Brain Research, 127 (1977) 219-234

219

© Elsevier/North-HollandBiomedical Press

MECHANISMS OF EXCITATION AND INHIBITION IN THE NIGROSTRIATAL SYSTEM

T. L. RICHARDSON, J. J. MILLER* and H. McLENNAN Department of Physiology, Faculty of Medicine, University of British Columbia, Vancouver B.C. I/6T 1 I¥5 (Canada}

(Accepted September 16th, 1976)

SUMMARY The extracellular responses of neurones in the neostriatum following single pulse stimulation of the substantia nigra were investigated in urethane anaesthetized rats. Low intensity stimulation ( < 10 V) evoked single large amplitude spikes while higher intensities (10-20 V) elicit a high frequency burst of small amplitude spikes or waves. When spontaneous or glutamate-induced large spikes are recorded, nigral stimulation causes their inhibition coincidentally with the development of a burst. If the burst is prevented, the inhibitory response disappears. Both the nigral evoked inhibition and burst response are unaffected by iontophoretically or systemically administered antagonists of dopamine or by chemical lesions of the dopamine-containing nigral neurones. The monosynaptic activation of large amplitude striatal neurones, which could also be identified antidromically by stimulation of the globus pallidus, was reversibly blocked by dopamine antagonists. It is concluded (a) that the burst responses are induced through the antidromic excitation of striatonigral axons within the striatum; (b) that the striatal neurones thus activated are inhibitory interneurones and (c) that the dopamine-containing neurones of the nigra make excitatory synaptic contact with a population of striatal output cells, some of which at least project to the globus pallidus.

INTRODUCTION There have been abundant anatomical and histochemical studies demonstrating a direct dopaminergic projection from neurones lying in the pars compacta of the substantia nigra (SN) to the neostriatuml,15,82,as,a6,41, 43. Although dopamine (DA) is presumed to be released as a synaptic transmitter from the terminals of this path* Scholar of the Medical Research Council of Canada.

220 way, the effect which it exerts on striatal neurones has remained controversial. The majority of such cells have been shown to be depressed by the iontophoretic application of DA which suggests that this substance mediates an inhibitory response; but at the same time a significant although smaller proportion of striatal cells which were excited by DA was also reported2,a,7,zo, 39,44,45. Inasmuch as a positive correlation between the inhibitory influences of DA applied iontophoretically and the effects of nigral stimulation on striatal neurones has been claimed, many investigators have concluded that DA acts exclusively as an inhibitory transmitter in the nigrostriatal system5,7,1o,19. On the other hand, a significant number of excitatory actions have been demonstrated following nigral stimulationll-14,16,39, 4~, and intracellular studies have indicated that this is by far the most common response elicited from this region a,21,22,zg,a°. These data are clearly not compatible with the view that the nigrostriatal dopaminergic pathway exerts an inhibitory effect on the neostriatum. Frigyesi and Purpura 16 postulated two functionally distinct nigrostriatal pathways subserving the inhibitory and excitatory actions on neostriatal cells, while Feltz and Albe-Fessard 12 suggested that there is a single excitatory input impinging on striatal target neurones and that the inhibitory component is due to intrinsic mechanism within the striatum. These suggestions have remained speculative in view of the lack of identification of the various cell types within the neostriatum which are responsive to nigral stimulation. The present experiments were designed to approach a solution to these problems first by identifying neurones electrophysiologically in the striatum which are responsive to nigral and pallidal stimulation; secondly by determining the effects of pharmacological agents administered both systemically and iontophoretically upon these neurones, and finally by observing the effects of electrolytic or chemical lesions interrupting the nigrosfriatal pathway. The results indicate (a) that dopamine is an excitatory transmitter of the pathway and (b) that the inhibition observed in the striatum following DA application or SN stimulation is due to a recurrent inhibitory system dependent upon interneurones located within the striatum itself. METHODS The experiments were performed on 68 male Wistar rats acutely prepared under urethane anaesthesia (1.5 g/kg, i.p.) and maintained at a body temperature of 36-37 °C. They were placed in a stereotaxic instrument and the calvarium removed so that the underlying cortex was exposed over an area roughly corresponding to boundaries 2.5 mm anterior and 4 mm posterior to bregma and 4 mm on either side of the sagittal suture. Concentric bipolar stimulating electrodes having a tip separation of 0.3-0.5 mm were positioned in the substantia nigra (SN), globus pallidus (GP) or along the nigrostriatal bundle using coordinates from the atlas of K6nig and Klippe131. Stimuli were single square-wave pulses of 0.1 msec duration and 0.1-0.4 mA intensity delivered through an isolation unit. Extracellular activity was recorded using either single glass capillaries filled with

