Functional anatomy of the basal ganglia: limbic aspects.

INTRODUCTION: The basal ganglia (BG) have been implicated in different processes that control action such as the control of movement parameters but also in processing cognitive and emotional information from the environment. Here, we review existing anatomical data on the interaction between the BG and the limbic system that support implication of the BG in limbic functions. STATE OF THE ART: The BG form a system that is fairly different from the limbic system, but have strong ties, both anatomical and functional, to the latter. Different models have been proposed. In the parallel model, five segregated circuits from the frontal cortex are individualized and terminate in different regions of the BG and thalamus, before projecting back to their cortical area of origin. Based on the extrafrontal cortical projections, another model has been proposed. It subdivides the cortico-striatal projection into three functional territories: limbic, associative and sensorimotor. In a third spiral model, propagation is possible between limbic information processed by the most medial striatal neurons to motor information processed by the most lateral neurons. PERSPECTIVES: Three main levels of interaction between the BG system and the limbic system are considered. (1) The BG receive direct afferences from several structures associated with the limbic system. Limbic cortical areas project to the striatum, of which the internal architecture is particularly complex, with significant cross-species differences: a compartmental striosome/matrix subdivision described mainly in primates, and a core/shell topographic subdivision of the nucleus accumbens more sharply marked in rodents. (2) Projections from the amygdala form a patchy dorso-ventral progressive gradient in the nucleus accumbens and ventral caudate. (3) Both shell and striosomes receive limbic information from cortical and subcortical limbic structures and project to the dopaminergic neurons of the substantia nigra pars compacta, which in turn modulates their activity. (4) There is a significant overlap between the ventral portions of the BG, nucleus accumbens and ventral pallidum, and the ventral subcortical structures of the limbic system, extended amygdala and nucleus basalis. CONCLUSION: Important interactions exist between the limbic system and the BG system but questions remain about the role that this information plays in the functional organisation of this system. Is limbic information processed separately in the BG, or is it integrated to motor and cognitive information? Do pathological conditions such as obsessive-compulsive disorders or Tourette syndrome result from abnormal afferent limbic input to the BG or abnormal processing within the BG?

Effects of pedunculopontine nucleus area stimulation on gait disorders in Parkinson's disease.

Gait disturbances are frequent and disabling in advanced Parkinson's disease. These symptoms respond poorly to usual medical and surgical treatments but were reported to be improved by stimulation of the pedunculopontine nucleus. We studied the effects of stimulating the pedunculopontine nucleus area in six patients with severe freezing of gait, unresponsive to levodopa and subthalamic nucleus stimulation. Electrodes were implanted bilaterally in the pedunculopontine nucleus area. Electrode placement was checked by postoperative magnetic resonance imaging. The primary outcome measures were a composite gait score, freezing of gait questionnaire score and duration of freezing episodes occurring during a walking protocol at baseline and one-year follow-up. A double-blind cross-over study was carried out from months 4 to 6 after surgery with or without pedunculopontine nucleus area stimulation. At one-year follow-up, the duration of freezing episodes under off-drug condition improved, as well as falls related to freezing. The other primary outcome measures did not significantly change, nor did the results during the double-blind evaluation. Individual results showed major improvement of all gait measures in one patient, moderate improvement of some tests in four patients and global worsening in one patient. Stimulation frequency ranged between 15 and 25 Hz. Oscillopsia and limb myoclonus could hinder voltage increase. No serious adverse events occurred. Although freezing of gait can be improved by low-frequency electrical stimulation of the pedunculopontine nucleus area in some patients with Parkinson's disease our overall results are disappointing compared to the high levels of expectation raised by previous open label studies. Further controlled studies are needed to determine whether optimization of patient selection, targeting and setting of stimulation parameters might improve the outcome to a point that could transform this experimental approach to a treatment with a reasonable risk-benefit ratio.

Effects of nigral stimulation on locomotion and postural stability in patients with Parkinson's disease.

