Contralateral pallidothalamic and pallidotegmental projections in primates: an anterograde and retrograde labeling study

212

Brain Research, 567 (1991) 212-223 1991 Elsevier Science Publishers B.V. All rights reserved. 0006-8993/91/$03.50

BRES 17256

Contralateral pallidothalamic and pallidotegmental projections in primates: an anterograde and retrograde labeling study Lili-Naz Hazrati and Andr6 Parent Centre de recherche en neurobiologie, Universitd Laval et H~pital de l'Enfant-Jdsus, Qudbec (Canada) (Accepted 30 July 1991)

Key words: Basal ganglia; Internal pallidum; Thalamus; Phaseolus vulgaris-leucoagglutinin; Fluorescent retrograde tracers; Pallidothalamic connection; Pallidotegmental connection; Saimiri sciureus; Reticular thalamic nucleus

Unilateral injections of the anterograde tracer Phaseolus vulgaris-leucoagglutinin (PHA-L) in the internal segment of the pallidum (GPi) of the squirrel monkey (Saimiri sciureus) led to anterograde labeling of fibers ipsilaterally in the following thalamic nuclei: ventral anterior (VA), ventral lateral (VL), centromedian (CM), and lateral habenula (Hbl). The labeled fibers reached these ipsilateral thalamic nuclei by coursing along or through the ansa lenticularis, the lenticular and thalamic fasciculi, and the Forel's fields. They arborized profusely in VA/VL nuclei where they displayed small glomerule-like formations. Numerous labeled fibers also occurred in the CM. Most of them were long, varicose and gave rise to shorter fibers that formed a dense terminal field covering a large portion of the CM. A small but dense terminal field composed of delicate fibers and extremely fine terminals was noted in the Hbl. A large contingent of labeled fibers were seen to cross the midline, principally at the rostral pole of the CM and in the supramammillary decussation, to reach the contralateral thalamus where they arborized profusely in the VA/VL and CM nuclei, but not in the Hbl. The patterns of termination of these contralateral pallidothalamic fibers were strikingly similar to those observed ipsilaterally. Other anterogradely labeled fibers were also noted bilaterally in the pedunculopontine nucleus (TPP) and ipsilaterally in the external segment of the pallidum (GPe) and in the putamen. Complementary, double-labeling, retrograde studies involving the injection of nuclear yellow in the VA/VL and CM nuclei and Fast blue in the TPP, confirmed the existence of contralateral pallidothalamic and pallidotegmental projections. The number of retrogradely labeled cells in the contralateral GPi amounted approximately to 10-20% that in the ipsilateral GPi. These experiments further indicated that contralaterally projecting pallidothalamic neurons exhibited a high degree of axonal collateralization, the majority of its neurons projecting also to the contralateral TPP. Cells retrogradely labeled with the tracer injected into the thalamus were also encountered bilaterally in the thalamic reticular nucleus. Taken together, the results of these anterograde and retrograde investigations indicate that the contralateral pallidothalamic projection involves a relatively small population of GPi neurons, but that these neurons arborize extensively in their contralateral thalamic targets. Furthermore, the presence of retrogradely labeled cells in the ipsi- and contralateral reticular thalamic nucleus indicates that the VA/VL and CM nuclei, which receive a massive input from the GPi, are under the bilateral influence of this perithalamic nucleus. Such contralateral projections could play a major role in the subcortical organization of the bilateral aspect of normal basal ganglia function. They may also serve as an important compensatory mechanism in various pathological conditions affecting the basal ganglia.

INTRODUCTION The internal or medial segment of the pallidum (GPi) in primates, the presumed homologue of the entopeduncular nucleus (ENT) in the non-primates, is a major output nucleus of the basal ganglia. The GPi in primates is known to project to the ventral anterior (VA)/ventral lateral (VL) thalamic nuclei, the centromedian (CM) nucleus, the habenula (Hb) and the pedunculopontine (TPP) nucleus of the midbrain-pontine tegmentum 1'11'12'18'19, The projections to VA/VL, CM and TPP nuclei were shown to arise mostly from axon collaterals of the same GPi neurons, whereas the projection to the H b originated principally from a distinct cell population located peripherally in the GPi 7'2°. In this study we used

the highly sensitive anterograde tracer Phaseolus vulgaris-leucoagglutinin (PHA-L) in order to provide a detailed description of the trajectory and patterns of termination of the pallidofugal fibers arborizing in the thalamus and the brainstem, with a particular emphasis on contralateral projections. We also employed the fluorescence, retrograde, double-labeling technique to evaluate the n u m b e r of GPi neurons giving rise to the ipsi- and contralateral pallidothalamic and paUidotegmental projections. Part of these results were published in a short form elsewhere 22. MATERIALS AND METHODS Two male, adult (body weight of 900 g and 1100 g) squirrel monkeys (Sairniri sciureus) were used in the present study. The first

Correspondence: A. Parent, Centre de recherche en neurobiologie, Hrpital de l'Enfant-Jrsus, 1401, 18e Rue, Qurbec (QC) Canada G1J 1Z4.

