THE JOURNAL OF C0MPAR.A"V.E NEUROLOGY 301:554-574

(1990)

SubdivisionsandNeuronTypesof the Nucleus of the Solitary Tract That Project to the Parabrachial Nucleus in the Hamster MARKC.WHITEHEAD Department of Oral Biology, The Ohio State University, College of Dentistry, Columbus, Ohio 43210

ABSTRACT The solitary nuclear complex (NST) consists of a number of subdivisionsthat differ in their cytoarchitectonicfeatures as well as in the amounts of inputs they receive from lingual afferent axons. In this study horseradish peroxidase (HRP) was injected into the parabrachial nucleus (PBN) of the hamster to determine which of these subdivisions contain cells that project to the pons. In the rostral, gustatory division of the NST, the rostral central subdivision contains the greatest number of labelled pontine-projection neurons. The rostral lateral subdivision contains moderate numbers of labelled cells; progressively fewer labelled cells are in the ventral, medial, and dorsal subdivisions. In the caudal, general viscerosensory division of the NST, the caudal central subdivision contains the majority of labelled cells, although fewer than its rostral counterpart. Progressively fewer cells are labelled in the medial, laminar, ventrolateral, and lateral subdivisions; none in the dorsolateral subdivision. Small horseradish peroxidase injections into the pons revealed that cells of the rostral central and rostral lateral subdivisions of the NST project to the medial subdivision of the PBN, predominantly to caudal and ventral parts of the subdivision. Cells of the caudal central and medial subdivisions of the NST project to the central lateral subdivision of the PBN, predominantly to intermediate and rostral-dorsal parts of the subdivision. Outside the NST, cells in the spinal trigeminal nucleus and parvicellular reticular formation were also labelled after PBN injections. Within the rostral central and rostral lateral (gustatory) subdivisions of the NST at least two types of neurons, distinguished on the basis of dendritic and cell body morphology, were labelled after HRP injections that included the medial PBN. Elongate cells have ovoid-fusiform somata and dendrites oriented in the mediolateral plane parallel to primary afferent axons entering from the solitary tract. Stellate cells have triangular or polygonal cell bodies and three to five dendrites oriented in all directions, although one or two often extend mediolaterally. These results indicate that cytoarchitectonic subdivisions of the NST are distinguished by their efferent ascending connections. For each subdivision within the rostral, gustatory NST there is a correlation between the density of lingual inputs it receives and the density of pontine-projection neurons it contains. Within the rostral central subdivision, which contains the densest lingual inputs and the largest collection of PBN-projection neurons, cell types previously identified in studies with the Golgi method were found to send their axons to the PBN. The presence of two types of pontine-projection cells in the rostral central subdivision provides a structural basis €or parallel information processing in the ascending gustatory system. Projections to the PBN from regions outside the NST provide opportunities for convergence,at the level of the pons, between inputs arising from gustatorylgeneral viscerosensory subdivisions of the NST and from trigeminal sensory nuclei and the reticular formation. Key words: parabrachial nucleus projection,cytoarchitecture, cell types

The nucleus of the solitary tract (NST) receives input from all of the gustatory and general visceral cranial nerves as well as a contribution from cranial nerve V (for review, see Norgren, '85). Major projections for neurons of the NST are to medullary nuclei involved in responses (e.g., oromoo 1990 WILEY-LISS, INC.

tor or cardiovascular reflexes), or to higher centers for perception. The higher centers compose, principally, two pathways: 1) through the basal forebrain, and 2) through Accepted August 10,1990.

NUCLEUS OF SOLITARY TRACT IN THE HAMSTER the thalamocortical system. Nearly all information ascending from the NST to the thalamocortical system, in rodents, undergoes an obligatory synapse in the ipsilateral parabrachial nucleus (PBN) of the pons (Norgren and Leonard, '73). The PBN is also a site of synaptic interruption in the pathway leading from the NST to the basal forebrain (Fulwiler and Saper, '84). Therefore, the PBN figures prominently in the ascending gustatory and general viscerosensory systems. It may also play a role in oromotor functions since in rats it is spared by mid-collicular transections of the neuraxis that leave intact taste-elicited oromotor behaviors (Grill and Norgren, '78). A role for the PBN in viscero-motor activities is also possible since many autonomic functions are not obviously compromised by midbrain section. Investigations of the NST-PBN pathways have described general features of the projections from the rostral (predominantly taste) and caudal (predominantly general viscerosensory) halves of the NST (Norgren, '78, in rat; Beckstead et al., '80, in monkey; Travers, '88, in hamster). More details are needed, however, for the NST is a complex structure consisting of a number of subdivisions with different cytoarchitectonic features (Loewy and Burton, '78, in cat; Kalia and Sullivan, '82, in rat; Whitehead, '88, in hamster), and there is little information available on how this structural heterogeneity is related to connections (but see Herbert et al., '90). The present study evaluates this in the hamster by using small, subnuclear, injections of the tracer horseradish peroxidase into the PBN and comparing the distributions of retrogradely labelled cells in the NST to its cytoarchitectonic subdivisions. Thus, this study investigates the projections to the PBN as a means of further characterizing and distinguishing the subdivisions of the NST. This information supplements the available cytological criteria. In addition, the labelled cells can be characterized on the basis of dendritic and cell body morphology to determine if more than one cell type in the NST projects to the PBN. Topographic features of the NST-PBN projec-

555

tions, demonstrated in the retrograde HRP experiments, were verified in anterograde autoradiographic experiments that analyzed the distribution of labelled axons following injections of 3Hamino acids into the NST.

MATERIALSANDMETHODS Horseradishperoxidasehistochemistry Twenty-three golden hamsters (Mesocricetus aurutus) were used in the horseradish peroxidase (HRP) study. Nine of the animals received microiontophoretic injections in the PBN. The injections were made by passing 2 pA current (5 Hz, on-off)through micropipettes (impedance, 2-8 MR; tip diameter, 6-10 bm), containing 8% HRP in 0.5 M KC1, for 1-4 minutes (Orona et al., '83). Fourteen of the animals received pressure injections through micropipettes connected to a Picospritzer (General Valve). The Picospritzer allowed the sizes of the injections to be regulated by varying the pressure, duration, and number of the ejection pulses. Using short (10 msec) durations, 40 psi pressure and small pipette tip diameters (8 pm), 1-3 pulses resulted in injection sites that were smaller (200-500 pm diam.) than ones obtainable with microsyringe injections, and more predictable in size than ones obtained with iontophoretic injections. To inject the PBN, the animals were anesthetized with sodium pentobarbital (45 mg/kg) and an opening was made in the skull to expose the anterior cerebellum. The PBN was located visually based on surface features of the dorsolateral pons after unilateral aspiration of the rostral cerebellum. The placement of injections involved an oblique dorsal approach of the micropipette that passed near or through a landmark eminence formed by the brachium conjunctivum on the dorsal pons. After a survival time of 1-2 days the animals were killed by transcardiac perfusion of fixative. A washout (50 ml. 0.1 M phosphate buffer, pH 7.2-7.4,0.05% lidocaine) preceded the fixative (150 ml. 1% paraformaldehyde, 3% glutaraldehyde, in the same buffer).

