The Substantia Nigra of the Rat: A Golgi Study JANICE M. JURASKA, CHARLES J. WILSON AND PHILIP M. GROVES Department of Psychology, University of Colorado, Boulder, Colorado 80309
ABSTRACT Three variants of the Golgi method were employed to examine the cell types, their dendritic fields and organization and axonal trajectories within the substantia nigra of albino and hooded rats. In both sagittal and coronal sections, large, medium and small neurons were classified on the basis of soma size, extent of dendritic fields and dendritic caliber. In general nigral cells have three to five primary dendrites that branch relatively infrequently. Some dendrites of all cell types have thinly scattered spines or varicosities. Small cells, found in all areas of the nucleus, have thin dendrites and small, nondirectional dendritic fields. These are considered to be interneurons. The medium cells found in pars compacta, presumed to be the dopaminergic cells of the nigroneostriatal pathway, send long dendrites into pars reticulata perpendicular to the course of pars compacta. In addition, these cells have a number of dendrites which remain in pars compacta. These cells have axons that run medio-dorsally. No axon collaterals were detected. Both large and medium cells are found in pars reticulata. Cells in the dorso-medial aspect of pars reticulata orient rostro-caudally and roughly perpendicular to the course of pars compacta, while cells in the peripeduncular area show a strict orientation which is parallel to the crus cerebri. Some pars reticulata cells emit axon collaterals while others remain unbranched for their observable length. Both large and medium cells are also seen in pars lateralis. These cells send long dendrites ventrally into pars reticulata where they run parallel to the crus cerebri, while some shorter dendrites remain in pars lateralis. In total, the substantia nigra appears to have a layered organization: the superior layer is the cellular pars compacta, the second is the dorso-medial area of pars reticulata where both pars compacta and pars reticulata dendrites run rostro-caudally and dorso-ventrally and the third layer is the peripeduncular region where dendrites from all areas run parallel to the crus cerebri.
While its very existence was in question as little as a decade ago, the dopaminergic nigro-neostriatal pathway has now been implicated in a number of phenomena of psychological, pharmacological and clinical importance, including eating and drinking behavior (Zigmond and Stricker, '72), self-stimulation and learning (e.g., Routtenberg and Holtzman, '73; German and Bowden, '74; Zis et al., '74), the behavioral effects of amphetamine and the antipsychotic drugs (e.g., Groves and Rebec, '76), the etiology of Parkinson's disease (e.g., Hornykiewicz, '66), and postlesion axonal regeneration (e.g., Stenevi et al., '73). The source and trajectory of the nigro-neostriatal tract have been visualized with clarity J. COMP. NEUR., 172 585-600.
using the catecholamine histofluorescence methods of Falk and Hillarp (e.g., Andkn et al., '64;Un erstedt, '71) and Lindvall and Bjorklund '74). It arises from fluorescent cells of medium size located primarily in the pars compacta region of the substantia nigra, whose axons are seen to course dorsally and medially to form a fairly compact fascicle in the medial forebrain bundle just rostra1 to the substantia nigra. These cells are almost certainly the medium-sized cells which are seen in Nissl stained preparations to make up the majority of cells in the pars compacta, and which are specificially ' Reprints may be obtained from Philip M. Groves, Depart-
8
ment of Psychology, University of Colorado, Boulder, Colorado 80309.
585
586
J. M. JURASKA, C. J. WILSON AND P. M. GROVES
labeled after intraventricular administration of labeled catecholamines (Parizek et al., '71; Sotelo, '71; Ljungdahl et al., '75). The restriction of the dopaminergic cells to this cytoarchitecturally defined subdivision of the substantia nigra suggests the presence of spatially organized, perhaps very specific, arrangements of interacting neurons of functionally different types within this brain stem nucleus. Thus the activity of dopaminergic neurons might be regulated in part by their interactions with other areas of the substantia nigra. While some evidence is available relating dopaminergic activity to that of certain nigral afferents (e.g., Precht and Yoshida, '71; Corrodi et al., '72; Kim et al., '71; Fonnum et al., '74; McGeer et al., '74) very little is known of the local integration which might be expected to modify or transform that input. The first requirement for such an analysis of local interactions, a detailed description of the geometric organization of cell types and the dendritic fields of the cells, is reported here. Special attention has also been paid to the initial axonal trajectories of the various cell t es insofar as they may be determined in t ick sections prepared by the Golgi methods.
YR
MATERIALS AND METHODS
Three variants of the Golgi method were applied to tissue blocks taken from the midbrains of male and female rats. The material from Simonsen rats, between 36 and 100 days of age was stained by the GolgiKopsch method (Colonnier, '64). Brains of some Long-Evans hooded rats aged 55 days were stained by the Golgi-Cox method described by Van der Loos ('56), while others aged 5 to 20 days were stained by the rapid Golgi method according to Valverde ('70). Both coronal and sagittal sections were cut at 100 pm. A representative sample of 130 neurons was chosen from a much larger sample of observed neurons and drawn at a magnification of x 625 with the aid of a Leitz camera lucida. OBSERVATIONS
The substantia nigra, present throughout
the midbrain, lies dorsal to the crus cerebri and ventral to the medial lemniscus. Medially it is bordered by the ventral tegmental area. The largest part of the nucleus is occupied by the dense neuropil of ars reticdata, which in Golgi material can e seen to consist of a network of small fibers coursing generally in the longitudinal and dorso-ventral directions, and dendrites with a similar organization. The neuropil is interrupted here and there by large fiber bundles from the crus cerebri, and by cell bodies of various sizes. This is in striking contrast to the pars compacta, which forms a thin densely cellular layer enclosing pars reticdata on all sides except for that bordering on the crus cerebri. Here the cells are of similar size and shape, and dendrites follow the contour of the cellular layer with dendritic fields overlapping extensively. The least distinct region is pars lateralis, an elli tical region lying on the lateral edge o f t e brainstem between the dorsolateral pars reticdata and the ventral lateral border of the medial geniculate.
1
K
Cell types Three general classes of cells could be distinguished on the basis of cell body size, thickness and branching of dendrites, and extent of dendritic field. Since the transition from soma to dendritic trunk is gradual, the measurement of cell body size is somewhat arbitrary in Golgi preparations. For this reason, two measurements were made on each neuron, one along its longest dimension, and one perpendicular to the long axis. Such measurements did, as expected, yield a generally trimodal distribution of cell body sizes. It was not possible, however, to unambiguously assign a cell to any distinct category on the basis of soma size alone, except in the case of the largest nigral neurons. Seen in pars reticulata and pars lateralis, these large neurons measure 45-74 p m along their longest axis. In other respects they resemble the medium-sized cells of these regions (19-46 pm). On the other hand, a catego of cells with thin delicate dendrites ra 'ating in all directions but with relatively small dendritic
';tl
RAT SUBSTANTIA NIGRA
fields can easily be distinguished from the other categories, despite the extensive overla between their cell body size (1126 pm and that of the medium-sized neurons. Cell body shape also varies widely. Small neurons tend to have round cell bodies. Medium and large cells are more heterogeneous in shape, with some being fusiform, triangular, polygonal, or ovoid. Aside from some tendency for mediumsized pars compacta neurons to be pyramidal in shape, there is no discernible segregation of cell body shape according to region in the substantia nigra. Many neurons of all types in the substantia nigra have some dendrites with varying degrees of beading or varicosities (figs. 1,4, 6, 11).The varicosities are generally not found on thick dendritic trunks, and are most obvious on dendrites located at some distance from the cell body. This is apparently related to dendritic caliber, cells having thin dendrites exhibiting varicosities closer to the cell body. Varicosities are also more commonly seen in younger animals. This was particularly noticeable in material prepared from the 5- and 10-day-old animals. Many neurons of all cell types also have thinly scattered dendritic spines (e.g., figs. 1, 4, 8). Spines are infrequent on the cell body itself. Although not a strict rule, dendrites with varicosities seem to have fewer spines, and vice versa. Some cells are seen to have both, and it was not possible to place cells into categories based on these features.