The physiopathology of gait and balance disorders in Parkinson's disease patients is still poorly understood. Levodopa treatment and subthalamic nucleus (STN) stimulation improve step length and walking speed, with less effect on postural instability. These disorders have been linked to dysfunction of the descending basal ganglia outputs to brainstem structures. In this study, we evaluated the effects of stimulation of the substantia nigra pars reticulata (SNr), on locomotion and balance in Parkinson's disease patients. Biomechanical parameters and leg muscle activity were recorded during gait initiation in seven selected patients operated for bilateral STN stimulation, out of 204 stimulated patients, with one contact of each electrode located within the SNr. Step length, anteroposterior and vertical velocities of the centre of gravity were studied, with special reference to the subjects' ability to brake the centre of gravity fall before foot-contact, and compared to seven controls. In Parkinson's disease patients, five treatment conditions were tested: (i) no treatment, (ii) levodopa treatment, (iii) STN stimulation, (iv) SNr stimulation and (v) combined levodopa treatment and STN stimulation. The effects of these treatments on motor parkinsonian disability were assessed with the UPDRS III scale, separated into 'axial' (rising from chair, posture, postural stability and gait) and 'distal' scores. Whereas levodopa and/or STN stimulation improved 'axial' and 'distal' motor symptoms, SNr stimulation improved only the 'axial' symptoms. Compared to controls, untreated Parkinson's disease patients showed reduced step length and velocity, and poor braking just prior to foot-contact, with a decrease in both soleus (S) and anterior tibialis (AT) muscle activity. Step length and velocity significantly increased with levodopa treatment alone or in combination with STN stimulation in both natural and fast gait conditions, and with STN stimulation alone in the fast gait condition. Conversely, SNr stimulation had no significant effect on these measures in either condition. In the natural gait condition, no fall in the centre of gravity occurred as step length was low and active braking was unnecessary. In the fast gait condition, braking was improved with STN or SNr stimulation but not with levodopa treatment, with an increase in the stance leg S muscle activity. These results suggest that anteroposterior (length and velocity) and vertical (braking capacity) gait parameters are controlled by two distinct systems within the basal ganglia circuitry, representing respectively locomotion and balance. The SNr, a major basal ganglia output known to project to pontomesencephalic structures, is postulated as being particularly involved in balance control during gait.

Gait is associated with an increase in tonic firing of the sub-cuneiform nucleus neurons.

In animals, the pedunculopontine (PPN) and the sub-cuneiform (SCU) nuclei located in the upper brainstem are involved during the processing of gait. Similar functional nuclei are suspected in humans but their role in gait is unclear. Here we show that, using extra-cellular recordings of the PPN/SCU region obtained in two parkinsonian patients, the SCU neurons increased their firing rate without modifying their firing pattern during mimicked steps. We conclude that SCU neurons are activated during gait processes.

Modeling the organization of the basal ganglia.

INTRODUCTION: Several models have been elaborated to describe the structure and function of the basal ganglia, but its different levels of architectural organization, macroscopic anatomy, connectivity, functional territorial subdivision and neuronal morphology have rarely been considered. In this review, I present some of these models and I analyze the functioning of the basal ganglia at the light of its architectural properties. STATE OF THE ART: The basal ganglia form an important neuronal system, which interacts with the cerebral cortex through a complex series of loop circuits. While the morphological, electrophysiological and biochemical properties of this system are progressively known better and better, they have led to various interpretations from which different, sometimes contradictory, models have been constructed. The basal ganglia are often analyzed as homogeneous nuclei that communicate through excitatory or inhibitory connections, the so-called "box and arrows" models. Among them, the dual-, triple- and five circuit models are the most popular. Analysis of the "inside of the boxes" provides, however, important data such as the functional subdivision into three, motor, associative and limbic territories and the demonstration that integrative properties are a characteristic of the basal ganglia. PERSPECTIVES: Considering these properties, the way cortical information is processed in the basal ganglia can be analyzed, which leads to modeling its organization and functioning. The striatum receives a compressed version of cortical information and transforms it through complex processes of activation/deactivation under a double dopaminergic and cholinergic control that enables behavioral reinforcement learning. The globus pallidus behaves as a keyboard on which various behavioral repertoires can be coded, from the simplest movement of a single joint to the most complex motor sequence involving the entire body and expressing an emotional content in a cognitive context. The role of the subthalamic nucleus must be considered at different scales. At the macroscopic scale, it works as a thermostat that regulates the level of execution of cortical commands. At a territorial scale, it can process separately motor, cognitive and emotional information. At the neuronal scale, it assures a much finer neuronal representation of cortical commands and can integrate motor, cognitive and emotional aspects. New experiments in both animal models and human clinical-research protocols are needed to demonstrate the neuronal mechanisms of these processes. CONCLUSION: A model is proposed that considers how neural information is processed in the basal ganglia during the execution of motor, cognitive and emotional activities.

