (1). The precise pharmacological characterization of the different kinds of calcium channels is still ambiguous [see (9, 11, 15, 16)]. Further experiments using calcium channel agonists like Bay K 8644 to try to enhance the status of the heavy-metal-blocked channel might further elucidate how the organic versus inorganic calcium channel blockers act on presynaptic nerve terminals. ACKNOWLEDGEMENTS Prof. F. Dreyer, Giessen, FRG, introduced us to the triangularis preparation and Nomarski optic use. Karolina Sveinsson prepared the figures skillfully. Our thanks are due to them.
REFERENCES 1. Anderson, A. J.; Harvey, A. L. Omega-conotoxin does not block the verapamil-sensitive calcium channels at mouse motor nerve terminals. Neurosci. Lett. 82:177-180; 1987. 2. Anderson, A. J.; Harvey, A. L. Effects of the facilitatory compounds catechol, guanidine, noradrenaline and phencyclidine on presynaptic currents of mouse motor nerve terminals. Naunyn Schmiedebergs Arch. Pharmacol. 338:133-137; 1988. 3. Brigant, J. L.; Mallart, A. Presynaptic currents in mouse nerve endings. J. Physiol. 333:619-636; 1982. 4. Chiu, Y.; Ritchie, J. L.; Rogard, R. B.; Stagg, D. A quantitative description of membrane currents in rabbit myelinated nerve. J. Physiol. 292:149-166; 1979. 5. Cooper, G. P.; Manalis, R, S. Influence of heavy metals on synaptic transmission: A review. Neurotoxicology 4(4):69-84; 1983. 6. Dreyer, F.; MUller, K. D.; Peper, K.; Sterz, R. The M. omohyoideus of the mouse as a convenient mammalian muscle preparation. A study of junctional and extrajunctional acetylcholine receptors by noise analysis and cooperativity. Pflugers Arch. 367:115-122; 1979. 7. Dreyer, F.; Penner, R. The action of presynaptic snake toxins on membrane currents of mouse motor nerve terminals. J. Physiol. 386:455-463; 1987. 8. Eckert, R.; Chad, J. E. Inactivation of Ca channels. Prog. Biophys. Mol. Biol. 44:215-267; 1984. 9. Hirning, L. D.; Fox, A. P.; McCleskey, E. W.; Olivera, B. D.; Thayer, S. A.; Miller, R. J.; Tsien, R. W. Dominant role of N-type Ca ÷ + channels in evoked release of norepinephrine from sympathetic neurons. Science 239:57-61; 1988. 10. Kita, H.; van der Kloot, W. Action of Co and Ni at the frog neuromuscular junction. Nature New Biol. 245:52-53; 1973. 11. Kostyuk, P. G. Diversity of calcium ion channels in cellular membrane. Neuroscience 28:253-261; 1989. 12. McArdle, J. J.; Angaut-Petit, D.; Mallart, A.; Bournaud, R.; Faille, L.; Brigant, J. L. Advantages of the triangularis sterni muscle of the
13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23.
24.