587
P
Fig. 1 A camera lucida drawing of a mediumsized pars compacta cell from a coronal section. The cell is oriented so that dorsal is toward the top of the figure and lateral is to the right.The dendrite running in a ventral direction is in pars reticulata for most of its course. The arrow points to the axon. GolgiKopsch. Bar, 100 Fm.
its superior surface so that in sagittal sections, cells at the rostral and caudal poles of the nucleus have dendrites oriented rostrocaudally (fig. 6) while dendrites from cells of the middle portion of the nucleus have a dorso-ventral orientation. In coronal secPars compacta tions the dendrites going into pars reticuThe neurons of pars compacta are pri- lata have a ventro-lateral or ventral course marily of medium size. They resemble the (fig.1).When only one dendrite is sent into medium-sized neurons of pars reticdata pars reticulata, it is frequently the cell’s and pars lateralis, but can be immediately largest and longest, giving many of these recognized by their dendritic branching neurons the appearance of an “upside pattern. Each characteristically sends one down yramidal cell” as described by or two dendrites deep into pars reticulata others TCajal, ’55; Grossman, ’66; Schwyn where they may branch one or two times. and Fox, ’74), with its apical dendrite These dendrites generally run parallel to oriented into pars reticdata and its basilar the course of the pars reticulata dendrites, dendrites remaining in pars compacta (e.g., and often show varicosities (figs. 1,6). The fig. 1). The basilar dendrites, of which orientation of these dendrites as they enter there are two to five, frequently branch pars reticulata is roughly perpendicular to one or two times and run parallel to the
588
J. M. JURASKA, C . J. WILSON AND P. M. GROVES
surface of pars compacta, again regardless of their position in the nucleus. Occasionally these dendrites exhibit a tufted branching pattern, as seen in sagittal section in figure 5. They may course for some distance within pars compacta, going in both directions and so forming a disk-like dendritic field. Since the cells of pars compacta are so close together, these dendritic fields overlap extensively. Except in the case of the most medially placed neurons, the dendrites of pars compacta cells do not extend outside of the substantia nigra. In the case of cells situated on the border of the ventral tegmental area however, dendrites are often seen to extend medially and run for some distance into the latter nucleus where they intermingle with the dendrites of cells of similar appearance there. While no attempt was made to characterize the cells of the ventral tegmental area in detail, this observation is consistent with the observation that this region, which is also rich in dopamine-containing neurons, is continuous with the pars compacta (e.g., Dahlstrom and Fuxe, '64). Axons arising from some pars compacta medium cells could be followed for as far as 165 pm before passing out of the plane of the section. They arise from cell bodies or primary dendrites in sagittal sections coursing initially in a dorsal or dorso-caudal direction. After a variable distance, they take a sharp medial turn and are lost from view. In coronal sections their course appears dorsal and/or medial. In no case was an axon from this type of cell seen to emit a collateral. The small neurons of pars compacta are much less frequently seen, representing only 18% of our sample of 61 neurons drawn from this area. They have small stellate dendritic fields consisting of three to five thin dendrites which generally branch one or two times, although in some cases u to four branchings have been seen (fig. l O f Although they are identical in all respects to pars reticulata small neurons, their dendrites seem to be restricted to pars compacta. Their axons arise from the cell body or a primary dendrite and appear to
have no preferred direction. Only a small number of axons arising from these cells were followed for more than a few microns. In some of these cases however, they could be seen to branch although it was not possible to determine the extent of their branching or the total size of their axonal trajectory.
Pars reticulata Cells of all three types are seen throughout pars reticdata. Large and mediumsized cells, which differ only in cell body size (figs. 2, 3, 7-9), will be considered together. In sagittal sections through the middle portion of the nucleus, dendrites of both of these are seen to course primarily in an anterior-posterior direction with some tendency to spread apart in the dorso-ventral direction, es ecially ventral toward the crus cerebri &g. 8). Thus the dendritic fields of these cells give the appearance of a narrow plate running nearly the entire length of the nucleus in the sagittal plane. In coronal sections most dendrites generally run in a dorso-ventral direction with little medio-lateral component (figs. 7, 9). However as in pars compacts, the absolute orientation of dendritic fields varies with location within the nucleus. This is evident in coronal sections such as the one reconstructed in figure 5 in
Fig. 2 A camera lucida drawing of a large-sized pars reticulata cell from a coronal section. The cell is oriented so that dorsal is toward the top and to the right of the figure and lateral is to the right and bottom. The arrow points to the axon. This cell lies in the medial area of pars reticulata and its dendrites display the mediolateral orientation typical of cells in this location. Golgi-Kopsch. Bar, 100 pm.
RAT SUBSTANTIA NIGRA
which it can be seen that the dendrites of pars reticulata cells are oriented roughly perpendicular to the course of the pars compacta cell layer. This results in most dendrites having a strongly dorso-ventral orientation (figs. 7, 9) that changes somewhat to mediolateral in the extreme medial areas (fig. 2). Near the crus cerebri, however, a different organization is seen. Those dendrites of pars reticulata neurons which approach the crus cerebri tend to ramify and send branches to run for some distance parallel to the surface of the peduncle. In this area also, a second type of medium-large-sized pars reticulata neuron is seen, whose dendritic field runs lengthwise along the crus cerebri (fig. 11). These peripeduncular cells are particularly prominent in coronal sections through the midportion of the nucleus where their dendrites, and the ventral branches of the more dorsal pars reticulata neurons give the nucleus a distinctly layered appearance. In the peripeduncular region, all dendrites appear to
589
run primarily in the coronal plane, and parallel to the surface of the brainstem, while the bulk of the pars reticulata, lying dorsal to this layer, has a more dorso-ventral orientation. Dendrites of pars reticulata tend not to leave this region. Cells located near the pars compacta or pars lateralis may send dendrites into those regions, but do not course through them for any great distance. Both large and medium cells of pars reticulata often have thick dendritic trunks arising from the cell body. These are especially noticeable in many of the larger cells (figs. 7, 8), and may give rise to long dendrites measuring up to 590 p m in length which are sometimes unbranched throughout most of their course. Generally they do however, branch once or twice near the cell body, and again if they emit a branch running near the crus cerebri. Dendrites in the peripeduncular layer often display prominent varicosities (fig. 11). Axons from both the large and medium
Fig. 3 A camera lucida drawing of a medium-sizedpars reticulata cell from a sagittal section. The cell is oriented so that dorsal is toward the top of the figure and rostra1 is to the right. The long rostrocaudal course of this cell’s dendrites can be noted, as well as the tendency to gradually spread out in a dorso-ventral direction. Golgi-Kopsch.Bar, 100 pm.
590
J. M. JURASKA,C. J. WILSON AND P. M. GROVES
cells in pars reticulata arise from cell bodies or primary dendrites. They immediately take a dorsal or dorso-medial course and can frequently be followed for several hundred microns, especially in coronal sections. In all cases, these were seen to travel dorsally and medially into the pars compacts. One, which was followed for some distance through the latter structure was seen to turn medially and follow a course very similar to that of the medium pars compacta neurons. Some axons show no branches in the substantia nigra (e.g., fig. 7). Many, however, emit an initial collateral within 8-84 p m of their initiation (fig. 9). Some of these turn back toward the cell of origin, while others remain in the general dorso-medial orientation of the main axon. It was not possible to determine the nature of the terminal field of such collaterals, nor whether they are restricted to pars reticulata. The initial trajectory of a few suggested the possibility that some may enter the pars compacta. Of the 63 cells sampled in pars reticulata, 29% were categorized as small neurons on the basis of soma size and dendritic field. These were identical in all respects to similarly sized neurons seen in other parts of the nucleus, being stellate in appearance and having no preferred axonal orientation
J Fig. 4 A camera lucida drawing of a small neuron in pars reticulata from a sagittal section. The cell is oriented so that rostra1 is toward the top of the figure and ventral to the right. The cell has a small, round soma, thin dendrites and a small dendritic field characteristic of small cells. It is presumed to be an interneuron. Golgi-Kopsch. Bar, 100 fim.
(fig. 4). Again, a few displayed some initial axonal branching, but it was not possible to determine their ultimate destination.
Pars lateralis As in pars reticulata, all three cell types may be observed in pars lateralis. The small neurons are identical to those described for pars compacta and pars reticulata. The large and medium cells, which again differ only in size, frequently have one or more long thick dendrites extending ventrally into pars reticulata where they join the dendrites running arallel to the peripeduncular layer (fig. 5 . These large dendrites often branch as they approach the CNS cerebri. Other dendrites remain in the pars lateralis, extending apparently in all directions, with some preference for the medio-lateral direction. The axons of large and medium cells of pars lateralis show an initial orientation in a dorsal and/or medial direction. Too few of these were observed for a sufficient distance to make any further generalizations concerning their trajectory, or to determine whether or not they possess initial collaterals.
