- Email: [email protected]
Heterogeneous distribution of the limbic system-associated membrane protein in the caudate nucleus and substantia nigra of the cat
0306-4522/91$3.00+ 0.00 Pergamon Press plc 0 1991IBRO
NeuroscienceVol. 40, No. 3, pp. 725-733, 1991 Printed in Great Britain
HETEROGENEOUS DISTRIBUTION OF THE LIMBIC SYSTEM-ASSOCIATED MEMBRANE PROTEIN IN THE CAUDATE NUCLEUS AND SUBSTANTIA NIGRA OF THE CAT *Department
M.-F. Cr-mssnLEr,*t~ C. GONZALES* and P. Luvtrrt University of Pennsylvania, 36 and Hamilton Walk, Philadelphia, PA
of Pharmacology, tDepartment
19104, U.S.A. of Anatomy, The Medical College of Pennsylvania, U.S.A.
Al&r&-The limbic system-associated membrane protein is a glycoprotein selectively associated, in the adult, with dendrites and cell bodies of neurons of the limbic system and related brain regions. In the present study, the distribution of the limbic system-associated membrane protein was studied by immunohistochemistry in the caudate nucleus and substantia nigra of the cat to determine how its expression relates to the compartmentalization of these areas. In all areas of the caudate nucleus, the pattern of limbic system-associated membrane protein immunoreactivity was highly heterogeneous, labeling zones that were in register with areas expressing neurochemical markers that classically identify striosomes. The extrastriosomal matrix exhibited low levels of staining. The results show that the limbic system-associated membrane protein is expressed by neurons within the target areas (striosomes) of subsets of limbic afferents (originating mainly from the basolateral nucleus of the amygdala and the prefrontal cortex), whereas regions of the caudate nucleus (extrastriosomal matrix) receiving inputs from other subdivisions of the limbic system, such as the cingulate cortex and the ventral tegmental area, contain relatively low levels of limbic system-associated membrane protein immunoreactivity. Thus the expression of this antigen may reflect the targeting of specific groups of limbic afferents to regions that are intimately associated with distinct components of the limbic system. The presence of limbic system-associated membrane protein in neurons of the substantia nigra pars compacta does not appear to be related to the presence or absence of the protein in their striatal target areas. In the substantia nigra, immunoreactivity to the limbic system-associated membrane protein was intense in the cell-sparse zone of the pars compacta, an area known to project to the extrastriosomal matrix of the caudate nucleus. This contrasted with the absence of immunostaining in areas containing dense clusters of dopaminergic neurons (densocellular zone), which project to the limbic system-associated protein-rich striosomes. By analogy to findings in the caudate nucleus and in other brain areas, the results suggest that subgroups of nigral dopaminergic neurons identified on the basis of their terminal fields in the caudate nucleus, may also differ in their limbic afferents.
Work in the nervous system of vertebrates and invertebrates has shown that some anatomically connected neurons share common membrane antigens. ‘,14.15~22 These “system-associated” antigens may serve as identification molecules in the development of neuronal networks involved in a common function.8.20,22In the rat, one such membrane antigen, the limbic system-associated membrane protein (LAMP), is a 64,000-68,000 mol. wt glycoprotein33 that is present during early development on axons, dendrites and cell bodies of neurons belonging to the limbic system,‘6*23an ensemble of brain structures involved in emotional behavior. LAMP immunoreactivity is also expressed in the adult central nervous system, but is restricted to neuronal cell bodies and their dendrites following synaptogenesis.‘6’22’23,33Staining has been found in such limbic areas as prefrontal, $To whom correspondence should be addressed. Abbreviations:
AChE, acetylcholinesterase; LAMP, limbic system-associated membrane protein; PBS, phosphatebuffered saline.
cingulate and perirhinal cortex, hippocampus, septum, amygdala, midline thalamic nuclei and hypothalamic regions. In addition, LAMP immunoreactivity is present in autonomic nuclei of the brainstem that are intimately associated with limbic forebrain regions. **,*’This location, and the presence of LAMP staining on cell bodies and dendrites, but not on axon terminals33 suggests that, in the adult, the molecule may signal areas of the central nervous system which receive afferents from subsets of limbic structures. The striatum (caudate nucleus and putamen) receives inputs from a variety of limbic areas.2~3,6,2’.2*,30 In adult rats, immunostaining for LAMP was much more intense in the ventromedial part of the striatum than in its dorsolateral region.** The area of intense LAMP immunoreactivity corresponds to the region of the caudate-putamen which receives inputs from the basolateral nucleus of the amygdala, a part of the limbic system, whereas the LAMP-negative region overlaps with an area receiving massive projections from the sensorimotor cortex.*’ This pattern of 725
726
M.-F.
~HiiSSELE’l’
LAMP immunoreactivity provides a strong argument in favor of the presence of LAMP in neuronal targets of limbic inputs, even in regions usually not classified as limbic structures themselves.22 Recent work on the connectivity of the caudate nucleus, however, has revealed a more complex organization of limbic inputs to this area. In the cat, afferents from different parts of the limbic system are segregated among two distinct striatal compartments. In particular, the basolateral nucleus of the amygdala projects to distinct patchy areas of the caudate nucleus called “striosomes”, but not to the extrastriosomal matrix surrounding them.30 The striosomes also receive inputs from some areas of the prefrontal cortex associated with the limbic system. zB-3oOther parts of the limbic circuitry such as the cingulate cortex and the ventral tegmental area, however, project preferentially to the extrastriosomal matrix and avoid the striosomes.is.** 3o A similar ~~egation of limbic inputs from cerebral cortex and mesencephalic dopaminergic cell groups to the striatum has been observed in the rat.2s’.7 Non-limbic inputs to the caudate nucleus are also distin~ished among these striatal subdivisions. In particular, subgroups of pars compacta neurons, which form one of the main afferent pathways to the caudate nucleus, project either to the striosomal or extrastriosomal compartments of the striatum.6,‘8 In view of the complex relationship between the striatum and distinct regions of the limbic system, we have now examined whether the distribution of LAMP immunorea~tivity in the caudate nucleus of the cat reflects the anatomical segregation of limbic inputs to distinct compartments in this structure. In addition, we examined the pattern of staining for LAMP in the substantia nigra to determine whether it was related to the organization of nigral projections to striatal subregions. In the caudate nucleus, we took advantage of the fact that the striatal compartments can be distinguished from each other based on the specific distribution of neurochemical markers. In particular, striosomes exhibit sparse staining for acetylcholinesterase (AChE)‘* and, in the dorsolateral striatum, intense substance P-like immunorea~tivity.“*13 Adjacent sections of caudate nuclei of adult cats were processed for LAMP localization using immunoperoxidase staining, and for AChE histochemistry or substance P-like immunorea~tivity, in order to identify the striosomes. In the substantia nigra, neurons of the pars compacta were identified by immunostaining with an antibody against tyrosine hydroxylase, the rate limiting enzyme of dopamine synthesis.”