Metabolic effects of nigrostriatal denervation in basal ganglia.

In the past, functional changes in the circuitry of the basal ganglia that occur in Parkinson's disease were primarily analyzed with electrophysiological and 2-deoxyglucose measurements. The increased activity of the subthalamic nucleus (STN) observed has been attributed to a reduction in inhibition mediated by the external segment of the globus pallidus (GPe), secondary to the loss of dopaminergic-neuron influence on D2-receptor-bearing striato-pallidal neurons. More recently, in situ hybridization studies of cytochrome oxidase subunit I have confirmed the overactivity of the STN in the parkinsonian state. In addition, this technique has provided evidence that the change in STN activity is owing not only to decreased inhibition from the GPe but to hyperactivity of excitatory inputs from the parafascicular nucleus of the thalamus and the pedunculopontine nucleus in the brainstem.

Golgi study of the primate substantia nigra. II. Spatial organization of dendritic arborizations in relation to the cytoarchitectonic boundaries and to the striatonigral bundle.

The spatial organization of Golgi-stained dendritic arborizations of the substantia nigra was studied in three dimensions by using a video computer system. Dendritic orientation was analyzed in relation to the cytoarchitectonic boundaries and to the direction of the axons of the striato-pallidonigral bundle. All the brains, humans and macaques, were sectioned according to the same ventricular planes. The striatal bundle is made up of dense fascicles of very thin parallel axons. Sixty neurons located in the pars reticulata, lateralis, and compacta were reconstructed from serial sections. In the anterior pars reticulata and lateralis, the dendritic arborizations spread in all directions inside the striatal bundle. Below the pars compacta fringes, the dendrites of pars reticulata neurons extend ventrolaterally in the bundle. Because one nigral arborization can cover the whole thickness of the striatal bundle, we are led to believe that nigral neurons exert a role of convergence of the corticostriatal information similar to that of pallidal neurons (Percheron et al., '84a,b). The pars reticulata neurons appear to receive information mainly from the associative striatal territory. The pars lateralis neurons, conversely, appear to receive information from the sensorimotor territory. The anterior pars compacta neurons are organized in such a way that their ventral dendrites, located inside the pars reticulata, are ventrolaterally oriented, perpendicular to the striatal bundle. Their dorsal dendrites remaining in the pars compacta can receive other input. At more caudal levels, the posterior pars compacta neurons have dendrites radiating outside the striatal bundle.

A Golgi analysis of the primate globus pallidus. II. Quantitative morphology and spatial orientation of dendritic arborizations.

The morphology of pallidal neurons was analyzed quantitatively in Golgi-impregnated brains of men and macaques (Macaca irus). Selected neurons were drawn with a camera lucida and reconstructed from serial sections. Dendritic arborizations were analyzed in three dimensions using a video computer microscope (Yelnik et al., '81). Morphological (topological and metrical) parameters were computed, and the overall geometry of arborizations was studied in three dimensions with the aid of principal component analysis (Yelnik et al., '83). Statistical tests were used in order to compare human with simian large neurons and the lateral with the medial pallidum. All neurons were found to belong to a single neuronal population. Particular neurites may be added randomly onto pallidal dendrites, mainly in the lateral pallidum. Large pallidal neurons are characterized by sparsely branched dendritic arborizations (4 stems, 13 tips) with thick, smooth, and long dendrites (longest dendrite = 1,000 micron, total dendritic length = 7,600 micron). All arborizations are discoidal in shape with mean dimensions of 1,500 X 1,000 X 250 micron. Pallidal discs are always parallel to the lateral border of each pallidal nucleus and thus perpendicular to striatal axons to which they present their greatest extent. They may be traversed by a large number of these axons. The existence of other pallidal neuronal groups, "intermediate" and "local circuit" neurons, identified in their fine morphological features by François et al. ('84), was confirmed quantitatively in the present study.

