J. Physiol. (I958) IV4, 29I-304

Transcription

1 291 J. Physiol. (I958) IV4, 29I-304 SOME EFFECTS PRODUCED BY ADRENALINE UPON NEUROMUSCULAR PROPAGATION IN RATS BY K. KRNJEVIC AND R. MILEDI* From the Physiology Department, Australian National University, Canberra (Received 25 November 1957) Some fifty years have passed since the first reports of a potentiating action of suprarenal extracts on the contractions of fatigued muscle (Dessy & Grandis, 1904; Panella, 1907). This subject has inspired a voluminous literature, which has been conspicuous for the variety of phenomena described and the diversity of explanatory hypotheses. A historical survey is out of the question, but reference may be made to four authors who give useful reviews covering the greater part of the field: Gruber (1919); Wastl (1928); Burn (1945) and Goffart (1952). A study of this literature gives a somewhat confusing impression, mainly because of the paucity of well established facts. It is, for instance, still not agreed whether adrenaline has an action of any importance on neuromuscular propagation (cf. Brown, Biilbring & Burns, 1948); or whether adrenaline really alters the membrane potential of muscle fibres (cf. Brown, Goffart & Vianna Dias, 1950; Hutter & Loewenstein, 1955). With this picture in mind it seemed to us that a detailed investigation of the effects of adrenaline at various critical points in neuromuscular propagation should provide a concrete answer to some fairly simple questions. We have suggested in a previous publication (Krnjevid & Miledi, 1958b) that two main types of failure of neuromuscular propagation are associated with fatigue. Post-synaptic failure, which commonly tends to occur first, is characterized by the failure of initiation of conducted impulses in the muscle fibres. It is apparently caused by two principle factors: a rise in the threshold of the fibre; and a reduced sensitivity of the end-plate to ACh. Presynaptic failure of conduction, on the other hand, prevents impulses from reaching the nerve ending. We have now studied the effects of adrenaline on these kinds of failure in rat muscle both in vitro and in situ. We have also looked for alterations in the resting potential of muscle fibres. * Fellow of the Rockefeller Foundation. 19-2

2 292 K. KRNJEVI6 AND R. MILEDI Like so many earlier authors, we have found the actions of adrenaline to some extent unpredictable. Nevertheless, certain trends can be clearly discerned. There is no doubt that there is a powerful defatiguing effect on neuromuscular propagation. A preliminary account has been given of some of the observations described here (Krnjevic & Miledi, 1957). METHODS All muscles were from albino rats of Wistar origin, weighing g. The recording and stimulating techniques were similar to those previously described (Krnjevi6 & Miledi, 1958b). Experiments in vitro In twenty-five experiments a phrenic-diaphragm preparation was mounted in a Perspex chamber with a capacity of about 20 ml. The bathing Ringer-Locke solution had the composition given by Liley (1956a); it was stirred vigorously with a mixture of 5% C02 in 02, and was kept either at room temperature (about 220 C) or at 370 C. The proximal portion of the phrenic nerve rested on platinum stimulating electrodes in a separate compartment filled with liquid paraffin. The electrical activity of single muscle fibres was recorded with conventional intracellular glass micro-electrodes filled with 3M-KCI. For direct electrical stimulation we used a non-polarizable electrode (Ag-AgCl-Agar-Ringer) with a diameter of 0-7 mm. The muscles were usually fully paralysed with D-tubocurarine chloride (Burroughs Wellcome and Co.). Artificial end-plate potentials were obtained by applying ACh (acetylcholine chloride, Roche Products) electrophoretically with micropipettes. The square pulses of current had a duration of 1-10 msec, and a magnitude of about 10-7 A. The concentration of ACh in the glass micropipettes was 2-4 M. Spontaneous diffusion was controlled by an adjustable inward current. Injection of adrenaline. In the earlier experiments a continuous flow of Ringer-Locke solution was maintained through the muscle chamber. Adrenaline was injected into the system at a convenient point upstream from the chamber. In most experiments the bathing solution was only changed whenever it was necessary to wash out expended adrenaline, etc. The drug was then injected directly into the chamber, either from a burette, or from a tuberculin syringe via a fine polythene catheter. This method of direct injection is preferable because there is little lag if stirring is adequate, and the final concentration is easily calculated. Experiments in situ Five rats were kept under pentobarbitone anaesthesia (Nembutal, Abbott Laboratories) with doses of 5 mg/100 g body weight injected intraperitoneally at intervals of about 1 hr. The left gracilis anticus was exposed and its nerve cut and mounted on platinum stimulating electrodes in a pool of liquid paraffin. Injection of adrenaline. A fine polythene cannula (PE 10, Clav-Adams, New York) was inserted into the right iliac artery so that its tip lay close to the bifurcation of the aorta. The tubing was filled with Ringer-Locke solution containing heparin. It was connected to a stopcock which permitted the injection of either an adrenaline or a flushing solution. The volume of adrenaline solution injected was of the order of 0-1 ml., in a concentration of 50,ug/ml. We soon found that there is a strict limit to the amount of adrenaline which can be injected into the blood stream without causing pulmonary oedema: according to Maire & Patton (1956), the threshold lethal dose in unanaesthetized rats is 12-5 ug/100 g body weight. We tested the effectiveness of this method of injection with small volumes of a solution of Evans Blue (1/1000). The dye was easily discernible with a microscope when it reached the vessels in the muscle. In preparations with a good circulation the dye appeared in the muscle with a lag of only 1-3 sec.