<‘f d
EXPERIMENTAL
PROCEDGRES
Four adult cats were used in this study. Three cats were perfused through the heart with a mixture of 4010 paraformaldehyde and 0.1% glutaraldehyde and one with 4% paraformaldehyde only. In all cases, the fixative was freshly prepared in sodium phosphate-buffered saline (PBS; 0.1 M, pH 7.4) and administered after a brief flush of saline. The brains were removed after I h postfixation in situ. cryoprotected and stored at 4°C in PBS containing sucrose (30%). The brains were frozen on dry ice and sections (35 pm thick) were cut on a sliding microtome, collected in PBS and processed either for immunohistochemistry or AChE histochemistry. Sodium sulfate-precipitated anti-LAMP was prepared from culture supernatants in which the hybridoma clone 2G9 was grown.22 The antibody precipitate was dissolved in water and dialysed extensively against PBS. The sections were incubated-for 24 h at 4°C in a I : 1000 dilution of the antibodv in PBS containing 4% non-fat drv milk. Followine extensive washing, the sections were incubated for 1h at room temperature in a 1: 50 dilution of horseradish peroxidase-conjugated goat anti-mouse IgG in 4% non-fat dry milk. Sections were rinsed and the peroxidase reaction product was developed using a standard 3,3’-diaminobenzidine-peroxidaseH,O, method.22,23,33 Substance P-like immunohistochemistrv was performed in tissue from the cat perfused with para~o~a~ehyde alone, using an antibody from INCSTAR (I:500 in Tris-buffered saline: 0.5 M, pH 7.4) and the avidin-biotin technique,” with a Vectastain kit (Vector Laboratories), following the instructions of the manufacturer. Sections incubated with primary antibody that had been preincubated overnight at 4°C with substance P (foe4 M) did not show any imnlunos~ining. Tyrosine hydroxylase-like immunoreactivity was detected using an antibodv from Eugene Tech. Inc. diluted 1:1200 in PBS and peroxid&-antipeyoxidase reaction as described.‘” Controls included omission of the primary antibody or of the peroxidase~nti~roxidase complex. AChE h~sto~hemist~ was performed as described by Joyce and MarshallI with slight modifications. Briefly, sections were incubated S-6 h in 0.1 M acetate buffer (pH 5.3) containing 0.1 M CuSO,-glycine solution. After incubation, sections were rinsed in dH,O, developed in 4.5% K, Fe(CN), in dH,O for 15 min then rinsed three times in dH,O before being mounted on gelatin-coated slides, air-dried, dehydrated and mounted with Eukitt.
RESULTS
In all areas examined, LAMP immunoreactivity was present within the neuropil between profiles of unstained somata but was not resolved to specific morphological structures. In thinner sections, however, LAMP immunoreactivity was detected pericellularly along cell bodies and processes. This is in agreement with previous ultrastructural studies in the adult rat which have identified the plasma membrane of neuronal somata and dendrites as the specific cellular compartments that exhibit LAMP immunoreactivity.33
Fig. 1. Photomicro~ap~s depict coincident striosomal (AChEpoor) localization of LAMP. Patches of low AChE activity (A, C: pale zones, examples at white arrows) and of dense LAMP immunolabeling (9, D: dark zones, examples at black arrows) in the caudate nucleus of the cat. A, 9; C, D are sets of adjacent sections. Note the matches between cholinesterase-poor and LAMP-rich areas (white and black arrows, respectively), except for areas indicated by open arrows in A and D. Scale bars = 500 pm.
Fig.
1
72-l
728
M.-F. CHESSELET cr ul
Fig. 2. Comparison between the heterogeneous patterns of substance P (A, C) and LAMP (I?, D) immunostaining in two sets of adjacent sections through the mediodorsal (A, B) and ventral (C, D) caudate nucleus. In A, arrows point to areas of dense substance P immunoreactivity that match the LAMP-rich zones in B (arrows); in C, arrows point to areas of low substance P staining matching the areas of dense LAMP immunoreactivity in D (arrows). CN, caudate nucleus; P. putamen; NA, nucleus accumbens. Scale bars = 500 pm.
729
LAMP in the basal ganglia of the cat Limbic system-associated membrane protein munoreactivity in the cat caudate nucleus
im-
A strikingly heterogeneous pattern of labeling with the anti-LAMP monoclonal antibody was observed in the caudate nucleus of all cases studied. At low magnification, fields of intense immunoreactivity contrasted with zones that were practically devoid of labeling (Figs IB, D; 2B, D). The irregularly shaped areas of immunolabeling were most intensely stained in the ventral and medial part of the caudate nucleus, still ~nspicuous in its dorsomedial region, and fainter in the dorsolateral quadrant. In most regions of the caudate nucleus, the shape and distribution of the LAMP-immunoreactive zones were clearly reminiscent of that of areas of low AChE activity, called striosomes.‘2 When the location of LAMP-positive areas and of AChE-poor zones were compared in adjacent sections, an almost perfect match was observed in the dorsomediai caudate nucleus. Areas of intense LAMP immunoreactivity indicated by the closed arrows in Fig. 1B line up with the zones of low cholinesterase activity in the adjacent section (Fig. 1A). A match between the LAMPpositive areas and the striosomes was also observed in the ventral half of the nucleus (Fig. lD, C). In some instances, LAMP staining tended to be more extensive than the AChE-poor areas. In particular, a band of intense LAMP immunoreactivity was often observed along the ventricular border of the caudate nucleus (Fig. lB, D, open arrow), an area usually rich in AChE (see Fig. IA, C). Patches of LAMP immunoreactivity in the dorsolateral part of the caudate nucleus could not be aligned with heterogeneities in the AChE staining, which normally are much less conspicuous than in the ventro- and dorsomedial part of the structure.‘3 In order to relate the pattern of LAMP immunostaining with the compartmentalization of the caudate nucleus in its dorsolateral region, sections adjacent to those stained for LAMP were processed for substance P-like immunoreactivity. It has previously been shown that substance P-rich areas line up with striosomes in the dorsolateral caudate nucleus but avoid the AChE-poor zones in the ventral part of the nucleus in the cat.“,i3 In accord with these results, discontinuous areas of intense substance P-like immunoreactivity were observed throughout the caudate nucleus, including its dorsolateral part (Fig. 2A, C). In this area, LAMP-positive zones were sometimes smaller than the areas of intense substance P staining but were located clearly within their boundaries (Fig. ZA, B). Because AChE-poor zones in the ventral caudate nucleus line up with LAMPrich areas, one could predict that LAMP and substance P would lie in non-overlapping aspects of this region. Indeed, we found that LAMP-positive zones were in register with areas devoid of substance P-like immunoreactivity in the ventral part of the caudate nucleus (Fig. 2C, D). LAMP immunoreactivity was
also present in the putamen and the nucleus accumbens (Fig. 2D). Its distribution showed some heterogeneity in these areas as well, but was not examined in detail in the present study. Limbic system -associated membrane prote~ immune reactivity in the substantia nigra
The pattern of LAMP immunostaining in the substantia nigra pars compacta was complex, and corresponded to recently described subdivisions of the A9 dopaminergic cell group in the cat.” These divisions are pa~i~ularly clear in the caudal substantia nigra, where a group of densely packed cells expressing tyrosine hydroxylase-like immunoreactivity represents the densocellular zone of the pars compacta (Fig. 3C). At this level, LAMP immunostaining was intense in the area of the dopaminergic cell groups AlO, A8 and in the cell-sparse part of the substantia nigra pars compacta. Intense labeling was also observed in the substantia nigra pars reticulata. In contrast, LAMP immunostaining was weaker in substantia nigra pars lateralis, and absent from the densocellular pars compacta (Fig. 3D). Rostrally, a band of intense labeling for LAMP was found in the cell-sparse pars compacta but was abruptly interrupted in zones lining up with clusters of tyrosine hydroxylase-positive neurons considered to be the rostra1 extension of the densocellular zone’* (Fig. 3A, B). LAMP immunostaining was much reduced in these cluster areas, as well as through the rostra1 substantia nigra pars reticulata (Fig. 3B). DISCUSSION
The present findings show that LAMP, a 64,OOO68,000 mol. wt membrane glycoprotein shared by subsets of limbic structures, is a common dete~inant of striosomes in all regions of the caudate nucleus in the cat. This indicates that the striosomes from both the ventral and the dorsal parts of the caudate nucleus display a unique relation at the molecular level, further suggesting that they are integrated in a functional network different to the surrounding extrastriosomal matrix. The striosomes receive inputs from a number of areas of the brain which are part of, or closely associated with, the limbic system. In particular, in the cat, striosomes are innervated by the prefrontal cortex (mostly in the dorsolateral caudate nucleus), and the basolateral nucleus of the amygdala (mostly in the ventral and medial aspects of the caudate nucleus).2”30 In the ventral portion of the caudate nucleus, the heterogeneous pattern of LAMP immunostaining observed here matches the distribution of striatal afferents from the basolateral nucleus in many respects. As we found for LAMP expression, this specific amygdalar projection overlaps with striosomes and is denser ventraliy and in a strip of labeling along the ventricular edge of the caudate nucleus.” The presence of distinct patches of LAMP im-
730
M.-F.
