THE JOURNAL OF COMPARATIVE NEUROLOGY 302:437-446 (1990)
TwoTranspallidalPathwaysOriinatingin the Rat Nucleus Accumbens D.S. ZAHM AND LENNART HEIMER Department of Anatomy and Neurobiology St. Louis University School of Medicine, St. Louis, Missouri 63104 (D.S.Z.) and Departments of Otolaryngology and Neurosurgery, University of Virginia School of Medicine, Charlottesville, Virginia 22908 (L.H.)
ABSTRACT The striatopallidal projection originating in the nucleus accumbens was investigated by using anterograde transport of PHA-L in combination with peptide immunohistochemistry in order to localize the injection sites and transported lectin with respect to neurochemically defined subterritories in the nucleus accumbens and subcommissural ventral pallidum. The results reported here supplement our previous observations, which indicated that the subcommissural ventral pallidum of the rat comprises two immunohistochemically defined subterritories (Zahm and Heimer, '88: J. Comp. Neurol., 272516435) which give rise to dichotomous downstream projection systems (Zahm, '89: Neuroscience, 30:33-50). The present data indicate that the neurotensin immunoreactivity-rich ventromedial district of ventral pallidum receives its accumbal input almost exclusively from the shell district of the nucleus accumbens. The accumbal core, alternatively, projects to the dorsolateral ventral pallidal subterritory that lacks appreciable neurotensin immunoreactivity and in many other respects more resembles the adjoining striatopallidal components of the caudate-putamen. In addition to direct topographic relationships in the frontal plane among the accumbal injection sites and ventral pallidal terminations, it was observed that more caudally placed core injections resulted in patches of striatopallidal terminations that were more caudally located in ventral pallidum. Shell injections, in contrast, produced columns of terminations that extended continuously from the rostralmost level that they appeared to the caudal end Df ventromedial ventral pallidum. The accumbal shell, its exclusive projection to the ventromedial subterritory in the subcommissural ventral pallidum, and the previously reported, almost exclusive projection of that pallidal subdistrict to the mesencephalicventral tegmental area are discussed in terms of a number of other neurochemical and hodological features that serve to distinguish them sufficientlyto suggest that they represent a uniquely specialized part of the basal ganglia. Key words: ventral pallidum, ventral striatum, neurotensin, substance P, PHA-L tract tracing, immunocytochemistry
Recent immunohistochemical studies of the ventral parts of the basal ganglia in the rat have demonstrated that ventral pallidum (VP)dorsal to the polymorph layer of the olfactory tubercle is divisible into two subdistricts, a ventromedial part that is rich in neurotensin (NT) immunoreactive axons and axon terminals, and a dorsolateral part that exhibits a significantlylesser amount of NT immunoreactivity (Zahm and Heimer, '88). Subsequently, investigation of wheat germ agglutinin HRP conjugate transported retrogradely from structures known to receive inputs from pallidum revealed that the projection systems arising in the two ventral pallidal subdistricts are different (Zahm, '89). Whereas dorsolateral VP projects relatively massively to the substantia nigra and subthalamic nucleus, well-known targets of pallidum in the rat (Hattori et al., '75; Bunney and Aghajanian, '76; Carter and Fibiger, '78; Haber et al.,
o 1990 WILEY-LISS, INC.
'85), the ventromedial district of VP gives rise to a mesencephalic projection mostly restricted to the ventral tegmental area (VTA) and projects insignificantly, if at all, to the subthalamic nucleus. Moreover, ventral pallidal neurons projecting to the substantia nigra and subthalamic nucleus are typical "pallidal" neurons in the sense that they receive a very dense afferent input that is reflected ultrastructurally in numerous boutons contacting even their proximal dendrites and perikarya in a pallisadelike fashion (for references, see Zahm et al., '87). The VTA-projecting
Accepted September 1,1990 Address reprint requests to D.S. Zahm, Department of Anatomy and Neurobiology, St. Louis University School of Medicine, 1402 S. Grand Blvd., St. Louis, MO 63104.
D.S. ZAHM AND L. HEIMER
438
neurons, conversely, are contacted by relatively few boutons at the level of their proximal dendrites and perikarya. The fact that the two ventral pallidal compartments give rise to dichotomous projections fostered a suspicion that the striatal inputs from the nucleus accumbens (Nauta et al., '78; Zaborszky et al., '82; Mogenson et al., '83; Groenewegen and Russchen, '84; Haber et al., '90) to the two compartments might also differ. Since recent observations indicate that the nucleus accumbens exhibits a rather conspicuous compartmentation reflected in "core" and "shell" subdistricts (Zaborzsky et al., '85; Voorn et al., '86, '89; Groenewegen et al., '89, '90; Meredith et al., '89; Heimer et al., '90a; Zahm, 'go), it seemed appropriate to determine if each of the accumbal compartments project to the different pallidal compartments identified on the basis of NT-IR. To address this question, the striatopallidal projection originating in the nucleus accumbens was investigated by using anterograde transport of PHA-L in combination with peptide immunohistochemistry to localize the injection sites and transported lectin with respect to the accumbal and pallidal compartments.
MATERTALSIWDMETHODS With the aid of a Kopf stereotaxic instrument, small injections of Phaseolus vulgaris leucoagglutinin (PHA-L, 17 rats) were placed unilaterally in the ventral striatum of pentobarbital anesthetized male Sprague-Dawley rats weighing 225-250 g. The tracers were delivered iontophoretically with positive pulses of 4-7 pA (7s on/7s off) from 2% solutions. Seven days later, the rats were killed under deep pentobarbital anesthesia by vascular perfusion with an aqueous solution containing 0.1% glutaraldehyde, 4% paraformaldehyde, and 2.5% polyvinylpyrrolidone (PVP40) in 0.07 M Sorenson's phosphate buffer (SPB, pH 7.41, preceded by a brief rinse with 2.5% PVP-40 and 0.5% procaine-HC1in sodium phosphate-buffered saline (NaPBS). Fifty micrometer frontal sections cut on the vibratome were subjected to an immunocytochemical protocol for PHA-L, neurotensin (NT) or substance P (SP). Several of the rats received subcutaneous injections of haloperidol (2 mg/kg in a solution containing 50% propylene glycol, 20% ethanol, and 30% distilled water) or the vehicle 18 and 4 hours before sacrifice to enhance the neurotensin immunoreactivity in the dorsolateral ventral pallidal subdistrict (Eggerman and Zahm, '88). Sections were pretreated with 1%sodium borohydride for 20 minutes, 1% H,O, for 10 minutes, and appropriate normal serum for ?hhour, after which they were placed without further rinsing into vials containing 1-2 ml of SPB with 1%appropriate normal serum and primary antisera at the following dilutions: anti-PHA-L, 1:2000; anti-NT, 15000; anti-SP, 1:lOOO-1: 2000. The sections were left on a rocker plate overnight at 4°C. The following morning the sections were rinsed and placed in vials containing SPB, 1% rabbit serum and biotinylated second antibodies at a dilution of 1:200. Bound anti-PHA-L was identified with antigoat, and anti-SP with antirat immunoglobulins (IgG) both produced in rabbit (Vector). Anti-NT was recognized with antirabbit IgGs made in goat. After three rinses, the sections were incubated for 1hour in SPB containing 1%appropriate normal serum and avidinbiotin-peroxidase complex (Vector) at a dilution of 1:200. Immunolabeling was produced with diaminobenzidine (DAB) by modification of the coi1.pled glucose oxidase reaction (Itoh et al., '79; Zahm and Heimer,
'88) and enhanced with osmium and thiocarbohydrazide (C.R. Gerfen, unpublished). The sections were then mounted in rostrocaudal sequence on glass slides coated with gelatin (0.2%) and chrome alum (0.02%) and coverslipped under Permount. Attempts to produce immunolabeling with preabsorbed antisera (50 pgiml) or buffer lacking primary antibody resulted in absence of staining, whereas antisera that were preincubated with an inappropriate antigen (e.g., anti-PHA-L pre-incubated with SP or the converse) produced normal staining patterns.