3 ADRENALINE AND NEUROMUSCULAR PROPAGATION 293 Adrenaline solution8 As a rule, we used 1/1000 adrenaline hydrochloride (D.H.A. Laboratories, Sydney). In control experiments, the solutions were prepared from the salt adrenaline bitartrate (British Drug Houses). In a few experiments, we also tried the effect of L-noradrenaline bitartrate (Lights and Co. London). RESULTS Presynaptic failure of neuromuscular propagation Prolonged repetitive maximal stimulation of a motor nerve is usually associated, sooner or later, with a failure of conduction somewhere near the nerve terminals, probably at a point of branching. If one records the electrical activity of a single muscle fibre in the region of the end-plate, intermittent failures are observed, which may become more frequent with time, especially at high rates of stimulation (Figs. 1, 3). In some cases, complete and apparently irreversible block may result; more often, the changes are easily reversed by lowering the frequency, or stopping the stimulation altogether E ~60-1 ~~~ Time (min) Fig. 1. Complete reversal by adrenaline HCI (60,ug) of intermittent presynaptic failure as recorded intracellularly from the end-plate of a muscle fibre stimulated indirectly at the rate of 10/sec. Abscissa: time after injection of adrenaline, the beginning and end of which are indicated by the arrows; ordinates indicate the corresponding percentage rate of transmission, 10/sec being taken as 100%. Each point is the mean rate of transmission over a period of 24 sec; at 10 min the rate of transmission was still 100%. Diaphragm at 250 C. Inset: examples of the potentials recorded (A) before and (B) 5 min after adrenaline. Experiments in vitro. We have found that adrenaline has a potent action tending to restore conduction in fibres which are failing intermittently or, sometimes, even completely. With experiments in vitro this was plainly evident in eighteen out of thirty trials. In six cases there was a rather small effect, the significance of which could not be established. In five trials there was clearly

4 294 K. KRNJEVIC AND R. MILEDI no change, and in one case there seemed to be an actual depression of propagation. The amount of change produced by adrenaline was highly variable. There might be only a very small increase in the number of transmitted impulses, or full conduction might be restored. When the latter happened, it was often possible to raise appreciably the frequency of stimulation without causing further failures (Fig. 2; cf. also Fig. 2, Krnjevi6 & Miledi, 1957). At room temperature there was relatively little change within the first minute of injection and the effect did not reach its peak until 2-3 min had elapsed. Moreover, it often persisted for some min even after replacing the solution in the bath. At 370 C recovery of transmission was first observed 100 C 0_ 75 E I Time (min) Fig. 2. Restoration of presynaptic conduction by adrenaline HCI (100 jig) in a curarized diaphragm at room temperature. The arrows indicate the time taken to inject adrenaline. The frequency of stimulation was 25/sec and the resting potential remained at 74 mv throughout. The system could follow without failure a frequency of 30/sec for a short period of time, approximately 3 min after the application of adrenaline. At 25/sec there was no failure even after 10 min. after sec, but the effect seemed to be more evanescent. Fig. 1 illustrates an experiment in vitro in which 100% conduction was restored by 60,ug adrenaline HCl. The fibre had been stimulated at a frequency of 10/sec, and was failing intermittently. The electrical activity recorded with an intracellular electrode consisted initially of intermittent spikes. Five minutes after the injection of adrenaline the presynaptic failures no longer occurred. It should be noted, however, that post-synaptic failure was now evident, many impulses failing to evoke anything more than a non-propagated e.p.p. Presynaptic failure of conduction occurs even when the muscular contractions are prevented by curare. Fig. 2 shows that under these conditions adrenaline is no less effective in improving conduction. The contractions can also be much reduced by dissecting and stimulating only a single fibre of the phrenic nerve (Krnjevi6 & Miledi, 1958a). Repetitive stimulation readily