CHESELET
PI d
Fig. 3. lmmunostaining for tyrosine hydroxylase (A. C) and LAMP (B, D) in two sets of adjacent sections of the rostra1 (A, B) and caudal (C, D) substantia nigra of the cat. Closed arrow in A points to a dense cluster of tyrosine hydroxylase-positive neurons in the rostra1 substantia nigra pars compacta, whrch corresponds to a zone of low LAMP immunostaining (B, closed arrow). In C, closed arrows point to the densocellular zone of the substantia nigra pars compacta, which is devoid of LAMP labeling (D, closed arrows). Open arrows indicate the areas of intense LAMP immunostaining (B, D) in the cell-sparse region of substantia nigra pars compacta (A, C). SNr, substantia nigra pars reticulata; Pe. cerebral peduncle. As, A, dopaminergic cell group; A,,,, dopaminergic cell group of the ventral tegmental area. Scale bars = 500pm. munoreactivity in the dorsolateral quadrant of the caudate nucleus suggests that LAMP is also present in the target areas of inputs from the prefrontal cortex, which projects much more massively to striosomes in this region than the basolateral nucleus of the amygdala28~“0 (see Fig. 4). Thus the pattern of immunoreactivity of LAMP in the striatum of the cat confirmed previous observations in the rat in which a ventromedial to dorsolateral gradient of intensity was described.” The much weaker patches of immunostaining in the dorsolateral striatum may have gone undetected previously in the rodent because of differences in sensitivity of the immunohistochemical procedure, since they have been recently detected in the mouse (Steindler and Levitt, unpublished observations). The quasi absence of LAMP immunostaining in the extrastriosomal matrix was in sharp contrast to
the dense immunolabeling of the striosomes, in particular in the ventro- and dorsomedial parts of the caudate nucleus. This was consistent with previous observation of a sharp distinction between labeled and non-labeled areas in the rat brain” but compartmental differences in striatal labeling had not been identified in this species. This may be related to a greater difficulty in identifying striosomes and extrastriosomal matrix in this species because of the presence of fiber tracts crossing the striatum, in the absence of markers of each compartment such as calbindin 28K (matrix) or opiate binding sites (striosomes).‘.’ The much weaker LAMP immunostaining in the extrastriosomal matrix of the caudate nucleus in the cat was surprising in view of projections to this area from regions associated with the limbic system such as the cingulate cortex, the ventral tegmental area, the
LAMP in the basal ganglia of the cat
_\ ++t
UYP
IAMP
l
.
**
*aim-
&-
A&LAN?.
--
%\
Fig. 4. Schematic diagram showing some connections of areas that express (LAMP+ and LAMP+++) or do not express (LAMP-) LAMP immunoreactivity in the caudate nucleus (CN) and substantia nigra pars compacta @NC)of the cat. Connections are based on data from Refs 18,29 and 30. See Results for details.
A8 dopaminergic cell group, and the lateral nucleus of the amygdala. ‘**28-30 A clue to the distinct pattern of immunoreactivity to LAMP in the caudate nucleus may reside in a comparison of the level of labeling of limbic inputs to the two striatal compartments. LAMP immunostaining was dense in the basolateral nucleus of the amygdala, and conspicuously absent from the lateral nucleus of the cat (unpublished observations). This further distinguishes two amygdaloid nuclei, which, while both part of the limbic system, differ in the type of afferents they receive. Projections from the autonomic-associated hypothalamic, thalamic and neocortical areas terminate principally in the central and basolateral nuclei,27 while the lateral nucleus is the primary recipient of polysensory (so-called non-limbic) inputs.27,3’ This finding highlights the fact that, as shown previously in the rat,16sz2not all components of the limbic system are equally immunoreactive for LAMP. In particular, other regions projecting to the extrastriosomal matrix such as the cingulate cortex and the ventral tegmental area express a significantly less intense staining for LAMP than areas of the brain projecting to striosomes such as the prefrontal cortex.22 Thus the present results in the cat confirm observations in the rat that LAMP is a molecular marker expressed with high intensity by only a subset of limbic pathways. This may reflect significant differences in the projection and afferent patterns of distinct limbic cortical and subcortical areas. In the rat, a preferential projection from the deep layers of the cerebral cortex to the striosomes has been described.4 It is as yet unknown whether a similar organization of cortico-striatal
731
inputs exists in the cat, and whether differences in the staining intensity for LAMP exist among cortical layers in this species. Recent data on the connections between the substantia nigra and the caudate nucleus in both cats and rats raised intriguing questions regarding the relationship of striosomes with the limbic system. Although the striosomal compartment clearly receives inputs from limbic areas such as the basolateral nucleus of the amygdala and the prefrontal cortex as indicated above, it is also innervated by a subpopulation of neurons of the substantia nigra pars compacta which is not considered part of the limbic system, whereas the ventral tegmental area, a limbic region, projects to the extrastriosomal matrix.6~18That areas of the brain connected to the limbic system are not exclusively part of closed circuits of limbic nature is further illustrated by the pattern of LAMP-poor and -rich areas observed in the striatum and the substantia nigra in the present study. According to Jimenez-Castellanos and Graybiel,‘* the densocellular zone of substantia nigra pars compacta and its rostra1 finger-like extensions project to the striosomes. Thus, our data show that dopaminergic neurons giving rise to afferents to the LAMP-rich striosomes do not stain heavily for LAMP, whereas LAMP immunoreactivity is intense in the subgroups of dopaminergic neurons projecting to the LAMP-poor extrastriosomal matrix (see Fig. 4). The complexity of LAMP expression in the regions projecting to the striatum may be due, in part, to the specific postsynaptic localization of LAMP in the adult, thus reflecting limbic-receptive regions of the central nervous system rather than specifically the cells of origin. ‘Q For example, in the rat, the olfactory bulb, which projects to parts of the limbic system but does not receive any substantial inputs from limbic structures, does not express LAMP immunoreactivity. 22Accordingly, the difference in the intensity of LAMP staining in subgroups of pars compacta neurons suggests that the cell-sparse and the celldense zones may receive different sets of inputs. This further emphasizes distinctions between subgroups of dopaminergic neurons also revealed in studies of receptor ligand binding in the cat,‘O and in the distribution of calcium binding proteins and tyrosine hydroxylase messenger RNA in the rat.5.b,32In addition, the presence of LAMP immunoreactivity in substantia nigra pars reticulata, particularly in its caudal aspect, may reflect the existence of a limbic input to this region. From studies in the rat, the best candidate for the origin of this pathway is the dorsal hypothalamus.’ It is unlikely that inputs from the amygdala, other hypothalamic areas or nucleus accumbens, contribute significantly to the innervation of the substantia nigra pars reticulata.9*24,25Further tracing studies in the cat will be needed to determine more precisely the origin of potential limbic inputs to the substantia nigra pars reticulata.
732
M.-F. C‘HESSELETEd al. CONCLUSION
The major implication of the pattern of LAMP labeling in the caudate nucleus is that LAMP expression is restricted in this structure to the target areas of a distinct subgroup of limbic afferents. Thus, a common molecular “cue” appears to be shared by different pathways projecting to striosomes located in distinct regions of the caudate nucleus. This suggests that LAMP and other newly discovered cell surface molecules provide a useful molecular approach for investigating the organization of the central nervous
system. In addition, the availability of a cell surl’ace marker for striosomal neurons in the striatum provides a unique opportunity to identify these neurons in future anatomical, molecular and cell culture experiments. Acknowledgements-We are grateful to Dr M. Goldberger for providing the perfused cat tissue. Supported by grants from the National Science Foundation BNS-8607645. the Dystonia Medical Research Foundation (MFC). NIMH grant MH45507, and a March of Dimes Basic Research Grant 1-919 (PL).
