The somatostatin systems of the guinea-pig brainstem

0306-4522/85 53.00 + 0.00 Pergamon Press Ltd c, 198.5 IBRO

Neuroscience Vol. 15, No. I, pp. 215-235, 1985 Printed in Great Britain

THE SOMATOSTATIN SYSTEMS OF THE GUINEA-PIG BRAINSTEM E. TAIW@IERCE*, E. LI~~N~NS~IN~ and S. C. FE~~AN~§ *Departments of Anatomy and Psychiatry, Harvard Medical School and Massachusetts General Hospital, Fruit Street, Boston, MA 02115, U.S.A.; tDepartment of Medicine, Yale University, New Haven, CT 06520, U.S.A.; SDepartment of Anatomy, University of Medicine and Dentistry of New Jersey, New Jersey Medical School, 100 Bergen Street, Newark, NJ 07103, U.S.A. Abstract---The tetradecapeptide somatostatin has been shown to have a widespread distribution in the rat brain. Except for its role in the inhibition of growth hormone secretion, the function of this molecule in the remainder of the central nervous system is unknown. To address this problem, the dist~bution of somatostatin-like immunoreactivity in the guinea-pig brainstem was examined systematically.” Of 116 nuclei and/or areas, 34 nuclei had somatostatin neurons, 32 did not have any immunoreactivity and the remainder had immunoreactive fibers and/or terminals. Cranial nerve motor nuclei-somatic, branchiomeric and visceral--did not contain somatostatin neurons; somatostatin fibers were present in all nuclei with the exception of the somatic motor nuclei which innervate the ocular muscles. Of the cranial nerve sensory nuclei-both somatic and visceral-somatostatin neurons were present only in the somatic nuclei nervi spinal trigeminal caudalis, interpolaris and oralis; all of these nuclei, however, contained substantial numbers of immunoreactive fibers. Somatostatin neurons and fibers were also present in the spinal somatic sensory nuclei cuneatus medialis and gracilis. Of the cranial nerve speciat somatic sensory nuclei, somatostatin neurons were present in two vestibular nuclei-spinalis and medialis-and in the ventral cochlear nucleus. Not all of these nuclei contained somatostatin fibers. Of the nuclei related to the auditory system, somatostatin neurons were present only in the dorsal and ventral trapezoid nuclei and in the cortex of the inferior colliculus. In nuclei of the visual system in the brainst~, ~matostatin neurons were present only in the superior colliculus. Of the raphe nuclei, four had somatostatin neurons--magnus, obscurus, pallidus and superior centralis; somatostatin fibers were present in all raphe nuclei. Of 24 nuclei in the reticular formation, 12 had somatostatin neurons-most notably nuclei gigantocellularis and ~ra~gant~Ilula~~and only two nuclei, paranigralis and sagulum, did not contain any immunoreactive ftbers. In the cerebellum, somatostatin fibers and terminals were restricted to the deep cerebellar nuclei. Of the 11 nuclei projecting to the cerebellum, five contained somatostatin neurons and the majority received somatostatin fibers. Of the limbic system nuclei, somatostatin neurons were confined to the central grey at both pontine and mesencephalic levels. Somatostatin neurons were present in the substantia nigraempactus and lateralis, but not in reticularis-and absent from the nucleus ruber. The area postrema contained somatostatin neurons and fibers but in the ependyma immunoreactivity was confined to a limited numbers of fibers. The results of the present study demonstrate that although the majority of areas in the brainstem contain somatostatin, the peptide does not appear to be an obligatory com~nent of any one system.

somatostatin, which has made characterization of Based on immun~hemi~ methods an extensive network of neurons, fibers and terminals, reported to receptors for the peptide difficult. contain somatostatin, has recently been demonIt is the very widespread anatomical localization of strated to exist within the brains of rat,‘z~‘617~m~2*~237 somatosta~n which makes it both difficult to study, bird (Melopsittucus undufutus)M and frog (Ram yet suggests its importance in the central nervous cotesbiuna).25 In these reports the distribution of system. Because of this, we have undertaken a deneurons containing somatostatin was widespread and tailed and systemic analysis of the distribution of occurred at all levels of the brain. Fiber pathways and immunoreactive-like somatostatin (I-somatostatin) in terminals were extensive. the guinea-pig brainstem (Cuuiu cobuyu). We have In spite of the widespread distribution of the organized this atlas to stress the concept of the peptide littie is known of its function outside of its brainstem as a system of functional and/or anarole in the regulation of growth hormone secretion.’ tomical complexes.9.‘0 Because of the complicated Although somatostatin has profound effects on a anatomy of many of these nuclei we have described, number of behaviors, 26 electrophysiological studies where appropriate, both the morphology and intrasuggest both excitatory” and inhibitory33 roles for the nuclear locahxation of I-somatostatin-containing molecule. To date there are no known inhibitors of neurons as well as the spatial arrangement of immunoreactive fibers and terminals. This detailed approach to the neuro~atomy-mo~hoiogy and §To whom all ~~~nden~ should be addressed. distribution-of somatostatin systems will be particuAbbr&ution: I-somatostatin, immunoreactive-like somatostatin. larly useful in future studies on the function of 21s

216

E. Taber-Pierce el LII

somatostatin. organization of transmitter and inter-relationships of brainstem nuclei

systems

cross-reactivity of our antiserum to molecules other than somatostatin is a problem.

EXPERIMENTAL PROCEDURES Tissue preparation Brains from three adult guinea pigs (Cavia cobaya) were used in this study. Animals received no pharmacological pretreatment before sacrifice, e.g. colchicine. Animals were anesthesized, then killed by cardiac perfusion with saline followed by Bouin’s solution. Tissue was postfixed in Bouin’s (12-24 h), dehydrated in a graded series of alcohols, cleared in cedar wood oil and xylene, then embedded in paraffin. Brains were sectioned in the coronal plane at 10 pm. Immunocytochemistry Somatostatin was localized immunocytochemically using a highly specific antiserum to the peptide and the unlabeled antibody-enzyme technique of Stemberger” (peroxidaseantiperoxidase technique). Details of the procedure and antibody specificity have been published.16.17.29In brief, sections, rehydrated to phosphate buffer (NaPG,, 0.1 M, pH 7.4). were sequentially exposed to (1) antiserum to somatostatin, I: 250 (48-72 h; 4°C); (2) goat anti-rabbit immunogiobulin G; I: 100 (0.5 h); and (3) peroxidaseantiperoxidase, 1: 400 (1 h) with buffer washes (2 x 5 min) between each step. Sections were developed in 3.3’-diaminobenzidine-HCI. 50 mn‘!! (in 0.05 M Tris buffer. pi 7.6) with 0.1% H,Oz. Alternat; sections were stained with cresyl-echt violet for the determination of nuclear groups. Controls The specificity of the antiserum for somatostatin was determined both by radioimmunoassay and in tissue sections. By radioimmunoassay, the antiserum was shown not to cross-react with other known peptides. In the immuthe specificity of nocytochemical analysis, the 3,3’-diaminobenzidine reaction product for somatostatin was determined by the absence of staining in neuronal elements in sections in which antiserum to somatostatin was replaced with either (1) antiserum to somatostatin which had been preadsorbed with synthetic somatostatin (4 pm) or (2) normal (preimmune) rabbit serum (1: 10, 1: 100). Because of the large number of substance P-containing neurons in the brainstem,‘O the possibility of cross-reactivity of the somatostatin antiserum with substance P was investigated. Staining of adjacent sections at 10 levels of the guinea pig brainstem with antiserum to either substance P (1: 300; Immunonuclear Corp., Minn.) or our antiserum to somatostatin resulted in staining of different populations of neurons and fibers. In addition, preadsorption of our antiserum to somatostatin with substance P (IO pm) did not affect either the distribution or number of immunoreactive elements. Thus, although there may be some neurons in which the two molecules coexist, it is unlikely that possible

