Thalamic input to parvalbumin-immunoreactive GABAergic interneurons: organization in normal striatum and effect of neonatal decortication

Pergamon

PII:

Neuroscience Vol. 88, No. 4, pp. 1165–1175, 1999 Copyright  1998 IBRO. Published by Elsevier Science Ltd Printed in Great Britain. All rights reserved 0306–4522/99 $19.00+0.00 S0306-4522(98)00265-6

THALAMIC INPUT TO PARVALBUMIN-IMMUNOREACTIVE GABAERGIC INTERNEURONS: ORGANIZATION IN NORMAL STRIATUM AND EFFECT OF NEONATAL DECORTICATION T. M. RUDKIN and A. F. SADIKOT* Department of Neurology and Neurosurgery, Montreal Neurological Institute, McGill University, 3801 University Street, Montreal, Canada, H3A 2B4 Abstract––The neocortex and thalamus send dense glutaminergic projections to the neostriatum. The neocortex makes synaptic contact with spines of striatal projection neurons, and also targets a distinct class of GABAergic interneurons immunoreactive for the calcium-binding protein parvalbumin. We determined whether the parafascicular thalamic nucleus also targets striatal parvalbumin-immunoreactive interneurons. The anterograde tracer biotinylated dextranamine was injected into the parafascicular nucleus of adult rats. Double-labeled histochemistry/immunohistochemistry revealed overlapping thalamic fibers and parvalbumin-immunoreactive neurons in the neostriatum. Areas of overlap within the sensorimotor striatum were analysed by electron microscopy. Of 311 synaptic boutons originating from the parafascicular nucleus, 75.9% synapsed with unlabeled dendrites, 22.5% with unlabeled spines, and 1.3% had parvalbumin-immunoreactive dendrites as a postsynaptic target. Only 4% of all asymmetric synapses on parvalbumin-immunoreactive dendrites were derived from the parafascicular nucleus. A separate group of animals underwent bilateral neocortical deafferentation on the third postnatal day, prior to injection of anterograde tracer into the parafascicular nucleus of adult animals. These experiments were performed with the dual purpose of (i) reducing the possibility that thalamic inputs to parvalbuminimmunoreactive neurons are the result of transsynaptic uptake of tracer by a thalamo-cortico-striatal route, and (ii) determining whether competitive interactions between developing corticostriatal and thalamostriatal fibers may account for the relatively sparse thalamic input onto parvalbuminimmunoreactive interneurons. In decorticates, 219 striatal synaptic contacts derived from the parafascicular nucleus, out of which 77.2% were on unlabeled dendrites, 20.9% were upon unlabeled spines, and 0.9% targeted parvalbumin-immunoreactive dendrites. We conclude that the thalamic parafascicular nucleus indeed sends synaptic input to parvalbuminimmunoreactive striatal neurons. Parafascicular nucleus inputs to striatal parvalbumin-immunoreactive interneurons are sparse in comparison to other asymmetric inputs, most of which are likely to be of cortical origin. The synaptic profile of thalamostriatal inputs to parvalbumin-immunoreactive neurons and unlabeled elements is unchanged following neonatal decortication. This suggests that competitive interaction between developing thalamostriatal and corticostriatal projections is not a major mechanism determining synaptic input to striatal subpopulations.  1998 IBRO. Published by Elsevier Science Ltd. Key words: thalamostriatal, parvalbumin, ultrastructure, synaptogenesis, development, plasticity.

The mammalian neostriatum receives afferents from a wide variety of sources including the cerebral cortex, the thalamus, and brainstem monoaminergic systems.16,20,45,55 The cerebral cortex11,12,18,29,32,38,58 and intralaminar thalamic nuclei3,6,10,13,21,29,32,48 50–53,56 send dense projections to the neostriatum in the form of glutaminergic asymmetric synapses. In the rat, the sensorimotor cortex and the parafascicular thalamic nucleus (Pf) each direct their input to the dorsolateral quadrant of the striatum, and preferen*To whom correspondence should be addressed. Abbreviations: ABC, avidin–biotin–peroxidase complex; BDA, biotinylated dextranamine; BSA, bovine serum albumin; DAB, diaminobenzidine; NiDAB, nickelintensified diaminobenzidine; PBS, phosphate-buffered saline; Pf, parafascicular thalamic nucleus; TMB, 3,3 ,5,5 -tetramethylbenzidine.

tially innervate the matrix compartment of the chemically heterogeneous striatal mosaic.20,21 Despite overlap at the light-microscopic level, cortical and thalamic projections to the neostriatum show distinct synaptic interactions. Cortical fibers synapse mainly with heads of spines on striatal medium spiny projection neurons, whereas thalamic terminals preferentially synapse with dendritic shafts and to a lesser extent with spines.6,13,29,36,53,55,56,67 One study suggests that cortical and thalamic afferents may even preferentially terminate upon different subpopulations of medium spiny neurons.13 It also appears that cortical or thalamic afferents have complementary input upon subclasses of striatal interneurons. For example, cholinergic interneurons in the ventral41,42 and dorsal11,36 striatum receive only sparse input from cortical fibers. On the other hand,

1165

1166

T. M. Rudkin and A. F. Sadikot

cholinergic interneurons in both striatal areas receive robust glutaminergic input of thalamic origin.36,41,42 Another population of striatal interneurons, which is immunoreactive for somatostatin and neuropeptide Y, receives input from the cortex but not the thalamus.25,64 To establish whether this segregation of afferent inputs exists for other populations of striatal neurons, we focus on synaptic input to a class of fast firing aspiny GABAergic interneurons identified by their immunoreactivity for the calcium-binding protein parvalbumin.2,5,8,15,27,34 Thus far, it has been established that parvalbumin-immunoreactive interneurons in the dorsal striatum receive inputs from the cerebral cortex,2,35 and those in the ventral striatum receive inputs from the amygdala.54 Glutaminergic input to inhibitory striatal parvalbumin interneurons may ultimately result in feed-forward modulation of striatal projection neurons.2,27,35,47,54 Other sources of potential excitatory inputs to parvalbumin neurons have not been studied. We first determined whether terminals from the Pf nucleus, the main source of thalamic afferents to the rat neostriatum, synapse upon parvalbuminimmunoreactive striatal interneurons. We also performed decortication experiments to ensure that thalamic input to parvalbumin neurons is not an artifact of trans-synaptic transport of anterograde tracer via a Pf-corticostriatal route.4,39,44,48 These experiments determined whether depriving the developing neostriatum of cortical input would alter the sparse Pf input to parvalbumin interneurons. Description of the thalamostriatal profile after decortication prior to the period of corticostriatal synaptogenesis,19,40,43,60 allowed us to comment on mechanisms by which developing striatal afferents come to acquire a well-ordered relationship with their postsynaptic targets. EXPERIMENTAL PROCEDURES