mouse for investigations of synaptic phenomena. J. Neurosci. Methods 4:109-115; 1981. Mallart, A. Electric current flow inside perineurial sheaths of mouse motor nerves. J. Physiol. 368:565-575; 1985. Mallart, A. A calcium-activated potassium current in motor nerve terminals of the mouse. J. Physiol. 368:577-591; 1985. Miller, R. J. Multiple calcium channels and neuronal function. Science 235:46-52; 1987. Miller, R. J. Calcium channels in neurons. In: Venter, J. C.; Triggle, D., eds. Structure and physiology of the slow inward calcium channel. New York: Alan R. Liss; 1987:161-246. Penner, R.; Dreyer, F. Two different presynaptic calcium currents in mouse motor nerve terminals. Pflugers Arch. 406:190-197; 1986. Silinsky, E. M. The biophysical pharmacology of calcium-dependent acetylcholine secretion. Pharmacol. Rev. 37(1):81-132; 1985. Weakly, J. N. The action of cobalt ions on neuromuscular transmission in the frog. J. Physiol. 234:597-612; 1973. Wiegand, H.; Csicsaky, M.; Kr~imer, U. The action of thallium acetate on neuromuscular transmission in the rat phrenic nervediaphragm preparation. Arch. Toxicol. 55:55-58; 1984. Wiegand, H.; Papadopoulos, R.; Csicsaky, M.; Kr~imer, U. The action of thallium acetate on spontaneous transmitter release in the rat neuromuscular junction. Arch. Toxicol. 55:253-257; 1984. Wiegand, H.; Lohmann, H.; Chandra, S. V. The action of thallium acetate on phasic transmitter release in the mouse neuromuscular junction. Arch. Toxicol. 58:265-270; 1986. Wiegand, H.; Lohmann, H.; Uhlig, S. The action of thallium, cadmium and cobalt on presynaptic calcium currents in mouse motor nerve terminals. Naunyn Schmiedebergs Arch. Pharmacol. 334(Suppl.): R16; 1986. Wiegand, H. Neurotoxicology of heavy metals: Synaptic transmission as influenced by mono- and divalent metal cations. In: Seemayer, N, H.; Hadnagy, W., eds. Environmental hygiene. Berlin: Springer Verlag; 1988:98-102.
0892-0362/90 $3.00 + .00
The Action of Cobalt, Cadmium and Thallium on Presynaptic Currents in Mouse Motor Nerve Endings HERBERT WlEGAND, STEFAN UHLIG, URSULA GOTZSCH AND HORST LOHMANN ~
Medical Institute o f Environmental Hygiene at the Heinrich-Heine-University Diisseldorf Department o f Neurotoxicology, Auf'm Hennekamp 50, D-4000 Diisseldorf 1, Federal Republic of Germany R e c e i v e d 3 O c t o b e r 1989
WIEGAND, H., S. UHLIG, U. GOTZSCH AND H. LOHMANN. The action of cobalt, cadmium and thallium on presynaptic currents in mouse motor nerve endings. NEUROTOXICOL TERATOL 12(4) 313-318, 1990.--The action of cobalt, cadmium and thallium on presynaptic currents was investigated by recording extracellular potentials from mouse motor nerve terminals. Recorded waveforms consisted of two negative deflections preceded by a small positivity. The second negative deflection could be blocked by the potassium channel blockers tetraethylammonium (TEA) and 3,4-diaminopyridine (3,4-DAP). Application of either divalent cations (cobalt, cadmium) or monovalent thallous ions to the bath, even in mM concentrations, did not change these waveforms significantly. After application of high TEA and 3,4-DAP concentrations (10 mM and 250 p,M, respectively), a prolonged positive-going wave arose, which could be blocked reversibly by bath application of cobalt and cadmium, but not thallium. The concentration-inhibition curve for cadmium suggested two apparent dissociation constants, whereas cobalt seemed to have only one apparent dissociation constant. It was concluded that the long-lasting positive wave is driven by calcium influx, since it was competitively antagonized by the application of cobalt and cadmium. A different, short-lasting positive wave arose using lower concentrations of potassium channel blockers, and this wave could not be blocked by cobalt, cadmium, or thallium. Presynaptic currents
Cobalt
Cadmium
Thallium
Synaptic transmission
METHOD
THE release of neurotransmitter substances elicited by membrane depolarization is calcium dependent. The release of acetylcholine (ACh) in response to a nerve impulse is governed by the calcium concentration in the extracellular space [for review, see (18)], whereas the spontaneous release of ACh is presumed to depend on intracellular, presynaptic calcium concentration. Divalent metal cations (e.g., cadmium and cobalt) block impulse-evoked release of ACh from motor nerve endings by competitively antagonizing the depolarization-induced calcium influx (5). In a recent series of papers (20-22) we demonstrated that thallium appeared to impede synaptic transmission in a manner similar to divalent metal cations. In particular, we found evidence that thallium can antagonize presynaptic transmitter release (20). However, thallium seemed to disturb the utilization of calcium by mitochondria and/or endoplasmic reticulum inside the presynaptic terminals, thus acting on a stage of transmitter release different from that affected by the divalent metal cations. In this report, we describe studies which examine more directly the relationship between thallium and calcium currents by using perineural recording techniques and blocking potassium currents with tetraethylammonium (TEA) or 3,4-diaminopyridine (3,4-DAP) (3, 13, 14, 17). Some details of these studies were presented previously (23,24).