RESULTS

In the guinea-pig brainstem somatostatin reactivity could be identified within many brainstem nuclei cornprizing most functional and anatomical areas. Fibers and terminals were widespread and there was no single characteristic type of somatostatin neuron. Given the complicated anatomy of these elements and of the brainstem nuclei. for ease of presentation the data are presented along the lines for established recognized systems of neuroanatomical organization: somatic motor and sensory column nuclei; branchiomeric motor column nuclei; visceral motor and sensory column nuclei; nuclei related to the vestibular system; nuclei related to the auditory system; nuclei related to the visual system; raphe nuclei; reticular formation nuclei: the cerebellum and nuclei with established efferent connections to the cerebellum; limbic system areas and nuclei; described extrapyramidal system nuclei and, finally, the ependyma and its related circumventricular organ, the area postrema. In the following text we describe, in detail, the morphology and distribution of I-somatostatincontaining elements throughout the extent of each nucleus. The results are summarized in Table 1 and presented in schematic form in Figs l-12 and in selected micrographs in Figs 13-24. Crunial nerve somatic nuclei Motor. No I-somatostatin-containing neurons were seen in these nuclei. No I-somatostatincontaining fibers were present in III, IV and VI. A dense band of immunoreactive fibers in association with III is shown in Fig. 11; this band was considered to be limited to the Edinger-Westphal nucleus (see below). In contrast to the other nuclei, XII was characterized by a dense plexus of immunoreactive fibers in all regions except for the most rostral. A few fibers continued beyond the borders of this nucleus into areas of the raphe, nucleus intercalatus and the adjacent region of the reticular formation. Sensory. Of the six sensory nuclei, I-somatostatincontaining neurons were present in three: V oralis, V

Table 1. Distribution of somatostatin

Nucleus or area 1. Cranial nerve somatic motor nuclei III IV VI XII 2. Cranial nerve branchiomeric motor nuclei V VII

in the guinea-pig brainstem Positive neurons

Positive fibers

-

+_ +++

-

+-l--l+++

Positive terminals on soma ++ +

217

Somatostatin in the guinea-pig brainstem Tgble 1 (cmtinaed~

Nucleus or area IX (Ambiguus) X (Ambiguus) XI (Ambiguus)

Positive neurons

Positive

-

-k-i-+ +++ +++

-

fibers

Positive terminals on soma -I+ -I-

3. Visceral motor nuclei III (Edinger-Westphal) VII (Superior salivatorius) IX (Inferior salivatorius) X (Vagi dorsalis)

-

+++ ++ ++ ++++

+ ? ?

4. Cranial nerve somatic sensory nuclei V Principalis V Oralis V Interpolaris V Caudalis, zonalis V Caudalis, substantia gelatinosus V Caudalis, magnocellularis V Mesencephalicus (primary neurons) Nuclei insulae cuneatilateralis (paratrigeminalis)

+ + + -

+ + + + ++++ + + +

-

+-

+-I-++ ++++ +

5. Cranial Tractus Tractus Tractus

nerve visceral sensory solitarii solitarii, commissuralis solitarii, gelatinosus

6. Spinal nerve somatic sensory Cuneatus medialis Gracilis 7. Cranial nerve special somatic sensory Vestibularis spinalis Vestibularis medialis Vestibularis lateralis Vestibularis superior Nucleus y Cochlearis dorsalis (outer) Cochlearis dorsalis (intermediate) Cochlearis dorsalis (inner) Cochlearis ventralis (anterior) Cochlearis ventralis (posterior) 8. Auditory relay Olivaris superior lateralis Olivaris superior medialis Trapezoidalis dorsalis Trapezoidalis medialis Trapezoidalis lateralis Trapezoidalis ventralis Lemniscus dorsalis Lemniscus ventralis Colliculus inferior Cortex Dorsomedialis Externalis Intercolliculus Interstitialis Principalis Corpus geniculati medialis Dorsalis Medialis Ventralis

+ f

+ +

f + -

+ + +

+ +

++++ +

+

+

++

+ ++

-

+

-

-

9. Visual system related Darkschewitsch Interstitialis (Cajal) Opticus radix basalis Parabigeminalis Colliculus superior

-

10. Raphe Dorsalis Linearis intermedialis Linearis rostralis

-

+

+ ? + + + + 7.

+

+

+ +

-

+ -

-

+

+ +

++ + ++ + ++

E. Taber-Pierce er al.

218

Table

Nucleus or area Magnus Obscurus Pallidus Pontis Superior centralis 11. Reticular formation

Annularis Cuneiformis Subcuneiformis Gigantocellularis Paragigantocellularis dorsahs Paragigantocellularis lateralis Kolliker-Fuse Locus coeruleus Subcoeruleus Medullae oblongatae centralis Dorsalis Ventralis Parabrachialis pigmentosus Parabrachialis lateralis Parabrachialis medialis Paranigralis Parvocellularis Parvocellularis compactus Peripeduncularis Pontis centralis caudalis Pontis centralis oralis Sagulum Tegmentalis ventralis (Tsai) Suprageniculatus Nucleus z

I (conrimed)

Positive neurons

Positive fibers

++t ++t + +

+++ +++ ++ + +

4-t-f ++ ++

++ ++ ++ ++ ++ ++ ++ + +

12. Cerebellum and nuclei, which project to the cerebellum Cortex of cerebellum Deep cerebellar nuclei Corporis pontobulbaris Griseum pontis Cuneatus lateralis Inferior olive + Intercalatus + Interfascicularis hypoglossi Paramedianus dorsalis Paramedian reticularis ventrahs Praepositus hypoglossi Reticularis lateralis Reticularis tegmenti pontis (Bechterew) + Roller + Nucleus x 13. Limbic system Griseum centrale mesencephali Griseum centrale pontis Tegmentalis dorsalis (von Gudden) Tegmentalis ventralis (von Gudden) Interpeduncularis 14. Extrapyramidal Ruber Substantiae nigrae Compactus Lateralis Reticularis 15. Ependyma and area postrema Area postrema Ependyma

Positive terminals on soma + +

+ ++ ++ +++ ++ ++ + ++ +

+ +++

++ ++ ++ ++++ ++++

+ +

++ ++ +++ ++ ++ ++ ++ ++

+ + + + +-

+ -

+

+ + + + + + + +

+ + +

++ + -

++ ++ -

-

-

+

-

+

+

+ + -

+ ++

-

++ +-

+ -

-

-

-

Relative density of somatostatin immunoreactivity in 116 areas of the guinea pig brainstem. For each area the presence of somatostatin neurons, fibers and terminals surrounding cell bodies is rated from + + + + (most dense) to - (absent). For areas marked with a “?” see description in text.