Stereotactic injections Thirty-five to 45-day-old Sprague–Dawley rats (Charles River, LaSalle, Quebec, Canada) were anesthetized with a combination of ketamine hydrochloride (10 mg/kg, i.p.) and xylazine (5–10 mg/kg, i.p.). Unilateral stereotactic injections of 0.5 µl of 10% biotinylated dextranamine (BDA, Molecular Probes, Eugene, OR, U.S.A.)49,63,66 in 0.01 M phosphate buffer (pH 7.2) were then made into the parafascicular thalamic nucleus using a 2-µl Hamilton syringe. BDA histochemistry allows detailed visualization of anterogradely labeled fibers and terminals at both the light microscopic49 and the electron microscopic54,56 level. Following a six to 14 day survival period, rats were given a lethal dose of ketamine/xylazine and perfused transcardially. An initial rinse with 100 ml heparinized saline (4C) was followed by 20–30 min perfusion using a fixative solution containing 4% paraformaldehyde and 0.5% glutaraldehyde in phosphate buffer (0.1 M, pH 7.4, 4C). Brains were removed, postfixed for 1 h, sectioned at 60 µm in the coronal plane on a vibrating microtome, and collected in phosphate-buffered saline (PBS, 0.1 M, pH 7.4). The core of the BDA injection site and the labeled neuropil was revealed by incubating sections overnight at

room temperature in a PBS solution containing 1% avidin– biotin–peroxidase complex (ABC, Vector Laboratories, Burlingame, CA, U.S.A.), 0.1% Triton X-100 and 1% bovine serum albumin (BSA, Sigma, St Louis, MO, U.S.A.). Labeling was revealed by exposing the sections for 10–15 min in a solution containing 3,3 -diaminobenzidine (DAB, 0.025%, Sigma), imidazole (1 ml/100 ml, Fisher, Montreal, Canada), and hydrogen peroxide (0.006%) dissolved in Tris buffer (0.05 M, pH 7.6). At the level of the injection site, adjacent sections were processed for acetylcholinesterase histochemistry to delineate the acetylcholinesterase-dense Pf from the surrounding thalamus.14,52 Decortications Sprague–Dawley rats were bilaterally decorticated on the third postnatal day (P2). Neonates were anesthetized by using methoxyfluorane and hypothermia until they were immobile. The scalp and the frontal and parietal bones were excised under a surgical microscope. Areas of the frontal and parietal cortices, comprising the sensorimotor central areas and immediately adjacent cortical areas,69 were removed by gentle aspiration using a glass micropipette. Care was taken not to extend the ventral aspect of the lesion beyond the subcortical white matter or the lining of the ependyma of the lateral ventricle. The skull flap was replaced and the scalp closed using glue. The rats were warmed for a few minutes under a heating lamp and were then returned to their mother. The rats were allowed to survive for 35–45 days, and then received stereotactic injections of BDA into the Pf using the atlas of Paxinos and Watson as a guide.46 The rats were allowed to survive for one to two additional weeks, and perfused under deep anesthesia as described above. Histochemistry/immunohistochemistry Light microscopy. Striatal sections were initially treated for 20 min in sodium borohydride (1% in PBS), and then rinsed extensively in PBS. Terminal labeling containing BDA was revealed histochemically by incubating overnight in a PBS solution containing 1% ABC, 0.1% Triton X-100 and 1% BSA. Sections were rinsed in PBS (3, 5 min/ wash), and then exposed to DAB as a chromogen as described above. Sections for BDA-parvalbumin doublelabeling were then preincubated for 1 h at room temperature in a PBS solution containing 0.1% Triton X-100 and 2% BSA, and then incubated overnight at room temperature in a PBS solution containing antibody to parvalbumin (Sigma; 1:5000), 0.1% Triton X-100, and 1% BSA. Sections were washed in PBS (3, 5 min/wash), and then incubated for 1 h in a PBS solution containing biotinylated anti-mouse IgG (1:200), 0.1% Triton X-100 and 1% BSA. Sections were rinsed in PBS (3, 5 min/wash), and incubated for 1 h in PBS containing 1% ABC, 0.1% Triton X-100 and 1% BSA. Following rinses in PBS (3, 5 min/wash), immunoreactivity was revealed using a nickel-intensified DAB (NiDAB) procedure,65 by exposing to Tris buffer (0.05 M, pH 7.6) containing nickel ammonium sulphate (0.37%, Fisher), DAB (0.025%), and hydrogen peroxide (0.0006%), yielding a blue-black reaction product. Sections were thoroughly rinsed in PBS, mounted out of distilled water on to gelatincoated slides, air dried, dehydrated, and coverslipped with Permount (Fisher). Electron microscopy. Striatal sections were initially treated for 20 min in sodium borohydride (1% in PBS), and then rinsed extensively in PBS. Following the sodium borohydride treatment, sections selected for electron microscopy were cryoprotected, freeze-thawed for 20 min at
80C, passed through a graded series of diluted cryoprotectant solutions, and then washed twice in PBS. Terminal labeling was revealed histochemically by incubating overnight in 1% ABC dissolved in PBS and containing 1% BSA. DAB was