213

;co !. VA""::",, "

)

,

:'; . . . . ~,..

u"

,. f -

.'.

,..k '~':..'vP"--.,

,

"':--

A

"""

I,

.. •".

"

i

2,,',

:

.,;

{

.... ~---::,,-.;

....,,

- -

I (5,0A)

N~'.":

X

-:'-,.--. I~, " r.;k

;" """',,

:" ":...~

,,

":. 3>

Ht5,SA)

S''/

"'::"::"

..-..~

'

"

J(4,0A)

.:

i

\""J

Fig. 1. Schematic drawings of transverse sections through the forebrain and upper brainstem of the squirrel monkey to identify the structures of interest in the present study. The drawings are displayed in a rostrocaudal order (from A to K) and the stereotaxic plane of each section (according to the atlas of Emmers and Akert 2) is indicated in the lower left.

animal was anesthetized with ketamine hydrochloride (Ketaset, 40 mg/kg, i.m.) and received unilateral iontophoretic injections of the anterograde tracer PHA-L into the GPi over two needle penetrations. A 2.5% solution of PHA-L (Vector Labs, Burlingame, CA) was prepared by dissolving 5 mg of PHA-L in 200 #l of phosphate buffer (0.01 M, pH 8.0). This solution was loaded in a glass micropipette with a tip diameter of 25-30 l~m and was iontophoretitally injected with a 7-10/~A positive (cathodal) current delivered in 7 s pulses every 14 s over a 20-30 rain period using a constant current generator (Midgard Electronics). The micropipette was attached to a micromanipulator that was stereotaxically driven and the stereotaxic coordinates were chosen according to the atlas of Emmers and Akert 2. Following a survival period of 12 days, the animal was deeply anesthetized with an overdose of pentobarbital and perfused transcardially with 250 ml of phosphate buffer (0.1 M, pH 7.4) containing 10% sucrose followed by 600 ml of a fixative containing 4% paraformaldehyde and 0.1% glutaraldehyde in phos-

phate buffer (0.1 M, pH 7.4), and finally with 400 ml of 4% paraformaldehyde in phosphate buffer (0.1 M, pH 7.4) at 4 °(2. Following perfusion, the head of the animal was placed in the stereotaxic apparatus and, after removal of the cranial vault, the brain was sectioned stereotaxically in 5-10 mm thick transverse slabs that were postfixed for 1 h in 4% paraformaldehyde fixative, and placed for 18-24 h in a phosphate buffer (0.1 M, pH 7.4) containing 30% sucrose at 4 *C. The brain slabs were then sectioned on a freezing microtome at 40/zm in the transverse plane. The sections were serially collected, kept in 0.1 M This-buffered saline (TBS, pH 7.4) at 4 °C, and then processed for the immunohistochemical localization of PHA-L. In addition, some alternate sections taken at forebraln and brainstem levels were stained with Cresyl violet. The presence of PHA-L was revealed immunohistochemically by pre-incubating the sections for 1 h in TBS containing 0.1% TritonX-100 and 2% normal rabbit serum (NRS). They were then incu-

214

\

Fig. 2A-E. (For legend see p. 215).

bated overnight at 4 °C with a goat anti-PHA-L (Vector labs) diluted 1:2000 in TBS containing 0.1% Triton-X-100 and 1% NRS. After this primary incubation, the sections were rinsed 3 times (10 rain each) in TBS, incubated for 1 h at room temperature with biotynilated rabbit anti-goat IgG (Vector Labs) diluted 1:200 in a TBS solution containing 0.1% Triton-X-100 and 1% NRS, rinsed 3 times (for 20 min) in TBS and incubated for 1 h at room temperature in a solution containing the avidin-biotin-peroxidase complex (Vector Labs, 1:100 dilution). After a 30 min rinse in TBS, the bound peroxidase was revealed by placing the sections in a solution containing 3,3"-diaminobenzidine tetrahydrochloride (DAB, 0.05%) and hydrogen peroxide (H202, 0.005%) in TBS buffer (0.05 M, pH 7.6) for 15-20 min at room temperature. The immunostaincd sections were then rinsed in TBS (5 x 5 min) and mounted onto gelatin-coated slides with Permount to be examined with a light microscope under both bright- and darkfield illumination. The second animal was also anesthetized with ketamine hydro-

chloride and received injections of the retrograde fluorescent tracers, Fast blue (FB, 2% solution, Sigma, St Louis) and Nuclear yellow (NY, 2% solution, Dr IUing K.G., E R . G . ) by means of 1 kd Hamilton microsyringes. NY was injected in quantities ranging from 0.4 to 0.6 yl over 2 needle penetrations into the VA/VL complex and the CM, whereas similar quantities of FB were delivered over two needle penetrations into the TPP area. The thalamic injections were made with a vertical approach, whereas a lateral approach was used to reach the pedunculopontine nucleus in order to avoid spillage of the tracer in the superior colliculus2. The injections of FB and NY were made 48 and 18 h before sacrifice, respectively. After the appropriate survival period, the animal was administered an overdose of pentobarbital and perfused with 10% sucrose in phosphate buffer, followed by one liter of a fixative containing 4% paraformaldehyde and 0.1% glutaraldehyde in phosphate buffer, followed by a solution of 4% paraformaldehyde. The brain was stereotaxically cut into transverse slabs, as described above, and