Somatostatin in the guinea-pig brainstem interpolaris, and V caudalis magnocellularis. The majority of these neurons were medium-sized, a few were seen in V interpolaris. large neurons Immunoreactive-like somatostatin-containing fibers were seen in all six nuclei but were especially dense in substantia gelatinosa. The majority of these fibers appeared to terminate in the neuropil. In V interpolaris, I-somatostatin-containing fibers formed two interrupted bands lateral and medial to the spinal trigeminal tract (Fig. 2). Special sensory. In the nucleus cochlearis (Figs 5 and 6) I-somatostatin-containing neurons were limited to the nucleus ventraiis, both anterior and posterior subdivisions. In the anterior nucleus these neurons, medium and large multipolar, were limited to the caudal half of the nucleus. In the posterior nucleus a few positive medium-sized neurons were located caudally and ventrally; a few large, multipolar I-somatostatin-containing neurons were observed at the level of the entrance of the cochiear

219

fibers into the ventral nucleus. Immunoreactive fibers were sparse in both the dorsal cochiear and ventral nuclei. In the anterior division of the ventral nucleus positive terminals surrounded the soma of large globular neurons while in the posterior division immunoreactive terminals were seen on octopus neurons (Fig. 16). In the vestibular nuclei, I-somatostatin~on~ining neurons were limited to the medial and spinal vestibular nuclei. These neurons were multipolar and small to medium-sized. The medial nucleus contained two groups of immunoreactive neurons: a group of small neurons dorsal to the ventral pole of the nucleus tractus solitarius and a second group in the most rostra1 area of the nucleus; this second group appeared to form a continuum with I-somatostatin neurons in the adjoining periaquiductal grey. In the spinal vestibular nucleus (Figs 4-6) the majority of the I-somatostatin neurons were located in the dorsal region. Small numbers of I-somatostatin-containing

Abbreviationsused in figures A Ap Be :d Cl Cm Cod cov cs CU cv D Dm DS Dt EW : Gc $? :c z Lc Li Lid Llv Lr Lro MGB Oi OS1 P Pb Pbl Pbm PC Pgd Pgl Pot

nucleus ambiguus area postrema brachium conjunct&urn cortex, inferior colliculus nucleus medullae oblongatae centralis dorsalis nucleus cuneatus lateralis nucleus cuneatus medialis nucleus cochlearis dorsalis nucleus cochlearis ventralis nucleus raphe superior centralis nucleus cuneiformis nucleus medullae obiongatae centralis ventralis nucleus Darkschewitsch nucleus trapezoidalis dosalis nucleus tegmenti ~unculo~ntinus dissipatus nucleus tegmentaiis dorsalis (von Gudden) nucleus Edinger-Westphal subnucleus gelatinosus nucleus gracilis nucleus ~gant~llu~a~s griseum centralis pontis corporis pontobulbaris griseum pontis nucleus interstitialis (Cajal) inferior colliculus nucleus interpeduncuiaris nucleus Kolliker-Fuse nucleus locus coeruieus nucleus raphe Iinearis intermediahs nucleus lemniscus lateralis dorsalis nucleus lemniscus lateraiis ventralis nucleus raphe rostralis linearis nucleus raphe linearis oralis medial geniculate body inferior olivary complex nucleus olivaris superior lateralis nucleus olivaris superior medialis nucleus principalis posterior nucleus parabigeminalis

nucleus nucleus nucleus nucteus nucleus nucleus

parabrachialis lateralis parabrachiahs medialis parvocellularis ~ra~gant~llu~~s dorsalis paragigantocellularis lateralis pontis centralis caudalis

PO0

!$Y Pv Rd Rm Ro Rob Rpa Rpo Rtp RU s Sa SC SC scu

sgt

Sn T n Tm TV X 2 III

IV V

Vi Vm vo VP VI VII VIII 1 VIII m VIII s VIII sp X XII

nucleus pontis centralis oralis nucleus peripeduncuhuis nucleus parabrachiahs pigmentosus nucleus praepositus hypoglossi nucleus parvocellularis compactus nucleus raphe dorsalis nucleus raphe magnus nucleus Roller nucleus raphe obscurus nucleus raphe pallidus nucleus raphe= pontis nucleus reticularis tegmenti (Bechterew) nucleus tuber nucleus tractus solitarius nucleus saguium colliculus superior nucleus subcoeruleus nucleus subcuneiformis nucleus suprageniculatus nucieus substantiae nigrae nucleus tegmentahs ventralis (Tsai) nucleus trapezoidalis lateralis nucleus trapezoidalis medialis nucleus trapezoidalis ventralis nucleus x nucleus z nucleus nervi oculomotorius nucleus nervi trochlearis nucleus nervi trigemini motorius nucleus net-vi tractus spinalis trigemini caudalis magnocellularis nucleus nervi tractus spinalis trigemini interpolaris nucleus trigemini mesencephalicus nucleus nervi tractus spinalis trigemini oralis nucleus nervi trigemini sensibihs principalis nucleus nervi abducentis nucleus net-vi facialis nucleus vestibularis lateralis nucleus vestibularis medialis nucleus vestibularis superior nucleus vestibufaris spinalis nucleus nervi vagi dorsalis motorius nucleus nervi hypoglossi

E. T&m-Pierce

81 ul.

Figs I-12. Distribution of ~~~tos~tin~o~t~nin~ neurons (left-half of each drawing) and fibers (ant-half of each drawing) at twelve levels of the guinea pig brainstem In these diagrams each filled cirde represents one somatostatin-cantaining neuron. Abbreviations: see “Abbreviations used in figures*‘.

Somatostatin

in the guinea-pig brainstem

221

222

E. Taber-Pierce

PI at.

Somatostatin in the guinea-pig brainstem

223

224

E. Taber-Pierce cl ul.

Somatostatin in the guinea-pig brainstem

fibers were present in the medial, lateral and spinal nuclei. A small, discontinuous band of positive fibem extended across the dorsal border of both the medial and spinal nuclei. In contrast to the disposition of I-somatostatin-containing terminals in the majority of brainstem nuclei, positive terminals in the lateral and spinal vestibular nuclei were limited to only two distinct types of vestibular cells, the large and giant neurons (Fig. 15). Cranial nerve branchiomeric motor nuclei No somatostatincontaining neurons were present in any of these nuclei. All three nuclei, however, received a dense and uneven innervation of somatostatin fibers. Relatively dense plexuses of Isomatostatin fibers were seen in V, VII, the accessory facial nucleus (along the medial side of the nerve root of VII) and nucleus ambiguus. Primary spinal somatic sensory nuclei, the dorsal column nuclei