221 4 M NaCI and having tip diameters of 1-2 #m or through the central barrel, containing 4 M NaC1, of 7-barrelled micropipette assemblies having an overall tip diameter of 4-8 #m. The recording electrodes were stereotaxically directed and lowered by a micromanipulator into the neostriatum. The following compounds were placed in the outer barrels of the multipipette assemblies; sodium L-glutamate (0.5 M, pH 8, Nutritional Biochemicals), dopamine hydrochloride (1 M, pH 4.0, Regis Chemical), a-flupenthixol (0.5 Mr, pH 4.0, H. Lundbeck and Co.). The drugs were ejected electrophoretically using appropriate anionic or cationic currents. Haloperidol (0.5-2.5 mg/ kg, McNeil Laboratories) and a-flupenthixol (0.5-2.5 mg/kg) were also administered intravenously. Electrical signals from the recording electrodes were led through a high-pass filter (1-10 kHz), amplified and displayed on an oscilloscope. Action potentials of single neurones were also passed through an amplitude-discriminating window and used to fire a Schmitt trigger, the output of which was integrated and displayed on a paper chart to give a continuous record of firing frequency. The same output could also be fed into a PDP-8L computer to produce poststimulus time histograms of the cellular responses. Stimulus sites were marked by passing anodal currents (10-20/,A) through the central core of the concentric electrode with subsequent identification of a Prussian blue spot in histological sections. Recording sites were marked by filling the electrodes with Pontamine sky blue (2 ~ in 4 M sodium acetate), expelling the dye electrophoretically and locating the spot histologically. Lesions interrupting perinigral pathways which might have mediated some of the observed influences on neostriatal activity were carried out in preliminary operations 3 weeks to 2 months prior to acute electrophysiological experiments. Extensive electrolytic lesions of the intralaminar-parafascicular region of the thalamus in three animals and of the SN in 5 animals was made by passing DC currents up to 2 mA for 15-30 sec, while in another 4 animals aspiration of the cerebral cortex dorsal and rostral to the head of the striatum was carried out. In 6 experiments selective depletion of striatal dopamine was achieved by using 6-hydroxydopamine hydrobromide (6OHDA, Regis Chemical). Animals were pretreated with desmethylimipramine (25 mg/kg, i.p.) 1 h prior to a unilateral injection of 6-OHDA (12 #g dissolved in 4 #1 of 0.15 M NaC1 containing 1 mg/ml of ascorbic acid) into the nigrostriatal bundle (AP -k 5.9, L 2.3, V q- 1.9). The effectiveness of the chemical lesions was confirmed by measuring the dopamine content of the ipsilateral striatum compared to the contralateral control according to the method of McGeer and McGeer za. Retrograde transport of horseradish peroxidase (HRP, type VI, Sigma Chemical) was examined in 5 preparations. The animals were anaesthetized with pentobarbital (50 mg/kg, i.p.), and 0.1/A of a 10 ~ solution of HRP in saline was injected unilaterally into the SN by a stereotaxically guided microsyringe. Following survival periods of 24 h the rats were killed and perfused with a solution of 3 ~o paraformaldehyde aod 2 ~ glutaraldehyde in 0.05 M phosphate buffer (pH 7.5, 24 °C). The brains were removed and allowed to stand for 24 h in phosphate buffer containing 5 ~ sucrose. Frozen sections were then cut at 50/zm, treated to reveal peroxidase activity according to

222 the method of Graham and Karnovsky TMand examined microscopically under darkfield illumination. RESULTS

Inhibitory responses of striatal neurones Action potentials with amplitudes > 300 #V have been recorded from more than 300 neurones of the neostriatum. Some were spontaneously active (mean frequency 5.2/sec, range < 1-20/sec); however the majority fired only when glutamate was ejected iontophoretically from a barrel of an electrode assembly. As shown in Fig. 1A, single maximal stimuli applied to SN inhibited the discharge of such spontaneously firing or glutamate-activated neurones for periods ranging from 60 to 300 msec (mean 175 msec). A slight 'rebound' excitation lasting 50-150 msec frequently followed the period of inhibition. The firing of these neurones was also depressed by the iontophoretic application of dopamine (Fig. IB). When records such as these were obtained, it was noted that io the 3-10 msec period following stimulation a response consisting of a high frequency burst of small amplitude (50-300 #V) spikes or waves was evident, either at the original recording site or within 50 #m of it. This response, examples of which are shown in Fig. 2 was A.

B.

50

DA60

DA60

25

0 ag o.

t~

09 25

0

DA50 ~1

DA50

~ 30 see. msee,

6(~0

Fig. 1. A: the response of two spontaneously discharging neostriatal cells to single pulse stimulation of the substantia nigra. Poststimulus histograms show the summed responses to 50 stimuli. In this and subsequent figures the shock artifact is indicated by the first large deflection of the histogram, and the bin width is 10 msec. B: ratemeter records of the same two neurones indicating the inhibitory action of iontophoreticaUy applied dopamine (DA, 60 nA, 50 nA). The periods of application are shown by the solid horizontal bars.

223 A.

B.

C.

