identification of a subpopulation of substantia nigra pars compacta ?-aminobutyric acid neurons that...

Identification of a Subpopulation ofSubstantia Nigra Pars Compacta

g-Aminobutyric Acid Neurons That isRegulated by Basal Ganglia Activity

M.O. HEBB AND H.A. ROBERTSON*Laboratory of Molecular Neurobiology, Department of Pharmacology, Dalhousie University,

Halifax, Nova Scotia B3H 4H7, Canada

ABSTRACTIn this report, the authors provide a novel description of a population of g-aminobutyric

acid-containing neurons in the substantia nigra, pars compacta (SNC). By using metabolicmapping of the immediate-early gene, c-fos, the activation pattern of these cells wascharacterized with respect to basal ganglia stimulation. Dopaminergic stimulation withd-amphetamine or apomorphine induced Fos expression in the central region of the SNC.However, lesions of the nigrostriatal dopamine pathway significantly reduced d-amphet-amine- and apomorphine-induced Fos expression in the ipsilateral and contralateral SNC,respectively. Suppression of stimulant-induced Fos expression in the striatum, using anti-sense oligodeoxynucleotides, also eliminated Fos expression in the ipsilateral SNC, indicatingthat striatal efferent projections are involved in the activation of these cells. Double-labelingimmunohistochemistry revealed that the Fos-positive cells did not express tyrosine hydroxy-lase but were immunoreactive for glutamic acid decarboxylase. Retrograde labeling ofnigrostriatal neurons, combined with Fos immunofluorescence, revealed that these Fos-positive cells did not project to the striatum. Thus, these neurons do not appear to comprise anondopaminergic nigrostriatal circuit but likely represent locally-projecting interneurons ofthe substantia nigra. J. Comp. Neurol. 416:30–44, 2000. r 2000 Wiley-Liss, Inc.

Indexing terms: c-fos; glutamic acid decarboxylase; interneuron; dopamine; antisense

oligodeoxynucleotides; 6-hydroxydopamine

The role of dopamine (DA) in the basal ganglia and itscontribution to motor and psychocognitive functioning hasbeen under investigation for several decades. Loss ofmesencephalic DA innervation to the caudate putamen isthe primary cause of motor and cognitive decline inParkinson’s disease (PD), whereas hyperactivity in limbicDA systems produces psychosis and may contribute to thesymptomatology of schizophrenia (Hornykiewicz, 1979;Snyder, 1982). The DAergic neurons of the substantianigra pars compacta (SNC) project to the striatum andrelease DA onto striatonigral and striatopallidal neuronsto activate D1 and D2 receptors, respectively (Robertsonand Robertson, 1987; Gerfen et al., 1990; Graybiel, 1990,1991; LeMoine et al., 1990, 1991). Because DA D1 recep-tors (D1R) are coupled positively, whereas D2 receptors(D2R) are coupled negatively, to both adenylate cyclaseand inositol phospholipid signaling pathways (Kebabianand Calne, 1979; Stoof and Kebabian, 1981; Mahan et al.,1990), stimulation of these receptors increases transmis-sion through direct striatal projections but decreases

transmission through indirect striatal projections. In addi-tion to modulating efferent transmission from the stria-tum, DAergic neurons extend dendritic arborizations intothe underlying substantia nigra pars reticulata (SNR),where the release of DA regulates SNR activity (Cheramyet al., 1981; Robertson, 1992).

Greater than 85% of neurons in the SNC are DAergic,whereas the remainder, known as secondary cells, arethought to be cholinergic or g-aminobutyric acidergic(GABAergic; Javoy-Agid et al., 1981; Lacey et al., 1989).The function of these secondary cells remains unclear.Electrophysiological studies have revealed similarities

Grant sponsor: The Parkinson Foundation of Canada; Grant sponsor:The Hereditary Disease Foundation; Grant sponsor: MRC of Canada.

*Correspondence to: H.A. Robertson, Laboratory of Molecular Neurobiol-ogy, Department of Pharmacology, Sir Charles Tupper Medical Building,Dalhousie University, Halifax, Nova Scotia B3H 4H7, Canada.E-mail: [email protected]

Received 13 August 1998; Revised 28 June 1999; Accepted 19 July 1999

THE JOURNAL OF COMPARATIVE NEUROLOGY 416:30–44 (2000)

r 2000 WILEY-LISS, INC.

between secondary neurons in the SNC and interneuronsof other brain regions (Miles and Wong, 1984; Madison andNicoll, 1988; Lacey et al., 1989; Yung et al., 1991). How-ever, van der Kooy et al. (1981) have shown that a smallpopulation of non-DAergic cells in the SNC sends projec-tions to the striatum, supporting the notion that secondarycells also may contribute to the regulation of the basalganglia.

In this report, we provide a novel description of apopulation of cells in the SNC, the activation of whichappears to be correlated with changes in basal gangliaactivity. These cells can be activated reliably by administra-tion of the psychostimulant drugs, d-amphetamine orapomorphine, and they are found consistently at the levelof the interpeduncular nucleus but not always at otherrostrocaudal levels of the SNC. The objectives of this studywere to determine the neurochemical phenotype of thesecells as well as to elucidate the mechanism by which theyare activated. We also were interested in determiningwhether these cells projected rostrally to the striatum. Toaddress these issues, we examined the effects of impairedbasal ganglia circuitry on the stimulant-induced activa-tion of these SNC neurons. The expression of the immedi-ate-early gene, c-fos, was used to metabolically map theactivation of cells in the SNC in response to direct (apomor-phine) and indirect (d-amphetamine) dopamine agonists.The effects of nigrostriatal DA depletion, using 6-hydroxy-dopamine (6-OHDA) lesioning (Ungerstedt, 1968), werecompared with those produced by direct suppression ofstriatal efferent transmission, using intracerebral infu-sions of antisense oligodeoxynucleotides targeted to c-fosmRNA (Chiasson et al., 1992, 1994; Dragunow et al., 1993,1994; Heilig et al., 1993; Sommer et al., 1993; Hooper et al.,1994; Hebb and Robertson, 1997a,b). Previous studieshave shown that infusion of anti-c-fos oligodeoxynucleo-tides into the striatum not only suppressed Fos inductionin this region but also reduced the release of GABA fromstriatonigral terminals and increased the activation ofneurons in the globus pallidus (Sommer et al., 1996; Hebband Robertson, 1997b). The latter effect was consistentwith a reduction in inhibitory, GABAergic striatopallidalactivity. Infusion of control oligodeoxynucleotides had noeffect on either parameter. Thus, although the mechanismremains unclear, it appears that suppression of striatalc-fos induction effectively reduced the efferent transmis-sion from this region. We have used antisense oligodeoxy-nucleotides to alter striatal activity without interferingwith DA transmission from the SNC. We provide a neuro-anatomic and neurochemical description of SNC neuronsthat are activated differentially after alterations in thestriatum and the substantia nigra.