REFERENCES 1. Dodd J. and Jesse1 T. M. (1986) Cell surface glycoconjugates and carbohydrate-binding signals in sensory neuron development. J. exp. Biol. 124, 225-238.
proteins: possible recognition
2. Donoghue J. P. and Herkenham M. (1986) Neostriatal projections from individual cortical fields conform to histochemically distinct striatal compartments in the rat. Brain Res. 345, 397403. 3. Gerfen C. R. (1984) The neostriatal matrix: compartmentalization of corticostriatal input and striatonigral output systems. Nature 311, 461L464. 4. Gerfen C. R. (1989) The neostriatal mosaic: striatal patch-matrix organization is related to cortical lamination. Science 246, 385-388. 5. Gerfen C. R., Baimbridge K. G. and Miller J. J. (1985) The neostriatal mosaic: compartmental distribution of calcium binding protein and parvalbumin in the basal ganglia of the rat and monkey. Proc. natn. Acad. Sci. U.S.A. 82, 87808784.
6. Gerfen C. R., Herkenham M. and Thibault J. (1987) The neostriatal matrix: II Patch- and matrix-directed mesostriatal dopaminergic and non-dopaminergic systems. J. Neurosci. 7, 3915-3934. 7. Gerfen C. R., Staines W. A., Arbuthnott G. W. and Fibiger H. C. (1982) Crossed connections of the substantia nigra in the rat. J. camp. Neural. 207, 283-303. 8. Goodman C. S., Bastiani M. J., Doe C. G., duLac S., Helfand S. L., Kuwada J. Y. and Thomas J. B. (1984) Cell recognition during neuronal development. Science 225, 1271-1279. 9. Gonzales C. and Chesselet M.-F. (1990) The amygdalo-nigral pathway: an anterograde study in the rat using Phaseolus vulgaris leucoagglutinin (PHA-L). J. camp. Neurol. 297, 1822200. 10. Graybiel A. M., Besson M. J. and Weber E. (1989) Neuroleptic-sensitive binding sites in the nigro-striatal system: evidence for differential distribution of sigma sites in the substantia nigra pars compacta of the cat. J. Neurosci. 9, 326338.
11. Graybiel A. M. and Chesselet M.-F. (1984) Compartmental distribution of striatal cell bodies expressing [Metlenkephalin-like immunoreactivity. Proc. natn. Acad. Sci. U.S.A. 81, 7980-7984. 12. Graybiel A. M. and Ragsdale C. W. (1978) Histochemically distinct compartments in the striatum of human, monkey and cat demonstrated by acetylcholinesterase staining. Proc. natn Acad. Sci. U.S.A. 75, 5723-5726. 13. Graybiel A. M., Ragsdale C. W., Yoneoka E. S. and Elde R. P. (1981) An immunohistochemical study of enkephalins and other neuropeptides in the striatum of the cat with evidence that the opiate peptides are arranged to form mosaic patterns in register-with the striosomal compartments visible by acetylcholinesterase staining. Neuroscience 6, 377-397. 14. Harrelson A. L. and Goodman C. S. (1988) Growth cone guidance in insects: fasciclin II is a member of the immunoglobulin superfamily. Science 24i, 76705. 15. Hendry S. H. C., Jones E. G., Hockfield S. and MacKay R. D. G. (1988) Neuronal populations stained with the monoclonal antibody CAT-301 in the mammalian cerebral cortex and thalamus. J. Neurosci. 8, 518-542. 16. Horton H. L. and Levitt P. (1988) A unique membrane protein is expressed on early developing _ - limbic system axons and cortical targets. J. Neurosci. 8, 46532661. 17. Hsu S. M.. Raine L. and Farmer H. (1981) The use of avidin-biotin neroxidase corn&x (ABC) in immunoneroxidase techniques; a comparison beiween ABC ‘and unlabelled antibody {PAP) procedures. j. Hisiochem. Cytochem. 29, 577-580.
18. Jimenez-Castellanos J. and Graybiel A. M. (1987) Subdivisions of the dopamine-containing A8-A9-A10 complex identified by their differential mesostriatal innervation of striosomes and extrastriosomal matrix. Neuroscience 23, 223-242.
19. Joyce J. N. and Marshall J. F. (1987) Quantitative autoradiography of dopamine D2 sites in rat caudate-putamen: localization to intrinsic neurons and not to neocortical afferents. Neuroscience 28, 773-795. 20. Keller F., Rimvall K., Barbe M. F. and Levitt P. (1989) A membrane glycoprotein associated with the limbic system mediates the formation of the septo-hippocampal pathway in vitro. Neuron 3, 551-561. 21. Kelley A., Domesick V. B. and Nauta W. J. H. (1982) The amygdalo-striatal projection in the rat. An anatomical study by anterograde and retrograde tracing methods. Neuroscience~7, 615630. _ 22. Levitt P. (1984) A monoclonal antibody to limbic svstem neurons. Science 223. 299-301. 23. Levitt P., Pawlack-Byczkowska E., Horton H. L. and Cooper V. (1986) Assembly of functional systems in the brain: molecular and anatomical studies of the limbic system. In The Neurobiology of Down Syndrome (ad. Epstein C. J.). pp. 195-209. Raven Press, New York. 24. Nauta W. J. H. and Domesick V. B. (1977) Cross-roads of limbic and striatal circuitry: hypothalamo-nigral connections. In Limbic Mechanisms: the Continuing Evolution qf the Limbic System Concept (eds Livingston K. E. and Hornykiewicz O.), pp. 75-92. Plenum Press, New York. 25. Nauta W. J. H., Smith G. P., Faull R. L. M. and Domesick V. B. (1978) Efferent connections and nigral afferents of the nucleus
accumbens
septi in the rat. Neuroscience
3, 285401.
LAMP in the basal ganglia of the cat
133
26. Pickel V. M., Joh T. and Reis D. J. (1976) Monoamine-synthetizing enzymes in central dopaminergic, noradrenergic and serotonergic neurons. J. Histochem. Cytochem. 24, 792-806. 21. Price J. L., Russchen F. T. and Amaral D. G. (1987) In Handbook of Chemical Neuroanatomy, Vol. 5: Integrated Systems in the CNS, Part 1 (eds Bjorklund A., Hiikfelt T., Swanson L. W.), pp. 279-388. Elsevier, Amsterdam. 28. Ragsdale C. W. and Graybiel A. M. (1981) The fronto-striatal projection in the cat and monkey and its relationship to inhomogeneities established by acetylcholinesterase histochemistry. Brain Res. MS, 259-266. 29. Ragsdale C. W. and Graybiel A. M. (1984) Further observations on the striosomal organization of frontostriatal projections in cats and monkeys. Neurosci. Absrr. 10, 514. 30. Ragsdale C. W. and Graybiel A. M. (1988) Fibers from the basolateral nucleus of the amygdala selectively innervate striosomes in the caudate nucleus of the cat. J. camp. Neural. 269, 506-522. 31. Russchen F. T. (1986) Cortical and subcortical afferents of the amygdaloid complex. In Excitatory Amino-acids and Epilepsy (eds Schwartz R. and Ben-Ari Y.), pp. 35-52. Plenum P&s<, New York. 32. Weiss-Wunder L. T. and Chesselet M.-F. (1989) Distribution and reaulation of tvrosine hvdroxvlase TTH) mRNA in . < II mesencephalic dopamine (DA) neurons. ieuro&. Absfr. 15, 904. 33. i&co A., Cooper V., Chantler P. D., Fisher-Hyland S., Horton H. L. and Levitt P. (1990) Isolation, biochemical characterization and ultrastructural analysis of the limbic system-associated membrane protein (LAMP), a protein expressed by neurons comprising functional neuronal circuits. J. Neurosci. 10, 73-90. (Accepted 14 September 1990)
MC
4013-F
NeuroscienceVol. 40, No. 3, pp. 725-733, 1991 Printed in Great Britain
HETEROGENEOUS DISTRIBUTION OF THE LIMBIC SYSTEM-ASSOCIATED MEMBRANE PROTEIN IN THE CAUDATE NUCLEUS AND SUBSTANTIA NIGRA OF THE CAT *Department
M.-F. Cr-mssnLEr,*t~ C. GONZALES* and P. Luvtrrt University of Pennsylvania, 36 and Hamilton Walk, Philadelphia, PA
of Pharmacology, tDepartment
19104, U.S.A. of Anatomy, The Medical College of Pennsylvania, U.S.A.