J c

I0 V

I HZ V = 4.50

jl 15V

20V

2 0 Hz

40HZ

Recovery

( 30 sec )

V=4.55

V ,4.63

V = 4.82

f

i._T5o ~v 2 rnsec,

Fig. 2. Characteristics of burst responses elicited by stimulation of the substantia nigra. A : the change in amplitude and latency associated with increasing stimulus intensity (10, 15, 20 V). B: the change produced by repetitive pulse stimulation at 1, 20 and 40 Hz. C: the variability in response with movements of the recording electrode along a vertical tract in the striatum extending from 4.50 to 4.82 mm below surface of cortex (320/~m). Each trace represents 5-10 sweeps; the initial deflection (marked by arrows) indicates shock artifact. encountered throughout the striatum. The temporal relationship between the appearance of this burst and the start of the period of inhibition of the larger amplitude spikes suggested that it may represent the discharge of inhibitory interneurones.

Identification of inhibitory interneurones The burst responses elicited by single pulse stimuli applied to the SN consisted of 2-8 'ripples' or 'spikes' of variable amplitude and the frequency within the burst ranged from 200 to 900 Hz. As the intensity of SN stimulation was increased, the amplitude of the individual components of the burst response usually became larger and the latency of onset was reduced (Fig. 2A). With increasing rates of SN stimulation the amplitude of the bursts decreased and they failed to follow at frequencies above 40 Hz (Fig. 2B). It was also observed that as the recording electrode was moved along a vertical track, the burst could be evoked for several hundred microns before a 'silent zone' was encountered. The variability in the burst response within an 'active zone' is shown in Fig. 2C. When the position of the tip of an electrode recording a burst was marked with pontamine blue, it was always located in the vicinity of cellular distribution which lies between the corticospinal fascicles running through the neostriatum. Conversely,

224

SN

A. 50-

°U

J

0i

i

IC

O.ImV ec.

msec co

5()0

c

I C Stimulalion (SN Lesion) Oi

0

I

rebec

I000

50-

2msec.

msec.

! 900

Fig. 3. The effects of iontophoretic application of glutamate (Glu), substantia nigra (SN) and internal capsule (IC) stimulation on the response pattern of large amplitude neostriatal neurones. A: poststimulus histogram (PSH) showing inhibition of a spontaneously firing neurone following single pulse stimulation of SN. Inset record shows prolongation of burst response following application of 10 nA of Glu. B : elimination of inhibitory period following depolarization block of burst response (inset) by 25 nA Glu. Note that spontaneous discharge of large amplitude spike was only slightly altered by Glu. C: PSH indicating inhibition from both SN and IC sites. Inset records of fast sweep speeds illustrate comparable burst responses evoked from both sites. Diagrammatic section of the rat brain at the level of the hypothalamus indicates effective sites (Q) in region of the IC which evoked burst response. Dotted line indicates extent of electrolytic lesions which eliminate the SN evoked responses in striatum. D: PSH showing inhibition of large spike produced by IC stimulation 3 weeks following electrolytic lesion of SN. Inset record shows IC elicited burst response recorded at same site. Arrows on inset recordings indicate shock artifact. LH, lateral hypothalamus; V, ventral thalamus.

burst responses were not recorded when the electrode tip was located within regions of high fiber density. It was noted too that m o v e m e n t of the stimulating electrode dorsal or caudal to SN abolished the burst response a n d that it was only when the electrode was in that structure that such responses appeared. These findings a n d the fact that the bursts are responsive to the iontophoretic application of glutamate (Fig. 3A) indicate (a) that they reflect n e u r o n a l rather t h a n axonal activity; (b) that they are synaptically produced a n d (c) that they represent the

225 quasisynchronously evoked responses of a number of neurones which are recruited as the intensity of SN stimulation is raised. Finally, to the earlier evidence that there is a temporal relationship between the appearance of a burst and the beginning of inhibition of the larger amplitude spikes, may be added the following. A burst may be prolonged by the iontophoretic application of a small quantity of glutamate (compare the inset record of Fig. 3A with the record of Fig. 2; however if the application of glutamate is increased the neuronal elements giving rise to the burst are inactivated by depolarization (Fig. 3B, inset) and the inhibition of the large spikes which had been observed previously was abolished (histograms of Fig. 3A and B). We conclude therefore that the bursts represent the activity of a population of inhibitory interneurones located in the striatum. These interneurones, synaptically activated from SN, in turn influence other striatal target cells identifiable by large extracellular action potentials.