P
DISCUSSION
The neural cell types found in our material correspond with those observed by Gulley and Wood (‘71) in one micron sections stained by toluidine blue from the rat. The difference in cell body diameter as measured here and those measured by Gulley and Wood is probably attributable simply to the thickness of the section in which measurement is taken. Taking this difference into account, the classification of cell types a rees with that found by Gulley and Wood f71). The distribution of cell types in the various areas of the nucleus in our material is also similar to that reported by Gulley and Wood. The only difference is that they did not find a significant number of medium-sized cells in pars reticulata as we report. Those authors found that 10%of pars compacta and 40% of pars reticulata neurons were of the small cell type. Con-
591
RAT SUBSTANTIA NIGRA
VTA
Fig. 5 A composite of representative cell types seen in the substantia nigra, their chracteristic dendritic fields and relationships within the nucleus. The cells are based on camera lucida drawings of actual neurons but for emphasis have been enlarged here relative to the nucleus. The view on the right is in a coronal plane and the left view in a sagittal plane. The text should be consulted for details. CC, crus cerebri, H2, Forel’s field H2, LM, medial lemniscus; VTA, ventral tegmental area.
sidering the variability inherent in the Golgi technique, our values of 18% and 29% respectively probably do not represent any significant discrepancy with their finding. Our observations are at variance, however, with the report of Hanaway et al. (’70) that small cells are not present in the pars compacta of rats at all. Although there are differences between our observations of the location of cells of various sizes and those made by Rinvik and Grofovh (’70), these may be attributed to species differences since these authors used cats. It is notable that in the rat we found cell types, d e h e d by their dendritic fields, comparable to those described by both Rinvik and Grofov6 (‘70) in cats and Schwyn and Fox (‘74) in monkeys. The dendritic varicosities present on many nigral neurons, which have also been reported in electron microscopic studies (e.g., Kemp and Powell, ’71), appeared initially to correspond to the fluorescent varicosities reported by Bjorklund and Lindvall (‘75). However, their ubiquity throughout the nucleus, including the dendrites of cells which certainly are not dopaminergic, and the failure of these to be seen consistently on the dendrites of pars compacta medium cells makes an exact correspondence appear unlikely. The “varicosities” seen with the fluorescence techniques probably do not represent actual changes in dendritic caliber, but
rather areas of high catecholamine accumulation. It may still be, however, that in the pars compacta medium cells dendritic varicosities are sites of accumulation of dopamine. The extensive overlap of dendritic fields in pars cornpacta which forms a dense network of dendrites in this area suggests that dendritic interactions could play a powerful role in determining the behavior of pars cornpacta neurons. Such interactions could take the form of dendro-dendritic synapses as suggested by Bjorklund and Lindvall (‘75). It is also possible, however, that dendritic release of dopamine occurs in the absence of dendro-dendritic contact between dopaminergic neurons as suggested by Groves et al. (‘75, ’76). The only instance of dendritic overlap with neurons of origin outside the substantia nigra was seen to occur between medium-sized pars compacts neurons and those of the also predominantly dopaminergic ventral tegmental area raising the possibility of dendritic interactions between the cells of these two areas as well. This overlap also suggests that pars compacta neurons could receive afferents from areas known to project to the ventral tegmental area, such as the nucleus accumbens (e.g., Bunney and Aghajanian, ’76). Several observations shown in figure 5 are worthy of comment. As reported in material from primates (Schwyn and Fox, ’74),
592
J. M. JURASKA, C. J. WILSON AND P. M. GROVES
most dendrites of pars reticulata neurons are seen to take a rostro-caudal course, with some descending to the edge of the crus cerebri. In addition, these dendrites have a component roughly perpendicular to the border of pars compacta, which generally gives them the appearance of a dorso-ventral orientation, but which may be nearly medio-lateral in the more medial areas of the nucleus. This can be seen in coronal sections, as well as the striking tendency for some pars reticulata cells to emit dendrites which align themselves along the crus cerebri. This latter tendency is also seen in the cells of pars lateralis, and in some pars compacta cells whose downward directed dendrites traverse the entire pars reticulata. This may reflect an orientation of all of these cells toward the peduncle, through which much of the descending input to the substantia nigra passes (e.g., Voneida, '60; Szabo, '62; Nauta and Mehler, '66). Pars lateralis has been considered a separate subdivision of the substantia nigra by some investigators, particularly those employin the Nissl techniques (Hanaway et al., '70 . This distinction is not often made in Golgi stained material (Cajal,'55; Rinvik and Grofova, '70; Schwyn and Fox, '74). In our observations the range of cell types was the same in pars lateralis as in pars reticulata. On the other hand, like those of pars compacta, pars lateralis cells send long dendrites into pars reticulata which join the course taken by some pars reticulata cells. Thus descending pars compacta dendrites are oriented rather perpendicular to the pars compacta cell layer as are the dendrites of some pars reticulata cells. In a similar fashion, pars lateralis dendrites orient with the pars reticulata cells of the peripeduncular layer. Pars lateralis then appears to share characteristics of both of the other areas of the nucleus. The small cells seen in all parts of the nucleus are usually presumed to be interneurons (Cajal, '55; Gulley and Wood, '71). They are characterized by their thin, delicate dendrites and small, nondirectional dendritic fields, the lack of preferred
7
orientation of their axons and at least some axon collaterals. While it is not possible from the present material to conclude the matter with certainty, it seems very likely that they do represent a variety of short axoned neurons. The medium-sized neurons of pars compacts, distinguishable from those of the other areas by their characteristic dendritic fields, are almost certainly the origin of the dopaminergic nigro-neostriatal tract. Their axonal trajectory, as far as it could be determined from our material, duplicates that seen in studies employing histofluorescence techniques, and chemically or electrolytically induced degeneration of the dopaminergic projections (Carpenter and Peter, '72; Hedreen, '71; Bjorklund and Lindvall, '75). Although our failure to demonstrate axon collaterals in Golgi preparations cannot stand as proof that they do not exist, axons of some of these neurons were followed well beyond the 8-84 p m distance in which pars reticulata neurons emitted their collaterals. These observations at least cast doubt on the presence of initial collaterals from medium pars compacta neurons. On the other hand, the medium and large neurons of pars reticulata often were seen to emit collaterals. At least some of these cells are presumed to be the origin of the nigro-thalamic tract due to their location and the orientation of their axons (Carpenter et al., '76; Faull and Carman, '68; Rinvik, '75; Afifi and Kaelber, '65). The functional role of the collaterals from these cells is unknown. In spite of their primitive, isodendritic appearance and similarity to cells of the reticular formation (Ramon-Mollinier, '75), the neurons of the substantia nigra are organized in three layers. The superior layer is the cellular pars compacta. Next is the dorsal aspect of pars reticulata, with dendrites running in dorso-ventral and rostrocaudal directions both from cells in pars compacta and pars reticulata itself. In the most ventral layer dendrites from all the areas of the nucleus run parallel to the crus cerebri. The significanceof this anatomical
RAT SUBSTANTIA NIGRA
organization may bear an important relation to the functions of the substantia nigra in a wide variety of behavioral phenomena. ACKNOWLEDGMENTS