Abbreviations A

AP B C cc Ce CN cl d dl DT em el

G IC ICP K L la

~~

LC LR LV M m me MCP

Mv MLF NST

nucleus ambiguus area postrema brachium conjunctivum cuneate nucleus caudal central subdivision, nucleus of the solitary tract cerebellum cochlear nucleus central lateral subdivision,parabrachial nucleus dorsal subdivision, parabrachial nucleus dorsolateral subdivision, nucleus of the solitary tract dorsal tegmental nucleus external medial subdivision,parabrachial nucleus external lateral subdivision, parabrachial nucleus gracile nucleus inferior colliculus inferior cerebellar peduncle Kolliker-Fuse nucleus lateral division, parabrachial nucleus laminar subdivision, nucleus of the solitary tract locus ceruleus lateral reticular nucleus lateral vestibular nucleus medial division, parabrachial nucleus medial subdivision,parabrachial nucleus medial subdivision,nucleus of the solitary tract middle cerebellar peduncle medial vestibular nucleus medial longitudinal fasciculus nucleus of the solitary tract

PC PH R RG rl RT sg

so

SVN T t V ve vl X 5m 5M 5P 5sc 5Si 5% 5% 5T 6 7 7g

7M 12

parvicellular reticular formation nucleus prepositus hypoglossi raphe nucleus gigantocellular reticular nucleus rostral lateral subdivision, nucleus of the solitary tract reticulo tegmental nucleus suprageniculate nucleus superior olive spinal vestibular nucleus solitary tract mesencephalic trigeminal tract ventral lateral subdivision, parabrachial nucleus ventral subdivision, nucleus of the solitary tract ventrolateral subdivision, nucleus of the solitary tract dorsal motor vagal nucleus mesencephalictrigeminal nucleus motor trigeminal nucleus principal sensory trigeminal nucleus spinal trigeminal nucleus, pars caudalis spinal trigeminal nucleus, pars interpolaris spinal trigeminal nucleus, pars oralis supratrigeminal nucleus spinal trigeminal tract trochlear nucleus facial nerve root facial nerve genu motor facial nucleus hypoglossal nucleus

M.C. WHITEHEAD

556

The brains were dissected and placed in 30%sucrose in 0.1 M buffer overnight. Transverse, 30 pm thick, frozen sections of the brainstem, from the medullospinal junction to the mid-colliculus, were cut and alternating sections reacted with either p-phenylenediamine/pyrocatechol(Hanker et d . , '77) or tetramethylbenzidine (Mesulam, '76) as chromagens. The former sections were counterstained with Cresyl Violet; the latter with Neutral Red. Analysis of the injection sites and the resulting labelled cells was performed with brightfield and darkfield optics. Besides noting the locations of labelled neurons, those in the NST were stuhed with attention to their sizes, shapes, and dendritic patterns. A collection of Golgi-impregnated, Nissl-, reduced silver- or myelin-stained material from normal animals was available for comparison with the HRP-labelled material. The nomenclature used for the subdivisions of the NST corresponds to that of Whitehead ('88). The nomenclature used for the subdivisions of the PBN corresponds to that used by Fulwiler and Saper ('84) in the rat.

Autoradiography Ten hamsters were used to trace anterograde transport from sites in and around the NST to the PBN. Injections of a 1:l mixture of tritiated proline and leucine (NEN-Dupont, 40-60 Ci/mmole), concentrated to 8 pCiIp1 or 100 p,Ci/pl, were made in anesthetized animals via a dorsal approach after removing the posterior occipital bone and aspirating the caudal, midline cerebellum. The NST was targeted using a coordinate system that relates surface landmarks of the medulla (e.g., the medullary midline and blood vessels, and the area postrema) to cytoarchitectonic regions with the head held in a fixed plane by a non-traumatic headholder. The injections were made with the Picospritzer. After a survival time of 2-4 days the animals were anesthetized and perfused with buffer as above followed by 10% formol-saline. Transverse sections of the brains were cut by frozen-sectioning at a thickness of 30 Fm, paraffinsectioning at 10 pm, or methacrylate plastic (JB-4)sectioning at 3 pm. The sections were coated with NTB-3 emulsion (Kodak) and processed according to the technique described by Cowan et al. ('72). Analysis of the anterograde connections was conducted by viewing the silver grains with darkfield optics. The injection sites were defined, using both light- and darkfield illumination, as the regions in which the density of silver grains over neuronal somata exceeded that over neuropil.

RESULTS Subdivisionsof the NST and PBN The mammalian NST has been traditionally divided into a rostral, gustatory part, and a larger, caudal, general viscerosensory part. The subdivision scheme proposed for the hamster, based on cytoarchitectonics and the Golgi method, similarly defines a rostral division containing rostrul luteral, rostral central, medial, dorsal and ventral subdivisions; and a caudal division containing caudal lateral, caudal central, dorsolateral, laminar, ventsolateral, and medial subdivisions (see Whitehead, '88, for photomicrographs and descriptions of these subdivisions). The PBN, in hamster as in rat (Fulwiler and Saper '84), consists of a medial division that lies ventromedial to the brachium conjunctivum, and a lateral division that lies dorsolateral to the brachium (Fig. 1A). The medial division

consists of the medial and external medial subdivisions. The medial subdivision, in the hamster, contains a variety of cell types, including large (12 pm, ave. diam.) and medium-sized (9 km, ave. diam.) round, oval, multipolar, and fusiform cells (Halsell and Frank, '89) (Fig. 2F). The medial subdivision is most prominent in the caudal twothirds of the PBN (Fig. 2B,C); it becomes less distinct rostrally (Fig. 2A), and is absent a t the rostral pole of the PBN. The medial division is bordered dorsally and laterally by the brachium conjunctivum, ventrally by the supratrigeminal nucleus, and medially by the locus ceruleus and mesencephalic trigeminal nucleus. The external medial subdivision (Fig. 2C,H) is a collection of medium-sized ovoid and multipolar cells located caudally, bordering the ventrolateral edge of the brachium conjunctivum and the subjacent supratrigeminal and Kolliker-Fuse nuclei. The lateral division of the PBN has been subdivided in the rat (Fulwiler and Saper, '84) and most of these subdivisions were identified in the hamster. The central lateral subdivision consists of sparsely distributed polygonal and ovoid-fusiform, medium-sized cells (Fig. 2D) that occupy a large portion of the lateral division at intermediate and rostral levels of the PBN (Fig. 2A,B); it is the only subdivision present at the rostral pole of the PBN. At intermediate and caudal levels the central lateral subdivision appears compressed against the dorsolateral edge of the brain and its constituent cells appear similarly flattened (Fig. 2B,C,G). The dorsal lateral subdivision is a collection of densely packed dark cells on the dorsolateral surface of the PBN that indents the central lateral subdivision throughout its caudal two-thirds (Fig. BB,C,G). Indenting the brachium conjunctivum predominantly at intermediate-rostra1levels of the PBN is the ventral lateral subdivision. It contains large (12 p,m, ave. diam.) oval cells (Halsell and Frank '89) more densely packed than those that lie in the adjacent central lateral subdivision (Fig. 2B,C,E). The external lateral and extreme lateral subdivisions, defined for the rat (Fulwiler and Saper, '831, are indistinguishable in the hamster. The external lateral subdivision, as identified here, is a small collection of medium-sized multipolar neurons dorsolateral to the ventrolateral edge of the brachium conjunctivum at intermediate rostrocaudal levels of the PBN (Fig. 2B). The Kolliker-Fuse nucleus is a collection of larger, darker triangular cells that lies ventral to the external medial and external lateral subdivisions (Fig, ZA-C,H).

Topgraphic featuresof the NST-PBN projection Since a major purpose of the present study was to map, in detail, the projection from the NST to the PBN, it was necessary to place HRP injections in different parts of the PBN in different animals. These injections were of various sizes and are categorized here as small (n = lo), medium (n = 51, or large (n = 8). Large injections provided general information on PBN afference and are presented in this section before the smaller injections. Large injections also resulted in more extensive filling of retrogradely labelled cells than the smaller injections, a feature that is presented in the last section of Results. The large injections included all of the subdivisions of the PBN and spread only into the immediately adjacent nuclei. Thus, while principally confined to the PBN the injections also involved the mesencephalic trigeminal nucleus, locus ceruleus, medial Kolliker-Fuse nucleus, dorsal border of the

