SYNAPSE 12:287-303 (1992)

Role of the Subthalamic Nucleus in the Regulation of Nigral Dopamine Neuron Activity IAN D. SMITH AND ANTHONY A. GRACE Departments of Behavioral Neuroscience and Psychiatry, Center for Neuroscience, University of Pittsburgh, Pittsburgh, Pennsylvania 15260

KEY WORDS

Substantia nigra, Burst firing, Bicuculline, Muscimol

ABSTRACT The influence of subthalamic nucleus (STN) afferents on dopaminergic (DA)neurons of the rat substantia nigra (SN) was investigated. Hemisections of the brain placed between the STN and the SN or located anterior to the STN caused an increase in the firing rate of DA cells without producing significant changes in their firing pattern. In contrast, electrolytic and ibotenic acid lesions of the STN resulted in 93% and 49% reductions, respectively, in the level of burst firing without affecting the firing rate of DA cells recorded in the lateral SN. Furthermore, procedures which interrupted the STN input to the SN produced rapid pacemaker-like firing in 18% of the lateral SN DA neurons recorded. Activation of the STN using single pulses of electrical stimulation caused: 1) a 20-50 msec inhibition of DA cell firing followed by an excitation, which in 35% of DA cells was accompanied by spikes occurring in a burst-like pattern, and 2) a short-latency inhibition lasting 5-25 msec in 75% of non-DA SN zona reticulata (ZR) neurons. On the other hand, stimulation of the STN for 1 minute at 20 Hz resulted in an initial decrease in DA cell burst firing followed by elevated firing rates and increased burst firing by 30-60 minutes after the stimulation. Pharmacological activation of the STN by infusion of bicuculline caused a rapid inhibition of DA cells followed by a two-fold increase in burst firing 6-14 minutes later, whereas SN ZR cells responded with an elevation in firing rate which dissipated in 6-14 minutes. Muscimol-induced STN inhibition produced complimentary biphasic changes in SN neuron firing: 1) an initial increase followed by a decrease in burst firing and firing rate of DA neurons and 2 ) a rapid inhibition followed by an excitation of ZR cells over a similar time course. Thus, the STN appears to exert a dual action on SN DA cells: 1) initial inhibition possibly mediated through STN excitation of the inhibitory SN ZR projections to DA cells, and 2 ) a facilitation of burst firing which may be a direct effect of excitatory STN afferents. 0 1992 Wiley-Liss, Inc

INTRODUCTION The subthalamic nucleus (STN) has been gaining increased recognition for its role in regulating extrapyramidal motor function. This attention is due, in part, to its widespread interconnections with structures of the basal ganglia as well as for its potential involvement in the motor deficits observed in Parkinson’s disease (Aziz et al., 1991; Bergman et al., 1990). It has been established that STN efferents are excitatory in nature (Hammond et al., 1978; Kitai and Kita, 1987; Nakanishi et al., 19871,with extensive collaterals innervating the entopeduncular nucleus, globus pallidus, and substantia nigra (SN),in addition to a comparatively minor projection to the striatum (Groenwegen and Berendse, lgS8; Van Der 1990; Takada et and Hattori’ 1980). Thus, STN efferents are organized in a manner 0 1992 WILEY-LISS, INC.

that would allow this nucleus to modulate information flow within the basal ganglia and their output nuclei. Neuroanatomical studies indicate that the presumed glutamatergic (Albin et al., 1989) STN efferents innervating the SN project preferentially to the SN zona reticulata (ZR) GABAergic output cells. However, dopaminergic (DA) neurons of the zona compacta (ZC) may also receive afferents from the STN (Damlama and Tepper, 1991; Groenwegen and Berendse, 1990; f i t a and Kitai, 1987; Ricardo, 1980). Electrophysiological evidence shows that stimulation of the STN evokes monosynaptic excitatory postsynaptic potentials in SN - -~

Received February 20,1992, accepted rn revised form April 17,1992 Address reprint requests t o AnthonyA Grace, Department of Behavioral Neuroscience, 458 Crawford Hall, Univer5ity of Pittsburgh, Pittsburgh, PA 15260