This work was supported in part by Grant MH 19515 and Research Scientist Development Award KO2 MH 70706 from the National Institute of Mental Health, and Grant DA 01467 from the National Institute of Drug Abuse to P.M.G., and J. M. J. acknowledges support from NIH grant EY 01500 to Eva Fifkovi. The authors thank Eva Fifkova for providing microscopic facilities and advice in various stages of this work. LITERATURE CITED Afifi, A., and W. W. Kaelber 1965 Efferent connections of the substantia nigra in the cat. Exper. Neur., 11: 474-482. Andi-n, N. E . , A. Carlsson, A. Dahlstrom, K. Fuxe, N. A. Hillarp and K. Larsson 1964 Demonstration and mapping out of nigroneostriatal dopaminergic neurons. Life Sci., 3: 523-530. Bjorklund, A., and 0. Lindvall 1975 Dopamine in dendrites of substantia nigra neurons: Suggestions for a role in dendritic terminals. Brain Research, 83: 531-537. Bunney, B. S., and G. K. Aghajanian 1976 The precise localization of nigral afferents in the rat as determined by a retrograde tracing technique. Brain Research, in press. Cajal, S. R. 1955 Histologie du Systeum Nerveux de 1'Homme et des Vertebres. Vol. 11.Translation by L. Azoulay. Instituto Ramon y Cajal, Madrid, pp. 275278. Carpenter, M. B., K. Nakano and R. Kim 1976 Nigrothalamic projections in the monkey demonstrated by autoradiographic technics. J. Comp. Neur., 165: 401-416. Carpenter, M. B., andP. Peter 1972 Nigrostriatal and nigrothalamic fibers in the rhesus monkey. 1. Comp. Neur., 144: 93-116. Colonnier, M. 1964 The tangential organization of the visual cortex. J. Anat., 98: 327-344. Corrodi, H., K. Fuxe and P. Lidbrink 1972 Interaction between cholinergic and catecholaminergic neurons in rat brain. Brain Research, 43: 397-416. Dahlstrom, A., and K. F u e 1964 Evidence for the existence of monoamine-containing neurons in the central nervous system. I. Demonstration of monoamines in the cell bodies of brain stem neurons. Acta. Physiol. Scand. (Suppl.), 232: 5-55. F a d , R. L. M., and J. B. Carman 1968 Ascending projections of the substantia nigra in the rat. J. Comp. Neur., 132: 73-92. Fonnum, F., I. Grofova, E. Rinvik, J. Storm-Matisen
593
and F. Walberg 1974 Origin and distribution of glutamate decarboxylase in substantia nigra of the cat. Brain Research, 71; 77-92. German, D. C., and D. M. Bowden 1974 Catecholamine systems as the neural substrate for intracranial self-stimulation: An hypothesis. Brain Research, 73: 381-419. Grossman, R. G. 1966 Mechanisms of generation of potentials produced in the substantia nigra following stimulation of the neostriatum. Neurology, 16: 320. Groves, P. M., and G. V. Rebec 1976 Biochemistry and behavior: Some central actions of amphetamine and antipsychotic drugs. Ann. Rev. Psych., 27: 91127. Groves, P. M., C. J. Wilson, S. J. Young and G. V. Rebec 1975 Self-inhibition by dopaminergic neurons. Science, 190: 522-529. Groves, P. M., S. J. Young and C. J. Wilson 1976 Nigro-striatal relations and the mechanisms of action of amphetamine. In: Cholinergic-monoaminergic Interactions in the Brain. L. L. Butcher, ed. Academic Press, New York, in press. Gulley, R. L., and R. L. Wood 1971 The fine structure of the neurons in the rat substantia nigra. Tissue and Cell, 3: 675-690. Hanaway, J., J. A. McConnell and M. G. Netsky 1970 Cytoarchitecture of the substantia nigra in the rat. Am. J. Anat., 129: 417-438. Hedreen, J. C. 1971 Separate demonstration of dopaminergic and nondopaminergic projections of substantia nigra in the rat. Anat. Rec., 169: 338. Hornykiewicz, 0. 1966 Dopamine (3-hydroxytyramine) and brain function. Pharmacological Reviews, 18: 925-964. Kemp, J. M., and T. P. S. Powell 1971 The site of termination of afferent fibres in the caudate n u cleus. Phil. Trans. R. Soc. London B., 262: 413-427. Kim, J. 4, I. J. Bak, R. Hassler and Y. Okada 1971 Role of y-aminobutyric acid (GABA) in the extrapyramidal motor system. 2. Some evidence for the existence of a type of GABA-rich strio-nigral neurons. Exp. Brain Res., 14: 95-104. Lindvall, O., and A. Bjorklund 1974 The organization of the ascending catecholamine neuron systems in the rat brain as revealed by the glyoxylic acid fluorescence method. Acta Physiol. Scand. (Suppl.), 412: 4-48. Ljungdahl, A., T. Hokfelt, M. Goldstein and D. Park 1975 Retrograde peroxidase tracing of neurons combined with transmitter histochemistry. Brain Research, 84: 313-319. McGeer, P. L., H. C. Fibiger, T. Hattori, V. K. Singh, E. G. McGeer and L. Maler 1974 Biochemical neuroanatomy of the basal ganglia. Adv. Behav. Biol., 10: 27-47. Nauta, W. J. H., and W. R. Mehler 1966 Projections of the lentiform nucleus in the monkey. Brain Research, 1 : 3-42. Parizek, J., R. Hassler and I. J. Bak 1971 Light and electron microscopic autoradiography of substantia nigra of rat after intraventricular administration of
594
J. M. JURASKA, C. J. WILSON AND P. M. GROVES
tritium labeled norepinephrine, dopamine, serotonin and the precursors. Z. Zellforsch., 115: 137148. Precht, W., and M. Yoshida 1971 Blockage of caudate-evoked inhibition of neurons in the substantia nigra by picrotoxin. Brain Research, 32: 229-233. Ramon-Moliner,E. 1975 Specialized and generalized dendritic patterns. In: Golgi Centennial Symposium. Proceedings. M. Santi, ed. Raven Press, New York, pp. 87-100. Rinvik, E. 1975 Demonstration of nigro-thalamic connections in the cat by retrograde axonal transport of horseradish peroxidase. Brain Research, 90: 313-318. Rinvik, E., and I. Grofova 1970 Observations on the fine structure of the substantia nigra in the cat. Exp. Brain Res., 11: 229-248. Routtenberg, A., and Holtzman, H. 1973 Memory disruption by electrical stimulation of substantia nigra, pars compacta. Science, 181: 83-86. Schwyn, R. C . , and C. A. Fox 1974 The primate substantia nigra: A Golgi and electron microscopic study. J. Hirnforsch., 15: 95-126. Sotelo, C. 1971 The fine structural localization of Norepinephrine- 3H in the substantia nigra and area postrema of the rat: An autoradiographic study. J. Ultrastructure Res., 36: 824-841.
Stenevi, IJ., A. Bjorklund and R. Y. Moore 1973 Morphological plasticity of central adrenergic neurons. Brain, Behav. and Evol., 8: 110-134. Szabo, J. 1962 Topical distribution of the striatal efferents in the monkey. Exp. Neur., 5: 21-36. Ungerstedt, U. 1971 Stereotaxic mapping of the monoamine pathways in the rat brain. Acta Physiol. Scand., Suppl. 367: 1-48. Valverde, F. 1970 The Golgi method: A tool for comparative structural analysis. In: Contemporary Research Methods in Neuroanatomy. W. J. H. Nauta and S. 0. E. Ebbesson, eds. Springer-Verlag, New York, pp. 12-31. Van der Loos, H. 1956 Une combinaison de deux vieilles methods histologiques pour le systeme nerveux central. Monat. Psychiat. Neur., 132: 331-334. Voneida, T. J. 1960 An experimental study of the course and distribution of fihers arising in the head of the caudate nucleus in the cat and monkey. J. Comp. Neur., 115: 75-82. Zigmond, M. J., and E. M. Stricker 1972 Deficits in feeding behavior after intraventricular injection of 6-hydroxydopamine in rats. Science, 177: 12111212. Zis, A. P., H. C. Fihiger and A. G. Phillips 1974 Reversal by L-dopa of impaired learning due to destruction of the dopaminergic nigro-neostriatal projection. Science, 185: 960-962.
PLATE 1 EXPLANATION OF FIGURES
6 A montage of photomicrographs of a medium-sized pars compacta cell from a sagittal section. The cell is oriented so that rostral is toward the top of the figure and dorsal is to the right. The cell is located in the rostral end of pars compacta, therefore its longest dendrites course caudally into pars reticulata. Golgi-Kopsch. Bar, 50 pm.
7 A montage of photomicrographs of a large-sized cell in pars reticulata from a coronal section. The cell is oriented so that dorsal is toward the top and left of the figure and medial is to the right and top. This cell emits an axon from a dorsal dendritic trunk that runs dorso-medially for several hundred microns without a collateral. The dendrites course mainly in a dorso-ventral direction. Golgi-Kopsch. Bar, 50 pm.
RAT SUBSTANTIA NICRA J. M. Juraska,C. J. Wilson and P. M.Groves
PLATE I
595
PLATE 2 EXPLANATION OF FIGURES
8 A montage of photomicrographs of a medium-sized pars reticulata cell from a sagittal section. The cell is oriented so that dorsal is toward the top of the figure and rostral is to the right. The dendrites of this neuron primarily run rostro-caudally although there is a dorso-ventral component. Some spines can be seen on far rostral branches. Golgi-Kopsch. Bar, 50 pm.
9 A montage of photomicrographs of a medium-sized pars reticulata cell from a coronal section. The cell is oriented so that dorsal is toward the top of the figure and lateral is to the right. The arrow points to the initiation of an axon collateral frequently seen on cells of this type. The dendrites of the cell are oriented in a characteristic dorso-ventral direction. Golgi-Kopsch. Bar, 50 Ctm.
596
RAT SUBSTANTIA NIGRA
I’LATE 2
J. M. Juraska, C . J. Wilson and P. M. Groves
597
PLATE 3 EXPLANATION OF FIGURES
10
A photomicrograph of a small-sized cell in pars compacta from a sagittal section. The cell is oriented so that dorsal is toward the top of the figure and rostral is to the right. The cell, which is probably an interneuron, has thin dendrites and a small nondirectional dendritic field. Golgi-Kopsch. Bar, 50 Pm.
11 A photomicrograph of part of pars reticulata in coronal section. The figure is oriented so that dorsal is toward the top and lateral is to the right. The figure clearly shows the orientation of dendrites parallel to the crus cerebri while dorso-medial to this peripeduncular area, dendrites are oriented in a more dorso-ventral direction. CC, crus cerebri. Golgi-Cox. Bar, 100 pm.