The nuclei cuneatus medialis and gracilis (Figs 1 and 2) contained a small number of positive, medium-sized neurons. Immunoreactive-like somatostatin fibers traveled in the dorsal regions of the nuclei and from there turned ventrally into the deeper regions. Their number was densest in the rostra1 regions of both nuclei. Terminal fibers were present in both nuclei, although no terminals were observed on the soma of neurons. Cranial nerve visceral nuclei Sensory. The nucleus tractus solitarius, including the region of the subnucleus commissuralis (Figs 1, 2 and 17-19), contained a dense plexus of positive fibers throughout its caudal-rostra1 extent, with the exception of the area of the sub-nucleus gelatinosus where the density of fibers was sparse. The nucleus probably contains a small number of somatostatin neurons, but the density of immunoreactive fibers make their validation difficult (Fig. 19). Motor. None of the cranial nerve visceral motor nuclei contained I-somatostatin. The small number of medium-sized immunoreactive neurons in the region of the parasympathetic divisions of VII and IX are probably part of the nucleus parvocellularis. The Edinger-Westphal nucleus (Fig. 12), in its position straddling the raphe, contained a dense plexus of fibers. Positive terminals were present on the soma of a small number of neurons. The dorsal vagal nucleus was both surrounded by, and contained, a dense plexus of positive fibers (Figs 1-3 and 17-19) the majority of which, however, appeared to terminate in the neuropil. Auditory relay nuclei

Within this complex of nuclei, I-somatostatincontaining neurons were limited to the ventral and dorsal trapezoid nuclei (Fig. 6) and the cortex of the inferior colliculus. The I-somatostatin neurons of the

225

dorsal and ventral trapezoid nuclei extended forward beyond the usually described rostra1 poles of these nuclei between the lateral border of the griseum pontis and the ventral nucleus of the lateral lemniscus, an area generally included at this level as part of the reticular formation (Figs 7 and 8). In both trapezoid nuclei the cells ranged from small to medium; the majority of somatostatin neurons were medium-sized (Fig. 14). In the cortex of the inferior small positive neurons and immucolhculus, noreactive fibers were present both in the outer zone and in the deepest region. In general, except for the superior olivary complex and the nuclei of the lateral lemniscus, the auditory relay nuclei were characterized by a scarcity of Isomatostatin fibers. Nuclei of the visual system

In the brainstem nuclei concerned with vision I-somatostatin-containing neurons were present only in the stratum cinerium of the superior colliculus. Two nuclei, Darkschewitsch and opticus radix basalis, contained neither immunoreactive neurons nor fibers. Raphe nuclei

Of the eight raphe nuclei, four-raphe pallidus, obscurus, magnus and superior centralis-ontained I-somatostatin neurons; immunoreactive fibers were seen in ail of the raphe nuclei. The raphe magnus (Figs 3-6) was characterized by large numbers of I-somatostatin-containing neurons of all sizes. These neurons were not uniformly distributed throughout the nucleus; the majority were found at caudal and rostra1 levels leaving a small I-somatostatin cell-free zone between these two groups. In contrast, Isomatostatintontaining neurons of variable size were evenly distributed in the nuclei raphe obscurus (Figs 2-5) and pallidus (Figs 2-4). In the nucleus superior centralis (Figs 8-12) immunoreactive neurons were limited to the small cells of the lateral region of the nucleus. In general, the distribution of I-somatostatincontaining fibers paralleled that of the neurons. Raphe nuclei with I-somatostatin neurons contained moderate to dense plexuses of immunoreactive fibers throughout their extent. In the remainder of the raphe nuclei the I-somatostatin-containing fiber plexuses were moderate to sparse. Nuclei of the reticular formation

In the guinea pig, 24 nuclei were identified in the reticular formation. Of these, 12 contained immunoreactive neurons and only in two nucleiparanigraiis and sagulum-was there no evidence of I-somatostatin-containing fibers and/or terminals (Table 1). Several nuclei were notable for their large number of immunoreactive neurons and fibers, These included: nucleus gigantocellularis (Figs 3, 4, 22 and 25), paragigantocellularis dorsalis and lateralis (Figs

226

E. Taber-Pierce er al.

Figs 13 and 14. Somatostatin neurons in the periaqueductal grey (Fig. 13) and in the trapezoid body (Fig. 14). In both figures note that the reaction product fills the cytoplasm and, in Fig. 13, the highly branched axon (arrows). Fig. 13 x 1000; Fig. 14 x600. Figs 15 and 16. Two patterns of innervation of neurons by somatostatin-containing

terminals and fibers.

Fig. 15. A neuron in the lateral spinal vestibular nucleus. Note the presence of somatostatin-containing “terminal-like” structures (arrows) limited to the outermost rim of the cell. The remainder of this cell, and the one next to it, are devoid of other somatostatin immunoreactivity. Cresyl-violet-stained section. x 1000. Fig. 16. An “octopus” neuron in the ventral cochlear nucleus. In this neuron, a thin, varicose-like somatostatin-containing fiber can be traced around the cell body (arrows); the neuron’s cytoplasm, and the surrounding neuropil, are devoid of somatostatin. Cresyl-violet stained section. x 600. Figs 17-19. At the level of Fig. 3, the more dorsal areas of the brainstem are characterized by dense plexuses of immunoreactive fibers and a few I-somatostatin-containing neurons. Fig. 17. Low-power micrograph to show the distribution of I-somatostatin elements. Dense plexuses of immunoreactive fibers characterize the area of the dorsal vagal nucleus (X) and nucleus tractus solitarius (S) and extend into the nuclues z. Details of areas indicated by “a” and “b” are shown in Figs 18 and 19, respectively. VIII, nucleus vestibularis spinalis. Cresyl-violet counter-stained section. x 450. Fig. 18. The area encompassed by the nucleus tractus solitaris (S) and the dorsal vagal nucleus (X) is characterized by a dense plexus of somatostatin fibers. Note that the fibers surround non-immunoreactive cell bodies. x 350. Fig. 19. The area indicated in Fig. 17 “b”. Within this dense plexus of immunoreactive fibers a small number of I-somatostatin-containing neurons (arrows) are demonstrable. The majority of neurons in this area do not contain the peptide although they may receive I-somatostatin terminals on their cell bodies. Cresyl-violet counter-stained section. x 340. Figs 20-23. Somatostatin immunoreactivity in the raphe magnus-nucleus gigantoceliularis. At this level of the brainstem (Fig. 4) there appears to be a continuous population of I-somatostatin cells and fibers extending from the raphe magnus (Rm) to the nucleus gigantocelhilaris (Gc) and the nucleus paragigantocelhrlaris lateralis (Pgl). Fig. 20. Low-power dark-field photomicrograph

of the area. x 60.

Fig. 21. A higher-power micrograph of nucleus paragigantocellularis lateralis to show the numerous I-somatostatin neurons in the area (arrows) and the dense plexus of immunoreactive fibers. These neurons are generally large and resemble cells in the nucleus gigantocellularis (see below). Cresyl-violet counterstained section. x 100. Fig. 22. Somatostatin neurons in the nucleus raphe magnus (arrows). These neurons are generally medium-sized and lightly staining. Cresyl-violet counter-stained section. x 400. Fig. 23. The nucleus gigantocellularis contains many large I-somatostatin-containing neurons (curved arrows) and a plexus of fine varicose immunoreactive fibers (small arrows). Cresyl-violet counter-stained. x 330. Fig. 24, Photomontage of the area surrounding the containing neurons (arrows) in the nucleus parabrachialis the brachium conjunctivum (fk) into areas occupied by coeruleus (Lc) and the midline periaqueductal

brachium conjunctivum (Bc). I-somatostatinlateralis (Pbl) send their fibers over and around the mesencephalic nucleus of V (Vm), the locus grey. m, towards the midline. x 150.