288

I.D. SMITH AND A.A. GRACE

DA neurons recorded in vitro (Nakanishi et al., 1987). Attempts to determine the significance of this projection to SN DA neurons in situ, however, have yielded mixed results. One of the primary afferents to the STN in rats is the inhibitory GABAergic projection from the globus pallidus (Canteras et al., 1990; Kita et al., 1983; Rouzaire-Dubois et al., 1980; Smith et al., 1990). For this reason, prior studies investigating the effects of STN efferents on their target nuclei have relied on manipulating GABAergic receptors in this nucleus to alter STN neuronal activity without activating fibers of passage (Feger and Robledo, 1991; Robledo and Feger, 1990). Studies in which the STN is pharmacologically activated report only small increases in the firing rate of some DA cells whereas other DA neurons showed inhibition (Robledoand Feger, 1990). Spontaneously active SN DA cells in the anesthetized rat are known to fire action potentials in one of two distinct patterns of activity: irregular single spiking or burst firing (Grace and Bunney, 1984a,b). Furthermore, DA cells recorded in unanesthetized behaving animals exhibit rapid transitions between single spiking and burst firing (Freeman et al., 1985) and are reported to fire bursts of spikes during locomotor tasks and when presented with reward-associated stimuli (Diana et al., 1989; Romo and Schultz, 1990; Schultz and Romo, 1990). It has been suggested that an intact afferent input to SN DA neurons is required for the generation of burst firing, since DA cells in the deafferented midbrain slice preparation discharge only in a highly regular pacemaker-like pattern (Grace and Onn, 1989; Kita et al., 1986; Pinnock, 1984; Silva and Bunney, 1988). Burst firing in DA neurons is believed to play a significant role in the function of this system. For example, stimulation of DA cell axons in patterns resembling normal DA neuron burst firing has been shown to cause a two- to three-fold increase in the amount of DA released per impulse as compared to stimuli delivered at the same frequency but at constant interstimulus intervals (Gonon, 1988). In addition, burst firing in DA neurons may also be involved in the release of peptide cotransmitters (Bean et al., 1990) as has been shown for burst firing in other systems (Lundberg et al., 1989). This last function may be of particular relevance for the DA system in view of immunocytochemical studies showing that cholecystokinin (CCK), neurotensin, and acetylcholinesterase may be colocalized with DA in subpopulations of midbrain neurons (Greenfield et al., 1980; Hokfelt et al., 1984; Seroogy et al., 1989). Additional evidence that the STN may be involved in the regulation of SN DA cell firing characteristics is provided by recent investigations into the pathophysiology of Parkinson’s disease. Metabolic mapping and electrophysiological measures have provided evidence for STN activation following large 1-methyl-4-phenyl1,2,3,64etrahydropyridine(MPTP)-induced dopamine depletions in monkeys (Miller and DeLong, 1987;

Mitchell et al., 1989; Palumbo et al., 1990). Similarly, in 6-OHDA-treated rats, DA neurons exhibit higher firing rates and more burst firing after depletions of 95% of striatal DA levels (Hollerman and Grace, 1990). Since lesions of the nigrostriatal pathway also result in an elevation of striatal unit discharge (Nisenbaum et al., 1988) and a reduction in neuronal activity in the globus pallidus (Pan and Walters, 1988),it is possible that the resultant disinhibition of STN neurons projecting to the SN could contribute to the increases in DA cell firing observed in rats with extensive DA depletions (Hollerman and Grace, 1990,1992). Several studies, including those examining striatonigral interactions, have provided examples in which stimulation of afferents to the SN can elicit opposite effects on DA cell firing. In particular, the GABAergic striatonigral projection is thought to indirectly excite DA neurons via the inhibition of GABAergic interneurons in the SN ZR as well as provide direct inhibitory input to DA cells (Grace and Bunney, 1979,1985).However, it is not known whether the excitatory STN projection to the SN affects DA cell firing directly or indirectly. The present study examines the net effects of manipulating STN neuron activity on the firing characteristics of SN DA neurons. Portions of these results have been presented previously (Smith and Grace, 1990,1991a,b). MATERIALS AND METHODS Male Sprague-Dawley rats obtained from Zivic Miller and weighing 250-400 g were anesthetized with chloral hydrate (400 mgkg intraperitoneally ti.p.1) and mounted in a stereotaxic instrument (Narishige). Body temperature was maintained at 37°C with a heating pad, and supplemental chloral hydrate was administered via a lateral tail vein. In experiments in which survival surgery was implemented, the initial surgery was performed using Equithesin anesthesia (0.3 mU0.1 kg, i.p.1. All surgical procedures were carried out in accordance with The Guide for the Care and Use of Laboratory Animals published by the US Public Health Service (USPHS) and were approved by the Institutional Animal Care and Use Committee of the University of Pittsburgh. DA cell identification and recording Single unit recordings from identified DA cells were made using glass microelectrodes pulled from 2.0 mm 0.d. Omegadot glass tubing (WPI) using a vertical microelectrode puller (Narishige) and broken back under microscopic control to a tip diameter of approximately 1 pm. Electrodes were filled with 2% Pontamine sky blue in 2 M NaC1, with electrode impedances ranging from 5-10 Megohms when measured at 135 Hz using a Micro Electrode Tester (Winston Electronics Co., BL-1000). Electrode potentials were amplified by a high-input impedance headstage amplifier and a preamplifier (Fintronics, Orange, CT), displayed on a Hitachi Storage