A Golgi analysis of the primate globus pallidus. III. Spatial organization of the striato-pallidal complex.

An atlas of transverse sections of the globus pallidus and striatum was established in macaque with reference to ventricular coordinates. The three-dimensional geometry of the striato-pallidal complex was investigated by means of sagittal and horizontal reconstructions. Both a personal case studied with autoradiography and data from literature were used to analyze the distribution of cortical axons into the striatum. One may distinguish two striatal territories: one, somatotopically arranged, sensorimotor territory extending over the major part of the putamen; and the other, an associative territory, comprising the caudate nucleus and antero-medial and postero-inferior parts of the putamen. The striato-pallido-nigral bundle was studied using Golgi, Perls, and Fink-Heimer techniques. The bundle is described in four parts: prepallidal (subdivided into caudato-pallidal and putamino-pallidal subparts), transpallidal, pallido-nigral, and nigral. The tracing of the limit between the caudate (associative) and putaminal (essentially sensorimotor) territories shows that the two components are of roughly the same size in the pallidum. The data were compared with geometry and orientation of the dendritic arborizations of large pallidal neurons analyzed in Yelnik et al. ('84). Each pallidal dendritic disc is able to receive axons from a wide region of the striatum. This leads to a convergence on pallidal neurons of striatal axons from different striatal somatotopic strips and from the sensorimotor and associative territories. This is an indication that the globus pallidus may have an integrative role.

Morphological taxonomy of the neurons of the primate striatum.

A quantitative taxonomy of primate striatal neurons was elaborated on the basis of the morphology of Golgi-impregnated neurons. Dendritic arborizations were reconstructed from serial sections and digitized in three dimensions by means of a video computer system. Topological, metrical, and geometrical parameters were measured for each neuron. Groups of neurons were isolated by using uni- and multidimensional statistical tests. A neuronal species was defined as a group of neurons characterized quantitatively by a series of nonredundant parameters, differing statistically from other groups, and appearing as a separate cluster in principal component analysis. Four neuronal species were isolated: (1) the spiny neuronal species (96% of striatal neurons) characterized by spine-free proximal dendrites (up to 31 microns) and spine-laden distal dendrites, which are more numerous, shorter, and less spiny in the human than in the monkey, (2) the leptodendritic neuronal species (2%) characterized by a small number of long, thick, smooth, and sparsely ramified dendrites, (3) the spidery neuronal species (1%) characterized by very thick dendritic stems and a large number of varicose recurrent distal processes, and (4) the microneuronal species (1%) characterized by numerous short, thin, and beaded axonlike processes. All striatal neurons give off a local axonal arborization. The size and shape of cell bodies were analyzed quantitatively in Golgi material and in materials treated for Nissl-staining, immunohistochemical demonstration of parvalbumin and histochemical demonstration of acetylcholinesterase. Only three types were distinguishable: small, round cell bodies corresponding to either spiny neurons or microneurons, medium-size elongated cell bodies, which were parvalbumin-immunoreactive and corresponded to leptodendritic neurons, and large round cell bodies, which were acetylcholinesterase-positive and corresponded to spidery neurons. Thorough analysis of previously elaborated classifications revealed that spidery neurons do not exist in rats and cats and that large cholinergic neurons in these species correspond to leptodendritic neurons. From this, it can be assumed that the dendritic domain of striatal cholinergic neurons is considerably smaller in primates than in other species. Computer simulations based on both the frequency of each neuronal species and their three-dimensional dendritic morphology revealed that the striatum consists of two intertwined dendritic lattices: a fine-grain lattice (300-600 microns) formed by the dendritic arborizations of spiny, spidery, and microneurons, and a large-grain lattice (1,200 microns) formed by the dendritic arborizations of leptodendritic neurons. This suggests that cortical information can be processed in the striatum through two different systems: a fine-grain system that would conserve the precision of the cortical input, and a large-grain system that would blur it.