We performed the experiments on the M. triangularis sterni of adult mice (12) under visual control of a Zeiss microscope equipped with Nomarski interference contrast optics [(6); see Fig. la]. We perfused the muscle preparations with either Krebs solution as described previously by Penner and Dreyer (17), or a modified solution for investigations of cadmium's effect with the following composition (mM): NaC1 145; KC1 5; CaC12 2.5; MgSO4 l; HEPES 5; glucose 11; aerated with 02. The experiments were performed at room temperature (21 - 2°C). Upon nerve stimulation through a suction electrode applied to the N. intercostalis, we recorded signals inside the perineuronal sheath from nerve bundles containing usually 2 to 4 nerve fibers. Glass microelectrodes filled with 0.5 M NaC1 had resistances between 5 and 15 MOhm. The reference electrode was a chlorided silver wire in the recording chamber. Neuromuscular transmission was blocked by d-tubocurarine (50 IxM) added to the bath solution in order to suppress postsynaptic muscle responses. Investigations on presynaptic "calcium-potentials" were performed in the presence ot TEA (1-10 mM) and 3,4-DAP (50-250 IxM). To prevent repetitive firing of the nerves occurring in the presence of these
1Present address: Faculty of Biology, Gener. Zoology and Neurobiology, Geb. ND7, POB 102148, D-4630 Bochum, 1, FRG.
313
314
WIEGAND, UHLIG, GOTZSCH AND LOHMANN
FIG. 1. (a) Neuromuscular junctions as viewed by Nomarski optics. An endplate with a nonmyelinated nerve terminal (arrow) is clear to identify. The perineuronal recording site is indicated by a microelectrode symbol. Magnification: × 400, Normarski interference contrast. (b) External current recorded perineuronally upon supramaximal stimulation from preterminal nerve bundles as indicated for a typical example in (a). First positive (upward) peak: stimulation artifact; second positive peak (single arrow): sodium-influx peak at the preterminal heminode; second negative peak (double arrow): potassium-influx peak at the terminal membrane [see Brigant and Mallart (3)].
potassium-channel blockers, procaine (100 IxM) was added to the bath solution. This technique allowed us to record potential changes between recording electrode and reference electrode through the open end of the nerve sheath where the synaptic terminals emerge. Under these special experimental conditions the potential changes reflect the sum of the longitudinal currents inside the perineuronal space flowing along the extracellular space of the axons, the block of
potassium channels unmasking a short- and a long-lasting, positive voltage deflection, which Penner and Dreyer (17) proved to be carried by influx of calcium via presynaptic voltage-gated ion channels. Because of the dependence of the calcium signal amplitude on stimulation frequency (17) we stimulated the intercostal nerves usually every 2 to 3 minutes with pulses of 50-1xsec duration and supramaximal voltage and thus obtained signals of sufficient signal to noise ratio without averaging.
PRESYNAPTIC CURRENTS AS INFLUENCED BY METAL IONS
315
In order to test the mode of action of either divalent cobalt and cadmium ions or monovalent thallium ions in the blockade of these calcium signals, we recorded from the perineurium of presynaptic axon bundles during superfusion either with Ringer's solution for a control period of usually half an hour, after which we treated the preparation with metal ion-containingRinger's solutions, until the perineuronal signals were reduced to a constant plateau. Subsequently untreated Ringer's solution washed out the metal ions from the preparations. Each muscle was treated only once, using a single concentration of metal. Signal analysis was performed using an LSI 11/03 computer with a modified software as described previously (22).