Thalamic projection on to parvalbumin-immunoreactive neurons used as a chromogen as described above. Sections were then incubated in a solution containing the primary antibody (anti-parvalbumin) as described above, except that Triton X-100 was excluded from the prepared solutions. The parvalbumin-immunoreactive elements were revealed using an ammonium tungstate-stabilized 3,3 ,5,5 tetramethylbenzidine (TMB) reaction product.37 Sections were then postfixed in a 1% solution of osmium tetroxide in phosphate buffer (0.1 M, pH 6.0). To improve the contrast under the electron microscope, sections were incubated initially in 50% ethyl alcohol and then immersed in the dark for 35 min in 1% uranyl acetate dissolved in 70% ethyl alcohol. The sections were then further dehydrated in increasing concentrations of ethyl alcohol, immersed in propylene oxide, and left overnight embedded in resin (Durcupan ACM, Fluka, Switzerland). Sections were then mounted on greased microscope slides, coverslipped, and cured for 48 h at 60C. Structures labeled with TMB could easily be distinguished from DAB-labeled elements. Tetramethylbenzidine forms crystalline deposits whereas DAB forms an amorphous electron-dense reaction product (Figs 3, 4). Tetramethylbenzidine-labeled neurons, dendrites, axons and terminals were observed in both experimental and control material. Some sections were processed for only DAB or TMB in order to verify specificity of labeling. In the case of DAB labeling, non-specific labeling was easily distinguished since the specific reaction product was much more electron dense, was confined to axons or terminal boutons, and often showed synaptic vesicles in terminal varicosities (Fig. 3A). To further control for non-specific labeling, some sections were processed using TMB as a chromogen for BDA histochemistry, and DAB was used as a chromogen for parvalbumin immunohistochemistry. Areas of overlap between thalamic afferent terminals and parvalbumin-immunoreactive elements in the dorsolateral caudate–putamen were selected under the light microscope. The coverslips were removed with a razor blade, and the sections were cut out from the slides and mounted onto resin blocks. Ultrathin sections were cut on an ultramicrotome (Reichert-Jung Ultracut E, Austria), and collected onto either Pioloform-coated single-slot copper grids or G-100 hexagonal copper grids (Electron Microscopy Sciences, Fort Washington, PA, U.S.A.). The grids were stained with lead citrate and examined using a JEOL-100 CX II (JEOL, Tokyo, Japan) transmission electron microscope. In order to avoid counting the same element at different levels, only one section in 10 was scanned for quantification of data. Synaptic interactions between DAB-labeled thalamic boutons and postsynaptic elements were recorded and photographs taken of both control and decorticated material. In unlesioned control material, an additional series of analyses were performed to quantify unlabeled asymmetric synapses on parvalbumin neurons. Neuropil fields which contained both DAB and TMB labeling, and in which thalamic afferents were observed to form at least 100 synaptic specializations were selected, and the number of asymmetric specializations made upon parvalbumin-immunoreactive elements was recorded. Asymmetric synaptic interactions were selected for analysis when presynaptic and postsynaptic membranes were preserved, there was an unequivocal postsynaptic density, and the presynaptic bouton contained vesicles and a presynaptic density. The following criteria were used to select animals for electron microscopic analysis: (1) the core of the BDA injection site in both normal and lesioned animals corresponded to the acetylcholinesterase-rich Pf nucleus. (2) after bilateral decortication, the lesion ipsilateral to the Pf injection did not invade the dorsal striatum. (3) tissue was well-preserved at the ultrastructural level. (4) the electron-dense deposits of TMB and DAB were readily distinguished from each other and from unlabeled structures.

1167

RESULTS

Three adult controls and two lesioned rats fulfilled the selection criteria and were included in the study. Camera lucida drawings depicting the cortical lesions are illustrated in Fig. 3. Injection sites Labeling of the BDA injection site was dense in both control and lesioned animals, and included filling of neurons and surrounding processes. In the animals selected for further histochemical/ immunohistochemical processing, the injection site filled cells within the pars lateralis of the Pf, lateral to the fasciculus retroflexus (Fig. 1A), and in all but one animal, extended medially to include the pars medialis of the Pf. In all cases, small portions of the BDA injection extended dorsally and laterally to include areas immediately surrounding the Pf, with additional minor contamination of the overlying hippocampus due to the injection tract. In no case was there retrograde labeling of cells in the substantia nigra pars reticulata. In control animals, cells in the cerebral cortex did not show retrograde BDAlabeling. Light microscopy Thalamostriatal projection. Following injections of BDA into the Pf, labeled fibers in both control and decorticated tissue sections were distributed throughout the ipsilateral neostriatum. Varicose fibers were found in high density in the dorsolateral striatum, interrupted by patches of light to absent labeling (Fig. 1B). Decorticate animals had ablation of frontal agranular cortex and parietal areas of both hemispheres. Lesions that were ipsilateral to the hemisphere injected with anterograde tracer did not extend into the striatum (Fig. 2). The density of varicose thalamostriatal fibre labeling in the dorsolateral striatum was similar in control and decorticate animals (Fig. 1C, D). Parvalbumin-immunoreactive neuron and neuropil density was comparable in unlesioned and decorticate animals, and distributed in both the patch and the matrix compartments of the neostriatum. The parvalbumin-immunoreactive neurons varied from fusiform to multipolar, and had long, branching, varicose dendrites (Fig. 1E, F). Parvalbumin-immunoreactive neurons showed their highest densities in the dorsolateral striatum, and areas of overlap between dense thalamostriatal varicosities and parvalbumin-immunoreactive neurons were easily identified in this area (Fig. 1C–F). Electron microscopy Character of diaminobenzidine and tetramethylbenzidine labeling. In all the material, DAB-labeled

1168

T. M. Rudkin and A. F. Sadikot

Fig. 1. (A) Photomicrograph of a BDA injection site in the Pf nucleus. The injection involves the bulk of the Pf which is lateral to the fasciculus retroflexus (fr), and also involves the medial part of Pf. (B) Photomicrograph of double-stained section using DAB for histochemical revelation of BDA and NiDAB as an immunostain for parvalbumin. Dense brown DAB-stained BDA-labeling of neuropil in the dorsolateral neostriatum after injection of anterograde tracer in Pf, overlapping with blue-black NiDAB-labeled parvalbumin neurons. Note areas (asterisks) that are devoid of DAB labeling, probably corresponding to patch/striosomes. (C) Medium-power photomicrograph of brown DAB-containing BDA-labeled varicose fibers in the dorsolateral striatum in an area containing blue-black NiDAB-labeled parvalbumin neurons. This is representative of areas used for electron microscopic analysis. (D) Similar area to that shown in C, but after neonatal decortication. By light microscopy, the character of Pf-striatal labeling and areas of overlap between BDA and parvalbumin neurons are similar in control and decorticate animals. (E, F) High-power photomicrographs (E, 400; F, 1000, oil immersion) of areas containing varicose brown DAB-containing thalamostriatal fibers overlapping with collections of blueblack NiDAB-labeled striatal parvalbumin neurons. Arrowheads in E and F denote branching dendrites of parvalbumin neurons which were well-delineated by immunohistochemistry. Scale bars: (A, B)=200 µm, (C, D)=100 µm, (E)=50 µm, (F)=20 µm.