Central complex of the primate thalamus: a quantitative analysis of neuronal morphology.

Neuronal morphology was analyzed in the central complex (centre median-parafascicular complex) of macaques and humans. Cell bodies were described from Nissl material. Golgi-impregnated dendritic arborizations were reconstructed from serial sections and digitized in three dimensions by computer-assisted microscopy. The central complex was subdivided into three parts on the basis of cytoarchitectonic and hodological criteria: pars parafascicularis (medial), pars media (intermediate), and pars paralateralis (lateral). The mean cross-sectional areas of cell bodies were identical (181 microns2) in the three parts in macaques. In humans they were larger in the pars parafascicularis (304 microns2) than in the other parts (248 and 240 microns2). Small local circuit neurons were found throughout the complex. Large projection neurons differed statistically in the three parts. In macaques, pars parafascicularis neurons had few dendritic stems and tips (3-11) and a short total dendritic length (2,000 microns). Pars paralateralis neurons had more ramified (5-60) and longer (5,800 microns) dendrites. They bore numerous axonlike processes. Pars media neurons had intermediate characteristics (5-19; 2,400 microns). In humans, pars parafascicular neurons had similar topological characteristics (3-12) but longer dendrites (3,000 microns) than in the monkey. Pars paralateralis neurons had more branched (6-71) and longer (9,000 microns) dendrites, with more numerous axonlike processes. Pars media neurons also had intermediate characteristics (4-25; 3,800 microns). The present study supports a tripartite subdivision of the primate central complex and demonstrates significant interspecies differences.

Golgi study of the primate substantia nigra. I. Quantitative morphology and typology of nigral neurons.

Neuronal morphology was analyzed in the pars compacta, reticulata, and lateralis of the substantia nigra of humans and macaques. Golgi-impregnated dendritic arborizations, reconstructed from serial sections, were described by using topological, metrical, and geometrical parameters measured in three dimensions. Morphological parameters were statistically analyzed. Cell bodies and axons were also described. The primate substantia nigra comprises few local circuit microneurons. It consists mainly of large projection neurons having large cell bodies and sparsely branched dendritic arborizations. In all subdivisions, "complex endings" and "thin processes" can be found on nigral dendrites. Axons of large neurons occasionally had initial collaterals that never form profuse arborizations. Pars reticulata neurons had a cell body surface of 520 micron2, 4 dendritic stems, and 13 dendritic tips. The total dendritic length (L) was 7,100 micron, the highest dendritic length (Lm) 1,200 micron, and the mean length of dendritic segments 320 micron. Pars lateralis neurons were similar except for their larger cell bodies (650 micron2) and longer dendritic segments (440 micron). Pars compacta neurons had larger cell bodies (860 micron2), thicker and more numerous (5 stems, 19 tips), and longer dendrites (L = 10,500 micron; Lm = 1,400 micron). Large neurons of monkeys had the same topological characteristics as human neurons but shorter dendrites. The overall shape of arborizations was highly variable and not characteristic in any subdivision. A hierarchical typology of nigral neurons is proposed, which comprises two neuronal species, the compacta and reticulata species, and a lateralis subspecies. Pallidal neurons (Yelnik et al., '84) belong to the reticulata species. The position of these species in relation to higher hierarchical levels is discussed.

A Golgi analysis of the primate globus pallidus. I. Inconstant processes of large neurons, other neuronal types, and afferent axons.