215

I/1

\ ~

.."'-.,,' /

/I v.',

/

I ~ ~ . , , ,--~.,~ .. ,'./,

•

~,.~ ,

.. ~ .

2 " - ~/' x\ w ~ /.- -':,.# : f # ? -f" /

~,

t,.." ...""

\

~

,...~

"-

'~"..~~?,'--"

.

.

.

.

,' i¢

f"... '..

,'7

.... , .-':v, ,.., ~...-' i'_ ", i

"'}'(,

! :"

( \%".

"x

/

I

<'./h

J Fig. 2 (Continued). Series of schematic drawings illustrating the extent of the injection site (stippled areas in B-E) and the distribution of the anterogradely labeled fibers (sinuous lines) in the monkey that received a phaseolus vuigaris-leucoagglutinin (PHA-L) injection into the internal pallidum. The photomicrograph shows the injection site at the level of its maximum mediolateral extent (Bar -- 150 gm).

placed in phosphate buffer containing 30% sucrose for 12 h at 4 °C. The brain slabs were then sectioned on a freezing microtome at 40 /zm in the transverse plane. The sections (1 out of 3) were serially collected in distilled water, mounted on slides without coverslips and air dried. They were examined with a Leitz Ploemopack fluorescence microscope equipped with filter mirror system A (360 nm), which allows the visualization of both NY and FB. Some sections were also stained with Cresyl violet in order to ensure a proper identification of brain structures.

RESULTS

Anterograde labeling experiment Injection sites. I n the first a n i m a l

two major injection

sites in c o n t i n u i t y w i t h o n e a n o t h e r o c c u r r e d in the G P i . T h e first o n e h a d its m a x i m a l e x t e n t at a level corres p o n d i n g to t h e s t e r e o t a x i c p l a n e A 11.5 (Fig. 1), acc o r d i n g to t h e atlas of E m m e r s a n d A k e r t 2. T h i s injection site c o v e r e d m o s t o f t h e c e n t r a l c o r e o f t h e G P i (Fig.

216

Fig. 3. Darkfield photomicrographs showing examples of anterogradely labeled fibers and terminal arborization observed ipsilaterally after PHA-L injection into the internal pallidum. A,B: terminal fields displaying glomerule-like features in the ventral lateral thalamic nucleus. C: numerous thin fibers forming a dense terminal field in the centromedian nucleus. The arrow indicates the habenulo-interpeduncular tract. D: fibers approaching the midline (on the right) en route to the contralateral thalamus. E: typical terminal field in the lateral habenular nucleus. F: weakly arborized fibers in the putamen. Bar = 200/~m (A), 360/zm (B-E), 100/tm (F).

2B,C). The second injection site had its maximal extent at stereotaxic level A 10 (Fig. 1) and was slightly larger than the first injection site (Fig. 2D,E). lpsilateral fiber labeling. Ipsilateral to the P H A - L injections, numerous anterogradely labeled fibers emerged from the injection sites in the GPi and coursed along the ventral and dorsal surfaces of the pallidal complex, that is either along the ansa lenticularis rostrally and the lenticular fasciculus caudally, en route to the thalamus. These fibers met at the medioventral tip of the GPi and

continued medially at the basis of the internal capsule. Rostrally, the fibers reached the thalamus mostly by coursing along the ansa lenticularis and arborized principally within the rostral portion of the VA nucleus (Fig. 2C,D). More caudally, they formed several fascicles that pierced the internal capsule and accumulated into the lenticular fasciculus where they formed a rather compact bundle (Fit) (Fig. 2E). This bundle broke-off into two different fiber systems. The first one curved laterodorsally and swept rostraUy along the thalamic fasciculus to

217

Fig. 4. Darkfield photomicrographs showing examples of anterogradely labeled fibers and terminal arborization observed contralaterally after PHA-L injection into the internal pallidum. A: photomontage depicting the numerous labeled fibers in the centromedian nucleus whose contour is delimited by a dotted line. B: higher power view of the decussating fibers as seen near the midline (left), before they reach the centromedian nucleus. C: terminal field displaying giomerule-like features in the ventral lateral thalamic nucleus. Bar = 0.5 mm (A) and 200 ~,m (B,C).