227

. ..,;,

.+ _

.-

229

----

_-...__._._. 230

231

Somatostatin in the guinea-pig brainstem 1, 2, 22 and 23); nuclei medullae oblongatae centralis and ventralis (Figs 1 and 2); nucleus parvocellularis (Figs 3-5); nuclei pontis centralis caudalis and oralis (Figs 5-8 and 21). The largest number of I-somatostatin-containing neurons in the reticular formation was found in the nucleus gigantocellularis (Figs 3, 4, 20 and 23). The majority of the neurons were concentrated in the caudal portion of the nucleus. The neurons ranged in size from small to large; the peptide was apparently excluded from all of the giant neurons. The nucleus also contained large numbers of somatostatin fibers, particularly in the ventral area where some of the large non-immunoreactive neurons appeared to receive an I-somatostatin input on their cell bodies. In both the nucleus paragigantocellularis dorsalis (Figs 3 and 4) and paragigantocellularis lateralis (Figs 4,20 and 21) I-somatostatin neurons and fibers were seen. As in the nucleus gigantocellularis, these neurons ranged from large to small (Fig. 21). The nuclei had only a moderate number of immunoreactive fibers, and synapses on neuronal cell bodies were rare. Both nucleus medullae oblongata centralis dorsalis and nucleus medullae oblongatae centralis ventralis (Figs 1 and 2) contained a small number of immunoreactive neurons and a large number of positive fibers. The nucleus dorsalis contained two wide bands of I-somatostatin fibers. The first crossed through the dorsal area in a zone ventral to the nuclei tractus solitarius, gracilis and cuneatus. The second band of immunoreactive fibers, located along its medial border, contained the nucleus ambiguus and extended from the ventral border of the nucleus tractus solitarius to the ventral border of the spinal trigeminal tract. Scattered somatostatin neurons, ranging in size from medium to small, were found in both bands. The subnucleus ventralis also contained a medium population of variably sized I-somatostatin neurons and a dense network of immunoreactive fibers, particularly in its lateral and ventral areas. The nuclei parabrachialis lateralis and medialis (Figs 8,9 and 24) contained an abundance of immunoreactive fibers which appeared to take origin from some medium-sized I-somatostatin-containing neurons located in the caudal part of the lateral nucleus (Fig. 24). Immunoreactive-like somatostatin fibers formed a rich plexus in this area. These axons appeared to sweep medially through the brachium conjunctivum into the medial nucleus and then into the area of the periaqueductal grey and the nucleus locus coeruleus (Fig. 24). Fibers from the lateral nucleus were also observed to enter ventrally into the nuclei subcoeruleus and Kolliker-Fuse. The nucleus parvocellularis (Figs 3-5), located laterally at levels of the hindbrain, contained both positive neurons and fibers. immunoreactive fibers were especially dense in regions of preganglionic neurons, including superior and inferior salivatory

neurons, the area just dorsal to the facial nucleus and in the rigion lying medid to the borders of the V oralis and V interpolaris. Small and medium-sized I-somatostatin nesrons were seen both within, and just outside of, this dense dorsal plexus of immunoreactive fibers and scattered in other regions of the nucleus. The nucleus peripeduncularis (Figs 11 and 12) contained a dense plexus of immunoreactive fibers in which some medium-sized I-somatostatin-containing neurons were located. These fibers appeared to extend dorsally along the medial and lateral borders of the medial geniculate body, whereas other fibres turned medially into the nucleus cuneiformis and nucleus subcuneiformis (see descriptions for these nuclei). Both nuclei pontis centralis caudalis and oralis (Figs 5-7) contained I-somatostatin neurons and fibers. The neurons of the caudal nucleus and a group of I-somatostatin neurons from the nucleus oralis were located on the medial border of the ventral nucleus of the lateral lemniscus. These neurons appeared to represent a rostra1 continuation of the cells in the medial area of the nucleus trapezoidalis ventralis. The nucleus tegmentalis ventralis (Tsai) (Figs 11 and 12) contained a sparse number of small or medium-sized I-somatostatin neurons and a moderately dense plexus of immunoreactive fibers which appeared to extend laterally into the nucleus pedunculopontinus dissipatus. In the nuclei tegmenti pedunculopontinus compactus and dissipatus, I-somatostatin-containing neurons were limited to the caudal area of the nucleus dissipatus (Figs 10 and 11). Both nuclei contained moderately dense plexuses of immunoreactive fibers. The nucleus z (Figs 3 and 17) contained a small number of immunoreactive medium-sized and small neurons. The majority of I-somatostatin fibers were seen to course through the dorsal area; relatively fewer positive fibers were present in the deeper portions of the nucleus. The

cerebellum

and

nuclei

in

the hindbrain

with

connections to the cerebellum

In the cerebellum, I-somatostatin was restricted to a small number of fibers in the deep cerebellar nuclei, many of which appeared to terminate on large neurons. The peptide was completely absent from the cerebellar cortex. Of the 11 nuclei with known connections to the cerebellum, four contained Isomatostatin neurons: nucleus intercalatus, which lies between nuclei X and XII; nucleus interfascicularis hypoglossi (Figs 2 and 3); nucleus praepositus hypoglossi (Figs 3-5), in its ventral and lateral regions; and nucleus Roller (Fig. 3). Except for the nucleus praepositus hypoglossi, which contained a moderately dense plexus of immunoreactive fibers, Isomatostatin-containing fibers were generally sparse.

E. Taber-Pierce rr u/

232 Limbic

system nuclei

and areas

Of the nuclei which comprise the limbic system in the brainstem, I-somatostatin neurons were found only in the periaqueductal grey (Figs 4-9, 13); immunoreactive fibers were also seen in the nucleus interpeduncularis (Figs l&12). The periaqueductal grey, griseum centrale mesencephali, contained an abundance of somatostatin fibers and neurons. The positive neurons (Fig. 13) were not uniformly distributed but were both clustered as well as scattered singly. Somatostatin fibers also were both scattered and arranged into dense plexuses. Within the region the distribution of I-somatostatin elements was not uniform. At pontine levels, at the level of the genu of VII rostra] to the locus coeruleus (Figs 4-9) immunoreactive fibers were scattered. At the level of the locus coeruleus there was an abrupt increase in the number of immunoreactive fibers which disappeared at the rostra] end of this region. At the level of the raphe dorsalis (Fig. 10) there was an increase in the number of fibers in the midline. Continuing forwards, at rostra1 levels of the inferior colliculus, there was a marked increase in the number of I-somatostatincontaining neurons to approximately seven to ten per section (Fig. 9) which decreased at the level of the superior colliculus (Figs I1 and 12). Extrapyramidal

nuclei of the midbrain

Immunoreactive-like somatostatin neurons were limited to the subnuclei compactus and lateralis of the substantia nigrae (Figs 11 and 12) and were small or medium-sized. Both of these subnuclei also contained plexuses of immunoreactive fibers, which terminated in the neuropil. In contrast the nucleus tuber (Figs 11 and 12) contained a small number of positive fibers, many of which terminated on the soma of large cells. The ependyma

and area postrema

The ependyma appeared to contain an occasional positive fiber but no positive neurons. The area postrema (Ap) (Fig. 2) contained a moderate to dense plexus of fine immunoreactive fibers which appeared to stream into the region from the area of the nucleus tractus solitarius. A sparse number of positive neurons were scattered throughout the area. DISCUSSION