-

NUCLEUS OF SOLITARY TRACT IN THE HAMSTER

557

Fig. 1. Distribution of labelled cells following a large HRP injection covering all subdivisions of the PEN. Sections are 30 pm thick. Many cells of the NST were labelled; they were more numerous in the rostral division (C-E)than in the caudal division (F,G). Labelled cells in the ipsilateral parvicellular reticular formation lie rostral, lateral, and ventrolateral to the rostral NST; fewer were ventrolateral to the caudal NST. A small collection of labelled cells was located contralateral and homotopic to the reticular neurons at all levels except the most rostral level (B), where they were absent. Labelled cells in the trigeminal sensory nucleus were located bilaterally in the dorsal third of pars oralis

and interpolaris (B-F); they were predominantly ipsilateral in lamina I of pars caudalis (G).Small collections of labelled cells were also medial to the motor trigeminal nucleus ipsilaterally (A), dorsal to the motor facial nucleus, ipsilaterally (ED),and dorsal to nucleus ambiguus, bilaterally (E,F). Rarely cells were labelled in the nucleus reticularis gigantocellularis (E-G). Cells located bilaterally in the supratrigeminal nucleus were probably labelled by ventral spread of the injection site. Wuuy lines, axons of labelled cells ascending past the motor trigeminal nucleus to the PEN. Scale = 1mm.

supratrigeminal nucleus, and the dorsolateral edge of the pontine tegmentum. An example of such an injection is shown in Figure 1A. These injections produced prominent collections of retrogradely labelled cells in the NST, in the medullary reticular formation, and in the spinal trigeminal

nucleus (SpV); the axons of the labelled cells coursed rostrally through the dorsolateral medulla (Fig. 1B) and passed close to and through the trigeminal motor nucleus (Fig. lA, wavy lines) en route to the PBN. Labelled cells in the NST were present at all rostrocaudal levels in the

558

M.C. WHITEHEAD

__ Fig. 2. Photomicrographs of Nissl-stained, 30 pm thick coronal sections through the PBN with the cytoarchitectonic subdivisions labelled (see list of abbreviations). Sections shown in A-C are in rostral to caudal order. Rostrully (A), at the level where the inferior colliculus meets the pons, the central lateral subdivision is prominent; its scattered cells are shown at high magnification in D. At an intermediate level (B) the relatively densicellular dorsal lateral and ventral lateral subdivisions appear in the lateral division. At high magnification, the ovoid cells of the ventral lateral subdivision flank more flattened cells that extend medially into the brachium conjunctiwm (El. The medial

-

~

-

_I

subdivision contains a heterogeneous population of cell sizes and shapes (F);it is bordered medially by large, dark mesencephalic trigeminal cells (on left in F). More caudally in the PBN (C), the dorsal lateral subdivision becomes compact (G,d),while the cells of the central lateral subdivision are more flattened than those of the subdivision at rostral levels (compare G , cl with D). Ventrally, the external medial subdivision forms a mediolaterally oriented strip of cells dorsomedial to the larger, darker cells of the Kolliker-Fuse nucleus (H, K ) . Scales = 0.5 mm in A-C; 50 p,m in D-H.

NUCLEUS OF SOLITARY TRACT IN THE HAMSTER ipsilateral nuclear complex but were more numerous in the rostral division (Fig. 1C-E) than in the caudal division (Fig. 1F-H). A few isolated labelled cells were seen in the contralateral, rostral, NST (Fig. lD,E). Labelled cells in the ipsilateral reticular formation were more scattered than those in the NST, with the exception of a dense collection in the dorsolateral reticular formation immediately rostral, lateral, and ventrolateral to the rostral division of the NST (Fig. 1B-El. This region corresponds to the parvicellular reticular formation (PCRF) (Paxinos and Watson, '86). The labelled cells collectivelyhave the appearance of a radially oriented wedge expanding more-or-less uninterrupted from the NST through the reticular formation and into the SpV nucleus. Ventral and ventromedial to this region more scattered labelled cells were lacated in the dorsal and ventral nucleus reticularis medullaris, and in the nucleus paragigantocellularis.Occasionalmedium-sized cells of the nucleus reticularis gigantocellularis were labelled. Labelled reticular neurons were more numerous at the level of the rostral division of the NST (Fig. 1C-E) than at the level of its caudal division (Fig. 1F-H); they extended rostral to the NST to form a cluster in the PCRF and dorsal SpV nucleus lateral to the facial genu (Fig. 1B). Labelled cells trailed off rostral to this point to form a modest collection ventromedial to the motor trigeminal nucleus. On the contralateral side of the reticular formation labelled cells were also present. These cells had nearly mirror-image distributions to those located ipsilateral to the injected PBN, but were fewer in number, especially at the level of the rostral pole of the NST. Labelled cells in the SpV nucleus were most numerous ipsilaterally in the dorsal and medial pars oralis, at the same rostrocaudal level that contained the largest collections of labelled cells in the NST and in the ipsilateral reticular formation (Fig. lC,D). Fewer labelled cells were present in the SpV nucleus pars intermedialis (Fig. 1E,F), and pars caudalis where they clustered in the dorsal part of the marginal subdivision (Figs. lG, 9). At all levels, labelled cells were conspicuously absent from the ventral and ventrolateral parts of the SpV nucleus; only one or two were present in the main sensory nucleus. Occasional contralateral labelled cells were found in sites homotopic to the ipsilateral labelled cells at all levels of the spinal trigeminal nucleus. Small injections of the PBN, in most cases, can be categorized as ones involving predominantly the medial PBN or ones involving the lateral PBN. Injections confined to the medial division of the PBN are represented by the case summarized in Figure 3. This injection resulted in labelled cells in the NST that were largely restricted to the rostral division (Fig. 3G-HI. Only a few cells were labelled in the caudal division. Outside the NST moderate collections of labelled cells were also present in the PCRF rostral and ventrolateral to the rostral NST and in the dorsal SpV nucleus at this level (Fig. 3F-I), Additional, more caudally located cells in the SpV nucleus and reticular formation were seen in one injection of the medial PBN that was also small but more caudolaterally placed than the previous one, bordering the supratrigeminal nucleus and including the external medial subdivision of the PBN (Fig. 4). Injections confined to the lateral division of the PBN are exemplified by the case in Figure 5 . The injection site involved the central lateral subdivision most heavily and spread beyond this subdivision only slightly into the dorsal lateral subdivision. Retrogradely labelled cells in the NST

559

were largely confined to the caudal division, where they clustered predominantly medial to the solitary tract and spread into the lateral area postrema (Fig. 5H-J). Few cells were labelled in the rostral division; only rare cells were labelled in the contralateral caudal NST. Labelled cells in the reticular formation and SpV nucleus were also rare after small lateral PBN injections. A larger injection of the lateral division of the PBN (Fig. 61, with extensive involvement of the central lateral subdivision resulted in more numerous labelled cells in the caudal division of the NST and in the area postrema, SpV nucleus and the reticular formation. No cells were labelled in the rostral division of the NST in this case. Same small injections labelled very few cells (sometimes only 2-3 cells per section), all in the NST, and served to define regions of the PBN to which only the NST projects, apparently lightly. Two injections into the dorsal edge of the PBN, one in the medial part of the rostral pole of the medial subdivision (Fig. 7A), the other in the dorsomedial border of the lateral division and medial subdivision (Fig. 7 0 , labelled only a few cells in the rostral NST (Fig. 7B,D). No cells were seen in the reticular formation or SpV nucleus. A small injection in the vsntrolateral edge of the lateral division of the PBN that spread into the ventrolateral subdivision labelled only a few cells in the caudal division of the NST (Fig. 7E,F); none outside the NST. Medium-sized injections that involved both the medial and lateral PBN labelled cells in both the rostral and caudal division of the NST as well as collections of cells in the reticular formation and SpV nucleus (Fig. 8). There were differences between cases in the number of cells labelled in the reticular formation and SpV nucleus depending upon the rostral-caudal location of the injection sites. Injections in the rostral PEN (Fig. 8A) labelled only modest collections of cells in the reticular formation and SpV nucleus (Fig. 8B,C). More caudally located injections (Fig. 8D) labelled many more cells outside the NST in the reticular formation and SpV nucleus (Fig. 8E,F). The caudalmost injections (Fig. 8G), compared with the rostral injections, of all the medium-sized injections, labelled the most cells in the reticular formation and SpV nucleus, not only at the level of the rostral NST (Fig. 8H), but also at the level of the caudal NST (Figs. 81,9). As with the large injections, in every case of a medium-sized injection, labelled cells outside the NST were more numerous at the level of the rostral NST than at the level of the caudal NST. All of the labelled cells in these cases are likely to project to the PBN and not the surrounding pontine nuclei because the small and medium-sized injections were well confined to the PBN. The casesjust presented indicate that the rostral division of the NST projects to the medial PBN, while the caudal division of the NST projects to the lateral PBN. The HRP data suggest further that some subdivisions of the PBN receive heavier projections from the NST than others. The rostral NST apparently projects most heavily to ventral and ventrolateral parts of the medial subdivision at intermediate rostrocaudal levels. The caudal NST projects heavily to the central lateral subdivision. This was confirmed by the autoradiography experiments. Two different cases will serve as examples. An injection of tritiated amino acids centered in the rostral division of the NST (Fig. IOA) resulted in labelling of axon terminals in the medial division of the PBN (Fig. IOB). Terminals were heaviest ventrally in the intermediate to caudal half of the medial division although a light distribution spread dorsally into the brachium

M.C. WHITEHEAD

560

.