STN EFFECTS ON DA CELL FIRING

Oscilloscope (V-134) and monitored on a Grass AM-8 audio monitor. Spikes were discriminated and counted using a time and amplitude window discriminator (Fintronics). Single unit extracellular recordings were made from DA neurons as previously described (Hollerman and Grace, 1990). Briefly, after drilling a burr hole in the skull overlying the SN (coordinates 2.9 mm anterior from lambda, 2.2 mm lateral to the midline), electrodes were lowered into the brain using a hydraulic microdrive (Kopf, model 640). Data were recorded from neurons in the ZC region of the SN that exhibited spike activity consistent with that reported for identified DA neurons. At least 3 minutes of spontaneous activity from SN DA neurons was recorded prior to stimulation or drug administration. DA cells were identified by their extracellular waveform and firing pattern (Bunney et al., 1973; Grace and Bunney, 1980, 1983; Guyenet and Aghajanian, 1978). Electrophysiological data were digitized and stored on VHS videotapes using a Neurodata Neurocorder connected to a videocassette recorder for subsequent off-line analysis of cell firing rate and pattern. Recordings were made from neurons in the region of the SN located between 1.9 mm and 2.5 mm anterior to lambda. In some experiments, the recorded DA cells were subdivided for analysis according to their relative location within the SN as follows: lateral = L2.7-L2.4, central = L2.4-L2.1, medial = L2.1L1.8, based on stereotaxically defined coordinates from Paxinos and Watson (1986). The average level of activity of the DA cells recorded in the SN was assessed by passing the electrode nine times through this region in a predefined pattern, analogous to that described previously (Bunney and Grace, 1978). Briefly, the electrode was passed through the three lateral planes of the SN at three anterior coordinates (A1.9,A2.2, and A2.5) in a randomized pattern, and the firing characteristics of each DA cell encountered was logged. Following recordings, current was passed through the electrode (-30 pA for 15 minutes) to eject dye, marking the recording site (Thomas and Wilson, 1965). The rats were then perfused transcardially with saline followed by 10% buffered formalin. After removing the brains and fixing them overnight in buffered formalin, the tissue was sectioned parasagittally into either 35 pm or 70 pm sections. Stimulation, lesion, injection, and recording sites were then identified in cresyl violet-stained sections. DA neuron firing pattern was assessed by determining the percentage of spikes fired in bursts. This analysis was performed on representative samples of 500 consecutive spike events from each neuron tested and interspike interval histograms (ISIHs) were generated using a CED 1401 interface and Spike2 software adapted for this purpose (Cambridge Electronic Design, Cambridge, England). The onset and offset of bursts of action potentials were defined as interspike intervals of less than 80 msec and greater than 160 msec, respec-

289

tively, as described previously (Grace and Bunney, 1984b). The data presented represent the average of these values from cells collected for that location/ treatment group across each animal tested. In a subset of experiments, the effects of the STN on non-DA cells of the SN ZR were investigated. These neurons were identified by the following criteria: spontaneous firing rates of 3-30 Hz, extracellular action potential duration of less than 1.7 msec, and a characteristic brief, rapid excitatory response to a foot pinch (Grace and Bunney, 1979, 1985). Only SN ZR neurons encountered within 300 pm of DA cells recorded in the SN ZC were included in this study.