MATERIALS AND METHODS

Experimental design

All experimental procedures were approved by the Ani-mal Care Committee at Dalhousie University. AdultSprague-Dawley rats weighing 275–400 grams were usedin this study. Surgery-naive animals were challenged withd-amphetamine (n 5 8). Also, two groups of animalsreceived a unilateral lesion of the nigrostriatal pathway byusing 6-OHDA and subsequently were challenged withd-amphetamine (6-OHDAamp group; n 5 4) or apomor-phine (6-OHDAapo group; n 5 5). A fourth group ofanimals received a unilateral infusion of antisense oligode-

oxynucleotides targeting c-fos mRNA into the striatumfollowed by administration of d-amphetamine (ASF group;n 5 5).

Surgical procedures

All surgeries were performed under halothane-inducedanesthesia with the animals mounted in a Kopf stereotaxicapparatus (Kopf, Inc., Tujunga, CA). Both 6-OHDA andoligodeoxynucleotide infusions were performed by pres-sure injection by using a CMA 100 (Carnegie-Medicin,Stockholm, Sweden) microinjection pump with 25-gaugesteel cannulae.All stereotaxic coordinates given are accord-ing to Paxinos and Watson (1997). Animals that received6-OHDA lesions were treated with desipramine (25 mg/kg,i.p.) 30 minutes prior to infusion of 6-OHDA (4 µl of 12 mMsolution at 0.5 µl/minute) into the right medial forebrainbundle [coordinates from Bregma: anteroposterior (AP),23.6 mm; lateral (LAT), 2.0 mm, dovsoventral (DV). 28.8mm). After infusion, the cannula was left in place for anadditional 2 minutes to allow for diffusion away from theinjection site. After surgery, the skin was sutured, andanimals were left to recover for 3 days, 7 days, or 21 days.Animals that received postoperative recovery periods of 3days and 7 days were used only to assess the rate of lesionprogression; Fos mapping and double-labeling studieswere performed only in animals that had 21-day recoveryperiods.

Animals in the ASF group received bilateral placementof cannulae into the striata (coordinates from Bregma: AP,1.0 mm; LAT, 63.0 mm; DV, 26.0 mm). The right striatumwas infused with anti-c-fos oligodeoxynucleotides (2 µl of 1mM solution at 0.25 µl/minute), whereas the left striatumreceived an equal volume of vehicle. After infusion, thecannulae were left in place for an additional 2 minutes toallow for diffusion away from the injection site. The skinwas then sutured, and the animals were left to recover for1 hour. This postoperative latency had been establishedpreviously as the period of optimal suppressive effects ofASF (Hebb and Robertson, 1997a).

Oligodeoxynucleotides

Oligodeoxynucleotides were purchased from Genosys(Houston, TX) and were single end capped (sulfur modifica-tion on the phosphate group between the first and last basepairs). Briefly, standard phosphoramidite chemistry foroligodeoxynucleotide synthesis was used with samples ofeach synthesis examined on polyacrylamide gels to verifyoligonucleotide quality. All oligodeoxynucleotides were ex-tracted and precipitated by the manufacturer (to removeorganics and salts) prior to lyophilization. Oligodeoxy-nucleotides were reconstituted in ultrafiltered (millipore)distilled water at a concentration of 1 nmol/µl.

The antisense oligodeoxynucleotide to c-fos mRNA, calledASF, was 15 bases in length and had been used previouslyin both partial and complete phosphorothioate forms (Chi-asson et al., 1992, 1994; Dragunow et al., 1993; Heilig etal., 1993; Sommer et al., 1993, 1996; Hooper et al., 1994;Hebb and Robertson, 1997a,b, 1999a,b). Its sequence,58-GsAA-CAT-CAT-GGT-CGsT-38, corresponded to bases129–143 on the mRNA transcript (GenBank accession no.XO6769) and spanned the initiation codon (the subscript‘‘s’’ denotes locations of the sulfur modifications). We havepreviously demonstrated the specificity of end-capped ASFin the striatum by the comparison of its suppressiveactions with the effects of vehicle infusions and two

GABA NEURONS IN THE SNC 31

distinct, random oligodeoxynucleotides (Hebb and Robert-son, 1997a,b, 1999a). We and others also have demon-strated the specificity of the same oligodeoxynucleotidesequence in both partial and complete phosphorothioateforms by using mismatch and sense controls (Chiasson etal., 1992; Heilig et al., 1993; Hooper et al., 1994; Sommer etal., 1996). Unilateral infusions of control oligodeoxynucleo-tides (random, sense, mismatch) or vehicle into the stria-tum produced negligible suppression of stimulant-inducedc-fos expression. Thus, the neurophysiological effects pres-ently described in the ASF group have been attributed tothe specific inhibition of c-fos expression in the striatumand were not produced by nonspecific oligodeoxynucleotideor vehicle infusions.

Immunohistochemistry

After the appropriate recovery periods, animals weregiven intraperitoneal injections of either d-amphetamine(5 mg/kg; 6-OHDAamp and ASF groups) or apomorphine(0.5 mg/kg; 6-OHDAapo group). Two hours after stimulantadministration, animals were deeply anesthetized withsodium pentobarbital (.100 mg/kg) and perfused throughthe left ventricle, initially with saline, followed by 4%paraformaldehyde in a 0.1 M phosphate-buffer solution,pH 7.4. Brains were removed subsequently and postfixedat 4°C until further analysis. Brains were blocked, cut into50-µm coronal sections on a Vibratome, and subsequentlyprocessed for c-Fos, tyrosine hydroxylase (TH), or glutamicacid decarboxylase (GAD; 67 kDa isoform; GAD67) immu-noreactivity.