RESULTS
~kJ
2ms
Perineuronal Signals Without Potassium Channel Blockade After application of d-tubocurarine to the muscle preparation, the penetration of the perineural sheath was indicated by the recording of a negative potential drop in the range of 2 to 6 mV as described by Penner and Dreyer (17). Applying supramaximal stimulation to the intercostal nerve we recorded a nerve signal as shown in Fig. lb. It comprised a double-peaked negative (downward) component, often preceded by a small positivity (upward; see double arrow and single arrow in Fig. lb, respectively). The latter is due to a passive current flux evoked by sodium entrance within the next proximal node of Ranvier, whereas the first and the second negativity peaks correspond to the sodium influx at the heminode producing an outward current at the nerve endings and a potassium efflux in the nerve terminals, respectively (2, 3, 17). Neither the first nor the second negativity of the presynaptic nerve signal in this stage of the experiment could be changed even by superfusion with concentrations of heavy metal ions (cobalt, cadmium, thallium) up to 1 mM.
Perineuronal Signals During Potassium Channel Blockade Blockade by TEA (10 mM) and 3,4-DAP (250 0.M) of the potassium channels of the terminals diminished the second negative peak and resulted in development of a long-lasting positivity, which seemed to consist of two components as shown in Fig. 2. How these signals were generated is difficult to interpret. Nevertheless Penner and Dreyer (17) identified these potentials by application of channel-blocking agents to the nerve terminals as related to potassium and calcium currents, respectively: As considerable longitudinal potential gradients are generated within the endoneuronal space because of the insulating properties of the perineuronal sheath, the potentials recorded at some distance from the neuromuscular junctions are driven by passive currents resulting from ion-flux changes within the presynaptic terminals. This membrane region of the presynaptic, mammalian terminals is devoid of sodium channels, but calcium channels are abundant (4). Penner and Dreyer (17) recorded perineuronal potentials from the nerve bundles and by iontophoretically applied calcium channel blockers evidenced that the positive potentials were carried by calcium ion-influx into the presynaptic nerve terminals. Therefore, the term "calcium plateau" [see (14)] for the long-lasting component seems to be justified. The waveforms recorded critically depended on the electrode position within the perineuronal space. Therefore we continuously monitored the waves obtained from the same recording site before and throughout application of metal ions. In control experiments we succeeded in recording the same waveforms without a variation of more than about 5% during a three-hour recording period. As a criterion of recording stability, we monitored the amplitudes of the first, sodium-carried negativity, and required less than about 5% variability during the recording period. This was justified because
FIG. 2. Externalcurrents from preterminalendoneuriumupon supramaximal stimulationduring superfusionwith potassium-channelblockers TEA (10 mM) and 3,4-DAP (250 ~M). The second negative peak begins to vanish shortly after applicationof TEA; in combinationwith 3,4-DAP a prolonged positive voltage deflection arises, carried by terminal calcium influx, i.e., calciumplateau signal.
heavy metal ions did not change the sodium-carried wave, even with the highest metal concentrations used.