RESULTS PHA-L injections into the nucleus accumbens resulted in transport of the lectin to axon terminations in ventral pallidum (VP) that when displayed in frontal sections exhibited direct topographic relationships to the positions of the injection sites (Figs. 1-4). Injections in the core of the nucleus accumbens as demonstrated with SP-IR (Figs. 1A-B') labeled axon terminal fields in dorsolateral parts of the subcommissural VP (Figs. 2, 3) and those in the shell (Fig. lC, C) labeled fields in ventromedial VP (Fig. 4). Following injections involving exclusively the core of nucleus accumbens (Figs. 1A-B'), dense aggregates of PHA-L labeled axon terminals and fibers were located almost exclusively in the dorsolateral compartment of ventral pallidum (VPl), although a sparse to moderate amount of labeled fibers and terminals could usually be observed more ventromedially as well (Figs. 2A, 3A). The terminations in VP1 were patchy and when the PHA-L sections (Figs. 2A,D, 3A,D) were compared to adjacent sections reacted to show NT-IR (Figs. 2B,E, 3B,E), it was noted that the PHA-L patches were in register with openings embellishing the dorsolateral margin of the dense NT-IR, which is a marker of the ventromedial pallidal compartment. Bundles of labeled axons issuing from injection sites in the accumbal core produced two aggregates of very dense terminations reflected as rostrolateral and caudomedial patches (Figs. 2A,D, 3A,D). These were separated by an interval containing much less dense terminations. Injections placed more caudally in the accumbal core resulted in dual aggregates of striatopallidal terminations that were located more caudally in ventral pallidum (Fig. 3). Beyond the caudal boundary of VPI the axon fascicles exhibited few terminals and proceeded to more distal sites of termination as described by Heimer et al. ('gob). The immunohistochemical compartmentation of ventral pallidum can be appreciated by comparing panels B and E of Figures 2-4, which show sections processed to display NT-IR with panels C and F, which show the adjacent sections processed to display substance P-IR and so exhibit the full extent of the forebrain region considered to represent ventral pallidum. Shell injections (Fig. lC,C'), conversely, resulted in transport of PHA-L to terminal areas that are located ventrally and medially in ventral pallidum (VPm) and, consequently, after shell injections the PHA-L labeled terminal fields were in register with the dense NT-IR exhibited in adjacent sections through VPm (Fig. 4A-F). Also after shell injections, patchiness and rostrocaudal topography were not observed. On the contrary, fascicles of labeled axons leaving injection sites in the shell entered VPm and produced an abundance of labeled terminals, which continued as a broad column through the length of VPm and into forebrain regions beyond pallidum (Heimer et al., '90b). Rostral and
TWO VENTRAL TRANSPALLIDAL PATHWAYS
Fig. 1. Frontal sections through the rat forebrain at a level displaying the caudate-putamen (CPu), nucleus accumbens (Acb), and olfactory tubercle (Tu). A,B and C indicate the positions of PHA-L deposition (arrows) that resulted in the anterograde transport illustrated in Figures 2 , 3 , and 4, respectively. A',B', and C' show substance
439
P (SP) immunoreactivity, which delineates the core and shell regions of the nucleus accumbens. Note that the PHA-L is present in the accumbal core in A and B and in the shell in C. The injection in B is at a more caudal level of the core than is the one in A. Abbreviation: ac-anterior commissure. Scale bar: 1.0 mm.
D.S. ZAHM AND L. HEIMER
440
Figure 2
TWO VENTRAL TRANSPALLIDAL PATHWAYS lateral parts of the shell project to the polymorph layer of the olfactory tubercle and to its caudal continuation represented by most ventral and lateral parts of VPm, i.e., subdistricts also innervated by the striatal parts of the olfactory tubercle (Heimer et al., '87). The distribution of labeled striatopallidal projections following shell injections was similar in untreated rats and rats that were acutely administered haloperidol (Fig. 4)or vehicle. This observation strongly suggests that the dramatic enhancement of NT-IR observed in VP1 following the drug treatment (Eggerman and Zahm, '88) is not a reflection of rapid sprouting of shell projections into VP1.
DISCUSSION The results of this study establish that two immunohistochemically distinct subterritories of the nucleus accumbens, its core and shell, project in a minimally overlapping fashion upon the dorsolateral (VPl) and ventromedial (VPm) subterritories, respectively, of the subcommissural ventral pallidum (VP)as they were defined immunohistochemically by Zahm and Heimer ('88).This observation gains support from recent investigations utilizing the retrograde transport of fluorogold (Churchill et al.,'90; Heimer et al., '90b). The consequent distribution among downstream structures of influences conveyed by these segregated striatopallidal projections depends upon the projection patterns of VP1 and VPm. Whereas it would appear that all of ventral pallidum sends fibers to the thalamic mediodorsal nucleus (Young et al., '84), a recent investigation (Zahm, '89) has indicated that VPl and VPm otherwise project to different structures in the diencephalon and mesencephalon. According to the results of the latter investigation, the more caudally directed projections originating in VPl terminate in the subthalamic nucleus (STN) and substantia nigra (SN), whereas those originating in VPm terminate in the midbrain ventral tegmental area (VTA), but not STN or SN. The summed results of these studies establish a case for the existence of distinct transpallidal pathways originatingin the accumbal core and shell as indicated diagrammatically in Figure 5 . Since the pallidal projections depicted in Figure 5 derive only from the results of retrograde labeling
Fig. 2. Adjacent sections through the ventral pallidum of the brain receiving the injection shown in Figure 1A. PHA-L, neurotensin (NT) and substance P (SP) immunoreactivities are shown in the upper (A,D), middle (B,E), and lower (C,F) panels, respectively. The left (A,B,C)and right (D,E,F) panels show, respectively, rostra1 (rost.) and caudal (caud.) levels through the pallidum. The striatopallidal projection pattern displayed by the anterogradely transported PHA-L (white in darkfield micrographs in A and D) from this core injection consists of two aggregates of axon terminals, a rostrolateral patch (A and D, number 1)and a caudomedial patch (D, number 21,which are confined to ventral pallidum. In D, only a sparse remnant of the rostrolateral patch is visible. The middle panels show adjacent sections reacted to display NT-IR and illustrate that following core injections the PHA-L labeled terminal fields are present in dorsolateral ventral pallidum (W1)and in register with openings in the NT-IR. The positions of the bundles of fascicles giving rise to the terminal fields are marked by the letter f i n the upper and middle panels. Note that in these sections, as following all core injections, the patches of PHA-L labeled terminals are dorsolateral to the bundles. The lower panels in Figures 2 4 show substance P immunoreactivity, which delineates all of ventral pallidum. Abbreviations: ac-anterior commissure; acp-posterior limb of anterior commissure; VPm-ventromedial ventral pallidum. Scale bar: 1.0 mm.
44 1
after injections in STN, SN and VTA (Zahm, '891, the possibility must be considered that additional projections from these subdistricts will be revealed following appropriate anterograde studies. Although synaptic contacts between accumbal afferents and dorsolateral or ventromedial ventral pallidal projection neurons have yet to be established experimentally, the presence of such synapses is favored by the high degree of registration of the core- and shell-derived striatopallidal terminals with the positions of one or the other subgroup of ventral pallidal projection neurons, as judged by the immunohistochemical compartmentation. Even in the unlikely absence of monosynaptic connectivity between these hodologically aligned structures, segregation of the "conduits" linking the immunohistochemically distinct subterritories should, on it own merit, bear major functional consequence. Ventral pallidum and ventral striatum, since they were first described nearly 20 years ago (Heimer and Wilson, '751, have increasingly gained acceptance as integral components of the basal ganglia and certainly the immunocytochemical staining patterns, ultrastructural features, and downstream connections of VP1, e.g., with the substantia nigra and subthalamic nucleus, support the affiliation of VP1 in this manner (Zahm and Heimer, '88; Zahm, '89). VPm, however, based upon its immunohistochemical and ultrastructural features, failure to project to the STN, and the concentration of its mesencephalic projection in the VTA, would appear to represent a somewhat different specialization of pallidum. This consideration is strengthened by the observation that the VTA-projecting cells in VPm represent one sector of a larger group of neurons scattered among the shared boundaries of several forebrain structures including the bed nucleus of stria terminalis, lateral and medial preoptic areas, and lateral hypothalamus. As described in greater detail in a subsequent paper (Heimer et al., 'gob), the projection from the accumbal shell appears suitably disposed to influence most, if not all, of the cells in the forebrain VTA-projecting cluster, a circumstance tending to implicate VPm, at least in part, as an appendage of a larger VTA-projecting system. The present results serve to associate, by virtue of their striatopallidal projections, the accumbal core with VP1 and the shell with VPm. Moreover, like VPl more resembles globus pallidus than does VPm, the accumbal core has more features tending to ally it with the adjacent caudateputamen than does the shell. First, the accumbal core sends to VP1 spatially distinct, dual aggregates of axon terminals reminiscent of the neostriatopallidal projection pattern (Chang et al., '81; Wilson and Phelan, '82; Smith and Parent, '86). Furthermore, immunoperoxidase preparations for substance P (Zahm and Heimer, '88; Voorn et al., '89; Zahm, '89, '90; Churchill et al., 'gob), enkephalin (Voorn et al., '87, '89; Zahm, '891, neurotensin (Zahm and Heimer, '88; Zahm, '891, cholecystokinin (Zaborszky et al., '86; Zahm and Heimer, '88), 28 kD calcium binding protein (Voorn et al.,'89; Zahm, '90) tyrosine hydroxylase (Voorn et al., '86; Zahm and Johnson, '891, dopamine (Voorn et al., '86, '891, and autoradiograms displaying receptor binding of gabaergic and opiatergic ligands (Hammer, '89; Churchill et al., '90a,b) each clearly distinguish core and shell, and the pattern in core more resembles that seen in caudateputamen in each case. Neurotensin and cholecystokinin immunoreactivities, which are poorly represented in caudate-putamen, at least in the rat, are richly distributed in a considerable portion of the shell (Kalivas and Miller, '84; Zaborszky et al., '86; Zahm and Heimer, '88). Finally, the
442
Fig. 3. Adjacent sections through the ventral pallidum of the brain receiving the injection shown in Figure 1B. Note that following this more caudally placed injection the initial aggregate of PHA-L labeled terminals is present at a more caudal level of ventral pallidum. Two
D.S. ZAHM AND L. HEWER
such aggregates, nevertheless, are present (1 and 2) and the terminal patches are present dorsal and lateral to the fiber bundle (0.Scale bar: 1.0 mm.