Our anlaysis of the distribution of somatostatinlike immunoreactivity reveals of 116 nuclei and/or areas examined in this study 34 nuclei or areas had somatostatin positive neurons, 32 areas did not have any somatostatin immunoreactivity and the remainder had immunoreactive fibers and/or terminals. Seven areas contained impressive numbers of immunoreactive neurons. The largest number of immunoreactive neurons was seen in three medullary nuclei: gigantocellularis, raphe magnus and raphe obscurous. Smaller numbers of I-somatostatin neu-

rons were present in the periaqueductai grey at midbrain levels; the peripeduncular nucleus located ventral to the medial geniculate body; the lateral parabrachial nucleus at pontine levels; and the ventral trapezoid nucleus, which contains a group of immunoreactive neurons extending from the rostra1 edge of this nucleus into the midbrain. In the remaining 27 nuclei the population of I-somatostatin neurons was small. When the distribution of Isomatostatin neurons is analysed it is apparent that the peptide is one of the most widely distributed neurotransmitter/neuromodulator substances in the CNS. However, although the majority of areas in the brainstem contain some somatostatin immunoreactivity, the peptide does not appear to be an obligatory component of any one system. Distribution

of somatostatin

Although the distribution of I-somatostatin immunoreactivity was extremely complicated certain patterns could be discerned. Thus taken as a group, the cranial nuclei were remarkable for the presence of I-somatostatin neurons only in sensory nuclei and the absence of immunoreactive fibers from the motor nuclei innervating the somatic extraocular musclesIII, IV and VI. In the cat,*’ as in the guinea-pig, I-somatostatin terminals in the oculomotor nucleus appear to be restricted to the midline, i.e. EdingerWestphal nucleus. These results suggest that the somatic motor nuclei concerned with vision are exclusively excluded form somatostatin influence. Two cranial nerve sensory nuclei-the substantia gelatinosa and the nucleus tractus solitarius-received a dense I-somatostatin innervation, suggesting a role for the peptide in pain and temperature regulation and in visceral sensation. The nuclei cornprizing both the auditory and brainstem visual systems were relatively devoid of I-somatostatin innervation. In the auditory system, I-somatostatin neurons were limited to the cortex of the inferior colliculus and to the ventral and dorsal trapezoid nuclei. This is in marked contrast to the distribution of another neuropeptide, delta-sleepinducing peptide, in which immunoreactive neurons were present in all components of the auditory system.18 In visual system nuclei located at brainstem levels, I-somatostatin neurons were restricted to the superior colliculus and fibers to the caudal region of the parabigeminal nucleus and the superior colliculus. Thus somatostatin, although a component of the retina where it is apparently limited to the axonless amacrine cells8 and the visual cortex and lateral geniculate bodies, 29is not associated with all aspects of visual processing. Somatostatin was not strongly associated with the cerebellum or the nuclei projecting it. Although six of the nuclei contained small populations of Isomatostatin neurons, they were absent from the griseum pontis and the inferior olive, which supply

Somatostatin in the guinea-pig brainstem

the major input to the cerebellum.‘0.42 A small number of I-somatostatin fibers and terminal8 wet-e present in the cerebellar nuclei. This sparse distribution of I-somatostatin in the cerebellum is not surpriting-to date, only a relatively few neuropepties, e.g. motilin,‘* substance pZ3.30and deltasl~p-inducing peptide’*---have been identified in the cerebellum. Thus unlike the basal gangiia,2’.30 the cerebellar outflow system does not appear to use peptides. In the forebrain, I-somatostatin neurons are associated with the majority of limbic system areasolfactory region,20 amygdala and hippocampus,’ hypothalamus. I7 In the guinea-pig brainstem, Isomatostatin neurons were confined to the periaqueductal grey. This restricted distribution of Isomatostatin neurons in the present study is different from that in the rat’7.20where I-somatostatin neurons have been reported both in the periaqueductal grey and in the interpeduncular nucleus. The raphe nuclei, besides being a major site for serotonin neurons’3.3* also contained large numbers of I-somatostatin neurons and fibers. In fact, of the three brainstem nuclei containing the most striking accumulations of immunoreactive neurons, two of these were in the medullary raphe-raphe obscurus and raphe magnus. Immunoreactive-like somatostatin neurons, in fewer numbers, were also present in nuclei superior centralis and pallidus, suggesting that the raphe nuclei, which like the periaqueductal grey, contain several different peptides20.22.Mrepresent a major concentration of different transmitters. Probably the largest concentration of Isomatostatin neurons and fibers was seen in the reticular formation, particularly at medullary levels. Of the 24 nuclei, 12 had I-somatostatin-containing neurons and all of these, except the sagulum and paranigralis, contained immunoreactive fibers. The number of immunoreactive cells and fibers was greatest in the medulla and then declined rostrally with the exception, at mid-brain levels, of a pocket of neurons in nucleus peripeduncularis. The reticular formation, with its bundles of criss-crossing fibers, is presumed to function as an integrator of many systems including sleep and arousal, consciousness, pain, temperature and visceral functions.

233

nuclei and only a few were seen in the nucleus gigantoceltdaris.” Other -differences between the two species included the lack of I-somatostatin neurons in :he following guinea-pig nuclei: locus coeruleus, cuneiform, dorsal and ventral lateral lemniscus, ambiguus and lateral reticular nucleus. In the rat, Isomatostatin neurons have been described in the nucleus parabrachialis pigmentosus but not in the nucleus compactus substantiae nigrae” whereas in the guinea-pig, the location of the immunoreactive neurons is reversed. There is only a small amount of variation in fiber distribution between the two species; the most striking examples being the almost total lack of immunoreactive fibers in the guinea-pig nuclei III, IV and VI and the only minimal input to the inferior olive. The significance of these species differences is unclear. However, investigation of the relationship of I-somatostatin association with the neurotransmitters serotonin and norepinephrine may be useful. Possible projecrions of immunoreactire-like somatostatin neurons in the brainstem