Thalamic projection on to parvalbumin-immunoreactive neurons

1169

Fig. 2. (A) Somatosensory and motor cortical areas (stippled regions) in a rat brain, modified from the stereotactic atlas of Paxinos and Watson.46 These areas include frontal agranular cortex and parietal areas.69 The coronal levels are represented with reference to the distance from the bregma. (B) Camera lucida drawings of coronal sections from the adult rat brains after neonatal decortication. The lesioned areas comprise the stippled areas shown in A.

thalamostriatal elements were well-differentiated from the parvalbumin-immunoreactive structures containing TMB deposits.37 The DAB-labeled neuropil included myelinated and unmyelinated axons, and synaptic terminals. Thalamostriatal boutons were variable in size and were observed to contain round synaptic vesicles and occasionally contained one or more mitochondria. The peroxidase reaction product was typically associated with the surface of cell organelles and the inner surface of the plasmalemma. DAB-labeled synaptic terminals formed asymmetric membrane specializations (Figs 3, 4). Parvalbumin-immunoreactive structures were labeled with the crystalline TMB-reaction product (Figs 3, 4A). TMB-labeled neuropil contained dendrites, cell bodies, myelinated and unmyelinated axons, as well as synaptic terminals. A technical limitation of using TMB as a reaction product for electron microscopic analysis is that the crystalline deposit can mask the postsynaptic density, sometimes making it difficult to determine if adjacent cells are forming synaptic interactions. There were however many examples of parvalbumin-immunoreactive dendrites receiving unequivocal inputs in the form of both symmetric and asymmetric synapses (Figs 3, 4A).

Character of labeled and unlabeled asymmetric inputs to parvalbumin-immunoreactive neurons. At the light microscopic level, anterogradely-labeled thalamostriatal fibres and boutons were distinguishable from the parvalbumin-immunoreactive structures by the light brown or blue-black colour of their reaction products, respectively (Fig. 1C–F). For electron microscopic analysis, we selected areas of dense overlap between DAB-labeled thalamostriatal fibres and TMB-labeled parvalbumin neurons in the dorsolateral neostriatum at the coronal level where the anterior commissure decussates. This area contained relatively high concentrations of parvalbumin interneurons, and a dense Pf-striatal projection (Fig. 1B–D). Synaptic relationships between DAB-labeled thalamic boutons, and TMB-labeled postsynaptic elements or unlabeled postsynaptic elements were studied. Only superficial ultrathin sections were used for analysis to ensure that both DAB and TMB reaction products were present in the same tissue sections. Three hundred and eleven asymmetric synapses formed by thalamic boutons were found in sections from unlesioned animals. Of these, 236 (75.9%) occurred on unlabeled dendrites, 70 (22.5%) synapsed with unlabeled spines, and 4 (1.3%) formed synapses with TMB-labeled parvalbumin-immunoreactive

1170

T. M. Rudkin and A. F. Sadikot

Fig. 3. High power electron micrographs of thalamic terminals in the dorsolateral striatum of normal adult rats. The terminals were anterogradely labeled using BDA injected into Pf and revealed with DAB. Parvalbumin-immunoreactive elements were revealed using TMB. (A) Asymmetric synaptic specialization (arrow) between DAB-labeled thalamostriatal bouton and parvalbumin-immunoreactive dendrite. The dendrite (d) contains crystalline TMB reaction product. (B, C) Asymmetric synapses (arrows) between Pf terminals and an unlabeled dendrite (‘‘d’’ in B) or unlabeled spine (‘‘s’’ in C). (D) An unlabeled terminal bouton (‘‘b’’) forming an asymmetric synapse with both a parvalbumin-immunoreactive dendrite containing crystalline TMB reaction product (asterisk) and an unlabeled spine (‘‘s’’). The parvalbuminimmunoreactive dendrite also receives input from an additional unlabeled terminal bouton (‘‘c’’). Scale bars=2 µm.

dendrites (Fig. 3). In decorticate animals, we observed 219 asymmetric synapses formed by thalamic boutons. Of these, 169 (77.2%) were on unlabeled dendrites, 42 (21.9%) occurred on unlabeled spines, and 2 (0.9%) synapses were seen on TMB-labeled parvalbumin-immunoreactive dendrites (Fig. 4). In a separate analysis, 68 parvalbuminimmunoreactive dendrites received 100 asymmetric

synaptic contacts. Only four of these asymmetric synapses contained DAB deposits, indicating that they were derived from the Pf. The rest were unlabeled boutons many of which were small and contained round synaptic vesicles. Within this area of analysis, the thalamic innervation was dense, containing a total of 114 identified synapses between Pf terminals and unlabeled dendrites, spines or parvalbumin-immunoreactive elements.