TWO VENTRAL TRANSPALLIDAL PATHWAYS
Fig. 4. Adjacent sections through the ventral pallidum of the brains receiving the injections shown in Figure 1C. The layout of the plate, symbols and abbreviations are identical to those in Figure 2 and 3. Shell injections give rise to a single column of terminals that extends the length of ventral pallidum and beyond (A and D, number 1) and is located in ventromedial ventral pallidum ( W m ) in register with the dense NT-IR. Note that the fascicles are lateral and/or dorsal to the
443
terminal fields after shell injections. Bands of strong NT-IR in B and E are the result of haloperidol administration and are denoted with hal. These strong dorsolateral bands of NT-IR enhance the distinction between ventromedial ( W m ) and dorsolateral (WI)ventral pallidum making it clear that the PHA-L labeled terminals in Figure 4 are present in W m . Scale bar: 1.0 mm.
Figure 5
P
m
TWO VENTRAL TRANSPALLIDAL PATHWAYS accumbal shell is the major striatal recipient of hippocampal input (Kelley and Domesick, '82; Groenewegen et al., '87; Quirk and Groves, '89) and recent evidence suggests that the single unit electrophysiological responses of striatal neurons to subicular stimulations differ in the core and shell (Boeijinga et al., '90). The synaptic organization and relative susceptibility to neurotoxic perturbation of the dopamine-containinginputs to core and shell also appear to differ significantly. Thus it was demonstrated that in accumbal core and in the caudateputamen, the frequency of axospinous and axodendritic synapses with smooth dendritic segments is about equal. In the shell, in contrast, synaptic contacts with the smooth segments of dendrites (between spines or on aspiny dendrites) are significantly more numerous than are contacts with spines (Zahm and Haycock, '89). Previous indications that the nucleus accumbens is less sensitive to the tyrosine hydroxylase immunoreactivity depleting effects of mesencephalic 6-hydroxydopamine (6-OHDA) injections (Voorn et al., '87) were subsequently confirmed (Zahm and Johnson, '89), and it was noted later (Zahm, '90) that the mesostriatal neurons apparently resistant to the toxin selectively innervate the accumbal shell and medial parts of the olfactory tubercle. Mesostriatal fibers and terminals in the accumbal core, like those in the caudate-putamen (with the exception of those innervating the striatal patches), are depleted of tyrosine hydroxylase immunoreactivity within 24 hours of the 6-OHDA injection. Moreover, the capacity of striatal neurons to accumulate detectable neurotensin immunoreactivity following haloperidol administration, an indirect indication of the perturbed status of dopaminergic neurotransmission (Eggerman and Zahm, '881, is apparently greatest in accumbal shell and olfactory tubercle, whereas the increases in numbers of detectable cells in accumbal core resemble those observed in most parts of the caudate putamen (Zahm, '90a). Collectively, these results establish the identities of two adjacent, if not parallel, systems of connections originating in the rat nucleus accumbens and projecting farther downstream via relays in the ventral pallidum (Fig. 5 ) . The accumbal core-VP1-substantia nigraisubthalamic nucleus connections, as described above, resemble other transpallidal subsystems of the basal ganglia, of which several are currently being considered in more or less specific functional terms (see Alexander et al., '86). The pathway originating in the accumbal shell and projecting via VPm, in contrast, fails to reach the substantia nigra and subthalamic nucleus, making it difficult to categorize it similarly.
Fig. 5. Summary diagram depicting five levels of the rat brain c u t in the frontal plane (A-E) and two adjacent transpallidal pathways originating in t h e nucleus accumbens. One pathway originates i n t h e accumbal core (broken line) and passes to t h e dorsolateral part of ventral pallidum (VPI). Neurons in VP1 can be retrogradely labeled after injections in t h e mediodorsal thalamic nucleus, subthalamic nucleus and substantia nigra, b u t not after injections confined to t h e ventral tegmental area. The other pathway originates in t h e accumbal shell (unbroken line) and projects to t h e ventromedial pallidal subterritory (VPm). Neurons in VF'm a r e retrogradely labeled after injections in t h e mediodorsal thalamic nucleus and ventral tegmental area, b u t not after injections involving t h e subthalamic nucleus. Abbreviations: ac-anterior commissure; Acb-nucleus accumbens; f-fornix; frfasciculus retroflexus; lot-lateral olfactory tract; MD-mediodorsal thalamic nucleus; ml-medial lemniscus; sm-stria medullaris; SNsubstantia nigra; STN-subthalamic nucleus; Tu-olfactory tubercle; VTA-ventral tegmental area.
445
Other efferent targets of the accumbal shell that are not treated in this study, e.g. in the hypothalamus and the sublenticular gray, are not innervated by accumbal core (Heimer et al., 'gob), buttressing the suggestion that accumbal shell, but not the core, contains in significant portion elements of the recently described "extended amygdala" system (Alheid and Heimer, '88; Heimer et al., '90a). The results reported here are limited to confirmation of a projection from accumbal core to VP1 and from accumbal shell to VPm. The latter projection is aligned appropriately to exert direct synaptic influence on a projection from VPm to the VTA. The neurochemical and hodological character of this transpallidal pathway is sufficiently different from that of other pathways leaving the striatal complex to indicate that it represents a uniquely specialized part of the basal ganglia.
ACKN0WLEM;;MENTS This work was supported by USPHS grants NIH NS23805 (D.S.Z.) and NIH NS-17743 (L.H.). The authors gratefully acknowledge the expert technical assistance of Evelyn Williams and the clerical support of Victoria Cardillo.