The present paper is a Golgi study, with high-power lenses, of the primate globus pallidus. Two kinds of inconstant processes of large neurons are first described: complex endings and thin processes. Complex endings are thick apparatuses terminally located on dendrites having many appendages of various types. Contacts were observed not only between striatal axons and these complex endings but also between complex endings and the soma, dendritic stems, dendritic portions or complex endings of other large pallidal neurons. Thin processes were usually beaded, very thin, and arose from any part of the dendritic tree. Contacts were seen between them and soma or dendrites of other large neurons. These thin processes were very similar to initial axonal collaterals and together constitute a common pool of processes. Complex endings and thin processes were essentially observed in the lateral nucleus of the pallidum where they apparently are evenly distributed inside the nucleus but randomly distributed on individual neurons. Two neuronal types other than large pallidal neurons were isolated: the smallest were considered to be local circuit neurons, while intermediate-sized neurons might be the origin of a particular efference. Many striatal axons gave no branches over long distances and collaterals were of two types and most frequently were short (less than 50 micron). Larger axonal arborization were rarely encountered. In addition to parallel contacts, numerous very short ones were observed. All these contacts between striatal axons and dendrites of large pallidal neurons seem to be irregularly distributed.

Dopaminergic cell group A8 in the monkey: anatomical organization and projections to the striatum.

The first part of the study was a quantitative analysis of the distribution of A8 neurons compared with that of A9 and A10 neurons by means of tyrosine hydroxylase and calbindin-D(28K) immunohistochemistry and image analysis in monkeys. Then the striatal projection of A8 neurons was studied using retrograde and anterograde tracing methods. It was compared with that originating in cell groups A9 and A10 by performing injections of the retrograde tracer wheat germ agglutinin conjugated to horseradish peroxidase into different regions of the striatum. Ten percent of all mesencephalic dopaminergic neurons are located in cell group A8. This cell group, along with A10 and the dorsal part of A9, constitutes the dorsal tier, which accounts for 28% of mesencephalic dopaminergic neurons. Double-staining experiments showed that the neurons located in the dorsal tier were calbindin positive, whereas those from the ventral tier were not. In terms of anatomical projection, the dorsal tier mainly projects to the ventral part of the associative striatum, with preferential projections of A8 neurons to the ventrocaudal putamen, of A10 neurons to the nucleus accumbens, and of dorsal A9 neurons to both. Conversely, the main targets of the ventral tier of mesencephalic neurons (ventral part of A9) are the sensorimotor putamen and the associative caudate nucleus. In conclusion, each mesencephalic cell group projects primarily to one specific striatal region but also participates, albeit to a lesser extent, in the innervation of all the remaining striatal parts.

Topography of the projection from the central complex of the thalamus to the sensorimotor striatal territory in monkeys.

The distribution of axons arising from the central complex (or centre médian-parafascicular complex) and terminating in the striatum was studied in seven macaques and one squirrel monkey. Deposits of anterograde tracers were made in the two lateral-most subdivisions of the central complex, i.e., the middle part (or pars media) and the lateral part (or pars paralateralis). All injections avoided the pars parafascicularis. The intrastriatal distribution of labeled axonal endings was mapped in relation to the standard ventricular (CA-CP) system of coordinates. Labeled endings were observed in the major posterior and dorsal parts of the putamen (excluding its anteromedial and ventral parts) and also in a restricted ventrolateral part of the caudate nucleus. The topography of the central territory of the striatum, defined as the striatal space receiving axons from the central complex, was found to correspond exactly to that of the cortical sensorimotor territory delineated after cortical injections. The termination pattern of the central axons within the striatum was patchy. Viewed as a whole, the irregular and hazy patches formed oblique streaks, parallel one with the other. The three-dimensional reconstructions of data from transverse sections revealed that the streaks were bi-dimensional pictures of three-dimensional parasagittal layers covering the whole anteroposterior extent of the cortical sensorimotor territory of the striatum. Our work shows that the pars media of the central complex, which receives selectively pallidal afferent axons (François et al., '88: Brain Res. 473:181-186), is the main source of the centroputaminal projection. The probable implication of this in a closed sensorimotor loop of the basal ganglia is discussed.