Thalamic projection on to parvalbumin-immunoreactive neurons

1171

Fig. 4. High-power electron micrographs of DAB-labeled terminal boutons in the adult dorsolateral striatum of neonatally decorticated rats. The terminals were anterogradely labeled with BDA after Pf injection and form asymmetric specializations (arrowheads) with: (A) a parvalbumin-immunoreactive dendrite (‘‘d’’) containing crystalline TMB reaction product (asterisk), (B) an unlabeled dendrite (‘‘d’’), and (C, D) postsynaptic structures that are likely unlabeled spines (‘‘s’’). Note also a second DAB-labeled terminal bouton in B (asterisk). Scale bars=1 µm. DISCUSSION

We demonstrate asymmetric synaptic input arising from the rat Pf thalamic nucleus to parvalbuminimmunoreactive interneurons in the dorsolateral neostriatum. In comparison to other asymmetric inputs to parvalbumin neurons, which likely arise mainly from the cerebral cortex,35 Pf inputs appear anatomically minor. After large injections of anterograde tracer into Pf, only 4% of asymmetric inputs to striatal parvalbumin neurons can be identified as arising from the thalamus. This synaptic input to parvalbumin neurons persists despite corticostriatal denervation prior to injection of the anterograde

tracer, excluding artifactual input mediated by a trans-synaptic Pf-corticostriatal route. Early corticostriatal denervation does not increase the frequency of thalamic synapses onto parvalbuminimmunoreactive striatal interneurons, suggesting that thalamostriatal inputs to parvalbumin neurons may develop as a result of tropic mechanisms, rather than on a purely competitive basis. Thalamic input to unlabeled elements We show that in the normal rat striatum, 75.9% of striatal synapses arising from Pf terminate on to unlabeled dendrites and 22.5% target unlabeled

1172

T. M. Rudkin and A. F. Sadikot

spines. These results are in broad agreement with other studies in the rat,13,36,67 cat6 and primate,53,56 which demonstrate that dendritic shafts of medium spiny projection neurons are the major postsynaptic targets of the Pf nucleus (centromedian– parafascicular complex in cats and primates). The proportions demonstrated here are very similar to those described in the rat,36 using an anterograde tracer to label thalamostriatal fibers. However, our results differ from other rat studies which have reported either a higher innervation onto dendrites,67 or no innervation onto spines.13 The discrepancy may be due in part to methodological differences between studies. In one study, spared striatum was examined following neurotoxic lesions.67 The other study used anterograde degeneration to identify thalamostriatal fibers.13 Our results are in agreement with studies that used anterograde tracers to identify thalamostriatal synapses.36,53,56 Thalamic elements

input

to

parvalbumin-immunoreactive

The present study indicates that parvalbuminimmunoreactive elements account for approximately 1.3% of all thalamostriatal targets. This is not surprising since it is estimated that parvalbumin interneurons account for only 3–5% of all striatal neurons.8,31 However, in comparison to the large number of unlabeled asymmetric synapses upon parvalbumin interneurons, synapses of thalamic origin are proportionately minor. Limitations of such an analysis include: (1) parvalbumin-immunoreactive products may not identify the complete extent of the dendritic tree.2,31,68 Many asymmetric synapses are formed on the smallest observable parvalbuminimmunoreactive dendrites,2,31 indicating that more distal dendrites not detected by the antibody may be postsynaptic targets. (2) Although the BDA injections into Pf are large, it is unlikely that all Pf-striatal fibers are labeled with the anterograde tracer. Relative proportions of postsynaptic targets of Pf would be altered in the unlikely event that different regions within Pf provide differential inputs to striatal projection neurons or parvalbumin-immunoreactive cells. Functional implications The thalamus exerts a direct excitatory influence on striatal output by virtue of synaptic input to dendrites and spines of medium spiny neurons.6,13,29,36,52,53,56,67 Our results suggest that the Pf also indirectly influences striatal output by means of a projection to GABAergic parvalbuminimmunoreactive interneurons. Since medium spiny neurons are major targets of parvalbumin interneurons,2 the net effect of thalamic activation of parvalbumin interneurons may be feed-forward inhibition of projection neurons. Furthermore, since

parvalbumin-immunoreactive dendrites are connected by gap junctions,31 even an anatomically minor innervation from the thalamus might result in a significant inhibitory influence upon striatal output. Of all asymmetric synapses to parvalbumin neurons in a striatal area dense in Pf terminals, 96% arise from unlabeled sources. Asymmetric synapses within the striatum may arise from the cerebral cortex, other areas of the thalamus, the amygdala,1,24,28,30,54 the hippocampal formation1,17,42,62 or serotonergic cells of the median raphe nuclei.57 The amygdala and hippocampus project mainly to the ventral striatum, and are not the likely sources of unlabeled asymmetric synapses with parvalbumin interneurons. Serotonergic inputs to the dorsolateral matrix compartment are relatively sparse compared to inputs from the thalamus and cortex, and only a minority of varicosities show membrane specializations.57 Other intralaminar nuclei besides Pf remain possible sources of input. However, since the intralaminar nuclei project to the striatum in a topographical manner, we would not expect other nuclei to contribute a large proportion of unlabeled asymmetric synapses in areas rich in Pf terminals. Thus, the cerebral cortex is the likely source of the bulk of unlabeled asymmetric inputs to striatal parvalbumin neurons.35 Only 4% of presynaptic boutons that form asymmetric synapses onto parvalbumin-immunoreactive dendrites in the dorsolateral striatum are DABlabeled, and therefore derived from the Pf. Furthermore, Pf inputs to parvalbumin-immunoreactive dendrites account for little more than 1% of the entire thalamostriatal projection. Together, these results suggest that parvalbumin-immunoreactive interneurons indeed receive synapses from Pf, but this influence is minor in comparison to the number of synapses received from other excitatory sources. One may theorize that the striatum distinguishes between the glutaminergic excitatory influences from the cortex and thalamus by segregating synaptic afferents to separate cell populations. Such a theory is supported by evidence suggesting that striatal cholinergic interneurons in the rat receive robust afferents from the thalamus, but only sparse or absent afferents from the cerebral cortex.11,36 Conversely, in the case of parvalbumin and somatostatinergic striatal interneurons, cortical afferents predominate over thalamic inputs.25,35,64 Selective glutaminergic inputs to different populations of interneurons also occur in the ventral striatum. For example, in the nucleus accumbens, the glutaminergic amygdalostriatal projection synapses upon parvalbumin-immunoreactive, but not calretinin-immunoreactive, GABAergic interneurons.54 The complementary nature of inputs to interneurons arising from the cortex or thalamus is especially relevant to an emerging view of distinct functional roles played by thalamostriatal and corticostriatal projections. For example, the Pf-striatal

Thalamic projection on to parvalbumin-immunoreactive neurons

projection, but not the corticostriatal projection, is implicated in N-methyl--aspartate mediated D1 dopamine receptor-dependent control of acetylcholine release in the striatum.7 This data correlates well with anatomical evidence suggesting that cortical input to striatal cholinergic interneurons is sparse compared to thalamostriatal input.36 Our present results predict that the corticostriatal projection should participate in parvalbumin interneuronmediated feed-forward inhibition of striatal output neurons to a greater extent than the thalamostriatal projection.