reach the caudal portion of the V A nucleus and the oral part of the V L nucleus (VLo). The second one was composed of several fascicles that separated from the main bundle (Fit) in a straight caudodorsomedial course to reach the CM, the V L and, less abundantly, the ventral most portion of the lateroposterior (LP) thalamic nuclei. Caudal to the Flt, numerous labeled fibers accumulated within the Forel's field H, and from there a large portion of the labeled fibers could be followed up to the caudal portions of CM and V L (Fig. 2H). A t the level of the caudal pole of CM, a smaller contingent of fibers ran

dorsocaudally close to the midline and reached the habenula (Fig. 2I,J). Other labeled fibers left the area of Forel's fields, swept caudally, and descended within the central portion of the midbrain tegmentum. These fibers were seen to reach the TPP area where most of them appeared to terminate (Fig. 2J,K). In addition, some labeled fibers ran into the cerebral peduncle and contributed to the innervation of the TPP. A significant number of anterogradely labeled fibers were also encountered within the GPe and the striatum after GPi injections (Fig. 2 A - F ) . The fibers were particularly numerous in the dorsal two-thirds of the GPe and

218

e

F Fig. 5. Schematicdrawings illustrating the Nuclear yellow (NY) injection sites in the thalamus (stippled areas) and the Fast blue (FB) injection site in the brainstem (hatched areas), as well as the distribution of retrogradely labeled cells in the basal forebrain and upper brainstem. The open circles indicate cells retrogradely labeled with the tracer (NY) injected in the thalamus, the filled circles indicate cells retrogradely labeled with the tracer (FB) delivered in the brainstem, and the asterisks represent doubly labeled cells (NY/FB). in the peripallidal zone of the putamen. No labeled fibers were found in the caudate nucleus after GPi injections. Patterns of terminal labeling. The pallidothalamic fibers displayed different patterns of arborization within each target structure. They displayed typical glomerulelike plexuses in the VA, VLo, LP and VL nuclei (Fig. 3A,B). These glomeruli occurred throughout the rostrocaudal extent of the VA, but abounded particularly in the dorsolateral sector of the nucleus. Similar plexuses were encountered in the dorsolateral region of the VLo and in much of the remaining portion of the VL, except for an area near the medullary lamina, which was devoid of such plexuses but contained numerous thin and rather

unbranched fibers heading for the dorsolateral portion of the VA/VL. Only a few glomerule-like terminal fields were noted in the LP, and most of them occurred at the rather fuzzy junction zone between LP and VL. In the CM most labeled fibers were long, varicose and directed mostly mediolaterally. These long and coarse fibers were intermingled with shorter and thinner fibers that gave rise to a rather dense terminal field covering a large portion of the CM (Figs. 2H and 3C). By comparison, the parafascicular nucleus (Pf) was virtually devoid of labeled fibers and terminals. In the dorsal portion of CM, several long and coarse fibers were seen to continue their course laterally beyond the limits of CM to reach

219

Fig. 6. Darkfield photomicrographs showing examples of single and doubly labeled cells disclosed after the injection of NY in the thalamus and FB in the brainstem. A: doubly labeled ceils in ipsilateral internal segment of the globus pallidus (GPi). B: doubly labeled cells in contralateral GPi. C: single (FB)-labeled cell in contralateral GPi. D: single (NY)-labeled cells in contralateral GPi. E: single (NY)-labeled cell in contralateral reticular thalamic nucleus. Bar = 30/~m (valid for all figures).

the VL nucleus. In the habenula the labeling appeared as a small but dense terminal field within the lateral portion of the lateral habenular nucleus (Hbl) (Figs. 2I,J and 3E). The labeled fibers that terminated in the GPe displayed long intervaricose segments proximally and branched rather frequently distally. At their distal ends, single labeled fibers were often in close apposition with several GPe neurons. Most of the positive fibers reaching the putamen were long, thin and poorly branched (Fig. 3F). By comparison to the numerous and highly ramified fibers that occurred in the thalamus after GPi injections, the labeled fibers in the TPP area were fewer and relatively unbranched. Most of these fibers exhibited a rather short and sinuous course. Contralateral fiber labeling. In addition to the innervation that occurred ipsilaterally to the injection sites, numerous labeled fibers were seen to cross the midline at both thalamic and brainstem levels. The majority of these

decussating fibers crossed the midline at the rostral pole of the CM and in the supramammillary decussation (Figs. 2F, G and 4B). Other fibers decussated more caudally in the core of the midbrain tegmentum and in the posterior commissure (Fig. 2I-K). Several fibers crossing at the CM level terminated within the contralateral CM, where their pattern of arborization was similar to that seen in the ipsilateral CM (Fig. 4A). It consisted of thin, long and varicose fibers directed mediolateraUy and giving rise to a dense terminal field composed of shorter fibers and axon terminals that occurred in the CM but completely avoided the Pf. Some coarse fibers were also noted among the thinner fibers in the dorsal portion of the CM. These fibers continued their course laterodorsally to reach the VA (Fig. 2E), the VLo (Fig. 2F), and the remaining portion of the VL (Fig. 2G). They formed glomerule-like plexuses identical to those found in the thalamus ipsilaterally (Fig. 4C). However, by comparison to the ipsilateral side, the contralateral pallidal pro-