For immunohistochemistry, the tissue was washed for10 minutes in 0.01 M phosphate-buffered saline (PBS)containing 0.1% Triton X-100 (PBS-TX; except for GADimmunoreactivity; see below). This was followed by a15-minute incubation in 1% hydrogen peroxide to inacti-vate endogenous peroxidase activity and three subsequent10-minute washes in PBS-TX at room temperature. Sec-tions were then incubated in a solution containing apolyclonal antibody to c-Fos (1:5,000; Genosys), TH (1:4,000; Pel-Freez, Rogers, AR), or GAD67 (1:5,000; Chemi-con, Mississauga, Ontario, Canada) for 16–24 hours at4°C. They were then washed three times for 10 minuteseach in PBS-TX and incubated in a 1:500 dilution of anappropriate, biotinylated secondary antibody (Vector Labo-ratories, Burlingame, CA) for 1–2 hours at room tempera-ture. Excess antibody was removed by washing threetimes for 10 minutes each in PBS-TX, and the boundsecondary was visualized by using the avidin-biotin com-plex (ABC) technique (ABC Elite kit; Vector Laboratories)with diaminobenzidine (DAB) as the chromogen. We foundthat addition of detergents markedly reduced the GADimmunoreactivity in cell somata; therefore, all immunohis-tochemical procedures using the GAD67 antibody wereperformed in solutions that did not contain Triton X-100.Processed sections were mounted on gelatin-coated slides,air-dried, dehydrated in graded alcohols, delipidated inxylenes, and coverslipped.

Some tissue sections were double labeled for Fos-likeimmunoreactivity (-LI) and GAD67 by using two-colorDAB immunohistochemistry. Immunolabeling of these twoproteins in the same tissue was technically challenging.Because the c-Fos antibody required the presence ofTriton-X, whereas the GAD67 antibody required detergent-free PBS, simultaneous incubation in a PBS solutioncontaining both c-Fos and GAD67 antibodies was not

feasible. Therefore, we performed complete immunohisto-chemical reactions with the GAD67 antibody by usingconventional (brown) DAB followed by immunohistochem-istry for c-Fos by using nickel-enhanced (blue) DAB, asdescribed above.

Fluorescent double labeling was performed with c-Fos/TH and c-Fos/GAD67 antibody combinations. All incu-bations were performed at 4°C. For the c-Fos/TH combina-tion, sections were incubated in a buffer solution of PBS-TXthat contained both primary antibodies at the dilutionsdescribed above. After a 24–48 hour incubation, the sec-tions were washed three times for 10 minutes each inPBS-TX and subsequently incubated in indocarobocya-nine-2 (Cy2)-conjugated donkey anti-sheep (c-Fos) andCy3-conjugated donkey anti-rabbit (TH) antibodies at adilution of 1:400 (BIO/CAN, Mississauga, Ontario, Canada)for 16–24 hours. Excess antibody was removed by washingthree times for 10 each minutes in PBS-TX. The sectionswere then mounted on gelatin-coated slides and visualizedusing filter sets to detect Cy2 (catalogue no.487710; Zeiss,Thorwood, NY) and Cy3 (catalog no. 487715; Zeiss) immu-nofluorescence.

For the c-Fos/GAD67 antibody combination, the sectionsinitially were incubated in a PBS (without TX) solutioncontaining the GAD67 antibody for 16–24 hours. Theywere then washed three times for 10 minutes each in PBSand incubated with a Cy3-conjugated anti-rabbit antibody,as described above. Sections subsequently were washedand incubated in a PBS-TX solution containing the c-Fosantibody for 16–24 hours. After rinsing, the tissue wasincubated with a Cy2-conjugated anti-sheep antibody (inPBS-TX) for 16–24 hours. After the removal of excessantibody, the fluorochromes were viewed as describedabove.

Selected sections also were stained for Nissl substance.Briefly, 50-µm sections were mounted on gelatin-coatedslides, air-dried, dehydrated in graded alcohols and xy-lenes, rehydrated, and stained with 0.1% cresyl violet.

Retrograde labeling of SNC neurons

To determine whether the Fos-positive cells that wereidentified in the SNC projected to the striatum, retrogradelabeling of nigrostriatal neurons was combined with c-Fosimmunofluorescence. Two naive animals were subjected tostereotaxic surgery, as described above, in which a single25-gauge cannula was inserted into the right striatum(coordinates from Bregma: AP, 1.0 mm; LAT 13.0 mm; DV,26.0 mm), and 0.5 µl of 4% FluoroGold (Fluorochrome Inc.,Englewood, CO) solution was infused at a rate of 0.1µl/minute. After infusion, the cannula was retracted, theskin was sutured, and the animals were left to recover for 5days. After this time, the animals received 5 mg/kgd-amphetamine (i.p.) and were placed in their cages for 2hours, after which they were anesthetized and perfused asdescribed above. Sections were cut through the SNC, andthe tissue was processed for c-Fos immunofluorescence, asdescribed above. FluoroGold fluoresced optimally under awide-band ultraviolet filter. However, because both Fluoro-Gold and the Cy2 fluorochrome fluoresced under the ZeissCy2 filter, we used a Cy3-conjugated secondary antibody tovisualize the c-Fos immunolabeling. There was no overlapbetween the emission wavelengths for FluoroGold and theCy3-conjugated antibody.

32 M.O. HEBB AND H.A. ROBERTSON

Quantification of Fos-positive nuclei inthe SNC

Sections were collected through the SNC of all animalsand processed for Fos-LI, as described above. After immu-nohistochemistry, duplicate sections were obtained fromthe SNC of each animal at the level of the interpeduncularnucleus and were scanned into a computer using a CCDcamera (JVC, Tokyo, Japan) to create digital images. Byusing NIH densitometry software (NIH, Bethesda, MD),both the ipsilateral and contralateral SNC were isolated intheir entire cross section, and the number of Fos-positivenuclei per area was calculated. Any degree of nuclear Foslabeling within a cell that was visible under a lightmicroscope qualified that cell as Fos-positive. The thresh-old of the NIH densitometry software was set to include allvisible nuclei. The number of Fos-positive nuclei percross-sectional area in each animal was calculated as themean of the two sample values. The mean and standarddeviation were calculated for the ipsilateral and contralat-eral SNC of all experimental groups. Within groups, thesevalues were subjected to a Student’s paired t-test todetermine whether there were significant differences inthe Fos-LI between the ipsilateral SNC and the contralat-eral SNC. Significance was assumed at P , 0.05. Computer-generated photomicrographs were assembled into figuresfor publication by using Adobe Photoshop (version 4.0;Adobe Systems, Mountain View, CA) with minimal alter-ations to the contrast and background. The regions shownin the figures represent the areas from which cell countswere obtained.

RESULTS

Extent of 6-OHDA lesions and ASF-inducedc-Fos suppression

TH immunoreactivity. Examination of TH immunore-activity after various postlesion recovery periods revealeda progressive deterioration of DA cells in the substantianigra. Three days after infusion of 6-OHDA into the medialforebrain bundle, there was substantial loss of TH-positivedendritic processes and an overall reduction in the numberof TH-positive cell bodies in the SNC (Fig. 1A,B). After 1week, TH immunoreactivity revealed extensive pyknosisin the SNC. At this time point, all evidence of dendriticarborization was lost, and the SNC appeared as a strip ofdensely clustered granules (Fig. 1C,D). By 3 weeks after6-OHDA infusion, there remained only scant traces of THimmunoreactivity in the substantia nigra, indicating anear total destruction of DAergic cells (Fig. 1E,F).