Effects of Metal Ions on Perineuronal Calcium Signals Application of cobalt to presynaptic terminals. In order to investigate the influence of the calcium channel blocker cobalt on the positive perineuronal signals, we superfused triangularis preparations, pretreated with low TEA and 3,4-DAP concentrations, d-tubocurarine and procaine as described in the Method section, with cobalt-containing Ringer's solutions. The result of a typical experiment is shown in Fig. 3. The top inset shows the waveform which we obtained using this moderate potassium channel blockade by application of 1 mM TEA and 50 IxM 3,4-DAP during a control superfusion. It consists of a brief positive signal (about 6-msec duration), followed by a prolonged positive signal (about 50-msec duration) at a stimulation rate of 1 per 2 or 3 minutes. The middle inset shows the signal 10 minutes after superfusion of 50 p,M cobalt-containingRinger's solution. The long-lasting positive waveform is fully blocked, whereas the short-lasting positivity is uninfluenced by cobalt application. The bottom inset shows recovery of the long-lasting positivity 12 minutes after resuperfusion with Ringer's solution without cobalt. Different cobalt concentrations were applied to different muscle preparations; the effect of cobalt on the positive-going signals was recorded to evaluate the area between zero-line and positivegoing waveform. From these values we calculated the percent inhibition by cobalt of the long-lasting positive wave and plotted the concentration inhibition-curve as shown in Fig. 4. The inhibition by cobalt of the long-lasting calcium influx into the presynaptic terminals was maximal in the range of 100 ~M, minimal inhibition being in the submicromolar range. Fifty percent inhibition of the long-lasting positive signal occurred at about 20 ixmolar. Application of cadmium to presynaptic terminals. If we conducted the same procedure using different concentrations of cadmium-containingRinger's solution for superfusion of triangularis muscles, an effect similar to that produced by cobalt was observed. Figure 5 shows a typical example with a long-lasting positivity produced by potassium channel blockade with higher TEA/3,4-DAP concentrations (10 mM TEA; 250 I~M 3,4-DAP) in the top inset as control. Analogous to the findings of Penner and
316
WIEGAND, UHLIG, GOTZSCH AND LOHMANN
O.5mV[ FIG. 3. Inhibition by cobalt (50 ~xM) of the calcium plateau of perineuronally recorded external currents. Top: control; middle: 10 rain after cobalt superfusion; bottom: 12 rain after wash. The recordings are from one and the same preparation.
Dreyer (17), these higher concentrations of potassium channel blockers TEA and 3,4-DAP resulted in a positive-going waveform, in which short-lasting and long-lasting signals could not be differentiated. Application of cadmium (10 IxM) fully blocked the long-lasting positive waveform leaving the short-lasting positive signal more or less unaffected. Resuperfusion with modified control-Ringer's solution resulted in an at least partially reversible long-lasting positivity. The areas between zero-line and positive waveforms were measured and percent inhibition produced by different concentrations of bath-applied cadmium was calculated.
100-
50. 10/.
3-tog
FIG. 4. Percent inhibition of calcium plateau signals 20 min after start of cobalt superfusion as normalized to controls is plotted versus molar concentrations of cobalt in the superfusion medium. Each point is mean "4-SEMof at least n=4 experiments. The curve was fitted by eye.
Figure 6 shows the resulting concentration-inhibition curve for cadmium. This inhibition curve for the slow positive signal shows evidence for the existence of two apparent dissociation constants for cadmium; the lower one being in the nanomolar range, the higher one being in the micromolar range. Application of thallium to presynaptic terminals. When TEA/ 3,4-DAP-treated muscle preparations were superfused with thallium-containing Ringer's solutions, we observed neither an effect on the sodium-influx-carried negative signal of presynaptic terminals nor an effect on the short-lasting or the long-lasting positive waveforms obtained after potassium channel blockade. This is demonstrated in Fig. 7. The top inset shows the control signal obtained after potassium blockade by application of the high blockade version (10 mM TEA; 250 txM 3,4-DAP). In our hands this kind of potassium blockade seems to be the maximum effect, because elevation of either TEA or 3,4-DAP concentration in the bath solution did not affect the signal form substantially. The middle inset shows the signal 21 minutes after application of a Ringer's solution containing thallium in a concentration of I mM. Whereas the inhibition of the long-lasting positive potentials was fully developed by this time in the experiments using either cobalt or cadmium, even this relatively high concentration of thallium affected neither the negative-going sodium signal nor the positivegoing long-lasting calcium waveform. If the duration of exposure to this high thallium concentration was extended to 1 hour or more (bottom record), the amplitudes of both the positive- and negativegoing waves declined, fading to zero during the next 2 or 3 hours of superfusion. However, we believe that this was the sign of nonspecific toxicity due to the extreme thallium concentration used since the effects were not reversible by resuperfusion with control Ringer's once the decline of signal had started. DISCUSSION