F

Fig. 3. A smaEl injection of HRP, centered in the medial division of the rostral PBN, labelled cells that were primarily confined to the rostral pole of the NST, the reticular formation immediately rostral, ventral, and lateral to the rostral NST (F-H); and the trigeminal sensory nucleus, pars oralis (F).Within the rostral NST, most labelled cells were located in the rostral central subdivision, the remaining cells

were located in the rostral lateral subdivision (D,E) (for comparable figures showing cytoarchitecture, see Figs. 1,2, Whitehead, '88). The medial half of the rostral NST was devoid of labelled cells. High magnification drawings in D and E are taken from levels similar to those depicted at lower magnification in G and H, respectively. Scale = 1mm in A-C and F-K; 130 pm in D,E.

conjunctivum and the medial part of the central and ventral lateral subdivisions and the dorsolateral subdivision (Fig. 10B). Injections of the isotope into the caudal division of the NST (Fig. 1OC) labelled terminals in the lateral PBN. The distribution of silver grains from caudal NST injections was

largely non-overlapping with that from rostral injections. Heaviest label was in the central lateral subdivision at. intermediate and rostral levels of the PBN. Light label extended ventrolaterally and into the dorsolateral subdivision. Labelled axom passed from ventromedial to dorsolat-

Fig. 4. A: Center of a small HRP injection site located in the caudolateral part of the medial subdivision (compare Fig. 2A); Hanker-Yates reaction. B At high magnification, injection site appears confined to the medial subdivision, ventrolaterally. Dashed line, border with the supratrigeminal nucleus and pontine tegmentum. C: Retrogradely labelled cells are most numerous in the rostral NST, including a few ventrally (arrow),and in the parvicellular reticular

formation and spinal trigeminal nucleus, at the level of the rostral NST, particularly. D: Moderate collections of labelled cells are also present at the level of the caudal NST in the parvicellular reticular formation, bilaterally, and in the spinal trigemind nucleus, ipsilaterally Scale = 1.3 mm in A 200 pm in B; 1mm in C,D. (compareFig. W).

C

562

G

H

Fig. 5 . A small injection of HRP, confined to the lateral division of the PBN, is centered in the central lateral subdivision, rostrally (A), and spreads into the dorsal subdivision at intermediate rostrocaudal levels (B).Retrogradely labelled cells were largely confined to the caudal division of the NST (H-J), primarily in the caudal central and

medial subdivisions and area postrema, with fewer cells in the ventrolateral subdivision (D,E). High magnification drawings in D and E are taken from levels similar to those depicted in H and I, respectively. Scale = 1mm in A-C, F-J; 130 Fm in D,E.

NUCLEUS OF SOLITARY TRACT IN THE HAMSTER

563

Fig. 6. A medium-sized HRP injection, centered slightly caudal to that of Figure 4,labels the lateral division of the PBN extensively while entirely sparing the medial division. The densest part of the injections site, in the central lateral subdivision, corresponds to the region of intense autoradiographic labelling after tritiated amino acid injections of the caudal NST (compareFig. 10D).Retrogradely labelled cells in the

NST were more numerous than after of the smaller injection of Figure 4 but were similarly confined to the caudal division (D-F); there were none in the rostral division (C). A few labelled cells were present bilaterally in the medullary RF and the marginal SpV nucleus. Scale = 1mm in A,B; 1.2 mm in C-F.

eral through the brachium conjunctivum. Label was lightbackground in the medial division of the PBN. Injections of the rostral and caudal NST also resulted in anterograde labelling of intramedullary projectionsbut these will not be considered here.

fill the rostral central subdivision with the majority of cells labelled (Figs. 11, 12A). The caudal central subdivision also exhibits many labelled cells after large injections in the PBN (Fig. 12B), but they were not as numerous nor as heavily labelled as those in the rostral central subdivision. Large HRP injections that involve the entire PBN label approximately60%of cells in the rostral central subdivision and 30% cells in the caudal central subdivision. In the rostral division of the NST, many heavily labelled cells were also seen in the rostral lateral subdivision (Fig. 12A), although they were less numerous than those seen in the medially adjacent rostral central subdivision. A few cells were seen in the ventral subdivision at this level with fewer still in the medial subdivision. Only rarely were cells labelled in the dorsal subdivision. In the caudal division of

Subnuclear locationsand morphologies of IabeIledcellsintheNST The retrogradely labelled cells in the NST were not homogeneously distributed throughout the nucleus but were concentrated in some subdivisions and sparse or absent in others. The majority of labelled cells, even after small injections of the PBN, were located in the rostral central subdivision. With large injections, that included virtually the entire PBN, the cells containing HRP virtually

M.C. WHITEHEAD

564 I

1

B

C

D

Fig. 7. Small injections of the PBN that retrogradely label only very few c e l l s a l l in the NST. Sections depicted in B,D, and F were the ones showing the greatest number of retrogradely labelled cells for each of three cases whose injection sites are depicted in A,C,and E,respectively. At the rostral pole of the PBN, injection of the medial extreme of the medial division (A) labelled cells in the rostral central subdivision of the NST at the level of the cochlear nucleus (B) (see Fig. 1,Whitehead, '88, for comparable cytoarchitecture). At an intermediate rostrocaudal

level of the PBN, injection of the medial extreme of the PBN (C) labelled cells in the rostral central subdivision at a level just caudal to the cochlear nucleus (D)(compare Fig. 2, Whitehead, '88). Caudally in the PBN, an injection of the ventral lateral and central lateral subdivisions (E) labelled cells in the caudal central subdivision of the NST (F) (compare Fig. 5, Whitehead, '88). Scale in E = 1mm in A,C,E; scale in F = 250 km in B,D,F.

the NST, the medial, laminar, and ventrolateral subdivisions contained modest scatterings of labelled cells, fewer than were present in the caudal central subdivision (Fig. 12B). Only rarely were cells labelled in the ventral and caudal lateral subdivisions. The dorsolateral subdivision was always devoid of labelled cells. Labelled cells in the rostral NST were sufficiently filled with reaction product after large PBN injections that their

dendritic morphology could be studied. Heavily labelled cells in the rostral central and rostral lateral subdivisions of the NST contained HRP that extended into primary and, in some cases secondary, dendrites. The labelled cells were of two types: elongate and stellate. Elongate cells had cell bodies ranging in shape from round to fusiform; most were ovoid-fusiform.The cells were oriented in the mediolateral plane. From opposite ends of the somata the elongate cells

D

Fig. 8. Medium-sized HRP injections, involving both medial and lateral PBN, centered at different rostrocaudal levels. The most rostral injection (A) labelled cells in the rostral (B)and caudal NST and area postrema (0.Moderate numbers of labelled cells outside the NST, in the parvicellular reticular formation and spinal trigeminal nucleus are present at the level of the rostral NST (A) but are rare caudally (0.Intermediate (D) and caudal (GI injections labelled

A

progressively more cells in the rostral and caudal NST, and in the reticular formation and spinal trigeminal nucleus (E,F,H), including many bilaterally, with a wide distribution in the reticular formation (compare I to C,F) after the caudalmost injection. Scale = 130 Frn in A,D,G; 1 mm in B,C,E,F,H,I.