In the present study, somatostatin immunoreactivity was present in neurons and fibers throughout the brainstem. Although we have not as yet traced the projections of these neurons there have been several recent studies on the projections of brainstem nuclei containing I-somatostatin neurons.3”.6.“,‘9”‘“4”2 Thus it is possible to at least speculate on the origin of I-somatostatin terminals and the projections of specific groups of Isomatostatin neurons. Several brainstem nuclei containing large populations of I-somatostatin neurons-nuclei raphe magnus, gigantocellularis paragigantocellularis and lateral parabachialproject, via pathways containing I-somatostatin fibers, to a number of areas which contain Isomatostatin terminals. These include, in part, nucleus tractus solitarius and dorsal vagal nucleus (Figs 17 and 18); the cerebellum, via the stria acoustica; the spinal cord; motor nuclei of V, VII and XII; and midbrain, particularly the periaqueductal grey, inferior and superior colliculi and nucleus-EdingerWestphal. It is possible that some I-somatostatin neurons may have more than one projection site. For examComparative distribution of immunoreacrive-like ple, in Golgi studies, Scheibel and Scheibe135 desomafostaatin in rhe rat and guinea-pig scribed cells in the nucleus gigantocellularis of the rat To date the distribution of I-somatostatin immuhaving bifurcating axons which projected to a numnoreactivity in the brainstem has been systematically ber of areas containing I-somatostatin terminalsinvestigated only in one other species, the rat.‘7.20 magnocellular reticular formation, nucleus gracilis, Comparison of the two rodents-rat and guineaventral horn of the spinal cord, reticular formation of pig-indicates that the distribution of I-somatostatin the pans and midbrain, periaqueductal grey and neurons and fibers is similar in both. The most thalamus. striking difference between the two brains was the The possibility of overlapping I-somatostatin presence of large numbers of I-somatostatin neurons projections from several brainstem nuclei may be in the guinea-pig nuclei raphe magnus, raphe ob- highly significant. For example, the nuclei raphe scurus and gigantocellularis. In the rat, I- magnus and gigantocellularis contain different transsomatostatin neurons were absent from the raphe mitters, i.e. serotonin’3.38 and acetylcholine~28 re-

234

E. Taber-Pierce et al.

spectively. However, the I-somatostatin neurons in these nuclei appeared to form a continuum (Fig. 24), strengthening the idea of a commonality of function which was base8 on morphological features and known projections. ‘1.40.42Detailed analyses of the projections of biochemically identified neurons may be useful in providing a framework around which to organize some brainstem nuclei. Itnmunoreactive-like mitter

somatostatin

and

other

trans-

systems

sleep-inducing peptide,” enkephalin,4’ vasoactive intestinal polypeptide” and cholecystokinin’-appears to rank with the hypothalamus and amygdala as a major site for peptidergic cell bodies. Our results suggest that I-somatostatin can coexist with a number of different substances, possibly more than one in :he same nucleus or area. Based on similar evidence we have suggested the same possibility in the hippocampus.‘6 Conclusion

The very widespread distribution of I-somatostatin neurons in the guinea-pig brainstemnuclei and/or areas-suggest that the peptide is a major neurotransmitter/modulator system. Several of these areas are known to contain two or three transmitters in the same neuron.*’ Based solely on numbers of I-somatostatin neurons this is particularly likely in the raphe, 20.22~30 in some of the dopaminergic cell groups in the lower brainstem’.” and in the cholinergic neurons of the nuclei gigantocellularis, paragigantocellularis and the trapezoid body.*.** Large numbers of I-somatostatin neurons were found in the periaqueductal grey. This area, which has now been shown to contain at least five different neuropeptides in addition to I-somatostatin-substance P,” delta-

We have demonstrated an extremely widespread distribution of somatostatin-like immunoreactivity in the guinea-pig brainstem. Our analysis suggests that while somatostatin is a component of most brainstem systems, it is not an obligatory requirement. The diversity in the location of these neurons and their widespread projections suggest that whatever the role of this putative neurotransmitter/modulator, it acts upon all levels of the CNS. Acknowledgements-The

authors would like to thank Ms. Marge Pascavage for her excellent secretarial help. Sup-

ported in part by a grant from the Foundation of the University of Medicine and Dentistry of New Jersey to S. C. Feldman.

REFERENCES I. Armstrong D. M., Ross C. A., Pickel V. M., Joh T. H. aird Reis D. J. (1982) Distribution of dopamine-, noradrenaline-.

and adrenaline-containing cell bodies in the rat medulla oblongata: demonstrated by the immunocytochemical localization of catecholamine biosynthetic enzymes. J. cotnp. Neural. 212, 173-187. 2. Armstrong D. M., Saper C. B., Levey A. I., Wainer B. H. and Terry R. D. (1983) Distribution of chohnergic neurons in rat brain: demonstrated by the immunocytochemical localization of choline acetyltransferase. J. camp. Neural. 212. 53-68. 3. Basbaum A. I., Clanton C. H. and Fields H. L. (1976) Ascending projections of nucleus gigantocellularis and nucleus raphe magnus in the cat. An autoradiographic study. Anor. Rec. 184, 354 (abstr.). 4. Basbaum A. I., Clanton C. H. and Fields H. L. (1978) Three bulbospinal pathways from the rostra1 medulla of the cat: an autoradiographic study of pain modulating systems. J. camp. Neural. 178, 209-224. 5. Beitz A. J.. Shepard R. D. and Wells W. E. The periaqueductal grey-raphe magnus projection contains somatostatin, neurotensin and serotonin but not cholecystokimn. Brain Res. i61, 132-137. 6. Bobillier P., Sequin S., Petitjean F., Salvert D., Touvet M. and Touvet M. (1976) The raphe nuclei of the cat brain stem: a topographical atlas of their efferent projections as revealed by autoradiography. Brain Res. 113, 449-486. 7. Brazeau P., Vale W.. Burgus R., Ling N., Butcher M., Rivier J. and Guillemin R. (1973) Hypothalamic polypeptide that inhibits the secretion of immunomactive pituitary growth hormone. Science, N.Y. 179, 77-79. 8. Brecha W.. Karten H. J. and Schenker C. (1981) The localization of neurotensin-like and somatostatin-like immunoreactivity within amacrine cells of the retina. Neuroscience 6, 1329-1340. 9. Brodal A. (1957) The reticular formation of the brainstem. Anatomical aspects and functional correlations. In Henderson Trust Lecture. Oliver and Boyd, Edinburgh. to. Brodal A. (1981) Neurologicnl Anatomy in Relarion to Clinical Medicine. Oxford, New York. II. Brodal A.. Taber E. and Walberg F. (1960) The raphe nuclei of the brainstem in the cat. II. Efferent Connections. J. camp. Neural. 114, 239-259. 12. Brownstein M.. Arimura A.. Sato H.. Schallv A. V. and Kizer J. S. (1975) The regional distribution of somatostatin in the rat brain. Endocrinology 96, 14561461. 13. Dahlstriim A. and Fuxe K. (1964) Evidence for the existence of monoamine-containing neurons in the central nervous system. I. Demonstration of monoamines in the cell bodies of brainstem neurons. Acfa physiol. scond. 62, Suppl. 232. l-55. 14. Dodd J. and Kelly J. S. (1978) Is somatostatin an excitatory transmitter in the hippocampus? Nature 273, 674-675. IS. Fahrenkrug J. and Schaffalitzky de Muckadel 0. B. (1978) Distribution of vasoactive intestinal polypeptide (VIP) in the porcine central nervous system. J. Neurochem. 31, 1445-1451. 16. Feldman S. C., Dreyfus C. F. and Lichtenstein . . . .E. S. (1982) _ Somatostatin a_ 1,. 1. neurons in the rodent hippocampus: an in vitro and in vivo immunocytochemical study. Neurosa. Let?. .ijr LY-~4. 17. Feldman S. C., Lichtenstein E. S. and Silverman A.-J. (1979) Distribution of somatostatin in the rat brain. Am. Ass. Anat. absrr., 535. IS. Feldman S. C. and Kastin A. J. (1984) Localization of neurons containing immunoreactive delta sleep-inducing peptide

in the rat brain: an immunocytochemical

study. Neuroscience 11, 303-317.