The aim of this paper was to analyze presynaptic events as influenced by heavy metals using the recording of typical nerve signals from inside the perineuronal space close to the nerve terminals. After application of d-tubocurarine to the neuromuscular preparations in order to block neuromuscular transmission the typical double negativity of the passing nerve action potential was recorded, using intraperineural microelectrodes (Fig. la). Preliminary experiments revealed that the amplitudes of the first negativity, which is carried by sodium influx in the next proximal node of Ranvier (17), did not change during superfusions with heavy metals. During these control experiments we further recognized that the amplitude of this sodium peak would serve as a measure of quality of recording conditions. Because of the different number of motor axons within the perineural sheath of different recording sites, amplitudes differed from one preparation to the other. The higher the number of axons, the larger the amplitude of the sodium peak was. The greater the distance of the recording site from the contributing neuromuscular junctions (this could be exactly measured by the use of Nomarski interference contrast optics), the lower the amplitudes of the recorded signals were. In analogy to the lack of influence of metal ions on the first negativity carried by sodium influx we never saw an influence on the second negativity, the so-called fast potassium current as described by Penner and Dreyer (17). This fast potassium negativity could be blocked by application of TEA and 3,4-DAP to the bath solution, the blockade developing within several minutes as shown in Fig. 2. No effects of the metal ions under investigation were seen on nodal or presynaptic membrane currents recorded in the absence of any channel blocker. In accordance with the report of Dreyer and Penner (7), high
PRESYNAPTIC CURRENTS AS INFLUENCED BY METAL IONS
C-----------0.5mVI lOms
FIG. 5. Inhibition by cadmium (10 I~M) of the calcium plateau of perineuronally recorded external currents. Top: control; middle: 20 min after cadmium superfusion; bottom: 20 min after wash. The recordings are from one and the same preparation. concentrations of 3,4-DAP consistently replaced the potassium negativity with a short positive-going wave due to the prolonged depolarization of the nerve terminals caused by blockade of the fast repolarizing potassium efflux. If in addition TEA was applied in high concentrations, a more or less pronounced slow positivity appeared which was evidenced pharmacologically to be due to a distinct set of slow calcium channels (17). Mallart (14), in contrast, suggested that the fast and the slow positive signal components are the mirror of one and the same calcium current becoming regenerative. In accordance with the findings of Penner and Dreyer (17) our results concerning the application of divalent heavy metal ions
~100.£ ..£ 5o-
10Concentration (M) FIG. 6. Percent inhibition of the calcium plateau signals 20 min after start of cadmium superfusion as normalized to controls is plotted versus molar concentration of cadmium in the superfusion medium. Each point is mean -+SEM of at least 4 experiments. The curve was fitted by eye.
317
05mVI lOms FIG. 7. Failure of inhibition by thallium (1 mM) of the calcium plateau of perineuronally recorded external currents. Top: control; middle: 20 min after thallium superfusion; bottom: 60 min after thallium superfusion. The recordings are from one and the same preparation.
clearly distinguished two different presynaptic calcium currents that had different thresholds of activation in mouse motor nerve terminals. The method by which we recorded the presynaptic currents does not unambiguously characterize which type of current is actually influenced. All currents activated during the action potential contribute to the waveforms. With this precaution in mind, and with the findings on the pharmacological dissection of the fast and slow positivity (17), these positive-going signals represented an appropriate tool for studying the influence of metal ions on presynaptic calcium currents as directly as possible in mammalian terminals. Aside from Penner and Dreyer (17), the authors are not aware of recordings from mammalian nerve terminals which were sufficiently stable to allow long-term pharmacological analysis of dose-response relationships. The effects of cobalt and cadmium on both the fast and the long-lasting positive components of the signals during potassium channel blockade indicate that these were due to calcium channel conductances. Among the divalent metal cations, cobalt and cadmium are usually thought to be typical inorganic calcium channel blockers (5, 10, 19). Both heavy metals affected the long-lasting calcium signal component in the ranges of concentrations applied here. Penner and Dreyer (17) blocked the fast calcium signal by application of far higher cadmium concentrations (as high as 15 mM). However, we observed a fading depression of the nerve spikes after application of metal concentrations higher than 1 mM and therefore did not systematically investigate the effects of such high metal concentrations. The action of cobalt on the calcium signals showed a fundamental difference between the two calcium potentials. As is clearly shown in Fig. 3, cobalt had no effect whatsoever on the fast calcium component, whereas the slow calcium component was blocked completely. The same holds, more or less, for cadmium. It is interesting that according to Penner and Dreyer (17), verapamil as well as diltiazem showed similar effects with regard to the fast and slow calcium component.