Nissl staining. Nissl staining of lesioned animalswith 3-week recovery periods revealed extensive loss oflarge- diameter somata in the SNC (Fig. 1G,H). Thisconfirmed that, at 3 weeks after 6-OHDA lesioning, the DAcells of the SNC were destroyed and did not merely reflecta suppression of TH expression. All of the 6-OHDA-lesioned animals used in subsequent analyses were exam-ined at this time point.

Fos immunoreactivity. Infusion of ASF reduced theinduction of Fos-LI in the striatum by approximately 65%(data not shown). Effective ASF-mediated suppression ofc-fos expression in this region has been described previ-ously (Hebb and Robertson 1997a, 1999a,c).

Examination of Fos-LI in striata of 6-OHDAamp ani-mals also revealed a significant reduction in expression on

the ipsilateral side. In contrast, 6-OHDAapo animals hadrobust Fos-LI in the ipsilateral striatum but negligibleinduction on the contralateral side. The effects of 6-OHDAlesioning on stimulant-induced c-Fos expression in thestriatum were consistent with those reported previously(Cenci et al., 1992; Hebb and Robertson, 1999c).

Differential expression of c-fos in the SNC

In naive animals, d-amphetamine (5 mg/kg) inducedbilateral expression of Fos-LI in the central region of theSNC at the level of the interpeduncular nucleus (Fig.2A–C). Animals that received a 6-OHDA lesion followed byapomorphine challenge had a significant reduction of thisexpression in the contralateral SNC, with robust inductionon the ipsilateral side (number of Fos-positive cells: ipsilat-eral, 41.3 6 9.8; contralateral, 11.9 6 1.6; P 5 0.035; Fig.2D–F). In contrast, the 6-OHDAamp (ipsilateral, 6.0 6 1.3;contralateral, 27.0 6 0.96; P , 0.001; Fig. 3A–C) and ASFgroups (ipsilateral, 18.3 6 5.9; contralateral, 35.3 6 9.2;P 5 0.027; Fig. 3D–F) had a near complete elimination ofFos-LI in the ipsilateral SNC, with normal induction onthe contralateral side. The results of the quantification ofFos expression in the SNC of the experimental groups areillustrated in Figure 4.

Characterization of Fos-positive cells inthe SNC

Fos and TH immunohistochemistry. The apomor-phine-induced Fos-LI in 6-OHDA-lesioned SNC, whichexhibited no TH immunoreactivity, indicated that theFos-positive cells were not DA neurons. However, to con-firm this hypothesis, we used double-labeling immunofluo-rescence for Fos-LI and TH in naive animals that had beenchallenged with d-amphetamine. The Fos-positive nucleiwere interspersed between large TH-containing neurons,but the two antigens were never colocalized in the samecells (Fig. 5A,B). Also, to confirm that Fos induction wasnot altered in the SNC after ablation of the DAergicneurons of this region, similar double-labeling experi-ments were performed on 6-OHDAapo animals. Althoughthe TH immunolabeling was eliminated in these animals,apomorphine still produced robust induction of Fos-LI inthe lesioned SNC (Fig. 5C,D).

Fos and GAD immunohistochemistry. Previous re-ports have described the distribution of GAD-positiveneurons in the SNC (Oertel et al., 1982; Yung et al., 1991).We were interested in determining whether the Fos-positive cells in the SNC also expressed GAD, indicating aGABAergic phenotype. We initially performed conven-tional DAB immunohistochemistry for Fos or GAD byusing serial sections through the SNC. Because the Fosantibody worked poorly on 14-µm-thick, slide-mountedsections (not shown), it was necessary to use 50-µm-thick,free-floating sections. This method revealed a high concen-tration of GAD-positive cell bodies in the vicinity of theFos-positive nuclei (Fig. 6).

To confirm Fos and GAD colocalization, we next per-formed sequential, two-color DAB immunohistochemistry.This technique revealed several double-labeled neurons inthe SNC, confirming our hypothesis that the Fos-express-ing cells were GABAergic (Fig. 7). Sequential immunofluo-rescent double labeling of Fos and GAD also demonstratedcolocalization of these proteins (Fig. 8).


Fig. 1. Progression of deterioration in the substantia nigra parscompacta (SNC) after 6-hydroxydopamine (6-OHDA) infusion. Photo-micrographs show tyrosine hydroxylase (TH) immunoreactivity (-IR;A–F) and Nissl staining (G,H) in the SNC, ipsilateral (A,C,E,G) andcontralateral (B,D,F,H) to the infusion of 6-OHDA into the medialforebrain bundle. By 3 days after 6-OHDA infusion, there was markedloss of dopaminergic (DAergic) dendrites in the SNR as well as cellsomata in the SNC (A,B). There was extensive pyknosis and a

complete absence of normal TH-positive somata by 7 days afterlesioning (C,D). At 21 days postlesion, TH-IR was virtually abolishedfrom the ipsilateral SNC (E,F). Nissl staining confirmed the markedloss of large-diameter cell bodies at the 21-day time point (G,H). Thetissue sections shown in E–H were taken from the same animal.Asterisks in G and H indicate corresponding regions of the SNC. Scalebar 5 100 µm.


Fig. 2. Fos-like immunoreactivity (-LI) in the SNC of naive and6-OHDA-apomorphine (apo) animals after psychostimulant challenge.A–C: Amphetamine-induced (amp) Fos-LI was observed bilaterally inthe SNC of naive animals at the level of the interpeduncular nucleus.The arrows in A indicate the corresponding regions shown magnified

in B and C. D–F: Apomorphine induced Fos-LI in the ipsilateral SNC(left), but not in the contralateral SNC (right) SNC of 6-OHDA-lesioned animals. The arrows in D indicate the corresponding regionsshown magnified in panels E and F. Scale bars 5 100 µm (C applies toB, F applies to E).


Fig. 3. Fos-LI in the SNC of the 6-OHDAamp animal group (A–C)and in the SNC of animals that received a unilateral infusion ofantisense oligodeoxynucleotides targeting c-fos mRNA into the stria-tum followed by administration of d-amphetamine (ASF group; D–F)after d-amphetamine challenge. Both groups showed a significant

reduction in Fos expression in the ipsilateral SNC (left arrows in A andD). However, neither treatment affected the induction of Fos-LI in thecontralateral SNC (right arrows in A and D). The arrows in A and Dindicate the corresponding regions shown magnified in B,C and E,F,respectively. Scale bars 5 100 µm (C applies to B, F applies to E).