LITERAmcm Alexander, G.E., M.R. DeLong, and P.L. Strick (1986) Parallel organization of functionally segregated circuits linking basal ganglia and cortex. Ann. Rev. of Neurosci. 9:351-381. Alheid, G.F., and L. Heimer (1988) New perspectives in basal forebrain organization of special relevance for neuropsychiatric disorders; the striatopallidal, amygdaloid, and corticopetal components of substantia innominata. Neuroscience 27: 1-39. Boeijinga, P.H., C.M.A. Pennartz, and F.H. Lopes da Silva (1990) Pairedpulse facilitation in the nucleus accumbens following stimulation of subicular inputs in the rat. Neuroscience 35~301-311. 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 Res. 11 7:423-435. Carter, D.A., and H.C. Fibiger (1978) The projections of the entopeduncalar nucleus and globus pallidus in rat as demonstrated by autoradiography and horseradish peroxidase histochemistry. J. Comp. Neurol. 177:113124. Chang, H.T., C.J. Wilson, and S.T. Kitai (1981) Single neostriatal efferent axons in the globs pallidus; a light and electron microscopic study. Science 213:915-918. Churcbill, L., R.P. Dilts, and P.W. Kalivas (199Oa) Changes in gammaaminobutyric acid, mu-opiodand neurotensin receptors in theaccumbenspallidal projection after discrete quinolinic acid lesions in the nucleus accumbens. Brain Res. 51 1:41-54. Churchill, L., R.S. Cross, T.L. Pazdernick, S.R. Nelson, D.S. Zahm, L. Heimer, R.P. Dilts, and P.W. Kalivas (1990b) Patterns of glucose use after bicuculline-induced convulsions provides a functional marker for the ventral pallidum. Neuroscience (in review). Eggerman, K.W., and D.S. Zahm (1988) Numbers of neurotensin-immunoreactive neurons selectively increased in rat ventral striatum following acute haloperidol administration. Neuropeptides 11.125-132. Groenewegen, H.J., and F.T. Russchen (1984) Organization of the efferent projections of the nucleus accumbens to pallidal, hypothalamic and mesencephalic structures: A tracing and immunohistochemical study in the cat. J. Comp. Neurol. 223:347-367. Groenewegen, H.J., E. Vermeulen-Van der Zee, A. te Kortschot, and M.P. Witter (1987) Organization of the projections from the subiculum to the ventral striatum in the rat. A study using anterograde transport of Phaseolus-uulgaris-leucoagglutinin. Neuroscience 23: 103-120. Groenewegen, H.J., G.E. Meredith, H.W. Berendse, P. Voorn, and J.G. Wolters (1989) The compartmental organization of the ventral striatum in the rat. In A.R. Crossman and M.A. Sambrook (ed): Neural Mechanisms in Disorders of Movement. London: John Libbey, pp. 45-54.
446 Groenewegen, H.J., H.W. Berendse, G.E. Meredith, S.N. Haber, P. Voorn, J.G. Wolters, and A.H.M. Lohman (1991) Functional anatomy of the ventral, limbic system-innervated striatum. In P. Willner and J. ScheelKruger (eds): The Mesolimbic Dopamine System: From Motivation to Action. Chichester, England: John Wiley & Sons, (in press). Haber, S.N., Groenewegen, H.J., E.A. Grove, and W.J.H. Nauta (1985) Efferent connections of the ventral pallidum: Evidence of a dual striatopallidofugal pathway. J. Comp. Neurol. 235:322-335. Haber, S.N., E. Lynd, C. Klein, and H.C. Groenewegen (1990) Topographic organization of the ventral striatal efferent projections in the rhesus monkey: An anterograde tracing study. J. Comp. Neurol. 293:282-298. Hammer, R.P., Jr. (1989) Cocaine alters opiate receptor binding in critical brain reward regions. Synapse 3:5540. Hattori, T., H.C. Fibiger, and P.L. McGeer (1975) Demonstration of a pallido-nigral projection innervating dopaminergic neurons. J. Comp. Neurol. 162r487-504. Heimer, L., and R.D. Wilson (1975) The subcortical projections of allocortex. Similarities in the neural associations of the hippocampus, the piriform cortex and the neocortex. In M. Santini (ed): Golgi Centennial Symposium Proceedings. New York Raven Press, pp. 177-193. Heimer, L., J. de Olmos, G.F. Alheid, and L. Zaborszky (1990a) “Perestroika” in the basal forebrain; opening the border between neurology and psychiatry. Progress in Brain Research (in press). Heimer, L., L. Zaborszky, D.S. Zahm, and G.F. Alheid (1987) The ventral striato-pallidothalamic projection: I. The striatopallidal link originating in the superficial layers of the olfactory tubercle. J. Comp. Neurol. 255571-591. Heimer, L., D.S. Zahm, L. Churchill, P.W. Kalivas, and C. Wohltmann (1990b) Specificity in the projection patterns of accumbal core and shell. Neuroscience (in press). Itoh, Z., K. Akiva, S. Nomura, N. Mizuno, Y. Hakamura, and T. Sugimoto (1979) Application of a coupled oxidation reaction to electron microscopic demonstration of horseradish peroxidase: Cobalt-glucose oxidase method. Brain Res. 175341-347. Kalivas, P.W., and J.S. Miller (1984) Neurotensin neurons in the ventral tegmental area project to the medial accumbens. Brain Res. 300;157160. Kelley, A.E., and V.B. Domesick (1982) The distribution of the projection from the hippocampal formation to the nucleus accumbens in the rat: An anterograde and retrograde horseradish peroxidase study. Neuroscience. 7:3221-2335. Meredith, G.E., B. Blank, and H.J. Groenewegen (1989) The distribution and compartmental organization of the cholinergic neurons in nucleus accumbens of the rat. Neuroscience 31:3278-345. Mogenson, G.J., L.W. Swanson, and M. Wu (1983) Neural projections from nucleus accumbens to globus pallidus, substantia innominata, and lateral preoptic area: an anatomical and electrophysiological investigation in the rat. J. Neurosci. 3:189-202. Nauta, W.J.H., G.P. Smith, R.L.M. Faul1,andV.B. Domesick(1978) Efferent connections and nigral efferents of the nucleus accumbens septi in the rat. Neuroscience 3:385-410.
D.S. ZAHM AND L. HEIMER Quirk, T.L., and P.M. Groves (1989) Hippocampal projections to the rat nucleus accumbens: A light and electron microscopic analysis using Neurosci Abstr. 15r1244. PHA-L. SOC. Smith, Y., and A. Parent (1986) Differential connections of caudate nucleus and putamen in the squirrel monkey (Saimiri sciureus). Neuroscience 18r347-371. Voorn, P., C.R. Gerfen, and H.J. Groenewegen (1989) Compartmental organization of the ventral striatum of the rat: Immunohistochemical distribution of enkephalin, substance P, dopamine and calcium-binding protein. J. Comp. Neurol. 289:189-201. Voorn, P., G. Roest, and H.J. Groenewegen (1987) Increase of enkephalin and decrease of substance P immunoreactivity in the dorsal and ventral striatum of the rat after midbrain 6- hydroxydopamine lesions. Brain Res. 412t391-396. Voorn, P., B. Jorritsma-Byham, C. Van Dijk, and R.M. Buijs (1986) The dopaminergic innervation of the ventral striatum in the rat: A light and electron microscopical study with antibodies against dopamine. J. Comp. Neurol. 251r84-99. Wilson, C.J., and K.D. Phelan (1982) Dual topographic representation of neostriatum in the globus pallidus of rats. Brain Res. 243r354-359. Young, W.S., Jr., G.F. Alheid, and L. Heimer (1984) The ventral pallidal projection to the mediodorsal thalamus: A study with fluorescent retrograde tracers and immunohistofluorescence. J. Neurosci. 4: 162% 1638. Zaborszky, L., G.F. Alheid, V.E. Alones, W.W. Oertel, D.E. Schmechel, and L. Heimer (1982) M e r e n t s of the ventral pallidum studied with a combined immunohistochemical-anterogradedegeneration method. SOC. Neurosci. Abstr. 8:218. Zaborszky, L., G.F. Alheid, M.C. Beinfeld, L.E. Eiden, L. Heimer, and M. Palkovits (1985) Cholecystokinin innervation of the ventral striatum: A morphological and radioimmunological study. Neuroscience 14t427453. Zahm, D.S. (1989) Ventral striatopallidal parts of the basal ganglia in the rat: 11. Compartmentation of ventral pallidal efferents. Neurosci. 30:33-50. Zahm, D.S. (1990) Compartmental and subterritorial organization of dorsal and ventral striatum revealed following acute perturbations of the mesostriatal projections in rats. J. Comp. Neurol. (in review). Zahm, D.S., and J.W. Haycock (1989) On-line morphometry of tyrosine hydroxylase-immunoreactive axon terminals in caudate putamen and nucleus accumbens in rat with a focus on synapses. SOC.Neurosci. Abst. 15r905. Zahm, D.S., and L. Heimer (1988) Ventral striatopallidal parts of the basal ganglia in the rat: I. Neurochemical compartmentation as reflected by the distributions of neurotensin and substance P immunoreactivity. J. Comp. Neurol. 272516-535. Zahm, D.S., and Johnson, S.N. (1989)Asymmetrical distribution of neurotensin immunoreactivity following unilateral injection of 6-hydroxydopamine in rat ventral tegmental area (VTA). Brain Res. 483r301-311. Zahm, D.S., L. Zaborszky, G.F. Alheid, and L. Heimer (1987) The ventral striatopallidothalamic projection: 11. The ventral pallidothalamic link. J. Comp. Neurol. 255;592-605.