Ce

566

M.C. WHITEHEAD

Fig. 9. Cells of lamina I and the promontorium in the spinal trigeminal nucleus, pars caudalis, labelled following a medium-sized HRP injection of the caudal PBN. Photomicrograph, from a case plotted in Figure 8 G I , is of a section adjacent to that plotted in Figure 8 I. Scale = 50 um.

extended two primary dendrites, one directed medially, the other laterally (Figs. 12C, 13). The primary dendrites formed oblique branches fairly close to the cell bodies. When labelled with HRP these y-shaped branches were seen both on the medial and lateral first-order dendrites (Fig. 13A). Many primary dendrites, however, were unbranched throughout the extent of their labelling. Occasionally a third, small primary dendrite emanated from the cell body, but these cells were classified as elongate cells based on the pronounced mediolateral orientation of their cell bodies and larger dendrites. Elongate cells were fairly common in the rostral central subdivision, but they were more evident in the rostral latersl subdivision. Within the rostral lateral subdivision their cell bodies were located in or bordering the solitary tract. They sent lateral dendrites toward and sometimes beyond the lateral edge of the NST. The medial dendrites were directed toward and sometimes into the rostral central subdivision. Dendrites of the elongate cells were oriented orthogonal to the axons descending in the solitary tract. The dendrites often coursed between the axons of the tract to reach lateral and central portions of the NST. Stellate cells had 3-5 primary dendrites that radiated in all directions from a multipolar cell body (Figs. 12D, 14). While some stellate cell dendrites were oriented in the mediolateral plane (Fig. 14B), many examples of oblique and dorsally oriented HRP-filled dendrites were also seen (Fig. 14D).In the rostral NST stellate cells were seen in the

rostral central and rostral lateral subdivisions. They were also present caudally in the NST. Only second-order dendritic branches that were close to the cell bodies were HRP-labelled and they were infrequently seen (Fig. 1 4 0 . With subtotal injections of the PBN there were substantially reduced populations of labelled cells in all of the subdivisions and they were less heavily labelled than after large injections. In every case, however, the central subdivisions contained more labelled cells than the subdivisions that surround them at their respective rostrocaudal levels of the NST (Figs. 3D,E, 5D,E, 7B,D,F). Small injections confined to the medial division of the PBN labelled cells that were largely restricted to the rostral division of the NST (Fig. 3). Within the rostral NST the retrogradely labelled cells were most numerous in the rostral central subdivision although the rostral lateral subdivision had substantial collections of labelled cells (Fig. 3D,E). Only occasionally were labelled cells present in the ventral subdivision (Fig. 4C, arrow) and only rarely was a cell seen in the medial or dorsal subdivisions after small injections of the medial PBN. Small injections of the lateral PBN resulted in labelled cells that were restricted to the caudal division of the NST. These cells were most prevalent in the caudal central subdivision (Figs. 5D,E, 7F). The medial subdivision at caudal levels of the NST contained a moderate collection of labelled cells. Cells in the laminar subdivision were only occasionally labelled. Labelled cells in the ventrolateral

Fig. 10. Photomicrographs of autoradiographic injection sites and anterograde labelling. A: Injection site heavily labels the rostra1 NST (outlined by arrowheads) with lighter spread into the descending vestibular nucleus above. B: Anterograde label is heaviest in the medial subdivision of the caudal PBN but spreads into the brachium conjunctivum and ventral part of the central lateral subdivision. Line, dorsal border of the lateral division. C: Injection site heavily

labels the entire caudal NST with lighter ventral spread into the hypoglossal nucleus. D Anterograde label is heaviest in the central lateral subdivision with very light spread into the ventrolateral and dorsolateral subdivisions. Arrowhead, preterminal axons passing through the brachium conjunctivum; line, ventral border of the medial subdivision. Scales = 1 mm in A,C; 200 Fm in B,D.

M.C. WHITEHEAD

568

Fig. 11. Photomicrograph of HRP-labelled cells in the rostral NST after a large injection of the ipsilateral PBN that involved all of its subdivision (comparable to Fig. 1A). In this coronal section labelled cells heavily populate the rostral central subdivision, with moderate numbers of labelled cells, including some with elongate morphologies

(arrows),in the rostral lateral subdivision (see Fig. 2, Whitehead, '88, for comparablecytoarchitecture).Labelled cells also extended ventrolaterally into the parvicellular reticular formation (arrowhead).The NST outline (solid lines) is indicated. Hanker-Yates reaction, Nissl counterstain. Scale = 100 pm.

subdivision were few and never included the very large neurons that characterize this subnuclear region. The caudal lateral subdivision contained only rare labelled cells, the dorsolateral subdivision contained none. The area postrema contained a substantial collection of labelled cells ipsilaterally after injections of the lateral PBN; only rare cells contralaterally (Figs. 51,6E).

chitectonic subdivisions and morphological cell types that compose the pontine projection system in the hamster. Additional observations of labelled cells in the reticular formation and spinal trigeminal nucleus extend our understanding of the organization of non-NST projections t o the PBN.

DISCUSSION

NSTprojectionstothePBN-Correlationof efferent connectionswith cytoarchiteclnnic

The present experiments using HRP and autoradiographic tracing techniques generally confirm previous reports (e.g., Norgren, '78; Loewy and Burton, '78; Ricardo and Koh, '78; Beckstead et al., '80; King, '80; Hermann and Rogers, '85; Travers, '88; Herbert et al., '90) that the rostral and caudal NST have differential projections to the medial and lateral PBN, respectively. Beyond this, the present study has evaluated the results of small injections of HRP placed in discrete regions of the PBN. These injections have made it possible to relate details of the pontine projections to the structural heterogeneity of the NST. Specifically, it has been possible to define the cytoar-

The NST consists of a rostral, predominantly taste, division and a caudal, predominantly general viscerosensory, division. The NST at all levels is, however, structurally heterogeneous, and these divisions have been subdivided based on cytoarchitectoniccriteria (Loewy and Burton, '78, in cat; Kalia and Sullivan, '82, in rat; Whitehead, '88, in hamster). The cytoarchitectonic subdivisionsdiffer not only cytologicallybut also in terms of their primary afferent and efferent connections, including those to the PBN. In the rostral NST, most labelled PBN-projection neurons were located in the rostral central subdivision. This

subdivisions

NUCLEUS OF SOLITARY TRACT IN THE HAMSTER

569

A

A?.. AP

.

*. *:*

-4 . *

I

D

Fig. 12. Subnuclear locations of labelled cells after a large injection in the PBN similar to that depicted in Figure 1A. Labelled cells in the rostral NST (A) are concentrated in the rostral central subdivision; those in the caudal NST (B) are concentrated in the caudal central subdivision (see Figs. 1and 5, Whitehead, '88, for cytoarchitecture of

NST at levels comparable to A and B, respectively).Drawings of labelled cells in the rostral central subdivision illustrate elongate cells (C) and stellate cells (D). All cells are oriented with the dorsal aspect of the medulla at the top of the figure and medial to the right. Scale in A = 100 pm for A,B; scale in C = 50 pm for C,D.