1173

aptic ordering of striatal afferents. As yet unidentified chemotropic mechanisms may facilitate thalamostriatal synapses with appropriate targets, and chemorepulsive factors may prevent thalamic terminals from establishing synapses with putative targets of the corticostriatal projection.26,61 Alternatively, filopodia from developing thalamostriatal sprouts may form transient appositions with inappropriate striatal dendritic filopodia, but stable synapses do not occur due to lack of necessary trophic factors.26,61

CONCLUSIONS

Postsynaptic targets of the thalamostriatal projection after neonatal decortication We performed cortical ablation prior to thalamic injections in order to exclude trans-synaptic transport of the tracer which could yield artifactual thalamic input to parvalbumin neurons. Since thalamic input to parvalbumin neurons persists despite cortical ablation, the relatively minor projection is not an artifact resulting from thalamo-cortico-striatal transsynaptic labeling. The observation that subsets of striatal interneurons receive distinct inputs from the cortex or the striatum raises important developmental questions about the mechanism by which glutaminergic afferents find their appropriate targets. Despite massive early cortical denervation, there is lack of a dramatic compensatory increase in Pf input to parvalbumin interneuron dendrites. Although the experimental design does not permit speculation on quantitative changes in thalamostriatal synaptic density,22,59 our study also suggests that the proportion of Pf-derived synapses upon unlabeled dendrites or spines of projection neurons is very similar in adult rats with or without neonatal decortication. Lesioninduced compensatory synaptogenesis9,22,23,33 involving other striatal afferents may indeed occur after cortical deafferentation, but does not appear to alter the character of Pf targets. Our results suggest that competition is not the sole factor involved in generation of well-organized syn-

The Pf nucleus of the thalamus provides massive excitatory input to dendrites and spines of GABAergic medium spiny projection neurons of the striatum. It also provides input to parvalbumin-immunoreactive GABAergic interneurons. However, in comparison to other excitatory asymmetric inputs to parvalbumin interneurons, Pf input is sparse. This suggests that the excitatory corticostriatal projection, but not the thalamostriatal projection, is the principal participant in feed-forward inhibition of striatal projection neurons by parvalbumin interneurons. Differences in synaptic organization provide structural evidence for distinct functional roles played by thalamic and cortical inputs to the striatum. The distinctive relationship between these two massive glutaminergic inputs to the striatum also appears to hold during synaptogenesis. Apparently, specific non-competitive developmental mechanisms determine striatal targets of the thalamus or cortex. Acknowledgements—We thank Dr Marie-Claude Be´langer and Mr Jim Dixon for excellent technical and research assistance. We also thank Dr Yoland Smith at Emory University, Atlanta for useful discussions. This work was supported by research grants from the Medical Research Council (MRC) of Canada, the American Association of Neurological Surgeons, and the Parkinson Foundation of Canada. T.M.R. is supported by a studentship from the Fonds de Recherche en Sante´ du Que´bec (FRSQ), and A.F.S. is a Scholar of the MRC of Canada.

REFERENCES

1. 2. 3. 4. 5. 6. 7.

Aylward R. L. and Totterdell S. (1993) Neurons in ventral subiculum, amygdala, and entorhinal cortex which project to the nucleus accumbens and their input from somatostatin-immunoreactive boutons. J. chem. Neuroanat. 6, 31–42. Bennett B. D. and Bolam J. P. (1994) Synaptic input and output of parvalbumin-immunoreactive neurons in the neostriatum of the rat. Neuroscience 62, 707–719. Berendse H. W. and Groenewegen H. J. (1990) Organization of the thalamostriatal projections in the rat, with special emphasis on the ventral striatum. J. comp. Neurol. 299, 187–228. Berendse H. W. and Groenewegen H. J. (1991) Restricted cortical termination fields of the midline and intralaminar thalamic nuclei in the rat. Neuroscience 42, 73–102. Bolam J. P., Clark D. J., Smith A. D. and Somogyi P. (1983) A type of aspiny neuron in the rat neostriatum accumulates (3H-)GABA: combination of Golgi-staining, autoradiography and electron microscopy. J. comp. Neurol. 213, 121–134. Chung J. W., Hassler R. and Wagner A. (1977) Degeneration of two of nine types of synapses in the putamen after centre median coagulation on the cat. Expl Brain Res. 28, 345–361. Consolo S., Baronio P., Guidi G. and Di Chiara G. (1996) Role of the parafascicular thalamic nucleus and N-methyl--aspartate transmission in the D1-dependent control of in vivo acetylcholine release in rat striatum. Neuroscience 71, 157–165.

1174 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 40.