Identification of the pathway mediating the burst response In three animals electrolytic lesions of the intralaminar and parafascicular nuclei of the thalamus and in 4 others extensive removal by aspiration of the cerebral cortex dorsal and rostral of the neostriatum were performed. After 3 weeks to 2 months the striatal responses to stimulation of SN were unchanged in these animals. Stimulating electrodes were positioned along the pathway between the SN and neostriatum at the level of the hypothalamus in 12 preparations. Effective sites located in the medial and lateral extent of the internal capsule (IC) elicited burst discharges with concomitant inhibitions of larger amplitude spikes which were indistinguishable from those obtained from SN except that the latency of onset was reduced (Fig. 3C). Electrolytic lesions placed at these sites in 5 experiments effectively eliminated the SN evoked responses. Actions of pharmacological agents on the burst response The foregoing observation prompted an examination of the possibility that the bursts were attributable to the excitation of the dopamine-containing neurones of SN. The results of 4 groups of experiments indicate however that this is not the case. The intravenous administration of haloperidol (0.5-2.5 mg/kg) or of a-flupenthixol (0.52.5 mg/kg), or the iontophoretic application of the latter failed to modify either the burst responses or their associated inhibition. In 6 animals unilateral depletion of striatal dopamine (90-95 ~) brought about by the injection of 6-OHDA into the nigrostriatal bundle was similarly ineffective. Thus although the burst response appears to be synaptically evoked and is elicited by stimulation of SN and not by neighbouring structures, it does not depend upon the dopaminergic nigrostriatal system. The remaining possibility that the bursts were due to antidromic activation of striatonigral fibres, collaterals of which make synaptic contact within the striatum with the interneurones giving rise to the inhibitions, was tested in five animals who survived 4 weeks after extensive lesions of the SN. In all cases subsequent stimulation of the IC revealed that the striatal responses were unchanged even though neostriatal dopamine contents were decreased below detectable levels (Fig. 3D).

226

A.

B.

C.

D.

J

Fig. 4. A: diagrammatic representation of the distribution of positively labelled H R P cells in the neostriatum. Each dot marks the position of one cell. Hatched area indicates region from which large amplitude neurones were synaptically excited by SN stimulation. AC, anterior commissure; CC, corpus callosum; S, septum. B: antidromic spike elicited by 100 Hz nigral stimulation. These responses were recorded from positions in the ventral and 'peripheral shell' regions corresponding to those shown in (A). Large deflection is shock artifact. C: photomicrograph of positively labelled HRP cells in neostriatum corresponding to area outlined in (A) and the injection site in the SN (D). Magnifications: C, x 250; D, x 40.

In view of the possibility that the burst response was mediated by axon collaterals of a descending striatonigral system, attempts were made to record antidromic responses in the striatum following SN stimulation. Several criteria were used to identify striatal output cells as being antidromically activated by the SN: (a) relatively short latencies of firing (!.9-4.5 msec); (b) constant latency of firing at threshold and (c) the ability to follow at least 100 Hz stimulus trains (Fig. 4B). In several neurones responding in this manner, IS-SD breaks were also observed and stimulus trains between 150 and 200 Hz resulted in eliciting only the IS component. None of the neurones which displayed these characteristics showed spontaneous firing. Histological verification of the recording placements indicated that these antidromically evoked cells were restricted to the ventral aspect and 'peripheral shell' of the neostriatum (Fig. 4A). Further confirmation that the neurones in these regions of the striatum presumably are the cells of origin of a descending striatonigral pathway were shown by the retrograde transport of HRP following injections into the region of the SN. Labelled cells were found in regions roughly corresponding to those from which the antidromic responses were recorded (Fig. 4A and C). There were no differences in the distribution of positively labelled neurones in the striatum following SN injections into either the anterior or posterior extent of this structure. Although the conclusions to be drawn from the retrograde transport experiments are limited because of the diffusion of HRP from the injection site (Fig. 4D), the coincidence of the topographical localization shown by

227 A.

B.

C.

5V SN I HZ

,or

GP I00 HZ

t

SN IOHz

t

t

~Tso/Jv I m6e¢.

20 V 2 msec.

, Tso,~v IO msec Fig. 5. Activation of large amplitude neurones in neostriatum following stimulation of SN. A: effects of increasing stimulus intensity from 5 to 20 V. Note that the large amplitude spike is blocked by the burst response at higher intensities. B: activation of large spike following SN stimulation at 1 and 10 Hz. The response failed to follow above 50 Hz. C: the same cell as in (B) evoked antidromically from globus pallidus (GP) stimulation and indicating ability of this response to follow 100 Hz stimulation.

Arrows indicate shock artifacts. both electrophysiological and histological procedures is strong evidence to suggest that the identified cells are those giving rise to the striatonigral pathway.

Properties of large amplitude spike discharges In contrast to the previously described effects of maximal SN stimuli, single pulses at low intensities evoked single spikes of large amplitude ( > 300 FV) with latencies ranging from 4 to 18 msec (mean 10.8 msec) (Fig. 5A). As the intensity of stimulation was increased, a high frequency burst was observed, and with further increases the burst response was augmented but the single spike was abolished. The large amplitude spike evoked in this manner followed stimulation frequencies between 20 and 50 Hz but with considerable variability in latency (Fig, 5B) and this, together with the fact that these latencies are long, suggests an orthodromic activation. Furthermore when the neurones were either spontaneously active or glutamate-activated and spikes occurred during the latent period preceding an evoked spike, there was no evidence of occlusion due to collision, again suggesting an orthodromic origin. Single pulse stimuli applied to the globus pallidus (GP) were observed to evoke large amplitude antidromic spikes in the striatum with latencies between 1.1 and 2.6 msec. The short latency of this response together with the constancy of latency with threshold stimuli and ability to follow stimulus trains of 100-200 Hz suggests that these neurones were antidromically evoked by GP stimulation (Fig. 5C). Further