TwoTranspallidalPathwaysOriinatingin the Rat Nucleus Accumbens D.S. ZAHM AND LENNART HEIMER Department of Anatomy and Neurobiology St. Louis University School of Medicine, St. Louis, Missouri 63104 (D.S.Z.) and Departments of Otolaryngology and Neurosurgery, University of Virginia School of Medicine, Charlottesville, Virginia 22908 (L.H.)
ABSTRACT The striatopallidal projection originating in the nucleus accumbens was investigated by using anterograde transport of PHA-L in combination with peptide immunohistochemistry in order to localize the injection sites and transported lectin with respect to neurochemically defined subterritories in the nucleus accumbens and subcommissural ventral pallidum. The results reported here supplement our previous observations, which indicated that the subcommissural ventral pallidum of the rat comprises two immunohistochemically defined subterritories (Zahm and Heimer, '88: J. Comp. Neurol., 272516435) which give rise to dichotomous downstream projection systems (Zahm, '89: Neuroscience, 30:33-50). The present data indicate that the neurotensin immunoreactivity-rich ventromedial district of ventral pallidum receives its accumbal input almost exclusively from the shell district of the nucleus accumbens. The accumbal core, alternatively, projects to the dorsolateral ventral pallidal subterritory that lacks appreciable neurotensin immunoreactivity and in many other respects more resembles the adjoining striatopallidal components of the caudate-putamen. In addition to direct topographic relationships in the frontal plane among the accumbal injection sites and ventral pallidal terminations, it was observed that more caudally placed core injections resulted in patches of striatopallidal terminations that were more caudally located in ventral pallidum. Shell injections, in contrast, produced columns of terminations that extended continuously from the rostralmost level that they appeared to the caudal end Df ventromedial ventral pallidum. The accumbal shell, its exclusive projection to the ventromedial subterritory in the subcommissural ventral pallidum, and the previously reported, almost exclusive projection of that pallidal subdistrict to the mesencephalicventral tegmental area are discussed in terms of a number of other neurochemical and hodological features that serve to distinguish them sufficientlyto suggest that they represent a uniquely specialized part of the basal ganglia. Key words: ventral pallidum, ventral striatum, neurotensin, substance P, PHA-L tract tracing, immunocytochemistry
Recent immunohistochemical studies of the ventral parts of the basal ganglia in the rat have demonstrated that ventral pallidum (VP)dorsal to the polymorph layer of the olfactory tubercle is divisible into two subdistricts, a ventromedial part that is rich in neurotensin (NT) immunoreactive axons and axon terminals, and a dorsolateral part that exhibits a significantlylesser amount of NT immunoreactivity (Zahm and Heimer, '88). Subsequently, investigation of wheat germ agglutinin HRP conjugate transported retrogradely from structures known to receive inputs from pallidum revealed that the projection systems arising in the two ventral pallidal subdistricts are different (Zahm, '89). Whereas dorsolateral VP projects relatively massively to the substantia nigra and subthalamic nucleus, well-known targets of pallidum in the rat (Hattori et al., '75; Bunney and Aghajanian, '76; Carter and Fibiger, '78; Haber et al.,
o 1990 WILEY-LISS, INC.
'85), the ventromedial district of VP gives rise to a mesencephalic projection mostly restricted to the ventral tegmental area (VTA) and projects insignificantly, if at all, to the subthalamic nucleus. Moreover, ventral pallidal neurons projecting to the substantia nigra and subthalamic nucleus are typical "pallidal" neurons in the sense that they receive a very dense afferent input that is reflected ultrastructurally in numerous boutons contacting even their proximal dendrites and perikarya in a pallisadelike fashion (for references, see Zahm et al., '87). The VTA-projecting
Accepted September 1,1990 Address reprint requests to D.S. Zahm, Department of Anatomy and Neurobiology, St. Louis University School of Medicine, 1402 S. Grand Blvd., St. Louis, MO 63104.
D.S. ZAHM AND L. HEIMER
438
neurons, conversely, are contacted by relatively few boutons at the level of their proximal dendrites and perikarya. The fact that the two ventral pallidal compartments give rise to dichotomous projections fostered a suspicion that the striatal inputs from the nucleus accumbens (Nauta et al., '78; Zaborszky et al., '82; Mogenson et al., '83; Groenewegen and Russchen, '84; Haber et al., '90) to the two compartments might also differ. Since recent observations indicate that the nucleus accumbens exhibits a rather conspicuous compartmentation reflected in "core" and "shell" subdistricts (Zaborzsky et al., '85; Voorn et al., '86, '89; Groenewegen et al., '89, '90; Meredith et al., '89; Heimer et al., '90a; Zahm, 'go), it seemed appropriate to determine if each of the accumbal compartments project to the different pallidal compartments identified on the basis of NT-IR. To address this question, the striatopallidal projection originating in the nucleus accumbens was investigated by using anterograde transport of PHA-L in combination with peptide immunohistochemistry to localize the injection sites and transported lectin with respect to the accumbal and pallidal compartments.
MATERTALSIWDMETHODS With the aid of a Kopf stereotaxic instrument, small injections of Phaseolus vulgaris leucoagglutinin (PHA-L, 17 rats) were placed unilaterally in the ventral striatum of pentobarbital anesthetized male Sprague-Dawley rats weighing 225-250 g. The tracers were delivered iontophoretically with positive pulses of 4-7 pA (7s on/7s off) from 2% solutions. Seven days later, the rats were killed under deep pentobarbital anesthesia by vascular perfusion with an aqueous solution containing 0.1% glutaraldehyde, 4% paraformaldehyde, and 2.5% polyvinylpyrrolidone (PVP40) in 0.07 M Sorenson's phosphate buffer (SPB, pH 7.41, preceded by a brief rinse with 2.5% PVP-40 and 0.5% procaine-HC1in sodium phosphate-buffered saline (NaPBS). Fifty micrometer frontal sections cut on the vibratome were subjected to an immunocytochemical protocol for PHA-L, neurotensin (NT) or substance P (SP). Several of the rats received subcutaneous injections of haloperidol (2 mg/kg in a solution containing 50% propylene glycol, 20% ethanol, and 30% distilled water) or the vehicle 18 and 4 hours before sacrifice to enhance the neurotensin immunoreactivity in the dorsolateral ventral pallidal subdistrict (Eggerman and Zahm, '88). Sections were pretreated with 1%sodium borohydride for 20 minutes, 1% H,O, for 10 minutes, and appropriate normal serum for ?hhour, after which they were placed without further rinsing into vials containing 1-2 ml of SPB with 1%appropriate normal serum and primary antisera at the following dilutions: anti-PHA-L, 1:2000; anti-NT, 15000; anti-SP, 1:lOOO-1: 2000. The sections were left on a rocker plate overnight at 4°C. The following morning the sections were rinsed and placed in vials containing SPB, 1% rabbit serum and biotinylated second antibodies at a dilution of 1:200. Bound anti-PHA-L was identified with antigoat, and anti-SP with antirat immunoglobulins (IgG) both produced in rabbit (Vector). Anti-NT was recognized with antirabbit IgGs made in goat. After three rinses, the sections were incubated for 1hour in SPB containing 1%appropriate normal serum and avidinbiotin-peroxidase complex (Vector) at a dilution of 1:200. Immunolabeling was produced with diaminobenzidine (DAB) by modification of the coi1.pled glucose oxidase reaction (Itoh et al., '79; Zahm and Heimer,
'88) and enhanced with osmium and thiocarbohydrazide (C.R. Gerfen, unpublished). The sections were then mounted in rostrocaudal sequence on glass slides coated with gelatin (0.2%) and chrome alum (0.02%) and coverslipped under Permount. Attempts to produce immunolabeling with preabsorbed antisera (50 pgiml) or buffer lacking primary antibody resulted in absence of staining, whereas antisera that were preincubated with an inappropriate antigen (e.g., anti-PHA-L pre-incubated with SP or the converse) produced normal staining patterns.