318
WIEGAND, UHLIG, GOTZSCH AND LOHMANN
In our investigation the concentration-effect curves of cadmium and cobalt clearly indicated an effect on the long-lasting calcium signal in the range of txmolar concentrations. In addition, the shape of the cadmium concentration-effect relationship points to a second binding site in the range of nano- or subnanomolar concentrations. These findings are in accordance with those of Penner and Dreyer (17). The slow calcium potential tends to decline with increasing stimulation frequency. However, since we used stimulation frequencies as low as 0.5-1 per minute, we believe that recovery time was ample. These long-lasting recovery periods should have prevented occurrence of an increase in intracellular presynaptic calcium concentration resulting from repetitive stimulation (8). The failure of thallium to block presynaptic calcium currents needs a further comment. In a recent series of papers (20-22) we tried to characterize the action of thallium on spontaneous and stimulated transmitter release. From these investigations the action of thallium on presynaptic nerve terminal ion channels remained inconclusive. The present investigation supports the view that neither presynaptic potassium channels nor presynaptic calcium channels could be influenced by thallium in acute superfusion experiments. Thus, the thallium-induced enhancement of spontaneous release of acetylcholine and the thallium-induced reduction of phasic transmitter release at the neuromuscular junction must result from other mechanisms.
As was stated by Penner and Dreyer (17), the calcium plateau signal of mouse motor nerve terminals recorded using the methods described cannot be unequivocally attributed to any presynaptic physiological event. If one used the model of Miller (15), there might be two kinds of calcium channels within a nerve terminal, a dihydropyridine sensitive (or L-type) channel and a dihydropyridine insensitive (or N-type) channel. Both L as well as N channels were strongly blocked by cadmium and omega-conotoxin [for review, see (15)]. The calcium plateau signal of presynaptic motor terminals fit neither of these schemes. Though it was strongly blocked by cadmium [(17), this paper], it was insensitive to 1,4-dihydropyridine blockade [(17), Wiegand and Meis, unpublished results] as well as for omega-conotoxin (1). The precise pharmacological characterization of the different kinds of calcium channels is still ambiguous [see (9, 11, 15, 16)]. Further experiments using calcium channel agonists like Bay K 8644 to try to enhance the status of the heavy-metal-blocked channel might further elucidate how the organic versus inorganic calcium channel blockers act on presynaptic nerve terminals. ACKNOWLEDGEMENTS Prof. F. Dreyer, Giessen, FRG, introduced us to the triangularis preparation and Nomarski optic use. Karolina Sveinsson prepared the figures skillfully. Our thanks are due to them.