Al&r&-The limbic system-associated membrane protein is a glycoprotein selectively associated, in the adult, with dendrites and cell bodies of neurons of the limbic system and related brain regions. In the present study, the distribution of the limbic system-associated membrane protein was studied by immunohistochemistry in the caudate nucleus and substantia nigra of the cat to determine how its expression relates to the compartmentalization of these areas. In all areas of the caudate nucleus, the pattern of limbic system-associated membrane protein immunoreactivity was highly heterogeneous, labeling zones that were in register with areas expressing neurochemical markers that classically identify striosomes. The extrastriosomal matrix exhibited low levels of staining. The results show that the limbic system-associated membrane protein is expressed by neurons within the target areas (striosomes) of subsets of limbic afferents (originating mainly from the basolateral nucleus of the amygdala and the prefrontal cortex), whereas regions of the caudate nucleus (extrastriosomal matrix) receiving inputs from other subdivisions of the limbic system, such as the cingulate cortex and the ventral tegmental area, contain relatively low levels of limbic system-associated membrane protein immunoreactivity. Thus the expression of this antigen may reflect the targeting of specific groups of limbic afferents to regions that are intimately associated with distinct components of the limbic system. The presence of limbic system-associated membrane protein in neurons of the substantia nigra pars compacta does not appear to be related to the presence or absence of the protein in their striatal target areas. In the substantia nigra, immunoreactivity to the limbic system-associated membrane protein was intense in the cell-sparse zone of the pars compacta, an area known to project to the extrastriosomal matrix of the caudate nucleus. This contrasted with the absence of immunostaining in areas containing dense clusters of dopaminergic neurons (densocellular zone), which project to the limbic system-associated protein-rich striosomes. By analogy to findings in the caudate nucleus and in other brain areas, the results suggest that subgroups of nigral dopaminergic neurons identified on the basis of their terminal fields in the caudate nucleus, may also differ in their limbic afferents.
Work in the nervous system of vertebrates and invertebrates has shown that some anatomically connected neurons share common membrane antigens. ‘,14.15~22 These “system-associated” antigens may serve as identification molecules in the development of neuronal networks involved in a common function.8.20,22In the rat, one such membrane antigen, the limbic system-associated membrane protein (LAMP), is a 64,000-68,000 mol. wt glycoprotein33 that is present during early development on axons, dendrites and cell bodies of neurons belonging to the limbic system,‘6*23an ensemble of brain structures involved in emotional behavior. LAMP immunoreactivity is also expressed in the adult central nervous system, but is restricted to neuronal cell bodies and their dendrites following synaptogenesis.‘6’22’23,33Staining has been found in such limbic areas as prefrontal, $To whom correspondence should be addressed. Abbreviations:
AChE, acetylcholinesterase; LAMP, limbic system-associated membrane protein; PBS, phosphatebuffered saline.
cingulate and perirhinal cortex, hippocampus, septum, amygdala, midline thalamic nuclei and hypothalamic regions. In addition, LAMP immunoreactivity is present in autonomic nuclei of the brainstem that are intimately associated with limbic forebrain regions. **,*’This location, and the presence of LAMP staining on cell bodies and dendrites, but not on axon terminals33 suggests that, in the adult, the molecule may signal areas of the central nervous system which receive afferents from subsets of limbic structures. The striatum (caudate nucleus and putamen) receives inputs from a variety of limbic areas.2~3,6,2’.2*,30 In adult rats, immunostaining for LAMP was much more intense in the ventromedial part of the striatum than in its dorsolateral region.** The area of intense LAMP immunoreactivity corresponds to the region of the caudate-putamen which receives inputs from the basolateral nucleus of the amygdala, a part of the limbic system, whereas the LAMP-negative region overlaps with an area receiving massive projections from the sensorimotor cortex.*’ This pattern of 725
726
M.-F.
~HiiSSELE’l’
LAMP immunoreactivity provides a strong argument in favor of the presence of LAMP in neuronal targets of limbic inputs, even in regions usually not classified as limbic structures themselves.22 Recent work on the connectivity of the caudate nucleus, however, has revealed a more complex organization of limbic inputs to this area. In the cat, afferents from different parts of the limbic system are segregated among two distinct striatal compartments. In particular, the basolateral nucleus of the amygdala projects to distinct patchy areas of the caudate nucleus called “striosomes”, but not to the extrastriosomal matrix surrounding them.30 The striosomes also receive inputs from some areas of the prefrontal cortex associated with the limbic system. zB-3oOther parts of the limbic circuitry such as the cingulate cortex and the ventral tegmental area, however, project preferentially to the extrastriosomal matrix and avoid the striosomes.is.** 3o A similar ~~egation of limbic inputs from cerebral cortex and mesencephalic dopaminergic cell groups to the striatum has been observed in the rat.2s’.7 Non-limbic inputs to the caudate nucleus are also distin~ished among these striatal subdivisions. In particular, subgroups of pars compacta neurons, which form one of the main afferent pathways to the caudate nucleus, project either to the striosomal or extrastriosomal compartments of the striatum.6,‘8 In view of the complex relationship between the striatum and distinct regions of the limbic system, we have now examined whether the distribution of LAMP immunorea~tivity in the caudate nucleus of the cat reflects the anatomical segregation of limbic inputs to distinct compartments in this structure. In addition, we examined the pattern of staining for LAMP in the substantia nigra to determine whether it was related to the organization of nigral projections to striatal subregions. In the caudate nucleus, we took advantage of the fact that the striatal compartments can be distinguished from each other based on the specific distribution of neurochemical markers. In particular, striosomes exhibit sparse staining for acetylcholinesterase (AChE)‘* and, in the dorsolateral striatum, intense substance P-like immunorea~tivity.“*13 Adjacent sections of caudate nuclei of adult cats were processed for LAMP localization using immunoperoxidase staining, and for AChE histochemistry or substance P-like immunorea~tivity, in order to identify the striosomes. In the substantia nigra, neurons of the pars compacta were identified by immunostaining with an antibody against tyrosine hydroxylase, the rate limiting enzyme of dopamine synthesis.”