220 CORTEX

CORTEX

o#

l

oo Z

,,,+

I

t

i

l

THALAMUS

I

I! R

f

I I ! I (

Fig. 7. Diagram illustrating the putative bilateral influence of the pallidothalamic input upon the ipsilateral and contralateral thalamocortical neurons (large circles). The interactions between the external pallidum, the reticular thalamic nucleus and the thalamocortical neurons have also been included in the diagram. The question marks indicate our lack of knowledge regarding the degree of axonal eollateralization of some of these pathways.

jection to the thalamus appeared more prominent in the VL than in the VA nuclei (Fig. 2 E - G ) . At brainstem levels some relatively short, sinuous and nonvaricose labeled fibers were found in the contralateral TPP area (Fig. 2K). No labeling was detected in the habenula, neither in the GPe and the striatum contralaterally. The exact contribution to the innervation of the thalamus and/or the brainstem of the few fibers crossing the midline at the level of the supramammillary decussation could not be established in the present material.

Retrograde labeling experiment Injection sites. In the animal used for the retrograde labeling experiments, the NY injection sites in the thalamus involved the dorsal two-thirds of the VA/VL and the central portion of the CM (Fig. 5C-E). These injection loci did not encroach upon the adjacent reticular nucleus and did not spread into the contralateral side. The FB injection site in the TPP area covered most of the rostrocaudal extent of the structure. The tip of cannula was centered upon the brachium conjunctivum, but the tracer had diffused within most of the dorsoventral extent of the TPP (Fig. 5F). Retrograde cell labeling. Retrogradely labeled cells were encountered bilaterally in the GPi, the substantia nigra, the reticular thalamic nucleus, the GPe and the zona incerta. In the ipsilateral GPi, the majority (70-

80%) of labeled cells contained both fluorescent tracers and these doubly labeled cells were distributed throughout the core of the structure (Figs. 5 and 6A). Single pallidothalamic cells were found to be particularly abundant dorsolaterally, whereas the single pallidotegmental cells appeared more numerous ventromedially in the GPi. In the contralateral GPi, the number of retrogradely labeled cells was approximately 10-20% that found in the ipsilateral GPi. The pallidotegmental cells were more numerous than the pallidothalamic cells, and some doubly labeled also occurred in the contralateral GPi (Figs. 5 and 6B-D). The retrogradely labeled cells in the reticular nucleus of thalamus contained NY only and were present bilaterally throughout most of the rostrocaudal and dorsoventral extent of the structure (Figs. 5 and 6E). In the zona incerta NY-labeled cells abounded rostrally, whereas FB-labeled cells predominated caudally. The bilateral labeling in the substantia nigra was confined to the pars reticulata (SNr), and the majority of the labeled cells contained the tracer (FB) injected into the TPP (Fig. 5E). In the substantia nigra the number of labeled cells was much greater ipsilaterally than contralaterally. In the reticular nucleus, the retrogradely labeled cells were also more numerous ipsilaterally although the number of contralateral NY-labeled cells was impressive (Fig.

5).

221 DISCUSSION In primates, studies with retrograde fluorescent tracers or with horseradish peroxidase (HRP) have revealed the exact cellular origin of the pallidothalamic projections 3'2°'21. These studies showed that neurons projecting to VA/VL nuclei occurred in the core of the GPi along its entire rostrocaudal extent, whereas those projecting to the CM formed two more or less continuous clusters lying in the ventrocaudal portion of the GPi. The pallidal neurons projecting to the TPP were also present in the core of the GPi, whereas those projecting to the habenula were located more peripherally and abounded particularly at the ill-defined junction between the GPi and the lateral hypothalamus 19'2°. Furthermore, antidromic invasion studies 7's and fluorescence retrograde doublelabeling investigations 2°'2~ have revealed that GPi neurons in primates display a high degree of axonal collateralization. In fact, the majority of pallidal neurons were found to send axon collaterals to either the VA/VL and the TPP, the VA/VL and the CM, or to the 3 of them 7"8'2°'21. In contrast, the pallidohabenular projection was found to arise predominantly from a distinct neuronal population. In rats and cats the pallidohabenular projection was reported to arise from paUidal neurons other than those projecting to the VA/VL, the CM and the TPP 4'26, but this projection in non-primates appeared much more prominent than that in primates ~6'26. As in primates, however, the pallidal neurons in the entopeduncular nucleus (ENT) of rats and cats that project to the VA/VL, the CM or the TPP were highly collateralized 4'26. The data obtained in the present study with the retrograde double-labeling technique confirm the high degree of collateralization of the pallidothalamic and pallidotegmental projections in primates. The first allusion to a crossed pallidothalamic projection was made by Nauta in his autoradiographic tracing studies of the pallidal projections in the cat 16'17. This investigator reported a very sparse contralateral projection attributable to a few ENT labeled axons, which crossed the midline in the supramammillary decussation and to a lesser extent in the massa intermedia, to sparsely innervate the contralateral CM and habenula ~6'17. Labeled fibers crossing the midline in the supramammiUary decussation were also observed in another radioautographic study of the ENT efferents in the cat 13, but these fibers were apparently too sparse and could not be followed further on the contralateral side. However, a more recent retrograde tracing study in the ecat has clearly revealed retrogradely labeled cells in the contralateral ENT after HRP injections into either the VA/VL or the CM 15. In primates, the existence of a contralateral pallido-