5 ADRENALINE AND NEUROMUSCULAR PROPAGATION 295 elicits presynaptic block in such a preparation, and here again adrenaline produces a marked improvement in conduction. Similar results were obtained with solutions of adrenaline bitartrate. Noradrenaline was only administered in one experiment. In two trials better conduction resulted; in two others (one fibre) there was apparently depression. Experiments in situ. Clear-cut results are less easily obtained with a gracilis in situ. Presynaptic failures cannot usually be recognized with certainty unless one records the activity of end-plates with intracellular electrodes. Adrenaline always has a powerful vasoconstrictor action, which causes the muscle to shrink suddenly, in most cases displacing the micro-electrode out of the muscle fibre, and often into another one. This may give a false impression that the nerve fibre is firing at '6 a higher or lower rate. We have only considered experiments in which it was quite clear that the electrode had remained in the same fibre throughout Time (sec) Fig. 3. Effect of adrenaline on presynaptic failure and on e.p.p. amplitude in a curarized gracilis ins8itu. About 10 minafterstartingstimnulation at 1/sec complete presynaptic failure occurred. At zero time adrenaline HCI (2 ug) was injected. Responses first appeared 10 sec later. Each bar represents one e.p.p. Successful experiments were made with three rats. A potentiating action similar to that already described was seen in seven trials (including one doubtful response). There was no effect in one case, and an apparent depression in another. Fig. 3 shows the return of propagation to a fibre of a curarized gracilis, which had ceased to give e.p.p.'s before adrenaline was injected; e.p.p.'s reappeared 10 sec after the intra-arterial injections, and did not vanish again for nearly 3 min. In all experiments in situ the effect had a latency of only sec. Post-synaptic failure of neuromuscular propagation Sensitivity of the post-synaptic membrane to ACh Amplitude of ACh-potentials. We have been unable to find any effect of adrenaline upon potentials evoked directly at end-plates in the diaphragm with microjets of ACh at rates of 0-2-5/sec. Various kinds of potentials were

6 296 K. KRNJEVIO AND R. MILEDI produced: some with short (1 msec) pulses; others with longer (10 msec) pulses of ACh. In several cases the adrenaline was added to the bath only after repetitive stimulation had caused the ACh potential to diminish greatly (Fig. 8). Adrenaline and noradrenaline consistently failed to have any significant effect. In two cases the potentials increased somewhat but the control potentials had not been sufficiently regular to make these observations significant, especially in view of the total lack of corroboration in other experiments. I iiia Fig. 4. Increase in threshold produced by adrenaline. A muscle fibre was stimulated directly with just-threshold condenser discharges (0.1 msec time constant) at a frequency of 1/sec; at the moment indicated by the dot, adrenaline HCI (40,ug) was applied. This record was obtained with a pair of extracellular electrodes from a diaphragm at room temperature. Amplitude of spontaneous miniature end-plate potentials (min.e.p.p.'s). A change in the sensitivity of the end-plate to the chemical transmitter ought to be reflected in a corresponding change in the mean amplitude of min.e.p.p.'s. In a few experiments in vitro we have looked for such an effect after the administration of adrenaline (three times) and noradrenaline (once). Each mean was calculated from the measurements of 100 min.e.p.p.'s after suitable magnification. In a typical experiment, at room temperature, the mean amplitudes were as follows: control, over a period of about 1 min, 1-11 mv; over a period of about 1 min, beginning 1-5 min after the addition of 100,ug noradrenaline to the bath, 1N10 mv. Changes in the electrical threshold of musclefibres (in vitro) The first procedure consisted in stimulating a small group of muscle fibres directly with non-polarizable electrodes, at a rate of about once per second. The diaphragm was curarized in some cases; in all cases the stimulating electrodes were placed in a region distant from end-plates. The activity of single fibres was recorded with extracellular electrodes to avoid damage. The intensity of stimulation was adjusted to a level near the threshold, so that some stimuli failed to excite the fibre. Under these conditions, extremely small changes in threshold are readily observed. A few trials failed to produce any effect. In most cases, however, adrenaline (in doses varying from 10 to 60,ug) had an unmistakable depressing action on the excitability of the fibre (Fig. 4). The change usually began within sec, but in some cases there was a delay of some minutes. The period of higher