<‘f d
EXPERIMENTAL
PROCEDGRES
Four adult cats were used in this study. Three cats were perfused through the heart with a mixture of 4010 paraformaldehyde and 0.1% glutaraldehyde and one with 4% paraformaldehyde only. In all cases, the fixative was freshly prepared in sodium phosphate-buffered saline (PBS; 0.1 M, pH 7.4) and administered after a brief flush of saline. The brains were removed after I h postfixation in situ. cryoprotected and stored at 4°C in PBS containing sucrose (30%). The brains were frozen on dry ice and sections (35 pm thick) were cut on a sliding microtome, collected in PBS and processed either for immunohistochemistry or AChE histochemistry. Sodium sulfate-precipitated anti-LAMP was prepared from culture supernatants in which the hybridoma clone 2G9 was grown.22 The antibody precipitate was dissolved in water and dialysed extensively against PBS. The sections were incubated-for 24 h at 4°C in a I : 1000 dilution of the antibodv in PBS containing 4% non-fat drv milk. Followine extensive washing, the sections were incubated for 1h at room temperature in a 1: 50 dilution of horseradish peroxidase-conjugated goat anti-mouse IgG in 4% non-fat dry milk. Sections were rinsed and the peroxidase reaction product was developed using a standard 3,3’-diaminobenzidine-peroxidaseH,O, method.22,23,33 Substance P-like immunohistochemistrv was performed in tissue from the cat perfused with para~o~a~ehyde alone, using an antibody from INCSTAR (I:500 in Tris-buffered saline: 0.5 M, pH 7.4) and the avidin-biotin technique,” with a Vectastain kit (Vector Laboratories), following the instructions of the manufacturer. Sections incubated with primary antibody that had been preincubated overnight at 4°C with substance P (foe4 M) did not show any imnlunos~ining. Tyrosine hydroxylase-like immunoreactivity was detected using an antibodv from Eugene Tech. Inc. diluted 1:1200 in PBS and peroxid&-antipeyoxidase reaction as described.‘” Controls included omission of the primary antibody or of the peroxidase~nti~roxidase complex. AChE h~sto~hemist~ was performed as described by Joyce and MarshallI with slight modifications. Briefly, sections were incubated S-6 h in 0.1 M acetate buffer (pH 5.3) containing 0.1 M CuSO,-glycine solution. After incubation, sections were rinsed in dH,O, developed in 4.5% K, Fe(CN), in dH,O for 15 min then rinsed three times in dH,O before being mounted on gelatin-coated slides, air-dried, dehydrated and mounted with Eukitt.
RESULTS
In all areas examined, LAMP immunoreactivity was present within the neuropil between profiles of unstained somata but was not resolved to specific morphological structures. In thinner sections, however, LAMP immunoreactivity was detected pericellularly along cell bodies and processes. This is in agreement with previous ultrastructural studies in the adult rat which have identified the plasma membrane of neuronal somata and dendrites as the specific cellular compartments that exhibit LAMP immunoreactivity.33
Fig. 1. Photomicro~ap~s depict coincident striosomal (AChEpoor) localization of LAMP. Patches of low AChE activity (A, C: pale zones, examples at white arrows) and of dense LAMP immunolabeling (9, D: dark zones, examples at black arrows) in the caudate nucleus of the cat. A, 9; C, D are sets of adjacent sections. Note the matches between cholinesterase-poor and LAMP-rich areas (white and black arrows, respectively), except for areas indicated by open arrows in A and D. Scale bars = 500 pm.
Fig.
1
72-l
728
M.-F. CHESSELET cr ul
Fig. 2. Comparison between the heterogeneous patterns of substance P (A, C) and LAMP (I?, D) immunostaining in two sets of adjacent sections through the mediodorsal (A, B) and ventral (C, D) caudate nucleus. In A, arrows point to areas of dense substance P immunoreactivity that match the LAMP-rich zones in B (arrows); in C, arrows point to areas of low substance P staining matching the areas of dense LAMP immunoreactivity in D (arrows). CN, caudate nucleus; P. putamen; NA, nucleus accumbens. Scale bars = 500 pm.
729
LAMP in the basal ganglia of the cat Limbic system-associated membrane protein munoreactivity in the cat caudate nucleus
im-
A strikingly heterogeneous pattern of labeling with the anti-LAMP monoclonal antibody was observed in the caudate nucleus of all cases studied. At low magnification, fields of intense immunoreactivity contrasted with zones that were practically devoid of labeling (Figs IB, D; 2B, D). The irregularly shaped areas of immunolabeling were most intensely stained in the ventral and medial part of the caudate nucleus, still ~nspicuous in its dorsomedial region, and fainter in the dorsolateral quadrant. In most regions of the caudate nucleus, the shape and distribution of the LAMP-immunoreactive zones were clearly reminiscent of that of areas of low AChE activity, called striosomes.‘2 When the location of LAMP-positive areas and of AChE-poor zones were compared in adjacent sections, an almost perfect match was observed in the dorsomediai caudate nucleus. Areas of intense LAMP immunoreactivity indicated by the closed arrows in Fig. 1B line up with the zones of low cholinesterase activity in the adjacent section (Fig. 1A). A match between the LAMPpositive areas and the striosomes was also observed in the ventral half of the nucleus (Fig. lD, C). In some instances, LAMP staining tended to be more extensive than the AChE-poor areas. In particular, a band of intense LAMP immunoreactivity was often observed along the ventricular border of the caudate nucleus (Fig. lB, D, open arrow), an area usually rich in AChE (see Fig. IA, C). Patches of LAMP immunoreactivity in the dorsolateral part of the caudate nucleus could not be aligned with heterogeneities in the AChE staining, which normally are much less conspicuous than in the ventro- and dorsomedial part of the structure.‘3 In order to relate the pattern of LAMP immunostaining with the compartmentalization of the caudate nucleus in its dorsolateral region, sections adjacent to those stained for LAMP were processed for substance P-like immunoreactivity. It has previously been shown that substance P-rich areas line up with striosomes in the dorsolateral caudate nucleus but avoid the AChE-poor zones in the ventral part of the nucleus in the cat.“,i3 In accord with these results, discontinuous areas of intense substance P-like immunoreactivity were observed throughout the caudate nucleus, including its dorsolateral part (Fig. 2A, C). In this area, LAMP-positive zones were sometimes smaller than the areas of intense substance P staining but were located clearly within their boundaries (Fig. ZA, B). Because AChE-poor zones in the ventral caudate nucleus line up with LAMPrich areas, one could predict that LAMP and substance P would lie in non-overlapping aspects of this region. Indeed, we found that LAMP-positive zones were in register with areas devoid of substance P-like immunoreactivity in the ventral part of the caudate nucleus (Fig. 2C, D). LAMP immunoreactivity was
also present in the putamen and the nucleus accumbens (Fig. 2D). Its distribution showed some heterogeneity in these areas as well, but was not examined in detail in the present study. Limbic system -associated membrane prote~ immune reactivity in the substantia nigra
The pattern of LAMP immunostaining in the substantia nigra pars compacta was complex, and corresponded to recently described subdivisions of the A9 dopaminergic cell group in the cat.” These divisions are pa~i~ularly clear in the caudal substantia nigra, where a group of densely packed cells expressing tyrosine hydroxylase-like immunoreactivity represents the densocellular zone of the pars compacta (Fig. 3C). At this level, LAMP immunostaining was intense in the area of the dopaminergic cell groups AlO, A8 and in the cell-sparse part of the substantia nigra pars compacta. Intense labeling was also observed in the substantia nigra pars reticulata. In contrast, LAMP immunostaining was weaker in substantia nigra pars lateralis, and absent from the densocellular pars compacta (Fig. 3D). Rostrally, a band of intense labeling for LAMP was found in the cell-sparse pars compacta but was abruptly interrupted in zones lining up with clusters of tyrosine hydroxylase-positive neurons considered to be the rostra1 extension of the densocellular zone’* (Fig. 3A, B). LAMP immunostaining was much reduced in these cluster areas, as well as through the rostra1 substantia nigra pars reticulata (Fig. 3B). DISCUSSION
The present findings show that LAMP, a 64,OOO68,000 mol. wt membrane glycoprotein shared by subsets of limbic structures, is a common dete~inant of striosomes in all regions of the caudate nucleus in the cat. This indicates that the striosomes from both the ventral and the dorsal parts of the caudate nucleus display a unique relation at the molecular level, further suggesting that they are integrated in a functional network different to the surrounding extrastriosomal matrix. The striosomes receive inputs from a number of areas of the brain which are part of, or closely associated with, the limbic system. In particular, in the cat, striosomes are innervated by the prefrontal cortex (mostly in the dorsolateral caudate nucleus), and the basolateral nucleus of the amygdala (mostly in the ventral and medial aspects of the caudate nucleus).2”30 In the ventral portion of the caudate nucleus, the heterogeneous pattern of LAMP immunostaining observed here matches the distribution of striatal afferents from the basolateral nucleus in many respects. As we found for LAMP expression, this specific amygdalar projection overlaps with striosomes and is denser ventraliy and in a strip of labeling along the ventricular edge of the caudate nucleus.” The presence of distinct patches of LAMP im-
730
M.-F.