T. M. Rudkin and A. F. Sadikot Cowan R. L., Wilson C. J., Emson P. C. and Heizmann C. W. (1990) Parvalbumin-containing GABAergic interneurons in the rat neostriatum. J. comp. Neurol. 302, 197–205. Descarries L., Soghomonian J. J., Garcia S., Doucet G. and Bruno J. P. (1992) Ultrastructural analysis of the serotonin hyperinnervation in adult rat neostriatum following neonatal dopamine denervation with 6-hydroxydopamine. Brain Res. 569, 1–13. Descheˆnes M., Bourassa J., Doan V. D. and Parent A. (1996) A single-cell study of the axonal projections arising from the posterior intralaminar thalamic nuclei in the rat. Eur. J. Neurosci. 8, 329–343. Dimova R., Vuillet J., Nieoullon A. and Kerkerian-Le Goff L. (1993) Ultrastructural features of the choline acetyltransferase-containing neurons and relationships with nigral dopaminergic and cortical afferent pathways in the rat striatum. Neuroscience 53, 1059–1071. Divac I., Fonnum F. and Storm-Mathisen J. (1977) High affinity uptake of glutamate in terminals of corticostriatal axons. Nature 266, 377–378. Dube´ L., Smith A. D. and Bolam J. P. (1988) Identification of synaptic terminals of thalamic or cortical origin in contact with distinct medium-size spiny neurons in the rat neostriatum. J. comp. Neurol. 267, 455–471. Geneser-Jensen F. A. and Blackstad T. W. (1971) Distribution of acetyl cholinesterase in the hippocampal region of the guinea pig. I. Entorhinal area, parasubiculum, and presubiculum. Z. Zellforsch. mikrosk. Anat. 114, 460–481. Gerfen C. R., Baimbridge K. G. and Miller J. J. (1985) The neostriatal mosaic: compartmental distribution of calcium binding protein and parvalbumin in the basal ganglia of the rat and monkey. Proc. natn. Acad. Sci. U.S.A. 82, 8780–8784. Gerfen C. R. (1988) Synaptic organization of the striatum. J. Electron Microsc. Tech. 10, 265–281. Groenewegen H. J., Room P., Witter M. P. and Lohman A. H. M. (1982) Cortical afferents of the nucleus accumbens septi in the cat, studied with anterograde and retrograde transport techniques. Neuroscience 7, 977–996. Hassler R., Haug P., Nitsch C., Kim J. S. and Paik K. (1982) Effect of motor and premotor cortex ablation on concentrations of amino acids, monoamines, and acetylcholine and on the ultrastructure in rat striatum. A confirmation of glutamate as the specific cortico-striatal transmitter. J. Neurochem. 38, 1087–1098. Hattori T. and McGeer P. L. (1973) Synaptogenesis in the corpus striatum of infant rat. Expl Neurol. 38, 70–79. Heimer L., Zahm D. S. and Alheid G. F. (1995) Basal ganglia. In The Rat Nervous System (ed. Paxinos G.), pp. 579–628. Academic, New York. Herkenham M. and Pert C. B. (1981) Mosaic distribution of opiate receptors, parafascicular projections and acetylcholinesterase in rat striatum. Nature 291, 415–418. Ingham C. A., Hood S. H., van Maldegem B., Weenink A. and Arbuthnott G. W. (1993) Morphological changes in the rat neostriatum after unilateral 6-hydroxydopamine injections into the nigrostriatal pathway. Expl Brain Res. 93, 17–27. Ingham C. A., Hood S. H., Taggart P. and Arbuthnott G. W. (1996) Synaptic plasticity in the rat neostriatum after unilateral 6-hydroxydopamine lesion of the nigrostriatal dopaminergic pathway. In Advances in Behavioural Biology, The Basal Ganglia (eds Ohye V. C. et al.), Vol. 47, pp. 157–164. Plenum, New York. Johnson L. R., Aylward R. L. M., Hussain Z. and Totterdell S. (1994) Input from the amygdala to the rat nucleus accumbens: its relationship with tyrosine hydroxylase immunoreactivity and identified neurons. Neuroscience 61, 851–865. Kachidian P., Vuillet J., Nieoullon A., Lafaille G. and Kerkerian-Le Goff L. (1996) Striatal neuropeptide Y neurones are not a target for thalamic afferent fibres. NeuroReport 7, 1665–1669. Katz L. C. and Shatz C. J. (1996) Synaptic activity and the construction of cortical circuits. Science 274, 1133–1138. Kawaguchi Y. (1993) Physiological, morphological, and histochemical characterization of three classes of interneurons in rat neostriatum. J. Neurosci. 13, 4908–4923. Kelley A. E., Domesick V. B. and Nauta W. J. H. (1982) The amygdalostriatal projection in the rat—an anatomical study by anterograde and retrograde tracing methods. Neuroscience 7, 615–630. Kemp D. E. and Powell J. F. (1971) The termination of fibers from the cerebral cortex and thalamus upon dendritic spines in the caudate nucleus: a study with the Golgi method. Phil. Trans. R. Soc. 262, 429–439. Kita H. and Kitai S. T. (1990) Amygdaloid projections to the frontal cortex and striatum in the rat. J. comp. Neurol. 298, 40–49. Kita H., Kosaka T. and Heizmann C. W. (1990) Parvalbumin-immunoreactive neurons in the rat neostriatum: a light and electron microscopic study. Brain Res. 536, 1–15. Kitai S. T., Kocsis J. D., Preston R. J. and Sugimori M. (1976) Monosynaptic inputs to caudate neurons identified by intracellular injection of horseradish peroxidase. Brain Res. 109, 601–606. Kolb B., Gibb R. and van der Kooy D. (1992) Cortical and striatal structure and connectivity are altered by neonatal hemidecortication in rats. J. comp. Neurol. 322, 311–324. Kubota Y., Mikawa S. and Kawaguchi Y. (1993) Neostriatal GABAergic interneurones contain NOS, calretinin or parvalbumin. NeuroReport 5, 205–208. Lapper S. R., Smith Y., Sadikot A. F., Parent A. and Bolam J. P. (1992) Cortical input to parvalbuminimmunoreactive neurones in the putamen of the squirrel monkey. Brain Res. 580, 215–224. Lapper S. R. and Bolam J. P. (1992) Input from the frontal cortex and the parafascicular nucleus to cholinergic interneurons in the dorsal striatum of the rat. Neuroscience 51, 533–545. Llewellyn-Smith I. J., Pilowsky P. and Minson J. B. (1993) The tungstate-stabilized tetramethylbenzidine reaction for light and electron microscopic immunocytochemistry and for revealing biocytin-filled neurons. J. Neurosci. Meth. 46, 27–40. McGeer P. L., McGeer E. G., Scherer U. and Singh K. (1977) A glutamatergic corticostriatal path? Brain Res. 128, 369–373. McGeorge A. J. and Faull R. L. M. (1989) The organization of the projection from the cerebral cortex to the striatum in the rat. Neuroscience 29, 503–537. Mensah P. L. (1982) An electron microscopical study of neuronal cell clustering in postnatal mouse striatum, with special emphasis on neuronal cell death. Anat. Embryol. 164, 387–401.

Thalamic projection on to parvalbumin-immunoreactive neurons 41. 42. 43. 44. 45. 46. 47. 48. 49. 50. 51. 52. 53. 54. 55. 56. 57. 58. 59. 60. 61. 62. 63. 64. 65. 66. 67. 68. 69.