RESULTS PHA-L injections into the nucleus accumbens resulted in transport of the lectin to axon terminations in ventral pallidum (VP) that when displayed in frontal sections exhibited direct topographic relationships to the positions of the injection sites (Figs. 1-4). Injections in the core of the nucleus accumbens as demonstrated with SP-IR (Figs. 1A-B') labeled axon terminal fields in dorsolateral parts of the subcommissural VP (Figs. 2, 3) and those in the shell (Fig. lC, C) labeled fields in ventromedial VP (Fig. 4). Following injections involving exclusively the core of nucleus accumbens (Figs. 1A-B'), dense aggregates of PHA-L labeled axon terminals and fibers were located almost exclusively in the dorsolateral compartment of ventral pallidum (VPl), although a sparse to moderate amount of labeled fibers and terminals could usually be observed more ventromedially as well (Figs. 2A, 3A). The terminations in VP1 were patchy and when the PHA-L sections (Figs. 2A,D, 3A,D) were compared to adjacent sections reacted to show NT-IR (Figs. 2B,E, 3B,E), it was noted that the PHA-L patches were in register with openings embellishing the dorsolateral margin of the dense NT-IR, which is a marker of the ventromedial pallidal compartment. Bundles of labeled axons issuing from injection sites in the accumbal core produced two aggregates of very dense terminations reflected as rostrolateral and caudomedial patches (Figs. 2A,D, 3A,D). These were separated by an interval containing much less dense terminations. Injections placed more caudally in the accumbal core resulted in dual aggregates of striatopallidal terminations that were located more caudally in ventral pallidum (Fig. 3). Beyond the caudal boundary of VPI the axon fascicles exhibited few terminals and proceeded to more distal sites of termination as described by Heimer et al. ('gob). The immunohistochemical compartmentation of ventral pallidum can be appreciated by comparing panels B and E of Figures 2-4, which show sections processed to display NT-IR with panels C and F, which show the adjacent sections processed to display substance P-IR and so exhibit the full extent of the forebrain region considered to represent ventral pallidum. Shell injections (Fig. lC,C'), conversely, resulted in transport of PHA-L to terminal areas that are located ventrally and medially in ventral pallidum (VPm) and, consequently, after shell injections the PHA-L labeled terminal fields were in register with the dense NT-IR exhibited in adjacent sections through VPm (Fig. 4A-F). Also after shell injections, patchiness and rostrocaudal topography were not observed. On the contrary, fascicles of labeled axons leaving injection sites in the shell entered VPm and produced an abundance of labeled terminals, which continued as a broad column through the length of VPm and into forebrain regions beyond pallidum (Heimer et al., '90b). Rostral and
TWO VENTRAL TRANSPALLIDAL PATHWAYS
Fig. 1. Frontal sections through the rat forebrain at a level displaying the caudate-putamen (CPu), nucleus accumbens (Acb), and olfactory tubercle (Tu). A,B and C indicate the positions of PHA-L deposition (arrows) that resulted in the anterograde transport illustrated in Figures 2 , 3 , and 4, respectively. A',B', and C' show substance
439
P (SP) immunoreactivity, which delineates the core and shell regions of the nucleus accumbens. Note that the PHA-L is present in the accumbal core in A and B and in the shell in C. The injection in B is at a more caudal level of the core than is the one in A. Abbreviation: ac-anterior commissure. Scale bar: 1.0 mm.
D.S. ZAHM AND L. HEIMER
440
Figure 2
TWO VENTRAL TRANSPALLIDAL PATHWAYS lateral parts of the shell project to the polymorph layer of the olfactory tubercle and to its caudal continuation represented by most ventral and lateral parts of VPm, i.e., subdistricts also innervated by the striatal parts of the olfactory tubercle (Heimer et al., '87). The distribution of labeled striatopallidal projections following shell injections was similar in untreated rats and rats that were acutely administered haloperidol (Fig. 4)or vehicle. This observation strongly suggests that the dramatic enhancement of NT-IR observed in VP1 following the drug treatment (Eggerman and Zahm, '88) is not a reflection of rapid sprouting of shell projections into VP1.
DISCUSSION The results of this study establish that two immunohistochemically distinct subterritories of the nucleus accumbens, its core and shell, project in a minimally overlapping fashion upon the dorsolateral (VPl) and ventromedial (VPm) subterritories, respectively, of the subcommissural ventral pallidum (VP)as they were defined immunohistochemically by Zahm and Heimer ('88).This observation gains support from recent investigations utilizing the retrograde transport of fluorogold (Churchill et al.,'90; Heimer et al., '90b). The consequent distribution among downstream structures of influences conveyed by these segregated striatopallidal projections depends upon the projection patterns of VP1 and VPm. Whereas it would appear that all of ventral pallidum sends fibers to the thalamic mediodorsal nucleus (Young et al., '84), a recent investigation (Zahm, '89) has indicated that VPl and VPm otherwise project to different structures in the diencephalon and mesencephalon. According to the results of the latter investigation, the more caudally directed projections originating in VPl terminate in the subthalamic nucleus (STN) and substantia nigra (SN), whereas those originating in VPm terminate in the midbrain ventral tegmental area (VTA), but not STN or SN. The summed results of these studies establish a case for the existence of distinct transpallidal pathways originatingin the accumbal core and shell as indicated diagrammatically in Figure 5 . Since the pallidal projections depicted in Figure 5 derive only from the results of retrograde labeling
Fig. 2. Adjacent sections through the ventral pallidum of the brain receiving the injection shown in Figure 1A. PHA-L, neurotensin (NT) and substance P (SP) immunoreactivities are shown in the upper (A,D), middle (B,E), and lower (C,F) panels, respectively. The left (A,B,C)and right (D,E,F) panels show, respectively, rostra1 (rost.) and caudal (caud.) levels through the pallidum. The striatopallidal projection pattern displayed by the anterogradely transported PHA-L (white in darkfield micrographs in A and D) from this core injection consists of two aggregates of axon terminals, a rostrolateral patch (A and D, number 1)and a caudomedial patch (D, number 21,which are confined to ventral pallidum. In D, only a sparse remnant of the rostrolateral patch is visible. The middle panels show adjacent sections reacted to display NT-IR and illustrate that following core injections the PHA-L labeled terminal fields are present in dorsolateral ventral pallidum (W1)and in register with openings in the NT-IR. The positions of the bundles of fascicles giving rise to the terminal fields are marked by the letter f i n the upper and middle panels. Note that in these sections, as following all core injections, the patches of PHA-L labeled terminals are dorsolateral to the bundles. The lower panels in Figures 2 4 show substance P immunoreactivity, which delineates all of ventral pallidum. Abbreviations: ac-anterior commissure; acp-posterior limb of anterior commissure; VPm-ventromedial ventral pallidum. Scale bar: 1.0 mm.
44 1
after injections in STN, SN and VTA (Zahm, '891, the possibility must be considered that additional projections from these subdistricts will be revealed following appropriate anterograde studies. Although synaptic contacts between accumbal afferents and dorsolateral or ventromedial ventral pallidal projection neurons have yet to be established experimentally, the presence of such synapses is favored by the high degree of registration of the core- and shell-derived striatopallidal terminals with the positions of one or the other subgroup of ventral pallidal projection neurons, as judged by the immunohistochemical compartmentation. Even in the unlikely absence of monosynaptic connectivity between these hodologically aligned structures, segregation of the "conduits" linking the immunohistochemically distinct subterritories should, on it own merit, bear major functional consequence. Ventral pallidum and ventral striatum, since they were first described nearly 20 years ago (Heimer and Wilson, '751, have increasingly gained acceptance as integral components of the basal ganglia and certainly the immunocytochemical staining patterns, ultrastructural features, and downstream connections of VP1, e.g., with the substantia nigra and subthalamic nucleus, support the affiliation of VP1 in this manner (Zahm and Heimer, '88; Zahm, '89). VPm, however, based upon its immunohistochemical and ultrastructural features, failure to project to the STN, and the concentration of its mesencephalic projection in the VTA, would appear to represent a somewhat different specialization of pallidum. This consideration is strengthened by the observation that the VTA-projecting cells in VPm represent one sector of a larger group of neurons scattered among the shared boundaries of several forebrain structures including the bed nucleus of stria terminalis, lateral and medial preoptic areas, and lateral hypothalamus. As described in greater detail in a subsequent paper (Heimer et al., 'gob), the projection from the accumbal shell appears suitably disposed to influence most, if not all, of the cells in the forebrain VTA-projecting cluster, a circumstance tending to implicate VPm, at least in part, as an appendage of a larger VTA-projecting system. The present results serve to associate, by virtue of their striatopallidal projections, the accumbal core with VP1 and the shell with VPm. Moreover, like VPl more resembles globus pallidus than does VPm, the accumbal core has more features tending to ally it with the adjacent caudateputamen than does the shell. First, the accumbal core sends to VP1 spatially distinct, dual aggregates of axon terminals reminiscent of the neostriatopallidal projection pattern (Chang et al., '81; Wilson and Phelan, '82; Smith and Parent, '86). Furthermore, immunoperoxidase preparations for substance P (Zahm and Heimer, '88; Voorn et al., '89; Zahm, '89, '90; Churchill et al., 'gob), enkephalin (Voorn et al., '87, '89; Zahm, '891, neurotensin (Zahm and Heimer, '88; Zahm, '891, cholecystokinin (Zaborszky et al., '86; Zahm and Heimer, '88), 28 kD calcium binding protein (Voorn et al.,'89; Zahm, '90) tyrosine hydroxylase (Voorn et al., '86; Zahm and Johnson, '891, dopamine (Voorn et al., '86, '891, and autoradiograms displaying receptor binding of gabaergic and opiatergic ligands (Hammer, '89; Churchill et al., '90a,b) each clearly distinguish core and shell, and the pattern in core more resembles that seen in caudateputamen in each case. Neurotensin and cholecystokinin immunoreactivities, which are poorly represented in caudate-putamen, at least in the rat, are richly distributed in a considerable portion of the shell (Kalivas and Miller, '84; Zaborszky et al., '86; Zahm and Heimer, '88). Finally, the
442
Fig. 3. Adjacent sections through the ventral pallidum of the brain receiving the injection shown in Figure 1B. Note that following this more caudally placed injection the initial aggregate of PHA-L labeled terminals is present at a more caudal level of ventral pallidum. Two
D.S. ZAHM AND L. HEWER
such aggregates, nevertheless, are present (1 and 2) and the terminal patches are present dorsal and lateral to the fiber bundle (0.Scale bar: 1.0 mm.