STN lesions In a subset of experiments, either electrolytic or excitotoxic lesions of the STN were made at least 24 hours prior t o recording. Bilateral electrolytic lesions of the STN were performed by passing 250 mA direct current for 20 seconds through a monopolar tungsten microelectrode (50 pm tip exposure) implanted in the STN. Rats receiving electrolytic lesions were recorded between 1-12 days post-lesion. Bilateral chemical lesions were performed by infusing 200 nl of 50 p M ibotenic acid (Sigma) (Schwarcz et al., 1979; Contestabile et al., 1984)dissolved in 0.9% saline into each STN at a rate of 100 nl per minute using a micropipette connected t o a Hamilton syringe. Rats were allowed to recover for at least 4 days following ibotenic acid lesions, and subsequent recordings were performed between 4-20 days post-lesion. For both lesion groups, no significant differences in cell firing were noted in rats recorded at earlier times post-lesion vs. those recorded at the longest post-lesion intervals. The position and extent of both electrolytic and excitotoxic lesions were confirmed histologically. Brain hemisections In another series of experiments, hemisections of the brain were performed in anesthetized rats at least 1 hour prior to recording. Hemisections were made with a microknife constructed from a glass coverslip and encompassed a plane beginning at approximately 1 mm lateral to the midline and extending through the lateral brainstem and cortical white matter. Fourteen animals were hemisected between the STN and the SN at approximately A46 and 12 animals received hemisections anterior to the STN at approximately A5.6 (Paxinos and Watson, 1986). Hemisections were considered complete if their ventral extent was sufficient to transect the fibers of the descending internal capsule. The site and extent of hemisections were verified histologically in sagittal brain sections. Electrical stimulation of the STN Electrical stimulation of the STN was performed using bipolar concentric stimulating electrodes (Rhodes Medical Instruments SNE-100) with a 50 pm tip expo-

290

I.D. SMITH AND A.A. GRACE

sure. The electrode tip was stereotaxically positioned in the STN a t A5.0, L2.4, and V8.0 (Paxinos and Watson, 1986). Single pulses of 30 p s duration and 30-200 pA amplitude were delivered at a frequency of 0.5 Hz. Stimulation was initiated using a current amplitude of 30 FA, and increased until an effective level of inhibition or excitation was evoked (typical current = 50 PA). Peri-stimulus time histograms (PSTHs; 400 bins, 1 msechin) were generated by the CED 1401 interface and software from approximately 100 stimulus sweeps collected for each neuron studied. A deviation of PSTH events by at least 25% from control values averaged over a prestimulus period of 50 msec was required for a cell to be classified as responding to the stimulation. In other experiments, DA neuron activity was recorded after stimulating the STN with 1-minute trains of constant current pulses with amplitudes ranging between 50 and 200 pA (50 ps duration) at a frequency of 20 Hz. Pharmacological activation and inhibition of STN neurons Microinjections into the STN were performed using a micropipette connected to a Hamilton syringe. Pipette tips were broken back under microscopic control to a diameter of approximately 20 pm. At least 3 minutes of baseline firing was recorded from nigral neurons before injecting the STN with 200 nl of a solution containing either 200 pM bicuculline methiodide (BMI) or 200 pM muscimol dissolved in 0.9% saline. Injections were performed over a period of 1 minute. In a prior study by Robledo and Feger (19901,larger doses of BMI (390 pM) and muscimol(900 pM) resulted in a sustained activation or inhibition of the STN which lasted for periods of up to 30 minutes. Smaller doses were chosen in the present study in order to limit drug diffusion and increase the likelihood of observing recovery from drug treatment during the recording period. Changes in discharge rate and pattern were quantified by calculating the firing frequency for a 1-3 minute period centering around the maximal or minimal level of firing exhibited by a particular cell following drug infusion.

which, for DA cells exhibiting a high level of burst firing, resulted in a distribution pattern characteristic for these events (Figs. 1A,B, 2A,B). No significant differences in firing frequency or burst firing were found among DA neurons recorded a t different locations within the SN (Fig. 2C). Consistent with previous reports (Grace and Bunney, 1984b), burst firing in DA neurons is poorly correlated with their baseline firing rate (r = 0.59, Fig. 3A). Nonetheless, except for a single extreme case, very slow firing DA neurons rarely exhibited burst firing and rapidly firing DA cells typically fired a t least a portion of their spikes in bursts (i.e., only 1 of 100 neurons recorded was firing faster than 5.0 spikes per second and had less than 5% of its spikes occurring in bursts). This is particularly evident when the analysis is restricted to DA cells recorded in the lateral SN in which the correlation between firing rate and burst firing is highest (i.e., all DA neurons firing above 5.0 spikes per second exhibit at least 20% burst firing, r = 0.82, n = 31; Fig. 3A).