Retrograde labeling and Fos immunofluorescence.

Five days after tracer injection, the FluoroGold had dif-fused throughout a substantial portion of the striatum.The fluorescence was confined within the striatum, exceptfor a small region of diffusion through the cannulae tractinto the overlying cortex (Fig. 9A). At this time point, therewas extensive retrograde labeling of cells in the SNC,producing a pattern of fluorescence similar to that seenwith TH immunofluorescence in naive animals (Fig. 9B;compare with Fig. 5A). Subsequent immunohistochemis-try revealed that the Fos-positive nuclei in the SNC werenot contained within FluoroGold-labeled neurons but wereinterspersed between these cells (Fig. 5E,F).

DISCUSSION

In this report, we describe the activation of a populationof neurons in the SNC by the DA agonists, d-amphetamineand apomorphine, in intact, 6-OHDA-lesioned, and ASF-infused animals. Nuclear expression of Fos-LI in the SNCwas not present in TH-positive cells but was colocalizedwith cytoplasmic GAD, indicating that these neurons wereGABAergic. In all groups, elimination of the stimulant-induced Fos-LI corresponded to a reduced expression ofc-fos in the striatum, suggesting that efferent striataltransmission had mediated the SNC response.

Because DA neurons of the SNC are active tonically, it isbelieved that the release of DA into the striatum maintainsa resting tone that can be modulated by changes in nigralactivity (Lacey et al., 1989; Yung et al., 1991). Also,dendritic release of DA into the underlying SNR has beenshown to facilitate efferent transmission from this regionand attenuate the inhibitory effects of GABA (Ruffieux andSchultz, 1980; Waszczak and Walters, 1983; Martin andWaszczak, 1996). Thus, the effects of DA in the striatumand in the SNR appear to counteract one another, increas-ing inhibitory tone to the SNR (via activation of directstriatonigral transmission) and facilitating SNR activity,respectively. It is possible that the extent to which each ofthese components affects SNR output is reflected by thelevel of stimulation of SNC neurons.

Although DA regulates the activity of neurons in thestriatum and SNR, stimulation of either of these regionsmarkedly suppresses neuronal activity in the SNC (Drayet al., 1976; Gerfen, 1984; Jimenez-Castellanos and Gray-biel, 1989; Tepper et al., 1995). For example, Tepper et al.(1995) used antidromic stimulation of SNR neurons while

simultaneously recording from the SNC to show thatincreases in reticulata firing rates were associated withdecreased SNC activity. The physiologic role of SNR-SNCinteractions remains unclear. One possibility is that thiscircuit provides a positive feed-back mechanism by whichincreased SNR activity reduces nigrostriatal DAergic trans-mission and, subsequently, decreases the inhibitory inputto the SNR from the striatum. Alternatively, increasedsuppression of DAergic neurons in the SNC would reducethe dendritic release of DA onto SNR terminals anddecrease the facilitory influence of DA on reticulata out-flow. This paradoxical influence of reduced SNC tone onSNR activity may be dependent on the degree of reticulatastimulation. For example, low-level SNC inhibition mayresult only in a decrease of dendritic DA release, increas-ing inhibitory tone to the SNR. Such localized inhibitionmay act as a physiologic barrier that prevents the transmis-sion of unwanted signals from the SNR to downstreamnuclei. Marked increases in SNR activity may overcomethis threshold and potentiate SNR output by reducing therelease of DA into the striatum and decreasing D1R-mediated striatonigral transmission.

Although the majority of studies that have investigatedthe SNC have focused on the neurophysiology of DAergicneurons in this region, several electrophysiological andimmunocytochemical studies have described at least twotypes of SNC neurons (Javoy-Agid et al., 1981; van derKooy et al., 1981; Lacey et al., 1989; Yung et al., 1991).Lacey et al. (1989) described two classes of neurons in theSNC that were distinguished on the basis of their electro-physiological properties. They found that 95% of SNCneurons, which they termed ‘‘principal neurons,’’ hadspontaneous, low-frequency action potentials of relativelylong duration. These neurons were inhibited significantlyby DA or baclofen (a GABA agonist). In contrast, theremaining 5% of neurons (called ‘‘secondary cells’’) exhib-ited rapid, high-frequency action potentials that wereunaffected by DA but were inhibited by baclofen. Yung etal. (1991) also described two populations of SNC neuronsthat were distinguished by their electrophysiological mem-brane properties. Consistent with previous reports, thoseauthors classified these cells as either bursting (15% oftotal cells) or nonbursting (85% of total cells). Double-labeling experiments revealed that all of the nonburstingcells were positive for TH immunoreactivity and, thus,were assumed to be DA neurons. None of the bursting cellswere TH-positive, and their distribution paralleled that ofGAD-positive cells in the SNC. A similar abundance ofGAD immunoreactivity in this region was reported previ-ously by Oertel et al. (1982). Yung and colleagues (1991)suggested that these burster cells were GABAergic neu-rons and may represent a component of the non-DAergicnigrostriatal pathway implicated in previous anatomic(van der Kooy et al., 1981) and electrophysiological (Guy-enet and Aghajanian, 1978) studies. Consistent with thedescription of secondary cells in the SNC (Lacey et al.,1989), it was found that the burster cells in this region alsowere DA-insensitive (Yung et al., 1991).

Striatonigral neurons that project to the SNC are local-ized predominantly to the striosomal compartment of thestriatum and are stimulated by D1R activation (Graybielet al., 1990). The stereotypic effects of d-amphetamine alsoare mediated primarily through D1R activation; conse-quently, the Fos-LI that is induced in the striatum isconfined mainly to striosomes (Graybiel et al., 1990).

Fig. 4. Histogram showing Fos-positive cells counts in the ipsilat-eral and contralateral SNC of the experimental groups. Asterisksindicate significant differences between ipsilateral and contralateralvalues (P , 0.05).