REFERENCES 1. Anderson, A. J.; Harvey, A. L. Omega-conotoxin does not block the verapamil-sensitive calcium channels at mouse motor nerve terminals. Neurosci. Lett. 82:177-180; 1987. 2. Anderson, A. J.; Harvey, A. L. Effects of the facilitatory compounds catechol, guanidine, noradrenaline and phencyclidine on presynaptic currents of mouse motor nerve terminals. Naunyn Schmiedebergs Arch. Pharmacol. 338:133-137; 1988. 3. Brigant, J. L.; Mallart, A. Presynaptic currents in mouse nerve endings. J. Physiol. 333:619-636; 1982. 4. Chiu, Y.; Ritchie, J. L.; Rogard, R. B.; Stagg, D. A quantitative description of membrane currents in rabbit myelinated nerve. J. Physiol. 292:149-166; 1979. 5. Cooper, G. P.; Manalis, R, S. Influence of heavy metals on synaptic transmission: A review. Neurotoxicology 4(4):69-84; 1983. 6. Dreyer, F.; MUller, K. D.; Peper, K.; Sterz, R. The M. omohyoideus of the mouse as a convenient mammalian muscle preparation. A study of junctional and extrajunctional acetylcholine receptors by noise analysis and cooperativity. Pflugers Arch. 367:115-122; 1979. 7. Dreyer, F.; Penner, R. The action of presynaptic snake toxins on membrane currents of mouse motor nerve terminals. J. Physiol. 386:455-463; 1987. 8. Eckert, R.; Chad, J. E. Inactivation of Ca channels. Prog. Biophys. Mol. Biol. 44:215-267; 1984. 9. Hirning, L. D.; Fox, A. P.; McCleskey, E. W.; Olivera, B. D.; Thayer, S. A.; Miller, R. J.; Tsien, R. W. Dominant role of N-type Ca ÷ + channels in evoked release of norepinephrine from sympathetic neurons. Science 239:57-61; 1988. 10. Kita, H.; van der Kloot, W. Action of Co and Ni at the frog neuromuscular junction. Nature New Biol. 245:52-53; 1973. 11. Kostyuk, P. G. Diversity of calcium ion channels in cellular membrane. Neuroscience 28:253-261; 1989. 12. McArdle, J. J.; Angaut-Petit, D.; Mallart, A.; Bournaud, R.; Faille, L.; Brigant, J. L. Advantages of the triangularis sterni muscle of the
13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23.
24.
mouse for investigations of synaptic phenomena. J. Neurosci. Methods 4:109-115; 1981. Mallart, A. Electric current flow inside perineurial sheaths of mouse motor nerves. J. Physiol. 368:565-575; 1985. Mallart, A. A calcium-activated potassium current in motor nerve terminals of the mouse. J. Physiol. 368:577-591; 1985. Miller, R. J. Multiple calcium channels and neuronal function. Science 235:46-52; 1987. Miller, R. J. Calcium channels in neurons. In: Venter, J. C.; Triggle, D., eds. Structure and physiology of the slow inward calcium channel. New York: Alan R. Liss; 1987:161-246. Penner, R.; Dreyer, F. Two different presynaptic calcium currents in mouse motor nerve terminals. Pflugers Arch. 406:190-197; 1986. Silinsky, E. M. The biophysical pharmacology of calcium-dependent acetylcholine secretion. Pharmacol. Rev. 37(1):81-132; 1985. Weakly, J. N. The action of cobalt ions on neuromuscular transmission in the frog. J. Physiol. 234:597-612; 1973. Wiegand, H.; Csicsaky, M.; Kr~imer, U. The action of thallium acetate on neuromuscular transmission in the rat phrenic nervediaphragm preparation. Arch. Toxicol. 55:55-58; 1984. Wiegand, H.; Papadopoulos, R.; Csicsaky, M.; Kr~imer, U. The action of thallium acetate on spontaneous transmitter release in the rat neuromuscular junction. Arch. Toxicol. 55:253-257; 1984. Wiegand, H.; Lohmann, H.; Chandra, S. V. The action of thallium acetate on phasic transmitter release in the mouse neuromuscular junction. Arch. Toxicol. 58:265-270; 1986. Wiegand, H.; Lohmann, H.; Uhlig, S. The action of thallium, cadmium and cobalt on presynaptic calcium currents in mouse motor nerve terminals. Naunyn Schmiedebergs Arch. Pharmacol. 334(Suppl.): R16; 1986. Wiegand, H. Neurotoxicology of heavy metals: Synaptic transmission as influenced by mono- and divalent metal cations. In: Seemayer, N, H.; Hadnagy, W., eds. Environmental hygiene. Berlin: Springer Verlag; 1988:98-102.