centromedian projection was categorically denied in a recent retrograde (HRP) study of the CM afferents from the GPi in macaque monkeys 3. Likewise, previous anterograde labeling studies of the pallidothalamic projections in monkeys made no mention of a contralateral contribution n'12'18, except one in which the occurrence of a sparse anterograde autoradiographic labeling in the contralateral habenula was noted 1. The present investigation provides evidence from PHA-L anterograde labeling experiments for the existence of a prominent crossed pallidothalamic and pallidotegmental projection in primates. The contralateral pallidothalamic projection was found to be much more profusely arborized than the pallidotegmental projection. However, the patterns of termination of these pallidofugal fibers in the VA/VL, the CM and the TPP were strikingly similar to those observed ipsilateraUy. The existence of such crossed pallidothalamic and pallidotegmental projections was confirmed by retrograde doublelabeling experiments, which revealed the presence in the GPi of neurons projecting contralaterally to the thalamus (VA/VL-CM) or the brainstem (TPP), and even to both thalamus and brainstem via axon collaterals. In contrast to the profuse contralateral labeling found in the VA/VL and the CM, no anterogradely PHA-Llabeled fibers or terminals were noted in the contralateral habenula and this projection has not been investigated with retrograde tracer in the present study. However, it must be recalled that the PHA-L injection sites were partial and covered mostly the central portion of the GPi. Since the majority of pallidal neurons projecting to the Hbl are located peripherally, it is likely that only a small portion of pallidohabenular neurons have taken up the PHA-L, which would explain the sparse labeling observed in ipsilateral Hbl. This point needs further investigation with retrograde labeling and more extensive injections of anterograde tracers. These future studies should not rule out the possibility that the contralateral pallidohabenular projection may originate from a distinct cell population in the GPi. Interestingly, the present retrograde labeling experiments have revealed that the number of labeled cells in the contralateral GPi amounted only to 10-20% that in the ipsilateral GPi. In the contralateral GPi only one third of all retrogradely labeled neurons contained the two tracers and the number of neurons containing the tracer delivered in the TPP were more numerous than those labeled with the tracer injected in the thalamus. These data confirmed the existence of a bilateral pallidotegmental projection, as demonstrated in a previous study 2°. The results also provide retrograde labeling evidence of a contralateral pallidothalamic projection in primates, and establish the existence of GPi neurons

222 projecting contralaterally to both the thalamus and the brainstem. On the other hand, the results of the anterograde and retrograde labeling experiments indicate that the contralateral pallidothalamic projection involves relatively few GPi neurons, but that these neurons arborize extensively in their contralateral thalamic targets. The percentages of retrogradely labeled cells mentioned above must obviously be taken with caution since they are based on data obtained in a single animal. However, these values are strikingly similar to those reported in earlier fluorescence retrograde double-labeling studies of the pallidal efferents in the same species, but involving several cases of thalamic and brainstem injections 2°'21. In these studies, 70-75% of pallidal neurons in ipsilateral GPi were found to be doubly labeled after thalamus and brainstem injections. Contralaterally, GPi neurons containing the tracer injected into the brainstem amounted to 15-20% those in ipsilateral GPi, whereas no pallidal neurons were found to contain the tracer injected into the thalamus. This absence of pallidal neurons retrogradely labeled after thalamic injection in the contralateral GPi may be explained by the difference in the sensitivity of the tracers used in these earlier studies, which were Evans blue (EB) and a mixture of DAPI-primuline