7 ADRENALINE AND NEUROMUSCULAR PROPAGATION 297 threshold always lasted about 3 min; it was not followed by an obvious overswing. In the second type of experiment muscle fibres were stimulated directly as before but the rate of excitation was such as to produce a gradual increase in threshold (Krnjevic & Miledi, 1958b). Adrenaline was then added. This was repeated at several frequencies, and control runs were done at each frequency without adding adrenaline. We found no appreciable difference between the test and the control curves. It must be realized, of course, that this procedure (unlike the first) would not reveal clearly small alterations in excitability (e.g. < 5 %). We also observed the time course of the recovery of excitability in all cases after ending stimulation. It was unchanged after adrenaline. Other effects of adrenaline Effect on e.p.p.'s in muscle fibres paralysed with curare, excess magnesium or fatigue. In the curarized diaphragm we have seen a significant increase in the amplitude of e.p.p.'s (recorded intracellularly) after the administration of adrenaline in a majority of trials (eight out of fifteen). By 'significant' is meant either a pronounced change, which is immediately obvious (e.g. Figs. 5, 6) or one which can be demonstrated statistically in terms of the 0 05 prob- E _ v ~~~~~~~> E2 80' C E_& o20 X j o ~~~~~~~~~~E ft t I -20 4," w Time (min) Fig. 5. Enhancement of e.p.p. amplitude (0) produced by adrenaline HCl (50,g) injected at the time marked by the arrows in a curarized diaphragm at 240 C. Each point corresponds to the mean of 10 responses; the frequency of stimulation was 5/sec. 0, membrane resting potential. ability limit, using Student's t-test. In four trials the results also suggested enhancement, but this was not statistically significant. These cases, and the three in which there was apparently no change at all, were associated with unstable resting potentials; these usually lead to substantial variations in e.p.p.'s, which would tend to mask any other changes of small magnitude. Commonly the increase ranged from 20 to 60 %; the greatest increase was

8 298 K. KRNJEVIO1 AND R. MILEDI 100 %. The effect reached its maximum in about 1 min, and to some degree it usually persisted for at least 5 min. We have also found a similar temporary increase of e.p.p.'s in muscles curarized in situ (e.g. Fig. 3). Adrenaline also increases the amplitude of e.p.p.'s in a diaphragm paralysed by a relatively high concentration of magnesium (15 mm). The effect tends to be obscured by the large variations in amplitude which characterize such e.p.p.'s: nevertheless, a significant change was found in three out of five trials. In the other two trials the change was either too small to be significant at the 0 05 level or negligible. E.p.p.'s in a muscle fibre fatigued by repetitive stimulation are apparently less sensitive to adrenaline. We have only seen an enhancement of amplitude in one case out of four in which we actually looked for such a change (see Fig. 2, in Krnjevi6 & Miledi, 1957). Fig. 6. Intracellular e.p.p.'s from a curarized diaphragm at 370 C obtained at a stimulation frequency of 1/sec. A, control (resting potential 72 mv); B, 2 min after the application of 50 pg adrenaline HCl (resting potential 72 mv); C, 7 min later (resting potential 70 mv). Effect on the resting membrane potential of muscle fibres. In all experiments which required intracellular recording, both in vitro and in situ, we noted the resting potential at frequent intervals (e.g. Fig. 5). The potential shown by the millivoltmeter could be read to within 1 mv. In addition, we made a series of five experiments with one diaphragm at room temperature in vitro, keeping a continuous record of the resting potential by means of an ink-writer, to make certain that we could not miss a transient change. With this method of recording, the limit of resolution was about 0-5 mv. We have failed to observe any consistent change of resting potential after the addition of adrenaline. Resting potentials seldom remain steady for a very long time. Usually some drift is evident, more or less random in direction when fibres are not injured. There was no reason to believe that adrenaline added anything to the spontaneous changes already taking place. In a few cases the resting potential varied by less than 1 mv during a period of 6-8 min, near the beginning of which adrenaline had been applied. Effect on thefrequency of spontaneous miniature end-plate potentials. Adrenaline regularly accelerated the discharge of min.e.p.p.'s in vitro, both at room