228

Con trol

Haloperidol (5 rain)

Recovery (30 rain )

i To.2 mV 1o mseo

Fig. 6. The effects of intravenous administration of haloperidol (0.5 mg/kg) on SN evoked responses. Left and right panels are recordings from two different cells. On the right simultaneous recording of burst response with large amplitude spike. Haloperidol blocks synaptically evoked spikes but not burst responses 5 rain following injection. Recovery of responses occurs after 30 rain. evidence of antidromicity was tested by using collision-extinction procedures on 27 neurones showing convergence from both SN and GP stimulus sites. The longest latent period at which the G P response was blocked by the preceding SN evoked orthodromic spike was 5.0 msec. This approximates the calculated critical latent period based on the conduction time of the antidromic spike evoked by GP plus the refractoriness of these neurones (1.0-2.0 msec) as determined by double pulse stimulation 9. Histological verification of the location of 15 of these neurones which showed a convergence from both the SN and GP indicated that they were restricted to the centromedial 'core' region of the striatum (Fig. 4A). That the single spike responses evoked from the SN are due to activation of the dopaminergic system is indicated by the results shown in Fig. 6, where in two different experiments the spikes were reversibly blocked by an intravenous administration of haloperidol. Note that in the second instance the burst recorded from the same site was unaffected by this procedure. Similarly the antidromic evoked spike from GP stimulation remained unaltered by haloperidol. The excitatory response elicited by low intensity stimulation of the SN was followed by a period of inhibition ranging from 40 to 220 msec in glutamate-excited or spontaneously active striatal neurortes. Similar periods of inhibition could be elicited following the pallidal antidromic response. The onset of the inhibitory period

229 elicited from both sites was accompanied in many instances by a burst response displaying similar characteristics to those evoked by supramaximal SN stimulation. The temporal relationship between the burst response and the inhibition again suggested that there were interneurones mediating the effect. In 7 preparations in which stimulating electrodes were positioned in both the SN and GP, attempts were made to determine whether the same interneurones could be activated by supramaximal SN stimulation and low intensity SN or GP stimulation. Nine of 35 burst responses recorded from the 'core' region of the striatum were observed to be activated by both sites. DISCUSSION Previous studies have emphasized the inhibitory action which stimulation of the substantia nigra exerts on extracellular unit activity in the neostriatum and the parallel effect of dopamine iontophoretically applied to striatal neurones. These findings have been interpreted to indicate that this compound is the synaptic transmitter of the nigrostriatal pathway which mediates these inhibitory responsesS,7,10,19. The results of the present investigation indicate that a different mode of inhibition appears more probable. Stimulation of the SN, in addition to producing inhibition of striatal neurones whose action potentials were of large amplitude, was observed also to evoke small amplitude, high frequency spikes or waves. The temporal correlation between this response and the onset of inhibition in the target cells, as well as the fact that elimination of the small amplitude spikes abolished the inhibition, suggest that they represent the discharge of inhibitory interneurones within the striatum. Three possible systems may be involved in the activation of these interneurones: (1) a direct innervation by orthodromic projections of the nigrostriatal pathway, (2) an indirect innervation by axon collaterals of the striatonigral system, or (3) an innervation from axons which pass through or close to SN from neurones situated elsewhere. The persistence of the burst response and the concomitant inhibition of the target cells following depletion of striatal DA by 6-OHDA lesions and after the administration of the antagonists a-flupenthixol and haloperidol, indicate that the dopaminergic nigrostriatal system is not involved in mediating either of these synaptically induced responses. The possibility that non-dopaminergic projections from the SN to the striatum are responsible is also untenable since electrolytic lesions of the SN, which produced a complete depletion of striatal DA as well as degeneration of all orthograde systems emanating from the nucleus, did not interfere with either of the responses when these were evoked by stimulation at the level of the internal capsule. These data, together with the fact that striatal neurones could be activated antidromically by SN stimulation at latencies which were slightly shorter than that required to elicit the orthodromic burst discharge, indicate that striatal inhibition is mediated by interneurones which are innervated by axon collaterals of the descending striatonigral pathway. Recent studies have demonstrated a projection system from the raphe nucleus to the striatum which courses through the region of and slightly dorsal to SN 8,4°.