Somatostatin in the guinea-pig brainstem

235

19. Fields H. L., Bassbaum A. I., Clanton C. H. and Anderson S. D. (1977) Nucleus raphe magnus inhibition of spinal cord dorsal horn neurons. Bruin Res. 126, 441453. 20. Finley J. C. W., Maderut J. L., Roger L. f. and Petrusz P. (1981) The immunocytochemical localization of somatostatin-containing neurons in the rat central nervous system. Neuroscience 6, 2173-2192. 21. Graybiel A. M. and Elde R. P. (1983)Somatostatin-containing immunoreactivity characterizes neurons of the nucleus reticularis thalami in cat and monkey. J. Neurosci. 3, 1308-1321. 22. Hiikfelt T., Ljungdahl A., Steinbusch H., Verhofstad A., Nilsson G., Brodin E., Pemow B. and Golstein M. (1978) Immunohistochemical evidence of substance P-like immunoreactivity in some S-hydroxytryptamine-containing neurons in the rat central nervous system. Neuroscience 3, 517-538. 23. Inagaki S., Sakanaka M., Shiosaka S., Senba E., Takagi H., Takatsuki K., Kawai Y., Matsuzaki T., Iida H., Hara Y. and Tohyama M. (1982) Experimental and immunohistochemical studies on the cerebellar substance P of the rat: localization, postnatal ontogeny and ways to the cerebellum. Neuroscience 7, 639645. 24. Inagaki S., Shiosaka S., Takatsuki K., Iida H., Sakanaka M., Senba E., Hara Y., Matsuzaki T., Kawai Y. and Tohyama M. (1982) Ontogeny of somatostatin-containing neuron system of the rat cerebellum including fiber connections: an experimental and immunohistochemical analysis. Deul Brain Res. 3, 509-527. 25. Inagaki S., Shiosaka S., Takatsuki K., Sakanaka M., Takagi H., Senba E., Matsuzaki T. and Tohyama M. (1981) Distribution of somatostatin in frog brain, Rana cutesbiana, in relation to location of catecholaminetontaining nervous system. J. camp. Neural. 202, 89-101. 26. Ioffe S., Havlicek V., Friesen H. and Chemick V. (1978) Effect of somatostatin (SRIF) and L-glutamate on neurons of the sensorimotor cortex in awake and habituated rabbits. Bruin Res. 153, 414-418. 27. Johansson O., HBkfelt T., Pemow B., Jeffcoate S. L., White N., Steinbusch H. W. M., Verhofstad A. A., Emson P. C. and Spindel E. (1981) Immunohistochemical support for three putative transmitters in one neuron: coexistence or 5_hydroxytryptamine, substance P- and thyrotropin releasing hormone-like immunoreactivity in medullary neurons projecting to the spinal cord. Neuroscience 6, 1857-1881. 28. Kimura H., McGeer P. L., Peng J. H. and McGeer E. (1981) The central cholinergic system studied by choline acetyltransferase immunohistochemistry in the cat. J. camp. Neural. 200, 151-201. 2Y. Laemle L. K., Feldman S. C. and Lichtenstein E. S. (1982) Somatostatin-like immunoreactivity in the central visual pathway of the prenatal rat. Bruin Res. 251, 365-370. 30. Liungdahl A., Hokfelt T. and Nilsson G. (1978) Distribution of substance P-like immunoreactivity in the central nervous system of the rat--I. Cell bodies and nerve terminals. Neuroscience 3, 861-943. 31. Martin R. F.. Jordon L. M. and Willis W. P. 11978) Differential oroiections of cat medullarv raahe neurons demonstrated by retrograde labeling following spinal cord lesions. J. iornj. Neurol. 182, 77-83. ’ 32. O’Donohue T. L., Beinfeld M. C., Chey W. Y., Chang T.-M., Nilaver G., Zimmerman E. A., Yajima H., Adachi H., Poth M., McDevitt R. P. and Jacobowitz D. M. (1981) Identification, characterization and distribution of motilin immunoreactivity in the rat central nervous system. Peptides 2, 467-477. 33. Renaud L. P., Martin J. B. and Brazeau P. (1975) Depressant action of TRH, LH-RH and somatostatin on activity of central neurons. Nature 255, 233-235. Saper C. B. and Loewy A. D. (1980) Efferent connections of the parabrachial nucleus in the rat. Brain Res. 197,291-317. ::: Schiebel M. E. and Schiebel A. B. (1958) Structural substrates for integrative patterns in brainstem reticular core. In Reticular Formation of the Bruin (eds Jasper H. H. and Proctor L. D.), pp. 31-55. Little Brown and Co., New York. 36. Shiosaka S., Takatsuki K., Inagaki S., Sakanaka M., Takagi H., Senba E., Matsuzaki T. and Tohyama M. (1981) Topographic atlas of somatostatin-containing neuron system in avian brain in relation to catecholamine-containing neuron systems. I. Mesencephalon, Rhombencephalon and Spinal Cord. J. camp. Neural. 202, 115-124. 37. Shiosaka S., Takatsuki K., Sakanaka M., Senba E., Kawai Y. and Tohyama M. (1981) Ontogeny of somatostatincontaining neuron system of the rat: Immunohistochemical observations. I. Lower brainstem. J. camp. Neural. 203, 173-188.

38. Steinbusch H. W. M. (1981) Distribution of serotonin-immunoreactivity in the central nervous system of the rat-cell bodies and terminals. Neuroscience 6, 557-618. 39. Stemberger L. (1979)Immunocytochemisrry. Wiley, New York. 40. Taber E., Brodal A. and Walberg F. (1960) The raphe nuclei of the brainstem in the cat. I. Normal topography and cytoarchitecture and genera1 discussion. J. camp. Neural. 114, 161-188. 41. Taber-Pierce E., Feldman S. C. and Lichtenstein E. S. (1981) Distribution of somatostatin in the guinea pig hindbrain. Sot. Neurosci. absrr. p. 898. 42. Taber-Pierce E., Hoddevik G. H. and Walberg F. (1977) The cerebellar projection from the raphe nuclei in the cat as studied with the method of retrograde transport of horseradish peroxidase. Anar. Embryo/. 152, 73-87. 43. Uhl G. R., Goodman R. R., Kuhar M. J., Childers S. R. and Snyder S. H. (1979) Immunohistochemical mapping of enkaphhn-containing cell bodies, fibers and nerve terminals in the brainstem of the rat. Brain Res. 166, 75-94.

(Accepted 2 Ocrober 1984)