Distribution and spatial geometry of dopamine interplexiform cells in the rat retina: I. Developing retina.

The morphology and distribution of dopaminergic interplexiform cells were analyzed in 9-day-old rat retinas processed as wholemounts for tyrosine hydroxylase immunohistochemistry. The mean number of dopaminergic interplexiform cells was estimated as about half of the total number of dopaminergic neurons in the immature retina, with a higher density in the temporal retina. Four interplexiform cells were individually analyzed and reconstructed with a computer system. Their arborizations could be divided into three different regions based on both their morphological features and their position within the retinal layers: (1) an internal arborization, spreading at the margin between the inner nuclear layer and the inner plexiform layer, composed of long, thick, somatofugal dendrites branching at acute angles, (2) an external arborization in the middle of the inner nuclear layer, formed by short, thin, varicose, recurved, axon-like processes branching at right angles, (3) and one or more scleral process(es), originating either from the cell body or from the internal arborization, running toward the outermost cell row of the INL, some of which reached the outer plexiform layer. Finally, analysis of the arborization network by computer simulations based on the 4 digitalized cells was compared with a nearest-neighbour analysis of cell body distribution. It showed that cell bodies are almost randomly distributed--at least in the inferior retina--but that an adjustment of dendritic growth and orientation probably occurs to ensure a homogeneous coverage of the retina with a constant degree of overlap in the adult. This report represents the first three-dimensional computer reconstruction of chemically identified neurons in the retina.

Calpastatin immunoreactivity in the monkey and human brain of control subjects and patients with Parkinson's disease.

Parkinson's disease is characterized by a selective loss of dopaminergic neurons in the nigrostriatal pathway. However, not all dopaminergic neurons degenerate in this disease, and calcium has been suspected of playing a role in this differential vulnerability. An overexpression of the calcium-dependent protease calpain II has recently been reported in the parkinsonian substantia nigra, suggesting that a rise in intracellular calcium concentrations may be involved in the mechanism leading to cell death. The proteasic activity of calpain is regulated by an endogenous inhibitory protein called calpastatin. Because little is known about the distribution of calpastatin in the primate brain, we first analyzed immunohistochemically the calpastatin expression in normal human and monkey brain. A ubiquitous distribution of calpastatin immunostaining was observed in both cases, but its expression was variable from one region to another. In the basal ganglia, staining was intense in the striatum, in the pallidal complex, and in some nuclei of the thalamus. The cerebellum was stained intensely, particularly in the granular and Purkinje cell layers. A dense, heterogeneous staining was observed in the hippocampal formation, mostly in the pyramidal and granular layers. The distribution of staining was similar in the different cerebral cortices studied, and it was most intense in layer V. In the brainstem, staining was particularly prominent in the substantia nigra pars reticulata and compacta, the central gray substance, the superior colliculus, and the cuneiform nucleus, and staining was moderate in the tegmenti pedonculopontinus nucleus and the griseum pontis. In the second part of the study, the authors compared calpastatin expression in the mesencephalon between patients with Parkinson's disease and control subjects. Sequential double staining revealed that some dopaminergic neurons coexpress calpastatin, the proportion of double-stained neurons ranging between 52% and 76% among the different dopaminergic cell groups. Quantitative analysis of the number of calpastatin-stained neurons evidenced a loss of both calpastatin-positive and calpastatin-negative neurons in the substantia nigra of patients with Parkinson's disease. These data suggest that calpain II overexpression in Parkinson's disease is not compensated for by a concomitant increase in calpastatin expression.

Dopaminergic innervation of the subthalamic nucleus in the normal state, in MPTP-treated monkeys, and in Parkinson's disease patients.