598
RAT SUBSTANTIA NlGRA
PLATE 3
J. M. Juraska,C. J. Wilsoa and P. M. Groves
599
ABSTRACT Three variants of the Golgi method were employed to examine the cell types, their dendritic fields and organization and axonal trajectories within the substantia nigra of albino and hooded rats. In both sagittal and coronal sections, large, medium and small neurons were classified on the basis of soma size, extent of dendritic fields and dendritic caliber. In general nigral cells have three to five primary dendrites that branch relatively infrequently. Some dendrites of all cell types have thinly scattered spines or varicosities. Small cells, found in all areas of the nucleus, have thin dendrites and small, nondirectional dendritic fields. These are considered to be interneurons. The medium cells found in pars compacta, presumed to be the dopaminergic cells of the nigroneostriatal pathway, send long dendrites into pars reticulata perpendicular to the course of pars compacta. In addition, these cells have a number of dendrites which remain in pars compacta. These cells have axons that run medio-dorsally. No axon collaterals were detected. Both large and medium cells are found in pars reticulata. Cells in the dorso-medial aspect of pars reticulata orient rostro-caudally and roughly perpendicular to the course of pars compacta, while cells in the peripeduncular area show a strict orientation which is parallel to the crus cerebri. Some pars reticulata cells emit axon collaterals while others remain unbranched for their observable length. Both large and medium cells are also seen in pars lateralis. These cells send long dendrites ventrally into pars reticulata where they run parallel to the crus cerebri, while some shorter dendrites remain in pars lateralis. In total, the substantia nigra appears to have a layered organization: the superior layer is the cellular pars compacta, the second is the dorso-medial area of pars reticulata where both pars compacta and pars reticulata dendrites run rostro-caudally and dorso-ventrally and the third layer is the peripeduncular region where dendrites from all areas run parallel to the crus cerebri.
While its very existence was in question as little as a decade ago, the dopaminergic nigro-neostriatal pathway has now been implicated in a number of phenomena of psychological, pharmacological and clinical importance, including eating and drinking behavior (Zigmond and Stricker, '72), self-stimulation and learning (e.g., Routtenberg and Holtzman, '73; German and Bowden, '74; Zis et al., '74), the behavioral effects of amphetamine and the antipsychotic drugs (e.g., Groves and Rebec, '76), the etiology of Parkinson's disease (e.g., Hornykiewicz, '66), and postlesion axonal regeneration (e.g., Stenevi et al., '73). The source and trajectory of the nigro-neostriatal tract have been visualized with clarity J. COMP. NEUR., 172 585-600.
using the catecholamine histofluorescence methods of Falk and Hillarp (e.g., Andkn et al., '64;Un erstedt, '71) and Lindvall and Bjorklund '74). It arises from fluorescent cells of medium size located primarily in the pars compacta region of the substantia nigra, whose axons are seen to course dorsally and medially to form a fairly compact fascicle in the medial forebrain bundle just rostra1 to the substantia nigra. These cells are almost certainly the medium-sized cells which are seen in Nissl stained preparations to make up the majority of cells in the pars compacta, and which are specificially ' Reprints may be obtained from Philip M. Groves, Depart-
8
ment of Psychology, University of Colorado, Boulder, Colorado 80309.
585
586
J. M. JURASKA, C. J. WILSON AND P. M. GROVES
labeled after intraventricular administration of labeled catecholamines (Parizek et al., '71; Sotelo, '71; Ljungdahl et al., '75). The restriction of the dopaminergic cells to this cytoarchitecturally defined subdivision of the substantia nigra suggests the presence of spatially organized, perhaps very specific, arrangements of interacting neurons of functionally different types within this brain stem nucleus. Thus the activity of dopaminergic neurons might be regulated in part by their interactions with other areas of the substantia nigra. While some evidence is available relating dopaminergic activity to that of certain nigral afferents (e.g., Precht and Yoshida, '71; Corrodi et al., '72; Kim et al., '71; Fonnum et al., '74; McGeer et al., '74) very little is known of the local integration which might be expected to modify or transform that input. The first requirement for such an analysis of local interactions, a detailed description of the geometric organization of cell types and the dendritic fields of the cells, is reported here. Special attention has also been paid to the initial axonal trajectories of the various cell t es insofar as they may be determined in t ick sections prepared by the Golgi methods.
YR
MATERIALS AND METHODS
Three variants of the Golgi method were applied to tissue blocks taken from the midbrains of male and female rats. The material from Simonsen rats, between 36 and 100 days of age was stained by the GolgiKopsch method (Colonnier, '64). Brains of some Long-Evans hooded rats aged 55 days were stained by the Golgi-Cox method described by Van der Loos ('56), while others aged 5 to 20 days were stained by the rapid Golgi method according to Valverde ('70). Both coronal and sagittal sections were cut at 100 pm. A representative sample of 130 neurons was chosen from a much larger sample of observed neurons and drawn at a magnification of x 625 with the aid of a Leitz camera lucida. OBSERVATIONS
The substantia nigra, present throughout
the midbrain, lies dorsal to the crus cerebri and ventral to the medial lemniscus. Medially it is bordered by the ventral tegmental area. The largest part of the nucleus is occupied by the dense neuropil of ars reticdata, which in Golgi material can e seen to consist of a network of small fibers coursing generally in the longitudinal and dorso-ventral directions, and dendrites with a similar organization. The neuropil is interrupted here and there by large fiber bundles from the crus cerebri, and by cell bodies of various sizes. This is in striking contrast to the pars compacta, which forms a thin densely cellular layer enclosing pars reticdata on all sides except for that bordering on the crus cerebri. Here the cells are of similar size and shape, and dendrites follow the contour of the cellular layer with dendritic fields overlapping extensively. The least distinct region is pars lateralis, an elli tical region lying on the lateral edge o f t e brainstem between the dorsolateral pars reticdata and the ventral lateral border of the medial geniculate.
1
K
Cell types Three general classes of cells could be distinguished on the basis of cell body size, thickness and branching of dendrites, and extent of dendritic field. Since the transition from soma to dendritic trunk is gradual, the measurement of cell body size is somewhat arbitrary in Golgi preparations. For this reason, two measurements were made on each neuron, one along its longest dimension, and one perpendicular to the long axis. Such measurements did, as expected, yield a generally trimodal distribution of cell body sizes. It was not possible, however, to unambiguously assign a cell to any distinct category on the basis of soma size alone, except in the case of the largest nigral neurons. Seen in pars reticulata and pars lateralis, these large neurons measure 45-74 p m along their longest axis. In other respects they resemble the medium-sized cells of these regions (19-46 pm). On the other hand, a catego of cells with thin delicate dendrites ra 'ating in all directions but with relatively small dendritic
';tl
RAT SUBSTANTIA NIGRA
fields can easily be distinguished from the other categories, despite the extensive overla between their cell body size (1126 pm and that of the medium-sized neurons. Cell body shape also varies widely. Small neurons tend to have round cell bodies. Medium and large cells are more heterogeneous in shape, with some being fusiform, triangular, polygonal, or ovoid. Aside from some tendency for mediumsized pars compacta neurons to be pyramidal in shape, there is no discernible segregation of cell body shape according to region in the substantia nigra. Many neurons of all types in the substantia nigra have some dendrites with varying degrees of beading or varicosities (figs. 1,4, 6, 11).The varicosities are generally not found on thick dendritic trunks, and are most obvious on dendrites located at some distance from the cell body. This is apparently related to dendritic caliber, cells having thin dendrites exhibiting varicosities closer to the cell body. Varicosities are also more commonly seen in younger animals. This was particularly noticeable in material prepared from the 5- and 10-day-old animals. Many neurons of all cell types also have thinly scattered dendritic spines (e.g., figs. 1, 4, 8). Spines are infrequent on the cell body itself. Although not a strict rule, dendrites with varicosities seem to have fewer spines, and vice versa. Some cells are seen to have both, and it was not possible to place cells into categories based on these features.