230 Since this pathway has been shown to exert a direct inhibitory effect on striatal target cells 4°, the possibility exists that the effects of SN stimulation may in fact reflect the activation of fibres of passage of this pathway. However in acute experiments carried out 3-4 weeks following electrolytic lesions of the raphe nuclei, responses in the striarum elicited by SN stimulation were unaltered. These data indicate that at least two independent systems are capable of exerting inhibitory effects on striatal neurones, and the conclusion that the inhibition elicited from SN is due to the antidromic activation of striatal inhibitory interneurones remains. There are some inconsistencies in the literature regarding the antidromic responses of striatal neurones elicited from SN. The latencies of those recorded in the present study agree with the results obtained by Frigyesi and Purpura TM and York 45. Others however have reported antidromic responses with inordinately long latencies, in the range 8-20 msec30, 34. We have, however, invariably found that responses occurring with such long latencies show all of the characteristics of orthodromic rather than antidromic activation. It is however possible that there may be two striatonigral projection systems exhibiting different conduction velocities 17. The present results have also shown that in addition to the inhibitory system described above, stimulation of SN can produce an excitation of striatal cells. Stimulation at intensities lower than that necessary to produce inhibition of target cells resulted in a single spike activation or an activation-inhibition sequence when the neurones were firing spontaneously. The fact that the excitatory component would follow relatively high rates of stimulation (30-50 Hz) before decomposition suggests that it is mediated by the monosynaptic nigrostriatal pathway. Since many of the striatal neurones excited by SN stimulation could also be evoked antidromically from sites in the globus pallidus, it would appear that they may be further characterized as the output cells for the striatopallidal pathway25, 28. In the period immediately following both the SN evoked excitation and the pallidal evoked antidromic spike, small amplitude burst responses similar to those described were recorded. When the cells were firing spontaneously stimulation at both of these sites produced inhibition following the initial excitation, and once again the onset of this period coincided with the discharge of the interneurones. This would suggest that a recurrent collateral system in the efferent striatopallidal pathway is involved in the control of inhibitory component of the activation-inhibition sequence elicited by SN stimulation. This interpretation is in agreement also with intracellular studies of striatal neurones where the response to SN stimulation is an EPSP frequently followed by an IPSp21,22, 29,3°. The lack of major convergence upon the interneurones of the pathway from SN and globus pallidus indicates that at least two populations are involved in controlling the activity of striatal target cells. Previous studies have emphasized the inhibitory action of caudate stimulation on pallidal neurones ~,37,~7. The results of the present investigation would, however, suggest that the striatopallidal projection is excitatory on target cells in the GP. Whether this discrepancy is due to a species difference between cats used in other studies and rats used in these experiments or to a sampling bias in the recorded neurones remains to be investigated.

231 In contradistinction to earlier claims, there is little to suggest that dopamine is involved in striatal inhibitions elicited from SN. Nevertheless the evidence for dopamine as a neurotransmitter in the nigrostriatal pathway is reasonably conclusive and in the present experiments administration of haloperidol was shown to reversibly block the synaptically evoked excitation of striatal cells without altering either the spontaneous activity or the amplitude of the pallidal evoked antidromic response. Furthermore in preparations chronically treated with 6-OHDA, single spike activation could not be elicited from SN. These data therefore indicate that dopamine mediates an excitatory rather than an inhibitory action on striatal target cells. It should be noted that this conclusion does not agree with that of Feltz and De Champlain 1~, who stated that 6 - O H D A lesions did not alter the excitatory response of striatal neurones. It seems possible however that these authors were observing interneuronal discharges which have been shown above to be non-dopaminergic. Several studies have demonstrated that a small number of neurones encountered in the striatum are excited by the iontophoretic application of dopamine 29,45,46, and our data infer that this may be the only action of this compound in the nigrostriatal pathway. There therefore remains the question how the depression produced in the majority of cells, and which has been consistently reported, may come about. A number of possible explanations may be offered. It is conceivable first of all that when dopamine is applied near a target cell a brief initial excitation would be missed in the ensuing inhibition induced by the collateral activation of inhibitory interneurones. Secondly, as York 45 has suggested it may be that striatal neurones possess two types of dopamine receptors, one giving rise to an excitatory response and the other to inhibition; since iontophoretically applied dopamine depresses neurones also in regions where a dopaminergic input is not sus!

CD

I

oP

I

SNC $NR

Fig. 7. A schematic illustration of the proposed synaptic arrangements of neostriatal neurones (CD) with the substantia nigra and globus pallidus (GP). Stimulation of the nigral compacta (SNC) region produces an activation-inhibition sequence of striatal target cells, the inhibitory component being mediated by recurrent axon collaterals impinging on inhibitory interneurones (shown in black). GP stimulation elicits antidromic spike in the same target cell and inhibition of spontaneous firing by the same axon collateral system. Stimulation of the reticulata region of SN (SNR) evokes antidromic responses mediated by the striatonigral pathway and inhibition of CD target cells by recurrent collaterals impinging on inhibitory interneurones.