The existence of a dopaminergic innervation of the subthalamic nucleus (STN) has been demonstrated in rats but has remained controversial in primates. The aim of the present study was first to demonstrate the existence of a dopaminergic innervation of the STN in monkeys using tracing methods and then to quantify the loss of dopaminergic fibers in the parkinsonian state in monkeys and humans. Following injection of Fluoro-Gold into the STN of a vervet monkey (Cercopithecus aethiops), retrogradely labeled neurons were found to be scattered in all dopaminergic areas of the mesencephalon. Injection of biotin dextran amine into dopaminergic areas A8 and A9 of two monkeys resulted in anterogradely labeled axons located throughout the whole extent of the STN. Labeled axons that also expressed tyrosine hydroxylase (TH) were reconstructed from serial sections. Some terminal axonal arborizations had profuse branching and occupied much of the STN, and others were restricted to small portions of the nucleus. In TH-immunoreactive sections, numerous sparse, fine, and varicose TH-positive fibers were observed in the STN of normal monkeys and humans. Quantification of these TH-positive fibers revealed a 51% loss of TH-positive fibers in MPTP-intoxicated monkeys and a 65% loss in Parkinson's disease patients compared with their respective controls. These findings demonstrate the existence of a dopaminergic innervation of the STN in primates. The loss of dopaminergic innervation in MPTP-intoxicated monkeys and in Parkinson's disease patients may directly affect the activity of STN neurons and could participate in the hyperactivity of the structure.

Distribution and spatial geometry of dopamine interplexiform cells in the retina. II. External arborizations in the adult rat and monkey.

The morphology and distribution of dopaminergic interplexiform cells in adult rat and monkey retinas were analyzed to determine any correlation with the function of dopamine in the outer retinal layers. The retinas were processed as whole mounts for tyrosine hydroxylase immunohistochemistry. There was a network formed by the sclerally directed processes of interplexiform cells in the inner nuclear, outer plexiform, and outer nuclear layers running throughout the retina. Their density was higher in the superior retina than in the inferior retina of the rat and was especially high in the superior temporal quadrant. The external network in this quadrant was significantly less dense in the monkey than in the rat, as are the interplexiform cells. The somata of interplexiform and other dopaminergic cells were about the same size in both rats and monkeys. Computer-assisted reconstruction of external arborizations of individual cells showed that external processes lay very close to horizontal and photoreceptor cells and also to blood capillaries. Because they were long, thin, and highly varicose; branched at right angles; and often arose from an axon hillock, the external processes were identified as axons. Therefore, we define the dopaminergic interplexiform cells as multiaxonal neurons, with at least one outwardly directed axon that reaches the outer plexiform layer. The function of the network of external processes from the interplexiform dopaminergic cells is discussed in terms of modulating the release of dopamine to external layers.

Parkin immunoreactivity in the brain of human and non-human primates: an immunohistochemical analysis in normal conditions and in Parkinsonian syndromes.

The etiology of Parkinson's disease is unknown, but the gene involved in an autosomic recessive form of the disease with early onset has recently been identified. It codes for a protein with an unknown function called parkin. In the present study we produced a specific polyclonal antiserum against human parkin. Immunohistochemical analysis showed that parkin is expressed in neuronal perikarya and processes but also in glial and blood vessels in the primate brain (human and monkey). Electron microscopy indicated that parkin immunoreactivity is mostly located in large cytoplasmic vesicles and at the level of the endoplasmic reticulum. Parkin was expressed heterogeneously in various structures of the brain. It was detectable in the dopaminergic systems at the level of the perikarya in the mesencephalon but also in the striatum. However, parkin was also expressed by numerous nondopaminergic neurons. The staining intensity of parkin was particularly high in the hippocampal formation, the pallidal complex, the red nucleus, and the cerebellum. Comparison of control subjects with patients with Parkinson's disease and control animals with 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine (MPTP)-intoxicated animals revealed a loss of parkin-immunoreactive neurons only in the substantia nigra pars compacta. Furthermore, the surviving dopaminergic neurons in the parkinsonian state continued to express parkin at a level similar to that observed in the control situation. These data indicate that parkin is a widely expressed protein. Thus, the degeneration of dopaminergic neurons in familial cases of Parkinson's disease with autosomal recessive transmission cannot be explained solely in terms of an alteration of this protein.