587
P
Fig. 1 A camera lucida drawing of a mediumsized pars compacta cell from a coronal section. The cell is oriented so that dorsal is toward the top of the figure and lateral is to the right.The dendrite running in a ventral direction is in pars reticulata for most of its course. The arrow points to the axon. GolgiKopsch. Bar, 100 Fm.
its superior surface so that in sagittal sections, cells at the rostral and caudal poles of the nucleus have dendrites oriented rostrocaudally (fig. 6) while dendrites from cells of the middle portion of the nucleus have a dorso-ventral orientation. In coronal secPars compacta tions the dendrites going into pars reticuThe neurons of pars compacta are pri- lata have a ventro-lateral or ventral course marily of medium size. They resemble the (fig.1).When only one dendrite is sent into medium-sized neurons of pars reticdata pars reticulata, it is frequently the cell’s and pars lateralis, but can be immediately largest and longest, giving many of these recognized by their dendritic branching neurons the appearance of an “upside pattern. Each characteristically sends one down yramidal cell” as described by or two dendrites deep into pars reticulata others TCajal, ’55; Grossman, ’66; Schwyn where they may branch one or two times. and Fox, ’74), with its apical dendrite These dendrites generally run parallel to oriented into pars reticdata and its basilar the course of the pars reticulata dendrites, dendrites remaining in pars compacta (e.g., and often show varicosities (figs. 1,6). The fig. 1). The basilar dendrites, of which orientation of these dendrites as they enter there are two to five, frequently branch pars reticulata is roughly perpendicular to one or two times and run parallel to the
588
J. M. JURASKA, C . J. WILSON AND P. M. GROVES
surface of pars compacta, again regardless of their position in the nucleus. Occasionally these dendrites exhibit a tufted branching pattern, as seen in sagittal section in figure 5. They may course for some distance within pars compacta, going in both directions and so forming a disk-like dendritic field. Since the cells of pars compacta are so close together, these dendritic fields overlap extensively. Except in the case of the most medially placed neurons, the dendrites of pars compacta cells do not extend outside of the substantia nigra. In the case of cells situated on the border of the ventral tegmental area however, dendrites are often seen to extend medially and run for some distance into the latter nucleus where they intermingle with the dendrites of cells of similar appearance there. While no attempt was made to characterize the cells of the ventral tegmental area in detail, this observation is consistent with the observation that this region, which is also rich in dopamine-containing neurons, is continuous with the pars compacta (e.g., Dahlstrom and Fuxe, '64). Axons arising from some pars compacta medium cells could be followed for as far as 165 pm before passing out of the plane of the section. They arise from cell bodies or primary dendrites in sagittal sections coursing initially in a dorsal or dorso-caudal direction. After a variable distance, they take a sharp medial turn and are lost from view. In coronal sections their course appears dorsal and/or medial. In no case was an axon from this type of cell seen to emit a collateral. The small neurons of pars compacta are much less frequently seen, representing only 18% of our sample of 61 neurons drawn from this area. They have small stellate dendritic fields consisting of three to five thin dendrites which generally branch one or two times, although in some cases u to four branchings have been seen (fig. l O f Although they are identical in all respects to pars reticulata small neurons, their dendrites seem to be restricted to pars compacta. Their axons arise from the cell body or a primary dendrite and appear to
have no preferred direction. Only a small number of axons arising from these cells were followed for more than a few microns. In some of these cases however, they could be seen to branch although it was not possible to determine the extent of their branching or the total size of their axonal trajectory.
Pars reticulata Cells of all three types are seen throughout pars reticdata. Large and mediumsized cells, which differ only in cell body size (figs. 2, 3, 7-9), will be considered together. In sagittal sections through the middle portion of the nucleus, dendrites of both of these are seen to course primarily in an anterior-posterior direction with some tendency to spread apart in the dorso-ventral direction, es ecially ventral toward the crus cerebri &g. 8). Thus the dendritic fields of these cells give the appearance of a narrow plate running nearly the entire length of the nucleus in the sagittal plane. In coronal sections most dendrites generally run in a dorso-ventral direction with little medio-lateral component (figs. 7, 9). However as in pars compacts, the absolute orientation of dendritic fields varies with location within the nucleus. This is evident in coronal sections such as the one reconstructed in figure 5 in
Fig. 2 A camera lucida drawing of a large-sized pars reticulata cell from a coronal section. The cell is oriented so that dorsal is toward the top and to the right of the figure and lateral is to the right and bottom. The arrow points to the axon. This cell lies in the medial area of pars reticulata and its dendrites display the mediolateral orientation typical of cells in this location. Golgi-Kopsch. Bar, 100 pm.
RAT SUBSTANTIA NIGRA
which it can be seen that the dendrites of pars reticulata cells are oriented roughly perpendicular to the course of the pars compacta cell layer. This results in most dendrites having a strongly dorso-ventral orientation (figs. 7, 9) that changes somewhat to mediolateral in the extreme medial areas (fig. 2). Near the crus cerebri, however, a different organization is seen. Those dendrites of pars reticulata neurons which approach the crus cerebri tend to ramify and send branches to run for some distance parallel to the surface of the peduncle. In this area also, a second type of medium-large-sized pars reticulata neuron is seen, whose dendritic field runs lengthwise along the crus cerebri (fig. 11). These peripeduncular cells are particularly prominent in coronal sections through the midportion of the nucleus where their dendrites, and the ventral branches of the more dorsal pars reticulata neurons give the nucleus a distinctly layered appearance. In the peripeduncular region, all dendrites appear to
589
run primarily in the coronal plane, and parallel to the surface of the brainstem, while the bulk of the pars reticulata, lying dorsal to this layer, has a more dorso-ventral orientation. Dendrites of pars reticulata tend not to leave this region. Cells located near the pars compacta or pars lateralis may send dendrites into those regions, but do not course through them for any great distance. Both large and medium cells of pars reticulata often have thick dendritic trunks arising from the cell body. These are especially noticeable in many of the larger cells (figs. 7, 8), and may give rise to long dendrites measuring up to 590 p m in length which are sometimes unbranched throughout most of their course. Generally they do however, branch once or twice near the cell body, and again if they emit a branch running near the crus cerebri. Dendrites in the peripeduncular layer often display prominent varicosities (fig. 11). Axons from both the large and medium
Fig. 3 A camera lucida drawing of a medium-sizedpars reticulata cell from a sagittal section. The cell is oriented so that dorsal is toward the top of the figure and rostra1 is to the right. The long rostrocaudal course of this cell’s dendrites can be noted, as well as the tendency to gradually spread out in a dorso-ventral direction. Golgi-Kopsch.Bar, 100 pm.
590
J. M. JURASKA,C. J. WILSON AND P. M. GROVES
cells in pars reticulata arise from cell bodies or primary dendrites. They immediately take a dorsal or dorso-medial course and can frequently be followed for several hundred microns, especially in coronal sections. In all cases, these were seen to travel dorsally and medially into the pars compacts. One, which was followed for some distance through the latter structure was seen to turn medially and follow a course very similar to that of the medium pars compacta neurons. Some axons show no branches in the substantia nigra (e.g., fig. 7). Many, however, emit an initial collateral within 8-84 p m of their initiation (fig. 9). Some of these turn back toward the cell of origin, while others remain in the general dorso-medial orientation of the main axon. It was not possible to determine the nature of the terminal field of such collaterals, nor whether they are restricted to pars reticulata. The initial trajectory of a few suggested the possibility that some may enter the pars compacta. Of the 63 cells sampled in pars reticulata, 29% were categorized as small neurons on the basis of soma size and dendritic field. These were identical in all respects to similarly sized neurons seen in other parts of the nucleus, being stellate in appearance and having no preferred axonal orientation
J Fig. 4 A camera lucida drawing of a small neuron in pars reticulata from a sagittal section. The cell is oriented so that rostra1 is toward the top of the figure and ventral to the right. The cell has a small, round soma, thin dendrites and a small dendritic field characteristic of small cells. It is presumed to be an interneuron. Golgi-Kopsch. Bar, 100 fim.
(fig. 4). Again, a few displayed some initial axonal branching, but it was not possible to determine their ultimate destination.
Pars lateralis As in pars reticulata, all three cell types may be observed in pars lateralis. The small neurons are identical to those described for pars compacta and pars reticulata. The large and medium cells, which again differ only in size, frequently have one or more long thick dendrites extending ventrally into pars reticulata where they join the dendrites running arallel to the peripeduncular layer (fig. 5 . These large dendrites often branch as they approach the CNS cerebri. Other dendrites remain in the pars lateralis, extending apparently in all directions, with some preference for the medio-lateral direction. The axons of large and medium cells of pars lateralis show an initial orientation in a dorsal and/or medial direction. Too few of these were observed for a sufficient distance to make any further generalizations concerning their trajectory, or to determine whether or not they possess initial collaterals.