232 pected it is further possible that the 'inhibitory' receptor is a pharmacological phenomenon but is of no synaptic importance. The synaptic arrangements which may be inferred from these experiments are represented in Fig. 7. Striatal target cells exhibit two types of responses following stimulation of the substantia nigra: (1) a monosynaptic excitation which is frequently followed by inhibition, and (2) a pure inhibition. While the nigrostriatal dopaminergic pathway evokes the excitatory response, the inhibitory components in both instances are mediated by recurrent collateral systems which innervate inhibitory interneurones, in the first case by collaterals of the striatopallidal system and in the second by those of the striatonigral pathway. Recordings of antidromically evoked responses following stimulation of SN and GP have demonstrated that the neurones giving rise to the two efferent systems are located in different regions of the striatum. The majority of SN antidromic responses were restricted to the ventral and peripheral 'shell' regions, and this distribution was confirmed by the presence of positively labelled neurones in the same areas following H R P injections into the SN. GP evoked responses on the other hand were localized mainly to the central 'core' region of the striatum. These data are in accord with what is known of the neuronal architecture of the striatum. Large and medium sized cells are believed to form the efferent fibre systems of this structure and these therefore presumably represent those neurones which are antidromically activated by SN and GP stimulation26, 2s,42. The axons of these larger neurones exhibit extensive collateral branching which terminate on smaller intrinsic cells distributed throughout the striatuma,24, 27. It is the discharge of a population of these latter cells which we correlate with the burst responses and with the inhibition of striatal target cell activity. ACKNOWLEDGEMENTS This work was supported by grants from the Medical Research Council of Canada to J. J. M. and H. McL. The technical assistance of H. Brandejs is appreciated.

REFERENCES 1 And6n, N. E., Dalstr6m, A., Fuxe, K., Larsson, K., Olson, L. and Ungerstedt, U., Ascending monoamine neurones to the telencephalon and diencephalon, Acta physiol, scand., 67 (1966) 313326. 2 Bloom, F. E., Costa, E. and Salmoiraghi, G. C., Anesthesia and the responsiveness of individual neurones of the caudate nucleus of the cat to acetylcholine, norepinephrine and dopamine administered by microelectrophoresis, J. PharmacoL exp. Ther., 150 (1965) 244-252. 3 Buchwald, N. A., Price, D. D., Vernon, L. and Hull, C. D., Caudate intracellular response to thalamic and cortical inputs, Exp. Neurol., 38 (1973) 311-323. 4 Cajal, S. Ram6n Y., Histologie du Syst~me Nerveaux de l'Homme et des Vertebras, VoL H, Maloine, Paris, 1911, pp. 504-514. 5 Connor, J. D., Caudate unit responses to nigral stimuli: evidence for a possible nigro-neostriatal pathway, Science, 160 (1968) 899-900. 6 Connor, J. D., Caudate nucleus neurones: correlation of the effects of substantia nigra stimulation with iontophoretic dopamine, J. Physiol. (Lond.), 208 (1970) 691-703.

233 7 Connor, J. D., Electrophysiology of the nigro-caudate pathway, Pharmacol. Ther., 1 (1975) 357-370. 8 Conrad, L. C. A., Leonard, C. M. and Pfaff, D. W., Connections of the median and dorsal raphe nuclei in the rat: an autoradiographic and degeneration study, J. camp. Neurol., 156 (1974) 179205. 9 Darian-Smith, I., Phillips, G. and Ryan, R. D., Functional organization in the trigeminal main sensory and rostral spinal nuclei of the cat, J. PhysioL (Lond.), 168 (1963) 129-146. 10 Feltz, P., Dopamine, amino acids and caudate unitary responses to nigral stimulation, J. PhysioL (Land.), 205 (1970) 8-9P. 11 Feltz, P., Sensitivity to haloperidol of caudate neurones excited by nigral stimulation, Europ. J. Pharmacol., 14 (1971) 360-364. 12 Feltz, P. and Albe-Fessard, D., A study of an ascending nigro-caudate pathway, Electroenceph. clin. Neurophysiol., 33 (1972) 179-193. 13 Feltz, P. and De Champlain, J., Persistence ofcaudate unitary responses to nigral stimulation after destruction and functional impairment of the striatal dopaminergic terminals, Brain Research, 43 (1972) 595-600. 14 Feltz, P. and MacKenzie, J. S., Properties of caudate unitary responses in repetitive nigral stimulation, Brain Research, 13 (1969) 612-616. 15 Fibiger, H. C., McGeer, E. G. and Atmadja, S., Axoplasmic transport of dopamine in nigrostriatal neurons, J. Neurochem., 21 (1973) 373-385. 16 Frigyesi, T. L. and Purpura, D. P., Electrophysiological analysis of reciprocal caudate-nigral relations, Brain Research, 6 (1967) 440-456. 17 Fuller, D. R. G., Hull, C. D. and Buchwald, N. A., Intracellular responses of caudate output neurons to orthodromic stimulation, Brain Research, 96 (1975) 337-341. 18 Graham, R. C. and Karnovsky, M. J., The early stages of absorption of injected horseradish peroxidase in the proximal tubules of mouse kidney: ultrastructural cytochemistry by a new technique, J. Histochem. Cytochem., 14 (1966) 291-302. 19 Gonzalez-Vegas, J. A., Antagonism of dopamine-mediated inhibition in the nigro-striatal pathway: a mode of action of some catatonia-inducing drugs, Brain Research, 80 (1974) 219-228. 20 Herz, A. and Zieglgansberger, W., The influence of microelectrophoretically applied biogenic amines, cholinomimetics and procaine on synaptic excitation in the corpus striatum, Int. J. Neuropharmacol., 7 (1968) 221-230. 21 Hull, C. D., Bernardi, G. and Buchwald, N. A., Intracellular responses of caudate neurons to brain stem stimulation, Brain Research., 22 (1970) 163-179. 22 Hull, C. D., Bernardi, G., Price, D. D. and Buchwald, N. A., Intracellular responses of caudate neurons to temporally and spatially combined stimuli, Exp. Neurol., 38 (1973) 324-336. 23 Hull, C. D., Levine, M. S., Buchwald, N. A., Heller, A. and Browning, R. A., The spontaneous firing pattern of forebrain neurons. I. The effects of dopamine and non-dopamine depleting lesions on caudate unit firing patterns, Brain Research, 73 (1974) 241-262. 24 Kemp, J. M., Observations on the caudate nucleus of the cat impregnated with the Golgi method, Brain Research, 11 (1968) 467-470. 25 Kemp, J. M., The termination of strio-pallidal and strionigral fibres, Brain Research, 17 (1970) 125-128. 26 Kemp, J. M. and Powell, T. P. S., The structure of the caudate nucleus of the cat: light and electron microscopy, Phil. Trans. B, 262 (1971) 383-401. 27 Kemp, J. M. and Powell, T. P. S., The synaptic organization of the caudate nucleus, Phil. Trans. B, 262 (1971) 403-412. 28 Kemp, J. M. and Powell, T. P. S., The connexions of the striatum and globus pallidus: synthesis and speculation, Phil. Trans. B, 262 (1971) 441-457. 29 Kitai, S. T., Sugimori, M. and Kocsis, J. D., Excitatory nature of dopamine in the nigro caudate pathway, Exp. Brain Res., 24 (1976) 351-363. 30 Kitai, S. T., Wagner, A., Precht, W. and Ohno, T., Nigrocaudate and caudato-nigral relationship: an electrophysiological study, Brain Research, 85 (1975) 44-48. 31 K6nig, F. C. and Klippel, R. A., The Rat Brain, Williams and Wilkins, Baltimore, Md., 1963. 32 Kuyper, H. G. J. M., Kievit, J. and Groen-Klevant, A. C., Retrograde axonal transport of horseradish peroxidase in rat's forebrain, Brain Research, 67 (1974) 211-218. 33 Levine, M. S., Hull, C. D. and Buchwald, N. A., Pallidal and entopeduncular intracellular responses to striatal, cortical, thalamic and sensory inputs, Exp. Neurol., 44 (1974) 448-460.