TWO VENTRAL TRANSPALLIDAL PATHWAYS
Fig. 4. Adjacent sections through the ventral pallidum of the brains receiving the injections shown in Figure 1C. The layout of the plate, symbols and abbreviations are identical to those in Figure 2 and 3. Shell injections give rise to a single column of terminals that extends the length of ventral pallidum and beyond (A and D, number 1) and is located in ventromedial ventral pallidum ( W m ) in register with the dense NT-IR. Note that the fascicles are lateral and/or dorsal to the
443
terminal fields after shell injections. Bands of strong NT-IR in B and E are the result of haloperidol administration and are denoted with hal. These strong dorsolateral bands of NT-IR enhance the distinction between ventromedial ( W m ) and dorsolateral (WI)ventral pallidum making it clear that the PHA-L labeled terminals in Figure 4 are present in W m . Scale bar: 1.0 mm.
Figure 5
P
m
TWO VENTRAL TRANSPALLIDAL PATHWAYS accumbal shell is the major striatal recipient of hippocampal input (Kelley and Domesick, '82; Groenewegen et al., '87; Quirk and Groves, '89) and recent evidence suggests that the single unit electrophysiological responses of striatal neurons to subicular stimulations differ in the core and shell (Boeijinga et al., '90). The synaptic organization and relative susceptibility to neurotoxic perturbation of the dopamine-containinginputs to core and shell also appear to differ significantly. Thus it was demonstrated that in accumbal core and in the caudateputamen, the frequency of axospinous and axodendritic synapses with smooth dendritic segments is about equal. In the shell, in contrast, synaptic contacts with the smooth segments of dendrites (between spines or on aspiny dendrites) are significantly more numerous than are contacts with spines (Zahm and Haycock, '89). Previous indications that the nucleus accumbens is less sensitive to the tyrosine hydroxylase immunoreactivity depleting effects of mesencephalic 6-hydroxydopamine (6-OHDA) injections (Voorn et al., '87) were subsequently confirmed (Zahm and Johnson, '89), and it was noted later (Zahm, '90) that the mesostriatal neurons apparently resistant to the toxin selectively innervate the accumbal shell and medial parts of the olfactory tubercle. Mesostriatal fibers and terminals in the accumbal core, like those in the caudate-putamen (with the exception of those innervating the striatal patches), are depleted of tyrosine hydroxylase immunoreactivity within 24 hours of the 6-OHDA injection. Moreover, the capacity of striatal neurons to accumulate detectable neurotensin immunoreactivity following haloperidol administration, an indirect indication of the perturbed status of dopaminergic neurotransmission (Eggerman and Zahm, '881, is apparently greatest in accumbal shell and olfactory tubercle, whereas the increases in numbers of detectable cells in accumbal core resemble those observed in most parts of the caudate putamen (Zahm, '90a). Collectively, these results establish the identities of two adjacent, if not parallel, systems of connections originating in the rat nucleus accumbens and projecting farther downstream via relays in the ventral pallidum (Fig. 5 ) . The accumbal core-VP1-substantia nigraisubthalamic nucleus connections, as described above, resemble other transpallidal subsystems of the basal ganglia, of which several are currently being considered in more or less specific functional terms (see Alexander et al., '86). The pathway originating in the accumbal shell and projecting via VPm, in contrast, fails to reach the substantia nigra and subthalamic nucleus, making it difficult to categorize it similarly.
Fig. 5. Summary diagram depicting five levels of the rat brain c u t in the frontal plane (A-E) and two adjacent transpallidal pathways originating in t h e nucleus accumbens. One pathway originates i n t h e accumbal core (broken line) and passes to t h e dorsolateral part of ventral pallidum (VPI). Neurons in VP1 can be retrogradely labeled after injections in t h e mediodorsal thalamic nucleus, subthalamic nucleus and substantia nigra, b u t not after injections confined to t h e ventral tegmental area. The other pathway originates in t h e accumbal shell (unbroken line) and projects to t h e ventromedial pallidal subterritory (VPm). Neurons in VF'm a r e retrogradely labeled after injections in t h e mediodorsal thalamic nucleus and ventral tegmental area, b u t not after injections involving t h e subthalamic nucleus. Abbreviations: ac-anterior commissure; Acb-nucleus accumbens; f-fornix; frfasciculus retroflexus; lot-lateral olfactory tract; MD-mediodorsal thalamic nucleus; ml-medial lemniscus; sm-stria medullaris; SNsubstantia nigra; STN-subthalamic nucleus; Tu-olfactory tubercle; VTA-ventral tegmental area.
445
Other efferent targets of the accumbal shell that are not treated in this study, e.g. in the hypothalamus and the sublenticular gray, are not innervated by accumbal core (Heimer et al., 'gob), buttressing the suggestion that accumbal shell, but not the core, contains in significant portion elements of the recently described "extended amygdala" system (Alheid and Heimer, '88; Heimer et al., '90a). The results reported here are limited to confirmation of a projection from accumbal core to VP1 and from accumbal shell to VPm. The latter projection is aligned appropriately to exert direct synaptic influence on a projection from VPm to the VTA. The neurochemical and hodological character of this transpallidal pathway is sufficiently different from that of other pathways leaving the striatal complex to indicate that it represents a uniquely specialized part of the basal ganglia.
ACKN0WLEM;;MENTS This work was supported by USPHS grants NIH NS23805 (D.S.Z.) and NIH NS-17743 (L.H.). The authors gratefully acknowledge the expert technical assistance of Evelyn Williams and the clerical support of Victoria Cardillo.