P
DISCUSSION
The neural cell types found in our material correspond with those observed by Gulley and Wood (‘71) in one micron sections stained by toluidine blue from the rat. The difference in cell body diameter as measured here and those measured by Gulley and Wood is probably attributable simply to the thickness of the section in which measurement is taken. Taking this difference into account, the classification of cell types a rees with that found by Gulley and Wood f71). The distribution of cell types in the various areas of the nucleus in our material is also similar to that reported by Gulley and Wood. The only difference is that they did not find a significant number of medium-sized cells in pars reticulata as we report. Those authors found that 10%of pars compacta and 40% of pars reticulata neurons were of the small cell type. Con-
591
RAT SUBSTANTIA NIGRA
VTA
Fig. 5 A composite of representative cell types seen in the substantia nigra, their chracteristic dendritic fields and relationships within the nucleus. The cells are based on camera lucida drawings of actual neurons but for emphasis have been enlarged here relative to the nucleus. The view on the right is in a coronal plane and the left view in a sagittal plane. The text should be consulted for details. CC, crus cerebri, H2, Forel’s field H2, LM, medial lemniscus; VTA, ventral tegmental area.
sidering the variability inherent in the Golgi technique, our values of 18% and 29% respectively probably do not represent any significant discrepancy with their finding. Our observations are at variance, however, with the report of Hanaway et al. (’70) that small cells are not present in the pars compacta of rats at all. Although there are differences between our observations of the location of cells of various sizes and those made by Rinvik and Grofovh (’70), these may be attributed to species differences since these authors used cats. It is notable that in the rat we found cell types, d e h e d by their dendritic fields, comparable to those described by both Rinvik and Grofov6 (‘70) in cats and Schwyn and Fox (‘74) in monkeys. The dendritic varicosities present on many nigral neurons, which have also been reported in electron microscopic studies (e.g., Kemp and Powell, ’71), appeared initially to correspond to the fluorescent varicosities reported by Bjorklund and Lindvall (‘75). However, their ubiquity throughout the nucleus, including the dendrites of cells which certainly are not dopaminergic, and the failure of these to be seen consistently on the dendrites of pars compacta medium cells makes an exact correspondence appear unlikely. The “varicosities” seen with the fluorescence techniques probably do not represent actual changes in dendritic caliber, but
rather areas of high catecholamine accumulation. It may still be, however, that in the pars compacta medium cells dendritic varicosities are sites of accumulation of dopamine. The extensive overlap of dendritic fields in pars cornpacta which forms a dense network of dendrites in this area suggests that dendritic interactions could play a powerful role in determining the behavior of pars cornpacta neurons. Such interactions could take the form of dendro-dendritic synapses as suggested by Bjorklund and Lindvall (‘75). It is also possible, however, that dendritic release of dopamine occurs in the absence of dendro-dendritic contact between dopaminergic neurons as suggested by Groves et al. (‘75, ’76). The only instance of dendritic overlap with neurons of origin outside the substantia nigra was seen to occur between medium-sized pars compacts neurons and those of the also predominantly dopaminergic ventral tegmental area raising the possibility of dendritic interactions between the cells of these two areas as well. This overlap also suggests that pars compacta neurons could receive afferents from areas known to project to the ventral tegmental area, such as the nucleus accumbens (e.g., Bunney and Aghajanian, ’76). Several observations shown in figure 5 are worthy of comment. As reported in material from primates (Schwyn and Fox, ’74),
592
J. M. JURASKA, C. J. WILSON AND P. M. GROVES
most dendrites of pars reticulata neurons are seen to take a rostro-caudal course, with some descending to the edge of the crus cerebri. In addition, these dendrites have a component roughly perpendicular to the border of pars compacta, which generally gives them the appearance of a dorso-ventral orientation, but which may be nearly medio-lateral in the more medial areas of the nucleus. This can be seen in coronal sections, as well as the striking tendency for some pars reticulata cells to emit dendrites which align themselves along the crus cerebri. This latter tendency is also seen in the cells of pars lateralis, and in some pars compacta cells whose downward directed dendrites traverse the entire pars reticulata. This may reflect an orientation of all of these cells toward the peduncle, through which much of the descending input to the substantia nigra passes (e.g., Voneida, '60; Szabo, '62; Nauta and Mehler, '66). Pars lateralis has been considered a separate subdivision of the substantia nigra by some investigators, particularly those employin the Nissl techniques (Hanaway et al., '70 . This distinction is not often made in Golgi stained material (Cajal,'55; Rinvik and Grofova, '70; Schwyn and Fox, '74). In our observations the range of cell types was the same in pars lateralis as in pars reticulata. On the other hand, like those of pars compacta, pars lateralis cells send long dendrites into pars reticulata which join the course taken by some pars reticulata cells. Thus descending pars compacta dendrites are oriented rather perpendicular to the pars compacta cell layer as are the dendrites of some pars reticulata cells. In a similar fashion, pars lateralis dendrites orient with the pars reticulata cells of the peripeduncular layer. Pars lateralis then appears to share characteristics of both of the other areas of the nucleus. The small cells seen in all parts of the nucleus are usually presumed to be interneurons (Cajal, '55; Gulley and Wood, '71). They are characterized by their thin, delicate dendrites and small, nondirectional dendritic fields, the lack of preferred
7
orientation of their axons and at least some axon collaterals. While it is not possible from the present material to conclude the matter with certainty, it seems very likely that they do represent a variety of short axoned neurons. The medium-sized neurons of pars compacts, distinguishable from those of the other areas by their characteristic dendritic fields, are almost certainly the origin of the dopaminergic nigro-neostriatal tract. Their axonal trajectory, as far as it could be determined from our material, duplicates that seen in studies employing histofluorescence techniques, and chemically or electrolytically induced degeneration of the dopaminergic projections (Carpenter and Peter, '72; Hedreen, '71; Bjorklund and Lindvall, '75). Although our failure to demonstrate axon collaterals in Golgi preparations cannot stand as proof that they do not exist, axons of some of these neurons were followed well beyond the 8-84 p m distance in which pars reticulata neurons emitted their collaterals. These observations at least cast doubt on the presence of initial collaterals from medium pars compacta neurons. On the other hand, the medium and large neurons of pars reticulata often were seen to emit collaterals. At least some of these cells are presumed to be the origin of the nigro-thalamic tract due to their location and the orientation of their axons (Carpenter et al., '76; Faull and Carman, '68; Rinvik, '75; Afifi and Kaelber, '65). The functional role of the collaterals from these cells is unknown. In spite of their primitive, isodendritic appearance and similarity to cells of the reticular formation (Ramon-Mollinier, '75), the neurons of the substantia nigra are organized in three layers. The superior layer is the cellular pars compacta. Next is the dorsal aspect of pars reticulata, with dendrites running in dorso-ventral and rostrocaudal directions both from cells in pars compacta and pars reticulata itself. In the most ventral layer dendrites from all the areas of the nucleus run parallel to the crus cerebri. The significanceof this anatomical
RAT SUBSTANTIA NIGRA
organization may bear an important relation to the functions of the substantia nigra in a wide variety of behavioral phenomena. ACKNOWLEDGMENTS