9 ADRENALINE AND NEUROMUSCULAR PROPAGATION 299 temperature and at 370 C (eight trials in five diaphragms). The maximum change resulted in a frequency not much greater than twice the initial average value. Fig. 7 shows the effect on the spontaneous discharges recorded simultaneously at two different junctions. The time courses of the changes in the two fibres were remarkably similar, even though the spontaneous discharges had differed substantially, and varied independently of each other, during the preliminary control period. E 80 C 7l - e, a. bo C E 60L 60 E 40 _-, C0 28 U_ Time (min) Fig. 7. Acceleration by adrenaline HCI (50 jg) of the spontaneous discharges of min.e.p.p.'s observed by simultaneous intracellular recording from two separate end-plates, * and 0. A and A illustrate the respective resting potentials. Diaphragm at 200C. In one of the two trials with noradrenaline there was a significant increase in frequency. A greater frequency of discharge followed the administration of adrenaline in situ almost as regularly as in vitro. There were only two results of doubtful significance out of nine trials in four rats. DISCUSSION Reversal of presynaptiefailure The most striking action of adrenaline tending to improve propagation during fatigue is the presynaptic recovery of conduction in motor nerve fibres (cf. Corkill & Tiegs, 1933). We can do no more than speculate on the mechanism involved. It has been claimed that the electrical threshold of nerve fibres may

10 300 K. KRNJEVIO AND R. MILEDI be diminished under the action of adrenaline (Biilbring & Whitteridge, 1941; Legouix & Minz, 1953). This is supported to some extent by the observation that adrenaline may lower the mechanical threshold of tactile receptors and Pacinian corpuscles (Loewenstein, 1956; Loewenstein & Altamirano-Orrego, 1956), but was not confirmed by Naess & Sirnes (1953), who studied rabbit nerves. Should the change in threshold really be caused by a reduction in membrane potential, as suggested by Legouix & Minz (1953), a ready explanation might also be available for the observed increase in the frequency of spontaneous min.e.p.p.'s. According to Liley (1956b) a simple, logarithmic relationship exists between the membrane potential of the nerve terminals and the spontaneous rate of discharge. From Liley's data one can deduce that depolarization of the nerve terminals by about 5 mv would be sufficient to double the frequency of the min.e.p.p.'s. We have previously pointed out (Krnjevic & Miledi, 1958b) that intermittent failure of propagation may well have a protective action, by preventing excessive bombardment of the muscle fibre. This suggestion receives some support from the paradoxical observation that the adrenaline-induced recovery of presynaptic function may ultimately result in post-synaptic failure (Fig. 1). It is possible, of course, that the post-synaptic failure is aggravated by a small decrease in the excitability of the muscle fibre also caused by the adrenaline. Post-synaptic changes Potentiation of ACh. We have been unable to confirm earlier reports that adrenaline enhances the action of ACh. Neither the potentials generated by microjets of ACh nor spontaneous min.e.p.p.'s showed any significant increase in amplitude in the presence of adrenaline. Several authors had found a depression of the action of ACh on muscle in situ (Frank, Nothmann & Guttmann, 1923; Bender, 1938). Luco (1939), however, demonstrated the importance of vascular changes in situ (cf. Wastl, 1928), which are apparently responsible for the observed depression. When this is suppressed by ergotoxine, the potentiation is revealed. This is in agreement with the observations of Dale & Gaddum (1930), Burn (1945), Torda & Wolff (1946), Ellis & Beckett (1955) and Hutter & Loewenstein (1955), which were based mostly on experiments with muscles in vitro. It is essential to realize that, with the exception of Hutter & Loewenstein, all these authors measured the contractions or contractures produced by ACh. Their results, therefore, do not permit one to distinguish between a potentiation of ACh and a direct enhancement of the contractile mechanism of the muscle (cf. Goffart & Ritchie, 1952). Hutter & Loewenstein (1955), on the other hand, demonstrated what seems to be a reversal of the desensitization which follows repeated applications of