(DP). Also worth noting is the presence of a certain amount of NY labeling of the glial cells surrounding some retrogradely labeled neurons in both ipsi- and contralateral GPi (Fig. 6). This glial labeling indicates that NY may leak out of labeled neurons and hence contribute to some of the neuronal labeling observed at pallidal level. However, we believe that this possibility is very unlikely because the proportions of single and double-labeled neurons observed in the present study are strikingly similar to those obtained in previous double-labeling studies with EB and DP, which are two tracers that fluoresce at different wavelengths and do not leak out of retrogradely labeled neurons 2°'2x. This study has also revealed the presence of retrogradely labeled neurons in the ipsi- and contralateral reticular thalamic nucleus after thalamic injections. This finding indicated that these reticular thalamic neurons may influence the activity of the thalamic nuclei receiving a projection from the ipsilateral GPi. In this perspective, it is worth recalling that a direct projection from the GPe to the reticular thalamic nucleus has been documented in a previous study 9. By virtue of its projection to the reticular thalamic nucleus, as well as to several basal ganglia components, it was therefore proposed to consider the GPe as a control structure of the output of the basal ganglia 9. The reticular nucleus is known to exert a powerful inhibitory action upon neurons of most thalamic nuclei by using y-aminobutyric acid (GABA) as a neu-

rotransmitter 2s. Since GPe neurons projecting to the reticular nucleus are also known to be GABAergic 24, the GPe is thus in a position to disinhibit the thalamocortical neurons including those receiving inputs from the GPi and the other major output structure of the basal ganglia, namely the substantia nigra pars reticulata (SNr). It may therefore be postulated that the GPe also influences the contralateral VA/VL and CM via the crossed projection of the reticular thalamic nucleus (Fig. 7). In addition, numerous labeled fibers were seen to terminate in the GPe after PHA-L injection in the GPi (Fig. 2A-F). These fibers displayed long intervaricose segments proximally and branch rather frequently distally making contact along their course with several GPe neurons. This pattern of terminal arborization suggests that single GPi axons arborizing within the GPe can influence several GPe cells in a rather diffuse manner by comparison to the more prominent cell-to-cell type of relationship displayed by GPe axons in the GPP °. A comparison of the patterns of innervation observed in the GPe and the GPi suggests that GPe neurons exert a strong inhibitory action upon GPi cells, whereas the GPi to GPe projection may act as a more diffuse feedback inhibitory mechanism 1°. The existence of a direct projection from the ENT to the globus pallidus (homologue of the primate GPe) has also been recently documented in the rat s . Although the exact role of the contralateral pallidofugal projections in the normal functioning of the basal ganglia is not known, it is worth noting that the other major output structure of the basal ganglia, the SNr also gives rise to a significant contralateral thalamic input 6. These contralateral projections could play a very important compensatory role in cases involving unilateral basal ganglia lesions, such as in hemiparkinsonism. Indeed, some changes in G A B A receptors and glucose metabolism have been noted in some contralateral basal ganglia components in monkeys rendered hemiparkinsonian by the unilateral intracarotid injection of the neurotoxin 1-methyl-4-phenyl1,2,3,6-tetrahydropyridine (MPTP) 14'23. These contralateral changes have been tentatively explained by complex loop systems involving the basal ganglia-thalamo-cortical circuit, and particularly by the bilateral corticostriatal projections. However, we believe that the more direct bilateral pallidothalamic and/or pallidotegmental projections demonstrated in the present study should be taken into account in the interpretation of these contralateral changes.

Acknowledgements. The authors are grateful to Carole Harvey and Lisette Bertrand for their technical assistanceand to Suzanne Bilodeau for typing the manuscript. We also thank Louise Bertrand for the artwork. This research was supported by Grant MT-5781 of the Medical Research Council of Canada to A.P. The financial support of

223 FRSQ and FCAR is also acknowledged, L.-N.H. was the recipient of

a Studentship from the FCAR.

ABBREVIATIONS

MD MM PC Pf PUT R SN St TO TPP VA VL VLo VP ZI

CA CD CM DBC Fit GL GPe GPi Hbl Hbm HI IC LP

anterior commissure caudate nucleus centromedian thalamic nucleus decussation of brachium conjunctivum lenticular fasciculns lateral geniculate body external segment of the globus pallidus internal segment of the globus pallidus habenula, lateral nucleus habenula, medial nucleus habenulo-interpeduncular tract internal capsule lateral posterior thalamic nucleus