This work was supported in part by Grant MH 19515 and Research Scientist Development Award KO2 MH 70706 from the National Institute of Mental Health, and Grant DA 01467 from the National Institute of Drug Abuse to P.M.G., and J. M. J. acknowledges support from NIH grant EY 01500 to Eva Fifkovi. The authors thank Eva Fifkova for providing microscopic facilities and advice in various stages of this work. LITERATURE CITED Afifi, A., and W. W. Kaelber 1965 Efferent connections of the substantia nigra in the cat. Exper. Neur., 11: 474-482. Andi-n, N. E . , A. Carlsson, A. Dahlstrom, K. Fuxe, N. A. Hillarp and K. Larsson 1964 Demonstration and mapping out of nigroneostriatal dopaminergic neurons. Life Sci., 3: 523-530. Bjorklund, A., and 0. Lindvall 1975 Dopamine in dendrites of substantia nigra neurons: Suggestions for a role in dendritic terminals. Brain Research, 83: 531-537. Bunney, B. S., and G. K. Aghajanian 1976 The precise localization of nigral afferents in the rat as determined by a retrograde tracing technique. Brain Research, in press. Cajal, S. R. 1955 Histologie du Systeum Nerveux de 1'Homme et des Vertebres. Vol. 11.Translation by L. Azoulay. Instituto Ramon y Cajal, Madrid, pp. 275278. Carpenter, M. B., K. Nakano and R. Kim 1976 Nigrothalamic projections in the monkey demonstrated by autoradiographic technics. J. Comp. Neur., 165: 401-416. Carpenter, M. B., andP. Peter 1972 Nigrostriatal and nigrothalamic fibers in the rhesus monkey. 1. Comp. Neur., 144: 93-116. Colonnier, M. 1964 The tangential organization of the visual cortex. J. Anat., 98: 327-344. Corrodi, H., K. Fuxe and P. Lidbrink 1972 Interaction between cholinergic and catecholaminergic neurons in rat brain. Brain Research, 43: 397-416. Dahlstrom, A., and K. F u e 1964 Evidence for the existence of monoamine-containing neurons in the central nervous system. I. Demonstration of monoamines in the cell bodies of brain stem neurons. Acta. Physiol. Scand. (Suppl.), 232: 5-55. F a d , R. L. M., and J. B. Carman 1968 Ascending projections of the substantia nigra in the rat. J. Comp. Neur., 132: 73-92. Fonnum, F., I. Grofova, E. Rinvik, J. Storm-Matisen
593
and F. Walberg 1974 Origin and distribution of glutamate decarboxylase in substantia nigra of the cat. Brain Research, 71; 77-92. German, D. C., and D. M. Bowden 1974 Catecholamine systems as the neural substrate for intracranial self-stimulation: An hypothesis. Brain Research, 73: 381-419. Grossman, R. G. 1966 Mechanisms of generation of potentials produced in the substantia nigra following stimulation of the neostriatum. Neurology, 16: 320. Groves, P. M., and G. V. Rebec 1976 Biochemistry and behavior: Some central actions of amphetamine and antipsychotic drugs. Ann. Rev. Psych., 27: 91127. Groves, P. M., C. J. Wilson, S. J. Young and G. V. Rebec 1975 Self-inhibition by dopaminergic neurons. Science, 190: 522-529. Groves, P. M., S. J. Young and C. J. Wilson 1976 Nigro-striatal relations and the mechanisms of action of amphetamine. In: Cholinergic-monoaminergic Interactions in the Brain. L. L. Butcher, ed. Academic Press, New York, in press. Gulley, R. L., and R. L. Wood 1971 The fine structure of the neurons in the rat substantia nigra. Tissue and Cell, 3: 675-690. Hanaway, J., J. A. McConnell and M. G. Netsky 1970 Cytoarchitecture of the substantia nigra in the rat. Am. J. Anat., 129: 417-438. Hedreen, J. C. 1971 Separate demonstration of dopaminergic and nondopaminergic projections of substantia nigra in the rat. Anat. Rec., 169: 338. Hornykiewicz, 0. 1966 Dopamine (3-hydroxytyramine) and brain function. Pharmacological Reviews, 18: 925-964. Kemp, J. M., and T. P. S. Powell 1971 The site of termination of afferent fibres in the caudate n u cleus. Phil. Trans. R. Soc. London B., 262: 413-427. Kim, J. 4, I. J. Bak, R. Hassler and Y. Okada 1971 Role of y-aminobutyric acid (GABA) in the extrapyramidal motor system. 2. Some evidence for the existence of a type of GABA-rich strio-nigral neurons. Exp. Brain Res., 14: 95-104. Lindvall, O., and A. Bjorklund 1974 The organization of the ascending catecholamine neuron systems in the rat brain as revealed by the glyoxylic acid fluorescence method. Acta Physiol. Scand. (Suppl.), 412: 4-48. Ljungdahl, A., T. Hokfelt, M. Goldstein and D. Park 1975 Retrograde peroxidase tracing of neurons combined with transmitter histochemistry. Brain Research, 84: 313-319. McGeer, P. L., H. C. Fibiger, T. Hattori, V. K. Singh, E. G. McGeer and L. Maler 1974 Biochemical neuroanatomy of the basal ganglia. Adv. Behav. Biol., 10: 27-47. Nauta, W. J. H., and W. R. Mehler 1966 Projections of the lentiform nucleus in the monkey. Brain Research, 1 : 3-42. Parizek, J., R. Hassler and I. J. Bak 1971 Light and electron microscopic autoradiography of substantia nigra of rat after intraventricular administration of
594
J. M. JURASKA, C. J. WILSON AND P. M. GROVES
tritium labeled norepinephrine, dopamine, serotonin and the precursors. Z. Zellforsch., 115: 137148. Precht, W., and M. Yoshida 1971 Blockage of caudate-evoked inhibition of neurons in the substantia nigra by picrotoxin. Brain Research, 32: 229-233. Ramon-Moliner,E. 1975 Specialized and generalized dendritic patterns. In: Golgi Centennial Symposium. Proceedings. M. Santi, ed. Raven Press, New York, pp. 87-100. Rinvik, E. 1975 Demonstration of nigro-thalamic connections in the cat by retrograde axonal transport of horseradish peroxidase. Brain Research, 90: 313-318. Rinvik, E., and I. Grofova 1970 Observations on the fine structure of the substantia nigra in the cat. Exp. Brain Res., 11: 229-248. Routtenberg, A., and Holtzman, H. 1973 Memory disruption by electrical stimulation of substantia nigra, pars compacta. Science, 181: 83-86. Schwyn, R. C . , and C. A. Fox 1974 The primate substantia nigra: A Golgi and electron microscopic study. J. Hirnforsch., 15: 95-126. Sotelo, C. 1971 The fine structural localization of Norepinephrine- 3H in the substantia nigra and area postrema of the rat: An autoradiographic study. J. Ultrastructure Res., 36: 824-841.
Stenevi, IJ., A. Bjorklund and R. Y. Moore 1973 Morphological plasticity of central adrenergic neurons. Brain, Behav. and Evol., 8: 110-134. Szabo, J. 1962 Topical distribution of the striatal efferents in the monkey. Exp. Neur., 5: 21-36. Ungerstedt, U. 1971 Stereotaxic mapping of the monoamine pathways in the rat brain. Acta Physiol. Scand., Suppl. 367: 1-48. Valverde, F. 1970 The Golgi method: A tool for comparative structural analysis. In: Contemporary Research Methods in Neuroanatomy. W. J. H. Nauta and S. 0. E. Ebbesson, eds. Springer-Verlag, New York, pp. 12-31. Van der Loos, H. 1956 Une combinaison de deux vieilles methods histologiques pour le systeme nerveux central. Monat. Psychiat. Neur., 132: 331-334. Voneida, T. J. 1960 An experimental study of the course and distribution of fihers arising in the head of the caudate nucleus in the cat and monkey. J. Comp. Neur., 115: 75-82. Zigmond, M. J., and E. M. Stricker 1972 Deficits in feeding behavior after intraventricular injection of 6-hydroxydopamine in rats. Science, 177: 12111212. Zis, A. P., H. C. Fihiger and A. G. Phillips 1974 Reversal by L-dopa of impaired learning due to destruction of the dopaminergic nigro-neostriatal projection. Science, 185: 960-962.
PLATE 1 EXPLANATION OF FIGURES
6 A montage of photomicrographs of a medium-sized pars compacta cell from a sagittal section. The cell is oriented so that rostral is toward the top of the figure and dorsal is to the right. The cell is located in the rostral end of pars compacta, therefore its longest dendrites course caudally into pars reticulata. Golgi-Kopsch. Bar, 50 pm.
7 A montage of photomicrographs of a large-sized cell in pars reticulata from a coronal section. The cell is oriented so that dorsal is toward the top and left of the figure and medial is to the right and top. This cell emits an axon from a dorsal dendritic trunk that runs dorso-medially for several hundred microns without a collateral. The dendrites course mainly in a dorso-ventral direction. Golgi-Kopsch. Bar, 50 pm.
RAT SUBSTANTIA NICRA J. M. Juraska,C. J. Wilson and P. M.Groves
PLATE I
595
PLATE 2 EXPLANATION OF FIGURES
8 A montage of photomicrographs of a medium-sized pars reticulata cell from a sagittal section. The cell is oriented so that dorsal is toward the top of the figure and rostral is to the right. The dendrites of this neuron primarily run rostro-caudally although there is a dorso-ventral component. Some spines can be seen on far rostral branches. Golgi-Kopsch. Bar, 50 pm.
9 A montage of photomicrographs of a medium-sized pars reticulata cell from a coronal section. The cell is oriented so that dorsal is toward the top of the figure and lateral is to the right. The arrow points to the initiation of an axon collateral frequently seen on cells of this type. The dendrites of the cell are oriented in a characteristic dorso-ventral direction. Golgi-Kopsch. Bar, 50 Ctm.
596
RAT SUBSTANTIA NIGRA
I’LATE 2
J. M. Juraska, C . J. Wilson and P. M. Groves
597
PLATE 3 EXPLANATION OF FIGURES
10
A photomicrograph of a small-sized cell in pars compacta from a sagittal section. The cell is oriented so that dorsal is toward the top of the figure and rostral is to the right. The cell, which is probably an interneuron, has thin dendrites and a small nondirectional dendritic field. Golgi-Kopsch. Bar, 50 Pm.
11 A photomicrograph of part of pars reticulata in coronal section. The figure is oriented so that dorsal is toward the top and lateral is to the right. The figure clearly shows the orientation of dendrites parallel to the crus cerebri while dorso-medial to this peripeduncular area, dendrites are oriented in a more dorso-ventral direction. CC, crus cerebri. Golgi-Cox. Bar, 100 pm.
598
RAT SUBSTANTIA NlGRA
PLATE 3
J. M. Juraska,C. J. Wilsoa and P. M. Groves
599