However, there also is substantial D1-mediated stimula-tion after d-amphetamine challenge of striatal neuronsthat project to the SNR. In the present study, administra-tion of d-amphetamine consistently induced Fos-LI inTH-negative, GAD-positive neurons of the central regionof the SNC. The mechanism by which d-amphetamine

induced Fos-LI in the SNC remains unclear. However, theipsilateral elimination of this expression in animals thathad received either a unilateral 6-OHDA lesion or anintrastriatal infusion of ASF indicated that striatal activa-tion was necessary for the SNC response to these stimu-lants. Also, because striatal projection neurons are

Fig. 5. Immunofluorescence and retrograde labeling of SNC cells.The low-magnification photomicrographs in A, C, and E are shown athigher magnification in B, D, and F. A,B: Double labeling for TH (red)and Fos-LI (green) after d-amphetamine treatment in naive animals.The Fos-positive nuclei were interspersed between TH-expressingcells, but the two proteins did not colocalize. C,D: Double labeling forFos-LI (green) and TH (red) after apomorphine challenge in 6-OHDA-lesioned animals. Apomorphine induced normal expression of Fos-LIin the ipsilateral SNC despite total ablation of DA neurons.E,F: Retrograde labeling of nigrostriatal neurons (blue-white) com-

bined with Fos immunofluorescence (red) in the SNC of naive animalsafter d-amphetamine challenge. Striatal infusion of FluoroGold pro-duced extensive labeling of SNC neurons. However, the Fos-LI thatwas induced by d-amphetamine was not colocalized with the tracer.Instead, the appearance of the fluorochromes was similar to thatobserved with Fos and TH immunofluorescence (A,B), such that theFos-positive nuclei were found interspersed between large, FluoroGold-filled neurons. Scale bars 5 100 µm (A applies to C,E; B applies toD,F).


GABAergic, the stimulation of the SNC must have beenproduced indirectly through secondary circuits. Resultsobtained from the 6-OHDAamp and ASF groups suggestedtwo possible mechanisms of SNC stimulation. The firstpossibility (hypothesis 1) is that, in naive animals, DA actsin the SNC to tonically inhibit activation of the GABAergicneurons, i.e., by D2R stimulation. Because d-amphet-amine increases the inhibitory input to the SNC from the

striatum, this influence effectively would reduce the DA-mediated inhibition of GABAergic neurons. Ablation ofDAergic neurons by 6-OHDA would remove permanentlythe inhibition of GABAergic transmission in the SNC,disregulating the activity of these neurons. Because c-Fosis induced in response to changes in neuronal activity, noexpression would be observed in these cells that had beenrendered tonically active. However, even if the DAergic

Fig. 6. Immunoreactivity of Fos and glutamic acid decarboxylase(GAD) in serial sections through the SNC. Low-magnification (A,B),medium-magnification (C,D), and high-magnification (E,F) photomi-crographs demonstrate the expression of Fos-LI (A,C,E) and GAD(B,D,F) in the SNC of a naive animal after d-amphetamine challenge.

The boxed areas in A and B indicate the corresponding regions that areshown magnified in C,E and D,F, respectively. Fos-LI was observed inthe central region of the SNC (arrows in C and D), where a highconcentration of GAD-positive cells was found. Scale bars 5 100 µm (Aapplies to B, C applies to D, E applies to F).


Fig. 7. Double labeling of Fos (dark nuclei) and GAD (light browncytoplasm) in the SNC using two-color diaminobenzidine (DAB)immunohistochemistry. A: Low-magnification photomicrograph show-ing the location of the double-labeled cells. B: Higher magnification ofthe region shown in A. Many double-labeled cells (arrows) were

present as well as GAD-positive cells that did not express Fos-LI(arrowheads). It is unclear whether these latter cells did not expressFos-LI due to a lack of stimulation, the inability to express Fosantigens, or simply insensitive immunohistochemical methods. Notethe abundance of GAD-IR in terminal boutons. Scale bars 5 100 µm.

neurons had remained intact, a similar reduction in stria-tonigral transmission, produced by an alternate mecha-nism, would be expected to suppress the stimulation of theGABAergic neurons in the SNC (by disinhibition of DAactivity). This was confirmed in the ASF group, in whichnigrostriatal DA systems remained intact, and only effer-ent transmission from the striatum was affected.

The second possibility (hypothesis 2) predicts that theGABAergic neurons of the SNC are regulated by GABA-

ergic influences from the SNR. This notion is supported byprevious studies that demonstrated that SNR activityinhibits activity of the SNC (Tepper et al., 1995). Thus, innaive animals, administration of d-amphetamine in-creases the inhibition of the SNR via striatonigral stimula-tion, which reduces the inhibitory tone of the SNR on theSNC, permitting induction of Fos-LI in the SNC. In boththe 6-OHDAamp group and the ASF group, the striatoni-gral influence was reduced drastically (in the ipsilateral

Fig. 8. Double immunofluorescent-labeling of Fos (green) and GAD(red) imaged by using confocal microscopy. Sections of 50 µm thicknesswere examined in incremental planes of 5 µm to reduce obscuring ofsoma labeling by dense GAD-positive fibers. These sections wereobtained from naive animals that had been stimulated withd-amphetamine. A: Consistent with the DAB immunohistochemistry,several double-labeled neurons were apparent in the SNC by using

this method (arrows). Fos-positive nuclei that were labeled lessintensely were evident in adjacent tissue planes (small arrowheads).Also, there appeared to be GAD-positive, Fos-negative cells in thisregion (large arrowhead). The absence of Fos-LI in the GAD-positivecells may have been due to several factors, as noted in the legend toFigure 7. B: High-magnification photomicrograph of a double-labeledneuron in the SNC. Scale bars 5 10 µm.


hemisphere), increasing both SNR outflow and inhibitorytone to the SNC. This second hypothesis is supportedfurther by the previous report that 6-OHDAamp animalshave a significant increase in Fos-LI in the ipsilateral SNRwith a concurrent elimination of the expression in the SNC(Hebb and Robertson, 1999c).

Expression of Fos-LI in the SNC of 6-OHDAapo animalsconflicted with the notion of DA-mediated regulation ofGABAergic neurons (hypothesis 1) and further supportedthe possibility that these cells in the SNC were influencedby SNR activity (hypothesis 2). For example, although it is

possible that apomorphine acted directly at DA receptorsin the SNC to induce Fos-LI, this notion contradictshypothesis 1, which predicts that DA produces inhibitionof these neurons. Furthermore, regions like the striatumthat are subjected to tonic DA influence develop supersen-sitivity to direct DA agonists after depletion of endogenoussources (Ungerstedt, 1971). This phenomenon also hasbeen observed in both the SNR and the entopeduncularnucleus in 6-OHDA-lesioned animals (Hebb and Robert-son, 1999c). Therefore, if the GABAergic neurons in theSNC were directly influenced by DA, then it would be

Fig. 9. Retrograde labeling of nigrostriatal neurons. A: FluoroGoldinjection site in the striatum. Diffusion of the tracer was apparentthroughout most of the striatum, with no penetration into surround-ing brain regions. Dashed line indicates the lateral border of the

striatum. cc, corpus callosum. B: FluoroGold-labeled neurons in theSNC 5 days after tracer injection into the striatum. Scale bars 5100 µm.


expected that a certain degree of sensitization would occurafter DA depletion. However, there was no evidence tosupport this in the current studies. Hypothesis 2, however,was supported by the robust expression of Fos-LI in theipsilateral striatum, but not the contralateral striatum, of6-OHDAapo animals. This was consistent with the previ-ous observation that induction of Fos-LI in the SNCparalleled that in the striatum. These animals also exhib-ited an increase in Fos-LI in the contralateral SNR with noexpression in the contralateral SNC. Thus, it appearedthat, as in the other animal groups, SNC stimulation wasmediated by increased striatonigral activity and a reduc-tion in SNR inhibition of the GABAergic neurons in theSNC.