subdivision is a relatively densicellular collection of smallmedium-sized cells in the dorsal half of the NST medial to the solitary tract at rostral levels. The finding that many rostral central neurons project to the PBN is consistent with a recent report of similarly located retrogradely labelled cells after HRP injections of PBN regions responding to taste stimulation of the anterior tongue in the hamster (Travers, '88). A similar central location in the rostral NST of the rat has been described for gustatory responsive PBN-projection neurons (Ogawa et al., '84). The rostral central subdivision in hamster is also the site of densest input from lingual afferent axons (Whitehead, '88). Chorda tympani, lingual (trigeminal), and glossopharyngeal (lingual branch) afferent axons all synapse heavily in the rostral central subdivision, less heavily in the rostral lateral subdivision, and only very lightly in the medial (glossopharyngeal only),ventral or dorsal subdivisions. Similarly,only moderate numbers of PBN-projection neurons were seen in

the rostral lateral and ventral subdivisions, and few were seen in the dorsal and medial subdivisions. Thus, for each subdivision of the rostral division of the NST, there is a correlation between the density of lingual inputs it receives and the density of PBN-projection neurons it contains. Little is known about the connections of subdivisions that do not project heavily to the pons, although a region corresponding to the ventral subdivision has recently been shown in hamster to project to sites in the medullary reticular formation adjacent to oromotor nuclei (Travers, '88). The number of cells in the rostral central subdivision that were labelled in relation to those that were not varied enormously depending on the size of the HRP injections. With large PBN injections (e.g., that depicted in Fig. 1A) that covered the medial subdivision, the Kolliker-Fuse nucleus, and spread into the pontine tegmentum, as many as 60% of cells in the rostral central subdivision were

570

M.C. WHITEHEAD

Fig. 13. HRP-labelled elongate cells seen in a coronal section of the rostral NST after injections of the medial PBN. A: Elongate cell from the rostral lateral subdivision contains diffuse reaction product that fills two mediolaterally oriented primary dendrites and their y-shaped

second-order branches. A stellate cell containing granular reaction product is indicated (arrow). B,C: Elongate cells from the rostral central subdivision. Scale = 40 prn.

labelled. Medium-sized injections (e.g., those in Fig. 8 ) that were subtotal for the medial division of the PBN labelled approximately 10-30% of cells in this subdivision while small injections (e.g., those in Figs. 3 and 4)label far fewer cells. An electrophysiological study reported that 34% of neurons in the central part of the rostral NST of the rat that responded to gustatory stimulation could be activated antidromically from the pons (Ogawa et al., '84). A previous study in the hamster reported 26% of cells in the lingual

afferent zone of the NST labelled after large pontine projections (Davis and Jang, '85). Thus, while the rostral central subdivision projects heavily to the medial subdivision of the PBN, the rostral central subdivision has many cells, including some that may function in gustation, that project elsewhere. Some cells of the rostral NST project medially within the lateral division of the PBN (see also Norgren and Leonard, '73; Norgren, '781, but this is a minor contingent based on injections restricted to the

NUCLEUS OF SOLITARY TRACT IN THE HAMSTER

571

Fig. 14. HRP-labelled stellate cells seen in coronal sections of the rostral NST. Varieties of stellate cells from the rostral central subdivision include one with five primary dendrites radiating in all directions (A);one with three dendrites, one medial, one lateral and one ventral

(B);one with two obliquely oriented dendrites one of which branches close to the cell body (arrow) (C);and one with three primary dendrites, two of which orient dorsoventrally (D). Scale = 40 bm.

lateral division of the PBN. A few cells in the rostral central together with a few in the ventral subdivision also project to the contralateral PBN, a finding that confirms similar bilateral connections of the NST in the rat (Norgren, '78) and cat (King, '80). Projections from the rostral central subdivision to the pontine tegmentum ventral to the PBN are possible but were not significant in the autoradiographic experiments and were not examined further in the HRP experiments. A number of neurons in the rostral central subdivision undoubtedly project to sites within the NST. These locally projecting cells probably include Golgi type I1 cells Whitehead, '88) and GABAergic cells (Lasiter and Kachele, '88).

Some details of the topography of the projection from the rostral central subdivision of the NST to the medial PBN were apparent in the present study. Subtotal injections of the medial PBN, caudally, labelled many more cells in the NST (Most in the rostral central subdivision) than similar sized injections of the medial PBN, rostrally (compare Fig. 8H with Fig. 3E,H). This supports previous electrophysiological (Norgren and Pfaffmann '75; Travers and Smith, '84) and anatomical (Norgren, '78) reports of heaviest representation of the gustatory NST caudally in the medial PBN. There is also evidence from the present study that the rostral central subdivision at the level of the cochlear nucleus projects predominantly to ventrolateral and central

572

parts of the medial PBN while the rostral central subdivision caudal to the cochlear nucleus projects to the dorsomedial part of the medial PBN. The rostral central subdivision, both anterior and posterior to the caudal edge of the cochlear nucleus receives taste input from the anterior tongue, but only the caudal part of the rostral central subdivision receives dense input from the posterior tongue (Stanford and Whitehead, '84). The posterior and anterior tongue, therefore, may be best represented dorsomedially and central-ventrolaterally within the medial PBN. This fits with the available electrophysiological data comparing locations of responses to chorda tympani and glossopharyngeal activation in the rat (Norgren and Pfaffmann, '75). In the caudal NST there were, in general, fewer labelled cells than in the rostral NST after large PBN injections. While it is difficult to determine for each injection the precise boundaries of the HRP uptake site in the PBN, both the dark centers and the paler edges are probably involved. This is because labelling of certain cell groups, e.g., supratrigeminal cells contralateral to the injection site in Figure 1,and caudal NST cells in Figure 81, likely result from the injection edges including the ipsilateral supratrigeminal nucleus (Fig. 1A) and the lateral division of the PBN (Fig. 8G), respectively. Of course, the pale edges of the injection sites may retrogradely label fewer cells than the denser central zones. Thus, injections of the PBN that are large, if centered medially as in Figure 1,may label only a subset of the entire population of pontine-projection cells in the caudal NST. Within the caudal NST most labelled cells were seen in the caudal central subdivision after HRP injections of the lateral PEN. Within the caudal central subdivision labelled cells were concentrated dorsally and medially and, in the cases of medium and large injections, spread medially into the medial subdivision. The caudal central subdivision probably receives general viscerosensory information via the vagus nerve which terminates medial to the caudal solitary tract in rat (Kalia and Sullivan, '82; Higgins et al., '841, cat (Gwyn et al., '79; Kalia and Mesulam, '80a,b), and monkey (Gwyn et al., '85). In addition, this region probably receives input from epiglottal taste buds via the vagus nerve (Sweazy and Bradley, '86). The medial subdivision also appears to receive projections from general viscerosensory primary afferent axons in the hamster (Miceliand Malsbury, '85), rat (Higgins et al., '84; Shapiro and Miselis, ,851, and cat (Cirielloet al., '81). It is not known to what extent vagal terminals and neurons projecting to the lateral PBN overlap in the hamster. It appears from the present study that the caudal central subdivision projects less heavily to the PBN than its rostral counterpart. Cells engaged in descending or intramedullary connections may, however, be more pronounced in the caudal central subdivision than in the rostral central subdivision. Ross et al. ('85) describe a central subdivision of the caudal NST in the rat that projects heavily to the ambiguus, retrofacial complex. We have identified similar intramedullary connections for the caudal central subhvision in the hamster (Davidovitchand Whitehead, '88). Evidence from the monkey suggests that ambiguus projection neurons overlap but predominate lateral to PBN projection neurons (Beckstead et al., '80). Other regions in the caudal NST that project to the PBN are the ventrolateral, caudal lateral, and laminar subdivisions, all of which contain modest collections of labelled cells after injections of the lateral PBN. Of these regions only the ventrolateral subdivision is an area of significant primary afferent input, receiving light projections from

M.C. WHITEHEAD lingual as well as vagal axons (Hamilton and Norgren, '84, in rat; Whitehead, '88, in hamster). The largest neurons in this subdivision are well characterized anatomically and electrophysiologically(Berger et al., '84) and have descending connections with respiratory centers; they were not labelled in the present study. The smaller cells in the ventrolateral subdivision that do project to the pons could provide information to the lateral PBN relevant to pneumotaxic functions (Lumsden, '23; King, '80). In all but the smallest, most ventrolateral injections of the lateral PBN, there were labelled cells in the lateral area postrema. Previous studies are contradictory on whether the area postrema projects to the PBN. Loewy and Burton ('78), in the cat, found HRP-positivecells in the area postrema after PBN injections, Kalia ('77) and King ('80) did not. Consistent with the present results, PBN-projection cells were demonstrated in the area postrema of the rat (Cedarbaum and Aghajanian, '78; Herbert et al., '90). No labelled cells were observed after lateral PBN injections in the dorsolateral subdivision of the NST, a region that is devoid of primary afferent input and, in the cat, reportedly projects to the inferior olive (Loewy and Burton, '78). No labelled cells were seen in the vagal motor nucleus of the hamster although they were seen in the cat after PBN injections (King, '80). These projection cells are smaller than the motor neurons and may not be present in the vagal motor nucleus of the hamster which contains very few small cells ( < 125 pm2,diam.). Most cells are distinctly large, averaging 210 pm2in dameter (Whitehead, '88).