11 ADRENALINE AND NEUROMUSCULAR PROPAGATION 301 ACh. This would be significant during fatigue, one feature of which is reduced effectiveness of ACh (Krnjevic & Miledi, 1958b). However, in spite of several attempts to produce such an effect, even with relatively large pulses of ACh and high frequencies of release, we have not been able to reverse with adrenaline the progressive diminution of ACh potentials (Fig. 8). Muscle fibre threshold. The only effect that we could find was a small but distinct increase in threshold during stimulation at a low frequency. There was no evident change produced by adrenaline while stimulating repetitively; the latter test, however, as already noted, is less sensitive. Nevertheless, a large change, such as the 62% reduction in threshold which Gruber (1914) claimed to have seen, would have been immediately noticed. WORM_ -. -0a I Fig. 8. A, control ACh potential evoked by a 2 msee pulse of 1-3 x 10-6 A. B and C, a continuous sequenceof ACh potentials in the same fibre, ata frequencyof 2/sec. The signal at the beginning of C indicates the time during which adrenaline bitartrate (60,ug) was injected into the muscle bath. After ending stimulation at 2/sec, test pulses were applied every 30 sec to illustrate the recovery of the sensitivity to ACh. Records obtained from an end-plate of a diaphragm at room temperature. The time constant of the R.C. amplifier was 0-3 see. The resting potential was 75 mv and was tending to increase. Other authors have observed either no change in the electrical excitability of the muscle after adrenaline (Luco, 1939) or a reduction (Obre, 1923); even Gruber (1922) later found depression in denervated muscle. Other effects of adrenaline Decurarizing action. The enhancement by adrenaline of neuromuscular transmission blocked by curare is now too well substantiated to be much in doubt (Gruber, 1914; Rosenblueth, Lindsley & Morison, 1936; Hutter & Loewenstein, 1955), in spite of the negative conclusion reached by Brown et al. (1948). We have not observed any synergism between adrenaline and curare, like that described by Naess & Sirnes (1953). It would seem essential to eliminate decisively the possibility that this phenomenon was secondary to the large and prolonged vascular changes produced by adrenaline in rabbit muscle in situ (Girling, 1951). Since adrenaline increases the amplitude of e.p.p.'s at junctions blocked by curare and magnesium (and possibly by fatigue), it does not seem likely that its action is due to a specific inhibition of curare. We have already shown that

12 302 K. KRNJEVIC AND R. MILEDI there is probably no sensitization of the end-plate to ACh. The most likely explanation is that there may be an increase in the amount of ACh released by each nerve impulse. Changes in resting potential. Our results do not show any significant effect of adrenaline on the muscle fibre resting potential, confirming the negative observations made by Hutter & Loewenstein (1955) on frog muscle in vitro. Brown et al. (1950) found a small change in demarcation potential in cat muscle in situ, which they interpreted as an increase in resting potential. They correlated the absence of such an effect in frog muscle in vitro with the failure of adrenaline to augment the twitch tension of unfatigued frog muscle; however, adrenaline has been shown clearly to have a potentiating action on the unfatigued rat diaphragm in vitro (Goffart & Brown, 1947; Goffart, 1949, 1952). 25_ ;\ ' 10_ is-~~~~~~~~~~~~~~~~ Time (min) Fig. 9. Effect of adrenaline HC1 (80 jig) applied at arrow, during a depression of neuromuscular propagation caused bythe repetitive indirect stimulation of a diaphragm at room temperature. The intracellular electrode was not in an end-plate region. The resting potential was 60 mv and the frequency of stimulation 12-5/sec. Each point is the mean percentage rate of transmission over a period of 30 sec. Conclusions Adrenaline has a number of actions which influence neuromuscular propagation in several ways. It produces a reversal of presynaptic failure of conduction, and more ACh may be released by each nerve impulse under its influence. On the other hand, it also decreases the electrical excitability of the muscle fibre. Since these effects conflict with each other, the total result would depend upon which predominates. As a further complication, during continued repetitive activity the restoration of presynaptic conduction abolishes the protective effect of the failure, and hence may eventually predispose to a more complete post-synaptic block. The over-all effect is usually probably beneficial as far as the contraction of the muscle is concerned. This is shown in Fig. 9 which illustrates the temporary improvement, under the influence of adrenaline, in the repetitive activity of a failing fibre recorded at a point distant from its end-plate. There is a striking