In their discussion of burster neurons in the SNC, Yunget al. (1991) suggested that these cells may comprise aGABAergic nigrostriatal pathway. From the similarities inelectrophysiological properties, the abundance and loca-tion between burster neurons and the ‘‘secondary cells’’described previously by Lacey et al. (1989), those authorsconcluded that they belonged to the same neuronal sub-population. More recently, Rodriguez and Gonzalez-Hernandez (1999) used immunohistochemistry combinedwith cellular tracing and electrophysiology to characterizea subpopulation of GABAergic neurons in the substantianigra. However, unlike the present findings, the cellsomata in the former study never were found in the SNCand could be filled retrogradely with a tracer that wasapplied to the striatum. Because the Fos-positive cellsidentified here were found consistently in the midst ofretrogradely-labeled neurons but were not labeled them-selves, it is unlikely that these cells projected to thestriatum but failed to take up the cellular tracer. Itappears, therefore, that the GABAergic cells in this studyare distinct from those characterized by Rodriguez andGonzalez-Hernandez (1999). However, the location andabundance of the Fos-positive, GABAergic neurons in theSNC are consistent with earlier reports and suggest thatthese cells are burster/secondary neurons (Lacey et al.,1989; Yung et al., 1991). Also, the demonstration that theburster cells are insensitive to DA but are inhibitedmarkedly by GABA agonists further supports the hypoth-esis that these cells are regulated through SNR transmis-sion and not directly by DAergic colaterals (hypothesis 2).

The GABAergic neurons in the SNC that were identifiedin this study do not appear to comprise a non-DAergic,nigrostriatal circuit. Retrograde labeling with FluoroGoldproduced an abundance of filled neurons throughout theentire extent of the SNC. We did not observe any cells inthis region that were double labeled for Fos-LI and Fluoro-Gold. In fact, the distribution of the Fos-positive nucleiamong the large, FluoroGold-filled neurons was similar tothat observed with double immunofluorescence for Fos andTH. Together with the previously reported similaritiesbetween secondary neurons and interneurons of otherbrain regions (Miles and Wong, 1984; Madison and Nicoll,1988; Lacey et al., 1989; Yung et al., 1991), these resultssuggest that the Fos-positive, GABAergic neurons re-ported in this study may belong to a class of interneurons.The activation of these cells by DA agonists suggests thatthey may play a role in inhibitory mechanisms that reducethe DAergic stimulation of striatal neurons. Whetherthese neurons project to the SNR or only arborize locallyremains unknown. However, results from studies thatdemonstrated a reciprocal relation between SNR activity

and SNC activity argue against a GABAergic interneuro-nal connection between these two regions (Grace andBunney, 1979; Waszczak et al., 1980). Determination of themechanism of stimulation and the function of theseGABAergic neurons will require electrophysiological meth-ods and may provide insight into regulatory processes thatexist between the SNC, the SNR, and the basal ganglia.

ACKNOWLEDGMENT

M.O.H. was supported by a Scholarship from the Hun-tington Society of Canada.

LITERATURE CITED

Cenci MA, Kalen P, Mandel RJ, Wictorin K, Bjorklund A. 1992. Dopaminer-gic transplants normalize amphetamine- and apomorphine-inducedFos expression in the 6-hydroxydopamine-lesioned striatum. Neurosci-ence 46:943–957.

Cheramy A, Leviel V, Glowinski J. 1981. Dendritic release of dopamine inthe substantia nigra. Nature 289:537–542.

Chiasson B, Hooper M, Murphy P, Robertson H. 1992.Antisense oligonucleo-tide eliminates in vivo expression of c-fos in mammalian brain. Eur JPharmacol 227:451–453.

Chiasson B, Armstrong J, Hooper M, Murphy P, Robertson H. 1994. Theapplication of antisense oligonucleotide technology to the brain: somepitfalls. Cell Mol Neurobiol 14:507–521.

Dragunow M, Lawlor P, Chiasson B, Robertson H. 1993. C-fos antisensegenerates apomorphine and amphetamine-induced rotation. Neurore-port 5:305–306.

Dragunow M, Tse C, Glass M, Lawlor P. 1994. C-fos antisense reducesexpression of Krox 24 in rat caudate and neocortex. Cell Mol Neurobiol14:395–405.

Dray A, Gonye TJ, Oakley NR. 1976. Caudate stimulation and substantianigra activity in the rat. J Physiol 259:825–849.

Gerfen CR. 1984. The neostriatal mosaic: compartmentalization of cortico-striatal input and striatonigral output systems. Nature 311:461–464.

Gerfen CR, Engber TM, Mahan LC, Susel Z, Chase TN, Monsma FJ, SibleyDR. 1990. D1 and D2 dopamine receptor regulated gene expression ofstriatonigral and striatopallidal neurons. Science 250:1429–1432.

Grace AA, Bunney BS. 1979. Paradoxical GABA excitation of nigraldopaminergic cells: indirect mediation through reticulata inhibitoryneurons. Eur J Pharmacol 59:211–218.

Graybiel AM. 1990. Neurotransmitters and neuromodulators in the basalganglia. Trends Neurosci 13:244–254.

Graybiel AM. 1991. Basal ganglia—input, neural activity and relation tothe cortex. Curr Opin Neurobiol 1:644–651.

Graybiel AM, Moratalla R, Robertson HA. 1990. Amphetamine and cocaineinduce drug-specific activation of the c-fos gene in striosome-matrix andlimbic subdivisions of the striatum. Proc Natl Acad Sci USA 87:6912–6916.

Guyenet PG, Aghajanian GK. 1978. Antidromic identification of dopaminer-gic and other output neurons of the rat substantia nigra. Brain Res150:69–84.