CHESELET
PI d
Fig. 3. lmmunostaining for tyrosine hydroxylase (A. C) and LAMP (B, D) in two sets of adjacent sections of the rostra1 (A, B) and caudal (C, D) substantia nigra of the cat. Closed arrow in A points to a dense cluster of tyrosine hydroxylase-positive neurons in the rostra1 substantia nigra pars compacta, whrch corresponds to a zone of low LAMP immunostaining (B, closed arrow). In C, closed arrows point to the densocellular zone of the substantia nigra pars compacta, which is devoid of LAMP labeling (D, closed arrows). Open arrows indicate the areas of intense LAMP immunostaining (B, D) in the cell-sparse region of substantia nigra pars compacta (A, C). SNr, substantia nigra pars reticulata; Pe. cerebral peduncle. As, A, dopaminergic cell group; A,,,, dopaminergic cell group of the ventral tegmental area. Scale bars = 500pm. munoreactivity in the dorsolateral quadrant of the caudate nucleus suggests that LAMP is also present in the target areas of inputs from the prefrontal cortex, which projects much more massively to striosomes in this region than the basolateral nucleus of the amygdala28~“0 (see Fig. 4). Thus the pattern of immunoreactivity of LAMP in the striatum of the cat confirmed previous observations in the rat in which a ventromedial to dorsolateral gradient of intensity was described.” The much weaker patches of immunostaining in the dorsolateral striatum may have gone undetected previously in the rodent because of differences in sensitivity of the immunohistochemical procedure, since they have been recently detected in the mouse (Steindler and Levitt, unpublished observations). The quasi absence of LAMP immunostaining in the extrastriosomal matrix was in sharp contrast to
the dense immunolabeling of the striosomes, in particular in the ventro- and dorsomedial parts of the caudate nucleus. This was consistent with previous observation of a sharp distinction between labeled and non-labeled areas in the rat brain” but compartmental differences in striatal labeling had not been identified in this species. This may be related to a greater difficulty in identifying striosomes and extrastriosomal matrix in this species because of the presence of fiber tracts crossing the striatum, in the absence of markers of each compartment such as calbindin 28K (matrix) or opiate binding sites (striosomes).‘.’ The much weaker LAMP immunostaining in the extrastriosomal matrix of the caudate nucleus in the cat was surprising in view of projections to this area from regions associated with the limbic system such as the cingulate cortex, the ventral tegmental area, the
LAMP in the basal ganglia of the cat
_\ ++t
UYP
IAMP
l
.
**
*aim-
&-
A&LAN?.
--
%\
Fig. 4. Schematic diagram showing some connections of areas that express (LAMP+ and LAMP+++) or do not express (LAMP-) LAMP immunoreactivity in the caudate nucleus (CN) and substantia nigra pars compacta @NC)of the cat. Connections are based on data from Refs 18,29 and 30. See Results for details.
A8 dopaminergic cell group, and the lateral nucleus of the amygdala. ‘**28-30 A clue to the distinct pattern of immunoreactivity to LAMP in the caudate nucleus may reside in a comparison of the level of labeling of limbic inputs to the two striatal compartments. LAMP immunostaining was dense in the basolateral nucleus of the amygdala, and conspicuously absent from the lateral nucleus of the cat (unpublished observations). This further distinguishes two amygdaloid nuclei, which, while both part of the limbic system, differ in the type of afferents they receive. Projections from the autonomic-associated hypothalamic, thalamic and neocortical areas terminate principally in the central and basolateral nuclei,27 while the lateral nucleus is the primary recipient of polysensory (so-called non-limbic) inputs.27,3’ This finding highlights the fact that, as shown previously in the rat,16sz2not all components of the limbic system are equally immunoreactive for LAMP. In particular, other regions projecting to the extrastriosomal matrix such as the cingulate cortex and the ventral tegmental area express a significantly less intense staining for LAMP than areas of the brain projecting to striosomes such as the prefrontal cortex.22 Thus the present results in the cat confirm observations in the rat that LAMP is a molecular marker expressed with high intensity by only a subset of limbic pathways. This may reflect significant differences in the projection and afferent patterns of distinct limbic cortical and subcortical areas. In the rat, a preferential projection from the deep layers of the cerebral cortex to the striosomes has been described.4 It is as yet unknown whether a similar organization of cortico-striatal
731
inputs exists in the cat, and whether differences in the staining intensity for LAMP exist among cortical layers in this species. Recent data on the connections between the substantia nigra and the caudate nucleus in both cats and rats raised intriguing questions regarding the relationship of striosomes with the limbic system. Although the striosomal compartment clearly receives inputs from limbic areas such as the basolateral nucleus of the amygdala and the prefrontal cortex as indicated above, it is also innervated by a subpopulation of neurons of the substantia nigra pars compacta which is not considered part of the limbic system, whereas the ventral tegmental area, a limbic region, projects to the extrastriosomal matrix.6~18That areas of the brain connected to the limbic system are not exclusively part of closed circuits of limbic nature is further illustrated by the pattern of LAMP-poor and -rich areas observed in the striatum and the substantia nigra in the present study. According to Jimenez-Castellanos and Graybiel,‘* the densocellular zone of substantia nigra pars compacta and its rostra1 finger-like extensions project to the striosomes. Thus, our data show that dopaminergic neurons giving rise to afferents to the LAMP-rich striosomes do not stain heavily for LAMP, whereas LAMP immunoreactivity is intense in the subgroups of dopaminergic neurons projecting to the LAMP-poor extrastriosomal matrix (see Fig. 4). The complexity of LAMP expression in the regions projecting to the striatum may be due, in part, to the specific postsynaptic localization of LAMP in the adult, thus reflecting limbic-receptive regions of the central nervous system rather than specifically the cells of origin. ‘Q For example, in the rat, the olfactory bulb, which projects to parts of the limbic system but does not receive any substantial inputs from limbic structures, does not express LAMP immunoreactivity. 22Accordingly, the difference in the intensity of LAMP staining in subgroups of pars compacta neurons suggests that the cell-sparse and the celldense zones may receive different sets of inputs. This further emphasizes distinctions between subgroups of dopaminergic neurons also revealed in studies of receptor ligand binding in the cat,‘O and in the distribution of calcium binding proteins and tyrosine hydroxylase messenger RNA in the rat.5.b,32In addition, the presence of LAMP immunoreactivity in substantia nigra pars reticulata, particularly in its caudal aspect, may reflect the existence of a limbic input to this region. From studies in the rat, the best candidate for the origin of this pathway is the dorsal hypothalamus.’ It is unlikely that inputs from the amygdala, other hypothalamic areas or nucleus accumbens, contribute significantly to the innervation of the substantia nigra pars reticulata.9*24,25Further tracing studies in the cat will be needed to determine more precisely the origin of potential limbic inputs to the substantia nigra pars reticulata.
732
M.-F. C‘HESSELETEd al. CONCLUSION
The major implication of the pattern of LAMP labeling in the caudate nucleus is that LAMP expression is restricted in this structure to the target areas of a distinct subgroup of limbic afferents. Thus, a common molecular “cue” appears to be shared by different pathways projecting to striosomes located in distinct regions of the caudate nucleus. This suggests that LAMP and other newly discovered cell surface molecules provide a useful molecular approach for investigating the organization of the central nervous
system. In addition, the availability of a cell surl’ace marker for striosomal neurons in the striatum provides a unique opportunity to identify these neurons in future anatomical, molecular and cell culture experiments. Acknowledgements-We are grateful to Dr M. Goldberger for providing the perfused cat tissue. Supported by grants from the National Science Foundation BNS-8607645. the Dystonia Medical Research Foundation (MFC). NIMH grant MH45507, and a March of Dimes Basic Research Grant 1-919 (PL).