Lesions of the STN Electrolytic lesions of the STN (n = 7 rats) were made at least 24 hours prior to recording. SN DA neurons recorded after STN lesions (66 cells) fired significantly fewer spikes in a bursting pattern when compared to control (10.3 2 2.3% spikes fired in bursts vs. 18.6 2 2.4%in control, P < .Ol), although no significant difference in firing rate was observed (Fig. 4A). Furthermore, DA neurons recorded in the lateral SN after STN lesions fired only 1.5 0.7% of their spikes in bursts, a value which was less than 7% of that observed in the lateral SN of intact rats (control = 16.8 2 4.1%,P < .01, Figs. 4A, 5A,B). In particular, a greater proportion (18%,4/22) of DA cells recorded in the lateral SN after lesions were categorized as nonbursting, rapidly firing neurons (i.e., firing rate >5 Hz with 6% spikes in bursts) whereas no lateral DA cells recorded in controls met this criterion ( P < .05; Figs. lC,D, 3B). In rats in which the electrolytic lesion was placed outside of the STN cell group and did not interrupt the STN-SN pathway (n = 41, no significant alteration in DA cell firing Statistics rate (3.9 ? 0.7 Hz, n = 32 cells) or burst firing (overall: Statistical significance levels were calculated using 16.1% 2 3.2%, n = 32; lateral cells: 12.6% 2 3.8%, n = Student’s t-test unless data were not normally distrib- 12) was found. Although electrolytic lesions were capable of elimiuted, in which case the Kolmogorov-Smirnov test was used. Statistics are reported as mean 2 standard error. nating STN cells within a small, discrete area, it was possible that an interruption of descending fibers of RESULTS passage by the lesion could have affected the results Baseline firing properties of SN DA neurons obtained. Therefore, a series of experiments was done The firing rate and burst firing data for SN DA cells in which the activity of DA cells (88 cells) was assessed were quantified in untreated control animals (n = 19 in rats in which the STN neurons were lesioned using rats). Consistent with previous reports (Grace and Bun- ibotenic acid injections to preserve fibers of passage (n ney, 1980,1983,1984a,b),spontaneously active SN DA = 6 rats). In rats recorded 4-20 days after ibotenic acid cells fired at rates below 10 Hz (mean = 4.3 2 0.2 Hz, n lesions of the STN, DA cells in the lateral SN exhibited = 100) and exhibited moderate levels of burst firing a 49% reduction in burst firing with respect to control (18.6 5 2.4% spikes fired in bursts). Burst firing in DA (P < .05; Figs. 4B, 5C).As found in rats with electrolytic neurons was also analyzed by constructing ISIHs lesions of the STN, this shift to a more regular firing

STN EFFECTS ON DA CELL FIRING

291

B

A I

1.07

I

20

0.5

e

$-

15

B

0.0

L

2 10

c)

E:

2 0

-0.5

5

-1.0 0 0

1

2

3

4

5

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

Seconds

Seconds

D

C

20 1

05

> E

O0

Seconds Fig. 1. A: Example of a rapidly firing DA neuron in an intact rat displaying burst firing, which consists of a series of spikes exhibiting a progressive decrease in spike amplitude and a post-burst quiescent period. This neuron is firing at 7.2 spikes per second with 75% of the spikes occurring within bursts. B: Analysis of this burst firing pattern by plotting its ISIH shows the broad distribution of the intervals between consecutive spikes in the train (standard deviation = 0.12

seconds). C: Following an electrolytic lesion of the STN, a DA neuron firing a t a comparable rate (7.0 Hz) exhibits a very regular firing pattern. D: The ISIH of this regularly firing neuron shows the narrow distribution of intervals between spikes that is characteristic of this firing pattern (standard deviation = 0.03 seconds). Bin width for B and D = 1 msec.

pattern occurred without a significant change in DA neuron firing rate. In addition, a greater proportion of DA cells recorded in the lateral SN after ibotenic acid lesions were found to fire rapidly in a nonbursting pattern (20%of cells firing >5 Hz [i.e., 6/30 cellsJhad