234 34 Liles, S. L., Single unit responses of caudate neurons to stimulation of the frontal cortex, substantia nigra and entopeduncular nucleus in cats, J. Neurophysiol., 37 (1974) 254-265. 35 Lindvall, O. and Bj6rklund, A., The organization of the ascending catecholamine neuron systems in the rat brain, Acta. physiol, scand., Suppl, 412 (1974) 1-48. 36 Maler, L., Fibiger, H. C. and McGeer, P. L., Demonstration of the nigrostriatal projection by silver staining after nigral injections of 6-hydroxydopamine, Exp. Neurol., 40 (1973) 505-515. 37 Malliani, A. and Purpura, D. P., lntracellular studies of the corpus striatum. II. Patterns of synaptic activities in lenticular and entopeduncular neurons, Brain Research, 6 (1967) 341-354. 38 McGeer, E. G. and McGeer, P. L., Catecholamine content of spinal cord, Canad. J. Biochem. Physiol., 40 (1962) 1141-1151. 39 McLennan, H. and York, D. H., The action of dopamine on neurones of the caudate nucleus, J. Physiol. (Lond.), 189 (1967) 393-402. 40 Miller, J. J., Richardson, T. L., Fibiger, H. C. and McLennan, H., Anatomical and electrophysiological identification of a projection from the mesencephalic raphe to the caudate-putamen in the rat, Brain Research, 97 (1975) 133-138. 41 Moore, R. Y., Bhatnaga, R. K. and Heller, A., Anatomical and chemical studies of a nigro-neostriatal projection in the cat, Brain Research, 30 (1971) 119-135. 42 Szabo, J., Topical distribution of striatal efferents in the monkey, Exp. Neural., 5 (1962) 21-36. 43 Ungerstedt, U., Stereotaxic mapping of the monoamine pathways in the rat brain, Acta. physiol. scand., Suppl. 367 (1971) 1-48. 44 York, D. H., The inhibitory action of dopamine on neurons of the caudate nucleus, Brain Research, 5 (1967) 263-266. 45 York, D. H., Possible dopaminergic pathway from substantia nigra to putamen, Brain Research, 20 (1970) 233-249. 46 York, D. H., Amine receptors in C.N.S. 1I. Dopamine. In L. L. lversen, S. D. Iversen and S. H. Snyder (Eds.), Handbook ofPsychopharmacology, Vol. 6, Plenum Press, New York, 1975, pp. 2361. 47 Yoshida, M., Rabin, A. and Anderson, M., Monosynaptic inhibition of pallidal neurons by axon collaterals of caudato-nigral fibers, Exp. Brain Res., 15 (1972) 333-347.