1175

Meredith G. E. and Chang H. T. (1994) Synaptic relationships of enkephalinergic and cholinergic neurons in the nucleus accumbens of the rat. Brain Res. 667, 67–76. Meredith G. E. and Wouterlood F. G. (1990) Hippocampal and midline thalamic fibers and terminals in relation to the choline acetyltransferase neurons in nucleus accumbens of the rat: a light and electron microscopic study. J. comp. Neurol. 296, 204–221. Misgeld U., Dodt H. U. and Frotscher M. (1986) Late development of intrinsic excitation in the rat neostriatum: an in vitro study. Brain Res. 392, 59–67. Pare´ D. and Smith Y. (1996) Thalamic collaterals of corticostriatal axons: their termination field and synaptic targets in cats. J. comp. Neurol. 372, 551–567. Parent A. (1996) The basal ganglia. In Carpenter’s Human Neuroanatomy (ed. Parent A.), pp. 795–863. Williams and Wilkins, Philadelphia. Paxinos G. and Watson C. (1986) The Rat Brain in Stereotaxic Coordinates. Academic, New York. Pennartz C. M. and Kitai S. T. (1991) Hippocampal inputs to identified neurons in an in vitro slice preparation of the rat nucleus accumbens: evidence for feed-forward inhibition. J. Neurosci. 11, 2838–2847. Purpura D. P. and Malliani A. (1967) I. Synaptic potentials and discharge characteristics of caudate neurons activated by thalamic stimulation. Brain Res. 6, 325–340. Rajakumar N., Elisevich K. and Flumerfelt B. A. (1993) Biotinylated dextran: a versatile anterograde and retrograde neuronal tracer. Brain Res. 607, 47–53. Royce G. J. and Mourey R. J. (1985) Efferent connections of the centromedian and parafascicular thalamic nuclei: an autoradiographic investigation in the cat. J. comp. Neurol. 235, 277–300. Sadikot A. F., Parent A. and Franc¸ois C. (1990) The centre median and parafascicular thalamic nuclei project respectively to the sensorimotor and associative-limbic striatal territories in the squirrel monkey. Brain Res. 510, 161–165. Sadikot A. F., Parent A. and Franc¸ois C. (1992) Efferent connections of the centromedian and parafascicular thalamic nuclei in the squirrel monkey: a PHA-L study of subcortical projections. J. comp. Neurol. 315, 137–159. Sadikot A. F., Parent A., Smith Y. and Bolam J. P. (1992) Efferent connections of the centromedian and parafascicular thalamic nuclei in the squirrel monkey: a light and electron microscopic study of the thalamostriatal projection in relation to striatal heterogeneity. J. comp. Neurol. 320, 228–242. Sadikot A. F. and Rudkin T. M. (1996) The amygdalostriatal projection: an analysis of synaptic inputs to GABAergic interneuron subtypes. In Advances in Behavioural Biology, The Basal Ganglia (eds Ohye V. C. et al.), Vol. 47, pp. 33–42. Plenum, New York. Smith A. D. and Bolam J. P. (1990) The neural network of the basal ganglia as revealed by the study of synaptic connections of identified neurons. Trends Neurosci. 13, 259–265. Smith Y., Bennett B. D., Bolam J. P., Parent A. and Sadikot A. F. (1994) Synaptic relationships between dopaminergic afferents and cortical or thalamic input in the sensorimotor territory of the striatum in monkey. J. comp. Neurol. 344, 1–19. Soghomonian J. J., Descarries L. and Watkins K. C. (1989) Serotonin innervation in adult rat neostriatum. II. Ultrastructural features and a radioautographic and immunohistochemical study. Brain Res. 481, 67–86. Somogyi P., Bolam J. P. and Smith A. D. (1981) Monosynaptic cortical input and local axon collaterals of identified striatonigral neurons. A light and electron microscopic study using the Golgi-peroxidase transport-degeneration procedure. J. comp. Neurol. 195, 567–584. Sterio D. C. (1984) Estimating numbers, mean sizes and variations in size of particles in 3-D specimens using dissectors. J. Microsc. 134, 127–136. Tepper J. M. and Trent F. (1993) In vivo studies of the postnatal development of rat neostriatal neurons. Prog. Brain Res. 99, 35–50. Tessier-Lavigne M. and Goodman C. S. (1996) The molecular biology of axon guidance. Science 274, 1123–1133. Totterdell S. and Smith A. D. (1989) Convergence of hippocampal and dopaminergic input onto identified neurons in the nucleus accumbens in the rat neostriatum. J. chem. Neuroanat. 2, 285–298. Veenman C. L., Reiner A. and Honig M. G. (1992) Biotinylated dextran amine as an anterograde tracer for single- and double-labeling studies. J. Neurosci. Meth. 41, 239–254. Vuillet J., Kerkerian L., Kachidian P., Bosler O. and Nieoullon A. (1989) Ultrastructural correlates of functional relationships between nigral dopaminergic or cortical afferent fibers and neuropeptide Y-containing neurons in the rat striatum. Neurosci. Lett. 100, 99–104. Wouterlood F. G., Bol J. G. and Steinbusch H. W. (1987) Double-label immunocytochemistry: combination of anterograde neuroanatomical tracing with Phaseolus vulgaris-leucoagglutinin and enzyme immunocytochemistry of target neurons. J. Histochem. Cytochem. 35, 817–823. Wouterlood F. G. and Jorritsma-Byham B. (1993) The anterograde neuroanatomical tracer biotinylated dextranamine: comparison with the tracer Phaseolus vulgaris-leucoagglutinin in preparations for electron microscopy. J. Neurosci. Meth. 48, 75–87. Xu Z. C., Wilson C. J. and Emson P. C. (1991) Restoration of thalamostriatal projections in rat neostriatal grafts: an electron microscopic analysis. J. comp. Neurol. 303, 22–34. Yelnik J., Franc¸ois C., Percheron G. and Tande D. (1991) Morphological taxonomy of the neurons of the primate striatum. J. comp. Neurol. 313, 273–294. Zilles K. (1985) The Cortex of the Rat. A Stereotaxic Atlas. Springer-Verlag, Berlin. (Accepted 24 April 1998)