Dif€erentcelltypespmjectfromtherostral NST to the PBN The present results show that two very different morphological classes of neurons of the rostral central and rostral lateral subdivisions of the NST have ascending projections to the PEN. Different types of neurons were identified in HRP-labelled material on the basis of cell body shape and orientation and on the basis of primary dendrite numbers, and orientation. Based on the results of subtotal PHN injections the cell types both project primarily to the medial division of the PBN. The most conspicuous pontine projection cell type is the elongate cell (Whitehead and Savoy, '87; Travers, '88). Elongate cells have been identified in Golgi material (Whitehead, '85, '88; Davis and Jang '88) and are distinguished as small-medium-sized (100-150 pm, soma diam.) round to fusiform cell bodies that have primary dendrites oriented in the mediolateral plane parallel to lingual afferent axons entering the NST from the solitary tract (Whitehead, '88). Primary afferent axons, including taste axons originating in the geniculate ganglion, appear to travel along and contact elongate cell dendrites and their sparse spines. Indeed, synaptic endings resembling those of geniculate afferent s o n s (Whitehead, '86) do contact the dendrites of elongate cells that were labelled following HRP injections into the PBN (Whitehead, '89). Thus, elongate neurons probably constitute a class of second-order gustatory neurons that project to the PBN. Evidence that this cell type functions in taste sensitivity derives from the observation that elongate cells were frequently HRP-filled when small iontophoretic injections of HRP were used to mark sites of anterior tongue taste-elicited single unit activity (Whitehead et a1 , '88). Stellate cells are a second type of PBN-projectionneuron. They too are frequently labelled when HRP is used to mark taste-responsive sites in the rostral central and rostral

NUCLEUS OF SOLITARY TRACT IN THE HAMSTER

573

lateral subdivisionsof the NST. Because the morphology of the stellate cell is not as distinctive as that of the elongate cell they have been harder to characterize in terms of synaptology at the electron microscopic level so it is not known whether they receive input directly from taste afferent-like synaptic endings. Stellate cells are, however, readily distinguished from elongate cells with light microscopy, and they constitute a second class of NST output cells with ascending axons. Two types of NST-PBN projection neurons have also been identified on the basis of cell nuclear morphology (Davis and Jang, '86). It remains to be determined whether there is a correlation between nuclear and dendritic morphology.

the dorsal third of the sensory trigeminal nucleus is a source of PBN-projection neurons for this is the major site outside the NST to which the glossopharyngeal nerve projects (Stanford and Whitehead, '84; Brining et al., '89, in hamster; Hanamori and Smith, '89; in rabbit). In summary, the connections from the reticular formation and spinal trigeminal nucleus establish a relationship between the traditionally defined gustatorylgeneral viscerosensory nuclei and other brain stem structures that project to the parabrachial nucleus.

Projectionsfrom the reticular formationand spinal trigeminal nucleus to the PBN

This research was supported by NIH grant DC00452 and The Ohio State University College of Dentistry.

The medullary reticular formation contained retrogradely labelled cells after most injections of the PBN. The majority of these cells were located in the PCRF, particularly in a wedge-shaped region lateral and ventrolateral to the rostral division of the NST and medial to the dorsal third of the SpV nucleus. This location of reticulo-pontine cells fits with descriptions in the cat (King, '80) and rat (Herbert et al., '90). Interestingly, the PCRF receives projections from the rostral (Norgren, '78; Loewy and Burton, "78) and caudal NST (Ross et al., '85) and thus may serve as a multisynaptic pathway ascending to the PBN analogousto the spino- or trigeminal-reticulothalamic pathway. Neurons of the PCRF, including those located medial to the motor trigeminal nucleus, also project to motor nuclei involved in orofacial movements (Travers and Norgren, '83). Therefore, there may be overlap in the locations of interneurons of an indirect ascending pathway and of a pathway to the oromotor nuclei. The possibility that individual neurons could form both ascending and intramedullary projections via axonal collaterals would not be without precedent in the reticular formation (Scheibeland Scheibel, '58). Whatever roles may be ascribed to neurons in the reticular formation that project to the PBN, the connectional data suggest that these cells function in the gustatory system for two reasons: First, most PCRF neurons are labelled after HRP injections into the medial PBN, the region that also receives projections from the rostral, gustatory, NST (see also, Herbert et al., '90). Second, the caudal part of the medial PBN, a region identified electrophysiologically as gustatory (Norgren and Pfaffmann, ' 7 9 , receives projections from more cells of the PCRF than the rostral PBN. The present data suggest that two ascending pathways, one direct from the gustatory NST, one after synaptic interruption in the PCRF, converge in the medial PBN. Clearly, the lateral, general viscerosensory, PBN does not have a similar arrangement of inputs; it lacks the PCRF input. The spinal trigeminal nucleus also projects to the PBN. This projection involves neurons in the dorsal third of the spinal trigeminal nucleus at all levels, includingthe promontorium and layer I of pars caudalis, that send s o n s predominantly to the caudal PBN. Many trigeminal neurons were seen to project to the medial PBN and to the lateral division as well. The greatest numbers of trigemind cells were labelled after total PBN injection. A similar finding was reported for the rat by Cechetto et al. ('85), who suggested that both gustatory (medial)and general viscerosensory (lateral) parts of the PBN may receive general somatosensory (e.g., pain) inputs that converge with taste and interoceptive inputs. It is particularly interesting that

Beckstead, R.M., J.R. Morse, and R. Norgren (1980) The nucleus of the solitary tract in the monkey: Projections to the thalamus and brain stem nuclei. J. Comp. Neurol. 190:259-282. Berger, A.J., D.B. Averill, and W.E. Cameron (1984) Morphology of inspiratory neurons located in the ventrolateral nucleus of the tractus solitarius of the cat. J. Comp. Neurol. 224r60-70. Brining, S.K., M.N. Lehman, and D.V. Smith (1989) Tracing the central connections of the IXth cranial nerve in the hamster usingperiganglionic placement of HRP or fluorescent tracers. Achems Abstr. 11t100. Cechetto, D.F., D.G. Standaert, and C.B. Saper (1985) Spinal and trigeminal dorsal horn projections to the parabrachial nucleus in the rat. J. Comp. Neurol. 240:153-160. Cedarbaum, J.M., and G.K. Aghajanian (1978) Afferent projections to the rat locus coeruleus as determined by a retrograde tracing technique. J. Comp. Neurol. 178:l-16. Ciriello,J.,A.W. Hrycyshyn, and F.R. Calaresu (1981) Glossopharyngeal and vagal afferent projections to the brain stem of the c a t A horseradish peroxidase study. J. Auton. Nerv. Syst. 4:63-79. Cowan, W.M., D.I. Gottlieb, A.E. Hendrickson, J.L. Price, and T.A. Woolsey (1972) The autoradiographic demonstration of axonal connections in the central nervous system. Brain Res. 37.2141. Davidovitch, M., and M.C. Whitehead (1988) Intramedullary connections of the hypoglossal nucleus in the hamster. J. Dent. Res. 67206. Davis, B.J., and T. Jang(1985) Gustatory recipient zone of the nucleus of the solitary tract in the hamster: Electron microscopic observations. Achems Abstr. 7:37. Davis, B.J., and T. Jang (1986) The gustatory zone of the nucleus of the solitary tract in the hamster: Light microscopic morphometric studies. Chem. Senses 11213-228. Davis, B.J., and T. Jang (1988) A Golgi analysis of the gustatory zone of the nucleus of the solitary tract in the adult hamster. J. Comp. Neurol. 278:388-396. Fulwiler, C.E., and C.B. Saper (1984) Subnuclear organization of the efferent connections of the parabrachial nucleus in the rat. Brain Res. Rev. 7:229-259. Grill, H.J., and R. Norgren (1978) The taste reactivity test. 11. Mimetic responses to gustatory stimuli in chronic thalamic and chronic decerebrate rats. Brain Res. 143381-297. Gwyn, D.G., R.A. Leslie, and D.A. Hopkins (1979) Gastric afferents to the nucleus ofthe solitary tract in the cat. Neurosci. Lett. 14~13-17. Gwyn, D.G., R.A. Leslie, and D.A. Hopkins (1985) Observations on the afferent and efferent organization of the vagus nerve and the innervation of the stomach in the squirrel monkey. J. Comp. Neurol. 239:163-175. Halsell, C.B., and M.E. Frank (1989) Cytoarchitecture of the pontine taste area in the golden hamster. Chem. Senses 14:707. Hamilton, R.B., and R. Norgren (1984) Central projections of gustatory nerves in the rat. J. Comp. Neurol. 222t560-577. Hanamori, T., and D.V. Smith (1989) Gustatory innervation in the rabbit: Central distribution of sensory and motor components of the chorda tympani, glossopharyngeal, and superior laryngeal nerves. J. Comp. Neurol. 282: 1-14. Hanker, J.S., P.E. Yates, C.B. Metz, and A. Rustioni (1977) A new specific, sensitive and non-carcinogenic reagent for the demonstration of horseradish peroxidase. Histochemistry 9t789-792