REFERENCES 1. Dodd J. and Jesse1 T. M. (1986) Cell surface glycoconjugates and carbohydrate-binding signals in sensory neuron development. J. exp. Biol. 124, 225-238.
proteins: possible recognition
2. Donoghue J. P. and Herkenham M. (1986) Neostriatal projections from individual cortical fields conform to histochemically distinct striatal compartments in the rat. Brain Res. 345, 397403. 3. Gerfen C. R. (1984) The neostriatal matrix: compartmentalization of corticostriatal input and striatonigral output systems. Nature 311, 461L464. 4. Gerfen C. R. (1989) The neostriatal mosaic: striatal patch-matrix organization is related to cortical lamination. Science 246, 385-388. 5. Gerfen C. R., Baimbridge K. G. and Miller J. J. (1985) The neostriatal mosaic: compartmental distribution of calcium binding protein and parvalbumin in the basal ganglia of the rat and monkey. Proc. natn. Acad. Sci. U.S.A. 82, 87808784.
6. Gerfen C. R., Herkenham M. and Thibault J. (1987) The neostriatal matrix: II Patch- and matrix-directed mesostriatal dopaminergic and non-dopaminergic systems. J. Neurosci. 7, 3915-3934. 7. Gerfen C. R., Staines W. A., Arbuthnott G. W. and Fibiger H. C. (1982) Crossed connections of the substantia nigra in the rat. J. camp. Neural. 207, 283-303. 8. Goodman C. S., Bastiani M. J., Doe C. G., duLac S., Helfand S. L., Kuwada J. Y. and Thomas J. B. (1984) Cell recognition during neuronal development. Science 225, 1271-1279. 9. Gonzales C. and Chesselet M.-F. (1990) The amygdalo-nigral pathway: an anterograde study in the rat using Phaseolus vulgaris leucoagglutinin (PHA-L). J. camp. Neurol. 297, 1822200. 10. Graybiel A. M., Besson M. J. and Weber E. (1989) Neuroleptic-sensitive binding sites in the nigro-striatal system: evidence for differential distribution of sigma sites in the substantia nigra pars compacta of the cat. J. Neurosci. 9, 326338.
11. Graybiel A. M. and Chesselet M.-F. (1984) Compartmental distribution of striatal cell bodies expressing [Metlenkephalin-like immunoreactivity. Proc. natn. Acad. Sci. U.S.A. 81, 7980-7984. 12. Graybiel A. M. and Ragsdale C. W. (1978) Histochemically distinct compartments in the striatum of human, monkey and cat demonstrated by acetylcholinesterase staining. Proc. natn Acad. Sci. U.S.A. 75, 5723-5726. 13. Graybiel A. M., Ragsdale C. W., Yoneoka E. S. and Elde R. P. (1981) An immunohistochemical study of enkephalins and other neuropeptides in the striatum of the cat with evidence that the opiate peptides are arranged to form mosaic patterns in register-with the striosomal compartments visible by acetylcholinesterase staining. Neuroscience 6, 377-397. 14. Harrelson A. L. and Goodman C. S. (1988) Growth cone guidance in insects: fasciclin II is a member of the immunoglobulin superfamily. Science 24i, 76705. 15. Hendry S. H. C., Jones E. G., Hockfield S. and MacKay R. D. G. (1988) Neuronal populations stained with the monoclonal antibody CAT-301 in the mammalian cerebral cortex and thalamus. J. Neurosci. 8, 518-542. 16. Horton H. L. and Levitt P. (1988) A unique membrane protein is expressed on early developing _ - limbic system axons and cortical targets. J. Neurosci. 8, 46532661. 17. Hsu S. M.. Raine L. and Farmer H. (1981) The use of avidin-biotin neroxidase corn&x (ABC) in immunoneroxidase techniques; a comparison beiween ABC ‘and unlabelled antibody {PAP) procedures. j. Hisiochem. Cytochem. 29, 577-580.
18. Jimenez-Castellanos J. and Graybiel A. M. (1987) Subdivisions of the dopamine-containing A8-A9-A10 complex identified by their differential mesostriatal innervation of striosomes and extrastriosomal matrix. Neuroscience 23, 223-242.
19. Joyce J. N. and Marshall J. F. (1987) Quantitative autoradiography of dopamine D2 sites in rat caudate-putamen: localization to intrinsic neurons and not to neocortical afferents. Neuroscience 28, 773-795. 20. Keller F., Rimvall K., Barbe M. F. and Levitt P. (1989) A membrane glycoprotein associated with the limbic system mediates the formation of the septo-hippocampal pathway in vitro. Neuron 3, 551-561. 21. Kelley A., Domesick V. B. and Nauta W. J. H. (1982) The amygdalo-striatal projection in the rat. An anatomical study by anterograde and retrograde tracing methods. Neuroscience~7, 615630. _ 22. Levitt P. (1984) A monoclonal antibody to limbic svstem neurons. Science 223. 299-301. 23. Levitt P., Pawlack-Byczkowska E., Horton H. L. and Cooper V. (1986) Assembly of functional systems in the brain: molecular and anatomical studies of the limbic system. In The Neurobiology of Down Syndrome (ad. Epstein C. J.). pp. 195-209. Raven Press, New York. 24. Nauta W. J. H. and Domesick V. B. (1977) Cross-roads of limbic and striatal circuitry: hypothalamo-nigral connections. In Limbic Mechanisms: the Continuing Evolution qf the Limbic System Concept (eds Livingston K. E. and Hornykiewicz O.), pp. 75-92. Plenum Press, New York. 25. Nauta W. J. H., Smith G. P., Faull R. L. M. and Domesick V. B. (1978) Efferent connections and nigral afferents of the nucleus
accumbens
septi in the rat. Neuroscience
3, 285401.
LAMP in the basal ganglia of the cat
133
26. Pickel V. M., Joh T. and Reis D. J. (1976) Monoamine-synthetizing enzymes in central dopaminergic, noradrenergic and serotonergic neurons. J. Histochem. Cytochem. 24, 792-806. 21. Price J. L., Russchen F. T. and Amaral D. G. (1987) In Handbook of Chemical Neuroanatomy, Vol. 5: Integrated Systems in the CNS, Part 1 (eds Bjorklund A., Hiikfelt T., Swanson L. W.), pp. 279-388. Elsevier, Amsterdam. 28. Ragsdale C. W. and Graybiel A. M. (1981) The fronto-striatal projection in the cat and monkey and its relationship to inhomogeneities established by acetylcholinesterase histochemistry. Brain Res. MS, 259-266. 29. Ragsdale C. W. and Graybiel A. M. (1984) Further observations on the striosomal organization of frontostriatal projections in cats and monkeys. Neurosci. Absrr. 10, 514. 30. Ragsdale C. W. and Graybiel A. M. (1988) Fibers from the basolateral nucleus of the amygdala selectively innervate striosomes in the caudate nucleus of the cat. J. camp. Neural. 269, 506-522. 31. Russchen F. T. (1986) Cortical and subcortical afferents of the amygdaloid complex. In Excitatory Amino-acids and Epilepsy (eds Schwartz R. and Ben-Ari Y.), pp. 35-52. Plenum P&s<, New York. 32. Weiss-Wunder L. T. and Chesselet M.-F. (1989) Distribution and reaulation of tvrosine hvdroxvlase TTH) mRNA in . < II mesencephalic dopamine (DA) neurons. ieuro&. Absfr. 15, 904. 33. i&co A., Cooper V., Chantler P. D., Fisher-Hyland S., Horton H. L. and Levitt P. (1990) Isolation, biochemical characterization and ultrastructural analysis of the limbic system-associated membrane protein (LAMP), a protein expressed by neurons comprising functional neuronal circuits. J. Neurosci. 10, 73-90. (Accepted 14 September 1990)
MC
4013-F