13 ADRENALINE AND NEUROMUSCULAR PROPAGATION 303 change in the rate of firing; however, it should be noted that the effect is more transient than is usually the case in vitro when recording at an end-plate (Figs. 1, 2). This suggests that post-synaptic failure is setting a limit to the main presynaptic effect. This multiplicity of, and mutual interference between, presynaptic and postsynaptic events may help to explain the variety of effects ascribed to adrenaline. SUMMARY 1. We have studied the action of adrenaline on neuromuscular propagation in the rat diaphragm (in vitro) and gracilis (in situ). 2. In a majority of trials adrenaline relieved at least in some degree the intermittent presynaptic failure of conduction caused by tetanic stimulation of the motor nerve. 3. Adrenaline produced no detectable effect in vitro on potentials evoked artificially at end-plates with microjets of ACh, or on the amplitude of spontaneous miniature end-plate potentials. 4. Adrenaline (in vitro and in situ) increased the amplitude of e.p.p.'s in muscles paralysed by curare, excess Mg, or even fatigue. This suggests that it may cause more transmitter to be liberated at the neuromuscular junction. 5. Adrenaline in vitro tended to decrease the electrical excitability of muscle fibres stimulated at a low frequency. At high frequency it failed to change in any significant way the usual progressive increase in threshold. 6. Adrenaline regularly caused an acceleration of the spontaneous discharge of miniature e.p.p.'s. There was no appreciable action on the level of the resting membrane potential of muscle fibres. 7. The over-all action of adrenaline depends upon which effect predominates. Full restoration of presynaptic conduction may predispose to an eventual complete post-synaptic failure. REFERENCES BENDER, M. B. (1938). Fright and drug contractions in denervated facial and ocular muscles of monkeys. Amer. J. Phy8iol. 121, BROWN, G. L., BUTBRING, E. & BURNS, B. D. (1948). The action of adrenaline on mammalian skeletal muscle. J. Phy8iol. 107, BROWN, G. L., GOFFART, M. & VIANNA DIAS, M. (1950). The effects of adrenaline and sympathetic stimulation on the demarcation potential of mammalian skeletal muscle. J. Physiol. 111, BifLBRING, E. & WHITTERIDGE, D. (1941). The effect of adrenaline on nerve action potentials. J. Phy8iol. 99, BURN, J. H. (1945). The relation of adrenaline to acetylcholine in the nervous system. Physiol. Rev. 25, COREILL, A. B. & TIEGS, 0. W. (1933). The effect of sympathetic nerve stimulation on the power of contraction of skeletal muscle. J. Phy8iol. 78, DALE, H. H. & GADDUM, J. H. (1930). Reactions of denervated voluntary muscle, and their bearing on the mode of action of parasympathetic and related nerves. J. Physiol. 70, DEssy, S. & GRANDIS, V. (1904). Contribution a l'etude de la fatigue. Action de l'adrenaline sur la fonction du muscle. Arch. ital. Biol. 41,