REFERENCES 1 DeVito, J.L. and Anderson, M.E., An autoradiographic study of efferent connections of the globns pallidns in Macaca mulatta, Exp. Brain Res., 46 (1982) 107-117. 2 Emmers, E. and Akert, K., A Stereotaxic Atlas of the Brain of the Squirrel Monkey (Saimiri sciureus), The University of Wisconsin Press, Madison, 1963. 3 F6nelon, G., Francois, C., Percheron, G. and Yelnik, J., Topographic distribution of paUidal neurons projecting to the thalamns in macaques, Brain Research, 520 (1990) 27-35. 4 Filion, M. and Harnois, C., A comparison of projections of entopeduncular neurons to the thalamus, the midbrain and the habenula in the cat, J. Comp. Neurol., 181 (1978) 763-780. 5 Fink-Jensen, A. and Mikkelsen, J.S., A direct neuronal projection from the entopeduncular nucleus to the globns pallidus. A PHA-L anterograde tracing study in the rat, Brain Research, 542 (1991) 175-179. 6 Gerfen, C.R., Staines, W.A., Arbuthnott, G.W. and Fibiger, H.C., Crossed connections of the substantia nigra in the rat, J. Comp. Neurol., 207 (1982) 283-303. 7 Harnois, C. and Filion, M., Pallidal neurons branching to the thalamus and to the midbrain in the monkey, Brain Research, 186 (1980) 222-225. 8 Harnois, C. and Filion, M., Pallidofugal projections to thalamus and midbrain: a quantitative antidromic activation study in monkeys and cats, Exp. Brain Res., 47 (1982) 277-285. 9 Hazrati, L.-N. and Parent, A., Projection from the external pallidum to the reticular thalamic nucleus in the squirrel monkey, Brain Research, 550 (1991) 142-146. 10 Hazrati, L.-N., Parent, A., Mitchell, S. and Haber, S.N., Evidence for interconnections between the two segments of the globns pallidus in primates: a PHA-L anterograde tracing study, Brain Research, 533 (1990) 171-175. 11 Kim, R., Nakano, K., Jayaraman, A. and Carpenter, M.B., Projections of the globns pallidus and adjacent structures: an autoradiographic study in the monkey, J. Comp. Neurol., 169 (1976) 263-290. 12 Kuo, J.-S. and Carpenter, M.B., Organization of pallidothalamic projections in the Rhesus monkey, J. Comp. Neurol., 151 (1973) 201-236. 13 Larsen, K.D. and McBride, R.L., The organization of feline entopeduncular nucleus projections: anatomical studies, J. Comp. Neurol., 184 (1979) 293-308. 14 Mitchell, I.J., Clarke, C.E., Boyce, S., Robertson, R.G., Peggs,

15 16 17 18 19 20

21 22

23

24

25

26

mediodorsal thalamic nucleus mammillary body cerebral peduncle parafascicular thalamic nucleus putamen reticular thalamic nucleus substantia nigra subthalamic nucleus optic tract pedunculopontine nucleus ventral anterior thalamic nucleus ventral lateral thalamic nucleus oral part of VL ventral posterior thalamic nucleus zona incerta

D., Sambrook, M.A. and Crossman, A.R., Neural mechanisms underlying parkinsonian symptoms based upon regional uptake of 2-deoxyglucose in monkeys exposed to 1-methyi-4-phenyl1,2,3,6-tetrahydropyridine, Neuroscience, 32 (1989) 213-226. Nakano, K., Kohno, M., Kawahira, J. and Tokushige, A., Entopeduncular nucleus projections to the contralateral thalamic nuclei: an HRP study, Brain Research, 262 (1983) 283-287. Nauta, H.J.W., Evidence of a paUidohabenular pathway in the cat, Y. Comp. Neurol., 156 (1974) 19-28. Nauta, H.J.W., Projections of the pallidal complex: an autoradiographic study in the cat, Neuroscience, 4 (1979) 1853-1873. Nauta, W.J.H. and Mehler, W.R., Projections of the lentiform nucleus in the monkey, Brain Research, 1 (1966) 3-42. Parent, A., Identification of the pallidal and peripallidal cells projecting to the habenula in monkey, Neurosci. Lett., 15 (1979) 159-164. Parent, A. and DeBellefeuille, L., Organization of efferent projections from the internal segment of giobns paUidns in primate as revealed by fluorescence retrograde labeling method, Brain Research, 245 (1982) 201-213. Parent, A. and DeBellefeuille, L., The pallidointralaminar and pallidonigral projections in primate as studied by retrograde double-labeling method, Brain Research, 278 (1983) 11-27. Parent, A., Hazrati, L.-N. and Lavoie, B., The paUidum as a dual structure in primates. In G. Bernardi, M.B. Carpenter, G. Di Chiara, M. Morelli and E Stanzione (Eds.), Basal Ganglia, III, Plenum, New York, 1991, pp. 81-88. Robertson, R.G., Clarke, C.A., Boyce, S., Sambrook, M.A. and Crossman, A.R., The role of striatopallidal neurones utilizing y-aminobutyric acid in the pathophysiology of MPTPinduced parkinsonism in the primate: evidence from [3I-I]flunitrazepan autoradiography, Brain Research, 531 (1991) 95-104. Smith, Y., Parent, A., S6gu61a, P. and Descarries, L., Distribution of GABA immunoreactive neurons in the basal ganglia of the squirrel monkey (Saimiri sciureus), J. Comp. Neurol., 259 (1987) 50-65. Steriade, M., Parent, A. and Hada, J., Thalamic projections of nucleus reticularis thalami of cat: a study using retrograde transport of horseradish peroxidase and fluorescent tracers, J. Comp. NeuroL, 229 (1984) 531-547. Van der Kooy, D. and Carter, D.A., The organization of the efferent projections and striatal afferents of the entopeduncular nucleus and adjacent areas in the rat, Brain Research, 211 (1981) 15-36.