LITERAmcm Alexander, G.E., M.R. DeLong, and P.L. Strick (1986) Parallel organization of functionally segregated circuits linking basal ganglia and cortex. Ann. Rev. of Neurosci. 9:351-381. Alheid, G.F., and L. Heimer (1988) New perspectives in basal forebrain organization of special relevance for neuropsychiatric disorders; the striatopallidal, amygdaloid, and corticopetal components of substantia innominata. Neuroscience 27: 1-39. Boeijinga, P.H., C.M.A. Pennartz, and F.H. Lopes da Silva (1990) Pairedpulse facilitation in the nucleus accumbens following stimulation of subicular inputs in the rat. Neuroscience 35~301-311. 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 Res. 11 7:423-435. Carter, D.A., and H.C. Fibiger (1978) The projections of the entopeduncalar nucleus and globus pallidus in rat as demonstrated by autoradiography and horseradish peroxidase histochemistry. J. Comp. Neurol. 177:113124. Chang, H.T., C.J. Wilson, and S.T. Kitai (1981) Single neostriatal efferent axons in the globs pallidus; a light and electron microscopic study. Science 213:915-918. Churcbill, L., R.P. Dilts, and P.W. Kalivas (199Oa) Changes in gammaaminobutyric acid, mu-opiodand neurotensin receptors in theaccumbenspallidal projection after discrete quinolinic acid lesions in the nucleus accumbens. Brain Res. 51 1:41-54. Churchill, L., R.S. Cross, T.L. Pazdernick, S.R. Nelson, D.S. Zahm, L. Heimer, R.P. Dilts, and P.W. Kalivas (1990b) Patterns of glucose use after bicuculline-induced convulsions provides a functional marker for the ventral pallidum. Neuroscience (in review). Eggerman, K.W., and D.S. Zahm (1988) Numbers of neurotensin-immunoreactive neurons selectively increased in rat ventral striatum following acute haloperidol administration. Neuropeptides 11.125-132. Groenewegen, H.J., and F.T. Russchen (1984) Organization of the efferent projections of the nucleus accumbens to pallidal, hypothalamic and mesencephalic structures: A tracing and immunohistochemical study in the cat. J. Comp. Neurol. 223:347-367. Groenewegen, H.J., E. Vermeulen-Van der Zee, A. te Kortschot, and M.P. Witter (1987) Organization of the projections from the subiculum to the ventral striatum in the rat. A study using anterograde transport of Phaseolus-uulgaris-leucoagglutinin. Neuroscience 23: 103-120. Groenewegen, H.J., G.E. Meredith, H.W. Berendse, P. Voorn, and J.G. Wolters (1989) The compartmental organization of the ventral striatum in the rat. In A.R. Crossman and M.A. Sambrook (ed): Neural Mechanisms in Disorders of Movement. London: John Libbey, pp. 45-54.
446 Groenewegen, H.J., H.W. Berendse, G.E. Meredith, S.N. Haber, P. Voorn, J.G. Wolters, and A.H.M. Lohman (1991) Functional anatomy of the ventral, limbic system-innervated striatum. In P. Willner and J. ScheelKruger (eds): The Mesolimbic Dopamine System: From Motivation to Action. Chichester, England: John Wiley & Sons, (in press). Haber, S.N., Groenewegen, H.J., E.A. Grove, and W.J.H. Nauta (1985) Efferent connections of the ventral pallidum: Evidence of a dual striatopallidofugal pathway. J. Comp. Neurol. 235:322-335. Haber, S.N., E. Lynd, C. Klein, and H.C. Groenewegen (1990) Topographic organization of the ventral striatal efferent projections in the rhesus monkey: An anterograde tracing study. J. Comp. Neurol. 293:282-298. Hammer, R.P., Jr. (1989) Cocaine alters opiate receptor binding in critical brain reward regions. Synapse 3:5540. Hattori, T., H.C. Fibiger, and P.L. McGeer (1975) Demonstration of a pallido-nigral projection innervating dopaminergic neurons. J. Comp. Neurol. 162r487-504. Heimer, L., and R.D. Wilson (1975) The subcortical projections of allocortex. Similarities in the neural associations of the hippocampus, the piriform cortex and the neocortex. In M. Santini (ed): Golgi Centennial Symposium Proceedings. New York Raven Press, pp. 177-193. Heimer, L., J. de Olmos, G.F. Alheid, and L. Zaborszky (1990a) “Perestroika” in the basal forebrain; opening the border between neurology and psychiatry. Progress in Brain Research (in press). Heimer, L., L. Zaborszky, D.S. Zahm, and G.F. Alheid (1987) The ventral striato-pallidothalamic projection: I. The striatopallidal link originating in the superficial layers of the olfactory tubercle. J. Comp. Neurol. 255571-591. Heimer, L., D.S. Zahm, L. Churchill, P.W. Kalivas, and C. Wohltmann (1990b) Specificity in the projection patterns of accumbal core and shell. Neuroscience (in press). Itoh, Z., K. Akiva, S. Nomura, N. Mizuno, Y. Hakamura, and T. Sugimoto (1979) Application of a coupled oxidation reaction to electron microscopic demonstration of horseradish peroxidase: Cobalt-glucose oxidase method. Brain Res. 175341-347. Kalivas, P.W., and J.S. Miller (1984) Neurotensin neurons in the ventral tegmental area project to the medial accumbens. Brain Res. 300;157160. Kelley, A.E., and V.B. Domesick (1982) The distribution of the projection from the hippocampal formation to the nucleus accumbens in the rat: An anterograde and retrograde horseradish peroxidase study. Neuroscience. 7:3221-2335. Meredith, G.E., B. Blank, and H.J. Groenewegen (1989) The distribution and compartmental organization of the cholinergic neurons in nucleus accumbens of the rat. Neuroscience 31:3278-345. Mogenson, G.J., L.W. Swanson, and M. Wu (1983) Neural projections from nucleus accumbens to globus pallidus, substantia innominata, and lateral preoptic area: an anatomical and electrophysiological investigation in the rat. J. Neurosci. 3:189-202. Nauta, W.J.H., G.P. Smith, R.L.M. Faul1,andV.B. Domesick(1978) Efferent connections and nigral efferents of the nucleus accumbens septi in the rat. Neuroscience 3:385-410.
D.S. ZAHM AND L. HEIMER Quirk, T.L., and P.M. Groves (1989) Hippocampal projections to the rat nucleus accumbens: A light and electron microscopic analysis using Neurosci Abstr. 15r1244. PHA-L. SOC. Smith, Y., and A. Parent (1986) Differential connections of caudate nucleus and putamen in the squirrel monkey (Saimiri sciureus). Neuroscience 18r347-371. Voorn, P., C.R. Gerfen, and H.J. Groenewegen (1989) Compartmental organization of the ventral striatum of the rat: Immunohistochemical distribution of enkephalin, substance P, dopamine and calcium-binding protein. J. Comp. Neurol. 289:189-201. Voorn, P., G. Roest, and H.J. Groenewegen (1987) Increase of enkephalin and decrease of substance P immunoreactivity in the dorsal and ventral striatum of the rat after midbrain 6- hydroxydopamine lesions. Brain Res. 412t391-396. Voorn, P., B. Jorritsma-Byham, C. Van Dijk, and R.M. Buijs (1986) The dopaminergic innervation of the ventral striatum in the rat: A light and electron microscopical study with antibodies against dopamine. J. Comp. Neurol. 251r84-99. Wilson, C.J., and K.D. Phelan (1982) Dual topographic representation of neostriatum in the globus pallidus of rats. Brain Res. 243r354-359. Young, W.S., Jr., G.F. Alheid, and L. Heimer (1984) The ventral pallidal projection to the mediodorsal thalamus: A study with fluorescent retrograde tracers and immunohistofluorescence. J. Neurosci. 4: 162% 1638. Zaborszky, L., G.F. Alheid, V.E. Alones, W.W. Oertel, D.E. Schmechel, and L. Heimer (1982) M e r e n t s of the ventral pallidum studied with a combined immunohistochemical-anterogradedegeneration method. SOC. Neurosci. Abstr. 8:218. Zaborszky, L., G.F. Alheid, M.C. Beinfeld, L.E. Eiden, L. Heimer, and M. Palkovits (1985) Cholecystokinin innervation of the ventral striatum: A morphological and radioimmunological study. Neuroscience 14t427453. Zahm, D.S. (1989) Ventral striatopallidal parts of the basal ganglia in the rat: 11. Compartmentation of ventral pallidal efferents. Neurosci. 30:33-50. Zahm, D.S. (1990) Compartmental and subterritorial organization of dorsal and ventral striatum revealed following acute perturbations of the mesostriatal projections in rats. J. Comp. Neurol. (in review). Zahm, D.S., and J.W. Haycock (1989) On-line morphometry of tyrosine hydroxylase-immunoreactive axon terminals in caudate putamen and nucleus accumbens in rat with a focus on synapses. SOC.Neurosci. Abst. 15r905. Zahm, D.S., and L. Heimer (1988) Ventral striatopallidal parts of the basal ganglia in the rat: I. Neurochemical compartmentation as reflected by the distributions of neurotensin and substance P immunoreactivity. J. Comp. Neurol. 272516-535. Zahm, D.S., and Johnson, S.N. (1989)Asymmetrical distribution of neurotensin immunoreactivity following unilateral injection of 6-hydroxydopamine in rat ventral tegmental area (VTA). Brain Res. 483r301-311. Zahm, D.S., L. Zaborszky, G.F. Alheid, and L. Heimer (1987) The ventral striatopallidothalamic projection: 11. The ventral pallidothalamic link. J. Comp. Neurol. 255;592-605.