14 304 K. KRNJEVI6 AND R. MILEDI ELLIS, S. & BECKETT, S. B. (1955). Depression of neuromuscular transmission by epinephrine. Fed. Proc. 14, 336. FRANK, E., NOTHMANN, M. & GUTTMANN, E. (1923). tber die toniseche Kontraktion des quergestreiften Saugetiermuskels nach Ausschaltung des motorischen Nerven. Pflhi.q. Arch. ge8. Physiol. 199, GIRLrNG, F. (1951). Effects of intravenous and intra-arterial adrenaline, and of adrenaline after priscoline, in hind limb of intact rabbit. Amer. J. Physiol. 164, GOFFART, M. (1949). Calcium et action potentiatrice de quelques amines sympaticomim6tiques sur la contraction du muscle strie non-fatigue de mammifere. Experientia, 8, 332. GOFFART, M. (1952). Recherches relatives a l'action de l'adr6naline sur le muscle strie de mammif6re. 1. Potentiation par l'adrenaline de la contraction maximale du muscle non-fatigue. Arch. int. Physiol. 60, GOFFART, M. & BROWN, G. L. (1947). Relation entre le potassium du milieu extracellulaire et l'action de l'adr6naline sur le muscle strie non fatigu6 du Rat. C.R. Soc. Biol., Paris, 141,958. GOFFART, M. & RrTcHIE, J. M. (1952). The effect of adrenaline on the contraction of mammalian skeletal muscle. J. Phy8iol. 116, GRUBER, C. M. (1914). Studies in fatigue. IV. The relation of adrenalin to curare and fatigue in normal and denervated muscle. Amer. J. Physiol. 34, GRUBER, C. M. (1919). The significance of epinephrin in muscular activity. Endocriaology, 3, GRUBER, C. M. (1922). Studies in fatigue. XI. The effect of intravenous injection of massive doses of adrenalin upon skeletal muscle at rest and undergoing fatigue. Amer. J. Physiol. 61, HUTTER, 0. F. & LOEWENSTEIN, W. R. (1955). Nature of neuromuscular facilitation by sympathetic stimulation in the frog. J. Physiol. 130, KRNJEVI6, K. & MILEDI, R. (1957). Adrenaline and failure of neuromuscular transmission. Nature, Lond., 180, KRNJEVIC, K. & MILEDI, R. (1958a). Motor units in the rat diaphragm. J. Physiol. 140, KRN.TEVIC, K. & MILEDI, R. (1958b). Failure of neuromuscular propagation in rats. J. Physiol. 140, LEGOUIX, J. P. & MINZ, B. (1953). ]thude de l'action de l'adr6naline sur le potentiel d'action du nerf de grenouille perfuse. C.R. Soc. Biol., Paris, 147, LILEY, A. W. (1956a). An investigation of spontaneous activity at the neuromuscular juinction of the rat. J. Physiol. 132, LILEY, A. W. (1956b). The effect of presynaptic polarization on the spontaneous activity at the mammalian neuromuscular junction. J. Physiol. 134, LOEWENSTEIN, W. R. (1956). Modulation of cutaneous mechanoreceptors by sympathetic stimulation. J. Physiol. 132, LOEWENSTEIN, W. R. & ALTAMIRANO-ORREGO, R. (1956). Enhancement of activity in a Pacinian corpuscle by sympathomimetic agents. Nature, Lond., 178, Luco, J. V. (1939). The defatiguing effect of adrenaline. Amer. J. Physiol. 125, MAIRE, F. W. & PATTON, H. D. (1956). Role of the splanchnic nerve and the adrenal medulla in the genesis of 'preoptic pulmonary edema'. Amer. J. Physiol. 184, NAESS, K. & SIRNES, T. (1953). A synergistic effect of adrenaline and d-tubocurarine on the neuromuscular transmission. Acta physiol. scand. 29, OBR*, A. (1923). Action de l'adrenaline et de l'extrait surr6nal sur l'excitabilit6 musculaire. C.R. Soc. Biol., Paris, 88, PANELLA, A. (1907). Action du principe actif surrenal sur la fatigue musculaire. Arch. ital. Biol. 48, ROSENBLUETH, A., LINDSLEY, D. B. & MORISON, R. S. (1936). A study of some decurarizing substances. Amer. J. Physiol. 115, TORDA, C. & WOLFF, H. G. (1946). Effect of epinephrine and physostigmine on the response of striated muscle to acetylcholine and potassium. Amer. J. Physiol. 146, WASTL, H. (1928). (ber den Einfluss des Adrenalins und einiger anderer Inkrete auf die Kontraktionen des Warmbluterskelettmuskels. Pflfig. Arch. ges. Physiol. 219,