ACKNOWLEDGMENTS

574 Herbert, H., M.M. Moga, and C.B. Saper (1990) Connections of the parabrachial nucleus with t.he nucleus of the solitary tract and the medullary reticular formation in the rat. J. Comp. Neurol. 293:540-580. Hermann, G.E., and R.C. Rogers (1985) Convergence of vagal and gustatory afferent input within the parabrachial nucleus of the rat. J. Auton. Nerv. Syst. 13:l-17. Higgins, G.A., G.E. Hoffman, S. Wray, and J.S. Schwaber (1984) Distribution of neurotensin-immunoreactivity within baroreceptive portions of the nucleus of the tractus solitarius and the dorsal vagal nucleus of the rat. J. Comp. Neurol. 226:155-164. Kalia, M. (1977) Neuroanatomical organization of the respiratory centers. Fed. Proc. 36~2405-2411. Kalia, M., and M.-.M. Mesulam (1980a) Brain stem projections of sensory and motor components of the vagus complex in the cat: I. The cervical vagus and nodose ganglion. J. Comp. Neurol. 193~435465. M i a , M., and M.-.M. Mesulam (1980b) Brain stem projections of sensory and motor components of the vagus complex in the cat: 11. Laryngeal, tracheobronchial, pulmonary, cardiac, and gastrointestinal branches. J. Comp. Neurol. 193:467-508. Kalia, M., and J.M. Sullivan (1982) Brainstem projections of sensory and motor components of the vagus nerve in the rat. J. Comp. Neurol. 21 1:248-264. King, G.W. (1980) Topology of ascending brainstem projections to nucleus parabrachialis in the cat. J. Comp. Neurol. 191:615-638. Lasiter, P.S., and D.L. Kachele (1988) Organization of GABA and GABAtransaminase containing neurons in the gustatory zone of the nucleus of the solitary tract. Brain Res. Bull. 21:623-636. Loewy, A.D., and H. Burton (1978) Nuclei of the solitary tract: Efferent projections to the lower brain stem and spinal cord of the cat. J. Comp. Neurol. 181:42 1 4 5 0 . Lumsden, A.D. (1923) Observations on the respiratory centres in the cat. J. Physiol. (Lond.)57~152-160. Mesulam,M.-.M. (1976) The blue reaction product in horseradish peroxidase neurohistochemistry Incubation parameters and visibility. J. Histochem. Cytochem. 24:1273-1280. Miceli, M.O., and C.W. Malsbury (1985) Brainstem origins and projections of the cervical and abdominal vagus in the golden hamster: A horseradish peroxidase study. J. Comp. Neurol. 237:65-76. Norgren, R. (1978)Projections from the nucleus of the solitary tract in the rat. Neuroscience 3907-218. Norgren, R. (1985) Taste and the autonomic nervous system. Chem. Senses 10: 143-16 1. Norgren, R., and C.M. Leonard (1973) Ascending central gustatory pathways. J. Comp. Neurol. 150:217-238. Norgren, R., and C. Pfafl’mann (1975) The pontine taste area in the rat. Brain Res. 91:99-117. Ogawa, H., T. Imoto, and T. Hayama (1984) Responsiveness of solitario-

M.C. WHITEHEAD parabrachial relay neurons to taste and mechanical stimulation applied to the oral cavity in rats. Exp. Brain Res. 54:349-358. Orona, E., J.W. Scott, and E.C. Rainer (1983) Different granule cell populations innervate superficial and deep regions of the external plexiform layer in rat olfactory bulb. J. Comp. Neurol. 21 7~227-237. Paxinos, G., and C. Watson (1986) The Rat Brain in Stereotaxic Coordinates. Orlando: Academic Press. Ricardo, J.A., and E.T. Koh (1978)Anatomical evidence of direct projections from the nucleus of the solitary tract to the hypothalamus, amygdala, and other forebrain structures in the rat. Brain Res. I53:l-26. Ross, C.A., D.A. Ruggiero, and D.J. Reis (1985) Projections from the nucleus tractus solitarii to the rostra1 ventrolateral medulla. J. Comp. Neurol. 242:511-534. Scheibel, M.E., and A.B. Scheibel (1958) Structural substrates for integrative patterns in the brain stem reticular core. In H.H. Jasper et al. (ed): Reticular Formation of the Brain. Boston: Little Brown, pp. 31-55. Shapiro, R.E., and R.R. Miselis (1985) The central neural connections of the area postrema of the rat. J. Comp. Neurol. 234:344-364. Stanford, T.R., and M.C. Whitehead (1984) Central organization of afferent fibers of the posterior tongue in the hamster. Neurosci. Abstr. 10:858. Sweazey, R.D., and R.M. Bradley (1986) Central connections of the lingualtonsillar branch of the glossopharyngeal nerve and the superior laryngeal nerve in lamb. J. Comp. Neurol. 245~471482. Travers, J.B. (1988) Efferent projections from the anterior nucleus of the solitary tract ofthe hamster. Brain Res. 457:l-11. Travers, J.B., and R.Norgren (1983) Afferent projections to the oral motor nuclei in the rat. J. Comp. Neurol. 220:280-298. Travers, S.P. and D.V. Smith (1984) Responsiveness of neurons in the hamster parabrachial nuclei to taste mixtures. J. Gen. Physiol. 84:221250. Whitehead, M.C. (1985) The neuronal architecture of the solitary nucleus of the hamster. Achems Abstr. 7:183. Whitehead, M.C. (1986) Anatomy of the gustatory system in the hamster: Synaptologyof facial afferent terminals in the solitary nucleus. J. Comp. Neurol. 244~72-85. Whitehead, M.C. (1988) Neuronal architecture of the nucleus of the solitary tract in the hamster. J. Comp. Neurol. 276:547-572. Whitehead, M.C. (1989) Ultrastructure of an identified cell type in the gustatory subdivisions of the nucleus of the solitary tract in the hamster. Acherns Abstr. 11~102. Whitehead, M.C. and L.D. Savoy (1987)The solitary nucleus of the hamster: Cytoarchitecture andpontine connections. Ann. N.Y. Acad. Sci. 510:707709. Whitehead, M.C., S.A. Ahmed, L.D. Savoy, and M. McPheeters (1988) Morphological types of neurons in taste-responsive sites in the hamster’s solitary nucleus. Chem. Senses 13~746.