Hebb MO, Robertson HA. 1997a. End-capped antisense oligodeoxynucleo-tides effectively inhibit gene expression in vivo and offer a low-toxicityalternative to fully modified phosphorothioate oligodeoxynucleotides.Mol Brain Res 47:223–228.

Hebb MO, Robertson HA. 1997b. Coordinate suppression of striatal ngfi-aand c-fos produces locomotor asymmetry and upregulation of IEGs inthe globus pallidus. Mol Brain Res 48:97–106.

Hebb MO, Robertson HA. 1999a. Synergistic influences of the striatum andthe globus pallidus on postural and locomotor control. Neuroscience90:413–421.

Hebb MO, Robertson HA. 1999b. Alterations of neuronal activity in thesuperior colliculus of rotating animals. Neuroscience 90:423–432.

Hebb MO, Robertson HA. 1999c. Motor effects and mapping of cerebralalterations in animal models of Parkinson’s and Huntington’s diseases.J Comp Neurol 410:99–114.

Heilig M, Engel J, Soderpalm B. 1993. C-fos antisense in the nucleusaccumbens blocks the locomotor stimulant action of cocaine. Eur JPharmacol 236:339–340.


Hooper M, Chiasson B, Robertson H. 1994. Infusion into the brain of anantisense oligonucleotide to the immediate-early gene c-fos suppressesproduction of Fos and produces a behavioral effect. Neuroscience63:917–924.

Hornykiewicz O. 1979. Dopamine in Parkinson’s disease and other neuro-logical disturbances. In: Horn AS, Korf J, Westerink BH, editors. Theneurobiology of dopamine. London: Academic Press, p 633–654.

Javoy-Agid F, Ploska A, Agid Y. 1981. Microtopography of TH, GAD andCAT in the substantia nigra and ventral tegmental area of control andParkinsonian brains. J Neurochem 36:1218–1227.

Jimenez-Castellanos J, Graybiel AM. 1989. Compartmental origins ofstriatal efferent projections in the cat. Neuroscience 32:297–321.

Kebabian JW, Calne DB. 1979. Multiple receptors for dopamine. Nature277:93–96.

Lacey MG, Mercuri NB, North RA. 1989. Two cell types in rat substantianigra zona compacta distinguished by membrane properties and theactions of dopamine and opioids. J Neurosci 9:1233–1241.

LeMoine C, Normad E, Guitteny AF, Fouque B, Teoule R, Bloch B. 1990.Dopamine receptor gene expression by enkephalin neurons in ratforebrain. Proc Natl Acad Sci USA 87:230–234.

LeMoine C, Normad E, Bloch B. 1991. Phenotypical characterization of ratstriatal neurons expressing the D1 dopamine receptor gene. Proc NatlAcad Sci USA 88:4205–4209.

Madison DV, Nicoll RA. 1988. Enkephalin hyperpolarizes interneurons inthe rat hippocampus. J Physiol 398:123–130.

Mahan LC, Burch RM, Monsma FJ, Sibley DR. 1990. Expression of striatalD1 dopamine receptors coupled to inositol phosphate production andCa21 mobilization in Xenopus oocytes. Proc Natl Acad Sci USA 87:2186–2200.

Martin LP, Waszczak BL. 1996. Dopamine D2 receptor-mediated modula-tion of the GABAergic inhibition of substantia nigra pars reticulataneurons. Brain Res 729:156–169.

Miles R, Wong RK. 1984. Unitary inhibitory synaptic potentials in theguinea-pig hippocampus in vitro. J Physiol 356:97–113.

Oertel WH, Tappaz ML, Berod A, Mugnaini E. 1982. Two color immunohis-tochemistry for dopamine and GABA neurons in rat substantia nigraand zona incerta. Brain Res Bull 9:463–474.

Paxinos G and Watson C. 1997. The rat brain in stereotaxic coordinates.Academic Press, New York.

Robertson GS, Robertson HA. 1987. D1 and D2 dopamine agonist syner-gism: separate sites of action. Trends Pharmacol Sci 8:295–299.

Robertson HA. 1992. Dopamine receptor interactions: some implications forthe treatment of Parkinson’s disease. Trends Neurosci 15:201–206.

Rodriguez M, Gonzalez-Hernandez T. 1999. Electrophysiological and mor-phological evidence for a GABAergic nigrostriatal pathway. J Neurosci19:4682–4694.

Ruffieux A, Schultz W. 1980. Dopaminergic activation of reticulata neu-rones in the substantia nigra. Nature 285:240–241.

Snyder SH. 1982. Neurotransmitters and CNS disease. schizophrenia.Lancet 2:970–974.

Sommer W, Bjelke B, Ganten D, Fuxe K. 1993. Antisense oligonucleotide toc-fos induces ipsilateral rotational behavior to d-amphetamine. Neurore-port 5:277–280.

Sommer W, Rimondini R, O’Connor W, Hansson AC, Ungerstedt U, Fuxe K.1996. Intrastriatally injected c-fos antisense oligonucleotide interfereswith striatonigral but not striatopallidal g-aminobutyric acid transmis-sion in the conscious rat. Proc Natl Acad Sci USA 93:14134–14139.

Stoof JC, Kebabian JW.1981. Opposing roles for D1 and D2 dopaminereceptors in efflux of cyclic AMP from rat neostriatum. Nature 294:366–368.

Tepper JM, Martin LP, Anderson DR. 1995. GABAA receptor-mediatedinhibition of rat substantia nigra dopaminergic neurons by parsreticulata projection neurons. J Neurosci 15:3092–3103.

Ungerstedt U. 1968. 6-Hydroxydopamine induced degeneration of centralmonoamine neurons. Eur J Pharmacol 5:107–110.

Ungerstedt U. 1971. Postsynaptic supersensitivity after 6-hydroxydopamine-induced degeneration of the nigro-striatal dopamine system in the ratbrain. Acta Physiol Scand 82(Suppl 376):69–93.

van der Kooy D, Coscina DV, Hattori T. 1981. Is there a non-dopaminergicnigrostriatal pathway? Neuroscience 6:345–357.

Waszczak BL, Walters JR. 1983. Dopamine modulation of the effects ofgamma-aminobutyric acid on substantia nigra pars reticulata neurons.Science 220:218–221.

Waszczak BL, Eng N, Walters JR. 1980. Effects of muscimol and picrotoxinon single unit activity of substantia nigra neurons. Brain Res 188:185–197.

Yung WH, Hausser MA, Jack JJ. 1991. Electrophysiology of dopaminergicand non-dopaminergic neurones of the guinea-pig substantia nigra parscompacta in vitro. J Physiol 436:643–667.


identification of a subpopulation of substantia nigra pars compacta ?-aminobutyric acid neurons that...

Documents